https://cpw.cvlcollections.org/files/original/ef4b1bda61f8d909a07939fd61df0438.pdf 0fe204f01803d23f1032246c7e4e7c94 PDF Text Text 1 (Supplemental State of JOB PROGRESS REPORT to report of 1 January - 31 December 1987) ~C=o~l~o~r~a~d='o~ _ Project: (SE-11-2) Period Covered: 1 January - 30 June, 1988 Personnel: : Peregrine Falcon Restoration G.R. Craig, Colorado Division of Wildlife Colorado College. Program and J.H. Enderson, The The previous Job Progress Report summarized activities conducted during the 1987 breeding season (generally that period from 1 April through 31 August). Since the reporting period has been amended from a calendar to a fiscal year, work accomplished during the breeding season would be split between two reporting periods in the future. Since it is ineffective to attempt to report on progress in the midst of the field wor k , the progress for each field season will be reported at its conclusion. Hence, this interim report merely states that investigations undertaken during this period will be compiled, summarized and reported in the next (1 July 1988 to 30 June, 1989) Job Progress Report. ��3 (Supplemental JOB PROGRESS REPORT to report of 1 January - 31 December State of Colorado Project: (W-151-R-1 ) Period Covered: 1987) Bald Eagle Nest Site Protection and Enhancement Program 1 July, 1987 - 30 June, 1988 Personnel: G.R. Craig, Colorado Division of Wildlife and R.L. Knight, Colorado State University. ABSTRACT Breeding bald eagles occupied 8 territories in Colorado in 1988. One new territory was located and a pair failed to return to another site. In all, 8 young were successfully fledged by 6 pairs. Artificial nests were constructed at 2 sites t~ replace nests lost to high winds, Samples of feather pulp and blood sera were obtained from 6 nestlings to determine genetic similarity of the Colorado population to adjacent populations. This Job Progress Report represents a preliminary analysis and is subject to change. For this reason, information presented herein MAY NOT BE PUBLISHED OR QUOTED without permission of the author. ��5 BALD EAGLE NEST SITE PROTECTION AND ENHANCEMENT PROGRAM Gerald R. Craig SEGMENT OBJECTIVES 1. Monitor nest site occupancy and reproductive 2. Document survival rates and mortality 3. Determine migration 4. Determine if philopatry 5. Determine nest site tenacity by individual breeding eagles. 6. Quantify document 7. Document pesticide contamination analysis of nonviable eggs. 8. Where necessary, occupancy. and wintering success. factors. areas. occurs in breeding eagles. nesting habitats and associated foraging areas in an effort to nest site parameters conducive to improved reproduction. implement through eggshell measurement actions to stabilize nests and chemical and maintain METHODS AND MATERIALS This work will be a coopeiative endeavor Knight of Colorado State University. between the Division and Or. Richard 1. Annually visit all documented breeding sites to determine the presence of bald eagles. Pairs at territories will be documented by DWMs and other field personnel. Previously unrecorded pairs will probably be revealed in the course of aerial eagle and waterfowl flights. DWMs will confirm actual inCUbation from ground visits. 2. Occupied territories will be visited by DWMs periodically throughout breeding season to determine hatch of young, nesting failures, etc. 3. In May and June, a Utility Worker will observe breeding eagles from a distance and endeavor to follow thei r movements to locate important foraging areas. Responses of eagles to various human activities and land u~es will be recorded. 4. In June, when the young are determined to be old enough to band, sites will be visited by Craig and Knight to place a federal band on one leg and a colored, alpha numeric marker on the other. The color markers will oermit identification if the young return in subsequent years. During the same nest visit the following will be recorded: Physical parameters such as tree species, height, DBH, condition, and dominance. Nest condition, size, and location. Vegetative community and land use practices. the �6 In addition, collect prey remains, nonviable eggs and eggshell fragments. 5. When necessary, remedial actions will be taken to stabilize nests that are threatened by wind throw. Should the tree be decadent and in danger of falling, an artificial nest base may be placed in a suitable, adjacent tree. Action will be taken only after it has been deemed desirable to encourage the eagles to nest at the same location. RESULTS AND DISCUSSION Territory Occupancy Said eagle nesting activities for Colorado are summarized in Table 1. In 1988, 8 territories (Adams, La Plata ~1, Moffat #1 and #2, Montezuma ~2 and #3, and Rio Blanco 1*1 and #3) were occupied. The territory at Moffat ~3 was not checked. A new pair was located in the vicinity of Cortez and the site was designated Montezuma #3. Montezuma #3 is similar to the Archuleta site in that it is upland and located in a grove of cottonwoods that are surrounded by irrigated pasture land. The Archuleta site was vacant in 1988 and it is possible that pair may have re located e 1 sewhe re in the vic inity since a detailed su rvey was not initiated. Land Status Ownership of the terr,tories have been reported previously (Craig 1987). The recently discovered pair (Montezuma #3) occupy private property. The landowners are very protective of the_site and discourage trespass. Reproduction Reproductive efforts of the 8 pairs are summarized 1n Table 2. In summary, 8 young were fledged in 8 nesting attempts (1.0 young per attempt) by 6 successful pairs (1.33 young per successful palr). The pair at the Adams County site failed when the entire tree blew Over while they were incubating. Egg remains could not be located since the tree was standing in 2 to 3 feet of water. In an effort to encourage renesting, an artificial nest was placed in an adjacent, live tree. Although the pair took over the nest and added sticks, they did not recycle. Failure to do so may be due in part to replacement of the original male with a new mate sometime in the interim that the first tree blew down and placement of the artificial nest. It is also possible that C~lorado ~ay be situated too far north for pairs to renest. Although the pair failed to recycle and produce a second clutch of eggs, they continued to frea~2n~ t~e site throughout the summer . .u.s previously reported (Craig 1987), the nest at Moffat Co. #2 blew out killed bo th young in 1987. By October of that same year, the pair reconst ructed a suc st ant i a 1 nest nearby. In 1988 the pa ir produced successfully fledged 2 young at the new nest. Moffat Co. #3 was not checked 1988. and had and in �7 When the nest at Montezuma Co. #3 was visited after severe winds, the nest was partially destroyed and in danger of disintegrating. One nestling was present and a sing1e, nonviable egg was discovered amidst the detritus on the ground, thus at 'least 2 eggs were produced. The nest was partially stabilized with cross bracing but despite the efforts, the it fell before the nestling fledged. Fortunately, the young was old enough to maintain its balance on the bare limbs and achieved independence. In September. an artificial structure was affixed in the fork where the previous nest was situated and the nest was rebuilt. The landowner had observed the ad~lts in the area as recently as 10 days previously so it is anticipated that the eagles will adopt the new nest. Reproductive success could not be confirmed at Rio Blanco Co. #3. Barry Dupire (OWM) observed the nest from a distance and confirmed 3 young earlier in the season, but the site could not be relocated from the ground on subsequent visits. It is poss i b 1e that the nest cou 1d have b 1cwn down, but that cou 1d not be confirmed despite extensive searching. Hence, it was assumed that the site failed. Culmen length and foot pad length measurements were obtained for 6 of the eaglets. The nestlings at Moffat Co. #2 were deemed too old to approach without danger of causing premature fledging, so no data were obtained. In addition to the measurements, feather pulp ana n Iood samples were procured from the 6 nestlings. The samples will be pooled with otheF samples collected throughout the U.S. and Canada for comparison to determine the affinity of Colorado's eagles with other subpopulations. Eggshell Condition One ncnv i ab 1e egg was co 11ected from Montezuma Co. #3. The egg '..•• as broken thereby invalidating pesticlde analysis, but the shell thickness measurements will be obtained. Nest Stabilization Efforts Actions to stabilize nests were reported above. Artificial nest platforms were placed at the Adams County site and the Montezuma Co. #3 site. The Adams County platform was constructed in an effort to maintain the pair on territory, induce production of a second clutch and shift the pair away from a grove of dead cottonwoods to a substantial, living tree. The effort was successful in all objectives save production of a second clutch of eggs. It is too early to ascertain results of the reconstruction of the Montezuma Co. #3 nest. Although it was ~revicusly suggested that Moffat Co. #2 be relocated (Craig 1987), the pa i r ret.urned in the fall of 1987 and constructed a new nest on a more substantial limb. Montezuma Co. #1 was also slated for stabilization, but the work was not accomplished this period. Work should be initiated to secure the present nest and build an artificial nest in a neighboring tree as an alternate should the first nest be destroyed. �8 LITERATURE CITED Craig, G.R. 1987. Bald Eagle nest site protection and enhancement Job Prog. Rep., Colo. Div. Wildl. Res. Rep., Jan., pp1-10. - Prepared by: C .~. ~ Gerald R. era:; Wildlife Researcher C program. �Table 1. Bald Eagle Nesting Efforts in Colorado ~- -. ---_. --- -- .. _- - - --_._. - Site 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 La Plata Co. #1 t-loffatCo. #1 La Plata Co. lt2 Grand Co. Montezuma Co. ttl Rio Blanco Co. #1 Rio Blanco Co. #3 Weld Co. Montezuma Co. #2 Moffat Co. #2 Moffat Co. #3 Adams Co. Archuleta Co. Montezuma Co. #3 1egg Total Young Total Pairs IA = Inactive -------- IA IA IA IA IA lyng 2yng 2yng 2yng 1yng -- 2yng 2yng 2yng·Oyng ---- Oy.ng Oyng -- ---- ------------ 0 1 1 2 4 4 2 2 -- --- --- -- ---- --- -- -- -- -- --------- -------- --------- 4 3 1 3 -- -- -- -- ? -IA IA A ----- -- ---- -0 1 eggs ? ? ? ? ? 2yng 2yng -- 2yng Oyng lyng IA IA IA IA IA IA IA IA IA IA IA IA A A IA IA IA A 1yng 1yng ? eggs Oyng 2yng -- 3yng 2yng 2yng 2yng Oyng ---- 2yng 2yng eggs --- -- -- 2yng 1yng ---- -- -- 1yng ---- --' -- 1egg --- -- -- -- eggs ----- -- eggs 2yng 2yng IA IA IA 2yng 1yng IA lyng Oyng IA 1egg 2yng -- -- -- -- -- -- -- 3 6 4 2 2 6 4 6 5 5 10 10 9 3 lyng 1yng IA IA IA 2yng A IA lyng 2yng ? eggs IA lyng 8 8 A = Active 1.0 �10 Table Bald Eagle Nesting 2. Colorado Age of 8 i rds t·1a 1e Female Slte Adult Adult Adult Adult :.1offat Co. ~1 Adult Adult ~1offat Co. ..•,,'-') Adult Adu l; 1-1ontezuma Co. #2 Adult Adult Adu 1t Adu It Adams Co. La Plata Co. #1 Montezuma Co. ;t3 Young Produced Young Fledged 0 0 ...• ~ ,.. least 2 L Rio Blanco Co. ,,~ Adult Adult .~? ,., ,:I, i1 2 eggs produced. .-' Aault v Nest tree blown down. .-..i,.. Adult o u Comments -, Co. j;1 Tetal - 1988 '- Rio Blanco ...~ Efforts u Not relocated. � https://cpw.cvlcollections.org/files/original/a6e96618e1103fe4aa74093f987848ae.pdf 4016036c44b36a9f7efca62a633314b8 PDF Text Text 1 JOB PROGRESS REPORT State of: Project : Colorado W-152-R Work Plan: _3_: Job Title: Job _ 13 Responses of Sage Grouse to Vegetation Fertilization Period Covered: Author: ---,A~v~i~a!..!.!n~R~e s::!..!e~a~r~cc!.!h 01 January through 31 December 1987 Orrin B. Myers Personnel: C. E. Braun, Colorado Division of Wildlife; O. B. Myers, G. C. White, Colorado State University; C. Cesar, L. Upham, U. S. Bureau of Land Management ABSTRACT The response of sagebrush (Artemisia tridentata) and,sage grouse (Centrocercus urophasianus) to nitrogen fertilization was studied in North Park, Jackson County, Colorado. Sagebrush plants responded to application of 112 kg-N/ha in Fall 1985 with increased growth, increased levels of foliar crude protein, and reduced levels of foliar coumarins. Sage grouse used fertilized study plots for feeding significantly more often than adjacent unfertilized plots. When presented with 8.1. wyomingensis (ATW) and fertilized ATW (FATW) in paired choice experiments, captive sage grouse consumed significantly more fertilized than unfertilized sagebrush (ATW). Captive grouse digested dry matter, crude protein, and energy in FATW and ATW at the same rates but the nitrogen balance of birds fed FATW was significantly improved over those fed ATW. Reproductive success was measured for radio-marked hens captured near fertilized plots and compared with those not associated with fertilized study areas. Apparent differences in reproductive success among grouse could not be distinguished from any pretreatment differences. ��3 RESPONSES OF SAGE GROUSE TO VEGETATION FERTILIZATION Orrin B. Myers P. N. OBJECTIVES The project is part of a two-phased study to on sage grouse winter distribution in mining ecology, and to 2) evaluate whether nitrogen grouse winter habitat, can be used as a tool available to sage grouse. The focus of this Phase 2. 1) collect baseline information areas and on grouse feeding fertilizer, when applied to sage to mitigate reduction in habitat portion of the project is on SEGMENT OBJECTIVES 1. Document chemical and growth response of sagebrush to nitrogen fertilizer, 2. Evaluate feeding preferences of sage grouse for fertilized and unfertilized sagebrush subspecies, 3. Estimate digestibility of fertilized and unfertilized sagebrush subspecies, and 4. Monitor reproductive parameters of radio-marked' sage grouse to learn if sage grouse fitness is affected by fertilizer treatment. DESCRIPTION OF STUDY AREA The study area is in North Park, Jackson County, Colorado. The Park is an intermountain basin at an elevation of about 2,500 m. It is drained to the northwest by the North Platte River, which is fed by many smaller streams. Topography is flat to rolling with numerous ridges and benches. Climate is cold and dry with an average annual frost-free period of 46 days. Sagebrushdominated grasslands cover upland sites in the Park, and grasses and sedges occur in native and irrigated meadows that border drainages. Artemisia tridentata is the dominant sagebrush species and includes 2 subspecies, 8· 1. wyomihgensis (ATW) and 8.1. vaseyana (ATV). Other species of sagebrush occurring with limited distribution in North Park are 8. longiloba, 8. cana, and 8. argillosa. METHODS In October-November 1985 ammonium nitrate fertilizer (33.5% N) was applied at a rate of 112 kg-N/ha to 330 ha of sagebrush rangeland (Fig. 1). Thirty-three 20-ha blocks were randomly-selected with 11 blocks distributed in each Area. The northern and southern 10 ha of each block were randomly assigned as fertilized or control and treated accordingly. In October 1987 fertilizer was applied to an additional 6 I-ha plots (Fig. 1). �4 A E3 8 B c Jackson County El 1987 1985 B2] • =er t i I ized D D Contiol El -.~ 1 km Fig.1. Colorado. Locations I of fertilized study plots in North Park, Jackson County, �5 Sagebrush foliage was collected each quarter from a ra~dom sample of the 33 experimental blocks. Four blocks from each of the maln study areas were sampled, and a total of 5 randomly-located plants of the 2 big sagebrush subspecies were clipped from the control and fertilized halves of each study block. In addition, individual plants of each subspecies were examined in control and fertilized study plots for evidence of browsing by sage grouse. Sage grouse were trapped at nocturnal roosts to attach radio packages and to provide birds for captive experiments. Fertilized study areas were scanned for radio-marked birds >3 times weekly prior to and during the breeding season. Radio-marked bTrds also were located periodically to assess any use of experimental blocks and to monitor reproductive success. Captive birds were placed into metabolic cages and fed fertilized and unfertilized sagebrush in single food digestibility trials and in paired choice experiments. Feeding trials were conducted during March-April 1987 to estimate feeding preferences for fertilized and unfertilized Wyoming big sagebrush, Artemisia tridentata wyomingensis (ATW). Similar amounts of control ATW and fertilized ATW (FATW) were placed on the left or right side of each cage prior to the morning feeding period. Positions of each treatment in the cages were determined randomly. Birds were allowed to feed for 60 minutes after which the sagebrush and spillage was removed and weighed. Digestion trials were accomplished by adding known amounts of fertilized or unfertilized sagebrush to cages at about sunrise. Additional measured amounts of sagebrush were added by about 1400 hours as needed. Sagebrush branches also were placed on top of cages in the morning and weighed at night to provide an estimate of water loss during the day. After the evening feeding period sagebrush was removed from the cages and weighed. Spilled sagebrush was removed from the bottom of the cage, weighed and discarded. Fecal droppings were collected and stored overnight in resea1able plastic bags. Fecal and caecal droppings were collected and placed into plastiC bags before additional sagebrush was added to cages the following morning and subsequently placed in frozen storage. A 8 c Twenty-meter transects in 5 m segments were used to sample sagebrush plants in fertilized and control study plots during early spring prior to the onset of growth TOTAL 2 (Apr-May). Transect segments were 3.20 3.11 3.53 PLANTS/m placed at random locations along lines walked through the plots. Fig. 2. Relative densities of sagebrush Plant subspecies, height, and line subspecies on each of 3 main fertilized intercept distance were recorded for study areas in North Park, Jackson all big sagebrush plants under the County, Colorado. line transect. Densities of sagebrush plants were similar on the 3 fertilized study areas, although the relative proportions of each subspecies were not the same for each area (Fig. 2). Each plant was examined for evidence of feeding by sage grouse and 3 growth leaders, including leaves, from 1986 were measured on each plant. Sagebrush plants clipped during early spring foliage collections also were examined for evidence of feeding activity by sage grouse. Q . �6 Numbers of male and female grouse attending leks were counted between about 0500 and 0730. From 01 April through the first 2 weeks of May. Ground searches were made for new or relocated leks during the period when known leks were active. Aerial searches for leks were conducted on 12-13 May. Sagebrush foliage samples were prepared for analyses by separating leaves from stems and pooling equal amounts of leaves from each plant into composite samples. Composite samples for each subspecies collected on the control and treatment halves of each study block were ground in a mortar and pestle by freezing with liquid nitrogen. Composite foliage samples collected during feeding trials were treated in the same manner. Grouse fecal and cecal droppings collected in the course of feeding trials were freeze-dried and then ground in a mortar and pestle. Coumarin content of winter-collected sagebrush leaves was indexed using the methods of Welch and McArthur (1986). In this method, percent transmittance values are produced which are assumed to be inversely related to the coumarin content of sagebrush leaves. Measurements of percent transmittance was made in triplicate on ATW and ATV plants. Dry matter content of foliage and grouse droppings was determined after drying overnight at 100 C. Samples were ashed overnight at 500 C to determine organic matter and ash content. Kjeldahl nitrogen (Horwitz 1980) was measured in addition to neutral detergent fiber, neutral detergent solubles, and acid detergent fiber (Van Soest 1963~, Q; Mould and Robbins 1981, Van Soest 1982). Soil samples were collected on 13 October from each of the 6 study plots to receive nitrogen fertilizer in Fall 1987. Adjacent control plots also were sampled. Each plot was subsampled at 6 random locations. Subsamples were placed into a plastic bucket and thoroughly mixed before removing about 750 cc for analyses. Analyses of soil samples were performed by the Soil Testing Laboratory at Colorado State University. AREA C -.~ I 33 34 4 3 1 km 10 On 05 August a wildfire burned' ·[Il BU"'n SIte ~ existIng approximately 75 ha of sagebrush in the southern half of S9 T8N R78W. The burn Fig. 3. Location of wildfire in destroyed approximately 80% of experimental block #16 in Area C (Fig. 3) relation to fertilized study blocks about 1.5 km west-southwest of the Perdiz in North Park, Jackson County, Lek. Colorado. PlOts RESULTS AND DISCUSSION The most obvious response by sagebrush to fertilization was increased growth (Table 1). New growth on nearly 400 plants was measured in March and April 1987, to determine the amount that growth of leaders was increased in �7 the first year of response. For this parameter the 2 dominant subspecies, ATV and ATW, responded similarly to the fertilizer treatment. New growth from fertilized plots was more than twice as long as that from unfertilized plots (Table 1). The size of individual leaves and the number of seed heads also appeared to have been increased by fertilization. Table 1. Length of annual growth increment (cm) of fertilized (F) and unfertilized (C) sagebrush plants. Subspecies Treatment n x SE C 118 96 68 101 3.0 7.1 2.9 7.3 0.1 0.2 0.2 0.3 8. 1· vaseyana F 8. 1· wyomingensis C F ANOVA. Source df MS .E value E. >.E Subspecies Treatment Subspecies x Treatment Error 1 1 1 379 0.21 1,665.04 1.18 4.30 .0.06 401. 41 0.27 0.80 0.0001 0.60 The secondary chemistry of sagebrush plants was influenced both by subspecies involved and fertil ization (Table 2). Coumarin level s measured by percent transmittance values were lower in ATW than in ATV, and coumarins were lower in fertilized plants than in unfertilized plants (Table 2). The reduction of coumarins in fertilized plants conforms to predictions by Bryant et al. (1983) that plant investments in chemical defenses are a function of the plant's carbon/nutrient balance. With increased nutrient availability, plants would be expected either to increase their investment in defensive substances or to increase growth without a proportional increase in defensive substances. If other defensive substances follow coumarin levels, then it appears that sagebrush plants adopt the strategy of maximizing growth rather than investing in defense compounds. The result of relatively lower levels of anti-herbivore defense chemistry in fertilized sagebrush could be that such plants contain higher levels of available nutrients. Levels of crude protein in winter sagebrush foliage were significantly influenced by fertilization (Table 3, Fig. 4). ATW samples contained more crude protein than ATV (E. = 0.0001) and each subspecies showed about the same proportional response to fertilization (E. > 0.9). There was considerable block-to-block variability in vegetation response. In Area B the terrain is more variable and contains more soil types than the other areas. If Area B was considered separately there was no increase in protein levels (E. > 0.05). �8 Table 2. Relative levels of coumarins in fertilized (F) and unfertilized (C) sagebrush plants as measured by percent transmittance. Subspecies Treatment 8· 1. vase~ana 8· 1. w~omingensis 4.4 4.2 2.5 3.5 32.4 44.3 38.4 61.8 15 15 15 15 C F C F SE X D. ANOVA Source df MS E value £ > Subspecies Treatment Subspecies x Treatment Error 1 1 1 56 2070.93 4,679.90 497.08 208.04 9.95 22.50 2.39 0.003 0.0001 0.13 E Table 3. ANOVA for effect of nitrogen fertilization on crude protein levels in winter sagebrush foliage. Source df Block Subspecies Treatment Subspecies x Treatment Error 11 1 1 1 33 MS 0.0003 0.0102 0.0023 0.0000 0.0001 E value £ 3.89 111.39 24.77 0.01 0.0012 0.0001 0.0001 0.9 > E Observations of feeding activities of free-ranging grouse and of captive birds indicated that grouse preferred fertilized sagebrush over unfertilized sagebrush. The number of transects in which feeding activity by grouse was found was greater on fertilized plots than on adjacent control plots (Fig. 5) (£ < 0.01). Twenty-eight percent of the transects on fertilized plots contained plants that had been fed upon by grouse compared to 3% on the unfertilized plots. Grouse did not appear to increase their use of ATV. Maintenance of captive birds was difficult. Birds did not consume enough sagebrush to maintain mass and did not acclimate behaviorally to captivity well enough to allow periodic weighings. Mealworms and a gamebird ration were added to cages in addition to sagebrush on days when digestive trials were not being conducted. Birds soon learned to feed upon mealworms contained in 500 ml beakers but did not consume any of the gamebird ration in the beakers. The �9 14~--------------------------' c=J Control 1:: :1 Fertilized 2 o ATV ATW Fig. 4. Crude protein levels in winter sagebrush foliage samples from fertilized study blocks in North Park, Jackson County, Colorado. combination of sagebrush and mealworms appeared to arrest declines in mass shown by the birds for a time, however some birds lost mass even with the meal worm supplement (Fig. 6). ~ m ~An "ATI E t ~ ao w ~ Captive bird studies began 7 February when a yearling female (Band # 5845) was confined. Additional birds were captured 5 and confined (Table 4). Several preliminary digestion and preference 0 r c r c trials were conducted prior to 1 April when 4 birds wer~ available for trials. us APi-HAY About 750 g of ATW and FATW were put into Fig. 5. Use of sagebrush plants by cages for the digestion trails. Larger grouse on fertilized (F) and amounts were added on days when trials unfertilized (C) study plots in were not conducted. North Park, Jackson County, Sage grouse selection for fertilized and Colorado. unfertilized ATW also was studied by allowing captive grouse to choose between FATW and ATW plants. Sage grouse consumed an average of 2.1 times more FATW than ATW (Table 5), and all birds in all trials consumed more FATW than ATW. Each treatment was intended to occur in the left and right cage positions with equal frequency but problems with birds and an incorrect treatment assignment prevented this arrangement. Unequal representation of treatments in the cage positions could have confounded feeding preference with a learned propensity to feed at a 15 10 �10 Table 4. Captive sage grouse used in feeding trials, 1987. Band Age Date captured 5845 Yrlg 5 Feb 4000 5848 5851 5852 5854 Ad Yrlg Yrlg Yrlg Yrlg 18 11 21 21 21 5858 Yrlg 31 Mar # Comments Released 2/22/87 with radio (151.089); radio recovered after raptor mortality on 3/11/87. Died in captivity 2/21/87. Died in captivity 3/27/87. Released 4/13/87. Released 4/13/87. Released 4/13/87 with radio (151.050); radio active as of 7/8/87. Released 4/13/87 with radio (150.989); radio signal last received 4/18/87. Feb Mar Mar Mar Mar particular cage position. An additional analysis was performed by partioning the data set and using observations such that treatments were distributed evenly among birds and cage positions, thereby removing any confounding of treatment and cage position. Analysis of this data set indicated that an average of 2.6 times more FATW than ATW was consumed by grouse (P < 0.005). Although birds consumed significantly more fertilized than unfertilized sagebrush in the choice experiments, the intake of dry matter by captive grouse was not affected by fertilization when a single food was offered (£ = 0.82). Daily consumption of dry matter averaged 51 g (range = 28 - 82 g) for ATW and 48 g (range = 27 - 68 g) for FATW. The daily intake of nitrogen was similar (.~AT\J = 0 -. 9 g, SEAT~ = 0.08; KFAT\J = 1.2 g, SEFAT\J = 0.08; £ =0.14) for the sagebrush treatments, Dut the amount of nitrogen retained by birds was significantly affected by fertilization (Table 6). This parameter also was affected by among bird differences (£ = 0.02). Three birds behaved Similarly and produced consistent results compared to a fourth bird, which had consumption rates that consistently were lower than the others. Table 5. Consumption (g) of fertilized and unfertilized Wyoming big sagebrush by captive sage grouse during one-hour preference trials in March - April 1987. Treatment n Unfert ilized Fertil ized 15 15 5.8 12.1 111 SE £ > 1.1 1.2 0.001 �11 1.5 • 1.4 , ,, , r=-; b1 ~ 1.3 • '-/ U) U) .<:C ::c >-< Q 1 .1 --. -, -, <, 1.0 0 iII ...._ 1.2 A 0.9 0.8 0 100 90 , .... E-< ::r:: H 8 6 ,, ~ Q 4 _ ......:1 12 14 16 18 10 12 14 16 18 ._ ..•.... 80 .•.. W ::;:: -< !:z:; 10 •,, •••........ <:» 2 -, 70 H Q H 60 B 0:: 0 50 0 2 4 6 8 DAYS IN CAPTIVITY Fig. 6. Absolute (A) and proportional sage grouse hens used in captive feeding (8) losses of body mass of 2 captive experiments. �1,2 Table 6. ANOVA for effects of experimental fertilization on nitrogen balance of captive sage grouse. Source df Bird Treatment Bird x treatment Laboratory rep. Error 3 1 3 1 31 MS .E value 0.686 1.072 0.122 0.013 0.185 3.71 5.81 0.66 0.07 £ > .E 0.02 0.02 0.6 0.8 True metabolizable energy values (TME) cannot be obtained from birds without estimates of excreted fecal metabolic energy and urinary endogenous energy. These estimates can be obtained by Sibbald's (1979) TME assay using fasted birds. Apparent metabolizable energy (AME) values can be obtained from the conditions of this study by comparing energy intake and energy retention and correcting energy retention to a zero nitrogen balance (Sibbald 1979). The slope of the line relating energy intake and the corrected energy retention provides a means for estimating AME. Analysis of covariance can be used to test for differences in slopes to estimate treatment effects. ATW and FATW treatments were retained by birds at the same rates (Fig. 7, Table 7) and treatments alone did not affect the amounts of energy retained (£ = 0.8). 300, • FATI 250 " - C' B :J< u 200 - •••• ,. ~ w aJ 150 f- .. 0 w :=: -c 100 I- •" JIll ~ ~ •• " An' • • .. 50 I- 0 0 100 200 300 "00 500 INTAt::E (KCtl I) Fig. 7. Apparent metabolizable energy adjusted to zero nitrogen balance of ATW and FATW fed to captive sage grouse hens. �! ._ _j 13 Table 7. Analysis of covariance for effects of fert il izat ion treatments amounts of energy retained from sagebrush. df Source Energy intake (kcal) Treatment Intake x treatment Error MS 54618.71 19.6 7.6 424.0 1 1 1 16 £ value £ > £ 133.07 0.05 0.02 0.0001 0.8 0.9 on The apparent digestibility of dry matter also did not differ with respect to fertilization (Fig. 8) (£ = 0.30). FATW had a mean dry matter digestibility of 52.5% compared to 48.9% for ATW, but all birds did not retain the same proportion of ingested dry matter (£ = 0.05). Sage grouse trapping began in the first quarter of 1987 and continued through mid-May. Initial trapping success was poor but success rates improved later in the season (Fig. 9). The first bird was captured on 6 February. A total of 41 female sage grouse was outfitted with radios by the conclusion of the trapping season (Table 8, Appendix A). Yearling birds comprised the bulk of the radio-marked sample (73%). 60 HALES FEl1ALES 50 ~ ~ p:: I=> CJ D 40 E-< p." <: u 30 U) ~ p:: H ~ 20 ZI 10 1 2 FEB Fig. 9. 3 4 5 6 7 B 9 HAR WEEK-LONG TRAPPING 10 11 12 13 14 15 16 APR PERIODS Numbers of sage grouse captured during Spring 1987. HAY �14 SO r- • •• ATW ,.. 40 IFATW r\ 0 30 I- w Z '" •• '" « I;j a: ..• ,.. • 0> u JIll!.. 20 • l- 6 ,.. 10 l- a 20 • • • '" • ",'" • 30 40 CM 70 60 SO I NTAKE 80 90 (g) Fig. 8. Retained dry matter in relation to intake by captive sage grouse fed fertilized (FATW) and unfertilized (ATW) sagebrush. Table 8. Numbers of adult and juvenile female sage grouse fitted with solirpowered radio transmitter packages in North Park, Colorado, 1987. 1- 2+ n Area Fertil ized Lake John Totals n % % Totals 1 4 33 20 14 16 66 80 21 20 11 27 30 73 41 Total body mass of birds and weights of radio packages were similar from fertilized and unfertilized study areas (Appendix B). Radio packages were < 2% of total body mass. �15 Nesting success was higher among birds captured near the fertilized study areas than those captured in the unfertilized area near Lake John (E = 0.04) (Table 9). However, it is not know to what extent if any the fertilizer treatment had in producing this effect. As birds were marked in Spring 1987, the fertilized study areas were intensively searched for radio-marked birds. Three birds (of 21 trapped in the area) were within 1 km of fertilized plots. One of these (freq. 151.191) used the plots on the McCallum Oil Field (Area A) before breeding and then disappearing after a failed nesting attempt. Two other birds were captured during the laying interval in the fertilized study area south of JC 12, along JC 23A. Both birds successfully hatched eggs in nests on about 1 June. One bird (freq. 150.065) nested about 10 m from the fertilized plot at the western end of what is now the Perdiz Burn (Area C). This bird moved to meadow areas northwest of the nest site and disappeared about 2 weeks post hatch. The other bird in Area C (freq. 151.111) nested in the first fertilized plot south of JC 12. This bird hatched 6 eggs and spent much of the summer and fall within about 400 m of the nest site. Table 9. Nesting success (n successful nests/n birds in sample) of radiomarked sage grouse in North Park, Colorado, 1987. The data includes several birds for which reproductive status was not determined because of failure to detect radio signals. Treatment Adults Fertilized Lake John 4/6 W4 6/13 4/15 10/19 (52.6%) 4/19 (21.0%) Overall 4/10 10/28 14/38 (36.8%) Yearlings Totals Clutch sizes could be determined accurately for 9 hens after chicks left nests (Appendix A) and ranged between 6 and 7 eggs/clutch. Clutch size was not correlated with distance to fertilized plots (Rz = 0.05, E = 0.54) (Fig 10). One bird was located while renesting with a clutch of 5 eggs. I was unable to reliably determine the nesting success or clutch sizes of several birds (Appendix A). Reasons for not locating birds include radio failures, mortality, and bird movements outside the range of ground-based radiotracking.' Subsequent observations indicate that several birds moved outside the study area into Wyoming. Therefore, several "missing" birds could have nested successfully. By including only those birds with complete histories through the nesting season, the proportion of successfully nesting hens was 9/17 for birds from the fertilized areas, 5/13 for birds from the unfertilized area (X = 0.62, E = 0.43), and about 47% overall. Therefore, it is unclear if fertilization improved the nesting success of marked or unmarked birds in the northeast portion of North Park. No new leks were identified near the fertilized study areas in 1987. One small lek (Ram) was inactive and another (Hawk) became active with a peak attendance of 8 males and 9 females. Other leks in the northeast portion of Jackson County most closely associated with the fertilized study areas �16 / ) ·-~"'r' 8 I U 5 .....J .A. 1- • 2+ 7 U Z in 8w 6 .A. zl .A. • .A. 5 o 10 20 o I STANCE 30 Renest 40 (km) Fig. 10. Clutch size of 2 age classes of radio-marked sage grouse hens in relation to distance from fertilized study plots. (Perdiz, Turkey, and Denmark) did not show increased numbers of males/lek (Fig 11) . Results of analyses on soils were received and summarized. The control and treatment plots did not differ prior to fertilization in any of the parameters measured (Appendix C). Two soil series were represented in the plots fertilized in Fall 1987. The Morset loam series is a dominant soil type in Area C and composes a substantial portion of Area B. This series contained higher levels of organic matter, potassium, and copper than Bosler loam (Appendix C). Bosler loam is a major soil series in the Area A study area. Bosler loam contained more zinc than Morset loam. Browse transects indicated that sage grouse used the fertilized plots for feeding more than they use adjacent control plots; a this finding supported by the captive grouse experiments. The hypothesis that nitrogen fertilization will improve grouse habitat to the extent that reproductive success or survival will be improved is based on work with red grouse (Lagopus lagopus scoticus) (e.g., Miller 1968, Miller et al. 1970). Red grouse are strongly territorial and monogamous, whereas sage grouse are not. Also, sage grouse do not carryover large energy reserves to the breeding season (Remington and Braun 1988), thus, exogenous nutrient sources are likely more important to egg and chick quality than endogenous sources. Because red grouse are more predictable in space and time than sage grouse, it may be possible to treat specific areas and measure responses. Moreover, red grouse are more likely to �l7 1985 1986 1987 -r;;.~,. /" ~ ....-'. .. ~~~~~~~~--~~-~~~-"~"~~_=- -~ .. Cc~~~~·_- ~ Fig. 11. Peak counts of male sage grouse on strutting grounds in North Park, Jackson County, Colorado. Leks most closely associated with fertilizer treatments are in the northeast portion of North Park. �18 remain in the treatment area (unless the treatment is deleterious or the bird disappears for other reasons that are independent of the treatment) so that any cumulative effects of the treatment on egg/chick quality or adult and chick survival can be realized. In this study, explicit sage grouse nesting areas could not be identified for fertilizer treatment and radio-marked hens apparently did not preferentially use fertilized areas. Thus, there was no discernable experimental manipulation of the radio-marked birds. An alternative approach to conducting the field portion of the fertilization project could have concentrated trapping at 3-4 leks having relatively large expanses of adjacent nesting habitat. In a pre-fertilization field season, intensive sampling of hens at treatment leks could be done and their nesting areas identified. Fertilizer could be applied the subsequent fall to the nesting areas identified. Birds with active radios could be monitored throughout the winter to document use of treatment areas prior to fertilizer response. In spring, intensive trapping could be repeated at treatment leks. In this way a sample of grouse using fertilized areas could be obtained. The effect of different soil types on fertilizer response could be studied by applying fertilizer to much smaller plots distributed among soil types. A revised trapping procedure also may have been useful to overcome negative bias for hens of the spotlighting technique (Giesen et al. 1982). For example, during 1979 and 1980 male grouse were captured at a rate of 1.98/hr compared to 0.86/hr for females (Giesen et al. 1982) from the North Park population, which is assumed to have about twice as many females as males (Braun 1986). Perhaps by using cannon nets or drop nets during hen attendance peaks, large numbers of hens may be captured at treatment leks with disturbance potentially limited to a single trapping event. This method also might offer some assurance that radio-marked birds would nest in the vicinity of treatment leks (Braun et al. 1977). LITERATURE CITED Braun, C. E. 1986. Responses of sage grouse to vegetation fertilization. Job Prog. Rep., Colorado Div. Wildl. Fed. Aid Proj. W-37-R-38. Pp. 115-141. , T. Britt, and R. o. Wallestad. 1977. Guidelines for maintenance of ------sage grouse habitats. Wildl. Soc. Bull. 5:99-106. Bryant, J. P., F. S. Chapin, III, and D. R. Klein. 1983. Carbon/nutrient balance of boreal plants in relation to vertebrate herbivory. Oikos 40:357-368. Giesen, K. M., T. J. Schoenberg, and C. E. Braun. 1982. Methods for trapping sage grouse in Colorado. Wildl. Soc. Bull. 10:224-231. Horwitz, W., ed. 1980. Official methods of analysis of the Association of Official Analytical Chemists. Assoc. Off. Anal. Chern., Washington, D. C. 1018pp. Miller, G. R. 1968. Evidence for selective feeding on fertilized plots by red grouse, hares, and rabbits. J. Wildl. Manage. 32:849-853. �19 _____ , A. Watson, and D. Jenkins. 1970. Response of red grouse populations to experimental improvement of their food. Pages 323-335 in A. Watson, ed. Animal populations in relation to their food resources. Brit. Ecol. Soc. Symp. 10. Blackwell Scientific Publ., Oxford and Edinburgh. Mould, E. D., and C. T. Robbins. 1981. Evaluation of detergent analysis in estimating nutritional value of browse. J. Wildl. Manage. 45:937947. Remington, T. E., and C. E. Braun. 1988. Carcass composition and energy reserves of sage grouse during winter. Condor 90:15-19. Sibbald, I. R. 1979. Metabolizable energy evaluation of poultry diets. Pages 35-49 in W. Haresign and D. Lewis, eds. Recent Advances in Animal Nutrition--1979. Butterworth's, London, U.K. Van Soest, P. J. 1963~. Use of detergents in the analysis of fibrous feeds. I. Preparation of fiber residues of low nitrogen content. J. Assoc. Off. Agric. Chern. 46:825-829. _____ . 1963h. Use of detergents in the analysis of fibrous 'feeds. II. A rapid method for the determination of fiber and lignin. J. Assoc. Off. Agric. Chern. 46:829-835. _____ . 1982. Nutritional ecology of the ruminant. 0 and B Books, Inc., Corvallis, Ore. 374pp. Welch, B. L., and E. D. McArthur. 1986. Wintering mule deer preference for 21 accessions of big sagebrush. Great Basin Nat. 46:281-286. Prepared by: Approved by: ~~ Orrin B. MYS Graduate Research Assistant ~2' ~ Clait E. Braun Wildlife Research Leader �20 Appendix A. Nesting success of female sage grouse fitted with radio transmitters during winter-spring 1987 and status of radio packages as of 30 June 1987. Band Treatment # Age Fertilized 5895 5894 5899 5893 5896 5727 5721 5722 5849 5857 5858 5728 5898 5720 5892 5846 5856 5887 5855 5719 5763 5900 5743 5742 5847 5854 5882 5724 5890 5886 5744 5897 5885 5725 5881 5891 5723 5883 5889 5761 5884 Yrlg Yrlg Ad Yrlg Yrlg Yrlg Yrlg Yrlg Ad Yrlg Yrlg Ad Ad Ad Ad Ad Ad Yrlg Yrlg Yrlg Yrlg Yrlg Yrlg Yrlg Ad Yrlg Ad Yrlg Yrlg Yrlg Yrlg Yrlg Yrlg Ad Yrlg Yrlg Yrlg Ad Yrlg Yrlg Yrlg Control S Nesting success Clutch size fail ed ? {fail ed)" ? successful successful 6 7 ? successful failed successful successful 4+ ? (5+)b 7 ? successful successful fail ed successful fail ed successful fail ed successful failed fail ed successful 6 7 6 6 ? (3+) 7 ? ? (failed) ? (failed) ? ? successful successful ? (failed) ? ? ? ? ? fail ed successful 7 ? fail ed fail ed renest ? (fail ed) ? Presumed failed. b Number of chicks observed with hen. 5 Radio frequency 150.024 150.084 150.096 150.104 150.124 150.145 150.165 150.185 150.874 150.983 150.989 151.028 151. 034 151.107 151.111 151.191 151.232 151. 266 151. 308 151.373 151. 432 150.045 150.225 150.671 150.936 151. 051 151. 073 151. 088 151. 099 151.142 151.173 151. 206 151.212 151.251 151. 283 151.313 151. 326 151.344 151. 404 151. 423 151.447 Radio status active active recovered active active recovered active missing active active missing active active recovered active missing active active active active active active missing active active active active active active active active missing active active active missing active active active active missing �21 Appendix B. Sage grouse body mass and weight of radio packages from fertilized and unfertilized (Lake John) study areas in North Park, Colorado, 1987. 111 Treatment n x SE Body mass (g) Unfertilized Fertil ized 20 21 1525 1546 26.3 17.2 0.5229 Radio weight (g) Unfert ilized Fertil ized 14 20 29.5 25.7 1.5 1.3 0.0691 Radio/body (%) Unfert ilized Fertil ized 13 19 1.95 1.68 1.01 0.86 £ > �N N Appendix C-l. Soil characteristics of experimental plots when treated with ammonium nitrate fertilizer in October 1987. Condo (mmhosLcm) 6 6.8 0.18 6 6.7 0.15 > > 0.5 > 0.4 n Control Treatment £ > 111 pH OM (%) N03-N P Zn K Fe Mn Cu (ggm) 3.1 0.6 4.2 231 0.8 2.9 0.7 3.6 205 0.8 0.5 > 0.5 > 0.4 > 0.5 > 0.6 23.9 3.0 2.2 2.1 24.0 2.9 > 0.8 > 0.8 > 0.8 Appendix C-2. Soil characteristics of experimental plots when treated with ammonium nitrate fertilizer in October 1987. Soil series Bosler loam Morset loam E > 111 Condo (mmhosLcm) 0.10 4 6.7 0.20 8 6.8 > 0.5 < 0.01 < n pH OM N~-N P K 2.1 3.4 0.01 0.5 0.7 0.08 Zn Fe Mn Cu (ggm) (%) 3.5 140 4.2 257 0.4 < 0.01 0.9 0.7 0.02 1.8 17.9 3.1 2.3 27.0 2.9 0.10 > 0.5 < 0.01 �23 JOB FINAL REPORT Colorado State of: Project: W-152-R (01-03-045) Avian Research 15 : Job --- Work Plan: 3 Job Title: Sage Grouse Resource Exploitation and Endogenous Reserves in Colorado Period Covered: Author: 1 January 1984 through 31 December 1987 Jerry W. Hupp Personnel: D. A. Hein, D. J. Nash, R. A. Ryder, Colorado State University; M. Blymeyer, J. Capodice , T. Reed, Bureau of Land Management: J. Barry, C. Molitoris, W. Shuster, W. Wallis, E. Zieroth, U.S. Forest Service; C. E. Braun, C. Coghill, T. Henry, J. Houston, P. Mason, T. Sherrill, Colorado Division of Wildlife ABSTRACT Winter habitat use and foraging ecology of sage grouse (Centrocercus urophasianus) were studied in the Gunnison Basin, Colorado between 1985 and 1986. Sage grouse foraging activity (N = 157 feeding sites) was not proportionally distrubted (p < 0.001) among topographic features. Topographic distribution of feeding activity was influenced by physiographic variation in shrub structure of mountain big sagebrush (Artemisia tridentata vaseyana) relative to snow depth. Most foraging (45-64% of feeding sites) occurred in drainages and on southwest slopes where sagebrush exposure above snow was maximized. Sage grouse rarely foraged (1-4% of feeding sites) on northeast aspects with slopes >50 because exposed sagebrush was not widely available. Distribution of foraging was not influenced by topographic variation in crude protein or monoterpene concentrations of mountain big sagebrush. Within feeding sites, sage grouse did not selectively forage on plants with high crude protein or low monoterpene concentrations. Sagebrush structural characteristics at 87 winter feeding sites were compared to 100 random locations. Sagebrush structural measures were not useful to identify winter habitat in the mesic terrains where sagebrush removal is most likely to occur. Spring lipid reserves of adult male sage grouse were determined from carcass analysis of 96 individuals collected from 2 Colorado populations between 1983 and 1985. Spring lipid reserves were affected by winter severity. Males depleted lipid reserves during the courtship season. Males in Jackson County mobilized 125-130 g of fat during courtship while males in Gunnison County used an average of 66 g. Strutting display rates of adult males were quantified in Jackson and Gunnison counties in 1986. Slower display rates, lack of evening display, greater variance in male attendance at leks suggest �24 reduced energetic investment in courtship among males in Gunnison County. However, behavioral differences could not be attributed to unequal size of endogenous reserves as lipid deposits (N = 10 adult males/population) during early courtship were similar between populations in 1986. Lipid catabolism likely provides <10% of adult male energetic needs during courtship. Lipids may primarily be mobilized during early courtship when male displays are most vigorous due to the presence of females on leks, and when male reproductive success is determined. �25 RECOMMENDATIONS Habitat Management Recommendations 1. Sagebrush in drainages throughout the Gunnison Basin should be maintained as winter habitat. Treatment of sagebrush in drainages would have a disproportionately severe affect on winter habitat availability. In harsh winters when snow depths may exceed 50 cm, exposed sagebrush is primarily available in drainage sites. Removal of sagebrush at those sites would restrict sage grouse to other terrains where sagebrush forage could be buried by snow; reduced survival due to greater exposure to winter weather and predators is possible. There should be no treatment of sagebrush in drainages to insure that exposed sagebrush is available during severe winters. 2. Sagebrush on >50 southwest terrains as well as 0-50 high sites should not be treated. Southwest slopes are frequently used as winter foraging sites by sage grouse and sagebrush on that terrain should be maintained. Sagebrush canopy cover is sparse «20%) on southwest slopes and 0-50 high sites. Due to the xeric soils on those sites, livestock forage response would be minimal following sagebrush removal and would likely not justify treatment. No treatment of sagebrush on sites with less than 15-20% canopy has been recommended by other authors (U.S. Dep. Agric. 1969, Braun et ale 1977). 3. Limited treatment of 0-50 low terrains and slopes with northeast aspects would likely have minimal impact on sage grouse winter habitat. Sagebrush removal on those sites should not be complete. Live sagebrush should be, interspersed with treated areas to insure some shrub availability for sage grouse and other sagebrush obligates. Herbicide treatments in strips 30-50 m wide are preferable. Treated strips should be alternated with untreated strips in a 1:1 ratio. Repeated herbicide applications should not occur until recovery of sagebrush in treated strips is complete and structure is similar to adjacent untreated strips. Prescribed burning should not be conducted on 0-50 low sites or northeast slopes if fire cannot be contained within those terrains, or if removal of>50% of sagebrush within the treated terrain is likely. No treatments of 0-50 low or northeast slopes should occur if breeding habitats will be affected. 4. There should be no treatment of sagebrush within 1 km of any active sage grouse strutting ground in the Gunnison Basin. Maintenance of sagebrush within 1 km of leks insures that diurnal habitats required by male sage grouse during spring courtship will be available. 5. Breeding activity is not widely distributed in the Gunnison Basin. Areas within 3 km of Ohio, Tomichi, Quartz, and Razor Creek drainages may be important breeding areas due to the availability of irrigated hay meadows as summer habitats. Breeding activity in areas south and west of Gunnison may be limited because mesic summer habitats are not widely available. To insure that nesting habitats are available in the Gunnison Basin, there should be no treatment of sagebrush within 3 km of any lek that is near «3 km) floodplain boundaries of the Ohio, Tomichi, Quartz, or Razor Creek drainages. Maintenance of undisturbed sagebrush within 3 km of leks is consistent with recommendations suggested by the Western States Sage Grouse �26 Committee (Braun et al. 1977, Autenrieth 1981), and will help insure adequate availability of nest sites. 6. There should be no sagebrush treatment within 3 km of other major sage grouse leks in the Gunnison Basin that are >3 km from the Ohio, Tomichi, Quartz, or Razor Creek drainages. Major leks are those at which >20 males were observed displaying during at least 1 year between 1984 and 1986. These leks include Sapinero 1, Sapinero 3, and Upper South Parlin strutting grounds. 7. There should be no sagebrush treatment within 200 m of riparian areas with permanent streams, boundaries of irrigated hay meadows, spring developments, or livestock ponds. During summer months sage grouse use sagebrush habitats adjacent to these moist soil sites. Maintenance of sagebrush cover near moist soil areas may be necessary to sustain summer use of those sites. 8. Summer habitats in areas south and west of Gunnison could be improved by increasing availability of moist soil sites. Maintaining water flow in stream drainages during summer months would be beneficial. Placement of gabion structures and check dams could improve surface water availability in intermittent streams, increase soil moisture, and improve herbaceous growth in riparian areas. Removal of sagebrush to stimulate herbaceous growth is not a valid technique to improve brood habitat if it impacts nesting or wintering habitats. INTERAGENCY CONSULTATION 1. The CDOW should be responsible for monitoring the sage grouse population in the Gunnison Basin. District Wildlife Managers should conduct annual surveys of male sage grouse lek attendance on known leks. Lek surveys should be conducted between 1 April and 15 May and should begin as quickly as access to strutting grounds is feasible. Ideally, each lek should be surveyed a minimum of 3 times during the display season to obtain a representative evaluation of male display activity. Multiple visits are especially important in the Gunnison Basin due to the high degree of day to day variation in numbers of males that attend leks (Hupp 1987). Lek counts may not provide a reliable measure of annual fluctuations in the Gunnison Basin sage grouse population. However they do give an indication of the distribution of breeding activity in the Gunnison Basin. Long term changes in numbers of males that attend a specific lek may indicate habitat related changes in breeding activity in local areas. Results of annual lek surveys should be summarized, recorded, and forwarded to local offices of the BLM, USFS, and SCS. 2. CDOW personnel should continue to search for undiscovered or newly formed leks throughout the Gunnison Basin. This study suggests that little breeding activity occurs in areas southwest of Gunnison and north of Blue Mesa Reservoir. Searches should continue to be conducted within those areas to confirm whether breeding activity is limited. Searches for leks would be most productive in early April because males remain on leks for longer periods during the early display season (J. W. Hupp, unpubl. data). Searches conducted in late April and May would be less likely to �27 locate leks because males display for shorter periods during morning hours. Lek locations should be confirmed during multiple visits in the year following discovery. Locations of discovered leks should be marked on 7.5-minute topographic maps upon confirmation. Copies of maps with discovered leks as well as numbers of observed males should be forwarded to local offices of the BLM, USFS, and SCS. 3. CDOW personnel should continue to monitor the sage grouse harvest in the Gunnison Basin through collection of wings from hunter-harvested birds. Wing barrels have proven to be an efficient technique of collecting large samples of wings. Placement of wing barrels prior to the opening of the sage grouse season should continue. Examination of wings provides information on age and sex structure of the fall population. While population trends cannot be evaluated from these data, age and sex ratios do provide information on reproductive success during the previous breeding season. Reproductive success can affect rates of population growth, and should be monitored. Collection of wings at wing barrels also provides information on distribution of hunting success in the Gunnison Basin as well as data regarding the timing of harvest. Reports of hunter harvest and sex and age structure of the sage grouse fall population should be reported to local offices of the BLM, USFS, and SCS. 4. Federal agencies should provide the CDOW with information regarding proposed sagebrush treatments. Information supplied to the CDOW should include (1) a 7.5-minute topograhic map with boundaries of the treatment area clearly defined, (2) information on the method and season of the proposed treatment, and (3) locations of previous sagebrush treatments that have been completed during the previous 10 years within 1 km of boundaries of the proposed project. The CDOW should be notified immediately when a treatment proposal is developed within a federal agency. Advanced notification should allow CDOW District Wildlife Managers and federal agency wildlife biologists adequate time to search for sage grouse leks within 3 km of boundaries of the proposed treatment during the breeding season (1 Apr to 15 May) prior to implementation of the project. Treatment boundaries should be adjusted or the treatment postponed for a year if any sage grouse display activity is observed within 3 km of the project boundaries. Numbers of males that attend discovered leks near proposed treatments should be assessed during mUltiple visits ~3) in the year following discovery. If the discovered lek is (1) within 3 km of the Ohio, Tomichi, Quarta, or Razor Creek floodplains, or (2) attended by large numbers (~20) of males, then the planned treatment should be cancelled, or the boundaries adjusted so that no treatment occurs within 3 km. If the numbers of males observed on leks in the year following discovery is small ~20) and the lek is not within 3 km of the Ohio, Tomichi, Quartz, or Razor Creek floodplains, then treatment to within 1 km may be permissible. 5. Definitions of potential climax plant communities on mountain loam and dry mountain loam sites should be re-evaluated to determine the validity of sagebrush removal to improve "range condition." Federal agencies should evaluate pre- and post-treatment vegetative composition at sagebrush treatment sites. Canopy cover or biomass of shrubs (by species), grass, and forbs should be estimated at sample sites within treatment areas. Sampling should be of sufficient intensity to construct adequate �28 confidence intervals on pre- and post-treatment vegetation composition, and to draw valid inference conclusions on treatment effects within the project area. Post-treatment sampling should occur periodically during 10-15 years to evaluate long-term effects of treatment and to evaluate recovery time of shrubs. Ideally, monitoring should occur concurrently on ~ nearby untreated control area. Reports of treatment effects on plant species composition should be provided to the CDOW. �29 ACKNOWLEDGMENTS This study was supported Additional Bureau technical and Colorado landowners encouragement develop was a constant Nash comments source Service, program, deal of assistance of information by the U. S. U. S. Forest access to ranchers to private for his support scientist on my graduate natural Groshek, cheerful lands and for challenging with my academic and expertise. and and and professional. committee during My deepest The support of Mike thanks management and Cliff Reynolds and friendship, A special resource to this project. assistance conditions. support were provided for providing my doctoral of Wildlife. me to Dr. Ronald program and Drs. Dale Hein and provided and many useful on my dissertation. contributed Barbara Division I am grateful to Dr. Clait Braun a great served Numerous Basin thanks Basin fully as a wildlife provided Donald State University. throughout more support Soil Conservation in the Gunnison My special Ryder and logistical of Land Management, Service, by the Colorado My sincere friendship, and for sharing management. The assistance his under for his unselfish insights and friendship difficult this study. acknowledged. assistance on wildlife and conservation of District and for his in suggesting is gratefully Griffin, competent, also to Joe Capodice and Terry Reed to Jim Houston to Robert of field research and for his initiative Blymeyer thanks in the Gunnison for their dedicated, long hours appreciation professionals Wildlife and Managers �30 Cliff Coghill, Tom Henry, Phil Mason, and Tom Sherrill appreciated. My thanks also to Jim Olterman and Rick Sherman for their support and interest. Wallis, Jim Barry, Carol Molitoris, stages of the study. allowed me to use Soil Conservation I am deeply grateful for sharing regarding sagebrush Rick Hooley, and Tom Washburn aspects of laboratory Chapman provided consultation Diane Hall provided I am especially Department excellent grateful secretarial their discussions Biology of biological assistance analyses. Jann Black and assistance. in the Biology for their friendship issues. and for Thanks also to Fishery and to make my doctoral experience. Finally, my deepest appreciation Hupp for their unconditional I have worked analyses. David Bowden and Phil faculty and staff for helping a rewarding laboratory t~ my fellow graduate students of Fishery and Wildlife and Thanks also to Joan also provided analysis. on statistical and Lynn Stevens chemical analysis, support. and Carol Ann We inland for conducting with various program Service snow mobiles. a great deal of laboratory Rick Bettinger, support Ken Lair and John Scott graciously to Don Dick, Tom Remington, their knowledge for providing Wildlife Bill Shuster, Bill and Elaine Zieroth of the U. S. Forest Service provided during various Demander is deeply to my parents, ~ayne and Betty love and support throughout towards my college degrees. the years that �31 TABLE OF CONTENTS ABSTRACT OF DISSERTATION iii ACKNOWLEDGMENTS -.......................... v LIST OF TABLES. .................................................... ix LIST OF FIGURES Chapter 1. xi SAGE GROUSE WINTER HABITAT USE AND FORAGING IN THE GUNNISON BASIN, COLORADO ECOLOGY Introduction. ................................................. 1 Study Area. ................................................... 3 Methods. ...................................................... 7 Topographic Distribution 7 Shrub Structure 10 Sagebrush 11 Chemical Analysis Results 14 Snow Depth Variation Topographic Sagebrush 14 Distribution Composition of Feeding Sites and Structure Chemical Variation Among Random Collection Sites 24 Foliage Chemical Differences Between Browsed Unbrowsed Sagebrush Plants Discussion 16 29 and 31 31 Sage Grouse Habitat Use 31 Sage Grouse Forage Selection 37 Management Implications 41 �32 Literature Chapter 2. Cited 43 ENDOGENOUS RESERVES AND COURTSHIP OF ADULT MALE SAGE GROUSE BEHAVIOR Introduction " Methods 48 49 Phase 1: Male Endogenous Reserves During Spring Courtship. ...................................... 49 Phase 2: Endogenous Reserves and Courtship Behavior ..... 52 Results 55 Phase 1: Male Endogenous Reserves During Spring Courtship.............................. Phase 2: Endogenous Reserves and Courtship ......... 55 Behavior ..... 58 Discussion 62 Sage Grouse Body Mass 62 Lipid Depostion 62 and Winter Severity Lipid Reserve Use During Courtship Male Display Rates a-q_dLipid Reserve The Adaptive Significance Literature Cited 76 Size of Male Lipid Reserves 65 67 69 �33 LIST OF TABLES 1.1. 1.2. 1.3. 1.4. 1.5. 1.6. 2.1. 2.2. 2.3. 2.4. Categories of slope and aspect used to classify terrain in randomly-selected sage grouse survey plots, Gunnison Basin, Colorado -....................... Gas chromatograph programming specifications used separate and quantify monoterpenes in Artemisia tridentata vaseyana and b. nova foliage from the Gunnison Basin, Colorado 8 to 14 Observed and expected distribution of sage grouse feeding sites among terrain categories in the Gunnison Basin, Colorado, 1985-86....................... 21 Big sagebrush structural characteristics at random and sage grouse feeding sites in the Gunnison Basin, Colorado, 1985-86....................................... 26 Chemical characteristics of big sagebru~h leaves collected at 32 random sites in 4 terrain categories (8 sites/category), Gunnison Basin, Colorado, March April 1986 -................................. 30 Chemical characteristics of leaves of browsed and unbrowsed big sagebrush plants at sage grouse winter feeding sites (N = 20), Gunnison Basin, Colorado, January -March 1986..................................... 32 Differences in sage grouse physical characteristics among years and between courtship periods (2-way ANOVA), Gunnison and Jackson counties, Colorado, 1983-85 . 56 Physical characteristics of adult male sage grouse during courtship in Gunnison and Jackson counties, Colorado, 1983- 85 . 57 Peak daily attendance of adult male sage grouse at 3 leks in Gunnison and Jackson counties, Colorado. Surveys were conducted during 5 mornings on each lek, April-May 1986 . 61 Winter (Nov-Mar) snowfall and mean temperature departure from 30-year average in Gunnison and Jackson counties, Colorado, 1983-86. Lipid reserves of adult male sage grouse during early courtship following winter . 63 �34 LIST OF FIGURES Figure 1.1. Gunnison 1.2. Mean winter (1 Jan - 15 Mar) snow depths within terrain categories in the Gunnison Basin, Colorado, 1985-1986............................................... 15 Distribution of l-km random survey plots in the Gunnison Basin, Colorado, 1985. Solid circles indicate survey plots with sage grouse feeding sites................................................... 17 Distribution of I-km random survey plots in the Gunnison Basin, Colorado, 1986. Solid circles indicate survey plots with sage grouse feeding sites................................................... 19 Observed and expected proportions of sage grouse winter feeding sites within terrain categories in the Gunnison Basin, Colorado, 1985-86. Confidence limits (95%) for observed use ·were calculated following Byers et al. (1984). Failure of an expected proportion to fall within a confidence limit indicates that feeding activity was different (E < 0.05) than expected.. ......................................... 22 Mean sagebrush height at feeding and random locations relative to snow depth within terrain categories in the Gunnison Basin, Colorado, 1985-86 34 2.1. Colorado 50 2.2. Strutting display rates (R ± SE.) of adult male sage grouse in Gunnison and Jackson counties, Colorado, April-May 1986.......................................... 1.3. 1.4. 1.5. 1.6. Basin, study Colorado................................ areas.................. .................. 4 59 �35 CHAPTER 1 SAGE GROUSE WINTER HABITAT USE AND FORAGING IN THE GUNNISON BASIN, COLORADO ECOLOGY INTRODUCTION Sage grouse dependence documented 1975). (Patterson Sagebrush wildlife To develop guidelines emphasized Previous measurement Beck 1977, Connelly Colorado assessments dependent on sagebrush of shrub structure (e.g., sagebrush 1982). More recently, in sagebrush have height, at use sites (Eng and Schladwieler 1982, Schoenberg for snow may limit shrub of sage grouse winter habitat based on taxonomic variation 1972, Remington in northern crude protein and content. Quantitative useful are of particular (1985) assessed winter feeding site selection monoterpene 1967, Higby for habitat maintenance, Winter habitats sage grouse are completely canopy cover, and density) and Braun to improve grass (Schneegas forage and cover during that season, and because availability. et al. have described biotic and abiotic characteristics of areas occupied by sage grouse. interest because 1972, Wallestad from attempts impact sage grouse habitat 1975). biologists for forage and cover is well 1952, Eng and Schladweiler removal resulting forage may severely 1969, Wallestad upon sagebrush descriptions in development of sagebrush of guidelines shrub structure for maintenance have been of winter habitat �36 (Braun et al. 1977, Autenrieth structure criteria vegetation applied in habitat evaluation sampling. to describe Other assessment species composition is available. topographic terrains. 1982). feeding physiographic distributed sagebrush structure, structural 1972, Beck in snow depth, or foliage chemistry. of sage grouse winter of southern Colorado between 1985 that sage grouse feeding sites were topographic or foliage chemistry of foraging activity. at winter variation among available physiographic was to evaluate whether structure topographic charact~ristics objective distribution in areas studies of sage grouse shrub structure, I tested the hypothesis proportionally (Winward and (Eng and Schladweiler However, previous sites in the Gunnison-Basin and 1986. features and Sage grouse use of sites with destinctive species composition, I evaluated structure of sage grouse winter habitat winter habitat have not quantified sagebrush criteria may be more easily Because sagebrush features has been observed 1977, Schoenberg intensive criteria could be useful to assess and distribution with heterogeneous use of shrub often requires are affected by physiographic 1977), topographic availability However, areas of sage grouse winter use and identify sites where winter habitat Tisdale et al. 1982). variation influenced I measured features. My in snow depth, topographic sagebrush shrub feeding and random sites to learn if distinctive characteristics of feeding areas existed. whether sage grouse selectively protein or low mono terpene concentrations I also evaluated foraged on sagebrush plants with high within feeding sites. �37 STUDY AREA The Gunnison (Fig. 1.1). Basin is an intermontane Elevation basin in southern Colorado in the study area is between 2,300 and 2,900 m. The Basin is semi-arid with a mean annual precipitation Gunnison. winter Approximately snowfall temperatures Oceanic 48% of the annual precipitation (Hunter and Spears 1975)-. Winters between -13 and 1 C from January and Atmospheric of 28 cm at Admin. 1986). occurs as are cold with mean through March (Natl. Snow cover typically persists through March. The Gunnison alluvial Basin is topographically diverse. flood plains are adjacent to major streams. plains occur at the mouths of tributary drainages. Steep-sloped are highly dissected by permanent irrigated and intermittent Shallow, eroded gulches are common on upland slopes. Soils and vegetation Alluvial Uplands are mesas with broad, flat tops occur in several areas of the Uplands streams. Level outwash to steeply rolling with slopes ranging from 5 to 300. moderately Basin. Broad (> 2 km) in the Gunnison Basin are also diverse. soils in flood plains are poorly drained and are used as grass hay meadows and pasture. well-drained soils derived from weathered (Hunter and Spears 1975). Upland sites are xeric with igneous and metamorphic Surface horizons rock of upland soils are fine grained or gravelly and stony loams. Mountain big sagebrush was the dominant upland vegetation. big sagebrush (~. ~. tridentata) wyomingensis) were not found in the Gunnison taxonomic classification classification and Wyoming big sagebrush Basin. follows Winward and Tisdale of big sagebrush herbarium Basin (~. ~. Big sagebrush (1977). Taxonomic specimens was confirmed by �38 0 C <t 0: 0 ...J 0 0 • ~ •••• rJ) .•. cr w •... ~ "'w -~-' NO o i g~ §~ S {J ~ ~. ~I\ ,0· Illl~ • I L ----- cr. _ �39 the U. S. Forest Service Shrub Science Laboratory pers. commun.). Growth forms of big sagebrush were highly variable and dependent on xeric south slopes was short cover < 20%), while sagebrush and vigorous (canopy cover> common on xeric ridgetops occurred were interspersed (Chrysothamnus dominated Quaking « 35 cm) and widely 30%). Black sagebrush and south slopes. and mountain with sagebrush viscidiflorus) snowberry occurred distributed bottlebrush above 2,900 m. Douglas (A. ~) bitterbrush oreophilus) rabbitbrush shrub that often had been removed. included western wheatgrass squirreltail (Pseudotseuga and Sandberg bluegrass (Eriogonum spp.), cinquefoil (Agropyron prairie in smithii), (Stipa comata), junegrass (Poa sandbergii). (Castilleja (Potentilla lupine (Lupinus argenteus), grasses needle and thread (Sitanion hystrix), forbs included paintbrush spp.). was (Symphoricarpos Common understory understory (Astragalus Antelope and Douglas-fir (Festuca arizonica), (Koeleria cristata), silvery tall (> 50 cm) (A. ~) was a common understory overstory (canopy on isolated mesic sites below 2,900 m, and were communities fescue spaced Silver sagebrush at some sites. Basin Big sagebrush on mesic sites was usually aspen (Populus tremuloidies) sagebrush Arizona on site conditions. sites where the sagebrush menziesii) widely in the Gunnison at a few mesic locations below 2,900 m. (Purshia tridentata) (E. D. McArthur, Common spp.), buckwheat spp.), Phlox spp., common fleabane (Erigeron Plant names follow Scott and Wasser spp.) and vetch (1980). �40 METHODS Topographic Distribution Ground searches between 3 January chosen throughout were established 7.5-minute winter for sage grouse feeding sites were conducted and 15 March in 1985 and 1986. Random points were the Gunnison plots around each point. topographic and searched rugged maps. (l-km diameter) of plots were drawn on Circular plots were visited features characteristics topography Boundaries on foot for feeding sites. the field by comparing terrain Basin and circular illustrated Plots were located in on topographic in the area to be surveyed. of the Gunnison once each Basin, boundaries maps with Because of the of survey sites were easily discerned. Terrain within each random survey plot was classified aspect (Table 1.1). level to verify categories Where necessary I used a compass slope aspect ~nd steepness. were drawn on topographic on foot for feeding sites. obtained at 5 points spaced at 50-m intervals category. detecting terrain Ideally, transects terrain Measures of snow depth were along transects However, of terrain categories, Therefore, of for each due to the high maintenance of straight Also, feeding sites in one topographic were often discovered type. assessed in each to do so were made using line (Burnham et al. 1982) in 1985. lines was not feasible. of terrain search effort and the likelihood Initial attempts degree of interspersion category Boundaries use sites would have been independently category. and an Abney maps and each terrain type was searched terrain by slope and when an observer was in a different it was not possible to measure search effort �41 Table 1.1 Categories of slope and aspect used to classify terrain in randomly selected sage grouse survey plots, Gunnison Basin, Colorado. Terrain 1. a - category Site description 50 slope, Areas of < 50 slope,broad low topography outwash plains adjacent topography,terraces 2. a - 50 slope, high topography Drainage to areas of higher on slopes. Areas of < 50 slope,broad ridge tops topographically surrounding 3. (> 100 m) flood or Narrow « (> 100 m) mesa or higher than terrain. 100 m) flood plains of permanent intermittent streams, shallow eroded gulches on slopes. 4. > 50 slope, Slopes> 50 with south or west aspectsa. Slopes> 50 with north or east aspectsb. southwest 5. > 50 slope, northeast aAspects with bearings of 136 to 3150. bAspects with bearings of 316 to 1350. and �42 independently searched for each terrain category. such that all terrain categories Feeding in snow. Track observation throughout flushed. Therefore, site terrain features were classified categories (Table 1.1). overlapped 2 terrain categories. On a few occasions, was assigned to the terrain take place. Because 1 foraging observations, terrain discovered feeding feeding large feeding sites « the observation activity appeared it was possible to To improve independence of site was not included in the analysis to of 200 m) to a previously site in the same plot. Discovery of adjacent sites rarely occurred. I used chi square analysis site distribution availability. field surveys pooled to 1 of 5 possible in which most foraging if it was close to At each sage grouse When that occurred, area within a plot. and then and it was possible feeding sites were discrete, a feeding distribution lande4 at a site, foraged, of most feeding areas. snow Track observations feeding sites were discrete the boundaries as well as tracks of use because each winter period. that sage grouse usually observe> were sampled. was a valid criterion indicated feeding each survey plot was sites were located from flock sightings cover was present identify Instead, among terrains Although was made between distribution Expected 7 terrain categories were identified categories 6-150 and> to during as either 6-150 or > 150), I in some terrain categories due to bias associated in some terrain that feeding in 1985 and 1986 was proportional (slopes were classified observations was necessary to test the hypothesis during analysis. with small numbers (Alldredge and Ratti 1986). This of observations No distinction 150 slopes during analysis and among 5 terrain categories use was based on a uniform (Table 1.1) was tested. distribution of feeding activity �43 across all terrain categories within survey plots where feeding activity was observed. not considered categories in calculation was estimated (95%) were calculated each terrain category. proportions differed Terrain categories in unused survey plots were of availability. from topographic Availability maps. Confidence for the observed proportions Confidence of terrain limits of total use within intervals were compared to expected to evaluate whether observed use of a terrain category from expected use (Byers et a1. 1984). Shrub Structure Shrub structure was measured sites. at a sample of sage grouse feeding Sites selected for measurement where approximate boundaries structure was not measured of the use area could be discerned. snow, or where feeding activity was not A 15-m tape was placed in a north-south center of the foraging area. transect was recorded direction Species of each shrub beneath and sagebrush was classified measurements shrubs beneath included axis of crown, (4) height the tape were measured. (1) length of transect (3) crown width perpendicular Canopy dimensions Individual intercept, shrub (2) longest to the longest axis, and Live sagebrush plants within 0.5 m of the 15-m transect were counted to obtain an estimate of density. crossed the from ground level to the tallest stem (excluding inflorescenses). dimensions over the as living or dead. Dead shrubs had living foliage on < 20% of stems. of individual Shrub at sites where sage grouse tracks were obscured by melted or windblown apparent. of shrub structure were those and density were measured Shrub along a 2nd, ls-m transect the center of the 1st line at a perpendicular angle. that �44 Shrub structure Gunnison Basin. was also measured Random sites were stratified in each of 5 terrain catagories within (Table 1.1). with 20 locations sage grouse winter distribution transects sites. were oriented Measurement followed of shrub structure the same procedures Summary statistics Summary values that was intercepted height of live big sagebrush (3) the standard proportion were selected assessment Connelly assumptions among variables plant height as a for height), (density). they have previously and These variables been used in (Eng and Schladweiler was examined. 1972, I used analysis of no shrub structure and between differences random and feeding sites. by Ko1omogorov-Smirnov of normality (mean 1982). to test hypotheses Sagebrush p1ants/m2 (2) the mean the transects of variation of sage grouse winter habitat were evaluated (canopy cover), of big sagebrush (coefficient for analysis because categories, at random sites for each use and random plot. plants beneath deviation 1982, Schoenberg variance at random used at sage grouse feeding sites. of live sagebrush Correlation Two, 15-m directions and density by transects of mean height (4) the number use. (1) the total length of live, big sagebrush canopy height), east-west were calculated included survey plots used to and habitat in north-south, sampled Sites were selected a random sample of the same 1-km diameter evaluate terrain at 100 random sites in the distribution of among All data tests to insure that were met (Conover 1980:357). Chemical Analysis Sagebrush for analysis foliage was collected of crude protein at 20 winter and monoterpene sites were located while searching feeding sites in 1986 content. l-km diameter Winter survey plots. feeding �45 Sagebrush foliage was collected their discovery. boundaries Sagebrush at winter feeding sites at the time of foliage was collected of the feeding area could be discerned. shrub structure measurements were obtained in the Gunnison an adequate Basin were often narrow sample of browsed plants Therefore, Sage grouse feeding sites S m) and linear. a perpendicular angle. shrub canopy dimensions The 2nd line crossed Snow cover beneath along the axis the birds the center of the 1st at transects was removed, and density were measured. each transect were examined for evidence woody material. leaves. from sagebrush browsed and unbrowsed these were not available sample was individually sealed in a plastic bag and frozen. samples were also obtained during March and April 1986. each of 4 terrain categories (0-50 low, 0-50 plant. plants that were totally covered by snow because Sagebrush if ~ 2 Leafy stems were collected from several areas of the crown of each browsed Samples were not obtained by the dark, exposed Plants were considered leaves had been clipped by sage grouse. plants and do not conswne Browsed plants were distinguished of clipped Sagebrush and of feeding activity. Sage grouse feed only on the leaves of sagebrush occurred To obtain (N ~ 10) at a feeding site, it was 1 transect was oriented traveled while feeding. mesophyll samples and to sample sagebrush plants along the path of feeding activity. beneath « Foliage along 2, lS-m transects located near the center of the feeding area. necessary only if approximate to sage grouse. at 32 randomly Sagebrush was collected located Each sites from 8 sites in where sage grouse foraging primarily high, drainage, > 50 southwest). Sites were located in the same l-km plots used to evaluate winter distribution habitat use. were obtained Sagebrush foliage and shrub measurements and �46 along 2, lS-m transects east-west). Foliage I obtained and unbrowsed feeding site. samples. site. bagged plants for each sagebrush Equal amounts of leaves (1 g minimum) and frozen. by grinding facilitate fragmentation Monoterpenes were extracted ether through apparatus. ether. prior to grinding to loss of monoterpenes. dripping in flasks and saved. was added and the volume' increased Approximately Liquid for 6 hours in a Soxhlet The extract was captured (0.625ug/ul) from each random from samples by continuously 5-g subsamples in pooled leaves in a mortar. immediately and to prevent a were picked from plants so that all plants were equally represented was added to the mortar standard (north-south, species within Pooled samples of leaves were also prepared nitrogen diethyl directions samples were individually Samples were homogenized diethy1 in cardinal a 15-30 g pooled sample of leaves from each group of browsed individual oriented 20-ml subsamples A carvone to 100 ml with of extract were retained for analysis. I used a Perkin-Elmer ionization Extract detector samples Individual to separate with a hydrogen and quantify monoterpenes (4 ul) were carried monoterpenes Identification gas chromatograph flame (Table 1.2). through a 30-m tubular column. were identified was made by comparison using mass spectrum to published analysis. spectral patterns (Epstein et al. 1976, Heller and Milne 1978). Approximately samples 5 g of homogenized for analysis unbrowsed plants. matter percentages microkjeldahl multiplied of crude protein leaves were retained in random site, browsed, Samples were freeze-dried were determined. procedures from pooled Nitrogen (Horowitz 1980). by 6.25 to obtain an estimate to remove moisture and and dry content was evaluated Nitrogen content was of crude protein. by �47 Table 1.2. Gas chromatograph programming specifications used to separate and quantify monoterpenes in Artemisia tridentata vaseyana ~. ~ foliage from the Gunnison Basin, Colorado. Carrier gas Helium Flow rate 1.5 ml/min Initial temperature 70 C Initial rate of temperature increase 130 C Final rate of temperature increase 20 C/min Post temperature The hypotheses 300 C of no differences and total monoterpene tests dependent crude protein, at feeding analysis in mean crude protein, pooled individual among random sites in different of variance upon normality. and total and individual not differ between plants contents were tested with analysis nonparametric (approximate) 2 C/min Final temperature categories and or Kruskal-Wallis The hypotheses monoterpene samples of browsed terrain that concentrations and unbrowsed did sagebrush sites were tested with ~ tests or Mann-Whitney dependent upon normality. RESULTS Snow Depth Variation In both years snow depths increased and decreased low sites, in March and> (Fig. 1.2). 50 northeast between January Snow was deepest aspects. and February, in drainages, Snow was shallow on > 50 0-50 �48 1985 - 40 o • 5 NORTI-£AST ~ ~ 30 :c ra... ~ 20 ~ o z 10 Cf) o • 5 °SOUTHWEST JAN FEB MAR 1986 50 -t5 40 :c ._ .._ ..-.._--_ .._-_ ..-..,.., /«0<""- ------- --; r- a... 30 :>:: -~:~_~::>'" UJ ....... , o ~ o z o • 5 NORn-EAST ~ 20 ",", 0-5°LOW ........... ". DRANAGE Cf) -. O-t HGH o • 5 SOUTHWEST o JAN FEB MAR Fig. 1.2. Mean winter (1 Jan - 15 Mar) snow depths within terrain categories in the Gunnison Basin, Colorado, 1985-86. �49 southwest January on 0-50 high sites. slopes while depths were intermediate - March precipitation averages for the Gunnison in 1985 and 1986 was similar Basin (Natl. Oceanic to 30-year and Atmospheric Admin. 1986). Topographic Distribution Ground searches of Feeding Sites for sage grouse feeding sites were conducted in 142 and 120, l-km diameter survey plots in 1985 (Fig. 1.3) and 1986 (Fig. 1.4), respectively. I observed 74 winter feeding sites within used survey plots in 1985 and 83 feeding sites within plots in 1986. Analysis of sage grouse distribution features was not accurate not provide combined if flock sightings similar estimates track and flock sightings There was no difference similar estimates observations Feeding often foraged respectively), although - for analysis Thus, each sighting of topographic in either winter in drainages (Fig. 1.5). track and flock was valid. (l < 0.001) distributed (Table 1.3). A high proportion on use was proportional sites) provided Sage grouse (23-33% of feeding sites in 1985 and 1986, sites also occurred of feeding of track and flock distribution even though that terrain category low sites was greater of analysis. criterion use, and combining did I the distributions distributions activity was not proportionally the survey plots feeding and track observations and compared in topographic of topographic among terrain categories among topographic among terrains with chi-square (0.25 > l > 0.10). sightings 52 used survey of use among terrain categories. 1985 and 1986 observations 51 comprised only 3-4% of (23-42%) of sage grouse > 50 southwest slopes in each year, to availability (Fig. 1.5). Use of 0-50 (37% of feeding sites) in 1985 than in 1986 (11% (Fig. 1.5). Feeding activity (11-16% of feeding �Vl o ~ A\ o 1 2 3 .• 5 KILOMETERS 0 • 0 I 0 0 0 0 0 • • • • 0 0 . • • • 0 0 - • • 0 Oorl'':.I."t o Fig. 1.3. Distribution of l-km random survey plots in the Gunnison Basin, Colorado, 1985. circles indicate survey plots with sage grouse feeding sites. Solid �.N' ~ 012345 KILOMETERS 0 ~-\\ ~ '\ o~ tr_ 0 I 0 • 0 • • =v-:» • • • • Parlin • 0 0 0 0 ~ 0 • o ~ 0 0 0 Dorle~1I e • o ~ Fig. 1.4. Distribution of l-km random survey plots in the Gunnison circles indicate survey plots with sage grouse feeding sites. Basin, Colorado, 1986. Solid \.Jl I-' �52 Table 1.3. Distribution of sage grouse feeding categories in the Gunnison Basin, 1985-86. sites among terrain 1986 1985 Terrain Observed Expected Observed Expected o - 50 low 27 17.2 9 15.9 o - 50 high 12 14.1 9 13.0 Drainage 17 3.0 27 3.9 > 50 southwest 17 16.1 35 25.6 > 50 northeast 1 23.6 3 24.7 Total X2 92.9 fa < 0.001 aprobability expected 74 that the observed distribution. 83 163.5 < 0.001 distribution is similar to the �53 0- 5°LOW 0.5 z 00.4 I- ex: ~96" t=~ 0.3 00.3 Q. o ex: 0.2 Lower 95" Q. coo1\dence .,.. 0.1 0.1 1985 o 1986 z o z 00.4 0.4 lCC 0.3 i= ~ 0.3 Q. 1 00.2 40.19 16 0.1 0. 1985 o. z )5 0.16 8: 0 23 0. O~f 0.33 0.1 ·00.05 00.04 1985 1986 1986 ° SOUTHWEST 1 0.3 0.2 0.2 0.5 ~ 0.4 g: q; 0.11 o Q. o 0.5 0 ex: 0 a. DRAINAGE 0.5 0.5 ex: a. Expected jX'oportion 0.42 )5 ° NORTHEAST Z 0.4 0 0.31 i= CC 0.3 0 Q. 0 . 0.22 00.32 00.30 0.2 CC a. 0.1 0.04 0.01+ 1985 1986 1985 1986 Fig. 1.5. Observed and expected proportions of sage grouse winter feeding sites within terrain categories in the Gunnison Basin, Colorado, 1985-86. Confidence limits (95%) for observed use were calculated following Byers et al. (1984). Failure of an expected proportion to fall within a confidence limit indicates that feeding activity was different (f < 0.05) than expected. �54 sites) on 0-50 high sites was proportional I observed aspects little (1-4%) feeding activity even though that terrain category survey plots Sagebrush sagebrush sites in different were apparent terrain categories. It was the dominant it was present. 50 southwest and northeast black sagebrush sagebrush (R - 35 cm, SE - 1.8). The closest variables provided of big sagebrush described provided stands. coefficient of variation in between with each other. canopy cover and (£ indicated weak correlations variable measured of overhead Mean plant height was a measure The coefficient structural < 0.2) different Canopy cover was a measure of sagebrush Of the 4 sagebrush was rarely present correlated in plant size relative an evaluation on some terrain categories. by big sagebrush. plant size. variation black (R - 20 cm, SE - 1.0) than big Black sagebrush Each structural of sagebrush concealment However, was also present were not strongly Other comparisons among variables. dominated On sites where both species (£ - 0.56) occurred association mean height. sites. was shorter the mesic 0-50 low and drainage Structural selected species at 9 of 13 0-50 high sagebrush Black sagebrush occurred, components Big sagebrush in (present at 13 of 20 random sites) on xeric 0-50 was abundant sites where 30-31% of the among randomly 50% of shrub canopy cover) most sites. high sites. xeric> comprised Among Terrain- Categories.--Differences species composition sagebrush on > 50 slopes with northeast and Structure Site Variation (comprised> (Fig. 1.5). (Fig. 1.5). Composition Random to availability of variation to the mean. for height Density dispersion. variables analyzed, for height was similar only the (f - 0.71) among random �55 sites in different differences terrains (Table 1.4). in canopy cover, mean height, random sites were rejected Big sagebrush 53 cm) than plants (R 32%) was greater and sagebrush in other terrain categories, 50 northeast than plants on 0-50 high sites and> canopy cover at 0-50 low sites to drainages. 50 southwest (R - southwest and 0-50 high sites sagebrush canopy cover in those catgegories respectively) relative Sagebrush density drainages and> Comparison did not occur. 50 southwest slopes. sites was sparse 29 and 31 cm, respectively). was sparse similar estimates Big (R - 16% and 12% to density Between Years.--Feeding at sites where foliage was collected line transect orientation measurements that different of shrub structure structure was calculated regardless of transect in sites. replicate sites indicated (R - on > 50 as well as at sites where collection Although feeding Big sagebrush was shortest of Feeding Site Shrub Structure 2 groups of samples, I measured but taller than on 0-50 high sites was high relative analysis, (R sites were shorter to canopy cover in other terrain categories. site shrub structure was measured for chemical Big (R - 29%) and similar to was vigorous Big sagebrush were taller and canopy cover in drainages, stands. Canopy cover on > SO northeast 22%) relative among other than 0-50 low. than in all categories 43 and 42, cm respectively) drainage density (f < 0.006, Table 1.4). plants on 0-50 low and> sagebrush of no topographic plants at random sites in drainages (R sagebrush Hypotheses differed between of structure 0.2). Feeding provided site shrub from all sites where shrubs were measured, orientation. shrub structure 1985 and 1986, respectively. at 37 and 50 winter Shrub structure feeding differences the at 9 winter transect orientations (f> of foliage sites in between �56 Table 1.4. Big sagebrush structural characteristics at random and sage grouse winter feeding sites in the Gunnison Basin, Colorado, 1985-86. 0-50 low 0-50 high Random 20 20 Feeding 25 10 Variable N > 50 SW > 50 NE 20 20 20 21 28 3 Drainage sites cmb Canopy, Random Feeding 1,078±55 367±54 959±83 478±66 965±94 1,127±62 1,007±76 < 0.001 Mean height, 636±82 9l5±107 < 0.001 0.11 e cm Random 43.3±2.9 3l. 6±2.l 52.8±1.9 28.8±2.5 41.5±3.5 Feeding 43. 9±l. 7 44.4±3.4 54.l±2.9 41.l±l.6 51.6±3.7 0.86 0.003 of variation Random 29.6±1.5 27.6±2.8 31. 2±1. 6 29.9±2.2 27.8±2.l Feeding 31. O±1. 0 34.5±2.3 31. O±1. 8 32.6±1.1 28.0±3.l 0.56 for height, % O.ll 0.94 0.23 0.71 plants/m2 Random 1. 3±0.1 2.0±0.2 1. 3±0.1 1.4±0.2 1.4±0.l Feeding 1. 8±0. 2 1.4±0 .1 f 1. 2±0.1 1.0±0.1 0.024 0.059 aprobability among terrain bMean canopy 0.001 < 0.001 0.72 Coefficient Density, 0.001 0.049 that mean values of structural variable are similar intercept. Percent categories. total length of big sagebrush cover - mean total length 0.006 = 30. canopy �57 Table 1.4. Cont. CMean ± 1 standard error. dprobability that mean values do not differ between feeding and random sites. eNo comparision northeast (R - site density differed between winters; 1.8±0.2) differed 1986 feeding site density density made in > 50 due to small sample of feeding sites. fFeeding density of feeding and random site structure (E - 0.56). 1985 feeding site from random site density (E - 0.006); (R - 1.3±0.1) was similar to random site �58 winters were evaluated observations drainage, within each terrain category were obtained in both winters > 50 southwest). Sagebrush terrain categories sites was higher exception feeding site shrub structure variation associated category (e.g., drainage Shrub structure data for comparison features, measured differed cm). random sites). to random Sagebrush Within the> height - 29 cm). density (R - 1.8 p1ants/m2) (R at 50 southwest Sagebrush big were canopy cover - 29%, category, R sage canopy cover (R - 33%) and (R - 41 cm) than occurred at random locations canopy cover (R - (R canopy 32%) and (R ~ 44 cm) at feeding sites in the 0-50 high category were also than at random locations Comparison sites. (R canopy cover - 12%, of feeding and random sites in the> terrain was not possible feeding the same terrain In the 0;50 low terrain category, at feeding than at random locations R at from 1985 feeding sites but was canopy cover (R - 36%) and density taller plants of shrub structure (Table 1.4). grouse used sites with greater big sagebrush greater With the with random sites. to random sites within to 1986 feeding sites. density - 1.1 plants/m2). height in drainage I pooled 1985 and 1986 at feeding sites in drainages 'was similar random sites in drainages cover - 16%, density feeding sites vs. drainage sites for 3 of 4 variables greater in drainages, with topographic sites was compared sagebrush Sagebrush (4 showed of Winter Feeding and Random Sites.--Because feeding similar between Only 1 of 16 comparisons in 1985 than in 1986 (f - 0.006). of density estimates Comparison differences structure variables) (f < 0.05) between winters. differences feeding X 4 sagebrush (0-50 low, 0-50 high, structure 1985 and 1986 feeding sites were minimal. for which due to the inadequate sample R height ~ 31 50 northeast (N - 3) of �59 I observed while black sagebrush it was present occurred, between Chemical of black sagebrush Within Among Random Foliage Collection Sites Big sagebrush plants were present 21 plants/site, SE 1.1). (R 20 plants/site, SE 2.6) and was also sampled. This reflected topographic vigor that was also observed differences there were no differences monoterpene was present at 7 sites Big sagebrush in other terrain variation in plant growth and at random sites where shrub structure was differences comprised in site conditions (Table and sagebrush 1.5). plants that grew in different There were also no topographically- (f > 0.05) in the 6 individual monoterpenes 80-90% of the total mono terpene content was no correlation (K - -0.21) or total monoterpene big sagebrush plants sampled at random sites. crude protein content each was found. differences (Table 1.5). There (R content (K = 0.26) in (R - 18.1%) and total monoterpene (R - 1.08%) were lower than protein monoterpene that (f > 0.05) between mean plant height and levels of crude protein Black sagebrush growth forms, (f > 0.05) in crude protein or total content among big sagebrush terrain categories content at each site (Table 1.4). Despite related Black sagebrush in drainages were taller than plants evaluated (f - 0.14) sagebrush plants at 8 r~ndom sites in each of 4 categories. categories. sites where it was similar (R plants feeding sites (R = 7%) and random sites (R = 13%). Variation I collected terrain at 31% of 100 random sites. canopy intercept feeding at only 8% of 87 winter (R 2.15%) of big sagebrush 21.3%) and total at the 7 sites where I did not attempt to evaluate black sagebrush among terrain categories due to small sample sizes. chemical �60 Table 1.5. Chemical characteiistics (% dry matter) of big sagebrush leaves collected at 32 random sites in 4 terrain categories (8 sites/category), Gunnison Basin, Colorado, Harch - April 1986. o - o - 5° 50 low R Category Crude protein Individual SE 19.7 Total monoterpenes high 1.0 2.08 R SE 22.2 1.5 > 50 Drainage R 21.6 SE 1.6 southwest :R 21.5 SE 0.9 0.54 0.2 1.86 0.3 2.23 0.2 2.06 0.1 0.62 monoterpenes Camphene 0.16 0.02 0.15 0.03 0.15 0.01 0.15 0.01 0.98 Artho1e 0.09 0.09 0.10 0.02 0.08 0.01 0.12 0.01 0.23 1,8 Cineole 0.46 0.06 0.40 0.06 0.48 0.03 0.48 0.04 0.62 Camphor 1.03 0.11 0.95 0.12 0.96 0.06 1.01 0.05 0.91 Borneol 0.10 0.02 0.08 0.01 0.10 0.01 0.10 0.01 0.58 28.0 minb 0.05 0.003 0.03 0.007 0.05 0.004 0.05 0.005 0.09 similar among terrains. aprobabi1ity bUnidentified retention time. that means were monoterpene, name designates gas chromatograph �61 Chemical Differences Sagebrush foliage samples were collected sites in 1986. sagebrush (R = Between Browsed and Unbrowsed Big sagebrush was present 15 plants/site, plants/site, SE = dominated = from 20 winter Plants feeding all feeding sites and black at only 1 feeding site. SE Sagebrush A total of 300 browsed 1.0), and 285 unbrowsed (R - 14.3 1.5) big sagebrush plants was collected from feeding sites. Crude protein content was similar 20.6%) and unbrowsed sites (Table 1.6). (£ - 0.80) between sites. monoterpenes Crude protein content of big sagebrush averaged Total monoterpene were also similar big sagebrush plants at at random foliage content and 6 individual (£ > 0.05) between browsed and unbrowsed plants at winter feeding sites (Table 1.6). Monoterpenes 2.0% of dry matter in each group of samples. monoterpene (R = (R - 20.3%) big sagebrush plants at winter feeding feeding sites was similar to values observed collection browsed content at feeding sites was similar Total to values observed at sites (Table 1.5). random foliage collection DISCUSSION Sage Grouse Habitat Use Topographic reflected variation physiographic in shrub structure differences plants were tall and vigorous 50 northeast height was intermediate relative conditions on 0-50 high and> short sagebrush in site conditions. in drainages Soils on 0-50 low and> at random locations Big sagebrush due to mesic conditions. were well drained and sagebrush to other terrain categories. 50 southwest plants with open canopies. sites resulted High sagebrush Xeric in stands of density on �62 Table 1. 6. Chemical characteristics (% dry matter) of leaves of browsed and unbrowsed big sagebrush plants at sage grouse winter feeding sites (N - 20), Gunnison Basin, Colorado, Janaury - March 1986. Browsed Category 2l Crude Protein 20.6 Total monoterpenes Unbrowsed SE 1.1 2l 20.2 SE fa 1.1 0.80 2.00 0.10 2.01 0.12 0.97 Camphene 0.14 0.01 0.15 0.01 0.64 Arthole 0.08 0.009 0.09 0.01 0.71 1,8 Cineole 0.41 0.02 0.41 0.03 0.94 Camphor 1.01 0.06 1.05 0.06 0.68 Borneol 0.10 0.01 0.10 0.01 0.61 28.0 minb 0.04 0.002 0.05 0.003 0.39 Individual monoterpenes aprobability that means were similar between browsed and unbrowsed plants. bUnidentified retention time. mono terpene , name designates gas chromatograph �63 0-50 high sites reflected the tendency of black sagebrush stands of small, closely-spaced plants. Sage grouse foraged at sites where sagebrush was maximized. Topographic distribution was greater than expected shrub growth, mid-winter drainages to snow depth. in both winters. availability of sagebrush than in other terrain categories plants provided were sheltered forage and concealment apparent differences important covers shorter Use of Due to vigorous was greater (Fig. 1.6). in Large sagebrush from avian predators. costs by foraging beneath in drainages especially above snow Drainages from wind and sage grouse may have reduced thermoregulatory structural exposure of sage grouse feeding sites was affected by sagebrush height relative drainages to occur in between random and feeding foraging Drainages may be sites in severe winters when deep snow Schoenberg plants on more arid terrains. that a high proportion winter use sites in Jackson Sagebrush sites were less than in more xeric sites. sagebrush (1982) also observed closed canopies. (32-46%) of sage grouse County, Colorado occurred in drainages even though that terrain comprised < 2% of his study area. A high proportion slopes, although southwest of feeding sites occurred use was not disproportional slopes was shallow relative Shrub structural differences .were apparent between and random sites. canopies more closed at foraging locations shallow small « at southwest snow cover Sagebrush southwest 1 ha) sites beneath Snow on slope plants were taller and shrub than at random sites. slope feeding sites was readily (Fig. 1.6) slopes above drainages. to availability. to other terrain categories. foraging Sagebrush on > 50 southwest Southwest slope foraging rock outcrops available above sites were often or snow cornices, These areas received more moisture and lower than �64 1::::::::::::::1 MEAN SAGEBRUSH HEIGHT AT FEEDING SITES (Z] MEAN SAGEBRUSH HEIGHT AT RANDOM SITES 1<::::::::1 1985 FEBRUARY MEAN SNOW DEPTH ~ 1986 FEBRUARY MEAN SNOW DEPTH 60 50 45 40 ~ o 30 20 10 o DRAINAGE 0_50 LOW 0_50 HIGH >5 o SW Fig. 1. 6. Mean sagebrush height at feeding and random locations relative to snow depth within terrain categories in the Gunnison Colorado, 1985-86. Basin, �65 surrounding sites and sagebrush with slopes> Southwest 12 observations) used in 1986 (15% of feeding sites). of steep southwest following terrain categories Most use (7 of slopes in 1986 occurred slopes became snow-free During the plants and made access to foliage difficult. plants on steep southwest Schladweiler sagebrush of more readily available (1972) observed more quickly of winter use sites on slopes> suggested because 60 in Jackson aspects. January County, Colorado only 13% and exposure Basin regularly slopes foliage readily available at mesic sites where shrub growth was vigorous. was due to differences in 1985 than in 1986. in snow depths between snow depth on 0-50 low sites was shallower than in 1986 (R - 32.9 cm). category rarely used Sage grouse likely f~raged on southwest Use of 0-50 low sites was greater probably Eng and (6-150) to steep (> 150) slopes with snow was shallow and sagebrush localized exploited Beck (1977) observed Sage grouse in the Gunnison areas of moderate southwest than that sage grouse rarely foraged on slopes to minimize to avian predation. occupied forage. that sage grouse in Montana stands on slopes during winter. in most Sagebrush in other areas and sage grouse may have temporarily these sites because during a 3- a severe winter storm in mid-February. storm, wet snow became lodged in crowns of sagebrush plants aspects 150 were rarely used in 1985 (1% of feeding sites) but were more heavily day period growth was vigorous. in 1985 occurred accumulations. This years. Mean in 1985 (R Many (48%) observations in January prior to heavier = 17.5 cm) in the 0-50 low mid-winter snow Use of 0-50 low sites was reduced when snow depths exceeded approximately winter. Sagebrush 30 em in February 1985 and throughout stands on 0-50 low sites consisted shorter plants than stands in drainages. the 1986 of slightly Snow was deeper than on �66 southwest of deep snow (> 30 em) foraging slopes. During periods were less available proportion on 0-50 low sites because of the sagebrush structure snow covered than in drainages Canopy cover and density were greater at winter to random feeding locations. However, and random locations. periods Uniform (Fig. 1.6). feeding sites relative soils and topography in sagebrush stands of tall sagebrush approximately of 0-50 low growth characteristics. that could be exploited of deep snow were not widely available snow depth exceeded a higher sagebrush height did not differ between sites did not promote heterogeneity Localized sites during on 0-50 low sites. When 30 cm, sage grouse were less likely to forage on 0-50 low sites and more likely to exploit other terrain categories reported where sagebrush was more available. greater winter use of flat « 50) Schladweiler 1972, Beck 1977). However, were shallow (3-25 cm) relative to Other researchers areas by sage grouse have (Eng and snow depths in those studies winter conditions in the Gunnison Basin during 1985 and 1986. Only 13% of feeding sites were observed ~ Average height canopies greater of big sagebrush than on southwest interspersed of its short height, by shallow aspects as winter habitat. were exposed on 0-50 high sites. completely Because even snow to wind may have made 0-50 high sites less Feeding sites on the 0-50 high category Because mesa and ridgetops snow tended to become drifted and vigorous was covered Small plant size, moderate in shallow depressions. to winds, and Snow depths were Black sagebrush black sagebrush was usually snow depths of 20-30 cm. occurred depressions (Fig. 1.6). with big sagebrush and greater exposure favorable usually plants at these sites was shorter were more open than in more mesic terrains. frequently depths, on 0-50 high sites. ;. sagebrush in shallow growth was favored. �67 I observed only 3 feeding sites on slopes with northeast even though that terrain category comprised of the Gunnison Basin. a large percentage Snow was deeper on northeast other terrain categories. aspects, Also, big sagebrush (30-32%) slopes than in most plants were smaller and canopies more open than in drainages. northeast slopes was covered as a result of deep snows (Fig. 1.6) and foraging sites were not available. Most sagebrush Northeast foliage on slopes may be used during mild winters with little snow. However, in normal winters when snow depths are> 30 cm on northeast slopes, little foraging is likely to occur in that terrain category. Sage Grouse Forage Selection Sage grouse foraged at sites where big sagebrush was the dominant shrub species. 8% of winter Black sagebrush was rarely consumed feeding sites. grouse in Idaho usually sagebrush resource was usually Dalke et a1. (1963) observed foraged on black sagebrush. snow covered and not readily in the Gunnison Basin. Black sagebrush ridges and mesas at sites where sagebrush to winds. Taxonomic concentrations subspecies Tisdale on subspecies concentrations in foliage protein has been observed 1979, 1981). is influenced 1977). that sage However, available primarily black as a forage occurred on fed on big sagebrush due content. variation (Welch and McArthur at only plants were short and exposed Sage grouse may have preferentially to its higher protein and occurred and monoterpene among subspecies Distribution by site conditions of big sagebrush and topography Sage grouse have been observed and Braun 1985). (Winward and to preferentially of big sagebrush with high protein (Remington of big sagebrush forage and low mono terpene Topographic distribution of �68 feeding activity physiographic distribution 1965, Remington subspecies Foliage in some sage grouse populations and Braun 1985). of big sagebrush crude protein big sagebrush distribution influenced plants in different variation Remington Selection big sagebrush Similarities plants in crude protein composition of 18 individual a site. 24% crude protein) of sagebrush sage grouse concentrations. (Hoffmann 1961. in the Gunnison concentrations composition were variable The magnitude These results was not necessarily and unbrowsed of sagebrush analysis plants collected and unbrowsed Basin. of browsed plants was for selection I evaluated did protein from 3 feeding sites. among plants that grew of variation within was similar to the total variation data). content 1970, Doerr et al. 1974, Hohf et al. in a preliminary concentrations within Hupp, unpubl. feeding sites feeding sites, and that opportunities However, together that within of big sagebrush, that chemical not exist. Crude protein Basin was primarily foliage chemistry. plants were not apparent within Physiographic in crude protein content of browsed could indicate uniform were similar among in other grouse species et al. 1970, Pulliainen Differences Basin. plants or plant parts with high protein observed (Gill foliage above snow and not by fed on plants with high crude protein has also been widely 1987). in the Gunnison in the Gunnison and Braun (1985) observed of individual subspecies terrain categories. in sagebrush the big sagebrush was the only concentrations of sagebrush of a single subspecies selectively Miller that occurred and monoterpene by availability big sagebrush Mountain of sage grouse foraging topographic comprised of preferred may reflect indicate among sites (J. W. that chemical uniform within sites (14- composition feeding sites and that �69 opportunities for sage grouse to selectively forage on higher protein plants within sites did exist. Protein levels of both browsed 20.2%) mountain relative big sagebrush to protein (R - 20.6%) and unbrowsed plants in the Gunnison concentrations in Wyoming (R - Basin were high 14.1%) and mountain (R ~ 10.8%) big sagebrush plants in northern Colorado Braun 1985). species Crude protein Miller relative nitrogen Herbivores lagopus scoticus), forage resources rich sites populations rich protein that forage in were greatest In the Gunnison leaves was higher exploited (Lagopus content between browsed at nitrogen-poor for high protein plants was weaker (Moss 1972, Lance 1983). forage resources that are low in In' red grouse in protein (Ca1luna vulgaris) of big sagebrush 1961, plants or plant parts than herbivores (Mattson 1980). the differences sites while selection (Hoffmann 1970, Moss 1972, Doerr et a1. 1974, Hohf concentrations heather and are also low (6-14% of Basin big sagebrush that exploit rich environments and unbrowsed content selection may be more likely to select individual with higher protein nitrogen protein to Gunnison et al. 1970, Pul1iainen et a1. 1987). (Remington levels of forages used by other grouse that have demonstrated dry matter) (R ~ than at nitrogen- Basin, protein protein levels of by other grouse species or sage grouse thus far studied. Gunnison Basin sage grouse exploited a forage and birds may have been less likely to selectively forage on higher protein plants within sites. Remington and Braun (1985) observed foraged on subspecies monoterpenes, of big sagebrush with low concentrat~ons and that within stands of mountain grouse selected Monoterpenes that sage grouse selectively big sagebrush, plants that were low in oxygenated have been observed of sage monoterpenes. to depress bacterial digestion in �70 ruminants (Nagy et al. 1964, Oh et al. 1968, Schwartz Remington and Braun monoterpenes could influence forage selection terpenes (1985) suggested bacteriostatic minimized has been observed al. 1980b, Farentinos Although avoidance vertebrates, browsed of monoterpenes Basin. Avoidance monoterpenes Individual (Schwartz et among herbivorous did not differ between plants at sage grouse feeding sites plant selection on the basis of content may be less likely to occur when concentrations monoterpenes are low. Remington and Braun for low mono terpene plants within sagebrush (mean mono terpene content - 1.4%) but did observe 2.6%). low stands of mountain Monoterpene (R - big sagebrush concentrations 2.0%) in the Gunnison study area. Although cautiously due to dissimilar equipment, monoterpene sagebrush stands of Wyoming big of mountain between big sagebrush to Remington studies gas chromatography concentrations effects procedures within Gunnison of monoterpenes known and need to be experimentally of sage grouse forage selection plants were and Braun's should be interpreted and Basin big may have been below levels at which selection The biological selection (mean mono terpene content - Basin relative differences of (1985) did not observe selection within of et al. 1987). has been observed big sagebrush of and that sage grouse in other species of herbivores total and individual in the Gunnison effects mono terpene consumption. et al. 1981, Personius and unbrowsed monoterpene avian cecal digestion et a1. 1980a). would occur. on sage grouse digestion determined before are not' the significance can be fully evaluated. �71 MANAGEMENT IMPLICATIONS Winter foraging activity not evenly distributed distribution topographic among physiographic variation in big sagebrush Instead, Physiographic exploited Sagebrush because vigorous cover availability. crude protein by or monoterpene sage grouse foraged at sites where sagebrush above snow was maximized. frequently features. Basin was of sage grouse feeding activity was not influenced concentrations. exposure of sage grouse in the Gunnison Southwest in drainages shrub growth insured forage and slopes were also frequently snow depth was shallow and sagebrush was foliage was available sites. Use of 0-50 low sites was variable depth. Less foraging and dependent used because at feeding upon snow occured on 0-50 low sites during periods of deep (> 30 cm) snow because sagebrush foliage was less available. Mesa and ridge tops were not heavily plants on exploited because these xeric sites were often covered by snow. foraged on northeast short sagebrush Sage grouse rarely slopes even though sagebrush plants on northeast slopes were taller and canopies more closed than stands in more xeric terrains. Deep snows typically limited shrub exposure on northeast slopes. Structural value measures in assessment Sagebrush structure of big sagebrush would be of relatively of sage grouse winter habitat relative would need to be considered for each topographic to distinguish drainages. depth. to topographic and separate category. random locations variation structural Structural variables Basin. in snow depth criteria established were not adequate from sage grouse feeding Use of 0-50 low sites was variable It would not be feasible in the Gunnison little sites in and dependent upon snow to define a range of structural �72 characteristics 50 low feeding measures that could be used to consistently sites among winters are potentially availability availability conditions whether severity. useful to identify winter on southwest is questionable of differing identify potential is necessary Structural feeding site slopes, and mesa and ridge tops. identification of winter on these terrains. of these sites, sagebrush 0- However, it feeding site Due to the xeric removal to improve grass forage is not likely to occur. Physiographic and distribution Basin. criteria of sage grouse winter Maintenance important appear to be useful to assess availability of sagebrush in drainages in drainages 10% of the sagebrush Higher percentages exposed vegetation and was available sagebrush During'the in the Gunnison to sage grouse (J. occurred 1985-86 winter~. in drainages. W. Hupp, unpubl. Drainages that was exposed are a small (3-4%) of the total land area in the Gunnison frequently the focus of sagebrush growth. Removal of sagebrush disproportionally severe effect on availability feeding habitat. Maintenance also be emphasized on that terrain. occupied due to the frequent has a of sage grouse winter on southwest observation These findings have implications by sage grouse in which terrain Basin, yet are due to the vigorous in drainages of sagebrush data). in the region were proportion sagebrush is available Basin was not snow- Sagebrush treatments Maintenance severe 1984 winter < (78-84%) of sagebrush vegetation during the milder in 1984 primarily in this area. assures that vigorous to sage grouse during severe winters. in the Gunnison is a particularly aspect of winter habitat management of sagebrush covered feeding habitat slopes should of foraging activity for other regions is heterogeneous, winter snows �73 typically exceed 30 cm on low flat areas, and where mountain sagebrush is the dominant shrub species. LITERATURE CITED Alldredge, J. R., and J. T. Ratti. statistical techniques Wildl. Manage. Autenrieth, Burnham, 1986. Comparison for analysis of resource from line transect Wildl. Monogr. Beck, T. D. I. 72. 1977. selection sampling availability J. Wildl. Manage. data. J. W. 1977. Estimation An ecological Ph.D. Thesis, Washington Practical and habitat Wild1. for Soc. Bull. 5:99-106. 1984. of utilization- 48:1050-1053. study of sage grouse in State Univ., Pullman. nonparametric statistics. 84pp. John Wiley and Sons, New York, N.Y. 49lpp. Dalke, P. D., D. B. Pyrah, D. C. Stanton, J. E. CrawforQ, Schlatterer. of populations. Guidelines for analysis J. Wildl. Manage. 1982. 1980. 42pp. 41:18-26. and P. R. Krausman. of a technique J. W., Jr. 1980. Sage grouse flock characteristics Byers, C. R., R. K. Steinhorst, Conover, Sage grouse management of biological of sage grouse habitats. Clarification J. 202pp. in winter. maintenance Idaho. 1982. and J. L. Laake. Braun, C. E., T. Britt, R. O. Wallestad. Connelly, selection. West. States Sage Grouse CQmm. Tech Bull. 1. K. P., D. R. Anderson, density of some 50:157-165 R., C. Braun, and W. Molini. practices. big 1963. sage grouse in Idaho. Ecology, productivity, and E. F. and management J. Wildl. Manage. 27:811-841. of �74 Doerr, P. H., L. B. Keith, D. H. Rusch, C. A. Fischer. Characteristics in Alberta. of winter feeding aggregations J. Wi1dl. Manage. Eng, R. L., and P. Sch1adweiler. and habitat 1974. of ruffed grouse 38:601-615. 1972. Sage grouse winter movements use in central Montana. J. Wildl. Manage. 36:141- 146. Epstein, W. M., L. R. McGee, C. D. Poulter, Mass spectral and L. L. Marsh. data for gas chromatograph identification 1976. - mass spectral of some irregular monoterpenes. J. Chern and Eng. Data 21:500-502. Farentinos, R. C., P. J. Capretta, 1981. Selective monoterpenes herbivory in ponderosa R. E. Kepner, in tassel-eared and V. M. Littlefield. squirrels: role of pines chosen as feeding trees. Science 213:1273-1275. Gill, R. B. 1965. Distribution and abundance grouse in North Park, Colorado. Univ., Heller, Fort Collins. data base. Higby, L. W. 1969. project. M. S. Thesis, Colorado sage State 187pp. S. R., and G. W. A. Milne. spectral of a population Editors. 1978. EPA/ErH mass U. S. Dep. Corom., Washington, D.C. 988pp. A summary of the Longs Creek sagebrush control Proc. Bien. West. States Sage Grouse Workshop 6:164- 168. Hoffmann, R. S. 1961. Wi1d1. Manage. Hohf, R. S., analysis Manage. The quality of winter food of blue grouse. 25:209-210. J. T. Ratti, and R. Croteau. of winter 51:159-167. 1987. Experim~nta1 food selection by spruce grouse. J. Wi1d1. J. �75 Horwitz, W. Editor. Anal. Hunter, Chern. A. N. D.C. Mattson, Miller, D.C. W. J., Jr. 1980. Pages populations 323-335 to their R. Ecol. Herbivory in A. Watson, (Lath.) Wildl. Brit. in relation 1970. in relation to plant 14:78-80. nitrogen Responses improvements Animal Ecol. Scand. 11:119-161. populations in relation 10. (Lagopus to chemical of red of their food. Soc. Symp. by red grouse Manage. Oceanic lagopus composition. J. Anim. Admin. 91:4-6. R. L. Denver, Administration. Nat1. and W. M. Longhurst. inhibition from relatively Microbio1. Effects on deer rumen microbial U. S. Dep. Comm., Oh, H. K., M. B. Jones, microbial 1964. of function. J. 28:785-790. and Atmospheric Colorado. isolated and G. M. Ward. oils of sagebrush data, Patterson, ed. Ornis 41:411-428. essential rumen Serv., sites by hen red grouse to experimental J. E., H. W. Steinhoff, National Off. of the Gunnison Soil Conserv. and D. Jenkins. Food selection scoticus) Soil survey during breeding. food resources. 1972. 1975. Rev. Eco1. and Syst. G. R., A. Watson, grouse Assoc. l018pp. of feeding scoticus Annu. of analysis. 89pp. Selection lagopus content. methods U. S. Dep. Agric., 1983. Lagopus Nagy, Washington, Colorado. Washington, Moss, Official W. R., and C. F. Spears. area, Lance, 1980. resulting unplalatable 1986. Oceanic 1968. and Atmospheric Comparison from various plant Climatological of essential oils species .. Appl. 16:39-44. 1952. Colo. The sage grouse 341pp. in Wyoming. Sage Books, Inc. �76 Personius, J. R. Stephens, and R. C. Kelsey. T. L., C. L. Wambolt, 1987. Crude terpenoid sagebrush. Pulliainen, E. influence on mule deer preference J. Range. Manage. 1970. capercaillie Composition for 40:84-88. and selection (Tetrao urogallus) of winter food by the in northeastern Finnish Lapland. Suomen Riista 22:67-73. Remington, T. E., and C. E. Braun. in winter, Schneegas, 1985. North Park, Colorado. E. R. 1967. Sage grouse food selection J. Wildl. Manage. 49:1055-1061. Sage grouse and sagebrush North Am. Wi1d1. and Nat. Resour. Conf. Schoenberg, T. J. 1982. Collins. Schwartz, and habitat M.S. Thesis, Colorado selection State Univ., Fort 86pp. C. C., J. G. Nagy, and W. L. Regeliri. yield, terpenoid deer. J. Wildl. Manage. concentration, _________ , W. L. Regelin juniper Trans. 32:270-274. Sage grouse movements in North Park, Colorado. control. 1980a. and antimicrobial Juniper oil effects on 44:107-113. , and J. G. Nagy. forage and volatile 1980b. Deer preference oil treated foods. J. Wildl. Manage. 44:114-120. Scott, T. G., and C. H. Wasser. plants for wildlife biologists. Washington, Wallestad, 1980. R. O. treatment. D.C. Checklist of North American The Wildlife Society, Inc., 58pp. 1975. Male sage grouse responses J. Wildl. Manage. ________ , J. G. Peterson, grouse in central Montana. to sagebrush 39:482-484. and R. L. Eng. 1975. Foods of adult sage J. Wildl. Manage. for 39:628-630. �77 Welch, B. L., and E. D. McArthur. crude protein uniform among Artemisia garden. 1981. subspecies in a uniform Winward, Variation complex Stn. Bull. 19. in Idaho. lSpp. subspecies of monoterpenoid of Artemisia J. Range Manage. A. H., and E. W. Tisdale. in winter levels grown of in a 32:467-469. and accessions garden. tridentata Variation tridentata J. Range Manage. _______ , and sagebrush 1979. 1977. content among tridentata grown 34:380-384. Taxonomy of the Artemisia Idaho For., Wildl., and Range Exp. �78 CHAPTER 2 ENDOGENOUS RESERVES AND COURTSHIP BEHAVIOR OF ADULT MALE SAGE GROUSE INTRODUCTION The energetic and nutritional many avian species. energetic (Korschgen mobilization defense demands associated 1977. Drobney of endogenous reserves has also been observed 1981; Brittas Hohman Females may use endogenous and nutritional incubation 1984; Alisauskas 1986). endogenous 1986). Female reserve Whether reserves to meet with follicle 1980. Ankney 1984). during courtship and Ankney reproductive growth and Male or territory 1979; Krapu 1985; Krapu et a1. 1985; success may be influenced size (Ankney and MacInnes by availability are high in (Ankney 1977. 1984; Raveling male reproductive also be mediated costs of r~production by 1978. Krapu 1981. Erikstad success or courtship of endogenous behavior reserves may is poorly understood. Species insights courtship displays of Tetraoninae regarding behavior. the relationships and reproductive and prolonged 1970) likely result use endogenous Opportunities that display periods to examine between to partially relationships male endogenous success. of courtship in high energetic reserves on leks may provide Elaborate among males: meet energetic between reserves. courtship in 1ek species demands useful (Hjorth Males may costs of courtship. endogenous reserves and �79 courtship exist in lek species because courtship be quantified Previous during concurrent studies of breeding Tetraoninae evaluation behavior of endogenous season endogenous reserves. reserves among male that display on leks have not been reported. the extent to which male energetic of males can Therefore, reserves are used during courtship is unknown I examined breeding season fat and protein sage grouse in 2 Colorado populations. reserves of adult male The study was divided into 2 phases. In the first phase, I tested the hypothesis mobilize lipid and protein reserves during the period on leks. I also compared differences years to evaluate deposition. that they display reserves among the effects of winter severity on lipid and protein In the second phase of the study I compared lipid reserves differences in endogenous that males and courtship behavior between populations in display behavior display occurred and determine whether in a population adult male to examine faster rates of with larger lipid reserves. METHODS Phase 1: Male Endogenous Reserves During Spring Courtship Adult male sage grouse were collected Gunnison counties, Colorado (Fig. 2.1). on leks in Jackson The study areas were (> 200 km) from each other. geographically separated high elevation (> 2000 m), sagebrush-dominated and Braun (1985). Both areas are intermontane Study areas have been described by Hunter and Spears (1977), and Remington and basins. (1975), Beck Males were collected in Jackson County during 2 discrete sampling periods courtship) from 1983 through 1985. In Gunnison annually (early and late County, collections �80 COLORADO JACKSON CO . • DENVER GRAND JUNCTION • ~GUNNISON co. Fig. 2.1. Colorado study areas. �81 during early and late courtship were obtained in 1984 and 1985. The early sampling period occurred during the first 2 weeks that males consistently displayed Early collections occurred prior to or during Severe winter weather populations; on leks and was usually between week of April. peak female attendance. in 1984 delayed courtship early courtship collection The late courtship 22 May in each year. collected 14 and during night-time Usually no more than 2- from a single lek within a sampling period. 8 and 10 adult males were collected during all years in both populations. in each sampling period Endogenous reserves of 96 adult male sage grouse were evaluated between Following in both sampling period was between trapping on or near leks (Giesen et al. 1982). Between initiation did not occur until the final Males were primarily 3 males were collected 1 and 15 April. collection, frozen in plastic bags. 1983 and 1985. (± 5 g) then sealed and males were weighed Prior to laboratory from a total analysis, carcasses were thawed and culmen, tarsus, and wing length from the carpals to the tip of the longest primary measured. distally from the carpals were removed and discarded. (~. pectoralis and~. excised and weighed. tibiotarsus) supracoracoideus) Leg muscles gizzard, intestines, Breast muscles on I side of the body were (all muscles attached were excised and weighed also removed and weighed. Muscles Feathers, head, tarsi, and wings in 1985. Gut contents to the femur or Internal organs were (ingesta in the esophagus, and caeca) were removed, weighed, and discarded. and internal organs were returned to the carcass. was refrozen homogenate then homogenized meat grinder. The was spread on aluminum foil and dried in a forced air oven at 80 C for 24 hours. carcasses 4-6x in a commercial The carcass Kerr et al. (1982) found that drying of Microtus at 120 C did not affect lipid reserves. The dried carcass �82 was further homogenized in a high speed Wiley Mill until it passed through Lipids were extracted a screen. 2-rnrn homogenate of with diethyl ether for 6 hours in a Soxhlet apparatus. Dobush et al. (1985) observed for carcass Repeated from lO-g subsamples analysis because extraction that diethyl ether was a suitable it did not extract of some samples for an additional remove more lipids from the homogenate. considered adequate for exhaustive structural Six hours was therefore extraction of lipids from sage Dobush et al. (1985) also observed ether exhaustively extracted determined carcasses lipids from homogenized within 6 hours. from subtraction lipids. 6 hours failed to grouse carcasses. caerulescens) solvent that diethyl snow goose (Chen Weight of lipids was of post-extraction mass from pre-extraction mass. Hypotheses lipid reserves, of no differences in live weights, or breast muscle weights periods were tested using 2-way analysis tarsus lengths, among years or courtship of variance for each population. Live weight was body weight at collection of ingesta. Lipid reserves were expressed Breast muscle weight supracoracoideus as percent (the combined weights of the reserves. In sage grouse, likely the major source of labile protein. mass would be indicative Phase 2: Endogenous I compared between Reserves the breast muscles Changes and Courtship are in breast muscle Behavior and courtship in 1986 to test the hypotheses rates differed between by 2) was used as of protein mobilization. adult male lipid reserves populations of live weight. tl. pectoralis and tl. from 1 side of the body multiplied an index to protein minus the weight populations behavior that male display and that male lipid reserves were �83 larger in a population collected with more rapid display. in each population during early courtship and lipid reserves were evaluated each population accessible for evaluation by vehicle as in Phase 1. behavior was observed between 9 April and 6 May. Evaluations in each population at leks. temperatures ranged between Evaluation adequate Numbers minutes nuptial behavior 1970) and strutting investment were made of 150-200 m. (about 0.75 hours before 20-minute for observation from yearlings plumage. intervals. of courtship Adult behavior. by their large size and Selected males were observed displays The strutting for 5 (Wiley 1973) observed display during is the predominant of male sage grouse on leks (Lumsden 1968, Hjorth rates were presumed in courtship. to be indicative Time that males were involved (chases, fighting, the 5 minutes 1.75-2.0 hours prior to Observations at approximately selected that period recorded. disputes arrived display began as soon as light was and number of strutting courtship when low of females and adult males on leks were recorded. Adults were distinguished distinctive were -8 and 2 C, winds were < 16 krnfhour, and males from females Lek counts were repeated males were randomly on each lek were during mornings Observers of courtship (17-27) of of display behavior scope from a distance to distinguish sunrise). Leks were during 5 mornings to record time of first display. through a spotting 3 leks in during the period of peak female All observations there was no precipitation. sunrise I selected and were attended by similar numbers Courtship attendance (1-15 Apr) in 1986 of courtship ?ehavior. males. initiated Ten adult males were of observation confronted displays was recorded. of energetic in territorial [Hjorth 1970]) during I also recorded time birds �84 spent in crouched positions and common ravens (Cor,~s corax). Female-male rates distance influences (Gibson and Bradbury male and the nearest 1985). Distance males continued maximum number Evaluation Number of males observed rates were corrected disputes or avian predator female distance (Norusis survey period. was calculated disturbance. between populations for each male. due to territorial Males, for which disruption that mean strutting categories of display rates Hotellings display did not differ between populations statistical Differences between 12 was used populations. procedure simultaneously for each lek- rates across lek- that tests across several variables in lipid reserves were evaluated with ~ tests. of faster display rates in a population male lipid reserves behavior selected the lek was recorded. category within a population. 1986:117). relationship period was on leks to facilitate for time of disruption 12 is a multivariate Observation selected 20 cm high) Mean display rates were calculated to test the hypothesis female distance « Small 2 minutes were not used for comparison populations. differences observation during the morning displays/minute Display Hotellings a randomly until the number of males on leks was < 25% of the of strutting time exceeded display of display rates of randomly Time at which 75% of males had abandoned between between female during the 5-minute estimation. avian predators sage grouse strutting stakes were placed at 10-m intervals distance between to potential as either < 10 m, 10-30 m, or > 30 m. recorded wooden in response would not necessarily indicate size of male energetic but would provide supportive a cause and effect reserves evidence with larger and courtship for the hypothesis that �85 courtship is affected by endogenous dissimilar display reserve size. Observ~tion of display rates when lipid reserves were equal, similar rates when reserves were unequal, population with larger energetic interpopulation relationship or slower display reserves would between indicate lipid reserve rates in a that an size am± courtship J behavior was not likely. RESULTS Phase I: Male Endogenous Reserves During Spring Courtship live weights "dLf'f er-ed Body Mass and Size. -- In each population, (r < 0.05) among years but not between courtship periods (~ble 2.1). Males collected in other years between in 1984 were lighter than individuals (Table 2.2). courtship Differences populations. from Jackson Gunnison periods County. 2.2). Couty weighed each population, among years and between among years. lipid reserves reserves early and late courtship Lipid reserves (Table 2.2). similar among years (Table 2.2). in all years (Table that indtcated reserves were not similar during early courtship than in other years of male during early there were also strong interactions between in early and lateccourtship In both study areas, lipid reserves However, 10%~shorter (Table 2.2). were larger than late courtship differences 25-30% less ethan males Tarsus lengths were approximately Reserves.--In sage grouse differed courtship in each population. in adult male body size were apparent between County males (Table 2.1). Tarsus lengths were similar among years and Males from Gunnison Endogenous co11etted Late courtship were lower in 1984 lipid rese~es were �86 Table 2.1. Differences in sage grouse physical characteristics among years and between courtship periods (2-way ANOVA) , Gunnison and Jackson counties, Colorado, 1983-85. ,_- ," County -- Year Variable Courtshin fa E neriod fb E Year x neriod fC E Jackson Lipidsd 24.4 < 0.001 2~8.l Live weighte 16.1 < 0.001 0.3 0.59 1.2 0.31 < 0.001 < 0.001 24.5 Breast musclef 3.0 0.06 0.7 0.42 3.1 0.05 Tarsus 1.6 0.21 1.7 0.20 3.1 0.06 8.5 0.006 Gunnison Lipids 14.5 0.001 < 0.001 70.7 "Live weight 5.8 0.02 2.8 0.10 1.0 0.76 Breast muscle 2.5 0.13 2.8 0.11 0.2 0.67 Tarsus 1.7 0.21 0.008 0.93 1.8 0.19 - _-s.. ...:... -"-- aprobability that means are similar among years. bprobability that means are similar between early and late courtship. cProbability that differences between early and late courtship means are similar among years. dLipids as percent of live weight - weight of digestive tract contents. eBody weight at time of collection - weight of digestive tract contents. fCombined weights of tl. nectoralis and tl. sunracoracoideus. �87 Table 2.2. Physical characteristics of adult male sage grouse during courtship in Gunnison and Jackson counties, Colorado, 1983-85. County Year Period Live weight {g)a SE H :R Li12ids {%)Q SE :R Li12ids {g) SE :R Breast muscle {g)£. Tarsus{mm) SE SE :R :R Jackson 1983 Early 9 2,863 32 5.4 0.38 155 10.8 654 8.4 74.1 0.58 10 2,799 45 1.1 0.20 30 5.3 612 13.2 72.6 0.85 Early 10 2,524 61 2.3 0.38 59 10.6 597 14.7 74.7 0.56 Late 10 2,598 41 0.8 0.10 20 3.1 614 11.4 72.9 0.94 Early 10 2,651 57 5.5 0.29 145 7.6 608 13.4 74.0 0.53 Late 10 2,699 37 0.5 0.10 13 2.3 607 10.8 75.1 0.45 10 1,897 20 2.2 0.44 42 8.2 437 11. 7 69.9 0.47 9 1,846 57 0.7 0.10 12 1.8 423 7.7 69.2 0.21 10 1,996 37 4.1 0.25 82 5.6 458 12.6 68.6 0.69 8 1,923 21 0.8 0.10 16 1.9 435 8.7 69.3 0.52 of digestive tract Late 1984 1985 Gunnison 1984 Early Late 1985 Early Late aBody weight at time of collection - weight contents. bLipids as percent of live weight - weight of digestive tract contents. cCombined weights of M. 12ectora1is and M. supracorecoideus. �88 In Jackson County, breast muscle mass differed years but did not vary between (Table 2.2). grouse in Gunnison periods. 1983 was high relative to Breast muscle mass of adult male sage County did not differ between Among Jackson among (Table 2.1). early and late courtship Breast muscle mass during early courtship other periods slightly years or courtship County males, leg muscle mass (~ - 123 g, SE 2.8) during early courtship was similar-(f - 0.14) to muscle mass during late courtship (f - 0.94) between differences (~ = early (~ - 102 g, SE - 3.0) and late County. Early courtship 1984. There were also no (~ -102 g, SE - 2.0) leg muscle mass among males from courtship Gunnison 130 g, SE - 3.3). (~- However lipid reserves were similar between populations in 1985, lipid reserves of Gunnison in County sage grouse 4.1%) were low (f < 0.01) relative to Jackson County males (~ - 5.5%). Phase II: Endogenous Reserves Male Display Rates.--A observations) grouse display and Courtship total of 23.6 hours was spent measuring in Gunnison and Jackson f - 0.023). Rates of strutting County males (Fig. 2.2). differences counties. with greater male-female well as among males> Other behavioral Leks in each population Differences populations in strutting (Hotellings r2 - 64.2, display were slower among Gunnison Strutting were most apparent (284 5 minute display rates of adult male sage rates were apparent between diminished Behavior rates in both populations distance. Interpopulation among males within 10 m of females as 30 m from females. differences were apparent between were visited populations. during evening hours approximately �89 G6.2.0.4 - JACKSON 6 GUNNISON 1J.J I- :::l G Z :E .... 5.2±0.1 <, en >- . ...•. ....•.... ......•. ...... "''G , ,, 4 -c .....J 4.2:t 0.2 a... " ,, ,.•. en 25 ,.•. ,.•. 2 .•. , .•. .•. ,, , '8 < 10 2.1±0.2 1.3±0.3 10- 30 DISTANCE TO FEMALE (m) Fig. 2.2. Strutting display rates (K ± S.E.) of adult male sage grouse in Gunnison and Jackson counties, Colorado, April-May 1986. �90 1-2 weeks after the peak of female attendance. not evaluated during evening periods although numbers males were recorded. displaying Male sage grouse in Jackson on each of 7 occasions when evening on S leks. Numbers were similar of males observed to numbers males were observed recorded leks where between grouse in Jackson courtship evaluation (12-4S) In Gunnison County, during morning. sagebrush Male sage County were rarely trapping. Males in Gunnison cover adjacent of strutting displays, to leks. sage grouse in Gunnison in adult male attendance However, populations (E - Gunnison than in the Jackson in Gunnison Variance in daily (Table differed between Peak numbers of adult male sage County leks were more variable County population Lipid Reserves.--Early did not differ between populations In 1986, early courtship weight in male attendance 9.43, f < 0.01). 1986 Early Courtship reserves during morning was similar among leks within populations variance grouse that attended mornings patterns were also apparent between populations. peak male attendance 2.3). 1.0-1.S sunrise. Differences courtship County Preceding County were heard flying to and landing on leks approximately hours before on roosted on leks at night during the males in Gunnison on leks during night-time morning courtship was recorded In each instance only 2 males were displaying However, in vigorous County were observed observed during morning. County typically roosted of displaying during evening display 36 and Sl males displayed season. rates were on only 2 of 9 occasions when evening displays were on S leks. observed Strutting among (Table 2.3). courtship in 1986 lipid (f - 0.27). lipid reserves were 4.1% (SE - 0.3) of live County male sage grouse. Early courtship lipid �91 Table 2.3. Peak daily attendance of adult male sage grouse at 3 leks in Gunnison and Jackson counties, Colorado. Surveys were conducted during 5 mornings on each lek, April-May 1986. County .§.2b Lek Range Jackson Railroad 27.0 0.5 26-28 Ridge Road 17.0 2.0 15-19 Deer Creek 19.6 4.3 16-21 ~.£c - 2.27 Gunnison Ohio Creek 17.8 20.7 11-23 Chance Gulch 20.6 29.8 12-26 South Parlin 19.8 13.7 14-24 ~.f. - 21.4 aMean daily peak attendance of adult males. bVariance in daily attendance of adult males on individual cVariance in daily attendance of adult males pooled across leks within a population. leks. �92 reserves were 3.6% (SE - 0.3) of live weight for adult male sage grouse collected in Jackson County. DISCUSSION Sage Grouse Body Mass Sage grouse body weights courtship within populations. observed that live weights courtship. probably The disparity attributable did not differ between Beck and Braun between their findings to differences in sage grouse body weights large samples (1978) previously of adult male sage grouse diminish results were based on a large sample within early and late and my results are in sample sizes. Beck and Braun's (N - 465) of adult males. during courtship due to variance during Changes may only be apparent within collection periods (Table 2.2). Lipid Deposition and Winter Severity Male sage grouse lipid reserves 1983). Reserve winter severity. size at the begining Early courtship County were highest reduced males in Gunnison winter. March following following increase during winter of courtship lipid reserves mild winters severe winter conditions (> 50 cm) persisted among males in Jackson in 1984. Lipid reserves the severe 1984 from December 1984 in each study area and temperatures averages is affected by in 1983 and 1985 but were County were also low following Deep snows (Remington 1983 through were below 30-year (Table 2.4). Reduced winter result of conditions due to increased endogenous reserves in avian species can be the that limit access to forage resources thermoregulation demands resulting or may be from low ambient of �93 Table 2.4. Winter (Nov-Mar) snowfall and mean temperature departure from 30-year averages in Gunnison and Jackson counties, Colorado, 19831986. Lipid reserves of adult male sage grouse during early courtship. Jackson 1984 1985 1983 Snowfall, cma Temperature departure, Lipids, %c Jackson 1986 l66b 186 97 90 1.2 -0.7 0.2 1~9 -3.1 1.7 2.3 5.4 2.3 5.5 3.6 2.2 4.1 4.2 Climatological bSnowfall Gunnison 1985 102 aData are from National Monthly 1984 163 124 C 1986 Oceanic and Atmospheric Data for Colorado, accumulations Administration 1983-86. in December 1985 are not available for County. cLipid reserves digestive expressed as percent of live weight - weight of tract contents. temperatures (Dugan et al. 1981, Mortensen 1984, Heitmeyer lipid reserves 1985). et a1. 1983, Whyte and Bolan Both factors likely contributed in male sage grouse in 1984. the sagebrush vegetation to reduced In Gunnison County, was covered by snow in 1984 (J. W. Hupp, unpub1. data) while in 1985 and 1986 only 16-22% of the sagebrush habitat was snow covered. both populations Access to sagebrush and in Jackson forced to forage on nutritionally (Remington and Braun 1985). costs during feeding. severe winters inferior subspecies Deep snows increase Distances forage was reduced in County, sage grouse may have been between of big sagebrush sage grouse locomotion exposed plants increased and, based on track observations sage grouse usually 93% of in Gunnison in County, sank 5-10 cm into soft snow while foraging. Low �94 temperatures reduced also increase thermoregulatory lipid deposition. Severe winter weather breast muscle and live weights Early courtship not initiate reduced display activities until late April. (R- 2,764 early April males captured gm). was delayed reserves Live weights in reduced in 1984 because males did It is unlikely in 1984 were an artifact of 20 adult male-s captured that of delayed in Jackson County in gm) were lower (f - 0.03) than live weights or collected It is doubtful resulted to in 1984. collection early courtship sampling. costs and may contribute in the final week of the month of 18 (R - 2,925 that male lipid reserves were larger in early April when body weights were lower. Early courtship lipid reserves County were also low in 1986. December resulted population. Heavy snow accumulations in November and in reduced lipid deposition 'among males in that Winter conditions severe and early courtship males collected of adult male sage in Jackson following in Gunnison County during 1986 were not lipid reserves were similar to reserves of the mild 1985 winter. Lipid Reserve Use During Courtship In both populations, during courtship. mobilized Following in response lipid reserves males in Jackson County County males used Sage grouse mobilization to the energetic of courtship demands of courtship. 66 g of of lipids was likely Energetic demands in male sage grouse are not known but are potentially Daily energetic urogallus) mild winters, 125-130 g of fat while Gunnison lipids during courtship. high. male sage grouse depleted requirements during courtship and are higher of male capercaillie (Tetrao are 2.1 times standard metabolic than winter existence requirements rate (SMR) (Linden 1984). �95 Invariant breast and leg muscle mass indicates were not used during courtship. During periods demand, birds are more likely to mobilize (Reinecke et al. 1982, Mortensen Heitmeyer 1985). investment differences the opportunity to correlate between conditions courtship in that in Gunnison 1986. Slower-display energetic investment County. Higher variance grouse in Gunnison similar between size and courtship County. in courtship in male 1ek attendance may have among some individuals investment courtship differences behavior and by male sage was not could not be attributed to in 1986, the display period with low lipid lipid reserves of males in Jackson were similar to reserves (4.1% of live weight). display was lower among males Due to severe early winter conditions Early courtship (3.6% of live weight) during early rates and lack of evening Although County entered size. similar mild winter of reduced energetic populations, reserve were observed between lek attendance County. lipid reserves. males in Jackson County that energetic lipid reserves provided reserve Despite in courtship behavior from inconsistant reserves. endogenous the 2 populations. could also be indicative unequal in early courtship were smaller among males in Gunnison populations resulted Siz~ in both study areas in 1985, lipid reserves Differences suggest lipids rather than protein could be affected by endogenous Interpopulation behavior of high energetic of lipids raised the possibility in courtship reserves et al. 1983, ~~yte and Bolan 1984, Male Display Rates and Lipid Reserve Male depletion that protein These results County of males in Gunnison fail to support the �96 hypothesis that slower display rates of males in Gunnison County were the result of lower lipid reserves. There are at least two explanations investment in courtship First, metabolizable energy of sagebrush in Jackson resources has been observed courtship in American Gunnison 1985). However, in crude protein County higher crude protein necessarily indicate that metabolizable in Gunnison County. Monoterpene in Gunnison County were higher sagebrush exploited Remington and Braun 1985). sagebrush in the Gunnison Dissimilar differences Gunnison Within a species, thermoregulatory investment by sage grouse in and content does not energy of forage was superior of big sagebrush than concentrations forage found in Wyoming big County (Chapter 1, of Basin may result in lower digestibility. investment in courtship. to interpopulation Male sage grouse in 25-30% less than males from Jackson County. costs are greater among small-bodied As a result of smaller body size, male County may have devoted a larger proportion to maintenance of body temperature. costs may have resulted in courtship than forage The higher mono terpene concentration (Peters 1983:71). their energy budget in (Chapter 1, Remington concentrations thermoregulatory sage grouse in Gunnison quality of food (Brodsky and body size may also have contributed in energetic County may investment content by sage grouse in Jackson County weighed individuals Energetic forage exploited by sage grouse in Jackson Braun 1985). County. to affect male energetic Sagebrush County. forage in Gunnison black ducks (Anas rubribes) County was higher exploited energetic among male sage grouse in Gunnison be lower than sagebrush Weatherhead for reduced among Gunnison some avian species have been observed Higher in reduced energetic County sage grouse. to moderate breeding Males of behavior of �97 dependent Higher upon thermoregulatory thermoregulatory explain why Gunnison remained costs resulting roost sites with overhead to higher and Bradbury from small body size may also cover at night. Males may have selected cover to reduce thermal conductance. in male attendance relative (Santee and Bakken 1987). County males did not roost on leks but instead in heavy sagebrush heterogeneity requirements in Gunnison thermoregulatory (1985) observed demands infrequent Greater County could be attributed in that population. attendance Gibson among male sage grouse during cold, windy weather when thermoregulatory costs were high. The Adaptive Significance of Male Lipid Reserves Among avian species, both the timing of lipid deposition size of reserves pressures accumulated are believed to be influenced (King 1972, Dawson and Marsh. 1986, Lima 1986). effect of lipid reserve size on courtship behavior the timing of male sage grouse lipid deposition courtship reserves lipid reserves immediately are likely adaptive. during winter and reserves prior to courtship. prior to periods is selectively deposition of endogenous 1983, Gauthier Vangilder to breeding demanding reach a winter-spring Accumulation of endogenous reserves prior to periods demanding (King 1972). activity. Avian of high energetic Male sage grouse accumulation fattening maximum reserves (West and Meng 1968, Mortensen is likely an adaptive an Male sage grouse accumulate et al. 1984, Krapu et al. 1985, Mainguy et al. 1986). Although and the size of early favored in bird species demand has been widely observed by selective was not apparent, of forage scarcity or energetically activities and the et al. and Thomas 1985, of lipids prior prior to an energetically �98 Size of avian endogenous selective pressures. positively resource Marsh 1986). population Traits and Thomas lipid reserves winters. Early courtship or forage expressed to stabilizing among individuals Intrapopulation suggest a stabilizing similarities selective were 4.1-4.2% influence and 1986. Early courtship among males within lipid reserves sampling Early courtship in Gunnison periods costs during the courtship county males obtained respectively metabolic season. Assuming (Ricklefs approximately from lipid metabolism rates for gallinaceous a caloric 1974), Jackson conversion of and Gunnison Standard birds can be estimated requirements were approximately demands from body mass body mass in 1986, SMR 143 and 118 kcal/day respectively. season energetic (Table 2.2). of the total energetic during courtship. Given the early courtship for display County males in 1985 1,170 kcal and 590 kcal of energy (Zar 1968). county males, while of male sage grouse were not large to meet a major proportion lipid metabolized in were not highly variable sufficiently 9.0 kcal/g mild of live weight in either population lipid reserves on early years following were 5.4-5.5% of live weight in a in early In each population, were similar between reserves with large County males after the mild 1983 and 1985 winters reserves to be 1985, Dawson and risks associated size in male sage grouse. courtship Gunnison size is believed demands that are subjected are similarly lipid reserves lipid reserve Jackson predation (Mayr 1970:175). courtship by stabilizing size may be limited by greater maintenance or increased pressures reserve energetic (Evans 1969, Mainguy (Lima 1986). selective may be affected endogenous with seasonal Maximum requirements reserves Minimum correlated scarcity reserves Using Linden's in capercaillie for Jackson and (1984) estimate (2.lx SMR), total �energetic demands during a 6-week display period were approximately 12,600 and 10,400 kcal/male counties, respectively. small proportion « Therefore, fat metabolism 10%) of the total energetic grouse during the courtship An adaptive for sage grouse in Jackson advantage depleted reserves to lipid deposition early courtship may be rapidly mobilized likely greatest. selectively attendance success to Lipid short period of in courtship is of lipids would be success during female of energetic demands of courtship. reserves that Size of early to meet energetic short period of female attendance is primarily If males demands when male determined. CITED R. T., and C. D. Ankney. energetics Ankney, accumulation reserves may only be adequate reproductive A1isauskas, investment was affected by the availability during the temporally LITERATURE . during the relatively if male reproductive could be used to meet energetic courtship to meet a large proportion short period. when male energetic Pre-breeding favored may exist even though reserves would be adequate demands during a temporally peak female attendance only a needs of male sage demands of sage grouse courthship. lipids rapidly, meet energetic could provide period. the size of lipid reserves was not adequate of the total energetic and Gunnison C. D. of reproduction 1985. in American 1977. The use of nutrient lesser snow geese Chen caerulescens 55:1984-1987. Nutrient coots. reserves reserves and the Auk 102:133-144. by breeding caerulescens. male Can. J. Zool. �100 1984. N~rient reserve dynamics and molting Auk 10k:361370. brant. 1978. '(:and C. D. MacInnes. _____ of breeding performance selection reserves and reproductive Auk 95:459-471. of female lesser snow geese. 1977. Beck, T. D. I. Nutrient Sage grouse flock characteristics in winter. 41:18-26. J. Wildl. Manage. _ __,.",--_,and C ."E.. :,Braun. 1978. Weights and habitat of Colorado sage grouse. Condor 80:2.41-243. Brittas, R. 1984. Seasonal and annual changes . lagopus. Swedish ". w~:llow grouse, Lagopus Br odsky, L. 11-,,! and.R .•J. Weatherhead. .const ra.Inrs 2n courtship 1985. in wintering in condition of the Finn. Game Res. 42: 5 -17. Time and energy American black ducks. Condor 87..:~3-36. Dawson, goldfinch Physiol. Dobush, 1986. W. R., and R. L. Marsh. and the possible G. R., C. D. Ankney, extraction 1980. R. D. role of temperature in the American in its regulation. and D. G. Krementz. time, and solvent 1985. The effect of type on lipid extractions Can. J. Zool. 63:1917-1920. of snow geese. Drobney, fattening 59:357-368. Zool. apparatus, Winter Reproductive bioenergetics of wood ducks. Auk 97:480-490. Dugan, P. J., Winter P. R. Evans, L. R. Goodyer, fat reserves in shorebirds: levels by severe weather Erikstad, K. E. 1986. and i~cubation and N. C. Davidson. disturbance conditions. Relationship rhythm in willow 1981. of regulated Ibis 123:359-363. between grouse. weather, Cinclus body condition, 9:7-12. �101 Evans, P. R. 1969. Winter fat deposition yellow buntings and overnight (Emberiza citrinella L.). survival of J. Anim. Ecol. 38:415- 423. Gauthier, G., J. Bedard, J. Huot, and Y. Bedard. accumulations 1984. Spring of fat by greater snow geese in two staging areas. Condor 86:192-199. Gibson, R. M., and J. W. Bradbury. sage grouse: phenotypic Ecol. and Sociobiol. 1985. correlates 1985. of male mating success. ~ and C. E. Braun. trapping sage grouse in Colorado. M. E. Hjorth, I. Wintering strategies Ph.D. Thesis, Univ. Missouri, 1970. reference Hohman, W. L. Reproductive to males. 1986. breeding behavior Viltrevy area, Colorado. for related in the upper Mississippi Columbia. 376pp. in Tetraonidae with special 7:381-587. 1975. D.C. drying temperature of Auk 103:181-188. Soil survey of the Gunnison U. S. Dep. Agric., Soil Conserv. Serv., 89pp. Kerr, D. C., C. D. Ankney, and J. S. Millar. on extraction of small birds and mammals. 1972. Methods of female mallards ducks (Aythya collaris). Hunter, W. R., and C. F. Spears. King, J. R. 1982. Changes in body weight and body composition ring-necked Washington, Behav. Wild1. Soc. Bull. 10:224-231. to dynamics of lowland hardwood wetlands Delta. in lekking 18:117-123. Giesen, K. M., T. J. Schoenberg, Heitmeyer, Sexual selection 1982. of petroleum The effect of ether soluble fats Can. J. Zool. 60:470-472. Adaptive periodic fat storage by birds. Proc. Int. Ornith. Congr. 15:200-217. Korschgen, C. E. 1977. Breeding Wildl. Manage. 41:360-373. stress of female eiders in Maine. J. �102 Krapu, G. L. 1981. The role of nutrient reproduction. spring. Lima, S. L. and C. M. Boise. and usage by arctic-nesting 1985. Fat sandhill cranes during Auk 102:362-368. 1986. determinants Linden, H. in mallard Auk 98:29-38. ________ , G. C. Iverson, K. J. Reinecke, deposition reserves Predation risk and unpredictable of body mass in birds. 1984. Annual patterns capercai11ie feeding conditions: Ecology 67:377-385. in the ecological Tetrao urogallus in captivity. energetics of the Finn. Game Res. 42:19-27. Lumsden, H. G. 1968. The displays of the sage grouse. Lands For. Res. Rep. 83. Mainguy, 94pp. S. K., and V. G. Thomas. buildup Ontario Dep. 1985. Comparisons of body reserve and use in several groups of Canada geese. Can. J. Zool. 63:1765-1772. Mayr, E. 1970. Populations, Press, Cambridge, Mortensen, in body compostion ptarmigan Lagopus ~ M. J. Ill. 1986. and A. S. B1ix. 1983. Seasonal and crop content of Spitsbergen hyperboreus. SPSS/PC+ Advanced Ornis Scand. 14:144-148. statistics. SPSS Inc., Chicago, 230pp. Peters, R. H. 1983. The ecological implications Cambridge Univ. Press, New York, N.Y. Raveling, Harvard Univ. 453pp. A., S. Unander, M. Kolstad, changes Norusis, Mass. species, and evolution. D. G. 1979. 329pp The annual cycle of body composition geese with special reference 96:234-252. of body size. to control of reproduction. of'Canada Auk �103 Reinecke, K. J., T. L. Stone, and R. B. Owen, Jr. 1982. carcass composition Maine. Condor 84:420-426. Remington, T. E. 1983. and energy balance Food selection, Seasonal of female black ducks in nutrition, and energy reserves of sage grouse during winter, North Park, Colorado. Colorado State Univ., Fort Collins. ________ , and C. E. Braun. 1985. North Park, Colorado. Ricklefs, R. E. 1974. M.S. Thesis, 89pp. Sage grouse food selection J. Wildl. Manage. Energetics 49:1055-1061. of reproduction 297 in R. A. Paynter, Jr., ed. in winter, in birds. Avian energetics. Pages 159- NuttallOrnith. Club Publ. 15. Santee, W. R., and G. S. Bakken. blackbirds costs. Vangilder, 1987. (Agelaius phoeniceus): Social displays sensitivity reserves of premigratory West, G. C., and M. S. Meng. fat and relation ptarmigan. Whyte, R. J., mallard Wiley, R. H. 1986. brant during spring. 1968. Seasonal Auk 103:237-241. changes of fatty acid composition Nutrient in body weight and to diet in the willow Wilson Bull. 80:426-441. and E. G. Bolen. body condition. 1973. pattern. birds. to thermoregulatory Auk 104:413-420. L. D., L. M. Smith, and R. K. Lawrence. Zar, J. H. in red-winged Impact of winter stress on Condor 86:477-482. The strut display of sage grouse: Behaviour 1968. 1984. a "fixed" action 47:129-152. Standard metabolism Condor 70:278. comparisons between orders of �104 Prepared by Approved by ~ W. ~ Jerry W. Hupp ~ ~-:;.t!a~~::.,:.-_____::C£=-.-I.~:.....e.....~_.:..:_ Clait E. Braun _ �JOB PROGRESS REPORT Colorado State of: Project: W-152-R Avian Research Work Plan: 3 Job Title: Response of Selected Avifauna to Prescribed Burning in the Big Sagebrush Type Period Covered: Author: Job 17 01 January through 31 December 1987 Ernesto J. Hernandez Personnel: E. J. Hernandez, Colorado State University and C. E. Braun, Colorado Division of Wildlife ABSTRACT Preparations for this study were initiated by C. E. Braun in July 1986. Consultations were held with personnel of the Bureau of Land Management and a delay in controlled burning in the Deer Creek area, North Park, Colorado was requested. Treatment (Deer Creek) and control (Railroad) leks were selected. Review of literature was initiated and study designs were discussed by C. E. Braun and D. C. Bowden with personnel of the Colorado Division of Wildlife and the Bureau of Land Management. Despite effects on study design, a portion of the treatment area east-southeast of the Deer Creek Lek was burned in September-October 1986. Male sage grouse (Centrocercus urophasianus) distribution during the 1987 breeding season was ascertained for 10 radio-marked birds at each the Deer Creek and Railroad leks. Distribution around leks was not random. Distance from lek, to feeding and loafing sites varied from the.immediate vicinity to 2.9 km , Vegetation at use sites averaged 6.1% grasses, 8·~6% forbs, and 31.4% shrubs. Shrubs were primarily Artemisia tridentata vaseyana with a mean height of 21.8 cm. Peak lek attendance of males occurred on 5 and 21 April with 23 and 31 birds at Deer Creek and Railroad leks, respectively. Surveys of passerine birds in these areas were inconclusive because of the large number of birds that were not identified. The planned Deer Creek Wildlife burn occurred in Fall 1987 and resulted in up to 70% vegetation removal. ��107 RESPONSE OF SELECTED AVIFAUNA TO PRESCRIBED BURNING IN THE BIG SAGEBRUSH TYPE Ernesto J. Hernandez P. N. OBJECTIVE The objectives of this study are to document the response of sagebrush-dependent avifauna to prescribed burning in the big sagebrush habitat type in North Park, Colorado. Specific objectives were to: 1. Assess the impact, if any, of prescribed burning on breeding season habitat selection by male sage grouse. 2. Estimate changes in relative abundance and diversity of passerine birds in response to prescribed burning. Hypothesis HOI: Male sage grouse use each habitat class in proportion to its occurrence in the total area. H02: Prescribed burning does not alter selection of habitat class by male sage grouse. H03: Relative abundance and diversity of selected passerine birds do not change following prescribed burning. SEGMENT NARRATIVES 1. Review available literature concerning sage grouse, composition of sagebrush, and techniques adaptable to studies of galliform birds. 2. Review available literature concerning burning of sagebrush rangelands. 3. Capture up to 30 sage grouse within the treated and untreated areas during March through May 1987 for attachment of tail clip or poncho-mounted radios. 4. Relocate radio-marked sage grouse from April into August for observation of patterns of habitat use. 5. Conduct up to 12 counts (6/lek) of males and females on the treated and untreated sites during April-May for observation of male strutting display rates and fluctuations in female numbers. 6. Locate nests of radio-marked female sage grouse within the treated area and those dispersing from the study site during May-June. Determine clutch size, nesting success, and nest site characteristics (i.e., vegetation composition, height, cover). �lOB 7. Periodically (3-5 days intervals) locate radio-marked hens after hatching to estimate brood survival from hatching until radio failure or 1 September, and to observe patterns of habitat use. B. Measure vegetation structure and composition on treated and untreated areas from June into August. 9. Establish at least 4 permanent line transects on the treated and untreated sites for estimation of breeding bird response to burning as to species composition, species richness, and relative abundance from April into June. 10. Use night lighting techniques to capture adult and yearling male sage grouse roosting on leks during March-May. Trap and band 200 males and at least 100 females within North Park from March through May. Obtain body weights from sage grouse at time of capture. 11. Collect population data through use of wing barrels and check stations during September-October. 12. Compile and analyze data and prepare progress reports. pertinent findings at appropriate scientific meetings. Present DESCRIPTION OF AREA The Deer Creek and Railroad leks are in T7N, R7BW, SW1/4 S31, and T7N, R80W, SWl/4 S2, Jackson County, Colorado. A physical description of the study site was presented in the Environmental Assessment prepared for the Deer Creek Wildlife Burn in August 1987 (Appendix A). METHODS Sage Grouse Capture and Instrumentation Male sage grouse were captured during April using spotlights and long-handled nets (Giesen et a1. 1982). Captured birds were fitted with tail-clip or poncho-mounted radio transmitters (Amstrup 1980). Radio-marked birds were located every other day beginning 9 May continuing through 1 June to evaluate seasonal habitat use. Shrub Structure and Vegetation Composition Measurements of vegetation at observed use sites were taken using Canfield's (1941) line-intercept method. Thirty 15-m transects were analyzed at the Deer Creek site. All sage grouse leks in North Park were counted during April and early May between 0.5 hour before and 2.0 hours after sunrise. Frequency of counts varied from once every other day to once every 10 days, with the study sites receiving the highest number of counts. Counting continued until 11 May when no birds were observed attending leks. Sage grouse were captured for banding in all areas of North Park. �109 Habitat Use Empirical data (i.e., telemetry locations) were compared to planimetric estimates of habitat classes within a 2-km radius of the Deer Creek Lek. Boundaries between different classes, arbitrarily designated A, B, C, D, and E, were delineated from aerial photographs of the area on the basis of vegetation density (Marcum 1980). A chi-square test of homogeneity (Mendenhall 1971) was used to test the hypothesis that male sage grouse use each habitat class in proportion to its occurrence. Breeding Bird Surveys Four 800-m transects were marked in the Deer Creek Lek area. Following 5 unsuccessful surveys, these were replaced in June with 7 fixed-count stations. The stations were 200 m apart along a north-south axis in observed sage grouse use areas. Three-minute counts were made at each station once every other day between 0530 and 0730 from 1 to 10 July. A l-minute rest period was allowed at each station prior to every count. This census method followed Reynolds et al. (1980) variable circular-plot method. One transect was marked in sage grouse feeding and loafing areas in the Railroad Lek area. Deer Creek Wildlife Burn A Burn Plan (Appendix B) and Environmental Assessment were prepared and .submitted for review. Metal stakes were placed at each of the 4 corners of the proposed burn site. The plot measured 1,221 x 303 m and encompassed an area of about 37 ha oriented along a NW-SE axis about 60 m east of the Deer Creek Lek. Sevent.y-five percent of the recorded telemetry locations for this area were contained within the plot. RESULTS AND DISCUSSION Twenty male sage grouse, 10/lek, were captured between 2 and 23 April. Sixteen solar-powered transmitters and 4 battery-powered tail-clip radios were divided equally between the 2 study areas to instrument the captured birds. Twenty-eight locations were recorded in the Deer Creek area between 9 May and 1 June. Locations were not systematically recorded prior to that date. Of these, 25 were within a 1,600 x 300 m strip immediately east of the 1ek. Two locations were 1.4 km south and a single location was 1.5 km southeast of the lek (Fig. 1). Six transmitters were recovered from birds marked at this 1ek (Table 1). Sixteen locations were recorded 2.9 km southeast of the Railroad Lek and an additional 7 were recorded 0.8 km east of the 1ek (Fig. 2). Two of the latter were immediately following morning 1ek dispersal and these sites possibly constituted staging areas prior to daily movements away from the lek. Two tail-clip radios were recovered in this area. The status of radio-marked grouse in both study areas is summarized in Table 1. No attempts were made to locate radio-marked birds during the last quarter of 1987. �llO . t-."" - c' -', tz: "<, e, ''''. 8405 I-------~--------~------+ -< ' \ ", '- , ,) 25 ,, 29 , ~ " "~ "'~3i'+_---',..... • ~~ : .•. ,. 36 ~ . .~~\)I ..•. .::::.:::;.:::.::::. ,. " " v I' ", 1/ " /: /: •• n , v n • '_ II \ .,.....1 6 , V' a "u \I If " e>- " ~, I' ,; " " 1/ 1/ \ -. ,-, \ ~~o ~ B:58;'\-_ - 4,,:- ;.../" "t""·,\... i Fig. 1. Male sage grouse use sites, Deer CreeK Lek, Jackson County, Colorado, spring 1987. �Table 1. 1987. Status of radio-marked male sage grouse in North Park, Jackson County, Colorado, 30 September Frequency Radio type ABe Band Capture site (lek) 150.811 150.836 150.862 150.925 150.951 150.966 150.979 150.999 151.039 151.236 151. 256 151.268 151.387 151.417 151.464 151.505 151. 525 151.545 151.564 151.825 Poncho Poncho Poncho Poncho Poncho Tail clip Tail clip Tail clip Tail clip Poncho Poncho Poncho Poncho Poncho Poncho Poncho Poncho Poncho Poncho Poncho 12+ 12+ 112+ 2+ 2+ 2+ 2+ 2+ 11112+ 12+ 2+ 215 119 214 120 213 121 117 116 792 115 123 186 188 187 212 131 759 122 135 171 Railroad Deer Creek Railroad Deer Creek Deer Creek Railroad Railroad Deer Creek Deer Creek Railroad Deer Creek Railroad Railroad Railroad Deer Creek Railroad Deer Creek Railroad Deer Creek Deer Creek Date of capture 23 05 23 05 22 06 04 03 03 02 18 16 21 21 22 13 15 06 14 15 Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Date of last observation 07 No 02 No No 07 No 25 24 27 12 19 25 27 02 26 02 26 17 17 Ju1 observations Ju1 observations observations Apr observations Apr May Hay May May May May Ju1 Aug Ju1 Aug Apr Apr Status Recovered, shot 13 Sep Recovered, 09 May Active Undetermined Recovered, 09 Maya Recovered, 28 Mayb Recovered, 18 Mayb Recovered, 18 Mayb Recovered, 26 Mayb Shot, 13 Sepc Undetermined Undetermined Undetermined Undetermined Active Active Active Active Recovered, 26 Maya Recovered, 12 Mayb aTransmitter attached to carcass. bTransmitter not attached to carcass, nor any evidence of carcass. cLocation of radio unknown. I-' I-' I-' �112 ., I ' eterson ;"i:~aru. , ~ ~:. __ ._..._--- =:.-: :: ~ / I .::::-::;;:::.=::=:::.::=:::... ' 36 I 31 I I f'.:: <_: __ T.;_;.;....:::8::;....:cN~.:._ +8°-._.,.....~;----------------- --1- ~5:.:.:-------"H,-ac-::kT:le:":'"Y.-5;;;:: I T. 7 ['" • J 1 ----.~~. =" •..• Ranch R.80W.! ~.79 w. : / ~ o I ------- 2 .,»: f 0.,"- a r- ; • eO RAILROAD LEK I 1 • • .0 ~. '. 1.,/ 6 ,, I c I I i " _ (,':'--- __________ , =".:7i-+· __j.,__ I +8 .•. .,50 _~=::;c.=.,...". - [>: i ~ " ~o , .0' ~ J:- /7 \,': .' 12 J , y' .. ':< - .- _ .. --- - - - ._-_ - - - - - i I •• e: ;' . 7 ! ,~ • i{;lOch ; - •• r- "Gr,,( . ( ':. ,..;. -- <, ) .;:/ , 11 .' t ( ~./.' - +b.d, ,.:, -'-"'---.,_...._tt_"D •..,--J • • \ \ ?1'" __ "'I . ',,_ c- Fig. 2. Male sage grouse use sites, Railroad Lek, Jackson County, Colorado, spring 1987. �113 Analysis of 30 vegetation transects in the Deer Creek area gave the following composition: Grasses 6.1% Forbs 8.6% Shrubs 31.4% range 1.5 - 13.3% range 0.8 - 16.8% range 17.8 - 54.7% Mean shrub height was 21.8 cm (range 11.7-35.3 cm) with an average of 19.5 plants per transect. Shrubs consisted of Artemisia tridentata vaseyana. Peak male attendance displaying; peak hen observed. Peak male birds observed. Hen at the Deer Creek Lek occurred on 5 April with 23 birds attendance was recorded on 21 April with 19 hens attendance at the Railroad Lek was on 25 April with 31 attendance reached a maximum of 48 on 7 April. Telemetry data indicated that male sage grouse did not use each habitat type in proportion to its occurrence within a 2-km radius of the Deer Creek Lek as all locations were within type B habitat. Planimetric proportions estimated for habitat types A-E were 0.1761, 0.3211, 0.3225, 0.1749, and 0.5400 respectively (Table 2). The calculated test statistic was 40.7 and the chi-square value (0.01 significance level, 4 df) was 13.2767. The hypothesis of homogeneity of habitat use (HOI) was rejected. A total of 84 passerines was recorded during the Railroad Lek surveys. Of these 69% were identified and classified by species and distance from observer. Distance estimates were also made for 26 birds that were not identified. Eighty-six observations were recorded in the Deer Creek Lek area of which 55 (64%) were identified and classified (Table 3). The proposed Deer Creek Wildlife Burn was conducted between 21 and 24 October. Sagebrush removal was estimated at 60-65% of the southern part of the area and <50% of the northern part of the area. Assisting with the burn were personnel from the Craig and Kremmling offices of the Bureau of Land Management and the Colorado Division of Wildlife. Most of the burning was accomplished by using drip torches. Subsequent attempts to increase the burn efficiency were made by repeatedly crossing the burn site with a propane torch. These took place over a 3-day period and were only partially successful. Total sagebrush removal following the use of both methods was <70%. CONCLUSIONS Telemetry data from the Deer Creek area suggested that male sage grouse distribution was not random and may be influenced by vegetation structure and composition. Proximity to the lek did not appear to be important as evidenced by comparing distances recorded from the treatment and control sites to feeding and loafing areas. Although the Deer Creek habitat use data suggested that the hypothesis of homogeneity (HOI) could be rejected, without more locations and extended (2-3 years) pre-treatment observations it was not possible to establish whether the observed feeding and loafing areas were consistently used during consecutive breeding seasons. Continued observations at Railroad Lek may provide information on this question and analysis of habitat preference for that area will become more important. The effects of the 1986 Deer Creek burn on habitat selection, if any, were unknown. �Table 2. Observed and expected values for planimetric estimates and radiotelemetry locations for male sage grouse, Deer Creek Lek, Jackson County, Colorado, 1987. Habitat class A Planimetric proportion Sage grouse locations Totals C B D E Obs. Exp. Obs. Exp. Obs. Exp. Obs. Exp. Obs. Exp. Totals 17.6 0.0 13.8 3.9 32.1 28.0 47.0 13.1 32.3 0.0 25.2 7.1 17 .5 0.0 13.7 3.8 0.5 0.0 0.4 0.1 100 28 17.6 60.1 32.3 17.5 0.5 128 I-' I-' .j::- �ll5 Table 3. Survey of passerine birds on the Deer Creek and Railroad breeding bird transects, Jackson County, Colorado, 30 June-lO July 1987. N Species observations Deer Creek Sage thrasher Vesper sparrow Western meadowlark Horned lark Brewer's sparrow Unidentified Subtotal 10 24 8 3 10 31 0-10 Distance (m) from observer 51-100 26-50 11-25 2 4 1 10 4 2 1 2 13 3 3 1 3 1 9 8 8 2 4 16 86 Railroad Sage thrasher Vesper sparrow Western meadowlark Horned lark Brewer's sparrow Other Killdeer Common raven Swainson's hawk Brewer's blackbird Unidentified Subtotal 21 12 7 7 7 1 1 1 1 26 21 6 7 4 5 1 6 10 1 1 1 10 84 Surveys of passerine birds were inconclusive. Observer inexperience was an important factor as it resulted in a large proportion of unidentified birds. The methods selected for the surveys were consistent and reliable; differences in rate of travel as well as distractions from noise made during travel were eliminated by making counts from fixed stations. LITERATURE CITED Amstrup, S. C. 214-217. 1980. A radio collar for game birds. J. Wildl. Manage. 44: Canfield, R. H. 1941. Applications of the line intercept method in sampling range vegetation. J. For. 39:388-394. Giesen, K. M., T. J. Schoenberg, and C. E. Braun. 1982. Methods for trapping sage grouse in Colorado. Wildl. Soc. Bull. 10:224-231. �116 Marcum, C. L. 1980. A non-mapping technique for studying habitat preference. J. Wildl. Manage. 44:963-968. Mendenhall, W. 1971. Introduction to probability and statistics. Duxbury Press, Belmont, Calif. 466pp. 3rd Ed. Reynolds, R. T., J. M. Scott, and R. A. Nussbaum. 1980. A variable circularplot method for estimating bird numbers. Condor 82:309-313. Prepared by §Jt,~ 9. ~CU\~ Ernesto J. Hernandez Graduate Research Assistant Approved by -=-~a..,.:.· -;~;---;';-~_ ..!...~ _ ___: Clait E. Braun Wildlife Research Leader _ �117 APPENDIX A PRESCRIBED p ~.~c. FIRE PLAN .... G..c..?~.k.... J".:L.),..:;:I.l:i,f. ..~ .....~.t"A.c.n. Burning 15.c.~n.Jn.JJ:i,.n.9 R.~.:?.9.l,lc..c,:::~ Ac.~.1?. Unit Resource .............. _._G.C1?:i,9 _ Area . District Pre par ed by: .... I.~c.r:X.... J"'.,_ ... H9.t..t..J ..~.:?. _..._. . . . . ..' Colorado State Forest Service Reviewed Reviewed by: ~ cu1i'<l(j Area I2~L .. Mal;'~;~;;'jlr. The approved Prescribed Fire Plan constitutes the authority to burn. No one has authority to burn without an approved plan or in a manner not in compliance with the approved Prescribed Fire Plan will be fully supported. Personnel will be held accountable-for actions taken which are not in compliance with elements of the approved plan regarding execution of the objectives in a safe and cost effective manner. This project is rated COMPLEX INTERMEDIATE _.~ .. NON-COMPLEX, pursuant to prescribed fire guidelines Estimated Approved Cost Per Acre: by, ~ f)~t: $.::? ,...QQ....1:: q ..._l'.4-_, ...:?'.Q. /%~ Distri~Manager �, ' ;-t,;. ",::118 ~ SERVICE Steamboat Springs District P,O, Box 713657 Steamboat Springs, Colorado 80471 October 12~ 1987 Deer Creek Wildlife Burn Narrative Depending on the wind direction'the day of the burn you should start on the leeward side of the block and in a corner and black line 15-20 feet in- from the beaten line. This should be done all along the leeward side or sides of the block. Paying close atteJil:tjonito~~\'lind direction 'you can take progressively wi der strip's with the firing crew working from the leeward side of the block toward the windward side. �119 Proj8ct: ~ P~oject * DEER CREEK WILDLIFE BURN Location T7N R78W ~ County: Sec. 30, 31 ~ 32 Jackson Purpose of the burn: To eliminate sagebrush from saqeqrouse This is a cooperative burn between DOW and BLM to evaluate the-impact of rescribe burnign on seasonal habitat on sagebrush. A. Haza~d Reductiorl G. Silviculture c. ~3it e F'~ep C'<T "'.t i on ~. __._1?.I~_!?.P_.1>~§:.LP.gI_=i_.I!!..~.t~.I .:.; "./ \,' \/ A~AnAA \; . _ .'./ _ ~. Insect/Disease F. Range tJ.. F:p ;::':C i ~~.J. i:1~).ni pLt 1 .:7.).1:.i I-:::;rl H. Oth2!"- C3P2Ci i:y. ) Fire Objectives: sagebrush throughout the burn area. The _ _ I) .. II. Evaluation lek COlltrol _ Management T2chniques Division _ . ._... . . .. . 1) Kill 90% or more of the (Give Description). of Wildlife will evaluate the burn using their or,·.Jncrii.:e,i~.:i. ;. " * Ill. .~. Burning Permit t:ol_lI-n~ ["4 __ ._- Fall ------------------_._-_._-------_._-------------_. --------------------------------------.. .... ...- ~..~ ) ' .' . �120 ~ A. Type Presc~ibed of Burn s~gebrush burn ·r 1,:::-.( l...._J.:.. Fuel (T I -5':) .__ )L~ ..__. (s) j·lode.l D. Fuel Loading (TA) E. Fuel Continui~y varIes F. Fuel ~;c",\ttC'r ·:?o Arrangement 1. Fuel o - 1/·4'1 with light vary understory grasses. :1. :20'1. :- I(:'! II '-:' DE: 1-;: P 'I::. h Cj -;: + or ranqe I) !.). { r'- ,';:tr-j q 'i~ F~' N/(~'I ["f'-;:;".n~:'i.gi"::.:'rnE'nt, conditions, trends, E:(j'C t orn S 1 op .;(-c) to 50% ":--., :1. bi 10% brush DistributiGn~ :l / .~.~ -.._ ~;" ._ G ~,: i ["'I C 1 Ltd (? 1- ":".n~J;2 present t i ort OJ Tor2qe production, e d.r'';'~·I.l\;i':; i·:::;. annU21 i r.-~: c:;r rn·:1 85!)(} .. ~) --~.3~.~ on Slop2 ;_:C)r·;_:;idl_::;I'·.:~.tj_Dn·;:; --------- N/A -------------------------------_. Due to fuel •• --.-.-------- _.__._.__ ?_b..Qy_~_?_y.~E.~e_::.~~rl_cl._.~p.~_~c!~_.:. continuity, :o. w i Ll. need • _ . _ -----------------------------------------------------------_ . ~ . . ." ..;~. �121 v. Durn Prescription T i en :.::: •..., .., •. , "J .{.: :_, ~::;. }' 3. uays Since Measurable . Moisture .-:.r:::: "::"','_' cc- •....' .. cl~ar du: ;:'::"i< PC:'Ct.E';(] r :i. nq to p2r~ly b uvn i. nf:j P(,·:,'j--'~ od ~ although cloudy! b) Fire Behavior flame It E.' lengths in .----_ .._----_. __ .__ .__ ._-_._._ ..._ .._-_ .. w i t h fronts w i nd s associated would be more desireable. FL, D€·::~:;i:-·E·'d 'iesd: f::;Ur-n h:c?·=:,:_:.i.·c.s Due to the time of the year, to burn the entire unit Hill be attempted. no fronts an attempt Narrative acceptable -_ _----_._ .. .. _ .._--._- ---._---_ .... __ .__ ._--- --,_--: , J, Ac' :._.- --_-_. __ ._---------_._ ..._--._ : ------_ ., 85 of .-, .LL i-ltJ. ITI i d i t. ~l .-:: 1''::; 15 c:: c· !..~ -j, II '. ; i !.._ .;_ -_._-----:------------- .-. i I I I I .. �122 : CI :"', -',' ., _--_ ._.-.-- _ _-'.-._ ..- _ .. : _-_, .._-----_ _ ....}_?2~_.__.::'-'-.,-~._-.. _?_~.~__. _ sage L...1 \.:C: 35% .l ()" ; , ;.•..} 1. .2t) ..:~ " 3. Stabil~ty ~2ctor '0:' ]. .... -;- ~:·;n]. 1. wind .... ";.... ; ,... . in i ].1 :..:.! j_ -::~t:,I~.:;~._:..;_:: The permittee C:i !<:;;.i n -:: / /~. Line ~Q -i ·,·.· .••1••••• \ J. '_I ::}j. t i::': (:"i r, t t"; ,:~ L '-~;;. .: : -'.-.\/ : - c;.:':'i.U concerning the use of water -f r: CJ en '!:,', h .::.:.~.:ur n to be contracted if';:i.]. E·:' will need l_. __ . : : " Construct •...._. .1- ,- . ',' ::"_._'; i,··1 C'Il ...!j. Prepar~tiGn i., .~":_';. _ J'. f._= _ i_,'. _ N!A _1... _....• '1_·· .: ;,. ,._"', . " ··1·_-:::' _ '_'. ':~' __ _...,'....~.• ,,-,·1 u _- _---------_._._- I·· C'" ••... __ j -- --_--- -.; 1,'.1 L'r- 1• _._----_-.- \/i:::' I" q ~.~:-..: , 1 • fii;::l'l: E't£5) • D !_:.c. ';;.-;~·s· T. I ---------------------- ------------------------------------------_._---------------------.--_ ....• ----_.-_.--_. Du;- II in O __ ._.- ..-- - .__ .- ---_._-- ----._ __ ..._----------._ .. . !\I/ n__ .. ._.._. . . .. ------_._-------_._----------- ..__.._. ------------------------- ------------------------------------------------------------------- �123 i;:~~ ~ c~i r 'j i.::.i ",. __ .•. __ ,_._ ,_ •••••••. _ •. _. ,,_, __ • ,,_ .•. *. _ •. r-.~,./ '-;1 ._ . __ .__ •__ ._ .• _ .•. _ •. _ .• ••.•. _. __ • . ", .__ ._ •. _. __ • I,.' -, .(..) ••. _ .••••. _ ..• _ .••. ,. T .L. .. _ ._. _ ;'-'j ---------------------------------------------------~-------------- ',::; i)" -;:-;. F' j..•.•C: p '';:~' ;=,:.:;~ '_ .. *2. Qurniny ~':3 • l-;(J 1. ,j i .r, '._I I ~ _ 1~ I~ {-l -1 1-:1 .. . _ ------------------------------------------------------------------------------------------------------------------------------------_i_~__9:~t_i~_iE~_t_~_~_: . ·,o ; : _: :,'.j i.:. common to al~ __ '--.': . .-, - •• !.... __ ,,_. .__. __ "C ,_ •••• • ::_~(~ H. _ . •••••••• 0 •• i ..:.t:.::.iI I ._.__ . . -------------_ ----------'.! v~hi~L~~. "\, -' �.. 124 1"', •.•,' ' •• ".- . :i .;. \/ s: ~::,:~ .1. • • '. j i ! -! cJ . i.._: C) n ,t..;) C t; COf:lP J. et {2d (initi.:"ls Jackson Ca. ~~ar 723-4404 i-" __ i_.,J...,J 2. Fire • Department ------------r-'! t., ,L.._! !.,.I ! -/ ~2.:~-_.,_','~.-'.- -------------- �125 " i···.ic.,j··· t ~'! .: .. i ~::t.:...., .... C'·: . i ·.••1 ~•• " :3 • C)t. h ('-2 ;, .. 'c.. West2~n Slope Fir2 Center 2~5-4088 Actual Obs2rv~d/Measur2d I~~ni [)C). )-.rs :: (:n _ ~GPY if avail.) t; i inc r::.' Forecasted W2athe~ (Burn Day) (Attach actual forec23t ~~il;I-I~_ ------------------_._------------------------------------------- ----------------------------------------------------7---------------------------------~---------------------------------------------------------------------------------------------------t,c: t. ;..:.·~;).l (F'CJll ~;.Jf:::' E." D •.... -!:L nq t roo, '7.~ 1""- ( 1::; u. r- n i",! i q j'-, 1: ./ I).~';\ .~/ ) f' c:' (- i, c} d ) __.,,_._ __.. _ _ ..~.. . ------------------------ ----------------------------------------------- . "r:./~ �126 .i. Holaing Plan 1 i nr:::s:. dU.r·· j. nq will be *Threat 'j 1_; .1. -:::1.,.... .- i t: and ~sed to patrol Contingency Fire '-~- '.-..\_, ~-.' :-...:. ,::, - linG 23 black line be fence Trigger F·: 1. i ....':":, i.':.p Plan and Property would t -::':.1 no life threatened, lines. Mechanisms and Confirmations Constant t~i:..tl"· 1'"1 the progress. of Life *Escaped . ,!":, in reserve Fire thrc2ten2d __ _,_ _!. ;_. }.i ESC~P2d ':.:,:j'L. ~ ! Narrative held operations f:.::·l"l radio contact should be '! "-'.'.1. -------------------------------------------------------------------;'·(1dditi.Cln.::ll E~:;c-3.pe :3uprwession F\f2~.GU;--·cCCj !:.iit:·, Loc,,:d:iGrl Lo bc::' C=."dl&~d for Fire �127 .' c:::· .•...·-.• ...1 ·; 1'.-,;;',11' " -,:.:..,. .. .~ :i .:;.~<:.; ,-.ill s: '•••• 'Il ,,. . , , '_1 n -----------------------------------------------------~-----------]_ J. ·1 -;~ J. -_'" .• !:.:;' .L ,_) '" �128 +', i.::) I ~Jn i 'i:. i t: n :~:~ f.:! q U. (-:~,r'i c: ~~:, .:~;. n (1 o \/;:.;:. t- I .;:;" '/-::;; ) ;-: i: J_ d j. r', g ~~., 1 .:;(;::_ c: :T1 C:;', r-;-;:: >1.:;;\ p ~:'. :,- . .. ", ',~ . ';._:.. ~~ :~ _ �129 \ \ ----~----~7_----+------ _ 8405 25 29 " ~ __:_~:·':8.'?6 I,.• " _-' . "T. 7N. ,, \ I I .. , ~~J6 • , \', ."0: I I '~:. ,• • ..• T. 6 N. I' I I . • • ..• •, c "" I / I , I' • I f , , , .. ·•, • I I • •• ,,I • • • ,, . 6 -~ 5 .." •· _. i,J... _--.~ ~ ~~-~ r toO - --'" ' .... S •••. 'lLW'Ay ~-', .:.S Raervoir ~OO~ " . ~.':'.... ",' •.. • •••• ##- •••• '-:-» .';541 • . ~ ' , .•• •• ,,",.- ' • J .•. t •• .' \' ~ L_ .. -",-~-,-, I .' I· ~ ,'b" \.. I .",,'" ~o •.... ." . '" I I ,- 't.. ~ . •• ~ '. \ ~ _.#( .- Appendix I - Location of the Deer Creek Wildlife Burn, North Park, Jackson County, Colorado, Fall 1987. (From USGS 7.5' series, Owl Ridge Quadrangle Jackson County, Colorado, 1955). ...•....•. ._ • �;' 130 APPENDIX B DEER CREEK WILDLIFE BURN Background This environmental assessment will analyze the impacts of using controlled fire to remove sagebrush structure on one site in North Park, Colorado. The objectives of this action is to evaluate the response of sage grouse to burning of habitats used during the breeding season. I. Purpose and Need The purpose of the proposed action is to remove all of the above-ground vegetative structure at observed sage grouse use sites. This is desired to evaluate the impact of prescribed burning on seasonal habitat use by male sage grouse. Accomplishment of this objective will provide information that will aid in developing a holistic management plan for sage grouse and big sagebrush habitats in Colorado. II. Alternatives A. Proposed Action A single unit of approximately 92 acres is proposed for burning in Fall 1987. The action would consist of constructing fire lines around the unit and strip firing into the wind. Firelines would be constructed using a rubber-wheeled vehicle pulling a rotary brush beater. Since it is desirable to have fire lines 24-30 feet wide, it will be necessary to make 3 passes with the brush beater. A brush rake will be used to clear the cut brush and pull it into the burning unit. Fire lines will be constructed around the entire unit. Firing techniques will consist of people on foot or in a vehicle lighting the fire using drip torches. The area would be burned by first firing the downwind edge of the unit to create a black line. The next strip would be fired 10-15 feet upwind of the first. Strip firing would continue in this manner until the entire unit was burned. Post-fire management of the area consists of resting the area from grazing by livestock for a minimum of 2 growing seasons to allow for establishment of grasses. No post-fire seeding is necessary, since the area currently has a good stand of perennial grasses to supply a seed source. A 60-100 m wide vegetation buffer zone is desired along the periphery of the treatment area to assure adequate snow deposition and moisture, and to help control wind erosion in the treatment area. A total of 92 acres is identified for burning as shown on the attached map (Appendix I). Within this burn site will be 3 circular plots of 20-m radius that will be used as controls for vegetation measurements. These unburned areas are estimated to comprise 1.0% ~f the total treatment area. �131 B. If the proposed action is not selected, removal of vegetation structure can be accomplished by brush beating and raking the proposed unit. This is expected to increase the cost of treatment, however, and the effects of fire, as a management tool, on sage grouse populations will continue to remain undocumented. Nutrient release associated with burning cannot be expected via this alternative. III. Affected Environment/Environmental Consequences A. Proposed Action Vegetation in the area of the proposed action consists of an overstory of mountain big sagebrush, with an understory of perennial grasses and annual and perennial forbs. Approximately 92 acres of this vegetation will be altered as a result of the action. It is expected that 99% of the sagebrush within the burn units will be killed, and will take many years to regenerate. The grass and forb community will be depleted temporarily, but within 2 years following the burn should respond to the release of competition with sagebrush by improving to a higher level of production and vigor than that found prior to the burn. The proposed project area provides habitat for a variety of wildlife dependent on the sagebrush ecosystem. The more important species in the area include sage grouse, pronghorn, ~ule deer and raptors. Several species of small birds and mammals also use the area. The proposed burn will eliminate sagebrush habitat until new sagebrush re-establishes. Feeding and loafing habitat for sage grouse will be eliminated until sagebrush is re-established. Understory vegetation in the burn units is expected to flourish and should provide highly nutritious spring forage for mule deer, elk, and pronghorn. The increase in understory should also provid~ livestock forage away from the riparian zone associated with Deer Creek. Soils in the treatment area are of the Cabin sandy loam series with medium runoff. The risk of water erosion is moderate and the hazard of wind erosion is severe. A map delineating the soils of concern is in Appendix II. Air quality in this area is excellent. During the burning process, quality will be affected adversely. This subject will be covered in greater detail in the Deer Creek Burn Plan, which will be prepared prior to burning and coordinated with all concerned agencies. Immediately after the burn, infiltration decreases due to the water repellency of burnt organic matter. When precipitation occurs on wet soil, infiltration reductions cause high runoff and a doubling in sediment loads. As grasses come in, infiltration will recover and runoff will be reduced. As most runoff occurs during spring melt, before grasses are established, the first year will see increases in sediment, phosphorous and possibly potassium. If �:' 132 grass cover occurs, the area will re-stabilize. (The intensity of the burn will greatly determine the extent of impact to soil and water). BLM public lands identified in the proposed action are within the North Park Extensive Recreation Management Area (RMA). The existing physical and social setting for recreational use are those described for a Roaded Natural Opportunity Class as defined in BLM Manual 8320. This area is characterized by a predominantly natural environment with moderate sights and sounds of man. Resource modification consists primarily of numerous access roads and trails, and range developments to benefit grazing. Springs, reservoirs, fences, and vegetative manipulation (a burn area to the east) are noticeable throughout the surounding area. These developments generally harmonize with the natural environment and indirectly benefit recreation resources by dispersing grazing activity and by protecting riparian habitat and water resources. The predomiant recreational activity is hunting for pronghorn, mule deer, and sage grouse. Associated with hunting is the use of Off Road Vehicles (ORV's) and camping. Use is moderate during these hunting seasons and indicental throughout the remainder of the year, with exception of years of heavy snow accumulation suited for use of snow machines. Concentration of tourists and sightseers may be high on the neighboring Colorado State Highway 125 during the winter ski season and summer vacation season. Travel on Jackson County Road 27 to the east may be moderate to high during the summer. The proposed action may create short-term impacts to recreation resources, but after establishment of perennial grasses and an overall vegetative cover, opportunities for hunting will be improved. By improving rangeland, dispersing grazing activity and enhancing wildlife habitat, the recreational setting and experience in a Roaded Natural Environment will be of a higher quality than before the burn. During the burn and in the time it takes to establish ground cover, scenic qualities will be significantly impacted and recreational opportunities, especially for hunting and sightseeing will be lost. The temporary loss of of public lands for recreational use during this period will not significantly affect opportunities provided on the majority of public lands in the North Park Extensive RMA. Public lands in T 7N, R 78W, Sections 30, 31, and 32, have been inventoried and classified as Visual Resource Management (VRM) Class IV. Resource modification in a VRM Class IV may create contrasts that attract attention and are a dominant feature of the landscape in terms of scale. Changes should repeat the basic elements of line, form, color, and texture inherent in the landscape. The entire area encompassed by the unit described in the proposed action consists of a flat sagebrush bench. VRM Class IV areas are of "moderate" sensitivity since these areas are seldom seen from major transportation condors. The entire area is at the low end of the Bcale for scenic quality. �133 Short-term impacts from the proposed action will be most signficiant in terms of visual resources. During and immediately following the burn, the landscape will shift to VRM Class V. This is an interim classification for landscapes in need of rehabilitation or where management practices have reduced landscape quality. The burn will create obvious contrasts in line, color, form, and texture. However, upon regrowth of vegetation and establishment of native grasses, the visual quality of the landscape will shift back to VRM Class IV approaching Class III. In the long term, visual resources could significantly benefit from the proposed action. The removal of overstory big sagebrush and regrowth of lush native grasses will change a rangeland landscape to a pastoral landscape which may be construed as more scenic compared to the surrounding sagebrush covered. In turn, diversity in vegetation contributes to visual quality. A Class III inventory has not been performed in the proposed burn area, however, there is an unsubstantiated report of a cultural site (AR 134) somewhere within the NE 1/4 NW 1/4 of Section 32, T 7N, R 78W. Prior to burning, a control test plot should be established in an area where cultural resources occur, so that the effects of burning can be evaluated on open lithic sites. Following the burn, the control test plot should be re-examined to assess the effects of the fire on lithic sites. Additionally, no burning should be planned in the vicinity of historic structures. Firelines should be given Class III inventory prior to construction due to surface disturbance caused by raking the brush into piles. B. No Action If the proposed action is not selected, brush beating will temporarily eliminate sagebrush habitat. Sage grouse feeding and loafing sites are expected to be eliminated until re-establishment of sagebrush. Understory vegetation is expected to flourish and should provide adequate spring forage for mule deer, elk, and pronghorn •. It should also benefit livestock provided the unit is kept free from livestock grazing for a minimum of 2 growing seasons. This alternative action will create slightly lesser impacts to recreation resources than treatment by fire, but the same consideration (i.e., resource inventory) should occur. The impacts from the increasing use of fire as a management tool, to sage grouse populations will remain undocumented. C. Critical Elements The following critical elements apply to both Alternatives A and B: 1. Wilderness - the area of the proposed action does not lie within existing wilderness or study areas. �134 2. ACEC - the area of the proposed action does not lie within an Area of Critical Environmental Concern. 3. Cultural Resources - see III.A. 4. Visual Resources - see III.A. 5. Threatened/Endangered Species - none is known to be in the area of the proposed action. 6. Prime or Unique Farmland - the area of the proposed action does not lie on this type of land. 7. Air/Water Quality - see III.A. 8. Flood Plains/Flood Hazard - the area of the proposed action does not lie in these areas. 9. Socio-Economic Impacts - no impacts. 10. IV. The proposed project is in conformance with the Kremmling Resource Management Plan. Mitigating Measures A. Proposed Action 1. The unburned buffer strip surrounding the unit should be at least 190 feet wide. 2. If sufficient natural seeding does not occur within a reasonable time period, artificial seeding should be done. 3. Rilling is not expected to occur. If excessive rilling occurs, watershed rehabilitation work will be considered. 4. Water sediment loads will be monitored by a hydrologist after the second year of the burn to assure stabilization is returning. 5. Following completion of the burn, re-evaluation of impacts to visual resources will be done. If it is determined that the burned area is highly visible to the user public, interpretive signs may be erected at suitable locations explaining the purpose and results of the action. 6. A Class III (100% pedestrian) cultural resource inventory shall be completed prior to construction by a qualified professional archaeologist along all constructed firelines. A report of the inventory will be submitted and approved by the BLM with stipulations as appropriate in order to comply with EO 11593 and Section 106 of the National Historic Preservation Act of 1966. �135 7. B. If, in its operation, the cnaw discovers any cultural remains, monuments or sites, or any object of antiquity subject to the Antiquity Act of June 8, 1906 (34 Stat. 225; 16 U.S.C. secs. 431-433), the Archaeological Resource Protection Act of 1979 (Public Law 96-95), and 43 CFR, Part 3, they shall immediately cease activity and report directly to the BLM Area Manager. The Bureau shall then take such action as required under the Acts and regulations thereunder. Brushbeating The above measures 5-7 apply. �136 c- •.•• ..~ R .75 'II. .:;:.~ 25 f L~~- I ....::. -'. " '" ~. 8438+ " c, \\ ~ ~ _ " ··~·8.•.• 1 'v"'" ,~' ,\ ) . , 'I " \\ \\ \\ \\ . ,/ 36 T. 70.1. " " II ", 1/ 1/ " '--" T G N. 1/ \, II '- , \ "" " " " ..,. •..'" •... Appendix I - Location of the Deer Creek Wildlife Burn, North Park, Jackson County, Colorado, Fall 1987. (From USGS 7.5' series, Owl Ridge Quadrangle Jackson County, Colorado, 1955). �137 , . 1 -\ \~\~ <, f Ca 25 Ca C Ca GeD C. Appendix II - Soils map of the proposed Deer Creek Wildlife Burn, North Park, Jackson County, Colorado, Fall 1987. �.; 138 SOIL SURVEY TABLE Map symbol Ab Ag B4 Bg BI< Bn Bo B. Bw Bx By C" Cd Cf CoD CoF Cr C. Ct CyF Do Du Eo EhE Fe Fh Fo GeD GkE Gn GoE I.-Acreage and proportionate extent of the soils Soil name Acres Aaberg-Barishrnan association _ Agneston gravelly coarse sandy loam, 5 to 40 percent slopes _ Badland _ Bangston fine sand, 1 to 10 percent slopes _ Blackwell loam _ Blevinton sandy loam, 8 to 20 percent slopes _ Boettcher-Bundyman association Bosler sandy loam _ Bowen gravelly sandy loam, 15 to 45 percent slopes _ Brinkert-Morset association _ Buffmeyer sandy loam, 4 to 18 percent slopes _ Cabin sandy loam _ Chedsey loam, 5 to 12 percent slopes _ Coalmont-Fluetsch complex _ Cowdrey loam, 4 to 10 percent slopes _ Cowdrey loam, 10 to 50 percent slopes _ Crespin-Carlstrom clays _ Crespin-Carlstrom stony clays _ Cryaquents _ Cryorthents, steep _ Dobrow loam _ Dune land _ Eachuston gravelly loam _ Ethelman sandy loam, 0 to 25 percent slopes _ Fleer loam _ Fluetsch- Tiagos association -----Forelle loam _ Gelkie sandy loam. 2 to 15 percent slopes _ Gelkie sandy loam. moderately deep variant, 0 to 25 percent slopes _ Girardot silty clay loam --------Gothic slopes loam, 0 to 20 percent _ Grafen-Rock outcrop complex ---Grimstone-Siebert association ---Handran extremely stony sandy loam, 20 to 40 percent slopes --Kather clay loam, 5 to 20 percent slopes _ Larand fine sandy loam, 3 to 25 percent slopes _ Larand fine sandy loam, 25 to 40 percent slopes _ 4,981 Percent II 0.7 Map symbol Le LuE 5,084 1,340 0.7 0.2 17,658 12,505 2.5 1.8 LyD Me MbF 10.821 5,703 22,880 1.5 0.8 3.2 Me 3,985 3,848 0.6 0.5 MrO MsO Mn Mo 2,577 23,326 0.4,1 MuE 3.3 NeE 1 2,817 16,661 2.31 NeF 10,924 1.51 Nr 6.012 8.348 1.889 23.567 39.198 9,585 1.202 4,775 1.340 3.608 94.226 11,234 0.4 I 10e 0.81 On 1.1 I P~F 0.31 3.3 PeE 5.51 1.3 I PeF 0.2 ! 0.71 Ph I Pr 0.2 Ra i 0.511 13.2 I 1.6 !I RhO Rk ,I Ro 14,360 962 7,661 2.0 I! RtE :j '0 0.1 :: Sr 1.1 ;1 SuE 7.970 9.275 9,482 1.1:iTe 1.3 :, Tn 1.3 ,i TeF 1,718 0.2 .! ~ITv !Wa 12,264 1.7:1 We 10,100 1.4i 2,508 !I YeF 0,4 :1 ;1 :i ! Appendix II (continued) - Explanation of symbols. _~_cres Soil name Leavitt loam Lulude cobbly loam, 10 to 25 percent slopes Lymanson cobbly loam, 4 to 10 percent slopes MacFarlane-Rock outcrop association Manburn gravelly coarse sandy loam, 10 to 40 percent slopes Mendenhall loam Mine pits and dumps :'Iirror-Rock outcrop complex Mord loam, 4 to 15 percent slopes Mcrset loam. 1 to 15 percent slopes Mugging slopes loam, 5 to 30 percent Nokhu loam. 0 to 25 percent slopes Nokhu loam, 25 to 50 percent slopes Norriston ~a •.•. elly sandy loam Owen Creek sandy loam Owen Creek-Norriston association Parkview very stony loam, 20 to 35 percent slopes Peeler sandy loam. 5 to 25 percent slopes Peeler sandy loam, 25 to 40 percent slopes Perceron-Hyannis association Pinkham-Rock outcrop associat.ion Randman sandy loam Rawah loam. 3 to 10 percent slopes Rock land _ 9,482 _ 4,741 _ 1,992 _ :"2,608 _ ., _ _ _ 1,992 3,348 800 7,386 5,394 _ 7,489 _ _ _ _ _ 2,061 1,615 4.947 4.500 _ 5,565 _ 2,-173 _ ~ 4,016 ; _ -. ~.501 ~ 3.S95 : 5.654 _ _ -:,092 ~.806 Rock outcrop ~------------Rouert g ravef ly sandy loam. 10 to 25 percent slopes _ Spicerton sandy loam _ Stumpp clay loam _ Sudduth loam, 5 to 15 percent slopes _ Tealson-Rock land association _ Tine sandy loam _ Troutville sandy loam. 15 to 45 percent slopes _ Troutville-Newcomb association __ 'Vaiden sandy loam _ Wichup loam _ Yochum gra velly sandy loam 35 to 65 percent slopes ..: _ ~ater _ Total-"!.. 859 _ _ s.sn ',417 .; :~.040 :,092 ::_':.711 : -.486 ~.943 '.057 ::.520 ~.516 :.905 :.928 : .1i9 7:'':.640 F ,::- �139 JOB PROGRESS REPORT State of: Colorado W-152-R Project: Avian Research 5 Job --- Work Plan: 8 Job Title: Population Inventory and Habitat Use by Lesser Prairie-chickens' in Southeast Colorado Period Covered: Author: 01 January through 31 December 1987 Kenneth M. Giesen Personnel: Clait Braun, Beth Dillon, Ken Giesen, Rick Hoffman, Gwen Kittel, Jennie Slater, Chuck Wagner, Bryant Will, Colorado Division of Wildlife ABSTRACT Lek surveys were conducted on a 4l.4-km2 study area and surrounding area of the Comanche National Grasslands. Both lek numbers and total numbers of lesser prairie-chickens (Tympanuchus pallidicinctus) counted increased from previous years. Analysis of the lek count data from the study area indicates a strong correlation between male breeding density and number of active leks (r = 0.91) but no relationship between male breeding density and average lek size (r = -0.014). Experimental roadside transect routes to measure observer efficiency in detecting active leks indicated 20.0-66.7% of known leks were missed under favorable weather conditions. Timing of transects and observer differences were found to affect the results. Experimental quadrat surveys resulted in 0.0-100% of active leks being located on individual quadrats with a higher proportion of leks (86.7%) being located during the peak of hen attendance than 2 weeks later (46.7%). A total of 27 prairie-chickens were captured and banded with 10 hens and 9 males being fitted with radio transmitters. Average home range documented was 347 ha. Hens dispersed 0.4 to 2.9 km from leks for nesting. Average clutch size was 10.8 eggs. Eightly random vegetation transects were measured to document residual vegetative structure and species composition. ��141 POPULATION INVENTORY AND HABITAT USE BY LESSER PRAIRIE-CHICKENS IN SOUTHEAST COLORADO Kenneth M. Giesen Both distribution and populations of lesser prairie-chickens in North America have decreased >90% from historic levels of the 1800's (Taylor and Guthery 1980). Although the exact historic distribution of lesser prairie-chickens is unknown, early reports (Bendire 1892, Judd 1905, Bent 1932, Baker 1953, Sands 1978) suggested they were abundant and widely distributed throughout their range. Although Aldrich (1963) indicated lesser prairie-chickens historically inhabited about 360,000 km2 in 5 states, recent e~timates suggest a current population of 50,000 birds existing on 125,000 km (Crawford 1980, Taylor and Guthery 1980, Johnsgard 1983). Although evidence suggests lesser prairie-chickens were historically peripheral in Colorado, they were thought to be common to abundant in 6 southeastern counties (Baca, Prowers, Bent, Kiowa, Lincoln, and Cheyenne), and peripheral in adjacent counties (Loeffler 1983). Recent surveys have documented breeding populations in Baca, Prowers, and Kiowa counties (Hoffman 1963, Loeffler 1983, Rash 1985). The lesser prairie-chicken is currently classified as a threatened species in Colorado. P. N. OBJECTIVES The objectives of this study are to evaluate 1ek surveys as indices to population trends, ascertain the accuracy of aerial and ground surveys in detecting leks, describe the seasonal floristic and structural characteristics of lesser prairie-chicken habitats in southeast Colorado, and contribute to preparation of a recovery plan for lesser prairie-chickens in Colorado. Segment Objectives 1. Review pertinent literature applicable to the objectives of this study. 2a. Locate all active leks within the 41.4 km2 primary study area and obtain at least 1 count/week (Mar-May) of all males and females on each lek. 2b. Survey all historic leks in Baca County and obtain at least 1 count of males on each active 1ek. 3. Select 4 20-km long roadside listening transects within the range of lesser prairie-chickens in Baca County and test the ability of naive observers to detect active lesser prairie-chicken leks within 1.6 km of the transects. 4. Select 2 or more quadrats, 40-50 km2 each, within the range of lesser prairie-chckens in Baca County, and test the ability of a naive observer to detect active lesser prairie-chicken leks from low-level aircraft flights. �\ 142 5. Trap and band lesser prairie-chickens on active leks within the primary study area. Up to 20 will be marked with miniature radio transmitters to facilitate their periodic location. 6. Locate lesser prairie-chicken nests by following radio-marked hens. Record clutch size, incubation period, and nest fate. 7. Locate radio-marked birds weekly for estimates of movement and home range. 8. Measure vegetative cover at grouse use and random sites. Height and canopy cover of shrubs, forbs, and grasses will be recorded. 9. Compile data, analyze results, and prepare annual progress report. METHODS Field surveys were conducted on the Comanche National Grasslands and adjacent areas in eastern Baca County from March through June using binoculars, a parabolic micrphone listening device, and a trained pointing dog to locate active lesser prairie-chicken leks. All leks known to be active in 1986 were surveyed as well as most known historic lek sites. Active leks were visited within 2 hours of sunrise to count grouse and classify them to sex. Four roadside transects were conducted by naive observers during the peak of lesser prairie-chicken breeding (14-15 Apr) and 2 weeks later. A 3-minute listening stop was taken approximately every 800 m to listen for displaying grouse. Distance and direction to detected leks were estimated and plotted on a map. Routes began 15 minutes prior to sunrise and were completed within 2 hours. Aerial quadrat searches for leks occurred during the peak of breeding activities (14-16 Apr) and 2 weeks later. Three quadrats, each approximately 40 to 48 km2 were plotted on topographic maps and boundaries flagged on the ground. A jet Bell Ranger helicoptor with a pilot and 1 observer/recorder flew north-south transects 400 m apart at approximately 50 m altitude. Grouse sightings were recorded on a portable cassette recorder and later plotted on a topographic map. Visual landmarks including topographic features, fencelines, windmills, and roads were recorded and used to verify lek locations. Because grouse were not detected unless flushed, a lek was defined as 2 or more grouse flushing from 1 site. Cannon-nets and funnel traps (Giesen et ale 1982) were used to capture grouse on leks. Each captured grouse was marked with a numbered aluminum band and a unique combination of colored plastic bandettes. Miniature solar- or battery-powered transmitters (weight 18-24 gms) were attached to all captured females and selected males using a poncho (Amstrup 1980). Radio-marked birds were located using a portable receiver and hand-held yagi antenna. Birds were approached on foot until they flushed. Vegetative structure and species composition were measured using line-intercept of canopy cover (Canfield 1941) and a range pole (Robel et ale 1970). Sand sage (Artemisia filifolia) density was measured on 0.001 ha circular plots. The minimum convex polygon method (Mohr 1947) was used to calculate home ranged. �143 RESULTS AND DISCUSSION Lek Surveys A total of 155 counts of 30 active leks was obtained in Baca County between 7 February and 21 May 1987 (Table 1). In addition, 11 historic lek sites were surveyed and found to be inactive. A minimum of 219 males, 22 females, and 281 total birds was counted for an average count of 9.4 birds/active lek. Summaries of lek count data since 1977 indicate the dynamic nature of lek sites with many becoming inactive and others becoming active (Table 2). Although there is an upward trend in numbers of males counted and numbers of leks active, the data should not be used to calculate absolute breeding population size because not all active leks are located and counted each year. Because numbers of grouse attending leks fluctuates during the breeding season and survey timing and intensity varied each year, the data (Tables 1 and 2) do not reflect actual variability. Analysis of lek surveys and lek count data on the 41.4-km2 study area (Table 3) indicates a strong positive correlation between male breeding density and number of active leks (r = 0.91) but not between male breeding density and average lek size (r = -0.014). If we assume the high count of males on leks is a constant proportion of the total population, then lek density is a better index to population changes than average lek size. These data support results of prairie grouse population surveys in other states (Cannon and Knopf 1981, Martin and Knopf 1981). Lek Transect Routes Results of experimental roadside transect routes (Table 4) indicated 20.0-66.7% of known active leks within 1.6 km were missed under good weather conditions (winds <10 km/hr, clear-partly cloudy skies, no precipitation). When winds exceeded 30 km/hr on 14 April, only 9.1% and 14.3% of active leks were detected on 2 transects. A higher proportion of leks was detected during the peak of hen attendance than 2 weeks later (62.5% vs. 53.8%) under similar weather conditions. Observer differences also occurred with 1 observer detecting 63.2% of the known leks and the 2nd observer detecting 46.7% of the leks under similar conditions. Although many state wildlife agencies use survey routes to monitor changes in numbers of prairie grouse the technique has not been evaluated for accuracy or preC1S10n. Weather conditions, observer efficiency, and timing of counts all appeared to influence results in Colorado in 1987. Other factors including lek size and topography might also be important but were not investigated. Additional testing should be conducted to document factors influencing lek detection before survey routes are implemented as a management strategy to monitor population trends. �144 Table l. Lesser Prairie-chicken lek count data, Baca County, 1987.a N Lek counts 2 3 4 5 6 7 12 14 17 18 25 28 31 33 35 36 37 38 39 40 42 44 86-1 86-2 86-3 87-1 87-2 87-3 87-4 87-5 10 6 6 20 Total Average 8 6 2 5 4 11 3 22 1 5 1 2 2 7 1 6 1 1 8 4 4 3 1 2 2 1 155 Count period 7 20 20 7 21 21 3 31 6 21 23 20 21 6 18 31 27 19 19 19 27 1 1 Feb-29 Mar-17 Mar-17 Feb- 9 Mar-17 Mar-17 Apr-21 Nar-16 Apr-23 Mar-24 Apr-21 Mar-II 18 Apr Mar- 9 18 Apr Apr-21 Apr-27 Mar-16 9 May Mar-16 18 Apr 16 May Mar- 7 Mar- 7 Mar- 5 Mar- 1 15 Apr May- 8 May- 8 13 May Apr May May May May May Apr Hay Apr Apr May May Apr May Apr May May May Apr Apr May May May Males 9 3 3 18 3 8 4 8 7 13 6 17 9 8 18 0 18 7 0 9 10 5 11 5 8 0 0 0 7 5 219 7.3 High count Females Totalsb 1 0 0 4 0 0 2 0 2 2 0 8 0 0 1 0 0 0 0 1 0 0 1 0 0 0 0 0 0 0 10 6 6 19 5 8 5 9 9 15 6 24 12 9 21 2 18 7 6 9 10 8 14 6 8 4 7 4 9 5 22 0.7 281 9.4 aLeks 8, 10, 13, 21, 23, 26, 27, 29, 32, 41 and 46 were surveyed with no birds 0 bserved. blnc1udes males, females, and birds not classified to sex. �145 Table 2. Lek 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 86-1 86-2 86-3 87-1 87-2 87-3 87-4 87-5 High counts of male ,lesserprairie-chickens,Baca County, 1977-87. 1977 4 18 8 9 24 22 15 1 0 0 7 0 7 17 17 16 2 3 3 5 __b 1978 1979 1980 1981 5 14 6 8 NC 15 11 NC NC NC NC 5 13 11 11 0 4 14 NC 0 4 13 16 6 NCa 10 2 7 12 15 11 0 0 0 NC 7 12 10 12 NC 3 NC NC 2 1 6 17 NC 0 1 17 12 15 14 11 0 0 0 0 17 8 9 9 0 0 9 0 0 0 4 19 0 3 9 7 7 0 6 15 7 16 14 6 0 0 0 0 20 8 8 0 0 0 9 0 0 0 3 30 0 6 4 4 7 4 18 14 6 8 6 aNo count. bLek not yet located. Year 1982 0 3 12 7 29 19 9 0 0 0 0 14 7 6 0 0 0 10 0 0 0 0 33 0 11 2 4 4 0 10 17 0 6 0 10 12 1983 1984 1985 1986 1987 0 12 11 9 26 14 8 0 0 0 0 16 5 12 0 0 0 9 0 0 0 0 23 0 3 2 3 4 0 0 13 0 0 0 16 7 12 7 10 4 0 9 7 7 23 17 7 0 0 0 0 18 0 11 0 0 2 5 0 0 0 0 21 0 2 3 0 0 0 6 11 0 0 0 21 0 10 8 18 5 3 14 0 7 7 0 18 6 5 0 0 0 0 0 0 11 0 0 5 9 0 0 0 0 16 0 0 0 4 10 2 2 7 5 0 0 13 0 12 3 9 7 7 7 7 3 4 4 NC 16 2 0 17 6 9 NC NC NC NC 5 0 10 NC NC 4 10 NC NC 0 NC 2 NC 2 0 0 12 0 0 2 0 0 0 10 0 12 2 11 8 3 0 5 0 0 1 15 6 7 NC 9 3 3 18 3 8 0 NC 0 NC 4 0 8 NC NC 7 13 NC NC 0 NC 0 NC 6 0 0 17 0 NC 9 0 8 NC 18 0 18 7 0 9 0 10 NC 5 NC 0 11 5 8 0 0 0 7 5 �146 Table 3. Lek count trends of lesser prairie-chickens on a 4l.4-km2 (16 2 mi ) study area, Baca County, Colorado, 1980-87. Year N leks N males 1980 1981 1982 1983 1984 1985 1986 1987 6 6 6 6 4 5 7 8 59 55 59 65 46 46 75 74 x Breeding density (males/km2) males/lek 9.8 9.2 9.8 10.8 11.5 9.2 10.7 9.2 1.42 1.33 1.42 1.57 1.11 1.11 1.81 1.79 Table 4. Leks detected during experimental roadside transects in Baca County, Colorado, 1987. Transect Length (km) 17.6 22.5 19.2 19.2 1 2 3 4 N leksa N detected 6 Survey 1 Survey 2r: 3 7 11 5 2 ld ld 3 1 3 6 4 aNumber of active leks within 1.6 km of transects. b14-l5 Apr. c30 Apr-l May. dWinds >30 km/hr. Aerial Surveys Aerial quadrat surveys were relatively effective in detecting active lesser prairie-chicken leks during the peak of mating activities but less effective 2 weeks later (86.7 vs. 46.7% of active leks detected, Table 5). Because the aircraft, pilot, observer, and procedures were identical for both survey periods, the causes of the discrepancy are unknown. Possibly hens are more likely than males to flush in response to low level aircraft flights. During experimental flights over 2 leks the observer failed to detect male lesser prairie-chckens on leks until the birds flushed despite knowing the approximate lek locations. Additional tests of factors affecting detection of lekking grouse from aircraft should be conducted to ascertain the relative importance of aircraft type (helicoptor or fixed-wing), altitude, flight speed, and presence or absence of hens on leks. The advantage of aerial �147 surveys of leks over ground surveys is the relative ease in which large parcels of rangeland can be surveyed to detect occupancy by lesser prairie-chickens. However, unless the survey is conducted during the peak of mating when hens are attending leks, the survey may underestimate lek density and relative population size. Table 5. Leks detected by experimental aerial surveys of lesser prairie-chicken occupied rangelands in Baca County, Colorado, 1987. Quadrat Quadrat size (km2) N leksa 1 2 3 47.9 46.0 40.1 3 3 9 N leks detected Flight 2c Flight 15 3 3 7 3 0 4 aNumber of active leks within quadrat boundaries. b14-l6 Apr. c28-30 Apr. Trapping and Banding A total of 27 lesser prairie-chickens (18 males, 9 females) was trapped on 5 of 8 active leks within the primary study area. All but 5 males were captured in funnel traps on leks. The remaining 5 males were captured using a cannon-net. Males quickly habituated to the presence of funnel traps and the chicken-wire drift fences and became difficult to capture. Few females were captured because they hesitated to enter the funnel opening after following the wire leads, perhaps because the funnels were too long and narrow. However, several hens escaped from traps when funnels were shortened. Apparently the small size of the traps «1.0 m diameter) allowed them to find the funnel opening and escape. All captured prairie-chickens were marked with a unique combination of colored bands to facilitate future identification. Ten males and all 9 hens were fitted with miniature radio transmitters to facilitate their periodic relocation. Although the transmitters were not thought to have a major affect on daily activity patterns or movement of birds, there is a possibility that radio-marked birds had higher than typical mortality from predation. This increased predation on radio-marked birds has been previously reported (Herzog 1979, Warner and Etter 1983, Hines and Zwickel 1985, Marks and Marks 1987). Most males captured (14 of 18, 77.8%) were 2 years of age or older suggesting that many yearling males may not regularly attend leks or that trapping methods were selective for more dominant and older males. There appeared to be no age bias for females captured (5 adults, 4 yearlings) although older hens were captured earlier in the season than yearlings. �148 Nesting The fates of 9 radio-marked hens were ascertained using radio-telemetry techniques. Four hens were killed by raptors prior to nest completion, 1 hen was not relocated after trapping, and 4 hens completed clutches. Two of these hens lost clutches prior to hatching and 2 hens successfully hatched their clutches. Average clutch size was 10.8 eggs (range 9-13) and average distance from ~ek of capture to nest site was 1.9 km (range 0.4-2.9 km). Home Range Home ranges were measured for 5 males and 5 females for which data were available (Table 6). Home range size varied among individuals (range 44-662 ha) but there was no significant difference between males and females (t = 17, p > 0.2) • Combined average home range size was 347 ha , Home range size of radio-marked lesser prairie-chickens in Baca Table 6. County, Colorado, 1987. Band Age Sex 335 340 2+ I- M F 343 344 345 349 351 354 355 356 I2+ 2+ 2+ 2+ 2+ 2+ 2+ F F F F M M M M Time interval Home range (ha) 3 Apr-14 Oct 6 Apr- 1 Sep 14 15 15 15 17 29 29 8 Apr-IO Apr-14 Apr-23 Apr-13 Apr- 1 Apr-13 Apr-14 May-14 236 181 May Oct Jun Oct Sep Aug Oct Oct 315 329· 459 226 468 44 662 550 Habitat Measurement A total of 80 random vegetation 1987 to complete the vegetative Analysis has not been completed sandsage density and height and complete analysis of vegetation final report. transects was measured sampling of the entire although there appears species composition of on the study area will on the study area in area begun in 1986. to be high variation in residual grasses. The be provided in the LITERATURE CITED Aldrich, J. W. 1963. Geographic orientation of American Tetraonidae. Wildl. Manage. 27:529-545. Amstrup, S. C. 214-217. 1980. A radio-collar for game birds. J. J. Wildl. Manage. 44: �149 Baker, M. F. 1953. Prairie chickens of Kansas. Misc. Publ. 5. 68pp. Univ. Kansas Mus. Nat. Hist. Bendire, C. E. 1892. Life histories of North American birds with special reference to their breeding habits and eggs. u.s. Natl. Mus. Spec. Bull. 1. 446pp. Bent, A. C. 1932. Life histories of North American gallinaceous birds. Natl. Mus. Bull. 162. 490pp. U.S. Canfield, R. H. 1941. Application of the line intercept method in sampling range vegetation. J. For. 39:388-394. Cannon, R. W., and F. L. Knopf. 1981. Lek numbers as a trend index to prairie grouse populations. J. Wildl. Manage. 45:776-778. Crawford, J. A. 1980. Status, problems, and research needs of the lesser prairie chicken. Pages 1-7 in P. A. Vohs and F. L. Knopf, eds. Proc. Prairie Grouse Symp. Oklahoma State Univ., Stillwater. Giesen, K. M., T. J. Schoenberg, and C. E. Braun. 1982. Methods for trapping sage grouse in Colorado. Wild1. Soc. Bull. 10:224-231. Herzog, P. W. 1979. Effects of radio-marking on behavior, movements, and survival of spruce grouse. J. Wildl. Manage. 43:316-323. Hines, J. E., and F. C. Zwicker. 1985. Influence of radio packages on young blue grouse. J. Wildl. Manage. 49:1050-1054. Hoffman, D. M. 1963. The lesser prairie chicken in Colorado. Manage. 27:726-732. Johnsgard, P. A. 1983. Lincoln. 4l3pp. The grouse of the world. J. Wildl. Univ. Nebraska Press, Judd, S. D. 1905. The grouse and wild turkeys of the United States and their economic values. U.S. Dep. Agric. BioI. Surv. Bull. 24. 55pp. Loeffler, C. 1983. The status and management of the lesser prairie chicken in Colorado. Unpubl. Rep., Colorado Div. Wi1dl., Colorado Springs. 9pp. Marks, J. S., and V. S. Marks. 1987. Influence of radio-collars on survival of sharp-tailed grouse. J. Wildl. Manage. 51:468-471. Martin, S. A., and F. L. Knopf. 1981. Aerial survey of greater prairie chicken leks. Wildl. Soc. Bull. 9:219-221. Mohr, C. o. 1947. Table of equivalent populations of North American small mammals. Am. MidI. Nat. 37:223-249. Rash, M. T. 1985. Survey of the lesser prairie-chicken in Colorado, 3 April25 May 1985. Unpubl. Rep., Colorado Div. Wildl., Colorado Springs. 22pp. �150 Robel, R. J., J. N. Briggs, J. J. Cebula, A. D. Dayton, and L. C. Hulbert. 1970. Relationships between visual obstruction measurements and weight of grassland vegetation. J. Range. Manage. 23:295-297. Sands, J. L. 1978. Game bird studies. Performance Rep., Proj. W-l04-R-19. New Mexico Dep. Game and Fish Proj. Albuquerque. 5pp. Taylor, M. A., and F. S. Guthery. 1980. Fall-winter movements, ranges and habitat use of lesser prairie chickens. J. Wildl. Manage. 44:521-524. Warner, R. E., and S. L. Etter. 1983. Reproduction and survival of radiomarked hen ring-necked pheasants in Illinois. J. Wildl. Manage. 47:369-375. Prepared by ==--..::..~~.;=::;:...:....:IJ1~...::.~=~. ::.:....:.. Kenneth M. Giesen Wildlife Researcher _ � https://cpw.cvlcollections.org/files/original/a5a6010f9e0e65933a7886aebe48fbee.pdf 58f0d8a1d0b519137b279d5b701ef96e PDF Text Text 151 JOB PROGRESS REPORT State of Colorado - project W-152-R Work Plan _9_ Job Title: Author: Personnel: Job _8_ Food Selection During Winter Period Covered: Avian Research 01 January and Nutritional through Ecology 31 December of Blue Grouse 1987 T.E. Remington R.W. Hoffman, T.E. Remington, Division of Wildlife M.L. Stevens, Colorado ABSTRACT Digestibility of neutral detergent fiber (NDF) of Douglas-fir (Pseudotsuga menziesii)., lodgepole pine (Pinus contorta), subalpine fir (Abies lasiocarpa), and Engelmann spruce (Picea engelmannii) by 4 captive blue grouse (Dendragapus obscurus) averaged 13.5, 14.3, 17.3, and 7.5%, respectively. Tannin activity was high for needles of all species but was not related to blue grouse preferences among needle groups. Polyvinylpyrrolidone (which binds tannin) added to spruce needles at 2 and 5% of the Dm did not reduce nitrogen excretions or affect intake (~ > 0.05). Nitrogen in fecal droppings was excreted primarily as ammonium salts (33 to 55%) and uric acid (9 to 21%) or bound to plant fiber (9 to 25%). Uric acid excretion was inversely related to intake and weight loss (~ <0.05) across all species and represented normal endogenous nitrogen losses as well as loss due to catabolism of protein during energy . shortages. Ammonium salt excretion was positively correlated with intake (P < 0.05) on all conifer species and highest when grouse fed on unpalatable species (4.5 mg/g unpalatable vs. 2.6 mg/g palatable). Ammonium salts may be a byproduct of metabolic detoxification of plant defense compounds, arising through deamination of amino acids to furnish substrates (glucuronic acid, methyl or sulfur groups) for conjugation with toxins. Defense compounds in conifer needles represent a metabolic cost in nitrogen and energy to detoxify or restrict intake rather than directly decreasing digestibility as originally hypothesized. ��153 FOOD SELECTION AND NUTRITIONAL ECOLOGY OF BLUE GROUSE DURING WINTER Thomas E. Remington Chemical analyses of samples and statistical analysis and interpretation of data generated from previous trials with captive blue grouse were continued. Another trial was conducted which assessed the role which tannins may play in reducing the metabolizable nitrogen (MN) and energy content of Engelmann spruce needles. The primary effect of ingested tannins is to bind to and precipitate protein (McLeod 1974). It was hypothesized that the poor palatability and low MN content of Engelmann spruce needles is caused in whole or in part by tannins. This was tested by treating spruce needles with polyvinylpropylene (PVP) , a inert polymer which binds to tannins and inhibits formation of tannin-protein complexes (Gray 1978, Asquith and Butler 1985). Palatability and MN of spruce needles should increase, if the hypothesis is correct, when tannin activity is decreased. In addition, needles were collected and processed and four blue grouse captured for trials to be conducted in January of 1988. P.N. OBJECTIVES The objectives of this study are to investigate (1) winter food habits, and (2) winter food preferences of blue grouse, and (3) measure the nutritional quality (protein, cell contents) and anti-quality components (terpenoids, phenolic resins, neutral detergent fiber [NDF], acid detergent fiber [ADF]) of blue grouse winter foods and their relationship to diet preferences. Specific objectives are to: 1. Identify winter foods of blue grouse. 2. Investigate blue grouse winter use of conifers for food by species, tree ages, and growth forms in relation to their' availability. 3. Quantify tree species composition characteristics of blue grouse winter and physical feeding sites. 4. Measure protein and NDF, ADF content of conifer needles trees fed upon by blue grouse and from randomly-located trees. 5. Measure tannin, phenolic resin, and mono-, di-, and sesquiterpene levels in conifer needles from trees fed upon by blue grouse and from randomly-located trees. from �154 6. Rank blue grouse preference for Douglas-fir, subalpine fir, lodgepole pine, limber pine, and Engelmann spruce as winter foods. 7. Measure protein, NDF, ADF, and DM digestibility by blue grouse of randomly-selected needles of Douglas-fir, lodgepole pine, limber pine, subalpine fir, and Engelmann spruce. 8. Investigate the deterrent effects of specific phenolic resins, and mono-, di-, and sesquiterpenes to browsing by blue grouse and to blue grouse digestibility of conifer needles. SEGMENT OBJECTIVES 1. Complete chemical analyses (DM, nitrogen, fiber, gross energy, uric acid, urea, ammonia, tannins) of feed and excreta generated during trials in previous segments. 2. Code and enter resulting files. data onto existing computer data 3. Collect needles from Douglas-fir and/or lodgepole pine feeding and random trees to determine distribution of nutrients/anti-nutrients and for trials with captive birds. 4. Capture 4 blue grouse for use in feeding trials. 5. Conduct trials to evaluate ME and MN (or DMD and nitrogen balance) of feeding and random trees of Douglas-fir and/or lodgepole pine. 6. Analyze and interpret data and prepare progress report. Present pertinent findings at appropriate scientific meetings. DESCRIPTION OF STUDY AREAS This study was conducted on Whitely Peak and adjacent Burnt Mountain, located approximately 30 km west of Kremmling in Grand County, Colorado. Characteristics of these areas have been described by Cade (1985). METHODS Conifer needles and fecal and cecal excreta samples were prepared and neutral detergent fiber, nitrogen, gross energy, and dry matter content measured using standard and previously described techniques (Remington 1987). Uric acid (Remington 1987) and ammonia levels in fecal samples were measured, ammonia by microkjeldahl distillation without prior sample digestion. Tannin activity in needle samples was estimated by measuring the amount �155 of bovine serum albumin (BSA) precipitated by methanol of plant material (Asquith and Butler 1985). extracts Trial birds were captured with telescoping noose poles (Zwickel and Bendell 1967) and acclimated to experimental conditions and diets for 1 week prior to data collection. Trial methodology followed Remington (1987). Treatments consisted of 3 levels of PVP added to needles of Engelmann spruce; 0 (control), 2%, and 5% of DM. PVP was dissolved in ethanol (2% = O.952g/10 ml, 5% = 2.380g/10 ml) and sprayed over 100g (wet wt.) of needles. Control needles were sprayed with alcohol only. The needles of all groups were placed in a fume hood overnight to allow the alcohol to evaporate. Variables measured were dry matter intake (palatability), dry matter digestibility (DMD), and nitrogen balance (nitrogen intake - nitrogen excreted). NDF digestibilities calculated from the conifer species trial conducted in 1986 did not agree closely with the preliminary data reported (Remington 1987) from the aborted 1985 trial, probably because of the extreme physiological stresses imposed on the birds by the 1985 experimental protocol. Revised estimates and discussion of the 1986 results are presented. RESULTS AND DISCUSSION NDF digestibility ranged from 7.5 to 17.3% among species and averaged 13.1% for all species (Table 1). These values are substantially below the 38% digestibility of cellulose and 44% digestibility of lignin in heather by red grouse (Lagopus lagopus scoticus) estimated by Moss (1977). The methods used by Moss to estimate fiber are likely responsible for his inflated digestibilities, although it is possible that red grouse digest fiber far better than blue grouse. NDF digestibility is a function of how much NDF is made available to cellulolytic fermentation (i.e. that enters the ceca), and the efficiency of digestion of NDF within the ceca. These parameters were evaluated to assess causes of low NDF digestibility. NDF digestibility was determined largely by how much NDF entered the caeca since essentially all (avg. = 96.8; range across species = 94.3 to 98.6%) NDF that entered the ceca was digested (Table 1). Most fiber was excluded from the ceca. NDF digestibility of Engelmann spruce needles was about half that of the other 3 species, primarily because of greater exclusion of NDF from the ceca (Table 1). The low NDF digestibility of spruce needles is exacerbated by their relatively high NDF content. �156 Table 1. Digestibility of NDF in conifer needles by blue grouse. NDF Digestibility speciesb Psme pico Abla Pien %DM 0.395 0.474 0.383 0.416 Total 13.48 14.31 17.31 7.48 (%) Available In ceca Amount (g) 98.5 98.6 94.3 95.8 4.7 5.1 3.5 2.0 aassuming no fermentation outside of ceca. bPsme = Douglas-fir, pico = lodgepole pine, Abla fir, Pien = Engelmann spruce. to cecaa ~ of total 0 13.68 14.51 18.22 7.69 = subalpine Tannin activity (protein precipitating ability) of extracts of conifer needles was measured to assess their relationship to preference, palatability, and metabolizable nitrogen. Tannin activity did not vary substantially across needle groups (Table 2) and was not consistently related to preference or palatability. For instance, lodgepole pine, a palatable and preferred species, had a tannin index about 20% higher than the unpalatable and non-preferred subalpine fir. Tannin indices tended to increase with needle age in old and young trees of Douglas-fir, consistent with preferences of blue grouse for these groups and with their nitrogen losses on them. All groups had relatively high tannin index levels. On average there was enough precipitation capacity to bind twice as much protein as was actually present in the plant sample (assuming plant protein has an affinity for tannins equal to BSA). Given this it is not surprising that the small amount of variation across needle groups is unimportant. The protein precipitation assay exhaustively extracts tannins under conditions of optimal tannin solubility which may not reflect solubility in the stomach or gut of herbivores. Perhaps much of the precipitating capacity of tannins and/or other polyphenols is not realized because of poor tannin solubility in the gut. To test this I extracted plant samples with HzO-HCL (pH = 2.0) instead of methanol-HCL and repeated the assay. Protein precipitation declined by about half for most groups (Table 2). While this was a significant decrease, there remained enough precipitation capacity to bind all the protein in these needles. Obviously most of this invitro capacity is not realized in-vivo. How grouse circumvent this potential loss of protein is not obvious. Robbins et ale (1987) found that reduction of protein digestibility in conifer forage fed to elk (Cervus elaphus) was only 12% of that estimated by in-vitro protein precipitation. Chickens and rats are known to secrete a proline~rich protein from their salivary glands in �157 response to dietary tannins (Mehanso et ale 1983). These proteins have a tremendous affinity for tannins and may represent a countermeasure to prevent tannins from binding dietary protein or from attacking the gut wall. This strategy should still lead to elevated protein loss unless the inactivated tannin-proline protein complex can be metabolized farther down the digestive tract. Table 2. Absorbancea (590 nm) of dyed BSA precipitated by 375 or 500 ul of methanol-HCL or HzO-HCL extracts of conifer needles. Needle group Douglas-fir Lodgepole pine Subalpine fir Engelmann spruce Douglas-fir (young) 1-2 year old needles 3-4 year old needles 5-7 year old needles Douglas-fir (old) 1-2 year old needles 3-4 year old needles 5-7 year old needles aAbsorbance is linearly precipitated. !izO-HCL 500 ul Methanol-HCL 500 ul 375 ul 1.29 1. 36 1.12 1.26 0.98 1.11 0.84 1. 02 0.72 0.80 0.78 1. 30 1.48 1.54 1. 01 1. 33 1. 43 0.63 0.75 0.75 1. 33 1. 38 1.36 0.98 1.16 1.14 related to the amount of BSA If PVP had mitigated the protein precipitating activity of tannins in spruce needles as hypothesized, then nitrogen excretion should have decreased, nitrogen balance should have increased, and intake potentially could have increased with addition of PVP. Addition of PVP to needles of Engelmann spruce had no effect on their palatability (defined as dry matter consumption; £ = 0.97), or the nitrogen balance of birds feeding on them (£ > 0.05, Table 3). Cecal nitrogen excretion remained constant over treatment levels (£ = 0.19, Table 3) as well. If tannin-protein complexes are diverted into the ceca, then a beneficial effect of PVP in preventing formation of tanninprotein complexes should have been demonstrated by a reduction in cecal nitrogen excretion. Tannins do not appear to be the cause of excessive nitrogen losses when birds feed on Engelmann spruce needles. �158 Table 3. Nitrogen balance and cecal nitrogen excretion (mg) by 4 blue grouse following feeding of Engelmann spruce needles untreated and treated with 2 levels of PVP. Average of 3 replicates/bird. PVP Oa 2 5 422 Cecal Bal. -399 -310 -189 aTreatment 168 194 181 level Bird 426 Bal. Cecal -455 -457 -359 183 191 216 427 Bal. Cecal -467 -568 -473 201 242 242 428 Bal. Cecal -677 -795 -1008 192 219 235 (%DM) . Endogenous nitrogen losses have been measured both to correct apparent metabolizable energy and to investigate sources of nitrogen losses on poor foods. The principle nitrogen excretion product of blue grouse is ammonium salts, followed by indigestible plant protein (NDF nitrogen), and uric acid (Table 4). Consumption of Engelmann spruce and subalpine fir (unpalatable species) resulted in elevated uric acid (Fig. 1) and ammonia (Fig. 2) excretions compared to an equivalent amount of Douglas-fir or lodgepole pine (palatable species). Uric acid is the primary endogenous nitrogen excretion product in most birds (Farner and King 1972:541). Uric acid seems to be largely a byproduct of protein catabolism in blue grouse since the amount of excreted uric acid is an inverse function of intake and a direct function of weight loss (Fig. 1). Moss and Hanssen (1980) found this to be true of red grouse as well. Table 4. Forms of nitrogen excretion (% of total) in blue grouse fecal droppings following feeding on 4 conifer species. species Douglas-fir Lodgepole pine Subalpine fir Engelmann spruce NH3 salts 46.5 32.8 39.0 55.4 Bound 15.2 25.0 9.0 10.1 Uric acid 9.0 9.7 20.7 15.4 ~ 0 70.7 67.6 68.8 80.9 Ammonia nitrogen is positively correlated with intake but the form of this relationship differs among species (Fig. 3). Ammonium salts are normally a minor component (~ 15%) of endogenous nitrogen excretion in birds (Farner and King �159 BIRD #422 """ C) 6 6 140 BIRD #424 0 6 90 - 120 E •.....• 75 - o Z 100 0 0 0 0 « 60 ~ 80 0 0 a:: :;) 60 ~ 45 ~ 0 0 40 I I I 40 50 60 INTAKE 0 * 0 0 I I 50 80 60 70 """ C) E t- t- 200 I- 0 t- 150 t- 0 « 125 t- 0 a:: 100 r- 75 r- 50 r- 25 I- (g) 0 400 0 350 Z 0 100 450 0 175 90 BIRD #397 0 225 80 INTAKE (g) BIRD #425 250 * * 0 /::, 30 ~ * * 70 * 0 300 250 0 0 200 :;) 6 06 150 * 6 0 I 30 I 50' 40 INTAKE I I 60 70 (g) I- 100 ~ 50 0 /::, I- S* * /::, ** I 80 I 40 50 60 70 INTAKE 80 (g) Fig. 1. Uric acid nitrogen excretion as a function of intake for 4 blue grouse feeding on Douglas-fir (*), lodgepole pine (0), subalpine fir (0), and Engelmann spruce (6). * /::, 90 �160 BIRD #422 BIRD #424 375 240 f- 220 f... 350 •...•. C) E ...., * £:::,. 325 * * * 275 - 4: 200 Z ~ ~ c: 300 - Z 0 c. 0 180 0 6 4: 0 * 0 * 250 225 - 0 0 160 200 f- 0 175 0 140 _l 1 1 1 40 50 60 _l 70 80 0 ,I I 1 60 50 70 c; 1 80 1 90 100 INTAKE (g) INTAKE (g) BIRD #397 BIRD #425 6 350 375 6 325 •...•. 300 - 350 6 C) E 275 ...., Z 0 250 4: Z 0 225 ~ ~ 4: 325 6 f- 300 I- 275 * 0 200 0 0 175 * ** 0 0 250 f- 225 I- 200 f- ** 6 0 0 150 125 0 0 1 30 40 50 60 70 80 feeding on Douglas-fir Engelmann spruce (6). excretion 50 60 70 INTAKE (g) INTAKE (g) Fig. 2. Ammonia nitrogen 40 as a function of intake for 4 blue grouse (*), lodgepole pine (0), subalpine fir (D), and 80 90 �161 1972:541). The level of ammonia nitrogen found in fecal droppings of blue grouse following ingestion of conifer needles was excessive (33-55% of excreted nitrogen). Ammonia nitrogen excretion may not represent a true endogenous loss, but may result from detoxification of plant defense compounds. This Three predictions can be nitrogen loss will be designated as Nd• advanced and tested with available data. 1) Nd excretion should increase as dry matter consumption of a particular food increases. Conversely, if ammonia nitrogen represents a endogenous loss resulting from protein catabolism then it should decrease as intake increases. 2) Nd excretion per gram of dry matter intake should be higher on unpalatable foods (presumably with more toxins) than on palatable foods. 3) Toxins, such as monoterpenes or benzoic acid, should elevate Nd excretion. These 3 predictions were tested and found true (£ < 0.05). Nd excretion was positively correlated with intake of Douglas-fir (Figs. 3 and 4), subalpine fir, and Engelmann spruce (Fig. 3) with 84, 44, and 91%, respectively, of the variance in ammonia excretion explained by dry matter consumption. Nd excretion was not strongly correlated (£ = 0.12) with intake of lodgepole pine (Fig. 3) needles but was negatively correlated with weight loss (£ < 0.05, r2 = 0.48). The unpalatable species, subalpine fir and Engelmann spruce, had substantially higher Nd excretions per gram of intake than the palatable species, Douglas-fir and lodgepole pine (4.1 and 4.9 vs. 2.7 and 2.6 mg/g DM, respectively) . Ammonia excretions per gram intake did not vary across tree and needle age groups where palatability also did not differ. Ammonia nitrogen excretion was positively correlated with intake (£ ~ 0.05; Fig. 4). Addition of benzoic acid at 0.33% DM to Douglas-fir needles elevated Nd excretion by 9%. Addition of monoterpenes at 2.5 and 5.0% of the DM to needles of Douglas-fir increased excretion of Nd by 5 and 39%, respectively. It seems clear that the excretion of nitrogen which can be recovered by micro-kjeldahl analysis is directly related to metabolic detoxification of toxins. Which class of toxins is primarily responsible for elevated nitrogen excretions is not certain, but an examination of toxicology literature and the data gathered so far can lend some insights. Xenobiotic compounds ingested in food present different problems to animals depending on whether they are lipophilic (hydrophobic) or hydrophilic. Lipophilic compounds (such as monoterpenes) are readily absorbed, have access to cell organelles, and are difficult to excrete. Their lack of solubility in water means they are not likely'to be leached from plant material. Absorption thus depends largely on the extent to which plant cells containing these compounds are physically disrupted by the animal's digestive apparatus. Hydrophilic compounds (such as phenolic acids) should be easily leached from plant material but will be absorbed readily only if they are relatively small compounds. �162 DOUGLAS FIR 300 .. Z ~ Z 0 ~ ~ ~ Y = -179.754 + 4.7773xR SUBALPINE FIR 350 = 0.92 280 260 240 Z ~ 0 • 220 ~ ~ ~ • 200 150 75 80 85 95 90 100 20 100 LODQEPOLE 70 100:-----.-----r----.-----r----~ 30 40 50 6C 70 80 PINE ENGELMANN 300 • Z ~ ~ ~ 140 z • 220 180 • • riA • 50 • SPRUCE Y = - 30.5648 + 5.3747xR = 0.96 350 Y = 89.2089 + 1.3423xR = 0.35 • 40 OM CONSUMPTION 400 260 30 60 OM CONSUMPTION Z • 250 • 180 70 0 • 300 Z 200 ~ Y = 58.0405 + 2.8833xR = 0.66 ~ • • • 300 Z 250 ~ ~ ~ 200 o 150 100+----.~--,-----~---r----r---~ 65 70 75 80 85 60 OM CONSUMPTION 90 OM CONSUMPTION Fig. 3. Relationship between ammonia nitrogen excretion and dry matter consumption when captive blue grouse fed on needles of 4 species of_conifers. �163 400 y = 34.104+ 3.3149xR"2 = 0.620 • • 300 • • <t Z 0 • • ::a: ::a: <t 200 - • • • • •• •• • I •• • • • 100~--~----~---r----r---~--~----~---' 50 60 70 40 80 .OM CONSUMPTION Fig. 4. Relationship between ammonia nitrogen excretion-(mg) and dry matter consumption (g) when captive blue grouse fed on 1-2, 3-4, and 5-6 year old needles of old and young Douglas-fir. �164 Once absorbed, xenobiotics undergo phase I and phase II biotransformation reactions (Fig. 5) mitigated by a complex host of enzymatic pathways (Klaassen et ale 1986). Phase I reactions expose or add functional groups which makes the compounds more hydrophilic (increasing their excretability) and makes them susceptible to phase II reactions. In phase II reactions the compound or a phase I derived metabolite is covalently linked to an endogenous molecule, producing a conjugate. These conjugates are more easily secreted than the parent compounds. Conjugating moieties utilized depend on the structure of the xenobiotic, but typically glucuronic acid, sulfates, or amino acids are employed. There are several important points to consider about detoxification. It is energetically expensive. There is an ATP requirement to drive the reaction and also to produce the substrates. Energy stored in the substrate used for conjugation is lost upon excretion. Glucuronic acid is synthesized from glucose via ascorbate (vitamin C), which means when a xenobiotic is conjugated to glucuronic acid a molar equivalent of glucose is lost and vitamin C requirements may rise as well. There is also a nitrogen cost in most cases. This is either because the molecule used for conjugation contains nitrogen (e.g. ornithine, glycine), or because the conjugate is derived from a nitrogen containing compound. For example, the sulfur group in the cofactor used for sulfotransferase reactions is donated from cysteine, and methionine donates methyl groups for similar methylation reactions. Most xenobiotics can enter several detoxification pathways. The route used at any given time depends on the concentration of the compound, the energy status of the animal and many other factors. For instance the sulfation pathway has a high affinity but low capacity for conjugation of phenols. The major alternative reactions, glucuronidation and amino acid conjugation, have low affinity but a much higher capacity. saturation of high affinity, low capacity pathways may have been indicated by the pattern of nitrogen excretion following ingestion of 3 levels of monoterpenes. Doubling monoterpene levels over levels already present in Douglas-fir needles increased nitrogen excretion only 5%. Doubling this level (4x baseline) increased nitrogen excretion by 39%. There is some information available on how specific defense compounds present in conifer needles are metabolized and excreted (summarized in Fig. 5). Chickens conjugate benzoic acid (found in spruce) primarily with ornithine and to a lesser extent with glycine and glucuronic acid (Baldwin et ale 1960, Nesheim and Garlich 1963, Bridges et ale 1970). More complex phenolic monomers, of which there are several in conifer needles, are oxidized to benzoic acid and then conjugated with ornithine (Crowdle and Sherwin 1923), or conjugated directly without transformation. Polyphenolic compounds such as tannic acid are hydrolysed in the gut to gallic acid, absorbed and methylated to 4-0-methyl gallic acid, or decarboxylated to pyrogallol (Potter and Fuller 1968, Kadirvel et ale 1969). Methionine and choline are the probable methyl group donors. Addition of methionine, �165 choline, cysteine and arginine to diets containing as much as 1% tannic acid alleviated the growth suppressing effects of tannic acid (Fuller et al. 1967, Rayudu et al. 1970). This implies that some conjugation of gallic acid (the hydrolysis product of tannic acid) with ornithine also occurs (the arginine serving as a source of ornithine) . The pathway of detoxification and excretion of monoterpenes has not been studied in birds. In mammals monoterpenes are conjugated with glucuronic acid and excreted (Dean et al. 1967, Eberhardt et al. 1975, Dash 1988). It is likely that this pathway or ornithine conjugation predominates in birds as well. The metabolism of diterpene resin acids, which make up from 1-3% of the DM of conifer needles, has also not been studied. Conjugation with glucuronic acid and/or amino acids following appropriate transformations seems probable. It is apparent that the major classes of plant defense compounds share simiiar detoxification pathways (Fig. 5) and that these pathways consume amino acids and elevate nitrogen excretion. The means by which this nitrogen is excreted as ammonia or ammonium ion is by no means clear. I propose 2 mechanisms which may account for elevated ammonia excretion following metabolism of toxins. 1) Amine groups cleaved off of amino acids (cysteine, methionine, choline) to generate carbon skeletons for methylation or sulfation of toxins are transported to the kidney and excreted as ammonium ion (NH4+). Low protein diets and weight loss/starvation, conditions prevalent on spruce and subalpine fir, increase NH4+ excretion in chickens relative to other urates (Skadhauge 1983). 2) Glucose needed to produce glucuronic acid for conjugation will likely result from metabolism of carbon skeletons produced by catabolism of protein reserves on toxic diets (i.e. spruce and fir). since low intakes and metabolizable energy contents are inadequate to meet daily energy needs dietary carbohydrates will not provide excess glucose for synthesis of glucuronic acid. Amine groups generated by this catabolic process could elevate ammonia excretion as above. Since retrograde transport of urine from the cloaca to the ceca has been demonstrated (Mortensen and Tindall 1981), it is possible this nitrogen would continually cycle until fixed by bacteria. This may be prevented by the strong tendency of ammonium ion to complex with and precipitate uric acid out of colloidal suspension (Skadhauge 1983), thereby preventing the whole complex from being washed back to the ceca. The conjugation of xenobiotics in conifer needles with glucuronic acid or amino acids may represent a potential bottleneck in detoxification, either because of enzyme kinetics or substrate limitation. Tannins, phenolic monomers, monoterpenes and diterpene resin acids together may comprise 5-10% of the DM of conifer needles and pose a formidable energy and nitrogen cost to detoxify (Fig. 5). If these compounds singly or in toto overwhelm detoxification pathways, grouse may respond by lowering intake to hold the detoxification load to a manageable level. �166 XENQBIQTIQ Monoterpenes Tannins -. PHASE 1 TRANSFORMATION Phenolics gut hydrolysis Resin acids / Add or remove functional groups to facilitate phase 2 reactions cysteine ~methionine PHASE 2 QQNJUGATIQN Glycine Ornithine sultate ~ fmethYI carbon skeleton plus NH 3 EXCRETED (Bile or urine) Hippuric acid Dibenzoyl ornithine Glucuronic acid conjugate Sulfate eOOH 4-0-methyl gallic acid eOOH ~9 o-~-o 8 ~ O-CH 3 ~ig. 5: Probable pathways of detoxification of xenobiotics In conlfer needles and structures of excreted conJ.ugt present a es. �167 This strategy exasperates low dietary nitrogen and energy contents, but may have been used by grouse eating spruce and subalpine fir (Remington 1987). Secondary compounds in conifer needles were originally hypothesized to have anti-digestibility effects on blue grouse. Trials with captive birds have indicated that these effects are minor. These compounds depress intake (anti-feedant) and incur sUbstantial metabolic costs of detoxification, particularly in nitrogen. Further experimental and analytical work will focus on elucidating which classes of compounds are primarily responsible for these costs. Analysis of detoxification products in excreta is a promising approach. LITERATURE CITED Asquith, T.N., and L.G. Butler. 1985. Use of dye-labelled protein as spectrophotometric assay for protein precipitants such as tannin. J. Chern. Ecol. 11:1535-1544. Baldwin, B.C., D. Robinson, and R.T. Williams. 1960. Studies in detoxification. The fate of benzoic acid in some domestic and other birds. Biochem. J. 76:595-600. Bridges, J.W., M.R. French, R.L. smith, and R.T. Williams. The fate of benzoic acid in various species. Biochem. 118:47-51. 19700 J. Cade, B.S. 1985. Winter habitat preferences and migration patterns of blue grouse in Middle Park, Colorado. M.S. Thesis, Colorado State Univ., Fort Collins. 101pp. Crowdle, J.H., and C.P. Sherwin. 1923. Synthesis of amino acids in-the animal organism. II. The synthesis of ornithine in the body of the fowl. J. BioI. Chern. 55:365. Dash, J.A. 1988. Effect of dietary terpenes on glucuronic acid excretion and ascorbic acid turnover in the brushtail possum (Trichosurus vulpecula). Compo Biochem. Physiol. 89B:221-226. Dean, F.M., A.W. Paice, A.P. Wade, and G.S. Wilkinson. Uroterpenol B-D-Glucuronide. J. Chern. Soc. London 19:1893-1896. 1967. (C) Eberhardt, I. H., J. McNamara, R.J. Pearse, and I. A. Southwell. 1975. Ingestion and excretion of Eucalyptus punctata D.C. and its essential oil by the koala, Phascolarctos cinereus (Goldfuss). Aust. J. Zool. 23:169-179. Farner, D.S., and J.R. King, eds. 1972. Avian biology. Vol. II. Academic Press, New York, N.Y. 612pp. �168 Fuller, H.L., S.I. Chang, and D.K. Potter. 1967. Detoxification of dietary tannic acid by chicks. Nutr. 91:477-481. J. Gray, M. 1978. Absorption of polyphenols by polyvinylpyrrolidone and polystyrene resins. Phytochemistry 17:495-497. Kadirvel, R., G.V.N. Rayudu, and P. Vohra. 1969. Excretion of metabolites of tannic acid by chickens with and without ceca. Poult. Sci. 48:1511-1513. Klaassen, C.D., Amdur, M.D., and J. Doull, eds. 1986. Casarett and Doull's Toxicology, 3rd ed. McMillan Publ. Co., New York, N.Y. 974pp. McLeod, M.N. quality. 1974. Plant tannins - Their role in forage Nutr. Abstr. and Rev. 44:803-815. Mehanso, H., A.E. Hagerman, S. Clements, L.G. Butler, J. Rogler, and D.M. Carlson. 1983. Modulation of proline rich protein biosynthesis in rat parotid glands by sorghums. Proc. Natl. Acad. Sci. 80:3948-3952. Mortensen, A., and A. Tindall. 1981. Cecal decomposition of uric acid in captive and free ranging willow ptarmigan (Lagopus lagopus lagopus). Acta Physiol. Scand. 111:129133. Moss, R. 1977. Digestion of heather spring. Condor 79:471-477. --- , and I. Hanssen. 1980. and Rev. B. 50:555-567. by red grouse Grouse nutrition. during the Nutr. Abstr. Nesheim; M.C., and J.D. Garlich. 1963. Studies on ornithine synthesis in relation to benzoic acid excretion in the .domestic fowl. J. Nutr. 79:311-317. Potter, D.K., and H.L. Fuller. dietary tannins in chicks. 1968. Metabolic fate of J. Nutr. 91:477-481. Rayudu, G.V.N., R. Kadirvel, P. Vohra, and F.H. Kratzer. 1970. Effect of various agents in alleviating the toxicity of tannic acid for chickens. Poultry Sci. 49:1323-1326. Remington, T.E. 1987. Food selection and nutritional ecology of blue grouse during winter. Job Prog. Rep., Colorado Div~ Wildl. Fed. Aid. Proj. 01-03-045 (W-37R). Apr. Pp. 177-198. Robbins, C.T., T.A. Hanley, A.E. Hagerman, O. Hjeljord, D.L. Baker, C.C. Schwartz, and W.W. Mautz. 1987. Role of tannins in defending plants against ruminants: reduction protein availability. Ecology 68:98-107. in �169 Skadhauge, E. 1983. Formation and composition of urine. Chpt. 6 in B.M. Freeman ed., Physiology and Biochemistry of the Domestic Fowl. Academic Press, New York, N.Y. Zwickel, F.C., and J.F. Bendell. 1967. A snare for capturing blue grouse. J. Wildl. Manage. 32:456-468. Prepared by Approved by ��171 JOB PROGRESS REPORT State of: Colorado Project: W-152-R Avian Research Work Plan: 12 Job Title: Chronology of Breeding and Nesting Activities of Wild Turkeys in Relation to Timing of the Spring Hunting Season Period Covered: Author: Job 15 --- 01 January through 31 December 1987 Richard W. Hoffman Personnel: C. E. Braun, R. W. Hoffman, R. L. Holder, and T. B. Lundt, Colorado Division of Wildlife ABSTRACT Efforts to trap wild turkeys (Meleagris gallopavo) in 1987 were unsuccessful. Few birds were located in the study area and those that were could not be attracted to bait sites. The decision was made to wait and attempt trapping on the study area in 1988 rather than to trap elsewhere in 1987. Hunter activity and harvest were monitored in 1987 through use of a permit system and wing collection program. Based on a 100% survey, 3,670 spring hunters harvested 575 turkeys (16% success) and 1,314 fall hunters harvested 278 turkeys (21% success). Las Animas County accounted for 25 and 29% of the spring and fall harvest, respectively. Public-land supported about 55% of the hunting pressure, but produced only 42 (spring) and 36% (fall) of the harvest. Hunter success was better on private (spring = 28%, fall = 40%) than public (spring = 12%, fall =·15%) land. The wing collection program sampled 66% of the spring harvest and 65% of the fall harvest. Merriam's dominated (>92%) in the·harvest samples. Bearded hens comprised 3% (N = 10) of the sample. Seventy-nine percent of the spring harvest of MerrIam's (males only) were adults compared to 59% for Rio G~andes. The fall harvest was comprised I}f 51% juveniles, 4% subadults, and 45% adults (excluding Rio Grandes due to inadequate sample). Sex ratios favored females in the fall harvest. ��173 CHRONOLOGY OF BREEDING AND NESTING ACTIVITIES OF WILD TURKEYS IN RELATION TO TIMING OF THE SPRING HUNTING SEASON Richard W. Hoffman P. N. OBJECTIVES 1. Document the timing of winter flock dispersal, onset of gobbling, peaks of gobbling, nest initiation, onset of incubation, and peak of hatch in relation to timing of the spring season. 2. Describe the gonadal cycle of females and compare the reproductive condition of females in relation to timing of the spring season. 3. Measure the abandonment rate of incubating females to varying levels of human disturbance around the nest. 4. ~bnitor hunter activity and harvest of wild turkeys on a statewide basis. SEGMENT OBJECTIVES 1. Review literature pertinent to the objectives of this study. 2. Trap and instrument 40 wild turkeys (6 males, 34 females) with tailor poncho-mounted radio trasmitters in December-February.- 3. Document timing of winter flock break-up. 4. Document onset of gobbling and peaks of gobbling activity. 5. Document onset of egg laying, incubation, and peak of hatch. 6. Measure effects of human disturbance on rate of nest abandonment. -7. Conduct a hunter questionnaire- wing collection program using a permit system and mail wing survey. 8. Compile data, analyze results, and prepare progress report. DESCRIPTION OF STUDY AREA The study area, located southwest of Trinidad, Colorado in Las Animas County, was previously described (Hoffman 1987). METHODS A detailed account of the methods used in this study were reported by Hoffman (1987). �174 RESULTS AND DISCUSSION Trapping Trapping operations were initiated in January 1987 and continued through mid-April 1987. Except for a flock of 20-30 birds in Sowbelly Canyon, no other turkeys or their sign were found in Longs Canyon or any of its tributary canyons. Efforts to trap the birds in Sowbelly Canyon were unsuccessful as they could not be attracted to bait sites. Instead, they preferred to feed on hillsides rather than in canyon bottoms where they could be trapped. Several different baits were offered to no avail. Cattle were also a problem as they consumed the oat hay and trampled the bait site, including the cannon nets and wire leads. The lack of use of bait sites was attributed to an abundant supply of pinon (Pinus edulis) nuts. Bait sites were placed at other known turkey use areas in Longs, Colorow, Saruche, Martinez, and Little Martinez canyons. No birds were observed at any of these sites. Flocks were found near Lake Dorothey and Spanish Peaks. However, because of access and landowner problems, it was not considered feasible to trap these flocks. In addition, the resulting data would not have been directly comparable with that collected in 1986. It was considered better to skip a year and try to trap at the same sites in 1988 than to trap at another location in 1987. Consequently, trapping operations were terminated on 15 April 1987. This was the latest date to trap and still meet the objectives of the study. Hunter Compliance - Permit System A total of 4,257 permits was issued for the spring season and 4,607 licenses were sold. Hunter compliance with the spring permit requirement was therefore 92%. Questionnaires were sent to all permit holders by 1 June 1987 of which 58% were returned. A followup questionnaire was mailed to the non-respondents 3 weeks later and 527 additional responses were received. The combined return rate for the first and second mailing was 70% (Table 1). There were more permits issued (1,630) for the fall season than there were licenses sold (892). The reason was that unsuccessful spring hunters who wanted to hunt the fall season were not required to purchase another license, but were required to obtain a fall permit. Consequently, total license sales and total permits issued were not directly comparable in evaluating hunter compliance. Instead, hunter compliance was measured as the proportion of new license buyers who also obtained a permit. The estimate was 87% (776 of. 897). Questionnaires were sent to the 1,630 fall permit holders plus the 121 license buyers who did not obtain a permit by 20 October. Fifty-nine percent were returned by 15 November 1987. On 20 November 1987, a followup questionnaire was sent to the 717 non-respondents and 217 replied. Seventy-two percent of the fall hunters responded to the questionnaire (Table 1). �175 Table 1. Response to the 1987 spring and fall turkey harvest survey. Surveys mailed Surveys returned Percent return Non-deliverable 1st Spring 2nd Totals 1st 4,257 2,472 58 83 1,785 527 30 38 4,257a 2,999 70 121 1,751 1,036 59 50 Fall 2nd Totals 717 217 30 6 1,75la 1,253 72 56 aTotal hunters sampled. Hunter Activity and Harvest - Questionnaire Projected estimates for the spring season indicated that 3,670 hunters harvested 575 turkeys for a success rate of 16% (Table 2). Total kill, including a reported crippling loss of 19%, was estimated at 713 birds. Spring hunters averaged 3.6 days afield over the course of the 30-day season from 18 April to 17 May. There were 64% fewer hunters that participated in the fall season (1,314 hunters) and 52% fewer birds (278) were harvested (Table 3). However, hunter success (21%) was higher and crippling loss (18%) was slightly lower. Hunters in fall spent 2.8 days afield or about one day less than hunters in spring, but hunters had only 16 days (19 Sep-4 Oct) of hunting which included just 3 weekends. The spring season lasted 14 days longer, allowing for 2 additional weekends of hunting opportunity. Most hunting pressure and harvest occurred on opening weekend during both seasons CTable 4). Pressure and harvest did not change substantially over the remainder of the fall season. There was an increase in pressure on weekends during the spring season, but this was not always associated with a corresponding increase in harvest. The Southeast Region accounted for over 77% of the spring and fall harvest (Tables 5 and 6). Las Animas County was the leading harvest area with 25 and 29% of the spring and fall harvest, respectively. Public land supported 55% of the spring hunting pressure and 54% of the fall pressure, but produced only 42 (spring) and 36% (fall) of the harvest CTable 7). Hunter success averaged 24% on the spring limited permit areas, exceeding the statewide success rate of 16% on all but 2 areas (Table 8). An eastern plains limited permit area (unit 96) supported the highest hunter success rate (33%). It was not advantageous in terms of success to hunt on a limited permit area during the ~all season as only 27 of 220 (12%) fall limited permit holders were successful (Table 8). However, hunter success on 2 of the 4 fall permit areas surpassed the statewide average. It was a definite advantage to hunt on private (spring = 28% success, fall = 40% success) vs. public land (spring = 12% success, fall = 15% success) in both seasons. �f-' Table 2. -...J 1987 spring turkey harvest and hunter activity. 0' Mailing Descriptive statistic N in sample N hunters % hunters N hunters observing turkeys % hunters observing turkeysb N successful hunt~rs (harvest) % successful huntersb N hunter dabs Days/hunter Crippling loss % crippling loss Total harvest 1st 2nd 2,472 2,151 87 1,441 67 394 18 7,720 3.6 79 17 473 527 450 85 274 61 53 12 1,647 3.6 22 29 75 aNon-respondents. bBased only on those license holders who actually hunted. L 2,999 2,601 87 1,715 66 447 17 9,367 3.6 101 18 548 Projected for Totals 1,258a 1,069 85 652 61 128 12 3,848 3.6 37 29 165 4,257 3,670 86 2,367 64 575 16 13,215 3.6 138 19 713 �Table 3. 1987 fall turkey harvest and hunter activity. Mailing DescriEtive statistic .!:! in sample N hunters % hunters N hunters observing turkeys % hunters observing turkeysb N successful hunters (harvest) % successful huntersb N hunter dabs Days/hunter Crippling loss % crippling loss Total harvest 1st 1,036 779 75 427 55 182 23 2,102 2.7 40 18 222 2nd 217 162 75 87 54 29 18 483 3.0 6 17 35 E 1,253 941 75 514 55 211 22 2,585 2.7 46 18 257 Projected for 498a 373 75 201 54 67 18 1,119 3.0 14 17 81 Totals 1,751 1,314 75 715 54 278 21 3,704 2.8 60 18 338 aNon-respondents. bBased only on those license holders who actually hunted. •..... -.....J -.....J �178 Table 4. Chronological distribution of hunting pressure and harvest during the 1987 spring and fall wild turkey seasons.a Date 1st 1st 2nd 2nd 3rd 3rd 4th 4th 5th weekend week weekend week weekend week weekend week weekend Totals Spring Hunter da~s N % Fall Harvest N % Hunter da~s % N Harvest N % 2,345 1,117 1,529 657 1,043 571 782 509 814 25 12 16 7 11 6 8 6 9 134 55 72 36 28 31 26 24 17 32 13 17 8 7 7 6 6 4 883 459 539 284 331 35 19 22 11 13 94 30 33 22 28 45 14 16 11 14 9,367 100 423 100 2,496 100 207 100 aBased on hunter days and harvest that could be assigned to specific time periods. Seventy-seven percent of the birds harvested during the spring season were taken before noon; 58% were harvested betwen 0500 and 0900 hours (Table 9). Only 29% were not associated with other birds at the time of harvest (Table 10). Most (50%) were taken from flocks comprise of 1-5 birds. The remaining 21% were taken from flocks with 6 or more birds. There was no indication that hunters encountered smaller flocks as the season progressed. Hens undoubtedly were present in the larger flocks, which suggests they were still receptive to males and not incubating. Wing Collection Program Limitations - The validity of any population indice calculated from harvest samples is dependent upon the assumption that the different age and sex classes are harvested in proportion to their occurrence in the population. Long-term population and harvest data are often necessary to test this assumption. Such data are not available for the Merriam's wild turkey in Colorado or elsewhere throughout its range. In addition, if wings are collected over a broad geographic area, they may not accurately reflect the characteristics of local populations. These limitations do not preclude the use of wing data as a useful tool in formulating management strategies. Biologists must be aware of them and use caution in interpreting the data. Compliance - Of the 447 successful spring hunters who responded to the questionnaire, 303 (70%) said they returned a wing. However, 377 wings were processed, meaning 74 successful hunters who did not respond to the questionnaire still returned a wing. The wing collection program sampled 66% (377/575) of the spring harvest. �179 Table 5. Wild turkey harvest by county, Game Management Unit, and Region, Spring 1987. County Las Animas Fremont Huerfano Custer Pueblo Dolores ] Montezuma Mesa E1 Paso Logan ] Washington Morgan Larimer Yuma ""1 Ki t Carson-! Archuleta Douglas Baca Teller Boulder Weld Park Chaffee Bent Jefferson Gilpin Costilla Delta Otero % GMU N % Region 25 li 10 SE SW 6 5 NW 31 26a 64 60 52 26 20 19 15 14 12 8 7 6 84 140 85 71 59 143 24 6 581 17 4 96 15 4 421 103 109 13 3 106 49 43 35 15 4 4 3 3 10 8 7 6 5 5 4 3 3 3 2 1 1 1 2 2 2 1 1 1 1 <1 <1 <1 <1 <1 <1 <1 511 20 2 2 8 8 8 8 2 851 19 51 136 7 2 58 6 1 137 39 125 46 52 83 135 130 30 Totals Unknown 425 22 100 314 74 37 31 9 7 6 25 18 4 425 100 2 2 86 95 40 li8 501 56 123 41 142 78 141 38 144 108. 146 Central % 4 69 77 NE N 2 2 2 5 1 5 5 5 5 1 1 1 1 4 1 4 1 1 4 3 2 2 2 1 1 1 1 1 1 1 1 1 1 1 1 425 22 <1 <1 <1 <1 <1 <1 <1 <1 <1 <1 <1 <1 <1 <1 <1 <1 100 22 aHarvest was pooled by area which included more than one county or GMU. �180 Table 6. Wild turkey harvest by county, Game Management Unit, and Region, Fall 1987. County N % GMU N % Region N % Las Animas 56 29 24 18 10 10 29 15 13 10 5 5 33 32 27 20 13 10 9 17 16 14 10 7 5 5 SE SW Central NW NE 170 10 8 6 2 87 5 4 3 1 Baca Douglas Mesa Kit Carson Teller Park Yuma El Paso Larimer Costilla 7 7 6 5 4 4 3 2 2 1 4 4 3 3 2 2 2 1 1 <1 84 85 140 143 59 77 58 103 108 109 8 4 Jefferson Chaffee 1 1 <1 <1 8 6 5 4 3 3 3 2 2 2 2 1 1 1 1 4 3 3 2 2 2 2 1 1 1 1 <1 <1 <1 <1 190 21 100 196 15 100 196 15 100 Erenont; Pueblo Huerfano Custer Archuleta Totals Unknowris Ja 69 51 421 123 851 86 46 144 511 501 19 136 40 56 38 .. aHarvest was pooled by area which included more than one county or GMU. Distribution of hunting pressure and harvest by land status during Table 7. the 1987 spring and fall wild turkey seasons. Spring Fall Harvest Hunters Land status Public Private Both Totals N % N % Hunters % N Harvest N % 1,395 794 370 55 31 14 164 224 42 58 491 336 84 54 37 9 74 134 36 64 2,559 100 388 100 911 100 208 100 �181 Hunter success on limited permit areas during the 1987 spring and Table 8. fall wild turkey seasons. Spring Permit areas Lake Dorothey Beaver-Skagway Units 103 and 109 Units 107, 108, 112, 113, 114, 115, 120, 121 Unit 96 Colorado Spgs SWA Spanish Peaks Fall Harvest Success (%) Permits issued Harvest Success (%) 75 20 40 16 4 11 21 20 27 75 30 25 7 7 8 9 23 32 20 70 5 __b 1 23 0 5 33 0 5 6 Permits issued __a __a __a 90 aClosed to fall hunting. bNo restrictions on numbers of hunters during the spring season. Distribution of harvest by time period during the 1987 spring and Table 9. fall wild turkey seasons. Spring Time period 0500 0600 0700 0800 0900 1000 1100 1200 1300 1400 1500 1600 1700 1800 1900 - 0559 - 0659 - 0759 - 0859 - 0959 - 1059 - 1159 - 1259 - 1359 - 1459 - 1559 - 1659 - 1759 .- 1859 - 1959 Fall N % N % 5 100 63 49 31 19 19 6 7 9 11 10 27 11 5 1 27 17 13 8 5 5 2 2 3 3 3 7 3 1 10 31 33 29 12 9 3 3 1 10 11 14 10 6 17 18 16 7 5 2 2 1 6 6 8 6 aMountain Daylight Time. �182 Table 10. Size of flocks from which turkeys were harvested during the 1987 spring season. Flock size 1 Date of harvest 18-22 Apr 23-27 Apr 28 Apr - 2 fuy 3-7 May 8-12 May 13-17 May Totals 2-5 >10 6-10 N % R % N % N % 46 17 17 12 14 12 28 19 38 34 38 39 79 51 24 18 20 12 47 57 53 51 54 39 19 18 3 2 2 3 11 20 7 6 5 9 24 4 1 3 1 4 14 4 2 9 3 13 118 29 204 50 47 12 37 9 A total of 180 wings was collected from the fall wing survey representing 65% of the estimated fall harvest. Seventy-five percent (160) of the successful hunters responding to the fall questionnaire indicated they returned a wing and 20 successful hunters not responding also returned a wing. Subspecies Composition - The Merriam's wild turkey was the dominant subspecies identified from inspection of wings. It comprised 91 and 96% of the spring (Table 11) and fall (Table 12) samples, respectively, and was harvested in 24 of Colorado's 63 counties. Rio Grande's were taken in 7 counties. Fremont, El Paso, and Pueblo were counties where both subspecies were harvested. The 34 Rio Grande wings collected during the spring season originated from the following areas: 15 from Unit 96, 10 from Units 103/109, 4 from Unit 123, 2 from Unit 95, 2 from Unit 84, and 1 from Unit 69. Two of these areas (Units 95 and 96) were closed to hunting in the fall. The result was that only 8 Rio Grande wi~gs were identified in the fall sample of 180 wings: 7 from Unit 103/109 and 1 from Unit 59. Sex COmposition - Ten (3%) wings examined from the spring sample were from females (Table 11). All were Merriam's and all were assumed to be from legal hens; i.e., bearded hens. Three percent of the total spring harvest (N = 17) is probably a reasonable estimate of the legal harvest of hens. At thIs level of harvest, spring hunting should have minimal impact on the female segment of the population. Examination of wings (Merriam's only) originating from the fall season revealed the sex ratio of adults and subadults combined favored females (1.6 females: 1 male), while the sex ratio of juveniles was about equal (1.2 females: 1 male). The biological implications of these ratios are difficult to interpret without knowledge of the population structure. Hoffman (1962) reported winter counts of 1.6 females: 1 male along the Front Range in southeastern Colorado. The sex ratio of 171 turkeys harvested on the Uncompahgre Plateau during the fall seasons of 1961 and 1963-67 was 1.7 females: 1 male) (Myers 1973). Both ratios were based on combined samples of adults, subadults, and juveniles. A comparable ratio using the 1987 fall harvest data would be 1.4 females : 1 male. Myers' data by age class shows �183 that the sex ratio of adults and subadults (2.5 females : 1 male) and juveniles (1.5 females : 1 male) still favored females. Sex ratios should favor females in a promiscuous species such as the wild turkey where most hens breed but only a portion of the males do. Table 11. Sex, age, and subspecies composition of the spring 1987 wild turkey harvest based on wing analyses. Males Subspecies Merriam's Rio Grande Totals Subadult Adult Females Subadult Adult Totals 60 (18)a 14 (41) 269 (79) 20 (59) 2 (1) 0 8 (2) 0 339 34 74 (20) 289 (77) 2 (1) 8 (2) 373b apercent. bFour additional Merriam's wings were processed but could not be classified to age or sex bringing the total wings collected to 377. Age Composition - Seventy-nine percent of the spring harvest (males only) of Merriam's wild turkeys were adults compared to only 59% for Rio Grande's. Rio Grande populations in Colorado are the result of recent (since 1981) introductions. Available evidence suggest these populations are still increasing and expanding into unoccupied ranges, while populations of Merriam's wild turkeys have remained stable or declined. The age composition of the Rio Grande harvest reflects good production, survival, and recruitment of young birds, attributes indicative of an increasing population. Estimates of nesting success based on observations of radio-marked Rio Grande (58% success) (Schmutz 1988) and Merriam's (20% success) (Hoffman 1987) hens further suggest differences in reproductive output that should be reflected by harvest samples. Subadult males in the increasing Rio Grande populations may not be suppressed by mature males from participating in breeding activities. If so, then subadults should be equally, if not more vulnerable due to their inexperience, to the calling tactics of hunters. The following discussion regarding the age composition of the fall harvest pertains to the Merriam's wild turkey. Samples of Rio Grande wings from the fall season were inadequate for meaningful interpretation. Subadults comprised jus~ 4% of the fall harvest including juveniles and 8% excluding juveniles (Table 12). Biologically, this could be interpreted as poor production in 1986 and subsequent low recruitment into the 1987 population. However, another factor contributing to the low number of subadults is that associated with the onset of primary molt and whether the key feathers (primaries IX and X) for separating adults and subadults are still present at the time wings are collected. If these feathers have already molted, then subadults cannot be distinguished from adults. This was the case as 44% of the males and 24% of the females had completed their primary molt by the fall season. The problem of separating adults and subadults was �I--' Tables 12. Age, sex, and subspecies composition of the 1987 fall harvest based on wing analyses. Adults Females Males Subspecies Merriam's Rio Grandes co ~ Total Subadu1ts Females Males Total Juveniles Females Males N % N % N % N % N % N % N % N % 32 41 25 46 59 75 78 4 45 50 0 0 0 0 7 1 100 100 7 1 4 12 40 0 46 0 47 54 100 1 3 3 Total N 87 3 % Sample size 51 38 172 8 Pou1ts/ hen 1.6 1.0 �185 compounded for males because they start and finish their primary molt sooner than females. A similar problem was confronted in interpreting wing data derived from harvest samples of blue grouse (Dendragapus obscurus) (Hoffman 1985). Misidentification of subadults is not a problem in spring when birds are just beginning their primary molt. Percent juveniles in the harvest (51%) and the ratio (1.6:1) of juveniles adult and subadult females were used as indicies to productivity. The 2 indices suggested only fair production of young in 1986. LITERATURE CITED Hoffman, D. M. 1962. The wild turkey in eastern Colorado. Game and Fish. Tech. Publ. 2. 49pp. Hoffman, R. W. inferences. to 1985. Blue grouse wing analyses: methodology Colorado Div. Wildl. Spec. Rep. 60. 2lpp. Colorado Dep. and population 19B7. Chronology of breeding and nesting activities of wild turkeys in relation to timing of spring hunting seasons. Prog. Rep. Colorado Div. Wildl. Fed. Aid Proj. W-152-R. Apr. 1987. Pp. 199-232. Myers, G. T. 1973. The wild turkey on the Uncompahgre Plateau. Colorado Vivo Wildl. Fed. Aid Proj. W-37-R. l53pp. Final Rep. Schmutz, J. A. 1988. Reproductive performance and habitat use of Rio Grande wild turkeys. M.S. Thesis, Colorado State Univ., Fort Collins. 5lpp. ��187 JOB FINAL REPORT Colorado State of: Project: W-152-R (01-03-045) : Job Avian Research 16 Work Plan: 12 Job Title: Reproductive Performance and Habitat Use of Rio Grande \Vild ur eys ~eriod Covered: Author: 01 January 1986 through 31 July 1988 Joel-A. Schmutz Personnel: W. Andelt, M. Baker, J. Ratti, G. White, Colorado State University; C. Braun, J. Corey, T. Davis, M. Etl, D. Hall, R. Hoffman, W. Miles, J. Schmutz, S. Steinert, Colorado Division of Wildlife , ABSTRACT Reproductive performance, nest habitat use, and brood habitat use of Rio Grande wild turkeys (Meleagris gallopavo intermedia) were studied along the South Platte River in northeast Colorado in 1986-87. Data were collected from 13 and 18 radio-marked hens in 1986 and 1987, respectively. Yearlings dispersed farther (K = 35.5 km) than adults (K = 14.3 km) from winter flocks to nest sites. Thirty of 31 hens were known to attempt nesting. Eighteen hens hatched eggs. Adult hens nested earlier than yearlings. Earlier first nest.attempts tended to be more successful than later first attempts. Among first nest attempts by adults, earlier nests had. larger clutches. Adults laid larger eggs than yearlings but there .. :was no difference between age classes in clutch size or number of eggs hatched. Egg size was not related to clutch size or time of season. Thirty-three of 35 nests were in ungrazed riverbottom. Early nests were more likely to be in snowberry (Symphoricarpos sp.) clumps (vs , mixed forbs and grasses) than later nests. Forbs and grasses provided little cover early in the season but did not differ from snowberry later in the season. Nest plots had greater overstory canopy cover, more shrubs, fewer grasses, and greater understory cover and height than adjacent random plots. Nesting habitat did not differ between age classes or nest fates. Minimizing probabilities of nest depredation may be the proximate cause of chronological shifts in nest habitat use. Seventy-eight percent of 191 macrohabitat locations were in the riverbottom. Use of riverbottom, agricultural uplands, and the edge between these habitat types was dependent on time of day and may be related to the thermal environment. Brood use of grazed areas increased with brood age. Microhabitat �188 use of 14 broods was examined at 47 locations within the riverbottom. Among these locations, grasses were more abundant on brood-use plots than on random plots. Young broods used areas with more shrubs and grasses than older broods. RECOMMENDATIONS The results of this research suggest management strategies but do not test the effect of such strategies. Further research involving manipulations to test specific management options is required. 1. Do not allow cattle grazing on state wildlife areas unless within the context of a carefully designed experiment. 2. Burning of standing dead grass on state wildlife areas should only be done on areas outside the riverbottom. 3. Open riverbottom from Sterling to Crook to hunting. 4. Timing of the spring hunting season for Rio Grande wild turkeys alo~g the South Platte River should remain as it was set in 1986-88. 5. Expansion of this population will occur without transplant operations. If more rapid expansion is desired, transplanted turkeys should be moved >30 km from the closest winter flock. The most immediate need in further research is to develop a suitable method to estimate turkey densities. Such a method would allow the Division of Wildlife to assess the impact of hunting seasons and would be an integral part of any additional research including testing of density effects on reproductive parameters and habitat manipulations. Aerial surveys during winter when there is snow cover is likely the best approach. The actual sampling methodology and analysis requires investigation. �189 ACKNOWLEDGMENTS This project required the cooperation of the Colorado Division of Wildlife (CDOW), Colorado State University, and the landowners along the South Platte River. Funding was provided by the CDOW through Federal Aid in Wildlife Restoration Project W-lS2-R. For varied assistance in all aspects of this project, I am indebted to numerous CDOW employees including C. Braun, J. Corey, T. Davis, M. Etl, D. Hall, R. Hoffman, W. Miles, and S. Steinart. Clait Braun deserves special thanks for his encouragement, candid discussions, and concern for my development as a biologist. I thank Dr. Bill Andelt for serving as my major advisor. much from his friendship and thorough critiques. I learned I also thank Drs. Mike Baker and Rick Knight for serving on my graduate committee. perspectives helped open my mind to new ideas. Their I appreciate Dr. John Ratti for bringing me to CSU and serving as my initial advisor. I am grateful to Dr. Gary White for all he taught me in the classroom and at the bar. His forthright approach, constructive criticisms, professional conduct, and fervent enthusiasm were inspiring. My time in the field was greatly enriched by Tim Davis and the whole Etl clan - Mike, Paula, Cara, Jamie, and Sadie. support and friendship helped me immeasurably. for her dinners. Their logistical My stomach thanks Paula �190 Many thanks go to my second family - my fellow graduate students. Morning coffee, FAC's, volleyball, basketball, and skiing were all great. In particular, I want to thank Tanya and CA for opening up their house and their hearts to me. And finally, I thank my first family - mom, dad, and all the way down to Namu. Their unflagging love, interest, and s~pport for me through all my pursuits provides me with the strength to follow my dreams. �191 Table of Contents Page ABSTRACT OF THES IS ACKNOWL EDGMENTS iii v LIST OF TABLES ..................................•.................. ix LIST OF FIGURES ....................•................................ x Chapter 1. REPRODUCTIVE PERFORMANCE OF RIO GRANDE WILD TURKEYS 1 Study Area 2 Methods 3 Results 4 Discussion 14 Literature Cited 2a Chapter 2. NEST HABITAT USE OF RIO GRANDE WILD TURKEYS 23 Study Area ..............•....................................... 24 Methods 25 Habitat Measurements ...........•............................. 25 Statistical Analyses 27 Resul ts 27 Discussion 32 Literature Cited 38 Chapter 3. BROOD HABITAT USE OF RIO GRANDE WILD TURKEYS 41 Study Area 42 Methods 43 Resul ts 45 Macrohabi tat Use 45 �192 Table of Contents (cant/d) Page Microhabi tat Use 46 Di scuss i on .......•.............................................. 46 Literature 50 Cited �193 List of Tables 1.1 Initiation dates of first nest attempts by wild turkeys in northeast Colorado, 1986-87 8 1.2 Body weights (kg) of wild turkeys captured in February 1986-87 in northeast Colorado ~.17 2.1 Habitat variables at wild turkey nest and random plots in northeast Colorado, 1986-87 31 2.2 Habitat variables at wild turkey nests in snowberry and other vegetation in northeast Colorado, 1986-87 33 2.3 Numbers of successful and depredated wild turkey nests in snowberry and mixed forbs and grasses in early and late season in northeast Colorado, 1986-87 •............................... 34 3.1 Microhabitat variables at wild turkey brood-use « 22 days old) and random plots in early and late season in northeast Colorado, 1986-87 47 �194 List of Figures Figure Page 1.1 Winter flock and nest locations of wild turkeys in northeast Colorado, 1986-87. Shaded portion of state map represents study area .............................................•..... 7 1.2 Cumulative nest initiation for wild turkeys in northeast Colorado, 1986-87 _ 10 1.3 First nest initiation as a predictor of clutch size in adult wild turkeys in northeast Colorado, 1986-87 13 1.4 Winter body weight as a predictor of nest initiation date in yearling wild turkeys in northeast Colorado, 1986-87 16 2.1 Probability of wild turkeys nesting in snowberry as a function of nest initiation date in northeast Colorado, 1986-87. Regression predicted values ± SE from y = 0.0414! - 6.0542, where initiation date was the Julian date 30 �195 Chapter 1 REPRODUCTIVE PERFORMANCE OF RIO GRANDE WILD TURKEYS The Rio Grande wild turkey (Meleagris gallopavo intermedia) is endemic to northern Mexico, Texas, Oklahoma, and southern Kansas (Aldrich 1967). Transplanting programs have extended the range of this subspecies from California (Graves 1975) to North Dakota (Aldrich 1967). Although several studies of Rio Grande wild turkeys have been reported from Texas (e.g., Watts 1968, Beasom and Pattee 1980, Ransom et al. 1987), little is known about the reproductive ecology of newly established populations in other habitats. Information on reproductive performance of this subspecies is necessary to understand its population dynamics. Beasom and Pattee (1980), working in south Texas, reported productivity of wild turkeys was correlated with the amount of previous fall and current spring rainfall. Greater soil moisture resulting in more vigorous plant growth of higher nutritional value was the hypothesized mechanism affecting reproduction. On the Welder Wildlife Refuge in south Texas, Ransom et al. (1987) postulated that predation, primarily of nests, was the major limiting factor for Rio Grande wild turkeys. In more northerly habitats, winter severity has been suggested as a major factor limiting Eastern wild turkey populations (M. g. sylvestris) (Porter et al. 1983, Gray and Prince 1988). I studied the reproductive performance of an introduced population of Rio Grande wild turkeys in riparian habitats in northeast Colorado �196 in 1986-87. Nesting rates, hen success, and temporal patterns in nest initiation were investigated to better understand the population dynamics of Rio Grande wild turkeys in a non-native, more northerly habitat. STUDY AREA The study was conducted along the South Platte River in Logan, Morgan, and Washington counties in northeast Colorado. Elevation varied from 1,130 to 1,322 m in a gradual east-west gradient. Annual precipitation averaged 36 cm; mean January and July temperatures were -4.5 and 23.5 C, respectively (Natl. Oceanic and Atmos. Adm. 1986). The riverbottom community extended to 1.0 km in width and was dominated by an open-canopied plains cottonwood (Populus sargentii) forest. Discrete clumps of woody shrubs, primarily western snowberry (Symphoricarpos occidentalis), occurred in an understory dominated by cheatgrass brome (Bromus tectorum), prairie cordgrass (Spartina pectinata), inland saltgrass (Distichlis stricta), and a variety of forbs (plant names follow Scott and Wasser [1980]). Lindauer (1983). provided a complete vegetative description of this particular community. Private lands adjacent to the riverbottom were primarily used for production of alfalfa, corn, wheat, and other small grains and row crops. Cattle were grazed at varying intensities both in and adjacent to the riverbottom. The Colorado Division of Wildlife owned approximately one-third of the riverbottom in the study area. lands were not grazed and were used for both consumptive and nonconsumptive recreation. These �197 Sixty Rio Grande wild turkeys from Kansas and Texas were introduced to the study area during 1980-83. No wild turkeys had previously existed in northeast Colorado. METHODS Turkeys were trapped in February 1986-87 with drop-nets and clover traps. Captured birds were classified to age and sex, weighed, measured for carpal and primary feather lengths, and fitted with aluminum leg bands. Turkeys were classified as yearlings « 1 year of age) or adults (> 1 year of age) based on characteristics of P IX and X (Petrides 1942). Females were fitted with transmitters mounted on ponchos (Amstrup 1980) or attached to the central pair of retrices (Bray and Corner 1972). Poncho and tail clip transmitters weighed 29- 32 g and 26-29 g, respectively, and had expected battery lives of 6 months. Turkeys were relocated at least once every 4 days prior to winter flock breakup and dispersal. Once dispersal ceased and individual hens localized, they were relocated every 1-2 days to ascertain nest initiation. Hens that moved in response to my approach were assumed to have not yet initiated incubation. Hens that remained in vegetation clumps as I circled to within 25 m were assumed to be incubating. Locations of incubating hens were flagged so that nests could be found after hatching, depredation, or abandonment. Dispersal distance was measured from the approximate center of the flock winter range to a nest site. Distances were measured parallel to the river because no turkeys were observed> 0.5 km from the riparian area along the river. Differences between ages and years were tested with Mann-Whitney tests. Clutch size and number of hatched eggs were determined from egg shell characteristics. When egg shells were sufficiently intact, �198 lengths and widths were measured with calipers. Clutch and egg size comparisons used data from successful first nest attempts and those unsuccessful first nest attempts whose clutches were observed prior to nest destruction. Differences in clutch size, numbers of eggs hatched, and egg width between age classes were tested using analysis of variance (ANOVA) with age, year, and age-year interactions specified as main effects. Egg widths within a clutch were averaged and the mean value entered into the ANOVA. Nest initiation date as a predictor of clutch size and egg width was examined with linear regression. Hen success was defined as hatching ~ 1 egg during all nest attempts. Differences in hen success between years and age classes were tested with Fisher's exact test. Abandonment caused by observer disturbance and females not attempting to nest were not included in hen success analyses. Nest initiation dates were estimated by counting egg shells and assuming 1 egg laid/day and a 28-day incubation period (Bailey and Rinnell 1967). Initiation dates for unsuccessful nests were estimated by calculating the egg-laying period from the first date of suspected incubation ascertained from activity data. Differences between age classes in nest initiation dates and nest fates were tested with median tests. The power of nest initiation date as a predictor of hen age and nest fate was examined with logistic regression analysis. The relationship between winter body weight and subsequent date of first nest initiation was analyzed using linear regression. The Statistical Analysis System was used for all analyses (SAS Inst. Inc. 1987). RESULTS Five adult and 8 yearling females, and 6 adult and 12 yearling females were radiomarked, respectively, in 1986 and 1987. In 1987 a �199 nest of an unmarked adult was found and the data included in all but the dispersal analyses (age ascertained from egg dimensions). Dispersal distances did not differ between years and the data were pooled. Dispersal began in mid-March with peak movements occurring from the third week in March through the first week in April. Yearlings dispersed farther (n = 18, ~ = 35.5 km, SE = 4.1 km, £= 0.009) from winter flocks to nest sites than adults (n = 11, ~ = km, SE 4.9 km). Three yearling hens dispersed> 14.3 60 km. Within age classes, dispersal was not uniform across the study area (Fig. 1.1). All 12 adult and 19 of 20 yearling females were known to attempt nesting. Adults and yearlings did not differ in date of first nest initiation in 1986, but adults initiated nests earlier than yearlings in 1987 (Table 1.1). Nest initiation date was not a significant predictor of hen age (£ = 0.163). Most (67%) nest initiation occurred between mid-April and mid-May (Fig. 1.2). Fifty-eight percent of nesting hens successfully hatched eggs. Fifteen of 36 nests were destroyed by predators, mostly raccoons (Procyon lotor) and striped skunks (Mephitis mephitis), and 3 were abandoned, 2 of which may have been observer induced. Five hens were attacked while on nests with 1 known mortality, probably by a coyote (Canis latrans). renested twice. Three yearlings were known to renest and 1 adult Hen success was not dependent on age (£ = 0.158). In 1987, 6 of 7 adults hatched eggs compared to only 5 of 12 yearlings, but this difference was not significant (£ = 0.147). of successful first nest attempts (median = Initiation dates 21 Apr) tended to be earlier than unsuccessful first attempts (median = 5 May) (£ = 0.073). Date of nest initiation was a marginally significant predictor of nest fate (f = 0.077). The first nest initiated had a 74 ± 13% (SE) �N o o ( --- --- -..-- .. - ---------------------------------------------_ .. - -- CROOK [J o YAY~( A ~\e DENVER P'" '(\~ y_Yi,,=p,J COLORADO Y A-NEST OF ADULT Y -NEST OF YEARLING F-WINTER FLOCK LOCATION N t r- o FORT MORGAN o l 20 40 KILOMETERS Fig. 1.1. Winter flock and nest locations of wild turkeys in northeast Colorado, 1986-87. of state map represents study area. Shaded portion �201 Table 1.1. Initiation dates of first nest attempts by wild turkeys in northeast Colorado, 1986-87. Year Age Median Range Adult 27 Apr 11 Apr - 6 Jun Yearling 24 Apr 17 Apr - 14 May 16 Apr 10 Apr - 12 Jun 6 May 27 Apr - 2 Jul 20 Apr 10 Apr - 12 Jun 5 May 17 Apr - 2 Jul 1986 0.575 1987 Adult Yearling 0.032 1986-87 Adult Yearling 0.006 a Median test for age differences. • �0 6 W l<l: I- 6 0.8 6 Z - 6 (J) I(/) W 6 0.6 LL 6 0 z 0.4 6 0 6 l- 0 n, • • 6 Z a: 6 0.2 • 0 a: n, • 6 1.0 ! - • • • • • • • • • • • • --• 6 1986 • 1987 6- • 0 30 15 APR Fig. 1.2. 30 15 MAY 30 15 JUN Cumulative nest initiation for wild turkeys in northeast Colorado, 1986-87. N 0 N �203 probability for success whereas the probability for success for the last nest was only 6 ± 9%. Clutch size of adults was not dependent on nest initiation date when all 11 data points were examined (r2 = 0.265, latest nesting adults initiated nests> adult. E= 0.101). The 2 1 month after the next latest The localized behavior of these 2 hens indicated they may have initiated earlier nests that I did not find. With these 2 outliers excluded, nest initiation date was inversely related to both clutch size (r2 = 0.723, £ = 0.004) (Fig. 1.3) and number of eggs hatched (r2 = 0.763, £ = 0.010). Adult clutch sizes (n = 11, ~ = 11.3, SE = 0.68) were not Significantly larger than yearling clutch sizes (n = 10, ~ = 10.0, SE= 0.82, E= 0.233). Numbers of hatched eggs also did not differ between age classes (adults, n yearlings, n = 7, ~ = 8.6, SE = 0.57, = 8, ~ = 10.0, SE 1.0; £ = 0.255). Eggs of adults were wider (n = 12, ~ = 47.97 mm, SE = 0.51, £ = 0.022) than eggs of yearlings (n = 20, ~ = 46.54 mm, SE = 0.31), but there was no relationship between egg width and initiation date or clutch size 0.1, E > 0.10). (£2 < Year and age-year interactions were not significant in any ANOVA (E > 0.10). Poncho-marked birds were not different from tail-clip marked birds in date of first nest initiation, clutch size, number of eggs hatched, or nest success (£ > 0.25 for all tests). Body weight of hens captured in February was a Significant predictor of subsequent first nest initiation date for yearlings but not adults. Lighter weight yearlings nested earlier than heavier yearlings (r2 = 0.226, £ = 0.024). Omission of the 2 latest nesting yearlings appeared justified as these hens did not nest until late June when the effects of winter body condition on nest initiation date were probably inconsequential. Winter body weight accounted for 41% of the �N a p.. 15 I •• 14~ <. 13-j w N (f) I 0 n=9 r2=0.723 • P=0.004 •• 12--i 11 -1 • l- :J 10-, _J • <, 0 • 9--i 8-1 7, ,8 I , 12 I I I , 20 16 " I , 24 I I 28 APR Fig. 1.3. First nest initiation date as a predictor of clutch size in adult wild turkeys in northeast Colorado, 1986-87. �205 variance in nest initiation date for the truncated data set (E = 0.006) (Fig. 1.4). When this data set was examined by years, the relationship was evident in 1986 (n = 7, r2 = 0.525, 10, r2 = 0.041, E= 0.573). E= 0.065) but not in 1987 (n = Mean weight of each age and sex class was lower in 1986 than 1987, although this difference was significant only for adult females (E = 0.056) (Table 1.2). low power due to small sample sizes. Most statistical tests had Thus, small, but real, differences between groups may have gone undetected. DISCUSSION Population dynamics in many birds are affected by differences in reproductive performance between age classes. Compared to yearlings, adult females generally nest more frequently, earlier (Hannon et al. 1982) and with greater success, lay larger clutches (Wallestad and Pyrah 1974, Giesen et al. 1980), and lay bigger eggs that produce young of higher survivorship (Ankney 1980). In wild turkeys, these reproductive parameters are poorly understood. Adult clutch sizes have been reported to be larger (Porter 1978}, smaller (Glidden and Austin 1975, Reagan and Morgan 1980), and not different from yearling clutch sizes (Vangilder et al. 1987). Nest (hen) success was greater in adults in some studies (Vangilder et al. 1987, Vander Haegen et al. 1988) but not in others (Porter 1978, Everett et al. 1980). The high rate of yearling breeding in this study (95%) is consistent with that reported by some investigators (Glidden and Austin 1975, 100%; Porter 1978, 88%; Everett et al. 1980,85%; Vander Haegen et al. 1988, 81%), but in contrast to others (Reagan and Morgan 1980, 31%; Lutz and Crawford 1987, 31%; Wertz and Flake 1988, 0%). In blue grouse (Oendragapus obscurus), non-breeding among yearlings has been attributed to interfemale aggression in high density �N o 0' 30- • ::l 20 0=17 [2=0.410 10 [>=0.006 z J >- « •• 30 20· ~ 10 • • ~ 30 a: Q_ « 20 • 10- J J---T I~ 2.9 3.0 3.1 3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 WEIGHT (kg) Fig. 1.4. Winter body weight as a predictor of nest initiation date in yearling wild turkeys in northeast Colorado, 1986-87. �207 Table 1.2. Body weights (kg) of wild turkeys captured in February 1986-87 in northeast Colorado. Inter-year Sex difference n Age X SE (%)a E_b 4.4 0.358 5.1 0.056 Females Adults 1986 5 3.96 0.11 1987 6 4.13 0.14 1986 11 3.42 0.07 1987 17 3.59 0.05 1986 1 8.00 1987 2 8.45 0.50 1986 8 5.51 0.20 1987 4 5.85 0.24 Yearlings Males Adults 5.6 Yearlings 6.1 0.336 a [(1987 mean weight/1986 mean weight) - 1)] x 100. b 1 test for differences in mean weights between years within age and sex classes. �208 populations (Zwickel 1980, Hannon et al. 1982). The interaction between population densities and reproductive performance, particularly the frequency of non-breeding by yearlings, merits further investigation. Delayed yearling breeding (within seasons) has not previously been addressed in the wild turkey literature, although it has been observed in several grouse species (e.g., sage grouse [Centrocercus urophasianus], Petersen 1980; blue grouse [Dendragapus obscurus], Zwickel 1977, Hannon et al. 1982). The long dispersal distances by yearlings observed in this study are consistent with the interfemale aggression hypothesis of delayed yearling breeding suggested by Hannon et al. (1982) for blue grouse, but data on social behavior are lacking. Also, whether these spring movements constitute dispersal vs. migration is not clear. Most temperate bird species are seasonally constrained to nesting when food is most abundant; specifically, the female must have enough body reserves and/or food available to endure the energetic drain of egg-laying and incubation (Perrins 1970). In Missouri, nest initiation dates of Eastern wild turkeys were correlated with spring temperatures over a 4-year period (Vangilder et al. 1987). In both years of my study, spring temperatures were similar and vegetative growth began in early to mid April, corresponding closely to the first observed nest attempts. Onset of wild turkey nesting is likely proximately controlled by spring vegetative growth or a correlate of this parameter, such as temperature. Winter conditions may affect timing of breeding (Murphy 1986) as well as over-winter survival (Haramis et al. 1986) in birds that rely on endogenous reserves. Although it is not clear to what extent wild �209 turkeys use endogenous reserves, one expects an inverse relationship between body condition (roughly indexed by body weight) and both survival and subsequent nest initiation date (Hannon et al. 1988). Porter et al. (1983) found that wild turkey hens in Minnesota that weighed < 4.3 kg were less likely to survive and less likely to breed than heavier hens. In my study, the low observed mortality (no birds died between 1 week after capture and early May in either year) and early nest initiation by lightweight hens suggests that winter conditions do not impact survival and timing of breeding in this population. Breeding too late may be proximally constrained by lower egg production, hen success, and chick survival. In this study, adults laid fewer eggs and experienced marginally higher rates of nest predation as the season progressed. Toft et al. (1984) suggested that ducks optimally apportion their reserves into either many small eggs or few large eggs, depending upon current probabilities of nest and duckling survivorship. They argued that decreasing survival of young in late season selectively favored large eggs that produced young of high survivorship. This inverse relationship between clutch and egg size, and the assumed optimal allocation of a limited energy supply, has been questioned by Rohwer (1988). He found little correlation between these 2 parameters for most waterfowl species examined. In this wild turkey population, the lack of a relationship between clutch and egg size suggests that egg production may not be limited by energetic constraints. Further, breeding early (Vangilder et al. 1987, Hannon et al. 1988) or laying fewer eggs in later nests would increase the time available to renest. Maximizing the opportunity to renest may be �210 important to these wild turkeys as 42% of all nests initiated were depredated. All 4 hens that were known to renest were eventually successful. Although population density and nest predation have been separately suggested as affecting yearling breeding and clutch and egg size and timing of breeding, interaction between these 2 parameters is probable. I suggest that variation in frequency of yearling breeding may impact wild turkey population dynamics as much as nest depredation. Further studies of reproductive performance in wild turkeys should address these parameters simultaneously. LITERATURE CITED Aldrich, J. W. 1967. Taxonomy, distribution, and present status. Pages 17-44 in O. H. Hewitt, ed. The wild turkey and its management. The Wildl. Soc., Washington, D.C. Amstrup, S. C. 1980. A radio-collar for game birds. Manage. 44:214-217. J. Wildl. Ankney, C. D. 1980. Egg weight, survival, and growth of lesser snow goose goslings. J. Wildl. Manage. 44:174-182. Bailey, R. W., and K. T. Rinnell. 1967. Events in the turkey year. Pages 73-92 in O. H. Hewitt, ed. The wild turkey and its management. The Wildl. Soc., Washington, D.C. Beasom, S. L., and o. H. Pattee. 1980. The effect of climatic variables on wild turkey productivity. Proc. Natl. Wild Turkey Symp. 4:127-135. Bray, o. E., and G. W. Corner. 1972. A tail clip for attaching transmitters to birds. J. Wildl. Manage. 36:640-642. Everett, D. D., D. W. Speake, and W. K. Maddox. 1980. Natality and mortality of a north Alabama wild turkey population. Proc. Natl. Wild Turkey Symp. 4:117-126. Giesen, K. M., C. E. Braun, and T. A. May. 1980. Reproduction and nest-site selection in white-tailed ptarmigan. Wilson Bull. 92:188-199. Glidden, J. W., and D. E. Austin. 1975. Natality and mortality of wild turkey poults in southwestern New York. Proc. Natl. Wild Turkey Symp. 3:48-54. �211 Graves, W. C. 1975. Wild turkey management in California. Proc. Natl. Wild Turkey Symp. 3:1-5. Gray, B. T., and H. H. Prince. 1988. Basal metabolism and energetic cost of thermoregulation in wild turkeys. J. Wildl. Manage. 52:133-137. Hannon, S. J., K. Martin, and J. O. Schieck. 1988. Timing of reproduction in two populations of ~illow ptarmigan in northern Canada. Auk 105:330-338. ______., L. G. Sopuck, and F. C. Zwickel. 1982. Spring movements of female blue grouse: evidence for socially induced delayed breeding in yearlings. Auk 99:687-694. Haramis, G. M., J. D. Nichols, K. H. Pollock, and J. E. Hines. 1986. The relationship between body mass and survival of wintering canvasbacks. Auk 103:506-514. Lindauer, I. E. 1983. A comparison of the plant communities of the South Platte and Arkansas river drainages in eastern Colorado. Southwest. Nat. 28:249-259. Lutz, R. S., and J. A. Crawford. 1987. Reproductive success and nesting habitat of Merriam's wild turkeys in Oregon. J. Wildl. Manage. 51:783-787. Murphy, M. T. 1986. Body size and condition, timing of breeding, and aspects of egg production in eastern kingbirds. Auk 103:465-476. National Oceanic and Atmospheric Administration. 1986. Climatological data, Colorado. U.S. Dep. Comm., Natl. Oceanic and Atmospheric Admin. 91(13):12. Perrins, C. M. 1970. The timing of birds' breeding seasons. 112:242-255. Ibis Petersen, B. E. 1980. Breeding and nesting ecology of female sage grouse in North Park, Colorado. M.S. Thesis, Colorado State Univ., Fort Collins. 86pp. Petrides, G. A. 1942. Age determination in American gallinaceous game birds. Trans. North Am. Wildl. Conf. 7:308-328. Porter, W. F. 1978. The ecology and behavior of the wild turkey (Meleagris gallopavo) in southeastern Minnesota. Ph.D. Thesis, Univ. Minnesota, Minneapolis. 121pp. , G. C. Nelson, and K. Mattson. 1983. Effects of winter -----conditions on reproduction in a northern wild turkey population. J. Wildl. Manage. 47:281-290. Ransom, D., Jr., O. R. Rongstad, and D. H. Rusch. 1987. Nesting ecology of Rio Grande turkeys. J. Wildl. Manage. 51:435-439. Reagan, J. M., and K. D. Morgan. 1980. Reproductive potential of Rio �212 Grande turkey hens in the Edwards Plateau of Texas. Wild Turkey Symp. 4:136-144. Proc. Natl. Rohwer, F. C. 1988. Inter- and intraspecific relationships between egg size and clutch size in waterfowl. Auk 105:161-176. SAS Institute, Inc. 1987. SAS/STAT guide for personal computers, version 6 edition. Cary, N.C. 1028pp. Scott, T. G., and C. H. Wasser. 1980. Checklist of North American plants for wildlife biologists. The Wildl. Soc., Washington, D.C. 58pp. Toft, C. A., D. L. Trauger, and H. W. Murdy. 1984. Seasonal decline in brood sizes of sympatric waterfowl (Anas and Aythya, Anatidae) and a proposed evolutionary explanation. J. Anim. Ecol. 53:75~92. Vander Haegen, W. M., W. E. Dodge, and M. W. Sayre. 1988. Factors affecting productivity in a northern wild turkey population. J. Wildl. Manage 52:127-133. Vangilder, L. D., E. W. Kurzejeski, V. L. Kimmel-Truitt, and J. B. Lewis. 1987. Reproductive parameters of wild turkey hens in north Missouri. J. Wildl. Manage. 51:535-540. Wallestad, R., and D. Pyrah. 1974. Movement and nesting of sage grouse hens in central Montana. J. Wildl. Manage. 38:630-633. Watts, C. R. 1968. Rio Grande turkeys in the mating season. Trans. North Am. Wildl. Conf. 33:205-210. Wertz, T. L., and L. D. Flake. 1988. Wild turkey nesting ecology in south central South Dakota. Prairie Nat. 20:29-37. Zwickel, F. C. 1977. Local variations in the time of breeding of female Blue Grouse. Condor 79:185-191. 1980. Surplus yearlings and the regulation of breeding density in blue grouse. Can. J. Zool. 58:896-905. �213 Chapter 2 NEST HABITAT USE OF RIO GRANDE WILD TURKEYS Avian nest habitat selection may be influenced by many factors including predation (Martin and Roper 1988), inter- and intraspecific competition (Orians 1980), and the thermal environment (Walsberg 1985). To understand habitat selection and the effect of such factors, patterns of nest habitat use must first be documented. For wild turkeys (Meleagris gallopavo), several investigators have recently reported quantitative data on nest habitat use (Lazarus and Porter 1985, Ransom et al. 1987, Wertz and Flake 1988). The variable habitats used for nesting and low sample sizes of these studies precluded elucidation of what criteria wild turkeys may use in choosing nest sites. Nest predation has been implicated by many investigators as a major limiting factor of wild turkey populations (Reagan and Morgan 1980, Speake 1980, Ransom et al. 1987), but the influence of predation on habitat choice is not clear. The objective of my study was to quantitatively document nest habitat use of an introduced population of Rio Grande wild turkeys (M. g. intermedia). I investigated nest site selection by (1) comparing nest sites to random sites at several levels, or scales, and (2) comparing the chronology and success of nests in different vegetative types. �214 STUDY AREA The study was conducted along the South Platte River in Logan, Morgan, and Washington counties in northeast Colorado. This riverbottom community extended to 1.0 km in width and was dominated by an open-canopied plains cottonwood (Populus sargentii) forest. Boxelder maple (Acer negundo), red ash (Fraxinus pennsylvanica), and Russian-olive (Eleagnus augustifolia) occurred in low but increasing frequencies. Common forbs included pepperweed or tall whitetop (Lepidium latifolium), poison hemlock (Conium maculatum), ragweed (Ambrosia spp.), sunflower (Helianthus spp.), and thistle (Cirsium spp.). Common grasses included cheatgrass brome (Bromus tectorum), prairie cordgrass (Spartina pectinata), inland saltgrass (Distichlis stricta), sand drop seed (Sporobolus cryptandrus), and wheatgrass (Agropyron spp.). Shrubs occurred in discrete patches and were predominately western snowberry (Symphoricarpos occidental is), although willows (Salix spp.) were common in mesic areas (plant names follow Scott and Wasser [1980]). Lindauer (1983) provided a complete vegetative description of this particular community. Private lands adjacent to the riverbottom were primarily used for production of alfalfa, corn, wheat, and other small grains and row crops. Cattle were grazed at varying intensities both in and adjacent to the riverbottom. The Colorado Division of Wildlife owned approximately one-third of the riverbottom in the study area. These lands were not grazed and were used for both consumptive and nonconsumptive recreation. Sixty Rio Grande wild turkeys from Kansas and Texas were introduced to the study area during 1980-83. previously existed in northeast Colorado. No wild turkeys had �215 METHODS Wild turkeys were trapped in February 1986-87 with drop-nets and clover traps. Captured birds were classified to age and sex and fitted with aluminum leg bands. Wild turkeys were classified as yearlings « 1 year of age) or adults (> 1 year of age) based on characteristics of P IX and X (Petrides 1942). Females were fitted with transmitters mounted on ponchos (Amstrup 1980) or attached to the central pair of retrices (Bray and Corner 1972). Poncho and tail-clip transmitters weighed 29-32 and 26-29 g, respectively, and had expected battery lives of 6 months. Habitat Measurements Hens were monitored daily, when possible, to ascertain nest initiation. Date of initiation was estimated by calculating the number of eggs laid and the incubation period (Chapter 1). Habitat variables were measured at both nest and random plots to document patterns of nest habitat use. Nest habitat variables were measured within 2 days after eggs hatched or were abandoned or depredated. Measurements of random habitat plots were distributed over the same periods as measurements of nest plots to minimize phenological bias. Eight nest plots and 31 associated random plots from 1986 were remeasured in April 1987 at the approximate time of nest initiation the previous year. All plots were 0.04-ha circles with 22.5-m diameters. Nest plots were centered on nest sites. Two types of random plots were established: adjacent random (AR) and study area random (SAR) plots. Up to 4 AR plots were associated with each nest. These plots were in random distances and directions from the nest with the constraint that they be within 79 m. Thus, in addition to the nest plot, an approximately 2-ha area was sampled around each nest. Study area �216 random plots were spaced at 2.5-km intervals throughout the linear study area. At each interval, SAR plots were established in the riverbottom at a random percentage of the riverbottom's width from the river at that interval. No SAR plots were within 300 m of a nest. Variables measured at nest and random plots were: canopy cover (overstory), understory cover, understory height, amount of shrubs, forbs, grasses, and bare ground, distance to nearest tree> 30 cm in diameter at breast height (DBH), and basal area of all trees and small « 25 cm DBH), medium (~ 25 and < 45 cm DBH), and large (~ 45 cm DBH) trees. Canopy cover was measured with a densiometer. A vegetation profile board (Nudds 1977) was used to estimate percent understory cover to 1 of 6 classes « and> 2.5, 2.5 - 25, 26 - 50, 51 - 75, 76 - 95, 95%) in each of 3 height categories « 0.5, 0.5 2.0 m). 1.0, and 1.1 The profile board was placed at the plot center and read from the plot perimeter in the 4 cardinal directions. Understory height was measured at these 4 perimeter locations and the plot center. Both understory height and basal area were measured in 5 cm classes. Meters of shrubs, forbs, grasses, and ground were estimated along 2 perpendicular, but randomly oriented, transects, each equal to the plot diameter of 22.5 m. In 1987, egg visibility was measured concurrent with habitat measurements at all 22 nests of radio-marked hens. Ten wild turkey eggs were placed in the nest bowl and number of visible eggs counted while standing above the nest, and while standing and crouching 2 and 5 m away in the 4 cardinal directions. were then averaged. The 4 directional measurements �217 Statistical Analyses Habitat measurements at AR plots for each nest site were averaged and these values then paired with associated nest plot data for analysis with Wilcoxon signed ranks tests. If < 2 AR plots were measured at a nest site, then the nest and associated AR plots were excluded from this analysis. Likewise, the 8 nests measured in April 1987 were paired with 1986 habitat data and analyzed with Wilcoxon signed ranks tests. Habitat differences between nest and SAR plots were tested with median tests because the distributional differences between these 2 groups prevented analysis with more powerful MannWhitney tests (Conover 1980). Habitat and egg visibility differences between nests of different age hens, fates (successful vs. unsuccessful), and vegetation types (snowberry vs. other) were tested with Mann-Whitney tests as were differences between AR and SAR plots. Nesting in snowberry vs. other vegetation as a function of nest initiation date was examined with logistic regression. Wilcoxon signed ranks tests were conducted using SPSS (Norusis 1986). The Statistical Analysis System was used for all other analyses (SAS Inst. Inc. 1987). RESULTS Thirty radio-marked hens initiated a known total of 35 nests. Thirty nests were in ungrazed"riverbottom, 3 nests were in riverbottom lightly grazed within the last year, and the 2 latest initiated nests (> 1 1/2 months after median nest initiation date [Chapter 1]) were approximately 200 m from the riverbottom edge in currently grazed pastures. Three hens were known to renest once and 1 hen twice. Excluding the hen that renested after laying a single egg, all 4 renests were in the opposite vegetation type (snowberry vs. other) from �218 the hens' previous attempts. Early nesting hens were more likely to nest in snowberry than late nesting hens (£ = 0.028) (Fig. 2.1). Nest plots measured at hatch (late May - Jun) had greater canopy and understory cover in all 3 height classes than the same nests measured at nest initiation (mid-Apr) the following year (£ < 0.05 for all tests). initiation Height of live grasses and forbs was much less at nest (X = 9, SE = 1) than at hatch (X = 56, SE = 3, £ < 0.001), but height of live shrubs did not vary between nest initiation 100, SE = 8) and hatch (X = 106, SE = 9, £ = 0.722). (X = Adjacent random plots also had greater canopy and understory (~ 1 m) cover at hatch than at nest initiation (£ < 0.05 for all tests). Thirty-one nests (includes 1 nest of an unmarked adult) were compared with their averaged AR plots. Nests were characterized by greater canopy cover, more shrubs, fewer grasses, and greater understory cover (~ 1 m) and height than AR plots (£ < 0.01 for all tests) (Table 2.1). Distance to large tree, basal area of trees, amount of forbs and bare ground, and understory cover> differ (£ > 0.05) between nest and AR plots. 1 m did not Comparing all 36 nests (including 1 nest of an unmarked adult) to SAR plots (n = 36), the same differences were found except that canopy cover and grass abundance did not differ. All significant differences were stronger than those between nest and AR plots. Additionally, AR plots had greater understory cover (> 0.5 and ~ 1.0 m) and height, more shrubs, and fewer large trees than SAR plots (£ < 0.05). Understory cover ~ 0.5 m also tended to be greater at AR plots (£ = 0.068). Other habitat variables did not differ between these 2 types of random plots (£ > 0.10). Nests in snowberry (n = 24) were characterized by significantly greater canopy cover, more large trees, being closer to a large tree, �>a: a: w co S 0 z (f) z CJ ._ z 0.9 0.80.7 -- - -, <, - -- .•...•. z LL 0 ._>- <, <, <, 0=36 P=0.028 <, <, <, <, •••.•...•..•.. <, <, <, <, <, <, <, <, <, ................ 0.6 <, 0.5 <, <, ~ CO 0 a: n, <, <, <, <, <, <, <, <, <, " <, -, 0.4 •..... <, 0.3 <, <, <, <, -, <, <, <, <, <, <, <, <, <, 0.2 _J co <, <, (f) w ...•.. 0.1 <, <, <, <, <, ....••. <, <, - .•...•. 0 1-----~------r-----r__--~----~------~----~----~~ 10 20 30 10 20 30 30 20 10 APR MAY JUN Fig. 2.1. Probability of wild turkeys nesting in snowberry as a function of nest initiation date in northeast Colorado, 1986-87. Regression predicted values ± SE from y = O.0414X - 6.0542, where initiation date was the Julian date. N f-' '0 �220 Table 2.1. Habitat variables at wild turkey nest and random plots in northeast Colorado, 1986-87. Nest Variable Canopy, % Shrubs, m Forbs, m Grasses, m Ground, m Understory cover, % < 0.5 m 0.5 - 1.0 m 1.1 - 2.0 m Understory height, cm Plot center Total plot Distance to large tree, m Basal area, m2/ha Small trees Medium trees Large trees Total trees Adjacent random (0._= 31) X SE Study area random (0._= 36) X SE 33.3 6.6 6.4 8.6 0.5 6.1 1.0 1.0 0.9 0.2 14.9 2.5 5.8 12.4 1.3 2.5* 0.5* 0.7 0.7* 0.3* 29.8 2.2 4.6 11.9 3.3 5.9 0.7+# 0.8 1.1 0.7+ 97 60 8 < 1 5 2 68 30 4* 4* 51 23 6+ 4+# - 6 1 11 2 94 70 6 4 57 54 4* 3* 36 41 5+# 3+# 24.7 7.6 0.7 2.8 6.3 9.9 0.3 0.9 1.9 2.1 27.5 6.9 0.3 1.8 3.5 5.6 0.1 0.5 0.6 0.8 25.1 5.7 0.8 2.1 0.5 3.7 1.3# 7.9 1.6 2.2 a Five additional nests included for statistical comparison to SAR plots. * £ < 0.05 for nest vs. AR plots. + £ < 0.05 for nest vs. SAR plots. # £ < 0.05 for AR vs. SAR plots. �221 more shrubs, and fewer forbs than nests in other vegetation (Table 2.2). Many snowberry clumps contained 1-5 large cottonwoods resulting in many of these differences. Nest failure due to predation was independent of habitat type (snowberry vs. other) when examined across entire seasons (~ test, ~ = 0.502, £ = 0.479). After splitting the data set into early (n =17) and late (n = 16) seasons and excluding 3 abandonments, early nests in snowberry tended to be more successful than late nests in snowberry, but this difference was not significant (Table 2.3). A greater proportion of nests in 1986 (11 of 13) were in snowberry than in 1987 (13 of 23). Nests in 1986 had greater understory cover (> 0.5 and ~ 1.0 m) than nests in 1987 (£ = 0.041) but did not differ for any other habitat variables. Nests of adult hens (n = 14) did not differ from those of yearlings (n = 22) nor did successful nests (n = 18) differ from unsuccessful nests (n = 16, 2 observer-induced abandonments excluded) for any of the measured nest habitat variables (£ > 0.05). Egg visibility from all angles did not differ between age classes, nest fates, or vegetation types. An average of 5.4 eggs was visible from above the nest, < 3 eggs were visible from 2 m, and < 1 egg was visible from 5 m. DISCUSSION Whereas floristic composition at nest sites varies greatly across the wild turkey's geographic range, most investigators have observed similar structural patterns in nest site vegetation. Nests are characterized by concealment in dense herbaceous or woody vegetation, both around and above the nest (Williams et al. 1968, Lazarus and Porter 1985, Wertz and Flake 1988). Similarly, I found that nests of Rio Grande wild turkeys were in denser and taller understory vegetation �222 Table 2.2. Habitat variables at wild turkey nests in snowberry and other vegetation in northeast Colorado, 1986-87. Snowberry ell = 24) SE X Variable Canopy, % Shrubs, m Forbs, m Grasses, m Ground, m Understory cover, % < 0.5 m 0.5 - 1.0 m 1.1 - 2.0 m Understory height, cm Plot center Total plot Distance to large tree, m Basal area, m2/ha Small trees Medium trees Large trees Total trees * £ < 0.05. Other 12) SE X (U = 40.2 9.3 4.4 7.9 0.4 7.2 1.0 0.9 1.2 0.2 14.1 0.5 10.6 10.6 0.5 7.2* 0.3* 1.8* 1.8 0.2 97 62 7 < 1 5 1 94 51 8 2 2 3 92 70 6 5 97 60 9 10 11.5 2.2 59.5 19.4* 0.9 3.3 8.6 12.8 0.4 1.1 2.3 2.5 0.1 0.8 0.5 1.3 0.1* 0.5 0.5* 1.0* �223 Table 2.3. Numbers of successful and depredated wild turkey nests in snowberry and mixed forbs and grasses in early and late season in northeast Colorado, 1986-87. Vegetation Season Fate Early Successful 9 1 Depredated 4 2 Successful 2 6 Depredated 6 3 Late Snowberry Forbs and grasses �224 than the surrounding environment. Low egg visibility and observation of incubating hens substantiated the concealing effect of these understory characteristics. Few nesting studies, however, have examined habitat use at multiple spatial and temporal scales. There is growing support that patterns and processes of habitat selection may differ among areal scales (Wiens 1985, Morris 1987, Martin and Roper 1988). In one wild turkey study that attempted to discern such dissimilarities across scales, Lazarus and Porter (1985) observed no habitat differences between nest sites and surrounding O.5-ha areas. These areas were, however, different from random plots sampled throughout the study area. They suggested that hens may be cuing on O.5-ha locales and then choosing a suitable nest site within this area. In my study area, hens may use a hierarchial nest selection process similar to that discussed by Wiens (1985). At the broadest scale, these hens exhibited a strong preference for riverbottom habitat (94% of all nests) vs. adjoining agricultural uplands. The comparatively sparse vegetation at nest initiation and the high disturbance levels in these agricultural areas were obvious disadvantages of nesting in this habitat type. On a more restricted spatial scale, differences in understory characteristics between AR and SAR plots suggest that perhaps hens were cuing on areas ~ 2 ha before making specific nest site choices. All nests in the riverbottom were surrounded by ~ 2 ha of currently ungrazed, relatively undisturbed land. Use of ungrazed areas may be related to grazing impacts on vegetation. Using unreplicated plots, Crouch (1978) found that cover and height of shrubs, forbs, and grasses were significantly greater on ungrazed than grazed plots in this riverbottom community. �225 Once within a 2-ha area, nest site selection on an even finer scale is suggested by habitat differences between nest and AR plots. Because differences between SAR and AR plots, and AR and nest plots were similar, such differential scalar selection processes should perhaps be envisioned as occurring on a gradient rather than at discrete levels. Alternatively, selection at the 2-ha level may not occur at all, and differences between SAR and AR plots may simply be due to specific nest site selection (i.e., SAR - nest plot differe~ces) and a correlation between AR and nest sites due to their proximity. Implied in this discussion is that hens cue on vegetative structure in the habitat selection process. Confounding influences such as population density, philopatry, and interspecific competition were not addressed in this study. Density effects on habitat selection similar to the Fretwell - Lucas (1969) model do not seem likely. This model requires a threshold density before suboptimal habitats fill and intraspecific nest site competition in optimal habitat occurs. The density of wild turkeys in this study area (-250-300 birds along -140 km of riverbottom in Jan 1987, T. Davis pers. commun.) was much lower than population estimates from other areas (Porter and Ludwig 1980, Hayden 1985, Kimmel and Kurzejeski 1985) and likely below a threshold density. Additionally, what a hen perceives as optimal habitat may change during a season. In mid April, when hens were first initiating nests, the amount of cover provided by snowberry was much greater than that provided by herbaceous vegetation, and thus, was probably more effective at deterring nest predators (Bowman and Harris 1980). As the season progressed, the cover value of forbs and grasses came closer to that of snowberry, and correspondingly, their use as nesting cover increased. �226 If indeed snowberry is more predator safe than mixed forbs and grasses in early season and less so in late season, as suggested by this study, then differential nesting success across seasons and between vegetation types would be expected. Direct interspecific competition with other ground nesting species is likely unimportant as the large body size of wild turkeys would be a great advantage in any aggressive encounter. However, parasitism of wild turkey nests by ring-necked pheasants (Phasianus colchicus) has been observed in this study area (Schmutz 1988). Indirect competition may occur through its effect on nest predation levels. Woody shrubs, primarily snowberry, were less abundant than mixed forbs and grasses, the other major vegetative type used for nesting by wild turkeys. Other species commonly nesting in snowberry were ring-necked pheasants, mallards (Anas platyrhynchos), red-winged blackbirds (Agelaius phoeniceus), and brown thrashers (Toxostoma rufum) (pers. observ.). If a greater density of birds' nests occurred in snowberry than the mixed forbs and grasses, then generalist nest predators would be expected to search these habitats with greater intensity. Additionally, if mixed forbs and grasses offered adequate cover, their greater abundances than shrubs would result in lower probabilities of nest predation in these habitats due to the increased amount of area (potential nest sites) a predator would need to search (Martin and Roper 1988). Therefore, hens may nest in snowberry early in the season because of the substantially greater cover offered by these woody shrubs. Later in the season when mixed forbs and grasses offer equivalent cover, hens may prefer this habitat type because of reduced predation. The observation that all renests after nest depredation occurred in the opposite vegetation type is further support that probability of �227 nest predation influences habitat selection in these wild turkeys. Changing habitat use in response to nest predation has also been observed in other ground nesting species (Storaas and Wegge 1987). Tall, dense understory vegetation, because of its moderation of nest predation, may be the primary cue by which wild turkeys select nesting habitat. LITERATURE CITED Amstrup, S. C. 1980. A radio-collar for game birds. Manage. 44:214-217. J. Wildl. Bowman, G. B., and L. D. Harris. 1980. Effect of spatial heterogeneity on ground-nest depredation. J. Wildl. Manage. 44:806-813. Bray, O. E., and G. W. Corner. 1972. A tail clip for attaching transmitters to birds. J. Wildl. Manage. 36:640-642. Conover, W. J. 1980. Practical nonparametric statistics. John Wiley & Sons, New York, N.Y. 493pp. Second ed. Crouch, G. L. 1978. Effect of protection from livestock on bottomland wildlife habitat in northeastern Colorado. Pages 118-125 in W. D. Graul and S. J. Bissell, tech. coords. Lowland river and stream habitat in Colorado: a symposium. Colorado Chap. The Wildl. Soc. and Colorado Audubon Counc., Greeley. Fretwell, S. D., and H. L. Lucas. 1969. On territorial behavior and' other factors influencing habitat distribution in birds. I. theoretical development. Acta Biotheor. 19:16-36. Hayden, A. H. 1985. Summer baiting as an indicator of wild turkey population trends and harvest. Proc. Natl. Wild Turkey Symp. 5:245-252. Kimmel, V. L., and E. W. Kurzejeski. 1985. Illegal hen kill - a major turkey mortality factor. Proc. Natl. Wild Turkey Symp. 5:55-65. Lazarus, J. E., and W. F. Porter. 1985. Nest habitat selection by wild turkeys in Minnesota. Proc. Natl. Wild Turkey Symp. 5:67-81. Lindauer, I. E. 1983. A comparison of the plant communities of the South Platte and Arkansas River drainages in eastern Colorado. Southwest. Nat. 28:249-259. Martin, T. E., and J. J. Roper. 1988. Nest predation and nest-site selection of a western population of the hermit thrush. Condor 90:51-57. �228 Morris, D. W. 1987. Ecological scale and habitat use. Ecology 68:362-369. Norusis, M. J. 1986. SPSS/PC+ for the IBM PC/XT/AT. Chicago, Ill. 643pp. SPSS Inc., Nudds, T. D. 1977. Quantifying the vegetative structure of wildlife cover. Wildl. Soc. Bull. 5:113-117. Orians, G. H. 1980. Some adaptations of marsh nesting blackbirds. Princeton Univ. Press, Princeton, N.J. Petrides, G. A. 1942. Age determination in American gallinaceous game birds. Trans. North Am. Wildl. Conf. 7:308-328. Porter, W. F., and J. R. Ludwig. 1980. Use of gobbling counts to monitor the distribution and abundance of wild turkeys. Proc. Natl. Wild Turkey Symp. 4:61-68. Ransom, D., Jr., O. J. Rongstad, and D. H. Rusch. 1987. Nesting ecology of Rio Grande turkeys. J. Wildl. Manage. 51:435-439. Reagan, J. M., and K. D. Morgan. 1980. Reproductive potential of Rio Grande turkey hens in the Edwards Plateau of Texas. Proc. Natl. Wild Turkey Symp. 4:136-144. SAS Institute Inc. 1987. SAS/STAT guide for personal computers, version 6 edition. Cary, N.C. 1028pp. Schmutz, J. A. 1988. Ring-necked pheasant parasitism of wild turkey nests. Wilson Bull. 100:In Press. Scott, T. G., and C. H. Wasser. 1980. Checklist of North American plants for wildlife biologists. The Wildl. Soc., Washington, D.C. 58pp. Speake, D. W. 1980. Predation on wild turkeys in Alabama. Natl. Wild Turkey Symp. 4:86-101. Proc. Storaas, T., and P. Wegge. 1987. Nesting habitats and nest predation in sympatric populations of capercaillie and black grouse. J. Wildl. Manage. 51:167-172. Walsberg, G. E. 1985. Physiological consequences of microhabitat selection. Pages 389-413 in M. L. Cody, ed. Habitat selection in birds. Academic Press, Inc. Orlando, Fla. Wertz, T. L., and L. D. Flake. 1988. Wild turkey nesting ecology in south central South Dakota. Prairie Nat. 20:29-37. Wiens, J. A. 1985. Habitat selection in variable environments: shrub-steppe birds. Pages 227-251 in M. L. Cody, ed. Habitat selection in birds. Academic Press, Inc. Orlando, Fla. Williams, L. L, Jr., D. H. Austin, N. F. Eicholz, T. E. Peoples, and R. W. Phillips. 1968. A study of nesting turkeys in southern Florida. Proc. Southeast. Assoc. Game and Fish Comm. 22:16-30. �229 Chapter 3 BROOD HABITAT USE OF RIO GRANDE WILD TURKEYS Many investigators have observed high use of clearings by wild turkey (Meleagris gallopavo) broods. These clearings included agricultural fields (Porter 1980), pastures (Everett et al. 1980), and old fields (Metzler and Speake 1985). Others have reported high use of woodlands and forests by broods (McCabe and Flake 1985, Mackey 1986). In particular, the importance of agricultural habitats is not clear. Whereas McCabe and Flake (1985) found little or no use of grasslands and nearby agricultural areas, Porter (1980) found that wild turkeys in Minnesota spent 45% of their diurnal activity in agricultural fields, particularly corn and alfalfa. Understanding brood habitat needs is vital for effective wild turkey management. The quality of brood habitat may affect brood survival (Metzler and Speake 1985), which can be quite low, and may affect wild turkey population dynamics (Everett et al. 1980, Speake et al. 1985). Habitat data on Rio Grande wild turkey broods (M. g. intermedia) are scant compared to other subspecies of wild turkeys. Although work has recently been conducted on Rio Grande wild turkeys in south Texas, poor nesting success limited data on brood habitat use (Ransom et al. 1987). I documented patterns of brood habitat use in an introduced population of Rio Grande wild turkeys in a Colorado riverbottom community bounded by intensive agriculture. �230 STUDY AREA The study was conducted along the South Platte River in Logan, Morgan, and Washington counties in northeast Colorado. This riverbottom community extended to 1.0 km in width and was dominated by an open-canopied plains cottonwood (Populus sargentii) forest. Boxelder maple (Acer negundo), red ash (Fraxinus pennsvlvanica), and Russian-olive (Eleagnus augustifolia) occurred in low but increasing frequencies. Common forbs included pepperweed or tall whitetop (Lepidium latifolium), poison hemlock (Conium maculatum), ragweed (Ambrosia spp.), sunflower (Helianthus spp.), and thistle (Cirsium spp.). Common grasses included cheatgrass brome (Bromus tectorum), prairie cordgrass (Spartina pectinata), inland saltgrass (Distichlis stricta), sand dropseed (Sporobolus cryptandrus), and wheatgrass (Agropyron spp.). Shrubs occurred in discrete patches and were predominately western snowberry (Symphoricarpos occidentalis), although willows (Salix spp.) were common in mesic areas (plant names follow Scott and Wasser [1980]). Lindauer (1983) provided a complete vegetative description of this particular community. Private lands adjacent to the riverbottom were primarily used for production of alfalfa, corn, wheat, and other small grains and row crops. Cattle were grazed at-varying intensities both in and adjacent to the riverbottom. The Colorado Division of Wildlife owned approximately one-third of the riverbottom in the study area. These lands were not grazed and were used for both consumptive and nonconsumptive recreation. Sixty Rio Grande wild turkeys from Kansas and Texas were introduced to the study area during 1980-83. previously existed in northeast Colorado. No wild turkeys had �231 METHODS Wild turkeys were trapped in February 1986-87 with drop-nets and clover traps. Captured birds were classified to age and sex and fitted with aluminum leg bands. Wild turkeys were classified as yearlings « 1 year of age) or adults (> 1 year of age) based on characteristics of P IX and X (Petrides 1942). Females were fitted with transmitters mounted on ponchos (Amstrup 1980) or attached to the central pair of retrices (Bray and Corner 1972). Poncho and tail-clip transmitters weighed 29-32 and 26-29 g, respectively, and had expected battery lives of 6 months. Wild turkey nests were located with biotelemetry and monitored daily, when possible, to ascertain hatching date. A hand-held Vagi antenna and receiver were used to locate broods to 1 of 3 precision levels: (1) visual location, (2) approaching brood hens to within 100 m, but without flushing, to estimate location coordinates, or (3) approaching only close enough to determine if the brood hen was in the riverbottom, adjacent agricultural areas, or the edge between these habitats. Macrohabitat use was examined by comparing frequencies of all locations across all habitat types, i.e., riverbottom, agricultural uplands, and edge. Edge was defined as being within 30 m of the discrete treeline that bounded the riverbottom. was also noted. Occurrence of grazing Individual broods were not located more than once per day with locations rotated among 3 diurnal periods of equal duration and starting 1/2 hour after sunrise and ending 1/2 hour before sunset. Multiple radio-marked hens in brood aggregations were counted as single data points except for measurement of nest to brood location distances. �232 Macrohabitat frequency data were analyzed with likelihood ratio tests (~ tests) (Sokal and Rohlf 1981:695-697). Microhabitat use was examined at 0.04-ha plots centered on visual locations in the riverbottom. Habitat variables measured at these plots were compared to 0.04-ha plots located randomly within the riverbottom but spaced systematically at 2.5 km intervals across the study area. Visual locations were not used for microhabitat analyses if location bias resulting from observer induced brood movement wa~ suspected. Collection of microhabitat data ceased when broods were 6 weeks old as location bias markedly increased at this age. Individual broods were not flushed more than once every 3 days. Variables measured at microhabitat plots were: canopy cover, understory cover, understory height, frequency of shrubs, forbs, grasses, and bare ground, distance to nearest tree> 30 cm in diameter at breast height (DBH), and basal area of all trees and small « 25 cm DBH), medium (~ 25 and < 45 cm DBH), and large (~ 45 cm DBH) trees. Canopy cover was measured with a densiometer. A vegetation profile board (Nudds 1977) was used to estimate percent cover to 1 of 6 classes « 2.5, 2.5 - 25, 26 - 50, 51 - 75, 76 - 95, and> 95%) in each of 3 height categories « 0.5, 0.5 - 1.0, and 1.1 - 2.0 m). The profile board was placed at the plot center and read from the plot perimeter in the 4 cardinal directions. Understory height was measured at these 4 perimeter locations and the plot center. Both understory height and basal area were measured in 5 cm classes. Meters of shrubs, forbs, grasses, and ground were estimated along 2 perpendicular, but randomly oriented, transects, each equal to the plot diameter of 22.5 m. Differential brood microhabitat use by brood age (young, < 22 days of age; old, ~ 22 days of age) was tested with I-way analysis of �233 variance of the ranked data with time of season blocked (Friedman test, Conover 1981:299). Locations prior to 1 July were classified as early season and all others as late season. Differential brood microhabitat use by time of season with brood age blocked was also tested as was young brood habitat vs. random habitat with time of season blocked. The Statistical Analysis System was used for all analyses (SAS Inst. Inc. 1987). RESULTS Macrohabitat Use Six broods in 1986 and 8 broods in 1987 were located 191 times. Seventy-eight percent of these 191 locations were in the riverbottom, 11% in agricultural uplands, and 11% on the edge between these 2 habitats. This distribution of locations was not constant throughout the day = 11.23, (§ £ = 0.024). Edge habitat was used more than expected in the evening, and agricultural uplands were used less then expected in the afternoon but more than expected in the morning. The contribution of each of these 3 use patterns to the overall § statistic was> 1.70. Use of the 3 habitat types was independent of brood age (£ = 0.228) but was mildly dependent on time of season (£ = 0.077). Agricultural habitats were used less than expected in early season but more than expected in late season. 1057 and her brood. This result was largely due to hen Eggs of this hen hatched on 21 July, approximately 1 1/2 months after the median hatch date (Chapter 1). of this brood's locations were in an alfalfa field. broods located> Eighty percent Of the 7 other 10 times, the next highest use of agricultural habitats was 15%, primarily pastures and old fields. Twenty-seven percent of all locations were on lands that were currently grazed or had visible signs of recent grazing. Grazing �234 intensity was not evaluated, but approximately 30-60% of the riverbottom was grazed. Within the riverbottom, broods used grazed areas more late in the season (£ < 0.001) and when they were older (£ < 0.001). Time of season and brood age were undoubtedly correlated. All 17 broods that hatched in the riverbottom were from nests in thickly vegetated, ungrazed land (Chapter 2). Hens with young broods stayed (X = within 1/2 km of their nest 451 m, SE = 106 m). At about 3 weeks of age many broods made movements to or across grazed areas. Ten brood hens each located> 10 times moved an average of 1,024 m (SE = 366 m) from their nest site to the geometric center of their older brood range. Microhabitat Use Forty-seven locations were obtained from 14 broods, each located 1-9 times. Young broods used habitat with fewer large trees (£ = 0.047) but more shrubs (£ = 0.030) and grasses (£ = 0.040) than older broods, irrespective of time of season. habitats with more shrubs (£ = 0.054) than late season broods. Early season broods used 0.021) and closer to large trees (£ = The limitations of ranked data analysis prohibited examination of a brood age - time of season interaction. Grasses were more abundant on young brood-use (n = 28) than random plots (n = 36, £= 0.035), but all other habitat variables did not differ between plots (£ > 0.10) (Table 3.1). DISCUSSION Daily activity patterns have not been quantitatively documented for wild turkey broods except with human-imprinted poults (e.g., Healy et al. 1975). In the wild turkey population I studied, brood habitat use varied with time of day. The low use of agricultural habitats in the afternoon may have been associated with high temperatures (daily �235 Table 3.1. Microhabitat variables at wild turkey brood-use « 22 days old) and random plots in early and late season in northeast Colorado, 1986-87. Random Brood Early (n = 25) SE X Variable Canopy, % Shrubs, m Forbs, m Grasses, m Ground, m Understory cover, < 0.5 m 0.5 - 1.0 m 1.1 - 2.0 m Understory height, cm Distance to large tree, m Basal area, m2/ha Small trees Medium trees Large trees Total trees * £ < 0.05. Late (n = 3) SE X Early (n_= 29) SE X Late 7) SE X tn = 22.2 0.8 4.0 14.5* 2.8 5.8 0.3 0.7 1.1 0.8 35.2 18.8 0.0 1.8 1.2 16.8 1.3 3.3 0.2 27.6 2.7 4.8 11.2 3.3 6.4 0.9 0.9 1.2 0.8 38.6 15.2 0.1 0.1 4.0 1.3 14.7 2.2 3.1 1.6 54 15 6 5 3 1 60 15 9 13 6 7 48 22 12 6 5 3 65 24 6 12 10 3 43 4 45 13 40 4 44 7 % 40.1 11.4 0.4 2.0 3.3 5.8 0.2 0.6 1.3 1.7 8.3 2.2 0.5 1.2 6.4 8.1 0.5 1.2 3.2 2.3 22.3 5.4 1.6 2.1 4.2 7.9 0.8 0.6 1.5 2.0 35.9 19.3 4.3 2.0 1.4 7.7 2.4 1.3 1.4 2.2 �236 maximum in Jul averaged 33 C). Assuming similarity to willow ptarmigan (Lagopus lagopus), young poults (~ 1 1/2 weeks of age) are not endothermic and thus highly susceptible to heat stress (Pedersen and Steen 1979). Because of their small size, relative to yearling and adult hens, even older poults may have difficulty maintaining thermoneutrality. Goldstein (1984) found that Gambel's quail (Callipepla gambelii) could only stay within the thermoneutral zone by staying inactive under shade from 1000 to 1600 hours. Also, warm temperatures may increase brood susceptibility to predation in open habitats. Red-tailed hawks (Buteo jamaicensis), abundant in the study area, soar and forage most often at temperatures between 28 and 35 C (Ballam 1984). Wild turkey poults switch from an invertebrate to primarily a vegetative diet at 2-4 weeks of age (Blackburn et al. 1975, Healy et al. 1975, Hurst and Stringer 1975). Changes in habitat use at this age have been attributed to this dietary switch (Williams et al. 1973, McCabe and Flake 1985). Healy (1985) reported that invertebrate abundance, poult feeding rate, and vegetation density were highly correlated. He suggested that life form, percent cover, and understory height could be used to predict adequate brood range. In this study, however, understory cover and-height did not differ between young and old brood-use sites. These lack of differences may have occurred because poults had not yet switched to a vegetative diet. Healy et al. (1975) did not observe this switch in human-imprinted poults until 5 weeks of age. Only 15% of the brood plots in my study were measured after this age. Vegetative life form differed between brood ages. Greater amounts of shrubs at young brood plots may have been associated with enhanced �237 protection from predators. Young poults are most susceptible to predation (Speake et al. 1985). Metzler and Speake (1985) observed differential survival of poults in relation to habitat. Additionally, structural complexity of vegetation has been shown to decrease avian predator efficiency (Temple 1987), particularly for snowberry (Sugden and Beyersbergen 1987), the most abundant woody shrub in the study area. Crouch (1978) studied the effects of grazing on vegetation in the South Platte riverbottom community. He found large differences in understory cover and height for shrubs, forbs, and grasses between an unreplicated grazed and ungrazed area. If grazed areas contain more preferred plant material then ungrazed areas, then increased used of grazed areas by older broods would be expected. Blackburn et al. (1975) found that pastures produced more inflorescences throughout the summer than 5 other habitat types sampled for poult forage availability in Alabama. Also, inflorescences in ungrazed areas may be too far off the ground for poults to obtain easily. Both dietary shift and reduced need for predator avoidance may result in higher use of grazed areas. Further research investigating proximate cues used by brood hens in selecting habitat is warranted. Heterogenous habitat use among individuals must be considered. Although numbers of locations were small, 1 brood hen (1057) exhibited distinctly different habitat use patterns than all other brood hens. Given such heterogeneity, habitat management for "average" habitat may not represent habitat used by any broods. Until fitness differences can be attributed to differential brood habitat use, management for single habitat types is not justified. �238 LITERATURE CITED Amstrup, S. C. 1980. A radio-collar for game birds. Manage. 44:214-217. J. Wildl. Ballam, J. M. 1984. The use of soaring by the red-tailed hawk (Buteo iamaicensis). Auk 101:519-524. Blackburn, W. E., J. P. Kirk, and J. E. Kennamer. 1975. Availability and utilization of summer foods by eastern wild turkey broods in Lee county, Alabama. Proc. Natl. Wild Turkey Symp. 3:86-96. Bray, O. E., and G. W. Corner. 1972. A tail clip for attaching transmitters to birds. J. Wildl. Manage. 36:640-642. Conover, W. J. 1980. Practical nonparametric statistics. John Wiley & Sons, New York, N.Y. 493pp. Second ed. Crouch, G. L. 1978. Effect of protection from livestock on bottomland wildlife habitat in northeastern Colorado. Pages 118-125 in W. D. Graul and S. J. Bissell, tech. coords. Lowland river and stream habitat in Colorado: a symposium. Colorado Chap. The Wildl. Soc. and Colorado Audubon Counc., Greeley. Everett, D. D., D. W. Speake, and W. K. Maddox. 1980. Natality and mortality of a north Alabama wild turkey population. Proc. Natl. Wild Turkey Symp. 4:117-126. Goldstein, D. L. 1984. The thermal environment and its constraint on activity of desert quail in summer. Auk 101:542-550. Healy, W. M. 1985. Turkey poult feeding activity, invertebrate abundance, and vegetation structure. J. Wildl. Manage. 49:466472. ______, R. O. Kimmel, and E. J. Goetz. 1975. Behavior of humanimprinted and hen-reared wild turkey poults. Proc. Natl. Wild Turkey Symp. 3:97-107. Hurst, G. A., and B. D. Stringer, Jr. 1975. Food habits of wild turkey poults in Mississippi. Proc. Natl. Wild Turkey Symp. 3:7685. Lindauer, I. E. 1983. A comparison of the plant communities of the South Platte and Arkansas River drainages in eastern Colorado. Southwest. Nat. 28:249-259. Mackey, D. L. 1986. Brood habitat of Merriam',s turkeys in southcentral Washington. Northwest Sci. 60:108-111. McCabe, K. F., and L. D. Flake. 1985. Brood rearing habitat use by wild turkey hens in southcentral South Dakota. Proc. Natl. Wild Turkey Symp. 5:121-131. Metzler, R., and D. W. Speake. 1985. Wild turkey poult mortality �239 rates and their relationship to brood habitat structure in northeast Alabama. Proc. Natl. Wild Turkey Symp. 5:103-111. Nudds, T. D. 1977. Quantifying the vegetative cover. Wildl. Soc. Bull. 5:113-117. structure of wildlife Pedersen, H. C., and J. B. Steen. 1979. Behavioral thermoregulation in willow ptarmigan chicks Lagopus lagopus. Ornis Scand. 10:1721. Petrides, G. A. 1942. Age determination in American gallinaceous birds. Trans. North Am. Wildl. Conf. 7:308-328. game Porter, W. F. 1980. An evaluation of wild turkey brood habitat in southeastern Minnesota. Proc Natl. Wild Turkey Symp. 4:203-212. Ransom, D., Jr., O. J. Rongst~d, and D. H. Rusch. 1987. Nesting ecology of Rio Grande turkeys. J. Wildl. Manage. 51:435-439. SAS Institute, Inc. 1987. SAS/STAT guide for personal computers, version 6 edition. Cary, N.C. 1028pp. Sakal, R. R., and F. J. Rohlf. 1981. Biometry. Freeman and Co., New York, N.Y. 859pp. Second ed. W. H. Scott, T. G., and C. H. Wasser. 1980. Checklist of North American plants for wildlife biologists. The Wildl. Soc., Washington, D.C. 58pp. Speake, D. W., R. Metzler, and J. McGlincy. 1985. Mortality of wild turkey poults in northern Alabama. J. Wildl. Manage. 49:472-474. Sugden, L. G., and G. W. Beyersbergen. 1987. Effect of nesting cover density on American crow predation of simulated duck nests. J. Wildl. Manage. 51:481-485. Temple, S. A. 1987. Do predators always capture substandard individuals disproportionately from prey populations. Ecology 68:669-674. Williams, L. E., Jr., D. H. Austin, T. E. Peoples, and R. W. Phillips. 1973. Observations on movement, behavior, and development of turkey broods. Pages 79-99 in G. C. Sanderson and H. C. Schultz, eds. Wild turkey management: current problems and programs. Univ. Missouri Press, Columbia. �240 Prepared by J"~ II .sG~;1"~ Joel A. Schmutz Approved by f!I~ f. ~ Clait E. Braun �241 JOB PROGRESS REPORT Colorado State of: Project: W-152-R Avian Research Job 9 Work Plan: 13 Job Title: Seasonal Habitat Use by Plains Sharp-tailed Grouse in Douglas County, COlorado Period Covered: Author: 01 January through 31 December 1987 Anthony Hoag Personnel: C. E. Braun, K. Demarest, J. Sarason, Colorado Division of wildlife; T. Hoag, E. Redente, Colorado State University ABSTRACT Plains sharp-tailed grouse (Tympanuchus phasianellus jamesii) movements and habitat were studied in Douglas County, Colorado during Harch-November 1987. Three leks were identified for trapping and radiomarking sharptails. Cherokee Lek had the highest peak count of sharptails visiting the lek with 13 while Dakin Lek had the highest peak count of hens (4) at one time. Sixteen sharptails were captured in walk-in traps and fitted with solar-powered radios. Spring movement data indicated the sharptails moved 2 km from Gambel oak (Quercus gambelii), and true mountainmahogany (Cercocarpus montanus) dominated areas to leks. Summer was the time of least movement with sharptails not moving outside a 2 km2 area. In fall sharptails moved into mountain mahogany and .o akbrush habitats. Home range was, on average, 15 km2 for the year. Habita_tmeasurements identified structural and vegetative components of sharp-tailed grouse use areaS. Canopy cover was highest at nests. Distance to clumps varied with nesting and non-nesting sharptails. The highest distance to cover, on average was 3.2 m,' Clump height, length, and width were highest at roosting sites in spring. ��243 SEASONAL HABITAT USE BY PLAINS SHARP-TAILED GROUSE IN DOUGLAS COUNTY, COLORADO Anthony Hoag INTRODUCTION Plains sharp-tailed grouse occupy suitable habitats from northeastern Colorado and western Nebraska into southern Canada (Aldrich 1963, Bailey and Niedrach 1965). The range of this subspecies has been greatly reduced along its southern periphery largely because former habitats have been altered by agricultural practices (Aldrich 1963, Miller and Graul 1980). Livestock grazing is a dominant land use in the remaining habitat of plains sharp-tailed grouse. Range management favoririg livestock has altered the mixture of grasses, forbs, and shrubs that sharptai1s use for cover (Sisson 1976). In Colorado, plains sharp-tailed grouse historically occupied suitable habitats east of the Front Range of the Rocky Mountains from Larimer County (Cooke 1897, Sc1ater 1912) south into E1 Paso County (Aiken and Warren 1914) and east to Kit Carson, Lincoln, and Yuma counties (Cooke 1897, Sc1ater 1912). Within this area, specimens are available in the Denver Huseum of Natural History from Arapahoe, Douglas, Elbert, and Yuma counties (Bailey and Niedrach 1965). This race of sharptai1s historically was apparently most abundant along the foothills in Larimer (Cooke 1897, Sclater 1912), Boulder (Henderson 1909), Douglas (Bailey and Niedrach 1965), and E1 Paso counties (Aiken and Warren 1914). llothCooke (1897) and Sc1ater (1912) reported that sharp-tailed grouse were not common in Colorado. Apparently, the species became uncommon between 1877 and 1887 (Cooke 1897). The plains race of sharptai1s (!. £. jamesii) was described from Colorado by Lincoln (1917) and this name was applied to all sharp-tailed grouse in eastern Colorado by Snyder (1939). Throughout the greater part of its range, sharp-tailed grouse are inhabitants of the ecotone between forest and prairie with woody vegetation in the form of shrubs and trees considered an essential component of its habitat (Grange 1948). Sclater (1912), Baumgartner (1939), Bailey and Niedrach (1965), Hillman and Jackson (1973), and Moyles (1981) reported sharptails used grassland and grassland-shrub transition zones throughout the year. In spring, courtship occurs on ridges and hills with short sparse vegetation (Baumgartner 1939, Douglas 1942, Kobriger 1965, Sisson 1976) at the center of the lek and escape cover within 200 m (Pepper 1972). When not on leks, hens and cocks feed or rest in ecotone edges (Baumgartner 1939). Schiller (1973) reported nests of sharptails in grassland, brush1and, and cropland. Artmann (1970) found extensive use of brush1and cover by nesting sharptails and Pepper (1972) found nests in native grass-shrub mixtures. Sharptail hens with broods usually select sites with some overhead cover nearby (Amman 1957) and in grassland interspersed with shrubs (Hamerstrom 1963, Bernhoft 1969, Pepper 1972, Schiller 1973, Sisson 1976, Kohn et al. 1982). Habitats used by sharptails during summer are dominated- by grass and shrubs (Aldous 1943, Hamerstrom and Hamerstrom 1951, Buss and Dziedzic 1955). �244 Moyles (1981) reported that sharptails use grassland or grassland-shrub communities during fall. These communities offer both food (insects and fruits) and roosting cover. In winters with snow, sharptails roost singly or singly in small flocks in snow beneath scattered trees (Baumgartner 1939, Hamerstrom and Hamerstrom 1951, Kirsch et ale 1973). In mild weather, flocks disperse and travel distances up to approximately 2 km (Hamerstrom and Hamerstrom 1951). Previous work on plains sharp-tailed grouse has provided valuable data for the subspecies in major areas of its range. However, data on the relict plains sharp-tailed grouse population in Colorado are limited (Stearns 1968). If this population is to be maintained in perpetuity, knowledge on how it presently survives in the remaining habitats is essential. The objectives of this study are to quantify seasonal habitat use and movements of plains sharp-tailed grouse in Douglas County, Colorado. Hypotheses HOI: H02: being tested are: No changes occur sites. changes among sharptail use Ha: Use site plant composition Hb: Vegetation at escape, nesting, brooding, feeding and loafing, and roosting sites has proportionately more shrubs than at random sites. seasonally. Plant physiognomy (canopy cover, plant height, clump distance, height, clump width, and clump length) does not differ between points or between use sites. Ha: H03: in plant species composition Nest, brood, feeding and loafing, and escape cover have greater canopy cover, plant height, clump height, clump width, clump length, and less distance to clumps than at random points or between use sites. Movements activity. Ha: clump random do not differ by age class, sex, season or biological Movements are greatest in spring and fall. P. N. OBJECTIVES The objectives of this study are to quantify seasonal habitat use and movements of plains sharp~tailed grouse in Douglas County, Colorado. �245 SEGMENT OBJECTIVES 1. Review literature relating to grouse, plains sharptails, habitat selection, movement, capture and marking techniques, and methods analyzing vegetation. trap, and radio mark 20 plains sharp-tailed for 2. Locate, 3. Quantify and compare use site vegetation with random point vegetation classified by season, activity, and sex and age of sharp-tailed grouse. 4. Quantify 5. Map the vegetation of the study site and identify used by plains sharp-tailed grouse. 6. Compile and analyze data and prepare aNnual report. daily and seasonal grouse on 4 leks. movement's of sharp-tailed grouse. the habitat components STUDY AREA The study area is in the foothills and tablelands of Douglas County, Colorado (Fig. 1). Elevation ranges from approximately 1,850 to 2,200 m. The climate is of the high inland continental type as modified by the Rocky Mountains and the Palmer Divide (U.S. Dep. Agric. 1974). General climatic characteristics include low precipitation, low humidity, variable winds, and a wide temperature range. The average annual precipitation is 37-44 cm. May is the wetest month receiving 16-20% of the annual moisture (U.S. Dep. Agric. 1974). Winter (Dec-Feb) is the driest period receiving only 8-9% of the annual moisture (U.S. Dep. Agric. 1974). Temperature ranges from -30 to 38 C with a mean annual temperature of 19 C (U.S. Dep. Agric. 1974). Upland areas are moderately rolling to broken tablelands dissected by tributaries of Plum and Cherry creeks. Soil parent materials vary widely. The primary parent materials are calcareous and arkosic fans and pedisediments; red arkose, sedimentary, and red sedimentary bedrock; and Eolian deposits. Soils are loam to clay in the western section and sandy to gravelly in the eastern section of the study site (U.S. Dep. Agric. 1974). Gambel's oak and true mountainmahogany are the dominant shrubs in upland areas. Fragrant sumac (Rhus aromatica) and rabbitbrush (Chrysothamnus spp.) are common understory shrubs. Understory forbs include Astragalus, Melilotus, uescurainia, and Yucca. Dominant grasses in upland areas include Bromus, Poa, Koeleria, and Agropyron. The primary land use is ranching, housing developments, and agriculture. METHODS Plains sharp-tailed grouse at 3 active leks were chosen for study in 1987. Walk-in traps with chicken wire leads connecting traps to direct sharptails into them were placed in the middle or at the periphery of the lek. Captured sharptails were fitted with solar-powered poncho-mounted radiQs (150-152 Mhz) (Amstrup 1980), serially-numbered aluminum leg bands, and plastic bandettes �N ..,.. 0'1 T6s NORTH I 7 8 • • KNOWN 9 ACTIVE LEKS DOCUMENTED BIRD PRESENCE 10 ~ 70 R71W Fig 1. 69 Plains sharp-tailed 68 67 grouse distribution 66 in Douglas 65 County, Colorado. STATUS UNKNOWN �247 color coded to designate lek of capture, sex, and age. Captured birds were classified to sex and age, and measurements were taken of weight, carpal length, and length of primary feathers I-X. In spring 1987, 12 males and 3 females were captured in walk-in traps (7 males and 1 female at Cherokee Lek, 2 males and 2 females at Dakin Lek, 3 males at Lincoln Mountain Lek). One male was captured using a spotlight and a long-handled net (Giesen et ale 1982) while roosting near the Dakin Lek. Radiolocations were obtained at leks, feeding and loafing, escape, roost, nest, and brood sites (activity sites) using a portable receiver and a 3-element Yagi antenna. Locations were determined by triangulation or by flushing the birds and plotted on 7.5-minute U.S. Geologic Survey topographic maps by season and activity. Size of study sites was delineated after movement data from the first year were available. Weather conditions, physiognomic features, general vegetation, and flock size were recorded at each site where birds were encountered. Vegetation measurements were taken separately for each biological activity and season. A polygon was drawn around concentrations (>2) of observations and the resulting areas were designated as use sites. A transect, started from a random point produced by a random number generator, was placed in each use polygon with vegetation measurements taken at 25-m intervals. If the edge of the transect was met before an adequate sample (Sokol and Rohlf 1981) was obtained, a second random transect was started. Each vegetation sample included canopy cover, highest plant height within 1 m of the random point, distance to clump cover (defined as any live or residual vegetation that could supply a sharp tail with at least 50% cover taken at one direction on a spherical densiometer), and height, width, and length of clump cover. Vegetation at random points and in clumps was identified following Harrington (1954). All measurements were taken once in each of the 4 cardinal directions from--the random point. Data analysis was by comparing means. RESULTS Trapping The most successful method used to trap plains sharp-tailed grouse was the walk-in trap. Twenty-five of 26 captured birds were caught in walk-in traps, while one was caught using a long-handled net and spotlight (9 of 26 captures were recaptures). The most effective set-up to direct sharptails into traps was 3 separate lines of traps placed in the center of the lek. Thirteen of 16 individuals captured were adult males with an average weight of 933 grams. Carpal length varied from 225 to 240 mm for males (x = 233 for adults, 225 for yearlings) and 210 to 222 mm for females. Average primary feather lengths ranged from III (PI, yearling males) to 168 mm (p VII & VIII, adult males) (Table 1). �248 Table 1. Measurements of plains sharp-tailed grouse in Douglas County, Colorado, 1987. Males Category Ad Yrlg Females Sample size Carpal length, mm Weight, <:>a Primary length, mm X IX VIII VII VI V IV III II I 11 233.2 933.2 3 225.0 886.7 3 217.0 775.0 128.8 154.8 168.0 168.0 166.0 161.0 137.0 126.0 120.0 113.0 124.0 145.0 162.0 162.0 159.0 147.0 131.0 123.0 113.0 111.0 122.0 147.0 160.0 162.0 163.0 155.0 135.0 123.0 117.0 112.0 The numbers of sharptails at each lek varied from a high of 11 males and 2 females at Cherokee to a low of 7 males and 0 females at Lincoln Mountain. The Dakin Lek had the most females on the lek at one time with 4 present for a 2-day period (2 were captured and fitted with radios). One possible reason more hens did not appear on the Cherokee Lek is that they were bypassing it for the Woodhouse Lek to the northwest. One radio-marked hen captured on the Cherokee Lek was found predated 2.4 km north of the Woodhouse Lek. Movements Radiotracking was used to identify the daily and seasonal movements and habitats of plains sharptails. The birds at Cherokee and Lincoln Mountain were initially found feeding, loafing, roosting, and escaping to tall dense oakbrush clumps (Tables 2, 3). Lek activity occurred in areas virtually devoid of vegetation. As the breeding period advanced, males stayed closer to leks selecting mountainmahogany for feeding, loafing, and escape cover. Roosting occurred on ridges with residual forb and grass cover that were near leks. Males spent most of the day on or near leks as the breeding period peaked, often displaying during 2-3 periods/day on the Cherokee Lek. When breeding ended, non-nesting sharptails moved to areas dominated by cool season grasses. Often birds were flushed from small dense clumps of cheatgrass (Bromus spp.) and flew to draws with mountainmahogany and oakbrush. Sharptails remained in habitats dominated by grasses throughout the summer and into the beginning of fall. As grasses senesced and began to lodge, sharptails moved into mountainmahogany and oakbrush habitats where they spent the fall (Tables 2, 3). �249 Table 2. Dominant plant species at sharp-tailed grouse use sites, 1987. Cherokee Lek Species % Spring (Mar-Hay) N = 38 Quercus gambelii Grasses Poa fendleriana KOeleria cristata Descurainia sophia Yucca glauca 0.440 0.237 0.158 0.078 0.0526 0.0263 Bromus spp. Poa fendleriana Symphoricarpos albus Cercocarpus montanus 0.788 0.091 0.091 0.031 Summer (Jun-Aug) N = 68 0.553 0.106 0.082 0.071 0.059 0.047 0.029 0.029 0.012 0.011 Fall (Sep-Nov) N = 36 Quercus gambelii Cercocarpus montanus Rhus aroma tica Poa fendleriana BrOmus spp , Yucca glauca % Spring (Mar-May) N = 33 Summer (Jun-Aug) N = 170 Bromus spp. Koeleria cristata Cercocarpus montanus Poa fendleriana YUCca glauca Quercus gambelii Stipa comata Helianthus spp. Astragalus drummondii Agropyron smithii Dakin Lek Species Poa fendleriana ChTysothamnus spp. Koeleria cristata Bromus spp. Forbs Cirsium spp. Quercus gambelii 0.309 0.206 0.191 0.118 0.103 0.044 0.029 Fall (Sep-Nov) N = 22 0.600 0.200 0.066 0.066 0.033 0.033 Bromus Stipa comata Lupinus spp. 0.550 0.275 0.131 �250 Table 3. Clump species at sharp-tailed grouse use sites, 1987. Cherokee Lek Species % Spring (Mar-May) N = 38 Quercus gambelii Yucca glauca Symphoricarpos albus Forbs 0.526 0.368 0.079 0.0263 Yucca glauca Cercocarpus montanus Symphoricarpos albus Forbs 0.510 0.358 0.091 0.062 Summer (Jun-Aug) N = 59 0.414 0.279 0.129 0.085 0.057 Chrysothamnus spp. Yucca glauca Cirsium spp. Forbs Lupinus spp. 0.356 0.288 0.186 0.119 0.033 Fall (Sep-Nov) N = 22 Fall (Sep-Nov) N = 36 Quercus gambelii Cercocarpus montanus Rhus aroma tica Yi.'iCCa glauca % Spring (Mar-May) N = 22 Summer (Jun-Aug) N = 140 Bromus spp. Yucca glauca Cercocarpus montanus Quercus gambelii Rhus aromatica Dakin Lek Species 0.600 0.300 0.066 0.033 Bromus spp. Stipa comata Lupinus spp. 0.550 0.275 0.181 Sharptail home ranges were calculated for m1n1mum convex and concave polygons and harmonic means transformation contours were drawn for each bird (Stuwe and Blohowisk 1985). Contours were calculated for 10, 25, 50, 75, 95, and 100% (Appendix). Home ranges at the Cherokee and Dakin sites increased significantly as the number of observations increased. Birds that survived the breeding season moved to summer habitats not included in spring locations leading to the increased size of the home range (Table 4). Home ranges of males were larger and more continuous than those of females. Hens moved considerable distances from leks to nest, and remained near the nesting site with their brood throughout the summer. Male sharp-tailed grouse remained within 1 km of the lek throughout summer and fall. �251 Table 4. Home range size estimates (ha) for plains sharp-tailed grouse in Douglas County, Colorado, 1987. Age Sex N· Convex polygon Concave polygon M M M M M M M 23 26 36 47 63 30 21 102.04 82.20 105.76 162.99 207.20 144.14 149.77 33.87 43.90 49.77 114.90 79.58 85.04 62.21 M M F F 18 26 13 47 46.20 50.25 132.60 179.96 18.00 34.63 1.90 10.58 Cherokee Ad Ad Ad Ad Ad Ad Yrlg Dakin Ad Yrlg Ad Ad During the breeding season male sharptails were not observed farther than 2 km from the lek. From June through September sharptails were sedentary, rarely moving outside a 2 km2 area. During October through November sharptails moved from grass/forb habitats used in summer into mountainmahogany and oakbrush habitats. Weekly radiolocations indicated sharptails moved through different oakbrush-dominated draws near the Cherokee Lek. Young produced near Dakin Lek separated from the hen in late August, but remained in the same valley as the hen throughout fall. HABITAT Habitat data showed wide variation in variables between activity sites, seasons, and general lek sites. The data from Cherokee Lek were based on 7 males in spring, 3 males in summer, and 2 males in fall. Data from Dakin were based on 3 males and 2 females in spring, and from 1 hen and 7 chicks in summer and fall. The 3 radio-marked males from Lincoln Mountain were predated before sufficient data had been collected. Seasonal Habitat Canopy cover was highest in areas used by the hen with chicks during summer. Cover at use sites in spring and fall for Dakin and Cherok~e leks was between 21 and 25%. Plant height was lowest for sharptai1s on leks and also at Cherokee Lek in summer where birds moved into habitats dominated by grasses (Bromus, Poa, Koeleria, and Agropyron). The spring and fall clump data reflected the dominance of mountain mahogany at Dakin Lek and oakbrush at Cherokee Lek. Clump distance was lowest for the hen and chicks near Dakin Lek during summer. All other clump distances were roughly equal for birds marked �252 at both leks for all seasons. Differences between clump width, length, and height were greatest between seasons with all 3 variables decreasing between spring and summer and increasing in fall (Table 5). Table 5. Seasonal habitat measurements at plains sharp-tailed grouse use sites, 1987. Spring Canopy cover, % Plant height, m Clump distance, m Clump width, m Clump length, m Clump height, m Summer Cherokee N_ = 110 Dakin N_ = 45 Cherokee N = 170 Dakin N_ = 60 25.2 0.816 2.55 4.15 3.99 1.19 25.0 0.722 3.66 '4.54 4.77 0.954 21.7 0.554 2.86 1.705 1.625 0.805 45.0 0.812 1.36 3.207 3.24 0.886 Activity Site Habitats Canopy cover was highest at feeding, loafing, and flushing sites (23%) and lowest for roosting sites at Cherokee Lek. Nesting canopy cover was highest near Dakin Lek (79%). Canopy cover at flush sites was lower than would have been predicted, but measurements were averaged over all flush areas including samples taken from shortgrass areas between clumps (Table 6). Plant height was highest for roosting sites at Cherokee Lek and feeding and loafing sites at Dakin Lek. Plant height averages were equal for all other activities (Table 6). Average distance to clumps was generally opposite for the Cherokee and Dakin leks. Distances were lowest for roosting at Cherokee Lek and highest at Dakin Lek. The highest distance to clump cover at Cherokee Lek was at flush sites. This was the second lowest distance at Dakin Lek (nesting was lowest distance to cover). For other activities, the furthest average distance to clump cover was 3.2 m (Table 6). Clump width, length, and height decreased from roosting sites to feeding and loafing sites, and to flush sites at both leks. A typical sharptail roosting site had greater plant height, clump height, width, and length, and less distance to clump cover than feeding, loafing, and flush sites. The average canopy cover (or plant density) at roosting sites was less than at all other activity sites except lek sites (Table 6). �253 Table 6. Characteristics of activity site habitat variables for plains sharp-tailed grouse, 1987. Characteristic Canopy cover Cherokee Dakin Roosting x N 19.44 28.6 33 34 Feeding and loafing x N 23.26 47.9 152 41 Flush x 23.1 40.1 Brood N 46 27 x 42.3 Nest N 50 x 71.5 Plant height Cherokee Dakin 0.829 0.694 0.590 0.767 0.597 0.7204 1.52 0.82 Clump distance Cherokee Dakin 1.303 3.2 2.95 1.28 2.87 0.903 1.54 0.00 Clump width Cherokee Dakin 4.44 4.27 1.85 2.98 1.50 2.71 2.85 15.25 Clump length Cherokee Dakin 4.37 4.56 1.66 2.86 1.64 2.23 3.04 12.29 Clump height Cherokee Dakin 1.33 0.91 0.86 0.814 0.79 0.856 0.825 0.82 N 2 �254 LITERATURE CITED Aiken, L. E. H., and E. R. Warren. 1914. The birds of El Paso County, Colorado. Colorado ColI. Sci. Series 12:455-603. Aldous, S. E. 1943. Sharp-tailed grouse in the sand dune country of north central North Dakota. J. Wildl. Manage. 7:23-31. Aldrich, J. W. 1963. Geographic orientation of American Tetraonidae. Wildl. Manage. 27:529-545. Amman, G. A. 1957. The prairie grouse of Michigan. Tech. Bull. 200pp. Amstrup, S. C. 215-216. 1980. A radio collar for game birds. J. Michigan Dep. Conserve J. Wildl. Manage. 44: Artmann, J. W. 1970. Spring and summer ecology of the sharptail grouse. Ph.D. Thesis, Univ. Minnesota, Minneapolis. l52pp. Bailey, A. M., and R. J. Niedrach. Mus. Nat. Hist., Denver, Colo. 1965. Birds of Colorado. 454pp. Vol. 1. Denver Baumgartner, F. M. 1939. Studies on the distribution and habits of the sharptail grouse in Michigan. Trans. Am. Wildl. Conf. 4:485-490. Bernhoft, L. S. 1969. Reproductive ecology of female sharp-tailed grouse (Pedioecetes phasianellus jamesii) and food habits of broods in southwestern North Dakota. M.S. Thesis, Univ. North Dakota, Grand Forks. 96 pp. Buss, I. 0., and E. S. Dziedzic. 1955. Relation of cultivation to the disappearance of the Columbian sharp-tailed grouse from south-eastern Washington. Condor 57:185-187. Cooke, W. W. l44pp. 1897. The birds of Colorado. Colorado Agric. ColI. Bull. 37. Douglas, D. W. 1942. A pralrle chicken booming grounds survey in central Michigan. Wilson Bull. 54:171-172. Giesen, K. M., T. J. Schoenberg, and C. E. Braun. 1982. Methods for trapping sage grouse in Colorado. Wildl. Soc. Bull. 10:224-231. Grange, W. B. 1948. Wisconsin grouse problems. Wisconsin Conserve Dep. Fed. Aid Wildl. Rest. Proj. 5R, Publ. 328. 3l8pp. Hamerstrom, F. N., Jr. 1963. Sharptail brood habitat in Wisconsin's northern pine barrens. J. Wildl. Manage. 27:792-802. _____ , and F. Hamerstrom. 1951. Mobility of the sharp-tailed grouse in relation to its ecology and distribution. Am. MidI. Nat. 46:174-226. �255 Harrington, H. D. Denver, Colo. 1954. Manual of the plants of Colorado. 666pp. Sage Books, Henderson, J. 1909. An annotated list of birds of Boulder County, Colorado. Univ. Colorado Studies 6:219-242. Hillman, C. N., and W. W. Jackson. 1973. The sharp-tailed grouse in South Dakota. South Dakota Dep. Game, Fish and Parks Tech. Bull. 3. 64pp. Kirsch, L. M., A. T. Klett, and H. W. Miller. 1973. grouse population relationships in North Dakota. 37:449-453. Land use and pralrle J. Wildl. Mange. Kobriger, G. D. 1965. Status, movements, habits, and foods of prairie grouse on a sandhills refuge. J. Wildl. Manage. 29:788-800. Kohn, S. C., R. L. Linder, and G. D. Kobriger. 1982. Sharp-tailed grouse nesting habitat in south-western North Dakota. Pages 166-164 in J. N. Peek and P. D. Dalke, eds., Proc. Wildlife-Livestock Relationships Symp. Univ. Idaho For., Wildl. and Range Expt. Stn., Moscow. Lincoln, F. C. 1917. A review of the genus Pedioecetes in Colorado. BioI. Soc. Washington 30:83-86. Proc. Miller, G. C., and W. D. Graul. 1980. Status of sharp-tailed grouse in North America. Pages 18-28 in P. A. Vohs, Jr., and F. L. Knopf, eds. Proc. Prairie Grouse Symp. Oklahoma State Univ., Stillwater. Moyles, D. L. J. 1981. Seasonal and daily use of plant communities by sharp-tailed grouse (Pedioecetes phasianellus) in the parklands of Alberta. Can. Field-Nat. 95:287-291. Pepper, G. W. 1972. The ecology of sharp-tailed grouse during spring and summer in the aspen parklands of Saskatchewan. Saskatchewan Dep. Nat. Resour. Wildl. Rep. 1. 56pp. Schiller, R. J. 1973. Reproductive ecology of female sharp-tailed grouse (Pedioecetes phasianellus) and its relationship to early plant succession in north-west Minnesota. Ph.D. Thesis, Univ. Minnesota, Minneapolis. l86pp. Sclater, W. H. 1912. A history of the birds of Colorado. London, U.K. 576pp. Sisson, L. 1976. The sharp-tailed grouse in Nebraska: Nebraska Game and Parks Comm., Lincoln. 88pp. Snyder, L. L. 1939. 56:184-185. Witherby and Co., a research study. Great plains races of sharp-tailed grouse. Sokal, R. R., and F. J. Rohlf. 1981. San Francisco, Calif. 859pp. Biometry. Auk W. H. Freeman and Co., �256 Stearns, F. D. 1968. Sharp-tailed grouse density in east-central Colorado related to selected habitat factors. M.S. Thesis, Colorado State Univ., Fort Collins. l33pp. Stuwe, M., and C. E. Blohowisk. 1985. Micro-computer analysis for animal locations. Conserve Res. Center, Natl. Zool. Park, Smithsonian lnst., Washington, D.C. 2lpp. u.S. Department of Agriculture. 1974. Soil Survey of Castle Rock area, Colorado. u.S. Dep. Agric., Soil Conserve Serv., Washington, D.C. ~~ Prepared by ~~~~~~~~~~~. - ~~~~~~/_. Anth~ ~ Graduate Research Assistant Approved by Clait E. Braun Wildlife Research Leader _ �257 APPENDIX �i r.---". ='~ I I··--~ 1 i ; '1 --J; 2 ) .-.-. ",'. ".'.I,. •• ' •••• I J •••• e-.'.'' ISI' !\ \ r ..- I I' 1 Ii ", e.; ..,....\ -G • .".J ~_ _~- l r= ~-=:,tl I . -eo •••. i .....' 258 .p' -c- // ~.~ J i' J! /', ~.:=: II i'~.. ¥. ;;;! .-__ I I.... ~ 1•• _. "1 , ~~I\i.. ~ ~J c: o ...~•. ':;' $~ (;'.::l ...,._ ~ . .>: • -~. '! ~, . _ .... .•-..-.--._-------."=, 1'-.-. -. le." I ~. j :- .;."'.... l_ -~ -" -4 ----------_. .....• ,/ 't ., /\ li il ,I Il l--.=.! 7.l ....~l ,~t .• .. .._- .-~-.-. --_.... .'-. ----.. ' -._--- .. LI"\ . r--. CO ...-l LI"\ o ~ �c;s ".-1 -.+-,;. n:;; ~ _ ..o:>.:.. v; ".c=s:.... ~ i:= 11'$ ~ ..•....lloJ ..••. ~;;... ._.. -= Cl ::E >. I"'lS -- ---_.----.._ II ~ CJl o ~ •.-1 -I.J CJl 1-1 Cll ..0 o zl . o U"\ ,....j ~ 259 �c 0 o_ ~ ::E n:: s:.. •••••• 0 e.... c: n:: s:.. t:-c: n:: 2: c.:> CLI ._ c: 0 :E s:.. n:: :::c 260 o . f~ I i, I ,I I " I ' or"" »<>: ------. I•• J " '. ,II / •••••• - ». I' _II I{'"'.;J!.~:-:·", -...•_.-.-- ·I.:-!~;"•...- .:•at:_(.~.-"'" .-".-_,., ~ ~.. ---~-------_/ .--~~ �.~-c c;J ""1 r:..1:= I(Y~ ~ r--. c= a.~ r,::i ::£,: ._c= o _- ;..£ :<-i .'S:; :::z::: __ --.---- --' M------- _'C::1 •• ~,: ---!--=. •.. ----------.,. ~ ."~"... i .-.-.=-.----~----------.------• - ... ~ >'- 1 -, .. ~ 261 �c:: i~. • -:-01 .-4-" .::£: 1"'-:: ' .t -0 ;;w..c: 1..1"': ~~ c: $:< ,.0'.:i ~ C ,"£:I f~ ~.:- ._. .~ :..£ c.::: ~ .~ t';:i -.- 262 jl .I II I I ,J ••• I l ; • . ' _ _ .. ..' '..•...._-_ i• ". '• '. _ -___ .-------- _ --_ a:a. a_~ __ • _ .-._,. -'J,. •••••. '~"··c. " 2. • iI .---_--_ '. fl ~.;.. -~II~·I~:: ...• '"_ .> ..-- .-•...•_ -. ..... "--. ,,~. -.,,_ I~ .j ,.-. _ ••' . ..... .... "..-, .• ". .. ...". •..• : --... . • .•- .,. .: .. s I I -----.-- -:. I ".. --. I S Ir I a ••I I I I ! .... '.. •• ~ --. I Ii( ,'I! I \~~====::_~_=--~.:~~~:::~~._ .._. co ....; �._ ;-.:;i ..•..•.. ;:.. :£ CI "-.• :.:-4 i..i""l - .. :0 ._ c=.~ :.!: :7.0<l~; .= _.- ..•- .----- _0- ----- _-"_ ---_ ..---~- -._ 263 �264 .__ ....,..• », ------_. t. {. I.. J '. ~ ~ i l c:: CI j , i I .,..... I"'CA .."._. ~~ ~:.:.. L-"'"1 ':s.. ..:.,,_ c: r-ci :;:.. ~ C a.,. ••• ..,:i (:.,:J -WiO!'- ._ -=c= o >.. ::E t"!S == �t~ j!z o >... E...-"'1 <f!5 c:: ~ E-4 c o -_ o ~ ~ •.ti M----- --.----. -~~-..--.----.- ~ 265 �o -<-. In C !I""!::' ~ ~ -_ 266 o M " ell C !> CIS o ~ ~ ~ <lJ ell ,.0 o ZI �c.: ",~ -_ , ...t:."i .....p...; .• Cl ~:£ $~ i..•, -•••..--4 c:: r:; ~ '::,--,:i c:: •...• .:J -~ Q ",_ -:=;I c: ., A .-.- •.••,Ai· ' ....-,::.~. ...•.•.• .----- .----- ---- .---------------- -- --- . -._. " w , 267 �r;:; ......,_.;. ,'::'"'1 ,-::.-- .'>.. :£ C:! 1,.:;''1 <..:;...c c: I"i"'; ;;:.. o ._ .:= o :..£ ~ ~ :::!::: 268 .• .: ----..:.~- .-----------~------..-.-~ �269 JOB PROGRESS REPORT State of: Project: Colorado Avian Research W-152-R Work Plan: 14 Job Title: Seasonal Movement and Habitat Use by Greater Prairie-Chickens Period Covered: Author: Job 3 ----- 01 January through 31 December 1987 Michael A. Schroeder Personnel: M. A. Schroeder and G. C. White, Colorado State University; C. E. Braun, J. F. Corey, F. Pusateri, and L. A. Robb, Colorado Division of Wildlife ABSTRACT Investigations were continued to examine seasonal movements and habitat use of greater prairie-chickens (Tympanuchus cupido) in northeastern Colorado. Preliminary results suggest that during the breeding season female prairie-chickens move throughout home ranges that often encompass 2 or more leks. Females nested an average of 2.86 km from the lek at which they were captured. Females also tended to have larger home ranges than males during most seasons. Although most males seemed to have site fidelity to a particular lek, some moved between 2 or more leks.· Habitat-use was variable within each season, and differences between observed and available habitat sites were generally unclear. Behavior (such as loafing, nesting, feeding) appeared to be more important than either sex or season for interpreting habitat use. ��271 SEASONAL MOVEMENT AND HABITAT USE BY GREATER PRAIRIE-CHICKENS Michael A. Schroeder P. N. OBJECTIVES 1. Quantify seasonal habitat use, movements, and lek attendance of greater prairie-chickens in northeastern Colorado. SEGMENT OBJECTIVES 1. Trap and mark as many greater prairie-chickens as possible in an area containing 8 to 10 leks. 2. Radiomark up to 50 (32 female and 18 male) greater prairie-chickens. 3. Relocate all radio-marked greater prairie-chickens at least 20 times per season (40 times during the late spring season, 1 Apr-15 May). 4. Analyze movement of greater prairie-chickens in relation to cover types as mapped on the entire study area. 5. Describe vegetation at all greater prairie-chicken relocation sites. Analyze habitat use for each season and sex. 6. Document reproductive parameters, movements, and survival for all radio-marked greater prairie-chickens. 7. Analyze the dispersion of female greater prairie-chickens during the breeding and nesting seasons to test hypotheses about lek evolution and territoriality. 8. Compile and analyze data, and prepare progress report. METHODS An area centered 20 km northwest of Wray, Colorado was chosen for research on greater prairie-chickens. A study area of approximately 200 km2 was monitored to determine lek density and male lek attendance. Trapping efforts were concentrated on a smaller core area of 75 km2• Birds were trapped at winter feeding sites using walk-in traps (baited with corn) and on leks using walk-in traps and cannon nets. All captured birds were banded with a numbered aluminum band and a unique combination of 3 colored plastic bands. In addition, the lengths of primary I to X, the diameter of primary IX (at the base, immediately below the barbs), the length of the pinnae, and weight were measured. Age was ascertained by examining patterns of feather wear (Ammann 1944). Bird ages were: yearlings, 5 to 17 months of age (1 Nov of 1st year to 31 Oct of 2nd year) and adults, older than 17 months of age (after 31 Oct of 2nd year). Battery- and solar-powered radio transmitters were attached to poncho-type markers (Amstrup 1980) and placed on greater prairJe-chickens in both 1986 and 1987 (Table 1). Radio weights ranged between 1.8 and 2.3% of each bird's body weight. �272 Table 1. Number of greater prairie-chickens banded and fitted with radio transmitters, and their fates, in northeastern Colorado during 1986 ad 1987. Capture data RadiosD Mortalities Fates of radio-marked birdsa Removed~ Losttl Predation Alive Cate~ory Bands Males Adults Yearlings 106 60 46 32 19 13 4 2e 2f 6 4 2 4 2 2 13 6 7 9 7 2 Females Adults Yearlings 126 62 64 81 36 4S 3 2f If 14 7 7 17 3 14 29 11 18 21 IS 6 Totals 232 113 7 20 21 42 30 aFates as of 31 Dec 1987. bBirds given radio transmitters were also banded. CAt least 14 radios were recovered after they fell off birds. dBirds potentially were lost because of several reasons including 1) long untrackable movements by radio-marked birds; 2) broken radios on the live birds; and 3) broken radios on dead birds. eTwo birds in a single trap died of heat stress. fRaptors killed S birds in traps in a 3-day period. Seasonal collection periods were defined as winter (1 Nov-14 Feb), early spring (IS Feb-3l Mar), late spring (1 Apr-IS May), early summer (16 May-30 Jun), late summer (1 Jul-IS Aug), and fall (16 Aug-3l Oct); designation of seasons was based on aspects of breeding behavior and movement (Robel et al. 1970). Radio-marked prairie-chickens were tracked using a portable receiver and a 3-element yagi antenna. Each bird was visually observed once every 21-28 days and numerous additional observations were obtained by triangulation; azimuths were obtained within 1.0 km of the target transmitters and at angles-of-incidence greater than 3So and less than l4So• Initial estimates of accuracy suggested that locations derived by triangulation had a 9S% probability of being within 200 m of the actual location. All locations were recorded using Universal Transverse Mercator cordinates (nearest 10-m interval). Home range size was estimated as the area within a 7S% probability contour generated with harmonic means for each radio-tracked bird for each season (Dixon and Chapman 1980). Habitat was examined at both observed and 'available' sites. Available sites were chosen relative to observed sites (modification of a stratified sampling method) and were randomly selected within a O.S-km circle (O.S km was representative of a typical prairie-chicken flight distance) centered over the observed site. One available site was recorded for each observed site. To eliminate problems associated with measuring habitat variables in a changing environment (e.g., snow cover, plant growth, grazing pressure), both sites were examined on the same day. Two l8-m perpendicular transects were established in the center of the site, orientation of the initial transect was randomly determined. Ten point-intercept locations, 2 m apart, were located �273 along each transect (total of 20 points). All plant species intercepted at each point were identified and recorded. A height-density-index (HOI) was recorded from a height of 1 m and at a distance of 4 m to one side of the transect for all 20 points. The HDI was recorded as the height of vegetation obstructed on a Robel Pole (to the nearest 5 cm). Heights of sand sagebrush (Artemisia filifolia), grasses, and forbs were recorded to the nearest 5 cm. A single 25-m2 circle was also centered on each site (2.82 m diameter) and all plant species within the circle were identified and recorded. Location, slope, and aspect were also recorded at each site. Bradbury's (1981) female preference hypothesis will also be tested. The female preference hypothesis predicts that most females should have home ranges that incorporate only 1 lek, and that each female should visit only the lek(s) within her home range. Two null hypotheses will be tested with X2 tests: 1) the probability of a female visiting 2 or more leks during aparticular breeding seasn is 50% (50% is a conservative interpretation of 'most'); and 2) the probability of a femal~ nesting closer to a lek that is different than the lek where she was captured is equal to 50%. RESULTS Three different trapping methods were used to capture 232 greater prairie-chickens during 1986 and 1987 (Table 1). Walk-in traps were particularly effective for capturing birds on both winter feeding areas and on leks (Table 2). Males were larger than females for all feather measurements and weights (P <0.01). In addition, adults were generally larger than yearlings for-each sex (Table 3). The results also indicated that certain measurements may be particularly effective for differentiating between adults and yearlings. Among those, the lengths of primaries I, II, and IX and the diameter of primary IX appeared to be best for discrimination. Although body weights also differed by age and sex, seasonal variations in weights were detected (Table 4). For example, females weighed more during the breeding season than in winter (~< 0.05); males did not differ. Table 2. Numbers of greater prairie-chickens captured by different techniques in 1986-87 in northeastern Colorado (including recaptures). Cannon nets at lek sites Total 61 42 19 21 14 7 141 81 60 41 25 16 101 49 52 6 2 4 148 76 72 100 162 27 289 Category Walk-in traps at winter feeding sites Males Adults Yearlings 59 25 34 Females Adults Yearlings Totals Walk-in traps at lek sites �N '-J Table 3. Measurements of greater prairie-chickens captured in northeastern Colorado in 1986 and 1987. Fema1esa Ma1esa Adults x SD 11.69 12.02 12.63 13.96 16.40 17.12 17.28 17.10 16.22 13.06 0.39 0.35 0.41 0.48 0.57 0.60 0.47 0.83 0.57 0.96 32 31 30 32 31 33 31 43 43 44 Primary IX diameter (mm) 39 3.41 0.11 Pinnae length (cm) 49 0.47 Measurement N Primary length (cm) I 54 II 54 III 54 IV 52 V 51 VI 54 VII 54 VIII 63 IX 62 X 65 7.93 +-- Yearlings x Adults SD pb N. 11.21 11.75 12.43 13.65 16.23 17.09 17.31 16.83 15.69 13.25 0.34 0.28 0.31 0.44 0.50 0.60 0.45 0.97 0.40 0.95 *** *** * ** 57 57 57 57 56 57 55 68 68 68 40 3.21 0.10 *** 26 7.72 0.53 N *** X SD 11.34 11.71 12.29 13.45 16.66 16.37 16.50 16.18 15.39 12.55 0.44 0.64 0.72 0.64 0.57 0.54 0.61 1.11 0.70 0.90 55 55 55 55 54 53 53 63 63 64 N Year1in~s x SD pb 10.85 11.32 11.96 13.05 15.32 16.18 16.45 16.19 14.97 12.52 0.31 0.31 0.38 0.39 0.49 0.46 0.41 0.63 0.53 0.58 *** *** ** *** ** *** 47 3.27- 0.11 46 3.14 0.12 49 3.78 0.50 50 3.62 0.52 *** Weight (g) 78 1030.7 59.4 57 1005.6 42.2 ** 76 901.9 59.5 69 877.7 54.6 aThe differences between measurements for males and females within each age category were all significant at p < 0.01 (t test). - bThe probability for differences in measurements (!test) between adults and yearlings within each category of sex were: * = R. < 0.05; ** = R. < 0.01; and *** = p < 0.001. * �275 Table 4. Comparison of weights (!test) for greater prairie-chickens captured during the breeding season (15 Feb-3l May) and winter (1 Nov-14 Feb). N Winter x SD N Breeding season SD x P Males Adults Yearlings 26 32 1033.7 1007.8 73.2 42.6 52 25 1029.2 1002.8 51.9 42.5 0.784 0.661 Females Adults Yearlings 26 16 879.6 845.0 56.9 47.3 50 53 913.5 887.5 58.0 53.1 0.018 0.005 Density Thirty-eight active leks were documented on the study area during both 1986 and 1987 (Table 5). Twenty-one percent (8 of 38) of the leks that were active in 1986 were not active in 1987 (8 new leks were found in 1987). The density of leks on the study area was approximately 0.19 leks/km2 (0.49 leks/miles2) in both 1986 and 1987. The mean lek attendance was 6.69 males/lek in 1986 and 6.95 males/lek in 1987. Density of displaying males was approximately 1.27 males/km2 (3.25 males/mile2) in 1986 and 1.32 males/km2 (3.38 males/mile2) in 1987. Male attendance at leks appeared to be fairly constant from early March to early June, whereas female attendance (as represented by trapping success) showed a dramatic peak during the first 3 weeks of April (Table 6). Survival Radio-marked greater prairie-chickens were monitored as long as possible past the date of capture (Table 1). Forty-two birds died; examination of kill characteristics suggested that at least 22 were taken by coyotes (Canus latrans), 12 by avian predators (at least 1 Swainson's hawk [Buteo swainsoni], 1 ferruginous hawk [Buteo rega1is], and 1 great-horned owl [Bubo virginianus], and 8 by unknown predators. Other predators were also observed at fresh kills of greater prairie-chickens, including golden eagles (Aquila chrysaetos), red-tailed hawks (Buteo jamaicensis), and a northern harrier (Circus c aneus). Birds were also killed in collisions with vehicles and power lines and or fences (1 such kill was carried to a farm house by a domestic cat). Twenty-one radio-marked birds were 'lost' because of broken radios, or because the birds moved too far from the study area to be found. If broken radios were a problem, then 'lost' birds should have been distributed equally among the age and sex-categories (Table 1). However, yearling females 'disappeared' more frequently than adults and males (X2 = 8.071; P <0.05), suggesting that they may be more likely to make long movements away-from the study area. �276 Table 5. Location of active greater prairie-chicken leks on the study area in northeastern Colorado and the median number of males attending the 1ek during the breeding season (a plus sign symbolizes an active 1ek without an accurate count of males). Universal Transverse Mercators (Zone 12) Meters North Meters East 714880 715830 716370 716610 716670 717050 717100 717150 717250 718350 718130 718690 718860 719340 719630 719920 720060 720620 720980 721900 722100 722440 722850 723020 723300 723450 724110 724270 724420 724830 724960 725280 725870 726290 726410 726520 726700 727280 727450 728280 728630 728730 730390 731000 732420 733080 4453420 4448010 4453120 4456480 4453840 4455600 4457710 4448870 4454290 4449920 4450100 4453530 4457680 4451710 4453920 4449630 4448250 4455520 4458240 4455230 4453700 4453940 4456700 4456460 4458750 4455900 4453510 4459920 4449180 4448130 4447950 4450980 4445950 4447050 4449940 4455730 4450140 4444810 4454630 4447390 4448610 4445330 4446390 4447120 4445100 4446330 TownshiE 3N 2N 3N 3N 3N 3N 3N 2N 3N 2N 2N 3N 3N 3N 3N 2N 2N 3N 3N 3N 3N 3N 3N 3N 3N 3N 3N 3N 2N 2N 2N 3N 2N 2N 2N 3N 2N 2N 3N 2N 2N 2N 2N 2N 2N 2N Legal location Section Ran~e 46W 46W 46W 46W 46W 45W 45W 45W 45W 45W 45W 45W 45W 45W 45W 45W 45W 45W 45W 45W 45W 45W 45W 45W 45W 45W 45W 45W 45W 45W 45W 45W 45W 45W 45W 44w 44w 44W 44W 44W 44w 44W 44W 44W 44W 44w 25 12 25 13 24 18 7 7 19 6 6 29 8 32 28 5 8 16 9 22 27 22 15 15 2 14 26 2 2 11 12 36 13 13 1 18 6 19 19 17 8 20 16 16 22 14 Quarter NW NW NE NE SE SW SW NW SW NE NE NW SW NE SW SE NE SW NE NW NW SE NE NE SW SW NE NE SE SE SW SW SE NE SE SW NW SE SE NW NW NW S.W NE SE SW Attendance 1986 1987 + + 3 5 6 3 6 7 13 8 0 5 0 3 0 14 0 0 5 15 11 0 5 3 4 0 2 5 9 11 7 7 2 2 0 11 9 8 + 3 8 9 7 8 6 4 3 9 0 9 9 2 4 15 12 8 2 0 2 0 2 17 2 2 6 8 12 4 4 5 0 2 0 5 8 6 7 8 5 0 2 7 8 9 11 0 7 10 18 10 4 0 �277 Table 6. Trapping success rate (percent captures per trap per day) of greater prairie-chickens in northeastern Colorado. Trapping week 4-10 11-17 18-24 25-31 1- 7 8-14 15-21 22-28 Mar Mar Mar Mar Apr Apr Apr Apr Trap days Males Females Totals 62 59 167 348 386 274 223 91 3.2 1.7 3.0 0.9 5.7 7.3 4.0 4.4 0.0 0.0 1.2 2.6 8.8 13.5 9.0 1.1 3.2 1.7 4.2 3.4 14.5 20.8 13.0 5.5 Reproductive Success In 1986, 20 nests had a mean date of hatch of 10 June (som~hatch dates were predicted based on the known onset of incubation). First nests had a mean hatch date of 1 June (N = 10), while probable, or known ~enesters (based on the presence of a brood patch on a female visiting a lek or the known failure of a female's first nest), had a mean hatch date of 19 June (!= 10). In 1987, the mean date of hatch was 4 June. First nests had a mean hatch date of 4 June (N = 23) and renests had a mean hatch date of 15 June (N = 2). Six nests (4-renests and 2 first nests) successfully hatched an av;rage of 4.8 chicks (33.3% hatching success) in 1986 and 11 nests (all first nests) successfully hatched an average of 6.7 chicks (40.7% hatching success) in 1987. Movement Home range size of greater prairie-chickens was estimated by sex and age class for each season (Table 7). Yearling males tended to have larger home ranges than adult males for all seasons. Anecdotal observations of lek visitation by males indicated that most males were observed attending the same lek where they were captured. Of 5 exceptions, 4 were yearlings. One yearling male visited at least 5 different leks in 1 breeding season. When age categories were combined within each sex and the sexes were compared with t tests, males were found to have smaller home ranges than females during the late spring (P = 0.003), early summer (p = 0.006), and fall (p =0.026). Results for the other seasons were not significant (p >0.05), however, the trends were all in the same direction. Comparisons of male home range size by season showed that male home range size was relatively stable throughout the year. Conversely, females had larger (p <0.05) home ranges during the early summer and small home ranges during the-late summer and early spring. The early summer increase in female movement was particularly noticeable since females made movements up to 48 km following predation loss of their nests. Behavior also appeared to affect the size of female home ranges during late summer and fall. Females with broods tended to have larger home ranges than females without (Table 8). �278 Table 7. Home range size (area within 75% probability contours generated with harmonic means, ha) of greater prairie-chickens in northeastern Colorado during 1986 and 1987 (years combined). Category Adults Males Yearlings Totals Adults Females Yearlings Totals Totals Early spring N Median x \. SD 5 38.8 45.1 44.4 1 133.4 133.4 6 39.3 59.8 53.6 3 256.2 308.4 175.4 2 90.1 90.1 89.7 5 165.0 221.1 178.0 11 120.9 133.1 145.6 14 79.4 92.0 35.4 5 158.7 394.0 597.8 19 92.9 171.4 213.5 25 315.9 370.1 319.4 31 252.5 453.5 520.2 56 260.5 416.3 430.6 75 212.8 354.3 375.6 9 144.0 172.3 99.3 4 357.2 435.5 247.9 13 198.3 253.3 167.9 16 240.3 1142.5 1418.7 16 167.2 519.5 990.0 32 196.4 831.0 1056.8 45 197.2 664.1 774.1 7 108.1 134.3 121.1 4 195.1 179.8 89.5 11 126.9 150.8 108.3 14 137.0 180.6 147.5 15 121.4 168.6 120.9 29 126.9 174.4 132.1 40 125.9 167.9 125.1 4 225.7 189.0 96.4 4 213.2 214.7 90.7 8 225.7 201.9 87.7 7 406.8 575.3 397.0 10 272.8 291.9 227.5 17 397.0 408.6 330.0 25 289.6 432.5 290.7 9 208.2 226.4 105.6 4 260.1 242.5 126.9 13 221.1 231.4 107.3 16 244.1 448.5 538.3 7 236.4 261.5 144.8 23 239.4 391.5 459.4 36 237.9 333.7 377 .8 Late spring N Median x SD Early summer N Median ~ SD Late' summer N Median x SD Fall N Median x SD Winter N Median K SD �27~ Table 8. Home range size (area within 75% probability contours generated with harmonic means, ha) of female greater prairie-chickens with and without broods in northeastern Colorado during the brood season (late summer and fall) of 1986-87 (years combined). Category Late summer N Median x SD Fall N Median x SD With broods Without broods Totals 12 229.9 225.3 152.5 17 120.0 138.5 105.9 29 126.9 174.4 132.1 6 442.3 592.0 391.6 11 284.7 308.6 257.6 17 397.0 408.6 330.0 Habitat Habitat variables were compared between the observed locations of radio-marked greater prairie-chickens and available sites for winter (Table 9), early spring (Table 10), late spring (Table 11), early summer (Table 12), late summer (Table 13), and fall (Table 14). No differences were found for any of the seasons, except late spring. One possible explanation for the lack of habitat differences was the large amount of variability in habitat use; greater prairie-chickens were frequently feeding in fields of corn stubble and loafing and/or roosting in residual grass and sand sagebrush cover. In late spring, birds were generally observed in areas with shorter cover (grasses and forbs), sparser grass cover, and more corn and bare ground (Table 11). As well, greater prairie-chickens often fed in fields of corn stubble during late spring and males were frequently on booming grounds (characterized by short and sparse vegetation). Observed and available habitat (all seasons combined) was examined for males (Table 15) and females (Table 16). Males were observed in areas with less corn cover than expected. Females were observed in areas with shorter grasses and forbs and with denser sand sagebrush. Examination of 35 leks indicated a considerable difference between observed and available habitat (Table 17). Leks were usually on hill or ridge-crests (relatively flat on top) with sparse, short cover. Nesting habitat also appeared to be relatively distinct (Table 18). Generally, nests were in areas with thicker and taller vegetation. Loafing (and/or roosting) habitat was similar to nesting habitat in that sand sagebrush cover was thicker and taller at observed sites than it was at available sites (Table 19). In addition, forb cover was greater and species richness was higher. Feeding sites covered a broad spectrum of habitat types, and hence, few differences between available and observed sites were detected (Table 20). Most qotable was the choice of areas with short grass cover. Sites of brood locations were characterized by greater warm season grass and forb cover (Table 21). �280 Table 9. Habitat at observed and available locations of greater prairie-chickens in northeastern Colorado during winter (1 Nov-14 Feb) 1986-87 and 1987-88. P-va1uea t Fma;l:l; Observed (26) x SD Available (26) SD )C 8.15 4.23 10.08 6.93 0.017 0.233 Height (cm) Sand sagebrushb Grasses Forbs Barec 35.38 77 .69 40.77 20.96 37.39 33.05 37.94 20.40 42.12 88.46 54.23 18.46 35.13 37.52 38.72 11.02 0.758 0.531 0.920 0.003 0.507 0.277 0.211 0.586 Cover (%) Bare Sand sagebrush Grasses Cool season grasses Need1e-and-threadd Sex weeks fescuee Warm season grasses Blue gramag Sand dropseedh Prairie sandreedi Corn Forbs 54.42 6.54 38.85 4.04 4.04 0.00 35.77 6.35 5.58 7.31 11.73 5.00 29.74 8.34 26.88 7.75 7.75 0.00 25.21 14.32 8.04 13.21 15.81 9.38 59.04 3.46 35.38 6.15 5.38 0.19 30.58 4.42 6.73 4.42 5.77 4.23 29.93 4.64 28.91 13.36 13.34 0.98 26.01 12.36 8.48 10.33 13.17 7.03 0.975 0.005 0.719 0.008 0.009 __f 0.876 0.466 0.793 0.226 0.367 0.156 0.580 0.108 0.657 0.489 0.659 0.327 0.468 0.607 0.617 0.385 0.146 0.739 Species richness (N) 8.19 6.40 9.38 5.84 0.650 0.486 Slope (degrees) 2.58 2.04 3.23 5.24 0.000 0.558 Habitat variable Height-density-index (cm) arf !max was significant, a! test for unequal variances was used. bSand sagebrush = Artemisia filifolia. CHeight of bare ground includes soil mounds, man-made objects, and snow. dNeed1e-and-thread = Stipa comata. eSix weeks fescue = Vu1pia octof10ra. fAt least one of the variances was undefined; hence the variances were assumed to be different. gB1ue grama = Boute10ua gracilis. ~Sand dropseed = Sporobo1us cryptandrus. 1Prairie sandreed = Ca1amovilfa longifo1ia. �281 Table 10. Habitat at observed and available locations of greater prairie-chickens in northeastern Colorado during early spring (15 Feb-31 Mar) 1987. Observed (5) x SD Habitat variable Available (5) x SD P-va1uea Lnax t 5.38 4.59 2.04 0.56 0.001 0.180 60.00 100.00 55.00 6.00 21.79 23.45 11.73 2.24 31.00 85.00 47.00 9.00 23.29 25.50 38.01 5.48 0.901 0.875 0.043 0.111 0.077 0.361 0.673 0.290 Cover (%) Bare Sand sagebrush Grasses Cool season grasses Need1e-and-thread Six weeks fescue Warm season grasses Blue grama Sand dropseed Prairie sandreed Corn Forbs 15.00 13.00 79.00 28.00 28.00 0.00 62.00 19.00 19.00 26.00 0.00 1.00 10.00 9.08 12.94 19.24 19.24 0.00 13.51 7.42 6.52 17.10 0.00 2.24 28.00 4.00 64.00 10.00 8.00 1.00 55.00 10.00 10.00 15.00 11.00 7.00 12.04 4.18 15.97 10.00 9.08 2.24 20.00 7.91 16.96 11.18 24.60 6.71 0.728 0.162 0.694 0.233 0.175 __c 0.056 0.100 0.079 0.141 0.101 0.069 0.374 0.535 0.101 0.300 0.263 0.374 0.094 Species richness (N) 12.80 3.27 12.00 6.81 0.178 0.821 3.20 1.64 2.40 1.14 0.496 0.397 Height-density-index (cm) Height (cm) Sand sagebrush Grasses Forbs Bareb Slope (degrees) 0.466 0.904 0.091 0.431 __ c alf !max was significant, a ! test for unequal variances was used. bHeight of bare ground includes soil mounds, man-made objects, and snow. CAt least one of the variances was undefined; hence the variances were assumed to be different. �282 Table 11. Habitat at observed and available locations of greater prairie-chickens in northeastern Colorado during late spring (1 Apr-15 May) 1987. Observed (38) SD x Habitat variable Available (38) SD x P-va1uea £nax t 3.12 2.43 3.26 2.53 0.805 0.800 Height (cm) Sand sagebrush Grasses Forbs Bareb 51.05 73.03 44.34 8.82 43.00 34.46 42.08 4.10 66.71 98.16 71.05 6.45 30.43 20.01 35.42 2.58 0.039 0.001 0.299 0.006 0.071 0.000 0.004 0.004 Cover (%) Bare Sand sagebrush Grasses Cool season grasses Need1e-and-thread Six weeks fescue Warm season grasses Blue grama Sand dropseed Prairie sandreed Corn Forbs 28.95 8.42 60.92 24.08 16.18 5.13 41.58 15.13 6.58 6.32 4.08 9.21 22.37 11.03 21.21 18.56 16.08 8.26 19.11 17.38 6.89 7.68 9.85 17.18 19.34 6.05 71.05 30.79 21.71 9.47 46.58 12.37 14.34 9.74 0.13 9.08 11.28 8.63 17.71 20.62 20.58 11.50 15.77 14.92 15.99 10.65 0.81 9.36 0.000 0.140 0.278 0.526 0.139 0.048 0.248 0.358 0.000 0.050 0.000 0.000 0.022 0.301 0.027 0.140 0.196 0.063 0.217 0.459 0.008 0.113 0.019 0.967 Species richness (N) 12.89 8.07 16.18 5.84 0.054 0.045 1.95 1.66 3.00 3.08 0.003 0.069 Height-density-index Slope (degrees) (cm) alf !max was significant, a ! test for unequal variances was used. bHeight of bare ground includes soil mounds, man-made objects, and snow. �283 Table 12. Habitat at observed and available locations of greater prairie-chickens in northeastern Colorado during early summer (16 May-30 Jun) 1987. Observed (33) Habitat variable x SD Available (33) x SD P-va1uea max t 3.46 2.77 3.38 1.92 0.041 0.894 Height (cm) Sand sagebrush Grasses Forbs Bareb 67.88 99.09 66.52 6.82 33.50 27.37 29.65 2.74 61.97 97.58 78.03 6.67 37.25 38.10 44.19 3.23 0.552 0.066 0.027 0.363 0.500 0.853 0.219 0.838 Cover (%) Bare Sand sagebrush Grasses Cool season grasses Need1e-and-thread Sex weeks fescue Warm season grasses Blue grama Sand dropseed Prairie sandreed Corn Forbs 18.64 6.06 71.36 32.12 16.21 10.76 48.64 14.24 8.48 8.03 1.06 13.03 15.53 7.04 16.92 20.95 15.46 10.01 16.02 14.31 9.72 9.60 4.29 12.74 19.09 6.21 70.00 33.79 22.12 8.48 46.97 13.33 7.42 12.58 1.67 10.91 20.78 7.61 24.65 26.34 20.69 12.90 20.08 16.04 9.36 14.37 5.40 11.95 0.104 0.667 0.037 0.201 0.104 0.157 0.207 0.523 0.834 0.025 0.197 0.720 0.920 0.933 0.794 0.777 0.194 0.427 0.711 0.809 0.653 0.136 0.615 0.488 Species richness (N) 21.06 7.57 19.15 7.95 0.786 0.322 1.79 1.47 2.27 2.14 0.039 0.288 Height-density-index Slope (degrees) (cm) alf !max was significant, a t test for unequal variances was used. bHeight of bare ground includes soil mounds, man-made objects, and snow. �284 Habitat at observed and available locations of greater Table 13. prairie-chickens in northeastern Colorado during late summer (1 Jul-15 Aug) 1986-87. Observed (69) SD x Habitat variable Available (69) SD x P-va1uea t lmax 4.78 3.60 3.88 2.73 0.024 0.100 61.67 102.83 68.26 4~35 32.10 22.86 30.96 3.92 50.94 100.65 64.64 5.14 32.22 23.07 31.72 10.98 0.975 0.939 0.843 0.000 0.052 0.579 0.498 0.572 Cover (%) Bare Sand sagebrush Grasses Cool season grasses Need1e-and-thread Six weeks fescue Warm season grasses Blue grama Sand dropseed Praire sandreed Corn Forbs 14.35 10.29 73.33 30.00 22.10 8.04 52.25 12.61 10.29 13.12 1.16 13.70 8.66 10.00 15.16 17.88 16.50 9.97 15.11 10.10 10.50 10.33 3.94 11.87 17.03 8.62 72.75 32.90 24.06 7.32 48.91 16.30 9.71 11.59 0.43 10.87 11.70 10.21 17.25 18.20 17.20 8.32 16.93 15.09 10.77 10.13 1.66 10.88 0.134 0.859 0.291 0.886 0.735 0.155 0.351 0.001 0.831 0.871 0.000 0.474 0.129 0.334 0.834 0.347 0.497 0.645 0.225 0.094 0.749 0.384 0.163 0.147 Species richness (N) 20.29 5.65 18.96 5.38 0.686 0.158 2.26 2.20 2.57 2.65 0.129 0.464 Height-density-index (cm) Height (cm) Sand sagebrush Grasses Forbs Bareb Slope (degrees) alf !max was significant, a ! test for unequal variances was used. bHeight of bare ground includes soil mounds, man-made objects, and snow. �285 Table 14. Habitat at observed and available locations of greater prairie-chickens in northeastern Colorado during fall (16 Aug-31 Oct) 1986-87. Observed (51) x Habitat variable SD Available (51) x SD P-va1uea Lnax t 6.74 12.00 12.17 21.68 0.000 0.122 Height (cm) Sand sagebrush Grasses Forbs Bareb 45.88 93.53 54.80 5.98 31.32 35.86 33.36 4.24 33.14 111.18 60.69 7.35 31.80 49.93 34.35 4.40 0.915 0.021 0.837 0.802 0.044 0.043 0.382 0.112 Cover (%) Bare Sand sagebrush Grasses Cool season grasses Need1e-and-thread Six weeks fescue Warm season grasses Blue grama Sand dropseed Prairie sandreed Corn Forbs 17.16 8.43 72.25 21.86 11.96 4.71 58.92 12.25 14.02 10.59 6.67 13.14 13.39 10.22 18.01 18.97 14.77 7.17 15.92 15.31 14.70 13.33 19.38 15.30 19.90 4.71 72.25 15.69 12.75 2.65 62.55 15.39 12.45 9.12 10.98 11.08 12.02 6.59 15.98 16~-76 16.01 4.93 17.56 18.86 14.26 13.85 23.47 11.85 0.449 0.002 0.401 0.384 0.570 0.009 0.489 0.143 0.832 0.789 0.179 0.074 0.279 0.031 1.000 0.085 0.798 0.094 0.277 0.359 0.586 0.586 0.314 0.449 Species richness (N) 16.47 6.72 14.98 7.36 0.522 0.288 2.51 3.26 2.49 2.96 0.503 0.975 Height-density-index Slope, degrees (cm) arf !max was significant, a t test for unequal variances was used. bHeight of bare ground includes soil mounds, man-made objects, and snow. �286 Table 15. Habitat at observed and available locations for male greater prairie-chickens in northeastern Colorado during 1986-87. Observed (92) x Habitat variable SD Available (92) x SD P-valuea .F-max t 5.82 9.68 8.40 16.61 0.000 0.199 Height (cm) Sand sagebrush Grasses Forbs Bareb 50.71 92.12 53.70 7.77 35.70 36.95 38.18 7.12 53.21 102.66 60.16 8.53 35.82 42.53 38.88 7.43 0.974 0.182 0.862 0.687 0.636 0.074 0.257 0.479 Cover (%) Bare Sand sagebrush Grasses Cool season grasses Needle-and-thread Sex weeks fescue Warm season grasses Blue grama Sand dropseed Prairie sandreed Corn Forbs 24.08 7.50 67.55 23.37 15.98 5.05 50.98 15.27 8.75 10.11 3.91 8.53 20.86 8.60 20.96 18.78 15.54 7.90 17.78 14.94 11.06 11.24 10.94 10.81 25.65 6.09 66.36 23.04 18.10 4.46 49.67 13.53 8.48 10.38 5.92 8.97 22.63 8.15 23.17 21.90 18.94 9.12 20.83 15.54 9.80 12.00 16.98 11.30 0.438 0.610 0.340 0.145 0.060 0.171 0.132 0.709 0.250 0.533 0.000 0.669 0.624 0.254 0.714 0.914 0.408 0.635 0.648 0.440 0.860 0.874 0.034 0.790 Species richness (N) 15.18 7.72 14.33 7.31 0.610 0.440 1.86 2.01 2.41 2.39 0.102 0.090 Height-density-index Slope (degrees) (cm) alf !max was significant, a ! test for unequal variances was used. bHeight of bare ground includes soil mounds, man-made objects, and snow. �287 Table 16. Habitat at observed and available locations for female greater prairie-chickens in northeastern Colorado during 1986-87. Observed (130) SD x Habitat variable P-va1uea Kma~ t 4.70 2.93 4.80 4.83 0.000 0.837 Height (cm) Sand sagebrush Grasses Forbs Bareb 56.38 91.96 59.85 7.88 35.86 28.15 33.09 10.49 47.23 98.81 68.77 7.19 33.81 28.64 34.17 9.38 0.505 0.846 0.716 0.206 0.035 0.053 0.033 0.575 Cover (%) Bare Sand sagebrush Grasses Cool season grasses Need1e-and-thread Six weeks fescue Warm season grasses Blue grama Sand dropseed Prairie sand reed Corn Forbs 21.96 9.27 66.19 25.04 15.85 6.77 48.81 10.73 10.69 10.31 4.27 13.42 21.11 10.38 22.25 20.19 16.59 9.44 19.81 13.02 11.06 11.84 13.12 15.28 22.08 6.31 68.08 26.50 18.31 6.77 49.12 13.38 11.85 9.88 2.38 10.31 19.40 8.43 22.89 21.-20 18.48 9.13 20.60 16.43 14.03 12.06 10.02 10.28 0.321 0.019 0.748 0.579 0.221 0.702 0.660 0.009 0.007 0.833 0.002 0.000 0.964 0.012 0.501 0.570 0.260 1.000 0.902 0.150 0.462 0.776 0.194 0.055 Species richness (N) 17.73 7.87 17.73 6.64 0.054 1.000 2.53 2.45 2.82 3.52 0.000 0.438 Height-density-index .Slope (degrees) (cm) Available (130) SD x alf !max was significant, a ! test for unequal variances was used. bHeight of bare ground includes soil mounds, man-made objects, and snow. �288 Table 17. Habitat at observed and available 1ek locations of greater prairie-chickens in northeastern Colorado during 1986-87. Observed (35) Habitat variable Height-density-index Y x SD l:-va1uea ¥max t 1.12 0.76 3.63 2.57 0.000 0.000 Height (cm) Sand sagebrush Grasses Forbs Bareb 32.29 60.00 39.43 6.86 16.99 28.18 38.40 2.99 52.71 95.00 50.57 7.29 32.21 31.67 35.70 2.80 0.000 0.500 0.673 0.706 0.002 0.000 0.213 0.538 Cover (%) Bare Sand sagebrush Grasses Cool season grasses Need1e-and-thread Six weeks fescue Warm season grasses Blue grama Sand dropseed Prairie sandreed Corn Forbs 28.00 3.57 68.86 26.57 17.00 8.29 47.43 25.43 5.71 8.43 0.00 2.43 9.64 4.13 10.51 13.33 12.44 11.56 15.16 18.29 7.49 7.93 0.00 6.23 19.14 4.86 75.29 32.00 25.14 8.71 51.00 14.14 17.57 11.14 0.57 6.43 14.53 7.12 14.09 21.67 19.83 13.47 16.66 14.01 15.55 12.31 3.38 7.33 0.019 0.002 0.092 0.006 0.008 0.379 0.586 0.125 0.000 0.012 0.345 0.004 0.359 0.034 0.212 0.044 0.887 0.352 0.005 0.000 0.277 0.324 0.017 Species richness (N) 11.69 5.95 13.97 4.99 0.308 0.086 0.54 0.56 2.40 2.34 0.000 0.000 Slope (degrees) (cm) SD Available (35) __c alf !max was significant, a ! test for unequal variances was used. bHeight of bare ground includes soil mounds, man-made objects, and snow. CAt least one of the variances was undefined; hence the variances were assumed to be different. �289 Table 18. Habitat at observed and available nest site locations for greater prairie-chickens in northeastern Colorado during 1986-87. Observed (53) x SD Habitat variable Available (53) x SD P-va1uea 1max .t. 5.16 2.12 2.79 2.00 0.689 0.000 74.34 108.77 67.08 3.49 30.60 19.34 34.38 3.19 59.91 98.11 61.60 5.47 32.59 26.77 34.13 15.07 0.652 0.021 0.959 0.000 0.021 0.021 0.413 0.353 Cover (%) Bare Sand sagebrush Grasses Cool season grasses Need1e-and-thread Six weeks fescue Warm season grasses Blue grama Sand dropseed Prairie sandreed Corn Forbs 11.79 11.98 74.72 38.02 28.21 10.75 43.77 10.85 11.79 11.51 0.28 13.96 7.08 8.28 15.20 21.54 19.59 12.95 17.97 11.21 11.14 9.18 1.52 11.28 21.32 6.89 67.92 31.23 21.32 10.85 43.30 10.94 8.02 10.00 0.28 11.13 18.22 7.29 20.20 19.71 18.99 12.08 18.00 11.69 10.89 11.05 1.52 10.27 0.000 0.359 0.043 0.525 0.824 0.618 0.991 0.766 0.870 0.184 1.000 0.500 0.001 0.001 0.053 0.093 0.069 0.969 0.893 0.966 0.081 0.446 1.000 0.180 Species richness (N) 21.13 5.86 19.98 6.06 0.811 0.323 1.85 1.73 2.94 3.40 0.000 0.040 Height-density-index (cm) Height (cm) Sand sagebrush Grasses Forbs Bareb Slope (degrees) alf !max was significant, a t test for unequal variances was used. bHeight of bare ground includes soil mounds, man-made objects, and snow. �290 Table 19. Habitat at observed and available loafing (and/or roosting) sites for greater prairie-chickens in northeastern Colorado during 1986-87. Observed (75) Habitat variable Height-density-index x (cm) SD Available (75) x SD P-va1uea Lnax t 5.94 7.53 5.93 11.14 0.001 0.994 64.60 103.80 65.13 6.53 33.35 28.19 33.35 5.00 47.20 99.67 68.40 7.00 34.78 36.72 37.75 5.58 0.718 0.024 0.288 0.350 0.002 0.441 0.575 0.590 Cover (%) Bare Sand sagebrush Grasses Cool season grasses Need1e-and-thread Six weeks fescue Warm season grasses Blue grama Sand dropseed Prairie sandreed Corn Forbs 19.67 10.80 67.53 25.73 15.60 6.20 49.73 11.80 11.13 10.47 2.27 13.00 18.42 10.37 20.49 18.32 15.77 8.09 19.10 12.72 10.70 11.28 9.60 13.00 22.47 5.40 68.00 25.60 19.80 4.93 49.13 10.73 9.87 10.60 3.73 9.07 19.55 7.06 22.63 22.60 19.84 7.60 20.11 14.13 10.23 11.91 12.97 10.80 0.612 0.001 0.395 0.073 0.050 0.593 0.659 0.368 0.700 0.639 0.010 0.114 0.368 0.000 0.895 0.968 0.153 0.325 0.852 0.628 0.460 0.944 0.433 0.046 Species richness (~) 18.81 6.86 15.63 7.26 0.630 0.006 2.83 2.91 2.88 3.76 0.028 0.923 Height (cm) Sand sagebrush Grasses Forbs Bareb Slope (degrees) alf !max was significant, a ! test for unequal variances was used. bHeight of bare ground includes soil mounds, man-made objects, and snow. �291 Table 20. Habitat at observed and available feeding sites for greater prairie-chickens in northeastern Colorado during 1986-87. Observed (112) Habitat variable Height-density-index x (cm) SD Available (112) X SD Lnax t 5.49 6.71 7.36 13.10 0.000 0.181 Height (cm) Sand sagebrush Grasses Forbs Bareb 49.64 87.77 56.25 9.11 38.01 33.02 36.14 12.08 47.59 100.71 64.87 8.57 35.07 36.77 36.03 11.08 0.397 0.259 0.974 0.361 0.675 0.006 0.075 0.730 Cover (%) Bare Sand sagebrush Grasses Cool season grasses Need1e-and-thread Six weeks fescue Warm season grasses Blue grama Sand dropseed Prairie sand reed Corn Forbs 25.49 7.90 64.11 22.72 14.87 5.49 48.17 8.97 8.97 9.73 6.34 12.10 23.91 9.53 23.77 21.50 16.61 8.85 20.19 11.55 11.45 12.34 14.78 15.06 25.67 6.29 65.36 22.46 15.22 5.22 40.24 13.17 11.43 9.20 4.11 10.76 23.33 8.10 24.92 19;85 17.01 8.42 21.77 15.51 14.66 11.52 14.37 10.83 0.798 0.089 0.619 0.403 0.802 0.599 0.429 0.002 0.010 0.471 0.765 0.001 0.955 0.175 0.701 0.923 0.874 0.817 0.703 0.023 0.164 0.737 0.253 0.446 Species richness (N) 16.13 8.71 16.79 7.16 0.040 0.536 2.29 1.91 2.54 2.79 0.000 0.451 Slope (degrees) alf !max was significant, a t test for unequal variances was used. bHeight of bare ground includes soil mounds, man-made objects, and snow. �292 Table 21. Habitat at observed and available brood locations for greater prairie-chickens in northeastern Colorado during 1986-87. Observed (37) SD x Habitat variable Height-density-index (cm) Available (37) SD x P-va1uea E.roa:Q; t 4.82 2.84 4.55 3.03 0.699 0.703 67.16 104.19 74.32 5.41 27.25 23.02 25.09 3.98 57.03 110.68 73.92 4.46 31.15 18.42 35.96 3.49 0.426 0.185 0.048 0.441 0.141 0.185 0.955 0.281 Cover (%) Bare Sand sagebrush Grasses Cool season grasses Need1e-and-thread Six weeks fescue Warm season grasses Blue grama Sand dropseed Prairie sandreed Corn Forbs 13.78 9.46 74.86 29.59 19.19 7.43 55.41 11.89 13.92 11.08 1.08 17.03 7.76 9.26 13.56 19.66 16.39 9.25 15.29 12.93 12.26 9.36 4.43 12.61 18.65 10.81 68.65 31.49 22.43 6.76 45.14 12.30 12.84 7.70 0.27 11.49 10.71 11. 70 16.31 18.33 15.62 8.84 17.26 16.36 13.31 8.13 1.15 9.12 0.057 0.166 0.273 0.675 0.733 0.785 0.471 0.163 0.623 0.401 0.000 0.056 0.028 0.583 0.079 0.670 0.387 0.749 0.008 0.906 0.717 0.102 0.288 0.034 Species richness (N) 22.54 5.45 21.38 5.40 0.950 0.360 2.59 2.02 3.19 3.65 0.001 0.389 Height (cm) Sand sagebrush Grasses Forbs Bareb Slope (degrees) alf !max was significant, a! test for unequal variances was used. bHeight of bare ground includes soil mounds, man-made objects, and snow. �293 Female Preference Hypothesis Overall, females moved a mean distance of 2.86 km (1.78 miles) between leks where they were captured and their nests (Table 22). Forty-one of 52 (78.8%) females nested closer to a different lek than the one at which they were captured (Fig. 1). When compared with the null hypothesis value of 50% (as predicted by Bradbury 1981), the actual result was higher (X2 = 9.439; P 0.003). Table 22. Distances (km) between nests and leks where females were captured and the lek nearest to their nest in 1986 (N = 21) and 1987 (N = 31) in northeastern Colorado. Category Capture lek Median x- SD Nearest lek Median X SD 1986 1987 Combined 2.09 2.65 2.11 1.77 3.01 4.33 1.91 2.86 3.43 0.83 0.91 0.55 0.96 1.07 0.72 0.92 1.01 0.65 Lek visitation by females was greater than predicted by Bradbury's (1981) female preference hypothesis (Table 23). Thirty-two of 38 (84.2%) females that visited at least 1 lek after being captured, visited more than 1 lek. When compared with the null hypothesis value of 50%, the actual result was higher (X2 = 10.074; P = 0.001). Table 23. Minimum number of leks visited during a breeding season by radio-marked female greater prairie-chickens in northeastern Colorado in 1986-87. N leks visited N % 1 2 3 4 26 18 11 3 44.8 31.0 19.0 5.1 All females Females observed on at least 1 lek after being captured % N 6 18 11 3 15.8 47.4 29.0 7.9 �294 1986 o o o o N o o Lek * o o o Nest o 0 o o 1987 0 0 0 0 0 0 0 00 0 0 0°1** 0 0 0 0 0 0 0 0 0 5km Fig. 1. Greater prairie-chicken nests in relation to respective leks of capture in northeastern Colorado in 1986 (~= 21) and 1987 (~= 31). �295 DISCUSSION Few data have been gathered on mortality, reproduction, lek attendance, movement, and habitat use by greater prairie-chickens in Colorado. Some sample sizes in this study are still too small to infer clear conclusions, however certain trends are apparent. Males apparently have site fidelity to a particular lek and can be found on the same lek every day. However, little is known about the proportion of males remaining faithful to a particular lek and those that move between leks. Reproductive success improved from 1986 to 1987. However, certain difficulties exist with interpretation of these data as the biases associated with trapping are not clearly understood. Trapping intensity was greater during the latter half of spring 1986, and consequently numerous females were caught that had already failed with their first nesting attempts. The bias toward late trapping success might partially explain why the estimated hatching success rate was only 30% (previous research suggests that 2nd nests are less successful than first nests, Robel and Ballard 1974). Likewise, late trapping success in 1986 may also explain the relatively large number of renests. Although numerous hypotheses have been proposed to explain the evolution of lek behavior from a dispersed mating system, most of these hypotheses have been rejected (Davies 1978, Wittenberger 1978, Bradbury 1981, Bradbury and Gibson 1984, Payne 1984). Two hypotheses are still debated: 1) female preference for clustered males (Bradbury 1981); and settlement of males on 'hotspots' of female traffic (Bradbury et al. 1986). Both predict a positive correlation between female home range size and spacing of male clusters (leks). The first hypothesis ('female preference' hypothesis), which suggests that each lek should have its own population of females, is not supported by the preliminary findings of this study which suggests that most· females visit more than 1 lek. Likewise, long female movements to nest sites indicate that most females are not restricted to the area around a single lek. The 2nd hypothesis ('hotspot' hypothesis) will be tested later in this study. Movement of females in relation to leks is also a significant aspect of greater prairie-chicken management. The fact that females nested an average of 2.86 km from the lek at which they were captured, suggests that nest locations may be more important than leks as indications of local habitat quality. Examination of habitat has been limited to a stratified sampling method that limits examination ·of available habitat to areas relatively close to the 'observed' habitat. Several differences were detected between observed and available habitat, especially for nesting, lekking, feeding, loafing, and brooding sites. Behavior appeared to be more important than sex and season for explaining differences in habitat use. In general, the habitat differences associated with sex and season were trivial. One possible reason for the lack of significant differences may be that greater prairie-chickens use a wide variety of habitats that are dependent on their behavior. �296 LITERATURE CITED Ammann, G. A. 1944. Determining the age of pinnated and sharp-tailed grouses. J. Wildl. Manage. 8:170-171. Amstrup, S. C. 214-217. 1980. A radio-collar for game birds. J. Wildl. Manage. 44: Bradbury, J. W. 1981. The evolution of leks. Pp. 138-169 in R. D. Alexander and D. W. Tinkle, ed., Natural selection and social behavior: recent research and new theory. Chiron Press, New York, N.Y. , and R. M. Gibson. 1984. --- Bateson, ed., Mate choice. Leks and mate choice. , and I. M. Tsai. 1986. ---Behav. Anim. 34:1694-1709. Pp. 109-138 in P. Hotspots and the disperson of leks. Davies, N. B. 1978. Ecological questions about territorial behavior. 317-350 in J. R. Krebs and N. B. Davies, ed., Behavioral ecology: evolutionary approach. Sinauer Associates, Sunderland, Mass. Pp. an Dixon, K. R., and J. A. Chapman. 1980. Harmonic mean measure of animal activity areas. Ecology 61:1040-1044. Payne, R. B. 1984. Sexual selection, lek and arena behavior, and sexual size dimorphism in birds. Ornith. Monogr. 33. 52pp. Robel, R. J., and W. B. Ballard, Jr. 1974. Lek social organization and reproductive success in the greater prairie chicken. Am. 2001. 14:121-128. ____ , J. N. Briggs, J. J. Cebula, N. J. Si1vy, C. E. Viers, and P. G. Watt. 1970. Greater prairie chicken ranges, movements, and habitat usage in Kansas. J. Wi1d1. Manage. 34:286-306. Wittenberger, J. F. 80:126-137. 1978. The evolution of mating systems in grouse. Prepared by .,.....,..,../.t...:,-:.7ldJ~..:......:..;...--lo~'---'.:...._:Jh~LI.Oc.::.~.!::..::>J2"-.c _ Michael A. Schroeder Graduate Research Assistant Approved by -=-,a~~_~-:, =-·--::-"-?_'---l..8..<..;7."",,~~ Clait E. Braun Wildlife Research Leader _ Condor �297 JOB PROGRESS REPORT State of: Project: Colorado W-152-R Avian Research Work Plan: 17 Job Title: Population Period Covered: Authors: Personnel: Job 7 --- Dynamics of White-tailed 01 January through 31 December Ptarmigan 1987 Clait E. Braun and Kenneth M. Giesen Shannon Lord and Kathy Martin, University of Alberta; Clait E. Braun and Kenneth M. Giesen, Colorado Division of \vildlife ABSTRACT Long-term studies of populations of white-tailed ptarmigan (Lagopus leucurus) were continued at hunted (Mt. Evans) and unhunted (Rocky Mountain Na~ional Park) areas in Colorado through 1987. Densities of b reed Lng ptarmigan decreased slightly at Mt. Evans but increased at Rocky Mountain National Park. Nesting success at both sites was good to excellent but brood size at Mt. Evans was poor (2.0 chicks/hen). There was no known harvest at Mt. Evans Ln 1987 because the highway was closed due to road construction. Apparently, no hunters walked the 3+ km into areas where ptarmigan were known to occur. ��299 POPULATION DYNAMICS OF WHITE-TAILED Clait E. Braun and Kenneth PTARMIGAN M. Giesen Long-term studies of trends in population size and investigation of reasons for fluctuations in size of tetraonid populations are lacking. Studies on the population dynamics of unhunted and hunted populations of white-tailed ptarmigan were init.iat.ed in Colorado in 1966 and have continued essentially unint.errupt.ed at. 2 sit.es. St.udies of t.he unhunted population (Rocky Mountain National Park) identified possible short-term cycles of 7-8 years with an amplit.ude of 25-30% bet.ween high and low breeding densities. Conversely, studies of t.he manipulat.ed populat.ion (hunt.ed) at Ht. Evans have no t indicated any cyclic pattern and it. would appear t.hat.controlled hunt.ing may mask any long-t.erm trend t.hat.may occur. ·This study is designed to examine the quest.ion whet.her whit.e-t.ailed ptarmigan are t.ruly cyclic and whether hunting affects the apparent oscillations. P. N. OBJECTIVES The goals of this investigation are t.o be able to predict the length and amplitude of cycles in whit.e-tailed ptarmigan in Colorado, to examine the impact. of hunting on cycles, and to clarify underlying causes of the apparent cycles. SEGMENT OBJECTIVES 1. Conduct breeding (May-Jun) and brood (Aug-Sep) censuses ptarmigan using tape-recorded calls of males (breeding) (broods). 2. Censuses will be conducted on previously established, defined study areas at Mt. Evans (hunted) and at Rocky Mountain National Park (unhunted). 3. Capture (noose poles) and band (aluminum and plastic color-coded bands) all umarked white-tailed ptarmigan encount.ered on study areas at Mt. Evans and at Rocky Mountain National Park. 4. Individually identify all ptarmigan observed on study areas at Mt. Evans and Rocky t·lountainNational Park through use of binoculars. 5. Make hunting season and bag limit recommendations for Mt. Evans and collect hunting data through use of volunteer wing barrels and hunter field checks. 6. Compile data, analyze results, and prepare progress of white-tailed and chicks reports. STUDY AREA AND METHODS Areas investigated were Mt. Goliath-Mt. Evans in Clear Creek County and at Tombstone Ridge-Sundance Mountain to Fall River Pass in Rocky Nountain National Park in Larimer County. The physiography, geology, location, and �300 vegetation of these study areas have been previously 1971; Braun and Rogers 1971; Giesen 1977). described (Braun 1969, Ptarmigan were located through use of tape-recorded calls (Braun et ale 1973), captured through use of telescoping noose poles (Zwickel and Bendell 1967) as described by Braun and Rogers (1971), and classified to age and sex and banded following Braun and Rogers (1971). Age of chicks was estimated following Giesen and Braun (1979). Numbered plastic bandettes were not used as in earlier years (Braun and Rogers 1971) as a color-code system using up to 4 different colored plastic bandettes was instituted in 1977-78. A check station was scheduled to be operated on the Mt. Evans highway during the opening weekend of the ptarmigan season in that area. A volunteer wing collection station was available to hunters in the area when the check station was not in operation until the season closed. A sample of hens was radio marked at Mt. Evans as part of another study (K. Martin, pers. commun.). RESULTS AND DISCUSSION Breeding Densities Mt. Evans.--Timing of breeding events in the Mt. Evans area was about 1 week later in 1987 than in 1986, probably because of a major storm on 3-5 May. During the May-early June interval, 9 pairs and 2 single males were identified. Thus, breeding densities decreased from levels documented in 1985786 (Table 1). This decrease was primarily the result of territories on lower Mt. Goliath and west Mt. Goliath being unoccupied. During the breeding season, 5 of 11 males identified were yearlings while 6 of 9 hens were yearlings. While recruitment of yearlings was good in 1987, overwinter survival of yearlings and adults identified in 1986 was poor. Kocky Mountain National Park.--Surveys of ptarmigan present on breeding territories along Trail Ridge Road in R~mP in May and June indicated the minimum breeding population was 33 birds, which was comprised of 13 pairs and 7 single males. This represents a 16 percent increase over the 28 birds (12 pairs and 4 single males) identified in 1986 (Table 1). The increased breeding densities reflect a return to average survival of adults and recruitment of yearlings. Survival of banded adults from 1986 was 63.6% (38 of 52 males, 11 of 25 females). Recruitment of chicks banded in 1986 was only 13.8% (4 of 29) although yearlings comprised 16.5% of all adult ptarmigan identified in 1987. Nesting Success and Brood Size Mt. Evans.--Eight hens were located during mid July-early September 1987 on or immediately adjacent to the study area. Five hens (62.5%) were with broods while 3 were apparently unsuccessful nesters (without chicks). Of 11 known first nests in the Mt. Evans area, 6 were successful (K. Martin, pers. commun.). The 1 known renest was also successful, thus 7 of 11 (63.6%) hens whose nests were found were successful. Average brood size to 1 September was poor (2.0 chicks/hen). Data from 10 chicks that were banded indicated hatch dates from 30 June to 23 July with most (60.0%) hatching from 8 to 19 July. Data from 7 successful known nests indicated hatching dates from 24 June to 18 July (renest) with 4 of 7 (57.1%) hatching on 5-7 july. �3Jl Table 1. 1966-87. White-tailed ptarmigan breeding densities (birds/km2), Colorado Study area Year 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1971) 1979 191)0 1981 191)2 1983 1984 1985 1986 1987 Rocky Mountain National Park (5.5 km2) 11.3 9.8 U.S 12.0 9.6 9.1 8.7 7.8 8.0 11.1 13.5 12.9 10.7 8.7 8.4 8.2 7.8 6.7 5.8 6.0 4.5 6.0 Mt Evans (4.0 km2) 3.0 2.7 2.7 2.2 2.0 4.2 7.5 6.2 6.2 6.2 6.7 >6.0 7.5 10.3 9.5 9.0 6.5 6.5 8.0 8.0 6.5 5.0 Rocky Mountain National Park.--Nest success was estimated at >90% (11 of 12 hens observed in Ju1-Aug were with broods) which is the highest since the study was initiated in 1966. Hatch dates were calculated from primary molt and growth for 44 chicks captured for banding. Hatch dates ranged from 16 June to 21 July (Median = 2 Ju1) and were among the earliest recorded. Survival of chicks to 1 September was higher than average with an average August brood size of 4.4 chicks. Harvest Mt. Evans.--The hunting season at Mt. Evans in 1987 opened on 19 September and closed on 4 October (16 days) with a bag and possession limit of 3 and 6. Thus, the season was delayed 1 week from the statewide opening as it was in 1981 and 1986. The season opening was delayed 2 weeks from 1978 to 1980 and 1982 to 1985. Prior to 1978, experimental seasons were in effect (1970-1976) or the season opened with the statewide grouse seasons (dates from 17 Aug to 14 Sep). While the season was open in 1987, the Ht. Evans road was closed (gate) to all vehicular traffic because of road rebuilding in the Lincoln Lake area. Thus, no check station was operated on opening weekend but a volunteer wing collection station was in place on the highway throughout the season. No �302 wings were received from ptarmigan, no birds were known to have been harvested, and no bands were reported. LITERATURE CITED Braun, C. E. 1969. Population dynamics, habitat, and movements of whitetailed ptarmigan in Colorado. Ph.D. Thesis, Colorado State Univ., Fort Collins. l89pp. _____ 1971. Habitat requirements of Colorado white-tailed ptarmigan. West. Assoc. State Game and Fish Comm. 51:284-292. , and G. E. Rogers. 1971. ----Colorado Div. Game, Fish and Proc. The white-tailed ptarmigan in Colorado. Parks Tech. Publ. 27. 80pp. _____ , R. K. Schmidt., Jr., and G. E. Roger~. 1973. Census of Colorado whitetailed ptarmigan with tape recorded calls. J. Wildl. Manage. 37:90-93. Giesen, K. M. 1977. Mortality and dispersal of juvenile white-tailed ptarmigan. M.S. Thesis, Colorado State Univ., Fort Collins. 55pp. , and C. E. ----white-tailed Braun. 1979. A technique for age determination of juvenile ptarmigan. J. Wildl. Manage. 43:508-511. Zwickel, F. C., and J. F. Bendell. J. Wildl. Manage. 31:202-204. Prepared by -=-=--,:&::;_",W:::--'~z_. --'~~ 1967. Clait·E. Braun Wildlife Research Leader Kenneth M. Giesen Wildlife Researcher B A snare for capturing blue grouse. _ � https://cpw.cvlcollections.org/files/original/ef585dbb03f427ad1f28eee4f55a5fa8.pdf dd74b22c36a0f839aff20687271a6f02 PDF Text Text JUJ JOB FINAL REPORT Colorado State of: Project: W-152-R (W-37-R) 23 Work Plan: 1 Job Title: Evaluation of No-till Wheat Farming Period Covered: Author: : Job Avian Research 01 January through 31 December 1987 Warren D. Snyder Personnel: W. Snyder, Colorado Division of Wildlife ABSTRACT The quantity, quality, and security of wheat stubble as nesting cover for ring-necked pheasants (Phasianus colchicus), mourning doves (Zenaida macroura), and other ground-nesting wildlife were documented under different cropping systems on a Colorado Division of Wildlife Property (1984-85) and private lands (1986-87) in eastern Colorado. Biennial wheat-fallow cropping rotations using conventional tillage methods for summer fallow continued to dominate the agricultural landscape in eastern Colorado but provided low security for nesting wildlife due to the timing of initial summer fallow tillage. A 3-year cropping rotation (wheat-row crop-fallow), termed ecofallow, is gaining rapid acceptance but comprises <5% of the cropland in most eastern Colorado localities. Its security for nesting wildlife is considered slightly better than that of conventional wheat-fallow rotations. A biennial wheat-fallow rotation using no-till fallowing has not yet gained significant acceptance among eastern Colorado farmers but has the potential of dramatically increasing both the quantity and security of nesting cover. However, recent trends toward increased use of semi-dwarf wheat varieties (2%"':19ia- to almost 60%;_1987) have progressively reduced the height-density index (HDI) of wheat stubble for nesting wildlife. Impacts have primarily affected fall to spring survival of pheasants and the nesting quality of wheat stubble under biennial no-till wheat-fallow cropping. Comparisons of wheat stubble quality with that of green wheat, native unmanaged and unpastured grass, unharvested alfalfa, cool-season tame grasses, and switchgrass (Panicum virgatum) revealed that the latter retained the highest HDI in early spring prior to vegetation growth. Green wheat and alfalfa possessed similar growth patterns. Native grass usually had a lower HPI than wheat stubble through early spring. Recommendations concerning use and management of winter wheat on Colorado Division of Wildlife properties are presented. �304 RECOMMENDATIONS 1. Biennial no-till cropping of winter wheat is recommended because it retains 100% of the wheatland in secure nesting cover for ground nesting wildlife and retains standing stubble for wildlife use from fall to spring. 2. The most practical and economical farming approaches available will be used on private lands and their impacts on wildlife usually are a minor consideration. However, farmers should be informed as to the impacts or benefits of different wheat farming approaches on wildlife including the potential values of biennial no-till cropping of winter wheat. 3. Fields of biennially-cropped winter wheat are recommended for use on Colorado Division of Wildlife (CDOW) properties in eastern Colorado when managing for pheasants and mourning doves because wheat fields provide combinations of covers for nesting, brood rearing, night-roosting, feeding, and escape. 4. Tall wheat varieties and a minimum stubble height (12 in. or 3 dm) should be required and included in CDaW sharecropping contracts. 5. Since minimal manpower is needed for no-till fallowing, the CDaw should consider eliminating sharecropping when possible and using proper equipment and property technicians to apply herbicides. Planting, fertilizing, and harvesting can be custom contracted allowing the CDaW greater control over wheat stubble quality. 6. Ecofallow and annual cropping of small grains should not be permitted on CDaW properties. �305 EVALUATION OF NO-TILL WHEAT FARMING Warren D. Snyder Wheat stubble and green wheat comprise the vast majority of cover available to ground nesting wildlife in eastern Colorado. However, the insecurity of these covers, especially wheat stubble, under conventional summer fallow has previously been documented (Snyder 1984). New reduced-till or no-till farming methods (Table 1) have been receiving increased use in eastern Colorado in efforts to increase crop production and to conserve soil, soil moisture, reduce tillage operations, and reduce labor and fuel costs. The approach that appeared to be of greatest potential benefit for wildlife was a biennial no-till winter wheat-fallow crop system where herbicides completely replace tillage in weed control management. Based on interest expressed by Division management biologists, a study to evaluate this approach on a Division of Wildlife property was implemented. However, several problems were encountered in trying to conduct this evaluation. Lack of essential equipment was a major factor. Therefore, the study and its primary objective were modified slightly and transferred to private farmland scattered over eastern (primarily northeastern) Colorado. This report presents the findings from both the 1st (1984-85) and 2nd (1986-87) phases of this study. Table 1. The amount of small grains, corn, and grain sorghum planted under no-till cropping systems in 1983 and 1986 in relation to the total amount planted in Colorado in 1986. Crop Small grains - Fall Spring Total Corn Grain Sorghum Ha planted under no-till 1983 1986 6,742 1,9l2a 1,264a 25,934 7,368 33,302 9,570a 7,056a Total ha 1986 1,434,506 329,406 1,763,914 404,836 193,126 % of Total 1.81 2.24 1.89 2.36 3.65 aThe majority of no-till corn and grain sorghum was assumed to be in wheat stubble under the ecofallow or wheat-row crop-fallow rotation. (Data extracted from Conservation Tillage Information Center 1983 and 1986 Colorado County Summaries, James E. Lake, Person. Commun.). P. N. OBJECTIVE Test biennial, annual, and ecofallow wheat and wheat-row crop no-till, and reduced tillage crop systems for increasing the quality, quantity, and security of nesting cover for ground nesting birds in eastern Colorado. The initial objective was to test a no-till wheat fallow biennial-cropping system on Colorado Division of Wildlife (CDOW) properties within eastern Colorado. A 2nd objective was to compare the quality of wheat stubble with other available or established covers for ground-nesting wildlife. �306 METHODS AND MATERIALS Background information concerning new farming methods, equipment, herbicides, wheat varieties, and soil moisture measurement techniques were obtained from the literature, personal contacts, and field demonstrations. Dr. D. Smika, Agronomist, U.S. Department of Agriculture, Agriculture Research Service (Akron, Colo.) who had worked extensively in developing no-till wheat farming systems, was a primary contact. Selection of study sites and fields during the initial phases was coordinated with CDOW management biologists. Precipitation data were obtained from U.S. Department of Commerce, National Climatic Data Center, Climatological Data Summaries for stations in eastern Colorado. These data were averaged to obtain mean monthly and mean annual totals. Precipitation accumulated over 16 months (Jan-Dec, Jan-Apr) combined with the departure of the mean April temperature from the long-term mean provided a precipitation - temperature index that was correlated with wheat growth and stubble height (Snyder 1984). Usually, all biennial no-till, annual no-till, and switchgrass fields that could be located were sampled. Other covers were randomly sampled after making an initial effort to stratify sampling among several locations and farmers within several northeastern Colorado counties. Sampling was extended into southeastern Colorado in 1986 to obtain an adequate sample of no-till treated fields (Fig. 1). §ampling was restricted to northeastern Colorado in 1987. Height-density indices (HDI) of residual and green vegetation were obtained following Robel et ale (1970) as modified by L. M. Kirsch (Unpubl. Rep., U.S. Fish and Wildl. Serv., Jamestown, N.D., 1977). Wheat stubble was sampled in early spring prior to major growth of volunteer wheat and forbs. Other covers were sampled at progressive intervals through spring to establish the pattern of growth. The starting point and direction of sampling were randomized along an accessible side of tracts or fields. A 2nd HDI was obtained within ecofallow fields within 2 weeks after they had been planted to a row crop. Progression of spring tillage of wheat stubble fields under conventional summer fallow was monitored along a route from Holyoke to near Fleming in 1986 and 1987 (Fig. 2). Monitoring was begun 1 April and continued at 2-week intervals until >95% of all fields had been cultivated. Progression of slot-planting of row crops or drilling of millet within ecofallow stubble fields was monitored at 7-10 day intervals along a route in northwestern Phillips County during 1986-87 (Fig. 2). Monitoring began in late April and continued into June. Definitions Annual Winter Wheat Cropping System.--The system by which winter wheat is harvested in mid-summer and planted in the subsequent early fallon an annual basis. Conventional tillage is used to reduce the residual stubble prior to planting or a herbicide is applied to kill forbs and volunteer wheat prior to drilling directly into the standing stubble (annual no-till). �307 .. ' -'--------,----B- E R T 34 C R B L 0 l -, Fig. 1. Height-density sampling locations in eastern Colorado, Spring 1986-87. Sampling was not conducted at sites 8 and 9 in 1987. �308 I -IJ ,..c: e H 0 c ''''; T I I Q) '-" ::l -IJ 0 H Q) e-, :> ::l H :3 (JJ 0 r-i r-i o til 4-< 0 Q) til c -'"0 I I' • '-" Q) ::l -IJ :>, 0 H Q) :> H ::l (JJ Q) . r-- 0 '"0 ~ r-i I \.0 00 0\ bDOO til ,...., r-i .,..; -IJ Q) 0 ,_.co r-i -IJ Q) co -IJ (JJ Q) ,_.e 0 U r-i ,.0 ,.0 :l UJ N · · eo ''''; ••• �309 Biennial No-till Winter Wheat - Fallow Cropping Rotation.--A system wherein herbicides are substituted in summer after harvest and/or during the subsequent spring as a replacement for tillage forb control within summer fallow. Conventional Winter Wheat - Fallow Cropping Rotation.--A system wherein discs, chisels, sweeps, or plows and other tillage implements are used to cultivate fallow ground to control forbs, loosen the soil, and prepare a seedbed during the summer fallow cycle. Tillage begins after harvest or is delayed until the following spring and is then continued as needed until wheat planting in fall. Ecofallow.--A crop system using a persistent herbicide, usually applied following wheat harvest, where corn or grain sorghum is slot planted directly into the standing stubble during the following spring (millet may be substituted for row crops). Conventional summer fallowing is usually conducted during the 3rd year prior to wheat planting in fall. Thus, 2 crops are harvested during a 3-year rotation. RESULTS AND DISCUSSION Quantity of Wheat Stubble for Nesting Wheat stubble represented about 75% of the available early spring pheasant nesting cover (assuming pastureland [20.4%] could be included as nesting cover) when approximately 1,000 mi2 was inventoried in extreme northeastern Colorado Tablelands in the mid-1960's (Snyder 1970). If short-grass pastures 'which usually offer extremely marginal nesting cover) were excluded, wheat stubble represented about 95% of the available nesting cover prior to onset of major green wheat growth. Starting in mid to late April, winter wheat rapidly becomes a co-dominant in importance for nesting with wheat stubble or replaces it since stubble is being lost to summer fallow tillage. By mid to late May green wheat becomes the dominant nesting cover. Wheat stubble and green wheat in combination represented about 90% of the available nesting cover for pheasants during a 1979-81 study in southeastern Sedgwick County (Snyder 1984). The amount of wheat stubble available in spring may be markedly altered by federal farm programs administered by the U.S. Department of Agriculture's Agricultural Stabilization and Conservation Service (ASCS) and by other factors. For many years wheat acreages have been reduced by annual "set-aside" programs. In 1986 and 1987 further reductions occurred because of the Conservation Reserve Program (CRP). Over 14% of eastern Colorado farmland had been committed to this 10-year retirement program by February 1987 and sign-ups will continue for several more years in most counties. The majority of these lands were previously farmed to winter wheat on a biennial cropping rotation. About 40% of stubble fields were tilled after wheat harvest in 1986 due to weed problems or other factors. Thus, the amount of standing stubble available in early spring 1987 was greatly reduced. This amount of late-summer tillage had not been noted in previous years in northeastern Colorado. However, the proportion of post-harvest tillage usually progressively increases from north to south across eastern Colorado. �310 In spite of these variables, wheat stubble remains the most abundant early spring nesting cover for pheasants in northeastern Colorado until green wheat attains adequate growth. Conventional summer fallow using tillage remains the major treatment of wheat stubble and represents over 90% of the area. Ecofallow has been rapidly gaining increased acceptance but its use in 1987 was restricted to <5% of the total area. Only about 1.9% of the fields in eastern Colorado were committed to the biennial no-till (wheat-fallow-wheat) cropping rotation in 1986 (Table 1) but the number of fields appeared to double in 1987. Some of this increase was possibly due to the cost share incentive payments farmers received under the ASCS administered Agricultural Conservation Program (ACP). A few farmers used annual no-till winter wheat farming in 1986 but area involved declined in 1987. Annual cropping conflicted with crop area restrictions which were mandatory if farmers were to receive payments for retiring cropland under ASCS administered federal farm programs. Quality of Wheat Stubble as Nesting Cover Weather-wheat Growth-Stubble Relationships.--Spring growth and quality of wheat is a product of soil moisture accumulated during the year of fallowing and the late winter and spring months when major growth occurs, and to a lesser degree, is a product of the departure of spring temperatures from long-term averages (Snyder 1984). wnen the mean April temperature (F) departure from the long-term mean was added to the l6-month accu~ulated· precipitation (in.), a direct positive relationship with 10 May wheat height (HDI) was ohtained. Inadequate samples were obtained in 1984, however, 1985-87 data provide continued evidence of this weather-wheat growth relationship (Fig. 3). The 1979-81 Sand Draw study (Snyder 1984) provided limited evidence that wheat stubble height was partially a product of wheat growth as influenced by weather. The precipitation-temperature index when compared with the 1984 and 1985 (South Republican) stubble heights closely fit this relationship (Fig. 4). However, subsequent 1986-87 stubble HDI samples, obtained from scattered locations in eastern Colorado (Fig. 1), showed deviation from data obtained in previous years (Fig. 4). Average HDI's for 1986-87 were lower than most preceding years (Fig. 5) even though the precipitation-temperature index was average or above. One possible explanation for the discrepancy between 1986-87 samples and previous data lies in changes in wheat varieties planted. Standard tall wheat varieties were planted in both the 1979-81 and 1984-85 study areas. However, a rapid shift to semi-dwarf varieties in east central and northeastern Colorado has occurred in recent years. Only 2% of wheat planted in 1978 was semi-dwarf varieties, this percentage increased to 50 in 1986 and 60% in 1987. Semi-dwarf varieties averaged <80 em taIlor about 33% shorter than standard wheat varieties (J. Echols, pers. commun.). These new, shorter varieties would be expected to yield shorter residual stubble (Figs. 4 and 5). Other Variables Determining Stubble Quality.--Most wheat fields appear uniform in height, however, farmers report many wheat heads are on shorter stalks forcing harvesters to cut lower leaving shorter stubble. The extent that this varies among farmers, wheat fields, wheat varieties, and years is unknown. Strong winds and light to moderate hail, produced by pre-harvest thunderstorms, �311 75 .1981 ~ :E ..- 0 60 Z 0 E o ._ I .1966· .1963 45 CJ .1966· 1980 W I ~1985-87 ~ w I 3: 30 Z W W a: CJ .1965 .1979 15 20 16 MONTH 30 25 PRECIPITATION + APRIL 35 TEMPERATURE Fig. 3. The relationship between whea t height on 10 Hay and an index combining precipitation (in.) over a l6-month (JanApr) interval plus departure of mean April temperatures from the long-term average, northeastern Colorado. Data include individual samples from 1963-64, 1979-81, 1985-87, and 2 samples for 1965-66 and 1968. �312 2.5 .1981 E -c 2.0 X W 0 ~ >ICi5 1.5 Z . W .1978 0 l- *1985 I CJ iIi :::r: 1.0 *1984 W -' • CD CD 1980 (HAIL) ::> I- *1986 en ~ 0.5 * w I 5:: 1987 .1979 o ~ 15 ~~ 20 ~ ~ 25 30 ~ 35 16 MONTH PRECIPITATION + APRIL TEMPERATURE DEPARTURE Fig. 4. The relationship between the whea t stubble heightdensity index and an index combining precipitation (in.) over a l6-month (Jan-Apr) interval plus departure of mean April temperatures from the long-term average, northeastern Colorado, 1979-87. �2.5 ............, EXCELLENT 2.0 ,-.... "d S '-" :>< ~(/) 1.5 z ~ GOOD r;-;-; A I E-< ~ ::r: ~~ ............ 1.0 ::r: FAIR r:-;-;-: ~ j:Q j:Q ~ (/) r.:: 0.5 E-< n ~ POOR .. <I; ~ ~ o 1979 1980 1981 1982 1984 1985 1986 1987 Fig. 5. Average height-density (dm) of wheat stubble within the Sand Draw (1979-82), South Republican (1984-85), and eastern Colorado (1986-87) study areas in relation to an arbitrary quality rating. w •..... w �314 may also lodge wheat forcing harvesters to cut lower. Wide variation among localities and years also occurs. Occasionally, wheat fields are intentionally harvested close to the ground so that straw can be baled for livestock or other commercial uses. Commercial fertilizers, used by most but not all wheat farmers, also impacts wheat growth and stubble quality as does soil type and quality. Eight of 21 stubble fields sampled in 1986 and 11 of 35 fields sampled in 1987 were farmed using no-till "herbicide" fallow methods during the respective years they were sampled. Few, if any, had been no-till fallowed previously so differences in stubble HDI resulting from past fallow treatment were not expected. Comparisons would need to be made between fields continuously subjected to no-till vs. conventional tillage farming before determinations as to the impact of farming method on stubble quality could be discerned. I assumed there were no significant HDI quality differences in stubble among treatments. Therefore, all stubble fields were combined and treated as samples from a single population. Wheat Stubble Quality-Wildlife Use Relationships.--The change to semi-dwarf wheat varieties has progressed rapidly and is expected to continue resulting in shorter wheat stubble than in the past. As a consequence, wheat stubble will become less attractive as early spring nesting cover for pheasants and some species of waterfowl (Anatidae). Under conventional summer fallow, tillage of stubble fields, which usually occurs during the s'prLng nesting season, will have less negative impact because fewer pheasant nests will have been started there. However, if the security of wheat stubble is enhanced using no-till methods, then shorter stubble means lower quality and reduced use by pheasants and other ground nesting wildlife. Nests placed in shorter stubble would also be less secure from predation. An arbitrary quality rating for pheasant nesting is provided (Fig. 5). Stubble HDI in 1986-87 was in the poor to fair range although stubble quality in most preceding years, when standard wheat varieties were used, had averaged from fair to good. The exception was in 1980 because of drought conditions in 1979 that severely impacted wheat growth. Wheat stubble has been used extensively by pheasants throughout much of the year for night-roosting, feeding, escape, protection from avian predators, and during blizzards and periods of deep snow. However, reduced stubble height, which may be more common in the future, will make cover conditions more marginal. Thus, pheasant survival will potentially be lowered and fewer hens will be present for nesting. The HDI quality ratings (Fig. 5) do not pertain to all wildlife species, most notably mourning doves. Personal observations, primarily during the 1979-81 Sand Draw Study, indicated that mourning doves used wheat stubble extensively for placement of first nests, especially in fields close to tree plantings. Dove nests typically were placed adjacent to a vertical clump of stubble usually in open or sparse stubble or stubble that had been undercut with a sweep plow. Soutiere and Bolen (1972) noted that most ground nests of doves had vertical cover on at least 1 side but overhead cover was not important within a burned Texas rangeland site. Doves were apparently attracted to the burn for nesting because of reduced litter and vegetation. Based on these observations and others in eastern Colorado rangelands, mourning doves would usually select against dense wheat stubble that retained minimal between-row �315 openings. Thus, the HDI rating at most dove nest sites would probably be low. Therefore, semi-dwarf wheat stubble fields which retain shorter more open stubble may be more attractive to nesting doves. These fields also provide open feeding sites for doves and other small avifauna. Little information is available in the literature to support the hypothesis that doves select against dense cover for nesting. Olson (1980) apparently conducted mourning dove nest searches in eastern Colorado after wheat stubble had been converted to fallow ground. Dove nesting there and in green wheat was reported as insignificant. Downing (1959) reported only 2 dove nests in 104 ha of wheat stubble searched soon after wheat harvest in northwestern Oklahoma; ground nest densities there were much lower than in nearby rangelands. Almost no dove nesting was noted in Illinois stubble fields (Hanson and Kossack 1963). It is assumed that wheat stubble there would be taller and more dense than in eastern Colorado. It also seems probable that dove use of stubble fields in eastern Colorado is highly variable based in part on stubble quality but also on the availability of other proximal life requirements. Quality of Nesting Cover in Annual No-till Fields.--Several northeastern Colorado fields were found that received an annual no-till winter wheat cropping treatment in spring 1986. Following wheat harvest in July 1985, these fields had been treated with a herbicide and fertilized or the latter was applied at the time of planting in September. Within the 5 sampled fields, height-density quality of wheat stubble and g~een wheat were listed separately and combined during i April sampling. Sites where wheat stubble was the dominant obstruction averaged 0.27 dm, those for wheat were 0.12 dm yielding a combined average of 0.18 dm. Thus, planted no-till wheat fields under annual cropping provided extremely poor nesting cover for pheasants but their value for first nests of mourning doves might be much better than in conventionally fallowed wheat fields. Growth of green wheat in these annually cropped fields was the same in 1986 as that in biennially cropped fields. Thus, their nesting quality tended to improve in the same time frame. Fall drilling into wheat stubble knocked nearly all of the stubble over or flattened it. However, the residual straw, both standing and flattened may have made annual no-till fields more attractive for nesting by pheasants, doves, horned larks (Eremophila alpestris), and other wildlife, whereas soil surface in conventionally fallowed wheat fields was nearly void of residual straw. Quality of Nesting Cover Under Ecofallow.--Slot planting of corn or grain sorghum within wheat stubble under the ecofallow (wheat-row crop-fallow) 3-year cropping system greatly reduced the quality of the remaining stubble as nesting cover. In both 1986 and 1987 mean indices averaged among fields were below 0.2 dm or they provided almost no standing concealment cover. Such fields would be of little value to nesting pheasants but might receive considerable use from mourning doves and other species with reduced cover requirements. Observations indicated post-plant HDI's were usually in direct relation to pre-plant stubble quality. The highest post-plant HDI obtained in 1986 was 0.26 dm from a field with a pre-plant HDI of 0.63 dm. In 1987 the highest post-plant HDI was 0.33 dm in a field where the pre-plant index was 1.1 dm. Observations indicatd a few fields existed containing considerably higher �316 pre- and post-plant HDI's. Where stubble fields contain much higher quality cover in bordering states to the east (where annual precipitation is greater), post-plant nesting quality would be expected to be much higher. Security of Wheat Stubble as Nesting Cover The primary reason wildlife managers have expressed an interest in new cropping systems involving winter wheat is because of their potential increased security ,for nesting. The insecurity of wheat stubble for pheasant nesting under conventional summer fallow has been documented (Snyder 1984). Eggs in only 1 of 18 nests hatched prior to tillage and the majority (11) were destroyed by tillage. Numerous other nests in the laying stage were suspected but could not be located and documented before they were destroyed. Stubble tillage progressed slowly from late April to early June along the tillage transect line during the 1979-81 Sand Draw study. As a consequence of this late tillage, 12 of 16 pheasant nests had advanced into incubation before being lost. Stubble tillage in 1986-87 progressed more rapidly (Fig. 6) and was comparable in timing to averages obtained from studies conducted in the 1960's and 70's within northeastern Colorado. Possibly fewer nests had advanced into incubation as a consequence but in 1987 numerous fields remained untilled into late May and early June. Surveys conducted in spring 1987 revealed that approximately 40% of the stubble fields had been disked or,undercut the previous year sometime after wheat harvest. Among fields tilled in spring, most were disked, however, individual farmers used different tillage implements including plows, chisels, and sweeps. In nearly all cases the majority Qf the standing cover was lost and the opportunity for a nest to survive intact was slight or nonexistent. The ecofallow cropping system, wherein a row-crop was slot planted directly into standing wheat stubble, appeared to be slightly more secure than conventional summer fallow tillage for ground nesting wildlife. Corn was planted beginning in late April or early May and most had been planted prior to when most pheasants would start incubation of clutches (Fig. 7). However, grain sorghum was usually planted from mid to late May which would be during the peak interval for pheasant incubation. Because of area restrictions for corn, grain sorghums, and wheat, millet planting has increased in recent years within eastern Colorado. Whereas, a nest had a limited opportunity of not being destroyed during slot planting of row-crops (probability <50%) millet was usually drilled in early to mid June (Fig. 7) yielding little chance that nests would not be destroyed. Most nests would be in incubation stages when lost. Some farmers apply fertilizer after their row crop has emerged which would cause further disturbance. A few also cultivate the crop to loosen the soil and to give the grain stalks additional support. Use of these practices does not appear extensive in northeastern Colorado. Since planting occurs directly into stubble in early fall, no disturbance occurs during the spring nesting seasn within annual no-till wheat. However, the quality of the cover over winter and prior to major wheat growth in late April is poor. �317 100 1986 80 ....... .. .. 60 ••• 1979-81 ...../ .. ... .... 40 20 ...• . ... o· o .....' o o 20 24 8 29 22 15 APR 5 HAY PROGRESSION OF 12 JUN STUBBLE TILLAGE Fig. 6. Progression of spring cultivation of wheat stubble fields during spring 1986 and 1987 compared with averages from previous intervals, northeastern Colorado. _,,' 100 r: 1987 " 80 o "-l e~ 60 c, :t: " ," CORN " MILLET GRAIN SORGHUM I" o u I I I I I e:z: "-l 40 u '""-lc, I I 20 I I ,J o 5 10 15 20 25 30 4 9 MAY PROGRESSION OF ROI,r-CROP PLANTIXG 14 19 24 JUN IN ECOFALLOI,r STUBBLE Fig. 7. Progression of row-crop planting stubble during 1986 and 1987, northeastern into ecofallow Colorado. wheat �318 The security of the biennial no-till wheat-fallow cropping rotation appears to be high and comparable to that of unfarmed covers. The only disturbance during the nesting season may be a late spring herbicide treatment. Since sprayers have wide booms, usually <10% of the ground will be contacted by machinery. Some nest abandonment might occur during spraying operations but its extent remains unknown. Most herbicides being used during spring applications have relatively low toxicity to wildlife or their eggs and little direct negative impact has been documented to date (Duebbert and Kantrud 1987). The Quality of Other Covers in Relation to Wheat Stubble Green Wheat.--The growth curve of semi-dwarf varieties of wheat may not be as steep as that for other wheat varieties, however, its value as nesting cover should not be markedly altered. Taller wheat varieties usually attain a HDI of 8-10 dm when headed. HDI values >3 or 4 dm are probably of little additional value for pheasants, the primary nesters in green wheat. Green wheat usually begins major growth in mid to late April and rapidly surpasses the HDI of wheat stubble in late April or early May (Fig. 8). Because of its uniform density and growth pattern, and lack of residual ground cover under conventional farming methods it receives little use by mourning doves or small passerine birds. Wheat planted directly into no-till stubble may be more attractive in early spring for doves and other species but this advantage is probably lost after growth >2.5 dm is attained. Wheat is a relatively secure nesting cover except when harvest occ~rs, normally during the first part of July in northeastern Colorado. Pheasants are the primary species affected because many hens are still incubating clutches placed there after previously being forced out of wheat stubble by tillage or predption (Snyder 1984). Native Grass.--Northeastern Colorado contains numerous small areas of unfarmed, ungrazed native vegetation within draws, odd corners, caliche knolls, right-of-ways, and other sites not suited for cultivation. These usually represent <5% of the available nesting cover but are moderately to highly secure from human disturbance. Dominant native grasses include blue grama (Boutelous gracillis), western wheatgrass (Agropyron smithii), sand dropseed (Sporobolus cryptandrus), and needle-and-thread (Stipa comata). Introduced annual bromes (Bromus spp.) often dominate in spring. The HDI quality of these tracts usually is low (Fig. 9) and if they surpass that of wheat stubble it isn't until late May when most stubble is already destroyed by tillage. Therefore, they are considered marginal in quality for pheasants. Short, high density vegetation within these tracts is not usually attractive to nesting mourning doves. However, these sites are considered essential to horned larks, western meadowlarks (Sturnel1a neglecta), grasshopper sparrows (Ammodramus savannarum), and lark buntings (Calamospiza melanocorys). A1falfa.--Numerous tracts in northeastern Colorado, planted to grass-alfalfa mixtures on CDOW leases and properties in the late 1970's and early 1980's, became strongly dominated by alfalfa. Sampling in 1986-87 revealed that alfalfa grew rapidly following a growth curve similar to that of wheat (Fig. 10) assuming average or better precipitation was received. Height was less than that of wheat but was adequate for nesting pheasants and the areas �319 5 ~~ ~6 I I I 4 I I e -e I 3 >- E-< >-< GREEN WHEAT (/) :z: c..J c I 2 E-< u >-< c..J = I·/HEAT STUBBLE 3 15 27 JUN HAY APR Fig. 8. He ight-densi t y (dm) progress ion of green whea t in relation to that of wheat stubble in 1986 and 1987, north- . eastern Colorado. 3 ..•..•. e "'" >•..... 2 E-< (/) :z: c..J c I t: G •....• c..J 87 0 J 15 APR L7 21 ':i MAY 2 14 JUN Fig. 9. Height-density (dm) pror,ression of native grass in relation to that of wheat stubble in 1986 nnd 1987, northeastern Colorado. �320 .,-,- . 8'3 , , ,, , .,~ "to >E-' H CIl :z: t<l ~~ 86 ~~ 2 ""E-'I == (.!) H t<l == WHEAT STUBBLE 0 3 IS APR 27 9 21 MAY 2 14 JUN Fig. 10. Height-density (drn) progression of alfalfa in relation to that of wheat stubble in 1986 and 1987, northeastern Colorado. �321 contained excellent dense foliage. Since a litter base was present along with considerable early spring diversity of height, its nesting value for pheasants was considered to be good. Past observations have shown that alfalfa usually tended to begin lodging by early summer and often lost most or all of its leaves during hot dry conditions. Late summer rains often stimulated regrowth. It also served as a food source for cottontails (Sylvilagus sp.) and plains deer (Odocoileus spp.), and provided brood habitat for pheasants. Loss of foliage and matting under winter snows has been a primary detractant to the value of alfalfa for wildlife, especially over winter. Old stands, >10-15 years of age, often become thin and codominated by annual bromes with reduced value for wildlife. Cool-Season Tame Grasses.--Tracts of unfarmed, ungrazed cool-season tame grasses were not abundant in northeastern Colorado during 1986-87. Nearly all that were located during the study were within CDOW leases or properties. Primary species included crested wheatgrass (Agropyron cristatum), intermediate wheatgrass (A. intermedium), and smooth brome (Bromus inermis). On the average, their early spring HDI quality ranked below that of wheat stubble, primarily because of over-winter lodging (Fig. 11). They slowly surpassed the HDI of stubble in mid to late April to attain moderate nesting cover value for pheasants by mid-May. Crested wheatgrass stands were open and sterile in appearance and of marginal value for most wildlife species. One tract containing smooth brome was within a grassed waterway "hich received above average moisture with resultant above average HDI. Past observations indicate this species routinely becomes sodded, nitrogen deficient, and offers marginal sterile cover on dryland sites in eastern Colorado within a few years after establishment. One bias within the sample was that most samples were >10 years of age and therefore of lower HDI quality and vigor than would be typified by younger stands. On the average, grass stands were older than alfalfa stands and some bias in HDI quality would be expected. Young stands of cool-season tame grasses «5 years of age) within CRP tracts would be expected to sustain higher HDI vigor approaching that of alfalfa but with a slower spring growth curve. Switchgrass.--The HDI quality of standing residual switchgrass, and in some tracts, other tall, warm season grasses, was much greater than that of wheat stubble or other sampled covers through early spring (Figs. 11, 12). Major new growth contributing to increased HDI was not attained until mid to late May. Average HDI indices/tract ranged from 1.1 to 3.2 dm within 5 tracts sampled in 1986 and from 0.5 to 4.1 dm within 8 tracts sampled in 1987. This illustrates the high variability among tracts. Younger plantings «10 years old) retained much higher HDI's and were potentially of high value for pheasant nesting, winter night roosting, and survival cover. Older stands retained considerable height diversity within individual tracts possibly increasing their attractiveness to passerine species. Prescribed burning or burning-tillage combinations could be used to restore HDI quality within the older stands, however, most were on private land. �322 3 STUBBLE 1) j LU ..! i JUN MAY APR Fig. 11. Height-density (dm) progression of switchgrass and tame grasses in relation to that of wheat stubble in 1986 and 1987, northeastern Colorado. . .... 5 o GREE:-I WHEAT :: e - . , ., ::," o 4 "'" 00 :, :, ., 3 .,; ALFALFA :, >- eCJl z W Q I 2 SIHTrHr.p·IC:C: e- = (.;) w ::: o j 15 APR 9 21 ~!AY 2 14 JUN Fig. 12. Height-density (dm) progression of major nesting covers averaged between 1986 and 1987, northeastern Colorado. �323 MANAGEMENT OPTIONS Wildlife managers have been urged to use winter wheat on CDOW properties in eastern Colorado when managing for pheasants and mourning doves (Rutherford and Snyder 1983). This is because it provides combinations of covers for nesting, brood rearing, night-roosting, and escape along with an excellent food source for pheasants, doves, and other wildlife. Wheat fields allow ease of movement by pheasants and create interspersions and diversity among perennial covers. Wheat fields provide both food and nesting cover whereas row crops provide only food. Wheat stubble stands well over-winter whereas many perennials lodge under snow. If winter wheat is to be of maximum benefit to wildlife, biennial no-till cropping should be used. However, farming practices including application of correct amounts of fertilizers and correct timing of herbicide application are essential to retain vigorous stands of both wheat and stubble. Tall wheat varieties and a minimum stubble height (12 in. or 3 dm) should be included in sharecropping contracts •• Sampling in 1986-87 revealed that wheat and stubble HDI was lower on some CDOW properties than on most adjacent private lands. Use of biennial no-till eliminates the existence of large bare fields over winter reducing snow drift and potentially increasing wintering wildlife survival. In spring and summer 100% of the land is in secure nesting cover and the vast amount of nesting cover available further increases nest security from predators. Division of Wildlife property technicians could readily farm biennially cropped no-till winter wheat on CDOW properties if they possessed necessary equipment and knowledge concerning the correct use and application of herbicides. Minimal manpower would be needed for herbicide application and planting. Fertilizer application and harvesting could be contracted to custom applicators and custom harvesters. Under this approach wheat stubble to 2 ft. tall (~6 dm) could be retained for yearlong use by wildlife. An alternate option for managing winter wheat on CDOW properties would be to require the sharecropper to apply short persistent herbicides in spring as a replacement for summer fallow tillage. The first tillage could then be delayed until mid-July or later, after the primary nesting season. Using this method, conventional rather than costly no-till drills could be used. Supplemental pre-plant fertilization would be essential to replace the nitrogen being used by decaying straw. Ecofallow or annual cropping of winter wheat should not be permitted on CDOW properties. Conventional summer fallowing should be discouraged along with stipulations as to wheat varieties used, stubble height retained, and fertilizer application. If conventional tillage is permitted, the field should be disced to reduce standing stubble, preferrably prior to 20 April or 1 May at the latest. Post harvest or spring sweep tillage should not be allowed. These stipulations should be written into sharecropping contracts. Little can be done to alter trends on private lands and a trend toward shorter, higher yielding wheat varieties is evident. Ecofallow is the primary new farming practice gaining acceptance in eastern Colorado. Biennial no-till or reduced tillage may result because of increased problems with annual bromes, jointed goatgrass (Adgilops cylindrica), and mustards (Brassicaceae) in wheat. �324 If crop surpluses and area restrictions ever become history, then a rapid trend toward an annual no-till winter wheat crop system may evolve in extreme eastern Colorado. At present there is no incentive in that direction. CONCLUSIONS There is evidence that wheat stubble HDI quality is deteriorating because of the rapid transition to semi-dwarf wheat varieties in eastern Colorado. This is not a concern for pheasant nesting under conventional summer fallow methods because nearly all nests placed there are subsequently lost to tillage or predation anyway. If it is less attractive for nesting, it may be better to prompt hens to place first nests elsewhere. The primary impact of shorter stubble is toward reduced fall to early spring survival of wildlife, primarily pheasants, because of greater exposure to blizzards and severe weather, and to increased exposure to avian predation. Reduced stubble HDI is primarily a concern for pheasant nesting under the bien;ial no-till winter wheat cropping system wherein stubble is not subject to tillage and therefore retains high security. Shorter stubble will retain lower HDI's and therefore be less attractive to nesting pheasants and nests placed there will be less secure from predation. In contrast, shorter, more open stubble may facilitate increased nesting use by mourning doves and some passerines. Reduced stubble HDI is of lesser importance for pheasant nesting under ecofallow but is a deterrent to winter survival of pheasants. Under annual cropping of winter wheat, reduced stubble HDI is of little significance since no undisturbed standing stubble is retained after fall planting. LITERATURE CITED Downing, R. L. 1959. Significance of ground nesting by mourning doves in northwestern Oklahoma. J. Wildl. Manage. 23:117-118. Duebbert, H. F., and H. A. Kantrud. 1987. Use of no-till winter wheat by nesting ducks in North Dakota. J. Soil and Water Conserv. 42:50-53. Hanson, H. C. and C. W. Kossack. 1963. The mourning dove in Illinois. Illinois Dep. Conserv. Tech. Bull. 2. Olson, T. E. 1980. Mourning doves and land use changes in Eastern Colorado. M.S. Thesis, Colorado State Univ., Fort Collins. 119pp. Robel, R. J., J. N. Briggs, A. D. Dayton, and L. C. Hulbert. 1970. Relationship between visual obstruction measurements and weight of grassland vegetation. J. Range Manage. 23:295-297. Rutherford, W. H., and W. D. Snyder. 1983. Guidelines for habitat modification to benefit wildlife. Colorado. Div. Wildl. Spec. Publ. 194pp. �325 Snyder, W. D. 1970. Pheasant hen harvest investigation •. Colorado Div. Game, Fish and Parks. Game Res. Rep" Apr il:3-81. 1984. Ring-necked pheasant nesting ecology and wheat farming on the -----High Plains. J. Wildl. Manage. 48:878-888. Soutiere, E. C., and E. G. Bolen. 1972. Role of fire in mourning dove nesting ecology. Proc. Tall Timbers Fire Ecology Conf. 12:277-288. Prepared by ~ a-UW ~ WarrenIf. snyaer Wildlife Researcher C ��31./ JOB FINAL REPORT Colorado State of: Project: 01-03-045 (W-37-R) 2 Work Plan: 21 Job Title: Dynamics of Cottonwood Regeneration Period Covered: Author: : Job Avian Research 01 January through 31 December 1987 Warren D. Snyder Personnel: C. E. Braun and W. D. Snyder, Colorado Division of Wildlife ABSTRACT Aerial photo interpretive methods were used by personnel of the Colorado State Forest Service (under contract) to obtain status and trends of cottonwoods (Populus spp.), shrubs, and other cover types along the lower portions of the Arkansas, South Platte, Colorado, and Rio Grande rivers in Colorado. A small segment of the South Fork of the Republican River was also inventoried. Aerial photos from earliest and most recent intervals yielded mean time lapse intervals of 25-37 years. Sampling intensity ranged from 15 to 27% using stratified random sampling procedures. Area occupied by cottonwoods declined >30% (P < 0.05) along the lower Arkansas River or at a rate of approximately l%/year, much greater than the 9% decline (p <0.10) along the South Platte and the 17.5% decline (P <0.10) along the Colorado River. In contrast, a slight increase (P>O.lO) In cottonwoods was recorded along the Rio Grande River. The South Platte retained a much greater area o;f cottonwoods per sample unit than any other river sampled. A marked trend toward more open stands of trees was noted along all rivers except the Rio Grande. The Colorado River had the youngest age structure among the rivers but it, like the other rivers, had a shift from younger to older trees over time. Thus, reproduction was not adequate to sustain existing stands of cottonwoods and most were in a decadent and declining status. Cottonwoods were nearly nonexistent along the sampled segment of the South Republican River in the 1930's but increased dramatically in recent decades. However, recent inventories indicate little natural regeneration of cottonwoods along this river. Tamarisk (Tamarix spp.) was replacing cottonwoods and willows (Salix spp.) along the Arkansas River whereas shrubs were declining along the South Platte and Rio Grande rivers but were apparently stable along the Colorado River. Declines in quantity of hay meadow occurred along all rivers from early to recent inventories, whereas dramatic increases were noted in quantities of agriculture and developed land. River channel declined dramatically along the lower Arkansas and Rio Grande rivers (p < 0.05), remained stable along the Colorado, and increased along the South-Platte. �328 Extensive snowpacks in the Colorado Rockies in 1982-83 and 1983-84 caused increased stream flows in 1983 and 1984 on most Colorado drainages. Flow volume along the South Platte and Colorado rivers was the highest recorded in recent history and caused flooding resulting in natural regeneration of cottonwoods, willows, and other species. Monitoring of intensive seedling transects along the lower South Platte revealed that extensive regeneration occurred but natural mortality, primarily because of desiccation, severely reduced seedling populations. However, some seedlings survived in most samples and rate of survival improved over time so that widely distributed natural regeneration of cottonwoods was evident into fall1987. In addition, flooding stimulated considerable reproduction of other woody species including green ash (Fraxinus pennsy1vanica) which showed evidence of gradually replacing cottonwoods along the lower South Platte. Little recent natural reproduction of cottonwoods was observed along the Rio Grande and Arkansas rivers. Altered stream flows, natural successional change, and other factors influencing reproduction of cottonwoods and willows were reviewed. Experiments with stem cutting planting of cottonwoods, willows, and other woody species into sites containing high groundwater tables yielded poor survival along the South Platte, South Republican, and Arkansas rivers. Several factors contributed to poor survival, but fluctuating groundwater levels were the primary deterrent. Poor survival and planting difficulties made stem cutting a questionable management tool in most sites. Alternate methods of stimulating reproduction or replacing trees, shrubs, and v-ines are summarized. Recommendations were provided concerning future inventory and monitoring of riparian habitats and wildlife populations in combination to fully assess impacts of changing environments. �329 RECOMMENDATIONS 1. Inventories of riparian habitats along the San Juan, Dolores, Gunnison, North Fork of the Gunnison, White, Yampa, and other rivers should be conducted in the near future on a priority basis. Limited monitoring should be extended to lower portions of tributary streams in eastern Colorado including the Purgatoire, Big Thompson, Cache la Poudre, and North Fork of the Republican River. Where habitat deterioration appears greatest, as along the Arkansas River, monitoring should be more frequent and sites where threatened or endangered wildlife occur should be closely monitored. Sample size should be increased in future surveys of the Arkansas River and monitoring should be extended to and potentially beyond Pueblo Reservoir. Inventories of status and trends along the Rio Grande, Colorado, and South Platte rivers should be conducted again in the 1990's. Colorado Division of Wildlife personnel should work closely with Colorado State Forest Service personnel to ensure that the oldest and most recent aerial photos available are used in sampling. 2. Photo interpretive monitoring of previously established 1.61 km sample units as to status and trends of key vegetation types, as used in this study, should be continued at intervals of 10-15 years. These data should .be supplemented with permanent transects within each sample unit. Standardized sampling procedures that yield indices of vertical and horizontal diversity (Anderson et ale 1978), composition of woody species, and composition of herbaceous vegetation (percentages of annual, biennial, and perennial forbs, and annual and perennial grasses) should be used. 3. Census of key stenotopic wildlife species (Svoboda and Graul 1980, Miller 1985) (density/sample units) should be periodically conducted (3-5 yrs) following Fitzgerald (1978), and Sedgwick and Knopf (1986, 1987). Replicates over time demonstrate trends and impacts of habitat changes along major rivers. 4. Inventories of the quantity and quality of riparian habitats in important tributary drainages are needed, especially in grazed low elevation locations. 5. Pothole blasting (Hopper 1978), use of heavy equipment, of other methods of creating small basins or- ditches within the flood plain could be used to promote establishment of cottonwoods, willows, and other phreatophytes where high «1.5 m) groundwater tables exist. Irrigation, either by ditching and flooding or sprinkling could be used but would need to be sustained to retain moist soil at or near the soil surface throughout the growing season for at least 2-3 years and probably longer. Relative availability of water, equipment, and manpower may dictate which, if any, approach is feasible in a specific location. 6. Use of stem cuttings of phreatophytes should not be completely discounted even though only poor survival was attained during this study as it has been used economically and successfully in many locations. However, monitoring of groundwater stability, and soil quality and structure is recommended for a year in advance of planting. �7. The groundwater level in riparian zones along small stable streams (e.g., below Bonny Reservoir) could be raised by using small metal check dams placed at frequent intervals along the stream. This approach combined with soil disturbance may be sufficient to enhance natural regeneration of phreatophytes in many locations. 8. Seepage areas often occur in locations between irrigation canals and the river and some have relatively stable water conditions. Some sites may have considerable value for establishing cottonwoods, willows, or other phreatophytes using either stem cuttings or rooted seedlings. 9. Planting of rooted seedlings of selected tree, shrub, and vine species obtained from nursery stock appears more practical than use of stem cuttings. However, neither method could be used over extensive areas. Individual seedlings could be planted in desired patterns to enhance vertical and horizontal diversity using polypropylene mulch placed 1-2 m around individual plants to reduce weed competition and retain soil moisture. Plastic tree protectors to reduce browsing by deer, rabbits and other wildlife is essential. 10. Livestock grazing of riparian sites, especially when intensive and/or during the grow~ng season is highly negative to both cottonwood and willow reproduction, and to the herbaceous understory. Grazing should be excluded when possible and, if permitted, should be restricted to light , intensity during the late winter dormant season. 11. Plowing and discing of disturbance tillage strips should be used, especially in ecotones between shrubs and/or trees and perennial herbaceous cover, to enhance the food base for wildlife and to increase horizontal diversity. These strips should be tilled early each spring or in alternate years' to retain seed producing annuals. Fireguards, surrounding grass-dominated meadows can serve as additional disturbance tillage strips. 12. Guidelines for conducting controlled burns within riverbottoms are: Burns should be limited to small tracts « 2 ha) in higher elevation sites dominated by grasses'. Efforts should be made to protect trees, and shrubs, especially those that have high quality as winter r'estLng sites for wildlife <, DiiiCing of fireguards, backfiring; use of wetting agents, and other methods should be used to reduce fire intensity.' . 13. Beaver (Castor castor) are <?- natural part of riparian ecosystems and should not be eliminated. However, trapping is suggested when needed to prevent overpopulation and extensive damage to streamside cottonwoods. �331 DYNAMICS OF COTTONWOOD REGENERATION Warren D. Snyder INTRODUCTION The cottonwood-willow riparian habitat is the most important ecosystem for wildlife in Colorado based on diversity (species richness) and density (Beidleman 1978, Crouch 1961, Fitzgerald 1978, Graul 1977, Hoover and Wills 1984, Schrupp 1978). Statewide, the cottonwood riparian ecosystem is used by 283 species of vertebrate wildlife, by far the highest number of any ecosystem. Most of these species are present because of food-cover associations, not because of associated aquatic environments. Fitzgerald (1978) listed 129 species of birds, 28 mammals, and 20 herpetofauna as occurring in the South Platte cottonwood riparian zone based on data from his field work and that of Beidleman (1954) and Crouch (1961) in northeastern Colorado. The cottonwood-willow riparian ecosystem is especially important because it comprises such a small portLon of Colorado, especially within the High Plains of eastern Colorado. There it is estimated to occupy <0.5% of the total land area. Statewide, cottonwoods occupy <0.2% of the total land area (Bottorff 1974). Riparian habitats have also been centers of man's act~vities since Colorado was first settled and before. Highways, railroads, towns and cities, reservoirs and irrigation diversions, stream channeling, diking, gravel mining, and farming are among the impacts that usually degrade this habitat for wildlife. In addition, intensive livestock grazing is a major impact especially along nearly all of the smaller streams or intermittant washes. Along with these problems man has introduced invader species including tamarisk (Tamarix spp.), Russian-olive (Elaeagnus angustifolia), and Canada thistle (Cirsium arvense). This study was initiated in the early 1980's within the Division's Nongame Research group which was later incorporated into the Avian Research group. The primary basis for the study was concern that cottonwoods were not being sustained by natural reproduction and were deteriorating. This concern has not been restricted to Colorado, but is widespread especially in western United States. Numerous studies, conferences, and symposia (Johnson and Jones 1977, Swanson 1979, Nelson and Peek 1981, Simons et ale 1981, Schmidt 1983, Hoar and Erwin 1985, Johnson et ale 1985) attest to this concern. P. N. OBJECTIVES 1. Quantify changes in stand density and changes in area of riparian cover types over a recent time span approximating 30 years within the South Platte, Arkansas, and South Republican (below Bonny Reservoir) rivers in eastern Colorado, the Rio Grande River in the San Luis Valley, and the lower Colorado River in western Colorado. Analyses and evaluations will be based on aerial photo interpretation by the Colorado State Forest Service (under contract). �332 2. Document conditions conductive to natural regeneration and survival of plains cottonwoods (Populus sargentii) and willows, including site characteristics and frequency of occurrence, in streamside riparian habitats of the lower South Platte River in northeastern Colorado following high water conditions in 1983 (potentially also in 1984). 3. Test methods for establishing woody vegetation within streamside riparian zones where natural propagation cannot be expected. These include: (a) using stem cuttings of dormant specimens and planting them so their bases extend into the groundwater as a method for propagation of selected tree, shrub, and vine species for use where natural propagation no longer occurs, (b) create exposed bare ground sites using tillage or scarification for natural establishment of plains cottonwoods, peachleaf willow, or other woody vegetation present near the site. ACKNOWLEDGMENTS Preliminary study planning and design were conducted by G. C. Miller, former Wildlife Researcher, under the supervision of W. D. Graul, former project leader' within the Nongame Research group. They coordinated contract agreements with T. Owens and D. Teska, of the Colorado State Forest Service; the latter personnel did the actua~ photo interpretive inventory. In 1983 the study was transferred to the Avian Research Section of the Colorado Division of Wildlife under the direction of C. E. Braun, Project Leader. D. C. Bowden, consulting statistician, assisted. in preliminary design as well as conducting major analysis of findings. Personnel assisting in field data collection included J. F. Corey, M. Stanley, J. Palic, and A. Palic. METHODS Pertinent literature was reviewed concerning facets of riparian ecology including past inventory methods for cottonwoods, stem cutting techniques, and ecology of natural cottonwood - willow regeneration. Inventory A stratified random sampling method was developed in consultation with D. C. Bowden, contract statistician, to inventory downstream portions of several major rivers in Colorado. River mile (1.61 km) sampling units were located with an electronic planimeter in mid channel. Strata were selected and river mile sample units were randomly selected for inventory. Sampled rivers included the South Platte, Rio Grande, Colorado and Arkansas (Fig. 1). In addition a non-random segment of the South Fork of the Republican River was selected. Personnel skilled in aerial photography interpretive methods for inventorying stands of trees within the mapping section of the Colorado State Forest Service (CSFS) were contracted to conduct inventories. They obtained aerial photos from the earliest and most recent intervals they could find for use in assessment of changes among cover types and stands of trees. A 25-year separation between the earliest and latest photography was the minimum interval used. A vegetation classification system (Table 1) was developed in discussions between cnow and CSFS personnel. Characteristics of inventoried segments, sampling intensity, and other criteria varied among rivers (Table 2). �r: -------------- _ -1 GRANO JUNCTION ~-,~ I I I , ,,I I I • I J • •• •/ LAKE PUEBLO CITY I --- I COLORADO ........ I ~---------------------Fig. 1. Distribution of cottonwoods (Map courtesy of C. C. lJiller) --------...:.---- PLAINS RIO COTTONWOOD GRANDE NARROWLEAF -- COTTONWOOD COTTONWOOD J I ---------- by species along major river drainages in Colorado. w w w �334 Table 1. Classification system and codes used by the Colorado State Forest Service during inventory of vegetation types along rivers in Colorado. Cover or vegetation type Symbol Hay meadow/emergents Grassland Shrubs communities (upland, mesic, & wetland) Conifers Agricultural Developed River Unvegetated/sandbar = River Channel Lakes, ponds, reservoirs Cottonwoods (plus other tree species) Size-age classes <1.5 dm - dbh «6") 1.5 - 3.8 dm - dbh (6-15") 4.1 - 7.6 dm - dbh (16-30") >7.6 dm - dbh (>30") Crown density classes (canopy cover) 10-34% 35-55% >55% H GR S J AG D R NV R C I" 2 3 4 A B C aMinimum polygon size for all types was 1 ha. Table 2. Sampling intervals (X years), linear river distance (km), sampling intensity (ha, %), number of sampling strata, upper and lower elevations (m), and upper and lower K daily stream flows (m3s) along segments of 4 rivers in Colorado. Variable Arkansas ~ sampling intervala Distance sampledb Area sampled % sampled Strata Elevations - upper lower Stream flows - upper lower 31.2 241.3 3,433 15 3 1,372 1,036 26.6 3.2 River South Platte Colorado 36.4 263.7 5,443 18 4 1,372 1,066 21.8 15.1 25.0 167.3 1,826 20 4 1,829 1,372 100.5 175.5 Rio Grande 36.8 117.4 3,263 27 3 2,438 2,286 25.3 7.0 aAerial photos were primarily from 1940's (early) and 1978-80 (late) except along the Colorado River where early photos were from 1951-57. bLinear study area distance was measured at the center of the river channel. �335 Field trips by CSFS personnel were used to verify deliniation accuracy on all rivers. It was not possible to distinguish between tamarisk and willow species from available photography. Therefore, areas and points within them were selected from aerial photos and among strata for field sampling. Shrubs within 43 0.004-ha (1/100 ac) circular plots were identified as to species, and the percentage of shrubs by species was calculated. Analysis of variance and paired t tests were used in preliminary analyses of changes within cover type and trees stands over time and among strata. D. C. Bowden, consulting statistician, was contracted to conduct a more detailed analysis of cottonwood age class and canopy cover changes among strata. Natural Reproduction Conditions conducive to natural regeneration and survival of plains cottonwoods along the South Platte River were documented using 2 primary approaches. The locations of new cottonwood seedlings were marked along several accessible areas of the lower South Platte including both grazed and ungrazed sites following 1983~84 high stream flows. Two dozen individual intensive transects, up to 25-m in length were randomly selected among these locations. The transect ends were marked with steel posts and, during sampling, a tape marked at l-m intervals was placed between the posts. Within high density seedling stands, seedlings were counted within a 0.09-m2 sampling frame at l-m intervals along 1 side of the tape during early fall intervals from 1984 through 1987. Within low density seedling stands, the tape bisected a l-m wid~ continuous sample strip sampled at l-m intervals. As seedling stands evolved from high to low density status, the procedure was adjusted from 0.09-m2 to l-m2 samples. Stands were monitored to document the primary factors (desiccation, grazing, etc.) responsible for cottonwood attrition. The relationship between seedling survival and water table level, and water level fluctuations was assessed. Site characteristic information recorded at each transect included: (1) location (edge of gravel bar, side channel, main channel, etc.), (2) overstory, (3) competition from other vegetation, and (4) livestock grazing season of use and intensity. Frequency of cottonwood seedling occurrence was sampled at 30 random transects along the lower South Platte River in northeastern Colorado. Odometer readings were used to randomly select starting points along accessible sides of the river; random selection among compass bearings toward the river was used to determine direction. Locations where the distance from the river channel to the edge of the flood plain was <100 m were excluded. A l-m2 plot was used as the sampling unit and seedlings of cottonwoods and other tree species were counted as present or absent at 5-pace intervals along the transect. Sampling was conducted in early fall each year from 1984 through 1987. Year of seedling establishment was recorded if it could be determined. The dominant cover type present in all sample frames was recorded where tree seedlings were absent. Categories included sand bar, bare ground, litter, low shrubs, tall shrubs, vines, annual grass, perennial grass, annual forbs, and perennial forbs. �336 Woody Cover Propagation Procedures One or more locations within several sites were selected to receive stem cutting treatments. These included the South Platte Wildlife Management Area (SPWMA) (Tamarack), Duck Creek, and Sedgwick Bar properties along the South Platte River, The South Republican Wildlife Area (SRWA) below Bonny Reservoir along the South Fork of the Republican, and the Rocky Ford, Escalante, Purgatoire, and John Martin properties along the lower Arkansas River. After preliminary extensive trials, more intensive procedures were concentrated at SPWMA and SRWA. Groundwater levels and fluctuations were monitored at stem cutting locations using perforated 3.B-cm diameter plastic pipe inserted into the water table. Monitoring of water levels and fluctuations occurred primarily during successive growing seasons. Stem cuttings of cottonwoods, several species of willows, and other selected trees, shrubs, and vines were primarily obtained from wild stock within CDOW properties. Most stem cuttings were treated with a root stimulator (indole 3 - butyric acid). Preparation of planting stock, pruning, treatment, and placement generally followed procedures described by McCluskey et al~ (1984) and Swenson and Mullins (1985). Holes were dug for placement using a tractor-mounted power augers approximately 1.0 and 2.5 dm in diameter to depths approximating 1-1.2 m. Hand augers were used to penetrate to greater depths and, in some instances, a rod or pipe was driven into the soil to facilitate placement of cuttings. Methods used in 1986 included inserting a l-dm diameter plastic pipe into a hole dug with a power auger. The pipe was repeatedly pushed down as sand was extracted with a hand-operated ·sand auger permitting stem cutting placement to greater depths into the water table. Within some 1986 test sites, competing vegetation was removed by tillage and, after stem cutting placement, the site was overlaid with 4 mil. black plastic sheeting and wood chips. Stream Profile Sampling A map of the South Republican River in Yuma County was gridded at 400-m intervals to bbtain stream profile data. Random selection among intervals permitted selection of 12 points respectively below and above Bonny Reservoir. A 30-m tape marked at l-m intervals was placed level across the stream and vertical distances from the tape to the stream were obtained at l-m intervals. A transit and rod were used to obtain stream profile samples along the Arkansas River. Samples were obtained at accessible lbcations within previously selected inventory random units. Twelve locations above and 12 locations below John Martin Reservoir were sampled. A line was placed across the channel facilitating sampling at I or 2-m intervals. Channel width and vertical distance from river level to cottonwood tree bases were also obtained. Evaluation of Controlled Burns Evaluation of controlled burning within the South Platte riverbottom included late May and early September post-burn inventory of survival and regrowth status of trees and shrubs within burns conducted by CDOW management personnel �337 in early spring 1985 and 1986. Visual obstruction sampling (Robel et ale 1970) of herbaceous vegetation during pre and post-treatment late winter intervals was conducted along random transects within the 1986 burns and proximal controls. Herbaceous Vegetation Sampling within Disturbance Tillage Sampling of dominant herbaceous vegetation within fireguards and other disturbance tillage strips was conducted during September 1986 and 1987 within CDOW properties along the South Platte River. Locations of starting points for transects were randomly selected. The dominant vegetation within 1-2 m2 samples was recorded at 5 or 10 step intervals. Scientific names of plants follow Harrington (1954), Scott and Wasser (1980), and McGregor (1986). RESULTS AND DISCUSSION Inventory Results of the inventory conducted by personnel of the Colorado State Forest Service were reported earlier (Snyder 1984, 1986a, b; 1987). These data were further examined in attempts to understand probable-reasons for changes in quantity, density, and ages of stands of cottonwoods and their potential impacts on wildlife. Arkansas River.--Historically, the Arkansas Riyer Valley through much of western Kansas and eastern Colorado was treeless based on accounts by early explorers (Lindauer 1970, Nadler 1978, Athearn 1985). One large open stand of mature cottonwoods, called Big Timbers, existed near the present town of Lamar extending upstream for 6.4 to 48 km (reports conflict). No underbrush was present. Trees also reportedly lined the riverbanks in the vicinity of Pueblo, but many other stretches along the river were treeless. Several small canals diverting irrigation water from the Arkansas were constructed along the valley in the 1870's and 1880's and priorities of 8 ditches were adjudicated by 1884 along the lower Arkansas (Nadler 1978). The Sante Fe Railroad was completed in 1873 which set the stage for development of towns along the valley. During the 1880's and 1890's larger and longer canals were constructed, funded by eastern investors, and by 1895 20 major diversions had been developed. Development of irrigation and intensified farming along the valley continued in the early 1900's including construction of several off-stream reservoirs and use of natural basins for water storage. The amount of land under irrigation in those early years is unknown but Nadler (1978) documented that irrigated lands along the valley declined from 220,725 ha in 1936 to 197,600 ha in 1969. John Martin Reservoir, a large main-stream impoundment, was constructed in the late 1940's and began storage of water in 1949. However, due to downstream irrigation demands the reservoir remained dry through much of its existence. Pueblo Reservoir, constructed above the city of Pueblo, began storage of water in 1974 (Fig. 2). Ten diversions of West Slope waters occurred between 1880 and 1972 (Nadler 1978) with the most recent diverting a considerable amount of water from the Frying Pan River. �SOUTHEASTERN COLORADO N en ct: en z -c ~ ------.--------------- J NEW MEXICO , OKLAHOMA ~ C') r" Fig 2. stratas Arkansas (I). River drainage showing major reservoirs, portion inventoried (__ ), and �339 The Arkansas River drains about 15,534 km2 of mountainous area and is fed by numerous tributaries draining an additional 50,227 km2 between Pueblo and the Kansas-Colorado border (Fig. 2). The semi-arid shortgrass plains of southeastern Colorado are dominated by shallow soils and marginal vegetation. Thus, runoff from infrequent but occasionally heavy rains can cause flooding. Major floods have been infrequent, occurring in 1909, 1921, and 1965. Lesser floods or high run-offs occurred in 1903, 1914, 1915, 1942, 1947, and 1956. Mean annual discharge peaked in the early 1920's, early 1940's, to a lesser extent in the mid 1960's (Nadler 1978), and increased flows were again noted in the 1980's. Nadler (1978) provided data showing the bankfull width of the Arkansas River in the LaJunta-Las Animas area (Fig. 2) averaged about 175 m in 1870 increasing to 215 m in 1892-1926. By 1952 the channel averaged 46 m in width and only 32 m by 1977 or about 15% of the 1892-1926 mean width. An average width of 87 m was obtained for 12 sites in 1986 between Pueblo and Las Animas. In western to central Prowers County the river was historically wider averaging 215 m (1870), 355 m (1892), but only 145 m by 1926. Mean width in 1952 was 54 m and shrinkage continued as the river was impacted by John Martin Reservoir. Stream widths downstream from the reservoir in 1986 averaged 35 m, about 10% of that in the late 1800's. The Arkansas River was originally much wider (375 m in 1870, 335 m in 1892) in eastern Prowers County but had declined to 137 m by 1952. Dramatic shrinkage has continued because widths ranged from 24 to 50 m in that vicinity in 1986. Apparently the river flowed intermittantly prior to being impacted b~ man, drying completely in late summer and fall. Continuous flows began in the early 1900's because of return flows and seepage from irrigation. Nadler (1978) reported that the river originally was nearly straight with numerous small shallow channels and islands. With dewatering and altered stream flows, the river has become more meandering and confined to a single channel. Arkansas River flow (K m3/s) (U.S. Geol. Survey \~ater-Data Reports) averaged annually since 1914 show flows have fluctuated considerably among decades (Table 3). At Nepesta, east of Pueblo, recent average flow rates were less than in earlier decades with no interval of clear decline evident. At Lamar, reduction in flow in recent decades is more evident. Stream flow rates along the Arkansas diminish rapidly due to irrigation diversions along the river. At Avondale, east of Pueblo mean daily flows averaged over 14 years from 1965 to 1984 were 26.6 m3/s. Downstream near Las Animas, flows averaged 6.0 m3/s. Flows from John Martin Reservoir were slightly higher at 7.3 m3/s due to supplemental water from the Purgatorie River. At Lamar, daily flows averaged 3.2 m3/s compared to 5.7 m3/s near the Kansas-Colorado Border. Photo interpretive inventory sampled 3,433 ha within 36.7 km of a 250-km segment of the lower Arkansas River beginning about 36.7 km east of Pueblo and continuing to the Kansas-Colorado Border (Fig. 2). John Martin Reservoir (25 km) was excluded from sampling. The 1st of 3 strata contained 50 km, the 2nd and 3rd, 100 km each above and below John Martin Reservoir respectively (Fig. 2). Average sampling intensity was 14.7%. Original year aerial photos ranged from 1940 to 1955 and all recent photos were from 1980 yielding average time intervals of 25.7, 29.1, and 38.6 years, respectively for the upper, middle and lower strata. The overall mean interval was 31.2 years. �340 Table 3. Mean stream flow rates (m3/s) at 10-year intervals at measuring stations along the Arkansas, South Platte, Colorado, and Rio Grande rivers, Colorado. Arkansas Period Nepesta 1891-1900 1901-10 1911-20 1921-30 1931-40 1941-50 1951-60 1961-70 1971-80 1981-86 22.8 21.0 14.1 23.6 17.4 17.9 17.9 28.0 South Platte Lamar Kersey 7.1 11. 7 2.9 8.7 1.6 3.1 1.4 4.2 19.0 21.3 25.5 13.2 23.0 16.0 21.3 30.5 36.6 Julesburg 14.8 15.4 14.3 7.6 15.1 9.2 12.3 17.9 28.2 Colorado Glenwood Springs Cameo 84.4 74.5 92.5 65.9 63.6 58.0 Rio Grande Del Norte Alamosa 23.5 26.9 31.1 31.1 21.1 28.6 19.9 23.0 22.7 24.1 108.6 118.6 105.6 101.7 116.7 146.6 12.9 11.3 3.9 7.7 3.1 4.6 5.9 Overall changes in vegetation types and comparati·ve changes among strata were variable (Tables 4, 5). Hay meadow and grassland both declined at modest rates from early to recent sampling. The majority of the decline in hay meadows occurred within the middle stratum whereas grassland declined markedly in the lower stratum apparently giving way to cropland (Table 5). Agriculture or cropland increased dramatically (p < 0.005), especially in the lower stratum: Developed land, while comprising only 1-2% of the total, nearly doubled in quantity (p < 0.005). Table 4. Proportions and changes (ha/l.6l km) in vegetation types over a 3l.2-year interval along the Arkansas River, southeastern Colorado (1940-80). Vegetation Cottonwood Shrub Hay meadow Grassland Agriculture Developed River channel Standing water *p < 0.05. haiL 61 km Early Recent 29.42 39.07 34.17 24.98 1.54 1.66 25.12 0.10 20.41 47.32 29.40 22.52 19.38 3.30 12.62 1.05 Change ha % .t. -9.01 8.25 -4.77 -2.46 17.84 1.64 -12.50 0.95 -30.6 21.1 -14.0 -9.8 1,155.2 98.8 -49.8 950.0 2.90* 1.78 0.87 0.66 3.00* 3.00* 5.45 1.10 �341 Table 5. Average change of vegetation types (ha/l.6l km) among strata from early to recent intervals along the Arkansas River, southeastern Colorado. Vegetation Upper Stratum Middle Cottonwood Shrub Hay meadow Grassland Agriculture Developed River channel -9.00 4.39 -2.17 4.85 7.37 0.93 -9.77 -7.30 9.94 -9.89 -1.28 11. 77 3.39 -6.86 Lower F value -11.20 9.37 0.41 -10.23 34.60 0 -22.10 0.17 0.12 0.29 1.28 1. 73 0.87 6.41* *p < 0.05. River and vegetated (sandbars and mudflats) were combined for analysis since differing water levels between photo periods could bias results.· There was a 50% decline in river channel (p < 0.05) with the greatest decline occurring in the lower stratum below John Martin Reservoir (P<0.05) (Tables 4,5). These dramatic declines in channel width supports Nadler's (1978) findings and are readily discernable from inspection of aerial photos. Tamarisk has invaded much of the former channel within the lower stratum. A channel width averaging 35 m in the lower stratum differed (P <0.05) markedly from the 87 m mean width upstream from John Martin Reservoir~ Another concern was whether the river channel had deepened appreciably below John Martin Reservoir in relation to upstream areas. However, dense tamarisk stands along the river made it extremely difficult to obtain stream profile samples comparing upstream and downstream sites based on a standard 100-m width because transit-level sightings could not be conducted through the dense growth. Stream profile samples that varied in length, but averaged 98.2 m wide, were conducted on 12 upstream samples. In comparison, 9 downstream sites averaged 97.2 m wide. The average of the 8 deepest profile measurements from upstream samples was 17.8 dm whereas the average downstream profile was 23.1 dm (t = 2.93, 19 df, P < 0.05). Thus, there was evidence that the downstream channel had deepened. Personnel of the U.S. Corp of Engineers at John Martin Reservoir provided historic degradation data from 13 sample points lying between the Reservoir and Lamar. Graphic presentations of those data showed that at several points within a few km of the Reservoir, a narrow channel had deepened about 0.61 to 1.2 m beyond the original channel depth. However, further downstream near Lamar deposition had occurred filling much of the original channel leaving a narrow channel at about the same elevation as that present in 1943-44 when monitoring was initiated. Nadler (1978) presented data showing the same type of deposition had occurred further downstream. A major flood occurred immediately below John Martin Reservoir and downstream in 1965 when the spillway was opened and Muddy Creek, Clay Creek, Butte Creek, and other streams rampaged and undoubtedly deposited considerable sediment in sites containing tamarisk and other vegetation. �342 The vertical distance from river water level to the base of nearby cottonwoods was measured within sampling sites at several locations upstream and below John Martin Reservoir. Trees were not present for sampling at several downstream locations so sample size was reduced. The average vertical distance from water to tree base among upstream samples was 13.6 dm contrasted to an average of 20.4 dm among downstream samples (~< 0.05). Occupancy by cottonwood stands decreased 31% in the inventoried areas or approximately l%/year with a mean loss of 9.0 ha/l.6l km during the interval (p = 0.002, Table 4). The decrease was greatest in the lower stratum (p = 0:029, Table 5). Initially, cottonwoods occupied 18.9% of the inventoried riverbottom but declined to 13.1% occurrence during the recent sample. Occurrence of trees was strongly skewed toward the upper stratum (Table 6, Fig. 2) where they declined from 52.3 to 43.3 ha/l.6l km during the sample interval. Within the 2nd stratum, trees initially occupied only about one-half (24.1 ha/l.6l km) the area of occurrence as in the upper stratum, and declined at a much greater rate to 16.8 ha/l.6l km. In the lower stratum below John Martin Reservoir, trees initially occupied only about one-third of the area and by the recent inventory had declined to 5.4 ha/l.6l km (Table 6). It is probable that cottonwoods were not highly abundant within the lower stratum because of the rapid invasion of tamarisk in the early 1900's. Table 6. Area occupied by cottonwoods (X ha/l.6l km) from early to recent sampling intervals among strata and rivers, Colorado. Stratum 1 2 3 4 Arkansas Early Recent South Platte Early Recent 52.3 24.1 16.6 63.7 47.2 68.2 72.3 43.3 16.8 5.4 50.6 56.9 55.0 72.5 Colorado Early Recent 3.6 38.6 12.6 14.9 0.3 31.6 11.9 13.1 Rio Grande Early Recent 23.7 48.1 8.0 29.6 51.8 6.9 aStratum 1 is the stratum located furthest upstream on all rivers. Decreases in occupancy were noted in all size-age classes of cottonwoods from early to recent inventories, but were most significant in the 4.1-7.7 (-58%, P = 0.05) and >7.6 (-15%, P = 0.0004) classes (Table 7, Fig. 3). Size-age class distributions of the cottonwood-dominated stands were similar during the 2 inventories with a preponderance in the 1.5-3.8 and 4.1-7.6 size-age classes (Fig. 3). Open stands of cottonwoods dominated the forested areas in both inventories (Fig. 3) and did not change substantially in occupied area (-3%, ~ = 0.36). However, stands of intermediate canopy closure decreased by 59% in area (p = 0.003) and closed stands declined by 51% (~ = 0.50). �343 AGE CLASS (DM -DBHI 14 12 ~ > 7.6 (LZ;] 4.1-7.6 ~ 1.5-3.8 10 ••• . to < 1.5 8 6 4 2 o EARLY RECENT 10-34' EARLY RECENT 36-55 EARLY RECENT <55 CANOPY COVER Fig. 3,~ Early to recent changes/sample (X ha/1.6l classes (dm-dbh) and canopy cover (%) of cottonwood lower Arkansas River, southeastern Colorado. km) by size-age stands along the �344 Table 7. Changes by age-class and canopy cover (~ ha/l.6l km) of cottonwood stands over a 3l.2-year intervala along the Arkansas River, southeastern Colorado, 1940-80. Age-class (dm-dbh) <1.5 Canopy cover (%) 10-34 35-55 >55 Subtotal 1.5-3.8 10-34 35-55 >55 Subtotal 4.1-7.6 10-34 35-55 >55 Subtotal > 7.6 10-34 35-55 >55 Subtotal All ages Total 10-34 35-55 >55 Recent Earl~ ha 0.76 0.51 0.61 1.87 4.32 5.04 1.18 10.55 8.60 4.71 1.76 15.08 0.70 1.15 0.09 1.93 14.38 11.41 3.64 29.43 % ha 6.4 1.73 0 0.08 1.81 35.9 5.23 2.21 0.74 8.18 51.2 6.71 2.35 0.78 9.84 6.6 0.23 0.18 0.17 0.58 48.8 38.8 12.4 13.90 4.74 1.77 20.41 % Change %!age ha class 8.9 0.97 -0.51 -0.53 -0.06 0.7 40.1 0.91 -2.83 -0.44 -2.37 -26.3 48.2 -1.89 -2.36 -0.98 -5.24 -58.2 2.9 -0.47 -0.97 0.08 -1.35 -14.9 -0.48 -6.67 -1.87 -9.02 -3.3 -58.5 -51.4 30.6 68.1 23.2 8.7 aThe time interval ranged from 25 to 40 years with a mean of 31.18 years. Thus, there is a pronounced trend toward fewer hectares of trees and those remaining are becoming more open. The size-age structure of the stands has not changed markedly over time except that large trees (>7.6 dm dbh) have declined dramatically. Trees in the intermediate age classes comprised between 85 and 90% of the total composition during both inventories (Table 7). Lindauer (1970) listed peachleaf willow (Salix amygdaloides) and Russian-olive as additional species within cottonwood communities along the Arkansas River. Cottonwoods averaged 64 trees/ha compared with 7.4 peachleaf willow and 12.4 Russian olive/ha. He noted the latter species was most prevelant in upstream samples. R. J. Niedrach reported observing tamarisk (salt cedar) near Lamar in 1913 and noted its subsequent expansion in 1921 and 1923 (Lindauer 1970). Lindauer's interviews with ranchers indicated floods in 1921 and 1937 promoted its �345 expansion. It has continued to move upstream and Bittenger and Stringham (1963) noted increased phreatophyte growth between Las Animas and La Junta between 1940 and 1961, to which they attributed diminished stream flows. Shrubs primarily tamarisk, were the most abundant vegetation type present during both sampling intervals and increased 21% over the 31.2 years (p < 0.10). Lindauer (1970) reported that tamarisk stands within the lower-stratum were older and the species was progressively spreading upstream. Their continued spread was evident during this study as greatest increases were noted in the middle stratum (Table 5). Colorado State Forest Service personnel could not distinguish tamarisk from other shrubs along the Arkansas using photo interpretation methods. However, .their 1985 sampling indicated that tamarisk dominated most sites containing shrubs and represented over 75% of the shrubs and small trees present. Sandbar willow (Salix interior) and coyote willow (S. exigua), not readily discernable as separate species, were the second most important shrubs with low densities of indigo bush amorpha (Amorpha fruticosa) and a few other species. Young Russian-olive may have been inventoried as shrubs. It is obvious that the quantity and quality of riparian habitats for wildlife along the lower Arkansas River have diminished rapidly in recent decades. The data point to 2 major causes: (1) dewatering and flood control, and (2) the invasion of tamarisk •. Either by itself would be devastating, but the 2 in combination provide little hope that highly struct~red wildlife habitats can be retained or economically restored. Lindauer (1970, 1983) documented the spread of tamarisk upstream along the Arkansas River since the early 1900's and noted its apparent capacity to raise the levels of soluable salts in the topsoil. As a consequence, only salt tolerant herbaceous species such as inland saltgrass (Distichlis stricta) and fireweed summer cypress (Kochia scoparia) grow in association with tamarisk. Diverse associations of food producing annuals were not present. Assuming tamarisk was not dominant along the Arkansas, it is probable there would be limited opportunity for cottonwoods and willows to become established. This is because several tributaries drain extensive areas of southeastern Colorado (Fig. 2) and infrequent heavy rains, like those in 1965, can still cause flooding. Flooding and high stream flows would probably not be sustained for long enough intervals to assure survival of new stands of seedlings. Pueblo Reservoir, John Martin Reservoir, and major irrigation diversions in combination make the potential of flooding from snowmelt a rarity. Flooding was not significant along the Arkansas in 1983 when major flooding was noted along other Colorado rivers. Flooding from snowmelt is usually sustained over longer intervals, a situation that is more conducive to cottonwood regeneration. Elimination of flooding and stabilization of reduced stream flows along the lower Arkansas has dramatically reduced opportunities for cottonwoods and willows to sustain themselves. Elimination of cottonwoods removes the upper portions of vertical structure essential to numerous passerines, raptors, fox squirrels (Sciurus niger), and other wildlife. Without trees, there is a reduction in both species richness and density or abundance. Elimination of �346 flooding results in plant growth on sandbars and mud flats and the elimination of edge, weed strips, and diversity essential to wildlife. As areas become established in continuous stands of perennial grasses, the weed-seed food base is eliminated which a majority of wildlife need to survive. This is especially critical in bottomland areas bordering semi-arid rangelands. Although tamarisk may provide nesting and protective cover for many wildlife species, densities will be severely restricted if a food base is lacking. An illustration of declining habitat quality on wildlife abundance is reflected by the apparent impact of John Martin Reservoir, and river dewatering, on downstream northern bobwhites (Colinus virginianus). Censuses of a downstream route in 1971~76 indicated on average of 1.0 calling males/station whereas above John Martin, the number of calling males/station during the same interval was 3.4 (Snyder 1978). The Arkansas still retains much more of a resemblance of a river in upstream areas. However, Pueblo Reservoir and its stream stabilizing impact will gradually degrade wildlife habitats downstream to John Martin Reservoir. The most recent aerial photos of the Arkansas River reviewed were taken in 1980. Field activities and observations from 1984 to 1987 indicated degradation of cottonwoods and the riverbottom continued, possibly at an accelerated rate. Cutting of old trees for firewood was frequently observed and has increased in recent years because of higher costs for natural gas, petroleum products, and electricity accompanied by a depressed farm economy. Loss of these large trees severely impacts several species of woodpeckers, fox squirrels, raptors, passerines, and other wildlife. Frequent fires in the bottomlands were apparent. In spring 1986, a large fire escaped in high winds to burn several kilometers of riverbottom west of Las' Animas killing hundreds of cottonwoods. Livestock grazing has long been a common use along the Arkansas and will continue into the future. South Platte River.--Reviews of historical records (Nadler 1978, Williams 1978, Crouch 1979, Smith 1980) indicate that cottonwoods were rare along the South Platte River on the plains of northeastern Colorado and western Nebraska during the 1800's. One grove of trees near the present town of Orchard was prominant and several lone trees in the Iliff-Proctor area were noted in early reports. Historically, the river was relatively straight, shallow, and broad with many small channels and islands. Mean widths from Greeley to Brush (Fig. 4) averaged over 400 m in 1867 (Nadler 1978) and Williams (1978) showed channel widths over 600 m in the Sedgwick-Julesburg vicinity. By 1952, Nadler noted that the mean width had decreased 86% to about 60 m. By 1977 the channel had widened to 90 m and observations following high water in 1983-84 indicated substantial width increases in downstream locations. Settlement of the valley, beginning in the 1860's brought ditch irrigation and diversion of river water. The senior adjudicated water right was dated 1859 (Nadler 1978). The Union Pacific railroad was completed in 1867 and stimulated establishment of numerous towns along the valley. By 1882 and elaborate network of canals had been completed along the valley and by 1885 residents were reportedly beginning to notice return flows to the river. In 1899, 287,226 ha were reportedly irrigated and irrigated land peaked at �WYOMING ---1---------- NEBRASKA I ------ .. _ ••• --~BURGI -------------- II lOr.. ' I I z m N CD :n » (/) A » ~~ ~A AIVt tI I NORTHEASTERN COLORADO I I I~ Z ('-1> i~ I(/) «'«'r w .p.. --.J Fig. 4. South Platte drainage in nortlleastern Colorado, inventoried portion (--), and strata (I). �348 500,372 ha in 1930 along with construction of numerous off-stream storage reservoirs (Smith 1980). Historians report that the South Platte became dry in late summer and fall prior to development of irrigation; the latter subsequently stimulated seepage and year-long flows. Intermittant flows may have been a factor in the lack of cottonwoods but it is more probable that browsing by buffalo (Bison bison) and periodic range fires were the primary deterrents to cottonwoods. Removal of buffalo, elimination of wildfires, and reduction of peak spring stream flows due to irrigation apparently allowed establishment of plains cottonwoods and peachleaf willows within areas that formerly were river channel. Major increases in numbers of plains cottonwoods apparently began after 1900 and the species peaked in density in the 1940's and 1950's (Sedgwick and Knopf 1987~). In more recent years, major diversion projects including the Colorado-Big Thompson (Adams tunnel - 1947) and 14 other diversions (Nadler 1978) brought water into the South Platte drainage to provide supplemental water for irrigation and urban development. Currently, Chatfield Reservoir, just south of Denver (Fig. 4), is the only major flood control storage reservoir on the South Platte after it leaves the mountains. The proposed Two Forks Reservoir, a few miles upstream from Chatfield Reservoir, may be constructed to provide additional water to the Denver metropolitan area in the near future. Other reservoirs occur along the South Platte in upstream mountainous areas and along a few of its tributaries. The South Platte drains about 59,910 km2 in central and northe?stern Colorado. The Narrows Reservoir, long proposed for construction on the South Platte a few miles west of Fort Horgan, has been only marginally funded and its future remains uncertain. If constructed it would potentially curtail high water and flood threats from the Cache la Poudre, Big Thompson, and Saint Vrain rivers and Crow Creek (Fig. 4). The intermittant Kiowa/Bijou Creek drainage would be the only major tributaries as Beaver Creek near Brush and Pawnee Creek near Sterling seldom flow significant water. A proposed diversion of South Platte excess flow, primarily winter flows, into the Frenchman Creek drainage is currently undergoing feasibility studies. Mean flow rates averaged among 83 years shows the rate of flow at Julesburg (Fig. 4) was 15.1 m3/s or 69% of the 21.8 m3/s rate upstream at Kersey. This reduction in flow was much less than that along the Arkansas River. However, considerable water is diverted from the South Platte and its tributaries before it gets to Kersey. Stream flows averaged among decades show no marked reduction from 1903 to the present with some increase in the last decade. Much of this flow retention has been attributed to seepage and return flows from irrigation and urban areas. An inventory of 5,445 ha of riverbottom within 36.4 km (17.7%) of 263.7 km was completed using stratified random sampling of linear mile (1.61 km) segments. Early year aerial photos were mostly from 1941 (a few from 1948-49) and recent photos were from 1979 (2 from 1978) yielding an average time span of 36.4 years. The 29 samples were distributed within 4 strata (Fig. 4). Hay meadow was the most abundant cover type (39.4%) along the South Platte during the early inventory (Table 8). However, it declined by 45% (p < 0.05). In contrast, grassland and cropland increased from early to recent intervals �349 (p < 0.05, Table 8). Developed land and standing water, representing small p;rcentages of the total, increased from early to recent inventories. River channel (river + unvegetated) also increased markedly (~<0.05) in contrast to dramatic declines along the Arkansas (Table 9). Grassland and river channel showed considerable variance among strata (p < 0.05) whereas most other vegetation types were fairly uniform among strata (Table 10). Table 8. Changes (ha/l.6l km) in vegetation types over a 36-year interval (1940's - late 1970's) along the South Platte River, northeastern Colorado. Vegetation Cottonwood Shrub Hay meadow Grassland Agriculture Developed River channel Standing water hall. 61 km Recent Earll 62.29 16.92 73.83 4.43 16.99 1.09 11.85 0.19 56.51 13.41 40.59 13.55 36.84 3.59 20.52 2.79 Change ha -5.78 -3.51 -33.24 9.12 19.85 2.50 8.67 2.60 % -9.3 -20.7 -45.0 205.9 116.8 229.4 73.2 1368.4 t 1.76* 1.67 5.33** 3.23** 3.87** 1.88* 5.45** 2.47** *p < 0.10 **p < 0.05 Hectares of shrub cover declined slightly (p > 0.10) along the South Platte. Ground inspections of the aerial photo classifications revealed some interpretive overlap between shrub and hay meadow classifications. However, both types declined from early to recent intervals (Table 8). Primary shrub species along the South Platte River include sandbar and coyote willow, a few diamond willow (Salix lutea), and large patches of western snowberry (Symphorocarpos occidentalis). Indigo bush, woods rose (Rosa woodsii), poison ivy (Toxicodendron radicans), and currant (Ribes aureum) were present in lower densities. Only a few scattered Tamarisk occur along the South Platte. Stands of plains cottonwoods containing peachleaf willow and occasionally a few other tree species occupied 33% of the inventoried bottomland during the early interval and 30% during the recent interval at which time they were the dominant cover (Table 8). The decline of 9.3% (p = 0.046) was not distributed among all strata as the 2nd from upper stratum showed an increase in area occupied over time whereas the 1st and 3rd strata showed considerable decline (P < 0.005, Table 10). However, hectares of cottonwoods were comparable among all strata (Table 6). Young trees «1.5 dm-dbh) showed an early to recent decline of 33.6% (p = 0.087) in occupied area but initially comprised only 10% of the composition (Table 11). Trees in the 1.5-3.8 and 4.1-7.6 dm age classes in combination comprised 88 to 90% of the composition during both inventories and showed more modest declines in occupied area (Table 11, Fig. 5). Large trees (>7.6 dm) occupied only 2-3% of the area and may have increased slightly (P > 0.10). �w V1 o Table 9. Early to recent changes/sample (ha/l.6 km) of vegetation types or land use occurring along Colorado's low-elevation riparian zones.a Vegetation/use Cottonwood stands Shrubs Hay meadow Grassland Agriculture/development River channel Arkansas Early Recent 29.4 39.1 34.2 25.0 3.2 25.1 20.4 47.3 29.4 22.5 22.7 12.6 River South Platte Early Recent 62.3 16.9 73.8 4.4 18.1 11.9 as tanding water and other minor covers were excluded. 56.6 13.4 40.6 13.5 40.4 20.5 Colorado Early Recent 18.0 15.3 23.6 5.0 9.8 15.0 14.9 16.3 18.0 . 6.5 13.3 14.2 Rio Grande Early Recent 28.0 10.6 110.4 1.4 1.3 9.9 30.6 8.0 87.7 5.0 23.4 6.2 �351 Table 10. Average change of vegetation types (ha/l.61 km) among strata from early to recent intervals along the South Platte River, northeastern Colorado. Stratum Vegetation 1 2 3 4 F value Cottonwood Shrub Hay meadow Grassland Agriculture Developed River channel Standing water -13.1 +0.5 -29.0 +11.0 +12.6 +3.5 +12.9 +3.3 +9.7 -6.9 -38.9 +9.7 +15.4 +4.9 +0.3 +5.3 -13.2 -2.6 -34.1 +1.9 +37.9 +0.1 +8.5 +0.8 +0.2 -8.6 -31.1 +20.3 +3.2 +1.5 +14.2 +0.4 3.92* 0.84 0.11 3.36 1.09 0.57 5.01* 0.85 *p < 0.05 Table 11. Changes (ha/l.61 km) by age class and canopy cover of cottonwood stands over a 36-year interval (1940's to late 1970's along the South Platte River, northeastern Colorado. Age-class (dm-dbh) <1.5 Canopy cover (%) 10-34 35-55 >55 Subtotal 1.5-3.8 10-34 35-55 >55 Subtotal 4.1-7.6 10-34 35-55 >55 Subtotal >7.6 10-34 35-55 >55 Subtotal All ages Total 10-34 35-55 > 55 Earl:t:Interval % ha 2.60 1. 74 1.89 6.23 13.79 8.45 3.93 26.17 17.28 6.02 5.26 28.56 0.67 0.27 0.38 1.32 34.34 16.48 11.46 62.28 Recent Interval % ha 10.0 2.15 1.23 0.76 4.14 42.0 10.17 8.34 4.52 23.03 45.9 15.05 9.12 3.47 27.64 2.1 1.04 0.41 0.24 1.69 55.1 26.5 18.4 28.41 19.10 8.99 56.50 Change %!age ha class 7.3 -0.45 -0.51 -1.13 -2.09 -33.6 40.8 -3.62 -0.11 +0.59 -3.14 -12.0 48.9 -2.23 +3.10 -1. 79 -0.92 -3.2 3.0 +0.37 +0.14 -0.14 +0.37 +28.2 -5.93 +2.62 -2.47 -5.78 -17.3 +15.9 -21.6 -9.3 50.3 33.8 15.9 �352 AGE CLASS (DM -DBHI 35- c1.5 30 25 ~ 1.5-3.8 0 4.1-7.6 ~ .7.6 ~ .•.. U) .•... ';a J: 20 15 - Z « w ~ 10 5 o EARLY RECENT EARLY REceIT EARLY. REcafT 10-34 35-55 >55 CANOPY COVER Fig. 5. Early to recent changes/ sample (E ha/1. 61 km) by s i ze+age clas~es (dm-dbh) and canopy cover (%) of cottonwood stands along the lower South Platte River, northeastern Colorado. �353 Open stands (10-34% canopy cover) of cottonwoods that initially comprised 55% of the total declined 17% (P = 0.02), whereas those of intermediate (35-55%) density increased by 16% (P-= 0.016) to occupy 34% of the area (Table 11, Fig. 5). Trees in dense stands-(>55%) initially comprised 18% of the area occupied by trees but declined 22% (~ = 0.42, Fig. 5). Based on these data there was a modest overall decline in cottonwoods along the South Platte but the decline was not distributed uniformly among strata (Table 10). However, stands of trees continued to be the dominant cover type within the inventoried bottomlands and averaged over 56 ha/l.6l km. The proportion of young replacement trees declined during the 36-year interval providing evidence of insufficient natural reproduction and survival needed to sustain former densities. Few large trees were present during either inventory. Open stands dominated during both inventories but there was no pronounced trend toward opening of stands. Within open stands of trees along the South Platte River, Lindauer (1983) found that peachleaf willow comprised 40% of the 95 trees/ha density. In closed stands, peachleaf willow comprised 25% of the composition which averaged 120 trees/ha. Data from Christy (1973) indicated cottonwoods averaged about 85 trees/ha and other minor species, green ash, American elm (Ulmus americana), and boxelder (Acer negundo) averaged about 7-8 trees/ha within sites classified as cottonwood communities. Thus, cottonwoods comprised over 90% of the tree composition. In comparison within mixed communities, plains cottonwoods averaged 78% and peachleaf willow represented 22% of the composition with respective densities of'about 85 and 24 trees/ha or a combined tree density of 109/ha. Sedgwick and Knopf (1986) recorded much iower cottonwood densities averaging 23.9 trees/ha during their recent studies on the South Platte Wildlife Management Area. Crouch (1979) reported 94.6 and 117.9 trees/ha respectively on grazed and ungrazed study areas in the same vicinity in 1961. By 1978 densities had declined to 43.5 and 79.1 trees/ha yielding declines of 53% on the grazed site and 33% on the ungrazed site during the l7-year interval. This rate of decline was dramatically greater than recorded over the 36-year extensive photo interpretive evaluation. Sedgwick and Knopf (1987b) considered the cottonwoods within their study to be in a decadent age class representative of declining tree stands. They noted that 43% of the trees were 35-55 years old. The South Platte riverbottom with respect to status and trends of cottonwoods and wildlife is in relatively good condition when contrasted with that of the Arkansas River. Cottonwoods occupy between 2 and 3 times greater area/l.6l km along the South Platte (Table 9) and its 0.75 to 1.0-km wide flood plain is markedly wider than that of the Arkansas or any other Colorado River. Opportunity for occasional to frequent flooding still occurs along the lower South Platte and tamarisk, although present in small numbers, does not appear to be a major threat. Observations following the severe winter of 1983-84 indicated extreme cold weather may be a major deterent to tamarisk in northeastern Colorado. Extensive winter kill of top growth was noted along both the Arkansas and South Platte. Although root sprouting and regrowth has been observed, tamarisk along the South Platte remains in poor vigor. Sedgwick and Knopf (1987b) reported that the last phase·of significant overstory development in-their South Platte study area occurred between 1929 �354 and 1949. After that cottonwood reproduction fell below levels necessary to sustain an equilibrium population. Data from Crouch (1978) and the extensive inventory provide additional evidence of decline although the latter data indicate the decline has been moderate (O.3%/year) rather than rapid. A few young trees were established following 1965 and 1973 floods but regeneration was not enough to offset reductions in stands of trees, their aging or the progression toward reduced canopy cover. Flooding and sustained high stream flows in 1983-84 markedly stimulated natural reproduction of cottonwoods, other trees, shrubs, and grape vines. Cottonwood establishment was widely distributed and subsequent survival through 1987 was relatively good providing hope that stands of cottonwoods along the lower South Platte received a substantial boost in the form of replacement trees. It is doubtful that these new seedlings will reverse long-term declines of cottonwoods along the South Platte but should slow the rate of decline. In some locations, e.g.,·the Dodd Bridge Wildlife Area, the river channel seems to have deepened too much in relation to the adjacent flood plain and little natural regeneration has occurred. In other locations, spring to early fall livestock grazing has eliminated all seedlings from both recent and previous decades so that cottonwood stands are decadent and deteriorating along with their understory shrubs. Data from Crouch (1978) indicates stands of cottonwoods deteriorated more rapidly within a grazed site than within an ungrazed site along the South Platte. Historically, the South Platte River was several times wider and straighter than it is today and was nearly treeless (Nadler 1978, Williams 1978). Devel.opment;of extensive LrrLgat.Lon along the Front Range and downstream apparently reduced peak stream flows and altered flows so that the river flowed yearlong. As a consequence, willows and cottonwoods invaded former river channel, possibly during drought years (Nadler 1978), so that the river channel was further restricted. Sites containing willows, cottonwoods, and herbaceous vegetation became collectors of sediments during subsequent floods so that these sites, over time gradually became more elevated from repeated flood deposits (Nadler 1978, u.S. Dep. Inter. 1981). Continued elevation increases first made these sites too mesic for willows to establish, and as elevations continued to increase they became too high for cottonwoods to become established upon. Some of these sites still contain gradually opening stands of mature to overmature cottonwoods, whereas other locations have given way to areas dominated by perennial grasses. This process along the South Platte in northeastern Colorado appears to be slow because the river still retains numerous usually dry side channels, and occasional floods, as in 1983-84, scour new locations and create new sites suitable for willow and cottonwood germination. Cottonwoods establish in narrow linear bands as evident from inspection of aerial photos. Although some regeneration of willows and cottonwoods continues, Sedgwick and Knopf (1988) suggest it is probably not at a rate sufficient to sustain these species over time. As the flood plain became too elevated to sustain cottonwood regeneration along the Platte River in Nebraska, other more shade tolerant mesic trees invade (U.S. Dep. Inter. 1981). These include eastern red cedar (Juniperus virginiana), box elder, mulberry (Morus spp.), and green ash. Sedgwick and Knopf (1988) presented data supporting the hypothesis that green ash would be the most probable mesophytic tree to achieve overstory dominance following decline of cottonwoods along the So'ut h Platte. As documented in this study, green ash shows evidence of rapid expansion apparently promoted in flooding �355 in 1983-84. Russian-olive also has shown evidence of limited invasion. However, flooding in 1983 killed some seedlings and it is anticipated natural expansion will primarily occur along flood plain edges and adjacent meadows. Green ash appears tolerant of poor soils which have high alkalinity and soluable salts and are low in nutrients, it is shade tolerant invading under overstory cottonwoods, it is not readily killed by flooding, and it is not highly succeptable to livestock browsing. Green ash along the South Platte are relatively young trees and few if any old trees exist within the flood plain. Personnel of the CDOW and DSFS have and are continuing to attempt to establish other tree species including oak (Quercus spp.), and walnuts (Juglans spp.) with no success to date. Poor soil conditions are believed a major hinderance to survival of these species. Colorado River.--The Colorado River and its numerous tributaries drain an area of about 46,196 km2 of primarily mountainous terrain. Tributaries including the San Juan, Delores, White, and Yampa, which join the Colorado River further downstream, also drain extensive areas of Western Colorado. The Colorado River and its tributaries are confined within mountain valleys through most of their passage in Colorado and frequently are subjected to steep gradients. Narrow valleys severely limit the extent of riparian habitat along the streams. Not until the Colorado River joins with the Gunnison River at Grand Junction does it enter into a relatively broad valley and with reduced stream gradient. Like the rivers in Eastern Colorado, the Colorado River valley was the the major focal point of early travel and settlement. Thus, the narrow riparian zone along the stream was severely impacted by railroads, highways, and settlements which crowded the riverbanks in many locations. In intermediate to higher elevations livestock grazing, mining, and lumbering were early developments impacting the river and its tributaries. Numerous small diversions, flood-irrigated hay meadows, and livestock have impacted bottom lands along the narrow valleys, through the decades. At lower elevations fruit orchards, alfalfa, and other crops are grown but opportunities for irrigation are much less in Western Colorado than in the eastern part of the state. Therefore, river flows, which were much greater than those for East Slope rivers, were not severely reduced by early settlement and development of local irrigation. Alteration and reduction of stream flows and water quality along the Colorado have primarily occurred within the last 50 years with the development of high mountain water diversions to the East Slope and the construction of numerous reservoirs to retain snowmelt for delayed release during peak irrigation downstream. However, most of the water entering the Colorado drainage exits the state to be impounded and used for urban areas and irrigation in the arid Southwest. Narrowleaf cottonwood (Populus angustifolia) is the dominant species along the Colorado River and its tributaries (Fig. 1), generally occurring between 1,524 and 2,438 m elevation. A less abundant species, lanceleaf cottonwood (P. acuminata), occurs within a slightly broader elevational range (Harrington 1954). The Rio Grande cottonwood (P. wis1izeni), the same as the plains cottonwood in general appearance, is confined to lower elevations in western Colorado. �356 Just when tamarisk was introduced to the lower Colorado River valley is unknown but it has spread extensively within the lower Colorado and Gunnison valleys. Russian-olive is also abundant in many locations along these West Slope rivers at lower elevations. No pre-settlement data were found in the literature but it is assumed that cottonwoods and willows were present all along the Colorado River probably to a greater extent than they are today. Photos taken within the Grand Valley in 1882-1904 show scattered cottonwoods along the Colorado. Flooding along the Colorado River is primarily a consequence of snow melt and above normal spring temperatures. Major reservoirs on the Gunnison River have severely restricted peak flows in recent years. Major peak flows or flooding were recorded frequently in the early 1900's along the Colorado River (1909, 1912, 1914, 1917, 1918, 1921, 1928, 1935, and 1938). Peak flows in more recent decades were recorded in 1957, 1983, and 1984. The Colorado River was sampled along 167.3 km from Glenwood Springs to the Utah-Colorado Border (Fig. 6). Within that distance it dropped about 457 m in elevation from 1,829 to 1,372 m. In recent years stream flows at the upper part of the sampled area averaged 3-4 times those of other inventoried rivers and, because of its confluence with the Gunnison River at Grand Junction, it left the state with a flow approximately 75% greater than at the upper end of the sampled area (Tables 2, 3). Inventory of 1,826 ha of riverbottom within 33.8 km (20.2% sample) was accomplished within 4 strata (Fig. 6). Original year aerial photos were from 1951-57 and recent photos were from 1978-80 yielding an average time- span of 25.1 years, considerably less than that along other sampled rivers. Hay meadow, the most abundant vegetation classified, declined by 23.7% (p < 0.10) during the sampled interval (Table 12) with the primary decrease occurring in the lower 4th stratum (Table 13). Grassland represented only 5.7% of the sampled area during early sampling but increased 31% in all but the 2nd stratum. Agriculture or cropland represented about 10% of the sampled area and changed little over the 25 years. The amount of developed land and standing water were insignificant in early year samples but increased to represent 10% of the area during the recent inventory (Table 12). Overall, the river channel changed little but variance among strata was evident with channel widening in the 2 upstream strata but apparently narrowing downstream (Table 13). Shrubs occupied about 17-18% of the sampled riverbotton and showed evidence of increasing slightly, primarily in the 2nd stratum. Only about 3-4 ha/l.6l km of shrubs were present/sample within the upper stratum. However, shrubs occupied over 20 ha/l.6l km within the 2nd and 3rd strata and about 15 ha/l.6l km within the lower stratum. Tamarisk was the dominant shrub, especially at lower elevations where dense stands similar to those along the Arkansas River were common. Field inspection revealed that Russian-olive was common in some locations and may be increasing. Willows were common especially within the upper 3 strata. �,------------- I ---------------- NORTHWESTERN -- WYOMING ---- _-- - ----- _----- _---- COLORADO N ~.t.- <;, I -c I::J V-l VI -.j Colorado River drainage in Western Colorado with i~entoried Fig. 6. portion (--I and strata (II. �358 Table 12. Changes (ha/l.6l km) in vegetation types over a 25-year interval along the Colorado River, western Colorado (1955-80). Vegetation Cottonwood Shrub Hay meadow Grassland Juniper Agriculture Developed River channel Standing water ha/l. 61 km Early Recent 17.98 15.28 23.58 4.99 0.14 8.78 1.06 14.93 0.21 Change % t -17.5 +6.5 -23.7 +30.9 -35.7 -7.4 +388.7 -5.2 +1661. 9 1.74* 0.93 1.94* 0.71 ha 14.84 16.28 18.00 6.53 0.09 8.13 5.18 14.15 3.70 -3.14 +1.00 -5.58 +1.54 -0.05 -0.65 +4.12 -0.78 +3.49 0.55 3.38** 0.87 3.00** *p < 0.10. **P<0.05. Table 13. Average change of vegetation types (ha/l.6l km) among strata from early to recent intervals (1950's-1980) along the Colorado River, western Colorado. Stratum Vegetation Cottonwood Shrub Hay meadow Grassland Agriculture Developed River channel Standing water 1 -3.4 -0.9 -6.0 +3.8 0 +4.9 +1.2 -0.1 2 -7.0 +3.1 -3.1 -7.2 -0.3 +5.0 +3.0 +0.9 3 -0.7 +0.2 +2.4 +3.1 -0.9 +0.5 -3.6 0 4 F value -1.9 +1.1 -14.8 +5.1 -1.1 +5.5 -2.7 +8.6 0.52 0.48 2.72 1.90 0 0.93 4.73* 5.75* *p < 0.05. Stands of cottonwoods occupied about 21% of the sampled Colorado River bottomlands during the early inventory and about 17.5% during the recent sampling (Table 12). The average rate of decline was 3.04 ha/l.6l km (p = 0.08). Within the upper stratum, extending about 30 km downstream fromGlenwood Springs (Fig. 6), stands of trees were especially sparse averaging only 3.6 ha/l.6l km during the early inventory. However, they declined over 90% to average only 0.26 km/l.6l km during the recent sample (p = 0.02). The greatest decline/sample unit was 7 ha within the 2nd stratum. -However, trees occupied 38.6 and 31.6 ha/l.6l km respectively in early and recent �359 inventories. Hectares/sample unit were much higher than within other downstream strata where cottonwoods averaged about 12-15 ha/l.6l km and declined at more modest rates. The age structure of cottonwoods along the Colorado appeared to be that of a healthy stand with 58% of the trees occupying the 2 younger age classes with a fair proportion of large trees (Table 14, Fig. 7). The percentage of young trees «1.5 dm-dbh) declined almost 50% during the 25-year interval but still remained nearly twice that within East Slope samples. Numbers of large old trees (>7.6 dm) also declined dramatically. Table 14. Changes by age-class and canopy cover (ha/l.6l km) of cottonwood stands over a 25-year interval (1950's to 1980) along the Colorado River, western Colorado. Age-class (dm-dbh) <1.5 Canopy cover CO 10-34 35-55 >55 Subtotal 1.5-3.8 10-34 35-55 >55 Subtotal 4.1-7.6 10-34 35-55 >55 Subtotal >7.6 10-34 35-55 >55 Subtotal All ages Total 10-34 35-55 >55 Recent Earll ha 1.89 1.34 1.49 4.72 2.10 1.94 1. 71 5.75 2.78 1.86 1.67 6.31 0.06 0.57 0.58 1.21 6.83 5.71 5.45 17.99 % ha 26.3 1.26 0.54 0.63 2.43 32.0 3.32 1.65 1.36 6.33 35.0 2.95 1.00 1.93 5.88 6.7 0.07 0.09 0.04 0.20 38.0 31. 7 30.3 7.60 3.28 3.96 14.84 % Change %!age ha class 16.4 -0.63 -0.80 -0.86 -2.29 -48.5 42.7 +1.22 -0.29 -0.35 +0.58 +10.1 39.6 +0.17 -0.86 +0.26 -0.43 -6.8 1.3 +0.01 -0.48 -0.54 -1.01 -83.4 +0.77 -2.43 -1.49 -3.15 -17.5 51.2 22.1 26.7 Changes in stand densities were more discouraging than changes in age structure. Within the early inventory samples, stands were well distributed among all canopy cover classes. However, by the recent interval open stands had increased 11% whereas intermediate stands (35-55% canopy cover) had declined 42% and dense stands had decreased 27%. Thus, the same general pattern as observed on the Arkansas and South Platte rivers was evident. �360 AGE CLASS (OM -OBH] 8 tsSJ >7.6 7 E ~ ••••• to ••••• <, m ~ 4.1-7.6 ~ 1.5-3.8 <1.5 6 5 4 ..c: Z 3 :E 2 « .. w 1 0 EARLY RECENT 10-34 EARLY RECENT 35-55 EARLY RECENT >55 CANOPY COVER Fig. 7'. Early to recent changes/sample (25: ha/l.6l km) by size-age class (dm-dbh) and canopy cover (%) of cottonwood stands along the lower Colorado River, western Colorado. �361 Extensive flooding occurred along the Colorado River in 1983-84 and extreme stream flows were recorded. Scouring of previously deposited sediments was noted in some locations widening the river channel and promoting a rocky streambed. However, natural regeneration of cottonwoods was noted in numerous locations along the river. Subsequent observations indicated loss of numerous trees to beaver (Castor canadensis). Since the riparian zone is narrow in many locations, the problem of loss to beaver is enhanced. Wildlife use of the riverbottom is severaly restricted in most areas by the relative narrow valley and restricted quantity and quality of phreatophytes. A heavily traveled interstate highway and railroad crowd the riverbank in many locations and numerous developments have been completed in recent decades. Rio Grande River.--The Rio Grande River drains approximately 19,194 km2 of which 7,612 km2 lies within a closed basin. River water is nearly exclusively from snow melt primarily from the San Juan Range with lesser amounts from the Sangre de Cristo Range to the east. The river enters into the western part of the S~n Luis Valley, a high elevation (2,286-2,438 m) relatively flat plateau and travels through a farming area for approximately 100 km before entering a canyon and exiting into New Mexico (Fig. 8). Drainages in the northern part of the San Luis Valley do not extend to the Rio Grande but enter the ground to subirrigate parts of the valley. Subirrigation, irrigation water from the Rio Grande and its tributaries, and artesian wells prompted rapid settlement and development of irrigation in this high, cold desert basin. Spanish immigrants explored and attempted to settle the area in prior centuries but their first permanent settlement of San Luis was established in 1851 - the earliest known settlement in Colorado. The first recorded water rights date from 1852 known as the San Luis People's Ditch at San Luis on Culebra Creek (Athearn 1985). Several more Spanish settlements which used irrigation water for farming developed along the Rio Grande and its tributaries in subsequent years. Several large diversion canals were constricted between 1880 and 1890 and 99% of the irrigation canals in the valley had been constructed by 1890 (Tipton 1939). Without reservoir storage to regulate water flow, farmers attempted to increase storage of water in the ground resulting in creation of numerous seeps and alkaline flats and the loss of farmground. LaJara Reservoir construction began in 1910 and by 1938, 235,000 ac. ft. of reservoir storage existed within the upper Rio Grande Drainage (E. Dumpf, Pers. Commun.). Additional major regulation of stream flow from high mountain reservoirs was initiated in 1949 and later when the San Luis Valley Project was implemented by the u.S. Bureau of Reclamation. More recently, development of center pivot irrigation has prompted rapid conversion to sprinkler irrigation from wells causing major declines in groundwater levels. However, due to 3 years of above normal stream flows, groundwater levels along the Rio Grande are above 1977 levels (E. Dumpf, Pers. Commun.). Rio Grande stream flows based on 10-year averages since 1890 have not shown major evidence of declining in recent decades at Del Norte in the west-central portion of the San Luis Valley (Fig. 8). There, mean daily flows have averaged about 25.3 m3/s. Partial regulation of flows following construction of upstream reservoirs in recent decades has undoubtedly altered flow patterns with increased volumes into late summer for irrigation. �w oN SAN LUIS VALLEY, DEL COLORADO NORTE N ---- -- -- ---- ------ -- -- -- -- -- ------ -- --- -- ----- ----NEW MEXICO Fig. 8. e I) . Rio Grande River drainaBe in southern Colorado witll inventoried portion (--) and strata �363 Further downstream at Alamosa, mea~ daily flows average only about 30% of those at Del Norte. Mean flow rates averaged over 11.2 m3/s between 1913 and 1930 at Alamosa, but declined to less than one-half that volume since 1930. Flooding or excessive flows along the Rio Grande were evident in 1907, 1911, 1917, 1921, 1927, 1941, 1977, 1979, 1981, and 1983 to 1987. Information concerning status of cottonwoods along the River in the 1800's and early 1900's could not be found. The Rio Grande River was sampled along 177.4 km from South Fork downstream to below Alamosa (Fig. 8). Sampling intensity averaged 27.4% including 20 1.6l-km units distributed among 3 strata containing 3,263 ha of riverbottom. Initial aerial photos were mostly from 1941 (1 from 1944) and recent photo data ranged from 1973 to 1983 yielding an average time span of 36.8 years. Hay meadow-dominated vegetation types (Table 15) and comprised 68 and 54% of the sampled area respectively during early and recent samples. The decline of 20.6% (p <0.05) occurred primarily within the 2 upper strata (Table 16). Grassland was a minor vegetation type initially ~ut increased substantially (P <0.10), primarily within the 2 upper strata. During the early sampling interval only 2 of 20 samples contained cropland but this number increased to 9 of 20 within the recent inventory and the percentage changed from 0.1 to 13.4% (p < 0.05). Developed l?nd and standing water were minor components in both sample intervals. River channel showed a 36.7% decrease in occupied area (p < 0.05) that was distributed among all strata (Table 15). Shrub cover declined by 25% from early to recent intervals (p < 0.10). Willows were the primary shrub component within the riverbottom and comprised only a minor percentage of the inventoried area during both sampling periods (Table 15). Table 15. Changes (ha/l.6l km) in vegetation types over a 36.8 year interval (1941 to 1973-83) along the Rio Grande River, San Luis Valley, Colorado. hall. 61 km Vegetation Cottonwood Shrub Hay meadow Grassland Agriculture Developed River channel Standing water *p < 0.10. **P < 0.05. Change Early Recent ha % t 27.99 10.54 110.43 1.40 0.18 1.12 9.91 1.63 30.56 7.93 87.68 5.00 21.75 1.63 6.27 1.87 +2.57 -2.61 -22.75 +3.60 +21.57 +0.51 -3.64 +0.24 9.2 24.8 20.6 257.1 1.46 2.31* 2.79** 2.42* 3.05** 1.43 6.86** 0.56 'V 45.5 36.7 14.7 �364 Table 16. Average change of vegetation types (ha/l.6l km) among strata from early to recent intervals along the Rio Grande River, San Luis Valley, Colorado. Vegetation 1 Stratum 2 4 F value Cottonwood Shrub Hay meadow Grassland Agriculture Developed River channel Standing water +5.9 -2.8 -44.4 +12.6 +30.6 -0.3 -2.1 +0.5 +3.6 -4.1 -30.4 +1.2 +32.5 +1.6 -5.3 +0.9 -1.0 -0.8 +1.4 0 +2.7 -0.2 -2.9 -0.7 0.85 0.53 3.18 2.89 3.31 2.92 4.48* 1.22 *P<0.05. Cottonwoods were moderately abundant within the upper stratum of the Rio Grande increasing from 23.7 to 29.6 ha/l.6l km (Table 6) a~d were more common within the middle stratum where they increased from 48.1 to 51.8 ha/l.61 km. However, within the lower stratum in areas downstream from Alamosa they were completely absent along several sample units. The average occurrence in the lower stratum was only 8.0 and 6.9 ha/l.61 km respectively from early to recent inventories (Table 6). The overall increase was modest, 9.2% (P = 0.12). Initially cottonwoods occupied 17.1% of the sampled area and this percentage increased to 18.8% by the 2nd interval. Young trees «1.5 dm) comprised 10.4% of the composition during both sampling intervals and increased 9.3% in occupied area (Table 17). Trees of intermediate size (1.5-3.8 dm) declined over the interval (P = 0.13) glvlng way to the next larger (4.1-7.6 dm) age class that increased 27.2% (p = 0.16) (Fig. 9). This latter group dominated among age classes during both-intervals (Table 17). Older trees (>7.6 dm) represented only about 3% of the total composition during both intervals but showed limited evidence of increasing in occupied area (p = 0.56). Open stands (10-34% canopy cover) initially occupied only 31% of the timbered area and declined to 25% by the 2nd inventory (P = 0.25). In contrast, stands of intermediate density increased from 33 to 40% occurrence (31%, P = 0.015). Dense stands increased more modestly (9%, P = 0.49) representing 35% of the total composition during both sampling intervals. Based on all 3 variables, area occupied, age structure, and stand density, cottonwoods along the Rio Grande River were by far the most healthy among the 4 inventoried rivers. However, Colorado Division of Wildlife personnel in the San Luis Valley report almost no recent reproduction of narrowleaf cottonwoods along the Rio Grande or along its tributaries (Alamosa, La Jara, and Conejos rivers) in the 1980's (E. Dumpf, pers. commun.). Floods along the Rio Grande in recent years have been short term and apparently not adequate to sustain survival of seedlings. Beaver were also reported to be causing considerable loss of cottonwoods within the Rio Grande drainage. �365 Changes (ha/l.6l km) of cottonwood stands by age class and canopy Table 17. cover over a 36.8-year interval (1941 to 1973-83) along the Rio Grande River, San Luis Valley, Colorado. Age-class (dm-dbh) <1.5 Canopy cover (%) 10-34 35-55 >55 Subtotal 1.5-3.8 10-34 35-55 >55 Subtotal 4.1-7.6 10-34 35-55 >55 Subtotal > 7.6 10-34 35-55 >55 Subtotal All ages Total 10-34 35-55 > 55 Early ha 1.16 0.71 1.04 2.91 3.53 3.67 2.50 9.70 3.93 4.83 5.81 14.57 0.17 0.15 0.50 0.82 8.79 9.36 9.85 28.00 Recent % ha 10.4 1.79 0.86 0.53 3.18 34.7 3.54 2.54 1.83 7.91 52.0 2.16 8.62 7.76 18.54 2.9 0.11 0.21 0.64 0.96 31.4 33.4 35.2 7.60 12.23 10.76 30.59 % Change %/age ha class 10.4 +0.63 +0.15 -0.51 +0.27 25.8 +0.01 -1.13 -0.67 -1. 79· -18.5 60.7 -1.77 +3.79 +1.95 +3.97 +27.3 3.1 -0.06 +0.06 +0.14 +0.14 +17.1 -1.19 +2.87 +0.91 +2.59 -14 +31 +9 +9.3 24.8 40.0 35.2 +9.3 Narrowleaf cottonwoods were by far the dominant species present although a few other species may occur. No Russian-olive were noted during 1985 inspections but the species may be present in certain locations. Narrowleaf cottonwoods were the primary trees planted at rural and urban residences in the San Luis Valley indicating the species was well adapted to the climate if supplemental water was available. South Fork of the Republican River.--The South Fork of the Republican River is a relatively small stream that drains about 4,725 km2 of east-central Colorado extending approximately 115 km west from Kansas (Fig. 10). It flows water for about 48 km through northern Kit Carson and southern Yuma counties before passing into Kansas and Nebraska where it joins with the Arikaree, North Fork of the Republican, and Frenchman Creek. Above where it flows in Colorado, a few intermittant seeps and springs can be found. A small recreation reservoir which fluctuates depending on intermittant flows from upstream rains, is on the Republican near Flagler and is managed by the Colorado Division of Wildlife. �366 AGE CLASS 10M -OBHI 12 lS.:Sl >7.6 rza 4.1-7.6 11 ~ 13' 10 9 ~ . to 1.5-3.8 <1.5 8 7 6 5 4 3 2 1 o EARLY RECENT EARLY RECENT 10-34 35-55 EARLY RECENT >55 ' CANOPY COVER Fig. 9. Early to recent changes/sample (~ ha/l.6l km) by size-age class (dm-dbh) and canopy cover (%) of cottonwood stands along the lower Rio Grande River, San Luis Valley, southern Colorado. �367 z m CD :::0 » (J') " » EAST-CENTRAL COLORADO N o IDALIA BONNY ". » z (J') » (J') o Fig. 10. Reservoir BURLINGTON Republican River drainage in eastern Colorado showing and the inventoried 'portion (--) of the South Fork. Bonny �368 Bonny Dam was constructed by the U.S. Department of Interior, Bureau of Reclamation in the late 1940's and early 1950's on the South Republican about 10 km west of Kansas (Fig. 10). The reservoir backs water about 5 km and is used primarily for flood control and recreation. Only limited irrigation is permitted. Therefore, water levels within the reservoir are stable except during infrequent floods. Downstream flows have been regulated by Bonny Reservoir since July 1950. The average discharge over 34 years (1952-85) was 0.55 m3/s near Hale below Bonny Reservoir. The maximum flow was 4.03 m3/s in a sample of 14 recent years (1965-84). In contrast, 2 major floods occurred prior to construction of Bonny Reservoir. In 1935 the greatest flood known occurred with a maximum discharge of 2,884 m3/s at a site a few km above Hale. A discharge of 106.1 m3/s was recorded in May 1947 while the dam was under construction. A continuous small flow is released from Bonny Reservoir whereas in previous years the stream often became dry in late summer and fall. The Colorado Division of Wildlife (CDOW) owns or leases most of the bottomland from Bonny Reservoir downstream to Kansas. Management of CDOW properties has not been intensive and spring-summer livestock grazing was permitted under a 25-year contract that terminated in 1987 on the property containing the Hale Ponds. Most private lands along the stream have been intensively grazed in recent years. Personnel of the Colorado State Forest Service were requested to inventory approximately 4.8 km of riparian habitat west from Kansas. Early photos were from 1961 while recent photos were from 1975. Subsequent contact with the East-Yuma County, U.S. Department of Agriculture Soil Conservation Service District office located aerial photos from 1938. Therefore, changes from 1938 to 1961 and 1975 were examined (Tables 18, 19). These data represent a specific area of the South Republican a few km below Bonny Reservoir and do not represent a random sample of the stream. Table 18. Changes (~ ha/l.6l km) of vegetation types among intervals from 1938, 1961, and 1975 in the Hale Ponds vicinity, South Republican River, Colorado. .Vegetation 1938 1961 1975 Cottonwood Shrub Hay meadow Grassland Agriculture Developed River channel 6.6 45.8 8.5 68.9 20.4 0.3 25.0 42.8 7.3 0.3 96.2 20.1 0 8.8 62.8 12.4 2.3 56~0 37.7 0 4.3 Grassland was the dominant vegetation type in 1938 and 1961 but decreased in the 1975 sample (Table 18). Hay meadow has been minor in occurrence and cropland has recently increased. Inspection of aerial photos clearly revealed �Table 19. Changes (ha/1.61 km) by age class and canopy cover within cottonwood stands in the vicinity of Hale Ponds, South Fork of the Republican River, Colorado among 1938, 1961, and 1975 sample intervals. Age class (dm-dbh) <1.5 Canopy cover (%) 10-34 35-55 >55 Subtotal 1.5-3.8 10-34 35-55 >55 Subtotal 4.1-7.6 10-34 35-55 >55 Subtotal >7.6 10-34 35-55 >55 Subtotal All ages Total 10-34 35-55 >55 1938 ha 1961 % ha 1975 % ha 5.8 0 0 0 0 1938-61 1961-75 1938-75 0 +0.65 -2.49 -1.84 33.7 14.07 7.16 0.69 21.92 34.9 +12.15 +7.51 +19.66 58.5 8.19 13.67 15.48 37.34 59.4 +23.86 +12.28 +36.14 5.7 -0.45 +2.73 +2.28 35.4 37.5 27.1 +13.19 +15.02 +8.00 +4.15 +8.01 +7.87 +17.34 +23.03 +15.87 27.9 1.88 0.62 0 2.49 34.2 7.41 5.80 1.20 14.41 18.2 8.40 8.68 7.98 25.06 0.88 0 0.42 1.30 19.7 0.42 0.42 0 0.85 2.0 0 2.70 0.88 3.58 4.92 0.50 1.18 74.5 7.6 17.9 18.11 15.52 9.18 42.3 36.3 21.4 22.26 23.53 17.05 1.34 0.50 0 1.84 2.26 0 0 2.26 0.44 0 0.76 .1,.20 6.60 42.81 % 62.84 LV 0\ \.0 �370 a marked decrease in river channel (Table 18). The 1938 photos depict a broad shallow sandy stream with extensive unvegetated sandbars. The 1935 flood may have been partially responsible for the broad channel. The Arikaree to the north where u.s. Highway 385 crosses at present typifies the appearance of the South Republican at the sampled location in 1938. No reservoirs occur on the Arikaree River. The impact of Bonny Reservoir and controlled water releases during the last 37 years have greatly altered the channel appearance downstream. Both perennial grasses and trees occupy what formerly was river channel and there is evidence that the narrow stable stream has deepened. Thirty-meter wide stream cross-section profiles obtained below and above Bonny Reservoir in Yuma County indicated the average depth was 6.6 dm below Bonny and 4.5 dm upstream (P <0.01). The 7 deepest measurements/sample averaged 7.7 dm upstream and 13.6 dm downstream (p <0.001) providing additional evidence of downstream channel deepening. Stream width averaged 15.4 m upstream from Bonny Reservoir vs. 5.4 m in downstream locations (P <0.001). Livestock grazing, which tends to erode streambanks and flatten channels was present in all upstream and downstream locations. Shrubs declined extensively from 1938 to 1961 and increased slightly from 1961 to 1975 (Table 18). Field inspections within recent years indicated shrubs (mostly Salix sp.) were a minor vegetation type restricted primarily to a narrow band along the stream channel and were moderately to severely impacted by'livestock. A few tamarisk were scattered at the lower end the CDOW Hale Ponds Property but were not healthy and apparently were not. increasing in number. A few indigo bush were also present. Russian-01ive, while still a minor species, has apparently increased. This species was common above Bonny Reservoir. In contrast to shrubs, the area occupied by plains cottonwoods increased dramatically since 1938 when few trees were present. Based on age class data for 1961 (Table 19), most of this increase apparently occurred prior to completion of Bonny Reservoir in 1950 although some production of young trees has occurred since the dam was built. Floods in 1935 and 1947 may have been responsible for establishment of some stands but the exact ages of the trees is unknown. Several factors could have induced additional cottonwood establishment since completion of Bonny Reservoir in 1950. Bonny Creek, which drains into the inventoried section of the river may have flooded. An irrigation ditch which extended from Bonny Reservoir to Bonny Creek. (Fig. 10) was completed along with the reservoir and irrigation drainage may have stimulated cottonwood establishment. It is also possible that spillway releases from the reservoir may have occurred in the 1950's promoting temporary high water or flooding. Age class comparisons (Table 19) indicated that production of young cottonwoods apparently declined to nearly zero in recent years with a trend toward older trees. Changes in canopy cover did not markedly change from 1961 to 1975. Plains cottonwood was by far the dominant species although a few peachleaf willows, primarily old trees, were present. Livestock grazing along the stream has been conducted primarily during growing seasons for many years and has severely impacted tree and shrub reproduction, and growth of shrubs and herbaceus vegetation. Most sites have poor horizontal and vertical diversity because only 2 covers, trees and grazed grass, are present. �371 Establishment and Survival of Cottonwoods Under Natural Conditions Intensive Transects.--The inventory suggested that cottonwoods and willows along riverbottoms in Colorado have not been reproducing in recent years due either to lack of seed germination, lack of seedling survival, or both. Therefore, seedling survival and the factors influencing it were examined. Flood conditions in 1983 and 1984 along the South Platte River (Fig. 11) caused by melting of extensive snow accumulations in high mountain headwater areas provided ,conditions suited for documenting natural reproduction and monitoring seedling survival. Cottonwoods and willows are both early successional, pioneering plant groups which are intolerant of shade have shallow root systems, normally must grow at sites close to the water table, and which have narrow requirements for natural reproduction including exposed (bare with little competition), wet soil (Christy 1973, Baker and Broadfoot 1977, Fitzgerald 1978, Snyder and Bovee 1983) short-lived seeds are released in conjunction with late spring-early summer peak streamflows. When streams overflow their banks, sediments picked up and carried, or scoured out of the channel are deposited in slow-moving backwaters where alluvial deposits (primarily silt and clay) accumulate (U.S. Dep. Inter. 1981). Persistent high water may kill, stunt, or cover existing vegetation leaving an exposed site for seed g~rmination. Slow declines in water levels along streambanks and side channels may leave an alluvial deposit containing seeds from which narrow, linear stands of new seedling germinate. Based on data from Sedgwick and Knopf (1988) and U.S. Geological Survey Water-Data Reports for Colorado, average South Platte river flows in 1983 and 1984 were the highest recorded in 80 years. River flow in 1983 was especially conducive to establishment and survival of cottonwood and willow seedlings because the river reached and maintained a low flood stage through May and June lowering slowly in July (Fig. 11). This prolonged inundation of perennial herbaceous vegetation retarded its growth and covered many perennials with silt deposits resulting in extensive mudflats. At the same time, seed production of cottonwoods and willows was retarded by below average temperatures through April, May, and June. Their seed disseminated slowly from late June into July as river levels slowly decreased. The result was extensive seedling germination, often in vast dense mats. The South Platte remained higher than normal through the 1983-84 fall-winter interval and increased to low flood stage again in May (Fig. 11) •. Many seedlings established in 1983 were inundated and lost. Flows decreased rapidly in late June 1984 and many seedlings on higher sites were desiccated by dry, late-summer conditions. As a consequence, within 11 preliminary transects established in late summer 1983, seedling survival to early fall 1984 was only 2.2% and 100% mortality occurred within 6 of 11 transects. Seedlings germinating in 1984 were primarily in linear bands along main stream and side channel locations at lower elevations than those established in 1983. Some were probably lost during sporatic flows in fall-winter 1984. Seedlings of peachleaf, coyote, and sandbar willows were generally less abundant than those of cottonwoods but were common following 1983-84 high stream flows. Seedlings of frost grape, green ash, and Russian-olive were less common and primarily were observed following high water in 1983. �372 400 ~ . t\ •• I •\ I \ • .\ . ,;.\ ,·v I' "'" 300 , .II -- \ !._" \ •• \ I \ I J 1 ,. , • • •• • · 1 , r ,. ~, ··· I tI'] <, rt'\ ~ ...•• 200 , • :3 ~• 0 • ~ FLOOD STAGE - ~ ~ tI'] 100 50 ., 11 MAY 21 . 1 11 JUN 21 1 11 JUL 21 1 AUG Fig. 11. Stream flow (m3/s) averaged over 5-day intervals at Julesburg during years within recent decades when the South Platte River flooded. �Table 20. Site Cottonwood seedling density along transects from fall 1984 through 1987, South Platte River, Colorado. Transect II Status Samp1es/ transect Grazed Ungrazed Ungrazed Ungrazed Grazed Grazed Grazed Grazed Grazed Ungrazed Ungrazed Ungrazed Ungrazed Ungrazed Ungrazed Ungrazed 25 9 25 25 25 20 7 17 25 25 16 25 25 25 25 25 1984 Total seedlings 1985 1986 1987 1984 5 14 1 0 8 0 5 90 66 2 65 34 0 47 15 179 2.76 9.11 1.12 0.12 2.16 1.35 5.00 10.12 7.96 0.52 5.75 2.04 10.72 2.68 21.33 Mean density/m 1985 1986 1987 Seedlings established in 1983 Sedgwick Bar Haxtun R&G Tamarack Bravo Tamarack Brush X density 4 2 4 2\07 4W1 4W2 4W3 6W1 6H2 21W 22W 24w 1 2 9Wa 1 69 82 28 3 54 27 35 172 199 13 92 51 268 67 48 29 24 4 1 26 3 24 129 119 8 81 44 0 54 27 226 26 14 1 0 13 0 12 99 85 4 75 37 0 54 33 184 3.31 1.16 2.67 0.16 0.04 1.04 0.15 3.43 7.59 4.76 0.32 5.06 1.76 0 2.16 12.00 9.04 3.21 1.04 1.55 0.04 0 0.52 0 1. 71 5.82 3.40 0.16 4.69 1.48 0 2.16 14.67 7.36 2.79 0.20 1.55 0.04 0 0.32 0 0.71 5.29 2.64 0.08 4.06 1.36 0 1.88 6.72 7.16 2.00 14.40 3.36 8.78 41. 78 86.67 31.33 96.00 148.44 47.11 53.10 0.08 0.04 2.83 0.44 42.22 5.89 31.11 16.44 27.56 14.07 0 0 2.17 0 42.22 4.56 (>.67 4.89 19.11 8.85 0 0 1.83 0 44.00 3.92 1.25 0.92 9.84 6.86 Seedlings established in 1984 Tamarack Bravo Sedgwick Bar Haxtun R&G Tamarack Jones X density 13W1 13W2 3 5a 3a 2Ea 6\073 a 6w4a 3a Grazed Grazed Ungrazed Grazed Ungrazed Ungrazed Grazed Grazed Ungrazed 25 25 18 25 25 17 20 25 25 360 84 158 94 195 48 216 334 106 2 1 51 1 95 9 56 37 62 0 0 39 0 95 7 12 11 43 aSamples obtained with a 0.09-m2 frame and corrected to m2 density. 0 0 33 0 99 6 2 2 22 w ---.J w �374 Plains cottonwood seedling survival within 24 randomly selected intensive transects established in 1984-85 and monitored through 1987 was variable (Table 20). Sixteen of the transects were dominated by 1983 seedlings and 9 by 1984 seedlings. Livestock grazed within 11 of the transects sometime during the year. Once transect had total mortality in 1985 when a sand dam was constructed upstream diverting stream flows to the main channel on the other side of the riverbottom. All cottonwoods died within 4 other transects in 1986 but 80% of the transects sustained at least some survival throughout the 4-year moni~oring interval (Table 20). Table 21. Mean survival rates (%) of cottonwoods seedlings within intensive transects, South Platte River, northeastern Colorado, 1984-87. Year of establishment 1983 1984 Survival rate, Sep to Sep 1984-85 1985-86 1986-87 47.4 19.7 79.1 65.9 77.7 79.4 3-yr survival rate 29.1 10.3 Seedlings established in 1983 within transects sampled in late summer 1984 had already survived through the last half of 1983 and most of 1934. Thus, major attrition had already occurred (2.2% survival). Sedgwick and Knopf (1988) reported <25% survival over the same interval among 300 tagged seedlings within their SP\~ study area. First-year survival among seedlings established in 1984 to early fall 1985 was 19.7%. Subsequent survival among successive years remained much higher, ranging from 47.4 to 79.1% (Table 21). The 3-year survival rate for 1983 seedlings were 29.1% compared to 10.3% for 1984 seedlings. Mean density Cm2) varied greatly among transects and among years (Table 20). Among 1983 seedlings, density declined to 2.0 seedlings/m2 by fall 1987; seedling density in 1987 on 1984 transects was 6.9/m2 (Table 20). Attrition and thinning through browsing, competition for space, and other factors was expected. It was more important that at least a few seedlings/transect survived for several years because chances of survival to maturity increased with each successive year and that seedlings were well distributed along the riverbottom. By fall 1987, surviving 1983 seedlings had obtained considerable growth, however height varied widely among transects apparently because of varying soil quality. In better sites seedlings averaged 2-3 m tall with some to 4 m. In poorer sites heights ranged from 0.5 to 2 m. Most mortality was attributed to desiccation during the first 2 growing seasons. Among 1983 seedlings extensive desiccation occurred in July-August 1984 as stream flows declined. Additional mortality attributed to desiccation was noted in late summer 1985. Broadfoot (1973) reported that soil moisture �375 must be available within 30-46 cm of the soil surface for the first year of seedling growth. Thus, the water table must be close to the soil surface with moist soil extending above it. Finer soils containing silt and clay appeared to retain moisture much better than coarse sands and gravels. This may be a major factor in the apparent better seedling survival and growth in sites containing better soils. Sedgwick and Knopf (1988) found soil moisture at sites containing germinated seedlings was greater than for random sites where no seedlings were present. The alluvial flood plains where cottonwoods and willows occur show little or no soil profile development and contain alternating bands or lenses of sand, silt, clay, and gravel with little humus (Lindauer 1970, Christy 1973). Extensive soil monitoring along the South Platte River showed most soils contained a high proportion of sand, a high pH (8.0), and moderate to high soluable salt content (Christy 1973). Sedgwick and Knopf (1988) found no relationship between cottonwood seedling establishment and 13 edaphic characteristics including soil pH, organic matter, and several soil nutrients (nitrogen, phosphorus, potassium, zinc, etc.). These findings imply that soil type and quality are usually not important considerations in seedling establishment. Moisture, not soil is the key factor in germination and survival. Crouch (1979) cited references stating that cottonwood seed required bare·, moist, mineral soil for germination. Sedgwick and Knopf (1988) found that germination in 1984 occurred under a widely varied conditions including areas where plant cover «1.0 m high) exceeded 95% and litter cover was >75% •. However, their data showed germination was more successful in areas where plant and litter cover was low. Likewise, seedling survival was lower in sites with increased litter, increased tall plant cover, and increased short plant cover. Survival was positively correlated with increased bare ground. They listed several factors including competition as responsible for a negative relationship between cottonwood seedlings and herbaceous cover. Numerous seedlings germinating in 1983 survived to 1987 in relatively thin stands distributed over ungrazed meadows on the SPw~ that contained dense herbaceous vegetation. Herbaceous cover had not been silted over or subjected to prolonged inundation during the 1983 flood. Many seedlings in these sites were subsequently lost to noxious weed control. This illustrates that moist soil is the primary requisite for cotton,mod seedling establishment. Several small sites within a proposed late winter 1986 controlled burn on the Brush Wildlife Area contained cottonwood seedlings established in 1983. One 25-m transect was established and monitored prior to and subsequent to burning. First-year post-burn survival was about 81% indicating that, although most seedlings had been burned to the ground, regrowth of seedlings was common. Favorable survival (97%) was recorded in 1987 (Table 20). Livestock also impacted cottonwood seedling survival on the transects. Seven transects were exposed to moderate winter cattle use in 1983 and 1984. Dormant seedlings were not noticeably browsed but some trampling of seedlings probably occurred. Sedgwick and Knopf (1988) observed first year (1983-84) fall to fall seedling survival at 2.6-18.5% within these and other moderately grazed winter plots whereas seedlings within ungrazed plots survived slightly better (3.2-27.4%). �376 Two transects containing 1984 seedlings were subjected to more intensive winter livestock use. Trampling and desiccation were the primary causes of 100% mortality on both transects. One other transect received summer livestock use in 1985 when low stream flow permitted cattle to cross the river. Extensive browsing was noted on the seedlings and mortality in 1985 approximated 80%. Surviving seedlings remained alive from basal regrowth in 1986 but were severely stunted. Observations indicate cottonwood seedlings are readily consumed by livestock when the plants are leafed out. In addition they are readily browsed by cottontails (Sylvilagus spp.), deer (Odocoileus spp.), and beaver. Based on these data, it evident that cottonwood seedlings may survive under light winter grazing when the seedlings are dormant. However, with spring through early fall livestock use, there is little chance for survival. Extensive Transects.--Preliminary sampling of cottonwood seedling occurrence over extensive areas was conducted within several walking transects on Division of Wildlife Properties along the South Platte in early fall 1983. Seedlings were found within 14.8% of 1,055 m2 samples. Thirty extensive random transects, distributed from the Nebraska-Colorado boundary upstream to Atwood were established in 1984 and were sampled in early fall from 1984 to 1987. New cottonwood seedlings were recruited into the samples during all 4 years and ages of older seedlings became progressively more difficult to determine. However, seedlings established from 1985 ·through 1987 occurred in low areas subject to frequent inundation and their chances of long-term survival were minimal. Attrition of older seedlings was noted and, whereas they represented 1.9% occurrence in 1984 they probably represented less than 1% occurrence in 1987 (Table 22). The number of transects containing cottonwood seedlings also declined over time. J Table 22. Number, age, and frequency of occurrence (%) of cottonwood and green ash seedlings within 30 random transects along the South Platte River, northeastern Colorado, 1984-87. Year 1984 1985 1986 1987 Co ttonwoo d Year established 1983 1984 1985 Total 22 5 10 12 l7a 17 9 21 8 32 26 38 2Sa % occurrence Transect m2 1.9 1.5 1.9 1.3 56.7 43.3 53.3 30.0 Green ash % occurrence Total m2 Transect 4 9 12 34 0.24 0.51 0.60 1.74 10.0 16.7 23.3 40.0 aAges of seedlings could not be accurately determined in 1986 and 1987. b1985 or later. In contrast to the decline of cottonwood seedlings, those of green ash, considered more mesic in adaptation, increased with time within the transects CTable 22). This species was represented by few, if any, old trees along the river but a few young to medium age trees were present. Green ash were found within 12 of 30 transects in 1987. Sedgwick and Knopf (1988) summarized reasons why this species (referred to as red ash) seemed to the most probable �377 species for future overstory dominance along the South Platte. Personal observations strongly support their hypothesis. Observations in Nebraska indicate green ash usually form dense closed canopies so that little understory vegetation exists, however, old trees contain numerous cavities. Russian-olive seedlings were found twice in 1987 transects and were observed spreading into the floodplain from the SPWMA meadows. Previously, mortalities of 2-3 year old seedlings had been noted as a consequence of prolonged inundation in 1983. Subsequent floods may restrict this species to the outer edges or higher elevations within the South Platte floodplain. Common hackberry (Celtis occidentalis) and silver buffalo berry (Shepherdia argentea) were planted within the Sedwick Bar Property along the South Platte floodplain in western Sedgwick County in the early 1950's. Little or no natural regeneration and spreading of these species has been observed in adjacent or downstream locations. A few other exotics occur along the South Platte including silver poplar (Populus alba) but none has shown evidence of becoming a major invader. One Juniperus species was recently reported growing wild within the SPWMA (M. Gardner, Pers. Commun.). Supplementing Natural Reproduction Cottonwoods, willows, and several other woody species including grape vines can be propagated by use of hardwood cuttings (Doran 1957, u.S. Dep. Agric. 1965). This technique has been used successfully in the southwestern United States (U.S. Dep. Agric. 1983, Swenson and Hullins 1985, York 1985). Therefore, evaluations of this technique were conducted to ascertain if stem cutting propagation could be used effectively to supplement natural reproduction along Colorado's rivers. trials were begun on the SPWMA in late winter 1984, primarily using cuttings from the previous year's growth. Stem cutting survival to mid summer appeared good but survival to early fall varied widely among species and was poor (Table 23). Stem cutting efforts were expanded to several additional locations along the lower South Platte, South Republican, and Arkansas rivers in 1985. Good survival occurred to early summer but survival was poor to early fall (Table 23)~ Plant vigor and growth among survivors also was poor and few survived until spring 1986. Reasons for poor survival were uncertain but 3 primary factors were suspected. First was competition with herbaceous vegetation present at most locations. Second was unstable ground water levels that rose to near the soil surface in late spring and decreased rapidly in mid summer when irrigation demands reduced stream flows. Within the Arkansas Valley, the water table rose to drown most of the 1985 cuttings placed there. The 3rd suspected reason for poor survival was that cuttings had not been placed deep enough and groundwater levels in mid to late summer fell below the average depth at which cuttings were placed (Fig. 12). Planting methods were modified in 1986, primarily on the SP~lliA,to identify which of the 3 variables impacted stem cutting survival. Preliminary �378 Table 23. Survival of stem cutting plantings from late winter to early fall within the South Platte, South Republican, and Arkansas River planting sites, 1984-86. Species Populus sargentii Salix amygdaloides S. interior!exigua lutea S. vitellina Amorpha fruiticosa Vitis vulpina Parthenocissus quinquefolia s. Annual total/average Planted 1984 Survival Planted 1985 Survival Planted 1986 Survival 31 6 12 35 39 15 12 4 0.35 0 1.00 0.43 0.59 0.13 0 0 80 40 53 70 15 55 40 20 0.11 0.05 0.26 0.04 0.87 0.27 0.12 0.05 34 24 5 33 5 30 13 0.41 0.29 1.00 0.52 0.60 0.50 0.31 154 0.41 373 0.17 144 0.45 Average planting depths of 10 dm in 1985 were increased in 1986 to range from 17.1 to 19.5 dm whereas lowest groundwater levels were in the 12-15 dm range (Fig. 12). In addition, two-thirds of the 1986 cuttings were placed within sites overlaid with a plastic-organic mulch (Snyder 1983) to eliminate competition with herbaceous vegetation. Groundwater levels were monitored more frequently to detect variations. Among the tree, shrub, and vine species tested all but one, golden willow tSalix vitellina), were native species which previously had shown promise of being able to survive as stem cuttings. Several other species, including New Mexico elderberry (Sambucus neomexicana), cotoneaster (Cotoneaster sp.), and Russian-olive, which had shown little evidence of survival, were not included in 1986 evaluations. Survival to fall among 97 plantings where herbaceous vegetation competition had been eliminated with mulch was 52%, whereas survival of 47 unmulched cuttings was 32% (P < 0.10). This provided evidence that competition was at least a partial factor reducing stem cutting survival. Survival among unmulched cuttings was 34% in 1986 compared to 22% in 1985 among the same combined species within the same locations. This provided little evidence that increased 1986 planting depth markedly enhanced survival. Attrition continued among the 1986 cut,tings over winter and by late summer 1987, only 33% of those surviving to fall 1986 were still alive. Other factors including browsing of new leaves by deer and/or cottontails in 1986 were possibly important. However, the evidence indicates that fluctuating groundwater was the primary variable contributing to low survival of stem cuttings. New roots forming in moist soil above the water table might be repeatedly inundated and desiccated and therefore severely weakened or killed, and the new cuttings had only limited energy reserves. Swenson and Mullins (1985) noted reduced survival among cuttings placed in fluctuating water tables in their New Mexico trials. Another factor making stem cutting planting questionable was the difficulty encountered in placement of cuttings into the water table. When planting was �o CHECK STATION 2 - TAMARACK MEADOW 11-E 4 en a: w I- w 6 ~ I (.) w o ...., J: Ia. w o a: 8 \ I \ I \ I \ , \ 10 w , ,, I I- « :r: ,•... .. , ', ,, ,, , ,, , ,, ,, ,, ,, " ,, ,, 12 ,I I ' ,,, • ,, , I I I " I , . ,,, , " ,, , , ,, '' • . ," ,, ,,, , I I , , \ \ \, " ," , I ,, , ,, ,, , ,'", I I , , , I , ,,, ,,, ,, , , I " ,, , , , , I I I . '--' I , ,, ," , ,, I ••I , 14 , " ,I , I I I I I \ I " , ,, , , ', • ,, ,, , ,' , ',, ,, ,, ,, , ,, , , ,, ,, " , ,,, , I , \'•' I I , ' ,_' , ,, I ,, ,, , .•. , I , I If • " 16 J F M A M J J A SON D J F M A M J 1984 Fig. 12. Groundwater fluctuations within stern cutting lower South Platte River, northeastern Colorado. J A SON D J F M A M J J A SON 1986 1985 planting D sites from 1984 through 1986 along the LV --..J \.0 �380 done in March, water tables were moderately high so that it was essential to place the cuttings several dm deeper to be positioned below the lowest late summer water levels (Fig. 12). However, when digging holes for the cuttings the sands would collapse preventing further penetration as soon as the water table was reached. A retaining sleeve (casing) had to be inserted and repeatedly forced down as the sand was extracted with a hand auger and the stem cutting had to be inserted before the sleeve was removed. This procedure was labor intensive and too costly to be practical on an extensive scale. Svoboda and Graul (1980) recommended scouring 15 x 30 m strips about 2-dm deep with a D-8 dozer blade in late fall or early spring to remove competitive vegetation permitting the natural establishment of cottonwoods. A dozer blade was not available for use during this study but a small tractor and rototiller were used to remove competitive herbaceous vegetation from 18 randomly selected sites within 6 locations on the SPWMA in 1984. Three additional disturbed sites along a pipeline installed in October 1984 were included in the test. In early summer 1985, cottonwood seedlings were observed on 3 of the sites but all other locations were too far above the water table to sustain moist soil conditions essential for seed germination. Decreased mid-summer river flows caused desiccation of all seedlings within the 3 sites. Sites were maintained weed free into the 1986 growing season but low groundwater levels, due to reduced river flows, again prevented seedling germination. Increasing problems with control of noxious weeds forced abandonment of this experiment. Most of the tilled sites were proximal to cottonwood seedlings that had established in 1983 indicating that if high sustained stream flows had occurred, the results would have been dramatically different. Understory Vegetation Along the South Platte River Natural Occurrence.--A frequency of occurrence index showing the relative abundance and importance of different vegetation and ground cover types was derived from the 30 extensive transects along the South Platte River from 1984 to 1987 (Table 24). Perennial grass was the dominant cover during all years with annual and biennial forbs ranking 2nd in 1984 and 1985, and perennial forbs ranking 2nd in 1986-87 (Table 24). Annual grasses, primarily Bromus spp., like annual forbs, showed declining trends through the 4 years. Low shrubs, primarily snowberry and poison ivy, and tall shrubs, primarily sandbar and coyote willow, were nearly equal in occurrence. In combination they comprised 20-25% of the dominant cover. Grape vines frequently were encountered and often were associated with patches of low shrubs whereas Virginia creeper (Parthenocissus quinquefolia) was infrequently encountered. During sampling within the SP~1A in 1972-73 (Snyder 1978), grasses represented 40% occurrence compared with 24% forbs, 20% brush and vines, 5% litter, and 11% sandbars or bare ground. These data (Table 24) show that vegetation along the lower South Platte was not only diverse in growth form, but was also interrupted by sand bars and bare ground increasing edge and interspersion (horizontal diversity). Overstory trees provided additional vertical diversity enhancing the quality of the riverbottom for wildlife. �381 Table 24. Frquency of occurrence (%) of major cover types within random transects along the South Platte River, northeastern Colorado, 1984-87. Year 1984 1985 1986 1987 Forbs 24.8 14.9 11.6 5.0 Grasses 12.0 8.0 13.7 19.1 9.1 8.4 4.4 5.3 25.0 32.9 37.1 37.7 Shrubs Low Tall 9.7 12.7 11.4 12.2 10.9 11.4 11.2 12.8 Vines Litter 2.0 2.3 1.9 2.1 5.8 7.6 Sand bar 0.7 1.9 1.6 1.2 7.1 4.8 aAnnual forbs include some biennials. bperennial grasses including common reed (Phragmites communis). CLitter includes downed timber and sand bar includes bare ground. As a consequence, of 1983 flooding, mud flats and sand bars were numerous. Limited fall 1983 sampling indicated about 15% of the area was bare. Stream flows were neither as high nor as prolonged in 1984 allowing initial recovery of herbaceous vegetation. The transition from annual and biennial forbs, and from total annuals (forbs + grass) toward perennial forbs and grasses through the 1984-87 interval was highly evident (P < 0.1, Table 25). Occurrence of annuals and perennials was nearly equal in 1984, each representing about 1/3 of the total composition. However, by 1987 perennials dominated 57% of the samples whereas annuals represented only 10%. Although pre-flood vegetation occurrence data were lacking it seems apparent that 1983 flooding promoted major regression of vegetation toward early seral stages. Host perennials were stunted rather than killed allowing for their rapid recovery. Table 25. Frequency of occurrence (%) of dominant annual and perennial herbaceous vegetation in m2 samples within extensive transects along the South Platte River, northeastern Colorado 1984-87. Year Forbs 1984 1985 1986 1987 24.8 14.9 11.6 5.0 Annuals and biennials Forbs + grasses 33.9 23.3 16.0 10.3 Perennials Forbs + grasses 37.0 40.9 50.8 56.8 These data illustrate that 1983 flooding by its scouring, silting, inundation, and general disturbance of bottomlands dramatically promoted production of annual herbaceous vegetation in addition to cottonwood reproduction. Food habits studies (l1artinet a1. 1951) have demonstrated that seeds of annuals are the primary food of a majority of wildlife species, especially upland game �382 and passerines. In riverbottoms, where flooding has been eliminated, e.g., below reservoirs, conversion to perennial grass monocultures occurs rapidly so that much of the food base essential for wildlife species richness (numerous species) is lost before the vertical and horizontal diversity, promoted and sustained by flooding, is lost. Disturbance Tillage.--Along the lower Sout.hPlatte River where flooding still occurs, numerous grass-dominated areas, usually elevated above the normal flood plain, occur which are seldom flooded. Disturbance tillage of narrow linear strips within these areas has been recommended to promote food producing wild annuals for northern bobwhite and other game and nongame wildlife (Rutherford and Snyder 1983). Personnel responsible for managing several CDOW properties along the lower South Platte have implemented limited disturbance tillage. Fireguards plowed out of grass sod and then disced to create mineral soil fire barriers also provided additional disturbance tillage strips. Frequency of occurrence sampling within 7 disturbed strips in fall 1986 and 1987 revealed major enhancement of annuals (Table 26). Annuals were not promoted as dominants in 1986 within one disturbance tillage site where snowberry and poison ivy dominated prior to plowing. These species resprouted profusely to retain dominance and suppress growth of annuals. Annuals dominated among samples in 4 of 5 other sites in 1936 and were common in all. Averaged among sites, annuals dominated 48.2% of the 1986 samples whereas .biennials dominated 25.3% and perennials 26.5%. Table 26. Frequency of occurrence (%) of annuals (A), biennials (B), and perennial (p) herbaceous vegetation within disturbance tillage strips, South Platte River, Northeastern Colorado 1986-87. 1987 1986 Location Brush Wildlife Area 1-86 2-86 1-85 2-85 Elliott West East Tamarack Interior Exterior P N A 17.7 39.7 23.1 19.4 12.7 9.5 18.4 49.3 70a 68a 50b 50b 31.4 30.9 36.0 54.0 55.3 42.6 2.1 40 75 100.0 100.0 30.9 14.6 25.0 54.4 75.0 40a N A 79 63 65a 67a 69.6 50.8 58.5 31.3 47 55 16 0 B 10.0 P B 27.2 10.3 12.0 8.0 41.4 58.8 52.0 38.0 0 0 0 0 40.0 50.0 aSampling was conducted 2 growing seasons after disturbance tillage. bSampling was conducted 3 growing seasons after disturbance tillage. �383 Among the annuals sunflower (Helianthus sp.) was the most frequent dominant often averaging 1.9 m tall and provided food as well as cover. Annual ragweed (Ambrosia trifida) and lambsquarter and orache (Chenopodium spp.) were less abundant but provided excellent food-tall cover combinations where present. The majority of 13 annuals, rated as dominants in one or more samples we~e considered good seed sources for wildlife whereas only western ragweed (Ambrosia psilostachya) of 12 perennials was considered important (Martin et al. 1951). A rapid transition to dominance by perennials was noted in 1987 within the Brush and Tamarack sites as disturbance tillage was not repeated and successional progression occurred (Table 26). Among 4 sites within the Brush Wildlife Area, annuals averaged 1.3 m in 1986 and 1.2 m in 1987. Biennials were 1.0 m tall both years and perennials averaged 0.7 m in 1986 and 0.5 m in 1987. In addition to being an excellent food source, annuals usually have improved height, they·remain erect, and usually retain an open understory making them an excellent food-cover combination for wintering wildlife. Prescribed Burning Within Riverbottom Prescribed burns were conducted by CDOW management personnel on several bottomland meadow sites along the South Platte River in early spring 1985 and 1986. Their primary objectives of removing dense matted residual cover and stimulating new growth of perennial grasses were achieved. Within the burned sites, average height-density indices (HDI) (Robel et al. 1970) increased from 0.57 to 1.25 dm respectively from late winter pre~ to postburn intervals. ~·lithinproximal unburned controls, HD Lncreaaed at a slower rate (0.35 to 0.61 dm, P< 0.05). Burned sites remained dominated by perennial grasses but first-year enhancement by fire of sunflowers, sweet clover (Melilotus spp.), and other forbs was noted. Although fire opened the herbaceous cover allowing increased use by wildlife in spring and summer, grass stands in fall and winter appeared generally as dense and difficult for small wildlife to penetrate as they were prior to burning. Early fall observations indicated most wildlife were concentrated along the forb-dominated firguards. A few plains cottonwoods, peachleaf willow, sandbar willow, indigobush, and other woody species were present in areas burned in 1985 and 1986. In 1985 most trees and shrubs were subjected to a rather intense headfire whereas in 1986 suppressed backfires were set around most mature trees. Among 34 primarily mature plains cottonwoods subjected to fire in 1985, 53% were not markedly impacted whereas 32% were moderately impacted (considerable basal scorching and branches killed to above 2 m) and 15% were severely impacted or killed. Six of 11 peachleaf willow, which were smaller trees than the cottonwoods, were severely impacted. The peachleaf willows root sprouted and sent up new growth to 2 m whereas no basal sprouting among mature cottonwoods was noted. Young cottonwood seedlings that had germinated in 1983 were burned off in both 1985 and 1986 fires but most resprouted to resume growing. Three of 4 young Russian-olive trees were entirely scorched by the 1985 fire. Resprouting was noted from basal interior areas but regrowth was not vigorous. Among shrubs, indigobush amorpha, and snowberry both showed rapid regrowth during the first growing season. In contrast sandbar willows appeared more severely impacted by fire and their regrowth was marginal. Regrowth of Virginia creeper was noted following 1986 burns. �384 Wildfires occur infrequently along the South Platte river bottom and their extent intensity, and the time of year determine their impact on cottonwoods and other vegetation. Postburn surveys of 3 wildfires in recent years show generally moderate damage to cottonwoods. Although the tops of many trees may live, branches close to the ground which provide screening and protection for many wildlife species are lost. Wildfires are more frequent along the Arkansas River and their impacts are much more severe. Many are set intentionally to open stands of tamarisk. However, tamarisk resprouts profusely and rapidly regrows. Observations indicate the burning tamarisk produces intense heat as well as flames that reach several m into the air killing any overstory cottonwoods. As a consequence, the transition from cottonwoods to pure stands of tamarisk is enhanced. Riverbottom Habitat-Wildlife Relationships The direct .impacts of cottonwood riparian habitat quantity and quality on wildlife species richness and density is apparent, but were not studied. However, before changes in quantity and quality of riverbottom habitat can really be meaningful their impacts on wildlife must be known. Past studies (Beidleman 1954, 1978; Crouch 1961; Bottorff 1974; Fitzgerald 1978; Rucks 1978; Ports 1980; Tubbs 1980; Hoover and Wills 1984; Knopf 1985, 1986; Knopf and Sedgwick 1987; Sedgwick and Knopf 1986, 1987, 1988) provide a base of general knowledge conc&rning wildlife species diversity and the importance of the cottonwood riparian ecosystem in relation to other habitats in Colorado and the Great Plains. Much of this knowledge is restricted to the South Platte River in northeastern Colorado and little data are available for wildlife along other Colorado rivers. Several of these references also provide information on densities of certain wildlife species within specific study sites and Fitzgerald (1978) associated densities with plant communities within his South Platte study areas. Apparently, little \wrk has been done in Colorado to relate densities of specific wildlife species to cottonwood riparian habitat quantity-quality indices over extensive areas. Such relationships would provide a basis for identifying the relative importance of specific food-cover associations, importance of vegetation structure, relative importance of plant communities, and the minimum requirements for specific wildlife species. Monitoring among seasons and over time would reveal habitat trends and their impacts on wildlife. Beidleman (1978) and Fitzgerald (1978) stressed the lack of wildlife data for cottonwood riparian habitats and emphasized the importance of broad based studies such as those conducted in Arizona (Anderson et ale 1978, Anderson and Ohmart 1980, 1985). Svoboda and Graul (1980) discussed application of a species-ecosystem approach to managing cottonwood riparian habitats for wildlife. Using this approach management would be directed toward wildlife species which were most sensitive to habitat changes or had the most narrow range of ecological requirements (stenotopic). The goal of species-ecosystem management would be to maintain cottonwood riparian habitats so no species was lost as a functional element of the ecosystem (Graul et ale 1976). However, in reviewing wildlife species present along the South Platte, Svoboda and Graul (1980) found few truly stenotopic species and recommended adding several species with widely separate ecological requirements to the list of stenotopic species so that management goals could be more comprehensive. Lack of stenotopic species along the South Platte was attributed to nearly all wildlife being invaders since development �385 of the cottonwood ecosystem. Knopf (1986) stated that 90% of the contemporary avifauna along the lower South Platte River was not present at the turn of the century. Whereas the general objective in riverbottom habitat may be to sustain or enhance habitats for all existing wildlife, more specific objectives directed toward a single species or group of species may be included under the species-ecosystem approach. However, this assumes no wildlife species will be eliminated by management directed toward another. Miller (1985) summarized initial attempts to implement ecosystem management within 2 CDOW properties along the South Platte River applying the process described in Hoover and Wills (1984). However, to date, little progress has been made toward this type of management within the cottonwood riparian habitats of Colorado. Currently, the northern bobwhite is the only eastern Colorado riparian species which is censused routinely, and for which general density information can be obtained among differing locations and years. Livestock Grazing Impacts on Wildlife Livestock grazing can impact natural regeneration of cottonwoods and willows, especially when conducted during the growing season. Most livestock impacts on wildlife are indirect through their effects on vegetation. Studies have generally shown negative impacts of grazing on wildlife. Crouch (1981) noted dramatically greater numbers of birds and mammals during 15 inven~ories of the ungrazed SPWMA in comparison to a proximal intensively grazed riverbottom. Taylor (1986) found that counts of birds were 5-7 times higher on an area ungrazed since 1940 than on 2 areas grazed annually until 1980, and 11-13 times higher than on a transect severely disturbed by extensive grazing and dredging activities. Total numbers of birds and mammals decreased with increased grazing pressure along streams in western Colorado (Rucks 1978). Numerous other studies have reported similar findings. However, the intensity, duration, and period of grazing influences the impact of grazing on wildlife. Light to moderate grazing during winter, when plants are dormant, does not destroy all cottonwood seedlings and its general impact on riverbottoms is less severe than summer grazing. Sedgwick and Knopf (1988) found that grazing at a moderate rate when vegetation was dormant on a SPWMA study site had no apparent negative effect on 6 migratory birds species that were predominantly insectivorous but foraged and/or nested within the grass-herb-shrub layer of the riverbottom. Their study did not include impacts on more numerous birds using riverbottoms that were primarily weed seed eaters. The first foods generally consumed by livestock are seed heads of herbaceous annuals, if available. Thus, livestock are in direct food competition with numerous wildlife species. Livestock also reduce cover quality and quantity needed by many wildlife species. Fall movements of northern bobwhite coveys from grazed to ungrazed sites for winter survival were noted during population dynamics studies within the SPWMA (Snyder 1978). In Colorado the problem of intensive livestock degradation of riparian habitat is not confined to riverbottoms. Hundreds of intermittant tributaries and small streams could support vast increases in wildlife habitat if livestock were excluded, especially during the growing season. �386 CONCLUSIONS Persons working with wildlife in Colorado should be concerned because stands of cottonwoods, willows, and wildlife habitat are deteriorating along all inventoried rivers. The same trend is probably occurring along other lower elevation streams, although the rate of habitat deterioration varies widely among streams. Among the inventoried rivers, wildlife habitats along the Arkansas River are deteriorating most rapidly and opportunities for curtailing the decline appear dim. Although habitat deterioration along the South Platte and Colorado rivers was evident, declines were at more moderate rates and recent flooding has stimulated regeneration of cottonwoods and willows. Habitat conditions along the Rio Grande River appear more stable. However, resident wildlife managers indicate little new natural reproduction was occurring. The Rio Grande River may be in the early stages of a general habitat decline. Construction of mainstream flood control reservoirs (e.g., John Martin Reservoir) has a devastating impact on downstream riparian habitats. Such reservoirs, by eliminating flooding, and reducing and stabilizing streamflows, curtail nearly all opportunity for natural regeneration of cottonwoods and willows, as well as allowing perennial grasses to increase eliminating food and horizontal diversity needed by wildlife. The river channel narrows and deepens so that its capacity to sustain high stream flows, if they ever occur, is greatly diminished. Construction of the Narrows Reservoir on the South Platte would be especially devastating to dOWF_stream wildlife habitats because few tributaries occur below the proposed dam site and opportunity for downstream flooding would be low. Construction of water storage reservoirs off the main channel would be much less environmentally damaging. Willows and cottonwoods represent seral stages of succession that have expanded along Colorado rivers in past decades, apparently because of man's reduction of peak seasonal stream flows, alteration of stream flows toward stability, disturbance of soils within riparian habitats, and possibly because of elimination of buffalo and wildfires. Willows and cottonwoods are intolerant of shade and are usually perpetuated by disturbance (flooding, soil scarification, etc.) in combination with high groundwater and moist soil conditions. Once willows and cottonwoods become established within a flood plain, flood water flows are slowed and sediments are deposited so that, over several to many successive floods, the ground level gradually becomes elevated above the streambed and the groundwater level. This in turn, reduces opportunity for natural regeneration as willows and cottonwoods tend to not readily perpetuate themselves in the same location. Assuming soil quality and soil moisture are satisfactory, seed sources are available, and livestock grazing permits, shade tolerant mesic tree and shrub species (eastern red cedar, green ash, etc.) invade under cottonwoods to replace them. If these criteria are not met, the site will progress toward perennial herbaceous vegetation. The streamside soil building successional process has probably occurred along most low elevation streams in Colorado and is partially responsible for the current degeneration of cottonwood stands. However, construction of �387 reservoirs to eliminate or reduce flooding, river dewatering for irrigation, invasions of tamarisk, and intensive livestock grazing especially from spring to fall have all severely impacted natural reproduction of willows and cottonwoods along Colorado streams. Natural succession is the primary process along the South Platte supplemented by localized grazing impacts. In contrast, all factors and especially flood control, tamarisk, and wildfires are impacting the Arkansas River floodplain. Cutting of old trees for firewood has also increased dramatically in recent years. The South Platte riverbottom is the only riparian location in Colorado where another overstory species may successionally replace cottonwoods in the near future. Green ash is anticipated as the primary replacement species along the South Platte and in many locations will occur in dense stands. Succession in most other locations will be either toward tamarisk or perennial herbaceous vegetation. Russian-olive may continue to spread along rivers in Colorado as an intermediate rather than overstory species (Knopf and Olson 1984). The probability of this species forming dense monoculture stands remains uncertain. Although Russian-olive, green ash, and other trees may compete with existing cottonwoods for space and water, their reproductive requirements are different than those of cottonwoods. Therefore, direct competition during reproduction and establishment is not considered important. Where dense stands of tamarisk have invaded floodplains, as along the lower Arkansas "aridColorado river drainages, soil deposition during flooding may be dramatically more rapid than in open cottonwood habitats. When soil building is rapid, opportunities for natural establishment of any woody species are reduced. If tamarisk is not perpetuated from root sprouts, such sites will gradually revert to perennial herbaceous vegetation. Available evidence indicates tamarisk is not tolerant of extremely cold temperatures, and therefore, will probably not be a major invasion threat in northeastern Colorado or in high elevation mountainous areas. Tamarisk is directly competitive with willows and cottonwoods during reproductive phases. Once established, there currently is no known economical method of eradicating tamarisk (Kerpez and Smith 1987). The CDOW does not have control over much of the riparian habitat along stream and riverbottoms in Colorado, nor does it have control over stream flows. Therefore, opportunities for management of riverbottom habitats are limited. If demands for outdoor recreation continue to increase while livestock and cropland values remain suppressed, demand should dictate enhancement of bottomlands for harvestable wildlife. The CDOW can assist private landowners in sustaining wildlife oriented habitat. Many hunting clubs already own or lease extensive areas along the South Platte and Arkansas rivers and, with suppressed land prices, conversion of these lands to recreational areas may continue. Fluctuating groundwater levels are the major cause of poor survival of cottonwood-willow stem cuttings. Therefore, stem cutting planting is not an economical method of supplementing natural reproduction. A second deterrent is that below groundwater level in most locations, sands collapse into holes so that stem cuttings can not be readily inserted. Hodified procedures have �388 been described wherein cuttings were inserted behind a high pressure water jet system (Schultze and Wilcox 1985). However, large, expensive equipment would be needed and the approach still is not practical if survival of cuttings can not be increased. Elimination of flooding not only curtails reproduction of woody species but also allows herbaceous perennials to become dominant as seed producing annuals decline. Thus, flooding is of value for the disturbance it promotes which retains seral stages of vegetation succession as well as for its stimulation of natural tree-shrub regeneration. LITERATURE CITED Anderson, B. W. 1985. Managing riparian vegetation and wildlife along the Colorado River: synthesis of data, predictive models, and management. Pages 123-127 in R. R. Johnson, C. D. Ziebell, D. R. Patton, P. F. Ffolliott, and~. H. Hamre, Tech. Coords. Riparian ecosystems and their management. First North American Riparian Conf. U.S. Dep. Agric., For. Servo Gen. Tech. Rep. ~~-120. , R. D. Ohmart and J. Disano. ----~ floodplain for wildlife. Pages 1978. Revegetating the riparian 318-331 in R. R. Johnson, and J. F. McCormick, Tech. Coords. Strategies for-Protection and management of floodplain wetlands and other riparian ecosystems. U.S. Dep. Agr.ic., For. Servo Gen: Tech. Rep. W-0-12. ----- , and R. D. Ohmart. 1980. Designing and developing a predictive model and testing a revegetated riparian community for southwestern birds. Pages 434-449 in R. M. DeGraff, Tech Coord. Management of western forests and grasslands~or nongame birds. U.S. Dep. Agric., For. Servo Gen. Tech. Rep. INT-86. Athearn, F. J. 1985. Land of contrast - A history of southeast Colorado. U.S. Dep. Inter., Bur. Land Manage. Cultural Resource Sera 17. 26lpp. Baker, J. B., and W. M. Broadfoot. 1977. A practical field method of site evaluation for eight important southern hardwoods. U.S. Dep. Agric., Servo Gen. Tech. Rep. SO-14. 3lpp. Beidleman, R. G. 1954. The cottonwood riverbottom community as a vertebrate habitat. Ph.D. Thesis, Univ. Colorado, Boulder. 4l5pp. 1978. The cottonwood-willow riparian ecosystem as a vertebrate habitat with particular reference to birds. Pages 192-195 in W. D. Graul and S. J. Bissell, Tech. Coords. Lowland river and stream habitat in Colorado: a symposium. Colorado Chap., The Wildl. Soc. and Colorado Audubon Counc. Bittinger, M. W., and G. E. Stringham. 1963. A survey of phreatophyte growth in the lower Arkansas River valley of Colorado. Civil Eng. Sect. Colo. Exp. Stn., Colorado State Univ., Fort Collins. 3lpp. �389 Bottorff, R. L. 28: 975-979. 1974. Cottonwood habitat for birds in Colorado. Am. Birds Broadfoot, W. M. 1973. Water table depth and growth of young cottonwood. U.S. Dep. Agric., For Servo Res. Note SO-167. 4pp. Christy, S. 1973. An analysis of the woody vegetation on the South Platte River flood plain in northeastern Colorado. M.S. Thesis, Univ. Northern Colorado, Greeley. 82pp. Crouch, G. L. 1961. Inventory and analysis of wildlife populations and habitat, South Platte River valley. Job Compl. Rep. Colorado Game and Fish Dep. Fed. Aid Proj. W-l04-R-1-2. 68pp. 1978. Effects of protection from livestock grazing on a bottomland wildlife habitat in northeastern Colorado. Pages 118-125 in W. D. Graul and S. J. Bissell, Tech Coords. Lowland river and stream habitat in Colorado: a symposium. Colorado Chap., The Wildl. Soc. and Colorado Audubon Counc , 1979. Long-term changes in cottonwoods on a grazed and an ungrazed plains bottomland in northeastern Colorado. u.S. Dep. Agric., For. Servo Res. Note RM-370. 4pp. 1981. Wildlife on ungrazed and grazed bottomlands on the South Platte River, northeastern Colorado. Pages 186-197 in L. Nelson, Jr. and J. M. Peek. Proc. Wildlife-Livestock Relationships Symp. Univ. Idaho For. Wildl. and Range Exp. Stn. Moscow. Doran, W. L. 1957. Propagation of woody plants by cuttings. Massachusetts Agric. Exp. Stn. Bull. 491. 99pp. Univ. Fitzgerald, J. P. 1978. Vertebrate associations in plant communities along the South Platte River in northeastern Colorado. Pages 73-88 in W. D. Graul and S. J. Bissell, Tech. Coords. Lowland river and stream habitat in Colorado: a symposium. Colorado Chap., The Wildl. Soc. and Colorado Audubon Counc , Graul, W. D. 1977. Avian management considerations for Colorado riparian habitat. 6pp in U.S. Dep. Agric., For. Servo Region 2 Avian Habitat Symposium (Unpublished). --- , J. Torres, and R. Denney. nongame programs. Harrington, H. D. Denver, Colo. 1976. A species-ecosystem approach for \-lildl.Soc. Bull. 4:79-80. 1954. Manual of the plants of Colorado. 666pp. Sage Books, �390 Hoar, A. R., and M. J. Erwin. 1985. Relationships between the expansion of agriculture and the reduction of natural riparian habitat in the Missouri River floodplain of northeast Montana, 1938-1982. Pages 250-256 in R. R. Johnson, C. D. Ziebell, D. R. Patton, P. F. Ffolliott, and R. H. Hamre, Tech. Coords. Riparian ecosystems and their management: Reconciling conflicting uses. First North American Riparian Conf. U.S. Dep. Agric., For. Servo Gen. Tech. Rep. RM-120. Hoover, R. L., and D. L. Wills, Eds. 1984. Managing forested lands for wildlife. Colorado Div. Wildl. and U.S. Dep. Agric., For. Serv., Rocky Mountain Region, Denver, Colo. 459pp. Hopper, R. M. Colorado. 1978. Evaluation of pothole blasting for waterfowl in Colorado Div. Wildl. Spec. Rep. 44. 2lpp. Johnson, R. R., and D. A. Jones, Tech. Coords. 1977. Importance, preservation and management of riparian habitat: a symposium. Agric., For. Servo Gen. Tech. Rep. RN-43. 2l8pp. U.S. Dep. , C. D. Ziebell, D. R. Patton, P. F. Ffolliott, and R. H. Hamre. 1985. reconciling conflicting uses. F~rst North American Riparian Conf. U.S. Dep. Agric., For. Servo Gen. Tech. Rep. RN-120. 523pp. ---,.".Riparian ecosystems and their management: Kerpez, T. ~., and N. S. Smith. 1987. Saltcedar control for wildlife habitat improvement in the southwestern United States. U.S. Dep. Inter., Fish and Wildl. Servo Resour. Publ. 169. l6pp. Knopf, F. L. 1985. Significance of riparian vegetation to breeding birds across an altitudinal cline. Pages 105-116 in R. R. Johnson,C. D. Ziebell, D. R. Patton, P. F. Ffolliott, and H. Hamre, Tech. Coords. Riparian ecosystems and their management: reconciling conflicting uses. First North American Riparian Conf. U.S. Dep. Agric., Servo Gen. Tech. Rep. RM-120. R: 1986. Changing landscapes and the cosmopolitism of the eastern Colorado avifauna. Wildl. Soc. Bull. 14:132-142. ____ , and T. E. Olson. 1984. Naturalization of Russian olive: to Rocky Mountain wildlife. Wildl. Soc. Bull. 12:289-298. J. A. Sedgwick. 1987. Latent population responses ----to, and a catastrophic, climatic event. Condor 89:869-873. implications of summer birds Lindauer, I. E. 1970. The vegetation of the flood plain of the Arkansas River in southeastern Colorado. Ph.D. Thesis, Colorado State Univ., Fort Collins. 99pp. 1983. A comparison of the plant communities of the South Platte and Arkansas River drainages in eastern Colorado. Southwest. Nat. 28:249-259. Nartin, A. C., J. S. Zim, and A. L. Nelson. 1951. American wildlife and plants: a guide to wildlife food habits. Dover Publ. Inc., New York, N.Y. 500pp. �391 McCluskey, D. C., J. Brown, D. Bornholdt, D. A. Duff, and A. H. Winward. 1983 (1984). Willow planting for riparian habitat improvement. U.S. Dep. Inter., Bur. Land Hanage. Tech. Note 363. 2lpp. HcGregor, R. L., Coord. Lawrence. 1,392pp. 1986. Flora of the Great Plains. Hiller, G. C. 1985. Ecosystem management. Wildl. Proj. 86-039-01-1-1. 28pp. Job Final Rep. Univ. Kansas Press, Colorado Div. Nadler, Jr., C. T. 1978. River metamorphosis of the South Platte and Arkansas rivers, Colorado. M.S. Thesis, Colorado State Univ., Fort Collins. l5lpp. Nelson, Jr., L., and J. M. Peek. 1981. Proc. Wildlife-Livestock Relationships Symp. Univ. Idaho For. Wildl., and Range Exp. Stn. Moscow. 6l4pp. Ports, M. A. 1980. Avian density and diversity in four habitat types in Morton County, Kansas. Kansas Ornithol. Soc. Bull. 31(2):17-21. Robel, R. J., J. N. Briggs, A. D. Dayton, and L. C. Hulbert. 1970. Relationships between visual obstruction measurements and weight of grassland vegetation. J. Range Hanage. 23:295-297. Rucks, J. A. 1978. Comparisons of riparian communities· influenced by grazing. Pages 100-113 in W. D. Graul and S. J. Bissell, Tech. Coords. Lowland river and stream habitat in Colorado: a symposium. Colorado Chap., The Wildl. Soc. and Colorado Audubon Counc. Rutherford, W. H., and W. D. Snyder. 1983. Guidelines for habitat modification to benefit wildlife. Colorado Div. Wildl. Rep. 194pp. Schmidt, Sr., R. A. 1983. Management of cottonwood-willow riparian associations in Colorado. Colorado Chapt., The Wildl. Soc., Denver. 53pp. Schrupp, D. L. 1978. The wildlife values of lowland river and stream habitat as related to other habitats in Colorado Pages 42-51 in W. D. Graul and S. J. Bissell, Tech. Coords. Lowland river and stream habitat in Colorado: a symposium. Color~do Chapt., The Wildl. Soc. and Colorado Audubon Counc. Schultze, R. F., and G. I. Wilcox. 1985. Emergency measures for streambank stabilization: an evaluation. Pages 59-61 in R. R. Johnson, C. D. Ziebell, D. R. Patton, P. F. Ffolliott, R. H-.-Hamre, Tech. Coords. Riparian ecosystems and their management: reconciling conflicting uses. First North American Riparian Conf. U.S. Dep. Agric., For. Servo Gen. Tech. Rep. RH-120. Scott, T. G., and C. H. Wasser. 1980. Checklist of North American plants for wildlife biologists. The Wildl. Soc ,, \-lashington,D.C. 58pp. Sedgwick, J. A., and F. L. Knopf. 1986. Cavity-nesting birds and the cavitytree resource in plains cottonwood bottomlands. J. Wildl. Hanage. 50:247-252. �392 , and ---cottonwood --a~ 1987. Breeding bird responses to cattle grazing of a bottomland. J. Wildl. Manage. 51:230-237. , and 1988. Demography, regeneration, and future projections for in R. R. Sharitz bottomland cottonwood community, Pages Wildlife Symp. Savannah and J. W. Gibbons, eds. Freshwater Wetland and River Ecol. Lab. Symp. (In Press). Simons, D. B., R. M. Li, P. Lagasse, and R. T. Milhous. 1980 (1981). Proc: Workshop on downstream river channel changes resulting from diversions or reservoir construction. U.S. Dep. Inter., FWS/OBS 81-48. 348pp. Smith, B. 1980. The South Platte. Colorado Outdoors 29(5):26-31. Snyder, B., and K. Bovee. 1983. Water supply for riparian vegetation. Pages 27-31 in R. A. Schmidt, ed. Management of cottonwood-willow riparian associations in Colorado. Colorado Chap., The Wildl. Soc., Denver. Snyder, W. D. 1978. The bobwhite in eastern Colorado. Tech. Publ. 32. 88pp. Colorado Div. Wildl. 1983. Shrub thicket .establishment in Colorado's High Plains. 'Colorado Div. Wildl. Game Inf. Leaf. 108. 4pp. 1984. Lowland riparian cottonwood studies. Job Pr?g. Rep. Div. Wildl. Fed. Aid Proj. N-4-'R-2. Jan: 295-370. Colorado 1986~ and~. 'Dynamics of cottonwood regeneration. Job Prog. Rep. Colorado Div. Wildl. Fed. Aid Proj. 01-03-045. Apr: 435-454 and 455-473. 1987. Dynamics of cottonwood regeneration. Job Prog. Rep. Div. Wildl. Fed. Aid Proj. 01-03-045 (In Press). Colorado Svoboda, P. L., and W. D. Graul. 1980. Application of the species-ecosystem approach to the lowland riparian zone within the Tamarack Ranch Wildlife Area, northeastern Colorado. Unpubl. Res. Rep. Colorado Div. \.Jildl. 39pp. Swanson, G. A. 1979. The mitigation symposium: a national workshop on ,mitigating losses of Fish and Wildlife habitats. u.S. Dep. Agric., For. Servo Gen. Tech. Rep. fu~-65. 684pp. Swenson, E. A., and C. L. Mullins. 1985. Revegetating riparian trees in southwestern floodplains. Pages 135-138 in R. R. Johnson, C. D. Ziebell, D. R. Patton, P. F. Ffolliott, R. H. Hamre, Tech. Coords. Riparian Ecosystems and their management: reconciling conflicting uses. First North American Riparian Conf. U.S. Dep. Agric., Servo Gen. Tech. Rep. RM-120. Taylor, D. M. 1986. Effects of cattle grazing on passerine birds nesting in riparian habitat. J. Range Manage. 39:254-258. Tipton, R. J. 1939. Conserv. Board. San Luis Valley Project. Unpubl. Rep. Colorado \vater �393 Tubbs, A. A. 1980. Riparian bird communities of the Great Plains. Pages 419-433 in R. M. DeGraff, Tech. Coord. Management of western Forests and grasslands for nongame birds. u.s. Dep. Agric., For. Servo Gen. Tech. Rep. INT-86. u.s. Department of Agriculture. 1965. Silvics of forest trees of the United States. u.s. Dep. Agric., For. Servo Handb. 271. 762pp. u.s. Department of Agriculture. 1983. Pormant stock planting for channel stabilization. U.S. Dep. Agric., Soil Conserve Servo Tech. Notes: Biol. Note 22, Arizona. 4pp. u.s. Department of Interior. 1981. The Platte River ecology study. u.s. Dep. Inter., Fish and Wildl. Serv., Spec. Res. Rep., Northern Prairie Wildl. Res. Cent., Jamestown, N.D. l87pp. Williams, G. P. 1978. Historical perspective of the Platte Rivers in Nebraska and Colorado. Pages 11-41 in \0/. D. Graul and S. J. Bissell, Tech. Coords. Lowland river and stream habitat in Colorado: a symposium. Colorado Chap., The Wildl. Soc. and Colorado Audobon Counc. York, J. C. 1985. Dormant stub planting techniques. Pages 513-514 in R. R. Johnson, C. D. Ziebell, D. R. Patton, P. F. Ffolliott, and R. H. Hamre, Tech. Coords. Riparian ecosystems and their management: reconciling conflicting uses. First North American Riparian Conf. U.S. Dep. Agric., For. Servo Gen. Tech. Rep. RM-120. Prepared bYw1t1~~ Wildlife Researcher C ��JOB PROGRESS REPORT State of: Project: Colorado W-152-R Avian Research 3 Job ----- Work Plan: 21 Job Title: Sandsage - Bluestem Prairie Renovation Period Covered: Author: 01 January through 31 December 1987 Warren D. Snyder Personnel: Warren D. Snyder, Colorado Division of Wildlife ABSTRACT Above average precipitation and warm spring conditions combined. to promote excellent vegetation growth in 1987 within the Tamarack Prairie. However, the potential impact of these conditions for enhancing the quality of the height-density index (HDI) of vegetation will not be known until sampling is completed in early spring 1988. Grass-forb and sandsag~ (Artemisia filifolia) HDI's had not fully recovered by early spring 1987 from controlled burns conducted on sites 1-84 and 3-84. In contrast, grass-forb cover recovered dramatically and fully within 2 years following the 1985 burns (P <0.05) based on spring 1987 HDI sampling. The 1987 HDI of grass-forb vegetation was reduced significantly from 1986 pre-treatment sampling within only 1 of the 3 burns conducted in 1986. These data provide considerable evidence that post-burn precipitation received during the growing season was a major factor influencing grass response to fire within the Tamarack Prairie. Crown cover sampling in August 1987 within the 1984 burns could not detect major vegetation changes other than the reduction of sandsage. Little change among species ~ad occurred since the last (1985) sampling was conducte4.. The amount of dead vegetation and litter remained below pre-treatment levels. Sampling indicated the revegetation strips, seeded primarily to sWitchgrass (Panicum virgatum) in spring 1985, had attained HDI's approaching 2 dm by early spring 1987 within 2 growing seasons. This HDI was >4 times that of grass-forb vegetation within undisturbed prairie. These tracts continued to improve in quality during the 1987 growing season. AnalYSis also documented that tillage-herbicide renovation of sites previously interseeded to tall, warm-season grasses dramatically enhanced these interseeded grasses (p < 0.05). Sampling within a site sprayed with a contact herbicide in June 1985 to reduce the density of sandsage showed that standing dead sandsage still contributed substantially to overall HDI. Perennial forbs continued to increase in frequency after most were killed by the herbicide treatment. Major HDI quality reduction was noted within a tract burned in 1986 within the sandsage spray site. ��397 SANDSAGE-BLUESTEM PRAIRIE RENOVATION Warren D. Snyder _ P. N. OBJECTIVES Test renovation and revegetation techniques for increasing standing residual height-density of grasses, increasing the proportion of tall, warm-season grasses within the composition, and for reducing the quantity of sand sagebrush to <30% canopy cover in an ungrazed sandsage-bluestem prairie on the South Tamarack, South Platte Wildlife Area in northeastern Colorado. SEGMENT OBJECTIVES Monitoring of environmental and vegetation conditions and changes continued as follows: 1. Precipitation was monitored throughout the year supplementing electronic rain gauge data with information from nearby weather stations through the winter months. 2. Soil moisture accumulations, plant phenology, and weather were monitored primarily in spring and especially at time of controlled burns. 3. Visual obstruction (height-density) measurements were obtained on treatments and controls where applicable in late winter and/or early spring prior to green-up. 4. Crown cover, species composition, and frequency of occurrence measurements were obtained from mid-summer to early fall. 5. Photos of treatments and controls were taken in October. Data compilation and writing the annual job progress report was conducted during fall and winter 1987-88. METHODS Approaches used within this study were previously summarized (Snyder 1986a, 1986b, 1987) and are outlined in the Segment Objectives. Segment objective #4 was conducted only within the 1984 burn sites and their controls after being omitted there the previous year. pampling within the 1985 and 1986 burns and their controls will be resumed in 1988. RESULTS AND DISCUSSION Environmental Conditions Precipitation.--The 4 automatic precipitation recorders, placed on the Tamarack Prairie in February 1987, provided meaningful data from April through �398 October. Data for the winter months were obtained from nearby U.S. Weather Service Stations. Two of the recorders malfunctioned (mouse damage, loose wires) for brief intervals during the summer and the collector on 1 site was destroyed and another damaged by deer in fall. Rainfall amounts varied among recorders (distributed within the Prairie), especially from June through September, thus data were averaged to provide monthly totals (Fig. 1). April 1987 was especially dry, however, subsequent rains from May through August were average or above making total 1987 precipitation almost 25% above average and the highest recorded since study initiation (Fig. 1). Except for an extremely hot-dry interval during the last one-half of July 1987, rainfall and weather conditions were favorable for vegetation growth. Precipitation received in October 1986 and February 1987 (Fig. 1) enhanced soil moisture conditions going into the 1987 growing season. April through early July 1987 soil probe sampling revealed good soil moisture conditions throughout the spring - early summer interval (Table 1). Most probe indices averaged over 1 m in depth throughout the interval. Since sandy soils retain about 2.5-3.8 cm of moisture/0.3 m (1-1.5 in./ft.) (D. Smika, pers. commun.), the Tamarack Prairie soils contained >5 cm 2-5 in. of moisture through spring 1987. Table 1. Average soil moisture accumulations based on soil probe samples (m) during spring-early summer 1987, Tamarack Prairie, CoLorado, Location 1 Apr West Tamarack Prairie W. rain gauge Tr. 2-21 Central Tamarack Prairie M. rain gauge Burn 3-84 4-4 3-5 3-85 14-59 2-85 17-54 10-48 3-86 29 Apr 0.7'9- 2 Jun 1.18 0.95 2 Ju1 0.74 1.31 1.26 1.29 1.06 1.17 1.14 0.97 1.35 1.49 1.21 1.24 0.69 0.79 0.57 East Tamarack Prairie E. rain gauge 3-86 24-65 3E-630 Spray Reveg. 112 113 1.29 1.12 1.14 1.00 1.08 Mean 1.21 0.91 ax of 4 samples/site. 6 May 0.86 1.12 0.72 1.01 1.29 1.10 0.74 �399 20 D N s 0 .•...•. .s• 15 0 z S 0 D N 0 H 8 -c ~ ~ 10 u J ~ A J J J J s J A J J A S A ~ a. ~ A S D J M M M A A J 5 M M A A M o F F J Long-term ~ M A M J 1984 1985 1986 1987 Fig. 1. Monthly and accumulated annual precipitation (in.) from 1984 through 1987 in relation to the long-term average, Tamarack Prairie, Colorado. �400 Temperature and Phenology.--March 1987 was slightly colder than normal, however, April was exceptionally warm and dry (Fig 2). May-June temperatures also were above normal promoting phenologically early growth of Tamarack Prairie vegetation. Comparisons of phenological data in 1987 (Table 2) with data from previous years of study show growth was well ahead of that in 1984, when April was colder than normal but slightly behind that of 1985 and 1986 which were both phenologically advanced (Fig. 2). The graphic data (Fig. 2) are believed to closely depict the relative phenological growth of prairie vegetation among years. Controlled burns, planned by management personnel for portions of the Tamarack Prairie in 1987 were cancelled due to equipment problems and dry, warm weather in April that promoted early green-up. These burns have been rescheduled for spring 1988. Their postponement has not impacted this study as no vegetation sampling was planned. Vegetation Sampling and Evaluation on Burned Sites Height-Density Sampling.--The controlled burns conducted in 1984 reduced the visual obstruction (height-density-index = HDI) quality of both grass-forb and sandsage vegetation in subsequent years (Table 3, Fig. 3). Vegetation had not fully recovered on either burn 1-84 or 3-84 by early spring 1987 although the recovery within burn 1-84, dominated by grass, appeared better than that within burn 3-84 where sandsage was much more abundant. Sandsage remained more severely impacted by fire than grass on both sites. Grass-forb, sandsage, and combined vegetation in 1987 remained markedly below pre-treatment (1984) levels (p <0.05) within burn 3-84 (Table 3). Pre-treatment to 1987 differences were not detected within the 1-84 burn. If post-treatment (1985 and 1986) transect means were difference from those in 1987 on the 2 sites, sampling could not detect those differences (Table 3). Within the 3 spring 1985 burns, fire severely reduced the HDI quality of grass-forb vegetation during the first post-treatment (1986) sampling. However, by early spring 1987 grass-forb vegetation had recovered dramatically to exceed pre-treatment levels (Table 4, Fig. 4). The 1987 grass-forb HDI was markedly above 1986 levels (p < 0.05) when tested among transect means and site means using analysis of covariance, and among site means using a paired t test (Table 4.). In contrast, significant (!>0.05) recovery of sandsage from 1986 to 1987 was not detected and as a consequence combined vegetation likewise did not make a significant recovery either among transects or among sites. However, the dominance of grass-forb cover within samples is illustrated by comparable trends of grass-forb and combined cover (Fig. 4). Controlled burns conducted on 3 sites in spring 1986 did not reveal strong evidence that the height-density quality of grass-forb vegetation was markedly reduced following the first post-bur~ (1986) growing season (Table 5, Fig. 5). Some height-density reduction apparently occurred but a difference (p< 0.05) was detected only within the 1986-87 comparison among transects. In contrast, the height-density quality of sandsage was reduced both among transects and among sites (p< 0.05, Table 5). Combined vegetation, dominated by grass-forb cover, was reduced based on analysis among transects, but the reduction could not be detected when comparing changes among site means. �401 TIME BURNED 80 ~ · JUN JUN 70 · JUN JUN MAY 60 - ~ MAY ~ JUN ~ MAY MAY 50 APR · APR MAY ~ APR APR 40 APR MAR MAR MAR MAR MAR 30 Long-term 1984 1985 1986 1987 x Fig. 2. Accumulating monthly average temperatures (F) from March through June, 1984-87 in relation to that of the long-term X, as an index to vegetation phenology, Sterling, Colorado. - �~ o Table 2. N Phenological conditions of selected vegetation during spring 1987, Tamarack Prairie, Colorado. Species 15 Allium textile Artemisia filifolia A. ludoviciana Abronia fragrens Astragalus sp. Cymopteris montanus Evolvulus nuttalianus Lathyrus polymorphus Leucocrinum montanum Mentzelia nuda Opuntia humifusa Penstemon angustifolius Phlox andicola Physalis subglabrata Psoralea lanceolata P. tenuiflora Sphaera1cea coccinea Tradescantia occidenta1is Tragopogan sp. 5-7.6a Basal leafed Agropyron smithii Boutelua gracilis Calmovi1fa longifo1ia Muhlenbergia pungens Panicum virgatum Paspa1um stramineum Sporobu1us cryptandrus Stipa comata Cyperus sp. aVegetation height, cm. bE = early; F = full; L Apr 23 May 29 Jun 18 6 ---2 Headed 0.5 leafed 2.5-5 Headed 0.5 leafed Bloom Fully leafed Bloom 5 5 5-7.6 Fully blooming Bloom 15.2 L.b bloom E.b bloom Bloom 5 2.5 Budded 2.5-5 L. bud L. bloom Emerging E. bloom F.b bloom 7.6-10.1 Bloom Seed Bloom 2.5-5 H.b bloom L. bloom L. bloom 10.1-15 Budding E. bloom 2.5-5 5.1 Leafed 7.6-10.2 Dormant Dormant Emerging Dormant 5-7.6 2.5-5 late; M 12.7-15.2 7.6-10.2 2.5-5 2.5-5 2.5-5 Dormant Dormant Emerging 10.2-12.7 F. bloom medium. F. bloom 5-7.6 7.6-10 E. bloom L. bloom 30 E. bloom Bud L. bloom 5.1 20-25 7.6 20-30 F. bloom 12.7-15 5-7.6 10-12.7 15-20 5-10 10-15 30 10-12.7 10-15 15-20 30-50 10 15 E. head 5 5-7.6 15 10 15-25 25-38 Near bloom L. bloom Bloom 38-50 F. bloom F. bloom Seed 30-38 38-60 50-60 15-20 F. head �403 Table 3. Height-density (dm) means within 1984 burns and their controls during early spring 1984-87 intervals and! values from analysis of covariance testing among years, Tamarack Prairie, Colorado. Vegetation 1984 Year Burned 1985 1986 1987 1984 Control 1985 1986 1987 0.253 0.814 0.334 0.295 0.687 0.355 0.301 0.643 0.368 0.224 0.847 0.309 0.183 0.935 0.531 0.191 0.797 0.503 0.200 1.044 0.629 0.216 0.688 0.424 BURN 1-84 Grass/forb Sandsage Combined 0.256 0.856 0.372 0.134 0.313 0.162 0.232 0.358 0.256 0.208 0.526 0.258 BURN 3-84 Grass/forb Sandsage Combined 0.222 0.827 0.493 F-Va1ue comEarisons 0.021 0.121 0.047 0.106 0.356 0.201 0.077 0.286 0.157 Pre-burn to Eost-burn 84-87 84-85 84-86 Post-burn to post-burn 85-87 86-87 85-86 BURN 1-84 Grass/forb Sandsage Combined 47.5a 40.1a 66.1a 4.24 14.98a 10.51a 0.52 1.28 2.05 8.19a 8.06a 0.02 0.64 0.41 1.48 0.04 1.40 0.01 2.32 0.91 0.14 2.77 4.45 0.00 4.34 4.68 1.46 BURN 3-84 Grass/forb Sandsage Combined ap <0.05. 22.5a 114.1a 187.0a 5.96 35.31a 53.18a 6.62a 13.55a 9.22a �404 0.7 'C.~o\. •••••• ••• co'(\ eA.·· ,'" 0.6 ••••••• 0.5 ••••••••••••• _ 0.4 •• •• •• .... • ..•• •• •• •• • • ....• 1-84 Control ---------- , '-, ...•....•. ...•. 0.3 0.2 0.1 o 1984 1985 1986 1987 Fig. 3. Height-density (dm) of combined residual vegetation in early spring pre- (1984) and post-treatment intervals within the 1-84 and 3-84 burns, Tamarack Prairie, Colorado. �Table 4. Height-density (dm) means within 1985 burns and their controls during early spring 1985-87 intervals with statistical values of relationships.a Site 1985 Treatment 1986 1987 1985 Control 1986 Statistical values 1985-86 1985-87 1986-87 1987 GRASS-FORB 1 2 3 0.507 0.381 0.180 0.138 0.119 0.053 0.419 0.472 0.244 0.374 0.346 0.158 0.325 0.280 0.180 0.322 0.376 0.156 £1,31 Il,3 t 96.62b 49.82b 4.35b 1.59 0.70 -0.91 21.60b 9.92b -12.55b -_-- --- 95.85b 14.39b 0.38 1.04 2.71 2.71 -2.06 115.25b 33.3b 2.64 0.18 0.01 0.81 2.02 0.87 1.25 SANDSAGE 1 2 3 1.008 0.679 0.642 0.417 0.342 0.438 0.523 0.591 0.769 0.872 0.615 0.537 0.828 0.745 0.767 0.925 0.560 0.648 Il,31 £1,3 t - COl-IBINED 1 2 3 0.565 0.407 0.302 0.142 0.130 0.088 0.425 0.476 0.330 0.444 0.369 0.323 0.391 0.335 0.465 0.383 0.393 ·0.411 £1,31 £1,3 t - all,31 represents an analysis of covariance comparing trends among the 17 treatment and 17 control transects ignoring sites. £1,3 is an analysis among sites. t tests at 2 d.f. represent paired tests of year to year differences among sites. bCP<0.05). "'" 0 V1 �406 0.5 0.4 0.3 ••..... .§ ..... )04 ~ {/) 0.2 z til 0 I ~ H til :z: 0.1 o 1985 1986 1987 Fig. 4. Height-density (dm) of grass-forb and combined residual vegetation in early spring pre- (1985) and posttreatment (1986-87) within the 1985 burned and control sites, Tamarack Prairie, Colorado. �Height-density (dm) means within 1986 burns and their controls during early spring 1985-87 Table 5. intervals with statistical values of re1ationships.a . Site 1985 Treatment 1986 1987 1985 Control 1986 1987 1985-87 1986-87 GRASS-FORB 1 2 3 0.326 0.314 0.259 0.304 0.337 0.256 0.120 0.437 0.202 0.294 0.316 0.233 0.277 0.312 0.213 0.234 0.404 0.269 £.1,35 fl.,3 0.885 0.833 0.671 £.1,35 fI,3 t t 3.13 0.27 1.27 5.26b 0.88 1. 78 28.22b 38.96b 3.80 26.42b 51. 22b 4.79b 8.40b 0.91 1.47 8.71b 0.70 1.43 SANDSAGE 1 2 3 0.621 0.375 0.395 0.838 0.500 0.650 0.169 0.250 0.250 0.829 0.417 0.618 0.756 0.300 0.690 - COMBINED 1 2 3 0.421 0.315 0.272 0.470 0.339 0.281 0.127 0.434 0.204 0.345 0.317 0.298 0.342 0.331 0.323 0.325 0.413 0.357 £.1,35 fI,3 t - a1985 and 1986 were both pre-treatment years. !:.l,35represents analysis among the 19 treatment and 19 control transects ignoring sites. !:.I,3is an analysis among sites. t tests at 2 d.f. represent paired tests of year to year differences among sites. bp<0.05. .j:'- 0 "'-l �408 0.4 .••. _hi.r\ed 'leg. COl'''"' ••• _ surn • •• ~ - - __ •.•• ~........ -•••Co~ed -•••••••••••••••••••••••• 0.3 -~ ..... V g e· - ••••• cont. - ------~-------------;;:~~~ Grass-Forb - Burn ---_ Grass-Forb - Control - ~ ~ Ul i5 0.2 Q I !i: ~ H ~ 0.1 o 1985 1986 1987 Fig. 5. Height-density (dm) of grass-forb and combined residual vegetation in early spring pre- (1985-86) and post-treatment (1987) intervals within the 1986 burned and control sites, Tamarack Prairie, Colorado. �409 It is evident that grass-forb vegetation within the 1985 burns recovered faster and to a greater extent after fire than that within the 1984 burns. In addition, the HDI of grass-forb cover was not reduced as much during the first year following the 1986 burns when compared with previous burns. The reasons why are not conclusively evident at this time but several factors are suspect. Precipitation was considered a primary factor influencing vegetation recovery. Rainfall remained below normal from May through August 1984, and although accurate measurements were not obtained, it appeared more deficient in burn 3-84 than in burn 1-84. Precipitation in 1985 remained slightly below average (Fig. 1), especially during the growing season, but strong biases in amounts received among sites were not detected. In 1986, above average precipitation was received but biases in distribution were noted. The eastern and central portions of the Tamarack Prairie received greater amounts than the west portion in April and July (Snyder 1987). Thus, the 1985 and 1986 burns would have received more during the growing season than burn 1-84 where vegetative growth was observably suppressed in 1986. This is believed to be a major factor in the slow recovery of vegetation within burn 1-84. Burn 3-84 in the south-central part of·the Tamarack Prairie, received more rainfall than burn 1-84 in 1986 but a marked recovery in HDI was not noted, possibly due to more sandy soils and grass-forb suppression by sandsage. That grass-forb vegetation did not show as drastic a reduction 1 year following 1986 controlled burning, as previous sites had shown, is additional evidence that increas~d precipitation promoted rapid recovery. Phenologically, the 1984 burns were conducted much earlier than those in 1985 and 1986 (Fig. 2). Sandsage had begun only minor leafing in 1984 when it was burned and it resprouted and grew rapidly. Subsequent phenologically later burns in 1985 and 1986 appeared to severely suppress sandsage and sprouting and regrowth were much slower. Thus, sandsage in 1984 was much more competitive with grass-forb vegetation after the burns, especially in burn 3-84 where it was the dominant species (Table 6). Grasshopper defoliation of grass-forb vegetation was also severe following the fire in burn 3-84. Total annual precipitation was plotted in relation to grass-forb and combined vegetation among controls for the several burns in an attempt to see if a precipitation - HDI relationship existed. No relationship is evident. Snow compaction (lodging) of vegetation over winter is probably of equal or greater influence on the residual cover and resulting HDI. Additional data collection may provide insight into the influences of timing of burns and subsequent precipitation amounts on grass-forb height-density quality.· Until then, the effectiveness of controlled burning as a method for enhancing height-density quality of grass-forb cover in the Tamarack Prairie remains uncertain. Crown Cover, Composition, and Frequency of Occurrence.--The 1984 burns and their controls, which had not been sampled in 1986, were replicated in August 1987 using the metric belt transect system previously used. Sampling was not conducted within the 1985 and 1986 burns on their controls in 1987. Crown cover, composition, and frequency of occurrence for burns 1-84 and 3-84 differed (Tables 7, 8). Mean crown cover through all sampling intervals for the 2 burns and their controls also differed (Table 9). A review of primary species based on these analyses follows. �410 Table 6. Percent total height-density samples obstructed by sandsage at the 1984, 1985, and 1986 burn sites from 1984 through 1987, Tamarack Prairie, Colorado. Year burned 1984 1 3 1985 1 2 3 1986 1984 Treatment 1985 1986 1987 1984 Control 1986 1985 1987 19.4a 44.7a 13.3 26.2 12.7 38.1 15.8 38.3 16.7 46.3 15.9 51.5 12.9 50.8 13.6 44.0 x l1.6a 8.7a 26.3a 13.8a 1.4 4.8 9.1 4.8 5.5 3.4 16.3 7.1 13.9 8.7 42.5 18.6 13.2 11.8 48.6 21.3 10.0 9.1 51.7 20.1 X 32.1 0.9 8.4 15.5 30.9a 1.2a 7.5a 15.2a 13.2 1.3 3.1 6.4 9.6 1.3 17.0 11.0 13.7 3.1 23.3 15.7 13.9 1.9 21.9 14.7 Site 1 2 3 aDenotes burn treatment occurred. Blue Grama (Bouteloua graci1is).--Data within burn 1-84 (Table 9) showed a trend similar to that from the 1985 and 1986 burns (Snyder 1987) indicating that controlled burns potentially enhanced this species. However, from Summer 1984 through 1987 sampling treatment and control transects have shown nearly identical trends. Since blue grama is a short species, and often in partial understory to other vegetation, removal of overs tory residual by fire may have exposed more blue grama to visual contact increasing sampling tallies during the first growing season. Blue grama within burn 3-84 comprised a much smaller percentage of the total composition, which combined with a small number of transects, provides less confidence in findings which conflict with those during first-year intervals within other burns. Needle-and-Thread (Stipas Comata).--Findings from both burn 1-84 and 3-84 show trends similar to data obtained from the 1985 and 1986 burns (Snyder 1987) indicating this species was set back during the first growing season after fire. The first-year set back is understandable as this cool-season species had made growth to 15.2 cm in 1984 when burned, and to 35-40 cm (was beginning to head) in 1985 when burned. In 1986 needle-and-thread height was 15-25 cm when burned. Within all burned sites the species made a dramatic recovery during the 2nd growing season including greatly enhanced seed production. Changes from 1985 to 1987 within burn 1-84 were similar between burned and control transects, whereas within burn 3-84, control transect means increased over those in the burn (!<0.05, Table 9). No conclusions can be derived from these conflicting findings. Prairie Sanddropseed (Sporobolus cryptandris).--Transect data within burns 1-84 and 3-84 showed conflicting first-year trends (Table 9). These and other conflicting data leave uncertainty as to potential fire impacts on this species. Following additional sampling, a more detailed evaluation will be attempted. �411 Table 7. Crown cover (0.01-m2), species composition (%), and frequency of occurrence of vegetation within treatment and control transects on Tamarack Prairie burn 1-84 during August 1987. Treatment Vegetation Bare ground Dead vegetation Boute1oua gracilis Stipa comata Sporobo1us cryptandrus Ca1amovi1fa longifo1ia Andropogon ha11ii Agropyron smithii Aristida sp. Paspa1um stramineum Muh1enbergia sp. Ko1eria cristata Panicum virga tum Festuca octof1ora Cyperus & Carex spp. Artemisia fi1ifo1ia Opuntia sp. Mammi1aria & Echinocereus Ambrosia psi10stachya Artemisia 1udoviciana Tradescantia occidenta1is Phlox andico1a Evo1vu1us nutta1ianus Lathyrus po1ymorphus Penstomen angustifo1uis The1esperma megapotimicum Liatris punctata Mentze1ia nuda Lygodesmia juncea Psora1ea tenuif10ra Physalis subg1abrata Erigeron sp. Astragalus sp. Chenopodium album Plantago purshii Eriogonum annum Pepidium densif10rum Cryptantha sp. Croton texensis Euphorbia sp. Conyza canadensis He1ianthus sp. Crown cover Compo Control Freq.! occur. 1,217.0 1,873.0 726.5 598.5 248.0 485.5 149.5 2.0 2.0 26.32 21.69 8.99 17.59 5.42 0.07 0.07 97.78 100.00 95.56 95.56 55.56 6.67 4.44 1.0 3.0 56.0 -20.0 359.0 9.0 0.04 0.11 2.03 0.73 13.01 0.33 4.44 2.22 62.22 13.33 80.00 22.22 2.0 0.07 4.44 40.0 4.0 8.0 10.5 1.45 0.15 0.29 0.38 40.00 11.11 17.77 11.11 4.0 0.15 11.11 1.5 0.5 1.5 0.05 0.02 0.05 2.22 2.22 6.67 1.5 7.0 0.5 1.5 1.0 2.5 13.0 1.0 0.05 0.25 0.02 0.05 0.04 0.09 0.47 0.04 6.67 24.44 2.22 4.44 4.44 13.33 13.33 2.22 Crown cover Compo Freq.! occur. 929.0 2,236.0 638.0 707.5 163.5 578.5 117.5 7.5 5.0 6.0 23.76 26.35 6.09 21.55 4.38 0.28 0.19 0.22 100.00 100.00 91.11 100.00 48.89 8.89 4.44 4.44 49.5 18.0 150.0 12.5 0.5 19.5 10.5 76.0 5.5 6.0 1.84 0.67 5.59 0.47 0.02 0.73 0.39 2.83 0.21 0.22 33.33 28.89 77 .80 28.89 2.22 17.78 6.67 55.56 8.89 8.89 4.0 17.0 0.5 1.0 1.0 4.0 0.15 0.63 0.02 0.04 0.04 0.15 8.89 20.00 2.22 2.22 2.22 4.44 9.5 0.5 1.5 4.0 10.0 0.35 0.02 0.06 0.15 0.37 13.33 2.22 6.67 11.11 11.11 1.5 20.0 10.5 28.0 0.5 0.06 0.75 0.39 1.04 0.02 6.67 31.11 8.89 28.89 2.22 �412 Table 8. Crown cover (0.01-m2), species composition (%), and frequency of occurrence of vegetation within treatment and control transects on Tamarack Prairie burn 3-84 during August 1987. Treatment Vegetation Bare ground Dead vegetation Boute1ua gracilis StiEa comata Sporobo1us crYEtandrus Ca1amovi1fa longifo1ia Andropogon ha11ii Agropyron smithii Panicum virgatum Paspa1um stramineum Muh1enbergia sp. Ko1eria cristata Festuca octof1ora Cyperus & Carex spp. Artemisia fi1ifo1ia 0Euntia sp. Mammi1aria & Echinocereus Ambrosia psi10stachya Artemisia 1udoviciana Tradescantia occidenta1is Phlox andico1a Evo1vu1us nutta1ianus Lathyrus po1ymorphus Penstomen angustifo1uis The1esperma megaEotimicum HaE10pappus spinu10sus Mentze1ia nuda Asclepias sp. Erigeron sp. IEomoea 1eptoEhy11a Argemone intermedia Cirsium sp. Physalis subg1abrata Chenpodium album Plantago purshii Croton texensis Euphorbia sp. CrYEtantha sp. Conyza canadensis Amaranthus sp. Crown cover Compo Control Freq.! occur. 748.0 572.0 140.0 175.0 160.0 192.0 31.5 7.0 ·9.5 30.0 0.5 1.0 6.0 16.0 320.5 15.5 0.5 0.5 2.5 19.5 4.0 16.0 10.5 10.94 13.67 12.50 15.00 2.46 0.55 0.74 2.34 0.04 0.08 0.47 1.25 25.04 1.21 0.04 0.04 0.20 1.52 0.31 1.25 0.82 0.90 1.00 1.00 1.00 0.45 0.20 0.20 0.50 0.05 0.05 0.30 0.65 0.95 0.40 0.05 0.05 0.05 0.40 0.25 0.40 0.10 0.5 1.0 7.0 3.5 21.5 21.0 0.04 0.08 0.55 0.27 1.68 1.64 0.05 0.05 0.15 0.10 0.60 0.20 0.5 2.0 3.0 12.5 38.0 2.0 0.04 0.16 0.23 0.98 2.97 0.16 0.05 0.15 0.20 0.35 0.80 0.15 9.5 0.74 0.15 Crown cover ComE· Freq.! occur. 552.5 665.5 128.5 286.5 227.0 131.5 37.0 0.5 5.5 15.5 4.0 9.30 20.73 16.43 9.52 2.68 0.04 0.40 1.12 0.29 0.75 1.00 1.00 0.90 0.45 0.05 0.05 0.50 0.15 24.0 401. 5 16.5 1.74 29.05 1.19 0~85 1.00 0.30 2.5 17.5 0.18 1.27 0.10 0.45 4.0 7.0 1.0 0.29 0.51 0.07 0.15 0.15 0.05 15.0 1.09 0.30 19.0 11.0 0.5 1.0 1.37 0.80 0.04 0.07 0.50 0.05 0.05 0.05 1.5 3.5 11.5 8.5 0.11 0.25 0.83 0.62 0.05 0.30 0.35 0.30 0.5 0.04 0.05 �Mean crown cover (0.01-m2) and analysis of covariance Table 9. samples during pre-treatment (1984) alld post-treatment (1984-87) Vegetation Pre-tr 1984 Burned transects Jul Aug 1984 1985 Aug 1987 relationships for selected species, species groups, and covers within burn and control intervals, burn 1-84 and burn 3-84, Tamarack Prairie, Colorado. Pre-tr 1984 Control transects Aug Jul 1984 1985 F-value Pre treatment 1984 to Aug 84 Aug 87 AUIl 84 . Ju1 85 to 85 Aug 84 to 87 1985 to 87 136.3 151. 2 34.9 123.6 25.1 150.0 32.1 33.0 16.2 198.5 477.8 21. 6a 38.7s 3.6 0.3 1.5 1.1 17.3a 8.1a 9.8a 132.7a 211. o- 14.89a 19.26a 9.82a 5.72a 1. 54 3.16 0.0 0.11 17.22a 166.6a 87.05a 61.8 1)7.7 109.2 63.2 17.8 91.1 193.0 37.3 12.7 265.6 320.0 2.3 6.8a 0.4 0.3 0.1 0.4 23.9a 2.6 0.7 280.2a 90.0 0.43 0.95 0.71 4.15 0.40 3.18 0.07 0.24 18.00a 94.49a 31. 81 Aug 1987 BURN 1-84 Bouteloua Ilraci1is Stipa ~ Sporobo1us sp. Calamovilfa sp , Andropollon hal1ii Comb. warm season Artemisia fl1ifo1ia Perennial forbs Annual forbs Bare ground Dead vegetation 362.4 257.7 44.9 185.7 28.7 213.7 102.9 16.3 3.1 117.3 124.8 310.2 86.5 75.4 197.3 47.7 245.8 75.8 40.8 4.1 386.6 7.5 116.5 103.3 58.5 73.3 16.5 89.9 84.0 12.5 15.9 458.5 301. 9 155.2 127.9 53.0 103.7 31.9 136.3 76.7 15.4 6.0 260.1 400.2 510.3 187.0 34.4 194.8 21.8 216.9 63.5 27.6 1.6 104.4 91.7 325.1 183.2 36.9 191. 5 29.2 221. 7 67.2 28.8 50.6 91.3 233.9 92.6 125.7 18.4 95.6 9.7 105.8 49.5 15.9 2.4 231. 4 6Q1.4 16.42a 22.36a 2.50 1. 32 0.02 0.64 1.30 2.01 4.04 10.75a 11.16a 4.24 3.37 15.56a 7.ns 0.02 5.55a 14.lOa 5.99a 7.92a 0.82 1.54 2.72 5.17~ 0.16 2.56 1.95 2.08 13.00a 10.19a 2.33 4.37 0.37 0.03 0.05 1. 09 0.43 0.74 0.94 3.75 6.36° 5.928 6.86° 11.18a 1.39 2.42 9.00a 3.72 0.49 4.43 0.23 0.03 2.23 4.99 1.05 3.23 1.53 0.30 3.55 0.07 1.95 12 .10a 0.02 16.78 1.18 0.00 0.36 0.88 5.79 9.45a 0.32 7.89a 14.04 0.04 2.40 0.39 1.15 3.16 7.08a 10.42a 0.04 0.0 0.08 0.47 0.08 0.46 0.08 0.52 BURN 3-84 Boute10ua Ilraci1is Stipa comata Sporobo1us sp. Calamovilfa sp , Andropogon hallii Comb. warm season Artemisia fi1ifolia Perennial forbs Annual forbs Bare ground Dead vegetation ap ~ 0.05. bTncludes Ca1amovi1fa 166.4 132.7 110.0 97.6 17.7 128.0 242.4 102.1 2.6 202.9 121. 6 87.5 23.7 69.4 68.2 9.'1 89.4 116.6 85.5 14.3 706.2 31. 4 longifo1ia, 35.1 59.4 53.1 86.1 4.6 95.0 196.4 50.2 63.0 520.2 153.1 Andropogon 67.3 84.1 76.9 92.3 15.2 126.4 154.1 52.0 32.3 359.6 275.0 ha11ii, 1)1.2 169.9 130.4 120.7 15.6 155.7 224.4 61.4 1.6 201. 2 141. 9 93.6 87.1 71.0 72.0 9.4 90.6 293.4 43.4 9.9 189.2 337.0 Panicum virgatum, 25.7' 51. 7 41.8 61.8 5.5 70.0 250.3 27.1 17.8 273.8 460.8 snd Paspalum stramineum. .pI-' lJJ �414 Prairie Sandreed (Calamovilfa longifolia).--Controlled burns in 1984 occurred ahead of greenup of this species and no major stimulus by fire was noted (Table 9). Enhancement of sandreed within 1984 burns was not noted in subsequent years. These findings conflict with results from the 1985 and 1986 burns where later burns and improved post-burn rainfall potentially stimulated increased growth within burned sites (Snyder 1987). Although this species does not stand well over winter, its abundance and quality is a major determinant of the height-density quality of grass-forb vegetation within the Tamarack Prairie. Sand Bluestem (Andropogon hallii).--This less abundant species appeared to be enhanced by fire within burn 1-84 during the first growing season and increased seed head production was noted. However, no evidence of enhancement beyond the first year was noted in burn 1-84 and no evidence of any enhancement from fire was noted within burn 3-84. Burn 1-84 findings were supported by data from the 1985 and 1986 burns (Snyder 1987). Sandsage.--Fire, by removing the growth of preceding y~ars reduced the crown cover especially after the first growing season. Within the 1984 burns rapid regrowth was noted for the first 2 years following fire but by the end of the 4th growing season (1987), crown cover on both sites was less than in July 1985 (Table 9). Crown cover within control transects on both 1984 burn sites continued to decrease from August 1984. General observations support these data indicating sandsage within t~is range possibly cannot compete favorably with ungrazed grasses and is gradually deteriorating. More years of data are needed to substantiate this finding. Review of height-density indices could not discern evidence of deterioration within controls. However, HD1 samples were obtained when the plants were dormant and would not detect deterioration rapidly since most dead branches remain standing for several years. General observations indicate that fires conducted when sandsage is dormant or just beginning to leaf out may stimulate new growth, whereas later spring burns stress but seldom kill the species. Additional data indicate still later burning of sandsage can inflict greater stress and some mortality. As reported earlier (Snyder 1987) 2 fires burned north from Interstate 1-76 on 5 July 1986. Late August 1986 sampling of sandsage survival at 2 locations showed only 14 of 50 plants (28%) resprouted at the upper site and 42 of 50 (84%) resprouted at the lower site. These sandsage plants were resampled on 19 May 1987. On the upper site, 30 of 50 (60%) showed new growth and 86% of the lower 50 were alive. This indicates mid-summer burning can induce mortality and those surviving are extremely stressed and slow to recover. However, burns at that time of year also severely stress many grass species~ Perennial Forbs.--Although perennial forbs seemed to be enhanced during the first growing season after fire within burn 1-84, burn 3-84 data did not support this conclusion. Perennial forbs declined more within burn transects than within controls during the 1985-87 growing seasons. Frequency of occurrence comparisons show no marked declines of perennials either by species or by occurrence within the 1984 sites. More detailed analysis and evaluation is needed. Annual Forbs.--Annual forbs have been highly variable as to species and occurrence within the 1984 burns and their controls during the 3-intervals of �415 post-burn sampling. Pre-treatment 1984 sampling was not considered highly accurate due to over-winter deterioration of most annuals. Fires in 1984 apparently reduced or eliminated some species such as lambsquarter (Chenopodium album) but no marked changes over several growing seasons have been noted. Precipitation patterns and amounts along with temperatures are considered primary factors in establishment of annuals within the Tamarack Prairie. Canada horseweed (Conyza canadensis) was a typical example. Climatic conditions in spring 1987 apparently were involved in its dramatically increased occurrence throughout unburned sections of the Prairie. Bare Ground and Dead Vegetation.--The amount of dead vegetation and/or litter made considerable recovery from 1985 through 1987 within both 1984 burn sites along with marked declines in the amount of bare ground. These trends would be expected. However, it should be noted that complete recovery to pre-treatment levels has not been attained within either burn site. Revegetation Treatments Ti1lage~Reseeding.--The 19 revegetation strips established in 1985 continued to improve in quality. The average HDI in early spring 1987 among 12 random transects was 1.83 dm (range 0.98 to 3.07) with the lowest sample averaging twice the HDI of unmanaged vegetation on the Tamarack Prairie. Four of the transects had been previously sampled in late September 1986 when they yielded a mean HDI of 2.60 dm whereas 4 randomly selected nearby transects domi~ated· by need1e-and-thread and prairie sandreed averaged 1.04 dm. By spring the HDI o~ revegetation transects dominated by switchgrass had declined 19% to 2.11 dm whereas the HDI of native prairie transects had declined 65.7% to 0.36 dm. This illustrates the observably greater over-winter residual value of switchgrass compared to most other grasses native to the Tamarack Prairie. A comparison of crown cover changes within the revegetation strips from 1985 through 1987 revealed differences (Table 10) and illustrates the marked increase in crown cover of switchgrass and b1uestems (p <0.05). Switchgrass and bluestems comprised 77.2% of the composition of revegetation samples in 1987. It is anticipated that as switchgrass crown cover continues to increase within the revegetation strips in future years, grass height will decrease resulting in a gradual decrease in HDI. However, use of fire, and if necessary, tillage, should help sustain the HDI of these revegetation strips at levels much greater than surrounding vegetation. Annual precipitation was above average in 1986 and 1987 promoting excellent growth, whereas the impact of drought remains unknown. Tillage Renovation of Interseeded Tracts.--Efforts to evaluate vegetation changes within a tract that previously had been interseeded (1981-82) and partially renovated (disced, harrowed, treated with atrazine herbicide) in spring 1986 (Snyder 1987) continued in 1987. A portion of this tract had been included within the 1-84 burn and pre- and post-treatment crown cover sampling had indicated interseeded switchgrass and b1uestems responded more favorably than other grasses to fire (l <0.05, Snyder 1986). Subsequent evaluation of the tillage-herbicide renovation was not biased by the previous burn because the burned and unburned portions and their respective crown cover transects were bisected by the renovation and control tracts in spring 1986. �416 Table 10. Crown cover among 1985-87 sampling intervals from 12 random transects within the 1985 revegetation strips, Tamarack Prairie, Colorado. Vegetation or cover Bare ground Dead vegetation 1985 Subtotal N 804 212 1986 Subtotal N 812 156 1,016 Bouteloua gracilis Stipa comata Ca1amovilfa longifo1ia Sporobo1us cryptandrus 25 968 4 17 5 5 2 5 1 475 o o o o o o o o 3 5 1 6 1 1 1 8 10 1 1 2 8 3 833 5 3 7 1 21 61 772 3 2 22 Croton texensis Conzya canadensis Other annuals 2 37 6 Psoralea tenuiflora Sphaeralcea coccinea Physalis subglabrata Ipomoea leptophylla Oenothera sp. Cichorium intybus Evo1vu1us notta1ianus Liatris sp. Mentze1ia nuda o 75 400 8 303 183 17 10 311 Artemisia filifo1ia Opuntia sp. 45 99 226 79 Andropogon sp. Panicum virgatum o 57 128 64 11 650 39 1 217 Agropyron smithii Panicum capillare Cyperus & Carex Paspalum stramineum 278 372 40 66 126 1987 Subtotal N 12 76 2 5 18 76 o 8 31 �417 Pre-treatment and first-year post-treatment HDI varied (Table 11). Grass-forb HDI declined from spring 1986 to 1987 on both the tillage-herbicide and control sites, possibly because of deficient localized rainfall during the 1986 growing season. The tillage-herbicide treatment completely removed all standing residual in early spring 1986 so the recovery through the 1986 growing season would have to be dramatic to equal the HDI on the control by the following spring. Sandsage was more severely impacted by tillage (Table 11) and subsequently also influenced comparisons of combined vegetation. Table 11. Height-density (dm) means of grass-forb, sandsage, and combined cover from pre-treatment (1986) to post-treatment (1987) intervals between a tillage-herbicide and untreated control within a previously interseeded location, Tamarack Prairie, Colorado. Year Tillage-herbicide Control Grass-Forb 1986 1987 0.382 0.314 0.393 0.271 Sandsage 1986 1987 1.234 0.227 0.696 0.608 Combined 1986 1987 0.579 0.306 0.544 0.375 Crown cover sampling in August 1987 revealed that interseeded bluestems and switchgrass increased at a greater rate within the renovated transects than within controls from 1986 to 1987 (p <0.10, Table 12). Averages were much greater than those obtained during the 1985 pre-treatment sampling showing that the interseeded grasses were markedly enhanced by tillage-herbicide renovation. In contrast, sandreed, another deep-rooted native showed only modest increase from 1985 to 1987 (Table 12). Other native grasses with more shallow root systems were reduced by the treatment and showed little recovery from 1986 to 1987. The occupancy by perennial forbs was little affected by the tillage-herbicide treatment. �418 Table 12. Average crown cover/transect of selected species and species groups from pre-treatment (1985) to post-treatment (1986-87) intervals within tillage-herbicide renovation of an inter seeded site and its control, Tamarack Prairie, Colorado. Species/sEecies group Andropogon - Panicum Calimovilfa longifolia Bouteloua, Stipa, Sporobolus, etc. Perennial forbs Tillage-herbicide 1985 1986 1987 1985 Control 1986 1987 13.7 11.3 17.6 24.2 12.4 5.4 36.3 13.8 6.4 16.4 10.7 17.1 16.6 10.6 16.8 22.1 10.2 19.8 1.5 0.6 1.2 2.2 1.7 3.4 85-86 F Values 85-87 86-87 34.68a 61.94a 1.04 2.76 62.11a 66.70a 1.76 2.92 3.62b 1.78 2.87 1.56 ap < 0.05. bp<O.lO. Early spring 1987 height density sampling was conducted on 4 randomly selected tracts of the 30 renovated interseeded tracts treated in spring 1986 •. Their HDI's averaged 0.23, 0.26, 0.42, and 0.73. The highest index was obtained on the most easterly tract that had received greater precipitation during the 1986 growing season. Bluestem seed was harvested in fall 1987 from that tract leaving a tall stubble. Observations indicated all renovation tracts made considerable additional recovery during the 1987 growing season that potentially will be revealed in 1988 HDI sampling. Management personnel used tillage-herbicide renovation on 4 additional interseeded tracts within the northeastern part of the Tamarack Prairie in April 1987. Two large tracts adjacent to the property boundary were reseeded and subsequent rains promoted dense switchgrass stands while diluting the atrazine allowing dense stands of annual forbs to establish. One tract was mowed by management personnel in early summer to curtail weed growth. Strip Spraying of Sandsage Pre-treatment (May 1985) sampling of crown cover was completed within the large tract sprayed in June 1985 with herbicide to partially kill sandsage, however, pre-treatment HDI sampling was not conducted. Height-density sampling on 2 of the transects was conducted in early spring 1986 and on 11 transects in 1987. No marked HDI change had occurred within the 2 transects from 1986 to 1987 and sandsage, while dead, remained standing to contribute considerably to the visual obstruction during early spring intervals (Table 13). A smaller tract within the spray site was burned in spring 1986 and 4 transects were sampled for HDI during pre-burn (1986) and 1987 early-spring intervals. Analysis indicated major reduction of grass, sandsage, and combined vegetation following the treatment (_!: < 0.05, Table 13). �Table 13. Height-density (dm) means within the 1985 sandsage spray site during 1986 and 1987 post-spray intervals and pre-burn (1986) to post-burn (1987) intervals within a portion of the spray site burned in 1986, Tamarack Prairie, Colorado. Year Transects Grass-forb Sandsage Combined vegetation 1985 SANDSAGE SPRAY 1986 1987 0.-246 0.272 2 11 0.879 0.800 0.468 0.469 1985 SANDSAGE SPRAY - 1986 BURN 1986 1987 0.279 0.122 4 4 t 4.9 with 3 df 0.713 0_.333 0.404 0.129 t = 8.68 with 3 df Crown cover sampling within the sandsage spray site from May 1985 (pre-treatment) through 1986 and 1987 intervals revealed differences (Table 14). All major grasses except sandreed have shown crown cover increases following herbicide treatment. Live sandsage was almost totally eliminated although the attempt was for a partial reduction through strip spraying (Snyder 1986b). Perennial forbs were nearly eliminated in summer 1985 following the herbicide treatment. However, by spring 1986 considerable recovery was noted and this recovery continued into 1987 (Table 14). Perennial sweet pea (Lathyrus po1ymorphus) has been the primary species to recover and has dominated all sampling intervals. The number of species counted remained much reduced from pre-treatment sampling (Table 14). �420 Table 14. Crown cover (point frame) of vegetation and ground cover within 11 random transects during pre- (1985) and post-treatment (1986-87) May intervals within the June 1985 herbicide treated sandsage spray tract, Tamarack Prairie, Colorado. Vegetation/cover Bare ground Dead vegetation Perennial grass Boute1oua gracilis Stipa comata Sporobo1us cryptandrus Ca1amovilfa longifolia Andropogon ha11ii Muhlenbergia pungens Ko1eria cristata Aristida sp. Agropyron smithii Carex sp. Annual grass Bromus tectorum Festuca sp. Pre-treatment 1985 Post-treatment 1987 1986 402 1,425 529 1,324 536 1,083 142 194 34 51 148 510 46 65 2 3 9 1 312 425 106 53 2 3 6 3 1 3 5 70 117 4 116 7 Artemisia fi1ifo1ia (alive) A. fi1ifo1ia (dead) 436 112 1 259 287 Opuntia & Echinocereus spp. 44 50 62 5 18 114 55 4 1 9 77 Perennia forbs Ambrosia & Artemisia spp. Tradescantia occidentalis Lathyrus polymorphus Psoralea tenuif10ra Phlox andico1a Evo1vu1us nutta1ianus Abronia fragrans Cymopteris montanus Allium textile Mentzelia nuda Leucocrinum montanum Penstemon angustifo1ius The1esperma megapotimicum Annual forbs Croton texensis Chenopodium album Pepidium & Lesquerella sp. Cryptanthia sp. Tragopogan sp. Plantag purshii Salsola kali Unid. forbs 6 2 2 ° 22 123 1 6 1 1 6 1 1 1 4 1 12 1 2 2 3 24 3 2 1 2 1 1 1 �421 LITERATURE CITED Snyder, W. D. 1986a. Sandsage-bluestem prairie renovation. Job Progress Rep., Colorado Div. Wildl. Wildl. Res. Rep., Fed. Aid Proj. 01-03-045 (W-37-R). Apr.:475-498. 1986b. Sandsage-bluestem pralrle renovation. Job Progress Rep., Colorado Div. Wildl. Wildl. Res. Rep., Fed. Aid Proj. 01-03-045 (W-37-R). Apr.:499-525. 1987. Sandsage-bluestem prairie renovation. Job Progress Rep., Colorado Div. Wildl. Wildl. Res. Rep., Fed. Aid Proj. W-152-R. In Press. Prepared by W~ Warren D. Snyder Wildlife Researcher C JyM ��423 JOB PROGRESS REPORT Co'Lo'rado State of: Project: W-152-R Avian Research 21 Job Title: Response of Selected Wildlife Species to Aspen Silvicultural Practices Period Covered: Author: Job 4 --- Work Plan: 01 January through 31 December 1987 Suzanne M. Joy Personnel: C. E. Braun and R. W. Hoffman, Colorado Division of Wildlife (CDOW); R. T. Reynolds, u.S. Forest Service (USFS); R. L. Knight and S. M. Joy, Colorado State University (CSU) ABSTRACT In August 1987, the CDOW, USFS, and CSU entered into a cooperative study to investigate nest site characteristics and foraging behavior of sharp-shinned hawks (Accipiter striatus) and other accipiter hawks fn mature aspen and conifer habitats. A literature review was initiated in preparation for writing a study plan. A study plan was prepared, peer reviewed, and approved by the cooperating agencies. The study plan is presented as part of this report. ��425 RESPONSES OF SELECTED WILDLIFE SPECIES TO ASPEN SILVICULTURAL PRACTICES Suzanne M. Joy P. N. OBJECTIVES 1. Examine and evaluate existing literature on aspen-wildlife relationships to identify information needs pertinent to the development of a detailed study plan. 2. Interview selected personnel of the CDOW, USFS, BLM, and Louisiana Pacific Corporation concerning research needs within the aspen type. 3. Prioritize research needs and identify funding sources. 4. Prepare a detailed study plan on an approved research topic dealing with aspen-wildlife relationships. SEGMENT OBJECTIVES 1. Review literature on aspen ecology, aspen silvicultural practices, aspen-wildlife relationships, effe~ts of aspen manipulation on wildlife populations, and chemical and nutritional characteristics of aspen. 2. Prepare a detailed study plan on an approved research topic. 3. Prepare annual progress report. ��427 PROGRAM NARRATIVE (Research) State: Colorado Project Title: Study Title: Project Number: Avian Research Work Plan: W-152-R 21 Job: 4 Accipiter Nest Site Selection and Foraging Behavior of Sharp-shinned Hawks in Mature Aspen and Conifer Habitats A. NEED: Quaking aspen (Populus tremuloides) dominated forests are widely distributed throughout the Central Rocky Mountains. Pure aspen and mixed conifer-aspen stands are extensive and provide nesting and foraging habitat for numerous avian species. that occur in aspen forest types. DeByle (1985) listd 135 bird species In Colorado, the largest pure stands of aspen are on the western slope, where they often occur in close association with Engelmann spruce (Picea engelmannii)-fir (Abies spp.) forests (Flack 1976, Johnston and Hendzel 1985, Mueggler 1985). Aspen stands in Colorado comprise over 60% (1.15 of 1.78 million ha) of the commercial aspen area (i.e., trees suitable for industrial wood products) in the western United States and over 25% (1.15 of 4.58 million ha) of Colorado's commercial timberland (Green and Van Hooser 1983). Aspen is harvested for a variety of wood products including pulp, paper, particle board, f1akeboard, and matches (Wengert et ale 1985). Little is known about habitat requirements of avian species in aspen. Scott et ale (1980) listed 18 species of both primary and secondary cavity-nesting birds that nested in aspen trees. Crouch (1983:1), found that removal of mature aspen overstory "adversely affected cavity-nesters and other species requiring mature forest". He also �428 suggested that clearcutting "should benefit (wildlife) species needing sparsely vegetated areas and forest edges". Presently, over 70% (1,484 of 2,004 stands surveyed) of aspen stands in Colorado are in mature (70-120 years) and overmature (120+ years) age classes (W. Shepperd, pers. commun.). Management practices, such as clearcutting and removal of invading conifers, have been prescribed to encourage sucker regeneration and growth of younger stands (Jones et al. 1985). Before the effects of regenerative harvests on aspen avifauna can be fully evaluated, pre-harvest data on bird species and their habitat associations in aspen must be collected. Flack (1976) suggested that abundance of avian species in aspen is regulated by features of the vegetation and that, although some species use apsen habitat opportunistically, for others aspen may be critical. Salt (1957) illustrated the value of aspen when he found 3 times the bird biomass in aspen compared to 5 other vegetation types inventoried. North American accipiters exist sympatrically throughout much of the western United States (Reynolds et al. 1982, Moore and Renny 1983, Fischer and Murphy 1987). atricapillus), The northern goshawk (Accipiter gentilis Cooper's hawk (A. coperii), and sharp-shinned hawk (A. striatus) coexist during the nesting season by exploiting different habitats for nesting (Reynolds et al. 1982, Moore and Renny 1983) and by foraging for birds and mammals of different sizes and taxa (Storer 1966, Reynolds and Meslow 1984). Accipiter nest-site characteristics have been quantified in conifer-dominated forests (Reynolds et al. 1982, Moore and Renny 1983). Fischer and Murphy (1987) provided some nesting data on accipiters in predominantly deciduous forests, but they compared the relative use of 6 �429 habitat types (including aspen) available to the species. To date, no studies have focused on nesting habitat of accipiters exclusively using aspen forests. Proposed aspen management by clearcutting and removal of invading conifers emphasizes the need for these data. Habitats used for foraging by accipiters must contain a sufficient number of available prey, vegetative cover for protection and concealment while hunting, and trees of suitable densities and sizes to permit uninhibited flight. Accipiters forage in structurally diverse habitats, with considerable interspecific and intersexual variation in preference (Fischer and Murphy 1987). forests opportunistically Because accipiters appear to use available (given that they are sufficiently open to allow flight), Reynolds (1988) suggested that differences in habitats used for foraging are more closely linked to prey availability than tree species composition or habitat structure. Alterations in habitat structure and/or composition affect bird and mammal populations, which may in turn affect the reproductive performance and abundance of one or more accipiter (Reynolds 1988). Few data exist on use of space in aspen forest types by foraging accipiters. Therefore, to better predict the effects of aspen harvest, analysis of accipiter foraging behavior relative to prey availability in this forest type is needed. Distribution The latitudinal range of North American accipiters includes Alaska and the Northwest Territories, Canada south to Costa Rica, with interspecific variations in breadth and seasonality of distribution. Sharp-shinned hawks are the most migratory accipiter, breeding as far north as the Arctic Circle and wintering north of the equator (Guatemala and Costa Rica) (Mueller and Berger 1967a). Cooper's hawks occur �430 throughout temperate zones of North America, breeding from southern Canada south to the Gulf of Mexico and into Columbia, South America (Brown and Amadon 1968, Wattel 1973); winter range includes Costa Rica north to New England and British Columbia (Bull and Farrand 1983). northern goshawk is circumboreal. The In the nearctic, goshawks occur throughout much of Canada and Alaska south to New England, Maryland, Michigan, New Mexico, Arizona, and northern Mexico (Mueller and Berger 1967£, Brown and Amadon 1968, Wattel 1973, Scott 1987). Accipiters are classified as uncommon residents of Colorado (Bailey and Niedrach 1965:193, 199, 203) and references on their distribution are few. Sightings occur predominantly in western and northern areas of the state (Bailey and Niedrach 1965, Shuster 1976). Seasonal Activities and Movements Sharp-shinned hawks migrate from breeding to wintering ranges presumably in response to prey availability (Newton 1979). The exact timing varies annually, but appears to be initiated by southward movements in their avian prey (Clarke 1984). Sharp-shinned hawks in Colorado begin migration during late August and early September (R. Reynolds, pers. commun.). Mueller and Berger (1967~) reported that immatures migrate approximately 1 month before adults, supposedly to reduce competition with adults for declining prey resources. Goshawks are the least migratory of North American accipiters; migrations occur primarily in response to declining prey populations (Mueller and Berger 1967b). Little information exists on timing of migration by Cooper's hawks; however, Brown and Amadon (1968) noted they are often seen migrating with other hawks. �431 Northward migration for all accipiters begins in early spring. pairs reach breeding sites by March or April. Most Sharp-shinned hawks at high elevations may not arrive at nests until late May (Reynolds 1988). Onset of reproductive behavior is signaled by courtship flights in all 3 species (Bent 1937, Brown and Amadon 1968). Aerial displays and occasional calling serve to attract the female or male (Bent 1937, Brown and Amadon 1968). In all species, one or both sexes construct the nest. When both sexes participate, Brown and Amadon (1968) reported male Cooper's hawks contributed most to the nest, whereas Clarke (1984) found that female sharp-shinned hawks provided the majority of nest material. Sharp-shinned hawks commonly return to the same nest site each year, but build new nests (Bent 1937, Reynolds 1983). Goshawks show high nest fidelity, occupying the same nest for several years (Brown and Amadon 1968, Reynolds 1983). Cooper's hawks return annually to nest in the same section of forest, but generally build nests on new sites (Brown and Amadon 1968). All species are determinate layers. Eggs are primarily incubated by the female and the male delivers prey to her (Brown and Amadon 1968). Prey items are exchanged and plucked at distinct "plucking posts". Sharp-shinned hawks have the shortest incubation period (30-34 days), followed by Cooper's hawks (35-36 days) and goshawks (35-38 days) (Newton 1979). However, Reynolds and Wight (1978) found no difference among incubation periods of accipiters in Oregon (30-32 days). Fledging of sharp-shinned hawks occurs 21-27 days post-hatch, compared to 27-34 days for Cooper's hawks, and 33-43 days for goshawks (Reynolds and lvight (1978, Newton 1979, Clarke 1984). Variations in fledging within a species result from differences in the maturation rate of female and male �432 nest1ings--ma1es are smaller, develop more quickly, and leave the nest earlier than females (Reynolds and Wight 1978, Newton 1979), and uneven growth rates associated with infrequent prey deliveries and poor attentiveness by the female (Newton 1978). Females seldom leave the nest during the nestling period except to defend the brood (Gromme 1935, Fitch et ale 1946, Brown and Amadon 1968, Clarke 1984). Approximately 3 weeks post-hatch, females leave the nest to hunt or perch nearby (Brown and Amadon 1968). Sharp-shinned hawk families separate and individuals disperse from the nest site 3-4 weeks after fledging (Clarke 1984). Platt (1973) noted that dispersal occurred after the flight feathers of young were fully developed. Southward migration begins not long after families disperse. Although post-fledging periods are somewhat longer, the other accipiters exhibit similar behavior. Nest Site Characteristics Nest site is defined as the "area (immediately) surrounding the tree •••used by a nesting pair during an entire nesting season, exclusive of foraging areas" (Reynolds et ale 1982:126). Sharp-shinned hawks tend to nest closer to the ground than their congeners (Moore and Henny 1983, Fischer and Murphy 1987). sharp-shinned Canopy cover at nest sites is high for both (82-98%) and Cooper's hawks (83-95%), whereas, canopy cover is somewhat lower at goshawk nests (68-88%) (Moore and Henny 1983, Fischer and Murphy 1987). Because sharp-shinned and Cooper's hawks typically place nests within the lower portion of the nest tree crown, as opposed to below the canopy as goshawks frequently do (Shuster 1980, Moore and Henny 1983), these percentages may reflect the level of nest concealment each prefers. Topographic features associated with accipiter �433 nest sites include gentle to moderate slopes (0-30%) with northerly (continental United States) or southerly (Alaska) aspects; a source of water occurs either within the nest stand or adjacent to it (Reynolds 1988). Goshawk nests are flat, untidy structures that may exceed 1.5 m diameter (Farley 1923, Brown and Amadon 1968). twigs, small branches, and bark chips. bark and conifer sprigs. Nest materials include The nest bowl is layered with Cooper's hawks nests are built with small twigs and lined with dry deciduous leaves, s~rips of bark, and conifer needles and sprigs (Fitch et ale 1946, Brown and Amadon 1968). Sharp-shinned hawk nests are broad platforms (0.5-m diameter) constructed with freshly cut twigs and lined with flakes of wood or bark (Bent 1937, Clarke 1984). Breeding Success Sharp-shinned hawks lay a mean of 4.4 eggs/clutch (range = 3-6), Cooper's hawks, 4.1 eggs (range = 3-5), and goshawks, 3.2 eggs (range 2-4) (Platt 1976, Reynolds and Wight 1978, Clarke 1984). Hatching success is lowest for sharp-shinned hawks (64.4%), followed by Cooper's hawks (74.0%), and goshawks (81.2%) (Reynolds and Wight 1978). Productivity or nesting success is defined as the number of young fledged per nest attempt. Based on a review of 8 studies, Reynolds (1988) reported that sharp-shinned hawks produce a mean of 3.0 (range = 2-4) fledglings per nest attempt, while Cooper's hawks fledge 2.3 young (range = 2-3), and northern goshawks fledge 1.9 young (range = 1-3). Foraging Behavior, Home Range, and Daily Activity Rhythms Descriptions of accipiter hunting behavior range from "sit and wait" strategies (where the hawks passively wait for prey to pass within range before attacking) to more aggressive "short-stay perched-hunting" �434 strategies (where they actively search their surroundings for potential prey items while perched briefly between flying bouts) (reviewed by Fischer and Murphy 1986). Distance traveled while searching for prey partly determines home range size. Reynolds (1983, 1988) summarized estimates of home ranges for all accipiter species. Nesting sharp-shinned hawks traversed areas of approximately 380, 706, 460, and 100-1,248 ha in Alaska, Ontario, Oregon, and Wyoming, respectively. Cooper's hawks traveled over larger areas during nesting (range = 173 ha in Wyoming to 1,830 ha in New York), while home range estimates for nesting goshawks were even greater (up to 2,463 ha). In a comparative study of daily activity rhythms of accipiters, Fischer and Murphy (1986) found sharp-shinned hawks were most active in the early morning (0500-0900 MDT) and evening (1700-2100). Peak activity of Cooper's hawks occurred during late morning (0900-1300) and remained fairly high throughout the day. A single male goshawk exhibited activity peaks in the late morning (ca. 1100) and late evening (ca. 2000). Equating "activity" with locomotion and, later, foraging, Fischer and Murphy (1986) found that hawk activity patterns corresponded to activity periods in their prey species. Fischer and Murphy (1986) also investigated duration of flights and perching bouts in accipiters. Sharp-shinned hawks exhibited shorter flights (10-20 sec) and perching bouts (2-3 min) than congeners, which ranged from 15-20 sec, and 2-6 min, respectively, for goshawks and Cooper's hawks combined. Nesting and Foraging Habitat in Western Montane Forests Sharp-shinned hawks nest in dense, young, even-aged conifer stands (Hennessy 1978, Reynolds et a1. 1982, Moore and Henny 19~3, Clarke 1984), but occasionally use dense, young stands in deciduous forests (Platt �435 1973, Reynolds et ale 1982, Fischer and Murphy 1987). Nest concealment is important to sharp-shinned hawks because they are ineffective in defending against predators such as owls and others hawks (Reynolds 1988). Nesting sharp-shinned hawks were only found in mature aspen stands when islands of conifers (Picea or Abies spp.) occurred within the stand (Reynolds, pers. commun.). Layered foliage associated with the deep, conical crowns of conifers provide the necessary cover not provided by the narrow, high crowns of mature aspen trees (Reynolds 1988). Consequently, mature aspen stands devoid of conifers are prob~bly unsuitable for nesting sharp-shinned hawks. Small birds (10-30 g) associated with the forest canopy comprise over 95% of sharp-shinned hawks' diet (Craighead and Craighead 1956, Storer 1966, Clarke 1984, Reynolds and Meslow 1984). Flack (1976) demonstrated that aspen canopies in Arizona, New Mexico, Colorado, Utah, California, and Wyoming contained 41.4% (818 of 1,975 individuals) of all breeding birds occurring in the 4 nesting guilds he defined (canopy, shrub, hole, and ground). Birds occurring in the upper-most stratum of mature Colorado aspen forests were more abundant (52.7%; 178 of 338 individuals) than in all other nesting guilds compared. Thus, the upper canopy of mature aspen forests contains ample food for the smaller sharp-shinned hawk, making this forest type ideal foraging habitat. Northern goshawks nest in mature or old-growth coniferous or deciduous forests (Shuster 1980, Reynolds et ale 1982). They require large trees that can support their heavy nests and stands with sufficient space to allow unimpaired flight. Reynolds and Meslow (1984) reported that goshawks take vertebrate prey (x = 307 ± 364 g SD, ~ = 218) from lower portions of the vegetative �436 column. Goshawk foraging success is a function of their ability to detect, pursue and seize prey--a function of the amount of obstructing vegetation and cover available to the prey. Many mature aspen forests in the Rocky Mountains support numerous undergrowth plants (forbs, grasses, and shrubs) (Mueggler 1985) that form a dense, tall (1-3 m) understory that may eliminate foraging by goshawks in this vegetative zone. Cooper's hawks nest in dense stands of young to mature deciduous and coniferous forests (Reynolds et al. 1982, Moore and Henny 1983, Fischer and Murphy 1987). Although Cooper's hawks are less vulnerable to nest predation than sharp-shinned hawks, nest concealment provided by high vegetation density is still an important habitat feature for this species. Cooper's hawks take some prey from the lower vegetative column, bu~ are less restricted to this zone than goshawks (Reynolds and Meslow 1984). Thus, dense understories in mature aspen may limit Cooper's hawks to foraging in the mid-zones in this habitat type. Mature aspen forests appear to provide suitable nesting habitat, but inferior foraging habitat for goshawks, poor nesting and marginal foraging habitat for Cooper's hawks, and poor nesting (when conifers are absent) and good foraging habitat for sharp-shinned hawks. Preliminary field surveys for nests of accipiters (1986-1987) in Colorado (R. Reynolds, pers. commun.) appear to confirm these hypotheses (additional data are needed to support or refute initial observations). Consequently, harvest of aspen (particularly stands that are succeeding to conifers) could reduce sharp-shinned hawk foraging and nesting habitat. Aspen harvests may also reduce the amount of foraging habitat available to other accipiters by decreasing the abundance of conifer or mixed aspen-conifer stands. �437 Research efforts will concentrate on sharp-shinned hawks, because· this species appears to be more abundant and sensitive to apsen cutting, forages closer to its nest, and initiates nesting at a later date than congeners thus allowing post-winter travel conditions to improve on the study area. However, foraging behaviors and nest site characteristics of goshawks and Cooper's hawks will be measured when practical. Study Area Portions of Gunnison, Grand Mesa, and Uncompahgre National forests in Gunnison, Delta, Mesa, Montrose, and Ouray counties, Colorado were selected for study. Climax aspen stands and conifer communities exist in these forests between 2,750 and 3,200 m, allowing comparative studies between mature aspen and conifer habitats. Both large (>500 ha) and small (100-200 ha) stands of mature aspen and mature spruce-fir occur in this area. At lower elevations, the forests are broken by large (100-500 ha) clearings with low shrub, herb, and grass cover. are dominated by Gambel's oak (Quercus gambelii). South-facing slopes Active nests will determine individual study areas within these regions. Morgan (1969) listed several species of plants occurring in aspen understories of Gunnison County. Dominant forbs include meadow rue (Thalictrum fendleri), butterweed groundsel (Senecio serra), Porter lovage (Ligusticum porteri), white geranium (Geranium richardsonii), Barbey larkspur (Delphinium barbevi), sweet cicely (Osmorhiza obtusa), American vetch (Vicia americana), and white-flowered peavine (Lathyrus leucanthus). Prominent low growing plants include elk sedge (Carex geyeri), wild strawberry (Fragaria ovalis), yellow prairie violet (Viola nuttallii), and fringed brome (Bromus ciliatus). The shrub component consists primarily of snowberry (Symphoricarpos spp.) mixed with �438 red-berried elder (Sambucus racemosa) and chokecherry (Prunus virginiana) (R. Reynolds, pers. commun.). In Gunnison, Grand 11esa, and Uncompahgre National forests, Johnston and Hendzel (1985) reported plant associations between Engelmann spruce/elk sedge, Douglas-fir/kinnikinnik Arctostaphylos (Pseudotsuga menziesii/ uvaursi), and subalpine fir (Abies lasiocarpa)/elk sedge where aspen is seral to conifers. Common plant names were obtained from Weber (1976). B. OBJECTIVES AND HYPOTHESES: The objectives of 'the study are to (1) measure sharp-shinned hawk daily activity budgets in mature aspen and conifer overstory types, (2) compare sizes and taxa of prey delivered to sharp-shinned hawk nests in each overs tory type through the breeding season, (3) measure time between prey deliveries to nests in relation to breeding phenology and prey abundance in the 2 overs tory types, and (4) describe nest site characteristics of sharp-shinned hawks, Cooper's hawks, and goshawks found in each overstory type. Null hypotheses to be tested are: Ha: Perch and flight times, and periods of peak activity do not differ between aspen and conifer cover types and/or among incubation, nestling, and fledgling stages of breeding phenology. Hb: Mean time between deliveries, and composition (size and taxon) of prey delivered to nests do not differ between aspen and conifer cover types. Hc: Mean time between deliveries, and composition of food delivered to nests do not differ among incubation, nestling, and early fledgling stages of food provisioning. �439 C. EXPECTED RESULTS AND BENEFITS Knowledge of sharp-shinned hawk foraging behavior and nesting accipiters in mature aspen and conifer forests will allow wildlife managers to better predict effects of logging practices and other perturbations on accipiter population numbers. New guidelines for aspen management practices may be developed based on sharp-shinned hawk foraging efficiency and/or accipiter nest site selection in forest type. In addition, this study may identify factors limiting accipiter distribution in these habitats. Reynolds (1988) called for increased work on foraging habitat of Accipiter spp. to better identify the habitat association of these forest hawks. Data concerning foraging rates and activity budgets of sharp-shinned hawks should contribute to a better understanding of accipiter foraging behavior in different habitats and may stimulate new hypotheses for research. D. APPROACH 1. Review of relevant scientific literature. 2a. Field work will begin in May when sharp-shinned hawks arrive from wintering areas and begin breeding. Aspen and conifer stands containing structural characteristics typical of accipiter nest sites in western montane forests will be searched for active nests. If few or no active nests are located, presumed "unsuitable" stands will be searched. Active nest sites will be distinguished from inactive sites by the presence of a breeding pair, nest, plucking posts with recent prey remains, and/or feces. Active sites will determine individual study areas. 2b. known. Locations of 4 previously occupied sharp-shinned hawk nests are Attempts will be made to locate additional nesting pairs for �440 radiomarking in pure stands of aspen and conifer (target = 4 nests/cover type/year). Nests" in designated forest cover types ~2 km from another cover type will be used for radiomarking since foraging sharp-shinned hawks generally do not travel >1.5 km from the nest (Reynolds 1983). Visual observations (% per cover type) will be used to confirm that radiomarked hawks are foraging predominantly (~90%) in the designated cover type. 3. The following nest site variables will be measured: 1) elevation (topographic map), 2) slope (clinometer), 3) aspect at the nest tree (compass), 4) nest height (clinometer) and directional exposure (compass), 5) nest tree species, height (clinometer), diameter breast height (dbh) (metric tape), and crown depth (clinometer), 6) canopy cover at the nest in the 4 cardinal compass directions (densiometer), 7) distance of nest tree to water and nearest other cover type (metric tape), 8) forest cover type (predominant overstory species), 9) 2-3 dominant species of understory vegetation, and 10) tree density (trees/ha) (Reynolds et ale 1982, Moore and Henny 1983, Fischer and Murphy 1987). 4. Aerial photographs (1:24,000) of nest sites and proximate cover types will be used to plot nest locations with respect to surrounding cover types and aid in locating nests on subsequent visits. 5. Trapping of hawks will occur near nest sites with mist nests or dho-gaza traps made from mist nets following Bloom (1987). Data collected on captured birds will include sex (ascertained by tail length, wing chord, and weight measurements) (Mueller et ale 1976), stage of molt (Mueller et ale 1979, Clarke 1984), and age (categorized as hatching or 2nd-year juvenile, or adult) (Mueller et ale 1979). be banded with butt-end aluminum bands. Captured birds will �441 6. Because molt in accipiters overlaps with nesting, sharp-shinned hawks will be fitted with back-pack mounted (Dunstan 1972~, 1972~, Kenward 1985), lithium-powered transmitters made by Biotrack (Dorset, England). Transmitter packages weigh 4 g and 7 g which are approximately 4% of male and female body weights, respectively. expectancy is 10-12 weeks. Battery life Pulse rates will be regulated by mercury tilt-switches that change with activity (flight vs. perch). A maximum transmitting range of 2-3 km is expected depending on vegetation density and topography. 7. Two 3-element Yagi or whip antennas will be positioned at and near each nest site. Antennas will be elevated at least 3 m above the ground and will be used simultaneously or alternately to monitor hawk activity. At the end of an activity sampling period the antennas will be disassembled and reassembled at the next nest site. 8. telemetry: The following measures of activity will be obtained using 1) mean duration of perching bouts, 2) mean flight duration, 3) percent time spent in flight, and 4) periods of peak activity. 9a. Radio-monitoring of male sharp-shinned hawk activity will begin soon after marking and continue until the young fledge. Females will be radiomarked and monitored when they begin to assist the male with prey provisioning, generally 3-4 weeks post-hatch. 9b. Nest sampling frequency will depend partly on the number of nesting pairs radiomarked (goal = 4 nests visited/week). Sexes will be monitored for an equal number of systematically chosen days. Sampling will occur between 0500 and 2100 MDT, and consist of 3 systematically chosen 10-minute periods of continuous monitoring each hour (focal-animal sampling) (Altmann 1974). �442 9c. Fischer and Murphy (1986) visually monitored sharp-shinned hawks fitted with back-pack mounts and determined that fast-pulse signals of ~5 seconds between pulses corresponded to prey handling or preening behavior [defined as stationary activity bouts (SAB)], and signals of >5 seconds corresponded to flight in this species. length should correspond to perching. designations will be tested. A slow pulse of any Duration of said activity The mean value obtained for periods of stationary activity will distinguish SAB from flight. 9d. Analysis of activity patterns will be conducted using multivariate analysis of variance (MANOVA). Differences among cover types, sex, breeding phenology (days from nest initiation), time of day, nests within a cover type, and all interactions will be evaluated for dependent variable (1) mean duration of perching bouts, (2) mean duration of flying bouts, and (3) percent time in flight. The number of data points needed at each level to achieve a specified MANOVA power, given desired differences among means and significance levels, cannot be calculated ~ priori because data similar to the above were not found in the published literature. However, approximately 288 lO-minute sampling periods should be achieved for each nest by the end of the 1st field season. 10. Diurnal periods not used for radio-monitoring will be spent collecting and identifying prey remains and casts, collapsing and reassembling antennas, and traveling between sites. lla. A nest site observer will record frequency of prey deliveries and prey species. Prey consumed between nest visits will be reconstructed, identified, and counted by matching uneaten remains (remiges, retrices, and bills of birds, and skull fragments, fur, and �443 feet of mammals) (Reynolds and Meslow 1984). To avoid overestimating species abundance in the diet, regurgitated pellets (casts) and pluckings will be collected from plucking areas and nests at the beginning and end of each visit. In addition, prey items delivered to nestlings in each cover type will be simultaneously recorded by time-lapse cameras (2 cameras per cover type) that will run from dawn to dusk for 3 consecutive days each week. frame/min. Cameras will be mounted at nest-height and film 1 These data will confirm observed prey deliveries and augment delivery data for nests not observed. Mean weights for prey species will be obtained from the literature. lIb. Mean time between deliveries, and grams of prey delivered per minute (dependent variables) will be calculated for sex, cover type, stage of breeding, time of day, brood size, and nests within a cover type (independent variables). Differences among the independent variables will be evaluated for the dependent variables using a MANOVA model, the power of which cannot be calculated ~ priori because similar data are unavailable from the literature. Two-tailed student's t tests will be used to test for differences among mean size of prey delivered to nests in the 2 cover types (Zar 1984). Differences among mean prey size within cover type will be evaluated using chi-square contingency tables (Zar 1984). Each nest site will be observed for approximately 6 16-hour periods during the 1st field season. 12. The U.S. Forest Service will provide extensive data on abundance and distribution of bird and small mammal populations in the study areas. Systematic sampling using variable circular plot (Reynolds et ale 1980) and mark-recapture methods will provide quantitative estimates of avian and small mammal species composition and abundance in �444 aspen, conifer, and mixed aspen-conifer cover types, respectively, during the breeding seasons of 1988 to 1989. Twelve study plots have been selected in the 2 overs tory types (9 in aspen, and 3 in conifer). Bird sampling will occur from June to mid-July followed by small mammal sampling between mid-July and September. 13. Compile data, analyze results, and prepare progress and final reports. E. TIME SCHEDULE: Approach Date 1 September 1987 - December 1989 2a, 2b, 3, 4, 7, 10, lla May 5, 6 May - July 1988, 1989 8, 9a, 9b, 9c June - August 1988, 1989 9d, lIb, 12 September - April 1988, September - August 1988, 1989 December 1989 13 F. September - December 1987, 1988, 1989 PERSONNEL ASSIGNMENTS: Person Days 1987-88 Suzanne M. Joy Graduate Research Assistant Richard W. Hoffman Wildlife Researcher 220 11 1988-89 Suzanne H. Joy Graduate Research Assistant Richard W. Hoffman Wildlife Researcher 220 11 1989-90 Suzanne M. Joy Graduate Research Assistant Richard W. Hoffman Wildlife Researcher 132 11 �445 G. H. ESTIMATED COSTS: 1987-88 $14,084.00 1988-89 $14,598.00 1989-90 $ 8,000.00 SUPERVISION AND COOPERATION: Project Leader: Richard W. Hoffman, Wildlife Researcher, CDOW Principal Investigator: Academic Supervision: Collaborator: Cooperation: I. Suzanne M. Joy, Graduate Research Assistant, CSU Richard L. Knight, Assistant Professor, CSU Richard T. Reynolds, Research Wildlife Biologist, USFS USDA Forest Service LOCATION OF WORK: Study Headquarters: Department of Fishery and Wildlife Biology, Colorado State University, Fort Collins, Colorado 80523 Field Investigation: Forests. Gunnison, Grand Mesa, and Uncompahgre National Living quarters at West Muddy Guard Station, Gunnison National forest. Location of accipiter breeding pairs will determine individual study areas. J. RELATED FEDERAL AID PROJECTS: None LITERATURE CITED: Altmann, J. 1974. Observational study of behavior: sampling methods. Behaviour 49:227-267. Bailey, A. M., and R. J. Niedrach. 1965. Denver Mus. Nat. Hist., Denver, CO. Bent, A. C. 1937. Birds of Colorado. Vol. 1. 454 pp. Life histories of North American birds of prey. U.S. Natl. Mus. Bull. 167. 409 pp. Part 1. �446 Bloom, P. H. 1987. Capturing and handling raptors. Pages 99-123 in B. A. G. Pendleton, B. A. Millsap, K. W. Cline, and D. M. Bird, eds. management techniques manual. Raptor Natl. Wildl. Fed., Washington, D.C. 420 pp. Brown, L., and D. Amadon. Vols. 1 and 2. 1968. McGraw-Hill Book Co., New York, NY. Bull, J., and J. Farrand, Jr. North American birds. Clarke, R. G. Eagles, hawks and falcons of the world. 1984. 1983. 944 pp. The Audubon Society field guide to Alfred A. Knopf, Inc., New York, NY. 784 pp. The sharp-shinned hawk (Accipiter striatus vieillot) in interior Alaska. M.S. thesis, Univ. Alaska, Fairbanks. Craighead, J. J., and F. C. Craighead, Jr. 1956. 130 pp. Hawks, owls and wildlife. Stackpole Co., and Wildl. Manage. Inst., Harrisburg, PA and Washington, D.C. 443 pp. Crouch, G. L. 1983. Effects of commercial clearcutting of aspen on understory and wildlife habitat values in southwestern Colorado. For. Servo Res. Paper. DeByle, N. V. 1985. Winokur, eds. States. Dunstan, T. C. RM-246. Wildlife. Aspen: U.S. Dep. Agric., 8 pp. Pages 135-152 in N. V. DeByle and R. P. ecology and management in the western United U.S. Dep. Agric., For. Servo Gen. Tech. Rep. RM-119. 1972a. A harness for radio-tagging raptorial birds. Inl. Bird Band. News 44:4-8. 1972b. Radio-tagging falconiform and strigiform birds. Raptor Res. 6:93-102. Farley, J. A. 1923. Breeding of the goshawk in Massachusetts. Auk 40:532- 533. Fischer, D. L., and J. R. Murphy. 1986. Daily activity patterns and habitat use of coexisting Accipiter hawks in Utah. Unpubl. ms. 28 pp. �447 ----- , and 1987. Utah. Unpubl. ms. Foraging and nesting habitat of Accipiter hawks in 33 pp. Fitch, H. S., B. Glading, and V. House. nesting and predation. Flack, J. A. 1976. America. 1946. Observations on Cooper hawk Calif. Fish and Game 32:144-154. Bird populations of aspen forests in western North Ornithol. Monogr. 19. Green, A. W., and D. D. Van Hooser. Mountain states. 97 pp. 1983. Forest resources of the Rocky U.S. Dep. Agric., For. Servo Resour. Bull. INT-33. 127 pp. Gromme, o. J. 1935. Wisconsin. Hennessy, S. P. The goshawk (Astur atricapillus atricapillus) nesting in auk 52:15-20. 1978. Ecological relationships of accipiters in northern Utah--with special emphasis on the effects of human disturbance. thesis. Utah State Univ., Logan. Johnston, B. C., and L. Hendzel. M.S. 66 pp. 1985. Examples of aspen treatment, succession, and management in western Colorado. Serv., Rocky Mtn. Reg., Lakewood, CO. U.S. Dep. Agric., For. 164 pp. Jones, R. J., R. P. Winokur, and W. D. Shepperd. 1985. Management overview. Pages 193-195 in N. V. DeByle and R. P. Winokur, eds. and management in the western United States. Aspen: ecology U.S. Dep. Agric., For. Servo Gen. Tech. Rep. RM-119. Kenward, R. E. 1985. Raptor radiotracking and telemetry. I. Newton and R. D. Chancellor, eds. Pages 409-420 in Conservation studies on raptors. Paston Press, Norwich, England. Moore, K. R., and C. J. Henny. 1983. Nest site characteristics of three coexisting accipiter hawks in northeastern Oregon. Morgan, M. D. 1969. Nat. 82:204-228. Raptor Res. 17:65-76. Ecology of aspen in Gunnison County, Colorado. Am. MidI. �448 Muegg1er, W. F. 1985. Vegetation associations. and R. P. Winokur, eds. United States. Aspen: ----- , and ecology and management in the western U.S. Dep. Agric., For. Servo Gen. Tech. Rep. RM-119. Mueller, H. C., and D. D. Berger. hawk. Pages 45-55 in N. V. DeByle 1967a. Fall migration of the sharp-shinned Wilson Bull. 79:397-415. 1967b. ----- Some observations and comments on the periodic invasion of goshawks. Auk 84:183-191. and G. Al1ez. goshawks. Age and sex variations in the sizes of Bird Band. 47:310-318. _____ , and 1979. shinned hawks. Newton, I. 1976. 1978. nestlings. 1979. Age and sex differences in size of sharp- Bird Band. 50:34-44. Feeding and development of sparrowhawk Accipiter nisus J. Zool., Lond. 184:465-487. Population ecology of raptors. Buteo Books. Vermillion, SD. 399 pp. Platt, J. B. 1973. Habitat and time utilization of a pair of nesting sharp- shinned hawks (Accipiter striatus velox). Univ., Provo, UT. 1976. M.S. thesis, Brigham Young 41 pp. Sharp-shinned hawk nesting and nest site selection in Utah. Condor 78:102-103. Reynolds, R. T. 1983. Management of western coniferous forest habitat for nesting accipiter hawks. U.S. Dep. Agric., For. Servo Gen. Tech. Rep. RM-I07. 1988. States. eds. The status of Accipiter populations in the western United Pages 000-000 in B. Pendleton, K. Steenhof, and M. N. Kockert, Proc. Western raptor management symposium and workshop. Wildl. Fed., Washington, D.C. (in press). Nat1. �44~ ----- , and E. C. Mes1ow. 1984. Partitioning of food and niche characteristics of coexisting Accipiter during breeding. ----- , and H. M. Wight. 1978. Distribution, density, and productivity of accipiter hawks breeding in Oregon. ----- , E. C. Mes1ow, and H. M. Wight. Accipiter in Oregon. ----- , for estimating bird numbers. 1957. 1987. 1980. A variable circular plot method Condor 59:373-393. Field guide to the birds of North America. J. A. Whelan, and P. L. Svoboda. management. Nesting habitat of coexisting Condor 82:309-313. Geograph. Soc., Washington, D.C. ----- , 1982. An analysis of avifaunas in the Teton Mountains and Jackson Hole, Wyoming. Scott, S. L. Wilson Bull. 90:182-196. J. Wi1d1. Manage. 46:124-138. J. M. Scott, and R. A. Nussbaum. Salt, G. W. Auk 101:761-779. Nat1. 464 pp. 1980. Cavity nesting birds and forest Pages 311-324 in R. M. DeGraff and N. G. Tilghman, eds. Management of western forests and grasslands for nongame birds. U.S. Dep. Agric., For. Servo Gen. Tech. Rep. INT-86. Shuster, W. C. 1976. Northern goshawk nesting densities in montane Colorado. West. Birds 7:108-110. 1980. Rockies. Storer, R. W. Northern goshawk nest site requirements in the Colorado West. Birds 11:89-96. 1966. accipiters. Watte1, J. Sexual dimorphism and food habits in three North American Auk 83:423-436. 1973. Geograpic differentiation in the genus Accipiter. Ornitho1. Club, Bull. 13, Cambridge, MA. Weber, W. A. CO. 1976. 479 pp. Rocky mountain flora. Nuttall 231 pp. Colo. Assoc. Univ. Press, Boulder, �450 Wengert, E. M., D. M. Donnelly, D. C. Markstrom, and H. E. Worth. utilization. Aspen: 1985. Wood Pages 169-180 in N. V. DeByle and R. P. Winokur, eds. ecology and management in the western United States. U.S. Dep. Agric., For. Servo Gen. Tech. Rep. RM-119. Zar, J. H. 1984. Cliffs, NJ. Prepared by Biostatistical analysis. 718 pp. __ ~=-=-()..,-:.(J..l._-:-:--'-'V--1~-L..&_'..;_. Q~~~ tTu Su~ H. Joy Graduate Research Assistant Wildlife Researcher Prentice-Hall, Inc., Englewood �JOB PROGRESS REPORT Colorado State of: Project: 01-03-045 : Job Avian Research 5 Work Plan: 21 Job Title: Evaluation of Habitat Quality on Conservation Reserve Lands in Eastern Colorado Period Covered: Author: 01 July through 31 December 1987 Warren D. Snyder Personnel: L. Benson, D. Bowden, C. Braun, L. Groshans, and W. Snyder, Colorado Division of Wildlife ABSTRACT The general distribution of Conservation Reserve (CR) i~ eastern Colorado through the first 4 signups was obtained and a·map was prepared. Field inspections revealed that the vast majority of the CR acreages were planted to temporary cover crops in 1987 and seeding began on these in November 1987 and will continue through April 1988. Among fields in southeastern Colorado seeded in fall to spring 1986-87, poor to fair grass stands were evident in fall 1987 and annual forbs dominated providing fair to good brood and winter cover on most. Planning and design for field sampling continued in cooperation with personnel of the National Ecology Center of the USDI Fish and Wildlife Service. ��453 EVALUATION OF HABITAT QUALITY ON CONSERVATION RESERVE LANDS IN EASTERN COLORADO Warren D. Snyder P. N. OBJECTIVES To determine the distribution and quantity of Conservation Reserve land in eastern Colorado in relation to the distribution of selected wildlife species; to evaluate the quality of the vegetation on these lands for selected wildlife species; to measure the response of selected wildlife species "to Conservation Reserve using existing annual surveys, and to evaluate the impact of the Colorado Division of Wildlife's cost share program on cover quality. SEGMENT NARRATIVES 1. Obtain the"legal description or plotted map location of all tracts signed into the Conservation Reserve Program through February 1987 in all eastern Colorado counties. Names of landowners, acreages, CR practice applied, seeding schedule, and other pertinent information will be obtained. 2. Design sampling scheme. for enacting stratified random sampling based on the intensity of commitment to CRP per county, the quality of the cropland, and the grasses seeded into the tract. 3. Evaluate habitat quality within selected CRP tracts (potentially using Natural Resource Inventory tracts as a basis for sampling) as Colorado's portion of a national monitoring program using appropriate habitat suitability index models for selected wildlife species. 4. Measure the response of selected wildlife species to conservation reserve using existing annual population surveys. 5. Monitor the height-density and general composition qualities of vegetation within a stratified random sample of CRP tracts in relation to randomly selected proximal controls. 6. Monitor the quality of tracts cost-shared by the CDOW (for enhancement of cover quality) with proximal tracts not cost-shared. 7. Show relationship of CRP distribution to that of selected wildlife species using computer mapping facilities. 8. Compile and analyze data, and prepare progress reports. RESULTS AND DISCUSSION District offices of the USDA Soil Conservation Service (SCS) within 23 eastern Colorado counties were contacted in spring and early summer 1987 to obtain information on the distribution of Colorado Reserve (CR) lands in eastern �454 Colorado. Some SCS personnel already had the distribution tallied by location or plotted on county maps whereas it was necessary to extract the essential data from individual files at some offices. The data were compiled per township and plotted on a map of eastern Colorado (Fig. la and lb) based on CRP as a percent of the total land per township. As the Figures illustrate, the majority of the CR lands were located within the southeastern quarter of Colorado in locations traditionally considered marginal for dryland farming and for farm game wildlife. The data summarize findings from the first 4 CRP sign-up intervals. Seven counties, located in southeastern Colorado (Fig. 2) had reached their respective quotas of eligible CRP acreage by the fourth sign-up. Two additional sign-up intervals have occurred since then but data are not yet available as to CR increases within counties that had previously not reached their quotas. A comparison of CR land in relation to total cropland/township would provide meaningful data but a source for cropland/township could not be found. Following discussion of study design and sample size with D. C. Bowden, consulting statistician, eastern Colorado was stratified into 7 units (Fig. 3). A random list of CR fields was obtained within each stratum for use in monitoring vegetation quality. The location and conditions within some of these tracts in extreme southeastern and northeastern parts of Colorado were obtained in fall 1987. A new random list of sample fields was provided to the Colorado Division of Wildlife by personnel of the National Ecology Center (NEC) of the USDI Fish and Wildlife Service. This list includes a minimum of 68 contracts and probably >100 fields so probably will preempt most of the previously obtained random sample. It is part of a coordinated effort to evaluate the Conservation Reserve for wildlife on a regional and national basis. Field data collection procedures, developed for the Southern Great Plains Region which includes Colorado, will be supplemented where deemed necessary to obtain data deemed pertinent to Colorado. Key indicator wildlife species for which sampling will be directed include the ring-necked pheasant (Phasianus colchicus), western meadowlark CSturnella neglecta), and cottontail (Sylvilagus spp.). Field inspections in fall 1987 indicated that the majority of the CR fields had been planted to a sorghum cover crop in summer 1987 and would be seeded to perennial grasses from November 1987 through April 1988. Sorghum cover was primarily of sterile varieties that varied by field and locality in height, density, and standability. Some fields provided good over-winter cover to pheasants and other wildlife but most would be classed as fair to marginal in protective quality. Up to a third or more of the CR tracts in the southeastern corner of Colorado had been seeded to permanent (perennial) cover. Side-oats (Bouteloua curtipendula) appeared to be one of the primary species being seeded. Seeded grass stand establishment ranged from poor to fair in most inspected fields and as would be expected, had not attained significant growth of cover value during their first growing season. Wild annuals dominated in all fields and varied widely in species, height, and density primarily among fields. Some provided excellent brood and winter habitat for pheasants, whereas others were sparse and/or short. Some had been mowed so provided little over-winter cover. Precipitation in southeastern Colorado had been generally favorable for grass stand establishment and for growth of annual weeds in 1987. �455 o '"0 ~ ~ o U o rl c (l) 1..; ~ til (l) rU ;C ~ l.; o c C 'rl ~ ~ (l) til (l) 0:: C o rU +l -rl til (!J :> ~ o C u c ~ o :s o -rl ~ til ~ ~ ~ o -rl �~ V1 (J"\ SYMBOL 0 • • - PERCENT CRP < N 1.0 1.0 - 2.4 2.5 - 4.9 5.0 - 9.9 10.0 - 14.9 15.0 - 19.9 20.0 - 24.9 25.0 & > Fig. 1bo Distribution of Conservation Reserve in southeastern Colorado through February 1987. �457 Fig. 2 Reserve Percent of eligible cropland Program by county in eastern committed Colorado, to the spring Conservation 1987. �458 WELD l F~g. 3 Strata deliniated for selection Conservation Reserve. tracts in eastern of random Colorado, samples 1987. of �459 Many of the CR tracts within southeastern Colorado were surrounded on one or more sides by shortgrass rangeland. During cover crop and early stand stages when they are dominated by annuals they will retain diversity. However, in future years when they revert to perennial grasses interspersion of covers will decline. Most of these fields are extremely marginal for farming so as they are reverted to perennial cover, wind erosion should also decline markedly. Monitoring of pheasant densities using crowing count indices was conducted, primarily by management personnel in spring, 1987 and census data were compiled for future use. Compilation of historic pheasant crow-count data also was initiated. However, few crowing censuses have been conducted in recent years so 1987 data will be the primary pre-CR data available. Maps of the breeding bird survey (BBS) and mourning dove (Zenaida macroura) census routes in eastern Colorado were acquired and historic BBS data were provided by Sam Droege (USDI Fish and Wildlife Service) and Clait Braun. Prepared by "UJ~ ~ warren D. Snyder Wildlife Researcher C ��JOB PROGRESS REPORT State of: Colorado Project: W-152-R Avian Research Work Plan: 22 Job Title: Avian Research Publications Period Covered: Author: Job 1 01 January through 31 December 1987 Clait E. Braun Personnel: C. E. Braun, K. M. Giesen, M. A. Graham, E. J. Hernandez, J. W. Hupp, N. J. Kitzmiller, O. B. Myers, T. E. Remington, M. Schroeder, Colorado Division of Wildlife ABSTRACT Publications accomplished under this job in 1987 are: Braun, C. E. 1987. Current issues in sage grouse management. Assoc. Fish and Wildl. Agencies 67:134-144. 1987. Are sage grouse cyclic? Worksho p 15:18. Proc. West. Trans. Bien. West. States Sage Grouse Giesen, K. M. 1987. Evaluation of aerial and ground transects to inventory lesser prairie-chickens in southeast Colorado. Proc. Annu. Conf. Central Mountains and Plains Sect., The Wildl. Soc. 32:Abstract. 1987. Evaluation of aerial and ground trasects to inventory lesser prairie-chickens in southeast Colorado. Proc. Prairie Grouse Tech. Conf. l7:Abstract. 1987. Habitat use and preference by Columbian sharp-tailed grouse in northwest Colorado. Trans. West. States Sage Grouse Workshop 15:20. 1987. Colorado. Hovement and habitat use by Columbian sharp-tailed grouse in Proc. Prairie Grouse Tech. Conf. l7:Abstract. Graham, M. A., and C. E. Braun. 1987. Changes in territory size of whitetailed ptarmigan. J. Colorado-I-lyomingAcad. Sci. 19(1):20. Hernandez, E. J. sage grouse. 1987. Fire in the big sagebrush type: potential impacts on Trans. Bien. West. States Sage Grouse l-lorkshop15:14. �Hupp, J. W. 1987. A test of the relationship between sage grouse lipid reserves and courtship behavior. Trans. Bien. West. States Sage Grouse Workshop 15:19. 1987. Winter distribution, foraging ecology, and habitat use of sage grouse in the Gunnison Basin, Colorado. Trans. Bien. West. States Sage Grouse Workshop 15:10. Kitzmiller, N. J., and C. E. Braun. 1904-1985. J. Parasit. 73:679. 1987. In memoriam, Robert M. Stabler, Myers, O. B. 1987. Sagebru~h fertilization: will sage grouse respond? Trans. Bien. '\~est.States Sage Grouse Workshop 15:11. Remington, T. E., and C. E. Braun. 1987. Carcass composition and energy reserves of sage grouse during winter. Condor 90:15-19. Schroeder, M. 1987. to 1ek location. Prepared by Movement of female greater prairie-chickens in relation Proc. Prairie Grouse Tech. Conf. 17:Abstract. t/dz ~ ~C~l-a~i~t~E~.~B~r-a-u-n--~~~~~~-----Wildlife Research Leader �JOB FINAL REPORT Colorado State of: Project: W-152-R (01-03-045) Avian Research Work Plan: 24 Job Title: Roost Site Selection of Whooping Cranes at Fall Migration Stopover Sites Period Covered: Author: : Job 1 01 January 1986 through 31 August 1988 Tanya M. Shenk Personnel: R. J. Behnke, D. A. Hein, B. Van Horne, Colorado State University; C. E. Braun, J. K. Ringelman, Tanya N. Shenk, Colorado Division of Wildlife ABSTRA.CT Roost site selection by whooping cranes (Grus americana) was investigated during the 1986 and 1987 fall migration periods at traditional stop~ver sites (TSS) and non-traditional stopover sites (NTSS) in Colorado. Common characteristics among 35 TSS roosting wetlands in the San Luis Valley (SLV) included areas of water depth <30 cm, extensive horizontal visibility, and close proximity to feeding sites, loafing areas, and similar wetlands. Other characteristics that showed no consistency among TSS roosting wetlands included wetland size, emergent vegetation composition and density, specific conductivity of water, and distances to power lines, fences, and potential disturbances such as roads and residences. Three roost sites were used by 2 Grays Lake whooping cranes at NTSS in eastern Colorado. One NTSS was used during fall 1985,.a second NTSS was used by the same whoping crane during fall 1985 and 1986 •. Characteristics common among 3 NTSS roosting wetlands included areas of water depth <30 cm, isolation from disturbance, and close proximity «2 km) to feeding sites, loafing areas, and similar wetlands. There was no consistency among the lITSS wetlands for percent of horizontal visibility or vegetation composition and density. Mean length of stay for 16 whooping cranes in the SLV was 37.3 days (SD = 8.6). Males and females stayed in the region for the same length of time, however, variance of stopover duration was greater for males than females (F = 16.63, ~ = 0.003), primarily due to the broader range of departure dates for males. Sexes did not differ in the distances flown from one activity site to the next. Distances flown from the roost sites to feeding sites were greater than distances flown from feeding sites to loafing areas. Greatest site fidelity was exhibited for roosting complexes. Site fidelity did not differ among feeding sites, loafing areas, and roost sites within a roosting complex. Most (65%) loafing areas were within roosting complexes. The �non-roost complex loafing sites, located nearer feeding sites than roost complexes, were used when greater distances were flown from roost site to feeding site (~ = 0.04). Behavior of a female subadult whooping crane was described at a NTSS (eastern Colorado) and a TSS (SLV) during the 1986 fall migration. Time budget data were collected on this crane during all diurnal activity periods at the NTSS, and during feeding periods at the TSS. The frequency of alert behavior increased and foraging time decreased when feeding alone at the NTSS compared to behavior when feeding with a flock at the TSS. Daily activity patterns at the NTSS were consistent with those of the whooping cranes of the Grays Lake flock using TSS. �465 ACKNOWLEDGMENTS This project was supported by the Colorado Division of Wildlife (CDOW) and Colorado State University. Funding was provided by the CDOW through Federal Aid in Wildlife Restoration Project 01-03-045, 16700. Special thanks are extended to Dr. J. K. Ringelman, CDOW, for the assistance, counsel, and encouragement given me throughout all aspects of the project. I am especially grateful for his patience and the instructive criticisms received during preparation of this thesis. Appreciation is expressed to Dr. D. A. Hein, Professor of Wildlife Biology, his thoughtful suggestions and comments concerning professional ethics will serve as guidelines in future decisions. I also thank my graduate committee members Dr. B. Van Horne, Assistant Professor of Biology, for her assistance in study design and thesis preparation and Dr. R. J. Behnke, Professor of Fishery Biology, for his unique, insightful comments on this study and ecology as a whole. Dr. C. E. Braun, CDOW, was instrumental in supporting and funding this project through to completion. I am grateful to Or. R. C. Drewien and W. M. Brown for sharing their extensive field knowledge of crane behavior, particularly whooping cranes of the Grays Lake flock. I will remember the 1986-88 graduate student cohort and faculty with a fond smile and sense of kinship. To share with them the experience of the satisfaction of accomplishments and the inevitable �466 disappointments of field work, course work, thesis preparation, and personal growth was one of the most rewarding facets of my degree. May we continue to play hard and work hard as biologists and as individuals. My parents, sisters, brother, nieces, and nephews unconditionally supported me with their love and faith. reach my goals. thank them for helping me �467 Table of Contents ABSTRACT OF THESIS ...........•....•.•.•....•.•..•••••.•......••.•... Pace iii ACKNOWLEDGMENTS ........•.•..••.••.•..••.......•..•..•....•••••••...... v Chapter 1. ROOST SITE SELECTION OF WHOOPING CRANES IN THE SAN LUIS ----VALLEy, COLORADO....••.•..••......•.•.•..••..............• 1 Summary .................•....................•..........•........ 1 Study 5 Area Methods ...................•...................................... 6 Resul ts ...........•......•.•.•.................•..•.. Discussion Literature 2. .................•.......................•...........• 38 A .•...••.....•••••....••..........••.•....••..•.••..... 41 WHOOPING CRANE BEHAVIOR AT TRADITIONAL AND NON-TRADITIONAL FALL MIGRATION STOPOVER SIT£S Summary. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. Study 9 27 Cited ..........•....•.........••................•...• Appendix Chapter " Area .•..........•........................................ 48 . 48 50 Methods ........•............................................... 51 Resul ts ..................•..................................... 53 Discussion 60 Literature ..................•.............•.................... Cited ••......•.........................•.....•...... 65 Append i x B...............................•.........•........... 68 Appendix 89 C..............................................•...... �468 Table of Contents (cont/d) Paae Chapter 3. ROOST SITE SELECTION OF WHOOPING CRANES AT NON-TRADITIONAL FALL MIGRATION STOPOVER SITES IN EASTERN COLORADO 95 Summary 95 Study Area 97 Methods '.' Resul ts 97 98 Discussion 103 Literature Cited 108 Appendix D Chapter 4. ; 110 MANAGEMENT IMPLICATIONS 114 �4b~ CHAPTER 1 ROOST SITE SELECTION OF WHOOPING CRANES IN THE SAN LUIS VALLEY, COLORADO SUMMARY Roost site selection invesfigated Luis Valley and loafing cranes (Grus americana) during the 1986 and 1987 fall migtation (SLV), Colorado. for roosting wetlands. Habitat characteristics Juxtaposition period. of individual Common characteristics included wetland conductivity and potential cranes during horizontal that showed no consistency size, emergent vegetation of water, disturbances of stay for 16 whooping sites, was investigated each visibility, to feeding sites, loafing areas, and similar Other characteristics specific feeding among 35 roost sites included areas of water depth < 30 cm, extensive close proximity in the San were evaluated of roost sites, whooping was periods areas as a factor in roost site selection using daily locations activity by whoo~~~g wetlands. among roost sites composition and distances and and density, to power lines, such as roads and residences. fences, I~ean length in the SLY was 37.3 days (SO = 8.6). cranes Males and females stayed in the region for the same length of time, however, of stopover females departure variance (£ = 16.63, £ = 0.003), dates for males. from one activity duration was greater for males than primarily due to the broader range of Sexes did not differ in the distances site to the next. Distances flown flown from roost sites �470 to feeding sites were greater to loafing areas. complexes. Greatest Site fidelity than distances site fidelity was exhibited sites, located roosting complexes. complex. sites for roosting did not differ among feeding areas, and roost sites within a roosting areas were within flown from feeding sites, loafing Most (65%) loafing The non-roost complex loafing nearer feeding sites than roost areas, were used when greater distances site (£ = were flown from roost site to feeding 0.04). Suitable roosting, stopover feeding, to complete rivers, migration a combination in good condition. roosting and feeding A primary delineate migration objective remaining (Whooping information (Whooping considerable about their ecolDgical studies Aransas-Wood as critical migration Team 1980). migration the potential periods crane recovery (Lingle plan is to stopover sites for Before additional habitat is needed on the characteristics Crane Recovery Despite suitable Crane Recovery areas can be designated specific thereby eliminating of the whooping cranes wetlands, throughout sites for cranes during stopover and protect this species corridor, of for whooping Unfortunately, and streams have been altered or destroyed cranes' Several which includes and loafing sites, is necessary whooping 1987). habitat, however, more of stopover sites Team 1980). research on whooping cranes, little is known requirements during migration (Howe 1987). have been conducted on the migration ecology Buffalo (AWB) population (Allen 1952; Johnson 1980; Lingle et al. 1984, 1988; Howe 1987; Ward and Anderson These studies were limited by too few observations of the and Temple 1987). each season, the �471 isolation of stopover sites (Johnson that was biased by the distribution 1981), or a scarcity of marked 1987). studies No intensive requirements migration marked respects ecology. have been conducted cross-fostered activity population through Colorado and Bizeau sandhill Wildlife and Derrickson cranes Refuge, NWR, New Mexico 1981). these cranes usually site for several and comparison of daily stay, and (2) cranes and the Rocky Mountain (Grus canadensis 1974, 1978, Kauffeld tabida) migrate grounds Idaho to their wintering (Drewien and Bizeau The San Luis Valley spring and fall migration spending of duration. use the SLV from mid-February November, and (3) effects stopover flock of whooping near Bosque del Apache is the major traditional (1) of each bird to each fall in route from their breeding Grays Lake National on sites and loafing and among birds over an extended of stopover of greater crane flock. exist to document: Additionally, (1) observation within The cross-fostered whooping (2) site fidelity behavior. located allowing: patterns documentation Erickson cranes; on individual weeks thereby (Howe on the migration feeding a season and over years; remain at a centrally (Ellis et al. since all birds of this flock are from the roost to specific areas used by individual disturbance data this flock better lends itself to studies For example, roost sites within locational within the population (Melvin et al. 1983) the opportunities flight distance 198Q), of trained observers individuals of the experimental Yet in several and Temple grounds 1974, (SLV) of Colorado stop for these cranes 1981, Brown et al. 1987). (Drewien Both species to early May and from late August more time in the Valley near to mid- than in either their summer �472 or winter ranges (Drewien and Bizeau 1974, Kauffeld 1981, Brown et al. 1987). Cross-fostered occasionally whooping Prog. Rep. Whooping Similar associations behavior larger wetlands migrate Crane Egg Transplant occur at roost sites in the SLV. may be needed to accommodate cranes as individuals and Derrickson 1981). or in pairs, families, reared flock have also been observed of sandhill cranes, pairs, families, 1981, Johnson whooping species. whooping and Temple 1980). cranes migrating sightings, Investigation and roosting with flocks and less frequently as and Derrickson et al. (1976) reported cranes in 14% of their requirements may be similar for the 2 of roost site selection provide of both naturally-reared may cranes and normally (Allen 1952, Erickson Archibald cranes could therefore Specifically, cranes within the naturally- migrating with sandhill hence migration requirements Such gregarious or small flocks (Erickson most often as individuals and larger groups 1-20). Naturally-reared than sandhill However, whooping (R. a flock, and isolation with added flock vigilance. are less gregarious or cranes Experiment may effect their roost site selection. be less important whooping as individuals, in groups of 2 or 3, within a flock of sandhill C. Drewien, roosting cranes usually migrate by the cross-fostered information on the ecological and cross;fostered whooping crane flocks. The objectives and biological whooping characteristics cranes juxtaposition of this study were to: (1) quantify the physical of roost sites used by cross-fostered in the SLV, and (2) relate roost site selection of habitats during the fall stopover used by individual period. to the cranes of this flock �473 STUDY AREA The study was conducted in and around the Monte Vista-Alamosa within the SLV of Colorado. of Colorado's The SLV is the largest 4 intermountain parks. the west by the San Juan Mountains Cristo Mountains, (Ryder 1951). and southern-most The Valley, which is bordered on and on the east by the Sangre de is 161 km long and 81 km across at its widest point The Rio Grande, flowing through the central parts of the valley in a southeasterly Alamosa and Conejos rivers, relatively NWR direction, is joined here by the its main tributaries. flat surface that rises gradually the north, east, and west. Elevations and southern The basin has a toward steep mountains to range from 2100 m at the south end of the Valley to about 2400 m at the base of the mountains (Hopper 1968). The high elevation produces a climate and rain shadow effect of the San Juan Mountains classified annual precipitation as "cold desert" (Anderson ranges from < 18 em in the central 1965). Mean part of the valley to 25 cm around the edges of the valley floor (Ryder 1951). Despite the arid climate, table and numerous (Anderson 1965). impervious artesian wetlands Most important sub-strata that are valuable are also associated The principle husbandry. Alfalfa oats predominate; industry and marshes, but oxbow ponds with the SLV river systems. is irrigated and irrigated to waterfowl are large tracts of flooded hay and pasture lands, as well as drain ditches and sloughs create a high water farming and cattle and sheep pasture grasses, other crops such as lettuce, potatoes, cabbage, peas, and spring wheat are grown in lesser amounts. barley and cauliflower, The alkaline and �474 saline conditions caused by the high water table restrict to only the most alkaline-tolerant The Monte Vista National shrubs and grasses. Wildlife south of Monte Vista in the western production habitat plant growth Refuge (MVNWR), located 3.8 km portion of the SLV, is managed area for waterfowl. The 5758-ha refuge also provides for cranes, waterfowl, and other birds during migration, winter habitat for waterfowl leucocephalus). Artesian control water and pump wells located throughout The Alamosa Alamosa. This 4520-ha wetlands and is bordered and and bald eagles (Haliaeetus levels and maintai~ and waterfowl. as a roosting the refuge and loafing areas for cranes NWR (ANWR) is located 1.9 km southeast refuge is composed of natural of riverbottom on the west by the Rio Grande River. METHODS Criteria information cranes for habitat evaluations concerning habitat requirements (Allen 1952; Johnson Howe 1987; Ward and Anderson measure were based on published 1987). Variables (1) distances as fences and power lines and to potential optical measured maps were consulted. with a portable from aerial profiles and farm buildings. and techniques to potential disturbance photographs derived each wetland. perpendicular 2) Specific conductivity meter. and/or topographic from transects established Water depth was measured distance to shoreline used to hazards such sites including If these objects were < 180 m, an range finder was used to measure distance, topographic whooping and Temple 1980; Lingle et al. 1984, 1988; them were as follows: roads, houses, of migrating otherwise conductivity 3) Wetland maps. of water size measured 4) Water depth at 10-m intervals along each transect; was recorded across then for water depths of �475 10, 20, 30, and 40 cm. Contour lines at these 4 water depths were drawn to create a water profile for each wetland. visibility divided measured with a 3-m high vegetation 5} Horizontal profile board that was into O.S-m sections, with the first 0.5 m section sub-divided into 0.2S-m intervals increments (Nudds 1977). Viewing points were set at 10-m along the 20-cm water depth contour of the wetland. viewing point, a 90-m transect was established shoreline perpendicular or stand of dense emergent vegetation, was measured. Habitat evaluations From the to the and percent visibility were conducted on 35 roost sites over 2 years. Whooping November cranes were observed in the SLV from 28 October through 8 1986 and from 23 September arrival, departure, and daily activities , recorded only during 1987. cranes during roosting, periods were obtained binoculars. through 21 November of individual Daily locations morning feeding, 1987. Dates of birds were of individual whooping loafing, and evening feeding from a vehicle with a 30x spotting Average length of stay was calculated males and compared with a t-test; variances scope and 7xSO for females and were compared with an F- test. Individual cranes were initially site, then the direction noted. at the morning roost of flight to the morning feeding site was Birds were located on feeding sites either by searching them from roads or by following habitat type to another. an individual loafing site. an individual for bird as it flew from one The location of feeding sites was recorded quarter section of the field. watching identified by Loafing areas were best located by bird fly from the morning feeding site to the Evening roost sites were located either by following an �476 individual driving bird from the evening feeding site to the roost or by into a roost area near sunset when birds were first flying in to roost, then observing The sequence which birds flew into that site. of diurnal habitat use by cranes (Ellis et al. 1981, Kauffeld approximately one-half is well documented 1981): (1) cranes fly from the roost site hour before sunrise to the morning where they remain until late morning; (2) thereafter they fly to a loafing area where they remain until late afternoon; fly to feeding sites (often the same location where they remain until approximately feeding site, (3) cranes then used in the morning) one-half hour after sunset; (4) after which they return to the roost site where they remain through the night. Inter-site distances which an individual distances site to morning area (AMFSLA), evening averaged Location site (AMRSFS), morning activity periods were measured over individuals. Distance AMRSFS, sexes and between Site fidelity different AMFSLA, site (PMLAFS), Distances consecutive activity was calculated roost on topographic measurements and flown between maps and to and from feeding field quarter section. PMLAFS, and PMFSRS were compared sites used expressed observations periods: feeding site to loafing loafing area to evening feeding site to roost site (PMFSRS). periods by mean activity sites were taken from the center of the recorded Distances diurnal cycles in data were also grouped flown between consecutive feeding feeding consecutive for complete bird was located during all 5 activity , within the same day. individual were measured between periods. for each bird as the number of as a percentage of the total number of for that bird durjng a given activity period (e.g. 22 �477 different feeding observations). fidelity. sites used by bird 76-7 for 55 feeding period Low values for this quotient Comparisons roost complexes, categorized feeding sites, and loafing areas. complex (RCLA), and (2) complex (NRCLA). Wilcoxon a roosting were used on a given day, averaged Because of small sample sizes rank-sum tests were used for all distance comparisons. in this study was set at ~ a roosting PMLAFS, and PMFSRS distances crane and compared. data and site fidelity Loafing areas were loafing areas located outside by loafing area catagories and non-normality, of roost sites, (1) loafing areas located within AMRSFS, AMFSLA, for each whooping high site were made among site fidelities by location: then separated indicated = Level of statistical significance 0.05. RESULTS One juvenile and 21 subadult and adult whooping cranes were observed using 18 roost sites in the SLV during the 1986 fall stopover period. In 1987, one juvenile cranes were observed and 17 subadult and adult whooping using 27 roosting wetlands. located in 1987 had also been used in 1986. Ten of the 27 roosts Of 8 roost sites used in 1986 but not used in 1987, 6 were dry during fall 1987. roost sites were located during the 2 stopover Locations the MVNWR, within of the 35 roosts included 9 wetlands A total of 35 periods. 1 site on the ANWR, 10 sites on located within 1 km of the Rio Grande River, 11 1-2 km of the river, and 1 roost site within river. The 3 remaining MVNWR. All roost sites were located within 1.1), which were defined roosting wetlands 2-3 km of the were within 0.5 km of the 10 roosting as a group of wetlands within complexes (Fig. 2 km of each �478 7S Roost Sites 11S I--------...J:..:.J:,L_- ___; 12S 1------_;___--__;__-;._JO,........_- __ --i ~ --i --Io::..I--l- 0 Feecmq Sites Loafing Sites r--r-; 16 Km Hwy 15 1E 2E 3E 4E 5E 6E 101 102 103 104 105 106 107 Fig 1.1. Distributicns of roost complexes, roost sites, feeding sites. ana loafing areas usea by whooping cranes in the San Luis Valley auring fail 1987. �479 other used as roost sites by whooping Whooping cranes used 1-10 wetlands Whooping in each complex. cranes used wetlands with a broad range of characteristics (Appendix A). ha, SO cranes and sandhill cranes. Size of wetlands ranged from 0.04 to 31.6 ha (~ = 7.2 9.6 ha). ~istance to potential (a roost adjacent to a road) to 1.9 km. = roads, and buildings were 982 m (SO 733 m (SO = 494 m) respectively. Mean distances wetlands Fences frequently (cells). 505 m), and ranged from 460 Two specific conductivities or were adjacent to ~istance from 0 m to 2220 m, with a mean distance first category = SO = 2293 mmhos/cm). dissected (~ = 84.4 m; SO = 123.9 m). Three catagories to houses, Specific conductivity sewage ponds from a starch factory registered >10,000 mmhos/cm. ranged from 0.0 km 535 m), 533 m (SO = (~ = 1162 mmhos/cm, to >10,000 mmhos/cm disturbance to power lines ranged of 471 m (SO = 431 m). of wetlands were used as roosts in the SLV. included 11 artificially created refuge wetlands Cells were flooded with well or ditch water, which collected in shallow uplands and in 2-m wide barrow pits behind dikes. depth at the barrow pit was> flooded area was controlled sites. (Eleocharis Greasewood spp.) were the dominant (Sarcobatus vermiculatus) these wetlands dominant especially generally flooded vegetation, visibility Size of the was only obstructed <0.5 m high (Fig. 1.2). into each sedges (Carex spp.), and the upland areas surrounding Horizontal < 20 cm. by the amount of water directed Baltic rush (Juncus balticus), spikerushes Water 40 cm, whereas water depth in the flooded uplands was < 30 cm with a large percentage cell. The plants on most of these was frequently present on and in some cases was the in recently flooded cells. by surrounding Five of these wetlands vegetation had dense �480 Refuge roosti1g wetlands (cells) 2.5-3.0 2.0-2.5 -E ._ 1.5-2.0 1.0-1.5 W I J i I I l .25-.50 / ""1 .> 0.5-1.0 (f) Z ....• j .i.> / 0-0.25 2 W 0 40 20 100 80 60 120 0: 0 Z Rio Grande River roosting oxbows 2.5-3.0 ..; \ I 0 0: 2.0-2.5 « 1.5-2.0 rn 1:0-1.5 0 \ J I j ./ .1 I / 0.5-1.0 W ...J u::: 0 a: a.. / .25-.50 ._ 0 <{ l- ui CJ W > / / /' 0-0.25 0 20 Flooded pasnre Z I 2.5-3.0 -I i 2.0-2.5 -J 40 80 100 80 120 and stock pond roosting wetlands I I 1.5-2.0 1.0-1.5 0.5-1.0 I J i I , I ,/ .~ / ! .25-.50 0-0.25 / i i / ! 0 / 20 40 60 80 100 120 PERCENT VISIBILITY Fig.l.2. Mean (± SO) horizontal visibility of 3 catagories of roosting wetlands from 90 m distance. �4bl cattail (IYQh£ latifolia) visibility and bulrush (Scirpus spp.) that obstructed up to 3.0 m high. Rio Grande oxbows made up the second category wetlands. Characteristic water depth profiles of roosting for oxbows deeply cut outer edge with a gently sloping inner bank. in the shallow water «30 cm) extending visibility obstructed by trees present along the river bank. vegetation was sparse probably obstructed at the 11 wetlands in this category of roosting wetlands consisted df low-growing stock ponds had no emergent Horizontal only obstructing and only was flooded pastures or submergent Water depths were always < 20 cm and generally types. was highly Understory as a result of grazing, Pasture vegetation grasses whereas Cranes roosted visibility. The third category stock ponds. a from the inner bank. Horizontal partially included visibility and meadow vegetation. < 10 cm in both wetland was high at all 13 sites, with vegetation view in the lowest 0.25 m (higher visual obstruction was a result of topography). Feeding sites were primarily and harvested potato fields. quarter section. geese. (roosting areas were located within roosting feeding intermittently Field size ranged from one-eighth to one- These fields were also used by feeding ducks and Loafing near distant cut, disced, or plowed barley fields sites (NRCLA). flooded greasewood complexes RCLA on the MVNWR were areas adjacent to roosting area A, on the MVNWR was grazed during fall 1987). river bottomland were flooded hay meadows NRCLA were irrigation within a feeding ditches adjacent (RCLA) or and cattle-grazed to a feeding site, and nearby greasewood areas. wetlands RCLA in pastures. site, flooded areas �482 Mean stopover duration for 16 whooping cranes using the SLV during the fall 1987 was 37.3 days (SO = 8.6 days; whooping eliminated from the mean calculation November; Pat 14 was also eliminated crane remained because it was found dead on 3 because I was informed that this in the SLV beyond the time of my last observation). (£ > 0.05) between the mean stopover duration There was no difference n of males (~ = 36.4, However, the variance = 9) and females (~ = 38.4, of stopover duration greater than for females dates for all whooping (s2 = 1.1; E n = 7, Fig. 1.3). for males (s2 = 16.63, £ = 0.003). cranes ranged from 23 September with 50% of the birds arriving October. crane 87-5 was 3.8) was Arrival to 12 October, in the 4-day period 29 September to 2 All females and 56% of the males departed within the 9-aay period 3-11 November. Fifteen individuals 1987 stopover period the 18 whooping observed were observed (Table 1.1). cranes present on roost sites. on roosts 174 times during the Three (Pat 7, Pat 10, and Pat 15) of in the SLV during fall 1987 were not I abandoned efforts to locate whooping cranes. Pat 7 and Pat 15 using areas near Hooper, Colorado because of the distance from the concentrated use area. only 17 days and was never observed Cranes exhibited individual = 0.002, £ Pat 10 remained in the SLV at a roost site. greater site fideli~y to a roost complex then to roost sites, feeding sites, or loafing areas (Table 1.2; 0.002, £ = 0.028 respectively). P Site fidelity did not differ among feeding sites, loafing areas, and roost sites within a roosting complex. hunting Roost complex fidelity was frequently (Table 1.3). hunting closure This effect was observed following between the first and second duck season. disturbed by the 12-day Opening day �483 16 15 14 (f) 0 0: rn LL 0 0: iu §2j FEMALES: 13 12 MALES 11 10 ;j m 6 ~5 =>4 Z 3 2 18 23 28 SEP 3 8 13 18 23 28 OCT 2 7 12 17 22 NOV Hunting Season DATE Fig. 1.3. Number of whooping cranes in the San Luis Valley on given dates during the 1987 stopover period. 27 �484 Table 1.1. Roost site locations of whooping cranes in the SLV during the 1987 fall stopover period. Bird Sex Date Roost Complex (Site) Use 76-7 M 23 SEP C(4) Hunted 25 SEP C(I) Hunted 26 SEP C(1) Hunted 27 SEP B(2) Non-hunted 29 SEP C( 1) Hunted 3 OCT o ( 1) Non-hunted 10 OCT O(l} Non-hunted 16 OCT C(2) Hunted 22 OCT 0(1) Non-hunted 23 OCT 0(2) Non-hunted 26 OCT H(1) Hunted 28 OCT 1(6) Hunted 30 OCT O(l) Non-hunted 1 NOV O(l) Non-hunted 4 NOV 78-10 M Last date observed 8 OCT 1(2) Hunted 14 OCT F(I) Hunted 18 OCT F(3) Hunted 21 OCT F(3) Hunted 27 OCT F(1) Hunted 30 OCT F(1) Hunted 4 NOV H(l) Hunted �485 Table 1.1 (cont.). Roost site locations of whooping cranes in the SLV during the 1987 stopover period. Bird Sex 78-10 M 79-7 83-8 M M Date Roost Complex (Site) Use 5 NOV H (1) Hunted 6 NOV H{l) Hunted 8 NOV H(1) Hunted 9 NOV 0(2) Non-hunted 10 NOV 0(2) Non-hunted 11 NOV 0(2) Non-hunted 12 NOV 0(2) Non-hunted 13 NOV Last date observed 30 SEP 1(2) Hunted 8 OCT I(2) Hunted 29 OCT I{S) Hunted 31 OCT I(5) Hunted 1 NOV I(2) Hunted 2 NOV I(9) Hunted 4 NOV Last date observed 16 OCT C(2) Hunted 22 OCT 0(2) Non-hunted 30 OCT 0(1) Non-hunted 31 OCT 0(1) Non-hunted 1 NOV 0(2) Non-hunted 6 NOV 0(1) Non-hunted 7 NOV 0(1) Non-hunted �486 Table 1.1 (cont.). Roost site locations of whooping cranes in the SLV during the 1987 stopover period. Bird Sex Date 83-8 M 10 NOV Pat 4 M 2 OCT Pat 6 M Roost Complex (Site) Use Last date observed I(unknown) Hunted 11 OCT F(2) Hunted 12 OCT F(I) Hunted 13 OCT F(2) Hunted 17 OCT F(2) Hunted 18 OCT F(3.) Hunted 19 OCT F(l) Hunted 20 OCT F(2) Hunted 27 OCT F(l) Hunted 29 OCT F(I) Hunted 30 OCT F(l) Hunted 1 NOV 0(1) Non-hunted 8 NOV O(l) Non-hunted 8 NOV Last date observed 19 OCT I(l) Hunted 20 OCT 1(1) Hunted 26 OCT I(1) Hunted 27 OCT I(2) Hunted 28 OCT 1(1) Hunted 5 NOV 0(2) Non-hunted 7 NOV 0(1) Non-hunted �487 Table 1.1 (cont.). Roost site locations of whooping cranes in the SLV during the 1987 stopover period. Bird Pat 6 Pat 9 Pat 16 Sex M M M Use Date Roost Complex (Site) 8 NOV 0(2) 8 NOV Last date observed Non-hunted 29 SEP C(1) Hunted 2 OCT A( 1) Non-hunted 3 OCT A( 1) Non-hunted 4 OCT A(2) Non-hunted 5 OCT A(1) Non-hunted 10 OCT . 0(2) Non-hunted 13 OCT 0(2) Non-hunted 18 OCT 8(1) Non-hunted 21 OCT A(1) Non-hunted 22 OCT A( 1) Non-hunted 25 OCT C( 4) Hunted 26 OCT 0(1) Non-hunted 27 OCT A(2) Non-hunted 30 OCT Last date observed 23 SEP C( 1) Hunted 25 SEP C( 1) Hunted 27 SEP 0(3) Non-hunted 5 OCT A(3) Non-hunted 8 OCT A(1) Non-hunted 9 OCT A(3) Non-hunted �488 Table 1.1 (cont.). Roost site locations of whooping cranes in the SLY during the 1987 fall stopover period. Bird Pat 16 Sex Oate Roost Complex (Site) Use M 13 OCT 0(1) Non-hunted 18 OCT B(l) Non-hunted 21 OCT A( 1) Non-hunted 31 OCT A(l) Non-hunted 1 NOV A(3} Non-hunted 3 NOV 0(1) Non-hunted 5 NOV O(l) Non-hunted 6 NOV 0(2} Non-hunted 10 NOV 0(2) Non-hunted 11 NOV 0(2) Non-hunted 12 NOV 0(2) Non-hunted 13 NOV O(2} Non-hunted 14 NOV 8(1) Non-hunted 15 NOV A(2) Non-hunted 16 NOV A(3) Non-hunted 16 NOV A(3) Non-hunted 17 NOV A(3) Non-hunted 18 NOV A(3) Non-hunted 19 NOV A(4) Non-hunted 20 NOV A(4) Non-hunted 21 NOV Last date observed �489 Table 1.1 (cont.). Roost site locations of whooping cranes in the SLV during the 1987 fall stopover period. Bird Sex Date Roost Complex (Site) Use 83-19 F 25 SEP C(2) Hunted 25 SEP C(4) Hunted 27 SEP C(4) Hunted 28 SEP C(2) Hunted 30 SEP C(2) Hunted 1 OCT C(2) Hunted 2 OCT C(1) Hunted 3 OCT C(2) Hunted 3 OCT 0(1) Non-hunted 4 OCT 0(2) Non-hunted 5 OCT C(2) Hunted 7 OCT C(2) Hunted 9 OCT 0(1) Non-hunted 13 OCT 0(2) Non-hunted 26 OCT H(I) Hunted 27 OCT 1(2) Hunted 28 OCT H(l) Hunted 3 NOV 86-15 F Last date observed 11 OCT F(1) Hunted 22 OCT 0(1) Non-hunted 29 OCT 0(2) Non-hunted 30 OCT 0(2) Non-hunted �490 Table 1.1 (cont.). Roost site locations of whooping cranes in the SLV during the 1987 fall stopover period. Bird Sex Date Roost Complex (Site) 86-15 F 5 NOV o (1) Non-hunted 6 NOV .0(2) Non-hunted 7 NOV 0(2) Non-hunted 10 NOV 0(2) Non-hunted 11 NOV Last date observed 87-5 Pat 3 F F Use 6 OCT O(l) Non-hunted 22 OCT O(l) Non-hunted 23 OCT 0(2) Non-hunted 30 OCT C( 4) Hunted 31 OCT 0(1) Non-hunted 1 NOV C( 4) Hunted 3 NOV Found dead by refuge personnel 9 OCT l(unknown) Hunted 10 OCT 0(1/2 ) Non-hunted 19 OCT 1(4) Hunted 20 OCT I(4) Hunted 28 OCT I(4) Hunted 29 OCT 1(9) Hunted 31 OCT 1(9) Hunted 1 NOV 1(6) Hunted 2 NOV 1(2) Hunted 8 NOV 1(6) Hunted �491 Table 1.1 (cont.). Roost site observations of whooping cranes in the SLV during the 1987 fall stopover period. Bird Pat 3 Pat 12 Pat 14 Pat 17 Sex Date Roost Complex (Site) Use F 9 NOV 1(2) Hunted 11 NOV 0.(2) Non-hunted 11 NOV Last date observed F F F 9 OCT 0(2) Non-hunted 21 OCT 0(2) Non-hunted 22 OCT 0(2) Non-hunted 5 NOV 0(1) Non-hunted 6 NOV 0(1) Non-hunted 7 NOV 0(1) Non-hunted 8 NOV Last date observed 11 OCT H(I) 11 OCT Last date observed Hunted 4 OCT 0(2) Non-hunted 6 OCT o ( 1) Non-hunted 9 OCT 0(1) Non-hunted 22 OCT 0(2) Non-hunted 25 OCT 0(2) Non-hunted 28 OCT 0(2) Non-hunted 30 OCT 0(2) Non-hunted 1 NOV 0(2) Non-hunted 4 NOV Last date observed �492 Table 1.2. Site fidelity for roosting complexes, roost sites within a roost complex, feeding sites, and loafing areas. Bird Roost Complexes Roost Sites Loafing Areas Feeding Sites Sites(Obs.) % Sites(Obs.) % Sites(Obs.) % Sites(Obs.) % 76-7 6(13) 46 8(13) 62 2(10) 20 22(55) 40 78-10 4(14) 29 5(14) 36 5(22) 27 22(56) 39 79-7 1(5) 20 3(5) 60 3(8) 38 10(27) 37 83-8 2(7) 29 3(7) 43 4(10) 40 12(29) 41 83-19 4 (15) 27 6(15) 40 9(22) 41 18(61) 30 86-15 2(9) 22 3(9) 33 8(9) 89 19(38) 50 Pat3 2(10) 20 ··5(10) 50 4(11) 36 15(51) 29 Pat4 2(12) 17 4(12) 33 5(13) ·38 14(40) 35 Pat6 3(8) 38 4(8) 50 3(17) 12 11(45) 24 Pat7 Pat9 3(3) 4(13) 31 7(13) 54 PatIO 100 3(15} 20 14(39) 36 2(2) 100 9(12} 75 16(22) 73 Patl2 1(5) 20 2(5} 40 6(9} 67 Pat14 1(1) 100 1(1) 100 2(2) 100 3(3) 100 2(2) 100 4(4) 100 Patl5 Patl6 4(25) 16 9(25) 36 10(25) Patl7 1(7) 14 2(7) 29 2(2) Totals 37(144) Means 62(144) 30 44 30(68) 44 100 18(33) 55 70(179) 48 240(586) 54 53 �493 Table 1.3. Hunted and non-hunted roost complexes used before and after opening day of hunting season. Roost Complexes Used1 Before hunting During hunting Bird Hunted Non-hunted 76-7 C H I B 0 78-10 F H 79-7 I I 83-8 C Pat4 Hunted Non-hunted Last Date Observed 0 4 Nov 0 13 Nov 4 Nov 0 10 Nov F I 0 8 Nov Pat6 I 0 8 Nov Pat9 C A 0 B Pat16 C A B D 83-19 C I 0 86-15 F 0 0 11 Nov Pat3 I D 0 11 Nov 0 D 8 Nov Pat12 Pat14 Patl7 0 30 Oct A B D 21 Nov 3 Nov H 11 Oct D D 4 Nov 1 Letters correspond to lettered roost complexes in figure 1. �494 of goose season and the second duck season occurred cranes using hunted roosting non-hunted roost complexes complexes on 31 October. before opening after opening day. Six day used only The 5 whooping cranes (83-8, Pat 16, 86-15, Pat 12, and Pat 17) using non-hunted roost complexes prior to opening day were unaffected by opening immediately day and remained continued in the non-hunted roosting complexes. to use hunted roost complexes Hunting pressure was not measured have been sufficiently H and I during opening hunting season. at these complexes, however, light that birds roosted undisturbed. 79-7 used roost sites II and 15 respectively, Following Three birds Pat 6 and prior to opening day. day these birds did not change roosting did change roost sites within a complex. and may Roost ~ites complexes but II and 15 were located at the outer edges of the complex and were also under different ownership than all other roost sites within this complex. pressure may have been greater on these 2 sites, thereby abandonment by roosting within 4 days following Sandhill All whooping Five whooping cranes causing left the SLV opening day. and whooping changed roosting 02. cranes. Hunting cranes remaining complexes and congregated cranes observed in the SLV after 9 November primarily on roosts 01 and from 9-13 November (78-10, Pat 3, Pat 6, and Pat 16) used these roosts even if they had not previously been observed crane at these sites. Pat 16 was the only remaining in the SLV along with several hundred sandhill November. bird roosted From 15 November in complex cranes until its departure A, also a non-hunted whooping after 13 on 21 November, area. this Roost sites 02 and A4 were kept open by pumped wells after other wetlands froze. �495 Eight whooping activity cranes periods within (22 observations) a day (Table 1.4). 24-hr cycle was 11,753 m (SE = 1142 m). between females and males same activity periods. and males for all distance feeding and evening distance 1.5). Sixty-five between percent flown between between areas and feeding roost and than morning sites f = were located When compared flown on days of RCLA use, AMRSFS distances NRCLA were used (Table 1.6; (Table All NRCLA were located of the MVNWR. 0.04). the were pooled over females Distances on NRCLA. the Rio Grande River, outside distances distance and evening were greater loafing all 5 flown in a of the loafing area observations on RCLA, with the remainder distances Mean distance distances comparisons. sites in both the morning during There was no difference in the inter-site Therefore, were located AMFSLA, south of with were greater PMLAFS, when and PMFSRS did not differ with NRCLA use. DISCUSSION Roost site selection investigated on 3 levels and biological another. cranes of the Grays Lake flock was (Fig. 1.4). characteristics level was site selection of suitable by whooping Level one examined of the roosting wetland. for a roosting roosting wetlands complex, the physical The second an area where occur within close proximity Level three was a landscape sites, loafing areas, and additional to one approach encompassing and second levels as well as the juxtaposition of suitable a number the first feeding roosting complexes. Roosting wetlands Habitat evaluations within the parameters Temple of the roosting of characteristics wetlands in the SLV fit well summarized by Johnson (1980) for roost sites used by the AWB population and of whooping �496 Table 1.4. Distances flown during complete diurnal cycles. Distance Flown (m) Bird Sex 78-10 M Date AMRSFS AMFSLA PMLAFS PMFSRS TOTAL 6 NOV 11 NOV 12 NOV 1545 7364 7455 1182 7727 636 1091 7515 636 1273 7455 7455 5091 30061 16182 X SE 17111 7223 79-7 M 2 NOV 7424 5030 7515 1394 21363 83-8 M 6 NOV 7333 364 4182 2848 14727 Pat 4 M 12 OCT 20 OCT 2182 2455 2303 2303 2303 2303 2455 2182 9243 9243 9243 0 10030 4981 Pat 6 M Pat 16 M 83-19 Pat 3 F F Combined 20 27 . 28 8 OCT OCT OCT NOV 2182 1333 2515 5697 485 485 485 6424 1303 485 1121 6424 1333 2515 909 6424 5303 4818 5030 24969 12 13 17 19 20 NOV NOV NOV NOV NOV 4667 8788 1697 1697 2030 4242 3606 152 2030 2121 4242 2333 152 2030 2121 4667 5242 1697 2030 2030 17818 19969 3698 7787 8302 11515 3134 3 OCT 4 OCT 28 OCT 3545 5333 2545 1879 2000 2485 1636 697 2515 3848 4424 1273 10908 12454 8818 10727 1054 29 OCT 2 NOV 9 NOV 939 5182 5273 1152 2970 6212 939 1909 2030 1364 4667 3030 4394 14728 16545 11889 3784 11753 1142 �497 Table 1.5. Comparisons of mean distances flown between consecutively used activity sites. Distance Mean Measured n (m) SE Ar~RSFS 13 4843 417 AMFSLA 16 2584 301 PMLAFS 14 2309 224 PMFSRS 15 4065 330 test of 'hypothesis that AMRSFS = AMFSLA 2 test of hypothesis that AMFSLA = PMLAFS 3 test of hypothesis that pr~LAFS= PMFSRS 4 test of hypothesis that PMFSRS = AMRSFS 0.00021 0.27062 0.00043 0.18164 �498 Table 1.6. Comparisons of distances between consecutive activity sites for days in which RCLA were used versus days in which NRCLA were used. Mean distance(SE) Inter-site distance Mean distance(SE) when RCLA were used n (m) when NRCLA were used n (m) ANRSFS 11 4080(538) 9 6255(853) 0.041 Ar~FSLA 15 2928(318) 13 2590(489) 0.252 pr~LAFS 13 2367(219) 11 2439(534) 0.973 PMFSRS 11 3064(458) 10 4745(731) 0.094 1 test of hypothesis that AMRSFS'for RCLA days = AMRSFS for NRCLA days 2 test of hypothesis that AMFSLA for RCLA days = AMFSLA for NRCLA days 3 test of hypothesis that PMLAFS for RCLA days PMLAFS for NRCLA days 4 test of hypothesis that PMFSRS for RCLA days = PMFSRS for NRCLA days �499 _---------aJ __ / ------/a2 LEVEL 3 I' LEVEL 1 , -, ~ Feeding site Roost site RCLA NRCLA Fig. 1.4. Three level approach of roost site selection. Levell is selection for a suitable roost site, level 2 selection for a roost complex, and level 3 is for selection based on juxtaposition of all required habitats. Distances al and a2 are flown by birds using RCLA, distances bl and b2 are flown by birds using NRCLA. �500 cranes. Shallow water and extensive horizontal visibility appear to be the most consistent characteristics among wetlands used by roosting cranes. Two variables measured in the SLV which were not consistent with site evaluations in the Central Flyway included distance from roost site to potential hazards and distance to disturbance . .Power line strikes account for 39% of all known losses of fledged birds in the cross-fostered whooping cranes (unpubl. data, Id. Coop Wildl. Res. Unit). In the SLV, agricultural fields are primarily irrigated by center pivot overhead sprinklers. transmission Recently, large power and distribution lines have been constructed to service the expansion of these systems (Brown et al. 1987). Many power lines cross traditional crane concentration areas. Roost sites are frequently closer to such hazards than sites used by AWB whooping cranes, but cranes in the SLV are not selecting against areas where these hazards exist. This may be a result of unavailability of areas free from these hazards or a lack of recognition of such threats by the birds. Whooping cranes are by nature wary and prefer to avoid human activity (Allen 1952). developments Roost sites are usually away from human such as roads, houses, and buildings. This isolation distance may vary in relation to level of activity associated with the development and visual isolation of the roost from the development. Whooping cranes of the Rocky Mountain population have become more habituated to human disturbance than the AWB population because of greater accessability of breeding, stopover, and wintering grounds (J. C. Lewis, pers. commun.). This is consistent with my observation that distance to potential disturbances were less in the SLV when compared �501 to AWB crane 'roost sites. Three roosting adjacent to roads. adjacent to roads, were visible from roads. explanation for the decreased difference stopover The majority complexes in familiarity of the remaining isolation appear as a potential of SLV roost sites may be the allowing birds to roost at whooping cranes were not from roads or houses and were more consistent described for the AWB flock (see chapter isolation of roost sites might result from additional overriding the necessity Naturally-reared and sometimes Howe 1987). 3). whooping cranes generally behavior additional flock that isolation migrate vigilance from disturbance site was greater than alert behavior and Derrickson whooping 1981, cranes sandhill crane Alert crane at a non-traditional documented flock in the SLV (see chapter for the same 2). Complexes Selection for roost sites in close proximity wetlands for cranes Temple as pairs, families, may not be as critical. stopover roosting disturbance. by a surrounding of a cross-fostered bird among a sandhill whooping decreased flock vigilance of Rocky Mountain behavior Roosting Alternativ~ly, small groups (Allen 1952, Erickson enough with roost sites to avoid areas of potential The gregarious may provide What may Roost sites used at 2 non-traditional sites by solitary cross-fostered visible The SLV is a traditional such as a road or house, may have proven over the years not to be a threat, stopover possible cranes each spring and fall. disturbance, these sites undisturbed. sites, although not Another with roost areas. site, used by migrating in the Valley were was observed were described (1983), in the SLV. to other suitable Similar roosting complexes by Lovvorn and Kirkpatrick and Littlefield (1986). Such complexes (1982), Melvin and potentially �502 provide several advantages over single roost sites. For example, cranes fly to roost approximately one half hour after sunset, allowing limited time to select a roost. If a bird approaches a roost site that appears disturbed, an alternate roost can be quickly selected. On several occasions I observed that the presence of humans at a roosting wetland caused birds to fly over the site and select a nearby undisturbed roost. Roosting complexes also provide an efficient escape if birds are disturbed off a roost after landing. Birds flushed from roosts flew to nearby roost sites within the same roosting complex. Disturbance by hunters affected the distribution of whooping cranes in the SLV and may have accelerated their migration chronology. The effect of hunting was documented by observing changes in roost complex use before and after opening of hunting season. Cranes usually abandoned hunted roost sites, switching to non-hunted roost complexes or low-pressure hunted complexes. Abandonment of hunted roost areas following opening of hunting season has been well documented for sandhill cranes (Lovvorn and Kirkpatrick 1981, Melvin and Temple 1983, Littlefield 1986). During fall 1987, roost area use before hunting consisted of 9 roosting complexes (five hunted complexes, and 4 nonhunted complexes). During hunting season only 2 hunted roost complexes were used by whooping cranes. Hunting pressure was not measured at hunted areas, so it is possible that the 2 hunted roost complexes used following opening day had minimal hunting pressure. A smaller scale response was observed in 2 whooping cranes abandoning roost sites within a hunted complex and switching to other non-hunted roosts within the same complex. The roost sites used prior to hunting were located at the outer edges of the complex and were also under different �503 ownership than all other roost sites within this complex. pressure may have varied sufficiently portion of the area was unsuitable responded to hunting by leaving wintering grounds. migration (Stahlecker Landscape scale Juxtaposition foraging (Caccamise describes Other factors, however, also affect the chronology of required habitats, also influences cranes. center (Weatherhead and Morrison 1986). insight components sites, communal by 3 (2) two- which hypothesis best sites are related for a particular important in roost selection. among roost sites, feeding sites, cranes in the SLY using inter-site Roost sites, and frequently to all daily activities been explained Understanding spatial relationships of the roosting between (Ward and Zahavi 1973), into factors and loafing areas for whooping distances. Relationships feeding 1983), and (3) patch-sitting how roosts and foraging I described of crane of roost sites, have most frequently (1) information species provides such as the third level of selection areas by whooping principal-strategies on opening 1986). and roosting hypotheses: on to the 9 days, and after the 11th day all females and food abundance and loafing Cranes may also have the SLY and continuing and 71% of the males were gone. investigation, for roosting. that only a Nine of 14 (64%) whooping cranes present day left the Valley within weather within the complex Hunting complexes. loafing areas, were considered Roosting complexes when time at roost and loafing were central areas were combined. Caccamise and Morrison (1988) describe these central locations for communal starlings (Sturnus vulaaris) as diurnal centers. A second type of habitat used by starlings activity was described as �504 supplemental feeding areas; patches of readily accessible.high energy food located outside the diurnal activity center. Caccamise and Morrison (1988) hypothesized that roosts form near locally superabundant food patches (patch-sitting). By estimating habitat type fidelity (diurnal activity center-based birds versus roost-based birds) and measuring distances flown between the diurnal activity center, roosts, and supplemental feeding areas, they were able to support this hypothesis. I applied this methodology to whooping cranes in the SLY and found evidence that cranes are roost-based birds. Roost site selection is then influenced by the fact that the roost is the center of operation of all daily activities at the stopover site. Factors such as minimal disturbance and distance to supplemental feeding areas should therefore affect suitability of the roosting wetland. The first criteria for establishing cranes as roost-based birds is that greater site fidelity is exhibited for roosting complexes than any other activity sites. Two factors affected site fidelity: and the numbers of cranes in the Valley. hunting, Cranes remaining in the SLV after 9 November congregated at 2 non-hunted roost complexes. Inter-site distances also supported the roost-based hypothesis. The primary activities of the cranes during roosting and loafing are alert and comfort behaviors (R. C. Drewien, Prog. Rep. Whooping Crane Egg Transplant Experiment 1-20). Although whooping cranes feed on a variety of invertebrates and aquatic plants at wetland sites while roosting and loafing, agricultural lands, primarily fields containing small grains, comprise over 70% of the sites where feeding is the primary activity (Johnson and Temple 1980). These agricultural sites, used during morning and evening feeding periods, are supplemental �505 feeding areas as described generally returned and evening feeding in the inter-site are used for minimizing feeding and Morrison to the same or adjacent for both the morning Patterns by Caccamise distance and Benham 1976, Reinecke data suggest to a nearby supplemental to loaf (Fig. 1.4). to select 1986). for whooping and Kirkpatrick The first strategy of cranes cranes (Lewis 1982, Krapu et is for cranes to fly feeding area then return to the roost complex With this strategy it is advantageous loafing habitat. The second strategy loafing areas by cranes feeding at distant allowing a bird to use a distant feeding used by whooping that 2 strategies The close proximity roost sites within roost complexes in returning feeding area 1981) and for sandhill and Krapu 1979, Lovvorn Cranes periods. daily movements. 1969, Kauffeld al. 1984, Littlefield energy supplemental sites to roost sites has been documented (Shields (1988). for cranes that also provide suitable involves use of non-roost feeding complex sites, thereby site but not expend extra / to the roost complex to loaf. Both strategies cranes using the same roost complex. apparently do not select a roost site in preference but rather select a site that affords the opportunity were Thus, cranes of one strategy, to employ either strategy. The use of NRCLA was only observed may have resulted areas. converted south of the river area. from the effects of different In the vicinity of the Rio Grande agricultural nearby feeding areas. land use in the 2 River more land has been to grain fields, thereby providing fewer loafing This more feeding sites but The area south of the river is also primarily but is interspersed with non-agriculturai land. sites are not limited south of the river, the Although �506 availability of suitable loafing habitat outside the roost complexes provides cranes with the opportunity to exploit distant feeding sites at no increased cost. Roost site selection by whooping cranes of the cross-fostered appears dependent characteristics flock not only on the physical and biological of a roosting wetland but also on the juxtaposition of other suitable roost sites, feeding sites, and loafing areas. Further studies should be conducted which relate the role of suitable loafing habitat aut-s-ideof roost complexes to the use of feeding sites. The influence of suitable loafing habitat adjacent to roost sites should also be investigated as a factor in roost site selection. L-ITERATURE CITED Allen, R. P. 1952. 3. 246pp. The whooping crane. Natl. Audubon Soc., Res. Rep. Anderson, D. R. 1965. Effects of water manipulation on waterfowl production and habitat. M. S. Thesis, Colo. State Univ., Fort Collins. 60 pp. Archibald, G. W., J. Baldwin, and P. Konrad. 1976. Is sandhill hunting a threat to whooping cranes? Pages 207-222 in J. C. Lewis, ed. Proc. Int. Crane Workshop. Okla. State Univ. Publ. and Print., Stillwater. Brown, W. M., R. C. Drewien, and E. G. Bizeau. 1987. Mortality of cranes and waterfowl from power line collisions in the San Luis Valley, Colorado. Pages 128-136 in J. C. Lewis, ed. Proc. 1985 Crane Workshop. Natl. Audubon Soc. Tavernier, Fl. Caccamise, D. F. and D. W. Morrison. 1986. Avian communal roosting: implications of diurnal activity centers. Am. Nat. 128:191-198. __ , and _ . 1988. Avi an communal roost ing: a test of the "patch-sitting" hypothesis. Condor 90:453-458. Drewien, R. C. 1973. Ecology of Rocky Mountain greater sandhill cranes. PhD. dissertation, Univ. Idaho, Moscow. 152 pp. , and E. G. Bizeau. 1974. Status and distribution of greater --sandhi 11 cranes in the Rocky Mountains. J. Wi 1dl. 11anage. 38: 720742. �507 , and 1978. Cross-fostering whooping cranes to sandhill crane foster parents. Pages 201-222 in S. A. Temple, ed. Endangered birds: management techniques for preserving threatened species. Univ. Wisconsin Press, Madison. Ellis, S. L., T. G. Shoemaker, H. Wen Shen, and W. Wang. Niobrara River whooping crane habitat study. Prepared Environmental Research and Technology, Inc. for Water Resour. Serv., U. S. Dep. Interior, Denver, Colorado. 1981. by and Power 356pp. Erickson, R. C. and S. R. Derrickson. 1981. The whooping crane. Pages 104-118 in J. C. Lewis, ed., Cr-ane research around the worl d, Int. Crane Foundation. Int. Counc. for bird preservation. Baraboo, WI. Hopper, R. M. 1968. Publ. 22. 88pp. Wetlands of Colorado. Colo. Div. Wildl. Tech. Howe, M. 1987. Habitat use by migrating whooping cranes in the Aransas -Wood Buffalo corri dor. Pages 303 -311 in J. C. l.ewis , ed., Proc. 1985 Crane WorkshQP. Natl. Audubon Soc., Tavernier, Fl. Johnson, K. A. and S. A. Temple. 1980. The migratory ecology of the whooping crane. Unpublished report, contract 14-16~009-78-034, U. S. Fish and Wildl. Serv., Washington, D. C. 120pp. Kauffeld, J. D. 1981. Management of migratory crane habitat on Alamosa and Monte Vista National Wildlife Refuges. Pages 117-121 in J. C. Lewis, ed. Proc. 1981 Crane Workshop. Natl. Audubon Soc., Tavernier, Fl. Krapu, G. L., D. E. Facey, E. K. Fritzell, and D. H. Johnson. 1984. Habitat use by migrant sandhill cranes in Nebraska. J. Wildl. Manage. 48:407-417. Lewis, J. C. 1976. Roost habitat and roosting behavior of sandhill cranes. Pages 93-104 in J. C. Lewis, ed. Proc. Int. Crane Workshop. Okla. State Univ. Publ. and Print, Stillwater. Lingle, G. R., P. J. Currier, and K. L. Lingle. 1984. Physical characteristics of a whooping crane roost site on the Platte River, Hall County, Nebraska. Prairie Nat. 16:39-44. Lingle, G. R. 1987. Status of whooping crane migration habitat within the Great Plains of North America. Pages 331-340 in J. C. Lewis, ed. Proc. 1985 Crane Workshop. Natl. Audubon Soc., Tavernier, Fl. Lingle, G. R., G. A. Wingfield, and J. W. Ziewitz. 1988. The migration ecology of whooping cranes in Nebraska, U. S. A .. Proc. Int. Crane Workshop (in press). Littlefield, C. D. 1986. Autumn sandhill crane habitat use in southeast Oregon. Wilson Bull. 98:131-137. �508 Lovvorn, J. R., and C. M. Kirkpatrick. 1981. Roosting behavior and habitat of migrant greater sandhill cranes. J. Wildl. Manage. 45:842-857. , and 1982. Field use by staging eastern greater sandhill --cranes.-J-.-Wildl. 11anage. 46:99-108. Melvin, S. M., and S. A. Temple. 1982. Migration ecology of sandhill cranes: a review. Pages 73-87 in J. C. Lewis, ed., Proc. 1981 Crane Workshop. , and 1983. Fall migration and mortality of Interlake, ---~Manitoba sandhill cranes in North Dakota. J. Wildl. Manage. 47:805-817. Melvin, S. M., R. C. Drewien, S. A. Temple, and E. G. Bizeau. 1983. Leg-band attachment of radio transmitters for large birds. Wildl. Soc. Bull. 11:282-285. Nudds, T. D. 1977. Quantifying the vegetative structure of wildlife cover. Wildl. Soc. Bull. 5:113-117 Reinecke, K. J., and G. L. Krapu. 1979. Spring food habits of sandhill cranes in Nebraska. Pages 13-19 in J. C. Lewis, ed. Proc. 1978 Crane Workshop. Colo. State Univ., Fort Collins. Ryder, R. A. 1951. Waterfowl production in the San Luis Valley, Colorado. 11. S. Thesis, Colo. A & M Col1., Fort Collins, 166pp. Shields, R. H., and E. L. Benham. 1969. Farm crops as food supplements for whooping cranes. J. Wild1. Manage. 33:811-817. Stahlecker, D. W. 1986. The 1985 fall crane migration in the Rio Grande Valley, New Mexico. Unpub1. Rep., New Mexico Audubon Soc., Albuquerque. 22pp. Ward, J. P. and S. H. Anderson. 1987. Roost site use versus preference by two migrating whooping cranes. Pages 283-288 in J. C. Lewis, ed., Proc. 1985 Crane Workshop. Ward, P. and A. Zahavi. 1973. The importance of certain assemblages of birds as "information centers" for food finding. Ibis 115:517534. Weatherhead, P. J. 1983. Two principal strategies in avian communal roosts. Am. Nat. 121:237-243. Whooping Crane Recovery Team. 1980. Whooping crane recovery plan. U. S. Dep. Inter., Fish and Wi1d1. Serv., Nebr. Game and Parks Comm., Texas Parks and Wildl. Dep.; Natl. Audubon Soc. 206pp. �Table AI. Habitat evaluations Luis Valley, Colorado. Hoosl [liH!!!!re Complex Size (S ite) (ha) --.. _._- ._------_ .... !!Q!!~g for 35 roosting wetlands located during fall 1986 and 1987 in the San tQ_IJQtlli.!U!!LiU~lllrbance llQM! fillllsLim (III) . _-_._-_. Specific Couduc t iv i t.y of Water Oist~nce to HQlenlliLh!!I!!rd (mmhos/cm) ._-_ .. __ . _ .._. _ .. __ - ._---_ POI'JerLine Fence (m) .. _.-- --------- A (1) 1.5 1470 72 092 1300 50 50 A(2) 1.5 944 208 922 600 192 192 -u -u A(3) 5.3 862 244 862 750 236 236 I-< ::t:> rn :z: 0 >< A(4) 8.4 504 226 468 600 222 222 13 ( I ) ]2.3 1424 15 1394 1600 485 485 [3(2) 11.0 909 970 423 975 0 970 C (1) 0.4 1273 394 1273 500 394 394 C(2) 0.2 136,1 333 1333 700 242 333 C(3) 0.6 1515 606 1405 600 0 606 C(4) 78. ] 1152 000 ))52 950 194 800 o(]) 58.2 1181 1)21 ]10) 950 0 1121 0(2) 73.4 939 800 939 460 34 800 0(3) 60.9 1333 24 1303 950 0 24 ::t:> I.Jl 0 1.0 �Table Al (cont.). Habitat evaluations San Luis Valley, Colorado. for 35 roosting wetlands located during fall 1986 and 1987 in the ---------- Roost Qistilnce tQ_Qote!lJial disturbance Complex Size (Site) (ha) l!Q!!~~ BQ_Q1 ~!!U1Ins of Hater Distance to potential hazard EOI'/erLi ne Fence (mmhos/crn) (III ) -_. Specific Conductivity (m) __ ------ E (1) 19.5 970 1121 606 > 10000 0 56 E(2) 27.4 940 112 j SIS > 10000 24 303 .[(3) - 1879 175n 1303 750 152 1091 r( ] } 2.7 121 174 56 950 44 174 F(2) 0.9 78 0 64 580 3 9 F(3} - 264 9 254 600 0 0 ]1.2 592 624 532 2600 2 624 G(2} 1.4 1126 2613 242 2000 90 268 G(3) 3.3 1350 42 188 II( 1) 9.6 456 574 420 560 3 828 11(2) 0.6 610 588 578 800 152 588 I(1) 1.6 1148 1000 848 680 ·0 658 1(2 ) 3.5 1242 455 727 500 273 697 G(l) W81 284 Vl I-' 0 �Table Al (cont.). Habitat evaluations San Luis Valley, Colorado. for 35 roosting wetlands located during fall 1986 and 1987 in the --------- Roost --- ni~ti)!1~g_tQ.__nQt~nttel ~U~1.!1rhanc~ Specific Conductivity Complex Size (Site) (ha) !hl!J~g fll2.i!.d fluiJging ([II) of I-I a tel' Oii1~nce tQ_p.ote!1tial_b~~Ar~ Fence Power Line (01) (1IIIIII1os/elll) --1(3) 0.2 758 788 697 3500 4 333 1(4 ) 9.3 424 455 424 . 650 71 424 1(5) 1.2 424 242 242 2200 0 242 1(6 ) 1.2 1485 152 333 600 0 394 1(7) 2.5 1758 182 576 480 0 576 I(8) 0.1 814 8GO 830 600 0 504 I(9) 5.3 333 2 303 5000 0 0 1(10) 0.04 212 182 70 2400 27 68 8.6 2515 2212 2303 800 16 2212 Alamosa _._------_. l.r1 I-' I-' ��513 CHAPTER 2 WHOOPING CRANE BEHAVIOR AT TRADITIONAL AND NON-TRADITIONAL FALL MIGRATION STOPOVER SITES SUMMARY Behavior described at a non-traditional stopover -located of a female subadult whooping stopover crane (Grus americana) site (NTSS) and a traditional site (TSS) during the 1986 fall migration. in eastern Colorado using a TSS on the Alamosa Luis Valley alert behavior Wildl~fe Daily activity Refuge and foraging to behavior patterns (NWR) in the San eastern Colorado. feeding periods increased at the NTSS compared at the TSS. time decreased Time budget activity periods The frequency of when feeding alone when feeding with a flock at the TSS. at the NTSS were consistent with those of the cranes of the Grays Lake flock using TSS. One result experiment National path of The same bird was observed on this crane during all diurnal at the NTSS, and during whooping crane flock. 6 days after it departured data were collected The NTSS was 170 km east of the u$ual migration the Grays Lake heterospecific was of the whooping is the communal during migration. Sandhill their large migration crane transplant behavior cranes and foster parent of Grays Lake whooping cranes (Grus canadensis) flocks and communal along the Platte and North Platte Rivers behavior are well known for at staging (Krapu 1979, Johnson areas and �514 Temple 1980, U.S. Fish and Wildlife Service 1981, Melvin and Temple 1982, Krapu et al. 1984, Currier and Ziewitz 1987, Lingle et al. 1988). Whooping cranes of the natural Aransas-Wood Buffalo population, however, migrate as individuals or in family units (Allen 1952, Erickson and Derrickson 1981, Howe 1987). The use of 2 non-traditional migration stopover sites (NTSS) during fall 1985 by solitary subadult whooping cranes of the Grays Lake flock, and the repeated use of a site near Hudson, Colorado (spring 1986, fall 1986), shows that fosterreared whooping cranes have the ability to survive as solitary birds despite rearing by gregarious foster parents. The opportunity to observe the behavior of the same bird while alone at the NTSS and as a member of a sandhill flock at the traditional stopover site (TSS) in the San Luis Valley (SLV) could provide insight into any behavioral changes accompanying the presence or absence of a surrounding ftock. The effect of flock size on vigilant behavior of individual birds within the flock has been investigated in numerous studies (Powell 1974, Abramson 1979, Barnard 1980, Bertram 1980, Weatherhead primary assumption 1983). A of these studies is that vigilance while feeding enhances the probability of detecting predators. Waite (1987), however, found evidence supporting the hypotheses that vigilance is predator- and conspecific-directed and is influenced both by the tendency to be social and by an individual's dominance status within the social group. Subordinate social foragers have the additional constraint on their foraging time of keeping higher-ranking under surveillance. flock mates Knight and Knight (1986) also provided evidence that vigilance behavior of birds that feed in social groups has the additional function of conspecific surveillance. �515 Biologists studying whooping cranes of the Grays Lake flock have noted a low level of alert behavior during daily activities compared to sandhill cranes of the same flock (R. C. Drewien, pers. commun.). Drewien hypothesized that this may result from a possible learning deficiency as a consequence of rearing by the foster greater sandhill parents. He felt that the whooping cranes were displaying behavior. chick-like This behavior could increase the possibility of predation because of decreased vigilance for predator protection (Tacha 1981). An alt~rrative hypothesis is that the whooping cranes are displaying dominant behavior within the flock and, as suggested by Waite (1987), exhibit decreased vigilance because of social status. Data were collected on an individual whooping crane with and without a surrounding sandhill flock to address these hypotheses.' STUDY AREA The NTSS was located 2 km south of Hudson, Colorado in southern Weld County. m. Topography is broadly rolling with an elevation of 1512 Irrigated farmland in the area supports the main crops of corn, alfalfa, wheat, and sugar beets. grown in lesser amounts. Malting barley and vegetables are Feeding sites used by subadult, female whooping crane #8422 were nearby fields of volunteer barley, corn, wheat, hay, and grazed pasture. as (system) palustrine, (class) emergent wetland, and (modifier) saturated/semi-permanent/seasonal 1979). The roosting wetlands were classified wetlands (PEMY; Cowardin et al. Both roost sites were located in grazed cattle pastures. Loafing areas were adjacent to the roosting wetlands. The TSS was located on the Alamosa NWR 1.9 km southeast of Alamosa, Colorado. The 4520-ha refuge is composed of natural riverbottom �516 wetland and ;s bordered on the west by the Rio Grande River. The feeding site used by #8422 was a cut barley field on the northwest corner of the refuge. greasewood Loafing areas were river bottomland and a (Sarcobatus vermiculatis) area of the refuge. The roosting wetland was a flooded slough maintained by an artesian well. METHODS Activities of bird #8422 were recorded at IS-second intervals following the protocol of focal-animal sampling (Altmann 1974). Bird #8422 was identified by a combination of colored leg bands (red band left leg, white band right leg) and a radio attached to the left leg band (R. C. Drewien, pers. commun.). Behaviors were classified 'according to the codes developed for sandhill and whooping cranes of the Grays Lake flock (R. C. Drewien, pers. commun.), coded numerically, (see Table 2.I) then recorded on data sheets. Observations were performed with a 30x spotting scope between dawn and dusk. observations were not made because the primary nocturnal activity of cranes is sleeping (R. C. Drewien, pers. commun.). corresponds Night-time This methodology to that used by W. M. Brown and R. C. Drewien for cranes of the Grays Lake flock. Time budget data were summarized by days and grouped into 9 general catagories of behavior: (1) feeding, (2) drinking, (3) alert, (4) resting, (S) comfort, (6) locomotion, and (9) out of sight. (7) agonistic, (8) vocalization, Each day of observation was divided into 5 activity periods: (1) a.m. roost, (2) a.m. feeding, (3) loafing, (4) p.m. feeding, and (5) p.m. roost. The number of behavioral codes recorded for each activity was divided by the total number of recordings during that activity period �517 Table 2.1. Code 01-09 01 02 03 lQ 11 12 13 14 15 16 17 18 19 Coded behaviors used in time budget data collection. Activity stop observation out of sight out of sight in standing Foraging corn alfalfa milo barley wetland natural upland plowed/disced field picking up gravel other Code 60 66 67 66 68 69 69 70 70 71 72 72 73 74 20 Drinkina 75 30 Defecation 76 77 40 'Standing (Alert) standing alert (tall alert) standing 1 leg (head horizontal) head down-looking/hunting standing upright, wings outstretched 41 42 43 44 50 51 52 53 54 55 60 61 62 62 62 63 64 65 ("'- 00 78 80 81 82 83 84 85 Restina/Sleepina/Loafina head tuck position standing on 1 leg (head down) bill tucked under wing laying/sitting down other 86 Comfort Movements oil i ng head rubbing head shake (neck/head fl i ck) head stretch (neck and head) scratching bathing wing flapping stretching-leg extended 92 87 88 90 91 93 94 95 96 97 Activity Comfort Movements stretching-wings over back, head and neck extended shaking-body preening preening radio thermoregulation (bill gape) yawning Locomotion hunting (head down) walking running running with wings outspread flight intention flyi ng 1 andi ng jumping/dancing aerial pursuit (chasing sandhills) butterfly Agonistic Behavior (Initiates, recipient=circle #) threat posture aggressive strut (stick walk) attacking (charging) pecks other bird is being pecked submissive behavior (avoidance=retreat, runs or backs away, head tuck) fighting - jumps in air other Call s nervous call (Kepler) guard/ alarm call unison call whistle-click (contact, distress, contentment) threat call precopulatory other �518 to obtain percentage of time spent per activity per period. small sample sizes (non-traditional site N=9, traditional Because of site N=4), Wilcoxon tests were used to detect differences in behavior between TSS and NTSS. RESULTS The daily activity pattern of #8422 at both the NTSS and TSS (Table 2.2) was similar to that observed in other whooping cranes of the Grays Lake flock during fall stopover periods at TSS in the SLV (Drewien 1982, Drewien and Brown 1983, 1984). This general daily activity pattern is to fly from the roost just prior to sunrise to the morning feeding site, where they remain until late morning or early afternoon, after which they fly to a loafing area. Cranes remain in the loafing area until late afternoon then fly to a feeding site (usually the same location used in the morning)· until just prior to sunset when they return to the roost where they remain through the night. The location category "other" includes observations of flight and at sites visited for a brief time between activity periods or following a flush. Whooping crane #8422 was first observed near Hudson, Colorado on 17 October. Daily behavioral observations began on 18 October and continued through 26 October, when I observed the bird spiral and leave the area. During the 9-day stopover period near Hudson, 45.7 hr of time budget data were collected: 1.5 hr at the a.m. roost site, 22.6 hr at the a.m. feeding site, 9.5 hr at the loafing site, 6.3 hr at the p.m. feeding site, 5.2 hr at the p.m. roost site and, 0.6 hr at other sites (Table 2.3). Daily summaries of time spent in diurnal activities are presented in Appendix B. �519 Table 2.2. Locations and activities of whooping crane #8422 during the 1986 fall stopover period in Colorado. Date Time Location1 17 Oct 1200 RS1, FS1 18 Oct 1205 1205-1841 FS1 RSl, FSI 19 Oct 20 Oct 21 Oct 22 Oct 23 Oct 24 Oct 700 715 RS1 RS1 1100 1309 1300 FS2 FS2 FSI 1552 1751 FS2 RSI 700 730 1049 RS1 FS2 FSI 1448 1539 1138 708 735 1110 1200 1526 1620 1812 1840 RS1 FS1 FS2 RS1 FS2 FS1 RS1 RSI FS2 FS2 RS1 1923 1943 1603 1617 1656 1703 RS1 FS2 FS2 WI WI W2 801 815 1225 1323 1425 FS7 FS7 FS1 FSI Activity #8422 first observed by landowner L. Long, reported to CDOW begin observation loafed and foraged in the area surrounding the roost site, fed in an adjacent hay field and volunteer wheat field begin observation flew from the roost in heavy fog, I was unable to locate for 4 hrs foraged in volunteer barley flew from the area foraged in volunteer wheat near wetland edge, loafed flew to feeding site flew to roost site, stop observation begin observation flew to feeding site flew to volunteer wheat, loafed and foraged in water stop observation foraging begin observation flew to feeding site flew to volunteer wheat field in water and surrounding area flew, not located for 45 min. located at feeding site flew to RSl, stop observation begin observation, too foggy to collect time budget data flew to feeding site located at feeding site hazed from the area by a dog located at Spayds' pond flew to SW end of Longs' pond 1 located on N side of Longs' pond 1, stop observation begin observation, seen flying from direction of Longs' pond 2 located in grazed pasture flew to RS1 located in volunteer wheat hazed by 2 dogs, flew to Longs' pond 2 �520 Table 2.2 (cont.). Locations and activities of whooping crane #8422 during the 1986 fall stopover period in Colorado. Date Time 24 Oct 25 Oct 26 Oct 1521 RS2 1842 1843 721 740 1038 1406 1619 RS2 RS2 RS2 FS2 FS2 RS1, FS1 RS2, FS7 1843 919 1140 RS2 FS2 FS2 2 Nov 3 Nov 4 Nov 5 Nov 6 Nov 1 Location FS 753 1306 1622 1650 1857 842 925 1040 1501 1713 1930 1046 1519 1716 837 916 1508 1716 FS FS FS FS FS LS LS FS FS FS FS FS FS FS FS FS FS FS Activity in and near edge of the water, foraged in grazed pasture stop observation begin observation flew towards feeding site located at feeding site f1ew towards RS1 located bird foragins in water and adjacent pasture stop observation begin observation observed spiralling and leaving Hudson area, stop observation first observed by Alamosa NWR personnel foraging in a barley field with a group of sandhill cranes and 1 adult whooping crane begin observation flushed from FS begin observation Flew to roost begin observation flew to adjacent greasewood area flew back to FS flew from feeding area begin observation flew to roost bird arrived at FS flew in SW direction begin observation flew to roost begin observation all birds flushed from area begin observation flew to roost RS1-roost site on Wilhelm property, used in previous stopovers FSl-volunteer wheat field adjacent to RSI FS2-volunteer barley field NE of RS1 WI-pond located SE of RSI on Spayd property W2-pond located SE of RS1 on Long property RS2-roost site located on Long property SE of RS1 FS7-grazed pasture adjacent to RS2 FS-barley field on Alamosa NWR LS-greasewood field adjacent to feeding site �Table 2.3. Average percent time spent in various diurnal activities by whooping crane #8422 from 18-26 October 1986 near Hudson, Colorado. Activities were recorded at IS-second intervals. Locatioll Ac tivity A.M~QQ~t ~ (SD) _._------_---_-------- --- Feeding 4.0 (7.4) Drillking 0.0 (-) Alert Comfort I OCOIIIO 0.0 10.7 t ion - x (SO) 8.3 (8.5) 0.0 (-) 0.) (0.1) 47.6 (7.3) (-) 0.0 (-) (13.3) 1.0 (1.2) - ) 62.0 (18.0) 0.1 IS.7 (5.3) 0.0 Vocal i z at ion 0.0 (- 0.0 1.1 (1.2) 0.7 of Sight ~ (9.1) 26.7 (12.2) 32.0 Agonistic Out (SO) £_J'L __f g e <1iu s Loafing - x (SO) --------- 54.6 (IS.8) Re s t i119 8_,_!L_I eed j us 0.0 (0.3) 16.1 (12.6) 12.8 (9.4) 0.0 (- (0.6) ) - ) tJ_L_BQQ~t - (SO) x QttJer ~ (SO) TolQl - -- ---------- 33.2 (21.8) 0.0 ( - ) 39.9 (2. I) ------ 1.7 (1.6) 1.4 0.5 (1.0) 0.7 (24.9) 40.4 (1.7) 0.0 63.8 1.3 (SO) X 25.7 (10.9) 0.0 (0.1) 49.9 (7.0) - 0.2 (0.4) 11.6 - 6.1 (4.8) 45.9 - 17.0 - 0.0 (-) 5.2 (9.2) 20.0 (9.5) 9.8 (9.2) 0.0 {- 0.0 (- 0.0 - 0.0 (- 0.0 - 0.0 0.0 - 1.0 21.5 (16.1) 0.0 (- 0.0 (- 0.0 0.6 (0.8) 1.8 (2_9) 1.4 (2.5) (5.8) (- ) (0.9) Vl N f-' �522 Alert behavior accounted for the majority of the bird's activity during 8 of the 9 days of observation. On the day that did not follow this pattern, the percent time spent in alert behavior tied with feeding as the primary activity. When days were divided into the 5 100% of the loafing observations, 50% of the p.m. feeding site observations and, 100% of the p.m. roost site observations. Feeding accounted for the second highest percent of diurnal activity, locomotion followed feeding, and comfort behaviors ranked fourth for total daily activities. The majority of comfort behaviors occurred at roost sites and loafing areas. Whooping crane #8422 was observed in the SLY from 2-8 November, and daily behavioral observations November. on this crane were conducted from 3-6 During the SLY stopover period, I observed this bird a total of 16.8 hr: feeding site. 10.4 hr at the a.m. feeding site and 6.4 hr at the p.m. Daily summaries of time spent in diurnal activities during the San Luis Valley stopover period are presented in Appendix C; average percent time spent in these activities are presented in Table 2.4. Time budget data were collected only during the a.m. and p.m. feeding periods, since approaching the birds to collect behavioral data at the roost site and loafing area might have caused the birds to flush. For 87% of all observations, the primary activity during the a.m. and p.m. foraging periods was feeding, followed by alert behavior, then locomotion. This time allocation is consistent with that observed in other whooping cranes of the Grays Lake flock during a traditional stopover period in the SLY (Drewien 1982, Drewien and Brown 1983, 1984). Behavioral data recorded by W. Brown and R. Drewien on other �523 Table 2.4. Average percent time spent in various diurnal activities by whooping crane #8422 from 3-6 November 1986 in the San Luis Valley, Colorado. Activities were recorded at IS-second intervals. Activity A. M. Feeding P. M. Feeding SD SD SD (%) (%) (%) Total Feeding 74.6 9.5 78.7 1.0 72.8 3.7 Drinking 0.0 0.0 0.0 0.0 0.0 0.0 17.0 2.5 11.1 1.4 15.2 3.2 Resting 0.0 0.0 0.0 0.0 0.0 0.0 Comfort 2.5 2.8 0.2 0.5 1.6 1.7 Locomotion 9.9 1.6 10.0 2.4 10.3 0 ..7 Agonistic 0.0 0.0 0.0 0.0 0.0 0.0 Alert , Vocalization 0.0 0.0 0.0 0.0 0.0 0.0 Out of Sight 0.3 0.2 0.0 0.0 0.2 0.1 �524 whooping cranes in the San Luis Valley were not available for statistical comparison. The behavior of whooping crane #8422 during both feeding activity periods at the Hudson stopover site differed (E < 0.05) from the behavior exhibited by the same bird in the corresponding activity periods at the TSS in the SLY (Table 2.5). During the a.m. feeding activity period, time spent feeding (~ = 32%, SO = 9.1) at the Hudson site was less than time spent feeding (~ (E 0.01). 74.6%, SO = 9.5) in the SLY = For the same activity period, time spent in alert behavior (~ = 47.6%, SO = 7.3) at Hudson differed (E = 0.01) from the time spent in alert behavior (~ = 17.0%, SO = 2.5) at the SLY site. Locomotor activity at the Hudson site (~ = 18.7%, SO = 5.3) was more frequent (E = 0.03, Wilcoxon test) than at the SLY (~ = 9.9%, SO = 1.6). Compafisons made among the 3 primary behaviors (feeding, alert, and locomotion) during the p.m. feeding period yielded no significant differences. than (E = Feeding time at Hudson (~ = 33.2%, SO = 21.8) was less 0.03) time spent feeding in the SLY (~ = 78.7%, SO = 1.0). At Hudson, time spent in alert behavior (~ = 39.9%, SO = 2.1) was greater than (E = 0.03) time spent alert in the SLY (~ 1.4). Locomotion activity averaged 20.0% (SO and 10.0~~ (SO significant = = 11.1%, SO 9.5) at the Hudson site 2.4) at the SLY site, however the difference was not (E = 0.06). DISCUSSION Communal avian behavior has most frequently been explained by 4 hypotheses: (1) the predation avoidance hypothesis, which proposes that the risk of predation for an individual is reduced by increased vigilance and dilution of prey (Lack 1968), (2) the information center �525 Table 2.5. Comparisons of 3 primary behaviors during the a.m. feeding and p.m. feeding activity periods at the traditional and non-traditional stopover sites. Activity Period Activity Non-Traditional X (%) SO Traditional X (%) SO A.M. Feeding Alert 47.6 7.3 17.0 2.5 0.0069 Feeding 32.0 9.1 74.6 9.5 0.0069 Locomotion 18.7 5.3 9.9 1.6 0.0253 Alert 39.9 2.1 11.1 1.4 0.0304 Feeding 33.2 21.8 78.7 . 1.0 0.0304 Locomotion 20.0 9.5 10.0 2.4 0.0606 P.M. Feeding pl denotes ~robability level for Wilcoxon test of no difference in percent of time spent in each behavior between stopover sites. �526 hypothesis, in which individuals find out about the location of good feeding sites by following others (Ward and Zahavi 1973), (3) the patch-sitting hypothesis, super-abundant which suggests that roosts form near locally food patches (Caccamise and Morrison 1986), and (4) the two-principal-strategies hypothesis (Weatherhead 1983), in which superior foragers roost communally because their status allows them access to central roosting positions that are buffered from predation by the surrounding subordinate individuals. Subordinate individuals pay the price of higher predation risk because of the foraging advantage they realize by following dominant individuals feeding locations (Weatherhead to superior 1983). Considering the "abundant forage available to sandhill and whooping cranes in the SLY, it seems unlikely that the information hypothesis or the patch-sitting communal behavior exhibited hypothesis would be the basis for in this location. Rather, time budget data on whooping crane #8422 support the predation avoidance principal strategies hypotheses. bird spent significantly center and two- During the period at the NTSS, the more time in alert behavior then when foraging at the TSS while surrounded by the sandhill flock. As a result of decreased alert behavior at the TSS the percent time spent in foraging significantly increased, increased vigilance as predicted by Krebs and Davies (1987). in the absence of a flock indicates the bird compensated for the loss of the flock vigilance by increasing alert behavior. Conversely, its own the decrease in alert behavior when it returned to the SLY indicates a recognition of decreased behavior. This need for alert A number of species allocate time to these behaviors in a manner dependent on flock size, as predicted by the hypothesis (Lazarus �527 1979, Caraco 1979, Barnard 1980, Bertram 1980). assumed that the percent time spent in alert behavior at the Hudson site was due to the absence of a surrounding flock and not influenced by unfamiliarity the area. of felt this assumption was justified because although it was a non-traditional site it was the third migration stopover period that #8422 had used the area. Decreased alert behavior in foster-reared whooping cranes compared to sandhill cranes of the same flock (R. C. Drewien, pers. commun.) may be the result of social status within the flock rather than a learnir.g deficiency. If whooping cranes were truly expressing chick-like behavior, assuming the necessity of alert behavior had not been learned, the percent time spent in vigilance would not have significantly increased when feeding alone. Thus, time budget data collected on #8422 as a solitary bird supports the concept of dominance position as a more probable explanation of decreased alert behavior by whooping cranes when foraging with a sandhill crane flock. Social dominance within a flock is known to influence the amount of alert behavior exhibited by individual birds within a flock. Waite (1987) found that subordinate social foragers exhibit increased vigilance when compared to dominant foragers within the same flock. He suggested that this increased vigilance is a result of subordinates keeping higherranking flock mates under surveillance. Caraco (1979) found subordinate yellow-eyed juncos (Junco phaeonotus) spent an increased amount of time watching over other flock members after being displaced by dominants. Incidents of dominance of whooping cranes over greater sandhill cranes has been observed within the Grays Lake flock (R. C. Drewien, pers. commun.). �528 Social interactions other than dominance vigilance also influence the unequal observed in individuals within a flock. Higher levels of alert behavior have been reported in males compared with females of several bird species. (Anas penelope) are more alert than females. found in ostriches guarding, Mayhew (1987) found that male European wigeons A similar result was (Struthis camelus) by Bertram (1980). Mate- territorial defense, increased predator risk for conspicuous males, and potential mate searching are hypotheses that have been tested to explain different vigilance patterns between sexes. All whooping cranes in the Grays Lake flock are currently non-breeders and could thus be classified as subadults, eliminating the need for increased vigilance by any individuals Bishop and Blankenship based on these social dynamics. (1982) investigated the dynamics of subadult flocks of whooping cranes at Aransas NWR, Texas. Subadults are the only age class that exhibit gregarious Chicks remain with behavior. their parents throughout their first winter and breeding pairs establish territories. and aggression Food abundance at Aransas NWR was not limiting among the subadults was minimal, thereby decreaSing need for conspecific surveillance. the Thus decreased alert behavior of the Grays Lake whooping cranes compared to sandhill cranes of the same flock may be an artifact of their non-breeding sandhill cranes in a breeding population as a result of conspecific surveillance avoidance exhibiting behavior. The foster-reared status. In contrast, may exhibit greater vigilance and not increased predator whooping cranes may be sufficient predator surveillance but not increased alert behavior associated with the social dynamics of a breeding population. �529 It is likely that the whooping vigilance and early warning situation was described (Vanellus vanellus), headed gulls Barnard vigilance and golden warning. Whooping feeding higher interactions apricaria), in mixed rate of responsiveness. plovers flocks and black- (Thompson and the greatest Gulls provided the benefit of added vigilance the and early cranes of the Grays Lake flock may also be in which of this study can only sugg~st which could begin to address Further investigation among sandhill better understand one species possible the social of individual cranes and whooping the possible avenues of dynamics behavior cranes of this and social is needed to effects of this cross-fostering exoeriment. Comparisons associations and interactions between and within the information needed to ascertain could provide A parallel rates of responsiveness. The results unique flock. flock. to alarm of lapwings (Pluvialis from a mixed species association investigation from the added Of the 3 species, gulls exhibited lapwings exhibits for the responsiveness golden plovers and highest benefiting benefit in the heterospecific (Larus ridibundus) 1983). cranes of time budget data and analysis of the 2 crane species the social structure of this flock. LITERATURE CITED Abramson, M. 1979. Vigilance as a factor influencing among curlews Numenius arquata. Ibis 121:213-216. flock formation Allen, R. P. 1952. 3. 246pp. Soc., Res. Rep. The whooping Altmann, J. 1974. Observational Behavior 49:227-267. crane. Natl. Audubon study of behavior: Sampling methods. Barnard, C. J. 1980. Flock feeding and time budgets in the house sparrow (Passer domesticus L.). Behaviour 74:113-127. �530 Bertram, B. C. R. 1980. Behav. 28:278-286. Vigilance and group size in ostriches. Anim. Bishoo, M. A. and D. R. Blankenship. 1982. Dynamics of subadult flocks of whooping cranes on the Aransas National Wildlife Refuge, Texas.. Pages 180-189 in J. C. Lewis, ed. Proc. 1981 Crane Workshop. Natl. Audubon Soc., Tavernier, Fl. Caccamise, D. F. and D. W. Morrison. 1986. Avian communal roosting: implications of diurnal activity centers. Am. Nat. 128:191-198. Caraco, T. 1979. Time budget ing and group size: Ecology 60:618-627. a test of theory. Cowardin, L. M., V. Carter, F. C. Golet, and E. T. LaRoe. 1979. Classification of wetlands and deepwater habitats of the United States. U. S. Dep. Inter., Fish and Wildl. Serv., Biol. Servo Prog., nJSjOBS-79j31. 103pp. Currier, P. J. and J. W. Ziewitz. crane model to the management Pages 315-325 in J. C. Lewis, Natl. Audubon Soc., Tavernier, 1987. Application of a sandhill of habitat along the Platte River. ed., Proc. 1985 Crane Workshop. Fl. Drewien, R. C. 1982. Whooping crane transplant experiment. Rep., Idaho Coop. Wi 1dl. Research Uni t. Job Prog. , and W. M. Brown. 1983. Whooping crane transplant experiment. -----Job Prog. Rep., Idaho Coop. Wildl. Research Unit. , . 1984. Whooping crane transplant experiment. ------JobProg. Rep., Idaho Coop. Wildl. Research Unit. Erickson, R. C. and S. R. Derrickson. 1981. The whooping crane. Pages 104-118 in J. C. Lewis. ed., Crane research around the world, Tntl. Counc. for bird preservation. Intl. Crane Foundation. Baraboo, \H. Howe, M. 1987. Habitat use by migrating whooping cranes in the Aransas-Wood Buffalo corridor. Pages 303-311 in J. C. Lewis, ed. Proc. 1985 Crane Workshop. Natl. Audubon Soc., Tavernier, Fl. Johnson, K. A. and S. A. Temple. 1980. The migratory ecology of the whooping crane. Unpublished report, contract 14-16-0009-78-034, U. S. Fish and Wildl. Serv., Washington, D. C. 120pp. Knight, S. K. and R. L. Knight. 1986. Vigilance patterns of bald eagles feeding in groups. Auk 103:263-272. Krapu, G. L. 1979. Sandhill crane use of staging areas in Nebraska. Pages 1-5 in J. C. Lewis, ed., Proc. 1978 Crane Workshop. Natl. Audubon Soc., Fort Collins, CO. �531 Krapu, G. L., D. E. Facey, E. K. Fritzell, and D. H. Johnson. 1984. Habitat use by migrant sandhill cranes in Nebraska. J. Wildl. Manage. 48:407-417. Krebs, J. R. and N. B. Davies. 1987. An introduction to behavioral ecology. Blackwell Scientific Publ., Oxford, England. 291pp. Lack, D. 1968. Ecological adaptations for breeding in birds. Methuen, London. Lazarus, J. 1979. The early warning function of flocking in birds: an experimental study with captive Ouelea. Anim. Behav. 17:855~ 865. Lingle, G. R., G. A. Wingfield, and J. W. Ziewitz. 1988. The migration ecology of whooping cranes in Nebraska, U. S. A .. Proc. Int. Crane Workshop (in press) Mayhew, P. W. 1987. Vigilance levels in European wigeon - sexual differences. Wildfowl 38:77-81. Melvin S. M. and S. A. Temple. 1982. Migration ecology of sandhill cranes: a review. Pages 73-87 in J. C. Lewis, ed. Proc. 1981" Crane Workshop. Natl. Audubon Soc., New York, N. Y. Powell, G. V. N. 1974. Experimental analysis of the social value of flocking by starlings (Sturnus vuloaris) in relation to predation and foraging. Anim. Behav. 22:501-505. Tacha, T. C. 1981. Behavior and taxonomy of sandhill cranes from midcontinental North America. Ph.D. Thesis, Oklahoma State Un;v., Stillwater. 112pp. Thompson, D. B. A. and C. J. Barnard. 1983. Antipredator responses in mixed -species associations of lapwings, golden plovers, and blackheaded gulls. Anim. Behav. 31:585-593. U.S. Fish and Wildlife Service. 1981. The Platte River ecology study. Spec. Res. Rep., Northern Prairie Wildl. Res. Ctr. Jamestown, N.D. 187pp. Waite, T. A. 1987. Vigilance in the white-breasted nuthatch: of dominance and sociality. Auk 104:429-434. effects Ward, P. and A. Zahavi. 1973. The importance of certain assemblages of birds as "information centers" for food finding. Ibis 115:517534. Weatherhead, P. J. 1983. Two principal strategies in avian communal roosts. Am. Nat. 121:237-243. ��Table Bl. Time spent in diurnal activities by whooping crane #8422 on 18 October 1986 near Hudson, Colorado. Activities were recorded at 15-second intervals. Act iv i ty it J:L__[g~Ji!)g 8..._l!:._ UQ!!~l n % n - - 42 26.6 II a 0.0 87 55. I 0 - 0 - Feeding Orinking Alert - LQafing P. M. reedj!.!~ P. ~1. Roo s t l_Qta 1 % n % 11. 1 304 50.0 3 2.9 360 37.2 0 0.0 0 0.0 2 2.0 2 0.2 54 54.5 223 36.7 45 44. I 409 42.3 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0.0 28 28.3 7 1.1 30 29.4 65 6.7 29 18.3 6 6.1 68 11.2 22 21.6 125 12.9 a 0.0 0 0.0 0 0.0 0 0.0 a 0.0 a 0.0 0 0.0 0 0.0 0 0.0 0 0.0 a 0.0 a 0.0 6 1.0 a 0.0 6 0.6 99 100.0 608 100.0 102 100.0 967 100.0 n % n % n % ):::0 "U \J Hesting rn Z 0 I-< Cotuf ur l Locomotion Agonistic Vocal izat ion - Gut of Sight - - . - - - - Total data __ "_ 158 >< CP .. 100.0 points Minutes of 39.516.3 24.8 10.2 212.0 62.9 25.5 10.6 241.8 100.0 observat ion (%) _.----- . __ __ ._ .. .. -._ -- --- ----- --_ ...... _._---_.- .- t/l w w �VI LV .&:- Table B2. Time spent in diurnal activities by whooping crane #8422 on 19 October 1986 near Hudson, Colorado. Activities were recorded at 15-second intervals. Activity 8.:.-.l:L_nQQ~! lL_M. Feedjng IQ1~1 43.9 0 0.0 397 30.2 ·0 0.0 0 0.0 1 O. 1 51.0 188 4 1.0 60 100.0 600 45.7 2 0.7 0 0.0 0 0.0 2 0.2 0.4 92 31.3 3 0.7 0 0.0 97 7.4 104 20.8 39 13.3 65 14.2 0 0.0 208 15.8 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 n Feeding - - 192 Drinking - Alert - Resting - Comfort - locomotion - Agonistic n Ll'L_Boo ~ t % % - f..:_l!Jgg!J in9 n n - loafing n % 38.3 4 1.4 201 0 0.0 1 0.3 202 40.3 150 0 0.0 2 % % n % Vocalization - 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 Ollt of Sight - I 0.2 6 2.0 1 0.2 0 0.0 8 0.6 294 100.0 458 100.0 60 100.0 1313 100.0 .--- .. ~~.--- Total data - SOl - 125.3 38.2 100.0 point.s Minlltes of observation -_._----_. (%) 73.5 22.4 114.5 3·1.9 15.0 4.6 328.3100.0 �Table 83. Time spent in diurnal activities by whooping crane #8422 on 20 October 1986 near Hudson, Colorado. Activities were recorded at I5-second intervals. - -- - .._--_--- -_-_ - -- --_ .. -- -_- _- .__---_ _--- -----_._- ------_- .. Activity .... 1l.:_lL_fl!2 0 s t L!:L_f£~gin9 n % % n _._--------------- loaiil}9 n .. IQL!l L11,_£f~ging r.!._lL_noo~l % n % n % n % - 370 28.4 - 0 0.0 544 41.8 0 0.0 59 4.5 ------ - feeding I 0.9 218 44.7 151 21.4 - Drinking 0 0.0 0 0.0 0 0.0 - 62 57.4 174 35.7 308 43.6 - - - Resting 0 0.0 0 0.0 0 0.0 - - - Comfort 6 5.6 1 0.2 52 7.4 - 37 34.3 92 194 27.4 323 24.8 A~Joll i s L ie 0 0.0 0 0.0 0 0.0 0 0.0 Vocal i za t ion 0 0.0 0 0.0 0 0.0 0 0.0 Gilt of Sight 2 1.8 3 0.6 1 0.3 6 0.5 488 ]00.0 706 ]00.0 ]302 100.0 Alert I ucomo t ion Total data ]08 ]00 10.0 - - - points Minlltes of 27.0 8.3 122.0 37.4 176.8 325.8 100.0 54.3 observa lion (%) ._. o__ _"_ _______ _ _ VI W VI �Table B4. Time spent in diurnal activities by whooping crane #8422 on 21 October 1986 near Hudson, Colorado. Activities were recorded at 15-second intervals. A. ,.1. Roost Activity n % A. M. Feeding % n Loafing n % P. M. Feeding % n -- P. ,.1. Roost II % Total n % - feeding 69 46.0 69 46.0 Orinking o 0.0 o 0.0 69 4G.0 9 46.0 o 0.0 o 0.0 J\lel'l Resting Comfort 0.7 0.7 Locomotion 9 6.0 9 6.0 Agonistic o 0.0 o 0.0 Vocal izalion o 0.0 o 0.0 2 1.3 2 1.3 Ollt of Sight ---- Tolal dala ---------- 150 100.0 150 100.0 points Minutes of 37.5 100.0 observation ------_ .. (%) _-- -----_. 37.5 100.0 Vl W o- �Table 8S. Time spent in diurnal activities by whooping crane #8422 on 22 October 1986 near Hudson, Colorado. Activities were recorded at IS-second intervals. Ac t i v ity tL_f:L_H(H1H 8.!_!L_[~Quing n L!~i![iOg n f_,__!:1!_£~~Ui!lg E,-_ !:L_nQQ~t n % n - - 389 19.3 - 0 0.0 - 1152 57.3 0 0.0 55 2.7 - 408 20.3 - - 0 0.0 0.0 - - 0 0.0 0 0.0 - - 7 0.4 295 100.0 2011 ]00.0 n n % Feeding 15 15.0 197 29.1 67 7.1 110 37.3 Drinking 0 0.0 0 0.0 0 0.0 0 0.0 43 43.0 299 44. 1 690 73.6 120 40.7 - Resting 0 0.0 0 0.0 0 0.0 0 0.0 - COlilfort 4 4.0 4 0.6 47 5.0 0 0.0 37 37.0 173 25.5 133 14.2 65 22.0 - Agonistic 0 0.0 0 0.0 0 0.0 0 0.0 Vocalization 0 0.0 0 0.0 0 0.0 0 Out of Sight ] 1.0 5 0.7 ] O. ] 100 100.0 678 100.0 938 100.0 Alert locomotion Total data % % I111i!1 % % points Minutes of observation 25.0 5.0 169.5 33.7 234.5 46.6 73.8 14.7 502.8 100.0 (%) VI VJ --.J �\Jl VJ Table 86. Time spent in diurnal activities by whooping crane #8422 on 23 October Hudson, Colorado. Activities were recorded at IS-second intervals. 1986 near ----- Activity A. M~Qll A. ~1. n % Feeding 0 0.0 449 Drinking 0 0.0 16 Resting LtL BQQ~t n % n % n 26.4 - - - - 0 0.0 - - - 80.0 812 47.8 - - - 0 0.0 0 0.0 - - - Comfort 2 10.0 64 3.8 - - - locolllotion 2 10.0 ·361 21.2 AC]onistic 0 0.0 0 0.0 - - - - Vocal ization 0 0.0 0 0.0 - - - - Out of Sight 0 0.0 14 0.8 - - - Alert. Total data 20 n feeding Loafing ~_lL_feeding % 100.0 1700 100.0 IQtd mb~r n % - 2 1.4 451 24.2 - - 1 0.7 1 0.1 - - 59 40.4 887 47.5 - 0 0.0 0 0.0 17 11.6 83 4.5 • 67 45.9 430 23.0 - 0 0.0 0 0.0 - 0 0.0 0 0.0 0 0.0 14 0.7 146 100.0 1866 100.0 - - % n % points Millllles of observation 5.0 (%) 1.1 425.0 91.1 36.5 7.8. 466.5 100.0 co �Table B7. Time spent in diurnal activities by whooping crane #8422 on 24 October 1986 near Hudson, Colorado. Activities were recorded at 15-second intervals: --- I\ct iv i ty LtL B!'!QH LtL_£~~~jj!!g Loafing n % n feeding - - 173 Orinking - - Alert - Resting __ --'---------- - L!L_£f_l1~U !EJ --- - - - e e.; - - -------------------- IQi!!l tL__BQQH n % n % n % n 30.6 I 0.4 - - 25 3.3 199 12.7 0 0.0 0 0.0 - - 0 0.0 0 0.0 - 269 47.6 217 87.5 - - 444 59.1 930 59.5 - - 0 0.0 0 0.0 - - II 1.5 II 0.7 Comfort - - 2 0.4 21 8.5 - - 144 19.2 167 10.7 Locomotion - - 115 20.3 8 3.2 - 11.8 212 13.5 I\gonistic - 0 0.0 0 0.0 - 0 0.0 0 0.0 Vocal izat ion - 0 0.0 0 0.0 - 0 0.0 0 0.0 Out of Sight - 6 1.1 I 0.4 - 38 5. I 45 2.9 Total data - 565 100.0 248 100.0 - 751 100.0 1564 100.0 15.9 - - - % 89 - - % . points Minutes of observation - 141. 3 36.1 62.0 107.8 48.0 391.0 100.0 (%) V1 w \0 �VI Table B8. Time spent in diurnal activities by whooping crane #8422 on 25 October 1986 near Hudson, Colorado. Activities were recorded at I5-second intervals. -------------------------------------- Activity A. 11. Roost n % A.I·1. Feeding n Loafing n % % P. t·1. Feeding n --------- P. H. Roost n % --------- -- --- -- % Iotal n % ._--- Feeding 0 0.0 182 26.6 2 1. 4 2 0.6 186 14.2 Drinking 0 0_0 a 0.0 o 0.0 0 0.0 o 0.0 52 30.0 365 53.3 61 41.2 177 52. I 655 50.0 Resting 0 0.0 o 0.0 o 0.0 12 3.5 12 0.9 Comfort 46 33.6 8 1.2 28 18.9 )27 37.3 209 16.0 Locomotion 35 25.5 117 17 . I 48 32.4 20 5.9 220 16.8 Agonistic 0 0.0 o 0.0 o 0.0 0 0.0 o 0.0 Vocal i za l i on a 0.0 o 0.0 o 0.0 0 0.0 a 0.0 Out of Sight 4 2.9 13 1.9 9 6. I 2 0.6 28 2. 1 100.0 1310 100.0 Alert --_---------_. ------- Total data ]37 100.0 685 100.0 )48 100.0 340 points Minutes of observation 34.3 ]0.5 171.3 52.3 37.0 11.3 (%) ------------ __ .. -- -- --_._. -_ .. ------------------------- ..--_ .. 85.0 25.9 327.5 ]00.0 ~ o �Table 89. Time spent in diurnal activities by whooping crane #8422 on 26 October Hudson,Colorado. Activities were recorded at 15-second intervals. Activity 8J:L_ BQ~!~i n Feeding % - Drinking A.M. feeding ---------.~ !.Q2f in9 £_,!:L__£QgJ in<J % n % n 98 19.4 - - - 0 0.0 n % 1986 near £ JL_BQQ~ t n % n - 98 19.4 0 0.0 - 297 58.7 - 0 0.0 9 1.8 - 102 20.2 - 0 0.0 - 0 0.0 - 0 0.0 - - JQL.!!l % Alerl - 297 Restillg - 0 0.0 Comfort - 9 1.8 102 20.2 - 0 0.0 - - 0 0.0 - - 0 0.0 - 506 100.0 150 100.0 126.5 100.0 126.5 100.0 I.ocolllotion Agonistic - Vocalization - Out of Sight - Tota 1 - 5£3.7 . - - - data points ~1inlltes of observation (%) VI .j::-- •..... �V1 po. N Time spent in diurnal activities by whooping crane N8422 from 18-26 October 1986 near Table BID. Hudson,Colorado. Activities were recorded at IS-second intervals. --------" -_._- ----- .. -.--.---.-~----_ Activity BJL_E!'!Q~! BJ:L_Fegg ina n n % % --_ LQa finn n % ~_,_lleding n % ~-~ r.J:L_.BQQ~l n IQli!l Q!J!~[ % n % n % 2.9 ---------- Feeding 16 4.4 J620 29.8 234 Drinking 0 0.0 0 0.0 I 173 47.4 2574 47.4 Alert 617 40.9 30 2.4 2 1.4 519 O. I 0 0.0 1 O. I 1 0.7 4 1419 62. I 592 39.2 726 57.9 59 40.4 5543 50.4 10.2 o. I Resting 0 0.0 0 0.0 2 0.1 0 0.0 23 1.8 0 0.0 25 0.2 Comfort 58 15.9 91 1.7 240 10.5 38 2.5 301 24.0 17 11.6 745 6.8 III 30.4 1102 20.3 380 16.6 246 16.3 131 10.6 67 45.9 2037 18.5 Agonistic 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 Vocalization 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 Out of Sight 7 1.9 44 0.8 9 0.4 16 1.1 40 3.2 0 0.0 116 1.1 365 100.0 5431 100.0 2285 100.0 1509 100.0 1253 100.0 146 100.0 10989 100.0 locomotion Total data points MillllLes of observation 90.3 3.3 (%) 1356.8 49.4 570.3 20.8 376.3 13.7 312.3 11.4 36.5 7.8 2746.3 100.0 �543 APPENDIX C Table CI. Time spent in diurnal activities by whooping crane #8422 on 3 November 1986 in the San Luis Valley, Colorado. Activities were recorded at IS-second intervals. Activity A. M. Feeding n Feeding % P. M. Feeding n % n % 718 67.0 87 79.1 805 68.2 0 0.0 0 0.0 0 0.0 218 20.4 14 12.7 232 19.6 Resting 0 0.0 0 0.0 0 0.0 Comfort 8 0.7 1 0.9 9 0.8 126 11.8 8 7.3 134 11.3 Agonistic 0 0.0 0 0.0 0 0.0 Vocalization 0 0.0 0 0.0 0 0.0 Out of Sight 1 0.1 0 0.0 1 0.1 1071 100.0 110 100.0 1181 100.0 Drinking Alert Locomotion Total data points Minutes of observation 317.8 (%) 90.7 27.5 9.3 355.3 100.0 �544 Table C2. Time spent in diurnal activities by whooping crane #8422 on 4 November 1986 in the San Luis Valley, Colorado. Activities were recorded at IS-second intervals. Activity A. M. Feeding n Feeding % P. M. Feeding n % n % 553 67.2 431 79.7 984 72.2 a 0.0 0 0.0 a 0.0 132 16.1 55 10.2 187 13.7 Resting a 0.0 0 0.0 a 0.0 Comfort 53 6.4 0 0.0 53 3.9 Locomotion 80 9.7 55 10.1 135 9.9 Agonistic a 0.0 0 0.0 a 0.0 Vocalization a 0.0 0 0.0 a 0.0 Out of Sight 4 0.5 0 0.0 4 0.3 822 100.0 541 100.0 1363 100.0 Drinking Alert Total data points Minutes of observation 205.5 (%) 60.3 135.3 39.7 340.8 100.0 �545 Table C3. Time spent in diurnal activities by whooping crane #8422 on 5 November 1986 in the San Luis Valley, Colorado. Activities were recorded at IS-second intervals. Activity A. t~. Feeding n Feeding % Total P. M. Feeding n % n % 353 77 .2 282 77 .3 635 77 .3 0 0.0 0 0.0 0 0.0 66 14.4 35 9.6 101 12.3 Resting 0 0.0 0 0.0 0 0.0 Comfort 1 0.2 0 0.0 1 0.1 36 7.9 48 13.1 84 10.2 Agonistic 0 0.0 0 0.0 0 0.0 Vocalization 0 0.0 0 0.0 0 0.0 Out of Sight 1 0.2 0 0.0 1 0.1 457 100.0 365 100.0 Drinking Alert Locomotion Total data 822 100.0 points Minutes of observation 114.3 (%) 55.6 91.3 114.4 205.5 100.0 �546 Table C4. Time spent in diurnal activities by whooping crane #8422 on 6 November 1986 in the San Luis Valley, Colorado. Activities were recorded at IS-second intervals. A. r~. Feed i ng Activity n Feeding % P. M. Feeding n % n % 134 87.0 400 78.0 534 80.1 0 0.0 0 0.0 0 0.0 12 7.8 76 14.8 88 13.2 Resting 0 0.0 0 0.0 0 0.0 Comfort 0 0.0 1 0.2 1 0.1 Locomotion 8 5.2 36 7.0 44 6.6 Agonistic 0 0.0 0 0.0 0 0.0 Vocalization 0 0.0 0 0.0 0 0.0 Out of Sight 0 0.0 0 0.0 0 0.0 154 100.0 513 100.0 667 100.0 Drinking Alert Total data points t~inutes of observation 38.5 (%) 23.1 188.3 76.9 226.8 100.0 �547 Table C5. Time spent in diurnal activities by whooping crane #8422 from 3-6 November 1986 in the San Luis Valley, Colorado. Activities were recorded at IS-second intervals. Activity A. M. Feedina n Feeding % P. M. Feedina % n n % 1758 70.2 1200 78.5 2958 73.3 0 0.0 0 0.0 0 0.0 428 17.1 180 11.8 608 15.1 Resting 0 0.0 0 0.0 0 0.0 Comfort 62 2.5 2 0.1 64 1.6 250 10.0 147 9.6 397 9.8 Agonistic 0 0.0 0 0.0 a 0.0 Vocalization 0 0.0 a 0.0 a 0.0 Out of Sight 6 0.2 0 0.0 6 0.2 2504 100.0 1529 100.0 4033 100.0 Drinking Alert Locomotion Total data points Minutes of observation 626.0 (%) 62.1 382.3 37.9 1008.3 100.0 ��549 CHAPTER 3 ROOST SITE SELECTION OF WHOOPING CP~ES AT NON-TRADITIONAL FALL MIGRATION STOPOVER SITES IN EASTERN COLORADO SUMMARY Roost site selection investigated by whooping at 2 non-traditional in eastern Colorado. cranes (Grus americana) fall migration stopover NTSS were used by 2 whooping fostered flock during fall 1985. One whooping cranes was sites (NTSS) of the cross- crane returned" to the same NTSS it used in fall 1985 during spring and fall migration stopovers of 1986. Two roosting near Hudson, Colorado Severence NTSS. wetlands were used at the NTSS located and 1 wetland was used for roosting Characteristics common among roosting included areas of water depth < 30 cm, isolation close proximity wetlands. conductivity « and distances There was no consistency horizontal Whooping from disturbance, Both Hudson roost sites were similar in size, However, these characteristics site. wetlands 2 km) to feeding sites, loafing areas, of water, visibility at the and similar specific to power lines and fences. were not shared by the Severence among the 3 wetlands or vegetation and composition migration greater sandhill crane (Grus canadensis of and density. cranes of the Grays Lake experimental exhibit the gregarious for percent roost flock usually behavior common to their tabida) parents. foster These cross- �550 fostered whooping cranes migrate as individuals, or occasionally in groups of 2 or 3, within the sandhill crane flock (R. C. Drewien, Prog. Rep. Whooping Crane Egg Transplant winters near Bosque-del-Apache Experiment 1-20). This crane flock National Wildlife Refuge (NWR), New Mexico and summers at Grays Lake NWR, Idaho (Stahlecker 1986). Normally they concentrate in the traditional spring and fall stopover sites in the San Luis Valley of southcentral Colorado. fall migration During the 1985 however, 2 subadult whooping cranes from the Grays Lake flock (Pat 4 and #8422) used 2 separate non-traditional (NTSS) in eastern Colorado (Dennis 1985). stopover sites Whooping crane #8422 returned to its previously used eastern Colorado NTSS during the 1986 spring and fall migration. Evaluating information cranes. habitat on these NTSS will add to the limited available on 'habitat requirements of migrating whooping Comparing roost site selection at NTSS with roost sites used at TSS could provide insights into habitat use changes associated with the presence or absence of a surrounding the use of non-traditional management flock. Information concerning stopover sites could also be applied to of whooping cranes during any occurrence outside of the normal range. The objectives biological and of this study were to: and physical characteristics (2) develop standardized methodology of non-traditional comprehensive sites. (1) measure selected of NTSS in eastern Colorado, for future habitat evaluation stopover sites to facilitate consistent, data collection for any future use of non-traditional �551 STUDY AREA The 2 NTSS were located in southern Weld County of eastern Colorado. The NTSS used by whooping crane Pat 4 during fall 1985 was located 1 km east of Severence, Colorado. was used at this site. One roosting wetland (S[l]) Two roosting wetlands (H[1], H[2]) were used at the NTSS used by #8422 during fall 1985 and spring and fall 1986. site was located 2 km south of Hudson, Colorado. This Topography is broadly rolling with an elevation of 1512 m at the Hudson site and 1477 m at the Severence site. Irrigated farmland in the area supports the main crops of corn, alfalfa, wheat, and sugar beets. vegetables are grown in lesser amounts. Malting barley and The 3 roosting wetlands were classified as (system) palustrine, (class) emergent wetland, and (modifier) saturated/semi-permanent/seasonal wetlands (PEMY; Cowardin et al. 1979). The Severence roosting wetland used by whooping crane Pat 4 was an irrigation reservoir surrounded by dense cattails (IYQh£ latifolia). The feeding site used at this NTSS was an open corn stubble field 0.1 km east of the roost site. Feeding sites used by whooping crane #8422 at the Hudson NTSS were fields of volunteer barley, corn, wheat, and hay, all located < 1.5 km from the roost sites. Both roost sites used at the Hudson NTSS were stock ponds located in grazed cattle pastures. Loafing areas were adjacent to the roosting wetlands at both NTSS. METHODS Habitat evaluations were conducted on the 3 NTSS roosting wetlands. Criteria for habitat evaluations were based on published literature concerning habitat requirements of migrating cranes (Allen 1952; Johnson and Temple 1980; Lingle et al. 1984, 1988; Howe 1987; �552 Ward and Anderson 1987}. Variables and techniques used for measurement were described in chapter 1. A standardized field form was developed by combining the U. S. Fish and Wildlife Service and Colorado Division of Wildlife whooping crane sighting report forms. Additional questions were prompted by current literature and field observations (Appendix D). RESULTS Whooping crane Pat 4 used the Severence NTSS from 23 September to 29 October 1985. S(I}, a 21.0-ha irrigation reservoir, was located only 6 m from a road and 75 m and 78 m from the nearest building and house respectively (Table 3.1). Visual isolation from these potential disturbances was proyide by dense cattails (Typha latifolia) surrounding the northern end of the wetland. These cattails were> 3 m high and resulted in only 10% horizontal visibility from the 20-cm water depth contour (Fig. 3.1). Pat 4 roosted in the shallow northern end of the wetland (Fig. 3.2), but the specific location of the roost was unknown. The greatest distance from shoreline to the 20-cm water depth contour was 84.0 m. Distance to feeding site was 0.1 km. Whooping crane #8422 used the Hudson NTSS from 11 September to 29 October 1985, 15-29 April 1986, and 17-26 October 1986. The same roosting wetland (H[l]) was used during all 3 stopover periods and an additional roosting wetland (H[2]) was used 2 nights during fall 1986. H(l), a l.l-ha stock pond, was characterized by extensive horizontal visibility (Fig. 3.1) and isolation from potential disturbance and hazards (Table 3.1). The wetland was surrounded by grazed pasture with sedges (Carex spp.) at waters edge. Horizontal visibility was extensive in all directions and visual obstruction by �Table 3.1. Habitat evaluations for the 3 roosting wetlands located at HISS in eastern Colorado. -- Roost Dis tance to 12Q1en tiAl__Qli!J!rbance Specific Conductivity location Size (Site) (ha) !IO!!5e Roag {luilging of IJater Oi~t~nce to potential Fence (nuuho sZcm) (Ill ) hazard Power line (m) --._._- S (1) 21.0 78 6 75 2600 0 45 II( 1) 1.1 660 630 600 4400 16 360 1I(2} 1.3 600 700 600 3700 0 700 VI VI W �554 Fig. 3.1. Mean (± SO) horizontal visibility wetlands in eastern Colorado. of 3 NTSS roosting �555 { .: ! .•. / " " " i ; water depth '\ , I I Wetland edge 10 cm contour 20 cm contour 30 cm contour 40 cm contour / (' j ,/" / / ./ I 20 m Fig. 3.2. Water depth profile of wetland S(l). �556 west end of the wetland. The roost location was in the area of most vegetation occurred only in the lower 0.2S-m section of the profile board. Other obstructions resulted from topography and a dike on the gradual slope and greatest distance (6 m) from shoreline to the 20-cm water depth contour (Fig. 3.3). The second roost site used at the Hudson NTSS was similar to H(l) in distance from potential disturbances and hazards, wetland size, and specific conductivity. Horizontal visibility was obstructed at the eastern end of H(2) by cattails and on the west end by a dike. The cattails had been beaten down by a hail storm and only obstructed view below 1.5 m (Fig 3.1). roosting. Whooping crane #8422 used 2 specific sites for Both roost sites occurred away from the cattails in are~s where sedges and spikerushes emergent vegetation. (Eleocharis spp.) were the dominant Water depth was < 40 cm throughout the wetland (Fig. 3.4). Four feeding sites were used by #8422 at the Hudson NTSS, all located within a 1.S-km radius of H(I). H(2) was 1.5 km. The distance between H(I) and When using roost H(2), ~8422 used the same feeding sites as when roosting at H(I). DISCUSSION Whooping cranes are evidently tolerant of a broad range of natural and human-modified habitats during migration (Johnson and Temple 1980, Erickson and Derrickson 1981, Lingle et al. 1984, Howe 1987, Ward and Anderson 1987). Despite evidence that migrating whooping cranes are habitat generalists certain patterns of habitat selection are apparent. Johnson and Temple (1980) summarized wetland characteristics for roost sites used by whooping cranes of the Aransas-Wood Buffalo flock. These �557 ... ,-.;- .................. ---------=----=-------:----::- I I \ ~.;.~7:_::..~:.:_:.:::::::-:.:-~:.:::-:.:-.--::-~:..~.~.~ _-=-_ -.~.~:.:::-;::::: . ::-::-:. _ . ~ - ... , . ,;; ':. " " /? . ;(- .' , '\ \, ". . . \ ' " " , .. ' , : : , , . , . : " ' \ 40 cm > ".. '-. , ':.:: \ . water depth . ,, ,:' /p ".' 7 \._, , " , " " '. " " " " " \. , ., ., '," " ""''''' " '. \. ," ,.". ,':, ' .' . \ \ \ /'/;/ I ~~.~:' f r ,,' ',,' " ./' : .' :.: . ...... - r , , r ', " .'/ , \ . ,, , ' ·· ., ·, , . ,' ,. ', . ' , ,, . . ' , , ,, '' , ' " " " " ' " ., " " ; j ,, , , , , : ,: ," , , , . Wetland edge 10 em contour 20 em contour 30 em contour 40 em contour " " ' ; ! " ., '' , .' , ' ,, ' ,, '. , , ,, , , ' , " ·· ,' : : I , , , . ~ , :: . ,,, , 5m , \ :~ \: \ I. ".: '. \ " \" , . ,: ....•• ' ••••--:-+-1 __ - Roost location •• Fig. 3.3. Water depth profile of wetland H(l). �558 ........................................... - ._- .... ---.--------. .................. ........................... -._------ ) .----------- ------------ ----------_------ ~ \ J '\::.',: '.\ .... . <.; ;~~.~ \ x.'. , ~ <, ~ ........ ••••.•••• Roost location , , , , '~': \ ' ~:, I ... : I ~: ; , '. " <, , : , , ; : ' .. : : , , < ' : ' : , : , : , : , .: ', : , : , \. 40 em water depth : \ \, [ 1 , .' / i :, I ' I' I: I: i\ \ I / '. , / / ' \ ',' i t / / Wetland edge 10 em contour 20 em contour 30 cm contour 40 em contour l- I 10 m Fig. 3.4. Water depth profile of wetland H(2). �559 included shallow water, extensive horizontal visibility, either sparse or short emergent vegetation, close proximity to feeding sites, and isolation from potential hazards and disturbance. characteristics The roost site of the 2 eastern Colorado NTSS used by foster-reared whooping cranes fit well within the parameters described by Johnson and Temple (1980). When compared with the TSS in the SLV used by cranes of the Grays Lake flock however, several characteristics inconsistent, particularly, were distances to potential disturbance and hazards. In the SLV, roost sites are frequently located near roads and power lines. The close pr~ximity of roost sites to power lines in this region is a result of extensive use of center pivot sprinklers for irrigation. Several possible explanations for the lack of isolation from disturbance in the SLV include: (1) the familiarity of the site as a result of traditional use, (2) the gregarious behavior of the Grays Lake whooping cranes provides additional flock vigilance, and (3) a habituation to human disturbance resulting from the greater accessability of breeding, stopover, and wintering grounds (see chapter 1). The roost sites used by Grays Lake whooping cranes at the NTSS were isolated from disturbance. Alert behavior of whooping crane #8422 increased at the NTSS when compared with behavior at the TSS due to the loss of a surrounding flock (see chapter 2). The greater isolation of the NTSS roosting wetlands might also be a response to loss of flock vigilance and unfamiliarity with the sight. Hudson NTSS were isolated by distance. The 2 roost sites at the The Severence NTSS roosting wetland was visually isolated from disturbances stand surrounding by a dense cattail the used portion of the wetland. Although whooping �560 cranes are thought to avoid areas of dense vegetation, emergent aquatic plants such as cattails and bulrush (Scirous spp.) might be an exception (W. M. Brown, pers. commun.). Suitable roosts usually consist of open water that enables whooping cranes to detect, either visually or audibly (splashing water), the approach of a predator (Ward and Anderson 1987). It would be difficult for a predator to approach a roosting crane through a stand of cattails or bulrush without detection because they are dense emergent aquatic plants. Whooping crane #8422 used a complex of 2 roost sites, 4 feeding sites, and 2 loafing areas at the Hudson NTSS. The bird also landed at 2 other wetlands for brief periods following hazing by dogs at a feeding site and a loafing area. Although Pat 4 used only 1 roost site and adjacent loafing area and 1 primary feeding site, it also infrequently used 2 surrounding pastures and 1 wetland while loafing and feeding. This pattern of using a complex of required activity sites (roost, feeding, and loafing areas) is similar to that observed in whooping cranes at the TSS in the SLY (see chapter 1). Use of such a complex provides cranes with sufficient habitat in the event disturbance occurs at anyone of the used sites. When flushed from an activity site or when humans were present at the site as cranes attempted to fly in, whooping cranes in the SLY flew to other nearby suitable activity sites (chapter 1). Whooping cfane #8422 flew to H(2) after being flushed from the loafing area adjacent to H(l). The bird remained to loaf at the new site and later returned there to roost after its evening feeding period. This use of complexes should be considered in management of future NTSS. Disturbance should be minimized at several suitable roosts, feeding sites, and loafing areas �561 to provide the whooping crane sufficient habitat in the event disturbance occurs on any of the used activity sites. Habitat use and behavior of whooping cranes at future NTSS in Colorado should be documented in a consistent manner. A standardized field form (Appendix 1) should be used for habitat evaluations to facilitate comparisons with other whooping crane stopover sites and to add to the limited information available on requirements of this species during migration. LITERATURE CITED Allen, R. P. 3.246pp. 1952. The whooping crane. Natl. Audubon Soc., Res. Rep. Cowardin, L. r1., V. Carter, F. C. Golet, and E. T.· LaRoe. 1979. Classification of wetlands and deepwater habitats of the United States. U. S. Dep. Inter., Fish and Wildl. S.erv., Biol. Servo Prog., FWS/OBS-79/31. 103pp. Dennis, J. A. 1985. Whooping cranes in eastern Colorado. Field Ornith. J. 19:77-78. Colorado Erickson, R. C. and S. R. Derrickson. 1981. The whooping crane. Pages 104-118 in J. C. Lewis, ed. Crane research around the world, Int. Counc. for bird preservation. Int. Crane Foundation. Baraboo, WI. Howe, M. 1987. Habitat use by migrating whooping cranes in the Aransas-Wood Buffalo corridor. Pages 303-311 in J. C. Lewis, ed. Proc. 1985 Crane Workshop. Natl. Audubon Soc., Tavernier, Fl. Johnson, K. A., and S. A. Temple. 1980. The migratory ecology of the whooping crane. Unpublished report, contract 14-16-0009-78-034, U. S. Fish and Wildl. Serv., Washington, D. C. 120pp. Lingel, G. R., P. J. Currier, and K. L. Lingle. 1984. Physical characteristics of a whooping crane roost site on the Platte River, Hall County, Nebraska. Prairie Nat. 16:39-44. Lingel, G. R., G. A. Wingfield, and J. W. Ziewitz. 1988. The migration ecology of whooping cranes in Nebraska, U. S. A .. Proc. Int. Crane Workshop (in press). Stahlecker, D. W. 1986. The 1985 fall crane migration in the Rio Grande Valley, New Mexico. Unpubl. Rep., New Mexico Audubon Soc., Albuquerque. 22pp. �562 Ward, J. P., and S. H. Anderson. 1987. Roost site use versus preference by two migrating whooping cranes. Pages 283-288 C. Lewis, ed. Proc. 1985 Crane Workshop in J. �563 APPENDIX D WHOOPING CRANE MIGRATION STOPOVER SITE FIELD FORM Date: Recorder: Name -------------------- Telephone Agency/Organization Number of birds at the site: _ Adul ts _ Young _ Accompanying sandhill cranes Color-markings observed (include order of colored bands and left/right 1eg) : Time and duration of sightings: Location (direction and distance from nearest town, and legal description of site; township, range, and section location) Roosting wetland: _ Feeding site: Loafing area: I. Roost Site Description 'n'etlandsize: Wetland classification (Cowardin et al. 1976): Distance to nearest: Potential disturbance: Potent ia1 hazard: 1 ) house 1 ) power 1ine road 2) fence 3) building 3) other 2) 4) other �564 Water depth profiles: Derive from transects established wetland. at 10-m intervals across the Measure water depth along each transect, then record perpendicular distance to shoreline for water depths of 10, 20, 30, and 40 cm. Distance from shore to water depth (m) Transect 10 cm 20 cm 30 cm 40 cm A B C etc. Horizontal visib,lity: Measured with a 3-m high vegetation profile board divided into 0.5- m sections, with the first 0.5 m section sub-divided into 0.25-m increments. Viewing points are set at 10-m intervals along the 20- cm water depth contour of the wetland. 90-m transect is established From the viewing point, a perpendicular to the shoreline and percent visibility is measured. Percent visibilitv Transect A B C etc. at profile board increments (m) 0-0.25 .25-.50 .50-1.0 1.0-1.5 1.5-2.0 2.0-2.5 2.5-3.0 �565 Describe surrounding upland habitat types, land use: II. Feeding Site Description land use (agricultural etc.): Available grain: Distance from roosting wetland: III. Loafing Area Descriotion Land use: Distance from rODsting wetland: Comments: ��567 CHAPTER 4 MANAGEMENT IMPLICATIONS The following management recommendations are based on data collected during fall 1986 and 1987 in the San Luis Valley and at 2 non-traditional stopover sites in eastern Colorado. Due to small sample sizes and local conditions, such as agricultural these recommendations geographic areas. practices, might require alteration to accommodate other Management of roost site habitat for migrating whooping cranes (Grus americana) in the San Luis Valley should include maintenance of a complex of suitable roost sites (> 2), feeding sites, and loafing areas located within an area of 5 km. characteristics The physical of the roosting wetlands should include: (1) sufficient area of shallow water « 30 cm) to allow cranes to roost away from the waters edge, and (2) extensive horizontal visibility within a 90-m radius of the roost site. Juxtaposition of required habitats (roost, feeding, and loafing areas) could be used to manipulate areas of use. Provision of loafing areas near distant feeding sites ( > 4 km from the roost complex) might improve use rates. to be considered If distant feeding sites are part of the complex, loafing areas should be located within 3 km of the feeding site. This strategy could be used to support a higher density of birds or to compensate feeding sites near the roost. for the loss of To prevent abandonment, hunting and �568 other potential disturbances minimized, especially should be prohibited or at least at the roost complex. Development migration stopover sites should incorporate the landscape of suitable approach of a complex of suitable roosting wetlands, feeding sites, and loafing areas. A complex should not be developed where power lines divide adjacent required habitats. Non-traditional stopover sites (NTSS) may require g~eater isolation to compensate for loss of familiarity and surrounding flock vigilance. Throughout the stopover duration, disturbance should be minimized at several suitable roosts, feeding sites, and loafing areas to provide the whooping crane alternate undisturbed sites. non-traditional Habitat evaluations stopover sites should· be standardized to ensure the compatibility of (see chapter 3) of information collected over years. Behavior and daily activity patterns of whooping cranes at NTSS should be documented for comparisqn naturally-reared with whooping cranes in the SLV as well as whooping cranes during migration. These data could provide information on the social status of the cross-fostered cranes within the Rocky Mountain greater sandhill crane (Grus canadensis tabida) flock. Approved by ~ X es K. Ringe Go~ a whooping � https://cpw.cvlcollections.org/files/original/3fdb05925ec8868447a51a6307e7fb68.pdf 886ac70fd31f450337144245c1d476c1 PDF Text Text ~U~ULciUU u~v~s~on or W1iQi1re Wildlife Research Report July 1988 1 JOB PROGRESS REPORT State of Colorado Project No. W-153-R-2 Mammals Research ------------------------ Work Plan No. 1 --~~---------------- Multispecies Investigations Job No. 7 Period Covered: July 1, 1987 - June 30, 1988 Authors: Personnel: Terrestrial Research Publication, Editing and Library Service M. W. Hershcopf, L. H. Carpenter R. B. Gill, N. McEwen, L. Lovett, K. Chociej, R. Harrison, and all Mammals Researchers ABSTRACT During the 1987-88 Segment, 11 books were purchased for permanent reference by DOW personnel. Sixty additional publications were located, ordered, and obtained free of charge for use. Twenty-five theses were purchased, obtained on Interlibrary Loan, or given to the library. An additional 890 individual references requested by Mammals Researchers were located by library staff and made available for use. About 25 of these requests were not available locally and were obtained through interlibrary loans. Seven manuscripts were published in various journals. ��3 MAMMALS PUBLICATION EDITING AND LIBRARY SERVICES Marian W. Hershcopf and Len H. Carpenter P. N. OBJECTIVE To provide a centralized support program for manuscript editing and library services to facilitate publishing results of research conducted in projects 01-03-047 - 11700 and 01-03-048 - 11700, 11400, 16700. SEGMENT OBJECTIVES 1. To provide coordinated and efficient editing and library services and publish findings for all Colorado Mammals Research programs. 2. To provide for the centralized support program for Mammals Research editing, library, and publishing services so that Mammals Research scientists can be most efficient in publishing results of their research. SUMMARY OF SERVICES Publications purchased with Mammals Research funds and placed in the Research Center Library Chicago manual of style. 13th ed., revised and expanded. Chicago Press. Chicago. 738pp. Endler, J. A. 1986. Natural selection in the wild. Press, Princeton, N.J. 336pp. University of Princeton University Inglis, J. M., B. A. Brown, C. A. McMahan, and R. E. Hood. 1986. Deer-brush relationships on the Rio Grande plain, Texas. Kleberg studies in Natural Resources. Texas A&M University, Texas Agricultural Experiment Station, College Station. 80pp. Hason, C. F., and S. M. MacDonald. 1986. Cambridge University P~ess, New York. Otters; ecology and conservation. 236pp. Metcalf, F. compo 1987. The Penguin dictionary of modern humorous quotations. Penguin Books, New York. 3l9pp. Miles, H. 1984. 160pp. The track of the wild otter. St. Martin's Press, New York. National Research Council, National Academy of Sciences. 1984. Developing strategies for rangeland management. Westview Press, Boulder, CO. 2022pp. Weber, W. A. 1987. Colorado flora: Western slope. University Press, Boulder, CO. 530pp. Colorado Associated �4 Wemmer, C. M., ed. 1987. Biology and management of the Cervidae (Research Symposia of the National Zoological Park, Front Royal, VA, 1982). Smithsonian Institution Press, Washington, DC. 577pp. Workman, G. W. 1985. Western elk management; a symposium. University, Logan. 213pp. Utah State Wrigley, R. E., and R. W. Nero. 1982. Manitoba's big cat: the story of the cougar in Manitoba. Manitoba Museum of Man and Nature, Winnipeg, Canada. 68pp. Publications obtained free or at low cost In addition to books purchased with Federal Aid Funds, about 60 free reports and short publications from state or federal agencies or from private sources were located, ordered, and obtained for use by Mammals Research personnel. Theses purchased, obtained on Interlibrary Loan or as gifts for use by Reserachers Banks, T. 1985. Nonconsumptive wildlife-associated recreation in Wyoming. M.S. Thesis, University of Wyoming, Laramie. 161pp. Bateman, M. C. 1972. Winter shelter: some effects on the behavior and physiology of penned white-tailed deer. M.S. Thesis, University of Maine, Orono. 108pp. Boyce, M. S. 1974. Beaver population ecology in interior Alaska. Thesis, University of Alaska, Fairbanks. 161pp. Bradley, P. V. 1986. Ecology of river otters in Nevada. University of Nevada, Reno. 112pp. M.S. M.S. Thesis, Busher, P. E. 1975. Movements and activities of beavers, Castor canadensis, on Sagehen Creek, California. M.A. Thesis, San Francisco State University, San Francisco. 86pp. 1980. The population dynamics and behavior of beavers in the Sierra Nevada. Ph.D. Thesis, University of Nevada, Reno. 147pp. Christensen, K. M. 1984. Habitat selection, food habits, movements, and activity patterns of reintroduced river otters in central Arizona. M.S. Thesis, Northern Arizona University. 72pp. Cooley, L. S. 1983. Winter food habits and factors influencing the winter diet of river otter in north Florida. M.S. Thesis, University of Florida, Gainesville. Davison, R. P. 1975. The efficiency of food utilization and energy requirements of captive female fishers. M.S. Thesis, University of New Hampshire, Durham. 1980. The effect of exploitation on some parameters of coyote population. Ph.D~ Thesis, Utah State University, Logan. 153pp. �5 Gerlach, T. P. 1987. Ecology of mule deer on the Pinon Canyon Maneuver Site, Colorado. M.S. Thesis, Virginia Polytechnic Institute and State University, Blacksburg. 6lpp. Graf, W. 1943. Natural history of the Roosevelt elk. State University, Corvallis. 222pp. Ph.D. Thesis, Oregon Horner, M. A. 1986. Internal structure of home ranges of black bears and analyses of home range overlap. M. S. Thesis, North Carolina State University, Raleigh. 54pp. Kopf, V. E. 1983. The relationship between food intake and thyroid hormone concentrations in white-tailed deer. Ph.D. Thesis, Virginia Polytechnic Institute and State University, Blacksburg. l44pp. Lauhachinda, V. 1978. Life history of the river otter in Alabama with emphasis on food habits. Thesis, Auburn University, Auburn, AL. l85pp. McAda, C. W. 1977. Aspects of the life history of three catostomids native to the upper Colorado River Basin. M.S. Thesis, Utah State University, Logan. l04pp. Marcum, C. L. 1975. Summer-fall habitat selection and use by a western Montana elk herd. Ph.D. Thesis, University of Montana, Missoula. l88pp. Pedevillano, C. 1986. Mountain goat behavior at the Walton Lick and Highway 2 underpasses in Glacier National Park. M.S. Thesis, University of Idaho, Moscow. 110pp. Reid, D. G. 1984. Ecological interactions of river otters and beavers in a boreal ecosystem. M.S. Thesis, University of Calgary, Alberta. 2l0pp. Renecker, L. A. 1987. Bioenergetics and behavior of moose (Alces alces) in the aspen-dominated boreal forest. Ph.D. Thesis, University of Alberta, Edmonton, Canada. 265pp. Sandell, M. 1985. Ecology and behaviour of the stoat Mustela erminea and a theory on delayed implantation. Ph.D. Thesis, University of Lund, Lund, Sweden. l15pp. Sayre, R. 1987. Effect of shrub stems on microhistological estimates of ruminant diets. M.S. Thesis, Colorado State University, Fort Collins. 52pp. Stenson, G. B. 1985. The reproductive cycle of the river otter, Lutra canadensis, in the marine environment of southwestern British Columbia. Ph.D. Thesis, University of British Columbia, Vancouver. Whitman, J. S. 1981. Ecology of the mink (Mustela vison) in west central Idaho. M.S. Thesis, University of Idaho, Moscow. lOlpp. Zackheim, H. S. 1982. Ecology and population status of the river otter in south-western Montana. M.S. Thesis, University of Montana, Missoula. l03pp. �6 Reference document location and delivery The Research Center Library staff also located and delivered about 890 individual articles on request for Mammals Researchers during this segment; about 25 were not available locally and were obtained through Interlibrary Loan procedures. Manuscripts published Job Progress Reports; Federal Aid. Carpenter, L. H. 5(3):25-28. 1988. All Studies. The big herds: Colorado's White River elk. Hobbs, N. T. 1987. Elk survival during winter: conservation. Bugle 5(1):60-64. 1988. Springtime in the Rockies: Bugle 5(2):25-26. Bugle a question of energy a critical period for elk. 1988. The legacy of the Elkhorn Ranch: country. Bugle 5(3);31-32, 34. research on elk in cattle ____ ~, and D. M. Swift. 1988. Grazing in herds: when are nutritional benefits realized? Amer. Nat. 131(5):760-764. Miller, H. W., and N. T. Uobbs. 1988. Ivermectin for treating lungworms in mountain sheep: an appropriate technology with a specific utility. Wildl. Soc. Bull. 16:236-238. Robbins, C. T., T. A. Hanley, A. E. Hagerman, O. Hjeljord, D. L. Baker, C. C. Schwartz, and W. W. Mautz. 1987. Role of tannins in defending plants against ruminants: reduction in protein availability. Ecology 68(1):98-107 • Prepared by .Mc-n'CJ\_ tJ . ~ ~,n;lco~~ Marian W. Hershcopf Librarian Len H. Carpenter Wildlife Research Leader �7 Wildlife Research Report July 1988 JOB PROGRESS REPORT State of Colorado Project No. W-153-R-2 -----------------------No. 1 --~------------------ Mammals Research Work Plan Multispecies Investigations Job No. 9 Period Covered: July 1, 1987 - June 30, 1988 Author: Mammals 1 Research Administration Len H. Carpenter ABSTRACT Research superv~s~on for 5, full-time employees involved in various aspects of deer and elk research was provided during the 1987-88 segment. Considerable effort was spent in an attempt to bring research planning, budgeting, documentation, and evaluation in line with proposed Division-wide planning systems. Assignment of planning and evaluation responsibilities for state-wide deer and elk programs significantly impacted time available for research administration. ��9 MAMMALS I RESEARCH ADMINISTRATION Len H. Carpenter P. N. OBJECTIVE To supervise and administer research on deer and elk in the Mammals Project. SEGNENT OBJECTIVE To supervise and administer research on deer and elk in the Mammals Project. METHODS The position of Research Leader is established to supervise and administer all research conducted for the wildlife species of concern. In January, 1988, this position was broadened to also include responsibilities of planning and evaluation for statewide species programs. RESULTS AND DISCUSSION Considerable time was spent during the latter part of the segment evaluating the last 2 big game seasons in preparation for recommendations to be made concerning a new 3-year season structure framework for 1989-91. Effort was expended throughout the segment working to bring research planning, budgeting, documenting, and evaluating into compatible formats with overall Division planning systems. By the end of the segment, all Mammals 1 research personnel except one were finally equipped with personal computers. Prepared by /1. (c../~~~ ~I 1! Len H. Carpen r Wildlife Research Leader ��Wildlife Research Report July 1988 11 JOB PROGRESS REPORT State of Colorado Project No. W-153-R-2 ------------------------No. 2 ----------------------- Mammals Research Work Plan Deer Investigations Job No. 7 Period Covered: July 1, 1987 - June 30, 1988 Author: Development of Census Methods for Deer in Plains Riverbottom Habitats R. C. Kufeld Personnel: D. Bowden, D. North. ABSTRACT Seventy-four mule and white-tailed deer were trapped in Clover traps and a drop net in the South Platte riverbottom between Gilcrest and Sedgwick during January 6-February 11, 1988. Twenty-four adults (11 mule deer and 13 whitetails) were radio-collared and eartagged. Remaining deer, mostly fawns, were eartagged. These instrumented deer plus 25 others, radio-collared in January and February, 1987, were located by aerial telemetry at 2-week intervals to determine movements and home range size. A summary of first-year movements of each deer instrumented in January and February, 1987, through March 4, 1988, and reports received about locations of eartagged (only) deer is presented. Movements of up to 145 and 82 airline km, respectively, were recorded for whitetails and mule deer. More locations are needed for calculation of home range size. ��13 DEVELOPMENT OF CENSUS METHODS FOR DEER IN PLAINS RIVERBOTTOM HABITATS Roland C. Kufeld P. N. OBJECTIVES 1. To determine seasonal movements and home range size of mule and whitetailed deer in plains riverbottom habitats. 2. To develop and test methods for estimating size of deer populations in plains riverbottom habitats. SEGMENT OBJECTIVE To determine seasonal movements and home range size of mule and white-tailed deer in plains riverbottom habitats. ACKNOWLEDGMENTS David North and numerous personnel from Colorado Division of Wildlife Northeast Region and Game Research Section were instrumental in capturing deer. STUDY AREA The study area is the South Platte River basin extending from Greeley, Colorado, to Nebraska. Riparian vegetation in the floodplain is dominated primarily by cottonwood (Populus sargentii), and willow (Salix spp.). The riverbottom is bordered along much of its length by agricultural lands, mainly cornfields. Other stretches are bordered by rangelands dominated by mixed prairie or sand sagebrush (Artemisia fililolia) (Costello 1954). Land status includes both private lands and state-owned wildlife areas. Results from the South Platte should apply to most other plains riverbottom habitats such as those found along the Arkansas River. METHODS AND MATERIALS Mule and white-tailed deer were trapped and marked in the South Platte riverbottom between Gilcrest and Sedgwick, Colorado, during the period January 6February 11, 1988. Most adult does and each adult buck received a radiocollar and 2 orange, numbered eartags. Deer were captured in Clover traps (Clover 1956) and a drop net. Each trap was baited with alfalfa hay and eared corn. The drop net was baited with hay, corn, and apple pulp. Adult white-tailed and mule deer, radio-collared during January and February, 1987 (Kufeld 1987), were located at approximately 2-week intervals throughout the segment. Those deer instrumented in Jqnuary and February, 1988, were located at about 2-week intervals beginning February 2, 1988. �14 RESULTS AND DISCUSSION Seventy-four deer were captured and marked in 1988, of which 24 adults (11 mule deer and 13 whitetails) were radio-collared (Tables 1 and 2). The following is a summary of first-year movements of 12 adult radio-collared whitetails, 13 adult radio-collared mule deer, and reports received about 43 other deer with only eargags from date of capture in January-February, 1987, along the South Platte River in eastern Colorado, through March 4, 1988. As of March 4, 1988, reports have been received on 11 of the 43 eartagged (only) deer. Five were reported shot by hunters, 2 were found dead, and 4 were observed alive. Of 6 adult radio-collared bucks, 1 mule deer and 2 whitetails were legally killed by hunters, and 1 mule deer was poached. The other 2 (1 mule deer and 1 whitetail) were still alive on March 4, 1988. Ten adult mule deer does and 9 adult white-tail does were radio-collared. One adult doe of each species was killed by an automobile. None were shot by hunters or died of other causes. During the first year, some individual bucks and does of both species made relatively long movements from their trapsites, while others stayed fairly close to their point of capture. Movements of up to 145 and 82 airline km, respectively, were recorded for whitetails and mule deer. RADIO-COll.ARED DEER White-tail Bucks (all tagged as adults) Eartag No. 28. Frequency 149.660. Caught 1-22-87 1.6 km west of Hardin Bridge. This deer stayed in or near the riverbottom and ranged from 9.7 river km ups t ream to 6.4 river km downstream of its trapsite through 3-4-88. Eartag No. 39. Frequency 149.770. Caught 2-3-87 1.6 km east of Merino Bridge. This deer stayed in or near the riverbottom and ranged from 4.8 river km upstream to 17.7 river km downstream of its trapsite. It was legally shot 12-3-87 6.1 km downstream of its trapsite. When shot, its antlers had 4 points on each side. Eartag No. 42. Frequency 149.780. Caught 2-3-87 5.6 km east of Hardin Bridge. Between 2-3-87 and mid-June, 1987, the deer stayed in the riverbottom in an area from 5.6 river km upstream to 2.4 river km downstream of its trapsite. In mid-June, 1987, it moved to an area centered about 9.7 airline km NW of Jackson Reservoir. It stayed within 1.6 km of that spot until mid-September. That area was treeless country with grassland, wheatfields, and some sorghum fields. The approximate center of this area was 27.4 airline km from its trapsite and 16.9 km from the Platte River. It stayed within about 1.6 km of the approximate center of the area until mid-September. Then it returned to the trapsite vicinity and occupied the same segment of riverbottom it used from 2-3-87 to mid-June of 1987. It was legally shot by a hunter on 12-13-87 within 0.6 km of the river and 2.4 km west of its trapsite. When shot, its antlers had 5 x 6 points. White-tail Does (all tagged as adults) Eartag No.8. Frequency 149.511. Caught 1-19-87 1.6 km west of Weldona Bridge. In late February, 1987, the deer moved 23.3 river km upstream to a site about 4.0 river km upstream from Orchard Bridge. It occupied a very �15 small area in the riverbottom there (about 1.9 km) until about 9-1-87. On 9-12-87 it was located at Goodrich, on 9-25 at Sterling, and on 10-13 it was in the riverbottom at the east end of the Bravo State Wildlife Area east of Sterling. This represented a movement of 121 river km downstream between 9-1 and 10-13. Between late October and 3-4-88, it stayed in the riverbottom between Dune Ridge Wildlife Area and Sterling. Eartag No. 16. Frequency 149.538. Caught 1-20-87 6.4 km east of Atwood Bridge. The deer stayed near the riverbottom and ranged from 2.4 river km upstream to 2.4 river km downstream of its trapsite until 7-10-87, when it was hit by a car and killed. Eartag No. 35. Frequency 149.731. Caught 2-3-87 4.0 km east of Merino Bridge. Between 2-3-87 and late August, 1987, the deer stayed in the riverbottom and ranged from the trapsite to 5.6 river km upstream. In late August, 1987, it became entangled in a barbed wire fence and slipped off its collar. Thus, it could no longer be radio tracked, but it is assumed its eartags remain intact. Eartag No. 36. Frequency 149.741. Caught 2-3-87 1.6 km east of Merino Bridge. This deer stayed in the riverbottom from 2-3-87 through 3-4-88. Except for 1 location on 4-18-87, which was 10.3 river km upstream from its trapsite, it remained in an area between 3.2 river km upstream and 1.6 river kID downstream of its trapsite. Eartag No. 43. Frequency 149.791. Caught 2-3-87 5.6 miles east of Hardin Bridge. From 2-3-87 until 4-17-87 the deer stayed in the riverbottom from 6.4 river km'upstream to 1.6 river km downstream of its trapsite. Sometime between 4-17 and 5-4 it left that area and spent the summer at an unknown location, which was apparently far from its trapsite. In addition to scanning its frequency on all regular flights to locate radio-collared deer, we made a special flight just to locate this deer. The flight route was down the South Platte River from Greeley to Julesburg, 48 km north of the river back to Greeley, 48 km south of the river to south of Brush, up Bijou Creek to Colorado Springs, then to Denver, Greeley, and Fort Collins. No signal was received during the flight. Sometime between August 27 and September 12 the deer returned to the trapsite vicinity. Between 9-12-87 and 3-4-88 it stayed in the riverbottom from 2.4 river km upstream to 4.0 river km downstream of its trapsite, except on 10-27-87 when it was located 22.5 river km downstream from its trapsite near the town of Orchard. Eartag No. 44. Frequency 149.800. Caught 2-4-87 1.6 km east of Merino Bridge. Between 2-4-87 and 5-4-87 the deer stayed in the riverbottom in an area from 0.8 river km upstream to 2.4 river km downstream of its trapsite. On 5-15 it was located at the southwest end of Prewitt Reservoir (11.3 airline km from its trapsite), and on 6-4 it was located about 6.4 km south and 1.6 km east of Woodrow, which is about 60 km from its trapsite and about 39 airline km south of the Platte River. The site was far out on the plains in an area of wheatfields and pastureland. A small creek (Beaver Creek) lined with willows, making a riparian zone only about 3 to 6 m wide, traversed the area. The deer stayed within 2.4 km of this site through 3-4-88 �16 Eartag No. 46. Frequency 149.809. Caught 2-4-87 5.6 km east of Hardin Bridge. Between 2-4-87 and 5-15-87 the deer stayed in the riverbottom from 2.4 river km upstream to 1.6 river km downstream of its trapsite. On 6-4-87 it was located near Banner Lakes Wildlife Area, 33.8 airline km from its trapsite. This is also 33.8 km from the Platte River. It stayed in that area until early August, when it returned to the vicinity of its trapsite. Between 8-18-87 and 3-4-88 it stayed in the Platte riverbottom from the trapsite to 3.2 river km downstream, except that on October 27 it was located 12.9 river km downstream from its trapsite. Eartag No. 62. Frequency 149.921. Caught 2-13-87 3.2 km east of Fort Morgan Bridge. This deer stayed in or within 0.4 km of the riverbottom and from 9.7 river km upstream to 2.4 river km downstream from its trapsite from 2-13-87 through 3-4-B8. Eartag No. 64. Frequency 149.951. Caught 2-17-87 4.0 km east of Masters Bridge. This deer stayed in the riverbottom from 0.8 river km upstream to 0.8 river km downstream of its trapsite between 2-17-87 and 5-15-87. On 5-30 it was located at the south end of Empire Reservoir, 7.2 airline km from its trapsite and 7.2 airline km from the river. It stayed in that area until the first week in August, when it returned to the vicinity of its trapsite. Between 8-18-87 and 3-4-88 it stayed in the riverbottom in an area 0.8 river km above to 0.8 river km below its trapsite. Mule Deer Bucks (all tagged as adults) Eartag No. 25. Frequency 149.641. Caught 1-21-87 at Hardin Bridge. Between 1-21-87 and 3-4-88 this deer remained in or near the riverbottom and ranged from 6.4 river km upstream to 0.8 river km downstream of its trapsite. During the fall of 1987, its antlers had 4 points on each side. Eartag No. 26. Frequency 149.651. Caught 1-21-87 at Hardin Bridge. Between 1-21-87 and 5-30-87 it ranged within 1.6 km of the riverbottom and from its trapsite to 4.0 river km downstream of its trapsite. On 5-30-87 it was located about 20.1 airline km south of Kersey and about 20.9 airline km from the Platte River. It stayed in that general area until last located on 10-28-87 12.9 km south and 3.7 km east of Kersey. It was killed illegally between 10-28, after 2:28 PM, and 11-9; and the collar was thrown in a pond 8.4 km south and 0.6 km east of La Salle. When last seen, its antlers had 4 points on each side. Eartag No. 33. Frequency 149.701. Caught 2-2-87 1.6 km east of Weldona Bridge. Between 2-2-87 and 10-24-87 this deer stayed within about 1.6 airline km of the riverbottom and ranged from 0.8 river km upstream to 2.0 river km downstream of its trapsite, except for 2 trips of short duration outside that area. On 3-6-87 it was located in the riverbottom 8.9 river km downstream from its trapsite, and on 5-15-87 it was located in Bijou Creek 7.2 airline km from its trapsite and 7.2 airline km from the South Platte River. It was shot by a hunter on 10-24-87. It was a very large buck, and when killed, its antlers had 7 x 6 points. Mule Deer Does (all tagged as adults) Eartag No. 12. Frequency 149.521. Caught 1-9-87 at Hardin Bridge. Between 1-19-87 and 3-4-88 this deer stayed in the riverbottom and ranged from its trapsite to 5.6 river km upstream from its trapsite. �17 Eartag No. 13. Frequency 149.531. Caught 1-19-87 at Hardin Bridge. Between 1-19-87 and 3-4-88 this deer stayed in the riverbottom and ranged from 4.8 river km upstream to 4.8 river km downstream from its trapsite. Eartag No. 22. Frequency 149.551. Caught 1-20-87 at Hardin Bridge. Between 1-20-87 and 3-4-88 this deer stayed within 3.2 km of the riverbottom and ranged from 1.6 river km upstream to 3.2 river km downstream of its trapsite. Eartag No. 23. Frequency 149.631. Caught 1-20-87 at Hardin Bridge. Between 1-20-87 and 5-20-87 this deer stayed in the riverbottom and ranged from 3.6 river km upstream to 4.5 river km downstream of its trapsite. On 5-30-87 it was located at Lord Reservoir, 3.2 km south and 5.6 km east of Keenesburg. This was 30.9 airline km from its trapsite and 27.4 airline km from the Platte River. It stayed within 1.6 km of Lord Reservoir during the summer and was still there on 10-13-87. On 10-27-87 it was located back in the Platte River near its trapsite. Between 10-27-87 and 3-4-88 it stayed in the riverbottom and ranged from 4.5 river km to 1.3 river km upstream of its trapsite. Eartag No. 29. Frequency 149.671. Caught 1-22-87 at Hardin Bridge. Between 1-22-87 and 3-4-88 this deer stayed in the riverbottom and ranged from 4.2 river km upstream to 1.6 river km downstream of its trapsite. Eartag No. 48. Frequency 149.859. Caught 2-9-87 3.2 km east of Orchard Bridge. On 2-9-87 and 3-6-87 it was located in the Platte riverbottom 2.9 river km upstream from its trapsite. On 3-18-87 it was located 3.2 airline km west of Jackson Reservoir and about 5.6 airline km from the Platte River. It stayed away from the river and lived on the western and southern sides of Jackson Reservoir until 3-4-88 except for 1 location on 12-7-87 when it was found at its trapsite in the riverbottom. Eartag No. 50. Frequency 149.839. Caught 2-12-87 1.6 km west of Hardin Bridge. Between 2-12-87 and 3-4-88 this deer stayed within 1.6 km of the river and ranged from 1.6 to 7.7 river km upstream of its trapsite. Eartag No. 57. Frequency 149.869. Caught 3.2 km east of Orchard Bridge. This deer was trapped in the riverbottom but soon left. On 2-23-87 it was located at the northwest corner of Jackson Reservoir. This was 7.7 airline km from its trapsite and 7.7 airline km from the river itself. From the time it left shortly after being caught through 3-4-88, it never returned to the river. It stayed in an area extending about 6.4 km to the west and north of the northwest corner of Jackson Reservoir. Eartag No. 61. Frequency 149.889. Caught 2-12-87, 1.6 km west of Hardin Bridge. Between 2-12-87 and 3-4-88 this deer stayed within 1.6 km of the river and ranged from its trapsite to 6.1 river km upstream. Eartag No. 67. Frequency 149.970. Caught 2-17-87 at Hardin Bridge. Between 2-12-87 and 6-30-87 the deer stayed in the riverbottom and ranged from 6.0 to 0.8 river km upstream from its trapsite. On 7-16-87 it was located in a cornfield about 1.6 km southwest of its trapsite and about 1.6 km from the river. It stayed in cornfields in that area until it was killed by an automobile on US 34 4.8 airline km from its trapsite and 3.2 airline km from the river about 8-1-87. �18 EARTAGGED (ONLY) DEER White-tail Bucks (all tagged as fawns) Eartag No.3. Caught 1-19-87 1.6 km east of Cooper Bridge (Hillrose). This deer was shot by a hunter on 10-1-87 4.8 km east of Meridan, Wyoming. This was about 145 airline km from its trapsite. Eartag No. 45. Caught 2-4-87 5.6 km east of Hardin Bridge. This deer was shot by a hunter on 12-5-87 in the Platte riverbottom about 3.9 river km downstream from Masters Bridge and 16.1 river km downstream from its trapsite. Eartag No. 63. Caught 2-13-87 3.2 km east of Fort Morgan Bridge. This deer was found dead in a roadside ditch on 10-16-87. It had been dead about a month. The spot was 6.4 km miles southeast of Gill, which is 82 river km upstream from its trapsite and 2.8 airline km north of the Platte River. White-tail Does (all tagged as fawns) Eartag No. 15. Caught 1-20-87 6.4 km east of Atwood Bridge. It was observed on 3-18-87 1.6 km east of Merino Bridge. This is 16.1 river km upstream from its trapsite. Eartag No. 31. Caught 2-2-87 1.6 km east of Merino Bridge. This deer was shot by a hunter on 12-1-87 near its trapsite 1.6 km east of Merino Bridge. Eartag No. 41. Caught 2-3-82 3.2 km east of Fort Morgan Bridge. observed 12-15-87 1.6 km east of its trapsite. It was Eartag No. 60. Caught 2-11-87 at Hardin Bridge. It was found dead in a field 0.8 km east of Jackson Reservoir on 12-4-87. It had been dead for several weeks. The site where found was about 33.8 airline km east of the trapsite and 4.8 airline km north of the South Platte River. Mule Deer Bucks (all tagged as fawns) Eartag No. 51. Caught 2-10-87 1.6 km east of Weldona Bridge. On 7-8-87 this deer was observed 4.0 km east of Nunn. This is approximately 82 airline km northwest of its trapsite and about 37 airline km north of the South Platte River. Eartag No. 65. Caught 2-13-87 3.2 km east of Orchard Bridge. This deer was observed on 4-25-87 1.6 km southwest of Jackson Reservoir and about 4.8 airline km northwest of its trapsite. On 10-18-87 it was killed by a hunter about 4.8 km west of Grover, which was about 59.5 airline km north of its trapsite and 59.5 airline km north of the South Platte River. Mule Deer Does (all tagged as fawns) Eartag No. 53. Caught 2-10-87 at Hardin Bridge. On 12-6-87 this deer was observed 6.4 km north of Black Hollow Reservoir. This is approximately 49.9 airline km northwest of its trapsite and about 37 airline km northwest of the South Platte River. �19 Eartag No. 55. Caught 2-10-87 at Hardin Bridge. On 12-3-87 this deer was shot by a hunter 4.2 km east and 5.6 km north of Galeton. This is approximately 26.6 airline km north-northwest of its trapsite and about 17.7 airline km north of the South Platte River. LITERATURE CITED Clover, M. R. 1956. Single-gate deer trap. Calif. Fish and Game 42:199-201. Kufeld, R. C. 1987. Development of census methods for deer in plains riverbottom habitats. Colo. Div. Wildl., Wildl. Res. Rep. July (1):13-19. Prepared by ~f-;-~_';--:'- ~~-:;(--::C:--:;--,-_;t&_·.,_t~_/ _;_._' __ Roland C. Kufeld Wildlife Researcher C �20 Table 1. Mule and white-tailed deer trapped and marked in the South Platte riverbottom between Gilcrest and Sterling, Colorado, during 1-6-88 through 2-11-88. -----------------------------------------------------------------------------.Species Sex Age Radio collared and eartagged Eartagged only Total ------------------------------------------------------------------------------ Mule deer Whi tetai 1 buck adul t fawn 4 0 0 9 4 9 doe adul t fawn 7 4 9 11 0 adult fawn 5 0 0 5 13 13 adul t fawn 8 0 24 5 10 13 10 74 buck doe TOTAL 50 9 �21 Table 2. Deer tagged along South Platte River between Platteville through 2-11-88. and Julesburg, Colorado, 1-6-88 -------------------------------------------------------------------------------------------------------Eartag Age when number Species Sex captured Date captured Capture 1ocati on Radioco11ar attached 1 mi. E. of Merino Bridge 1 mi. E. of Merino Bridge 1 mi. E. of Merino Bridge 1 mi. E. of Merino Bridge 2.S mi. E. of Merino Bridge 2.5 mi. E. of Merino Bridge 1 mi. E. of Merino Bridge 1 mi. W. of Proctor Bridge 1 mi. W. of Proctor Bridge 1 mi. W. of Proctor Bridge 1 mi. W. of Red Lion Bridge 2 mi. E. of Red Lion Bridge 2 m1. E. of Red Lion Bridge 2 mi. E. of Atwood Bridge 2.S mi. E. of Merino Bridge 2.5 mi. E. of Merino Bridge 2.5 mi. E. of Merino Bridge 1 mi. E. of Merino Bridge 1 mi. E. of Merino Bridge 1 mi. W. of Red Lion Bridge 1 mi. E. of Merino Bridge 2 mi. E. of Red Lion Bridge 2 mi. E. of Red Lion Bridge 2 mi. E. of Red Lion Bridge 1 mi. W. of Proctor Bridge 2 m1. E. of Red Lion Bridge 2 mi. E. of Red Lion Bridge 1 mi. E. of Merino Bridge 1 mi. E. of Merino Bridge 1 mi. Eo of Heri no Bri dge 2 mi. E. of Atwood Bridge 1 mi. E. of Merino Bridge 1 mi. E. of Merino Bridge 1 mi. E. of Merino Bridge 1 mi. E. of Merino Bridge 1 mi. W. of Proctor Bridge 2.5 mi. E. of Merino Bridge 1 mi. W. of Proctor Bridge 1 mi. E. of Merino Bridge 1 mi. w. of Proctor Bridge 1 mi. w. of Proctor Bridge 1 mi. W. of Proctor Bridge 1 mi. E. of Merino Bridge 1 mi. E. of Merino Bridge 1 mi. E. of Merino Bridge 1 mi. E. of Merino Bridge 1 mi. E. of Merino Bridge 1 mi. E. of Merino Bridge Hardin Bridge 1 mi. W. of Hardin Bridge 1 mi. W. of Hardin Bridge Hardin Bridge 2.5 mi. E. of Masters Bridge 3.5 mi. W. of Evans Bridge Hardin Bridge 1. 5 mi. E. of Greeley Airport 2.5 mi. E. of Masters 8ridge 2.5 mi. E. of r~asters Bri dge 2.5 mi. E. of Masters Bridge Masters Bridge 2.5 mi. E. of Mas ters Bridge Masters Bridge 2.5 mi. E. of Masters Bridge Masters Bridge Masters Bridge 2.5 mi. E. of Masters Bridge Hardin Bridge 3.5 mi. W. of Evans Bridge 2.5 mi. E. of Masters Bridge 2.5 mi. E. of Masters Bridge Masters Bridge 2.5 mi. E. of Masters Bridge 2.5 mi. E. of Masters Bridge 3.5 mi. W. of Evans Bridge 149.361 149.262 -------------------------------------------------------------------------------------------------------70 whitetail buck adult 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 90 91 92 93 94 95 96 97 98 99 100 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 151 152 153 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 whitetail whi teta il whitetail whitetail whitetail whitetail mule deer mule deer mule deer whi teta il mule deer mule deer whi te te i1 whitetail whitetail whitetail whitetail whitetail whi teta il whitetail mule deer mule deer mule deer mule deer mule deer whi tetai 1 whi tetail whitetail whi tetail whitetai 1 whi tetail whteitail whi tetail whitetail mule deer whi tetail mule deer whitetail mule deer mule deer mule deer whi tetai 1 whi teta il whitetail whi tetail whitetail whi tetail mule deer mule deer mule deer mule deer whi tetai1 mule deer mule deer mule deer mule deer whitetai 1 mule deer whitetail whitetai 1 whi teta il mule deer whitetail mule deer mule deer mule deer mule deer whi tetai 1 whitetail mule deer mule deer mule deer mule deer doe doe buck doe doe buck buck doe doe doe doe doe buck doe buck doe doe doe buck doe doe doe doe buck doe doe buck buck doe doe doe doe buck doe buck doe buck buck doe doe doe buck doe buck doe buck doe buck buck doe buck buck buck buck buck doe doe doe doe buck buck buck buck doe doe doe buck doe buck doe doe doe buck adult fawn fawn adult fawn adult adult adult fawn adult adult adult fawn adult fawn fawn adult adult fawn adult fawn adult adult fawn fawn fawn fawn fawn fawn adult fawn fawn fawn fawn fawn adult fawn fawn fawn fawn adult fawn adult adu1 t fawn adult fawn fawn fawn adult fawn fawn adult fawn adult adult adult adult adult fawn adult fawn fawn fawn adult adult fawn adu1 t fawn fawn fawn adult adult 1-18-88 1-18-88 1-18-88 1-18-88 1-18-88 1-18-88 2- 3-88 1-18-88 1-18-88 1-18-88 1-18-88 1-18-88 1-18-88 1-20-88 1-20-88 1-20-88 1-20-88 2- 5-88 1-20-88 1-20-88 1-21-88 1-21-88 1-21-88 1-21-88 1-22-88 1-22-88 1-22-88 1-20-88 1-26-88 1-27-88 1-28-88 1-18-88 1-29-88 1-29-88 1-29-88 1-29-88 1-18-88 2- 3-88 2- 5-88 2- 5-88 2- 5-88 2- 5-88 2-11-88 1-27-88 1-18-88 2-11-88 2-11-88 2-11-88 1-11-88 1-11-88 1-12-88 1- 9-88 1- 9-88 1-11-88 1-11-88 1-11-88 1- 7-88 1- 7-88 1- 7-88 1- 7-88 1- 7-88 1- 6-88 1- 6-88 1- 7-88 1- 6-88 1- 6-88 1- 8-88 1- 8-88 1- 8-88 1- 8-88 1- 8-88 1- 8-88 1- 8-88 1-13-88 149.252 149.370 149.382 149.291 149.282 149.538 149.651 149.971 149.271 149.351 149.390 149.731 149.431 149.452 149.332 149.301 149.341 149.311 149.411 149.321 149.780 149.461 -------------------------------------------------------------------------------------------------------- ��Colorado Division of Wildlife Wildlife Research Report July 1988 23 JOB FINAL REPORT State of Colorado Project No. W-153-R-2 -----------------------No. 2 ---------------------- Work Plan Job No. 10 Period Covered: Author: Mammals Research Deer Investigations Compensatory Mortality in Mule Deer Populations July 1, 1987 - June 30, 1988 R. M. Bartmann Personnel: T. A. Abbott, L. H. Carpenter, B. T. Helmich, J. S. Osborn, L. Renner, T. D. Schnurr, W. F. Tysinger, D. L. Weybright, E. Weber, and G. C. White ABSTRACT Three pastures at the Little Hills Wildlife Area were stocked with mule deer to simulate removal rates of 0, 33, and 66%. Forty-five to 47 radio-collared fawns were placed in each pasture along with enough adults to achieve target densities. No fawns in the high density pasture survived. Eleven and 15% of the fawns in the low and medium density pastures, respectively, survived. Snow cover that persisted into late March may be partly responsible for the similarity in survival rates in these pastures. ��25 COMPENSATORY MORTALITY IN MULE DEER POPULATIONS Richard M. Bartmann P. N. OBJECTIVES To estimate the degree of compensatory winter mortality forces operating within the mule deer fawn age class in Piceance Basin. SEGMENT OBJECTIVES 1. Stock 50 radio-collared fawns plus enough uncollared adults in each of three pastures at the Little Hills Wildlife Area to achieve three different deer densities. 2. Monitor radio-collared fawns for mortality 5-7 days per week throughout winter and identify cause of death for each mortality. The study is divided into two parts: a field experiment and a pasture experiment. The field experiment is the primary responsibility of cooperators at Colorado State University, and the pasture experiment is the primary responsibility of the Division of Wildlife. Therefore, only the pasture experiment is reported here. METHODS AND MATERIALS Three pastures at the Little Hills Wildlife Area were stocked in late November and early December with deer trapped in the immediate area. Target densities for pasture stocking were 44, 89, and 133 deer/km2 which simulated removal rates of 67, 33, 0%, respectively. Fifty fawns were put fawn was weighed and adults were added to were neither weighed in each pasture for estimation of survival rates. Each radio collared before being placed in a pasture. Enough each pasture to attain the target densities, but they nor marked in any way. Fawn radio collars contained a motion-sensor set with a 4-hour delay to enable detecting mortalities. Radio signals were monitored 5-7 days per week during winter and all mortalities checked for cause of death as soon as possible. In the spring when deer started to migrate, gates in all pastures were opened and the deer allowed to leave voluntarily. Monitoring for suvival continued at 2-week intervals until June 15, 1988. RESULTS Pasture stocking occurred from November 9 through December 5, 1988. Several early deaths in each pasture reduced the number of fawns available for survival rate estimates (Table 1). ��27 Table 1. Stocking summary for 3 pastures at the Little Hills Wildlife Area during winter 1987-88. . ------------------------------------------------------------------------------Stocking densi tya Low t~ediurn High Number of deer Fawns Adults M F M F 23 20 24 27 23 22 2 8 5 27 36 36 Fawn:adu1t ratio 162:100 109:100 112:100 Pasture size Hectares Acres 170 101 66 aStocking densities simulate removal rates as follows: Medium = 33%, and High = 0%. Deer/area km2 mi 2 419 249 45 91 116 237 162 132 344 Low = 67%, Table 2. Fate of mule deer fawns stocked at 3 removal rates in patures at the Little Hills Wildlife Area during winter 1987-88. Percentages are in parentheses. Removal rate (%) 67 33 o Starvati on 47 47 45 40 (85) 39 (83) 43 (96) Mortal ity Unknown 2 (4) 1 (2) 2 (4) Total 42 40 45 (89) (85) (100) Survived 5 (11) 7 (15) (0) o ��Colorado Division of Wildlife Wildlife Research Report July 1988 29 JOB PROGRESS REPORT State of Colorado Project No. W-153-R-2 -----------------------No. 2 --~--------------- Work Plan Job No. 11 Period Covered: Author: Mammals Research Deer Investigations Testing of Mule Deer Census Methodology July 1, 1987 - June 30, 1988 R. M. Bartmann Personnel: L. H. Carpenter, D. L. Weybright, and G. C. White ABSTRACT Three line transect surveys were completed in the pastures during November and December, 1987. A paper entitled, "Estimating l1ule Deer Densities in PinyonJuniper Woodland Using Aerial Line Transects" was submitted to The Journal of Wildlife Management. ��31 TESTING OF MULE DEER CENSUS METHODOLOGY Richard M. Bartmann P. N. OBJECTIVES 1. Conduct three areial line transect surveys during the stocking process at the Little Hills Wildlife Area to estimate deer numbers at three density levels. 2. Compare true deer densities with estimated densities generated by five line transect estimators to determine which is most efficient. 3. Analyze data and publish results in The Journal of Wildlife Management. SEGMENT OBJECTIVES Same as P. N. Objectives. RESULTS Three line transect surveys were completed in the pastures during November and December, 1987 •. A paper entitled, "Estimating Mule Deer Densities in PinyonJuniper Woodland Using Aerial Line Transects" was submitted to The Journal of Wildlife Management. -: Prepared by / ~ "M--- /~/{~ ~hard M. Bartmann Wildlife Researcher C - ��33 Wildlife Research Report July 1988 JOB PROGRESS REPORT State of Colorado Project No. W-153-R-2 -----------------------No. 3 ---------------------- Work Plan Mammals Research Elk Investigations Job No. 2 Period Covered: July 1, 1987 - June 30, 1988 Author: Trapping, Transporting, and Maintenance of Elk at Livestock-Elk Grazing Study G. D. Bear Personnel: D. Baker, T. Hobbs, M. Miller, D. Haskins, T. Lines, C. Reichert, J. .Nad Lson , J. Papez, B. Dupire, C. Woodward, J. Haskins, M. Bauman, K. Navo ABSTRACT A total of 154 elk was trapped at locations near Meeker, Craig, and Maybell to stock the experimental pastures on the Little Snake River Wildlife Area north of Maybell, Colorado. Pastures were stocked with 54 adult females in December and January; the elk were removed from the pastures in mid-April. Snow accumulations persisted in the pastures throughout the winter making it difficult for the elk to forage. Eleven elk died of malnutrition (10 in "high-use pastures" and 1 in a "low-use pasture") and had to be replaced. Average weights of 51 elk calves at the Meeker, Craig, and 11es Grove trapsites were: 253, 255, and 238 lbs, respectively. ��35 TRAPPING, TRANSPORTING, AND MAINTENANCE OF ELK AT LIVESTOCK-ELK GRAZING STUDY G. D. Bear P. N. OBJECTIVE To provide assistance to the Livestock-Elk Grazing Study near Maybell, Colorado, by capturing and maintaining an experimental elk herd. SEGMENT OBJECTIVE 1. To trap, mark, stock, and maintain 54 adult female elk in the livestockelk grazing pasture complex during the period December 15 to April 20. 2. To determine seasonal movements of radio-collared elk after they are released from the pastures. METHODS AND RESULTS Capturing and Handling of Elk Trapsites were established at the following sites: (1) northeast of Craig (5 mi east of Ralph White Reservoir near the Don Cook Ranch; (2) southwest of Hamilton (1 mi south of lIes Grove in Axial Basin); (3) 4 mi southeast of Meeker near the Environmental Plant Center; (4) south of Lay at the junction of ~organ Gulch and Moffat Co. Road 17; (5) 15 mi north of Maybell at Godiva Rim; and (6) southwest of Hayden. Portable group-traps baited with alfalfa hay and salt were used to capture elk. Details for handling elk at the trapsites and in the corral at the livestock-elk grazing pasture complex are presented in the previous Federal Aid Report. A total of 144 elk were captured from December 9, 1987, through January 14, 1988, to meet the stocking quota (54 adult female elk) for the experimental pastures (Table 1). An additional 10 adult female elk were trapped in late January through mid-February and maintained at the pasture corrals for substituting in experimental pastures as needed (Table 3). Calf elk were weighed at the trapsites before they were released. A platform scale was used to obtain the weights. Average weight of calves at the Meeker, Craig, and lIes Grove sites were: 253, 255, and 238 lbs, respectively (Table 2). Average weight of the lIes Grove site was slightly lower than average weights for calves at the other sites, however, this difference was not significant (P = .05). Elk in the Pastures Stocking experimental pastures with elk at the Little Snake Study Area proceeded on schedule during December and January, 1987-88. Average stocking date was January 4, 1988. All animals released into pastures tested negative for serologic evidence of exposure to brucellosis and to 5 serovars of leptospirosis. �36 Weather conditions during the 1987-88 winter were very difficult for elk in the pastures. This was in marked contrast to 1986-87, which had very light snow accumulations. There was very little snow in early December, then a series of storms in late December and January resulted in 4-6 ft snow drifts in the low areas of the pastures and 6-12 in of snow on the ridgetops. Air temperatures ranged from -15 to -400F at night and only a maximum of l5-250F during the daylight hours in January through February. Due to these low temperatures, there was not a noticeable decrease in the snow accumulations during those months. The elk were forced to paw for food, even on the ridgetops. After February 20 thre was a week of sunny weather with daytime temperatures at 38-400F, which opened holes in the snow cover. By February 26 there was an estimated 30-50% bare ground (mostly on the ridge- tops). On March 10-11 a storm with extremely strong winds left 6-12 in of snow drifted across the pastures. Only 10% of the pastures (ridges) were free of snow cover. Another severe storm was in the area on March 16; again making it very difficult for the elk to secure food. The snow cover was 90% melted by March 22. In late March new growth on Poa was approximately 1 in tall; Stipa, Agropyron, and Festuca were showing green shoots at the base. Severe weather conditions in January and February, then again in mid-March, severely stressed the elk. Nine elk died of malnutrition from January 25 to February 27; then an additional elk died on April 1 and April 12. The elk that died April 1 was in a "low-use pasture" while the remaining 10 were in the "high-use pastures" (3 cows in #2, 5 cows in #5, 2 cows in #9, and 1 cow in #11). All were replaced with animals held in the corrals. Animals crossed the electric fences at an accelerated rate last winter. Ten elk moved through the fences and were returned to their pastures. It is speculated that snow insulated the animals, thus reducing the electrical shock they received when attempting to push between the wires. Deer readily crossed fences (exceeding 20 crossings) and made it difficult to keep them cleared from the pastures. Only 1 antelope crossed the fence into the pastures during the 1987-88 winter. Elk were removed from the experimental pastures during the week of April 15 to April 21. Thirty of them were transported to Little Hills Experiment Station, near Meeker, Colorado; while the remaining ones were released at the site. Elk at Little Hills are to be maintained in a 360-acre pasture until next winter, when they will again be taken back to the Little Snake River Experimental Pastures. Elk Movements Minimal effort was conducted on this aspect of the study. Over half of the elk were moved to the Little Hills pasture and many of the elk released at the study site were in poor condition. It was thought that the added stress of collaring the elk would be detrimental, so they were released uncollared. Two cows radio collared the previous year returned to the Godiva winter range in January and left during the first week in April. '- 0) Prepared by ~~~~ ~~LJ~~~~ Geo~Wildlife Researcher _ �37 Table 1. Summary of elk trapped during the period December 9,1987, January 14, 1988.* Cows Yearlings Ferna1e r~a1e Calves Female Nale through Bull s Total ------------------------------------------------------------------------------Meeker NE Craig 11es Grove Morgan Gulch Total 14 8 22 15 59 1 3 3 1 8 2 0 4 1 3 20 1 4 6 29 13 9 19 0 41 0 1 0 0 1 34 22 67 144 ------------------------------------------------------------------------------*E1k trapped January 28 and February 18 are not included in this summary because adult cows were sorted and all other elk released without sexing or aging. ELK TRAPPING - MOFFAT COUNTY December 9, 1987 - January 14, 1988 Location of Traps Portable-group traps were set at the following locations: (1) northeast of Craig (5 miles east of Ralph White Reservoir near Don Cook Ranch; (2) southwest of Hamilton (1 mile south of Iles Grove in Axial Basin); (3) south of Lay (Morgan Gulch and Moffat Co. Rd. 17), and (4) southeast of Meeker (near Envir. Plant Center). Additional sites were tried south of Hayden and north of Maybell without success. f~arki n9 of E1 k Elk were marked with metal, yellow colored eartags stamped with the DOW Grand Junction Office address and an identification number (86 - prefix). Trapping Crew George Bear Dale Haskins Hi ke Nil 1 er Dan Baker Tom Hobbs Division of Wildlife Tom Lines Chuck Reichert Jeff Madison Joe Papez Barry Dupi re Other Contributors Tim Dole Ramsy Buffham Chip ~lcIntire Don Cook Kris Maneotis Dana McIntire Chuck Wood\'iard Jim Haskins I~ike Bauman Kirk Navo �w co Table 2. Weights of elk calves trapped during the 1987-88 winter. Male Female Both sexes r·1a 1 e Female Both sexes Male Female Both sexes -----------------------------------------------------------------------------------------------------------t-iean Std. dev. Range Number 251 32 187-286 7 258 14 246-274 4 253 26 187-286 11 255 32 208-289 8 250 - - 1 255 30 208-289 9 245 46 153-302 16 229 31 159-268 15 238 40 153-302 31 �39 Table 3. Elk trapped in Moffat and Rio Blanco Counties during Oecember, 19B7, to February 1988. --------------------------------------------------------------------------------------------------Eartag no. Date Sex Age location Remarks --------------------------------------------------------------------------------------------------1 12/9 2 M F 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 12/9 12/9 12/9 12/9 12/9 12/9 12/9 12/9 12/12 12/12 12/12 12/12 12/12 12/12 12/12 12/12 12/12 12/12 12/12 12/12 12/12 12/12 12/12 12/12 12/12 12/12 12/12 12/12 12/12 12/12 12/12 12/12 12/12 12/31 12/31 12/31 12/31 12/31 12/31 12/31 12/31 12/31 12/31 12/31 12/31 12/31 12/31 12/31 12/31 12/31 12/31 12/31 12/31 12/31 1/7 57 58 59 60 61 62 63 64 65 1/7 1/7 1/7 1/7 1/7 1/7 1/7 1/7 1/7 M F F M M 66 67 68 69 70 71 72 73 74 75 1/7 1/7 1/7 1/7 1/7 1/7 1/7 1/7 1/8 1/9 F F F M F F F 76 77 78 1/9 1/9 1/9 F 3 4 5 6 7 8 9 10 11 M F F M M F M M M M F F M F M M F F F F F M F M F F F M F F M F F F M F F F M M M F F M F M F ~1 M F M F M F F M F F F F F M F calf adult calf adult adul t calf calf adult yearling yearling calf calf adult adul t calf calf calf calf adult adult adult adult adult calf adult calf calf calf adult calf calf yearling calf adult adult adult calf adult adult calf .calf calf cal f yearling yearling calf adult calf adult adult calf yearling calf calf adult Meeker Meeker Meeker Meeker Meeker ~leker Meeker Meeker Meeker Meeker Meeker Meeker Meeker Meeker Meeker Meeker Meeker Meeker Meeker Meeker Meeker Meeker Meeker Meeker Meeker Meeker Meeker Meeker Meeker Meeker Meeker Meeker Meeker Meeker NE Craig NE Crai g NE Craig NE Craig NE Craig NE Craig NE Crai g NE Craig IJE Craig IlE Cra ig NE Crai g NE Craig NE Craig NE Craig NE Craig NE Craig NE Craig NE Craig NE Craig NE Craig NE Craig Iles Grove calf yearling calf calf calf calf yearling adult calf Iles Iles lIes Ll es lIes lIes Iles IIes lIes adult calf adult calf calf calf calf adult adult adult lIes Grove Iles Grove Iles Grove lIes Grove Iles Grove Iles Grove Iles Grove lIes Grove NE Craig lIes Grove yearling calf adult Iles Grove Iles Grove Iles Grove adul t Grove Grove Grove Grove Grove Grove Grove Grove Grove trans. Marvine trans. elk pens trans. Marvine trans. elk pens trans. elk pens trans. Marvine trans. Marvine trans. elk pens released released trans. Marvine - 272 1bs. trans. Marvine - 282 1bs. trans. elk pens trans. elk pens trans. Marvine - 260 1bs. trans. Marvine - 266 lbs. trans. Marvine - 244 1bs. trans. Marvine - 247 1bs. trans. elk pens - RC trans. elk pens - RC trans. elk pens - RC trans. elk pens trans. elk pens trans. Marvine - 255 lbs. trans. e1 k pens trans. Marvine - 276 lbs. trans. Marvine - 247 lbs. trans. Marvine - 246 lbs. trans. elk pens trans. Marvine - 286 lbs. trans. Marvine - 274 lbs. released trans. Marvine - 187 1bs. trans. elk pens trans. elk pens - RC-17l31 trans. elk pens - RC-16740 released - 289 lbs. trans. elk pens - RC-16737 trans. elk pens - RC-17l38 released - 250 lbs. released - 208 lbs. released - 212 lbs. released - 290 1bs. trans. elk pens released - broken incisors released - 244 1bs. trans. elk pens released - 272 1bs trans. elk pens released - 4x4 - 2-yr. old released trans. elk pens released - 276 1bs. trans. elk pens released - 250 lbs. trans. elk pens - RC-3334 died 1/11/88 released - 153 lbs. released released - 208 lbs. released - 254 1bs. released - 178 lbs. released - 241 lbs. released trans. elk pens released - inside of leg injured trans. elk pens - very old released - 232 lbs. trans. elk pens released - 233 lbs. released - 216 lbs. released - 218 lbs. released - 195 1bs. trans. elk pens trans. elk pens - RC-17l48 trans. elk pens - RC-25253 very thin trans. elk pens - RC-23944 released - 275 lbs. trans. elk pens �40 Table 3. - continued --------------------------------------------------------------------------------------------------Eartag no. Sex Date Age Location Remarks --------------------------------------------------------------------------------------------------79 F 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 1/9 1/9 1/9 1/11 1/11 1111 1111 1/11 1/11 1/11 1/11 1/11 1111 1/11 1/11 1/11 1/11 1/11 1/11 1/11 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 1/11 1111 1/11 1/11 1/11 1111 1/11 1/11 1/11 1/11 1/11 1/11 1/11 1/11 1/11 1/11 1111 1/11 1111 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 None None None None None None None None None 1/13 1/13 1/13 1/13 1/13 1/13 1/14 1/14 1/14 1/14 1/14 1/14 1/14 1/14 1/14 1/14 1/14 1/14 1/14 1/14 1/14 1/14 1/14 1/14 1/14 1/14 1/14 1/28 1/28 1/29 2/18 2/18 2/18 2/18 2/18 2/18 F F F F F F F F F M F F F F F F M M F F M M M M M F M M F F M M M M F F M F F F F M F F F F F F F F F F F F F F F F F F F F M F F F F F F F F F F f calf adult cal f calf adult adult calf adult adult yearling calf calf calf calf calf adult adult yearling calf adult Iles Iles Iles 11es Iles 11es I1es 11es 11es Iles Iles Iles Iles Iles Iles Iles Iles Iles Iles Iles Grove Grove Grove Grove Grove Grove Grove Grove Grove Grove Grove Grove Grove Grove Grove Grove Grove Grove Grove Grove calf calf calf calf calf calf adult calf calf calf adult calf calf yearling calf adult calf calf adul t Iles Iles Iles Iles Iles Iles Iles Iles Iles Iles Iles 11es Iles Iles Iles Iles Iles Iles Iles Grove Grove Grove Grove Grove Grove Grove Grove Grove Grove Grove Grove Grove Grove Grove Grove Grove Grove Grove calf calf adult calf adult adult adult adul t yearl ing adult adult adult adult adul t adult adult calf adult calf adult adult calf calf adult yearl ing adult adult adult adult adul t adult adult adult adult adult adult Iles Grove Iles Grove Iles Grove 11es Grove Iles Grove Iles Grove t10rgan Gul ch Morgan Gulch tiorgan Gul ch Morgan Gulch Morgan Gulch Morgan GulCh Morgan Gulch Morgan GulCh Morgan Gulch Morgan Gulch Morgan Gulch Morgan GulCh t~organ Gul ch Morgan GulCh Morgan Gulch Morgan Gulch Morgan Gul ch Morgan Gulch Morgan Gulch Morgan Gulch Morgan Gulch Morgan GulCh Morgan GulCh Morgan Gul ch Godiva Rim Godiva Rim Godiva Rim Godiva Rim Godiva Rim Godiva Rim released - 260 1bs. trans. elk pens released - 260 lbs. released - 213 lbs. trans. elk pens trans. elk pen s released - 268 lbs. trans. elk pens trans. elk pens trans. elk pens released released released released released trans. elk pens trans. elk pens released released - 286 lbs. trans. elk pens - white eartag '4 released - 271 lbs. released - 290 lbs. released - 251 lbs. released released - 301 lbs. released - 212 lbs. trans. elk pens released - 283 lbs. released - 225 lbs. released - 159 lbs. released - very poor condo released released - 178 lbs. released released - 259 lbs. trans. elk pens - RC-17l49 released - 252 lbs. released - 246 lbs. trans. elk pens - RC-25245 purple eartag '55 released - 214 lbs. released - 230 lbs. trans. elk pens released - 302 lbs. trans. elk pens trans. elk pens trans. elk pens trans. elk pens - RC-171S0 released trans. elk pens trans. elk pens trans. elk pens trans. elk pens trans. elk pens - very old trans. elk pens trans. elk pens released trans. elk pens released trans. elk pens - very old trans. elk pens released released trans. elk pens - very old released trans. elk pens trans. elk pens trans. elk pens trans. elk pens trans. elk pens trans. elk pens trans. elk pens trans. elk pens trans. elk pens trans. elk pens trans. elk pens --------------------------------------------------------------------------------------------------- �co~orado Division of Wildlife Wildlife Research Report July 1988 41 JOB PROGRESS REPORT State of Colorado Project No. W-153-R-2 -----------------------No. 3 ---------------------- Mammals Research Work Plan Multispecies Investigations Job No. 5 Period Covered: July 1, 1987 - June 30, 1988 Author: Impact of Elk Winter Grazing on Livestock Production D. L. Baker Personnel: G. Bear, M. Miller, L. Carpenter, B. Gill, K. Navo, C. Woodward, B. Seely, H. Seely, L. Lovett ABSTRACT All research results for this segment are summarized in the Mammals 2, WP9A,Jl Report. ��Co~orado Division ot Wildlife Wildlife Research Report July 1988 43 JOB PROGRESS REPORT State of Colorado Project No. W-153-R-2 ~~~------------------ Work Plan No. __.....:...3 _ Mammals Research Elk Investigations Job No. 6 Period Covered: July 1, 1987 - June 30, 1988 Author: Effect of Elk Harvest Systems on Elk Breeding Biology D. J. Freddy Personnel: Dr. M. Miller, DVM, D. Hopper, C. Wetherill, M. Cousins, and G. Byrne - Colorado Division of Wildlife; L. Wyman and Wyman Ranch personnel; E. Ryland and Forbes-Trinchera Ranch personnel ABSTRACT Rectal palpation and rectal ultrasound similarly assessed the pregnancy status of elk. Ultrasound, by providing images of fetuses, allowed detection of different sized fetuses and likely different conception dates. Heifercheck blood assays did"not perform well and may have been complicated by low progesterone levels of pregnant elk. Serum progesterone RIAs were more reliable in detecting pregnancy provided threshold values approaching 1 ng/ml for pregnant elk were considered. Reproductive tracts from elk and mule deer were collected on the ForbesTrinchera Ranch. Pregnancy rates for elk were 0% for yearling and 90% for mature cows, and fetal rates were 0.90/mature cow. Fetal sex ratio was 12M:24F which differed from the 17M:6F ratio observed in 1986. Conceptions occurred in the 62-day interval 18 September to 18 November 1987 with 97% of . the concept.Lous occurring in the 33-day interval 18 September to 20 October. Concept.Ions peaked 23 September to 2 Oc t obe r , which .was similar to 1~86. Pregnahcy rates for deer were 0% for yearling and 100% for mature .does, and fetal rates were 1.79 fawns/doe. FetCil sex ratio was 13M: 21F, which differed from 20M:13F ratio observed in 1986. Conceptions occurred in the 19-day interval from 30 November to 18 December 1987 and peaked 6-11 December, 'which was similar to 1986. Most does were in poor body condition when collected in late March, 1988, having kidney fat indexes of 16% and Kistner body condition indexes of 35. -. Reproductive tracts were collected from elk in GMU 33, a portion of the White River elk population. Pregnancy rates were 0% for yearling and 97% for mature cows, and the fetal sex ratio was 20M:15F. Conceptions occurred in the 40-day interval 4 September to 12 October and peaked 18-27 September 1987. ��45 EFFECT OF ELK HARVEST SYSTEMS ON ELK BREEDING BIOLOGY David J. Freddy P. N. OBJECTIVE To evaluate effects of harvest systems on breeding biology of elk. SEGMENT OBJECTIVE Pregnancy test cow elk from the White River (DAU-E6) and Forbes-Trinchera Ranch to compare conception rates in herds with low and high bu11:cow ratios. METHODS AND MATERIALS Large scale pregnancy testing of wild elk in the White River DAU could not be accomplished because of unfavorable trapping conditions. Collections of reproductive tracts from hunter harvests in December, 1987, and January, 1988, provided limited information on pregnancy and fetal rates and conception dates for elk in GMU 33, a portion of the White River DAU. Evaluation of rectal palpation (Greer and Hawkins 1967, Follis 1972), rectal ultrasound (White et a1. 1985), and serum progesterone assays (Weber et a1. 1982, Wood et a1. 1986) to determine pregnancy status continued using captive domesticated elk at the Wyman Elk Ranch near Craig, Colorado, and wild elk on the Forbes-Trinchera Ranch in southern Colorado. Reproductive tracts from both elk and deer were obtained on the ForbesTrinchera Ranch to provide baseline data on pregnancy and fetal rates and estimated conception dates from herds having chronically high bu11:cow and buck:doe ratios. Postseason ratios on the Forbes-Trinchera Ranch in December, 1987, were 34 bu11s:100 cows:56 calves and 46 bucks:lOO does:70 fawns. Comparative ratios for elk in the White River DAU after 3 years of antler-point restrictions were 21 bu11s:100 cows:55 calves. RESULTS AND DISCUSSION Pregnancy Testing Rectal Palpation and Ultrasound On 5 December 1987, 20 female elk on the tvyman Ranch were pregnancy tested. Elk were held in a cattle squeeze-chute, blindfolded, and pregnancy status was determined by rectal palpation and rectal ultrasound. Elk did not generally resist the pregnancy testing. Pregnancy rates were 65% and 67% based on palpation and ultrasound, respectively. The two techniques provided the same results on those elk where both techniques were used (Table 1). One yearling cow and one mature cow that were palpated did not receive an ultrasound scan. The yearling was deemed too small (palpation was difficult), and the mature cow became excited after being palpated. �46 Fetal size was variable, but all fetuses were located within the upper pelvic region of the cow where the ultrasound probe could detect their presence. Fetuses varied in size from about 20 mm in total body length to fetuses having heads 35 mm in length. The differences in fetal size suggested conception dates differed by at least 1 estrus cycle and that ultrasound could be used to judge relative conception dates. The use of ultrasound to judge conception dates appears feasible if conducted in early December as opposed to late January when fetal development and subsequent movement into the abdominal cavity of the cow precludes using ultrasound to judge fetal size (Freddy 1987a). However, even in early December, 4 of 12 fetuses had progressed in their development to the point that they were moving into the abdominal region of the cow where detection by the ultrasound probe would likely have become difficult in a few weeks. Further testing of ultrasound with cows bred on known dates is needed to adequately judge the use of ultrasound to estimate conception dates. Blood Assays The commercial blood progesterone pregnancy testing kit "Heifercheck" (American Diagnostic Sales, Inc., Westport,CT) was evaluated for accuracy using blood obtained from elk at the Wyman Ranch and on the Forbes-Trinchera Ranch. Blood was obtained via the jugular vein of Wyman elk during pregnancy testing. Antlerless elk were harvested by licensed hunters on the Forbes Ranch 12-21 December 1987. Hunters were given blood viles and instructed to collect blood from the chest cavity of their elk shortly after its death and to keep viles copl by placing them in snow, if possible. Blood viles were collected at check stations alQng with the reproductive organs of the elk so pregnancy status was known. Blood collected from Wyman elk was kept cool and centrifuged within 12 hrs and serum was then frozen. Blood from Forbes elk was kept cool and centrifuged within 24 hrs of the animal's death, and serum was then frozen. Heifercheck tests and radioimmunoassays (RIAs) were completed 16 and 18 March 1988, respectively. Four observers simultaneously and independently judged the pregnancy status of Wyman elk using Heifercheck. There were no differences among observers in judging the pregnancy status of elk (X2 [3] ~0.40, ! > 0.50). Observers correctly identified >92% of the 13 pregnant elk but only 43-71% of the 7 nonpregnant elk (Table 2). Five observers similarly judged the pregnancy status of Forbes elk, and there were also no differences among observers (X2 [4] <4.5, P > 0.30). Observers correctly judged 40-77% of the 35 pregnant adult elk, 78-89% of the 9 nonpregnant adult elk, and 40-50% of the 10 nonpregnant' calves (Table 2). These results were discouraging compared to similar, but not true blind tests conducted in 1987 (Freddy 1987a). In 1987 pregnancy status was correctly determined in 100% of 27 Wyman elk, but blood was collected from these elk in late January, 1987, and thus, later in gestation. Pregnancy status was correctly determined in 91% of 23 Forbes elk collected in early December 1986. Most errors were associated with progesterone levels of 1-2 ng/ml, which corresponds to the lower threshold of sensitivity for Heifercheck. The most common error among Wyman elk was judging nonpregnant elk as pregnant (Table 2). All observers incorrectly judged 3 yearlings, and various observers were incorrect on 3 additional elk. RIAs of these 6 females averaged 0.95 + 0.08 �47 (SE) ng/ml (range 0.67-1.23). For Forbes elk, the most common error was judging pregnant elk as nonpregnant (Table 2). RlAs for 11 elk on which nearly all observers erred averaged 2.11 + 0.36 (SE) ng/ml (range 0.75-4.34). Four of the 10 calves were incorrectly judged as pregnant by all 5 observers, and RlAs of these calves averaged 1.85 + 0.56 (SE) ng/ml while RlAs of the 5 calves correctly judged as nonpregnant by all observers averaged 0.76 + 0.46 (SE) ng/ml. Two of the calves on which all observers erred had serum cortisol levels >19 ng/ml suggesting that some calves may have had elevated progesterone levels because of stress from the hunting season (Plotka et ale 1983). Threshold levels of serum progesterone for pregnant wild ungulates have been reported as 2.0 ng/ml for mule deer (Wood et ale 1986), >2 but <3 ng/ml for mountain sheep (Ramsay and Sadleier 1979), and 3.7 ng/ml for elk (Weber et ale 1982). In particular, levels reported by Weber et ale (1982) may have been strongly influenced and elevated by the stress of trapping and drugging their elk with succinylcholine chloride (Wesson et ale 1979). Our data indicate that elk can be pregnant with serum progesterone levels <3 ng/ml. Serum progesterone averaged <3.53 ng/ml for all groups of pregnant adult elk and <O.35ng/ml for all nonpregnant adult elk. Overall ranges of serum progesterone were 0.75-14.86 ng/ml for pregnant and 0.05-1.23 ng/ml for nonpregnant elk (Table 3). Using thresholds of 3, 2, and 1 ng/ml for pregnant adult elk would have resulted in error rates of 43%, 16% and 6% (n = 69); respectively. A threshold of 1 ng/ml would have caused an 9% error rate for nonpregnant adult elk. A threshold value for pregnant elk of 1 ng/ml may therefore be more appropriate than the 3.7 ng/ml proposed by Weber et ale (1982). Summary Rectal palpation and rectal ultrasound both correctly determined pregnancy status of elk. Ultrasound, by providing images of fetuses, allowed detection of different sized fetuses and likely different conception dates. Further testing of ultrasound on pregnant elk with known breeding dates is warranted. Heifercheck blood assays did not perform well and may have been complicted by low progesterone levels of pregnant elk. Serum progesterone RlAs were more reliable in detecting pregnancy provided threshold values approaching 1 ng/ml for pregnant elk were considered. Fetal Collections Forbes Elk Reproductive tracts from 52 female elk were collected by hunters on the Forbes-Trinchera Ranch from 12-14 and 19-21 December 1987. Instructions explaining how to collect the intact uterus and ovaries along with a blood sample and a median incisor were sent via mail prior to the season to all 100 antlerless elk permittees. Reproductive tracts from 3 calves, 7 yearlings, and 42 >2-year females were obtained at check stations by study personnel, usually-within a few hours after elk were harvested. Average age of collected female elk was 4.7 + 0.5 (SE) years with a range of 0.5-16 years (Appendix A). Pregnancy was determined from the presence of fetuses, and rates were 0% for calves, 0% for yearlings, and 90% for cows >2 years old. Unlike 1986, no uteri were judged to be infected and not capable of supporting an embryo �48 (Dr. M. Miller, DVM, pers. comm.). One yearling female may have been reabsorbing an embryo while a 16-year-old female was possibly in the early stages of pregnancy, but both were considered to be nonpregnant. An extremely small embryo was found in a 15-year-old female. There were 38 fetuses observed in reproductive tracts from 49 cows >1 year old. Unlike 1986, no twins were observed. Fetal rates were 0 fetuses/yearling cow and 0.90 fetuses/cow >2 years old. Fetal rates have been lowest for cows aged as yearlings or >13 years and relatively constant for cows 2-12 years old (Table 4). Fetal sex ratios were 12M(33%):24F with 2 of unknown sex. Collecting fetuses approximately 2 wks later in 1987 compared to 1986 aided greatly in determining fetal sex as most fetuses had developed genitalia. Mid-December should be the standard period to collect this data on the Forbes-Trinchera Ranch. Fetal sex ratios differed from unity in both 1986 and 1987 (X2 (1] > 4.00, P < 0.04), differed between years (X2 (1] = 9.2, P = 0.002), but did not differ from unity when pooled among years (X2 (1] = 0.02, P = 0.60)(Table 5). The shift towards female fetuses in 1987 was primarily due to female fetuses accounting for 78% of the fetuses produced by cows 3-6 years old. Conception dates were estimated from fetal crown-rump measurements (Morrison et ale 1959). A median collection date was not used as fetuses were collected in 2, 3-day intervals 1 wk apart so fetuses were aged from the date of collection. Conceptions occurred in the 62-day interval 18 September to 18 November with 97% of the conceptions occurring in the 33-day interval 18 September to 20 October. Con~eptions peaked (59%) 23 September to 2 October (Fig. 1). In 1986, conceptions peaked (54%)·23 September to 2 October and occurred during the 35-day interval 19 September to 23 October (Fig. 1). Conception dates for elk in the San Juan Basin in southwestern Colorado in 1965 and 1966 peaked 2 October (Boyd and Ryland 1971). There was no relationship between estimated fetal age and age of cow (r = -0.075) suggesting that conception date was not influenced by age of cow. However, the ages of the 10 cows bred after 13 October during both years were 1 yearling, 3 2-year olds, and 3 >8 years old suggesting that first-time breeders or old cows can conceive later in the rut. Limited data suggest that elk on the Forbes-Trinchera Ranch and from a portion of the White River elk population are breeding at similar rates and over similar time intervals. Fetuses were collected from 37 cow elk harvested during a late-season hunt in GMU 33, which represents part of the White River elk population (Freddy 1987b)(data courtesy of G. Byrne, Biologist, Cnow). The fetal sex ratio was 20M (57%):15F, which did not differ from unity (X2 [1] = 0.7, P = 0.40). Pregnancy rates were 0% for yearlings (n = 2) and 97% for cows ~ 2 years old. Twin fetuses were not observed. Conception dates peaked (58%) 18-27 September and conceptions occurred during the 40-day interval 4 September to 12 October (Fig. 1). Male fetuses were substantially larg~r than female fetuses in body weight, crown-rump length, and hind foot length (P < 0.002, Table 6). These results contrasted with 1986 when there were no differences in measurements of male and female fetuses (p > 0.50, Freddy 1987a). Fetuses of both sexes were larger in 1987 than 1986 probably reflecting a collection date about 14 days later in 1987 (~~ 0.01, Table 6). Male fetuses were estimated to have been �7/88 49 conceived about 8 days earlier than female fetuses (P = 0.002). Eviscerated body weights of cows having male and female fetuses were not different (P > 0.30), and male fetuses came from all age classes of cows suggesting that differences in fetal size were not due to overall condition or age of cow at breeding. Fetuses collected throughout December from GMU 33, White River, were larger than fetuses from the Forbes-Trinchera, but this may primarily reflect the time of collection and not inherent differences between populations (Table 6). For the Forbes fetuses, regressions between hind foot length and crown-rump length (linear) and crown-rump length and body weight (curvilinear) were highly correlated (r2 > 0.90). Regressions may allow detecting yearly differences in fetal growth rates. Forbes Deer Nineteen adult (> 1 yr old) female mule deer were selectively shot Forbes-Trinchera-Ranch 29 March-l April 1988 to obtain reproductive indices of body condition. Reproductive tracts were obtained from and 18, >2-year-old females. Average age of these females was 5.4 years (Appendix B). on the tracts and 1 yearling + 0.8 (SE) All females> 2 years old were pregnant, but the yearling doe was not pregnant. The 34 fetuses produced fetal rates of 1.79 fawns/doe, 0 fawns/yearling doe, and 1.89 fawns/doe >2 years old. Fawns occurred as 3 singletons, 14 twins, and 1 set of triplets. The triplets occurred in a doe 8 yrs old. Fetal rates have been lowest for does aged as yearlings or >10 years and nearly constant for does aged 2-9 (Table 7). Fetal rates of Forbes-Trinchera deer have been exceeded only by fetal rates of deer in Middle Park for herds in Colorado where such data is available (Table 8). Fetal rates by age class of doe have been similar for Forbes-Trinchera and Middle Park (Table 7). Frequencies of twinning and triplets in ForbesTrinchera deer met or exceeded rates in Middle Park, the Poudre River, and the Piceance Basin. Quadruplets have been measured only in Middle Park, and triplets have not been observed in the Piceance Basin (Table 9). Fetal sex ratios in 1988 were l3M(38%):2l F overall and lM:2F for singletons, IlM:17F for twins, and lM:2F for triplets. Fetal sex ratios did not differ from unity in 1987 or 1988 (X2 [1] = 3.35 P = 0.06) but sex ratios were dif2 ferent between years (X [1] = 3.35, P = 0.06) (Table 8). Doe deer 2-4 yrs 2 old tended to produce more female fetuses in 1988 than 1987 (X [1] = 3.25, P = 0.07) while fetal sex ratios were not different between years for does 5+ years old (P > 0.54). Fetal sex ratios did not vary from unity in the Piceance, Poudre River, and Forbes-Trinchera populations when all years of data were pooled within each herd (X2 [1] = 0.0, P > 0.40) but ratios favored 2 males in 11iddle Park (X [1] = 11.3, ! = 0.001) (Table 8). The significant yearly shifts in fetal sex ratios for Forbes-Trinchera and Middle Park populations may reflect yearly differences in weather and nutrition (Gill 1972b). Conception dates were estimated from fetal crown-rump measurements (Hudson and Browman 1959) using a median collection date of 30 March 1988. Conceptions peaked (50%) between 6 and 11 December 1987 (Fig. 2). All conceptions were estimated to have occurred in the 19-day interval from 30 November-18 December. In 1986, conceptions peaked (67%) 3-8 December and occurred during the 12-day interval from 29 November-lO December (Fig. 2). There was no relationship between estimated fetal age and age of doe (r = 0.19, n = 36) suggesting that conception date was not influenced by age of doe. �so Average conception dates for Forbes-Trinchera deer were similar to estimated dates for other deer herds in Colorado (Table 10). Estimated conception intervals for Forbes-Trinchera deer were the shortest measured of any population in Colorado. This short interval cannot be attributed to high numbers of bucks available for breeding as buck:doe ratios in Middle Park and the Poudre River were also high, but breeding itervals were longer in Middle Park and the Poudre River (Table 10). There were no differences (p > 0.50) in body weight, crown-rump length, and hind foot length between male and female fetuses, although females tended to be lighter in weight (Table 11). Comparisons between male and female fetuses produced similar results in 1987 (Freddy 1987a). Fetuses were larger in 1988, probably because they were collected at a later median date of 30 March as opposed to 18 March 1987. Linear relationships occurred between crown-rump length and body weight and between hind foot length and crown-rump length (r2 > 0.93). These relationships could be used to compare fetal growth among years but only if there is a standard collection date. Mid-March is the preferred time to collect deer from the Forbes-Trinchera. Fetuses from the Forbes-Trinchera have been comparable in size to fetuses collected in Middle Park, the Poudre River, and the Piceance Basin during March of various years (Table 1). Average fetal size in the different populations was influenced by the year of collection and the days in March when collections occurred. The Forbes-Trinchera Ranch consists of 2 major areas referred to as the Blanca and Trinchera. Deer populations in these 2 large areas may function independently as observations have suggested that deer, for the most part, utilize different winter and summer ranges. Nutritional qualities of seasonal ranges are hypothesized to be different, and such differences could be reflected in sizes of fetuses. In 1988, fetuses from does collected in the Blanca were larger in weight, crown-rump length, and hind foot length (p < 0.01) and were estimated to have an earlier conception date (p = 0.003) than-fetuses from the Trinchera (Table 12). These differences were due to Trinchera female fetuses being smaller than Blanca female fetuses for all measurements (p < 0.02) as there were no differences in male fetuses between portions of the-Ranch (P > 0.10). In 1987, comparisons between areas of the Ranch were hindered by few collections from the Trinchera, but Trinchera fetuses were larger in weight, crown-rump length, and hind foot length (P < 0.05) and were estimated to have an earlier conception date (p = 0.03)(Table-12). These differences were due to Trinchera male fetuses being larger than Blanca male fetuses for all measurements (P < 0.01). This alternating in fetal size between areas may represent random noise or differences in annual production between areas of the Ranch. Further collections in forthcoming years may clarify differences in fetal size between areas of the Ranch. Body condition of adult does was variable ranging from emaciated to fair. Kistner body condition indexes averaged 35 ~ 3 (SE), which was lower than 1987 (X2 [2], P = 0.02) and placed most does in the poor condition category (Kistner et ale 1980; Table 13). Percent kidney fat (Anderson and Medin 1965a) was 16+ 2.3 (SE) for both kidneys, which was lower for both kidneys than 1987 (P < 0.04). Percent marrow fat (dry weight basis) was 92 + 0.03, which was hIgher than 1987 (p < O.OOl)(Table 13). Frequencies of marrow texture and color ratings were not different between years (X2 [2] <2.6, P > 0.26), which suggested possible problems with fat analytical procedures �51 in 1987. The generally lower indexes in 1988 compared to 1987 likely reflect the later median collection date in 1988 and not necessarily poorer nutrition in 1988. Eviscerated body weight, percents right and left kidney fat, and percent marrow fat did not differ (p> 0.08 for body weight, > 0.17 for % fat) between adult females collected in the Blanca and Trinchera portions of the Ranch in 1987 and 1988, although Trinchera does were slightly heavier and fatter both years. Body condition indexes suggest that Trinchera and Blanca females overwinter similarly while fetal size data suggest summer and fall forage may change yearly between areas of the Ranch. Body condition indexes suggest that adult female deer on the Forbes-Trinchera are performing physiologically as well or better than deer in Middle Park, the Poudre River, and Piceance Basin, Colorado (Table 13). Caution should be used when interpreting this comparative population data as collections occurred in different years and significant changes in winter conditions have been thought to affect animal performance in Middle Park and Piceance Basin (Gill 1972b, Bartmann 1986). Litter size appears related to body condition of adult females. Adult female eviscerated body weight, percent fat right kidney, and Kistner condition index increased with increasing litter size during both 1987 and 1988 (Table 14). This data is suggestive that deer reach threshold body weights to conceive multiple births and, therefore, nutritional condition prior to breeding could influence fetal rates (Julander et al. 1961). Forbes-Trinchera Hunter Harvests Elk Hunters paying a fee harvested 72 and 93 bull elk during fall 1986 and 1987, respectively. Average age of bull harvested was 4.8 + 0.1 (SE) years with no difference in average age between years (p > 0.50). Numbers of bulls harvested by age class also did not differ between years (X2 [6 ] = 1.9, P = 0.41; Fig. 3). Antler score and weight increased with age of bull with the-highest scoring and heaviest antlers coming from bulls ~8 years old (Fig. 3). Hunters having access to Forbes-Trinchera under the Ranching for Wildlife Program harvested 36 and 71 antler1ess elk during December, 1986 and 1987, respectively (includes bull calves). Average age of female elk harvested, including female calves, was 4.7 + 0.4 (SE) years with no difference in average age between years (p = 0.27). Numbers of females harvested by age 2 class also did not differ between years (X [5] 7.2, f = 0.21; Fig. 4). Eviscerated body weights of 37 adult female and 13 calf elk harvested in 1988 were 143.4 + 3.1 and 64.5 + 2.5 (SE) kgs, respectively (Fig. 4). Whole body weights for-calves were about 87 kg, assuming eviscerated body weight equaled 75% of whole body weight (Blood and Lovass 1966). Whole body weights of male and female calves from elk herds in North Park, Durango, and Creede, Colorado, in January or February, 1986, averaged 116, 105, and 91 kgs, respectively (CDOW unpubl. data), suggesting Forbes-Trinchera calves may be undersized. �52 Deer Hunters paying a fee harvested 87 and 110 bucks during fall 1986 and 1987, respectively. Average age of buck harvested for both years was 6.1 + 0.13 (SE) years, but average age differed between years (p < 0.001) with ages being 5.6 + 0.20 and 6.6 + 0.17 (SE) in 1986 and 1987, respectively. Numbers of bucks harvested by age class also differed between years (X2 [6] = 18.0, P = 0.006) primarily due to more bucks >8 years old being harvested in 1987 (Fig. 5). Antler score and weight increased with age with maximum scores and weights occurring at ages 3 and 7, respectively (Fig. 5). Hunters having access to Forbes-Trinchera under the Ranching for Wildlife Program harvested 65 and 139 antlerless deer (including buck fawns) in 1986 and 1987, respectively. Average age of female deer harvested, including female fawns, was 3.1+ 0.17 (SE) years with no difference between years (p > 0.50). Numbers of females harvested by age class also did not differ between years (X 2 [7] = 5.7, ~ = 0.58; Fig. 6). Eviscerated body weights of 108 adult female and 21 fawn deer harvested in 1988 averaged 40.8 + 0.6 and 23 + 1.1 (SE) kgs, respectively (Fig. 6). Whole body weight of fawns was about 31 kg, assuming eviscerated body weight equaled 75% of whole body weight (Anderson et ale 1974). Whole body weights of fawns collected in December in the Piceance Basin averaged 33 kg (Trobit et ale 1988) where starvation over winter has accounted for > 50% of the fawn mortality (White et ale 1987). This weight data suggests that Forbes-Trinchera fawns may be undersized and subject to low survival rates. Summary Morphological measurements of harvested deer and elk, determining age structures of harvested animals, fetal collections, and scientific collections of female deer should be continued to allow monitoring changes in the populations potentially caused by the harvest systems used on the ForbesTrinchera Ranch. Body sizes of fawns and calves harvested in December suggested these age classes may be undersized and subject to low winter survival rates. Pregnancy rates and fetal rates of both elk and deer indicated that any problems in recruitment are likely associated with survival and not production of young. Nutritional factors may be influencing both elk and deer as significant yearly shifts in fetal sex ratios have occurred although body condition of adult female deer in late winter has been satisfactory. Assigning Ages to Elk and Deer ~~ny antlered and antlerless elk and deer were assigned an age based on both dental cementum and replacement-wear techniques during 1986 and 1987. Many serious discrepancies occurred (Freddy, D. J., unpubl. rep.). There appear to be significant problems in aging male elk and deer. Ages based on dental cementum were consistently younger than those based on replacement-wear. Which technique was correct is unknown, but corroborative antler score data suggest that replacement-wear provided a more reasonable estimate for both male elk and deer. Dental cementum provided accurate estimates for known-aged antlerless elk and generally agreed with replacement-wear estimates of age for mature cows. Dental cementum did not accurately age fawn and yearling antlerless deer but generally agreed with replacement-wear estimates of age �53 for female deer >2 years old. Because of these problems, ages of animals harvested in 1986 and 1987 at Forbes-Trinchera were estimated as follows: 1) ages of male elk and deer were estimated from replacement-wear, 2) ages of antlerless elk were estimated from dental cementum, and 3) ages of antlerless deer were estimated from replacement-wear for fawns and yearlings and from dental cementum for females >2 years old. LITERATURE CITED Anderson, A. E., and D. E. Medin. 1964. Cache la Poudre deer herd, Colorado. Res. Rep. January(1):238-268. An ecological investigation of the Colo. Game, Fish, and Parks Game , and -----deer herd 1965a. Two conditions indices of the Cache la Poudre mule and their application to management. Colo. Div. Game Info. Leaflet 23. 3pp. _____ , and 1965b. An ecological investigation of the Cache la Podure deer herd, Colorado. Colo. Game, Fish, and Parks Game Res. Rep. January (1):165-195. , and -----deer herd, 1965c. An ecological investigation of the Cache la Poudre Colorado. Colo. Game, Fish, and Parks Game Res. Rep. January (4):531-552. , and -----deer herd, 1966. An ecological investigation of the Cache la Poudre Colorado. Colo. Game, Fish, and Parks Game Res. Rep. January (2):275-290. , -----mule , and D. C. Bowden. 1972. Indices of carcass fat in a Colorado deer population. J. Wildl. Manage. 36:579-594. , , and -----selected bones, 1974. Growth and morphometry of the carcass, organs, and glands of mule deer. Wildl. Monogr. 39. 122pp. Armstrong, R. A. 1950. Fetal development of northern white-tailed deer. Amer. Mid. Nat. 43:650-666. Bartmann, R. M. 1972. Piceance deer study - productivity and mortality. Colo. Div. Wildl. Game Res. Rep. July(3):345-350. 1973. Piceance deer study - productivity and mortality. Wildl. Game Res. Rep. July(2):255-261. Colo. Div. 1986. Growth rates of mule deer fetuses under different winter conditions. Great Basin Nat. 46:245-248. Blood, D. A., and A. L. Lovass. 1966. Measurements and weight relationships in Manitoba elk. J. Wildl. Manage. 30:135-140. �54 Boyd, R. J., and E. E. Ryland. 1971. estimated by fetal growth curves. Info. Leaflet 88. 2pp. Breeding dates of Colorado elk as Colo. Div. Game, Fish, and Parks Game Follis, T. B. 1972. Reproduction and hematology of the Cache elk herd. Div. Wildl. Resouces Publ. 72-8. 133pp. Utah Freddy, D. J. 1987a. Effect of elk harvest systems on elk breeding biology. Colo. Div. of Wildl. Game Res. Rep. July:lOl-120. 1987b. The White River elk herd: Div. Wildl. Tech. Bull. 37. 64pp. a perspective 1960-1985. Gill, R. B. 1969. ~liddle park deer study-productivity Game, Fish, Parks Game Res. Rep. July(1):123-l40. Colo. and mortality. Colo. 1970. Middle Park deer study - productivity and mortality. Game,Fish, Parks Game Res. Rep. July (3):337-354. Colo. 1971. Middle Park deer study - productivity and mortality. Game, Fish, Parks Game Res. Rep. July(2):189-207. Colo. 1972a. Middle Park deer study - productivity and mortality. Div. Wildl. Game Res. Rep. July(2):179-198. Colo. 1972b. Productivity studies of mule deer in Middle Park, Colorado. Proc. Second. Mule Deer Workshop. Greer, K. R., and W. W. Hawkins. 1967. Determining pregnancy in elk by rectal palpation. J. Wildl. Manage. 31:145-149. Hudson, P., and L. G. Browman. 1959. Embryonic and fetal development of the mule deer. J. Wildl. Manage. 23:295-304. Julander, D., W. L. Robinette, and D. A. Jones. 1961. Relation of summer range condition to mule deer herd productivity. J. Wildl. Manage. 25:54-60. Kistner, T. P., C. E. Trainer, and N. A. Hartmann. 1980. A field technique for evaluating physical condition of deer. Wildl. Soc. Bull. 8:11-16. Morrison, J. A., C. E. Trainer, and P. L. Wright. 1959. Breeding seasons in elk as determined from known-age embryos. J. Wildl. Manage. 23:27-34. Plotka, E. D., U. S. Seal, L. J. Verme, and J. J Dsoga. 1983. The adrenal gland in white-tailed deer: a significant source of progesterone. J. Wildl. Manage. 47:38-44. Ramsay, M. A., and R. M. F. S. SadLeI r , 1979. Detection of pregnancy in living bighorn sheep by progestin determination. J. Wildl. Manage. 43:970-973. Roper, L. A. 1972. Middle Park deer study-physical characteristics and food habits. Colo. Div. Wildl. Game Res. Rep. July(2):199-209. �55 Torbit, S. C., L. H. Carpenter, R. M. Bartmann, A. W. Alldredge, and G. C. White. Calibration of carcass fat indices in wintering mule deer. J. Wildl. Manage. (In press). Weber, B. J., M. L. Wolfe, and G. C. White. 1982. Use of serum progesterone levels to detect pregnancy in elk. J. Wildl. Manage. 46:835-837. Wesson, J. A., III, P. F. Scanlon, R. L. Kirkpatrick, and H. S. Mosby. 1977. Influence of chemical immobilization and physical restraint on steroid hormone levels in blood of white-tailed deer. Can. J. 2001. 57:768-776. White, I. R., A. J. F. Webster, I. A. Wright, and T. K. Whyte. 1985. Realtime ultrasonic scanning in the diagnosis of pregnancy and estimation of gestational age in cattle. The Vet. Record 117:5-8. White, G. C., R. A. Garrott, R. M. Bartmann, L. H. Carpenter, and A. W. Alldredge. 1987. Survival of mule deer in northwest Colorado. J. Wildl. Manage. 51:852-859. Wood, A. K., R. E. Short, A. Darling, G. L. Dusek, R. G. Sasser, and C. A. Ruder. 1986. Serum assays for detecting pregnancy in mule and white-tailed deer. J. Wildl. Manage. 50:684-687. Prepared by -:L:~.,/~.t'L.~~~~~C"";;;~f-;-~"'=;'(;-7;("-(_' David' J,. Pr eddy.>" Wildlife Researcher _ �56 Tabl e 1. Pregnancy status of CON elk on the Ny.an Elk Ranch as determined by rectal palpation, rectal ultrasound, and blood assays, 5 Decelllber 1987. Age Elk No.- (years) Pregnanc~ Status Palpation Ultrasound Bloodb RIA Seruac Progesterone ng/.l BL OR OR GR 6R YE YE 1.5 2.5-3.5 2.5 1.5 1.5 1.5 2.5 no no no no no no no no no no unknown no no no yes yes no yes yes no yes 1.08 0.67 0.12 0.99 0.80 0.26 1.23 105 YE 121 YE 138 YE 142 YE 143 YE 145 YE 153 YE 164 YE 174 YE 175 YE 177 YE 180 YE no tag 2.5-3.5 3.5-4.5 2.5-3.5 3.5 4.5 3.5 3.5 yes yes yes yes yes yes yes yes yes yes yes yes yes yes yes unknown yes yes yes yes yes yes yes yes yes yes yes yes yes yes yes yes yes yes yes yes yes yes yes 2.65 4.35 0.91 2.04 2.76 2.38 4.71 3.80 3.55 3.62 3.07 5. 14 3.99 4 11 37 45 50 173 200 10 r: ..J • ..J 3.5 2.5-3.5 4.5 4.5 2.5-3.5 -Elk eartag nUlber and color: BL=blue, OR=orange, 6R=green, YE=yellow bMHeifercheckM co~~ercial blood test, done 16 Harch 1988, result is Most frequent score of 4 independent observers; observers agreed 1004 except on salllpies 11 OR, 138 YE, 200 YEo cRadioi.munoassays done 18 ~arch 1988. �57 Table 2. Frequencies of correct and incorrect judg.ents of pregnancy status of elk by oblervers using Heifercheck pregnancy assay. Herd Pregnancy Status Observer Score O~server C A kno",n R F '" Wyun Forbes (Adults) Forbes (Calves> Pregnant Pregnant Non-Preg. Non-Preg. Pregnant Pregnant Non-Preg. Non-Preg. Non-Preg. Non-Pr-eg. Correct Incorrect Correct Incorrect 13 Totals Correct Incorrect Correct Incorrect Table 3. 1987. 20 20 20 22 13 14 21 25 10 27 8 1 8 8 7 2 44 44 4 4 6 5 5 5 5 4 6 10 10 10 10 10 20 3S 9 Total s 44 Correct Incorrect 10 Seru. progesterone 20 12 1 7 Totals 13 0 2 3 13 0 2 5 13 0 2 5 10 4 5 8 1 4-4 level! lIeasured in different 44 Pregnancy Status Age Class n NYllan Jan 1987 Pregnant Non-Preg. Adult Adult 21 6 3.35 0.31 0.29 0.01 1. 15 0.05 6.14 0.59 WYlllan Dec 1987 Pregnant Non-Preg. Adult Adult 13 7 3.31 0.74 0.32 0.16 0.91 0.26 5.14 1.23 Forbes Dec 1987 Pregnant Non-Preg. Non-Preg. Adult Adult Calf 35 10 3.53 0.35 1.24 0.41 0.08 0.38 0.75 0.06 O. 11 14.86 0.80 2.92 -Tille when blood collected. 44 6 groups of elk in Herdl Year- 9 23 12 8 1 RIA Serull Progeterone nglill SE ' I'Iean l'Iax "in �58 Table 4. Fetal rates and sex ratios for age classes of fe.ale elk collected on the Forbes-Trinchera Ranch in Deceaber 1986-87. Feule Age (yrs) Number Collected Feul es Fetuses 0 4 1 13 2 3-4 5-6 7-8 9-10 11-12 13-16 11 21 13 9 4 3 8 0 2 11 21 12 8 5 3 5 Fetusesl Feule Fetal Sex F:Unk ": 0.00 0.15 1.00 1.00 0.92 0: 0:0 1: Osl 4: 4:3 7:12:2 5: 7:0 5: 2: 1 1: 3: 1 2: 1:0 4: 1:0 Q.89 1.25 1.00 0.63 Table 5. Fetal sex ratios from adult female elk collected on the Forbes-Trinchera Ranch in December 1986 and 1987. Fetal Sex Unknown No. Year Fetal Sex Fellate "ale No. (% ) No. (7.) 1986 1987 17(74) 12(33) 6(26) 24 (76) 6 2 29(49) 30 (51) 8 Totals �59 Table 6. l1easure.ents of elk fetuses from the Forbes-Trinchera in 1986-87 and fro. G"U 33, 1987, White River, Colorado. Herdl Year Body Weight (g) t1ale Feule Cro"n-Ru.p length (•• ) "ale Feule Ranch Hind Foot length (u) "ale Fuale ForbesTrinchera 29 Nov-5 Dec 1986 "ean SD .in lIa)( n 27.6 13.8 12.2 58.5 17.0 20.9 7.0 11.9 29.0 6.0 90.2 16.7 63.S 115.5 17.0 86.0 9.5 72.0 94.0 5.0 20.5 4.8 13.5 28.5 17.0 19.8 2.7 16.5 23.0 5.0 118.7 54.7 66.0 242.0 12.0 60.4 33.2 12.0 135.0 24.0 141. 4 21.1 113.0 180.0 12.0 115.4 23.3 66.0 151. 0 24.0 42.3 10.9 31.0 68.0 12.0 30.3 8.8 14.0 46.0 24.0 239.2 139.0 140.0 510.0 6.0 184.8 42.0 105.0 230.0 14.0 188.3 33.7 165.0 255.0 6.0 56.6 16.3 25.0 75.0 14.0 59.8 12.3 47.0 80.0 5.0 12-14 &: 19-21 Dec 1987 tiean SD lIin lDax n GMU 33 White River 1-30 Dec 1987 Mean SD lIin fftax n 257.9 123.6 49.0 415.0 14.0 �60 Table 7. Fetal rates for age classes of adult female .ule deer Forbes-Trinchera Ranch in 1987-88 and in Middle Park 1969-72. Age of Doe (yrs) Number Does Forbes-Trinchera 1 2-3 4-5 6-7 8-9 L 10 Middle 1 2 3-7 L8 Collected Fetuses Fetuses/ Doe 1987-88 4 4 13 6 6 4 25 12 13 8 4 5 1.00 1.92 2.00 2.17 2.00 1.25 Park 1969-72 26 18 88 40 33 38 163 69 1.27 2. II 1. 85 1.72 on the �61 Tabl e 8. Pregnancy and fetal rates for Forbes-Trinchera lule deer populations in Colorado. Herd Year ForbesTrinchera 1987 1988 Totals or Ave. Middle Parka 1969 1970 1971 1972 Totals or Ave. Piceanceb Basin 1971 1972 1973 . Totals or Ave. Poudrec River 19611965 No. Deer Collected Yearl. L_2yrs No. Fetuses Fetal Rate Fetal Sex Ratio P1:F 3 1 15 18 100 0 100 100 33 34 1.83 1.79 20:13 13:21 4 33 75 100 67 1.81 33:34 2 7 11 6 39 35 40 33 79 69 81 73 1.93 1.64 1.59 1.87 43:25 47:23 43:32 29:27 26 147 302 1.75 162: 107 5 5 19 12 15 46 100 100 79 100 93 89 28 29 94 1.65 1.45 1.45 13:15 16:13 45:47 29 73 86 92 151 1.48 74:75 86 1.72 36135 --50-- a Gill 1969, 1970, 1971, 1972 Sarhann 1972, 1973 c Anderson 1966 It Percent Deer Pregnant Yearl. L_2yrs and other �62 Table 9. Frequencies lule deer populations of litter sizes for Forbes-Trinchera in Colorado. and other Frequency of Litter Size Herd Year 0 1 2 3 4 ForbesTrinchera 1987 1988 0 1 4 13 3 14 1 1 0 0 1 3 7 19 27 73 2 0 5 0 0 5 1 0 10 7 19 5 25 28 31 34 5 2 0 0 0 0 0 6 3 41 24 118 68 7 1 4 1 0 1 9 6 0 0 18 11 10 38 0 0 0 10 10 33 59 32 58 0 0 0 0 0 15 34 0 30 68 Totals Percent I'tiddleParka 1969 1970 1971 1972 Totals Percent Piceance Basin" 1971 1972 1973 Totals Percent Poudre Riverc 1961- 9 1 0 0 1965 Percent a Gill 1972 " Bart.ann 1972, 1973 c Anderson 1966 2 0 �63 Table 10. Esti.ated conception dates for the Forbes-Trinchera other .ule deer populations in Colorado. and Conception Interval (days) Breeding Buck: Doe Ratios (":100F) Herd Fetal Collect Year Average or Peak Date(s) Of Conception Range in Conception Dates ForbesTrinchera 1987 1988 6 Dec 1986 9 Dec 1987 29 Nov-tO Dec 30 Nov-18 Dec 12 2t 52 46 Middle Park- 1969 1970 1971 1972 28 Nov 1968 1 Dec 1969 28 Nov 1970 24 Nov 1971 11 16 11 11 Nov-2 Dec Nov-15 Jan Nov-15 Dec Nov-5 Dec 40 61 35 25 53 4S 45 55 Piceanceb Basin 1971 1972 1973 4 Dec 1970 24 Nov-29 Dec 14 Nov-3 Dec 12 Nov-9 Jan 36 28 Nov 1971 4 Dec 1972 59 28 19611965 25 Nov-8 Dec 11 Nov-10 Feb 92 37-49 Poudre Riverc - Sill 1972a Bart.ann 1972,1973 c Anderson 1966 b 20 �b4 Table 11. Heasure.ents of fetal size for Forbes-Trinchera aule deer populations in Colorado. Herd/ Year Body Weight (g) Hale Feule CroMn-Ru.p length (fIR!) Hale Feule and other Hind Foot length (II1II) Hale Feule ForbesTrinchera March 1987 Mean SO lIin e ax n 243.9 54.8 164.0 382.0 20.0 234.2 68.7 156.0 388.0 13.0 188.1 11.2 175.0 213.0 20.0 188.4 13.9 171. 0 215.0 13.0 66.8 5.6 60.0 81.0 20.0 67.1 7.8 56.0 82.0 13.0 436.8 138.0 226.0 704.0 13.0 422.9 98.2 242.0 6(17.0 21.0 228.B 22.1 IBB.O 269.0 13.0 231. 4 18.4 192.0 263.0 21.0 87.2 11.9 65.0 107.0 13.0 88.3 10.0 70.0 106.0 21.0 389.2 171. 4 112.0 742.0 43.0 342.8 160.8 104.0 742.0 32.0 225.1 34.8 154.0 284.0 43.0 213.5 34.6 156.0 288.0 32.0 85.1 18.0 50.0 117.0 43.0 79.5 17. 1 48.0 116.0 32.0 217.4 30.7 158.0 280.0 14.0 227.6 33.5 155.0 283.0 20.0 226.6 21.7 200.0 251.0 5.0 212.1 33.8 146.0 March 1988 Mean SO min lIIax n Middle Parka Harch 1969-72 Mean SO sin max n Piceance Basinb March 1971-73 !'lean SD min IIi)( n Poudre Riverc March 1961-65 Mean SD .in lIlax n 398.4 121. 3 289.0 589.0 5.0 389.0 154.2 98.0 S74.0 9.0 a 6i 11, R. B., unpubl i shed data b Bart.ann, R. H., unpublished data c Anderson 1964, 1965a, 1965b, 1966 247.0 7.0 84.6 10.8 75.0 98.0 5.0 84.0 18.4 43.0 102.0 9.0 �65 Table 12. Sizes of .ule deer fetuses fro. the Trinchera Ranch in 1987-88. portions of the Forbes-Trinchera CrotolnAreal Year Trinchera 1987 Stat. Fetal "eight Ru.p (I}) length .. Hind Foot ( ) Length and Blanca < •• ) Fetal Age (days) Mean SD n 287.4 87.4 8.0 195.2 15.6 8.0 71.5 9.6 105.3 4.3 B.O B.O 1988 !'lean SD n 381. 7 128.5 17.0 221.6 22.3 17.0 83.2 11.5 17.0 111.5 S.2 17.0 Blanca 1987 Mean SD n 225.0 39.5 25.0 185.9 10.0 25.0 65.4 4.3 25.0 102.6 2.5 25.0 1988 Mean 474.7 73.0 17.0 239.1 11.6 17.0 92.6 7.2 17.0 116.0 2.8 17.0 SD n �66 Table 13. Indexes of body condition for adult fe.ale .ule deer for the Forbes-Trinchera and other deer populations in Colorado. Herdl Year Whole Body wt. (kg) Evis.Body Wt. (kg) Kidney Fat (X) Ri ght left l'IarroN Fat ('X) Kistner Fat Index Forbes-Trinchera "arch 1987 Mean SD lIin III ax n 57.4 7.0 45.0 71.0 18.0 40.0 5.0 30.0 2.0 18.0 26.0 18.0 7.0 62.0 18.0 28.0 22.0 7.0 83.0 18.0 21.1 20.1 4.0 87.0 18.0 53.0 18.0 25.0 85.0 18.0 57.3 7.7 40..0 73.0 19.0 38.3 5.4 28.0 50.0 19.0 16. 1 10. 1 6.0 45.0 19.0 16.4 10.0 5.0 41.0 19.0 92.1 12.3 42.0 98.0 19.0 35.0 13.6 10.0 55.0 19.0 21.0 12.3 17.0 28.5 17.7 10.2 15.8 23.0 88.9 69.6 83.4 81.9 58.7 42.2 18.2 69.1 34.6 16.3 93.4 87.9 43.3 "arch 1988 Mean SD lBin lIay. n Middle Park" March n = 8/yr Mean-1969 l1ean-1970 "ean-1971 l1ean-1972 60.6 Piceance Basine 1982-1983 n = 6/.on Decellber February April 63.7 67.3 59.4 Poudre Riverct Winter n=25 !'lean1961-1965 40.8 ---29.4--- 90.0 • Eviscerated weight was whole "eight less internal organs and fluids . •• Roper 1972 Torbit et ala 1988 ct Anderson et al. 1972 C �67 Table 14. Li tter size in relation to body condition indexes of adult fe~ale .ule deer on the Forbes-Trinchera Ranch, 1987-88. No. Fetuses Eviscerated Bod~ tit.(kg) SE "ean Right Kidne~ Fat % SE "ean Kistner Index SE "ean Year n 1987 1988 0 1 28 1987 1988 4 3 34.5 37.3 2. 1 1.2 11 9.7 2. 1 1.9 35 26.7 4.6 8.8 2 1987 1988 13 14 40.5 38.4 1.4 1.2 27.9 17.5 4.7 3.0 56.9 37.9 4.3 3.4 3 1987 1988 1 0 1 45 50 15 15 60 17 80 40 �68 CONCEPTION DATES FOR ELK 1~86, 1~81 40 35 P E R 38 2$ 20 ~ T ~ 15 to ? J 5 • I 9/8 v v V V Iv J7l ~ VII Sl18 Sl28 1118 10/18 11121 MONTHANDDAY(Begin 5-dav intenals) .1986 n=26 fli}ISl1 n=31 1'1 un U/17 CONCEPTION DATES FOR ELK, GHU 33, 1987 50r------------------------------------, 45 4. P 35 R 38 F 25 20 ~ 15 E 10 5 o 9/8 9/t8 9128 1011 10118 18/28 MONTHANDDAY(Begin 5-d~\I intena1s) .1987 n=36 lin 11117 Fig. 1. Estimated conception dates for elk on the Forbes-Trinchera Ranch in 1986 and 1987 (TOP) and in GMU ..,....,. ..;,..;> , White River, Colorado, 1987 (BOTTOM). �69 CONCEPTION DRTES FOR DEER 1986. 1987 40r-------------~~--------------------~ 35 P E R E N T 25 20 15 10 5 I Nov 27 Nov 30 Dec 3 Dec 5 Dec S Dec 12 Dec 15 Dec II MONTHANDDAY(Begin 3-d." intenals) glS15 n=11 E'lJ1S11 n=11 Fig. 2. Estimated conception dates for mule deer on the Forbes-Trinchera Ranch in 1986 and 1987. �70 AGES OF BULLS HARVESTED 1986, 1987 n=165 ~ E Q U E N C Y 50 45 40 35 30 25 28 IS 10 5 0 0 2 I 3 4 , 5 7 8 9 BULL AGE <YEARS) GROSS ANTLER SCORES 1987, n=85 BULLS 350 A 300 T l E R 250 S C 0 R E 150 N 200 lOa 50 0 0 1 2 3 4 5 6 7 8-9 7 1-9 BULL ACE (YEARS) • AVE. / CLASS WEIGHTS OF BULL ANTLERS 1987, n=85 20 W iC H T 18 16 14 12 10 8 l B S s 4 2 0 0 1 2 3 4 5 6 BULL AGE (YEARS) _AVE. / CLASS Fig. 3. Ages (TOP), gross Boone and Crockett antler scores (MIDDLE), and weights of antlers (BOTTOM) from bull elk harvested on the ForbesTrinchera Ranch in 1986 and 1987. �71 AGES OF FEMALE ELK HARVESTED 1986, 1987 n-96 16 14 F R E Q 12 18 N •, Y 4 U E C 2 0 1 0 2 3 6 4 5 FEf1AlE AGE (YEARS) 7 1 !-28 WEIGHTS OF ANTLERLESS ELK HARVESTED 1987 n=50 W E I G H T K G S 160 140 120 100 80 60 40 20 0 0 1 2-3 ANTlERlESS 4-5 6-1 '-20 ACE (YEARS) Fig. 4. Ages of female elk harvested (TOP) and eviscerated body weights of antlerless elk harvested in December (BOTTOM) on the Forbes-Trinchera Ranch in 1986 and 1987. �72 RGES OF BUCKS HRRVESTED 1986, 1987, n=87, 109 F R E Q U E N C Y 1 2 3 4 5 , 7 8 !-18 BUCK ACE (YEARS) .1986 mas81 GROSS ANTLER SCORES 1987, n=102 BUCKS A N T L E R S C o R E 180...-----160 140 120 100 80 60 40 20 e..L...-,-~- o 345 BUCK AGE (YEARS) IIAVE.! CLASS WEIGHTS OF BUCK ANTLERS 1987, n=102 w i H T 7~--------------6 5 4 3 hS 2 I O..L...-,---..- o 1 345 BUCK AGE (YEARS) • AVE. / CLASS Fig. 5. Ages (TOP), gross Boone and Crockett antler scores (MIDDLE), and weights of antlers (BOTTOM) from buck mule deer harvested on the ForbesTrinchera Ranch in 1986 and 1987. �73 RGES OF FEMRLE DEER HRRVESTED 1986, 1981, 0=188 40 35 F 30 E II U 25 20 N 15 Y 10 R E C 5 0 0 1 2 3 4 FEMALE AGE (YEARS) 5 6 1-1' WEIGHTS OF RNTLERLESS DEER HRRVESTED 1981, n=129 Fig. 6. Ages of female mule deer harvested (TOP) and eviscerated .body weights of antlerless deer harvested (BOTTOM) on the Forbes-Trinchera Ranch in December 1986 and 1987. �74 Appendix A. Reproductive leasureaents frol felale elk collected and 19-21 Decelber 1987 on the Forbes-Trinchera Ranch. Elk ID No. ElkAge 30087 30187 30387 30887 31287 31487 31587 31887 31987 32087 32187 32287 32387 32487 32687 32787 32887 32987 33087 33287 33587 33987 34087 34187 34287 34387 34687 34787 34887 37887 37987 38087 38187 38587 42087 42287 46187 33387 a 3 2 4 9 3 3 3 3 3 2 13 4 4 3 6 2 4 3 8 6 3 3 2 6 5 6 8 5 2 10 5 7 6 6 3 16 9 Sex F F F F F F F " F F 11 F'" " I't 11 -M F F 11 Body Weight Fetal "easure.ents" CrownHind (g) RUlp (•• ) Foot (l1li) 12-14 Hind Leg (II) Agee Days 26 52 135 58 50 67 38 75 67 21 93 111 34 85 66 144 161 59 62 242 124 91 110 151 118 110 127 100 123 127 84 133 139 107 131 113 161 162 114 122 180 145 20 30 46 30 28 34 26 34 32 19 39 41 24 37 31 49 50 31 32 68 42 29 45 69 49 43 50 43 52 51 32 61 66 40 58 50 78 77 52 54 95 68 69 75 86 77 75 80 72 79 80 67 82 83 74 81 78 89 89 76 79 93 84 99 69 77 13 98 182 104 33 72 145 121 132 70 136 165 141 104 124 91 148 135 118 122 66 12 43 32 37 14 39 52 41 24 32 22 43 36 32 29 14 65 53 57 23 84 79 81 62 82 90 83 73 79 69 85 82 77 79 60 33 '" F I't F F F '" F F 'F" 26 F F F F F 114 92 62 50 12 0.5 64 81 64 41 52 37 70 60 52 48 21 3 - Age in years for fe.ales L7 years based on dental ceeentul and for felales ~1 year old based on dental celentul or replacelent and wear . •• See Ar~strong 1950 for description of leasurelents. e Age based on ,",orrison et al. 1959. �75 Appendix B. Reproductive .easure.ents froM fe.ale aule deer collected 29 Harch-l April 198B on the Forbes-Trinchera Ranch. Deer ID No. IBB 188 18B 288 2BB 38B 388 488 488 588 5BB 6B8 68B 788 788 888 988 988 1088 1188 1188 1288 1388 1388 1488 1588 1588 1688 1688 1788 1788 1888 1888 19B8 1988 • b c DeerAge 8 B 8 7 7 3 3 2 2 6 6 3 Sex 11 F F F 11 11 F H F 1'1 F F 3 M 7 7 12 2 2 10 10 10 1 2 2 8 3 3 3 3 5 5 9 9 2 2 F 11 F F 11 F F F F 11 11 F F M 11 F F 11 F F F Body Weight Fetal Heasure.entsb CrownHind (g) Ruap (•• ) Foot 240 349 406 301 291 431 404 477 432 476 491 362 361 607 704 40B 272 226 242 493 518 198 221 226 206 204 225 227 239 233 245 244 225 226 263 269 226 203 188 192 245 250 473 490 446 409 390 456 442 533 507 638 585 341 359 244 237 230 225 222 226 235 252 250 252 257 224 224 (u) Hind Leg (•• ) Agee Days 110 110 110 .107 90 83 82 105 107 90 71 65 70 93 96 97 105 114 105 94 112 110 liB 118 119 121 109 111 138 137 117 92 88 94 129 135 94 92 88 87 87 91 89 106 93 105 106 81 81 124 121 120 113 114 123 116 124 116 136 142 107 110 116 116 114 112 112 114 114 119 119 120 120 112 112 72 eo 85 79 74 87 8a 90 89 92 Age in years for does L 2 yrs based on dental ce.entu. ~ 1 yr old based on replace.ent and wear. See Armstrong 1950 for description of .easureMents. Age based on Hudson ~nd Brow_an 1959. and for dOe! 107 113 113 115 115 117 117 113 113 122 122 113 105 105 104 118 118 ��July 1988 JOB PROGRESS REPORT State of Colorado Project No. Mammals Research W-153-R-2 ------------------------ Work Plan No. 3 --~------------------ Elk Investigations Job No. 7 Period Covered: July 1, 1987 - June 30, 1988 Author: Elk Census Methodology D. J. Freddy Personnel: R. Firth, B. Thompson, G. White, R. Bartmann ABSTRACT Number of elk in a portion of GMU 18, Middle Park, Colorado, was estimated by flying line transects with a helicopter. Estimated density of elk was 14.5 + l2.4/mi2, and the projected population was 1,276 + 1,091. The minimal number of elk in the sampled area was known to be 744. The high variance about the mean estimate was caused by the small number of groups seen (33) and the large range in group size (1-83). The small number of groups seen precluded assessing the utility of line transects to estimate numbers of elk. Next year, transects will be replicated 4 times to increase the number of observed groups to 130+, which should allow for better analyses of line transect data. ��7/88-3,7 79 ELK CENSUS METHODOLOGY David J. Freddy P. N. OBJECTIVE Evaluate methods to estimate numbers of elk during winter. SEGMENT OBJECTIVES 1. Determine the distribution and group sizes of elk in a selected herd during early and late winter using concentrated aerial surveys. 2. Use computer simulation techniques to test the effectiveness of various sampling systems to estimate numbers of elk from the known distribution of elk. METHODS AND MATERIALS A portion of Game Management Unit (GMU) 18 within Middle Park, Colorado, was selected to estimate numbers of elk. The area between Highway 125 and Troublesome Creek, about 88 mi2, was surveyed with 24 line transects (Burnham et ale 1980) spaced at 3,280 ft (1,000 m) intervals. Transects were 3 to 9 mi in length and vegetation included sagebrush, aspen, aspen-mixed conifer, and heavy conifer. Line transects were flown with a Bell-Soloy helicopter on 15 and 22 January because inclement weather precluded all transects being flown in one day. Bright sunshine, minimal wind, and excellent snow background occurred on both days. Air speed was about 50 mph, flight elevation was about 200 ft above ground, and 4 hrs were needed to fly the 137 mi of transects. Two observers and the pilot searched for elk, although the primary responsibility of the observer seated in the middle of the aircraft was to navigate. Numbers of elk in each group on both the right and left sides of the aircraft were counted. Efforts to count elk on the left side were less intense than efforts to count elk on the right side of the transect. The observer seated on the right counted and estimated perpendicular distances to those groups of elk located from the transect center line to the right. The middle observer counted and estimated perpendicular distances to those elk to the left of the transect center line. Markers placed at known distances from a transect line were used by observers to practice estimating perpendicular distances prior to flying each day. Total number of elk in the sampled area was estimated using PROGRAM TRANSECT (Laake et ale 1979). Groups of elk observed to both the left and right side of the aircraft were included in analyses (G. White, pers. comm.). RESULTS AND DISCUSSION Estimated density of elk was 14.5 + 12.4/mi2 (95% CI). The projected elk population was 1,276 ~ 1,091. The-minimal number of elk in the sampled area was 744 based on a sex-and-age classification flight flown 23 January 1988. �80 The high variance about the mean estimate was caused by the small number of groups seen (33) and the range in group size (1-83)(Table 1). Average number of elk/group was 12.788. The best sighting functions were the exponential polynomial and the negative exponential (Table 2). These sighting functions also provided the best fit for distribution of deer groups on line transects in the Piceance Basin (G. White, pers. comm.). The data suggested that elk were moving off the center line of the transect prior to being detected by observers. Flying transects, estimating sighting distances to elk, and counting elk went reasonably well for this first attempt. However, the low numbers of groups seen precluded any meaningful conclusions about the utility of using line transects to estimate numbers of elk. Next year, transects will be flown in at least 4 replicates to expand the number of groups seen to 130+, which should allow for more meaningful analyses of the line transect data (G. White, pers. comm.). LITERATURE CITED Burnham, K. P., D. R. Anderson, and J. L. Laake. 1980. Estimation of density from line transect sampling of biological populations. Wildl. Mono. 72. 202pp. Laake, J. L., K. P. Burnham, and D. R. Anderson. 1979. User's manual for program transect. Utah State Univ. Press, Logan. 26pp. �81 Table 1. Groups of elk observed on 24 line transects flown 15 and 22 January, 1988, Middle Park, Colorado. Transect no. No. elk in group Perpendicular distance (yds) Transect side 7 83 12 10 175 350 0 330 right left right right 6 43 8 5 1 1 17 3 130 430 30 75 65 50 50 50 50 right right right left left right right left right 4 5 1 60 170 140 right right left 6 6 33 5 1 13 3 4 27 5 3 420 120 60 80 50 550 60 0 120 150 250 225 right right right left left left right right right right right right 13 450 left 6 52 325 325 1eft left 19 9 500 125 left right 1 2 2 2 3 4 5 5 6 6 6 6 6 6 7 8 9 9 9 8 10 11 12 13 13 14 14 14 14 15 15 15 16 17 18 19 20 20 21 22 23 24 3 �82 Table 2. Estimated densities and variances for numbers of elk groups based upon 5 mathematical models used for analyzing line transect data. Data was grouped and the initial cut-point was 55 yards. ------------------------------------------------------------------------------ Model and no. parameters Fourier Series Expo. Polynomial Expo. Power Series Neg. Exponential Half-normal 2 2 2 1 1 Density estimate Stand. error % Coef. variation 95% CI 0.8401 1. 1340 1.0690 1.0000 0.6366 0.2320 0.4264 0.8352 0.2934 0.1673 27.61 37.60 78.15 29.34 26.28 0.3601-1. 320 0.2519-2.016 -0.6593-2.797 0.3931-1.607 0.2905-0.9828 � https://cpw.cvlcollections.org/files/original/aeef9a070b35561f0bb9fe3005c6cbd4.pdf efea667834b35784aff5647fb5fd1d37 PDF Text Text Colorado Division of Wildlife Wildlife Research Report July 1988 83 JOB PROGRESS REPORT State of Colorado Project No. . Mammals W-153-R-2 ~~~~--------------- Research Work Plan No. 1A Multispecies Job No. 1 Animal and Pen Support Facilities for.Big Game Research Period Covered: July 1, 1987 - June 30, 1988 Author: Investigations P. H. Neil ABSTRACT Thus far this fiscal year, 8 pronghorn antelope fawns and 7 bighorn sheep lambs have been successfully reared and are being trained for food habits and other nutritional studies. Seasonal intake investigations with pronghorn antelope continued throughout the year as well as investigations of various stress levels in bighorn sheep. The entire east side of the Foothills Facility was rebuilt to include a new alleyway, weighing platform, and new fencing. One elk suffered rattlesnake bite during mid-June but has recovered. All animals at the facility are presently healthy with the exception of 4 bighorn lambs that are being treated for pneumonia-related problems. _" ��85 ANIMAL AND PEN SUPPORT FACILITIES FOR BIG GAME RESEARCH Paul H. Neil P. N. OBJECTIVES To provide and maintain populations of captive animals and pen facilities support Mammal and Avian Research programs. to SEGMENT OBJECTIVES 1. Maintain animal research facilities. 2. Coordinate all rearing, training, maintenance, under one support facility manager. 3. Integrate all animal and physical plant support facility manpower and monetary requirements under a single budget. 4. Maintain 11 elk, 1 Rocky Mountain goat, up to 30 pronghorn antelope, up to 30 mountain sheep, and 11 domestic cows in suitable health and responsiveness to perform required research experiments. and research activities METHODS AND MATERIALS Routine neonate rearing procedures were used to han~rear 8 pronghorn antelope .. this fiscal year. Eight bighorn sheep lambs were born at the facility and ieft on the ewes to raise. One died from pneumonia. Daily contact is made with the animals to enhance tameness and, in most cases, seems to be a satisfactory procedure for our needs. Animals are hand-reared on undiluted, canned, evaporated milk. Seven pronghorn antelope were born at the facility, two of which were stillborn. One pronghorn orphan was received at the facility, and two fawns were captured from the wild in northeastern Colorado. The entire alleyway complex on the east side of the facility was torn down and rebuilt due to aging and weather damage. Several other repair projects and maintenance -eontlnued during the year. RESULTS AND DISCUSSION Two pronghorn antelope were captured from the wild in northeastern Colorado and incorporated into the captive research herd this fiscal year. One orphan was received from eastern Colorado, and seven fawns were born at the facility from captive does. Two of the fawns born at the facility were stillborn. The eight remaining fawns are being hand-reared on canned, evaporated milk and trained for forage selection studies. Details and approaches to this study are discussed under Work Plan lA, Job 4 (Bruce Gill, Principal Investigator). �86 RESULTS AND DISCUSSION Two pronghorn antelope were captured from the wild in northeastern Colorado and incorporated into the captive research herd this fiscal year. One orphan was received from eastern Colorado and seven fawns were born at the facility from captive does. Two of the fawns born at the facility were stillborn. The eight remaining fawns are being hand-reared on canned, evaporated milk and trained for forage selection studies. Details and approaches to this study are discussed under Work Plan lA, Job 4 (Bruce Gill, Principal Investigator). Seasonal intake measurements were taken on adult pronghorn antelope. Daily intake and weekly animal weights were recorded throughout the year. Progress and results are reported under Work Plan lA, Job 4 (Bruce Gill, Principal Investigator). Research being conducted on the various levels of stress in bighorn sheep continued and has incorporated several approaches and designs. Progress of this study is explained under Work Plan 2A, Job 4 (Mike Miller and Tom Hobbs, Principal Investigators). The entire alleyway complex on the east side of the facility was torn down and rebuilt due to aging and weather damage. All animal shelters were also rebuilt and new fencing installed. In addition, a weighing platform was installed into the alleyway to simplify the weighing procedure for pronghorn antelope. Rebuilding of many of the shelters and parts of the alleyway on the west side of the facility has begun and will continue throughout the next fiscal year. All animals at the facility are presently healthy with·-the exception of four bighorn lambs, which are being treated for pneumonia-related problems. One elk suffered a rattlesnake bite during mid-June but recovered in a short period of time. The big game research herd presently consists of 11 elk, 23 bighorn sheep, 23 pronghorn antelope, 1 Rocky Mountain goat, and 11 domestic cows. Other activities at the Foothills Wildlife Research Station consisted of educational tours for wildlife personnel from various federal and state agencies, Colorado State University personnel, local Boy/Girl Scout troops, and Wildlife Commissioners. Prepare by """"I1";"':"C?..;;..,r-","~~~O.,aw.....a....;;...- ..=".";t,__]I.00::'....e-,=-'...::l£_' __ p~Nerr Wildlife Technician III ,~ �Colorado Division of Wildlife Wildlife Research Report July 1988 87 JOB PROGRESS REPORT State of Colorado Project No. W-153-R-2 ----------------------- Work Plan No. Job No. LA --~~--------------. 3 Period Covered: Author: Mammals Research Multispecies Investications Mammals Research 2 Administration July 1, 1987 - June 30, 1988 R. B. Gill Personnel: L. E. Lovett ABSTRACT All 7 objectives listed in the Segment Narrative for Mammals Research 2 for FY 1987-88 were met. ��89 MAMMALS RESEARCH 2 ADMINISTRATION R. Bruce Gill P. N. OBJECTIVE Administer research within the Mammals Research 2 Unit for the highest productivity at the lowest cost. SEGMENT OBJECTIVES 1. Require candidate research projects to go through the Division's research project selection process before they are considered for funding. 2. Require all research study plans to be peer reviewed, reviewed and approved by the Research Leader before research is initiated. 3. Require quarterly progress reports and annual progress reports of each research study. 4. Subject all research projects to periodic review according to the Division's formal administrative review process. 5. Allocate fiscal resources and track expenditures assure fiscal responsibility and accountability. 6. Maintain a manual on current standard administrative and distribute to each Wildlife Researcher. 7. Develop, implement, and maintain an effective information transfer process that assures research results are disseminated to appropriate information users in a form that maximizes the probability that the information will be used. of each research study to operating procedures RESULTS AND DISCUSSION Nine new projects were proposed for consideration in the Division's planning and resource allocation process. All would have required additional funding to initiate in FY 1988-89. None received funding, so further pl~in8 was . discontinued. These proposed studies were: a. Replication of the pronghorn-w1nter wheat damage study in the Craig area to test how general the results of the New Raymer study are. b. Develop a market demand study for pronghorn hunting which would measure demand for pronghorn hunting on private lands and estimate how much hunters would be willing to pay for that privilege. c. Study the impacts of pesticide control of Russian wheat aphid on pronghorn survival and reproduction. �90 d. Study the response of selected mountain sheep populations to controlled, nonlethal removals (for translocations) to simulate hunting removals. e. Develop a market demand study for mountain sheep hunting which would characterize demand by category of animal (any sheep, trophy rams, ewes, etc.) and estimate demand by animal category. f. Test a catch-per-unit-effort population size. g. Devise and conduct a management experiment to test the feasibility of managing urban beaver populations using nonlethal control strategies such as vasectomization, translocation, and habitat management. h. Devise and conduct a management experiment in urban wildlife habitat development and test real wildlife responses against predicted responses. i. Devise and conduct a management experiment in backyard wildlife landscaping using Division office complex properties as the sites for the experiment. Test real wildlife responses against predicted responses. j. Develop a cooperative project with Watchable Wildlife, Information and Education, and Research which would target research specifically for its commercial value as a salable video production. The idea would be to conduct research on a species of high public appeal, use the research project results as the basis for a script and for video footage, and produce a made-for-television video to be sold for profit which would then offset both production and research costs. approach to estimating mountain lion -. Objective 2. Draft Program Narratives were prepared for two research studies (MarkRecapture Estimates of Mountain Sheep Numbers and Development of River Otter Reintroduction Procedures) and processed through the peer review process. Both projects are now in the active research phase. Objective 3. Quarterly reports were received and reviewed for each Mammals Research'2 study active during FY 1987-88. Annual Reports were received and are submitted to the Federal Aid Planner in time to meet the Division's reporting obligations to the Fish and Wildlife Service. Objective 4. Every research study except the mountain lion research study was reviewed in the field at least once during FY 1987-88. �91 Objective 5. Following the Division's formal resource allocation process, final budget allocations were prepared and submitted to each Wildlife Researcher as their target spending limit for FY 1988-89. These budget allocations are prepared in standard Federal Aid Segment Narrative formats. During FY 1987-88, expenditures for each research study were monitored via the Mammals Research Section's DBASE budget tracking data base. Wildlife Researchers received quarterly statements of budget status until the beginning of the fourth quarter, whereafter they received monthly reports until the close of the fiscal year. By July 14, 1988, the FY 1987-88 resource allocation for Mammals Research 2 Project of $576,604.00 had been 99% expended. Objective 6. Each member of the Mammals Research Staff (both Mammals Research 1 and 2) have been issued a manual briefly describing Division administration standard operating procedures. These have been prepared in notebook form to facilitate continual updating. All manuals are current. Objective 7. During FY 1987-88 the Mammals 2 Research staff utilized a variety of methods to communicate information to Division users. Scientific and technical information was conveyed via publication in professional journals. Implications of scientific and technical information to Division management programs was conveyed through training schools, one-on-one contacts, advisory input to ongoing Division management programs (e.g. habitat management initiatives), regulation formulation and review, long range plan and operation plans, standard operating procedures, simulation models, and biologist cordination meetings. In addition, responsibilities of all Terres'tria1 Wildlife Research Leaders were expanded to include not only research planning and evaluation, but management planning and evaluation as well. ��Colorado Division of Wildlife Wildlife Research Report July 1988 93 JOB PROGRESS REPORT State of Colorado Project No. W-153-R-2 ----------------------No. LA --------------------- Work Plan Mammals Research Multispecies Investigations Job No. 4 Period Covered: July 1, 1987 - June 30, 1988 Author: Wild Ruminant Forage Selection Dynamics R. B. Gill and B. J. Maynard Personnel: M. W. Miller, M. A. Wild-Bowser, P. N. Lehner, and P. H. Neil ABSTRACT Birth weights, weekly weight gains, and milk intake are compared for pronghorn fawns raised in summer 1987 vs. summer lS2~, Birth weights were nearly identical for the two years. Fawns rear e.. :;': 1987 consumed slightly more milk per day but gained weight slightly less r~p~'ily than fawns reared in 1988. Rearing and training protocol is described and research agenda is outlined. ��95 WILD RUMINANT FORAGE SELECTION DYNAMICS R. B. Gill and B. J. Maynard P. N. OBJECTIVE To evaluate the role of color vision in diet selection foraging ruminant--the pronghorn. of a small, selectively SEGMENT OBJECTIVES 1. Rear and train up to six female fawns. 2. Develop procedures, technology, and sample size requirements to test the response of pronghorn diet selection to experimental manipulation of light wavelengths. METHODS Pronghorn rearing procedures were similar to those described by Neil et al. 1979. Pronghorn fawns were removed from the dam as soon after initial nursing as possible, usually within 12-18 hours to assure imprinting on the handler. Two handlers reared all fawns, and after approximately one month of age, a single handler fed all supplemental food. Hand-rearing protocol closely followed the protocol outlined by Brinkley 1987. Particularly, newborns were isolated for ~he first three weeks after removal from the dam to facilitate imprinting on the handler. A variety of food rewards was tested for use in operant conditioning training. Initially, all fawns w.ere trained to wear training masks to accustom them to headgear. The origInal experimental procedure was to have been to study nutritional quality of native forage diets while fawns were subjected to a number of experimental treatments denying them access to specific wavelengths of light. Light wavelength treatments were to be experimentally manipulated by fitting headgear with camera cutoff filters that filtered light of specific wavelengths. Training of pronghorn for this experimental approach was discontinued when my duties were expanded, and I could no longer afford to invest the required amounts of time necessary to train the animals. Prior to fawning in spring 1988, a contract was drawn between Division to assume this study. Eight new fawns were acquired. and training proceeded as follows: esu and the Fawn rearing Fawn Rearing Five female and three male pronghorn fawns were bottle-raised at the Foothills Wildlife Research Station. One male and four females were born at the station between June 9 and June 23, 1988. The remaining two males and one female were caught from the wild between June 6 and June 22, 1988. Captive-born fawns were removed from their mothers approximately 24 hours after birth, allowing time for sufficient colostrum intake. Wild fawns were caught when they were between two and seven days of age. �96 Upon arrival, fawns were weighed, vaccinated with a seven-way vaccine against Enterotoxemia and other clostridial diseases, and placed in individual isopens. Raising fawns in isolation discouraged imprinting on other fawns and encouraged imprinting on the handler. Fawns were fed canned, evaporated milk ad libitum from 24Q-ml baby bottles with standard baby bottle nipples. Holes in the nipples were enlarged for greater milk flow. Milk diets were supplemented once daily with 1 ml liquid baby vitamins without iron. Captive-born fawns also received a daily supplement of 1 ml CUS04 solution (100 mg/ml) in order to prevent copper deficiencies encountered in the previous year's captive-born and hand-reared fawns. A schedule of five feedings per day was maintained until all fawns were at least one week old. Four feedings per day were given for the next three weeks. Fawns currently are receiving three feedings pe~ day. In addition to milk, fawns have ad libitum access to alfalfa hay, calf-manna, and water. RESULTS AND DISCUSSION Fawn Rearing Fawns generally maintained good health. Problems that were encountered were successfully treated. Two captive-born fawns initially had low blood protein levels when they were removed from their dams. These two received supplemental bovine colostrum for their first six feedings, followed by a 50/50 mixture of colostrum and milk for the next two feedings, and then 75 milk/25 colostrum for one feeding. One of these two siblings also developed lameness in her right front leg as a result of an umbilical infection. She was placed on a schedule of 0.3 cc penicillin administered subcutaneously, twice dialy for one week. She has subsequently recovered. At three weeks of age, one wild-caught buck fawn developed severe diarrhea. His milk was replaced with Revive electrolyte solution containing a dissolved antibiotic (240 mg Sulfamethoxazole/Trimetoprim) administered twice daily for four consecutive days. Bismuth Subsalicylate (lO-ml doses) was administered periodically until feces began to solidify. After two-and-one-half days on Revive, the diarrhea subsided and the fawn was taken off the Revive treatment and fed evaporated milk again. Other fawns which developed occasional bouts of diarrhea improved after a few feedings, and less severe cases recovered untreated. Birth weights averaged 3.23 1987 and 2.77 1987 and 2.95 of kg kg kg captive-born fawns were similar for 1987 and 1988. Males in 1987 and 3.65 kg in 1988. Females averaged 2.88 kg in in 1988. Average overall birth weights were: 2.96 kg in in 1988 (Iable 1). Fawns gained weight more rapidly in 1988 than in 1987 (0.84 vs. 0.78 kg/wk). This occurred despite slightly greater milk intake in 1987 than in 1988 (570 vs. 564 ml/day)(Iable 2). The eight fawns reared in 1988 are to be used in an investigation of the role of pronghorn color vision in diet selection. Electroretinograms will be used �97 to measure retinal reception of varying wavelengths of light throughout the "visible" light spectrum and continuing into the near infrared light spectrum (700-750 nm). Perception of light wavelengths in this range will be tested in discrimination trials in which an animal will have to choose a correct wavelength stimulus in order to obtain a food reward. In preparation for discrimination trials, a Y-maze apparatus has been constructed in the fawn rearing pen. At each feeding the fawns are led through the maze and fed at the end of either branch. In an attempt to find a sufficiently attractive food stimulus, fawns have been offered rolled oats, raisins, strawberries, blueberries, apples, and milk-soaked Grape Nuts. Only rolled oats and raisins have been eaten by any of the fawns thus far. Licorice, marshmallows, and granola are among the food rewards still to be tested. LITERATURE CITED Brinkley, K. 1987. 14(8):225-237. Pronghorn hand-rearing protocol. Anim. Keepers' Forum Neil, P. H., T. N. Woodard, and D. L. Baker. 1979. Procedures for rearing wild ruminants in captivity. Colo. Div. Wildl. Game Info. Leaflet No. 106. 4pp. Prepared Wildlife Research Leader �98 Table 1. Mean birth weights of pronghorn fawns born at Foothills Facility. II of fawns 1987 Mean weight {kg} II of {SO} fawns 1988 Mean weight (kg) (SO) -----------------------------------------------------------------.-------------Male Female Overall mean 2 7 3.23 2.88 (0.10) (0.280) 1 4 3.65 2.77 (0) (0.436) 9 2.96 (0.482) 5 2.95 (O.546) Table 2. Mean intakes and weights of pronghorn fawns raised at Foothills Facil ity. ------------------------------------------------------------------------------1988 1987 Weight (kg) (SO) Intake (ml) (SO) Weight (kg) (SO) Intake (ml) (SO) ------------------------------------------------------------------------------Week Week Week Week Week 1 2 3 4 5 4.0 4.9 6.0 6.9 4.9 (O.61 ) (O.71) (0.76 ) (1.5) (1.8) 390 560 630 620 650 (91) (110) (90) (180) (120) 4.1 4.9 6.0 7.1 8.3 (O.59) (0.78) (0.75) (0.79) (0.86) 340 520 600 620 740 (52) (110) (74) (120) (120) ..--------------------------------------------------------------------.-~-~---,---.- �Colorado Division of Wildlife Wildlife Research Report July 1988 99 JOB PROGRESS REPORT State of Colorado Project No. Work Plan W-153-R-2 Mammals Research -----------------------No. 2A ---------------------- Mountain Sheep Investigations Job No. 4 Period Covered: July 1, 1987 - June 30, 1988 Author: Experiments to Identify and Manage Stress in Mountain Sheep S. A. Gottlieb, K. A. Trust ABSTRACT We used simple water extraction to measure fecal cortisol concentrations. Water extraction yielded about 1.3X more cortisol than ether/water extraction. Assay variation was also reduced using water extraction (cv = 4%). Extraction efficiency for the water method was about 33%. We demonstrated parallelism of radioimmunoassays (RIAs) for cortisol in plasma, urine, and feces from bighorn sheep; mean slopes of logit-transformed binding data for cortisol assays from each of these media did not differ (P > 0.05) from mean slope of a standard assay curve. Repositol adrenocorticotropic hormone elevated (P < 0.05) plasma cortisol concentrations (PCCs) in treated bighorns within 2 hrs posttreatment, and urine cortisol:creatinine ratios (UCCRs) showed similar trends 4-8 hrs after treatment. Correlation between PCC and UCCR was strongest (r2 = 0.72, P < 0.001) when corrected for lag time in urine cortisol excretion. Cortlsol:creatinine ratios from urine- stained snow (SCCR) provided repeatable (cv = 9.2%) estimates of UCCRs measured directly from voided urine (r2 = 0.87, P < 0.001), supporting use of urine cortisol measurements in monitoring wild bighorns. We were unable to cause weight loss or e£fect an increase in cortisol excretion in bighorns using a relatively severe malnutrition treatment (intake restricted to 35% of calculated metabolizable energy requirement) during summer. ��101 PLANE OF NUTRITION AND BIGHORN SHEEP POPULATION PERFORMANCE M. W. Miller N. T. Hobbs P. N. OBJECTIVE To treat bighorn sheep to control disease where necessary. SEGMENT OBJECTIVES 1. Estimate the effects of animal density and plane of nutrition on endocrine responses and immune competence in mountain sheep within + 10% of the true responses with 90% confidence. - 2. Assess the utility of urine and fecal samples to indicate environmental stress in mountain sheep populations. 3. Publish results in peer-reviewed, scientific journals. REFINING FECAL CORTISOL EXTRACTION TECHNIQUES Our previous attempts to quantify cortisol in feces of bighorns were plagued by excessive inter- and intra-assay variation with average cvs two to three times greater than those for urine or serum cortisol assays. After improving both feces preparation and radioimmunoassay (RIA) performance with little overall reduction in this variability, the extraction technique itself appeared the most likely source of error. In particular, volumes of solvent used in the original protocol seemed excessive and may have rendered the extraction procedure relatively inefficient. Because evolution of the accepted extraction protocol is unclear (and apparently undocumented), we designed two experiments to examine alternatives for recovering cortisol from feces. EXTRACTION METHODS FOR FECAL CORTISOL DETERMINATION Methods and Materials We compared three recovery methods using £our~ecal pools from a single tame bighorn (A82) before (12.1, 12.2) and after (14.1, 14.2) imposing alternateday adrenocorticotropic hormone (ACTH) administration. Each pool sample had previously been lyophilized and ground in a Wil~y mill then mixed thoroughly. Nine 0.5-gm subsamples of each pool were placed in glass test tubes. For each pool, we randomly assigned three aliquots to each of three treatments (n - 12 samples/treatment): -5.0 ml deionized water was added to 0.5 gm feces. Samples were agitated for 1.0 hr, then centrifuged at 3000 RPM for 15 mins. All supernatant (x - 2.8 ml) was collected and fro~en until submitted for RIA. �102 -Samples were treated as above with diethyl ether substituted for deionized water. Ether supernatant was dried under nitrogen and residue rehydrated at 2.0 ml PBS gel, then frozen until assayed. -0.5 gm feces was rehydrated with 5.0 ml deionized water. After adding 5.0 ml diethyl ether, samples were agitated for 1.0 hr, then frozen (-200C) for 1 hr. Ether extracts were poured into clean glass tubes, and the process was repeated with an additional 5.0 ml diethyl ether. Combined extracts from each aliquot were dried under nitrogen, and residues rehydrated in 2.0 ml PBS gel, then frozen until assayed. Cortisol concentrations were determined by RIA (extracted "bench" method, previously described). We converted results from ng/ml to ng/gDM feces; coversion factors were about 10.0 for water and 4.0 for ether and combined methods. Results and Discussion Extractions using ether alone recovered virtually no cortisol (pg quantities instead of ng), and data were not analyzed further. The other two methods produced comparable results (Table 1). Water alone consistently recovered more cortisol than the combined extraction; estimated fecal cortisol concentrations from water extracts averaged about 1.3X estimates from ether/water extracts. Variation among replicates was substantially lower for water extraction (cv - 4% for water vs. cv -18% for ether/water). Both methods showed similar trends in cortisol excretion, with fecal cortisol concentrations rising more than two-fold in response to ACTH injection (sample 14.2). The magnitude of change in fecal cortisol concentration was less than was reflected by urine cortisol:creatinine ratios (Table 1) from samples collected during the same periods. Differences in urine and fecal responses may be an artifact of sampling periods used or may indicate greater integration of plasma cortisol changes in fecal cortisol excretion. Our results suggest that simple water extraction provides repeatable but possibly biased estimates of fecal cortisol concentrations. EFFICIENCY OF FECAL CORTISOL EXTRACTION METHODS Methods and Materials Based on results of previously described protocol comparisons, we used radioisotope techniques to estimate extraction efficiencies of water and ether/water protocols for fecal cortisol determinations. Ten 0.5-gm aliquots were t... __ en from each of four lyophilized and ground fecal pools; these were randomly ~~Bigned to one of two extraction protocols (five replicates per pool X protocol). We rehydrated all aliquots with 2.0 m1 deionized water containing 20 ~l 3H-cortisol colution (about 152 CPM/~l) and uncubated them overnight at 4oC. The next day we added 3.0 ml deionized water to each aliquot. We performed respective extraction methods as described previously, except that ether/water residues were rehydrated with 3.0 ml deionized water. We placed 1.0 ml of aqueous extract from each aliquot into 10.0 ml Toluene-Triton X (TTX) scintillation liquid, agitated vials, and allowed these mixtures to extract overnight. In addition to extracted samples, we included two total count controls (20 ~ 3H-cortisol in 10.0 ml TTX) and two color quench �103 controls for the water extractions (20 ~l 3H-cortisol and 1.0 supernatant in 10.0 m1 TTX). The following day we ran 10-min extracts using a liquid scintillation counter. We calculated extraction efficiencies (%) using scintillation data corected and background. m1 fecal counts of TTX average for quenching Results and Discussion Virtually no 3H-cortisol was recovered from feces using ether/water extraction; counts from ether/water extracts did not differ from background (about 35 CPM). Counts from water extracts were slightly higher (x - 50.05 CPM). However, supernatants from water extraction contained pigmented that caused heavy color quenching (about 91.8%) in quenched controls. After correcting for quenching effects, efficiency of water extraction was about 33.25%. Estimated precision of water extract results (based on cvs from replicates) was similar to previous observations (cv - 4.8%). These data, combined with results previously described, suggest that simple water extraction is preferable to ether/water for recovering cortisol from feces. Water extraction should provide a precise, although somewhat biased, estimate of fecal cortisol concentrations in bighorns. VALIDATION OF RIA FOR MEASURING CORTISOL IN EXCRETA Two approaches are commonly used to validate RIA methods: parallelism is one indicator of assay specificity, quantitiativ~ recovery is a measure of assay sensitivity. We began by comparing average slope of standard assay curves with those of plasma, urine, and fecal cortisol assays from bighorns. A valid RIA for these experimental media should produce slopes that do not differ from slope of the standard curve (Le. the sample and standard curves are parallel). Quantitative recovery demonstrates the assay's ability to measure additions of cortisol to experimental media; these determinations are in progress. PARALLELISM OF PLASMA, URINE, AND FECAL CORTISOL RIAs IN BIGHORNS Methods and Materials We examined para1le~ism between standard curves for cortisol RIA and curves generated by diluti~g pl~sma, urine, and fecal supernatants (water extraction previously described) from bighorns. A standard curve was created using 100 ~ of stock cortisol solution (100 ng/ml) in an 8-step, 2.5-fold dilution series; four replicate series were constructed. For plasma, urine, and fecal supernatants we created similar curves using 200 ~ of samples in 5-step, 2-fold dilution series; three replicate series were constructed for each sample medium, using a different sample for each replicate. Aliquots from standard, urine, plasma, and fecal series were extracted and assayed using standard "bench" cortisol RIA procedures. In addition, unextracted fecal supernatants were assayed. �104 We plotted logit transformations of percent buffer control bound (y-axis) against natural log of sample amount assayed. Regressional analyses generated "average" standard and sample assay curves. We compared slopes of sample curves to slope of the standard curve using a small-sample t-test for parallelism. Results and Discussion Average slopes from logit-transformed cortisol assays of bighorn plasma (Sp - -0.938)(Fig. 1), urine (Bu - -0.870)(Fig. 2), and extracted (Be -0.8l7)(Fig. 3) or unextracted (Bf - -O.763)(Fig. 4) fecal supernatant did not differ (p > 0.05) from average slope of the standard curve (Bs - -0.848). Performance of urine and extracted fecal supernatant assays most closely approximated the standard curve. These data support use of any of these assays in determining cortisol concentrations, and allow comparison of results derived from these methods. Some confoundment still exists in interpreting results from fecal cortisol assays. Cortisol concentrations from unextracted aqueous supernatant are about 2.5 times greater than from extracts of aqueous samples; moreover, cvs for unextracted assays are invariably smaller. We suspect that water-soluble cortisol metabolites are being detected in the unextracted (i.e. aqueous) assay, contrary to more restrictive antibody specificity claims for this RIA. Alternatively, cortisol may have greater solubility in aqueous solutions than has been previously recognized. In either case, all results to date suggest that unextracted assays of aqueous fecal supernatants recover more cortisol than extracted assays, and provide more precise estimates of fecal cortisol concentrations. CORRELATION OF CIRCULATING AND EXCRETED CORTISOL LEVELS IN BIGHORNS Methods and Materials Experiments to date have demonstrated that urine and fecal cortisol levels can be used to detect adrenal responses to exogenous ACTH and mild stressors in bighorns under laboratory conditions. However, these techniques must be adaptable to field conditions and their inherent sampling limitations if they are to be used as management tools. Here, we conducted an experiment to define temporal relationships between plasma, urine, and fecal cortisol concentrations and further ~~amine utility of urine and feces for remote sampling. Eight bighorns were randomly assigned to one of four ACTH treatments: 0 (control), 0.25, 0.50, or 1.0 U/kgj we used repositol ACTH to provide a sustained cortisol response. Sheep were housed in digestion cages (5 m2) for about 24 brs prior to injection. We collected blood, urine, and feces at 0, 2, 4, 8, 12, and 24 hrs after ACTH administration. Urine-stained fecal pellets were discarded. Blood was drawn with little or no restraint, usually in less thanl min « 2.0 mins in all cases), and we assume plasma corticoid levels to be at or near baseline in controls and at time O. We rotated bleeding order to equalize cortisol responses attributable to anticipation of bleeding, and bleeding of all animals was completed in 15-20 mins. We �105 centrifuged blood samples within 30 mins of collection and harvested plasma. Urine and feces from each collection were weighed, and urine samples were filtered. All samples were stored at -20oC between processing steps. Feces were prepared as previously described and extracted using 5.0 ml of water. We measured cortisol using RIA; urine creatinine was measured by a colorimetric method. We used analysis of variance to assess treatment effects and dose-dependence, as well as regressional analyses to examine relationships between circulating and excreted cortisol levels. For the latter, we also adjusted data to account for lag times in cortisol excretion, and again compared these to circulated levels. Results and Discussion Plasma cortisol concentrations elevated (P < 0.025) within 2 hrs after repositol ACTH treatment and remained elevated for at least 8 hrs (Fig. 5). A similar trend was seen in urine cortisol excretion, although peak response was delayed 2-6 hrs (Fig. 6). Fecal cortisol analyses are in progress. Urine cortisol:creatinine ratios (UCCR) from samples collected in the same period as plasma loosely correlated with plasma cortisol concentrations (PCC) (r2 - 0.46, P < O.Ol)(Fig. 7). We observed a better relationship (r2 0.72, P < 0.001) between UCCR and PCC by comparing PCCs with UCCRs from the succeeding sampling period (i.e. 0 hr PCCs with 2-hr UCCRs) (Fig. 8). Our results are consistent with other observations on the integrative nature of urine cortisol excretion as compared to more dynamic and instantaneous plasma cortisol levels. These data underscore the utility of urine cortisol as a potential estimator of chronic adrenal responses to environmental stressors. UTILITY OF URINE-STAINED SNOW FOR ESTIMATING URINE CORTISOL:CREATININE RATIOS Methods and Materials Urine-stained snow offers the most viable approach to field sampling of urine cortisol. To examine reliability of cortisol:creatinine ratios from snow to estimate CCRs from urine, we conducted a snow recovery experiment. We collected urine from ACTH-treated bighorns in conjunction with the previous experiment. Aliquots (about 65 ml) from each collection were warmed to about 380C in a water bath. After removing a 4-ml subsample, aliquots were poured into 17 em of fresh, lightly packed snow. In ~epositing urine into snow, we tried to si;ulate urination patterns of ~dult bighorn ewes. Urine-stained snow was coilected using empty l2-cc syringe casings pushed downward into the snow. Snow packed into the flared opening of the casing, and a relatively consistent amount was recovered with each sampling. We did not collect snow randomly from urine deposits, but instead attempted to collect from areas of heaviest deposition, as would be done in field sampling. Snow samples were thawed for 10-15 mins, mixed thoroughly, and 4-m1 sub samples retained for assay. In most cases, melting yielded 5-10 ml of liquid, but volumes ranged from about 2.5 to about 15 ml. To determine variability of this sampling method, each of 10 l80-ml aliquots were divided evenly among 5 deposit sites and sampled as described above. �106 We measured cortisol in urine and snow using an extracted RIA. Creatinine was determined by a colorimetric method using aliquots of snow diluted 1:20 and 1:5, respectively. We expressed results as cortisol:creatinine ratios (ng:mg), and compared urine and snow CCRs using regressional analyses. Results and Discussion Snow CCRs provided relatively consistent estimates of UCCRs from voided samples, although we detected a tendency for underestimation of UCCRs by SCCRs (r2 - 0.87, f < O.OOl)(Fig. 9). Our sampling procedure and cortisol assay produced minimal variability (cv - 9.2%) among replicate SCCR estimates. From these results, we conclude that urine cortisol excretion can be estimated using CCRs obtained from urine-stained snow. It follows that our techniques should be readily applied to field situations where adrenal responses of bighorns are being measured. ADRENAL RESPONSE TO CHRONIC MALNUTRITON IN BIGHORNS Methods and Materials During June-August, 1987, we attempted an experiment evaluating adrenal responses of bighorn to chronic malnutrition. We used 5 pairs of sheep and imposed treatment using a low quality ration; controls received calculated maintenance allotments of "high energy" wafers, and treatment animals were provided the same amount (by weight) of a "low energy" wafer with 70% DE of the high quality diet. We weighed bighorns weekly and collected urine and feces for 24 hrs each week immediately prior to weighing. After 2 wks of treatment, no weight loss had occurred in the treatment group. We halved the amount of low quality feed provided to treatment sheep (effectively providing -35% of calculated maintenance requirements), and continued for another week. With no weight loss measured after 1 wk of what we judged to be a severe malnutrition treatment, we decided to abandon the experiment. Results and Discussion We were unable to cause weight loss in captive bighorns using the malnutrition treatment described. Cortisol determinations on two pairs of bighorn showed no differences between control and "malnourished" bighorns; in fact, cortisol was undectable in a majority of the samples run. Failure to impose a treatment effect could be attributed to several factors. Water consumption may have increased as feed supply diminished and would account for unchanging weights; excessive urine volumes, as compared to winter collections, support this explanation. High solubility and subsequent digestibility of the low energy pellet may have reduced severity of treatment. In addition, it may be that bighorns are as efficient as their domestic counterparts in extracting energy from whatever foodstuffs are available, particularly during warmer months when energetic costs are low in nonlactating individuals. From our results, it appears difficult to nutritionally stress nonlactating bighorns during the summer, at least under our experimental conditions. This �107 leads to some question about the relative importance of malnutrition in contributing to autumn outbreaks of pneumonia in wild bighorns as compared to other unidentified factors. Table 1. Comparison of two extraction methods for estimating fecal cortisol concentrations (ng/gDM) and corresponding urine cortiso1:creatinine ratios (ng:mg). ------------------------------------------------------------------------------Water Ether/Water Urine cortiso1:creatinine (ng:mg) tv (%) x (5E) cv (%) -------------------------------------------------------------------------------- x Sample 12. 1 12.2 14.1 14.2 Prepared (5E) 11.25 (O.05 ) 12.57 (0.40) 15.58 (0.55) 36.63 (1. 07) by: 1.0 5.0 6.0 5.0 10.13 (0.92) 8.40 (1.27) 13.33 (0.67) 24.67 (2.75) ~~a~~~, ~/J~%~/~~' j~,:~,,>__ ~hael w. Mi'1ler . ~?~. 16.0 27.0 9.0 19.0 Research Associate N. Thompson Hobbs Wildlife Researcher 14.0 17.0 2.0 93.0 �108 PLASMA 4 • 3 ..J 0 -,'. ". ~"'. ~",'.-,~'" "" ....• ". -,<::::::::-. 2 ~ 0 0 1 H: 0 s => r:c -1 LL 0 1= -2 o 9 -3 -4 -2 -1 1 0 2 3 4 5 6 LOG OF uL ADDED Figure 1. Parallelism of cortisol assay for bighorn plasma. Open and closed dots represent individual observations for sample and standard curves, respectively. URINE 4 ..J 3 ~ 2 8 1 o t=> 0 r:c -1 LL o I:: -2 o 9-3 -4~ __~ __~ __~ __~ __+-__~~ -2 -1 0 1 2 3 4 __~_ 5 6 LOG OF uL ADDED Figure 2. Parallelism of cortisol assay for bighorn urine. Open and closed dots represent individual observations for sample and standard curves, respectively. �109 EXTRACTED FECES 4 o 3 ~ 2 8 1 ...J >,.:".,...,.....".. <,<, •..•••• ~........ <:::~.... •...:.•... ,: .• , ·..·········'O···::·.·.··&···-,Q .•....., -, •..~... ••.... ~ 0 :J - ..•..•.• 0···... 0..-., rc -1 - .:::::: L1. o t:: -2 (!) 9 -3 -, .,<., .... ' 0·<. . .. 0 -, ". -: -4~ __~ __~ __~ __~ __~ __~ __~ __~ -2 -1 0 1 2 3 4 5 6 LOG OF uL ADDED Figure 3. Parallelism of extracted cortisol assay for bighorn feces. Open and closed dots represent individual observations for sample and standard curves, respectively. UNEXTRACTED FECES 4 ...J 0 ~ t .... - 3 ' ..-, ,.... •. ...•.•. .... 2 8 :J "<, -, .. 6'..' 1 0 rc -1 L1. 0 t:: -2 9-3 -4 -2 -1 0 1 2 3 4 5 6 LOG OF uL ADDED Figure 4. Parallelism of unextracted cortisol assay for bighorn feces. Open and closed dots represent individual observations for sample and standard curves, respectively. �110 PLASMA 90 ~, I , I I I I ~ .•••.••• .••• .•..• \ , fr· \ Ii <, \ .\\ A.. \\ f' iQ-EI " Ii: \. It Lj Ii \ \ I} \ \\ \~ \ \ '. \ \ di 'tt \ •.._ \ -....__ -_ •.•. ". -,-,~ O~r- ,- o 10 ·tI EJ ~ 20 ~ 30 HOURS POSTIREATMENT ACTli DOSE (units/kg) e-e-e 0.00 •.•..• 0.25 •...••.•••1.00 0'0'0 0.50 Figure 5. Changes in plasma cortisol concentrations in bighorns responding to repositol ACTH administration. URINE 400 'Si .€ 9 1\ 1 \ ~300 / \\ W Z Z ell ,i\ ~200 a: o " /: I: • I I I y..... t111.i» ..J ~100 8 1 I I I I \ \ \ \ \ \ ~ ,;"', t:!., "" " "'~'" '. II I II I I. \ ".\A , \ " \.\ ~~e~ --~" ~ __.ij\ _-" ,,~ J' =~ O~r---------~--------~--------~10 20 30 o HOURS POSITREATMENT ACTli DOSE (unitsIkg) ~ ._ 0.00 0.50 .,.0-0 _ 0.25 1.00 Figure 6. Changes in urine cortisol:creatinine ratios in bighorns responding to repositol ACTH administration. Note lag time of peak responses (compare to Fig. 5). �111 o ... .. o o o ~200 § ~ 100 § CO o o (>._/.-/ -»/ ...... .r: _ r:? •• CJ·· ..... »>" ••••••• _. 0 .., , , ~.- .. »>: ••_ O.,/~······~~.,···~··o 00' ..".. 0 o o o 10 20 30 40 50 60 70 80 90 100 110 PLAMSA CORTlSOL CONCENTRATION (nglmQ Figure 7. Correlation of pees with UeeRs from samples collected during the same sampling period. (r2 - 0.46, p < 0.01). HOURS + 500 o o ..' o ._. __ ., '3-//-0 - •. ... o· o 10 20 30 40 SO 60 70 80 90 100 110 PLASMA CORTISOL CONCENTRATION Figure 8. collected 0.72, P < secretion (nglmQ Correlation of pees with UeeRs from sam~les during the succeeding sampling period (r 0.001). Correcting for lag in urine cortisol improved correlation (compare to Fig. 7). �112 400 I -300 w I.. a: U200 5 en E 8100 ~ z en 0 0 100 200 300 URINE CORTlSOL : CREATlNINE 400 SOO (nWmg) Figure 9. Correlation of CCRs estimated from urinestained snow with UCCRs from voided samples. (r2 - 0.87, P < 0.001). �Colorado Division of Wildlife Wildlife Research Report July 1988 113 JOB PROGRESS REPORT State of Colorado Project No. W-153-R-2 ------------------------ Work Plan No. __.....;;;2;;,;;A _ Mammals Research Mountain Sheep Investigations Job No. 5 Period Covered: July 1, 1987 - June 30, 1988 Author: Tests of the Mark-Recapture Method to Estimate Mountain Sheep Numbers D. F. Reed ABSTRACT Trickle Mountain located in the Southwest Region was selected as the study area. Baiting was initiated in December, 1987, and the first drop was made at the site northwest of the Dabney Ranch 7 January 1988. All yearling and adult females in this drop were allocated to the SE Region for translocation, hence no telemetry collars were placed on sheep during this occasion. The second dropnet site was then baited, and a drop was made on 1 February 1988 where 19 yearling or adult females were trapped and radio collared. Two of these were subsequently lost to possible latent capture myopathy. Thereafter sheep response to bait was tentative, and plans were made for alternative capture methods. Three helicopter flights were made combining netgun and capture gun/chemical immobilization methods. During these flights, 11 additional sheep were captured and radio collared bringing the total of radio-collared sheep to 28. Additional simulations were run, and it was decided by the coprincipal investigators and graduate research assistant that number of helicopter flights would be increased to 15 versus capturing the additional 12 animals and possibly influencing their behavior and/or distribution. ��115 TESTS OF THE MARK-RECAPTURE METHOD TO ESTIMATE MOUNTAIN SHEEP NUMBERS Dale F. Reed P. N. OBJECTIVE Test the fundamental assumptions mountain sheep numbers. of a mark-recapture procedure in estimating SEGMENT OBJECTIVES 1. Select study area. 2. Quantify the extent to which mark-recapture in mountain sheep population estimates. 3. Test the robustness data. of those assumptions model assumptions via simulations are violated with actual field ACKNOWLEDGMENTS I thank co-principal investigators R. B. Gill and G. C. White, graduate assistant A. K. Neal, J. H. Olterman, J. Cuttingham, J. H. Gould, and B. J. Poole for their continued concern and/or efforts to capture and radio collar sheep. Division personnel G. C. Wetherill, M. Cousins, E. Dumph, G. A. Hinshaw, J. Johnson, D. W. Kenvin, Jr., J. T. Rauch, R. L. Weldon; USFS personnel S. Sylva, J. Jaminet; and volunteers D. Payne, J. Colman, P. White, L. Schmidt, W. Nye, A. Nye, K. Nye, and L. Nye assisted during trapping with the dropnet. M. W. Miller did the capture gun/chemical immobiliztion work, and R. Dick piloted the Soloy helicopter. DESCRIPTION OF AREA The Trickle Mountain study area is located in south-central Colorado approximately 20 km west of Saguache. The study area includes the Trickle Mountain sheep winter range as reported by Shepherd (1975:4). Areas adjacent to this winter r%nge may be included if radio-collared sheep are located in such areas during c.he 1989 winter season wten replicate helicopter counts are planned. Most of the lands within the winter range are administered by the BLM. BLM lands are characterized by mesas topped by rolling hills and steeply dissected by drainages. Volcanic rock outcrops are associated with mesa edges as "rimrock" cliffs and ledges and with some ridge tops. Open grassy slopes are common on the mesa tops and lower slopes and hilltops. The creek bottoms are mostly narrow with flat floodplains bordered by talus slopes climbing steeply to mesa rimrocks. Surface rocks of the study area are predominantly Tertiary volcanic rocks overlying Precambrian crystalline and sedimentary rocks (Shepherd 1975). Vegetation of the winter, intermediate, and lambing ranges has been described (Shepherd 1975:20, 36, and 54, respectively). �METHODS AND MATERIALS Three methods were used to capture mountain sheep: dropnet, netgun (Coda Enterprises, Inc., 1038 E. Norwood, Mesa, AZ 85203), and capture gun using darts containing carfentanil. The dropnet used at the two Trickle Mountain sites was approximately 27 m2• These sites were baited daily to habituate the sheep to the site, vehicle, and personnel and, eventually, to the dropnet, which was typically put in place after the sheep were "hooked" on the fermented apple pulp. Once 20-30+ sheep regularly came "under the dropnet, plans were made to assemble a similar or greater number of personnel on a given day when the net would be dropped. On that day, every effort was made to keep conditions the same as before so that the sheep coming under the net would not be affected. Once the net was dropped, personnel untangled and stabilized animals in the net as soon as possible. The animals were treated for lungworm, radio collared, and released. The netgun and capture gun/chemical immobilization techniques were utilized during helicopter flights. The netgun was used only during the first helicopter flight, whereas use of both net and chemical immobilization was incorporated into the subsequent flights. First, a search was conducted with the helicopter to locate sheep, then preferably to haze a small group or individual into the open and shoot the net over one selected animal (usually at a full run). If instead of running the sheep sought cover and "held" under pinyon (Pinus edulis) - juniper (Juniperus spp.), the capture gun option was chosen; and 1-2 animals were darted from either 1) the hovering helicopter, or 2) the ground while the helicopter hovered overhead keeping the sheep in cover and "holding." In the former case, the helicopter was landed as soon as possible, and the netted animal collared and released quickly (occasionally within 3-4 mins). In the latter case, the helicopter was "backed off" and the darted animal observed until the drug had affected immobilization (usually 4-6 mins), then the helicopter was landed. The animal was then collared, given the chemical releaser, and released as mobility was regained. In addition to trapping and radio collaring sheep, 3 fixed-wing flights were conducted to check on mortality and location of the animals in relation to the study area. RESULTS AND DISCUSSION Trickle Mountain was selected as the study area. Trapping operations included capturing sheep for a transplant that had been planned by the SW Region. Baiting was in"'.;;'iated in December, 1987, and the first drop was made at the site northwest of the Dabney Ranch 7 January 1988 (Table 1). All yearl':'ngand adult females in this drop were needed for the transplant, hence no te~ewetLy collars were placed on sheep during this occasion. The second dropnet site located northeast of the Dabney Ranch was then baited, and an attempt was made on 28 January 1988 to drop the net. However, an insufficient number of sheep came under the net (some appearing tentative and leaving before others arrived), and the drop was planned for the following week on 1 February 1988. On 1 February 1988 the drop was made, and 19 yearling and adult females were trapped and radio collared (Table 1). Subsequently, two of these sheep, Black M (Channel 10, 173.237 MHz) and Black NI (Channel 11, 173.262 MHz), died. It was estimated, based on the position of the carcass in one and condition of the striated muscle tissue in the other, that these deaths resulted from capture myopathy. �117 Baiting was continued throughout February and into early March with sheep response being both tentative and sporadic. Alternative capturing methods were planned using the netgun during the first of 3 helicopter flights and either the netgun or capture gun/chemical immobilization during the second and third helicopter flights. During these flights on 17 March, 23 March, and 3 April, 4 sheep were captured with the netgun and during the flights on 23 March and 3 April, 7 sheep were captured with the capture gun, bringing the total number of radio-collared sheep to 28 (Table 1). Although sample sizes of captured animals were small, the comparison of dropnet and netgun results appear to be consistent with those reported (Koch et ale 1983). Nineteen of the 28 telemetry collars had mortality sensors (Telonics, 932 E. Impala Ave., Mesa, AZ 85204-6699) hence, the mortality status of each animal was readily determined. For the remaining 9 collars without mortality sensors, changes in location were used in estimating no deaths by the end of the segment. In addition to the 3 helicopter flights, 3 fixed-wing flights were made with SW Region Biologist Jim Olterman on 5 May, 7 June, and 22 Jun 1988. No mortality signals were received, and all but Red 5 (Channel 15, 172.637 MHz) (Table 2) had been located by the last flight (Table 3). Preliminary information on distribution was determined during these flights (Figs. 1 and 2). Movements of individual animals will be analyzed after additional data points are collected. The principal program narrative for Work Plan 2A, Job 5 was reported previously (Reed 1987:317-321). The program narrative for Work Plan 2A, Job 5 (supplement) by graduate research assistant A. K. Neal titled, "Computer simulations to statistically evaluate the mark-recapture method with mountain sheep" is included as Appendix A. LITERATURE CITED Kock, M. D., D. A. Jessup, R. K. Clark, C. E. Franti, and R. A. Weaver. 1987. Capture methods in five subspecies of free-ranging bighorn sheep: an evaluation of drop-net, drive-net, chemical immobilization and the net-gun. J. Wildl. Diseases 23:634-640. Reed, D. F. 1987. Improve matching harvest regulations with available mountain goat and sheep populations and improve methods for obtaining better popula~ion estimates. Colo. Div. of Wildl. Game Res. Rep. July Part 2:291-321. Shepherd, H. R. 1975. Vegetation of two dissimilar bighorn sheep ranges in Colorado. Colo. Div. of Wildlife, Div. Rep. No.4. 223pp. c=------t_J to Ii Prepared by Da e F. Reed Wildlife Researcher �I-' I-' 00 Table 1. Mountain sheep trapping results at Trickle Mountain during 1988. Number Capture date 7 1 17 23 Location NW NE N N Jan Feb Mar Mar 3 Apr Dabney Dabney Dabney Dabney Ranch Ranch Ranch Ranch N Dabney Ranch and N Guard Station --1 Method Dropnet Dropnet Netgun3 Netgun Capture (n-3) Netgun Capture (n-4) ~s collared for study Mortality 201 0 0 0 19 2 Unk 22 0 0 17 19 4 0 4 0 23 5 0 5 0 ------------ 28 Trapped -35 45+ 2 Transplanted Collars active (n-l)/ gun4 (n-l)/ gun ----------------------------------- Included 10 yearling and adult females. 2B1ack M (Channel 10, 173.237 MHz) and black NI (Channel 11, 173.262 MHz) deaths estimated due to latent capture myopathy. 3 Netgun was used from helicopter (n-4). 4Capture gun was used from helicopter (n-4) or from ground with helicopter hovering overhead (n-3); chemical immobilizer was carfentanil used with naloxone releaser. �119 Table 2. Number assigned to animal, collar description, channel, frequency, pulse, activation date, and status for telemetered mountain sheep in the Trickle Mountain area. ---------------------------------------------Activation Pulse per Animal Collar Frequency no. description Channel (MHz) min. (ppm) Status date I 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 -- Blue white Blue orange Blue square Black A Black E Black F Black H Black J Black K Black M Black NI Black P Black R Black S Black T Black V Black ~ Black X Black Y Black Z Black NIl Black triangle Red dot Red 1 Red 2 Red 3 Red 4 Red 5 Red 6 Red 7 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 00 11 09c 10 11 12 13 14 15 16 17 172.787 172.812 172.862 172.912 172.962 173.012 173.062 173.112 173.162 173.237 173.262 172.837 172.887 172.937 172.987 173.037 173.087 173.137 173.187 173.212 173.237 172.487 172.512 172.537 172 .562 172.587 172.612 172.637 172.662 172.687 60-110b 60-110 60-110 60-110 60-110 60-110 60-110 60-110 60-110 60-110 60-110 60-110 60-110 60-110 60-110 60-110 60-110 60-110 60-110 60-110 60-110 60 60 60 60 60 60 60 60 60 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 17 3 17 23 23 23 23 '3 3 3 3 Feb Feb Feb Feb Feb Feb Feb Feb Feb Feb Feb Feb Feb Feb Feb Feb Feb Feb Feb Mar Apr Mar Mar Mar Mar Mar Apr Apr Apr Apr 88 88 88 88 88 88 88 88 88 88 88 88 88 88 88 88 88 88 88 88 88 88 88 88 88 88 88 88 88 88 2A 3A 3A 4A 3A 3A 2A 3A 4A 3D 4D 6-12A 6A 4A 5A 2A 4A 4A 3A 4A 6A 2A 3A 5A 3A 2A 6A IV lA 4A aNo. - estimated age, A - active (collar transmitting), D - animal died, and U - unknown. bMortality sensors: 60 ~pm a alive, 110 ppm - dead (no movement for period of 4 hrs). Sheep 22730 ~~ not have mortality sensors. c Second series of numbers is for a second Te10nics TR 1 receiver. �120 Table 3. Flight date, aircraft type, number of mountain sheep observed, number of mountain sheep with transmitters observed, and number of transmitters located and not located during 1988. -------------------------------Number mountain .: Flight date 17 23 3 5 Mar Mar Apr May 7 Jun 22 Jun Aircraft type Soloy helicopter Soloy helicopter Soloy helicopter Cessna 185 Cessna 185 Cessna 185 Observed Observed wI transmitters 39 41 46 5 20-30 39 --------.---- sheep Transmitters located Transmitters not located 14 12 13 1 1 5 4 16 27 24 12 1 �4235~----------------------------------------~ SpanIsh Creek Rd . 4231 ~ !s 1:: J 3l·< .., ~ North ' • Pass ". 4227 ,.._,-. o i :u a. ~.•. '?-Pass Buffalo ~ Campground Ben ny ~ TrlckJe <::l 4223 Guard station -----. ~~ 421 9 I , . 358 I I I I iii 362 r 366 i r i r 370 iii '.\ Mountain A. ....._ .•. .. iiiii;;;,;__._./ 771 iii 374 i J r 378 I I Houghland t:::, , 382 I i I •.•.•••• HiD i I I 386 i I I 390 UTM east/west km . Fig. 1. Location of 16 radiocollared mountain sheep (+) the 7 June 1988 fixed-wing flight (Map by A. K. Neal). as determined from •... N •... �I-' N N 4235,.-----------------------------------------~ Spanish Creek Rd . ~ 3 ..~r .. 4231 l .. I North , ••,. Pass +- \ ,,:t;. .:lJ Q, .. ...•..•. ""'- ~ ~ ~ f Q <- 4227 Buffalo Pass ~ Campground Benny 8 ~Z(_/.' .,..'..a.. .. ~ Trickle <:) ~CKnltaln 4223 ~ + + . I ',--r--r . . 4219 358 362 i I I ..::-;-...,.. / IifJhwp' -._. ••~ •.•...._•• -..._ I I + qjl I I I I 366 I iii 370 i I 374 I I I 111 i 378 I HOlJghland ."" •• I ~Hill i 382 f iii i 386 UTM east/west km . Fig. 2. Location of 27 radiocollared mountain sheep (+) the 22 June 1988 fixed-wing flight (Map by A. K. Neal). ..4 +0+ \ ..•.. ...-\ • • + ., ... •••. • as determined from , I I 390 I �123 APPENDIX A PROGRAM NARRATIVE State: Colorado Project No.: W-125-R-2 Mammals Research Work Plan No.: _..2::..:.A.:._ Job No.: _ Mountain Sheep Investigations 5 (supplement) Computer Simulations to Statistically Evaluate the Mark-Recapture Method with Mountain Sheep A. NEED One method to estimate population numbers of mountain sheep (Ovis canadensis canadensis) is mark-resighting, based on mark-recapture (Caughley 1978) with a lincoln-Petersen estimate (Petersen 1896, lincoln 1930). This method involves first capturing and marking the animals, and subsequently recapturing or resighting them. Identifying and counting marked and unmarked individuals then provide estimates of population size. Quantitatively, the basic eitimator of population size (N) with one marking occasion and one recapture occasion is (nO (n2) (equation 1) N - where n1 - the number of animals initially caught and marked, n2 s the number of animals caught in the second occasion, and m2 - the number of marked animals caught iim the second occaston.. ',"' However, this estimator becomes more biased as sample size decreases (White et al. 1982) and may be corrected for (n1 + n2) > N by modifying equation 1 as follows (Chapman 1951, Seber 1982:60): N c (n1 + 1)(n2 + 1) ---------------(m2 +'.1) 1. (equation 2) �124 The associated estimated sampling variance is (n1 + 1){n2 + 1){n1 - m2)(n2 - m2) var (N) • ---------------------------------{m2 + 1)2 (m2 + 2) (equation 3) (Seber 1982:60). When using this method, several critical assumptions must be satisfied to accurately estimate abundance. The first assumption states that the population is closed demographically and geographically. No individual may enter through birth or immigration, or leave through death or emigration; given only one marking occasion the proportion of marked individuals must remain constant. This assumption is rarely met in a natural population, but its effects can be minimized (e.g. by shortening the time period between marking and recapturing). Some models have been developed incorporating deaths and migration (Seber 1982, Seber 1986). Kufeld et ale (1987), using radio-collars on mule deer, determined which proportion of marked animals were in the study area at any given . occasion and thus maintained a "closed" population. A second assumption of the mark-recapture method states that all animals (marked and unmarked) have an equal and constant probability of being captured and of being recaptured or resighted on subsequent occasions. Animals that are captured on a given occasion should not be more or less susceptible to being recaptured or resighted than others. In a study where animals are recaptured, if heterogeneity in capture probability occurs, the estimates can be significantly biased (Otis et ale 1978:11,25,30,34). In a study where animals were not marked but were sighted twice, Pollock and Kendall (1987) reported heterogeneity i~ sighting probability led to negative bias in the estimates. In a study where animals were live-captured and marked and then "recaptured" by resighting from the air, White and Garrott (1987:220) reported that capture heterogeneity was avoided because one method of capture and a different method of -recapture" were~ used, and that individual sighting heterogeneity did not bias estimates of population size but did underestimate the ~ariance. One explanation for this heterogeneity in probabilities may be disproportionate marking so that marks are not randomly diitributed among the animals but on2 segment of the population receives a greater proportion of marks (Rice and Harder 1977). Alternatively, the researcher may not be able to regulate heterogeneity of probabilities because trapping probabilities and reactions vary per individual animal. . 2 �125 A third assumption states that samples be random. Two possibilities exist to minimize this problem: randomize the capture effort, or randomize the recapture effort. This assumption is violated if one animal's probability of being sighted is not independent of another's (i.e. animals form more or less stable social associations). If marked animals are not randomly dispersed throughout the population and are accumulated in higher densities around the initial trapping areas, traditional variance calculations are not valid (Kufe1d et a1. 1987). The variance of the estimate increases with clumping (Rice and Harder 1977). Researchers have little control over this aspect if the animals being studied characteristically band together, but if groups tend to be randomly distributed, then groups can be counted and mean group size (number of animals per group) calculated separately to reduce the bias and still get an estimate of abundance (Bergerud and Manuel 1969, Eberhardt et a1. 1979, Seber 1986). A fourth assumption states that animals do not lose their marks. If they do, a greater proportion of unmarked individuals will be counted in the population than is actually present. A final assumption states that marked and unmarked animals be correctly identified, counted, and recorded. The observing and recording must be precise and unbiased. Violations of these assumptions result in biased estimates, and in reality, violations often occur: population demographics change; capture and sighting probabilities differ per individual; animals group together; animals lose their marks; and recording errors are made. An open population with recruitment and mortality results in an estimate that is too high (Otis et a1. 1978:10, White et a1. 1982:3). Carothers (1973) noted an underestimation of the population when probabilities of capture differ. Eberhardt (1969) reviewed several studies, noted the bias involved with unequal capture probabilities, and suggested revision of trapping methods (i.e. randomization of trap locations and shifting of traps) and modification of the Lincoln Index ~ ( i.e. N - (n1 + n2)(nl + n2 - m2) /2m2 where variables are defined as previously). In contrast, Magnusson et al. (1978) feel that violations of the equal capture probability assumption w!:l not bias the estimate if the probability of being marked and recaDtured are independent and based on beta distributlons. However this condition will only apply to populations where capture methods differ from recapture methods. In regard to sighting probabilities, Strandgaard (1967) observed that variations in behavior of individual roe deer affected the animal's likelihood of being sighted, and ·consequently the estimate was depressed. Little research has been done with regard to clumping, however Quinn (1980), using schooling populations of fish, provided a mark-recapture estimate accounting for clumping given specific marking schemes or population characteristics • ._ 3 �126 Interactions of one or more of these violations of assumptions result in an estimate that is biased, imprecise, or both. Another variable affecting the mark-recapture estimator includes the number of animals marked. For example, given a single marking session, Bartmann et ale (1987) advised that at least 45% of a mule deer population be marked to achieve a reliable estimate for small populations. In a study where roe deer were marked as they were caught, Strandgaard (1967) concluded at least 66% of the population must be marked to obtain estimates that were acceptably precise and unbiased. The number of recaptures or resightings can also affect the estimator. Improved estimates are achieved through multiple recaptures or resightings. Schnabel (1938) and Darroch (1958) described this method of mark-recapture with multiple recaptures. At each recapture, the proportion of marked animals was recorded, unmarked animals were marked, and all were then returned to the population. The proportion of marked animals over all recapture occasions was then used to estimate the population size and variance. Rice and Harder (1977) used a similar multiple recapture method with white-tailed deer but did not mark after the initial marking because resightings were done via helicopter surveys. These researchers emphasized the importance of multiple resightings to strengthen the population estimate by increasing precision. When all animals caught in a single capture occasion are marked and recaptures result from multiple resightings, separate Lincoln-Petersen estimates result from each resighting. These individual estimates must be combined to get an overall population estimate. This estimate is effected best by using a joint maximum likelihood estimate with a hypergeometric distribution, or JOMLEHD (Chapman 1951, Seber 1982:59, Seber 1986, Bartmann et ale 1987). Bartmann et ale (1987) and White and Garrott (1987:216) explained the values of this estimate over the mean or the median: minimum variance and narrower confidence interval width. Using JOMLEHD reduces bias and increases precision of the estimate (for definitions see White et ale 1982:18-19). Specifically, White and Garrott (1987) define the MLE as the value of N where the following is maximized: . k+l nl m'1 N - nl ni - mi ------------ (equation 4) N 1 • 2 n'1 and the terms are defined for all i • 2 to .k+l number of sightings. This can be found iteratively with computer. 4 �127 Aerial mark-resighting methods have been used to estimate numbers of free-ranging mountain sheep (McQuivey 1978, Leslie, Jr. and Douglas 1979, 1986). By using two different methods for capture and resighting, confounding factors (time differences, animal affinity or aversion to traps, and individual variability in these reactions) are eliminated (Otis et al. 1978:11). In comparison with fixed-wing aircraft, a helicopter also allows greater maneuverability in the air and optimum airspeed and height above the ground (Seber 1986). However, questions about the effectiveness and accuracy of helicopter mark-resighting methods exist (Furlow et al. 1981, DeYoung 1986). When using the hypergeometric model (sampling without replacement), animals, both marked and unmarked, cannot be counted more than once (Rice and Harder 1977). Visibility bias can lead to erroneous population estimates (Pollock and Kendall 1987). If the collars or marks make the animals more conspicuous to the observer, the population would be underestimated because unmarked individuals would not be noticed as readily. Alternatively, Bear et al. (1987) overestimated an elk population when marked animals were misclassified. Packard et al. (1985) noted that habitat, weather conditions, and observer experience can all lead to an underestimation of population abundance. Floyd et al. (1979) in studies of deer and Samuel et al. (1987) in studies of elk examined the effects of observer bias and evaluated the effects on counts of observers, changing ground and weather conditions,and different cover types. With moose, LeResche and Rausch (1974) noted the inaccuracy of the counts given observer experience, number of observers, snow conditions, habitat and terrain, and time of day. Finally, Caughley (1974) listed several factors that can affect the observer and the probability that an animal will be seen: thickness of cover, background, lighting, animal's color, animal's movement, animal's size, observer's eye sight, level of fatigue, speed of travel, altitude, and strip width. Because this method of mark-resighting,·with one marking period and multiple aerial resightings, is used fairly extensively, the estimator of population size should be precise and unbiased. To date, few studies have mathematically assessed the extent of problems inherent in this method. To improve estimates, this mark-resighting method needs to be evaluated quantitatively with simulations, using realistic data. With a known population size, the effectiveness of JOMLEHD estimator and its robustness to each individual asswdption need to be simulated and tested. Similarly, violations of the assumptions need to be simulated and tested with populations of known size. 5 �128 B. OBJECTIVES 1. Document the performance (precisi~n and bias) of JOMLEHD with varying capture probabilities (proportions of population marked) and varying number of sighting occasions (flights), given that all assumptions are met. HI: JOMLEHD performs acceptably despite variabilities in capture probabilities and number of resighting occasions. 2. Simulate a continuous range of heterogeneities in the sighting probabilities of individuals to evaluate the performance of the estimator. H2: JOMLEHD performs acceptably despite heterogeneous sighting probabilities of individual animals. 3. Simulate discrete probabilities· of sighting (e.g. habituation to the helicopter might cause the proportion of marked animals seen to increase or decrease sequentially per flight) to evaluate the performance of the estimator. H3: JOMLEHD performs acceptably despite changes in sighting probabilities across resighting occasions. 4. Simulate a range of aggregations among marked animals (degrees of independence of sighting probabilities) to evaluate the performance of the estimator .. H4: JOMLEHD performs acceptably despite nonrandom associations (aggregations) among individuals. .. 5. . Develop guidelines on the design and methods of captureresighting studies to improve precision and reduce the bias of the estimates. 6 �129 C. EXPECTED RESULTS OR BENEFITS Colorado requires a reliable estimator of mountain sheep numbers before intensive management systems can be devised, tested, and implemented. If mark-resighting estimates prove to be precise, unbiased, and robust to violations of model assumptions, mark-resighting estimation could be used as the standard procedure for tracking mountain sheep numbers through time. If estimates prove to be imprecise, biased, or nonrobust, suggestions will be made to improve the estimator. D. APPROACH Field Studies Several variables related to mountain sheep mark-resighti~g methods need to be determined from field studies. Reed (1987) proposes to study 28 radio-collared ewes to estimate capture probabilities, sighting probabilities across multiple resighting occasions, probability distributions for these sightings, average group size, and frequency of aggregation. Up to fifteen surveys per year will be flown to determine population estimates and to conduct these studies. This information will then be used to develop a set of realistic conditions for the simulations. Simulations The Statistical Analysis System (SAS) will be used with all MonteCarlo simulations to derive JOMLEHDs and their respective variances and confidence intervals. Four population sizes (50, 100, 200, and 500) will be used in all simulation sets. Objective 1 - Document the performance of JOMLEHD with varying capture probabilities and varying number of sig~ting occasions, given that all assumptions are met. The first objective evaluates the overall estimator performance given that all assumptions (equal and constant capture and sighting probabilities, and randomly distributed animals with no clumping) are 7 �130 met. The null hypothesis to be tested is that neither capture probabilities (proportion of population marked) nor number of sighting occasions affect the JOMLEHD. Capture probabilities are assumed uniform for the whole population and will be 0.1, 0.3, and 0.5. The number of sightings occasions (flights flown) wilf be 5, 10, 15, and 20. For this set of simulations, sighting probabilities will be 0.1, 0.3, 0.5, and 0.7 and constant. Objective 2 - Simulate a continuous range of heterogeneities in the sighting probabilities of individuals to document the performance of the estimator. The second objective consists of testing heterogeneity of sighting individuals. The null hypothesis to be tested is that heterogeneity of sighting individuals does not affect the JOMLEHD. Population sizes, capture probabilities, and number of sighting occasions used in the initial simulation will be used here. However, to simulate heterogeneity of individuals, the mean probability of sightings will tentatively be set at 0.1, 0.3, 0.5, and 0.7, and each probability of sighting will be based on a beta distribution. For each probability of sighting, three beta distributions are used (symmetrical distribution [1], skewed left distribution [2], and distribution with greater proportions in the extremes [3]). [3 ] o 0.2 0.4 0.6 0.8 SfQtrtjng Probabimty Fig. 1. Beta distributions that will-be simulated for the sighting probabilities (figure from Burnham et ale 1987) 8 1 �131 From the specified beta distribution, SAS will randomly pick a value of probability of sighting for each animal on all sj~hting occasions. This will be repeated for all animals marked with the result that each marked animal was "seen" a proportion of the total counts. All of these values will then be used to calculate an estimate of the population for that simulation, and the JOMLEHD will be calculated. Objective 3 - Simulate discrete probabilities the performance of the estimator. of sighting to document For objective '3, to detect effects on the estimator if sighting probability is variable (e.g. habituation to the helicopters), as compared to the continuous distribution, the mean sighting probability will be set at 0.5 but three discrete values (PO - 0.7, PI - 0.5, P2 0.3) for a decreasing step-wise distribution to be incorporated. The simulations will then be repeated. An increasing step-wise distribution will also be simulated to demostrate the opposite effects. The null hypothesis to be tested is that variations among flights in sighting probabilities will not affect the JOMLEHD. Objective 4 - Simulate a range of aggregations among marked animals to document the performance of the estimator. To meet the fourth objective to determine effects of aggregations of animals, average group size (mean with its variance) and fidelity to the group (percent of time together) will be incorporated into the simulations instead of a random distribution of independent animals. Sighting probability will be 0.1, 0.3, 0.5, and 0.7 with uniform distributions. The average group size and fidelity of the group will be determined from the field studies, and are expected to follow uniform rather than beta distributions.· The null hypothesis to be tested is that clumping of animals does not affect the JOMLEHD. 9 �132 In summary, the following values will be used for simulations for each objective: Objective 1.:. L a, ~ Designated population size (n) 50 100 200 500 50 100 200 500 50 100 200 500 50 100 200 500 Capture probability (c) 0.1 0.3 0.5 0.1 0.3 0.5 0.1 0.3 0.5 0.1 0.3 0.5 Number of sighting occasions (f) 5 10 15 20 5 10 15 20 5 10 15 20 5· 10 15 20 Mean sighting probabil ity (s) 0.1 0.3 0.5 0.7 0.1 0.3 0.5 0.7 0.5 0.1 0.3 0.5 0.7 Distribution of sighting probability uniform 1,2,3 Average group size (g) 1 1 1 (field data) Variance of group size (varg) 0 0 0 % of time together (%) 0 0 0 (field data) ""(field data) discrete uniform One set of simulations is composed on one value for each of t~e variables (i.e., one set for objective 1 is n=50, c=O.I, fa5, s=O.1 with constant distribution of sighting probability, g=l, varg=O, and %=0). Given the variables and their possible values, at least 624 sets are possible. For each set, one-thousand replications will be simulated. Results of the field studies will be used to determine a realistic base to indicate which values of the simulations can be expected. 10 �133 For the final analysis, for each set of simulations, the JOMLEHD will be determined, and means and variances with confidence intervals and coefficients of variation will be calculated. Also, percent relative bias (PRB), variance, and mean squared error of the bias estimate between the JOMLEHD and the true population size will be calculated. The 95% coverage of the estimator will be determined. The PRB of each method can be plotted with a given parameter to note differences between methods. Finally, goodness-of-fit tests with the estimates and the actual values will be performed. The above data will be used to check the overall performances of the estimators given different combinations of capture probabilities and numbers of sighting occasions, and violations of assumptions of constant sighting probabilities or random animal distribution. Schedule Activity Period Design study Capture and collar sheep, and conduct resighting samples Helicopter counts Conduct population estimation simulations Completion date August - December 1987 January - May 1988 November 1988 - February 1989 May 1988 - 1989 May 1989 Personnel Andrea K. Neal R. Bruce Gill Dale F. Reed Gary C. White Graduate Research Assistant Co-Principal Investigator Co-Principal Investigator Co-Principal Investigator Estimated Annual Cost ETE Requirements Expenditure CategorY Costs Contract (01) Personal Services: Contract $13,000.00 -; 11 �134 E. LOCATION The study will be located on Trickle Mountain, 3S009'N l06020'W, Saguache County, Colorado. The simulations will take place in the Fishery and Wildlife Biology Department at Colorado State University, Fort Collins, Colorado. F. RELATED FEDERAL PROJECTS This is a companion study to Colorado FA Project W-lS3-R-2 Work Plan 2A Job 5. Job 5 will provide real estimates of simulated values to define real-world ranges to be expected among simulated values. Job 5 (supplement) will simulate responses of the JOMLEHD both to preselected values and to real-world values. 12 �135 literature Cited Bartmann, R.M., G.C. White, l.H. Carpenter, _and R.A. Garrott. 1987. Aerial mark-recapture estimates of confined mule deer in pinyon-juniper woodland. J. Wildl. Manage. 51(1):41-46. - Bear, G., G.C. White, l.H. Carpenter, and R.C. Garrott. 1987. Observability bias in mark-resighting estimates of elk populations. J. Wildl. Manage. (submitted) Bergerud, A.T. and F. Manuel. 1969. Aerial census of moose in central Newfoundland. J. Wildl. Manage. 33(4):910-916. Burnham, K.P., D.R. Anderson, G.C. White, C.Brownie, and K.H. Pollock. 1987. Design and analysis methods for fish survival experiments based on release-recapture. Am. Fish. Soc. Monogr. 5. Bethesda, Maryland. 437pp. Carothers, A.D. 1973. Capture-recapture methods applied to a population with known parameters. J. Anim. Ecol. 42(1):125-146. Caughley, G. 1974. Bias in aerial survey., J. Wildl. Manage. 38(4):921-933. Caughley, G. 1978. Analysis of vertebrate populations. John Wiley and Sons. New York, New York. 234pp. Chapman, D.G. 1951. Some properties of the hypergeometric distribution with applications to zoological sample censuses. Univ. Calif. Publ. Stat. 1:131-160. Darroch, J.N. 1958. The multiple-recapture census. I. Estimation of a closed population. Biometrika 45(3/4):343-359. DeYoung, C.A. 1986. Accuracy of helicopter surveys of deer in south Texas. Wildl. Soc. Bull. 13:146-149. Eberhardt, l.l. 1969. Population estimates from recapture frequencies. J. Wildl. Manage. 33:28-39. Eberhardt, l.l., D.C. Chapman, and J.R. Gilbert. 1979. A review of marine mammal census methods. Wildl. Monogr. 63:1-46. Floyd, T.J., l.D. Mech, and M.E. Nelson. 1979. An improved method of censusing deer in deciduous-coniferous forests. J. Wildl. Manage. 4391):258-261. Furlow, R.C., M. Haderlie, and R. VandenBerge. 1981. Estimating a bighorn sheep population by mark-recapture. Desert Bighorn Council, Trans. 1981:31-33. 13 �136 Kufeld, R.C., D.C. Bowden, and D. L. Schrupp. 1987. Estimating mule deer density by combining mark-recapture and telemetry data. J. Mamm. 68(4):818-825. . LeResche, R.E. and R.A. Rausch. 1974. Accuracy and precision of aerial moose censusing. J. Wildl. Manage. 38:175-182. - Leslie, Jr., D.M. and C.L. Douglas. 1979. Desert bighorn sheep of the River Mountains, Nevada. Wildl. Monogr. 66:1-56. Leslie, Jr., D.M. and C.L. Douglas. 1986. Modeling demographics of bighorn sheep: current abilities and missing links. N. Amer. Wildl. natur. Res. Conf. Transc. 51:62-73. Lincoln, F.C. 1930. Calculating waterfowl abundances on the basis of banding returns. U.S. Dept. Agric. Circ. 118. 4pp. Magnusson, W.E., Caughley, G.J., and G.C. Grigg. 1978. A double-survey estimate of population size from incomplete counts. J. Wildl. Manage. 4291):174-176. McQuivey, R.P. 1978. The desert bighorn sheep of Nevada. Nevada Fish, Game Biol. Bull. No.6 81pp. Otis, D.L., K.P. Burnham, G.C. White, and D.R. Anderson. 1978. Statistical inference from capture data on closed animal populations. Wildl. Monogr. 62:1-135. Packard, J.M., R.C. Summers, and L.B. Barnes. 1985. Variation of visibility bias during aerial surveys of manatees. J.Wildl. Manage. 49(2):347-351. Petersen, C.G.J. 1896. The yearly immigration of young plaice into Limfjord from the German sea. Rept. Danish Biol. Stn. Pollock, K.H. and W~L. Kendall. 1987. Visibility bias in aerial surveys: a review of estimation procedures. J. Wildl. Manage. 51(2):502-510. Quinn, T.J. II. 1980. Sampling for the abundance of schooling population ,with line-transect, mark-recapture, and catch-effort methods. _ Ph.D. dissertation. University of Washington. Rice, W.R. and J.D. Harder. 1977~ Application of multiple aerial sampling to a mark-recapture census of white-tailed deer. J. Wildl. Manage. 41:197-206. Samuel, M.D., E.O. Garton, M.W. Schlegel, and R.G. Carson. 1987. Visibility bias during aerial surveys of elk in northcentral Idaho. J~ Wildl. Manage. 51(3):622-630. Schnabel, Z.E. 1938. The estimation of the total fish population of a lake. Amer. Math. Mon. 45(6):348-352. 14 �137 Seber, G.A.F. 1982. The estimation of animal abundance and related parameters (2nd edition). Macmillan Publ. Co., Inc. New York, New York. 654pp. Seber, G.A.F. 1986. A review for estimating animal abundance. Biometrics 42:267-292. Strandgaard, H. 1967. Reliability of the Petersen method tested on a roedeer population. J. Wildl. Manage. 31:643-651. White, G.C., D.R. Anderson, K.P. Burnham, and D.l. Otis. 1982. Capturerecapture and removal methods for sampling closed populations. los Alamos National laboratory. lA-8787-NREP. los Alamos, N.M. White, G.e. and R.A. Garrott. 1987. Analysis of biotelemetry data - a primer. (in prep) 15 ��Colorado Division of Wildlife Wildlife Research Report July 1988 139 JOB FINAL REPORT State of Colorado Project No. W-153-R-2 -----------------------No. 3A ---------------------- Mammals Research Work Plan Pronghorn Investigations Job No. 1 Period Covered: July 1, 1987 - June 30, 1988 Author: Pronghorn Population Dynamics Study T. M. Pojar ABSTRACT The final phase of the pronghorn census study was completed. The objective was to quantify the biases due to pronghorn movement during the quadrat search. Two helicopters were used; one with the usual crew of pilot, primary observer, and navigator/observer flew the quadrat in the standard fashion (i.e. at 15-20 m altitude and 65-100 km/hr). The other helicopter with 2 experienced observers was flown at a higher altitude (100-200 m) to broaden the visual perspective and permit the observers to track movements of pronghorns observed. The density estimate based on data from the upper crew was 4.85% less than the estimate from the lower flight. Observations from the upper helicopter indicated that 4 groups (18 animals total) were missed, and 1 group (6 animals) was erroneously included in the sample of the lower helicopter crew. These corrections resulted in a population density estimate 2.10% greater than the uncorrected estimate. Data analysis and manuscript preparation is progressing. One manuscript has been completed and submitted to The Journal of Wildlife Management.. Appendix A is a draft copy. The remainder of the project findings will initi8.l:iybe included in a single manuscript, the outline of which is Appendix B. ��141 PRONGHORN POPULATION DYNAMICS STUDY Thomas M. Pojar P. N. OBJECTIVE Develop standard operating composition inventories. procedures for pronghorn density and sex/age SEGMENT OBJECTIVES 1. Quantify the responses 2. Evaluate of pronghorns to helicopter the effects of flight responses overflights. on quadrat census bias. ACKNOWLEDGMENTS Special recognition is given to M. Bauman of the Colorado Division of Wildlife for his interest, participation, and contributions to this research effort since its inception. V. Graham, also of the CDOW, has been a consistent and valuable contributor throughout the duration of this project. D. Bowden of the Statistics Department (Colorado State University) provided statistical consultation and creative suggestions. D. Bartmann of the CDOW served as an observer during this final phase of the project, and his participation is gratefully acknowledged. PRONGHORNS' RESPONSE TO HELICOPTER QUADRAT SEARCH Justification Comparison of density estimates using random strip transects and random quadrat sampling systems has resulted in similar precision levels but a large (ca 2X) and consistent difference in estimated density (Fig. 1). The line transect method was applied to a portion of the study area for 2 consecutive years with highly inconsistent results. However, the line transect data does strongly indicate that the observability of animals decreases sharply with increasing distance from the flight-line (Fig. 2). It may be assumed, then, that the strip transect method results in an underestimation basis. The current data set does not provide any information on what biases, if any, are affecting the random quadrat density estimation. Pollock and Kendall (1987) reviewed some methods of estimating bias in aerial surveys and suggest that the best method is to have a total enumeration of a subsample of the population. Experiments censusing pastures with known numbers of mule deer (Odocoileus hemionus) resulted in an approximate underestimation bias of 33% (Bartmann et ale 1986). These experiments were conducted in pinyon (Pinus edulis)-juniper (Juniperus osteosperma) habitat type in the Piceance Basin, Colorado. Experienced observers flown over moose (Alces alces) enclosures in birch (Betula papyrifera)-spruce (Picea glauca) cover underestimated the known numbers of moose by 32% (LeResche and Rausch 1974). �142 Pronghorn (Antilocapra americana) inhabit open rolling terrain with shortgrass or sagebrush steppe vegetative types. Having evolved in this habitat type, their principal defense mechanism to disturbance is high-speed flight. The low level helicopter flights over quadrats is relatively intense and initiates the flight reaction in pronghorn. During the search of a quadrat, it is assumed that each group of animals observed either is driven from the quadrat or that it is possible for the observers to keep track of the groups (visually or by group size and composition) so they will not be counted more than once. Because of the amount of relief in the terrain, it is not possible to keep all areas of the quadrat in sight at all times. Pronghorns are notorious for their speed and endurance (Hoover et al. 1959). Sprints of 60 miles per hour are possible, and longer runs at 30 miles per hour are not uncommon (Carr 1927). It takes at least 4 mins to fly the perimeter of a quadrat, during which time a pronghorn could travel about 2 mi. This mobility, which is not encountered in other species censused with quadrats, may lead to biases in the estimates. Biases due to object mobility during quadrat censusing may be caused by one or more of the following: a. b. c. d. Group departs the quadrat before detection. Group enters the quadrat before detection. Detected group joins another group while out of sight of observers. After detection, a group may split into 2 or more groups, leading the observers to count them as new groups. It is critical to have some measure of accuracy by which to judge (and adjust, if necessary) the results of a census technique. The quadrat sampling unit lends itself to this sort of evaluation because it is a relatively small area compared to strip and line transect areas. In mule deer, pinyon-juniper habitat, it was possible to fence sample areas and stock them with known numbers of animals. This approach was considered but dismissed for pronghorns because of the expense of fencing and because fences would obviously alter the pronghorns' natural movements in response to the search aircraft. The latter is the more crucial objection since results based on experimentation that alters the pronghorns' natural flight response would not be applicable in natural situations. METHODS AND MATERIALS Approximately one-third of the Great Divide study area was desi~ted as the test area. On the study area map, it is Block 3, which is 442 mi2 in size. There are 40 randomly selected (and marked) quadrats in this area from previous years' quadrat surveys. Given certain assumptions about the variance of pronghorn density on sample quadrats, a sample of 40 quadrats was calculated to be sufficient to detect a 15% error at the 90% confidence level. The quadrat corners were marked with 2' x 3' flags of nylon reinforced, plastic tarp attached to the existing marker posts. This marking was in addition to the l' x 2' plastic panels that were in place. The intent was to make the quadrat corners visible to observers in a helicopter flying at 100200 m altitude. �143 These 40 quadrats were searched with exactly the same procedure used for the entire quadrat census during the past 4 yrs. We used a Bell-Soloy helicopter with the same observer (myself) and the same observer/navigator (Mike Bauman) as in past years. The search consisted of flying the perimeter at 65-100 km/hr, 15-20 m altitude, and transecting the interior to accomplish a thorough search. Groups of pronghorns seen on the quadrat were pursued and classified as has been done in past censuses. A second helicopter, with a pilot and 2 observers, followed the primary helicopter at an altitude of approximately 100-200 m giving a broader perspective of the primary search. There was no communication between the 2 aircraft to ensure independent observations. The observers in the upper aircraft recorded location, movement, and group sizes of all groups detected. Movement of each group was detailed throughout the search verbally on tape recorder and on a map of each quadrat. Movements of groups across quadrat boundaries and mixing and splitting of groups were recorded. RESULTS All 40 quadrats were searched in the prescribed manner on 18 August 1987. The flag markers were only partially successful in delineating the quadrat corners for the upper crew, so it was necessary for them to key on the turns of the lower helicopter to estimate the location of the actual corners. The pronghorn density estimate based on data from the upper crew was 4.85% less than the estimate from the lower flight. Observations from the upper helicopter indicated that 18 animals (4 groups of 8, 7, 1, and 2) were missed by the lower crew. It was also determined that the lower crew erroneously included 1 group of 6 animals. This group was bedded on the exterior of the quadrat and was disturbed during the perimeter flight. The lower crew was intent on watching to the interior of the quadrat and did not see the group while it was bedded. After being disturbed, the group ran onto the quadrat where the lower crew first observed them. After the above corrections were incorporated into the data set, the density estimate was increased by only 2.10%. There was no significant (p - .348) difference in the number of animals seen by quadrat between the upper and lower crews as determined by a paired t-test (t - .954, 39 df). The evidence seems strong that prongho~~ groups, upon being disturbed, do not manuever or regroup leading to their being counted more than once by the primary search team. LITERATURE CITED Bartmann, R. M., L. H. Carpenter, R. A. Garrott, and D. C. Bowden. 1986. Accuracy of helicopter counts of mule deer in Pinyon-Juniper woodland. Wildl. Soc. Bull. 14(4):356-363. . Carr, C. 1927. The speed of pronghorn antelope. J. Mammal. 8(3):249-250. �144 Hoover, R. L., C. E. Till, and S. Ogilvie. 1959. The antelope of Colorado. 1l0pp. Colo. Dept. Game and Fish. Tech Bull. No.4. LeResche, R. E., and R. A. Rausch. 1974. Accuracy and precision of aerial moose censusing. J. Wildl. Manage. 38:175-182. Pollock, K. H., and W. L. Kendall. 1987. Visibility bias in aerial surveys: A review of estimation procedures. J. Wildl. Manage. 51:502-509. Prepared bY~fZr OJif~ Wildlife Reseracher �145 GREAT DIVIDE 20 - 18 0 0 0 .•... 16 ~ w 14 le{ ~ ~ 12 C/) w l- 10 e{ a: 0 8 e{ y= ::> 0 -269.99 + 2.15x R2 6 = .991 O~ 0 5 6 7 8 TRANSECT ESTIMATE (x 1000) Figure j. 1 :,':. __ Comparison of strip vs. quadrat population size estimates. Great Divide pronghorn DAU. :r----------------------....._---------..., 1 10 CRAIG 1986 Strip Tr~In:;"gt Di:;tong" E;timot,,:; ~ o o Z £(l :;Q 10 o ~~~~~~~~--~~~~~~~~~~-ee 100 1$0 ~OO .25Q ~OO seo .oKlO 4$Q ~~~--~~~~~.---~~~ 60Xl sso £(00 65(l 700 ~, eoo Di:5tonl:le(meter:;) from f1i~ht-line Figure 2. ~umber of groups detected by estimated distance from the flight-line. Great Divide strip transects, 1986. • �146 Figure 3. Test area (Block 3) is 442 m2 and has 40 marked quadrats. �147 APPENDIX A Estimation of Proportions and Ratios from Line Transect Data Tom Drummer Assistant Professor of Mathematics Department of Mathematical Sciences Michigan Technological University Houghton, Michigan 49931 Tom Pojar Researcher, Colorado Division of Wildlife Fort Collins, Colorado 80526 Lyman L. McDonald Professor of Statistics and Zoology University of Wyoming Laramie, Wyoming 82071 �148 ABSTRACT If the population of interest in a line transect survey consists of distinct sub-populations, or classes, line transect data can be used to estimate the proportion of the population in each class, or the ratio of the number of individuals in one class to another class. However, if the various classes have unequal overall detect ability then the observed ratio or proportion is biased and should be adjusted. Procedures for examining the equal detectability hypotheses are discussed, and estimators for the desired proportions and ratios proposed and studied. The problem is further complicated when the population consists of groups of individuals, and the size of the group influences its detectability. In this case, a post stratified ratio estimator of the desired ratio is proposed. An application of this procedure to data from a pronghorn antelope survey is given, although in this illustration, the impact of the post-stratification is minimal. KEYWORDS: line transect, proportions, ratios �149 Tom Drummer Assistant Professor of Mathematics Department of Mathematical Sciences Michigan Technological University Houghton, Michigan 49931 Introduction Line transect data are traditionally used to estimate the total number of individuals in a study area. However, line transect data can also be used to estimate attributes of a population in addition to its total number. Specifically, if the population consists of several sub-populations, or classes, the estimator may desire an estimate of the proportion of the population in each class, or of the ratio of the number of items in one class to the number of items in another class. For example, if a manager is using line transect sampling to estimate the size of a game population, he or she may also need an estimate of the ratio of the number of male animals to female animals. The observed male to female ratio may be biased if, say, female animals choose habitat which makes them more likely to be detected (sampled) than males. Due to the over-representation of females in the data, the observed male to female ratio would tend to underestimate the true ratio. When using line transect data in this manner, it is important to consider the possibility that the various classes of objects may be sampled with unequal probability. Clearly, individuals at varying distance from the transect are sampled with varying probability. The concern here is that the average probabilities of detection for the various classes may differ substantially. In this instance, given sufficient data, \Ierecommend the use of the estimated sub-population sizes (densities) to estimate the desired ratios (proportions), rather than simply using the observed number of individuals in the sub-populations to obtain these estimates. In the grouped population case, it is possible that both the group's size and composition could influence its detectability. If group size alone �150 Drummer influences detectability, then, given sufficient data, we recommend estimating sub-population densities by post-stratifying on group size, and using these estimates to obtain estimates of the desired ratios (proportions). where composition influences detectability is unresolved. The case We suspect, however, that in many cases, group size and group composition are related, so adjusting for group size influence is sufficient. �151 Drummer Estimators The reader is referred to Burnham, Anderson, and Laake (1980) for a comprehensive discussion of line transect assumptions and estimation procedures. For our purposes, we assume that the population of unknown size N consists of k classes with N.~ items in the ith class (i = 1,2, •..k). The probability of detecting an individual in the ith class at perpendicular distance x from the transect is given by the detection function g.~ (x), 0< g.(x) ~ 1.0. ~ Let c. ~ =J co g.(x) dx, with 0 ~ 2Lc. i A denoting the average probability of detection for items in the ith class, = --~ n (1) where A is the size of the study area and L is the total length of transects randomly located therein. The n umb er 0f't ~ ems d e tecte d' ~n the ~.th c l'ass ~s n., w h'~ch we assume t 0 ~ have a binomial distribution with·E(n.) ~ = = N.n. (l-n.) ~~ ~ N.n. and Var(n.) ~~ ~ The n. are assumed to be independent of one another. (Seber, 1982). ~ k The total number of items detected is n = ~ ni. i=l An estimate of the number of items in the ith class is given by A -1 n.A c.~ ~ N. = n . t«, = 2L ~ ~ ~ (Seber, 1982), wit~ estimated variance given by '" Var(Ni) A A = '" () (A/2L)2 {('" ci_1)2 Var ni (Burnham and Anderson, 1976). A A Var (n.) ~ = A A (2) + ()2 ni Var ci-I)} A Under the binomial assumption, A n.(l-n.) (Quinn and Gallucci, 1980) with n.~ » I ~ (3) ('" = '" n./N. ~ ~ = '" 2L c./A. ~ �152 Drummer If the size of the study area, A, is unknown, then the population density, Di = Ni/A, is estimated. Also, Var (ni) can be estimated via " replication or, using a Poisson assumption, Var(n i)'= ni. We employ the binomial formula throughout. Let r1.) .. = N./N. denote the ratio of the number of items in class i to 1.) the number of items in class j. An·obvious estimate of rij is given by the observed ratio, r .. 1.) = n./n. . (4) 1.) If, overall, the two classes are equally detectable, then the observed ratio is approximately unbiased. Assuming that ni = nj = n, and pooling data across both classes to estimate n, an estimate of the approximate variance is given by " V~r(; ..) 1.J = (; .. 1.) 1\ + (l-n)} )2 {(l-n) n. n. 1. (5) ) If the two classes are not equally detectable, then the relative bias in r.. 1.) is approximately proportional to the ratio n. /n ., which can lead to 1. substantial bias. J Adjusting the observed ratio by estimates of ni and nj yields the adjusted estimator 1\ " r .. * 1.) 1. with estimated variance given by " " " " "2 far(N.) Var(r ..*) = (r..*) ,,~ 1.J 1\ N./N. 1.) ". (N.) 1. (6) ) ,., ,., Var (N.) } + A ~ • (7) (N .) ) The adjusted estimator is only feasible if there is sufficient uata to obtain reliable line transect population estimates for each of the classes involved. Although the adjusted estimator is approximately unbiased, its variance is large compared to the observed ratio. A simple formula expressing the relationship between the variances is not possible for the general case. study the unbiasedness for precision trade-off via a limited Monte Carlo We �153 Drummer computer Let class. simulation. Pi - N./N denote the proportion of ite~ 1. The observed p. - n./n, proportion, 1. classes have equal detectability. the bias depends in the population is approximately unbiased If this is not the case, then the extent of in the n. and the size upon the disparity Assuming equal detectability all data to obtain n, if all k 1. (N.) of each of the 1. classes. in the ith an estimate (n 1 - 1. n2 - - nk - n), and pooling of the approximate variance of p. is given 1. by Var(p.) c 1. (~.)2{(1-;) _ (1-;)} . n. n (8 ) 1. 1. The "adjusted adjusts each of the n. by n., proportion" 1. yielding 1. A A p. n. In. * k 1. l: i=l k l: where N A = N. --:-_1. __ 1._ IN ( 9) 1. n. In . 1. l. k N., Var(N) c 1. i=l l: Var (N.) 1. i=l The experimenter needs sufficient data to obtain a reliable density estimate in each of the individual et al. (1980) discuss the estimation classes. simulation variance methods of the adjusted '" * ) Var (p. 1. A - Again, the adjusted exceeds assume that N - (p. 1. proportion We note also that Burnham of total population data across classes with varying detectability. N .• line transect size when pooling The variance An estimate formula and of the approximate 1. is given by * 2 (10) ) proportion is approximately that of the observed proportion. unbiased but its variance �154· Drummer Simulation Procedure A limited computer simulation was performed in order to determine under A A what detectability conditions the "adjusted" estimators, rij* and Pi*' are A A preferrable to the "observed" estimators r .. and P., respectively. ~J .•. The simulation was limited in that we simulated line transect sampling from only two detection curves, the half-normal model (Quinn, 1977), and negative exponential model (Gates, Marshall and Olson, 1968). These curves represent the extremes in terms of shape, with the half-normal possessing a broad shoulder and the negative exponential having a sharp spike near the origin. " The shape of the curve near the origin influences Var(N). All simulation results are based on R=500 replications. For each estimator, we estimated the percent.relative bias (% BIAS), variance, coefficient of variation (CV), mean squared error (MSE), and confidence interval coverage (% COV) as detailed below. A Let 8 9 denote the parameter of interest, with from the ith replicate. 9. 1 denoting the estimate of Then: R A E(9) ::::9" A = ~ e./R i=1 1 A % BIAS :::: (9;8) x 100% ; A R " Var (9) :::: 1: i=1 cv (9) :::: I Var ~9) E A (9.-8. \ 2 -1 R x 100% ; (9) A MSE (9) For each replication, an approximate 95% confidence interval was computed as 9.~± 2 . I V~r (9.~ ) . The percent coverage is the percentage of these �155 Drummer intervals that captured the true value of e. " to the theoretical variance and also to the mean We also compared Var(9) of the variance estimates generated by the replicates. For all four estimators these quantities were in close agreement, indicating that the variance formulae provide approximately unbiased estimators of the true sampling variance. Estimation of Ratios Two populations, each of size N.~ = 1000, were randomly generated, with the parameter of interest being r12 = rij = 1.00. In other simulation runs, the ratio N1/N2 was varied, but this had no appreciable effect on the simulation results. Procedures for simulating line transect sampling are discussed by Gates (1969). A similar procedure was employed here. While each population was sampled with the same type of detection function, either half-normal or negative exponential, the parameter values of the two curves, 1, O2) were varied in order to simulate varying overall (0 probabilities of detection, with a total of 30 pairs 1, O2) examined. We chose 5 values of 01, which generated 5 different overall detection rates for population one, ranging from n1 = 0.03 to nl = 0.09. sample sizes ranging from E(n1) - 30 to E(nl) - 90. (0 This results in expected The bias in r" .. (when ~J n .) is primarily a function of the ratio n. In .. For each value of J ~ . 1 we chose 0 2 to obtain the ratios n1/n2 = 0.50; 0.60; 0.70; 0.80; 0.90; The case n1/n2 = 1.00 is the equal detectability case. 6 values of 1.00. J n._ ~ O Results In general, results for the negative exponential model paralleled those of the half-normal, so only the latter will.be examined in detail. Results �156 Drummer for the simulation are summarized in Table 1 and Figures lA-lC. The horizontal axis in the figures represents the ratio nl/n2. From Table 1 and Figures 1A and 1B, lt is clear that both % BIAS and A % COV for the observed ratio (r.. 1.) c A n./n.) are proportional to n1/n2• 1.) A The A % COV for the adjusted ratio (rij* = Ni/Nj) is consistently close to the target value of 95%, while r1.) ..* exhibits a slight positive bias that A diminishes with sample size. Perhaps most importantly, the reader should A note that the bias in the observed ratio r1.) .. does not diminish with increased sample size. A Column (6) of Table 1 A displays the ratio of Var (r..*)/Var (r..). For 1.) 1.) the equal detectability case (n1/n2 = 1.0), this ratio is in the neighborhood of 1.55. This implies that if the two classes are equally detectable and the experimenter unnecessarily uses the adjusted ratio, he will unnecessarily increase the standard error of the estimate by a factor of approximately 11.55 , or roughly 25%. Also, this ratio tends to increase sharply as n1/n2 A declines. However, MSE A (r..*) /MSE (r..) is < 1.) sample size considerations. 1.) 1. 0 for n1/n2 ~ 0.80, subject to As the disparity between detactability increases, A the bias component of MSE (r..) becomes large, while the bias component of 1.) A MSE (r..*) is consistenly close to zero. 1.) A A A .., as CV(r ..*) As expected, r 1.) ..* is relatively less precise than r1.) 1.) A exceeds CV (r..), typically by 3-8%. 1.) The disparity decreases as sample sizes increase. A Using mean squared error as the criteria, it appears that r 1.) ..* is A preferrable to rij if n1/n2 < 0.80. Of course, the experimenter does not know A A the value of n1/n2, but can obtain nl/n2 and use its value as a rough gauge. Given the potential for severe bias, and poor confidence interval coverage, A the relatively small loss in relative precision makes r1.) ..* an attractive �157 Drummer alternative if there is sufficient data available. Estimation of p, - N,/N -------~1.--1.- For this populations. study we simulated The expected line transect value sampling of the observed from three " proportion, p, - n,/n, 1. k approximate " of N,w,/ L N,w. 1. 1. • 1 J J J= value has 1. If W1 - W2 - ... - wk' then Pi is ,., approximately unbiased. When this is not the case, the bias in p. is a 1. function N . k in the Wi and of the population of the disparity We have summarized the simulation results N , N , 1 2 sizes, in terms ..., of the ratio k N.n./n-BAR, where w-BAR L - 1. 1. know any of the values them. Note k that ,.,,, statistic ] = in this n./(n/k) Of course J J ratio, such an estimate N.n. / ( L N.n./k 1. 1. . 1 J J J= this N.w,/k j=l = but given simplifies kn./n. 1. the experimenter sufficient data does not can estimate to The experimenter can use the value of 1. to judge whether or not it is advisable to use the estimator p.* - N./N or p. - n./n. 1. 1. 1. We present with expected population simulation sample two, results and found exponential comparable model The results (a , a , a ), 1 2 3 O.is to 1.50. sizes results for a situation ranging from 30 to 180. with P2 - 1000/3000 We also examined equal 1. N 1 - N in which Results - N3 - 1000, 2 We arbitrarily - 1/3, as the npopulation for situations results. where chose of interestn. the population for the half-normal sizes were not and negative were similar. presented of parameter here are from a simulation values. Only the half-normal These generated model results involving values 60 triplets, of W /W-BAR 2 are presented here. from �15~ Drummer Results We have displayed horizontal axis the most represents important maintains proportion, the targeted proportional ~ is essentially of confidence. The bias 1.0 implies that than in at least one other are less likely value = 95% at the equal interval, CV 2C, using is preferred the adjusted (p.*) > CV y! var(~.*)/var(~.) ~ ~ As previously ~ chosen in class class, than items equal case, a mean to decide over p.* ~ The squared error in P2 - n /n 2 is preferred. likely of to be detected < 1.0 implies that they one other class. A ~ rapidly elsewhere. the observed For values Although sample For the equal is N2 - N , a value 3 criteria, ~ 1.05. and The % COV for p. reaches detectability. and declines unbaised outside of this not displayed, sizes, detectability indicating a very case, ~ 1.26. stated, which • precision. a ratio in at least if .95 ~ n /n-BAR 2 proportion c 2 are more while ~ loss in relative kn./n since Nl (P.) by only 2-3% for comparable ~ small 1.0 implies detectability From Figure proportion items to be detected of n /n-BAR 2 ~ In this case, n /n-BAR> 2 items 2A-2C. ~ P.* - N./N, ~ level to n /n-BAR. 2 in Figures the ratio n /n-BAR. 2 ~ The adjusted results the experimenter estimator to use. can use the value of the statistic If kn./n ~ 1.00, then p. would ~ ~ be �159 Drummer TABLE 1 SUMMARY OF SIMULATION RESULTS FOR ESTIMATiON OF RATIOS (1) 111/112 .5 .5 .5 .5 .5 .6 .6 .6 .6 .6 .7 .7 .7 .7 .7 .8 .8 .8 .8 .8 .9 .9 .9 .9 .9 1 1 1 1 1 NOTE: -----------------------------------------------------(6) (5) (4) (7) RATIO OF VAR'S RATIO OF HSE'S 6.09 5.71 7.34 6.64 5.76 4.66 4.04 4.48 4.11 4.45 2.84 3.15 3.14 3.34 3.16 2.31 2.59 2.43 2.51 2.35 1.99 2.07 2.14 1.90 2.13 .1.55 1.46 1.59 1.61 1.49 0.32 0.19 0.18 0.12 0.09 0.53 0.35 0.23 0.19 0.17 0.77 0.60 0.41 0.39 0.31 1.50 1.05 0.79 0.68 0.56 1.88 1.58 1.49 1.47 1.45 1.57 1.46 1.60 1.61 1.50 (2) (3) EXPECTED SAMPLE SIZES n1 n2 ~ BIAS ------30 45 60 75 90 30 45 60 75 90 30 45 60 75 90 30 45 60 75 90 30 45 60 75 90 30 45 60 75 90 60 90 120 150 180 50 15 100 125 150 43 64 86 107 129 38 56 75 94 113 33 50 67 84 100 30 45 60 75 90 r" .. ~J "r ..* ~J ------------ -48.2 -49.4 -49.5 -49.8 -50.0 -38.7 -38.4 -39.7 -40.0 -39.4 -28.6 -28.5 -29.7 -29.0 -29.2 -16.4 -19.4 -19.1 -19.8 -20.2 -7.0 -9.6 -9.9 -8.4 -8.8 4.3 2.7 1.9 1.4 1.8 3.9 1.8 2.5 0.1 -0.2 3.5 3.5 0.1 0.8 1.2 0.7 2.6 0.7 2.2 1.6 5.9 2.1 1.2 0.6 0.7 5.8 1.5 1.0 2.8 2.4 6.6 2.9 3.2 1.3 2.5 r" .. ~J r" ..* ~J r" .. ~J r" ..* ~J ---------- ---------- 5.6 0.8 0.2 0.0 0.0 26.6 15.4 5.0 2.0 1.0 53.6 41.0 30.4 25.4 18.8 77 .8 68.8 66.0 56.2 50.2 88.6 88.0 86.6 85.4 84.8 93.2 93.6 94.4 95.6 96.4 21.7 17.9 15.6 13.4 12.8 22.5 19.1 15.2 13.5 13.0 24.4 19.2 16.3 14.8 13.5 25.7 19.7 16.3 15.0 14.2 25.3 19.0 16.6 16.1 13.7 28.7 21.6 18.6 15.0 14.5 COLUMN (5) IS VAR(r" ..*)/VAR(r" ..) ~J CV ~ COV .~J A ..*)/MSE(r ..) COLUMN (7) IS MSE(r"~J ~J 94.8 95.2 93.4 93.8 94.4 94.0 96.0 94.2 93.0 94.4 93.8 94.6 95.8 94.0 95.0 93.0 93.4 96.0 95.0 94.8 95.4 95.0 94.0 95.6 94.6 93.2 94.2 95.6 94.6 96.4 26.7 21.2 20.8 17.3 15.4 28.8 22.8 19.4 17.5 16.4 29.2 23.1 20.1 18.8 16.7 30.8 25.1 20.3 19.0 17.2 31.4 24.4 21.6 19.8 17.8 35.0 26.1 23.2 19.0 17.6 �160 Drummer ESTIMATION OF RATIOS FIGURE 1A - PERCENT BIAS 10 % BIAS e 'StRVED • 0 t i UTIO -30 - + -40 - A.DJUSTED RATIO i ~ ; * -10 -20 - + ~ I + * + -50 -60 0.5 0.6 0.7 O.B nl/Tr2 0.9 1 FIGURE 18 - CONFIDENCE INTERVAL COVERAGE 100 % COV e ISE:RVEO 8 ~ t UTIO e , ADJUSTED ! + 80 - + + 60 + 40 - :t + + 20 0 + + + + + .1 0.5 0.6 0.7 O.B nl/n2 FIGURE 1C - COMPARISON 2.25 + 0.9 1 OF MSE ,",SE(AOJ)!WSE(OBS) 21.751.50 1.25 10.75 0.50 0.25 0 , • •I •• •• 0.5 0.6 • • • I • I • • • 0.7 0.8 nl/Tr2 0.9 • •• • RATIO ¢ �1hl Drummer ESTIMATION OF PROPORTIONS fiGURE 2A - PERCENT BIAS 60 50 40 30 20 10 0 - A~J. PROP. Q "" + _.. .•. .•. -10 -20 -30 1- OBS. P~OP. % BIAS - ••• ••• •• Q •• "''''' 1 •• ~ ..:I .•••• "" , 0.7 O.B ! .2 1.3 n2/TT- ~ A'~ 0.9 1.1 1.4 1.5 FIGURE 28 - CONFIDENCE INTERVA:... COVERAGE I # 0 80 + 60 - !. t:t + *'8- T- + + t + + + + + + + + + * 0.7 O.B + + + + , 0.9 : .2 1.1 • .;- + , I n2/TT- '~H 1.3 ,l. 1.4 + "'SE(AO~)NS£(OeSI 1 • 75- ,! 1 • 501.25 - • • 10.75 - 0.50 - 0.25 - -• •• -•• • 0 0.7 O.B :).9 - I •• • .- , -is 1.1 n2/TT- 1.2 I 1.3 (lAIt.. 1.4 I I 1.5 FIGURE 2C - COMPARISON OF MSE 2 ¢ .;- + + ADJ. °ROP. + + + + 20 0 ~ + 40 - -t CBS. P~~P. % COV 100 • 1.5 I �162 Drummer Detecting Unequal Sampling Probabilities From previous discussion, it is important to determine which, if any, of the classes differ regarding their respective overall probabilities of detection. We desire to test the hypothesis HO : ft1 c ft2 = •.• = k• ft If it can be assumed that all k detection functions are of the same parametric form, then the above hypothesis can be tested by testing HO : c1 = c2 = ... ck ' which can be accomplished by, say, a likelihood ratio test, or other procedures, depending upon the postulated detection function. For example, if one assumes that all detection functions are of the negative exponential type, the hypothesis tests the equivalence of k exponential distribution parameters. At the minimum, the analyst should compare the estimated probabilities of detection for the classes. A number of procedures can be used to test for the equivalence of the distance distributions for the classes, although this is not equivalent to testing the hypothesis of equal detectability. The observed distance distributions are affected by both the overall detect ability and shape of the detection curve. homogeneity. One such test procedure is the chi-square test of The procedure is particularly useful because it can be extended to examine the influence of several factors on detectability through the use of log-linear analysis (Bishop, Fienberg and Holland, 1975). Let g(x,y) represent the probability of detection of an item at perpendicular distanc~ x and possessing attribute Y which may influence detectability. For example, Y might represent an activity level (say a bedded animal vs. a standing animal) or group size if the population of interest consists of groups of individuals. If the attribute Y does influence detect ability, then under fairly general conditions, this attribute and the distance data will exhibit some sort of relationship. If the attribute Y does �163 Drummer not influence detectability, then it is easily shown that X and Y will be independent. This argument extends easily to the case where several attributes, say Y1, Y2, ••• , may be influencing detectability. Thus, an attribute exhibiting a relationship with the distance data may be influencing detectability. The general conditions previously referred to require that the detection function g(x,y) is not factorable. That is, if distance and Y independently influence detect ability so that g(x,y) remain independent. = g1 (x) . g2(y), then X and Y will This may occur at small distances from the transect. For example, if Y represents group size, it may not exhibit much of an influence at small distances. However, at greater distances, it is likely that this influence increases, so that X and Y interact, and g(x,y) is most likely not factorable. �164 Drummer The Clustered Population Problem We assume that the population consists of N identifiable groups, or clusters of items, and that each group may consist of items from more than one of k classes. n umb er 'tems 0f ~ Let the random variable T,~ (i = 1, 2, .•. N) denote the total 'th e ~,th custer, 1 ~n the jth class (j = 1,2, and 1et Yij denote the numb~r of items of ••. k) in the ith cluster, with T, = I: y, '. It is ~ j=l ~J desired to estimate the ratio of the number of items in, say, class A to the number of items in class B, = YA/YB (12 ). ' where YA and YB denote the mean number of items in class A and B, respectively, per group. For example, small groups of large game mammals may consist of animals in various age-sex classes, and it may be desired to estimate the ratio of young to adult female animals. The group size problem in line transect data has been addressed by Quinn (1979), Pollock, (1985), Rao and Portier (1981), and Drummer and McDonald (1987). It is assumed that the probability of detecting a group is a function of the number of individuals in the group and also the perpendicular distance from the transect to the geometric center of the group. In order to estimate the total number of individuals in the population, an estimate of the mean group size, E(T), is required. Since the probability of detection of a group is a function of its size, the observed group size distribution is distorted, �165 Drummer and hence the observed mean group size is a biased estimator of E(T). Quinn (1979) proposes post-stratification by group size as a means to alleviate the group size influence and to obtain an unbiased estimate of the mean group size. The other approaches involve the use of bivariate detection functions which incorporate both the distance and the group size variables. k For the ith detected group of size ti 1: y, , , let ~) E j=1 y., = ~ (yil' Yi2' ..., Yik) denote the group mix, or composition. It is possible that the probability of detecting a group is a function of its mix, as well as its size and distance from the transect. For example, in the previous discussion, the chance of detection could vary with the proportion of young animals in the herd. Let g(') represent the probability of detecting a group from the transect. We consider three cases for the form of the detection function g(') . In each instance, the observed group composition and group size distributions are examined in order to determine an appropriate estimator. The relevant distributions are derived in Appendix A. Case I: g(') = g(x) ; the probability of detection of a group is a function only of its perpendicular from the transect, x, and is independent of both group size and group mix. In this case, both the observed group composition and group size distributions are equivalent to the respective true distributions. Thus, the observed ratio n r" AB = 1: YiA i=1 -..-........n~= YA/YB 1: i=l is the recommended estimator. Y'B ~ The variance of "rAB can be estimated by the usual variance estimate of a ratio estimator. .. " �166 Drummer Case II: g(.) = g(x, t) ; the probability of detecting a gr01lpis a function of its distance from the transect and its total size, but is independent of group composition. In Appendix A, it is shown that both the size and composition distributions are distorted by the detection function. However, the conditional distribution of group composition given group size is not distorted. This suggests that post-stratification by group size will sufficiently adjust for the group size influence. Let m denote the number of group size classes formed by poststratification, with corresponding overall probability of detection 2L th nr = cr (r = 1, 2, ... m). In the r class, nr groups are detected, and x- YrA and YrB are the mean number of items in class A and B per group in the rth group size class. An estimate of the ratio rAE is given by m I: r=l " rAE * = m I: r=l nr YrA " -1 cr " " = NA/N B " -1 nr YrB (oH) cr Quinn (1979) notes that m I: r=l m I: n r . YrA nr . c"r -1 " -1 cr r=l " * is an unbiased estimator of YA (conditional on c"r-1). Thus rAE is the ratio - of two approximately unbiased estimators. An estimator of the approximate variance of r" AB* is given in the Appendix. Case III: g(.) = g(x, X, t) ; detect ability is influenced by both a group's size and its composition. It is shown in the Appendix that unless ~g(x, X, t)dx can be postulated, then the relevant distributions cannot be determined. research. Thus, this case is unresolved and remains a problem for future �167 Drummer An Illustration The study area is in NW Colorado characterized by the sagebrush steppe habitat type. The small community of Great Divide, located roughly in the center of the area has the highest elevation of 2,240 m. The elevation generally decreases in all directions from Great Divide and there are numerous dry and intermittent drainages radiating in all directions. The lowest point of the area is at the village of Maybell with an elevation of 1,804 m. The area is generally semi-arid with cattle and sheep grazing as the predominant land use. The topography is moderately rolling, and, where the terrain permits, winter wheat is grown in very irregular shaped fields. These data are from a modified strip transect survey of pronghorns (Antilocapra americana) conducted by the Colorado Division of Wildlife, during August 1985. A Bell-Soloy helicopter flown 30 m above the ground and at 60 knots served as the observation platform. The area was arbitrarily divided into 3 roughly equal strata to ensure distributive sampling. Within each stratum, transects 1.6-km wide were randomly selected with the constraint that one 1.6-km wide transect must separate any two adjacent selected transects. This was to maintain the independence of the observations on each transect. Approximately one-third of the total area was sampled. The center-line of the transect was the line-of-flight and 0.8-km on each side of the helicopter was searched. The crew consisted of the pilot, an observer/navigator, and the primary observer. When a group of pronghorns was located, the primary observer estimated its distance from the flight-line in 100-m increments and recorded the activity (bedded, standing, or running) of the group when it was first spotted. Unlike the strict line transect methodology, we diverted from the flight-line and pursued the group to within close range to obtain an accurate �168 Drummer count and classification of the group members. Any groups observed while returning to the flight-line were recorded but not included in the sample. Thus, the data could be treated as a line transect data set. The information, by group, was recorded on a portable tape recorder and later transcribed. Flights were begun at sun-up and continued through the day. A total of 237 herds, ranging in size from 1 to 24 animals, were detected during the sampling process. either bucks, does, or fawns. fawns to does were desired. The individuals in the herds were classified as Estimates of the ratio of bucks to does and A large number of herds (64) could not be accurately classified as to activity. Since it was possible that this phenomenon did not occur at random, these herds were included in the detectability analysis. The first phase of the analysis consisted of a log-linear analysis of four variables to determine which of these factors influenced detectability. The four variables examined were herd size, distance, herd activity, and herd composition. herd. Herd composition was defined as the proportion of bucks in the For the purposes of the log-linear analysis, cut-points were established for the distance, herd size, and herd composition variables. Numerous sets of cut-points were tried in order to examine their influence on the analysis. Also, three-dimensional analyses were performed to guard against spurious results due to sparse data. However, the results of the hypothesis tests were not influenced by the choice of cut-points nor sparse data. Based on the tests of simultaneous interaction and marginal and partial association (Brown, 1976), statistically significant relationships were found between the distance and herd size variables, and also between the herd size and herd composition variables. A two-dimensional analysis revealed that larger herds tended to be detected at greater distances from the transect �169 Drummer than did smaller herds, indicating that the probability of detection was an increasing function of group size. Also, the large herds tended, on average, to be comprised primarily of does and fawns, while, smaller herds tended to consist of bucks. Herd activity exhibited no relationship with the distance variable, and exhibited a weak relationship with the group size variable. The three way relationship between herd size, herd composition, and distance was insignificant, indicating that, conditional on herd size, distance and herd composition were independent. It was concluded that post-stratification by herd size could be used to obtain an overall density estimate and estimates of the sex-age ratios. Table II contains the summary statistics for the herd data. Assuming equal detect ability for groups of all sizes, the buck to doe ratio can be estimated by ~ rBD with a standard error of 0.05435. YB/YD = = 0.43 The fawn to doe estimate, unadjusted for group size influence, is ~ rFD = YF/YD = 0.86 with a standard error of 0.0555. Since the does and fawns tended to exist in large groups and therefore were more likely to be detected, one would suspect that the observed buck to doe ratio is low. Table III contains the summary statistics for the post- stratified ratio estimators. (m = Four herd-size categories were defined as 4), and the Fourier Series estimator (Crain et al., 1978) was used to estimate D r , the herd density for the rth herd size class (r = 1, 2, 3, 4). A The program Transect (Laake et al., 1978) was used to obtain Dr and its variance estimate. We note from column (3) of Table III that the estimated overall probability of detection of herds tends to increase with herd size, as �170 Drummer expected, although herds of size 3-4 had the highest estimated probability of detection. A A A In Table III, the symbols DB ' DO ' and OF denote the estimates of the ~A~~ buck, doe and;:~sity, respectively. The post-stratifcation ratio estimates are A A DB/Do = 0.48 (SE = 0.0663) and A * A A rFD = OF/Do = 0.87 (SE = 0.0505) As expected, the post-stratified estimate of the buck to doe ratio is greater than the observed ratio, although the difference here is very small. The disparity between the estimators depends upon the variability in herd sizes, the extent of the herd size's influence on detectability, the degree that members of the different classes associate with one another, and the level of pooling required in order to obtain reliable density estimates. The observed fawn to doe ratio and post-stratified estimates are nearly identical. Recall that fawns and does tended to associate with one another, and thus were sampled with approximately equal probability. The large correlation (0.726) between the number of fawns and does per herd is evidence of this association. The standard errors of the post-stratified estimates are comparable to those of the observed ratios. This comes about because even-though the post- stratified estimators require additional parameters, the post-stratification tends to reduce the variance in the mean herd size variable. is needed regarding properties of this estimator. Further research �171 Drummer Table II SUmmary Statistics for Sex-Age Ratio Estimators Statistic Does - Fawns ~ mean per group 2.26 1.95 0.975 std. dev , 2.58 2.25 1.484 Bucks ~ Fawns 1.000 -.105 -.162 1.000 .726 Correlation Xatrix per herd) (!I Bucks Does Fawns 1.000 Table III Summary Statistics (1) Herd Size for Post-Stratified (2) (3) " of Herds Estimated PR(Detection) Sex-Age Ratio Estimators (4) (5) Estimated Herd Density(SE) 0 Xean 2er He rd Bucks Does YrB r Fawns YrO :rF 1-2 74 0.0834 0.2797 (0.03695) 0.7432 0.3514 0.2703 3-4 68 0.1509 0.1420 (0.02267) 0.5588 1.3529 1.3382 5-7 40 0.0906 0.1392 (0.02463) 1.2750 2.3500 2.3250 8-24 55 0.1309 0.1324 (0.02555) 1.5816 5.8727 4.6727 4 Os = ) D • v r=l r . rB - 0.6742 (SE ; 0.0847) - 1.3951 (SE :II: 1.2079 CSt •• 0.1499) 4 ~ oD K ~ L r=l D'v r 4 ~ D F '\ - . 0 • v ;.. r - rF K 0.1756) r=l ~ * rFD ·rD = ~ OF/DO - 0.87 (SE ••0.0505) • �172 Drummer APPENDIX A Derivation'of Distribution Results for the Clustered Population Problem Let variable Y. h(t) T. denote the probability density function for the group size and let We note that g(y) g(y) can also be represented through the use of a contagious distribution. g(Y) where fey: t) E denote the density function for the group mix That is. ff(ylt)h(t)dt (1) • 1s the density function of the group mix conditioned upon group size. Let (x , w(x. y. denote the joint observed, or weighted distribution of t) Y, t), .,ith corresponding marginal distributions Case I: g(o) = (y. t) , w2(~) • and g(x) From the theory of .,eighted distributions w(x, y. t) '10'1 (~ahfoud and Patil. 1982); = g(x)f(y't)h(t) :: ..':5(X)f(y t)h(t)dxd..:::. dt = g(x)f(y:t)h(t) (2 ) c where c = :g(x)dx. The joint observed distribution of (.:::_. t) is given by (3) which yields 'W (v) 2 ~ = g(y) - and w (t) 3 z h(t) . Thus, i~ this case, the group size and group mix distributions are not distorted by the sampling procedure. Thus YA and YB are unbiased estimators of YA and YB, respectively. �173 Drummer Case II: g(.) ·g(x,t) We aga~n consider the observed distributions. From the previous discussion, w( x -r- t) • -:-::-_-=-g_(-:-x_, _t)~f_(1.~1 t_)_h_( t_) _ ff ... fg(x,t)f(v'" it)h(t)dxdv dt _ = = where ~(t) == g(x,t)f(v't)h(t) "-:. f(y·t)h(t)Ht)d~ dt g(x,t)f(;lt)h(t) fg(x,t)dx and (4) e e = :f(~:t)h(t) ·:'(t)d~ dt. JO ••• This yields the following marginal distributions: wI (r, t) = f (.:::_: :)h(t) ~(t) _- e .·:(':::_!t)h(t) ~(t)dt ( 6) e and w (t) = :f(Z.:t)h(t)~(t)dv 3 = h(t)~(t) e e It is evident that both the group mix and group size distributions torted by the sampling procedure. distri~ution ( 7) are dis- Consider, however, the observed group mix conditioned on group size. w 4 (v:t) = f(v't)h(t)!(t)!8 .•.... h(t):;'(t)/e From (5) and (7). = f(~.t) ( 8) Thus, the conditional distribution of sroup composition given group size is not distorted ~y the sampling procedure. ratio esti~ator. This suggests the post-stratified �174 l.)rummer Case III: g(.) - g(x.y,t) Let 'f(l..t)- fg(x,z,t)dx, and let y - :: ••• fg(x,Z,t)f(Z!t)h(t)dx d_y. Then w(x,y,t) - g(x,Z,t)f(ylt)h(t) y wl(y,t) - '-¥(v,t)f(y't)h(t) y w2(z) • :f(Z,t)f(y!t)h(t)dt y w 3 (t) = h (t): ~(y.t)f (: 't) d::::. and y This yields w 4 (vit) ._, = -ev,t)fev '~(l' Thus, unless the f cr-a of distributions t) . ~C.;:_, t )f(Z-t....• )~d-.;:_ t) = :g(x, y. t)dx is known, the desired cannot be determined. APPE:\DIX B Variance Estimation for Post- Stratified Ratio Estimators Let m n herds in the with r denote the number r th of post-stratification class (r = herd size classes, m). 1, 2, ... v - rA Let and ? S- rA denote the sample mean and variance of the nu~ber of class A items per herd in the r th herd size class. Let D r denote the estimated density of ~ o. herds in the rth herd size class, with of the overall density of class A items. independent, and that observations independent, then D = A r=l D r . Assuming that .'rA , the estir.tace D r and YrA in different herd size classes are are �175 ". Drummer m m Var(DA) • Var( I D·y )r rA r-l I r-l Var(D r .y rA) (1) + Var(D ).Var(Y ) } , r rA where Var(yrA) • S;A/nr. The desired ratio estimator is with approximate variance estimated by ~* Var(rAB ) ~* = ( rAn ) 2 { Var ~ (D.;) ~ . + ? (D )A - 2 Cov( DA, DB) ( 2) CDA • DB) m !!l ; \' - D . D·v : Cov(YiA, YjB) = 0 for r=l By assumption, (i, j) E Cov(Di, D ) j { 1, 2, ... m } c r V • _ - r A.' r= 1 • Taking expectations (and cancellation) yields m + I r=l Var(D ) • E(YrA) • E(YrB) r r - rB i * j , where �176 Drummer Substituting the observed sample moments yields m cOv(DA' DB) · r-l I [C~v(yrA' m + I Var(D ) r rzl ~ ~ YrB) . { Var(D ) + r } . YrA . YrB (3) ~ ~ Var(D ), Var(D ) , and CQV(D , DB) A B A the desired estimate. Substitution of (Dr ) 2 into (2) yields �177 Drummer Literature Cited Bishop, Y. M. M., S. E. Fienberg, and P. W. Holland (1975). Discrete Multivariate Analysis. MIT Press, Cambridge, MA. 557 pp. Brown, M. B. (1976). Screening effects in multi-dimensional contingency tables. Applied Statistics (25), 37-46. Burnham, K. P. and D. R. Anderson (1976). Mathematical models for nonparametric inferences from line transect data. Biometrics 32: 325-336. Burnham, K. P., D. R. Anderson and j. L. Laake (1980). Estimation of density from line transect sampling of biological populations. Wildlife Monograph No. 72. 202 pp. Crain, B. R., K. P. Burnham, D. R. Anderson and J. L. Laake (1978). A Fourier Series Estimator of Population Density for Line Transect Sampling. Utah State University Press, Logan, Utah. 23 pp. Drummer, T. D. and L. L. McDonald (1987). Size-bias in line transect sampling. Biometrics 43(1): 13-21. Gates, C. E., W. H. Marshall, and D. P. Olson (1968). Line transect method of estimating grouse population densities. Biometrics 24(1): 135-145. Gates, C. E. (1969). Simulation study of estimators for the line transect sampling method. Biometrics 25(2): 317-328. Laake, J. L., K. P. Burnham and D. R. Anderson (1979). User's manual for program transect. Utah State University Press, Logan, Utah. 26 pp. Mahfoud, M. and G. P. Patil (1982). On Weighted Distributions. In Statistics and Probability: Essays in Honor of ~ ~ Rao, Kallianpur, ~, P. R. Krishniah and J. K. Ghosh (eds.). North-Holland Publishing Company, New York. pp. 479-492. Pollock, K. H. (1985). Unpublished manuscript; 12 pp. Quinn, T. J. II (1977). The effects of aggregation on line transect estimators of population abundance with application to marine mammal populations. Thesis. University of Washington, Seattle, Washington, USA. Quinn, T. J. II (1979). The effects of school structure on line transect estimators of abundance. In Contemporary Quantitative Ecology and Related Ecometrics, Patil, G. P. and M. L. Rozenweig (eds.). Internatt. Co-op. Publ. House, Fairland, MD. pp. 473-491. �178 Drummer Quinn, T. J. II and V. F. Gallucci (1980). Parametric models for line transect estimators of abundance. Ecology 61: 293-302. Rao, P. v., K. M. Portier, and J. A. Ondrasik (1981). Density estimation using line transect sampling. In Estimating Number of Terrestial Birds, Ralph, C. J. and J. M. Scott (eds.). Studies in Avian Biology No.6: 441-444. Cooper Ornithological Society, Allen Press, Lawrence, Kansas. Seber, G. A. F. (1982). The Estimation of Animal Abundance, MacMillan Publishing Company, New York, New York, 654 pp. �179 APPENDIX B EVALUATION OF CENSUS TECHNIQUES Thomas M. Pojar, David C. Bowden, FOR F'RONGH2RNS and Bruce Gill OUTLIN~ I. Introduction ~. Status and need. B. Crecits II. Description of study arE~S A. Shortgrass pralrie B. Sagebrush steppe III. Methods A. Transect strip width B. Sampling systems and intensities 1. Strip transects anj line transe=ts 2. Random quadrats C. Searc~ techniques a,d data recording 1. Strip transects 2. Random quadrats 3. Line transects D. Test fer accuracy in quadra~ SE~rch IV. Results A. Transect strip width 1. 800m width vs. 1600m width 2. Comparison with data from line tr~ns2cts shortgrass prairie. B. Strip transect vs. random quadrat 1. Population size and preCision estimates 2. Herd structure and precision estim~te5 3. Differences in results from two habitat types C. Quadrat search accuracy V. Conclusions and recommendations VI. Literature Cited in ��Colorado Division ot W11d11ie Wildlife Research Report July 1988 181 JOB PROGRESS REPORT State of Colorado Project No. W-153-R-2 -----------------------No. 3A ---------------------- Mammals Research Work Plan Pronghorn Investigations Habitat Selection and Population Performance of a Pioneering Pronghorn Population Job No. 2 Period Covered: July 1, 1987 - June 30, 1988 Author: T. M. Pojar ABSTRACT Except for rate sightings of 1 or 2 pronghorns (Antilocapra americana), Middle Park Colorado has been devoid of this species since the 1920s. Beginning in the early 1970s, pronghorn sightings were reported with increasing frequency, and by 1979 pronghorns were reported seasonally throughout the year. The largest number sighted since the report of 7 pronghorns in the 1920s was on 21 August 1979, when a group of 13 was seen 10 mi north of Kremmling in the area of Antelope Pass. The population expanded through reproduction, which indicates that this is likely a bona fide pioneering popultion. As of December, 1986, the herd totaled 80 animals with a structure of 36Bucks:lOODoes:77Fawns (determined from trapped animal). A portion of the herd was trapped using a corral trap, and 9 radio collars were put on yearling and adult females, the remainder of the adult does were collared with numbered, yellow, plastic collars and numbered eartags. Five of the 8 yearling/adult males captured were neck collared and eartagged; the remaining 3 males and all fawns were eartagged only. .In total, 47 animals were trapped and marked. The radioed animals have been.located bi-weekly beginning in January, 1981' (ca 40 locations. .on each animal). One radioed doe was killed in an automobile collisibIi·0014 July·19B7 ab~ut,l mi east of KremUu.ing•. AIiother radio ceased to function in April, 1988. In .October, 1987, a sample of 63 pronghorns were classified from the ground; the ratios were 54B:lOOD:7lF. A ground count of the total popul.atIon on 14 December 1987 was 122, which was verified by aerial photography frow a helicopter within 2 hrs of the ground count. ��183 HABITAT SELECTION AND POPULATION PERFORMANCE OF A PIONEERING PRONGHORN POPULATION Thomas M. Po jar P. N. OBJECTIVE Describe population dynamics and habitat use of a pioneering, expanding pronghorn population. SEGMENT OBJECTIVES 1. Describe seasonal and annual distribution of the Middle Park pronghorn population. 2. Determine sample sizes of radio-collared animals and observations necessary to describe habitat preferences. 3. Monitor population dynamics of Middle Park pronghorns with: a. Ground counts to describe changes in population size. b. Ground counts to quantify population sex-and-age composition. ACKNOWLEDGMENTS The trapping operation was facilitated through the efforts of S. Steinert, J. Gerrans and Area 9 personnel, R. Spowart, J. Claassen, R. Firth, A. Chappell, B. Thompson, B. Sigler, M. Middleton, J. Hicks, and J. Frank. C. Cesar and other personnel from the Kremmling Bureau of Land Management District provided able assistance during trapping. J. Hill permitted the trapping operation on his property. J. Ritchie, B. Adrian, and M. Miller assisted in collection and handling of pronghorn blood samples. J. Ellenberger and the CDOW NW Region contributed in the form of helicopter time and 5 radio transmitters. STUDY AREA Middle Park is a·high mountain park that is bordered on the west by the Gore Range, on the north by the Rabbit Ears Range, and on the east and southeast by the Continental Divide (Tiedeman et ale 197). Kremmling (2,361 m) is more or less centrally located and is the major municipality in the Park. The lower elevation of the Park is predominantly big sagebrush (Artemisia tridentata)/western wheatgrass (Agropyron smithii) plant community and ranges in elevation from 2,200-2,750 m with slopes from 0-20%. Aspen (Populus tremuloides) and conifer communities are found above 2,750 m (Tiedeman et ale 1987) • �184 Precipitation averages 35-40 cm annually and is evenly distributed throughout the year. Mean annual temperature is 4.60C (Tiedeman et ale 1987). Severe winter conditions occur periodically in Middle Park with deep snow and long perods of extremely cold temperatures. These conditions result in population mortality of 50% or greater on other ungulate species in the park (Carpenter et al. ND). METHODS AND MATERIALS The pronghorns were trapped using a coral trap following the trapping and handling procedures sanctioned by the Pronghorn Antelope Workshop (Anonymous 1984). Tracking was done either from the ground or from an airplane. In either case, animal location was recorded on a 15 min topographic map to the nearest quarter mile. Herd structure estimates were made from the ground during late summer and early autumn using 8 x 44 binoculars and a 20-40X spotting scope. Attempts were made to classify at least 50% of the population. Total population size was estimated from the ground in winter when it was assumed that all Middle Park pronghorns were on the wintering area. RESULTS Trapping and Marking The trapping operation was conducted 16 December 1986, and 47 animals were captured. Radio transmitters (Telonics) were put on 9 yearling/adult does. Light-weight, yellow, plastic collars with numbers were put on all remaining yearling/adult does. This same type of collar was put on 5 of the 8 yearling/ adult bucks captured, and the remaining bucks and all fawns were eartagged with numbered yellow tags (Appendix A). All 47 animals were marked in some fashion. There were no trap mortalities or injuries, and all of the neck-collared animals were verified to have survived at least through September, 1987. All of the radioed does were alive as of 30 June 1988 with the exception of radio frequency 149.370. This animal was killed by an autmobile on a county road about 1 mi northeast of Kremmling on 14 July 1987. The doe with radio frequency 149.290 has been verified as being alive, but the radio ceased functioning during April, 1988. Observations of Marked Animals Pronghorns with yellow eartags were observed in the Grandby area, which is approximtely 30 air mi east of the trap site. In early May, 1987, Bob Thompson (DWM Kremmling) reported seeing 4 animals, 3 with eartags, North of Grandby; and Mac and Mrs. Rienhoff reported 4 pronghorns, 3 does, and 1 buck (horns about the length of the ears), all with eartags. In early June, 1988, Ron Scott (USFS) saw "about" 12 pronghorns near Grandby. He did not notice any eartags or collars, but he only observed them as he drove past in a vehicle. An unmarked buck was apparently killed by a vehicle on Highway 9 near Green Mountain Reservoir in late August, 1987) (J. Liewer, pers. comm.). The location of the kill (TlS,R80W,Sec 2) is about 12 air mi from the trap site. �185 A doe with ear tag number 23 was observed near Coalmont (T7N,R8lW,Sec 24) (S. Steinert, pers. comm.), which is in North Park, about 35 air mi north northeast of the trap site. This is the first evidence of interchange between Middle and North Park, although it is presumed that the original Middle Park pronghorns immigrated from North Park. Locations of Radioed Animals All of the radioed does remained within 10 mi of where they were trapped, most much closer. Based on about 40 location points for each animal, the movements were mostly to the north (Figs. 1 and 2). The legal description of each location will be converted to points on a Universal Transverse Mercator grid (U.S. Army 1973). The task of converting to the UTM grid is completed for locations through November, 1987. When the conversion process is completed, this series of points will be used in the program McPAAL (Stuwe NO) to delineate the seasonal and annual area of habitation for each animal based on the minimum convex polygon. An example of the results of this procedure is shown in Figure 1, which includes location points for the doe with radio frequency 148.390 from January through November 1987. The winter (Jan-Mar) area of habitation for this animal was 7.5 kmi (2.9 mi2) and the yearlong (Jan-Nov) area was 66.2 km2 (25.6 mi2). The distribution of radioed animals on 16 June 1987 is shown in Figure 2. All of them, except 2, are within 3 mi of the trap site. Typical winter distribution was within 3 mi of the trap site for all of the animals (Fig. 3). Population Size and Structure Prior to the trapping operation, the presumed total population in Middle Park was 80 animals based on counts of the major group inhabiting the wintering area. The sex-and-age ratios of the 47 trapped animal was 36B:lOOD:77F; if it is assumed this is representative of the population, then there were 14 bucks, 37 does, and 29 fawns in December, 1986. The winter population for 1987-88, based on several ground counts and aerial photography (14 January 1988), was estimated to be 122 (Fig. 3). The herd structure obtained from classification (ground) of 63 animals on 14 October 1987 was 54B:lOOD:7lF. Applying these ratios to the population of 122 gives 29 bucks, 55 does, and 38 fawns. Of the 55 does, 40 would be of reproducing age, which gives a fawn:lOO mothers ratio of 95:100. During the winter of 1987-88, several attempts were made to get a total count of yearling and mature bucks. The highest count of 25 was obtained on-14 January 1988. Assuming that this count is more accuratp. than extrapolating from the 14 October buck:doe ratio estimate, then the total number of does and fawns would be 97 (122-25-97). This could be apportioned into 40 fawns (20 males and 20 females) and 57 does using the 14 October classification of 7lF:lOOD. This restructuring of the population results in ratios of 44B:lOOD:70F. Using these ratios and assuming no natural mortality (which is not realistic), the population size and structure of the fall 1988 population can be projected. In June of 1988, yearlings and 57 tribute 95 fawns by fall of 1988, the female side of the population would be composed of 20 reproducing-aged animals. If reproducing-aged females conper 100 does, then 54 fawns will be added to the population and the total will be 176 animals composed of: �186 Bucks Does 25 Mat. 20 Yr1. 57 Mat. 20 Yrl. 54 45 77 54 Fawns Total 176 If it is assumed that the count of 80 pronghorns in December of 1986 was, in fact, the total population in Middle Park and that there was no immigration or emigration, then the annual rate of increase between 1986 and 1987 was 55%. LITERATURE CITED: Anonymous. 1984. Management guidelines: Trapping and translocation. of Pronghorn Antelope Wkshp. 11:237-251. Proc. Carpenter, L. H., R. B. Gill, D. L. Baker, and N. T. Hobbs. ND. Colorado's big game supplemental winter feeding program, 1983-84. Colo. Div. Wi1d1. Xerox, Appendix. 92pp. Stuwe, M. ND. McPAAL micro-computer programs for the analysis of animal locations. Program Documentation. Cons. and Res. Cntr., Nat1. Zool. Park, Smithsonian Inst., Front Royal, VA. Printout. lapp. Tiedeman, J. A., R. E. Francis, C. Terwilliger, Jr., and L. H. Carpenter. 1987. Shrub-steppe habitat types of Middle Park, Colorado. USDA For. Serv., Res. Paper RM 273. 20pp. U.S. Army. 1973. Technical manual: Universal transverse mercator grid. H.Q., Dept. of Army, Washington, DC. No. 5-241-8. 64pp. ™ Prepared bY'ib&, e~~ Thomas M. P jar Wildlife Researcher z �187 Figure 1. Winter (Jan-Mar, solid line) and year-long (1987, dashed line) home range of mature pronghorn doe with radio 148.39~. Minimun convex pqlygon (Program McPAAL, Stuwe ND) is 2.9 miles 2 (7.5 km2) for winter and 25.6 miles2 (66.2 kmf for year-long area of habitation. �188 KRE""'LI~;C QUADRANGLE ':~:"jP":"J-C~"'''': .::- _~ 0j)2 ".--~ .. 11fS.~ ... ",1-': - . ..r-~_:-' =>:..•. Figure 2. Location of the radioed pronghorns on 16 June 1987. This distribution represents typical summer dispersal. The numbers are radio frequencies. �189 'J ~~., D,-"-' /"',,: ... 13:':1 '" '-'j "'. .,...-tI .'..;; r\ ~ 18/ a"", .•..•••.• 'I'~ ......._ I .~. -.. __-------...::,_,-~jl . -.;;<'- II ;,---) [1"'\ I .7656. ~ - •.~ ''-, ..-. -, iI .~ '-:'[;.' " \,""J"! :_;,....!~,~. ':> i -: ;-;;-;.:-,:--UT/2 .•.• _u 124 I' \ ,: 1ci ,/ J " P, •.~•.• \.c·'""t' ,_:...- . .' - ..,., - .,'" ---- :-'- --.ii:i,i;::L--1:;,:_;" ~'I .~:n =]. --~- - ~- : !"r-'==r-=-~ , Kremmling • ,/ , /~ j,;c. _ L Figure 3. Dates and total counts of the wintering group of pronghorns during January, February, and March 1988. All radioed animals were in this group. ��APPENDIX A 191 BANDING AND TAGGING RECORD Species: Pronghorn (Antilocapra americana) Date: 16 Dec 1986 Trap site Name:~M~i~d~d~l~e~P~a~r~k~ _ Trap site Location:~C~u~r~ry~~R~a~n~c~h~ _ Legal Description: R80W,T2N,Sec 36 (SW 1/4) Nearest Town:~K~r~e~mm==1~1~'n~g~ ~ capture Method: Wing Trap wI curtains Purpose: Define Tag Number Neck Ear seasonal and annual home range Sex Age Comments 1 2 3 4 radio radio none radio F F M F A A F A Blood yes yes yes yes 5 7 1 3 radio M F F A A A yes yes yes 8 9 2 radio M F Y A yes yes 10 radio F A yes 11 none F A no 12 radio F A no 13 14 15 16 17 18 19 20 radio 4 5 11 6 7 8 radio F F F F F Y A A A A Y A A yes yes yes no 21 22 23 24 25 9 10 none 12 13 Y A F A Y yes no no no no 6 M M F F M F F F _ Freq. 148.420 Freq. 148.390 Obs. Freq. 148.400 12/17/86 - ok Observed 12/17/86 - ok Freq. 148.440 12/17/86 - ok Obs. Obs. Freq. 148.430 12/17/86 - ok Obs. Freq. 149.210 12/17/86 - ok Weak when released - shock and/or exhaustion. Obs. Freq. 149.250 12/17/86 - ok Freq. 149.370 ye'F yes yes yes Freq. 149.290 12/17/86 - ok Obs. �192 Tag Number Ear Neck 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 14 none none none none none none 15 none none none none none 16 17 none none none none none none none Sex F F F A F F M A F F F M M F F Comments Age A Y F M M Y M F F F F M F M F F F F F F A A F F F F F F F Blood no no no no no no no no no no no Void tag number no no Tag in left ear. no no no no no no no no no SUMMARY OF CAPTURE Fawns 6 MALES Yearlings 3 Mature 5 = 14 FEMALES Fawns Yearlings Mature 11 4 18 = 33 Total pronghorns trapped = 47 RATIOS OF CAPTURED PRONGHORNS BUCK:I00DOE = 36.4 FAWN:I00DOE = 77.3 �Colorado Division of Wildlife Wildlife Research Report July 1988 193 JOB PROGRESS REPORT State of Colorado Project No. W-153-R-2 -----------------------No. 6A ---------------------- Mammals Research Work Plan Mountain Lion Investigations Job No. 1 Period Covered: July 1, 1987 - June 30, 1988 Author: Mountain Lion Population Dynamics A. E. Anderson Personnel: See Acknowledgments ABSTRACT Five puma (Felis concolor) were captured for the first time, and 5 radiocollared puma died. Seven, mature radio-collared males were recaptured for blood, tissue, and semen sampling, and a mature female was recaptured and recollared. Since 1981, 57 puma have been handled, and 29 have died. The major causes of death were sport hunting outside the study area (7) and capture (9). The mean + SD number of days from capture to death for 11 male and 9 female puma were 307.0 + 221.2 and 458.2 + 358.6, respectively. Young «24 months) puma comprised 61.4% of the total sample. Sex ratios for individual and combined age classes did not differ (P < 0.05) from equality for samples of either living or dead puma. ��195 MOUNTAIN LION POPULATION DYNAMICS Allen E. Anderson P. N. OBJECTIVE To assess the effects of sport hunting on mountain lion populations. SEGMENT OBJECTIVES 1. Capture and mark up to 12 mountain lions. 2. Monitor mountain lion movements. ACKNOWLEDGMENTS I thank G. Bock, D. Bowden, G. Byrne, G. Cheney, D. Coven, T. Davis, J. Ellenberger, V. Giba1di, J. Gray, L. Gray, L. Green, G. Hanson, M. Hargis, C. Heacock, R. Henkle, M. Herschcopf, J. Humphrey, J. Kane, D. Kattner, D. Masden, D. McCauley, K. Miller, J. 01terman, S. Parker, B. Poehnert, M. Potter, M. Roelke, D. Rowley, P. Shearwood, M. Stevens, N. Ting, A. Wacker, and J. Wolfe for their assistance. I am also grateful to D. Miller, VicePresident, Research and Education Department, National Wildlife Federation for facilitating financial assistance from the N.W.F. D. Kattner, Project Puma Hunter, deserves recognition for his exemplary efforts as does M. Roelke and staff of the "Collaborative Wild Cougar Project." METHODS AND MATERIALS Methods were described in Anderson (1983a). Eighty-six days were spent hunting from November 18, 1987, to April 29, 1988. Puma were located with aerial telemetry at approximate weekly intervals. Only those locations subjectively judged, on the basis of signal quality, as being within 1.5 km of the actual location are presented in this report. Locations were plotted (+ km) as Universal Transverse Mercator coordinates on USGS county (1:50,000 scale) topographic maps. In an assessment of location accuracy, the arithmetic mean, SD, median, and range of 25 map distances (+ 0.1 km) between estimated and actual point locations were 0.55, 0.88, 0.20, and 0.0 - 4.5 km, respectively, on 5 livi~g and 8 dead radio-collared puma. Capture effort was allocated proportion~l to surface area among 4 pre¥!~usly established strata (Fig. 1 and Table 1). The intent was to ensure that each resident puma might have a similar probability of capture. RESULTS AND DISCUSSION Allocation of Capture Effort As in previous years, we did not achieve the desired proportional allocation of capture effort (Table 1). Variation in hunting conditions and accessibility among strata was largely responsible. �196 Puma Capture, 1987-88 We captured, radio collared, and released 5 puma for the first age, 14 days were required to radio collar 1 puma using hounds Details on their capture and body measurements plus those of 3 handled are in Table 3. In addition, 7 mature radio-collared, recaptured for sampling blood, skin, and semen by personnel of "Collaborative Wild Cougar Project." Puma Telemetry, time. On aver(Table 2). additional puma male puma were the 1987-88 Twenty-six puma were radio tracked during this fiscal year (Table 4). However, only 14 of these were tracked yearlong; the usual locations of 6 were outside the study area, and by 7-14-88, 5 had died. A mature male (#42) was recaptured and released in GMU 25 about 112 km from its capture site in Cottonwood Canyon by P. Shearwood, a volunteer houndsman, D.W.M. L. Green, and others on 1-21-88. Subsequently, Wildlife Biologist G. Byrne obtained 8 aerial locations of #42 from 2-18-88 to 7-12-88 while radio tracking elk in GMU 25 and vicinity. Individual aerial locations for each of the 26 puma are listed in Appendix. Puma Telemetry, 1981-88 Fifty-seven puma have been handled since the study began. They include 25 puma who were (as of 7-14-88) either alive or fate unknown (Table 5) and 29 dead puma plus 3 puma whose deaths were suspected but never verified (Table 6). Only 1 of 9 puma caught in leghold traps, ostensibly set for bobcat or coyote by recreational trappers, was large enough (> 27.4 kg body weight) to radio collar. Hence, marking those animals was generally limited to ear tattoo numbers. The 2 young puma killed by our hounds during pursuit were not marked. They are identified by their dates of death. Puma Mortality, 1981-88 Twenty-nine or 50.9% of 57 puma handled since 1981 have died (Table 7). Among 8 generalized sources of puma mortality (capture-related, legal kill, and unknown causes), accounted for 9, 7, and 6 or 75.9% of confirmed deaths, respectively. Shaw (1977) reported that 11 of 16 puma (68.8%) died and that capture-related causes and sport hunting accounted for 4 and 7 deaths, respectively. His study conditions were similar to ours; that is, the study area was closed to legal hunting, but radio-collared puma were vigorously hunted outside the study area. There were no differences- (p > 0.05) in the equality of the sex ratios of each age class or for both age classes combined. Thus, it is unlikely that the tabulated mortality factors were sex specific. Age-and-Sex Classes, 1981-88 Age-and-sex classes of annual captures include all puma handled (Table radio-collared puma (Table 9), and the survivors radio tracked in 1988 10). Annual sample sizes were highly variable and generally too small analysis of sex ratios within individual or combined age classes. The quencies presented are likely influenced by our deliberate and largely 8), (Table for fre- �197 successful selection of puma weighing at least 27.4 kg for pursuit with hounds. For example, the different age class frequencies of female puma between Tables 8 and 9 are largely due to the exclusion in Table 9 of trapped puma whose body weight was generally >27.4 kg. In Table 8, 35 (61.4%) were <24 months of age and 22 (38.6%) >24 months whereas in Table 9, 27 (55.1%) were <24 months of age and 22 (44.9%) at >24 months. As in our sample of 29 dead puma (Table 7), there were no differences (P < 0.05) either in the equality of the sex ratios of each age class or for combined age classes (Tables 8 and 9). Logan et ale (1986:651) tabulated sex-and-age classes of 58 puma estimated from about 46 captures using hounds. They reported 29 (50.0%) >24 months and 29 (50%) >24 months of age in contrast with our data (Table 7) of 35 (61.4%) and 22 (38.6%) puma, respectively. However, a test for independence of the Logan et al. and our data was not" significant (X2 - 1.95, df - 1, P > 0.05). Logan et al. (1986:650) stated: "The high proportion of young lions on the study area suggested a high reproductive rate, rapid population turnover and frequent immigration of young transients." Whether such an interpretation is valid for puma is unknown. It is instructive to observe that even with more <24-month-old puma captured, this age class comprised only 9.5% of the total known survivors (Table 10). Estimates of Survival Telemetry data can be used to estimate survival (Trent and Rongstad 1974, White 1983, Sievert and Keith 1985). Some of the essential vital statistics for such an effort are listed in Table 11. Mean age and number of days monitored are described statistically for 20 male and 21 female puma captured in GMU 62 (Table 12). The mean number of days from capture to death was 307.0 for 11 males and 458.2 for 9 females. About 21 male and 23 female puma are available for estimates of survival to be presented in future publications. Home Range A manuscript on the dynamics of home range of 2 male and 5 female resident puma through March, 1986, was submitted to the Canadian Journal of Zoology on March 29, 1988. Other than an acknowledgment card dated April 6, 1988, nothing has been learned from the Editor as of July 30, 1988. The strongest data for future analysis of resident puma home range characteristics are identified in Table 13. Cervids Killed by Puma Eight mule deer killed by puma included 1 whose sex could not be determin~d and another whose" sex and age were not determined (Table 14). Of unusual interest is the l6-l7-month-old female found 11-28-87. A st~dy of tracks and the use of radiotelemetry by D. Kattner indicated a small puma had actually killed the deer and ate the heart, lungs, liver, and some ribs. That night, however, the much larger #52 ate much of the remainder. Four days later, only a portion of hide and lower legs remained. Only the smaller puma track was found in the vicinity. Also on 3-3-88, #56, a 35.4 kg female puma, killed a large, mature female mule deer but fed little on the carcass. On 3-4-88 some viscera had been eaten, and 4 days later only the skeleton and hide remained. Tracks in the vicinity indicated that only one puma (#56) had probably fed on the carcass. �198 Since 1981, 68 mule deer and 3 elk carcasses have been recorded as puma victims within GMU 62 (Table 15). Both sex and age could be identified for 48 deer carcasses. Sex ratios differed (P - 0.95) from equality for deer <17 months of age (40:100), 18+ months of age (17.4:100), and for the total sample (26.3:100). "Collaborative Wild Cougar Project" The goal of this Florida-Washington, D.C., based study is to examine genetic diversity among subspecies of Felis concolor and how the endangered Florida form compares to other presumably heterogeneic species. The project is staffed by M. Roelke, D.V.M. - Florida Panther Project Veterinarian, Florida Game and Fresh Water Commission, Gainesville, Florida; S. J. O'Brien, Ph.D. - Laboratory of Viral Carcinogenetics, National Cancer Institute, Frederick, Maryland; and D. E. Wildt, Ph.D.; E. Jacobson, D.V.M., Ph.D.; and J. G. Howard, D.V.M. - Department of Animal Health, National Zoological Park, Washington, D.C. M. Roelke, assisted by S. Parker, and V. Gibaldi, D.V.M., spent 12 days with us from April 17, 1988, to April 29, 1988. D. Kattner and myself (assisted by Roelke, Parker, and Gibaldi) captured 7 mature, male, radio-collared puma, and they processed the puma for blood samples, skin biopsies, and semen samples. In addition, Dr. Roelke performed a field postmortem and obtained clotted blood and tissue and fecal samples from #50 who had died from unknown cause(s) during March, 1988. I expect to receive copies of pertinent data during the next fiscal year from M. Roelke. Because of health and person reasons, I plan to retire July 1, 1989. I will spend the next year analyzing data and preparing publications on relevant information gained during this study. LITERATURE CITED Anderson, A. E. 1983a. Program Narrative Proj. 45-01-503-15050, Work Plan 6, Job 1, Mountain lion population dynamics. 7pp + 3 tables and Appendix A. Colo. Div. Wildl., Fort Collins. 1983b. A critical review of literature on puma (Felis concolor). Colo. Div. Wildl. Spec. Rep. 54. 9lpp. Ashman, D. et al. 1983. The mountain lion in Nevada. Nev. Fish and Game Dept., Fed. Aid Wildl. Restor., Final Rep., Proj. W-48-l5. 75pp. Logan, K. A., L. L. Irwin, and R~ L. Skinner. 1986. Characteristics hunted mountain livn population. J. Wildl. Manage. 50:648-654. of a Sievert, P. R., and L. B. Keith. 1985. Survival of snowshoe hares at a geographic range boundary. J. Wildl. Manage. 49:854-866. Shaw, H. G. 1977. Impact of mountain lion on mule deer and cattle in northwestern Arizona. Pages 17-32 in R. L. Phillips and C. 'Jonkel, eds. Proc. 1975 predator symposium, Mont-.-For. and Cons. Exp. Stn., Univ. Montana, Missoula. �199 Trent, T. T., and o. J. Rongstad. 1974. Home range and survival of cottontail rabbits in southwestern Wisconsin. J. Wildl. Manage. 38:459-472. White, G. C. 1983. Numerical estimation of survival rates from band-recovery and biotelemetry data. J. Wildl. Manage. 47:716-728. Prepared by C. 4-n~ ~ Allen E. Anderson . Wildlife Researcher �200 rab1e 1. ChroDOlogyof p.-a huntil18 effort by .trata, DIlcoapabgrePlateau (QfU 62), represents 1 day of hunting effort. Underlined dates are explained in footDOtes 1987-88. Iach date Strata and date. Month 2 1 o o Hov 1 27 2,8,9,10,11, 4 3 10 12,15,29,30, :n .r Other GMt] Grand total r 4 o o 21 18,24,25 3 28,.!! 2 16,17,18, 19,22,23, 26 7 3,13,20,21 6 o 2,4,8,11,12, 13,14,26 8 18 1 25 1 o 10 Feb 17,18,24 3 4,5,8,11,12, 13,15,16,29 9 10,23 2 o o o 14 Mar 2 1 3,4,8,11,14, 23,29,30 8 24,25,31 3 o o o 12 Aprc 4,6,7,11,12, 28,29 7 5,13,26,15, 16 - 17,18,19, ! ! 23 1 86 rota1 11 ~capture and recollared b Capture and kill .2 4 41 20 21,22,23 27 13 112 on 12-11, 12-87. .heep killer 129 on 1-18 and iudvertent treeing of $52 on 1-25 and 11-29. cRecapture and assist in obtaining ti.sue, blood, and se.en samples fro. 7 aature, aale, radio collared puu in cooperation with M. &oelke, D.V.M., ·Collaborative Wild Couaar Project. 147 recaptured and reco1lared on GMU 40 on 4-27-88.· rab1e 2. iate of radiocollaring GMt] 62£ 4-9-81 to 4-15-88. To No. radiocollared No. days hunted 4-29-81 2-21-82 5 -6-83 5-17~ 6-28-85 5-17-85 5- 7-87 4-15-88 1 3 7 5 7 5 14 5 16 32 91 Period 'roll 4- 9-81 12-14-81 1- 2-83 11-19-83 11-19-84 11-18-85 11-17-86 11-18-87 puaa using hounds, Uuco~re 47 III 77 136 97 70b 630b Plateau, ladio collaring rateS 16.0 10.7 13.0 22.2 11.0 27.2 6.9 !hl 13.4 aNo. day. to radio collar 1 puaaj does DOt include replacing or refitting radio collar. or puaa captured by ilethad. other than pursuit by hound. and drua ~blll&atlon. . . bDoeaDOt1nc.lude 4 day•• pent 1114dvertently pur.uing radio-collared puaa or 12 days recapturing 7 rad1o-collared .ale puaa for the ·Collaborative Wild Oouaar Project·; 4-17-88 to 4-29-88. �201 Table 3. Details on the capture and body measurements Plateau (GMU 62). of 8 puma handled during FY 1987-88, Uncompahgre Ear tattoo nuaber Sex Date of capture Est. age (80nths) Legal descr., capture site 1/4,S T R. U.T.M., capture site Elevation (m) capture site Radio collar aerial no. Transmitter freq. (MHz) Drug (Ketamine:rompun, 180: 90 mg/ml) Dosage (cc injected) Induction time (ains) Body we (kg) Measurements (em) Total body length Tail length Head-body length Chest girth Neck circumference Height of shoulder Head length Zygomatic breadth Ear length Hind foot Hind paw length Heel pad, max. dia. Left front, 1engt:h width Left rear, length width Right rear, length width ~captured bProbab1e 53 54 F F 12-19-87 48 1-11-88 14 NEB 48N llW 753-4257 1,981 6571-01 148.7285 NW27 50N 14W 724-4273 1,861 3955-01 148.0600 55 M a 56 57 47 F F M F F 4-15-88 26 4-27-88 24 1-18-88 27 12-12-87 117 NE23 SON 13W 736-4274 1,920 3956-01 148.0700 SW24 103W 135 685-4309 2,316 3964-01 148.1500 SW34 50N 11W 753-4271 1,800 SE3S 15 13 734-4286 1,768 12903-01 149.9000 2-5-88 30 .3-3-88 14 NW2 49N 13W 736-4269 2,073 12902-01 149.8890 SE36 51N 14W 728-4280 1,859 3966-01 148.1810 12 3.0 4.0 3.0 5.0 shot 49.0 3.8 60.0 36.3 61.7 35.4 45.4 51.8 38.0 4.0 4.0 42.7 211 81 130 76 42 74 18.7 13.5 9.2 27.3 9.5 204 80 124 71 38 70 19.8 13.0 8.1 26.0 8.5 212 77 135 81 49 76 23.3 15.6 8.3 28.3 10.1 195 66 129 63 36 67 20.1 12.5 7.7 26.0 9.7 214 77 137 81 40 75 20.6 14.6 8.8 28.6 9.2 207 72 135 75 41 71 21.9 15.1 9.0 26.9 190 79 111 66 39 64 20.2 14.0 8.4 26.0 7.5 195 73 123 73 38 69 20.2 14.6 7.8 27.9 9.5 4.9 5.7 4.1 5.7 4.8 5.9 4.2 5.2 4.6 6.5 4.8 5.7 4.3 5.8 4.9 3.8 4.8 6.5 4.2 5.8 4.5 and radio collar replaced in conjunction with ~Co1laborative 4.2 6.2 4.1 5.2 4.5b 6.0 8.6d 6.3 9.2 5.7 3.5 5.2 3.3 4.0 d Wild Cougar Project.~ healed fracture of left tibia. cRight front paw missing. Ear-tattooed only at 5.9 kg body weight after release from 1egho1d trap on 12-17-85. Animal pursued with bounds and shot at site of sheep killing. ~ght foot, left foot injured. �202 Table 4. Number of aerial telemetric locations of 26 puma radio tracked 7-19-87 to 7-14-88; 22 were being tracked at end of the fiscal year. Ear tatoo no. 4 5 6 7 12 20 21 22 32 34 36 37 38 40 42 44 45 46 47 50 52 53 54 55 56 57 Total Capture date 1-21-82 2- 7-83 2-10-83 3- 4-83 3-25-83 11-26-84 12- 7-84 1- 5-85 1- 8-86 2-13-86 2-13-86 12- 9-86 12-12-86 1- 3-87 1-10-87 1-17-87 1-20-87 1-22-87 1-29-87 2-17-87 4-30-87 12-19-87 1-11-88 2- 5-88 3- 3-88 4-15-88 Sex F M F F F M F M F F F F Est. age (months) on 6-8aa 101 125 94 111 124 44 53 56 53 27 M M M M F F M M M F F M F F 50 30 54 48 26 43 54 19 34 17 28 GMU 10cations No. locations 1987-88 31 51 50 49 51 45 51 49 51 51 7 48 1 50 8 47 48 4 36 27 51 30 27 20 16 13 ---------Majority Minor 62 62 62 62 62 62 62 65 62 62 65 62 62 62 25 62 62 62 40 62 62 62 411 62 62 62 62,64 40,61 62,UtahC 62 912 ---------------------------------------------------------------------------aDash indicates mortality during fiscal year except for #37 who died 7-3-88. bDistribution of aerial telemetric locations, 2-18-18 - 7-12-88; 3 in Eagle and 5 in Garfield counties. Reported by Wildlife Biologist Gene Byrne. cAll Utah locations in Little Dolores drainage. �203 Table 5. 7-14-88. Twenty-five puma captured on the Uncompahgre Capture No. Fate 4 5 Body wt. Est. age Capture method" (kg) (months)b Hounds 1-21-82 4-14-84 11-17-86 42.5 24 51 82 x 2- 7-83 5-17-85 s- 9-87 4-20-88 64.0 68.1 61 85 109 121 X X X X 34.0 12 42 X X 48 74 97 X X X X X X Sex F date M 6 F 2-10-83 8-24-85 7 F 3- 4-83 5-14-85 4-25-87 12 F Plateau (GMU 62) who were alive or fate unknown as of 46.0 3-25-83 4-30-85 12-12-87 44.0 42.7 60 85 117 Trapped Comments 9 males. 16 females listed herein X X 20 M 11-26-84 1-12-87 4-25-87 44.5 62.7 57.7 10 36 51 X X X 21 F 12- 7-84 5-12-87 36.8 49.9 12 41 X X M 1- 5-85 1- 7-87 4-22-88 56.0 69.8 69.7 15 39 54 X X X Too exhausted and hot to weigh Possibly pregnant Too exhausted and hot to weigh Capture site too steep to weigh Missing 2 1/2 toes. left front paw 25 Unk F 1-24-85 14.0 4 X Not radio collared. too small 30 Unk F 12-17-85 5- 3-86 6.4 21.8 2 6 X X Not radio collared. too small Not radio collared. too small Possibly lactating 32 F 1- 8-86 40.0 30 x 34 F 2-12-86 28.1 7 x 40 M 1- 3-87 4-22-88 70.8 70.3 32 47 X X 42 M 1-10-87 49.8 12 X 44 M 1-17-87 59.9 36 X 45 F 1-20-87 46.3 30 X M 1-29-87 4-27-88 39.0 51.8a 9 24 X X Since 1-21-88 in GMU 25 X Severe trap injury. left leg Severe trap injury, left leg 48 Unk F 2- 4-87 17.2 5 X Litter made of 49, not radio collared 49 Unk F 2- 4-87 19.5 5 X Not radio collared 52 M 4-30-87 56.3 30 X 53 F 12-19-87 49.0 48 X 54 F 1-11-88 36.3 14 X 55 M 2- 5-88 4-26-88 62.6 54.9 30 32 X X 56 F 3- 3-88 35.4 14 X F 4-15-88 45.4 26 X 57 a'22 recaptured and recollared Possibly pregnant in GMU 65 and #47 in GMU 40. bGross approximations based on dental characteristics measurements (Anderson 1983) at time(s) of capture(s). cPursuit with hounds and drug immobilization commercial trappers. (Ashman et al. 1983) and body wt. and or caught in leghold traps set for bobcat or coyotes by �204 Table 6. Thirty-two puma captured on the Uncompahgre ausl2ected as of 7-14-88. Ear Body vt. tattoo Sex Date 1 2 3 8 10 13 F F F M M M 33.1 43.5 38.5 25.9 26.3 36.3 52.0 14 15 16 17 18 19 23 24 M F M M M F M M 4-16-81 1- 5-82 1- 8-82 2-20-83 3-23-83 4-13-83 10-11-83 12-12-83 12-15-83 1-15-84 4-14-84 4-16-84 7-18-84 1- 9-85 1-23-85 26 M 3-20-85 27 28 None 29 31 33 35 36 37 38 39 None 41 43 F 46 50 51 F M no. Plateau (GMU 62) whose deatha were confirmed or P P F P P M P P M M M M F M (~) Body Eat. age Cal2ture methodb (IIOnths)a Hounds vt. Tral2l2ed (kS) 12 40 26 6 6 10 16 84 60 30 8 36 7 12 18 X X X X X X X X X X X X 44.0 10 X 4-24-85 11-20-85 12-14-85 12-17-85 12-30-85 2- 4-86 2-13-86 2-13-86 12- 9-86 12-12-86 12-19-86 4- 3-86 1- 8-87 1-16-87 36.3 43.1 14.5 5.9 12.9 23.6 30.9 25.4 27.8 X X X 32.2 33.1 53.5 39.9 14 24 3 2 4 6 6 6 6 36 7 8 14 24 X X X X X X X X 61.3 33.1 53.5 39.9 1-22-87 2-17-87 3-16-87 46.3 51.8 29.1 60 13 7 X X X 29.1 48.0 55.0 36.2 64.3 28.0 44.0 32.2 X X X X X X aGross approximations based on dental charscteristics measurements (Anderson 1983b) at time of capture. bpursuit with hounds and drug immobilization coyotes by commercial trappers. 14.5 38.0 40.0 Date Cause(s) Unconfirmed 1-10-82 Illegal kill Unconfirmed 8- 4-83 Capture trauma 5-29-8JC UnkDovn Unconfir.ed Area GMU 62 GMU 62 GMU 62 6-l6-8JC Possible predacide GMU 62 3-26-86c Prob. capture trauma GMU 62 c GMU 62 3-27-84 Puma GMU 42 1-11-85 Legal kill GMU 65 3-18-86 Legal kill Sbot as aheep killer GMU 70 1-15-86 Legal kill GMU 40 12-12-85 GMU 62 Prob. capture 1-23-85 trauma - puma attack injuries 2-18-87 Legal kill w. of Ibnticello, Utah 12- 9-85 Legal kill GMU 61 11-22-86 Legal kill GMU 40 12-14-85 Hounds GMU 62 1-18-88 Sbot as sheep killer GMU 62 1- 7-86 Trsp injuries GMU 62 11-6-86 Illegal kill GMU 65 Unknovn Unknovn GMU 40 8- 7-8r Unknovn GMU 62 7- 3-88 Sbot as sheep killer GMU 62 c 5-l8-87 Unknovn GMU 62 12-22-87 Legal kill GMU 40 4- 3-86 Hounds GMU 62 1- 8-87 Fall from tree GMU 62 1-16-87 Probably captureGMU 62 related 8- i-sz'' Unknovn GMU 62 c 2-l9-87 Unknovn GMU 62 3-16-87 Hounds GMU 62 (Ashman et al. 1983) and body ¥t. and (Ho) or caught in legbold traps set for bobcst or cOate of death judged as ~ 3 days of the actual; all other dates are actual dates of death. �205 Table 7. 1981-88. Mortality among puma captured on the Uncompahgre Ages estimated at the time of death. Less than 24 months Male Female Plateau, GMU 62, 24 months and older Male Female Total Radiocollared Unconfirmed Illegal kill Legal killa Unknown causes Suspected predacide Intraspecific strife Predator managementb Capture-related o o 1 1 o 3 2 1 1 o o 2 2 1 1 o o o 1 1 o 1 3 1 1 1 2 7 6 1 1 1 2 2 6 28 o o 4 o Subtotal 9 5 7 7 Not radio collared Capture-related Predator managementb 1 2 Subtotal Total o o o o o 1 o 3 1 1 2 o 1 4 10 7 7 8 32 a3 puma killed in GMU 40, 1 in GMU 42, 1 in GMU 61, and 1 in GMU 65, and 1 west of Monticello, Utah. bpuma shot after killing domestic sheep; 2 shot in GMU 62 and 1 in GMU 70. �206 Table 8. Age and sex classes of 57 puma at first capture on the Uncompahgre Plateau, GMU 62, 1981-88. Includes all puma handled. Year 1981-82 1982-83 1983-84 1984-85 1985-86 1986-87 1987-88 Less than 24 months of a~e 24 months and older Male Male 0 3 1 5 2 6 0 Female 1 1 0 4 7 3 2 Total 17 18 --------------------------------- Female Total 0 1 3 0 0 4 1 3 2 1 0 2 3 2 4 7 5 9 11 16 5 9 13 57 Table 9. Age and sex classes of 49 puma radioco11ared at first capture, Uncompahgre Plateau, GMU 62, 1981-88. Excludes 8 puma handled but not radiocollared. Less than 24 months of age Year Male 1981-82 1982-83 1983-84 1984-85 1985-86 1986-87 1987-88 Total Female 24 months and older Male Female o 1 o 3 3 1 1 2 1 Total 4 7 1 o 5 1 3 3 o 2 6 5 o 2 2 4 3 1 2 14 5 15 12 9 13 49 3 o o 5 8 Table 10. Age and sex classes of 21 puma alive and being radio tracked as of 7-14-88. Includes 1 male (#42) radio tracked by Wildlife Biologist G. Byrne in GMU 25. Less than 24 months of age Year 1987-88 Male o Female 2 24 months and older Male 9 Female 10 Total 21 �207 Table 11. Some vital statistics of 21 male and 23 female puma captured on the VncomEahgre Plateau (GMU ~2), 1981-88. Datesa Ear Est. age tattoo no. (months) Capture Death 61 6 6 9 84 30 7 36 10 15 12 10 36 7 32 12 36 9 13 30 30 2- 7-83 2-20-83 3-23-83 4-13-83 12-12-83 1-15-84 4-14-84 4-16-84 11-26-84 1- 5-85 1- 9-85 3-20-85 12-12-86 12-19-86 1- 3-87 1-10-87 1-17-87 1-29-87 2-17-87 4-30-87 2- 5-88 A 8- 4-83 5-29-83 Uc 6-16-84 3-28-84 1-11-85 3-18-86 A A 12-12-85 2-18-87 5-18-87 12-22-87 A A A A 2-18-88 A A Dal:s monitored MALE 5 8 lOb 13 14b 16b 17 18 20 22 23 26 38b 39 40 42 44 47 50 52 55 1,983 165 66 186 72 262 701 1,326 1,292 335 699 157 368 558 549 538 526 366 441 160 FEMALE 1 12 4-16-81 U 1-10-82 1- 5-82 2 40 3 26 U 1- 8-82 A 1-21-82 4 24 6 12 2-10-83 A 3- 4-83 A 48 7 12 A 61 3-25-83 15b 12-15-83 3-23-86 60 1-15-85 7-18-84 19 7 21 12 A 12- 7-84 12- 9-85 4-24-85 27 14 28 24 11-20-85 11-22-86 1-18-88 12-17-85 29 2 32 30 1- 8-86 A 11- 6-86 2- 4-86 7 33 34 6 2-12-86 A 36b 2-13-86 8- 7-87 6 7- 3-88 12- 9-86 37 6 45 A 30 1-20-87 A 12-19-87 48 53 54 A 14 1-11-88 A 3- 3-88 56 14 57 26 A 4-15-88 --------------------------------------------------a"A" indicates puma alive at last telemetric signal during July, bPuma whose date of death approximated i3 5 2,371 1,980 1,957 1,936 1,193 181 1,315 229 367 762 918 275 883 540 572 541 208 155 127 91 1988. days from the actual date. c"U" indicates death of puma was unconfirmed; numbers, 1, 3, 13. d Dates of death for pumas #35 and #46 are unknown. �208 Table 12. Temporal factors in radio tracking puma captured on Uncompahgre Plateau (GMU 62), 1981-88. Statistic Sex and variable SD Min. Max. 24.1 537.5 20.4 487.0 6 66 84 1,983 11 307.0a 221.2 66 701 21 21 23.4 790.8 18.4 729.5 2 5 61 2,371 458.2a 358.6 5 1,193 N X 20 20 MALE Est. age (months) No. days monitored No. days from capture to death FEMALE Est. age (months) No. days monitored No. days from capture to death 9 ------------------------------------------------------------------------------aMeans did not differ (p < 0.05) from each other. Table 13. Number of resident puma on Uncompahgre Plateau related to days monitored as of 7-14-88. Male 365 730 731 - 1,096 1,097 - 1,462 4 0 2a 2 1 2 6 1 4 1 - 4 5 7 9 ----- 1,463+ ---------------------aApproximate1y nonresident. Female Total Days monitored 1,000 days monitored as a resident, 200-300 days as a 16 �209 Table 14. Eisht .ule deer killed bl ~umal Unco.E!hgre Plateau Date (GMU62)1 1987-88. Lesal deecri2tion Eat. ase (.os. ) 1/4 Habitat Co_nU piDe~alt red on by 2 pu.a, DDt covered Part1ally covered Covered, ld11ed by pu.a found Sex 11-28-87a F 16 SE 16 46N 9W 12-12-87 3- 3-88· F F aature aature SE SW 35 36 15S 5lN 98W 14W 4- .5-88 4- 8-88 4-18-88 4-25-88 F SE NW SW SW 11 19 22 21 49N l3W 48N 11W 47N 8W 49N 11W P-J P-J P-J P-J, 4-26-88 F SE 18 49 P-J S I. T P-J, .parae P-J, roadside # 156 aSee 10 10 10 F H 12W Not covered Only hair aDd drag aarlt found Covered ria text. Table 15. Age claaaee (GMU62) 1981-88.a and aell: 0 f cervids 13-17 .os ~12 .os Fiscal lear SE!!e1es Unco-.pah&re Plateau 1t1l1ed by p~, 18+ .oa H F H F H F Sex unit Age unit Sell: and ale unit Total 1980-81 .ue ellt deer 0 0 1 0 0 0 0 0 0 0 2 0 0 0 0 0 0 0 3 0 1981-82 .w.e deer elk 1 0 0 0 0 0 0 0 2 0 6 0 0 0 0 0 0 0 9 0 1982-83 .w.e deer ellt 0 0 1 0 1 0 2 0 0 0 5 0 0 0 0 0 0 0 9 0 -1983-84 lIUle deer ellt 3 0 4 0 0 0 1 1 0 0 2 1 1 0 1 0 0 0 12 2 1984-85 .w.e deer ellt 1 0 1 0 0 0 0 0 1 0 0 0 0 0 0 0 4 0 7 0 1985-86 .w.e deer ellt 0 0 1 0 0 0 1 0 1 0 2 0 4 1 0 0 0 0 9 1 1986-87 .w.e deer aU 0 0 0 1 0 0 0 0 0 0 3 0 6 0 0 0 2 0 11 1 1987-88 .w.e deer alit 0 0 2 0 0 0 1 0 0 0 3 0 1 0 0 0 1 0 8 0 Total .w.e deer S 10 1 S 4 23 12 1 7 68 Total ellt 0 1 0 1 0 1 1 0 0 aThe aajority of 1t1l1a found fro. late No_her to early Hay while purau1D& paa with bounds. bOne ellt (1986-87) fro. GMU6S (Billy CIt) probably 1t1l1ed by pu.a 122. 4b �21ID o 10 20 I I I KM Figure 1. Puma study area GMU 62, Uncompahgre Plateau, showing subunits. �211 APPENDIX �212 AERIAL TELEMETRY LOCATIONS OF 26 PUMA Legal description Date 1/4 S T U.T.M. R X Y Appro~. elev (m) Distance (IuD) between locstions Rating Major drainase S1snal Location ADULT FEMALE PUMA 14 7-1H7 7-24-87 8- 5-87 8- 9-87 8-16-87 8-28-87 9- 4-87 9-13-87 9-18-87 9-26-87 10- 2-87 10- 9-87 10-16-87 10-23-87 10-29-87 11- 6-87 11-13-87 11-20-87 11-28-87 12- 4-87 12-11-87 12-21-87 12-29-87 1- 8-88 1-15-88 1-22-88 1-29-88 2- 4-88 2-12-88 2-19-88 2-26-88 3-12-88 3-18-88 3-25-88 4- 2-88 4- 8-88 4-15-88 4-16-88 4-22-88 4-29-88 5- 7-88 5-20-88 5-27-88 6- 4-88 6-10-88 6-17-88 6-23-88 6-30-88 7- 8-88 7-14-88 7-20-88 ININ IN SW SW HE SE HE IN IN IN SW IN SE HE IN HE SE IN SW SE HE SE HE SW IN SW IN SW HE SW No No No No No No No No IN No No No No No No No No No No No -- 9 47 13 47 29 47 47 13 34 48 33 48 7 47 33 48 8 47 33 48 33 48 8 47 17 47 32 48 8 47 7 47 18 47 17 47 47 8 4 47 4 47 22 48 32 48 48 21 32 48 28 48 23 47 28 48 47 8 - 33 48 34 48 location location location location location location location signal 14 50 location location signal location signal location location location location location location - 10 239 4248 2,286 1.0 Spring G G 758 4246 11 2,499 4.9 Spring G G 4242 10 762 2,408 4.7 E Fk Spring G G 4245 11 758 2,621 4.0 Spring G F 10 241 4250 2,073 8.2 G G Spring 10 240 4251 2,134 0.9 Spring G F 10 761 4247 2,377 4.8 Spring G F 10 240 4251 2,134 -G 4.9 G Spring 10 762 4247 2,347 4.3 G G Spring 10 240 4251 2,164 4.3 Spring G G 10 4251 239 2,196 G 0.7 Devinny G 10 762 4247 2,377 4.4 G Spring G 10 762 4246 2,316 1.5 G G Spring 10 762 4250 2,256 4.2 G Devinny G 10 4248 762 2,316 G 2.1 Spring G 10 761 4247 2,377 2.0 G Spring G 10 761 4246 2,438 1.4 G Spring G 10 239 4245 2,438 2.2 G Spring G 10 762 4247 2,286 2.6 G Spring G 10 239 4248 2,134 1.4 Spring G G 10 4248 240 2,256 1.0 G G Spring 4254 10 242 1,890 6.4 G Spring G 10 762 4250 2,286 5.6 Spring G G 10 240 4254 4.7 2,012 Lindsay G G 762 4251 10 2,256 4.5 Devinny G G 4252 10 239 2,103 2.7 Lindsay F G 243 4244 10 2,316 9.4 Happy G F 4252 239 1,829 9.5 10 Lindsay G F 761 4247 10 2,377 5.8 Spring G G 4250 10 240 2,073 4.5 Spring G G 4250 10 241 2,164 0.9 Spring G G Fair momentary signal over Potter and Criswell Drainages Faint momentsry signals from Upper Little Dominguez to Spring Ck Fsint momentary signal Escalante to Spring Ck Good intermittent signal over southcentral Dry Mesa Good intermittent signal over IN portion Dry Mesa Mostly poor signals IN Jct Big Dom-Guonison R, Sawmill Mesa, Monitor and 7 N Mesa No signals over same route as 4-15-88 4775 Roubideau 12 745 G Strong signal over Granite Can (GMU 40) and a less strong signal over Dry Mesa Signals of varying quality from Montrose north to Big Dominguez Signals good quality, momentary, P just HE Blake Field, Delta Signals of varying quality from Dry Ck Basin to Upper Escalante Poor signals in Horsefly Ck vicinity Good momentary signal heard over Ladder Can, GMU 40 Fair signal over Spring Ck Basin Fair momentary signal IN of Delta; rim Grand Mesa Fair momentary signal Log Hill Mesa ADULT MALE PUMA 85 7-19-87 7-24-87 8- 5-87 8- 9-87 8-16-87 8-28-87 9- 4-87 9-13-87 9-18-87 9-26-87 10- 2-87 10- 9-87 10-16-87 10-23-87 10-29-87 11- 6-87 11-13-87 11-20-87 11-28-87 12- 4-87 12-11-87 12-21-87 12-29-87 1-10-88 1-15-88 1-22-88 1-29-88 2- 5-88 IN NW HE HE HE SW SW IN IN SE IN SW SW IN IN SW SE IN SE HE HE HE HE HE HE SE SW HE 25 10 9 8 6 4 21 17 29 36 27 29 26 1 30 31 31 23 8 31 8 9 35 7 10 1 33 33 49 49 48 49 48 49 49 49 49 50 50 49 49 49 50 50 50 50 48 49 49 49 49 49 49 49 49 49 13 13 13 12 12 13 13 12 12 14 13 12 13 13 12 13 12 13 12 12 12 13 13 11 12 12 11 11 738 734 735 741 741 732 732 740 740 728 734 741 736 738 738 739 739 736 743 739 741 733 737 749 744 748 752 753 4262 4267 4257 4267 4260 4268 4263 4206 4263 4269 4272 4262 4261 4269 4272 4270 4270 4274 4256 4261 4268 4267 4261 4268 4268 4269 4260 4261 2,316 2,256 2,682 2,042 2,347 2,377 2,408 ~,O42 2,377 2,103 2,225 2,256 2,499 2,256 2,042 2,134 2,103 2,073 2,499 2,408 1,981 2,316 2,438 1,981 1,951 1,951 2,103 2,134 1.9 6.3 10.7 12.2 7.9 8.9 5.0 8.3 3.2 13.2 5.9 13.5 5.0 7.4 3.5 2.3 0.3 5.3 19.2 6.2 7.1 8.7 7.3 14.7 5.3 3.6 9.0 1.1 Criswell Little Monitor Criswell Criswell Traver Cottonwood Monitor Criswell Traver Dry Fk Escalante Cottonwood Traver Criswell Monitor Monitor Potter Potter Cottonwood Bull Traver Criswell Cottonwood Criswell Big Sandy Wash Roub1deau Roatcap Gulch Piney Piney G G G G G G G G G G G G G G G G G G G G G G G G G G G G G F F F F F F G G F G F G G F G G G G G G G G F G F G G �213 Legal description Date 1/4 S T R U.T.H. X Y Approx. dev (m) Distance (km) between locations Rating Major drainase Sisna1 Location Adult Male Puma IS - continued 2-12-88 2-19-88 2-26-88 3- 4-88 3-12-88 3-18-88 3-25-88 4- 2-88 4- 8-88 4-15-88 4-22-88 4-29-88 5- 7-88 5-14-88 5-20-88 5-27-88 6- 4-88 6-10-88 6-17-88 6-23-88 6-30-88 7- 8-88 7-14-88 SE SW NW SW SE SE SE SE SW NW SW SE SE SW SE NE SE SW SW SW SE SW NE 1 1 12 24 18 15 16 21 36 7 1 10 9 11 31 16 15 27 24 34 23 20 12 48 48 48 48 48 49 49 49 50 48 48 48 48 49 49 48 49 49 49 49 49 49 48 11 11 11 11 10 11 11 11 12 11 12 12 12 13 12 13 13 13 13 13 13 12 13 759 759 758 759 761 754 753 752 747 750 749 747 744 737 739 735 735 734 738 734 738 741 740 4259 4258 4257 4253 4255 4266 4265 4264 4270 4257 4259 4256 4256 4266 4260 4255 4264 4262 4263 4260 4263 4263 4257 1,951 2,012 2,073 2,256 2,103 1,951 1,981 1,981 2,012 2,316 2,286 2,377 2,469 2,225 2,408 2,775 2,347 2,408 2,347 2,621 2,377 2,256 2,591 6.6 0.8 0.7 4.3 2.8 13.0 1.1 1.6 8.1 13.3 2.3 2.8 2.2 13.2 7.0 7.0 9.9 3.3 3.8 4.4 4.4 4.0 6.3 Coal Coal Coal Spring Coal Cushman Cushman Cushman Roatcap Piney Cushman Roubideau Roubideau Monitor Traver Moore Potter Potter Criswell Potter Criswell Roubideau Wright G G G G G G G G G G G G G G G G G G G G G G G G G G Criswell Moore Criswell Criswell Criswell Moore Criswell Criswell Potter Criswell Criswell Roubideau Terrible Traver Moore Traver Traver Criswell Criswell Criswell Moore Criswell Traver Potter Potter Criswell Potter Potter Criswell Criswell Potter Potter Potter Potter Potter Potter Potter Criswell Traver Traver Bull Criswell Criswell Terrible Wright Terrible Criswell Roubideau Roubideau Roubideau G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G F F G F F F F G G F G G G G F G F G G G F ADULT FEMALE PUMA #6 7-19-87 7-24-87 8- 5-87 8- 9-87 8-16-87 8-28-87 9- 4-87 9-13-87 9-18-87 9-26-87 10- 2-87 10- 9-87 10-16-87 10-23-87 10-29-87 11- 6-87 11-13-87 11-20-87 11-28-87 12- 4-87 12-11-87 12-21-87 12-29-87 1- 8-88 1-15-88 1-22-88 1-29-88 2- 5-88 2-12-88 2-19-88 2-26-88 3- 4-88 3-12-88 3-18-88 3-25-88 4- 2-88 4- 8-88 4-15-88 4-22-88 4-29-88 5- 7-88 5-14-88 5-20-88 5-27-88 6- 3-88 6-17-88 6-23-88 6-30-88 7- 8-88 7-14-88 SW NW NW SW SE HE SW NW SE SW SE NW SW NW SE SE SE SW SE SW SW SW SW NW SW SW SW SE SW NE NE NW HE SE SW NW NW NW NW SE HE SW SE NW NW NW HE HE NW NW 25 30 25 24 35 36 25 36 22 25 18 27 6 32 19 29 30 19 35 19 30 30 29 14 12 13 32 1 27 33 28 16 31 31 7 14 5 9 31 2 17 30 4 7 6 7 35 9 16 15 49 49 49 49 49 49 49 49 49 49 49 49 48 49 49 49 49 49 49 49 49 49 49 49 49 49 50 49 50 50 50 50 50 50 49 49 49 49 49 48 48 49 48 48 48 48 49 48 48 48 13 12 13 13 13 13 13 13 13 13 12 12 12 12 12 12 12 12 13 12 12 12 12 13 13 13 12 13 12 12 12 12 12 12 12 13 12 12 12 13 12 12 13 12 12 12 13 12 12 12 738 739 738 738 737 738 738 738 734 738 740 ·744 741 740 741 741 740 739 738 739 740 739 740 736 739 739 741 738 743 742 743 742 739 739 739 735 741 742 739 738 743 739 735 741 741 740 737 746 744 745 4262 4262 4263 4264 4260 4261 4262 4261 4263 4261 4265 4263 4258 4261 4263 4262 4267 4263 4260 4263 4261 4262 4262 4266 4266 4265 4270 4268 4272 4271 4272 4276 4270 4270 4267 4266 4269 4268 4261 4258 4255 4262 4258 4257 4259 4257 4261 4257 4255 4255 2,347 2,347 2,377 2,377 2,469 2,499 2,408 2,438 2,377 2.377 2,073 1.951 2.469 2,316 2.286 2.256 2.438 2.256 2.469 2.256 2.347 2,347 2.408 2.286 2.134 2.256 2.042 2.256 1.981 2.103 2,042 2.012 2,103 2.134 2,225 2,316 2.164 2.103 2,438 2,560 2,499 2,347 2,012 2,469 2,408 2,560 2,469 2,347 2,530 2,134 1.2 2.1 1.4 1.3 3.3 1.6 1.6 0.6 3.5 3.5 4.1 4.7 4.8 2.4 2.3 1.8 1.2 1.8 3.9 3.6 1.8 0.5 1.7 6.1 1.9 1.7 6.0 3.9 6.6 1.1 1.3 3.5 5.5 0.7 3.1 3.9 6.3 1.8 7.3 3.2 5.3 7.6 5.7 5.2 2.0 2.5 5.4 8.8 2.5 1.4 G F F G F G G G F G F G F G G G G G G G F G F G G F G G G F F G F G F G G G G G G G F G G F F G G �214 Legal deacription Date 1/4 S T D.T.M. R X Distance Y Appro>:. elev (m) (kill) between locations Rating Major drainase Sisnal Location ADULT FEMALE PUMA 17 7-19-87 7-24-87 8- 5-87 8- 9-87 8-16-87 8-28-87 9- 4-87 9-13-87 9-18-87 9-26-87 10- 2-87 10- 9-87 10-16-87 10-23-87 10-29-87 ll- 6-87 11-13-87 11-20-87 11-28-87 12-11-87 12-21-87 12-29-87 1- 8-88 1-15-88 1-22-88 1-29-88 2- 5-88 2-12-88 2-19-88 2-26-88 3- 4-88 3-12-88 3-18-88 3-25-88 4- 2-88 4- 8-88 4-15-88 4-22-88 4-29-88 5- 7-88 5-14-88 5-20-88 5-27-88 6- 3-88 6-10-88 6-17-88 6-23-88 6-30-88 7- 8-88 7-14-88 NW NW SE SW NW NW SE HE NW HE HE NW SW SE HE SW NW HE SE SW SW SW SE SE HE NW HE SE HE HE HE HE NW SE NW SE HE NW SE SW No SE NW NW NW NW NW NW NW HE 20 16 10 12 70 15 17 5 4 24 7 23 32 10 16 35 3 10 19 4 3 33 14 32 21 22 27 6 . 14 6 17 4 20 19 29 30 34 25 13 35 signal 36 17 34 36 23 7 11 6 6 49 49 49 49 49 49 49 49 49 49 49 49 50 49 49 50 49 49 50 49 49 50 50 50 50 50 50 50 50 50 50 50 50 50 50 50 50 50 49 50 14 14 14 14 14 14 14 13 13 14 13 14 13 14 14 14 14 14 13 13 14 13 13 13 13 13 13 12 13 12 13 13 13 13 13 13 14 14 14 14 721 723 725 727 725 724 722 731 732 729 730 726 731 725 724 726 724 725 730 732 724 732 736 732 732 733 734 739 736 739 732 733 730 730 730 730 725 ·727 729 726 4264 4265 4266 4266 4267 4265 4264 4269 4268 4264 4267 4264 4269 4266 4265 4269 4269 4267 4273 4268 4268 4270 4275 4269 4274 4274 4272 4278 4275 4279 4275 4279 4274 4273 4272 4271 4270 4272 4265 4269 2,499 2,256 2,438 2,316 2,377 2,469 2,316 2,316 2,256 2,499 2,408 2,499 2,377 2,560 2,438 2,196 2,377 2,408 2,164 2,103 2,316 2,347 2,073 2,316 2,134 2,225 2,073 1,951 2,073 1,859 2,073 1,981 2,164 2,196 2,225 2,225 2,225 2,377 2,499 2,103 6.2 2.1 2.8 2.0 2.5 2.7 2.1 8.0 0.6 6.3 4.0 5.0 8.0 2.6 1.6 4.6 1.9 1.6 8.0 5.2 8.0 8.5 6.9 7.3 4.3 1.2 1.7 7.2 4.0 5.2 8.5 3.7 5.7 0.6 1.4 0.5 5.8 3.3 7.0 5.0 E Fk Escalante E Fk Escalante E Fk Escalante Dry Fk Escalante Middle Pk Escalante E Pk Escalante E Fk Escalante Cottonwood Cottonwood Cottonwood Cottonwood Dry Fk Escalante Cottonwood E Pk Escalante E Fk Eacalante E Pk Eaca1ante E Pk Eacalante E Fk Escalante Dry Fk Escalante Cottonwood E Fk Escalante Cottonwood Cottonwood Cottonwood Dry Fk Escalante Grade G Cottonwood Cottonwood Cottonwood Cottonwood Dry Fk Escalante Dry Fk Escalante Dry Fk Escalante Dry Fk Escalante Dry Fk Escalante Dry Fk Escalante E Fk Escalante E Fk Escalante Cottonwood E Fk Escalante G G G G 50 49 50 50 49 49 49 49 49 14 13 14 14 14 13 14 13 13 728 729 724 727 726 729 726 729 730 4269 4265 4271 4270 4264 4267 4267 4269 4269 2,103 2,469 2,347 2,377 2,499 2,438 2,499 2,408 2,408 3.1 4.2 6.5 2.7 6.3 4.6 3.5 4.0 0.5 Dry Fk Escalante Cottonwood Middle Fk Escalante E Fk Escalante Dry Fk Escalante Dry Fk Escalante E Fk Escalante Dry Fk Escalante Dry Fk Escalante G G G G KiddIe Middle Kelso Middle KiddIe Middle Middle G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G F F P F G F .F G G F G G G F G G G G G G P G P G G P F G G F F F P F P F G G F G P G F G G F G F G ADULT FEMALE PUMA #12 7-19-87 7-24-87 8- 5-87 8- 9-87 8-16-87 8-28-87 9- 4-87 9-13-87 9-18-87 9-26-87 10- 2-87 10- 9-87 10-16-87 10-23-87 10-29-87 11- 6-87 11-13-87 11-20-87 11-28-87 12- 4-87 12-11-87 12-21-87 12-29-87 1- 8-88 1-15-88 1-22-88 1-29-88 SW SE NW SW NW HE SW SE SE SW SW HE HE NW SW HE SW SE SE SE SE SW SE SE HE SE NE 30 29 19 30 31 79 30 36 31 36 17 19 29 13 10 25 22 1 26 12 35 22 6 33 1 29 4 50 50 50 50 50 50 50 50 50 50 50 50 50 50 50 50 50 50 51 50 15 51 50 51 50 51 50 14 14 14 14 ·14 14 14 15 14 15 14 14 14 14 14 15 14 14 14 14 13 13 13 13 14 13 13 719 722 719 719 719 722 719 718 720 118 720 720 721 727 724 118 724 728 726 728 734 734 729 733 728 131 133 4271 4271 4273 4270 4270 4271 4271 4269 4270 4269 4274 4273 4271 4275 4275 4271 4273 4278 4281 4276 4286 4283 4278 4280 4278 4281 4279 2,560 2,103 2,134 2,499 2,469 2,377 2,560 2,316 2,256 2,347 2,225 2,134 2,377 2,256 2,256 2,560 2,196 2,164 2,134 2,286 1,768 1,951 2,196 2,042 2,134 1,981 1,951 2.4 2.6 4.5 2.7 0.4 3.2 3.0 1.7 2.0 2.2 5.3 0.8 2.2 6.8 3.0 7.3 6.1 2.2 4.3 5.3 12.0 3.6 6.6 4.2 4.8 3.7 2.8 ~.!::!;!ls Fk Escalante Fk Escalante Fk Fk Fk Fk Escalante Escalante Escalante Escalante rll Escalante Middle Fk Escalante KiddIe Fk Escalante Kelso Kelso Middle Fk Escalante Escalante Kelso Kelso KiddIe Fk Escalante Escalante Escalante Escalante Escalante Escalante Tatum D Dry Fk Escalante Escalante Escalante Dry Fk Escalante G G G G G G G G G G G G G F G F F G F G G G F F G F G G G F G G G G G G G G G G G G F G G G G G F G �215 Legal description Date 1/4 S T U.l.M. R X Y Approx. dev (m) Distance (km) between locations Rating Major drainalle Sillnal Location Adult female 2um8 112 - continued 2- 5-88 2-12-88 2-19-88 2-26-88 3- 4-88 3-12-88 3-18-88 3-25-88 4- 2-88 4- 8-88 4-15-88 4-22-88 4-29-88 5- 7-88 5-14-88 5-20-88 5-27-88 6- 4-88 6-10-88 6-17-88 6-23-88 6-30-88 7- 8-88 7-14-88 NW NE NE SE HE NW NW NW SE SW NW NW NW h'W NW SE SW HE HE SE SW SE SW NW 4 28 10 28 14 4 34 15 29 10 20 30 15 33 2 3 28 29 6 18 14 10 34 21 SO 51 SO 51 SO 50 51 50 51 50 50 50 50 50 49 49 SO SO 49 SO 50 50 SO 50 13 13 14 13 14 13 13 13 13 14 14 14 14 15 15 15 15 15 15 14 14 14 15 14 732 732 724 733 727 732 734 733 731 724 721 719 724 713 716 715 713 713 712 720 725 725 715 722 4278 4282 4276 4281 4275 4279 4280 4275 4281 4276 4273 4271 4275 4278 4268 4268 4271 4271 4268 4274 4274 4276 4269 4273 2,164 2,073 1,981 2,073 2,225 2,042 2,042 2,073 1,981 2,164 2,256 2,499 2,286 2,714 2,347 2,377 2,408 2,408 2,560 2,256 2,134 2,134 2,621 2,377 1.2 6.9 9.2 9.5 9.0 7.0 2.2 4.7 6.7 8.8 4.0 2.5 5.8 11.7 3.2 0.7 4.4 0.4 3.2 10.9 5.2 1.8 12.2 8.7 Dry Fk Escalante Eecalante Kelso Tatwa D Escalante Dry Fk Escalante Dry Fk Escalante Dry Fk Escalante Eecalante Kelso Ke1ao Kdso Eac:alante Kdso Middle Fk Escalante Middle Fk Escalante Ke1ao Kelso Ke1ao Kelso E Fk Escalante Escalante Middle Fk Escalante Kelso G G G G G G G G G G G G G G G G G G G G G G G G 1,829 2,316 2,316 2,256 2,377 2,377 1,951 2,042 2,073 1,981 2,286 1,981 2,042 2,469 2,408 2,347 2,196 1,890 2,073 1,981 2,042 2,286 2,134 1,920 2,073 1,981 1,981 1,920 1,920 1,981 6.1 18.1 1.0 5.8 1.6 7.7 15.4 6.7 8.3 3.3 3.5 13.1 12.8 3.5 0.8 6.2 13.7 15.2 12.7 11.6 7.9 21.1 18.0 13.6 16.5 0.9 0.2 17.1 12.5 3.6 Criswell 'IIFk Dry 'IIFk Dry Coa1bank Roubideau 'IIFk Dry Roubideau Criswell Roubideau Roubideau Roubideau Roubideau Roubideau Roubideau Roubideau 'IIFk Dry Moore Dry Ck Roubideau 'IIFk Dry Roubideau Spring Piney Roubideau E Fk Dry Dry Ck E Fk Dry Roubideau Dry Dry G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G 1,859 2,134 2,012 2,103 2,103 1,981 2,286 2,316 2,286 2,073 2,164 2,286 2,377 2,408 2,408 5.4 7.7 28.0 3.4 7.4 6.8 5.3 4.3 2.8 4.4 5.2 9.5 12.3 2.8 6.5 Cushman 'IIFk Dry Ck Roubideau Roubideau Coa1bank Roubideau Roubideau Big Sandy IJash Roubideau Roubideau Roubideau E Fk Dry Ck Long Roubideau Traver G G G G G G G G G G G G G G G Dry Fk Escalante E Fk Escalante Cottonwood Escalante C F G G G G G G G G F G G F F F G G F F G G G F G F F F F G MALE PUMA #20 7-19-87 8- 5-87 8- 9-87 8-16-87 8-28-87 9- 4-87 9-13-87 9-18-87 9-26-87 10- 2-87 10- 9-87 10-16-87 10-23-87 10-29-87 11-13-87 11-28-87 12- 4-87 12-11-87 12-21-87 12-29-87 1- 8-88 1-15-88 1-22-88 2-12-88 2-19-88 2-26-88 3- 4-88 3-12-88 3-18-88 3-25-88 4- 2-88 4- 8-88 4-15-88 4-22-88 4-29-88 5- 7-88 5-14-88 5-20-88 5-27-88 6- 3-88 6-10-88 6-17-88 6-23-88 6-30-88 7- 8-88 7-14-88 34 NW 50 12 48 NW 19 11 SE 48 11 18 48 NW 2 12 SE 49 3 12 48 11 NW 17 49 NW 2 12 HE 49 18 12 HE 4 48 12 SE 49 28 12 NW 3 48 12 HE 26 50 12 HE 4 48 12 48 SE 9 12 48 HE 9 12 SE 48 18 11 SW 49 17 12 11 NW 48 3 49 12 SW 28 48 HE 8 11 48 12 SW 3 47 10 NE 17 49 SW 33 11 SW 34 50 12 48 11 HE 9 SE 4 48 11 48 11 SE 4 SW 26 50 12 NW 49 11 23 11 49 SW 35 No signal - Extensive NW 15 48 11 48 SE 5 11 49 12 HE 9 NW 21 49 ::.2 SE 49 11 19 49 SE 28 12 48 SE 12 3 49 12 SW 26 48 12 NW 3 49 12 HE 28 48 12 SW 3 48 11 SW 21 48 12 NW 8 48 12 HE 9 49 13 SE 31 743 4271 750 4254 751 4255 747 4259 745 4260 752 4256 745 4269 741 4265 745 4259 743 4262 746 4259 746 4272 745 4260 744 4256 745 4257 751 4256 741 4265 755 4259 742 4262 753 4257 745 4258 4246 238 752 4260 744 4270 4257 755 755 4259 4258 755 745 4272 754 4264 755 4261 search 4266 753 753 4258 742 4268 4264 7;';' 749 4264 743 4262 746 4258 745 4261 746 4259 743 4262 745 4258 754 4253 742 4257 745 4257 740 4260 Ck Ck Ck Ck Ck Ck Ck F F F G F F F F F F F G G G G G G F G F F G G G G G G G G F F G F G F F F F F F G G G G G ADULT FEl-1ALEPUMA 121 7-19-87 7-24-87 8- 5-87 8- 9-87 NW SW NW HE 10 3 6 22 49 49 49 50 14 14 13 14 724 124 729 725 4267 4268 4268 4273 2,286 2,347 2,408 2,196 0.7 1.5 5.0 6.5 F F �216 Legal description Date Adult 1/4 female S T U.I.M. R X y Approx. elev (m) Distance (Itm) between locations Rating Major drainal:le S11:lnal Location 2um8 121 - continued 8-16-87 8-28-87 9- 4-87 9-13-87 9-18-87 9-26-87 10- 2-87 10- 9-87 10-16-87 10-23-87 10-29-87 11-6-87 11-13-87 11-20-87 11-28-87 12- 4-87 12-11-87 12-21-87 12-29-87 1- 8-88 1-15-88 1-22-88 1-29-88 2- 5-88 2-12-88 2-19-88 2-26-88 3- 4-88 3-12-88 3-18-88 3-25-88 4- 2-88 4- 8-88 4-15-88 4-22-88 4-29-88 5- 7-88 5-14-88 5-20-88 5-27-88 6- 3-88 6-10-88 6-17-88 6-23-88 6-30-88 7- 8-88 7-14-88 29 7 7 34 HE 1 HE 33 NW 12 HE 36 NW 2 SE 9 SE 36 SE 3 SE 6 SW 27 HE 8 SW 19 NW 7 SE 1 SW 34 SE 32 HE 8 HE 16 NW 4 HE 5 ME 9 SE 16 SW 20 NW 8 NW - 3 NW 16 SE 20 SW 16 SE 24 SW 25 SW 6 SE 32 NW 5 SW 1 SE 6 NW 17 ME 34 SW 30 SE 28 SW 30 NW 21 HE 16 SE 1 50 49 49 50 49 50 49 50 49 49 50 49 50 51 7-19-87 7-24-87 8- 5-87 8- 9-87 8-16-87 8-28-87 9- 4-87 9-13-87 9-18-87 9-26-87 10- 2-87 10- 9-87 10-16-87 10-23-87 10-29-87 11- 6-87 11-13-87 11-20-87 11-28-87 12- 4-87 12-11-87 12-21-87 12-29-87 1-10-88 1-15-88 1-22-88 1-29-88 2- 5-88 2-12-88 2-19-88 SE 46 46 48 48 48 46 48 50 48 48 48 49 48 50 50 50 47 50 47 48 50 48 50 46 48 50 50 50 49 50 ME SW SE SE SO SO 50 50 51 51 50 50 50 50 50 50 50 50 50 50 SO 50 50 50 50 51 49 49 49 49 50 50 50 50 49 49 49 14 14 14 15 15 14 15 15 15 15 15 15 13 13 13 13 13 14 13 13 13 13 13 13 13 13 13 13 13 13 13 13 14 14 13 13 14 15 14 14 14 14 14 14 15 15 15 722 719 720 715 719 724 718 718 716 715 718 715 729 734 731 729 729 728 733 731 731 733 732 731 733 733 732 730 734 732 731 732 728 727 729 731 721 718 720 721 725 719 723 719 716 714 719 4272 4266 4266 4269 4269 4270 4266 4270 4268 4268 4269 4267 4277 4281 4277 4272 4277 4277 4280 4279 4277 4275 4279 4278 4277 4274 4274 4277 4278 4275 4273 4274 4272 4271 4278 4279 4269 4267 4267 4265 4270 4271 4271 4271 4268 4265 4267 259 258 263 261 261 257 260 258 261 261 262 266 261 260 261 261 262 264 262 262 258 263 261 264 261 255 252 252 263 259 4232 4235 4252 4253 4253 4236 4254 4271 4254 4254 4254 4260 4256 4270 4276 4270 4245 4275 4243 4253 4271 4257 4270 4238 4254 4274 4274 4277 4269 4271 2,377 2,469 2,560 2,621 2,316 2,316 2,255 2,530 2,469 2,377 2,438 2,560 2,196 2,073 2,164 2,256 2,225 2,196 2,012 2,073 2,103 2,073 2,073 2,134 2,012 2,134 2,164 2,256 2,012 1,920 2,134 1,859 2,377 2,286 2,256 2,164 2,408 2,408 2,377 2,499 2,256 2,530 2,225 2,591 2,499 2,621 2,196 3.1 6.8 0.9 5.4 4.4 5.2 7.3 3.7 2.6 1.2 3.3 3.7 12.5 5.4 4.4 5.1 4.4 1.2 6.5 2.1 2.3 2.6 3.8 1.0 2.2 3.4 0.9 3.7 4.0 3.7 2.1 1.5 4.3 1.5 6.5 2.3 15.0 3.0 1.8 2.3 6.0 6.0 6.1 4.4 4.2 3.8 5.4 Middle Fk Escalante Middle Fir.Escalante Middle Fk Escalante Middle Fk Escalante Middle Fk Escalante Middle Fir.Escalante Middle Fk Escalante Middle Fir.Escalante Middle Fk Escalante Middle Fk Escalante Middle Fk Escalante Middle Fir.Escalante TatUII IatUII Dry Fk Escalante Dry Fk Escalante Tatum D ESCalante Dry Fk Escalante Tatum D Dry Fk Escalante Dry Fk Escalante Dry Fk Escalante Dry Fk Escalante Dry Fir.Escalante Dry Fir.Escalante Dry Fk Escalante Dry Fk Escalante Dry Fk Escalante Dry Fk Escalante Dry Fk Escalante Dry Fk Escalante Dry Fk Escalante East Fk Escalante Escalante Tatum D Middle Fk Escalante Middle Fk Escalante Middle Fk Escalante East Fir.Escalante East Fk Escalante Middle Fk Escalante Middle Fk Escalante Kelso Middle Fk Escalante Middle Fk Escalante Middle Fk Escalante G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G F F F G G G G Uncompahgre Uncompahgre Beaton Dry Cedar CIt Dry Cedar Ck Fisher Dry Cedar Cir. Gunnison R Dry Cedar Ck Dry Cedar Ck Dry Cedar Cit. Cedar Cir. Dry Cedar Ck Red Rock Can Gunnison Cedar Ck Billy Gunnison Onion Dry Cedar Red Rock Can Dry Cedar Red Rock Can Burro Dry Cedar Loutsenhizer Loutsenhizer Gunnison Cedar Ck Red Rock Can G G G G G G G G G G G G G G G G G F G F F F G F F G G G G G G G G G G F F G F F G G G G G G F G G F F G G G F F G G F F F G F F F ADULT MALE PUMA 122 HE HE SW NW SW SW SW SW SE SW ME SE SE NE HE SW SE SE NW SE SW SE SE NW ME NE SE SE ME 28 20 26 22 22 17 15 28 15 15 14 32 10 34 15 35 23 13 22 23 28 1 35 1 22 19 23 11 1 34 8 8 8 8 8 8 8 8 ·8 8 8 7 8 8 8 8 8 8 8 8 8 8 8 8 8 8 9 9 8 8 2,196 2,196 2,438 2,286 2,256 2,073 2,073 2,196 2,164 2,134 2,408 2,316 2,286 2,196 2,377 2,377 2,347 2,438 2,408 2,286 2,256 2,438 2,256 2,377 2,164 2,225 1,981 1,768 2,499 2,316 8.9 3.2 18.0 2.1 0.3 18.0 19.5 17.2 18.0 0.4 1.5 6.9 6.9 15.0 5.6 5.3 28.0 40.0 32.0 8.0 18.8 15.9 13.0 32.3 16.0 21.3 2.8 2.7 13.6 4.5 G G G G G G G G G G G G G G G G G F G G G G G G G G G F G G F F F G G �217 Legal description Date 1/4 S T U.T.M. R X Y Approx. e1ev (m) Distance (0) between locations Rating Major drainalle Sillo&l Location Adult male I!UDIa'22 - continued 2-26-88 3- 4-88 3-12-88 3-18-88 3-25-88 4- 2-88 4- 8-88 4-15-88 4-22-88 4-29-88 5- 7-88 5-14-88 5-20-88 5-27-88 6- 3-88 6-10-88 6-17-88 6-23-88 6-30-88 7- 8-88 7-14-88 7-20-88 SE 14 50 9 251 4275 2,042 NW 26 48 262 8 4252 2,499 NW 49 262 1 8 4269 2,347 SE 48 22 8 261 4253 2,286 SE 45 5 8 258 4229 2,256 Scanner aalfunction - Could DOt traclt SE 46 18 256 8 4235 2,225 SE 27 50 8 260 4272 2,347 SW 46 29 8 258 4232 2,256 NW 46 33 8 259 4231 2,225 HE 49 1 8 263 4269 2,377 SW 17 46 8 257 4235 2,164 SW 46 18 8 255 4235 2,286 SE 18 46 8 256 4235 2,225 SW 35 47 8 261 4239 2,164 SW 8 46 257 8 4237 2,134 SW 30 46 265 8 4231 2,256 SE 46 28 259 8 4232 2,134 SW 46 10 8 260 4237 2,196 No location - Fair mollentary signal SW Montrose No location - Fair momentary signal SW Montrose NW 4 46 8 258 4239 2,164 8.8 26.5 17.4 16.5 24.2 Loutsenhizer Dry Cedar Cedar Dry Cedar Dallas G G G G G F G F G G 6.6 37.8 40.5 1.8 39.0 35.0 1.8 1.3 6.6 9.0 5.5 5.0 Fisher Gunnison Uncompahgre Uncompahgre Jones D Fisher Fisher Fisher Chaffee Uncollpahgre Cow Uncompahgre Cow G G G G G G G G G G G G G F G G G G G G F G F G G G 3.2 Uncompahgre G F G G G G G G F G F G F G s.o, ADULT FEMALE PUMA 132 7-19-87 7-24-87 8- 5-87 8- 9-87 8-16-87 8-28-87 9- 4-87 9-13-87 9-18-87 9-26-87 10- 2-87 10- 9-87 10-16-87 10-23-87 10-29-87 11- 6-87 11-13-87 11-20-87 11-28-87 12- 4-87 12-11-87 12-21-87 12-29-87 1-10-88 1-15-88 1-22-88 1-29-88 2- 5-88 2-12-88 2-19-88 2-26-88 SE NW HE SW HE SE NE SW SW SW NE SW NW SE SE SE SE SE NE SW SE SE SW SW NW SW SE SW SW NW SW 28 26 ·23 24 18 24 24 24 24 24 24 17 18 5 17 17 20 20 31 20 30 30 8 17 17 18 18 18 23 19 19 46 46 46 46 46 46 46 46 46 46 46 46 46 46 46 46 46 46 46 46 46 46 46 46 46 46 46 46 46 46 46 8 9 9 9 8 9 9 9 9 9 9 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 9 8 8 259 252 252 253 256 254 254 254 254 254 254 257 256 258 258 258 258 257 256 257 256 256 257 257 257 255 256 255 252 255 255 4232 4233 4233 4233 4236 4234 4234 4234 4234 4234 4234 4235 4236 4238 4234 4235 4234 4233 4231 4233 4232 4233 4237 4235 4236 4235 4235 4235 4233 4234 4233 2,103 2,438 2,438 2,377 2,196 2,316 2,316 2,377 2,347 2,347 2,377 2,073 2,196 2,134 2,073 2,196 2,134 2,225 2,377 2,256 2,316 2,286 2,164 2,196 2,134 2,256 2,225 2,286 2,377 2,316 2,377 5.8 7.7 0.6 1.8 3.8 3.3 0.4 0.8 0.1 0.0 0.5 3.5 0.9 2.6 2.5 0.4 1.2 0.6 2.8 2.4 1.3 1.0 5.2 1.8 1.0 2.7 1.3 1.2 4.2 3.2 1.2 Uncompahgre Fisher McKenzie Fisher Fisher Fisher Fisher Fisher Fisher Fisher Fisher Fisher Fisher Uncompahgre Fisher Uncompahgre Uncompahgre Uncompahgre Uncompahgre Uncompahgre Uncompahgre Uncompahgre Uncompahgre Fisher Uncompahgre Fisher Fisher Fisher McKenzie Fisher Fisher 3- 4-88 3-12-88 3-18-88 3-25-88 4- 2-88 4- 8-88 4-15-88 4-22-88 4-29-88 5- 7-88 5-14-88 5-20-88 5-27-88 6- 3-88 6-10-88 6-17-88 6-23-88 6-30-88 7- 8-88 7-14-88 NW HE HE 28 30 18 30 13 13 8 18 19 30 18 13 32 7 21 27 28 1 35 13 46 46 46 46 46 46 46 46 46 46 46 46 47 46 46 46 46 45 46 46 8 8 8 8 9 9 8 8 8 8 8 9 8 8 8 9 8 9 9 9 258 256 256 256 254 254 257 256 255 256 256 254 257 256 258 251 259 253 252 254 4233 4233 4236 4233 4235 4235 4236 4235 4235 4232 4236 4235 4240 4237 4234 4233 4231 4229 4231 4235 2,164 2,316 2,196 2,3J6 2,286 2,286 2,164 2,225 2,316 2,347 2,225 2,286 2,196 2,377 2,196 2,469 2,225 2,438 2,377 2,286 3.6 2.3 3.8 2.8 3.2 0.4 3.4 1.9 0.4 2.6 4.0 2.1 5.4 3.0 4.5 7.3 8.3 6.6 2.8 4.5 Uncompahgre R Fisher Fisher Fisher McKenzie McKenzie Uncompahgre Fisher Fisher Uncompahgre Fisher McKenzie Uncompahgre McKenzie Uncompahgre Fisher Uncompahgre Pleasant Valley Cit Fisher McKenzie !'!E HE SE SW SW NW SE HE SE SW NW SW NE SE SW HE SE R R R R R R R R R R G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G F F G G G G F G G F G G Visual G also G G G ··F G G G G G G G G G G G G G G G G G G G F G G G G G F G G G G G G G G G F �218 Legal description Date 1/4 S NW SW SE SE NE SW '19 19 22 22 27 5 8 4 27 7 7 19 4 8 8 12 13 12 31 24 35 24 14 19 18 12 30 26 13 . 12 18 12 6 13 7 13 33 12 13 7 35 18 6 5 36 5 35 20 31 34 6 T U.T.K. R X Approx. elev (m) Y Distaoce (0) between locations Ratiog Major drainase Sisnal Location SUBADULT FEMALE PUMA 134 7-19-87 7-24-87 8- 5-87 8- 9-87 8-16-87 8-28-87 9- 4-87 9-13-87 9-18-87 9-26-87 10- 2-87 10- 9-87 10-16-87 10-23-87 10-29-87 ll-6-87 ll-13-87 ll-20-87 ll-28-87 12- 4-87 12-ll-87 12-21-87 12-29-87 1- 8-88 1-15-88 1-22-88 1-29-88 2- 5-88 2-12-88 2-19-88 2-26-88 3- 4-88 3-12-88 3-18-88 3-25-88 4- 2-88 4- 8-88 4-15-88 4-22-88 4-29-88 5- 7-88 5-14-88 5-20-88 5-27-88 6- 3-88 6-10-88 6-17-88 6-23-88 6-30-88 7- 8-88 7-14-88 liE NW liE NW SW SE SW liE liE SE NE SE SW HE NW HE SE SW NW SE SW NE NW SE NW SE NW HE NW NW SW SE NW SW NE NE SE SE HE HE HE SW SE NW SE 47 47 47 47 47 46 46 46 47 46 46 46 46 46 46 47 48 47 48 47 47 47 47 47 47 47 47 47 47 47 47 47 47 47 47 47 47 47 47 47 47 47 46 46 47 46 47 48 47 46 46 9 9 10 10 10 9 9 9 9 9 9 9 9 9 9 10 10 10 9 10 10 10 10 9 9 10 9 10 10 10 9 10 9 10 9 10 9 10 10 9 10 9 9 9 10 9 10 9 9 9 10 245 245 242 242 241 247 248 249 251 246 246 246 249 248 248 245 245 245 246 244 242 244 243 246 245 245 245 243 244 245 245 245 246 244 246 244 249 245 244 245 243 247 246 248 245 248 243 247 246 250 246 4245 4244 4244 4244 4243 4238 4238 4239 4243 4238 4237 4234 4239 4238 4237 4247 4246 4247 4250 4244 4241 4244 4245 4244 4246 4247 4242 4243 4246 4246 4246 4247 4249 4246 4248 4246 4239 4247 4245 4247 4241 4245 4238 4238 4241 4239 4241 4253 4240 4241 4238 2,256 2,256 2,377 2,408 2,438 2,438 2,377 2,377 2,103 2,469 2,499 2,621 2,316 2,408 2,438 2,225 2,196 2,134 2,012 2,286 2,469 2,196 2,256 2,256 2,134 2,103 2,316 2,316 2,164 2,134 2,134 2,196 2,042 2,225 2,134 2,196 2,256 2,134 2,256 2,134 2,469 2,164 2,408 2,377 2,347 2,347 2,438 1,890 2,408 2,164 2,438 4.1 1.0 3.6 0.4 1.1 7.0 1.5 1.4 4.1 7.9 0.6 3.7 5.5 1.2 0.5 9.8 0.7 0.4 3.6 5.0 4.3 4.1 1.5 2.8 2.3 1.5 5.0 2.0 3.5 1.0 0.4 0.7 2.7 3.5 2.3 2.7 7.9 8.0 2.5 1.5 6.0 5.5 6.8 2.0 4.4 3.3 4.7 12.1 5.0 4.1 5.1 Dolores DoloreB Happy Happy Happy W FIt Horsefly E FItHor8efly E Fk HorBefly HorBefly W FItHorsefly W FItHorsefly E FIt Horsefly E FItHorsefly E FIt Horsefly E FIt Horsefly Happy Happy Happy Happy Happy Happy Happy Happy DoloreB Happy Happy Dolnres Happy Happy Happy Happy Happy Happy Happy Dolores Happy Horsefly Happy Happy Happy Dolores Dolores W Fk Horsefly Cottoowood Dolores W Fk Horsefly Dolores Happy Wildcat Horsefly W FIt Horsefly G G G G G G G G F F G G G G F G a a a a a a a a a a a a a a G G G G G a a a a a a G a G a G a a a F F a a a a a a a a F a a a a F F a G a a a a G G G G G a G a a a G G a a G a G a a F G F a G G F G F a a S UBADULT FE.'!ALEPUMA 136 47 23 7-19-87 SE 8 14 47 7-24-87 SE 8 48 8- 5-87 NW 25 8 48 22 8- 9-87 NW 8 48 22 8-16-87 SW 8 48 21 8-28-87 SE S 48 21 9- 4-87 SE 8 22 48 SW 8 9- 5-87 Cause of death DOt established: 4243 262 2,286 1.6 Ooioo Ck 4244 262 2,347 1.5 Onioo Ck 4251 2,499 263 7.0 Beatoo 4254 261 2,225 3.5 Dry Cedar 4253 260 2,256 0.7 Dry Cedar 260 4253 2,134 0.4 Dry Cedar 4253 2,134 260 0.0 Dry Cedar 4253 261 2,256 0.0 A sharp booe (aviao 7) about 12" 100g va. found io a G G Ck Ck Ck Ck a a G ~brtality Fouod Dead stomach. F F F F F F SlgUoi SUBADULT FEMALE PUMA 137 7-19-87 7-24-87 8- 5-87 8- 9-87 8-16-87 8-28-87 9- 4-87 9-13-87 9-18-87 9-26-87 10- 2-87 10- 9-87 SE SE NE HE NW SW SE SE NE SW SE SE 13 23 II 11 24 24 35 11 24 35 13 26 48 48 48 48 48 48 48 48 48 48 48 48 12 12 12 12 12 12 12 12 12 12 12 12 750 748 748 747 749 749 748 748 749 748 749 748 4255 4253 4257 4257 4254 4253 4250 4256 4254 4250 4254 4253 2,469 2,560 2,408 2,377 2,438 2,256 2,621 2,499 2,408 2,591 2,469 2,560 0.7 1.5 3.9 0.4 4.2 0.6 3.1 6.4 3.1 6.6 4.8 1.7 West Fk Dry Ck Piney Cushman Cushman West FItDry Ck West FItDry Ck arsys Ck Cushmsn W Fk Dry Ck W Fk Dry Ck W FItDry Ck Pi~ey a G G a G G G a a a a a F F F G G F G G G a F G �219 Legal description Date Subadult 1/4 S T U.T.M. R X Y Appro:.:. elev (m) Distance (u) between locations RatiDg Major draioase Sisnal Location female ~uma 137 - continued 10-16-87 10-23-87 10-29-87 11-6-87 11-13-87 11-20-87 11-28-87 12-11-87 12-21-87 12-29-87 1- 8-88 1-15-88 1-22-88 1-29-88 2- 5-88 2-12-88 2-19-88 2-26-88 .3- 4-88 .3-12-88 3-18-88 3-25-88 4- 2-88 4- 8-88 4-15-88 4-22-88 4-29-88 5- 7-88 5-14-88 5-20-88 5-27-88 6- 3-88 6-10-88 6-17-88 6-23-88 6-30-88 7- 3-88 NW NW SE HE HE 5W sw 7-19-87 7-24-87 8- 5-87 8- 9-87 8-16-87 8-28-87 8-30-87 SE NE NE SE SE NE NE 24 48 12 36 48 12 23 48 12 24 48 12 48 26 12 24 48 12 19 48 11 NW 24 48 12 5E 24 48 12 5w 19 48 11 SE 48 18 11 SW 8 48 11 NW 17 48 11 HE 18 48 11 NW 48 17 11 NW 22 48 11 NW 9 48 11 5W 48 16 11 5W 48 22 11 HE 24 48 11 5E 7 48 11 5W 48 12 1 5W 48 12 1 5W 48 12 13 48 5E 12 35 5W 48 25 12 48 HE 22 12 5E 2 47 12 48 SE 12 35 5W 48 36 12 SW 36 48 12 48 SE 13 12 NW 24 48 12 SE 48 12 35 NW 47 12 12 SW 47 12 1 Shot as sheeplt111er 748 749 748 749 748 749 750 750 749 750 751 752 752 751 752 755 754 753 755 750 751 749 749 749 748 749 747 748 748 749 748 749 749 748 749 749 4254 4251 4253 4254 4252 4253 4253 4254 4253 4253 4255 4256 4255 4255 4256 4254 4257 4255 4254 4254 4256 4258 4258 4255 4250 4251 4254 4248 4250 4250 4250 4254 4254 4249 4248 4248 2,530 2,499 2,560 2,469 2,560 2,530 2,469 2,469 2,377 2,377 2,286 2,196 2,256 2,256 2,042 2,134 2,012 2,196 2,196 2,408 2,256 2,347 2,347 2,438 2,621 2,499 2,499 2,714 2,621 2,591 2,652 2,530 2,499 2,652 2,591 2,621 0.6 2.6 1.8 1.8 2.7 1.6 1.4 1.2 0.7 1.2 1.4 2.7 0.7 0.8 1.4 3.3 3.1 2.0 2.6 6.0 2.8 3.4 0.2 3.1 4.6 1.6 3.7 5.2 1.4 1.1 0.2 4.7 0.3 4.3 1.3 0.9 w FIt Dry Cit W Fit Dry Cit Piney w FIt Dry Cit Piney w FIt Dry Cit w FIt Dry Cit w FIt Dry Cit w FIt Dry Cit w FIt Dry Cit w FIt Dry Cit W FIt Dry Cit W FIt Dry Cit w FIt Dry Cit W FIt Dry cit E FIt Dry CIt E FIt Dry Cit E FIt Dry CIt E FIt Dry CIt w FIt Dry Cit W FIt Dry Cit Cushman Cushman Piney Gray's Cit W FIt Dry Cit Cushman W FIt Dry Cit Gray's Cit W Fit Dry Cit I< FIt Dry Cit Piney Piney W fit Dry Cit I< FIt Dry CIt I< Fit Dry Cit I< FIt Dry Cit G G G G G G G G G G G G G G G G G G G G G G G E E E E E E E G G G G F G G F G G F F G G F G G G G F G F G G G G G G G G G G G G G G G G G G G G G G G G G G G F G G G F F G ADULT MALE PUMA 138 34 3 3 34 34 3 3 Probably 48 47 47 48 48 48 48 11 11 11 11 11 11 11 756 756 756 756 756 756 756 4250 4249 4250 4250 4250 4250 4250 2,438 2,438 2,408 2,560 2,560 2,560 2,469 died between 5-15 and 5-22-87; 0.4 0.8 0.6 0.3 0.0 0.0 FIt Dry FIt Dry FIt Dry FIt Dry FIt Dry fk Dry fit Dry Ck Cit Cit Cit Ck Ck Ck F G C G G Mortality Signal Nonfunctional cause of deatb unknown ADULT MALE PUMA 140 7-19-87 7-24-87 8- 5-87 8- 9-87 8-16-87 8-28-87 9- 4-87 9-13-87 9-18-87 9-26-87 10- 2-87 10- 9-87 10-16-87 10-23-87 10-29-87 11- 6-87 11-13-87 11-20-87 11-28-87 12- 4-87 12-11-87 12-21-87 12-29-87 1-10-88 1-15-88 1-22-88 NW NW NE NE 51< NW SW SI< NW 5E NW NW NE NE NW NW NE NW sw NW NW NE HE NE 5W SE 9 17 1 28 23 7 28 22 8 18 8 27 8 5 27 20 10 8 5 32 17 17 2 5 21 20 45 45 45 46 46 45 46 46 45 46 46 46 45 45 46 46 45 45 46 46 45 46 45 45 46 46 9 9 10 9 9 98 9 9 8 8 9 9 8 9 8 9 9 8 8 9 9 9 8 8 8 248 247 244 250 252 245 259 250 246 257 257 250 248 258 250 257 250 247 257 255 247 248 252 257 258 258 4228 4227 4229 4233 4233 4228 4232 4233 4228 4235 4238 4233 4228 4229 4233 4234 4228 4228 4238 4239 4227 4236 4229 4229 4233 4233 2,714 2,438 2,836 2,530 2,408 2,928 2,134 2,499 2,745 2,225 2,225 2,530 2,621 2,347 2,499 2,103 2,499 2,652 2,134 2,377 2,438 2,438 2,499 2,438 2,196 2,164 7.9 1.8 3.7 6.4 2.1 8.7 14.4 9.1 6.3 11.9 2.5 8.1 5.4 10.3 8.2 6.9 9.5 3.7 15.0 7.3 9.5 9.9 8.5 4.8 4.6 0.5 Pleasant V Pleasant V McKenzie McKenzie McKenzie Pleasant V Uncompahgre McKenzie Pleasant V Fisher McKenzie McKenzie Pleasant V Dallas McKenzie Uncompahgre Pleasant V Pleasant V Uncompahgre Uncompahgre Pleasant V E fk IIorsefly Dallas Dallas Uncompahgre Uncompahgre G G G G G G G G G G G G G G G G G G G G G G G G G G F F G F F G G G G G G G G G G G G G G F G G G C r G �220 Legal description Date 1/4 S T U.I.M. R X Di8t8nce Appro>:. e1ev (m) y (lcm) between Rating locations Major drainase 4.1 3.2 8.7 11.8 8.5 7.0 13.0 13.5 0.6 3.7 4.3 0.4 6.2 7.7 0.2 0.3 1.0 10.7 13.0 3.1 6.4 9.1 0.8 1.0 Uncompahgre Fisher W FIt Horsefly Pleasant V Dallas Pleasant V Horsefly Pleasant V Pleasant V Dallas Uncompahgre Uncompahgre McJ(e=ie E FIt Horsefly E FIt Horaefly E FIt Horsefly Busted Boller D Pleasant V Cit Uncompahgre Fisher Dallas E FIt Horsefly E FIt Horsefly Busted Boller Sisna1 location Adult male 2WD8 140 - ~ont1nued 1-29-88 2- 5-88 2-12-88 2-19-88 2-26-88 3- 4-88 3-12-88 3-18-88 3-25-88 4- 2-88 4- 8-88 4-15-88 4-22-88 4-29-88 s- 7-88 5-14-88 5-20-88 5-27-88 6- 3-88 6-10-88 6-17-88 6-30-88 7- 8-88 7-14-88 HE SE NW HE ME HE NW HE NW SW NW SW HE NW NW HE HE SE HE SW HE NW SW HE 8 18 4 9 5 10 34 10 11 6 33 28 7 16 16 6 16 18 32 17 1 16 16 16 46 46 46 45 45 45 46 45 47 45 46 46 46 46 46 46 46 45 46 46 45 46 46 46 8 8 9 9 8 9 9 9 9 8 8 8 8 9 9 9 9 9 8 8 9 9 9 9 258 256 248 249 258 251 250 251 251 255 259 259 256 249 249 749 249 246 259 257 254 249 249 249 4238 4235 4240 4228 4229 4228 4241 4228 4228 4229 4231 4032 4238 4236 4236 4236 4236 4225 4231 4235 4329 4236 4236 4236 2,103 2,225 2,316 2,560 2,316 2,591 2,164 2,530 2,499 2,316 2,256 2,225 2,256 2,408 2,438 2,438 2,438 2,714 2,286 2,164 2,408 2,377 2,438 2,377 G G G G G G G G G G G G G G G G G G G G G G G G G F G G G F G G G G G G G F G G F G G F G G G F SUBADULT MALE PUMA #42 1-21-88 1-22-88 2-18-88 3-17-88 4-20-88 5-10-88 5-25-88 6-10-88 6-29-88 7-12-88 SE HE NW SE ME HE SW NW SW NW aAl1 locations 31 29 33 1 10 27 36 20 10 23 3 3 2 3 2 4 4 4 5 5 by Wildlife 86a 86 86 88 86 87 87 87 87 87 318 318 323 310 328 314 319 311 '311 315 4401 4403 4410 4410 4418 4389 4391 4381 4386 4370 124.0 3.0 8.0 14.0 19.0 27.0 4.0 10.0 10.0 2.0 Sweetwater Horse E FIt Red Cit Sweetwater Derby Deep Colorado R Deep Tie Gulch Jaclt Cit Recapture Recapture Biologist G. Byrne ADULT MALE PUMA #44 7-19-87 7-24-87 8- 5-87 8- 9-87 8-16-87 8-28-87 9- 4-87 9-13-87 9-18-87 9-26-87 10- 2-87 10- 9-87 10-16-87 10-23-87 10-29-87 11- 6-87 11-13-87 11-20-87 11-28-87 12- 4-87 12-11-87 12-21-87 12-29-87 1- 8-88 1-15-88 1-22-88 1-29-88 2- 5-88 2-12-88 2-19-88 2-26-88 3- 4-88 3-18-88 3-25-88 4- 2-88 4-15-88 4-22-88 SE SE SW NW SW NW NW SW ME SE SW SW HE HE SW SE SE SW SE HE HE SE SE ME SW SE NW NW NW NE NE NE SW NW ME NW SW 34 24 17 16 7 29 30 17 9 34 31 25 10 17 17 18 15 30 21 4 33 21 19 17 21 8 10 36 16 25 16 19 27 10 2 29 18 15 15 51 15 51 15 15 15 15 15 15 15 15 15 15 15 51 15 15 50 14 15 15 15 15 15 15 14 15 15 15 15 14 51 15 51 51 101 101 15 101 15 100 100 100 101 101 100 101 100 100 101 100 17 100 100 17 100 15 99 100 100 99 100 100 99 100 99 99 100 15 99 15 14 703 706 712 701 709 708 706 708 702 702 707 705 712 709 699 707 696 707 711 694 711 711 717 709 710 719 712 715 720 715 720 718 711 714 724 715 719 4285 4289 4284 4291 4285 4288 4288 4291 4293 4285 4285 4285 4293 4291 4290 4291 4283 4287 4289 4277 4296 4289 4289 4292 4289 4292 4293 4297 4292 4288 4292 4290 4298 4285 4295 4281 4284 2,530 2,560 2,438 2,591 2,560 2,408 2,377 2,286 2,530 2,560 2,499 2,499 2,225 2,377 2,652 2,438 2,682 2,560 2,286 2,256 2,286 2,316 2,408 2,286 2,377 2,286 2,286 2,134 2,286 2,469 2,196 2,408 2,286 2,377 2,073 2,408 2,164 1.5 5.0 2.7 13.2 11.0 3.6 1.0 2.6 8.5 7.7 5.2 2.5 9.2 3.7 10.0 8.5 13.9 12.2 1.8 20.0 25.2 6.9 6.3 8.8 3.2 10.2 7.5 4.7 0.6 5.4 6.1 3.2 9.9 12.9 13.7 18.7 8.1 Lafair Big Dominguez Keith N ern Cit Keith Cit Big Dominguez Big Dominguez Big Dominguez West Beaver Big Dominguez Smith GUbler Big Dominguez Gill Big Dominguez Cow Cit Big Dominguez Big Dominguez Indian Gibb1er Big Dominguez Wlldhorse D Roclty Pitch Gulch Big Dominguez Dry FIt Big Dominguez Gibb1er GUbler Little Dominguez Wlldhorse D Little Dominguez Little Dominguez Farmer's Little Dominguez Dry Fit Big Dominguez Keith Cit Rose Cit G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G F F F F G G G G G F G G G G G G G G G F F G G F F F F F F G F F G F F G F �221 Legal description Date ---------1/4 S T R U.T.M. X Y Approx. e1ev (m) Distance (km) between locations Rating Major drainalle Si~a1 Location Adult male Euma 144 - continued 4-29-88 5- 7-88 5-14-88 5-20-88 5-27-88 6- 3-88 6-17-88 6-23-88 6-30-88 7- 8-88 SW SE SW SW HE HE SE NW HE SW 16 4 35 30 35 17 4 2 31 22 51 50 51 51 15 50 15 15 15 15 15 15 15 15 101 15 99 101 100 100 713 714 715 710 705 712 720 703 708 711 4283 4277 4279 4280 4285 4275 4294 4294 4286 4289 2,408 2,682 2,621 2,499 2,469 2,560 2,134 2,408 2,499 2,225 6.4 5.9 2.6 5.8 8.0 13.5 20.8 17.2 9.1 5.0 Keith CIt Corral G Corral G Keith Smith H fit Escalante Big Dominguez West Big Dominguez Big Dominguez G G G G G G G G G G F G F F G F F F G G Traver Criswell Monitor Potter Terrible Moore Traver Criswell Traver Coa1bank Roubideau Cushman Criswell Criswell Wright Roubideau Criswell Moore Roubideau Potter Moore Criswell Potter Criswell Roubideau Potter Criswell Criswell Roubideau Criswell Roubideau Potter Roubideau Criswell G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G F F F F F G G G G G G G F G G G G G F G G G G G G G F F G Roubideau Roubideau Roubideau Roubideau Roubideau Roubideau Roubideau Roubideau Moore Roubideau Traver Roubideau Wright Traver G G G G G G G G G G G G G G F F F F F G F F F F F G G F G F G F F ADULT FEMALE PUMA 145 7-19-87 7-24-87 8- 5-87 8- 9-87 8-16-87 8-28-87 9- 4-87 9-13-87 9-18-87 9-26-87 10- 2-87 10- 9-87 10-16-87 10-23-87 10-29-87 11- 6-87 11-13-87 11-20-87 11-2li-87 12- 4-87 12-11-87 12-21-87 12-29-87 1- 8-88 1-15-88 1-22-88 1-29-88 2- 5-88 2-12-88 2-19-88 2-26-88 3- 4-88 3-12-88 3-18-88 3-25-88 4- 2-88 4- 8-88 4-15-88 4-22-88 4-29-88 5- 7-88 5-14-88 5-20-88 5-27-88 6- 3-88 6-10-88 6-23-88 6-30-88 7- 8-88 7-14-88 SE SW 29 49 12 741 4262 2,196 2.9 23 50 12 745 4273 12.2 1,859 15 49 13 734 4265 2,438 13.3 SE 22 49 13 734 4263 2,377 2.2 NW 5 48 12 742 4259 2,377 8.4 SE 9 49 12 742 4266 1,951 7.2 NW 29 49 12 740 4262 2,408 4.5 NW 24 49 13 737 4264 2,316 3.2 NW 29 49 12 740 4263 2,377 3.1 SE 9 49 11 752 4267 1,890 12.6 SW 10 50 12 744 4276 1,890 11.9 SE 29 49 11 751 4262 2,164 15.7 NE 5 49 12 741 4269 2,103 11.7 SW 19 49 12 740 4263 2,286 6.4 NW 33 49 12 742 4261 2,164 3.1 SE 4 49 12 743 4268 1,951 7.4 SW 4 49 12 742 4269 1,951 1.1 sO' 20 49 12 740 4263 2,256 5.3 NI,' 16 49 742 12 4266 2,012 3.1 SW 31 50 12 740 4270 1,981 4.8 NE 17 49 12 742 4266 2,164 4.4 NW 4 49 12 742 4269 1,920 4.0 NE 32 50 12 741 4271 2,012 1.6 NW 4 49 12 742 4269 1,890 1.9 HE 35 50 12 745 4271 1,890 5.1 SW 31 12 50 739 4270 2,103 6.4 SW 27 50 12 743 4272 1,920 4.6 NW 9 49 12 742 4267 2,103 4.7 SW 34 50 12 743 4270 1,981 2.9 SE 33 50 12 743 4270 1,859 1.0 NE 3 49 12 744 4269 1,798 2.1 SE 21 50 12 742 4273 1,951 4.3 SE 4 49 12 743 4268 1,829 4.7 HE 33 50 12 743 4271 1,829 2.5 No location - Poor signal over Criswell, Moore, and Roubideau No signal - Extensive search HE 3 49 744 12 4269 1,768 2.0 SE 4 49 12 743 4268 1,951 1.6 SE 4 49 12 743 4269 1,951 0.2 NW 3 49 12 743 4270 2,316 1.2 SW 744 3 49 12 4269 1,829 1.0 NW 21 49 12 742 4265 2,164 4.4 SE 49 9 12 743 4266 1,859 2.0 NE 16 49 12 743 4266 1,920 0.2 SW ·16 49 742 4265 12 2,073 5.5 HE 16 49 4266 12 743 1,920 1.2 SE 49 29 12 742 4261 2,196 4.5 SE 21 49 12 743 4264 1,951 2.7 NW 33 49 12 742 4261 2,042 2.7 SW 28 49 12 742 4262 1,981 0.7 NW F F F G G ADULT FEMALE PUMA 146 7-19-87 SE 32 51 15 711 4279 2,438 ll.6 Red Ck 7-24-87 SE 16 51 15 714 4284 2,316 5.4 Keith Ck 8- 5-87 HE 28 51 714 15 4281 2,316 2.4 Red Ck 8- 9-87 SI:: 32 51 702 16 4279 2,682 11.5 Big Dominguez Found dead during first week in August by C. Black - Cause of death unknown G G G Black had removed collar and took same to his summer camp just south of Coldwater Ranger Station where I aerially located it on 8-9-87. G. Bock picked up skull from carcaas in mid-August. The approximate location was SE 1/4, 517, T51N, Rl5W 712-4283 7,900' on N Fk of Keitb Ck by Stockpond. �222 Legal description Date 1/4 S T R U.I.M. X Y Appro",. e1ev (m) Distance (kID) between locations Rating Major drainal!e Sis:!!al Location SUBADULI MALE PUMA #47 7-19-87 7-24-87 8- 5-87 8- 9-87 8-16-87 8-28-87 9- 4-87 12- 4-87 12-11-87 12-21-87 12-29-87 1-10-88 1-15-88 1-22-88 1-29-88 2- 5-88 2-12-88 2-19-88 2-26-88 3- 4-88 3-12-88 3-18-88 3-25-88 4- 2-88 4-15-88 4-22-88 4-29-88 5- 7-88 5-14-88 5-20-88 5-27-88 6-10-88 6-17-88 6-23-88 6-30-88 7- 8-88 SW SW 29 36 12 8 7 24 21 29 35 32 22 1 26 11 15 7 3 31 32 23 27 32 5 6 34 33 25 3 33 30 23 3 14 3 29 35 13 13 14 13 13 13 13 13 13 13 13 14 12 12 12 12 12 11 11 11 11 20 21 21 12 12 13 14 13 13 13 13 13 13 13 11 103 679 104 676 103 685 102 688 102 688 103 685 102 690 103 679 104 675 103 679 104 673 103 686 103 684 103 684 103 684 103 677 103 682 102 686 102 688 103 684 103- 681 26 667 26 667 26 665 102 691 102 691 101 685 104 672 102 691 101 697 102 694 104 673 102 243 102 692 101 685 103 _684 4306 4304 4303 4312 4312 4309 4308 4306 4305 4306 4309 4304 4317 4322 4320 4322 4323 4324 4325 4328 4327 4321 4319 4319 4315 4315 4207 4304 4306 4308 4309 4312 4310 4314 4207 4325 7.0 3.9 10.1 11.0 - 0.6 4.3 1.0 11.8 4.2 4.4 6.7 14.0 15.0 4.7 1.4 5.9 4.7 5.0 1.5 5.0 2.5 15.0 1.6 2.3 25.2 0.9 9.2 13.2 18.3 6.6 3.0 21.3 20.6 3.6 8.9 17.8 Briar Can KcXetl%ie Sheep Clark Wash Clark Wash Dierich Ck Little Dolores Briar McKenzie Briar Hill T Dodson Can Trail Can Sieber Can - Sieber Can Little Dolores R Seiber Can Seiber Can 28-Hole Wash Kee Knowles Jones Little Dolores Little Dolores Clark Wash Payne Wash Dierich Granite Little Dolores Ladder Clark Wash Coates Clark Wash Clark Wash Dierich 28-Hole Wash G G G G G G G G G G G 7-19-87 SE 10 47 11 756 4247 2,438 1.7 E Fk Dry Ck 7-24-87 SW 21 48 11 754 4253 2,316 0.7 E Fk Dry Ck 8- 5-87 HE 11 47 11 758 4249 2,469 6.8 E Fk Dry Ck 8- 9-87 SW 2 47 11 4248 757 2,438 1.9 E Fk Dry Ck 8-16-87 HE 4 47 11 755 4249 2,438 2.8 E Fk Dry Ck 8-28-87 SW 27 48 11 755 4252 2,286 3.3 E Fk Dry Ck 9- 4-87 SE 10 47 11 756 4246 2,560 4.5 E Fk Dry Ck 9-13-87 34 11 SW 48 755 4250 2,408 3.4 E Fk Dry Ck 9-18-87 HW 27 48 11 756 4252 2,164 2.4 E Fk Dry Ck 9-26-87 SE 34 48 11 756 4250 2,438 2.0 E Fk Dry Ck 10- 2-87 HE 47 11 753 5 4249 2,560 3.2 E Fk Dry Ck 10- 9-87 NW 4244 20 47 11 752 2,806 4.6 E Fk Dry Ck 10-16-87 SW 33 48 11 754 4250 2,530 5.7 E Fk Dry Ck 10-23-87 HE 4245 19 47 11 752 2,682 4.6 E Fk Dry Ck 10-29-87 SW 11 755 4250 34 48 2,408 6.2 E Fk Dry Ck 11-6-87 4251 HE 34 48 11 756 2,316 1.2 EFkDzyCk 11-13-87 11 4247 SE 10 47 756 2,499 3.8 E Fk Dry Ck 11-20-87 4244 HW 19 47 10 760 2,438 5.3 KiddIe Fk Spring 11-28-87 4254 HE 21 48 11 754 2,103 11.8 E Fk Dry Ck ;., 11 12- 4-87 4241 SW 759 36 2,682 5.6 KiddIe Fk Spring 12-21-87 4240 SW 241 34 47 10 2,530 7.6 Happy 12-29-87 SW 4240 32 47 10 762 2,682 4.1 E Fk Spring 1- 8-88 HW 4251 35 48 11 757 2,377 11.5 W Coal Ck 1-15-88 HE 29 4243 47 238 2,499 10 6.1 E Fk Spring 1-22-88 No location - Poor signal vicinity of 1-15-88 location; 1/2-hr search 1-29-88 SW 4250 31 48 10 760 2,286 8.0 Lindsay Can 2-12-88 HE 4240 5 46 9 248 2,286 16.5 W Fk Horsefly 2-19-88 HW 4237 9 46 249 2,347 9 2.4 E Fk Horsefly 2-26-88 HE 4239 5 46 8 248 2,347 2.2 W Fk Horsefly Found dead from unknown causes 2-27-88; skull and radioco11ar recovered G G G G G G HW HW HE HW SW SW SE HW HE HE HW HW HE HW HW SW HW SW HW HW HW SW SW HE HW HW HW HW NE SE SW SW HW HW 2,438· 2,560 2,682 2,134 2,164 2,225 2,316 2,438 2,499 2,377 2,316 2,560 2,012 1,920 1,951 1,890 1,890 1,981 1,981 1,829 1,829 1,646 1,707 1,586 2,073 2,073 2,408 2,499 2,377 2,652 2,377 2,196 2,347 2,196 2,408 1,890 G G G G G G G G G G G G G G G G G G G G G G G G G F F F F G F G F F F F G F F F G G G F F G F F G F F G G G F F G G G G G SUBADULI MALE PUMA #50 G G G G F F F G G G G G G G G G G F F G G G G G G G G G G G G G G G Mortality Mortality G G G G G G G F G G F G F G �223 Legal description U.T.M. Approx. e1ev (m) 1/4 5 r R 7-19-87 7-24-87 8- 5-87 8- 9-87 .8-16-87 8-28-87 9- 4-87 9-13-87 9-18-87 9-26-87 10- 2-87 10- 9-87 10-16-87 10-23-87 10-29-87 11- 6-87 11-13-87 11-20-87 11-28-87 12- 4-87 12-11-87 12-21-87 12-29-87 1- 8-88 1-15-88 1-22-88 1-29-88 2- 5-88 2-12-88 2-19-88 2-26-88 3- 4-88 3-12-88 3-18-88 3-25-88 4- 2-88 4- 8-88 4-15-88 4-22-88 4-29-88 5- 7-88 5-14-88 5-20-88 5-27-88 6- 3-88 6-10-88 6-17-88 6-23-88 6-30-88 7- 8-88 7-14-88 SE HE SE SW HE SW HE SW HE HE SW HE NW NW SE NW NW SW NW NW SW SE SE SW SE NW NW SE NW NW NW HE SE NE SW NW NW NE SW HE HE SW SW HE NW SE SE HE HE SW SE 5 8 33 4 8 4 35 26 22 18 28 7 16 30 7 9 34 5 4 34 30 19 9 28 18 4 7 5 34 21 18 27 18 21 32 29 4 22 33 1 24 19 24 5 8 9 5 16 24 9 8 46 46 47 46 46 46 47 47 47 47 47 46 46 47 46 47 47 46 46 47 47 47 46 47 47 46 47 46 47 47 47 47 47 47 47 47 46 47 47 46 47 47 47 46 46 46 46 46 46 46 47 9 9 9 9 9 9 10 10 9 9 9 9 9 9 8 9 9 9 9 9 9 9 9 9 9 9 9 9 10 9 9 9 9 9 8 9 9 9 9 10 10 9 10 9 9 9 8 9 10 9 9 248 248 249 249 248 249 243 242 252 247 249 246 249 245 256 249 251 247 249 250 246 246 249 249 246 249 245 248 250 249 245 251 747 250 256 247 249 251 249 245 244 245 244 248 247 249 258 249 245 249 247 4239 4238 4240 4239 4238 4238 4241 4242 4244 4246 4242 4238 4236 4243 4237 4247 4241 4239 4239 4240 4242 4244 4236 4241 4245 4239 4247 4238 4241 4244 4246 4243 4245 4244 4240 4243 4239 4244 4240 4239 4244 4244 4243 4240 4237 4237 4238 4236 4234 4237 4245 2,377 2,377 2,347 2,316 2,408 2,377 2,438 2,438 2,042 2,225 2,225 2,438 2,408 2,316 2,225 2,012 2,134 2,408 2,316 2,164 2,347 2,347 2,347 2,256 2,134 2,256 2,134 2,408 2,164 2,225 2,134 2,103 2,256 2,134 2,196 2,286 2,256 2,073 2,316 2,438 2,256 2,256 2,256 2,377 2,438 2,408 2,164 2,408 2,621 2,438 2,196 12-19-87 12-21-87 12-29-87 1- 8-88 1-15-88 1-22-88 1-29-88 2- 5-'88 2-12-88 2-19-88 2-26-88 3- 4-88 3-12-88 3-18-88 3-25-88 4- 2-88 4- 8-88 4-15-88 .4-22-88 4-29-88 5- 7-88 5-14-88 5-20-88 5-27-88 6- 3-88 6-10-88 &-17-88 6-23-88 6-30-88 7- 8-88 7-14-88 HE SW SW SE 8 19 27 9 16 27 9 4 9 9 36 16 22 16 27 28 28 27 9 16 17 8 21 16 17 9 22 21 18 20 20 48 48 48 48 48 48 48 48 48 48 49 48 48 48 48 48 48 48 48 48 48 48 48 48 48 48 48 48 48 48 48 11 11 11 11 11 11 11 :11 11 11 11 753 751 756 754 754 756 755 754 754 755 756 755 756 754 755 754 755 755 753 753 752 753 754 754 753 754 755 754 751 753 753 4257 4253 4252 4257 4255 4251 4258 4258 4258 4257 4260 4255 4254 4255 4253 4253 4252 4252 4256 4256 4256 4257 4254 4255 4256 4257 4253 4253 4255 4253 4254 1,981 2,397 2,408 1,951 2,012 2,347 2,103 1,981 2,073 2,073 2,073 2,134 2,225 2,134 2,256 2,286 2,286 2,408 2,134 2,164 2,012 2,073 2,256 2,134 2,164 1,920 2,256 2,256 2,196 2,286 2,256 Date X Y Distance (Iaa) between locations Rating Major draInase SIsna1 Locatio) ADULT MALE PUMA 152 1.6 0.9 2.1 1.3 1.1 0.6 5.7 1.4 9.8 5.1 5.7 3.6 3.6 8.3 12.4 13.0 ~.6 4.4 2.5 2.8 4.7 1.5 7.0 4.2 4.3 6.4 9.0 9.5 3.6 3.6 4.0 6.5 4.8 3.1 7.7 9.7 4.1 5.5 4.2 3.9 4.7 1.1 1.2 2.3 2.3 2.3 8.2 8.5 5.1 4.5 8.5 Horsefly Horsefly Horsefly Horsefly Horsefly E Fit Horsefly Dolores Happy Horsefly Dolores Wildcat W FIt Horsefly E FIt Horsefly Dolores McKenzie Dolores Horsefly W FIt Horsefly E FIt Horsefly Horsefly Dolores Dolores E FIt Horsefly Wildcst Can Dolores Horsefly Happy E FIt Horsefly Horsefly Horsefly Happy Horsefly Dolores Wildcst Uncompshgre Horsefly E FIt Horsefly Horsefly Horsefly Dolores Happy Happy Happy W FIt Horsefly E FIt Horsefly E Fk Horsefly Uncompahgre Busted Boiler D Cottonwood E FIt Horsefly Dolores G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G F F F P F G G G G F G G G G F G G G G G G G F G G F F G F G G G F G F G G G F G G G G F F G G F G G FEMALE PUMA #53 sw SE HE SW NW HE SW SW NW SW HE NE NE SW SW NW NW HE NW SW HE NW SW SW HE SE HE II 11 II 11 II 11 II 11 II 11 II 11 11 11 II 11 II 11 II 11 Capture 5.2 5.3 5.3 1.5 4.2 6.1 1.1 0.3 1.2 3.8 5.9 1.4 2.0 2.2 0.9 0.7 0.6 4.5 0.7 1.2 1.5 3.6 1.2 1.0 1.7 4.2 0.9 3.2 2.7 1.0 W FIt Dry w FIt Dry E FIt Dry E FIt Dry E FIt Dry E FIt Dry E Fit Dry Dry Cit W FIt Dry E FIt Dry Dry Cit E FIt Dry E Fit Dry E FIt Dry E FIt Dry E FIt Dry E FIt Dry E FIt Dry E FIt Dry E Fit Dry W Fit Dry W Fit Dry E Fit Dry E Fit Dry E Fit Dry E Fit Dry E FIt Dry E FIt Dry W Fit Dry E Fit Dry E FIt Dry Cit CIt CIt CIt CIt CIt cit CIt Cit CIt Cit Cit Cit Cit Cit Cit Cit Cit Cit Cit Ck Cit CIt Cit Cit CIt Cit Cit Cit G G G G G G G G G G G G G G c .,; G G G G G G G G G G G G G G G G G G G G G G F G F F G F G G G F F G F G G G G F G G G G �224 Legal d~scription Date 1/4 S T U.T.M. R X Y Appro". ~lev (m) Distanc~ (0) between locations Rating Major drainalle Si!l!!!!l Location FEMALE PUMA 154 1-11-88 1-15-88 1-22-88 1-29-88 2- 5-88 2-12-88 2-19-88 2-26-88 3- 4-88 3-12-88 3-18-88 3-25-88 4- 2-88 4- 8-88 4-15-88 4-22-88 4-29-88 s- 7-88 5-14-88 5-20-88 5-27-88 6- 4-88 6-10-88 6-17-88 6-23-88 6-30-88 7- 8-88 7-14-88 NIl NIl NIl SW HE HE NIl NIl SW SE HE SE SE SE NIl SW SE HE SE HE SE SE NIl SE NIl SE SE NIl 27 27 26 12 25 30 26 26 12 12 8 4 34 4 36 4 5 31 36 34 13 33 19 23 19 19 24 19 50 50 50 50 15 14 14 14 14 14 14 14 14 14 13 14 14 13 13 13 13 12 12 12 12 12 12 12 14 14 14 14 98 3 98 98 97 97 97 96 96 96 96 97 96 96 97 96 96 95 94 95 95 95 96 95 724 724 726 727 735 735 732 732 743 744 747 749 750 749 753 748 748 746 744 751 754 759 244 243 755 755 754 754 4272 4272 4272 4276 4289 4296 4298 4298 4303 4303 4304 4304 4305 4305 4307 4303 4304 4306 4306 4307 4311 4216 4320 4320 4320 4320 4319 4320 1,861 1,859 2,012 2,225 1,737 1,556 1,951 1,646 1,890 2,012 2,134 2,347 2,316 2,225 2,134 2,256 2,256 2,621 2,438 2,377 2,316 2,438 2,591 2,682 3,111 3,050 3,111 3,172 Capture 0.2 1.7 4.4 11.6 7.2 3.2 0.2 12.8 1.0 3.3 1.7 2.3 1.6 4.3 5.6 0.3 3.0 1.6 6.6 5.0 7.2 6.5 1.7 7.4 1.0 1.4 1.0 Middle FIt Esca1ant~ Middle FIt Escalant~ E FIt Escalante Esca1ant~ Escalant~ Gunnison Gunnison Gunnison Point Cit Point CIt Gunnison Dry Gulch Negro Dry Gulch Beebe Gulch Gunnison Gunnison Alkali Point W FItDoughspoon Oalt Sand Kiser Ward Dirty George Dirty George Dirty George Dirty George G G G G G G G G G G G G G G G G G G G G G G G G F G G F F G F G G G F G G G F F F F G F G F G F F F F F G G MALE PUMA 155 2- 5-88 2-12-88 2-19-88 2-26-88 3- 4-88 3-12-88 3-18-88 3-25-88 4- 2-88 4-15-88 4-22-88 4-29-88 5- 7-88 5-14-88 5-20-88 5-27-88 6- 4-88 6-10-88 6-17-88 6-30-88 7- 8-88 7-14-88 NIl SW SW SW NIl 2 49 13 736 4269 34 51 13 734 4280 3 50 13 734 4278 7 50 12 738 4277 30 12 50 4272 738 110 location - Momentary wealt signal NIl 23 50 13 4274 735 SE 19 50 12 740 4273 NIl 29 50 13 731 4272 SE 31 50 12 739 4270 NE 2 50 13 736 4269 SE 49 11 13 736 4267 NE 30 50 13 4272 730 SW 23 49 13 735 4263 SE 11 49 13 736 4266 NE 10 49 14 725 4267 SE 5 49 13 732 4268 33 NE 49 13 4261 733 SW 15 49 13 734 4264 NIl 27 49 4263 13 734 NE 11 49 14 727 4267 NIl 20 49 12 741 4264 2,073 Capture Little Monitor 2,073 10.8 Dry FIt Escalante 1,829 2.2 Dry FIt Escalante 1,951 4.5 Cottonwood 2,042 4.8 Monitor over Monitor and Potter drainages 2,164 3.6 Cottonwood 2,103 4.4 Potter 2,196 9.0 Dry FIt Escalante 2,134 8.8 Potter 2,225 2.8 Monitor 2,438 2.5 Monitor 2,073 8.5 Dry FItEscalante 2,377 10.9 Potter 2,286 3.6 Potter 2,377 11.5 E FIt Escalante 2,256 7.0 Cottonwood 2,499 7.2 Potter 2,499 4.0 Monitor 2,438 1.8 Potter 2,377 9.0 Dry FItEscalante 2,256 14.0 Moor~ G G G G G G G G G G F G F G G G G G G G G G F G F G G F G G G G F G G G G G G G FEMALE PUMA 156 3- 3-88 3- 4-88 3-12-88 3-18-88 3-25-88 4- 2-88 4- 8-88 4-15-88 5-14-88 5-20-88 5-27-88 6- 4-88 6-10-88 6-17-88 6-23-88 6-30-88 7- 8-88 SE SW NIl NE NE SW NIl SE NE SW SE SW NE NIl SE NIl NIl 36 31 36 14 16 9 24 34 22 30 26 7 6 36 23 21 28 51 51 15 51 51 51 50 50 50 51 51 50 50 51 51 51 51 14 13 99 14 14 15 14 14 15 15 16 15 15 16 15 15 15 728 729 725 726 723 712 727 725 715 710 707 710 711 708 713 713 713 4280 4279 4287 4285 4284 4285 4274 4270 4274 4280 4280 4276 4279 4288 4282 4282 4281 1,859 2,103 2,134 2,134 2,316 2,530 2,196 2,196 2,530 2,530 2,621 2,682 2,560 2,714 2,377 2,347 2,134 Capture 1.5 8.4 2.5 6.6 10.2 18.5 9.1 10.5 9.0 2.6 5.5 3.0 3.4 11.4 4.8 2.3 Escalante Escalante Little Doalnguez Palmer Little Dominguez Keith Escalante E FIt Escalante Kelso Keith Big Dominguez N Fk Escalante Red Ck Keith Red Ck Keith Red Ck G G G G G F F F F F G F F G G F G G G G G G G G G G G F G G G F �225 Legal description Date -------1/4 T R S U.T.M. X Y Approx. elev (m) Distance (Itm) between locations Rating Major drainage Signal Locat Ior, FEMALE PUMA 157 4-15-88 4-22-88 4-29-88 5- 7-88 5-14-88 5-20-88 5-27-88 6- 4-88 6-10-88 6-17-88 6-23-88 6-30-88 7- 8-88 7-14-88 HE I,E I;W HE SW NW HE SW NW NW SE SW SW SW 23 20 14 14 30 26 26 26 7 7 28 5 20 28 50 50 50 50 SO 50 50 15 50 48 49 48 49 51 13 12 13 13 12 13 13 99 12 12 12 12 12 13 736 741 735 736 738 735 736 723 739 740 743 742 740 732 4274 4274 4275 4276 4272 4272 4272 4288 4278 4257 4262 4257 4263 4281 1,920 2,012 2,164 2,042 2,073 1,890 2,196 2,012 1,920 2,560 1,951 2,316 2,256 2,134 Capture 4.8 6.3 1.3 4.6 3.0 1.2 20.0 18.5 20.9 6.0 4.8 6.1 19.4 Cottonwood Potter Grade Gulch Cottonwood Monitor Cottonwood Cottonwood Little Doainguez Cottonwood Terrible Traver Long Moore Escalante G G G G G G G G G G G G G F G G F F F G F F G F G F ��227 Wildlife Research Report July 1988 JOB PROGRESS REPORT State of Colorado Project No. W-153-R-2 -----------------------No. 8A ---------------------- Work Plan Mammals Research Small Carnivorous Mammal Investigations Job No. 1 Period Covered: July 1, 1987 - June 30, 1988 Author: Development of River Otter Reintroduction Procedures T. D. I. Beck Personnel: S. Boyle, C. Haynes ABSTRACT Agreement was reached to trade for river otters with Oregon Department of Fish and Wildlife. Formal agreements between Colorado Division of Wildlife and BLM were completed. Fish surveys provided data on fish species distribution in the Dolores River but were inadequate to estimate fish biomass. Crawfish were widespread in distribution throughout the study area. River otter prey appears to be adequate for reintroduction. A detailed study plan was prepared, peer reviewed, and approved. ��229 DEVELOPMENT OF RIVER OTTER REINTRODUCTION PROCEDURES Thomas D. I. Beck P. N. OBJECTIVE Develop procedures for river otter reintroductions in Colorado and establish a self-sustaining population of river otters from which to collect river otters for future translocations. SEGMENT OBJECTIVES 1. Introduce up to 20 river otters into the Dolores River drainage. 2. Monitor all river otter release sites in Colorado to evaluate success of past reintroductions. 3. Develop techniques to monitor survival, reproduction, dispersal of river otters after reintroduction. dispersion, and METHODS AND MATERIALS Surveys of fish distribution were conducted unevenly along the 98-mi study stretch of the Dolores River. Hoop, seine, and gill nets were used to sample various habitats. Passive capture nets were set in late afternoon and pulled 12-16 hrs later. All captured fish were measured (cm) and weighed (gms) and some collected for reference collections. Crawfish were also collected and preserved in formalin. Nearly all sampling was canoe-based as road access is limited to a 10-mi stretch below Dove Creek Pump Station and 4 road crossings. Substantial time was devoted to revising the study plan and coaxing the Environmental Analysis Report (EAR) through the BLM channels. RESULTS AND DISCUSSION An approved study plan was completed in May, 1988 (Appendix A). The EAR was approved by the Colorado Division of Wildlife and the Colorado State Office of BLM in June, 1988. River otters were to be received fro~ the Oregon Department of Fish and Wildlife in late summer 1988. Procedures (surgeons, anesthesia, operating room) were completed for the surgical implantation of radio transmitters into the river otters. Crawfish samples (n ~ 15) were sent to Dr. Horton H. Hobbs, Jr. at the Smithsonian Institute for identification. The samples from the Dolores River were identified as "probably" Orconectes causeyi (H. H. Hobbs, Jr., pers. comm.). The 3 species of the Virilis Group (Q. nais, Q. causeyi, Q. virilis) of the subgenus Gremicambarus are nearly indistinguishable from morphological characters (Fitzpatrick 1987). It is unknown if crawfish were native to the �230 Dolores River, but it is known that the Colorado Division of Wildlife released crawfish (species and origin unknown) into the drainage in 1954 (Klein 1955). Pending more detailed studies of the Virilis Group, the crawfish of the Dolores River will be considered as o. causeyi. Crawfish were collected throughout the 98-mi study stretch of the Dolores River. Twenty-two nights of gill netting resulted in the capture of 172 round tail chubs (Gila robusta), 99 flannelmouth suckers (Catostomus latipinnis), 41 rainbow trout (Salmo gairdneri), 15 brown trout (Salmo trutta), 14 bluehead suckers (Catostomus discobolus), 13 channel catfish (Ictalurus punctatus), 3 green sunfish (Lepomis cyanellus), 1 carp (Cyprinus carpio), 1 mottled sculpin (Cottus bairdi), and 1 Snake River cutthroat (Salmo clarki). Gill netting was the most effective technique tried. The abundance of boulder-dominated river bottoms made all netting attempts relative inefficient. The deep, slow pools of Slickrock Canyon (RM 103-70) were poorly sampled, so samples of carp, channel catfish, and bullheads (Ictalurus spp.) were small. Flannelmouth suckers, bluehead suckers, and roundtail chubs were common throughout the study stretch (RM 166-69). Channel catfish and carp were found from RM 144.5 downstream to RM 69. Green sunfish were collected in a short stretch, RM 146-135. Crawfish were relatively abundant from RM 166 down to RM 129 and were present in apparently lower numbers down to RM 69. Rainbow trout were common from RM 167 to 148 with a few specimens down to RM 137.5. Brown trout were common throughout RM 167 to 135.6, with a single specimen caught down at R}l 112. All species caught appeared to be in good body condition. Ponderal indexes (KTL) (Carlander 1969) were calculated for each fish. Values for rainbow trout compared favorably to those of fish taken in the Dolores River immediately downstream of McPhee Dam [K = 1.15 (n = 37) study area vs. K = 1.12 (n = 47) catch and release area]. Comparisons of nongame species between rivers is limited by the lack of published data on suckers and roundtail chubs in the West. Reference collections of fish scales and selected bones were developed and a key to separate the Dolores River fishes using scales only was developed. Review of the literature on fish biomass estimation and our experience on the Dolores River in 1987 leads to the conclusion that reliable estimates of river otter prey base are unobtainable. We must resign ourselves to making general observations about the prey base which do not convey much information; e.g. there are at least 10 species of fish in the Dolores River which can attain a size of >10 cm and be considered potential river otter prey. Two river otters killed in beaver traps were collected during the year, a, female on the N. Fk. Gunnison and a male on the Fraser River (proximate'to Rocky MOuntain National Park release site). Based on studies of chese carcasses, an external radio transmitter package was developed. Utility of the new design will be tested on 2 river otters maintained in a large holding pen (2.5 x 5 m) for 30 days prior to release in the wild. The holding pens for 2 river otters and 12 transport cages were built. The transport cages measure 40 x 40 x 76 cm and are made of 5 x 5 cm framing covered with 0.6 cm thick pegboard (for ventilation) lined with 0.6 cm2 hardware cloth. The cages have waterproof bottoms of epoxy coating, a vertically sliding door, and a movable rear panel to force the animals out of the cage. The cages exceed the minimum requirements for live animal transport �231 as described in the lATA Live Animals Regulations issue date May, 1987). (Section 4, pages 29-30, A survey for river otter sign was conducted on the Gunnison River in January, 1988. Unusually cold weather had maintained a constant snow cover along the banks for over a month, extending to the confluence with the Colorado River. The survey began at the confluence of the Gunnison and N. Fk. Gunnison Rivers (RM 72.8) and ended at RM 36.5. Surveys were done from canoes and were ended when slush ice completely filled the river channel. Ambient temperature at sunrise varied from -26oC to -16oC. River otter sign (scat, tracks, fish remains) were seen regularly between RM 72.8 and RM 64.0. No river otter sign was observed between RM 64.0 and RM 36.5. Because of land use practices, the Gunnison River between RM 64.0 and 51.0 has little bank cover, and adequate cover for river otters appears to be lacking. LITERATURE CITED Carlander, K. D. 1969. Handbook of freshwater fishery biology. State Univ. Press, Ames. 752pp. Vol. I. Iowa Fitzpatrick, J. F., Jr. 1987. The Subgenera of the crawfish genus Orconectes (Decapoda: Cambaridae). Proc. BioI. Soc. Wash. 100(1):44-74. Klein, W. D. 1955. Rep. 1123. 4pp. Prepared by Crayfish Introductions. L ,PrJ4J.. Colo. Div. Wild1. Spec. Purpose ~T~h~o~m~a-s~D~.~I~.~B-e-c~k-------------Wildlife Researcher ��Colorado Division of Wildlife Wildlife Research Report July 1988 233 JOB PROGRESS REPORT State of Colorado Project No. W-l53-R-2 ------------------------- Work Plan No. Mammals Research 9A --~~----------------- Job No. 1 Elk Investigations Impact of Elk Winter Grazing on Livestock Production and Work Plan No. 3 ----------------------- Job No. 5 Period Covered: July 1, 1987 - June 30, 1988 Author: N. T. Hobbs, D. L. Baker Personnel: G. Bear, M. Miller, L. Carpenter, B. Gill, K. Navo, C. Woodward, B. Seely, H. Seely, L. Lovett ABSTRACT Elk grazing during winter influenced forage production and cattle performance on sagebrush-grassland range during spring. The magnitude of that influence depended on year for cattle responses, but was largely independent of year for vegetation responses. Elk grazing caused linear declines in daily gains and weights of heifers at the end of the spring grazing season during the first year of treatment. During year 2, we observed similar, proportional effects of treatment on rates of gain and weights of cows in late spring. During year 2, birth dates of calves whose mothers were in the high density elk treatment during year 1 averaged 6 days later than,calves whose mothers were in controls the previous year. During both years we observed linear declines in the standing crop of live and dead, perennial grass in relation to elk density. Utilization of dead perennial irass Lnceased in propor t Ion to treatment and approached 80% in the high density treatment during year 2. Utilization of live perennfa.l grass reached an asymptote at about 20% for the moderate and high density treatments during both years. Although we occasionally observed a weak stimulation of primary production of grasses and forbs by elk grazing, the overall influence on production was negligible. Large differences in precipitation between year land 2 caused marked reductions in forage production. The magnitude of these reductions did not depend on elk density. ��235 IMPACTS OF WINTER GRAZING BY ELK ON CATTLE PRODUCTION P. N. OBJECTIVES 1. To test the hypothesis that elk grazing during winter influences the productivity and botanical composition of herbage on sagebrush grassland ranges during spring. 2. To test the hypothesis that elk grazing during winter influences the body weights and rates of gain of cows and calves using sagebrush grassland ranges during spring. METHODS AND MATERIALS Study Area We conducted experiments on the Little Snake Wildlife Management Area in northwestern Colorado (township 9 north, range 95 west, sections 9, 10). The area is about 35 km (19 mi) north of Maybell, Colorado on County Road 19. Although this area does not typically contain high concentrations of elk during winter, it is representative of areas that do have those high densities. Topography of the area includes level ridge tops, rolling hills, and deep gullies, ranging in elevation from 1,800 to 2,000 m (5,900 to 6,600 ft). Aspects are southern and southwesterly with an average slope of 15 degrees. Soils are generally sandy and sandy loam. Climate of the area is dry and cold. The growing season averages only 81 days. Annual mean temperature is 6.06 C (42.9 F). Annual precipitation averages 27.5 cm (12.5 in). Vegetation is dominated by big sagebrush (Artemisia tridentata) with an understory predominated by needle and thread (Stipa comata), western wheatgrass (Agropyron smithii), Indian ricegrass (Oryzopis hymenoides), Junegrass (Koleria cristata), and cheatgrass (Bromus tectorum). Important forbs include wallflower (Erysimum asperum), peppergrass (Lepidium perfoliatum), silver lupine (Lupinus argenteus), and scarlet globe mallow (Sphaeralcea coccinea). Experimental Design We observed effects of elk grazing on forage and cattle response in a randomized complete block design with four levels of elk density (0 elk/km2, 8 elk/km2, 15 elk/km2, and 30 elk/km2) and three replications per level. There were three blocks, each consisting of four pastures. Each pasture within a block was stocked with one level of elk density such that each block contained all levels. During year 1, the twelve available pastures were blocked by pretreatmet biomass of perennial grasses with the four lowest grass biomass pastures forming one block, and the four highest grass biomass pastures forming a second block, and the remaining four pastures serving as the third block. The four levels of elk density were randomly assigned to pastures within each block during year 1. �236 Procedures We stocked pastures with elk in December and January 1987-88. Average date of release into pastures was January 3. All elk were removed from pastures during April 18-21. Overwinter mortality approached 30% in the high density pastures, but was nil in the medium and low density ones. Replacement animals were substituted for those that died to maintain a consistent stocking rate across replications. We introduced 7 cow-calf pairs and one dry heifer into each pasture on May 9. With the exception of the heifers and few replacements required by death losses, etc., cows during year 2 were the same animals we observed during year 1 and were assigned to the same pastures they were in previously. Cows and calves were weighed to the nearest 0.5 kg when they were introduced to pastures and were reweighed 6 weeks later when they were removed. Animals were allowed free access to feed and water before weighing. We also observed the birth dates of all calves and weighed them to the nearest 0.5 kg immediately after birth. We estimated canopy cover of herbs shrubs immediately after removing cattle from pastures. Cover was estimated from the summed length of interception by each plant along 30, l2-m transects randomly placed in each pasture during year 1. We estimated standing crop, productivity, and utilization of forbs, perennial grass, and annual grasses by harvesting samples from 40 pairs of 1/4 m2 plots in each pasture on each of three sample dates. Pastures were sampled immediately after the elk were removed (April 27-29), at the midpoint of the spring grazing season (May 30-June 1), and at its end (June 30-July 2). Samples were dried at 600C for 48 hrs, separated by hand into live and dead, and weighed to the nearest 0.01 g. RESULTS All results described in this report are preliminary and are subject to revision as analyses are checked for accuracy. Here, we report our findings to provide a timely summary of progress, but we are obliged to thoroughly revisit them later. This is particularly the case for data on forage production, which are based on pooled estimates from composites of material harvested from individual plots. We have not yet been able to check and summarize data for individual plots from year 2. This process will probably modify our results. . Effects of Elk on Cattle Performance Elk grazing during winter influenced the performance of cattle during spring, and the magnitude of that influence depended on year (Table 1). Weights of adult cows (defined as animals 2 yrs or older with calves) at the end of the grazing season during year 2 declined in direct proportion to elk density (Fig. 1 linear effect P=0.03, Table 1). This decline in weights resulted from the effects of treatment on cow rates of gain (Fig. 2, linear effect p=0.045). We observed no such effects during the first year (linear effect P > 0.28). �237 Average weights of cows at the end of the grazing season were lower during the second year than the first (Fig 1. P-O.OOl), and the difference between years was greatest in the high density pastures (Year x level P=O.OOl). The effect of year on average daily gain of cows only approached significance (P=0.09). The coefficient of variation (CV) for cow weights into pastures was 3.1%, for weights out of pastures was 3.4%, and for average daily gain was 27%. Because CV's on weights were small, the large variability in average daily gain resulted in large part from variability induced by treatment. Although we saw linear declines in weights out of pastures and rates of gain of heifers (defined as 1 year old cows without calves) in response to elk density during the first year (Fig 3., Fig. 4, Table 1, linear effect P < 0.05), we were unable to detect any effect of treatment during year 2 (p > 0.48). Heifers weighed more at the end of the grazing season during year 2 than year 1 (Fig. 3, P=O.06) as a result of more rapid rates of gain (Fig. 4, P=0.05). Using analysis of covariance to remove the effect of heifer weights at the beginning of the grazing season did not change effects of treatment or year. Calves during year 2 whose mothers were in the high density treatment during year 1 were born an average of 6 days later than calves whose mothers were in the control (P=O.07, Table 1), and birth dates were generally delayed in proportion to elk density (linear effect P=O.lO). During both years, birth weights of calves tended to decline relative to elk density. However, this tendency resulted at least in part from a spurious effect of the initial randomization. On average, larger calves were assigned to the control and low density pastures than to the medium and high ones (Fig. 5) These assignments resulted purely by chance. That is, there was no opportunity for treatment to affect birth weights of calves during year 1. The effect of these assignments persisted when calves were weighed into and out of pastures (Figs. 6, 7, 8). When we removed the effect of birth weight from later weights using analysis of covariance, treatment effects on weights out of pasture and average daily gain were not significant during either year (p > 0.23). MOreover, we saw no significant (P=O.13) linear trend in weights out the pastures or daily gain (P=O.23) during year 2 even when the effects of birth weights were included in the analysis. Effects of Elk on Vegetation Elk grazing reduced the standing crop of live (Fig. 9, P-0.13) and dead perennial grass at the beginning of the grazing season (Fig. 10, P=0.03) during both years (Table 2). The effect of treatment did not depend on y.ear for either response (p > '0.45). Effects of treatment on live and dead forbs and dead annual grass were also consistent between years (Figs. 11, 12, 1.3, treatment x year interaction P >0.67). Elk grazing at the highest density tended to reduce the standing crop of dead forbs available to cattle during year 2 (Fig. 11, control v~3l elk/km2 P=0.13). Otherwise, we observed no effects on live or dead forbs. We consistently observed a significant quadratic effect on the standing crop of dead annual grass (Fig. 13, quadratic P=O.04, year x quadratic P=0.63), but could detect no treatment effects on the standing crop of live annual grass (Fig. 14, P > 0.29). Standing crops of all forage classes except dead forbs were lower during year 2 than during the previous year. �238 Effects of elk grazing on standing crops of perennial live and dead perennial grasses resulted from differences in rates of utilization during winter. During both years we observed highly significant linear relationships between elk density and utilization of standing dead perennial grass (Fig. 15, linear effect P=O.OOOl, year x linear effect P=0.35). Although rates of utilization during year 2 appeared to be less intense those during year one, this appearance was influenced by negative utilization values in the control. Negative values are artifacts of field methods. If values are adjusted such that the control utilization rates = 0.0 for both years, utilization for year 2 exceeds that for year 1 by about 15 percentage points. During both years, utilization of live perennial grass increased between controls and the medium elk density (Fig. 16, linear effect P~ 0.07, year x linear effect P=0.98). We observed no difference in utilization of live perennial grass between the medium and high densities during either year (P>O.78). Values for primary production of all forage classes during year 2 exceeded values for year 1 (Figs. 17, 18, 19, Year effect P < 0.005). In general, we could not detect any effect of treatment on primary production of perennial grass, annual grass, or forbs before May 1 (Figs. 17,18, 19, Table 2). However, during year two, forb production before May 1 in the 15 elk /km2 treatment tended to exceed control values (P=0.08, Fig. 19, Table 2) and this led to a quadratic effect that approached significance (P=0.15). During May I-May 30 of year 1, forb production in the 8 elk/km2 treatment tended to exceed control values (Fig. 20, P=0.15, Table 2), and we also observed a weak stimulation production of perennial grass by moderate and light elk grazing (Fig. 21, quadratic effect P=0.09, Table 2). Otherwise, we could detect no effect of treatment on primary production during May 1-30 (Figs. 20, 21, 22, Table 2). DISCUSSION Differences in weather contributed to year effects. Winter weather during year 2 was far more severe than during the first year. Crusted snow covered most of the herbaceous forage during January to March and this coverage probably reduced the level of utilization of grasses and forbs by elk that we would have observed in the absence of snow accumulation. Although snowfall was heavier during year 2, precipitation during the growing season was substantially reduced. Rainfall during year 2 was half that of year 1, and this difference was reflected in rates of primary production. A confounding influence emerged in our measurements of cattle performance. Our cooperator decided to breed his cows to both Hereford and Angus bullsdu~ing year 2. This resulted in a mixture of Herefords and Angus/Hereford crosses within pastures. Fortunately, this mixture was balanced across treatments with the exception of the 15 elk/km2 level, which tended to have more crossbred calves. In the future we will do our best to assure that the experiment is balanced for breed. At the end of two years of study we have observed significant effects of elk density on cow weights, but not on weights of calves. This is not unexpected because the performance of calves is buffered from environmental influences by the milk they receive from their mothers. However, we did observe that increasing elk density delayed birth dates of calves. If this delay increases �239 during succeeding years, then we anticipate that elk density will eventually compel reductions in calf weights in and out of pastures as a consequence of differences in calf age. Prepared WCWt £8k Dan L. Baker Wildlife Researcher C N. Thompson Hobbs Wildlife Researcher C �240 Table 1. Contrasts testing for differences between the control and treatment levels for cattle responses to elk density. Response/contrast OF Contrast SS Mean Square F Value Pr F Cow Weight Into Pasture (kg) Year 1 control vs others control vs 8 elk/km2 control vs 15 e1k/km2 control vs 31 e1k/km2 1inear effects quadratic effects 1 1 1 1 1 1 53.413091 0.010280 26.601413 159.842136 201.075503 6.029824 53.413091 0.010280 26.601413 159:.842136 20L075503 6.029824 0.17 0.00 0.08 0.50 0.63 0.02 0.6970 0.9957 0.7828 0.5062 0.4581 0.8953 Year 2 control vs others control vs 8 elk/km2 control vs15elk/km2 control vs 31 elk/km2 linear effects quadratic effects 1 1 1 1 1 1 184.227639 97.854023 125.174601 148.030479 118.333249 60.476155 184.227639 97.854023 125.174601 148.030479 118.333249 60.476155 1.16 0.61 0.79 0.93 0.74 0.38 0.3236 0.4631 0.4096 0.3724 0.4219 0.5605 Cow Weight Out of Pasture (kg) Year 1 control vs others control vs 8 elk/km2 control vs 15 elk/km2 control vs 31 elk/km2 linear effects quadratic effects 1 1 1 1 1 1 7.113415 268.050598 95.255095 0.006296 50.729412 1.930841 7.113415 268.050598 95.255095 0.006296 50.729412 1.930841 0.02 0.92 0.33 0.00 0.17 0.01 0.8809 0.3742 0.5880 0.9964 0.6908 0.9377 Year 2 control vs others control vs 8 elk/km2 control vs 15 elk/km2 control vs 31 elk/km2 1inear effects quadratic effects 1 1 1 1 1 1 1799.89964 774.22459 828.12000 2239.01067 1968.06072 142.97798 1799.89964 774.22459 828.12000 2239.01067 1968.06072 142.97798 7.23 3.11 3.33 9.00 7.91 0.57 0.0361 0.1282 0.1179 0 •.0240 0.0307 0.4771 0.00065118 0.01259802 0.10833942 0.02383994 0.00748295 0.11349828 0.00065118 0.01259802 0.10833942 0.02383994 0.00748295 0.11349828 0.00 0.08 0.67 0.15 0.05 0.70 0.9514 0.7893 0.4438 0.7139 0.8366 0.4337 Cow Gain in Pasture (k9/d) Year 1 control vs others control vs 8 elk/km2 control vs 15 elk/km2 control vs 31 elk/km2 1inear effects quadratic effects 1 1 1 1 1 1 �241 Tab 1e 1. Cont. Response/contrast OF Contrast SS Mean Square F Value Pr 0.47190932 0.18230404 0.17538005 0.70046568 0.63561598 0.00990830 0.47190932 0.18230404 0.17538005 0.70046568 0.63561598 0.00990830 4.70 1.82 1.75 6.98 6.33 0.10 0.0733 0.2265 0.2345 0.0385 0.0455 0.7641 F Cow Gain in Pasture (k~/d) Year 2 control vs others control vs 8 elk/km2 control vs 15 elk/km2 control vs 31 elk/km2 linear effects quadratic effects 1 1 1 1 1 1 Heifer Weight Into Pasture (kg) Year 1 control vs others control vs 8 elk/km2 control vs 15 elk /km2 control vs 31 elk/km2 1inear effects quadratic effects 1 1 1 1 1 1 10.163503 224.906289 134.514711 92.691145 299.063614 0.009916 10.163503 224.906289 134.514711 92.691145 299.063614 0.009916 0.03 0.61 0.36 0.25 0.81 0.00 0.8750 0.4713 0.5732 0.6383 0.4103 0.9961 Year 2 control vs others control vs 8 elk/km2 control vs 15 elk/km2 control vs 31 e1k/km2 linear effects quadratic effects 1 1 1 1 1 1 233.417170 99.958351 766.495899 271.526096 628.596732 180.883030 233.417170 99.958351 766.495899 271.526096 628.596732 180.883030 0.28 0.12 0.92 0.33 0.76 0.22 0.6190 0.7430 0.3812 0.5925 0.4244 0.6606 Heifer Weight Out of Pasture (kg) Year 1 control vs others control vs 8 e1k/km2 control vs 15 e1k/km2 control vs 31 e1k/km2 1inear effects quadratic effects Year 2 control vs others control vs 8 e1k/km2 control vs 15 elk/km2 control vs 31 elk/km2 linear effects quadratic effects 1 1 1 1 1 1 477.173149 652.814414 552.890342 19.744859 2.468597 933.421845 477.173149 652.814414 552.890342 19.744859 2.468597 933.421845 0.58 0.79 0.67 0.02 0.00 1.13 0.4761 0.4083 0.4446 0.8822 0.9582 0.3288 1 1 1 14.283029 6.718737 0.137117 131.769513 177.470636 54.535202 14.283029 6.718737 0.137117 131.769513 177.470636 54.535202 0.01 0.01 0.00 0.11 0.15 0.05 0.9149 0.9416 0.9916 0.7466 0.7081 0.8349 1 1 1 �242 Table 1. Cont. Response/contrast Contrast SS Mean Square F Value 1 1 1 1 1 1 0.10939538 0.07322377 0.03029204 0.13017558 0.09882086 0.00798611 0.10939538 0.07322377 0.03029204 0.13017558 0.09882086 0.00798611 6.56 4.39 1.82 7.81 5.93 0.48 0.0505 0.0902 0.2355 0.0382 0.0590 0.5197 1 1 1 1 1 1 0.01803490 0.03109231 0.11718076 0.01416643 0.05661465 0.03524869 0.01803490 0.03109231 0.11718076 0.01416643 0.05661465 0.03524869 0.18 0.31 1.16 0.14 0.56 0.35 0.6897 0.6024 0.3300 0.7230 0.4871 0.5798 1 1 1 1 1 1 0.1111111 8.1666667 1.5000000 6.0000000 14.4857143 2.5021645 0.1111111 8.1666667 1.5000000 6.0000000 14.4857143 2.5021645 0.02 1.14 0.21 0.83 2.01 0.35 0.9051 0.3275 0.6639 0.3962 0.2056 0.5768 1 1 1 1 .1 1 53.5452255 23.9048564 40.5352419 44.4629630 38.2010582 17.9242424 53.5452255 23.9048564 40.5352419 44.4629630 38.2010582 17.9242424 5.37 2.40 4.07 4.46 3.83 1.80 0.0597 0.1725 0.0904 0.0792 0.0980 0.2285 1 1 1 1 1 1 9.3578780 4.0265907 1.8530518 17.0176295 15.0600917 0.0379540 9.3578780 4.0265907 1.8530518 17.0176295 15.0600917 0.0379540 2.30 0.99 0.45 4.18 3.70 0.01 0.1804 0.3585 0.5251 0.0869 0.1028 0.9262 OF Pr F Heifer Gain in Pasture (kg/d) Year 1 control vs others control vs 8 elk/km2 control vs 15 elk/km2 control vs 31 e1k/km2 1inear effects quadratic effects Year 2 control vs others control vs 8 elk/km2 control vs 15 elk/km2 control vs 31 elk/km2 1inear effects quadratic effects Cal f Birth Date Year 1 control vs others control vs 8 elk/km2 control vs 15 elk /km2 control vs 31 elk/km2 1inear effects quadratic effects Year 2 control vs others control vs 8 elk/km2 control vs 15 e1k/km2 control vs 31 elk/km2 linear effects quadratic effects Calf Birth Weight (kg) Year 1 control vs others control vs 8 elk/km2 control vs 15 e1k/km2 control vs 31 elk/km2 linear effects quadratic effects �243 Tab1e 1. Cont. Response/contrast OF Contrast SS Mean Square F Value 1 1 1 1 1 1 3.23735720 0.11659716 1.40784233 8.29033618 9.71352236 0.16260247 3.23735720 0.11659716 1.40784233 8.29033618 9.11352236 0.16260247 1.60 0.06 0.69 4.09 4.79 0.08 0.2534 0.8185 0.4368 0.0897 0.0713 0.7866 Pr F Calf Birth Weight Year 2 control vs others control vs 8 elk/km2 control vs 15 elk/km2 control vs 31 elk/km2 linear effects quadratic effects Calf Weight Into Pasture (kg) Year 1 control vs others control vs 8 elk/km2 control vs 15 elk/km2 control vs 31 elk/km2 linear effects quadratic effects 1 1 1 1 1 1 11.3499362 0.7163668 23.9430206 17.6860045 29.2917260 2.8187408 11.3499362 0.7163668 23.9430206 17.6860045 29.2917260 2.8187408 1.11 0.07 2.35 1.73 2.87 0.28 0.3320 0.7998 0.1763 0.2359 0.1410 0.6179 Year 2 control vs others control vs 8 e1k/km2 control vs 15 elk /km2 control vs 31 elk/km2 1inear effects quadratic effects 1 1 1 1 1 1 39.4916543 9.6612743 18.4466873 63.8394316 63.8986943 0.6081260 39.4916543 9.6612743 18.4466873 63.8394316 .63.8986943 0.6081260 3.83 0.94 1.79 6.20 6.20 0.06 0.0980 0.3703 0.2294 0.0472 0.0472 0.8162 1 1 1 1 1 1 63.285596 . 1.454672 54.690238 118.480129 147.873890 1.187521 63.285596 1.454672 54.690238 118.480129 147.873890 1.187521 2.13 0.05 1.84 3.98 4.97 0.04 0.1950 0.8323 0.2240 0.0930 0.0673 0.8483 1 1 1 1 100.916318 64.652102 21.261620 142.926366 110.844169 1.117207 100.916318 64.652102 21.261620 142.926366 110.844169 1.117207 2.82 1.81 0.59 4.00 3.10 0.03 0.1439 0.2272 0.4698 0.0925 0.1287 0.8655 Calf Weight Out of Pasture (kg) Year 1 control vs others control vs 8 elk/km2 control vs 15 elk /km~ control vs 31 elk/km~ 1inear effects quadratic effects Year 2 control vs others control vs 8 elk/km2 control vs15e1k/km2 control vs 31 elk/km2 linear effects quadratic effects 1 1 �244 Tab1e 1. Cont. Response/contrast OF Contrast SS Mean Square F Value Pr F 1 1 1 1 1 1 0.00518067 0.00110539 0.01122251 0.00137811 0.00165435 0.00820535 0.00518067 0.00110539 0.01122251 0.00137811 0.1)0165435 0.00820535 0.77 0.16 1.66 0.20 0.24 1.21 0.4150 0.7000 0.2450 0.6675 0.6384 0.3128 1 1 1 1 1 1 0.00403162 0.01097058 0.00211109 0.00002346 0.00072119 0.00754712 0.00403162 0.01097058 0.002111 09 0.00002346 0.00072119 0.00754712 0.40 1.08 0.21 0.00 0.07 0.74 0.5516 0.3384 0.6643 0.9632 0.7987 0.4215 1 1 1 1 1 1 0.00793583 0.00045334 0.00105839 0.02702202 0.02995751 0.00294858 0.00793583 0.00045334 0.00105839 0.02702202 0.02995751 0.00294858 1.11 0.06 0.15 3.78 4.19 0.41 0.3327 0.8096 0.7137 0.0999 0.0866 0.5445 1 1 1 1 1 1 0.01231732 0.01072953 0.00165587 0.01627587 0.01101600 0.00013509 0.01231732 0.01072953 0.00165587 0.01627587 0.01101600 0.00013509 2.01 1.75 0.27 2.65 0.2063 0.2342 0•.6220 0.1545 0.2288 0.8869 Calf Gain: Birth-May (kg/d) Year 1 control vs others control vs 8 elk/km2 control vs 15 elk/km2 control vs 31 elk/km2 linear effects quadratic effects Year 2 control vs others control vs 8 elk/km2 control vs 15 elk/km2 control vs 31 elk/km2 linear effects quadratic effects Calf Gain in Pastures (kg/d) Year 1 control vs others control vs 8 elk/km2 control vs 15 elk/km2 control vs 31 elk/km2 1inear effects quadratic effects Year 2 control vs others control vs 8 elk/km2 control vs 15 elk/km2 control vs 31 elk/km2 1inear effects quadratic effects 1.80 0.02 �245 Table 2. Contrasts testing for differences between control and treatment levels for vegetation responses. Response/Contrast OF Contrast SS Mean Square F Value Standing Cro~, Dead Annual Grass, May 1 (g/m2) Year 1 control vs others 1 0.47129494 0.47129494 control vs 8 elk/km2 1 0.30943434 0.30943434 control vs 15 elk/km2 0.58847048 0.58847048 1 control vs 31 elk/km2 0.12831499 0.12831499 1 linear effects 1 0.07703806 0.07703806 quadratic effects 1 0.56074783 0.56074783 3.54 2.32 4.42 0.96 0.58 4.21 0.1091 0.1784 0.0803 0.3644 0.4759 0.0861 2.98 2.45 4.78 0.23 0.04 5.93 0.1353 0.1688 0.0714 0.6516 0.8396 0.0508 0.72 0.21 0.70 0.61 0.60 0.28 0.4297 0.6645 0.4348 0.4644 0.4667 0.6181 Pr F Year 2 control vs others control vs 8 elk/km2 control v s 15 elk /km2 control vs 31 elk/km2 linear effects quadratic effects 1 1 1 1 1 1 0.16496547 0.13565296 0.26496534 0.01250473 0.00247503 0.32885805 0.16496547 0.13565296 0.26496534 0.01250473 0.00247503 0.32885805 Standing Cro~, Live Annual Grass, May 1 (g/m2) Year 1 control vs others 0.32469754 0.32469754 1 control vs 8 elK/km2 1 0.09412864 0.09412864 control vs 15 elk /km2 0.31714935 0.31714935 1 control vs 31 e1k/km2 0.27647616 0.27647616 1 0.27344231 1inear effects 1 0.27344231 quadratic effects 0.12508288 0.12508288 1 Year 2 control vs others control vs 8 elk/km2 control vs 15 elk/km2 control vs 31 elk/km2 1inear effects quadratic effects 0.00033478 0.00080655 0.00001969 0.00043498 0.00013739 0.00000890 0.00033478 0.00080655 0.00001969 0.00043498 0.00013739 0.00000890 0.56 1.35 0.03 0.73 0.23 0.01 0.4832 0.2902 0.8622 0.4271 0.6491 0.9070 Standing Cro~, Dead Forbs, May 1 (g/m2) Year 1 control vs others 0.00164312 1 0.01543621 control vs 8 e1k/km2 1 0.00014597 control vs 15 elk /km2 1 0.00286577 control vs 31 elk/km2 1 1 0.01109843 1inear effects quadratic effects 1 0.01069027 0.04787047 control vs others 1 0.00164312 0.01543621 0.00014597 0.00286577 0.01109843 0.01 069027 0.04787047 0.06 0.61 0.01 0.11 0.44 0.42 1.17 0.8096 0.4715 0.9426 0.7509 0.5383 0.5456 0.3295 1 1 1 1 1 1 �246 Table 2. Cont. Response/Contrast DF Contrast SS Mean Square F Value 0.03899230 0.00070888 0.12196144 0.08104926 0.01331676 0.95 0.02 2.97 1.97 0.32 0.3745 0.9006 0.1454 0.2190 0.5936 Pr F Standing CroE, Dead Forbs, May 1 <g/m2) Year 2 control vs 8 e1k/km2 control vs 15 elk /km2 control vs 31 e1k/km2 1 i near effects quadratic effects 1 1 1 1 1 0.03899230 0.00070888 0.12196144 0.08104926 0.01331676 Standing CroE, Dead Perennial Grass, Mall <9/m2) Year 1 control vs others control vs 8 elk/km2 control vs 15 elk/km2 control vs 31 elk/km2 1inear effects quadratic effects control vs others 1 1 1 1 1 1 1 66.630425 1.984317 47.229185 137.207396 166.839643 0.054346 39.0738233 66.630425 1.984317 47.229185 137.207396 166.839643 0.054346 39.0738233 3.01 0.09 2.13 6.19 7.53 0.00 1.90 0.1336 0.7748 0.1946 0.0473 0.0336 0.9621 0.2174 1 1 1 1 1 12.1342352 5.1821243 91.2345670 86.4497340 3.1659021 12.1342352 5.1821243 91.2345670 86.4497340 3.1659021 0.59 0.25 4.43 4.20 0.15 0.4717 0.6337 0.0799 0.0863 0.7084 Year 2 control vs 8 e1k/km2 control vs 15 elk/km2 control vs 31 e1k/km2 1inear effects quadratic effects Standing CroE, Live Perennial Grass, May 1 <g/m2) Year 1 control vs others control vs 8 elk/km2 . control ys15elk/km 2 control vs 31 e1k/km2 linear effects quadratic effects 1 1 1 1 1 1 11.6388369 4.5603544 4.4460805 16.9130156 15.2895613 0.3052176 11.6388369 4.5603544 4.4460805 16.9130156 15.2895613 0.3052176 3.66 1.43 1.40 5.32 4.81 0.10 0.1042 0.2761 0.2816 0.0605 0.0708 0.7671 1 1 1 1 1 1 1.45132601 0.61690690 0.35887224 2.45370589 2.17558971 0.00013527 1.45132601 0.61690690 0.35887224 2.45370589 2.17558971 0.00013527 0.53 0.23 0.13 0.90 0.80 0.00 0.4931 0.6511 0.7292 0.3794 0.4061 0.9946 Year2 control vs others control vs 8 e1k/km2 control vs 15 e1k/km2 control vs 31 e1k/km2 linear effects quadratic effects �247 Table 2. Cont. Response/Contrast OF Utilization by Elk, Dead Year 1 control vs others control vs 8 e1k/km2 control vs 15elk/km2 control vs 31 elk/km2 linear effects quadratic effects Year 2 control vs others control vs 8 e1k/km2 control vs 15e1k/km2 control vs 31 elk/km2 1inear effects quadratic effects Contrast SS Mean Square F Value Pr F Annual Grass (%) 1 1 1 1 1 1 17681.2165 19327.4114 7702.0705 9786.3438 4496.4282 7571.6199 17681.2165 19327.4114 7702.0705 9786.3438 4496.4282 7571.6199 3.25 3.55 1.42 1.80 0.83 1.39 0.1214 0.1084 0.2790 0.2283 0.3983 0.2827 1 1 1 1 1 1 111283.252 1881.937 65687.808 267755.739 326844.691 1714.683 111283.252 1881.937 65687.808 267755.739 326844.691 1714.683 0.80 0.01 0.47 1.92 2.34 0.01 0.4067 0.9114 0.5187 0.2157 0.1771 0.9154 Utilization by Elk, Live Annual Grass (%) Year 1 control vs others control vs 8 elk/km2 control vs 15e1k/km2 control vs 31 elk/km2 1inear effects quadratic effects 1 1 1 1 1 1 594.70031 1305.19823 123.58818 946.90596 271.10804 70.02368 594.70031 1305. 19823 123.58818 946.90596 271.10804 70.02368 0.33 0.72 0.07 0.53 0.15 0.04 0.5904 0.4334 0.8038 0.5009 0.7140 0.8514 Year 2 control vs others control vs 8 elk/km2 control vs 15elk/km2 control vs 31 elk/km2 linear effects quadratic effects 1 1 1 1 1 1 2998.81464 918.75391 1787.84014 3663.91965 3183.25371 374.56788 2998.81464 918.75391 1787.84014 3663.91965 3183.25371 374.56788 1.06 0.32 0.63 1.29 1.12 0.13 0.3511 0.5940 0.4634 0.3074 0.3380 0.7312 45.50647 1494.09141 8606.41961 2080.30270 376.27049 7312.36452 0.01 0.26 1.48 0.36 0.06 1.25 0.9339 0.6394 0.2912 0.5825 0.8120 0.3255 Utilization b~ Elk, Dead Year 1 control vs others control vs 8 elk/km2 control vs 15elk/km2 control vs 31 elk/km2 1inear effects quadratic effects Forbs (%) 1 1 1 1 1 1 45.50647 1494.09141 8606.41961 2080.30270 376.27049 7312.36452 �248 Table 2. Cont. Response/Contrast OF Contrast SS Mean Square F Value Pr F 1744.8801 4670l.4050 4706.9936 3898.5673 26949.6347 4047.1604 0.15 3.91 0.39 0.33 2.26 0.34 0.7217 0.1191 0.5641 0.5982 0.2074 0.5916 Utilization bl Elk, Dead Forbs (tt) Year 2 control vs others control vs 8 e1k/km2 control vs 15e1k/km2 control vs 31 e1k/km2 linear effects quadratic effects 1 1 1 1 1 1 1744.8801 46701.4050 4706.9936 3898.5673 26949.6347 4047.1604 Utilization by Elk, Live Forbs (tt) Year 1 control vs others control vs 8 elk/km2 control vs 15e1k/km2 control vs 31 e1k/km2 1inear effects quadratic effects 1 1 1 1 1 1 1604.21647 3414.11249 4.49406 2008.33744 723.57430 1.00352 1604.21647 3414.11249 4.49406 2008.33744 723.57430 1.00352 3.06 6.52 0.01 3.84 l.38 0.00 0.1404 0.0510 0.9298 0.1075 0.2926 0.9668 Year 2 control vs others control vs 8 elk/km2 control vs 15e1k/km2 control vs 31 e1k/km2 1inear effects quadratic effects 1 1 1 1 1 1 751.85970 100.03421 1397.63184 250.76455 344.60546 953.32716 751.85970 100.03421 1397.63184 250.76455 344.60546 953.32716 0.72 0.10 l.34 0.24 0.33 0.91 0.4348 0.7694 0.2995 0.6448 0.5905 0.3832 Utilization by Elk, Dead Perennial Grass (tt) Year 1 control vs others control vs 8 elk/km2 control vs 1Se1k/km2 control vs 31 elk/Km2 1inear effects quadratic effects 1 1 1 1 1 1 703.303989 169.003609 627.390022 724.270722 744.309184 180.283581 703.303989 169.003609 627.390022 724.270722 744.309184 180.283581 5.05 1.21 4.51 5.20 5.35 1.29 0.0657 0.3128 0.0780 0.0627 0.0601 0.2986 1 1 1 1 1 1 3250.78443 1072.52751 595.27664 6808.14667 6389.01475 96.88756 3250.78443 1072.52751 595.27664 6808.14667 6389.01475 96.88756 10.99 3.63 2.01 23.02 21.60 0.33 0.0161 0.1055 0.2058 0.0030 0.0035 0.5879 Year 2 control vs others control vs 8 e1k/km2 control vs 15e1k/km2 control vs 31 e1k/km2 1inear effects quadratic effects �249 Table 2. Cont. Response/Contrast DF Contrast SS Mean Square F Value Pr F Utilization by Elk, Live Perennial Grass (~) Year 1 control vs others control vs 8 e1k/km2 control vs 15e1k/km2 control vs 31 elk/km2 linear effects quadratic effects 1 1 1 1 1 1 703.303989 169.003609 627.390022 724.270722 744.309184 180.283581 703.303989 169.003609 627.390022 724.270722 744.309184 180.283581 5.05 1.21 4.51 5.20 5.35 1.29 0.0657 0.3128 0.0780 0.0627 0.0601 0.2986 Year 2 control vs others control vs 8 e1k/km2 control vs 15e1k/km2 control vs 31 e1k/km2 linear effects quadratic effects 1 1 1 1 1 1 675.906321 113.824158 706.497634 698.727628 772.063497 196.954482 675.906321 113.824158 706.497634 698.727628 772.063497 196.954482 1.79 0.30 1.87 1.85 2.04 0.52 0.2299 0.6032 0.2208 0.2231 0.2031 0.4978 Primary Production Before May 1, Annual Grass (~/m2) Year 1 control vs others control vs 8 e1k/km2 control vs 15 elk/km2 control vs 31 elk/km2 1inear effects quadratic effects 1 1 1 1 1 1 1.63474990 0.79878348 1.85937607 0.76477842 0.61345324 1.32772458 1.63474990 0.79878348 1.85937607 0.76477842 0.61345324 1.32772458 1.27 0.62 1.44 0.59 0.48 1.03 0.3029 0.4609 0.2748 0.4702 0.5159 0.3491 Year 2 control vs others control vs 8 elk/km2 control vs 15 elk/km2 control vs 31 elk/km2 1inear effects quadratic effects 1 1 1 1 1 1 0.00813918 0.00710854 0.00332782 0.00623897 0.00379828 0.00211770 0.00813918 0.00710854 0.00332782 0.00623897 0.00379828 0.00211770 1.62 1.42 0.66 1.24 0.76 0.42 0.2498 0.2788 0.4465 0.3074 0.4177 0.5399 0.82229708 0.57647929 0.06021183 1.48003922 1.13582494 0.01522518 0.46 0.33 0.03 0.83 0.64 0.01 0.5213 0.5893 0.8599 0.3962 0.4541 0.9292 Primarl Production Before May 1, Forbs (9/m2) Year 1 control vs others control vs 8 elk/km2 control vs 15 elk/km2 control vs 31 elk/km2 1inear effects quadratic effects 1 1 1 1 1 1 0.82229708 0.57647929 0.06021183 1.48003922 1.13582494 0.01522518 �250 Table 2. Cont. Response/Contrast OF Contrast SS Mean Square F Value Pr F Primar~ Production Before May 1, Forbs (9/m2) Year 2 control vs others 1 0.99457260 0.99457260 control vs 8 e1k/km2 0.00868384 0.00868384 1 control vs 15 e1k/km2 4.57474795 4.57474795 1 control vs 31 elk/km2 0.15773177 0.15773177 1 linear effects 0.51305536 1 0.51305536 quadratic effects 2."82640924 2.82640924 1 0.94 0.01 4.34 0.15 0.49 2.68 0.3687 0.9306 0.0823 0.7121 0.5114 0.1525 Primary Production Before May 1, Perennial Grass (9/m2) Year 1 control vs others 1.00640940 1.00640940 1 control vs 8 e1k/km2 1 1.47080961 1.47080961 control vs 15 e1k/km2 1.49935759 1.49935759 1 control vs 31 elk/km2 1 6.09615984 6.09615984 4.30457156 4.30457156 1inear effects 1 quadratic effects 1 3.88238390 3.88238390 0.13 0.19 0.19 0.79 0.56 0.50 0.7307 0.6781 0.6752 0.4089 0.4839 0.5053 Year2 control vs others control vs 8 e1k/km2 control vs 15 elk/km2 control vs 31 elk/km2 1inear effects quadratic effects 0.00201849 0.20890438 0.49781410 0.12855703 0.01889927 0.40857930 0.00 0.14 0.34 0.09 0.01 0.28 0.9716 0.7188 0.5813 0.7771 0.9133 0.6165 Primary Production Ma~ 1-30 (9/m2), Annual Grass Year 1 control vs others 1 0.22959601 0.22959601 control vs 8 elk/km2 1 0.18318803 0.18318803 control vs 15 e1k/km2 0.52763668 0.52763668 1 control vs 31 elk/km2 0.76618604 0.76618604 1 1.39683114 1.39683114 1 linear effects 1 0.00100032 0.00100032 quadratic effects 0.01 0.01 0.03 0.04 0.08 0.00 0.9148 0.9239 0.8713 0.8453 0.7925 0.9944 Year 2 control vs others control vs 8 e1k/km2 control vs 15 elk/km2 control vs 31 elk/km2 1inear effects quadratic effects 0.68 0.68 0.21 0.55 0.31 0.13 0.4412 0.4416 0.6664 0.4856 0.5971 0.7274 1 1 1 1 1 1 1 1 1 1 1 1 0.00201849 0.20890438 0.49781410 0.12855703 0.01889927 0.40857930 0.85931557 0.85766560 0.25951166 0.69744880 0.39334837 0.16869738 0.85931557 0.85766560 0.25951166 0.69744880 0.39334837 0.16869738 �251 Table 2. Cont. Response/Contrast OF Contrast SS Mean Square F Value Pr F Primary Production Mal 1-30, Forbs (9/m2) Year 1 control vs others control vs 8 e1k/km2 control vs 15 e1k/km2 control vs 31 elk/km2 1inear effects quadratic effects 1 1 1 1 1 1 6.7629642 24.0068970 0.2103269 1.0236826 0.3766138 4.3059728 6.7629642 24.0068970 0.2103269 1.0236826 0.3766138 4.3059728 0.91 3.24 0.03 0.14 0.05 0.58 0.3764 0.1220 0.8718 0.7229 0.8291 0.4748 1 1 1 1 1 1 1.45569471 0.01005750 3.15690296 1.16274811 1.78582235 1.10490316 1.45569471 0.01005750 3.15690296 1.16274811 1.78582235 1.10490316 0.14 0.00 0.30 0.11 0.17 0.11 0.7212 0.9762 0.6016 0.7495 0.6930 0.7556 Year 2 control vs others control vs 8 elk/km2 control vs 15 elk/km2 control vs 31 elk/km2 1inear effects quadratic effects Primary Production May 1-30, Perennial Grass (9/m2) Year 1 control vs others control vs 8 elk/km2 control vs 15 e1 k/km2 control vs 31 elk/km2 1inear effects quadratic effects 1 1 1 1 1 1 0.0410311 1.0721369 5.2584121 14.6285968 17.7366068 22.4860639 0.0410311 1.0721369 5.2584121 14.6285968 17.7366068 22.4860639 0.01 0.19 0.94 2.62 3.18 4.03 0.9345 0.6765 0.3692 0.1566 0.1249 0.0915 1 1 1 1 1 1 2.39128109 1.27640911 0.70339062 3.31009767 2.72990111 0.05901847 2.39128109 1.27640911 0.70339062 3.31009767 2.72990111 0.05901847 0.54 0.29 0.16 0.75 0.62 0.01 0.4887 0.6094 0.7030 0.4189 0.4607 0.9115 Year 2 control vs others control vs 8 elk/km2 control vs 15 elk/km2 control vs 31 elk/km2 1inear effects quadratic effects �252 Cow Wt. Out of Pastures 500 490 480 'S e 470 •.. s: .2l 460 ~ f450 440 430 420 0 10 20 30 40 Elk Density (animals/km2) Year-1 --- 2 Figure 1. Effects of elk grazing during winter on weight~ of adult cows at the end of the spring grazing season. Vertical bars c + Z standard errors of the ~ean. Cow Gain In Pastures 1.90 1.80 1.70 ~1.60 e1.50 .j 1.40 Cl 1.30 >'a 1.20 c CD 1.10 1.00 ~0.90 0.80 0.70 0.60 ff 0 10 20 30 40 Elk Density (animalslkm2) Year-1 --- 2 Figure Z. Effects of elk grazing during winter on average daily gain of adult cows during the spring grazing season. Vertical bars - !Z standard errors of the ~ean. �253 Heifer Wt. Out of Pastures 310 300 290 280 ~270 1:260 .~ 250 :: 240 !230 220 210 200 190 10 0 20 30 40 Elk Density (animals/km2) Year -1 --- 2 Figure 3. Effects of elk grazing during winter on weights of heifers at the end of the spring grazing season. Vertical bars ± Z standard errors of the ~ean. Vertical bars = ± standard errors of the ~ean. Heifer Gain in Pastures 2.30 2.20 2.10 '6'2.00 ~1.90 .:s.1.80 c ~ 1.70 ~ 1.60 'iiI 1.50 C 1.40 CD ~1.3O --------.__ ! 1.20 < 1.10 1 1.00 0.90 0.80 '--T-----r-----,-----.------,-o 10 30 40 Elk Density (animalslkm2) Year -1 Figure 4. Effeci5 of elk grazing gain of heifer5 during the spring f Z standard errors of the ~ean. -- 2 during winter on average grazing season. Vertical daily bars; �254 Calf Birth Wt. 38 sr - 36 01 - C 35 s: 01 .~ 34 :: t '_ 33 32 " ..••....... -~ " 31 30 0 10 20 30 40 Elk Density (anlmals/km2) Year -1 --- 2 Figure 5. Effects of elk grazing during winter on birth welghts of calves. During year 1, there was no opportunity for treatMent to influence calf weights. Vertlcal bars· • Z standard errors of the Mean. Calf Wt Into Pastures 59 57 -g_ 55 s: 01 51 :: 49 .~ t "" <, 53 -, ""- ------ 47 45 1 43 41 0 10 20 30 40 Elk Density (anlmals/km2) Year -1 -2 Figure 6. Effects .of elk grazing du'ring winter on weights of calves at the ~'" ~ '''':';of the spring grazing season. During year 1, there was no opportunity for treatMent to influence calf weights. Vertical bars· + Z standard errors of the Mean. �255 Calf Wt. Out of Pastures 100 90 ---_ ---_ -_ - _ ... 70 6O~r-------~------~------~------~o 10 20 30 40 Elk Density (animals/km2) Year -1 --- 2 Figure 7. Effects of elk grazing during winter on weights of calves at the end of the spring crazing season. Vertical bars Z standard errors of the ~ean. f Calf Gain in Pastures 0.99 0.94 ---- '00.89 ~0.84 .j 0.79 -_-_ -_ ___ (!) ~0.74 ~069 CD' i C) 0.64 <0.59 0.54 0.49 '----T-----.-----.-----,...;------r- o 10 20 30 40 Elk Density (anlmalslkm2) Year -1 --- 2 Figure 8. Effect5 of elk grazing during winter on average daily gain of calves during the spring grazing ~eason. Vertical bars = ~ Z standard errors of the ~ean. �256 ForageClass - Perennial Grasses Age- Standing Uve 12 11 ~10 ••... •.. 9 18 ci S7 ~6 c ~ 5 1 J: r-----+----+-------~-l 2~r_------_r--------,_------_.--------._ o 20 10 30 40 Elk Density (animaJSl1<m2) Year -1 --- 2 Figure 9. Effects of elk grazing during winter on th~ standing crop of live perennial grass at the beginning of the spring grazing season. Vertical bars = Z 5tandard errors of the ~ean. Forage Class - Perennial Grasses Age- Standing Dead 30 ci e o f 10 1 }----- t- ---+-----'--------1 O~r_------_r--------r-------_r--------r_ o 20 10 30 40 Elk Density (animalslkm2) Year -1 --2 FiQure 10. Effects of elk grazing during winter on the standing crop of dead perennial cra5S at the beginning of the 5pring grazing season. Vertical bars = Z standard error5 of the ~ean. �257 Forage Class - Forbs Age - Standing Dead 1.3 12 ~ 1.1 S1.0 iO~ N~~ ~O.6 '§ 0.5 - f-- __ --! <, <, <, •......•.•......••.••....•••.. ~ 0.4 1 0.3 (!J 02 0.1 0.0'--r----r-----r------.-------r-o 10 20 30 40 Elk Density (animal~) Year-1 --- 2 FiQure 1,. Effects of elk grazing during winter on the standing crop of dead forbs at the beginninQ of the spring grazing season. Vertical bars l standard errors of the Mean. E Forage Class - Forbs Age - Standing Live 8 O~r-------_r--------~------_,--------~o 10 20 30 40 8k Density (animalslkm2) Year-1 --- 2 Figure 12. Effect~ of elk grazing during winter on the standing crop of live forbs at the beginning of the spring grazing season. Vertical bars = 2 standard errors of the Mean. �258 Forage CIass- Annual Grasses Age- Standing Uve 3 •... o .,_---- __,_ o ---- 10 -s 30 20 40 Bk Density (animaiS/krn2) Year-1 --- 2 Figure 13. Effects of elk grazing during winter on the stall.'ng crop of live annual grass at the beginning of the spring grazing season. Vertical barz = Z ztandard errors of the Mean. Forage Class - Annual Grasses Age - Standing Dead 1.6 1.5 •••..1.4 t•... 1.3 -1.2 >- 1.1 ~ 1.0 ~0.9 o 0.8 2'0.7 'g 0.6 ~ 0.5 i 0.4 ~ 0.3 e 0.2 •• "- "- "- "- 0.1 0.0 0 10 20 30 40 Elk Density (animalsJkm2) Year-1 --2 Figure 14. Effects of elk grazing during winter on the standing crop of live ennual grass at the beginning of the 5pring grazing 5eason. Vertical bar5 c Z standard errors of the Mean. �259 Forage Class - Perennial Grasses Age - Standing Live 2~r--------r--------~------~--------'-o 10 20 30 40 Elk Density (animalsJkm2) Year -1 --- 2 Figure 15. Effects of elk density on utilization of standing dead grass. Vertical bars c Z standard errors of the ~ean. Forage Class - Perennial Grasses Age-Standing Live 50 40 30 l ~ W ~ e ~ .§ 20 10 0 ~ -10 -20 -30 -40 0 10 20 30 40 Elk Density (animals/km2) Year -1 Figure graS5. -2 16. Effects of elk density on utilization of standing Vertical bars = Z standard error5 of the ~ean. live �260 Forage Class - Perennial Grasses Age - Standing Dead 70 60 50 l .¥ W 40 / / // 30 / / ..t' 20 c:: 0 = 10 ~ 5 0 / r/ I -10 -20 / / / / ---t ' / / / / / / / -30 o 30 20 10 40 Elk Density (anirnaJs/km2) Year -1 --- 2 Figure 17. Effects or elk grazing during winter o~ pri~ary production of perennial grass before May 1. Vertical bars standard errors of the Mean. Forage Class - Annual Grasses = Age - Standing Uve 5 •... o 1I--------1- o ------ 10 J 20 30 40 Elk Density (anirnaJ&'km2) Year -1 -2 Figure 18. Effects of elk grazing during winter on priMary production of annual grass before May 1. Vertical bBrs Z standard errors of the Mean. c 2 �261 Forage Class - Forbs Age - Standing Uve 7 is .....•.. Is i4 I· (2 it 1~---------r--------~-------.---------ro 20 10 30 40 Bk Density (animalslkm2) Year --- 1 --- 2 Figure 19. Ef Fe c t s of e lk grazing production of forbs before May I. errors of the !'lean. Forage Class - Forbs during winter on pri!'lary Vertical bars Z standard 2 Age - Standing Uve 14 13 -. 12 -- ~ 11 Z; 10 ..-I 9 ::::I t! 7 t ~ l~ ~ a. "I: 8 8 5 4 3 2 1 0 -1 0 20 10 30 40 Elk Density (animals/km2) Year -1 --- 2 Figure 20. Effects of elk grazing during winter on pri!'lary production of forbs during May 1-30. Vertical bars = Z standard errors of the ~ean. �262 Forage CW-Perennial GRsses Age-Standing Live }-------t-+-----]- , 8 7~~------_r--------r_------~--------~ o 10 20 30 40 Elk Density (animalSl1an2) Year-1 --- 2 Flgure 21. Effects of elk grazing during winter on pri~ary production of perennial grass during May 1-30. Vertical bars standard error5 of the ~ean. Forage Class-Annual Grasses Age-Standing Uve 18 17 _16 ~ 15 :914 C; 13 I 12 .•.. 11 ~10 C 9 8 :l 7 ,g Ie jI:05 ~ 4 "C 3 a. 2 1 r-------t------t--------------1 O~r- o ~--------~-------'--------~ 10 20 30 40 Elk Density (animaJs/km2) Year-1 --- 2 Figure ZZ. Effects of elk graz'ing during winter on pri~ary production of annual grass during Hay 1-30. Vertical bar5 ~ 2 standard errors of the ~ean. 2 � https://cpw.cvlcollections.org/files/original/e4a71e5f401e12ce56e1447ce0b09d0b.pdf 54fcf7e5120b532221feb3ed7845e73c PDF Text Text 257 Wildlife Research Report July 1988 JOB PROGRESS REPORT State of Colorado Project No. Mammals Research Work Plan Hultispecies Investigations W-153-R-2 ------------------------No. LA ----------------------- Consulting Services for Mark-Recapture Analysis Job No. 5 Period Covered: July 1, 1987 - June 30, 1988 Author: G. C. White Personnel: L. H. Carpenter, R. B. Gill, R. M. Bartmann, G. D. Bear ABSTRACT Progress towards the objectives of this job include: 1. The last six years of survival data from the Piceance mule deer study have been summarized to provide survival estimates from radio-collared animals. 2. A manuscript has been completed on an evaluation of line transect methods for estimating mule deer density. The manuscript is currently under peer review. 3. A manuscript has been completed on an evaluation of capture resighting methods for estimating population size of elk. The manuscript has been submitted for publication to The Journal of Wildlife Management. 4. A variety of consultations were held with CDOW personnel on the application of microcomputer technology to wildlife analysis problems. 5. Simulations of mark-recapture estimates of mountain sheep populations have been developed. 6. An analysis of elk and deer age structure estimates from populations have been developed. 7. An alternative method to compute variances and confidence limits for age-and-sex ratios was investigated. ��259 CONSULTING SERVICES FOR MARK-RECAPTURE ANALYSES Gary C. White P. N. OBJECTIVE Model and simulate population estimates of deer, elk, mountain sheep, and mountain goats with mark-recapture data. SEGMENT OBJECTIVES 1. Evaluate line transect methodology for estimation of mule deer population size. 2. Evaluate capture-resighting methodology for estimation of elk population size. RESULTS MiD DISCUSSION Line transect methodology for estimating mule deer (Odocoileus hemionus hemionus) densities with helicopter surveys was evaluated in pinyon pine (Pinus edulis)-Utah juniper (Juniperus osteosperma) woodland in Piceance Basin, northwestern Colorado. Eight sighting distance models were applied to eight line transect trials flown in enclosures of 194-482 ha where density (21-76 deer/km2) was known. Based on a standardized analysis, the exponential polynomial estimator performed best. It had small bias (-10%, P = 0.261) and deviated least from nominal 95% confidence interval coverage with 6 of 8 (P = 0.057) intervals covering true density. The negative exponential estimator was unbiased (p = 0.821), and the exponential power series estimator was biased (p = 0.033). Both had only 5 of 8 (P = 0.0006) confidence intervals that covered true density. The Fourier serIes estimator underestimated the true population in all 8 cases (P = 0.004), although adding one term beyond the stopping rule proposed by Burnham et ale (1980) lowered, but did not eliminate, this bias (P = 0.024). Confidence interval coverage for the Fourier series (5 of 8, P ~ 0.006) did not change with one additional term. The half-normal estimator-had the greatest bias (-25%, P = 0.006) and only 3 of 8 (P < 0.001) confidence interval covering true density. The hazard function a~d Hermite polynomial estimators had large bias (both -20%, P < 0.02) with 5 of 8 (P = 0.006) and 4 of 8 (P = 0.001) confidence intervals, respectively, covering true density. Deer density estimates from line 2 transect surveys on Piceance Basin winter range, were higher than from a helicopter quadrat census. To obtain the same precision with line transects as with quadrat counts at a density of at least 25 deer/km2, less than half the helicopter time was needed. Density estimates from line transect estimators were biased as much as 25% low, probably because deer on the centerline were not always seen, but this bias is less than for quadrat surveys. A manuscript (included as Appendix A) has been written for submission to The Journal of Wildlife Management. Mark-resighting techniques were evaluated for estimating elk population size. Misc1assifying marked animals as unmarked caused overestimates of population size. Coefficients of variation (4.410.0%) gave precise but biased estimates. Yellow neck collars were found to �260 be more visible during aerial surveys than orange neck collars. Costs for the procedure were high due to the expense of trapping and mUltiple helicopter surveys, but mark-resighting appears to be more reliable than total counts or trend counts in estimating elk populations. A manuscript (included as Appendix B) has been written and submitted for publiction to The Journal of Wildlife Management. An in-depth analysis of age structure data from deer and elk harvests was concluded. The objective of this analysis was to test for differences between seasons (first, second and third) in estimates of age structure from dental cementum data. We hypothesized that data would be biased because of differential vulnerability of young vs. older animals during the first season with that bias getting progressively smaller through the second and third seasons. A memo (included as Appendix C) was prepared and distributed for review and comments. As a result of discussions with Carpenter, Freddy, Bartmann, and Gill, an analysis was undertaken of the Bowden et ale (1984) methodology for estimating variances and computing confidence ratios about sample values of age-and-sex ratios for deer and elk populations. A memo (included as Appendix D) was prepared and distributed for review and comments. Consultations were held with a variety of CDOW personnel on the application of microcomputer technology to specific wildlife analysis problems. Among those consulted were: L. H. Carpenter, R. M. Bartmann, R. B. Gill, G. D. Bear, D. J. Freddy, D. F. Reed, T. D. I. Beck, and N. T. Hobbs. �261 ,. APPENDIX May 20, 1988 Gary C. White Department of Fishery and Wildlife Colorado State University Fort Collins, CO 80523 303-491-6678 RH: Aerial AERIAL Deer Line Transects' LINE TRANSECT IN PINYON-JUNIPER GARY C. WHITE, RICHARD LEN H. CARPENTER, (Burnham et al. 1980, Buckland best. from nominal intervals Laboratory, for estimating Colorado. 317 Center, 317 West Colorado State analysis, (Juniperus St. Paul, MN 55108. osteosperma) Eight sighting (-10%, f = distance in woodland to 8 line transect polynomial 0.261) trials was known. estimator and deviated least with 6 of 8 (f = 0.057) The negative exponential estimator of Fisheries and Wildlife, coverage in models (21-76 deer/km2) the exponential interval Department mule deer (Odocoileus surveys was evaluated 1985) were applied true density. address: Center, 80525 juniper It had small bias covering Research Research of 194-482 ha where density 95% confidence Ipresent Minnesota, CO Ecology with helicopter Basin, northwestern performed State 80526 of Wildlife Resource Piceance Based on a standardized Colorado CO 80526 (Pinus edulis)-Utah flown in enclosures of Wildlife CO methodology densities Biology, 80523 Division Fort Collins, hemionus) and Wildlife Division Natural Line transect pine CO Fort Collins, Fort Collins, University, pinyon Colorado Colorado ROBERT A. GARROTT,l hemionus of Fishery Fort Collins, West Prospect, Abstract: OF MULE DEER DENSITIES WOODLAND M. BARTMANN, Prospect, Biology White et al. ESTIMATION Department University, A was University of �262 White (£ unbiased = (£ = 0.033). covered Both had only 5 of 8 (£ true density. stopping rule proposed eliminate, 0.004), although = 0.024). estimator and Hermite Confidence intervals polynomial estimators respectively, true density. covering surveys on Piceance quadrat census. as with quadrat the helicopter coverage (-25%, £ the true the for the = true density. The hazard (both -20%, £ < 0.001) confidence Deer density estimates Basin winter range were higher To obtain the same precision term. 0.006) and only = had large bias 0.006) and 4 of 8 (£ transects interval covering 0.02) with 5 of 8 (£ helicopter underestimated that adding 1 term beyond had the greatest bias 3 of 8 (f < 0.001) confidence transect intervals (5 of 8, f - 0.006) did not change with 1 additional The half-normal function 0.006) confidence was biased by Burnham et al. (1980) lowered, but did not (£ this bias series = power series estimator The Fourier series estimator in all 8 cases (£ population Fourier 0.821) and the exponential intervals, from 2 line than from a with line counts at a density of at least 25 deer/km2, time was needed. Density estimates <1/2 from line transect estimators were biased as much as 25% low probably because deer on the centerline were not always seen, but this bias is less than for quadrat surveys. ~ Key Words: Aerial census, estimator, exponential estimator, negative function hemionus. estimator, Colorado, line transect, power series estimator, exponential Hermite WILDL. MANAGE. estimator, polynomial Fourier exponential half-normal estimator, OO{O):OOO-OOO series polynomial estimator, Odocoileus hazard hemionus �263 APPENDIX B August 12, 1987 George Bear Colorado Division of Wildlife P.O. Box 39 Maybell, CO 81640 (303)272-3247 RH: Bear et al. -- Observability bias OBSERVABILITY BIAS IN MARK-RESIGHTING ESTIMATES OF ELK POPULATIONS George D. Bear, Colorado Division of Wildlife, 317 W. Prospect Street, Fort Collins, CO 80526 Gary C. White, Department of Fishery and Wildlife Biology, Colorado State University, Fort Collins, CO 80523 Len H. Carpenter, Colorado Division of Wildlife, 317 W. Prospect Street, Fort Collins, CO 80526 R. Bruce Gill, Colorado Division of Wildlife, 317 W. Prospect Street, Fort C9llins, CO 80526 David Essex, Rocky Mountain National Park, Estes Park, CO 80517 J. WILDL. MANAGE. 00(0):000-000 Abstract: Mark-resighting techniques were evaluated for estimating elk population size. Misclassifying marked animals as unmarked caused over estimates of population size. Coefficients of variation (4.4-10.0%) gave precise but biased estimates. Yellow neck collars were found to be more visible during aerial surveys than orange neck collars. Costs for the procedure were high due to the expense of trapping and multiple helicopter surveys, but mark-resighting appears to be more reliable than total counts or trend counts in estimating elk populations. Key words: Biotelemetry, mark-resighting, capture-recapture, captureresighting, Cervus elaphus, elk, hypergeometric, mark-recapture, population estimation. ��265 APPENDIX C M E M 0 RAN DUM June 19, 1988 To: Distribution From: Gary C. White Q._cJ. Re: Analysis of Elk and Deer Age Structure from Tooth Collections Introduction An ongoing debate has existed concerning the validity of age structure estimates made from harvested animals. The population age structure is estimated from ages determined by sectioning incisors collected from animals harvested by sport hunters. Several potential problems exist. First, ages determined from tooth sectioning may have questionable accuracy. Second, the sample of animals collected by sport hunters may not reflect the population age structure because of either (or both) selection by hunters or by differences in vulnerability of animals by age. This memo addresses the second question. For any reasonable population, the true age structure is not known. Thus, no means for testing the sample against the truth is available. However, if a representative (random) sample of the population is taken by sport hunters, there should not be differences between 2 consecutive samples from the same population. That is, the 2 age structure estimates from 2 different samples should be the same. The season structure of deer and elk seasons in Colorado allows for testing this hypothesis because males are harvested in 3 separate seasons, whereas females are harvested in 2 separate seasons. In addition, some later seasons are occasionally allowed, providing more data for testing the hypothesis. Methods A total of 500 deer and 793 elk teeth were sectioned and aged for 1986, and 163 deer and 949 elk for 1987. The breakdown of the sample by DAU, GMU, year, sex, and season are provided in pages 1-34 of the accompanying SAS listing. Because few older age animals occur in the sample, pooling of age classes was necessary for chi-square tests. The distribution of ages across all areas are shown on pages 35-38. For deer, pooling of ages 5 and greater was used. For elk, ages 5-8 were pooled, and 9 and greater. The sample was not collected uniformly across seasons within areas, so that most comparisons of seasons were not cleanly delineated to specific areas. Further, most potential comparisons lacked adequate sample size to provide reasonable statistical power. Thirty tests were performed across �266 Elk and Deer Age Structure from Tooth Collections -- June 19, 1988 Page 2 samples taken on a OAU basis, but most were based on too small of a sample. Further, some of the samples were from different GMU's within the same OAU between seasons, i.e., the samples from 2 seasons may not have been from the same population. Thus, interpretation of the results must be made with care. To determine if the samples for the seasons came from the same areas, refer to the output on pages 1-21, where the harvest by GMU is presented. Several statistics are output from SAS to determine if the null hypothesis of equal age distributions is rejected. The familiar Pearson chi-square is provided, plus the slightly more powerful likelihood ratio chi-square. I have used the likelihood ratio statistic for testing the null hypothesis. Neither of these statistics are valid for small samples, so the Fisher exact test has also been computed for small samples. This statistic is not totally valid for this situation because the marginals of the tables are random variables. However, for the many tables with small samples, this statistic is useful to help interpret the likelihood ratio test. Results Several of the tests for males were significant (£ < 0.05). Generally, this result is not particularly interesting because many more yearling males were harvested in the third season, when antler point restrictions were removed for some OAU's. The best example of this phenomena is for the 1986 E-43 data (page 78). Another test that rejected the null hypothesis is for 1986 0-7 (page 50). However, the distribution of the harvest across GMU's was not uniform within the OAU for each season, so this test probably represents differences due to sampling effort. The best test of the null hypothesis was provided by the data from DAU E-6, White River. Over 400 females were aged in 1987. In 1986, both males and females rejected the null hypothesis (pages 82-84), suggesting strong differences between the 2 seasons for females and 3 seasons for males. In 1987, a late hunt was conducted in E-6 (GMU 33), labeled as season 4 in the output. When the late hunt is included in the analysis, the null hypothesis is rejected for females (pages 85-86), but not males (page 87). However, if the late hunt is not included, the null hypothesis is no longer rejected (£ = 0.250) for females (pages 40-41). The rejection for females may be because of differences in age structure of GMU 33 from the rest of the unit, and not necessarily that the late season produced a different age structure. Data were pooled for the entire state to test the null hypothesis, with several of these tests significant (pages 88-105). However, because of differences in sampling areas and intensities between seasons, I doubt that these tests contribute much information concerning the question being addressed here. Conclusions �267 Elk and Deer Age Structure from Tooth Collections -- June 19, 1988 Page 3 The 1986 E-6 data suggest strongly that differences in age structure estimates between seasons occur. Not rejecting the null hypothesis does not contribute evidence that it is true, particularly for the small sample sizes (and hence low power) in the data available. Thus, the many tests that were not significant contribute little information relative to the bias. Rejection of the null hypothesis for data where areas and seasons are not confounded does contribute evidence that a biased sample of age structure is obtained from hunter harvest. Such is the case for 1986 E-6. Not rejecting the null hypothesis in 1987 for E-6, where the largest sample sizes occur, suggests that not every year has major differences between seasons. Suggestions for Future Work The tooth collection should be limited to a few GMU's, so that adequate sample sizes are obtained within a small area for all seasons. Much of the tooth sectioning effort is wasted when teeth from across the state are aged, but adequate samples are not available from multiple seasons in a single GMU. Confounding of seasons and areas severely limits the interpretation of the test results. The coming year presents some excellent opportunities for testing the hypothesis with deer, because the number of female permits is greatly increased. I suggest that several key GMU's be selected, and teeth solicited from only these units. In particular, I would suggest that GMU's 22 (0-7), 10 (0-6), and 33 (0-42) be selected to complement existing and planned research. Elk teeth should continue to be collected from E-6, particularly if several GMU's could be emphasized instead of pooling data from the entire OAU to achieve large sample sizes. I will incorporate dBase programs into the database I am currently developing to include the tooth age data in the data base. Then, standardized recording and reporting procedures will be available. Instead of recording the season as 1-4, the season codes used by Harlan Riffel should be incorporated in the tooth database. I will attempt to do this with the existing data, but probably will not be able to separate limited hunt males from over-the-counter hunt males. A final observation is worth presenting. Collection of teeth is only feasible when the management biologists are interested in the project, and understand the significance. I believe a journal paper could be developed from such data, particularly if they were collected to minimize all the problems discovered in the current data. This is the type of project that could be used to motivate several of the area/district biologists, and further improve the cooperation between the research section and management biologists. Oistribution: Len Carpenter ~ce Gill Jim Lipscomb �268 Comparison of Age Distribution Across Seasons Number of Animals by GMU Analysis DAU D-1 Variable GMU 2 : AGE YEAR SEX SEASON 1986 F M D-10 20 30 42 521 D-13 43 D-11 D-12 1986 1986 1987 1987 1986 F M M M F F M D-15 48 56 1987 M 1986 F M 1987 F M D-16 49 1986 F 57 1987 F 1986 F M 58 1987 F 2 M 1 2 3 3 3 3 3 1 3 3 3 1 2 1986 F M 1986 F M 1987 F M 46 1987 M D-19 51 62 1987 M 1986 M D-2 3 1986 1987 1986 1986 4 D-20 53 2 M M D-17 2 2 2 3 2 3 2 2 2 2 2 2 1 2 3 2 3 1 2 3 3 3 2 2 3 1987 F 581 11:05 SUnday, June 19, 1988 M M M M 2 1 3 2 3 3 3 N Obs 1 1 2 1 2 6 1 1 1 1 1 6 5 2 1 9 1 2 10 1 1 1 2 1 2 1 2 1 2 2 10 4 14 4 2 25 4 2 2 5 1 1 3 6 1 1 4 Minimum 0 3 1 1 2 2 5 5 4 4 5 1 2 2 4 3 2 4 2 Maximum 0 3 1 1 2 3 5 5 4 4 5 6 5 3 4 10 2 5 8 2 2 4 4 2 2 2 3 3 2 2 8 2 3 1 1 1 1 2 1 2 2 1 2 2 4 2 1 3 4 1 3 3 8 2 3 8 3 6 4 2 8 8 3 6 7 13 2 5 2 4 3 2 3 4 4 �269 Comparison of Age Distribution Across Seasons Ntnnber of Animals by GMU Analysis D.AU Variable: GMU AGE YEAR SEX SEASON Ir21 54 Ir22 55 1986 M 1987 M 1986 M Ir24 551 70 1986 M 1986 F 71 1986 F 66 1987 M 1986 M 67 1986 M M M Ir25 Ir26 68 Ir27 Ir29 Ir3 681 38 73 16 161 1987 1986 1987 1987 1987 1987 1986 1987 1986 M M M M M M M M M 17 1986 F 171 1986 M 83 1986 F M Ir31 11:05 SUnday, June 19, 1988 M 1987 M D-32 112 85 1986 M 1986 M 851 D-33 143 D-34 84 1987 1986 1987 1986 F M M F M 1 2 1 2 3 2 2 3 3 3 3 2 1 2 1 2 3 2 3 2 2 2 2 1 3 1 2 3 3 3 1 2 3 3 1 2 3 2 3 3 2 3 3 N Obs Min:ilnum Maximum 1 1 3 2 2 1 2 1 2 3 1 3 1 2 1 3 5 5 2 1 1 3 3 5 2 2 2 2 3 2 2 2 3 0 2 1 1 3 0 5 1 1 2 1 2 3 2 2 8 3 1 2 3 5 3 3 4 2 4 5 5 2 4 7 3 3 5 2 2 2 2 3 4 2 5 3 10 6 9 10 3 0 5 7 1 9 2 2 2 1 2 2 3 1 7 2 1 2 2 22 1 2 6 2 3 1 1 1 5 4 2 1 1 1 2 1 1 1 3 3 3 1 61 21 35 34 1 1 1 2 1 3 1 2 1 2 5 5 5 3 2 2 3 2 2 �270 Corrparison of Age Distribution Across Seasons Number of Animals by GMU Analysis I:llill D-34 Variable GMU 86 : AGE 1987 F M F M D-35 80 81 1986 F M 1987 M 1986 M 1987 F M D-36 76 79 1986 1986 M M D-37 D-39 82 63 1987 1987 1986 D-4 19 1986 M M F M F M 1987 191 1986 8 1987 1987 9 D-40 D-42 D-43 64 33 25 D-43 26 34 34 11:05 SUnday, June 19, 1988 YEAR SEX SEASONN Obs 1986 1986 1987 1987 1987 1987 1986 1986 1986 1987 F M F M M M M M M F F M M F M F M 5 2 2 2 3 1 2 3 2 2 3 2 3 2 3 2 1 3 2 2 3 2 1 2 2 2 3 1 2 3 3 3 3 3 1 1 2 1 3 2 2 2 2 2 2 2 2 2 5 2 3 3 1 3 4 1 9 7 5 2 5 5 2 1 3 2 1 12 4 1 3 11 2 1 2 3 2 2 1 1 43 2 3 2 2 1 1 1 1 1 1 1 1 1 2 Minimum Maximum 0 4 2 0 4 2 1 1 1 1 2 1 1 1 2 4 1 1 1 1 2 9 1 1 0 3 1 1 2 2 2 0 1 3 2 2 2 2 2 4 3 3 2 3 4 11 2 4 4 6 1 4 6 4 2 3 8 6 4 2 3 1 8 6 9 5 5 0 3 1 2 4 2 2 0 8 4 3 4 3 2 2 4 3 3 2 2 2 2 2 2 2 4 2 �271 Corrparison of Age Distribution Across Seasons Nlnnber of Animals by GMU Analysis I:l.AU 0-44 0-46 0-49 Variable GMU 881 114 105 : AGE M M F M 0-5 0-7 0-8 1986 F 22 23 1986 1986 1987 1986 1986 M M M M M M 24 15 1987 1986 1986 87 11 12 M M F M 0-8 35 35 36 1987 1986 1987 1986 F M F F M 1987 0-9 18 181 1987 1986 F M M F M 27 1986 F M 28 1987 37 1986 1987 F M M F M 371 1986 F M E- E-1 109 114 143 89 1 11:05 SUnday, June 19, 1988 YEAR SEX SEASON NObs 1986 1986 1987 1987 1986 1986 1987 1987 1986 1987 F M F M M F M F M 7 3 2 2 2 2 2 2 3 3 3 1 2 3 2 1 2 1 2 3 2 2 2 2 2 2 3 2 2 2 2 2 3 3 2 2 2 2 2 3 2 2 2 3 3 2 2 2 1 1 2 2 1 1 2 1 1 1 15 2 3 1 5 1 1 3 1 1 2 1 2 3 3 1 2 1 2 2 2 1 1 1 1 1 1 1 1 2 1 4 2 2 1 1 1 Minllnum 4 4 2 2 2 1 1 1 2 2 1 2 1 2 2 0 1 3 3 1 3 0 2 3 2 2 0 2 2 2 4 6 5 5 3 5 6 2 4 1 1 2 3 3 1 1 5 Max:irnum 4 4 3 6 2 1 1 1 2 2 3 3 2 2 3 0 1 3 3 1 7 0 2 7 3 2 1 2 2 4 6 6 5 5 3 5 6 2 4 1 1 4 3 5 1 1 5 �272 Corcparison of Age Distribution Across Ntnnber of Animals by GMU Analysis Variable: 11: 05 SUnday, June 19, 1988 AGE GMU YEAR SEX SEASON NObs 201 1987 E-I0 22 1987 E-12 30 31 32 36 1986 1987 1987 1986 1987 E-13 28 1986 1987 D.AU F M F M M F F F F M F F M E-13 E-14 E-15 E-16 E-17 37 1987 371 371 42 1986 1987 1987 421 1987 521 43 471 44 444 45 1987 1987 1987 1987 1986 1986 1987 481 1986 56 1986 F M M F F M F M F F F M M F F M F M F M E-18 E-18 50 50 500 1987 1986 1986 M F M 1987 F M M 1986 Seasons 2 2 2 3 3 3 3 2 2 2 2 2 2 3 2 3 3 3 2 3 2 2 2 1 3 3 3 3 2 1 2 2 3 2 2 2 2 3 1 3 2 3 1 3 2 2 3 1 1 1 1 1 1 1 1 1 1 2 4 1 4 1 6 1 3 1 1 1 1 2 1 1 1 1 1 1 1 1 11 1 3 1 1 1 2 3 5 1 2 2 3 1 1 1 Minllnum 2 2 1 16 0 5 3 6 5 3 3 2 1 0 1 0 0 0 1 2 9 1 3 2 1 3 4 0 1 3 0 1 4 0 4 1 3 4 1 1 2 1 1 1 2 4 1 Maximum 2 2 1 16 0 5 3 6 5 3 3 5 1 1 1 3 0 0 1 2 9 1 3 2 1 3 4 0 1 3 0 11 4 3 4 1 3 18 5 4 2 1 1 2 2 4 1 �273 Comparison Analysis ]),AU E-2 Variable GMU of Age Distribution Across Number of Animals by GMU : AGE SEX SEASON N Cbs 1987 F M F F F M 1986 1986 1986 441 5 1987 1986 61 1986 F F M F 62 1987 1986 M M E-22 49 1987 1986 M F E-22 49 1986 1987 M F E-20 M M 57 1986 F 1987 M F M 58 11: 05 SUnday YEAR 14 3 4 1986 F M 1987 F M E-23 E-23 581 581 1986 1986 E-24 70 1987 1986 F F M F F 1987 F Seasons 2 2 2 2 3 2 3 3 3 3 2 2 2 2 3 2 2 3 2 2 3 1 2 3 2 3 2 2 3 2 3 2 3 2 3 2 3 1 2 3 2 3 3 2 2 3 2 3 1 2 2 4 1 1 1 1 1 1 1 2 4 1 1 11 6 1 19 2 6 8 1 11 12 2 26 2 3 1 13 5 7 7 11 5 2 6 4 2 1 5 1 28 15 2 I Mininn.nn 2 2 0 5 1 2 2 1 3 6 2 3 2 2 2 4 0 0 5 0 1 1 1 3 0 1 1 0 3 0 1 0 0 1 1 0 3 2 0 1 0 8 1 1 0 0 6 June 19 I 1988 MaxiIm.nn 4 2 1 8 10 2 2 1 3 6 2 3 3 3 2 4 16 13 5 11 6 9 11 3 9 11 3 10 7 3 1 10 11 3 4 10 8 3 2 2 1 8 4 1 10 13 8 �274 Comparison of Age Distribution Across Seasons Number of .Anilnal.s by GMU Analysis Di\U E-25 Variable: YEAR SEX SEASON N Obs 71 1986 F 66 1987 1986 F F M F 67 1986 1987 68 1986 1987 M F M F F M F M M F M 681 681 1986 1987 E-27 86 1986 E-28 E-3 69 16 1987 1986 1986 F M M F F 161 1986 M F 17 1986 M F E-26 M 171 1987 1986 F F M 6 1986 F E-31 78 1986 E-32 80 1986 F M F M E-3 11: 05 SUnday, June 19, 1988 GMU 1987 E-26 AGE 2 3 2 2 2 2 3 2 3 3 2 3 2 2 2 2 3 2 2 3 3 1 3 3 2 3 3 2 3 3 2 3 1 2 3 3 2 3 2 3 2 3 2 2 3 2 3 13 24 1 3 2 12 5 2 4 2 14 6 3 1 5 1 1 2 1 1 1 1 1 3 11 7 1 7 7 3 3 3 1 1 4 1 9 7 1 3 1 6 4 5 5 3 1 Mi.niJnum Maximum 0 0 16 2 2 1 3 2 2 1 0 0 0 1 1 5 2 1 4 1 0 3 4 3 0 0 2 0 0 1 3 1 4 2 1 5 0 0 2 2 1 0 2 2 0 2 1 16 12 16 6 4 17 10 2 6 2 6 7 3 1 5 5 2 12 4 1 0 3 4 17 9 5 2 9 6 2 7 4 4 2 1 5 4 5 2 4 1 3 10 3 7 3 1 �275 Comparison of Age Distribution Across Seasons Number of Animals by GMU Analysis IWJ Variable GMU 11:05 SUnday, Jrme 19, 1988 : AGE YEAR SEX SEASON N Obs Minimum Maxi.mLnn --------1987 81 1986 F M F M 1987 F M E-33 83 1986 F M E-33 83 85 1987 1987 1986 851 1986 M M F M F M E-34 76 1986 F M 1987 79 1986 1987 F M F M F M F F M E-38 29 38 1987 1987 E-39 39 1986 F 1987 F M 46 1987 F 2 2 2 3 1 2 3 2 1 2 3 3 1 2 3 1 3 2 2 2 3 1 2 3 2 3 2 3 2 3 3 3 3 2 2 3 2 1 2 2 3 2 3 1 2 3 2 4 9 2 5 6 4 8 13 3 4 1 35 30 27 4 1 1 2 2 3 6 9 12 7 6 7 9 3 1 2 1 8 12 8 3 1 3 1 1 4 8 14 1 2 6 3 13 2 0 1 1 2 1 1 0 2 1 2 0 2 2 2 3 0 6 3 3 2 3 3 1 2 1 1 1 6 0 1 0 1 0 2 6 3 3 2 0 0 0 5 3 0 2 0 5 4 2 9 3 2 3 15 3 2 2 15 7 7 4 3 0 7 3 12 10 8 7 7 14 17 6 3 6 3 1 18 2 9 4 6 6 3 2 9 9 14 5 5 3 12 6 �276 Corrparison Analysis Dt\U E-4 Variable GMU 19 YEAR 1986 1987 of Age Distribution Across Number of Anilnals by GMU : AGE M F M F M 8 1986 1987 M F E-41 9 54 1987 1986 E-41 54 1986 1987 F F M M F E-42 E-43 53 55 1986 1986 M M F M 1987 F M 551 1986 F M 1987 F M E-6 12 1986 F E-6 12 1986 1987 M F M 13 11:05 SUnday, June 19, 1988 SEX SEASON N Obs 1986 F 1987 M F 3 2 2 2 2 3 1 3 2 2 3 2 2 1 3 2 3 2 3 2 3 1 2 3 2 3 2 3 2 1 3 2 3 2 3 2 3 2 2 3 1 3 2 3 3 2 3 Seasons 5 5 6 1 11 5 1 2 1 13 6 1 1 1 5 15 18 1 1 2 17 4 7 13 16 10 4 1 1 5 2 20 4 1 5 12 10 1 62 18 1 1 11 23 1 40 20 Mi.ni1num Maximum --------- 2 2 0 3 0 2 2 0 2 0 0 5 3 3 1 0 0 0 4 0 0 2 2 1 0 0 0 1 1 1 1 0 1 0 0 0 1 1 0 0 0 0 0 0 2 0 0 8 5 7 3 11 5 2 1 2 13 7 5 3 3 2 14 7 0 4 5 10 5 4 3 14 10 3 1 1 5 1 16 5 0 1 5 13 1 20 17 0 0 7 13 2 16 15 �277 Comparison of Age Distribution Across seasons Number of Animals by GMU Analysis Di\U Variable: GMU YEAR SEX SEASONN Obs M 131 E-6 1986 F 1987 M F M F M 23 1986 23 1987 1987 F F M 24 1986 F M 1987 F 1986 M F 1987 F 25 M 26 1986 1987 E-6 33 1986 33 1987 F M F M F M F M 34 E-7 15 1986 11:05 SUnday, June 19, 1988 AGE 1987 F M F 1986 M M 2 3 2 3 2 2 3 2 2 1 2 3 2 3 2 3 2 1 2 3 2 3 3 2 3 2 3 2 3 2 2 2 3 3 2 3 2 3 4 2 4 2 2 2 3 2 1 1 1 9 13 2 41 10 2 3 7 8 3 63 28 1 2 1 18 8 1 34 5 1 2 5 16 11 1 2 4 2 1 4 1 3 7 33 9 54 2 1 4 1 13 1 3 1 Minimum 1 0 0 0 1 0 1 0 0 2 2 2 0 0 0 0 1 2 1 2 0 1 2 0 1 0 1 0 0 0 2 0 0 0 0 1 0 1 0 0 0 0 3 2 2 0 4 MaxiJ.nurn 1 0 9 9 2 14 8 0 1 2 2 3 14 16 0 0 1 3 2 2 16 9 2 3 3 19 15 0 2 6 4 0 3 0 6 12 10 17 17 0 0 5 3 7 2 3 4 �278 Conparison Analysis Illill E-8 E-9 0-1 Variable GMU YEAR 27 18 1987 1986 1987 181 1986 20 1986 1986 0-10 1986 0-11 0-12 1986 1987 0-13 1986 0-15 1986 1987 of Age Distribution Across Seasons Number of AniInals by GMU : AGE SEX M M F M F M F F M F M M F M F M F M F M 0-16 1986 M 0-16 1986 1987 M F M 0-17 1987 M 0-19 1986 M 0-2 1986 M 1987 1986 1986 1987 1986 M M M M M 0-20 0-21 0-22 11: 05 SUnday, June 19, 1988 SEASON 3 1 3 3 2 2 3 2 2 2 3 2 3 2 2 2 2 2 1 2 3 2 3 1 2 3 2 3 1 2 3 2 3 1 2 3 1 2 1 3 2 3 3 3 1 2 1 N Obs Minimum 1 1 1 2 6 1 2 1 1 2 1 2 6 1 1 1 1 6 5 2 1 9 1 2 11 1 1 7 2 2 37 4 8 1 2 16 2 6 1 3 6 1 1 4 1 2 2 1 1 2 1 0 4 3 0 3 1 1 2 2 5 5 4 4 1 2 2 4 3 2 4 2 2 3 1 2 3 1 2 1 8 2 1 1 2 4 2 1 4 3 1 1 1 3 Maximum ------1 1 2 6 7 4 3 0 3 1 1 2 3 5 5 4 4 6 5 3 4 10 2 5 8 2 3 8 3 8 6 3 7 8 3 13 2 5 4 3 2 4 3 4 1 2 3 �279 Comparison of Age Distribution Across Seasons Number of Animals by GMU Analysis DMJ Variable GMU YEAR SEX SEASON N Cbs 0-24 1986 F 0-25 1987 1986 M M M 0-27 0-29 0-3 1987 1986 1987 1987 1987 1986 M M M M M F M 1987 1986 M 0-31 0-26 F M 1987 M 0-32 1986 M 0-33 0-34 1987 1987 1986 F M F M 1987 F M 0-35 1986 F M 1987 F M 0-36 11:05 SUnday, June : AGE 1986 M 1987 M 2 3 2 3 3 2 1 2 3 2 3 2 2 2 3 1 2 3 3 3 1 2 3 2 3 2 3 3 2 2 2 3 2 3 1 2 3 2 2 3 2 1 2 3 2 3 2 10 1 2 3 24 1 4 9 1 1 1 9 2 1 1 6 4 5 1 61 21 35 34 1 1 3 2 3 2 6 2 2 3 3 1 3 4 1 14 12 2 1 5 5 3 12 4 Minimum 1 2 2 1 1 3 1 2 5 5 2 1 3 3 3 2 2 2 2 0 2 1 1 3 0 1 1 2 2 0 4 2 2 0 4 2 1 1 1 1 2 4 2 1 1 1 2 19, 1988 Maximum 8 2 3 3 5 3 3 4 5 5 2 7 3 3 3 5 5 4 2 10 6 9 10 3 0 7 5 9 5 5 4 3 11 2 4 4 6 1 4 8 6 4 4 2 3 8 6 �280 Comparison of Age Distribution Across Seasons Number of Animals by GMU Analysis [l?ill Variable GMU : AGE YEAR SEX SEASON N Obs D-37 D-39 1987 1986 D-4 1986 M F M F D-4 1986 M 1987 F M 1987 1987 1986 M F F M F M M M F M F M M D-40 D-42 D-43 1987 D-44 D-46 D-49 1986 1986 1987 D-5 1986 D-7 1986 D-7 D-8 1987 1987 1986 M M F M 1987 F D-9 11:05 SUrrlay, June 19, 1988 1986 1987 M F M F M 1 2 2 2 3 1 2 3 3 1 2 3 2 2 2 2 2 2 3 2 2 2 2 2 1 2 3 2 3 2 1 2 2 3 2 2 2 2 3 2 3 E- 1986 1987 E-l 1986 F M M F M 2 2 2 3 3 2 1 3 11 2 2 4 3 45 2 5 2 2 1 1 1 2 2 3 1 1 2 2 1 1 20 4 5 1 1 2 1 6 5 1 3 5 5 3 2 1 2 1 2 4 2 2 1 Mi.niJm.nn 9 1 1 0 0 1 1 1 2 2 2 2 2 4 2 2 2 2 4 4 2 2 2 1 1 1 1 2 2 0 1 1 3 3 2 0 2 3 4 5 2 1 1 2 3 3 1 Maximum 9 5 5 0 3 3 2 8 2 4 4 2 2 4 2 2 3 4 4 4 3 6 2 1 3 3 2 2 2 0 1 3 7 3 3 6 5 6 6 5 5 1 1 4 3 5 1 �281 Comparison of Age Distribution Across Number of Anilnals by GMU Analysis Di\U E-10 E-10 Variable GMU YEAR SEX SEASON NObs F M 1986 1987 1987 M F F M E-12 1986 1987 E-13 1986 F F M F M 1987 F M E-14 1987 F M E-15 E-16 1987 1986 F F M 1987 F M E-17 E-17 1986 1986 F F M E-18 1987 1986 M F M 1987 F M E-2 1986 F M E-20 1987 1986 F F M E-22 11:05 SUnday, June : AGE 1987 1987 1986 M F 2 2 3 2 3 3 2 2 2 2 2 2 3 2 3 2 3 1 2 3 3 2 1 2 3 2 2 3 1 2 3 2 3 1 3 2 2 2 3 2 3 3 2 2 3 2 2 Seasons 2 2 1 2 2 1 1 1 2 4 1 1 6 1 9 3 1 1 1 1 2 1 1 11 1 4 2 2 3 1 5 1 2 2 4 4 2 4 5 1 2 1 1 5 1 3 35 Minimum 1 2 5 1 3 0 5 3 3 2 1 1 0 1 0 3 3 2 1 1 0 0 3 1 4 0 3 4 1 1 1 2 1 1 1 2 2 0 1 2 2 1 2 2 2 2 0 19, 1988 Maxinrum 2 5 5 6 16 0 5 3 3 5 1 1 2 1 3 9 3 2 1 1 4 0 3 11 4 3 4 18 5 1 4 2 1 1 2 4 4 8 10 2 6 1 2 3 2 4 16 �282 Con'prrison of Age Distribution Across Seasons Nmnber of Anilnals by GMU Analysis IlMJ Variable GMU : AGE YEAR SEX SEASONN Cbs M E-22 11:05 SUnday, Jl.ll1e 19, 1988 1987 F M E-23 1986 F E-24 1987 1986 M F F E-25 1987 1986 F F M 1987 F E-26 1986 M F M E-26 1987 1987 F M E-27 1986 E-28 E-3 1987 1986 1986 F M M F F M E-31 1987 1986 E-32 1986 F F M F M 1987 F M 3 2 3 2 3 1 2 3 2 3 3 2 2 3 2 2 3 2 3 2 3 2 2 2 3 2 2 3 3 1 3 3 2 3 1 2 3 3 2 2 2 3 1 2 3 2 1 23 10 7 56 9 8 17 6 2 1 5 1 41 39 3 3 4 2 2 26 11 2 3 1 1 7 2 1 1 1 1 3 31 30 1 2 11 1 4 5 2 10 6 7 9 17 3 Minimum 0 1 1 0 1 1 0 1 0 8 1 1 0 0 6 2 2 2 1 0 0 2 0 1 2 1 4 1 0 3 4 3 0 0 4 2 1 5 2 2 1 0 2 1 1 0 2 Maximt.nn 13 5 4 11 8 9 11 3 1 8 4 1 16 13 16 6 6 4 2 17 10 2 3 1 2 12 5 1 0 3 4 17 9 6 4 2 4 5 10 3 2 9 3 3 3 15 3 �283 Comparison of Age Distribution Across Nlnnber of AniInals by GMU Analysis J:li\U E-33 Variable GMU : AGE F M E-34 1987 M 1986 F M 1987 F M E-38 1987 F M E-39 1986 F 1987 F E-39 1987 M E-4 1986 F M F 1987 M E-41 E-42 E-43 11:05 SUnday, June YEAR SEX SEASON N Cbs 1986 1986 F M 1987 F 1986 1986 M M F M Seasons 2 3 2 3 1 2 3 1 3 2 3 2 3 2 3 2 3 2 3 1 2 2 3 2 3 1 2 3 2 2 2 3 1 3 2 1 2 3 2 3 2 3 2 3 1 2 3 Minimum 13 1 5 41 39 41 11 1 1 6 15 9 15 9 2 3 1 3 1 1 1 4 8 27 6 2 11 3 6 2 25 11 1 2 1 1 5 5 15 18 1 1 3 17 9 7 15 0 2 3 0 2 2 1 3 0 2 0 1 1 0 0 2 1 3 6 3 2 0 0 0 2 3 0 2 0 2 0 0 2 0 3 3 1 1 0 0 0 4 0 0 1 2 1 19, 1988 Maximum 4 2 12 15 8 7 7 3 0 14 18 6 3 9 3 4 1 6 6 3 2 9 9 14 8 5 5 12 7 3 13 7 2 1 3 3 3 2 14 7 0 4 5 10 5 4 3 �284 Corrparison Analysis I::lt\U E-43 Variable GMU of Age Distribution Across Number of .Animals by GMU : AGE F F M E-6 1986 F M 1987 F M E-7 E-8 1986 1987 1986 M M F M 1987 F M E-9 11:05 SUnday, June YEAR SEX SEASON N Obs 1987 1987 1986 F Seasons 2 3 2 3 2 3 1 2 3 2 3 4 1 2 3 4 1 3 2 2 3 3 3 36 14 5 6 49 51 25 22 12 303 106 54 1 10 8 1 2 1 6 1 1 2 2 Minimum 0 0 0 0 0 0 2 1 1 0 0 0 0 0 0 0 1 1 0 4 2 1 3 19, 1988 Maxilnurn 16 10 3 1 9 13 3 4 12 20 17 17 0 3 2 0 4 1 7 4 2 6 3 �285 Comparison of Age Distribution Across Seasons Distribution of AniInals by Age 11:05 SUnday, June 19, 1988 SPECIFS=Deer YFAR=1987 AGE 0 1 2 3 4 5 6 7 8 9 10 ------- Frequency Percent 21 89 149 90 60 47 26 5 9 2 2 4.2 17.8 29.8 18.0 12.0 9.4 5.2 1.0 1.8 0.4 0.4 CUmulative Frequency 21 110 259 349 409 456 482 487 496 498 500 CUmulative Percent 4.2 22.0 51.8 69.8 81.8 91.2 96.4 97.4 99.2 99.6 100.0 SPECIFS=Deer YFAR=1987 AGE 0 1 2 3 4 5 6 7 8 9 10 11 13 Frequency Percent 2 11 54 41 19 14 9 4 4 2 1 1 1 1.2 6.7 33.1 25.2 11.7 8.6 5.5 2.5 2.5 1.2 0.6 0.6 0.6 CUmulative Frequency 2 13 67 108 127 141 150 154 158 160 161 162 163 CUmulative Percent 1.2 8.0 41.1 66.3 77.9 86.5 92.0 94.5 96.9 98.2 98.8 99.4 100.0 �286 comparison of Age Distribution Across Seasons Distribution of Animals by Age 11:05 SUnday, June 19, 1988 -----AGE 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 -----AGE 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 SPECIFS=Elk YFAR=1986 CUmulative Frequency Percent Frequency 67 121 197 138 82 58 34 27 16 16 9 5 5 8 2 2 2 2 2 8.4 15.3 24.8 17.4 10.3 7.3 4.3 3.4 2.0 2.0 1.1 0.6 0.6 1.0 0.3 0.3 0.3 0.3 0.3 67 188 385 523 605 663 697 724 740 756 765 770 775 783 785 787 789 791 793 SPECIFS=Elk YFAR=1987 CUmulative Frequency Frequency Percent 121 139 193 148 88 62 56 35 27 21 13 9 6 5 7 4 8 4 1 1 1 12.8 14.6 20.3 15.6 9.3 6.5 5.9 3.7 2.8 2.2 1.4 0.9 0.6 0.5 0.7 0.4 0.8 0.4 0.1 0.1 0.1 121 260 453 601 689 751 807 842 869 890 903 912 918 923 930 934 942 946 947 948 949 CUmulative Percent 8.4 23.7 48.5 66.0 76.3 83.6 87.9 91.3 93.3 95.3 96.5 97.1 97.7 98.7 99.0 99.2 99.5 99.7 100.0 CUmulative Percent 12.8 27.4 47.7 63.3 72.6 79.1 85.0 88.7 91.6 93.8 95.2 96.1 96.7 97.3 98.0 98.4 99.3 99.7 99.8 99.9 100.0 �287 Corrparison of Age Distribution Across Seasons comparison Across 3 Seasons for Elk in D.MJ E-6, 1987, Males 11:05 SUnday, June 19, 1988 TABlEOF SEASON BYAGE SEASON AGE Frequency Cell Chi-Square ReM FCt cal 1Yea Total 13 Y 12 y 1 1 0.0561 100.00 0 0.0526 0.00 0 'O~'1053 0.00 0 0.0526 0.00 1 2 8 0.0014 80.00 1 0.4263 10.00 0 1.0526 0.00 1 0.4263 10.00 10 3 6 0.0158 75.00 0 0.4211 0.00 2 1.5921 25.00 0 0.4211 0.00 8 15 1 2 1 19 --Total STATISTICSFORTABIEOF SEASON BYAGE statistic Chi-Square Likelihood Ratio Chi-Square Mantel-Haenszel Chi-Square Phi Coefficient COntingency Coefficient cramer's V = DF Value Prob 6 6 1 4.623 6.097 0.212 0.493 0.442 0.349 0.593 0.412 0.645 Sample Size 19 WARNING:83% of the cells have expected counts less than 5. Chi-Square may not be a valid test. �288 Comparison of Age Distribution Across Seasons Cornparison Across 3 Seasons for Elk in ]),t\IJ E-6, 1987, Females 11:05 SUnday, June 19, 1988 TABlE SEASON Frequency Cell <lri -Square Row Pet OF SEASONBY AGE AGE II IYea cal Total 13 Y 12 y -+ 2 33 0.227 10.89 50 0.0027 16 •.50 55 0.0732 18.15 46 0.0145 15.18 303 3 8 0.6489 7.55 17 0.0076 16.04 22 0.2094 20.75 15 0.0414 14.15 106 41 67 77 61 409 Total (Continued) TABLE OF SEASONBY AGE SEASON Frequency Cell Chi -Square Row Pet AGE I I4 Y 15-8 Total 1>8 2 30 0.6391 9.90 65 0.0006 21.45 24 1.0709 7.92 303 3 5 1.827 4.72 23 0.0016 21.70 16 3.0611 15.09 106 35 88 40 409 Total STATISTICSFOR TABlE statistic Chi -Square Likelihcx:x:l Ratio <lri -Square Mantel-Haenszel <lri -Square Phi Coefficient Contingency Coefficient cramer's V Sample Size = 409 OF SEASONBY AGE DF Value Frob 6 6 1 7.825 7.837 2.787 0.138 0.137 0.138 0.251 0.250 0.095 �289 Comparison of Age Distribution Across Seasons Comparison for I.li\U=D-15 & year=1987 & sex=F 11:05 SUnday, June 19, 1988 TABlE OF SEASON BY AGE AGE Frequency cell Chi-Square Row Pet 2 Y 13 Y Total 1>4 14 Y 2 0 0.9 0.00 5 0.0556 55.56 1 0.0111 11.11 3 0.0333 33.33 9 3 1 8.1 100.00 0 0 0.1 0.00 0 0.3 0.00 1 0.5 0.00 1 5 1 3 10 Total STATISl'ICSFOR TABLE OF SEASON BY AGE statistic DF Value Frob 3 1 6.502 2.462 0.090 0.117 0.200 -----------------------------------3 10.000 0.019 Chi-Square LikelihCXJd Ratio Chi-Square Mantel-Haenszel Chi-Square Fisher's Exact Test (2-Tail) :Rrl Coefficient Contingency Coefficient Cramer's V 1.000 0.707 1.000 sample Size = 10 WARNING: 100%of the cells have expected counts less than 5. Chi-Square ma.ynot be a valid test. �290 Comparison of Age Distribution Across Seasons Comparison for ~D-16 & year=1986 & sex=M 11:05 SUnday, June 19, 1988 TABlE OF SEASON BY AGE AGE SEASON Frequency Cell au.-Square Row Pet Yea o 1 2 12 13 Y 14 Y 1 1>4 Y Total -+ o o 0.3415 0.00 0.2439 0.00 0.3415 0.00 2.3439 50.00 ---+ 1 0.2927 0.00 0.2111 50.00 0.5378 50.00 0.2927 0.00 0.6341 0.00 0.5378 50.00 6 12 0.0061 32.43 8 7 4 0.1163 21.62 0.0738 18.92 0.0581 10.81 ----------~------_+------_r------_r------~----___+ o o o 1 3 0.0633 16.22 1 ----------~------_+------_r-------_r----,--~------_+ 10 7 13 Total 5 6 STATISTICS FOR TABLE OF SEASON BY AGE statistic au.-Square Likelihcxxi Ratio au.-Square Mantel-Haenszel Chi-Square Rri Coefficient Contingency Coefficient Cramer's V DF Value Frob 8 8 1 6.095 6.924 0.096 0.386 0.360 0.273 0.637 0.545 0.756 Sample Size = 41 WARNING: 73% of the cells have expected counts less than 5. au.-Square may not be a valid test. 2 2 37 41 �291 CoIrparison of Age Distribution Across Seasons CoIrparison for DAU=D-16 & year=1987 & sex=F 11:05 SUnday, June 19, 1988 TABIEOF SEASON BYAGE SEASON AGE Frequency Cell au.-Square ReM Pet Yea 13 Y 12 y Total 1>4 -+ 2 0 0.6667 0.00 2 0.0667 50.00 2 0.3333 50.00 0 0.3333 0.00 4 3 2 0.3333 25.00 3 0.0333 37.50 2 0.1667 25.00 1 0.1667 12.50 8 2 5 4 1 12 Total STATISTICSFORTABIEOF SEASON BYAGE statistic au. -Square Likelihood Ratio au. -Square Mantel-Haenszel Chl.-Square Fisher's Exact Test (2-Tail) Fbi Coefficient contingency Coefficient cramer's V DF Value Prob 3 3 1 2.100 3.001 0.017 0.552 0.391 0.896 0.838 0.418 0.386 0.418 Sample Size = 12 WARNING: 100% of the cells have expected counts less than 5. Chi-Square may not be a valid test. �292 Comparison of Age Distribution Across Seasons Comparison for I:lrill=D-16 & yea:r=1987 & sex=M 11:05 SUnday, June 19, 1988 TABlEOF SEASON BYAGE SEASON AGE Frequency cell Oli-Square Yea Row Pet 13 Y 12 y Total 1>4 14 Y 1 0 0.0526 0.00 0 0.1579 0.00 0 0.4737 0.00 0 0.1053 0.00 1 2.9605 100.00 2 0 0.1053 0.00 1 1.4825 50.00 1 0.0029 50.00 0 0.2105 0.00 0 0.4211 0.00 3 1 0.0296 6.25 2 0.1096 12.50 8 0.0234 50.00 2 0.0592 12.50 9 '2 1 -+ 2 ----+ 3 0.0403 18.75 16 ---+ Total 1 3 4 STATISTICSFOR TABIEOF SEASON BYAGE statistic Oli-Square Likelihood Ratio Chi-Square Mantel-Haenszel Chi-Square Fisher's Exact Test (2-Tail) Fbi Coefficient Contingency Coefficient Cramer's V DF Value Prob 8 8 1 6.234 5.797 1.093 0.621 0.670 0.296 0.641 0.573 0.497 0.405 Sample Size = 19 WARNING:93% of the cells have expected counts less than 5. Chi-Square may not be a valid test. 19 �293 Comparison of Age Distribution Across seasons Comparison for mU=D-3 & year=1986 & sex=M 11:05 SUnday, June 19, 1988 TABLE OF SEASON BY AGE SEASON AGE Frequency Cell Chi-Square ReMPet 2 Y 13 Y Total 1>4 14 Y 1 5 0.0818 83.33 0 0.4 0.00 0 0.4 0.00 1 0.05 16.67 6 2 3 0.0015 75.00 P 0 0.2667 0.00 1 0.4083 25.00 4 0.2667 0.00 3 0.1212 60.00 1 1.3333 20.00 1 1.3333 20.00 0 0.6667 0.00 11 1 1 2 -t 3 5 -t Total 15 STATISTICSFOR TABLE OF SEASON BY AGE Statistic Chi-Square Likelihocxi Ratio Chi-Square Mantel-Haenszel Chi-Square Fisher's Exact Test (2-Tail) Ihi Coefficient Contingency Coefficient Cramer's V OF Value Frob 6 6 1 5.330 6.307 0.026 0.502 0.390 0.871 0.707 0.596 0.512 0.421 Sample Size = 15 WARNING: 100%of the cells have expected counts less than 5. Chi-Square may not be a valid test. �294 Cc::Irrparisonof Age Distribution Across Seasons Cc::Irrparisonfor DAU=D-31 & year=1986 & sex=M 11:05 SUnday, June 19, 1988 TABIEOF SEASON BY AGE SEASON AGE Frequency Cell ari-Square Yea Row Pet 13 Y 12 y 14 Y Total 1>4 1 0 1.1667 0.00 3 1.1435 14.29 4 0.3081 19.05 4 0.0424 19.05 10 0.0501 47.62 21 2 1 0.4587 2.86 2 0.1916 5.71 6 0.1764 17.14 7 0.0205 20.00 19 0.069 54.29 35 3 4 2.3595 11.76 2 0.157 5.88 3 0.7437 8.82 8 0.0942 23.53 17 0.0082 50.00 34 5 7 13 19 46 ---+ Total STATISTICSFOR TABIEOF SEASON BY AGE statistic ari-Square Likelihood Ratio ari-Square Mantel-Haenszel Chi-Square :R1i.Coefficient Contingency Coefficient Cramer's V = DF Value Frob 8 8 1 6.990 7.564 0.084 0.279 0.268 0.197 0.538 0.477 0.771 Sample Size 90 WARNING:60% of the cells have e:xpected counts less than 5. Chi-Square may not be a valid test. 90 �295 Comparison of Age Distribution Across Seasons Comparison for IYill=D-36 & year=1986 & sex=M 11:05 SUnday, June 19, 1988 TABLEOF SEASON BY AGE SEASON AGE Frequency Cell au.-Square Row Pet Yea 13 Y 12 y Total 1>4 14 Y 0 0.2 0.00 2 2 1 66.67 0 0.8 0.00 1 0.05 33.33 0 0.2 0.00 3 3 3 0.25 25.00 4 0.2 33.33 3 0.0125 25.00 1 0.05 8.33 1 0.05 8.33 12 5 4 4 1 1 15 ----+ Total STATISTICSFORTABIEOF SEASON BYAGE statistic au. -Square Likelihood Ratio au. -Square Mantel-Haenszel au.-Square Fisher's Exact Test (2-Tail) Phi Coefficient Contingency Coefficient Cramer's V DF Value Prob 4 4 1 2.813 3.783 0.735 0.590 0.436 0.391 0.824 0.433 0.397 0.433 Sample Size = 15 WARNING: 100% of the cells have expected counts less than 5. Ori-Square may not be a valid test. �296 Ccln'parison of Age Distribution Across Seasons Ccln'parison for I:Ilill=D-4 & year=1986 & sex=M 11:05 SUnday, June 19, 1988 TABIEOF SEASON BYAGE SEASON AGE Frequency Cell Chi-Square Row Pet Yea 13 Y 12 y Total 1>4 14 Y 1 2 0.0559 50.00 1 0.2344 25.00 1 0.9846 25.00 0 0.1538 0.00 0 0.1538 0.00 2 2 0.4207 66.67 1 0.0369 33.33 0 0.2885 0.00 0 0.1154 0.00 0 0.1154 0.00 3 18 0.0566 40.00 19 0.0376 42.22 4 0.0247 8.89 2 0.0419 4.44 22 21 5 2 4 -+ 3 -+-------+ 2 0.0419 4.44 45 -----+ Total 2 STATIsrICS FOR TABIEOF SEASON BYAGE statistic Chi-Square Likelihood Ratio Chi-Square Mantel-Haenszel Chi-Square !hi Coefficient Contingency Coefficient Cramer's V = DF Value Prob 8 8 1 2.762 3.255 0.398 0.230 0.225 0.163 0.948 0.917 0.528 Sample Size 52 WARNING:87% of the cells have expected counts less than 5 • Chi-Square may not be a valid test. 52 �297 Comparison of Age Distribution Across Seasons Comparison for D.AU=D-7& year=1986 & sex=M 11:05 SUnday, June 19, 1988 TABIEOF SEASON BYAGE SEASON AGE Frequency Cell au-Square ReM Pet Yea Total 13 Y 12 y 1 2 1.6562 10.00 12 0.2648 60.00 6 0.2847 30.00 20 2 2 1.1084 50.00 1 0.5523 25.00 1 0.0012 25.00 4 3 3 2.664 60.00 2 0.1329 40.00 0 1.2069 0.00 5 7 15 7 -+ Total 29 STATISTICSFOR TABIEOF SEASONBYAGE statistic au-Square Likelihood Ratio Chi-Square Mantel-Haenszel au-Square Fisher's Exact Test (2-Tail) :Ehi Coefficient Contin:]ency Coefficient Cramer's V DF Value Prob 4 4 1 7.871 8.610 5.684 0.096 0.072 0.017 7. 64E-02 0.521 0.462 0.368 Sample Size = 29 WARNING:89% of the cells have expected counts less than 5. Ori-Square may not be a valid test. �298 Cclrrparison of Age Distribution Across Seasons Comparison for ~E-13 & year=1987 & sex9M 11:05 SUnday, June 19, 1988 TABIEOF SEASON BYAGE SEASON AGE Frequency Cell Chi-Square ReM Pet cal 1Yea 12 y 2 0 0.4 0.00 1 0.9 100.00 0 0.1 0.00 3 4 0.0444 44.44 ~ 0.1 33.33 0.0111 11.11 4 4 Total Total 13 Y 0 0.1 0.00 1 1 1 9 0.0111 11.11 1 10 1 STATISTICSFOR TABLEOF SEASON BYAGE statistic Chi-Square Likelihocx:i Ratio Chi-Square Mantel-Haenszel Chi-Square Fisher's Exact Test (2-Tail) Fhi Coefficient Contingency Coefficient Cramer's V = DF Value Prob 3 3 1 1.667 2.003 0.011 0.644 0.572 0.916 1.000 0.408 0.378 0.408 Sample Size 10 WARNING: 100% of the cells have expected counts less than 5. Chi-Square nay not be a valid test. �299 Comparison of Age Distribution Across Seasons Gamparison for ~E-16 & year=1987 & sex=F 11:05 SUnday, June 19, 1988 TABLEOF SEASON BYAGE SEASON Frequency cell au.-Square RO'itl Pet Yea 13 Y 12 y 15-8 14 Y 2 2 0.0152 18.18 2 0.0152 18.18 1 0.0076 9.09 1 0.3788 9.09 3 0 0.1667 0.00 0 0.1667 0.00 0 0.0833 0.00 1 4.1667 100.00 2 2 1 2 Total 1>8 ---t-------t 4 1 0.0303 0.0076 36.36 9.09 -----t 0 0 0.3333 0.0833 0.00 0.00 ---t--------t 4 1 STATISTICSFORTABLEOF SEASON BYAGE statistic au. -Square Li.kelihcxxl Ratio au. -Square Mantel-Haenszel Chi-Square Fisher's Exact Test (2-Tail) fhi Coefficient Contingency Coefficient Cramer's V = DF Value Prob 5 5 1 5.455 4.111 0.000 0.363 0.533 1.000 0.667 0.674 0.559 0.674 Sarrple Size 12 WARNING: 100% of the cells have expected courrts less than 5. Chi-Square may not be a valid test. Total 11 1 12 �300 Cc::Irrparisonof Aqe Distribution Across Seasons Cc::Irrparisonfor IlMJ=:E-22 & year=1986 & sex=M 11:05 SUnday, June 19, 1988 TABIEOF SEASON BY AGE SEASON AGE Frequency cell Chi-Square Row Pet Yea 13 Y 12 y Total 15-8 14 Y 2 5 0.0184 50.00 0 0.5882 0.00 4 0.0627 40.00 0 0.5882 0.00 1 0.2882 10.00 10 3 3 0.0263 42.86 1 0.8403 14.29 2 0.0896 28.57 1 0.8403 14.29 0 0.4118 0.00 7 8 1 6 1 --+ Total 1 STATISTICSFORTABlE OF SEASON BYAGE statistic Chi-Square Likelihocx:i Ratio Chi-Square Mantel-Haenszel Chi-Square Fisher's Exact Test (2-Tail) Phi Coefficient Contingency Coefficient Cramer's V DF Value Prob 4 4 1 3.754 4.812 0.008 0.440 0.307 0.928 0.593 0.470 0.425 0.470 Sample Size = 17 WARNING: 100% of the cells have expected counts less than 5. Chi-Square may not be a valid test. 17 �301 Col'rprrison of Age Distribution Across Seasons Col'rprrison for ~E-22 & year=1987 & sex=F 11:05 SUnday, June 19, 1988 'ITffiIEOF SEASON BYAGE SEASON AGE Frequency 1 Cell Ori -Square ReM Pet cal I 1Yea 12 y Total 13 Y -t 2 5 0.1113 8.93 9 0.0172 16.07 15 0.3338 26.79 4 0.6838 7.14 56 3 0 0.6923 0.00 1 0.106~ 11.11 0 2.0769 0.00 3 4.2549 33.33 9 5 10 15 7 65 Total (Continued) TABIEOF SEASON BYAGE SEASON Frequency Cell Ori -Square ReM Pet AGE 1 I4 Y 15-8 Total 1>8 2 6 0.1155 10.71 12 0.0659 21.43 5 0.1113 8.93 56 3 2 0.7188 22.22 3 0.4103 33.33 0 0.6923 0.00 9 8 15 5 65 Total statistic STATISI'ICSFOR TABIEOF SEASON BYAGE DF Value Ori-Square Likelihood Ratio Ori-Square Mantel-Haenszel Ori-Square Ihl. Coefficient Conti.rgency Coefficient Cramer's V = 6 6 1 10.391 12.209 0.296 0.400 0.371 0.400 Frob 0.109 0.057 0.586 Sample Size 65 WARNING:64% of the cells have expected counts less than 5. Ori-Square may not be a valid test. �302 Comparison of Age Distribution Across Seasons Comparison for DMJ=E-22 & year=1987 & sex=M 11:05 SUnday, June 19, 1988 TABIEOF SEASON BYAGE SEASON AGE Frequency 1 Cell au. -Square ReM Pet cal I 1Yea 12 y Total 13 Y 1 0 0.7742 0.00 1 0.1942 12.50 4 0.4751 50.00 1 0.1942 12.50 8 2 3 1.1157 17.65 3 0.025~ 17.65 4 0.6847 23.53 4 0.1531 23.53 17 3 0 0.5806 0.00 2 0.6057 33.33 3 0.3563 50.00 1 0.0224 16.67 6 3 6 11 6 -+ Total (Continued) SEASON Frequency Cell au.-Square ReM Pet AGE 15-8 4 Y Total 1>8 1 1 0.4536 12.50 0 0.2581 0.00 1 0.4536 12.50 8 2 1 0.0085 5.88 1 0.3719 5.88 1 0.0085 5.88 17 3 0 0.3871 0.00 0 0.1935 0.00 0 0.3871 0.00 6 2 1 2 31 Total 31 �303 STATISTICS FOR TABlE OF SEASON BY AGE statistic DF Value Frob Chi-Square Likelihood Ratio Chi-Square Mantel-Haenszel Chi-Square Fisher's Exact Test (2-Tail) :Rri Coefficient Contingency Coefficient Cramer's V 12 12 1 7.704 9.732 1.339 0.808 0.639 0.247 0.907 0.499 0.446 0.353 Sample Size = 31 WARNING: 95%of the cells have expected counts less than 5. Chi-Square may not be a valid test. �304 Comparison of Age Distribution Across Seasons Comparison for DAU=E-25 & year=1987 & sex=F 11:05 SUnday, June 19, 1988 TABIEOF SEASON BYAGE SEASON AGE Frequency 1 Cell Ori -Square ReM Pet cal I 1Yea 12 y 13 'Ibtal Y 2 3 0.0127 11.54 3 0.3773 11.54 4 0.5031 15.38 8 0.1347 30.77 26 3 1 0.0301 9.09 0 0.891~ 0.00 0 1.1892 0.00 2 0.3184 18.18 11 4 3 4 10 -+ 'Ibtal (Continued) SEASON AGE Frequency 1 Cell Ori -Square ReM Pet 4 Y I 15-8 'Ibtal 1>8 2 4 0.0111 15.38 3 1.2226 11.54 1 0.1169 3.85 26 3 2 0.0262 18.18 5 2.8897 45.45 1 0.2764 9.09 11 6 8 2 37 'Ibtal 37 STATISTICSFOR TABIEOF SEASON BYAGE statistic Ori -Square Likelihood Ratio Ori -Square Mantel-Haenszel Ori -Square Fisher's Exact Test (2-Tail) Fbi Coefficient Contingency Coefficient Cramer's V DF 6 6 1 Value 8.000 9.531 3.572 Frob 0.238 0.146 0.059 0.268 0.465 0.422 0.465 Sample Size = 37 WARNING:86% of the cells have expected counts less than 5. Ori-Square may not be a valid test. �305 Comparison of Age Distribution Across Seasons Comparison for IlAU=E-3 & year=1986 & sex=F 11:05 SUnday, June 19, 1988 OF SEASON BY AGE TABIE SEASON Frequency cell Chi-Square Row Pet AGE 1 Ical Total 1 Yea 2 7 0.0234 22.58 3 0.0873 9.68 2 0.6818 6.45 5 0.1978 16.13 31 3 6 0.0242 20.00 4 0.0902 13.33 5 0.7045 16.67 7 30 0.2044 23.33 13 7 7 12 Total (Continued) TABIE SEASON AGE Frequency 1 cell Chi-Square Row Pet 4 Y I OF SEASON BY AGE 15-8 Total 1>8 2 6 0.4447 19.35 5 0.0013 16.13 3 1.4278 9.68 31 3 3 0.4596 10.00 5 0.0014 16.67 0 1.4754 0.00 30 9 10 3 61 Total statistic STATISl'ICSFOR TABIE OF SEASON BY AGE DF Value Chi-Square Likelihood Ratio Chi-Square Mantel-Haenszel Chi-Square !hi Coefficient Contingency Coefficient Cramer's V = 61 6 6 1 5.824 7.045 1.731 0.309 0.295 0.309 Frob 0.443 0.317 0.188 Sample Size 61 WARNING:64% of the cells have expected counts less than 5. Chi-Square may not be a valid test. �306 Comparisonof Age Distribution Across Seasons Comparisonfor Illill=E-32 & year=1986 & sex=M 11:05 SUnday,June 19, 1988 TABlE OF SEASON BY AGE AGE Frequency Cell Chi-Square RowPet Yea Total 13 Y 12 y 1 0 0.8182 0.00 3 0 50.00 3 0.3068 50.00 6 2 1 0.0022 14.29 4 0.0714 57.14 2 0.1169 28.57 7 3 2 0.4865 22.22 4 0.0556 44.44 3 0.0227 33.33 9 3 11 8 22 Total STATISTICS FOR TABIE OF SEASON BY AGE statistic Chi-Square Likelihood Ratio Chi-Square Mantel-Haenszel Chi-Square Fisher's Exact Test (2-Tail) Rri Coefficient Contingency Coefficient Cramer's V DF Value Prob 4 4 1 1.880 2.596 1.049 0.758 0.627 0.306 0.902 0.292 0.281 0.207 Sample Size = 22 100%of the cells have expected counts less than 5. Chi-Square maynot be a valid test. WARNING: �307 Ccln'parison of Age Distribution Across Seasons Ccln'parison for IW.J=E-32 & year=1987 & sex=M 11:05 SUrrlay, June 19, 1988 TABIEOF SEASON BYAGE SEASON AGE Frequency Cell ali-Square Row Pet cal 1Yea 13 Y 12 y Total 14 Y 1 0 0.3529 0.00 0 0.1765 0.00 2 0.0018 66.67 1 3.8431 33.33 0 0.3529 0.00 3 2 2 0.1448 15.38 1 0.0724 7.69 8 0.0202 61.54 0 0.7647 0.00 2 0.1448 15.38 13 3 0 0.1176 0.00 0 0.0588 0.00 1 0.1925 100.00 0 0.0588 0.00 0 0.1176 0.00 1 2 1 11 1 2 17 -+ Total STATIsrrCS FORTABIEOF SEASON BYAGE statistic ali-Square Likelihood Ratio ali-Square Mantel-Haenszel ali-Square Fisher's Exact Test (2-Tail) !hi Coefficient Contingency Coefficient Cramer's V DF Value Frob 8 8 1 6.420 6.339 0.236 0.600 0.609 0.627 0.653 0.615 0.524 0.435 Sample Size = 17 WARNING: 93% of the cells have expected counts less than 5. ali -Square may not be a valid test. . �308 Comparison of Age Distril:ution Across Seasons Comparison for IW.FE-33 & year=1986 & sex=F 11:05 SUnday, June 19, 1988 TABIEOF SEASONBY AGE SEASON Frequency Cell ari.-Square Row R::t AGE Cal 1Yea 13 12 y Total Y 2 0 0.2174 0.00 0 0.4348 0.00 0 1.3043 0.00 1 0.0196 20.00 5 3 2 0.0265 4.88 4 0.053 9.76 12 0.1591 29.27 7 0.0024 17.07 41 2 4 12 8 46 Total (Continued) TABIEOF SEASONBY AGE SEASON Frequency Cell ari. -Square Row R::t I AGE 1 4 Y 15-8 Total 1>8 2 1 1.3928 20.00 2 1.0671 40.00 1 0.0196 20.00 5 3 2 0.1698 4.88 7 0.1301 17.07 7 0.0024 17.07 41 3 9 8 46 Total STATIsrrCS FOR TABIEOF SEASONBYAGE Statistic DF ari.-Square Likelihood Ratio ari. -Square Mantel-Haenszel ari.-Square !hi Coefficient Conti_nJency Coefficient Cramer's V = 6 6 1 Value 4.999 6.217 1.001 0.330 0.313 0.330 Frob 0.544 0.399 0.317 Sample Size 46 WARNING:71% of the cells have expected courrts less than 5. ari.-Square may not be a valid test. �Comparison of Age Distribution Across seasons, Comparison for Dlill=E-33 & year=1986 & sex=M: 11: 05 SUnday, Jrme 19, 19, :' TABIE SEASON OF SEASON BY AGE AGE , Frequency Cell Chi-Square '?n .n: Yea Row Pet , 13 Y 12 y 15c.~ ,. 14 Y : 14 0.3.333 35-~90 Total 1 0 0.4286 0.00 3 1.1868 7.69 10 0.1172 25.64 12 0.4658 30.77 2 0 0.4505 0.00 7 0.223 17.07 13 0.1411 31.71 9 0.1792 21.95 3 1 6.3936 9.09 3 1.2987 27.27 3 0.0065 27.27 2 0.219 18.18 2 0.5664 18.18 11 1 13 26 23 28 91 Total .12 Q,.iJi)3 29'.27 STATISTICSFOR TABIE OF SEASON BYAGE statistic Chi-Square Likelihood Ratio Chi-Square Mantel-Haenszel Chi-Square Rri Coefficient Contin;Jency Coefficient Cramer's V DF Value ~IProb 8 8 1 12.040 9.029 5.406 0.364 0.342 0.257 {O.149 19•340 lID.020 :i ( ( Sample Size = 91 t WARNING:47% of the cells have expected cormts ~ess than 5• Chi-Square may not be a valid t:est. 39 41 �310 Cc::llrparisoft".ofAge Distribution Across seasons CoIrparison for DMJ=E-34 & year=1986' -&_ seX:=F ,'_ 11: 05 Stmday, Jrme 19, 1988 TABIEOF SEMONBY AGE SEASON Frequency Cell Chi-Square RowPet AGE cal Q_ ..- -00.5714 0.,00 2 3 12 1 Yea 2 1.5238 33 •.33 0.2857 0.•00 2 0.2286 : _. 13.33 1 o 0.6095 6.67 15 1 0.1143 6.67 3-- - 1 6 0.2857 0.00 1 0.1143 6.67 Total· _ (Cont.inued) Total Y 21 1 ,'::". .__ TABLE. QF'" SEASON:.. BYAGE:: :" , ~ ' SEASON Frequency Cell Chi-Square RowPet AGE 15-8 4 Y 1 0.0238 16.67 2 Total 1>8 'O;;~~~ --r 1:' 0.2976 16.67 ~ "2 6 -. ~ .. : .' n',' ," ---+ 3 2 0.0095 13.33 Total 5 0.119 33.33 3 15 3 0.0914 ' 20.00_ 6.--- 5 .: 'i,.' 21 '7 ,,- statistic STATISTICSFUR TABIEOF SEASONBYAGE DF" Value Chi-Square Likelihcx:xl Ratio Chi-Square Mantel-Haenszel Chi-Square Fisher's Exact Test (2-Tail) Fbi Coefficient Contingency Coefficient Cramer's V = 6 6 1 4.503 5.352 0.285 - Prob 0.609 0.499 0.594 0.787 0.463 0.420 0.463 Sample Size 21 WARNING: 100% of the cells have expected cormts less than 5. Chi-Square may not be a valid test. �311 comparison of Age Distribution Across Seasons comparison for ~E-34 & year=1986 & ~ 11:05 SUnday, June 19, 1988 TABLEOF SEASON BYAGE SEASON AGE Frequency cell Chi-Square Yea Row Pet 12 y 13 Y 15-8 14 Y Total --+ 2 3 0.7212 33.33 2 0.0083 22.22 2 0.1667 22.22 1 1.0417 11.11 1 1.0417 11.11 9 3 10 0.4327 66.67 3 0.005 20.00 2 0.1 13.33 0 0.625 0.00 0 0.625 0.00 15 13 5 4 1 1 24 Total STATISl'ICSFOR TABIEOF SEASON BYAGE statistic Chi-Square Likelihood Ratio Chi-Square Mantel-Haenszel Chi-Square Fisher's Exact Test (2-Tail) !hi Coefficient Contingency Coefficient Cramer's V DF Value Prob 4 4 1 4.767 5.434 4.188 0.312 0.246 0.041 0.351 0.446 0.407 0.446 San'ple Size = 24 WARNING:90% of the cells have expected counts less than 5 • Chi-Square ma.ynot be a valid test. �312 canparison of Age Distribution Across Season, canparison for I::lAU=E-34 & year=1987 & sex=e 11:05 SUn' :' " TABlE SEASON OFSEASON BY AGE AGE Frequency Cell Chi-Square Row Pet cal 13 Y 15-8 14 Y 1>8 ,,·· .. _·--··'r 2 2 0.0842 22.22 3 1 0.3788 50.00 I 2 3 0.2475 11.11 0.0808 22.22 3 .. 0.1212 33.33 I .,:1.1 I 0 0 0.3636 0.00 0.5455 0.00 0 0.:' tn,S C.OCl r 2 2 3 I , ,..,._.... ",.,."._ + '_ Total , I o. t!,';:·~;4: 1 1.1136 50.00 . .1. '_". I 1 I co.,' ~f' " STATISTICS FOR TABlE OFSEASON BY AGE statistic Chi-Square Likelihood Ratio Chi-Square Mantel-Haenszel Chi-Square Fisher's Exact Test (2-Tail) Ib.i Coefficient ContinJency Coefficient Cramer's V DF Value 4 4 1 3.157 3.839 1.250 ..... , ',0" 0.536 0.472 0.536 Sample Size = 11 WARNING: 100%of the cells have expected counts ] s· :~.,; than 5. Chi-Square may not be a valid i.r:,· '",:~. .', �313 I .._,-, \ .. Comparison of Age Distribution Across Seasons Comparison for ~E-39 & year=1987 & sex=F 11:05 SUnday, June 19, 1988 TABlEOF SEASON BYAGE SEASON Frequency Cell au-Square AGE Rc¥ Pet cal . 2 -f_· =:: ... .0:. ;,')!' 3 \"", Total 13 Y 3 0.1212 11.11 4 0.1616 14.81 4 0.9899 14.81 3 0.1212 11.11 27 0 0.5455 0.00 0 4 0.727,?> 4.4545 0.00 66.67 0 0.5455 0.00 6 3 33 Total~ ...... "; (Continped) 3 '0' r 12 y 1Yea 4 8 TABlEOF SEASON BY AGE SEASON AGE Frequency Cell PJ,j;~Square ReM Pet~,. 4 Y : _,~", 3 Total Total 1>8 .•. \ .••. ..'.~> ~.~. 2 15-8 _. statistic 6 0.2424 22.22 4 0.1684 14.81 3 0.1212 11.11 27 o 2 0.7576 33.33 0 0.5455 0.00 6 1.0909 0.00 6 6 3 33 STATISTICSFOR TABlEOF SEASON BYAGE DF Value au-Square Likelihood Ratio au-Square Mantel-Haenszel au-Square Fisher's Exact Test (2-Tail) Fbi Coefficient Contingency Coefficient Cramer's V = 6 6 1 10.593 12.565 0.141 Frob 0.102 0.050 0.708 0.142 0.567 0.493 0.567 Sample Size 33 WARNING:93% of the cells have expected counts less than 5 • au-Square may not be a valid test. �314 Comparison of Age Distribution Across Seasons Comparison for Dt\U=E-39 & year=1987 & sex=M 11:05 SUnday, June 19, 1988 TABIEOF SEASONBYAGE SFASON' AGE Frequency Cell <lri-Square Row Pet cal 1 2 1Yea 15-8 13 Y 12 y 1>8 0 0.125 0.00 0 0.25 0.00 0 0.75 0.00 1 0.5 50.00 1 2.25 50.00 1 0.142 9.09 2 0.2841 18.18 5 0.1856 45.45 2 0.2045 18.18 1 0.1023 9.09 0 0.1875 0.00 0 0.375 0.00 1 0.0139 33.33 1 0.0833 33.33 0 0.375 0.00 1 2 6 4 2 ------+ 0 Total 2 0.125 0.00 ------+ 0 11 0.6875 0.00 ---+ 3 Total STATISTICSFOR TABLEOF SEASONBYAGE Statistic DF Value Frob ari-Square Likelihood Ratio Chi-Square Mantel-Haenszel <lri-Square Fisher's :Exact Test (2-Tail) Fbi Coefficient Contingency Coefficient Cramer's V 10 10 1 10.162 10.108 0.976 0.426 0.431 0.323 0.551 = 0.797 0.623 0.564 Sample Size 16 WARNING: 100% of the cells have expected courrcs less than 5. Chi-Square may not be a valid test. 1 3.5208 33.33 ----+1 3 16 �315 Corrparison of Age Distribution Across Seasons Comparison for DAU=E-41 & year=1987 & sex=F 11:05 sunday, June 19, 1988 TABlE OF SEASON BYAGE t : SEASON Frequency. Cell au.-Square Row Pet 2 3 AGE cal 1Yea Total 13 Y 12 y 1 0.3682 6.67 3 0.0104 20.00 4 0.2104 26.67 15 0.297 13.33 3 0.3068 16.67 1 0.2475 5.56 4 0.0087 22.22 3 0.1753 16.67 18 4· 3 7 7 33 Total (Continued) 2 " TABLEOF SEASON' BYAGE SEASON Frequency Cell au.-Square Row Pet AGE 4 Y 15~8 Total 1>8 2 0 0.9091, 0.00' 3 0.1114 20.00 2 1.3091 13.33' 15 3 2 0.7576 11.11 5 0.0928 27.78 0 1.0909 0.00 18 2 8 2 33 Total STATISTICSFORTABlEOF SEASONBYAGE statistic au. -Square Likelihcx:d Ratio au. -Square Mantel-Haenszel au. -Square Fisher's Exact Test (2-Tail) Fhi Coefficient Contingency Coefficient Cramer's V DF Value Prob 6 6 1 5.895 7.450 0.801 0.435 0.281 0.371 0.575 0.423 0.389 0.423 Sample Size = 33 WARNING: 100% of the cells have expected counts less than 5. au.-Square may not be a valid test. �316 Ccmprrison of Age Distribution Across Seasons comparison for DAU=E-43 & year=1986 & sex=F 11:05 SUnday, June 19, 1988 TABIEOF SFASONBYAGE SEASON Frequency Cell Chi-Square Row Pet AGE Cal 1Yea 13 12 y Total Y 2 1 0.0833 33.33 1 4.8167 33.33 0 0.75 0.00 0 0.3 0.00 3 3 4 0.0147 23.53 0 0.8~ 0.00 5 0.1324 29.41 2 0.0529 11.76 17 5 1 5 2 20 Total (Continued) TABIEOF SFASONBYAGE SEASON Frequency Cell Chi-Square Row Pet AGE 4 Y 15-8 o 2 0.15 0.00 3 1 0.0265 5.88 Total statistic Total 1>8 o 1 0.2667 33.33 3 0.3 0.00 3 0.0471 17.65 2 0.0529 11.76 17 4 2 20 1 STATISTICSFORTABIEOF SFASONBYAGE DF Value Chi-Square Likelihood Ratio Chi-Square Mantel-Haenszel Chi-Square Fisher's Exact Test (2-Tail) Ihi Coefficient Contingency Coefficient Cramer's V = 6 6 1 7.843 7.406 0.423 Prob 0.250 0.285 0.515 0.430 0.626 0.531 0.626 Sample Size 20 WARNING: 100% of the cells have expected counts less than 5. Chi-Square may not be a valid test. �317 Ccrnparison of Age Distribution Across Seasons Ccrnparison for IYill=E-43 & year=1986 & sex=M 11:05 SUnday, June 19, 1988 TABlEOF SEASON BYAGE SEASON AGE Frequency Cell Chi-Square 12 Yea Row Pet 1 14 Y Total 15-8 Y o 4 0.0289 44.44 2 0.6057 22.22 o 2.0323 0.00 6 2.0157 85.71 0.9032 0.00 2.6544 14.29 8 3.0511 53.33 5 0.7025 33.33 2 0.0022 13.33 o o 0.4839 0.00 0.9677 0.00 9 15 4 1 0.9956 11.11 o 2 3 13 Y Total 0.2903 0.00 2 3.4695 22.22 9 -+ 1 1 o 2 STATISTICSFOR TABIEOF SEASON BYAGE statistic Chi-Square Likelihood Ratio Chi-Square Mantel-Haenszel Chi-Square Fisher's Exact Test (2-'l'ail) fhl Coefficient Contingency Coefficient Cramer's V = 7 0.4516 0.00 OF Value Frob 8 8 1 18.655 20.493 6.980 0.017 0.009 0.008 1. 14E-02 0.776 0.613 0.549 Sample Size 31 WARNING:93% of the cells have expected courrts less than 5. Clri-Square may not be a valid test. 15 31 �318 Corrparison of Age Distribution Across Seasons Corrparison for J:ll!JJ=E-43 & year=1987 & sex==F 11:05 Sunday, June 19, 1988 TABlE OF SEASONBYAGE SEASON Frequency Cell Chi-Square Row Pet AGE cal 1Yea 12 y Total 13 Y 2 5 0.6722 13.89 3 0.4033 8.33 7 0.0417 ·19.44 7 0.0417 19.44 36 3 5 1.7286 35.71 3 1.0371 21.43 2 0.1073 14.29 2 0.1073 14.29 14 10 6 9 9 50 Total (Continued) TABlE OF SEASONBY AGE SEASON Frequency Cell Chi-Square Row Pet AGE 4 Y 15-8 Total 1>8 2 4 0.4356 11.11 6 0.1829 16.67 4 0.0444 11.11 36 3 0 1.12 0.00 1 0.4702 7.14 1 0.1143 7.14 14 4 7 5 50 Total statistic STATISTICSFORTABLEOF SEASONBY AGE DF Value Chi-Square Likelillcx:xi Ratio Chi-Square Mantel-Haenszel Chi-Square Fhi. Coefficient Contingency Coefficient Cramer's V 6 6 1 6.507 7.300 2.911 0.361 0.339 0.361 Frob 0.369 0.294 0.088 Sample Size = 50 WARNING:71% of the cells have expected counts less than 5. Chi-Square may not be a valid test. �319 Comparison of Age Distribution Across Seasons Comparison for ~E-43 & year=1987 & sex=M 11:05 SUnday, June 19, 1988 TABIEOF SEASON BY AGE SEASON AGE Frequency Cell <lrl-Square Row Pet cal 1Yea 12 y Total 13 Y 2 2 0.0182 40.00 1 0.7127 20.00 1 0.6545 20.00 1 0.6545 20.00 5 3 2 0.0152 33.33 ~ 0.5939 66.67 0 0.5455 0.00 0 0.5455 0.00 6 4 5 1 1 11 Total STATISl'ICSFOR TABIEOF SEASON BYAGE statistic Clri-Square Likelihood Ratio <lrl-Square Mantel-Haenszel Chi-Square Fisher's Exact Test (2-'l'ail) Fbi Coefficient Contingency Coefficient Cramer's V DF Value Frob 3 3 1 3.740 4.609 0.871 0.291 0.203 0.351 0.264 0.583 0.504 0.583 Sample Size = 11 WARNING: 100%of the cells have expected counts less than 5 • Chi-Square may not be a valid test. �320 Comparison of Aqe Distribution Across Seasons Comparison for I:li\U=E-6 & year=1986 & sex=F 11:05 SUnday, June 19, 1988 TABIE OF SEASONBYAGE SEASON Frequency Cell au.-Square RcM~ AGE cal 1Yea 13 12 y Total Y 2 11 2.4985 22.45 10 1.4373 20.41 8 0.0131 16.33 7 0.09 14.29 49 3 3 2.4005 5.88 4 1.3809 7.84 9 0.0126 17.65 9 0.0865 17.65 51 14 14 17 16 100 Total (Continued) TABIE OF SEASONBYAGE SEASON Frequency Cell au.-Square RcM~ 2 3 Total statistic AGE 4 Y 15-8 Total 1>8 3 0.0539 6.12 9 0.4572 18.37 1 2.6368 2.04 49 4 0.0518 7.84 14 0.4393 27.45 8 2.5334 15.69 51 7 23 9 100 .... STATISTICSFOR TABIE OF SEASONBYAGE OF Value au. -Square Likelihood Ratio Chi -Square Mantel-Haenszel au. -Square Rrl Coefficient Contingency Coefficient Cramer's V 6 6 1 14.092 15.223 12.298 0.375 0.351 0.375 Prob 0.029 0.019 0.000 Sample Size = 100 WARNING:29% of the cells have expected COlIDts less than 5. Chi-Square may not be a valid test. �321 Comparison of Age Distribution Across Seasons Oornparison for DAU=E-6 & year=1986 & sex=M 11:05 SUnday, June 19, 1988 TABIE OF SEASON BY AGE SEASON AGE Frequency Cell au.-Square Row Pet Yea o 1 1.6949 0.00 2 3 1.5256 13.64 3 13 12 Y 23 1.0177 92.00 17 0.0214 77.27 Y 1>8 15-8 14 Y Total 2 o o o 0.0066 8.00 0.8475 0.00 0.8475 0.00 0.8475 0.00 1 0.4008 4.55 1 0.0867 4.55 o 0.7458 0.00 --+ o 4 2 2 2 2.737 33.33 0.9503 16.67 0.8651 8.33 6.2401 16.67 6.2401 16.67 4 44 5 2 2 1 22 0.7458 0.00 ----+ 0.0427 8.33 1 25 12 -+-----+ Total STATISTICS FOR TABIE OF SEASON BY AGE statistic DF Value Prob au. -Square Likelihood Ratio au. -Square Mantel-Haenszel au. -Square Phi Coefficient Contingency Coefficient Cramer's V 10 10 1 25.863 25.385 10.881 0.662 0.552 0.468 0.004 0.005 0.001 Sample Size = 59 WARNING: 83% of the cells have expected counts less than 5. au.-Square may not be a valid test. 2 59 �322 Cclrrparisonof Age Distribution Across Seasons Cclrrparisonfor Jl.lill=E-6 & year=1987 & sex=F 11:05 SUnday, June 19, 1988 TABIE OF SFASON BY AGE SEASON AGE Frequency Cell au.-Square RaN Pet cal 1Yea Total 13 Y 12 y 2 33 0.8392 10.89 50 0.3832 16.50 55 0.0071 18.15 46 0.0505 15.18 303 3 8 0.3456 7.55 17 0.0592 16.0:4 22 0.3315 20.75 15 0.0207 14.15 106 4 2 1.8127 3.70 3 3.2665 5.56 8 0.3694 14.81 7 0.1093 12.96 54 43 70 85 68 463 Total (Continued) SEASON AGE Frequency Cell au.-Square ReM Pet 4 Y 15-8 Total 1>8 2 30 0.7854 9.90 65 0.4562 21.45 24 2.3246 7.92 303 3 5 1.7287 4.72 23 0.1204 21.70 16 1.8109 15.09 106 4 4 0.0662 7.41 20 4.3519 37.04 10 2.9797 18.52 54 39 108 50 463 Total STATISI'ICS FOR TABLE OF SFASON BY AGE statistic DF Value Frob --------------------- --------------------22.219 0.035 12 au.-Square Likelihcxxi Ratio au.-Square Mantel-Haenszel au.-Square fbi Coefficient Contingency Coefficient Cramer's V Sample Size = 463 12 1 22.946 14.244 0.219 0.214 0.155 0.028 0.000 �323 Cclrrparison of Age Distribution Across Seasons Cclrrparison for D.2ill=E-6 & year=1987 & sex=M 11:05 SUnday, June 19, 1988 TABIEOF SEASON BY AGE SEASON AGE Frequency Cell Chi-Square Row Pet cal 1Yea Total 13 Y 12 y 1 1 0.05 100.00 0 0.05 0.00 0 0.1 0.00 0 0.05 0.00 1 2 8 0 80.00 1 0.5 10.00 0 1 0.00 1 0.5 10.00 10 3 6 0.025 75.00 0 0.4 0.00 2 1.8 25.00 0 0.4 0.00 8 4 1 0.05 100.00 0 0.05 0.00 0 0.1 0.00 0 0.05 0.00 1 16 1 2 1 20 Total STATISTICSFOR TABLEOF SEASON BYAGE statistic Chi-Square Likelihood Ratio Chi-Square Mantel-Haensz~l Chi-Square Fisher's Exact Test (2-Tail) Fhi. Coefficient Contingency Coefficient Cramer's V = DF Value Frob 9 9 1 5.125 6.556 0.023 0.823 0.683 0.880 0.571 0.506 0.452 0.292 Sample Size 20 WARNING:88% of the cells have expected counts less than 5. Chi-Square may not be a valid test. �324 Cclrrparison of Age Distribution Across Seasons .. COrrparison Across Seasons for Deer with all Di\Us 11:05 SUnday, June 19, 1988 -------- SPECIES=Deer YEAR=1986SEX=F---------TABIEOF SFASONBYAGE SEASON AGE Frequency Cell Chi-Square Row Pet Faw 1Yea 13 Y 12 y 1>4 14 Y --------+ 2 8 0.3208 23.53 8 1.372 23.53 6 0.0463 17.65 4 0.4644 11.76 4 0.0943 11.76 3 13 0.1454 17.33 9 0.622 12.00 15 0.021 20.00 14 0.2105 18.67 21 17 21 18 Total 4 0.993 11.76 34 7 0.0427 9.33 17 0.4502 22.67 75 11 21 109 STATISTICSFORTABLEOF SFASONBYAGE statistic Chi-Square Likelihcxrl Ratio au-Square Mantel-Haenszel Chi-Square Rri Coefficient Contingency Coefficient cramer's V Sample Size = 109 Total DF Value Frob 5 5 1 4.782 4.812 2.602 0.209 0.205 0.209 0.443 0.439 0.107 ------+ ----+ �325 Ccarparison of Age Distribution Across Seasons Ccarparison Across Seasons for Deer with all Di\IJs 11:05 SUnday, June 19, 1988 ------- SPECIES=DeerYFAR=1986SEX=M-------------TABlEOF SEASON BYAGE SEASON AGE Frequency Cell Chi-Square Row 'Pet Yea 12 Y 13 Y 14 Y 1>4 Total 8 1.525 11.94 26 0.7539 38.81 16 1.0872 23.88 5 1.3739 7.46 12 218E-8 17.91 67 2 19 0.8227 14.96 38 0.3075 29.92 27 0.5584 21.26 18 0.273 14.17 25 0.2253 19.69 127 3 45 2.0979 22.84 64 0.0037 32.49 29 1.4594 14.72 26 0.0697 13.20 33 197 0.1459 16.75 1 -~------+-------+-----~------+ 72 128 72 49 70 Total STATISTICSFORTABlEOF SEASON BYAGE statistic Chi-Square Likelihood Ratio Chi-Square Mantel-Haenszel Chi-Square !hi Coefficient Contingency Coefficient Cramer's V Sample Size = 391 DF Value Frob 8 8 1 10.704 11.002 1.113 0.165 0.163 0.117 0.219 0.202 0.292 391 �326 Comparison of Age Distribution Across Seasons -.Comparison Across Seasons for Deer with all DAUs 11:05 SUnday, June 19, 1988 ,-------- SPECIES=DeerYFAR=1987 SEX=F---------TABIEOF SEASON BYAGE SEASON AGE Frequency Cell au.-Square Row Pet Faw 1Yea 12 y 13 Y Total 1>4 14 Y -----+ 2 0 0.6154 0.00 0 1.2308 0.00 6 1.1308 18.75 13 1.0102 40.63 3 0.1178 9.38 3 1 0.9846 5.00 2 1.9692 10.00 9 1.8092 45.00 3 1.6163 15.00 1 0.1885 5.00 4 0.356 20.00 20 1 2 15 16 4 14 52 10 0.2225 31.25 32 ----+ Total STATISTICSFOR TABIEOF SEASON BYAGE statistic au. -Square Likelihood Ratio au-Square Mantel-Haenszel au -Square :Rri Coefficient Contingency Coefficient Cramer's V OF Value Prob 5 5 1 11.251 12.410 5.412 0.465 0.422 0.465 0.047 0.030 0.020 Sample Size = 52 'WARNING: 50% of the cells have expected counts less than 5. au-Square may not be a valid test. �-.:r< ~_;: ..~.'-~~', (~~~':'~~~~;., (~:~i;1~(:,'J;~~:'~.~ ':_:.;,~: :.r\ {f- '"i ..;.' /:r::~.~ ",~:~,~::: -' "~'~ '.. ' . <' I H • ;I{", ,_,.. •. tc -_ , . t.. ;\~ .. ;( . 1.;' ~i~ .. 'f I, - .•..•... '~.•• f;.;_.... -'.",,·.·,.~- •.. ..- .. t:~.""" ... 01'. ~.16/). i I C.!._·d 7.69 .1i ~-.. ~ 0.~~9~i. \ ~,:·._.5~ '~ :'1 t . ;:;..538 23r>()8 ..___~ ,,) o. \J6'.:' G 2,(:: .;!,,:' :':.;~,: ~ =, .1 j --.~-~,--i--...,',"'--'-"""'"';'" - -r: ") 1..) 0,,5766- J..:~:..o. 0.00 ..:. ... .•,.•.......•...•.. ~"" ... , ..•.. l .....•.. ',' " ',._ .. c- ....,:: ..•.. .,:: " J.., ~)7T 2<-94 1,.·:'{::5~~' O~~'Lj,,,~'I fl." .~_':~~~: i' i ,; n 2'g.41! I 2S;,,4:'. I ,.''''-'.-...,-+--_. _,_."....,+,. 'j;:1. ,,- • 0 • -. 9 :2:), 25· _. "", V~L.UC C~.13·--(~(1·O::.:L~(~ ]~.Jk:2.1~;l1-·:~:r_;'_~-~i·,'1_;;t.. ~j:~·_:~;t l~_';~_'·~:;·~~C~ll~~::~.:! ft,;_:~.·i;·':'~:;~~. '-}~:lt:!,n£~~et ..v:( '.(jj.~...Gtfi_.k·.j..l "',:: i:.·· h,l, ('!..''y.;;,:ff ·~.\.·.!it:n~·:: (!r}rlt:irB'e~r(:~ ~ (1;':?;f\~j_,,~:i-f~.{It~ .,:... c.~jT"t:'J.:"'t'!~\7 Sa!nr:}lt~~'~ .sj..~:..:~ :;';:'11.:' tb2 c:e~:.:tS hiiV,,,, e:'\.3:'y?'G"'~Xi C'::. ,-Int,:;: Less (~hj_--S(:r~a"rf~' '_'."~'::;;,z a ··'/3.1~·_a t.(~.t.;. t-l1D"."1 50 '.;,' ~':i-' ,~_ I ~'T?"::~'m~:;: ::.')6\ o:f' ._. 'i. z- �328 Corrparison of Age Distribution Across Seasons Corrparison Across 4 Seasons for Elk with all D.MJs 11:05 SUnday, Jrme 19, 1988 -------------- SPECIES=ElkYFAR=1986 SEX=F------TABlE SEASON AGE Frequency Cell Chi-Square Cal RCM Pet OF SEASON BYAGE IYea 12 y Total 13 Y 2 42 4.2613 19.53 . ~29· 31 0.7104 0.3415 13.49 14.42 26 1.6092 ·12.09 215 3 25 3.5789 9.77 28 0.2868 10.94 48 0.5967 18.75 47 1.3515 18.36 256 67 57 79 73 471 Total (Continued) TABLE SEASON AGE Frequency Cell Chi-Square Row Pet 4 Y OF SEASON BYAGE 15-8 Total 1>8 2 24 0.4292 11.16 44 0.0121 20.47 19 0.787 8.84 215 3 22 0.3605 8.59 54 0.0101 21.09 32 0.6609 12.50 256 46 98 51 471 Total STATISTICSFOR TABLE OF SEASON BY AGE statistic Chi-Square Likelihcxxl Ratio au-Square Mantel-Haenszel Chi-Square !hi Coefficient Contingency Coefficient Cramer's V Sample Size = 471 DF Value Prob 6 6 14.996 15.079 4.513 0.178 0.176 0.178 0.020 0.020 0.034 1 �329 Corrparison of Age Distribution .Across Seasons Comparison.Across 4 Seasons for Elk with all Di\Us 11:05 SUnday, June 19, 1988 -------- SPECIES=Elk YFAR=1986 SEX=M ------------TABIE OF SEASON BY AGE SEASON AGE Frequency Cell Chi-Square Row Pet Yea 1>8 12 y 14 Y 15-8 13 Y -----0 6 20 14 17 1 33 7.9007 105E-7 0.1848 1.5411 4.287 0.559 22.22 15.56 18.89 0.00 6.67 36.67 -----0 28 14 14 2 20 49 0.9447 0.2225 0.3034 442E-7 0.0092 0.7764 16.00 22.40 11.20 11.20 0.00 39.20 Total 90 125 --I- 3 38 13.165 35.51 36 0.263 33.64 17 0.9794 15.89 8 1.3127 7.48 6 3.223 5.61 2 2.6833 1.87 64 118 65 36 37 2 ------ 107 -+ Total STATISTICS FOR TABIE OF SEASON BY AGE statistic DF Value Prob Ori -Square Likelihocxi Ratio Ori-Square Mantel-Haenszel Ori-Square Fbi Coefficient Contingency Coefficient Cramer's V 10 10 1 38.356 39.184 14.923 0.345 0.326 0.244 0.000 0.000 0.000 Sample Size = 322 322 �( . >:-. ~ . f' . _\; I.;' ,~.:/. .:~.', f'~'~ 0\ " r ., '. ....•. :, ". t.~:" •...~ ..·-···-''''····.,..····.....,·'' ••..•. ·~t ~-, ", e· ., (>~~. ~."~?~ ., 11.,0': .'",. -'-, ...;.- ,,, ~"""""'" .•...•••........ ~~ ".. .•.... .: F~·-} ~.~ ... .. ..... ~~ .....•.. 1:' ,~ , ,.'. (}. ". ... .... " () .~-•.. - f~'L " . :.)~:.:, '):,'.: - , - •• r . l..: :' ~ ••••".- •.. """t' . ,. 3f '.~. ••.~\,1 . .-~ ::.L·.< ~-- • r"" •...,.. ---- ....•..-~. , " . ,~, .•.. ·rr i .•.:' •... ," ..,-+- ,. ~~.~•••- .:. J.:}:L ((: ",:.·i~.:.5"\ll7C:.~' t-.\ !'.t·:1,t .,..11C;·'· c-: I. '~·:tj_-~·qu<:rre I l~ ..'•....• ".•... ,•• ' ...L... r ..! , ,. ~) . /;,f; ;.:' lO_C~;; '.; .. '/) _ i I. :~ ..;:: 0;.. "/8::~1. .. ..•...• ,... ",.0 . ,_,_" ...•.. ," -" ~-~7A~. .. ..•. ,_,,'", , ,4;;, ....~,.52 '" .-.~ ..-+. . .:.~ 77 ~ .J 1'.~.'"'{) ..f ),':'~ ~(;,;-) ::'..":'~,·~·~:·~.·~;.3:,:~:~_.: I<.-::/:.J,o , . 7S:J ..:i �331 ..,.jf._. :", .. Cc:atparisoh 'of ,Age Distribution AcrosS, SeiasOns. Coniparison Aci"oss '4. Seasons for Elk With al;l'-mus . ..' l··./..... 11:05 SUnday, Jm.te 1~! ~9~~, > • ,.., SPECIES=Elk YEAR=1987SEX::::M ----------:. TABlE OF SEASON BY AGE AGE SFASON Frequency L; _ Cell·.Chi -Square Row Pet ..,.. .; .(j_r ).~t. call ._ . . _..>--., _. "''' ..-.. , I- " : .J flO rl Yea _. • - :~~ 1 ';2,'; i' 12 .Y r! ~.i13 I Y ; Total 1 - 85 2 . .. . \ .. ,46 .' U' 3 ~..... .. o o 4 0.18 ,:.<~.OO ,0.3 0 . 0.16'. ,;-<~;b.oo \( '~~~.oo' ,) Total .'(~':c (Continued) s : •..•.•••.• .. ", "~~'~-,...•. . :, 150 t " • :-:~••.I -.. - ..• -:.~ " .- .' ..' c']-.I·" �332 Cc:IJ:rparisonof Age Distribution Across seasons Cc:IJ:rparisonAcross 4 seasons for Elk with all DAUs 11:05 SUnday, June 19, 1988 --------- SPECIES=ElkYFAR=1987 SEX=M -------------TABlE OF SEASON BY AGE AGE Frequency Cell Ori-Square RowPet 4 Y 15-8 Total 1>8 1 1 0.0333 5.56 1 0.1089 5.56 1 1.1378 5.56 18 2 8 0.9608 9.41 4 0.1059 4.71 1 0.2882 1.18 85 3 1 1.3928 2.17 1 0.3835 2.17 1 0.007 2.17 46 4 0 0.0667 0.00 0 0.04 0.00 0 0.02 0.00 10 6 3 Total STATISTICS Ori-Square Likelihood Ratio Ori-Square . Mantel-Haenszel Ori-Square Fbi Coefficient Contingency Coefficient Cramer's V Sample Size 150 FOR TABLE OF SEASON BY AGE statistic WARNING: 1 DF Value Frob 18 18 26.717 27.090 10.413 0.422 0.389 0.244 0.084 0.077 0.001 1 = 150 64% of the cells have expected counts less than 5. Ori-Square may not be a valid test. �333 CoIrparison of Age Distribution Across Seasons CoIrparison Across 3 Seasons for Elk with all DAUs 11:05 SUnday, June 19, 1988 ------- SPECIES=ElkYFAR=1986 SEX=F--------TAmE OF SEASON BYAGE SEASON AGE Frequency Cell Chi-Square Row Pet Cal 1Yea 12 y Total 13 Y 2 42 4.2613 19.53 29 0.3415 13.49 31 0.7104 14.42 26 1.6092 12.09 215 3 25 3.5789 9.77 28 0.2868 10.94 48 0.5967 18.75 47 1.3515 18.36 256 67 57 79 73 471 Total (Continued) TABIEOF SEASON BY AGE SEASON AGE Frequency Cell Chi-Square Row Pet 4 Y 15-8 Total 1>8 2 24 0.4292 11.16 44 0.0121 20.47 19 0.787 8.84 215 3 22 0.3605 8.59 54 0.0101 21.09 32 0.6609 12.50 256 46 98 51 471 Total STATIsrICS FOR TAmE OF SEASON BYAGE statistic Chi-Square Likelihocd Ratio Chi-Square Mantel-Haenszel Chi-Square Ihi Coefficient Contingency Coefficient Cramer's V Sample Size = 471 DF Value Frob 6 6 14.996 15.079 4.513 0.178 0.176 0.178 0.020 0.020 0.034 1 �334 Corrparison of Age Distribution Across Seasons Corrparison Across 3 Seasons for Elk with all Di\Us 11:05 SUnday, June 19, 1988 --------------- SPECIES=Elk YEAR=1986 SEX=M --------, TABlE OF SEASON BY AGE AGE Frequency Cell Chi-Square Yea Row Pet 13 Y 12 y 15-8 14 Y 1 Total 1>8 --t 1 6 7.9007 6.67 33 105E-7 36.67 20 0.1848 22.22 14 1.5411 15.56 17 4.287 18.89 2 20 0.9447 16.00 49 0.2225 39.20 28 0.3034 22.40 14 442E-7 11.20 14 0.0092 11.20 0 0.7764 0.00 3 38 13.165 35.51 36 0.263 33.64 17 0.9794 15.89 8 1.3127 7.48 6 3.223 5.61 2 2.6833 1.87 64 118 65 36 37 2 0 0.559 0.00 90 --t- 125 -t- 107 -t- Total STATISTICS FOR TABLE OF SEASON BY AGE statistic DF Value Frob Chi-Square Likelihood Ratio Chi-Square Mantel-Haenszel Chi-Square Fhi Coefficient Contingency Coefficient Cramer's V 10 10 1 38.356 39.184 14.923 0.345 0.326 0.244 0.000 0.000 0.000 Sample Size = 322 322 �335 canparison of Age Distribution Across Seasons canparison Across 3 Seasons for Elk with all I::lA.Us 11:05 SUnday, June 19, 1988 -------- SPECIES=ElkYFAR=1987 SEX=F---------------TABlEOF SEASON BY AGE SEASON AGE Frequency Cell ari-Square Row Pet cal 1Yea 12 y 'Ibtal 13 Y 2 61 0.0247 11.05 83 0.062 15.0~ 104 0.0007 18.84 87 0.0011 15.76 552 3 23 0.0705 11.92 26 0.1773 13.47 36 0.002 18.65 30 0.0032 15.54 193 84 109 140 117 745 'Ibtal (Continued) TABlEOF SEASONBYAGE SFASON Frequency Cell ari-Square Row Pet AGE 4 Y 15-8 'Ibtal 1>8 2 59 0.3172 10.69 110 0.1477 19.93 48 0.0544 8.70 552 3 15 0.9073 7.77 44 0.4223 22.80 19 0.1555 9.84 193 74 154 67 745 'Ibtal STATISTICSFOR TABlEOF SEASONBYAGE statistic Clri-Square Likelihood Ratio ari-Square Mantel-Haenszel ari-Square Rri Coefficient Contingency Coefficient Cramer's V Sample Size = 745 DF Value Prob 6 6 2.346 2.399 0.264 0.056 0.056 0.056 0.885 0.880 0.608 1 �336 Comparison of Age Distribution Across Seasons Comparison Across 3 Seasons for Elk with all I:li\Us 11:05 Sunday, JlU1e 19, 1988 SPECIES=ElkYFAR=1987 SEX=M TABlEOF SEASON BYAGE AGE SEASON Frequency Cell Chl-Square ReM Pet Cal Total 1Yea 12 y 13 Y 1 1 2.3508 5.56 1 1.5683 5.56 8 1.2091 44.44 5 1.522 27.78 18 2 18 0.1005 21.18 11 1.2585 12.94 29 0.4317 34.12 14 0.007 16.47 85 3 15 1.9321 32.61 15 5.3283 32.61 8 2.4994 17.39 5 0.7835 10.87 46 45 24 149 Total (Continued) Frequency Cell Chl-Square ReM Pet 4 Y 34 27. 15-8 Total 1>8 1 1 0.0358 5.56 1 0.1045 5.56 1 1.1217 5.56 18 2 8 0.9235 9.41 4 0.0973 4.71 1 0.2957 1.18 85 3 1 1.4112 2.17 1 0.3922 2.17 1 0.0059 2.17 46 10 6 3 149 Total statistic STATISTICSFORTABlEOF SEASON BYAGE DF Value Prob Chl-Square 23.379 12 0.025 Likelihood Ratio Chl-Square 12 24.158 0.019 0.002 Mantel-Haenszel .Chl-square 1 9.321 0.396 Fhi Coefficient Contingency Coefficient 0.368 0.280 cramer's V Sample Size = 149 WARNING: 52% of the cells have expected counts less than 5. Chl-Square may not be a valid test. �337 NOTE: Copyright(c) 1985,86,87 SAS Institute Inc., Cary, NC 27512-8000, U.S.A. NOTE: SAS (r) Proprietary Software Release 6.03 Licensed to COLORADO STATE UNIVERSITY, Site 09521001. NOTE: OM statements are only valid in OMS ~ode. NOTE: AUTOEXEC processing completed. 1 2 3 4 5 title 'Comparison of Age Distribution Across Seasons'; libname library' .'; proc dbf db3=tooth out=library.tooth; 6 proc format library=library; NOTE: 2561 observations written to the output SAS data set. NOTE: The PROCEDURE DBF used 20.00 seconds. 7 value elkfmt O='Calf' 8 1='Yearling' 9 2='2 years' 10 3='3 years' 11 4='4 years' 12 5-8='5-8 years' 13 9-high='>8 years'; NOTE: Format ELKFMT has been output. 14 value deerfmt O='Fawn' 15 1='Yearling' 16 2='2 years' 17 3='3 years' 18 4='4 years' 19 5-high='>4 years'; ~OTE: Format DEERFMT has been output. 20 21 data valid; NOTE: The PROCEDURE FORMAT used 14.00 seconds. 22 set library.tooth; 23 if sexA=' '; 24 if seasonA=' '; 25 if sex='F' & season='I' then delete; 26 27 proc sort; ~OTE: The data set WORK. VALID has 2405 observations and 7 variables. 10TE: The DATA statement used 30.00 seconds. 28 by dau gmu year sex season~ 29 30 proc means maxdec=O min max; :OTE: The data set WORK. VALID has 2405 observations and 7 variables. 10TE: The PROCEDURE SORT used 25.00 seconds. 31 class dau gmu year sex season; 32 var age; 33 title2 'Number of Animals by GMU'; 34 35 proc sort; :OTE: The PROCEDURE MEANS used 1.03 minutes. 36 by dau year sex season; 37 38 proc means maxdec=O min max; OTE: The data set WORK. VALID has 2405 observations and 7 variables. 'OTE: The PROCEDURE SORT used 25.00 seconds. 39 class dau year sex season; 40 var age; 41 title2 'Number of Animals by DAU'; �3.38 42 43 proc sort; NOTE: The PROCEDURE MEANS used 48.00 seconds. 44 by species year sex season; 45 . 46 proc freq; NOTE: The data set WORK. VALID has 2405 observations and 7 variables. NOTE: The PROCEDURE SORT used 25.00 seconds. 47 by species year; 48 tables age; 49 tit1e2 'Distribution of Animals by Age'; 50 51 proc sort; NOTE: The PROCEDURE FREQ used 36.00 seconds. 52 by dau year sex season; 53 54 proc freq; NOTE: The data set WORK. VALID has 2405 observations and 7 variables. NOTE: The PROCEDURE SORT used 25.00 seconds. 55 where dau='E-6' & year='1987' & sex='M' & seasonA='4'; 56 tables season*age / chisq ce11chi2 nopercent noco1; 57 format age e1kfmt.; 58 tit1e2 'Comparison Across 3 Seasons for Elk in DAU E-6, 1987, Males' ; 59 60 proc freq; NOTE: The PROCEDURE FREQ used 36.00 seconds. 61 where dau='E-6' & year='1987' & sex='F' & seasonA='4'; 62 tables season*age / chisq ce11chi2 nopercent noco1; 63 format age e1kfmt.; 64 title2 'Comparison Across 3 Seasons for Elk in DAU E-6, 1987, Females'; 65 66 data compare; ~OTE: The PROCEDURE FREQ used 37.00 seconds. 67 set valid; 68 by dau year sex; 69 file 'temp.sas'; 70 retain nseason nrecs; 71 drop nseason nrecs; 72 if first. sex then do; 73 nseason=O; 74 nrecs=O; 75 end; 76 nrecs=nrecs+l; 77 if lag(season) A= season then nseason=nseason+1; 78 if last.sex then do; 79 if nrecs >= nseason*5 & nseason > 1 then do; 80 put "proc freq;"; 81 put" where dau='" dau ", & year='" year ", & sex='" .ex"';"; 82 if nrecs > 40 then 83 put" tables season*age / chisq ce1lchi2 nopercent oco 1 ; " ; 84 else 85 put " tables season*age / chisq exact ce11chi2 .opercent noco 1 ; " ; 86 if species="Deer" then 87 put" format age deerfmt.;"; 88 else put" format age e1kfmt.;"; �339 put" title2 'Comparison for DAU=" dau " & year=" year" 89 & sex=" sex"';"; 90 end; 91 end; 92 run; NOtE: The file 'temp.sas' is file C:\CDOW\TOOTH\TEMP.SAS. NOTE: 170 records were written to the file C:\CDOW\TOOTH\TEMP.SAS. The minimum record length was 10. The maximum record length was 60. NOTE: The data set WORK.COMPARE has 2405 observations and 7 variables. NOTE: The DATA statement used 42.00 seconds. 93 94 %include 'temp.sas'; NOTE: The PROCEDURE FREQ used 3S.00 seconds. NOTE: The PROCEDURE FREQ used 33.00 seconds. NOTE: The PROCEDURE FREQ used 34.00 seconds. NOTE: The PROCEDURE FREQ used 3S.00 seconds. NOTE: The PROCEDURE FREQ used 3S.00 seconds. NOTE: The PROCEDURE FREQ used 33.00 seconds. NOTE: The PROCEDURE FREQ used 34.00 seconds. NOTE: The PROCEDURE FREQ used 33.00 seconds. NOTE: The PROCEDURE FREQ used 36.00 seconds. NOTE: The PROCEDURE FREQ used 34.00 seconds. NOTE: The PROCEDURE FREQ used 35.00 seconds. NOTE: The PROCEDURE FREQ used 3S.00 seconds. NOTE: The PROCEDURE FREQ used 34.00 seconds. NOTE: The PROCEDURE FREQ used SO.OO seconds. NOTE: The PROCEDURE FREQ used 44.00 seconds. NOTE: The PROCEDURE FREQ used 34.00 seconds. NOTE: The PROCEDURE FREQ used 36.00 seconds. NOTE: The PROCEDURE FREQ used 3S.00 seconds. NOTE: The PROCEDURE FREQ used 34.00 seconds. NOTE: The PROCEDURE FREQ used 33.00 seconds. NOTE: The PROCEDURE FREQ used 36.00 seconds. NOTE: The PROCEDURE FREQ used 34.00 seconds. ~OTE: The PROCEDURE FREQ used 34.00 seconds. ~OTE: The PROCEDURE FREQ used 37.00 seconds. ~OTE: The PROCEDURE FREQ used 35.00 seconds. ~OTE: The PROCEDURE FREQ used 42.00 seconds. ~OTE: The PROCEDURE FREQ used 36.00 seconds. 'IOTE:The PROCEDURE FREQ used 41.00 seconds. 10TE: The PROCEDURE FREQ used 33.00 seconds. 10TE: The PROCEDURE FREQ used 34.00 seconds. lOTE: The PROCEDURE FREQ used 34.00 seconds. 10TE: The PROCEDURE FREQ used 33.00 seconds. 10TE: The PROCEDURE FREQ used 36.00 seconds. 265 266 proc sort; :OTE: The PROCEDURE FREQ used 35.00 seconds. 267 by species year sex season; 268 269 proc freq; ;OTE: The data set WORK.COMPARE has 2405 observations and 7 variables. :OTE: The PROCEDURE SORT used 27.00 seconds. 270 where spec;es='Deer'; 271 by species year sex; 272 tables season*age / chisq cellchi2 nopercent nocol; 273 format age deerfmt.; 274 title2 'Comparison Across Seasons for Deer with all DAUs'; 275 �340 276 NOTE: 277 278 279 280 281 282 283 NOTE: 284 285 286 287 288 289 290 NOTE: NOTE: proc freq; The PROCEDURE FREQ used 52.00 seconds. where species='Elk '; by species year sex; tables season*age I chisq cellchi2 nopercent nocol; format age elkfmt.; title2 'Comparison Across 4 Seasons for Elk with all DAUs'; proc freq; The PROCEDURE FREQ used 1.05 minutes. where species='Elk ' & seasonA='4'; by species year sex; tables season*age I chisq cellchi2 nopercent nocol; format age elkfmt.; title2 'Comparison Across 3 Seasons for Elk with all DAUs'; run; The PROCEDURE FREQ used 1.05 minutes. SAS Institute Inc., SAS Circle, PO Box 8000, Cary, NC 27512-8000 �341 I .j I M E M 0 RAN DUM June 21, 1988 To: Distribution From: Gary C. White C~-<...-JI Re: Variance of Age and Sex Ratios Introduction Bowden et al. (1984) proposed th~t age and sex ratios for mule deer should be based on a sample taken from the population of groups of deer. The estimator they proposed is the same as that for a sample of the population of individuals. However, the variance of the estim~te is considerably different. Bowden et al. (1984:504) compare the estimators for 10 samples, and conclude that the standard error of the proportion of fawns to fawns plus does is about 70% too large for the estimator based on individuals, but that the standard error of the proportion of bucks to bucks plus does is about the same for both the group and individual estimators. The conclusions of Bowden et al. (1984) are based on strictly theoretical grounds, because the true variance of the estimates is unknown. The comparison of the 2 competing estimators does not include a comparison to a known variance. Although they conclude that the estimator based on groups should be used, they present no evidence that the smaller variance froOo this estimator is not too small, and that confidence intervals calculated with this variance estimate will perfo~m as stated. I have populations performance relative to simulated the 2 competing estimators with 6 empirical where the true variance can be determined empirically, and thus of the 2 competing variance estimators can be determined the empirical variance of the point estimator. Note that this analysis does not examine the question of whethe~ aerial surveys performed by CDOW biologist provide an unbiased estimate of the proportion of the age and sex classes. That is, if the population is not sampled in a representative fashion, estimates will be biased. What this analysis examines is, given the observed sample from a sex and age ratio survey, which of 2 competing estimators of the variance of the point estimates should be used to compute confidence intervals. Methods Age and sex ratios are difficult to work with because a ratio may be distributed from 0 to a limit of infinity, i.e., if no adult females are observed, then, the number of young divided by the number of. females is �•L , 342 Variance of Age and Sex Ratios -- June 22, 1988 undefined. quant ity Thus, as doe~ Bowden et al. (1984), I will consider the Ey = 'jj (Y + f), where Ey is the proportion of young to young plus females, I is the number of adult females, and Y is the number of young in the sample. Likewise for adult males, EM=M/U1+f), where Ey is the proportion of males to males plus females, and M is the number of males in the sample. As noted by Bowden et al. (1984), the ratio of young to females (Ry) is just Ey/(1 - Ey), so the ratio can be easily constructed from the proportion, as well as a confidence interval. To assess each of the variance estimators, a sample of groups of animals was taken as an empirical population. That is, the true population is taken as an observed sample of actual animals. Then, this population is sampled repeatedly to form estimates of Ey and EM. The mean of the estimates of Ey and EM provides an estimate of the expected values of Ey and EM' Further, the standard deviation calculated for the estimates of Ey and EM provides an empirical estimate of the true standard error for a single estimate. Four samples of deer and 2 of elk were used to compare the 2 estimators (groups and individuals) of the standard errors of Ey and EM' The 6 samples are described in Table 1. Table. 1. Description of the 6 data sets used as a known population from which samples were taken to estimate Ey and EM. Mean Group Size Group Size Range Sma 11 Large Species Area Sampled Year Number Groups Deer Piceance Basin 1976 567 3.3 1 20.:' :0.376 0.163 Deer Piceance Basin 1977 815 4.2 1 25 0.406 0.149 Deer Gt·1Us52, 62, 1986 63, 64, and 70. 816 9.1 1 65 0.322 0.131 Deer GNUs 52, 62, 1987 63, 64, and 70. 958 8.6 1 66 0.355 0.164 Elk GNUs 54, 64, 1986 66, 70, and 521. 135 35.4 1 428 0.307 0.097 Elk GNUs 54, 64, 1987 66, 70, and 521. 448 20.5 1 410 0.317 0.146 Ey �343 Variance of Age and Sex Ratios -- June 22, 1988 For the Piceance Basin samples, males were classified as either immature or mature males, so that EM could be further subdivided. For each population, 10Q_0 replicate estimates of Ey and EM ~ere made. For each replicate, 100 groups were drawn from the population at :~nd_g!", with replacement. Results The estimates of Ey and EM were not biased (E > 0.50) for any of the deer populations, but were biased high (E < 0.00l) for both Ey and EM for both elk populations. This can be seen by comparing the true values of £y and EM given in Table 1 with the average of the 1000 estimates of Ey and EM given in Table 2. The relative bias of £y is calculated as 100(~y - £y}/£y and likewise for EM. The relative bias of £y and EM is less than 1% for all 4 deer populations. However, the relative bias of Ey is 7% and 5% for the 2 elk populations, and for EM, relative bias is 19% and 12% for the 2 elk populations. Thus, a significant bias of the point estimator occurs for the large group sizes encountered in elk populations. Table 2. Results of 1000 simulations for each of 6 populations, where 2 estimators of the standard errors of ~y and ~ are compared. AVE is the average of the 1000 replications, SO is the standard deviation for the 1000 replications. Species Year A~E _y A~E -M SD(Ey) Individuals AV~ AV~ SD(_t~) SE(_y) SE(m} GrouQs AV~ AV~ SE(_y) SE(.rH) Deer 1976 0.375 0.162 0.0231 0.0280 0.0_?82 0.0248 0.0232 0.0273 Deer 1977 0.406 0.149 0.0199 0.0243 0.0253 0.0218 0.0198 0.0238 Deer 1986 0.322 0.131 0.0147 0.0166 0.0163 0.0133 0.0144 0.0171 Deer 1987 0.355 0.163 0.0150 0.0194 0.0172 0.0151 0.0147 0.0194 E1k .. 1986 0.329 0.115 0.0212 0.0165 0.0088 0.0069 0.0206 0.0160 E1k 1987 0.332 0.163 0.0238 0.0219 0.0116 0.0101 0.0210 0.0208 The variance estimator based on groups performed much better than the estimator based on individuals! partic~lar..1.Y(f9~elk (Table 2). T~e. performance of each of the est1mators 1S made by comparing the emp1r1cal estimate of the standard deviation (i.e., SD(Py» with the average of the 1000 estimated standard errors (i.e., AVE SE(~y» for each of the two estimators (individuals and groups). The square root of the-variance �.. 344 Vari ance of Age and Sex Ratios estimate gives standard error estimate of the estimate. This sizes cause the sma11, whereas to the empirical error based on values for elk deviation. - - June 22, E: ; the standard error for each tr : '.1_ , cr: for groups is generally much (' ",:,;;." ",) standard deviation, and thus .·:":'~\'·.~f., ; is particularly true for the 0.t u:tG. vari ance estimate based on i r ~ \') ,:IU , : ::; I: ' the average standard error fo: l,'n! p..; ~ ,; ., standard deviation. However even ~.~ groups appears biased low, as: 4 averl~~ are less than their corr-espon. ':I':J c,K;j{ !._c.i "i_, r (.0 The final conclusions about the use ful nr. ;)'1 fcC;', (:I ,. \, estimators comes down to how good the conf idt c,' ';n;~u·· i.:~·~ ..•: •. each of them. Results are presented in Table ~, -rot (t" [; 1';,:,;1' where a 95% confidence interval is computed t-l';·~·~~.~} ·;.i~...l"' ...··.~L.~. -t';' groups variance estimators. .' ;,\ . Tabl e 3. Resul ts for confidence i nterva 1 cov ,-i'~:' "t :,::YJ each of 6 populations. A 95% confidence tntev-: ... L,;:,.,;. estimate. The interval either is entirely a ',:\;,! (1;;;.<' ~Ti; t:.:: entirely below (Be.) the true value, or else (,),(""<; C·",.' value. ,< .: .. Individuals ~~ v- '.1 '. ,_;'.': Species ______ Year Deer 1976 0.7 1. 2 98.1 2.7 6.3 91.0 Deer 1977 0.8 0.7 98.5 3.1 5.8 91.1 Deer 1986 1.1 1. 9 97.0 5.1 7.3 87.6 Deer 1987 1.3 0.9 97.8 4.7 8.S 86.8 E1 k 1986 59.0 3.8 37.2 58.7 1.7 39.6 f" E1 k 1987 34.8 3.8 61.4 43.5 3.8 :~. I ,I -.Y·,(_ --:--_'_~Y-- --~PM~-- . :...:.Ab=-.=--..=:B.:::..e..:.,.. _,C::..::o:....:,v_,_. _-,-A.:..::b~. ____;B=e::....:.:.......,,:.C.:::..ov.!-:-. _,.. r '.,._ ..J~.;.~ ... : ;:,\i . --------------------_ " ....'. \ ...• J. .:., " I ';, ,-' ~. " • ~. t. ' .. ,.' . " ~ } } i ., ..\ 52.7 ...•...... . .'} . < , . .: " ,~ .z -.... ----. _ •. _ •....•...•... _. If the estimator of the parameter is unb i.» :~d (::;1";01.:)/ n + "i:~.. 1 case for e 1k}, and if the est i mated standard eq . ( .::; COd'", '> .o- ;, true for the estimator based on groups), then ';;,:: ;'0(. fi d: »c. " should cover the true parameter value 95% of U ' L;;:f~•. i,;~ -.~" ~'," '!, the confidence interval for the standard error ;_.;~;e(; if'! ~. l' .. p: G I~,; relatively well for deer, i.e., the lowest COVL;:~:W \t,'·~('·.1: \~,:",. ;~' ';~~ 1976 deer data, where 92.9% of the computed cor;'; ..';~~p 11','.": <'r j' 'j~r" the true parameter value. I consider this lew, "{::",C", >.? ..:; :~ f ',I' . good. In contrast, for deer, the coverage for ':".~~: ~.:·~~:·:·f"_-':I;,.:)~ ~)(.~(;:~ ··.'r· :~r ~~~!)i1~~1!y~e~~~ ~~ ~~ot~~r~~ehf~~'~: t~~rc~~;,J';:i.::: ~,'~,~~~~'".'~:~;, ;~;.:..;:(:Ugfj �345 Variance I o( Age ~nd Sex R~tios -- June 22, 1988 individuals is verypoor , giving only 61.4% coverage at best for ~y. Th'i estimator t>~sedon groups~does much better, with the worst coverage for £y el k.;in- }986~' Much of thi s - 1ack of coverage is due to the bias of the poi nt est tmator , because most of the confidence intervals missed the true paramete~, by being above' tt , - .: The fnl l ow'ing is a' summary graph demonstrating the effects of an increase inmean qroup .s ize on both bias and precision. The increase in bi as is shown, for the 1ar-.ger group sizes found for elk. Further, the breakdown of the estimator of the standard error based on individuals is also clearly demonstrated for the 2 elk data points. '. " Relative Bias Males . .' , , Mean .' Group Size - '\ SD Young SD Males Ind. SE Young Ind. SE Males Grp. SE Young Grp. SE Males o~~~--~------~--~~~~--3.3 4.2 8.6 9.1 20.5 35.4 "1 ' t. Mean Group Size �346 Variance of Age and Sex Ratios -- June 22, 1988 Discussion One problem with the analysis presented here is that populations with larger mean group sizes also had the smaller numbers of observations. The correlation between mean group size and number of groups in the population is 0.83, so that the 2 effects are confounded. Additional simulations should be done where the larger populations (all deer) are subsampled to be sure that the observed changes in bias and precision as a function of group size are due to group size, and not due to a problem of sampling a small population. Conclusions Results of this analysis suggest that for deer, use of standard errors based on the binomial estimator to ~ompute confidence intervals would not be particularly unsatisfactory. However, this estimator is totally unacceptable for elk data. Old data sets existing in the files for deer can still be used to compute an approximate confidence interval, but such an approach is not useful for elk. Future work should incorporate the additional information available from recording group sizes. Recommendations Based on this analysis, I suggest that biologists collecting age and sex data should record the data by groups for both deer and elk, so that confidence intervals based on groups can be computed. I will modify the population database that I have been developing to incorporate this information, and the software will provide confidence intervals on age and sex ratios based on the standard error computed from groups when this information is available. The advantages gained for deer are not as great as far elk. but I believe that a standardized procedure for both species offsets the effort required for deer. Literature Cited Bowden, D. C., A. E. Anderson, and D. E. Medin. 1984. Sampling plans for mule deer sex and age ratios. J. Wildl. Manage. 48:500-509. Distribution: ~ Carpenter Bruce Gill Jim Lipscomb Dave Freddy Dick Bartmann Dave Bowden " ; . � https://cpw.cvlcollections.org/files/original/e331294ea03e68cc875af71cb23b5b01.pdf 03c56be0cffbd6d754dfcf202b5dff3e PDF Text Text 1 Colorado Division of Wildlife Wildlife Research Report December 1988 JOB PROGRESS REPORT State of __~C~o~l~o~r~a~d~o~ Project 01-03-212 (W-88-R) Work Plan Job Title: _1_ Avian Research - Migratory Birds 14 Ecological studies of the flightless period of ducks in Colorado Period Covered: Author: : Job _ 01 January 1987 through 31 March 1988 James K. Ringelman Personnel: M. Szymczak, Colorado Division of Wildlife ABSTRACT No field work or extensive data analyses were conducted during this reporting period. Plans for next segment include comprehensive data analyses, completion of a final report and preparation of manuscripts. ��3 Colorado Division of Wildlife Wildlife Research Report December 1988 JOB PROGRESS REPORT State of __~C~o~l~o~r~a~d~o~ Project 01-03-212 (W-88-R) Work Plan Job Title: 1 Avian Research - Migratory Birds 15 Development and use of a physiological condition index for monitoring wintering mallard nutrient reserves Period Covered: Author: : Job _ 01 January 1987 through 31 March 1988 James K. Ringelman Personnel: M. Szymczak, Colorado Division of Wildlife ABSTRACT No field work or extensive data analyses were conducted during this reporting period. Plans for next segment include comprehensive data analyses, completion of a final report and preparation of manuscripts. ��Colorado Division of Wildlife Wildlife Research Report December 1988 JOB PROGRESS REPORT State of __~C~o~l~o~r~a~d~o~ Project 01-03-212 (W-88-R) Work Plan 1 Job Title: Avian Research - Migratory Birds 16 Field feeding ecology of mallard ducks Period Covered: Author: : Job _ 01 January 1987 through 31 March 1988 James K. Ringelman Personnel: M. Szymczak; Colorado Division of Wildlife; D. Chapman, W. Dean, Colorado State University. ABSTRACT Feeding trials using pen-reared mallards (Anas platychynchos) were conducted in standing corn stubble containing whole ears only from 30 January to 18 February 1987. Mallard feeding rates on corn kernals varied by corn density, which ranged from 50 to 742 kg/ha. Feeding rates also differed among individual birds across all corn densities. Corn consumption rates on whole ears were similar to feeding rates on kernals at low (100-200 kg/ha) densities, but appreciably lower than feeding rates on kernals at higher densities. Local enhancement, wherein social interactions c~use .birds to observe and join successful foragers, minimized variatiorl among individuals and across densities. ��7 FIELD FEEDING ECOLOGY OF MALLARD DUCKS James K. Ringelman Man has drastically altered waterfowl habitat during the last 100 years. Widespread drainage of inland wetlands and alteration of coastal areas has had detrimental effects on waterfowl, but water development projects and irrigated agriculture have created new wintering and migration habitat. In Colorado, relatively few waterfowl were present in winter until irrigated cereal grain crops were introduced in the mid-1800's (Steinel 1926:174). Within a decade, waterfowl were being "short-stopped" along their traditional migratory routes, attracted by newly-created reservoirs and abundant waste cereal grain (Buller 1975). Wintering waterfowl, particularly the mallard, became so plentiful in the 1940's that a special season was held on mallards in Colorado during 1942-43 to help alleviate damage to grain crops (Wagar 1946). Mallards remain the most common wintering duck species today and accounted for 65% of the total duck harvest in 1981 (Colo. Div. Wildl. 1981). Wintering mallards, like all inland wintering waterfowl populations, are highly dependent upon waste products of agriculture for food (Girard 1941, Reed 1971, Sugden et al. 1974). In Colorado and adjacent states to the east and south, waste corn comprises from 50 to 90% of the winter diet (Gordon 1981, Jorde 1981, Baldeassaree and Bolen 1984). Despite the importance of corn to waterfowl, little is known of the factors tha~ regulate its use by ducks. Harvester efficiency, slope of the land, and corn moisture content all affect the amount of corn remaining in the field after harvest (Baldassarre et al. 1983). This "wastage", typically 3-5% of the pre-harvest corn crop, contributes the bulk of cereal grains necessary to sustain waterfowl throughout the winter. The amount of this wastage used by ducks varies depending on post-harvest treatments including discing, plowing, burning, and cattle grazing (Jorde 1981, Baldassarre and Bolen 1984), Economic forecasters predict radical changes in eastern Colorado irrigated agriculture within the next 7-17 years. Four factors will play important roles in this transformation: (1) depletion of ground water aquifers, (2) rising energy costs and related impacts on the cost of operating pumps, (3) conversion of farmland to urban use, especially along the Front Range, and (4) corn prices. Nearly 13% of all Colorado cropland is planted to irrigated corn and it ranks second behind wheat in crop value (Colo. Dep. Agric. 1982), However, under scenarios of rising energy costs and groundwater depletion, corn is the first irrigated crop to become economically unfeasible (Sharp 1979). The break-even point is when electricity to power pumps reaches $0.12fkwh (Oamek 1981:64), and some scenarios predict the disappearance of irrigated corn from the region overlying the Ogallala Aquifer by 1990 (Young et al. 1982:129). Front Range area croplands, which benefit from gravity flow irrigation water transported by canals, are under increasing pressure from urban growth. In the last 2 decades, Colorado:has seen over 1.3 million acres converted from agricultural use (Colo. Dep. Agric. 1980). Much of this development occurs on irrigated land (U.S. Dep. Agric. 1978). The key to preserving wintering waterfowl and subsequent harvest opportunity is the management of water areas coupled with more intensive land management practices, particularly in relation to cornfields on Division of Wildlife and private lands. Based on historical evidence and knowledge of wintering waterfowl food habits, it is believed that loss of cornfields will �resul t in maj or shifts of birds out of an area. In the 'case of the Front Range, this would result in severe loss of hunting opportunity where demand is greatest. This would also be true elsewhere in eastern Colorado where hunting pressure and demand is somewhat less pronounced, but nevertheless important. Little is known of how mallards use cornfields and waste grain, so it is not possible to determine the area of cornfields needed to sustain a given mallard population. In the past, such information was unneeded because of abundant corn crops. However, with the expected declines in irrigated agriculture in the near future, these data will be necessary to maintain wintering waterfowl. Past field-feeding studies have focused on the timing of flights and the relationships between flight times and weather (Hochbaum 1955, Swinebroad 1956, Bossenmaier and Marshall 1958, Winner 1959, Gordon 1981, Jorde 1981), or on ways to alleviate crop damage by ducks (Gollop 1950, Hammond 1952, MacLennan 1973). Only recently have studies documented the potential value of post-harvest treatments in making fields more attractive to ducks. Jorde (1981:50) described a commensal relationship between field-feeding mallards and cattle which exposed corn in times of heavy snowfall. Baldassarre et al. (1983) reported that treatments such as burning of discing of fields with moderate amounts of waste corn made these fields more attractive to ducks than untreated cornfields of standing stubble with waste corn densities many times greater. This suggests an "optimal foraging" strategy of field-feeding ducks in which feeding rate is balanced with other considerations such as search time, handling time, and distance to field. A large body of information on optimal foraging in birds exists in the ecological literature, but thus far these findings have not been applied to the management of field-feeding ducks. Efforts must be made to evaluate methods of making waste corn more available to waterfowl, thereby optimizing feeding rates and winter physiological condition. P. N. OBJECTIVES 1. Identify the relationships between feeding rates of mallards and waste corn density, group foraging size, sex, and age of duck. 2. Relate post-harvest cornfield treatments to foraging efficiency. 3. Document the foraging habitat requirements of wintering mallards. 4. Develop a model for post-harvest management of cornfields for wintering mallards. SEGMENT OBJECTIVES 1. Contact local farmers to arrange a cornfield lease. ears from 1 ha of corn during late summer. Hand pick and remove 2. In fall 1986, "harvest" the cornfield plot using standard techniques leaving canes, leaf litter, and stubble in place. 3. During January-March 1987, quantify the feeding rates of mallards as a function of kernal density on whole ears in standing stubble. �9 4. Summarize and compare feeding rates of mallards by cornfield treatment and corn density. 5. Derive functional response curves of feeding rates by treatment and corn density. 6. Monitor use of cornfields by mallards radio-marked in conjunction with Work Plan 1, Job 17. Evaluate duck use in relation to post-harvest treatment, hunting disturbance, snowfall, and proximity to roost and refuge areas. 7. Compile and analyze data and prepare progress report. METHODS A cornfield located just east of Wellington, Colorado was selected for use in field-feeding experiments. Whole ears were hand-picked from a 8 x 500 m area during early September 1986. Corn was harvested in mid-September, and permanent plot markers were established shortly thereafter. An electric fence was erected to exclude cattle from the plots. Panels measuring 1.2 x 2.0 m were used to enclose square plots 7.3 m on a side. Whole ears of corn were randomly dispersed in the enclosure. Mallards, which were previously conditioned to eat corn from ears in mock feeding trial situations, were deprived of food 24 hours before a trial. Eight birds of each sex were weighed to the nearest gram on an electronic digital balance immediately before each trial. Mallards were released gn ~ into the enclosure, allowed to feed for precisely 5 minutes, then rounded up and reweighed. The amount of corn ingested was assumed to equal the difference between starting and ending body mass. A BASIC computer program was written to estimate the weight of kernals on a whole ear of corn based on weight and linear measurements of the ear. This mathematical relationship was derived using stepwise, forward linear regression models that related the measured weight of kerna1s on an ear to linear combinations of ear length and weight prior to removing kernals. The following equation was used to estimate total kernal weight: K - E - (0.03643 where: K X * kernal weight (g), E ear length (cm) R (X * 3.1416 * R2) - 28.722) weight of the whole ear (with kernals) (g) (ear diameter at center (cm) + 5) / 2 The Pearson Correlation coefficient between estimated and true kernal weight was 0.94 on the sample of 10 ears used to establish this relationship. Several ears were used to achieve kernal densities of 50 to 742 kg/ha, which corresponds to the range of density values used in prior experiments with kernals. Since mallards depleted a significant proportion of available corn, particularly at lower initial densities, an "ending density" was calculated by subtracting the total corn consumption for the trial from the initial dosage. An "average density" was then computed as the mean of the initial and ending densities. Average density was used to test for �10 differences in feeding rates by density and treatment. Birds with feeding rates above the median were used to quantify maximum feeding rates. RESULTS Twelve feeding trials were conducted using whole ears in standing stubble. A wide variation in feeding rate within density, reflecting differential foraging success among individual birds, was apparent in every trial. In considering birds only above the median, variation was greatest at kernal densities less tha 400 kg/ha. In contrast, variation in feeding rates with kernals in standing stubble was highest at densities between 200 to SOO kg/ha. When averaged across all densities, feeding rates differed among birds (~ 0.05; Table 1). The feeding rate of the highest bird (9) was 79% greater than that of the lowest bird (5). Trends in individual feeding rates by sex were not apparent. Table 1. Feeding rate of individual mallards on whole ears of corn dispersed in standing stubble, averaged across all corn densities and trials. Bird 9 7 10 14 13 1 12 11 16 S 6 4 17 3 2 5 # Mean feeding rate (g/min) 5.95 5.63 5.54 5.53 5.50 5.1S 5.12 4.S3 4.65 4.51 4.40 4.24 4.23 4.14 3.S6 3.32 a a,b a,b a,b a,b a,b,c a,b,c a,b,c a,b,c a,b,c a,b,c a,b,c a,b,c a,b,c b,c c Sex M F M M M F M M M F F F M F F F a Means with the same letter do not differ (P>0.05). Feeding rates increased rapidly between densities of 0-200 kg/ha, but this rate of increase declined sharply above 300 kg/ha (Fig. 1). Asymptotic mean feeding rate was approximately 6.5 g kerna1s/minute on whole ears, although unexpectedly high feeding rates were achieved at densities ranging from 50-339 kg/ha (Table 2). �.il. -:E z <, 12 10 o w ~ « a: "za « a: 8 ....~~-------------------------- 0 D ~~~~ --------0-----------0---------- 6 0 4 0 LL 2 0 300 100 500 700 900 1100 MEAN CORN DENSITY (KG / HA) Fig. 1. Foraging rates of mallards on whole ears of corn. -z 12 o :E ........10 w ~ « a: o z a« a: ____ --------0 o e o 8 o o ------0------------------------------------------------6 4 o LL 2 o 100 300 500 700 900 MEAN CORN DENSITY (KG / HA) Fig. 2. Foraging rates of mallards on corn kernals. 1100 �12 Table 2. kerna1s. Foraging rates of mallards on whole ears of corn by density of Density (kg/ha) Significance8 Mean foraging rate (g/min) 50 62 67 83 109 147 158 335 339 367 535 742 6.92 4.66 4.23 4.29 4.07 6.54 5.20 7.34 6.64 4.41 6.80 6.61 a b b b b a· b a a b a a 8Means with the same letter do not differ (P>0.05). DISCUSSION Unlike loose kernals, wherein feeding rates continue to increase at densities above 300 kg/ha (Fig. 2), mallards did not increase their maximum feeding rates above this density (Fig. 1, Tables 2,3). These two feeding response curves quickly diverge at densities above 300 kg/ha, reaching asympotic feeding rates of 9.3 and 6.5 kg/ha with kernals and ears, respectively. Table 3. Estimated foraging rates of mallards by density of corn kernals and treatment (from previous experiments). Foraging rate (g/min) by treatment Density (kg/ha) 100 200 300 400 600 800 1000 1200 Bare ground 10.3 11.8 12.7 13.3 13.8 13.9 13.9 14.0 Kernals Disced stubble8 4.7 6.3 7.2 7.8 8.6 9.0 9.3 9.4 8 Corn kernals applied to sample plots prior to treatment. 0.1 0.1 0.3 0.7 1.7 2.0 5.0 �Optimal foraging theory (Pyke et al. 1977) can be helpful in understanding the behavior of feeding birds and makinz management decisions about post-harvest treatments. Handling time, which is an important factor in the selection of barley and wheat strains (Clark et al. 1986), was insignificant for corn kernals on dirt substrate and minimal for kernals in standing stubble, where search time was the primary time factor. With whole ears however, birds frequently struggled to remove kernals. Maximum feeding rates were apparently limited by both search and handling time of ears, thus accounting for differences observed among treatments. Feeding rate on whole ears was 30% lower at 800 kg/ha than the feeding rate on loose kernals at the same density, suggesting that about one-third again as much time is needed to forage on whole ears owing to increased handling time. In the range of normal waste corn densities (200 to 400 kg/ha, Baldessarre and Bolen 1984), this handling time resulted in a large decrease in feeding efficiency. Since mallards consume from 34 to 64 g of corn/feeding bout (Whyte and Bolen 1985) and kernals on ears comprise 70% of the total corn available (Baldassarre et al. 1983), minimum feeding time in a typical cornfield should be about 5-10 minutes. Assuming 2 field-feeding flights/day, total daily feeding time could be less than 20 minutes (not including flight time). Other factors such as flock size, which affect opportunities for social interactions and local enhancement for food-finding, may also interact with forging rates. Thus, under normal conditions, mallards on high plains wintering areas can afford to adopt a strategy of time minimization (Pyke et al. 1977), thereby conserving energy and affording time for courtship as well as other, less demanding behaviors. LITERATURE CITED Baldassarre, G. A., and E. G. Bolen. 1984. Field-feeding ecology of waterfowl wintering on the southern high plains of Texas. J. Wildl. Manage. 48:63-71. ____________________ , R. J. Whyte, E. E. Quinlan, and E. G. Bolen. 1983. Dynamics and quality of waste corn available to postbreeding waterfowl in Texas. Wildl. Soc. Bull. 11:25-31. Bossenmaier, E. F., and W. H. Marshall. 1958. Field-feeding by waterfowl in southwestern Manitoba. Wildl. Monogr. 1. 32 pp. Buller, R. J. 1975. Redistribution of waterfowl: influence of water, protection, and feed. Proc. Int. Waterfowl Symp. 1:143-154. Clark, R. G., H. Greenwood, and L. G. Sugden .. 1986. Influence of grain characteristics on optimal diet of field-feeding mallards. J. Appl. Ecol. 23:763-771. Colorado Department of Agriculture. 1980. Agricultural land conversion in Colorado. Resour. Analysis Sect., Colorado Dep. Agric., Denver. 8 pp. �i4 1982. Colorado agricultural statistics. Livestock Rep. Serv., Denver. Bull. 1·82. 94 pp. Colo. Crop and Colorado Division of Wildlife. 1981. Colorado small game, furbearer and varmint harvest, 1981. Colorado Div. Wi1d1., Denver. 228 pp. Girard, G. L. 1941. The mallard: Wildl. Manage. 5:233·259. its management in western Montana. J. Gollop, B. J. 1950. Report on investigation of damage to cereal crops by ducks in the prairie provinces. Unpubl. Rep., Can. Wildl. Serv., Ottawa. 11 pp. Gordon, D. H. 1981. Condition, feeding ecology, and behavior of mallards wintering in northcentral Oklahoma. M.S. Thesis, Oklahoma State Univ., Stillwater. 68 pp. Hammond, M. C. 1952. Waterfowl damage and control measures, Lower Souris Refuge and vicinity. Unpubl. Rep., U.S. Dep. Inter., Fish and Wildl. Serv., Washington, D.C. 5 pp. Hochbaum, H. A. 1955. Press, Minneapolis. Travels and traditions of waterfowl. 301 pp. Univ. Minnesota Jorde, D. G. 1981. Winter and spring staging ecology of mallards in southcentral Nebraska. M.S. Thesis, Univ. North Dakota, Grand Forks. 116 pp. MacLennan, R. 1973. A study of waterfowl crop depredation in Saskatchewan. Can. Wildl. Serv., Saskatoon, Saskatchewan. Wildl. Rep. 2. 38 pp. Oamek, G. E. 1981. Economic adjustments to rising energy costs: the case for pump irrigators. M.S. Thesis, Colorado State Univ., Fort Collins. 90 pp. Owen, R. B., Jr., and K. J. Reinecke. 1979. Bioenergetics of breeding dabbling ducks. Pages 71-93 in T. A. Bookhout, ed. Waterfowl and wetlands--an integrated review. La Cross Printing Co., La Crosse, WI. Pyke, G. H., H. R. Pulliam, and E. L. Charnov. 1977. Optimal foraging theory:a selective review of theory and tests. Quart. Rev. BioI. 52:137-154. Reed, L. W. 1971. Use of western Lake Erie by migratory and wintering waterfowl. M.S. Thesis, Michigan State Univ., East Lansing. 71 pp. Sharp, R. L. 1979. Economic adjustments to increasing energy costs for pump irrigation in northeastern Colorado. M.S. Thesis, Colorado State Univ., Fort Collins. 72 pp. Steinel, A. T. 1926. History of agriculture in Colorado. ColI., Fort Collins, Colo. 659 pp. State Agric. �Sugden, L. G., W. J. Thurlow, R. D. Harris, and K. Vermeer. 1974. Investigations of mallards overwintering at Calgary, Alberta. Can. Field-Nat. 88:303-311. Swinebroad, J. 1956. Some aspects of the role of weather in bird migration. Ph.D. Diss., Ohio State Univ., Columbus. 315 pp. u.S. Department of Agriculture. 1978. Urbanization of rural lands in the northern Colorado Front Range. U.S. Dep. Agric., Washington, D.C. 23 pp. Wagar, J. V. K. 1946. Colorado's duck-damage, grain-crop problem. North Am. Wi1d1. Conf. 11:156-162. Proc. Whyte, R. J., and E. G. Bolen. 1985. Corn consumption by wintering mallards during morning field-flights. Prairie Nat. 17:71-78. Winner, R. W. 1959. Field-feeding periodicity of black and mallard ducks. Wildl. Manage. 23:197-202. J Young, R. A., L. R. Conklin, R. A. Longenbaugh, and R. L. Gardner. 1982. Energy and water scarcity and the irrigated agricultural economy of the Colorado high plains: direct ecnomic and hydrologic impact forecasts (1979-2020). Colorado Water Resour. Res. Inst., Colorado State Univ., Fort Collins. 362 pp. Prepared bY~Z_~~ es K. Ringe m n Wildlife Researcher C ��17 Colorado Division of Yildlife Wildlife Research Report December 1988 JOB PROGRESS REPORT State of __~C~o~l~o~r~a~d~o~ Project York Plan 01-03-212 (Y-88-R) 1 Job _ Avian Research - Migratory Birds 17 Job Title: Habitat use by wintering mallards along the Front Range of Colorado Period Covered: 01 January 1987 through 31 March 1988 Authors: James K. Ringelman and Michael R. Szymczak Personnel: K. Ragotzkie, Colorado Division of Yildlife; J. Barnett, L. Benson, R. Yestfall, Colorado State University. ABSTRACT Twenty-five mallards were captured and radio-marked at Chestnut Slough, 10 km south of Greeley, Colorado, during early December 1986 and 1987. Movements and habitat use of marked and unmarked mallards were measured from December through February to determine relationships among hunting activity, weather and habitat selection. Focal animal time budget data of unmarked mallards were collected to aid in understanding habitat selection by instrumented ducks. Yinter temperatures and snowfall differed between years, thereby influencing habitat availability, usage, and duck behavior. Of the 367 wetlands (1,266 ha) on the study area, most were frozen and unavailable during winter. Home range sizes averaged 154 ha in 1986-87 and 120 ha in 1987-88. Yarm-water wetlands, lakes, and wetlands closed to hunting were key components in the home range' of mos t; birds •. Although they represented'<1% of the surface water on the study area, warm-water wetlands were used extensively by ducks and were highly selected during both winters. During .themild winter of 1986-87, portions of lakes and small ponds remained open and received greater use than in 1987-88 when these habitats were mostly frozen. Mallards used rivers, holding ponds and ditches extensively in 1987-88. Yhen cornfields became snow-covered for an extended period in 1987-88, ducks switched to feeding in and around cattle feedlots. Adult females used lakes more frequently than males, and adults made greater use of rivers than did immatures of either sex. Resting was the most. common behavior during both winters, followed by swimming and water feeding. Courtship displays were most common on reservoirs and warm-water wetlands, peaking during December and January. February behaviors included more resting but less water feeding than in January. Behaviors suggest that lakes and warm-water sloughs serve as courtship arenas during the period of pair formation. ��HABITAT USE BY WINTERING MALLARDS ALONG THE FRONT·RANGE OF COLORADO James K. Ringelman Michael R. Szymczak This report documents the accomplishments of two and 1987-88. Parenthetical ~eferences to years refers or results from a single field season. The absence of implies objectives, methods or findings common to both field seasons, 1986-87 to objectives, methods such a reference field periods. P. N. OBJECTIVES 1. Relate mallard habitat use to the physical characteristics of wetlands, aquatic and upland plant coIllIllUD1tiesr we1t.land . macro invertebrate populations, vea~eT. and bunting regimes. 2. Characterize mallard use of wetlands through time budget techniques. 3. Document the composition of the wetl~~ community within the study area, and relate wetland use by mallards to availability. SEGMENT OiJ.ECTIVES 1. Capture and radio-mark 25 mallards during early December. marked sample equally by age anm SEX. Divide the 2. Monitor the diurnal movements of ~t~ented mallards using aerial and vehicle-mounted tracking systems. <code bird locations using a Universal Tram;·'erse Mercator (UTM) grid sySte1n_ 3. Conduct aerial tracking flights once a week, or more frequently if weather conditions or bird movemen~ warran~~ 4. Classify all study wetlands ~y geneTal we~and type, then digitize information for ease of retrie7al by computer systems (1986-87). Obtain aerial photographs of study we~lamilisto aid in delineating physical characteristics of each. 5. Determine the availability of ice-free wetlands daily using a representative sample of wetland types (1987-88). 6. Monitor hunting regimes on key wetlands to determine the effect of hunting disturbance on movemen~s of Ta~io-marked mallards. 7. Obtain National Weather Service records for daily temperature, wind and snowfall for a representative site in the study area. 8. Measure the microhabitat temperatures at warm-water wetlands for comparison with ambient conditions. 9. Quantify behavior of marked and unmarked mallards in relation to wetland type, social status, and time of day using focal animal sampling. , �20 STIJPY AREA The 1,089 km2 study area is located along Colorado's Front Range near Greeley, Colorado. The area is rectangular in shape (36.2 km east to west, 30.5 km north to south) and is bordered on the west by Interstate Highway 25, on the north by State Highway 392, on the south by State Highway 66, and on the east along the UTM line 542000• Upland habitats are typical of the Colorado Piedmont (Fenneman 1931) portion of the Great Plains physiographic province. Elevations average about 1430 m, with generally flat to rolling terrain interrupted by occasional steep bluffs along major river systems. The study area is drained by 4 rivers: the Cache la Poudre, Big Thompson, St. Vrain, and South Platte. Climate is characteristically continental (Gittings 1941), with December, January, and February daily temperatures averaging -1.3, -3.9 and -0.4 C, respectively. Average precipitation during these same months totals 0.74, 0.76 and 0.71 cm. Fog is common during warm winter mornings and evenings. Native grasses consist of blue grama (Bouteloua gracilis) and buffalo grass (Buchloe dactyloides), with rubber rabbitbrush (Chrysothamnus nauseosus) and snakeweed (Gutierrezia sarothrae) the dominant shrubs. Cottonwood (Populus sargentii) , willows (Salix spp.), and green ash (Fraxinus pennsylvanica) are common along watercourses. Aquatic vegetation varies by wetland type and water regimes. Large wetlands used for irrigation water storage are mostly devoid of vegetation due to extreme water level fluctuations during the growing season. Small wetlands often contain submergent watermilfoil (Myriophyllum exalbescens), pondweeds (Potomogeton spp.), coontail (Ceratophyllum dernersum)and muskgrasses (Chara spp.). Smartweeds (Pologonum spp.), sedges (Carex spp.), cocklebur (Xanthium strumarium), and cattails (Typha spp.) occur around wetland margins and in shallowly flooded areas. Watercress (Nasturtium officionale), a common aquatic of warm-water sloughs, harbors high densities of pond snails (Physa spp.) and other macroinvertebrates consumed by mallards. Most land in the study area is devoted to irrigated agriculture or livestock husbandry. Corn, alfalfa, wheat and sugar beets are the dominant irrigated crops. Malting barley and vegetables are grown in lesser amounts. Several large cattle feedlot operations exist in the northern portion of the area. Winter waterfowl populations typically range from several thousand to >10,000 mallards (Anas platyrhynchos), and up to 10,000 Canada geese (Branta canadensis). Pintails (Anas acuta), green-winged teal (Anas crecca) and wigeon (Anas americana) are present in lesser numbers. METHODS Mallards were captured at Chestnut Slough on 3 December 1986 and 2 December 1987 using Salt Plains bait traps (Szymczak and Corey 1976). In each year, 25 mallards were instrumented with back-mounted radio transmitters (Dwyer 1972) weighing 24 g. Measurements were made of wing length and body weight, and a standard U.S. Fish and Wildlife Service band affixed to each bird prior to release. Truck-mounted, precision direction finding antennae arrays were used to locate birds from the ground. When a signal was detected, 2 or more azimuths were taken from a known location then plotted on 1:24,000 topographic maps to a resolution of 1 ha. A transparent grid overlay was used to code the �21 location of each bird in the standard UTM grid system. Concurrent with information on the bird's location, the date, time, activity status, air temperature, precipitation, wind speed, wetland name, and habitat type were also recorded. In 1987-88, a portable computer (Tandy Model 102) was used to tabulate and error check data in the field. Standard field forms were used in 1986-87. All data were entered into an IBM-PC for subsequent analyses. In 1986-87, mallards were located as frequently as possible, usually 2-4 times/day. Tracking schedules in 1987-88 alternated between (1) locating as many individuals as frequently as possible during the day, and (2) constant monitoring of 1 or 2 individual birds during a single day. Aerial tracking was conducted only when several marked birds could not be located after extensive ground searches. Flights were made over the entire study area and, if necessary, along the Front Range from Wellington south to Barr Lake, and from Boulder east to Jumbo Reservoir. Responses of marked birds to hunting disturbance were obtained by constant monitoring at Chestnut Slough. Personnel arrived at Chestnut prior to arrival of hunting parties in the morning, and remained at the slough until either mid-afternoon or dark, depending upon scheduling of other activities. Activities on Chestnut Slough were ca~egorized by disturbance factors in the following manner: no disturbance (factor 0), shooting audible from Chestnut but not originating on the slough (factor 1), shooting on Chestnut >200 m from the subject bird (factor 2), driving near the subject bird (factor 3), walking near the subject bird (factor 4), and shooting <200 m from the subject bird (factor 5). Continuous monitoring of birds and hunters on Chestnut was conducted during the first 3 days of the hunting season and periodically thereafter. Wetlands were categorized into 5 habitat types. Small ponds and reservoirs were defined as wetlands <40 ha in surface water area. Lakes were wetlands >40 ha in size. Holding and sewage ponds were artificial wetlands created for treatment of wastewater. Gravel pits were wetlands created as a byproduct of surface mining for aggregates, generally along river courses. Warm-water wetlands included warm-water sloughs, which are maintained by subterranean seeps, ditches that carry a warm water during the winter (usually as a result of reservoir discharges), and ponds that remain open as a result of warm water provided by either sloughs or ditches. Ditches were man-made systems built to convey water for agriculture. Rivers included natural watercourses in the study area. The availability of wetlands varied by type and local weather conditions. Consequently, in 1987-88 a sample of wetlands typical of habitat types on the study area were monitored for ice conditions at least once a day. Percent open water for each wetland was estimated and recorded along with the date, time, and air temperature. Temperature data were also obtained at Chestnut Slough and Union Ditch. At the latter location, maximum and minimum temperatures were obtained 5 cm above the surface and 15 cm under water. Time budget data were obtained using focal animal sampling (Altmann 1974) stratified by habitat type, time period, gender, and social status (paired or unpaired) of the subject bird. Habitat types were reservoirs, rivers, warm-water wetlands, and other (mostly gravel pits). Time periods were sunrise-0900, 0900-1200, 1200-1500, and l500-sunset. Spotting scopes (40X) or binoculars were used for all time budget collections. Focal animals were selected by pointing the scope or binocular at the flock, then selecting the closest individual of the desired gender and social status. Behaviors were recorded continuously (Tacha et al. 1985) on either a hand-held tape recorder or directly on a portable computer (Tandy Model 102). The following �22 behavior classifications were used: (0) out-of-sight, (1,) rest, (2) swim, (3) walk, (4) preen, (5) feed in water, (6) feed on land, (7) courtship, (8) aggression, and (9) other. Time budget bouts were 15 minutes unless circumstances prohibited collection of data for the entire period. A Fortran computer program was written to select subsets of time budget data and summarize overall behaviors. Statistical analyses were performed on an IBM-PC using the Statistical Analysis System (SAS). RESULTS Weather and Ice Conditions Temperature and snowfall had a profound influence on wetland habitat and food availability, thereby mediating movements and habitat use of radio-marked birds. The 1986-87 field season was typified by above average temperatures and below average snowfall. Brief periods of cold occurred in mid December and January, but daily high temperatures were usually above freezing throughout the winter (Fig. 1). The 1987-88 winter was more typical of those along the Front Range, with daily low temperatures dipping below 0 F on several occasions during late December through January (Fig. 2). Daily high temperatures ranged from 13 to 30 F during this same period. Compared to temperatures regimes at Greeley, Chestnut Slough experienced colder low temperatures but warmer high temperatures (Fig. 3). Greeley temperatures are more typical of those experienced throughout the study area. Snow on the ground was less during 1986-87 than during 1987-88, reflecting differences in daily temperatures and snowfall between years. In 1986-87, snow cover was absent during late December into early January, and again from late January into early February (Fig. 4). In contrast, continual snow cover existed from 22 December 1987 until 18 February 1988 (Fig. 5). During this latter period, cornfields remained snow covered throughout most of the winter. Wetland Habitats Low level aerial photographs were used in conjunction with field observations and 1:24,000 topographic maps to characterize the surface area and number of wetland habitat types. Wetlands were digitized on computer, then displayed using commercially available software (Freelance) and hardware (HP Laserjet Series II, Fig. 6). Total wetland area was 1,266 ha for the 367 wetlands enumerated on the study area. Small ponds and reservoirs were the most numerous and comprised the largest number within a classification (Table 1). Ditches and canals were the second most numerous type, but occupied only 7.8% of the surface water area. Rivers and streams were next in abundance, followed by gravel pits, lakes, and holding/sewage ponds, respectively. Warmwater wetlands were few (n-IS), comprising <1% of the total surface water area (Table 1). The number and distribution of ice-free wetlands varied during 1987-88, reflecting daily temperature regimes and snowfall (Fig. 7). Except for occasional open water areas or "ice holes", small ponds and reservoirs became unavailable after 12 December. Ditches and rivers had dynamic ice conditions, responding more quickly to warming trends (Fig. 7). Riverine habitats averaged about 60% open water "during winter. Most ditches containing open water during winter carried warm-water sources. Warm water wetlands, by definition, never froze. �60 -- 50 LL W . ~. "7 -{ ~ - 40 a: :J t- < a: w Q. 30 20 ~ W t- \ • -,',- . "" ,-, .. , ____ O 't • ,.,,'~," :~ ,""-~iV ~' \: ~.' I' I, " '. ' .•' , 10 I I ," ,, I-~ I' \:. , ',".," I' I' , I. - :-:,~, ,,: ""' , ! :'! '- ", I , . ' .: ~ " ",I II ,,' ... , I, I , I,· •• • . ,- I' I I I, I,' ,, ,,, ,'" ,',. ,. '" , • I ' ~~L If " f ", . .' _ -10 1 10 20 30 Dec. Fig. 1. Dally maximum Colorado, 1986-87. 20 10 30 10 Feb. DATE and minimum temperatures 20 at Greeley 60 -LL 50 W a: :J 40 < cr 30 tW Q. - - - _,.__. ." t- , .. I 10 ,, ,, " " 0 ,~, 1-' •, 'I.,.... ,, 20 ~ W • I' I " , .' :. , ~,i ' I' ,I,.', ' ~,- , ! "~' I 'If ~: .1., . .'" I'" .•. , r , ",' I" • I ,, ,, , ~:~ .: "~ : ''f: V, -----~--~~~-~~~--~~------~------" I, I ~' -10 , f • _ '" ',,' -I l' I' " I, I' I \ • I • ~, I 1 10 20 Dec. Fig. 2. Daily maximum Colorado, 1987-88. 30 10 20 30 20 Feb. DATE and minimum temperatures 10 at Greeley �24 60 -- 50 LL W CC 40 ::> 30 ed: CC W 20 f- a. ~ W ,. - ,,' 10 1'\' I f- -- "",. •• , " • , , .. .n , -20 11 20 Dec. 30 , , I • I '.'.. 10 . , . , I , tI • I, I . . ,, .', . .', ~J- ~"""P"---- ' ·· ' • ,' "~ 30 10 Feb. DATE Fig. 3. Ambient maximum and minimum air temperatures Chestnut Slough, 1987-88. , I I ••~ ,.., '. 20 .' I' , I " I, I', .&.,--~-'-~~ 'f ',' ~ .'~ I , , I -10 , It •• ' I' . ,.. . " " • I , , " " , I' 7,t......!---~--_+'r+ I r' I. 1,·1 0 I " ~ at 20 �- :i ~ 8 c Z :J o a: 6 ~ 4 CJ ~ o Z CIJ 2 o 1 10 20 30 10 20 30 10 DATE Dec. 20 28 Feb. Fig. 4. Snow on ground at Greeley, Colorado, 1986-87. -~ 8 c Z :J o a: 6 Z 4 CJ o ~ o z CIJ 2 o 1 10 20 Dec. 30 10 20 DATE Fig. 5. Snow on ground at Greeley, Colorado, 1987-88. 30 10 20 Feb. 28 �N 0\ Wlnd.or tJ o () Pond rr R ~,~~" o --- ~ - .._, ~ .C1 ~ ~ o· ~\ "-J ~ 0 Greeley o o /~-f' r::) -r IJ v D <$>0 \. ?'10~~ ••........ ~ "V ,,\.(.I.\\e~ 0'1 ~ ~ Q c:::? Latham Res. Milton Res. o Ch•• tnut Slough o o 10 km ~ o p~ 1// / / / / / / / / / / / / / /1 I) 1:0'0 U ~ c. o~ c. Plett.vlll. Fig. 6. Digitized wetland habitat map of the Greeley study area. �27 100 --, ,,' I'I ...•.. ,...•.• 1 I I I 80 ...•.. "'I ,, 60 I \ \ , I / / / / I I \ \ \ \I , 20 I I \ I ,- -- ---, , ,, ' \/ ,I II 40 ,.- \ \ / \ I I -"'\ \ l DITCHES / ,, / I I \ I \/ 0 100 80 c:c 60 r- <X 40 ~ 20 W Z W CL 0 r- Z W RIVERS 0 100 80 U c:c W 60 RESERVOIRS CL 40 .; .. .. 20 .......... "'. ....... 0 100 /\ ....... ·.::··· ••••••••••• 0 •••••••••••••••••••••••••••• 0 •••••••••••••••••••••••••• -- -, • 80 • • • 60 I • • • •• 40 • 20 ",'. I, I. . ',,- ••• ,' " PONDS ,,,,', ',' '", I :. ' t " ..." •.~• - .•- - - •.•. - - - - - •.- •.•. -: 0 10 20 30 10 , :~ I ·20 • !_ - - •.- - - - - - - - - - , ' ...- - •. : ~- - - - •.- 30 Dec. 20 Feb. DATE Fig. 7. Percent open water during 1987-88 10 for 4 wetland types. �28 Table 1. Surface area and number of wetlands by habitat type on the Greeley study area. Habitat type Area (ha) Small ponds/reservoirs , Area (ha) , 513 40.5 130 35.4 112 8.9 3 0.8 84 6.6 20 5.5 112 8.9 24 6.5 8 0.6 15 4.1 Ditches/canalsC 99 7.8 107 29.2 Rivers/streamsd 338 26.7 68 18.5 1,266 100.0 367 100.0 Holding/sewage ponds Gravel pits Warm-water wetlandsb Totals ·Wetlands >40 ha. bCalculated as 33 km total length x 2.5 m width. cCalculated as 494 km total length x 2 m width. dCalculated as 268 km total length x 12.6 m width. Characteristics of Instrumented Birds Twenty-five mallards were captured and radio-marked each year at Chestnut Slough. The 1986 sample was stratified by age and sex as follows: 5 adult males, 8 adult females,S immature males and 7 immature females (Table 2). In 1987, 7 adult males, 6 adult females, 6 immature males and 6 immature females were marked (Table 3). The percentage of marked males was increased in 1987 because of the higher mortality among males during 1986-87. Thirteen birds were monitored for the duration of the first field season (27 Feb.), but only 6 mallards were present at the conclusion of the 1987-88 field season (Tables 2 and 3). Increased severity of weather and higher hunting mortality contributed to the loss of instrumented birds during the latter year. Mallards instrumented in 1986 were lighter (X-l095 g) than those marked in 1987 (X-1133 g, Tables 2 and 3). Table 2. Characteristics of mallards radio-marked on 3 December 1986 at Chestnut Slough. Bird No. Age Sex Weight (g) Wing "(mm) Fate 816 Immature Female 1,033 288 Died 2/10 837 Adult Male 1,054 301 Last Contact 2/27 �29 Table 2 (cont.). Bird No. Age Sex 'Weight (g) 'Wing (mm) Fate 862 Adult Female 972 281 Last contact 2/27 887 Adult Female 999 289 Last contact 2/27 900 Immature Female 1,060 281 Last contact 2/13 916 Immature Female 1,202 289 Last contact 2/27 936 Adult Male 1,173 296 Bad Radio 946 Adult Female 1,076 281 Last contact 2/27 963 Immature Female 266 Died 12/17 979 Immature Female 985 277 Last contact 2/27 995 Immature Female 896 271 Last contact 2/27 1020 Adult Male 1,068 308 Last contact 2/27 1063 Adult Male 1,208 309 Died 12/20 1099 Adult Female 1,111 283 Died/last con. 1/30 1111 Adult Female 1,140 281 Last contact 2/27 1126 Adult Male 1,236 300 Last contact 2/10 1142 Immature Female 873 282 Last contact 12/3 1155 Immature Male 1,116 287 Last contact 1/5 1174 Immature Male 1,205 296 Last contact 2/25 1189 Adult Female 1,021 275 Last contact 2/27 1202 Adult Female 1,110 275 Last contact 2/27 1232 Immature Male 1,226 297 Last contact 2/6 1249 Adult Female 1,082 284 Last contact 12/27 1714 Immature Male 1,214 291- Last contact 2/27 1732 Immature Male 1,210 285 Died 12/15 �30 Table 3. Characteristics of mallards radio-marked on 2 December 1987 at Chestnut Slough. Bird No. Age Sex Weight (g) 030 Immature Male 1,228 286 Last contact 1/2 048 Immature Male 1,163 285 Died 1/23 070 Adult Male 1,139 302 Shot 1/2 090 Immature Female 1,034 272 Last contact 2/25 109 Adult Male 1,333 297 Died 1/17 128 Immature Female 969 264 Last contact 2/26 150 Immature Female 1,040 270 Died 2/4 170 Adult Female 1,133 283 Radio removed 12/20 184 Adult Female 1,043 284 Last contact 12/31 220 Adult Female 1,102 284 Radio fell off 1/15 240 Immature Female 1,162 281 Last contact 12/23 266 Adult Female 1,042 279 Radio removed 12/22 282 Adult Male 1,246 283 Died 1/3 300 Adult Male 1,188 295 Last contact 2/26 312 Immature Male 1,235 290 Died (powerline) 12/29 330 Immature Male 1,204 280 Last contact 12/30 350 Immature Male 1,142 287 Last contact 12/31 372 Adult Female 1,040 275 Last contact 2/26 392 Immature Female 899 267 Shot 1/1 420 Immature Female 980 279 Last contact 2/26 438 Adult Female 1,012 279 Last contact 1/17 464 Immature Male 1,140 291 Last contact 12/27 506 Adult Male 1,406 307 Last contact 2/18 532 Adult Male 1,249 307 Died 1/6 Wing (mm) Fate �31 Table 3 (cont.). Bird No. 568 Age Sex Weight (g) Adult Male 1,190 Wing (mm) 298 Fate Last contact 12/27 Movements Instrumented ducks were re-located 1,388 times in 1986-87 and 1,406 times in 1987-88. The time of telemetry fixes was skewed to evening in 1986-87 (Fig. 8), but nearly equally distributed throughout the day in 1987-88 (Fig. 9). Radio-marked birds dispersed in all directions from Chestnut Slough during both winters. In 1986-87, most telemetry locations were in the Chestnut Slough vicinity and eastward to Latham Reservoir, Beebee Draw and Milton Reservoir (Fig. 10). During 1987-88, radio-marked mallards were widely dispersed from around Windsor in the northwest, Lone Tree Creek in the northeast, Latham Reservoir in the southeast, and the St. Vrain River in the southwest (Fig. 11). Milton Reservoir received no use in 1987-88, but usage in the northeast portion of the study area increased markedly. Home range sizes, computed using the minimum area polygon method, differed among age/sex classes. In 1986-87, home ranges for adult male, adult female, immature male, and immature females averaged 219, 170, 131 and 94 ha, respectively (Table 4). Comparable home range sizes for 1987-88 were 133, 78, 126 and 141 ha (Table 5). Average overall home range sizes were 150 ha and 124 ha in seasons 1 and 2, respectively. Overall home range sizes using the Jennrich-Turner method with 95% confidence ellipses were 340 ha in 1986-87 and 233 ha in lSi:n -88, both considerably larger than those home ranges computed using the mL; 'llumarea polygon method (Tables 4 and 5). The shape of home ranges varied among birds, but often included heavily utilized wetlands that served as anchors or corners for minimum area polygons. In 1986-87, Chestnut Slough, Kodak ponds, and Seeley, Latham, and Milton reservoirs were important in delineating home range boundaries (Figs. 12-15). Kodak ponds, Latham Reservoir, Chestnut Slough and Norgren's Slough served the same functions in 1987-88 (Figs. 16-19). Home ranges of mallards commonly overlapped or in some instances included the entire range of another bird. Habitat Use Habitat use varied among birds in both 1986-87 (Table 6) and 1987-88 (Table 7). When unknown habitat type locations were eliminated and birds pooled within years, warm-water wetlands received the highest use in both years (Table 8). During the mild winter of 1986-87 portions of lakes and small ponds remained open, and these wetlands received considerable use by instrumented ducks (Table 8). In contrast, during 1987-88 many lakes and ponds were completely frozen, and duck use shifted to rivers, holding/sewage ponds and ditches (Table 8). Cornfield use decreased but feedlot use increased during 1987-88 compared to the previous year. �32 240 220 t/) 200 Z 0 180 i= <t 0 0 160 u, 120 140 ..J 0 0: 100 m 80 W :E ::) Z 60 40 20 0 0500 1000 1500 2000 TIME OF DAY Fig. 8. Hourly distribution of telemetry locations, 1986-87 t/) 120 Z 0 i= 100 ..J 80 <t 0 0 u, 0 0: 60 W m :E ::) 40 Z 20 0 0500 1000 1500 TIME OF DAY Fig. 9. Hourly distribution of telemetry locations, 1987-88 2000 �o a a + ~ =t* + 0# + + !of- + D a E .,..0 .lJt. c 33 •..o I: (II - III CJ .2 - .2 Q) E Q) •..>- G) td •.. til u:: �34 + o o + a o a o I ex) ex) " ex) en .•.. •.. VI C o o III ;:: o s CD >- -•.. I- E CD 'i ...• •... �i / '\ G a -. - ___ --_ ,," { o ~ o D 0 ~ ? r ~ ~ <:l pt::( ~ -} o N co ~ ~ o N o I I I I I ,.._ M co ci 35 ,.._ I co CD CO (1) ~ iii II) '0 ~ ~ CI) E II) E "5 '0 CI) (I) e II) - C'l t: II) E CI) ~ :::r::: e o tJ. C'l N - u: �36 L) o : G ,I ,( " , ,'I ,'1. \ / ./ /' • ,J ' .. : /.': '/ . /. : ;/ en CD ~ ~ ~ ~ ~ c:i CD co ('II • I I I I CD CD 0 ,... ('II o ('II . -----r.#f~~~ / "~,,.rr.:.':.~~1 ~ / . f /'/ / : .: oQ : : : . : . --!_ o P. o : : I / y' ;/: I ~, ;~: . 0 ;/ ,i /: / \',, /. \',/.' .i \'"/ ;/ E \' ,~: D ./ ~ ~ f\ I./~ ·I!l / '\\ . . 0 ,... I CD co CD (1) .•... o CI> g) (I) c: III •.. CI> •.. E o :::t M til u:: �4J rt "--t I , I I I I \ I ( )if tt) / (' \) /' a / / o q ( ,," " .- ;" 0 (8 /' r,-, II o ~: 1 /\. / e .. ~Jo~~/ o o ~ . : LO LO 'P'" 'P'" 37 Q) •..ftI c: Cl ~ (/) o - ............ [J .... o E o :I: �w 00 o ~ 1I '. o <> jl-~,_'_~~~ '. o _, c:::> -. o o " ~ r <> O\:J C> 0.816 ~h -- --- -L, " " 10 km ~)\ 0.900 ......... 0.916 Flg.15. Home ranges of Immature female mallards, 1986-87. 0.963 _ .._ ••_ 0.979 0.995 �I , ' .... " ~ ...... ..... Co (' .. ~ .... 'of ' ,/~,' ' ••• ''U . ,II' , ,': \ \'~. ~ \ ~. , • "'t:~;' I,' ~: t~\.. -.~ ././i ...\' '. ~ ~ \ ...~..... \' ,,;;., / .' ' _,/ ....::/ ':~A\,' ~\" \\0 ,I // ../'..... v., / »,> ".': 1/:'. 1-' f.• ' I I " \ ,' ., i": . \ , :.~ " \,," -1-. \ : : D : \~ :~"" l ~ ).' h<{J : !';.., ~ ~ ..~~------~ .".._".)( J ( ·\ "\1,! • \~ (I \ " "._'",,/' !) :,"..".","'/ ~,/ ,/ 1'('/ I if ,' ' (" /: ,'.) l:( I 'I -» " ', r/ ' I . , \' \ . _' / "f? G,' I \ \ :\ \ :~ '" G, ,./' (\' I • , • c ../>.:., 'l..' ,\ '.'. /', \ ~ ~~ ) , ' II ~ 'I": \-,.' of', / ,'1 V', ~" • : , : I i'· \ : , (• \ " i : \ \ \ ..,'/ . . r . \~' \~..,"'I o ,/ _..---,.,." , ," '.!,/ ,/,/ \t' (V / o . , }f' r;- -, '-1::;:- _0 0 t •""' ; r! r r } o co ci CD It) (\I ci C'? It) CD o ci It) o o M ci co (\I ci (\I I 39 aj co co "" O'l .•.. II) ... ('J '0 ('J E CLl E uS •... o ~ E LL �40 U o " _-/' ' I ~ : /: . : ~ . / :/ i/ • (J ,, ... " : : ' ~~:~ // ! ,I ) '" / (',1) 0 J . ,~ :, ., :: 0 '" . 1 . c- :. : '" ","" I' <:: f...... ',\.", ...,<-. .: ., , . ::':. i: '.: I I~: \. t: t \~,:. ::' : ,t.; ,:(: o ~D\ ~:~~--------------J "" .. I o o E 0 ~ 'P'" .__, o N •.... ci C') . IX) ex; ,....I 0 N N E I'D ~ a> ,... IX) ci I I I I o •.... 'P'" en o 'C I'D '3 CI.l -CI.l Cl C I'D •.. CI.l E o z Cl •..,... u: �) i · ; f I , I ~· .'.' · J it.,I. «, I \ I ' I Q I I a I .;.~ },'~\ ~~~ ~ ,~//' '~"'i~/": I <t ~ ' .:. r-.<~.~" { , ,I I.' II ~ c> 0 " ~a/~ "s, " " D " " 0 0 " ./ "" 0' ~ ' E ::.: o ", ' ~, o Q fl --.,~~-----•••• -' -or "", ',' ~ '\ ~" ~, ...' ~-J~,'''-\'''. "'. ~+-...f.1/ .,l . ,..:' .' ", =:': .-,-, "-: ~l '\, -' ! ~~ " o 0 o o CO) II) o CO) N .•.. o 41 CI> III E CI> ~ :J E III .5 - CI) .0 CI> til CI> c: III ~ E 0 X ... til ccS u:: �42 D D o o ('oj ci V o v ('oj ci o ci L() (Xl ci ('oj o OJ o ci cO (Xl ,....I (Xl OJ - o en CI) 01 C ,._III CI) o E z ai •... E LL �Table 4. 1986-87. Home range sizes of radio-marked mallards during winter, Home range size (Ian) Bird No. No. fixes Minimum area polygon method Jennrich-Turner method (95% ellipse) 816 40 146.5 285.0 837 113 328.9 858.1 862 66 90.8 177.7 887 54 137.7 195.2 900 23 0.1 0.3 916 46 159.9 492.6 946 105 455.0 842.2 963 10 1.2 4.9 979 43 127.5 351.9 995 64 129.8 312.9 1020 81 130.4 268.9 1063 16 0.6 1.3 1099 23 32.0 87.8 1111 57 141.2 372.0 1126 46 414.6 624.0 1155 13 4.5 29.4 1174 95 162.7 418.8 1189 54 209.7 519.4 1202 51 126.3 310.3 1232 71 214.5 441.6 1714 16 140.2 550.0 150.2 340.2 Means �44 Table 5. 1987-88. Home range sizes of radio-marked mallards during winter, 2 Home range size (km ) No. fixes Minimum area polygon method Jennrich-Turner method (95% ellipse) 048 79 143.5 139.0 090 120 432.0 380.7 128 114 87.1 76.5 150 55 28.4 36.1 170 11 3l.3 135.9 220 22 135.6 528.6 Bird No. , 240 20 6l.9 150.8 282 123 100.3 125.9 300 173 92.6 87.2 312 42 45.6 97.2 350 42 312.6 545.8 372 79 67.4 137.4 420 169 95.6 101.1 464 25 2.9 4.8 506 71 125.3 220.0 532 66 148.6 387.8 568 12 197.4 800.0 124.0 232.6 Means �Table 6. Habitat use by individual radio-marked mallards, 1986-87. Percentage of locations by habitat typeBird No. Unk. Pond Lake Hold. Gravel Warm-W. Ditch River Corn. No. obs. 92 102 9.8 31.5 1.1 0 0 22.8 0 29.4 5.4 816 6.8 4.6 20.4 0 0 63.6 0 4.6 0 837 5.6 9.5 36.5 0 0 39.7 0 1.6 7.1 126 862 13.9 11.4 1.3 0 0 54.4 1.3 12.7 5.1 79 887 20.0 7.1 2.9 0 0 62.9 0 0 7.1 70 900 11.5 3.9 84.6 0 0 0 0 0 0 26 916 18.3 0 5l.7 0 0 21.7 0 1.7 6.7 60 946 15.4 16.2 16.2 0 0 34.6 0 13.1 4.6 130 963 50.0 0 0 0 0 35.0 0 15.0 0 20 979 16.9 2.8 28.2 0 0 35.2 0 12.7 4.2 71 995 15.5 11.9 34.5 0 0 27.4 0 3.6 7.1 84 1063 18.2 4.6 0 0 0 63.6 0 13.6 0 22 1099-16.7 3.3 30.0 0 0 43.3 0 6.7 0 30 1111 11.6 10.1 60.9 0 0 10.1 0 2.9 4.4 69 7.1 5.4 7.1 0 5.4 60.7 0 3.6 10.7 56 1155 24.0 0 20.0 0 0 52.0 0 4.0 0 25 1174 14.3 25.9 20.5 0 0 30.4 0 2.7 6.3 112 1189 7.7 13.9 67.7 0 0 4.6 0 0 6.2 65 1202 11.4 12.9 17.1 0 0 41.4 0 12.9 4.3 70 1232 11.2 15.7 14.6 0 0 36.0 0 9.0 13.5 89 0 0 0 0 0 0 0 0 2 1126 1249 100 44- �46 Table 6 (cont.). Percentage of locations by habitat type· Bird No. Unk. Pond Lake Hold. Gravel Warm-W. Ditch River Corn. No. obs. 1714 40.S 7.1 23.8 0 0 23.8 0 4.8 0 42 1733 SO.O 0 0 0 0 0 0 SO.O 0 4 •Key to habitat abbreviations: Unk. - unknown habitat type Pond - Small ponds/reservoirs; Lake - lakes; Hold. - holding/sewage ponds; Gravel - gravel pits; Warm-W. - warm-water wetlands; Ditch - ditches; River - rivers; Corn. - cornfields. Table 7. Habitat use by individual radio-marked mallards, 1987-88. Percentage of locations by habitat type· Bird No. Unk. Pond Lake Hold. 'Warm-'W.Ditch River Corn. Feed. No. obs. 30 50.0 30.0 20.0 0 0 0 0 0 0 10 48 13.8 2.3 1.2 0 65.5 6.9 10.3 0 0 87 49 0 0 0 0 0 100.0 0 0 0 1 SO 0 0 0 0 100.0 0 0 0 0 3 70 75.0 0 25.0 0 0 0 0 0 0 4 90 7.5 17.9 2.2 0 53.7 4.5 6.7 3.0 4.5 134 109 33.3 0 33.3 0 0 33.3 0 0 0 3 128 8.7 3.9 0 0 4.7 7.9 52.8 1.6 20.5 127 ISO 3.S 1.8 14.0 0 66.7 1.8 12.3 0 0 S7 170 0 0 0 0 30.8 23.1 46.2 0 0 13 0 0 0 0 0 0 0 0 2 . 184 100.0 220 13.3 13.3 0 0 63.3 0 10.0 0 0 30 240 0 4.8 19.1 0 57.1 4.8 9.5 4.8 0 21 , �47 Table 7 (cont.). Percentage of locations by habitat type· Bird No. Unk. Pond Lake Hold. lJarm-lJ.Ditch River Corn. Feed. No. obs. 266 o 100.0 0 0 0 0 0 0 0 1 282 2.2 0 2.2 0 91.8 0 3.7 0 0 134 300 2.8 32.6 0.6 0 32.0 2.8 13.3 13.3 2.8 181 312 12.0 16.0 0 2.0 56.0 4.0 8.0 2.0 0 50 330 50.0 0 0 0 50.0 0 0 0 0 2 350 12.2 42.9 2.0 0 26.5 2.0 10.2 4.1 0 51 360 100.0 0 0 0 0 0 0 0 0 1 372 7.3 6.1 11.0 0 51.2 2.4 4.9 17.1 0 82 392 25.0 0 62.5 0 0 0 12.5 0 0 8 400 0 0 0 0 100.0 0 0 0 0 1 420 3.9 16.9 0 0.6 52.3 0 14.0 12.4 0 178 438 18.2 0 45.5 0 18.2 0 9.1 9.1 0 11 464 3.7 0 81.5 0 3.7 7.4 3.7 0 0 27 506 19.6 0 5.4 35.9 17.4 1.1 10.9 0 9.8 92 532 5.5 46.6 4.1 0 39.7 0 2.7 1.4 0 73 568 27.3 18.2 22.7 0 31.8 0 0 0 0 22 •Key to habitat abbreviations: Unk. - unknown habitat type Pond - Small ponds/reservoirs; Lake - lakes; Hold. - holding/sewage ponds; lJarm-lJ.- warm-water wetlands; Ditch - ditches; River - rivers; Corn. - cornfields; Feed. - feedlots. �48 Table 8. Habitat use by radio-marked mallards 1986-87 Habitat type during 1986-87 and 1987-88. 1987-88 No. locations No. locations , Small ponds/reservoirs 167 14.0 204 15.8 Lakes 344 29.0 79 6.1 o o 35 2.7 3 0.3 o o 488 41.0 627 48.6 0.1 42 3.2 108 9.1 186 14.4 77 6.5 72 5.6 o o 46 3.6 Holding/sewage Gravel ponds pits Warm-water Ditches Rivers Cornfields Feedlots Totals wetlands 1 1,188 100.0 1,291 100.0 Differences in habitat use were apparent when instrumented birds were pooled by age and sex class. In 1986-87, adult males used warm-water wetlands more than other age/sex groups (Table 9). Females used lakes more frequently than males, and were located in cornfields more often than birds of other age/sex classes. During both field seasons, adults made greater use of riverine habitat than did immatures. In 1987-88, adult males used small ponds and reservoirs more than other age/sex groups, whereas immature females made greater use of rivers (Table 10). Temporal changes in habitat use, mostly in relation to the availability of ice-free wetlands, are apparent when use data are summarized by month. Lakes and small ponds/reservoirs received use whenever these habitats were ice-free, as they were in December, January and February 1987 (Tables 11 and 12). However, when temperatures were coldest, warm-water wetlands received their highest use. Use of rivers was nearly constant across months and years except for a large increase in use during February 1988. Ditches were used infrequently during early and late winter, 1987-88. Mallards used feedlots only during January and February 1988 (Table 12). �49 Table 9. Mallard habitat use by age-sex class during winter, 1986-87. , , , , 1m, male No. obs. Habitat type Ad, bmale No. obs. Ad, male No. obs. 1m, flilma1e No. obs. Unknown 73 14.2 24 8.1 52 17.1 51 18.8 Small pondsj reservoirs 61 11.8 45 15.2 15 4.9 46 16.9 131 25.4 51 17.2 111 36.4 51 18.7 0 0 3 1.0 0 0 0 0 184 35.7 119 40.2 96 31.5 89 32.7 1 0.2 0 0 0 0 0 0 Rivers 40 7.8 34 11.5 18 5.9 16 5.9 Cornfields 25 4.9 20 6.8 13 4.3 19 7.0 Lakes Gravel pits Warm-water wetlands Ditches ·Co1umn percentage Table 10. Mallard habitat use by age-sex class during winter, 1987-88. , Habitat type Ad, female No. % obs. Ad. male No. obs. 1m, female No. % obs. 1m, male % No. obs. Unknown 14 10.0 40 7.9 32 6.1 31 13.4 Small pondsj reservoirs 10 7.1 97 19.1 61 11.6 36 15.6 Lakes 14 10.0 19 3.7 20 3.8 26 11.3 Holding/sewage ponds 0 0 33 6.5 1 0.2 1 0.4 Warm-water wetlands 67 47.9 233 45.8 222 42.2 103 44.6 Ditches 5 3.6 7 1.4 18 3.4 12 5.2 Rivers 15 10.7 41 8.1 111 21.1 19 8.2 Cornfields 15 10.7 25 4.9 29 5.5 3 1.3 0 0 14 2.7 32 6.1 0 0 Feedlots Column percentage �50 Table 11. Mallard habitat use by month during winter, December Habitat type No. obs. Unknown 1986-87. January % • No. obs. February % No. obs. % 131 21.4 53 3.8 16 15.0 31 5.1 79 18.7 57 16.2 104 17.0 57 13.5 183 52.0 0 o 3 0.7 o o 277 45.2 194 45.9 17 4.8 1 0.2 o o o o Rivers 49 8.0 33 7.8 26 7.4 Cornfields 20 3.3 41 9.7 16 4.6 Small ponds/reservoirs Lakes Gravel pits Warm-water wetlands Ditches Column percentage Table 12. Mallard habitat use by month during winter, December 1987-88. January February a Habitat type No. obs. Unknown No. obs. % % No. obs. % 71 10.1 28 6.4 19 7.1 170 24.2 27 6.1 7 2.6 79 11.3 o o o o 7 1.0 25 5.7 3 1.1 289 41.2 245 55.7 93 34.8 Ditches 25 3.6 11 1.4 6 4.1 Rivers 51 7.3 41 9.3 94 35.2 Cornfields 10 1.4 38 8.6 24 9.0 o o 30 6.8 16 6.0 Small ponds/reservoirs Lakes Holding/sewage Warm-water ponds wetlands Feedlots Column percentage �51 Response to Hunting Disturbance Responses to hunter disturbance were recorded for 15 radio-marked birds at Chestnut Slough from 13 December 1986 to 3 January 1987 (Table 13). Gross analysis indicated little difference in the age/sex proportion of usable radio-marked birds within the study area using Chestnut sometime during the hunting season (3 of 4 AM, 6 of 7 AF, 3 of 31M, 3 of 5 AF), although adult females seemed more likely to return to the Chestnut roost following disturbance. Table 13. Presence or absence of radio-marked birds at Chestnut Sloughs on consecutive days during and following disturbance at Chestnut, 1986-87 (P - Present, A - Absent) Date December Bird January Age/Sex AM AF 1M IF • a No. 13 14 15 837 1126 1063 P P P a 20 21 a a 27 22 28 29 A P A P a 3 4 A P P A A A A A A P A P P P A A A A P b P P P P P P P P P P ~c A A A A A A A A P P P A A A A A P b P P A A A P P A A A P P A A A A A A A A A A P P A A A A A A A A A P A A A P A A P A A A A P P A A A A 862 887 1202 1099 946 1111 A A 1174 1155 1232 P P A A P A A 979 916 963 P A A P P P P P P P P A P P A A P No hunting on Chestnut bDeparted Chestnut prior to any disturbance cHarvested A A A A A A A A A A A b �lJ1 N TABLE 14 • Hunting season disturbance birds at Chestnut Sloughs 1986-87. factors (see methods), bird reaction Disturbance Age/Sell AM Bird No. 0.837 Date 12113 12114 12120 1.126 1.063 AF 0.862 0.887 1.202 AF .. 1.099 0 * * 1 28 7 12113 12114 12127 12128 12/13 12114 U Accu.ulated rating before departurt! Hours before return 40 7 3 10 11 >25$.203 111 9 10 7 23* 1 1 26 3 10 Harvested 3 18 U U 6 21 7 0 11 )22910 )11522 10 * 2* 1 7 1 1 1 113 22 11 2 92 10 11 5.600 7 1 1 1 43 21 13 1 10 $..202 10 1 1 8 4 * 28 19* U 12121 2* 12/13 12114 12120 s 3 2 * 4 6* 4* 12120 12/13 12/14 12/20 3 1 12127 12128 12/14 12120 2 and roost homing for radio-marked * 28 * 2 18/ 4 U B* U 11 HarvestE.·d �Disturbance AgelSex Accu.ulated rating before departure 0 1 2 3 0.946 12114 12128 1/3 * 19 1 1 24 0 0 1 3 0 1.174 1.155 ,1.232 IF 5 Date 1.111 1M 4 Bird No. 0.979 12127 12128 12113 12/14 12120 * * * 22 18 12121 12128 12/13 12114 12121 12127 1/3 3 2* * 1 U 1 * 12113 12121 6* U U 3 * U 9 11 11 <24 Did not return 11 10 ~672 953 Did not return 2 3 948 6 10 10 90 1 1 21 2 3 3 1 2 * * 6 2 2 18* 37 21 3 Hours before return 8 11 10 0.916 12/14 17 U 20 10 0.963 12/14 17 U 20 10 \J1 W �Vt "" TABLE 15 • Hunting season disturbance Chestnut Sloughs 1987-88. factors, bird reaction and Disturbance Age/Sex AM Bird No. 0.532 Date 12126 112 1/2 112 5 * AF 0.438 12112 * 1M 0.0L!8 12113 12113 1211L! 12115 12117 12/18 12119 12/20 12121 12122 * * * * * * * 0.350 0.312 12112 12113 12126 12117 12118 12119 12120 12121 12123 12128 1 0 2 3 4 5 4 3* 1 1 4 U 2 2 U 2 4 3 3 ** * * * * * ho.ing for 1 1 1 1 1 1 U 1 5 3 '1* 5 U radio-IIIarked birds Acculllulated rating before departure 35 4 28 45 0 U * roost at Hours before return >96~166 0.25 3 <18 Did not return 7 1 Did not leave 0 3 0 0 0 36 0 0 5 >115.28 11 ~11 90 ~4 90 >24916 5 8 0 >240912 8·· Did not return 0 0 0 0 0 11 11 11 >10<12 o· >7920 0 3'1 Did not return (Found dead) �II' c: ~ :s QI +J QI QI +J :s ~ .II) QI ~ 0 ~ ~ ~ :s 0 % 0- ...•c: ~ ~ 'II +J 'II QI '1:) CL. QI QI '1:) +J ...•'II ..., o::r :s ~ E 0 :s ~ u QI u .&J ~ QI u c: :s I. 'II .&J 1/1 N .... +J II:> o o :z '1:) ..•~ al :s c: ~ ~ +J 0 Q '1:) ...• e 0 +J QI o::ro .... vi o::r ..., N "- Oo::r * .... * * N N en N ....•.. N ...• ...• N "_ ...• o c;r N . o 0- o . o o 55 �56 Eight birds (1 AM, 4 AF, 21M, 1 IF) were present on Chestnut during the opening morning of the hunting season (Table 14). Two adult males (887, 1202) remained on the area throughout all disturbance although they left the area on an afternoon feeding flight. Birds were tolerant of shooting that occurred off of Chestnut Slough but could be heard on Chestnut (disturbance factor 1), since only 1 response was noted in 116 occurrances. Hunters driving in vehicles (disturbance actor 3) resulted in 3 of 7 birds leaving. Two of 4 left in response to shooting on Chestnut >200 m away from the birds (disturbance factor 2). Four birds (Table 14) were generally tolerant of all disturbance types recorded on opening day. No birds returned to Chestnut during daylight hours, but all but 1 bird (1M 1155) returned for night roosting. On 14 December 1987, 3 of the 5 birds remaining on Chestnut that had been present on 13 December responded to hunters driving in; 2 of the same birds had responded in the same manner on 13 December. Birds experiencing their second consecutive day of disturbance were less tolerable of minor disturbances, with all birds leaving prior to any shots being fired on Chestnut. Four birds experiencing their first day of Chestnut disturbance (1063 AM, 946 AF, 916 IF, 963 IF) responded in a similar manner, although the last 2 birds to leave were among this group_ No birds returned to Chestnut during daylight hours, but again all but 1 bird (862 AF) returned there to roost at night. Hunting occurred at Chestnut on 17 December, 1987 but monitoring of bird/hunter interactions did not. Six birds (2 AM, 3 AF, 11M) were on Chestnut on the morning of 12 December 1988. Five of six experienced at least their third disturbance bout on that day. Three of 6 birds left the area in response to the initial disturbance which was hunters driving into the area. None of the 3 birds had left in response to eriving in previous bird/hunter interactions, indicating a "learned" response. A fourth bird left shortly after "drive-in" with no disturbance evident, while a fifth bird left when the first shots were fired away from Chestnut. A sixth bird (AF 1202) which had been very tolerant of dist~rbance the previous weekend remained on Chestnut throughout the hunting bout (Table 14) but left later in the day. Only one (AF 887) of the 6 birds returned to Chestnut for night roost; 2 were shot by hunters, and 3 were noted on Chestnut again for about one week, including the bird that was tolerant of disturbance (AF 1202). Two birds (1M 1232, IF 979) joined 887 on the night roost on 20 December 1986. Immature male 1155 arrived about 0.5 hours before sunrise on 21 December, 1986. There were no hunters on Chestnut on 21 December, 1986. Bird 1155 left in response to the second occurrence of shooting away from Chestnut and eventually left the study area without returning to Chestnut. The remaining 3 birds experienced no additional ~isturbance and remained on the area for at least 2 hours before departing. The latter 3 birds were not noted on Chestnut for a considerable amount of time after their 21 December departures (Table 14), but monitoring of birds on Chestnut was sporadic during the latter part of December. Birds were hunted but no monitoring occurred on 23, 24 and 25 December, 1986. Even though considerable disturbance occurred at Chestnut prior to the next monitoring session on 27 December, 4 birds were present on that morning. Two of the birds (AM 1126, AF 1111) were first noted at Chestnut on the afternoon of 26 December 1986 and may not previously been subject to disturbance at Chestnut. Adult female 862 had not been noted at Chestnut for about 2 weeks and the previous Chestnut occurrence noted for 979 was on 21 �57 December. All birds tolerated the initial drive-in by hunters but began to leave within 5 minutes after drive-in and before additional disturbances. One of the "new" birds (1111) left early while the second (1126) tolerated about 40 minutes of sporadic disturbance and was the last of the 4 birds to leave. All 4 birds returned to Chestnut after dark along with 947, but 979 left after about a 10 minute stay. On 28 December 1986, 3 of 5 birds left about 20 minutes before sunrise prior to any disturbance. Bird 1232 left when hunters drove in about 1 1/2 hours after sunrise, but 1126 tolerated some shooting on Chestnut before leaving when shots were fired close by. All birds except 1111 returned that evening. Two birds (862 and 1232) returned before sundown but left again within 20 minutes and only 1232 returned. Bird 947 which returned about 40 minutes after sunset left about 15 minutes later, leaving only 1126 and 1232 at roost on Chestnut from the original morning disturbance group. Chestnut was very attractive to radio-marked birds on the evening of 28 December as 4 other birds also roosted there. On 3 January 1987, 946 was the only bird on Chestnut, and it left before any disturbance, about 40 minutes before sunrise and returned about 40 minutes after sunset. Radio-marked birds continued to use Chestnut Slough through the 13 December - 3 January 1986-87 period in spite of disturbance, although the number of birds using the area declined as the season progressed (Table 14). Fifteen marked birds experienced a total of 31 disturbance/days (1 bird disturbed on 1 day) throughout the period. Those birds returned to the Chestnut roo;;t.generally within 10 hours following 81% of those disturbances (after Table 14). Only 2 birds (87%) failed to return to roost following their first exposure to disturbance while 3 of 11 (79%) did not return following their second bout with Chestnut disturbance. With one exception, the birds did not return to Chestnut within the same daylight period following disturbance. Radio-marked birds did not choose to remain on Chestnut exclusively when not disturbed. In 3 instances marked birds left Chestnut before any disturbance occurred and 13 departures took place after some disturbance. Although the departures took place after some disturbance, the departures were not recognized as a responses to a specific disturbance. The period of disturbance monitoring followed a short period of abnormally cold temperatures but was characterized by moderate temperatures. Responses to hunter disturbance were recorded for only 7 radio-marked birds at Chestnut Slough from 12 December 1987 through 2 January 1988. Only 9 radio-marked birds (2 AM, 2 AF, 31M, 1 IF) were recorded on Chestnut during the hunting season, and 2 of those were thought to be behaving atypically. Attempts to monitor early morning hunter and bird activity occurred on 17 of 22 mornings at Chestnut. Because of the absence of radio-marked birds and early morning hunting and the movement of birds away from the Chestnut roost before any disturbance, disturbance/bird interactions were recorded for only 8 bird-days during the 1987-88 period. By leaving the roost site early and returning after dark, 3 of 7 birds using Chestnut never experienced any hunting disturbance (Table 15). Early morning departures were typical of immature male 312 which roosted on Chestnut 7 nights in late December. Immature male 48 left the roost before disturbance on 5 of 9 days in mid to late December. Those birds experiencing harassment on Chestnut usually left after minor disturbance except on 13 December (1M 48), 20 December (1M 48), 26 December (AM 532, 1M 350) and 2 January (AM 532, IF 90). The later 2 days had �58 temperatures below 0° F. Adult male 532 returned to Chestnut after leaving in response to disturbance~ twice on 2 January. In contrast to birds monitoreu Uuring the 1986-87 season, 1987-88 birds in general used Chestnut Sloughs le~s .nd .ere less tolerant of disturbance. Whereas a greater use of Chestnut and more tolerance of disturbance by 1987-88 birds would have been predicted, given the colder temperatures (Fig. 2) and greater snow cover (Fig. 5) during 1987-88. Behavior Time budget observations totaling 4,759 minutes were obtained on 4 wetland habitat types during 1986-87 and 1987-88 (Table 16). Observation times were similar within pair status-sex-habitat categories. Resting was the most common activity each winter, followed by swimming (Figs. 20 and 21). Time devoted to preening was constant both years, but water feeding decreased in 1987-88 while courtship and other activities increased. Males and females behaved similarly in 1986-87 (Fig. 22), but in 1987-88 females spent more time in courtship and less time swimming than did males (Fig. 23). Table 16. Behavioral observation time by hai>itat type, sex, and pair status. 1986-87 and 1987-88. Observation Time (min.) Habitat Sex Status Reservoir Female Paired Unpaired Paired Unpaired 312.1 287.1 244.8 288.2 Paired Unpaired Paired Unpaired 453.0 236.6 521.1 286.6 Paired Unpaired Paired Unpaired 306.0 302.6 246.7 427.9 Paired Unpaired Paired Unpaired 261.6 158.2 165.7 260.5 Male Riverine Female Male 'Warm-water Female Male Female Other Male Total 4,758.7 Behaviors differed among habitat types, but a habitat-year interaction �59 Fig. 2O. Time budget for all mallards, 1986-87. Rest Swim Preen Water Feed Fig. 21. Time budget for all mallards, 1987-88. Courtship Other �60 Fig. 22. Time budget for females (left) and males (right) 1986-87. Rest Swim Preen Water Feed Courtship Fig. 23. Time budget for females (left) and males (right) 1987-88. Other �61 • Re.t ~ Re.ervolr. Warm-waters River. ~ Swim Preen ~ Water Feed IQQ Court.hlp rza Other Others Fig. 24. Time budgets by habitat type, 1986-87 . • Rest ~ Re.ervolr. River. ~ ~ ~ rza Warm-waters Other. Fig. 25. Time budgets by habitat type, 1987-88. Swim Preen Water Feed Courtship Other �62 • Rest ~ Wj ~ Jan. 1987 F.b.1987 Fig. 26. Time budgets for Jan.• nd Feb. 1987. Dec. 1987 Swim Preen Water Feed 50j Courtship rza Other • Rest ~ Swim ~ Preen ~ Water Feed ~ Courtship o Other Jan. 1988 Feb. 1988 Fig. 27. Time budgets for Dec., Jan.• nd Feb. 1987-88. �63 was apparent (Figs. 24 and 25). Generally, warm-water wetlands were used for resting more than other wetland types. Courtship displays were most common on reservoirs and warm-water wetlands. Water feeding, common on reservoirs and rivers during 1986-87, was less apparent ~n these habitats in 1987-88. Monthly time budget summaries indicate temporal changes in behavior. During 1986-87, February behaviors included more resting but less water feeding than in January (Fig. 26). Courtship behavior was most common in December 1987 and January 1988, but decreased in February 1988 (Fig. 27). Swimming decreased whereas water feeding and preening increased from December through February, 1987-88 (Fig. 27). DISCUSSION Differences in snowfall and temperature regimes between the two field seasons affected habitat availability on the study area. Consequently, mallard habitat use and behavior differed between years. During 1986-87, when most wetlands remained unfrozen due to mild temperatures, home range size was larger than in the colder winter of 1987-88. With fewer wetlands available, birds reduced the number and frequency of their movements. In Nebraska, size of activity centers (home ranges) differed between adult and immature female mallards (Jorde 1981). However in 1986-87, all age-sex groups appeared to differ in home range size. In 1987-88, immatures occupied larger home ranges than adults. Jorde (1981) found that immature males and females occupied activity centers 27.8% and 28.2% larger than their adult counterparts, respectively. Nebraska mallards occupied primary activity centers ranging from 210 to 842 ha. These values are similar to the home range sizes of mallards reported here. Relative to their availability (0.6% o£ the sur£ace water area), warmwater wetlands were highly selected by radio-marked uallards in both winters. Jorde et al. (1984) speculated that such areas may provide thermal refugia important for minimizing thermoregulatory energy loss. Surprisingly, maximum air temperatures 5 cm above the water surface at Chestnut Slough, a warm-water wetland, were slightly lower than ambient air temperatures 1.5 m above the ground (Fig. 28). Minimum temperatures at 5 cm height were, however, warmer than-ambient conditions. Thus, in this instance a thermal advantage was realized only during late night and earl~ morning periods when diurnal temperatures were at a minimum. Nevertheless, water temperatures at 15 cm subsurface depth were always well above air temperatures (Fig. 29), thereby providing a favorable microclimate for minimizing heat loss (eg. Kilgore and Schmidt-Nielsen 1975). These data suggest that aside from simply providing water, warm-water wetlands provide a thermal habitat beneficial to wintering waterfowl. Mallards respond to a hierarchy of habitat selection that, at its highest levels, is dependent upon the availability of ice-free wetlands weighted by such disturbance factors as hunting (Ringelman et al. 1989). When weather allows, mallards preferred lakes and $mall ponds over rivers, holding ponds and ditches. Under snowfree conditions, ducks preferred to feed in cornfields rather than feedlots. However, mallards showed plasticity in their response to these and other conditions, adapting their movements and behavior to shortterm weather events. Temporal and spatial variability in habitat use may also relate to social events that occur during winter. Most mallards establish pair bonds during mid-winter in Colorado, and wetlands such as lakes and warm-water sloughs are �64 60 -u, W 40 a: ::J t~ a: w 0. ::E w t- . 30 I I I 20 , • • ,_ , " ,;, 'I I. I ' ~ ,'I II , 10 • " : .' " • " I' , • I' , I' .' .' , i ,'. I. I~'I,. I ~ I •• • I " 'I I,' '• ,I ,. I • • " ••••• I I " ~ ,""'''..' " __ ~~~ ,I \ " '. • 1" • __ , '. 1 I I I -10 ,,' 10 1 >" _-. , .~ I,' I • .: 1 30 I , • t '. i,. " •, I ~~~~ ',. 20 Dec. ". ,I I .' ,I,. 11 f' II , • • I '. ':. " , .' • •• I I' .' !: __ ~_L_~~~ • O .1 II .' ,. •••.• . , • • ' ~~ _ .' " ., ' I I 20 10 30 20 Feb. DATE Fig. 28. Maximum and minimum dally temperatures 5 em above the water at Chestnut Slough, 1987-88. 60 -- .' 50 ,, '' U. W , 40 a: ::J t~ 30 w 20 ::E w 10 a: 0. t- I '. ,'"" - ~, ..... '" , •.'... ',_, ,', " ~ '\ ' ",,,\ I I' " , " ",,"'''' ,•. -----------------~--~-----~------',' , •.•._ ,'---' , ...' ...--,' I, ' ••...."', ,--, '---",," O -10 11 20 Dec. 30 10 .20 30 DATE Fig. 29. Maximum and minimum dally temperatures 15 em under water at Chestnut Slough, 1987-88. 10 20 Feb. �65 thought to provide ·courtship arenas" for mallards on the High Plains (Ringelman et al. 1989). After pairing, mallards vacate such wetlands in favor of locations that provide greater isolation from conspecifics. Since adult male mallards outnumber females in Colorado and competition for mates is acute, extensive use of lakes and ponds by adult male mallards in 1987-88 may reflect the role of these wetland types in the social behavior of mallards. Behavior of mallards differed between years, but in general was dominated by resting (includes sleeping), swimming, water feeding and preening. Jorde et a1. (1984) found that mallards wintering in Nebraska behaved similarly, with behaviors dominated by sleeping, water feeding and preening. Resting, swimming and preening also dominates the behavior of mallards wintering in Oklahoma (Gordon 1977). As with habitat use, time budgets reflect social changes occurring in the mallard population. For example, courtship behavior remained high during December and January of 1987-88, but declined in February when numbers of unpaired birds presumably declined. LITERATURE CITED Altmann, J. 1974. Observational study of behavior: sampling methods. Behaviour 49:227-267. Dwyer, T. J. 1972. an adjustable radio-package for ducks. Bird-banding 43:282-284. Fenneman, N. M. 1931. Physiography of the western United States. Hill Book Company, New York and London. 534pp. McGraw- Gittings, E. B. 1941. Climate of Colorado. Pages 798-808 in Yearbook of Agriculture: Climate and Man. U.S. Dep. Agric., Washington, D.C. Gordon, D. H. 1981. Condition, feeding ecology, and behavior or mallards wintering in northcentral Oklahoma. M.S. Thesis, Oklahoma State University, Stil-1water. 68pp. Jorde, D. G. 1981. Winter and spring staging ecology of mallards in south central Nebraska. M.S. Thesis, Univ. North Dakota, Grand Forks. 116pp. Jorde, D. G., G. L. Krapu, R. D. Crawford, and M. A. Hay. 1984. Effects of weather on habitat selection and behavior of mallards wintering in Nebraska. Condor 86:258-265. Kilgore, D. L., Jr. and K. Schmidt-Nielsen. 1975. feet immersed in cold water. Condor 77:475-517. Heat loss from ducks Ringelman, J. K., W. R. Eddleman, and H. W. Miller. 1989. High plains reservoirs and sloughs. Pages in L. M. Smith, R. L. Pederson, and R. M. Kaminski, eds. Habitat management for migrating and wintering waterfowl in North America. Texas Tech Press, Lubbock. Szymczak, M. R. , and J. F. Corey. 1976. Construction and use of the �66 Salt Plains duck trap in Colorado. Tacha, T.C., P.A. Vohs, and G.C. Iverson. 1985. A comparison of interval and continuous sampJing methods for behavioral observations. J. Field Ornithol. 56:258-264. Prepared b~ l(~ ~Ringelman Yildlife Resear:::I , ��Colorado Division of Wildlife Wildlife Research Report December 1988 JOB PROGRESS REPORT State of __~C~o~l~o~r~a~d~o~ Project 01-03-212 (W-88-R) Work Plan 1 : Job _ Avian Research - Migratory Birds 18 Job Title: Winter survival and reproductive success of female mallards' Period Covered: Author: 01 January 1987 through 31 March 1988 James K. Ringelman Personnel: D. R. Anderson, J. Armstrong, J. Boulanger, K. Frye, D. Gilbert, M. Gilbert, W. Schraeder, Colorado Cooperative Fish and Wildlife Research Unit; C. Jeske, S. Borthwick, R. Knight, J. Laake, E. Rexstad, R. Scarpella, T. Shenk, D. Smith, J. Schmutz, K. Wilson, Colorado State University; J. Corey, R. Hopper, M. Szymczak, Colorado Division of Wildlife; R. Croft, R. Kirby, M. Nail, U.S. Fish and Wildlife Service. ABSTRACT Adult female mallards were captured and radio-marked in the San Luis Valley, Colorado, to assess (1) the magnitude of winter mortality, and (2) the relationship between winter body condition and subsequent reproductive performance. Hens were assigned to either the "good" (upper one-third of the condition range) or "poor" (lower one-third of the condition range) condition groups based on an index to fat reserves. In 1987, 105 mallards were instrumented with back-mounted transmitters, then located from the ground or air weekly. Of the birds marked in January, twenty-eight good condition and 37 poor condition hens died during 17 January through 4 July" for minimal mortality rate estimates of' 0.538 and 0.699, respectively. Mortality rates for, females instrumented in February were 0.277 for females in "good" condition'and,0.588 for females in "poot" condition. Of 31 instrumented'birds in the SanLuf.s Valley during May, only 2 females were known to have nested. In 1988, back-pack radio transmitters were affixed to 189 adult female mallards. from January through March. Most radios were applied to females in the upper or lower third of the condition index distribution. No differences in survival rates of females in "good" and "poor" condition were detected. A pooled 70-day survival rate of 0.279 was observed. Two of the instrumented females nested. Fate of 1 nest was unknown, the other was destroyed by avian predators. Instrumented females were collected from 2-5 weeks after instrumentation to determine effects of instrumentation on weight dynamics. Of 40 females collected, only 2 had gained weight, and some had lost a substantial percentage of their initial weight. Increased search effort over 1987 and an outbreak of avian cholera allowed us to collect 4,195 mallard carcasses, of which 900 were analyzed for lipids in the ulna. Ulna lipids indicated that 6.7% of the carcasses analyzed had depleted body lipids at the time of death. �69 YINTER SURVIVAL AND REPRODUCTIVE SUCCESS OF FEMALE MALLARDS James K. Ringe1man Michael R. Szymczak This report documents the accomplishments of 2 field seasons, 1986-87 and 1987-88. Parenthetical references to years refers to objectives, methods, or results from a single field season. The absence of such a reference implies objectives, methods, or findings common to both field periods. P. N. OBJECTIVES 1. Quantify winter survival rates of adult female mallards good condition. 2. Determine the sources of winter mortality Luis Valley. 3. Investigate the relationship between winter body condition and reproductive performance of mallards wintering in the San Luis Valley. 4. Measure the winter energy balance of mallards by quantifying components of energy acquisition and depletion. for mallards in poor and in the San SEGMENT OBJECTIVES 1. Capture and radio-mark a minimum of 50 "poor" and 50 "good" condition adult female mallards in the San Luis Valley during January. 2. Quantify differences trapping. 3. Affix nasal saddle markers (1986-87) and wing bands (1987-88) to mallards to assist in behavioral observations and mortality estimations. 4. Measure the effects of radio instrumentation mallards (1987-88). 5. Conduct scheduled ground and aerial tracking to locate dead, radiomarked birds in winter and nesting birds in spring. 6. Measure clutch parameters mallards. 7. Collect remains of marked and unmarked carcasses, causes of mortality in the carcass sample. 8. Measure the standard metabolic rate of mallards, and perform digestion trials to assess the digestible energy of mallard foods. in mallard condition by age, sex, and date of on weight dynamics of and record nesting history of radio-marked and partition �70 9. Conduct winter counts of raptors to assess their temporal abundance and potential relationship to mallard mortality. STUDY AREA This study was conducted in the San Luis Valley (SLV), located in south-central Colorado (Fig. 1). The SLV is a 12,960 km2 intermountain basin, bounded by the San Juan Mountains to the west and the Sangre de Cristo Range to the east. As a result of the surrounding mountains and 2,286 to 2,438 m elevation, January temperatures average -4 C, and snow and ice fog are common. Upland areas are dominated by greasewood (Sarcobatus vermiculatis) and. rabbitbrush (Chrysothamnus spp.). Common wetland vegetation includes baltic rush (Juncus balticus), cattail (~ latifolia), hardstem bulrush (Scirpus acutus), coontail (Ceratophyllum demersum) and pondweeds (Potamogeton spp.). Further descriptions of the SLV can be found in Hopper et al. (1975) and Szymczak (1986). An average of 14,700 mallards winter in the SLV (1982-84 average, Colo. Div. Wildl., unpubl. data), with most residing on the Monte Vista National Wildlife Refuge (MVNWR). MVNWR is 5,670 ha in area and has over 200 ponds and impoundments. Wintering ducks are concentrated on those water areas which do not freeze as a result of continuous pumping or artesian flow. Mallards breed throughout the SLV and surrounding mountains (Rutherford and Hayes 1976, Szymczak 1986), but nesting is concentrated on the MVNYR (pers. obs.). METHODS Capture - 1986-87 and Instrumentation Mallards wintering on the MVNWR were captured with Salt Plains bait traps (Szymczak and Corey 1976) 17-20 December 1986, 5-14 January, 20-22 February, 5 and 18 March 1987. Each bird had its absolute and relative lipid stores assessed, it's condition index (Ringelman and Szymczak 1985) computed, and molt status scored for adult females. Most captured birds were classified by age and sex, then banded with standard USFYS leg bands. Adult females were assigned to 1 of 3 condition classes based upon their condition index relative to the distribution of the population in the SLV. Population distribution was determined by ordering the condition indices of each age-sex class on the first trapping day. Birds with a condition index in the upper 33% of the population were termed in "good" condition, while those in the lower 33% were considered to be in "poor" condition. A sample of 52 adult females in "good" condition and 53 adult females in "poor" condition were radio-marked with 26-g transmitter packages attached with a Dwyer harness (Dwyer 1972). High mortality of instrumented birds allowed another sample of adult females to be instrumented in February. The same procedure was followed in February, with 18 adult females in "good" condition and 17 in "poor" condition being instrumented. Radios were placed on females with condition indices less than 12.5 and greater than 14.0 in January. In February, the limits were set at 14.0 and 16.5. The February sample was treated separately from the January sample. In January, 100 each adult males, immature males, and immature females were marked with nasal saddles (Greenwood 1977) depicting their condition: , �ngood" ("+" saddle) or "poor" ("0" saddle). Birds in "average" condition were not marked. Approximately 60 adult females which were not instrumented were marked with nasal saddles. For the first three age-sex classes mentioned, criteria for determining which birds were marked were similar to those for radio instrumentation. Adult females marked with nasal saddles were in the middle 33% of the population and birds above and below the median were marked appropriately. These birds were used for focal animal time budget observations (Altmann 1974) when they could be located. Radio Tracking From 17 January through 1 May, instrumented birds were located from the ground or air weekly. Ground tracking employed scanning receivers with a 14element directional antenna or a 3-element null peak antenna. Aerial tracking utilized a Cessna 182 with 2-element "H" antennae attached to the wing struts (Dodge 1985). The date, time, location, and other comments were recorded with each contact. If a signal was slowed,-weak, or the signal had come from the same location for three consecutive contacts, hand-held antennas were used to locate and visually determine whether the bird was alive or dead. Locations were recorded and will be plotted using the UTM Coordinate system. Supplemental data, such as wind speed and direction, ambient temperature, and precipitation was recorded to be used as covariates in the survival and time budget analyses. Throughout the nesting period (May-July) birds were located as often as possible to monitor nesting chronology. If a duck was in the same location several consecutive times, indicating an incubating female, we attempted to find the nest site. If a nest was located, then weight, length, breadth and fertility of each egg and incubation stage was recorded. From this, an estimate of the nest initiation and probable hatching date could be made. Each nest of an instrumented female was marked with flagging. After the estimated hatching date, nest fate was checked. Nests of unmarked females were also measured in a similar manner, but no nest fate checks were made. Besides recording mortality of marked birds, we collected remains (usually just wings) from all waterfowl carcasses and attempted to determine the cause of death whenever possible. Most mallard wings were marked and frozen for later analysis. Additionally, a technique was developed to measure the lipid levels in carcass samples. First, a sample of 12 mallards were captured; some were killed immediately and others deprived of food for varying periods of time. These birds had the ulna excised, and were then subjected to whole body lipid extraction. The ulnas were cracked open and a seperate lipid extraction was performed on these bones. A relationship similar to that reported by Hutchinson and Owen (1984) was found between whole body lipids and ulna lipids. We were therefore able to assume that low levels of ulna lipid are indicative of starvation, but no quantitative measure of body condition can be made. Time Budget Time budget data were obtained using focal animal observations (Altmann 1974). Diurnal observations were made with a 40 power spotting scope, and nocturnal observations made with a light enhancing night vision scope. Birds was watched continuously for 10 minute intervals. Bird status (sex, marked or �72 not, and paired or not), habitat type, location, ambient conditions (temperature, wind speed,snowing or not, and location), date, and time observations started were recorded with the aid of a portable computer. Exact times of activity changes by the focal animal were entered on the portable microcomputer (Hensler et al. 1986). METHODS Yeight Changes - 1987-88 of Marked and Unmarked Mallards Mallards wintering on MVNWR were captured with Salt Plains bait trap$ (Szymczak and Corey 1976) and cannon nets (Dill and Thornberry 1950) during 4 trapping periods, from 11-18 January, 25 January-7 February, 8-22 February, and 29 February-4 March, 1988. Data on age, sex, weight, and wing length were obtained for all birds, and a condition index (Ringelman and Szymczak 1985) calculated for each. All mallards were banded with standard USFYS leg bands. During the 11-18 January trapping period, most mallards also had monel bands affixed to the shaft of primary feather VIII of both wings. Mean weights were plotted by trapping period and by week for each age-sex class. The UNIVARIATE Procedure (SAS Institute 1985) was used to test normality of weight and condition index distributions by age-sex class for each trapping period. Duncan's multiple range test was used to compare weight and condition index means by trapping period and age-sex class. To examine the relationship between weights and trapping method, weights of birds captured in cannon nets during the first trapping period were compared with weights of birds captured in bait traps during that same period. Yeights of birds recaptured within 7 days of initial capture were plotted as percent initial body weight for visual inspection. Linear regression was used to test for any relationship between duration of time between recapture and differences in initial and recapture weights and condition indices. Weight and condition indices of mallards originally banded in 1987 and recaptured in 1988 were examined to see if a bird in relatively "good" or "poor" condition in 1987 was in a similar relative condition class in 1988. Mean weight and condition index were placed in relative categories by subtracting mean weight and condition for the respective age-sex class from each respective trapping period. Instrumentation During the 11-18 January and 29 February-4 March trapping periods, adult females were assigned to 1 of 3 condition classes based upon their condition index relative to the distribution of the population in the SLV. Population distribution was determined by ordering the condition indices of adult females on the first trapping day. As in 1986-87, birds with a condition index in the upper 33% of the population (condition index >18.5) were considered to be in "good" condition, while those in the lower 33% (condition index <15.0) were considered to be in "poor" condition. A sample of 52 adult females in "good" condition and 51 in "poor" condition were radio-marked with 26 g transmitter packages attached with a back-pack harness (Dwyer 1972). Females in "average" condition were processed and released. , �73 High mortality of instrumented females during 1987 prompted us to anticipate high mortality again this season. Initially, we planned to reapply radios collected from dead females the week after the radio was collected. To reapply radios, we planned to trap weekly, then assign radios to adult females based upon their condition index and the cut-offs used during the 11-18 January trapping period. During this period, radio transmitters were reapplied to 23 adult females, 6 in "good" condition, 1 in "average" condition, and 16 in "poor" condition. Unfortunately, extremely high mortality of instrumented females, coupled with difficulties in trapping and changes in the condition of the population, made it impractical to reapply radios each week. Consequently, we decided in early February to reapply all remaining recovered radios during an intensive trapping effort 29 February-4 March. Procedures for the March trapping period were similar to those used in January. Condition index cut-offs were initially set at <12.5 for "poor" and >15.5 for "good" condition, but we were unable to trap a sufficient number of "poor" condition females. Therefore, later in the week, radios were also applied to females in "average" condition (12.5 to 15.5). Twenty-two females in "good" condition, 17 in "average condition, and 24 in "poor" condition were instrumented. Radio Tracking Radio tracking during 11 January through 26 March 1988 was conducted using both aerial search and ground crews. Procedures for both methods followed those used in 1986-87. When an instrumented birds was found dead, date, estimated date of death, likely cause of death, site description, UTM coordinates, and any other comments were recorded. Carcass remains found with the radio were collected and frozen for later analyses. From 27 March through 11 June, instrumented birds were visually located biweekly to determine nesting status. Nest parameters and estimated nest initiation date were measured as in 1987. Survival Rate Estimation The Kaplan-Meier method (Lee 1980) was used to estimate survival rates through 26 March of females in "good" and "poor" condition. Survival rate estimates were bracketed by plotting Kaplan-Meier estimates assuming censored birds had lived or died. Comparisons of weekly survival between condition classes were made using 2x2 chi-square tests. The square root of the weekly chi-square values were calculated to produce a z value, which was given a sign depending upon the difference in survival rates of "good" and "poor" condition classes. Pooled z values were used to test for an overall relationship between survival and body condition. Comparisons between condition class survival were made using the CATMon Procedure (SAS Institute 1982). Estimates of relative survival rates of wing banded birds in "good" and "poor" condition were made by a ratio of birds recovered divided by the number marked for each age-sex and condition class (Burnham et al. 1987:97). "Good" and "poor" condition were defined as the upper and lower third of the condition index distributions for the respective age-sex classes. Percentiles , �74 used to define "good" and "poor" condition were also changed to the upper and lower 25 percentiles and above and below the median to determine whether the percentile used to define condition classes affected inferences concerning any relationships between survival and condition class. Recovery rates of "good" and "poor" birds were tested with 2x2 chi-square tables by age-sex class, and for ages combined and overall. Wing band retention was checked on recaptured birds. Total mortality between 28 January and 4 February was estimated from differences in ducks counted from photos of roost areas taken at 154 m elevation. Collection of Instrumented Females At intervals of 2-3 and 4-5 weeks after the 11-17 January trapping period, 10 instrumented females (5 in "good" and 5 in "poor" condition) were collected for analyses on the effects of instrumentation on weight dynamics of free-living mallards. Similar collections were made 2-3 weeks after the March trapping effort. Birds were captured by hand-held nets, cannon nets, shotguns or rifles. Collected birds were immediately bagged and the following recorded: weight, wing length, condition index, blood smear sample, and any pertinent comments recorded before the carcass was frozen. Carcasses were later thawed and examined for signs of avian cholera. Wing and Carcass Collections Wintering waterfowl concentration areas and raptor perches on MVNWR were walked 5 days a week to collect remains of dead waterfowl. Species, age, sex, marsh unit where collected, and band numbers (if bands were present) were recorded. In addition, every fifth mallard wing was retained and frozen for later analysis of lipids in the ulna and Pasturella in the marrow of the radius. All wings found with wing bands were retained for the same analyses. A 4x13 chi-square table was used to test for a random distribution of mortality by age-sex class, using marsh units in which more than 5 wings were collected and for which the age and sex of carcass remains could be determined. Similarly, a 2xl3 chi-square table was used to test for a random distribution of recovered, banded-birds. Condition indices of wing-banded birds were analyzed for normality using the UNIVARIATE procedure (SAS Institute 1985), as were the condition indices of recovered wing-banded birds. Sample distributions were then compared to determine whether the distribution of recovered birds differed from the distribution of those marked. Wilcoxon rank sum tests were used to test for changes in the condition distribution of recovered birds compared with the condition distribution of birds banded. Wings retained for analyses were labeled and frozen in airtight bags. To expedite analysis, 100 ulnas were excised and cracked open at 1 end, then the exposed marrow was touched to a clean sheet of white paper to evaluate lipid content. The resulting blot was allowed to dry and then the paper was held up to a light source to evaluate the amount of residual lipid. A translucent blot indicated appreciable lipid ~esidue. If no visible blot remained, the ulna was presumed to have no lipids in it. To verify this technique, the 100 sample ulnas were cracked completely open, oven-dried, and lipids extracted for 6-8 hours using petroleum ether in a Soxhlet extraction apparatus. Blots were compared with extraction results to determine classification errors. A 2x2 chi-square table was used to determine if similar , �75 percentages of banded and unbanded birds appeared to have starved. The radius of each wing was cracked open and the marrow smeared on a clean microscope slide. The smears were then stained and examined for Pasturella. To determine the percentage of carcasses collected in areas searched, 25 carcasses with wing bands were placed around 3 water areas searched every 3 days by field technicians. Carcasses were placed in areas where carcasses had been previously recovered by technicians. Only the person planting the carcasses knew when and where they had been placed. Carcasses were checked 24 hours after placement to determine the initial rate of removal. Thereafter, carcasses collected by technicians were checked for wing bands of the planted carcasses, and wing band and marsh unit recorded. Fifteen carcasses were used to test whether Pasturella could be cultured from wing bones if the wing was not recovered immediately after death. One wing was removed form each carcass and placed in a wire cage to prevent scavengers from removing it, then exposed to the elements for 1 to 21 days. The carcass was frozen soon after the wing was removed. After the wing had been exposed for the required number of days, it was placed in the bag with the carcass from which it had been removed. Carcasses were later examined for gross lesions of cholera, and the liver cultured for Pasturella. Bone marrow from the radius of the exposed wing was cultured for Pasturella. Smears of the marrow were also made and stained to check for the presence of Pasturella. Standard Metabolic Rate Standard metabolic rates were measured using the open circuit oxygen consumption method. Captive adult mallards, which were held in a pen with indoor and outdoor access and subject to natural photoperiod, were supplied with food and water ad libitum. However, birds were denied access to food and water a minimum of 8 hours prior to taking measurements. Birds were placed in a darkened respirometer between 2000 and 0400 hrs and room air pumped through the chamber at the rate of 1.6 l/min. For analysis, air from the respirometer was.passed through columns of soda lime and drierite to remove CO2 and H20, resp.ectively. Oxygen content of the dried effluent air was measured with a Beckman F3 oxygen analyzer. Room air and a standard gas with 20.27% oxygen were used for calibration. All measurements were adjusted to standard temperature and pressure. Time Budget Diurnal time budget data were obtained from focal animal observations (Altmann 1974) made February through April 1988. Focal animals were initially selected by pointing a spotting scope at a mallard flock, then selecting a bird within the field of view. The bird's pair status and sex, as well as ambient temperature, wind speed, location, a~d habitat type, were recorded on a portable computer (Hensler et al. 1986). Ac~ivities were categorized as out of sight, rest, swim, walk, preen, dabble, tip~up, feeding on land, courtship, aggression, alert, fly, drink, bathe, and other. Percentage of time spent in the various activities was calculated and graphed. �76 Metebolizable Energy Ten captive adult mallards (5 males and 5 females) were fed the test food ~ libitum for a minimum of 21 days prior to testing to assure that gut morphology had adjusted to the diet (D. Johnson, pers. comm.). Birds were placed in isolation cages and trials were not started until all birds were eating and had returned to their approximate pre-trial weight. Prior to the 7-day trial, 8 meals were weighed to the nearest gram and divided into 2 portions, 1 to be fed in the morning and 1 at night. Meals exceeded ~ libitum intakes. Birds were weighed the day prior to trial initiation and again at trial completion. Ducks were denied access to food for 8 hours prior to starting the trial to allow food in the intestinal tract to be passed. Throughout the trial, birds were maintained at thermoneutrality. Dry weights of meals fed, dry weights of feces produced, and dry weights of food not eaten were recorded. Fecal samples from days 3-6 were combined to provide composite samples. After drying, composite grain and feces samples were frozen then later analyzed for gross energy with a bomb calorimeter. Nitrogen content was determined by the Kjeldahl method, then multiplied by 6.25 to give protein content. The difference between amounts of food fed and grain spilled was considered intake. Differences between intake and amount excreted was termed apparent digestible energy (Miller and Reinecke 1984). Raptor Counts Raptor predation was identified as an important mortality factor of mallards wintering on MVNWR in 1987. To monitor trends in the abundance of raptors utilizing MVNWR, a survey route was established which included most perches available on MVNWR. The route was driven weekly, from 14 January through 24 March 1988, in the early morning before human activity disturbed the birds. The route was not driven when visibility was less than 1.5 km. Raptors were classified by species, and bald eagles (Haliaeetus leucocephalus) further classified as either adult or immature. Trends in raptor populations were identified from a plot of the number of raptors observed from the route. Reproduction In addition to monitoring the instrumented birds through the nesting season, a survey route was established on MVNWR and driven weekly from 19 April through 15 June for the purpose of enumerating and the number of lone male and paired male mallards. Broods encountered were aged (Gollop and Marshall 1954), and from these data hatching dates were estimated. Weather Daily temperature maxima and mlnlma, and snow cover were obtained from National Oceanic and Atmospheric Administration reports (National Oceanic and Atmospheric Administration 1988a, 1988b, 1988c) for the Alamosa airport reporting station. Temperatures and snow on the ground from 1 January- 31 March were plotted to determine periods of weather change. , �77 RESULTS - 1986-87 One hundred and eighty-nine mallards were trapped, measured, and banded on the Monte Vista National Wildlife Refuge 17-20 December 1986. From 5-15 January 1987, 121 adult females, 354 adult males, 169 immature females, and 334 immature males were weighed and banded. More males and immature females were captured but not processed due to time limitations. All adult females captured were processed. For the 20-22 February trapping period 56 adult females, 165 adult males, 69 immature females, and 216 immature males were processed. Similar to the January trapping period, only a limited number adult and immature males, and immature females were processed. Condition Indices Mean condition indices for for all birds trapped were greater December than in January or February (Table 1). Female condition in Table 1. Mean monthly condition indices of mallards trapped in Salt Plains Bait Traps on the Monte Vista National Wildlife Refuge during winter, 1986-87. Month Mean Condition Adult Female Indices (Standard Deviation) Adult Male Immature Female Immature Male Dec 19.9 (4.4) 18.0 (3.6) 17.1 (4.6)' 16.5 (3.8) Jan 13.6 (4.7) 12.4 (3.9) 13.7 (4.0) 11.5 (4.2) Feb 14.8 (3.7) 10.4 (3.5) 13.3 (3.8) 9.7 (3.6) indices were higher than similar age class males (P<O.OOOI for both adults and immatures, ~-testt 895 df for adult comparison and 973 df for immature comparison). Adult males were in better condition than immature males (P<O.OOOl, ~-test, 1,348 df). Adult and immature female condition indices were similar (P-0.17, ~-test, 520 df). Condition indices for the four age-sex classes were greater for birds trapped in December than in January (P<O.OOOl for adult and immature males and adult females, P<0.0007 for immature females, ~-test, 429 df, 394 df, 137 df, and 197 df, respectively). Condition indices were not different between January and February females (P>0.05, ~-test, 134 df for adult and 236 df for immature). January trapped males had higher condition indices than February males (P<O.OOOl, ~-test, 516 df for adult and 506 df for immature). Recaptured individuals showed similar patterns of weight loss between the December-January and January-February trapping periods (Table 2). , �78 Table 2. Mean body mass changes of birds banded one trapping period and recaptured during the next trapping period. Month of Capture First Second Body Mass Age Sex n first ~apture Recapture ~ifference December Janauary A M 2 1162 973 -189 December January I F 4 922 841 -81 December January I M 4 1137 1024 -113 January February A F 4 955 928 -27 January February A M 24 1055 977 -78 January February I F 1 1014 862 -152 January February I M 22 1020 888 -132 Condition indices were approximately normally Shapiro-Wilk statistic) for all age-sex classes and periods with the exception of adult females trapped indices of adult females trapped in January differed n-l2l) due to kurtosis rather than skewness. distributed (P>0.05, the three main trapping in January. Condition from normal (P-0.026, Radio Tracking Most instrumented birds were contacted at least weekly (Table 3) until spring thaw began in mid-March. As ponds thawed throughout the SLV the birds became more dispersed and time-consuming to locate from the ground, so we became more dependant upon locating birds from the air and making ground checks to determine a bird's status. Flights from the Rio Grande River south to LaJara were generally most productive in terms of number of birds located (Table 4). No birds were located in the area bounded by highway 285 on the west, highway 17 on the east, 5 North Road on the South, and Highway 112 on the North. This highly agricultural area has little habitat other than irrigation ditches. �79 Table 3. Number of birds contacted by week from 18 January 1987 through 2 May 1987. Radio-tracking was not directed at finding as many birds as possible after 1 May since we attempted to determine the nesting stage of birds near MVNWR. Dates Number of Birds Contacted January Marking February Marking 18-24 Jan 69 25-31 Jan 38 1-7 Feb 42 8-14 Feb 59 15-21 Feb 10 22-28 Feb 20 32 1-7 Mar 22 11 8-14 Mar 29 25 15-21 Mar 31 24 22-28 Mar 7 5 29 Mar-4 Apr 28 21 5-11 Apr 19 23 12-18 Apr 18 17 . 19-25 Apr 17 15 5 7 26 Apr- 2 May �80 Table 4. Dates, locations, the 1986-87 field season.1 Date of Flight and numbers Locations of birds located from aircraft Flown during Number of Birds Located 2/6/87 Valley-wide 23 3/3/87 Russell Lakes South to LaJara 36 3/19/87 Rio Grande South to LaJara 13 3/26/87 MVNWR North to Russell Lakes 4/21/87 Valley-wide 29 5/7/87 MVNWR South to Antonito 19 5/11/87 Rio Grande South to LaJara 13 5/26/87 Rio Grande North to Saguache 6 6/4/87 Rio Grande Upstream to Creede 6 6/12/87 San Juans and MVNWR (all on MVNWR) 6/17/87 Rio Grande South to Antonito 7 6/18/87 Bueno Vista to Gunnison 4 6/19/87 Rio Grande North to Saguache 8 6/30/87 Rio Grande South to Antonito 11 9 18 Throughout the field season, only 1 bird was not contacted at least once after release. That bird was instrumented in February, had a condition index of 12.3, and was shot in October in northern New Mexico. Eight birds were not contacted in the Valley after mid-February, including 5 in "good" and 3 in "poor" condition. Twenty birds were not contacted after April. Four of these were instrumented in February (all in "good" condition). Birds instrumented in January which were not contacted after this t-ime included 8 in "good" and 9 in "poor" condition. Two birds which were subsequently located in the Gunnison Drainage were both in "good" condition and instrumented in January_ One was last contacted in the Valley on 27 February and the other was last contacted 20 ApriL , �81 Mortality Instrumented birds experienced high mortality. Twenty-eight of the females in "good" condition and 37 of those in "poor" condition when instrumented in January died during the period 17 January through 4 July, for minimal mortality rates of 0.538 and 0.699, respectively (Fig. 2). Mortality rates for females instrumented in February were 0.277 for females in "good" condition and 0.588 for females in "poor" condition. Two peaks in mortality for females instrumented in January occurred (Fig. 3). Causes of mortality during this period included eagle predation and starvation. Most mortality occurred on or around MVNWR (Figs. 4 and 5). Table 5. Daily survival rates for the period 18 January through 4 July 1987 obtained through MICROMORT for adult female mallards instrumented 5-15 January 1987 in the San Luis Valley. Condition Class Mean Daily Survival Rate Standard Error Good 0.99612 0.00008 Poor 0.99057 0.00011 Table 6. Daily survival rates obtained from MICROMORT of adult female mallards instrumented 5-15 January and 20-22 February 1987 in the San Luis Valley. Daily survival rates are calculated for the period 22 February to 4 July 1987. Condition Class Mean Daily Survival Rate Standard Error Jan Good 0.99761 0.00008 Jan Poor 0.99433 0.00016 Feb Good 0.99748 0.00011 Feb Poor 0.92308 0.00234 �82 State of Colorado Figure 1. Location of the San Luis Valley, Colorado. w > ::::; <[ I- 100 100 80 80 60 60 40 40 JAN HIGH 20 20 0 0 100 100 JAN LOW Z w a: w e, 0 80 ~ 80 60 I- 60 40 ~ 40 20 ~ 0 FEB HIGH 20 28 JUN 18 JAN 0 FEB LOW 18 JAN WEEKL Y INTERVAL Figure 2. Proportion of Instrumented by week, for the 1986-87 females potentially field season. alive, 28 JUN �83 ~ FEBRUARY HIGH [ ::}!:':j FEBRUARY LOW w D JANUARY HIGH > o __ JANUARY LOW 12 UJ a: w o W 8 a:: LL o a:: w m ~ => Z 4 o 18 JAN 28 JUN WEEK OF RECOVERY Figure 3. Mortality of Instrumented femal. mallards by week and condition class, 1986-87. �84 70,---------------------------------------------------~ Monte Vista ® + + Alamosa + + ® • • La Jar~ford ® @) 104-----~------r_----~----~------~----~----~----~ 115 125 105 '95 135 Figure~. Spatial distribution of recoveries for female mallards instrumented in January, 1987. Squares and pluses denote good and poor condition females, respectively. 70~--------------------------------------------------__, of: Monte Vista (!) 60 + AJamosa 50 + 40 @l. • 30 La Jara ® Sanford ® 20 Manassa @ 104-----~----~r-----~----,_----_r----_.----~r_--__, 125 115 95 105 135 Figure S. Spatial distribution of recoveries for female mallards instrumented in February, 1987. Squares and pluses denote good and poor condition females, respectively. , �85 Mea~ daily survival rates obtained from MICROMORT for the period of 18 January through 4 July 1987 were higher for birds in "good" than "poor" condition in January (Table 5). Females in "good" condition in either January or February did not survive at different rates (Table 6). Females in "poor" condition in January survived better than females in "poor" condition in February, yet poorer than females in "good" condition in either month (Table 6). Logistic regression of survival based upon month of banding or condition index, using the Statistical Analysis System (SAS) Categorical Modeling procedure (CATMOD), identified a significant affect of condition index on survival (P-O.0103). No month of marking or month-condition index interact"ion was found (P>O.5). As with mortality estimates based upon the instrumented birds, band recoveries suggest higher mortality of birds in "poor" condition at the time of banding (Table 7). Recoveries of nasal saddles also suggest this, but only 4 nasal saddles were recovered. (3 "0" and 1 "+"). Predation and scavenging made it extremely difficult to locate bands or nasal saddles. Table 7. Recoveries of banded birds based upon month of banding and condition. Number Month of Banding Condition Class Banded Recovered % December Good Poor 126 29 4 2 3.1 6.9 January Good Poor 346 462 5 13 1.4 3.8 February Good Poor 55 413 1 2 1.8 0.5 Reproduction Of the 31 birds which were known to be alive and in the" San Luis Valley in early May, only 2 females were found incubating. The first nest found was produced by a female instrumented 22 February in "good" condition (CI of 16.3). The nest was located in an alfalfa field about 200 m south of Highway 160 and 1 km west of Monte Vista. Her clutch consisted of 12 eggs with a mean egg mass of 50.4 g. The nest was found too early to determine fertility of the eggs. It was initiated about 20 May, and was predated in mid-incubation by an avian predator. The only other nest located was produced by a female with a condition index of 8.5 on 15 January. The nest was about 30 m up a talus slope north of the Rio Grande about 2 km downstream from South Fork. Her clutch consisted of �86 8 fertile eggs with a mean mass of 47.8 g. She initiated this clutch about 30 May, and was also unsuccessful in hatching any eggs. It appeared as if avian predators destroyed the nest just before hatching. Though many females remained on the MVNWR through the nesting period, no nest of instrumented birds were found. Four females appeared to be in the laying stage (in nesting cover at various times, generally early morning), but the nests must not have survived into incubation. Data was recorded for 18 nests of unmarked females. Eleven of these nests were located in cattails (~ latifolia), often over water, 2 were in flooded Baltic rush (Juncus balticus), 2 were in mixed Baltic rush and greasewood (Sarcobatus vermiculatus), and 2 were in dry Baltic rush. Mean clutch size of 10 incubated nests was 8.4. Mean egg mass of 6 clutches, containing 39 eggs known not to have been incubated was 48.4 g. Time Budget Time-budget observations were limited to only 27.5 hours of daylight and 2.3 hours of night observations. Females comprised about 40% of the total. Although most observations were planned for nasal-saddled birds, these marked birds were located too infrequently. Generally, birds alternated between foraging in the barley fields and resting on water areas. Night time observations revealed that the major activity was resting. The extensive amount of daylight hours spent foraging in the fields suggests that these birds have difficulties in obtaining enough energy for maintenance. RESULTS Weights Changes of Marked and Unmarked - 1987-88 Mallards Weight and condition indices were approximately normally distributed. Overall, adult males weighed more than immature males, which weighed more than aduit females, while immature females weighed the least (Duncan's Multiple Range test, 2679 df, P<O.Os). Adult females were in the best condition, followed by adult males, immature females, and immature males in the poorest condition (Duncan's Multiple Range test, 2679 df, P<O.Os). Mallards weighed more during the 10-18 January trapping period than the other trapping periods (Duncan's Multiple Range test, 2679 df, P<O.Os). Birds were in better condition during the first trapping period, followed by the second period, then the fourth period, with the poorest condition indices recorded during the third trapping period (Duncan's Multiple Range test, 2679 df, P<O.Os). Adult males weighed more than immature males, which weighed more than adult females during each trapping period (Table 8). During the 25 January-7 February trapping period, adult and immature females were similar in weight, otherwise adult females weighed more (Table 8). Generally, adults were in better condition then immatures and females in better condition than males (Table 9). �87 Table 8. Mean and standard error for weights ~f mallards captured in the San Luis Valley. Letters a-d indicate significant differences among means (P-0.05) within an age-sex class. Letters w-z compare among age-sex classes within a trapping period. Trapping Period Adult Immature Male Female 10-18 Jan 1l01(±98)aw n-783 963(±76)ay n-274 1029(±85)ax n-495 915(±79)az n-390 25 Jan7 Feb 1032(±74)bw n-34 898(±80)bcy n-43 953(±97)bx n-16 889(±70)aby n-45 8-22 Feb 1002(±65)bw n-46 889(±47)cy n-18 940(±49)bx 827(±78)cz n-22 1019(±76)bw n-152 927(±70)by n-9l 946 (±71)bx n-144 23 Feb4 Mar Female Male n=Ll, 871 (±80)bz n-122 Table 9. Means and standard errors for condition indices of mallards captured in the San Luis Valley, Colorado, during 4 trapping periods in January-March 1988. Letters a-d indicate significant differences among means (P-0.05) within an age-sex class. Letters w-z compare among age-sex classes within a trapping period. Trapping Period - Adult Male Immature Female Male Female 10-~8 Jan l5.4(±4.0)aw l5.9(±4.4)aw l3.9(±3.8)ay l4.8(±4.3)ax 25 Jan7 Feb l3.3(±3.4)bw l3.0(±5.2)bcw 10.6(±4.6)bx l4.0(±3.3)abw 8-22 Feb 10.9(±3.0)cw l2.l(±3.l)cw 23 Feb4 Mar l2.l(±3.5)bcx l4.6(±4.0)abw 9.7(±2.9)bw 9.7(±4.9)cw 10.0(±3.8)by l2.6(±4.5)bx Two hundred and twenty-one mallards wer~ captured with cannon nets and 1,426 captured in bait traps during the first trapping period. Mallards captured in cannon nets weighed more than birds captured in bait traps (Duncan Multiple Range Test, 1,639 df, P<0.05, Table 10). Analysis of variance indicated significant trap method (F-14.55, P-O.OOOl, 1 df), age (F-59.52, P-O.OOOl, 1 df), sex (F-8.56, P-0.0035, 1 df), and age-sex interaction (F-322.42, P-O.OOOl, 1 df) effects. Age-trap, sex-trap, and age-sex-trap interactions were not significant. , �88 Table 10. Mean weights Luis Valley, Colorado, (SE) of mallards captured 11-18 January 1988 in the San in cannon nets and Salt Plains bait traps. Trap Method Male Female Male Female Cannon 1128(86) 970(72) 1045(78) 952(73) n-88 n-63 n-38 n-32 1090(101) 961(79) 1026(85) 913(78) n-537 n-191 n-380 n-318 Net Bait Trap Adult Immature Mallards banded and recaptured during the 1988 season had lower weights and condition indices at recapture than at the initial capture (F-16.957, P<O.OOOl, 205 df and F-5.833, P-0.0166, 205 df, for weight and condition indices, respectively). Although mean weights of unmarked birds declined with time, recaptured immatures weighed consistently less than unmarked birds (Figs. 6 and 7). Adult male recaptures were similar to unmarked birds (Fig. 8). Too few banded adult females were recaptured to make any comparisons with unmarked birds captured (Fig. 9). Generally, mean weights of birds recaptured within 7 days of initial marking weighed less than at initial capture (Figs. 10 and 11). Weight and condition indices were not different between initial capture and recapture (F-3.026, 115 df, P-0.0846, 115 df and F~0.9l7, 115 df, P-0.3402, for weight and condition index, respectively). One hundred and twenty mallards banded in 1986-S7 and recaptured during 1988 were used in analyses of relative condition be~ween years. Only 4 females classified as adults in 1986-87 were recaptured in 1988, and no relationship between condition in 1986-87 and 1988 was detectable (F-O.OOO, P>O.9892, 3 df). A positive relationship (Fig. 12) was detected for 1986-87 adult males (F-8.934, P<0.0044, 49 df) and 1986-87 immature males (r-18.320, P<0.0002, 30 df, Fig. 13) Immature females captured in 1986-87 and recaptured in 1988 did not show a relationship (Fig. 14) between relative condition in 1986-87 and 1988 (F-l.587, P<0.2634, 6 df). Survival Rate Estimation Kaplan-Meier 10 week survival estimates for females instrumented in January and 3 week estimates for females instrumented March are given in Table 11. Cumulative survival rate estimates for January-marked females were not changed much when censored birds were assumed to ha~ lived (Fig. 15) or died (Fig. 16), because most birds were contacted weekly (Appendix B). Most instrumented females died within 3 weeks of release (Fig. 17). �_._ UNMARKED - .- RECAPTURE 1200 1150 " OJ •... ~ :x: 1100 (144) 1050 CJ iii (11) 1000 ~ >- 150 Q 0 --- --- --- --- ---- .----- --- •. • 00 CD 850 (21) ( 1) 800 750 700 10-18.1AN 25.1AN-7 FEB 8-22 FEB 23 FEB-4 MAR TRAPPING PERIOD Figure 6. Mean weights of Immature male mallards trapped In the San Luis Valley, Colorado, during 4 trapping periods In 1988. Sample sizes are shown In parentheses. _._ UNMARKED -.- RECAPTURE 1200 1150 "om ...... ',00 ~ 1050 :x: CJ iii 1000 (45) (122) ~ >- .50 0 CD .00 Q 850 ... 800 _---- ------- 750 ( 10) (3) 700 10-18 .IAN U.lAN-7 FEB 8-22 FEB 23 FEB-4MAR TRAPPING PERIOD Figure 7. Mean weights of Immature female mallards trapped In the San Luis Valley, Colorado, during 4 trapping periods In 1988. Sample sizes are shown In parentheses. , �90 _._ (783) UNMARKED 1200 - • •.... 1150 ~ 1100 ~ 1050 - RECAPTURE (4.) (152) CJ iii ~ 1000 >- .50 o .00 C (30) (5) ID ISO 100 760 700 10-18 JAN 26 JAN-7 FEB '-22 23 FEB-4MAR FEB TRAPPING PERIOD Figure 8. Mean weights of adult male mallards trapped In the San Luis Valley, Colorado, during 4 trapping periods In 1988. Sample sizes are shown In parentheses. _._ UNMARKED 1200 •.... Cl 1150 •... 1100 ••• x: 1050 iii 1000 - • - RECAPTURE (11) CJ ~ >- 150 0 .00 C ". ID 860 800 ( 1) 760 -- . (5) 700 10-18 JAN 26JAN-7 FEB 8-22 FEB TRAPPING PERIOD Figure 9. Mean weights of adult female mallards trapped In the San Luis Valley, Colorado, during 4 trapping periods In 1988. Sample sizes are shown In parentheses. UFEB-4MAR �91 ~ ::z:: CJ iii 110 ~ 105 r- • aDULT o IMMATURE ~ > o o (23) (24) m (33) ..J oct Z 100 (5) I a:o (12) ~ Z au a:: au (I) ~ s o . (1.) .0 ~ (4) D. I 85 , J I 2 5 DA YS BETWEEN RECAPTURE Figure 10. Weights (as % of Initial capture weight) of adult and Immature male mallards recaptured within 7 days of Initial capture In the San Luis Valley, Colorado, January-March 1988. Sample sizes are shown In parentheses. ~ ::z:: CJ iii 110 ~ aDULT D. IMMATURE ~ > 0 0 105 m ..J oct Z 100 (3) (3) (; a:0 f .5 ~ Z au 0 a:: au (4) .0 I (4) D. 85 1 2 3 DA YS BETWEEN " RECAPTURE Figure 11. Weights (as % of Initial capture weight) of adult and Immature female mallards recaptured within 7 days of Initial capture In the San Luis Valley, Colorado, January-March 1988. Sample sizes are shown In parentheses. 5 �92 18 CO CO 0) •.. I ~ Z • 0 t= 3 is z 0 ,·O.~O 18CONDITION.O.~'O(17CONDITION)-O. 710 12 0 o W > t= « ...I W -3 -. -I c:t -12 -. -15 -15 -12 -I -3 0 3 1 I 12 15 I 12 115 RELATIVE CONDITION IN 1986-87 Figure 1 2. Relative condition 1986-87 and recaptured Indices of adult males captured In In 1988 In the San Luis Valley, Colorado. 15 88CONDITION.0.'U(B7CONDITION)-0.813 co co ,·0.12 12 I 0) <pO ~ I Z 0 3 is z 0 t= 0 o W > t= « ...I W c:t -3 . -I - -12 -15 -15 -I -12 -. -3 o 3 • RELATIVE CONDITION IN 1986-87 Figure 13. Relative condition 1986-87 and recaptured Indices of Immature males captured In In 1988 In the San Luis Valley, Colorado. , �93 18CONDITION-0.3.0(17CONDITION)+0.3.. 12 CO CO en .•.. i: z 0 ,-0 •• ' • • • 3 ~ 0 Z 0 (J W > ~ _, oCt w a:: 0 -3 • -. -. -12 -. -15 -15 -12 -I o -3 3 I • 12 RELATIVE CONDITION IN 1986-87 Figure 14. Relative condition Indices of Immature females captured in 1986-87 and recaptured in 19881n the San Luis Valley, Colorado. ~ -GOOD· CONDITION 0.' _, > sa:: - -. 0.8 -. ·POOR" CONDITION oCt =» w 0.7 0.1 (h > ~ 0.5 oCt _, 0.• =» ~ =» (J 0.3 0.2 0.1 0 0 WEEK Figure 15. Cumulative females Instrumented weekly survival rates and standard errors of In January, assuming censored birds survived. , �94 -e- ·GOOD" CONDITION - -. ·POOR" CONDITION 0.' _, 0.8 - ~ > 5= 0.7 :;) 0.' a: (/) IU > i= 0.5 :E :;) 0.3 _,:;)~ o 0.• 0.2 0.1 0 0 WEEK Figure 16. Cumulative females Instrumented weekly survival rates and standard In January, assuming censored errors of birds died. 25 (/) 0 a: iii 20 0 ~ IU 0 u, 15 0 a: IU CD 10 Z 5 :E :;) 0 WEEKS AFTER INSTRUMENTATION Figure 1 7. Number of weeks between Instrumentation and recovery 97 Instrumented female mallards found dead In the San Luis Valley, Colorado, In 1988. of , �95 Table 11. Weekly Kaplan-Meier survival estimates and standard errors for adult female mallards instrumented in the San Luis Valley, Colorado, 1988. March Marking "Poor" "Good" Week January Marking "Good" "Poor" 13-20 Jan 0.867(0.047) 0.727(0.057) 21-27 Jan 0.676(0.062) 0.875(0.068) 28Jan-3Feb 0.806(0.062) 0.875(0.059) 4-10 Feb 0.885(0.063) 0.800(0.056) 11-17 Feb 0.864(0.061) 0.926(0.061) 18-24 Feb 1. 000(0.066) 0.846(0.055) 25 Feb-2 Mar 1. 000(0.066) 0.905(0.057) 3-9 Mar 1. 000(0.066) 0.947(0.057) 0.833(0.070) 0.857(0.071) 10-16 Mar 0.947(0.064) 0.944(0.056) 0.923(0.096) 0.941(0.086) 17-23 Mar 1.000 (0.064 ) 1.000(0.058) 0.909(0.113) 0.875(0.107) In only one week did survival rates differ by condition class; from 20-27 January significantly more "good" condition females died (X2-S.l3, 1 df, P~0.023s). During the other weeks, as well as the pooled chi square, no significant differences in survival were detected between birds in "good" and "poor" condition. Similarly, CATMOD (P-0.2143) and the z-test (z-0.737) failed to detect a difference in survival between the two groups. Since no differences were detected between survival rates of the 2 condition classes, estimates were pooled (Table 12, Fig. 18). Relative survival ratios (Table 13) were not different from 1:1 overall, by age, or for any age-sex class of wing banded birds (Table 14). When the cut-off values defining "good" and "poor" condition are changed to above or below the median and the upper and lower 25 percentiles, only immature females show differences in the number of recoveries of birds in "good" and "poor" condition (Table 15). More immature females in "poor" condition were recovered using both definitions (Table 15). As with instrumented females, most wing bands were collected within three weeks after release (Figs. 19-22). Wing band loss probably did not differ according to condition class, but age did effect band loss. Of 65 wing-banded birds recaptured and checked, 18% of the birds had lost 1 wing band, but none of the recaptured birds had lost both wing bands. Of the 24 birds for which age and sex were recorded, 6 had lost I wing band. All 6 birds were immatures, 3 males and 3 females. Ten adult males, 6 immature males, and 2 immature females had retained both bands. No adult females were checked. �96 0.' ..J - 0.11 ~ ,.OOLED SURVIVAL ct > 0.7 ~ 0.' > i= 0.11 ::E ~ 0.3 sa::: en w ct ..J ~ 0 0.4 0.2 0.1 0 0 WEEK Figure 18. Cumulative weekly survival rates and standard errors of pooled survival estimates. 20 0 15 W t- o W ..J ..J 0 0 10 ~ ·POOR" CONDITION ~ ·AVERAGE· •• CONDITION ·GOOD" CONDITION a::: w m ::E ~ Z II o 7 • • WEEK OF COLLECTION Figure 19. Number of weeks between release and recovery banded adult female mallards In the San Luis Valley, Colorado, of wlng- 10 11 �97 20 ~ 0 III t'8Z8 11 •• t0 III ..I ..I 0 0 10 ·POOR" CONDITION ·AVERAGE" -GOOD· CONDITION CONDITION a: III CD ~ ::l Z I o WEEK OF COLLECTION Figure 20. Number of weeks between release and recovery of wlng- banded adult male mallards In the San Luis Valley, Colorado, 1188. 20 l?Z3 0 15 ~ III •• t0 W ..I ..I 0 0 10 ·POOR" CONDITION ·AVERAGE" "GOOD" CONDITION CONDITION a: w CD ~ ::l Z I o WEEK OF COLLECTION Figure 2 1. Number of weeks between release and recovery of wlng- banded Immature female mallards In the San Luis Valley, Colorado, 1188. , �98 20 PZa 0 •• W t- 0 _,_, W 0 0 10 CONDITION ·AVERAGE"' ~ 111 ·POOR- ·GOOD- CONDITION CONDITION a:: w m ~ :l Z I o WEEK OF COLLECTION Figure 22. Number of weeks between release and recovery of wlngbanded Immature male mallards In the San Luis Valley, Colorado, 1888. 1200 ____ -Good" 1100 t- X 1000 c" (2) iii ~ .00 g aoo >o ------ ~ (3) 700 .00 o 7 14 21 28 35 DAYS BETWEENINSTRUMENTATION AND COLLECTION Figure 23. Mean body weight and standard error for females collected less than 14 days, 15-28, and 30-33 days after Instrumentation. Also given are mean body weights at capture for the collected birds. Sample sizes are shown In parentheses. , �99 Table 12. Pooled weekly Kaplan-Meier survival estimates and standard errors for adult female mallards instrumented in the San Luis Valley, Colorado, 1988. Week January Marking March Marking 13-20 Jan 0.797(0.037) 21-27 Jan 0.776(0.046) 28 Jan-3 Feb 0.841(0.043) 4-10 Feb 0.843(0.042) 11-17 Feb 0.895(0.043) 18-24 Feb 0.923(0.043) 25 Feb-2 Mar 0.953(0.044) 3-9 Mar 0.974(0.044) 0.845(0.050) 10-16 Mar 0.946(0.043) 0.932(0.064) 17-23 Mar l.000(0.043) 0.892(0.078) Table 13. Relative survival ratio, by age and sex, of mallards wintering the San Luis Valley, Colorado, and marked with monel wing bands. Condition Class Number Banded Number Recovered Survival Ratio SE Age Sex Adult Male Good Poor 251 246 21 22 l.069 0.312 Female Good Poor 82 80 6 l.879 0.907 11 Pooled Good Poor 333 326 27 33 l.248 0.309 Male Good Poor 169 165 13 20 l.576 0.534 Female Good Poor III 12 18 l.586 0.550 105 Good Poor 280 270 25 38 l.576 0.383 Good Poor 613 596 52 71 1.404 0.243 Immature Pooled Overall in , �100 Table 14. Comparisons of recoveries of wing-banded birds in the upper and lower thirds of the condition index distribution at the time of banding. Number of recoveries are given in Table 7. X2 Value Comparison df P Power of Test Adult Male 0.0439 1 0.8340 0.056 Adult Female 1.4454 1 0.2293 0.264 Immature Male 1. 5083 1 0.2194 0.271 Immature 1. 3667 1 0.2424 0.267 0.6730 1 0.4120 0.147 2.8497 1 0.0914 0.475 3.1710 1 0.0750 0.508 Adult, Female Pooled Immature, Pooled Overall Table 15. Comparisons of recoveries of wing-banded birds using different portions of the condition index distribution to define "good" and "poor" condition. All tests are Chi-Square tests with 1 df. Power refers to the power of the statistical test. Above/Below P X2 Comparison Median Power Upper And Lower 25% X2 P Power Adult Male 0.389 0.532 0.11 0.199 0.656 0.08 Adult Female 2.688 0.101 0.45 0.820 0.365 0.17 Immature Male 0.003 0.959 0.05 0.057 0.811 0.06 Immature 4.953 0.026 0.74 5.656 0.017 0.82 Female Adult, Pooled 0.097 0.756 0.06 0.014 0.907 0.05 Adult, Immature 2.884 0.090 0.48 2.843 0.092 0.48 1.952 0.162 0.34 1. 722 0.189 0.31 Overall �101 The 28 January duck count from aerial photographs was 23,570, with possibly 590 more birds present but untallied. The 4 February count was 20,864 with up to 1,475 more ducks untallied. These counts produce a 7-day survival rate estimate of 0.885 (or 0.925 in the additional untallied ducks are included). The corresponding survival rate of instrumented females was 0.841 (±0.043). These rates are similar (z-1.023). Assuming a rate of 0.90 (or 0.985/day) based on aerial counts, under a regime of constant mortality, this corresponds to a 70-day survival rate of 0.349. This rate is bracketed by the 0.3-0.4 70-day survival rate experienced by instrumented females (Figs. 15 and 16). Collection of Instrumented Females Eight instrumented females in the "good" condition class and 7 in the "poor" condition class were collected during the first period. During the second period, 5 females of each condition class were collected. During the third period, 3 "good", 4 "average", and 3 "poor" condition females were collected. Mean weights of all birds collected up to 33 days after instrumentation declined (Fig. 23, Table 16). None of these birds exhibited gross lesions typical of fowl cholera. During the first period, females collected an average of 15.0 (±2.2) days after instrumentation lost a mean of 238 g (range 59 g to -410 g). Condition indices declined an average of 15.4 (range 4.4 to -27.8). Weight and condition of collected females did not differ according to condition class (Table 17). Females of the two condition classes showed similar weight losses and declines in condition indices (Table 17). Table 16. Mean and standard error at time of collection for weight, condition index, and weight and condition index changes of instrumented female mallards collected 2-3 and 4-5 weeks after instrumentation. Collection Period Condition Class 1 "Good" 752(117) 2.3(8.0) -298(86) -18.9(6.8) "Poor" 720(115) 1.8(6.8) -170(106) -11.5(7.2) "Good" 920(63) 13.1(3.3) -151(27) - 8.3(2.1) "Poor" 799(30) 6.1(1.5) -112(23) - 7.4(1.3) "Good" 805(31) 7.0(2.3) -186(35) -12.0(2.3) "Average" 784(29) 5.4(3.3) -152(30) -10.2(2.6) "Poor" 817(59) 7.6(4.0) - 54(31) - 3.20.7) 2 3 Body Weight Condition Index Weight Change Condition Change �102 Table 17. Comparisons of weights and condition indices at collection and weight and condition index changes for adult female mallards instrumented in "good" and "poor" condition 11-18 January 1988 and collected 2-3 and 4-5 weeks after instrumentation. Collection Period 2-3 weeks 4-5 weeks Comparison ~ value df P Weight 0.53 13 0.6069 Condition 0.11 13 0.9154 Weight Change -2.05 13 0.0611 Condition -1.61 13 0.1323 Weight 3.81 8 0.0052 Condition 4.31 8 0.0026 Weight Change -2.16 8 0.0632 Condition -0.74 8 0.4799 Change Change During the second period, females were collected 28.3 (±2.6) days after instrumentation and lost an average of 132 g (range -71 g to -177 g) and condition indices declined an average 10.1 (range -4.4 to -10.1) in the 4-5 weeks. "Good" condition females weighed more (Table 11) and were in better con~ition (Table 17) at the time of collection than "poor" condition females. Weight declines were similar between the two groups, as were declines in condition (Table 17). Females instrumented in March were collected 18.1 (±4.5) days after instrumentation and had lost an average of 133 g (range -32 g to -209 g). Condition indices declined 8.6 (range -1.4 to -14.4). Weight and condition changes differed between "good" and "poor" condition females and between "average" and "poor" condition females (Table 18). During the first period only 1 banded adult female was recaptured. She weighed 888 g and had a condition index of 13.4. Unmarked adult females captured during this period weighed more than the collected birds (~--6.45, 51 df, P-0.0001) and were in better condition (1--6.51, 51 df, P-0.0001). The only banded adult female recaptured during the second collection period weighed 795 g and had a condition inde~ of 3.4. Four unmarked adult females were captured during this period. Neither their mean weight of 805 g nor their condition was different from the mean weight or condition of the collected birds (1-0.96, 12 df, P-0.3574 and 1-0.78, 12 df, P-0.449l, respectively). �103 Table 18. Comparisons of weights and condition indices at collection and weight and condition index changes for adult female mallards instrumented in "good", "average" and "poor" condition 29 February-4 March 1988 and collected 2-3 weeks after instrumentation. Condition Classes Compared Comparison "Good" and "Poor" Weight and "Poor" P 4 0.7702 -0.24 4 0.8230 Weight Change -4.87 4 0.0082 Condition -4.35 4 0.0122 -0.91 5 0.4028 -0.73 5 0.5001 Weight Change 1. 38 5 0.2257 Condition 0.81 5 0.4541 -0.98 5 0.3708 -0.83 5 0.4455 Weight Change -4.18 5 0.0087 Condition -3.47 5 0.0179 Index Change Weight Condition "Average" df -0.31 Condition "Good" and "Average" .! value Index Change Weight Condition Index Change Wing and Carcass Collections An outbreak of avian cholera began on MVNWR in early December 1987, and prompted closure of the refuge to waterfowl hunting before the scheduled season closure. Between 4 January and 1 April 1988, we collected 4,604 remains, of which 4,193 were mallards. Other carcasses collected were primarily Canada geese (Branta canadensis) and northern pintails (Anas ~). Throughout January and early February more than 400 carcasses per week were collected (Fig. 24). Although the number of carcasses collected per week declined through late February and early March, this was more a function of reduced search effort (Fig. 25) than reduced mortality rates. Even though cholera was the suspected cause of death for most birds, some waterfowl apparently died from lead poisoning, starvation, predation, and collisions with wires. The distribution of mallard wings by age-sex class was evaluated using 1,641 wings. More males were collected in unit 7, immature males in units 14 and 20, adult males in unit 18, and adult females in unit 21 than expected �104 700 .00 800 (I) C a:: iii c oct 400 300 W C 200 100 0 WEEK Figure 24. Number of dead waterfowl collected weekly between January and March 1988 on the MVNWR. 18 .•.... a:: 18 •.... • 14 0 12 J: <, ~ -a:: u.. u, W ~ Z :l 10 I <, (I) C a:: iii c I 4 oct W C 2 WEEK Figure 25. Number of dead waterfowl collected per hour of search time, January through March 1988, on the MVNWR. �105 (P-O.OOOl, Chi-square-75.37l3, 36 df). Fewer immature males were collected in units 16 and 23, immature females in unit 1, and adult females in unit 18. More banded birds were collected in units 16 and 23 than were expected (P-0.0397, Chi-square-21.8l23, 12 df). Condition indices of wing banded birds were normally distributed for all age-sex classes except adult females (P>0.5 for all age-sex classes except adult females, W-0.964, P-O.OOl, for adult females, Figs. 22-25). Adult females were negatively skewed and kurtotic. Condition indices of recovered immature females and adult males were normally distributed (W-0.968, P-0.299 and W-O.965, P-O.119, for immature females and adult males, respectively, Figs. 26 and 27). Condition indices of recovered adult females and immature males were negatively skewed and kurtotic (W-0.894, P-0.008 and W-O.936, P-0.028, for adult females and immature males, respectively, Figs. 28 and 29). Wilcoxon sign rank test detected differences in the distribution of condition indices of recovered birds compared with the distribution of banded birds for adult males (z-2.25, P-O.0244). Other age-sex classes did not show any differences (z--0.37, P-0.7086; z--O.16, P-0.87l5; z-0.99, P-0.3242 for adult females, immature females and immature males, respectively). Of the 100 wings from which lipids were extracted from the ulna, 8 were from birds that appeared to have starved. No classification errors occurred when fat blots were compared with extraction results. Because no classification errors were apparent, the remaining wings were only sampled with the blot technique. Of the 900 wings sampled, 60 (6.7%) appeared to have been from birds that had starved. Bone marrow smears from the radius bones have been made, but not yet examined. Twenty-nine of 201 banded birds (14.4%) found dead and analyzed for relative ulna lipid levels appeared to have starved. This is a significantly greater percentage than for unmarked birds analyzed (31 of 699, 4.4%; P-O.OOOO, Chi-square-20.87l9, 1 df). Ulna lipids were extracted from the wings of 52 dead, instrumented females. Twelve (23%) appeared to have come from birds that had starved. Of the 25 planted carcasses, 17 (68%) were eventually recovered. Carcasses were recovered an average of 6 days (SE-4) after placement. Only 1 carcass was recovered in a marsh unit other than where it was placed; that carcass was planted in unit 15 and recovered in unit 22. Five carcasses were taken by scavengers from the area they were placed in within 24 hours, and only 1 was later recovered. Of 24 carcasses necropsied, 17 had positive liver cultures for cholera and 7 were negative. Of the 17 that tested positive, 16 (94%) showed gross lesions characteristic of cholera, 13 (76%) had positive bone marrow cultures, and 12 (71%) showed positive marrow smears. None of the carcasses with negative liver cultures were positive for any of the other tests. Determination of detectability of cholera in wings which had been exposed to the elements for varying periods of time have not been completed yet. Standard Metabolic Rate Only 11 usable SMR measurements were obtained out of the expected range (Fig. 30). Lower critical determined from these measurements. and several temperature of these are can not be �106 eo o &IJ a: CJ ALL BIRDS ~ RECOVERED BIRDS 35 &IJ ~ 30 o &IJ U &IJ m 20 ~ 15 a: a: Z 10 5 o o 5 10 15 20 25 CONDITION CLASS Figure 26. Condition Indices of Immature females banded 11-18 January 1988 on MVNWR and condition Indices of banded Immature females found dead January-March 1988 on MVNWR. eo 75 ~ 70 ~ ~ ~ ~ ~ g- 85 0 &IJ 10 a: 55 > 0 0 41 &IJ a: a: 40 m 30 :l 215 &IJ &IJ ::; Z 10 315 20 15 10 5 0 [ CJ ~ ALL BIRDS RECOVERED BIRDS ~ ~ ~ g!;g~ ~ ~ !; g- nn.. o n n r1n 5 a ~ ~ ~ [51 ~ ~ ~ ~ ~ 10 CONDITION CLASS Figure 27. Condition Indices of adult males banded 11-18 January on MVNWR and condition Indices of banded adult males found dead January-March 1988 on MVNWR. bt A 20 ~ ~n n _n 215 �107 40 35 0 W 30 CJ ~ ALLIIRDS "ECOVERED IIRDS a:: W > 0 25 0 W a:: a:: 20 w m 15 Z 10 :E ::l 5 0 CONDITION CLASS Figure 28. Condition Indices of adult females banded 11-18 January 1988 on MVNWR and condition females found dead January-March Indices of banded adult 1988 on MVNWR. 10 55 0 50 a:: 45 > 0 40 0 IU 35 W c=J ~ ALL BIRDS RECOVERED BIRDS IU a:: a:: 30 w u :E ::l 20 m Z 15 10 5 0 25 CONDITION CLASS Figure 29. Condition Indices of Immature males banded 11-18 January 1988 on MVNWR and condition Indices of banded Immature males found dead January-March 1988 on MVNWR. , �108 225 I- 200 I- 175 I- 150 I- 125 I- 100 I- >- • • ••• • ~ 0 <, .J ~ 0 ~ 711 I- 110 I- 25 I- 0 -30 • • • •• I I I I I I I -20 -10 o 10 20 30 "'0 AMBIENT TEMPERATURE Figure 30. Standard metabolic rate measurements mallards made In the San Luis Valley, Colorado. of captive, adult �109 Time Budget From 18 February through 1 April 1988, 4.27 hours of focal animal time budget data were obtained. All observations were of birds around open water and environmental conditions were very similar. Swimming, feeding, and resting accounted for 73.4% of the observation time (Fig. 31). Metabolizable Energy Two grains grown in the SLV were measured for metabolizable energy, steptoe barley and field peas. Gross energy of steptoe barley was 18.85 (±0.24) kj/g and field peas was 18.37 (±O.05). Overall, birds maintained on field peas lost an average of 19.3 (± 19.8) g while those maintained on steptoe barley gained 13.2 (± 21.6) g. Apparent metabolizable energy was similar for the 2 foods (Table 19, ~-O.26, 18 df, P-0.7977). Intake energy did not differ between the 2 foods (Table 19, ~--1.86, 18 df, P-O.0791), but fecal energy of barley was greater than for peas (Table 10, ~--2.29, 18 df, P-0.0345). Daily protein intake was 7.31 (±1.58)g and 2.14 (±1.50)g was retained for birds fed steptoe barley. Analysis of the field pea and fecal samples from birds fed field peas have not been completed yet. Table 19. Daily intake, fecal, and apparent metabolizable energy (in Kj/day) of field peas and steptoe barley grown in the San Luis Valley, Colorado and fed to adult mallard ducks. Food Intake energy Fecal Energy Apparent Metabolizable Energy Peas 787.9(232.8) 193.6(67.0) 13.91(0.42) Barley 984.0(213.6) 258.2(51.9) 13.85(0.56) Raptor Counts Raptor populations peaked on MVNWR during the week of 31 January-6 February when 98 raptors were observed (Fig. 32). Total raptor populations were primarily influenced by eagle numbers (Fig. 32 and 33). Adult bald eagles outnumbered immatures on all counts, although some of the unidentified eagles were probably immature bald eagles which could not be definitively separated from golden eagles (Aguila chrysaetos). Reproduction Two instrumented females were found associated with nests. One female, instrumented 11-17 January with a condition index of 14.0, nested in baltic �110 -es observations, behavior A •• t ~ Maintenance Courtahlp other of 4.7 hours of focal animal time budget made In February categories. 'eed ~ t7J lZI Figure 31. Proportion Swim and March 1988, spent In different Male and female observations are combined. 100 10 eo Q W 70 W eo ->a: en al 0 en a: 0 ~ 50 "0 Q. < a: 30 20 10 0 1/14 1120 1128 2/" 2/11 2/11- 2/25 3/3 3/10 3/18 3/24 DA TE OF COUNT Figure 32. Total raptors counted MVNWR, January-March 1988. from a roadside census route on , �111 __._ TOTAL EAGLES --b - ADULT aALD •••• .•• IMMATURE 3/10 3/18 70 c w > a: w en CO 0 .0 eo CO en w ..J 30 CJ < w aALD &._--- 20 .& , ,, ,,' • ' ..... A •........•.. ......•.. 10 ,, ,, , .........•....... \,- ~-- \ , \ . .' .... /k-- •........••• ' .....<:'~ .... .. ..... 0 1I1~ 1/20 1128 2/~ 2/11 2/1. 2/25 3/3 3/24 DATE OF COUNT Figure 33. Number of adult, Immature, and total eagles counted from a roadside census route on MVNWR, January-March 1988. 220 LONE MALE 200 180 c w 180 Z 1~0 o o 120 w 100 ~ ;:) en ...J < ~ 80 80 ~o 20 o ~/1' C/28 1/17 • 1/2~ ./3 . DATE OF COUNT Figure 34. Number of male mallards counted from a census route on MVNWR and classified as lone or paired from mid-April through mid-June 1988. ./8 ./15 �112 rush south of the Avocet Trail in unit 20. She had 4 eggs and very little down when the nest was located. This nest was destroyed by avian predators. The second female was instrumented in late January ana had condition index of 8.6. Her nest was located in early June in cattail in unit 18. It had 7 eggs surrounded by abundant. The fate of this nest is unknown because it could not be located for a follow-up check. The ratio of lone:paired males increased until mid-May, and remained fairly constant thereafter (Fig. 34). Unlike most mallard populations, a substantial percentage of the males appeared to be paired and the females not nesting throughout the nest initiation period. Fifty-four mallard broods were encountered and aged. Unlike 1987, mallard hatch dates showed a peak during the week of 29 May-4 June (Fig. 35). Weather Average January temperatures were 6.3 C below normal, February temperatures were 3.1 C below normal and March temperatures normal. Precipitation was normal for all 3 months. Daily temperatures at Alamosa airport display wide differences in maxima and minima and large day-day variations (Fig. 36). Snow cover persisted until late February (Fig. 37). DISCUSSION - 1986-87 Environmental conditions in the SLV during the 1986-87 winter probably were more severe than normal in the SLV. Mean daily temperatures were below normal and precipitation above normal in January and February (Table 8). Snow depth influenced foraging patterns of mallards wintering on the MVNWR. Snow covered all grain fields within about 15 km radius of the refuge. Consequently, the only grain available was that unharvested grain grown on the refuge. In these fields, the grain heads protruded above the snow, but were quickly removed by the estimated 17,000 ducks and 2,000 Canada geese (Branta canadensis) wintering on the refuge. When the 35 cm of snow fell during the 15-17 January, very little grain was available to foraging birds. Refuge personnel bladed snow off portions of the fields several days after the snow fell, but I suspect that grain made available was rapidly depleted. As a result, I believe that mallards wintering on the MVNWR were both thermally and nutritionally stressed during January and February. Although mortality rates for instrumented birds are high and represent minimal estimates, they may not be excessively higher than population mortality rates in the SLV. Approximately 250 wings were collected during radio-tracking and carcass searches. A crude Lincoln-Petersen index based upon the number of mortalities of instrumented birds found during these collections without using the transmitter to locate the carcass suggests that only about 10% of the carcasses were found. Therefore, some 2,000 to 2,500 mallards are estimated to have died on MVNWR from December through April. Refuge counts (Ed Merrit, personal comm.) do not show noticeable declines during this period, but because immigration and emigration of birds occurs regularly, weekly counts may be unreliable for quantifying winter mortality. Fifteen mortalities occurred during the week 8-14 February (Fig. 2). Mild weather during this period made it unlikely that environmental stress was proximate cause. Since the number of raptors in the SLV increased during this period, predation rates may have increased thereby reducing survival of instrumented birds. Mortality which occurred the week after all the birds were , �113 20 fn 0 0 0 ex: CD 16 II. 0 ex: &LI CD 10 :;! :;) Z 6 o MAY 16 JUNE 1 JUNE 16 JULY 1 WEEK Figure 35. Estimated hatching dates of mallard broods encountered In the San Luis Valley, Colorado, In 1988. Deily Mexlmum 20 Deily Minimum .•... o .•... &LI a: :;) 10 o ••• <t a: &LI e, # ,.,1 " -10 &LI It ,. -20 •~ •• • ,,\ , I , I, -30 I., " , " • .' ' , 1", • ., : I ,: '"'. ,,' , , \,' , ,.' • ", , • , , ,I \' 'I' ' ' I .' " " • _, I' - ',' '..! ",, .' .-' , v , " .len 16 , Feb 1 F.b 16 Mar 1 DATE Figure 36. Dally temperature '"' ' , " ,.! " .,' I I !" ...,' ,, I, " ,,, '. , ", . .,.,, , " " II L I. " •• , • :;! ••• ",",_. " • \ I , ,~ I, maxmlma and minima recorded Alamosa airport, Alamosa, Colorad, 1988. at the Mer 16 Mar 30 �114 ~ ~ ~ ~ ~ ~ ~ ~ 0 ~ % --, L-, ~ .... ~ 10 •••• 0.. rI L...., ~ W 0 0 == Z ~ , ...___, (/) 15 U - ~ , o I I I I Jan 1 Jan 115 Feb 1 Feb 115 , I Mar 1 I I Mar 115 I Mar 30 DATE Figure 37. Depth of snow on the ground at the Alamosa airport during winter, 1988. , �115 instrumented probably resulted from 18 inches of snow which covered foraging areas, low ambient temperatures, and remnant effects from handling stress. Of the approximately 70 birds alive in May at the start of the nesting season, only 31 could be located; 29 in the SLV, and 2 in the Gunnison Drainage. Most of the remaining birds probably migrated from the SLV, although the possibility of these birds not being located due to radio failure or unlocated mortality can not be discounted. Searches of North and South parks, intermountain wetlands north of the SLV, and a portion of the Front Range around Fort Collins failed to located any instrumented birds. The reasons for low winter survival rates and poor reproductive effort among the marked population are not apparent. The physical effect of the transmitter package is one potential reason, although the transmitter package was identical to those successfully used in other studies of breeding ducks (eg. Ringelman 1980, Cowardin et al 1985). Nevertheless, the transmitter package may have influenced both the mortality estimates and the reproductive effort in subtle ways. For example, the harness which hold the transmitter in place may make the bird reluctant to fly for a few days after attachment. Consequently, instrumented birds may be more likely to forego field-feeding flights until they have adjusted to the harness. During this time the bird is utilizing endogenous reserves. Thus, birds in "poor" condition which do not have reserves to survive long without foraging become weak, and are more likely to die than birds with more reserves or those making regular field-feeding flights. "Good" condition birds may become "poor" as their lipid reserves are depleted. It is unlikely that a bird could maintain a positive energy balance without making field-feeding flights. Counts of pairs and lone drakes indicated a prolonged nesting period for mallards in the SLV. This is reflected in prolonged hatching dates, lasting from mid-May through mid-July. Such a hatching curve may result from nonValley residents arriving on the nesting grounds in good condition and initiating nests earlier than mallards resident during winter. DISCUSSION - 1987-88 Condition indices and weights were similar to those recorded in 1987. Whether mean weights and condition indices are representative of the population or biased low is unclear. Although sample sizes are relatively large, most birds were caught in bait traps, which appear to capture more "poor" condition birds. Although weights of birds recaptured less than 7 days after initial capture were not significantly different between the 2 captures, the possibility of a handling effect can not be discounted. Weights of immatures recaptured during different trapping periods tended to be lower than unmarked birds, yet recaptured adult males had similar weights to unmarked birds. This may result from lower condition immatures being more likely to be retrapped or handling stress may cause a weight decline. A similar relationship was not observed among adult males. Too few adult females were recaptured to identify any relationship. The tendency for a male in a relative condition class 1 year to be in the same relative class the next year suggests that some males are be more efficient than others in balancing their energy budget and that these differences continue through time. The lack of any relations for females is probably a result of small sample sizes. As in 1987, winter mortality of mallards in the SLV was extremely high. Although the outbreak of fowl cholera and an increased field effort caused us �116 to find more dead mallards this year than last, survival rates of instrumented females were similar between the 2 years (65 of 105 (62%) instrumented in January died in 1987 and 69 of 108 (64%) instrumented in January and February 1988). Assuming that 30% of the dead birds were found, we project that 14,000 mallards died from January-March 1988. Given that 24,000 mallards were censused on 28 January,' the mortality rate exceeds 50% assuming no birds migrated into the SLV. However, an unknown number of mallards migrate through the SLV during spring. Pintails begin migrating into the SLV in early February, and migrant mallards probably begin arriving at the same time. If an appreciable number of migrants succumbed to avian cholera in the SLV, that might account for the significant difference in ulna lipids of marked and unmarked wings. The banded birds were known to winter in the SLV, and 14.6% of them had starvation levels of lipids in the ulna. Only 4.6% of the unmarked wings had starvation levels. Perhaps the difference results from migrants in better condition than winter residents forming the bulk of the February and March cholera fatalities. Without more information on the arrival dates and numbers of mallards migrating into the SLV it is difficult to evaluate any effect the mortality may have on the number of mallards wintering in the SLV. As in 1987, only 2 instrumented birds nested, both in "poor" condition when captured. Whether this low number of winter residents nesting is representative of the population or an artifact of instrumentation remains to be answered. The large number of paired males throughout the nesting season (Fig. 28) suggests there may be a significant non-nesting component in the population. If so, then the instrumented females may be representative of the population, and most of the production occurring on MVNWR results from migrants. Alternatively, data from the instrumented females which were collected suggest the radio package may influence the weight dynamics of the instrumented birds (Fig. 16). An instrumented female may drop in weight below some threshold needed for breeding. These two hypotheses are not mutually exclusive, and the true "answer" is probably a combination of poor reproductive success of winter residents and the radio package influencing weight dynamics. Unlike the 1987 results, no apparent relationship between survival and con9ition at the time of instrumentation was found this year, although power of the statistical tests used were generally low (Tables 8 and 9). Cholera may have masked any relationship since body condition is known not to alter susceptibility to cholera (Wobeser 1981). With the large number of carcasses available, predators such as bald eagles probably scavenged more than they hunted. Consequently, predation, which may have been related to the survival process in 1987, was probably minimal. The standard metabolic rate measurements suggest that acclimation of metabolic rate to low environmental temperatures occurs in mallards. The mean standard metabolic rate at -9 to -10 C of 673 Kj/day is 49% of the rate estimated from Smith and Prince (1973) at -10.9 C. This provides a substantial energetic savings to mallards wintering in the SLV. To meet this minimum daily requirement, a mallard would have to consume 57 g of steptoe barley or field peas for maintenance. Unfortunately, without information on time in field and flight for individuals, no reliable estimate of energy expenditure can be made. Estimates of these two variables, as well as additional time budget data in fields and at roosts, should be top priorities in developing an energy budget for mallards wintering in the SLV. �117 LITERATURE CITED Altmann, J. 1974. Observational 49:227-267. study of behavior: sampling methods. Behaviour Burnham, K. P., D. R. Anderson, G. C. White, C. Brownie, and K. H. Pollock. 1987. Design and analysis methods for fish survival experiments on release-recapture. Am. Fisheries Soc. Monogr. 5:1-437. Dill, H. H., and W. H. Thornberry. 1950. A cannon projected capturing waterfowl. J. Wildl. Manage. 14:132-137. based net trap for Dodge, W. E. 1985. A telemetry antenna mount for Cessna-type aircraft -construction details. U. S. Fish Wildl. Serv., Res. Inform. Bull. 85-126. 4pp. Dwyer, T. J. 1972. An adjustable 43:282-284. radio package for ducks. Bird-Banding Gollop, J. B., and W. H. Marshall. 1954. A guide for aging duck broods field. Mississippi Flyway Tech. Sect. 14 pp. Greenwood, R. J. 1977. Evaluation Manage. 41:582-585. in the of a nasal marker for ducks. J. Wildl. Hensler, G. L., S. S. Klugman, and M. R. Fuller. 1986. Portable microcomputers for field collection of animal behavior data. Wi1dl. Bull. 14:189-192. Soc. Hopper, R. M., A. D. Geis, J. R. Grieb, and L. Nelson, Jr. 1975. Experimental duck hunting seasons, San Luis Valley, Colorado, 1963-70. Wildl. Monogr. 46. 68pp. Hutchinson, A. E., and R. B. Owen. 1984. Bone marrow fat in waterfowl. Wil~l. Manage. 48:585-591. Lee, E. T. 1980. Statistical methods for survival Learning Publ., Belmont, CA. 557 pp. data analysis. Miller, M. R., and K. J. Reinecke. 1984. Proper expression energy in avian energetics. Condor 86:396-400. National Colorado Oceanic and Atmospheric section. 92(1):1-14. Administration. J. Lifetime of metabolizable 1988a. Climatological 1988b. Climatological data. Colorado section. 92(2):1-14. 1988c. Climatological data. Colorado s~ction. 92(3):1-14. data. Ringelman, J. K. 1980. The breeding ecology of the black duck in south-central Maine. Ph. D. diss., Univ. Maine, Orono. 68pp. Ringelman, J. K., and M. R. Szymczak. 1985. A physiological for wintering mallards. J. Wi1d1. Manage. 49:564-568. condition index �118 LITERATURE CITED (CONT.) Rutherford, W. H., and C. R. Hayes. 1976. Stratification improving waterfowl surveys. Wi1d1. Soc. Bull. 4:74-78. SAS Institute, Inc. 1982. SAS users guide: statistics. Cary, NC. 584 pp. -----. 1985. SAS procedures Cary, NC. 373 pp. guide for personal as a means for SAS computers, Institute, version Inc., 6 edition. Smith, K. G., and H. H. Prince. 1973. The fasting metabolism of subadu1t mallards acclimatized to low ambient temperatures. Condor 75: 330-335. .. Szymczak, M. R. 1986. Characteristics of the duck populations in the intermountain parks of Colorado. Colorado Div. Wi1d1., Tech. Pub1. 35. 88 pp. -----and J. F. Corey. 1976. Construction and use of the Salt Plains duck trap in Colorado. Colorado Div. Wi1dl., Div. Rep. 6. 13pp. Wobeser, 300pp. Prepared G. A. 1981. Diseases of wild waterfowl. b~uZ~L~K. Wildlife Ringe~ Researcher C Plenum Press, New York, NY. �119 Appendix A. Condition index means for 9186-87 by date and age-sex class. The October and November samples are from hunter-killed birds. all other birds were trapped in Salt Plains Bait traps. Date Adult female Male n Mean(SD) n Mean(SD) n 10/4/86 7 21.2(3.8) 19 21.2(3.6) 3 10.0(12.2) 8 14.6(3.1) 11/1/86 1 26.2 13 22.2(2.2) 5 24.7(4.8) 7 23.0(2.1) 12/17/86 2 14.7(1.7) 9 16.4(1.9) 12/18/86 5 18.0(2.9) 25 19.3(2.5) 10 17.3(3.9) 10 15.8(4.7) 12/19/86 8 23.7(2.4) 16 18.7(5.2) 3 22.4(2.7) 14 18.1(3.3) 12/20/86 3 16.2(3.2) 28 17.0(3.4) 17 16.1(4.8) 39 16.1(3.7) 1/5/87 22 13.0(6.6) 134 12.6(4.5) 28 13.7(4.1) 82 12.3(4.9) 1/6/87 19 13.0(4.4) 41 13.4(3.0) 44 12.6(3.3) 46 12.4(3.7) 1/7/87 13 14.2(4.3) 66 11.8(3.6) 34 13.4(4.0) 75 10.0(3.8) 1/8/87 13 15.9(3.5) 102 12.1(3.5) 52 15.3(3.8) 122 11.5(3.8) 13.8(3.6) 11.7(6.2) Immature Male female n Mean(SD) Mean(SD) 1/9/87 3 13.6(7.3) 9 1/10/87 13 15.7(3.5) 1 1/11/87 10 14.9(5.4) 1/13/87 9 11.5(3.7) 1/14/87 6 12.4(4.2) 1/15/87 13 11.9(2.9) 2/5/87 13 12.5(3.8) 49 11.4(3.6) 28 13.7(3.7) 2/20/87 23 14.9(3.4) 52 11.8(3.0) 67 13.3(3.8) 40 11.3(3.4) 2/21/87 33 14.8(4.0) 112 9.7(3.5) 2/22/87 13 13.4(3.2) 3 6.7(1.4) 3 16.0(2.7) 3/18/87 1 19.4 9 8 14.8(2.0) 6.9 1 1 2.4 15.8 2 13.8(3.5) 36 12.0(3.0) 2 14.3(2.3) 176 9.3(3.6) 3 12.2(2.5) 7 12.7(4.5) �120 Appendix B. Capture histories of instrumented birds from week 1 (Jan. 11-18) through week 11 (March 20-26), 1988. One indicates that the bird was contacted and alive during the week, 0 that the bird was not contacted, and -1 that the bird was found dead during the week. Band 145795333 139719963 139719970 145796710 145795521 145796805 139719567 145796729 145795202 145796788 139719863 145796305 145795731 145796736 139719562 145796791 139719851 145795408 145796295 145796730 139719534 145796711 145795597 139719596 145796251 145795157 145796745 145795198 145796055 145795620 145796287 145796636 145796707 145795058 139719969 145796802 145795548 145796874 145795093 145796307 139719902 145795605 freq 150.010 150.021 150.050 150.050 150.072 150.072 150.100 150.100 150.140 150.140 150.180 150.180 150.187 150.187 150.200 150.200 150.210 150.219 150.219 150.219 150.259 150.259 150.281 150.290 150.290 150.300 150.300 150.311 150.320 150.341 150.341 150.341 150.353 150.362 150.370 150.370 150.402 150.402 150.420 150.420 150.435 150.452 CI 14.3 8.3 11.2 19.3 19.6 10.0 19.2 16.5 7.2 14.0 18.8 12.6 19.3 15.0 22.1 9.8 6.9 13.2 12.2 18.1 11.1 16.0 14.6 20.0 20.5 14.3 18.3 13.7 22.7 23.9 17.6 18.0 17.6 21.0 11.8 14.9 18.7 10.5 21.4 13.0 7.8 19.1 week: 1 2 3 4 5 6 7 8 9 1 1 1 -1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 -1 1 1 -1 1 1 -1 1 1 1 -1 1 1 -1 1 1 10 11 1 -1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 0 1 -1 1 1 1 -1 1 -1 1 1 1 1 -1 1 1 1 -1 1 1 1 -1 1 1 1 1 1 0 -1 1 1 1 0 1 0 -1 1 -1 1 1 1 0 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 -1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 -1 1 1 -1 1 1 1 1 -1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 �121 Appendix Band 145796744 145795405 145795595 145795604 145795403 139719810 145796797 145795277 145795079 145795203 145796706 139719846 145796732 145795488 145796615 139719589 145795110 145796302 145796795 145795409 145795274 139719932 145796719 145795275 145796306 145796689 145795300 145796705 145795220 145796724 145795291 145796309 145796865 139719668 145795476 145796833 145795067 145796470 145795059 145795603 145796631 139719685 145795592 145796622 145795104 145796275 139719590 B (cont.). freq 150.452 150.483 150.491 150.534 150.541 150.560 150.560 150.570 150.595 150.601 150.601 150.617 150.617 150.620 150.620 150.630 150.652 150.652 150.652 150.662 150.670 150.680 150.680 150.700 150.700 150.700 150.710 150.710 150.719 150.719 150.730 150.730 150.730 150.740 150.753 150.753 150.761 150.761 150.770 150.782 150.782 150.802 150.815 150.815 150.838 150.838 150.849 CI 19.3 14.4 14.9 19.7 13.8 14.5 9.6 10.0 21.5 14.1 17.1 18.8 11.7 21.0 18.3 11.6 19.4 10.6 10.0 12.7 20.9 10.9 16.9 8.0 10.4 19.2 9.9 15.7 2.3 15.8 19.9 19.4 8.1 13.8 15.0 14.4 21.0 16.7 18.9 11.9 16.0 18.9 25.3 19.5 9.7 8.6 7.3 week: 1 2 3 4 5 6 7 8 9 10 11 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 -1 1 1 1 1 1 1 1 -1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 0 1 0 1 0 1 1 0 1 1 0 1 1 0 1 1 0 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 -1 1 1 -1 1 1 -1 1 1 -1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 -1 1 -1 1 1 1 -1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 i 1 1 1 1 -1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 -1 1 1 1 1 �122 Appendix B (cont.). Band 145796301 145796735 145795490 145796283 145795622 145795589 145796876 139719820 145796688 139719514 145796276 145795332 145796836 145795065 145795404 145796274 145796769 139719958 145796310 145796692 145795047 145796737 145795480 145795371 145796625 145795276 145795106 145796734 139719832 139719848 139719951 145796838 139719859 139719831 139719699 145796786 139719955 145796293 145795086 145796279 145796294 145796794 139719906 139719968 139719533 145796501 145795064 freq 150.849 150.849 150.864 150.864 150.874 150.884 150.884 150.900 150.900 150.928 150.928 150.945 150.945 150.952 150.964 150.964 150.964 150.977 150.977 150.977 150.987 150.987 150.999 151.006 151.006 151.020 151.029 151.029 151.041 151.050 151.074 151.074 151.080 151.093 151.110 151.110 151.122 151.122 151.141 151.141 151.141 151.141 151.154 151.161 151.170 151.170 151.184 CI 12.7 15.3 18.6 9.7 21.6 22.3 7.3 19.7 19.5 10.3 ·1.0 9.5 14.4 20.0 11.7 18.7 8.0 14.9 6.9 21.8 14.8 16.8 12.4 13.0 22.8 22.1 19.0 19.7 20.2 19.0 19.3 23.2 13.5 20.1 13.6 8.7 8.9 3.0 9.1 21.1 13.3 13.5 9.2 14.0 21.8 9.6 18.4 week: 1 2 3 4 5 6 8 9 10 11 1 1 1 1 1 0 1 0 1 0 1 0 1 1 1 1 1 1 1 1 1 1 1 1 ·1 1 1 1 1 1 ·1 0 1 1 -1 1 1 ·1 7 1 ·1 1 ·1 1 1 1 1 1 1 1 1 1 0 1 ·1 1 1 1 ·1 1 0 1 0 1 ·1 1 1 0 1 1 1 1 1 1 1 ·1 1 1 1 1 1 1 1 ·1 1 ·1 1 1 ·1 1 1 -1 1 ·1 1 1 1 1 1 1 -1 0 -1 1 1 1 1 1 1 1 1 1 1 1 -1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 0 1 0 1 0 1 1 1 1 1 1 1 0 1 -1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 0 1 0 1 1 -1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 ·1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 ·1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 ·1 1 -1 1 ·1 , �123 Appendix B (cont.). Band freq 145796304 151.184 145796453 151.184 139719876 151.196 145796503 151.200 145796474 151.211 139719621 151.212 145796498 151.212 139719971 151.222 145796796 151.222 139719674 151.230 145795323 151.240 145796443 151.250 145795062 151.259 145796478 151.261 145795413 151.270 145796852 151.270 139719627 151.280 145796297 151.280 145796725 151.280 145795281 151.289 145796458 151.289 145796451 151.290 145795066 151.310 145795458 151.328 145796811 151.328 145795288 151.356 145796857 151.356 139719908 151.377 145795546 151.386 1457..96303 151.386 139719845 151.400 145796626 151.400 139719967 151.441 145796292 151.441 145795103 151.472 145796747 151.472 145796500 151.519 145796465 151.529 145795337 151.540 145796863 151.540 145795558 151.570 145796476 151.575 145795701 151.607 145795594 151.623 145796806 151.623 145795339 151.635 139719673 151.730 CI 7.3 8.4 14.3 16.2 8.2 12.3 18.3 23.5 12.6 20.3 19.0 16.8 19.1 7.4 19.6 12.8 11.3 8.4 23.0 12.6 18.1 8.9 21.0 14.2 9.4 20.0 22.9 13.2 19.4 11.0 13.9 24.5 11.5 20.5 19.3 11.3 15.8 10.2 10.3 15.0 23.6 21.2 21.5 21.9 12.8 17.9 13.0 week: 1 2 3 4 1 1 1 5 6 1 1 -1 7 8 9 10 11 1 1 1 1 1 1 1 0 1 1 1 1 1 1 -1 1 1 0 1 1 1 1 1 1 1 0 0 1 1 0 0 1 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 -1 1 1 1 1 0 -1 1 1 1 1 1 0 1 1 0 -1 1 1 -1 1 1 0 1 1 0 1 -1 1 0 1 0 1 -1 1 1 1 1 -1 1 1 1 -1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 -1 1 1 1 1 1 -1 1 1 1 1 1 1 1 -1 1 0 -1 1 1 1 1 1 1 1 0 1 1 1 0 1 1 1 1 -1 1 0 0 0 0 0 1 0 0 0 1 1 .1 1 1 1 1 1 1 1 1 1 1 1 -1 -1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1: 1 1 1 -1 1 1 -1 1 -1 �124 Appendix B (cont.). Band 145796311 145796467 145795590 145796742 139719850 145795577 145795335 145796184 freq 151.730 151.730 151.773 151.773 151.876 151.912 151.924 151.924 CI 11.2 7.1 14.5 18.1 21.1 19.8 20.3 19.6 week: 1 2 3 4 1 5 6 7 8 9 10 11 0 -1 1 -1 1 1 0 -1 1 1 1 1 0 0 1 -1 1 1 1 1 1 1 1 1 1 1 o 1 -1 1 1 1 1 1 1 1 1 1 1 , � https://cpw.cvlcollections.org/files/original/d5c034181db0298fc427c4ccfa1e11d7.pdf ff3220abcca37ff4c81ac14b26e86227 PDF Text Text 1 Colorado Division of Wildlife Wildlife Research Report April 1989 JOB PROGRESS REPORT State of: Colorado Project: __ W~-~1~5:2_-~R Work Plan: 3__ : Job Job Title: Responses Period Covered: Author: _ Upland Bird Research 13 of Sage Grouse to Vegetation 01 January through 31 December Fertilization 1988 Orrin B. Myers Personnel: C. E. Braun, S. Porter, K. Snyder, Colorado Division o. B. Myers, G. C. White, Colorado State University; of Wildlife; C. Cesar, L. Upham, U.S. Bureau of Land Management ABSTRACT The response of sagebrush (Artemisia tridentata) and sage grouse (Centrocercus urophasianus) to nitrogen fertilization was studied in North Park, Jackson County, Colorado. Sagebrush plants responded to application of 112 kg-Nfha in Fall 1985 with increased growth, increased levels of foliar crude protein; and reduced levels of foliar coumarins. Application of fertilizer in Fall 1987 produced a less dramatic response. Sage grouse used fertilized study plots for feeding more (f < 0.05) often than adjacent unfertilized plots. When presented with fertilized and unfertilized A. ~. wyomingensis (ATW) and A. ~. vaseyana (ATV) in paired choice experiments, captive sage grouse consumed more (f < 0.05) ATW treatments than ATV treatments. Sagebrush treatments were digested at similar rates. Clutch sizes were not related to distance to nearest fertilized plot and the survival distribution for hens near fertilized plots was not different from birds using other areas of North Park. ��3 RESPONSES OF SAGE GROUSE TO VEGETATION FERTILIZATION Orrin B. Myers P. N. OBJECTIVES The project is part of a two-phased study to on sage grouse winter distribution in mining ecology, and to 2) evaluate whether nitrogen grouse winter habitat, can be used as a tool available to sage grouse. The focus of this Phase 2. 1) collect baseline information areas and on grouse feeding fertilizer, when applied to sage to mitigate reduction in habitat portion of the project is on SEGMENT OBJECTIVES 1. Document chemical fertilizer, and growth response of sagebrush 2. Evaluate feeding preferences of sage grouse for fertilized unfertilized sagebrush subspecies, 3. Estimate digestibility subspecies, and 4. Monitor reproductive parameters of radio-marked sage grouse to learn if sage grouse fitness is affected by fertilizer treatment. of fertilized to nitrogen and unfertilized and sagebrush DESCRIPTION OF STUDY AREA The study area is in North Park, Jackson County, Colorado. The Park is an intermountain basin at an elevation of about 2,500 m. It is drained to the northwest by the North Platte River, which is fed by many smaller streams. Topography is flat to rolling with numerous ridges and benches. Climate is cold and dry with an average annual frost-free period of 46 days. Sagebrushdominated grasslands cover upland sites in the Park, and grasses and sedges occur in native and irrigated meadows that border drainages. Artemisia tridentata is the dominant sagebrush species and includes 2 subspecies, ft. ~. wyomingensis (ATW) and ft. ~. vaseyana (ATV). Other species of sagebrush occurring with limited distribution in North Park are ft. longiloba, ft. cana, and ft. argillosa. METHODS In October-November 1985 ammonium nitrate fertilizer (33.5% N) was applied at a rate of 112 kg-Nfha to 330 ha of sagebrush rangeland (Fig. 1). Thirty-three 20-ha blocks were randomly-selected with 11 blocks distributed in each area. The northern and southern 10 ha of each block were randomly assigned as fertilized or control and treated accordingly. In October 1987 fertilizer was applied to an additional 6 l-ha plots (Fig. 1). The 3 main study areas contain similar densities of sagebrush plants, but the proportions of ATV and ATW differ between areas (Fig. 2). �4 A JACKSON COUNTY 1987 1985 Gill II D 0 FERTILIZED CONTROL ~ 1 km 1. Colorado. Fig. A Fertilized + study plots in North Park, Jackson County, B c D ATW ATV 3,20 3,11 3,53 TOTAL 2 PLANTS/m 2. Relative densities of sagebrush subspecies on each of 3 main fertilized study areas in North Park, Jackson County, Colorado. Fig. �5 Sagebrush foliage was collected each quarter from a random sample of the 33 experimental blocks. Four blocks from each of the main study areas were sampled and 5 randomly-located plants of the 2 big sagebrush subspecies were clipped from the control and fertilized halves of each study block. Sage grouse use of study blocks was estimated by examining individual plants for evidence of feeding activity. The characteristic appearance of leaves that have been fed upon by grouse has been described (Remington and Braun 1985). Equal numbers of each subspecies were examined along pace transects through the control and fertilized portions of randomly selected study blocks. Sampling points were at random distances along transects. Sage grouse were trapped at nocturnal roosts to attach radio transmitters and to provide birds for captive experiments. Fertilized study areas were scanned periodically for radio-marked birds prior to the breeding season and ~ 3 times weekly during the breeding season. Radio-marked birds also were located periodically to assess any use of experimental blocks and to monitor reproductive success. Survival probabilities of female sage grouse were estimated for radio-marked hens. Survival functions were estimated using the Kaplan-Meier product-limit method (SAS Inst. Inc. 1988). Survival time (days) from 15 April to date of death (failure date) were used in computations. Annual survival probability estimates were bracketed by right censoring observations on the date of last radio contact and then by setting failure date equal to date of last radio contact. The procedure assumes animals are sampled randomly, that survival times are independent for different animals, that radio packages do not influence survival times, that the censoring mechanism is random, and that emigration or radio failure is zero (Pollock et al. 1989). Male and female grouse were counted on leks between about 0500 and 0730 from 01 April through the first 2 weeks of May. Ground searches were made for new or relocated leks. Feeding trials were conducted with captive sage grouse to estimate feeding preferences for and digestibility of Wyoming big sagebrush, Artemisia tridentata wyomingensis, (ATW) and mountain big sagebrush, a. ~. vaseyana, CATV). Two levels of fertilization also were tested: no fertilization (ATW and ATV) and 112 kg-Njha (FATW and FATV). The 6 possible 2-way combinations of the 4 sagebrush treatments were presented to each bird in random order to evaluate feeding preference for the different treatments. Sagebrush foliage was removed from stems and given to birds in a 400 ml beaker. For each trial similar amounts of 2 treatments were placed on the left or right side of each cage prior to the morning feeding period. Positions of each treatment in the cages were determined randomly and switched in subsequent trials. Birds were allowed to feed for 60 minutes after which the sagebrush and spillage were removed and weighed. Digestion trials were accomplished by adding known amounts of sagebrush to cages at about sunrise. Sagebrush foliage also was placed on top of cages in the morning and weighed at night to provide an estimate of water loss during the day. After the evening feeding period sagebrush was removed from the �6 cages and weighed. Spilled sagebrush was removed from the bottom of the cage, weighed and discarded. Fecal droppings were collected and stored overnight in resealable plastic bags. Fecal and cecal droppings were collected and placed into plastic bags before additional sagebrush was added to cages the following morning and subsequently placed in frozen storage. Between trials, birds were maintained on sagebrush and a commercial ration (12.5% crude protein and 10.0% crude fiber). Sagebrush foliage samples were prepared for analyses by separating leaves from sterns and pooling equal amounts of leaves from each plant into composite samples. Composite samples for each subspecies collected on the control and treatment halves of each study block were ground using a mortar and pestle after freezing with liquid nitrogen. Composite foliage samples collected during feeding trials were treated in the same manner. Grouse fecal and cecal droppings collected during feeding trials were freeze-dried and ground using a mortar and pestle. Coumarin content of winter-collected sagebrush leaves was indexed following Welch and McArthur (1986). In this method, percent transmittance values were produced and were assumed to be inversely related to the coumarin content of sagebrush leaves. Dry matter content of foliage and grouse droppings was determined after drying overnight at 100 C. Samples were ashed overnight at 500 C to determine organic matter and ash content. Kjeldahl nitrogen (Horwitz 1980) was measured in addition to neutral detergent fiber, acid detergent fiber, and lignin (Van Soest 1963~, Q; Mould and Robbins 1981, Van Soest 1982). Soil samples were collected from each of the 6 study plots prior to receiving nitrogen fertilizer in Fall 1987. Adjacent control plots also were sampled. Each plot was subsampled at 6 random locations. Subsamples were placed into a plastic bucket and thoroughly mixed before removing about 750 cc for analyses. Analyses of soil samples were performed by the Soil Testing Laboratory at Colorado State University. RESULTS Feeding Trials Two birds each from Colorado and from Wyoming were used to collect data on feeding choices by grouse and on forage digestibility. The Wyoming birds were from a captive flock maintained at the University of Wyoming at Laramie and the Colorado birds were captured in North Park. The Colorado birds were 2 of 4 birds captured in North Park for use in trials and were those that adapted best to captive conditions. All birds lost weight initially and then gained weight as they acclimated to captive conditions. Colorado birds were slower to recover lost body mass than Wyoming birds (Fig. 3). For the Colorado birds an important factor in maintaining condition was their acceptance of the commercial ration. Body mass was monitored daily during the digestion trials. Birds lost more weight when fed ATV treatments than when fed ATW treatments, regardless of fertilizer status (Table 1). However, birds also consumed less sagebrush when given FATV or ATV than when fed FATW or ATW (Table 2). When body mass changes were plotted against daily intakes (Fig. 4) the slopes were similar (~ = 0.8, ANCOVA test for heterogeneity of slopes). Thus, for a given level of intake changes in body mass were the same regardless of the treatment consumed. �7 1500 WYOtfING COLORADO 1400 1300 1\ r» '-.J (f) (f) 1200 --- "', « ::2 >0 <, 1100 ~ ----, 0 rn \. \ 1\ \ / \ v 1000 \ \ \ \ ,- 900 -- <, \ J " \{ _-\ -> ,/ 800 02/07 02/27 03/ 18 04/07 DATE Fig. 3. Performance of sage grouse used in feeding trials during 1988. Birds were obtained from wild stock in North Park and from the University of Wyoming captive flock. Table 1. treatments Body mass changes (g/d) of captive sage grouse fed 4 sagebrush during 1988 digestion trials. Subspecies fed [1. SE !,. wyomingensis Control Fertilized [1. .& (n (n 4) 4) -44 -41 7 5 (n (n 4) 4) -61 -60 9 1 !,. vaseyana Control Fertilized Source Fertilizer (F) Subspecies (S) F x S interaction df 1,12 1,12 1,12 > f f value f 0.13 8.52 0.04 0.7 0.01 0.8 �8 Table 2. Consumption of sagebrush treatments sage grouse during 1988 digestion trials. Subspecies (g fresh weight) by captive fed SE ~ . .t. wyomingensis Control Fertilized (n = 4) (n 4) 55.8 58.6 6.8 6.6 (n (n 39.8 36.2 6.7 4.2 ~ . .t. vaseyana Control Fertilized 4) 4) Source df Cage Fertilizer (F) Subspecies (S) F x S interaction 3,9 1,9 1,9 1,9 E value f > E 2.67 0.01 13.83 0.38 0.11 0.94 0.01 0.55 -20 -30 r--. -'10 bl • ATV <> ." ATVI rATV A FAT¥ c X = 0.45~ - A 68.5 <:» Ii:j CJ -50 !z; ..:r: ~ -60 E-< ::r:: Cl -70 U X= H Ii:j :;;t: 0.55X - Bl.6 -80 • -90 20 30 '10 INTAKE Fig. 4. sagebrush 50 50 70 (g) Daily changes in body mass of captive sage grouse fed 4 treatments in relation to daily intakes. �9 Dry matter intake during digestion trials also was not influenced by fertilization (£ > 0.7) but birds consumed more of the ATW treatments than of the ATV treatments (£ < 0.01). When adjusted for different intakes, the amount of dry matter retained was not affected by subspecies or by fertilization (ANCOVA £ > 0.5). Dry matter digestibility coefficients were similar for each subspecies (£ > 0.1) and for each fertilizer treatment (£ > 0.5). Dry matter digestibility coefficients for ATV treatments were between a and 10%, compared to ATW treatments, which were between 25 and 32%. Nitrogen balance and amounts of retained energy also were not affected (£ > 0.05) by nitrogen fertilization or by subspecies. Sagebrush foliage used in 1988 feeding trials did not show any effects of fertilization treatment upon crude protein and gross energy (Table 3). ATW contained more crude protein (£ < 0.01) and less gross energy (£ <0.05) than ATV. Fertilized sagebrush contained similar amounts of acid detergent fiber (ADF) and natural detergent fiber (NDF) , and ATW contained more ADF than ATV (Table 4). Coumarin levels in foliage samples collected in May 1987 showed no effect due to fertilization (£ > 0.05), but subspecies were different (£ < 0.05) (Table 5). Table 3. Crude protein (% dry matter) and gross energy (cal/g dry matter) in sagebrush samples fed to captive sage grouse during digestibility and feeding preference trials. Three samples were analyzed for each treatment. Subspecies Treatment Crude 12rotein SE ~ Gross energy SE ~ a. _t. vaseyana Control Fertilized 9.6 10.1 0.6 0.2 5747 5628 61 21 a. _t. ~omingensis Control Fertilized 12.4 13.3 0.2 0.4 5520 5494 56 90 Table 4. Neutral (NDF) and acid detergent fiber (ADF) fractions in sagebrush samples fed to captive sage grouse during digestibility and feeding preference trials. Three samples were analyzed for each treatment. ADF NDF ~ SE R SE Control Fertilized 0.318 0.316 0.005 0.003 0.219 0.210 0.008 0.001 Control Fertilized 0.351 0.357 0.005 0.017 0.233 0.246 0.002 0.004 Subspecies Treatment a. _t. vaseyana a. _t. ~omingensis �10 Table 5. Coumarin levels (% transmittance) in sagebrush plants collected during May 1987 from fertilization study plots in North Park, Jackson County, Colorado. R Subspecies Treatment 6· .t. vaseyana Control Fertilized 10.9 11.0 1.9 1.6 t:, . .t. :'£lomingensis Control Fertilized 21. 9 25.3 2.4 1.9 SE Sage grouse fed upon some sagebrush treatments in greater amounts than on others (Fig. 5). Wyoming and Colorado birds consumed treatments in the same proportions. Birds consumed more (f < 0.05) FATW than ATV or FATV. Selection for ATW was weaker than for FATW when tested against ATV and FATV. Within a subspecies, birds consumed equal amounts of fertilized and control treatments. The apparent lack of selectivity for fertilized plants in contrast to 1987 trials could be due to smaller differences in plant chemistry between treatments than in 1988. Also, in the 1988 trials only leaves were presented to birds, so they were not able to use plant morphological differences to help choose plants. A D FATV mill ATW o = 0.2 P = 0.06 P c ~ FATW 1m ATW B D 1IIIIIIII ATW .~~:~w P = 0.001 P = 0.7 E8 ~ FATW FATY D FATV 0.5 P = 0.001 DATV ". D .. P = ATY F Fig. S. Relative consumption of sagebrush grouse during 1988 feeding trials. treatments by captive sage �11 Browse Transects A total of 240 sagebrush plants was examined for evidence of feeding activity by grouse (Table 6). Sage grouse fed upon ATW more often than ATV (X2 - 11.8, £ - 0.001) and fed upon fertilized ATW plants more often than on unfertilized plants (X2 - 10.4, £ = 0.001). Low numbers of ATV plants were fed upon in all study blocks sampled. Table 6. Numbers of plants examined along transects showing evidence of feeding activity on control and fertilized study plots in North Park, Jackson County, Colorado, 1988. Fed u~on b~ grouse No Yes Subspecies n a. ~. a. ~. % n % ~omingensis Control Fertilized 58 46 97 77 2 14 3 23 vase~ana Control Fertilized 60 58 100 97 0 0 2 3 Nesting Three marked hens from the North Park wintering population nested in southern Wyoming. One of the birds had been captured for use in the captive feeding trials and was released near the Walden Coal Company mine on 22 February. The bird had lost 25% of its original body mass when released. It is not known if any other birds left the study area to nest. Nesting sites were located for 21 hens. None of the nests was on fertilized plots, including 1 bird that nested in a fertilized plot in 1987. At the end of May, 4 nests were active, 11 had been destroyed, eggs in 2 nests had hatched, and eggs in 4 nests were collected for captive propagation. Four clutches were believed to have hatched. Egg predation by ground squirrels was apparent at most failed nests. The relationship between clutch size and proximity to fertilized plots was examined. A negative correlation between clutch size and distance to fertilized plots may indicate a positive treatment effect on clutch size. The slope for the regression of clutch size on distance was not different from zero (£ > 0.2) when years were considered separately or when combined (Fig. 6). �12 10 9 8 7 Ii:J N H (JJ 6 0 0 5 P:: U 8 ~ ,_:j U 4 o 1987 •••. 1988 3 2 1 0 0 DISTANCE Fig. 6. Colorado, 40 30 20 10 (kID.) Clutch size of sage grouse nests in North Park, Jackson County, in relation to distance to nearest fertilization study plot. Leks No new leks were located near the fertilized study areas. Peak male counts from leks active in 1988 were summed for all for each quadrant of North Park years since 1973 to provide an index to grouse population changes. Six leks in the NE quadrant were included and all were within about 5 km of fertilizer treatments. The number of leks summed in the NW, SE, and SW quadrants were 6, 2, and 10, respectively. If it is assumed that lek counts track population levels, leks in the NE (fertilized) fluctuated in much the same manner as leks in other areas of North Park (Table 7, Fig. 7). Table 7. Peak male counts on selected leks in quadrants of North Park, Jackson County, Colorado. Fertilized study blocks were in the NE Quadrant. Longterm R Quadrant NE NW SE SW (n (n (n (n 6) 6) 2) 10) 104 188 90 340 1985 1986 1987 1988 70 124 111 332 47 94 54 266 108 177 99 363 99 199 92 342 �13 800 NE SE NV FERTILIZER APPLIED SW 600 I \ I I I \ / \ -- I -- - \ / I \ \ 400 / /" / \ \ / / \ \ I 200 ,,/ / __ .... v \ ....... -._ ._-_._.-._.-- _ .-._ -., . / \ I \ \/:.~.~,,,/ -s;»: v''_'-'' _ -, o 72 74 76 78 80 82 84 86 88 90 YEAR Fig. 7. Total numbers of male sage grouse counted at selected leks in North Park quadrants. Fertilizer was applied in Fall 1985 in the NE quadrant. Survival Radio locations were obtained from 63 birds on 110 different dates during 1987 and 1988. Survival distributions did not differ between years (f> 0.2), and survival probabilities for birds captured in the northeast (0.612) and the northwest quadrants (0.700) were similar (f> 0.4). The annual survival probability for 15 April 1986 to 14 April 1987 was 0.67 (95% confidence interval = 0.52 - 0.0.83) compared to an estimate of 0.58 (95% confidence interval = 0.33 - 0.82) from 15 April through 31 December 1988. The combined annual survival probability was 0.300 (95% confidence interval = 0.184 0.416) when dates of last radio contact were treated as failure dates (Fig. 8), which was not different from the estimates for 1987 and 1988. Soil Chemistry The amount of nitrate-nitrogen (N03-N) in soils from plots fertilized in Fall 1987 was increased (f = 0.01) by fertilizer treatment (Tables 8, 9). Other soil characteristics were not affected by the treatment (f> 0.1). The amount of organic matter (OM) in Morset Loam was greater than in Bosler Loam (f = 0.03; Tables 10, 11), and Morset Loam contained more K than Bosler Loam (f < 0.01). Soil chemistry values also differed (f < 0.05) between Fall 1987 and June 1988 for pH, conductivity (cond.), OM, N03-N, P, Zn, Mn, and Cu. �14 1 0.9 >- f- 0.8 + + + _J m 0,7 « CO 0 + + 0.6 II 0... _J 0.5 « > > a: ::) 0.4 0.3 + (f) 0,2 0 100 200 DAYS AFTER 300 400 INSTRUMENTATION Fig. 8. Survival probabilities of sage grouse in Jackson County, Colorado and exponential fit to data that assumes last radio contact equals time of death. Table 8. Soil characteristics of experimental plots when treated with ammonium nitrate fertilizer in October 1987. n Control Treatment 6 6 pH 6.8 6.7 Condo (nmhos/cm) OM 0.18 0.15 3.1 2.9 N03~-N~__ ~P ~K~ __ ~Z~n ~F~e ~M~n~ C~u~ 3.0 2.9 2.2 2.1 (ppm) (%) 0.6 0.7 4.2 3.6 231 205 0.8 0.8 23.9 24.0 Table 9. Soil characteristics in June 1988 from experimental plots treated with ammonium nitrate fertilizer in October 1987. n Control Treatment 6 6 pH 7.5 7.2 Condo (nmhos/cm) 0.22 0.28 OM N03-N P K Zn (ppm) Fe Mn Cu 0.8 7.0 2.7 2.9 194 204 0.4 0.5 14.9 19.7 2.0 2.4 1.9 1.7 (%) 1.8 1.9 �15 Table 10. Soil characteristics in experimental plots when treated with ammonium nitrate fertilizer in October 1987. Soil series Bosler loam Morset loam n 4 8 pH 6.7 6.8 Condo (nos/em) 0.10 0.20 OM N03-N P K Zn (ppm) Fe 0.5 0.7 3.5 4.2 140 257 0.9 0.7 17.9 27.0 (%) 2.1 3.4 Mn Cu 3.1 2.9 1.8 2.3 Table 11. Soil characteristics in June 1988 from experimental plots treated with ammonium nitrate fertilizer in October 1987. Soil series n pH Bosler loam Morset loam 4 8 7.3 7.4 Vegetative Condo (nos/em) 0.25 0.25 OM (%) N03-N P K Zn (ppm) Fe Mn Cu 1.3 2.1 2.5 4.6 2.6 2.9 172 212 0.4 0.4 13.5 19.2 2.4 2.1 1.8 1.8 Response Both sagebrush subspecies responded to fertilization with increased length of stems and by producing larger leaves. Average plant height did not increase (f > 0.05) due to large variation among plants. Whole plant morphology, however, was affected by fertilization in other ways. The number of ephemeral reproductive stems was increased by fertilization, which was not included in plant height measurements. The effect of lengthened stems, more numerous and longer reproductive stems, and larger leaves caused plants on fertilized plots to appear substantially taller and more lush than plants on adjacent control plots. The impact of fertilization on sagebrush plants apparently was greater for the 1985 treatment than when fertilizer was applied in 1987. The appearance of 1987 plots seemed less lush than the appearance of plots fertilized in 1985. In each case length of new stems was increased by fertilization (Fig. 9), but response to fertilization by ATW plants was less dramatic for the 1987 treatment than for the 1985 treatment. It is possible that vegetative response was dependent on soil moisture conditions during the growing season after fertilizer application. Soil moisture levels were higher during the 1986 growing season (extensive snow fell during and after fertilization in Fall 1985) than during the 1988 season. �16 80 70 ,....., FI FI t 60 t t P=1 s- 50 c..? :z; I'Ll ....:l j 40 ~ I'Ll A ...: 30 PLl ....:l 20 t t t + 10 0 .L_L _g________r ATW ATV 1987 _g________r _g________r AT" ATV 1985 Fig. 9. Length of sagebrush new growth after 1 growing season. Following fertilization with 112 kg-N(ha. Data are means and 95% confidence intervals. DISCUSSION Captive sage grouse showed affinities for feeding on some sagebrush treatments over others. One method for estimating palatabilities of forages is to measure intakes during digestibility trials. This method demonstrated that sage grouse fed preferentially upon ATW treatments over ATV treatments and that fertilization was not a factor in feeding choices. The 1988 choice trials support the finding that within a subspecies fertilization was not important (Fig. 5C and 5D). Additional insight into sage grouse feeding choices is gained when other treatment combinations are examined (Fig. 5). When fertilizer was applied to ATW, sage grouse increased their propensity for feeding on ATW over ATV, even when ATV also was fertilized (Fig. SF). When ATV was tested against FATV, their relative consumption was not different. Thus, fertilization did not seem to increase use of ATV plants by grouse. In 1987 grouse consumed more FATW than ATW (f < 0.01). The lack of any feeding preference for FATW over ATW in 1988 could be a function of several factors. Birds could use morphological cues associated with the entire plant to select food plants within a population. Also, primary and secondary plant chemistry of FATW plants may have changed between the 1st and 2nd years of fertilizer response due to a natural diminution of any fertilizer effects or due to drought conditions during the 1987 growing season. Drought may have prevented FATW plants from expressing characters that elicit increased feeding activity by grouse. �17 Grouse feeding activity measured on fertilized plots may have increased slightly from 1987. In 1987 10% of the ATW on fertilized plots was fed on compared to 23% in 1988. The use of ATV and FATV, and ATW on control plots was about the same between 1987 and 1988. The apparent increased use in 1988 of FATW plants could be due to an overall increase in the numbers of birds using the NE quadrant of North Park in response to weather conditions. Winter 1986-87 was extremely mild with no permanent snow cover in most areas after mid-February. Permanent snow cover began accumulating in December 1987 and eventually completely covered many sagebrush stands, a situation more typical for North Park winters than winter 1986-87. These conditions generally cause birds to move to areas, like the NE quadrant, where windswept ridges and benches provide access to food. Grouse also may have learned that superior feeding sites existed in fertilized plots and increased their use of these areas. Sage grouse in the NE in winter 1987-88 made extensive use of areas north and south of the Walden Coal Company mine. The mine was closing and earth-moving activities during the winter caused dust to settle about the mine. Much dust was deposited north of the mine which, on sunny days, caused the snow to melt and thereby exposed ATW stands to grouse. These exposed areas were heavily used by sage grouse for feeding and nocturnal roost sites. It is not known if grouse used these sites traditionally or whether this use represents a learned response to habitat alteration. The observed survival distribution (Fig. 8) resembles that found by Kurzejeski et al. (1987), with more mortality occurring in spring and fall than at other times of year. Although survival of birds in the northeastern (NE) and northwestern (NW) quadrants of North Park was similar, most documented mortalities in the NW were due to hunter harvest, whereas in the NE most mortalities were caused by predators in spring. That harvest mortality is substantial in the NW is not surprising given hunting pressure for this part of North Park. High spring mortality rates for the NE is less easy to explain. The NE may contain more avian predators than the NW: there have been 2 active golden eagle nests in the NE and numerous corvids visit the county landfill in the NE portion of the Park. A total of 67% of the radios had to be censored. The assumption that the censoring mechanism was random may be violated. The center of gravity on poncho-mounted solar radio transmitters is such that if dropped, the radios fall face down (thereby covering the solar panels) more frequently than they fall face up. Thus, some mortalities may not be found because radios quit transmitting if they fall face (solar panels) down. Although radio-marked grouse did not use fertilized plots to any appreciable extent, grouse use of experimental blocks was detected from transect data. The occurrence of a drought during 1987 and heavy snow cover in winter 1987-88 made classifying grouse use and nonuse difficult at times. Drought caused many plants to abort leaves late in the 1987 growing season. Whole stems with leaves dropped sometimes looked similar to stems fed upon heavily by sage grouse, which can be fairly common when few other stems were exposed above the snow for grouse to feed upon. Such plants seemed to occur mostly on fertilized plots and were classified as not used by grouse unless 2 or more clipped petioles remained on the plant. �18 Within treatment variability during the digestion trials was greater in 1988 than in previous trials. One possible source of this variation was that 2 sources of birds were used, but this does not seem to be the cause because the data showed no consistent pattern with respect to bird source. A more probable source of variation was that insufficient time and resources were available to allow birds to fully acclimate to sagebrush treatments before measurements were collected. Between trials, birds were fed combinations of sagebrush treatments to help minimize acclimation problems. In previous trials birds were fed only ATW treatments and did not have to adapt to ATV diets during the digestion trials. All birds used in digestion trials eventually consumed the commercial ration. Use of this ration to maintain birds condition also could have caused digestibility of sagebrush treatments to be impaired. Sage grouse have proven to be difficult subjects for nutrition experiments. Both wild-caught and captive birds have lost weight rapidly when fed only sagebrush, so that prolonged diet acclimatization and sample collection periods are impractical. Field studies have suffered because only easily measured chemical parameters have been estimated and because plant chemical parameters are uncontrolled and confounded (Remington and Braun 1985, Welch et al. 1988). The problem of feeding native forage to captive grouse, while maintaining birds in good condition, has hampered others (e.g., Hill et al. 1968, Servello et al. 1987). A commonly used compromise for studying nutritional value of foods is to evaluate the relative digestibility of forage-ration mixtures. This approach sometimes produces absurd partial digestibility coefficients. for the forage portion of the diet, which limits this method's usefulness for estimating true digestibility of forages. If the research goals are to estimate the relative nutritive value of forages, this approach may be appropriate. LITERATURE CITED Hill, D. C., E. V. Evans, and H. G. Lumsden. 1968. Metabolizable energy of aspen flower buds for captive ruffed grouse. J. Wildl. Manage. 32:854858. Horwitz, W. Editor. 1980. Official of Official Analytical Chemists. C. 1018pp. Kurzejeski, E. W., L. D. Vangilder, turkey hens in North Missouri. methods of analysis of the Association Assoc. Off. Anal. Chern., Washington, D. and J. B. Lewis. 1987. Survival J. Wildl. Manage. 51:188-193. of wild Mould, E. D., and C. T. Robbins. 1981. Evaluation of detergent analysis in estimating nutritional value of browse. J. Wi1dl. Manage. 45:937-947. Pollock, K. H., S. R. Winterstein, C. M. Bunck, and P. D. Curtis. 1989. Survival analysis in telemetry studies: the staggered entry design. Wildl. Manage. 53:7-15. J. �19 Remington, T. E., and C. E. Braun. winter, North Park, Colorado. 1985. Sage grouse food selection J. Wildl. Manage. 49:1055-1061. in SAS Institute Inc. 1988. Additional SAS/STAT procedures, release 6.03. Tech. Rep. P-179. SAS Institute, Inc., Cary, N.C. 255pp. Servello, F. A., R. L. Kirkpatrick, and K. E. Webb, Jr. metabolizable energy in the diet of ruffed grouse. 51:560-567. SAS 1987. Predicting J. Wildl. Manage. Van Soest, P. J. 1963~. Use of detergents in the analysis of fibrous feeds. I. Preparation of fiber residues of low nitrogen content. J. Assoc. Off. Agric. Chem. 46:825-829. 1963Q. Use of detergents in the analysis of fibrous feeds. II. A rapid method for the determination of fiber and lignin. J. Assoc. Off. Agric. Chem. 46:829-835. 1982. Nutritional ecology of the ruminant. Corvallis, are. 374pp. a and B Books, Inc., Welch, B. L., and E. D. McArthur. 1986. Wintering mule deer preference 21 accessions of big sagebrush. Great Basin Nat. 46:281-286. _____ , J. C. Pederson, and R. L. Rodriguez. by sage grouse. Prepared by, 1988. Great Basin Nat. 48:274-279. ~orrinB.~ Graduate Research Assistant Approved by: Clait E. Braun Selection for of big sagebrush ��Colorado Division Wildlife Research April 1989 of Wildlife Report 21 JOB PROGRESS REPORT State of: Colorado Project: W-152-R Upland Work Plan: 3__ Job Title: Response Period Covered: Author: Personnel: Job Bird Research 17 of Selected 01 January through Avifauna to Fire in the Big Sagebrush 31 December Type 1988 Lee A. Benson C. E. Braun, C. Gefroh, C. Poley, Colorado L. A. Benson, Colorado State University Division of Wildlife; ABSTRACT The response of sage grouse (Centrocercus urophasianus) and passerine birds to fire was studied in Jackson and Moffat counties, Colorado. Lek counts were conducted during the breeding season. Radio-marked sage grouse were located during spring and summer months. Average daily spring movements were 0.8-0.9 km at all leks. Few locations of radio-marked sage grouse were obtained within burned areas during the breeding season. Movements to summering areas ranged from 1.1 to 28.5 km in North Park and averaged 1.6 km in Moffat County. Breeding bird surveys showed a large decrease in density of passerines at all burn sites. ��23 RESPONSE OF SELECTED AVIFAUNA TO FIRE IN THE BIG SAGEBRUSH TYPE Lee A. Benson INTRODUCTION Sagebrush (Artemisia ~.) dominates major portions of the western United States. Reports of total area covered by sagebrush vary from 35 million ha (Schroeder and Sturges 1975) to 60 million ha (Beetle 1960). More than 10% of all sagebrush lands have been altered by man (Braun et al. 1976). Among the methods used, fire is becoming increasingly more attractive as a management tool for reducing sagebrush (Frandsen 1985). This study was initiated as additional information is needed on the impacts of fire on wildlife so that any impacts can be more fully considered in management of the sagebrush type. P. N. OBJECTIVES l. Review 2. Capture 6 to 10 male sage grouse per lek and equip with poncho-mounted tail-clip transmitters. 3. Locate and flush birds at spring feeding/loafing 4. Obtain 5. Monitor 6. Establish 7. Measure vegetation structure and composition areas and at sage grouse use sites. literature lek counts pertaining to the impacts of burning on avian species. or sites. (minimum of 3) from study area leks. post-breeding transects movements of radio-marked and conduct breeding sage grouse. bird surveys at study areas. on treated and untreated STUDY AREAS Four areas were chosen for study. Three areas, Deer Creek, Perdiz, and Fish Hatchery leks were in North Park, Colorado. Another site, Thornburg Well, was in northwestern Colorado, approximately 8 km west of Maybell, in Moffat County. The Thornburg Well and Perdiz sites were burned by wildfire in 1987 and were approximately 1200 and 120 ha in size, respectively. The Deer Creek study area was burned by prescribed fire in fall 1987. Sagebrush at the wildfire sites has been reduced by nearly 100%. Some small areas of sagebrush survived within the Thornburg Well burns. The Deer Creek prescribed burn removed approximately 50-70% of the sagebrush cover in a 38 ha area. A more detailed description of study sites will be provided in later progress reports. �24 METHODS Sage Grouse Capture and Radio-marking.--Sage grouse were captured at night using a spotlight and a long-handled net on or near leks (Giesen et al. 1982). Birds were banded with serially numbered aluminum leg bands and plastic bandettes color coded by year. Age, sex, weight, and primary molt were recorded for each bird captured. Thirty-seven sage grouse were fitted with either poncho (Amstrup 1980) or tail-clip (Bray and Corner 1972) radio transmitters obtained from Wildlife Materials (Carbondale, IL), Telemetry Systems (Mequon, WI), and Advanced Telemetry Systems (Isanti, MN). Radio package weights ranged from 19.0 to 36.6 g and represented 0.7-1.4% of male body weights and 1.2-1.7% of female body weights. Relocations.--Radio-marked sage grouse were relocated on spring ranges from 14 April to 6 June. Birds were flushed and the location was marked to facilitate habitat measurements. Group size and flushing distance were also noted. An effort was made to flush all birds in the immediate area. Locations were plotted on 7.S-minute U.S.G.S. topographic maps. UTM map coordinates were obtained and used for distance calculations. Radio-marked sage grouse in North Park were relocated at approximately weekly intervals from June through September. Four attempts were made to locate radio-marked sage grouse in Moffat County resulting in a limited sample size for the Thornburg Well Lek. Average daily spring movements of sage grouse were obtained for 29 male sage grouse during the breeding season. Nineteen sage grouse had active radios during the summer months. Lek Counts.--Counts of male and female sage grouse were conducted between S April and 23 May. Birds present on leks were counted 3-S times during a lS-20 minute interval during each lek visit. Counts occurred within 1 hour of sunrise and were made from a distance of 30-1000 m depending upon weather, access, and observer preference. Breeding Bird Surveys Transects for breeding bird surveys were established at the Deer Creek, Perdiz, and Thornburg Well study sites. Transects varied in length from 4S0 to 1000 m with lengths varying because of features of the habitats to be sampled. Width of the transect was 140 m. Transects were marked at SO-m intervals and counts were conducted between 27 April and 21 June. Counts started O.S hour before sunrise and were completed within 2 hours of sunrise. A grid form was used to plot the estimated distance of birds from the transect. Singing males and non-singing individuals were recorded. Calculation of species diversity, evenness, and richness followed Wiens and Rotenberry (1981). Only singing males were used for calculation of birds per 40 ha, re1acive abundance, species diversity, evenness, and richness on all transects. Counts were averaged and then grouped by burn, burn edge, and unburned. These values were then converted to birds per 40 ha. The only exception was the unequal �25 length burn transects at the Deer Creek site. birds per 40 ha and then averaged. These values were converted to Habitat Measurements , Vegetation measurements were obtained from sage grouse spring feeding/loafing sites, a limited sample of random sites, and from 10-11 plots within each burn. Measurements were taken using Canfield's (1941) line-intercept technique and a modified Daubenmire (1968) plot. Two 10-m lines were measured at each location. Lines were oriented north-south and east-west centered over sage grouse flush locations or at random locations. Percent shrubs, forbs, and grasses were estimated to the nearest 5%, or if greater than 5%, at 20 0.5-m2 plots (lO/line at l-m intervals). Percentages less than 5% were estimated to the nearest 1%. Canfield's technique was used to measure the canopy and subcanopy while the 0.5-m2 plots were used to examine ground cover at or just above the soil surface. Measurements were summarized for male sage grouse locations and 31 burn plots within the 3 burns. Percent sagebrush was derived from line transect data. Percent shrub values were derived from 0.5m2 plots and included all shrubs and half shrubs. RESULTS AND DISCUSSION t Sage Grouse Lek Counts Lek counts at all study sites were similar to counts of past years (Table 1). Trends over the past 10 years suggest that sage grouse populations in North Park were low in 1988. systematic counts at Thornburg Well Lek had not been conducted for 6 of 11 years since initial location making assessment difficult. However, the trend also appears to be down. Counts in post-burn years should provide a clearer picture of population levels following fire. Sage Grouse Relocations Thirty-seven sage grouse were captured between 17 April and 11 May. Eleven adult and yearling male sage grouse (8 and 3, respectively) were radio-marked at Fish hatchery, 11 (7 and 4) at Deer Creek, 4 (3 and 1) at Perdiz, and 8 (7 and 1) at Thornburg Well. Radio-marked sage grouse were relocated on spring use sites from 14 April to 6 June for a total of 196 locations. Based on 75 relocations, average daily movements of radio-marked male sage grouse at the Deer Creek study site were 0.9 km (0.2-1.8 km). While several grouse were relocated near the burn edge, only 1 radio-marked sage grouse was located within the burn (Fig. 1). Average daily movements at the Fish Hatchery site (Fig. 2) were 0.9 km with a range of 0.1-2.2 km (n - 67). A small sample of relocations was obtained for the Perdiz study site. The average daily movement at Perdiz was 1.0 km (0.71.5 km) (Fig. 3). Average daily movements of radio-marked male sage grouse at the Thornburg Well site (Fig. 4) were 0.8 km (0.1-1.8) (n - 41). Four radio-marked male sage grouse at Thornburg Well were observed within the burn early in the breeding season (14 Apr). These birds were observed within small patches of unburned sagebrush. �26 Table 1. 1988. Sage grouse lek counts, Jackson and Moffat counties, April-May, High count Males Females N counts Dates of high count Males Females Peak male counts 1987 1986 1985 1984 1983 1982 1981 1980 1979 1978 1977 1976 1975 1974 1973 Deer Creekb Fish Hatchery (contro1)b 18 13 36 4 13 2 20 15 4 5 5 5 11 May 29 Apr 29 Apr 22 Apr 23 21 45 14 47 66 52 28 43 41 31 36 27 11 37 35 32 26 16 63 50 67 78 82 97 64 81 69 62 67 Perdizb 14 May 22 Apr Thornburg Well 25 Apr 5 Apr 16 6 10 8 21 27 23 8 16c 17a 15a Sa NC 8a 8a 20b 28b 3Sb 32b 36b,c aNo systematic counts, NC - no count. bSystematic counts. cYear of initial location. Radio-marked sage grouse were relocated on summer ranges from June through September. Most marked sage grouse in North Park moved to (or adjacent to) low lying wet meadow areas. Average distances of individual sage grouse relative to 1ek areas were: Deer Creek (n - 5) Fish Hatchery (n - 8) Perdiz en - 3) Thornburg Well (n = 3) 2.4 14.9 4.0 1.6 km km km km (1.1-7.2) (8.7-28.5) (3.7-4.1) (1.5-1.7) Long distance movements were observed at the Fish Hatchery study site. birds from this 1ek were followed to an area 2.4 km northwest of Rand, Two �> \: 1- ••••. ~ "" , 19 "" \. \. "----' , ~' " u u \~ "~~.j~ I ••,> ~.• ~.•o ~~~ o ...,. \ \ \ " ... ~ 27 ,-' \\ ')~ R7Bv)\" •.•. ~ v. " ~,. '\. , ') ~"" ~. \ -/. f ~ ~~ ~~ ~ " , ,.,. 8'''.5 ", '. {-.fa. •· ..f· ,...,.~ ,. \ D r/ t .. '. ... , ,..~ ,,'11';" ¢I.,. . '71-- I t .' ." ."\ .. ' \.' . "',\ " . .•.... .-..., "~ ,.j ' .'11\ '. '~~ \\ ' " --~"~-'f-~~~~~~~~--~~--~~--~~~~ __ -J'-"--~~-'r~~' 5,.•. " d ... "\\ \ ,I J • ,. __ • seu. ••• ){j~~<r_,- ,__ ";_._";..":.L._.J_. 7Wir _ ~ \' Fig. 1. Male sage grouse use sites (n = 80), Deer Creek study area, Jackson County, Colorado, spring 1988. Hatched outline represents proposed study area. Dotted line is approximate location of 1987 controlled burn. �28 4 -- ,..___ - ,.- .18 -lOr"",..., "> ..» I , ,1 16 --.. ~_..: •• r- •.••~ , i _, ~ .~, tlO' .....--, J_~ =.. ~\/) ',:> \II~ " ..-r: /_.' ,f ~ ". l8,t \ ":" ~ 1.,.. I j. II .... _ "02 - 'l!~ •••• ! " ",_I~I~" '\ ( --) -: .../') r Fig. 2. Male sage grouse use sites (£ = 67), Fish Hatchery study area, North Park, Colorado, spring 1988. Hatched outline represents proposed study area. �29 . \ _-:~ l .. ......~ .....:. .: '. •••IfJ1S 6.14 <') ,l ,. , . .,' ..J __,. - "",5 _ .' . J:. ; ~'. . I,. '. ," / I /i \ \ ( •.•.~~,.. .... : ....•.....•• ••••• I ~ •• •• , _..I' 17'.~,'-.:.•............ ..•••.....• ) ~- .: Fig. 3. Sage grouse use sites (Q = 12), Perdiz study area, Jackson County, Colorado, spring 1988. Hatched outline represents proposed study area. Line with XIS is approximate outline of 1987 wildfire. �, « t \ t fig. 4. Male sage grouse use sites (~= 37), Thornburg Well study area, Moffat County, Colorado, spring 1988. Line with XIS is approximate west edge of the 1987 wildfire. Hatched outline represents proposed study area. �31 Colorado, a distance of 26 km. One of these birds moved east 4.6 km and was observed in a flock with radio-marked sage grouse from the Deer Creek Lek. This bird was shot during the 1988 grouse hunting season. Only 1 additional radio-marked bird was recovered during the open season. Sage grouse in North Park were not observed in burns during the summer. Sage grouse at the Thornburg Well site did not exhibit a similar summer movement pattern. Radio-marked grouse were observed over summer on 4 occasions. Male sage grouse were observed either in the burn or in unburned areas of sagebrush within the burn on 3 different dates during June and July. One successful nesting hen was observed in a brood flock adjacent to the burn on 2 dates. This hen was predated (raptor) and her radio was recovered on 19 August. The other radio-marked female (also successful) was located from the air on 17 June approximately 3 km northwest of the lek area and 2 km south of the Yampa River valley. No locations were obtained after this date for this hen. Passerines r Breeding Bird Survey.--Breeding bird surveys were conducted between 27 April and 21 June. Transects at the Thornburg Well study site were counted 4 times between 27 April and 20 May. Transects at the Deer Creek study site were counted 4 times between 11 May and 10 June while those at the Perdiz study site were counted 5 times between 6 May and 21 June (Tables 2-4). D • Table 2. Birds (total) observed during 4 surveys of transects at the Deer Creek study site, Jackson County, Colorado, May-June 1988. Burned Unburned Species 1 2 3 Brewer's sparrow Vesper sparrow Sage thrasher Horned lark Green-tailed towhee Brewer's blackbird Mountain bluebird 11 9 3 1 3 3 4 0 4 2 0 15 16 1 4 5 0 0 0 0 0 Totals 27 41 13 I ~ 4 8 15 1 16 0 1 2 43 �32 Table 3. Birds (total) observed during 5 surveys of transects at the Perdiz study site, Jackson County, Colorado, May-June 1988. Unburned Species 1 Vesper sparrow Brewer's sparrow Sage thrasher Green-tailed towhee Horned lark Brewer's blackbird Mountain bluebird Barn swallow 8 70 7 3 Totals Edge 3 2 9 86 7 6 1 0 Burned 4 5 2 1 17 8 0 10 1 0 0 0 1 0 1 0 0 2 0 0 0 11 9 0 0 90 112 38 22 3 0 0 6 5 2 0 14 , , Table 4. Birds (total) observed during 4 surveys of transects at the Thornburg Well study site, Moffat County, Colorado, April-May 1988. Unburned Species 2 1 Vesper sparrow Brewer's sparrow Sage thrasher Horned lark Green-tailed towhee Sage thrasher Western meadowlark Lark sparrow Lark bunting 23 Totals Edge 3 15 7 Burned 4 5 0 0 0 \ 3 ~ 4 7 11 0 18 13 10 0 2 2 5 6 0 0 7 1 1 0 9 0 0 0 3 0 3 3 0 0 0 0 2 54 52 34 11 2 0 1 The most obvious finding was a substantial reduction in birds per 40.5 ha on burned transects in comparison with unburned transects (Table 5). At Deer Creek, birds per 40.5 ha were 40 in the unburned area and 18 in the burned area, a difference of 56%. At Perdiz, this value varied from 39 in unburned transects to 7 for the edge transect to 4 birds/40.5 ha in the burn. Data from Thornburg Well show a similar pattern of decreased density within the burn. 0 0 0 1 0 �33 Table 5. Species diversity, evenness, richness, and density on breeding bird transects in Jackson and Moffat counties, Colorado, Spring, 1988. Deer Creek Burned Unburned Species Diversity Evenness Richness Birds/40.5 ha 2.96 0.84 5.00 40.00 3.52 0.91 5.00 18.00 Unburned Perdiz Edge Burn 1.74 0.76 4.00 39.00 2.90 0.98 3.00 7.00 1.00 1.00 1.00 4.00 Thornburg Well Unburned Edge 4.00 0.84 6.00 48.00 3.44 0.85 5.00 32.00 Burn 3.44 0.87 5.00 25.00 Relative abundance (Table 6) varied markedly between burn and unburned areas. Vesper sparrows (Spizella breweri), and green-tailed towhees (Pipilo chlorurus) had lower relative abundances between burned and unburned transects. Relative abundance and density of horned larks (Eremophila alpestris) were greater in burned transects at all study sites. Species diversity (Table 5) was greater within the burn transects at Deer Creek but was lower at the wildfire sites. r t Table 6. Relative abundance (X) of breeding birds at study sites in Jackson and Moffat counties, Colorado, Spring, 1988. Deer Creek Unburned Burned Vesper sparrow Brewer's sparrow Sage thrasher Green-tailed towhee Horned lark Sage sparrow Lark sparrow Western meadowlark NP = 29 48 2 12 9 NP NP NP 28 28 3 6 35 NP NP NP Unburned 13 74 4 9 NP NP NP NP Perdiz Edge NP 29 29 NP 42 NP NP NP Burn NP NP NP NP 100 NP NP NP Thornburg Well Edge Unburned 40 20 12 4 NP 17 NP 7 12 12 43 6 NP 27 NP NP Burn 8 NP 23 NP 42 23 4 NP not present. Large differences were observed between transects in burned versus unburned areas at the 2 wildfire sites. Differences at the controlled burn site were less distinct. Larger numbers and species of breeding birds were observed during later counts at the wildfire study sites. This appeared to be due to the transition from mostly bare ground to a partial grass cover. Habitat Measurements A total of 177 locations was measured for habitat parameters at all study sites (Table 7). Averages for most variables appear to be within the ranges previously reported for North Park (Schoenberg 1982, Hernandez 1988). The percentage of sagebrush was somewhat lower because of some locations within burns. �34 Table 7. Vegetative parameters at sage grouse use sites in Jackson and Moffat counties, Colorado, 1988. Deer Creek (n - 59) Sagebrush Livea Fish Hatchery (n - 41) Thornburg Well (n - 20) 2.1 (0.4-5.0) 0.5 (0-1.8) 28.2 (3.3-50.9) 1.5 (0.8-2.8) 0.4 (0.2-1.0) 27.4 (4.2-44.2) 1.4 (0.4-2.3) 0.3 (0-1.0) 27.2 (7.9-44.8) Ave height, cm Ave width, cm Ave length, cm 34.3 (17.0-72.2) 43.9 (17.2-84.4) 35.7 (11.9-65.3) 36.4 (18.1-56.1) 50.6 (16.8-73.9) 42.7 (16.1-61.0) 50.6 (30.8-77.0) 60.5 (38.5-98.8) 51.9 (33.2-88.1) Ave space, cmb,c 75.0 (29.9-321.8) 85.3 (36.5-175.6) 99.9 (53.0-240.9) Ave shrub, %d Ave forb, %d Ave grass, %d 9.3 (0.1-21.0) 7.0 (0-24.4) 13.0 (3.4-73.2) 13.0 (5.4-24.4) 5.5 (0.5-20.6) 7.4 (2.2-24.0) Dead" Pe r c errtf 7.9 (3.2-14.2) 2.5 (0-7.2) 6.4 (3.5-12.1) , 1 aSagebrush p1ants/m2. bLine-transect data. cDistance between sagebrush plants. dDerived from 0.5-m2 plots. Vegetation measurements for burn sites varied (Table 8). Burns were characterized by a lack of sagebrush and low percentages of grass and forbs. Broom snakeweed (Gutierrezia sarothrae) and rabbitbrush (Chrysothamnus §£Q.) were the predominate shrubs within the burns. Table 8. Habitat parameters on burn plots in Jackson and Moffat counties, Colorado, 1988. Sagebrush Livea Dead" Pe rc errt" Ave shrub, %c Ave forb, %c Ave grass, %c Deer Creek Perdiz Thornburg Well 3 (1.9-5.3) 0.8 (0-4.8) 0.06 (0-0.2) 2.4 (2.0-3.2) 0.5 (0-2.2) 0.07 (0-0.3) 1.5 (0.4-2.2) 0.4 (0-3.0) 2.5 (0-10.4) 1.3 (0.2-3.4) 5.1 (2.3-7.3) 0.4 (0-1.9) 0.7 (0.1-1.9) 3.9 (1.6-5.8) 0.4 (0-1.5) 1.4 (0-3.8) 4.4 (2.8-7.1) o aSagebrush plants/m2. bLine-transect data. cDerived from 0.5-m2 plots. I C �35 Measurements of random locations and more complete habitat analysis will be provided in future progress reports. LITERATURE CITED Amstrup, S. C. 1980. 44:214-217. A radio collar for game birds. J. Wildl. Manage. Beetle, A. A. 1960. A study of sagebrush. The section (Tridentatae of Artemisia. Wyoming Agric. Exp. Stn. Bull. 368. Braun, C. E., M. F. Baker, R. L. Eng, J. S. Gashwiler, and M. H. Schroeder. 1976. Conservation committee report on effects of alteration of sagebrush communities on the associated avifauna. Wilson Bull. 88:165171. Bray, O. E., and G. W. Corner. 1972. A tail clip for attaching transmitters to birds. J. Wi1dl. Manage. 36:640-642. Canfield, R. H. 1941. Applications of the line interception method in sampling range vegetation. J. For. 39:388-394. Daubenmire, R. 266. 1968. Ecology of fire in grasslands. Adv. Ecol. Res. 5:209- Frandsen, O. A. 1985. Fire as a management tool in southeast Idaho - a case study. Pages 85-87 in K. Sanders and J. Durham, eds. Rangeland fire effects: a symposium. U.S. Bur. Land Manage., Boise, Idaho. Giesen, K. M., T. J. Schoenberg, and C. E. Braun. 1982. Methods for trapping sage grouse in Colorado. Wi1dl. Soc. Bull. 10:224-231. Hernandez, E. J. 1988. Response of selected avifauna to prescribed burning in the big sagebrush type. Colorado Div. Wildl. Job Progress Rep., Fed. Aid Proj. W-1S2-R. Apr:105-l38. Schroeder, M. H., and D. L. Sturges. 1975. The effect on the Brewer's sparrow of spraying big sagebrush. J. Range Manage. 28:294-297. Schoenberg, T. J. 1982. Sage grouse movements and habitat selection in North Park, Colorado. M.S. Thesis, Colorado State Univ., Fort Collins. 86 pp. Wiens, J. A., and J. T. Rotenberry. 1981. Habitat associations and community structure of birds in shrubsteppe environments. Eco1. Monogr. 5:21-41. Prepared by l...~C:~~.LL...L:::.~~:::::::===::_ Lee A. Benson Graduate Research Assistant Approved by C1ait E. Braun Wildlife Research Leader ��37 Colorado Division of Wildlife Wildlife Research Report April 1989 JOB PROGRESS REPORT State of: Colorado Project: W-152-R Avian Research Work Plan: 8__ : Job Job Title: Population Inventory and Habitat Use by Lesser Prairie-chickens in Southeast Colorado Period Covered: Author: 5__ 01 January 1988 through 31 December 1988 Kenneth M. Giesen Personnel: Clait Braun, Dave Clarkson, Ken Giesen, Rick Hoffman, Diane Picken, Jennie Slater, Chuck Wagner, Bryant Will, Colorado Division of Wildlife ABSTRACT Lek surveys were conducted on a 4l.4-km2 study site and surrounding area of the Comanche National Grasslands. Both lek numbers and total numbers of lesser prairie-chickens (Tympanuchus pallidicinctus) counted increased from previous years, suggesting the population increased in 1988. There was a positive correlation between male breeding density and number of active leks (I = 0.94) but not between male breeding density and average number of males/lek (I = 0.30). Experimental roadside lek surveys resulted in 77 to 100% of active leks being detected. Leks not detected were along the transect with the highest lek density. Thirty-eight prairie-chickens were trapped and banded with 8 males and 19 females being fitted with radio transmitters. Average home range was 158 ha. Hens dispersed 0.4 to 2.4 km from leks where marked to nest sites (R = 1.4 km). Average size of initial clutches was 11.2 eggs. Two hens attempted renests. Residual vegetation from 160 random transects on the study area was analyzed to quantify available habitat. Existing vegetation was measured ~t all nest sites, 51 grou~e_flus~ sites, and 51 paired random sites to document habitat preference. Nest sites hadtal1er vegetation than random sites while no apparent differences were detected between 7 habitat variables measured at grouse flush and random sites. ��39 POPULATION INVENTORY AND HABITAT USE BY LESSER PRAIRIE-CHICKENS IN SOUTHEAST COLORADO Kenneth M. Giesen Both distribution and populations of lesser prairie-chickens in North America have decreased> 90% from historic levels of the 1800's (Taylor and Guthery 1980). Although the exact historic distribution of lesser prairie-chickens is unknown, early reports (Bendire 1892, Judd 1905, Bent 1932, Baker 1953, Sands 1978) suggested they were abundant and widely distributed throughout their range. Aldrich (1963) indicated lesser prairie-chickens historically inhabited about 360,000 km2 in 5 states while recent estimates suggest a population of 50,000 birds on 125,000 km2 (Crawford 1980, Taylor and Guthery 1980, Johnsgard 1983). Although evidence suggests lesser prairie-chickens were historically peripheral in Colorado, they were thought to be common to abundant in 6 southeastern counties (Baca, Prowers, Bent, Kiowa, Lincoln, and Cheyenne), and peripheral in adjacent counties (Loeffler 1983). Recent surveys have documented breeding populations in Baca, Prowers, and Kiowa counties (Hoffman 1963, Loeffler 1983, Rash 1985, this study). Efforts to transplant lesser prairie-chickens into Lincoln county in 1968 apparently failed and the outcome of a 1988 transplant to Pueblo County is unknown. Because of apparent small population and restricted range, the lesser prairie-chicken is classified as a threatened species in Colorado. P. N. OBJECTIVES The objectives of this study are to evaluate lek surveys as indices to population trends, ascertain the accuracy of aerial and ground surveys in detecting leks, describe the seasonal floristic and structural characteristics of lesser prairie-chicken habitats in southeast Colorado, and contribute to preparation of a recovery plan for lesser prairie-chickens in Colorado. SEGMENT OBJECTIVES 1. Review pertinent literature applicable to the objectives of this study. 2a. Locate all active leks within a 41.4 km2 primary study area and obtain at least 1 count/week (Mar-May) of all males and females on each lek. 2b. Survey all historic leks in Baca County and obtain at least 1 count of males on each active lek. 3. Select 4 20-km long roadside listening transects within the range of lesser prairie-chickens in Baca County and test the ability of naive observers to detect active lesser prairie-chicken leks within 1.6 km of the transects. �40 4. Trap and band lesser prairie-chickens on active leks within the primary study area. Up to 20 will be marked with miniature radio transmitters to facilitate their periodic location. 5. Locate lesser prairie-chicken nests by following radio-marked Record clutch size, incubation period, and nest fate. 6. Locate radio-marked range. 7. Measure vegetative cover at grouse use and random sites. Height and canopy cover of shrubs, forbs, and grasses will be recorded. 8. Compile birds weekly data, analyze results, for estimates and prepare of movement annual progress hens. and home report. METHODS Field surveys were conducted on the Comanche National Grasslands and adjacent areas in eastern Baca County from March through June using binoculars, a parabolic microphone listening device, and a trained pointing dog to locate active lesser prairie-chicken leks. Leks known to be active in 1987 were surveyed as well as most known historic lek sites. Active leks were visited within 2 hours of sunrise to count grouse and classify them to sex. Lek density was defined as the number of active leks on the 41.4 km2 study area and breeding male density was the sum of the high counts of males on each lek on the study area expressed as male/km2. Two roadside transects were conducted by naive observers during the peak of lesser prairie-chicken breeding (19-20 Apr). A 3-minute listening stop was taken approximately every 1.6 km to listen for displaying grouse. Distance and direction to detected leks were estimated and plotted on a map. Routes began 15 minutes prior to sunrise and were completed within 2 hours. Cannon-nets and funnel traps (Giesen et al. 1982) were used to capture grouse on leks. Each captured grouse was marked with a numbered aluminum band and a unique combination of colored plastic bandettes. Miniature solar- or battery-powered transmitters (weight 18-24 g) were attached to all captured females and selected males using a poncho (Amstrup 1980). Radio-marked birds were located using a portable receiver and hand-held yagi antenna. Birds were approached on foot until they flushed. Vegetative structure and species composition were measured using line-intercept of canopy cover (Canfield 1941) and a range pole (Robel et al. 1970). Nomenclature follows Harrington (1964). Sand sagebrush (Artemisia filifolia) density was measured on 0.001 ha circular plots. The minimum convex polygon method (Mohr 1947) was used to calculate home range sizes. Study Area The primary study area was a 41.4 km2 (16 mi2) area of rangeland on or adjacent to Pasture lAE of the Comanche National Grasslands. Included were sections 21-28 and 33-36 of T34S, R44W and sections 1-4 of T35S, R44W. Approximately 1160 ha was privately owned and was rangeland except for a 16 ha milo (fallow in alternate years) field. The topography was rolling hills bisected by Murray Draw and Mitchell Draw. Elevation ranged from 1240 m on �41 the west to 1070 m on the east. Soils were predominately fine sands in the draws (U.S.D.A. Soil Conserv. Service). sandy loams with Topography, grazing, ana revegetation have resulted in a diverse vegetative community. Sand sagebrush, broom snakeweed (Gutierrezia sarothrae), and yucca (Yucca glauca) were the predominant shrubs. A variety of grasses occurred with sand dropseed (Sporobolus cryptandrus), 3-awn (Aristida longiseta), sideoats grama (Bouteloua curtipendula), and blue grama (~. gracilis) being most widespread. Seven habitat variables were measured in March-May 1986-87 to quantify existing cover (Table 1). Except for forbs, the height measurements were from residual vegetation. Table 1. 1986-87. Habitat characteristics of a 41.4 km2 lesser prairie-chicken study area, Baca County, Colorado, Location Sec. Sec. Sec. Sec. Sec. Sec. Sec. Sec. Sec. Sec. Sec. Sec. Sec. Sec. Sec. Sec. 21, T34S, R44W 22, T34S, R44W 23, T34S, R44W 24, T34S, R44W 25, T34S, R44W 26, T34S, R44W 27, T34S, R44W 28, T34S, R44W 33, T34S, R44W 34, T34S, R44W 35, T34S, R44W 36, T34S, R44W 1, T35S, R44W 2, T35S, R44W 3, T35S, R44W 4, T35S, R44W Mean Heightdensity Bare ground Shrub Height ~cm2 Forb 40.5 44.6 40.0 33.6 31.5 31.0 29.2 29.8 35.5 36.6 22.2 33.8 28.1 31.1 42.9 32.8 3.3 3.8 3.6 3.1 4.6 4.5 1.5 1.7 3.2 4.7 5.2 1.9 2.0 4.5 3.6 6.0 6.0 27.1 22.5 16.9 13.9 17.2 6.6 7.4 11.0 15.5 12.0 4.4 9.7 9.9 17.4 11.2 0.28 0.90 1.16 1.46 0.87 1.58 0.32 0.79 1.67 1.18 0.97 0.78 1.33 0.57 1.07 0.78 81.2 59.9 65.4 76.1 86.7 70.4 55.2 84.5 81.9 72.0 34.6 3.6 13.0 0.98 Grass X Sand sagebrush Canopy cover Density 79.4 89.8 74.2 73.4 78.6 0 2360 4160 7480 2467 2840 2067 4266 2800 1560 2880 533 667 1520 1267 3080 0.0 7.1 10.0 5.4 3.1 4.6 0.6 3.0 5.1 2.6 3.5 0.2 0.0 0.9 1.9 1.5 75.4 2497 3.1 n.8 RESULTS AND DISCUSSION Lek Surveys A total of 185 counts of 26 active leks was obtained in Baca County between 18 March and 17 May 1988 (Table 2). In addition, 6 historic lek sites were surveyed and found to be inactive. A minimum of 202 males, 24 females, and 289 total birds was counted for an average 11.1 birds/active lek, up 18% from 1987. Several long-term active leks south of the Cimarron River were not surveyed in 1988. Summaries of lek count data since 1977 (when written records became available for analysis) indicate the dynamic nature of leks with many becoming inactive and others being established (Table 3). Fifteen of 57 leks (26%) active one or more years since 1977, were initially located in 1985-88 as a result of more search effort and possibly increasing lesser prairie-chicken populations. �42 Because numbers of grouse attending leks fluctuates during the breeding season, and survey timing and intensity varied each year, the data (Tables 2, 3) may not reflect actual population changes. Table 2. Lesser prairie-chicken lek count data, Baca County, 1988.4 Lek N counts 2 3 4 5 6 7 12 14 17 18 2S 27 28 33 35 39 40 86-1 86-3 87-1 87-2 87-4 87-5 88-1 88-2 88-3 18 IS 16 24 9 6 1 2 3 8 1 7 23 9 1 2 2 16 1 1 1 11 2 2 3 1 Total 185 Mean Count period 18 18 18 18 01 01 31 07 18 07 18 06 01 05 18 30 10 07 21 Mar-17 Mar-30 Mar-03 Mar-17 Apr-OS Apr-OS 10 May Mar-07 Apr-27 Mar-05 10 Apr Apr-20 Mar-13 Apr-OS 07 Apr Apr-05 May-06 Mar-10 28 Apr 23 Apr 04 May Mar-06 Apr-19 May-28 Apr-06 06 May May Apr May May May May May Apr May Apr May May Apr May May Males High count Females Tota1sb 12 11 6 23 6 12 2 6 11 11 5 7 20 11 16 7 3 1 1 4 12 2 15 6 5 9 2 4 3 5 1 14 4 3 6 8 202 24 1 1 1 5 2 15 13 9 25 6 16 4 13 14 13 8 7 25 15+ 16 14 6 3 9 May Apr Apr May 9.2 2.0 289 11.1 aLeks 8, 13, 21, 32, 38, 86-2, were surveyed and no birds observed. bIncludes males, females, and birds not classified to sex. Analyses of 1ek surveys and 1ek count data (Table 4) indicates a strong positive correlation between males breeding density and number of active leks (£ ~ 0.94) but not bet~een male breeding density and average lek size (£ 0.30). If the high count of males on leks was a constant proportion of the total population, lek density is a better index to population changes than average lek size. These data support results of prairie grouse population �43 Table 3. Lek High counts of male lesser prairie-chickens, Baca, COU'lty,1977-88. 1977 1978 4 18 8 9 24 22 15 1 0 5 14 6 8 NC 15 11 NC HC NC NC 5 13 11 11 0 4 14 HC 0 4 13 16 6 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 86-1 86-2 86-3 87-1 ' 87-2 87-3 87-4 87-5 88-1 88-2 88-3 0 7 0 7 17 17 16 2 3 3 5 aHo bLek b 1979.- 1980 HCa 10 2 7 12 15 11 0 0 0 NC 7 12 10 12 NC 3 NC HC 2 1 6 17 NC count. not yet located. 0 1 17 12 15 14 11 0 0 0 0 17 8 9 9 0 0 9 0 0 0 4 19 0 3 9 7 7 1981 Year 1982 1983 0 6 15 7 16 14 6 0 0 0 0 20 8 8 0 3 12 7 29 10 9 0 0 0 0 14 7 6 0 0 0 0 0 9 0 0 0 3 30 0 6 4 4 7 4 18 14 6 8 6 0 10 0 0 0 0 33 0 11 2 4 4 0 10 17 0 6 0 10 12 0 12 11 9 26 14 8 0 0 0 0 16 5 12 0 0 0 9 0 0 0 0 23 0 3 2 3 4 0 0 13 0 0 0 16 7 12 7 10 4 1984 1985 1986 1987 1988 0 9 7 7 23 18 7 0 0 0 7 7 0 18 6 5 0 0 HC 16 2 0 17 6 9 NC NC NC NC 5 0 10 NC NC 4 10 NC HC 0 HC 2 HC 2 0 0 12 0 0 2 0 0 0 10 0 12 2 11 8 3 0 5 HC 9 3 3 18 3 8 HC 12 11 6 23 6 12 0 0 0 18 0 0 0 11 0 0 5 9 0 0 0 0 11 0 0 2 5 0 0 0 0 21 0 2 3 0 0 0 6 11 0 0 0 21 0 10 8 18 5 3 14 0 16 0 0 0 4 10 2 2 7 5 0 0 13 0 12 3 9 7 7 7 7 3 4 4 0 0 1 15 6 7 0 0 NC 0 HC 4 0 8 NC HC 7 13 HC HC 0 HC 0 HC 6 0 0 17 0 HC 9 0 8 HC 18 0 18 7 0 9 0 10 NC 5 NC 0 11 5 8 0 0 0 7 5 HC HC HC 2 0 6 NC NC 11 11 HC NC 0 HC HC HC 5 HC 720 HC HC HC 0 11 HC 16 NC NC 0 7 6 HC HC HC NC NC HC 12 0 0 0 0 HC 9 4 3 5 0 �44 surveys in other states (Cannon and Knopf 1981, Martin and Knopf 1981). These results suggest lesser prairie-chicken population changes are reflected in the total number of active leks rather than the average number of birds/1ek. Thus, inventory methodology should incorporate plans to document annual status (active/inactive) of historic leks and efforts to search for additional leks. Table 4. Lek count trends of lesser prairie-chickens study area, Baca County, Colorado, 1980-88. N N Year leks males ~ ma1es/1ek 1980 1981 1982 1983 1984 1985 1986 1987 1988 6 6 6 6 4 5 7 8 9 59 55 59 65 46 46 75 74 100 9.8 9.2 9.8 10.8 11.5 9.2 10.7 9.2 11.1 on a 41.4-km2 (16 mi2) Breeding density (ma1es/km2) 1.42 1.33 1.42 1.57 1.11 1.11 1.81 1.79 2.42 Lek Transects Lek survey routes are used by many state wildlife agencies to monitor trends in prairie grouse populations. Although lek density correlates well with male breeding density, there have been no studies to evaluate the accuracy or precision of transects in detecting active leks. Surveys of 8 transects in 1987 resulted in only 40.7% of active leks being detected. Weather and observer bias were detected suggesting that 1ek survey rates may not be useful in monitoring lesser prairie-chicken populations unless the count variability can be controlled. In 1988, observers detected 76.9 to 100% of active leks within 1.6 km of the transect (Table 5). The increase in detection rate was attributed to favorable weather conditions (no wind) and conducting surveys during the peak of hen attendance when display activity was highest. All active leks were detected on the transect with low grouse density but 7.7 to 23.1% were missed where grouse densities were higher. At high lek densities, it was possible to detect 3 or more leks from some listening stops. In those situations, the larger or closest leks appeared to overwhelm the display sounds from smaller or farther leks, thus causing them to be missed. Naive observers (having no knowledge of lek locations) were used to evaluate survey accuracy in these experiments. For management purposes, the observers would most likely be the District Wildlife Manager or Area Biologist who would know the location of historic leks, thus needing only to verify their current status (active/inactive) and search for newly established leks. This should increase the accuracy of the survey routes. �45 Table 5. Leks detected County, Colorado, 1988. Transect 1 2 Length during experimental aNumber of active leks within Trapping transects in Baca n leks detected Observer Observer 1 (km) 19.2 19.2 roadside 13 4 12 4 2 10 4 1.6 km of transects. and Banding A total of 33 lesser prairie-chickens (19 males, 14 females) was trapped on 5 of 9 active leks within the primary study area. Five females were trapped on 2 leks adjacent to the study area. Most females (16 of 19, 84.2%) were captured between 13 and 23 April which corresponded to the peak of hen attendance at leks. Five of 9 hens captured prior to 18 April were adults compared to 1 of 10 hens captured after 18 April suggesting that adult hens breed earlier than yearling hens. All captured prairie-chickens were marked with a serially numbered aluminum band and unique combinations of colored bands to facilitate future identification. Eight males and all 19 hens were fitted with poncho-mounted miniature radio transmitters (18-24 g, 2.6-3.3% of grouse weight) to facilitate their periodic relocation. Although the transmitters were not thought to affect the daily activity patterns or movements of grouse, at least 10 radio-marked grouse (2 males, 8 females) were killed by predators within 30 days. Radio contact was lost with 5 other hens because of radio-failure or predation. Increased predation due to radio-marking has been reported for other birds (Herzog 1979, Warner and Etter 1983, Hines and Zwickel 1985, Marks and Marks 1987). Nesting Fates of 19 hens were ascertained using radio-telemetry techniques. Radio signals from 7 hens were lost before their nests were located. It was most likely that radio failure or undetected predation occurred although long range (> 15 km) dispersal may have occurred. Radio signals were not detected from these hens in subsequent searches through October. Six hens were killed by predators prior to incubation and 4 hens were depredated during incubation prior to hatch. Of 8 nests located (including 2 renests), none was successful. Causes of high predation on hens and nests were unknown as nests were visited initially at the estimated onset of incubation to ascertain clutch size and subsequently when hen mortality or nest depredation was suspected. The average distance from capture site on leks to first nests was 1442 m (n 6) with 2 hens nesting closer to another lek. The average distance between initial clutches and renests was 700 m. Average clutch size of 4 complete first clutches was 11.2 eggs; the 2 renests contained 5 and 10 eggs. �46 Hens selected nest sites which had greater cover than surrounding areas. Forb height, grass height, and height-density were greater at nest sites than random sites (Table 6). Average shrub height, sandsage density, and canopy cover were also higher ~t nest sites but differences were not significant (f > 0.05). Because nesting occurs early in the season before significant vegetative growth, it may be beneficial to manage rangeland to increase the height and density of residual grass cover. Nest site measurements suggested that hens selected for clumps of tall grasses (> 30 cm) and height density measurements of > 2.8 dm. The low recorded nest success observed may indicate that high quality nesting cover may be limiting in the study area. Table 6. Habitat variables at lesser prairie-chicken nest sites, 10-m nest transects, and random 10'm transects in Baca County, Colorado, 1986-88. (Sample sizes in parentheses; letters within rows indicate a [~ < 0.05] difference for that variable). Variable Nest site Shrub height, ~ Forb height, ~ Grass height, ~ Height-density, dm Sandsage density, plants/ha Sandsage canopy cover, % Bare ground, % (16) (16) (16) (16) Nest transect 42.9 18.2A 31.4A 2.88AB (16) (16) (16) (16) (16) (16) (16) 34.4 14.9B 25.5B 1.86A 3292 6.4 73.9 Random transect (141) (150) (151) (158) (158) (158) (158) 34.4 3.6A3 13.3AB 1.18B 2524 3.0 75.2 Home Range Size Home range sizes were measured for 5 males and 5 females which were observed for more than 30 days (Table 7). Average home range size was 158 ± 145 ha with no difference in mean home range size between males and females. Home range sizes of males included their lek site while female home ranges (after breeding) were spatially separate from leks. Table 7. Home range size of radio-marked County, Colorado, 1988. lesser prairie-chickens in Baca Band Time interval Home range size (ha) 360 361 362 366 368 369 373 381 390 393 Age 2+ 2+ 1+ 1+ 2+ 2+ 2+ 1+ 1+ 1+ Sex M F M M F M F F M F 7 8 8 12 IS 15 15 19 22 22 Apr-22 Apr-23 Apr-ll Apr-27 Apr-25 Apr- 3 Apr-12 Apr-25 Apr-14 Apr- 1 May Aug Oct Oct May Jul Sep May Jun Jun 80 227 258 130 32 374 392 7 26 57 �47 Habitat Preference Habitat preference was investigated by comparing vegetative measurements at grouse flush sites and adjacent random sites. Preliminary analysis indicated no differences (~ > 0.05) in mean values for the 7 habitat variables measured at flush and random sites (Table 8). The height-density and sandsage canopy cover variables had different (~ < 0.05) distribution functions between the grouse flush sites and random sites suggesting that the grouse may have been avoiding the extreme values of these 2 habitat components. Table 8. Habitat variables at lesser prairie-chicken flush and paired random sites in Saca County, Colorado, 1987-88. Mean Variable Grouse-flush site Random site 37.8 13.3 16.6 1.89 5189 9.95 85.5 35.1 12.9 15.2 1.51 3~3 5.15 85.6 Shrub height, cm Forb height, cm Grass height, cm Height-density, dm Sandsage density, plants/ha Sandsage canopy cover, X Bare ground, X awilcoxin rank-sum test, Ho: !1 = f > Prob f > Ilia 0.1829 0.3661 0.5183 0.0727 0.1405 0.0783 0.8513 0.2829 0.7717 0.5043 0.0478 0.0789 0.0145 0.9619 !2' There may be several reasons I failed to detect habitat selection by lesser prairie-chickens. The grouse flush and paired random sites were relatively close (293 ± 108 m, range 55-541 m) and measured on the same day. If habitat was fairly uniform over large areas there would be little need for selection at the macro level. However, measurements of flush and random sites indicated high variation for each habitat component. Because of small sample sizes each month, the data were pooled. Only grass height and forb height increased throughout the season and the differences between flush and random sites remained constant. Because the study area contained the highest known density of lesser prairie-chickens in Colorado, all of the habitat may have been equally good. Sampling in areas with different grouse densities may provide better insight into habitat preference. Furthermore, habitat preference may be strongest when some components may be in short supply at critical times of the year (i.e., nesting cover, winter habitat). Summer habitat measurements may not detect these differences. Additional sampling should be done at other seasons. LITERATURE CITED Aldrich, J. W. 1963. Geographic Wildl. Manage. 27:529-545. Amstrup, S. C. 1980. 44:214-217. orientation A radio-collar of American for game birds. Tetraonidae. J. Wildl. Manage. J. �48 Baker, M. F. 1953. Misc. Publ. 5. Prairie chickens of Kansas. 68pp. Univ. Kansas Mus. Nat. Hist. Bendire, C. E. 1892. ~ife histories of North American birds with special reference to their breeding habits and eggs. U.S. Nat1. Mus. Spec. Bull. 1. 446pp. Bent, A. C. 1932. Life histories of North America gallinaceous Natl. Mus. Bull. 162. 490pp. Canfield, R. H. 1941. range vegetation. birds. U.S. Application of the line intercept method in sampling J. For. 39:338-394. Cannon, R. W., and F. L. Knopf. prairie grouse populations. 1981. Lek numbers as a trend index to J. Wildl. Manage. 45:776-778. Crawford, J. A. 1980. Status, problems, and research needs of the lesser prairie chicken. Pages 1-7 in P. A. Vohs and F. L. Knopf, eds. Proc. Prairie Grouse Symp. Oklahoma State Univ., Stillwater. Giesen, K. M., T. J. Schoenberg, and C. E. Braun. 1982. Methods sage grouse in Colorado. Wildl. Soc. Bull. 10:224-231. Harrington, H. D. 1964. Manual of the plants of Colorado. Swallow Press, Inc., Chicago. 666pp. for trapping Sage Books, Herzog, P. W. 1979. Effects of radio-marking on behavior, movements, survival of spruce grouse. J. Wildl. Manage. 43:316-323. Hines, J. E., and F. C. Zwickel. 1985. Influence of radio packages blue grouse. J. Wildl. Manage. 49:1050-1054. Hoffman, D. M. 1963. The lesser prairie chicken in Colorado. Manage. 27:726-732. Johnsgard, P. A. 1983. Lincoln. 413pp. The grouse of the world. J. Univ. Nebraska and on young Wildl. Press, Judd, S. D. 1905. The grouse and wild turkeys of the United States and their economic values. U.S. Dep. Agric. Biol. Surv. Bull. 24. 55pp. Loeffler, C. 1983. The status and management of the lesser prairie chicken in Colorado. Unpubl. Rep., Colorado Div. Wildl., Colorado Springs. 9pp. Marks, J. S., and V. S. Marks. 1987. Influence of radio-collars of sharp-tailed grouse. J. Wildl. Manage. 51:468-471. on survival Martin, S. A., and F. L. Knopf. 1981. Aerial survey of greater prairie chicken leks. Wildl. Soc. Bull. 9:219-221. Mohr, C. O. 1947. Table of equivalent populations mammals. Am. Midi. Nat. 37:223-249. of North American small �49 Rash, M. T. 1985. Survey of the lesser prairie-chicken in Colorado, 3 April25 May 1985. Unpub1. Rep., Colorado Div. Wi1d1., Colorado Springs. 22pp. Robel, R. J., J. N. Briggs, J. J. Cebula, A. D. Dayton, and L. C. Hulbert. 1970. Relationships between visual obstruction measurements and weight of grassland vegetation. J. Range Manage. 23:295-297. Sands, J. L. 1978. Game bird studies. New Mexico Dep. Game and Fish Proj. Performance Rep., Proj. W-l04-R-19. Albuquerque. 5pp. Taylor, M. A., and F. S. Guthery. 1980. Fall-winter movements, ranges and habitat use of lesser prairie chickens. J. Wi1d1. Manage. 44:521-524. Warner, R. E., and S. L. Etter. 1983. Reproduction and survival of radiomarked hen ring-necked pheasants in Illinois. J. Wi1d1. Manage. 47:369-375. Prepared by _1iwd..=..==;.......!...rtJ . ~~=' ~ Kenneth M. Giesen Wildlife Researcher _ ��Colorado Division of Wildlife Wildlife Research Report April 1989 51 JOB FINAL REPORT State of: Colorado Project: W-152-R Work Plan: .... _9.... Job .... _8 .... Job Title: Food Selection and Nutritional Ecology of Blue Grouse During Winter Period Covered: Author: Upland Bird Research 01 July 1983 through 31 December 1988 T. E. Remington Personnel: C. E. Braun, S. F. Brinkman, R. W. Hoffman, T. E. Remington, M. L. Stevens, Colorado Division of Wildlife ABSTRACT Food use, food selection, and energy and nutrient use within foods by blue grouse (Dendragapus obscurus) were investigated. Douglas-fir needles (Pseudotsuga menziesii) were most used, most preferred, and the most nutritious needles. Lodgepole pine (Pinus contorta) and limber pine (f. flexilis) needles were also fed-upon by blue grouse. Blue grouse avoided monoterpene-treated chow but monoterpenes were not related to preferences among needle groups. The youngest needles from old (> 100 years) trees were most preferred and contained the most metabolizable energy and nitrogen. Energy and nutrients obtainable from different needle groups were a function of the initial nutrient or energy content, and more importantly the energy or nutrient costs of detoxifying the toxins within them. Monoterpenes and phenolic acids were important toxins but the primary toxins were apparently not identified. Six manuscripts were prepared for publication. Recommendations are presented for maintenance of blue grouse winter habitats. �52 RECOMMENDATIONS Blue grouse in Colorado are dependent on coniferous forests for food in winter. Consequently, logging within coniferous forests may impact blue grouse negatively. The results of this research suggest ways in which impacts of logging on blue grouse can be prevented or minimized. Recommendations are: 1. Cutting within high-elevation stands of Douglas-fir should be minimized. Needles of Douglas-fir are the most preferred, arguably the most used, and yield the greatest metabolizable energy and nitrogen of available conifers. 2. If high elevation (> 2500 m elevation) conifer stands must be cut, then a portion of the Douglas-fir should be left uncut. Douglas-fir most frequently used by blue grouse exceed 150 years of age while Douglas-fir less than 100 years old are seldom fed-upon. Therefore old, decadent trees should be left which will have the greatest value to blue grouse at the lowest sacrifice of merchantable timber. 3. Lodgepole pine needles are also valuable food resources for blue grouse, and lodgepole stands are frequently logged. Fortunately, most lodgepole stands within Colorado are large and blue grouse use only small portions of them. If cuts are to be large (> 50 hectares), impacts to blue grouse can be minimized by identifying (by searching for dropping accumulations), and leaving uncut, sections which contain evidence of winter occupancy by blue grouse. Scattered subalpine fir (Abies lasiocarpa) and/or Engelmann spruce (Picea en~elmanii) should also be left since blue grouse prefer to roost in these species (rather than lodgepole) and they loose less heat in them. 4. Attempts to characterize the quality of conifer needles as food for blue grouse by standard nutritional analyses are likely to lead to erroneous conclusions. Any nutritional assay must be validated by a bioassay, preferably with wild-captured birds. Palatability is an important parameter to assess in winter foods of blue grouse, and it can be evaluated only by observation or bioassay. �53 FOOD SELECTION AND NUTRITIONAL ECOLOGY OF BLUE GROUSE DURING VINTER Thomas E. Remington The results of this study have been prepared in a series of 6 papers. These manuscripts will be submitted for internal review and subsequent publication in technical journals. Publication will be supported through Work Plan 22, Job 1. The 6 manuscripts are: 1. Why don't blue grouse eat aspen in winter?; 2. Why do grouse have ceca: 3. Food habits and food selection of blue grouse during winter; 4. Temporal patterns 5. Metabolizable grouse; and 6. Costs of detoxification Submitted by ~ a test of the fiber digestion theory; of feeding by blue grouse during winter; energy and nitrogen content of conifer needles of secondary compounds t-- t. ~ Thomas E. RemingtOn Wildlife Researcher A to blue to blue grouse. ��Colorado Division of Wildlife Wildlife Research Report April 1989 55 JOB PROGRESS REPORT State of: Colorado Project: W-152-R Upland Bird Research Work Plan: 12 Job Title: Chronology of Breeding and Nesting Activities Relation to Timing of Spring Hunting Seasons Period Covered: Author: Job 01 January 15 through 31 December of Wild Turkeys in 1988 Richard W. Hoffman Personnel: C. E. Braun, R. W. Hoffman, R. L. Holder, and R. K. Mueckler, Colorado Division of Wildlife; E. N. Hall, Colorado State University ABSTRACT Gobbling and nesting activities of 52 radio-marked (10 subadult males, 3 subadult females, 39 adult females) Merriam's wild turkeys (Meleagris gallopavo merriami) were monitored in relation to timing of the spring hunting season in southeastern Colorado. Females moved 11.8 ± 5.7 km from wintering to breeding areas compared to movements of 8.7 ± 3.1 km for subadult males. Subadult males gobbled 62 times during 85 hours of morning (50 hrs) and evening (35 hrs) listening periods from one half hour before to one half hour after sunrise and sunset. Subadult males did not gobble during most morning (74%) and evening (97%) indices, but the tendency was to gobble more in the morning (1.2 gobblesjbird/hr) than evening (0.03 gobblesfbird/hr) and more on (0.93 gobblesjbird/hr) than off (0.56 gobblesjbird/hr) the roost. Subadult males gobbled most (1.7 gobblesjbird/hr) when only accompanied by hens. Eleven (31%) of 36 hens surviving into the nesting season were successful (hatched at least 1 egg). Seven hens initiated incubation prior to the end of the spring hunting season, 5 of which started the last week of the season. Twelve hens started incubation between 16 and 31 May and 4 hens started between 1 and 8 June. The peak period of incubation was 11-25 May; 16 hens (70%) initiated incubation during this period. The harvest survey indicated 3,819 spring hunters harvested 671 turkeys (18% success) and 1,225 fall hunters harvested 355 turkeys (29% success). Las Animas County was the leading harvest area. Public land supported 54% of the hunting pressure, but produced only 42 (spring) and 34% (fall) of the harvest. Hunter success was better on private (spring = 29%, fall = 56%) than public (spring = 14%, fall = 21%) land. The wing collection program sampled 74% of the spring harvest and 70% of the fall harvest. Merriam's dominated (> 90%) the harvest samples. Sixty six percent of the spring harvest of Merriam's were adults compared to 50% for Rio Grandes. The fall harvest (Merriam's only) was comprised of 39% juveniles. The poult to hen ratio was 1:1. ��57 CHRONOLOGY OF BREEDING AND NESTING ACTIVITIES OF WILD TURKEYS IN RELATION TO TIMING OF SPRING HUNTING SEASONS Richard W. Hoffman P. N. OBJECTIVES 1. Document t~m~ng of winter flock dispersal, onset of gobbling, peaks of gobbling, nest initiation, onset of incubation, and peak of hatch in relation to timing of the spring hunting season. 2. Describe the gonadal cycle of females and compare the reproductive condition of females in relation to timing of the spring season. 3. Measure the abandonment rate of incubating human disturbance around the nest. 4. Monitor hunter activity and harvest females to varying levels of of wild turkeys on a statewide basis. SEGMENT OBJECTIVES 1. Review literature pertinent to the objectives 2. Trap and instrument 40 wild turkeys (6 males, 34 females) with tailor poncho-mounted radio transmitters in December-February. 3. Document timing of winter flock break-up. 4. Document onset of gobbling and peaks of gobbling activity. 5. Document onset of egg laying, incubation, 6. Measure 7. Conduct a hunter questionnaire- wing collection using a permit system and mail wing survey. 8. Compile data, analyze results, and prepare progress effects of human disturbance of this study. and peak of hatch. on rate of nest abandonment. program in Fall 1988 report. DESCRIPTION OF STUDY AREA Trapping operations were confined to Longs Canyon and 2 tributary canyons, Sowbelly and Martinez, 17 km southwest of Trinidad, Colorado, in Las Animas County. Birds dispersed into an area encompassing over 448 km2. This area was bounded by 1-25 on the east, Lorencito Canyon on the west. Colorado Highway 12 on the north, and the Canadian River in New Mexico on the south. Topography was mountainous, varying in elevation from 1800 to 2600 m. Four large canyons in excess of 30 km in length occurred within the area, each with numerous side canyons and adjacent smaller canyons. Major vegetation types included pinyon pine-juniper (Pinus edulis-Juniperus spp.), mountain shrub, and ponderosa pine (f. ponderosa). The mountain shrub type was dominated by �58 Gambel's oak (Quercus gambelii), which extended into both the pinyon-juniper and ponderosa pine types. Douglas-fir (Pseudotsuga menziesii) and occasional white fir (Abies concolor) occurred on north slopes within the ponderosa pine type. Most of the area was under private ownership except for 1100 ha of land around Trinidad Reservoir administered by the Colorado Division of Parks and Outdoor Recreation. Use of private lands was limited to cattle grazing and some logging. METHODS Trapping, Marking and Radio-tracking Wild turkeys were baited with oat hay and livetrapped in February and March using drop nets or cannon nets. Captured birds were classified to age (Hoffman 1962) and sex (Hoffman 1962), and banded with serially-numbered aluminum leg bands and patagial wing tags. Ages were recorded as subadult (810 months) and adult (> 18 months). Body weight, and length of primaries, carpal, spur, and beard were measured on each bird. Fifty-two birds were equipped with lithium battery powered transmitters attached with a poncho collar (Amstrup 1980) or tail-clip (Bray and Corner 1972). Tracking was primarily conducted from the ground using a 3-element Yagi antenna and Telonics TR-2 receiver. Aerial searches were made on 6 May and 1 June. All locations were verified by visual observation and recorded to nearest 50 m as Universal Transverse Mercator (DTM) coordinates. Gobbling Indices Flocks containing instrumented birds were monitored a m~n~mum of 3 times per week beginning on 29 February to determine onset of gobbling and period of flock dispersal. Gobbling indices were conducted from 27 March to 10 June and categorized as preseason (27 Mar-1S Apr), hunting season (16 Apr-1S May), and postseason (16 May-10 Jun). An attempt was made to conduct 3 valid indices per week per time period (i.e., AM and PM). A gobbling index was considered valid if (1) positive identification was made of the instrumented bird(s) that was gobbling, (2) the bird(s) was not disturbed prior to or during the gobbling index, (3) the time the bird left (AM gobbling index) or went (PM gobbling index) to roost was known, and (4) the index included time on and off the roost. While it was not considered necessary to know the exact number of other birds present during the index, it was necessary to know whether the male being monitored was alone or associated with other males, females, or both sexes. A gobbling index lasted 1 hour from one half hour before to one half hour after sunrise (AM index) or sunset (PM index). The 1 hour period was divided into 6 10-minute listening periods so gobbling intensity could be quantified within the i-hour interval. Gobbling was also recorded in relation to presence or absence of other birds and whether the bird was on or off the roost. Instrumented males were monitored on a rotating basis, the initial order being randomly selected. For AM indices, the gobbler was located on the roost either the evening before or 1 hour before sunrise on the monitoring morning. Gobblers selected for a PM index were located at least one hour before sunset. �59 Nesting Radio-marked hens were located at least once every 3 days from time of capture until flock dispersal. Subsequent locations varied depending on how far the hen moved from wintering to breeding area. Birds moving longer distances were relocated less frequently because it took considerable searching time to find them initially. Priority was given to locating nests and documenting hatching dates. Timing of other nesting activities (i.e., nest initiation, onset of incubation) was approximated by knowing the date of hatch. Clutch size, fertility, and nesting success were determined from egg shell characteristics after the eggs hatched or after the nest was abandoned or depredated. Hens were not approached closer than 5 m while on the nest nor were they deliberately flushed off the nest to ascertain clutch size. Permits, Questionnaires, Hunters were required and fall seasons: and Wing Collections to obtain I of 2 types of permits during the 1988 spring 1. Special Unlimited Hunting Permit - Unlimited in number and free of charge, these permits were available to any holder of a valid turkey license. Whereas the license could be purchased at any license agent, the permits were only available from CDOW offices either on a walk-in basis or by mail application. The special unlimited permits were valid for one season (spring or fall). Unsuccessful spring hunters could hunt in the fall without purchasing a new license, but before doing so, they were required to obtain another permit valid for the fall season. Successful spring hunters who wanted to hunt the fall season needed to purchased another license in addition to obtaining a special unlimited fall permit. By requiring spring and fall permits, hunters could be categorized as spring only, fall only, or spring-fall hunters and surveyed accordingly. Special unlimited permits for the spring season were available from 1 March to 15 May 1988 (end of season); those for the fall, from 1 August to 2 October 1988. Special unlimited permits were valid for all areas not requiring a limited permit. 2. Limited Hunting Permits - Limited in number, free of charge, and available by public drawing. Only mail applications were accepted for limited permits. The drawing and issuance of permits was handled through the Denver office. Hunters could apply for a limited permit without first purchasing a license. However, if they succeeded in drawing a limited permit, they needed to purchase a license before going hunting. Drawings were held on 25 March and 26 August 1988 for the spring and fall seasons, respectively. Limited permits were only valid for the area, season, and time period indicated on the permit. Holders of a limited permit were required to obtain a special unlimited permit if they hunted outside the area for which their limited permit was valid. Consequently, some hunters obtained 2 permits and were subsequently mailed 2 harvest questionnaires. This problem was identified as the questionnaires were being processed and the duplicates were excluded. Questionnaires were mailed to all permit holders immediately after the spring and fall seasons. Non-respondents were mailed a followup questionnaire approximately 3 weeks later. Mean values calculated from responses to the �60 second mailing were used to project responding to either questionnaire. answers for those permit holders not Every hunter that obtained a limited or unlimited permit was also issued a wing envelope coded by permit number. The instructions on the envelope requested each successful hunter to (1) complete the questionnaire printed on the envelope including their name, address, time of harvest, and location of harvest (county, small game management unit, and nearest town), and (2) to remove the least damaged wing and 3 or 4 breast feathers as depicted by 2 schematic diagrams, place them in the envelope, and mail the postage-paid envelope. The envelopes were addressed to the Wildlife Research Center in Fort Collins. Upon delivery, the wings were frozen until they could be processed. Information provided by hunters was transcribed onto a standardized form. The envelope's contents were examined to determine if both a wing and breast feathers were enclosed. The samples were then identified to subspecies and classified to age and sex. Whenever possible, the following measurements and feather characteristics were recorded for each wing: length, condition (growing, fully grown, empty, broken, worn, pointed) and status (adult or juvenal) of primaries I through X (numbered proximal to distal), quill diameter of P IX and P X at their insertion into the follicle, stage of primary molt, carpal length, and status (adult or juvenal) of the first secondary and tertials. Since the range of the Merriam's and Rio Grande wild turkey does not overlap in Colorado, except possibly along the Arkansas River west of Pueblo, the 2 subspecies were identified from wing samples based on the location of harvest. However, inspection of the wings suggested that subspecies could be distinguished by wing color. Rio Grande's tended to have blacker primaries and fewer white bars than Merriam's, thus, giving the wing a darker appearance. Sex was ascertained by the presence of buffy (female) of black (male) tipping on the breast feathers (Hoffman 1962). Wings were assigned to age classes based on the shape, color, wear, and barring pattern of primaries IX and X (Hoffman 1962). Juvenal primaries IX and X were pointed, grayish-brown, and lacked white barring near the tip. Comparatively, adult primaries IX and X were rounded and black or blackish-brown, with white barring extending almost to the tip. Subadults possessed the outer primaries characteristic of juveniles except they were faded and worn as a result of being retained longer. Wings obtained from the spring season were classified as adults (> 22 months) or subadults (10-11 months). Those from the fall were separated into 3 age classes: adults (> 26 months), subadults (14-16 months), and juveniles « 6 months). RESULTS AND DISCUSSION Capture and Marking Eighty-six turkeys (32 males, 54 females) were trapped in late February-early March with drop nets (27) or cannon nets (59). Three subadult hens, 39 adult hens, and 10 subadult males were equipped with leg bands, wing tags, and radios, and released at the trap site (Table 1). Two birds died during �61 Table 1. Wild turkeys trapped and equipped with radios in Longs Canyon, Colorado, February-March 1988. Band Beard Spur Sex Age Weight (kg) Carpal # (nm) (nm) (mn) 139 144 181 182 183 184 185 186 189 190 M M M M M M subad Subad Subad Subad Subad Subad Subad Subad Subad Subad 6.2 4.9 5.3 5.8 5.1 6.5 5.2 5.5 5.8 5.2 501 487 480 475 480 486 484 504 492 495 98 49 38 60 48 73 63 66 72 73 Subad Subad Subad 3.4 3.9 3.5 415 425 415 Ad Ad Ad Ad Ad Ad Ad Ad Ad Ad Ad Ad Ad Ad Ad Ad Ad Ad Ad Ad Ad Ad Ad Ad Ad Ad Ad Ad Ad Ad Ad Ad Ad Ad Ad Ad Ad Ad Ad 4.6 4.7 4.2 4.4 4.0 4.1 3.5 3.7 4.6 3.6 4.4 4.0 4.0 4.2 4.6 4.2 4.2 4.7 4.6 4.5 4.5 4.2 4.5 4.8 4.0 3.8 4.8 4.1 4.4 4.7 4.5 4.8 4.7 5.0 4.6 4.5 4.1 4.2 4.4 440 434 446 440 410 435 431 435 460 438 437 415 424 441 437 434 434 441 435 444 446 417 438 435 442 435 453 450 442 438 446 440 442 435 440 445 443 442 437 153 179 180 27c 61c 128 129 132 133 135 136 140 141 142 151 152 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 1n 178 187 188 M M M M F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F aT = tail mount, P = poncho mount. bRadio recovered from bird 190 (mortality) and put on bird 169. cRecaptures. dRadio recovered from bird 175 (mortality) and put on bird 183. 3 3 5 2 2 3 2 3 3 4 Radio frequencya 151.159P 150.278T 150.555T 150.815T 151.596P 150.939T 150.513T 151.057T 150.762T 151.458pb 150.238T 151.662P 150.017T 175 60 159 173 170 160 150.156T 150.480T 151.424P 150.100T 150.064T 150.660T 151.678P 151.722P 150.906P 151.258P 150.079T 150.415T 151.418P 151.068P 151.513P 150.034T 151.436P 151.393P 150.869T 150.356T 151.656P 151.740P 150.400T 150.213T 150.140T 151.233P 150.305T 151.458P 150.576T 151.477P 151.574P 151.638P 151.542P 151.596pd 150.700T 151.494P 150.739T 151.617P 150.536T �62 trapping. The remalnlng 32 birds were leg banded and released at the trap site. Efforts to trap adult males were unsuccessful. No adult males were observed at the bait sites and no flocks of adult males consistently remained in one area where they could be attracted to bait. Seven (13%) of the 54 hens captured had beards. All bearded hens were adults. The beard length was 149 ± 44 mm (& ± SD). The shortest beard was 60 mm, but it was still clearly visible through the breast feathers. Beard length for subadult males was 64 ± 17 mm. Average beard length for adult males captured in 1986 was 216 ± 17 mm and for adult females 134 ± 52 mm (Hoffman 1987). Subadult males weighed 5.6 ± 0.5 kg compared to 4.3 ± 0.3 kg for adult females and 3.6 ± 0.3 kg for subadu1t females. Weights recorded in 1986 were 7.6 ± 0.6 kg for adult males, 3.9 ± 0.2 kg for adult females, and 3.0 ± 0.5 kg for subadu1t females (Hoffman 1987). Onset of Gobbling and Flock Dispersal Gobbling of adult males was first heard on 18 March and continued through 10 June when gobbling indices were terminated. Subadu1t males were first heard gobbling on 15 April and last heard on 3 June. Subadult males continually associated with hen flocks through the first week of April, whereas, flocks of adult males remained intact through March and did not start associating with hens until the first week of April. During the preseason (27 Mar-15 Apr) monitoring period, radio-marked subadult males were observed in mixed flocks (adult males and females) on 15 of 23 (65%) occasions, but were seen alone or with other jakes more often (21 of 42 observations, 50%) during the season (16 Apr-IS May) and postseason (16 May-l0 Jun) periods (Table 2). They were not seen with adult males only because adult males were usually observed with hens. However, even in mixed flocks, when adult males separated from hens or roosted separately, subadult males stayed with the hens. Table 2. Observations of subadult presence or absence of other birds. Flock composition Females and adult males Females only Adult males only Jakes only Alone (jake) male wild turkeys Preseason % II 15 1 0 3 4 65 5 0 13 17 in relation Number of observationsa Season % II 4 6 0 5 8 17 26 0 22 35 to Postseason % n 4 7 0 8 0 aConsecutive observations, where a male was indexed in the PM and again in the AM of the following day, were counted only once. 21 37 0 42 0 �63 Hens remained in winter groups into the second week of April. Smaller groups of hens were observed through late April into early May. Occasional lone hens or hens unaccompanied by adult males were encountered in late April. Such observations became more common in early to mid-May. Gobbling Indices Eighty-five (50 morning, 35 evening) valid gobbling indices were obtained on 9 subadult, radio-marked males between 27 March and 10 June. Subadult males gobbled a total of 61 times during 13 of 50 morning indices (1.2 gobbles/birdjhr) and only once (0.03 gobbles/birdjhr) during 35 evening indices (Table 3). They did not gobble during most (morning ~ 74%, evening 97%) indices. The number of gobbles per indice (n - 13) when they did gobble ranged from 1 to 28 (median - 2). They gobbled more during and after the hunting season than before the season, more on than off the roost, and more when only accompanied by hens than when alone or with other subadult males or with hens accompanied by adult males (Table 3). There was no indication that some subadult males gobbled more than others. Table 3. Gobbling activity of subadult male wild turkeys in relation to timing of the spring hunting season, time of day, roosting activity, and presence or absence of other birds. Hours of observationa Total gobbles Timing of Spring Season Preseason (27 Mar-15 Apr) Season (16 Apr-15 May) Postseason (16 May-10 Jun) 38 35 22 4 36 22 Time of Day AM PM 50 35 61 1 Roosting On Off 38.5 46.5 36 26 19 34 23 9 33 18 11 Category Activity Sex Compositionb Females Females and adult males Subadult males Alone a aListening intervals occurred from one-half hour before to one-half hour after sunrise and sunset. bSex of other birds associated with the radio-marked subadult males. Subadult males were not observed with only adult males. �64 In 1986, 94 valid gobbling indices were obtained on 7 radio-marked adult males (Hoffman 1987). Gobbling was heard during 80 (85%) of the 94 indices and averaged 25 (preseason), 35 (season), and 12 (postseason) gobblesfbird/hour. Adult males also gobbled more on (64 gobbles/hr) than off (12 gobbles/hr) the roost and more in the morning (31 gobblesfhr) than evening (7 gobbles/hr). However, unlike subadults, adult males gobbled most when alone (25 gobbles/hr) rather than when they were with females only « 2 gobblesfhr). Roosting Activity Males tended to leave the roost later as the observation intervals progressed from preseason (& - 13 min before sunrise), to season (& - 1 min after sunrise), to postseason (R - 6 min after sunrise) periods. There were only 4 of 26 morning indices prior to 28 April when males left the roost at or after sunrise. Starting on 2 May, males came off the roost after sunrise during 19 of 22 indices. Males went to roost after sunset during the preseason (R ~ 6 min) and before sunset during the season (R - 3 min) and postseason C& - 13 min) periods. Mortality Thirteen radios were recovered between 1 March and 15 July; 11 from hens and 2 from males. Eleven recoveries (10 females, 1 male) were classified as from birds that died of natural causes and 2 (1 female, 1 male) were from birds that lost their transmitters. Capture myopathy was not suspected as a cause of any mortalities in 1988. Five mortalities (including the 1 male) were documented between 7 and 27 March, 2 between 1 and 17 May, and 4 between 1 and 4 June. Only one hen was killed on the nest. Total loss of hens from late winter (1 Mar) through the nesting period (15 Jul) was 24% (10/42). Only 1 male (10%) died during this period. Both transmitters thought to have been lost were tail mounts. The nuts holding the transmitters to the tail came loose on one of the tail mounts. The other tail mount was found still attached to the 2 central tail feathers. No other feathers were found near the transmitters, nor had the transmitters sustained any damage (i.e., tooth marks, scratches, etc.). Nesting Eleven (31%) of 36 hens surviving into the nesting season nested successfully; whereas, 12 attempted to nest but failed, 2 were suspected of laying when predated, 1 was predated on the nest, and 10 either did not attempt to nest or lost or abandoned their clutch prior to onset of incubation. One renest attempt was documented. This hen lost her first clutch between 23 and 31 May and was relocated on a second clutch on 24 June. The second clutch was predated on -8 July. Clutch size of first nest attempts was ascertained from egg shell fragments for 8 nests (& = 9.2, range - 8-11); 70 (95%) of the 74 eggs hatched successfully. The earliest known date for onset of incubation for a first clutch was 5 May and the latest was 8 June. Seven hens initiated incubation prior to the conclusion (15 May) of the spring hunting season, 12 started between 16 and 31 May, and 4 started between 1 and 8 June. Five of the 7 hens starting before the season ended initiated incubation the last week of the season. The peak period for onset of incubation was 11-25 May; 16 hens (70%) initiated incubation during this period. �6S Movements Nine subadult males moved 8.7 ± 3.1 km from wintering to breeding areas. Adult males in 1986 moved 3.7 ± 1.7 km (Hoffman 1987). Females moved longer distances (1988 - 11.8 ± 5.7, 1986 - 9.6 ± 4.3) than adult or subadult males. Hunter Compliance - Permit System A total of 4,569 permits was issued for the spring season and 5,078 licenses were sold. Hunter compliance with the spring permit requirement was 90%. Questionnaires were sent to all permit holders by 1 June 1988. A followup questionnaire was mailed to the non-respondents 3 weeks later. The combined return rate for the first and second mailing was 72% (Table 4). There were more permits issued (1,599) for the fall season than there were licenses sold (889). The reason was that unsuccessful spring hunters who wanted to hunt the fall season were not required to purchase another license, but were required to obtain a fall permit. Consequently, total license sales and total permits issued were not directly comparable in evaluating hunter compliance. Instead, hunter compliance was measured as the proportion of new license buyers who also obtained a permit. The estimate was 88% (782 of 889). Questionnaires were sent to the 1,599 fall permit holders of which 58% were returned. A followup questionnaire was sent to the 669 non-respondents and 233 additional responses were received. Seventy-three percent of the fall hunters responded to the questionnaire (Table 4). Table 4. Response Surveys mailed Surveys returned Percent return Non-deliverable aTotal hunters Hunter Activity to the 1988 spring and fall turkey harvest survey. 1st SRring 2nd Totals 1st Fall 2nd 4,569 2,527 55 90 2,042 740 36 35 4,569a 3,267 72 125 1,599 930 58 26 669 233 35 14 Totals 1,599a 1,163 73 40 sampled. and Harvest - Questionnaire Projected estimates for the spring season indicated that 3,819 hunters harvested 671 turkeys for a success rate of 18% (Table 5). Total harvest including a reported crippling loss of 13%, was estimated at 771 birds. Spring hunters averaged 3.4 days afield over the course of the 30-day season from 16 April to 15 May. There were 68% fewer hunters that participated in the fall season (1,225 hunters) and 47% fewer birds (355) were harvested (Table 6). However, hunter success (29%) was higher. Crippling loss (14%) was similar. Hunters in fall spent 2.8 days afield, but had only 16 days (17 Sep-2 Oct) of hunting which included just 3 weekends. The spring season �66 lasted 14 days longer, opportunity. Table 5. allowing for 2 additional weekends of hunting 1988 spring turkey harvest and hunter activity. Descriptive statistic 1st Mai lins 2nd 1: Projected for Totals in sample hunters hunters hunters observing turkeys hunters observing turkeysb successful huntersb(harvest) successful hunters N hunter da~s Days/hunter Crippl ing loss % crippling loss Total harvest 2,527 2,349 93 1,658 71 451 19 7,715 3.3 58 11 509 740 533 72 373 70 79 15 1,948 3.7 15 16 94 3,267 2,882 88 2,031 70 530 18 9,663 3.4 73 12 603 1,302a 937 72 656 70 141 15 3,186 3.4 27 16 168 4,569 3,819 84 2,687 70 671 18 12,849 3.4 100 13 771 1: Projected for Totals 1i 1i % 1i % 1i % aNon-respondents. bBased only on those license holders who actually hunted. Table 6. 1988 fall turkey harvest and hunter activity. Mail ins Descriptive statistic 1st in sample hunters hunters hunters observing turkeys hunters observing turkeysb N successful hunters (harvest) b % successful hunters N hunters days Days/hunterb Crippl ing loss % crippling loss Total harvest 930 709 76 389 55 246 35 1,985 2.8 34 12 280 1i 1i % 1i % 2nd 233 180 77 100 56 38 21 486 2.7 9 19 47 1,163 889 76 489 55 284 32 2,471 2.8 43 13 327 436a 336 77 188 56 71 21 907 2.7 17 19 88 1,599 1,225 77 677 55 355 29 3,378 2.8 60 14 415 aNon-respondents. bBased only on those license holders who actually hunted. Most hunting pressure and harvest occurred on opening weekend during both seasons (Table 7). Pressure and harvest did not change substantially over the remainder of the fall season. There was an increase in pressure on weekends during the spring season. The Southeast Region accounted for 76% of the spring and fall harvest (Tables 8, 9). Las Animas County was the leading harvest area with 21 and 29% of the spring and fall harvest, respectively. Public land supported 54% of the spring and fall hunting pressure, but produced only 42% (spring) and 34% (fall) of the harvest (Table 10). �67 Table 7. Chronological distribution of hunting pressure the 1988 spring and fall wild turkey seasons.a Date 1st 1st 2nd 2nd 3rd 3rd 4th 4th 5th weekend week weekend week weekend week weekend week weekend Totals SQring; Hunter da::x:s Harvest % % N N and harvest Fall Hunter da::x:s % N during Harvest % N 1,880 979 1,094 772 1,388 812 1,197 572 969 20 10 11 8 14 8 13 6 10 178 60 70 37 48 30 28 27 26 35 12 14 7 10 6 6 5 5 927 400 467 322 242 39 17 20 14 10 109 42 47 36 49 38 15 17 13 17 9,663 100 504 100 2,358 100 283 100 aBased on hunter days and harvest periods. that could be assigned to specific time Hunter success averaged 26% on the spring limited permit areas, exceeding the statewide success rate of 18% on all but 2 areas (Table 11). The eastern plains limited permit areas continued to produce high hunter success rates. Hunter success (27%) on the fall limited permit areas approximated the statewide average of 29%. However, fall hunter success on both eastern plains permit areas surpassed the statewide average. It was a definite advantage to hunt on private (spring = 29% success, fall = 56% success) vs. public land (spring ~ 14% success, fall - 21% success) in both seasons. Seventy-three percent of the birds harvested during the spring season were taken before noon; 54% were harvested between 0500 and 0900 hours (Table 12). Only 19% were not associated with other birds at the time of harvest (Table 13). Most (47%) were taken f~om flocks comprised of 1-5 birds. The remaining 34% were taken from flocks with 6 or more birds. There was no indication that hunters encountered smaller flocks as the season progressed. Hens undoubtedly were present in the larger flocks, suggesting they were still receptive to males and not incubating. Wing Collection Program Limitations - The validity of any population index calculated from harvest samples is dependent upon the assumption that different age and sex classes are harvested in proportion to their occurrence in the population. Long-term population and harvest data are often necessary to test this assumption. Such data are not available for the Merriam's wild turkey in Colorado or elsewhere throughout its range. In addition, if wings are collected over a broad geographic area, they may not accurately reflect characteristics of local populations. These limitations do not preclude the use of wing data as a tool in formulating management strategies. Biologists must be aware of them and use caution in interpreting the data. �68 Table 8. ~ild turkey harvest by County, Game Management Unit, and Region, Spring 1988. Harvest Harvest County !!. Las Animas Harvest X GMU !!. X 110 21 Fremont 51 10 Pueblo 50 10 84 85 140 143 711 421 96 59 73 54 43 31 27 26 23 23 14 11 8 6 5 103 ] 109 107 69 581 851 86 71 58 511 42 20 123 144 136 78 135 130 83 771 38 118 147 40 77 19 51 501 56 41 142 39 145 128 46 95 99 461 57 751 146 137 19a 4 18 13 12 12 11 10 9 8 8 8 6 5 5 5 4 4 4 4 4 4 4 3 3 3 3 3 3 3 3 2 2 2 1 1 1 1 1 1 1 3 3 2 2 2 2 2 2 2 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 < 1 < 1 < 1 < 1 <; 1 <; 1 < 1 < 1 < 1 < 1 514 7 100 Huerfano 41 8 Custer 35 7 31 6 Washington Morgan 23a 4 Dolores 22 4 19a 4 Kit Carson El Paso 18 3 Baca MontezUlla 17 17 3 3 Archuleta Teller 12 11 2 2 Larimer Garfield 10 2 2 Jefferson 6 Otero 6 Chaffee 4 Bent 4 Costilla Parle: 4 3 Clear Creek Delta Douglas Gi lpin Adams Weld Boulder La Plata 2 2 2 2 1 1 1 1 Totals Unknowns 514 7 Mesa lo,oo ] Yuma t incctn ] 8 < < <; < < < <; < 1 1 1 1 1 1 1 1 100 Region !!. SE 369 72 S~ 58 11 NW 39 8 NE 34 6 Central 14 3 514 7 100 % 5 4 4 aHarvest was pooled by area which included more than one county or GMU. �69 Table 9. Wild turkey harvest by County, Game Management Unit, and Region, Fall 1988. County Las Animas Fremont Pueblo Custer Huerfano Yuna ] Kit Carson Mesa El Paso Logan ] lIashington Morgan Larimer Archuleta I !! X GMU !! X 81 39 29 22 21 12a 29 14 10 8 7 4 47 41 27 22 19 19 12a 17 15 10 8 7 7 4 11 8 4 3 84 85 140 69 59 143 103 ] 109 86 96 421 851 130 20 19 136 58 n1 581 9 3 3 3 3 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 8a 3 8 7 3 2 ~ Baca 6 2 P Otero 5 2 Bent 4 Douglas 3 Teller 3 Chaffee 3 Costilla 2 Garfield 2 Jefferson Gi lpin LaPlata Clear Creek Totals Unknown rr < < < < 278 9 51 56 511 40 144 135 147 42 461 741 83 39 146 38 41 123 1 1 1 1 100 8 8 8 6 4 4 4 4 4 3 3 3 3 2 2 2 2 2 2 1 1 1 1 1 1 1 1 < < < < < < < < Region !! SE NE Nil 233 16 13 10 6 84 6 5 3 2 278 9 100 511 Central 100 278 9 % aHarvest was pooled by area which included more than one county or GMU. Table 10. Distribution of hunting pressure and harvest by land status during the 1988 spring and fall wild turkey seasons. Fall 512ring Hunters Land status Public Private Both Totals !! Harvest X Hunters Harvest !! x !! x !i % 1,521 1,026 288 54 36 10 219 302 42 58 438 327 50 54 40 6 94 183 34 66 2,835 100 521 100 815 100 zn 100 �70 Table 11. Hunter success on limited permit areas during the 1988 spring and fall wild turkey seasons. Spring Permit areas Permits issued Lake Dorothey Beaver-Skagway Units 103 and 109 Units 107, 108, ll2, ll3, ll4, ll5, 120, 121 Unit 96 Colorado Spgs. SWA Spanish Peaks Unit 42 East Harvest 75 30 40 20 15 6 70 22 Fall Success (%) 13 Permits issued 17 10 37 30 65 30 20 31 20 20 1 10 6 60 250 66 26 3 5 b Success (%) Harvest 8 8 12 12 27 60 7 35 23 28 a a 81 t a 1 Q Totals aC10sed to fall hunting. bNo restrictions on numbers of hunters 216 58 27 during the spring season. Table 12. Distribution of harvest by time period during the 1988 spring and fall wild turkey seasons. Fall Spring Time period 0500 0600 0700 0800 0900 1000 1100 1200 1300 1400 1500 1600 1700 1800 1900 - 0559 0659 0759 0859 0959 1059 1159 1259 1359 1459 1559 1659 1759 1859 1959 aMountain Daylight N 7 88 107 64 46 33 18 10 18 9 10 21 20 29 15 Time. % N % 1 18 22 13 9 6 4 2 4 2 2 4 4 6 3 16 34 22 32 24 12 10 7 6 10 25 27 22 6 14 9 13 10 5 4 3 2 4 10 11 9 �71 Table 13. Size of flocks from which turkeys were harvested spring season. during the 1988 Flock size oa , 2-5 Date of harvest N % 16-20 Apr 21-25 Apr 25-30 Apr 1-5 May 6-10 May 11-15 May 37 14 18 8 9 8 Totals 94 > 10 6-10 N % 39 15 19 8 9 8 103 45 22 16 21 23 19 230 N % N % 45 20 9 7 9 10 39 14 9 12 9 7 43 16 10 13 10 8 37 18 8 7 3 5 47 23 10 9 4 6 47 90 18 78 16 I ~ aBird was alone. Compliance - Of the 530 successful spring hunters who responded to the questionnaire, 402 (76%) said they returned a wing. However, 499 wings were processed, meaning 97 successful hunters who did not respond to the questionnaire still returned a wing. The wing collection program sampled 74% (499/671) of the spring harvest. A total of 249 wings was collected from the fall wing survey representing 70% of the estimated fall harvest. Seventy-nine percent (213/284) of the successful hunters responding to the fall questionnaire indicated they returned a wing and 36 successful hunters not responding also returned a wing. Subspecies Composition - The Merriam's wild turkey was the dominant subspecies identified from inspection of wings. It comprised 90 and 92% of the spring (Table 14) and fall (Table 15) samples, respectively, and was harvested in 27 of Colorado's 63 counties. Rio Grande's were taken in 8 counties. Fremont, El Paso, and Pueblo were counties where both subspecies were harvested. The 48 Rio Grande wings collected during the spring season originated from the following areas: 19 from Unit 96, 17 from Units 103/109, 5 from Unit 123, 3 from Unit 951, 3 from Unit 118, and 1 from Unit 84. Only 19 Rio Grande wings were identified in the fall sample of 249 wings: 8 from Unit 96, 7 from Units 103/109, 3 from Unit 84, and 1 from Unit 59. Sex Composition - Seven (2%) wings examined from the spring sample were from females (Table 14). All were Merriam's and all were assumed to be from bearded hens. Two percent of the total spring harvest (n - 13) is probably a reasonable estimate of the legal harvest of hens. At this level of harvest, spring hunting should have minimal impact on the female segment of the population. �72 Table 14. Sex, age, and subspecies based on wing analyses. composition Males of the 1988 spring harvest Females Subadu1t Adult Subspecies Subadu1t Adult Merriam's Rio Grande's 147(33)a 24(50) 294(66) 24(50) 0(0) 0(0) 7(1) 0(0) 448b 48 Totals 171 (34) 318(65) 0 7(1) 496 Totals apercent. bThree additional Merriam's wings were processed but could not be classified to age or sex bringing the total wings collected to 499. t I ~ Table 15. Age, sex, and subspecies composition of the 1988 fall harvest based on wing analyses. Subspecies Males x Ii Adults Females % Ii Merriam's Rio Grande's 56 42 2 29 76 5 58 71 Total x Ii 132 57 7 37 Males Ii % 0 0 0 0 Subadults Females Illil Ii x Ii % 9 100 9 1 100 1 4 5 Males x Ii 52 58 5 45 Juveni les Females Total % % Ii Ii 37 42 6 55 89 39 11 58 Sample size Poul ts/ hen 230 19 1.0 1.8 Examination of wings (Merriam's only) originating from the fall season revealed the sex ratio of adults and subadults combined favored females (1.5 females: 1 male), while the sex ratio of juveniles favored males (1.4 males 1 female). The biological implications of these ratios are difficult to interpret without knowledge of the population structure. However, in a promiscuous species such as the wild turkey, sex ratios should favor females. Age Composition - Sixty-six percent of the spring harvest (males only) of Merriam's wild turkeys were adults compared to only 50% for Rio Grande's. Rio Grande populations in Colorado are the result of recent (since 1981) introductions. Available evidence suggest these populations are still increasing and expanding into unoccupied ranges, while populations of Merriam's wild turkeys have remained stable or declined. The age composition of the Rio Grande harvest reflects good production, survival, and recruitment of young birds, attributes indicative of an increasing population. Subadult males in the increasing Rio Grande populations may not be suppressed by mature males from participating in breeding activities. If so, subadults should be equally, if not more vulnerable due to their inexperience, to the calling tactics of hunters. The following discussion regarding the age composition of the fall harvest pertains to the Merriam's wild turkey. Samples of Rio Grande wings from the fall season were inadequate for meaningful interpretation. �73 Subadults comprised 4% of the fall harvest including juveniles and 6% excluding juveniles (Table 15). Biologically, this could be interpreted as poor production in 1987 and subsequent low recruitment into the 1988 population. However, anpther factor contributing to the low number of subadults is associated with the onset of primary molt and whether key feathers (primaries IX and X) for separating adults and subadults are still present at the time wings are collected. If these feathers have already molted, subadults cannot be distinguished from adults. This was the case as 100% of the males and 51% of the females had completed their primary molt by the fall season. The problem of separating adults and subadults was compounded for males because they start and finish their primary molt before females. Misidentification of subadults is not a problem in spring when birds are beginning molt of primary feathers. Percent juveniles in the harvest (39%) and the ratio (1:1) of juveniles to adult and subadult females were used as indices to productivity. The 2 indices suggested poor production of young in 1988. LITERATURE Amstrup, S. C. 44:214-217. 1980. A radio-collar CITED for game birds. J. Wildl. Manage. Bray, O. E., and G. W. Corner. 1972. A tail clip for attaching to birds. J. Wildl. Manage. 36:640-642. Hoffman, D. M. 1962. The wild turkey in eastern Colorado. Game and Fish. Tech. Publ. 2. 49pp. transmitters Colorado Dep. Hoffman, R. W. 1987. Chronology of breeding and nesting activities of wild turkeys in relation to timing of spring hunting seasons. Job Prog. Rep. Colorado Div. Wildl. Fed. Aid Proj. W-152-R. Apr. 1987. Pp. 199-232. Wildlife Researcher ��75 Colorado Division of Wildlife Wildlife Research Report April 1989 INTERIM FINAL REPORT State of: Colorado Project: W-152-R Upland Bird Research Work Plan: 13 Job Title: Seasonal Habitat Use by Plains Sharp-tailed County, Colorado Period Covered: Author: Job _9_ 01 January through 31 December Grouse in Douglas 1988 Anthony W. Hoag Personnel: C. E. Braun, K. Demarest, D. Prenzlow, M. Wertz, Colorado Division of Wildlife; A. W. Hoag, E. Redente, Colorado State University ABSTRACT The status and habitat use by plains sharp-tailed grouse (Tympanuchus phasianellus jamesii) were investigated during 1986-88 in Douglas County, Colorado. This grouse once occupied suitable habitats in northeastern Colorado, but the range of this subspecies in Colorado has been greatly reduced because former habitats have been altered by man. Within Colorado, the estimated population size in 1986-88 was l75~225 birds existing only in Douglas County in an area with 6 known leks and 2 historic lek sites. Habitat use was studied by capturing 23 grouse on 4 leks and fitting them with solar-powered radios. Seven habitat variables and 7 sharp-tailed grouse activities were studied. Habitat use varied by season and sex. Shrub communities were selected in fall, winter, and for escape and nesting. Transition zones between shrubs and grasslands were selected in spring, and for roosting and nesting. Grasslands were selected in summer and for feedingloafing, and mating activities. Home range size varied between 103 and 363 ha for males, and 306 and 884 ha for females. Management of private_l,?-nds with plains sharp-tailed grouse in Colorado should focus on mainta:Lning residual vegetation from the previous year to increase cover for nesting, brooding, feeding-loafing, and roosting, A draft thesis incorporating 2 manuscripts for publication has been submitted to fulfill requirements for the Master of Science degree in Range Science. ��Colorado Division Wildlife Research April 1989 of Wildlife Report 77 JOB PROGRESS REPORT State of: Colorado Project: W-152-R Upland Work Plan: 14 Job Title: Seasonal Period Covered: Author: Job 3__ Movement 01 January Bird Research and Habitat Use by Greater through 31 December Prairie-Chickens 1988 Michael A. Schroeder Personnel: M. A. Schroeder and G. C. White, Colorado State University; C. E. Braun, F. Pusateri, and L. A. Robb, Colorado Division of Wildlife ABSTRACT Investigations were continued to examine seasonal movements and habitat use of greater prairie-chickens (Tympanuchus cupido) in northeastern Colorado. Habitat use was variable within each season, and differences between observed and available habitat sites generally were unclear. Behaviors such as nesting, foraging, and breeding appeared to be more important than either sex or season for interpreting use of habitat. Leks were associated with areas of sparse, short cover and nests were in areas dominated by dense, tall cover. As with habitat use, home range size frequently was unclear. However, females typically had larger home ranges than males and yearling males had larger home ranges than adults. Relatively large movements by females during the early summer season dramatically increased average home range size and its standard deviation. Whether these movements represent dispersal or migratory movements is not clear. The evolution of lek behavior has been explained by the female preference hypothesis which suggests that females select large clusters of males for mating purpos.es. Predictions that females should have spring home range diameters equal to Lnce r -Lek distances and that most females should visit only 1 lek during. the breeding season were tested by monitoring (with radio telemetry) female greater prairie-chickens during the breeding seasons of 1986-88 in northeastern Colorado. In contrast to the predictions, 66 of 90 females (73%) had their nests closer to a lek other than that where first observed or captured. Females nested an average of 3.67 km from the lek where first observed or captured, and an average of 1.02 km from the nearest lek. Likewise, 59 of 79 (75%) visited more than 1 lek during the breeding season. These data clearly illustrate that the female preference hypothesis is not applicable to greater prairie-chickens in northeastern Colorado. In addition, female movement in relation to lek dispersion is an important management consideration. The designation of an arbitrary zone of habitat around leks may not be valid for determining critical habitat for greater prairiechickens. ��79 SEASONAL MOVEMENT AND HABITAT USE BY GREATER PRAIRIE-CHICKENS Michael A. Schroeder P. N. OBJECTIVES 1. Quantify seasonal habitat use, movements, and lek attendance prairie-chickens in northeastern Colorado. of greater SEGMENT OBJECTIVES 1. Relocate all radio-marked greater prairie-chickens at least 20 times during the late summer (1 Jul - 15 Aug) and periodically through the winter. 2. Analyze movement and habitat use of greater prairie-chickens to cover types as mapped on the entire study area. 3. Describe vegetation at all greater prairie-chicken Analyze habitat use for each season and sex. 4. Document reproductive parameters, marked greater prairie-chickens. 5. Analyze the dispersion of home ranges of female greater prairie-chickens during the breeding and nesting seasons to test hypotheses about lek evolution and territoriality. 6. Compile and analyze data, and prepare movements, relocation and survival in relation sites. for all radio- annual report. METHODS An area centered 20 km northwest of Wray, Colorado was chosen for research on greater prairie-chickens. A study area of 301 km2 was monitored in 1986-88 to estimate lek density and determine male lek attendance (Fig. 1). Two or more displaying males were considered a lek. Attendance for each lek (within a year) was estimated by using the maximum count of males at that site. Trapping efforts were concentrated on a core area of approximately 75 km2 (Fig. 1). Birds were trapped at winter feeding sites using walk-in traps baited with corn and at leks using walk-in traps and cannon nets. All captured birds were banded with a numbered aluminum band and a unique combination of 3 colored plastic bands. Age was ascertained by examining patterns of feather wear (Ammann 1944). Bird ages were: yearlings, 5-17 months of age (1 Nov of 1st year to 31 Oct of 2nd year) and adults, older than 17 months of age (after 31 Oct of 2nd year). Battery- and solar-powered radio transmitters were attached to poncho-type markers (Amstrup 1980) and placed on greater prairie-chickens in 1986-88. Radio weights were between 1.8 and 2.3% of each birds' body weight. �80 N MEAN YEARLY LEKATTENDANCE o 0-1 MALES o >103MALES o o o o >3-8MALES >11-10 MALES >10-15 MAl.!8 >15MAl.!8 CAPTURE LOCATIONS * FEEDING AREAS • LEKS R47W! •• -:- R46W ~ - =- ~ ••.•• -:- R45 W ~ - ~ ~ •• .,. -:- , .. ~ _,_ • _:_ •... ~ _!_ .,._:_ ~ - R44 W •••• ~ -!- ~-:- ~..~ -:-1 4. _:_ ••• ,_ _: •• ., _: ••.•. _ •• _ iii R 47 W i i R 46 W R 45 W R 44 W o 2 4 km Fig. 1. Greater prairie-chicken study area in 1986-87 (301 km2 outlined bv solid boundary) and in 1988 (932 km2 outlined by dashed boundary) in north-eastern Colorado. �81 Seasonal collection periods were early spring (15 Feb -31 Mar), late spring (1 Apr - 15 May), early summer (16 May - 30 Jun) , late summer (1 Ju1 - 15 Aug), autumn (16 Aug - 31 Oct), and winter (1 Nov - 14 Feb); designation of seasons was based on aspects of breeding behavior and movement (Robel et al. 1970Q). Time periods were defined as morning (0.75 hours before - 2.75 hours after sunrise), mid-day (2.75 hours after sunrise - 2.75 hours before sunset), and evening (2.75 hours before - 0.75 hours after sunset). Radio-marked prairie-chickens were tracked using a portable receiver and a 3-element yagi antenna. Each bird was visually observed once every 21-28 days and additional observations were obtained by triangulation; azimuths were obtained within 1.0 km of the target transmitters and at angles-of-incidence greater than 35° and less than 145°. Initial estimates of accuracy suggested that locations derived by triangulation had a 95% probability of being within 200 m of the actual location. All locations were recorded using Universal Transverse Mercator coordinates (nearest 10-m interval). Home range size was estimated as the area within a 75% probability contour generated with harmonic means on a grid size of 25 X 25 for each radio-marked bird for each season (Dixon and Chapman 1980). Habitat was examined at both observed and 'available' sites. Three observations were collected for each sex, age, season, and time period. Available sites were chosen relative to observed sites (modification of a stratified sampling method) and were randomly selected within a 0.5-krn circle (0.5 krn was representative of a typical prairie-chicken flight distance) centered over the observed site. One available site was measured for each observed site. To eliminate problems associated with measuring habitat variables in a changing environment (e.g., snow cover, plant growth, grazing pressure), both sites were examined on the same day. Two 18-m perpendicular transects were established in the center of the site with orientation of the initial transect being randomly determined. Ten point-intercept locations, 2 m apart, were located along each transect (total of 20 points). All plant species intercepted at each point were identified and recorded. Plant names follow Scott and Wasser (1980) and Harrington (1964) (Appendices A, B). Statistical analyses were used to compare general categories and/or common species including; sand sagebrush, needle-and-thread, sixweeks annual fescue, blue grama, sand dropseed, and prairie sandreed. Height-density-index (HOI) was recorded from a height of 1 m and at a distance of 4 m to one side of the transect for all 20 points. HOI was recorded as the height of vegetation obstructed on a Robel Pole (to the nearest 5 cm) (Robel et a1. 1970~). Heights of sand sagebrush, grasses, and forbs were recorded to the nearest 5 cm. A single 25-m2 circle was also centered on each site (2.82 m diam) and all plant species within the circle were identified and recorded. Location, slope, and aspect were also recorded at each site. Habitats were statistically compared with Lmax and ~ tests (assuming unequal variances). Habitat on the main 301 krn2 study area was also mapped according to dominant habitat types. Since the study area was generally divided in management units that were portions of sections (1 section - 1 mi1e2 - 640 acres), mapping was initiated by first dividing each section into 64 smaller units of 10 acres (about 4 ha). Each unit was classified as 1 of 45 different dominant habitat types. The only deviation from this method was that roads and areas of center-pivot irrigation were mapped without regard to the underlying division of units. Each habitat is described in detail in this report. �82 Female Preference Hypothesis Areas believed to be important for greater prairie-chickens, and grouse in general, are often selected by delineating an arbitrary circumference of habitat around all known leks. These areas are considered crucial for nesting habitat, a factor considered when for habitat disturbance or loss. However, little is known about movements and nest sites in relation to lek location. lekking generally mitigating female Research on lekking systems in other species has provided testable hypotheses that may help in understanding the evolution of a lekking breeding behavior from its hypothesized territorial precursor. One particular theory, the 'female preference' hypothesis, has been proposed by Bradbury (1981). Bradbury suggested that increase in female home range size is responsible for the transition from a territorial to a lekking breeding system. He further hypothesized that lek dispersion is based on female home range size and the detectability of leks. Bradbury also suggested that females should prefer to visit a cluster of males, rather than a single male, thereby improving their opportunities for mate choice. Bradbury (1981) specifically predicted that: 1) the diameter of female home ranges should be less than inter-lek distances; 2) most females should have only 1 lek within their home range; 3) females located between leks will be exceptions and will visit either or both leks within their range of detection; and 4) each lek should have an exclusive population of females. Thus, it would be expected that most females should visit only 1 lek. I examined Bradbury's predictions for the 'female preference' hypothesis by determining lek - nest distances (an indirect measure of home range size) and lek visitation behavior of female greater prairie-chickens. RESULTS Three different trapping methods were used to capture 373 (including 122 recaptures) greater prairie-chickens during 1986-88 (Table 1). Since cannon nets were considered to cause more disturbance than other trapping methods, they were used only sparingly in 1986. Walk-in traps were extremely effective during both winter and breeding seasons. Their success was positively correlated with snow cover in winter and date during the breeding season (Table 2). A total of 243 different greater prairie-chickens was captured and banded in 1986-88 (Table 3); radios were fitted to 145 birds. Measurements of primary length and diameter, pinnae length, and weight demonstrated that adults were generally larger than yearlings within each sex (Table 4). In addition, males were larger than females within each age category for all feather measurements and weights (f < 0.01). The lengths of primaries I, II, and IX and the diameter of primary IX were the best measurements for discriminating between sex and age categories. Although weight also may be a useful measurement, weight was variable with respect to season (Fig. 2). For example, females weighed more during the breeding season than in winter (f < 0.05); males did not differ. Weight variation within the breeding season was also examined (Fig. 3); weights for males declined and weights for females increased. �83 Table 1. Number of greater prairie-chickens captured by different in 1986-88 in northeastern Colorado (including recaptures). Walk-in traps at winter feeding sites Category Walk- in traps at lek sites techniques Cannon nets at lek sites Totals Males Adults Yearlings 76 32 44 86 56 30 21 14 7 183 102 81 Females Adults Yearlings 52 30 22 132 61 71 6 2 4 190 93 97 128 218 27 373 Totals Table 2. capturing Colorado, Trapping Trapping success rate (percent captures per trap per day) for greater prairie-chickens with walk-in traps on leks in northeastern 1986-88. period 4 Mar - 10 Mar 11 Mar - 17 Mar 18 Mar - 24 Mar 25 Mar - 31 Mar 1 Apr - 7 Apr 8 Apr - 14 Apr 15 Apr - 21 Apr 22 Apr - 28 Apr > 28 Apr Trap days Males Females All captures 62 59 173 384 435 389 283 91 39 3.2 1.7 3.5 0.8 6.9 6.7 4.2 4.4 5.1 0.0 0.0 1.2 2.1 9.2 12.3 11.7 1.1 0.0 3.2 1.7 4.6 2.9 16.1 19.0 15.9 5.5 5.1 �84 Table 3. Number of greater prairie-chickens banded and fitted with radio transmitters, and their fates, in northeastern Colorado, 1986-88. Fates of radio-marked birdsa Capture data Bands Radiosb Males Adults Yearl ings 108 61 47 34 20 14 5 38 2f 7 4 Females Adults Yearl ings 135 65 70 111 48 63 Totals 243 145 Category Removedc Lostd Predated 3 9 5 4 13 6 7 4 4 0 3 2f 1f 19 11 8 44 17 27 35 14 21 13 6 7 8 26 53 48 17 Mortalities aAs of 31 Dec 1988. bBirds with radio transmitters were also banded. CAt least 20 radios were recovered after premature loss. dBirds potentially were lost because of several reasons including: inoperative radios on live birds, and inoperative radios on dead birds. 9Three birds died of heat stress. fRaptors killed 5 birds in traps in a 3-day period. Table 4. long untracked movements, Measurements of greater prairie-chickens captured in northeastern Colorado, 1986-88. Malesa Adults Measurement At ive Femalesa Yearl ings Adults e. !!. -~ SO 0.36 0.28 0.31 0.46 0.49 0.58 0.43 0.46 0.39 0.57 *** *** * ** ** *** * 61 61 61 61 60 60 59 75 75 76 11.34 11.72 12.28 13.44 16.65 16.37 16.51 16.37 15.40 12.36 3.19 0.11 *** 59 7.72 0.52 49 ~ SO !!. ~ SO 11.69 12.07 12.66 14.01 16.43 17.16 17.31 17.27 16.28 12.93 0.42 0.38 0.41 0.50 0.56 0.59 0.46 0.45 0.57 0.53 35 34 33 35 35 36 35 56 55 57 11.23 11.76 12.45 13.70 16.27 17.10 17.33 17.04 15.70 13.19 Primary IX diameter (mm) 52 3.39 0.12 55 Pinnae length (em) 51 0.47 27 !!. Primary length (cm) I 59 II 59 III 59 IV 57 V 56 VI 59 VII 59 VII I 73 IX 72 X 75 7.93 Yearlings b e.b !!. ~ SO 0.44 0.63 0.70 0.63 0.56 0.53 0.59 0.62 0.69 0.45 58 58 58 58 57 56 56 76 75 77 10.84 11.32 11.96 13.04 15.33 16.19 16.45 16.23 14.94 12.47 0.31 0.33 0.38 0.39 0.48 0.45 0.40 0.40 0.52 0.39 *** *** ** *** *** * 3.26 0.10 61 3.11 0.12 *** 3.78 0.50 50 3.62 0.52 *** lJeight (g) 100 1034.8 60.8 78 1011.1 46.8 ** 93 899.7 58.6 93 871.0 60.5 aMeasurements for males and females within each age category differed (e. < 0.01, !test). bThe probability for differences in measurements (1 test) between adults and yearlings within each category of sex were: * = e. < 0.05, ** = e. < 0.01, and *** = e. < 0.001. ** �1200 1100 ,............. 0> "'--" 1000 l- I (!J W 900 N=32 N=27 N=46 N=73 800 N=67 ~ N=16 700 N=26 N=77 ~~--------------------------------------------------------------- WINTER SPRING ADULTS WINTER SPRING YEARLINGS MALES WINTER SPRING ADULTS WINTER SPRING YEARLINGS FEMALES J1ig. 2. Mean (horizontal bar), range, (vertical bar), and standard error (enclosed horizontal bars) of seasonal weigllts for each sex and age category of greater prairie-cllickens in nortlleastern Colorado, 1986-88. OJ VI �• 1'- 1150 IHlO ICI50 ............_. 0> '-"'" lI (!J w ~ IHlO ••• .t. • •••••••• ~,~ •• - - •• ~ ICIIIO .. . 1t•• _ I • _ II - • YEARUNG MALES l21li • • .II ••••• ..~-= ;.;;. ... •- •.• ' ~+----r--~---;r_--~--or--~---,~--~--~--, II II .II • 51 5 6 .III • • -t---,---,---,---,---,---,---,---,---,---, , .III 10 ~ aD •• • 51 • •• I lID • • lIID ••+----r--~---,r_--~--,_---r--_.----r_--~--, II o. S • •• • •• - • • • • •• • •• " •• • 00 YEARUNG FEMALES 1150 -.- •• • •• • 1Il10 • .. • ••• .. •• - • - ••• • . . •• .:.• ..• -L.. .. .::-,~• .. - .~.~.i . •I·.·· • • • . - • •• ••+---_r---.r_--~--_r--_,r_--~--_r--_,r_--~--~ .III 10 • .III II 51 • II • ICI50 • ... .'~~ s·. •• ICIIIO-i. ADULT FEMALES ADULT MALES • APR .III 10 • MAy .III • II II .II ••••• •••• • 51 10 iii aD 2Ii APR DATE Fig. 3. Weights for adult male (~= 73, y = -l.08x + 1075.89, ~2 = 0.071, f = 0.023), yearlng male (n = 46, y = -0.67x + 1039.03, r2 = 0.072, P = 0.072), adult female (n = 66, y = 2.l8x + 859.76, r2 0~134, f = 0.003), and yearling-female (~=-74, y = 2.83x + 708.77, ~7 = 0.154, P = 0.001) greater prairie-chickens captured in northeastern Colorado during the 1986-88 breeding s;asons. �87 Density Sixty-five leks were documented on the 301 km2 inner study area during 198688; 36 additional leks were found on the surrounding 631 km2 area in 1988 (Fig. 4, Table 5). Eleven additional display sites with single males were also documented. The density of leks on the inner study area was relatively stable at 0.14 leks/km2 (0.37 leks/mile2) (Table 6). The mean distance between nearest neighboring leks was 1.31 (SD - 0.56), 1.20 (SD - 0.70), and 1.18 (SD - 0.62) km for each year, respectively (medians of 1.42, 1.32, 1.15 km). The slight increase in lek density between 1987 and 1988 was mirrored by a significant increase in maximum attendance of males at leks (~= 4.760, f = 0.001). The 1988 density of leks on the surrounding area was 0.06 leks/km2 (0.15leks/mile2). Twenty-two percent (9 of 41) of leks active in 1986 were inactive in 1987 (Fig. 5). Twenty-four percent (10 of 42) of leks active in 1987 were inactive in 1988. Twenty-eight leks were active for all 3 years, 9 for 2 years, and 28 for 1 year. The lek disappearance rate was 22.9%, less than the 41.5% disappearance rate estimated by Miller (1984) (X2 = 4.594, f < 0.05). Reproduction Eggs in 41% of all nests hatched (Table 7). Unlike other studies, nest success did not vary with respect to year, age, and nest order (i.e., first nests or renests) (f> 0.05). The mean date for initiation of incubation for successful nests was 15.1 May (n - 30) and for unsuccessful nests 12.9 May (n = 35) (~- 1.057, f - 0.294). Wrangham (1980) suggested that proximity of nests to leks could affect their probability of success. In this study, the distance between each nest and the lek where the particular female was first observed was determined. Successful nests were a mean of 4.58 km (n = 30) and unsuccessful nests 3.20 km (n - 45) from the first lek visited (~ - 1.108, f 0.271). Distance between each nest and the nearest lek was also determined (Fig. 6). Successful nests were a mean of 1.18 km (n = 28) and unsuccessful nests 0.95 km (n = 44) from the nearest lek (~- 1.760, f - 0.083). Successful first nests tended to be further from leks than successful renests; the mean distance for 21 first nests was 1.32 km and for 7 renests was 0.77 km from the nearest lek (~- 3.389, f - 0.002). When nest success was not considered, there was no apparent relationship between time of initiation of incubation and distance to the nearest lek (Fig. 7). The overall mean date for initiating incubation was 13 May. There were no differences in incubation date associated with either age or year (f> 0.05). Adults had a slightly higher tendency to renest which may explain why differences in incubation dates between adults and yearlings were not detected (Fig. 8); 48.3% (n - 89) of first nesters were adults and 62.5% (n = 16) of .renesters were adults (X2 - 1.092, f> 0.05). The mean duration of incubation was 24.4 days (n = 29, SD - 0.6). Adults had a mean of 10 eggs in their first nests (n - 13); yearlings had 8.65 eggs (n - 17) (~= 1.045, f = 0.305). Only first nests were considered in the previous analysis since the mean clutch of 9.23 eggs (n - 30) for a first nest was larger (~ - 3.450, f - 0.001) than the mean clutch of 4.89 eggs (n - 9) for renests. Yearlings did not nest further from the nearest lek than adults (~- 0.253, f 0.801); yearling nests were a mean of 1.01 km (n - 43) and adult nests 0.98 km (n = 38) from the nearest lek (Fig. 9). �88 N o R 47 W i DISPLAY SITES R 46 W ,- ~ -, - ~ -;- ~ - ~ -~ -~ ! , R 45 W - ~ ~ ~ -;- ~ - R 44 W ! ~ -; - -~- ~ ~ ~ , ! T4N T4N T3N T2N T2N T1N T1N R47W R46W R45W R44W o Fig. 4. Colorado 2 4km Display sites of male greater prairie-chickens in northeastern during the early - late spring seasons (IS Feb-IS May) 1986-88. �89 Table 5. Location of display sites for male greater prairie-chickens in northeastern Colorado and the maximum number of males attending each display site during the 1986-88 breeding seasons (a plus sign symbolizes an active lek without an accurate count of males; a minus sign symbolizes a site that was not adequately monitored). Universal Transverse Mercators (Zone 12) Lek 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 # Male attendance Legal location East (m) North (m) Township 707970 709090 709830 710230 710480 710650 711180 711380 711750 712900 713270 714390 714540 714880 715420 715820 715830 716370 716610 716670 717050 717100 717150 717180 717250 717730 717750 717780 717930 718030 718130 718350 718690 718860 719030 719040 719100 719340 719590 719630 719890 719920 720060 720470 720620 720980 721230 721250 721620 721750 721900 722100 722370 722440 722510 4448270 4449010 4447340 4454050 4449370 4454850 4449000 4453280 4450530 4458300 4462440 4464320 4460620 4453420 4454310 4453760 4448010 4453120 4456480 4453840 4455600 4457710 4448870 4454700 4454290 4441630 4451570 4447490 4464480 4448140 4450100 4449920 4453530 4457680 4453800 4440690 4443610 4451710 4464240 4453920 4464190 4449630 .4448250 4456320 4455520 4458240 4455970 4450580 4445270 4461370 4455230 4453700 4463250 4453940 4456230 3N 2N 2N 3N 2N 3N 2N 3N 3N 3N 4N 4N 4N 3N 3N 3N 2N 3N 3N 3N 3N 3N 2N 3N 3N 2N 3N 2N 4N 2N 2N 2N 3N 3N 3N 1N 2N 3N 4N 3N 4N 2N 2N 3N 3N 3N 3N 2N 2N 4N 3N 3N 4N 3N 3N Range 461J 4611 4611 4611 4611 4611 4611 4611 4611 4611 4611 4611 4611 461J 461J 4611 461J 4611 4611 4611 4511 4511 4511 4511 4511 4511 4511 451J 4511 451J 4511 4511 4511 4511 4511 4511 4511 4511 4511 4511 4511 4511 4511 4511 4511 4511 4511 4511 4511 4511 4511 4511 4511 451J 4511 Section 7 5 8 20 5 21 4 28 33 10 27 23 35 25 24 24 12 25 13 24 18 7 7 19 19 31 31 7 19 7 6 6 29 8 20 5 29 32 20 28 21 5 8 16 16 9 16 4 21 34 22 27 27 22 15 Quarter NIJ SII SE SE NE Nil SII NE SE NE NE NE SE NIJ SII SIJ Nil NE NE SE SII SII NIJ Nil SII Nil NE SE NE SE NE NE Nil SIJ SII Nil Nil NE NE SII Nil SE NE Nil SIJ NE SE NE NE NIJ Nil Nil NE SE NE 1986 1987 + 14 3 + 3 5 6 3 6 7 0 14 3 3 0 0 9 0 9 9 2 4 15 0 12 0 0 0 2 0 0 8 8 0 0 0 2 9 0 2 0 + + 5 0 0 2 14 0 0 0 6 0 0 18 2 0 2 6 0 0 + + 18 11 10 14 0 0 5 1 + 0 0 + 1988 8 3 3 5 1 2 17 10 3 2 7 4 10 0 2 5 19 3 10 17 0 12 17 1 12 6 2 0 9 4 15 12 26 8 8 8 19 0 4 0 3 24 0 8 0 16 8 8 3 6 12 14 6 0 0 �90 Table 5. Continued. Universal Transverse Mercators (Zone 12) Lek # 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 Male attendance . Legal location East (m) North (m) Township 722850 723020 723300 723320 723440 723450 723780 724110 724120 724180 724270 724420 724450 724830 724960 725280 725750 725870 726290 726410 726520 726700 727160 727280 727410 727450 727590 727760 728090 728280 728300 728590 728630 728730 728740 728830 728940 729030 729820 730150 730360 730390 730430 730550 730970 731000 731320 732060 732420 732470 733080 734040 734150 734620 734700 735240 735690 4456700 4456460 4458750 4463170 4450180 4455900 4444340 4453510 4454510 4463360 4459920 4449180 4455370 4448130 4447950 4450980 4447080 4445950 4447050 4449940 4455730 4450140 4444880 4444760 4461430 4454630 4456700 4462440 4458320 4447390 4460090 4448100 4448610 4445330 4448260 4464820 4448110 4457150 4466220 4458700 4449870 4446390 4450070 4464090 4455770 4447120 4445470 4462010 4445100 4454660 4446330 4459030 4451990 4459970 4453650 4448720 4456600 3N 3N 3N 4N 2N 3N 2N 3N 3N 4N 3N 2N 3N 2N 2N 3N 2N 2N 2N 2N 3N 2N 2N 2N 4N 3N 3N 4N 3N 2N 3N 2N 2N 2N 2N 4N 2N 3N 4N 3N 2N 2N 2N 4N 3N 2N 2N 4N 2N 3N 2N 3N 3N 3N 3N 2N 3N Range 45101 45101 45101 45101 45101 45101 45101 45101 45101 45101 45101 45101 45101 45101 45101 45101 45101 45101 45101 45101 44101 44101 44101 44101 44101 44101 44101 44101 44101 44101 44101 44101 44101 44101 44101 44101 44101 44101 44101 44101 44101 44101 44101 44101 44101 44101 44101 44101 44101 44101 44101 44101 44101 44101 44101 44101 44101 Section 15 15 2 26 2 14 23 26 23 26 2 2 23 11 12 36 13 13 13 1 18 6 19 19 31 19 18 30 8 17 5 8 8 20 8 20 8 8 16 9 4 16 4 21 16 16 21 34 22 22 14 2 35 1 25 12 13 Quarter NE NE SIJ NIJ NIJ SIJ SIJ NE SE NE NE SE NE SE SIJ SIJ NIJ SE NE SE SIJ NIJ SE SE NE SE NE SE NIJ NIJ NIJ SIJ NIJ NIJ NIJ NE SE SE NIJ NIJ SIJ SIJ NIJ SE SE NE NE NE SE SE SIJ SE NE NIJ NIJ NIJ NE 1986 1987 1988 4 4 5 5 6 0 0 0 0 2 0 0 4 0 0 0 5 10 0 12 7 7 0 2 2 0 12 9 8 0 6 8 0 6 8 8 0 5 0 2 9 9 9 0 4 6 0 2 4 1 5 0 1 1 13 10 1 14 0 8 10 0 0 1 9 8 10 + + 12 4 0 0 8 9 0 0 7 10 0 0 0 + 2 0 7 0 0 18 0 8 0 10 0 6 4 4 0 + + 7 6 17 9 1 8 0 1 1 0 10 10 8 1 5 2 3 1 26 4 6 9 14 4 6 5 4 0 2 3 8 10 3 1 �91 Table 6. Colorado, Indices 1986-88. of male greater prairie-chicken abundance in northeastern Index 1986 1987 1988 n leks 41 42 47 ~ displaying males/lek 7.11 ± 3.72 7.26 + 4.35 10.66 ± Lek density Leks/km2 Leks/mile2 0.14 0.35 0.14 0.36 0.15 0.40 Displaying male densitya Males/km2 Males/mile2 0.97 2.52 l.02 2.64 l. 69 4.38 aMales displaying 6.46 alone were used in the analysis. Table 7. Success rate and date for initiation of incubation for female greater prairie-chicken nests in northeastern Colorado, 1986-88. Success rate Initiation Category n Year 1986 1987 1988 21 26 32 33.33 42.31 43.75 22 25 36 39 40 4l.03 40.00 Nest order First Renests 64 15 Totals 79 of incubation ~ date SD 12 May 12 May 11 May 15.3 May 11.4 May 12.8 May 11.5 4.8 44 39 11 May 13 May 13 .1 May 12.9 May 9.2 6.6 39.06 46.67 68 15 10 May 25 May 10.1 May 26.3 May 4.2 8.0 40.51 83 12 May 13 .0 May 8.1 Rate (%) n Median date 7.2 Age Adults Yearlings �R_W \D RoMW R«iW N T4N T4N N 1986 8 SIZE OF LEKS TliN T3M 2-5 MALES o 6-10 MALES o 11-15 MALES o >16MALES T2M 8 8 WRAY o , R .• W R .• W R_W 1987 1988 R4fiW T3M 4kn1 , RoMW T4N ·-------~·~·t...~.... O"~. ~ "'1 ' 8 TaM I T4N T4M T4M 2, RoMW R«iW RoMW R«iW , T3N : ....•. j > : 8 T3N [J ····i·· ..... o ~ ~l ~ T2M ( ,..;. :0 ·····i T2N : . ., ..•.. Fig. 5. Distribution of greater ).9.L : .... :... 8 : s R«iW ····~···O;··...:..... , ..... j : Q ~0 0 .O R_W o : ® RoMW prairie-chicken R_W leks in northeastern R«iW Colorado, WRAY RoMW 1968-88. �93 24 ~ UNSUCCESSFUL (n=65) II SUCCESSFUL (n=28) 1.6-2.0 2.0-2.4 20 16 en ~ 12 W Z 8 4 o 0.0-0.4 0.4-0.8 0.8-1.2 1.2-1.6 DISTANCE (KM) Fig. 6. Distances between nest sites and the nearest lek for successful and unsuccessful greater prairie-chicken nests in northeastern Colorado, 1986-88. �94 18 I EARLY (n=48) [] MID (0=25) 15 I LATE (n=6) 12 ~ w 9 z 6 3 o 0.0-0.4 0.4-0.8 0.8-1.2 1.2-1.6 1.6-2.0 2.0-2.4 DISTANCE (KM) Fig. 7. Distances between nest sites and the nearest lek vs. date of nest initia tion (early = 28 Apr - 12 May, mid = 13 - 27 May, late = 28 May - 11 Jun) for greater prairie-chickens in northeastern Colorado, 1986-88. �95 20 ~ ADULTS (1J=44) • YEARLINGS (D =39) 16 en W 12 _J <C ~ W LL 8 4 28 APR • 4 MAY 5 MAY· 12 MAY· 19 MAY· 26 MAY· 11 MAY 18 MAY 25 MAY 1 JUN > 1 JUN DATE OF NEST INITIATION Fig. 8. Colorado, Date of nest initiation 1986-88. for greater prairie-chickens in northeastern �96 16 ~ ADULTS (n=38) II YEARLINGS (n=43) 12 en UJ ....J « 8 ~ UJ u.. 4 o 0.0-0.4 0.4-0.8 0.8-1.2 1.2-1.6 1.6-2.0 2.0-2.4 DISTANCE (KM) Fig. 9. chickens Distances between nest sites and the nearest in northeastern Colorado, 1986-88. lek for greater prairie- �97 Movement Home range size of greater prairie-chickens was estimated by sex and age class for each season (Table 8). Yearling males tended to have larger home ranges than adult males for all but the late summer season (f> 0.05). However, there were small sample sizes of yearlings in each case. For females, the only difference detected was that adults had larger home ranges than yearlings during autumn (~- 2.734, f - 0.739). However, these results require cautious interpretation because of the large number of statistical comparisons. When ages were combined, females had larger home ranges than males during late spring (~~ 2.236, f - 0.027) and autumn (~~ 2.889, f - 0.006). The trends were in the same direction for early summer, late summer, and winter. Season also was associated with differences in home range size (f < 0.05). Males had larger home ranges in winter than in late summer and autumn. Females had smaller home ranges in late summer than in autumn, winter, and late spring, and smaller home ranges in early spring than in late spring. Some differences in home range size apparently were associated with behavior such as brood status of females (Table 9). For example, during late summer, females with broods tended to have larger home ranges than females without broods (!= 1.864, f - 0.072). Habitat Comparisons Comparison of habitat characteristics between observed and available sites for early spring revealed no differences (Table 10). Males were observed in areas with shorter sand sagebrush, grass, and forbs, lesser sand sagebrush cover, and greater blue grama cover (Table 11). Many of these differences appeared to reflect the frequent use of leks by males. Further examination of habitat by time period showed that differences existed in the height of sand sagebrush and forbs, cover of bare ground, grass, sixweeks annual fescue, warm season grass, sand dropseed, corn, and forbs, and species richness (Table 12). In most cases, the differences apparently reflected the regular use of relatively open, over-grazed habitats by males during their morning and evening display periods. Few differences were found between habitat at observed and available sites for late spring (Table 13). Two exceptions were that grasses and forbs were shorter at observed sites. Males were observed in areas with shorter forbs, lesser blue grama and forb cover, and lesser species richness and slope than females (Table 14). Further examination of habitat by time period showed that differences existed in sand sagebrush height, sand sagebrush, grass, cool season grass, needle-and-thread, forb cover, and slope (Table 15). As with early spring, many of these differences appeared to reflect the regular use of lek sites by males, particularly during morning and evening periods. Furthermore, lek site habitat differed dramatically from available habitat in that lek sites had short and sparse cover (Table 16). Lek sites were located on low, over-grazed ridgetops. These data were supported by an additional comparison between permanent (active 1986-88) and temporary (active < 3 years) leks which showed that permanent leks had shorter forbs, greater blue grama cover, and lesser forb cover .and species richness (Table 17). No differences were found between habitat at observed and available sites for early summer (Table 18). Females were observed in areas with taller forbs, greater forb cover, and species richness than males (Table 19). There were no �98 Table 8. Home range size (area within 75% probability contours generated with harmonic means on a grid size of 25 X 25, ha) of greater prairie-chickens in northeastern Colorado, 1986-88. Males Females All Category Ad Yrlg All Ad Yrlg All birds Early spring n Median ~ SO Late spring n Median ~ SO Early summer n Median ~ SO 12 107.9 167.5 164.2 2 535.2 535.2 568.1 14 127.2 220.2 255.9 22 165.0 175.9 97.2 8 134.9 313.6 439.2 30 160.8 212.6 239.3 44 155.5 214.9 241.7 19 91.0 108.0 97.3 6 250.2 401. 2 482.7 25 92.9 178.3 268.3 48 264.5 367.3 307.6 47 276.2 899.6 2671.9 95 266.0 630.6 1900.7 120 232.0 536.4 1703.6 15 160.5 162.4 82.0 5 291.5 377 .4 316.5 20 164.2 216.2 187.5 26 212.3 777.1 2866.2 28 174.5 694.0 1429.4 54 379.1 734.0 2217.6 74 180.5 594.1 1906.1 13 108.1 130.2 109.9 4 122.6 106.7 62.1 17 118.3 124.7 99.4 21 111.8 153.7 129.9 21 121.4 158.5 110.9 42 115.9 156.1 119.3 59 118.3 147.0 114.0 10 111.3 121.8 87.9 4 213.1 210.6 95.7 14 127.5 147.2 95.9 14 397.3 469.5 342.5 16 114.0 191.2 178.1 30 289.2 321.1 298.2 44 186.7 265.7 263.6 9 208.2 226.4 105.6 4 260.1 242.5 126.9 13 221.1 231.4 107.3 16 244.1 448.5 538.3 7 236.4 261.4 144.8 23 239.4 391.5 459.4 36 237.9 333.7 377.8 Late summer n Median z SO Autumn n Median ~ SO Winter n Median ~ SO �99 Table 9. Home range size (area within 75% probability contours generated with harmonic means, ha) of female greater prairie-chickens with and without broods in northeastern Colorado during late summer, 1986-88. Category With broods n 19 212.3 194.1 139.6 Median ~ SD Without broods All females 23 108.6 124.7 91.1 42 115.9 156.1 114.1 Table 10. Habitat at observed and available locations of greater prairie-chickens in northeastern Colorado during early spring (15 Feb Mar) 1986-88. Observed Habitat variable ~ (36) SO Available ~ - 31 f (36) SD Lax .t 3.17 2.87 3.15 2.52 0.438 0.972 Height, cm Sand sagebrush Grass Forbs 59.31 95.42 57.50 40.27 37.95 44.33 66.53 97.22 60.00 35.19 24.94 37.34 0.429 0.015 0.315 0.421 0.812 0.797 Cover, % Bare Sand sagebrush Grass Cool season grass Needle-and-thread Sixweeks annual fescue Warm season grass Blue grama Sand dropseed Prairie sandreed Corn Forbs 29.72 8.47 65.56 18.19 16.53 l. 53 50.97 15.69 11.11 10.00 4.31 2.22 18.93 10.27 19.23 16.39 16.47 3.55 16.98 15.31 12.82 12.87 12.54 4.22 29.72 7.78 63.19 20.00 17.08 2.64 46.11 12.92 15.83 6.67 3.33 4.31 18.59 9.74 20.43 18.78 16.10 7.61 16.00 12.33 16.06 9.18 11.65 5.37 0.915 0.757 0.722 0.425 0.894 0.000 0.726 0.204 0.188 0.049 0.664 0.159 1.000 0.769 0.615 0.665 0.885 0.430 0.215 0.399 0.172 0.211 0.734 0.071 1l.03 5.73 12.17 5.11 0.504 0.377 l. 50 l. 56 2.28 2.39 0.014 0.107 Height-density-index, Species richness, Slope, degrees n em �100 Table 11. Habitat at locations prairie-chickens in northeastern Mar) 1986-88. of radio-marked male and female greater Colorado during early spring (15 Feb - 31 Males Habitat variable ~ (18) SD Females ~ (18) f SD Imax .t 2.51 3.48 3.83 1. 99 0.027 0.173 Height, em Sand sagebrush Grass Forbs 44.17 74.44 39.44 32.10 38.07 38.04 74.44 116.39 75.56 42.70 24.18 43.69 0.249 0.070 0.574 0.022 0.000 0.012 Cover, % Bare Sand sagebrush Grass Cool season grass Need1e-and-thread Sixweeks annual fescue Warm season grass Blue grama Sand dropseed Prairie sandreed Corn Forbs 31.11 4.44 67.22 22.22 19.72 2.50 50.56 22.50 7.50 10.00 2.50 1. 39 17.95 5.39 17.84 18.25 18.43 4.62 17.23 16.56 8.27 12.00 10.61 2.87 28.33 12.50 63.89 14.17 13.33 0.56 51. 39 8.89 14.72 10.00 6.11 3.06 20.29 12.40 20.90 13.64 14.04 1. 62 17.22 10.51 15.57 14.04 14.30 5.18 0.619 0.001 0.521 0.240 0.272 0.000 0.999 0.069 0.013 0.526 0.228 0.020 0.666 0.019 0.610 0.143 0.250 0.107 0.885 0.006 0.094 1.000 0.396 0.243 9.83 5.28 12.22 6.05 0.581 0.216 1.17 1.04 1. 83 1.92 0.016 0.206 Height-density-index, Species richness, Slope, degrees n cm �101 Table 12. Habitat at locations of radio-marked greater prairie-chickens in northeastern Colorado during morning, afternoon, and evening periods during early spring (15 Feb - 31 Mar) 19S6-SS. Significant differences (f < 0.05) between means for each habitat variable are indicated by different superscripts. Morning Habitat variable Height-density-index, cm Height, cm Sand sagebrush Grass Forbs Cover, % Bare Sand sagebrush Grass Cool season grass Need1e-and-thread Sixweeks annual fescue Warm season grass Blue grama Sand dropseed Prairie sandreed Corn Forbs Species Slope, richness, degrees n (12) (12) Evening (12) SD SD so 1. 94 3.77 2.52 S9.SSa,b 90.S3a 49.17a 45.30 44.15 40.61 21. 2Sa 10.42a 71.67a 14.17a 10.00a 3.7Sa S9.SSa 17.S0a 17.0sa 10.S3a 12.0S 13.05 13.20 13.62 12.61 4. S2 15.59 19.36 16.SS 15.20 77.0Sa 10S.00a S6.2Sb 3.26 27.92a,b 7.0Sa 67.S0a,b 21.67a 21.67a O.OOb 50.42a,b 14.SSa 12.0Sa 12.0sa O.OOa 4.Ssa 4.SS 14.0Sa O.OOa 1. 67a,b Afternoon 1. 29 30.41 34.1S 36.44 41. 2Sb 90.42a 37.0Sa 3S.50 36.27 42.93 22.00 S .11 24.73 20.26 20.26 40.00b 7.92a 57. SOb lS.7Sa 17.92a O. S3a,b 42.92b lS.00a 4.17b 7.0Sa 12.92b 0.42b 17.71 9.64 16.58 15.09 14.69 2.89 14.22 15.23 5.15 8.38 19.48 1. 44 17.90 11.57 11.37 14.53 S.S2 5.21 5.24 1.40 1. 78 �102 Table 13. Habitat at observed and available locations of greater prairie-chickens in northeastern Colorado during late spring (1 Apr 1986-88. Observed Habitat variable ~ Height-density-index, cm (36) SO Available ~ - (36) 15 May) f so Lnax ! 3.56 3.73 4.62 6.53 0.001 0.402 Height, cm Sand sagebrush Grass Forbs 55.28 76.94 45.28 42.83 34.61 40.32 66.67 97.92 67.50 28.74 23.71 36.95 0.021 0.028 0.608 0.190 0.004 0.017 Cover, % Bare Sand sagebrush Grass Cool season grass Need1e-and-thread Sixweeks annual fescue Warm season grass Blue grama Sand dropseed Prairie sandreed Corn Forbs 29.72 9.03 60.56 23.89 17.22 5.00 41.39 14.86 7.22 6.25 2.50 7.50 23.69 10.54 24.17 18.86 16.10 8.28 19.66 16.88 7.22 7.40 7.32 13.86 24.58 6.39 65.56 28.19 20.56 7.92 43.47 11.25 11.39 10.00 0.14 8.89 20.89 8.75 23.57 22.81 21.77 11.49 17.40 14.51 10.11 10.82 0.83 9.42 0.460 0.275 0.883 0.266 0.079 0.057 0.472 0.375 0.049 0.028 0.000 0.025 0.332 0.252 0.377 0.386 0.463 0.221 0.636 0.334 0.049 0.091 0.062 0.621 12.92 7.58 15.33 6.45 0.341 0.150 1.97 1.68 3.06 3.09 0.001 0.070 Species richness, Slope, degrees n �103 Table 14. Habitat at locations prairie-chickens in northeastern 1986-88. of radio-marked male and female greater Colorado during late spring (1 Apr - 15 May) Males Habitat variable X Height-density-index, cm (18) SD Females X (18) f SD Emax ! 3.09 4.87 4.04 2.11 0.001 0.453 Height, em Sand sagebrush Grass Forbs 47.78 72.50 26.94 39.79 38.74 32.95 62.78 81. 39 63.61 45.54 30.38 39.36 0.584 0.326 0.471 0.300 0.449 0.005 Cover, % Bare Sand sagebrush Grass Cool season grass Needle-and-thread Sixweeks annual fescue Warm season grass Blue grama Sand dropseed Prairie sandreed Corn Forbs 31. 39 8.89 61. 67 19.72 14.17 4.44 46.39 22.50 6.39 7.78 2.22 2.50 22.28 10.79 22.94 12.42 12.75 5.66 18.21 18.65 7.44 7.71 6.69 5.22 28.06 9.17 59.44 28.06 20.28 5.56 36.39 7.22 8.06 4.72 2.78 12.50 25.56 10.61 25.95 23.27 18.75 10.42 20.28 10.74 7.10 6.96 8.08 17.76 0.578 0.946 0.618 0.013 0.122 0.016 0.663 0.029 0.851 0.677 0.444 0.000 0.679 0.938 0.787 0.192 0.261 0.694 0.129 0.006 0.496 0.221 0.824 0.033 10.22 6.49 15.61 7.81 0.452 0.031 1.06 1.11 2.89 1. 68 0.098 0.001 Species richness, Slope, degrees n �104 Table 15. Habitat at locations of radio-marked greater prairie-chickens in northeastern Colorado during morning, afternoon, and evening periods during late spring (1 Apr - 15 May) 1986-88. Significant differences (f < 0.05) between means for each habitat variable are indicated by different superscripts. Morning (12) Habitat variable & SD Afternoon (12) & SD Evening (12) R SD 2.34a 1.93 5.634 5.16 2.72a 2.64 Height, cm Sand sagebrush Grass Forbs 49.584 67.08a 32.08a 29.88 26.24 40.75 87.08b 92.50a 64.17a 43.61 34.01 44.36 29.17a 71.25a 39.58a 34.03 39.49 30.49 Cover, % Bare Sand sagebrush Grass Cool season grass Needle-and-thread Sixweeks annual fescue Warm season grass Blue grama Sand dropseed Prairie sandreed Corn Forbs 24.174 6.254 72.084 35.424 27.504 6.254 45.004 18.334 7.084 8.75a O.Ooa 1.674 13.62 7.11 13.89 22.81 17.90 10.90 14.77 14.20 4.98 9.56 0.00 3.26 30.424 16.25b 53.33b 19.17b 12.92b 4.58a 37.504 7.924 9.58a 5.83a 3.33a 11.67b 26.92 12.99 25.08 14.75 13.22 8.11 18.40 10.97 9.64 5.15 8.07 15.57 34.58a 4.58a 56.25a,b 17.08b 11.25b 4.17a 41. 67a 18.33a 5.00a 4.17a 4.1]4 9.17a,b 28.56 6.89 28.53 13.22 12.45 5.57 25.35 22.50 6.03 6.69 9.73 17.30 Species richness, n 11.42a 4.93 15.08a 7.75 12.25a 9.52 1.67a 1.72 3.08b 1.62 1.17a 1.11 Height-density-index, Slope, degrees cm �105 Table 16. Habitat at observed and available lek locations for greater prairie-chickens in northeastern Colorado during late spring (1 Apr - 15 May) 1986-88. Observed Habitat variable ~ SD Available ~ (36) ~ SD Lnax .£ 1.04 0.77 3.00 2.17 0.000 0.000 Height, cm Sand sagebrush Grass Forbs 38.46 67.38 53.08 24.65 31.66 42.72 58.15 94.31 63.23 34.49 30.67 40.58 0.008 0.800 0.682 0.000 0.000 0.167 Cover, % Bare Sand sagebrush Grass Cool season grass Need1e-and-thread Sixweeks annual fescue Warm season grass Blue grama Sand dropseed Prairie sandreed Corn Forbs 30.92 2.77 66.62 26.85 16.77 6.85 43.69 22.62 8.15 6.31 0.31 2.00 12.40 3.75 12.57 16.76 13.09 8.32 16.85 17.35 9.75 7.41 2.48 4.98 21. 31 4.92 72.38 30.23 21. 62 7.92 48.54 14.85 16.54 8.85 1.15 6.15 15.87 6.22 15.57 20.68 17.37 12.90 16.43 14.71 15.48 11.48 5.36 8.42 0.051 0.000 0.089 0.095 0.025 0.001 0.844 0.190 0.000 0.001 0.000 0.000 0.000 0.019 0.022 0.307 0.075 0.573 0.099 8:007 0.000 0.137 0.251 0.001 11.62 5.60 14.29 5.74 0.847 0.008 0.45 0.73 2.05 1. 98 0.000 0.000 Height-density-index, Species richness, Slope, degrees n cm (36) �106 Table 17. Habitat at permanent (active in 1986-88) and temporary (inactive for at least 1 year) 1ek locations for greater prairie-chickens in northeastern Colorado during late spring (1 Apr - 15 May). Permanent Habitat (28) Temporary so variable Height-density-index, cm (37) so Lax 0.88 0.67 1.16 0.83 0.254 0.148 Height, cm Sand sagebrush Grass Forbs 33.21 58.93 28.93 20.29 27.26 28.23 42.43 73.78 71.35 27.10 33.57 43.01 0.122 0.265 0.026 0.137 0.061 0.000 Cover, % Bare Sand sagebrush Grass Cool season grass Need1e-and-thread Sixweeks annual fescue Warm season grass Blue grama Sand dropseed Prairie sandreed Corn Forbs 31.43 3.57 66.96 26.07 18.57 7.14 44.82 28.21 6.07 6.25 0.00 0.71 12.01 4.48 11.97 13.43 10.70 7.13 16.47 15.59 8.96 6.18 0.00 2.24 30.54 2.16 66.35 27.43 15.41 6.62 42.84 18.38 9.73 6.35 0.54 2.97 12.84 3.01 13.16 19.06 14.64 9.21 17.30 17.60 10.13 8.30 3.29 6.18 0.724 0.027 0.615 0.062 0.095 0.170 0.800 0.517 0.512 0.115 0.000 0.778 0.158 0.847 0.749 0.338 0.805 0.642 0-.022 0.135 0.957 0.324 0.046 9.61 3.34 13 .14 6.48 0.001 0.006 0.39 0.57 0.49 0.84 0.039 0.593 Species richness, Slope, degrees n �107 Table 18. Habitat at observed and available locations of greater prairie-chickens in northeastern Colorado during early summer (16 May Jun) 1986-88. Observed Habitat variable ~ Height-density-index, SO Available ~ (36) ~ SO Enax .t 4.09 4.07 2.94 1.77 0.000 0.128 Height, cm Sand sagebrush Grass Forbs 64.86 91. 25 68.19 38.81 33.35 36.76 54.03 84.03 69.58 41. 28 42.34 44.50 0.717 0.163 0.263 0.255 0.424 0.886 Cover, % Bare Sand sagebrush Grass Cool season grass Need1e-and-thread Sixweeks annual fescue Warm season grass Blue grama Sand dropseed Prairie sandreed Corn Forbs 22.22 5.97 67.08 32.36 17.92 10.83 43.61 13 .06 7.22 6.53 3.06 13.33 21. 30 7.54 24.42 24.83 17.50 10.79 18.88 14.99 7.79 8.35 10.57 17.24 29.44 5.14 60.69 28.06 16.94 7.78 41. 67 10.14 5.56 10.56 2.78 11.39 27.85 6.15 29.64 26.73 18.91 12.73 22.04 13.76 8.93 13.82 8.40 17.01 0.117 0.233 0.256 0.664 0.649 0.333 0.365 0.615 0.423 0.004 0.179 0.939 0.221 0.609 0.322 0.481 0.822 0.276 0.689 0.393 0.401 0.140 0.902 0.632 Species 20.69 8.90 16.75 8.75 0.919 0.062 1. 58 1. 25 1.81 2.03 0.005 0.578 richness, Slope, degrees n cm (36) - 30 �108 Table 19. Habitat at locations prairie-chickens in northeastern Jun) 1986-88. of radio-marked male and female greater Colorado during early summer (16 May - 30 Males Habitat variable ~ Height-density-index, cm (18) SO Females ~ (18) £ SO Lax .t 4.93 5.46 3.25 1.71 0.000 0.227 Height, cm Sand sagebrush Grass Forbs 53.06 83.06 51.11 43.66 36.18 35.59 76.67 99.44 85.28 30.00 28.95 29.97 0.132 0.367 0.487 0.067 0.143 0.004 Cover, % Bare Sand sagebrush Grass Cool season grass Needle-and-thread Sixweeks annual fescue Warm season grass Blue grama Sand dropseed Prairie sandreed Corn Forbs 26.67 5.56 66.67 32.22 18.89 8.06 42.50 14.72 7.22 6.39 6.11 7.22 28.08 7.65 25.90 24.45 17.95 7.88 18.49 16.31 6.69 9.04 14.51 12.97 17.78 6.39 67.50 32.50 16.94 13.61 44.72 11.39 7.22 6.67 0.00 19.44 10.18 7.63 23.59 25.91 17.50 12.70 19.74 13.81 8.95 7.86 0.00 19.09 0.000 0.993 0.705 0.814 0.918 0.057 0.791 0.499 0.241 0.569 0.121 0.220 0.746 0.920 0.974 0.744 0.124 0.730 0.513 1.000 0.922 0.092 0.031 15.33 8.76 26.06 5.06 0.029 0.000 1.61 1.14 1. 56 1. 38 0.446 0.896 Species richness, Slope, degrees n �109 differences associated with time period (Table 20). Some differences between males and females may have related to selection by females of nesting areas while males were still frequenting leks. Examination of observed and available habitat at nests sites showed that observed sites had greater height-density-index, taller sand sagebrush and grass, lesser bare cover and slope, and greater sand sagebrush, grass, and prairie sandreed cover (Table 21). Nests within 1 km of leks had shorter sand sagebrush and forbs and less species richness (Table 22). Despite the fact that nests generally were in areas with relatively dense, tall cover, successful nests differed little from failed nests (Table 23). In addition, few differences in habitat were detected for either age (adults and yearlings) (Table 24) or nest order (first nests and renests) (Table 25). Comparison of habitat characteristics between observed and available sites for late summer revealed essentially no differences (Table 26). Likewise, comparison of habitat by sex (Table 27) and period (Table 28) also failed to reveal any differences. One possible reason for the lack of differences may be that most birds used relatively extensive and homogenous areas of sand sagebrush and grass cover. During autumn, observations of greater prairie-chickens suggested that many birds were moving closer to agricultural areas, even though they were rarely found feeding in them. Habitat comparisons between observed and available sites supported this suggestion (Table 29). Greater prairie-chickens were observed in areas with denser and taller sand sagebrush, less bare ground and corn cover, and greater species richness. No differences were detected when observations were divided by sex (Table 30) or period (Table 31). Although differences in habitat generally were not detected between males and females for early summer - autumn (16 May - 31 Oct), numerous differences occurred among habitat variables at locations of brood females and non-brood females (Table 32). Broods used areas with greater height-density-index and species richness, and taller sand sagebrush, grass, and forbs. Essentially no differences were detected between observed and available habitat for winter (Table 33). Likewise, habitat for males and females did not differ (Table 34). Since greater prairie-chickens are frequently in large mixed-sex flocks during winter, lack of sex differences in habitat possibly should be expected. However there were several differences apparently associated with diurnal patterns of behavior (Table 35). For example, afternoon habitat generally had greater height-density-index and slope, taller grass, and less corn and forb cover. Some differences may represent afternoon loafing habitat and morning and evening foraging habitat in corn fields (Table 35). Habitat Description The study area was classified and mapped into 10 general habitat types (Fig. 10). Mapping revealed that 85.23% of the study area consisted of rangeland. However, only 16.61% of the area was classified as good-quality rangeland (mid-tall grass or sage/mid-tall grass). Additional classification of the area into 45 specific habitat types (Figs. 11-14, Table 36) revealed that most habitat types were delineated by units of management (i.e., they were frequently bordered by fences and managed with different intensities and types of grazing pressure or agriculture). Numerous plant species were found throughout the study (Table 37). Habitat types recognized were: �110 Table 20. Habitat at locations of radio-marked greater prairie-chickens in northeastern Colorado during morning, afternoon, and evening during early summer (16 May - 30 Jun) 1986-88. None of the differences was significant. Morning (12) Habitat variable Height-density-index, cm Afternoon (12) SD Evening (12) SD SD 3.17 l. 95 4.12 2.75 4.98 6.30 Height, cm Sand sagebrush Grass Forbs 66.67 90.00 83.75 38.40 25.50 4l.40 73.33 100.42 65.83 33.26 2l.47 24.29 54.58 83.33 55.00 44.90 47.59 39.20 Cover, % Bare Sand sagebrush Grass Cool season grass Needle-and-thread Sixweeks annual fescue Warm season grass Blue grama Sand dropseed Prairie sandreed Corn Forbs 22.50 4.58 73.33 37.50 26.25 14.17 45.00 12.92 5.42 6.25 2.50 8.75 17.77 5.42 19.23 29.58 2l.86 14.28 17.32 15.73 3.34 9.08 8.66 7.72 15.83 6.67 70.83 33.33 13.33 10.83 48.33 14.58 8.33 7.50 0.00 16.25 12.76 7.78 19.17 19.58 12.49 9.25 18.38 14.05 6.85 8.66 0.00 14.16 28.33 6.67 57.08 26.25 14.17 7.50 37.50 1l.67 7.92 5.83 6.67 15.00 29.64 9.37 3l.51 25.15 15.05 7.54 20.73 16.28 11.37 7.93 16.00 25.50 Species richness, D 22.17 8.02 22.83 6.60 17.08 11.07 l.42 l.08 2.17 l. 53 l.17 0.94 Slope, degrees �ill Table 21. Habitat at observed and available nest locations prairie-chickens in northeastern Colorado, 1986-88. Observed Habitat variable 1i (87) SD Available 1i for greater (87) f SD Lax .t. 5.85 2.59 3.06 l. 97 0.012 0.000 83.79 111.32 82.64 29.69 16.49 34.20 71.32 102.24 75.57 33.95 22.71 34.32 0.216 0.003 0.973 0.011 0.003 0.175 Cover, % Bare Sand sagebrush Grass Cool season grass Needle-and-thread Sixweeks annual fescue Warm season grass Blue grama Sand dropseed Prairie sandreed Corn Forbs 11.44 13.74 72.87 36.26 26.61 9.60 44.20 10.17 10.52 12.41 0.17 15.34 7.15 9.80 15.82 21.09 19.16 11.42 17.34 10.77 11.23 10.59 l.19 12.12 21.72 6.78 66.21 31.67 21. 95 9.02 40.57 10.06 7.53 9.14 0.17 12.70 15.90 7.11 18.03 20.82 18.67 11. 20 18.54 11.70 11.43 10.15 l.19 10.61 0.000 0.003 0.229 0.650 0.809 0.860 0.534 0.443 0.870 0.696 1.000 0.220 0.000 0.000 0.010 0.143 0.106 0.738 0.185 0.946 0.084 0.039 1.000 0.128 Species 23.15 6.11 21.94 6.22 0.868 0.198 l. 63 1. 57 2.66 2.96 0.000 0.005 Height-density-index, cm Height, cm Sand sagebrush Grass Forbs richness, Slope, degrees II �112 Table 22. Habitat at greater prairie-chicken nest sites vs. distance each nest and the nearest 1ek in northeastern Colorado, 1986-88. < 1 Habitat variable kIn ~ (45) > 1 SO ~ kIn (42) between f so !:max .t 5.59 2.79 6.12 2.37 0.290 0.342 75.56 109.89 75.11 33.65 18.60 31.42 92.62 112.86 90.71 21. 90 13.93 35.57 0.006 0.064 0.419 0.006 0.400 0.033 Cover, % Bare Sand sagebrush Grass Cool season grass Need1e-and-thread Sixweeks annual fescue Warm season grass Blue grama Sand dropseed Prairie sandreed Corn Forbs 11.22 12.44 75.00 39.00 28.89 9.89 44.11 10.67 10.11 11.00 0.00 13.00 7.62 10.64 16.20 22.73 19.54 12.41 17.81 11. 90 10.42 8.96 0.00 10.89 11.67 15.12 70.60 33.33 24.17 9.29 44.29 9.64 10.95 13.93 0.36 17.86 6.69 8.73 15.27 19.02 18.67 10.39 17.02 9.52 12.16 12.02 1. 71 12.98 0.399 0.204 0.704 0.252 0.773 0.255 0.771 0.153 0.315 0.057 0.255 0.774 0.205 0.196 0.212 0.253 0.807 0.963 0.660 0.·729 0.199 0.183 0.061 Species 21. 53 5.18 24.88 6.60 0.118 0.010 1.49 1. 38 1. 79 1. 76 0.110 0.382 Height-density-index, cm Height, cm Sand sagebrush Grass Forbs richness, Slope, degrees n �113 Table 23. Habitat at successful and unsuccessful nest sites prairie-chickens in northeastern Colorado, 1986-88. Successful Habitat variable ~ Height-density-index, cm (29) SO Failed ~ for greater (40) f SO Imax ! 5.31 1. 85 6.09 2.57 0.073 0.167 81.03 113.28 90.52 32.41 15.77 33.09 85.63 108.63 80.88 27.08 15.93 33.80 0.296 0.969 0.921 0.525 0.234 0.242 Cover, % Bare Sand sagebrush Grass Cool season grass Need1e-and-thread Sixweeks annual fescue Warm season grass Blue grama Sand dropseed Prairie sandreed Corn Forbs 12.76 11.21 73.28 35.34 29.14 5.52 44.48 8.97 12.52 14.66 0.00 17.24 7.74 7.87 15.13 22.68 22.56 8.38 15.02 9.20 12.27 12.24 0.00 12.65 10.88 16.38 73.63 40.13 27.13 11.75 42.50 10.75 10.75 11.63 0.25 12.50 6.78 10.50 17.32 20.89 18.84 12.53 18.67 11.69 11.41 10.15 1. 58 10.06 0.439 0.114 0.460 0.628 0.295 0.029 0.232 0.189 0.665 0.277 0.185 0.287 0.029 0.931 0.369 0.688 0.016 0.639 0.497 0.936 0.266 0.324 0.087 Species 25.00 6.03 22.43 5.95 0.926 0.082 2.07 1. 75 1. 23 1. 27 0.064 0.032 Height, cm Sand sagebrush Grass Forbs Slope, richness, degrees n �114 Table 24. Habitat at nests sites for adult and yearling prairie-chickens in northeastern Colorado, 1986-88. Adults Habitat variable ~ Height-density-index, cm Height, cm Sand sagebrush Grass Forbs Cover, % Bare Sand sagebrush Grass Cool season grass Needle-and-thread Sixweeks annual fescue Warm season grass Blue grama Sand dropseed Prairie sandreed Corn Forbs Species Slope, richness, degrees n (43) Yearlings SD ~ female greater (44) f SD Emax ~ 5.92 2.28 5.78 2.90 0.122 0.798 8l.16 112.67 78.60 30.80 16.99 35.23 86.36 110.00 86.59 28.68 16.07 33.09 0.643 0.718 0.685 0.417 0.453 0.279 10.70 12.21 75.93 40.70 31.05 10.93 44.30 10.93 12.21 13.95 0.35 14.42 7.04 8.82 17.60 21.42 17.58 13.15 17.17 10.76 11.46 9.79 1. 69 13.68 12.16 15.23 69.89 31. 93 22.27 8.30 44.09 9.43 8.86 10.91 0.00 16.25 7.27 10.56 13 .40 20.06 19.84 9.40 17.69 10.85 10.88 11.22 0.00 10.46 0.836 0.244 0.079 0.670 0.435 0.031 0.846 0.959 0.737 0.380 0.083 0.343 0.152 0.075 0.052 0.032 0.287 0.955 0.520 0.166 0.182 0.183 0.484 21. 91 6.96 24.36 4.93 0.027 0.062 1. 53 1.44 1. 73 1. 70 0.272 0.571 �us Table 25. Habitat at first and second nesting attempts within a single in northeastern Colorado, breeding season by female greater prairie-chickens 1986-88. First Habitat variable ~ Height-density-index, cm Height, cm Sand sagebrush Grass Forbs Cover, % Bare Sand sagebrush Grass Cool season grass Need1e-and-thread Sixweeks annual fescue Warm season grass Blue grama Sand dropseed Prairie sandreed Corn Forbs Species richness, Slope, degrees n (73) SD Second ~ z (14) SD Lax .t 6.05 2.69 4.81 1.71 0.073 0.102 86.71 111.44 85.68 28.76 16.61 34.10 68.57 110.71 66.79 30.85 16.39 31. 23 0.667 1.000 0.763 0.035 0.881 0.058 11.16 14.79 72.81 35.89 26.30 10.48 44.79 9.93 10.82 11.85 0.21 15.48 6.85 10.02 15.90 20.45 19.15 11.91 17.11 10.52 11.58 10.26 1. 30 12.22 12.86 8.21 73.21 38.21 28.21 5.00 41.07 11.43 8.93 15.36 0.00 14.64 8.71 6.39 16.01 24.93 19.87 7.07 18.83 12.31 9.44 12.16 0.00 12.00 0.201 0.074 0.896 0.288 0.786 0.040 0.579 0.392 0.420 0.356 1.000 0.420 0.021 0.931 0.708 0.734 0.027 0.465 0.637 0.567 0.259 0.182 0.815 23.32 5.96 22.29 7.01 0.377 0.567 1. 66 1. 63 1. 50 1. 22 0.247 0.733 �116 Table 26. Habitat characteristics at observed use and available locations of greater prairie-chickens in northeastern Colorado during late summer (1 Ju1 15 Aug), 1986-88. Observed Habitat cm Height, cm Sand sagebrush Grass Forbs Cover, % Bare Sand sagebrush Grass Cool season grass Needle-and-thread Sixweeks annual fescue Warm season grass Blue grama Sand dropseed Prairie sandreed Corn Forbs richness, Slope, degrees Available SD variable Height-density-index, Species (36) n (36) SD Lax 4.97 4.35 3.81 3.17 0.064 0.200 53.06 101. 53 65.00 30.57 25.43 30.52 46.11 96.81 57.08 28.84 23.43 29.94 0.732 0.629 0.910 0.325 0.415 0.270 14.31 10.14 73.19 32.36 22.64 10.69 50.83 13.75 7.08 14.31 1. 81 13.61 8.71 10.72 15.91 18.11 16.28 10.01 16.01 9.44 7.21 10.63 5.23 10.32 18.47 6.81 71.25 30.83 23.33 6.11 48.47 14.72 8.33 12.64 0.56 12.78 12.41 7.09 18.84 18.34 17.77 7.08 18.39 12.70 8.78 11.31 1.59 12.79 0.040 0.016 0.322 0.939 0.607 0.044 0.416 0.083 0.247 0.719 0.000 0.211 0.104 0.125 0.638 0.723 0.863 0.029 0.563 0.714 0.511 0.522 0.178 0.762 21. 06 5.04 19.00 4.70 0.687 0.078 2.67 2.67 3.03 3.31 0.212 0.612 �117 Table 27. Habitat characteristics at locations of radio-marked male and in northeastern Colorado during late summer female greater prairie-chickens Jul - 15 Aug), 1986-88. Males Habitat variable ~ Height-density-index, cm Height, cm Sand sagebrush Grass Forbs Cover, % Bare Sand sagebrush Grass Cool season grass Need1e-and-thread Sixweeks annual fescue Warm season grass Blue grama Sand dropseed Prairie sandreed Corn Forbs Species richness, Slope, degrees n (18) SD Females R (18) (1 f SD Emax .t 4.96 5.40 4.98 3.14 0.032 0.991 53.89 108.33 7l. 94 36.56 29.61 32.46 52.22 94.72 58.06 24.21 18.90 27.61 0.098 0.073 0.512 0.873 0.109 0.176 13.61 7.50 75.00 34.44 23.61 10.28 5l.11 13.33 5.56 12.50 l. 39 14.44 8.54 7.52 14.45 18.78 17.64 10.36 16.41 10.43 5.39 10.33 4.13 11.23 15.00 12.78 71.39 30.28 2l. 67 11.11 50.56 14.17 8.61 16.11 2.22 12.78 9.07 12.86 17.47 17.70 15.24 9.93 16.08 8.62 8.54 10.92 6.24 9.58 0.806 0.033 0.442 0.810 0.554 0.866 0.934 0.439 0.066 0.820 0.099 0.520 0.639 0.144 0.504 0.498 0.726 0.807 0.919 0.795 0.208 0.315 0.640 0.635 2l. 33 4.41 20.78 5.71 0.293 0.746 2.61 2.91 2.72 2.49 0.527 0.903 �118 Table 28. Habitat characteristics at locations of radio-marked greater prairie-chickens in northeastern Colorado during morning, mid-day, and evening during late summer (1 Ju1 - 15 Aug), 1986-88. None of the differences was significant. Morning Habitat cm Height, cm Sand sagebrush Grass Forbs Cover, % Bare Sand sagebrush Grass Cool season grass Need1e-and-thread Sixweeks annual fescue Warm season grass Blue grama Sand dropseed Prairie sandreed Corn Forbs richness, Slope, degrees Afternoon SD variable Height-density-index, Species (12) n (12) Evening (12) so SD 4.31 2.51 6.28 6.35 4.32 3.28 54.58 102.50 57.50 32.71 19.13 25.54 54.17 110.42 76.67 34.10 34.14 31. 79 50.42 91. 67 60.83 27.01 18.26 32.74 15.00 12.50 72.50 32.92 25.00 10.00 51. 25 14.58 7.50 14.58 3.33 14.17 10.44 12.15 17.52 14.37 16.10 9.77 17.60 7.22 8.39 12.33 7.49 11.25 12.92 11.67 72.08 30.83 21.25 7.50 49.17 14.58 7.08 13.75 1. 25 13.75 6.89 13.03 19.24 19.75 17.73 8.39 16.35 12.33 8.38 9.80 4.33 8.56 15.00 6.25 75.00 33.33 21. 67 14.58 52.08 12.08 6.67 14.58 0.83 12.92 9.05 4.83 11.08 21. 03 16.14 11.17 15.29 8.65 4.92 10.54 2.89 11.77 19.08 4.23 22.33 5.16 21. 75 5.45 1. 92 1. 31 2.42 2.07 3.67 3.87 �119 Table 29. Habitat characteristics at observed use and available locations of greater prairie-chickens in northeastern Colorado during autumn (16 Aug - 31 Oct), 1986-88. Observed (36) Habitat variable R Available (36) £ SD R SD 8.05 14.09 15.03 25.05 0.001 0.151 48.19 100.69 58.19 29.23 34.97 32.27 29.31 115.97 54.72 29.62 55.93 35.13 0.938 0.007 0.619 0.008 0.170 0.664 Cover, % Bare Sand sagebrush Grass Cool season grass Need1e-and-thread Sixweeks annual fescue Warm season grass Blue grama Sand dropseed Prairie sandreed Corn Forbs 17.36 8.47 71.39 2l.81 12.08 5.14 58.89 10.97 15.97 11.67 2.22 14.31 10.92 10.27 16.72 16.87 13.70 7.51 14.69 13.30 14.87 13.31 10.45 12.77 20.83 3.06 72 .22 17.08 15.42 2.78 6l.25 12.22 10.56 8.61 12.36 10.00 11.98 4.97 15.97 18.80 17.94 5.40 16.96 17.17 12.47 13.23 24.48 10.14 0.586 0.000 0.787 0.525 0.115 0.055 0.400 0.135 0.301 0.974 0.000 0.178 0.203 0.006 0.829 0.266 0.379 0.130 0.530 0.731 0.098 0.332 0.027 0.118 Species richness, n 17.83 6.10 14.14 7.69 0.175 0.027 2.58 3.73 2.44 3.31 0.481 0.868 Height-density-index, cm Height, cm Sand sagebrush Grass Forbs Slope, degrees Emu ! �120 Table 30. Habitat characteristics at locations of radio-marked male and female greater prairie-chickens in northeastern Colorado during autumn (16 Aug - 31 Oct). 1986-88. Males Habitat variable (18) Females £nax 4.19 2.04 0.000 0.109 32.19 44.07 37.51 47.22 98.61 56.11 26.86 23.81 26.98 0.463 0.015 0.185 0.845 0.727 0.704 18.06 8.06 71. 39 16.67 6.39 3.61 60.83 10.83 16.39 9.44 4.44 12.22 13 .19 9.72 17.64 16.18 8.88 7.44 14.48 13.75 17.72 13.38 14.64 11.27 16.67 8.89 71. 39 26.94 17.78 6.67 56.94 11.11 15.56 13.89 0.00 16.39 8.40 11.06 16.25 16.37 15.46 7.48 15.06 13.23 11.87 13.23 0.00 14.12 0.071 0.602 0.739 0.962 0.028 0.983 0.872 0.878 0.108 0.964 0.362 0.709 0.812 1.000 0.067 0.012 0.227 0.435 0.951 0.869 0.324 0.215 0.335 16.39 6.74 19.27 5.17 0.283 0.159 1. 67 1. 78 3.50 4.87 0.000 0.148 SO ~ 11.92 19.32 49.17 102.78 60.28 Cover, % Bare Sand sagebrush Grass Cool season grass Need1e-and-thread Sixweeks annual fescue Warm season grass Blue grama Sand dropseed Prairie sandreed Corn Forbs Species cm Height. cm Sand sagebrush Grass Forbs richness, Slope. degrees II f so ~ Height-density-index. (18) ~ �121 Table 31. Habitat characteristics at locations of radio-marked greater prairie-chickens in northeastern Colorado during morning, mid-day, and evening during autumn (16 Aug - 31 Oct), 1986-88. None of the differences was significant. Morning (12) so Habitat variable , Evening (12) so so 3.17 2.32 13.88 22.50 7.11 7.51 Height, cm Sand sagebrush Grass Forbs 35.42 92.08 43.75 24.35 25.54 25.77 53.33 118.33 67.92 28.39 45.14 35.45 55.83 91.67 62.92 32.39 26.31 32.22 Cover, % Bare Sand sagebrush Grass Cool season grass Needle-and-thread Sixweeks annual fescue Warm season grass Blue grama . Sand dropseed Prairie sandreed Corn Forbs 17.08 7.08 70.42 21.25 17.92 4.17 56.67 11.25 16.67 16.25 0.00 18.33 13.73 10.10 18.27 17.47 18.02 5.15 14.51 14.48 10.94 15.54 0.00 14.67 20.83 7.92 70.00 23.33 7.50 5.42 57.08 10.00 10.42 10.00 6.67 10.83 9.73 11.17 15.81 18.38 10.55 8.65 14.22 14.30 14.53 8.53 17.75 11.84 14.17 10.42 73.75 20.83 10.83 5.83 62.92 11.67 20.83 8.75 0.00 13.75 8.48 10.10 17.21 16.07 9.96 8.75 15.73 12.12 17.69 14.64 0.00 11.51 17.92 7.62 18.83 5.44 16.75 5.31 2.33 2.84 3.67 5.61 1. 75 1.60 Height-density-index, I Afternoon (12) Species richness, Slope, degrees n cm �122 Table 32. Habitat characteristics at locations of radio-marked female greater prairie-chickens with and without broods in northeastern Colorado during early summer - autumn (16 May - 31 Oct). 1986-88. Brood (48) Habitat variable ~ SD No brood (56) R ~ SD Lax t 4.78 2.69 3.54 1.91 0.015 0.009 69.17 103.33 74.17 28.48 23.60 26.74 52.95 88.93 55.18 34.62 25.93 32.61 0.172 0.510 0.166 O.Oll 0.004 0.002 Cover. % Bare Sand sagebrush Grass Cool season grass Needle-and-thread Sixweeks annual fescue Warm season grass Blue grama Sand dropseed Prairie sandreed Corn Forbs 14.69 10.63 73.33 29.79 18.23 9.27 53.75 11.15 12.81 11.15 0.83 15.94 8.72 10.70 16.22 21.88 16.87 10.96 17.67 12.04 ll.80 10.22 3.90 11.97 18.04 8.04 69.46 29.38 19.46 7.41 49.02 11.25 9.55 11.96 5.27 13.57 12.53 9.37 18.77 20.25 19.13 9.63 18.35 12.87 11.41 13.54 17.04 17.70 0.012 0.344 0.307 0.578 0.378 0.353 0.795 0.642 0.807 0.050 0.000 0.007 o .ll3 Species richness. n 22.25 5.38 17.82 7.21 0.041 0.001 1 2.38 2.03 2.25 3.06 0.005 0.804 I Height-density-index. cm Height. cm Sand sagebrush Grass Forbs Slope. degrees 0.191 0.268 0.920 0.730 0.359 0.185 0.-966 0.156 0.732 0.063 0.421 �123 Table 33. Habitat at observed and available locations of greater prairie-chickens in northeastern Colorado during winter (1 Nov - 14 Feb) , 1986-88. Observed (36) Habitat variable Height-density-index, cm R SD Available (36) ~ f SD Lax .£ 9.63 6.46 13.52 8.56 0.100 0.033 Height, cm Sand sagebrush Grass Forbs 36.25 89.17 44.31 41.99 35.81 46.15 39.17 89.03 52.64 43.40 34.76 48.68 0.847 0.862 0.755 0.773 0.987 0.459 Cover, % Bare Sand sagebrush Grass Cool season grass Need1e-and-thread Sixweeks annual fescue Warm season grass Blue grama Sand dropseed Prairie sandreed Corn Forbs 67.36 5.14 27.50 2.22 2.22 0.00 25.69 2.36 5.42 2.78 10.00 1.39 21.46 9.06 20.96 4.99 4.99 21.97 8.02 22.46 10.24 10.24 0.891 0.473 0.685 0.000 0.000 18.37 6.92 12.09 8.57 12.31 3.71 70.28 5.00 24.31 2.92 2.92 0.00 21.67 1.81 4.44 2.08 9.86 1.53 19.49 7.09 7.15 4.53 14.37 4.44 0.727 0.887 0.003 0.000 0.364 0.288 0.571 0.945 0.535 0.716 0.716 1.000 0.370 0.737 0.680 0.669 0.965 0.886 Species richness, 11 6.00 5.40 5.97 4.94 0.600 0.982 Slope, degrees 1.89 1.67 3.08 9.84 0.000 0.477 �124 Table 34. Habitat at locations prairie-chickens in northeastern 1986-88. of radio-marked male and female greater Colorado during winter (1 Nov - 14 Feb), Males Habitat variable ~ Height-density-index, SD Females & (18) f SD Lax ~ 10.91 7.25 8.36 5.47 0.255 0.241 Height, cm Sand sagebrush Grass Forbs 38.61 88.89 40.83 46.58 42.06 46.03 33.89 89.44 47.78 38.06 29.50 47.35 0.413 0.154 0.908 0.741 0.964 0.658 Cover, % Bare Sand sagebrush Grass Cool season grass Need1e-and-thread Sixweeks annual fescue Warm season grass Blue grama Sand dropseed Prairie sandreed Corn Forbs 71.67 6.94 21. 39 1.11 1.11 0.00 20.56 1. 67 2.50 1.11 10.28 0.56 16.09 11.39 16.87 3.66 3.66 25.50 5.69 23.25 5.94 5.94 0.066 0.007 0.196 0.053 0.053 15.52 7.07 4.29 2.74 12.18 1.62 63.06 3.33 33.61 3.33 3.33 0.00 30.83 3.06 8.33 4.44 9.72 2.22 19.94 6.89 16.27 11.74 12.77 4.92 0.311 0.916 0.000 0.000 0.848 0.000 0.234 0.240 0.080 0.186 0.186 1.000 0.094 0.555 0.157 0.256 0.895 0.187 4.94 4.83 7.06 5.87 0.432 0.247 1. 39 1.46 2.39 1. 75 0.459 0.072 Species richness, Slope, degrees I! em (18) I I �125 Table 35. Habitat at locations of radio-marked greater prairie-chickens in northeastern Colorado during morning, afternoon, and evening periods during winter (1 Nov - 14 Feb) 1986-88. Differences (f < 0.05) between means for each habitat variable are indicated by different superscripts. Morning Habitat variable (12) Afternoon SD Height-density-index, cm 3.36 (12) SD 10.42a,b 6.06 41.25 36.69 50.07 25.83a 72. SOb 33.33a 41. 50 30.64 41. 03 32.76 10.94 3l. 65 76.25a 5.428 19.178 1.25a l. 25a 8.01 9.88 9.25 2.26 2.26 43.66 30.46 47.70 Cover, % Bare Sand sagebrush Grass Cool season grass Need1e-and-thread Sixweeks annual fescue Warm season grass Blue grama Sand dropseed Prairie sandreed Corn Forbs 69.588 3.338 26.678 0.42a o .42a 0.008 26.258 0.42a 5.00a 0.42a 12.08a,b 0.42a,b 10.10 6.15 11.93 l.44 l.44 5.088 4.58 6.45 1.008 l.21 l. 62 Species richness, Slope, degrees n 56.25a 6.6]8 36.67a 5.00a 5.008 0.008 32.92a 6.67a 10.42a.b 6.67a 3.33a 3.75a SD 12. SOb 34.58a 83.75a.b 42.50a 48.33a 111.25a 57.08a (12) 7.78 Height, cm Sand sagebrush Grass Forbs 11.89 l.44 6.03 l.44 13.39 l.44 Evening i.69 7.69 O.OOa 26.58 10.94 19.36 14.20 7.49 5.69 17.92a 10.10 O.OOa 0.83b l. 25a 14.58b 1.95 2.26 13 .05 O.OOb 4.17 2.258,b 1. 86 �.~ HABITAT TYPES N ~ ACHCU.1UE tSd NH.WA'OIIS • 1HOIrr_'" m m UOfIItQn'-UD 0IIA8I lIT] till] o o - - - o o 0MI8 1tICIII'-TAU. 0RASa IIN£JlIHtWT-TAU.~ foD.TAU. GRAIl ~TAU.OItASS CXUffY ROADS SCALE IN MLES 1 1 2 3 234 SCALE IN KLOMETERS Fig. 10. 1986-88. Distribution of 10 habitat types in greater prairie-chicken use areas in northeastern Colorado, I--' N (J\ �_-- ....-- N HABIT A T TYPES II TREESIRESDENTIAL AGRICU. TlIE ~ ~ • AJlNJAlIFOFB SHORT -M> (JtASS SAGElstK>RT -M) GRASS '·.iiii.iiil SHORT-TALL mASS I':::.H SAGE/SHORT-TALL k> I ~ r ::: I GRASS TALL GRASS SAGEJUD- TAU. GRASS - - - - COlHTY ROADS SCALE IN MILES 012 I I o 1 I I I 2 3 I SCALE N KLOMETERS Fig. 11. Distribution of 45 specific (delineated by number) and 10 general habitat quarter of a greater prairie-chicken study area in northeastern Colorado, 1986-88. types in the northwestern I-' N --..J �N HABITAT TYPES TREESIRESDENTIAL ~ AGRIClJ._TlIE ~ MNJALlFOfB SHORT -M) ~ss I~~:~!:H SAaElSOORT oM) GRASS SHORT-TAU. GRASS mm'=:j SAOEISHORT-TAU. MASS I))j til>- TAU. GRASS j: : : : I - - - - SAGEJt.4D- TALL GRASS COl.MY ROADS SCALENMLES 012 I o I 1 I I I 2 3 I SCALE IN KLOMETERS Pig. 12. Distribution of 45 specific (delineated by number) and 10 general hAhitat quarter of a greater prairie-cllicken study area in northeastern Colorado, 1986-88. ---~ types in the southwestern t-' tv oo �---.~ HABITAT TYPES N II TREESlFlESIemAL AGRIClL 1\R: ~ MNJAL.JFOfB ~ 1': : SHORT -M> GRASS •• n u n L.J r c· )jd - - - - SAOEISHORT-MD SHORT-TAll. GRASS GRASS SAGE/SHORT - TAll. GRASS MD- TAll. QRASS SAGEI~ TALL GRASS COlNTY ROADS SCALE IN MLES 012 I I o 1 I I I 2 3 I SCALE IN KLOMETERS Fig. 13. Distribution of 45 specific (delineated by number) and 10 general habitat quarter of a greater prairie-chicken study area in northeastern Colorado, 1986-88. types in the southeastern I-' N \0 �N HABIT AT TYPES II TREWRESOOffiAL ~ AGiiClLME ~ NNlAUFOfB • SHORT-Wl GRASS SAOEIstDU-MD SHORT-TMJ. aRASS 1:'::::':1 SAGE/SHOOT-TMJ. 1\/1 w}-TALLGRASS 1 - - - - o GRASS 11:1:11111:I'j I: :: I-' W GRASS SAGEItJO-TAll. GRASS COLmY ROADS SCALE t.J MLES 012 I I o 1 I I I 2 3 I SCALE IN KILOMETERS Fig. 14. Distribution of 45 specific (delineated by number) and 10 general habitat quarter of a greater prairie-chicken study area in northeastern Colorado, 1986-88. types in the northeastern �131 Table 36. Area (%) of different habitat types on a 301 km2 greater prairiechicken study area in northeastern Colorado. Categorya Habitat type Trees/residential 35 Trees/residential 0.98 0.98 Corn Wheat Alfalfa Millet Soybean Rye Sorghum Sudan 13.25 8.98 0.94 0.23 1.95 0.48 0.05 0.07 0.60 Annuals Hairy goldenaster 1.83 1.78 0.05 Agriculture 1 3 4 6 7 8 9 10 Annual/forb 5 29 Short-mid 2 11 34 37 grass Short-mid warm season grasses/annuals Smooth brome/orchardgrass Cool and short-mid warm season grasses/annuals Psoralea/short-mid warm season grasses/annuals Sage/short-mid grass 12 Sand sagebrush/cool and short-mid warm season grasses/annuals 13 Sand sagebrush/cool and short-mid warm season grasses 15 Sand sagebrush/cool and short-mid warm season grasses/muhly 16 Sand sagebrush/Yucca/cool and short-mid warm season grasses 30 Sand sagebrush/Psoralea/cool and short-mid warm season grasses 43 Yucca/cool and short-mid warm season grasses Short-tall 39 40 41 42 Area grass Cool and short-tall warm season grasses Psoralea/short-tall warm season grasses Short-tall warm season grasses Short-tall warm season grasses/muhly 4.58 2.67 0.76 1.06 0.10 26.36 9.87 7.23 6.14 0.37 2.61 0.15 2.80 1.62 0.17 0.84 0.16 Sage/short-tall grass 33.05 14 Sand sagebrush/cool and short-tall warm season grasses 5.74 18 Sand sagebrush/Psoralea/cool and short-tall warm season grasses/muhly 0.55 21 Sand sagebrush/short-tall warm season grasses 12.90 22 Sand sagebrush/short-tall warm season grasses/muhly 7.91 23 Sand sagebrush/Yucca/cool and short-tall warm season grasses/annuals 0.29 25 Sand sagebrush/Yucca/cool and short-tall warm season grasses/muhly 1.09 28 Sand sagebrush/Yucca/short-tall warm season grasses 0.73 31 Sand sagebrush/Yucca/short-tall warm season grasses/muhly 0.03 38 Sand sagebrush/cool and short-tall warm season grasses/muhly 3.77 44 Yucca/cool and short-tall warm season grasses 0.05 �132 Table 36. Continued. Categorya Mid-tall grass 19 20 33 36 Habitat type Mid-tall warm season grasses Mid-tall warm season grasses/muhly/annuals Cool and mid-tall warm season grasses/annuals Psoralea/mid-tall warm season grasses/muhly Sage/mid-tall grass 17 Sand sagebrush/mid-tall warm season grasses/annuals 24 Sand sagebrush/mid-tall warm season grasses 26 Sand sagebrush/Yucca/mid-tall warm season grasses 27 Sand sagebrush/Yucca/mid-tall warm season grasses/muhly 32 Yucca/mid-tall warm season grasses County roads 45 County roads Area 5.27 4.73 0.13 0.21 0.20 11.34 4.01 5.79 0.9l 0.55 0.08 0.53 0.53 aCeneral categories of habitat types are referenced on maps by both name and number. I �133 Table 37. Common plant species on a greater prairie-chicken study area in northeastern Colorado, 1986-88. Undisturbed Disturbed Agriculturea RangelanJ' Sandhill C Relatively flatd smooth brome Cheat grass brome Stinkgrass Orchardgrass \lheat Quackgrass Crested wheatgrass Bottlebrush squirreltail Foxtail barley Rye \Iild oat Sand dropseed Needle-and-thread Showy chloris lIindmillgrass False buffalograss \litchgrass panicum COlT1llOn barnyardgrass Yellow bristlegrass Green bristlegrass Foxtai l bristlegrass Mi llet Sorghum Sudan Mat sandbur Corn COlT1llOn Russianthistle Fireweed summercypress Lambsquarters goosefoot Sand amaranth Redroot amaranth Common purslane Prairie pepperweed Alfalfa \lhite clover Yellow sweetclover IIhite sweetclover Soybean COIT1llOn puncturevine Leafy spurge COlT1llOn mil kweed Sand mi lkweed Bush morningglory Bracted verbena Buffalobur nightshade Canada horseweed Prairie sunflower \lestern ragweed Bull thistle Smooth brome Cheatgrass brome Kentucky bluegrass Stinkgrass Purple lovegrass Blowout grass Quackgrass Crested wheatgrass IIild oat Sand dropseed Red threeawn lIindmillgrass False buffalograss lIitchgrass panicum Mat sandbur lIillow Chinese elm Veiny dock COIT1llOn Russianthistle Fireweed sUlll1lercypress Lambsquarters goosefoot Sand amaranth Sandverbena COIT1llOn purslane Intermediate pricklepoppy Roughseed clalTlllYweed Rocky Mountain beeplant Flixweed tansymustard Pinnate tansymustard Sl imflower scurfpea American vetch Lambert crazyweed Common puncturevine Texas croton Fendler euphorbia Scarlet globemallow Russian-olive Stemless cymopterus Sand milkweed Bush morningglory Bracted verbena American nightshade Buffalobur nightshade Buffalogourd Cheatgrass brome Sixweeks annual fescue Blowout grass Prairie junegrass Prairie sandreed Sandhill I1'lJhly Sand dropseed Indian ricegrass Needle-and-thread Red threeawn Blue grama Hairy grama Switchgrass panicum Sand bluestem Little bluestem Sand sedge Needle spikerush Prairie dayflower Prairie spiderwort Sma II soapweed Prairie onion Annual eriogonum Veiny dock Fireweed sl.lTlllercypress Narrow-leaved goosefoot Slender snakecotton Sandverbena Plains larkspur Silvery bladderpod Arkansas rose Bushpea IIhite prairieclover Silky prairieclover Purple prairieclover Lead plant Silverleaf scurfpea Lemon scurf pea American vetch Missouri milkvetch Fendler euphorbia Tenpetal mentzelia Brittle pricklypear COIT1llOn pricklypear Ballcactus Scarlet gaura Shrubby eveningprimrose Prairie eveningprimrose Stemless cymopterus Siberian dogbane Plains milkweed Sand mi lkweed Nuttall evolvulus Cheatgrass brome Sixweeks annual fescue Sand lovegrass Prairie sandreed Sandhill muhly Sand dropseed Needle-and· thread Red threeawn Blue grama Sideoats grama Sand paspalum Switchgrass panicum Big bluestem Sand bluestem Little bluestem Yellow Indiangrass Sand sedge Needle spikerush Prairie spiderwort Sma II soapweed Prairie onion Annual eriogonum Fireweed sUlll1lercypress Narrow-leaved goosefoot Prairie flameflower Plains larkspur Silvery bladderpod Arkansas rose Silky prairieclover Purple prairieclover Lead plant Lemon scurfpea Sl imflower scurfpea Lambert crazyweed Missouri milkvetch Groundplum milkvetch Texas croton Tenpetal mentzelia Brittle pricklypear COIT1llOn pricklypear Scarlet gaura Shrubby eveningprimrose Prairie eveningprimrose Stemless cymopterus Plains milkweed Nuttall evolvulus Hood phlox IIhite borage ClalTlllY groundcherry Narrow-leaved �134 Table 37. Continued. Undisturbed Disturbed Agriculturea Canada thistle Platte thistle Prickly lettuce Common dandelion a1ncluding blncluding c1ncluding dlncluding 44. habitats habitats habitats habitats Rangelandb Sandhi IIC Hairy goldenaster Canada horseweed Louisiana sagewort Prairie sunflower Western ragweed Platte thistle Prickly lettuce Rush skeletonplant Pink-throated morningglory Hood phlox Hairy puccoon White borage Clammy groundcherry Narrow-leaved penstemon Gilia penstemon Downy painted-cup Louisiana broomrape Woolly plantain Dotted gayfeather Missouri goldenrod Ironplant goldenweed Spreading fleabane Sand sagebrush Common sagewort Hooker palafoxia Nippleweed Prairie sunflower Western ragweed Relatively flatd penstemon Gil ia penstemon Woolly plantain Dotted gayfeather Blazing star Hairy goldenaster Missouri goldenrod Ironplant goldenweed Spreading fleabane Louisiana sagewort Sand sagebrush Old plainsman Prairie coneflower Nippleweed Prairie sunflower Western ragweed Common pearlyeverlasting Platte thistle Yellow salsify Wavyleaf microseris 1, 3, 4, 6, 7, 8, 9, 10, 11, and 45. 2, 5, 29, 34, 35, 37, and 45. 12, 13, 14, 15, 17, 18, 20, 21, 22, 24, 25, 26, 27, 28, 31, 32, 36, 38, and 39. 12, 13, 14, 16, 17, 18, 19, 21, 22, 23, 24, 26, 28, 30, 33, 38, 40, 41, 42, 43, and I I �135 Trees/residential: Chinese elm, Russian-olive, willow, and Kentucky bluegrass frequently occurred in this habitat type. Several additional species were common including needle-and-thread, sand dropseed, cheatgrass brome, sand lovegrass, sand sagebrush, blue grama, prairie sandreed, switchgrass panicum, and crested wheatgrass. These habitats were relatively moist, shady, and disturbed. Trees/residential (35): Same as above description. Agriculture: Agriculture was common throughout the study area, especially on relatively level sites conducive to center-pivot irrigation. Corn was the most common crop on irrigated land while millet was most common on non-irrigated land. Wheat, rye, sorghum, sudan, alfalfa, and soybeans were planted, but in smaller quantities. Other plants included witchgrass panicum, green bristlegrass, mat sandbur, redroot amaranth, and common dandelion. None of the crops provided nesting cover, but most provided grain, forage and/or cover in autumn and winter. Corn (1): Corn was commonly planted in circular fields with center-pivot irrigation in mid to late May and harvested in October. Fields were left unplowed throughout the winter and corn stubble and kernels were abundant. During the peak of the growing season (Jul), corn reached a height of 2.5 m and provided 95-100% canopy cover. The understory, particularly along the edges of fields, consisted'of witchgrass panicum, green bristlegrass, mat sandbur, redroot amaranth, and common dandelion. I I Wheat (3): Wheat was planted in center-pivot irrigated fields in August or September. It was frequently grazed during autumn and winter (0.1 m in height at the time) and harvested in June. Wheat reached a height of about 1 m during May and June (year following planting) , but generally provided only 75-85% canopy cover. Additional species included witchgrass panicum, green bristlegrass, mat sandbur, redroot amaranth, and common dandelion. Alfalfa (4): Alfalfa was planted in center-pivot irrigated fields, cut periodically, and was generally available all year. Alfalfa ranged in height from 0.1 to 0.8 m and provided 80-90% canopy cover. Additional species frequently include witchgrass panicum, green bristlegrass, mat sandbur, redroot amaranth, and common dandelion. Millet, was commonly planted as a cover crop with alfalfa to increase alfalfa seedling growth (usually planted in May). Millet (6): Millet was planted in non-irrigated fields in May and harvested in August or September. Stubble and waste millet were abundant throughout the winter. Millet reached a height of 0.5 m, but generally provided only 75-85% canopy cover. Additional species included witchgrass panicum, green bristlegrass, mat sandbur, redroot amaranth, and common dandelion. Alfalfa was often grown as an understory plant with millet in center-pivot irrigated fields (millet was not harvested in the latter case). Soybean (7): Center-pivot irrigated May and harvested in September. fields were planted with soybeans Virtually no stubble or waste in �136 remained throughout the winter. Soybeans reached a height of 1 m and provide 95-100% canopy cover. The understory, particularly on the edges of fields, consisted of witchgrass panicum, green bristlegrass, mat sandbur, redroot amaranth, common dandelion, and corn (planted in earlier rotations). Rye (8): Rye was planted in center-pivot irrigated fields in August or September. Rye was often grazed during autumn and winter (about 0.1 m in height at the time) and harvested in June (not always harvested). Rye reached a height of 1 m during May and June (year following planting), but generally provided only 80-90% canopy cover. Additional species included witchgrass panicum, green bristlegrass, mat sandbur, redroot amaranth, and common dandelion. Sorghum (9): Sorghum was planted (usually in non-irrigated fields) in May and harvested in August or September. Stubble and waste sorghum were abundant throughout the winter. Sorghum reached a height of 1 m, but generally provided only 75-85% canopy cover. The understory consisted of witchgrass panicum, green bristlegrass, mat sandbur, redroot amaranth, common dandelion, and millet (millet may be mixed with sorghum seeds). Sudan (10): Sudan, a hybridized version of sorghum and millet, was occasionally planted in center-pivot irrigated fi~lds. Sudan tolerated dry conditions and was commonly planted around the edges of irrigated crop fields. It was usually planted in May and grazed during late autumn or winter when the adjacent crop stubble was also grazed. Although stubble was abundant throughout the winter, waste sudan was readily eaten by cattle and generally not available for wildlife. Sudan reached a height of 2 m and provided about 85-95% canopy cover. The understory consisted of witchgrass panicum, green bristlegrass, mat sandbur, redroot amaranth, common dandelion, and millet (millet may be mixed with sudan seeds). Annual/forb: This habitat was dominated by annual and/or non-grass species; it was typical of disturbed sites on the study area. Most of this habitat was either around windmills or in fallow center-pivot irrigated fields. A variety of annual species dominated this type including cheatgrass brome, wheat, witchgrass panicum, green bristlegrass, millet, mat sandbur, common Russianthistle, lambsquarters goosefoot, sand amaranth, redroot amaranth, Rocky Mountain beeplant, pinnate tansymustard, Texas croton, bracted verbena, hairy goldenaster, prairie sunflower, and prickly lettuce. The variety of dominant plants was largely a response to variables such as moisture condition, disturbance, and previous vegetation. Annuals (5): Most of this habitat was in fallow center-pivot irrigated fields. Several species dominated including cheatgrass brome, wheat, witchgrass panicum, green bristlegrass, millet, mat sandbur, common Russianthistle, lambsquarters goosefoot, sand amaranth, redroot amaranth, Rocky Mountain beeplant, pinnate tansymustard, Texas croton, bracted verbena, prairie sunflower, and prickly lettuce. The variety of dominant plants was largely a response to variables including moisture, disturbance, and previous vegetation. For �137 example, in relatively moist and disturbed areas around windmills, bracted verbena (Jun-Jul), pinnate tansymustard (Jun), Rocky Mountain beeplant (Jul-Aug), cheatgrass brome (May), and lambsquarters goosefoot (Jul) dominated. Drier disturbed areas frequently had Texas croton and sand amaranth as dominant species. Hairy goldenaster (29): Hairy goldenaster was a perennial forb that dominated 1 relatively dry disturbed site on the study area. Few other species were present. Short-mid grass: This habitat was typical of disturbed and/or heavily overgrazed areas. The orchardgrass/smooth brome area was an exception, likely because it was specifically planted to be heavily grazed. Disturbance, such as previous agriculture, spraying, and/or fire, had eliminated sand sagebrush from this habitat. Sand lovegrass, sand dropseed, red threeawn, blue grama, and sand paspalum dominated (except in planted areas). Additional species included cheatgrass brome, needleand-thread, green brist1egrass, mat sandbur, common Russianthistle, lambsquarters goosefoot, redroot amaranth, Rocky Mountain beeplant, slimflower scurfpea, Texas croton, stemless cymopterus, bracted verbena, buffalobur nightshade, Louisiana sagewort, prairie sunflower, and western ragweed. This habitat provided essentially no cover for prairiechickens. " Short-mid warm season grasses/annuals (2): This habitat was typical of disturbed and/or heavily over-grazed areas. Sand lovegrass, sand dropseed, red threeawn, blue grama, and/or sand paspalum dominated. Rarely were all 5 species in the same location. Numerous species of annual grasses and forbs were common including cheatgrass brome, lambsquarters goosefoot, Rocky Mountain beeplant, Texas croton, stemless cymopterus, bracted verbena, Russian thistle, prairie sunflower, and western ragweed. Orchardgrass/smooth brome (11): This habitat occurred only in centerpivot irrigated fields. It was used entirely for May-October cattle grazing. Smooth brome and orchardgrass co-dominated and provided vegetation throughout the grazing season. The vegetation was typically less than 0.1 m in height. Additional common species included green bristlegrass, mat sandbur, redroot amaranth, and buffalobur nightshade. Cool and short-mid warm season grasses/annuals (34): This type was typical of disturbed and/or heavily over-grazed areas. It was dominated by sand lovegrass, quackgrass, sand dropseed, needle-andthread, red threeawn, blue grama, and/or sand paspalum. Rarely were all 5 species in the same location. In particular, quackgrass was localized. In addition to perennial grasses, numerous species of forbs and annual grasses were common including cheatgrass brome, common Russianthistle, Texas croton, prairie sunflower, and western ragweed. Psoralea/short-mid warm season grasses/annuals (37): Cheatgrass brome, blue grama, slimflower scurfpea, and Louisiana sagewort dominated this overgrazed habitat type. Additional species such as brittle �138 pricklypear, common pricklypear, and Platte thistle were also common. Although slimflower scurfpea grew to 1 m, it provided little cover. Grasses rarely reached a height of 0.1 m. Sage/short-mid grass: Needle-and-thread, blue grama, and sand sagebrush dominated this heavily overgrazed habitat type. Additional dominant species frequently included sand lovegrass, sandhill muhly, sand dropseed, hairy grama, sand paspalum, little bluestem, small soapweed, and slimflower scurfpea. Other common plants included sixweeks annual fescue, sand sedge, annual eriogonum, brittle pricklypear, common pricklypear, woolly plantain, prairie sunflower, western ragweed, and Platte thistle. This habitat provided little cover, except in the immediate vicinity of sand sagebrush or small soapweed. Sand sagebrush/cool and short-mid warm season grasses/annuals (12): Sand lovegrass, sand dropseed, needle-and-thread, blue grama, sand paspalum, and sand sagebrush dominated this habitat. Numerous other species were common, including sixweeks annual fescue, sand sedge, annual eriogonum, Texas croton, brittle pricklypear, common pricklypear, woolly plantain, prairie sunflower, western ragweed, and Platte thistle. This type provided little cover, since most of the perennial grasses had foliage that rarely reached a height of 0.1 m. , Sand sagebrush/cool and short-mid warm season grasses (13): This habitat type was dominated by quackgrass, needle-and-thread, blue grama, and sand sagebrush. Numerous other species were common, including sixweeks annual fescue, quackgrass, annual eriogonum, brittle pricklypear, common pricklypear, woolly plantain, and Platte thistle. Sand sagebrush/cool and short-mid warm season grasses/muhly (15): Sand sagebrush, sandhill muhly, sand dropseed, needle-and-thread, blue grama, hairy grama, sand paspalum, and little bluestem dominated this relatively steep habitat type. Hairy grama and sandhill muhly were most common on the steepest sites. Sixweeks annual fescue, sand sedge, annual eriogonum, brittle pricklypear, common pricklypear, woolly plantain, and western ragweed were also common. Sand sagebrush/Xucca/cool and short-mid warm season grasses (16): Sand dropseed, needle-and-thread, blue grama, small soapweed, and sand sagebrush dominated this habitat type. Additional species included sixweeks annual fescue, prairie sandreed, red threeawn, sand sedge, brittle pricklypear, common pricklypear, and woolly plantain. Sand sagebrush/Psoralea/cool and short-mid warm season grasses (30): Sand dropseed, needle-and-thread, blue grama, little bluestem, slimflower scurfpea, sand sagebrush, prairie sunflower, and western ragweed dominated in this type. Additional common species included sixweeks annual fescue, prairie sandreed, red threeawn, sand sedge, annual eriogonum, Rocky Mountain beeplant, brittle pricklypear, common pricklypear, and woolly plantain. Yucca/cool and short-mid warm season grasses (43): Sand dropseed, needle-and-thread, blue grama, and small soapweed dominated this limited habitat type. Additional species included sixweeks annual �139 fescue, pra~r~e sandreed, sand sedge, brittle pricklypear, pricklypear, and woolly plantain. common Short-tall grass: Much of this overgrazed habitat appeared to have undergone sagebrush removal, either through land use or spraying. Sand dropseed, blue grama, and little bluestem frequently dominated. Other dominant species included quackgrass, prairie sandreed, sandhill muhly, needleand-thread, hairy grama, switchgrass panicum, sand paspalum, sand bluestem, and slimflower scurfpea. Other common species included sand sedge, annual eriogonum, common pricklypear, woolly plantain, spreading fleabane, prairie sunflower, and western ragweed. Cool and short-tall warm season grasses (39): Sand lovegrass, quackgrass, prairie sandreed, sand dropseed, needle-and-thread, blue grama, hairy grama, sand paspalum, switchgrass panicum, sand bluestem, little bluestem, and/or hairy grama dominated. Because of the diversity of this habitat type, few of the species dominated in any given location, even though all were usually present. Common annuals included sixweeks annual fescue and western ragweed. Psoralea/short-tall warm season grasses (40): Blue grama, little bluestem, and slimflower scurfpea dominated on this restricted type. This habitat was relatively flat and less sandy than most sites. Other common species included sand sedge, prairie'spiderwort, annual eriogonum, common pricklypear, woolly plantain, spreading fleabane, and western ragweed. Short-tall warm season grasses (41): Prairie sandreed, needle-andthread, sand dropseed, blue grama, sand paspalum, switchgrass panicum, and/or little bluestem dominated. Common annuals included sixweeks annual fescue, prairie sunflower, and western ragweed. Short-tall warm season grasses/muhly (42): Prairie sandreed, muhly, sand dropseed, and blue grama dominated. sandhill Sage/short-tall grass: Prairie sandreed, sand dropseed, needle-and-thread, blue grama, little bluestem, and sand sagebrush (and/or small soapweed) dominated most of this habitat type. Additional common species included prairie sunflower and western ragweed on relatively flat areas and sandhill muhly, hairy grama, and lemon scurfpea on steep sites. This habitat provided cover for prairie-chickens, especially in the immediate vicinity of sand sagebrush plants. Most of this type appeared to be relatively undisturbed by past land management practices, other than moderate to heavy cattle grazing. Sand sagebrush/cool and short-tall warm season grasses (14): Prairie sandreed, sand dropseed, needle-and-thread, blue grama, little bluestem, and sand sagebrush dominated. Common annuals included cheatgrass brome, prairie sunflower, and western ragweed. Sand sagebrush/Psoralea/cool and short-tall warm season grasses/muhly (18): Prairie sandreed, sand dropseed, needle-and-thread, blue grama, little bluestem, slimflower scurfpea, and sand sagebrush dominated on relatively flat sites while prairie sandreed, sandhill �140 muhly, hairy grama, little bluestem, sagebrush dominated on steep sites. lemon scurfpea, and sand Sand sagebrush/short-tall warm season grasses (21): Sand lovegrass, prairie sandreed, sand dropseed, needle-and-thread, blue grama, little bluestem, switchgrass panicum, and sand sagebrush dominated. Common annuals included sixweeks annual fescue, prairie sunflower, and western ragweed. Sand sagebrush/short-tall warm season grasses/muhly (22): Prairie sandreed, sandhill muhly, sand dropseed, needle-and-thread, blue grama, little bluestem, and sand sagebrush dominated. Sand sagebrush/1ucca/cool and short-tall warm season grasses/annuals (23): Sand dropseed, needle-and-thread, blue grama, sand paspalum, little bluestem, small soapweed, and sand sagebrush dominated. Common annuals included cheatgrass brome and western ragweed. Sand sagebrush/Xucca/cool and short-tall warm season grasses/muhly (25): Prairie sandreed, sandhill muhly, sand dropseed, needle-and-thread, blue grama, hairy grama, little bluestem, small soapweed, and sand sagebrush dominated on these relatively steep sandy areas. Sand sagebrush/Xucca/short-tall warm season grasses (28): Prairie sandreed, sand dropseed, needle-and-thread, blue grama, little bluestem, small soapweed, and sand sagebrush dominated. This type is similar to # 44 except for the presence of sand sagebrush. Sand sagebrush/Xucca/short-tall warm season grasses/muhly (31): Prairie sandreed, sandhill muhly, sand dropseed, needle-and-thread, blue grama, hairy grama, little bluestem, hairy grama, small soapweed, and sand sagebrush dominated on these relatively steep sandy areas. Sand sagebrush/cool and short-tall warm season grasses/muhly (38): Prairie sandreed, sandhill muhly, sand dropseed, needle-and-thread, blue grama, hairy grama, little bluestem, and sand sagebrush dominated on these relatively steep sandy areas. Yucca/cool and short-tall warm season grasses (44): Prairie sandreed, sand dropseed, needle-and-thread, blue grama, little bluestem, and small soapweed dominated. Mid-tall grass: Prairie sandreed and sand bluestem typically dominated in most of this type. Additional dominant species were sand lovegrass, sandhill muhly, sand dropseed, needle-and-thread, blue grama, hairy grama, switchgrass panicum, big bluestem, sand bluestem, little bluestem, yellow Indiangrass, and lemon scurfpea. Many of these habitats appeared to had undergone sand sagebrush removal, either through general patterns of land use or spraying. Some areas had been seeded with a variety of tall grasses (such as big bluestem and yellow Indiangrass). Mid-tall warm season grasses (19): Sand lovegrass, prairie sandreed, sand dropseed, needle-and-thread, blue grama, sand paspalum, switchgrass panicum, big bluestem, sand bluestem, little bluestem, I �141 and yellow Indiangrass dominated this diverse habitat type. Common annuals included cheatgrass brome and prairie sunflower. This habitat was generally the result of seeding efforts and was usually on flat areas formerly used for agriculture and/or intensive grazing. Although quite variable with respect to grazing treatment, much of this habitat provided tall dense cover. Mid-tall warm season grasses/muhly/annuals (20): Sand lovegrass, pra~r~e sandreed, sandhill muhly, sand dropseed, needle-and-thread, blue grama, hairy grama, switchgrass panicum, sand bluestem, and little bluestem dominated on these relatively steep sandy sites. Common sagewort and western ragweed were also common. Cool and mid-tall warm season grasses/annuals (33): Prairie sandreed, sand dropseed, needle-and-thread, sand paspalum, and switchgrass panicum dominated. Cheatgrass brome and Texas croton were also common. Psoralea/mid-tall warm season grasses/muhly (36): Prairie sandreed, sandhill muhly, hairy grama, sand bluestem, and lemon scurfpea dominated on these steep sandy areas. Sage/mid-tall grass: Sand lovegrass, prairie sandreed, s~ndhill muhly, sand dropseed, needle-and-thread, blue grama, sand bluestem, and little bluestem dominated. Sandhill muhly, hairy grama, witchgrass panicum, and switchgrass panicum were also dominant, but were more restricted in distribution. Many of these sites had been seeded with a variety of tall grasses (such as big bluestem and yellow Indiangrass). Sand sagebrush/mid-tall warm season grasses/annuals (17): Sand lovegrass, prairie sandreed, sandhill muhly, sand dropseed, needleand-thread, blue grama, hairy grama, sand bluestem, and little bluestem dominated. Sandhill muhly and hairy grama were more typical on steep sandy areas while sand lovegrass was common on flat areas. Cheatgrass brome, gilia penstemon, and western ragweed were common throughout. Sand sagebrush/mid-tall warm season grasses \(24): Cheatgrass brome, sand lovegrass, prairie sandreed, sand dropseed, needle-and-thread, blue grama, switchgrass panicum, big bluestem, sand bluestem, little bluestem, yellow Indiangrass, sand sagebrush, prairie sunflower, and western ragweed dominated. The diversity of species was partly a result of seeding of big bluestem, yellow Indiangrass, and other species. Other common species included cheatgrass brome, prairie sunflower, and western ragweed. Sand sagebrush/Yucca/mid-tall warm season grasses (26): Sand lovegrass, prairie sandreed, sand dropseed, needle-and-thread, blue grama, switchgrass panicum, sand bluestem, little bluestem, small soapweed, and sand sagebrush dominated. Sand sagebrush/Yucca/mid-tall warm season grasses/muhly (27): Sand lovegrass, prairie sandreed, sandhill muhly, sand dropseed, needleand-thread, hairy grama, switchgrass panicum, sand bluestem, little �142 bluestem, small soapweed, and sand sagebrush dominated this steep habitat type. Prairie sunflower was also a common species. Yucca/mid-tall warm season grasses (32): Prairie sandreed, sand dropseed, needle-and-thread, blue grama, switchgrass panicum, sand bluestem, little bluestem, and small soapweed dominated on this steep habitat type. County roads: Most county roads on the study area were slightly elevated (1-2 m), gravel, and approximately 10 m wide. Each road had a 5-m wide bank on either side generally bounded by a fence. Since most roads and banks were not affected by grazing pressure, they had numerous dense species of mid and tall grass species, such as prairie sandreed, sand dropseed, needle-and-thread, switchgrass panicum, green bristlegrass, big bluestem, sand bluestem, little bluestem, and yellow Indiangrass. Other common species included sandhill muhly, blue grama, sand sedge, small soapweed, annual eriogonum, common pricklypear, tenpetal mentzelia, Nuttall evolvulus, spreading fleabane, sand sagebrush, common sagewort, and prairie sunflower. County roads Female Preference (45): Same as above description. Hypothesis Distance between nest locations and both the nearest lek and the lek where each female was first observed was determined for 1986-88 (Fig. 15). Eightynine females had a mean distance of 3.67 km to the 1ek where first observed and 1.02 km to the nearest lek (Table 38). Nest-first 1ek distances ranged between 0.23-29.40 km (Fig. 16) and nest-nearest 1ek distances ranged between 0.23-2.39 km (Fig. 17). There was no difference in movement attributable to year (f> 0.10). However, yearling females moved further between the lek where first observed and their nest site than adults (~- 2.414, f = 0.019); the difference may be partially attributable to yearling dispersal. For instance,S of 6 females that nested more than 10 km from the lek where first observed or captured were yearlings. When 'disturbed' females (captured during the breeding season) were compared with 'undisturbed' females (captured prior to the breeding season), disturbed females moved further (~- 2.105, f 0.039) between the first visited lek and their nest site. As with yearling movements, movement by undisturbed females may be caused by migration/dispersal. Since undisturbed females were caught prior to the breeding season, some may have been caught prior to normal movements between seasonal home ranges. There was no difference (f > 0.05) for nest-nearest lek distance attributable to either age or disturbance status. Seventy-three percent (n = 90) of the females nested closer to a lek that was different from the lek on which they were first observed or captured. There was no difference (f > 0.10) associated with year. If Bradbury's (1981) prediction that most females (presumably more than 50%) should visit only 1 lek is true, then most females should have nested closer to the lek on which they were first observed. However, when compared with a predicted value of 50%, the observed value of 73% was higher (X2 - 10.364, f < 0.01). Disturbed females, 74.3% (n = 74), did not differ (X2 - 0.209, f> 0.10) from undisturbed females, 68.8% (n = 16). If disturbance was a factor, disturbed females should be expected to avoid leks where captured. �143 1986 o o o N o LEKS * NESTS o o o o oo 0 1987 o 1988 o ~o o 2 4km Fig. 15. Distribution of female greater prairie-chicken nests in relation to the lek where each female was first observed or captured during spring in northeastern Colorado, 1986-88. �144 Table 38. Distances (m) between each greater prairie-chicken nest, nearest lek, and the lek where each female was first observed and/or captured in northeastern Colorado, 1986-88. Lek visited Category first R Lek nearest to nest 11 Median R 2112.2 4327.9 6922.9 21 30 31 843.0 963.0 1139.0 962.1 999.8 1075.2 527.6 574.0 465.8 2348.4 4847.6 2248.8 6685.8 38 43 966.5 980.0 984.8 1021. 1 506.1 509.1 2211.4 3940.1 1874.9 5599.4 15 67 1166.0 849.0 1231. 2 971.1 577 .4 497.1 1018.7 518.0 11 Median Year 1986 1987 1988 22 31 36 2092.5 1768.0 1855.0 2855.8 3006.4 4734.5 Agea Adult Yearling 42 47 1690.5 2041.0 14 75 1767.0 1980.0 SO SD St.at.us" Undisturbed Disturbed " Totals 89 1874.0 3668.2 5224.6 82 975.0 aOne female of unknown-age was excluded from the sample for nearest lek nest distances. bFemales captured prior to the period of lek visitation were considered undisturbed and those captured on leks during the period of lek visitation were considered disturbed. Although direct examinations of home range size are difficult, if not interpretable (Beehler and Foster 1988), home range sizes were estimated to give an additional indication of the relevancy of Bradbury's (1981) prediction of home range size and lek distribution. Female home range size was estimated for both the early spring (prior to most lek visitation) and late spring (during lek visitation and nest establishment) (Table 39). By using a 75% probability contour, a conservative estimate of home range size was obtained. There were no differences (£ > 0.05) in home range size associated with sex, age, or disturbance status. However, home range size during late spring (R 624 ha) was larger (~ - 2.098, £ - 0.038) than in early spring (R - 213 ha) (Table 39). Since Bradbury (1981) predicted that females should have a diameter of their respective home ranges that was less than the distance between neighboring leks, home range diameter was calculated by considering the home range as a perfect circle (conservative estimate of diameter). The diameter of an average home range was 1.65 km (1.43 km using the median home range) for early spring and 2.82 km (1.84 km using the med~an home range) for late spring. �145 18 15 en w ..J < :E w u.. 12 9 a 3 o o 5 10 15 20 25 30 DISTANCE (KM) Fig. 16. Distribution of distances between 89 female greater prairie-chicken nest sites and leks where each female was first observed or captured in northeastern Colorado, 1986-88. �146 10 8 CI) W 6 --I < ~ w LL 4 2 o 0.0 0.4 0.8 1.2 1.6 2.0 2.4 DISTANCE (KM) Fig. 17. Distribution of distances between 82 female greater prairie-chicken nest sites and the nearest lek in northeastern Colorado, 1986-88. �147 Table 39. Home range size (ha, area within 75% probability contours generated with harmonic means on a grid size of 25 X 25) for female greater prairie-chickens in northeastern Colorado, 1986-88. Late spring Early spring n Median 2£ 3 2 25 165.0 328.7 156.8 149.3 328.7 210.9 Age Adult Yearling 22 8 165.0 134.9 Statusa Undisturbed Disturbed 27 3 Totals 30 Category SD n Median 2£ SD U5.6 247.9 252.7 21 35 39 252.5 262.9 312.1 526.9 349.9 938.4 749.8 292.0 2896.2 175.9 313.6 97.2 439.2 48 47 264.5 276.2 367.3 307.6 899.6 2671.9 156.8 165.0 219.6 149.3 249.6 U5.6 18 77 470.1 2~.8.0 649.0 626.3 981.0 2062.2 160.8 212.6 239.3 95 266.0 630.6 1900.7 Year 1986 1987 1988 aFemales captured prior to the period of lek visitation were considered undisturbed and those captured on leks during the period of lek visitation were considered disturbed. These estimates are greater than distances between neighboring leks (mean range of 1.18-1.31, median range of 1.15-1.42). Bradbury (1981) also suggested that detection distance (distance at which males can be readily detected by females) could be added to the diameters of female home ranges. If this procedure was used, females would have much larger home ranges than predicted. seventy-nine radio-marked females were observed on leks at least twice during the breeding season; 74.7% of these were observed on more than 1 lek (Fig. 18). When compared with the hypothesized upper limit of 50% (Bradbury 1981), the observed value was greater (X2 - 10.251, ~ < 0.01). If females visited leks several times during the breeding season as indicated, the number of females visiting more than 1 lek could have been higher than 85% (Fig. 18). As the number of visits to leks increased, the number of visits to different leks also increased (Fig. 19). One female was observed on 6 different leks in 8 lek visits. Another female was observed visiting the same lek on 9 consecutive observations. Undisturbed females (n = 17, 70.6%) did not differ (X2 = 0.192, ~ > 0.10) from disturbed females (n = 62, 75.8%) in their likelihood to visit more than 1 lek. If disturbance was a factor, disturbed females should have visited more leks than undisturbed females. A possible disturbance factor was examined by �---(f. -- n=15 ~ 6----- ----------------- 100 w ...J .•.... o TOTAL z 6 UNDISTURBED -c J: •••• o 95 , " 0: ::E C) n=34 90 z n=79 6.. - ,, o ~=56 CI) CI) W n=22,/ ,A · " O. .---- ..- .... --.-.--~ ... •••• > , ,, DISTURBED W o n=6 .- ... _- .._- 85 /\,-,' ~ ...J D. ·· -c ::E w 80 o z o 75 u.. u.. Ii: o 0.. o 0: 0.. 70 r 2 -- ---r-----r-----r-----~------. 3 4 5 6 7 MINIMUM NUMBER OF LEK VISITS Fig. 18. Proportion of female greater prairie-chickens visiting more than 1 lek in relation of times observed on a lek during spring in nortlleastern Colorado, 1986-88. to number I-.j> oo �o 7 TOTAL 6. UNDISTURBED o o 6 DISTURBED W I- en :> 5 MAXIMUM NUMBER OF DIFFERE.., ~ ~ LEK VISITS POSSIBLE 4 I- Z ~ 3 W LL LL n=25 ················~~=22 (5 2 n=79 0=56 ·.D=34 1 I 1 2 3 4 5 >5 LEKVISIT t-' p\0 Fig. 19. Number of different leks visited by female greater prairie-chickens times observed on leks in spring in northeastern Colorado, 1986-88. in relation to number of �150 compiling data on consecutive lek visits (Fig. 20). If disturbed females were adversely affected by being captured, they should have avoided the capture lek on their next lek visit. However, there was no difference between disturbed and undisturbed females after their first lek visit. The likelihood of visiting a different lek declined with later visits to leks. The probability of visiting a different lek in visits 3 - > 5 (n - 137 combined, 55.5%), declined (X2 - 7.889, f < 0.01) from the 2nd lek visit (n - 79, 74.7%). The trend to visit the same lek(s) in later visits may represent selection of leks by females, as opposed to effects of disturbance. DISCUSSION Reproductive success increased from 1986 to 1987 and then stabilized. These trends were reflected in the increase in both lek density and male attendance at leks in 1987 and 1988. Despite these trends, the lack of quantified density measurements, makes detailed interpretations of these results tenuous. Research has begun to show the importance of lightly grazed rangelands for providing critical nesting, roosting, and foraging habitat. Mapping of the study area may help design a management strategy that will help insure the long-term security of greater prairie-chickens in northeastern Colorado . Female Preference Hypothesis . ' Most females (73%) nested closer to a lek other than that where captured or first observed and most (79%) visited more than 1 lek during the breeding season. Measures of both home range size and lek visitation suggested these results were conservative. -These results contradict Bradbury's (1981) predictions that females should generally visit only 1 lek and have home ranges that include only 1 lek. Therefore, the 'female preference' hypothesis is rejected for greater prairie-chickens in northeastern Colorado. The most important aspect of Bradbury's (1981) theory is the relationship between female home range size and tendencies for males to form leks. However, even in Bradbury's (1981) paper on lek evolution, female home range size was interpreted as a correlative factor, with little justification given for possible causation. Considerations of home range size are often disregarded because of problems associated with measuring home ranges (Beehler and Foster 1988). However, sizes and distributions of home ranges are probably the most important considerations because they provide insight into why some species may form leks and others may not. Other hypotheses, 'least costly male' (Wrangham 1980), 'hotspot' (Bradbury and Gibson 1983), and 'hotshot' (Beehler and Foster 1988), have incorporated the idea of female home range size (sometimes substituted with female choice) into their respective theories explaining the evolution of lek behavior. However, none differ with respect to the expected positive correlation between ability of females to choose males and/or leks and tendency of males to form leks (Schroeder, in prep.). While Bradbury (1981) suggested that female preference could result in a predicted distribution of leks and female home ranges, his proposed predictions did not provide an adequate test for the idea that increasing home ranges lead to increasing concentrations of males. Bradbury's (1981) original explanation for disappearance of unselected leks (resulting in �I- Z W a: w LL LL 0 -'#. en - 0 w ~ 0 t: 80 w 70 en > > w ~ _J ~ w _J w ~60 I- > w I ~ I- wOO I TOTAL D UNDISTURBED D DISTURBED A , 0=79 , .,~ 0 a: o n=56 /\ , , .. ' -,_ ;' 0J.- .' '. n=25 , , " 0.. 0=34 ···········0··········· , ... ,: , o . I j::: I:J Z m « I « m 0=22 , , ,, 1-40 ~ 0 a: 0.. 2 3 4 5 >5 LEK VISIT ,_. ,_. lJ1 Fig. 20. Likelihood of a female greater prairie-chicken visiting a lek different than the previous relation to the consecutive number of lek visit in spring in northeastern Colorado, 1986-88. lek in �152 the predicted lek distribution) may have been too simplistic as it did little to interpret the effects of male-male competition and variable male breeding potential (ideas subsequently incorporated into the hotspot hypothesis (Bradbury and Gibson 1983)). Female movement also may have implications for greater prairie-chicken management. Land use practices in areas critical to lekking species may be modified to conform to specifications relating to lek locations. One way to do this would be to delineate an arbitrary circumference of habitat around all known leks. These areas may be considered the most critical for nesting habitat, a factor considered when mitigating for habitat disturbance or loss. However, little information is available describing female movements and nest sites in relation to lek location. This problem may be of particular importance for greater prairie-chickens in northeastern Colorado. The reestablishment of greater prairie-chickens in formerly occupied range may necessitate additional information on the relative importance and proximity of nest and lek habitat. LITERATURE CITED Ammann, G. A. 1944. Determining the age of pinnated grouses. J. Wildl. Manage. 8:170-171. Amstrup, S. C. 44:214-217. 1980. A radio-collar Beehler, B. M., and M. S. Foster. preference in the organization 131:203-219. for game birds. and sharp-tailed J. Wildl. Manage. 1988. Hotshots, hotspots, and female of lek mating systems. Am. Nat. Bradbury, J. W. 1981. The evolution of leks. Pages 138-169 in R. D. Alexander and D. W. Tinkle, eds., Natural selection and social behavior: recent research and new theory. Chiron Press, New York, N.Y. _____ , and R. M. Gibson. 1983. Leks and mate choice. Pages 109-138 Cambridge Univ. Press, Cambridge, P. Bateson, ed., Mate choice. Mass. Dixon, K. R., and J. A. Chapman. 1980. Harmonic mean measure activity areas. Ecology 61:1040-1044. Harrington, H. D. 1964. Manual of the plants of Colorado. Swallow Press Inc., Chicago, Ill. in of animal 2nd ed. The Miller, G. C. 1984. Development of a preservation program for insular populations of prairie grouse. Colorado Div. Wildl. Fed. Aid Rep. N-l-R. Jan. Pp. 129-170. Robel, R. J., J. N. Briggs, A. D. Dayton, and L. C. Hulbert. 1970g. Relationships between visual obstruction measurements and weight of grassland vegetation. J. Range Manage. 23:295-297. �153 ______ , J. J. Cebula, N. J. Silvy, C. E. Viers, and P. G. Watt. 1970g. Kansas. Greater prairie chicken ranges, movements, J. Wildl. Manage. 34:286-306. and habitat usage in Scott, T. G., and C. H. Wasser. 1980. Checklist of North American plants for wildlife biologists. The Wildl. Soc., Washington, D.C. Wrangham, R. W. 1980. Female choice of least costly males; a possible factor in the evolution of leks. Z. Tierpsychol. 54:357-367. r Prepared by __;_VL;L)~' ....•••••.•...:...........o~'-'-',.....__",.£""""""'L=.o.o:>iA~"'-- __ Michael A. Schroeder Graduate Research Assistant Approved by Clait E. Braun Wildlife Research Leader ��155 Appendix A. Colorado. Plants on a greater prairie-chicken Scientific r name Solanum americanum Vicia americana Eriogonum annuum Rosa arkansana Salvia azurea Mamillaria vivipara Andropogon gerardi Liatris sguarrosa Redfieldia flexuosa Bouteloua gracilis Verbena hastata Cryptantha minima Sitanion hystrix Verbena bracteata Opuntia fragilis Solanum rostratum Cirsium vulgare Ipomoea leptophylla Hoffmanseggia jamsii Solidago canadensis Conyza canadensis Cirsium arvense Mollugo verticillata Nepeta cataria Bromus tectorum Ulmus parvifolia Physalis heterophylla Echinochloa crusgalli Taraxacum officinale Oenothera strigosa Asclepias speciosa Anaphalis margaritacea Opuntia humifusa Tribulus terrestris Portulaca oleracea Salsola kali Artemisia campestris Helianthus annuus Aster tanacetifolius Agropyron desertorum Polygonum lapathifolium Rumex crispus Grindelia sguarrosa Oenothera coronopifolia Liatris punctata Chrysothamnus viscidiflorus Castilleja sessiliflora study area in northeastern Common name American nightshade American vetch Annual eriogonum Arkansas rose Azure sage Ballcactus Big bluestem Blazing star Blowout grass Blue grama Blue verbena Borage Bottlebrush squirreltail Bracted verbena Brittle pricklypear Buffalobur nigh~shade Bull thistle Bush morningglory Bushpea Canada goldenrod Canada horseweed Canada thistle Carpetweed Catnip Cheatgrass brome Chinese elm Clammy groundcherry Common barnyardgrass Common dandelion Common eveningprimrose Common milkweed Common pearlyeverlasting Common pricklypear Common puncturevine Common purslane Common Russianthistle Common sagewort Common sunflower Common tansy-aster Crested wheatgrass Curl top ladysthumb Curly dock Curlycup gumweed Cut-leaf eveningprimrose Dotted gay-feather Douglas rabbitbrush Downy painted-cup �156 Appendix A. Continued. Scientific name Munroa sguarrosa Euphorbia fendleri Senecio fendleri Antennaria neglecta Kochia scoparia Descurainia sophia Mirabilis linearis Hordeum jubatum Setaria italica Solidago gigantea Penstemon ambiguus Dalea ~ Setaria viridis Astragalus crassicarpus Chrysopsis villosa Bouteloua hirsuta Lithospermum caroliniense Apocynum cannabinum Phlox hoodii Oryzopsis hymenoides Argemone intermedia Haplopappus spinulosus Cristatella jamesii Poa pratensis Oxytropis lambertii Chenopodium album Amorpha canescens Euphorbia esula Psoralea lanceolata Andropogon scoparius Orobanche ludoviciana Artemisia ludiviciana Cenchrus pauciflorus Solidago missouriensis Astragalus missouriensis Sparganium eurycarpum Chenopodium pratericola Penstemon angustifolius Stipa comata Eleocharis acicularis Thelesperma megapotamicum Evolvulus nuttallianus Hymenopappus tenuifolius Dactylis glomerata Palafoxia sphacelata Rumex altissimus Ipomoea longifolia Descurainia pinnata Common name False buffalo grass Fendler euphorbia Fendler groundsel Field pussy toes Fireweed summercypress Flixweed tansymustard Four-o'clock Foxtail barley Foxtail bristlegrass Giant goldenrod Gilia penstemon Golden dalea Green bristlegrass Groundplum milkvetch Hairy goldenaster Hairy grama Hairy puccoon ~ Hemp dogbane Hood phlox Indian ricegrass Intermediate pricklypoppy Ironplant goldenweed James cristatella Kentucky bluegrass Lambert crazyweed Lambsquarters goosefoot Lead plant Leafy spurge Lemon scurfpea Little bluestem Louisiana broomrape Louisiana sagewort Mat sandbur Missouri goldenrod Missouri milkvetch Narrow-leaved bur-reed Narrow-leaved goosefoot Narrow-leaved penstemon Needle-and-thread Needle spikerush Nippleweed Nuttall evolvulus Old plainsman Orchardgrass Hooker palafoxia Pale dock Pink-throated morningglory Pinnate tansymustard �157 Appendix A. Continued. Scientific name Delphinium virescens Asclepias pumila Cirsium plattense Ratibida columnifera Commelina crispa Oenothera albicaulis Talinum parviflorum Koeleria cristata Allium textile Lepidium densiflorum Calamovilfa longifolia Tradescantia occidentalis Helianthus petiolaris Lactuca scariola Polygonum aviculare Eragrostis spectabilis Petalostemon purpureum Agropyron rep ens Aristida longiseta Amaranthus retroflexus Cleome serrulata Tridens elongatus Polanisia trachysperma Chrysothamnus nauseosus Lygodesmia juncea Elaeagnus angustifolia Tragopogon pratensis Amaranthus torreyi Andropogon hallii Sporobolus cryptandrus Eragrostis trichodes Asclepias arenaria Paspalum stramineum Artemisia filifolia Psoralea digitata Carex gravida Muhlenbergia pungens Abronia fragrans Gaura coccinea Sphaeralcea coccinea Chloris virgata Oenothera serrulata Apocynum sibericum Bouteloua curtipendula Petalostemon villosus Psoralea argophylla Lesguerella ludoviciana Festuca octoflora Common name Plains larkspur Plains milkweed Platte thistle Prairie coneflower Prairie dayflower Prairie eveningprimrose Prairie flameflower Prairie junegrass Prairie onion Prairie pepperweed Prairie sandreed Prairie spiderwort Prairie sunflower Prickly lettuce Prostate knotweed Purple lovegrass Purple prairieclover Quackgrass " Red threeawn Redroot amaranth Rocky Mountain beeplant Rough tridens Roughseed clammyweed Rubber rabbitbrush Rush skeletonplant Russian-olive Salsify Sand amaranth Sand bluestem Sand dropseed Sand lovegrass Sand milkweed Sand paspalum Sand sagebrush Sand scurfpea Sand sedge Sandhill muhly Sandverbena Scarlet gaura Scarlet globemallow Showy chloris Shrubby eveningprimrose Siberian dogbane Sideoats grama Silky prairieclover Silverleaf scurfpea Silvery bladderpod Sixweeks annual fescue �158 Appendix A. Continued. Scientific name Froelichia gracilis Psoralea tenuiflora Gaura parviflora Yucca glauca Bromus inermis Helenium autumnale Euphorbia marginata Erigeron divergens Cymopterus acaulis Eragrostis cilianensis Asclepias incarnata Panicum virgatum Mentzelia decapetala Croton texensis Psoralea linearifolia Senecio longilobus Euphorbia serpyllifolia Euphorbia dentata Cycloloma atriplicifolium Amaranthus graecizans Rumex venosus Nothocalais cuspidata Ambrosia coronopifolia Cryptantha jamesii Trifolium repens Penstemon albidus Petalostemon compactus Melilotus alba Cucurbita foetidissima Avena fatua Salix sp. Chloris verticillata Panicum capillare Plantago patagonica Setaria lutescens Sorghastrum nutans Tragopogon dub ius Melilotus officinalis Common name Slender snakecotton Slimflower scurfpea Small-flowered gaura Small soapweed Smooth brome Sneezeweed Snow-on-the-mountain euphorbia Spreading fleabane Stemless cymopterus Stinkgrass Swamp milkweed Switchgrass panicum Tenpetal mentzelia Texas croton Thread-leaved alfalfa Threadleaf groundsel Thyme-leaved eURhorbia Toothed euphorbia Tumble ringwing Tumbleweed amaranth Veiny dock Wavy leaf microseris Western ragweed White borage White clover White penstemon White prairieclover White sweetclover Buffalogourd Wild oat Willow Windmillgrass Witchgrass panicum Woolly plantain Yellow bristlegrass Yellow indiangrass Yellow salsify Yellow sweetclover �159 Appendix B. Phylogenetic list of scientific and common names for plants (Class Angiospermae) on a greater prairie-chicken study area in northeastern Colorado. Subclass Family Tribe Scientific name Common Monocotyledoneae Sparganiaceae Sparganium eurycarpum Gramineae Festuceae Bromus inermis Bromus tectorum Festuca octoflora Poa pratensis Eragrostis cilianensis Eragrostis spectabilis Eragrostis trichodes Redfieldia flexuosa Dactylis glomerata Tridens elongatus Hordeae Agropyron rep ens Agropyron desertorum Sitanion hystrix Hordeum j uba tum Aveneae Koeleria cristata Avena fatua Agrostideae Calamovilfa longifolia Muhlenbergia pungens Sporobolus cryptandrus Oryzopsis hymenoides Stipa comata Aristida longiseta Chlorideae Chloris virgata Chloris verticillata Bouteloua gracilis Bouteloua curtipendula Bouteloua hirsuta Munroa sguarrosa Paniceae Paspalum stramineum Panicum capillare Panicum virgatum Echinochloa crusgalli Setaria lutescens Narrow-leaved bur-reed Smooth brome Cheatgrass brome Sixweeks annual fescue Kentucky bluegrass Stinkgrass Purple lovegrass Sand lovegrass Blowout grass Orchardgrass Rough tridens Quackgrass Crested wheatgrass Bottlebrush squirrel tail Foxtail barley Prairie junegrass W'ild oat Prairie sandreed Sandhill muhly Sand dropseed Indian ricegrass Needle-and-thread Red threeawn Showy chloris W'indmillgrass Blue grama Sideoats grama Hairy grama False buffalograss Sand paspalum W'itchgrass panicum Switchgrass panicum Common barnyardgrass Yellow bristlegrass �160 Appendix B. Continued. Subclass Family Tribe Scientific name Setaria viridis Setaria italica Cenchrus pauciflorus Andropogoneae Andropogon gerardi Andropogon hallii Andropogon scoparius Sorghastrum nutans Cyperaceae Carex gravida Eleocharis acicularis Commelinaceae Commelina crispa Tradescantia occidental is Liliaceae Yucca glauca Allium textile Dicotyledoneae Salicaceae Salix sp. Ulmaceae Ulmus parvifolia Polygonaceae Eriogonum annuum Polygonum lapathifolium Polygonum aviculare Rumex venosus Rumex crispus Rumex altissimus Chenopodiaceae Cycloloma atriplicifolium Salsola kali Kochia scoparia Chenopodium pratericola Chenopodium album Amaranthaceae Amaranthus graecizans Amaranthus torreyi Amaranthus retroflexus Froelichia gracilis Nyctaginaceae Abronia fragrans Mirabilis linearis Aizoaceae Mollugo verticillata Common Green bristlegrass Foxtail bristlegrass Mat sandbur Big bluestem Sand bluestem Little bluestem Yellow indiangrass Sand sedge Needle spikerush Prairie dayflower Prairie spi.derwo r t; Small soapweed Prairie onion Willow Chinese elm Annual eriogonum Curl top ladysthumb Prostate knotweed Veiny dock Curly dock Pale dock Tumble ringwing Common Russianthistle Fireweed summercypress Narrow-leaved goosefoot Lambsquarters goosefoot Tumbleweed amaranth Sand amaranth Redroot amaranth Slender snakecotton Sandverbena Four-o'clock Carpetweed �161 Appendix B. Continued. Subclass Family Tribe Scientific name Common Portulacaceae Portulaca oleracea Talinum parviflorum Common purslane Prairie flame flower Delphinium Plains larkspur Ranunculaceae virescens Papaveraceae Argemone intermedia Intermediate pricklepoppy Capparidaceae f Cristatella jamesii Polanisia trachysperma Cleome serrulata James cristatella Roughseed clammyweed Rocky Mountain beeplant Lepidium densiflorum Lesguerella ludoviciana Descurainia sophia Descurainia pinnata Prairie pepperweed Silvery b l.adde rpod Flixweed tansymustard Pinnate tansymustard Rosa arkansana Arkansas Hoffmanseggia jamsii Petalostemon compactus Petalostemon villosus Petalostemon purpureum Amorpha canescens Psoralea argophylla Psoralea digitata Psoralea linearifolia Psoralea lanceolata Psoralea tenuiflora Dalea ~ Vicia americana Trifolium repens Melilotus officinalis Melilotus alba Oxytropis lambertii Astragalus missouriensis Astragalus crassicarpus Bushpea White prairieclover Silky prairieclover Purple prairieclover. Lead plant Silverleaf scurfpea Sand scurfpea Thread-leaved alfalfa Lemon scurfpea Slimflower scurfpea Golden dalea American vetch White clover Yellow sweetclover White sweetclover Lambert crazyweed Missouri milkvetch Groundplum milkvetch Tribulus Common puncturevine Cruciferae Rosaceae rose Leguminosae Zygophyllaceae terrestris Euphorbiaceae Croton texensis Euphorbia esula Euphorbia fendleri Euphorbia marginata Euphorbia serpyllifolia Texas croton Leafy spurge Fendler euphorbia Snow-on-the-mountain euphorbia Thyme-leaved euphorbia �162 Appendix B. Continued. Subclass Family Tribe Scientific name Common Euphorbia dentata Toothed euphorbia Malvaceae Sphaeralcea coccinea Scarlet globemallow Loasaceae Mentzelia decapetala Tenpetal mentzelia Cactaceae Opuntia fragilis Opuntia humifusa Mamillaria vivipara Brittle pricklypear Common pricklypear Ballcactus Elaeagnus Russian-olive Elaeagnaceae angustifolia Onagraceae Gaura coccinea Gaura parviflora Oenothera coronopifolia Oenothera serrulata Oenothera strigosa Oenothera albicaulis Scarlet gaura Small-flowered gaura Cutleaf eveningprimrose Shrubby eveningprimrose Common eveningprimrose Prairie eveningprimrose Cymopterus Stemless Umbelliferae acaulis cymopterus Apocynaceae Apocynum Apocynum sibericum cannabinum Siberian dogbane Hemp dogbane Asclepiadaceae Asclepias Asclepias Asclepias Asclepias speciosa pumila arenaria incarnata Common milkweed Plains milkweed Sand milkweed Swamp milkweed Convolvulaceae Evolvulus nuttallianus Ipomoea longifolia Ipomoea leptophylla Nuttall evolvulus Pink-throated morningglory Bush morningglory Phlox hoodii Hood phlox Lithospermum caroliniense Cryptantha m1n1ma Cryptantha jamesii Hairy puccoon Borage White borage Verbena Verbena Blue verbena Bracted verbena Polemoniaceae Boraginaceae Verbenaceae hastata bracteata Labiatae Salvia azure a Nepeta cataria Azure sage Catnip �163 Appendix B. Continued. Subclass Family Tribe Scientific name Common Solanaceae Physalis heterophylla Solanum americanum Solanum rostratum Scrophulariaceae Penstemon angustifolius Penstemon albidus Penstemon ambiguus Castilleja sessiliflora Orobanchaceae Orobanche ludoviciana Plantaginaceae Plantago patagonica Cucurbitaceae Cucurbita foetidissima Compositae Eupatorieae Liatris punctata Liatris squarrosa Astereae Chrysothamnus viscidiflorus Chrysothamnus nauseosus Chrysopsis villosa Grindelia squarrosa Solidago gigantea Solidago canadensis Solidago missouriensis Haplopappus spinulosus Conyza canadensis Erigeron diver gens Aster tanacetifolius Anthemideae Artemisia ludiviciana Artemisia filifolia Artemisia campestris Helenieae Palafoxia sphacelata Hymenopappus tenuifolius Helenium autumnale Heliantheae Ratibida columnifera Thelesperma megapotamicum Helianthus annuus Helianthus petiolaris Clammy groundcherry American nightshade Buffalobur nightshade Narrow-leaved penstemon White penstemon Gilia penstemon Downy painted-cup Louisiana broomrape \.loollyplantain Buffalogourd Dotted gayfeather Blazing star Douglas rabbitbrush Rubber rabbitbrush Hairy goldenaster Curlycup gumweed Giant goldenrod Canada goldenrod Missouri goldenrod Ironplant goldenweed Canada horseweed Spreading fleabane Common tansy-aster Louisiana sagewort Sand sagebrush Common sagewort Hooker palafoxia Old plainsman Sneezeweed Prairie coneflower Nippleweed Common sunflower Prairie sunflower �164 Appendix B. Continued. Subclass Family Tribe Scientific name Senecioneae Senecio longilobus Senecio fendleri Ambrosineae Ambrosia coronopifolia Inuleae Antennaria neglecta Anaphalis margaritacea Cynareae Cirsium vulgare Cirsium arvense Cirsium plattense Cichorieae Tragopogon dubius Tragopogon pratensis Lactuca scariola Nothocalais cuspidata Taraxacum officinale Lygodesmia juncea Common Threadleaf groundsel Fendler groundsel Western ragweed Field pussy toes Common pearlyeverlasting Bull thistle Canada thistle Platte thistle Yellow salsify Salsify .. Prickly lettuce Wavyleaf microseris Common dandelion Rush skeletonplant 1 �165 Colorado Division of Wildlife Wildlife Research Report April 1989 JOB PROGRESS REPORT State of: Colorado Project: W-152-R Upland Bird Research Work Plan: 17 Job Title: Population Period Covered: Author: Job _7_ Dynamics of White-tailed Ptarmigan 01 January through 31 December 1988 Clait E. Braun and Kenneth M. Giesen Personnel: Clait E. Braun and Kenneth M. Giesen, Colorado Division Wildlife of ABSTRACT Long-term studies of populations of white-tailed ptarmigan (Lagopus leucurus) were continued at hunted (Mt. Evans) and unhunted (Rocky Mountain National Park) areas in Colorado through 1988. Densities of breeding ptarmigan increased at Mt. Evans but decreased at Rocky Mountain National Park. Nesting success at both sites was good to excellent and brood size increased at Mt. Evans. There was no known harvest at Mt. Evans in 1988 because the highway was closed due to road construction. Apparently, no hunters walked the 3+ km into areas where ptarmigan were known to occur. ��167 POPULATION DYNAMICS OF WHITE-TAILED PTARMIGAN Clait E. Braun and Kenneth M. Giesen Long-term studies of trends in population size and investigation of reasons for fluctuations in size of tetraonid populations are lacking. Studies on the population dynamics of unhunted and hunted populations of white-tailed ptarmigan were initiated in Colorado in 1966 and have continued essentially uninterrupted at 2 sites. Studies of the unhunted population (Rocky Mountain National Park) identified possible short-term cycles of 7-8 years with an amplitude of 25-30% between high and low breeding densities. Conversely, studies of the manipulated population (hunted) at Mt. Evans have not indicated any cyclic pattern and it would appear that controlled hunting may mask any long-term trend that may occur. This study is designed to examine the question whether white-tailed ptarmigan are truly cyclic and whether hunting affects the apparent oscillations. P. N. OBJECTIVES The goals of this investigation are to be able to predict the length and amplitude of cycles in white-tailed ptarmigan in Colorado, to examine the impact of hunting on cycles, and to clarify underlying causes of the apparent cycles. SEGMENT OBJECTIVES 1. Conduct breeding (May-Jun) and brood (Aug-Sep) censuses ptarmigan using tape-recorded calls of males (breeding) (broods). 2. Censuses will be conducted on previously established, defined study areas at Mt. Evans (hunted) and at Rocky Mountain National Park (unhunted). 3. Capture (noose poles) and band (aluminum and plastic color-coded bands) all unmarked white-tailed ptarmigan encountered on study areas at Mt. Evans and at Rocky Mountain National Park. 4. Individually identify all ptarmigan observed on study areas at Mt. Evans and Rocky Mountain National Park through use of binoculars. 5. Make hunting season and bag limit recommendations for Mt. Evans and collect hunting data through use of volunteer wing barrels and hunter field checks. 6. Compile data, analyze results, and prepare progress of white-tailed and chicks reports. STUDY AREA AND METHODS Areas investigated were Mt. Goliath-Mt. Evans in Clear Creek County and at Tombstone Ridge-Sundance Mountain to Fall River Pass in Rocky Mountain �168 National Park in Larimer County. The physiography, geology, location, and vegetation of these study areas have been previously described (Braun 1969, 1971; Braun and Rogers 1971; Giesen 1977). Ptarmigan were located through use of tape-recorded calls (Braun et al. 1973), captured through use of telescoping noose poles (Zwickel and Bendell 1967) as described by Braun and Rogers (1971), and classified to age and sex and banded following Braun and Rogers (1971). Age of chicks was estimated following Giesen and Braun (1979). Numbered plastic bandettes were not used as in earlier years (Braun and Rogers 1971) as a color-code system using up to 4 different colored plastic bandettes was instituted in 1977-78. A check station was scheduled to be operated on the Mt. Evans highway during the opening weekend of the ptarmigan season in that area. A volunteer wing collection station was available to hunters in the area when the check station was not in operation until the season closed. RESULTS AND DISCUSSION Breeding Densities Mt. Evans.--Timing of breeding events in the Mt. Evans area was about 1 week earlier in 1988 than in 1987. During the May-early June interval, 14 pairs and 2 single males were identified. Thus, breeding densities increased from levels documented in 1986-87 (Table 1). This increase was primarily the result of territories on lower Mt. Goliath and west Mt. Goliath being occupied. During the breeding season, 8 of 16 males identified were yearlings while 9 of 14 hens were yearlings. Recruitment of yearlings was good in 1988. Rocky Mountain National Park.--Surveys of ptarmigan present on breeding territories along Trail Ridge Road in RMNP in May and June indicated the minimum breeding population was 30 birds, which was comprised of 12 pairs and 6 single males. This represents a slight decrease from the 33 birds (13 pairs and 7 single males) identified in 1987 (Table 1). The decreased breeding density reflect a return to average survival of adults and recruitment of yearlings. Survival of banded adults from 1987 was 53.2% (31 of 56 males, 11 of 23 females). Recruitment of chicks banded in 1987 was 18.6% (8 of 43) although yearlings comprised 33% of all adult ptarmigan identified in 1988. Nesting Success and Brood Size Mt. Evans.--Sixteen hens were located during mid July-early September 1987 on or immediately adjacent to the study area. Nine hens (56.3%) were with broods while 7 were apparently unsuccessful nesters (without chicks). Average brood size to 1 September was good (3.3 chicksjhen). Data from 29 chicks that were banded indicated hatch dates from 1 July to 9 August with only 34% hatching before 12 July. Rockv Mountain National Park.--Nest success was estimated at > 75% (all 9 hens observed in Aug-Sep were with broods). Hatch dates calculated from primary molt of chicks captured for banding indicated the median hatch date was 8 July (range 4 to 21 Jul). Average brood size to 1 September was 3.8 chicks. �169 Table 1. 1966-88. White-tailed ptarmigan breeding densities (birds/km2), Colorado Study area Year Rocky Mountain National Park (5.5 km2) Mt. Evans (4.0 km2) 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 11.3 9.8 11.5 12.0 9.6 9.1 8.7 7.8 8.0 11.1 l3.5 12.9 10.7 8.7 8.4 8.2 7.8 6.7 5.8 6.0 4.5 6.0 5.4 3.0 2.7 2.7 2.2 2.0 4.2 7.5 6.2 6.2 6.2 6.7 > 6.0 7.5 10.3 9.5 9.0 6.5 6.5 8.0 8.0 6.5 5.0 7.5 Harvest Mt. Evans.--The hunting season at Mt. Evans in 1988 opened on 17 September and closed on 2 October (16 days) with a bag and possession limit of 3 and 6. Thus, the season was delayed 1 week from the statewide opening as it was in 1981 and 1986-87. The season opening was delayed 2 weeks from 1978 to 1980 and 1982 to 1985. Prior to 1978, experimental seasons were in effect (19701976) or the season opened with the statewide grouse seasons (dates from 17 Aug to 14 Sep). While the season was open in 1988, the Mt. Evans road was closed (as in 1987) (gate) to all vehicular traffic because of road rebuilding above Ptarmigan Flats. Thus, no check station was operated on opening weekend but a volunteer wing collection station was in place on the highway throughout the season. No wings were received from ptarmigan, no birds were known to have been harvested, and no bands were reported. �170 LITERATURE CITED Braun, C. E. 1969. Population dynamics, habitat, and movements of whitetailed ptarmigan in Colorado. Ph.D. Thesis, Colorado State Univ., Fort Collins. l89pp. 1971. Habitat requirements of Colorado white-tailed West. Assoc. State Game and Fish Comm. 51:284-292. ptarmigan. Proc. _____ , and G. E. Rogers. 1971. The white-tailed ptarmigan in Colorado. Colorado Div. Game, Fish and Parks Tech. Publ. 27. 80pp. _____ , R. K. Schmidt, Jr., and G. E. Rogers. 1973. Census of Colorado whitetailed ptarmigan with tape recorded calls. J. Wildl. Manage. 37:90-93. Giesen, K. M. 1977. Mortality and dispersal of juvenile white-tailed ptarmigan. M.S. Thesis, Colorado State Univ., Fort Collins. 55pp. _____ , and C. E. Braun. 1979. A technique for age determination white-tailed ptarmigan. J. Wildl. Manage. 43:508-511. Zwickel, F. C., and J. F. Bendell. J. Wildl. Manage. 31:202-204. Prepared by Clait E. Braun Wildlife Research Leader Kenneth M. Giesen Wildlife Researcher B 1967. A snare for capturing of juvenile blue grouse. �iii Colorado Division of Wildlife Wildlife Research Report April 1989 JOB PROGRESS State of: Colorado Project: W-152-R Upland Bird Research Work Plan: 21 Job Title: Sandsage-Bluestem Period Covered: Author: REPORT Job _3_ 01 January Prairie Renovation through 31 December 1988 Warren D. Snyder Personnel: B. J. Hof and W. D. Snyder, Colorado Division of Wildlife ABSTRACT Precipitation was above average within the Tamarack Prairie during May and July 1988, below average during June, August, and October through December, averaging near to above normal for the year. Phenology was slightly behind that of 1986 and 1987 through April, but advanced rapidly in May and June. The 1984 controlled burns failed to increase the height-density index (HDI) of residual nesting cover. However, by spring 1988, the HDI had recovered to preburn levels. In contrast, the HDI of residual grass-forb vegetation within the 1985 and 1986 controlled burns, was greater than within the control transects in spring 1988 (f < 0.05). HDI among all samples increased dramatically from preceding years, apparently because of favorable precipitation and vegetative growth in 1987 and lack of heavy snow over winter. Crown cover of most warm season grasses increased within 1-2 years after the 1985 and 1986 burns but effects of controlled burns began diminishing by summer 1988. The HDI within the revegetation strips established in 1985 increased to 4.44 dm in spring 1988, far above that of burns or controls. Tall warm season grasses continued to increase in dominance along with their standing residual in these areas. The disk-harrowatrazine herbicide' renovation treatment,. applied in 1986 to a previously interseeded site, showed dramatic increases in HDI over controls along with increased crown cover of interseeded grasses. Impacts of spraying sand sagebrush (Artemisia filifolia) in 1985 could not be detected based on increased grass HDI nor crown cover sampling. However, the sprayed area burned in 1986 had dramatically greater grass HDI than controls in early spring 1988 (f < 0.05). Greater prairie-chickens (Tyrnpanuchus cupido) were observed several times during spring 1988 within the Tamarack Prairie in contrast to previous years when no birds were observed on the property. ��173 SANDSAGE-BLUESTEM PRAIRIE RENOVATION Warren D. Snyder P. N. OBJECTIVES Test renovation and revegetation techniques for increasing standing residual height-density of grasses, the proportion of tall warm-season grasses within the composition, and for reducing the quantity of sand sagebrush to < 30% canopy cover in an ungrazed sandsage-bluestem prairie on the South Tamarack, South Platte Wildlife Area in northeastern Colorado. SEGMENT OBJECTIVES Monitoring follows: of environmental and vegetation conditions and changes continued as 1. Precipitation was monitored throughout the year supplementing rain gauge data with information from nearby weather stations winter months. electronic through the 2. Soil moisture accumulations, plant phenology, and weather were monitored primarily in spring and especially at the time of controlled burns. 3. Visual obstruction (height-density) measurements were obtained treatments and controls where applicable in late winter and/or spring prior to green-up. 4. Crown cover, species composition, and frequency of occurrence measurements were obtained from mid-summer to early fall. 5. Photos of treatments and controls were taken in October. Data compilation and writing the annual job progress report was conducted during fall and winter 1988-89. on early METHODS Approaches used were described by Snyder (1986~, 1986g, 1987, 1988) and are outlined in the Segment Objectives. Segment Objective #4 was conducted within the 1985 and 1986 burn sites and their controls after not being conducted in 1987 and was deleted within the 1984 burns. RESULTS AND DISCUSSION Environmental Conditions Precipitation.--The 4 automatic precipitation recorders operated continuously through spring, summer, and fall 1988. Data from these recorders were supplemented from nearby weather stations during November through March. Precipitation recorded at 3 rain gauges near the north boundary of the study area was near to slightly. above average and ranged from 38.2 to 40.3 cm (15.05 - 15.86 in.). Precipitation at the southcentral rain gauge was 49.2 cm (19.38 �174 in.) with higher than average amounts in May and July. Rainfall in May and July for the entire area was above average but was markedly deficient in June when compared with the long-term mean (Fig. 1). 22 DOCT-OEC 20 18 ~SEP Ld 16 r. ~ALG Vl UJ I ~ v z 14 m...JUL 12 0 -f-c f- !=::·:·Ol...JUN 10 c, l.) ur a:: o, 8 6 _APR 4 2 _...JAN-MAR 0 x 84 85 86 87 88 YEARS Fig. 1. relation Monthly and annual precipitation (in.) from 1984 through to the long-term mean, Tamarack Prairie, Colorado. 1988 in Soil moisture, based on soil probe sampling, was favorable for vegetative growth throughout early spring (Table 1). Samples indicated> 5 cm of moisture was available through the early spring growing season. Table 1. Soil moisture accumulations (m)a based on soil probe spring 1988, Tamarack Prairie, Colorado. Location 4 West North-central South-central East 0.72 l.17 l.07 0.95 A r 18 0.61 l.11 0.95 0.93 aMean depth (m) of 4 probes/location precipitation gauges. samples during May 27 0.88 l. 30 l.19 0.78 4 1.11 l. 34 l. 36 1. 29 taken near each of the 4 22 1. 21 l. 34 1. 53 1.29 �175 Average monthly temperatures at Sterling indicated March through May were slightly warmer than average but June was 6.80 F above the long-term average (Fig. 2). Spring-early summer intervals since 1985 have averaged above normal whereas spring 1984 was colder than normal. Phenology of plant growth on 27 April 1988 (Table 2), when 2 controlled burns were conducted by management personnel on the Tamarack Prairie, was slightly behind that of 1986 and 1987. Based on mean average temperatures the phenology should have been similar to that in 1987 (Fig. 2). However, several hard freezes in mid to late April retarded vegetation growth. By mid to late May phenology was approximately the same as in 1987 (Table 2). 90 ~ • 80 MAY 1\ UL ~ JUN . APR 70 W IT ~ ~ ~ MAR 60 IT W ~ 2 w ~ 50 40 30 x 84 85 86 87 88 YEAR Fig. 2. Monthly average temperature (F) from March through June, 84-88 in relation to the long-term mean, as an index to vegetation phenology, Sterling, Colorado. �176 Table 2. Phenological conditions of selected vegetation during spring 1988, TamaracK Prairie, Colorado. Ma A r Species Artemisia filifolia ~. ludoviciana Astragalus sp. Cymopteris mont anus Evolvulus nuttalianus Lathyrus polymorphus Leucocrinum montanum Mentzelia nuda Penstemon ansustifolius Phlox andicola Psoralea lanceolata Sphaeralcea coccinea Tradescantia occidentalis Tragopogan sp. Agropyron smithii Bouteloua gracilis Calmovilfa longifolia Panicum virgatum Paspalum stramineum Sporobolus cryptandrus S t i pa .£Q!!!lli Cyperus sp. avegetation height (cm). E = early, M = medium, L Height-Density 18 27 04 22 E. budded Basal leaf Emerg. M. budded Basal leaf 5.0 L. budded 2.5 leaf 5-7.6 F. bloom 5-7.6 leaf 5-7.6 Bloom E. bloom F. bloom 05 s·7.6a Budded L. bud Bloom 2.5 5-10 10 E. bloom Basal whorl 5 5 Leafing Leafed 5 Leafed 5 5-7.6 5-7.6 Dormant Dormant Dormant Dormant Dormant 5-10 2.5-5 late, F 5-10 basal Leafed 5 2.5 20-25 E. bloom 10-15 15 7.6-10 10-12.7 2.5-5 Emerging Emerging 5 2.5-5 5 5 5 5-10 5-10 5 Emerging 10-12.7 Bloom 5-7.6 15 L. bloom 5-7.6 15-20 Headed < 7.6 5 L. bloom Headed 15 F. bloom F. bloom 7.6-10 Budded E. bloom Seeded 15-20 7.6-10 20 15-20 5-10 10 E. heading full. Sampling Within Burned Sites There was no evidence that the controlled burn conducted at site 1-84 in spring 1984 enhanced growth of grass-forb vegetation in subsequent years based on visual obstruction readings (VOR - HDI, Table 3, Fig. 3). Fire reduced the HDI but a full recovery in relation to the controls was evident by spring 1988 (end of the 3rd growing season). Grass-forb HDI increased dramatically from previous years within both the burned and control transects, primarily because of above average precipitation during the 1987 growing season (Fig. 1). Lack of significant lodging of vegetation by winter snow cover was also a factor enhancing HDI in spring 1988. Sand sagebrush had almost completely recovered its original status in the burned site by spring 1988. Grasses recovered slowly through 1986 yielding a low HDI in late winter 1987 following controlled burning within the 3-84 burn site. This area was more sandy and contained a much higher density of competing sandsage than burn 184. A dramatic regrowth of grass-forb vegetation was noted during 1987 (Table 4, Fig. 4). Based on comparisons with 1984, pretreatment HDI of grass-forb vegetation had not completely recovered by early spring 1988 although it and sandsage were approaching complete recovery. There is no evidence that either of the 1984 burns enhanced standing residual nesting cover needed by nesting prairie grouse (Tetraoninae). There is no significant analysis of covariance change from preburn to 1988 status for either burn. �177 Table 3. Mean height·density (dm) within burn 1-84 and it's controls during spring 1984-88 intervals, Tamarack Prairie, Colorado. Years Grass/fb Burn Sandsage Combined Grass/fb Control Sandsage Combined 1984 1985 1986 1987 1988 0.256 0.134 0.232 0.208 0.589 0.856 0.313 0.358 0.526 0.836 0.372 0.162 0.256 0.258 0.627 0.253 0.295 0.301 0.224 0.587 0.814 0.687 0.643 0.847 1.045 0.334 0.355 0.368 0.309 0.622 £. Values 1984-88 1985-88 1986-88 1987-88 :e.e. 0.00 20.39a 7.37b 0.15 < < 0.60 1.33 0.22 0.32 0.01 9.82a 6.73b 0.10 0.010. 0.025. 0.6 O.S r-. ~ u >- 0.4 lll) z UJ 0 I I- I CONTROL 0.3 "'- .)3----~(.. (CJ , "- UJ J: 0,2 'BURN 0.1 o 84 B5 86 87 88 YEARS Fig, 3. Height-density (dm) of residual grass-forb vegetation from 1984 (pretreatment) to 1985-88 (post-treatment) within the 1-84 burn and control sites, Tamarack Prairie, Colorado, �178 Table 4. Mean height-density (dm) within Burn 3-84 and its controls during spring 1984-88 intervals, TamaracK Prairie, Colorado. Years Grass/tb 1984a 1985 1986 1987 1988 0.222 0.021 0.106 Burn Sandsage Combined Grass/tb Control Sandsage Combined 0.827 0.121 0.356 0.286 1.215 0.493 0.047 0.201 0.157 0.831 0.183 0.191 0.200 0.216 0.625 0.935 0.797 1.044 0.688 1.494 0.531 0.503 0.629 0.424 1.059 o.on 0.466 .E Values 1984-88 1985-88 1986-88 1987-88 5.68 2.89 1.73 0.70 3.n 0.85 0.13 1.27 0.19 0.02 0.89 0.00 apretreatment. 0.7 0.6 I I 0.5 I J J J ~ 0.4 (Jl z UJ o f I ~ 0.3 J (.:J CONTROL UJ ---......." I 0.2 ..,;'" - - - ~~ - - - 84 85 J..>.----;/) ... 0.1 o 86 87 88 YEARS Fig. 4. Height-density (dm) of residual grass-forb vegetation (pretreatment) to 1985-88 (post-treatment) within the 3-84 control sites, Tamarack Prairie, Colorado. from 1984 burn and �179 The 1985 and 1986 burns enhanced grass-forb HDI by early spring 1988 (Figs. 5, 6). Burning suppressed grass-forb HDI through the 1st growing season within the 1985 burns. The grass-forb HDI had recovered by the end of the 2nd growing season and was ahead of the controls by the end of the 3rd growing season. Mean grass-forb HDI within the burns in early spring 1988 was markedly greater than within the controls (Tables 5, 6, f < 0.01). Sandsage did not show this rate of recovery or enhancement of HDI (Table 5, f> 0.10). Thus, the combined vegetation HDI was not as dramatic as that for grass-forb alone. 0.8 f\ E D ~ ~ ~ ~ , ,, BURN , Z W 0 I I I ~ I ~ - ,, 0.6 I ~. I 0.4 , I W I I ---It-~ \ CONTROL 0.2 o 85 87 86 88 YEARS Fig. 5. Height-density (dm) of residual grass-forb vegetation from 1985 (pretreatment) to 1986-88 (post-treatment) within the 1985 combined burn and control sites, Tamarack Prairie, Colorado. �180 0.8 1\ E D ~ ~ ~ ~ BURN 0.6 ~ I ~ I I I Z W 0 I ~ I ~ 0.4 W I CONTROL 0.2 o 85 86 87 88 YEARS Fig. 6. Height-density (dm) of residual grass-forb vegetation from 1985-86 (pretreatment) to 1987-88 (post-treatment) within the 1986 combined burn and control sites, Tamarack Prairie, Colorado. The impact of burning on 1st year (1987) HDI was not as severe as in previous years apparently because of increased precipitation received in 1986 (Figs. 1, 6). The HDI within the 1986 burned sites by the end of the 2nd growing season was markedly ahead of that within the controls (Tables 7, 8). The burns attained greater HDI by spring 1988 than the controls in comparison to preburn sampling in 1985 and 1986 (f < 0.05) (Table 8). The recovery from 1987 to 1988 was even more dramatic (f < 0.005). Sandsage had not recovered from burning by early spring 1988 or within 2 growing seasons (Table 7). Because of the dominance of grasses over sandsage within the burns, the combined vegetation HDI also showed evidence of stimulation by fire. �181 Table 5. Mean height-density (dm) within 1985 burns and their controls during spring 1985-88 intervals, Tamarack Prairie, Colorado. Burn Year 2 Control z 3 2 3 z 0.374 0.325 0.322 0.511 0.346 0.280 0.376 0.775 0.158 0.180 0.156 0.443 0.323 0.277 0.325 0.644 0.872 0.828 0.925 1.070 0.615 0.745 0.560 1.732 0.537 0.767 0.648 1.253 0.627 0.771 0.670 1.317 0.444 0.391 0.383 0.581 0.369 0.335 0.393 0.859 0.323 0.465 0.411 0.878 0.380 0.382 0.394 0.782 GRASS-FORB 1985a 1986 1987 1988 0.507 0.138 0.419 0.780 0.381 0.119 0.472 1.028 0.180 0.053 0.244 0.667 0.379 0.110 0.408 0.877 SANDSAGE 1985a 1986 1987 1988 1.008 0.417 0.523 0.938 0.679 0.342 0.591 1.167 0.642 0.438 0.769 1.144 0.744 0.388 0.672 1.125 COMBINED 1985a 1986 1987 1988 0.565 0.142 0.425 0.786 0.407 0.130 0.476 1.037 0.302 0.088 0.330 0.766 0.429 0.124 0.426 0.899 apretreatment. Table 6. Height-density relationships controls among years, Tamarack Prairie, Vegetation between the 1985 burns Colorado. 1985-88 and their 1986-88 1987-88 17.29b l.47 6.80a l. 53 2.43 E VALUE Grass-forb Sandsage Combined 7.78b l. 31 3.53 3.04 PAIRED !. Grass-forb Sands age Combined ap b~ < 0.025. < 0.010. 7.39a 2.84 0.64a 13.34b 0.81 6.66a 14.8sb 0.74 l. 33 �182 TabLe 7. Mean height·density (dm) within 1986 burns and their controLs during spring 1985-88 intervaLs, TamaracK Prairie, Colorado. Burn Year 2 ControL 3 2 3 ~ 0.294 0.2IT 0.234 0.387 0.316 0.312 0.404 0.858 0.233 0.213 0.269 0.644 0.275 0.265 0.289 0.600 0.829 0.756 0.885 1.265 0.417 0.300 0.833 1.667 0.618 0.690 0.671 1.197 0.682 0.693 0.750 1.237 0.345 0.342 0.325 0.494 0.317 0.331 0.413 0.873 0.298 0.323 0.357 0.743 0.320 0.332 0.357 0.679 .! GRASS' FORB 1985a 1986a 1987 1988 0.326 0.304 0.120 0.543 0.314 0.337 0.437 1.359 0.259 0.256 0.202 0.824 0.292 0.290 0.226 0.855 SANDSAGE 1985 1986 1987 1988 0.621 0.838 0.169 0.904 0.375 0.500 0.250 1.000 0.395 0.650 0.250 1.194 0.566 0.793 0.189 0.945 COMBINED 1985 1986 1987 1988 0.421 0.470 0.127 0.616 0.315 0.339 0.434 1.355 0.272 0.281 0.204 0.834 0.335 0.367 0.224 0.863 a1985 and 1986 were both pretreatment years. Table 8. Height·density relationships controls among years, Tamarack Prairie, Vegetation 1985-88 E Grass·forb Sands age Combined between the 1986 burns and their Colorado. 1986-88 1987-88 4.58a 5.19 3.60 33.88C 0.19 5.69b 2.16 0.45 0.05 4.68a 1.45 5.37a VALUE 4.69a 1l.15a 3.31 PAIRED .t. Grass-forb Sandsage Combined ap < 0.05. cr. < 0.025. < 0.001. bf 2.14 0.76 1. 59 �183 Increased precipitation received in 1986 and 1987 (Fig. 1) was the probable reason for faster recovery of grass-forb vegetation in the 1985 and 1986 burns and the increased standing residual over-winter into early spring 1988. The 1985 and 1986 fires were phenologically later than those in 1984 which also may have been a factor. How long this enhanced HDI will persist remains to be seen. Height-density of grass-forb vegetation increased from pretreatment to 1988 intervals by an average of 0.54 dm within combined 1985-86 burn sites compared to an average increase of 0.32 dm within their respective controls (paired~, f < 0.05). In contrast, the HDI of sandsage increased more (f < 0.05) within the 6 controls (0.63) than within the burns (0.34). Thus, analysis of total vegetation did not detect a significant (f > 0.05) pretreatment to 1988 change. From 1st-year post-treatment to 1988 within the combined 1985-86 burn sites, grass-forb vegetation increased 0.69 dm contrasted to an 0.33 dm increase within the controls (f < 0.001). The recovery of sandsage within burned sites (0.75 dm) compared with controls (0.58 dm) was less dramatic (f> 0.05). Total vegetation within the burned sites increased 0.71 dm from 1st-year post-treatment to 1988 intervals vs. 0.56 dm within the controls (f < 0.001). Thus, grass-forb vegetation responded favorably to controlled burns in both 1985 and 1986 in contrast to results of the 1984 burns. Crown Cover, Composition, and Frequency of Occurrence Sampling Point frame sampling, conducted from 1984 through 1986, was resumed in summer 1988 within the 3 sites burned in 1985 and 3 sites burned in 1986. The 1985 burns represented 17 transects (432 samples/transect) and 17 control transects, whereas the 3 1986 burns contained 19 transects and 19 controls. Data on species crown cover, composition, and frequency of occurrence per burn site and combined burns (or controls) varied (Tables 9-12). Numbers of transects were not distributed equally among burns either year and comparisons among burn or control sites should not be made. Ignoring sites, mean tallies per transect are presented for the 1985 burns (Table 13) along with statistical values of crown cover relationships (Table 14). Data for the 1986 burns and their controls are also presented (Tables 15, 16). A general review by pertinent species or groups is presented below. Blue Grama.--Evidence of enhancement of blue grama was apparent from both the 1985 and 1986 burns, but enhancement occurred in the 2nd growing season for the 1985 burns, whereas it occurred in the 1st growing season within the 1986 burns (Tables 13-16). Greater precipitation in 1986 than in 1985 (Fig. 1) may have been the primary difference. Mean crown cover/transect began converging from 1986 to 1988 indicating burning enhancement of blue grama was being overcome with time and that enhancement was not sustained. Findings from the 1985 and 1986 burns compliment findings from burn 1-84 for blue grama (Snyder 1988). Blue grama composition ranged from 9.4 to 19.2% and was recorded within nearly all transects (Tables 9-13). Needle-and-Thread.--Burns in 1984, 1985, and 1986 all had the same impact as they retarded crown cover through the 1st post-treatment growing season. Within the 1985 burns, retardation continued through the 2nd growing season even though 1986 was an especially good year for growth and seed production. Thus, enhancement from burning was not sustained over time. Needle-and-thread was one the most abundant grasses with composition ranging from 17.2 - 22.1% and occurred on 100% of the transects (Tables 9-13). �184 Table 9. Crown cover (point frame), species composition (%), and frequency of occurrence of vegetation within treatment transects on 3 sites burned in 1985, Tamarack Prairie, August 1988. Crown Cover Category/species Bare ground Dead vegetation Bouteloua gracilis Stipa ~ Sporobolus cryptandrus Calamovilfa longifola Andropogon hallii Agropyron smithii Aristida sp. Panicum virgatum Bromus sp. Cyperus & Carex spp. Artemisia filifolia Opuntia sp. 144 980 162 213 97 331 26 23 2 16 1 33 2 3 Total Comp. Freq. Occur. 205 1,586 140 306 160 433 79 155 179 658 150 97 80 35 85 20 2 83 2 191 1 11 82 15 528 3,224 452 616 337 799 190 198 4 83 16 14 306 16 12.58 17.15 9.38 22.24 5.29 5.51 0.11 2.31 0.45 0.39 8.52 0.45 0.94 1.00 1.00 0.94 0.76 0.88 0.12 0.12 0.06 0.24 0.82 0.29 226 16 4 10 6 1 6 8 2 17 6.29 0.45 0.11 0.28 0.17 0.03 0.17 0.22 0.06 0.47 0.35 0.12 0.24 0.24 0.12 0.06 0.12 0.24 0.06 0.12 5 4 19 220 1 10 1 1 1 1 1 1 0.14 0.11 0.53 6.12 0.03 0.28 0.03 0.03 0.03 0.03 0.03 0.03 0.12 0.12 0.53 0.76 0.06 0.35 0.06 0.06 0.06 0.06 0.06 0.06 Ambrosia psilostachya Artemisia ludoviciana Tradescantia occidentalis Evolvulus nuttalianus Erigeron sp. Psora lea tenuiflora Physalis subglabrata Thelesperma megapotimicum Ipomoea leptophylla Mentzelia nuda 65 14 3 Helianthus petiolaris Croton texensis Chenopodium album Conyza canadensis Euphorba sp. Lactuca sp. Plantago purshii Pepidium densiflorum Cryptantha sp. Cirsium sp. Cichorium intybus Unidentified 1 4 5 34 4 2 2 1 4 161 16 2 2 5 2 9 156 4 1 8 6 4 2 5 30 1 2 • �185 Table 10. Crown cover (point frame), species composItIon (X), and frequency of occurrence of vegetation within 3 1985 control sites, Tamarack Prairie, August 1988. Crown Cover Category/species ~ , Bare ground Dead vegetation Bouteloua gracilis Stipa .£Q!!!ill Sporobolus cryptandrus Calamovilfa longifola Andropogon hallii Paspalum stramineum Agropyron smithii Muhlenbergia sp. Aristida sp. Koleria cristata Panicum virgatum Bromus sp. ~ & Carex spp. Artemisia filifolia ~. filifolia (dead) Opuntia sp. Ambrosia psilostachya Tradescantia occidentalis Evolvulus nuttalianus Phlox andicola Lathyrus polymorphus Erigeron sp. Psoralea tenuiflora Physalis subglabrata Thelesperma megapotimicum Sphaeralcea coccinea Lygodesmia juncea Mentzelia nuda Asclepias sp. Helianthus petiolaris Tragopogan sp. Croton texensis Chenopodium album Eriogonum ~ Conyza canadensis Euphorbiasp. Lactuca sp. Plantago purshi i Pepidium densiflorum Salsola kal i Cirsium sp. Polygonum sp. 148 884 148 251 91 262 38 1 15 2 3 Total Compo Freq. Occur. 226 1,569 115 409 199 275 34 5 134 352 539 80 73 109 64 82 726 2,992 343 733 399 601 154 6 151 2 4 2 3 12 19 554 35 36 9.46 20.22 11.00 16.57 4.25 0.17 4.16 0.06 0.11 0.06 0.08 0.33 0.52 15.28 0.97 0.99 1.00 1.00 1.00 0.88 0.47 0.12 0.82 0.06 0.06 0.06 0.06 0.06 0.35 0.82 0.53 0.41 77 27 18 3 8 9 3 29 30 5 1 3 3 10 1 13 106 2 166 7 28 1 3 7 8 4 2.12 0.74 0.50 0.08 0.22 0.25 0.08 0.80 0.83 0.14 0.03 0.08 0.08 0.28 0.03 0.36 2.92 0.06 4.58 0.19 0.77 0.03 0.08 0.19 0.22 0.11 0.47 0.47 0.24 0.12 0.06 0.18 0.06 0.24 0.29 0.12 0.06 0.06 0.06 0.18 0.06 0.24 0.41 0.06 0.59 0.18 0.35 0.06 0.12 0.24 0.18 0.12 2 2 4 2 3 7 177 19 7 2 63 5 3 65 12 7 1 13 12 10 314 11 26 11 2 11 3 8 3 5 3 5 29 22 5 1 3 3 3 1 3 2 1 5 10' 1 12 103 163 1 26 1 2 6 7 4 2 1 2 �186 Table 11. Crown cover (point frame), species composition (X), and frequency of occurrence of vegetation within treatment transects on 3 sites burned in 1986, TamaraCK Prairie, August 1988. Crown Cover Category/species Bare ground Dead vegetation Bouteloua gracilis Stipa comata Sporobolus cryptandrus Calamovilfa longifola Andropogon hallii Agropyron smithii Aristida sp. Panicum virgatum Muhlenbersia sp. Bromus sp. Festuca sp. Cyperus & Carex spp. Artemisia filifolia ~. fi lifolia (dead) Opuntia sp. Ambrosia psilostachya Artemisia ludoviciana Tradescantia occidental is Phlox andicola Evolvulus nuttalianus Lathyrus polymorphus Erigeron sp. Psoralea tenuiflora Thelesperma megapotimicum Ipomoea leptophylla Mentzelia nuda Sphaeralcea coccinea Helianthus petiolaris Croton texensis Chenopodium album Conyza canadensis Euphorba sp. Lactuca sp. Pepidium densiflorum Cryptantha sp. Cirsium sp. Salsola Kali 426 1,041 321 291 130 263 133 21 2 3 Total Compo Freq. Occur. 127 791 69 151 123 270 11 46 338 1,277 418 288 108 339 16 125 5 2 3 22 891 3,109 808 730 361 872 160 192 5 2 2 3 3 30 19.20 17.35 8.58 20.72 3.80 4.56 0.12 0.05 0.05 0.07 0.07 0.71 1.00 1.00 0.95 0.95 0.53 0.68 0.11 0.05 0.05 0.05 0.05 0.26 49 1 20 278 3 21 6.61 0.07 0.50 0.58 0.11 0.21 296 62 8 1 19 4 34 3 4 3 3 56 7.03 1.47 0.19 0.02 0.45 0.10 0.81 0.07 0.10 0.07 0.07 1.33 0.63 0.16 0.26 0.05 0.37 0.05 0.53 0.11 0.16 0.05 0.05 0.37 13 17 73 129 7 2 1 1 1 1 0.31 0.40 1.73 3.07 0.17 0.05 0.02 0.02 0.02 0.02 0.16 0.21 0.58 0.63 0.21 0.11 0.05 0.05 0.05 0.05 2 3 8 227 2 1 2 115 3 22 181 62 3 1 7 4 16 3 1 3 3 31 2 4 3 1 6 22 73 1 1 13 9 47 53 5 1 2 3 12 11 7 2 • �187 Table 12. Crown cover (point frame), species composition (%), and frequency of occurrence of vegetation within 3 1986 control sites, TamaraCK Prairie, August 1988. Crown Cover Category/species Bare ground Dead vegetation Bouteloua gracilis Stipa comata Sporobolus cryptandrus Calamovilfa longifola Andropogon hallii Agropyron smithii Muhlenbergia sp. Aristida sp. Panicum virgatum E:yperus & ~ spp. Artemisia filifolia ~. filifolia (dead) Opuntia sp. • Ambrosia psilostachya Artemisia ludoviciana Tradescantia occidentalis Evolvulus nuttalianus Phlox andicola Lathyrus polymorphus Erigeron sp. Psora lea tenuiflora Physalis subglabrata Thelesperma megapotimicum Sphaeralcea coccinea Lygodesmia ;uncea Penstemon angustifolius Croton texensis Chenopodium album Eriogonum .!!D.D.!:!!!l Conyza canadensis Euphorbia sp. Lactuca sp. Pepidium densiflorum Cryptantha sp. Polygonum sp. Argemone sp. Unidentified 250 1,324 327 286 68 630 11 6 2 3 Total Compo Freq. Occur. 114 688 63 290 95 257 19 21 296 1,351 199 349 205 347 24 53 8 4 660 3,363 589 925 368 1,034 54 80 8 4 7 42 553 11 56 14.07 22.10 8.79 24.71 1.29 1.91 0.19 0.10 0.17 1.00 13.21 0.26 1.34 1.00 1.00 1.00 1.00 0.47 0.68 0.05 0.11 0.05 0.05 0.79 0.26 0.68 86 20 30 9 4 2 9 4 5 12 40 3 1 2.05 0.48 0.72 0.22 0.10 0.05 0.22 0.10 0.12 0.29 0.96 0.07 0.02 0.26 0.11 0.58 0.26 0.11 0.05 0.21 0.11 0.21 0.16 0.11 0.16 0.05 11 59 1 111 9 26 8 1 1 1 1 0.26 1.41 0.02 2.65 0.22 0.62 0.19 0.02 0.02 0.02 0.02 0.21 0.63 0.05 0.63 0.21 0.32 0.11 0.05 0.05 0.05 0.05 7 170 7 22 2 2 80 6 2 18 2 4 1 1 2 9 3 20 2 6 1 3 3 7 9 41 39 6 23 1 42 382 4 32 6 20 6 5 2 5 3 2 3 20 3 1 12 69 3 �188 Table 13. Crown cover (mean point frame tallies/transect) for selected species and species groups among pre- and post-treatment samples within combined 1985 burns and their controls, Tamarack Prairie, Colorado, 1984-88. Species/category Bouteloua gracilis Stipa comata Sporobolus cryptandrus Calamovilfa longifolia Andropogon hallii Agropyron smithii Selected warm season Total warm season Total cool season Artemisia filifolia Comb. peren. forbs combined annual forbs Bare ground Dead vegetation 19843 28.5 31.4 4.8 36.8 1.7 0.6 39.4 72.7 32.1 22.5 4.4 4.1 30.9 260.1 Treatment 1985 1986 21.2 27.7 21.1 54.5 5.1 1.8 61.8 104.0 29.5 8.2 3.4 2.8 110.3 170.1 20.6 48.2 21.1 58.8 9.6 10.4 70.5 112.2 58.6 11.9 5.7 3.5 90.7 145.1 1988 1984 26.6 36.2 19.8 47.0 11.2 11.6 63.1 109.5 47.9 18.0 17.5 15.5 31.1 189.6 30.8 31.3 8.1 33.1 2.1 2.2 35.9 74.8 33.5 24.7 2.9 3.6 42.2 249.3 Control 1985 1986 21.9 55.2 22.7 28.5 4.1 3.5 32.8 77.4 58.7 22.6 2.1 3.5 47.1 217.9 10.5 67.2 22.1 26.4 4.6 11.0 31.1 63.8 78.2 23.8 4.2 4.2 67.7 187.9 1988 20.2 43.1 23.5 35.4 9.1 8.9 44.6 88.2 52.0 34.6 12.7 20.9 42.7 176.0 apretreatment. Table 14. Crown cover relationships for selected species and species groups among years, 1985 burns, Tamarack Prairie, Colorado. £. Value Species/category 8415-85 84~-86 8411-88 Bouteloua gracilis Stipa comata Sporobolus cryptandrus Calamovilfa longifolia Andropogon hallii Agropyron smithii Selected warm season Total warm season Total cool season Artemisia filifolia Combined perennial forbs Combined annual forbs Bare ground Dead vegetation 0.11 15.99c 0.46 17.11c 0.68 5.91b 4.70a 0.64 20.77c 5.65a 4.27a 1.12 0.06 1.74 0.00 8.00c 6.98b 0.07 1.69 4.93a 19.02c 18.19c 16.99c 30.01c 0.18 0.10 92.20c 32.35c 37.20c 41.44c 2.98 15.96c 0.19 0.34 30.10c 14.74c 3.78 8.98c 0.14 11.35c 0.00 0.70 0.44 1.09 6.96b 22.7Oc 9.77c 3.06 0.06 0.09 3.80 3.68 ap < 0.05. 0.025. cp < 0.010. dpretreatment. bp < 85-86 85-88 86-88 6.02b 5.34a 0.35 3.41 0.17 4.67a 0.22 0.50 9.65c 5.09a 0.19 1.08 13.62c 4.65a 0.04 1.00 0.60 11.51c 2.98 2.14 5.87b 0.06 3.60 0.06 0.00 0.68 23.25c 10.77c �189 Table 15. Crown cover (mean point-frame tallies/transect) for selected species and species groups among pre- and post-treatment samples within combined 1986 burns and their controls, Tamarack Prairie, Colorado, 1984-88. Species/category 1984~ Bouteloua gracilis Stipa ~ Sporobolus cryptandrus Calamovilfa lonsifolia Andropogon hsllii Agropyron smithii Selected warm season Total warm season Total cool season Artemisia filifolia Comb. peren. forbs Combined annual forbs Bare ground Dead vegetation 37.3 42.5 10.3 26.9 1.9 28.9 76.5 46.1 18.1 7.2 5.2 39.5 234.2 Treatment 1985~ 1986 36.1 40.0 13.7 25.3 2.4 4.3 27.6 77.4 44.2 16.2 4.7 1.6 41.1 241.4 38.1 38.5 24.5 44.9 6.8 9.5 51.7 114.3 47.9 7.9 15.4 3.2 182.4 57.2 1988 1984 42.5 38.4 19.0 45.9 8.4 10.1 54.4 115.9 48.5 14.8 25.9 12.9 46.9 163.6 40.6 28.1 4.6 39.5 0.6 40.1 85.3 29.1 14.4 5.1 3.1 28.9 263.9 Control 1985 1986 34.6 52.1 14.8 33.8 1.1 1.8 34.9 84.3 53.7 19.3 3.0 1.3 32.6 226.6 18.0 61.5 15.3 37.2 1.2 5.3 38.3 71.6 66.8 22.6 6.3 2.3 46.6 204.8 1988 31.0 48.7 19.4 54.4 2.8 4.2 57.6 108.0 52.9 29.7 11.7 12.2 34.7 177.0 apretreatment. Table 16. Crown cover relationships for selected species and species groups among years, 1986 burns, Tamarack Prairie, Colorado. f. Value Species/category Bouteloua gracilis S t i pa £Q!llill Sporobolus cryptandrus Calamovilfa longifolia Andropogon hsllii Agropyron smithii Selected warm season Total warm season Total cool season Artemisia filifolia Combined perennial forbs Combined annual forbs Bare ground Dead vegetation ap < bi) < c< 0.05. 0.025. 0.010. Pretreatment. i. 6.94b 14.35c 0.06 1.40 9.27C 4.21a 0.01 0.40 0.16 85d-86 85d-88 8.93c 6.14b 8.11b 17.00c 7.67b 2.59 0.34 0.01 0.27 2.30 0.79 1.04 2.43 0.05 6.46b 3.19 0.09 1.04 2.61 20.83c 28.62c 2.33 21.16c 4.26a 0.15 180.89c 137.04c 86-88 3.01 1.07 3.41 6.38b 0.61 1.74 8.44b 11.77c 2.18 5.39a 0.40 0.17 13.42c 3.32 �190 Prairie Sand Dropseed.--There has been no marked evidence that fire impacted this species either positively or negatively based on crown cover sampling of the burns and their controls. First-year stimulation of seed production was evident but crown cover was little affected. Composition ranged from 8.6 to 9.4% in 1988 and this species occurred in nearly every transect. Prairie Sandreed.--This species responded favorably to fire with increased crown cover and seedhead production during the first growing season after both the 1985 and 1986 burns. Enhancement continued through the 2nd growing season within 1985 burns, however, by 1988, enhancement that had occurred was mostly over (Tables 13-16). Composition varied more widely than that of the other dominant grasses (16.6 - 24.7%) but sandreed occurred in nearly every transect (Tables 9-13). Sand Bluestem.--Although this less abundant species increased from 1984 through 1988 within both 1985 and 1986 controls, controlled burns showed evidence of stimulating crown cover growth and seed head production. The response appeared similar to that for prairie sandreed except the major crown cover increase within burns occurred during the 2nd growing season following the 1985 burns and during the 1st growing season following the 1986 burns. Increased precipitation in 1986 probably complimented stimulation by fire. Impacts of 1985 fire enhancement apparently ended by 1988 as treatment and control means converged. Divergent means continued through 1988 within the 1986 burns (Tables 13-16). Sand bluestem ranged from 1.3 to 5.3% of the total composition within the 1985 and 1986 burns in 1988 and frequency ranged from 0.47 to 0.76. Selected Warm Season Grasses.--This group, dominated by prairie sandreed, included sand bluestem and occasional occurrences of switchgrass and sand paspalurn. Response was similar to that of the dominant species with declines (1985 burn) or less increase (1986 burn) (f < 0.05) of crown cover from 1986 to 1988 in contrast to increases within the controls (Tables 13-16). Total Warm Season Grasses.--This group included all species within the selected warm season group plus blue grama and sand dropseed. Response to fire appeared dramatic for 2 years following the 1985 burn and for at least 1 year following the 1986 burn (sampling was not conducted in 1987). However, by 1988 the impacts of the burns were greatly diminished and sample means were converging (Fig. 7). Composition of total warm season grasses ranged from 41.5 to 52.4% (Tables 9-13) with approximately 50% occurring in 3 of the 4 sites. Western Wheatgrass.--This cool season grass was often going into dormancy by the time sampling was conducted. Thus, some variation among years was probably due to moisture and sampling time variances. Controlled burns possibly stimulated growth of western wheatgrass, but sample sizes were relatively small and the data are questionable. Data over time indicate the species, like sand bluestem, may be increasing with deferment from grazing, or during years of above average precipitation. Western wheatgrass ranged from 1.9 to 5.5% composition and occurred within 68 to 88% of the transects (Tables 9-13). Total Cool-Season Grasses.--Needle-and-thread dominated over western wheatgrass and other cool-season grasses. Western wheatgrass did not show the same response to fire as needle-and-thread. Other cool season grasses were too scarce to provide meaningful data. Combined cool-season grasses represented 22.1 to 24.6 of the total vegetative composition within the 1985 and 1986 burns and controls. �191 120 110 100 IU W (J) z « 0:: l<, 90 (J) W I CONTROL ......J I _j « l- 80 Z --' « w 2 70 ". • '" I I ,, ,, , I I I ,, I • 60 I I 50 84 86 85 87 YEARS Fig. 7. Crown cover (R tallies/transect) of total warm season grasses from 1984 (pretreatment) to 1985-88 (post-treatment) within the 1985 combined burns and controls, Tamarack Prairie, Colorado. Sand Sagebrush.--Favorable moisture apparently enhanced growth of sandsage during the last 3 years as crown cover means within 1985 and 1986 control transects increased. The 1985 and 1986 fires were cooler than those in 1984, but were conducted phenologically later in spring. They therefore, potentially stressed sage plants more since leafing (and use of energy reserves) was more advanced than in 1984. Rapid sage regrowth immediately after burns did not occur in 1985 and 1986 as it had in 1984. However, most plants survived and by 1988 had made considerable recovery. Sandsage was more prevalent within the controls, representing 13.2 to 15.3% of the composition compared to 6.6 to 8.5% within the burned sites in 1988. It occurred on 58 to 82% of the transects. Combined Perennial Forbs.--There was no evidence that burning enhanced combined perennial forbs within the 1985 burns (Tables 13-14). However, fire seemed to �192 stimulate increased perennial forb abundance within the 1986 burns which carried through to 1988 (Tables 15, 16). Since sample sizes were small and species were diverse, results are uncertain. Combined perennial forbs represented 5.4 and 6.0% of the composition within the controls but 8.3 and 11.7% in the burns in 1988 (Tables 9-13). In general, perennial forbs increased on both burned and control sites since 1985. Some of this variation was due to varying weather conditions that stimulated different species in different years. Western ragweed, a seed producing perennial potentially important to prairie grouse, increased from 68 to 522 between 1984 and 1988 in burned areas. The increase from 49 to 163 within the controls was less dramatic. Small sample sizes prevented detection of significant differences if they existed, but there was evidence that fire stimulated increased occurrence in both 1985 and 1986. Western ragweed is a warm season species that attained little if any growth prior to the burns. Combined Annual Forbs.--In combination, annual forbs increased on all treatments and controls from 1986 to 1988. Horseweed was extremely abundant in 1987 (Snyder 1988) and while less common in 1988, was still dominant among annuals (Tables 9-12). Generally, in 1987, it was most prevalent on unburned sites that possessed considerable litter accumulation. It may be an indicator of changing range conditions as a result of long-term grazing deferment. However, its increased occurrence within unburned 1985 and 1986 controls was not evident in 1988. Lambsquarter ranked 2nd and most other annuals ranked far behind. Impacts of fire on abundance could not be detected. Bare Ground and Dead Vegetation.--The amount of bare ground within both 1985 and 1986 burns declined to, or near pretreatment levels by 1988. However, the amount of dead vegetation was still considerably below pre-treatment levels (Tables 13-16). Revegetation Treatments Tillage-Reseeding.--The average HDI in early spring 1987 was 1.83 dm, but 1 year later, it had increased to 4.44 dm within 12 transects sampled among 19 revegetation strips seeded in 1985 (f < 0.001). In contrast, the mean combined vegetation HDI's within proximal 1985 and 1986 burn site controls in 1988 were 0.78 and 0.68 dm respectively. The reseeded strips thus provided HDI's 6 - 7 times greater than adjacent native range. Dramatic increases in standing residual resulted primarily because of favorable precipitation in both 1986 and 1987 (Fig. 1). Crown cover changes for major vegetation groups among years (Table 17) within the revegetation strips varied. Bare ground declined rapidly whereas dead vegetation, primarily standing residual within the crowns of switchgrass, increased dramatically. Occurrence of seeded tall warm season grasses remained nearly equal to that of the previous year, but totals of other grasses declined. Perennial forbs increased in contrast to 1987, but a major increase was not evident and the increase was distributed among several species. Switchgrass and bluestems comprised 84% of the vegetation compos~t~on in 1988, up from 77% in 1987. One of the 12 transects was included within a burn conducted by management personnel in spring 1988. �193 Table 17. Crown cover among vegetation 12 1985 revegetation transects, Tamarack Vegetation/category Bare ground Dead vegetation/litter Dominant native grasses Lesser grasses and sedges Seeded tall grasses Sandsage and cactii Perennial forbs Annual forbs I i groups from 1985 through Prairie, Colorado. 1988 within 1985 1986 1987 1988 804 212 217 79 311 6 22 76 812 156 226 37 475 10 12 278 372 183 21 833 8 2 31 120 620 98 13 829 15 20 13 a Tillage Renovation of Interseeded Tracts.--Evaluation of vegetation changes within a site previously interseeded (1981-92) and partially renovated (disked, harrowed, treated with atrazine herbicide) in spring 1986 (Snyder 1987) continued in 1988. Whereas the HOI of residual grass-forb vegetation increased from early spring 1987 to 88 within the controls, that within the treated transects increased much more dramatically (f < 0.025, Table 18). Sandsage, which was much less abundant within the transects, made considerable growth within both the treatments and controls. Combined vegetation within the renovated portion of the site yielded an HOI about twice that of the control starting into the 1988 growing season indicating favorable response, primarily of grasses, to renovation. Crown cover of bluestems and switchgrass (primarily interseeded) in combination increased steadily within the controls from 1985 through 1988. With renovation, the 1985-88 increase was markedly greater (f < 0.005, Table 19). In contrast, prairie sandreed, another deep-rooted tall warm season grass present prior to interseeding, did not respond to renovation (f> 0.10). However, the species was not negatively impacted by renovation (Table 19). Other more shallow-rooted native grasses, including blue grama, sand dropseed, and needle-and-thread, were markedly reduced by renovation, an impact that continued through 1988 (f < 0.005). Perennial forbs were not abundant prior to renovation and, although they may have been set back temporarily, they showed evidence of complete recovery by 1988 (Table 19). The interseeded bluestem-switchgrass mix had been planted in furrows, approximately 1.07 m apart. While native vegetation had been disturbed by interseeding, it had recovered quickly to provide competition to the young interseeded bluestem and switchgrass seedlings. Renovation in 1986 removed much of the competing shallow-rooted vegetation while suppressing annual forbs. Prairie sandreed, deep-rooted like the seeded grasses, was not retarded, but responded poorly to renovation. Management personnel have used the disk-harrow-herbicide technique on numerous interseeded tracts on the Tamarack Prairie from 1986 through 1988. In every instance dramatic release of tall warm season grasses was attained. It is probable that this renovation technique could be repeated with good results in future years on tall warm season grass stands. �194 Table 18. Mean height-density (dm) of grass-forb, sandsage, and combined cover from pre- (1986) to post-treatment (1987-88) intervals between a tillage-herbicide and untreated control within a previously interseeded location, Tamarack Prairie, Colorado. Year Tillage-herbicide Control Grass-forb 1986 1987 1988 0.382 0.314 l. 554 0.393 0.271 0.551 Sandsage 1986 1987 1988 l. 234 0.227 1.136 0.696 0.608 l. 268 Combined 1986 1987 1988 0.544 0.375 0.677 0.579 0.306 1.525 Table 19. Average crown cover/transect of selected species and species groups from pre- (1985) to posttreatment (1986-88) intervals within tillage-herbicide renovation of an interseeded site and it's control, Tamarack Prairie, Colorado. Species/group AndroP9gon/Panicum Calamovilfa longifolia Bouteloua, Stipa, etc. Perennial forbs 1985 13.7 11.3 17.6 1.5 Tillage-herbicide 1986 1987 24.2 12.4 5.4 0.6 36.3 13.8 6.4 1.2 1988 1985 Control 1986 1987 1988 39.8 11.2 6.8 1.6 16.4 10.7 17.1 2.2 16.6 10.6 16.8 1.7 22.1 10.2 19.8 3.4 24.6 8.1 20.4 1.7 .E Values 85-88 Andropogon/Panicum Calamovilfa longifolia Bouteloua, Stipa, etc. Perennial forbs ae_ < 0.005. 51.93a 2.03 38.83a 0.03 86-88 3.54 0.34 �195 Strip Spraying of Sandsage The HOI of grass-forb residual vegetation increased from 0.27 dm in 1987 to 0.61 dm in 1988. The HOI of standing dead sandsage also increased from 0.80 dm in 1987 to 1.13 dm in 1988. Since the sage was dead and slowly deteriorating, increased HOI was attributed to increased grass-forb growth within the sage plants. Pretreatment HOI's were not obtained within the spray site so enhancement of grasses due to the spray treatment is unknown. No difference was detected from 1987 to 1988 when the HOI of grass was compared with that within burn 3-85 and 3-86 controls. However, th~t comparison is extremely weak and significant change would not be expected within several years post-treatment. Grasses within the portion of the 1985 spray site that was burned in 1986 made dramatic recovery in 1987 with favorable precipitation, and provided excellent residual cover (HOI = 1.13 dm) in spring 1988 (Table 20). Growth from 1986 preburn to 1988 for 4 spray-burn transects, when contrasted to 14 controls, showed marked impact of burning (f < 0.001). Grass HOI increased a full decimeter from 1987 to 1988. Table 20. Mean height-density (dm) within the 1985 sandsage spray site during 1986 and 1987 post-treatment intervals and preburn (1986) to post-burn (1987) intervals within a portion burned in 1986, Tamarack Prairie, Colorado. Years N Transects Grass-forb Sands age Combined 1985 SANOSAGE SPRAY SITE 1986 1987 1988 2 11 11 0.246 0.272 0.613 SANOSAGE 1986 1987 1988 4 4 4 0.879 0.800 1.127 0.468 0.469 0.737 0.713 0.333 1.250 0.404 0.129 1.133 SPRAY - 1986 BURN 0.279 0.122 1.131 Point frame sampling of 11 transects within the 1985 spray site was conducted at a later date in 1988 and vegetation was advanced phenologically over previous years. Increased occurrence of needle-and-thread and prairie sandreed, and declines of both bare ground and dead vegetation probably resulted (Table 21). The decline in cheatgrass brome was not influenced by time of sampling. Live sage plants, if present, were not recorded but a steady increase in prickly pear since 1985 was recorded (Table 21). Perennial forbs remained at approximately the same level as in 1987 and, while still below 1985 pretreatment levels, were not significantly lower. �196 Table 21. Crown cover (point frame) of vegetation and ground cover within 11 random transects during pre(1985) and post-treatment (1986-88) spring intervals within the June 1985 herbicide-treated sandsage spray tract, TamaracK Prairie, Colorado. Vegetation/cover Bare ground Dead vegetation Perennial grass Bouteloua gracilis Stipa comata Sporobolus cryptandrus Calamovilfa longifolia Other perennial grasses Annual grass Bromus tectorum ~sp. Pretreatment 1985 1986 402 1,425 529 1,324 536 1,083 387 990 142 194 34 51 148 510 46 65 15 312 425 106 53 19 307 683 161 120 15 117 4 116 7 38 10 4 70 Post-treatment 1987 1988 o Artemisia filifolia (alive) A. filifolia (dead) 436 112 1 259 287 ° 210 Opuntia & Echinocereus spp. 44 50 62 83 Perennial forbs Ambrosia & Artemisia spp. Tradescantia occidentalis Lathyrus polymorphus Psoralea tenuiflora Evolvulus nuttalianus Phlox andicola Allium textile Mentzelia nuda Leucocrinum mont anum Penstemon angustifolius Thelesperma megapotimicum Cymopteris montanus Abronia fragrans 5 18 114 55 6 4 2 12 1 2 2 2 1 9 77 22 123 17 121 1 1 1 6 1 8 Annua l forbs Croton texensis Chenopodium album Pepidium & Lesguerella sp. Cryptanthia sp. Tragopogan sp. Plantago ~ Salsola Kali Conyza canadensis Lactuca sp. Unid. forbs Wildlife Occurrence 1 6 1 7 1 1 3 4 24 3 2 1 2 1 2 3 1 On the Tamarack Prairie Greater prairie-chickens were observed several times during spring 1988 within the Tamarack Prairie although searches for grouse were not conducted. This was the 1st year that this species, which is the primary focus of management and research efforts, has been observed frequently within the Tamarack Prairie. Management personnel and research assistants have located several leks along the south property line of the Prairie. Thus far, no leks have been found on the ungrazed Prairie itself. In addition to wildlife species �197 reported as occurring within the Tamarack reports, short-eared owl (Asio flammeus), and Cassin's sparrow (Aimophila cassinii) Prairie in previous progress great horned owl (Bubo virginianus), were observed in 1988. LITERATURE CITED Snyder, W. D. 1986~. Sandsage-bluestem prairie renovation. Job Progress Rep., Colorado Div. Wildl., Wi1dl. Res. Rep., Fed. Aid Proj. 01-03-045 (W-37-R). Apr. :475-498. 1986Q. Sandsage-bluestem praLrLe renovation. Job Progress Rep., Colorado Div. Wildl., Wildl. Res. Rep., Fed. Aid Proj. 01-03-045 (W-37R). Apr. :499-525. 1987. Sandsage-bluestem praLrLe renovation. Job Progress Rep., Colorado Div. Wildl., Wi1dl. Res. Rep., Fed. Aid Proj. W-152-R. Apr. :331-356. 1988. Sandsage-bluestem praLrLe renovation. Job Progress Rep., Colorado Div. Wildl., Wildl. Res. Rep., Fed. Aid Proj. W-152-R. In Press. Prepared by Warren D. Snyder Wildlife Researcher ��Colorado Division of Wildlife Wildlife Research Report April 1989 199 JOB PROGRESS REPORT State of: Colorado Project: W-152-R Upland Bird Research Work Plan: 21 Job Title: Accipiter Nest-site Selection and Foraging Behavior shinned Hawks in Mature Aspen and Conifer Habitats Period Covered: Author: Job 01 January 4__ through 31 December of Sharp- 1988 Suzanne M. Joy Personnel: C. E. Braun, R. W. Hoffman, Colorado Division of Wildlife; R. T. Reynolds, U.S.D.A. Forest Service; R. L. Knight, S. M. Joy, Colorado State University ABSTRACT Eleven active accipiter nests were found between 27 May and 26 August. These included nests of 8 sharp-shinned hawks (Accipiter striatus), 2 Cooper's hawks (a. cooperii), and 1 northern goshawk (a. gentilis). Nest cover types ranged from total conifer (Picea and Abies spp.) (1 goshawk, 1 sharp-shinned hawk) to total quaking aspen (Populus tremuloides) (1 Cooper's hawk), with 4 sharpshinned hawk nests in insular stands of conifers surrounded by aspen. The remaining nests were in mixed aspen-conifer habitat. A live great horned owl (Bubo virginianus)/mist net array was most successful of 5 methods used as 7 sharp-shinned hawks were captured. Six hawks (4 females, 2 males) were radiomarked with back-pack mounted transmitters (~4% of body weight). Hawk activity was monitored between 0600 and 2000 MDT. Percent slow pulse rate differed (£ = .0005) between sexes during the fledgling period. Prey deliveries were monitored at 4 sharp-shinned hawk nests (3 in aspen, 1 in conifer). Mean time between deliveries was greatest during incubation at all nests and least_during the nestling period at 3 of 4 nests. Inter-prey delivery periods .d.idnot differ (~';;"0.76). between habitat .types. Remains of prey plucked at nest si t es were collected from the 11 accipiter nests. ��201 ACCIPITER NEST-SITE SELECTION AND FORAGING BEHAVIOR OF SHARP-SHINNED HAWKS IN MATURE ASPEN AND CONIFER HABITATS Suzanne M. Joy P. N. OBJECTIVES The objectives of this study are to (1) measure sharp-shinned hawk daily activity budgets in mature aspen and conifer overstory types, (2) compare sizes and taxa of prey delivered to sharp-shinned hawk nests in the 2 overstory types through the breeding season, (3) measure time between prey deliveries to nests in relation to breeding phenology and prey abundance in each overstory type, and (4) describe nest-site characteristics of sharpshinned hawks, Cooper's hawks, and goshawks. SEGMENT OBJECTIVES 1. Locate accipiter nests in Gunnison 2. Measure sites. 3. Capture, radiomark, shinned hawks. 4. Monitor sharp-shinned 5. Monitor frequency 6. Collect and sort accipiter 7. Begin reconstruction 8. Compile vegetative and topographic and Grand Mesa National characteristics and obtain morphological forests. at accipiter measurements nest from sharp- hawk daily activity. of prey deliveries to sharp-shinned hawk nests. prey remains. and identification of prey remains. and analyze data and prepare progress report. DESCRIPTION OF STUDY AREA Portions of Gunnison and Grand Mesa National forests in Gunnison, Delta, and Mesa counties were selected for study. The study area was bordered on the west by Colorado Highway 65, northwest by Colorado Highway 330 extending east to the border between Pitkin and Gunnison counties (northern border), east by the Ruby Range, and south by the southern border of Grand Mesa National Forest. Climax aspen stands and conifer communities exist between 2,750 and 3,200 m. Large (> 500 ha) and small (100-200 ha) stands of mature aspen and conifer were inventoried for presence of accipiter hawks. Principal conifers were subalpine fir (a. lasiocarpa), Engelmann spruce (f. engelmannii), and blue spruce (f. pungens). At lower elevations, forests contained large (100500 ha) clearings with low shrub, forb, and grass cover. Understory vegetation types varied greatly depending on elevation, slope, and aspect. Dominant forbs included butterweed groundsel (Senecio serra), white �202 geranium (Geranium richardsonii), Barbey larkspur (Delphinium barbeyi), whiteflowered peavine (Lathyrus leucanthus), and monkshood (Aconitum co1umbianum). Prominent low-growing plants included elk sedge (Carex geyeri), wild strawberry (Fragaria ova1is), yellow prairie violet (Viola nuttal1ii), and fringed brome (Bromus ciliatus). The shrub component consisted primarily of snowberry (Symphoricarpus spp.), common chokecherry (Prunus virginiana), and true mountainmahogany (Cercocarpus montanus). Plant names follow Weber (1976). Sheep and/or cattle grazing occurred on much of the area. All study sites had been grazed by the end of the field season. METHODS Ground searches for accipiter nests began in May and were periodically conducted throughout the 1988 breeding season. Mature aspen, conifer, and mixed aspen-conifer habitats near ephemeral (drainages, marshes) and permanent (streams, creeks, ponds) sources of water were searched. Search areas were selected from aerial photographs (1:24,000) and topographic maps (7.5 min series). Stands possessing characteristics of accipiter nest sites in western montane forests (Shuster 1976, Reynolds et al. 1982, Moore and Henny 1983, Fischer and Murphy 1987) were selectively searched, although "unsuitable" stands were also explored en route to more "suitable" stands. In addition, 4 sharp-shinned hawk nest sites occupied in previous years (1 in 1986 and 1987, 1 in 1986 but not 1987, and 2 in 1987) (R. Reynolds, pers. commun.) were examined. Active nest sites were identified by the presence of a nest, breeding pair, or plucking areas with recent signs of activity (e.g., feathers, fur, casts, feces, and molted accipiter feathers). Vegetation and physiographic characteristics were measured at nest sites of breeding accipiters. Habitat variables measured included: 1) nest tree species, height, and diameter at breast height (dbh) , 2) nest-site elevation, aspect, canopy cover, and slope (the latter 3 measured by point-centered quarter method) (Cottam et al. 1953), 3) nest height and directional exposure, 4) distance to water,S) 2-3 species of dominant understory vegetation, and 6) average stand age. Stand age was calculated by coring the nest tree and 3-4 randomly selected trees in the nest site, then counting and averaging the core rings (+10 yr for tree growth to breast height). Sharp-shinned hawk nest sites occurring in pure, mature aspen, and conifer cover types ~ 2 km from another cover type were used for radiomarking, because foraging sharp-shinned hawks generally do not travel> 1.5 km from the nest (Reynolds 1983). Visual observations and/or radiolocations (% per cover type) were used to confirm that hawks were foraging predominantly (~ 90%) in the designated cover type. I used variations of trapping methods described by Bloom (1987) to identify the most effective technique for capturing sharp-shinned hawks. Techniques included: 1) mist nets and dho-gaza traps placed 0.5-2.0 m upslope from plucking posts in flight paths of adults returning with prey, 2) prey (2-weekold chickens) tethered 0.3-0.7 m from the center of mist nests and dho-gazas at several locations in and around the nest site, 3) a ba1-chatri baited with -2-week-old turkey poults placed near adult perches in the nest site, 4) a plastic great horned owl decoy with tape-recorded call placed within 0.3-1.0 m from a dho-gaza, and 5) a live great horned owl tethered at the nest site 0.5 �203 m downslope from V-arranged mist nets. Nets and traps used in techniques 1-4 were set between 0530 and 0630 or 2000 and 2100 MDT and checked every 30-40 minutes during daylight hours. The live owl/mist net array was installed before 0430 (sunrise was at -0530) and monitored continuously from a blind. Immediately after capturing a hawk, it was weighed, measured (wing chord, tail length, stage of molt), classified to age and sex, leg banded, and fitted with a back-pack mounted (Kenward 1985) lithium-powered transmitter with a variable pulse rate. Transmitter packages weighed 4 and 7 g, which were ~ 4% of male and female body weights, respectively. Pulse rates were regulated by mercurytilt switches that responded to changes in posture. Horizontal postures resulted in a rapid pulse (2.5 beeps/sec for females, 3.5 beeps/sec for males), whereas vertical postures produced a slow pulse (1 beep/sec for females, 2 beeps/sec for males). Activity (flying, perching, preening, plucking, eating) was monitored indirectly by recording changes in signal rate. Efforts were made to distinguish flight from other rapid-pulse producing non-flight activities (preening, eating, plucking prey) by comparing signal patterns with data gathered while simultaneously observing sharpshinned hawks. Focal-animal sampling (Altmann 1974), where an animal is intensively observed for a given period of time, consisted of radio-monitoring each hawk during every other 10-minute period in a day. Sampling occurred between 0600 and 2100. When a transmitter signal was lost, the next 10-minute sequence began as soon as the signal was audible again, regardless of the previous monitoring sequence. Systematic monitoring in this manner was later abandoned due to sampling problems (e.g., bird's transmitter signal beyond range, equipment failure) that resulted in inconsistent and scarce monitoring periods. Signals were then monitored from the time found until lost. To learn whether periods of data collection were balanced through the day, total minutes monitored were plotted against time of day (hour) for each nest, sex, and stage of breeding phenology (nestling, fledgling, independence). Dependent variables (percent slow pulse rate, mean length of rapid and slow pulse rates, number of pulse changes per minute) were plotted as a function of time of day for nest, sex, and stage of breeding to learn whether response variables reflected monitoring intensity. A multivariate analysis of variance (~iner 1971) (MANOVA 1) was used to test for differences between nests, sexes, and hour (0600-1900) for the response variables (percent slow pulse rate was normalized using arcsin transformation). All response variables were weighted by total time monitored. No test for differences among stages of breeding phenology was made because data sufficient to run MANOVA 1 were obtained only from the fledgling period. Differences in activity between cover types (aspen and conifer) were not tested because only 1 nest in each habitat was monitored during the fledgling period. If variable cover type had been used in the model, any significant cover effect may have been due to nest differences instead. To learn whether manipulating the time variable would affect MANOVA results, the above model was used after consolidating the independent variable 'time of day' into 3 time blocks (0600-1000, 1100-1400, and 1500-1900) (MANOVA 2). Prey deliveries to 4 sharp-shinned hawk nests (3 in aspen, 1 in conifer habitat) were recorded through the breeding season. Prior to and after radiomarking, a nest-site observer was used to record deliveries by non- �204 radiomarked birds; however, prey deliveries by radiomarked birds could be ascertained from signal strength and direction and from instantaneous begging by the young (late-nestling and fledgling stages of breeding only). The following criteria were used to distinguish prey deliveries from non-prey related nest visits by non-radiomarked birds: Incubation 1. Male entered the nest site and called faintly ("keekee"). Female flew to meet him and returned 5-7 minutes later, perched near the nest, and cleaned her bill. 2. Male entered the nest site, female flew to meet him while begging persistently (long, whining "screech"). Flutter of wings coming together was heard (pair was presumed to have exchanged prey in air, as was occasionally observed). Female returned to the nest -5 minutes later. 3. Male entered the nest site and called faintly. Female flew in the direction of the male and was later observed plucking and eating prey. 4. During a cold, light rain or when another raptor was near the nest site, the female occasionally refused to leave the nest when the male arrived with prey and called faintly. These events were recorded as prey deliveries because delivery was attempted. Nestling 1. 1-3 above. 2. Female begged persistently from near the nest, male (calling faintly) entered the site. Multiple, intense begging calls were heard from the nest. Then all begging stopped. 3. Male entered the nest site and called faintly, 2-5 minutes later female flew to the nest and fed the young. 4. Male or female flew to the nest while young begged loudly. After the parent flew off, the young wrestled with and appeared to be eating something. Fledgling 1. Parent and fledgling met and appeared to touch talons in the air. fledglings later chased the fledgling that met the parent. Other 2. Fledglings, begging intensely, chased a parent that flew through the nest site. Later 1 fledgling was seen eating prey. 3. Begging by fledglings, which had been periodic and slow, suddenly increased in intensity and speed. Moments later, a fledgling flew through the nest site with prey in its talons. Siblings chased the fledgling. �205 I tested for differences in mean time between prey deliveries among stages of breeding phenology (incubation, nestling, fledgling) using a 3-factor ANOVA (Winer 1971) with habitat type, nest within habitat type [nest (habitat)], stage of breeding, and all possible interactions as independent variables. Nest (habitat) and nest x breeding phenology (habitat) were used as random effects error terms. The response variable was normalized using log transformation. Early in the study, efforts were made to identify incoming prey by taxon and/or species. However, prey composition was difficult to ascertain because food items were either plucked beyond recognition before being delivered or were transferred from male to female (or parent to young) in locations beyond view. Consequently, efforts to identify prey were abandoned. Remains of plucked prey and regurgitated pellets were collected from the 4 sharp-shinned hawk nest sites approximately once a week as well as prior to and after monitoring periods, and every 2-3 weeks at all other nests. Prey are currently being reconstructed, identified, and counted following Reynolds and Meslow (1984), wherein uneaten remains (retrices, remiges, bills, and feet of birds, and fur, tails, skull fragments, and feet of mammals) are matched and compared with museum specimens. RESULTS AND DISCUSSION Nest Sites Approximately 11, 5, and 5 days were spent searching for accipiter nests in contiguous conifer, aspen, and mixed aspen-conifer stands, respectively. Eleven nests were found between 27 May and 26 August, including 1 goshawk in conifer habitat, 2 Cooper's hawk in aspen and mixed habitats, and 8 sharpshinned hawk in aspen (4), mixed (3), and conifer (1) habitats. Of 4 sharpshinned hawk nest sites occupied in previous years, 1 was reused. Eggs in 4 sharp-shinned hawk nests hatched between 5 and 12 July (the remaining nests were found post-hatch). Assuming a 30-32 day incubation period (Reynolds and Meslow 1984), incubation was initiated in early June and egg laying in late May. Fledging at the 4 nests occurred between 22 July and 12 August. An average of 3 (SD - 0.5) fledglings were observed at the 8 sharp-shinned hawk nest sites. Some young may have dispersed or been depredated before being counted. All sharp-shinned hawk nesting attempts were successful. Nest-site characteristics were measured at 8 of 11 nests (Table 1). Time constraints precluded measuring all nests. Those not completed will be measured in 1989. Accipiter nest sites ranged from 2,180 to 2,897 m in elevation and 3 to 180 m from water. Goshawks nested closer to water and higher in elevation than other accipiters; however, conclusions based on 1 nest are weak. Preliminary evidence also suggests that sharp-shinned hawks nest farther from water and use smaller, shorter trees, and steeper slopes than congeners. Most sharp-shinned hawk nests (83%) were in mature (70-120 yr-old) stands; 1 nest (17%) was in a young (50 yr) stand. The goshawk and Cooper's hawk nested in stands of approximately 116 and 85 years in age, respectively. �206 Table 1. Nest-site, western Colorado. nest-tree, and nest characteristics Nest-site characteristics Variable Distance to water (m) Slope (%) ~ SD ~ SD Elevation, Accipiter m Sharp-shinned 2,180-2,861 hawk of accipiters 80 66 45 28 in Aspect NE (2), (n) NiJ (4) (n ~ 6) Cooper's hawk (n = 1) (n - 1) Goshawk 2,655 160 26 NE 2,897 3 17 SW Nest-tree characteristics Variable Height (m) Spec i.es" ~ SD Accipiter Sharp-shinned hawk (n - 6) Cooper's hawk (n - 2b) (n - 1) Goshawk Sharp-shinned = Goshawk hawk 30 10 1, 4 26 45.6 4 26 47.5 6 13 100 1 Exposure NE (2), SE (3) , S (1) 6) Cooper's 7 Nest characteristics Variable Height (m) Canopy cover (%) ~ SD ~ SD Accipiter (n 19 1 (3), 2 (2), 3 (1) Dbh (cm) ~ SD hawk (n - 1) (n - 1) 17 92 18 100 SW a1) (Abies 1asiocarpa), 2) (Picea enge1mannii), 3) (E. pungens), 4) (Populus tremu1oides). bNest-tree species was the only characteristic measured at 1 nest. �207 Nest height was greater for goshawk and Cooper's hawk nests than those of sharp-shinned hawks. Canopy cover at the nest tree did not appear to differ among species. Nest tree species and aspect, nest exposure, and cover type at the nest and surrounding the nest site varied among nests. However, sharpshinned hawks consistently nested in conifers (67% were in small, insular stands of conifers surrounded by extensive stands of aspen), whereas the goshawk nested in aspen and the Cooper's hawks in aspen and conifers. Dominant understory vegetation varied among nest sites. Sharp-shinned hawk nest-site characteristics were qualitatively compared with measurements of sharp-shinned hawk nest sites in other western states (Table 2). Sharp-shinned hawks in Colorado appear to nest in trees similar in height and dbh to nest trees in Oregon, but in taller and larger trees than in Utah. Nest height in Colorado was also within range of heights reported for birds in Oregon; however, sharp-shinned hawks in Colorado generally nested much higher in trees than did sharpshins in Utah. Differences between site characteristics of birds in Utah versus Colorado and Oregon may be the result of more deciduous trees at nest sites in Utah compared to primarily conifers at nest sites in Colorado and Oregon (Fischer and Murphy 1987). Table 2. Nest characteristics of sharp-shinned hawks in Colorado, Oregon, and Utah. Nest tree Nest sitea Canopy closure ~%l ~%l SO SO ~ Nest height ~ml so ~ Height ~ml SO ~ O~ ~ ~cml SO 12 7 19 7 30 10 38 18 96 4 This study 18 6 27 6 41 21 26 9 80 90 14b 9c Reynolds et al. (1982) E Oregon n = 10 13 4 11 6 23 13 25 16 ~ 24 Reynolds et al. (1982) NE Oregon n = 15 8 3 29 14 25 19 98 2d Moore and Henny (1983) Utah 3 18 4 19 6 82 7 Fischer and Murphy (1987) Location Colorado n= NW Slo~ ~ 6 Oregon = 6 n n Source = 9 9 2 aMeasured using (1) point-centered quarter method in this study and Reynolds et al. (1982), (2) point sampling along transects in 4 cardinal directions in Moore and Henny (1983), and (3) variations on methods 1 and 2 igs:~~~e;r~~h~U~~Y3~1987). COld growth, n =-2. dNest tree only. Slope gradient appeared steeper and canopy cover more dense at nest sites in Colorado than at nests in Utah and Oregon, with the exception of canopy cover reported by Moore and Henny (1983), which was for the nest tree only. In the studies examined, a majority of sharp-shinned hawks nested in sites with northerly aspects. Nests in Colorado were usually built on southeasterly exposures (Table 1, Fig. 1); however, in Utah and northeastern Oregon nest placement was random. Reynolds et al. (1982) reported a tendency towards southeast to southwest nest placement for hawks in northwestern and eastern Oregon. �208 w Fig. 1. Colorado. Directional exposure • of sharp-shinned • • • • • hawk nests in western Trapping Trapping occurred between 14 June and 24 July. Trapping techniques 1-4 (discussed in methods) were unsuccessful. When nets were placed in the path of ground-based plucking areas, the hawks plucked their prey in nearby trees. Prey tethered in front of mist nets were sometimes attacked, killed, and plucked, but hawks avoided the nets when approaching and leaving the site. Male sharp-shinned hawks appeared interested in the bal-chatris baited with turkey poults (observing and/or swooping by it), but only 1 male made contact with the trap. He escaped as I approached. No hawk reacted to the plastic great horned owl decoy. These trapping methods were used alone or in combination for nearly 6 weeks. Several possible unsuccessful: explanations exist for why capture methods 1-4 were 1. Hawks saw the nets and avoided them. 2. Nesting pairs were unaccustomed to hunting near the nest and suspicious of immobile, conspicuous prey placed there. Natural prey were relatively scarce at nest sites, due to either hunting pressure or nest-site avoidance. Consequently, hawks may have learned (or evolved) to hunt away from the site. �209 3. Prey used behavior. (domestic 4. The plastic fowl) were not appropriate owl decoy was not realistic stimuli and, therefore, for hunting posed no threat. Seven sharp-shinned hawks (4 females, 3 males) were captured in 4 days using a live great horned owl with V-arranged mist nets. Six were radiomarked; 1 male escaped after being weighed and could not be retrapped. During trapping, if the owl was seen from the nest, the female attacked it at first light (-0545). Males usually did not see the owl until the second prey delivery of the morning (0600-1000). One nest was in the early-nestling and 3 were in latenestling stages of breeding phenology when trapping occurred. Most birds were processed and released within 30-45 minutes of capture. Wing chord and tail length measurements were similar to those reported by Mueller et al. (1979) for birds trapped in Wisconsin and to wing chords of museum specimens reported by Storer (1966). Females had larger wing chords, longer tails, and weighed nearly 77% more than males (Table 3). Stage of molt varied among birds, but females tended to be farther along in their molt sequence than males. All radiomarked birds flew well upon release. Table 3. Measurements of sharp-shinned Male SD Measurement Weight, g Right wing chord, em Left wing chord, em Tail length, em Sharp-shinned hawks captured 103.3 17.8 17.7 3.1 0.3 0.1 14.6 in Colorado, R 3 2 2 1 182.5 20.7 20.8 15.5 1988. Female SD 8.9 0.5 0.3 0.9 on 4 4 4 4 Hawk Activity Six birds were radiomarked, including 2 nesting pairs (1 in aspen, 1 in conifer habitat) and 2 females in aspen habitat (1 female lost her transmitter after 1 week and could not be retrapped). At the latter 2 nests, I was unable to trap 1 male and could not retrap the other after it escaped. Data were collected only for the fledgling period at 2 nests (where the pairs were radiomarked) and the nestling period at 1 nest (female only) because birds were radiomarked late in the breeding season and nests were at different stages of breeding phenology. At the nest where data were obtained for the nestling period only, young fledged nearly 10 days later than other nestlings. In addition, the female lost her transmitter prior to or just after the young fledged (I believe she removed it while preening). All females lost their transmitters before the end of the field season. Females were perhaps better able to weaken and remove transmitter harnesses because they devoted less time to foraging than males and more time to preening (S. Joy, pers. observ.). Presently, I am designing a more durable transmitter harness that will weaken and falloff within 2-3 months of placement on the bird. �210 Initial efforts to distinguish flight from other non-flight, rapid-pulse producing activities were abandoned. Only 1 female was observed and simultaneously radio-monitored during the field season. Generalizations made from her behavior to the sharp-shinned hawk population would be biased. Other females were more secretive and male visits to the nests were too brief to allow behavioral observations. Plots of total minutes monitored per hour of daylight for each nest, sex, and stage of breeding phenology (nestling, fledgling, independence) revealed that most periods of data collection were not evenly distributed through the day or within categories. Lack of balanced categorical sampling was due to (1) an inability to trap hawks until late in the breeding season and (2) female transmitter removal, whereas lack of balanced monitoring through the day was due to (1) an inability to receive a bird's signal at certain times of day, (2) receiver failure, and (3) abandonment of systematic sampling. In the 1989 field season, charts of telemetry sampling effort will be maintained to ensure balanced sampling. Response variables plotted against time of day revealed that response values reflected monitoring intensity. Dependent variables were therefore weighted by total time monitored. Due to sampling problems and the monitoring scheme change, 20-46% of the total time hawks were monitored consisted of incomplete signal bouts. That is, the first and last signals received during a monitoring period were often truncated. As a result, the response variables (mean length of rapid and slow pulse rates) were biased by the length of time a hawk was monitored. During the 1989 field season, telemetry equipment and techniques will be improved so that fewer signals are lost, and the monitoring schedule will be altered (each bird will be monitored intensively for 3-hour periods; Appendix A) so that unbiased estimates of the variables will be obtained. Sex contributed significantly to MANOVA 1 for dependent variables (1) percent slow pulse rate and (2) mean length of slow pulse rate (Table 4). During the fledgling period, females spent more time perching than males (Table 5). No other main effects or interactions were significant. Consolidating time of day into 3 time blocks improved the strength of the model (MANOVA 2, Table 6) -- sex and nest, as well as some interactions involving sex, nest, and hour contributed significantly to the model. Blocking in this fashion, however, is subjective. One may achieve desired levels of significance by simply manipulating the size of the time block. Neither model is adequate given that monitoring period lengths varied greatly during 1988. Ideally, length of the time variable should be determined by a monitoring period length that is constant for all birds (see monitoring schedule proposed for 1989, Appendix A). A revised MANOVA model, one in which the time of day variable is determined by monitoring period length, will be used to analyze data collected during the 1989 field season. Nest differences in activity appeared to be small (Table 4). If nests (1 of which was in aspen and the other in conifer habitat) were at all indicative of habitat differences, numerous samples (> 10; G. White, pers. commun.) in each habitat would be required to detect significant habitat differences in activity. It is not feasible to radio-monitor at a large number of nests. Consequently, null hypothesis Ha in the Program Narrative addressing habitat differences in activity will be abandoned. Differences between sexes (Tables 1 �211 Table 4. MANOVA 1 -- measures of sharp-shinned hawk activity with nest, sex, time of day (hour), nest x sex, nest x hour, sex x hour, and nest x sex x hour as the effects. Source of variation df ss I f Percent slow pulse rate&,b Nest Sex Hour Nest*sex Sex*hour Nest*hour Nest*sex*hour Error 5.75 25.98 9.93 0.33 6.09 13.72 15.06 49.56 1 1 l3 1 11 11 8 29 3.36 15.20 0.45 0.19 0.32 0.73 l.10 0.08 0.0005* 0.94 0.66 0.97 0.70 0.39 Mean length slow pu1sea r Nest Sex Hour Nest*sex Sex*hour Nest*hour Nest*sex*hour Error 1 1 13 1 11 11 8 29 32326.82 1189976.78 294573l. 89 241603.36 2653564.49 2574733.80 302l386.96 7783310.33 l.61 4.43 0.84 0.90 0.90 0.87 l.41 0.21 0.04* 0.61 0.35 0.55 0.58 0.24 l.89 2.78 0.23 0.60 0.40 0.20 0.60 0.18 0.11 0.996 0.45 0.94 0.996 0.77 0.04 3.80 0.44 0.78 0.60 0.63 0.89 0.84 0.06 0.94 0.39 0.81 0.79 0.54 Mean length rapid pu1sea Nest Sex Hour Nest*sex Sex*hour Nest*hour Nest*sex*hour Error 1 1 13 1 11 11 8 29 12889.63 18962.43 20519.43 4078.84 30296.54 15l32.32 32863.79 191166.98 Activity changes Imina Nest Sex Hour Nest*sex Sex*hour Nest*hour Nest*sex*hour Error 1 1 l3 1 11 11 8 29 *f < 0.05. aWeighted by total time monitored. bArcsin transformed. 1.40 120.17 18l.18 24.55 209.31 217.65 225.27 918.05 �212 Table 5. Activity measures for male and female sharp-shinned hawks during nestling, fledgling, and independence stages of breeding phenology. Phenology Nestling Sex (,n) Female (3) n Variable & SD Slow pulse Slow pulse Change rate 87.25 10.90 178.03 l.4l 11.22 7.23 231.95 l.22 Slow pulse Rapid pulse Slow pulse Change rate 5l.37 24.48 19.56 2.84 24.69 16.98 9.11 0.96 Slow pulse 90.00 8.28 117.67 l.41 9.36 7.23 115.16 0.76 Slow pulse Rapid pulse Slow pulse Change rate 64.94 19.25 42.30 2.45 15.23 13.69 4l. 32 l.08 Slow pulse R Rapid pulse R Slow pulse Change rate 76.45 28.20 134.38 l.64 23.29 34.19 162.40 l.61 Observ. 18 % & Rapid pulse R Male (2) 7 % R R Fledgling Female (2) 23 % ~ Rapid pulse ~ Slow pulse Change rate Male (2) 24 % R R Independence Male (2) 22 % , �213 Table 6. MANOVA 2 -- measures of sharp-shinned hawk activity with nest, sex, time of daya (hour), nest x sex, nest x hour, sex x hour, and nest x sex x hour as the effects. Source of variation df ss Percent slow pulse rateb,c Nest Sex Hour Nest*sex Sex*hour Nest*hour Nest*sex*hour Error 1 1 2 1 2 2 2 64 13.57 64.04 6.39 l.67 l.22 5.82 8.57 82.56 10.52 49.65 2.48 l.29 0.47 2.26 3.32 0.002* 0.0001* 0.09 0.26 0.62 0.11 0.04* 11.70 23.06 3.05 6.86 5.03 2.23 4.19 0.001* 0.0001* 0.05 0.01* 0.009* 0.12 0.02* Mean length slow pu1seb Nest Sex Hour Nest*sex Sex*hour Nest*hour Nest*sex*hour Error 1 1 2 1 2 2 8 64 2980305.86 5874996.18 1553897.32 1747778.94 2562400.62 1135556.72 2133912.13 16308256.04 Mean length rapid pu1seb Nest Sex Hour Nest*sex Sex*hour Nest*hour Nest*sex*hour Error 1 1 2 1 2 2 2 64 33211.87 53115.34 9155.07 15065.46 9707.89 3337.40 8017.61 398478.48 5.25 8.40 0.72 2.38 0.77 0.26 0.63 0.03" 0.005* 0.49 0.13 0.47 0.77 0.53 Activity changes /mi.n" Nest Sex Hour Nest*sex Sex*hour Nest*hour Nest*sex*hour Error 1 1 2 1 2 2 2 64 5.22 334.97 7l.19 76.62 155.22 12.48 53.72 1756.68 0.19 12.20 l.30 2.79 2.83 0.23 0.98 *~ < 0.05. aMeasured in 3 time blocks (0600-1000, 1100-1400, 1500-1900). bWeighted by total time monitored. cArcsin transformed. 0.66 0.0009* 0.28 0.10 0.07 0.80 0.38 �214 4, 6) and breeding phenology (Table 5), however, the following null hypotheses will be tested: appear strong. Therefore, Hol: Percent time spent perching, mean length of rapid and slow pulse rates, and activity changes/minute do not differ between sexes or among nests, males, females, or incubation, nestling, and fledgling stages of breeding phenology. Ho2: Diurnal patterns of activity do not differ between sexes or among nests, males, females, or incubation, nestling, and fledgling stages of breeding phenology. Prey Deliveries Frequency of prey deliveries were recorded during each nest visit. During early stages of nesting (incubation, early-nestling), males delivered prey to females at the nest site. The female in turn plucked and ate or fed the prey to the young. As the nesting season progressed into late-nestling and fledgling stages, females were often away from the nest and males delivered prey directly to the young. Females also provided the young with food at this time. Prey deliveries rates did not differ among stages of breeding phenology (Table 7). However, deliveries appeared least frequent during incubation at all nests, and most frequent during the nestling stage at 3 of 4 nests (Fig. 2, Table 8). This trend, which reflected greater foraging effort by males and the initiation of foraging by females, did not differ (£ - 0.76) between aspen and conifer habitat types (Table 7). At nest BR (aspen habitat), where prey deliveries were most frequent during the fledgling stage, young were fed primarily by a subadult male until post-fledging, when the female began to consistently supplement prey deliveries. The BR male seemed less competent than adult males at other nests, as he would often fail to return for long periods or returned without prey. Prey deliveries occurred over 3 times more often during the nestling than incubation period (Table 8). At 3 of 4 nests, prey delivery rates decreased slightly during fledgling stages (Table 8). Contributions by habitat x breeding phenology and nest x breeding phenology (habitat) interactions were insignificant (Table 7). Habitat differences in prey delivery rates will be difficult to detect given a reasonable number of nests; therefore, null hypothesis Hb of the Program Narrative addressing habitat differences will be replaced by the following: Ho3: Mean time between prey deliveries does not differ among nests or incubation, nestling, and fledgling stages of breeding phenology. Food Habits Prey remains were collected from all accipiter nest sites found (62 sampledays). Approximately 35 sample-days of prey have been sorted by probable species. Avian prey from 1 sharp-shinned hawk nest in conifer (9 sample-days) and miscellaneous avian samples from 3 nests in aspen (-5 sample-days/nest) were identified. Few prey species were cornmon at both nest sites (Table 9), 'Twice as many species appeared in the diets of birds in aspen than birds in conifer habitat (Table 9). This trend may be biased by the greater number of �215 Table 7. Three-factor ANOVA nests during the 1988 breeding type [nest(habitat)], breeding and all possible interactions and nest x brphen8(habitat) as Source for prey delivery rates to sharp-shinned hawk season with habitat type, nest within habitat phenology (incubation, nestling, fledgling), as the effects. The model uses nest(habitat) random effects error terms. of variation df MS .E f Habitat Error [nest (habitat) ] 1 4.18 0.02 0.15 0.10 0.76 Nest(habitat) Error [nest*brphen 2 6.11 0.16 0.26 0.60 0.58 (habitat) ] Brphen Error [nest*brphen 2 5.42 1. 20 0.28 4.35 0.07 (habitat) ] Habitat x brphen Error [nest*brphen 2 5.74 0.04 0.27 0.13 0.88 (habitat) ] 0.32 0.15 2.20 0.07 Nest x brphen Error aBreeding (habitat) phenology. 1.2 1.0 0: I 4 152 .~. RA BZ Clf BR ASPEN HABITAT D . CONIfER HABITAT 0.8 <, (f) w - 0: 0.6 W > .....J W 0 0.4 INCUBATION NESTLII'-t:; FLECGLING STAGES OF BREEDING PHENOLOGY Fig. 2. Prey delivery rates of sharp-shinned hawks to nests in aspen and conifer habitats as a function of breeding phenology. (Total hours of observation are at the top of each bar). Nest acronyms stand for the nearest source of water to which nests were located: RA (Ruby Anthracite Creek), BZ (Buzzard Creek), BR (Bird Creek), CW (Cow Creek). �216 Table 8. Mean time (min) between prey deliveriesa nests during the 1988 breeding season. Stage of breeding BZ RA :& SD :& SD to sharp-shinned Nestb BR SD :& :& 208 128 (9, 47) 185 (2, CW SD hawk Total SD :& Incubation Hrd) (n 350 (2, 149 17) 262 (3, Nestling (n, Hr) 66 (38, 78 48) 85 (21, 70 39) 93 (26, 70 54) 67 (12, 28 17) 70 77 (97, 158) Fledgling (n, Hr) 91 (18, 74 38) 175 147 (6, 28) 71 (16, 77 23) 94 (11 , 65 24) 95 87 (51, 113) C , 76 24) 113 21) 233 121 (16, 109) aMeasured to the nearest 5 min. bNest acronyms stand for the nearest source of water to which the nest was located: RA (Ruby Anthracite Creek), BZ (Buzzard Creek), BR (Bird Creek), CW (Cow Creek). cNumber of inter-prey delivery bouts. dHours of observation. nest sites sampled in aspen habitat or the unique plucking habits of birds in either habitat. Sorting and identification of mammalian and avian remains and regurgitated pellets continue. Differences in prey composition (size, taxon) between habitat types and stage of breeding phenology (incubation, nestling, fledgling) appear to exist; therefore, null hypothesis Hc of the Program Narrative will be replaced by the following: Ho4: Composition and number of prey delivered to nests do not differ between mature aspen and conifer habitat types or among incubation, nestling, and fledgling stages of breeding phenology. �217 Table 9. Avian prey at sharp-shinned hawk nest sites in aspen and conifer habitat during the 1988 breeding season. Common name Scientific lSapsucker spp. lDowny woodpecker 2Violet-green swallow 2Western flycatcher Mountain chickadee 2Black-capped chickadee Brown creeper House wren American robin lTownsend's solitaire Hernmit/Swainson's thrush Golden-crowned kinglet 2Solitary vireo Yellow-rumped warbler 2Yellow warbler lEvening grosbeak 2House sparrow Pine siskin lGreen-tailed towhee 2Vesper sparrow 2Song sparrow 2Black-throated sparrow Dark-eyed junco White-crowned sparrow 2Lincoln's sparrow name Sphyrapicus spp. Picoides pubescens Tachycineta thalassina Empidonax difficilis Parus gambeli atricapillus Certhia americana Troglodytes aedon Turdus migratorius Myadestes townsendii Catharus spp. Regulus satrapa Vireo solitarius Dendroica coronata Q. petechia Coccothraustes vespertinus Passer domesticus Carduelis pinus pipno chlorurus Pooecetes gramineus Melospiza melodia Amphispiza bilineata Junco hyemalis Zonotrichia leucophrys Melospiza lincolnii z. lConifer habitat only. 2Aspen habitat only. LITERATURE Altmann, J. 1974. Observational Behavior 49:227-267. CITED study of behavior: sampling methods. Bloom, P. H. 1987. Capturing and handling raptors. Pages 99-123 in B. A. G. Pendleton, B. A. Millsap, K. W. Cline, and D. M. Bird, eds. Raptor management techniques manual. Natl. Wildl. Fed., Washington, D.C. 420 pp. Cottam, G., J. T. Curtis, and B. W. Hale. 1953. Some sampling characteristics of a population of randomly dispersed individuals. Ecology 34:741-757. Fischer, D. L., and J. R. Murphy. 1987. Foraging and nesting habitat Accipiter hawks in Utah. Brigham Young Univ. Unpubl. ms. of �218 Kenward, R. E. 1985. Raptor radio tracking I. Newton, and R. D. Chancellor, eds. Paston Press, Norwich, England. and telemetry. Pages 409-420 in Conservation studies on raptors. Moore, K. R., and C. J. Henny. 1983. Nest site characteristics of three coexisting accipiter hawks in northeastern Oregon. Raptor Res. 17:65-76. Mueller, H. C., D. D. Berger, and G. Allez. 1979. Age and sex differences size of sharp-shinned hawks. Bird Band. 50:34-44. in Reynolds, R. T. 1983. Management of western coniferous forest habitat for nesting accipiter hawks. U.S. Dep. Agric., For. Servo Gen. Tech. Rep. RM-107. 7 pp. ________ . and E. C. Meslow. 1984. Partitioning of food and niche characteristics of coexisting Accipiter during breeding. Auk 101:761779. ________ , and H. M. Wight. 1982. Nesting habitat of coexisting accipiter hawks breeding in Oregon. J. Wi1dl. Manage. 46:124-138. Shuster, W. C. 1976. Northern West. Birds 7:108-110. goshawk nesting densities Storer, R. w. 1966. Sexual dimorphism and food habits American accipiters. Auk 83:423-436. Weber, W. A. 1976. CO. 479 pp. Rocky mountain flora. Winer, B. J. 1971. Statistical principles Hill Book Co., New York, NY. 907 pp. in montane Colorado. in three North Colo. Assoc. Univ. Press, Boulder, in experimental design. McGraw- 1 �219 Appendix A. Sharp-shinned hawk activity monitoring schedule for the 1989 field season. Sampling constraints used to develop the schedule: 1. 3 antennas, receivers, and towers (T) -- when 2T are used at 1 nest site, only 1T may be used at the following nest site because time periods overlap. 2. 3 field assistants (FA) -- the 3rd FA will monitor prey deliveries at nests where lT is used. 3. 5 nests (N), 1 breeding pair at each nest (MF). Activity N #T Sex N #T Sex N #T Sex N #T Sex N #T Sex 6-7 7-10 10-11 Hike/set-up Radio-monitor Prey remains 1 2T MF 2 2T MF 3 2T MF 4 2T MF 5 2T MF 9-10 10-13 13-14 Hike/set-up Radio-monitor Prey remains 2 1T MR 3 1T MR 5 lT F 1 lT 12-13 13-16 16-17 Hike/set-up Radio-monitor Prey remains 3 2T MF 4 2T MF 5 2T MF 15-16 16-19 19-20 Hike/set-up Radio-monitor Prey remains 4 lT MR 5 lT MR 1 1T F 2 1T F 17-18 19-21 20-21 Hike/set-up Radio-monitor Prey remains 5 2T MF 1 1T MR 2 2T MF 3 lT F Time r 4 lT F 1 2T MF 2 2T MF 3 1T Prepared by _yL"'_'.;;_<v.""¥~' t_"-U--( __ !.....;.1A~_~-,.'H,'-+/ _ Suzan~M. Joy ~ Graduate Research Assistant pR 4 2T MF Rrhe 1st sex monitored during each pair of 1T visits to nests will be randomly selected. I} pR ��221 Colorado Division of Wildlife Wildlife Research Report April 1989 JOB PROGRESS REPORT State of: Colorado Project: W-152-R Upland Bird Research Work Plan: 21 Job Title: Evaluation of Habitat Quality Eastern Colorado Period Covered: Author: Job 01 January 5__ on Conservation through 31 December Reserve Lands in 1988 Warren D. Snyder Personnel: T. Abell, C. Braun, D. Johnson, of Wildlife and W. Snyder, Colorado Division ABSTRACT Vegetation sampling to monitor habitat quality within 104 Conservation Reserve (CR) fields was initiated in early spring 1988 as part of a regional and national evaluation effort. Qf 104 sampled fields, 1 was seeded in 1986, 42 were seeded in 1987, 57 were seeded in 1988, and 4 remained to be seeded following establishment of satisfactory cover crops. Visual obstruction readings (VOR) during early spring (pre-greenup) had an average index of 0.39 dm, however, the VOR index in late June (nesting) was lower (0.20 dm). Planting 22 fields to grass and application of herbicides to numerous fields were primarily responsible for the lower June index. Vegetation canopy cover averaged 27% during early spring and 24% in late spring with annual forbs dominating. Average late June height was 9.7 cm. Rangeland and CR fields bordered 62% of the sampled CR field edges, thus, interspersion of covers was not enhanced in a majority of samples. The quantity, quality, and security (for nesting wildlife) of croplands associated with CR were monitored from early to late spring. Small grain stubble was nearly equal in quality to CR vegetation in early spring but much less secure. '.Green wheat attained VQR readings much greater than those within CR t.rac t stby early May .. Approximately 44% of the CR contracted during the 1st 4 signup interv-~ls'was outside . pheasant (Phasianus colchicus) range in eastern Colorado, whereas the remainder was within remnant (33%), low (17%), fair (2.5%), and moderate to good pheasant range (4%). Approximately 6% of the CR in eastern Colorado was within the range of northern bobwhite (Colinus virginianus) and about 16% lay within the range of scaled quail (Callipepla squamata). Mapping of CR distribution in relation to census routes continued. A general assessment of the values of the CR fields for several wildlife species was made. ��223 EVALUATION OF HABITAT QUALITY ON CONSERVATION RESERVE LANDS IN EASTERN COLORADO Warren D. Snyder INTRODUCTION The Colorado Division of Wildlife (CDOW) was represented at a meeting in Lubbock, Texas in February 1988, where personnel of the U.S. Fish and Wildlife Service's National Ecology Center (NEC) coordinated a regional meeting of state delegates. The purpose of the meeting was to; standardize data collection procedures among states, select representative species as a basis for evaluation, and provide a data base for use as part of a nationwide evaluation (by NEC) of the impacts of the Conservation Reserve Program (CRP) on wildlife. Subsequent to the meeting, NEC personnel provided each state representative with a randomized list of CR contracts stratified to sample fields retired from representative crop bases (wheat, corn, sorghum, etc.). In Colorado, the sample was restricted to the eastern "High Plains". Therefore, efforts by myself to obtain a stratified random sample of CR fields were cancelled. Three wildlife species, ring-necked pheasants, western meadowlarks (Sturnella neglecta), and cottontails (Sylvilagus spp.) were selected as the primary indicator species to monitor impacts of the CRP on wildlife within the Southern Great Plains Region in which Colorado was included. P. N. OBJECTIVES Determine the distribution and quantity of Conservation Reserve land in eastern Colorado in relation to the distribution of selected wildlife species, evaluate the quality of the vegetation on these lands for selected wildlife species, measure the response of selected wildlife species to the Conservation Reserve Program using existing annual surveys, and evaluate the impact of the Colorado Division of Wildlife's cost share program on cover quality. SEGMENT OBJECTIVES 1. Conduct evaluations of randomly selected CR tracts within eastern Colorado as part of a regional and national assessment of the Conservation Reserve Program being coordinated by the National Ecology Center (NEC) of the U.S. Fish and Wildlife Service. 2. Conduct intensive visual obstruction readings (VOR) from a stratified random sample of CRP fields and proximal controls. 3. Conduct intensive visual obstruction readings (VOR) from a sample of tracts cost-shared by the CDOW (for enhancement of cover quality) for comparison with CRP tracts not cost-shared. 4. Conduct pre- and post-treatment assessments of a randomly selected sample of CR fields based on Pat Rec and HSI models for a selected sample of wildlife species. �224 5. Determine the response of selected wildlife Reserve based on existing annual population 6. Compare the distribution wildlife species. 7. Compile Center, and analyze and prepare species to the Conservation surveys. of CRP in eastern Colorado data, submit necessary progress reports. to that of selected forms to the National Ecology METHODS The contract data base provided by the NEC was used during contacts of county offices of the Agricultural Stabilization and Conservation Service (ASCS) to obtain photocopies of CR contracts, aerial photos of fields, legal descriptions, and cropping history for a random sample of fields. An authorization letter form the State ASCS Director was obtained to facilitate release of data from county ASCS offices. Landowners and/or operators within 18 eastern Colorado counties representing 78 contracts and 104 CR fields were contacted to gain permission for access to fields and explain the purpose of the study. Pre-greenup vegetative sampling began in late March and extended into early April 1988. Data were recorded on a standardized data collection form provided by the NEC (APPENDIX A). The visual obstruction reading (VOR) Robel et. al (1969) was used to measure nesting cover quality obtained from CR fields during pre-greenup and late spring (mid to late Jun) nesting intervals (Table 1). The NEC requested 8 VOR samples per field however, 20 samples per field were obtained in Colorado to reduce within-field variability. Sampling was conducted on an accessible side of each field and extended into the field based on random selection of target objects (if available). The 1st sample was obtained 25 paces from the field edge and subsequent samples were obtained at 10 pace intervals thereafter. This VOR or height-density index (HDI) sampling technique has been used extensively in eastern Colorado (Snyder 1984Q, 1985, 1987, 1988) for herbaceous vegetation. The VOR index is based on procedures developed by L. M. Kirsch (unpubl. rep., U.S. Fish and Wildl. Serv., Jamestown, N.D., 1977). The NEC required a different (higher) interpretive reading of VOR for use in regional evaluations than the Kirsch method previously used in Colorado. To maintain continuity in Colorado sampling, the Kirsch method was retained as the basic method and an adjustment was made to correct it to the NEC method. Both VOR interpretations are listed for pre-greenup and nesting interval samples (Table 1). The Kirsch method is used in subsequent comparisons and discussions. The percent canopy cover of herbaceous vegetation within a 0.5 x 1.0 m (0.5 m2) Daubenmire frame was estimated twice at each of the 4 locations used in sighting toward the Robel pole. The 0.5 m2 plot lying immediately to the front and right of the sighting stick was sampled and the sample frame was flipped forward to obtain a 2nd sample. The 1st 8 samples were obtained at the 25 pace location and a 2nd group were tallied at the last location along the VOR sample transect. Perennial and annual grasses were distinguished from annual and perennial forbs. �225 Table 1. Visual obstruction readings (VOR, dm) among conservation reserve fields during pre-greenup and nesting intervals, eastern Colorado, 1988. N Sample fields Seeded in 1986 Seeded in 1987 Seeded in 1988 Before pre-greenup After pre-greenup Not seeded Overall mean Pre-~reenu~ VOR NECa Kirschb 1 42 57 35 22 4 104 Nestin~ NEC 0.00 0.46 0.38 0.09 0.84 0.01 0.65 0.39 VOR Kirsch 0.48 0.30 0.14 0.01 0.41 0.20 • aVOR required for use by the National bVOR developed by L. M. Kirsch. Ecology Center. A standard form provided by the NEC was used for data collection during the nesting interval (APPENDIX B). Procedures generally followed those used during the pre-greenup sampling interval. Nesting interval sampling was from mid to late June progressing from south to north across eastern Colorado. Sampling was conducted from early spring through early summer 1988 to provide data on quantity, quality, and security (for nesting wildlife) of land use types, including CR in eastern Colorado. Sampling through spring was nonrandom and restricted by time constraints, primarily to the eastern tier of counties. Pastures and rangelands were not included within spring samples but were included in the June samples. During mid- to late June sampling, roadside listings of cover types were obtained to 20 km from each sampled CR field during travel from one field to the next. Since CR fields were randomly selected, this provided a random distribution of associated cover types. RESULTS AND DISCUSSION Evaluation of Conservation Reserve Vegetation Quality Of the 104 fields sampled, 1 had been seeded in spring 1986, 42 fields were seeded in 1987, 57 were planted from winter through late spring 1988, and 4 remained unplanted awaiting establishment of an adequate cover crop prior to seeding grass. Of the 57 fields seeded in 1988, 35 (61%) had been planted prior to pre-greenup sampling whereas 22 (39%) had not yet been seeded and still retained a sorghum cover crop. The average VOR index among 104 pre-greenup 1988 field samples was 0.39 dm (Table 1). A ranking, based on experience in past studies (Snyder 1984Q, 1988) rated 73.1% of the fields as extremely poor « 0.26 dm), 9.6% as poor (0.26-0.50 dm), 6.7% as fair (0.51-1.0 dm), 4.8% as moderate (1.1-2.0 dm), and 5.8% as good (> 2.0 dm) nesting cover for pheasants during the pre-greenup �226 interval (Table 2). Some fields with high VOR indices had previously been seeded to grass but stands of residual annual forbs provided the primary cover. Most high VOR's were derived from 22 fields that retained a standing sorghum cover crop into which grasses had not yet been seeded (Table 1). Perennial grasses did not yield high VOR's within any field. Most grass stands, where present, were not yet developed and offered minimal cover. Table 2. Ranking of Conservation Reserve fields for nesting pheasants by VOR classification (Kirsch, dm) during pre-greenup and nesting intervals, eastern Colorado, 1988. Sample Pre-greenup Seeded in 1986-87 Seeded in 1988 Prior to sampling After sampling Not yet seeded Totals Nesting Seeded in 1986-87 Seeded in 1988 Not yet seeded Totals < N Fields 0.51-1.0 0.26 0.26-0.50 30 42 32 10 4 5 5 3 2 4 3 76 29 54 1.1-2.0 2.0 Totals 2 3 2 4 3 3 4 43 57 35 22 4 10 7 5 6 8 3 2 1 o 1 > 43 57 4 87 104 4 8 4 3 2 104 The average VOR index for the 104 fields previously sampled declined to 0.20 dm during the nesting interval. Of the fields sampled, 83.7% were rated extremely poor, 7.7% poor, 3.8% fair, 2.9% moderate, and 1.9% good (Table 2). The primary reason for the decline was that grasses were seeded between sample intervals into 22 fields that previously had contained standing cover crops (Table 1). This operation destroyed most standing residual lowering VOR indices. Chlorsulfuron (Glean) and other herbicides were applied to at least 31 fields (the total number treated is unknown) either in 1987 or 1988. Chlorsulfuron was effective in preventing growth of annual forbs on many fields leaving them almost void of green vegetation at time of nest-interval sampling. At least 20 fields, and probably more, had been mowed in 1987. Among the few sampled CR fields that possessed high VOR indices most were within stratum 3 and 4 in southeastern Colorado and contained either tall unmowed weeds or a standing cover crop (Fig. 1). These fields increased the average pre-greenup VOR's to 0.86 dm (stratum 3) and 0.66 dm (stratum 4). In contrast, VOR's in the other 4 strata were below 0.25 dm during pre-greenup. By the nesting interval, VOR's in strata 3 and 4 had declined to 0.28 and 0.07 dm respectively. Fields in northeastern Colorado possessed the highest VOR indices (0.33 dm in stratum 1, Fig. 1). However, comparisons by stratum will not be meaningful until perennial covers have developed. Vegetation canopy cover varied widely among fields and treatments, but averaged 27.4% for the early spring sample and 24.0% for mid to late June. Within the latter sample, 22.1% was attributed to grasses (both annual and 4 �227 WEt. 1 0 I l. A SAN I MAS i d·····_ .. ....~._ :V._, . ~~ ~. .•..__;.l..•..__ ..l~ Fig. 1. Stratification for random sampling conservation eastern Colorado, 1988. _ reserve fields in �228 perennial) whereas 77.9% was attributed to forbs, primarily annuals. Average height of vegetation in the 104 fields in late June was 9.68 em (3.8 in.). Annual precipitation in 1987 was above average at 7 of 9 locations distributed in eastern Colorado (Table 3). In contrast, only 1 of these (Springfield) recorded above average precipitation in 1988. However, precipitation was not markedly deficient during 1988 in eastern Colorado. Only a trace of moisture was received in November-December 1988 (actual data not available) and the data (Table 3) represent the actual amounts received for most locations. Precipitation was deficient in April but above average in May and June at most locations both years. However, June 1988 was extremely warm and dry in northeastern Colorado. July rainfall averaged slightly below normal during both years. Above average rainfall in May during both years was probably the most important factor promoting grass establishment on CR fields. Precipitation was not markedly different than would be expected in eastern Colorado and fair grass establishment was observed in numerous fields. Establishment of perennial grasses is increasingly difficult from north to south in eastern Colorado and above average precipitation in both 1987 and 1988 in Baca County undoubtedly aided grass establishment there. Table 3. Annual eastern Colorado, precipitationa 1987-88. Location average Quality, and Security locations in 1988 16.65 15.40 12.50 15.09 17.41 17.34 12.45 17.61 14.96 21.84 15.10 11.51 19.84 18.22 19.85 15.00 20.06 14.92 16.88 11.21 9.05 13.53 12.21 15.74 10.93 16.22 14.36 15.49 17.37 13.35 ~ aNovember and December amounts are missing amounts were extremely small. Quantity, Bureau 1987 Long-term Springfield Lamar Rocky Ford Limon Burlington Akron Greeley Holyoke Sterling Combined (in.) at 9 U.S. Weather for 1988 but in general of Land Use Types Percentages of major crop types available to wildlife within portions of several counties or groups of counties varied (Table 4). There was an inverse relationship between fallow/mulched stubble and standing small grain stubble with a much greater percentage of fallow/mulch in southern counties and greater proportions of standing stubble in northeastern Colorado (Table 4). Mid-April sampling was not conducted in southern Prowers and Baca counties (extreme southeastern Colorado). Observations indicated most wheat stubble in these areas had been cultivated in fall 1987. Thus, percentages should be similar to those for northern Prowers and Kiowa counties. Observations in previous years indicated that most wheat stubble in southeastern Colorado was �229 tilled in late summer after harvest whereas most was left standing overwinter in northeastern Colorado. The combination of fallow/mulch and standing small grain stubble was the dominant cover type representing 46% of the total cropland sampled. Green wheat comprised 37% of the mid-April cropland and exceeded 30% in all but one sample. Conservation Reserve acreage ranged from o to 23.1% within the sample groups, decreasing from south to north. Millet stubble averaged < 5% of the available cropland whereas the percent of dryland and irrigated rowcrops (corn, sorghum, sunflowers) varied widely among counties (Table 4). Table 4. & Cropland cover types (X) in eastern Colorado, mid-Apri l 1988. County/location fields !!. Fallow/ mulch Wheat stubble Green wheat Conservation Reserve Mi llet stubble Row crop N.Prowers-Kiowa Cheyenne Kit Carson S. Yuma Phillips/Sedgwick Phillips/Logan Logan/lJashington SIILogan/Morgan 143 169 161 99 367 250 312 168 30.8 20.1 39.1 29.3 13.6 12.4 25.0 22.0 4.9 23.1 6.8 12.1 33.0 30.8 36.5 26.2 33.6 40.2 37.3 37.4 40.3 38.4 21.8 44.0 23.1 14.2 2.5 2.0 0 3.2 4.2 3.0 4.2 2.4 0 0 1.9 5.6 8.3 4.2 3.5 0 14.3 19.2 11.2 7.2 4.2 0.6 The VOR indices among 7 standing wheat stubble fields sampled in early spring ranged between 0.28 and 0.51 dm and averaged 0.40 dm. This was similar to the 0.39 dm measured in CR fields. Sampling of standing wheat stubble in southeastern Colorado was inadequate to provide meaningful data but general observations indicated VOR's there would have been lower. The VOR's within 6 mulched stubble fields ranged from 0.05 to 0.30 dm and averaged 0.17 dm, much below standing stubble and extremely marginal for pheasant nesting. Based on past experience (Snyder 1984Q, 1988), standing stubble fields, examined during mid-April, were arbitrarily rated as to their nesting quality for pheasants (Table 5). Most were rated fair except in parts of southeastern Colorado where fields rated poor dominated the sample. Wheat stubble was the only vegetation type, other than CR, providing suitable residual nesting cover in early spring. Standing small grain stubble (primarily winter wheat) was not secure nesting cover and nearly all was lost to tillage during the spring nesting season. Tillage progression varied among localities, but a survey conducted along a previously established Holyoke-Fleming route (Snyder 1988) provided a general index for northeastern Colorado (Fig. 2). Progression of tillage advanced in 1988 at about an average rate when compared to data from previous years (Snyder 1988). Wheat stubble was much more abundant than CR in northeastern Colorado and about equal in nesting quality during spring 1988. However, due to tillage disturbance, it was much less secure. Mulched stubble, while common in southeastern Colorado, was much lower in quality and security. Where tall annual forbs or sorghum existed within CR fields over winter in pheasant range in southeastern Colorado, CR provided valuable winter habitat for pheasants and cottontails. �230 Table 5. Ranking (%) of wheat stubble quality for pheasant nesting in eastern Colorado, spring 1988. County/location N Poor Fair Good N. Prowers-Kiowa Cheyenne Kit Carson S. Yuma Phillips-Sedgwick Phillips-Logan S. Logan-N. Wash. SW Logan-Morgan 7 39 11 12 121 77 114 44 57.1 12.2 45.4 0 10.7 19.5 30.7 9.1 28.6 7l.8 54.6 83.3 71.9 68.8 6l.4 54.5 14.3 15.4 0 16.7 17.4 15.1 65'.9 19.0 ~ 11.7 7.9 36.4 100 90 FALLOW & MULCH 80 I70 I- ~ ... Vl a '- 60 I- ...J "- UJ u:: "- 50 -, -, '- '- 40 " \ \ 30 I- STANDINS STU8BLE \ " \ \ 20 I- " \ \ 10 I- ECOFALLOW 0 19 APR 28 11 26 MAY Fig. 2. Status and conversion of small grain stubble fields to fallow, Holyoke - Fleming route, northeastern Colorado, 1988. Q �231 A 2nd extensive roadside sample of crop type occurrence was conducted on 9 and 10 May 1988 primarily in southeastern Colorado extending into Baca County (Table 6). Spring tillage of stubble fields had been in progress prior to this survey and fallow/mulched fields dominated along with green wheat. By this sample interval, green wheat and other small grains had attained considerable growth and were the dominant nesting cover along with CR fields. Green wheat did not attain significant growth in northeastern Colorado prior to 15 April 1988. Progression of growth had advanced beyond the average for CR fields by late April and was 3.7 dm by late May within northeastern Colorado fields (Fig. 3). Samples in Baca and Prowers Counties on 9-10 May showed wheat growth was more advanced there than in northeastern Colorado and a hypothetical growth curve for that area is illustrated (Fig. 3). Sampling (VOR) of maturing wheat fields in late June showed those in northeastern Colorado averaged> 6 dm compared to 4.7 dm among several fields in southeastern Colorado. Table 6. Cropland cover types County/location (%) !i fields N.Prowers-Kiowa Cheyenne Kit Carson S. Yuma in several eastern Colorado counties, 9-10 May 1988. Fallow/ f1J.Jlch 41.2 28.3 43.3 46.2 119 177 141 78 Wheat stubble Green wheat Conservation Reserve Mi llet stubble 2.5 14.7 1.4 7.7 33.6 42.4 39.7 34.6 19.3 13.0 2.1 0 3.4 1.1 1.4 5.1 Row crop 2.7 0.6 12.1 6.4 4 I 3.5 I I I I 3 I / / 2.5 5 \.J a: GREEN wrEAT - SE •..••. / / / I 2 I ~ I I 1.5 I I GREEN WHEAT - NE I I 1 I CR 0.5 ...........;.:.-.; . ~ o 1S APR 21 28 5 11 26 MAY Fig. 3. Progression of wheat growth based on visual obstruction indices (dm) in southeastern and northeastern Colorado in relation to that of 104 conservation reserve fields, 1988. �232 Canopy cover of wheat fields averaged 58.6% in late June in northeastern Colorado and 44.6% in southeastern Colorado. Canopy cover within sampled CR tracts was 24% in late June. Wheat height was 68 cm in northeastern Colorado and 59 cm in southeastern Colorado in contrast to an average vegetative height of 9.7 cm within CR fields. Green wheat was marginally secure as nesting cover for pheasants in 1988. Hot, dry weather in June advanced maturity and harvest began around 25 June whereas the average harvest initiation date is after 4 July in northeastern Colorado. Pheasant brood counts conducted by cnow personnel during August 1988 revealed extremely poor pheasant reproduction potentially attributed to hot dry weather and the early harvest in combination (Snyder 198412.). More extensive sampling of vegetation and cover types distributed over eastern Colorado was obtained during the mid- to late June 1988 CR nesting cover sample interval. Pasture and rangeland was included within the inventory (4,170 samples). Since millet was still being planted, an accurate assessment of its occurrence could not be made. Thus, percent fallow may be slightly inflated. However, I believe the percentages within county groups are representative (Table 7). Percent rangeland and CR increased in southern areas of the State. Exclusion of rangeland provided more meaningful comparisons of crop types and CR showing the dramatically greater amount of CR in southern parts of eastern Colorado (Table 8). Percentages of cover types bordering the 104 sampled CR fields provided an additional index of proportional occurrence of cover types (Table 9). However, since CR fields are more abundant in marginal farmland, percent occurrence of rangeland is possibly inflated. Table 7. Major vegetation types (%) among grouped county locations in eastern Colorado, June 1988. Vegetation NEa NWb ECc SWd SEe Rangeland Fallow Small grains Conservation Reserve Row crops & millet Alfalfa Other 20.0 32.2 30.0 1.6 13.1 0.1 3.0 17.3 37.9 37.7 2.3 2.7 0 2.1 22.4 38.4 27.0 6.4 4.9 0.3 0.6 53.2 18.7 9.0 9.5 2.3 0.5 6.7 38.2 21.2 13.9 16.3 6.0 4.4 0 aLogan, Phillips, Sedgwick, and Yuma counties. bArapahoe and Weld counties. CCheyenne, Kit Carson, and Washington counties. dElbert, El Paso, Lincoln, and Pueblo counties. eBaca, Bent, Kiowa, and Prowers counties. 31.2 28.2 21.7 9.1 6.2 1.8 1.8 �233 Table 8. D Major crop types (X) within grouped county locations in eastern Colorado, June, 1988. Vegetation NEa NWb ECc SWd SEe Fallow Small grain Conservation Reserve Sorghum/millet Corn Beans/beets Alfalfa Other 40.2 37.5 2.0 0.3 15.7 0.3 0.2 3.8 45.8 45.6 2.8 0.6 1.3 1.3 0 2.6 49.5 34.8 8.3 2.2 3.5 0.6 0.4 0.7 40.0 19.2 20.4 2.7 0 2.3 1.2 14.2 34.5 22.4 26.4 6.9 2.7 0 7.1 0 40.9 31.6 13.2 3.2 5.3 0.6 2.6 2.7 aLogan, Phillips, Sedgwick, and Yuma counties. bArapahoe and Weld counties. CCheyenne, Kit Carson, and Washington counties. dElbert, El Paso, Lincoln, and Pueblo counties. eSaca, Bent, Kiowa, and Prowers counties. In addition to providing secure nesting cover, CR fields should increase edge and interspersion of cover types which many wildlife species require for feeding, brood rearing, etc .. However, 33.4% of the CR field borders were adjacent to rangeland or pasture and another 28.8% were adjacent to other CR fields (Table 9). This grouping, represented 62.2% of the total edge and indicated that diversity was not greatly enhanced. Residual stubble, fallow, and small grains bordered 32.0% of the sampled CR tract edges and sorghum and corn bordered an additional 3.6%. Developed land (building sites, etc.) and other (mostly minor crops) represented 2.2% (Table 9). Based on general observations and assessments, approximately 32% of the 104 fields increased diversity by extending grass into cropland, whereas 38% decreased diversity by converting former cropland within rangeland into grass. About 30% of the CR tracts did not markedly change diversity. Approximately 22% of the sampled CR tracts extended rangeland into cropland, 24% reduced cropland in a range dominated site, 9% eliminated cropland in a range dominated site, and an additional 45% provided no change. Percent rangeland varied widely among strata from 24% within stratum 1 to 57% in stratum 4 (Table 9). Adjacent CR was highest in strata 3 and 5. Highest associations with stubble and fallow were in the northern part of eastern Colorado whereas the highest association with small grains was within the eastcentral (stratum 2) part of the state. Conservation Reserve fields, primarily within the eastern tier of southeastern Colorado counties, were evaluated during a nonrandom inspection on 18-19 July 1988. The 162 CR fields were subjectively rated as containing none, low, medium, or high quality cover for pheasant broods. Cover conditions were rated low within 94 (58.0%) of the fields, whereas 19.8% were rated medium, and 1.9% were rated high (Table 10). One-fifth (20.4%) contained no significant brood cover primarily because they had been treated with herbicides earlier in the year. Most, but not all, were within historic pheasant range. �234 Table 9. Cover type associations (X) with Conservation Reserve based on percent of edge (104 fields sampled within 6 strata), eastern Colorado, 1988. Cover type Rangeland Cons. Reserve Stubble/fallow Small grains Corn Sorghum Othera Developedb 23.8 15.0 30.2 16.1 7.8 1.9 4.3 0.9 2 3 Stratum 4 5 6 ~ 25.2 16.2 21.6 28.0 7.8 0.5 0.2 0.5 22.2 39.0 20.5 13.9 0.6 2.5 0.8 0.5 57.3 21.2 14.0 6.3 0 0.9 0 0.3 37.9 38.7 9.8 7.6 0.7 2.0 2.4 1.0 31.0 16.2 32.2 19.6 0 0 0.2 0.9 33.4 28.8 19.0 13.0 2.0 1.6 1.5 0.7 aprimarily alfalfa, trees, and other minor crops. bprimarily farmyards. Table 10. Rating of Conservation Reserve southeastern Colorado, 18-19 July 1988. High Value Medium Kiowa Eastern Western 0.0 0.0 7.1 16.7 Prowers 0.0 Baca Two Buttes Other fields for pheasant (% of total) Low brood habitat, N None fields 78.6 58.3 14.3 25.0 14 24 26.7 43.3 30.0 30 0.0 5.3 17.9 21.1 67.8 55.3 14.3 18.4 28 38 Bent 0.0 0.0 66.7 33.3 3 Cheyenne 4.0 24.0 56.0 16.0 25 1.9(3)a 19.8(32)a 58.0(94)a 20.4(33)a 162 County All fields "Tocal, fields. Conservation Reserve Association with Wildlife Distribution Conservation Reserve acreage distribution within eastern Colorado, based on data through the 1st 4 CR signups, was compared with the distribution and general density classifications for ring-necked pheasants, northern bobwhite, and scaled quail. Data were contrasted to pheasant range based on 1982-83 distribution-density estimates (Snyder 1985). Analysis indicated 43.8% of the �235 CR in eastern Colorado was outside pheasant range, whereas 32.6% occurred within range classed as remnant, or it contained only scattered occurrences of the species. About 16.9% of the CR was within range classed as low, 2.5% was within fair range, and 4.2% was within moderate to good pheasant range. Since remnant range in eastern Colorado provides almost no hunting opportunity, only 23.6% of the CR acreage occurred within huntable range. Most pheasant range rated moderate to high in 1982-83 has deteriorated to low to fair status in recent years. A higher proportion of land recently committed to CR within signups 5 through 7 is within pheasant range classed as low to good. However, the amount of land committed to the CRP during recent signups was much less than that during the 1st 4 signups. These data illustrate that much of eastern Colorado is extremely marginal for both farming and pheasants limiting the value of CR for the species. Since pheasants are primarily found within the better farmlands and CR was more extensively committed to marginal farmland, an inverse relationship between pheasant abundance and CR was noted. One exception was in portions of Baca County where extensive conversion to CR occurred within what is presently Colorado's best pheasant range. Comparison of CR distribution with that of northern bobwhite was based on bobwhite distribution presented in Snyder (1984~). Only 6% of the CR acreage was within northern bobwhite range, primarily remnant or marginal habitats. The best bobwhite range is along riparian streamside habitats in Colorado and, therefore, is separated from CR, especially along the South Platte River. Scaled quail distribution and density were based on Hoffman (1965) and modified to exclude range he classed as very poor, since most of this was farmland or shrub less shortgrass unsuited for quail habitation. Hoffman rated a large proportion of the scaled quail range as good based on habitat characteristics rather than population densities existing there. As a consequence, 2.2% of CR acreage was within poor scaled quail range, 7.4% was within fair range, 5.5% was within good range, and 1.0% was within range classed as very good or excellent. Thus, 16.2% of the CR was within or closely associated with scaled quail range. This analysis used townships as the basic unit of measure and scaled quail range was often isolated along rangeland-dominated drainages distributed as fingers of habitat within sections and townships. If a more refined analysis was conducted, the percentage of CR acreage within actual scaled quail range would be much smaller. If southeastern Colorado scaled quail range was reclassified based on long-term densities, a vast majority would be rated poor or low rather than good range. Quotas (-25% of total cropland) of CR were attained in most southeastern Colorado counties during the 1st 4 signups. Thus, no significant change in the CR-scaled quail range relationship is anticipated. Percentages of the 104 sampled CR fields within the ranges of 7 eastern Colorado game species were recorded during June 1988 (Table 11). Only 4 fields (3.8%) were within the range of prairie grouse (Tetraoninae), whereas 40.4% were within the range of Pronghorn (Antilocapra americana) and all fields were within the range of mourning doves (Zenaida macroura) and nongame passerines. When the CR field was within the range of a wildlife species, the overall habitat quality of the tract was rated (poor, fair, good, excellent) for that species (Table 11). For most game species, the habitat quality rating was in the poor classification although the fair classification dominated for pronghorn. Only a few fields were rated good and none were rated excellent. �236 Table 11. Occurrence of selected wildlife in association with 104 Conservation Reserve fields and the quality rating of these fields to wildlife, eastern Colorado, June 1988. Occurrence Species/group x Poor Ring-necked pheasant Northern bobwhite Scaled quail Prairie grouse Cottontail Pronghorn Mourning dove Nongame passerines 63.5 11.5 19.2 3.8 66.3 40.4 100.0 100.0 72.7 83.0 85.0 75.0 81.2 35.7 67.0 76.9 Rating of habitat gual it~ Fair Good 22.7 17.0 15.0 25.0 18.8 61.9 31.1 21.2 Excellent 4.5 0 0 0 0 2.4 1.9 1.9 0 0 0 0 0 0 0 0 An attempt was also made to rank the 104 fields as to their relative importance to selected wildlife species (or species groups). Most fields were rated of highest value to nongame passerines with some fields having value for pronghorn and mourning doves (Table 12). These ratings were subjective and based on June 1988 conditions within and surrounding the fields. These ratings would undoubtedly change during subsequent inspections. Table 12. Relative importance of 104 Conservation Reserve fields to selected wildlife, eastern Colorado, June 1988. lJildlife R-n pheasant Northern bobwhite Scaled quail Prairie grouse Cottontail Pronghorn Mourning dove Nongame passerines 2 2 0 1 0 0 17 4 80 17 0 0 0 6 11 41 22 Relative imQQrtance among s~ciesa 4 5 3 24 1 0 0 17 7 49 2 10 2 8 3 38 3 a a 11 5 10 1 7 3 2 a 6 7 8 2 3 1 0 1 0 0 0 0 1 0 0 0 0 0 0 0 0 0 a a 0 0 , aHighest value = 1, lowest value = 8 Progress on Other Phases of Study Preliminary work was begun on plotting CR (and other cover types) along pheasant crowing, mourning dove, and breeding bird survey census routes in eastern Colorado. However, extensive work continues and population changes of selected wildlife species in relation to CR distribution may not be detectable (if detectable) for several years. The Colorado Division of Wildlife cost shared to promote planting of perennial vegetation considered of greater value to wildlife on numerous fields, primarily within the eastern tier of counties. A random sample of these tracts was obtained with the assistance of Ernest Kaska and Tim Davis in late fall 1988. Sampling is to be initiation in 1989. However, additional time �237 may be needed to discern vegetation quality differences, development of perennial grasses is slow. if they exist, since Not all the variables sampled within the NEC data forms have been summarized and therefore are not included here. A summary of these data may be forthcoming from NEC personnel. LITERATURE CITED Hoffman, D. M. 1965. The scaled quail in Colorado - Range, population, status, and harvest. Colorado Dep. Game, Fish and Parks Tech. Bull. 18. 47pp. Robel, R. J., J. N. Briggs, A. D. Dayton, and L. C. Hulbert. 1970. Relationship between visual obstruction measurements and weight of grassland vegetation. J. Range Manage. 23:295-297. Snyder, W. D. Colorado. 1984~. Management procedures for northern bobwhites Colorado Div. Wildl. Spec. Rep. 56. 22pp. 1984g. High Plains. in eastern Ring-necked pheasant nesting ecology and wheat farming on the J. Wildl. Manage. 48:878-888. 1985. Management procedures for ring-necked Colorado Div. Wildl. Spec. Rep. 59. 53pp. pheasants in Colorado. 1987. Evaluation of sandsage-bluestem pra~r~e. Job Progress Rep., Colorado Div. Wildl., Wildl. Res. Rep., Fed. Aid Proj. W-152-R Apr. :331356. 1988. Evaluation of no-till wheat farming. Job Final Rep., Colorado Div. Wildl., Wildl. Res. Rep., Fed. Aid Proj. W-152-R. Apr. In press. Prepared by '%d~¥ Warren D. Snyder Wildlife Researcher C ��239 APPENDIX A PRE-GREENUP PERIOD DATA FORM Southern Great Plains Region l. Your Name: 2. County: 4. Contract 3. #: 5. State: Year to start: 6. CP#: ASCS Office 7. Owner/operator: 8. Field ID: 11. Contract Telephone 9. Field 2 CP# Acres 12a. Contracted planting forestry; wildlife, Description 13. Possible 10. Field acres: field descriptions: Field 1 CP# Acres 12b. Contracted Signup #: No.: (e.g., cool season/legume; or other): if "other," temporary pheasant Field 5 CP# Acres warm season grasses; or seed mix if CPl or CP2: cover crop or residue winter Description/location: Field 4 CP# Acres Field 3 CP# Acres (if any): cover on surroundings: Distance: �240 15. Directions for getting to the field: 16. Attach xerox copies of the contract, crop history form, and the ASCS photo of the field and surroundings. If the photo is not available, draw a sketch on the back of this form. Field sampling 21. Date of field work: 22. Location 23. Description (Starting of access point: of landmarks used. Show location on photo (14). landmark): (Target landmark): 24. Location of point 1 (# of steps): Starting point or notes: 25. Direction notes 26. Location Daubenmire Herb. canopy cover Robel reading Plot #1 Shrub canopy cover Daubenmire Herb. canopy cover Plot #2 Shrub canopy cover dm ---% --- % --- % % dm --- % --- % --- % --- % dm --- % --- % --- % --- dm ---% --- % --- % --- % of point 2 (# of steps): Notes: % �241 PRE - GREENUP DATA FORM Contract # 25. Direction notes Robel reading Whole field characteristics 31. Daubenrnire Plot #1 Herb. Shrub canopy canopy cover cover dm --- % dm --- % dm --- % dm --- % Daubenrnire Plot #2 Herb. Shrub canopy canopy cover cover ---% ---% ---% ---% --- % --- % --- % --- % --- % --- % ---% --- for this CP % of CP area undisturbed: amount (%) : mowed: amount (%) : grazed: amount (%) : burned: amount (%) : other: amount (%) : 100 TOTAL 32. % When was the permanent Is the permanent contracted contracted What are the dominant cover planted? cover present (year) on this CP? (Y/N): species?: 33. What is the quality class of winter cover for pheasant on the CP? (Poor: 0/10; Fair: 1-3/10; Good: 4-6/10; Excellent: 7-10/10): 34. Is there winter food for pheasants? a. On this CP? (Y/N): b. Close to this CP? (Y/N): « 1/4 mile-CO, cover for pheasants KS, NE, WY; < 1/2 mile-OK. 35. Is there winter within 2 miles of this CP? 37. Which of the following is the single most limiting life requisite pheasants within 2 miles, but not including the field? (N: nesting cover; WF: winter food; WC: winter cover): NM, TX) (Y/N) for �242 APPENDIX B NESTING PERIOD DATA FORM Southern Great Plains 1. Your Name: 2. County: 4. Contract 8. Field Field 3. #: 6. CP#: ID: Sampling 21. Date of field work: 22. Location 23. Description of access (Starting 24. State: point: of landmarks used. Show location on photo (14). landmark): (Target landmark): Location of point 1 (# of steps): Starting point or notes: 25a. Direction notes Robel reading 25b. Distance to meadowlark 26. of point Location Daubenmire Herb. canopy cover Plot #1 Shrub canopy cover Daubenmire Herb. canopy cover Plot #2 Shrub canopy cover dm --- % --- % --- % % dm --- % --- % --- % % dm --- % --- % --- % % dm --- % ---% --- % % perch site: 2 (# of steps): left side of line: right side of line: Notes: M M �243 27a. Robel reading Direction notes 27b. Distance Daubenmire Herb. canopy cover Daubenmire Herb. canopy cover -_% --- % ---% --- % dm ---% --- % -_% --- % dm --- % --- % --- % --- % dm -_% --- % --- % % perch site: Whole field characteristics for this CP left side of line: right side of line: M M % of CP area undisturbed: "amount mowed: amount (%): grazed: amount (%) : burned: amount (%) : other: amount (%) : (%) : TOTAL 32. Plot #2 Shrub canopy cover dm to meadowlark 31. Plot #1 Shrub canopy cover When was the permanent Is the permanent contracted contracted cover planted? 100 (year) cover present on this CP? (Y/N): What are the dominant species?: 36. What winter cover quality class will the present cover in this CP likely supply for pheasants next winter? (Poor: 0/10; Fair: 1-3/10; Good: 4-6/ 10; Excellent: 7-10/10) 37. Which of the following is the single most limiting life requisite pheasants within 2 miles, but not including the field? (N: nesting cover; WF: winter food; WC: winter cover): 38. Is there brood cover for pheasants? a. On this CP? b. Within (Y/N): 1/2 mile of this CP? (Y/N): for �244 NESTING DATA FORM 39. Contract # Number of cover type patches adjacent to the field: Ungrazed grassland Corn Range, native grass, shrubs Cotton Range, introduced grass Small grains Hayland, timber, orchards Sorghum Woodland, timber, orchards Wetland Fallow cropland, stubble, bare ground Open water Alfalfa and clover Developed land Beans �Colorado Division of Wildlife Wildlife Research Report April 1989 245 JOB PROGRESS State of: Colorado Project: W-152-R Upland Bird Research Work Plan: 22 Job Title: Avian Research Period Covered: Author: REPORT Job 01 January 1__ Publications through 31 December 1988 Clait E. Braun Personnel: C. E. Braun, K. M. Giesen, R. W. Hoffman, Snyder, Colorado Division of Wildlife T. E. Remington, W. D. ABSTRACT Progress was made on preparing and submitting manuscripts journals. Two manuscripts were published. Giesen, K. M. 1988. Status of lesser prairie-chickens Field-Ornith. 22:57-58. to technical in Colorado. Remington, T. E., and C. E. Braun. 1988. Carcass composition reserves of sage grouse during winter. Condor 90:15-19. Colo. and energy � https://cpw.cvlcollections.org/files/original/3d0701ffef240ca8cb52c149f7c4d644.pdf ae355eac0a2912a10e0c3d568477ef78 PDF Text Text 1 Colorado Division of Wildlife Wildlife Research Report July 1989 JOB PROGRESS REPORT State of ~C~o~l~o~r~a~d~o~ Project No. W-153-R-3 Work Plan No. Job. No. Personnel: ~l~ 7 Period Covered: Authors: _ Mammals Research _ Multispecies Investigations Terrestrial Research Publication, Editing and Library Service July 1, 1988 - June 30, 1989 M. W. Hershcopf, L. H. Carpenter R. B. Gill, N. McEwen, L. Lovett, K. Chociej, all Mammals Researchers S. Eby, and ABSTRACT During the 1988-89 Segment, 16 books were purchased for permanent reference by DOW personnel. Fifty additional publications were located, ordered, and obtained free of charge for use. Fourteen of these were purchased, obtained on Interlibrary Loan, or given to the library. An additional 1,265 individual references requested by Mammals Researchers were located by library staff and made available for use. About 30 of these requests were not available locally and were obtained through interlibrary loans. Twelve manuscripts were published in various journals. ��3 MAMMALS PUBLICATION, EDITING AND LIBRARY SERVICES Marian W. Hershcopf and Len H. Carpenter P. N. OBJECTIVES To provide a centralized support program for manuscript editing services to facilitate publishing results of research conducted 01-03-047 - 11700 and 01-03-048 - 11700 and 16700. and library in projects SEGMENT OBJECTIVES 1. To provide coordinated and efficient editing and library services publish findings for all Colorado Mammals Research programs. and 2. To provide for the centralized support program for Mammals Research editing, library, and publishing services so that Mammals Research scientists can be most efficient in publishing results of their research. SUMMARY OF SERVICES Publications purchased with Mammals Research funds and placed in the Research Center Library Allred, M. 1986. Beaver behavior: architect of fame and bane. Naturegraph Publishers, Inc. Happy Camp, CA. 110pp. Brown, L., and J. F. Downhower. 1988. Analyses in behavioral ecology: a manual for lab and field. Sinauer Assoc., Inc., Sunderland, MA. 194pp. Eastern alack Bear Workshop on Research and Management. Proceedings 5th, 1980, Wrightsville Beach, NC, 292pp.; Proceedings 7th, 1984, Homosassa, FL, 85pp.; Proceedings 8th, 1986, Williamsburg, VA, 249pp. Hobbs, H. H., III, and J. P. Jass. The crayfishes and shrimp of Wisconsin. Milwaukee Public Museum, Milwaukee, WI. l77pp. Hunt, C. E. 1988. Down by the river: projects and policies on biological DC. 260pp. the impact of federal water diversity. Island Press, Washington, Huntley, A. C., et al., eds. 1987. Approaches energetics. Soc. Marine Mammal., Lawrence, Kurten, B., and E. Anderson. 1980. Pleistocene Columbia Univ. Press, New York, NY. 442pp. to marine mammal KS, Spec. Publ. No.1. mammals 253pp. of North America. �4 Novak, M., et a1., eds. 1987. Wild furbearer management and conservation in North America. Ontario Trappers Assoc. and Ontario Ministry of Natural Resources, Toronto, Ontario, Can. 1150pp. Putman, R. 1988. The natural history Press, Ithaca, NY. 191pp. of deer. Cornell University Renecker, L. A., ed. 1987. Focus on a new industry. Proceedings of the Alberta Game Growers' Association Conference, Red Deer, Alberta, Alberta Game Growers Assoc., Edmonton, Alberta, Can. ll0pp. Rolston, H., III. natural world. 1988. Environmental ethics: duties to and values Temple Univ. Press, Philadelphia, PA. 391pp. in the Soule' ,M. E., ed. 1987. Viable populations for conservation. Cambridge Univ. Press, New York, NY. 189pp. Starfield, A. M., and A. L. B1eloch. 1986. Building models for conservation and wildlife management. Macmillan Publishing Co., New York, NY. 253pp. Zeveloff, S. I. 1988. Mammals of the intermountain Utah Press, Salt Lake City, UT. 365pp. Publications obtained west. Univ. of free or at low cost In addition to books purchased with Federal Aid Funds, about 50 free reports and short publications from state or federal agencies or from private sources were located, ordered, and obtained for use by Mammals Research personnel. Theses purchased. obtained on Interlibrary Loan or as gifts for use by Researchers Methods and applications in estimating mortality Barlow, J. P. 1982(?). Ph.D. Diss., Univ. of California, San Diego, CA. other vital rates. 177pp. Collins, T. C. 1976. Population characteristics and habitat relationships of beavers, Castor canadensis, in northwest Wyoming. Diss., Univ. of Wyoming, Laramie, WY. 172pp. and Ph.D. Greer, S. Q. 1987. Home range, habitat use, and food habits of river otters in southwestern Montana. M.S. Thesis, Montana State Univ., Bozeman, MT. 9lpp. Griess, J. M. 1987. River otter reintroduction in Great Smoky Mountains National Park. M.S. Thesis, Univ. of Tennessee, Knoxville, 148pp. Harmoning, A. K. 1976. White-tailed deer dispersion and habitat utilization in central North Dakota. M.S. Thesis, North Dakota State Univ., Fargo, ND. 56pp. TN. �5 Hardesty, J. Y. (n.d.) Riparian vegetation at three sites along the Gila River in southwest New Mexico. M.S. Thesis, New Mexico State Univ., Las Cruces, NM. 101pp. Kafcas, E. N. 1987. Census and exploitation of a discrete beaver population in Michigan. M.S. Thesis, Central Michigan Univ., Mount Pleasant, MI. 88pp. Liewer, J. A. 1988. Pronghorn grazing impacts on winter wheat. M.S. Thesis, Colorado State Univ., Fort Collins, CO. 32pp. Marcum, C. L. 1975. Summer-fall habitat selection and use by a western Montana elk herd. Ph.D. Diss., Univ. of Montana, Missoula, MT. l88pp. Olson, W. W. 1975. Effects of controlled burning on grassland within the Tewaukon National Wildlife Refuge. M.S. Thesis, North Dakota State Univ., Fargo, ND. 137pp. Polechla, P. J., Jr. 1987. Status of the river otter (Lutra canadensis) population with special reference to reproductive biology. Ph.D. Diss., Univ. of Arkansas, Fayetteville, AR. 383pp. Robb, L. 1987. Gastropod intermediate hosts of lungworms (Nematoda: Protostrongy lidae) on a bighorn sheep winter range: aspects of transmission. M.S. Thesis, Univ. of Alberta, Edmonton, Alberta, Can. lllpp. Shelton, L. R., III. 1978. Values of big game to western slope ranchers of Colorado. Ph.D. Diss., Colorado State Univ., Fort Collins, CO. 17lpp. Silflow, R. M. 1989. Comparison of pulmonary host defense mechanisms in Rocky Mountain bighorn sheep (Ovis canadensis canadensis) and domestic sheep. M.S. Thesis, Washington State Univ., Pullman, WA. l08pp. Reference document location and delivery The Research Center Library staff also located and delivered about 1,265 individual articles on request for Mammals Researchers during this segment; about 30 were not available locally and were obtained through Interlibrary Loan procedures. Manuscripts published Job Progress Reports; Federal Aid. All studies. Anderson, A. E., and R. J. Tully. 1988. Status report, mountain lion, Colorado. Proceedings Third National Mountain Lion Workshop. Prescott, AZ. (in press). , D. C. Bowden, and D. M. Kattner. 1988. Dynamics of home range size of unhunted mountain lion (Felis concolor hippolestes) in southwestern Colorado. (Abstract). Proceedings Third National Mountain Lion Workshop, Prescott, AZ. (in press). �6 Anderson, A. E., D. C. Bowden, and D. M. Kattner. 1988. Survival of mountain lion (Felis concolor hippolestes) in an unhunted population in southwestern Colorado. (Abstract). Proceedings Third National Mountain Lion Workshop, Prescott, AZ. (in press). Bear, G. D. 1989. Seasonal distribution and population characteristics of elk in Estes Valley, Colorado. Colorado Div. of Wildlife, Spec. Rep. No. 65. l4pp. Dailey, T. V., and N. T. Hobbs. 1989. Travel in alpine terrain: energetics of locomotion by mountain goats and bighorn sheep. Zool. (in press). Can. J. Hobbs, N. T. 1988. Obligations and expectations of your peers: manuscript review at the Journal of Range Management. J. Range Manage. 41:368-369. 1988. Estimating habitat carrying capacity: an approach for planning reclamation and mitigation for wild ungulates. Pages 3-7 in Issues and technology in the management of impacted wildlife: proceedings of a national symposium, Colorado S~rings, CO, 1987. Thorne Ecological Inst., Boulder, CO. 1989. Linking energy balance to survival in mule deer: development and test of a simulation model. Wildlife Monograph 37pp. 101. Kufeld, R., D. C. Bowden, and D. Schrupp. 1988. Habitat selection and activity patterns of female mule deer in the Front Range, Colorado. Range Manage. 41:515-522. Miller, M. W., M. A. Wild, B. J. Baker, and A. T. Tu. 1989. Snakebite captive Rocky Mountain elk. J. Wildl. Diseases 25:392-396. Torbit, S. C., L. H. Carpenter, R. M. Bartmann, A. W. Alldredge, and G. C. White. 1988. Calibration of carcass fat indices in wintering deer. J. Wildl. Manage. 52:582-588. J. in mule White, G. C., R. M. Bartmann, L. H. Carpenter, and R. A. Garrott. 1989. Evaluation of aerial line transects for estimating mule deer densities. J. Wild1. Manage. 53:625-635. Prepared by Len H. Carpenter Wildlife Research Leader �7 Colorado Division Wildlife Research July 1989 of Wildlife Report JOB PROGRESS REPORT State of Colorado Project No. Work Plan No. Job. No. ~l 9 Period Covered: Author: Mammals Research W-1S3-R-3 _ Multispecies Mammals Investigations 1 Research Administration July 1, 1988 - June 30, 1989 Len H. Carpenter ABSTRACT Planning and evaluation responsibilities for Colorado's deer and elk programs consumed a large portion of the Wildlife Research Leader's time during FY 88-89. An analysis and summary of the deer and elk seasons for the 3 years 1986-88 was prepared. ��9 MAMMALS I RESEARCH ADMINISTRATION Len H. Carpenter P. N. OBJECTIVE To supervise Project. and administer research on deer, elk, and moose in the Mammals METHODS The position of Research Leader is established to supervise and administer all research conducted for the wildlife species of concern and also to plan and evaluate all statewide deer, elk, and moose programs. SEGMENT OBJECTIVE To supervise Project. and administer research on deer, elk, and moose in the Mammals RESULTS AND DISCUSSION Research administration and supervision for 5 full-time researchers was provided until June 1, 1988. At that time, the Research Leader was temporarily assigned to the position of Acting State Wildlife Manager for Terrestrial Resources in Denver for 6 months. A significant amount of time was devoted to analyzing and summarizing results of the past 3 years of big game hunting seasons in order to make recommendations for the next 3-year season structure. This summary was presented to the Wildlife Commission in May, 1988, and was the basis for Division recommendations for years 1988-90. Prepared by Len H. Carpenter Wildlife Research Leader �Colorado Division of Wildlife Wildlife Research Report July 1989 JOB FINAL REPORT State of Colorado Project No. W-lS3-R-3 Work Plan No. Job. No. _ 5 Period Covered: Author: ~3 Mammals Research Multispecies Investigations Impact of Elk Winter Grazing on Livestock Production July 1, 1988 - June 30, 1989 D. L. Baker, N. T. Hobbs Personnel: G. Bear, M. Miller, B. Gill, L. Carpenter, B. Petch, C. Woodward, J. Ritchie, C. Mehaffy, M. Stevens, B. Seely, H. Seely, L. Lovett ABSTRACT All research results for this segment are summarized in the Mammals 2, WP9A,Jl Report. �11 Colorado Division Wildlife Research July 1989 of Wildlife Report JOB PROGRESS REPORT State of Colorado Project No. _W~-=15~3~-~R~-~3~ _ Mammals Work Plan No. _ Deer Investigations Job No. 7 Period Covered: Author: ~2~ Research Development of Census Methods for deer in Plains Riverbottom Habitats July 1, 1988 - June 30, 1998 R.C. Kufeld Personnel: D. Bowden and numerous Region Employees. Colorado Division of Wildlife Northeast ABSTRACT Seven deer (6 mule deer and 1 whitetail) were radio-collared and eartagged in the South Platte Riverbottom near Sedgwick and near Sterling, Colorado on January 25, and February 17, 1989. Deer radio-collared in the South Platte Riverbottom between Platteville, Colorado, and the Nebraska State Line during January and February, 1987,88, and 89, were located at approximately 2-week intervals throughout the year to determine movements and home range size. Deer in that stretch of the South Platte Riverbottom were counted and classified by sex and age 3 times by 2 observers in a helicopter between January 29, and February 13, 1989. Counts 2 and 3, however, were limited to certain river segments where radio-collared deer occurred. Observers also recorded the number of radio-collared deer they saw during each count. Radiocollared'deer were located by fixed wing aircraft while counts were in progress, so the number of radio-collared deer in the census area was known. The mean portion of radi.o-collared deer seen by observers on 3 counts was 93% for mule deer and 69% for whitetails. Data from count 1 suggest an estimated population, along the South Platte river between Platteville, Colorado, and the Nebraska State Line, of 1890 white-tailed deer and 529 mule deer. The population mean based on. 3 counts was estimated with 95% confidence within 28% for whitetails, and within 29% for mule deer. On count 1 buck/doe ratios were 24/100 and 26/100, and fawn/doe ratios were 75/100 and 93/100, respectively, for white-tailed and mule deer. Based on 3 counts the mean buck/doe ratio was estimated within 56% and 45%, respectively, for whitetails and mule deer, and the mean fawn/doe ratio was estimated, respectively, for the 2 species within 8% and 17%. Misc1assification of radio-collared bucks as does because of lost antlers suggest, however, that these sex and age ratios may be inaccurate. A December time frame for determination of sex and age ratios is recommended. ��13 DEVELOPMENT OF CENSUS METHODS FOR DEER IN PLAINS RIVERBOTTOM HABITATS Roland C. Kufeld P. N. OBJECTIVES 1. To determine seasonal movements mule deer in plains riverbottom and home range size of white-tailed habitats. 2. To develop and test methods for estimating size of deer populations in plains riverbottom habitats. and SEGMENT OBJECTIVES Same as P. N. objectives. ACKNOWLEDGMENTS Numerous personnel from Colorado Division of Wildlife Northeast Region were instrumental in capturing deer and in making aerial deer counts. STUDY AREA The study area is the South Platte River basin extending from Platteville, Colorado, to Nebraska. Riparian vegetation in the floodplain is dominated primarily by cottonwood (Populus sargentii), and willow (Salix §QQ.). The riverbottom is bordered along much of its length by agricultural lands, mainly cornfields. Other stretches are bordered by rangelands dominated by mixed prairie or sand sagebrush (Artemisia filifolia) (Costello 1954). Land status includes both private lands and state-owned wildlife areas. Results from the South Platte should apply to most other plains riverbottom habitats such as those found along the Arkansas River. METHODS AND MATERIALS White-tailed and mule deer were trapped and marked in the South Platte riverbottom near Sedgwick, Colorado, and at Dune Ridge State Wildlife Area near Sterling on January 25, and February 17, 1989. Each deer received a radio-collar and 2 orange, numbered eartags. Deer were captured in Clover traps (Clover 1956) and a drop net. Alfalfa hay and eared corn were used as bait. White-tailed and mule deer, radio-collared in the riverbottom during January and February, 1987 and 1988 (Kufeld 1987, 1988), were located at approximately 2-week intervals throughout the fiscal year. Those deer instrumented in January and February, 1989, were located at about 2-week intervals for the remainder of the fiscal year. �14 Deer were counted 3 times between January 29, and February 13, 1989, in the South Platte Riverbottom between Platteville, Colorado, and the Nebraska State Line, a distance of approximately 310 KID. Counts were made by 2 experienced observers (in addition to the pilot) in a Bell Jet Ranger Helicopter. As the helicopter moved along the river it flew at treetop altitude and followed a zig-zag course from the outer edge of the trees on one side of the river to the outer edge of the trees on the other side. The census area averaged approximately 0.8 KID wide and was partitioned into 31 river segments with a mean length of 10.0 Km (range 3.5 to 21.2 Km). Segment boundaries were bridges that crossed the river. A total count of deer was made in each segment. Each deer was classified by species, sex, and age, and a notation made if a deer was wearing a radio-collar. Every day helicopter counts were conducted a fixed-wing airplane flight was also made to locate deer that had been radio-collared during January and February, 1987 and 1988, and January, 1989, in the South Platte Riverbottom between Platteville, Colorado and the Nebraska State Line. Thus, the number of radio-collared deer available to be seen in the census area while counts were in progress was known, and the proportion of radio-collared deer of each species seen by observers in the helicopter provided an estimate of the accuracy of their count. During flight 1 deer were counted and classified in all river segments to provide continuity with data from counts conducted in prior years. During flights 2 and 3 deer were counted and classified in only those river segments where radio-collared deer occurred in order to provide replicate counts for testing of deer observability. Those were: Highway 34 Bridge in Greeley to Kersey Bridge, Kuner Bridge to Orchard Bridge, Highway 6 Bridge near Merino to Sterling Bridge, Iliff Bridge to Proctor Bridge, and Crook Bridge to Sedgwick Bridge. Percentages of radio-collared white-tailed and mule deer observed during the 3 counts were compared using paired-t tests. Unpaired-t tests were used to analyze sex and age ratio differences between deer species and between counts within species. Differences were considered significant if P ~ 0.05. RESULTS Seven deer were captured and radio-collared in 1989 (Table 1). A computer system for analyzing data from tracking deer that were radio-collared during the course of this study is being prepared. This analysis will be presented in a future report. Deer Counts When helicopter counts of deer along the South Platte River were conducted 32 of the radio-collared deer (18 mule deer and 14 whitetails) were alive. Most radio-collared deer of both species were in the riverbottom census area during all 3 counts (Fig. 1). The remaining radio-collared deer were on the plains away from the river. Presence of snowcover was considered a prerequisite to conducting aerial counts based on previous years of experience in counting deer along the South Platte River. Conditions for counting deer during the 3 helicopter flights were judged mostly very good in that there was 100% snow cover during flights �15 1 and 2. During flight 3 snow cover was 100% from Highway 34 Bridge in Greeley to Masters bridge, 80% from Masters Bridge to Orchard Bridge, 25% from Highway 6 Bridge near Merino to Atwood Bridge, and 100% from Atwood Bridge to Sedgwick Bridge. The percent of radio-collar~d deer seen during the 3 helicopter counts was consistently higher and less variable for mule deer than for whitetails (Table 2). Despite the relatively large difference between deer species in mean percent of radio-collared animals seen, the difference was not significant. Plans called for 5 counts, however, adequate snowcover existed only long enough to complete 3 counts. This lack of significance may have been a result of low sample size. The number of deer of each species recorded during a count was adjusted upward, based on the percent of radio-collared deer seen during that count, to provide an estimate of the total deer population in the area censused. Relative precision of population estimates is based on adjusted total deer seen during the 3 counts in river segments that had radio-collared deer (Table 2). The population mean was estimated within 28% with 95% confidence for whitetails, and within 29% of the mean with 95% confidence for mule deer. On flight 1, which provided a count of deer in all river segments, 529 mule deer were recorded. This represents an estimate of the total mule deer population since 100% of the radio-collared mule deer were seen. On flight 1, 1304 white-tailed deer and 69% of the radio-collared whitetails were observed. The estimated adjusted total number of whitetails is 1890 deer. Population estimates based on flight 1 suggest a composition of 78% white-tailed deer and 29% mule deer along the South Platte river between Platteville, Colorado, and the Nebraska State Line. This compares with mean population estimates (Table 2), that suggest a composition of 71% white-tailed deer and 29% mule deer in those river segments which had deer with radio-collars. Deer Sex and Age Classification Mule deer and whitetail buck/doe ratios in all river segments (flight 1) were 26/100 and 24/100, respectively. No significant differences in buck/doe ratios occurred between counts for either deer species or between mule deer and whitetails on individual counts conducted in river segments which had radio-collared deer (Table 2). Fawn/doe ratios were higher for mule deer (93/100) than for whitetails (75/100) in all river segments (flight 1), as well as in river segments which had deer with radio-collars (Table 2). Fawn/doe ratio differences between mule and white-tailed deer during the 3 counts were not significant, however, nor were differences between counts in fawn/doe ratios for either deer species (Table 2). In the case of both deer species estimating precision for fawn\doe ratios was much better than for buck\doe ratios. Based on 3 counts the mean fawn\doe ratio was estimated within 8% and 17% of the mean, respectively, for whitetails and mule deer with 95% confidence. Conversely, for whitetails and mule deer, respectively, the mean buck/doe ratio was estimated within 56% and 45% of the mean. �16 DISCUSSION The South Platte riverbottom in eastern Colorado presents a rather unique situation for aerial censusing of deer in that the area is very long, narrow and winding. Two species of deer occur, and distribution of both species throughout the area is non-uniform ..The area, because of its shape and deer distribution patterns, is not suited for implementation of a randomized helicopter quadrat census system (Kufeld et al. 1980), or an aerial line transect system (Burnham et al. 1980). Thus, it was deemed more efficient to census the entire area. DeYoung (1985), Beasom et al. (1986), and Bartmann et al. (1986) have shown that during aerial counts of white-tailed and mule deer somewhat fewer than 100% of the deer present are actually seen and recorded. Beasom et a1. (1986) reported accuracy of aerial counts for whitetails was unaffected by sample intensity, but precision increased with percent of the area sampled. Since we sampled the entire area of interest findings of Beasom et al. (1986) suggest the only way to increase our precision is with replicate counts. Replicate counts are subject to economic constraints, and limited by the relatively short period that adequate snow cover normally occurs on the eastern Colorado plains. Sightabi1ity of animals during aerial counts is influenced by factors such as speed, height above ground, transect width, observers, group size and vegetation cover (Caughley et al. 1976, Barnes et al. 1986, Samuel et al. 1987). Aerial deer counts in plains riverbottom habitats should be designed and conducted with these factors in mind. For example, the helicopter should fly at a speed of from 56-72 Kmjhr, and altitude of about 30 m. Zig-zag paths taken by the helicopter across the census area should be sufficiently narrow that there is good visibility of the area between subsequent passes. Observers should be experienced and familiar with the area. Counts should be conducted only in winter when there is 100% snow cover and when leaves are gone from deciduous trees and shrubs. Leon et a1. (1987) found no sex or age bias in the composition of whitetailed deer encountered during helicopter counts in Texas. In our study, however, 2 of 6 (33%) of the radio-collared bucks were repeatedly misclassified as does during all replicate flights where they were seen because they had lost their antlers. This was a result of the late date counts were made. Although the sample of radio-collared bucks was small this suggests that buck/doe classifications made in January and February may be quite inaccurate. Misclassifying bucks as does not only results in a low estimate of the buck/doe ratio, it also causes a low estimate of the fawn/doe ratio. Counts to determine population trends can be made in January and February, but classification counts should be made in December, when bucks still have their antlers and fawns are smaller, so both can be more easily distinguished from does. �17 LITERATURE CITED Barnes, A.G., G.T.E. Hill, and G.R. Wilson. 1986. Correcting for incomplete sightability in aerial surveys of Kangaroos. Aust. Wildl. Res. 13:339348. Bartmann, R.M.,L.H. Carpenter, R.A. Garrott, and D.C. Bowden. 1986. Accuracy of helicopter counts of mule deer in pinyon-juniper woodland. Wildl. Soc. Bull. 14:356-363. Beasom, S.L., F.G. Leon III, and D.R. Synatzske. 1986. Accuracy and prec~s~on of counting white-tailed deer with helicopters at different sampling intensities. Wildl. Soc. Bull. 14:364-368. Burnham, K.P., D.R. Anderson, and J.L. Laake. 1980. Estimation of density from line transect sampling of biological populations. Wildl. Monogr. 72. 202pp. Caughley, G., R. Sinclair, and D. Scott-Kemmis. 1976. Experiments in aerial survey. J. Wildl. Manage. 40:290-300. Clover, M. R. 1956. Single-gate deer trap. Calif. Fish and Game 42:199-201. DeYoung, C.A. 1985. Accuracy of helicopter surveys of deer in south Texas. Wildl. Soc. Bull. 13:146-149. Kufeld, R. C. 1987. Development of census methods for deer in plains riverbottom habitats. Colo. Div. Wild1., Wildl. Res. Rep. July (1):1319. Kufeld, R. C. 1988. Development of census methods for deer in plains riverbottom habitats. Colo. Div. Wildl., Wild1. Res. Rep. July (1):1121. Kufeld, R.C., J.H. 01terman, and D.C. Bowden. 1980. A helicopter quadrat census for mule deer on Uncompahgre Plateau, Colorado. J. Wildl. Manage. 44:632-639. Leon, F.G. III, C.A. DeYoung, and S.L. Beasom. 1987. Bias in age and sex composition of white-tailed deer observed from helicopters. Wildl. Soc. Bull. 15:426-429. Samuel, M.D., E.O. Garton, M.W. Schlegel, and R.G. Carson. 1987. Visibility bias during aerial surveys of elk in northcentral Idaho. J. Wildl. Manage. 51:622-630. Prepared by ~c/4tdA Roland C. Kufeld Wildlife Researcher C. �18 Table 1. Deer tagged along South Platte River near Sedgwick, Colorado, and at Dune Ridge Wildlife Area near Sterling, Colorado, January 25, and February 17, 1989. Eartag number 121 122 123 124 125 126 Species mule mule mule mule mule mule deer deer deer deer deer deer Sex doe doe doe doe doe buck Age when Date captured captured adult adult fawn adult adult adult 1-25-89 1-25-89 1-25-89 1-25-89 1-25-89 1-25-89 Capture Location 2 2 2 2 2 2 mi. mi. mi. mi. mi. mi. E. E. E. E. E. E. Red Red Red Red Red Red Radio -collar attached Lion Lion Lion Lion Lion Lion Bridge Bridge Bridge Bridge Bridge Bridge 149.538 149.252 149.170 149.520 149.262 149.770 Table 2. White-tailed and mule deer seen on January 29 - February 13, 1989, helicopter counts of deer in sections of the South Platte River between Greeley and Sedgwick which had deer with radio collars. Species Flight Whitetail 1 2 3 x SE Mule deer 1 2 3 x SE % of radiocollars seen Total deer seen Adjusted total deera Bucks/ 100 does Fawns/ 100 does 69 83 54 728 748 610 1,055 901 1.129 25 23 16 74 74 70 69% 8% 695 43 1,028 67 21 3 73 1 385 436 335 385 468 385 24 17 19 88 77 81 385 29 413 27 20 82 3 100 93 87 93% 4% 2 aAdjusted upward for the percent of radio-collars seen. Example: 83% of radio-collars seen and 748 total deer seen: 748/.83 - 901 adjusted deer seen. �19 L5 120 _ ~ COUNT 1 Imm:::ul COUNT 2 IZ2I ..... COUNT 3 80 60 40 LL. o UJ " ~ 20 z UJ o a:: ~ 0 WHITETAILS MULE DEER Fig. 1. Percentage of radio-collared deer that were in the South Platte riverbottom census area during all 3 helicopter counts. The remaining deer were on the plains away from the riverbottom. ��21 Colorado Division of Wildlife Wildlife Research Report July 1989 JOB FINAL REPORT State of Colorado Project No. ~W_-~1~53~-R~-~3~ Work Plan No. _:2 Job. No. 10 Period Covered: Author: _ Mammals Research _ Deer Investigations Compensatory Mortality in Mule Deer July 1, 1988 - June 30, 1989 R. M. Bartmann Personnel: T. A. Abbott, J. A. Bendykowski, C. A. Blumberg, J. A. Bonamico, L. H. Carpenter, J. N. Caulkins, K. M. Chociej, S. N. DeRuseau, P. M. H. S. L. J. L. M. Dixon, D. L. Dodd, L. J. Draski, E. M. Grody, K. Holzer, Hooker, S. M. Hotchkiss, A. A. Lippacher, D. F. Norman, Shackleford, D. M. Sinor, S. T. Spon, T. R. Stevens, Tandeski, E. K. Weber, G. C. White, and L. C. Zimmerman ABSTRACT Data from the CB and Ridge Study Areas and the pastures are being analyzed for publication as a Wildlife Monograph. A Program Narrative for a new study of the compensatory effects of harvest in a mule deer population was prepared. An experimental late season was established 1-31 December 1989 on the west end of the Ridge Study Area to begin the population reduction required for the new study. Only 33 deer were checked at 2 check stations around the Ridge Study Area during opening weekends of the 3 regular deer seasons. Sixty-four fawns were radio collared on the Ridge Study Area. Forty mortalities were recorded plus 1 radio failure and 1 premature collar dropoff. Excluding these last 2 deer, fawn survival was 35%.' . ��23 COMPENSATORY MORTALITY IN MULE DEER Richard M. Bartmann P. N. OBJECTIVES 1. Accomplish groundwork necessary to conduct an experimental female deer in a portion of GMU 22. 2. Trap and place radio collars on 120 mule deer fawns each winter years. 3. Operate check stations each year during hunting and hunter participation. 4. Monitor fawn survival rates each winter. 5. Analyze data, publish results, seasons and transfer knowledge harvest for for 5 to measure harvest to management. SEGMENT OBJECTIVES Same as P. N. Objectives. METHODS AND MATERIALS Planning was done in preparation for an expanded study of compensatory mortality in a mule deer population. Some preliminary fieldwork was done to aid in developing this study plan. Two check stations were operated during opening weekends of the 3 regular deer seasons to obtain preliminary data on hunting pressure and deer harvest on the Ridge Study Area. One station was at the Little Hills headquarters and the other at the mouth of Hay Gulch. Sex, estimated age, and weight were recorded for each deer checked. Mule deer fawns were trapped with dropnets and radio collared on the Ridge Study Area in November, 1988, to enable continued estimation of survival rates. Radio signals were monitored several days per week throughout the winter and spring and mortalities checked for cause of death. RESULTS AND DISCUSSION Data collected from the CB and Ridge Study Areas and the pastures are being compiled and analyzed to test relationships relating to the mechanisms of compensatory mortality in Piceance mule deer population. Results will be published as a Wildlife Monograph. A Program Narrative (Work Plan 2, Job 15) was prepared for the study of the compensatory effects of harvest in a mule deer population. This is an �24 expansion of previous work under controlled conditions in the pastures. plan was sent through internal review and submitted to Federal Aid. The A late season was established 1-31 December 1989 on the west end of the Ridge Study Area to begin the population reduction required for the new study. There were 375 licenses issued with each licensee allowed to take 2 antlerless deer. Only 23 of the 33 deer checked at the 2 check stations were killed on the Ridge Study Area. Hunter success may have been negatively influenced by the unusually warm, dry weather that prevailed through all 3 seasons. Two-thirds of the deer checked were killed during the second season (Table 1). The few animals checked provided a grossly inadequate sample for any kind of harvest analysis. Until hunting seasons are liberalized to allow a greater harvest of deer, the expense of equipment and personnel required to staff check stations is not warranted. The same information can be obtained more efficiently from field checks of hunters. Sixty-four fawns were radio collared on the Ridge Study Area in November, 1988, to maintain continuity of survival information until the new study can be implemented. As of 15 June 1989, 40 mortalities were recorded. There was 1 radio failure and 1 premature collar dropoff. With these 2 deer removed from the sample, fawn survival was 35%, which is close to average of the previous 6 yrs. There is a possibility of an additional 4-6 radio failures with fawns that have not been contacted for different lengths of time. A final decision will be made on status of these animals after the 1989 trapping session. Table l. Mule deer checked at 2 check stations around the Ridge Study Area during the 1988 regular deer seasons. Season Bucks First 3 Second 7 14 Third 5 2 Prepared by 'RiChard M. Bartmann Wildlife Researcher Does Fawns 1 1 �Colorado Division Wildlife Research July 1989 of Wildlife Report JOB FINAL REPORT State of Colorado Project No. ~W~-~1~5~3_-~R~-~3 _ Mammals Research Work Plan No. ~2~ _ Deer Investigations Job. No. 11 Period Covered: Author: Testing of Mule Deer Census Methodology July 1, 1988 - June 3D, 1989 R. M. Bartmann Personnel: D. J. Freddy ABSTRACT Due to unfavorable snow conditions on the study area in Middle Park, no work was done on this project. A paper entitled, "Evaluation of Aerial Line Transects for Estimating Mule Deer Densities" was published in Vol. 53, No.3 (July, 1989) of The Journal of Wildlife Management. ��27 TESTING OF MULE DEER CENSUS METHODOLOGY Richard M. Bartmann P. N. OBJECTIVES 1. Select area to be censused. 2-. Fly quadrat count to estimate 3. Fly line transects. 4. Analyze data and publish deer numbers results for comparison. in The Journal SEGMENT of Wildlife Management. OBJECTIVES Same as P. N. Objectives. RESULTS Due to unfavorable snow conditions on the study area in Middle Park, no work was done on this project. A paper entitled, "Evaluation of Aerial Line Transects for Estimating Mule Deer Densities" was published in Vol. 53, No.3 (July, 1989) of The Journal of Wildlife Management. Prepared s by Wildlife Researcher ��29 Colorado Division Wildlife Research July 1989 of Wildlife Report JOB PROGRESS REPORT State of Colorado Project No. _W~-~1~5~3~-R~-~3 _ Mammals Research Work Plan No. _ Elk Investigations ~3~ Job. No. 2 Period Covered: Author: Trapping, Transporting, and Maintenance of Elk at Livestock-Elk Grazing Study July 1, 1988 - June 30, 1989 G. D. Bear Personnel: D. Haskins, C. Woodward, C. Reichert, B. Dupire, J. Madison, R. Harthan J. Haskins, M. Bauman, ABSTRACT A total of 257 elk were trapped near Meeker, Craig, and Hayden, Colorado, to provide cows for stocking the experimental pastures on the Little Snake River Wildlife Area north of Maybell, Colorado. The pastures were stocked with 54 adult cows in late December, which were removed in mid-April. Snow accumulations and cold temperatures made it very difficult for the elk to obtain food. Telemetry-collared elk and visual observations indicated elk occupied the Little Snake River winter range from December through April. The collared elk then migrated to summe r ranges 50~6'O mi east: and 40-60 mi southeast of the study area'. ��31 TRAPPING, TRANSPORTING, AND MAINTENANCE OF ELK AT LIVESTOCK-ELK GRAZING STUDY George D. Bear P. N. OBJECTIVE To provide assistance to the Livestock-Elk Grazing Study near Maybell, Colorado, by capturing and maintaining an experimental elk herd. SEGMENT OBJECTIVE 1. To trap, mark, stock, and maintain 54 adult female elk in the livestockelk grazing pasture complex during the period December 15 to April 20. 2. To determine seasonal movements released from the pastures. of radio-collared elk after they are METHODS AND RESULTS Capturing and Handling of Elk The usual amount of effort was put into repa1r1ng the fences and corrals the elk pens, repairing the elk traps, securing hay and other trapping supplies during late summer and fall. at In late November and December trapsites were established at the following locations: (1) 5 mi southeast of Meeker near the Environmental Plant Center; (2) Trapper Mine - 5 mi south of Craig; and (3) Senica Mine - 10 mi southeast of Hayden. Other sites were baited around Craig with minimal success at attracting and holding elk. Late hunting seasons made the Axial Basin sites unacceptable this year. Portable group-traps, baited with alfalfa hay and salt, were used to capture elk. Details for handling elk at the trap sites and in the corral at the livestock-elk grazing pasture complex are presented in previous Federal Aid Reports. To decrease the stress on the elk while they were removed from the trap, the following activities were eliminated: collection of blood samples, weighing of calves, and eartagging. The results were very positive. Cows taken to the pastures were in much better condition, and it required less manpower and time to handle the elk. A group of 20-30 elk in the trap could be sorted and loaded into stock trailers in approximately 30 mins by 3-4 people. A total of 257 elk (102 cows, 126 calves, 9 yearling cows, 10 spikes, 10 adult bulls) were captured from November 27 to February 3. Fifty-four cows were initially used to stock the experimental pastures (December 12 - January 1). These cows were maintained in the pole-corral 2 days to calm them down, then they were released into the "hub" for 4 days to condition them to electric fences before releasing into the experimental pastures. The elk were in good body condition, and the releases went very well. Additional cows were kept in the pole-corral and hub to be used as replacements, when needed. �32 Snow conditions made it very difficult for the elk in the pastures. There were 12 in of snow on the ground after the pastures were stocked; then an additional 18 in fell from January 1 to January 9. A series of clear days reduced this to 8 in with a very heavy crust by late January. In early February, 30 in of snow fell, and heavy winds created drifts 6-8 ft deep in the ravines. Very little vegetation was exposed, only the taller sagebrush plants. Night-time temperatures ranged from 50° F below zero to 32° F below zero. By late February the snow was reduced to 12-16 in deep with a hard crust. Mild temperatures in early March reduced the snow cover so approximately 5% bare ground was showing. Then several unusually warm days (over 60° F) reduced the snow cover to less than 10% in mid-March. Poa and Stipa showed 1/2 -1 in of new growth by late March. Many of the elk in the high-density pastures showed a very rapid decline in body condition. In late January and early February these elk (30 cows) were removed and new elk released into the pastures. In spite of this exchange, it was necessary to replace 8 elk in the high-density pastures and 3 elk in the low-density pastures during March. Elk were successfully removed from the pastures (April 15-25) by baiting them back into the corral, then releasing them down the alleyway. Only 8 elk crossed the pasture fences this year. Six of these crossings occurred during the February blizzard when snow drifts covered the fences. It was necessary to shovel the drifts away from the fences to prevent more crossings. Decline in the number of elk attempting to cross the fences was attributed to the extended training period in the "hub." Elk Movements Telemetry collars were placed on 6 "replacement elk" released from the pens on March 29. They intermingled with large herds of elk, congregating on ridges 10-20 mi south of the experimental pastures; thus remaining in the area until April. Large herds of elk, 50-500 head, were commonly observed in the vicinity of the study area and along Godiva Rim (approximately 7 mi southeast of the experimental pastures) after December 1, then throughout the winter. As the snow melted and green grass became apparent in April, the elk started moving eastward. By May 1, only a few small bands totalling 30-50 elk could be found in the Godiva Rim area. These scattered remnants had moved from the area by mid-May. The 6 newly collared elk and 5 elk collared the previous year were interspersed in the larger herds. They moved along the juniper-covered ridges south of the study area toward Lay, Colorado (approximately 20-30 mi). One of the newly collared elk died enroute. Then 5 collared elk were tracked eastward to the high summer range on Black Mountain and Slater Creek north of Craig and Hayden, Colorado; an airline distance of 50-60 mi. The other 5 collared elk journeyed southeast to Axial Basin, then eastward up Marapos Creek and Williams Fork to high summer ranges near Sand Peak and Pagoda Peak (40-60 mi from the winter range). These migration patterns are similar to patterns reported previous years for these elk herds. Prepared by ~~ GeOrge: sea; Wildlife ~ Researcher �35 JOB PROGRESS REPORT State of Colorado Project No. Mammals W-Is3-R-2 Research Work Plan No. 3 Elk Investigations Job No. 6 Effect of Elk Harvest Systems on elk breeding biology Period Covered: Author: July 1, 1988 - June 30, 1989 D. J. Freddy Personnel: Dr. M. Miller, DVM, D. Hopper, C. Wetherill, M. Cousins, J. alterman, and G. Byrne - Colorado Division of Wildlife; E. Ryland and Forbes-Trinchera Ranch personnel; N. Walsh, Graduate Student, Colo. St. Univ. ABSTRACT We continued evaluating methods for detecting pregnancy in elk. Serum progesterone concentrations averaged 2.98 ± 0.16 for pregnant and 0.30 ± O.lO(SE) ng/ml for non-pregnant cows ~l yr old based on radioimmunoassays. A threshold level of progesterone in pregnant elk may be 1.5 ng/ml which is considerably lower than published estimates. Reproductive tracts from antlerless elk were collected on the Forbes-Trinchera Ranch and from Game Management Unit 33, White River in December-January 198889. For Forbes-Trinchera, pregnancy rates were 29% for yearlings and 83% for mature cows. Fetal sex ratio was 26M:lsF which differed from the l2M:24F ratio of 1987. Conceptions occurred in the 44-day interval from 14 Sep-27 Oct and median conception date was 26 Sep. For White River, pregnancy rates were 83% for yearlings and 92% for mature cows. Fetal sex ratio was 7M:9F. Conc ep t Lons occur redTn the 39-day interval from 15 Sep-23 Oct and median conception date was 26 Sep . . Lower pregnancy rates and more variable fetal sex ratios suggest Forbes elk have a lower quality plane of nutrition than White River elk. Differences in reproduction between herds appear to be independent of differences in numbers or composition of breeding bulis. Small body weights of yearling cows, calves, and mule deer fawns on Forbes are evidence supporting the hypothesis that forage quality may be limiting reproduction. ��37 EFFECT OF ELK HARVEST SYSTEMS ON ELK BREEDING BIOLOGY David J. Freddy P. N. OBJECTIVE To evaluate effects of harvest systems SEGMENT 1. Evaluate 2. Determine Ranch. methods to determine reproductive on breeding biology of elk. OBJECTIVES pregnancy status of elk. status of both elk and deer on the Forbes-Trinchera 3. Provide assistance to the graduate student project investigating bugling behavior of bull elk on the Forbes-Trinchera Ranch. rutting- INTRODUCTION During the mid-1980s, there was statewide concern that elk reproductive rates were declining because numbers of mature bulls were inadequate for breeding (Freddy 1987b). In response, antler-point restrictions to restrict hunting pressure to branch-antlered bulls were instituted to increase numbers of ~2 year old bulls available for breeding. Objectives of this project were to evaluate methods for pregnancy testing large numbers of elk and to compare reproductive rates in herds subject to different hunting systems. The two herds selected for monitoring were White River and Forbes-Trinchera. In 1985, antler-point restrictions in conjunction with unlimited public hunting pressure were implemented in the White River elk herd. The Forbes-Trinchera Ranch has allowed limited private fee hunting for male elk and deer for the last 20 years. Post-season sex ratios since 1986 have averaged 21 bulls:100 cows:56 calves in the White River and 40 bulls:100 cows:57 calves at Forbes-Trinchera. Yearling bulls comprised 83% and 50% of the post-season bulls observed in the White River and Forbes-Trinchera, respectively. METHODS Pregnancy Testing-Blood AND MATERIALS Assays Blood was collected from antlerless elk by hunters on the Forbes-Trinchera Ranch in the San Luis Valley of southern Colorado from 10-12 and 17-19 December 1988. Hunters were instructed to obtain blood from the thoracic entrance to the chest cavity immediately after harvesting the animal and keep the non-heparinized vials of blood cool until depositing vials at check stations. Blood was received from hunters usually within a few hours of when animals were harvested and kept cool until centrifuging within 24-hrs at which time serum was frozen and stored at -20C. Radioimmunoassays (RIA's) for progesterone concentrations were completed in May 1989 by the Physiology Laboratory, Colorado State University, Ft. Collins, Colorado. �38 Fetal Collections Reproductive tracts of female elk were collected by hunters on the ForbesTrinchera Ranch on 10-12 and 17-19 December 1988 and in Game Management Unit 33 (GMU) , White River, in December 1988 and January 1989. GMU 33 represents part of the White River elk population and is located near Rifle, westcentral, Colorado. Tracts were deposited at check stations and kept cool until processing. Pregnancy status was determined from the presence of fetuses, embryos, and developed uterine tissue. Questionable uteri were preserved for later examination. Fetal measurements were made on fresh specimens (subsequently preserved) and followed definitions of Armstrong (1950). Fetal age was determined from growth curves of Morrison et al. (1959). Compliance with the request to collect reproductive organs and blood was excellent and generally, hunters had few problems understanding collection instructions they had previously received by mail (Freddy 1987a). Disease Survey Blood serum samples from elk and deer on the Forbes-Trinchera Ranch were screened in May 1989 for presence of brucellosis and leptospirosis. Elk samples were collected in December 1987 and 1988 and deer samples were obtained in March 1988. All samples were kept at -20C prior to processing. Brucellosis tests were completed by the USDA Laboratory, Denver, CO and leptospirosis tests were done by the Veterinary Diagnostic Laboratories, Colorado State University, Ft. Collins, CO. Forbes-Trinchera Elk and Deer Harvests Male elk and deer were harvested during seasons running from early September into early-December by hunters paying a fee. Eviscerated body weights (head, legs, and hide attached), antler weights, and antler scores (Boone & Crockett) were obtained for bulls and bucks in 1987 and 1988. Antler weights included the frontal bone. Ages of males harvested 1986-88 were based on replacement and wear with dental cementum used secondarily (Freddy 1988). Female elk and deer were harvested in mid-December by hunters participating in the Wildlife Ranching Program. Eviscerated body weights, body measurements, and ages were obtained at mandatory check stations in 1987 and 1988. For 1986-88, ages of antlerless elk were based on dental cementum while ages of antlerless deer were based on replacement and wear for fawns and yearlings and dental cementum for females ~2 yrs old (Freddy 1988). RESULTS AND DISCUSSION Pregnancy Testing Serum progesterone concentrations averaged 2.98 ± 0.16 for pregnant and 0.30 O.lO(SE) ngjml for non-pregnant cows ~l year old collected during December on the Forbes-Trinchera Ranch (Table 1). Progesterone levels in calves averaged 0.85 ± 0.48 and 1.26 ± 0.74(SE) ngjml for females and males, respectively with maximum levels of 6.03 ngjml. The erratic and elevated levels in calves may reflect increased cortisol levels associated with stress from the hunting season (Freddy 1988, Plotka et al. 1983). Progesterone levels were not related to age of pregnant cow (r - -0.08, slope - 0, P>0.50). ± �39 Weber et al. (1982) reported threshold levels of progesterone in pregnant elk of 3.7 ng/ml. Progesterone levels measured in several groups of elk (Freddy 1988) suggest pregnancy commonly occurs with levels below 3 ng/ml. Progesterone concentrations can deteriorate if samples are not properly cooled. A concern, therefore, with samples collected by hunters is that progesterone levels decline between time of death and storage. In 1988, hunter samples were assigned to 3 categories of time between death of the cow and freezing of serum: <24, <14, and <8 hrs. Progesterone concentrations were not affected by time since death (P>0.50, ANOVA) using these coarse demarcations of time. Therefore, we consider these lower levels of progesterone to be accurate. A more appropriate threshold level for progesterone in pregnant elk may be 1.5 ng/ml (Fig. 1). At this time, serum RIA's are preferred over qualitative pregnancy tests (Freddy 1988) because more objective criteria for determining pregnancy status can be developed. We also recommend not using serum progesterone to judge pregnancy status of calves. Fetal Collections-Forbes Pregnancy Elk Rates Pregnancy rates for adult cows (~l yr old) averaged 78% with a range of 7782% from 1986-88 (Table 2). Pregnancy rates approached 90% for prime-aged adults but were 20% for yearlings and <70% for cows aged 11+ years. Pregnancy was not observed in calves. Litter size was 1 in all but 2 (2%) of 113 litters. Two sets of twins, consisting of 2 females and 1 male and 1 female, occurred in cows 9 and 12 years old, respectively, in 1986. Infected uteri not capable of supporting pregnancy occurred in 4 (3%) of 144 adults examined. Infections were in 1 yearling and 3 cows ~13 years old with 3 found in 1986 and 1 in 1988. Fetal Sex Ratios Fetal sex ratios deviated from unity in 1986 and 1987 (X~ P~0.04) and tended to favor males in 1988 (X~ P-0.08). Pooled among all years, fetal ratios were 55M:45F which did not differ from 50M:50F (X~P>0.25) (Fig. 2, Table 3). However, ratios differed among years (X~P-0.004), favoring males in 1986 and 1988 and females in 1987. Fetal sex ratios in mule deer collected on the Ranch also alternated from male to female between 1986 and 1987 (Freddy 1988). Yearly changes in fetal sex ratios could reflect differences in age or body condition of collected cows as fetal sex ratios favoring males are associated with older and dominant cows (Clutton-Brock 1986). Ages of adult cows collected at Forbes-Trinchera were not different among years (ANOVA P>0.50, or X~ P>0.30) (Fig. 3). However, ratios shifting from female to male were associated with increasing weights of cows from 1987 to 1988 (Fig. 4). Changes in weights of cows (P-0.12) represent changes in tissue and not skeletal mass because total body and hind foot lengths were not different between years (P>0.40). Body weight influencing fetal sex was suggested by a marginal interaction of cow weight with fetal sex (P-0.13) and a main effect of yearly weight (P~0.06, ANOVA). �40 Fetal Size Although male fetal weights and crown-rump lengths tended to be larger than females, males were significantly larger only in 1987 (P~0.002; Figs. 5, 6). Male fetal weights and lengths were disproportionatly larger in 1987 (P~0.04, year x sex interaction, ANOVA) when body weights of cows were smallest and fetal sex ratios favored females (Figs. 2, 4). Why male weights and crownrump lengths were greater in 1987 than 1988 (P~0.005), unlike female weights and lengths that remained unchanged (P>0.30), cannot be explained. Fetuses were smaller in 1986 because collections occurred 14 days earlier than in 1987-88. Conception Dates Conceptions occurred primarily between 14 September and 27 October during all 3 years. Median dates of conception were 26, 28, and 28 September for 198688, respectively and 28 September for all years (Fig. 7). At least 97% of the conceptions occurred yearly within a 33-44 -day interval. Conception date was not related to cowage based on linear regression of fetal age on cowage (r- -0.003, n-l09). However, the data suggested that a curvilinear model warrants further analysis. Nevertheless, late breeding cows were arbitrarily defined as the last 25% of the cows breeding yearly. For all years, 64% of these cows (n-25) were ~4 years old and 8% were 11+ years old. This may suggest that young cows have difficulty achieving a threshold body mass adequate for breeding, either initially or after sustaining a pregnancy and lactation. Fetal age was the same for male and female fetuses in both 1986 and 1988 (P >0.40) but in 1987, male fetuses were about 7 days older than females (P=0.002). These comparisons assume identical growth rates for male and female fetuses. Fetal Collections-White Pregnancy River Elk Rates Pregnancy rates for adult cows (~l yr old) collected in GMU 33 averaged 91% with a range of 89-92% from 1987-88 (Table 4). Pregnancy rates were 63% for yearling cows, greater than 90% for prime-aged cows, and 100% for cows 11+ years old. Rates for yearlings and old cows were higher than those observed at Forbes-Trinchera (Table 2). These rates may not be typical of the entire White River elk population, as elk collected in GMU 33 were associated with agricultural damage problems and may have a higher plane of nutrition than elk subsisting on native forage. Neither twins nor infected uteri were observed. Fetal Sex Ratios Fetal sex ratios did not differ from unity in either year or years pooled (X~ P>0.40) and were not different between years (X~P>0.40). The pooled fetal sex ratio was 53M:47F (Table 5). �41 Fetal Size Comparing fetal weights between sexes and years was precluded by the variable dates over which fetuses were collected. Fetuses from GMU 33 (Appendix 1) were collected primarily in the last 15 days of December in 1987 and 1988 and weighed about 250g at an average fetal age of 90 days. At this fetal age, Forbes-Trinchera fetuses would weigh about 2l0g based on a fetal weight vs. age growth curve. This preliminary comparison suggests fetal growth rates may be slower on the Forbes Ranch. Conception Dates Conceptions occurred between 3 September and 27 October in both 1987 and 1988. Median dates of conception were 20 and 26 September for 1987 and 1988, respectively and 20 September for all years (Fig. 8). Conceptions occurred within a 39-day interval both years. Variable median dates occurred when ages of cows remained unchanged (P>0.40), although average cow ages were 5.9 and 5.0 yrs in 1987 and 1988, respectively. Late breeding cows were again arbitrarily defined as the last 25% of the cows breeding yearly. For both years, 58% of these cows (n-12) were ~3 years old and 8% were 11+ years old. Summary There appear to be both differences and similarities in reproduction between Forbes-Trinchera and White River elk populations. Pregnancy rates for yearling and old cows were higher, fetal sex ratios were more consistent, and median conception dates were more variable in the White River than the ForbesTrinchera population. However, fetal sex ratios pooled over years tended to favor males in both populations and "late breeding" cows were composed of similar aged cows. A higher plane of nutrition in the White River is offered as a potential hypothesis for differences between popUlations. A principal premise of this hypothesis being that reproduction in yearling and old cows best reflects nutritional capabilities of habitats. Yearlings and old cows in both populations conceive but conception was higher for these age classes in the White River. Many of the "late breeders" in both populations were yearlings, young cows, and old cows. A second hypothesis follows that with better nutrition more young and old cows conceive and the breeding interval may be lengthened or become more variable among years as habitats could vary in quality yearly. Variable median breeding dates in the White River, compared to Forbes-Trinchera, were somewhat supportive of this argument. The consistency in median breeding dates at Forbes-Trinchera could therefore reflect a static lower quality plane of nutrition and that only prime aged cows consistently conceive. An alternative explanation could argue that the higher percentage of mature bulls .on the Forbes-Trinchera synchronizes breeding which results in more consistent median breeding dates, but this argument cannot explain the lower conception rates in yearling and old cows observed in this population. If late breeding is a function of available bulls, then antler-point restrictions that recruit 2-year old bulls for breeding still result in variable breeding dates as suggested by the White River data. �42 Fluctuations in fetal sex ratios at Forbes-Trinchera may also reflect variable nutritional status of cows. The general trend for both populations to conceive more male calves follows the pattern observed in many other ungulate populations. Post-season calf:cow ratios were nearly identical for both popUlations. Assuming a higher pregnancy rate in the White River, this may suggest that survival rates for calves are lower in the White River. Fetal Collections-Forbes Deer Collections of female mule deer for reproductive and physiological measurements did not occur as planned in March 1989. A large-scale poaching "raid" was conducted by the U.S. Fish and Wildlife Service Colorado Division of Wildlife in the San Luis Valley in early March Scientific collection of deer during the aftermath of this raid was appropriate. antiand the 1989. not deemed Disease Surveys All serum samples from elk and deer on the Forbes-Trinchera Ranch tested negative for brucellosis. The 72 elk samples were from 48 adult females, 11 yearling females, 10 female calves, and 3 male calves. The 12 deer samples were from 11 adult females and 1 yearling male. All serum samples from elk and deer also tested negative for the following 5 strains of leptospirosis: L. hardjo, L.ictero, L. canico, L. grippo, and ~ pomona. The 55 elk samples were from 44 adult females, 7 yearling females, 1 female calf, and 3 male calves. The 12 deer samples were from 11 adult females and 1 yearling male. Forbes-Trinchera Hunter Harvests Age distribution of bulls harvested by fee hunters was consistent from 198688 (Fig. 9). Average a~e and frequency by age class were not different among years (P>0.50, ANOVA, XlY. Average age of all bulls harvested was 4.8 ± O.l(SE) yrs. Age distribution of female elk harvested during late seasons was consistent from 1986-88 (Fig. 10). Average age and frequency by age class were not different among years (P>0.40, ANOVA, X~0. Yearly average ages of female elk ranged from 4.3-5.3 years, inclusive of calves. Average age of male deer harvested by fee hunters was younger in 1986 than 1987-88 (P<O.OOl, ANOVA) reflecting that more bucks in age classes 8+ were harvested in 1987-88 (P<O.Ol, X~2 (Fig. 9). Although total harvest increased among years, the shift to older aged males is thought to reflect changes in timing of seasons. In 1986, deer hunting ceased in early-November, prior to rut, but in 1987-88 seasons continued to early-December during the rut when older bucks were assumed to be more vulnerable. Average age of all bucks harvested was 6.3 ± O.lO(SE) yrs. �43 Age distribution of female deer harvested during late seasons was consistent from 1986-88 (Fig. 10). Average age and frequency by age class were not different among years (P>0.40, ANOVA, X~O). Average age of female harvested was 3.1 yrs, inclusive of fawns. Forbes-Trinchera Antler Scores and Body Weights Gross antler score, antler weight, and eviscerated body weight of bulls steadily increased through age 7 (Fig. 11). At age 8+, antler measurements and body weight tended to decline depending on collection year (Fig. 11). Eviscerated body weights of antlerless elk stabilized at lSO-lS5kg at age 4 (Fig. 13), or 200-207kg whole body weight assuming eviscerated weight equals 75% of whole body weight (Blood and Lovass 1966). Projected whole body weights of yearling females were 125 and l49kg for 1987 and 1988, respectively. These projected weights are below the l52kg threshold reported by Greer (1966) for pregnant yearlings in Yellowstone National Park and is commensurate with the low pregnancy rates observed in yearlings (Table 2). Projected whole body weights for male and female calves combined was 86kg in both 1987 and 1988. Whole body weights for male and female calves combined averaged 110, 116, lOS, and 9lkg, respectively for White River (GMU 23, Dec 1987), and North Park, Durango, and Creede, Colorado, (Jan-Feb 1986) (Bear 1988, CDOW unpubl. data). This general statewide north-south decline in calf weights may reflect differences in total forage/animal unit and/or genetic differences among herds. Gross antler score, antler weight, and body weight tended to increase with age of buck, but not to the degree that occurred with bulls (Figs. 11, 12). Marginal increases in antler score and body weight occurred after ages 6 and 5, respectively. Increases in antler weight but not score after age 6 reflects heavy beamed, short height, and multiple-tined antlers common to older bucks. The trend for body weight of bucks to stabilize at 80kg across all ages (Fig. 12) might suggest nutritional limitations on growth when compared to the steady growth curve presented by Anderson et al. (1974) for bucks in the Poudre River of northern Colorado. However, the 80kg weight is greater than weights projected by Anderson et al. (1974) for all age classes of males and therefore, likely reflects the bias of hunters to select larger bodied bucks from all age classes. Eviscerated body weights of antlerless deer stabilized at 43-45kg at age 3 with a marginal decline occurring at ages 7+, while eviscerated weights of fawns and yearlings were 23-24kg and 34-38kg, respectively (Fig. 13). Anderson et al. (1974) projected peak eviscerated weights of 24kg for fawns, 37kg for yearlings, and 52kg for adult females in the Poudre River which suggests that Forbes-Trinchera adult females could be undersized. Projected whole body weights of fawns was 33kg, assuming eviscerated weight equals 71% of whole body weight (Anderson et al. 1974). Weights of fawns, therefore, probably did not exceed the 33kg fawns of the Piceance Basin, Colorado (Torbit et al. 1988) where starvation in moderate winters accounted for >50% of the over-winter fawn mortality (White et al. 1987). �44 Summary Stability in ages of bulls, bucks, cows, and does harvested suggests that the Forbes-Trinchera populations are capable of supporting current levels of hunter harvest and selection (Appendix 2). The shift to older aged bucks harvested in 1987 and 1988 compared to 1986 demonstrates the effect of season timing on age structure of harvest. Steady increases with age in body size, antler weight, and antler score of bulls is commensurate with expected trends. However, the apparent stability in antler score of bucks after age 6 might argue for short-term increases in the harvest of older bucks to lower the average age of the buck population. A nutrition-density relationship that may be negatively affecting both deer and elk is suggested by elk calves and deer fawns smaller than those measured in other populations in Colorado and yearling elk cows weighing less than considered adequate for breeding. Reducing animal densities or increasing quality and quantity of forage appear to be appropriate management strategies to consider. Literature Cited Anderson, A. E., D. E. Medin, and D. C. Bowden. 1974. Growth and morphometry or the carcass, selected bones, organs, and glands of mule deer. Wi1dl. Monogr. 39. 122pp. Armstrong, R. A. 1950. Fetal development Amer. Mid. Nat. 43:650-666. Bear, G. D. 1988. livestock-elk 40. of northern white-tailed deer. Trapping, transporting, and maintenance of elk at grazing study. Colo. Div. Wildl. Game Res. Rep. Ju1y:33- Blood, D. A., and A. L. Lovass. 1966. Measurements and weight in Manitoba elk. J. Wildl. Manage. 30:135-140. relationships Clutton-Brock, T. H., S. D. Albon, and F. F. Guiness. 1986. Great expectations: dominance, breeding success and offspring sex ratios in red deer. Anim. Behav. 34:460-471. Freddy, D. J. 1987a. Effect of elk harvest systems on elk breeding Colo. Div. Wildl. Game Res. Rep. Ju1y:101-120. 1987b. The White River elk herd: Div. Wild1. Tech. Bull. 37. 64pp. a perspective 1960-1985. 1988. Effect of elk harvest systems on elk breeding Div. of Wild1. Game Res. Rep. July:43-75. biology. Colo. biology. Greer, K. R. 1968. Special collection-Yellowstone elk study, 1967-68. Dep. Fish and Game Fed. Aid Job Comp1. Rep. Prog. W-83-R-11. Colo. Mont. �4:) Morrison, J. A., C. E. Trainer, and P. L. Wright. 1959. Breeding elk as determined from known-age embryos. J. Wi1d1. Manage. seasons in 23:27-34. Plotka, E. D., U. S. Seal, L. J. Verme, and J. J. Ozoga. 1983. The adrenal gland in white-tailed deer: a significant source of progesteFone. J. Wildl. Manage. 47:38-44. Torbit, S. C., L. H. Carpenter, R. M. Bartmann, A. W. Alldredge, and G. C. White. Calibration of carcass fat indices in wintering mule deer. J. Wildl. Manage. 52:582-588. Weber, B. J., M. L. Wolfe, and G. C. White. 1982. Use of serum progesterone levels to detect pregnancy in elk. J. Wildl. Manage. 46:835-837. White, G. C., R. A. Garrott, R. M. Bartmann, L. H. Carpenter, and A. W. Alldredge. 1987. Survival of mule deer in northwest Colorado. J. Wild1. Manage. 51:852-859. /1 Prepared 1/<;' '/ //! by ~/~:.~/~_/~_(~t_t __ 1~/,_/~t_~~~_~~_//'_-+t! David J / Freddy / Wildlife Researcher _ �46 Table 1. Serum progesterone (ng/ml) in elk collected by hunters on the Forbes-Trinchera Ranch in December 1987 and 1988. Statistic 1987 Mean SE min max n 3.20 0.24 0.75 6.79 34 statistic Mean SE min max n a Adults Pregnant 1988 2.82 0.21 0.03 6.31 43 Non-Pregnant Male Adults Non-12regnant 1987-88° 2.98 0.16 0.03 6.79 77 0.30 0.10 0.02 2.14 21 Calves 1987-88 Female 1.26 0.74 0.02 6.03 8 aExcludes obvious outlier of 14.86 ng/ml bNo difference between years P = 0.24 1987-88 0.85 0.48 0.06 2.70 5 �47 Table 2. Pregnancy rates for age classes of female elk collected on the Forbes-Trinchera Ranch in December 1986-88. Agea (yrs) 1986 n Preg 0.5 1 2 3-4 5-6 7-8 9-10 11-12 13-17 Totals Totals 1+ 1 6 7 6 3 4 1 2 5 0 2 7 6 3 4 1 35 34 1987 n Preg Year 1988 n Preg 1986-88 Preg %Preg n 3 3 7 5 15 10 5 3 1 3 0 0 5 15 9 4 3 0 2 2 7 6 19 12 8 2 0 7 0 2 4 17 11 7 2 0 4 6 20 18 40 25 17 6 3 15 0 4 16 38 23 15 6 2 9 0 20 89 95 92 88 100 67 60 28 28 52 49 38 38 63 61 47 47 150 144 113 113 78 2 aAge for ~2 from replacement cementum. and wear, for 3+ from dental Table 3. collected Fetal sex ratios for age classes of female elk on the Forbes-Trinchera Ranch in December 1986-88. Agea (yrs) 1986b M F 1 1 1 4 3 3 2 3-4 5-6 7-8 9-10 11-12 13-17 Totals Total Fetuses 0 2 3 17 0 2 1 0 0 2 1 U M 1 4 1 0 1 0 3 3 2 0 0 0 0 2 1 0 1 6 7 12 30 1987 F 1988 F U M 1 2 8 8 5 1 0 2 5 3 0 0 1 0 0 1 0 0 1 0 0 1 24 2 26 0 2 11 7 2 1 0 38 U 1986-88 M U F Male 2 4 4 0 1 1 0 2 0 0 0 2 0 6 6 15 17 13 10 10 4 2 4 2 1 5 3 0 100 50 47 57 71 33 67 63 15 3 55 45 12 55 2 1 44 1 0 2 0 0 112 for ~2 from replacement and wear, for 3+ from dental cementum. bFetal sex M=male F=female U=unknown. aAge 9,-0 �48 .Table 4. Pregnancy rates for age classes of female elk collected in GMU 33 . White River. in December 1987-88. Year Agea (yrs) 1987 n Preg 0.5 1 2 3-4 5-6 7-8 9-10 11-12 13-19 0 2 6 7 13 5 0 0 4 0 0 6 7 12 5 0 0 4 0 6 2 4 1 2 0 1 2 0 5 1 4 1 2 0 1 2 0 8 8 11 14 7 0 1 6 0 5 7 11 13 7 0 1 6 100 100 37 37 34 34 18 18 16 16 55 55 50 50 91 Totals Totals 1+ 1988 n Preg 1987-88 Preg %Preg n a Ages for 1987 from dental cementum and wear. 63 88 100 93 100 and for 1988 from replacement Table 5. collected Fetal sex ratios for age classes of female elk in GMU 331 White Riverl in December 1987-88. Agea (yrs) 1987b M F U M 1988 F U 1987-88 U M F 1 2 3-4 5-6 7-8 9-10 11-12 13-19 Unkown 0 4 2 6 4 0 0 2 2 0 1 5 6 1 0 0 2 0 0 1 0 0 0 0 0 0 0 2 0 1 0 1 0 1 2 0 3 1 3 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 2 4 3 6 5 0 1 4 2 Totals 20 15 1 7 9 0 Total Fetuses 36 16 % Male 3 2 8 7 2 0 0 2 0 0 1 0 0 0 0 0 0 0 40 67 27 46 71 100 67 100 27 24 1 52 .52 for 1987 from dental cementum and for 1988 from replacement and wear. bFetal sex M=male F=female U=unknown. aAge �49 Appendix 1. Measurements of elk fetuses from Forbes-Trinchera Ranch, 1986-88, and from white River GMU 33, 1987-88, Colorado. Herd/ Year/ Statistic Body Weight Cg) Male Female Crown-Rump Length (mm) Male Female Hind Foot Length (mm) Male Female Forbes 29 Nov-5 Dec 1986 Mean SD min max n 27.6 13.8 12.2 58.5 17.0 20.9 7.0 11.9 29.0 6.0 90.2 16.7 63.5 115.5 17.0 86.0 9.5 72.0 94.0 5.0 20.5 4.8 13.5 28.5 17.0 19.8 2.7 16.5 60.4 33.2 12.0 135.0 24.0 141.4 21.1 113.0 180.0 12.0 115.4 23.3 66.0 151. 0 24.0 42.3 10.9 31.0 68.0 12.0 30.3 8.8 66.7 29.7 23.0 120.0 15.0 120.0 19.9 77.0 163.0 26.0 121. 7 17.9 87.0 150.0 15.0 32.6 7.7 17.0 47.0 26.0 239.2 139.0 140.0 510.0 6.0 184.8 42.0 105.0 230.0 14.0 188.3 33.7 165.0 255.0 6.0 56.6 16.3 59.8 12.3 25.0 47.0 75.0 14.0 80.0 5.0 279.0 166.1 62.0 470.0 6.0 152.8 18.9 140.0 180.0 182.5 44.0 122.0 221.0 6.0 44.5 9.1 36.0 57.0 58.8 21.5 30.0 23.0 5.0 Forbes 12-14 & 19-21 Dec 1987 Mean SD min max n 118.7 54.7 66.0 242.0 12.0 14.0 46.0 24.0 Forbes 10-12 & 17-19 Dec 1988 Mean SD min max n 74.5 35.2 22.0 162.0 26.0 33.1 7.0 20.0 45.0 15.0 White River GMU 33 4-31 Dec 1987 Mean SD min max n 257.9 123.6 49.0 415.0 14.0 White River GMU 33 17-31 Dec 1988 Mean SD min max n 136.3 42.5 103.0 195.0 4.0 4.0 4.0 80.0 6.0 �50 Appendix 2. Numbers of elk and deer harvested 1986-88. Trinchera Ranch a Reg. Season on the Forbes- Late sea son'' Species Adult Male Adult Male Adult Female 1986 1987 1988 Elk Elk Elk 72 93 87 3 6 3 34 51 65 1 6 8 1 4 7 0 0 0 1986 1987 1988 Deer Deer Deer 87 109 108 8 6 5 50 118 117 1 7 6 11 14 9 3 Yr Male Young Female Young Unk. 0 0 Regular season refers to fee hunting seasons: 13 Sep-9 Nov 1986, 5 Sep-9 Oct and 14 Nov-l1 Dec 1987, 10 Sep-9 Dec 1988. a Late-season refers to public hunting seasons under the wildlife Ranching Program: 29 Nov-5 Dec 1986, 12-14 and 19-21 Dec 1987, 10-12 and 17-19 Dec 1988. All animals processed at check stations. b �51 100 ,, ,, I- Z UJ ,----------' 80 o a: UJ o, 60 UJ > ~ => :E => o 40 PREGNANT 20 ------. o o n NON-PREGNANT 2.0 3.0 4.0 PROGESTERONE ng/ml 1.0 = 78 n = 21 5.0 6.0 Fig. 1. A threshold of 1.S ng/ml serum progesterone may be appropriate for differentiating pregnant and non-pregnant adult elk. 60 MALE 50 en UJ en 40 tuu. 30 :5 w 20 ~ FEMALE => 10 o 1986 1987 1988 ALL YEAR Fig. 2. Fetal sex differed from SOM:50F yearly (P~ 0.08) but not over all years (P>0.40). Fetal sex alternately favored males and females (P < 0.004) on the ForbesTrinchera Ranch, 1986-88. �52 10 -ea:n 9 w 7 -c C 8 o w 6 -c 5 3: 4 o 3 0 ~ ...J w 28 38 44 1987 1988 2 1 0 1986 YEAR Fig. 3. Ages of elk cows with fetuses were stable among years (P>O.50) on the ForbesTrinchera Ranch, 1986-88. Sample sizes above vertical lines showing + 2(SE). en -::r: 160 41 (!) ~ I- o w 31 150 3: 3: 0 o ~ ...J 140 w 130 1987 1988 YEAR Fig. 4. Weights of elk cows with fetuses tended to increase in 1988 (P=O.12) on the Forbes-Trinchera Ranch. Sample sizes above vertical lines showing +2(SE). �53 160 -(!) J- J: (.!J ....J 80 UJ 60 u.. ~ ....J UJ ~ FEMALE 120 100 s MALE 140 UJ ;: 12 40 26 15 45 17 6 20 0 1986 1987 1988 ALL YEAR Fig. 5. Male and female fetal weights were different only in 1987 (P<O.001) on the Forbes-Trinchera Ranch. Sample sizes above vertical lines showing +2(SE). -::E ~ a.. ::E ::> MALE 140 120 ;: 100 o a: o 80 ....J 60 ~ 40 u.. FEMALE 160 a:, z ~ 20 o 1986 1987 1988 ALL YEAR Fig. 6. Male and female fetal crown-rump lengths were different only in 1987 (P<O.002) on the Forbes-Trinchera Ranch. Sample sizes above vertical lines showing +2(SE). �54 50 III en Z 0 ln, W ~ YEARLY MEDIAN DATES 40 1986 ~ n = 26 30 1987 n 37 = o 1988 n= 44 o Z 0 o I- 20 Z w o c: W 10 a. 0 9/8 9/18 9/28 10/8 10/18 10/28 11/7 11/17 40 • en z 0 MEDIAN DATE 30 •• ICl. UJ o Z 0 o 20 1986-1988 n = 107 I- Z UJ o 0: UJ Cl. 10 o 9/8 9/18 9/28 10/8 10/18 10/28 11/7 11/17 MONTH AND DAY (Begin 5-day intervals) Fig. 7. Median conception date for elk was consistent among years (TOP) and was 28 September for all years (BOnOM) on the Forbes-Trinchera Ranch, 1986-88. �55 50 I 40 I )c )c )c )c z I0W 0 Z 0 30 f- ~ ~ x 0 >: x 20 x r-- 0 Z W 0 a: 10 w f-- ~ 00 I I L...- 9/8 tx tx tx tx tx tx tx ~ )< )< )< n = 16 ~ ~ ~ ~II I 9/18 1988 )c )< )< )c )< )< )c )c ~ ~ ~ n = 36 x )< )< )< )< Dr: 111987 )c x x x ~ I- YEARLY MEDIAN DATES ~ en 0 -==. 9/28 I 10/8 I 10/18 I I 10/28 r T T 11/7 I 11/17 40 • en Z MEDIAN DATE 30 1987-88 n = 52 0 b:UJ 0 Z 20 0 0 I- Z UJ 0 a: 10 UJ 0- o 9/8 9/18 9/28 10/8 10/18 10/28 11/7 11/17 MONTH AND DAY (Begin 5-day intervals) Fig. 8. Median conception date for elk varied about 1 week between years (TOP) and was 20 September for all years (BOnOM) in GMU 33, White River, 1987-88. �56 - 1986 n = 72 ~1987 AVG.AGE '" 4.8 30 n = 93 [2LJ1988 n = 87 AVG. AGE = 4.8 AVG. AGE = 4.9 BULLS ~ Z 20 UJ ::> o UJ c:: L1. 10 o 2 3 5 4 1986 n = 87 30 ~ ~ 7 8+ 1987 n = 109 ~ AVG. AGE •• &.& AVG. AGE", 5.& 25 6 1988 n = 108 AVG. AGE. = 6.8 BUCKS 20 Z UJ ::> o 15 LU c:: L1. 10 5 o 1-3 4 5 6 7 8 9+ AGE (YEARS) Fig. 9. Ages of bulls harvested on the Forbes-Trinchera Ranch were consistent among years (P>O.50) (TOP) while younger bucks were harvested in 1986 (P=O.01) (BOnOM). �57 30 1986 n = 35 FEMALE ELK AVG. AGE = 5.3 B;83 1987n=61 AVG. AGE = 4.3 20 ~ 1988 n = 71 ~ AVG. AGE = 4.9 Z W :::> 0 w a: u, 10 0 2-3 0 4-5 6-8 9-20 50 III FEMALE DEER 1986n=61 AVG. AGE = 3.1 40 1987 n = 127 ~ AVG. AGE = 3.1 0 ~ Z W 1988 n = 124 30 AVG. AGE = 3.2 :::> 0 w c: U. 20 10 o o 2 3 4 5 6 7+ AGE (YEARS) Fig. 10. Ages of female elk (TOP) and deer (BOnOM) harvested on the ForbesTrinchera Ranch were consistent among years (P>0.40). Male calves and fawns not included. �58 400 w 350 CC 300 o en 250 CC 200 o w .=z « I)•• BULLS 1987 n = 83 ~ 1988 n = 83 150 100 50 o - 2 3 4 5 6 15 IE 1987 n BULLS -J: 7 ~ ~ 8+ = 83 1988 n = 83 •••• 10 (!) w 3: c:: ~ 5 !z-c o 2 3 300 4 5 6 7 8+ 7 8+ & 1987n = 83 6250 -s: :::.:::: 200 o w 150 3: >Q o CO 100 50 o 2 3 456 AGE (YEARS) Fig. 11. Average gross antler score (TOP), antler weight (MIDDLE), and eviscerated body weight (BOTTOM) for bulls harvested on the Forbes-Trinchera Ranch. �59 200 c: 180 o en 160 L1J 1987 n = 102 - BUCKS 1988 n = 102 ~ o c: ~ 140 I- Z ~ 120 100 - 5 2 3 4 5 6 BUCKS 7 8 1987 n = 102 1988 n = 102 ~ ~ 9+ I- 4 J: 3 (!:) L1J ~ c: 2 ~ !z~ 1 o 140 2 3 4 5 6 7 BUCKS 8 IQQg 9+ 1987 n 1988 n 8'120 = = 102 102 ~ -100 a I- ~ s g 80 60 40 20 o 2 3 456 7 8 9+ AGE (yEARS) Fig. 12. Average gross antler score (TOP), antler weight (MIDDLE). and eviscerated body weight (BOnOM) for bucks harvested on the Forbes-Trinchera Ranch. �60 175 ~ ANTLERLESS ELK ,. 1987 n = 50 ~ 1988 n = 85 150 125 l- I C!J 100 w ~ >- 75 a co 50 c 25 o 50 8' - o 2-3 1 4-5 9-20 6-8 III 1987 n = 129 ~ ANTLERLESS DEER 1988 n = 12 40 ~ l- I C!J 30 UJ ~ >- o a co 20 10 o o 1 234 5 6 7-15 AGE (YEARS) Fig. 13. Average eviscerated body weights for antlerless elk (TOP) and deer (BOnOM) on the Forbes-Trinchera Ranch. Male and female young pooled for age class o. �JOB PROGRESS REPORT State of Colorado Project No. W-lS3-R-2 Mammals Research Work Plan No. 3 Elk Investigations Job No. 7 Elk Census Methodology Period Covered: Author: Personnel: July 1, 1988 - June 30, 1989 D.J. Freddy G. White, L. Carpenter, R. Bartmann ABSTRACT Line transects and quadrats were not flown because heavy snowfall severely concentrated elk and deer. Line transects and quadrats will be flown in December-January 1989-90. A brief description of methods to be used is included for reference. ��63 JOB PROGRESS REPORT ELK CENSUS METHODOLOGY David J. Freddy P.N. OBJECTIVE Evaluate methods to estimate numbers of elk during winter. SEGMENT OBJECTIVES 1. Determine efficacy of line transect census methodology of elk and deer on winter ranges shared by both species. to estimate numbers 2. Compare estimates of deer density based on line transects and quadrats (existing census system) in an area where both methods can be employed. STUDY AREA The Troublesome portion between Highway 125 and selected for evaluating estimate numbers of elk numbers of elk and deer efficiently sampled. of Game Management Unit (GMU) 18 within Middle Park, Troublesome Creek and north of the Colorado River, was aerial line transects (Burnham et al. 1980) to and deer during winter. This area has significant residing in a relatively small area that can be METHODS AND MATERIALS The sampled area was limited to those areas inhabited by elk and deer during January and February. Deer winter range in the Troublesome area for this time period was previously defined by Gill (1969) and encompassed l66km2 (64mi~ of sagebrush habitat. Elk winter range was based on distribution of elk as observed during sex and age classification surveys done in January and February in recent years (pers. comm. B. Thompson, R. Firth) and encompassed 246km2 (95mi~ of sagebrush, aspen, aspen-conifer, and heavy conifer habitats. In general, deer winter range occurred from 2257-2685m (7400-8800ft) and elk winter range included the deer winter range and areas to 2990m (9800ft.) in elevation. The 2 defined winter ranges were outlined on 1:24000 topographic maps. Two sets of line transects each containing 26 transects were selected and marked on topographic maps. Initial starting points for sets of transects were 500m apart. Transects within each set were systematically spaced at 1000m intervals, oriented on true north bearings, and were l-15km long. Each set of transects simultaneously sampled the defined elk and deer winter ranges. Beginning points of transects were not marked with flight markers. �64 We expected to encounter 130 and 150 groups of elk and deer, respectively, if each set of line transects were flown twice (Freddy 1988). This number of groups would allow reasonable evaluation of sighting curve functions and provide expected coefficients of variation of + 15% about mean densities of elk and deer (White et al. 1989 In press). We wanted to compare estimates of deer densities based on line transects and quadrats. Deer densities in Middle Park have been estimated since 1968 using a stratified sampling system that employs the 2.6km2 (lmi~ quadrat as the sampling unit (Gill 1969). The Troublesome strata coincides with the area sampled by line transects. In addition to the 10 existing quadrats in the Troublesome, 20 additional quadrats were selected at random and marked on topographic maps. These 30 quadrats and the line transects sampled the same 166km2of deer winter range in the Troublesome area. Key corners of all quadrats were marked on the ground with orange markers to aid in navigating to quadrats. We calculated that a difference in deer densities between transects and quadrats of 50% or more could be detected with 95% confidence if we flew 30 quadrats. These calculations were based on variances about average densities of deer (10/km~26/mi~ measured on 10 quadrats in the Troublesome in 12 previous years. We also assumed a sightabi1ity of 60% for deer on quadrats (Bartmann et a1. 1986) and observation of 150 groups of deer on transects. Quadrats were to be flown once. All flights were to be done using a Bell-Soloy helicopter. A navigator and observer, in addition to the pilot, were responsible for observing elk and deer. The observer seated on the right estimated perpendicular distances to and counted those groups of elk or deer located from the transect center line to the right. The middle observer (navigator) estimated perpendicular distances to and counted only elk to the left of the transect center line. Markers placed at known distances from a practice transect line were used by observers to practice estimating perpendicular distances prior to flying each day. Transects and quadrats were to be flown within a 3-day period. Both sets of transects were to be flown on both the first and third days with sets of transects randomly selected to be flown in either the morning or afternoon. Quadrats were to be flown throughout the second day. RESULTS AND DISCUSSION Line transects and quadrats were not flown because heavy snowfalls in late January and early February severely concentrated elk and deer just prior to the scheduled timing of the census. One major storm left 75cm (30in) of snow on portions of the winter range. Many elk concentrated around haystacks or in feed yards of domestic cattle. Censusing animals under these conditions would have resulted in undue harassment of elk, deer, and domestic cattle, damage to private property, and poor estimates of animal densities. Extreme conditions persisted until the end of March. Line transects and quadrats will be flown in late December or early January of 1989-90. �6.5 LITERATURE CITED Bartmann, R.M., L.H. Carpenter, R.A. Garrot, and D.C. Bowden. 1986. Accuracy of helicopter counts of mule deer in pinyon-juniper woodland. Wildl. Soc. Bull. 14:356-363. Burnham, K.P., D.R. Anderson, and J.L. Laake. 1980. Estimation of density from line transect sampling of biological populations. Wildl. Mono. 72. 202pp. Freddy, D.J. 1988. Elk census methodology. July (1):78-82. Colo. Div. Wildl. Game Res. Rep. Gill, R.B. 1969. A quadrat count system for estimating game population. Colo. Game, Fish, & Parks Game Info. Leaflet 76. 2pp. Gill, R.B. 1969. Middle Park deer study population density and structure. Colo. Game, Fish, & Parks Game Res. Rep. July (1):106-122. White, G.C., R.M. Bartmann, L.H. Carpenter, and R.A. Garrot. Evaluation of aerial line transects for estimating mule deer densities. J. Wildl. Manage. 53(3): In press. 7 - /; ./~/-//1 Prepared by .a(~.~/_·Z~-~i_~d_·_·~~~.JI-+ ~~~c~~~:IL~;"/_L~~~ David J .•Freddy 7 Wildlife Researcher _ �� https://cpw.cvlcollections.org/files/original/6c11c8f2c0318d707b31e9287f89ce18.pdf cb2dc74a27f04d7fe6153f2ca220a149 PDF Text Text 67 Colorado Division Wildlife Research July 1989 of Wildlife Report JOB PROGRESS REPORT State of Colorado Project No. ~W_-~1~5~3_-~R~-~3 _ Mammals Research Work Plan No. _ Multispecies ~l~A~ Job. No. 1 Period Covered: Author: Wild Ruminant Dynamics Investigations Forage Selection July 1, 1988 - June 30, 1989 P. H. Neil Personnel: M. W. Miller, R. B. Gill, M. L. Stevens, D. L. Baker, K. A. Trust, K. A. Sheehan, B. J. Maynard, M. A. Wild, H. J. Lucking, and T. Ritchie ABSTRACT Eight pronghorn fawns were successfully raised and are being used in controlled experiments to test the light wavelength perception range of pronghorn. Facility repairs continued, and 3 shelters in the bighorn sheep pens were rebuilt and 1 shelter in the elk pens. An additional section of alleyway was constructed to accommodate a portable squeeze chute for elk. All animals at the facility are presently healthy. ��69 ANIMAL AND PEN SUPPORT FACILITIES FOR MAMMALS RESEARCH Paul H. Neil P. N. OBJECTIVES Provide and maintain populations of captive animals and pen facilities support Mammal and Avian Research Programs. AGREEMENT to OBJECTIVES 1. Maintain and improve animal research 2. Coordinate 3. Maintain up to 15 elk, 30 mountain sheep, 30 pronghorn antelope, 15 mule deer, and 11 domestic cows in suitable health to perform required research experiments. 4. Conduct management experiments efficiency and efficacy. all rearing, facilties. training, maintenance, and research activities. to increase feeding and maintenance METHODS AND MATERIALS Routine neonate rearing procedures were used to hand rear 8 pronghorn. Eight bighorn sheep lambs were born at the facility and left on the ewes to raise. Daily contact was made with the animals to enhance tameness and, in most cases, seems to be a satisfactory procedure for our needs. Three shelters were rebuilt in the bighorn sheep pens using a different design to cut down on wind damage and animal damage. RESULTS AND DISCUSSION Eight pronghorn antelope are being used in controlled experiments to test the light wavelength perception range of pronghorn, specifically to test for visual perception in the region of near infrared wavelengths. Details of progress in this experiment are described under WPlA,J4. The 11 elk that were used in the elk-cattle competition study returned to the facility in good shape. Details of their use are described under WP9A,Jl and WP3,J5. Expansion of the pronghorn pens began during the fiscal year and is near completion. This project will relieve some pressure on our existing pens. �70 All animals at the facility are presently healthy. The big game research herd presently consists of 11 elk, 31 bighorn sheep, 1 Rocky Mountain goat, 11 domestic cows, and 23 pronghorn antelope. Prepared by \~A \~ ;_._~C i IA.' __ ~~ Paul H. Neil Wildlife Research Tech III _ �71 Colorado Division of Wildlife Wildlife Research Report July 1989 JOB PROGRESS REPORT State of ~C~o~l~o~r~a~d~o~ Project No. _ W-lS3-R-3 Mammals Research Work Plan No. __~l~A~ _ Multispecies Job No. __~ __-=3 _ Mammals Research Period Covered: Investigations 2 Administration July 1, 1988 - June 30, 1989 Author(s): R.B. Gill Personnel: L. Lovett, N. McEwen ABSTRACT All administrative objectives listed in the Segment Narrative for Mammals Research 2 for FY 1987-88 were accomplished. In addition a draft River Otter Recovery Plan and a draft Black Bear Management Plan were prepared and submitted for review. ��73 MAMMALS RESEARCH 2 ADMINISTRATION R. Bruce Gill P.N. OBJECTIVE Administer research within productivity at the lowest the Mammals cost. SEGMENT Research 2 Unit for the highest OBJECTIVES 1. Require candidate research research project selection funding. 2. Require all research study plans to be peer reviewed, reviewed and approved by the Research Leader before the research is initiated. 3. Require quarterly research study. 4. Subject all research projects to periodic review Division's formal administrative review process. 5. Allocate fiscal resources and track expenditures of each research to assure fiscal responsibility and accountability. 6. Maintain a manual on current standard administrative operating procedures and distribute to each Wildlife Researcher. 7. Develop, implement, and maintain an effective information transfer process that assures research results are disseminated to appropriate information users in a form that maximizes the probability that the information will be used. progress projects to go through the Division's process before they are considered for reports RESULTS and annual progress reports according of each to the study AND DISCUSSION Objective 1. A draft river otter recovery plan and draft black bear management plan were prepared. These plans will be used as the basis for selecting research on the 2 species. Objective 2. disease Planning was begun on a study to evaluate the role of in the ecology and management of mountain sheep. Objective 3. Quarterly reports were received and reviewed for each Mammals Research 2 study active during FY 88-89. Annual Job Progress Reports were prepared and submitted to the Federal Aid planner by the September 15 deadline. Objective 4. Mountain sheep, pronghorn, river otter, and elk/cattle grazing studies were reviewed in the field during the Segment. Field work has ended for the mountain lion study so it was not reviewed. Objective 5. The resource allocation to the Mammals Research 2 cost �74 center for FY 88-89 was further allocated to each Wildlife Researcher for each research study. These individual study allocations were prepared in Segment Narrative format and submitted to Federal Aid Division of the U.S. Fish and Wildlife Service for approval and P-R funding. Budget status reports were prepared quarterly for each Wildlife Researcher. A final budget status report issued 7/24/89 indicated that the Mammals Research 2 Section expended 100.1% of the $576,114 allocated for FY 88-89. Objective 6. updated The Standard Administrative Operating at least quarterly during the segment. Procedures manual was Objective 7. An adaptive assessment simulation model was developed during the segment to assist DOW in developing plans to co-manage sympatric populations of mountain sheep and mountain goats. The model integrated published literature on mountain sheep and mountain goat ecology and quantified "best guess" scenarios where experimental data were lacking. This model is being prepared for publication. It has been reviewed several times with DOW field personnel. The Black Bear Management Plan is now in it 4th draft version and incorporates the results of Colorado's black bear research study with those of other bear studies to identify critical management issues and strategies. This plan was reviewed and incorporated revisions from DOW personnel and nearly all public groups which expressed an interest in citizen participation in the black bear planning process. Prepared bY:_~__'_'_~~\;.J...--) R. Bruce Gill Wildlife Research ~_ Leader �75 Colorado Division of Wildlife Wildlife Research Report July 1989 JOB PROGRESS REPORT State of Colorado Project No. Work Plan No. ~l~A~ Job. No. 4 Period Covered: Authors: Personnel: Mammals Research W-1S3-R-3 _ Multispecies Investigations Wild Ruminant Forage Selection Dynamics July 1, 1988 - June 30, 1989 B.J. Maynard P.N. Lehner, M. Miller, S. Roberts, R.B. Gill ABSTRACT Much of the'Segment was spent training pronghorn to respond consistently to white light stimuli in a nonrandom fashion. Trial and error modifications were made in light source, brightness, and food reward. Additional improvements were made in training/conditioning protocol to sharpen the ability to detect differences in pronghorn light wavelength percepcion. Initial physiological tests (electroretinograms) were scheduled to determine pronghorn light wavelength recepcion at the retina. ��77 'WILD RUMINANT FORAGE SELECTION DYNAMICS B.J. Maynard P.N. OBJECTIVE To evaluate the role of color vision foraging ruminant--the pronghorn. in diet selection SEGMENT a detailed of a small, selectively OBJECTIVES 1. Prepare Program Narrative. 2. Foster-rear 3. Conduct controlled experiments to test the light wavelength perception range of pronghorn, specifically to test for visual perception in the region of visual and near infrared wavelengths (ca 500->700 nm) or (ca 500-800 nm). up to 8 pronghorn fawns to use in research studies. METHODS Eight hand-reared pronghorn were maintained at the Colorado Division of Wildlife's Foothills Wildlife Research Facility (FWRF) on a diet of alfalfa hay, pellets, water, and mineralized salt, fed ad libiLum. Pronghorn were weighed weekly to monitor overall health and growth performance (Fig. 1). All animals appeared to grow normally and were maintained in good health throughout the segment. Wavelength discrimination trials were continued throughout the segment (See Appendix A for specific discrimination trial methodology). For each trial, individual pronghorn were led into the start box of a Y-maze and required to choose between a "light on" branch and a "light off" branch. The "light on" branch held a food reward within a goal box at the end of the branch; the "light off" branch" did not offer a food reward. The intention is for pronghorn to learn the light/food association and to enter the "light on" branch significantly more often than one would expect if choices were made at random. Once pronghorn have learned to consistently discriminate "light on" from "light off", training will be extended to wavelengths of narrow color bandwidths. These trials will evaluate which wavelength bands pronghorn perceive. More importantly, once the protocol is established and pronghorn have learned to make light-related choices in the maze, the nature of the stimuli can be adapted to generate behaviorally a spectral sensitivity curve for the pronghorn visual system. This information will be fundamental to unraveling the role vision plays in pronghorn diet selection. �78 PRONGHORN GROWTH RATES 1988-1989 WEIGHT (KG) 50 r-----------------------------------------~ 45 40 ----- -----------.-----.-------.-.---------.---.. ---.-- - ---.-----.--------.. -.------.. 30 25 20 15 10 5 0 0 10 Summer 1988 Fig. 1. Pronghorn 20 30 40 WEEKS FROtl BIRTH growth rates, summer 1988-summer 50 Summer 1989 1989. RESULTS In July an analysis of Y-maze trial data revealed no discernable pattern to the choices pronghorn were making in the maze. There was no indication that the animals were learning the light/food association nor was there any recognizable pattern or strategy to their choices. At this time several changes were made to increase both light saliency and subject motivation. A light stimulus not obvious to a subject in the maze could inhibit learning of the light/food association. Through July 1989, the light sources had been high intensity lamps which shone directly through one-inch diameter holes in the end wall of each maze branch. In July the diameters were changed to 3 inches, and white plexiglass "screens" were inserted in front of the lamps. The high intensity lamps were extremely bright to the human eye, to the point of being mildly painful if the line of sight was directly into the beam of the light. The plexiglass decreased the brightness of the lamps so that they were still readily visible to a human subject, but not at all painful, and so �79 should have eliminated any possibility that the brightness of the lights was in fact aversive to the pronghorn. Also in an attempt to increase light saliency, the light was flashed while the pronghorn were in the start box. In order to further enhance the light/food association, several more changes in protocol are necessary. Currently, the lights in each branch of the maze are 32 inches above the food trays, which are recessed into feed boxes. Lights will be lowered to 12 inches above the food trays, facilitating the association between light and food. In addition to inadequate stimulus saliency, inadequate motivation of the animals to "solve" the problem could have led to poor performance. The animals were fed hay and pellets ad libi~um throughout the day and were rewarded with raisins and marshmallows in the maze. Cafeteria trials run during Summer 1988 indicated that these 2 foods were the best rewards for pronghorn. In July 1989, more cafeteria trials, offering sugar cubes, fruit drink, and gummy bears, were run in an attempt to find a better reward. The decision to offer these foods was based on their similarity to the chewy consistency and the sweetness of raisins and marshmallows. No food was found that the animals would ingest as frequently and rapidly as the marshmallows, raisins, and hay and pellets. Another procedure used in operant conditioning research to enhance stimulus/reward (light/food) associations classically is first to condition the pronghorn to associate light with food. The animals are placed in an enclosed area and presented light numerous times just prior to food presentation. Thus, the light signals food arrival and pronghorn should soon cue light with food. If the other changes made to enhance light saliency do not result in pronghorn responding to the light, then this procedure will be carried out. Paired with the behavioral research of pronghorn light percep~ion by the brain, are physiological tests to determine pronghorn recep~ion of light by the retina. An electroretinogram (ERG) will be conducted at the Colorado State University Veterinary Teaching Hospital on August 8. Dr. Steve Roberts (D.V.M.) will conduct the ERG and Dr. Mike Miller (D.V.M.) and Margaret Wild will administer anesthesia. Data obtained from this and future ERGs will be. combined to generate a spectral sensitivity curve for the pronghorn retina. Prepared by ��81 Appendix A Pronghorn Reception and Perception of Visible and Infrared Light Study Plan Revised July 1, 1989 Barbara J. Maynard A. NEED Pronghorn(Antilocapra americana) inhabit the semi-arid grasslands of North America, where precipitation is low ( < 25 cm annually; Beale and Smith, 1970; Boyle, 1981). In this environment, water might be a limiting resource of pronghorn populations. Indeed, Beale and Smith(1970) correlated pronghorn fitness(number and health of offspring) with water availability. In addition, water consumption from troughs decreased as forage succulence and temperature increased. As forb (a primary diet ingredient) succulence and volume decreased, pronghorn switched to the most succulent forage available. Beale and Smith offer no evidence that these relationships are causative rather than correlative; however, the indication is that water is an important limiting resource for pronghorn. Based on pronghorns' relatively small size and high metabolic rate, quality and digestibility of forage also should be important factors in pronghorn diet (Boyle, 1981). Krueger (1986) found that pronghorn foraged in areas where forbs and shrubs had higher nitrogen content than other areas. Likewise, Schwartz, et al. (1977) found that pronghorn were more selective grazers than sheep, bison, and cattle. Pronghorn ate forage of higher crude protein and lower fiber content than did the other three species. Numerous studies similar to those mentioned above have documented that pronghorn are selective foragers; how pronghorn select their food, however, is not known. Visible and infrared reflectances by plants provide a reliable index of plant succulence and productivity(discussed below). The present study is designed to investigate the capabilities of the pronghorn visual system in the visible and infrared spectra. Vegetation differentially reflects incident electromagnetic radiation in the visual and infrared spectra based on different plant qualities. Below 400 nanometers (nm) and above 2700nm, leaf reflectances are generally less than five percent of incident light and are fairly uniform (Knipling, 1970). Within this 400--2700nm range, however, reflectances vary from plant to plant and reveal features of plant quality. Water content is responsible for absorption of radiation in the mid-infrared range, 1300--2400nm. That is, plants with higher water content absorb more mid-infrared radiation than do plants with lower water content (Knipling, 1970; Hunt, et aI., 1987). Reflectance in wavelengths other than the mid-infrared band can also provide information about plant quality. Plots treated with different levels of nitrogen fertilizer can be distinguished in the red(640--660nm) and near-infrared(820--830nm), as well as midinfrared, bands (Everitt, et aI., 1987). Fertilized plots reflect a higher percentage of red and near-infrared radiation and absorb a higher percentage of mid-infrared than do unfertilized �82 plots. The different red and near-infrared reflectances can be attributed to higher productivity resulting in higher chlorophyll content. The mid-infrared reflectivities can again be attributed to higher water content, as a result of higher productivity (Everitt, et aI., 1987, 1986). The presence of visual pigments sensitive to long wavelengths has been demonstrated in some vertebrates including the freshwater fish, the rudd (Scardinius erythrophthalmus, 620nm, Loew & Lythgoe, 1978) and the goldfish (Carassius auratus, 625nm, Harosi & MacNichol, 1974). Color vision has been demonstrated in mammals such as giraffes, pigmy goats, red deer cows, and domestic goats, (Backhaus, 1959b; Backhaus, 1959a; Buchenauer & Fritsh, 1980). Pronghorn visual capabilities, however, are not well enough studied to understand their role in pronghorn forage selection. B. EXPECTED RESULTS AND BENEFITS This study will provide an understanding of pronghorn visual capabilities in the infrared and visual spectra, which will then generate further hypotheses and research into pronghorn foraging strategies. C. OBJECTIVE To determine radiation. the extent of pronghorn reception and perception of visual and infrared D.APPROACH Subjects Five female and three male pronghorn, born in June 1988, were bottle-raised at the Foothills Wildlife Research Station in preparation for this study. Five of the animals were captive-born; two were wild-caught; and one was found as an orphan. Apparatus Y-Maze: A T-high, wooden Y-maze has been constructed in the pronghorns' pen. The buzzer, the starting gate, and the two exit gates can all be controlled by an experimenter behind the start box. The T-shaped start gate, which raises and lowers in a channel iron track, is of wire mesh so that subjects can see through it, but not pass it. A video camera mounted above the maze gives a view from the start gate to slightly beyond each "choice" line. These "branch" and "choice" lines are wooden slats painted white and buried to ground level. The top of the maze is eighty percent covered by plywood, which prevents most, but not all, incident sunlight from shining into the maze. A three-sided "sun shield" will surround the light stimulus in each branch in order to prevent sunlight from directly interfering with the stimuli. The areas inside these shields will be painted flat black in order to accentuate the stimuli and to reduce reflections. Light Sources and Filters: High intensity tungsten light sources(Central Scientific Company, Franklin Park, IL, model no. 85262-01} will be mounted on top of the feed boxes behind the end wall of each branch of the maze and enclosed in black plexiglass boxes. Tungsten bulbs were selected for their high radiation in longer wavelengths (Riggs, 1966a). All filters will be 1"-diameter glass. Narrow bandpass filters with bandwidths of 10nm �83 and central wavelengths of 500, 550, 600, and 650nm will selectively transmit light of narrow wavelength bands. A long wave pass filter with cut-on wavelength (50% transmission) of 700nm will transmit radiation of wavelength longer than 700nm and will block radiation of wavelength less than 700nm. Procedure This study involves two phases: (1) electrophysiological tests (electroretinography) will determine reception of visual and infrared light wavelengths by the retina; and (2) discrimination trials will test perception of visual and infrared light wavelengths by the brain. Because the electrophysiological work involves anesthetization, which always carries with it a certain risk to the animals, the behavioral trials will be completed first. Discrimination Trials: Training & Trials: In order to accustom the pronghorn to the Y-maze, they were fed within this enclosure throughout bottle-feeding (Summer 1988). Fawns were led through the maze to the end of a branch, where they received the bottle. As the fawns began to take an interest in solid food, they were trained to eat marshmallows from the two feed boxes at the ends of the branches. Throughout all trials, a light-on cue will function as the SD+, light-off as the SD-. The two branches will be identical excepting the light source turned on in one branch, and off in the other. Subjects will be led into the start box with the start gate down. The buzzer, acting as a warning cue, will sound for 3 seconds. Ten seconds after the buzzer ceases, the gate will lift, and the subject will be allowed to make a choice. A subject's entire body must completely pass the choice line in order to constitute a choice. At this point, an experimenter observing the trial via video will lower the gate, retaining the animal in the chosen branch. The feed box associated with the SD+ will contain 10 miniature marshmallows in a metal tray available to the subject in the corresponding branch. The SD--associated branch will also contain 10 marshmallows, but this tray will be covered by a screen, preventing the subject in this branch from receiving the reward. A lip of the food box keeps the trays, screens, and marshmallows hidden from the subject until a choice has been made. Following a "correct" (SD+) choice, the subject will be released from the maze immediately after the choice(Le., crossing the choice line), but will be allowed to re-enter the branch to eat the reward. In the event of an "incorrect" (SD-) choice, the subject will be held in the maze 60 seconds; the extra time will act as a punishment. Light stimuli will be extinguished when the subject begins to feed. The pronghorn will be subjected to a correction procedure during trials. If a subject makes an incorrect choice, that subject will immediately repeat the same trial until a correct choice is made. Only the original incorrect choice will be considered a trial result. If a subject does not choose the SD+ after 4 attempts, then the subject will be led through the maze to the correct branch and allowed to take the reward. Subjects will undergo one or more trials daily. Assignments of stimuli to branches will be based on a random numbers table and will be done independently for each animal. A branch may be associated with the SD+ no more than three consecutive trials for any animal. Correct vs. incorrect trials will be assessed by the experimenter recording from the video monitor at the time of the trials. Number of crosses of each "branch" line before decision, as well as time from gate raising to choice will be assessed by the experimenter at a later date from video recordings of the trials. The criterion for a branch line cross is the same as for a choice; the entire body must cross the line. In order to avoid experimenter bias, an observer naive to the correct branch will periodically check the experimenter's results on all aspects of data collection. �84 Trials will begin with an SD + of white light. After this discrimination has been established, a narrow bandpass filter with central wavelength of 500nm will be placed in front of each light source. As before, only one light source will be turned on for each trial. This 500nm vs. no light discrimination will be followed by 550nm vs. no light, 600nm vs. no light, 650nm vs. no light, and finally by infrared vs. no light discriminations. If pronghorn are capable of seeing in the infrared range, they will be able to make this last discrimination. If the discrimination is made, then narrow bandpass filters in the infrared range will be used to determine which wavelengths in the infrared spectrum pronghorn are able to perceive. Over 20 trials, a performance level of 75% correct significantly differs from a level of 50% correct expected by chance (p = 0.05, Bruning & Kintz, 1987). Each subject will be required to demonstrate this 75% level correct over the subject's last 20 trials before advancing to the next wavelength. Data Analysis: Attainment of the 75% level correct criterion will demonstrate a subject's ability to discriminate a given wavelength. Electroretinography: Electroretinography (ERG) generates a graph of retinal potential over time (Riggs, 1966b). Light of a known wavelength and intensity is directed at the subject's eye and the elicited potential is measured. In this study, ERG will determine reception by the pronghorn eye of different wavelengths of light under both scotopic(dark-adapted) and photopic(light-adapted) conditions. This procedure will produce data for a graph of stimulus intensity required to produce a retinal potential of given amplitude versus wavelength of stimulus. This graph will show spectral sensitivity of the pronghorn retina. ERG data will also be analyzed for the presence of the scotopic and photopic waves in order to establish the presence or absence of rods and cones in the pronghorn retina. Electroretinography will be performed under the direction of Steve Roberts, D.V.M. and Mike Miller, D.V.M. The pronghorn will be transported to the Colorado State University Veterinary Teaching Hospital, via horse trailer, where they will be anesthetized before undergoing the procedure. A detailed protocol for transportation and handling of the animals will be established beforehand using three castrated pronghorn. E.SCHEDULE November 1988-January December 1988-December 1989 1989 Train castrated pronghorn to enter horse trailer. Discrimination trials. January 1989 Establish ERG procedure. January 1990 ERG on 8 pronghorn. February March 1990 1990-May Data analysis. 1990 Thesis. May 1990 Completion of coursework. August 1990 Graduation. �85 F. LITERATURE CITED Backhaus, V.D.(1959a). Experimentelle Untersuchungen uber die Sehscharfe und das Farbsehen einiger Huftiere. Z. Tierpsychologie, 16:445-467. ____ ~.(1959b). Giraffe(Giraffa 16 :468-477. Experimentelle Prufung des Farbsehvermogens einer Massaicamelopardalis tippelskirchi Matschie 1898). Z. Tierpsychologie, Beale, D.M., & A.D. Smith(1970). Forage use, water consumption and productivity of pronghorn antelope in western Utah. J. Wildl. Manage., 34(3):570-582. Boyle, S.A.(1981). Summer habitat of pronghorn in the Red Desert of Wyoming. M.S., Dept. of Fishery & Wildlife Bio., Colorado State University. Bruning, J.L. & B.L. Kintz(1987). Computational Foresman, & Co., Glenview, IL. Handbook of Statistics, third edition. Scott, Buchenauer, V.D. & B. Fritsh(1980). Zum Farbsehvermogen von Hausziegen(Capra Z. Tierpsychologie, 53:225-230. hircus L.). Everitt, J.H ..• D.E. Escobar, C.H. Blazquez, M.A. Hussey, & P.R. Nixon(1986). Evaluation of the mid-infrared(1.45-2.0 m) with a black-and-white infrared video camera. Photogram. Eng. & Rem. Sens., 52:1655-1660. _________ , M.A. Alaniz, & M.R. Davis(1987). Using airborne middleinfrared(1.45-2.0 m) video imagery for distinguishing plant species and soil conditions. Rem. Sens. of Environ, 22:423-428. Harosi, F.I. & E.F. MacNichol, Jr.(1974). Visual pigments of goldfish cones, spectral properties and dichroism. J. of Gen. Physiology, 63:279-304. Hunt, Jr., E.R., B.N. Rock, & P.S. Nobel(1987). Measurement of leaf relative water content by infrared reflectance. Rem. Sens. of Environ, 22:429-435. Knipling, E.B.(1970). Physical & physiological basis for the reflectance of visible and nearinfrared radiation from vegetation. Rem Sens of Environ, 1(3) :155-159. Krueger, K.(1986). Feeding relationships among bison, pronghorn, and prairie dogs:an experimental analysis. Ecology, 67(3):760-770. Loew, E.R. & J.N. Lythgoe(1978). The ecology of cone pigments in teleost fishes. Vision Res., 18:715-722. Riggs, L.A.(1966a). Light as a stimulus for vision, pp. 1-38. In Vision & Visual Perception, C.H. Graham(ed.). John Wiley & Sons, Inc., New York. Riggs, L.A.(1966b). Electrophsiology of vision, pp. 81-131. In Vision & Visual Perception, C.H. Graham(ed.). John Wiley & Sons, Inc., New York. �86 Schwartz, C.G., J.G. Nagy, & R.W. Rice(1977). Pronghorn dietary quality relative to forage availability and other ruminants in Colorado. J. Wildl. Manage., 41 (2):161-168. �87 Colorado Division Wildlife Research July 1989 of Wildlife Report JOB PROGRESS REPORT State of Colorado Project No. -=W_-~1~5~3_-~R~-~2 _ Mammals Research Work Plan No. __~l~A~ _ Multispecies Investigations Consulting Services for Mark-Recapture Analysis Job No. 5 Period Covered: Author: July 1, 1988 - June 30, 1989 G. C. White Personnel: L. H. Carpenter, T. D. 1. Beck R. B. Gill, R. M. Bartmann, G. D. Bear, ABSTRACT Progress toward the objectives of this job includes: 1. A manuscript summarizing the results of the Piceance Basin deer population studies has been developed, with publication intended as a Wildlife Monograph of The Wildlife Society. 2. A study of compensatory effects of harvest on the Piceance Basin mule deer population has been designed, with the first experimental harvest in December, 1989. Radio collars will be placed on fawns during November, 1989 to monitor over-winter survival of the animals. 3. A Monte Carlo simulation study to evaluate mark-resighting methods to estimate mountain sheep numbers is ongoing, with a final report expected by December, 1989. 4. Consultation has been provided in the design and analysis of an experimental isotope mark-recapture study to estimate river otter popUlation size. 5. A manuscript has been published in the July, 1989 issue of Journal of Wildlife Management on an evaluation of line transect methods for estimating mule deer density. 6. A manuscript will appear in the October, 1989 issue of Journal Wildlife Management on an evaluation of capture-resighting methods for estimating population size of elk. 7. A model of a Colorado black bear population has been developed and used to simulate various experimental harvest strategies. of ��89 CONSULTING SERVICES FOR MARK-RECAPTURE ANALYSES G. C. White P. N. OBJECTIVES Model and simulate population estimates of deer, elk, mountain mountain goats with mark-recapture methods. sheep, and SEGMENT OBJECTIVES 1. Evaluate various experimental harvest strategies for a Colorado bear population using a Monte Carlo simulation model. black RESULTS AND DISCUSSION A Monte Carlo computer simulation model has been developed of a Colorado black bear population using the Statistical Analysis System (SAS 1985). It is a stochastic model patterned after the grizzly bear model of Knight and Eberhardt (1985). The structure of the model is appropriate for Colorado black bears, with parameter estimates taken from the work in southwestern Colorado of T. D. I. Beck. The stochasticity in the model is of the form of binomial processes for births and deaths. Thus, for a 90% survival rate, the model does not multiply the current population size by 0.9. Rather, th~ N animals in the population are each assigned a fate (lived or died) according to the probability 0.9. The RANBIN function of SAS is used to perform this process. The structure of the model is based on 3 arrays. The array subadlt{O:l} holds the number of cubs and yearlings of both sexes without mothers. The array males{2:20} holds the number of males of ages 2 to 20 (the maximum age attainable). The array fem{2:20, O:6} holds the number of females of ages 2 to 20, and also their reproductive status. Category o means no cubs or yearlings, 1 means 1 cub, 2 means 2 cubs, 3 means 3 cubs, 4 means 1 yearling, 5 means 2 yearlings, and 6 means 3 yearlings. Although the yearlings are not actually with the mother, the mother is not allowed to breed in 2 consecutive years (unless the cubs are lost). Thus categories 4-6 provide a mechanism to store females for one year until they are ready to breed again. Two major extensions were made to the grizzly bear model developed by Knight and Eberhardt (1985). First, Beck (Pers. Comm.) has shown that survival of cubs and yearlings and reproductive rates are affected by the berry crop. Late spring frosts curtail mast and berry production. Therefore, 2 sets of values for these parameters were included in the model depending on whether the current year was a "good" berry crop or a "bad" berry crop. Bad berry crops occurred at random with a frequency of 1 year out of 10. Second, the model included various harvest strategies. A major consideration of the various experimental harvests was the proportion of males versus females in the harvest. Because of the differential survival of males and females, a straight percentage of the population �90 would have about 78% females. Therefore, all the experimental harvests were set up to take 10% of the adult population, but with differential vulnerability of females and males. A harvest vulnerability factor was developed to change the proportion of females depending on the time of year. For example, an early spring harvest might have only 30% females in the harvest. To obtain this level of harvest, only about a 3.3% harvest of the females would be needed. A female vulnerability factor of 0.25 was used to obtain the 3.3% (=[O.25/{1-0.25}]*0.1) harvest of adult females and a 30% (=[{1-0.25)/0.25]*0.1) harvest of males. The vulnerability factor was juggled to obtain values that generated appropriate sex ratios in the harvest. The cubs of any females harvested in a spring season (April-June) are all assumed to die, whereas cubs of females harvested in the fall (September-October) are assumed to survive at the rate of cubs without mothers. In addition, an increase in vulnerability to harvest of the entire population was included in the model. On average, 1 year out of 10 had a double success rate (20% harvest instead of 10%). This doubling of the harvest was included in the model to represent the effects of variable weather on harvest observed by Beck (Pers. Comm.) The following parameter values were used in the simulations. Survival Rates Age Class Females Good and Bad Berry Crops Adults, > 4 years old 0.71 0.52 4-year olds 0.52 3-year olds 0.52 2-year olds Yearlings (with mother) 0.90 0.90 Yearlings (without mother) 0.92 0.90 0.90 0.90 0.90 0.90 Good Berry Crops Cubs (with mother) Cubs (without mother) 0.57 0.57 0.57 0.57 Bad Berry Crops Cubs (with mother) Cubs (without mother) 0.33 0.33 0.33 0.33 Reproductive Parameter Parameters Definition (Barren means cubs born more than 1 year ago) Prop. of barren 2 year olds having cubs Prop. of barren 3 year olds having cubs Prop. of barren 4 year olds having cubs Prop. of barren 5 year olds having cubs Prop. of barren 6+ year olds having cubs Good Berry Years 0.00 0.25 0.33 0.50 0.80 Bad Berry Years 0.00 0.00 0.00 0.00 0.20 �91 Prop. of litters with 1 cub Prop. of litters with 2 cubs Prop. of litters with 3 cubs 0.19 0.59 0.22 0.19 0.59 0.22 Sex ratio at time of 2nd birthday 0.50 0.50 In addition to the above effects from bad berry years, all cubs born during bad years had reproductive rates as if it were a bad year until they were 6 years old. After a bad berry year, all barren adults had 0.90 reproduction instead of 0.80, a compensation for the previous bad year (unless the second year was also a bad berry year). The initial population sizes were constructed by making 5 runs of the model for 30 years with no harvest, and using the resulting population means for the various age and sex classes. The exact values for each of the age and sex classes are initialized the same in each simulation. A summary of these initial conditions are: Age and Sex Class Cubs (both sexes) Yearlings (both sexes) Adult Males Adult Females Total Population Size 197 93 80 296 666 Thus, these initial conditions repr~sent a fairly large population, approximately 5-10% of the state-wide population. This size of bear population is equivalent to the population of a major black bear Data Analysis Unit (DAU). Six harvest regimes were simulated: (1) no harvest; (2) an early spring season, with the expected harvest consisting of 30% females; (3) a later (or longer) spring season with the expected harvest consisting of 35% females; (4) a fall season with the expected harvest consisting of 45% females, (5) a fall harvest with the expected harvest consisting of 40% females, and (6) a fall harvest with the expected harvest consisting of 35% females. The difference between an early and late spring season is just the value of the selection factor, and similarly for the 3 fall seasons. The difference between the spring and fall seasons is the value of the selection factor and also that cubs survival is not zero for cubs of females that are harvested. All simulations were performed for 30 years, i.e., the population was projected forward for 30 years. This process was replicated 100 times to provide a mean and standard deviation for each of the parameters measured on the simulated population. Figure 1 provides a summary of the annual rate of increase for the simulations with constant survival and reproduction, berry crop failures, heavy harvest, and both berry crop failures and heavy harvest. Only the fall season with 45% females in the harvest is included. The effect of berry crop failures on the model population was to increase the variability of population size after 30 years and to �92 Expected Annual Population Increase (%) 5 4 3 2 1 o None D Early Spring Late Spring Fall Harvest System Constant Survival and Berry Cropp Reproduction Failures Heavy Both Berry Crop Harvest Fail. and Heavy Hrv. Figure 1. Annual rate of increase (percent) for simulations with constant parameters, berry crop failures, double the normal harvest rate, and a combination of berry crop failures and double the normal harvest. slightly change the expected population size. The impact of good and bad berry crops seems to affect the populations harvested in the spring more than those harvested in the fall. As an example, the late spring harvest under no berry failures had a 1.2% increase, whereas with berry failures, the population growth rate was 0.7%. These 2 values are significantly different (f < 0.05). The final population for no berry failures was 961 (SD 110, SE 11). The final population for the berry crop simulations was 855 (SD 206, SE 21). The population growth rates are very close to significantly different (f < 0.05) for the early spring harvest, but are not different for the fall harvest simulations. The impact of a 10% chance of double harvest significantly (f < 0.05) reduces the rate of population growth compared to where harvest is always the same from year to year. Another impact of the occasional double harvest is a slight increase in the variability of the final populations. For example, the standard deviations for the constant harvest simulations are 106, 110 and 105, but 124, 114, and 100 for the occasional double harvest simulations for early spring, late spring, and fall seasons, respectively. Part of the reason that the fall SD is smaller is that the population was much lower, so that the coefficient of variation would be quite a bit larger for the occasional double harvest simulations. Figure 2 provides a summary of the expected population size after 30 years for 4 of the harvest scenarios. One of the striking differences �93 noted in Fig. 2 is that the annual rate of population growth can be quite different (e.g., 1.2% versus 0.7%), and yet the differences in expected population size after 30 years are not all that great. This phenomena is caused by the linear scale on which the rate of increase is measured, whereas the 30 year population size is on an exponential scale. Another result evident in Fig. 2 is the greater variability of the 100 simulations for cases where berry crop failures occur with probability 0.1. The minimum and maximum values show a greater spread for both the cases where berry crop failures occur. Population 'L 000 3,500 3,000 2,500 2,000 1.500 1. 000 500 after 30 Years l1aximum l1ean l1inimum o Rate of Increase I I "---11~~-d-lmmmm-mm~mm-mm~I~-~J ; N ELF N ELF CDIlStut Beny B:eI."fl' !'allures Bl.nest Figure 2. Population size after 30 years. simulations. N ELF The minimum N ELF Both L r- (%) Icme Barl,. Bpr1q' Lata Bpr1q' Fall and maximum values are obtained from 100 The occasional double harvest can be quite important in limiting population growth. The current harvest levels are close to the maximum that the population can stand, particularly for the fall harvest. Berry crop failures will complicate this management process. The system is more variable, so would be harder to manage because the harvest (and population) would be more variable. Results of simulations of the 6 harvest scenarios with both berry crop failures and double harvest included in the model are presented in Table 1. The slight effect of no cub survival of females harvested in the spring is evident by comparing the late spring season with 35% females in the harvest to the fall season with 35% females. Because cub suvival is low to start with (0.51), and because the probability that the �94 mother will be harvested is low «5%), the effect of total mortality cubs of females harvested in the spring is not a particularly large effect. of Table 1. Results from the black bear simulation model for the case double harvest 10% of the time, plus berry crop failures 10% of the years. No compensation in cub survival as a function of population size was modeled. Values presented are the means of 100 30-year simulations. Hunting Type Season Annual Rate of Pop. Increase (%) Final Pop. Size Females in Harv. (%) Final Harvest Size No harvest 4.5 2480 0.0 0 Early spring l.l 932 30.0 46.03 Late spring 0.6 813 35.0 42.60 Fall 0.3 721 45.0 37.94 Fall 0.6 804 40.0 39.60 Fall 0.9 875 35.0 42.80 LITERATURE CITED Knight, R. R., and L. L. Eberhardt. 1985. Population Yellowstone grizzly bears. Ecology 66:323-334. dynamics of SAS Institute Inc. 1985. SAS Language Guide for Personal Computers, Version 6 Edition. SAS Institute Inc., Cary, NC. 429pp. Prepared by ./~ Gary c. C~ Whlte �95 Colorado Division Wildlife Research July 1989 of Wildlife Report JOB PROGRESS REPORT State of Project Colorado No. ~W_-~1~5~3~-R~-~3 Work Plan Mammals Research 2A Mountain Job No. 4 Experiments to Identify and Manage Stress in Mountain Sheep Period Covered: July 1, 1988 - June 30, 1989 Authors: Personnel: No. _ M. W. Miller Sheep Investigations and N. T. Hobbs M. A. Wild, K. A. Trust, H. J. Lucking, K. M·. Sheehan, Sousa, T. M. Nett, K. W. Mills, and E. S. Williams M. C. ABSTRACT Repositol adrenocorticotropic hormone (ACTH) injected every 48 hrs elevated urinary (P ~ 0.006) and fecal (P ~ 0.014) cortisol excretion in tame bighorns 0-24 hrs after treatment. Urinary cortisol excretion dropped below (P ~ 0.007) control levels >24-48 hrs after ACTH injection, but fecal cortisol excretion by treatment and control bighorns did not differ (P - 0.06) during that time. Our experimental data suggest that sampling excreta from bighorns offers a promising tool for detecting responses to environmental stressors. We reported results of our stress experiments in a doctoral thesis and two manuscripts submitted to peerreviewed scientific journals. Nonhemolytic Pasteurel1~ haemolytica biotype T was recovered .in nasal swabs from tame bighorn ewes prior to lambing, from l-week-old bighorn lambs, and from ewes and lambs during a pneumonia epizootic .in July 1988 that affected 7 of 8 lambs and 1 of 10 ewes; a similar shedding and transmission pattern appears to be developing again in 1989. Epidemiological findings from pasteurellosis studies in our captive bighorn herd offer insights into patterns of lamb mortality following pneumonia outbreaks in wild bighorns. ��97 PLANE OF NUTRITION AND BIGHORN POPULATION PERFORMANCE SHEEP M. W. Miller N. T. Hobbs P. N. Objective To treat bighorn sheep to control disease where necessary. SEGMENT OBJECTIVES 1. Assess the utility of urine and fecal samples to indicate environmental stress in mountain sheep populations. 2. Publish results in peer-reviewed USE OF FECAL CORTISOL RESPONSES scientific journals. TO DETECT SIMULATED CHRONIC STRESS Applying techniques for measuring cortisol in feces that we developed and described earlier (Miller and Hobbs 1988, Miller 1988), we detected marked changes in fecal cortisol excretion in bighorns treated with ACTH; details of the design and preliminary results of that experiment are reported elsewhere (Miller and Hobbs 1987, Miller 1988). Acute responses to alternate-day ACTH treatment were characterized by more than a 4-fold increase (P < 0.006) in mean urine cortisol:creatinine ratios (UCCRs) for the period from 0-24 hr after treatment (Fig. la). Fecal cortisol excretion followed a similar pattern of lesser magnitude; FCCs increased (P < 0.014) by more than 50% in ACTH-treated bighorns (Fig. lb). We detected no evidence for cumulative effects of chronic ACTH treatment on acute urine or fecal cortisol responses (time*treatment P > 0.05). Cortisol excretion 25-48 hr after ACTH treatment was lower than we anticipated. We observed no progressive increase in cortisol excretion between successive treatments (time*treatment P > 0.05). Instead, 25-48 hrs after they received ACTH, mean UCCRs were substantially lower (P < 0.007) in treated bighorns as compared to controls (Fig. la). The trend toward reduced FCes in treated sheep approached significance (P - 0.063) (Fig. lb). Our results suggest that cortisol excreted in feces, as in urine, may be used to detect adrenal responses in bighorn sheep. �98 EXPERIMENTS TOWARD DETECTING AND MANAGING STRESS IN BIGHORNS Results of the stress experiments in bighorns to date were summarized in a doctoral dissertation and two manuscripts submitted to peer-reviewed journals. The abstracts of these follow: Miller, M. W. 1988. Experiments toward detecting and managing stress in Rocky Mountain bighorn sheep (Ovis canadensis canadensis). Ph. D. Thesis. Colorado State University, Ft. Collins. 106 pp. Abstract: Endocrine responses of Rocky Mountain bighorn sheep (Ovis canadensis canadensis) to environmental stressors are believed to predispose them to disease. Methods for detecting adrenal responses in bighorn sheep were evaluated in a series of experiments. Average maximum serum cortisol concentrations (SCC) (ng/ml) from tame bighorns responding to physical restraint (hobbling) or exogenous adrenocorticotropic hormone (ACTH) (20 U iv) did not differ (P = 0.11). Bighorns responding to restraint with high maximum SCCs responded to ACTH with SCCs of similar magnitude (r2 = 0.84, P < 0.05). In a ser-cate study, plasma cortisol concentrations (PCCs) (ng/ml) and ur Lr;., cortisol: creatinine ratios (UCCRs) (ng /mg) from tame bighorns rose in response to repositol ACTH treatment. PCCs and UCCRs were correlated (r2 - 0.72, P < 0.001) after adjusting for lag in urinary cortisol excretion. Alternate-day repositol ACTH administration was used to simulate chronic stress responses in tame bighorns. Adrenal responses to ACTH were reflected by elevated mean 24hr UCCRs (P < 0.002) and fecal cortisol concentrations (FCCs) (ng/g dm) (P < 0.014) from bighorns on treatment days. Mean 24hr UCCRs from days between treatments were lower (P < 0.007) in ACTH-treated bighorns than in controls, and mean 24hr FCCs showed similar but weaker (P - 0.06) trends. Treated bighorns showed depressed lymphocyte blastogenic responses to lipopolysaccharide (P - 0.02), but not concanavalin A or phytohemagglutinin, by day 41 of each treatment period. However, no cumulative pattern was evident in elevated urinary or fecal cortisol excretion accompanying these changes. Urine-stained snow samples provided reliable estimates of UCCRs from voided bighorn urine (r2 - 0.87, P < 0.001). Tame bighorns exposed to a novel environment showed elevated cortisol excretion in both urine and feces. UCCRs and FCCs from 10-day composite samples were higher (P <0.05) in nonhabituated bighorns than in sheep previously habituated to metabolic cage confinement. Most (13 of 16) 24hr UCCRs from 4 tame bighorns that developed pneumonia equaled or exceeded upper 95% confidence limits for mean 24hr UCCRs from 8 clinically normal bighorns sampled over the same time interval. Elevated UCCRs preceded clinical signs of pneumonia by 5 12 days. �99 These results support further development of urine and fecal cortisol measures as noninvasive techniques for detecting adrenal responses in bighorn sheep, and perhaps other species. Both approaches offer potential use in testing hypotheses related to stress and the pneumonia complex of bighorn sheep. They may have broader application in studies of stress responses in a variety of other species. Miller, M. W., N. T. Hobbs, and M. C. Sousa. 19 Sampling adrenal responses in bighorn sheep (avis canadensis): reliability of cortisol concentrations in urine and feces. Can. J. Zool., in review. Abstract: We evaluated the reliability of cortisol concentrations in urine and feces to indicate responses to stress in captive Rocky Mountain bighorn sheep (avis canadensis canadensis). Tests for parallelism and quantitative recovery of cortisol in bighorn plasma, urine, and fecal supernatant revealed that radioimmunoassay reliably measures cortisol in these media. In a randomized complete block experiment, we observed cortisol responses of 8 bighorn sheep to 2 levels of adrenocorticotrophic hormone (ACTH) injections (0.0 and 0.50 U/kg) delivered at 48 hr intervals over a 29-day period. Treatment elevated (P < 0.01) cortisol in urine and fecal supernatant during 0-24 hrs following ACTH injections, but levels dropped below those of controls during 24-48 hrs posttreatment. Magnitude of ACTH effects did not change with time during the experiment (time X treatment P > 0.05). In a separate experiment, we observed short-term (0, 2, 4, 8, 12, 24 hr) cortisol responses in plasma, urine, and feces from 8 bighorns following a single challenge with 1 of 4 levels of ACTH (0.0, 0.25, 0.5, 1.0 U/kg). We failed to observe a dose response above the 0.5 U/kg level in any media. Concentrations of cortisol (ng/ml) in plasma were correlated (r2 - 0.76) with cortisol:creatinine ratios (ng/mg) in urine. We conclude that sampling excreta of bighorn sheep offers a promising alternative to blood sampling as a means of detecting responses to environmental stressors. Miller, M. W., N. T. Hobbs, and E. S. Williams. 19 Spontaneous pasteurellosis in captive Rocky Mountain bighorn (avis canadensis canadensis): clinical, laboratory, and epidemiological observations. J. Wildl. Dis., in review. sheep Abstract: We observed clinical signs, compared adrenal responses, and performed diagnostic tests on captive Rocky Mountain bighorn sheep (avis canadensis canadensis) during a spontaneous outbreak of pasteurellosis. Urine cortisol levels were measured for 12 bighorns sampled three times between 20 October and 1 November 1986. By 6 November, four of these developed pneumonia, four �100 showed only mild rhinitis, and four remained clinically normal; a similar pattern emerged in five unsampled sheep. Bighorns that ultimately developed pneumonia showed higher urinary cortisol excretion (P - 0.0034) over the l2-day sampling period. Elevated urine cortisol:creatinine ratios (UCCRs) preceded clinical pneumonia in affected bighorns by ~ 16 days. Mean UCCRs from pneumonic bighorns were 1.5- to 2 times those of mildly affected or unaffected sheep (P < 0.05) in two of three 48-hr. collection periods, but mean UCCRs from mildly affected and unaffected bighorns were indistinguishable (P > 0.05). Pasteurella haemolytica biotype T (nonhemolytic variant) was isolated from nasal and pharyngeal swabs from all bighorns with pneumonia, and from three of four sheep with mild rhinitis. We detected no evidence of parainfluenza 3, bovine respiratory syncytial virus, or Chlamydia psittaci using fluorescent antibody and/or serologic tests. Susceptibility to pneumonia in this outbreak appeared related to lack of previous exposure to pasteurellosis. In light of these data, we offer a refined hypothesis to explain epizootiology of the bighorn pneumonia complex. EPIDEMIOLOGY PNEUMONIA OUTBREAKS IN CAPTIVE BIGHORN LAMBS Poor lamb production and survival often follow pneumonia epizootics in bighorn herds. This carryover effect on recruitment typically prevents immediate recovery of bighorn numbers after an all-age die-off, and can lead to formation of small remnant populations. Lamb mortality usually results from pneumonia, but epidemiology for this facet of the pneumonia complex in bighorns remains undescribed. During the summers of 1988 and 1989, we've had the opportunity to gain some insights into pneumonia outbreaks in bighorn lambs. Methods and Materials: We hypothesized that bighorn ewes carrying Pasteurella haemolytica shed bacteria in nasal secretions during the periparturient period. As a result, their lambs are infected with Pasteurella early in life, and may succumb to pneumonia later in the summer. Using our tame bighorn herd, we examined epidemiology and diagnosis of pasteurellosis in bighorn lambs. During late April-August 1988, we simulated formation of a nursery band as described for wild bighorns, and monitored the spread of P. haemolytica through this ewe/lamb herd (n-l8); rams (n-6) maintained separately served to some extent as a control group. We held 6 tame ewes in isolation pens (about 50 mZ) from late April until 14 days after respective lambing dates, and allowed 2 wild ewes to lamb in a separate pasture. A single nursery group was constructed by adding ewe/lamb pairs as lambs reach 14 days of age. We added 2 barren ewes (ages 1 and 10 yr) to the nursery group about 1 week after the last pair entered. The resulting nursery band occupied 2 interconnected 4-ha pastures. We provided water, alfalfa hay, and a pelleted ration ad libitum in a single location in the pen complex, and natural forage was also available. �lOl We collected nasal swabs from tame ewes (n~6) prior to parturition, and from tame ewe/lamb pairs weekly until they joined the nursery band. Thereafter, ewes and lambs were sampled about every 2 weeks. We transported swabs in modified Amies medium with charcoal, and submitted them within 12 hrs for bacterial culture. Bacterial growth was screened for presence of Pasteurella spp. We observed ewes and lambs daily for signs of respiratory disease, and weighed each animal weekly. As lambs developed pneumonia, we removed them from the study and initiated antibiotic therapy (amoxicillin and gentamicin for 5 days followed by long-acting oxytetracycline for 7 - 10 days). We continued to monitor bacterial shedding in nasal swabs from tame bighorns by sampling in October 1988 and March 1989 in conjunction with other handling procedures. From mid-April 1989 to present, we began monitoring prevalence and transmission of P. haemolytica between isolated and semi-isolated bighorn ewe-lamb pairs. We maintained 8 tame ewes in isolation pens (about 50 Water, alfalfa hay, and a pelleted ration m2) beginning in mid-April. were provided ad libitum. We isolated 4 ewes in pens spaced about 5 m apart with individual water sources. Four others ewes were paired and held in 2 sets of adjacent pens; each set shared a common water source, and pens were separated by a pole fence that allowed nose-to-nose contact. We have weighed ewes and collected nasal swabs and serum weekly. Lambs were weighed, bled, and nasal swabs collected within 12 hr of birth; we sampled lambs weekly thereafter on the same schedule used for ewes. We are observing all bighorns daily for signs of respiratory disease. Beginning 17 July, we will immobilize lambs weekly with xylazine (about 3-4 mg IV) and collected tonsillar swabs and tracheal aspirates in addition to blood and nasal swabs. Ewes will be immobilized once in late July using ketamine and xylazine (about 91 mg and 9 mg IV), and we will collected tonsillar swabs and biopsies in addition to nasal swabs. We will screen bacterial growth for presence of Pasteurella spp., and submit isolates to the Wyoming State Veterinary Laboratory for biotyping and serotyping. Lambs that develop pneumonia will be removed from the study and treated with antibiotics as described previously. Results and Discussion: In 1988, 2 of 6 tame ewes (C83, T82) shed Pasteurella spp. before and after lambing; one of these (T82) shed intermittently throughout the summer. Aside from serous to mucopurulent nasal discharge, no clinical signs of respiratory disease were seen in these ewes. Clinical signs of pneumonia first appeared on 1 July in a 6-week-old ram lamb (C88). By 9 August, 7 of 8 lambs and a yearling ewe (L87) had developed pneumonia. In all, 8 of 8 lambs and 4 of 10 tame ewes shed Pasteurella spp. in nasal secretions during the outbreak. All lambs and the yearling affected by pneumonia recovered uneventfully after antibiotic therapy. The only mortality resulted from euthanasia of the index case in order to perform a complete diagnostic workup. �l02 The most common Pasteurella spp. isolate in 1988 was nonhemolytic, did not ferment indole, and did not produce catalase; this culturaljbiochemical profile is characteristic of the "bighorn strain" of P. haemolytica biotype T (Onderka and Wishart 1988, Onderka et al. 1988). Another Pasteurella isolate, similar to the first but catalase positive, was recovered occasionally. Pasteurella multocida (nonhemolytic, indole positive) grew from lung of the index lamb. We detected no serologic titers to Parainfluenza 3 (PI3) or Bovine Respiratory Syncytial (BRSV) viruses; no lungworm (Protostrongylus spp.) larvae were shed in feces from lambs. We concluded that Pasteurella spp. caused this outbreak, and that parasitism and viruses were not involved. We isolated Pasteurella spp. from 5 of 17 nasal swabs collected from tame sheep on 28 October 1988, more than 2 months after the summer outbreak subsided. Four of 7 lambs (including the lamb that did not develop clinical pneumonia) and 1 of 8 ewes were shedding bacteria at that time. In March 1989, we again isolated non-hemolytic P. haemolytica from 2 of 20 nasal swabs collected from our tame bighorns. Two of 7 lambs (including the lamb that did not develop clinical pneumonia) were still shedding bacteria at that time; none of the 13 swabs from adult sheep (rams or ewes) were positive. These isolates displayed metabolic profiles resembling P. haemolytica biotype A, in contrast to most previous isolates. Findings from 1989 are still preliminary, and no cases of pneumonia had developed by 1 July. However, we recovered Pasteurella spp. in nasal swabs from 3 of 8 ewes prior to lambing, and from 3 of 8 lambs (not necessarily related to shedding ewes) sampled at 1 week of age. As in 1988, prevalence of Pasteurella spp. recovered from nasal swabs was relatively low during June. Isolates recovered to date have been predominantly non-hemolytic and catalase negative, resembling the biotype T strain present in 1988. Early attempts to serotype recent isolates suggest nonspecific cross-reactivity with both type 3 and 4 antisera. Several features of these investigations bear emphasis. Bighorn ewes can carry and shed Pasteurella spp. without developing pneumonia; at least 1 lamb (the daughter of a carrier ewe) shed without clinical signs, as well. Furthermore, we detected preclinical shedding in lambs up to 1 wk before signs were noted. Both of these conditions shed some light on how Pasteurella spp. persist in and spread through bighorn herds. No consistent pattern for initiation of these outbreaks has emerged, although the first cases of pneumonia developed in early July -- that timing may coincide with loss of colostral antibody protection against Pasteurella spp. Once started, lamb-lamb transmission seems relatively important in sustaining epizootics. Most lambs affected in 1988 would probably have died in the wild; those not succumbing to pneumonia would have been quite vulnerable to predation, accidents, or exposure. Field accounts suggest that lambs "disappear" in mid-summer following dieoffs, and our findings provide a plausible explanation for these observations. Over two summers, the only adult ewe that developed pneumonia was a yearling that had not been exposed to previous outbreaks in our herd. We suggest that persisting immunity to Pasteurella spp. may protect adult bighorn from pneumonia, and that "all-age" dieoffs may result from a �103 rapid accumulation of immunologically naive sheep in a population where some carriers have survived. It follows that preventing pneumonia in bighorns may rely more on controlling growth rates of herds rather than total numbers. The role of stress in these outbreaks remains unclear; initiation of shedding by carrier animals seems to be a logical alternative to invoking some "population-wide" response to environmental adversity as the inciting mechanism. If carriers are less "fit" or "vigorous" than other individuals, they might be forced into less desirable social and environmental niches. The cumulative stresses associated with these suboptimal conditions could then lead to immunosuppression and, subsequently, induce Pas~eurella spp. shedding. In the presence of enough susceptible individual (adults or lambs), a pneumonia outbreak may ensue. LITERATURE CITED Miller, M. W. 1988. Experiments toward detecting and managing stress in Rocky Mountain bighorn sheep (Ovis canadensis canadensis). Ph. D. Thesis. Colorado State University, Ft. Collins. 106 pp. Miller, M. W., and N. T. plane of nutrition Colo. Div. Wildl., 26-P, WP2, J4, Job Hobbs. 1987. Mountain sheep investigations: and bighorn sheep population performance. Wildl. Res. Rpt., Part 2. Fed. Aid Proj. FWProg. Rpt., Ft. Collins, pp. 213-224. Miller, M. W., and N. T. plane of nutrition Colo. Div. Wildl., l53-R-2, WP2A, J4, Hobbs. 1988. Mountain sheep investigations: and bighorn sheep population performance. Wildl. Res. Rpt., Part 2. Fed. Aid Proj. WJob Prog. Rpt., Ft. Collins, pp. 99-112. Onderka, D. K., and W. D. Wishart. 1988. Experimental contact transmission of Pasteurella haemolytica from clinically normal domestic sheep causing pneumonia in Rocky Mountain bighorn sheep. J. Wildl. Dis. 24: 663-667. Onderka, D. K., S. A. Rawluk, and W. D. Wishart. 1988. Susceptibility of Rocky Mountain bighorn sheep and domestic sheep to pneumonia induced by bighorn and domestic livestock strains of Pasteurella haemolytica. Can. J. Vet. Res. 52: 439-444. Michael W. Miller Wildlife Researcher �104 • A 40, B 10g. 8· 7· ......J 6- (j) 5, o tco o ......J (5 UJ LL 12345678 12345678 f- ACTH -1 t- CONTROL ~ TIME = 0 - 24 hrs iZZl 25 - 48 hrs 12345678 12345678 '-. ACTH ---: '_CONTROL - TIME ;:::;::::J 0-24 hrs ~ 25-48 hrs Figure 1. Responses of tame bighorns (n = 8) to alternate-day ACTH (treatment) or saline (control) injections. We initiated treatments at the beginning of sampling period 4. Sampling periods were 6 days long and included 2 days of urine and fecal collections. Urine and fecal collections were divided into samples representing 0-24 and >24-48 hrs after injection. Treatment elevated cortisol excretion in both (A) urine (P ~ 0.006) and (B) feces (P ~ 0.014) in the first 24 hrs after ACTH injection. �105 Colorado Division Wildlife Research July 1989 of Wildlife Report JOB FINAL REPORT State of Project Colorado No. W-1S3-R-3 Work Plan No. -=2~A~ Job. No. 5 Period Covered: Author: Mammals _ Research Mountain Sheep Investigations Tests of the Mark-Recapture Method to Estimate Mountain Sheep Numbers July 1, 1988 - June 30, 1989 D. F. Reed ABSTRACT Fixed-wing flights were made on 10 November and 16 December 1988 to locate and determine the status of the 28 radio-collared female mountain sheep prior to the 15 replicate helicopter counts. During the flights and a ground telemetry check on 5 January 1989, 3 collars (Black E, Black F, and Black triangle) were found transmitting mortality signals. They were later confirmed to be mortalities. This reduced the total number of telemetry marks in the study population to 25. The lS helicopter counts were conducted as planned from 13 January through 23 February 1989. Preliminary analyses indicate variable resighting probabilities for the 25 marks. Overall, 57.5% (range = 35.7 85.7) of the marks that were in the study area were resighted during 14 of the 15 counts. Fifteen of the 25 marks were always in the study area, while 10 were out of the study area for varying occasions (n - 1-10). The mean number of animals counted arid marks resighted were' 58.5 (± 14.1 [SD]) and 11.5 (± 3.6) (ranges - 30-87 and 6-10), r~spective1y. Movements of the te1emetered animals before, during, and after the 15 'counts varied between individuals. ��107 TESTS OF THE MARK-RECAPTURE METHOD TO ESTIMATE MOUNTAIN SHEEP NUMBERS Dale F. Reed P. N. OBJECTIVE Test the fundamental assumption mountain sheep numbers. of a mark-recapture procedure in estimating SEGMENT OBJECTIVES 1. Complete 15 helicopter count replications. 2. Quantify the extent to which mark-recapture in mountain sheep population estimates. 3. Test the robustness data. of those assumptions model assumptions via simulations are violated with actual field ACKNOWLEDGMENTS I thank co-principal investigators R. B. Gill and G. C. White, graduate assistant A. K. Neal, and J. H. alterman. B. Metcalf piloted the Soloy helicopter, and R. Broomfield provided helicopter support. Saguache District Ranger J. Krugman provided the USFS Upper Saguache Guard Station for a housing and work center. DESCRIPTION Description of area was presented OF AREA in Reed (1988). METHODS AND MATERIALS Fixed-wing flights were made on 10 November and 16 December 1988 to locate and determine the status of the 28 radio-collared mountain sheep (Table 1). Following these counts and observations and telemetry from the ground, an estimate was made as to the area to be covered during the censuses or counts (Fig. 1). The first of the 15 counts was to be the flight path and area determinate i.e. each of the following flights would replicate the first flight path and covered to the greatest extent possible. The same helicopter and pilots used for each flight. The observers occupied the same positions in the helicopter (observer 1 betweenpilot and outboard observer (2]) during each of the flights. Both observers used hand-held taperecorders. Additionally, observer 1 carried binoculars for collar identification if needed and observer 2 a 35mm camera with 100-300mm zoom lens for recording animal groups for later classification, collar identification, and/or animal behavioral response to helicopter. �108 Each observer voice-recorded the number of animals in the group, number and identification of collars, and location. The aircraft intercom system and gestures (pointing) were used to alert pilot and observers to animals when they were first sighted. To reduce or eliminate error, observers conferred before or during voice-recordings. Observers agreed on whether needed information had been obtained before the pilot was instructed to "break-off" from the group and continue on the flight path. Upon completion of each count, the observers transcribed information from the recordings to field forms and marked locations onto a "master" 1:64,000 quad map. Then an additional flight was made using telemetry with 2 directional-finding antennas (1 fixed position and 1 hand-held) and observer 1 to verify the location ("in" or "out" of study) and status (alive or dead) of all marks that had not been sighted during the count in order to satisfy the "closed population" assumption. Fixed-wing flights were made on 23 May and 6 June 89 to locate and determine status of the 25 radio-collared animals after the counts and winter. RESULTS AND DISCUSSION Two fixed-wing flights were completed before the 15 helicopter counts were initiated on 13 January 1989 (Table 2). Three mortalities occurred before the counts were begun (Table 3). All 15 counts were conducted in the mornings. They were begun between 0814-1002 and completed between 1001-1141 DST. The verification flights (for determining "in" or "out" and mortality status) were begun between 1137-1309 and completed between 1259-1404 DST. Weather conditions were generally mild with a few occasions having light winds and one (6th count) having a low cloud ceiling and snow toward the end of the count. Snow cover varied from near 0 to 100% newly fallen snow. Most analyses will be done during the next segment. However, preliminary analyses indicate variable resighting probabilities for the marked animals en = 25) in this study. Overall, 57.5% (range - 35.7 - 85.7) of the marks that were in the study area were resighted during 14 of the 15 counts (Table 4). Fifteen of the 25 marks were always in the study area, while 10 were out of the study area for varying occasions (n - 1 - 10). The mean number of animals counted and marks resighted were 58.5 (± 14.1 [SD]) and 11.5 (± 3.6) (ranges 30 - 87 and 6 - 10), respectively (Table 5). Locations of the telemeterd animals before, during, and after the 15 helicopter counts suggest variable movements and sizes and areas of winter and summer home ranges (Figs. 2 - 27). A question that arose during the counts was, "Did given animals move after being counted as a result of the aerial procedures?". LITERATURE CITED Reed, D. F. 1988. sheep numbers. Prepared ~E-~ Dale Tests of the mark-recapture method to estimate mountain Colo. Div. Wi1dl. Game Res. Rep. July, Part 2:113-137. F. Reed Wildlife Researcher �109 Table 1. Number assigned to animal, collar description, channel, pulse, activation date, and status for telemetered mountain sheep Trickle Mountain study. Animal no. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 Collar Channel description Blue white 01 Blue orange 02 Blue square 03 Black A 04 Black E 05 Black F 06 Black H 07 Black J 08 Black K 09 Black M 10 Black NI 11 Black P 12 13 Black R 14 Black S Black T 15 Black U 16 Black ~ 17 18 Black X 19 Black Y Black Z 00 Black NII 11 Black triangle 09c Red dot 10 Red 1 11 Red 2 12 Red 3 13 Red 4 14 Red 5 15 Red 6 16 Red 7 17 aNo. - estimated Frequency (MHz) 172.787 172.812 172.862 172.912 172.962 173.012 173.062 173.112 173.162 173.237 173.262 172.837 172.887 172.937 172.987 173.037 173.087 173.137 173.187 173.212 173.237 172.487 172.512 172.537 172.562 172.587 172.612 172.637 172.662 172.687 age, A - active Pulse per min. (ppm) 60-110b 60-110 60-110 60-110 60-110 60 -110 60 -110 60-110 60 -110 60 -110 60-110 60-110 60 -110 60 -110 60-110 60-110 60 -110 60-110 60-110 60-110 60-110 60 60 60 60 60 60 60 60 60 (collar frequency, in the Activati0~ date 1 Feb 1 Feb 1 Feb 1 Feb 1 Feb 1 Feb 1 Feb 1 Feb 1 Feb 1 Feb 1 Feb 1 Feb 1 Feb 1 Feb 1 Feb 1 Feb 1 Feb 1 Feb 1 Feb 17 Mar 3 Apr 17 Mar 23 Mar 23 Mar 23 Mar 23 Mar 3 Apr 3 Apr 3 Apr 3 Apr transmitting), 88 88 88 88 88 88 88 88 88 88 88 88 88 88 88 88 88 88 88 88 88 88 88 88 88 88 88 88 88 88 Status 3A 4A 4A SA 4D 4D 3A 4A SA 3D 4D 6-12A 7A SA 6A 3A SA SA 4A SA 7A 3D 4A 6A 4A 3A 7A 2U 2A SA and D - animal died. bMorta1ity sensors: 60 ppm - alive, 110 ppm - dead (no movement period of 4 hrs). Sheep 22-30 do not have mortality sensors. cSecond series of numbers is for a second Te10nics TR 1 receiver. for �110 Table 2. Flight date, aircraft type, number of mountain sheep observed. number of mountain sheep with transmitters observed, and number of transmitters located and not located. Flight date 17 23 3 5 7 22 10 16 13 14 17 18 19 27 28 1 2 8 9 15 16 22 23 16 6 Mar Mar Apr May Jun Jun Nov Dec Jan Jan Jan Jan Jan Jan Jan Feb Feb Feb Feb Feb Feb Feb Feb May Jun Aircraft 88 88 88 88 88 88 88 88 89 89 89 89 89 89 89 89 89 89 89 89 89 89 89 89 89 Table 3. stud Soloy Helicopter Soloy Helicopter Soloy Helicopter Cessna 185 Cessna 185 Cessna 185 Cessna 185 Cessna 185 Soloy Helicopter Soloy Helicopter Soloy Helicopter Soloy Helicopter Soloy Helicopter Soloy Helicopter Soloy Helicopter Soloy Helicopter Soloy Helicopter Soloy Helicopter Soloy Helicopter Soloy Helicopter Soloy Helicopter Soloy Helicopter Soloy Helicopter Cessna 185 Cessna 185 Mortality Observed wi transmitters t::lI1e Observed 39 41 46 5 20-30 39 5 79 68 68 62 66 57 52 61 87 48 64 30 69 35 52 57 28 16 of collared Transmitters located 14 12 13 1 1 5 1 8 7 11 11 14 11 10 9 19 8 13 8 14 6 13 17 11 0 female mountain Transmitters _E_ot located 4 16 27 26 26 24 23 24 24 23 24 24 25 24 23 25 25 24 25 25 21 22 24 12 1 0 0 1 2 1 1 2 1 1 0 1 2 0 0 1 0 0 4 3 sheep in the Trickle Mountain Date No. 10 11 5 6 22 Collar Black Black Black Black Black M N1 E F triangle Age 3 4 4 4 3 Marked 1 1 1 1 17 Feb Feb Feb Feb Mar 88 88 88 88 88 RecoveredL:Death < 5 Feb 88 < 23 Mar 88 < 10 Nov 88 < 10 Nov 88 16 Dec 88-5 Jan 89 Estimated cause Capture myopathy Capture myopathy Illegal kill Unknown Unknown �ill Table 4. Frequency, Frequency Cell Chi -Square Percent Row Pct Col Pct chi-square, and resighting I I I I Inot sigh1sighted Ited I I I Frequency Cell Chi-Square Percent Row Pct Col Pct Total ---------------+--------+--------+ bear 2 12 2.6168 1.9307 0.70 4.21 14.29 85.71 1.65 7.32 - - - - - - - - - - - - - - -+- - - - - - - -+- - - - - - --+ besq 4 10 0.6357 0.469 1.40 3.51 28.57 71.43 3.31 6.10 14 14 1 0,9399 0.35 16.67 0.83 5 0.6935 1.75 83.33 3.05 4 10 0.6357 0.469 1.40 3.51 28.57 71.43 3.31 6.10 - - - - - - - - - - - - - - -+- - - - - - - -+- - - - - - --+ bkfe 5 9 0.1499 0.1106 1.75 3.16 35.71 64.29 4.13 5.49 ---------------+--------+--------+ bkh animal. I I I I Inot sigh-Isighted Ited I 3 3 0.0804 1.05 50.00 2.48 0.0593 1.05 50.00 1.83 7 0.3972 2.46 53.85 5.79 - - - - - - - - - - - - - - -+- - - - - - - -+6 2.11 14 4.91 14 4.91 I I Tota. 2.1 ---------------+--------+--------+ bkj 4.91 ---- - ---- - - --- -+- ----- --+- ---- ---+ bka ~2T~~C ---------------+--------+--------+ 4.91 ---- ----- - ---- -+- -------+- ---- ---+ bewh percent per 7 0.1877 2.46 50.00 5.79 - - - - - - - - - - - - - - -+- - - - - - - -+bkn 9 1.5714 3.16 64.29 7.44 - - - - - - - - - - - - - - -+- - - - - - - -+bkp 9 1.5714 3.16 64.29 7.44 - - - - - - - - - - - - - - -+- - - - - - - -+- bkk 6 0.2931 2.11 46.15 3.66 - - - - - --+ 7 0.1385 2.46 50.00 4.27 - - - - - --+ 5 1.1594 1.75 35.71 3.05 - - - - - --+ 5 1.1594 1.75 35.71 3.05 - - - - - --+ 1. 4.Si V4.91 14 4.91 14 4.91 �112 Frequency , Ce11 Chi -Square' Percent , Row Pct , Col Pct 'not sigfrlsighted I I ted I I Total 6 8 0.0005 I 0.0004 2.11 I 2.81 42.86' 57.14 4.96 I 4.88 14 1 0.2871 0.35 25.00 0.83 I I I I 3 0.2118 1.05 75.00 1.83 ---------------+--------+--------+ redl I 2 4 I 0.1176 I 0.0868 I 0.70 I 1.40 I 33.33 I 66.67 I 1.65 I 2.11, 50.00 I 4.96 I 2.11 50.00 3.66 I - - - - - - - - - - - - - - -+- - - - - - - -+- - - - - - --+ I 12 14 I 2.6168 I 1.9307 I 0.70 I 4.21 I 14.29 I 85.71 4.91 I 4 bks 1.40 2 1.65 I 7.32 4 0.3556 1.40 57.14 3.31 3 7 I 0.2624 I 1.05 I 42.86 I 1.83 2.46 ---------------+--------+--------+ 6 bkt 2.11 3 0.0463 1.05 37.50 2.48 5 0.0341 1.75 62.50 3.05 8 2.81 ---------------+--------+--------+ 4 bku 1.40 6 8 0.0005 I 0.0004 2.11 I 2.81 42.86 I 57.14 4.96 I 4.88 14 4.91 ---------------+--------+--------+ 12 I 0.1609 , 0.1187 I Total ---------------+--------+--------+ 2.44 ---------------+--------+--------+ red2 I 1 3 I 0.2871 I 0.2118 I 0.35 I 1.05 I 25.00 I 75.00 I 0.83 I 1.83 ---------------+--------+--------+ red3 I 6 6 , bkr 4.91 ---------------+--------+--------+ bkz I I ---------------+--------+--------+ ---------------+--------+--------+ bky Frequency I Cell Chi-Square I Percent I Row Pct I Col Pct Inot sigh-I sighted Ited I 4.21 bkx 5 9 0.1499 0.1106 1.75 3.16 35.71 64.29 4.13 5.49 - - - - - - - - - - - - - - -+- - - - - - - -+- - - - - - --+ 14 4.91 �1.13 Frequency Cell Chi-Square I Percent I Row Pct I Col Pct [rio t; sigh-jsighted I Ited I I Total ---------------+--------+--------+ red4 8 0.7113 2.81 57.14 6.61 6 0.5248 2.11 42.86 3.66 14 4.91 ---------------+--------+--------+ redS 8 0.7113 2.81 57.14 6.61 6 0.5248 2.11 42.86 3.66 14 4.91 ---------------+--------+~-------+ red6 6 1.2425 2.11 66.67 4.96 3 0.9168 1.05 33.33 1.83 9 3.16 ---------------+--------+--------+ red7 7 0.1877 2.46 50.00 5.79 7 0.1385 2.46 50.00 4.27 14 4.91 ---------------+--------+--------+ redo 5 0.1499 1.75 35.71 4.13 9 0.1106 3.16 64.29 5.49 14 4.91 --- - ---- - ------+- - -- ----+- ----- --+ Total 121 42.46 164 57.54 285 100.00 �Table s. Number of mountain sheep counted, marks (radio collars) out of the study area, and classification of the animals counted. No. marks Date 13 14 17 18 19 27 28 1 2 8 9 15 16 22 23 Counted Jan Jan Jan Jan Jan Jan Jan Feb Feb Feb Feb Feb Feb Feb Feb x 68 68 62 66 57 52 61 87 48 64 30 69 35 53 57 58.5 Sighted sighted, marks Classification Out Adult female .£ Young female 8 11 11 14 11 10 9 19 8 13 8 14 6 13 17 7 11 4 4 5 5 5 3 7 2 7 5 8 5 4 13 32 30 26 26 25 27 39 24 31 19 33 20 26 32 5 11 15 14 12 12 15 25 12 16 6 17 9 14 18 2 4 5 4 6 3 5 8 2 5 1 3 3 7 4 ll.s 5.5 26.9 13.4 4.1 Young male Adult femrl1e 2 0 0 0 1 1 0 0 0 2 1 0 0 13 13 12 15 11 11 14 10 9 9 3 12 3 6 3 33 8 0 7 1 0 0 5 9.6 4.0 0 O· 0.5 Unknown 1 1 0 4 0 0 0 �4235-.--------Spanish Creek Rd ~ -'3 -.···l· ..".\ E ~ 4231-1 ~ o North,\ .. , -, Pass .s: \. _ -, . ~ s •.....• ..c 1:: 4227 g 'Il ~ ~ ?- Buffalo Pass Campground t- ::> Trickle ~ Mountain 4223 RaM,. VUf( Ca/7YO/7 421 9--1 358 362 366 370 a. ~O' ~ ~ Benny ~ :::0 Q'C}~ ~ ,.>.' .•... ~ ~C'f' ~, -Y('/ . . ,/ .' -"-", l~ •.... '\ ~~ .{) Houghland "'" ..••••.....•••.• - ..- ..~ . '\ \0 Hill Hi9hW(l· -._./ .Y 711--r-r---.-r--.---r---'-'-1---'---' 374 378 382 386 390 UTM east/west km Fig. 1. The areas covered (shaded) during the 15 helicopter counts. Each flight departed tile Guard Station, first searching the area west and southwest of the Guard Station, then the main Trickle Mountain area from the eastern most part near Lone Tree Gulch and proceedinq clockwise, then counterclockwise over the area between Lone Tree Gulch and Jack's Creek Road, then north, east, and south over the horseshoe shaped area south of Middle Creek, and then north and northeast of the Guard Station during the return. r-' r~ \J1 �4235~----~~~~~--------------------------------~ SpanIsh Creek Rd . /~( 4231 3.~.. . +.. , l \fJ«' North Pass E ~ \ '~' "iJ %; \ .-- ..-, t:: s \ 9- Campground 'il Benny ~ ~ ~ , ", Rabbit~~~. al]YOn , I I I I 362 I I I , 366 I , I 370 , •• •• •• Trickle ~, Guard Mountain, '.. •••••...~ .• - •••.•••.••~ I I I , 374 UTM east/west \ ",' .-:~.t \ " 'x _._ .~. _. -.- Stta" Ion : .'\ k .... fl'" .. ....." 1A" " "1 358 Q ~--.l ~ 4227 4219I ( 0' ", ~ s :i: ..__~ !~ i t-" t-' 378 km ~. ....-\. '<. HOIjghland Hill 'x /' J.ng/}Wa·-·- . I I 1:Y ,111 I •. : _;'P"\'. ~ I I I 382 I I I 386 I I I 390 * = • = Nov 10. 1988 • = June 7. 1988 June 22. 1988 ~ = Dec 16. 1988 Fig. 2. Black F (Channel 6,173.01 MHz) (---), Black triangle (Channel 22,172.49 MHz) (_._), and Black E (Channel 5,172.96 MHz) (- - -) locations were determined by telemetry during fixed-wing flights on 7 and 22 Jun 88, 10 Nov 88, and 16 Dec 88. Black E and F were trapped with the dropnet at "X" 1 reb 88 and were transmitting mortality siqnals by 10 Nov 88. Black trianqle was captured with the net gun from helicopter at "x" 17 Mar 88 and was transmitting a mortality signal by 5 Jan 89. I I �4235 Spanish Creek Rd 4231 ~'~~ E Pass' .!Ii::: s: ::; 4227 Ca s ,. <r">. ~ Buff'a 0 Pass Campground <:;) ~ t- :::l 4223 I / -.-=:-13',(\12,14 4 / / J..~. "'t' ~ / 0' .' . / . ••••••• ,.. ••_ C1!1,yOf) •• _ 'f- " / / '~1O . ~ 1 .~ ' ..•...0 '\. 2 . 7(.., ;)~~ .•.._. / Trickle'" Mountain ~, ~ 3 /' __ -- .••.. --~-- / ~ 9/11:tJ /~ Benny \ ...•- {\", . ~-:s2 \ o- Rllbb;t C -i 4219 358 -- /L}: Q('i) _ -. / "7{~/~ <p. ij) , ~ 0 5 ~OL -o9>. round , Cochetopa Pass ~._ ~ Ludsr's Crss< g ••.... s: t ~.. North ,.. ~ / 7'(/'" 6• '\' _,)- //r ~ ••~' -" Y " ' Hifih'. x-: /.....•..'~ WeY-1'-' ~.~ '........ _: • y • ..--. --, I Hougl1land ~ Hill ",. 386 390 . 14 I 362 366 370 374 378 382 UTM east/west km Fig. 3. Blue white (Channell, 172.79 MHz) locations before (_._) the counts (---), and after (- ). Locations before (17 Mar, after (16 May, 25 t~ay, and 6 Jun 89) were determined by observation wing aircraft. the 15 helicopter counts, during 22 Jun, 10 Nov, and 16 Dec 88) and or telemetry from the ground or fixed1-- ~~I �f-' f-' co 4235~------------------------------------------~ Spanish Creek Rd E ~ s: ~ ?: i 3.:~.. 4231 .. ~.. .-..~ l o ~? North """ . Pass s 4227 s \._ 0:-, j ~ ~ '.q,_ _ - -. ?~ ~-;g. \~ -,.:" ~ Benny ~ ..\",~'~ <. ::> '., 4223- ~ Rllbbit~ C.'lh -IYo'!) 4219 I 358 I I • 0 . ~ ~O' (S) ~~ ~~ Campground t- ~~ Trickle ~ ~ountain "'- ua,:£!~_~.._~~'. ~tlOn -.-.._ •....•... "_" ~ -: \--:- 370 374 378 Houghland ( <, H'II • \' \\ I I 382 1::- _". \ /( ':'X'8/ U:-;-....:.:36~ I1I9hwa·-· . Y 111 7,10 14'· I 366 /"_"_"\~' r 11,15 <. 4: ~~. 5 2 9· I 362 / .A Buffalo Pass 'Il za a. O'(}~ ~~ _ ~ .•.... ..c 1:: ~C}f-. ~. '-"-> I I I " I I 386 I' .---- 390 UTM east/west km Fiq. 4. Blue orange (Channel 2, 172.81 ~1Hz) locations before (_0_), during (- -) the15 he l i cop te r counts. Locations before (23 Mar 88, 2 Apr 88, and and 6 Jun 89) were determined by observation or telemetry from tile ground or location during the 1st count was undetermined. Locations during the 12th and general area southeast of Trickle Mountain and are not represented with dots (numbered dots), and after 5 Jan 89) and after (16 May fixed-winq aircraft. The 13th counts were in the or corresponding nUlllbers. �4235-.---------- ----------------, Spanish Creek Rd ·"1-"> ~.. ~~~oo E ..:.::: s: ~ 8 ....... ..c 1:: 4231-1 North \00'00 '40f. ~ Pass \'_'" ....• ~ ?-- . Buffalo Pass 4227 s Campground 'i) I- "' .... ,"\ Cochetopa Pass ::> ?f ~Q' ~ ~ ~ ~ Benny ~ <S' C}.~ ':b,.>.' .'~ ••••... );;lckla ~ M~·h\in I) ."Guard •••.••, station '.' RllhJ..,. I I I I I 362 I I I I 366 r I r I 370 r r r r 374 r r}' I "-",1~ . \ $/ .. ...--'- • __ ._. ,111 J78 r r -, Houghla~d ~ HIli ··- D:nh ..~12·:\·?X·8/ .15 b llli;p 1Wa 4219 I 358 • • 4 5 ,,;,... 3' lj~. 2. •.•• ;-._ "VIr Cal1YOfl "--~.~;;;- I 382 I I I I 386 r ""'" I r I j 390 UTM east/west kIn Fig. 5. Blue square (Channel 3,172.86 MHz) locations before (_._), during (--and numbered dots) and after (dots not numbered) the 15 helicopter counts. Locations before (23 Mar, 3 Apr, 22 Jun, 10 Nov, and 16 Dec 88) and after (15 and 26 May 89) were determined by observation from ground or helicopter, or by telemetry from fixed-wing aircraft. Locations during the 9th-11th and 14th counts were in the general area of southeast Trickle Mountain and are not represented with dots and corresponding numbers. I-" 1-" .o �r~ Ij o 4-235-,------------------------Spanish Creek Rd E ~ s: -'3 ···l· .."..~... j 4231 ~ ~ North \"'" ~ Pass \._. 1:: s • .•••••• Trickle ~ 4 2 19 I "" VIJIC I , , I _"_. Cal1Yon , I , I 362 366 370 I I I 374 UTM east/west I ri;:;:--. . """'6 I I . .. 13-' 1·2 •• - •••••• / 6 /1 ~ Rllhl,.,. 'J • __ station , - _\-</'{ \ "' ~/~/,)./ ..~A~ •.. , ,~~·T ~/'>C·.......•.. ~.,., C::I Guard Mountain 4223 , \~ • .~Q' Benny :E .., ~ Campground 'Il () CD • v:~ ?- . Buffalo Pass 4227 358 Qf-.?a . ...•.• "5 ~ ••... s: ..'. {i ~ lOX ', /'/ .__:I.V Y l-fl~114 378 .~ .7/'1~/' I • "'_14,15 A ' Houghland \..:') HII ..••.•, . 386 390 I 382 km Fig. 6. Black A (Channel 4, 172.91 MHz) locations before (_._) the 15 helicopter counts, during the counts (---), and after (- - -). Locations before (7 and 22,Jun, 10 Nov, and 16 Dec 88) and after (16 May and 6 Jun 89) were by telemetry from fixed-wing aircraft. The location during the l l th count was in the general area of south Trickle Mountain and is not represented with a dot and corresponding number. �4235~----------~-----------------------------Spanish Creek Rd 4231 3_'''' .. · .., l ..!Ir::: =§ ~ 12 cL: "<, \ ~C}f-' 1:: s 2 :E • •••••• Trickle C) Mountain CClJ]YOf) 5' • .• _ ••_ ••:-;-......, ~ .{4 r II I II: ° \ RaM!.t / _. t II a.\y 1,,8 _""'{_ -_ _o 10 .,... ~\----rk/.(5~.........__~ ~ 5 "VOJIf ~-/ I I I 6"~\{ / I.:t( ~ Campground Benny •. / ~ '?Buffalo Pass ,"91415 13' 0' C} .._,-. s .•.... .c :?: o ~ North Pass E , , . • 1/ l\ X /' "-"_"\ . ~ ..._,;". -,. , .-\. Houghland ~ Hill "" Hi9hW8'-'-~ 421 9 I I 358 I I I I 362 I I I I 366 I I I I 370 I I I I 374 I I ,}' ,111'-r--r--'--·"---'-r---T 378 382 J86 I I I I 390 UTM east/west km Fig. 7. Black H (Channel 7,173.06 MHz) locations before (_._) the 15 helicopter counts, during the counts (---), and after (- - -). Locations before (7 and 22 Jun, 10 Nov, and 16 Dec 88) and after (16 May and 6 Jun 89) were determined by telemetry from fixed-wing aircraft. 1-' [ ..) t-·· �f-' '"'-.,j ,-' 4235~----------------------------------------~ SpanishCreek Rd 4231 3.''' .. · l ~ ~ .J::. .., ,..... 1:: i ~Q(. .. ..~...... ~.:--- 4227 Buffalo~~ •• · ~ CampgrolM~' ~ Benny • • ..•• . \. Q ~ :=! "'." -. \ ••••••••••••• ~ •• '1 t- ::> 4223 F/4 "- ~ bbit E ~ t l~ ~ ~,9 .... , · Trickle',," '\ "--- Guard ¥~mlain \. sr.liOi1 • ~~!. • s,\ .. -"-. ~(/" J 362 366 370 I 374 I J J ,111 378 .\" ~ f 15 any"" .•.•••.. - •• f.n~. 8-':~:· 'X6: / .,--.--_ , , , , , lVay--' 4219358 I , a. -5>0' G 'il 8 :n ~C}.~ o "''''' ~ ....•.. ..c: ~ o ..., North ." v. -.::: Pass •• -, . E s ~ 'r.:, \ ..~ I I Houghland W Hill ~ ", -, I 382 I I ,,----rI 386 I 390 UTM east/west km Fig. 8. Black J (Channel 8, 173.11 MHz) locations before (_._) the 15 helicopter counts, during the counts (---), and after (- - -). Locations before (23 ~1ar, 2 Apr, 7 and 22 Jun, 10 Nov, and 16 Dec 88) and after (16 May and 6 Jun 89) were determined by observation or telemetry from the helicopter, ground, or fixed-wing aircraft. The locations during the 2nd-4th and 12th-14th counts were in the general area south or southeast of Trickle Mountain and are not represented with dots or corresponding numbers. �4235---.-------Spanish Creek Rd /4231-0. ...- ..>.~ .._ -. I j E ..!:t.:: s: "5 g ••..•... ..c t g ;E I- ::l ~ '''\A' '" b" fT1P =. --..! Buffalo ).61ss" round ( Pass ~ ~ "v 0' Or. _'''-_ ,,-L -...~",~O' ~ \' <SI ~.~ ~ kl N-IC ~ M0J1l\altl l:'5~t'-.-.... Guard '," "., ,,/ / .._\_.., •• Sluuon ....___\ - -,.:1_ <1!)Yon n. -. ~ . ! <> / ~ ''\'.... "'- <,~ - ~ ,. c ' ,.A • . Rllbbit ~<l10 0 CampgrOllN~ ':;: ~, ~ Benny ,.,... 'Il 0 ( th -<l \ ..•...•.• '" ~ ~\ ~ North '\00'0 Pass \ _"'" • C7 "..___........J ~ ~ 0 \.. Luder's Cram 4227- \ ~ -~ .. _ .. _.. ~----,,- C) \--;. ;' ,/ ) .r /'/ 15 12. . '....... 113 8 6'~X' / / v,' •...•. ----~------- .' /' -... ( \~ \ r' I Houghland ~ Hill . <, Hin0 ...:_::. 14 / 4219 I I 358 ~/W8y , ,-.-. 362 , , , , 366 11-' ,-----.-.-.-,--.--'-,--'1-,-'1-1"--·-·1·-,-,---'--r-- 370 374 378 382 "'--'---'--"-1-,-'-- 386 390 UTM cast/west km Fig. 9. counts 88) and ground, south, numbers. Black K (Channel 9, 173.16 r~llz) l oca t ions before (-·-)the 15 he l icop tcr counts, during the (---), and after (- -). Locations be fo re (23 r1ar, 2l\pr, 7 and 22 Jun , 10 Nov, and 16 Dec after (16 and 26 May, and 6 Jun 89) were detennined hy observation or telemetry from the helicopter, or fixed-wing aircraft. The locations durinq the 2nd-5th and 9th-11th were in the general area southeast, or southwest of Trickle ttoun ta in and are not rcrresc>nted vii t.1I 'lots and corresponcling The 16 MilY, 26 ~1ay, and G Jun B9 l oce t i ons \t/(~re at the s arue si t c . ,_. ,. .) W �f-' Iv -I> 4235~--. Sp anish . Creek Rd 3.·r\ I E ~ .c t :J -o \ 2'~ COChetopac"t{ 11 ~t Beri //l 22 3-1 ::u a. ~\O'OI'{l~ ~ ••>. C) j ~Of' ~. ~iJI '.-. 4227 :'! o ~ " s I- ~ .., ~ North '" Pass '-.'. ~ .c ~ o 4231=i g ...•..• ?: <, -" 0>0' ~ ~ Buffalo Pass .~'% Campground \ ~ ~::s p" ny Pass o ~ " \ . __..-..: Trickle ~ Guard '\sta __..•.•..•. -e;. 00 ~\<--::t~.-.- '1~bbitCuf1Yon ----~ Mountain .•..•.... "-"-~-. • /ii,~ '.~ 15 .•.• "'--0. '. : .\) . /' ".- yI 1li2''':'''X »:' '\ Houghland \0 Hill . '" 8 Ohw~·-·-· 4219--,-r-.--r-,-.---J 358 362 r--r-;-r--r-r--y--,--,---,-,-,-,-,-.,-}',l~11---.-r-.--,----r--'--r-'--T-'--'----i 366 370 374 378 382 386 390 UTM east/west km Fig. 10. Black N (Channel 11, 173.26 MHz) locations before (_._) and during (numbered dots) the 15 helicopter counts. Locations before (27 May 88, 22 Jun 88, 10 Nov 88, and 5 Jan 89) were detennined by observation and telemetry from the ground and fixed-wing aircraft, respectively. Locations we re not determined during the 2nd-7th, 9th-10th, and 13th counts. �4235~----------------------------------------~ Spanish Creek Rd ~ 4231 . / ···l-·_)_''..+.., ~ o North '. Pass E ~ s: ~ :?:: ,. \ .•.•..... ~Q' Buffalo Pass Campground 'il ' Cochetopa Pass Trickle • VfJ1l: CC1!7YOI} 1 358 I I I C) Mountain Rllhk'-' 4219 I Y("/'-"-'\~ Benny ~ ~ I 362 I I I 1 366 x: a. ~C)..~ ..,.". .....• <-- s 4227.c t s i '4~. I 1 I 1 370 I ..,.. •• _.. •..•.. I . .' ~5 .!1 - .. ' 8 ~~ Houghla~d """'" ~ HIli ••~' "X"/ Hi,:-:-"""" 10 • Ohw(l·-·_· I I 374 I I . I }' ,111 378 I I 1 382 ,----o-r----r----r--r--r--.. 386 ,390 UTM east/west km Fig. 11. Black P (Channel 12, 172.84 MHz) locations before, during, and after the 15 helicopter counts. Locations before (5 May, 22 Jun, 10 Nov, and 16 Dec 88) and after (15 and 26 May 89) were determined by observation from the ground or telemetry from fixed-wing aircraft. Locations during the lst-3rd, 6th-7th, 9th, 12th-13th, and 14th-15th counts were in the general area of south or southeast Trickle Mountain and are not represented with dots and corresponding numbers. ,__. I J UI �t-' 10 0' 4235~-----------------------------------------~ Spanish Creek Rd 4231 ~ .c ~ ~ ••..... ..c 1:: .. /· · 3,~ l . J o 4227 '. \,~ '. ~Cf.i> s <) 'b., ~O' ~ ~ss~ ~ ....' '\ ~.~~...... 5 '" 4223 RaN.> ..,jf , 358 " I , I Gu~rd'~~J~ •••••••• , I I I 362 366 370 '. ii[j;--'i";3'x:'/"~ WaYl-' Houghland \0 Hill .._ .••_. - I 374 I .-"- 'Y , 4'112 stl!tiOll_ ilIIyO!} ~J(/"/ "\~ 1\ ~ • Trlcl$'''' \____ ;'- E ~ , '" x: o, q.~ ~ "').. Buffalo ':) Campgroul\d ._. ~ nny • Be , ,~, V~ ~ 4219 I i ~'" North " Pass '" E ~ •_'I' , .13: 5/" I I I 378 11- -r1~1- 382 ~\ \ I I -,--~I 386 I -, '. I I 390 UTM east/west km Fig. 12, Black R (Channel 13,172.89 MHZ) locations before (_._), during, and after (- - - ) the 15 helicopter counts, Locations before (23 Mar, 2 Apr, 7 and 22 Jun, 10 Nov, and 16 Dec 88) and after (16 May and 6 Jun 89) were determined by observation from helicopter or ground, or by telemetry from fixed-wing aircraft, Locations during the 1st, 9th, and 14th counts were in the general area of northnortheast, south, and south Trickle Mountain, respectively, and are not represented with dots and corresponding numbers, �4235~----------------------------------------~ Spanish Creek Rd o / · ... .. .., 4231 ···.1··>."" ~ s: '1.:J ~ North Pass 13 ~ v-,v-, ",. \ 5 C}f-.s.~\/'K~ .. .....•. "",- s 4227 t s 9 /....:~~3,4 <- ••.•..... / Q .t::. Buffalo Pass ~ Campground Benny %.;:s / / / , ;! '~~, ' -, 'rO" \,(\ <:»: • .)a., ' , I~ \' ~\ \ ~. _ ~4_- ~ •. 1,2 'il 11 ."","-' 5 C·/ .-~.- ..~ 1• uard Moun~4n .1JtQ1 station 10 G '.:"1' -, • ITlllu..,. CanYon 421 9358I , ' , , , , , , , ....• ~I I I "'IJfC I 362 366 I 370 ..- ,:0;;;--. ,: / ..~ i I Iii i 374 378 UTM east/west ). . X W8Y-l ._. I '1 iii 382 Houghla~d \0 HII "'. . i I I I 386 I iii I 390 km Fig. 13. Black S (Channel 14,172.94 MHz) locations before (_._), . .r i nq (----), and after ( ) the 15 helicopter counts. Locations before (7 and 22 Jun, 10 Nov, and 16 Dec 88) and after (16 May and 6 Jun 89) were determined by telemetry from fixed-wing aircraft. Locations during the 6th and 7th counts were in the general area northeast and southeast of Trickle Mountaili ud are not represented with dots and corresponding numbers. I- ~'-' ,.J �t-' I·.) OJ 4235--,------~ Spanish Creek Rd 4231~;31'_'..' , .. E ~ North P ass .!r::: .s: ~ 8 •....... ..c t 14 Ludsr's CreEk 4227 Cam::>around s '. '-. \ ,,-y' 13,." • I- Buffulo Puss 4223- Pass Guard station R~bh:4. Vir 4219-~ 358 CanYon ~ ~<, 6 r ,)( / ~ <, ~ • /'~\~ ~J$>O' ~:::s COChetopa( ::> / , 0~/ v ?-- \ •• th", ..~....•.... -," •• 9 •• , i1q~ , '\) Campground Benny ~ ,.. " / 8 \ /12 -.l1. 1,,57 . »<c: Trjckl~ ~ ..,:;;~1f.j~;/,/ -:-..:. ~ Mountain." ,..'~#._.\ . ~ ~_ .7///1 • .--~ -"- / t/ 2 ;;;:•...~ I \ •....... .(/~ .• ~ . X1oh/ lii9hwp' -._./ ~)t •. \ \\ Houghland ~ Hill 366 370 374 378 . 111- I 362 "" 382 ,- 386 390 UTM east/west km Fig. 14. Black T (Channel 15,172.99 t·1Hz) locations before (_._) the 15 helicopter counts, dur i nq the counts (---), and after (- - -). Locations before (7 and 22 Jun, 10 Nov, and 16 Dec 88) and after (16 t~ay 89) were determined by telemetry from fixed-wing aircraft. The location during the 11th count was in the general area of southeast Trickle Mountain and is not represented with a dot and corresponding number. �4235~----------------------------- / 3_.".. 4231 Spanish Creek Rd -I" E ~ :§ 8 •........ ..c t • ~ o i ".•.. " -t .••, North .. ,. Pass ~ . _ ~ ~" -x -s-, ~ v' ••_. .t:-." ••••, .••, ,..-..!. 4227 s , q" -, Buffo!~~ 0iJI."". ~" ass .'~"" . '" CamPll<. ,_ ~ r :J Benny ' ,' ,.'~..••. •'\ R. 4219~~~~~~~~~~~~ 358 362 366 370 )' // / / :l'''cklainc,/ / // 15, Mou!l)a / 374 ltii1>waj7i r I I 378 / _"-'" "', \ ,, r \:" / •.•.••x •••..•.. // /" .' r>, ,., ~.~ '~~::::~ abb;r CiJI)yon / ~ .', ~ "/ • • . Pass Q~ x: o, : _~," Houghland. / 11', , , '-'T:;-8 6 382 __/ . ., ,) 9 () '- UTM east/west km Fig. 15. Black U (Channel 16,173.04 MHz) locations before (---), during (---), and after (- - -) the 15 helicopter counts. Locations before (23 Mar, 2 Apr, 7 and 22 Jun, 10 Nov, and 16 Dec 88) were determined by observation or telemetry from helicopter, ground, or fixed-wing aircraft. Locations during the 1st, 5th-7th, 10th, and 12th counts were in the general area of south, southeast, and southwest Trickle Mountain and are not represented with dots and corresponding numbers. ,_. I·.) '-D �I-" W o 4235~------------'------------------------------~ Spanish Creek Rd 4231 E ..!a::: 3... ····· l . ~ Ludsr's g o '~"\ North ", P!Ie~" ~. .c :?: or R' .•..•• CrB~ \ -e s ~ b'.-"&' a"i~"""ss-- 'il ~ Camp13rp'tna: Benny ~ ~ ~ () •.•...••........•• ~0·.. .. ~ C arroaround .c ~ijI .....•. ".._"".. U'C}~ , ~ 7'.... '. ~ ..•.•. ~ '-: ·'~, \ 't>: Trickle ~ •~·i ~ountain "". • Guar•d...... •. 4219- _ .._. 12 • ~C_)l.!.9:~:='.15 __ -.j ".' • -"_" a. "'-y." ..(.. !/'...... '\~ •. ~ . 5 4' Rllbbit~' Ca17YOf) t» ~Q' ~ (SI ,........ !:i i ~q. <- 2 ro iX' /' k~._._. ~ •• •. ..', ~ - - - --'_' --r-H;ughland ~ Hill ---'- "'" '-'-~~'-'-~~-'-'-'-'-'-'-'-'-'II~.-rY-Tl~11-,,-·~~~~~~ 358 362 366 370 374 378 382 386 390 UTM east/west km Fig. 16. (- - -) and after aircraft. southeast Black female (Channel 17,173.09 MHz) locations before (_._), during (---), and after the 15 helicopter counts. Locations before (17 and 23 Mar, 7 and 22 Jun, 10 Nov, and 16 Dec 88) (16 May and 6 Jun 89) were determined by observation or telenletry from helicopter or fixed-wing Locations during the 1st, 6th-7th, 11th, and 13th-14th were in the general area of south and Trickle Mountain and are not represented by dots and corresponing numbers. �4235 \ Spanish Creek Rd ~ 4231 8 .••... .c 1:: -"-i··f _'."..".., North Pass E ..!&::: .s: ~ . ~ '. ~ ~ o ,.\ i '"\ '.-'. ..•.... \ 2'-Buffalo Pass 4227 s Campground 'il Benny ~ ~~~0f-~C} Q ~ ~ ~~ ~ ~ Q '\ . ~ ~ %~ \ .• ! t- ::> \ 4223 358 Trickle () 'i• •••••••••• ~.- •• -.. 366 370 374 UTM east/west Fig. 17. Black X (Channel the 15 helicopter counts. 6 Jun 89) were determined were in the general area numbers. NJ/ ~ i-» ~~ 6 / .... / I/// -·12/ / 'r_;,~;Z'!:~ »>' P:1_/~ ..•_.~. Hi9hW8. _._ }' .. • /8 ~7 /;1'\ x-...?: /_/ \\ ,91415 "Houghland \0 Hill .3/ 1 1 1--.--,-.-1-.-. 378 1,2 \;;... ~..".(~~ ~ ~<T "/-.tJj_'!I- .J" "~ y/ Y''''' Mounta!n.:.:": ~/ " 4· ~ -~-r--r--Y--Y--r--r---r---r--r--r---r--r----.---, 362 ~5._.. ~. //;. Guard t::l ~~ : '\ t r 1711bbit .••... ~.-/'~_!!.lOn Caf1YOf) 4219 x: a. 382 386 "\ .. ~". . "" r--,.. ---r---,-- 390 km 18,173.14 MHz) locations before (---), during (---), and after (- - -) Locations before (7 and 22 Jun, 10 Nov, and 16 Dec 88) and after (16 r>1ayand by telemetry from fixed-wing aircraft. Locations during the lOth and 11th counts southeast Trickle Mountain and are not represented by dots and corresponding ~w r~ �1-' W 1'0 4235~------------------------------Spanish Creek Rd 4231 E ~ ?: 3·\ "~ .. , .c ~ 8 4227 " t .c I North Pass or o \ ~ Campground Benny :E t- :> 4223 Trickle () Mountain Rdb&t~\ ~ . Guard ~tlOn 366 370 . I? .3. . r;J 8 .; ".! :.,,~ ..' ~ ·.••..··;:h.-··-·.~I?.2:_./ C""YOI) 362 ~~- ~'j .. .",.,. ..•........ Cochctopa( Pass : 4219358 c: '>t .f0' -"~ 2-BUffalo Pass /' <> 0. oJ q"{k, ~..x " -' :lJ ~, ..~. 'V i '4Qf-. '. ..•.... Luder's Crsek Carroaround s ~ \l Hi{J/}lVdy 114--- 374 378 1----~ <, ./ '\ .-'-. Houghland ~Hill _ '--'"38-6' "" _,---,_~~ j 90 382 UTM east/west krn Fiq. 18. Black Y (Channel 19, 173.19 MHz) locations before, during, and after the 15 helicopter counts. Locations before (3 Apr, 27 May, 7 and 22 Jun, 10 Nov, and 16 Dec 88) and after (15 May 89) are not differentiated from the helicopter count locations because of the tigllt cluster. Locations during the 2nd, 6-7th, 11-12th, and 14th counts and the fixed-wing flight of 6 Jun 89 were in the general area of soutll or southwest Trickle Mountain and are not represented with dots and corresponding numbers. �4235~-------------------------------------------~ Spanish Creek Rd ~ 4231 ·"~l··>_'\.~.... Pass 15 "-t ..c ~ . 0 ~ / ........• /9 cc. '-." ""'T 'lIJr -$'", TrickJe tl Mountain 4223 58 , ' ' !YO!) 1 j j I j I •••••• _ .,-..' liiih'. j j j j 362 366 I' 370 . 374 j I r 378 11 V ;,).0 • •. -/-.(.-(~ ". ~ / ~,<..,:.-" ' .15. X I " ~". lOV, ... \ /' J~ ""8Y-1'-' j I 3 .' . »«: /~""""6 13' .,,.-- Benny s 4'''' Q ..---- .....•. ~,,7 y.... , .<\ •.• Buffalo Pass ~ Campground :E 'y, ,..---81112 ,..f.3 2'-- 4227 s i' 5 North '\ .• E ..!II:: .c g . / Houghland • j j 382 \0 Hill I I r I 386 I ", _.~ I -, I I I 390 UTM east/west km Fig. 19. Black Z (Channel 00,173.21 MHz) locations before (_._), during (---), and after (- - _) the 15 helicopter counts. Locations before (7 and 22 Jun, 10 Nov, and 16 Dec 88) and after (16 and 25 May, and 6 Jun 89) were determined by observations from ground or telemetry from fixed-wing aircraft, t-" W '-'-' �r-' W -1-' 4235~-----------------------------------------~ Spanish Creek Rd 4231 E ~ 3"" .. ·, l ~ ?: a (j) o j .. North Pass ", \ ~ :0 ..""". 8 •....• ..c '?- Buffalo Pass Campground 1::: s Q.. 'il 13' Benny ~ (jp/ I- ::> Trickle (:) Mountain 362 366 370 374 UTM east/west 8/ :'//i ~ -"-"~ T •• { i~.•..... 15 Hi9hWa' - 421 9-358 ./~ ,. ~ 6.>f.·,?:14 tf'1 •••••.••• 1/ \,.' - ~ .' Rc')blw~~n Cal1YOI) i :(1/12 •• _. }' 378 .:»: . \ Houghland \0 Hill '-. "" 11-1-,--,-r-r---r--r-r-T-T-r-r- 382 386 390 krn Fig. 20. Red dot (Channel 10,172.51 MHz) locations before (_._) the 15 helicopter counts, during the counts (---), and after (- - - ). Locations before (22 Jun and 16 Dec 88) and after(25 May 89) were determined by telemetry or observations from fixed-wing aircraft or ground. Locations were not determined during the 2nd, 5th, and 10th counts. The location during the 11th count was in ttle general area of southeast Trickle Mountain and is not represented with a dot or corresponding number. �4235 Spanish Creek Rd 3 / · l ~~ .. , 4231 ~ North Pass E ..!x: s: A 15 g .•....•. .c -e .. l.uder's Craffi 4227 s ~ "<, \ ..__,-. i o """ ~CIL '1.' ~. '?_ Buffalo Pass ~ Campground Benny t- I --j \ ~.A' Ii>' . ~ 7,13 : ~ ~ ~ ~ ~ / Q'/ -:! ~ ~.;:s / / '.'. / '4 , ?' )", ...._~·/~~11~ ..-.-"'-:-.x'r.::. . .• , .. """ x._( /~ • " ~~. "/ ,." : ..•_ .. _.. -~-.-.-,,- .. /\) 11 / '\ __ l11-r--'-~ 4219- R'-~~~1I-:;;'7"""""-r:::T'-'-'--'-...-r~Hh~~n~~~. ~ 358 362 374 366 370 378 382 UJ1Wll';-'. 10 »:»: •. • 8,12 "~" . ~ / ~1·~· ~ Rabbit C al1yon ~ 119~ 0' C}: Trickl~ ~ Mountain ::> 4223 ~ J . L ~.~ Houghland ~ 386 Hill "'. . m UTM east/west km Fig. 21. Red 1 (Channel 11,172.54 MHz) locations before (_._), during (---), and after the 15 helicopter helicopter counts. Locations before (22 Jun, 10 Nov, and 16 Dec 88) and after (6 Jun 89) were determined by telemetry from fixed-wing aircraft. ~~ l,-J UI �t--' W Q\ 4235~-----------Spanish Creek Rd ~ (j) 09 .., . 4231 ~ ~~. 0' E ..:.:: s: ::J s 4227 s a 4 .,.~~ .•.•.. .c 1:: t- ::> Trickle Mountain /{ ,'{J_ 1 1 I'/ <::) RlJN,.", VUfC CanYon ~'" - , ~( staten ......... A_ .._ ..~ /' •... 1 7 '~ ~ 4219-1 358 II I .:..l4- 6 I__ ~~ 8~\ .'3 . 131. I ~ 5· -_ II IO'l' \ :~ \ .. \~ . ~ (\\ \, .._ _ •.- -" ~~ -...•..•....;.. ....." --.,....-------.;:....: ... 14 : .1 ~1,2 ...-~. Houghland ~ Hill "" Hi9hWa'-'-' y 111'-r--~ 362 366 370 374 UTM east/west 378 382 .,----,- 386 --.----. 390 km Fig. 22. Red 2 (Channel 12,172.56 MHz) locations before (_._), during (---), and after ( ) the 15 helicopter counts. Locations before (22 Jun, 10 Nov, and 16 Dec 88) and after (16 May and 6 Jun 89) were determined by telemetry from fixed-wing aircraft. �4235~-------------------------------------------~ Spanish Creek Rd ~ g: (j) 4231j ~'~" E .!I:: s: ~ ~ .•.•... .c. 1::: .. ~ North o , ~ "'" \ Pass 4227 g ~...0 ~\II "-' 2'-- . <) ~~ -r-~ ..•... Buffalo Pass Campground 'Il ~" Benny ~ ::> 5,9 35 8 '" 362 4 21 9 0- ' /3/\ \ ~O' 2 ~ 11 "Y On I \ ~,1 li \/' /.v'( I " •. 'l/·x \. Trickle ~tI, _I.I. /, M \ '" '-... Cdh I ')" _ .....• \ a. <J'q.~ -0> ,......... flabbit~ I , . ... :IJ :10, I- 4223 ,_ i ~Qf' <:~' 136 6' 1 I , ' I 370 ~6' :ouar(l"ounla~ station 7"', \ . ••••.•. 12 I I _' I 374 <, ~(" Hi(;;;;:.8Yl'_' 15 /' 1"\ I \';)~ V /'~ 378 11 382 I •._ .. " : \ I "_" I /'_ I ' 'I Houghland \0 Hill -~ -. -, r-,--------,-- 386 ,)90 UTM east/west km Fig. 23. Red 3 (Channel 13,172.59 MHz) locations before (_._), during (---), and after ( ) the 15 helicopter counts. Locations before (22 Jun, 10 Nov, and 16 Dec 88) and after (16 May and 6 Jun 89) were determined by telemetry from fixed-wing aircraft. .W -.J �I-' ,,) rx: 4235~------------------------------------------~ Spanish Creek Rd 4231 15 t \.. Pass ..c 8 3 ···",· l .. iD o North '\ .. ,. E ~ ........ ..c :?: [ \ x: Q. O'q.~ ".,- .......• t)lQ' 2"-Buffalo Pass 4227 s ~ '4Cf· \) Campground Benny ::::=! I- ::J 4223 =. TrickleC) Pass R~bM<, ~ uit allYOII 66 4219358 362 E ~ 3 • --- Guard Mountain .1';1 ~(// : '1station ~_ •• -.'- • •• ·A F HiOh'~:13.'8/· ~ I: ••_ ••.(4; "'5 ..~~X.J 370 IVaYI'-' 374 378 11 382_ ..\';)~ t, " A Houghland \..:)Hill -, -. ~I·--'---'386-'--' 390 UTM east/west km Fig. 24. Red 4 (Channel 14,172.61 MHz) locations before (_._), during, and after the 15 helicopter counts. Locations before (27 May, 22 Jun, 10 Nov, and 16 Dec 88) and after (15 May and 6 Jun 89) were determined from observation from the ground or telemetry from fixed-wing aircraft. Locations during the 1st, 3rd, 7th, 9th-12th, and 14th-15th counts were in the general area of south or southeast Trickle Mountain and are not represented by dots and corresponding numbers. �4235 Spanish Creek Rd 3 / · "I"~ .. 4231 .. ~ North 'v., E .!o:: s: ~ '-. Pass \ .....-. ...•.. s ..c 1:: s ~ ~ o <fI 'il ~ I- 4223 I ~ • " - ,~ Cl t(", -~./ "_~ountain station " -.11 CanYon -j ~ s·, "., . _ .. _.. d·:.X .. __ H'",~1.14 _ 4. 362 Trickle Guard \ Rltbb;t 4219 358 ~Q' .~ ~ Campgr~" Cochetopa Pass ::> 366 CI.. •... ~ &1tfaIQ fass Benny :::0 U'C}.~ 0 ij) . ?- ••••..... ~ ~~. ~ 370 - 374 UTM east/west IS : 8" / -"_" .. \") .c.>:' ·K. ' .-~ . Hougtlland "" ~ Hill '~JWLl,;-,_. ~J 11 378 1 '382'"386 y)Q~ km Fiq , 25. Red 5 (Channel 15, 172.64 r~HZ) locations before (_._) the 15 helicopter counts, during the counts (numbered black dots), and after (- - -). Locations before (10 Nov and 16 Dec 88) and after (6 Jun 89) were determined by telemetry from fixed-winq aircraft. Locations during the 2nd-7th, 9th-lath, and 13th counts were in the general area of south or southeast Trickle Mountain and are not represented with dots or corresponding numbers. The location on 6 Jun 89 was in Saguache Park about 19 km off the map to the southwest. ,_. w -0 �,_. +-' o 4235----r--------- .-----------------, .SpanishCreek Rd 4231~31'.'.'. ~ .J::. ~ 8 •....• .J::. t or o .., North \", Pass E ..:.: . -, ~ \ Cochetopac I- ::> U" _-' '.~~:-...r ,~. 'V Campgrouncr, nny ~ 358 1 I''. .-$>0' ~ ~ T .•k1-' ~~ .....• ~rJc ../ ..- ..- ..'\-:'l a () ( ~!1 ~ ountai1 • 'I -OUl!l!Q. - _ .8 Can 1 362 _"" ~ .'Yon -+-'-1 ~($) --.- ~- \ Rabbit 4-2 1 9 •• \\~~. ~~~, Pass 366 370 x: a. O'q.~ •• .- '§:.__ -->-.." ~ ~ 13 15\ Carrooround s :-..~ "-'.~~. 14 ~ ~('f' ..•~ ~r. Luder's CreEl< 4227 ~ g: . \\ ;tatk>n '~-..., . - .. :-;--....... • 1ii00W8.·l_ ._ . Y 111 374- 378 .' \ : './2 .-'- I .' \) / '\ 1 1 »> '..... Houghland ' . (,- HOII • "'-.-> I 382 I ,-r-l-r-r-T-r---,-- 386 390 UTM east/west km Fig. 26. Red 6 (Channel 16,172.66 MHz) locations before (_._), during (---), and after ( ) the 15 helicopter counts. Locations before (7 and 22 Jun, 10 Nov, and 16 Dec 88) and after (16 May and 6 Jun 89) were determined by telemetry from fixed-wing aircraft. Locations during the 1st, 3rd-4th, 6th, and 9th counts were in the general area south, southwest, and northwest of Trickle Mountain and are not represented with dots or corresponding numbers. �4235~------------Spanish Creek Rd 0l"~" 4231 ..._.f '. i ~C}f-. \ .:u ~.:.., ~ .•.•... .c Buffalo ~~S ~ Campgroun~ • Benny g ~ ~ / a. O'(}~ Q \.~~ s 4227 t: o ~ Pass 1.:i fl (j) .~ ~ North"". E ~ .c: s: -$»0' ,'"i. ~.~, '\ l--_______' X M~ain Guard· station .\ . 11 ._ .._ Rabbit ~ CanYon 5 •...• 2 12· . 14158 .~ \ . .~.~ • /' Houghland ~Hill --'- '<,-, Hi9hwa' - ._ . 4219 I , 358 J J J J 362 J J J , 366 J I J , 370 J J J J 374 J J J:.Y 1 ,11 378 J J , 382 J , 386 390 UTM east/west km Fig. 27. Red 7 (Channel 17, 172.69 MHz) locations before (--.--), during, and after the 15 helicopter counts. Locations before (22 Jun, 10 Nov, and 16 Dec 88, and 5 Jan 89) were determined by observation from the ground and telemetry from fixed-wing aircraft. Locations during the 1st, 3rd-4th, 6th-7th, 9th10th, and 13th counts were in the general area south or southeast Trickle Mountain and are not represented with dots and corresponding numbers. ,_. .... .1> ��l'4J~ Colorado Division of Wildlife Wildlife Research Report July 1989 JOB PROGRESS REPORT State: Colorado Project No.: W-12S-R-2 Work Plan No. :__~2~A~ Job No.: Author: 5 (supplement) Mammals Research _ Mountain Sheep Investigations Computer Simulations to Statistically Evaluate the MarkRecapture Method with Mountain Sheep A. K. Neal Personnel: A. K. Neal, G. C. White, D. F. Reed, R. B. Gill, J. Olterman ABSTRACT Monte Carlo simulations were used to evaluate the mark-resight method for estimating population sizes of mountain sheep. Specifically, the precision and bias of the estimates were assessed given violations of the assumptions of mark-resight. With sighting probabilities changing on each flight, the variability and the bias are greatly decreased. With heterogeneity in individual sighting probabilities, the confidence interval coverage of the estimates is too low. With decreasing sighting probabilities over time, coverage is much more variable. With groups of animals, the estimates are more variable, coverage is poorer, and overestimation is a problem. This Job Progress Report represents a preliminary analysis and is subject to change. For this reason, information presented herein MAY NOT BE PUBLISHED OR QUOTED without permission of the Director. ��1:;5 INTRODUCTION Mark-resight, a method based on mark-recapture (Caughley 1978) with a Lincoln-Petersen estimate (Petersen 1896, Lincoln 1930), is one procedure to estimate the size of mountain sheep (Ovis canadensis canadensis) populations. This method involves first capturing and marking the animals, and subsequently recapturing or resighting them. Identifying and counting marked and unmarked individuals then provide estimates of population size. Quantitatively, the basic estimator of population size (N) with one marking occasion and one recapture or resight occasion is N (equation 1) (m2) where n1 - the number of animals initially caught and marked, n2 - the number of animals caught or sighted in the second occasion, and m2 = the number of marked animals caught or sighted in the second occasion. However, this estimator becomes more biased as sample size decreases (White et al. 1982:18) and may be corrected when (nl + n2) > N by modifying equation 1 as follows (Chapman 1951, Seber 1982:60): (n1 + 1) (n2 + 1) -1. N (m2 + (equation 2) 1) Bias is defined as the difference between the estimated and the true population sizes. The associated estimated sampling variance is: (n1 + 1) (n2 + 1) (n1 - m2) (n2 - m2) var (N) (equation 3) (m2 + 1)2 (m2 + 2) (Seber 1982:60). When using this method, several critical assumptions must be satisfied to estimate abundance accurately (i.e. with good precision and without bias). The first assumption states that the population is closed demographically and geographically. No individual may enter through birth or immigration, or leave through death or emigration; given only one marking occasion the proportion of marked individuals must remain constant. This assumption is rarely met in a natural population, but its effects can be minimized (e.g. by shortening the time period between marking and recapturing). Some models have been developed incorporating deaths and migration (Seber 1982, Seber 1986). Kufeld et al. (1987), using radio-collars on mule deer, determined which proportion of marked animals were in the study area at any given occasion and thus maintained a "closed" population. A second assumption of the mark-recapture method states that all animals (marked and unmarked) have an equal and constant probability of being �146 captured and of being recaptured or resighted on subsequent occasions. Animals that are captured on a given occasion should not be more or less susceptible to being recaptured or resighted than others. In a study where animals are recaptured, if heterogeneity in capture probability occurs, the estimates can be significantly biased (Otis et al. 1978:11,25,30,34). In a study where animals were sighted twice using two observers in one plane, Pollock and Kendall (1987) reported heterogeneity in sighting probability led to negative bias in the estimates. In a study where animals were live-captured and marked and then "recaptured" by resighting from the air, White and Garrott (1989:275) reported that capture heterogeneity was avoided with one method of capture and a different method of "recapture", and that individual sighting heterogeneity did not bias estimates of population size but did underestimate the variance. One explanation for this heterogeneity in probabilities may be disproportionate marking so that marks are not randomly distributed among the animals but one segment of the population receives a greater proportion of marks (Rice and Harder 1977). Alternatively, the researcher may not be able to regulate heterogeneity of probabilities because trapping probabilities and reactions vary per individual animal. A third assumption states that samples are random. To meet this assumption, either the capture or the recapture effort must be randomized. This assumption is violated if one animal's probability of being sighted is not independent of another's (i.e. animals form more or less stable social associations). If marked animals are not randomly dispersed throughout the population and are accumulated in higher densities around the initial trapping areas, traditional variance calculations are not valid (Kufeld et al. 1987). The variance of the estimate increases with clumping (Rice and Harder 1977). Researchers have little control over this aspect if the animals being studied characteristically band together. If groups tend to be randomly distributed, then groups can be counted and mean group size (number of animals per group) calculated separately to reduce the bias and still get an estimate of abundance (Bergerud and Manuel 1969, Eberhardt et al. 1979, Seber 1986). A fourth assumption states that animals do not lose their marks. If they do, a greater proportion of unmarked individuals will be counted in the population than is actually present. A final assumption states that marked and unmarked animals are correctly identified, counted, and recorded. The observing and recording must be precise and unbiased. These assumptions are often violated in real-world situations, resulting in biased estimates. Population demographics change, capture and sighting probabilities differ per individual, animals group together, animals lose their marks, and recording errors are made. In an open population with recruitment and mortality, the estimate is too high (Otis et al. 1978:10, White et al. 1982:3). Carothers (1973) noted population size may be underestimated when probabilities of capture �differ. Eberhardt (1969) reviewed several studies, noted the bias involved with unequal capture probabilities, and suggested revision of trapping methods (i.e. randomization of trap locations and shifting of traps) and modification of the Lincoln Index ( i.e. N - (nl + nZ)(nl + nz - mz) /2mz where variables are defined as previously). In contrast, Magnusson et al. (1978) felt that violations of the equal capture probability assumption will not bias the estimate if the probability of being marked and recaptured are independent and based on beta distributions (i.e. where capture methods differ from recapture methods). However this condition will only apply to populations where capture methods differ from recapture methods. In regard to sighting probabilities, Strandgaard (1967) observed that variations in behavior of individual roe deer affected the animal's likelihood of being sighted; consequently the estimate was depressed. Little research has been done with regard to clumping, however Quinn (1980), using schooling populations of fish, provided a mark-recapture estimate accounting for clumping given specific marking schemes or population characteristics. Interactions of these violations of assumptions result in an estimate that is biased, imprecise, or both. In addition to problems associated with violations of model assumptions, the mark-recapture estimator is influenced by the number of animals marked. For example, given a single marking session, Bartmann et al. (1987) recommended that at least 45% of a mule deer population be marked in order to achieve a reliable estimate for small populations. In a study where roe deer were marked as they were caught, Strandgaard (1967) concluded that at least 66% of the population must be marked to obtain estimates that were acceptably precise and unbiased. The number of recaptures or resightings can also influence the estimator. Improved estimates are achieved through multiple recaptures or resightings. Schnabel (1938) and Darroch (1958) described this method of mark-recapture with mUltiple recaptures. At each recapture, the proportion of marked animals was recorded, unmarked animals were marked, and all were then returned to the population. The proportion of marked animals over all recapture occasions was then used to estimate the population size and variance. Rice and Harder (1977) used a similar multiple recapture method with white-tailed deer but did not mark after the initial marking because resightings were done via helicopter surveys. These researchers emphasized the importance of multiple resightings to strengthen the population estimate by increasing precision. When all animals caught in a single capture occasion are marked and recaptures result from multiple resightings, separate Lincoln-Petersen estimates result from each resighting. These individual estimates must be combined to obtain one overall population estimate, and the estimator chosen also affects the accuracy of the estimate. The joint maximum likelihood estimator from a hypergeometric distribution, or JOMLEHD (Chapman 1951, Seber 1982:59, Seber 1986, Bartmann et al. 1987), possesses the optimal properties of any of the estimators. Bartmann et al. (1987) and White and Garrott (1989:273) noted the advantages of this �148 estimate over the mean or the median: the variance is minimized, the confidence interval width is narrower, and the lower confidence bound is never lower than minimum number alive. Using JOMLEHD reduces bias and increases precision of the estimate (for definitions see White et al. 1982:18-19). Specifically, White and Garrott (1989:273) and Seber (1982:59) define the maximum likelihood estimate (MLE) as the value of N where the following is maximized: k+l IT i-2 [ ::] [:,::: 1 (equation 4) [ :, 1 and the terms are defined for all i - 2 to k+l number of sightings. This estimate can be found by iterative numerical methods. Aerial mark-resighting methods have been used to estimate numbers of free-ranging mountain sheep (McQuivey 1978, Leslie and Douglas 1979, 1986). By using one method for capture (e.g. drop nets) and another for resighting (e.g. aerial surveys), confounding factors (time differences, animal affinity or aversion to traps, and individual variability in these reactions) are eliminated (Otis et al. 1978:11). While flying, a helicopter, instead of a fixed-wing aircraft, allows greater maneuverability in the air and optimum airspeed and height above the ground (Seber 1986). However, questions about the effectiveness and accuracy of helicopter mark-resighting methods exist (Furlow et al. 1981, DeYoung 1986). When using the hypergeometric model (sampling without replacement), animals, both marked and unmarked, cannot be counted more than once (Rice and Harder 1977). Visibility bias can lead to erroneous population estimates (Pollock and Kendall 1987). If the collars or marks make the animals more conspicuous to the observer, the population will be underestimated because unmarked individuals will not be noticed as readily. Alternatively, Bear et al. (1989) overestimated an elk population when marked animals were misclassified. Packard et a1. (1985) noted that habitat, weather conditions, and observer experience can all lead to an underestimation of population abundance. Floyd et al. (1979) in studies of deer and Samuel et al. (1987) in studies of elk examined the effects of observer bias and evaluated the effects on counts of observers, changing ground and weather conditions,and different cover types. With moose, LeResche and Rausch (1974) noted the inaccuracy of the counts given observer experience, number of observers, snow conditions, habitat and terrain, and time of day. Finally, Caughley (1974) listed several factors that can affect the observer and the probability that an animal will be seen: thickness of cover, background, lighting, animal's color, animal's movement, animal's size, observer's eye sight, level of fatigue, speed of travel, altitude, and strip width. �149 Because this method of mark-resighting, with one marking period and multiple aerial resightings, is used fairly extensively, the estimator of population size should be precise and unbiased. To date, few studies have mathematically assessed the extent of problems inherent in this method. To improve estimates, the mark-resighting method needs to be evaluated quantitatively with simulations, using realistic data. With a known population size, the effectiveness of JOMLEHD estimator and its robustness to each individual assumption need to be simulated and tested. Similarly, violations of the assumptions need to be simulated and tested with populations of known size. METHODS For all sets of simulations, the precision and bias of the JOMLEHD estimator were assessed using known population sizes, different sighting probabilities, different capture probabilities (proportion of population marked), and different numbers of sighting occasions (flights). The Statistical Analysis System (SAS Institute Inc. 1985) was used for Monte Carlo simulations. Population sizes of 50, 100, 200, and 500 were used with 5, 10, 15, and 20 sighting occasions. Capture and sighting probabilities were assumed to be equal and constant for the whole population and were assigned values, respectively, of 0.1, 0.3, and 0.5, and 0.1, 0.3, 0.5, and 0.7. For each simulation, the number of marked and unmarked animals, and the number of marked and unmarked animals seen per flight were generated as random numbers from the binomial distribution. The number of marked animals was restricted to be greater than or equal to 10 and less than 80% of the true population size to simulate reasonable expectations of actual field trapping occasions. For each simulation, a JOMLEHD estimate, the 95% lower and upper confidence interval values, the confidence interval length or CIL (the difference between the lower and upper confidence interval values), and the coverage of the true population size by the confidence interval (above, below, or covering) were then determined. To minimize possible variations due to sampling, 1000 replications were performed for each of the 192 possible scenarios; i.e. 1000 JOMLEHD estimates were calculated for each scenario of true population size, number of flights, sighting probability, and capture probability, resulting in 192000 estimates. Simulations Several sets of simulations were run. First, the overall JOMLEHD estimator performance was evaluated given that all assumptions were met. The null hypothesis tested was that neither capture probabilities, sighting probabilities, true population size, nor number of sighting occasions affect the JOMLEHD estimate. For this set of simulations, sighting probabilities were equal and constant for the entire population. Second, sighting probabilities that vary per flight were simulated. null hypothesis tested was that sighting probabilities that change The �150 between flights do not affect the JOMLEHD. Sighting probabilities changed each flight with a value between 0.05 to 0.95 and a mean for all flights of 0.5, and were generated as random numbers from the uniform distribution. Third, a continuous range of heterogeneity in sighting probabilities of individuals was simulated. The null hypothesis tested was that heterogeneity of sighting probabilities of individuals does not affect the JOMLEHD. Four beta distributions and their corresponding parameters were used: [SR) skewed to the right (a - 2, ~ - 4), [SL) skewed to the left (a - 5, ~ - 2), [BE) symmetrically bell-shaped (a - 3, ~ - 3), and [EX) distribution with greater proportions in the extremes (a - 0.5, ~ 0.5). From the specified beta distribution, SAS randomly generated a sighting probability for each animal from the gamma distribution: Yl - rangam (seed, a) Y2 - rangam (seed, ~) sighting probability - Yl / (Yl + Y2)' 2.5 2 , i i "\ "\ i , ...• (II 1'1 41 .." ~ .., 1.5 ...• ...• ...• .t:I all .t:I ~ i ..,~ .... - ..• " I I BE " , ")~ / / M J " "", / , I "'- " " / I II I 0.5 \ "\ .~.~,_._. ' ._o.,_·-_._. __ '_'_'_'_'_'", EX I " I I 0 a "" "\ \ / "\ "\ \ , I , " '.'·t.... 0 0.. j 11"----., / " ,\ I I I " I '. 1 1 , SR ! ! \ .•...~.'" \ \ \ \ "" , \ ,.. / 0.2 O.i 0.6 . 0.8 "\ "-, •••... 1 Sighting proDaDl1lty Figure 1. Beta distribution used in simulations with heterogeneity individual sighting probabilities in �1S Fourth, probabilities of sighting that change exponentially were simulated. The null hypothesis tested was that habituation to the helicopters leading to increased or decreased sighting probabilities with increased number of flights does not affect the JOMLEHD. Maximum sighting probabilities were set at 0.55, 0.70, and 0.90. Minimum sighting probabilities were set at 0.0.10, 0.30, and 0.45. The rates of change were set at lambda values of 0.15, 0.35, and 0.60: sight probability All possible - maximum - ((maximum (1 - e -lambda combinations of the variables * minimum) x flight number were )). simulated. Fifth, a range of aggregations among marked animals were simulated to evaluate the degrees of independence of sighting probabilities. The null hypothesis tested was that nonrandom associations among animals does not affect JOMLEHD. Group size was randomly generated from a uniform distribution ranging from 1 to 24. Sighting probabilities were assumed constant for all group sizes and were given values of 0.1, 0.3, 0.5, and 0.7. Analysis To evaluate the estimator, a mean of the JOMLEHD estimates for each scenario was used. With PROG UNIVARIATE in SAS (SAS Institute Inc. 1985), standard deviations, variances, standard errors, coefficients of variation, and confidence intervals for the mean estimates were computed. To compare the estimates within a given population size, mean squared error (MSE) was calculated using the following equation: MSE - variance + (bias)2. MLEs are assumed to have asymptotically normally distributed estimates; goodness-of-fit tests using the Shapiro-Wilk statistic, W, were used to test the normality assumption by the fit of the simulation results to the normal distribution. Percent relative estimates: bias (PRB) was used to evaluate (estimate - true population the bias of the size) x 100. PRB true population size The Students's t value was used to test the hypothesis that bias equalled zero. The analysis of variance procedure indicated which variables affected the estimate effects: proc anova; class pop flight sight capture; model prb-poplflightlsightlcapture; �)_ l~? where pop - true population size, flight = number of flights, sight = sighting probability, and capture - capture probability (SAS Institute Inc. 1987). The general linear models procedure was used for regression on the variables: proc glm; model prb-poplflightlsightlcapture; Another GLM and a response surface regression was used on all variables and all two-way interactions for lack of fit information, and for analysis of linear, quadratic, crossproduct, and total regression surfaces for optimum response: proc glm; model prb-pop flight sight capture pop*pop flight*flight sight*sight capture*capture pop*flight pop*sight pop*capture flight*sight flight*capture sight*capture; proc rsreg; model prb-pop flight sight capture / lackfit; Tukey's studentized range test was used to detect differences in the mean values by population size. This test was chosen over other range tests because it is the most conservative and assumptions are not violated so robustness is not a consideration. Internal SAS problems prevented an overall Tukey's test but the test could be run by population size: proc glm; class flight sight capture; model prb-flight sight capture flight*flight sight*sight capture*capture flight*sight flight*capture sight*capture; means prb-flight sight capture flight*flight sight*sight capture*capture flight*sight flight*capture sight*capture / tukey; by pop;. The difference between the upper and lower 95% confidence interval values (CIL) was used as a measure of precision. The same programs to evaluate PRB were used here. used Coverage was used as a general indicator of both bias and prec~s~on of the estimate. The percent of the estimates' confidence intervals covering, above, or below the true population size was recorded. For the analysis, above and below categories were reclassified as 'not covers'. Categorical data modeling was used for logistic regression with maximum-likelihood analysis on covers or not: �133 proc catmod; direct pop flight sight capture; model covers-poplflightlsightlcapture ml;. / pred=prob Logistic regression without the direct statement indicated nonlinear effects of variables on the estimate. Insufficient memory prevented overall regression with this method, but with true population size removed, the analysis could be completed: proc catmod; model covers-flight by pop;. This analysis was repeated an I sight I capture / pred-prob ml; for all sets of simulations. RESULTS AND DISCUSSION Basic simulations In general, the precision of the estimate increases and the bias decreases as true population size, number of sighting occasions, capture probability, or sighting probability increase. The percent coefficients of variation of the mean estimate decrease as the four variables increase. Also, the differences between values for upper and lower confidence intervals of the mean estimate decrease, indicating a narrower distribution of estimates around the mean. Within a given population size, the variances of the estimates, the standard errors of the means, and MSE decrease as number of flights, capture probability, or sighting probability increase. At a population size of 50, all sets of estimates are significantly different from the normal distribution (P<O.Ol). At a population size of 100, all but 2 of the 48 scenarios are not normally distributed; at 200, 32 are not normally distributed; and at 500, 18 are not normally distributed. As population size and consequently the sample size increases, more distributions approach the normal distribution and the MLE normality assumption is not violated as often. No other trends in normality are evident as number of flights, sighting probability, or capture probability changed. Percent relative bias (PRB) was used to assess bias. PRB ranges from 60 to 548%. PRB can be negative if the estimate is lower than the true population size. As population size, number of flights, sighting probability, and capture probability increase, the mean PRB estimates, the standard deviations of these estimates, and the differences between the maximum and the minimum values decrease. The coefficients of variation increase as the four variables increase. From the ANOVAs, the sighting and capture probabilities are the main factors affecting the PRB. All variables and all interactions have a highly significant (P<O.OOl) effect on PRB. When GLM for regression was used, all variables and interactions affect the estimate (P<O.Ol, R-square-0.0249 , �154 and F Value-326.64). From the GLM and RSREG with all two-way interactions, the total regression on all variables and two-way interactions fits better than linear, quadratic, crossproduct, or total regression with all variables and all interactions (P<O.Ol, Rsquare-0.025l, and F Value-353.10). All of the variables affect the estimate (P<O.Ol). The lack of fit is high (P<O.OOl) so a high amount of variability aside from the pure error prevents a well-fitted model. The standard errors of the estimates of the regression equation reflect that capture probability and the square of capture probability have the most variability with PRB. Sighting probability, its square, and sighting probability*capture probability also have some variability. With the Tukey's test, 5, 10, and 15 flights have significantly different (P<0.05) PRB with population size 50, and 5 and 10 flights with greater population sizes. The sighting probabilities of 0.1, 0.3, and 0.5 are significantly different with all population sizes except 500, where only 0.1 and 0.3 are different. Capture probabilities 0.1 and 0.3 are significantly different for all population sizes. Thus, minimum variability occurs with 15 flights with population size 50 or with 10 flights with larger population sizes, with sighting probabilities of 0.5 or 0.3 with population size 500, and with capture probabilities of 0.3. The estimated regression equation for PRB is: PRB - 19.40 - (0.02*pop) - (0.80*flight) - (30.ll*sight) (25.89*capture) + (O.Ol*flight*flight) + (13.48*sight*sight) + (10.84*capture*capture) + (O.Ol*pop*sight) + (O.Ol*pop*capture) (0.49*flight*sight) + (0.45*flight*capture) + (18.20*sight*capture) where pop-population size, flight-number of flights, probability, and capture-capture probability. + sight-sighting The confidence interval length (CIL) was used to evaluate precLsLon. CIL values vary from'l to 12001. In general, the variability decreases as the four variables increase. The mean elLs and the coefficients of variation decrease as number of flights, capture probability, or sighting probability increase. The coefficients of variation decrease also as population size increases. The maximum values of PRB decrease. All four variables and all interactions have a highly significant impact on CIL (P<O.OOl). Sighting probability, capture probability, true population size, and number of flights, in that order, have the greatest F values. When GUM for regression was used, all four variables and all interactions except true population size*flight and true population size*flight*capture probability affect the CIL (P<O.Ol, R-square-0.2329, and F Value-3885.ll). From the response surface regression and GLM with all two-way interactions, the total regression on all variables and twoway interactions fits better than linear, quadratic, crossproduct, or total regression with all variables and all interactions (P<O.Ol, Rsquare-0.2460, and F Value-4473.4l). All of the variables affect the estimate (P<O.Ol). The lack of fit is high (P<O.OOl) so a high amount of variability aside from the pure error prevents a well-fitted model. The standard errors of the estimate reflect that capture probability and the square of capture probability have the most variability. Sighting �155 probability, its square, and capture probability*sighting probability also have some variability. With Tukey's studentized tests for differences in the means, 5, 10, and 15 flights significantly differ for population sizes 50 and 100 (P<0.05). With larger population sizes, all of the flights have different CILs. Sighting probabilities of 0.1, 0.3, 0.5, and 0.7 all differ in their CILs at all population sizes. Capture probabilities of 0.1, 0.3, and 0.5 all differ in their CILs at all population sizes. Minimum variability is found with 15 to 20 flights, 0.7 sighting probability, and 0.5 capture probability. The estimated regression equation for CIL is: CIL - 357.69 + (0.49*pop) - (17.l0*flight) - (635.86*sight) (637.3l*capture) + (0.28*flight*flight) + (328.70*sight*sight) + (369.l8*capture*capture) - (0.24*pop*sight) - (0.32*pop*capture) + (10.87*flight*sight) + (11.46*flight*capture) + (472. 25*sight*capture) where pop-population size, flight-number of flights, probability, and capture-capture probability. sight-sighting The coverage was used to assess bias and precision. Percent coverage ranges from 93.6 to 96.7; percent of intervals above the true population size ranges from 1.3 to 3.5; and percent of intervals below the true population size ranges from 1.1 to 3.8. An approximately equal number of the estimates are above (4399 of 192000 - 0.02%) as are below (4538 of 192000 - 0.02%) the true population size. No trends were observed in coverage as true population size, number of flights, sighting probability, or capture probability change. A 95% confidence interval for coverage is calculated as 93.6 to 96.4%. Only four of the scenarios fall above this interval and none fall below. There is no apparent trend with these scenarios. With logistic regression, neither population size, number of flights, sighting probability, capture probability, nor any of their interactions affect coverage (P>O.Ol and likelihood ratio-0.3l88). When the model was tested to include possible nonlinear effects, none of the variables nor their interactions affect cover (P>O.Ol). Apparently, coverage is not predictably variable by using the factors population size, number of flights, sighting probability, and capture probability. Thus population size, number of flights, capture probability, and sighting probability affect the JOMLEHD. Regression equations for PRB and CIL can be used as indicators of bias and precision given these four variables. However, the lack of fit is high and the R-squares are low, especially for PRB. Consequently, the regression equations should only be used as general predictors. Coverage, to reflect both bias and precision, cannot be approximated based on the variables population size, number of flights, sighting probability, and capture probability. �156 Sighting probabilities changing per flight Sighting probabilities are the same for all animals within one flight, but vary per flight between 0.05 and 0.95. The mean sighting probability is 0.5. Similarly to the basic simulations, the variances of the estimates, the standard errors of the mean, the percent coefficients of variation, and the mean squared errors decrease with increasing number of flights or capture probability. However in this scenario, percent coefficient of variation increases with increasing population size. The variability increases faster per population size with changing sighting probabilities than with constant. The decreased variability can be accounted for in the few flights where sighting probability is very high and the precision of the estimates is thus very good. For this same reason, the basic simulations with sighting probability equal to 0.5 are more biased than the simulations with sighting probability changing per flight (P<O.Ol). PRB ranges from -36.5 to 78.0%. The mean PRB estimates and the standard errors of the means decrease as population size, number of flights, or capture probability increase. The coefficients of variation increase as population size or capture probability increases, and remains fairly constant as number of flights changes. Changes per flight do not have as much variability as constant sighting probabilities in maximum and minimum values. The main differences in PRB between basic simulations and changing sighting probabilities falls in small number of flights and small capture probabilities. CIL ranges from 0 to 344 and is significantly lower for changing sighting probabilities than for the basic simulations (P<O.Ol). The mean CIL, the standard errors of the mean, and the coefficients of variation decrease as number of flights or capture probability increase. The coefficient of variation also decreases as population size increases. All of the variables have a significant impact on lowering the CIL for changing sighting probabilities as compared to constant. Coverage ranges from 77.1 to 96.6%; percent of intervals above the true population size ranges from 1.3 to 20.8%; and percent of intervals below the true population size ranges from 1.1 to 3.7%. Percent coverage increases as population size increases and decreases as number of flights or capture probability increase. Percent above changes opposite that of percent covers. Percent below increases with increasing population size or capture probability and decreases with decreasing number of flights. Coverage is especially low for small populations size 50. More confidence intervals are above (2073 of 48000 - 4.3) than are below (1145 of 48000 - 2.4) the true population values. One scenario fell above the 95% confidence interval for percent coverage (population size 500), and 13 fell below. With sighting probabilities changing per flight, those flights with higher sighting probabilities greatly decrease the variance and the bias. The precision is greater and the bias lower here compared to the �1 -~), basic simulations because these simulations as high as 0.95 while the basic simulations Heterogeneity of individual had sighting probabilities had values only up to 0.7. sighting probabilities A continuous range of heterogeneity in individual sighting probabilities was simulated. The beta distribution that was skewed to the right (0 - 2, ~ - 4), SR, had a mean sighting probability of 0.3. The beta distributions that was symmetrically bell-shaped (0 - 3, ~ - 3), BE, and the distribution that was distributed with the greatest proportions in the extremes (0 - 0.5, ~ - 0.5), EX, had mean sighting probabilities of 0.5. The beta distribution that was skewed to the left (0 - 5, ~ - 2), SL, had a mean sighting probability of 0.7. Like the basic simulations, the precision of the estimate generally increases and the bias generally decreases as true population size, number of sighting occasions, capture probability, or mean sighting probability increase. The same trends as the basic simulations occur; percent coefficients of variation and normality increase as four variables increase, and the variances of the estimates, the standard errors of the means, and MSE decrease within a population size with increases in the three variables. With more flights, EX becomes more variable than SR. The variability (percent coefficients of variation, variances of the estimates, the standard errors of the means, and MSE) is greater for heterogeneity of individuals than the basic simulations when comparing the mean sighting probabilities. With 5,10, and 15 flights, a capture probability of 0.1 in the basic simulations has the greatest variability. With 20 flights, the EX distribution from heterogeneity in individual sighting probabilities has the most variability. PRB ranges from -49 to 298%. As with the basic simulations, the mean PRB estimates and the standard errors of the means decrease as population size, number of flights, capture probability, or mean sighting probability increase. The coefficients of variation increase and the differences between the maximum and the minimum values decrease (except for number of flights) as the four variables increase. When compared to the same mean sighting probabilities from the basic simulations, PRB appears lower for constant sighting probabilities than for heterogeneity in sighting probabilities in all scenarios. elL ranges from 0 to 895. The mean elL, the standard errors of the means, the coefficients of variance, the maximum values, and the differences between the extreme values decrease with increasing number of flights, means sighting probability, or capture probability. elL appears slightly lower in basic simulations than for heterogeneity for all scenarios except EX when comparing mean sighting probabilities. Percent coverage ranges from 42.5 to 91.6; percent of intervals above the true population size ranges from 3.4 to 29.8%; and percent of �158 intervals below the true population size ranges from 4.0 to 32.0%. The EX distribution has especially low coverage compared to the other distributions. An approximately equal number of the estimates are above (23397 of 192000 - 12.2%) as are below (24575 of 192000 - 12.8%) the true population sizes. As number of flights increase, percent covers decreases and percent above and below increase. None of the scenarios fell in the 95% confidence interval for coverage; all fell below the 93.6 to 96.4% range. With heterogeneity in individual sighting probabilities, the population estimate is less variable. However, the estimates tend to be lower than the true value and coverage is too low. The resulting estimate would be biased but precise. Populations composed of many animals that are seen frequently or rarely lead to especially biased estimates. A greater number of flights would discern the lack of coverage. Animals becornin~ helicopter shy In this set of simulations, sighting probabilities decrease with increasing number of flights. Sighting probability is a function of a maximum value, a minimum value, and lambda (rate of decrease). Precision increases and bias decreases as population size, number of flights, or capture probability increase. In regard to sighting probability, as the rate of changes increases, the variability and the bias increase; with greater maximums or minimums, the variability and the bias decrease. Mean sighting probability ranges from 0.204 to 0.743. PRB ranges from -44.0 to 450.5%. As with the basic simulations, PRB and its variability decrease with increasing population size, number of flights, or capture probability. The difference of the extremes decreases as number of flights or capture probability increase. elL ranges form 0 to 4049. As number of flights or capture probability increase, the elL and its variability decrease. Likewise, the differences in the extremes decrease as these variables increase. As popUlation size increases, the elL and the variability increase. This variability is reflected in the minimum values of elL at higher popUlation sizes. . Percent coverage ranges from 91.8 to 97.8%; percent of intervals above the true population size ranges from 0.6 to 4.6%; and percent of intervals below the true population size range from 0.4 to 5.0%. Percent below tends to be greater than percent above. Coverage is lower in these simulations than in the basic simulations (29496 of 1295000 2.3% above and 30462 of 1295000 - 2.4% below). Thus with animals becoming less visible with time, the coverage of the estimates decreases. The bias or the elL are not apparently different �159 from those of the basic simulations. The initial sighting probability is important as this is the highest sighting probability. Groups of individuals together all the time -(all group sizes have equal sighting probabilities) In this set of simulations, groups remained together for all flights and sighting probabilities were constant for all group sizes. Similar to the other simulations, the precision increases and the bias decreases as population size, number of flights, capture probability, or sighting probability increase. The same general trends occur in these simulations as the basic simulations. Percent coefficients of variation decrease as the four variables increase; within one population size, the variances of the estimates, the standard errors of the means, and MSE decrease as number of flights, capture probability, or sighting probability increase; and normality increases as population size increases. In general, the basic simulations are more variable than the group simulations. With low sighting probabilities, the estimates with groups of individuals have more variability than the basic simulations without groups; with high sighting probabilities, the basic simulations have more variability than the group simulations. With more flights, only sighting probabilities of 0.1 are more variable in the group simulations than the basic simulations. At higher population sizes, as capture probability increases, the basic simulations become more variable. PRB ranges from -78.0 to 760.0%. As with the basic simulations, the mean PRB estimates and the standard errors of the means decrease as population size, number of flights, capture probability, or sighting probability increase. The coefficients of variation decrease as all but sighting probability increase; sighting probability has no trend in coefficient of variation. The minimum values continue to increase as the four variables increase but the maximum values did not follow a trend. There is no significant difference (P<O.Ol) in PRB between the basic simulations and the group simulations. CIL ranges from 0 to 17236. The mean CIL, and standard errors of the means, and the coefficients of variation of estimates increase as number of flights, capture probability, or sighting probability increase. In general, the maximum value for CIL also decreases as these three increase. CIL is significantly lower for basic simulations than for group simulations (P<O.Ol). The main differences occur in the number of flights, lower capture probabilities, and sighting probabilities. Percent coverage ranges from 22.7 to 96.6%; percent of intervals above the true population size ranges from 1.1 to 76.5%; and percent of intervals below the true population size ranges from 0.6 to 3.9%. Coverage is much poorer at low population sizes with high sighting probabilities. Percent coverage increases with increasing population �160 size. Percent coverage decreases with increasing number of flights and sighting probability. The percent above (13404 of 192000 - 7.0%) changes more than the percent below (4659 of 192000 - 2.4) with changes in percent coverage. As population size increases, percent above decreases; as number of flights, capture probability, or sighting probability increases, the percent above increases. With capture probability, percent above changes inversely with percent below. Percent coverage increases with population size, but decreases with number of flights or sighting probability. Two scenarios fell above the 95% confidence interval for percent coverage (in population sizes 200 and 500) and 47 fell below (all but one in population sizes 50 and 100). Coverage is much poorer with groups of individuals than with the basic simulations. Percent below is similar between the two sets but percent covers is lower with small population sizes, all number of flights, all capture probabilities, and all sighting probabilities (especially higher values). With groups of individuals, the population estimate may appear to be less variable but the coverage is much poorer. With small population sizes, the coverage is low. With low capture probabilities, low sighting probabilities, or with few flights, the population may be overestimated. Thus, violations of the assumptions of mark-resight with mountain sheep lead to estimates that are more variable, more biased, and less precise than would be expected. All of the simulations are being analyzed more rigorously and in more detail to find ways to avoid or limit these problems. Table I. Trends of the simulations where all assumptions are met. as compared Scenario PRB ell x Sighting probabilities changing per flight lower lower Heterogeneity of individuals higher Helicopter shy Groups to basic simulations X above X below lower higher same same lower higher higher same same more variable more variable more variable same higher lower higher lower covers �l6l LITERATURE CITED Bartmann, R.M., G.C. White, L.H. Carpenter, and R.A. Garrott. 1987. Aerial mark-recapture estimates of confined mule deer in pinyonjuniper woodland. J. Wi1d1. Manage. 51(1):41-46. Bear, G., G.C. White, L.H. Carpenter, and R.C. Garrott. 1989. Observabi1ity bias in mark-resighting estimates of elk populations. J. Wild1. Manage. (in press) Bergerud, A.T. and F. Manuel. 1969. Aerial census of moose in central Newfoundland. J. Wi1d1. Manage. 33(4):910-916. Carothers, A.D. 1973. Capture-recapture methods applied to a population with known parameters. J. Anim. Ecol. 42(1):125-146. Caughley, G. 1974. Bias in aerial survey. J. Wildl. Manage. 933. Caughley, G. 1978. Analysis of vertebrate Sons. New York, New York. 234pp. populations. 38(4):921- John Wiley and Chapman, D.G. 1951. Some properties of the hypergeometric distribution with applications to zoological sample censuses. Dniv. Calif. Publ. Stat. 1(7):131-160. Darroch, J.N. 1958. The multiple-recapture census. I. Estimation closed population. Biometrika 45(3/4):343-359. of a DeYoung, C.A. 1986. Accuracy of helicopter Texas. Wild1. Soc. Bull. 13:146-149. surveys of deer in south Eberhardt, L.L. 1969. Population estimates J. Wildl. Manage. 33:28-39. from recapture frequencies. Eberhardt, L.L., D.C. Chapman, and J.R. Gilbert. 1979. A review of marine mammal census methods. Wi1dl. Monogr. 63:1-46. Floyd, T.J., L.D. Mech, and M.E. Nelson. 1979. An improved method of censusing deer in deciduous-coniferous forests. J. Wi1d1. Manage. 43(1):258-261. Furlow, R.C., M. Hader1ie, and R. Van den Berge. 1981. Estimating bighorn sheep population by mark-recapture. Desert Bighorn Council, Trans. 1981:31-33. a Kufe1d, R.C., D.C. Bbwden, and D. L. Schrupp. 1987. Estimating mule deer density by combining mark-recapture and telemetry data. J. Mamm. 68(4):818-825. LeResche, R.E. and R.A. Rausch. 1974. Accuracy and prec~s~on of aerial moose censusing. J. Wildl. Manage. 38:175-182. Leslie, D.M., Jr. and C.L. Douglas. 1979. Desert bighorn sheep of the River Mountains, Nevada. Wildl. Monogr. 66:1-56. �l62 Leslie, D.M., Jr. and C.L. Douglas. 1986. Modeling bighorn sheep: current abilities and missing Wildl. Natur. Res. Conf. Transc. 51:62-73. demographics of links. N. Amer. Lincoln, F.C. 1930. Calculating waterfowl abundances on the basis of banding returns. U.S. Dept. Agric. Circ. 118. 4pp. Magnusson, W.E., Caughley, G.J., and G.C. Grigg. 1978. A double-survey estimate of population size from incomplete counts. J. Wildl. Manage. 42(1):174-176. McQuivey, R.P. 1978. The desert bighorn Game BioI. Bull. No.6 8lpp. sheep of Nevada. Nevada Fish, Otis, D.L., K.P. Burnham, G.C. White, and D.R. Anderson. 1978. Statistical inference from capture data on closed animal populations. Wildl. Monogr. 62:1-135. Packard, J.M., R.C. Summers, and L.B. Barnes. 1985. Variation of visibility bias during aerial surveys of manatees. J.Wildl. Manage. 49(2):347-351. Petersen, C.G.J. 1896. The yearly immigration of young plaice Limfjord from the German sea. Rept. Danish BioI. Stn. into Pollock, K.H. and W.L. Kendall. 1987. Visibility bias in aerial surveys: a review of estimation procedures. J. Wildl. Manage. 51(2):502510. Quinn, T.J. II. 1980. Sampling for the abundance of schooling population with line-transect, mark-recapture, and catch-effort methods. Ph.D. dissertation. University of Washington. Rice, W.R. and J.D. Harder. 1977. Application of multiple sampling to a mark-recapture census of white-tailed Wildl. Manage. 41(2):197-206. aerial deer. J. Samuel, M.D., E.O. Garton, M.W. Schlegel, and R.G. Carson. 1987. Visibility bias during aerial surveys of elk in northcentral Idaho. J. Wildl. Manage. 51(3):622-630. SAS Institute version Inc. 1985. SAS procedures guide for personal computers, 6 edition. Cary, NC: SAS Institute Inc. 373pp. SAS Institute Inc. 1987. SAS/STAT guide for person computers, edition. Cary, NC: SAS Institute Inc. 1029pp. version Schnabel, Z.E. 1938. The estimation of the total fish population lake. Amer. Math. Mon. 45(6):348-352. of a 6 �163 Seber, G.A.F. 1982. The estimation of animal abundance and related parameters (2nd edition). Macmillan Publ. Co., Inc. New York, New York. 654pp. Seber, G.A.F. 1986. A review for estimating animal abundance. Biometrics 42:267-292. Strandgaard, H. 1967. Reliability of the Petersen method tested on a roe-deer population. J. Wildl. Manage. 31:643-651. White, G.G., D.R. Anderson, K.P. Burnham, and D.L. Otis. 1982. Gapturerecapture and removal methods for sampling closed populations. Los Alamos National Laboratory. LA-8787-NREP. Los Alamos, N.M. White, G.G. and R.A. Garrott. primer. (in press) 1989. Analysis Prepared by Andrea K. Neal Graduate Research Assistant of biotelemetry data - a ��165 JOB PROGRESS State of Colorado Project No. 01-03-048 Work Plan No. 3 Job No. 2A Period Covered: Author: REPORT Mammals CW-153-R-2) 2 Research Pronghorn Investigations Habitat Selection and population Performance of a Pioneering Pronghorn Population July 1, 1988 - June 30, 1989 T. M. Pojar ABSTRACT Seven radioed pronghorn (Antilocapra americana) remain of the original 9 that were radioed in Dec 1986. Another trapping operation was conducted in Dec 1988, during which 67 pronghorns were captured. Twenty new radios were deployed (15 on adult females and 5 on adult males). Data are analyzed on the original 7 radioed animals for the period Dec 1986 through Sep 1988. Seasonal and yearlong areas of habitation were calculated. The wintering area of the pronghorn population did not overlap any of the critical winter range for mule deer (Odocoileus hemionus) in Middle Park. During the 2 years the pronghorn population has been monitored, it has increased at an annualized rate of 41.4%. ��- 1 .- HABITAT _ SELECTION AND POPULATION PERFORMANCE PRONGHORN POPULATION OF A PIONEERING Thomas M. Pojar P .N. OBJECTIVE Describe population dynamics pronghorn population. and habitat use of a pioneering, expanding SEGMENT OBJECTIVES 1. Describe seasonal population. and annual distribution of the Middle 2 Determine necessary 3. Monitor population dynamics of Middle Park pronghorns with: a. Ground counts to describe changes in population size. b. Ground counts to quantify population sex and age composition. sample sizes of radio-collared animals to describe habitat preferences. Park pronghorn and observations ACKJmWLEDGMENTS The trapping operation was facilitated through the efforts of Steve Steinert, Joe Gerrans and Area 9 personnel: J. Liewer, J. Claassen, R. Firth, A. Chappell, B. Thompson, B. Sigler, M. Middleton, J. Hicks, and J. Frank. Personnel from the Kremmling Bureau of Land Management district provided assistance during trapping as did several local Kremmling residents. K. Cushman and S. Fairbanks served as recorders and M. Atkins managed marking supplies during the trapping operation. M. Miller assisted in blood sample collection and processing for both hematological diagnostic panel and leptospirosis testing. The CDOW NW Region contributed helicopter time for the trapping operation. D. Schrupp, A. Cade, and J. Miller were instrumental in the data processing and mapping of seasonal areas of habitation. STUDY AREA The study area is described in Pojar (1988:183-184). METHODS AND MATERIALS The trap used was a corral trap with peripheral curtains and an interior diagonal curtain (same as in 1986). Trapping and handling procedures followed those sanctioned by the Pronghorn antelope Workshop (Anonymous 1984). ') . �163 Radio tracking was done mostly from the ground to increase the probability of observing and~dentifying animals with numbered plastic collars. Fixed wing aircraft was used if an animal could not be located after reasonable effort from the ground. Legal descriptions of animal locations were recorded to the nearest quarter mile then converted to UTM (U.S. Army 1973) coordinates for computer processing. All areas of habitation are based on m~n~mum convex polygons as calculated by the program McPAAL (Stuwe ND). The map plots of pronghorn habitat, areas of habitati ~, mule deer critical winter range, and individual point locations were accomplished through the SAGIS (Bartholow 1981) software package. The base map used was: C.S. Department of Agriculture Forest Service R~lltt National Forest, Colorado, Sixth Principal Meridian, 1975 (reprint 198 The original scale of 1:126720 was reduced to 1:230400 via photographic tec~niques at the Colorado State University Photo Lab. RESULTS Trapping and Marking On 15 Dec 1988, 67 pronghorns were captured in 2 catches at a trap site east of Kremmling on BLM land (T1N,R80W,NE 1/4 S 11). There were no pronghorn fatalities or injuries as a result of the trapping operation. Twenty new radio transmitters (Advanced Telemetry Systems) were deployed, 15 on adult females and 5 on adult males. Tables 1 through 4 detail the current status of marked pronghorns in Middle Park. In summary, there are 24 functional radio transmitters (4 remaining from the 1986 trapping) (Table 1), 29 animals with blue plastic collars and either a blue or yellow ear tag (Table 2), and 14 animals with a blue ear tag only (fawns were not collared because of potential growth) (Table 3). Of the 47 animals trapped in 1986, 13 were recaptured in the 15 Dec 1988 trapping operation (Table 4). Population Size and Structure The winter population in 1986-87, the first year of this study, was 80 animals. The herd structure at that time, based on the sample of 47 trapped animals, was 36B:100D:77F. The following year, 1987-88, the winter population increased by 55% to 122 with a structure of 44B:100D:70F. This represents production of 95 fawns per 100 producing-aged does. Lower production, 61F:100 mothers, in 1988 resulted in a lower population growth rate of 31% and a winter 1988-89 population of 160. An estimate of the herd structure based on a sample of 108 animals on 26 Oct 1988 was 40B:100D:32F and the sample of 67 animals trapped 15 Dec 1988 was 40B:100D:40F. From these estimates and the total count of 160 animals in the 1988-89 winter population, it was projected that the population should contain 37 yearling and mature bucks. On several occasions during winter, it was possible to get an accurate count of bucks. These counts were consistent at 35-36 bucks. This offers support that it may be possible to obtain accurate herd structure ta if surveys are done shortly after the breeding season when the breeding se~~on social structure has broken down. �Tests for Leptospirosis A random sample of the blood samples (n-10) collected during the 15 Dec 1988 trapping operation was tested for titers against 5 strains of the leptospirosis bacteria (Leptospira sp.). The sample was limited only to females age 3+ because they would be the most likely to have been exposed to the disease if it were present in the population. They were all negative (Table 5). The tests were done at the Diagnostic Laboratories, Colorado State University. Areas of Habitation Uith few exceptions, all radioed animals have been located every 2 weeks. Animal with radio frequency 149.370 was killed in a collision with an automobile on 14 July 1987 resulting in a minimal data base for her. Radio frequency 149.290 ceased to function curing April 1988. The bulk of the location analysis contained in this report, therefore, has been done on the remaining 7 radioed animals, all of which are mature does. The legal descriptions of radioed pronghorn locations were converted to UTM coordinates. Thus far, location data have been analyzed for the time period of Dec 1986 through Sep 1988. These data are reflected in Figures 1· - and Table 6. True seasons based on equinox and solstice did not fit the seasonal movement patterns of the pronghorns. The location data were arbitrarily assigned to seasons as follows: winter-Dec, Jan, and Feb; spring-Mar, Apr, and May; summer-Jun, Jul, and Aug; and autumn-Sep, Oct, and Nov. A finer delineation of seasons and transition periods will be possible as the data base of locations expands. Uinter Area of Habitation The winter area of use by the Middle Park pronghorn population is of particular interest of managers because of the potential overlap with mule deer use areas. The areas of "critical winter range" for mule deer, as defined by Tiedeman et al. (1987), are included in the digitized map for the pronghorns' winter area of habitation (Figure 1). The critical areas for mule deer are south-facing slopes and the area of use by pronghorn is mostly windblown ridgetops. For the winter of 1987-88, which was moderate to light in terms of severity, there was no overlap between the 2 species. Pronghorns are quite sedentary in winter (Figure 5). The average area of habitation for 7 radioed animals was 18 km2 (7 mi~ with a coefficient of variation of 14.22% which is much lower that for any other season (Table 6). The area described for the radioed animals will almost exactly describe the area for the entire population because they are usually in one large group during this time of year. Spring Area of Habitation Spring is a time of much movement and large variation among individuals (Figure 5). Spring areas ranged from 22 km2 (8.5 mi~ for one animal to 83 km2 (32 mi~ for another with the average for all animals of 51 km2 (48 mi~ (Table �1:-0 6). The dispersion of the radioed animals (Figure 2) does not accurately depict the tf'"uedispersion of the population. Reliable reports have been received of marked individuals (plastic neck collar and/or ear tag) in the Grandby/Hot Sulphur Springs and Muddy Pass areas and one ear-tagged animal was seen in North Park (See Pojar 1988 for details) during spring and summer of 1988. In 1989, sightings of several pronghorns were reported (J. Claassen, pers comm) in the Coffee Divide area approximately 3 miles North of Grandby. This group included one animal with a blue plastic collar. These sightings verify that a certain segment of the Middle Park population summer in the far eastern and northern reaches of the park but return to an area immediately north and east of Kremmling to winter (Figure 1). The new set of radios deployed in Dec 1988 include several animals that summer in the Northern part of the park (Muddy Pass area) but unfortunately, no radioed animals are included in the group that summers in eastern area of the park (Grandby area). Summer Area of Habitation During fawning and for several weeks thereafter females remain on relatively small fawning areas. The average area of habitation for this arbitrary definition of season is 24 krn2 (9 mi~ (Table 6). This season includes the time interval during which there are erratic movements of some does just prior to fawning and movements of does with fawns to join larger doe/fawn groups during the latter part of this period. The area of use by does during the fawning period is much smaller than for this entire season and will be much better described as more data points become available. The disbursement of animals is maximized during this season but some of them remain on portions of the wintering area yearlong (Figure 3). Once again, other sightings of ear-tagged or plastic collared animals verify that the summer dispersion of the entire population is greater than Figure 3 indicates. Autumn Area of Habitation This season encompasses the breeding season and the early stages of large wintering group formation. Location of the 7 radioed animals in Oct/Nov and their companions results in observing two-thirds or more of the entire population. Consequently, the distribution depicted in Figure 4 probably closely resembles the distribution of the whole population. Yearlong Area of Habitation A composite of all location points for 7 radioed animals was used to calculate the "yearlong" areas presented in Table 6. The average yearlong area of habitation was 79 krn2 (31 mi~ and was moderately variable between individual animals (c.v.-42.8l%) (Figure 6). �SUMMARY This population has had an annualized rate of increase of 41.4% during the 2 years thus far monitored. If this rate is maintained, the population will exceed 300 animals by the autumn of 1990. Management plans are to begin hunting when the population reaches 300 animals. Although the data base for radioed animals is minimal at this point, there are some trends of interest emerging. Wintering areas of pronghorns and mule deer did not overlap although they were in close proximity. This is in agreement with findings of a study in eastern Montana (Wood 1989) where these species had some habitat use overlap but maintained a high degree of spacial segregation. LITERATURE Anonymous, 1984. Proceedings CITED Management guidelines: Trapping and Translocation. of the Pronghorn Antelope Workshop 11:237-251. Bartholow, J. 1981. SAGIS-WINDOW documentation. U.S. Fish and Wildl. Serv., Western Energy and Land Use Team. BioI. Servo Program, Ft. Collins, CO. 40pp. Pojar, T.M. 1988. Habitat pioneering pronghorn 181-192. selection and population performance of a population. Colo. Div. Wildl. Res. Rep. July, pp Stuwe, M. ND. McPAAL micro-computer programs for the analysis of animal locations. Program Documentation. Cons. and Res. Center, Natl. Zool. Park, Smithsonian Inst., Front Royal, Virginia. Printout. 18 pp. Tiedeman, J.A., R.E. Francis, C. Terwilliger, Jr., and L.H. Carpenter. Shrub-steppe habitat types of Middle Park, Colorado. USDA Forest Service, Res. Paper RM-273, 20 pp. 1987. U. S. Army. 1973. Technical Manual: Universal transverse mercator grid. Headquarters, Dep. of the Army, Washington D.C. TM No. 5-241-8, 64 pp. Wood, A.K. 1989. mule deer. Prepared Comparative distribution J. M.amm. 70(2):335-340. ( Ul---v;_ A ili bY~7~ Wildlife Researcher C and habitat use by antelope and �1 ..l./":" " Table 1. Record of radio transmitters deployed as of 15 Dec 1988. includes 4 radio? that were deployed 16 Dec 1986. Species: Pronghorn Trap Site Location: Capture Freq. 148.390 148.400 148.410 148.430 148.440 148.450 148.460 148.520 148.530 148.540 148.550 148.560 148.570 148.580 148.590 148.600 148.620 148.630 148.640 148.650 148.700 148.750 148.850 148.950 Date: americana) BLM land, T1N,R80W,NE 1/4 S 11 15 Dec 1988 Nearest Town: Kremmling Wing trap wi curtains Method: Define seasonal and annual home range Sex Age Ear Tag F F F F F F F M F F F F F F M M F F F M F M F F 3+ 3+ 3+ 3+ 3+ 3+ 3+ Yr1 3+ Yr1 Purpose: Radio (Anti1ocapra The list Y 2 Y 4 B 25 Y 9 Y 7 Y 12 B 26 B 2 Y 1 B 6 2 B 8 2 Yr1 Yr1 3+ 3+ 3+ Yr1 3+ Yr1 Yrl 3+ 3+ Yrl B 27 B 5 Bll B 3 Y 3 B B B B B 14 23 24 16 17 Y 19 Y 10 B 10 Comments Radio put on 16 Dec 1986. Radio put on 16 Dec 1986. New radio. Radio put on 16 Dec 1986. Radio put on 16 Dec 1986. Replaces radio 149.250. New radio. New radio. Magnet left on. Replaces radio 148.420. New. New. New. New. New. New. Ear tgd as a fawn 12/16/86. New. New. New. New. New. Replaces Yellow # 8 collar. Replaces radio 149.210. New. �2. Table Species: Record of neck collars Pronghorn Trap Site Location: Capture (Antilocapra put on 15 Dec 1988. americana) BLM land. T1N.R80W,NE Define seasonal Ear tag Sex B 1 B 2 B 19 M M Y Y 27 B 21 Y 22 M M Y B 5 B 40 M 2 B 6 Y 28 M 3+ B 7 B 47 Y B 8 B 9 B 51 Y 32 B 28 3+ B 30 B 50 Y 42 Y 41 Y 17 M M M F F F F F F F F F F F F F F F F F F F F F Y 18 Y 21 F Bll 12 B 29 Bl3 B 14 B 15 B 16 B 30 B B 17 B 18 B 19 B 20 B21 B 22 B 23 B 24 B 25 B 26 B 27 B 28 B 29 Note: S 11 Nearest Town: Kremmling and annual home range Neck collar B 4 1/4 15 Dec 1988 Wing trap wi curtains Method: Purpose: B 3 Date: B 31 B 32 Y 39 B 34 B B B B B 36 37 38 41 39 B 42 B 43 B 45 B 44 B 46 B 49 An animal aged as "adult" Age 3+ 3+ 3+ 3+ Y Y 3+ 3+ 2 3+ 2 3+ 2 3+ 2 3+ Y Y Y 2 3+ Y 3+ 3+ 3+ Comments New Ear tagged as fawn Dec '86. New Replaces Yellow neck #10, "Adult" in Dec '86. New Ear tagged as fawn Dec '88. New New "Adult" in Dec '86. New New New New New Ear tagged as fawn Dec '86. New New New New New New New New New New New New New Ear tagged as fawn Dec '86. "Adult" in Dec '86. Collar not replaced. Replaces yellow neck # 9, Yearling in Dec '86. in 1986 was a yearling or older. �1 -, J.. /-+ Table 3. Record of blue ear tags put on 15 Dec 1988. Species: Pronghorn (Antilocapra Trap Site Location: Capture Bo land, T1N.R80W,NE 1/4 Date: S 11 15 Dec 1988 Nearest Town: Kremmling Wing trap wi curtains Method: Define seasonal Purpose: Note: americana) This list contains and annual home range only fawns of the year (one exception). Neck collars were not put on any fawns. Ear Tag B 1 B 4 B 7 B 9 B 12 B 13 B 15 B 18 B 20 B 22 B 33 B 35 B 48 B 52 BB yy Sex F M F M M M F M F M M M M M F F Comments New New New This is a yearling, did not get collar. New New New New New New New New New New Blue tag in each ear. Marked summer '88. Yellow tag in each ear. Marked summer '88. �Table 4. Species: S~~~us of animals marked 16 Dec 1986. Trap Site Location: Neck/Radio Collar Ear Tag 148.390 148.400 148.420 148.430 148.440 149.210 149.250 149.290 149.370 Y 1 Y 2 Y 3 Y 4 Y 5 Y 6 Y 7 Y 8 Y 9 Y 10 Yll Y 12 Y13 Y 14 Y 15 Y 16 Y 17 None None None None None None None None None None None None None None None None None None None None None Date Trapped: Pronghorn Jody Hill's, T2N,R80W, SW 1/4 S 36 Sex Age F F F F F F F F F M F M F F F F F F F 2+ 2+ 2+ 2+ 2+ 2+ 2+ 2+ Yr1 2+ Yr1 2+ 2+ 2+ 2+ Yr1 2+ Yr1 2+ 2+ 2+ Yr1 2+ Yr1 2+ 2+ Y 3 M F Yll Y 23 Y 27 Y 28 Y 29 Y 30 Y 31 Y 32 Y 34 Y 35 Y 36 Y 38 Y 39 Y 42 Y 43 Y 44 Y 45 Y 46 Y 47 Y 48 F F F(M) F(M) M F M 2+ M 2+ F M F Y 2 Y 4 Y 1 Y 9 Y 7 Y 10 Y 12 Y 20 Y13 Y 5 Y 8 Y 6 Y 14 Y 15 Y 17 Y 18 Y 19 Y 21 Y 22 Y 16 Y 24 Y 25 Y 26 Y 33 Y 40 Y 41 16 Dec 1986 M M F F F F M M M F F M F M F F F F F F 2+ F F Yrl F F F F F F F F F F Date Last Observed 12/29/88 12/29/88 12/29/88 12/29/88 12/29/88 12/29/88 12/29/88 8/11/88 7/14/87 10/12/88 9/30/87 3/17/88 11/22/88 4/7/88 11/22/88 10/2/88 12/29/88 12/15/88 12/15/88 11/22/88 12/6/88 11/22/88 4/21/88 12/6/88 12/6/88 12/15/88 12/29/88 not seen summ '88 12/15/88 12/15/88 9/27/88 10/12/88 not seen 12/15/88 not seen not seen not seen not seen 12/15/88 12/15/88 not seen not seen 6/13/88 10/20/87 not seen not seen Comments Raised 2 fernfawns '88 New radio #148.530. New radio #148.850. New radio #148.450. Radio quit 4/7/88. Killed by auto. New radio #148.750. New yellow neck #18. New blue neck #4. Trapped and released. New radio #148.600. In N. Park, Steinert Blue neck #2. A male!! Blue neck #6. A male!! New blue neck #9. New blue neck #16. New blue neck #30. �176 Table 5. Results of tests for titers against 5 strains of the leptospirosis bacteria (LeI;!tosI;!'ira sp. ) from Middle Park pronghorns trapped 16 December 1988. Animal 1.D. I Sex B14 B24 B25 B26 B46 Yl Y10 Y12 Y2l Y42 F F F F F F F F F F Age 3+ 3+ 3+ 3+ 3+ 3+ 3+ 3+ 3+ 3+ Hardjo Ictero Neg. Neg. Neg. Neg. Neg. Neg. Neg. Neg. Neg. Neg. Neg. Neg. Neg. Neg. Neg. Neg. Neg. Neg. Neg. Neg. Strain of Le2tos~ira Ganico Grippo Neg. Neg. Neg. Neg. Neg. Neg. Neg. Neg. Neg. Neg. Neg. Neg. Neg. Neg. Neg. Neg. Neg. Neg. Neg. Neg. Pomona Neg. Neg. Neg. Neg. Neg. Neg. Neg. Neg. Neg. Neg. 1 Ear tag, B-blue and Y-yellow. Table 6. Seasonal areas (km~ of habitation for 9 radio collared pronghorns in Middle Park for December 1986 through September 1988. Winter-DJF, spring-MAM, summer-JJA, autumn-SON. Season Animal Winter Spring Summer Autumn Total Yearlong 148.39 148.40 148.42 148.43 148.44 149.21 149.25 149.29 149.37 16.29 19.10 12.80 18.64 19.14 14.61 18.64 20.40 9.66 67.76 41.50 31.42 24.32 32.73 56.28 83.24 76.64 21.65 16.45 12.41 38.86 38.86 14.25 26.96 11.30 37.86 n.d.3 18.44 24.02 l.38 3.08 13.79 n.d. 3 31.74 50.98 n.d. 3 118.94 97.03 84.47 84.91 79.91 97.84 144.91 185.89 31.31 80.29 54.22 32.17 8l.28 6l.41 88.55 129.04 113.84 27.66 Ave.4 S.D. 5 G.V. 6 17.86 2.54 14.22% 51.09 24.14 47.25% 24.28 13.42 55.27% 20.49 17.25 84.19% 113.72 39.40 34.65% 78.89 33.77 42.81% 1 Total of all seasons, there is some overlap between seasons which results in this total being larger than the yearlong area. 2 This area represents the minimum convex polygon of all points on a yearlong basis. 3 n.d.- no da ta . 4 The average does not include data for 149.21 and 149.37. 5 Standard deviations calculations do not include data for 149.21 and 149.37. 6 G .V .- coeffic ient of variation. �- ., " . \ ;~;-,-- -~~._:'.- -'-,.!'. Figure 1. Winter area of habitation of 7 radioed pronghorns during the period Dec 1986-Sep 1988. Triangles represent pronghorn locations for Dec, Jan, and Feb; and outlined areas are critical winter range for mule deer as defined by Tiedeman et a1. (1987). �p.ro 0 Ii ~ '-< III ~() ..." ::s III Ii H') 0 Ii ::s 0 p. (I) P. ..." 0 III ::s ::s c III Ii ~ro H')()Q 0 Ii rt ::r' ::s ti•...' (/l ::s 0 ..." P. rt ~O 0 ti 0 III (/l ::s~ ::r'ti 0 0 ti ::s ()Q"d "d Ii 0 rt ::s ro --..J (/l ro Ii "d Ii ro Ii (/l ro III OQ rr ~..." ::s ..." ..." III rr ::r' 01 III Ii 0- \Dill 00 000 H') ~(I) "d ro VlOQ I 0\ ~Vl \D"d 00 Ii ro P.ON ::s III . ,..., I-rj »"d "d ro ..." Ii liOQ ~.loI.UVl •• I -I 'J'\ I ~ ...",. ~. ,... : """""'-1 ~·':':"I Ii:'.' I (lOOt ....K~•.• :. "'I\llI''' ...•.• " .1 '! ~: II 1 •• .\ II U .. 1.,.."'IJ. /' N'":. " I':.~ u. .j ., •" .~. ,r I ..I I -- , ... l' .....-_.~I.:.;:,;-~., r'! i ., I I .~c.• o: - I t-' �L -,~ Figure 3. Summer area of habitation for 7 radioed pronghorns during the period Dec 1986-Sep 1988. Triangles represent pronghorn locations for Jun, Jul, and Aug. �180 Figure 4. Autumn area of habitation Triangles period Dec 1986-Sep 1988. Oct, and Nov. for 7 radioed pronghorns during the represent pronghorn locations for Sep, �250 i. ql WINTER SUMMER k <I AUTUMN 200 150 100 50 o 1.••. 31 1.•8..•.• 1"8.42 1.• 8 ..• 0 148.43 141.21 141.25 AVE. ANIMAL 1.0. Figure 5. Middle Park pronghorns seasonal area of habitation for Dec 1986Sep 1988. Winter-DJF, spring-MAM, summer-JJA, and autumn-SON. One square mile - 2.59 square kilometers. �140 120 100 s ~ 0 en 80 Z « w a: « ...J « I- 10 0 I- 40 20 o 148.3' 148.42 148.40 148.44 148.43 141.2' 141.25 AVERAGE ANIMAL 1.0. Figure 6. Yearlong during Dec 1986-Sep area of habitation 1988. of 7 radioed pronghorns in Middle Park �183 Colorado Division Wildlife Research July 1989 of Wildlife Report JOB FINAL REPORT State of __~C~o~l~o~r~a~d~o __ Project No. 01-03-048 (W-1S3-R-2) Mammals 2 Research Work Plan No. 3A Pronghorn Job No. 1 Pronghorn recruitment and survival in Escarpment data analysis unit. Period covered: Author: Investigations July 1, 1988 - June 30, 1989 T.M. Pojar ABSTRACT This was an exploratory project to make a preliminary assessment of this pronghorn (Antilocapra americana) population's reproduction. A ground survey by vehicle was made to estimate the herd structure. The survey was completed during 29-30 August, 1988 using 6, 2-person crews. A total of 2,374 miles of route were driven and 70S pronghorns were classified yielding bucks:100does:fawns ratios of 61:100:66. The 90% confidence intervals were B:100D ± 34% and F:100D ± 17%. From this meager information it was concluded that an extensive fawn mortality study is not merrited at this time. It is suggested that general popUlation mortality factors and/or movements of segments of the population would provide a more fertile area of investigation. To accomplish this, more reliable estimates of population parameters must be collected by using more precise census methods. Radio telemetry methodology should be used to determine animal movements annually and by season. ��185 PRONGHORN RECRUITMENT AND SURVIVAL IN ESCARPMENT DATA ANALYSIS UNIT Thomas M. Pojar INTRODUCTION This project was initiated as a preliminary assessment of the reproductive performance of the Escarpment DAU (data analysis unit) pronghorn population. This is a large (4300 square miles) and diverse area (Figure 1) with land use practices that include irrigated agriculture in the southern part of the area, dryland wheat farming in the central and east, and domestic livestock grazing of native shortgrass prairie in the north and west portions of the area. ACKNOWLEDGEMENTS The following persons participated in the herd structure survey: M. Ball (U.S. Forest Service), and the following from the Colorado Division of Wildlife, C. Crawford, D. Crawford, L. Crooks, T. Davis, J. Dennis, K. Dillinger, A. Duvall, J. Rogstad, L. Ragstad, G. Schoonveld, S. Steinert, B. Smith, and J. Wagner. METHODS AND MATERIALS The initial effort was to get an estimate of the herd structure and get an indication of how it may vary across the entire area. To accomplish this, the area was stratified into high, medium, and low pronghorn density based on the best judgement of management personnel. These strata were then each arbitrarily subdivided into 3 "blocks" each, that were roughly equal in size, to ensure distribution of the sample units (see Figure 1). Given the amount of time and number of personnel available, it was determined that it would be possible to examine a 10% sample of the total area. To ensure the most efficient use of time the total sample was allocated to each stratum based on the optim~ allocation formula in Cochran (1963:97, Sampling Techniques, John Wiley & Sons Inc. New York 413 pp). The weights assigned were: High density, 3; medium density, 2; and low density, 1. The sample unit was a square mile and the mode of search was from a vehicle using binoculars and a 20+ power spotting scope. Each crew consisted of a driver and a spotter and the method of search was to drive from sample unit to sample unit by the shortest route and classify all pronghorns seen both enroute and on each sample unit. The number of miles traveled by each crew was recorded to get a general index of relative pronghorn density between strata. �186 RESULTS Stratification and optimum allocation of the total sample resulted following distribution of the sampling effort. Strata Total Area No. Samples % Sampled High Medium Low 1710 1109 1488 252 116 81 14.7 10.4 5.4 Totals 4307 449 10.4 Relative density as follows. Strata as derived Miles from number Driven of pronghorns No. Pronghorns in the seen per mile driven is Pronghorns/Mile High Medium Low 1135 556 683 515 109 81 .45/mile .19/mile .l1/mile Totals 2374 705 .30/mile The classification survey was completed during 29-30 August 1988 with the results for the entire Escarpment DAU being 61B:100D ± 34% and 66F:100D ± 17% (Table 1). The confidence limits are calculated at the 90% level. The results by strata are included in Table 1. The ratios for medium and low density are unreliable because of the small sample. prepared~1t !2(iw omas M. POJa Wildlife Researcher �187 Table 1. Herd structure for Escarpment pronghorn data analysis unit obtained from vehicle ground routes during Aug 29-30, 1988. Areal Sample Size Bucks: 100Does 90% C.L.2 Fawns: 100Does 90% C.L.2 Low 81 60 87% 60 37% Medium 109 35 100% 34 83% High 515 69 49% 76 13% Total 705 61 34% 66 17% 1 The total area of 4,247 square miles was partitioned into low density (1488 mi2) , medium density (1109 mi2), and high density(16s0 mi2). 2 The 90% confidence limits are in terms of plus or minus the given percentage. �I--' co oo E S()-O>1p/t"S / 13 f-\ K ",t. 1I"'t'~) Figure 1. Escarpment pronghorn DAU stratified and blocked for purposes of herd structure sampling. numbers in parentheses are the total nwnber of sample units (square miles) drawn at random the total number of potential sample units in the area (no. sample units/total no. units). The over �189 Colorado Division Wildlife Research July 1989 of Wildlife Report JOB PROGRESS REPORT State of Project Colorado No. ~Y_-~1~5~3~-R~-3~ York Plan No. Job. No. 1 Period Covered: Author: Personnel: ~6~A~ _ Mammals _ Mountain Lion Investigations Mountain Dynamics Lion Population Research July 1, 1988 - June 30, 1989 A. E. Anderson See acknowledgments ABSTRACT The results of preliminary data analyses are briefly described. A final, comprehensive, peer-reviewed report should be completed by November, 1990. ��191 MOUNTAIN LION POPULATION DYNAMICS Allen E. Anderson P. N. OBJECTIVE To assess the effects of sport hunting on mountain lion populations. SEGMENT OBJECTIVE To prepare final reports for publication. ACKNOWLEDGMENTS All statistical analyses were performed by Dr. D. C. Bowden, Statistics, Colorado State University, Fort Collins. METHODS Department of AND MATERIALS General methodology was described in Anderson (1983a). Aerial telemetry and all other field work were terminated on July 14, 1989. Checking and assembling data, planning data analyses and preliminary examination of home range dynamics and survival rates of radio-collared puma occupied much of the segment. D. Bowden conducted simulation studies to validate expressions relative to the direct application of nonparametric tolerance (NPT) region results and the attachment of a utilization statement (U). U is defined as the proportion of an animal's time during the time period of interest that the animal spends within the minimum convex polygon area. I investigated the feasibility of using the methodology of Marcum Loftsgaarden (1980) to quantify habitat preferences of puma. and RESULTS AND DISCUSSION Segregation of Data The radio-collared puma were segregated into 2 groups for analysis. First, based on both the scanty literature and our grossly approximate age assignments (Anderson 1983b), I set 24 months as the probable minimum age of sexual maturity in wild puma. Also, because spatial dynamics over the annual cycle were of interest, I included only those presumably sexually active puma whose length of telemetric surveillance was ~ 12 months. The resultant number of aerial locations from those "resident" puma available for analysis by year and period for 7 males (Table 1) and 10 females (Table 2) totaled 2,317. Second, I combined both resident and nonresident puma (Table 3) whose aerial �192 locations and other observations suggested: (a) interactions between motheryoung and litter-mates and (b) dispersion in young puma of each sex. Some hypotheses which could be tested include the following: (1) The generally larger home range area of male resident puma begins during the postnatal period. (2) The core area of dependent males is at a greater distance from the core area of the mother than that of dependent females. (3) Dispersing subadult males travel at greater rates than dispersing subadult females. Dynamics of Home Range Size and Survival Rates Preliminary analyses of the dynamics of home range size of resident male and female puma is given in Appendix 1. Survival rates estimated for male and female puma from about 2 to 84 months of age are in Appendix 2. Simulation Results The simulation results indicated that the expressions were valid for the direct application of NPT and to provide a lower bound for U. In fact, the expressions were also applicable to similar applications for mule deer (Bowden and Kufeld, unpubl. ms.). Habitat Preferences of Puma It is now uncertain whether the job of estimating habitat preferences of resident puma can be completed. The time-consuming feature of the Marcum and Loftsgaarden (1980) method is the very large number of random locations required (5 to 10 times greater) than the number of actual locations. Manuscripts in Press and Publication Goals Two additional reports were completed and publication likely during the 198990 fiscal year (Anderson and Tully 1989, Anderson 1989). My intent is to complete a comprehensive, peer-reviewed publication prior to my retirement during November, 1990. LITERATURE CITED Anderson, A. E. 1983a. Program Narrative Proj. 45-01-503-15050, Work Plan 6, Job 1. Mountain lion population dynamics. 7pp. (+3 tables and Appdendix A). _____ 1983b. A critical review of literature on puma (Felis conco1or). Colo. Div. Wildl. Spec. Rep. 54. 9lpp. 1989. Sexing and aging mountain lion. (draft submitted.) Colo. Div. Wildl. S.O.P. , and R. J. Tully. 1989. Status report - Colorado. Proceedings Third National Mountain Lion Workshop, Prescott, AZ. (in press). �193 Anderson, A. E., D. C. Bowden, and D. M. Kattner. 1989. Dynamics of home range size of unhunted mountain lion (Felis conco1or hippo1estes) in southwestern Colorado. (abstract). Proceedings Third National Mountain Lion Workshop, Prescott, AZ. (in press). Anderson, A. E., D. C. Bowden, and D. M. Kattner. 1989. Survival in an unhunted mountain lion (Felis conco1or hippolestes) population in southwestern Colorado. Proceedings Third National Mountain Lion Workshop, Prescott, AZ. (in press). Marcum, G. L., and D. D. Loftsgaarden. 1980. A nonmapping technique for studying habitat preferences. J. Wildl. Manage. 44:963-968. Prepared by e ~~ ~ Allen E. Anderson Wildlife Researcher �194 Table 1. Numbers of aerial locations of 7 resident male puma radio tracked >12 months at aEEroximate weeklx intervals, 1983-88. Year 83 84 85 86 87 88 Pe r Lod" Period Period Period Period Period 1 2 1 2 1 2 1 2 1 2 1 2 All No. N N N N N N N N N N N N N 5 18 20 22 40 44 52 12 23 24 27 25 23 23 All 12 24 24 22 4 26 47 52 20 25 22 25 9 5 23 13 15 24 18 20 22 23 19 15 3 18 22 21 22 22 21 24 25 23 26 9 8 8 7 9 255 89 103 126 73 67 60 51 75 58 107 127 144 50 773 Mos. monitored 65.0 25.0 30.6 42.5 18.0 17.0 15.0 aperiod 1 - Large prey species mainly on winter and transitional ranges (Nov. 16 - May 15). Period 2 - Large prey species mainly on summer and transitional ranges (May 16 - Nov. 15). �195 Table 2. Numbers of aerial locations of 10 resident female puma radiotracked ~12 months at approximate weeklI intervals, 1982-88. Year 82 Perioda 1 2 N N No. 4 6 7 12 15 21 28 32 34 45 6 All 6 21 21 84 83 Period 1 2 N N 20 12 10 8 50 25 24 25 22 96 85 86 Period 1 2 N N Period 1 2 N N Period 1 2 N N 21 22 25 22 17 25 27 26 25 25 23 24 24 24 18 21 23 18 22 20 23 23 27 107 115 26 26 26 24 23 129 125 20 20 20 19 20 14 20 174 134 87 Period 88 Period 1 N 2 N 1 N 2 N 25 25 24 25 22 22 22 22 14 25 23 25 8 9 9 2S 22 1 25 22 1 22 15 22 167 175 25 9 25 25 23 184 9 9 8 61 All N Mos. monitored 270 255 257 248 110 122 38 119 57 68 1 544 85.2 65.0 64.2 63.7 27.0 42.S 12.0 30.1 13.0 17.7 aperiod 1 - Large prey species mainly on winter and transitional rang•• (Nov. 16 - May 15). Large prey species mainly on summer and transitional ranges (May 16 - Nov. 15). Period 2 - Table 3. Twelve male 9 female resident and nonresident puma whose aerial locations and inferred familial relationshiEs lend themselves to testins hz:egthesesrelative to interaction and disEersion. Approx. date No. mos. monitored when ~24 mos. within ase classes Age (mos.) Dates of locations attained Sex <24 mos. >24 mos.a No. at caEture First Last Male Female 8 13 14 23 26 35 38 42 47 50 55 17 6 10 84 12 10 6 36 12 9 13 30 8 2-25-83 4-15-83 12-16-83 1-14-85 3-22-85 2-17-86 12-19-86 1-16-87 1-30-87 2-20-87 2-12-88 4-20-84 8- 4-83 2- 3-84 6-22-84 11-22-85 2- 6-87 10-10-86 5-15-87 4-17-87 7- 8-88 2-12-88 7-14-88 7- 6-84 5.5 10.0 0.0 10.2 13.0 7.7 0.0 3.0 14.0 10.0 0.0 2.5 0.0 0.0 5.0 0.0 9.5 0.0 5.0 0.0 3.2 1.7 5.0 0.0 3 27 46 53 54 56 57 33 26 14 60 48 14 14 26 7 6 1-11-82 4-27-85 1-24-87 12-21-87 1-15-88 3- 4-88 4-15-88 2-10-86 2-17-86 4- 8-83 11-22-85 8- 5-87 7-14-88 7-14-88 7-18-88 7-14-88 11- 6-86 8- 9-87 0.0 6.7 0.0 0.0 6.0 4.0 0.0 9.0 17.0 5.0 0.0 6.5 6.7 0.0 0.0 3.0 0.0 1.7 36 a~24 mos. may b. the minimum age of reproductive activity in wild puma (Anderson 1983). 4-25-86 3-30-88 12-21-87 6- 5-87 �196 Appendix 1 (Abstract) DYNAMICS OF HOME RANGE SIZE OF UNHUNTED MOUNTAIN HIPPOLESTES) IN SOUTHWESTERN COLORADOa,b Allen E. Anderson, Colorado Montrose, CO 81401 Division David of Statistics, Donald C. Bowden, Department Collins, CO 80523 M. Kattner, of Wildlife, 63478 Ida Road, Montrose, LION (FELIS CONCOLOR 206 South 5th Street, Colorado CO State University, Fort 81401 Five adult female and 2 adult male mountain lion were radio collared and aerially located at approximate weekly intervals on Uncompahgre Plateau, over periods ranging from 23 to 49 months each. Home range size was estimated for years and periods using the minimum convex polygon (MCP) as an application of the nonparametric tolerance region (NPT) technique and by the harmonic mean (HM). Multiresponse permutation procedures were used to detect temporal differences in their spatial distribution, which were described by dispersion and median location concentration estimators. At the 90% confidence level, the proportion (P) of time 5 female puma spent within MCP regions of size (A, km2) were at least (P,A) (.908, 219.5), (.867, 140.0), (.873, 238.0), (.841, 186.0), and (.875, 128.0). During the period of maximum average female home range size (November 16 - May 15) (P,A) were (.793, 187.5), (.795, 92.0), (.713, 164.0), (.737, 175.5), and (.708, 104.5) at the 90% confidence level. Among female mountain lions, use was most concentrated during the heavy snowfall year of 1985 and during the May 16 - November 15 period. Shifts in location of females were more than twice as large between periods as between years. For the total sample of 519 female locations, median MCP and HM home ranges approximated 186 km2• Home range sizes (MCP) of 2 male mountain lions 2 were 596.5 km and 279.0 km2 who spent at least 0.842 and 0.790, respectively, of their time within those home ranges at the 90% confidence level. aA1though closed to sport hunting of mountain 1ipn in 1982, 11 mountain lions marked on the study area were killed, 1 illegal and 2 livestock depredation kills within the study area, and 1 illegal and 7 legal kills outside the study area. grants bContribution from Colorado from the National Wildlife In press, Federal Aid Project W-53-R assisted by Federation and Pope and Young Club. Proc. Third Nat1. Mountain Lion Workshop, Prescott, AZ, Dec. 1988. �197 Appendix SURVIVAL IN AN UNHUNTED HIPPOLESTES) POPULATION 2 (Abstract) MOUNTAIN LION (FELIS CONCOLOR IN SOUTHWESTERN COLORADO&,b Allen E. Anderson, Colorado Montrose, CO 81401 Division David C. Bowden, Department Collins, CO 80523 of Statistics, Donald M. Kattner, of Wildlife, 206 South 5th Street, Colorado 63478 Ida Road, Montrose, CO State University, Fort 81401 Twenty male and 21 fe~ale mountain lion were captured on a 3,263-km2 portion of the Uncompahgre Plateau and fitted with transmitters equipped with mortality switches, 1981-88. Gross approximations of their ages at capture based on dental and physical characteristics ranged from 2 to 84 months. Animals were monitored with aerial telemetry at approximate weekly intervals over periods ranging from 0.2 to 79 months per individual. The survival of each sex was examined separately by the life regression procedure (LIFEREG) in SAS which allows for censored observations (animals alive at the end of the study). Tests of goodness of fit for each sex between the Weibul1 and exponential models indicated that the Weibull model did not provide a significant improvement of fit over the exponential model. Sexes were combined because a statistical comparison of male and female data suggested similarity (Z - 1.006, P - 0.32) assuming an exponential model and mu1tiresponse permutation procedures (a nonparametric test) indicated that ages at mortality of each sex were similar. A Q-Q plot did not reveal any serious departure of the data from an exponential model. Thus, 25%, 50%, 75%, 90%, and 95% of male and female mountain lions have died by ± SE) 27.35 ± 6.27, 65.90 ± 15.12, 131.80 ± 30.24, 218.92 ± 50.22, and 284.82 ± 65.34 months of age, respectively. Fifty percent of the mountain lions died by 65.9 months of age with 95% confidence limits of 42.0 - 103.3 months. The annual survival rate was 88.1% with approximate 95% confidence limits of 82.1 - 92.3%. (x &Although closed to sport hunting of mountain lion in 1982, 11 mountain lions marked on the study area were killed, 1 illegal and 2 livestock depredation kills within the study area, and 1 illegal and 7 legal kills outside the study area. bContribution from Colorado Federal Aid Project W-53-R assisted by grants from the National Wildlife Federation and Pope and Young Club. In press, Proc. Third Nat1. Mountain Lion Workshop, Prescott, AZ, Dec. 1988. ��199 Colorado Division Wildlife Research July 1989 of Wildlife Report JOB PROGRESS REPORT State of ~C~o~l~o~r~a~d~o~ _ Project _ Mammals _ Small Carnivorous No. ~W_-~1~5~3~-R~-~3~ Work Plan No. Job No. Period __~8~A~ I Covered: Author: Personnel: July Research Mammal Investigations Development of River Otter Reintroduction Procedures 1, 1988 - June 30, 1989 T. D. I. Beck T. Benjamin, L. Malville, R. B. Gill ABSTRACT Seven river otters (Lutra canadensis) were received from Oregon Fish & Wildlife Dept. Six were released into the Dolores River and 5 survived. Radiotelemetry indicates winter ranges varied from 7-12 mi in length. All river otters nearly doubled range use in May-June by taking long excursions. One male disappeared in April, presumably downriver. A draft river otter recovery plan was submitted to the state office, with appendices covering the history of prior releases. A general survey of crayfish (Orconectes virilis group) was conducted over 99 mi of river. Catch per trap night was significantly different between 4 sections of the river. The upper 44 mi appear to have relatively high densities of crayfish. Turbidity was measured biweekly at 5 sites between RM 167 and RM 102 for 1 yr. Water temperature was recorded ··continuously for 1 yr at RM 127.3, with a 10-week break in service in mid-winter because of operator error. General habitat mapping of river ·bed components, bank components, and bank vegetation was completed at each 0.1 mi from RM 166.7 to RM 69.7. ��201 DEVELOPMENT OF RIVER OTTER REINTRODUCTION PROCEDURES Thomas D. I. Beck P. N. OBJECTIVE Develop procedures for river otter reintroductions in Colorado and establish a self-sustaining population of river otters from which to collect river otters for future translocations. SEGMENT OBJECTIVES 1. Introduce up to 20 river otters into the Dolores River drainage. 2. Monitor all river otter release past reintroductions. sites in Colorado to evaluate 3. Develop techniques to monitor survival, reproduction, dispersal of river otters after reintroduction. METHODS River Otter Recovery success dispersion, of and AND MATERIALS Plan A thorough review of North American and European literature was conducted, informal meetings with regional staff were held, and all available records from earlier Colorado releases were collected prior to preparing a draft recovery plan. Appendices were attached detailing the history and status of the 4 prior release sites: Cheesman Reservoir, Gunnison River, Piedra River, and N. Fk. Colorado (RMNP). Dolores River Release River otters were captured in Oregon and Alaska and kept at the Portland Zoo for periods of 3-26 days. All river otters were shipped on one flight conducted by Bob Allen of Project Lighthawk. The intent was to use a private plane to enable us to fly direct and avoid long layovers on commercial flights. Weather did not cooperate and a most circuitous route from Portland to Cortez was found. Upon arrival the animals were taken to Cortez Animal Clinic where David Herrick, D.V.M., conducted the surgery to place the intraperitoneal radio transmitters in place. Surgical procedures basically followed those outlined by Hoover (1984). The radio transmitters, Telonics Model 400, weighed 110 gms and measured 9.5 em in length and 3.2 em in diameter. Gas sterilization techniques were used on the transmitters at least 10 days prior to surgery. Initial immobilization was accomplished with intramuscular injections of ketamine (100 mg/ml) and xylazine (20 mg/ml) at a dose rate of 15 mg ketarnine per kg of animal. Halothane gas was used to maintain narcosis until surgery was complete. Following surgery each river otter was tagged with a Monel #3 metal tag in the web between 2 front toes and measured for total length; tail length; neck, chest and head circumference; and weight. All injuries were recorded and white blood cell count, hematocrit, and a fecal smear for parasites were checked. �202 Radio tracking of the released river otters was conducted from ground, canoe, and aerial searches, dependent on season, and occurred at least once each 5 days for each individual. Because of the steep canyon terrain, triangulation was not feasible so all signals were tracked to the source; either the animal or more usually a beaver (Castor canadensis) den. Locations were recorded by river mile to an accuracy of 0.1 mi. Prey Base Studies An impression from the 1987 fish surveys was that crayfish (Orconectes virilis group) were surprisingly abundant in some stretches of the Dolores River. A relative abundance survey was conducted between 14 and 29 July 1988. Large minnow traps (78 X 23 cm) with funnels at both ends were baited with 0.5 kg of Purina Crawfish Bait. Sample sites were at 1.0 mi intervals beginning at RM 167.7 and ending at RM 69.7. A single trap was set at each site and left for 24 hrs. All crayfish were sexed and carapace length was measured. A measuring board was built by gluing a 8 rom diameter dowel into a 5 X 20 cm board and gluing a clear plastic rule IS cm long against the dowel. The rostrum of a crayfish was placed against the dowel and a vertical sight taken at the end of the carapace. Carapace length was recorded to the nearest mm. The traditional way for measuring crayfish is to use calipers and record to nearest 0.1 rom and then report all data to nearest rom. The measuring board was much faster and proved sufficiently accurate. The statistic for comparing relative abundance was total number of crayfish per trap day. Habitat Studies River flows in cubic feet per second (cfs) were obtained below McPhee Dam (RM 179.5) from the Bureau of Reclamation, USDI. Water temperature was recorded at RM 127.3 with a Ryan Automatic Recorder and was tabulated at 2 hr intervals for each day. Turbidity was measured at 5 sites at biweekly intervals for 1 yr. Turbidity was measured in Formazin Turbidity Units (FTU) with a Hach Portable Colorimeter DREL/lC. Sample sites were at RM 166.8, 147.4, 124.3, 118.6, and 102.4. A general habitat survey was conducted at 0.1 mi intervals between RM 166.7 and RM 69.7. The survey consisted of 3 parts: river bed components, bank components, and bank vegetation. River bed components were: 1) primary habitat 2) secondary habitat 3) substrate 4) maximum stream depth 5) stream width 6) boulders 7) islands 8) bars. Primary habitats were main channel, chute channel, side channel, and tributary with only 1 type per site. Secondary habitats were run, pool, riffle, shoreline eddy, obstacle eddy, rapid, rubble flats, and backwater; and as many as 3 types might be present per site. Substrate categories were silt, sand, gravel, cobble, boulder, and bedrock with the 2 major substrates being recorded. Width was recorded as <11 m, 11-30 m, and >30 m. Boulder categories were none, large and inaccessible to river otters, and large and accessible to river otters. Island categories were none, head of island, main body of island, and tail of island. Bar categories were none, cobble, sand, silt, and boulder. Bank components were 1) bank configuration at high and low water 2) bank height at high and low water 3) bank stability 4) bank cover 5) logs 6) �203 beaver activity. Bank configuration categories were sandbar, cobblebar, <30 degrees, 30-60 degrees, >60 degrees, rock cliff, and sand cliff. Bank height was recorded as <1 m, 1-2 m, 2-3 m, and >3 m. Bank stability was rated as stable «10% sloughing), moderate (10-30 % sloughing), and unstable (>30% sloughing). Bank cover was rated as sparse, moderate, or dense. The categories were based on the probability of seeing a river otter at 5 m distance. Sparse cover would always allow one to see a river otter at 5 m, moderate cover could possibly hide the river otter but most likely not, while dense cover would always hide the river otter at 5 m. Log categories were none, single from bank to water, 1-5 logs from bank to water, debris pile, and submerged log island. Beaver activity was categorized as none, active slides, active slides and dens, inactive slides, and inactive slides and dens. Bank vegetation was inventoried following the procedures described by O'Brien and Van Hooser (1983). The plot was 30 m along the bank to a distance 10 m back from the bank. By using the plant cover categories of 1-5%, 6-25%, 2650%, 51-75%, 76-95%, and >95% the dominant vegetation could be recorded from a canoe in mid-river. Part I lists dominant species by group--trees, shrubs, forbs, and graminoids-- with no more than 4 species by group. Canopy cover must exceed 5% for a plant species to be recorded in this part. The vertical layer of each species was also recorded as Layer 1 «0.5 m), Layer 2 (0.5-2.0 m), and Layer 3 (>2.0 m). Part II provides a cover class rating for the entire plant group within each layer and includes plants not abundant enough to show up in Part I. However, there were no species specific values given in Part II. Bank vegetation was described from both banks at each site. RESULTS AND DISCUSSION River Otter Recovery Plan A draft river otter recovery plan for Colorado was submitted to the state office in November 1988. The plan included a brief historical summary, statewide objectives, and a 7-step program to achieve recovery of the species within the state. Four appendices were included covering the history of the earlier river otter releases at Cheesman Reservoir, Gunnison River, Piedra River, and the N. Fk. Colorado (RMNP). Dolores River Release Seven river otter arrived at Cortez at 2130 on 2 November 1988 (Table 1). Surgical implantation of radio transmitters, tagging, and measuring were conducted at the Cortez Animal Clinic with the last animal completed at 0330 on 3 Nov. Six animals were released at RM 139.7 on the Dolores River at approximately 0800. RO-6 never recovered from anesthesia. He was extremely lethargic when removed from the shipping crate prior to surgery; a noted contrast to the rather excited nature of the adult river otters. His WBC count and hematocrit were normal. Perhaps the ubiquitous demon of stress was the cause of death. Five of the river otters swam off quite vigorously upon release but an adult female, RO-2, stayed in the shipping crate for over an hour before moving off. She had very restricted movements and died on 5 November. She was a very sick animal upon arrival, suffering from foot injuries which included broken bones and chewing, and vaginitis. Her WBC was 29,000 and hematocrit was 32%. She was treated with 150 ml half-strength �204 lactated Ringer's solution and dextrose as well as 500 mg Kefzol. It was decided to release her rather than maintain her in captivity since many of her wounds appeared to be self-inflicted, possibly a response to captivity. Table 1. River otters released ID Length (cm} W'T Total Tail in the Dolores River, CO. Circumference (cm} Head Neck Chest Tag 1988 Age Sex RO-1 AD M 7.3 127 47 25 24.5 34.5 1 148.02 Cordova, AK RO-2 AD F 5.8 123.5 43 24 27 33 2 148.03 Cordova, AK RO-3 SAD M 5.4 woke up too soon 3 148.04 Laketown, RO-4 AD M 6.0 woke up too soon 4 148.05 Cordova, RO-5 AD F 8.0 116 40 25 28 40 5 148.06 Laketown, OR RO-6 JUV M 3.6 95 35 22 20 27.5 6 148.07 Laketown, OR RO-7 AD M 7.3 109 41 24 27 37 7 148.08 Laketown, OR # RT Freq Origin OR AK Of the surviving 5 river otters, the 2 Alaskan males have been quite difficult to keep up with. In the 30 day period post-release RO-l was only located 5 times and had moved 10 mi downstream while RO-4 was also located 5 times and also moved 5-10 mi downstream. RO-3 decided to try an overland route from the release site and climbed up the canyon wall until becoming ledged out at the base of a vertical wall of sandstone approximately 900 ft above the river. We packed food and water to him for 2 days in attempts to lead him down. Finally on 8 Nov. we captured him on the ledge and packed him down to the river in a shipping crate. He remained the remainder of the month within 5 mi of the release site and upstream. RO-5 and RO-7 spent most of the first month together in the area near the release site and up to 4 mi downstream. The reasons for the difficulty in locating the 2 Alaskan males are unknown but it is suspected they make frequent long movements and do not show a strong affinity for one area. Thus if one checks for them in a 10-mi stretch one day and then moves downriver the otters likely moved upriver through the 10-mi stretch overnight. Whatever the reasons, they have been difficult to follow. RO-1 was only located 24 times in 8 months and ranged from RM 124 to RM 143. He is still within the study area. RO-4 was only located 27 times in 6 months, ranging from RM 128 to RM 158. He has not been located since 20 April although extensive boat and aerial searches have been conducted of the entire Dolores River and much of the San Miguel River. A sighting of a river otter by a commercial rafter was reported in mid-May from RM 80 but the same week a search of RM 102 to RM 70 did not find sign or a radio signal. �205 RO-3 was located 108 times during the 8 months, ranging from RM 140 to RM 156. However, 103 of the locations were between RM 140 and RM 147.5. The 5 locations outside this area were in May and June and he went upriver to RM 155-156. RO-7 was located 93 times during the 8 months, ranging from RM 132 to RM 156 with over 80% of the locations being between RM 137 and RM 147. There were no locations from RM 148 to RM 155 and he was located in RM 155 several times during May and June. The female, RO-5, was located 57 times, ranging between RM 124 and RM 156. Fifty of the locations were in the stretch between RM 132 and RM 143. Her movements downstream of this core stretch occurred in May and the upstream movements occurred in June. The river otters were released approximately 3 weeks before most of the Dolores River iced up so it was not surprising to see relatively limited movements throughout the winter. In reviewing the radio telemetry data from the N. Fk. Colorado for the recovery plan it was found that the river otters there made extensive movements, usually downstream to the big lakes, in May and June. As spring flooding occurs during this period it was thought that the expanded movements were a response to that. Now on the Dolores River we see greatly expanded movements again in May and June. However, high water in 1989 was from 1 April to 15 May and then the flow was reduced to the base summer level of 78 cfs. All the long movements occurred after the drop in water. Expanded effort will be made in 1990 to closely follow all river otters in May and June to better document this pattern. All introduced river otters used beaver bank dens extensively. In winter it was not unusual for river otters to exit a beaver den under the ice, forage, then return to the den without ever surfacing. River otters could be using a pool for several days and never leave any surface evidence of their presence. Extensive use of beaver dens was even more noticeable during the hot summer months when most river otter activity was at night and the entire day was spent in the dens. Temperatures in bank dens have been shown to be nearly constant during the day and significantly cooler than ambient temperature during the summer and constant but warmer than ambient in the winter ( Buech et al. 1989). No attempt was made to assess food habits from scat collected and such activity will not be attempted until after the controlled feeding experiment. This experiment could not be conducted in 1988 because the river otters were received so late in the year. Our impression is that crayfish are utilized extensively in all months of the year. The only surprise is that crayfish hibernate under rocks throughout the winter and the river otters have shown expertise at digging them out. The most common fish remains found were heads from large channel catfish (Ictalurus punctatus). It is not known if this is because of selective predation on the catfish or because of the massive bony catfish head which cannot be eaten by river otters. Prey Base Studies First a note on the taxonomic status of the crayfish in the Dolores River. is not known if any species of crayfish was native to the Dolores River It �206 drainage and large scale trans locations of crayfish into the basin have occurred. The Colorado Div. of Wildlife placed crayfish, species unreported, into numerous reservoirs in the basin in 1956. Sampling of these reservoirs in 1988 found the species present was in the Orconectes virilis group. Crayfish were reported in the Dolores River during fish surveys in 1977 but none were collected. Additionally crayfish, ~ virilis group, were transplanted into McPhee Reservoir on the main stern of the Dolores River, in the mid-1980's by Colorado Div. of Wildlife. Not surprising then that our surveys of the river below McPhee Darn found ~ virilis group. This group is composed of 3 species which are morphologically nearly indistinguishable: ~ virilis, ~ nais, ~ causeyi (Fitzpatrick 1987). Discussions with H. H. Hobbs, Jr., of the Smithsonian Institute, resulted in an identification of "probable" ~ causeyi because of the geographic location. It was my impression that Dr. Hobbs actually thinks that all the beasts are just normal variants of ~ virilis. Because of the many translocations of crayfish with poor record keeping it is unlikely if this taxonomic problem will ever be resolved. Until resolution, the decapod in question will be called ~ virilis group in our reports. Number of crayfish caught per trap night was plotted against river mile and 4 distinct sections of the river were apparent. Section 1 was from RM 167.7 to RM 146.7 (Bradfield Bridge to Dove Creek Pump Stn.), section 2 was from RM 145.7 to RM 124.7 (Dove Creek Pump Stn. to Disappointment Creek), section 3 was from RM 123.7 to RM 95.7 (Disappointment Cr. to 7 mi below Little Gypsum Valley Bridge), and section 4 was from RM 94.7 to RM 69.7 (ending at BLM ramp upstream from Bedrock Bridge). Plotting of the data suggested non-normal, negative binomial distributions (clumping) so the Mann-Whitney U Test was used to compare the similarity of medians and distributions. In comparing the catch per trap day data (Table 2), the null hypothesis was that Section X has the same distribution and median value as Section Y. Section 1 was compared to Sections 2,3, and 4 with Ho rejected in all cases (p-0.05, 0.01, 0.001 respectively). Section 2 was compared to Sections 3 and 4 with Ho rejected both times (p-0.001 in both) and Section 3 was compared to Section 4 with Ho rejected again (p-0.001). Table 2. Crayfish caught per trap day among sections of Dolores River, 1988. Std. Error Section No. Sample Sites Median Range Mean 1 22 34.5 10-51 31. 8 12.8 2 22 49 1-106 51.2 30.6 3 29 17 1-46 19.7 11.4 4 26 2 0-21 3.3 4.6 Age classes of captured crayfish were estimated from graphic plots of frequency of lengths based on the 2,490 crayfish measured (1645 males, 845 females). Age classes will not be reported yet because I am not confident of �207 the class categories. Our data were so at variance with much of the information in the literature that I want to carefully reexamine our analyses. Such analyses have been hampered by changes in our library staffing that has slowed down procurement of hard-to-find papers. The most perplexing problem is the apparent large percentage of Age Class III in all sections but especially Section I. There are also apparent growth rate differences between sections. Work in 1989 will focus on studies of density, growth, and trap vulnerability of crayfish. What can we say about crayfish abundance in the Dolores River? Unfortunately, simple catch per unit effort data does not tell much about absolute abundance, which is really what is of interest to a river otter. However, some strong trends were apparent. I have no explanation for the differences observed between Sections 1 and 2. The river appears to be similar habitat for crayfish throughout these stretches. More large crayfish were caught in Section I and perhaps the aggressive nature of the larger animals in the trap kept others out. The sharp decline in captures below Disappointment Creek is most likely a reflection of changing habitat caused by the side canyon discharges in spring and late summer. The most noticeable changes below RM 125 are increased turbidity, warmer water temperatures, more sand and silt substrate, and the absence of green algae (Cladophora sp.). The influence of turbidity to crayfish is likely expressed through the consequent reduction in green algae and phytoplankton production. The catch in section 3 exceeded my expectations from prior work on the river. The morning following the flash flood on 31 August 1988, when the river rose 1.3 m vertically overnight and turbidity went up to 9,000 FTU's, crayfish were observed leaving the water and crawling into beaver bank dens. Groups of 5-30 crayfish could be seen in all the beaver dens. The river cleared up to less than 200 FTU's within 48 hrs and presumably the crayfish went back into the river at that time. Our knowledge of crayfish life history in the Dolores River is too scanty to allow the generation of causative explanations for the marked decline in catch in Section 4. The break does not occur immediately downstream of a major side canyon nor does the bank or substrate appear to change much. Another survey will be conducted in 1990 to compare to the 1988 findings. Perhaps that survey combined with the physical habitat survey will provide more clues. Habitat Studies Dolores River flows in 1988 strongly reflected the influences of the McPhee Dam and the 50% snowpack. The Bureau of Reclamation had maintained a base winter flow of 78 cfs until 28 April. Discharge was rapidly increased over 5 days to peak at 1200 cfs for the week of May 3-8 when it was dropped back to 194 cfs. On 18 Maya release of roughly 600-700 cfs was begun and continued through 7 June. By 18 June the base summer flow of 78 cfs was in force and continued throughout the year. In comparing the pattern to what might have been had there been no dam it is apparent that the minimum flows were greater throughout 1988 than would have been expected. The peak flow was close to the time and amount one would have had naturally. The biggest difference was the yo-yo pattern in May. The major impact to aquatic and riparian life was probably the sustained flows of 78 cfs throughout the summer. Water temperatures were continuously recorded at a point 52.3 mi downstream of the dam and 3.5 mi upstream of Disappointment Creek. This point was selected �208 based on the 1987 fish surveys and represents my estimate of the downstream end of trout habitat. The diel pattern for July 1988 showed an average daily minimum of 17.4 C and an average daily maximum of 21.8 C. The minimum temperature occurred at 1000 and the maximum at 1800. An average of 8 hrs per day had water temperatures exceeding 21 C but the mean is somewhat misleading as there was a mixture of 0-3 hr days with 10-12 hr days and few in between. The diel pattern for August was similar but with minima and maxima occurring 2 hrs earlier. Average daily minimum temperature was 16.6 C and average daily maximum temperature was 21.2 C. The average period above 21 C per day was less than 1 hr. September was a period of marked cooling with average extremes of 11.5 C and 15.4 C. An unusual pattern of warming was evident in that the water released from the reservoir varied substantially in temperature during the day and the water in the river rapidly reached maximums of 21 C but then maximum daily temperature did not increase much over the next 47 mi. Data from the Bureau of Reclamation, Durango Projects Office, was only available in graph form and this precluded more detailed comparisons. However, the daily variance in water temperature for the first 2 weeks of July 1988 could be compared at 3 sites: RM 174.6, RM 127.3, RM 116.8. Water temperature at the upper site, 5 mi below the dam, varied from 12-21 C while the site 52.3 mi below the dam varied from 16-22 C. The maximum daily temperatures were quite similar. At the lower site, 62.8 mi below the dam, daily variation was between 20 and 26 C. This pattern of warming likely will have an influence on spatial and temporal distribution of spawning fishes and most likely will displace spawning of some species downstream. Attempts will be made to obtain more complete data sets from BuRec and then to analyze possible impacts to the fishes currently living in the Dolores River. The influence of McPhee Dam was quite evident on the turbidity of the river down to the confluence with Disappointment Creek (Table 3). Tributaries above RM 127 contributed fairly clear water even during summer storms. Flash floods scoured Tree Frog Canyon several times in 1988 as evident by the 3 high readings in August at RM 124.3. The situation below Disappointment Creek is dominated by side canyon discharge with peak turbidity during spring runoff in low elevation drainages and during the August-September monsoon season. The major influence of the ephemeral side canyons is evident in turbidity and flow. Attempts by BLM hydrologists to correlate flows at Dolores (RM 191) with flows at Bedrock (RM 69) have been to no avail because of the large flow increases' from summer storms (D. Murphy, pers. comm.). The 2 largest events occurred on 31 July and 31 August 1988. The Dolores River came up 98 cm vertically during the first and 140 cm during the second at RM 102. These storms had no impact on the upper river. During the 31 Aug. storm no rain actually fell on the Dolores River channel, all came from tributaries. Certain anomalies like downstream stations being clearer on a day result from afternoon melt and the timing of sampling. I usually sampled the downstream areas first. The general habitat survey was completed and all data entered into a dBaseIII file. All data entry has been proofed and errors corrected. Approximately 6 mi of the river had to be duplicated in June 1989 to correct a technician error on a shrub species. No analyses have been conducted with this data base yet. The description of the data collected was provided in this report so that federal cooperators will know what data is available. �209 Table 3. Turbidity at 2-week intervals on Dolores River, CO. 1988-89 Date 7-16-88 8-01-88 8-16-88 9-01-88 9-16-88 9-30-88 10-16-88 10-31-88 11-15-88 12-01-88 12-15-88 1-07-89 1-20-89 2-04-89 2-16-89 3-02-89 3-17-89 4-02-89 4-19-89 5-02-89 5-15-89 6-01-89 6-18-89 7-09-89 RM 166.8 8 0 8 4 8 12 2 2 2 2 2 8 1 2 10 8 10 12 19 8 4 2 1 4 FORMAZIN TURBIDITY UNITS RM 118.6 RM 147.4 RM 124.3 8 18 16 15 19 15 8 8 8 10 8 6 5 8 10 9 25 16 19 20 5 4 6 8 21 615 35 270 200 30 10 12 12 10 10 10 8 15 18 14 95 30 45 22 7 10 10 15 28 1240 40 3600 520 30 12 19 38 13 13 14 24 80 28 500 1300 135 160 25 14 16 12 28 RM 102.4 55 5500 75 9000 1010 35 13 19 275 14 15 18 22 80 26 1250 900 175 575 40 25 24 21 40 LITERATURE CITED Buech, R. B., D. J. Rugg, and N. L. Miller. 1989. Temperature in beaver lodges and bank dens in a near-boreal environment. Can. J. Zool. 67: 1061-1066. Fitzpatrick, J. F., Jr. 1987. The Subgenera of the crawfish genus Orconectes (Decapoda: Cambaridae). Proc. BioI. Soc. Wash. 100(1):44-74. Hoover, J. P. 1984. Surgical implantation of radiotelemetry devices in American river otters. J. Am. Vet. Med. Assoc. 185(11): 1317-1320. O'Brien, R. and D. D. Van Hooser. 1983. Understory vegetation inventory: an efficient procedure. Res. Pap. INT-323, USDA For. Ser. 6 p. Prepared by Thomas D. I. Beck Wildlife Researcher ��211 Colorado Division of Wildlife Wildlife Research Report July 1989 JOB PROGRESS REPORT State of Colorado Project No. _W~-~1~53~-R~3~ _ Mammals Research Work Plan No. ~9~A~ _ Elk Investigations Job No. 1 Impact of Elk Winter Grazing on Livestock Production and Work Plan No. 3~ Job No. 5 _ Period Covered: July 1, 1988 - June 30, 1989 Author: N. T. Hobbs, D. L. Baker Personnel: G. Bear, M. Miller, A. McIntosh, R. Reid, L. Carpenter, B. Gill, B. Petch, C. Woodward, B. Seely, H. Seely, L. Lovett ABSTRACT Elk grazing during winter influenced forage production and cattle performance on sagebrush-grassland range during spring. The magnitude of that influence depended strongly on year for cattle responses, but was largely independent of year for vegetation responses. Elk grazing caused linear declines in weights of cows and calves at the end of the spring grazing season· during years 2 and 3 of our studies. These declines resulted·from reductions in rates of daily gain and from treatment-:-induceq. delays in calf birth dates. During year ~.and 3, birth dates of cal.ves whose . mothers werei-n the high density eik treatment during the previous year occured an average of 4-6 days later than birth dates of calves whose mothers were in controls. During all years we observed linear declines in the standing crop of live and dead perennial grass in relation to elk density. Utilization of dead perennial grass approached 80% in the high density treatment during year 2 and 3. Utilization of live perennial grass reached an asymptote at about 20% for the moderate and high density treatments during all years. Although we occasionally observed a weak stimulation of primary production of grasses and forbs by elk grazing during year 1, the overall influence on.production was negligible. Large differences in precipitation among year interacting with the cumulative effects of cattle grazing were apparently responsible for marked annual differences in forage production and cattle performance. ��213 IMPACTS OF WINTER GRAZING BY ELK ON CATTLE PRODUCTION P. N. OBJECTIVES 1. To test the hypothesis that elk grazing during winter influences productivity and botanical composition of herbage on sagebrush grassland ranges during spring. 2. To test the hypothesis that elk grazing during winter influences the body weights and rates of gain of cows and calves using sagebrush grassland ranges during spring. METHODS the AND MATERIALS Study Area We conducted experiments on the Little Snake Wildlife Management Area in northwestern Colorado (township 9 north, range 95 west, sections 9, 10). The area is about 35 km (19 mi) north of Maybell, Colorado on County Road 19. Although this area does not typically contain high concentrations of elk during winter, it is representative of areas that do have those high densities. Topography of the area includes level ridge tops, rolling hills, and deep gullies, ranging in elevation from 1800 to 2000 m (5900 to'6600 ft). Aspects are southern and southwesterly with an average slope of 15 degrees. Soils are generally sandy and sandy loam. Climate of the area is dry and cold. The growing season averages only 81 days. Annual mean temperature is 6.06 C (42.9 F). Annual precipitation averages 27.5 cm (12.5 in). Vegetation is dominated by big sagebrush (Artemisia tridentata) with an understory predominated by needle and thread (Stipa comata), western wheatgrass (Agropyron smithii), Indian ricegrass (Oryzopis hymenoides), Junegrass (Koleria cristata), and cheatgrass (Bromus tectorum). Important forbs include wallflower (Erysimum asperum), peppergrass (Lepidium perfoliatum), silver lupine (Lupinus argenteus), and scarlet globe mallow (Sphaeralcea coccinea). Experimental Design We observed effects of elk grazing on forage and cattle response in a randomized complete block design with four levels of elk density (0 elk/kmz, 8 elk/kmz, 15 elk/kmz, and 31 elk/kmz) and three replications per level. There were three blocks, each consisting of four pastures. Each pasture within a block was stocked with one level of elk density such that each block contained all levels. During year 1, the twelve available pastures were blocked by pretreatment biomass of perennial grasses with the four lowest grass biomass pastures forming one block, and the four highest grass biomass pastures forming a second block, and the remaining four pastures serving as the third block. The four levels of elk density were randomly assigned to pastures within each block during year 1. �214 Procedures We stocked pastures with elk in December and January 1987-89. Average date of release into pastures was January 3. All elk were removed from pastures during April 18-21. Overwinter mortality approached 15% in the high density pastures, but was nil in the medium and low density ones. Healthy animals were for those in pour nutritional condition to minimize mortality and to maintain a consistent stocking rate across replications. We introduced 7 cow-calf pairs and one dry heifer into each pasture on May 9, 1989 and removed them 3 weeks later. This represents a departure form previous years when animals remained in the pastures for 6 weeks. However, a marked reduction in forage production compelled us to reduce the stocking rate this year in order to preserve consistent utilization rates in control pastures. With the exception of the heifers and replacements required by death losses etc, cows during year 3 were the same animals we observed during year 2 and were assigned to the same pastures they were in previously. We observed the birth dates of all calves and weighed them to the nearest 0.05 kg immediately after birth. Cows and calves were weighed to the nearest 1 kg when they were introduced to pastures and were reweighed 3 weeks later when they were removed. We estimated canopy cover of herbs shrubs immediately after removing cattle from pastures. Cover was estimated from the summed length of interception by each plant along 30 12m transects randomly placed in each pasture during year 1. We estimated standing crop, productivity, and utilization of forbs, perennial grass, and annual grasses by harvesting samples from 40 pairs of 0.70 m 2 plots in each pasture on each of three sample dates. Pastures were sampled immediately after the elk were removed (April 2729), at the midpoint of the spring grazing season (May 30-June 1), and at its end-(June 30-July 2). Samples were dried at 60% for 48 hrs, separated by hand into live and dead, and weighed to the nearest 0.01 g. We used a repeated measures analysis of covariance to analyze weight responses of cattle. Calf birthweight was used as a concomitant observation for calf weights out of pastures and for rates of gain. Cow weights into the pastures were used similarly for adult cattle and heifers. We believe that this analysis offers conservation tests of hypotheses because any cumulative, depressing effect of treatment on the covariate would tend to make it more difficult to detect significant differences among treatments. When use of covariance improves the sensitivity of analyses, we can conclude that this improvement results primarily from reducing experimental error. Vegetation responses were analyzed with a repeated measures analysis of variance for a randomized complete block design. In the future, we plan to use pretreatment biomass of forage as a covariate to improve precision in these analyses. �215 RESULTS All results described in this report are preliminary and are subject revision as analyses are checked for accuracy. Here, we report our findings to provide a timely summary of progress, but we are obliged thoroughly revisit them later. Effects to to of Elk on Cattle Performance Elk grazing during winter influenced the performance of cattle during spring, and the magnitude of that influence depended on year. Weights of cattle at the end of the grazing season during years 2 and 3 declined in direct proportion to elk density (Fig. 1). This decline in final weight resulted from the effects of treatment on cow rates of gain during spring (Fig. 2). We observed no such effects during the first year (linear effect P > 0.78). Average weights of cows at the end of the grazing season as well as rates of gain were substantially lower during the third year that during the first two (Fig. 1, 2). Effects of year were observed across all treatments, including the control, and hence can be attributed the effects of drought and to the cumulative effects of cattle grazing rather than to cumulative effects of elk grazing (year x treatment P-0.28). We observed significant linear declines in calf weights at the end of the grazing season during all years (Fig. 3). These effects approached significance (p-0.06) when the effects of birth weight were removed and were highly significant (p-O.Ol) when the effect of birth weight was included in the analysis. Effects of elk density on calf weights out of the pastures resulted in part from effects of treatment on calf rates of gain (Fig. 4, linear effect P - 0.07) and resulted partially from treatment-induced delays in birth date (Fig. 5, linear effect P-0.08). Calves during year 2 whose mothers were in the high density treatment during year 1 were born an average of 4 days later than calves whose mothers were in the control and were born an average of 6 days later the following year. (linear effect P-0.10). During years 1 and 2, birth weights of calves tended to decline relative to elk density (Fig. 6 P < 0.10) However, this tendency resulted at least in part from a spurious effect of the initial randomization. On average, larger calves were assigned to the control and low density pastures than to the medium and high ones. These assignments resulted purely by chance. That is, there was no opportunity for treatment to affect birth weights of calves during year 1. This effect emphasizes the importance of removing birth weight from the effects of treatment on subsequent weights. The linear effects of treatment on birth weight did not persist in year 3 (P - 0.52); however cows in the control still tended to produce larger calves than those in the treated pastures (Fig. 6 P-0.12). It is unclear whether this persistent effect id due to the initial randomization or to impacts of treatment. However, when significance is detected in subsequent weight responses (e.g. weight out of pastures) when the effects of birth weight are removed by covariance, we can be confident that those responses result from treatment. �216 Effects of Elk on Vegetation Of all begato responses, elk grazing during winter had greatest impact on the residual biomass remaining from the previous years growing season. During all years, elk grazing reduced the standing crop of residual perennial grass (Fig. 8 P~O.02) as well as the residual herbaceous biomass (forbs + annual grass + perennial grass, Fig. 9, P-0.02). These reductions were proportional to density. Standing crops of residual forage declined with year (Fig. 8, 9) but the effect of treatment did not depend on year for either response (P>0.45). Effects of elk density on live forage available to cattle at the beginning of the spring grazing season only approached significance (Fig. 10, 11, P<0.13). We observed no effect of treatment on production of perennial grass or total herbs (Fig. 12, 13), but production consistently declined with year across all treatments (year x treatment P - 0.78). We attribute this decline to effects of drought interacting with cumulative effects of cattle grazing. Effects of elk grazing on standing crops of perennial live and dead perennial grasses resulted from differences in rates of utilization during winter. At the highest densities, elk used about 50% of the residual biomass of perennial grass and total herbs (Fig. 14, 15) and used up to a third of the live biomass that was available at the beginning of the spring grazing season. Elk use accounted for as much as 25% of the total forage produced during the growing season (Fig. 16, 17). The combined effects of elk and cattle utilization of dead perennial grass and herbaceous biomass increased with increasing elk density during all years (Fig. 18, 19). However, we saw no effect of elk density on total utilization of live biomass except for perennial grasses during year 2 (Fig. 20, 21). The proportion of the total utilization of perennial grass that resulted from elk grazing increased with density and year (Fig. 22). Despite reducing the stocking rate from 7 to 14 acres per animal unit month during year 3, total utilization in the control pastures remained relatively constant with year (Fig. 18, 19, 20, 21). The absence of treatment effects on total utilization of live forage resulted from relatively constant utilization of live biomass by cattle across all treatment levels during year 1 and 2 and declines in use during year 3 (Fig. 23, 24). These patterns in utilization were driven by differences in removal of live forage by cattle, which tended to decline relative to treatment, particularly during year 3 (Fig. 25, 26). We observed pronounced effects of elk density on cattle removal of standing dead forage (Fig. 27, 28). Apparently, elk grazing can reduce the availability of live and dead perennial grass to the point that cattle substitute other forages in their diets, or reduce their total intake. During year 1, elk grazing reduced the average canopy cover of vegetation present at the end of the spring grazing from 47% in the control to 31% in the high density treatment (Fig. 29). During subsequent years we observed no effect of treatment except during year 3 when controls tended to have higher canopy coverage than other levels (Fig. 29). Canopy cover declined sharply between years 1 and 2, but �217 remained relatively constant at about 22% thereafter. Of all forage classes, annual grasses declined most steeply in response to treatment and year, declining from a high of 10.8% in the controls during year 1 to virtually 0 in all levels during the third year (Fig. 30). Perennial grasses tended to increase in coverage at intermediate levels of grazing during all years (Fig. 31, P-0.07) and declined sharply at treacment all levels as the study progressed. During all years, shrub canopy cover was higher in the controls than in the treatments grazed by elk (Fig. 31). We observed no consistent effects of elk grazing on canopy coverage of forbs, but did see strong effects of year (P-0.03) DISCUSSION A tendency for elk grazing during winter to reduce the performance of cattle during spring has clearly emerged in our studies during 1988-89. However, although this pattern was substantially more clear than it had been in previous years, there were several responses that nevertheless only approached statistical significance. This emphasizes the imperative need to continue our measurements to assure that our tests of hypotheses offer unambiguous results. That is, we wish to avoid failure to reject null hypotheses as a result of inadequate temporal replication. Forage production and canopy coverage data from control pastures revealed substantial impacts of serial drought interacting with cattle grazing in determining vegetative responses. Although this trend has made our job more difficult and has required us to adjust 'cattle stocking rates, we believe it is a fortunate outcome for our studies. The effects of elk grazing will be more economically and ecologically significant when forage supplies are limited. Because annual variation in forage availability due to variation in can consistently be expected on semi-arid rangelands, we are please that studies happened to coincide with a period of forage scarcity. This coincidence will allow our results to be far more relevant to policy decisions than they would have been if we had only been able to observe abundant forage conditions. The reduction of standing dead perennial grass that resulted form elk grazing appeared to be one the most important determinants of system responses. We surmise that this was the case because residual biomass represents an important "buffer" against forage scarcity during the growing season. Thus, although elk grazing had no discernable impact on the rate of forage production, its effect on the standing crop of live and dead forage was apparently sufficient to drive discernable differences in cattle performance. Prepared by Wildlife Researcher Dan L. Baker Wildlife Researcher �218 .-;t 460 '· I<.. -!-50 <1l ; 440 -'in 0...~ 430 ._ <, 0 I ...._..···_·····-1·········· .._.._._.... 420 - .•...••..•.•......•.•...•.•..•.......•...... -' 0 410 -' -= 'Q) 400 l---------t------+___________ I QD == ~ 390 u ~ 380 0 --- o: ....;: 370 5 0 10 15 20 25 30 35 Elk Density (animals/km2) IYear Means adjusted Contrast v. v. v. v. v. v. v. v. Year 1 eontro~ other. control. 8 .l.lc/1aa2 15 _l.lc/_ control. control. 31 _l.lc/Ial linear err.eta quadratic .rrect. v. v. v. v. Y_r 2 control other. control. 8 .l.lc/~ eontrol 15 .l.lc/_ contro.l 31 .l.lc/Ial lin •• r .r~.ct. quadratic _rr_eta v. v. v. v. Y.ar 3 control. others control. 8 .l.lc/~ 15 _l.lc/Ial control. control. 31 .1.1c/_ lin.ar •rrecta quadratic _rrecta to common 3 weight into pasture. Contrast ss M_an Square 1 1 1 1 1 1 539.563509 198.853090 65.917054 1202.530974 1103.359103 35.157516 539.563509 19••853090 65.917054 1202.530974 1103.359103 35.157516 1.42 0.52 0.17 3.17 2.91 0.09 0.2777 0.4960 0.6911 0.1251 0.138. 0.7709 1 1 1 1 1 1 0.519559 26.513162 99.655.10 43.550571 16.47••62 110.549671 0.519559 26.513162 99.655.10 43.550571 16.478862 110.549671 0.00 0.13 0.51 0.22 0.0. 0.56 0.9607 0.7260 0.5031 0.6544 0.7819 0.4816 1 1 1 1 1 1 724.70973 2.4.92550 261.59071 10.1.60889 977.07580 13.69219 724.70973 284.92550 261.59071 10.1.60.89 977.075.0 13.69219 5.06 1.99 1.83 7.55 6.82 0.10 0.0655 0.2080 0.2252 0.0334 0.0400 0.7676 1 1 1 1 1 1 15••55129. 5.73953. 61.959127 423.377231 493.53479. 11.944060 15••551298 5.739538 61.959127 423.377231 493.53479• 11.944060 1.01 0.04 0.39 2.69 3.13 0.0. 0.3544 0.8549 0.5536 0.1522 0.1271 0.7922 DP Al.l. Year. eontrol. other. eontro~ 8 .l.lc/~ eontro~ 15 _l.lc/_ control 31 .l.lc/Ial 1in_ar errect. quadratic .rr.eta 1 ....- ...•., 2 ----- P Va~u. Pr > P Figure l. Effects of elk grazing on average weights of cows at the end of the spring grazing season. Means are adjusted to remove the influence of weights at the beginning of the season. Vertical bars give ± 1 standard error. Tests of specific hypotheses are shown in the table of orthongonal contrasts. �219 - 2 >. al "C <, ~ ~ •.. ...•..• i1) - •...........•............... -------+ ··················f····· J------ _ 1 al c, .:: ~ C!! . ------- ---_---1 (/l :Il ............................ I 0 -_ -_ 3: 0 U ::a \/J ...• o 5 15 10 20 25 30 35 Elk Density (animals/km2) I Year Means adjusted Contrast DP Year. control v. others control v. .~k/kw.2 control v. 15 .~k/Icoo control v. 3~ .~k/Icoo linear _:tracts quadratiC .~~ecta 1 1 1 1 1 Y.ar 1 control v. other. control •.lk/1ca2 c::ontrol v. 15 .~k/Icoo control 31 •.lk/Ic:a lin.ar e~~ecta quadratic .~~.cta 1 1 1 1 1 Y.ar 2 control v. others contro.l 8 .1k/kw.2 control vs 15 .1k/1coo control v. 31 .~k/Icoo ~in_'" .~~ecta quadratic .~~.cta A~~ • v. • v. v. v. v. Y.ar 3 control other. control. v. 8 .~k/1ca2 control v. 15 .~k/Ic:a control 31 .~k/Ic:a lin.ar .~~ecta quadratic .~~ecta ~ Contr••t ss 1 .- ....- •.. 2 ----- to common weight into pasture. Xean square po Value L22 Pr > P 0.4.0713 .•.• 0.13.61138 0.05999403 0 •.•• 0713 .• 4 0.1386~138 0.05999403 ~.16.72259 1.~20.3.92 0.0.• 9720.9 0.049720.9 2.9. 2 ••6 0.~3 0.73"1 0.09414.10 0.0000026. 0.0176826" 0.09043016 0.026.0733 0.0085564. 0.09"14810 0.00 0.15 0.75 0.22 0.07 0.78 0.9964 0.71"9 0.4~97 0.6539 0.7989 0.4109 1 1 1 1 1 1 0.3702,,713 0.16"30520 0.1233300. 0.53.65963 0.47077113 0.00752465 0.37024713 0.16"30520 0.1233300& 0.53865963 0.47077113 0.00752465 4.26 1.&9 1.42 6.19 5.41 0.09 0.0847 0.21.5 0.278. 0.0"73 0.0590 0.7786 1 1 1 1 1 1 0.352.9471 0.01134.6. 0.13970791 0.95025955 1.11321026 0.02759750 0.352.9471 0.01134.6. 0.13970791 0.95025955 1.11321026 0.02759750 1.00 0.03 0.39 2.6. 3.14 0.0. 0.3567 0.&63. 0.5530 0.1525 0.1266 0.7.95 1 0.0000026. 0.0176.26.• 0.09043016 0.026.0733 0.00.55648 3 1.~6872259 1.120.3.92 0.35 0.15 0.310. 0.5740 0.7093 0.1352 0.1"20 Figure 2. Effects of elk density on average daily gain of cows during the spring grazing season. Means are adjusted to remove the influence of weight at the beginning of the season. vertical bars give ± 1 standard error. Test of specific hypotheses are shown in the table of orthogonal contrasts. �220 - 100 -•• I:U) ..!I: <lJ !---.-.-..·.·..... 90 :;, ··.I I " - _.... ....J 7l ] ·················1 :'il Q., 80 '0 ....J :;, C ...J ---_ ..:: 70 .- I:U) <lJ li= ':tI u ::E 60 ----------_ --1 \I'J ....• o 5 15 10 25 20 30 35 Elk Density (animals/km2) I Year --- 1 Means adjusted Contrast DF All Year. control other. control 8 elk/02 control 15 elkin control 31 elkin linear .rrect. quadratiC e~~ecte v. v. v. v. v. v. v. v. Y.ar 1 control others control 8 elk/02 control 15 elkin control 31 elkin lin.ar .~1!ect. quadratic e~~ecta 1 1 1 1 1 1 Contrast ss _ 2 ----- 3 to common birth weight. lIeanSquare 86.3.76491 86.3876491 1••4070203 18.4070203 26.0353602 26.0353602 17••8624231 17••8624231 181..1.1.54137181.1154137 1.0260403 1.0260403 P Value Pr > F 2.49 0.53 0.75 5.1.5 5.21 0.03 0.1659 0.4941 0.4200 0.063. 0.0625 0.8692 1 1 1 1 1 1 9.3496115 4.5452424 25.3253283 21..0622934 39.7246.30 1..0022327 9.3496115 4.5452424 25.3253283 ;;11.062293" 39.72.• 6.30 1.0022327 0.31 0.15 0.84 0.70 1.32 0.03 0.5978 0.7113 0.39.•• 0.435" 0.29.•• 0.8614 Year 2 other. control control 8 elk/02 control. 15 elkin control 31 elkin lin.ar e~rect. quadratiC e~r.cta 1 1 1 1 1 1 14.3580781 2.2623351 2.2369114 39."620550 ..0.9202607 2.3"16."5 1".3580781 2.2623351 2.236911.• 39."620550 .• 0.9202607 2.3.• 168.• 5 0.75 0.12 0.12 2.05 2.12 0.12 0."211 0.7"35 0.7.•.• 9 0.2023 0.1953 0.7393 Y_r 3 control other. control 8 .lk/n2 eontrol 15 elkin control v. 31 ·:UC/a lin.ar e~~ecte quadratiC .~recte 1 1 1. 1 1 1 .5.592.• 6. 64.9"6494 5.33.• 7 .• 2 151.120379 112.575111 1.501."01. .5.592468 7.5" 5.72 64.9"6494 0 •.• 7 5.334742 13.30 151.120379 11.2.57511.1. 9.91. 0.1.3 1.501401 v. v. v. v. v. v. v. 0.0335 0.0539 0.5187 0.0107 0.0199 0.72.6 Figure 3. Effects of elk density on average weights of calves at the end of the spring grazing season. Means are adjusted to remove the influence of birth weights. vertical bars give ± 1 standard error. Test of specific hypotheses are shown in the table of orthogonal contrasts. �221 1.011] ~0.961 ~0.911l If······· ...........j 1 I·· .!t: -;0.861 •• ........... .30.811 ;' ;' 'J) ~ ............ <, :tl c, 0.761 / ........;' .~ 0.711 ~ C-' 0.661 . ;'.... ;';'........ ;';' . .•...•.. ' . . . <,.•.. """''- / I := 00.611 ~ 0.561 o 5 10 15 20 25 30 35 Elk Density (animals/km2) I Year 1 Means adjusted Cont.rast 3 to common birth weight. DP Contrast ss ]. ]. ]. 0.0],7].7905 0.0].036040 0.000],2449 0.0530949], 0.044],],54. 0.00903823 0.0].7].7905 0.0],036040 0.000].2449 0.0530949]. 0.044].].548 0.00903.23 1.&6 ],.],2 0.0], 5.75 4.78 0.9& 0.22],4 0.330], 0.9].],3 0.0534 0.07].4 0.3606 0.00422800 0.00002394 0.00043785 0.0],78],042 0.020824],0 0.00299605 0.60 0.00 0.06 2.53 2.96 0.43 0.467& 0.9554 0.8],],4 0.],62. 0.],362 0.5383 A~~ Vear. control v. other. control v. 20 al.k/.q control v. 40 al.k/sq control. vs 80 a~k/.q 11near .:t~ect. ]. ]. quadratic ]. arracta 2 ----- Mean square F Value Pr > P Year ]. control v. others control v. 20 al.k/.q control v. 40 a~k/sq control v. 80 a~k/.q linear e:tr.ct. ]. ]. ]. ]. ]. quadratic ]. 0.00422&00 0.00002394 0.00043785 0.0],78].042 0.020824],0 0.00299605 ]. ]. ]. ]. ]. ]. 0.006],0].05 0.00543534 0.00046.39 0.00920845 0.00630068 0.00000].80 0.006],0],05 0.00543534 0.00046839 0.00920845 0.00630068 0.00000].80 ],,],4 1.02 0.09 ]..73 ]..].8 0.00 0.3259 0.35],6 0.7769 0.2368 0.3],87 0.9859 ]. 0.00703680 0.00954],56 0.00383068 0.02879422 0.0],963].4], 0.0],],79],44 0.007036.0 0.00954].56 0.003.306. 0.02879422 0.0].963].4]. 0.0].].79].44 ]..30 ]..76 0.7]. 5.3], 3.62 2.].7 0.29.0 0.2329 0.4328 0.0607 0.],057 0.],907 Yaar .rracta 2 control. v. other. control v. 20 .~k/.q control v. 40 al.k/.q control 80 a~k/.q lin.ar _:tract. v. quadratic arract. Y.ar 3 control. v. ·other. control 20 a~k/.q control v. 40 a~k/.q control 80 a~k/.q lin •• r arracta v. v. quadratic arracta ]. ]. ]. ]. ]. Figure 4. Effects of elk density on average daily gain of calves during the spring grazing season. Means are adjusted to remove the influence of birth weights. vertical bars give ± 1 standard error. Test of specific hypotheses are shown in the table of orthogonal contrasts. �222 05/08,--------------------------------- __ ~ ;>, ~ 04/28 <, --c' o,... ... r~~~~~~_t--~~~~~~----------------i ... ~ 04/18 -' (tI Q ..c ......•.••...•...•.•. -' ... - ::04/08 ······£·····....··....····..····..·····..···..····I (tI U o 5 10 15 20 25 30 35 Elk Density (animals/km2) !Year contra.t AJ.l.Year. others control control 8 el.k/Jca2 ],5 el.k/Ino control. control v. 3]. el.k/Ino l.inea.re:t!:t!.ct. .:t!:t!eeta quadratic v. v. v. v. v. Year ]. other. control control v. 8 el.k/Jca2 ],5 elk/Ino control. cont.J:'ol. v• 31 el.k/Ino lin •• r •:t!:t!ecta quadratiC e:t!:t!ecta Y.ar 2 control v. other. control. 8 elk/ka2 control. 15 .l.k/Ino control 31 elk/Ino l.inear e:t!:t!ecta quadratic .:t!:t!ecta v. v. v. Year 3 control other. control 8 .1k/ka2 control v. 15 .l.k/Jca control v. 31 .l.k/Jca linear e:t!:t!ect. quadratic .:t!:t!.eta v. v. 1 .•.••-.•.•.. 2 ----- or Contrast ss Mean Square ]. ]. ]. 90.75000000 34.72222222 72. OOOOOOOO 80.22222222 72.8642857]. 28.2857].429 90.75000000 34.72222222 72. OOOOOOOO 80.22222222 72.8642857]. 28.2857],429 ]. ]. ]. F Va~u. Pr > P 5.66 2.].6 4.49 5.00 4.54 ]..76 0.0549 0.].9].7 0.0785 0.0667 0.077]. 0.2326 0.6944444 6.0000000 2.6666667 8.1666667 17.2023810 1.5779221 0.6944444 6.0000000 2.6666667 8.1666667 17.202381.0 1.5779221 0.].]. 0.93 0.41 1..27 2.67 0.24 0.7539 0.3719 0.5438 0.3033 0.1534 0.6383 1 1 1 1 1 1 42.2500000 54'0000000 16.6666667 20.1666667 7.2023810 20.3658009 42.2500000 54.0000000 16.6666667 20.1666667 7.20238],0 20.3658009 ]..31 1..68 0.52 0.63 0.22 0.63 0.2958 0.2431 0.4991 0.4590 0.6531 0.4569 1 1 1 84.0277778 28.1666667 80.6666667 66.6666667 63.2595238 35.4632035 84.0277778 28.1666667 80.6666667 66.6666667 63.2595238 35.4632035 4.51 ]..51 4.33 3.58 3.39 ]..90 0.0779 0.2650 0.0827 0.].075 0.],],50 0.2170 ]. ]. 1 1 1 1 ]. 1 ]. 3 Figure 5. Effects of elk density on birth date of calves. vertical bars give ± 1 standard error. Test of specific hypotheses are shown in the table of orthogonal contrasts. �223 38.0,,-----------------------------------------, 37.0] I 36.0, i I j 35.01 ....; == I I 34-.0, j : ,~ 32.0i ~ 1 ------.f ··"'-<::: , I -- ------ --- -................. ] ;: 33.0j I I 31.0 30.0 o 5 10 15 20 25 30 35 Elk Density (animais/km2) 1 /Year Contrast A~~ Years control vs others control va 8 e~k/kla2 control VB 15 ••~k/kla control vs 31 e~k/kla linear erfecta quadratic erfecta OF Contrast S5 Mean Square 2 ----- F Value Pr > 3 F 1 1 1 1 1 1 22.71210354 10.28149578 11.10698537 26.36201103 22.76904171 2.47885934 22.71210354 10.28149578 11.10698537 26.36201103 22.76904171 2.47885934 8.38 3.79 4.10 9.73 8.40 0.91 0.0275 0.0994 0.0894 0.0206 0.0274 0.3759 control vs others control va 8 e~k/k1a2 control va 15 e1k/kla control VB 31 e~k/kla linear e~fect8 quadratic effects 1 1 1 1 1 1 9.3578780 4.0265907 1.8530518 17.0176295 15.0600917 0.0379540 9.3578780 4.0265907 1.8530518 17.0176295 15.0600917 0.0379540 2.30 0.99 0.45 4.18 3.70 0.01 0.1804 0.3585 0.5251 0.0869 0.1028 0.9262 Year 2 control VB others control. va 8 elk/kla2 control va 15 elk/kla control va 31 e~k/kla linear effects quadratic effects 1 1 1 1 1 1 3.23735720 0.11659716 1.4078.233 8.29033618 9.71352236 0.16260247 3.23735720 0.11659716 1 ••0784233 8.29033618 9.71352236 0.16260247 1.60 0.06 0.69 4.09 4.79 0.08 0.2534 0.8185 0.4368 0.0897 0.0713 0.7866 Year 3 control va others control va 8 elk/kla2 control va 15 e~k/kla control va 31 elk/kla linear effects quadratiC effecta 1 1 1 1 1 1 11.5337773 10.2763934 10.3982832 3.5664152 1.6063703 11.0560698 11.5337773 10.276393. 10.3982832 3.5664152 1.6063703 11.0560698 3.28 2.92 2.95 1.01 0.46 3.14 0.1203 0.1384 0.1365 0.3531 0.5246 .).1268 Year 1 Figure 6. Effects of elk density on birth weight of calves. vertical bars give ± 1 standard error. Test of specific hypotheses are shown in the table of orthogonal contrasts. �224 Forage Class=Perennial Grasses Age=Standing Dead 30.------------------------------------------ >, (':l ?O ;:;:0.. 0 •.. U ~ -... 10 ··············... I 1 ~ ....... ··················-1·············· 1 ....• tr: '" I--___ ~ . "0 c:: .. .................. ······················ ...r Q.) N •.. c.:l -- 0 ---f--------r---- --------------4 _I_ 10 5 0 15 20 25 30 35 Elk Density (animals/km2) I Year contrast All Year. contro~ v. other. control v. 8 elk/ka2 control v. 15 elk/ke control v. 31 elk/ka linear e~~ec:ta quadratic e~~ec:t. Year 1 contro~ v. other. control v. 8 elk/ka2 control v. 15 elk/_ control v. 31 .~k/ka linear e~~ec:t. quadratic e~~ecta Dr 1 1. ]. ]. ]. 1 ]. ]. ]. ]. ]. l. 1.09.5203835 ],9.0505832 .2.352.932 2].7.9].].3339 227.7736.9::2 0.2387],09 ••63 0.8]. ]..79 9.22 9.6. 0.0], 0.07.9 0 ••039 0.2292 0.0229 0.02l.0 0.9232 66.630.25 l..9 •• 3l.7 .7. 229l.85 l.37.207396 166 ••396.3 0.05.3.6 66.630.25 1.9 ••3],7 .7.229185 l.37.207396 ],66.8396.3 0.05.3.6 3.0]. 0.09 2.].3 6.l.9 7.53 0.00 0.l.336 0.77.8 0.19.6 0.0.73 0.0336 0.962], 39.0738233 l.2.l.3.2352 5.l.82].2.3 9].•.23.5670 86 •••973.0 3.1659021 1.90 0.59 0.25 •••3 ••20 0.15 0.217. 0 .•7].7 0.6337 0.0799 0.0.63 0.70 •• ],3.783•.273 7.l.171.0. ••5079799 18.5157027 ].5••129369 0 ••899007 1 Year 3 control v. others control v. 8 elk/ka2 control v. l.5 elk/ke control. va 31 el.k/_ linear e~~ec:ta quadratic e~rec:ta 1 1 1 1 1 l. 13.7.34273 7.1171.08 •• 5079799 1•• 5157027 15 ••],29369 0 •••99007 Val.ue ::27.].2 ]. ••00 8.87 36 ••3 30.33 0.96 P>." > 3 r 1.09.5203835 19.0505832 .2.352.932 217.9],13339 227.7736.92 0.2387l.09 39.0738233 1.2.13.2352 5.1.212.3 91.23.5670 86 •••973.0 3.1659021 ]. r _an l. 1 Square ----- Contr ••• t ss Year 2 control v. other. control v. 8 elk/ka2 control v. 1.5 elk/ka control v. 31 elk/ke linear e~~ec:ta quadratiC e~~ec:ta ]. ]. 1 ...........• 2 0.0020 0.0096 0.02.7 0.0009 0.00l.5 0.36.], Figure 7. Effects of elk density on standing dead perennial grass present at the begining of the spring grazing season. vertical bars give ± 1 standard error. Test of specific hypotheses are shown in the table of orthogonal contrasts. �225 Forage Class=Total o 5 10 Herbs 15 Age=Standing 20 Dead 25 30 35 Elk Density (animals/km2) IYear Contrast ll~ Year. control others control. 8 el.l</ka2 contro~ 15 e~lc/k1o 31 e1.lc/1ea control linear .-rr.ct. quadratic e~~.cta v. v. v. v. Year l. control other. 8 .~lc/lnl2 control control 15 .1lc/lclo control 31 .~k/k1o linear e~~.cta quadratic e~~ecta v. v. v. v. v. v. v. v. Year 2 control other. control 8 ellc/ka2 control 15 e~lc/k1o control 31 .~k/k1o 1inear e~~.cta quadratic e~~ecta v. v. v. v. Year 3 control. other. cantZ'ol a e~k/ka2 control 1.5 elk/lea control 31 e1k/1ea 1ine.r .~~ecta quadratic .~~ecta OF 1 1. 1. 1. 1 1 Contra.t 1 ...•.....•.. 2 ----- sa 128.8152054 26.262140. 54.0355911. 234.8669923 241.0767642 0.0932021. Mean Square 128.8152054 26.2621408 54.0355911 234.8669923 241..0767642 0.0932021. F Va~u. Pr > J! 4.88 1.00 2.05 8.90 9.14 0.00 0.0692 0.3569 0.2023 0.0245 0.0233 0.9545 1 1 1 1 1. 78.611.376 4.194883 5•• 546209 144.437949 1.71.82541.4 1..1.265.1 78.61.1.376 4.1.94.83 58.546209 144.437949 1.71.82541.4 1..1.26581 3.27 0.17 2.44 6.01. 7.1.6 0.05 0.1.204 0.6905 0.1.694 0.0496 0.036. 0 ••357 1 1. 1 1 1 1 46.970164 1.6.0.2031. 7.64274:Z 100.2548:Z4 93.378115 1.841037 46.9701.64 1.6.082031. 7.64:Z74:Z 1.00.254.:Z4 93.378115 1 ••41037 2.17 0.74 0.35 4.64 4.3:Z 0.09 0.1.909 0.4215 0;5738 0.0747 0.0.29 0.7802 1 1 1 1 1 1 15.5112791 7.9399379 5.3639:Z76 20.3703635 16.9.665:Z9 0.6793407 15. 511:Z791 7.9399379 5. 3639:Z76 :Z0.3703635 16.9a665:Z9 0.6793407 l. 31.53 16.14 10.90 41.40 34.5:Z 1.3. 3 0.0014 0.0070 0.0164 0.0007 0.0011 0.2.45 Figure 8. Effects of elk density on the standing dead biomass of herbaceous forage (annual grass + perennial grass + forbs) present at the beginning of the spring grazing season. vertical bars give ± 1 standard error. Test of specific hypotheses are shown in the table of orthogonal contrasts. �226 Forage Class=Perennial Grasses Age=Standing Live 12 CI,.;! c 11 "';n 10 - >, 9 :tl ::E 8 c, 0 s.. u "'0 .~-. c 7 6 5 ············ ~ _----1r--_______ C'il -' 4 t C/) "'0 I!.l N .. C'il c;, 3 1--:.:::::.......................]'1 ••••••••••• ••••.•••••••••••. ••••••.• ---------::.::1 2 1 0 5 10 15 20 25 30 35 Elk Density (animals/km2) I Year Contrast Y.ar. control v. others eontrol. v. 8 .l.k/ka2 control va 1.5 .l.k/ka control v. 31. .l.k/ka linear e:t'r.ct. quadratie .rr.et. 1 .-.-.-.-.-.-2 Contr ••t ss X.an 1 1 1 1 1 1 10.1.8740572 4.4551.7556 3.16250164 15.43814621 13.55941251 0.1.1..75685 1.0.18740572 4.45517556 3.162501.64 15.43814621 13.55941251 0.11.75685 1 1 11..638.369 Of' SqUare F Val.u. ----- 3 Pr > f' lll. Y.ar 2.36 1.03 0.73 3.57 3.14 0.03 0.1.757 0.3493 0.4253 0.1077 0.1270 0.8738 11.6388369 4.5603544 4.4460805 16.9130156 15.289561.3 0.30521.76 3.66 1. 43 1.40 5.32 4.81 0.1.0 0.1042 0.2761 0.2816 0.0605 0.0708 0.7671 1. 1 1. 0.61690690 0.35.87224 2.453705.9 2.1.7558971 0.0001.3527 1.45132601 0.61690690 0.35887224 2.45370589 2.17558971 0.00013527 0.53 0.23 0.13 0.90 0.80 0.00 0.4931 0.6511. 0.7292 0.3794 1 1 1. 1 1. 1. 0 ••3179256 0.5401.6100 0.13.79112 1.26.9945. 0.9.560106 0.0010750. 0.831.79256 0.54016100 0.1.3.791.12 1.26.9945. 0.98560106 0.00107508 1.56 1.01 0.26 2.38 1.85 0.00 1 control va others control v. a .lk/~ control va 1.5 elk/ka control. v. 31..l.k/ka linear err.eta quadratic .rr.eta Y.ar 2 control v. other. control v. 8 .lk/~ control v. 15 .lk/ka control. v. 31..l.k/ka l..1n.ar.rr.eta quadratic .rraeta Year 3 control v. other. eontrol. v. 8 .l.k/ka2 control. v. 1.5 .l.k/ka control. v. 31. .1.k/ka linear err.eta quadratie .rraeta I. 1 1. 1 1 1. I. 4.5603544 4.4460805 1.6.9130156 15.2.95613 0.3052176 1.451.32601 0.4061 0.9946 0.2581 0.3529 0.62.0 0.1737 0.2227 0.9656 Figure 9. Effects of elk density on the biomass perennial grass present at the beginning of the season. Vertical bars give ± 1 standard error. hypotheses are shown in the table of orthogonal of live spring grazing Test of specific contrasts. �227 Forage Class=Tot al Herbs Age=Standing Live 21r-----------------------------------------~ c:?20 E 19 <, 18 ~17 - 16 ~15 ::::E14 0.13 12 e ull "0 ~10 c 9 8 7 6 5 4 3 ----- ---...... r I~~=·· I ...··=~·1[~·~=·~~.-------------i ............. . { 2~~~~~--~~~~~~--~~--~~~~--~~ 0 5 10 15 20 25 30 35 Elk Density (animals/km2) I Year COntrast DP ----- 2 1 ............ ss F Val". Pr > F 1.41 0.41 0.24 3.16 3.04 0.08 0.2801 0.5468 0.6411 0.1257 0.1320 0.7925 36.73281.05 23.0398723 9.7840209 47.8565178 37.1404557 1.2054461. 3.37 2.11. 0.90 4.39 3.41. 0.1.1 0.11.61 0.1.963 0.3801. 0.0811 0.1.1.45 0.7509 1.257891.79 1.53718415 0.09212698 3.27948980 2.1.5265073 0.55268375 1..25789179 1..53718415 0.09212698 3.27948980 2.15265073 0.55268375 0.25 0.30 0.02 0.64 0.42 0.11 0.6371 0.6028 0.8975 0.4531 0.5399 0.7532 0.06655416 5.36298994 0.00143659 2.70927976 6.78867097 3.83760183 0.06655416 5.36298994 0.00143659 2.70927976 6.78867097 3.837601.83 0.04 3.02 0.00 1.52 3.82 2.16 0.8529 0.1330 0.9782 0.2631 0.0984 0.1921 COntrast X •.•. n square All year. control. v. other. control. v. 8 .lk/~ control. v. 15 .lk/lal control. v. 31 elk/lea linear e.t:tecta quadratic .rr.eta 1 1 1 1 1. 1 15.98214635 4.62274020 2.73095877 35.87858778 34.45593751 0.8581.1.735 15.98214635 4.62274020 2.73095877 35.87858778 34.45593751. 0.85811.735 Year 1. control. v. other. control va 8 .1.k/~ control v. 15 e1.k/1ea control v. 31 e1.k/1ea linear _:t:tect. quadratic .rr.eta 1. 1. 1. 1 1. 1 36.73281.05 23.0398723 9.7840209 47.85651.78 37.1404557 1..2054461. Year 2 control. v. other. control v. 8 .1.k/~ control v. 15 elk/lea contz'ol. v. 31 elle/Iea 1in•.."..rrecta quadratic .rrecta 1. 1 1 1 1 1 Year 3 contzol. va other. control. va 8 elle/~ control. v. 15 elk/lea control va 31 .lle/Iea lin.ar erreeta quadratiC errecta 1 1 1 1 1 1 3 Figure 10. Effects of elk density on the standing live biomass of herbaceous forage (annual grass + perennial grass + forbs) present at the beginning of the spring grazing season. vertical bars give ± 1 standard error. Test of specific hypotheses are shown in the table of orthogonal contrasts. �228 Forage Class=Perennial Grasses Age=Standing Live 40r-------------------------------------------~ C\l c <, ~30 ;: c -' C) :: -g 20 t 1······················f· >. 1-0 c:l E 'c ~ -' Q.) ···1 f················· 10 ~--------r-------------------------- I z 0 5 0 10 15 20 25 30 35 Elk Density {animals/km2} I Year Dr contrast A~~ 'tear. contro~ v. contro~ v. .contro~ v. contro~ other. 8 .~k/ka2 15 e~k/ka 3~ elk/ka ~ ~ quadratic .rracta 1 1 1 1 Year 1 control v. other. CODtrO~ v. 8 e~k/ka2 eontro~ 15 .1k/~ contro~ 31 .~k/ka lin.ar .'t1!.cta quadratic erreet. 1 1 1 1 1 1 'tear 2 control v. others control V. 8 e~k/ka2 control V. 15 e~k/ka contro~ v. 3~ e~k/ka ~inear erracta quadratic eireeta 1 1 1 1 1 1 'tear 3 control v. other. contro~ v. 8 e~k/ka2 contro~ 15 e~/ka contro~ V. 31 e~k/ka linear erreeta quadratic erreeta 1 1 1 1 1 1 v. ~in.ar .~t.ct. v. v. V. Contrast 1 .'.'-'-"'-' 2 ss 0 •• 61742259 0.43003319 4.79307473 9.71015692 8.~5222059 12.05896492 Xean Square 1! Va~\le 0.41742259 0.43003319 4.79307473 9.71015692 8.15222059 12.05896492 0.05 0.05 0.54 1.09 0.91 0.00 0.00 0.0013672 0.0000816 0.0013672 0.0000816 19.9414759 19.9414759 20.6763684 18.9607403 20.6763684 18.9607403 44.3297312 44.3297312 0.~8278700 0.27001437 0.16558095 0.01456.37 0.18278700 0.27001437 0.16558095 0.01456837 0.00006162 0.29060014 0.00006162 0.29060014 1.35 Pr > 0.8360 0.8336 0.4915 0.3373 0.3763 0.2894 0.9907 2.05 4.80 0.02 0.03 0.02 0.00 0.00 0.03 0.8873 0.8633 0.8927 0.9681 0.9979 0.8583 0.3627 0.3240 2.27891792 0.97 2.71071165 2.71071~65 1.16750100 1.16750100 0.94252471 0.34003452 0.94252471 0.34003452 1.39808833 1.15 0.50 0.40 0.14 0.60 3 r 0.9977 0.1922 0.1853 0.2020 0.0711 2.16 2.24 2.27891792 1.39.08.33 ----- 0.5073 0.549. 0.7167 0.4697 Figure 11. Effects of elk density on primary production of perennial grass during the spring grazlng season. Vertical bars give ± 1 standard error. Test of specific hypotheses are shown in the table of orthogonal contrasts. �229 Forage Class=Total Herbs Age=Standing live 50 ..• CIl ~40 ~ -..• 0 c:; 30 ~ -e e ?O >,- Q.. I ······················1··········· 1L.................... .I . ca"" ...S ~------- ...•v 10 Q.. "" ...•. f- ------ -t------------------I Z 0 0 5 10 15 20 25 30 35 Elk Density (animals/km2) I Year 2 ----- 3 1 ............ ss Mean square F Value Pr > 1" 1 1 1 1 1 1 0.19.639 •• ••3.782892 12.02569359 6.06798621 2.21315195 13.06577911 0.19.639 •• ••3.782892 12.02569359 6.06798621 2.21315195 13.06577911 0.01 0.15 0 ••2 0.21 0.0. 0 ••5 0.9373 0.7118 0.5.32 0.6633 0.7915 0.5268 Y.ar 1 contro~ v. others control. V. 8 el.k/ka2 control 15 el.k/kD control. V5 31 el.k/ka l.inear e~f'ecta quadratic e~~ecta 1 1 1 1 1 1 8.68.7262 5.2721022 32.9573929 0.6696918 1.0352021 .5.5193910 8.6••7262 5.2721022 32.9573929 0.6696918 1.0352021 .5.5193910 0 .•1 0.25 1.55 0.03 0.05 2.1. 0.5.67 0.6365 0.2599 0.8651 0.832. 0.1961 Year 2 control v. other. control. v. • el.k/ka2 control v. 15 el.k/ka control. 31 el.k/ka l.inear e~f'ecta quadratic e~~ecta 1 1 1 1 1 1 1.1131773 0 ••063575 16 ••6.9585 0.3605059 0.0030192 1:l.3:l36379 1.1131773 0 ••0.3575 1•••689585 0.3.05059 0.003019:l 12.3236379 0.05 0.02 0.62 0.01 0.00 0.5:l 0 ••350 0.9000 0 ••626 0.90.1 0.9913 0 ••966 1 1 1 1 1 1 22.7166972 :l7.7917863 1.2.519152. ••.2067006 2.6056826 15.972:l698 2:l.7166972 27.7917863 1.2.519152. 8.2067006 2.6056826 15.9722698 1.38 1.69 0.76 0.50 0.16 0.97 0.2.69 0.261. 0 ••170 0.5069 0.70.7 0.3630 Contrast DI" Al.l.Year. eontro~ v. others control. V5 8 el.k/ka2 control. V. 15 e1k/kD control. 31 e1k/ka 11n.ar .:t~.cta quadratic e~~ecta V. V. V. V. Year 3 control oth.1:'S control v. a .~/ka2 control. 15 el.k/ka control. v. 31 el.k/ka l.inear ef'~ecta quadratic e~f'ecta V. Contrast Figure 12. Effects of elk density on primary production of herbaceous biomass (annual grass + perennial grass + forbs) during the spring grazing season. vertical bars give ± 1 standard error. Test of specific hypotheses are shown in the table of orthogonal contrasts. �230 Forage Class=Perennial Grasses Age=Standing Dead _100r---------------------------------------~ ~~ ~ 90 ~ >, 80 .a v 70 ... 0 60 I:Ul (ll t:. I:Ul !: 50 ·C c, 40 .•. ... CFl 0 !: ...-' 0 •...•............ 30 »: 20 as N := ;:::, .. .... .. .. ' 10 ' ' ••.•..•..•.••• »: 0 0 5 10 15 20 25 30 35 Elk Density (animals/km2) I Year contra.t DP v. v. v. v. All Year. contro~ other. control. 8 elk/~ control 15elk/~ control 31 elk/lea linear e1"1"ecta quadratiC .r~.c:ta v. v. v. v. Year 1 control others control. 8 elk/~ control. 15elk/1ea2 31 elk/lea eontro.l linear e1"1"ect. quadratiC e1"1"ecta v. v. o~.r. v. v. Year 2 con'C:rol control 8 elk/1ea2 15e1.k/~ contro~ control. 31 e1.k/1ea linear e~r.ct. quadratic e1"l!ecta v. v. v. v. Year 3 control other. control 8 e1.k/~ control. 1.5elk/~ control. 31.el.k/Iea linear el!l!ecta quadratic el!l!ecta 1 1 1 1 1 1 contra.t S5 7307.84260 1712.23323 2754.63154 13347.91139 13390.08478 0.77549 1 .•...•.•.... 2 Xean SqUare 7307.84260 1712.23323 2754.63154 13347.91139 13390.08478 0.77549 F Va1u. ----Pr> 3 P 31..65 7.42 11.93 57.81 57.99 0.00 0.0013 0.0345 0.0136 0.0003 0.0003 0.9557 1 1 1 1. 1 1 1772.20140 249.99284 1162.60339 2831.231.45 3096.20466 63.5.878 1772.20140 249.99284 1.1.62.60339 2831.23145 3096.20466 63.5••78 14.26 2.01 9.35 22.78 24.91 0.51 0.0092 0.2059 0.0223 0.0031 0.0025 0.5013 1 1 1. 1 1 1 3250.78443 1072.52751 595.27664 6.0 ••14667 63.9.01475 96.8.756 3250.78443 1.072.52751 595.27664 6808.14667 6389.01475 96.8.756 10.99 3.63 2.01 23.02 21..60 0.33 0.0161 0.1055 0.2058 0.0030 0.0035 0.5879 1 1 2396.37510 534.078.2 1050.45371 41.45••7817 4205.5641.5 11.52041. 2396.37510 534.07 ••2 1.050.45371. 41.45 ••7817 4205.56415 1.1.52041 1.4.55 3.24 6.3. 25.18 25.54 0.07 0.0088 0.1218 0.0449 0.0024 0.0023 0 ••002 1 1 1 1. Figure 13. Effects of elk density on elk utilization of standing dead perennial grass. vertical bars give ± 1 standard error. Test of specific hypotheses are shown in the table of orthogonal contrasts. �23l Forage Class=Total Herbs Age=Slanding Dead _100 ~ ~ 90 ~ >. ~ ~ ::£ ~ :... 0 •••QIJ ::1 j 60j 1 c 50 ·c ._aCJJ. 40 1 30] C a _, :..s .--:.:: N ;:J 20 10 0 0 5 10 15 20 25 30 35 Elk Density (animals/km2) I Year Contrast All Year. control. v. other. control v. 8 elk/1<1L2 eontro~ va 15.1.k/k1II2 control. va 31 e1.k/1<II lin.ar .~~ect.a quadratic .~~ect.a Y.ar 1 control. v. other. control. v. 8 .lk/1<II2 control. v. ].5.1.k/1<II2 control. v. 3], .lk/I<II linear .~t'.ct.a quadratic e~~.ct.a v. v. Year 2 control. other. control 8 .].k/1<1L2 control v. ].551.k/1<II2 control v. 31 .lk/_ 1.inear .~~ect. quadratic e~~ect.a Y.ar 3 control v. others control 8 .1.k/1<1L2 control v. 15elk/1<1L2 3], elk/lao control linear e~~~ qua<ft.~c ~, v. v. 1 .....,...... 2 OF Contrast 5S Mean 1 6299.69136 1369.16039 2765.78365 10988.1.9107 ].1189.06440 25.84663 6299.69136 1369.16039 2765.78365 10988.19107 ]. ]. 1 1 1 1 ]. ]. ]. 1 ]. ]. ]. ]. ]. 1 1 1 ]. 1 1 ]. ]. Square ].].],89.06440 25.84663 F l1al •• e ----- Pr > 1! 20.74 4.51 9.11 36.].& 36.84 0.09 0.0039 0.0779 0.0235 0.00],0 0.0009 0.7803 1627.91193 276.89855 1034.86445 2502.09457 267]..45338 70.75377 1627.91],93 276.89855 ].034.86445 2502.09457 2671.45338 70.75377 ].0.05 ]..7]. 6.39 15.45 ],6.49 0.44 0.0193 0.23&9 0.0448 0.0077 0.0066 0.5332 2073.04&32 364.356].9 570.934.9 469&.378&], 4.64.4292], 73.47566 2073.04832 364.356].9 570.93489 4698.37 ••]. 4864.4292]. 73.47566 7.50 ]..32 2.06 ],6.99 1.7.59 0.27 0.0338 0.294& 0.200. 0.0062 0.0057 0.62.7 2662.],2372 .04.36042 ].226.&3770 396 ••~7l7' 3.].7. 0,.7. 2662.].2372 804.3&042 1226.83770 396& ••737& 18.49 5.59 8.52 27.57 26.52 0.56 0.005]. 0.0560 0.0267 0.0019 0.0021 0.4831 &0.3!r~ 3817.00979 80.3&~1.1 3 Figure 14. Effects of elk density on elk utilization of standing dead herbaceous biomass (annual grass + perennial grass + forbs). Vertical bars give ± 1 standard error. Test of specific hypotheses are shown in the table of orthogonal contrasts. �232 Forage Class=Perennial Grasses Age=Standing Live ~100r---------------------------------------------~ se ~ w 90 >, 80 ~ II) I:lD 70 1 ~ •..0 ~ I:lD .-•c.. tr: ....0 c 601 50 40 -----_ ..... ------- 30 0 :.:: 20 .--:~.: N ;:::l 10 ............... »: _---I 1 1-············ ·.·····.·.·.·.·.·.·-.· 0 0 5 10 15 20 30 25 35 Elk Density (animals/km2) I Year 1 ..._.,_..... 2 ----- DF Contraat S8 X.an Squa",. F Va~u. Pr > F A.U Year. control v. other. control V& 8 .l.k/1aI2 contro~ v. 15elk/Jca2 contro.l V& 31 elk/_ linear .~r.ct. quadratic e~~ecta 1 1 1 1 1 1 2151.386070 501.592860 1677.844067 2525.766.71 2593.558235 357.835612 2151.386070 501.592860 1677.844067 2525.766871 2593.558235 357.835612 6.23 1.45 4.86 7.32 7.51 1.04 0.0467 0.2734 0.0696 0.0353 0.0337 0.3478 Year 1 control V& other. control vs 8 elk/1aI2 control. vs 15el.k/_2 cont.%'ol V& 31 .1k/_ linear e~~ect& quadratiC e~~ecta 1 1 1 1 1 1 703.303989 169.003609 627.390022 724.270722 744.309184 180.283581 703.303989 169.003609 627.390022 724.270722 744.309184 180.283581 5.05 1.21 4.51 5.20 5.35 1.29 0.0657 0.3128 0.0780 0.0627 0.0601 0.2986 Year 2 control vs others control vs 8 .1k/1aI2 contro.l v. 15elk/Jca2 control V& 31 el.k/_ linear e~~ecta quadratiC .~~ecta 1 1 1 1 1 1 675.906321 113.824158 706.497634 698.727628 772.063497 196.954482 675.906321 113.824158 706.497634 698.727628 772.063497 196.954482 1.79 0.30 1.87 1.85 2.04 0.52 0.2299 0.6032 0.2208 0.2231 0.2031 0.4978 Year 3 control V& oth.ra control. 8 el.k/Jca2 control. v. 15elk/1aI2 control V& 31 .lk/_ l.1n ••",.~~ecta quadratic .~~ecta 1 1 1 1 1 1 773.93561 228.68989 373.24742 1135.82189 1098.26436 28.12600 773.93561 228.68989 373.24742 1135.82189 1098.26436 8.92 2.64 4.30 13.10 12.66 0.0244 0.1555 0.0834 0.0111 0.0119 contraat v. 28.1.2600 0.32 3 0.5897 Figure 15. Effects of elk density on elk utilization of live'perennial grass available at the beginning of the spring grazing season. vertical bars give ± 1 standard error. Test of specific hypotheses are shown in the table of orthogonal contrasts. �233 Forage Class=Perennial -~ ~ Grasses Age=Standing Live 90 w 80 •.... ,.:::l QD 70j ell 1.0 60, ~ 0 ::- ell ~ 0 ... ::: .-~ .--..;::l E- 0 50 40 30 _-----1 t------------r------- ~r----rI-------------------------------------------:-I 0 ell 20 N 10 0 wo:'•• __ .•• _-_ ••• __ .•..•_..... o • ••••••• 5 -_ .•.•. 10 15 20 25 30 35 Elk Density (animals/km2) I Year contrast DF Contrast 1 ------------2 ----- ss All Year. control. v. other. contro~ v. 8 .lle/~ control v. 15 .lle/Iao control v. 31 .lle/Iao lin.ar err.cta quAdratic .~r.cta 1 1 1 1 1 1 Year 1 oontrol. v. other. control. v. 8 elle/~ control va 15 elle/Iao control v. 31 elle/Iao lin.ar errecta quadratiC erreeta 1 1 1 1 1 1 83.269562 13 ••1773. 113.202014 63.924133 72.920421 46.76 •• 44 Year 2 control v. others control. va .lle/~ contro1 v. 15 .lle/Iao control v. 31 .lle/Iao linear erreot. quadratic errect. 1 1 1 1 1 1 1 1 1 1 1 1 • Year 3 cont.rol control control control lin.ar v. v. v. v. .rr.ct. quadratic other. 8 elle/~ 15 .lle/Iao 31 elle/Iea errecta 498.8454173 118.4665741 412.7378664 552.6633931 567.1075782 101.764.429 Mean Square 498.8454173 118.4665741 ,012.7378664 552.6633931 567.1075782 101.764.429 F tJalu. 3 Pr > F 6.48 1.54 5.36 7.18 7.36 1.32 0.0438 0.2612 0.0599 0.0366 0.0349 0.2941 83.269562 13.81773. 113.202014 63.924133 72.920421 46.76 ••• 4 2.41 0.40 3.28 1..5 2.11 1.35 0.1716 0.5505 0.1203 0.2227 0.1966 0.2 •• 9 56.875.031 9.7590.2. 64.39611.3 53.6460500 59.564.610 20 ••539205 56 ••754031 9.7590.2. 6 •• 3961183 53.6460500 59.5648610 20.8539205 2.12 0.36 2.40 2.00 2.22 0.78 0.1959 0.56.7 0.1725 0.2073 0.1.70 0.4122 ...4••05395 14 •• 26107a 273.039161 6.5.100172 624.496136 36.812425 4.4.805395 1.4.26107. 273.039161 645.100172 62 •• 496136 36 ••12425 a.66 2.5. 4.a. 11.52 11.15 0.66 0.0259 0.1596 0.0693 0.01.6 0.0156 0 ••••• Figure 17. Effects of elk density on elk utilization of live perennial grass.produced during the spring grazing s~a~on. vertical bars g1ve ± 1 standard error. Test of spec1flc hypotheses are shown in the table of orthogonal contrasts. �234 ;< 100 ~ Forage Class=Total I Herbs Age=Standing 15 20 Live 90 80 1 ~ ~ 701 ~ 0 '"' 60 ""Qll 50 = .i: e, en 40 ••... 0 30 = 0 :.:; 20 ~ ._- .. IS ~ 10 0 5 0 10 25 30 35 Elk Density (animals/km2) I Year 1 ................. 2 contrast ss X.an Square 1 1 1 1 1 1 10.9.039735 5••4919.4 101.2.415967 1711.323529 2075.77325. 79.620253 10.9.039735 5••491.9.4 1012.415967 1711..323529 2075.77325. 79.620253 Year 1 control. va others control. 8 e~k/1ca2 control. v. 15elk/_2 control. v. 31 .lk/_ lin.ar .rrecta quadratic .rr.eta 1 1 1 1 1 1 400.59.720 234.583704 133.54.602 490.799560 3••••998.3 23.951714 •••• ar 2 control v. other. control v. 8 .~k/1ca2 control v. 1.5.1k/_2 v. 31 .~k/_ control l.in.ar .rrecta quadratic .rr.eta 1 1 1 1 1 1 Year 3 control. v. other. control v. 8 e~k/1ca2 contro.l v. 15e~k/1ca2 control. 31 .~k/lca l.inear .rrecta quadratiC .rrecta 1 1 1 1 1 1 contraat DF year. contro~ v. other. elk/_2 contro~ v. contro~ v. 15.1k/_2 contro.1 31 .~k/Ica lin.ar .rrect. quadratic errecta ----- P Va~\l. Pr > P 2.73 0.15 2.54 4.30 5.21. 0.20 0.1493 0.7147 0.1619 0.0.35 0.0625 0.6705 400.59.720 234.5.3704 133.548602 490.799560 388.899.83 23.951714 2.44 1.43 0.81 2.99 2.37 0.15 0.1692 0.2770 0.4018' 0.1345 0.1746 0.7156 623.965817 122.45351.6 640.284454 61.5 ••75737 663.172472 195.236549 623.965.1.7 122.45351.6 640.284454 61.5.875737 663.172472 195.236549 1..07 0.21 1.09 1.05 1.13 0.33 0.3415 0.6634 0.335. 0.3444 0.3279 0.5.45 147.973.3 172.5351.5 333.10142 609.14638 1118.27967 11.63926 147.97383 172.53515 333.10142 609.14638 1118.27967 11.63926 0.73 0 ••5 1.64 3.00 5.50 0.06 0.4264 0.3925 0.2478 0.1342 0.0574 0.81.9 3 A~~ v. • v. v. Figure 16. Effects of elk density on elk utilization of herbaceous biomass (annual grass + perennial grass + forbs) available at the beginning of the spring grazing season. vertical bars give ± 1 standard error. Test of specific hypotheses are shown in the table of orthogonal contrasts. �235 Forage Class=Total ~ ~ c.J E v ~ ~ 0 Herbs Age=Standing live 100 90 j 80j 70j 60.j ----' 5°1 ~ 0 E- ... ._:: 40 .-'~_N 20 0 30 0 -' ;::l k----------- ..1-~- ..----. .",.. 5 0 1 .....-::::-~;t--.---.--..-.-----..--.--.-.-.-__._. .•T ._..~==.t1;:: 1 10 0 ______ J 10 15 25 20 ~I 30 35 Elk Density (animals/km2) IYear Cofttraat COntrast '---'---'--- 2 P Value ----- 3 Pr > r 58 Mean SqUare 1 1 1 1 1 1 l86.5201431 8.6198144 312.5201998 350.1005845 542.5730463 16.9503603 186.5201431 8.6198144 312.5201998 350.1005845 542.5730463 16.9503603 1.74 0.08 2.91 3.26 5.06 0.16 0.2355 0.7864 0.1388 0.1209 0.0656 0.7048 Year 1 control. v. other. control v. 8 el.k/1aL2 control vs 15 e1k/lal control v. 31 e11</1al linear errecta quadratic .rrecta 1 1 1 1 1 1 46.7139613 23.4546333 33.2353891 37.6218458 30.8083009 15.7492966 46.7139613 23.4546333 33.2353891 37.6218458 30.8083009 15.7492966 1.74 0.87 1.24 1.40 1.15 0.59 0.2356 0.3864 0.3088 0.2816 0.3256 0.4731 Year 2 control va other. control vs 8 ell</1aL2 control va 15 el.l</1ao control.. vs 31 elle/Iao lin •• r .~r.ct.a quadratic errect. 1 1 1 1 1 1 75.1692342 13.9617614 97.7165007 57.9945235 64.5393735 40.6552085 75.1692342 13.9617614 97.7165007 57.9945235 64.5393735 40.6552085 1.25 0.23 1.63 0.96 1.07 0.68 0.3063 0.6470 0.2495 0.3640 0.3401 0.4424 Year 3 control v. other. eontrol v. 8 el.lc/1aL2 eontrol vs 15 elk/lao control vs 31 elk/lao lin.ar errecta quadratic errecta 1 1 1 1 1 1 66.426561 186.725680 224.083877 348.168760 716.142489 10.327788 66.426561 186.725680 224.083877 348.168760 716.142489 10.327788 0.40 1.12 1.34 2.09 4.29 0.06 0.5513 0.3307 0.::1905 0.1986 0.0836 0.8l18 All Year. control vs others control va 8 el.k/1ao2 control vs 15 el.k/Iao control.. vs 31 elk/lao linear e~~.et. quadratic erreeta Df' 1 Figure 18. Effects of elk density on elk utilization of herbaceous biomass (annual grass + perennial grass + forbs) produced during the spring grazing season. vertical bars give + 1 standard error. Test of specific hypotheses are shown in the table of orthogonal contrasts. �236 Forage Class=Perennial Grasses Age=Standing Dead 100.r---------------------------------------~ 901 1 801 70j c c 60 _---: - ...••-......•.... .",.;~;~ 00 .... 30 .. ---- .".... »> .... ",. ---. .>" ',,,/' 20 10 o 5 10 15 25 20 30 35 Elk Density (animals/km2) I Year contrast OF Al.l Year. control. v. o~.rs contro~ v. 8 eUc/k112 control va 15 .lk/Iao control. v. 31 elk/lao linear err_eta quadratic ettects Y.er 1 control v. others control v. 8 .lk/k112 control. v. 15 elk/lao control. va 3],elk/lao linear .~r.ct. quadratic .tt.eta v. v. v. v. Year 2 control other. contro.l a .lk/1ao2 control 15 .lk/Iao contro.l 31 .1k/1ao lin.ar .tt.cta quadratic .tt.cta v. Y.er 3 con1:.rol. others control. va 8 .lk/k112 control. v. 15 .lk/Iao control v. 31 .lle/Iao lin.ar err.eta quadratic ett.eta ]. ]. ]. 1 1 ]. ]. ]. 1 ]. ]. 1 1 ]. 1 1 1 1 ]. 1 1 1 ]. 1 Contra.t SS 3209.554674 ].].54.937],45 8].9.878149 5799.263352 5351.615983 3.225230 1 ...•.••..... 2 ----- Mean SqUare 3209.554674 ],154.937145 8].9.878149 5799.263352 5351.615983 3.225230 F Value Pr > 3 F 7.74 2.79 1.98 ].3.99 12.91 0.01 0.03],9 0.1462 0.2093 0.0096 0.0115 0.9326 8.73 1.37 3.02 1••74 19.77 0.12 0.0255 0.2.69 0.1329 0.0049 0.0043 0.73.6 762.77615 119.35737 264.06424 1638.29524 1727.6473. 10.6.],29 762.776],5 ],],9.35737 264.06424 1638.29524 1727.6473. 10.68].29 849.48449 444.38825 115.36251 1565.89674 ].317.1196. 9.89976 849.4.449 444.38.25 115.36251 1565 ••9674 1317.1196. 9 ••9976 2.23 1.17 0.30 4.12 3.46 0.03 0.1.57 0.3213 0.6017 0.08 •• 0.1121 0.8771 1710.76843 721.3011. 510.93875 2688.77607 2386.38163 10.91665 1710.76.43 721.30118 5],0.93875 2688.77607 2386.38163 10.91.665 5.03 2.1.2 1..50 7.91. 7.02 0.03 0.1955 0.2661 0.0307 0.0380 0.8637 0.0660 Figure 19. Effects of elk density on total utilization (use by elk + cattle) of standing dead perennial grass. Vertical bars give ± 1 standard error. Test of specific hypotheses are shown in the table of orthogonal contrasts. �237 Forage Class=Total Herbs Age=Standing Dead 100 1 90 1 ~ 801 1, 704 ._::c 60] -' <'0 ._ -...• N -_. 50 I 40 ~ c 30 E- 20 10 0 0 5 10 DF Contrast 20 5S Xean square F Va~u. 30 25 Elk Density (animals/km2) 1 ._.__.._-_.-2 ----- IYear Contra.t 15 35 3 Pr > P All Years control v. other. contro.! va 8 elle/ka2 control va 15 elle/lea control va 31 elle/lea ~in.ar e~1!.ct. quadratic .~~.cta 1 1 1 1 1 1 2022 ••31634 707.970840 4•••124365 3778.161235 3507.023206 7.577182 2022 ••31634 707.970840 4•••124365 3778.161235 3507.023206 7.577182 3.64 1.27 0.8. 6.80 6.31 0.01 0.1051 0.3022 0.3.49 0.0403 0.0458 0.9109 ~•.• r 1 control v. other. control va 8 .11e/lea2 control va 15 .lle/lea control va 31 elle/lea linear .~1!.cta quadratic e~~ecta 1 1 1 1 1 1 557.23873 112.26394 125.60198 1297.42052 1316.74219 29.9955. 557.23.73 112.26394 125.60198 1297.42052 1316.74219 29.99558 4.87 0.98 1.10 11.34 11.51 0.26 0.0694 0.3601 0.3351 0.0151 0.0146 0.6269 Vear 2 va others control control "a 8 elle/ka2 control va 15 .lle/Jew control va 31 elle/Jew line.., e~~ecta quadratic e~~ecta 1 1 1 1 1 1 235.851221 103.293322 24.942.69 504.464916 445.773523 13.0454.4 235.851221 103.293322 24.942.69 504.464916 445.773523 13.0454.4 0.52 0.23 0.06 1.11 0.9. 0.03 0.4975 0.649. 0.8222 0.3318 0.3593 0.870. Y.ar 3 control va other. control va 8 elle/ka2 control "a 15 elle/lea control va 31 elle/lea line.., .~~ecta quadratic e~~ecta 1 1 1 1 1 1 1516.10493 641.46327 4.6.89139 2302.41740 2040.50720 1••67030 1516.10493 641.46327 486.89139 2302.41740 2040.50720 18.67030 5.24 2.22 1.68 7.97 7.06 0.06 0.0619 0.1869 0.2420 0.0303 0.0377 0.8079 Figure 20. Effects of elk density on total utilization (use by elk + cattle) of standing dead herbaceous biomass (annual grass + perennial grass + forbs). Vertical bars give ± 1 standard error. Test of specific hypotheses are shown in the table of orthogonal contrasts. �238 Forage Class=Perennial Grasses Age=Standing Live 100.r---------------------------------------~ 901I 80i ............····t··· 70 ::: :I H 60] 1-······· I1-- 50 .-_. 4°1 30 1 z: i " , ······1 I ------tt-------------------i _----- :0 o Eo- 20 10] O~~~~~ o __ ~ __ ~~~~~--~~~~~~~ 5 10 15 20 25 30 35 Elk Density {animals/km2} I Year Contrast OF AU. Year. vs others control contro.l vs II e1.k/1aa2 vs 15 elkin control vs 31 elkin control linear errecta quadratic effects contrast 1 ............ 2 ----- 55 X•••• SqUare F value 1 1.51. •.•• 39905 1.5•• 6.90.26 36.4231.93.• 132.6110063 73 ••20295 .• 21.921111,,77 control vs others control. vs II elk/n2 control vs 15 elk/Ial vs 31 elk/_ control linear effects quadratiC effects 1. 1 1 1. 1. 1. 1.••3 .• 1.372 0 •.• 92.005 1..3675.31 1.••32 .• 6 .• 62 13.5"11.723 ••.•2591.51. 1.113,,1.372 0.492.005 1..3675.31 1. .••32 .• 6 .• 62 13.5"11.723 ••.•2591.51. 0.02 0.01 0.02 0.1.7 0.1.6 0.1.0 Y.ar 2 control v. other. control va II elk/1aa2 contro.l va 15 elk/_ control vs 31 elkin linear effec:t.a quadratic effects 1. 1. 1 1. 1. 1. 1.7•• 1100.7. 1.9".3"1.299 63.2 .••.• 35 117.9"3573 5•• 2.1.61.3 55.01.32711 1.7 •• 1100.711 194.341.299 63.2 .••.• 35 1.17.9"3573 5•• 2.1.61.3 55.01.32711 11..25 12.23 3.9. 7."2 3.67 3.46 Year 3 control vs otb.r. elk/1aa2 control. vs v. 15 elkin control control v. 31. elkin linear effects quadratiC .rrecta 1. 1. 1. 1. 1. 1. .• 3."5355.9 51 •.• 99755 .• 1.3.4671.21.5 2.·.09.2.• 25 1.2.7269796 12.9:JS0962 .• 3 •.• 5355.9 51.•.• 99755 .• 1.3."671.21.5 2••09.2"25 1.2.7269796 12.9350962 Year 1. 1. 1 1 1. 1.51•.•• 39905 1.5 •• 6.90.26 36 •.• 231.93.• 132.6110063 73.11202954 21..9211."77 3 Pr > P 1..24 1..30 0.30 1.09 0.61 0.1.11 0.307. 0.2975 0.60"5 0.3373 0."662 0.6.6 .• 0.11••0 0.941.. 0.9032 0.6957 0.7036 0.7636 .1 • 0.63 0.75 0.20 0."1. 0.1.11 0.19 0.0153 0.0129 0.0931 0.03"5 0.1.0"0 0.1.1.22 0 •.• 57 .• 0."205 0.6739 0.5"67 0.6.24 0.6.00 Figure 21. Effects of elk density on total utilization (use by elk + cattle) of live perennial grass. vertical bars give ± 1 standard error. Test of specific hypotheses are shown in the table of orthogonal contrasts. �239 Forage 100] Class= Total Herbs Age=Standing Live 90 j 80 ..•._~c N so I __ j 1 501 ------j--------------.- .. -I.... -----.- ..--... -..... ----... -.- ···--·---·1 ------.-------- r: T. 40 r . I 1 1'..... -" 1-- <; L /,// 30 _ -------------- '( 20 10 o 5 10 15 20 25 30 35 Elk Density (animals/km2) I Year OF COntrast Years eontrol v. other. eontrol v. 8 e~k/D4 eontrol v. 15 elk/1cwI control. vs 31. e~k/Ial 11n •• r err.ct. quadratic .rrecta Contraat ss 1 X.an .---------.- 2 ----- 3 Pr > F Square F value 3.1705653 35.8760060 106.2739624 0.001.7904 11.733~752 44.1.432630 3.1705653 35.&760060 106.2739624 0.001.7904 11.7331.752 44.1.432630 0.01 0.10 0.30 0.00 0.03 0.13 0.9275 0.7605 0.602& 0.99&3 0.8612 0.7355 1 1 1 1 1 1 29.8476507 6.3093769 33.7684775 25.5973844 27.4187952 12.6736876 29.&476507 6.3093769 33.7684775 25.5973844 27.41&7952 12.6736.76 0.56 0.12 0.63 0.48 0.51 0.24 0.4838 0.7433 0.4576 0.51.54 0.5014 0.6441 1 1 1 1 1 1 155.064081 102.070530 159.507443 60.366419 38.996241 136.714032 155.064081 102.070530 159.507443 60.366419 38.996241 136.714032 1.09 0.72 1.13 0.43 0.2& 0.97 0.3357 0.4285 0.3294 0.5380 0.61&6 0.3638 1 1 1 1 1 1 219.978420 528.504963 0.342350 162.707372 30.781127 14.022595 219.97&420 528.504963 0.342350 162.707372 30.781127 14.022595 0.36 0.&7 0.00 0.27 0.05 0.02 0.5698 0.3875 0.9819 0.6237 0.8296 0.8844 1t.~~ Year 1 control control control control I1n.ar v. other. v. 8 elk/D4 v. v. .rr.ct. quadr.tic 15 31 e~lc/1cwI e~lc/Ial err.ee. Y •• r 2 control. other. eontrol v. 8 e~lc/D4 control 15 e~k/Ial eontrol v. 31. e~k/Ial line.,...rrecta quadratic errecta v. v. v. v. v. v. Year 3 control others control 8 e~k/D4 contro.l 15 e~k/Ial control. 31 e~k/Ial lin •• r .:rt'ect. errecta quadratic 1 ~ 1. 1. 1 1. Figure 22. Effects of elk density on total utilization (use by elk + cattle) of live herbaceous biomass (annual grass + perennial grass + forbs). vertical bars give ± 1 standard error. Test of specific hypotheses are shown in the table of orthogonal contrasts. �240 Forage Class=Perennial --' co 50 0 E- 40 ... 0 I1.l 30 t:lJl -' 20 L I1.l Co) l.o I1.l 10 0.. 0 Age=Standing Live ----I co = Grasses ........ . ........ I .... --- .•_-- ..••..._-- ...-_ ... _ 0 + ~ ~~ ~~ ~~~r-------------- 1-_ ...•- .........•••.•....•...----1---------------------------------------------1" 5 10 15 20 30 25 35 Elk Density (animals/km2) I Year Contra.t OJ! v. v. v. v. All Years others control 8"ellc/ka2 eontrol. 15 ellc/a control control. 31 ellc/a linear el!"l!"ect. quadratie el!"l!"ecta v. v. • v. v. Year 1 control other. elk/ka2 control control 15 ellc/a contro~ 31 ellc/a linear el!"l!"eeta quadratie el!"reeta v. v. • v. v. Year 2 control other. elk/ka2 eontrol control. 15 elk/a control. 31 ellc/a lJ.n•• r .l!"reeta quadratie .l!"l!"eeta v. v. v. v. • Y.ar 3 control. oth.Z'. .11c/ka2 eontrol control 15 .llc/a eontrol 31 .llc/a linear .l!"recta quadratie .l!"l!"ecta COntraat 1 ------------2 ----ss M.an Square 7! Val •• e Pr> 3 P 1 1 1 1 1 1 49._.45 •• 173 11•• 4665741 412_737.66 •• 552.6633931 567.10757.2 1.01_76".429 49•••••5•• 173 11 •• 4665741 "1.2.737.66" 552.6633931 567.10757.2 101..76••••• 29 6.4. 1.54 5_36 7.1. 7.36 1.32 0.0 •• 3. 0.261.2 0.0599 0.0366 0.03"9 0.2941. 1 1 1 1 1 1 .3.269562 13 ••1773. 11.3.20201" 63.9241.33 72.920"21. 46.76."44 .3.269562 13 ••1773. 113.202014 63.924133 72.920421 46.76."44 2.41 0.40 3.2. 1 ••5 2.11 1.35 0.1716 0.5505 0.1203 0.2227 0.1966 0.2 ••9 1 1 1 1 1 1 56 ••75 •• 031 9.7590.2. 6••• 39611.3 53.6"60500 59.56".610 20_.539205 56 ••754031 9.7590.2. 6••• 39611.3 53.6"60500 59.56".610 20 ••539205 2.12 0.36 2.40 2.00 2.22 0.7. 0.1959 0.56.7 0.1725 0.2073 0.1.70 0.4122 1 1 1 1 1 1 ••••••• 05395 14"_261.07. 273.039161 6"5.100172 62".4961.36 36_.12425 4•••••05395 14 ••• 26107. 273.039161 645.100172 62 ••••• 96136 36 ••12425 ••66 2.5. 4.a. 11.52 11.15 0.66 0.0259 0.1596 0.0693 0.0146 0.0156 0 •••••••• Figure 23. Effects of elk density on the proportion of total utilization (use by elk + cattle) of live perennial grass that resulted from elk grazing. vertical bars give ± 1 standard error. Test of specific hypotheses are shown in the table of orthogonal contrasts. �241 Forage Class=Perennial Grasses Age=Standing Live 100 1 1 ~ --'-' ".) 90J I 801 I j :: 601 u >. 50 C -'~ 401 :-l 30 :::l 20 .--' ·········-1 .s->: _._._•••.•••.• - •...•.•..•.•.. , .•••.. :••••••••••••.•••••••.••••• - I ...::l C I·· 70i }-r---__ f---___ + __ r --- -- -_ -I .•..•.--s-. _ ---------------1 10 0 0 5 10 15 20 25 30 35 Elk Density (animals/km2) I Year Contrast Year. control v. other. control va 8 .~k/ka2 contro~ va 1.5 e~k/kJI contro~ va 31. e1.k/kJI ~in.ar .:r~.ct. quadratic el!'l!'ects Year 1 control va other. contro.! va 8 .1.k/ka2 control v. 1.5 e1.k/kJI control v. 31. e1.k/1al linear err.eta quadratic .l!'l!'ecta A~~ Year 2 control va other. control v. 8 .1.k/ka2 control va 15 e~k/kJI control v. 31 .1.k/kJI 1.in.ar el!'l!'ecta quadratic .l!'l!'ecta Year 3 control v. other. control va 8 .1.k/1aL2 control va 1.5 .1.k/kJI eontro.l v. 3~ .~k/kJI lin •• r .~t'ecta quadratiC el!'l!'ecta 1 ...•........ 2 Contraat ss lI.an Squar. 1. 1. 1. 1. 1. ~00.5,,00~38 2.93"1.990 203.9"09045 1.43.8346569 231..71.37538 29.21."36,,9 ~00.5,,001.38 2.93"1.990 203.9409045 1.43.83"6569 231..7137538 29.21..3649 1. 1. 1. 1. 1. 1. 60.3.7081. 9.091.578 1.39.45.384 1.7.72.020 23.61..860 94 ••9659. 1. 1. 1. ~ OF ~ ~ ~ ~ 1. 1. 1 1 1 l! Va~u. ----- 3 Pr > F 2.05 0.06 4.~5 2.93 4.72 0.59 0.2025 0.81.51. 0.0877 0.1.379 0.0729 0 ••699 60.387081 9.091578 1.39••5"3.4 1.7.72.020 23.61.••60 94 ••96598 0.99 0.~5 2.28 0.29 0.39 1..55 0.3587 0.71.31. 0.1.81.. 0.6097 0.5572 0.2593 33.989673 1.1.7.00060. 0.005160 ~2.S021.40 0.0069.7 •• 1.25261. 33.989673 1.1.7.000604 0.0051.60 1.2.5021..0 0.006987 8.125261. 1.••• 5.10 0.00 0.5" 0.00 0.35 0.2693 0.0647 0.9.85 0.4 •• 3 0.9.66 0.5735 237.972752 23.372780 1.65.22••• 9 403.93~401. .5•• 92056. 6.~0477. 237.972752 23.3727.0 165.228 •• 9 403.931.01. 45•. 920568 6.~04778 1.2.53 ~.23 •. 70 2~.27 206.~7 0.32 0.01.22 0.3097 0.0256 0.0036 0.0027 0.5913 .0 Figure 24. Effects of elk density on cattle utilization of live Test of perennial grass. Vertical bars give ± 1 standard error. specific hypotheses are shown in the table of orthogonal contrasts. �242 Forage Class=Tot al Herbs Age=Standing Live 100 90j ,0 '" --'-' 1) :tl u >. ..c ::: 80J , J 70 1 ]._ _ ::1 _+._ -+ __ _.. r 1 c ~ 40j := .--.-' N 30 - 20 -, 10 0 5 0 10 15 20 25 Elk Density (animals/km2) I Year 1 -----------2 ----- DP Contraat ss X.an Al.l. Year. contro~ va other. contro~ va 8 .l.lc/1cJI2 eontro1 va 15 .l.lc/1aI control. va 31 el.lc/1aI linear .~rec:t. quadratic .rr.eta 1 1 1 1 1 1 141_0543311 9.3251216 54.3067964 348.5189300 394.7304766 6.3855174 141.0543311 9.3251216 54.3067964 348.5189300 394.7304766 6.3855174 Year 1 control. va other. contro.l va 8 .l.lc/1aI2 control. va 15 .l.lc/1aI control va 31 .l.lc/1aI l.in.ar e~r.cta quadratic .~~ecta 1 1 1 1 1 1 1.88090498 5.43426527 0.00212068 1.15407577 0.09.73838 0.166.9353 y•.•. r 2 control. va other. control. va 8 .l.lc/1cJI2 control va 15 el.lc/1aI control "a 31 .l.lc/1aI 1. inear e~~ecta quadratic .~~ecta 1 1 1 1 1 1 14.3069554 40.5317115 Y.ar 3 control. va other. control va 8 el.lc/1ca2 control va 15 el.lc/1aI eontro.l va 31 el.lc/1aI 1. inear e~recta quadratic e~recta 1 1 1 1 1 1 Contrast Square P Val.". 35 3 Pr > F 0.65 0.04 0.25 1.60 1.81 0.03 0.4520 0.8430 0.6356 0.2531 0.2271 0.8698 1.88090498 5.43426527 0.00212068 1.1.5407577 0.09873838 0.16689353 0.04 0.10 0.00 0.02 0.00 0.00 0.8572 0.7603 0.9952 0.8878 0.9671 0.9572 0.0237682 3.2003173 28.263405. 14.3069554 40.5317115 7.5320602 0.0237682 3.2003173 28.2634058 0.1.7 0.48 0.09 0.00 0.04 0.33 0.695. 0.5156 0.7760 0.9872 0.8525 0.5851 528.16859 .6.94584 241.94365 9.6.89976 1043.86585 0.2.195 528.16859 86.945'4 241.94365 986.89976 1043.86585 0.28195 1.36 0.22 0.62 2.54 2.69 0.00 0.2880 0.6530 0.4602 0.1622 0.1524 0.9794 7.5320602 30 Figure 25. Effects of elk density on cattle utilization of live herbaceous biomass (annual grass + perennial grass + forbs). Vertical bars give ± 1 standard error. Test of specific hypotheses are shown in the table of orthogonal contrasts. �243 Forage Class=Perennial Grasses Age=Slanding Live 2 0.40r-----------------------------------------~ ;:::-0.38 '-0.36 ~ 0.34 -'::0.32 ~0.30 ~ 0.28 ::; 0.26 2'30.24 3°·22 _ 0.20 ~ 0.18 E 0.16 ~ I v 0.14 0::0.12 ~0.10 ~ 0.08 ~0.06 ...~ :r... ••. .s-> i- I .•..•.. ......••..••....••.••.......•.....••.....•..••...••.....•••••..... I ............ t ...... _ -------~-------- J ---------1 0.04~~--~----~~--~----~----~--~~----~ o 5 10 15 20 25 30 35 Elk Density (animals/km2) I Year Contra.t Of' --- Contraat S5 1 Kean .•••_._._...2 ----- P Value Pr > Square 3 r All Year. eontrol vs other. control vs 8 elk/Jca2 control va 15 elk/klo control vs 31 elk/klo l1n.ar err.ct. quadratic et't'ecta 1 1 1 1 1. 1 2.98963061 0.05.87150 1.58210222 7.47941064 9.031.41804 0.09059563 2.98963061 0.05887150 1.58210222 7.47941064 9.03141804 0.09059563 1.36 0.03 0.72 3.41 4.1.2 0.04 0.2.71 0.8752 0.4281. 0.1142 0.0887 0.8456 Y.ar 1 control vs others control v. 8 elk/klo2 control vs 1.5elk/klo control. v. 31 elk/klo linear .rrect. quadratic .rrecta 1 1 1 1. 1 1 5.8454912 1.5789892 1.2553817 12.5686217 12.2287822 0.1.812986 5.8454912 1.5789892 1.2553817 1.2.568621.7 1.2.2287822 0.1812986 1.75 0.47 0.38 3.76 3.66 0.05 0.2343 0.51.77 0.5626 0.1006 0.1.044 0.8236 2 control v. other. control. v. 8 elk/Jca2 control v. 1.5e1.k/klo control vs 31 elk/klo lin •• r ert'ecta quadratic errecta 1 1 1 1 1 1 0.87246429 3.1.4115391 0.01610665 0.1.511.0586 0.04087717 0.49295989 0 ••7246429 3.1.41.1.5391 0.01610665 0.1.511.0586 0.04087717 0.49295989 0.1.8 0.66 0.00 0.03 0.01 0.10 0.6832 0.4471 0.9555 0.8643 0.9291 0.7582 1 1 1 1 1 1. 2.2.349490 0 ••761090. 1.40440466 2.49766183 2.26823336 0.36792826 2.28349490 0.87610908 1.40440466 2.497661.83 2.26823336 0.36792826 7.64 2.93 4.70 8.36 7.59 1.23 0.0326 0.1.376 0.0732 0.0276 0.0330 0.3096 Y.ar v. v. v. Y.ar 3 control other. control v. 8 elk/Jca2 control 1.5elk/kII control 31..1k/kII lin •• r err.eta quadratic .rrecta Figure 26. Effects of elk density on removal (intake + trampling) of live perennial grass by cattle. vertical bars give ± 1 standard error. Test of specific hypotheses are shown in the table of orthogonal contrasts. �244 Forage Class=Total Herbs Age=Standing Live 0.6 '0 <, ,.. 0,5 - 0.4 N :: <, ~ ~ ....• ....• :'0 u .••••..••...••.....•••..•.......•.....••.........•.••.......•.............•......... 0.3 >-. ,.Q <:Il .•....•... 0.2 > 0 6 Q.) 0.1 .................... t ' ~-------------- J ----------1 c:: Q.) bD 0: ••0 0.0 ~ -0.1 0 5 10 15 20 25 30 35 Elk Density (animals/km2) I Year Contra.t OF Years control VII others contro~ VII e~k/1aa2 control VII ~5 .~k/lQI control VII 3~ e~k/n linear e-rteeta quadratic e~~ect. A~~ " ~ Y.ar control va others control V. 8 elk/lcJo2 control V. ~5 elkin control V. 31 •• ~k/n linear ettect. quadratic e~~ecta Year 2 control V. others control V. elk/n;ol control V. ~5 elkin control V. 31 elkin ~inear e~~ect. quadratic e~~ecta • Y.ar 3 control V. other. control. v. 8 elk/lcJo2 control. v. ~5 elk/n control v. 31 elk/n ~inear .~~ecta quadratiC e~~ecta ~ ~ ~ ~ ~ ~ Contrast --- ss 1.2oo2479~ 2.~7722749 1 .•.......... 2 ----- Mean Square P Value 3 Pr > P 8.55739~96 5."~979833 6.55285327 ~.2002479~ 2.~7722749 2.94907263 ".55739~96 5.8~979833 6.55285327 O.~O 0.~9 0.25 0.74 0.50 0.56 0.7589 0.6804 0.6325 0.4240 0.5059 O.48~3 0.55967568 0.06866966 6.84439401 0.27209880 0.023.5962 6.56847783 0.55967568 0.06866966 6.84439401 0.27209880 0.02385962 6.56847783 0.07 O.O~ 0.85 0.03 0.00 0.81 0.8013 0.9296 0.3930 0.8605 0.9584 0.4021 ~ 2.~399662 0.7603231 9.85~0021 0.1826105 0.1577694 12.266.994 2.~399662 0.760323~ 9.85~0021 0.1826105 0.1577694 ~2.2668994 0.18 0.06 0.84 0.02 0.01 1.04 0.6843 0.8076 0.3950 0.904. 0.9~15 0.3461 1 1 1 1 1 1 ~6•••oo9.0 ~0.0212246 7.730527;01 16.9563705 ~3.15350.~ 2.661.578 ~6.88oo980 ~0.02~2246 7.7305272 ~6.9563705 13.1535081 2.84 ~.69 ~.30 2 ••6 2.;012 0.45 0.1427 0.24~5 0.2973 0.~420 0.~.72 0.5;01110 ~ ~ ~ ~ ~ 1 1 ~ ~ 1 1 2.94907263 2.6618578 Figure 2~. Effects of elk density on removal (intake + trampling) of herbaceous biomass (annual grass + perennial grass + forbs) by cattle. Vertical bars give ± 1 standard error. Test of specific hypotheses are shown in the table of orthogonal contrasts. �245 Forage Class=Perennial _ Grasses Age=Standing Dead 0.16 0.15 'N :: 0.14 0.13 <, 0.12 ~ 0.11 ..:: 0.10 0.·09 0.08 £ .----- ·······_ ..· 1· g:g~ <, • 0.05 i 0.04 ~ 0.03 0:: 0.02 ~ 0.01 ~ 0.00 ~ -0.01 -0.02~~~~~~~~~~ "--'" <, ~ s t--------t-----~--------_~~~~···~···········.11 0 5 -------~1 ~~ 10 __ ~~ 20 15 25 ~ ~ 30 35 Elk Density (animals/km2) I Year 1 -----------2 ----- Contra.t ss X.an 1 1 l. l. l. 1 14.14093013 3.21423414 8.79616897 19.82479149 20.26095478 0.96845746 14.14093013 3.21423414 8.796l.6897 19.82479l.49 20.26095478 0.96845746 7.61 1.73 4.74 10.67 10.9l. 0.52 Vear l. control v. other. control v. 8 81k/02 control v. 15 .1k/laa control. v. 31 .1k/ka 1in •• r erreeta quadratiC .rrecta l. 1 1 1 1 1 8.0424847 0.8677961 10.6707754 7.5537260 9.0774631 3.3735283 8.0424847 0.8677961 10.6707754 7.5537260 9.0774631 3.3735283 7.23 0.78 9.59 6.79 8.16 3.03 0.0361 0.4112 0.0212 0.0404 0.0290 0.1323 Vear 2 control v. other. contro~ v. 8 .1k/02 control v. 15 .1k/ka contro.! v. 31 .1k/a 11n.ar errect. quadratiC .rrioc:ta 1 1 1 1 1 1 9.6538543 3.7260751 2.1417777 17.7125711 16_1520238 0.03559.6 9.6538543 3.7260751 2.1417777 17.7825711 16.1520238 0.0355986 3.71 1.43 0.82 6.84 6.21 0.01 0.1023 0.2765 0.3992 0.0399 0.0470 0.9107 Vear 3 control v. other. 8 .1k/02 contro1 control l.5.1.k/a control 31 .1.k/a 11n •• r etteeta quadratiC .rrecta 1 1 1 1 1 1 0.32522533 0.05924.41 0.16554957 0.557441.74 0.58445366 0.00318937 0.32522533 0.05924841 0.l.6554957 0.55744174 0.58445366 0.00318937 2.74 0.50 1.39 4.70 4.92 0.03 0.1489 0.5063 0.2823 0.0733 0.0683 0.8752 Contrast DP Years control va other. contro~ v. 8 e1k/02 control v. l.5e1k/a control va 3l.e1k/ka 11near e~t.ct. quadratiC erreeta Square P Va~u. 3 Pr > P A~~ v. v. v. 0.0329 0.2364 0.0724 0.0171 0.0l.64 0.4974 Figure 28. Effects of elk density on removal (intake + trampling) of standinq dead perennial grass by cattle. vertical bars give ± 1 standard error. Test of specific hypotheses are shown in the table of orthogonal contrasts. �246 Forage Class=Total Herbs 0.181 0.171 -e <, 0.161 N 0.151 0.141 <, 0.131 QD 0.121 i!; 0.111 ::: 0.101 -' ~ u 0.091 >. 0.081 ~ 0.071 0.061 (!l > 0.051 0 0.041 v 0.031 0:: 0.021 v 0.011 QD 0.001 (!l ~ -0.009 '- -0.019 -0.029 Age=Standing Dead a a ~---+---o 5 I Year Contrast DP Contrast SS Al.l.Year. eontrol. va other. control va 8 el.k/D2 control. va 15 el.k/leJD control. va 31 el.k/1nI linear err.ct. quadratic e~~ecta 1 1 1 1 1 1 Yea.r 1 v. others control control. v. 8 el.k/D2 control v. 15 elk/1nI contro.l v. 31 .l.k/1nI linear .~~.cta quadratic err.eta 1 1 1 1 1 1 11.36637"7 1.55060"5 16.01,,7389 9.06701"6 10.69.9170 6.251"323 Y •• ", 2 contro1 v. other. contro~ 8 .l.k/D2 control v. 15 el.k/1nI control 31 .1.k/1nI 1.1n.a", .~~ect. quad",atic .~~.cta 1 1 1 1 1 1 Ye.., 3 control v. other. control. v. 8 .1.k/D2 control 15 el.lc/1nI control 31 .1.k/1nI 1in •• r .~~acta .~~acta quadratic 1 1 1 1 1 1 v. v. v. v. 22.30"70565 5.26771120 15.09237238 29.03"59160 29.56826921 2.303.•.• 050 10 15 20 25 Elk Density (animals/km2) 1 Mean Square F Va~u. Pr> 0.0174 0.1652 0.0368 0.0100 0.0096 11.3663747 1.55060"5 16.01"7389 9.06701"6 10.6989170 6.251"323 11.30 1.5.• 15.93 9.02 10.6" 6.22 0.0152 0.2606 0.0072 0.0239 0.0172 0.0"69 16.377563. 5.375.320 ".6586515 29.5"91690 27 ••156076 0.0052••1 16.3775638 5.3758320 4.6586515 29.5"91690 27.8156076 0.0052.81 6.02 1.98 1.71 10.&6 0.0"96 0.2095 0.5.03075. 0.1693312" 0.32329962 0.5.030758 0.76274"90 0.04047418 0.1693312" 0.32329962 0.7.47783. 0.76274490 0.0"047418 10.22 0.00 4.45 1.30 2.48 6.02 5 ••5 0.31 35 3 P 10.57 2.50 7.15 13.76 14.01 1.09 0.7 ••• 7783. 22.30"70565 5.26771120 15.09237238 29.03"59160 29.56826921 2.303.•.• 050 2 ----_ 30 0.336" 0.2386 0.0165 0.0187 0.9663 0.079" 0.2979 0.166 •• 0.0"96 0.0519 0.5975 Figure 29. Effects of elk density on removal (intake + trampling) of dead herbaceous biomass (annual grass + perennial grass + forbs) by cattle. Vertical bars give ± 1 standard error. Test of specific hypotheses are shown in the table of orthogonal contrasts. �247 Forage Class=Total Cover 2 50 ~ 40 :> e c..; >, ;-30 c :0= U J:::::::::"'"",.I-.... - -.t ------ ···---H --------4t--------------- 20 o 5 10 15 20 25 30 35 Elk Density {animals/km2} YEAR Contrast 1 .----- 2 --- 3 lI.an SqWlr. P Val.••• 2114.7018126 11 .•• 3343004 2211.7779789 240.63385110 214.9306032 96.5266871 40.23 16.16 32.33 34.01 30.37 13.6" 0.0007 0.0070 0.0013 0.0011 0.0015 0.0102 394.3948911 117.773307 294.584713 425.577572 414.180467 78.143150 394.3941198 117.773307 294.584713 425.577572 414.180467 711.143150 11.70 3.49 8.74 12.63 12.29 2.32 0.0141 0.1108 0.0254 0.0120 0.0127 0.17117 1 1 1 1 1 1 0.027586511 0.19245666 0.2111104501 0.095311302 0.30655470 0.012.4671 0.027586511 0.19245666 0.281104501 0.095311302 0.30655470 0.012.4671 0.00 0.03 0.05 0.02 0.05 0.00 0.9490 0.11661 0.11366 0.9054 0 ••316 0.9652 1 1 1 1 1 1 114.63211640 65.71115077 72.21272.3 35.1631581 20.1390304 65.0256343 114.632.640 65.71.5077 72.2127283 35.16315111 20.1390304 65.0256343 4.97 3.116 4.24 2.07 1.1. 3.112 0.0673 0.0971 0.0.51 0.2007 0.3185 0.09115 OF Contra.t ss All Year. control va others contro~ va II .1k/ka2 control va 15 elk/lao control VIS 31 elk/lao 11n.a.r e~~ect. quadratic effect. 1 1 1 1 1 1 2114.70111126 114.3343004 228.77797119 240.633115110 214.9306032 96.52661171 Y.ar 1 control vs others control vs 8 .1k/ka2 control va 15 elk/lao control va 31 e1k/1ao linear effects quadratic effects 1 1 1 1 1 1 Year 2 control vs other. control vs 8 .1k/ka2 control vs 15 elk/lao eontrol VIS 31 elk/lao linear effects quadratic effects Year 3 control vs other. control vs elk/ka2 control vs 15 el.k/Iao control vs 31 elk/lea linear effects quadratic effects • - Pr > F Figure 30. Effects of elk density on canopy coverage of herbaceous biomass (annual grass + perennial grass + forbs) at the end of the spring grazing season. vertical bars give ± 1 standard error. Test of specific hypotheses are shown in the table of orthogonal contrasts. �248 Forage Class=Annual Grass 17 16 15 1413 -12 ~11 :... il) >0 u 10 9 >. 8 7 c 6 :tl U 5 4 Co 0 3 2 1 0 5 0 15 10 20 25 30 35 Elk Density {animals/km2} YEAR COntrast A1.1.Year. contro.! v. others control. v. 8 e1.k/-.z control v. 15 elk/lao control v. 3], e1.k/1ao ~in.ar .rreet. quadratiC e~~ecta Dl" ]. 1 1. ]. 1. 1. contrast 55 - 1 .•.•.• 2 Xean SqUare 37.55320290 35.02282750 29.46399955 25.4231.1.011 1.8.9682451.3 25.03768704 37.55320290 35.02282750 29.46399955 25.423].1.01], 18.9682451.3 25.03768704 --- F Value 3 Pr > 1" 4.10 3.82 3.21. 2.77 2.07 2.73 0.2922 0.30],], 0.3239 0.3443 0.3868 0.3464 1.04.993802 90.1.65598 84.8271.35 74.970905 59.004997 65.933473 7.31. 6.28 5.90 5.22 4.1.], 4.59 0.2256 0.24],8 0.2485 0.2627 0.29],8 0.2780 V •• r 1. control v. other. control. v. 8 elk/1aI2 control v. 1.5 el.k/Iao control. va 3], el.Jc/1ao linear e~~eeta quadratic e~~eeta 1. 1. 1. 1.04.993802 90.165598 84.827],35 74.970905 59.004997 65.933473 Vear 2 contro.l va other. contro.! va 8 el.k/-.z control va ].5 el.k/Iao contro.! va 3]. el.k/Iao lin •• r e~~ect. quadratiC e~~eeta 1. 1. 1. 1. 1. 1. 0.1.1427738 0.52489969 0.0271.4899 0.00276740 0.02431.1.35 0.26979644 0.],1.427738 0.52489969 0.027],4899 0.00276740 0.0243],],35 0.26979644 0.05 0.25 0.01. 0.00 0.01. 0.1.3 0.8543 0.7052 0.92.0 0.9769 0.9319 0.7.].2 ]. 0.000.6585 0.0009141. 0.0007],651 0.0004.625 0.00032302 0.00075169 0.00086585 0.00091.418 0.0007].65]. 0.00048625 0.00032302 0.00075169 91..41. 96.52 75.65 5],.34 34.1.0 79.36 0.0663 0.0646 0.0729 0.0 •• 3 0.1.0.0 0.07],2 Y.ar 3 control va other. control. va 8 elk/-.z control va 1.5 elk/lao control v • 3]. e1.k/1ao .lin•• r e~:feeta quadratic .ereeta ]. ]. ]. 1. 1 ]. 1 ]. Figure 31. Effects of elk density on canopy coverage of annual grass at the end of the spring grazing season. Vertical bars give ± 1 standard error. Test of specific hypotheses are shown in the table of orthogonal contrasts. �249 Forage Class=Perennial Grass 14 13 12 11 _10 );~ .'- •.. QJ >- 9 8 0 u 7 >.. 0. 6 0 c 5 :c u 4 1······················ 1----__ 3 2 1 ················· ·····1····· . ----t----- ---------- +1 . ·················-I ---- --- ] ---1 0 0 5 10 15 25 20 30 35 Elk Density (animals/km2) YEAR Contrast Dl" - 1 Contr ••t 5S Mean SqUare .•.•.• 2 P Va~u. --- 3 Pr > 1" 1o.l.l. Year. contro.l va other. contro1 va 8 el.k/~ contro~ va 15 el.k/n contro.l v. 31 el.k/n 1in •• r .~-r.ct..s quadratic .~~ecta 1 1 1 1 1 1 0.00621144 1.99284448 2.71596508 10.58038808 14.50247489 15.99523287 0.00621144 1.99284448 2.71596508 10.58038808 14.50247489 15.99523287 0.00 0.58 0.79 3.06 4.19 4.63 0.9676" 0.4765 0.4096 0.1309 0.0865 0.0750 Year 1 va other. control control v. 8 el.k/~ contro.! v. 15 elkin control v. 31 .l.k/_ l.ln ••r e~~acta quadratic .~~ecta 1 1 1 1 1 1 0.210574. 9.4623374 0.6877041 7.735'422 15.9662046 14.0105597 0.2105748 9.4623374 0.6877041 7.7358422 15.9662046 14.0105597 0.07 3.06 0.22 2.50 5.16 4.53 0.8029 0.1309 0.6540 0.1649 0.0635 0.0774 va other. contro.! v. 8 .l.k/~ control. v. 15 .l.k/_ contro.! v. 31 .l.k/n lin •• r .~rec:t. quadratic .~racta 1 1 1 1 1 1 0.09272954 0.05915866 3.72415676 2.03668851 1.936'9716 7.2007'969 0.09272954 0.05915866 3.72415676 2.03668851 1.93689716 7.20078969 0.08 0.05 3.18 1.74 ]..65 6.15 0.7879 0.8297 0.1249 0.2354 0.2459 0.0479 Y.ar 3 control v. other. control v. 8 el.k/~ contro.! v. 15 .l.k/n control v. 31 .a/Iao l.ln •• r erracta quadratiC .rracta 1 1 1 1 1 1 0.809'3284 0.76424804 0.00909450 2.0319520' 1.46051563 0.25067455 0.80983284 0.76424.04 0.00909450 2.0319520. 1.46051563 0.25067455 0.53 0.50 0.01 1.34 0.96 0.17 0.4923 0.5043 0.9408 0.2909 0.3642 0.6983 Y.ar 2 control Figure 32. Effects of elk density on canopy coverage of perennial grass at the end of the spring grazing season. Vertical bars give ± 1 standard error. Test of specific hypotheses are shown in the table of orthogonal contrasts. �250 Forage Class=Shrub 29 28 27 26 25 _24 ~23 ,",22 ~ 21 820 >'19 0. o 18 al 17 u 16 15 14 13 12 11 5 0 10 15 20 25 30 35 Elk Density (animals/km2) YEAR Contrast - 1 .•.••• 2 -_. 3 ,. DP Contra.t 55 Xean SqUare P Va~u. lll. years contro~ other. control v. 8 el.k/1<a2 control v. 15 e1k/1nI control va 31 el.k/1nI 11near ertecta quadrat1c errecta 1 1 1 1 1 1 107 ••96165. 66 ••943027 134.5477783 32.0949135 20.6.28141 122.6617930 107.896165. 66.8943027 134.5477783 32.0949135 20.6.28141 122.6617930 15.62 9.69 19.48 4.65 3.00 17.76 0.0075 0.020. 0.0045 0.0745 0.1342 0.0056 Year 1 control v. other. v. 8 el.k/1<a2 control control 15 el.k/1nI control 31 el.k/1nI 11near err.cta quadratic ertects 1 1 1 1 1 1 86.6968369 54.7192451 73.9467323 46.3.97612 32.5130263 53.7411101 86.6968369 54.7192451 73.9467323 46.3897612 32.5130263 53.7411101 14.46 9.13 12.34 7.74 5.42 8.97 0.0089 0.0234 0.0126 0.0319 0.0587 0.0242 Year 2 control v. other. contro.! 8 al.k/1<a2 control v. 15 a1k/1nI control v. 31 al.k/1nI l.1near attects quadratic ettecta 1 1 1 1 1 1 0.10102930 0.17972678 5.54215152 1.32632669 0.64876543 6 ••08l.l.95O 0.10102930 0.17972678 5.54215152 1.32632669 0.64876543 6.8081l.950 0.02 0.03 1.06 0.25 0.12 1.30 0.8942 0.8593 0.3439 0.6332 0.7372 0.2983 1 1 1 1 1 1 69.9294057 5l..738611. .3.4933737 17.24855.9 ••••35211 .5.4307350 69.9294057 51.7386114 83.4933737 17.24.5589 8.8835211 85.4307350 3.72 2.75 4.45 0.92 0.47 4.55 0.1019 0.14.1 0.0796 0.3749 0.5173 0.0769 v. v. v. v. v. v. Y.ar 3 control. other. control 8 al.k/1<a2 control 15 el.k/1nI control 31 al.k/_ 11naar ettecta quadratic attects v. v. Pr P Figure 33. Effects of elk density on canopy coverage of shrubs at the end of the spring grazing season. Vertical bars give ± 1 standard error. Test of specific hypotheses are shown in the table of orthogonal contrasts. �251 Forage Class=Perennial Forb 61 5J j j 1 1 ~ 4. ;jJ ":"'1 > o· '[ u3 ~ 0 __1__ ~c? U I 1 fr ~~~-~ r ~--~-~_ ------- .::.·.·.····_···-1-··.............. I::::':'·-::=~-:::::.-:::::,,· fi ·························:::::1 •. 0 0 5 10 15 25 20 30 35 Elk Density (animals/km2) YEAR - 1 "'." 2 __ a 3 Dr Contrast 5S Xean Square A11 Y.ar. control. v. others contro.! va 8 e1k/n2 control v. 15 e1k/n control. va 31 •• 1k/n linear .t!~ect. quadratic errecta 1 1 1 1 1 1 0.82055529 0.27061652 0.02933343 2.33288970 2.18138675 0.22755608 0.82055529 0.27061652 0.02933343 2.33288970 2.18138675 0.22755608 0.63 0.21 0.02 1.78 1.67 0.17 0.4586 0.6653 0.8859 0.2302 0.2442 0.6912 Year 1 control v. other. control. v. 8 e1k/n2 control v. 15 elkin contro.l v. 31 elkin 11n •• r errecta quadratic errecta 1 1 1 1 1 1 2.11198217 0.67938777 0.77630054 3.43891518 3.25651384 0.01496561 2.11198217 0.67938777 0.77630054 3.43891518 3.25651384 0.01496561 1.63 0.52 0.60 2.65 2.51 0.01 0.2488 0.4962 0.4682 0.1543 0.1639 0.9179 Year 2 control. v. other. control. v. 8 e1k/n2 control v. 15 e1k/n contro.! 31 e1k/_ 1inear errecta quadratiC e·rrecta 1 1 1 1 1 1 0.03767560 0.00284659 0.00094345 0.20503965 0.21556661 0.06848496 0.03767560 0.00284659 0.00094345 0.20503965 0.21556661 0.06848496 0.14 0.01 0.00 0.77 0.81. 0.26 0.7198 0.9210 0.9545 0.4141 0.4030 0.6302 1 1 1 1 1 1 0.00614631 0.00054.67 0.30660056 0.11441679 0.0.368452 0.47179741 0.00614631 0.00054867 0.30660056 0.11441679 0.08368452 0.47179741 0.02 0.00 0.8. 0.33 0.24 1.36 0.8985 0.9696 0.3.35 0.5.67 0.6408 0.2878 Contrast v. v. Year 3 control. other. control. v. 8 e1k/ka2 control. v. 15 e11c/_ control v. 31 elk/_ lin •• r errecta quadratiC errecta F Value Pr > r Figure 34. Effects of elk density on canopy coverage of forbs at the end of the spring grazing season. vertical bars give ± 1 standard error. Test of specific hypotheses are shown in the table of orthogonal contrasts. �252 Appendix 1. Summary of canopy coverage data by species. Elk Density 0 (animals/km2) 8 15 31 Canopy Canopy Canopy Canopy Cover (% ) Cover ( % ) Cover ( % ) Cover ( %) MEAN S.E. MEAN S.E. MEAN S.E. MEAN S.E. Forage Class species YEAR Annual Forb Alyssum desertorum 1 0.17 0.05 0.06 0.05 0.14 0.05 0.03 0.02 2 0.01 0.01 0.00 3 0.00 0.00 0.00 0.00 0.01 0.00 0.01 1 10.5 4.65 5.94 1.08 3.55 0.44 3.14 1.84 2 0.70 0.29 0.56 0.20 0.56 0.35 1.05 0.61 3 0.04 1 0.02 0.01 0.02 0.00 0.04 0.04 0.01 0.01 2 0.00 1 0.02 Annual Grass Bromus tecto rum Festuca octoflora · · · 0.03 0.03 0.00 0.00 0.00 0.00 0.00 0.01 0.01 0.00 Perennial Forb Allium acuminatum · 2 Antenaria parvif. 1 Arenaria fendleri Artemisia frigida Astragalus drum. 3 · 0.02 1 · 0.10 Values shown as . · 0.01 3 0.01 1 0.07 0.05 0.15 · 0.13 · 0.01 2 0.11 0.10 0.13 · 0.02 3 0.08 · 0.06 · 0.07 · 0.07 · 0.03 1 0.08 · 0.04 0.02 0.07 0.02 0.03 0.01 2 0.07 (CONTINUED) * 0.00 indicate species was not present. �253 Appendix 1. Summary of canopy coverage data by species. Elk Density 0 (animals/krn2) 8 15 31 Canopy Canopy Canopy Canopy Cover (% ) Cover ( % ) Cover (% ) Cover (% ) MEAN S.E. MEAN S.E. MEAN S.E. MEAN S.E. Forage Class Species YEAR Perennial Forb Astragalus drum. 3 Astragalus spat. 1 0.01 3 0.04 1 0.17 0.15 0.03 0.00 0.07 0.02 0.01 0.00 2 0.04 . 0.01 0.00 3 0.13 . 0.03 0.00 0.03 0.01 Aster spp. 1 0.01 Balsamorhiza sag. 1 0.14 2 0.07 cela 2 0.13 Chaenactis doug. 1 Cirsium undulatum 1 · 0.55 2 · 0.03 3 · 0.28 Astragalus purse Commandra umbo 0.01 · 0.06 0.05 · 0.06 . 0.01 · 0.10 · 0.25 0.08 · 0.01 · 0.01 0.01 . 0.02 0.01 1 0.23 0.09 0.11 0.04 0.28 0.09 0.11 0.05 2 0.09 0.05 0.09 0.02 0.09 0.02 0.04 0.03 3 0.05 0.04 0.06 0.02 0.04 0.01 0.04 0.02 (CONTINUED) * · 0.01 Values shown as . indicate species was not present. �254 Appendix 1. Summary of canopy coverage data by species. Elk Density 0 (animals/km2) 8 15 31 Canopy Canopy Canopy Canopy Cover (% ) Cover (% ) Cover (% ) Cover ( % ) MEAN S.E. MEAN S.E. MEAN S.E. MEAN S.E. Forage Class Species YEAR Perennial Forb Plox hoodii 2 0.15 0.08 0.15 0.07 0.46 0.03 0.06 0.04 3 0.06 0.03 0.10 0.05 0.12 0.06 0.05 0.01 Plox longifolia 1 2 0.01 3 0.01 · 0.01 · 0.00 · 0.02 · 0.00 · 0.01 · 0.01 · · 0.00 0.00 · · 0.01 Plantago patgonica 1 Psoralidiurn lanc. 1 0.86 · 0.02 2 0.02 · 0.05 3 0.10 1 0.17 0.06 0.21 0.07 0.18 0.07 0.15 0.02 2 0.10 0.07 ·0.09 0.03 0.07 0.03 0.09 0.01 3 0.14 0.04 0.08 0.04 0.20 0.10 0.09 0.04 Sparalcea cocine Tragopogon dubius · 0.03 · 1 1 0.61 0.61 0.03 0.00 0.01 2 Unknown (CONTINUED) · 0.01 · 0.00 · 0.09 0.03 2 Trifolium gymnocar 0.08 0.07 · 0.00 · · 0.01 0.00 0.00 0.00 0.00 3 0.01 1 0.08 0.07 0.81 0.73 0.12 2 0.06 0.04 0.04 0.00 0.01 3 0.02 0.02 0.02 0.00 0.07 0.03 0.07 · · 0.04 0.02 0.04 �255 Appendix 1. Summary of canopy coverage data by species. Elk Density o (animals/km2) 15 8 31 Canopy Canopy Canopy Canopy Cover (%) Cover (%) Cover (%) Cover (%) MEAN S.E. MEAN S.E. MEAN S.E. MEAN S.E. Forage Class Species YEAR Perennial Forb Cryptantha flava 1 · 0.02 · . 3 Erigeron eatonii 1 Erigeron spp. Lupinus argenteus Lupinus polyphy. Plox hoodii · 0.01 · 0.11 0.04 0.06 0.03 0.06 0.02 1 0.06 2 0.04 0.03 0.06 0.02 0.05 0.02 0.01 0.00 3 · 0.06 1 · 0.02 2 · 0.01 · 0.01 0.00 0.01 0.00 1 0.56 0.24 0.31 0.14 0.43 0.32 0.23 0.02 2 0.22 0.10 0.06 0.05 0.21 0.17 0.20 0.09 3 0.28 0.07 0.01 2 0.33 0.28 0.53 0.28 0.23 0.07 0.24 0.13 3 0.39 0.20 0.68 0.36 0.15 0.03 0.26 0.09 · 0.59 0.06 0.22 0.05 1 0.31 0.14 0.36 0.15 0.33 0.22 0.14 0.07 (CONTINUED) * 0.02 0.16 2 Eriogonum oval. 0.34 0.10 0.03 Values shown as . indicate species was not present. �256 Appendix 1. Summary of canopy coverage data by species. Elk Density 0 (animals/km2) 8 15 Canopy Canopy Cover (% ) Cover (%) 31 Canopy Canopy Cover (% ) Cover ( %) MEAN S.E. MEAN S.E. MEAN S.E. MEAN S.E. Forage Class Species YEAR Perennial Grass Agropyron smithii 1 1. 28 0.75 1.02 0.66 0.88 0.58 0.17 0.05 2 0.36 0.11 0.44 0.22 0.30 0.07 0.14 0.05 3 0.12 0.06 0.05 0.01 0.52 0.39 0.02 0.01 Agropyron spicatum 1 2 Elymus cinereus 1 0.07 0.03 0.61 2 0.01 1 · 0.02 · 0.19 Koleria macrantha oryzopsis hymenoid. Poa spp. (CONTINUED) · 0.01 · 0.92 0.87 . · 3 Hilaria jamesia 1.09 0.02 0.02 3 Agropyron trachy. · · 0.00 0.16 0.03 1 0.02 2 0.15 3 0.11 1 1. 39 0.37 1.24 0.36 1.43 0.38 0.75 0.34 2 0.16 0.02 0.19 0.06 0.26 0.15 0.18 0.06 3 0.03 0.01 0.02 1 0.35 0.18 0.62 0.20 0.53 0.32 0.18 0.09 2 0.21 0.09 0.21 0.07 0.03 0.01 0.05 0.02 3 0.11 0.03 0.05 1 0.72 0.13 0.54 0.13 0.58 0.16 0.37 0.08 · · 0.03 0.01 0.05 0.02 0.02 . 0.04 0.01 �257 Appendix 1. Summary of canopy coverage data by species. Elk Density 0 (animals/km2) 8 15 31 Canopy Canopy Canopy Cover ( % ) Cover (% ) Cover ( %) Canopy Cover ( % ) MEAN S.E. MEAN S.E. MEAN S.E. MEAN S.E. Forage Class species YEAR Perennial Grass Poa spp. 2 0.30 0.09 0.31 0.07 0.39 0.13 0.21 0.03 3 0.09 0.04 0.03 0.01 0.06 0.03 0.09 0.02 1 0.15 0.08 0.54 0.07 0.12 0.03 0.13 0.06 2 0.19 0.05 0.09 0.02 0.02 0.01 0.03 0.00 3 0.06 0.02 0.03 0.01 0.01 1 5.88 0.93 8.10 0.69 5.97 1. 40 5.95 1.62 2 2.78 0.17 2.97 0.60 4.60 0.84 2.25 0.38 3 2.41 0.28 2.00 0.55 2.25 0.78 1. 47 0.59 1 0.13 0.00 0.14 2 0.43 3 0.06 1 15.4 2.93 14.8 0.79 11. 2 1.66 14.6 0.68 2 13.1 1.98 14.0 2.24 10.2 2.10 14.5 0.71 3 16.7 4.31 14.2 1.57 11. 3 2.53 17.0 1.65 1 0.29 0.08 0.32 0.32 0.19 0.13 0.30 0.06 2 0.24 0.12 0.13 0.06 0.22 3 0.22 0.04 0.56 1 1.66 0.47 0.81 0.46 1.93 0.96 0.88 0.29 2 0.83 0.23 0.33 0.21 0.92 0.38 0.64 0.16 3 1.79 1. 24 0.31 0.12 0.54 0.24 0.42 0.18 sitanion histrix Stipa comata Shrub Artemisia cana Artemisia triden. Chrysothamnus nas. Chrysothamnus vic. (CONTINUED) . . 0.47 0.05 · 0.02 0.00 · 1.64 · 0.56 · · 0.50 · 0.43 0.43 · 0.17 · · 0.14 0.11 0.38 0.15 0.16 0.08 �258 Appendix 1. Summary of canopy coverage data by species. Elk Density 0 (animals/km2) 8 Canopy Canopy Cover (% ) Cover (%) 15 31 Canopy Cover ( %) Canopy Cover ( %) MEAN S.E. MEAN S.E. MEAN S.E. MEAN S.E. Forage Class Species YEAR Shrub Eriogonum mic. 1 0.11 0.01 0.14 0.06 0.14 0.10 0.11 0.02 2 0.10 0.05 0.10 0.05 0.05 0.02 0.06 0.01 3 0.12 0.01 0.11 0.06 0.16 1 0.10 0.02 0.03 0.02 0.14 0.02 0.32 0.24 2 0.05 0.02 0.10 0.07 0.24 0.05 0.09 0.04 3 0.18 0.09 0.10 0.08 0.18 0.06 0.06 0.03 Gutierrezia saro. Haplopappus accal. 1 · 0.02 2 · · 0.02 3 Leptodactylon pun. Opuntia polyacan. Purshia· tridentata Tetrademia spinosa Values shown as . 0.10 0.08 0.06 1 4.37 4.23 0.62 0.40 1.51 0.98 0.54 0.41 2 0.24 0.15 0.27 0.21 0.77 0.32 0.30 0.13 3 1.64 1. 42 0.25 0.23 0.49 0.35 0.07 0.05 1 0.60 0.39 0.19 0.14 0.99 0.92 0.09 0.06 2 0.57 0.41 0.48 0.43 0.56 0.50 0.13 0.12 3 0.78 0.63 0.66 0.46 0.54 0.52 0.16 0.15 1 0.20 2 0.05 0.03 0.05 3 0.14 0.06 1 0.54 0.41 0.05 0.02 0.15 0.07 0.20 0.08 2 0.13 0.03 0.03 0.01 0.34 0.04 0.19 0.13 . 0.09 (CONTINUED) * · indicate species was not present. · 0.16 · · 0.38 · 0.19 0.41 �259 Appendix 1. Summary of canopy coverage data by species. (an i.ma Ls zkmz ) Elk Density 0 8 Canopy Cover ( %) 15 31 Canopy Canopy Canopy Cover (% ) Cover (% ) Cover ( % ) MEAN S.E. MEAN S.E. MEAN S.E. MEAN S.E. Forage Class Species YEAR Shrub Tetrademia spinosa 3 All species Total Cover 0.16 0.09 0.08 1 46.8 4.47 37.9 2 20.9 1. 08 21.2 3 25.4 4.81 18.8 . 0.42 0.08 0.16 0.07 1. 76 32.8 0.72 30.0 3.54 0.84 20.5 2.38 20.6 0.89 0.22 18.5 1. 81 20.6 0.82 �260 Appendix 2. Summary of canopy coverage data by forage class. Elk Density 0 (animals/km2) 8 15 Canopy Canopy Canopy Cover (% ) Cover (% ) Cover (%) 31 Canopy Cover (% ) MEAN S.E. MEAN S.E. MEAN S.E. MEAN S.E. Forage Class YEAR Annual 1 0.17 0.05 0.06 0.05 0.14 0.05 0.03 0.02 2 0.01 0.01 0.00 3 0.00 0.00 0.00 0.00 0.01 0.00 0.01 1 10.6 4.64 5.95 1.07 3.58 0.47 3.15 1.85 2 0.70 0.29 0.56 0.21 0.56 0.35 1.05 0.61 3 0.04 1 3.36 0.14 2.69 1. 28 2.64 0.88 1.85 0.22 2 0.94 0.38 0.89 0.39 0.96 0.08 0.57 0.21 3 0.99 0.16 0.97 0.45 1.45 0.31 0.72 0.17 1 9.83 1.69 12.3 0.95 10.5 0.29 7.56 1.65 2 4.01 0.18 4.21 0.93 5.59 1.14 2.85 0.53 3 2.85 0.26 2.13 0.58 2.92 1.00 1.68 0.65 1 22.9 2.71 16.9 1.02 15.9 1.15 17.4 0.14 2 15.2 1.06 15.6 1.88 13.3 0.87 16.2 0.37 3 21.6 4.64 15.7 0.81 14.1 1.24 18.2 1. 33 1 46.8 4.47 37.9 1.76 32.8 0.72 30.0 3.54 2 20.9 1.08 21. 2 0.84 20.5 2.38 20.6 0.89 3 25.4 4.81 18.8 0.22 18.5 1.81 20.6 0.82 Annual Forb Grass Perennial Forb Perennial Grass Shrub Total Cover . . 0.03 0.03 0.00 0.00 0.00 0.00 0.00 0.01 0.01 � https://cpw.cvlcollections.org/files/original/5e8871869a86359d187d1d8bb8d38b57.pdf 1e106350dfaaebd97d677d936a24cc3e PDF Text Text 1 Colorado Division of Wildlife Wildlife Research Report October 1989 JOB FINAL REPORT State of __~C~o~l~o~r~a~d~o~ Project Migratory Game Bird Research 01-03-212 Work Plan _1_ Job Title: : Job 14 Ecological studies of the flightless period of ducks in Colorado Period Covered: Author: _ 01 April 1988 through 31 March 1989 Michael R. Szymczak Bersonnel: Michael R. Szymczak and James K. Ringelman, Wildlife Colorado Division of ABSTRACT All data from this project have been collected and partially analyzed. This study addressed survival, habitat use, behavior, energetics, condition, food habits, weight dynamics, and the duration of the flightless period of ducks in North Park, Colorado. Data were gathered on all species, but most information were obtained for gadwall (Anas· strepera) or mallards (Anas platyrhynchos). Because of the comprehensive nature of the study, a monograph entitled "The ecology, behavior, and energetics of post-breeding waterfowl in Colorado" is currently being planned. All additional analysis and publication preparation will be conducted under Work Plan 22, Job 2, Migratory Bird Publications . • Prepared by: ?/I,;.R.,f 7? 400/ Michael R. Szymczak Wildlife Researcher C James K. Ringelman Wildlife Researcher C ��3 Colorado Division of Wildlife Wildlife Research Report October 1989 JOB PROGRESS REPORT State of __~C~o~l~o~r~a~d~o~ Project 01-03-212 Work Plan 1 Job Title: : Job Migratory Game Bird Research 15 Development and use of a physiological condition index for monitoring wintering mallard nutrient reserves Period Covered: Author: _ 01 April 1988 through 31 March 1989 Michael R. Szymczak Personnel: Michael R. Szymczak and James K. Ringelman, Colorado Division of Wildlife ABSTRACT Data for this project have been collected and most of it has been analyzed. To date, 2 manuscripts have been published, another is in draft form ready for review, and 2 others are planned using data collected under this study. All additional analysis and publication preparation will be conducted under Work Plan 22, Job 2, Migratory Bird Publications. Completed and planned publications are: Ringelman, J. K., and M. R. Szymczak. 1985. A physiological condition index for wintering mallards. J. Wildl. Manage. 49:564-568. Szymczak, M. R., and J. K. Ringelman. 1986. Differential habitat use of patagial-tagged female mallards. J. Field Ornitho1. 57:230-232. Ringelman, J. K., D. C. DeLong, and M. R. Szymczak. Geographical variation in body mass· and size of wintering mallards (draft). Ringelman, J. K., and M. R. Szymczak. Proximate and ultimate determinants of field-feeding behavior in mallards (planned). Szymczak, M. R., and J. R. Ringelman. The influence of weather on body mass dynamics of northern wintering mallards (planned). Prepared by: yn;_g.I7? ~ f Michael R. Szymczak Wildlife Researcher C ~~~ JamesK.~ Wildlife Researcher C ��5 Colorado Division of Wildlife Wildlife Research Report October 1989 JOB FINAL REPORT State of __~C~o~l~o~r~a~d~o Project 01-03-212 Work Plan Job Title: _1_ : Job MigratohY Game Bird Research 16 Field-feeding ecology of mallard ducks Period Covered: Author: _ 01 April 1988 through 31 March 1989 Michael R. Szymczak Personnel: Michael R. Szymczak and James K. Ringelman, Wildlife Colorado Division of ABSTRACT All data from this project have been collected and analyzed. An outline for a manuscript entitled "Foraging rates of mallards in cornfields: the effects of post-harvest treatments" has been prepared. Additional publication preparation will be conducted under Work Plan 22, Job 2, Migratory Bird Publications. Michael R. Szymczak Wildlife Researcher C James K. Ringelman Wildlife Researcher C ��7 Colorado Division of Wildlife Wildlife Research Report October 1989 JOB PROGRESS REPORT State of __~C~o~l~o~r~a~d~o~ Project Work Plan 01-03-212 <W-88-R) 1__ Job _ Migratory Game Bird Research 17 Job Title: Habitat use by wintering mallards along the Front Range of Colorado Period Covered: Authors: I April 1988 through 31 March 1989 James K. Ringelman and Michael R. Szymczak Personnel: R. Falise, M. James, R. Westfall, Colorado State University; L. Roberts, Front Range Land and Cattle Company; J. Corey, Colorado Division of Wildlife. ABSTRACT Twenty-five mallards were radio-marked at Lonetree Creek and Chestnut Slough during December 1988, and an additional 11 maliards were instrumented in January on Chestnut Slough. Movements and habitat use of radio-marked and unmarked mallards were measured from December through February to document relationships among hunting activity, weather and habitat selection. Disturbance was further evaluated by simulating hunting activity during January and February, then evaluating mallard behavioral responses. Scan sampling was used to document group behavior by habitat type and time of day. Temperatures were above normal in January and far below normal in February. Consequently, habitat availability caused by ice conditions decreased rapidly in late winter, concentrating birds on the remaining open water areas. In response to these conditions, along with habitat management practices intended to minimize disturbance at Chestnut Slough, the average home range size of radio-marked birds was smaller than in previous winters. Mallards preferred warm-water sloughs and rivers, but made infrequent use of ponds and most reservoirs. Responses of instrumented birds to hunting disturbance, measured using an accumulative disturbance index, varied among individuals and in relation to weather conditions. Cold weather reduced the duration of avoidance following disturbance. Shots fired from areas adjacent to the study wetland generally did not elicit movements, and birds often moved independent of hunting disturbance. During simulated hunting disturbance, most mallards returned to roost at Chestnut Slough 10 to 13 hours following departing. As in previous years, resting and swimming were the principal activities, f~llowed by feeding, preening and walking. ��9 HABITAT USE BY WINTERING MALLARDS ALONG THE FRONT RANGE OF COLORADO James K. Ringe1man Michael R. Szymczak P. N. OBJECTIVES 1. Relate mallard habitat use to the physical characteristics of wetlands, aquatic and upland plant communities, wetland macroinvertebrate populations, weather, and hunting regimes. 2. Characterize mallard use of wetlands through time budget techniques. 3. Document the composition of the wetland community within the study area, and relate wetland use by mallards to availability. SEGMENT OBJECTIVES 1. Capture and radio-mark 24 mallards during early December. marked sample equally by age and sex. Divide the 2. Monitor the diurnal movements of instrumented mallards using aerial and vehicle-mounted tracking systems. Code bird locations using a Universal Transverse Mercator (UTM) grid system. 3. Conduct aerial tracking flights once a week, or more frequently if weather conditions or bird movements warrant. 4. Determine the availability of ice-free wetlands daily using a representative sample of wetland types. 5. Monitor hunting regimes on key wetlands to determine the effect of hunting disturbance on movements of radio-marked mallards. Simulate the effects of hunting with controlled disturbance during the post-season period. 6. Obtain National Weather Service records for daily temperature, wind and snowfall for a representative site in the study area. 7. Quantify behavior of marked and unmarked mallards in relation to wetland type, social status, and time of day using scan sampling. SruDY AREA The 1,089 km2 study area is located along Colorado's Front Range near Greeley, Colorado. The area is rectangular in shape (36.2 km east to west, 30.5 km north. to south) and is bordered on the west by Interstate Highway 25, on the north by State Highway 392, on the south by State Highway 66, and on the east along the UTM line 542000• Upland habitats are typical of the Colorado Piedmont (Fenneman 1931) portion of the Great Plains physiographic �10 province. Elevations average about 1430 m, with generally flat to rolling terrain interrupted by occasional steep bluffs along major river systems. The study area is drained by 4 rivers: the Cache 1a Poudre, Big Thompson, St. Vrain, and South Platte. Climate is characteristically continental (Gittings 1941), with December, January, and February daily temperatures averaging -1.3, -3.9 and -0.4 C, respectively. Average precipitation during these same months totals 0.74,,0.76 and 0.71 cm. Fog is common during warm winter mornings and evenings. Native grasses consist of blue grama (Boute1oua gracilis) and buffalo grass (Buchloe dactyloides), with rubber rabbitbrush (Chrvsothamnus nauseosus) and snakeweed (Gutierrezia sarothrae) the dominant shrubs. Cottonwood (Populus sargentii), willows (Salix spp.), and green ash (Fraxinus pennsy1vanica) are common along watercourses. Aquatic vegetation varies by wetland type and water regimes. Large wetlands used for irrigation water storage are mostly devoid of vegetation due to extreme water level fluctuations during the growing season. Small wetlands often contain submergent watermilfoi1 (Myriophyllum exa1bescens), pondweeds (Potomogeton spp.), coontail (Ceratophyllum demersum) and muskgrasses (Chara spp.). Smartweeds (Po10gonum spp.), sedges (Carex spp.), cocklebur (Xanthium strumarium), and cattails (Typha spp.) occur around wetland margins and in shallowly flooded areas. Watercress (Nasturtium officiona1e), a common aquatic of warm-water sloughs, harbors high densities of pond snails (Physa spp.) and other macroinvertebrates consumed by mallards. Most land in the study area is devoted to irrigated agriculture or livestock husbandry. Corn, alfalfa, wheat and sugar beets are the dominant irrigated crops. Malting barley and vegetables' are grown in lesser amounts. Several large cattle feedlot operations exist in the northern portion of the area. Winter waterfowl populations typically range from several thousand to >10,000 mallards (Anas platyrhynchos), and up to 10,000 Canada geese (Branta canadensis). Pintails (Anas acuta), green-winged teal (Anas crecca) and wigeon (Anas americana) are present in lesser numbers. METHODS Mallards were captured in Salt Plains bait traps (Szymczak and Corey 1976) on Lonetree Creek, adjacent to Farr Feedlot (in the northwest corner of the study area) on 7 December 1988, and at Chestnut Slough on 8 December 1988 and 11 January 1989. Six birds captured at Lonetree Creek and 30 birds from Chestnut Slough were instrumented with back-mounted radio transmitters (Dwyer 1972) weighing 24g. Mallards marked in December provided information on habitat selection and movements, whereas those marked in January were used primarily to evaluate simulated hunting disturbance. Measurements were made of wing length and body weight, and a standard U.S. Fish and Wildlife Service band affixed to each bird prior to release. Truck-mounted, precision direction finding antennae arrays were used to locate birds from the ground. When a signal was detected, 2 or more azimuths were taken from a known location then plotted on 1:24,000 topographic maps to a resolution of 1 ha. A transparent grid overlay was used to code the location of each bird in the standard UTM grid system. Concurrent with information on the bird's location, the date, time, activity status, air temperature, precipitation, wind speed, wetland name, and habitat type were also recorded. A portable computer (Tandy Model 102) was used to tabulate and �11 error check data in the field. All data were transferred to an IBM-PC for subsequent analyses. Mallards were located as frequently as possible, usually 2-4 times/day. Tracking schedules alternated between (1) locating as many individuals as frequently as possible during the day, and (2) constant monitoring of 1 or 2 individual birds during a single day. Responses of marked birds to hunting disturbance at Chestnut Slough were obtained by continuous monitoring of birds and hunters during the hunting season and recording the reaction of birds during simulated hunts after the hunting season. During the season, personnel arrived at Chestnut prior to arrival of hunting parties in the morning, and remained at the slough as long as hunters and/or marked ducks were present. During this period, activities on Chestnut Slough were categorized by disturbance factors in the following manner: no disturbance (factor 0), shooting audible from Chestnut but not originating on the slough (factor I), shooting on Chestnut >200 m from the subject bird (factor 2), driving near the subject bird (factor 3), walking near the subject bird (factor 4), and shooting <200 m from the subject bird (factor 5). During simulated hunts, shots were fired <200 m from the subject bird(s) and the bird's response recorded. Chestnut was then monitored during duck activity hours to determine when or if the bird returned to the Chestnut roost. All shots were fired over the heads of ducks within legal shooting hours. Wetland availability was determined by monitoring ice conditions of various habitat types daily. Percent open water for each wetland was estimated and recorded along with the date, time, and air temperature. Daily maximum and minimum temperatures and snowfall records were obtained from the u.S. Department of Commerce, Climatological data for the Greeley station. Time budget data were obtained using scan sampling (Altmann 1974) stratified by habitat type and time period. Habitat types were reservoirs, rivers, warm-water wetlands, and other (mostly gravel pits). Time periods were sunrise-0900, 0900-1200, 1200-1500, and l500-sunset. Spotting scopes (40X) or binoculars were used for all time budget collections. Subject animals were selected by pointing the scope or binocular at a flock, then scanning through the flock until 50 individuals were observed. Behaviors were recorded on a tape recorder then immediately transferred to a portable computer. The following behavior classifications were used: (0) out-of-sight, (1) rest, (2) swim, (3) walk, (4) preen, (5) feed in water, (6) feed on land, (7) courtship, (8) aggression, and (9) other. RESULTS Weather and Ice Conditions Temperatures were variable through the 1988-89 winter period, ranging from near average in December, considerably above average in January (+6.3 F), and far below average in February (-9.7 F). Average February temperatures were influenced dramatically by an extreme cold period during the first 10 days of the month (Fig. 1). Precipitation was nearly 0.5 inches above average for all 3 months. The number of days with measurable snow on the ground (Fig. 2) was greater than in 1986-87 and similar to 1987-88 (Ringelman and Szymczak 1988). However, snow cover was divided into 2 segments (Dec and Feb) in 198889 rather than the continuous 50+ day snow cover experienced in 1987-88. Warm January temperatures in 1988-89 kept snow from accumulating even though precipitation was above average. Heavy snow cover in early February corresponded to extreme cold temperatures. �12 60 -U. 50 40 W a: :::l I- <t a: w a.. ~ W --------- ---- -----~-,\ 30 1\ , 20 \ '\ .•.I' , I I , 10 I I 0 I, \, , ,. , \' I' , \, I I \ ,', • rl\",J I ' I, I I I, ,I I "I' '1 ",', , ~ \ , I ----------------~-----------r-;---------------~--------------------1 " -10 \ I "", I \ 'V I', / ' I,' ' " I' I ~ I ' I \ I I \ \ , '1 ~ I ,I II I,I " I' I' I I II I \ : I (\ I J ,,' ,,, \ '" ~ I I- \ i , , - ,\ \ \ \ \ \ -- ----1------------------------ I , I \ I I I , I t» ~I , '''v -20 1 10 20 30 Dec. 10 20 30 DATE 10 Feb. 20 Fig. 1. Dally maximum and minimum temperatures at Greeley Colorado, 1988-89. Z ~ 10 8 c Z :::l o 6 a: CJ z o ~ o 4 2 Z til o 1 10 20 Dec. 30 10 20 DATE Fig. 2. Snow on ground at Greeley, Colorado, 1988-89. 30 10 20 Feb. 28 �13 Ice conditions of rivers varied seasonally and between areas. Except during early February, some river habitat was available in the study area throughout winter (Fig. 3). The Cache 1a Poudre River, on a percentage basis, provided less open water than the Big Thompson or the South Platte. The South Platte remained ice-free in most of the study area throughout the field season. Most reservoirs froze in mid-December and remained frozen except for small open water areas in the middle of some reservoirs that were maintained by the activity of roosting waterfowl (Fig. 3). Most small ponds remained frozen. The January warm period had a slight influence on the amount of open water on some reservoirs, but had a more pronounced affect on ponds (Fig. 3). Characteristics of Instrumented Birds Of the 25 mallards radio-marked during December, 7 were adult males, 5 were adult females, 7 were immature males, and 6 were immature females (Table 1). The initial sample of marked birds was intentionally skewed, since males experience higher hunting season mortality rates than females. By 10 January, only 10 radio-marked birds remained in the study area. Eight radios had been recovered from birds that had been shot or had died from natural causes (Table 1). When traps were re-dep10yed on Chestnut Slough on 11 January, an additional 4 adult male, 3 adult female, 2 immature male, and 2 immature female mallards were radio-marked. Movements Instrumented ducks were re-10cated 1,159 times, with the time of telemetry fixes nearly equally distributed throughout the day (Table 2). Thirty-eight percent (446) of the telemetry locations were obtained in December, 39% (452) were recorded in January, and 23% (260) were collected in February. Radio-marked birds remained concentrated near Chestnut Slough and Lonetree Creek, the two capture sites (Fig. 4). Despite this apparent affinity for wetlands in the vicinity of the capture sites, many birds used both areas (Figs. 5-33). Radio-marked mallards were also located along and near the South Platte River near Kersey and Gilcrest, on the Cache 1a Poudre River near Windsor, the Big Thompson River southwest of Greeley, and warmwater sloughs near Platteville (Fig. 4). Overall, home ranges were smaller in 1988-89 than in 1986-87 or 1987-88 (Table 3). Home range sizes for 1988-89, computed using either the minimum area polygon method or the Jennrich and Turner confidence ellipse, did not differ according to bird age, sex, or age/sex interaction (P-0.37, Table 4). Minimum area home ranges (± SD) were 143.2 km2 (± 181.5) for adult females, 76.4 krn2 (± 95.5) for adult males, 49.3 km2 (± 34.5) for immature females, and 45.0 km2 (± 53.7) for immature males. The total number of locations per bird did not correlate with either the minimum area home range (r-0.09, p-0.63) or the Jennrich-Turner 95% confidence ellipse (r--0.11, p-O.57), suggesting that the number of telemetry locations was sufficient to provide a stable estimate of home range size. �14 100 80 60 40 RNERS 20 0 a: W 100 I~ 3: z w 80 60 RESERVOIRS 0- 0 IZ w a: w Q. 40 20 0 :.......................... :", 0 -0 0 :" ••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••• 100 80 ,\ 60 i'" , I\ \ I \ 40 I 20 PONDS 1 1\ I / / /' \ \ \ '10 / \,." \/ 0 /' r--' 20 30 DEC / \ ,. \,. ,. ", , " \/ \ .•.. ,. 10 20 30 JAN DATE Fig. 3. Percent open water during 1988-89 for 3 wetland types. - ,. ...•• - - -" 10 FEB 20 �15 Table 1. Characteristics of radio-marked mallards during 1988-89. Capture location Lonetree Lonetree Lonetree Lonetree Lonetree Lonetree Chestnut Chestnut Chestnut Chestnut Chestnut Chestnut Chestnut Chestnut Chestnut Chestnut Chestnut Chestnut Chestnut Chestnut Chestnut Chestnut Chestnut Chestnut Chestnut Chestnut Chestnut Chestnut Chestnut Chestnut Chestnut Chestnut Chestnut Chestnut Chestnut Chestnut Date of capture Crk. Crk. Crk. Crk. Crk. Crk. 7 7 7 7 7 7 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 11 11 11 11 11 11 11 11 11 11 11 Dec Dec Dec Dec Dec Dec Dec Dec Dec Dec Dec Dec Dec Dec Dec Dec Dec Dec Dec Dec Dec Dec Dec Dec Dec Jan Jan Jan Jan Jan Jan Jan Jan Jan Jan Jan Bird number Age Sex 1384 1400 1420 1448 1480 1500 1360 1530 1560 1590 1630 1650 1670 1690 1708 1730 1778 1800 1824 1850 1906 1927 1940 1960 1980 048 220 266 281 532 2420 2560 2590 2650 2670 2927 Adult Immature Adult Immature Adult Immature Adult Immature Immature Immature Adult Immature Immature Adult Immature Immature Adult Adult Adult Adult Adult Immature Immature Immature Adult Adult Immature Adult Adult Adult Immature Immature Immature Adult Adult Adult Male Male Male Female Female Female Male Female Female Female Female Male Male Male Male Male Female Male Male Female Female Male Male Female Male Female Male Male Female Male Male Female Female Male Female Male Fate Last contact 15 Feb Died 10 Feb Died 3 Jan No contact Last contact 22 Jan Died 20 Jan Died 22 Jan Last contact 11 Feb Shot 17 Dec Shot 23 Dec Died 14 Jan Shot 17 Dec Radio removed 29 Dec Last contact 20 Feb Shot 28 Dec Died 16 Dec Last contact 24 Dec Last contact 21 Dec Last contact 22 Dec Last contact 22 Jan Last contact 26 Dec Shot 27 Dec Died 1 Feb Last contact 27 Dec No contact Last contact 5 Feb Last contact 20 Feb Last contact 11 Feb Last contact 16 Feb Last contact 15 Feb Died 3 Feb Last contact 18 Jan Died 9 Feb Last contact 11 Feb Last contact 20 Feb Died 11 Feb �16 Table 2. Hourly distribution of telemetry locations during 1988-89. Time interval 0400-0459 0500-0559 0600-0659 0700-0759 0800-0859 0900-0959 1000-1059 1100-1159 1200-1259 1300-1359 1400-1459 1500-1559 1600-1659 1700-1759 1800-1859 Frequency Percent Cumulative frequency 1 21 95 126 88 95 75 94 80 74 76 87 98 127 22 0.1 1.8 8.2 10.9 7.6 8.2 6.5 8.1 6.9 6.4 6.6 7.5 8.5 11.0 1.9 1 22 117 243 331 426 501 595 675 749 825 912 1010 1137 1159 Cumulative percent 0.1 1.9 10.1 21.0 28.6 36.8 43.2 51.3 58.2 64.6 71.2 78.7 87.1 98.1 100.0 Table 3. Comparative mean home range sizes for radio-marked mallards during winter. 2 Home range size (km ) Year No. fixes Minimum area polygon method Jennrich-Turner method (95% ellipse) 1986-87 1,388 150.2 340.2 1987-88 1,406 124.0 232.6 1988-89 1,159 83.8 159.5 Habitat Use As in previous years, birds used warm-water wetlands and rivers extensively but made infrequent use of ponds and reservoirs, most of which remained frozen during winter (Table 5). Seven of 34 birds used cattle feedlots, particularly during cold weather. Females of both ages used warmwater wetlands less than males (8.1% of female locations versus 18.0% of male locations). Except for an apparent increase in use of riverine habitat with declining temperatures, habitat use during 1988-89 did not correlate with weather variables. �Windsor + D 0 + 0 ~ ~~~ +. Greeley 0 + ++ Ii 0 Q +f" +'Iir v ~ c:::;:> . _~ $'~.IIIIIIIm~,?IF ~ ci) ,:7 laIhamRes . ~ 0 J+~ SIou!jl 0 ~ 0 ~ n.: ~~ + ...".._.:,~- C> o~ c, ~ sf ;.l \\/.._ II " ",'VI) rzzzzzt,?z~~zzzn 1M Platteville Fig. 4. Locations of all radio-marked mallards, 1988-89. D Milton Res. •.... -...I �18 o D o o o = 10 km I Z 2 2 2 2 Z 2 2 2 II Fig. 5. Locations of adult female 048, 11 Jan to 5 Feb, 1989. c o = "\,0 10 km kZlllZZZIZI Fig. 6. Locations of adult female 281, 11 Jan to 16 Feb, 1989. �19 ..................................... ........... ~ o <:0 ..... .... ... ....... ... ...... ... 10 km e 7 7 7 7 7 7 7 7, 21 Fig. 7. Locations of adult female 1480, 7 Dec to 22 Jan, 1989. ,. o o 10 km Fig. 8. Locations of adult female 1630, 8 Dec to 14 Jan, 1989. ::;... 1" : �20 o ,. o ~ l \ . . o ..~. .•. :1> = o 10 km eZZZZ27ZZZI Fig. 9. Locations for adult female 1778, 8 Oec to 24 Oec 1988. o o D ......'" ;'il o ... 0 = .. .. .. ·0 ' .. '\. ' ' o .~ .9° 10 km tl121lZlZZI Fig. 10. Locations for adult female 1850,8 Dec to 22 Jan, 1989. �21 , o -, ". ". ,. .~. ,. <::> o .~ • <;>0 10 km e Z 7 Z 7 7 Z 2 2 2. Fig. 11. Locations of adult female 1906, 8 Dec to 26 Dec, 1988. �22 <::> o 10 km elll?IIIZZI Fig. 13. Locations of adult male 266, 11 Jan to 11 Feb, 1989. o D ................... ~ ·..············ ..···············0········ "\. o 10 km t . Z ? ? ? 2 Z 2 2 21 6 Fig. 14. Locations of adult male 532, 11 Jan to 15 Feb, 1989. �23 o ,. o o '. +" ++ = o 10 km ., 2 Fig. 15. Locations of adult male 1360,8 ? 2 2 ? ? Z ? d D Dec to 22 Jan, 1989. ,. o D Fig. 16. Locations of adult male 1384, 7 Dec to 15 Feb, 1989. �24 = o .~ ....• 10 km eZZZZ2Z222i Fig. 17. Locations of adult male 1420,7 Dec to 3 Jan, 1989. ,. D o -, -, o -, II o ... -, .... 0.. -, -, ,. .,. = o .~ ..... 10 km '2 ? 2 2 2 ? ? ? 2 A Fig. 18. Locations of adult male 1690,8 Dec to 20 Feb, 1989. �25 ,. o o .~ ,.•. 10 km e 2 2 2 2 2 2 2 ? 21 Fig. 19. Locations of adult male 1800,8 Dec to 21 Dec, 1989. 0 /I .., .> = 7 " ~ cJ 0 e "'- •• .~ 0 o~ 6 ~/~ .•. , 10 km i?2Z????ZZd U Fig. 20. Locations of adult male 1824,8 Dec to 22 Dec, 1989. �26 ,. ~ 0 1I .•. ..,. c: J Q 0 0 ~o o~ C> ~f~ .~ .9' U 10 km rz z z z z z z Z Z ZI Fig. 21. Locations of adult male 2650, 11 Jan to 11 Feb, 1989. "'" o e 10 km It ? ? ? Z Z 2 2 2 A Fig. 22. Locations of adult mate 2927, 11 Jan to 11 Feb, 1989. c �27 o = 0° .0· .0· o o .~ •••• 10 km e Z ? 2 Z 2 2 2 Z 21 Fig. 23. Locations of Immature female 1500,7 Oec to 20 Jan, 1989. ~ = .:: ~···~····· ...: ..·:o:·:·::::~ .... .... .~ •••• 10 km ., ? Z Z 2 ? Z 2 Z a Fig. 24. Locations of Immature female 1530,8 Oec to 11 Feb, 1989. �28 , I) o 't... .... ,.: . ........ = '0 ". '" '" ". '" '" ". '" ". ". '" '0. • ..... :"t:;; 0"" 10 km e 2 ? Z ? 2 Z Z Z 21 Fig. 25. Locations of Immature female 1590, 8 Dec to 23 Dec, 1988. o () o o .. .... 7····.... ..' .... ./\ . 0 ~ = .. .. ' .. ' c ' .~ 0"" 10 km 12 Z Z Z 2 ? 7 Fig. 26. Locations of Immature female 1960,8 7 7 A Dec to 27 Dec, 1988. �29 ~ o () 4 0 0 .. 7 .., GJ <::> 0 10 km vI? ? 2 Z Z Z ? 21 Fig. 27. Locations of Immature female 2560, 11 Jan to 18 Jan, 1989. ~ o <::> o .~ .9' 10 km Fig. 28. Locations of Immature male 220, 11 Jan to 20 Feb, 1989. �30 o D o /I = o .~ .,,' 10 km eZ?22????" Fig. 29. Locations of Immature male 1400,7 Dec to 10 Feb, 1989. ,. 0 /I .., • 7 ~ GJ = 0 ...•.. 0 O~ Co ~f~" .~ .", 10 km 17 7 7 Z Z Z Z Z 7 d U Fig. 30. Locations of Immature male 1670,8 Dec to 29 Dec, 1989. �31 ,. 0 II • Q 0 e .~ .9' 0 "\. ~/~ ¥. o~ Co U 10 km 12 2 :2 Z Z Z Z Z Z 21 Fig. 31. Locations of Immature male 1708, 8 Dec to 28 Dec, 1988. 0 •. II .., ~ GJ = 0 e .~ .9' 0 "\. O~ Co ~f~ 10 km Ii 2 ? ? Z Z 2 Z ? A U Fig. 32. Locations of Immature male 1730,8 Dec to 16 Dec, 1988. �32 o o o .~ 0"" 10 km Q21?2??ZZZi D Fig. 33. Location. of Immature male 1940,8 Dec to 1 Feb, 1989. �33 Table 4. 1988-89. Home range sizes of radio-marked mallards during winter, 2 Home range size (km ) Bird No. 048 220 266 281 532 1360 1384 1400 1420 1480 1500 1530 1590 1630 1670 1690 1708 1778 1800 1824 1850 1906 1940 1960 2420 2560 2650 2670 2927 Means No. fixes Minimum area polygon method Jennrich-Turner method (95% ellipse) 12 44 55 25 25 33 25 129 14 33 15 12 14 20 11 142 28 18 17 15 22 21 61 36 25 10 52 51 61 17.6 54.6 4.0 81.8 37.0 226.8 257.2 129.2 19.4 454.5 36.2 32.9 99.5 24.9 3.8 118.4 20.5 29.8 10.1 0.7 414.7 69.7 4.4 72.4 67.4 0.1 83.7 53.1 6.6 67.0 129.7 8.2 196.0 82.3 476.8 668.1 96.3 62.7 774.9 83.8 160.1 193.6 94.4 14.4 76.4 23.6 118.6 36.1 2.4 629.8 136.5 7.8 153.3 163.3 0.3 112.2 48.7 8.1 83.8 159.5 �34 Table 5. Habitat use by individual radio-marked mallards, 1988-89. Percentage of locations by habitat typea Bird No. Unk. Pond Lake Hold. Warm-W. Ditch River Corn. Feed. No. obs. 048 26.7 0 0 0 60.0 6.7 6.7 0 0 15 220 10.0 4.0 0 0 44.0 10.0 24.0 0 8.0 50 266 6.8 0 0 0 62.7 1.7 28.8 0 0 59 281 10.7 0 0 0 50.0 7.1 28.6 0 3.6 28 532 0 20.0 0 0 28.0 8.0 32.0 4.0 8.0 25 1360 3.0 6.1 33.3 0 30.3 3.0 12.1 9.1 3.0 33 1384 21.9 9.4 0 0 34.4 6.2 21.9 6.2 0 32 1400 8.5 2.1 2.8 2.8 52.5 4.3 26.2 0.7 0 141 1420 0 0 7.1 0 50.0 0 42.9 0 0 14 1480 8.3 5.6 0 2.8 63.9 2.8 8.3 8.3 0 36 1500 0 0 0 0 73.3 0 26..7 0 0 15 1530 20.0 26.7 6.7 0 26.7 13.3 0 6.7 0 15 1560 50.0 0 0 0 25.0 0 25.0 0 0 4 1590 7.1 3.6 3.6 0 64.3 10.7 10.7 0 0 28 1630 8.3 0 0 0 58.3 25.0 8.3 0 0 24 1650 0 0 0 0 60.0 0 40.0 0 0 5 1670 8.3 0 0 0 83.3 8.3 0 0 0 12 1690 5.2 0 2.6 0 51.6 5.2 32.3 3.2 0 155 1708 6.7 0 6.7 0 60.0 10.0 13.3 3.3 0 30 1730 30.0 0 0 0 40.0 20.0 10.0 0 0 10 1778 5.0 5.0 30.0 0 45.0 5.0 0 10.0 0 20 �35 Table 5 (cont.). Percentage of locations by habitat typea Bird No. Unk. Pond Lake Hold. Warm-W. Ditch River Corn. Feed. No. obs. 1800 10.5 10.5 15.8 o 47.4 o 10.5 0 5.3 19 1824 21.1 0 0 o 36.8 31.6 o 5.3 5.3 19 7.4 7.4 o 63.0 o 11.1 0 o 27 o 23 1850 11.1 1906 8.7 0 0 o 60.9 o 21.7 1927 25.0 0 0 o 50.0 o 25.0 0 o 4 1940 6.2 0 0 o 40.6 3.1 50.0 0 o 64 1960 12.2 0 0 o 75.6 2.4 9.8 0 o 41 2420 0 0 0 o 56.0 4.0 40.0 0 o 25 2560 0 0 0 o 100.0 o o 0 o 10 2590 0 0 0 o 83.3 16.7 o 0 o 6 2650 5.6 1.9 0 o 72.2 o 18.5 1.9 o 54 2670 3.8 0 0 o 13.2 18.9 50.9 1. 9 11.3 53 2927 0 0 0 o 60.3 4.8 33.3 1. 6 o 63 Means 10.0 3.0 3.4 0.2 53.6 6.7 19.8 2.0 1.3 34 4.3 a Key to habitat abbreviations: Unk. - unknown habitat type Pond - Small ponds/reservoirs; Lake - lakes; Hold. - holding/sewage ponds; Warm-W. - warm-water wetlands; Ditch - ditches; River - rivers; Corn. - cornfields; Feed. - feedlots. Response to Hunter Disturbance Responses to disturbance were recorded for 14 radio-marked birds at Chestnut Sloughs during the final portion of the 1988-89 duck hunting season (Tables 6 and 7). Thirteen of these birds had been marked at Chestnut and one at Lonetree Creek. Eleven of the Chestnut-marked birds (3 of 5 AM, 4 of 4 AF, 3 of 6 1M, 2 of 4 IF) were present at Chestnut either during mid-morning hours on 16 December, the day before the hunting season, or prior to sunrise on 17 December, probably indicating the birds had been using the area on a regular basis since they were marked on 8 December. �36 TabLe 6. Presence or absence of radio-marked birds at Chestnut SLough, by day, during and (P = Present, BLank = Absent) foLLowing disturbance 1988-89. Date December Age/Sex AM AF 1M IF Bird No. 16 1690 1800 1824 1630 1778 1850 1906 17 18 19 20a P P P pb P P p p P P p pb p p 1590 1960 P P P 23 24 25 26 pb 27a 28 P P 29 30a 31 P P P 2 pb P P P 1400c 1650 1670 1708 1940 21a 22 Jan. pb pb p P P P pb pB P P P p P a No disturbance on Chestnut. b Departed without or prior to any disturbance on that date. c Radio-marked on lonetree Creek. p �37 Table 7 . Age/Sex AM Hunting season disturbance factors (see methods), bird reaction, and roost homing for radio-marked birds at Chestnut Sloughs 1988-89. Bird No. 1690 1800 Disturbance 1 2 3 4 5 10 2 * 1630 12/16 * 1778 12/16 12/17 *13 1906 1400 12/16 12/17 12/19 12/22 12/28 12/?8 12/29 12/30 12/31 1/2 1/2 1650 12/17 1670 12/16 12/17 12/19 1708 12/20 12/21 12/22 1 1* 3 2 2 2* 10 1 12/17 12/22 12/16 12/17 12/17 3 * * * 1850 1M 12/18 12/19 12/20 12/26 12/27 12/28 0 12/17 12/19 1824 AF Date 1 10 0 94 Did not return 4 22 Did not return 1 3* 1 1 15 24 21 Did not return 2 4 1* 22 (Did not depart) 10 3 0 >384~792 2 4 * 7* 1 2 1* * * * 9* 2 1 2* 1* 1* 10 * * 648 1* 6 * * 7 125 9 34 Did not return 1 1 2 10 (Did not depart) 30 0 0 0 (Did not depart) 16 19 4 2 1 5 1 1 Hours before return 1* 1 4 16 Accumulated rating before departure 1 2 22 (Did not dap ar t.) 31 >24~44 0 >48~60 0 Did not return 9 0 29 5 (Did not depart) 9 10 4 1 3 2 9 3 5 1* 19 Harvested12/17 >8~19 1* 22 0 5 4 41 Did not return (Radio removed 12/29) 0 (Did not depart) 0 18 0 Did not return �38 Table 7. Age/Sex (Continued) 1940 1M 1590 IF 1960 * Indicates Table 8. Disturbance Bird No. a Date 12/16 12/17 12/21 1 2 * 18 * 8 3 4 2 12/16 12/17 12/18 12/19 1 * 8 7 4* 1 * 18 * * type of disturbance 4 1 1986-87 1987-88 1 2* 8 4 1 1 Hours before return 22 (Did not depart) 43 which occurred Did not return 11 22 4 16 22 3 25 Harvested12/23 22 (Did not depart) 43 1 a .48~67 just prior Did not return to bird departure. A summary of the response to disturbance of radio-marked Chestnut Sloughs during the hunting season. See methods explanation of disturbance categories. Stay (%) Leave (%) Total Stay (%) Leave(%) Total 300 (96.9) 10 (3.1) 322 and res120nse to disturbance 2 1 Response 3 35 (89.7) 20 (69.0) ~ (10.3) ...2. (31.0) 39 29 12 (100.0) 15 (100.0) __Q __Q 12 15 94 a a No. of occurrances Year Accumulated rating before departure 1* 2 2 12/16 12/17 12/17 12/20 5 4 birds on for factor: 5 a a Q 1 (100.0) a 1 7 (63.6) 13 (71.2) 8 (88.9) ~ (36.4) ...l (18.8) 1 (11.1) 16 11 9 1988-89 Stay (%) Leave (%) Total 127 (100.0) 49 (96.1) 21 (95.5) 34 (82.9) 9 (60.0) _1 (0.7) (3.9) _l (4.5) ...2 (17.1) _j_ (40.0) 128 51 41 22 15 All Stay (%) Leave (%) Total .u Years .z 439 (97.6) 99 (94.3) 48 (77.4) (2.4) _6 (5.7) 14 (22.6) 462 105 62 47 (82.5) 17 (68.0) (32.0) 10 (17.5) 57 25 .a �39 Eight birds were subject to m1n1mum harassment on 16 December (Table 6) when some walking and driving (Table 7) occurred on Chestnut. Three birds did not depart in response to driving or walking, and of the 5 that did depart only one did not return. All others returned to Chestnut to roost. Of the 10 birds on Chestnut on opening day (17 December), only one (1670) departed prior to any disturbance (Table 6). Most others tolerated a considerable amount of disturbance in terms of shooting away from Chestnut. Four of the remaining 7 tolerated shots fired within 200 m, but all eventually left the area. Only one bird (1590) returned to Chestnut on the evening of 17 December, and 3 of the birds did not return to Chestnut for the remainder of the hunting season. After 17 December, 10 of the original 13 birds returned to Chestnut at some time, but most visits were brief. Before the second weekend of the hunting season, all but one bird had discontinued the use of Chestnut (Table 6). The birds did not return even though very cold temperatures toward the end of the month (Fig. 1) reduced the number of open water roosts in the study area. As in previous years, birds generally did not respond· to shots fired from areas adjacent to the Chestnut property. Responses to other disturbance factors (Table 8) have varied considerably between years and have been most likely confounded by the influence of weather and the effect of accumulative disturbance. Future analysis will weigh the effects of these other variables. It is clear that birds will leave Chestnut and move to other wetlands during the hunting season without being disturbed. Of 104 radio-marked bird departures from Chestnut during the hunting season after the 3 year period, 32% occurred without any prior disturbance on that particular day, and additional 21% were not direct responses to harassment (Table 9). Table 9. Comparison of the number of radio-marked bird departures initiated without disturbance with those immediately following some type of disturbance. See methods for description of disturbance categories. Disturbance 1986-87 No disturbance at all Number of de~artures 1988-89 1987-88 All Years 4 16 13 33 Some disturbance, but not preceding departure 12 3 7 22 Disturbance prior to departure (categories 1-5) 24 8 17 49 40 27 37 104 Total �40 Response to Simulated Hunting Disturbance When simulated hunting disturbance began on 25 January, 3 ducks marked on 11 January and 2 birds marked on December 7-8 remained on Chestnut Slough. Shots fired at sunrise were the primary disturbance factor used during simulated hunting, but sometimes birds flushed from the investigator walking just prior to the shot being fired. Only 2 shots were fired in the afternoon. On many occasions birds left Chestnut prior to sunrise with no disturbance (Table 10), even during the extreme cold experienced during the test period. Of 45 radio-bird day occurrences on Chestnut, birds remained on Chestnut on 6 occasions, and departed prior to disturbance 22 times. On 5 of 6 occasions, birds remaining on Chestnut experienced 1 shot greater than 200 m away, while a shot was fired in the afternoon near the sixth bird. Usually a shot fired near the bird stimulated departure from Chestnut. Once leaving, the Chestnut roost birds did not return to Chestnut during the daylight hours. They did fairly consistently return to roost at night 10 to 13 hours after leaving. The rate at which birds returned to roost the first night was independent of whether the bird left in response to disturbance or not (x2 - 1.50, P > 0.10). However, on the 3 occasions that birds were shot at multiple times before departing, the birds abandoned Chestnut for an extended period (Table 10). Table 10. Simulated hunting disturbance, bird reaction, and roost homing for radio-marked birds at Chestnut Slough, post-hunting season, 1989. Disturbance Facto! At De12arture AgeLSex AM Bird # Date 266 1/25 1/27 1/29 1/30 1/31 2/1 2/2 2/3 1690 1/25 1/26 1/27 1/28 1/29 1/30 1/31 2/1 2/2 2/12 2/13 0 *a * * * * * * * * 2 * * 1* * * * 1 Hours Acc. Rating Tem12 F before return 0 30 34 0 16 28 0 2 10 5 17 10 1* 0 37 11 0 13 11 0 12 -4 0 Did not return -14 0 30 10 15 5 10 1* 0 16 10 0 25 10 2 (Did not 2b depart) 17 <8 2 11 0 37 13 11 5 1* 227 10 -4 2* 30 13 5 1* 20 0 12 5 �41 Table 10 (cont.). D;LstuIbance Eactor At De~aIture Age/Sex Bird # 2650 AM 2650 Date 1/25 1/26 AF 1M 281 220 1/28 1/29 * 1400 1M a b c d 1400 2/1 1/31 2/1 2/2 1/25 1/26 1/27 1/28 1/30 1/31 2/1 2/2 5 1* * 1/25 1/26 1/27 1/29 2 1 1/27 1/30 1/31 2/1 2/2 2/5 2927 0 1 1 1 * 1* 1* 3* 1* 1 * 1*c * 1c * 2* Acc. Rating Temp F 5 30 2 (Did not lOb depart) 5 16 0 25 2 (Did not 2 depart) 2 17 " 0 37 5 11 5 -4 15 -27 5 2 0 5 0 5 0 10 " 30· lOb 16 30 " 13 40 13 -4 1* 0 5 0 0 0 0 5 30 15 16 25 17 37 13 1* 5 -4 * 1* * * * * Hours before return 10 10 10 11 10 34 Did not return 10 34 Did not return " 10 Did not return 11 9 9 -34 11 11 11 Did not returnd Indicates disturbance factor associated with departure from Chestnut Minimum temperature on that date (Greeley NOAA) Disturbance in the afternoon Died on 2/10/89 �42 Behavior Thirty-two scan samples were used to catagorize the behavior of 1,600 ducks during winter 1988-89 (Fig. 34). As in previous years, resting and swiming were the predominate behaviors. The quantity of behavior data collected during 198889 were insufficient to examine for differences by habitat type, sex or age. - SWIM REST OTHER LAND FEED WALK PREEN WATER FEED Fig. 34. Behavioral time budget of wintering mallards, 1988-89. �43 DISCUSSION Subtle changes in habitat availability and management may have been responsible for the small home range sizes observed in 1988-89 compared to previous years. The property manager and hunters at Chestnut Slough made efforts to reduce needless disturbance to ducks, with apparent success. Telemetry locations were clustered around Chestnut and the adjoining South Platte River more than in previous years, suggesting that habitat requirements could be met within this local area. This effectively reduced the home range size of many instrumented birds. Additionally, no use was made of Milton Reservoir and the Beebe Draw region just north of Milton, and little use waS made of Latham Reservoir. Because these lakes are located along the eastern edge of the study area, avoidance of these wetlands reduces travel distances and associated home range size. Relative to their availability (0.6% of the surface water area), warm-water wetlands were once again highly selected by radio-marked mallards in 1988-89. As noted by other researchers, these wetlands provide food resources and thermal refugia for waterfowl during winter. Mallards respond to a hierarchy of habitat selection that, at its highest levels, is dependent upon the availability of ice-free wetlands weighted by such disturbance factors as hunting (Ringelman et al. 1989). When weather allows, mallards prefer lakes and small ponds over rivers, holding ponds and ditches. Under snowfree conditions, ducks prefer to feed in cornfields rather than feedlots. However, mallards showed plasticity in their response to these and other conditions, adapting their movements and behavior to short-term weather events. Temporal and spatial variability in habitat use also relate to social events that occur during winter. Most mallards establish pair bonds during midwinter in Colorado, and wetlands such as lakes and warm-water sloughs are thought to provide "courtship arenas" for mallards on the High Plains (Ringelman et al. 1989). After pairing, mallards vacate such wetlands in favor of locations that provide greater isolation from conspecifics. LITERATURE CITED Altmann, J. 1974. Observational study of behavior: sampling methods. Behaviour 49:227-267. Dwyer, T. J. 1972. an adjustable radio-package for ducks. Bird-banding 43:282-284. Fenneman, N. M. 1931. Physiography of the western United States. Book Company, New York and London. 534pp. McGraw-Hill Gittings, E. B. 1941. Climate of Colorado. Pages 798-808 in Yearbook of Agriculture: Climate and Man. U.S. Dep. Agric., Washington, D.C. Ringelman, J. K., W. R. Eddleman, and H. W. Miller. 1989. High plains reservoirs and sloughs. Pages in L. M. Smith, R. L. Pederson, and R. M. Kaminski, eds. Habitat management for migrating and wintering waterfowl in North America. Texas Tech Press, Lubbock. �44 LITERATURE CITED (cont.) Ringelman, J. K., and M. R. Szymczak. 1988. Habitat use by mallards along the Front Range of Colorado. Job Progress Report Project 01-03-212, Colorado Division of Wildlife, Fort Collins, Colorado. Szymczak, M. R. , and J. F. Corey. 1976. Salt Plains duck trap in Colorado. Prepared by: ~ j(. ~ James K. Ringelman Wildlife Researcher C Construction and use of the W 71J;,£JP Michael R. Szymczak Wildlife Researcher C �45 Colorado Division of Wildlife Wildlife Research Report October 1989 JOB PROGRESS REPORT State of __~C~o~l~o~r~a~d~o~~ Project Work Plan _ Migratory Game Bird Research 01-03-212 1 : Job 18 Job Title: Winter survival and reproductive success of female mallards Period Covered: Authors: 01 April 1988 through 31 March 1989 Clinton W. Jeske, James K. Ringe1man. and Michael R. Szymczak Personnel: D. Anderson, J. Corey, D. Gilbert, M. Gilbert, M. Miller, Colorado Cooperative Fish and Wildlife Research Unit; M. Albrecht, A. Archuleta, C. Jeske, J. Laake, E. Rexstad, R. Scarpe1la, D. Smith, K. Trust, K. Wilson, Colorado State University; J. Corey, D. Freddy, R. Hopper, K. Navo, J. Ringelman, M. Szymczak, Colorado Division of Wildlife; M. Nail, R. Schneiderbeck, U.S. Fish and Wildlife Service. ABSTRACT Mallards were captured on the Monte Vista National Wildlife Refuge during 12-13 December, 12-17 January, 9-10 February, and 27 February-2 March. Age, sex, and condition index data were obtained for each bird. Monel wing bands were affixed to mallards captured in January, and 54 adult female mallards were radio-marked with neck collar transmitters in March. In addition to monitoring behavior and mortality of instrumented ducks, systematic carcass searches were conducted throughout the winter to provide samples used to examine (1) differential mortality rates based on age, sex and condition (determined from wing bands), (2) percentage of carcasses with indications of starvation, (3) percentage of carcasses showing traces of lead poisoning, and (4) percentage of carcasses testing positive for avian cholera. Within each age-sex class, mean body weights differed little among banding periods. However, individual birds recaptured during subsequent trapping periods lost weight. Mallards tended to get their bills caught in the flexible neck collars originally used for radio attachment; 67% of the radios were recovered within a month. Subsequent collar designs proved better, but only 3 instrumented females remained alive by the May nesting period. Of the 1,677 carcass remains (wings) collected, 131 were from banded birds. Most remains were of adult males (40%), followed by adult females (28%) and immatures (16% each sex). Banded birds found dead did not differ in condition from the sample of birds at banding, but starvation was the cause of death among 24% of the banded birds compared to 7% of the unbanded ducks. Aerial counts from color slides taken on 4, 11 and 18 January indicated an average population of 24,340 birds. Since only about one-third of the existing carcasses were detected by search crews, the 1,677 wings recovered by search crews extrapolates to an estimated 4,736 dead birds, or about 20% of the January population. Avian cholera and predation were the principal causes of mallard mortality. ��47 VINTER SURVIVAL AND REPRODUCTIVE SUCCESS OF FEMALE MALLARDS Clinton W. Jeske James K. Ringelman Michael R. Szymczak P. N. OBJECTIVES 1. Quantify winter survival rates of adult female mallards in poor and good condition. 2. Determine the sources of winter mortality for mallards in the San Luis Valley. 3. Investigate the relationship between winter body condition and reproductive performance of mallards wintering in the San Luis Valley. 4. Measure the winter energy balance of mallards by quantifying components of energy acquisition and depletion. SEGMENT OBJECTIVES 1. Capture and radio-mark 80 adult female mallards during the second week of January. After estimating the distribution of condition indices for the population, affix radios to 40 birds in the upper one-third and 40 radios to birds in the lower one-third of the condition distribution. 2. Mark a minimum of 1,000 mallards (250 of each age and sex) with wing bands and U.S. Fish and Wildlife Service leg bands. 3. Monitor radio-marked ducks by vehicle and aerial tracking throughout the January-March period. Document movements and status (alive, dead, paired, unpaired) of all birds. Attempt to ascertain the cause of mortality. 4. Conduct systematic surveys to obtain carcass remains (wings) of dead mallards. Evaluate wings for (1) wing bone fat, Pasturella bacteria, and lead in the wing bone, and (2) the presence or absence of wing bands. 5. Estimate winter mortality rates with maximum-likelihood procedures using competing risks theory to account for radio-marked birds whose fate is unknown. 6. Conduct respirometry experiments to determine the lower critical temperature and basal metabolic rates of winter-acclimated mallards. 7. During April-June, locate radio-marked mallards during nesting and obtain information on reproductive performance. Measurements will include egg mass, egg dimensions and number, total clutch mass, fertility(hatching success, nesting success, and renesting frequency. Relate reproductive performance to mid-winter body condition. �48 8. Quantify the nutritional value of barley through (1) digestabili ty experiments with penned ducks, and (2) periodic sampling of barley fields to evaluate losses in food value as a result of weather-induced nutrient leaching. STUDY AREA This study was conducted in the San Luis Valley (SLV), located in south-central Colorado. The SLV is a 12,960 km2 intermountain basin, bounded by the San Juan Mountains to the west and the Sangre de Cristo Range to the east. As a result of the surrounding mountains and 2,286 to 2,438 m elevation, January temperatures average -4 C, and snow and ice fog are common. Upland areas are dominated by greasewood (Sarcobatus vermicu1atis) and rabbitbrush (Chrysothamnus spp. ). Common wetland vegetation includes baltic rush (Juncus balticus), cattail (Typha 1atifo1ia), hardstem bulrush (Scirpus acutus), coontail (Ceratophyllum demersum) and pondweeds (Potamogeton spp.). Further descriptions of the SLV can be found in Hopper et al. (1975) and Szymczak (1986). An average of 14,700 mallards winter in the SLV (1982-84 average, Colo. Div. Wi1d1., unpubl. data), with most residing on the Monte Vista National Wildlife Refuge (MVNWR). MVNWR is 5,670 ha in area and has over 200 ponds and impoundments. Wintering ducks are concentrated on those water areas which do not freeze as a result of continuous pumping or artesian flow. Mallards breed throughout the SLV and surrounding mountains (Rutherford and Hayes 1976, Szymczak 1986), but nesting is concentrated on the MVNWR (pers. obs.). METHODS Weight Changes of Marked and Unmarked Mallards Mallards wintering on the Monte Vista National Wildlife Refuge (MVmlR) were captured with Salt Plains bait traps (Szymczak and Corey 1976) and cannon nets (Dill and Thornberry 1950) from 12-13 December 1988, 12-17 January, 9 -10 February, and 27 February-2 March 1989. Data on age, sex, weight, and wing length were obtained for all birds, and a condition index (Ringe1man and Szymczak 1985) calculated for each. All mallards were banded with standard USFWS leg bands. During the January trapping period, most mallards also had monel bands affixed to the shaft of primary_feather VIII of both wings. Mean weights were plotted by trapping period for each age-sex class. The UNIVARIATE Procedure (SAS Institute 1985) was used to test normality of weight and condition index distributions by age-sex class. Duncan's mu.lt Lp Le range test was used to compare weight and condition index means by trapping period and agesex class. To examine the relationship between weights and trapping method, weights of birds captured in cannon nets during the 9-10 February trapping period were compared with weights of birds captured in bait traps during that same period. Mean weights of birds recaptured within a banding period, as well as mean weight differences of birds recaptured during subsequent trapping periods, were compared with Duncan's Multiple Range test. Weights and condition indices of mallards originally captured in 1987 and 1988 and recaptured in 1989 were examined to see if a bird in relatively "good" or "poor" condition remains in that relative weight or condition class in subsequent years. Weight and condition indices were corrected for different mean condition values between years by subtracting mean weight and condition index for the respective age-sex class from each respective trapping period. �49 Instrumentation Five male and 5 female mallards were instrumented with subcutaneous implant radio transmitters 12-13 December 1988. Birds were anesthetised, then a 9 g transmitter was implanted dorsally at the base of the neck with the 16 cm whip antennae inserted subcutaneously to the hip joint. Anterior and posterior sutures on the transmitter helped hold it in place. During the same trapping period, 5 male and 5 female mallards were instrumented with radio packages similar to those used in 1987 and 1988. The latter birds were instrumented to aid in assessing the comparative effects of the subcutaneous implant radios and to provide an estimate of foraging time through continuous monitoring. During the January trapping period, 5 adult females were instrumented with 9 g transmitters glued on a plastic-coated nylon neck collar. These birds were monitored to evaluate the use of the collar for attaching the transmitters. Poor survival of mallards instrumented with implant transmitters prompted the use of a neck collar radio attachment. During the March trapping period, 54 neck collar-attached transmitters were placed on adult females. Instrumented birds were monitored regularly for the first 3 weeks after instrumentation and then not located again until mid-May. Due to high mortality of females instrumented with the flexible collar, we designed a collar made of polyvinal chloride (PVC) that would reduce mortality resulting from the bird getting its bill caught in the collar. Eight of these units were deployed during April 1989. During May, instrumented females were located from the ground and air as in previous years (Ringelman and Szymczak 1988). When an instrumented female was located, a ground check was made to determine whether the birds was nesting. If the female was nesting, egg length, breadth, and weight were recorded for each egg. Nest location, cover type, and incubation stage were also recorded. After the estimated hatching date, the nest was checked to determine nest success and to collect a down sample. Carcass Collection As in 1987-88, wintering waterfowl concentration areas and raptor perches were searched for waterfowl wings or carcasses from 20 January - 1 April 1989. Species, age, sex, marsh unit where collected, date, and band numbers (if bands were present) were recorded. If an ulna was available, one wing was retained and frozen from all mallards collected by our reseach crew, and from all banded birds collected independently by the refuge staff. Frozen wings were later analyzed for lipids in the ulna (Ringelman and Szymczak 1988). The January condition indices of the wing-banded sample of mallards were compared, by age and sex, to the condition indices of the wing-banded mortalities to determine whether the condition distributions of those birds dying differed from the distributions of those birds marked. Each sample was analyzed for normality using the UNIVARIATE procedure (SAS Institute 1985) and a Wilcoxon rank sum test to compare condition distributions. Wings retained for analyses were labeled and frozen in airtight bags. Ulnas were excised, cracked open, and the marrow touched to a clean sheet of white paper. The resulting blot was examined as in 1987-88 to evaluate lipid content of the ulna (Ringelman and Szymczak 1988). A 2x2 chi-square table was used to determine if similar percentages of banded and unbanded birds appeared to have starved. �50 To determine the percentage of carcasses collected in the areas searched, 65 carcasses with wing bands were placed around 8 water areas. Carcasses were placed in areas where remains had previously been recovered, and in some cases marked carcasses were exchanged for unmarked carcasses. Only the person planting the carcasses knew when and where carcasses were planted. Afterwards, collected carcasses were examined for wing bands from the planted carcasses, and wing band and marsh unit number recorded. Aerial Photographs Beginning 28 December 1988 and ending 15 February 1989, weekly aerial censuses were conducted on the MVNWR using a Cessna 182. Color, 35mm slides were taken from 500 ft. elevation through an open window with a Minolta Fl equipped with a 200 mm lens and skylight filter. An attempt was made to overlap photos. After development, transparencies were projected on a screen, overlap areas excluded, and ducks counted using a mechnical marker pen counter. Numbers of identifiable waterfowl were recorded for each flight. Mallard sex ratios were determined from transparencies taken on 28 December 1988 and 18 January 1989. Field-feeding Trials Wire panels were used to enclose 64 m2 of standing barley on MVNWR. Twelve captive mallards (6 of each sex) were transported to the enclosure daily until they appeared to forage normally. Thereafter, the enclosure was moved to a site with abundant standing grain, and the birds starved for 18 hr. Six mallards (3 of each sex) were weighed to the nearest gram on an electronic digital scale immediately before the trial. Mallards were released gn ~ into the enclosure which had been moved to a site with abundant standing grain. Test birds were allowed to forage for precisely 5 minutes then reweighed. The amount of barley ingested was assumed to equal the difference between starting and ending body mass. Particle Marker Fluorescent particles (AX series pigments, Day-Glo Color Corporation) were placed in the inflow to the impoundment in Unit 24 on the MVNWR on 3 and 5 January 1989. Particle marke r was mixed as suggested by Godfrey (1987) and siphoned into the inflow. Other water areas on the MVNWR were allowed to freeze so the only open water was in Unit 24. During subsequent trapping periods birds were examined under a high-intensity ultraviolet light for fluorescent particles. Initially, a bird was considered marked if one particle was found. During the March trapping period, a bird was considered marked if multiple particles were found. i Pair and Brood Counts A survey route was established on MVNWR and driven weekly from 22 April through 14 June 1989. Male mallards encountered were classified as paired (with female), lone male (single isolated drake without a visible associated hen), or grouped (2-4 associated drakes). Groups of 5 or more males were excluded. Broods encountered along the route or at any other time were recorded as to 5pecie~, number, and age. Brood ages were used to back-date to the estimated hatching date. �51 Capture of Nesting Females Mallard nests were located on MVNWR. During the initial encounter with the nest, the clutch size and incubation stage were recorded. During the final 10 days of incubation, we attempted to capture the female with a hand-held net, net gun, or nest trap. When a female was captured, greater secondary coverts 5 and 11, and primary V were removed for later analysis to determine the female's age (Gatti 1983). Clutch age, and egg length, breadth, weight, and fertility status were recorded. Nests were checked after the estimated hatch date to determine fate. Down Collection Differences in the trace element composition of grain consumed in winter could result in differences in the trace element composition of female basic plumage, which is normally produced during winter (Weller 1976). Such differences could then be used to evaluate the productivity of San Luis Valley winter residents compared to migrants originating south of the SLV. To address this hypothesis, barley grown in the San Luis Valley, corn grown near the Bosque del Apache National Wildlife Refuge, New Mexico, and corn grown near Lubbock, Texas, was harvested and analyzed for trace elements using an inductively coupled plasma machine at the Colorado State University Soils lab (Table 1). The general area of grain collections are locations in which mallards nesting in the San Luis valley may winter. Table 1. Analysis of grain samples for selected elements. All values are in ppm. Analyses done on 19 January 1988 at the Colorado State University Soils Lab using an inductively coupled plasma machine. Element Location Grain Mn San Luis Valley Barley San Luis Valley Barley Parmer, Texas Cu Ti Sr Ba 15.8 7.2 0.3 3.47 1.38 18.7 5.8 0.8 3.58 2.28 Corn 7.6 2.0 <0.1 0.56 0.19 Castro, Texas Corn 6.0 2.9 <0.1 0.38 0.11 New Mexico Corn 5.0 1.6 0.2 0.59 0.20 New Mexico Corn 7.4 1.5 <0.1 0.47 0.07 New Mexico Corn 7.2 1.6 <0.1 0.32 0.03 New Mexico Corn 7.3 1.6 <0.1 0.65 0.12 New Mexico Corn 6.8 1.6 <0.1 0.39 0.09 �52 Captive females were fed barley (from the SLV) or corn during the prebasic molt. After molt was completed, down was harvested from the live birds, then washed in an ultrasonic bath for 5 minutes each with diethyl ether, distilled water, and acetone. Acid digestion and analysis of down samples were conducted at the Soils lab in the same manner as with grain samples. Down samples were then compared for differences in trace element composition similar to those detected in grain samples. Nest bowls containing down from wild females nesting on MVNWR were collected after hatch or predation and placed in plastic bags. These samples were processed in a manner similar to the down collected from the captive birds. RESULTS Weight Changes of Marked and Unmarked Mallards Weight and condition indices were approximately normally distributed. During each banding period, males weighed more than females and adults weighed more than immatures (Duncan's Multiple Range tests, P<O.OS). Mallards weighed more during the December trapping period (Duncan's Multiple Range test, P<O.05), whereas weights during the other 3 periods were similar (Table 2). Condition indices were highest during December, followed by March and February. The lowest condition indices were recorded during the January trapping period (Duncan's Multiple Range test, P<O.OS). Adults were in better overall condition than immatures. Adult males and females had similar condition indices during each trapping period except in March, when adult females were in better condition (Table 3). Generally, immature males and females were in similar condition during each trapping period (Table 3). During February, 54 mallards were captured with bait traps and 139 with cannon nets. Weights of these mallards did not differ by capture method (Duncan's Multiple Range test, P-O.9S32). Mean weights of mallards were similar in all trapping periods (Figs. 14). Twenty-eight adult females, 184 adult males, 39 immature females, and 58 immature males were recaptured within 7 days. These mallards tended to weigh less than their original weights (T--2.20, P-O.0367, T-2.70, P=O.0076, T=-2.14, P=O.0387, andT--4.80 P-O.0001, for adult females, adult males, immature females, and immature males, respectively; Figs. 5 and 6). Mallards captured during one period and recaptured during su~sequent trap periods also tended to have lower weights at recapture (Table 4). Fifty-seven mallards banded in 1986-87 were recaptured in 1988-89: 1 adult female, 33 adult males,S immature females, and 18 immature males. Condition indices in 1988-89 were positively related to condition indices in 1986-87 for adult males (F-9.24, 32 df, P-O.0048), but not for immatures of either sex (F-2.S7, P-O.1283 and F-l.56 , P-O.300l for males and females respectively). Relative condition indices were positively related for adult males (F=9.13, P=O.0050, Fig. 7) but not for immatures (F-3.43, P-O.0827 and F-l.40, P=O.3224, Figs. 8 and 9 for immature males and females, respectively). One hundred and sixty-three birds (18 adult females, 92 adult males, 14 immature females, and 39 immature males) captured in 1987-88 were recaptured in 1988-89. Condition indices for adult females in 1989 were not related to those in 1988 (F-2.22, 17 df, P-O.1553), but they were positively related for the other age-sex classes (F-S.44, 91 df, P-O.02l9, F-17.l4, 38 df, P-O.0002, and F-15.18, 13 df, P-O.002l for adult males, immature males, and immature females, respectively). Relative condition indices in 1987-88 were positively related �53 1200 1150 r- 1100 r- 1050 r- (43) I- 1000 r- e 950 ~ 900 r- 850 f- 800 r- - 0) J: iii > o o [Xl 750 .•. UNMARKED 6. RECAPTURE .. ( 181) - (417) (50) T .- t ~ [ (21) (8) (30) - 700 DEC JAN TRAPPING FEB PERIOD MAR Fig. 1. Mean weights (± SE) of adult female mallards captured in the San Luis Valley, Colorado, during 4 trapping periods in 1988-89. Sample sizes are given in parentheses with unmarked samples above and recaptures below the standard errors. 1200 (78) (82) 1150 ( 1158) 1100 C, •...• 1050 ~ 1000 ~ iii 950 ~ > o 900 [Xl 850 o [195) t - (83) I UNMARKED 800 RECAPTURE 750 (22) 700 DEC JAN TRAPPING FEB MAR PERIOD Fig. 2. Mean weights (± SE) of adult male mallards captured in the San Luis Valley, Colorado, during 4 trapping periods in 1988-89. Sample sizes are given in parentheses with unmarked sample above and recapture sample below the standard errors. �54 1200 .• UNMARKED D RECAPTURE 1150 -- 1100 m 1050 I- ::r:: 1000 iii 950 <!J (39) 3: >- I 900 0 0 III 850 -I 800 750 (34) - (85) T (306) II r•• - 1 X I f- c: II (4) (10) (45) 700 DEC JAN TRAPPING MAR FEB PERIOD Fig. 3. Mean weights (± SE) of immature female mallards captured in the San Luis Valley, Colorado, during 4 trapping periods in 1988-89. Sample sizes are given in parentheses with unmarked sample above and recapture sample below the standard errors. 1200 C, 1150 E 1100 ~ 1050 I- ::r:: 1000 Q w 950 3: >- c 0 III 900 (45) P ._A ~I I•... 800 750 700 TII D RECAPTURE A 1 (32) I ~ (9) L t UNMARKED (440) I 850 ii i A -- (131) I II .6. I I Ii i 1 I I 1 - I I (33) (65) I DEC JAN TRAPPING FEB MAR PERIOD Fig. 4. Mean weights (± SE) of immature male mallards captured in the San Luis Valley, Colorado, during 4 trapping periods in 1988-89. Sample sizes are given in parentheses with unmarked sample above and recapture sample below the standard errors. �55 110 ADULT t- ::t: (!J IMMATURE (10) ~ 105 (72) >Q o m I t f I ~) (85) (34) (2) ..J ~ 100 z a a:o t- Z w 95 ( 17) o IX: W e, (28) (1) ••• (10) 90 1 2 3 4 DA YS BETWEEN 5 6 7 RECAPTURE Fig. 5. Weights (as percent of initial capture weight) of adult and immature male mallards recaptured within 7 days of initial capture in the San Luis Valley, Colorado, during 198889. Sample sizes are given in parentheses with adults above and immatures below the standard errors. 110 t- ::t: (!J W == 105 >Q o m ..J ~ 100 - a a: o t- Z w 95 o IX: i- (1) r I 90 6 IMMATURE I r r II (7) (9) 6. 1 tT- tT( 14) e, AOULT ( 11) (7) W ••• t L 6 (1 ) (7) (10) I 1 2 3 4 DAYS BETWEEN 5 6 7 RECAPTURE Fig. 6. Weights (as percent of initial capture weight) of adult and immature female mallards recaptured within 7 days of initial capture in the San Luis Valley, Colorado, during 198889. Sample sizes are given in parentheses with adults above and immatures below the standard errors. �56 en co I co co .•.. 0) 15 89CONDITIONoO.40(87CONDITION)-0.34 '00.48, poO.OOSO 10 ~ )( UJ 5 a ~ z 0 0 z -5 • Ea 0 0 UJ > i= -10 « ..J UJ a: -15 -15 -10 -5 o 5 10 15 RELATIVE CONDITION INDEX IN 1986-87 Fig. 7. Relative condition indices of adult male mallards captured in 1986-87 and recaptured in 1988-89 in the San Luis Valley, Colorado. en co I co co en .•.. 15 89CONDITIONoO.97(87CONDITION).0.O 1 RoO.42, poO.0827 10 ~ )( UJ 5 a ~ z o o ~ -5 E a o UJ > i= -10 « • ..J UJ a: -15 -15 -10 -5 o 5 10 RELATIVE CONDITION INDEX IN 1986-87 Fig. 8. Relative condition indices of immature male mallards captured in 1986-87 and recaptured in 1988-89 in the San Luis Valley, Colorado. 15 �57 0) co I co co ,.. 0) 15 88CONDITION.-0.82(87CONDITION)+2.311 ,.0.51, P·0.3224 10 ~ >< 5 UJ Q ~ Z 0 j: 0 Z 0 -5 is o UJ > j: -10 <t ...I UJ c:: -15 -15 -5 -10 o 5 10 15 RELATIVE CONDITION INDEX IN 1986-87 Fig. 9. Relative condition indices of immature female mallards captured in 1986-87 and recaptured in 1988-89 in the San Luis Valley, Colorado. 15 0) co 8I1CONDITION·0.82(88CONDITION).1.33 ,·0.23, P·0.02110 I co co 10 ,.. 0) ~ >< UJ 5 Q ~ Z o o ~ C ~ o -5 UJ > ~ -10 <t • ...I UJ c:: -15 -15 -10 • • -5 0 5 10 RELATIVE CONDITION INDEX IN 1987-88 Fig. 10. Relative condition indices of adult male mallards captured in 1987-88 and recaptured in 1988-89 in the San Luis Valley, Colorado. 15 �58 for adult males (F-4.92, 91, df, P-0.0290; Fig. 10), immature males (F-13.68, 38 df, P-0.0007; Fig. 11), and immature females (F-16.05, 13 df, P-0.0017; Fig. 12). The relationship for adult females approached significance (F-4.30, 17 df, P-0.OS46; Fig. 13). Table 2. Mean (± standard error) for weights of mallards captured in the San Luis Valley during winter 1988-89. Letters a-d indicate significant differences among means (P-O.OS) within an age-sex class. Letters w-z compare age-sex classes within a trapping period. Trapping Period Adult Immature Female Male Female 12-13 Dec 1063(82)aw n-73 927(92)ay n-43 985(87)ax n-45 882(77)az n-38 12-17 Jan 1025(82)bw n-llS8 893(76)by n-4l7 958(88)bx n=440 83l(88)bz n-306 9-10 Feb 1022(135)bw n-78 894(64)by n-50 953(62)bx n-32 871(68)ay n-34 27 Feb- 2 Mar 1022(79)bw n-254 942(64)ax n-l61 942(75)bx n-13l 855(71)aby n-8S Male Table 3. Mean (± standard error) for condition indices of mallards captured in the San Luis Valley during winter 1988-89. Letters a-d indicate significant differences among means (P-0.05) within an age-sex class. Letters w-z compare age-sex classes within a trapping period. Trapping Period Adult Male Female Immature Male Female 12-13 Dec l4.5(4.l)aw 14.2(S.7)bw ll.9(4.2)ax 12-17 Jan 11.8(4.0)bw 12.0(4.7)cw 9-10 Feb l2.3(3.1)bw l2.4(3.8)cw 10.4(2.9)abx 11.8(4.1)awx 27 Feb- 2 Mar l2.1(3.6)bx lS.7(3.6)aw 10.1(3.9)by 9.9(4.6)bx l3.4(4.S)awx 9.8(4.S)bx l2.3(4.1)ax �59 0) co I co co 0) .,.. 15 S8CONDITION-0.44(SSCONDITION)·0.72 ,-0.52, P-0.0007 10 ~ >< w 5 ~ Z 0 0 0 i= C z 0 -5 i= -10 c w > « • ..J w a:: -15 -15 -10 -5 o 10 5 15 RELATIVE CONDITIONINDEX IN 1987-88 Fig. 11. Relative condition indices of immature male mallards captured in 1987-88 and recaptured in 1988-89 in the San Luis Valley, Colorado. 15 0) CO 88CONDITION-O.88( SSCONDITION)-0.30 I ,-0.78, P- O.0 017 CO co .,.. 0) 10 ~ >< W 5 o ~ z o o ~ -5 Eo o w > ~ -10 -e ..J w a:: -15 -15 -10 -5 o 5 10 RELATIVE CONDITIONINDEX IN 1987-88 Fig. 12. Relative condition mallards captured in 1987-88 San Luis Valley, Colorado. indices of immature female and recaptured in 1988-89 in the 15 �60 0) co I co co .•. en 15 88CONDITION·O.42(88CONDITION)-1.08 P.0.0548 10 ~ >< IJJ ,.0.48, •• 5 C 3; Z o i= o ~ -5 • is o •• IJJ > i= <C -10 ..J IJJ I%: -15 -15 -10 -5 • o 5 10 15 RELATIVE CONDITION INDEX IN 1987-88 Fig. 13. Relative condition indices of adult female mallards captured in 1987-88 and recaptured in 1988-89 in the San Luis Valley, Colorado. 250 , LIMITED SEARCH TIME , LIMITED SEARCH TIME 200 ~ 150 I%: co o <C IJJ c 100 50 \ o FEB 1 FEB 15 MAR 1 MAR 15 APR 1 WEEK Fig. 14. Number of dead waterfowl collected weekly between 17 January and 1 April 1989 on the Monte Vista NWR. �61 Table 4. Mean differences (± standard error) in weights of mallards trapped during one banding period and recaptured during subsequent periods. Means with different letters are significantly different (P<O.OS). Initial Capture Period Recapture Period Male Adult Female Dec Jan -46(66)a n-24 -54(37)a n-4 Dec Feb Dec Mar -60(64)a n-3 -38(12)a n-3 Jan Feb Jan Mar -64(201)a n-22 -26(66)b n-40 -9(34)a n-6 36(66)a n-14 Feb Mar l27(369)a n-8 Sl(30)a n-2 Immature Male Female 1(81)a n-9 -102 n-l -251b n-l -42(36)a n-7 -13(54)a -14(43)a n-8 n-5 -23(67)b 3(6l)ab n-29 n-7 16a n-l -31(26)a n-2 Instrumentation Poor signal reception and high mortality of the mallards instrumented with the subcutaneous transmitters in December prevented use of these birds for monitoring foraging behavior. Two of the birds with subcutaneous implant radios were killed by hunters. One of the back-pack harness radios applied in December was shot. Each of four birds instrumented with back-pack transmitters were monitored for 16 hr periods. They spent 48.5 (± 15.2) min field-feeding, while the field-feeding flights on these days lasted 133.6 (± 39.5) min. Thirty-six of 54 females (67%) instrumented in late February and early March were recovered during the month. All birds had their bill caught in the neck collar. Those collars recovered without accompanying carcasses also showed evidence of the bill being caught (top edge rolled over in 2 places approximately 3/4 of an inch apart). Of the 18 instrumented females not recovered, only 9 could be located during the final week of the March. Radio packages on the 9 missing birds may have failed because of broken antenna; 13 radios recovered had broken antenna. In March, 6 birds (3 wing-clipped and 3 free-flying) were equipped with radio collars constructed of PVC pipe material. Two of the 3 wing-clipped birds released on 21 March were recaptured 24 March. These birds, a male and a female, had lost lS8g and 116g, respectively. The third bird, a female that was crippled when released, was found dead. The 3 free-flying birds, all females, were subsequently located with other mallards, but were not making f i eLd-feeding flights as most other mallards appeared to be doing. An additional five females were equipped with these radio collars in April. �62 Three females (2 with nylon collars and 1 with a PCV collar) were alive in May. Only 1 female nested. The female with the PVC collar, instrumented in April, nested on the MVNWR in baltic rush (Juncus balticus). The nest was located on 7 June 1989, had 5 eggs, appeared to have been partially predated, and was at about 17 days incubation. She ultimately hatched 3 eggs. Carcass Collection One thousand, seven hundred and thirty-eight wings were collected by the maintenance crew and our research crew. Of these, 131 wings were from banded birds. Numbers of carcasses collected on a weekly basis fluctuated (Fig. 14), primarly due to variable search effort. Suspected causes of death included avian cholera, starvation, lead poisoning, predation, and collisions with powerlines. Of the 591 carcasses whose age and sex could be determined, 283 (40.3%) were adult males, 96 (16.2%) were immature males, 162 (27.4%) were adult females, and 95 (16.1%) were immature females. The distribution of condition indices for banded birds found dead were not different from the distribution of condition indices of their cohorts during January banding (z--1.64, P-0.1000; z--O.72, P-0.4689; z--l.33, P-0.l845; z--0.72, P-0.472l Wilcoxon sign rank test for adult females, adult males, immature females, and immature males, respectively), indicating death was independent of condition. No apparent peaks in mortality related to condition class at banding were apparent (Figs. 15-18). Of the 1,017 ulnas analyzed for lipid content, 68 (6.7%) were from birds that had starved. Twenty-six of 106 banded birds (24.5%) found dead and analyzed for ulna lipids appeared to have starved. Banded birds apparently starved at a greater rate than unmarked birds (42 of 911 samples, 4.6%; X2-60.38, P-O.0000) . Fifteen of 20 (75%) samples from instrumented females appeared to have starved. Of 65 carcasses wing-banded and placed out to determine recovery rates, 41 wings (31.5% of the 130 wings), representing 23 birds (35.4%) of the birds, were recovered. Eight leg bands recovered from an eagle roost were from birds also wingbanded. Of these eight birds, only 1 (12.5%) had been recorded from our searches of the refuge. Recovery of the planted carcasses suggest that the 1,677 wings recovered represent 4,736 dead birds. Aerial Photographs Counts from aerial photographs were too variable to give identifiable population trends (Table 5). Sources of inconsistencies included bird movements to areas not photographed, photographs too blurred for accurate counts, inclusion of other waterfowl species in the counts, and 1 roll of film lost by the developer. Sex ratio counts of 1,830 mallards on 28 December 1988 and 1,135 on 18 January 1989 indicated 59.9% and 58.8 % males, respectively. The decrease in the number of males was significant (X2-6.4846, P-0.0109). Field-feeding Trials When trials in which birds appeared to lose weight are excluded, females obtained 0.92 g/min (± 0.77, range 0-2.4, n-27) and males 1.50 g/min (±O 99, range 0-4, n-31) of barley from fields on MVNWR. There were differences in the rate of consumption between sexes (F-6.05, P=0.0170) and individuals (F=8.69, P-0.0047) but no individual*sex interaction (F-0.89, 1 df, P-0.3507). Males consumed standing grain heads while females were only observed foraging on shelled grain. �63 20 0 15 ~ "GOOD" CONDITION ~ "AVERAGE" CONDITION •• LIJ I0 LIJ ..J ..J 0 10 "POOR" CONDITION 0 CC LIJ £D ~ ;:, Z 5 o JAN 15 FEB 1 FEB 15 MAR 1 MAR 15 APR 1 WEEK OF COLLECTION Fig. 15. Number of weeks between release and recovery of wing-banded adult female mallards in the San Luis Valley, Colorado, 1989. 20 ~ "GOOD" CONDITION ~ "AVERAGE" CONDITION •• "POOR" CONDITION o 15 LIJ I- o LIJ ..J ..J o o 10 CC LIJ CD ~ ::::) Z 5 o JAN 15 FEB 1 FEB 15 MAR 1 MAR 15 APR 1 WEEK OF COLLECTION Fig. 16. Number of weeks between release and recovery of wing-banded mallards in the San Luis Valley, Colorado, 1989. �64 20 0 15 w .... .... 0 "GOOD" CONDITION ~ "AVERAGE" CONDITION •• w t0 0 ~ "POOR" CONDITION 10 a:: w CD ~ ::) Z 5 o JAN 15 FEB 1 MAR 1 FEB 15 MAR 15 APR 1 WEEK OF COLLECTION Fig. 17. Number of weeks between release and recovery of wing-banded immature female mallards in the San Luis Valley, Colorado. 20 !2Z2l "GOOD" CONDITION ~ "AVERAGE" CONDITION •• "POOR" CONDITION o 15 w t- O w .... .... o 10 o a:: w CD :E ::) Z 5 o JAN 15 FEB 1 FEB 15 MAR 1 MAR 15 APR 1 WEEK OF COLLECTION Fig. 18. Number of weeks between release and recovery of wing-banded immature male mallards in the San Luis Valley, Colorado. �65 Table 5. Number of waterfowl counted from color slides taken during aerial flights. Flight date Number counted Nunmber + possible missed 28 December 1988 32,325 32,841 4 January 1989 22,697 23,001 11 January 1989 21,955 22,012 18 January 1989 27,413 28,011 25 January 1989 14,119 14,464 1 February 1989 26,350 27,056 8 February 1989 Developer lost 1 box - it is still missing 15 February 1989 21,003 21,557 Particle Marker Most if not all of the mallard population concentrated on the wetland with particle marker during early January, thus we assumed that all birds were exposed to and initially marked. By March, birds showed large patches of marker on the wing lining and around the vent. As a test to verify that only winter residents were marked, blue-winged teal (Anas discors) and other species not resident during winter in the SLV were also examined. Unfortunately, some individuals of these species showed flecks of color similar to those evident on mallards. Further investigation revealed that the source of the marker appeared to be the wetland where the marker was initially applied; captive mallards penned in this wetland were still being marked in March. Subsequent examinations identified a faint, colored sheen that was apparent only on some mallards, suggesting that this marking pattern might be indicative of winter-marked birds. However, lacking a control in the form of known winter-marked birds, the technique was abandoned as a means to identify winter residents. Pair and Brood Counts Numbers of paired, lone, and grouped males varied among census periods (Fig. 19). In mid-June, mixed flocks of males and females were observed, but no large concentrations had developed. A hatching curve (Fig. 20) developed from observations of 142 mallard broods indicated a prolonged hatch with no discernable peak, similar to that which occurred in 1987 (Ringelman and Szymczak 1988). The initial estimated hatch dates of the first week in May precede that of 1987 and 1988 by 2 weeks. �66 250 •• PAIR 200 ~ LONEMALE 175 ~ GROUPED MALES 225 o w ~ 150 o o 125 ::l en w ..J <t 100 ::E 75 50 25 o 4/22 4/30 5/6 5/14 5/22 5/30 6/8 6/14 DATE OF COUNT Fig. 19. Number of male mallards counted from a census route on the Monte Vista NWR and classified as paired, lone, or grouped from mid-April through mid-June, 1989. 30 25 en 0 0 0 20 a: CD IL 0 15 a: w CD ::E ::l 10 Z 5 0 MAY 1 MAY 15 JUNE 1 JUNE 15 JULY 1 WEEK Fig. 20. Estimated hatching dates of mallard broods encountered in the San Luis Valley, Colorado, in 1989. JULY 1 5 �67 Capture of Nesting Females Thirty-four incubating females were captured. Measurements of the feathers collected from the nesting females will be made during the fall of 1989, and any analyses relating female age to clutch parameters made then. Down Collection Preliminary analyses of grain samples suggested identifiable differences between grain grown in the SLV and grain grown in New Mexico, and near Texas. Initial analyses of down samples of captive birds fed these grains did not prove useful in determining which grain sample a bird was fed (Table 6). Further analyses will be conducted in the fall of 1989. Table 6. Concentrations (ppm) of selected trace elements found in down samples collected from 6 captive female mallards fed barley from the San Luis Valley or corn from Colorado. Element Grain Bird Mn Cu Ti Sr Ba Barley 5 5.2 7.0 2 2 50 Barley 8 3.4 5.7 1 1 <1 Barley 46 3.5 5.1 2 1 <1 Corn 1 3.7 6.3 3 1 1 Corn 7 3.4 5.0 1 1 <1 Corn 39 4.0 6.6 3 1 1 DISCUSSION Temperatures were mild (Fig. 21), snow cover light (Fig. 22) and food readily available in 1988-89. Therefore, it is interesting that mean weights of mallards captured in 1988-89 were similar to those of birds captured in 198788, during a more severe winter. The relationship of relative condition indices and relatively constant weights and condition indices between years suggest that condition indices are not completely controlled by environmental factors. The relationship of relative condition indices between years suggest that there may be a genetic component influencing winter condition of mallards in the San Luis Valley. The tendency for weights to be less at recapture may be the result of a handling effect, or that birds in relatively "poorer" condition may be more likely to re-enter a trap. Since many of the year to year recaptures were in �68 DAILY MAXIMUM 20 \ DAILY MINIMUM 10 IU a: 0 ::::) /\ I- f \:'\ ~ a: IU -10 e, \:". ~ .........J\ ,/: IU I- -20 . .... ' " ._ t·, \t :.; -30 DEC 1 DEC 15 JAN 1 JAN 15 FEB 1FEB 15 MAR 1MAR 15 DATE Fig. 21. Daily temperature maxima and minima recorded at the Alamosa airport, Alamosa, Colorado, 1988-89. 10 ""' E CJ ....•. :t: I- fuc 5 - r-1 - :: o Z ._, en o DEC 1 DEC 15 JAN 1 JAN 15 '-----.. FEB 1FEB 15 MAR 1MAR 15 DATE Fig. 22. Depth of snow on the ground at the Alamosa airport, Alamosa, Colorado, 1988-89. �69 relatively "poor" condition, and there are no obvious differences in survival related to body condition, it seems most likely that birds in relatively "poor" condition are more likely to re-enter a trap. As has been previously noted (Ringelman and Szymczak 1988), winter mortality in the SLV is high. Although avian cholera was not as prevalent in 1988-89 as in 1987-88, substantial numbers of birds still died during the winter period. It is difficult to estimate of the total number of ducks dying during January to April 1989. One third of the planted carcasses were recovered, while only 12.5% of the leg bands recovered from an eagle roost were represented by wings (wingbands) recovered during carcass searches. The recovery rate of planted carcasses is probably atypically high since carcasses were placed in areas where they should have been found. The recovery rate of wingbands compared to corresponding bands found at the eagle roost probably is closer to the actual recovery rate of ducks which died on MVNWR. Pair and brood counts suggest that the nesting period was advanced about 2 weeks in 1989 compared to 1987 and 1988. This may be the result of the mild winter and spring in 1989. The prolonged hatching curve (Fig. 20) suggests that female mallards came into breeding condition over a prolonged period, possibly because of a wide variation in winter contion. Field-feeding trials suggest that males may be more efficient at foraging in standing barley fields. This may result from the observation that females foraged on shelled grain while males readily took standing grain. To consume the 57 g of steptoe barley needed for maintenance at -9 to -10 C (Ringelman and Szymczak 1988), a male would have to forage for 38 minutes, while females must forage 62 minutes. Activities, such as flight and courtship, would increase energetic requirements thereby increasing foraging time. LITERATURE CITED Dill, H. H., and W. H. Thornberry. 1950. A cannon projected capturing waterfowl. J. Wildl. Manage. 14:132-137. net trap for Gatti, R. C. 1983. Spring and summer age separation techniques for the mallard. J. Wildl. Manage. 47:1054-1062. GOdfrey, R. D., Jr. 1987. Experimental mass-marking of waterfowl wintering on the Southern High Plains of Texas. M.S. Thesis, Texas Tech Univ., Lubbock, TX. 66 pp. Hopper, R. M., A. D. Geis, J. R. Grieb, and L. Nelson, Jr. 1975. Experimental duck hunting seasons, San Luis Valley, Colorado, 1963-70. Wildl. Monogr. 46. 68pp. Ringelman, J. K., and M. R. Szymczak. 1985. A physiological condition index for wintering mallards. J. Wildl. Manage. 49:564-568. Ringelman, J. K., and M. R. Szymczak. 1988. Winter survival and reproductive success of female mallards. Wildl. Res. Rep. October 1988. pp 69-124. Rutherford, W. H., and C. R. Hayes. 1976. Stratification as a means for improving waterfo.wl surveys. Wildl. Soc. Bull. 4:74-78. SAS Institute, Inc. 1985. SAS procedures guide for personal computers, version 6 edition. Cary, NC. 373 pp. �70 LITERATURE CITED (cont.) Szymczak, M. R. 1986. Characteristics of duck populations in the intermountain parks of Colorado. Colorado Div. Wi1d1. Tech. Pub1. 35. 8pp. Szymczak, M. R., and J. F. Corey. 1976. Construction and use of the Salt Plains duck trap in Colorado. Colorado Div. Wi1d1., Div. Rep. 6. 13 pp. Weller, M. W. 1976. Molts and plumages of waterfowl. pp. 34-38 IN Bellrose. Ducks, geese & swans of North America. Stackpole Harrisburg, PA. Prepared by: )11J"ql72 ~ Michael R. Szymczak Wildlife Researcher C F. C. Books, g-K~ James K. Ringelman Wildlife Researcher C �71 Colorado Division of Wildlife Wildlife Research Report October 1989 JOB PROGRESS REPORT State of __~C~o~l~o~r~a~d~o~ Project 01-03-212 Work Plan 22 Job Title: : Job Migratory Game Bird Research _2_ Migratory Bird Publications Period Covered: Author: _ 01 April 1988 through 31 March 1989 Michael R. Szymczak Personnel: James K. Ringelman and Michael R. Szymczak, Wildlife Colorado Division of ABSTRACT The following list contains those articles that were published or completed and ready for publication during this segment: Ringelman, J. K. 1988. Examining waterfowl condition: skewed ideas on the normal procedure. Pages 277-285 in M. W. Weller, ed. Waterfowl in winter. University of Minnesota Press, Minneapolis. Ringelman, J_. K., W. R. Eddleman, and fl. W. Miller. 1989. l:lighplains -__-_ in L. M. Smith, R, L. Pederson and reservoirs··and sloughs. Pages R. M. Kaminski, eds. Habitat management for migratiI).gand .wintering waterfowl in North America. Texas Tech Press, Lubbock. (In press).· Prepared by: ~ Z; ~ James K. Ringelman Wildlife Researcher C � https://cpw.cvlcollections.org/files/original/6ddb090ef09bd5990c93850e5662be2d.pdf 64e879cd02f114870691d0a815306a6d PDF Text Text JOB PROGRESS REPORT State of ~C~o~l~o~r~a:d~o _ Project: (W-150-R-3) Period Covered: 1 July, 1989 - 30 June, 1990 : Peregrine Falcon Restoration Program Personnel: G.R. Craig, Colorado Division of Wildlife and J.H. Enderson, The Colorado College. ABSTRACT Peregrine falcon population recovery continued in Colorado. In the 1990 breeding season, 44 territories were occupied by 37 breeding pairs that fledged 62 young. The West Slope subpopulation consisted of 35 occupied sites that continued to exhibit good productivity (1.5 young per total pair) without remedial activities. On the East Slope, only 5 of the 9 occupied sites were successful but averaged 1.38 young per total pair which was sufficient for population maintenance. Eggshell thicknesses in 1990 averaged 10% thin (n = 42). Due to insufficient representation (n = 4) East Slope thickness averages were not calculated. Contents of 11 nonviable eggs were submitted for organochlorine analysis. Hacking efforts were confined to three East Slope sites. Fifteen falcons were released and 12 successfully achieved independence. All the hacked young resulted from recycling 5 wild breeding West Slope pairs. This Job Progress Report represents a preliminary analysis and is subject to change. For this reason, information presented herein MAY NOT BE PUBLISHED OR QUOTED without permission of the author. ��PEREGRINE FALCON RESTORATION PROGRAM Gerald R. Craig SEGMENT OBJECTIVES 1. Annually monitor the number of breeding pairs of peregrines and their reproductive success in Colorado. 2. Annually monitor organochlorine peregrines in Colorado. 3. Monitor breeding population turnover through band recoveries, presence of color markers, and telephotographic identification of individual breeding adults. 4. Augment poor wild production by placement of captive hatched wild young and captive produced young into occupied wild nests. 5. Release captive hatched and captive produced young at potential and vacant wild territories. 6. Monitor recruitment of reintroduced pereg:--inesinto the wild breeding population of Colorado. pesticide levels in wild breeding METHODS AND MATERIALS 1. Visit all known peregrine breeding territories throughout Colorado and obzerve them from a distance to establish the presence of breeding adults. Breeding pairs will be kept under surveillance to determine initiation of egg laying. Depending upon the individual female's reproductive history and eggshell condition (obtained through measurement of previous year's eggshell thicknesses) and availability of captive hatched young for release, breeding pairs either will monitored or manipulated as outlined in approach 4. Those pairs not designated to be manipulated will be revisited periodically throughout the nesting season to document reproductive success. When a pair's behavior indicates that egg laying has occurred and tncubat i on is underway, the eyrie will be visited to document the number of eggs produced. The eggs wi 11 be candled to ascertain viability and approximate age. All nonviable eggs will be collected for chemical analysis. A second visit will be made to determine productivity, band nestlings, and collect eggshell fragments and unhatched eggs for thickness measurement and analysis under 2a and 2b. 2a. Eggshell fragments encountered during eyrie visits described in approaches 1 and 4a will be measured for index to thickness following standardized procedures. 2b. Whole, nonviable eggs which are encountered during eyrie visits will be collected, preserved and submitted to the appropriate Fish and Wildlife Service approved laboratory for pesticide analysis. Eggs collected from the wild in the course of Approaches 4a, 4b and 4c that are artificially �at the Peregrine Fund's Boise, Idaho facility also will be submitted for shell thickness measurement and chemical analysis. 3. Peregrines present at breeding territories will be examined to determine the presence of bands or color markers. Band confirmation will be accomp 1ished through observation from a distance with telescopes and concealed remote controlled cameras. When banded falcons are encountered, every effort will be made to read band numbers without trapping or handl ing the birds. It is possible this can be accompl ished in most situations with a Questar field model telescope (80-130x). When band numbers cannot be discerned, attempts will be made to trap and examine the falcon at a time when capture wi 11 have least impact upon breeding activities. 4a. In accordance with an annual release plan developed and approved by the State, u.s. Fish and Wildlife Service, Bureau of Land Management, National Park Service, and the Forest Service, a predetermined number of wild breeding pairs will be manipulated to augment natural productivity. Pairs with a history of reduced clutch size, cracked eggs, or infertile or dead eggs will be candidates for fostering efforts. 4b. On occasion, it may be necessary to recycle several early breeding pairs in order to delay them until captive hatched young of the proper age are available for placement into wild sites. No later than 10 days after the last egg has been deposited, the eyrie will be visited and the entire clutch removed without replacement. Approximately 14 days after removal of the clutch, the pai r wi 11 recycle, select another nest ledge, and deposit a second clutch of eggs. If the eggs are.thin shelled, they may be replaced with plastic replicas and treated as outlined in approach 4a. This technique also works well to augment captive production with wild produced eggs. 4c. At times, pairs will select inferior eyrie ledges that may compromise nest success such as ledges that are too narrow to support a brood of large nestlings, the site may be vulnerable to predators, or it may be exposed to the elements. If the ledge cannot be mechanically improved, pairs can be relocated to other ledges through the recycling method described in approach 4b since they invariably relocate and select a new ledge when recycled. 5. In accordance with an annual release plan developed and approved by the State, U.S. Fish and Wildlife Service, Bureau of Land Management, National Park Service, and the Forest Service, a predetermined number of captive produced falcons will be released at unoccupied or potential sites through the technique of hacking. This technique is employed at locations that do not have the benefit of protection or care from adults. Young falcons of about 35 days of age will be placed in a hack box on a suitable cliff 1edge at the reint roduct ion site. They wi 11 be fed and cared for by attendants until they are flying and capable of fending for themselves. This technique assures that the young become familiar with their surroundings and hopefully will return to the site as adults and take up residency. Hacking requires constant attendance and observation to protect the vulnerable young and assure they have sufficient food while they are dependent upon the hack site. Whi 1e the hack sites wi 11 be �operated by the State, actual costs to operate the sites will be borne by the appropriate land administering agency (Forest Service, Bureau of Land Management, and National Park Service). 6. Confirmed breeding territories and selected potential breeding sites will be surveyed annually to document the presence of released falcons and ultimately determine the success of recovery efforts. RESULTS AND DISCUSSION Territory Occupancy Territory occupancy increased from 39 sites in 1989 to 44 in 1990 (Table 1). The expansion was due to reoccupancy of one historical site (site 12) and discovery of 7 previously unreported sites (sites 58 through 64). All the recently discovered sites were situated on cliffs that had been surveyed and found to be vacant in the past. West Slope nesting pairs continued to expand (Table 2) with 35 sites occupied by 33 breeding pairs. East Slope site occupancy increased to 9 sites with addition of a new pair (site 60). Reproduction Peregrine productivity in 1990 averaged 2.00 young fledged per successful pair (31 pairs fledged 62 young) and 1.55 young fledged per total pair (40 pairs fledged 62 young)(Table 3). The 35 occupied sites on the West Slope fledged 51 young (1.55 young per occupied territory) while the 9 East Slope sites produced 11 fledglings (1.22 young per occupied territory) (Table 2.). Six sites (sites 9, 16, 29, 34, 43 and 46) were manipulated to produce second clutches of eggs for release on the East Slope. In all 19 eggs were obtained and 15 young were hatched. The original female at site 46 was replaced and although the new female laid and incubated eggs, they failed to hatch. Site 43 produced a second clutch of eggs, put they also failed to hatch. The second clutch exhibited no signs of development. The other four the pairs produced 9 young (2.25 young per pair), but do to failure of sites 43 and 46, the manipulated sites averaged an overall productivity of 1.5 young. The 15 young that were hatched from the recycle effort were hacked to the wild at 3 East Slope sites. For the third year the adult male at site 39 was paired with a pralrle falcon. their nesting attempt failed. Although they frequented a particular ledge, there was no evidence that they produced eggs. Eggshell Condition Eggshell fragments representing 40 first clutch eggs originating from 17 wild nests were collected in 1990. Shell thickness (with membrane) averaged 9.9% thin (0.324 mm) which continues to be an improvement over eggshell thicknesses observed prior to 1989. Unfortunately, within clutch thicknesses continued to vary wildly. As an example, Site 9 produced eggshells that were +1.7%, -1.7%, �and -12.0% thin, respectively. Shells from two eggs collected at Site 31 were 0% and -18.4% thin. The poorest eggshells came from Site 43, which averaged 14.2, -16.7, and -17.5% thin (despite their condition, the two thinnest were hatched ;n captivity). The thinnest egg encountered was on the East Slope (site 2) and measured -19.8 % thin. Organochlorine Residue in Eggs Results of chlorinated hydrocarbon scans of 28 nonviable whole eggs encountered at wi ld nests from 1986 through 1989 have not yet been received. Eleven additional intact nonviable eggs were collected in 1990 and submitted to the Fish and Wildlife Service for analysis. Hacking Efforts Hacking efforts continued along the East Slope in order to bolster the wild population. In 1990, 15 falcons were released at 3 sites and 12 achieved independence (Table 4.) for a success rate of 80% (Table 5.). In 1990, the Fish and Wildlife Service did not apportion any of the Peregrine Fund's captive production for release in Colorado. The only way falcons could be obtained to augment East Slope production was through recycling wild breeding pairs. In all, 6 pairs (sites 9, 16,29,34, 43, and 46) were recycled. and 15 viable young were hatched from 19 eggs. All 15 falcons were returned to the wild through hacking. The hacking effort in downtown Denver was not activated in 1990. Although pair of falcons frequented the release site and although they did not breed, their territorial defense jeopardized any hacking effort. After the hack box at the Air Force Academy had been opened for a week, it was closed and a male that had been held back from the 1989 release due to impaired feather development, was placed in the box and re-released. The effort was unsuccessful and the male was picked up several days later suffering more feather damage and poor condition. He will be permanently retained in captivity. Return of Released Falcons In the course of 1990 field investigations, observers attempted to determine the presence of bands or color markers on territory occupying falcons. In all, 44 of 87 falcons were checked for the presence of bands or markers and 9 (20%) were marked. Prepared By: _ Gerald R. Craig Wildlife Researcher C �Table 1. OCCUPANCY OF PEREGRINE BREEDING TERRITORIES IN COLORADO SITE PRE NO. 1964 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1 2 3 4 5 6 7 8 P P P P P A A A M P P P ? P P P P P P P P M P V P P V V M V P P P P P P V V1 V P P P V V A V M P P P P P P P P P P P p V 9 + 11 V P P A 12 M V 13 + A 14 + V F P V 15 + 16 .t 17 t V 18 + V 19 + 20 + V A 22 23 + V --------------------------------------------------------(Historical 24 25 26 27 28 29 30 31 32 33 34 35 37 + 38 39 40 41 42 43 44 45 46 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 V V P P p V M V V Ps P V V V V V V A A V V P P V V F4 V V V V V V V V V V V V V V V V V V V A V V V V V V V P P P V V V V V V V V V1 V V V V P P P1 P2 P V V V M P V V V P V V1 P V V P P V V V V V V V V V P1 V, P V1 V V V P M M F V P7 V V F V V V P1 P V P V V V V P P V P P V V V V V P P F V V V V P V P P V V V V V V V V V V M V P V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V sites above P P P P P P P P P P ? V V V V, V V V V V V V V V V V V V V V V V with fenale prairie P P P P M P V V V V V P P P P P P P P V V V V V P, P V p6 P V V P P P P P P P V V V V V V P P P V V V V V V V P V V P P P P V V P P1 P V V V V V V V V V V V V V V V V V V V V V P V V V V V V V V V V V V P P V V V V V V P V P V P V P V V Y V Y V V V V V V F V V Y P V P V P V V V P P V P P P P V P V P V V P P V M P P P, P P P P, P Pg P P P P P P P P P V P V P P P P V V P P V V P P V V V V V P P P P V V P P P1 P P P, P P P V P P P P P P V P P1 P p P P V V P P V falcon. V P V V V this line)-----------------------------------------------------V V M V M V V V V P P P P P P V3 P M V V v V V Va P P F P M P P P V V V V M P V V V V V V V P A V P P P P P, P P P P P P P M P P P M P P P P P P P P P P P P, P P P V V V V A: Lone adult. F: Lone adult female. M: Lone adult male. p: Adult pair. V: Vacant site. 1 Ad. male & imm. female. 2:1&111. male & ad. fellale. 3:Ad. female replaced by il1l1. female midway. 4:Dead ad. female 5 Ad. male replaced by imm. male midway. 6:Ad. male dead in Vicinity. 7:18111. nale & fe;ale. 8:1mll. male. 9 Ad. male paired P V V V F V V P P P1 P P P P P P P P P, P2 P V P P Pg P V V V F found in vicinity. P P P P9 P P P P P P P P M P Pg P P P P V P V P P P ~9 P V P P P P P P P P V P P F V P P P P V P P P P P P P P �Table 2. COMPARISON OF EAST AND WEST SLOPE SITE OCCUPANCY CUM West Slo~e Occupled sites Breeding Pairs Young Produced Young Hacked 73 74 75 76 77 78 79 80 81 82 83 84 11 9 7 7 10 9 11 11 11 11 13 13 1 4 4 3 8 6 4 5 6 6 8 11 o 10 5 4 11 16 12 16 15 20 21 26 0 0 0 0 0 0 5 7 10 8 15 18 East Slo~e Occupled Sites Breeding Pairs Young Produced Young Hacked 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 CUM 3 2 1 1 2 2 2 2 0 0 1 1 1 2 5 8 7 9 3 2 1 1 1 0 0 0 0 0 0 1 1 1 3 2 4 6 2 3 0 3 0 0 0 0 0 0 0 3 2 2 10 5 6 11 47 0 0 0 0 0 4 5 4 3 5 8 9 12 10 19 26 25 12 110 85 86 87 88 89 90 13 19 24 24 32 35 12 17 20 22 27 33 26 32 45 44 60 51 12 4 0 0 0 0 360 79 �Site 1 2 3 4 5 7 9 11 12 16 18 25 27 29 30 31 32 33 34 35 38 39 40 42 43 44 45 46 47 48 49 50 52 53 54 55 57 58 59 60 61 62 63 64 Table 3. Summary of 1990 Peregrine Production Age Eggs Young Young Male Female 1st 2nd Hatched Fostered Clutch Clutch A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A I A A A A A A A A A A A A * A A A A A A A A A A A A A A A A A A A A A A A I A A A A A A 1+ 2+ 4 1 3 3 2+ o a o 2 1+ 3+ 2 2+ 1 3+ Young Fledged 3 3 1+ 3+ 3 2 o a 4 3 3 3 2 3 2 2 2 2 3 2 3 3+ 2 3 3+ 3+ 2 2+ 4 o 3 4 1 2+ 2 2 3 3 3 a o 2+ 1 3 1 3 1+ 2+ 2+ 2 1+ 3 4 3 3 4 4 3 4 3 3+ 3 + 1 a o 3 3 2+ 2+ 1 3 3 3 1+ 1+ 1 1 2+ 1+ 2+ 3+ o o o a 2+ 1+ 1 2 3 2 o o o 1+ 1+ 1+ 1+ 3+ 3+ 1+ 1+ 91+ o o 14+ 73+ 1 1 3 1 o 62 Total Sites Occupied: 44 Total Adult Pairs: 40 Total Breeding Pairs: 37 Total Successful Pairs: 31 Total Young Produced: 73+ Total Young Fledged: 62 Average Fledged Brood Size:2.001 Young Fledged Per lotal Pair: 1.55 Young Fledged Per Total Unmanipulated Pair: 1.55 * Total Fledglings Divided by Total Successful Paired with female prairie falcon. Pairs. �Table 4. Site Name Air Force Academy Twin Mountain Deer Mountain TOTAL 1990 HACKING RESULTS Supporting Agency Young Released Young Achieving Independence USFS BLM NPS 5 5 5 4 15 3 5 12 �Table 5. Site Sheep Mtn. Hermosa Conejos II Perins Peak Chimney Rock Moraine Park Deer Lodre Conejos Wolf Creek Beaver Creek Royal Gorge Big Hole Adobe Peak Deer Mtn. Twin Mtn. Natural Arch Trout Creek Denver Academy TOTAL SUCCESS 1978 1979 1980 1981 5/4 5/5 5/5 -- 4/4 --- ------------------5/4 80% --- --------------- -- 4/3 4/4 ---------------- 4/4 3/3 3/3 3/3 -------------- COLORADO HACK SITE SUCCESSES 1918-90 1982 1983 1984 1985 1986 1987 1988 1989 1990 5/5 -- 4/4 --- 5/5 --5/5 -- 5/5 -- 3/3 ------- ------------ --------- -------------- --- -5/3 4/4 4/0 -- 4/0 5/4 4/4 4/4 4/4 4/4 ------------- -- ---------- 4/4 4/3 -- 5/2 5/5 5/4 4/4 ---------- 5/2 -- 5/4 5/3 515 5/0 ---- ------ -3/1 ---- -3/3 ---5/5 4/3 4/1 4/3 4/4 3/3 -- ---- ---- -- SIS 4/2 ----- 5/4 5/0 5/4 4/4 4/4 SIS 5/5 -- 10/10 12/11 13/13 22/13 25/23 32/27 35/24 20~14 22/19 33~26 100% 92% 100% 59% 92% 84% 69% 0% 86% 9% -- 4/2 ---- 4/3 5/5 4/4 515 5/3 3/3 5/5 5/3 ---- 5/4 31/25 15/~ 81% 8 '\ TOTAL SUCCESS 36/35 8/7 28/24 11/6 8/5 13/6 22/20 14/11 25/23 5/0 13/10 13/5 19/17 18/14 8/8 10/10 10/8 8/7 97% 100% 87% 86% 54% 62% 46% 91% 78% 92% 0% 77% 38% 89% 78% 100% 100% 80% 88% 274/221 81% SIS ��13 JOB PROGRESS REPORT State of Project: ~C~o~l~o~r=a=do=_ _ ~(=W-_1~5~1~-~R_-~3)~ _ Bald Eagle Nest Site Protection and Enhancement Program Period Covered: 1 July, 1989 - 30 June, 1990 Personnel: G.R. Craig, Colorado Division of Wildlife and R.L. Knight, Colorado State University. ABSTRACT Although 10 Colorado territories were occupied by breeding bald eagles once again in 1990, one of the 1989 territories was abandoned and a new pair was discovered in La Plata County. Productivity was 1.3 young per nesting attempt. Attrition to windthrow continued with loss of one nest in Rio Blanco County. Due to scheduling difficulties and hazardous placement of several nests, only 2 sites were visited during the nestling phase and only 4 young were banded and color marked. This Job Progress Report represents a preliminary analysis and is subject to change. For this reason, information presented herein MAY NOT BE PUBLISHED OR QUOTED without permission of the author. ��15 BALD EAGLE NEST SITE PROTECTION AND ENHANCEMENT PROGRAM Gerald R. Craig SEGMENT OBJECTIVES 1. Monitor nest site occupancy and reproductive success. 2. Document survival rates and mortality factors. 3. Determine migration and wintering areas. 4. Determine if philopatry occurs in breeding eagles. 5. Determine nest site tenacity by individual breeding eagles. 6. Quantify nesting habitats and associated foraging areas in an effort to document nest site parameters conducive to improved reproduction. 7. Document pesticide contamination through eggshell measurement and chemical analysis of nonviable eggs. 8. Where necessary, occupancy. implement actions to stabilize nests and maintain METHODS AND MATERIALS This work will be a cooperative endeavor between the Division and Dr. Richard Knight of Colorado State University. 1. Annually visit all documented breeding sites to determine the presence of bald eagles. Pairs at territories will be documented by DWMs and other field personnel. Previously unrecorded pairs will probably be revealed in the course of aerial eagle and waterfowl flights. DWMs will confirm actual incubation from ground visits. 2. Occupied territories will be visited by DWMs periodically throughout thebreeding season to determine hatch of young, nesting failures, etc. 3. In May and June, a Util ity Worker wi 11 observe breeding eagles from a distance and endeavor to follow thei r movements to locate important foraging areas. Responses of eagles to various human activities and land uses will be recorded. 4. In June, when the young are determined to be old enough to band, sites will be visited by Craig and Knight to place a federal band on one leg and a colored, alpha numeric marker on the other. The color markers will permit identification if the young return in subsequent years. During the same nest visit the following will be recorded: Physical parameters such as tree species, height, DBH, condition, and dominance. Nest condition, size, and location. Vegetative community and land use practices. �16 In addition, collect prey remains, nonviable eggs and eggshell fragments. 5. When necessary, remedial actions will be taken to stabilize nests that are threatened by wind throw. Should the tree be decadent and in danger of falling, an artificial nest base may be placed in a suitable, adjacent tree. Action will be taken only after it has been deemed desirable to encourage the eagles to nest at the same location. RESULTS AND DISCUSSION Territory Occupancy Bald eagle nesting activities for Colorado are summarized in Table 1. In 1990, 10 territories (Adams, La Plata #1 and #3, Moffat #1, and #2, Montezuma #2 and #3, Rio Blanco #1 and #3, and Weld #1) were occupied. The pair that first occupied Weld #2 in 1989 did not return in 1990. However, Weld #1 was occupied which suggests that the pair may alternate between the two sites. After several years of search, a new pair was discovered at La Plata #3. As far back as 1978, adult eagles were intermittently observed in the region throughout the summer, but nesting was not confirmed. The nest was located in an old growth ponderosa pine secluded in the bottom of a forested canyon. Land Status The territory of La Plata #3 is on U.S. Forest Service land. As reported previously, (Craig 1988) Weld #1 occupies private property. The landowner is aware of the site and is protect ive of the pair. Land use in the area is livestock ranching. Reproduction Reproductive efforts of the 10 pairs are summarized in Table 2. In all, 15 young were hatched by 10 pairs (1.5 young per pair), 13 young were fledged by 8 pairs (1.62 young per successful pair) which yielded an overall productivity of 1.3 young per nesting attempt. Both young that were lost died at 5 to 6 weeks of age of undetermined causes. Unlike the previous year, no broods of 3 young were observed. Because of scheduling difficulties, 5 sites (La Plata #1 and #3, Montezuma #2 and #3 and Weld #1) were not visited during the nestling phase. Nests at Moffat #1, Rio Blanco #1 and #3 were too precarious to cl imb without endangering the cl imber and the nest, so no data were obtained. As a result, two sites (Adams and Moffat #2) were visited to band nest 1ings and co11ect eggshe 11 fragments. On 1y 4 nestlings were marked with an unanodized aluminum U.S. Fish and Wildlife Service band on the left leg. A red anodized band with vinyl yellow tag affixed was attached to the right leg. Blood samples, culmen length and foot pad length measurements were obtained from the 4 eaglets. �17 Eggshell Condition Due to the minimal number of nest visits, no nonviable eggs or eggshell fragments were encountered in 1990. Nest Stabilization Efforts In the fall of 1989, the nest at Rio Blanco #3 was reported blown down. Upon inspection, it was determined that the entire main trunk had broken. An artificial nest platform was constructed one of the remaining forks of the same tree. The original nest materials were hoisted and placed in the new structure. Weight of the old nest was 280 lbs. Several groups of eggshell fragments were encountered in the old nest cup and were retained for thickness measurement. Artificial nest platforms at Adams, and Montezuma #3 continued to be used. The pair at Moffat #1 again used the extremely vulnerable nest they selected in 1989. If at all possible, an artificial nest should be constructed in an adjacent tree and the present nest should be removed. This nest is in a dead tree top and will be lost to windthrow in the near future. Should windthrow occur during the spring or summer, that pair's reproduction will be lost. Inspection broken and the Fall, additional of the platform at the Adams site revealed that a support wire had the additional mass of the nest required remedial stabilization. In the site will be visited to install plastiC coated cable and add supports. LITERATURE CITED Craig, G.R. and R.L. Knight 1989. Bald Eagle nest site protection and enhancement program. Job Prog. Rep., Colo. Div. Wildl. Res. Rep., Jan., pp1-7. Craig, G.R. 1988. Bald Eagle nest site protection and enhancement program. Job Prog. Rep., Colo. Div. Wildl. Res. Rep., Jan., pp1-6. Prepared by: c;~__Ic2 __ ~~~~ Gerald R. Cr i Wildlife Researc er C _ �r' co Table 1. Bald Eagle Nesting Efforts in Colorado Site 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 La Plata Co. #1 Moffat Co. #1 La Plata Co. #2 Grand Co. Montezuma Co. #1 Rio Blanco Co. #1 Rio Blanco Co. #3 Weld Co. #1 Montezuma Co. #2 Moffat Co. #2 Moffat Co. #3 Adams Co. Archuleta Co. Montezuma Co. #3 Weld Co. #2 La Plata Co. #3 1egg IA IA IA IA 1yng 2yng 2yng 2yng 2yng 2yng 2yng Oyng o 1 A ? ? ? -IA IA 2yng IA IA' A 1yng 2yng IA IA A 1yng 3yng A Total Young Total Pairs IA = Inactive IA 1yng Oyng Oyng = 1 1 Active 4 4 2 2 4 3 1 3 o 1 3 3 6 4 ? -IA IA A ? 2yng IA IA IA ? eggs 2yng 2yng 2yng 2 6 2 4 ? Oyng IA IA IA Oyng 2yng 2yng 2yng 6 5 eggs 1yng IA IA IA 2yng Oyng eggs 1yng 1yng 1egg eggs eggs 5 10 2yng 2yng A IA IA 2yng 1yng IA 1yng Oyng IA 1egg 2yng 10 9 1yng 1yng IA IA IA 2yng A IA 1yng 2yng 2yng 3yng IA IA IA 2yng 2yng IA 1yng 3yng ? eggs eggs 2yng IA IA 1yng 1yng eggs 8 8 16 10 Oyng 2yng IA IA IA 2yng 1yng Oyng 1yng 2yng IA 2yng IA 1yng IA 2yng 13 10 �19 Table 2. Colorado Bald Eagle Nesting Efforts - 1990 Age of Birds Male Female Site Young Produced Young Fledged Adams Co. Adult Adult 2 2 La Plata Co. #1 Adult Adult 0 0 La Plata Co. Adult Adult 2 2 Moffat Co. #1 Adult Adult 2 2 Moffat Co. #2 Adult Adult 2 2 Montezuma Co. #2 Adult Adult 2 1 Montezuma Co. Adult Adult 1 Rio Blanco Co. #1 Adult Adult 2 2 Rio Blanco Co. Adult Adult Adult Adult 1 0 10 10 15 13 Weld Co. ttl Total #3 #3 #3 Comments Half feathered young dead in nest. �JOB PROGRESS REPORT State of ~C~o~l~o~r=a~d~o _ Project: (W-156-R-1) Period Covered: 1 July, 1989 - 30 June, 1990 : Pawnee Grasslands Raptors Personnel: G.R. Craig, Colorado Division of Wildlife and Douglas G. Leslie, Colorado State University ABSTRACT Original 1972 field notes were obtained from R.R. Olendorff and field work was initiated on the same study area originally delineated by Olendorff. The cumulative number of nesting raptors increased from 158 in 1972 to 235 in 1990. The 33% increase primarily resulted from a doubling of the number of nesting Swainson's hawks and great horned owls. Site parameters and habitat features were recorded for each nest site. This Job Progress Report represents a preliminary analysis and is subject to change. For this reason, information presented herein MAY NOT BE PUBLISHED OR QUOTED without permission of the author. ��PAWNEE GRASSLANDS RAPTORS Gerald R. Craig and Douglas G. Leslie SEGMENT OBJECTIVES 1. Describe the summer raptor community on the Pawnee National Grassland including species richness, relative abundance, distribution, reproductive effort, and food habits. 2. Describe the habitat and nest-site structure use of this raptor community. 3. Compare objectives 1 and 2 with similar information collected for this raptor community 2 decades earlier. 4. Document the effects,if any, of recreational activity, grazing and mineral development on raptor productivity, and habitat and nest-site usage. METHODS AND MATERIALS a. Study Area The study area will be the same as used by Jameson and Bement (1969) and Olendorff (1972). This is a 1,073 km2 site of mixed (private and USFS) ownership. Geographically, the site corresponds to the western portion of the Pawnee National Grassland and includes 111/2 townships. The area is part of the shortgrass prai rie biome, however, parts of it are in agricultural crops. Areas to the west and south are in winter wheat while on the east is a large creek bottom with nearly continuous band of cottonwoods (Popu7os spp.) and other riparian plants. The north is bounded by a band of bluffs. Bluffs, large creeks, and large tracts of cultivated land do not occur inside the study area. The IBP Pawnee Site and the Central Plains Experimental Range are near the northwest corner of the site. In addition to the intensive study area, we will examine raptors nesting on f larger area that Olendorff (1972) also used. This site is 5870 km and comprises the northern portion of Weld County, Colorado. b. Methods 1. The study area will be divided into townships and searched on foot and with vehicles (i.e., truck and motorcycle) for nests and eyries. Nests and sightings will be plotted on detailed maps of the study area. Active nests will be monitored until their reproductive fate is determined. Nests will be visited to band young and determine food habits. 2. At each nest a series of nest and nest structure characteristics will be measured (Table 1). In the vicinity of each nest a series of vegetation and human variables will be measured (Table 1). This information will be compared with similar measurements collected by Olendorff and others 2 decades earlier. In addition, geographical information systems (GIS) technology will be used to document land use changes during this time and �compare this information with abundance and distribution. 3. the present and past raptor species Efforts will be made to examine whether recreational activity, livestock grazing or mineral development effect raptor reproductive success by looking for correlations between nesting success and proximity to the developments. c. Analysis 1. Species richness, relative abundance, and food habit information will be presented qua 1itat ive 1y. This informat ion wi 11 be contrasted with s imilar data collected 2 decades earlier. Distribution patterns will be analyzed using spatial statistics. Reproductive success will be analyzed using a modified Mayfield estimator (Steenhoff and Kochert 1982). 2. A GIS software package (program PMAP) will be used to compare raptor distribution and land use changes over the 2 decade period. 3. Stepwise multiple regression analysis and logistic regression analysis will be used to examine whether recreational activity, grazing or mineral development effect raptor reproductive success. For stepwise multiple regression analysis, the dependent variable will be the number of young fledged and the independent variables will be those listed in Table 1. For the logistic regression analysis, the dependent variable will be success or failure of the nest (regardless of the number of young fledged) and the independent variables will be those listed in Table 1. RESULTS AND DISCUSSION Olendorff's original field notes for 1972 were obtained and duplicated. The study area was resurveyed and all nesting raptors were inventoried (Table 1.). The number of nesting Swainson's hawks (Buteo swainsonii) and great horned owls (Bubo vi rginianus) nearly doubled and red-tailed hawk (Buteo jamiacensis) breeding pair tripled. Ferruginous hawk(Buteo regal is) pairs fell by half. The actual nesting status of prairie falcons (Falco mexicanus) and golden eagles (Aquila chrysaetos) remained uncertain since a significant population occupied the Eagle Rock Ranch and access was denied in 1990. Habitat data and productivity for all nesting raptors was recorded. A suitable database structure was formulated and data entered. Digitizing of nest locations, roads, houses and other geographic data into the ARC/INFO geographic information system will be accomplished in the next segment. LITERATURE CITED Jameson, D.A. and R.E. Bement. 1969. General description of the Pawnee site. USIBP Tech. Rept. 1. Colo. State Univ., Fort Collins, CO 24pp. Olendorff, R.R. 1973. The ecology of the nesting birds of prey of northeastern Colorado. U.S. IBP Grassland Biome Tech. Rep. No. 211. 233p. �Steenhoff, K. and M.N. Kochert. 1982. An evaluation of methods used to estimate raptor nesting success. J. Wildl. Manage. 46:885-893. Prepared By: _ Gerald R. Craig Wildlife Researcher C �Table 1. Variables to be measured in the vicinity of raptor nests in the Pawnee National Grasslands Nest-Site And Structure Variables Nest structure type (if tree, species; if cliff, geology) Nest height Nest structure height Nest elevation Nest relief Nest aspect Nest structure aspect Nest cliff length (for cliffs) Nest cliff area (for cliffs) Nest cliff verticality (for cliffs) Diameter breast height (for trees) Diameter nest height (for trees) Nest opening width (for cliffs) Human Disturbance Variables Linear distance from nest to nearest human dwelling, to nearest community Pavement distance from nest to nearest human dwelling, to nearest community Pavement time from nearest human dwelling, to nearest community Foot distance from nearest road to nest Number of roads within 5 km radius from nest Linear distance from nest to nearest paved road Linear distance from nest to nearest road of any nature Land use within a 5 km radius of the nest. �Table 2. Comparison of occupancy and productivity of six species of raptors inhabiting the Pawnee National Grasslands in 1972 and 1990. Species No. of Pairs 1972 1990 % Successful Pairs 1972 1990 Young Fledged 1972 1990 Swainson's Hawk 68 151 56.8 49.7 80 152 Ferruginous Hawk 26 16 61.6 43.8 48 15 Red-tailed Hawk 8 24 62.5 8 26 Great Horned Owl 30.5 64 29.1 60.9 18 75 Golden Eagle 12 6 41.6 50.0 6 4 Prairie Falcon 14 6 78.6 83.3 41 15 ? � https://cpw.cvlcollections.org/files/original/16b2f9b4ac09de45d0f33b1c1dbdf2ac.pdf 57bdeaf8b26a7b98b18e72707c3fc2b5 PDF Text Text .•.1 JOB PROGRESS REPORT State of: Colorado Project: __~W_-~1~5~2_-~R~ _ Work Plan: 3__ : Job Job Title: Sage Grouse Band Recovery 13a Period Covered: 01 January Author: A. Zablan Marilet Personnel: Upland Bird Research Analysis through 31 December and Survival Estimation 1989 C. E. Braun, Colorado Division of Wildlife; Zablan, Colorado State University G. C. White, M. A. ABSTRACT Survival estimates were calculated for sage grouse (Centrocercus urophasianus) using program BROWNIE for males and program ESTIMATE for females. Survival rate estimates for males banded and released as yearlings (young) ranged from 15.0% (95% confidence interval: 0-37.5%) in 1986 to 100% (27.3-100%) in 1974. The mean survival rate estimate of males banded as yearlings was 51.7% (40.063.3%). Survival rate estimates for males banded and released as adults ranged from 0% (0-19.9%) in 1985 to 84.6% (33.8-100%) in 1979 and averaged 38.4% (32.9-43.8%). Since no differences in survival rate estimates were detected between females banded as yearlings and adults, the data were pooled. Survival rate estimates for females did not differ by year (f - 0.47). The survival rate estimate for females was 54.2% (standard error - 2.42, 95% confidence interval - 49.5-59.0%). A preliminary analysis with survival as a function of harvest regulations was conducted for females. Addition of harvest regulation information did not significantly improve the goodness-offit and the simplest model (constant survival and recovery) adequately described the data. ��.) .) SAGE GROUSE SURVIVAL ESTIMATION, NORTH PARK, COLORADO Marilet A. Zablan SEGMENT OBJECTIVES 1. Analyze sage grouse banding data. 2. Monitor sage grouse population 3. Estimate annual survival rates by age and sex class from band recovery and recapture data from 1973 through 1989. 4. Extend analyses of recovery and recapture 5. Develop models from band recovery data with survival as a function available weather variables. 6. Extend the analysis of band recovery data for upland game birds. 7. Collect sage grouse harvest data. 8. Present findings at appropriate 9. Compile and analyze data and prepare quarterly size and weather variables. management data of birds. and scientific of meetings. and annual reports. METHODS A database was created using the "dBASE II" database management software package. The following information was entered for each recapture (live) and/or recovery (dead) of a sage grouse banded in North Park, Colorado from 1973 through 1988: 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. band number, date banded, age when banded, sex, site where banded (verbatim from banding form), North Park quadrant where banded (l=NW, 2=NE, 3=SE, 4=SW), section where banded, range (west) where banded, township (north) where banded, date when recaptured or recovered, age when recaptured or recovered, site where recovered or recaptured (verbatim from recovery North Park quadrant where recaptured or recovered, section where recaptured or recovered, range (west) where recaptured or recovered, township (north) where recaptured or recovered, whether bird was recaptured or recovered, method of recovery (for dead birds only), and whether or not bird was radio-marked. file), There were 1,467 different entries including recaptures and recoveries from 1973 through 1989. The database file created in dBASE II was converted into a dBASE III Plus file. �The data base was reviewed and missing, incomplete, or incorrectly entered data were noted. I reviewed the database with C. E. Braun (Colo. Div. Wildl.) and, in addition to clarifying locations of many of the recaptures and recoveries, an age-classification scheme for assigning ages of birds when recaptured or recovered was devised (Table 1). Additional recovery data (1989) obtained through use of check stations, volunteer wing collection barrels, and mail-in returns will be entered into the database. Table 1. Age classification of sage grouse banded, recaptured, or recovered. Bird banded as Will be classified If recovered during 11- 1+ 2+ 1st fall (01 Aug-3l Oct) 2nd spring (01 Nov-3l Jul) 2+ 2+ 1st fall 2nd spring 2+ 3+ 1st fall 2nd spring etc. 1+ 1+ etc. 2+ 2+ etc. The recapture/recovery database was transferred to the "dBASE III Plus" database management software package. Within dBASE III Plus, programs were written to convert dates (date banded, date recaptured or recovered) to an acceptable format for analysis. Program SAS was not used to edit the database, as applications in dBASE III Plus were acceptable. Programs SAS (SAS Inst. 1985) and SURVIV (White 1983) were used for data summary and analysis. Ages at banding and recapture/recovery were converted to decimal values (e.g., from 1- to 0.9, from 1+ to 1.2, from 2+ to 2.2, etc.) for ease of data summary and analysis. Of the 1,467 different entries for 1973 through spring 1988 including recaptures and hunting season recoveries, 2 entries were deleted for lack of banding data. Total banding data (1973-88) were summarized. This summary consists of total number of birds banded per year by age and sex as these data are needed to operate programs such as BROWNIE and ESTIMATE. Climatological data for North Park (Walden and Spicer weather stations) were summarized. Precipitation and temperature data for 1973-89 (Jan-Jul) were included, for future inclusion in survival models. Program SAS was used to summarize the North Park sage grouse recovery information by sex and year. Numbers of grouse banded by age and sex varied (Table 2). Recovery matrices were generated (Tables 3-6) and used as input for FORTRAN program BROWNIE. The 1988 recovery data were summarized and added to the database after the analyses were conducted. �Table 2. Sage grouse banding data, North Park, Colorado, 1973-88. 3umber banded Year 1- Males 2+ 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 88 54 138 120 183 106 111 127 110 110 152 102 163 104 117 ll5 99 88 153 114 123 98 146 173 190 190 157 92 88 51 85 88 Table 3. Colorado, Recoveries 1973-87. All 1- 179 142 291 234 306 204 257 300 300 300 309 194 251 155 202 203 41 22 62 71 101 32 48 72 31 42 33 123 46 33 71 48 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 68 27 68 74 133 22 52 94 38 69 32 III 28 30 29 59 of adult male sage grouse by hunting Recoveries Year banded Females 2+ All Totals 109 49 130 145 234 54 100 166 69 101 65 234 74 63 100 107 288 191 421 379 540 258 357 466 369 401 374 428 325 218 302 310 season, North Park, bX hunting; season 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 7 4 8 1 5 10 0 1 4 16 1 0 2 3 12 0 0 0 2 3 10 0 0 1 0 2 9 14 0 0 1 0 3 3 9 9 0 0 0 0 0 0 3 5 16 0 0 0 0 0 0 3 2 5 19 0 0 0 0 0 0 0 1 2 6 15 0 0 0 0 0 0 0 0 0 2 3 8 0 0 0 0 0 0 0 1 1 1 0 5 10 0 0 0 0 0 0 0 0 0 0 0 1 1 8 0 0 0 0 0 0 0 0 0 1 0 0 0 1 10 �6 Table 4. Recoveries of yearling Park, Colorado, 1973-87. male sage grouse by hunting Recoveries Year banded 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 bX hunting 73 74 75 76 77 78 79 80 6 4 6 6 5 18 1 2 6 17 0 1 6 5 20 1 0 2 6 9 0 0 0 2 6 0 0 1 1 2 3 14 4 13 4 13 81 0 0 0 1 1 1 0 5 13 season 82 0 0 0 0 1 0 1 3 5 7 83 84 85 0 0 0 0 0 0 0 1 4 1 15 0 0 0 0 0 0 0 0 0 3 10 12 0 0 0 0 0 0 0 0 0 1 2 Recoveries 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 4 16 Table 5. Recoveries of adult female sage grouse by hunting Park, Colorado, 1973-87. Year banded season, North 73 74 75 5 1 2 0 0 6 76 1 0 2 2 77 1 1 4 2 15 bX hunting 86 0 0 0 0 0 0 0 0 0 1 0 0 4 5 87 0 0 0 0 0 0 0 0 0 0 0 0 1 2 8 season, North season 78 79 80 81 82 83 84 85 86 87 0 0 1 0 0 1 3 5 0 3 0 0 0 0 1 0 3 7 0 0 0 1 1 0 0 5 0 0 0 0 0 0 0 2 1 5 0 0 0 0 1 0 0 2 0 2 1 0 0 0 0 1 0 1 0 0 0 1 6 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 2 2 1 0 0 0 0 0 0 0 0 1 1 0 0 0 1 6 1 3 1 4 2 1 4 1 �Table 6. Recoveries of yearling Park, Colorado, 1973-87. female sage grouse by hunting Recoveries Year banded 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 bX hunting season, North season 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 2 2 1 4 2 6 1 1 2 2 2 0 1 4 7 0 0 2 2 6 5 0 0 0 0 0 0 1 1 4 8 0 0 0 1 0 0 0 4 4 0 0 0 0 1 0 0 1 3 2 0 0 0 0 1 0 1 2 0 0 0 0 0 0 0 0 0 0 2 2 6 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 7 6 0 0 2 1 3 0 0 0 1 6 0 7 0 1 0 2 4 1 8 Brownie et a1. (1985) presented band recovery models for birds banded as young and adults. The models were presented as hypotheses with certain assumptions about the relationship of the band recovery data to the actual population: Hvpothesis Ho 1) Young and adults have identical survival (fi); and 2) survival, hunting, specific. Hxpothesis reporting, (Si) and recovery and hence recovery, rates are year- Hal 1) Annual recovery and survival rates are age-dependent year of life only; and 2) annual recovery and survival rates are otherwise to year. Hxpothesis rates for the first constant from year H02 1) Annual reporting year-specific; and harvest rates (and hence recovery 2) annual survival and harvest rates are age-dependent year of life only; and 3) annual survival rates are otherwise constant rates) are for the first from year to year. �8 Hypothesis Hi 1) Annual survival, reporting, are year-specific; 2) annual survival and harvest rates are age-dependent for the first year of life only (i.e., young and adult birds have different survival and harvest rates); and 3) reporting Hypothesis and harvest rates are not dependent rates (hence recovery rates) on time of release. H2 1) Annual survival, harvest, and reporting rates are year-specific; 2) annual survival and harvest year of life only; and 3) in any year, the reporting rate for newly released birds is different from that for survivors of previous releases. rates are age-dependent for the first (H2 is an extension of Hi in that the first year adult recovery rate in year I is different from the recovery rate in year I of previously banded adults. (The soliciting of bands from hunters by conservation officers near banding sites may give rise to this situation.)] Hypothesis H3 1) Annual survival and recovery rates are age-dependent for the first 2 years of life (i.e., are different for young, subadults, and older birds). This embraces assumption 3 of model H2 for the experimental situation we are concerned with, because of the unidentifiability of the reporting rate; and 2) annual survival and recovery rates are year-specific. (H3 assumes the parameters Sand f are age-specific for 3 age classes (young, subadult, adult), but only 2 age classes are recognized during banding, because subadults and adults are usually indistinguishable.] Program BROWNIE (Brownie et al. 1985) was used to analyze the sage grouse band recovery data from North Park, Colorado for 1973-87. Survival estimates were calculated using programs BROWNIE and ESTIMATE (Brownie et al. 1985). Program SURVIV (White 1983) was used to model survival as a function of harvest regulations (Table 7). Harvest regulations were quantified as an index, the theoretical maximum bag possible for 1 hunter (i.e., season length (days) x maximum number of birds per day). For example, in 1987 the theoretical maximum bag possible was 23 days x 3 birds per day - 69 birds, while in 1973, it was 3 days x 2 birds per day = 6 birds. Appropriate cell probability functions were assigned in program SURVIV. �9 Table 7. 1973-89. Sage grouse season length and bag limits, North Park, Colorado, Years Season length (days) Bag/possession limit Maximum bag per hunter 1973-74 1975 1976 1977-80 1981 1982-84 1985-88 1989 3 9 9 16 23 30 23 30 2/4 2/4 3/6 3/6 3/6 3/6 3/6 3/6 6 18 27 48 69 90 69 90 Changes in harvest regulations were incorporated into the user-specified cell probability functions for survival rate estimation with program SURVIV (White 1983). The maximum bag possible (Table 7) was incorporated as a covariate in a logistic model: where Si was a function of the parameters possible in a given season. ~o and ~l' and Hi, the maximum bag The following models were tested for females with program SURVIV: Model 3, constant annual survival and band recovery rates; Model 2, constant annual survival rate and year-specific band-recovery rates; and Model H, similar to Model 2, but changes in harvest regulations were incorporated. Model 3 was the simplest model tested with only 2 parameters estimated, while Model H was the most general model tested. RESULTS The null hypothesis that adult males and females had the same recovery and survival rates was rejected (f - 0.013) with program BROWNIE (Brownie et al 1985) (Table 8). Closer examination of the data (Table 8) revealed that, although a difference was found hetween male and female parameters, this difference was not constant over years. Survival rates for males differed by year (f - 0.007) and statistical tests of different models indicated that Model Hl (Brownie et al. 1985) fit the data for males (Table 9). The assumptions of Model Hl are: �10 Table 8. Contingency chi-square grouse, North Park, Colorado. Matrix Year 1 2 3 4 5 6 Contingency table 8 9 10 11 12 13 14 15 due to sex of sage Matrix 1 X2 statistic with 1 d.f. 2 x2 Contingency table statistic with 1 d. f. 13 8 86 60 0.068 14 3 74 24 0.378 18 14 135 54 2.960 6 21 93 65 l.311 21 5 10 9 15 8.887 20 27 103 106 0.696 18 23 12 19 0.196 22 76 21 15 3.717 6 19 14 1.056 117 45 l.056 26 12 22 7 9 0.052 18 16 155 78 2.400 25 11 15 14 2.131 24 7 166 31 0.903 24 11 15 10 0.471 29 10 161 59 29 0.023 15 12 3.784 18 3 139 29 0.118 9 9 4.713 14 12 99 0.875 9 10 12 0.820 11 3 77 25 0.064 18 8 3 7 4.573 9 2 42 28 10 l. 941 2 3 0.788 10 75 23 l.428 1 7 test for differences 29 Total chi-square = 6 5 3 0.151 8 3 0.189 10 11 2.833 12 3 16 8 24 78 6 7 6 13 48.58 with 29 d.f., f(X2 6 > 48.58) ~ 0.013 �11 Table 9. Statistical tests of models grouse, North Park, Colorado. Chi-sguare Model 75.49 75.85 34.34 19.46 Ho vs Hl Hl vs Hz Hz vs H3 Case HOl vs Hoz Hoz vs Hl program BROWNIE, of fit tests d.f. Z calculated X sage XZ calculated) 0.073 0.011 0.268 0.303 59 50 30 17 not done tests between d.f. 41.16 14.88 15.45 Likelihood ratio XZ calculated 27.57 46.78 male f(xZ > not done Chi-sguare Case goodness Z X calculated Ho Hoz Hl Hz H3 under models P(Xz 0.067 0.315 0.280 29 13 13 tests between d.f. 28 26 > XZ calculated) models P(Xz > XZ calculated) 0.487 0.007 �12 1) annual survival, reporting, are year-specific, and harvest 2) annual survival and harvest rates are age-dependent for the first year of life only (i.e., young and adult birds have different survival and harvest rates), and 3) reporting rates are not dependent For females, band recovery Model Ho2 are: rates (hence recovery on time of release. data fit Model Ho2 (Table 10). 1) annual reporting specific, and harvest 2) annual survival and harvest year of life only, and 3) annual survival rates) The assumptions of rates (hence recovery rates) are year- rates are age-dependent for the first rates are otherwise constant from year-to-year. Survival estimates were calculated using programs BROWNIE (males) and ESTIMATE (females). Survival rates for males (Table 11) banded and released as yearlings (young) ranged from 15.0% (95% confidence interval, 0-37.5%) in 1986 to 100% (27.3-100%) in 1974. Mean survival of males banded as yearlings was 51.7% (40.0-63.3%). Survival rates for males banded and released as adults (Table 12) ranged from 0% (0-19.9%) in 1985 to 84.6% (33.8-100%) in 1979 and averaged 38.4% (32.9-43.8%). Since no differences in survival estimates were detected between females banded as yearlings and adults (Table 10), the data were pooled. Survival rates for females did not differ by year (f - 0.47). The survival rate for females was 54.2% (standard error - 2.42, 95% confidence interval ~ 49.5-59.0%). A preliminary analysis incorporating year-specific harvest regulations into survival models was conducted for females. Testing of 3 models under program SURVIV showed that the simplest model (3) was adequate to describe the data, and that the parameters added under Models 2 and H did not significantly improve the goodness-of-fit (Table 13). Both the relatively high goodness-offit under Model 3 (0.918) and the relatively low value of the Akaike Information Criterion (-2 x log-likelihood + 2 x number of parameters estimated - 280.041) indicate that Model 3 was the most appropriate model tested. While the results obtained with program SURVIV may have been expected to occur with the female data (Model H02 was earlier found to be appropriate), the results of model testing with the unpooled male data may be quite different and will be investigated. �13 Table 10. Statistical tests of models under program grouse, North Park, Colorado. Model Chi-sguare 2 calculated X Ha Ha2 Hl H2 H3 53.63 35.61 23.91 17.35 not done Case Ha vs Hl Hl vs H2 H2 vs H3 Case Hal vs H02 H02 vs Hl Chi-sguare X2 calculated 29.72 6.56 16.89 goodness of fit tests d.f. tests between d.f. 0.602 0.581 0.686 0.289 not done models f(X2 > X2 calculated) 0.428 0.923 0.204 29 13 13 28 26 female sage f(X2 > X2 calculated) 57 38 28 15 Likelihood ratio tests between X2 calculated d. f. 30.94 25.93 BROWNIE, models f(X2 > X2 calculated) 0.320 0.467 �1 ! ~'+ Table 11. yearlings, l. 000) . Survival estimates (program BROWNIE) of male sage grouse banded as North Park, Colorado, 1973-87 (upper confidence limit bounded at Year Estimate 1973 0.890 0.322 0.259 - 1974 1.201 0.474 0.273 - 1.000 1975 0.568 0.178 0.219 - 0.917 1976 0.738 0.234 0.280 - 1.000 1977 0.447 0.128 0.195 - 0.699 1978 0.370 0.140 0.096 - 0.644 1979 0.413 0.202 0.016 - 0.809 1980 0.541 0.202 0.145 - 1981 0.521 0.189 0.151 - 0.891 1982 0.454 0.206 0.049 - 0.858 1983 0.490 0.181 0.134 - 0.845 1984 0.291 0.165 -0.032 - 0.613 1985 0.160 0.085 -0.008 - 0.327 1986 0.150 0.114 -0.074 - 0.375 Average 0.517 0.059 S.E. 95% C.l. 0.400 - 1.000 0.938 0.633 �Table 12. Survival estimates (program BROWNIE) of male sage grouse banded adults, North Park, Colorado, 1973-87 (upper confidence limits bounded at 1.000) . S.E. as 95% C.r. Year Estimate 1973 0.360 0.167 0.032 - 0.688 1974 0.645 0.242 0.171 - 1.000 1975 0.220 0.079 0.064 - 0.375 1976 0.448 0.148 0.158 - 0.737 1977 0.325 0.102 0.126 - 0.524 1978 0.558 0.155 0.254 - 0.863 1979 0.846 0.259 0.338 - 1.000 1980 0.284 0.098 0.093 - 0.475 1981 0.319 0.098 0.127 - 1982 0.430 0.144 0.149 - 0.712 1983 0.226 0.090 0.050 - 0.403 1984 0.395 0.171 0.061 - 0.729 1985 0.090 0.056 -0.020 - 0.199 1986 0.230 0.156 -0.075 - 0.535 Average 0.384 0.028 0.329 - 0.438 0.511 �16 Table 13. Statistical tests of models grouse, North Park, Colorado, 1973-87. Model Log-likelihood d.f. (program SURVIV) of female sage Number of parameters estimated Akaike Information Criterion Goodnessoffit 3 -138.02l 118 2 280.041 0.918 2 -127.832 104 16 287.663 0.977 H -127.272 103 17 288.543 0.978 f(X2 General submodel 2 Reduced submodel calculated d. f. 2 3 20.378 14 0.119 H 3 21. 498 15 0.122 H 2 1.120 1 0.290 X LITERATURE > X2 calculated) CITED Brownie, C., D. R. Anderson, K. P. Burnham, and D.S. Robson. 1985. Statistical inference from band recovery data - a handbook. Second ed. U. S. Dep. Inter., Fish and Wildl. Servo Resour. Publ. 156. 305pp. SAS Institute, Inc. 1985. SAS Procedures 6 ed. SAS Inst., Cary, N.C. 373pp. Guide for Personal Computers, Ver. White, G. C. 1983. Numerical estimation of survival rates from band-recovery and biotelemetry data. J. Wildl. Manage. 47:716-728. Prepared by: Approved by: ~~ Mari1et A. Z an Graduate Research Assistant Clait E. Braun Wildlife Research Leader �17 JOB PROGRESS REPORT State of: Colorado Project: 'W-152-R 'Work Plan: Job Title: _3_: Responses Period Covered: Author: Job Upland Bird Research __lTh_ of Sage Grouse to Vegetation 01 January through 31 December Fertilization 1989 Orrin B. Myers Personnel: C. E. Braun, S. Porter, K. Snyder, Colorado Division O. B. Myers, G. C. 'White, Colorado State University; Upham, U.S. Bureau of Land Management of Wildlife; C. Cesar, L. ABSTRACT The response of sagebrush (Artemisia tridentata) and sage grouse (Centrocercus urophasianus) to nitrogen fertilization was studied in North Park, Jackson County, Colorado. Sagebrush plants responded to application of 112 kg-N/ha in Fall 1985 with increased growth and increased levels of foliar crude protein. Foliar crude protein levels remained significantly elevated over controls in samples collected in Spring 1989. A second application of fertilizer in 1987 produced a less dramatic response by sagebrush. Sage grouse used fertilized study plots for feeding significantly more often than adjacent unfertilized plots, however the magnitude of the effect declined over time. When presented with fertilized and unfertilized ~. ~. wyomingensis (AT'W) and ~. ~. vaseyana (ATV) in paired choice experiments, captive-reared sage grouse consumed significantly more of the AT'W treatments than of the ATV treatments, but fertilization did not increase the palatability of ATV. Wild-caught sage grouse performed poorly on all treatments used in digestibility trials, but birds were better able to maintain body mass on AT'W treatments than on ATV tre.atments.· Imperfect knowledge. of grouse me t abo Ld c .rates, thermoregulatory regime, and: forage intake and digestibility confounded attempt.s to assign an importance to detoxification cos t s associated with intake of mono t e'rpenes . Clutch sizes were not related to distance to nearest fertilized plot. Nitrogen fertilization can be used to locally increase grouse use of particular sites for feeding. Mechanisms involved in grouse selecting such sites cannot be elaborated due to short-comings in sampling strategies of this and all other sagebrush-herbivore studies. ��RESPONSES OF SAGE GROUSE TO VEGETATION FERTILIZATION Orrin B. Myers P. N. OBJECTIVES The project is part of a two-phased study to on sage grouse winter distribution in mining ecology, and to 2) evaluate whether nitrogen grouse winter habitat, can be used as a tool available to sage grouse. The focus of this Phase 2. 1) collect baseline information areas and on grouse feeding fertilizer, when applied to sage to mitigate reduction in habitat portion of the project is on SEGMENT OBJECTIVES 1. Document chemical fertilizer, and growth response of sagebrush 2. Evaluate feeding preferences of sage grouse for fertilized unfertilized sagebrush subspecies, 3. Estimate digestibility subspecies, and 4. Monitor reproductive parameters of radio-marked sage grouse to learn if sage grouse fitness is affected by fertilizer treatment. of fertilized to nitrogen and unfertilized and sagebrush DESCRIPTION OF STUDY AREA The study area is in North Park, Jackson County, Colorado. The Park is an intermountain basin at an elevation of about 2,500 m. It is drained to the northwest by the North Platte River, which is fed by many smaller streams. Topography is flat to rolling with numerous ridges and benches. Climate is cold and dry with an average annual frost-free period of 46 days. Sagebrushdominated grasslands cover upland sites in the Park, and grasses and sedges occur in native and irrigated meadows that border drainages. Artemisia tridentata is the dominant sagebrush species and includes 2 subspecies, ~. I. wvomin~ensis (ATW) and~. I. vaseyana (ATV). Other species of sagebrush also occurring in North Park are ~. lon~iloba, ~. cana, and~. ar~illosa. METHODS In October-November 1985 ammonium nitrate fertilizer (33.51 N) was applied at a rate of 112 kg-N/ha to 330 ha of sagebrush rangeland. Thirty-three 20-ha blocks were randomly-selected with 11 blocks distributed in each Area. The northern and southern 10 ha of each block were randomly assigned as fertilized or control and treated accordingly. In October 1987 fertilizer was applied to an additional 6 l-ha plots. The 3 main study areas contain similar densities of sagebrush plants, but the proportions of ATV and ATW differ between areas. �Sagebrush foliage was collected each quarter from a stratified, random sample of the 33 experimental blocks. Four blocks from each of the main study areas were sampled, and 5 randomly-located plants of the 2 big sagebrush subspecies were clipped from the control and fertilized halves of each study block. Sage grouse use of study blocks was estimated by examining individual plants for evidence of feeding activity. The characteristic appearance of leaves that have been fed upon by grouse has been described (Remington and Braun 1985). Equal numbers of each subspecies were examined along pace transects through the control and fertilized portions of randomly selected study blocks. Sampling points were at random distances along transects. Sage grouse were trapped at nocturnal roosts to attach radio packages and to provide birds for captive experiments. Radio-marked birds were located periodically to assess any use of experimental blocks and to monitor reproductive success. Male and female grouse were counted on leks between about 0500 and 0730 from 01 April through the first 2 weeks of May. Ground searches were made for new or relocated leks. Feeding trials were conducted with captive sage grouse to estimate feeding preferences for and digestibility of Wyoming big sagebrush, Artemisia tridentata wyomin~ensis, (ATW) and mountain big sagebrush, a. ~. vaseyana, (ATV). Two levels of fertilization also were tested: no fertilization (ATW and ATV) and 112 kg-N(ha (FATW and FATV). The 6 possible 2-way combinations of the 4 sagebrush treatments were presented to birds in random order to evaluate feeding preference for the different treatments. Sagebrush foliage was removed from stems and given to birds in 400 ml beakers. For each trial similar amounts of 2 treatments were placed on the left or right side of each cage prior to the morning feeding period. Positions of each treatment in the cages were determined randomly and then switched in subsequent trials. Prior to the start of trials, birds were given all treatments to avoid predisposing birds to preferring anyone treatment. Birds were allowed to feed for 60 min after which the sagebrush and spillage was removed and weighed. Digestibility trials were conducted by presenting to birds a single treatment at a time and collecting all excreta. Wild-caught grouse were used to estimate digestibility coefficients. Treatments were presented as whole sterns because wild birds could not be conditioned to feed upon picked leaves. The experimental design was 2-4x4 latin squares with assignments randomized by row and column separately for each latin square. Sagebrush stems and foliage were added to cages at about sunrise. Sagebrush also was placed on top of cages in the morning and weighed at night to provide an estimate of water loss during the day. Minimum and maximum ambient temperatures were recorded each day. After the evening feeding period sagebrush was removed from the cages and weighed. Spilled sagebrush was collected, weighed, and discarded. Diurnal fecal droppings were collected and stored overnight in resealable plastic bags. The following morning fecal droppings were collected and added to those already collected. Cecal droppings were placed into separate plastic bags. All droppings were frozen for storage. Sagebrush sterns and samples. treatment foliage samples were prepared for analyses by separating leaves from pooling equal amounts of leaves from each plant into composite Composite samples for each subspecies collected on the control and halves of each study block were ground in liquid nitrogen in a �mortar and pestle. Composite foliage samples collected during feeding trlalS were treated in the same manner. Grouse fecal and cecal droppings collec:ed in the course of feeding trials were freeze-dried and then ground in a mor:ar and pestle. Dry matter content of foliage and grouse droppings was determined after drying overnight at 100 C. Samples were ashed overnight at 500 C to determine organic matter and ash content. Kjeldahl nitrogen and gross energy (Hor~itz 1980) content of samples was measured. Soil samples were collected during early June from study plots fertilized in Fall 1987. Soil chemistry analyses were performed at the Colorado State University Soil Testing Laboratory, Fort Collins. Data were analyzed as randomized block experiments with year and fertilizer treatments handled in a factorial arrangement. This arrangement ignores the correlation of measurements made on the same experimental units in different years. I developed a model to evaluate the relative importance of factors influencing sage grouse energetics. Important tasks in developing the model were developing estimates of sage grouse food intake, partitioning of intake energy, animal maintenance requirements, and reproductive requirements. Dry matter intake (DMI) of sage grouse was estimated from captive feeding trials (Myers unpubl. data) and allometric equations (Moss and Hanssen 1980). A base dry matter intake of 100g/day was used. Sagebrush energy content, fecal energy values (PFE) , methane production (0.5% of DE [Gasaway 1976]) and excretion of fecal metabolic energy plus urinary endogenous energy were estimated so that intake energy could be partitioned on a true metabolizable energy basis (Sibbald 1976, Robbins 1983). The specific dynamic action was assumed to be 12% of intake energy (Ricklefs 1974). Energy costs for maintenance-type activities was calculated as the sum of energy for basal metabolism (BMR) , energy for thermoregulation, and energy lost through detoxification of monoterpenes. Estimates of BMR were produced using allometric prediction equations of Zar (1968) and Andreev (1988) for BMR and standard metabolic rate (SMR). Three equations were used for prediction: an equation for non-passerine birds [ZAR 1: Y(kJ/day) - 78.5 x 4.184 x (maSSkg)O.7nJ' an equation for galliform birds [ZAR_2: Y(kJ/day) - 72.6 x 4.184 x (maSSkg)o.98], and a second equation specific for galliform birds [ANDREEV: Y(kJ/day) - 1.35 x (mass~)O.855]. Predictions were compared with SMR estimates for male sage grouse dur~ng breeding (Fig. 1). Equations from Zar (1968) agreed well with estimates produced by Vehrencamp et al. (1989), with the equation for non-passerine birds being the best predictor of male SMR. However, measurements were performed on birds captured from the wild that were held for a short period of time, so estimates of basal metabolism likely are biased high. Female BMR was calculated from ZAR_l using a mean body mass for adult females in spring (1745 g) (Beck and Braun 1980). Thermoregulatory costs were calculated assuming that female grouse have the same lower critical temperature and thermal conductance as male sage grouse. These assumptions ignore sage grouse sexual size dimorphism (Beck and Braun 1980), which should affect these parameters. During April-May, adult males weigh an average of 3190 g, yearling males weigh 2890 g, adult females weigh 1745 g, and yearling females weigh 1550 g. Lower critical temperature was set at 7 C and thermoregulatory costs below 7 C were estimated by Y - 141.3-20.3 (degrees C) (Fig. 2, adapted from Vehrencamp et al. 1989). �1400 I I r! Ii .ulORHV 1200 r- , ZAR 1 ! i i i ZAR 2 ; I :-! 1000 >- < ~ I VEHRElICAl!? , 0 I I I f- ! <, ...., -'"' 0: x:: en 3CC r --- I --- ~ I I 500 r --- _ BODY Fig. 1. Estimates predictive equations al. 1989). ........ __ ... .i.>: ...... 2500 1500 1000 . --~! 3000 3500 :-lASS (3") of grouse standard metabolic rates and doubly-labelled water (Vehrencamp by et 1300 1200 ~ t rico ..:~ >- 1:000 ...,..•.. 900 .>< 800 700 SMR ~ -----------------------------------------------------------~---- 600 ~----~----~------~'------'~----~I------~----~----~ -15 -25 -20 -10 -5 5 10 15 AnBI~NT T~HPERATURE CC) Fig. 2. Energy costs of thermoregulation for male sage grouse in spring (adapted from Vehrencamp et al. 1989). In northcentral Colorado, sage grouse feed preferentially upon one subspecies of sagebrush. Wyoming big sagebrush, ~. t. wyomingensis (ATW) , plants are preferred forage and contain about twice the monoterpene concentration as do plants of the under-used subspecies, a. ~. vaseyana, CATV - mountain big sagebrush) (Remington and Braun 1985). The absolute amounts are small, however with about 1.5 to 3% of leaf dry matter for ATW and ATV, respectively. I assumed that monoterpenes are 90% digestible (Foley et al. 1987), even though they may not be that highly digestible in grouse (Remington 1983) to provide the most powerful test of whether monoterpene detoxification competes with reproductive costs. I �Monoterpenes are relatively small molecule, and probably are detoxified bv Phase I and Phase II reactions that hydroxylate and conjugate the molecul~s (Sipes and Gandolfi 1986). Ornithine traditionally has been cited as the mos: common conjugate in birds, but Remington (In prep.) did not detect ornithine conjugates in droppings from blue grouse, Dendragapus obscurus, fed several monoterpenes. Instead, he found elevated glucuronic acid conjugates in droppings. Glucuronic acid conjugation of terpenes also was detected in brushtail possum, Trichosurus vulpecula (Dash 1988). Sage grouse were assumed to excrete all digested monoterpenes as glucuronic acid conjugates. Phase I and Phase II detoxification reactions are coupled along the microsomal membrane and should result in the net loss of one ATP in addition to the energetic equivalent of one mol of glucose for each conjugate formed. Glucose catabolism by glycolysis and the citric acid pathway is about 46% efficient at recovering the combustive energy in glucose (2870 kJ/mol) at ATP, with the balance lost as heat (Martin et al. 1985:137). Therefore, 2907 kJ/mol total energy or 1357 kJ/mol or 37 ATP are lost for each -onjugate produced and excreted. Loss of carbon skeletons may not be complete, however, as gut. microbes containing beta-glucuronidase are able to remove glucuronic acid from the conjugate. Once released, glucuronic acid may be absorbed and recycled by enterohepatic recirculation (Sipes and Gandolfi 1986). Energy cost to produce a clutch was estimated from the energy value for a typical galliform egg (1.64 kcal/g fresh x 4.184 kJ/kcal) and 30% inefficiency for energy deposition (Ricklefs 1974). Average egg mass times these multipliers was summed across the number of eggs in the clutch. Number of days to complete a clutch (NDAYS) was estimated using a typical laying sequence (Patterson 1952) and an assumption that the number of rapidly growing follicles is equal to mean clutch size (8) (Hannon et al. 1979). NDAYS was the number of daily time steps that NEp was summed over for comparison with energy required for completion of the clutch. The difference between total NEp and total energy for egg formation was the net energy available for activity. RESULTS Inadequate numbers of captive hens were available to conduct feeding preference trials in a 6 x 6 latin square, so 3 hens and 1 cock were used in randomized block design. In most cases relative consumption of sagebrush was strongly (f < 0.01) affected by treatments presented (Table 1, Fig. 3). In trials where an ATW or FATW treatment was paired with FATV or ATV treatments, the ATW-FATW treatments were preferred (f < 0.01, Table 1) indicating a subspecies effect on feeding preferences. On average, relative consumption of ATW-FATW treatments was 95.5% of total intake compared to 4.5% for ATV-FATV. The effect of fertilization on feeding preference was weaker than the effect of subspecies. The ATV and FATV treatments were consumed equally and FATW was favored over ATW only if a comparison error rate of alpha - 0.06 is acceptable. Over the 4 trials where fertilized treatments were compared with control treatments (Table 1, trial numbers 2, 3, 4, and 5) FATW-FATV treatments and ATW-ATV treatments comprised 52.7% and 47.3% of measured intakes, respectively. Consumption of ATV treatments relative to ATW was low (trial 1) and did not increase when FATV was paired with ATW (trial 4). Cage position did not significantly affect consumption of sagebrush treatments (E > 0.25). 0 �Table 1. Relative consumption (X intake during trials) of sagebrush during i-hour preference trials conducted in February-March 1989. Consumption of treatments was not affected by cage position (f > 0.25). Probability values should be considered significant at f < 0.0083 (experiment alpha - 0.05/6). f > I Treatment Trial 1 ATV ATIl 7.5 92.5 0.0001 2 ATV FATV 5l. 6 48.4 0.9 3 ATV FATIl 3.6 96.4 0.0001 4 ATIl FATV 96.5 3.5 0.0001 5 ATIl FATIl 37.5 62.5 0.06 6 FATV FATIl 3.6 96.4 0.0001 P < 0.01 P > 0.2 P < 0.01 P > 0 ATW AT\/ 1 FATW P < 0.01 P < 0.01 D FATV 3. Relative consumption of sagebrush treatments by captive sage grouse. Probability values are for the Ho: that consumption was equal. Fig. �- _' Intake of sagebrush treatments did not differ for the digestion trials CE > 0.5, Table 2), and sage grouse lost an average of about 30 g/day during their period of captivity. Birds lost an average of 31.3 g/day while on ATV-FArv treatments regardless of dry matter intake. The regression coefficient for g body mass change/day on dry matter intake of ATV-FATV was not different from 0 CE > 0.8, Fig. 4). When ATW-FATW treatments were considered, the linear regression of body mass change CY) on intake CX) was Y ~ -55.3 + 0.8X CE 0.0003, R2 ~ 0.5966). The equation predicts a daily mass change - 0 at 69,1 g-DM intake/day. Variability about the regression line undoubtedly was increased by variable gut fill, fluctuations in ambient temperature, and by different activity levels of birds. Moreover, differing acclimation to captivity probably influenced losses of body mass. Table 2. treatment Intake of dry matter Cg) by captive sage grouse during digestibility trials conducted in April-May 1989. ~ SE 35.4 32.6 35.9 46.4 7.4 7.4 8.3 7.8 Treatment ATV ATW FATV FATW single 3C ·20 ...J 0" 10 U'l CI'l ATV T~:::ATl::::~TTS * 0·........................ · ATV TRL.A.TliENTS * * -10 >Q 0 -20 ;z; 0* 0* * co i I I I * *° ................... *........................................... ':.: ........ '"" r-. -30 / H ::u i -*----------, ' j 0 <C :c * '...1 -40 lJ ;z; <C -50 U -60 ° * * ° >-....J -70 H <C Q * -80 0 20 40 DRY Fig. 4. treatments. Daily change MATTER in body mass 90 60 =NTAKE of (g) captive grouse fed sagebrush �26 Seasonal sagebrush foliage samples were collected. Grouse feeding activity in study blocks for all years was summarized (Fig. 5). In 1987 transects were conducted twice--once in February and again in April-May. Measured use of ATN plants increased between February (10% of FATW plants browsed) and April-May 1987 (33% of FATW plants browsed). Feeding activity on control plots was low throughout the study (Fig. 5) and was low on ATV-FATV plants relative to ArJFATW. In each year FATW plants received more browsing pressure than ATW plants (f < 0.05), but this effect diminished. The number of ATV and FATV plants found with evidence of grouse feeding activity was too small to establish annual trends. 41] F ATTN ~ 32. 5 I - 30 (', ~ ~ 1 ATW r""'A .:» T '--1' A ~~r , . l '/ CJl :s= 0 .--V ~ ~ ')D '-' ' (-v""\ 10 o 1987 1988 1989 YEAR Fig. 5. Sage grouse feeding activity on North Park study plots. �Estimates of soil chemistry parameters were similar on control and fertilized plots in 1987 before fertilizer was applied (Table 3). Levels of NO)-N on control plots were not different (f > 0.25) from fertilizer plots in 1987, bue increased significantly on fertilizer plots following fertilization (f < 0.05). Soil nitrogen decreased (f < 0.05) between 1988 and 1989, but a treatment effect was still evident in 1989 (f < 0.05) although N03-~ had approached background levels (Table 4). Soil nitrogen levels on control plots did not change significantly with respect to year. Other soil parameters apparently were not affected by fertilization (f > 0.05, Table 4). Estimated pH, OM, P, Zn, and Cu changed significantly (P < 0.05, Tables 3 and 4) between 1988 and 1989. Organic matter, Zn, and Cu increased in 1989 compared to 1988 and pH and P declined. Differences between years in soil chemistry could be due to phenological differences between years. Sampl~~g was timed to the calendar rather that to phenology. Also, differences between years could be artifacts of sampling or laboratory procedures. Fertilizer treatment increased sagebrush foliar crude protein levels over controls but not in all sampling periods (Fig. 6, Table 5). The fertilizer effect was greatest the first spring after treatment and was much reduced in subsequent springs when grouse might accrue fitness benefits due to enhanced levels of dietary protein. The probability that treatment plots had evidence of grouse feeding was related to the foliar crude protein levels of plants collected on the plot (Fig. 7). Few ATV plants showed any evidence of grouse feeding, so a second analysis was performed on ATW plants only. The relationship between foliar crude protein and the probability approached linearity. Each foliage sample analyzed, however, represented a mean of 5 25 --~ A,TV 20 r ~ FATV A,TW FATW z 15 I- 10 I- 5 I- W f- o IT Q_ w o ~ rr u . ii! o 86/05 871 D5 86/11 86/08 871 G2 88/04 88/01 88/10 88/06 89/04 DATE Fig. 6. Foliar crude protein levels of sagebrush plants on North study plots. Plots were fertilized in Fall 1985 with 112 kg-Nfha. Park �.28 Table 3. Soil chemistry measured on 1-ha study plots in North Park, Jackson County, Colorado. Digits are not significant beyond 1 decimal place. Year Treatment 1987 Control Variable N pH Condo OM N03 -N P 6 6 6 6 6 6 6 6 6 6 6.8333333 0.1833333 3.1000000 0.5833333 4.2333333 231.1666667 0.8166667 23.8833333 2.9500000 2.1500000 0.1626175 0.0401386 0.3559026 0.0833333 0.4302454 33.0982544 0.0872417 5.5341616 0.2140872 0.1607275 6 6 6 6 6 6 6 6 6 6 6.7333333 0.1500000 2.8666667 0.6666667 3.6666667 205.1666667 0.7666667 24.0333333 2.9333333 2.1333333 0.0802773 0.0223607 0.3190263 0.1054093 0.6221825 28.4375534 0.0557773 4.2359310 0.2076322 0.1646545 6 6 6 6 6 6 6 6 6 6 7.4833333 0.2166667 1.8000000 0.7500000 2.6500000 193.8333333 0.3500000 14.9166667 2.0000000 1.8500000 0.0945751 0.0166667 0.2463060 0.2500000 0.4958158 22.0838050 0.0341565 2.5536140 0.0683130 0.1746425 6 6 6 6 6 6 6 6 6 6 7.2333333 0.2833333 1.8500000 7.0000000 2.8500000 203.1666667 0.4833333 19.6500000 2.4333333 1.7333333 0.2139574 0.0307318 0.3008322 2.0330601 0.6946222 23.6056726 0.0703167 5.4224687 0.2485514 0.1475730 K Zn Fe Mn Cu Fertilized pH Condo OM N03-N p K Zn Fe Mn Cu 1988 Control pH Condo OM N03-N p K Zn Fe Mn Cu Fertilized pH Condo OM N03-N P K Zn Fe Mn Cu Mean SE �Table 3. (continued) . Year Treatment 1989 Control Variable pH Condo OM NO)-N P K Zn Fe Mn Cu Fertilized pH Condo OM NO)-N p K Zn Fe Mn Cu N Mean SE 6 6 6 6 6 6 6 6 6 6 7.1000000 0.2166667 2.566666, 1.0000000 1.9833333 217.5000000 0.5500000 22.1166667 2.1833333 8.5500000 0.5300419 36.3737543 0.0562731 5.5205928 0.1077549 0.6222272 6 6 6 6 6 6 6 6 6 6 7.0166667 0.1833333 2.3000000 1.6666667 2.2666667 207.6666667 0.5333333 21.0000000 2.1333333 8.6833333 0.1301708 0.0307318 0.2875181 0.2108185 0.4558265 28.2284333 0.0421637 4.1108799 0.1819646 0.4150636 0.1238278 0.0600925 0.3783003 0 �30 Table 4. Analysis of soil chemistry measurements on I-ha study plots in North Park, Jackson County, Colorado. YEAR tests for changes in soil chemistry between 1988 and 1989. TREATMENT tests for changes in soil chemistry due to fertilization with 112 kg-Nfha in Fall 1987. pH Source BLOCK YEAR TREATMENT YEAR*TREATMENT DF Type III SS Mean Square 5 1 1 1 0.60833333 0.54000000 0.16666667 0.04166667 0.12166667 0.54000000 0.16666667 0.04166667 OF Type III SS Mean Square 5 1 1 1 0.07500000 0.01500000 0.00166667 0.01500000 0.01500000 0.01500000 0.00166667 0.01500000 OF Type III SS Mean Square 5 1 1 1 9.11708333 2.22041667 0.07041667 0.15041667 1.82341667 2.22041667 0.07041667 0.15041667 DF Type III SS Mean Square 5 1 1 1 32.55208333 38.76041667 71.76041667 46.76041667 6.51041667 38.76041667 71.76041667 46.76041667 OF Type III SS Mean Square 5 1 1 1 32.94375000 2.34375000 0.35041667 0.01041667 6.58875000 2.34375000 0.35041667 0.01041667 f .r. 0.91 4.05 l.25 0.31 0.4994 0.0626 0.2813 0.5846 I f > I 2.29 2.29 0.25 2.29 0.0983 0.1511 0.6214 0.1511 I f > I > I Conductivity (mmhos/cm) Source BLOCK YEAR TREATMENT YEAR*TREATMENT Organic matter (1) Source BLOCK YEAR TREATMENT YEAR*TREATMENT NOrN 12.48 15.20 0.48 l.03 0.0001 0.0014 0.4981 0.3263 (ppm) Source BLOCK YEAR TREATMENT YEAR*TREATMENT I l.03 6.14 11.37 7.41 f > I 0.4344 0.0256 0.0042 0.0157 P (ppm) Source BLOCK YEAR TREATMENT YEAR*TREATMENT I 27.70 9.85 1.47 0.04 f > I 0.0001 0.0068 0.2436 0.8371 �Table 4. (continued) . K (ppm) Source BLOCK YEAR TREATMENT YEAR*TREATMENT OF Type III SS Mean Square 5 1 1 1 80113.20833 1190.04167 0.37500 55l.04167 16022.64167 1190.04167 0.37500 55l.04167 OF Type III SS Mean Square 5 1 1 1 0.12708333 0.09375000 0.02041667 0.03375000 0.02541667 0.09375000 0.02041667 0.03375000 OF Type III SS Mean Square 5 1 1 1 2037.002083 109.653750 19.620417 5l. 333750 407.400417 109.653750 19.620417 51. 333750 OF Type III SS Mean Square 5 1 1 1 2.01875000 0.02041667 0.22041667 0.35041667 0.40375000 0.02041667 0.22041667 0.35041667 OF Type III SS 5 1 1 1 12.0020833 279.4837500 0.0004167 0.0937500 .E 16.20 l.20 0.00 0.56 f > .E 0.0001 0.2899 0.9847 0.4669 Zn (ppm) Source BLOCK YEAR TREATMENT YEAR*TREATMENT Fe .E f > .E l.86 6.87 l. 50 2.47 0.1609 0.0192 0.2400 0.1366 .E z (ppm) Source BLOCK YEAR TREATMENT YEAR*TREATMENT 13.23 3.56 0.64 1.67 > .E 0.0001 0.0787 0.4372 0.2163 Mn (ppm) Source BLOCK YEAR TREATMENT YEAR*TREATMENT .E f > .E 4.60 0.23 2.51 3.99 0.0096 0.6365 0.1338 0.0641 Mean Square .E f > .E 2.4004167 279.4837500 0.0004167 0.0937500 5.67 660.24 0.00 0.22 0.0039 0.0001 0.9754 0.6447 Cu (ppm) Source BLOCK YEAR TREATMENT YEAR*TREATMENT �.32 Table 5. sagebrush Analysis of variance summary of foliar crude protein plants fertilized in Fall 1985. levels of Effect f > .E May 1986 Block Subspecies Fertilizer S x F interaction 0.1254 0.0001 0.0001 0.0087 August Block Subspecies Fertilizer S x F interaction 0.4531 0.0001 0.0001 0.1673 Sample date 1986 November 1986 Block Subspecies Fertilizer S x F interaction 0.0071 0.0001 0.0383 0.1163 February 1987 Block Subspecies Fertilizer S x F interaction O. 0012 0.0001 0.0001 0.9385 Block Subspecies Fertilizer S x F interaction 0.8278 0.0001 0.2082 0.3021 Block Subspecies Fertilizer S x F interaction 0.1479 0.0001 0.1530 0.8106 April 1988 Block Subspecies Fertilizer S x F interaction 0.0306 0.0001 0.0015 0.8584 June 1988 Block Subspecies Fertilizer S x F interaction 0.2720 0.0011 0.4610 0.0845 October Block Subspecies Fertilizer S x F interaction 0.0033 0.0001 0.9139 0.1349 May 1987 January 1988 1988 �Table 5. Sample April (continued). Effect f > .E Block Subspecies Fertilizer S x F interaction 0.0042 0.0001 0.0400 date 1989 plants and sage grouse the pooled herbivores 0.3183 provides no information on the chemistry of individual plants that select or avoid. All studies of herbivory on sagebrush have used plant sample approach and, thus, it cannot be established whether are selecting/avoiding particular sites or plants within sites. ATV ATY Only o B CJ"I C . ..-4 -g 0 6 :u ~ QJ 'o~ z: o 0.4 o 2 o 6 8 10 foliar 12 Crude 14 Protein 16 18 20 (%Dtl) Fig. 7. Probability that plants on North Park study plots had been fed upon by sage grouse in relation to foliar crude protein levels. Lek sites and fertilized plots were mapped to describe spatial relations between plots and leks. Six leks were located in proximity to study blocks: Buteo, Hawk, Denmark, Turkey, Raven, and Perdiz. Overall, 85% of fertilized plots lie partially or wholly within 3-km radii around leks (Fig. 8). Important grouse nesting habitat generally is assumed to exist within 3 km of leks (Braun et al. 1977). Eight of the 11 plots in Area A (NW cluster of �plots) were within 3 km of a 11 Area C plots (SE cluster) (Fig. 8). Thus, female sage opportunity to choose a nest other factors being equal. lek, all Area B plots (central cluster), and 9 of were at least partially within the 3-km buffer grouse attending these leks should have had ample site in close proximity to fertilized plots, all j::<:"': ..... *::::. .:l ., 5) (, ~... , ~~ i .. ! , Jt ,I :::: w- -~- ,/ 4518~ 4S16~ /\ E } V>~ 451L1 <t ::J. u I --= h85?>~/. < .: 45 12 :.::......... (/ . / _.. \" f- a: o 4510 :::2 4508 z f- ~ ...'. '.8 .••. ·.~.:::.~ ••.• ••.••.••.••.• ~ •••..•.••••• ~ :.:: ~ ..........•......... '~":: :J - 4506 .......... ......... ............... .. ........... • 4504 -:-:.:::::-:«.: ............ PLOTS * .....,... . -:-:-:-:.»»»> ~:'::~::':'.'. /.. ':-: :<':-.':':-:-:<-:':' ••••••••. «<.:-:-:-:.»» -:-::::::::::-:::::lIt:: LEKS 4502~~~~~~--~~~~~~~~ ]9] ]97 UTtv1 EA.ST 401 405 C kmJ Fig. 8. Locations of study plots with Filled circles are 3-km radii from leks. respect to leks. I located 33 nests, of which 14 were within 1.S km of a fertilized plot (Fig. 9). During egg-laying, movements of hens to feeding/loafing areas have averaged 1.S km and then decreased to 0.1 km during incubation (Petersen 1980). Additional marked hens attempted to nest in NE North Park, but their nests were not found due to early nesting failure. I attempted to locate nests after incubation was underway to reduce incidence of observer-caused �3) abandonment. Only 1 nest by a marked hen was in a fertilized plot. Although radio-marked hens had opportunity to nest in fertilized plots (about 85% of plots were within 3 km of leks), most did not. * 45::J4 , ~ _ f"JESTS PLOTS 4520 4516 I) I f- \ nO z . [ I 4512 . .. .. . 2 f- :=) 4508 cr{_ --.- 4504 2 km 4500 390 394 UT~1 Fig. 9. Locations 398 402 406 EAST of sage grouse nests with respect to North Park study plots. There was no linear relationship between clutch size and distance to fertilized plots (f - 0.09, R2 - 0.1281). When nests within 3 km of fertilizer plots were considered, no relationship was found between clutch size and distance (f> 0.5, R2 - 0.0519). �36 The 3 years of estimated harvest prior to fertilization (NEHARV LSM£&~ - 30 birds) was lower than for the 3 years of post-fertilization data (NEHARV LSMEAN - 42 birds) (f < 0.05). However, if pre-fertilization data spans 6 years there is no difference in NEHARV estimates (f > 0.7). An examination of the residuals for the multiple regression of NEHARV on OHARV and NEHUNT (f < 0.01, R2 ~ 0.7119) presented by year reveals an apparent correlation of errors with respect to year. County-wide production indices were available for inclusion in the model, but they did not improve the fit (f > 0.05). Correlated errors could indicate fluctuations in habitat conditions and survival that influence the number of birds available for harvest in the fall. Changes in hunter access also could have influenced harvest. Additional variables will be examined to evaluate reasons for correlated residuals. Eighteen comparisons were made of the differences between control and treatment leks during the years before and after fertilization (Table 6). If the experiment error rate is controlled to alpha - 0.05 by applying the Bonferrroni inequality (Sokal and Rohlf 1986), none of the comparisons was significant (f < 0.0028 required for significance). If a comparison error rate of alpha - 0.05 is acceptable, then counts of males at Coalmont lek increased over those at Raven lek after fertilization relative to before fertilization and the difference between counts at Ridge Road lek and Tur:. lek became significantly greater after fertilization. Other treatment-control lek pairs did not show significant changes from pre-fertilization periods. Table 6. Differences between treatment before and after fertilizer application. and control lek counts for periods n Treatment lek Control lek (Before/after) f > III Buteo Fish Hatchery Thrasher 5/3 5/3 0.3006 0.2313 Denmark Boettcher Delaney Butte Fish Hatchery Migan 9/3 8/3 9/3 5/2 0.8934 0.1425 0.2161 0.1483 Perdiz Alkali Lake Aspen Deer Creek Lost Creek 8/3 8/3 8/3 7/3 0.0908 0.2960 0.2753 0.2686 Raven Arapaho Cheyenne Coalmont Delaney Butte Fish Hatchery Lost Creek Migan 9/3 8/3 9/3 8/3 9/3 9/3 6/2 0.0307 0.0473 0.0039 0.4449 0.5068 0.0588 0.0124 Turkey Ridge Road 5/3 0.0210 �,- ) I Adequate energy was available for maintenance, reproduction, and activity under the initial conditions of the model (DMI - 100 g/day, Monoterpenes 1.5% of leaf DM, and bird at thermoneutrality). Daily N~ after 7 days was adequate to form a clutch of 9 eggs if activity was held to zero. The daily energy available for activity averaged about 260 kJ, compared to a BMR requirement of 490 kJ/day. Detoxification costs were small (29 kJ/day) compared to activity costs, which can be considered to be somewhat flexible. When monoterpene concentrations were raised to the level found in the unpalatable subspecies (3.0% of leaf DM), detoxification costs increased to 58 kJ/day, or about 5% of ME intake. I reasoned that thermoregulatory costs likely were important contributors to energetic costs incurred by sage grouse. Birds may also be able to control demands for thermogenesis by selecting appropriate roost sites, e.g., under a sagebrush plant and out of the wind. I added thermoregulatory costs to my grouse energy budget by randomly drawing temperatures from a uniform distribution of temperatures over a range I observed in Jackson County during April-May 1989. Temperatures are daily medians and are assumed to approximate average daily conditions. During a typical 19-day period to form and lay 9 eggs, an average adult female grouse used 94 kJ/day for thermoregulation, which was 1.6 times the detoxification costs associated with ATV. Daily thermoregulatory costs ranged between 0 and 261 kJ/day (0-21% of ME). When I also allowed terpene levels to vary between 1 and 4.3% DM, the amount of energy left after egg production averaged 230 kJ/day and detoxification costs remained small relative to the simulated cost of thermoregulation (Fig. 10). 300 f DETOXIFICATION . 1\ 1\ I \ I \ I , j \ I \ 250 ~ I I I I • II II I ~ I' I I , , \ \ \ \ I I I 200 : I I \ I/ I I ~ II 1\ A \ \ \ \ : II I I , \ , \ ,\ : I I I I r I \, I I I I f J I \ , I I r r I I \ I I I I I \ ~ I ·'~·--·-·"---"--'·"·'···'·'-··'·"-----··········T·························T··· I I : I I I I I I : ~ I I I I I I I I \ I I I I I I I i ~ I J \ \ \ I \ I I :\ \ "~ I 'I ,\ \ I I I I '" II I I I I I I I , I I J I 150 100 THER.tfOREGULATION I \ I \ I I I I I L I I I I I I , I I I \ i 1\ 1\ I I \ I I I I I I I I I It It 1\ , \ I \ I I I I I \ I I I I I I I \ \I I :MEAN : ·····-t···········.,··············"1'·························T·· I I " I I L , , '" I I 50 I I I I I L I I I f I I I I I , I I \I .... ··,·············· \ I I I \ I , , I , I I J I ,, o 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 DAY Fig. 10. Relative energy costs due to detoxification of monoterpenes and thermoregulation during a typical period of egg production and laying (clutch size - 9). �38 Energy costs for egg production are not uniform through the laying interval (Fig. 11). During the middle of the period when eggs are deposited, total daily costs can reach 270 kJ/day, leaving little energy left for activity if thermoregulatory costs are high. Under such circumstances grouse likely exploit endogenous lipid reserves (about 3-6% of live body mass) to meet energy needs. Interestingly, by extending the period of egg formation and laying from 1 egg/day to 1 egg/l.3 day the maximum daily cost is reduced by 50 kJ, while the total cost remains the same. This cost is about equal to detoxification costs. 70 Clutch Size: 9 LaYlnq Rat. = eqq/l 60 3 day 50 20 10 o 1 2 3 4 5 6 7 B 9 10 11 12 13 14 15 16 17 18 19 20 DAY Fig. 11. Daily energy costs to female sage grouse while forming a clutch of 9 eggs when typical laying sequence involving skipped days is followed. I was interested in how my model inputs affected estimates of detoxification costs. Sensitivity analyses1 were performed on model outcomes produced by factorial sampling at 3 levels for the effects of clutch size (CLUTCH), temperature (TEMP), concentration of monoterpenes in sagebrush foliage (PTERP), female body mass (BMASS), dry matter intake rates (DMI), forage digestibility (l-PFE), digestibility of monoterpenes (PTERP), and the method used to calculate BMR requirements (METHOD) (Table 7). 1 Several small changes to the model inputs were made after I performed the sensitivity analyses. I do not believe these modifications changed the outcome of the analyses, as sources of uncertainty were rather obvious a priori. �39 Table 7. Parameters and their values used in factorial sensitivity analysis. Parameter Levels Values CLUTCH TEMP PTERP BMASS DMI PFE METHOD 3 3 3 3 3 3 3 7 8 9 0 7 -14 o 0.03 0.015 1570 1745 1920 108 120 132 0.361 0.425 0.489 ANDREEV ZAR 1 ZAR 2 Average daily energy available for activity was most strongly influenced by temperature, BMR estimation method, diet digestibility, and dry matter intake rates (Table 8). TEMP and METHOD are components of maintenance-type costs, and DMI and PFE are strong modifiers of the amount of energy available for metabolism. BMASS exerted strong effects due to its entry into the equations that estimated BMR requirements. An additional sensitivity analysis was performed to examine the effects of CLUTCH, PTERP, AND DTERP on energy for activity (Table 9). Based on relative f values, foliar mono terpene concentrations were more important than their digestibility or clutch. Table 8. Factorial sensitivity analysis of parameters in sage grouse bioenergetics model. Dependent variable is daily energy for activity averaged over laying period. Source Model CLUTCH TEMP PTERP BMASS DMI PFE METHOD DTERP DF Sum of squares Mean square f value 16 514600279.1 32162517.4 99999.99 2 2 2 189259.7 185233522.9 492459.0 8017645.3 58383244.5 118919741. 0 143241291.8 123114.7 94629.9 92616761.5 246229.5 4008822.7 29191622.3 59459870.5 71620645.9 61557.4 315.25 99999.99 820.29 13354.93 97248.51 99999.99 99999.99 205.07 300.2 2 2 2 2 2 Error 6544 1964348.6 Corrected Total 6560 516564627.7 �40 Table 9. Factorial sensitivity analysis of parameters in sage grouse bioenergetics model. Dependent variable is daily energy for activity averaged over laying period. Sum of squares Mean square I value 6 3124.525965 520.754328 30.83 2 2 2 591.300834 2026.580105 506.645026 295.650417 1013.290052 253.322513 17.50 59.99 15.00 Error 20 337.797657 16.889883 Corrected Total 26 3462.323622 OF Source Model CLUTCH PTERP DTERP Flexibility in grouse behavior may be the dominant means that grouse use to make adjustments for nutritional constraints (King and Murphy 1985). For example, male sage grouse displaying at leks show a wide range of daily energy expenditures that were correlated with number of reproductive displays (struts) (Fig. 12). The few males that were successful breeders « 51 of population) had daily energy expenditures that were double that of males that did not display (Vehrencamp et al. 1989). Not surprisingly after the effect of display behaviors was statistically controlled, male sage grouse that 2.400 2.200 2.000 >< -< 1.BOO 0 <, I--) -'< 1.600 1.400 1.200 1.000 a 200 400 600 BOO 1.000 STRUTS/DAY F1g. 12. Daily energy expenditure of male sage grouse in spring in relation to strutting intensity (from Vehrencamp et al. 1989). �,, ~l foraged the farthest from display sites during the day also had daily energy expenditures that were greater than those that traveled less. Yith such a broad expanse in the amount of energy expended toward activity it seems unlikely that 60 kJjday cannot be made available for detoxification of monoterpenes. Compounds other than monoterpenes may add to the energy drain due to detoxification, however, and other plant defense compounds could substantially influence grouse energetics by acting quantitatively to reduce forage digestibility. Sagebrush secondary chemistry also could have qualitative effects on grouse. The parameters that influenced sage grouse energy balance estimates the most were components of BMR, thermoregulation, and food intake and processing. Reliable estimates for basal metabolism and forage digestibility should not pose great problems using captive grouse, but parameters that include a behavioral component, like thermoregulation, will be more difficult to obtain. LITERATURE CITED Andreev, A. V. 1988. Ecological energetics of Palaearctic Tetraonidae in relation to chemical composition and digestibility of their winter diets. Can. J. Zool. 66:1382-1388. Beck, T. D. I., and C. E. Braun. Condor 80:241-243. 1980. Weights of Colorado sage grouse. Braun, C. E., T. Britt, and R. O. Wallestad. 1977. Guidelines for maintenance of sage grouse habitats. Wildl. Soc, Bull. 5:99-106. Dash, J. A. 1988. Effect of dietary terpenes on glucuronic acid excretion and ascorbic acid turnover in the brushtail possum (Irichosurus vulpecula). Compo Biochem. Physiol. 89B:22l-226. Foley, Y. J., E. V. Lassak, and J. Brophy. 1987. Digestion and absorption of Eucalyptus essential oils in greater glider (Petauroidel yolans) and brushtail possum (Irichosurus vulpecula). J. Che•. &col. 13:2115-2130. Gasaway, Y. C. 1976. Methane production in rock ptaraigan (Li&opus ~). Comp Bioche •. Physiol. 54A:lS3-1S5. Hannon, S. J. B. R. Simard, and F. C. Zwickel. 1979. gonadal cycles of adult and yearling blue grouse. 1289. . t Differences in the Can. J. Zool. 57:1283- Horwitz, Y., ed. 1980. Official methods of analysis of the Association of Official Analytical Chemists. Assoc. Off. Anal. Che•. , Washington, D. C. 1018pp. King, J. R., and M. E. Murphy. 1985. Periods of nutritional stress in the annual cycles of endotherms: fact or fiction? Am. Zool. 25:955-964. Martin, D. Y., P. A. Mayes, V. Y. Rodwell, and D. K. Granner. 1985. Harper's review of biochemistry. Lange Medical Publ., Los Altos, Calif. 718pp. �Moss, R., and I. Hanssen. 50B:555-567. Patterson, R. L. 1952. Colo. 34lpp. 1980. Grouse nutrition. The sage grouse Nutr. Abstr. in wyoming. Sage Books, Rev. Denver, Petersen, B. E. 1980. Breeding and nesting ecology of female sage grouse in North Park, Colorado. M.S. Thesis, Colorado State Univ., Fort Collins. 86pp. Remington, T. E. 1983. Food selection, nutrition, and energy reserves of sage grouse during winter, North Park, Colorado. M.S. Thesis, Colorado State Univ., For,- Collins. 89pp. In prep. grouse. Costs of detoxification _____ , and C. E. Braun. 1985. Park, Colorado. J. wildl. of secondary compounds Sage grouse food selection Manage. 49:1055-1061. Ricklefs, R. E. 1974. Energetics Ornithol. Club 15:152-297. Robbins, C. T. 1983. wildlife York, N. Y. 342pp. of reproduction Sibbald, I. R. 1976. A bioassay Poultry Sci. 55:303-308. in winter, in birds. feeding and nutrition. for true metabolizable to blue North Publ. Nuttall Academic energy Press, New in feedstuffs. Sipes, I. G., and A. J. Gandolfi. 1986. Biotransformation of toxicants. Pages 64-98 in C. D. Klaassen, M. O. Amdur, and J. Dou11, eds. Casarett and Doull's toxicology. Macmillan Publ. Co., New York, N.Y. Sokal, R. R., and F. J. Rohlf. York, N.Y. 859pp. 1986. Biometry, 2nd ed. W. H. Freeman, New Vehrencamp, S. L., J. w. Bradbury, and R. M. Gibson. 1989. The energetic cost of display in male sage grouse. Anim. Behav. 38:885-896. Zar, J. G. Condor Prepared Approved 1968. Standard 70:278. metabolism comparisons &-+-_~ _ {f)_7:B __ by __ Orrin B. Myers Graduate Research Assistant by Clait E. Braun Wildlife Research Leader among orders of birds. �JOB PROGRESS REPORT Colorado State of: Project: __~W~-Al~5~2~-R~ Work Plan: 3__ : Job Job Title: Response Period Covered: Author: _ Upland Bird Research 17 of Selected 01 January through Avifauna to Fire in the Big Sagebrush 31 December Type 1989 Lee A. Benson Personnel: C. E. Braun, T. D. Abell, C. E. Spraker, C. E. Poley, Colorado Division of Wildlife; L. A. Benson, Colorado State University ABSTRACT The response of sage grouse (Centrocercus urophasianus) and passerine birds to fire was studied in Jackson and Moffat counties, Colorado. Sage grouse numbers and habitat use were monitored following a fall prescribed burn in Jackson County, Colorado. Sage grouse were also monitored at 2 wildfire sites, one in Jackson County and one in Moffat County. Lek counts were conducted during the breeding season. Radio-marked sage grouse were located during spring and summer. Few locations of radio-marked sage grouse occurred within burned areas during the breeding season. Movements to summering areas ranged from 3.4 to 25.3 km in North Park and averaged 1.6 km in Moffat County. Diversity and relative abundance of passerines were determined for the prescribed burn site and an adjacent unburned area in 1988-89. Passerines were also sampled at the Jackson County wildfire site (burned 1987). The prescribed fire site burned in a mosaic pattern and resulted in a 50-60% reduction in big sagebrush (Artemisia tridentata) within a 38-ha area. Both wildfires consumed nearly all sagebrush habitat. During spring 1987, 75% of 28 locations of radio-marked male sage grouse were within the proposed prescribed burn plot. Use of the burned plot by male sage grouse in 1988-89 was minimal (9 of 130 relocations). Peak male sage grouse counts decreased 43% at the lek closest to the burn from 1987 to 1989. Passerine responses differed greatly at the 2 burn sites in Jackson County. At the wildfire site, all passerines were eliminated except for horned larks (Erernophila alpestris). Species richness did not vary along burned and unburned transects at the prescribed burn site. Relative abundance at the prescribed burn site varied slightly between burned and unburned areas. ��RESPONSE OF SELECTED AVIFAUNA TO FIRE IN THE BIG SAGEBRUSH TYPE Lee A. Benson INTRODUCTION Sagebrush (Artemisia ~.)-dominated rangeland occupies major portions of the western United States. Reports of total area covered by sagebrush vary from 35 (Schroeder and Sturges 1975) to 60 million ha (Beetle 1960). More than 10% of all sagebrush lands have been altered by man (Braun et al. 1976). Among the methods used, fire is becoming increasingly more attractive as a management tool for reducing sagebrush (Frandsen 1985). This study was initiated as additional information is needed on the impacts of fire on wildlife so that any impacts can be more fully considered in management of the sagebrush type. The major uses of sagebrush ranges are two-fold. Cattle grazing, mainly during the summer months, is by far the predominant use of sagebrush regions. Recreation is also a factor in management of these areas. Hunting is the primary recreational activity occurring on rangelands in the western United States. Sage grouse dependence on sagebrush regions is well known. The original range of sage grouse parallels that of big sagebrush (~. tridentata), although the current range is reduced over that of previous decades (Johnsgard 1983). Settlement of the western United States appears to have resulted in a reduction in the range of sage grouse. Attempts to reduce sagebrush have generally not been successful in its eradication but have caused declines in sagebrush dependent wildlife. Sage grouse dependence or sagebrush is evident from life history studies of this species ...Food habits studies by Patterson (1952) show a high proportion of sagebrush in the diet of sage grouse. Winter food is restricted to sagebrush alone. During late spring and early summer months, sage grouse include other food items in their diet. Few studies have examined the effect of burning on sage grouse. Gates (1983) reported greater use of burned areas one year following a prescribed burn. Klebenow (1972) suggested that small irregularly shaped burns would benefit sage grouse in Nevada. Braun et al. (1976) recommended that sagebrush treatments should be small in size and done on productive sites. Connelly et al. (1971) located 3 leks on disturbed sites in Idaho. Several studies have assessed the impact of other forms of habitat manipulation on sage grouse. The general conclusion is that removal of sagebrush decreases sage grouse populations. Martin (1970) observed fewer grouse on chemically treated areas. Wallestad (1975) reported declines in sage grouse populations following sagebrush control in Montana. Ploughing of small areas of sagebrush resulted in declines of sage grouse in south-central Montana (Swenson et al. 1987). Other species of interest are Brewer's sparrow (Spizella breweri), sage sparrow (Amphispiza belli), and sage thrashers (Oreoscoptes montanus); these are considered to be sagebrush obligates (Braun et al. 1976). Also of �46 interest are green-tailed towhees (Chlorura chlorura), vesper sparrows, (Pooecetes gramineus) and horned larks. Braun et al. (1976) consider greentailed towhees and vesper sparrows to be near obligates of sagebrush habitat. P. N. OBJECTIVES 1. Review 2. Capture 6-10 male sage grouse per lek and equip with poncho-mounted tail-clip transmitters. 3. Locate and flush radio-marked 4. Obtain 5. Monitor 6. Establish 7. Measure vegetation structure and composition areas and at sage grouse use sites. literature pertaining lek counts on avian species. birds at spring feeding/loafing or sites. (minimum of 3) from study area leks. post-breeding transects to the impacts of burning movements of radio-marked and conduct breeding sage grouse. bird surveys at study areas. on treated and untreated STUDY AREAS Field research for this project was conducted in 2 northern Colorado counties. Three study sites were located in Jackson County in north-central Colorado. An additional site was located in Moffat County in northwest Colorado. 0 0 0 Jackson County lies between 106 42' and 105 50' west longitude and 40 20' and 41 north latitude. North Park comprises a large proportion of Jackson county. North Park lies between the Never Summer and Medicine Bow ranges on the east and .the Park Range on the western edge of the park. The southern border is comprised of the Rabbit Ears Range. Independence and Sentinel mountains mark the northern boundary. 0 Three sage grouse leks served as the approximate centers for the Jackson County study sites during the April-May interval. Deer Creek Lek was in the southeastern corner of North Park in the NW quarter of Section 31, Township 7 North, Range 78 West. Perdiz Lek was 14 km southeast of Walden in the NE quarter, Section 10, Township 8 North, Range 78 West. Fish Hatchery Lek was 10 km southwest of Walden in the southwest quarter of Section 9, Township 8 North, Range 80 West. Topography varied greatly between sites. The Deer Creek area was large, relatively flat expanses of two basic types. Bench top areas of 2590-2615 m in elevation form the area of the lek site. These benchtops extend north of the Deer Creek Lek 1.5 km with major areas to the south. Lower lying flat areas occurred directly west of the lek and off the north end of the bench on which the Deer Creek Lek is located (elevation 2548-2566 m). Areas of 20-40 lie between the benchtops and lowland areas at Deer Creek. 0 �47 The topography at the Fish Hatchery Lek was a series of small west to east ridge tops. Small drainages lie between these ridges. Also present were large expanses of low lying, alkali flats. This topography occurred east and northeast of the lek and forms the western edge of Case Flats. Elevations ranged from 2487 to 2511 m. The Perdiz Lek was in a depression with small scattered ridges sloping southwest to northeast towards Bolton Draw. To the west and north of the Perdiz Lek about 2 km were large areas with little topographic change. Elevations ranged from 2515 m near the lek to 2579 m in the large flat areas. The Moffat County site was 9 km west of Maybell. The lek was in the southwest quarter of Section 33, Township 7 North, Range 96 West. Elevations ranged from 1844 to 1963 m. The area was a series of large ridges sloping south to north. Large to moderate sized drainages separated ridges. METHODS Sage Grouse Capture and Radiomarking.--Sage grouse were spotlight and a long-handled net on or near were banded with serially numbered aluminum color coded by year. Age, sex, weight, and each bird captured. captured at night using a leks (Giesen et al. 1982). Birds leg bands and plastic bandettes primary molt were recorded for Twenty-nine sage grouse were fitted with either poncho (Amstrup 1980) or tailclip (Bray and Corner 1972) radio transmitters obtained from Wildlife Materials (Carbondale, IL), Telemetry Systems (Mequon, WI), and Advanced Telemetry Systems (Isanti, MN). Radio package weights ranged from 19.5 to 34.0 g and represented 0.7-1.4% of male body weights and 1.3-1.4% of female body weights. Relocations.--Radio-marked sage grouse were relocated on spring ranges from 10 April to 28 May. Birds were flushed and the location was marked to facilitate habitat measurements. Group size and flushing distance were also noted. An effort was made to flush all birds in the immediate area. Locations were plotted on 7.S-minute U.S.G.S. topographic maps. UTM map coordinates were obtained and used for distance calculations. Average daily spring movements of sage grouse were obtained for 29 male sage grouse during the breeding season. Radio-marked sage grouse in North Park were relocated at approximately weekly intervals from June through September. Four attempts were made to locate radio-marked sage grouse in Moffat County resulting in a limited sample size for the Thornburg Well Lek. Nineteen sage grouse had active radios during the summer months. Lek Counts.--Counts of male and female sage grouse were conducted between 10 April and 28 May. Birds present on leks were counted 3-5 times during a 15-20 minute interval during each lek visit. Counts occurred within 1 hour of sunrise from a distance of 30-1,000 m depending upon weather, access, and observer preference. Numbers of male and female sage grouse were obtained. �Breeding Bird Surveys Breeding bird surveys were conducted during June and on 2 July at the Deer Creek and Perdiz study sites. Transects at Deer Creek were counted 5 times between 1 and 28 June. Transects at the Perdiz study site were counted 3 times (29 and 30 Jun, 2 Jul). Two counts were obtained at the Thornburg Well site but did not constitute an adequate sample. These counts are not included in this report. Counts started 0.5 hour before sunrise and were completed within 2 hours of sunrise. A grid form was used to plot the estimated distance of birds from the transect. Singing males and non-singing individuals were recorded. Calculation of species d~versity, evenness, and richness followed Wiens and Rotenberry (1981). Only singing males were used for calculation of birds/40 ha, relative abundance, species diversity, evenness, and richness on all transects. Counts were averaged and then grouped by burn, burn edge, and unburned. These values were then converted to birds/40 ha. The only exception was the unequal length burn transects at the Deer Creek site. These values were converted to birds/40 ha and then averaged. Transects were slightly altered from those counted in 1988. An additional 500 m transect was established and counted in unburned habitat at the Deer Creek study site. The edge and unburned habitat transects were reduced to 500 m in length at the Perdiz study site. As in 1988, width was limited to 70 m from each side of the transect. Singing and non-singing individuals were recorded. Habitat Measurements Vegetation measurements were obtained from sage grouse spring feeding/loafing sites and from 10 plots within each burn in North Park. Measurements were taken using methods described by Canfield (1941) and Daubenmire (1968). Two 10-m lines were measured at each location. Lines were oriented north-south and east-west centered over sage grouse flush locations or at random locations. Shrubs, forbs, and grasses were estimated to the nearest 5%, if greater than 5%, at 20 0.5-m2 plots (lO/line at I-m intervals). Percentages less than 5% were estimated to the nearest 1%. Canfield's technique was used to measure the canopy and subcanopy while the 0.5-m2 plots were used to examine ground cover at or just above the soil surface. RESULTS AND DISCUSSION Sage Grouse Lek Counts Counts of sage grouse present were obtained for Thornburg Well (n - 6), Deer Creek (n = 5), Fish Hatchery (n - 4), and Perdiz (n - 4) leks (Table 1). Peak male sage grouse counts at Deer Creek were 43% lower than in 1987 (preburn) and 28% lower than in 1988. Low juvenile male recruitment appeared to be responsible for the reduction in male lek attendance. It is difficult to relate this to the burn as counts at 11 of 23 known leks in North Park decreased in lek attendance, ranging from 13 to 63% (~ - 35%) for the 1987-89 �interval. Peak hen attendance attendance. remained consistent with 1988 peak hen Peak male attendance at Perdiz decreased 50% from 1987 to 1989. Other grouse leks in the Perdiz area had varied trends in peak male counts. at Turkey decreased 29% while those at Raven increased 7%. Counts at indicated an increase in peak male attendance of 96%. Counts of males Hatchery remained constant during the 1987-89 interval (+1 male). sage Counts Denmark at Fish Peak male sage grouse attendance decreased slightly at Thornburg Well between 1988 to 1989. This lek was not systematically counted in 1987 but was counted at least once. High counts for male sage grouse in 1987 and 1989 were 17 and 16, respectively. Table 1. Counts of sage grouse on leks, Jackson May, 1989 and 1973-88. Deer Creek High count Males Females N counts Dates of high count Male Female Peak male counts 1988 1987 1986 1985 1984 1983 1982 1981 1980 1979 1978 1977 1976 1975 1974 1973 and Moffat Fish Hatchery counties, Perdiz April- Thornburg Well 13 13 36 27 8 5 16 2 5 4 4 6 9 Apr 16 Apr 12 Apr 12 Apr 18 23 21 45 14 47 66 52 28 43 41 31 36 27 36 35 32 26 16 63 50 67 78 82 97 64 81 69 62 67 11 37 aNo systematic counts, NC - no count. byear of initial location. 7 & 17 May 17 Apr 13 16 6 10 8 21 27 23 8 16b 22 Apr 22 Apr 20 17a 15a Sa NC 8a 8a 20 28 35 32 36b �50 Sage Grouse Relocations Twenty-nine sage grouse were captured and radiomarked during spring 1989. Twelve sage grouse (9 males, 3 females) were captured and radiomarked at Deer Creek. Twelve (10 males, 2 females) were captured and radiomarked at Fish Hatchery. Four males and 1 female were captured and radiomarked at Thornburg Well. No sage grouse were radiomarked at the Perdiz study site in 1989. Survival of radio-marked sage grouse was relatively good during Spring 1989. Eight male sage grouse provided 9-15 locations each at the Deer Creek study site for a total of 55 independent observations. A total of 152 grouse was observed while flushing radio-marked male sage grouse during April and May. Male sage grouse habitat use patterns at Deer Creek were similar to those documented in 1988. Average distances from the lek for breeding season locations was 0.8 km (0.7-1.1 km) in 1989 and 0.9 km (0.2-1.8) in 1988. Areas of observed use were similar between years. Several 1988-89 locations were within meters of each other. The north and east sides of the bench adjacent to the Deer Creek Lek were the most commonly used areas in both years. Use of the burn plot was limited but increased slightly over 1988. Eight of 55 observations of sage grouse were within the burn plot in 1989. In 1988, only 1 of 75 locations was within the burn plot. Two of 3 radio-marked hens at the Deer Creek study area successfully hatched clutches in 1989. All hens nested, but one nest (hen 2341) was destroyed. Average distance from the lek to the 3 nest sites was 2.4 km and ranged from 1.1 to 4.1 km. Accurate clutch size information was not obtained but remains of at least 5-6 eggs were found at each nest site. Both successful hens used meadow areas north and east of the Deer Creek Lek following hatching of clutches. No more than 1 chick(hen was observed. Following initial movements from nest sites, hens with chicks moved only short distances between observations. June locations were obtained for 7 male sage grouse from the Deer Creek Lek. A total of 121 grouse was observed during 44 locations of radio-marked male sage grouse. Radio-marked male 151.059 was captured in 1988 and was not located until after the breeding season. The other 6 males were captured in 1989. Males were located 4-9 times following the breeding season and prior to 1 July. Use areas were similar to areas used by sage grouse in 1988. Average distance from the lek of capture of male sage grouse was 4.8 km and ranged from 2.3 to 8.1 km. Group sizes ranged from 2 to 16 grouse with an average of 4. During the July-September interval ten males and 3 females were relocated at the Deer Creek study site. Average distances from the lek were 5.2 and 3.4 km for males and females, respectively (range 3.0-7.2). As in 1988, use was intensive near meadow areas for most birds. One radio-marked male captured in 1988 was located along the Illinois River drainage west of the Deer Creek Lek. This bird appeared to use the same area as in summer 1988. The radio from this bird was recovered on 15 August 1989. No sign of mortality was observed. Other sage grouse were also located primarily west and south of the Deer Creek Lek, primarily along the Illinois River drainage. Another area of intensive use was observed north and east of the Deer Creek Lek and Owl Ridge, along the Owl Creek drainage and adjacent upland areas. One radio-marked hen was shot during the 1989 hunting season (Band #2341). A total of 450 grouse was seen during 75 observations of radio-marked sage grouse. �-, )L Ten male and 2 female sage grouse were captured and radio-marked at the Fish Hatchery study site in 1989. Three male sage grouse captured in 1988 were present in 1989, although during the 1989 breeding season independent observations were only obtained for 2 of these birds. Spring locations were obtained on a total of 11 male sage grouse from the Fish Hatchery Lek. Fifty independent observations were obtained for male sage grouse during the breeding season. A total of 230 sage grouse sightings was recorded while obtaining independent observations. The average distance from the lek was 1.1 km in 1989 compared to 0.9 km in 1988. Areas of use were similar in both years. Nesting activity was not observed for the radio-marked female sage grouse at the Fish Hatchery study site. Movements from the Fish Hatchery Lek to summer range were similar to those observed in 1988. Ten male sage grouse were located 2-5 times during June for a total of 35 locations. Two hundred and twelve sage grouse were observed during these locations. Group sizes ranged from 2 to 70 birds and averaged 9. The average distance from the lek for 10 male sage grouse was 12.7 km and ranged from 6.5 to 24.7 km. As observed in 1988, 2 male sage grouse (926, 1988; 1149, 1989) moved in excess of 20 km southeast to the Willow Creek drainage northwest of Rand. Ten radio-marked sage grouse (9 males, 1 female) from the Fish Hatchery study site were located in July-September while 2 were not located. Grouse were observed along drainages and adjacent upland areas of the North Platte River, Grizzly Creek, Little Grizzly Creek, Willow Creek, and Chedsey Creek. Average distances from the lek for male sage grouse were 13.0 (range 6.4-25.2) and 4.8 km for the hen. A total of 563 sage grouse was observed on 38 occasions. One radio was recovered just north of the Chedsey Creek drainage (Band #1139). This bird was subsequently shot during the fall hunting season. The hen was also shot during the hunting season. Four male and 1 female sage grouse were captured and radiomarked at Thornburg Well during April 1989. Five male sage grouse captured at Thornburg Well in Spring 1988 were located on and near the lek in 1989. A broken poncho radio package was found on the lek early in the breeding season from male #990. Locations were obtained from 7 male sage grouse. As documented in North Park, breeding season habitat use patterns were similar in 1988 and 1989. Use of the burn was not observed until later in the spring. One male was observed in a remnant patch of sagebrush within the burn on 2 dates in May. A total of 82 sage grouse was observed at 37 locations. Average distance from the lek for male sage grouse during the breeding season was 0.65 km compared to 0.8 km in 1988. Group size averaged 2 male sage grouse and ranged from 2 to 14; single birds were commonly observed. The radio-marked female sage grouse (band 2340) nested about 1 km SE of Thornburg Well Lek. This hen was inadvertently flushed from a nest of 5 eggs when first located. An additional egg was laid and the nest was abandoned. A renesting attempt was also unsuccessful. Following the breeding season, most male sage grouse at the Thornburg Well study site moved east and south. One male (band 2292) moved north, near the Yampa River, and was predated between 16 May and 22 June. Sage grouse were observed in the burn near the lek and in other burned areas. Only 4 radios (3 males, 1 female) were active during July-September at the Thornburg Well study site. Two male sage grouse were not located. This site was visited approximately once per month. Sage grouse were observed to use areas along the Yampa River drainage. Birds were also found in areas south and east of Thornburg Well. Average distances from the Thornburg Well Lek �52 were 7.3 km (range 5.1-9.6) for males and 5.8 km for females. grouse was observed on 6 occasions. A total of 60 Passerines Breeding Bird Survey.--Breeding bird surveys were conducted 5 times at the Deer Creek study site during June. Dates of counts were 16, 20, 26, 27, and 28 June. The 20 June count was an evening count. Density estimates between 1988 and 1989 varied (Tables 2, 3). The maximum number of males observed during all counts was used to derive density estimates. This should be equivalent to the maximum number of breeding pairs/40.5 ha. However, additional analysis of the data is necessary before final conclusions are drawn. Transects at the Perdiz study site were counted on 29 and 30 June, and 2 July. Table 2. Density (males/40.5 ha) of breeding birds observed at the Deer Creek study site, Jackson County, Colorado, summer 1988 and 1989. Data are based on highest number of singing males observed during 3 counts in 1988 and 5 counts in 1989. Species Brewer's sparrow Vesper sparrow Sage thrasher Horned lark Green-tailed towhee Totals 1988 Unburned 1989 % change 6 6 12 8 +9 -23 -67 0 -33 82 68 -17 23 35 6 25 27 2 1988 Burned 1989 % change 6 18 4 10 3 18 10 3 15 3 +200 -44 -25 +50 0 41 49 +20 Density varied markedly between burn and unburned areas. Vesper sparrows and green-tailed towhees had lower densities between burned and unburned transects. Relative abundance and density of horned larks were greater in burned transects at all study sites. Species diversity was greater within the burn transects at Deer Creek but was lower at the wildfire sites. Habitat Measurements Measurements were summarized for male sage grouse locations and 20 burn plots. Averages for most variables appear to be within ranges previously reported for North Park (Schoenberg 1982, Hernandez 1988). The percentage of sagebrush was somewhat lower because of some locations within burns. Vegetation variables measured for spring grouse locations at the Deer Creek study site were similar to those observed in 1988, although percent sagebrush cover was lower than in 1988 (Table 4). Burns were characterized by low vegetation cover although increases were observed for all variables except based on highest number of singing males observed during 5 counts in 1988 and 3 counts in 1989. number of dead plants/m2 (Table 5). Young sagebrush plants were encountered at a higher rate than in 1988. Seventy-seven seedlings were �Table 3. Density (males/40.5 ha) of breeding birds observed at the Perdiz study site, Jackson County, Colorado, summer 1988 and 1989. Data are species 1988 Brewer's sparrow Vesper sparrow Sage thrasher Horned lark Green-tailed towhee 40 12 6 1 6 65 Totals Unburned 1989 X change 1988 43 12 3 0 9 +8 0 -50 -100 +50 3 12 6 6 0 67 +3 27 Edge 1989 Burned 1989 X change X change 1988 23 12 0 0 0 +667 -23 -100 -100 0 0 0 0 3 0 0 0 0 9 0 0 0 0 +200 0 35 +30 3 9 +200 V' �54 observed in burn plots at the Deer Creek site. Heights for these plants ranged from 0.8 to 9.5 cm and averaged 3.2 cm. Average widths were 2.6 cm (range 0.4 - 8.6). Forty-seven sagebrush seedlings were encountered in the Perdiz sample. Average heights were 5.2 cm (range 1.5 - 12.5) and widths averaged 5.8 cm (range 1.2 -16). Table 4. Vegetative site, Jackson County, parametersa at sage grouse use sites, Deer Creek study Colorado, 1988-89. Sagebrush Liveb Deadb 1988 1989 (n = 59) (n - 49) 2.1 (0.4-5.0) 0.5 (0-1.8) Pe r c erit.? 3.2 (1.2-7.8) 0.8 (0.1-2.8) 20.8 (10.1-35.2) 28.2 (3.3-50.9) Ave height, cm Ave width, cm Ave length, cm 34.3 (17.0-72.2) 43.9 (17.2-84.4) 35.7 (11.9-65.3) 25.7 (11.1-55.1) 33.3. (16.2-55.5) Ave space, cmc,d' 75.0 (29.9-321. 8) 52.8 (29.6-105.2) Ave shrub, Xe Ave forb, Xe Ave grass, Xe 9.3 (0.1-21.0) 7.0 (0-24.4) 13.0 (3.4-73.2) aMean (Range). bSagebrush p1ants/m2. CLine-transect data. dDistance between sagebrush eDerived from 0.s-m2 plots. Table 5. 9.8 (2.4-21.5) 7.4 (0.6-21.9) 8.0 (2.1-26.5) plants. Habitat variables8 On burn plots, Jackson county, Colorado, 1988-89. Deer Creek Perdiz 1988 Sag~br~sh LIve Deadb PercentC 3 (1.9-5.3) 0.8 (0-4.8) Ave shrub, Xd Ave forb, Xd Ave grass, Xd 2.5 (0-10.4) 1.3 (0.2-3.4) 5.1 (2.3-7.3) o aMean (Range )• bsagebrush plants/m2. CLine-transect data. dDerived from 0.5·m2 plots. 1989 1988 0.2 (0.1'0.5) 3.9 (2.3-5.6) 0.06 (0'0.2) 2.4 (2.0-3.2) 0.5 (0-2.2) o 1.8 (0-5.6) 2.7 (1.1-5.2) 12.6 (9.1'17.4) 0.4 (0-1.9) 0.7 (0.1-1.9) 3.9 (1.6-5.8) 1989 0.1 (0'0.4) 3 (1.6'3.9) o 1.5 (0'3.5) 1.7 (0'4.8) 14.0 (10.2'21) �55 LITERATURE Amstrup, S. C. 44:214-217. Beetle, A. A. Artemisia. 1980. CITED A radio collar for game birds. J. Wildl. Manage. 1960. A study of sagebrush. The section Tridentatae Wyoming Agric. Exp. Stn. Bull. 368. 83 pp. of Braun, C. E., M. F. Baker, R. L. Eng, J. S. Gashwiler, and M. H. Schroeder. 1976. Conservation committee report on effects of alteration of sagebrush communities on the associated avifauna. Wilson Bull. 88:165171. Bray, O. E., and G. W. Corner. 1972. A tail clip for attaching to birds. J. Wild1. Manage. 36:640-642. Canfield, R. H. 1941. Applications of the line interception sampling range vegetation. J. For. 39:388-394. Connelly, J. W., W. J. Arthur, recently disturbed sites. Daubenmire, 266. R. 1968. transmitters method in and O. D. Merkham. 1971. Sage grouse leks on J. Range Mange. 34:153-154. Ecology of fire in grasslands. Adv. Ecol. Res. 5:209- Frandsen, O. A. 1985. Fire as a management tool in southeast Idaho - a case study. Pages 85-87 in K. Sanders and J. Durham, eds. Rangeland fire effects: a symposium. U.S. Dep. Inter., Bur. Land Manage., Boise, Idaho. Gates, R. J. 1983. Sage grouse, 1agomorph, and pronghorn use of a sagebrush grassland burn site on the Idaho National Engineering Laboratory. M.S. Thesis, Montana State Univ., Bozeman. 135 pp. Giesen, K. M., T. J. Schoenberg, and C. E. Braun. 1982. Methods for trapping sage grouse in Colorado. Wi1d1. Soc. Bull. 10:224-231. Hernandez, E. J. 1988. Response of selected avifauna to prescribed burning in the big sagebrush type. Colorado Div. Wi1d1. Job Progress Rep., Fed. Aid Proj. W-152-R. Apr:105-138. Johnsgard, P. A. 1983. Lincoln. 413 pp. The grouse of the world. Univ. Nebraska Press, K1ebenow, D. A. 1972. The habitat requirements of sage grouse and the role of fire in management. Proc. Tall Timbers Fire Eco1. Conf. 12:305-315. Martin, N. S. 1970. Sagebrush control related occurrence. J. Wi1dl. Manage. 34:313-320. Patterson, R. L. CO. 341 pp. 1952. to habitat The sage grouse in Wyoming. and sage grouse Sage Books Inc., Denver, �56 Schroeder, M. H., and D. L. Sturges. sparrow of spraying big sagebrush. 1975. The effect J. Range Manage. on the Brewer's 28:294-297. Schoenberg, T. J. 1982. Sage grouse movements and habitat selection M.S. Thesis, Colorado State Univ., Fort Collins. Park, Colorado. Swenson, J. E., C. A. Simmons, and C. D. Eustace. 1987. Decrease grouse (Centrocerus urophasianus) after ploughing of sagebrush Biol. Cons. 41:125-132. wal1estad, R. 1975. Male sage grouse Wild1. Mange. 39:482-484. responses to sagebrush Wiens, J. A., and J. T. Rotenberry. 1981. Habitat structure of birds in shrubsteppe environments. Prepared Approved by Lee A. Benson Graduate Research Assistant C1ait E. Braun Wildlife Research Leader by in ~orth 86 ?p. of sage steppe. treatment. J. associations and community Ecol. Monogr. 5:21-41. �JOB PROGRESS State Colorado of: Project: T.J-152-R Work _8_: Plan: Job Title: Period Personnel: Upland Kenneth 5ird Researc~ Job _5_ Population Inventor'! and Habitat in Southeast Colorado Covered: Author: REPORT 01 January M. through 31 December Use by Lesser Prairie-chicker:s 1989 Giesen Dave Clarkson, Steve Wagner, Bryant Will, Fellinger, Ken Giesen, Jennie Colorado Division of Wildlife Slater, Chuck ABSTRACT Lek surveys were conducted on a 41.4-~~2 study site and adjacent area near Pasture lAE of the Comanche National Grasslands in Baca County, Colorado. Numbers of leks and lesser prairie-chickens (Tympanuchus pallidicinctus) declined slightly on the study site although numbers of males counted on all leks surveyed in Baca County increased slightly. There was a strong p osi ri ve corr0.ation between lek density and breeding male density (£ = 0.93) but not between average lek size an~ breeding male density (£ = 0.41). Twenty-nine lesser prairie-chickens (14 males, 15 females) were trapped and banded with 11 males and all females being fitted with radio transmitters. Females dispersed an average of 1.39 km from lek of capture to their nest site although the nearest lek to nest site distance was 0.98 km. Clutch size averaged 10 eggs but: none of the nests of radio-marked hens produced chicks. Calculated average home range size of males was 0.92 km2 and that of females was 3.75 km2. Hens selected taller vegetation and denser sandsage stands for nests than at random sites but preference of males and females for summer feeding and loafing s i t es did not .di f f er from random sites for 7 variables measured. ��POPULATION INVENTORY AND HABITAT USE BY LESSER PRAIRIE-CHICKENS IN SOUTHEAST COLORADO Kenneth M. Giesen Within ~he past century the range of lesser prairie-chickens in North America decreased by 92% (Taylor and Guthery 1980) and the population declined by 974 (Crawford 1980). Although the exact historic distribution and population size are unknown, early reports (Bendire 1892, Judd 1905, Bent 1932, Baker 1953. Sands 1978) suggested lesser prairie-chickens were abundant and widely distributed throughout their range. Aldrich (1963) indicated lesser prairiechickens historically inhabited about 360,000 km2 in 5 states while recent estimates suggest a population of 50,000 birds on 125,000 km2 (Crawford 1980, Taylor and Guthery 1980, Johnsgard 1983). Although evidence suggests lesser prairie-chickens were historically peripheral in Colorado, they were thought to be common to abundant in 6 southeastern counties (Baca, Bent, Prowers, Kiowa, Lincoln, and Cheyenne), anc peripheral in adjacent counties (Loeffler 1983). Recent surveys have documented breeding populations in Baca, Prowers, and Kiowa counties (Hoffman 1963, Loeffler 1983, Rash 1985, this study) and unconfirmed sightings in Lincoln County (T. Mathieson, pers. commun.). The number of known active leks is < 50 and the estimated population is 1,200 to 1,500 birds. Lesser prairiechickens were transplanted into Pueblo County in 1988 and 1989 but the success of the introduction has not been documented. Because of its limited distribution and small population size, the lesser prairie-chicken is classified as a threatened species in Colorado. P. N. OBJECTIVES The objectives of this study are to evaluate lek surveys as indices to population trends, ascertain the accuracy of aerial and ground surveys in detecting leks, describe the seasonal floristic and structural characteristics of lesser prairie-chicken habitats in southeastern Colorado, and contribute to preparation of a recovery plan for lesser prairie-chickens in Colorado. SEGMENT OBJECTIVES 1. Review pertinent literature applicable to the objectives of this study. 2a. Locate all active leks within a 4l.4-km2 primary study area and obtain at least 1 count/week (Mar-May) of all males and females on each lek. 2b. Survey all historic leks in Baca County and obtain at least 1 count of males and females on each active lek. 3. Trap and band lesser prairie-chickens on active leks within the primary study area. Up to 20 will be marked with miniature radio transmitters to f~cilitate their periodic location. �-, /' .") ;-' 4. Locate Record lesser clutch 5. Locate radio-marked movements. 6. Measure vegetation cover height of shrubs, forbs, 7. Compile data, prairie-chicken nests by following radio-marked size, incubation period, and nest fate. birds analyze weekly for estimates of seasonal at grouse use sites and random and grasses will be recorded. results, and prepare annual home sites. progress hens. range a~G VOR and report. METHODS Field surveys were conducted on the Comanche National Grasslands and ad':_cent areas in eastern Baca County from mid-March through June using binocula __ -.,a parabolic microphone listening device, and a trained pointing dog to locate active lesser prairie-chicken leks. Leks known to be active in 1988 were surveyed as were most known historic lek sites. Active leks were visited within 2 hours of sunrise to count grouse and classify them to sex. Lek density was defined as the number of active leks on the 41.4 km2 study area and breeding male density was calculated as the sum of the high counts of males on the study area and expressed as males/km-. " Cannon nets and funnel traps (Giesen et al. 1982, Haukos et al. 1990) were used to capture grouse on leks. Each captured grouse was marked with a numbered aluminum band and a unique combination of colored plastic bandettes. Miniature solar- or batterypowered transmitters (weight 18-24 gms) were attached to all captured females and selected males using a poncho (Amstrup 1980). Radio-marked birds were located using a hand-held 3-element yagi antenna and portable receiver. Birds were approached on foot until they flushed and locations were plotted on topographic maps (scale 1:2400). The minimum convex polygon method (Mohr 1947) was used to calculate home range. Vegetative structure and species composition were measured using line-intercept of canopy cover (Canfield 1941) and a range pole (Robel et al. 1970). Plant nomenclature follows Harrington (1964). Sand sagebrush (Artemisia filifolia) density was measured on O.OOl-ha circular plots. Study Area The primary study area was a 41.4-km1 (16 mi1) area of rangeland on or adjacent to Pasture 1AE of the Comanche National Grasslands in Baca County approximately 20 km east of Campo, Colorado. Included were sections 21-28 and 33-36 of T34S, R44W and sections 1-4 of T35S, R44W. Approximately 1,160 ha was privately owned and was rangeland except for a 16-ha milo field (fallow in alternate years). The topography was rolling hills bisected by Murray and Mitchell draws. Elevation ranged from 1,240 m on the west to 1,070 m on the east. Soils were predominately sandy loams with fine sands in the draws. Topography, grazing, and revegetation efforts have resulted in a diverse vegetative community. Sand sagebrush, broom snakeweed (Gutierrezia sarothrae), and yucca (Yucca glauca) were the predominant shrubs. A variety of grasses occurred with sand dropseed (Sporobolus cryptandrus), 3-awn (Aristida longiseta), sideoats grama (Bouteloua curtependula), and blue grama (~ gracilis) being most widespread. Vegetative structural characteristics of the study area were summarized in a previous report (Giesen 1989). �RESULTS AND DISCUSSION Lek Surveys A total of 283 counts of 25 active leks was obtained in Baca County between 22 March and 25 May 1989 (Table 1). In addition, 8 historic leks were surveyed and found to be inactive. High counts on all leks indicated 241 males, 43 females, and 312 total birds for an average of 12.5 birds/active lek. Hen attendance on leks was not synchronous and a minimum of III females was counted between 31 March and 20 May with most Cn = 79, 71.2%) being observed between 12 and 19 April. Although fewer active leks are being surveyed, the total number of lesser prairie-chickens counted and average number of birds/active lek have increased since the study was initiated in 1986 (Table 2). Since most of the same areas are surveyed each year, this suggests that populations may be increasing. Table 1. Lesser prairie-chicken n Lek 2 3 5 6 7 12 14 15 17 18 25 27 28 33/39 35 37 40 42 86-1 86-3 87-2 87-4 87-5 88-2 88-3 Totals Mean counts 21 24 33 11 9 1 1 7 1 11 1 12 38 11 7 7 7 1 26 7 2 28 - 13 3 283 Count period 22 22 23 25 11 1 29 23 29 29 1 1 30 29 16 16 29 31 30 1989a. lek count data, Baca County, Mar-25 Mar-25 Mar-25 Mar-19 Apr-18 28 Apr 30 Apr Apr-18 05 Apr Mar-21 05 Apr Mar-25 Mar-25 Mar-19 Apr-18 Apr-18 Mar-20 01 Apr Mar-25 Apr- 25 Apr-26 Mar-25 OS Apr Mar-2S Mar-03 May May May May May Apr May May May May Apr Apr May May May Apr May May May Males High count Females 13 9 21 7 16 3 5 17 11 11 6 3 15 10 22 11 13 6 10 1 9 1 2 9 13 10 21 9 20 3 8 21 13 17 7 4 19 14 25 15 16 2 2 5 6 2 5 1 2 2 2 5 1 9 3 12 2 1 241 43 10.5 To t a l s" 12 11 10 10 3 13 9 312 12.5 2.5 aLeks 4, 9, 13, 19/38, 23, 86-2, 87-1, and 88-1 were surveyed birds were observed. brnc1udes males, females, and birds not classified to sex. and no �Table 2. Lesser prairie-chicken lek survey data, Baca County, D birds counted D active 1986-89. ~ Year leks Males Females Totals 1986 1987 1988 1989 31 30 26 25 197 219 202 241 22 22 24 43 261 281 289 312 birds/lek 8.4 9.4 11.1 12.5 Summaries of lek count data since 1977 (when written records became available for analysis) indicate the dynamic nature of leks with many becoming inactive and others being established (Table 3). Fifteen of 57 (26.3%) leks being active one or more years since 1977 were initially located in 1985-89 as a result of more search effort and possibly increasing populations of lesser prairie-chickens. Only 13 of 28 leks (46.4%) active from 1977 to 1980 were still active in 1989, yet the total number of active leks in the area has increased. Because numbers of grouse attending leks fluctuates during the March through May breeding season, and survey timing and effort varied each year, the data (Tables 1, 3) may not reflect actual population changes and should be used only to indicate long-term trends. Analysis of lek surveys and lek count data on the 41.4 km2 study area (Table 4) indicate a strong positive correlation between male breeding density and number of active leks (£ = 0.93) but not between male breeding density and average lek size (£ = 0.41). If the high count of males on leks is a constant proportion of the total population, lek density is a better index to population changes than average lek size. These data support results of prairie grouse population surveys in other states (Cannon and Knopf 1981, Martin and Knopf 1981). These results suggest lesser prairie-chicken population changes are better reflected in total number of active leks rather than average number of birds/lek. Thus, inventory methodology should incorporate plans to document annual status (active/inactive) of historic leks and increase efforts to search for additional leks. Trapping and Banding A total of 27 lesser prairie-chickens (14 males, 13 females) was trapped on 4 of 7 active leks within the primary study area. Two females were trapped on another lek adjacent to the study area. Seven of 13 males classified to age were adults while only 2 of 15 hens were adults. The skewed age ratio of hens may indicate a higher susceptibility of yearlings to trapping. All hens captured after 13 April were yearlings suggesting that adult hens may attain breeding condition first or be dominant to yearlings on leks. �High counts of male lesser prairie-chickens, Table 3. Lek 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 86- 1 86-2 86-3 87-1 87-2 87-3 87-4 87-5 88-1 88-2 88-3 19n 1978 1979 1980 1981 1982 4 18 8 9 24 22 15 1 0 0 7 0 7 17 17 16 2 3 3 5 1" 6 8 NC 15 11 HC HC HC HC 5 13 11 11 0 4 14 HC 0 4 13 16 6 NCa 10 2 7 12 15 11 0 0 0 HC 7 12 10 12 NC 3 HC NC 2 1 6 17 HC 0 1 17 12 15 14 11 0 0 0 0 17 8 9 9 0 0 9 0 0 0 4 19 0 3 9 7 7 0 6 15 7 16 14 6 0 0 0 0 20 8 8 0 0 0 9 0 0 0 3 30 0 6 4 4 7 4 18 14 6 8 6 0 3 12 7 29 19 9 0 0 0 0 14 7 6 0 0 0 10 0 0 0 0 33 0 11 2 4 4 0 10 17 0 6 _:b 0 10 12 Baca County, 19n·89. Year 1983 1984 1985 1986 1987 1988 1989 0 12 11 9 26 14 8 0 0 0 0 16 5 12 0 0 0 9 0 0 0 0 23 0 3 2 3 4 0 9 7 7 23 17 7 0 0 0 0 18 0 11 0 0 2 5 0 0 0 0 21 0 2 3 0 7 7 0 18 6 5 0 0 0 0 0 0 11 0 0 5 9 0 0 HC 9 3 3 18 3 8 0 HC 0 NC 4 0 8 HC HC 7 13 NC NC 0 NC 0 0 6 11 0 0 0 21 0 10 8 18 5 3 14 4 10 2 2 7 5 0 0 13 0 12 3 9 7 7 7 7 3 4 4 NC 12 11 6 23 6 12 0 NC HC HC 2 0 6 POC NC 11 11 NC HC 0 NC NC NC 5 NC 7 20 NC HC HC 0 11 HC 16 HC HC 0 7 6 NC HC HC HC NC HC 12 0 0 NC 13 9 0 21 7 16 HC 0 0 NC 16 2 0 17 6 9 HC HC NC HC 5 0 10 NC NC 4 10 NC NC 0 HC 2 NC 2 0 0 12 0 0 0 9 HC 9 0 0 13 0 0 0 16 7 12 7 10 4 0 0 16 0 0 0 0 2 0 0 0 10 0 12 2 11 8 3 0 5 0 0 1 15 6 7 0 NC 6 0 0 17 0 NC 9 0 8 HC 18 0 18 7 0 9 0 10 HC 5 NC 0 11 5 8 0 0 0 7 5 aHO bLek count. not yet located. 0 HC 9 4 3 5 0 0 NC NC 3 0 5 17 HC 11 11 0 HC NC HC 0 HC 6 NC 3 15 HC HC HC HC 10 HC 22 HC 11 0 0 13 HC 6 HC HC NC HC 10 0 0 0 12 2 �Table 4. Lek count trends of lesser prairie-chickens study area, Baca County, Colorado, 1980-89 n n Year leks males 1980 1981 1982 _983 1984 1985 1986 1987 1988 1989 6 6 6 6 4 5 7 8 9 7 59 55 59 65 46 46 75 74 100 82 ~ males/lek 9,8 9.2 9.8 10.8 ll.5 9.2 10.7 9.2 11.1 ll.7 on a 41.4-km2 Breeding density (males/km:) . ", (16 m i. -,1 Lek dens it:r 1.42 1.33 1.42 1.57 1.11 1.ll 1.81 1. 79 2.42 1.98 0.14 0.14 0.14 0.14 0.10 0.12 0.17 0.19 0.22 0.17 Eleven males and all females were fitted with poncho-mounted mlniature radio transmitters to facilitate their periodic relocation. Although transmitters were not observed to have detrimental effects upon grouse, they may have increased predation as only 7 marked grouse were known to have survived at least 3 months (Table 5). Increased predation of radio-marked birds has been reported in other studies (Herzog 1979, Warner and Etter 1983, Hines and Zwickel 1985, Marks and Marks 1987, Haukos 1988). Table 5. Colorado, Fates of radio-marked 1989. lesser prairie-chickens n in Baca County, birds Fate Males Females Dispersal or radio failure Transmitter lost from bird Depredated Survived> 3 months 1 4 2 4 4 1 6 4 5 5 8 8 II 15 26 Totals Totals Nesting Fates of 15 hens were ascertained using radio telemetry. Two hens were killed by predators prior to nesting, signals from 3 hens were lost within 2 weeks, 3 hens were located periodically but did not localize and no nests were found, and 7 hens were known to nest. Predation caused failure of all nests and none �of the hens was known to attempt renesting. Nests were only visited once to ascertain clutch size and were discretely marked with flagging 10 m from the nest. Predation was usually not detected until the hen was absent from the nest for 2 consecutive radio relocations (usually a week or more). Thus, the identity of predators was difficult to ascertain. None of the hens was killed during incubation. The average distance from capture site on leks to 7 nests was 1.39 krn (range 0.79 - 2.84 krn). The average distance to the nearest lek was 0.98 krn (range 0.36 - 1.76 km) with 5 of 7 hens nesting closer to a lek different from that of capture. Sell (1979) reported a similar range of movements between leks and nest sites. If hens nest within their spring home range, we can assume that the closest lek may not be the one selected for mating. Clutch size ascertained from 6 nests averaged 10.0 eggs (range 6 - 11) but the clutch of 6 eggs may not have been complete. Haukos (1988) reported clutch sizes from 3 to 11 but smaller clutches may have been incomplete or the result of renests. Habitat Characteristics Lek Sites Vegetative compos~t~on at 7 lek sites was recorded using a visual estimate of species composition. Density and height of sandsage was measured along transects radiating from the center of the 1ek to the periphery along the 4 primary compass directions. Short grasses predominated with forbs and shrubs less common (Table 6). The density of sands age was 15.9% of that measured on random transects in the study area (Giesen 1989) although height was similar. Table 6. Vegetative County, Colorado. Lek characteristics Sandsa~e DensityA ~ hgt. at lesser prairie-chicken Bogr 10 65 25 20 5 5 0 889 111 111 0 1556 111 n/a 37.4 37.0 47.0 n/a 27.3 64.0 45 10 30 50 20 Ave. 397 42.5 22.1 in Baca Ve~etative com£osition~ Bocu Spcr Sihy Arlo Buda 2 3 5 27 28 86-1 87-4 leks 5 10 5 5 10 5 10 5 5 40 5 5 40 55 40 90 14.3 17.1 5.0 aplants(ha. bpercent of vegetative cover. cBuda - Buchloa dactyloides, Bogr - Bouteloua ~racilis, curtependu1a, Spcr - Sporobolus cryptandrus, Sihy - Sitanion Aristida lon~iseta, FoSh - forbs & shrubs. 9.3 21.4 Bocu - ~. hystrix, Arlo FoSh 10 10 15 15 5 15 10 10.7 - �Nest sites Seven habitat variables were measured at each nest site. Although complete statistical analysis of the dat~ has not been completed, height-density, heights of shrubs, forbs, and tall grasses, and sandsage density were higher at nest sites than at random sites. Percent bare ground did not differ between nests and random sites. Because nesting occurs prior to new growth of forbs and grasses, the quality of residual cover may limit nest sites. Selection of taller clumps of residual cover for nesting by lesser prairiechickens has also been reported in other areas (Copelin 1963, Suminski 1977, Davis et al. 1979, Haukos 1988). The primary factor affecting residual vegetation height on the Comanche National Grasslands appears to be livestock grazing. Grazing management to benefit lesser prairie-chickens might incluce provisions to minimize grazing beyond the growing season or to remove cattle from certain pastures when a certain height-density (3 dm) is reached. These hypotheses should be tested experimentally. Summer Range Preliminary vegetative analysis of grouse flush and random sites indicated no differences (f > 0.05) in mean values for the 7 habitat variables measured. Because the distribution of all variables were not normally distributed, it is possible that grouse are avoiding extremes of habitat variables and using most of the remaining habitats. It is also likely that the study area represents the best habitats available and that little selection (or avoidance of poor habitats) is necessary. If similar habitat preference measurements were conducted in lower quality habitat, in more patchy habitat, or at a different time of year (i.e., winter) a difference between use and random sites might occur. A complete analysis will be provided in the final report after suitable statistical tests have been conducted. Home Range Home range sizes were calculated for 9 males a~i 7 females which were observed for more than 30 days (Table 7). Although males tended to have smaller home ranges than females (0.92 ± 0.76 vs 3.75 ± 2.88 km2), differences were not significant. Because no radio-marked hen was successful in hatching a clutch of eggs they may have tended to move greater distances in summer than would hens with chicks. Hens also tended to summer alone with little apparent home range overlap whereas males were often located in small flocks and tended to remain adjacent to leks where they displayed. �Home range size of radio-marked Table 7. County, Colorado, 1989. Band 299 403 404 405 406 407 408 411 413 418 419 422 423 424 425 426 Sex Age 2+ 112+ 11111112+ 12+ 1- in Baca Time interval Home ran~e size (km+) 06 14 16 21 15 18 17 18 18 25 25 05 03 05 06 07 M M F F M F M F F M F F M M M M ? lesser prairie-chickens Apr -13 Apr-18 Apr-2l Apr-21 Apr-ls Apr-25 Apr-20 Apr- 21 Apr-21 Apr-ls Apr-25 May-09 May-21 May-26 May-2s May-11 Jul Aug Sep Sep Aug May Jun Sep Sep Aug May Jun Sep Ju1 Ju1 Jul 0.08 1. 53 2.31 6.29 1. 23 0.35 0.08 6.68 7.10 2.51 0.76 2.79 0.86 0.46 0.68 0.85 LITERATURE CITED Aldrich, J. W. 1963. Geographic Wildl. Manage. 27:529-545. Amstrup, S. C. 1980. 44:214-217. Baker, M. F. 1953. Misc. Pub1. 5. orientation A radio-collar Prairie 68pp. chickens of American for game birds. of Kansas. Tetraonidae. J. Wildl. Univ. Kansas J. Manage. Mus. Nat. Hist. Bendire, C. E. 1892. Life histories of North American birds with special reference to their breeding habits and eggs. U. S. Natl. Mus. Spec. Bull. l. 446pp. Bent, A. C. 1932. Life histories Natl. Mus. Bull. 162. 490pp. Canfield, R. H. 1941. range vegetation. of North American gallinaceous Application of the line intercept J. For. 39:338-394. Cannon, R. W., and F. L. Knopf. prairie grouse populations. method birds. in sampling 1981. Lek numbers as a trend index to J. Wildl. Manage. 45:776-778. Copelin, F. F. 1963. The lesser prairie-chicken Wild1. Conserv. Dep. Tech. Bull. 6. s8pp. in Oklahoma. U.S. Oklahoma �Crawford, J. A. 1980. Status, problems, and research needs of the lesser prairie chicken. Pages 1-7 in P. A. Vohs and F. A. Knopf, eds. Proc. Prairie Grouse Symp. Oklahoma State Vniv., Stillwater. Davis, C. A., T. Z. Riley, R. A. Smith, H. R. Suminski, and M. J. wisdom. 1979. Final report: habitat evaluation of lesser prairie chickens in eastern Chaves County, New Mexico. New ~exico State Univ. Agric. Exp. Stn., Las Cruces. l41pp. Giesen, K. M. 1989. Population inventory and habitat use by lesser pralrlechickens in southeast Colorado. Job Prog. Rep., Colorado Div. wildl. Fed. Aid Proj. w-152-R. Apr:37-49. _____ , T. J. Schoenberg, grouse in Colorado. and C. E. Braun. 1982. Methods wildl. Soc. Bull. 10:224-231. Harrington, H. D. 1964. Manual of the plants Swallow Press, Inc., Chicago. 566pp. for trapping of Colorado. Sage Books, Haukos, D. A. 1988. Reproductive ecology of lesser Thesis, Texas Tech. Univ., Lubbock. 81pp. prairie-chickens. ______ , L. M. Smith, and G. S. Broda. 1990. Spring prairie-chickens. J. Field Ornitho. 61:20-25. trapping E., and F. C. Zwickel. 1985. Influence grouse. J. Wildl. Manage. 1050-1054. Hoffman, D. M. 1963. The lesser Manage. 27:726-732. Johnsgard, P. A. 1983. Lincoln. 4l3pp. Judd, prairie The grouse chicken of the world. of radio packages in Colorado. Univ. M.S. of lesser Herzog, P. W. 1979. Effects of radio-marking on behavior, movements, survival of spruce grouse. J. wildl. Manage. 43:316-323. Hines, J. blue sage and on young J. wildl. Nebraska Press, S. D. 1905. The grouse and wild turkeys of the United States and their economic values. U.S. Dep. Agric., Biol. Surv. Bull. 24. 55pp. Loeffler, C. 1983. The status and management of the lesser prairie chicken Unpupl. Rep., Colorado Div. wildl., Colorado Springs. 9pp. in Colorado. Marks, J. S., and V. S. Marks. 1987. Influenc2 of radio-collars of sharp-tailed grouse. J. Wildl. Manage jl:468-47l. Martin, S. A., and F. L. Knopf. 1981. Aerial survey chicken leks. Wildl. Soc. Bull. 9:219-221. Mohr, C. O. 1947. Table of equivalent populations mammals. Am. Midl. Nat. 37:223-249. of greater of North on survival prairie American small �) Rash, M. T. 1985. Survey of the lesser prairie-chicken in Colorado, 3 April - 25 May 1985. Unpubl. Rep., Colorado Div. Wildl., Colorado Springs. 22pp. Robel, R. J., J. N. Briggs, J. J. Cebula, A. D. Dayton, and L. C. Hulbert. 1970. Relationships between visual obstruction measurements and weight of grassland vegetation. J. Range Manage. 23:295-297. Sands, J. L. 1978. Game bird studies. Performance Rep., Proj. W-104-R-19, New Mexico Dep. Game Albuquerque. 5pp. Sell, D. L. 1979. Spring and summer movements and habitat pra~r~e chicken females in Yoakum County, Texas. M.S. Tech. Univ., Lubbock. 41pp. and Fish Proj. use by lesser Thesis, Texas Suminski, H. R. 1977. Habitat evaluation for lesser prairie-chickens eastern Chaves County, New Mexico. M.S. Thesis, New Mexico State Las Cruces. 80pp. in Univ., Taylor, M. A., and F. S. Guthery. 1980. Fall-winter movements, ranges and habitat use of lesser prairie-chickens. J. Wildl. Manage. 44:521-524. Warner, R. E., and S. L. Etter. radiomarked hen ring-necked 47:369-375. Prepared 1983. Reproduction and survival of pheasants in Illinois. J. Wildl. Manage. by Kenneth M. Giesen Wildlife Researcher B ��71 JOB FINAL REPORT State of: Colorado Project: Work Plan: Job Title: 12 Job 15 Chronology of Breeding and Nesting Activities Relation to Timing of Spring Hunting Seasons Period Covered: Author: Upland Bird Research W-152-R of Wild Turkeys in 01 July 1985 through 31 June 1990 Richard W Hoffman Personnel: T. D. Abell, J. L. Aragon, C. E. Braun, R. W. Hoffman, R. L. Holder, B. S. Linkhart, T. B. Lundt, R. K. Mueck1er, and T. J. Spezze, Colorado Division of Wildlife ABSTRACT Gobbling, nesting, roosting, and movement activities of Merriam's wild turkeys (Meleagris gallopavo merriami) were studied in southcentra1 Colorado during spring 1986, 1988, and 1989. Mean dates for onset of incubation were 14 (1989), 18 (1986), and 21 May (1988). Peak of incubation occurred between 16 and 25 May, after the spring hunting season. Nesting success was only 8% for subadults and 22% for adults; most (92%) subadults and 38% of the adults either did not attempt to nest or lost their clutch during the laying period. Few (7%) hens attempted to renest. Average clutch size was 10.4 ± l.9(SD) eggs; 89% of the eggs in successful nests hatched. Gobbling was sporadic and differed among males. Subadult males seldom gobbled. Adult males gobbled more (~= 0.007) in morning (AM) than evening (PM), more (~= 0.003 for AM, ~ = 0.006 for PM) on than off the roost, more (~ = 0.016) in the absence than presence of hens, and more during than before (~ = 0.01) or after (~ = 0.034) the hunting season. Two peaks of gobbling were identified.. The 'second peak (11-20 May) approximated peak incubation. The gobbling and nesting data' supported a late April to late May season. Subadult males moved farther (8.7 ± 3.1 km, ~ = 0.03) from winter to breeding areas and occupied larger (12.3 ± 4.9 km2, ~ < 0.01) spring home ranges than adult males (movements = 5.3 ± 3.8 km, home range = 5.2 ± 3.3 km2). Adult (10.4 ± 5.5 km) and subadult (10.4 ± 5.9 km) females moved similar (~ = 0.91) distances. Adult females moved farther (~ < 0.01) than adult males, but subadult females did not move farther (~ = 0.61) than subadult males. The median distance between morning and evening roosting sites used on the same day was 996 m for subadu1t males and 1,073 m for adult males. The frequency with which adult males were found at previously-used spring roosts was 19%. Subadult males returned to the same roosts more often (29%, ~ = 0.14) than adult males. Departure and roosting times during the spring hunting season indicated shooting hours should start at sunrise and end one-half hour before sunset to discourage roost shooting. Based on a 100% survey conducted from 1986 to 1988, an average of 3,888 spring �72 hunters harvested 603 turkeys for a success rate of 16%. In comparison, 1,318 fall hunters harvested 315 turkeys annually for a success rate of 24%. Las Animas County was the leading harvest area. The legal harvest of bearded hens accounted for < 3% of the total spring harvest. Merriam's wild turkeys dominated wing samples and were harvested in 29 of the state's 63 counties. Rio Grande wild turkeys (M. g. intermedia) were harvested in 10 counties, including 3 counties where both subspecies were harvested. �73 RECOMMENDATIONS 1. The spring hunting season for Merriam's turkeys should open in late April and extend through late May to minimize disturbance during peak of mating, bracket peak incubation, and include the second peak of gobbling. 2. Following recommendations of Schmutz and Braun (1989), timing of the spring hunting season for Rio Grande turkeys should be from mid-April to mid-May. Because Rio Grande's primarily occur in eastern Colorado (i.e., east of 1-25) and Merriam's in western Colorado, and because most hunting of Rio Grande's is by permit only, seasons could be structured to open and close on different dates corresponding to the peak incubation periods of the 2 subspecies. 3. A statewide wing collection program should be implemented to monitor age, sex, and subspecies composition of the harvest. The feasibility of a mandatory checking program, similar to that for black bears, should be investigated. 4. Areas open to hunting of Rio Grande's should be by permit only so that harvest and hunter activity for Rio Grande's can be monitored separately from Merriam's. Hunters successful in drawing a permit should be restricted to hunting in the area where the permit is valid. A hunter successful in harvesting a Rio Grande turkey on a permit area should be allowed to pu~chase another license to hunt Merriam's turkeys in any nonpermit area. 5. A 100% survey of license buyers following the end of the spring enhanced by the implementation license agents must be directed weeks after the season ends. 6. A quantifiable method~for separating Rio Grande and Merriam's wing characteristics needs to be developed and tested. 7. The almost total lack of reproductive output by yearling hens and the apparently low nesting rate and success of adult hens requires further study. 8. Managers should avoid releasing Rio Grande turkeys into areas where they may come in contact with native populations of Merriam's turkeys. 9. Shooting hours during the spring season should start at sunrise and end one-half hour before sunset to discourage roost shooting and promote ethical hunting practices. 10. Use of gobbling counts to monitor relative abundance and trends in population growth is questionable; however, gobbling surveys can be used to detect presence or absence when measuring range expansion of introduced populations. Gobbling surveys should be conducted while the birds are still on the roost, preferably during the 45 minutes preceding sunrise. Tape recorded calls could be used to stimulate gobbling. should be conducted within 4 weeks and fall seasons. This process has been of the 2 license system. Accordingly, to return all license stubs within 2 turkeys by ��75 CHRONOLOGY OF BREEDING AND NESTING ACTIVITIES OF WILD TURKEYS IN RELATION TO TIMING OF THE SPRING HUNTING SEASON Richard W. Hoffman INTRODUCTION Wild turkeys have been legally hunted in Colorado since 1949 (Burget 1957). Fall-only hunting was permitted until 1964 when the first spring season was held: subsequent spring seasons were held in 1965, 1967, 1968, and 1970 (Colo. Div. Wi1d1. 1984). A continuous 23 to 30-day season, opening in mid-April and extending through early to mid-May, has been held every year since 1973,except in 1990, when the season was extended 2 weeks in May. Fall seasons have ranged from 9 to 16 days, opening in mid- to late September and closing in early to mid- October. In 1990, the fall season was lengthened to 30 days by opening a week earlier in September and closing a week later in October. In 1973, an estimated 1,301 hunters (spring = 496, fall = 805) harvested 265 turkeys (spring = 64, fall = 201). By 1983, 4,894 hunters (spring = 3,030, fall = 1,864) were harvesting 1,147 turkeys (spring = 645, fall = 502). The Colorado Division of Wildlife (CDOW) responded to this growing demand by focusing their management efforts on (1) restoring populations of Merriam's wild turkeys in western Colorado, and (2) establishing new populations of Rio Grande wild turkeys in eastern Colorado. The success of transplant programs plus increases in hunter participation generated a need for more refined data upon which to base management recommendations. Paramount was the need for more reliable estimates of harvest by area, subspecies, age, and sex, and for better documentation of the chronology of breeding and nesting activities. Previous studies of wild turkeys in Colorado (Burget 1957; Hoffman 1962, 1966, 1968, 1973; Myers 1973) failed to address these needs. Thus, no quantitative data existed to justify the timing and length of spring seasons and, as a result, seasons were set more on tradition than biological evidence. P. N. OBJECTIVES 1. Document timing of winter flock dispersal, onset of gobbling, peaks of gobbling, nest initiation, onset of incubation, and peak of hatch in relation to timing of the spring hunting season. 2. Describe the gonadal cycle of females and compare the reproductive condition of females in relation to timing of the spring season. 3. Measure the abandonment rate of incubating human disturbance around the nest. 4. Monitor hunter activity and harvest females to varying levels of of wild turkeys on a statewide basis. �76 STUDY AREA Trapping was confined to Longs Canyon and 2 tributary canyons, Sowbelly and Martinez, located approximately 17 km southwest of Trinidad, Colorado in Las Animas County. From here, radio-marked birds ranged over 448 km2 of surrounding areas during the breeding season. This area was bounded by 1-25 on the east, Lorencito Canyon on the west, Colorado Highway 12 on the north, and the Canadian River in New Mexico on the south. This topographically diverse area varied in elevation from 1,800 to 2,600 m and was intersected by 4 large canyons in excess of 30 km in length, each with numerous side canyons and adjacent smaller canyons. Major vegetation types included pinyon pine-juniper (Pinus edulis-Juniperus spp.), mountain shrub, and ponderosa pine (f. ponderosa). The mountain shrub type was dominated by Gambel oak (Quercus gambelii), which extended into the pinyon-juniper and ponderosa pine types. Douglas-fir (Pseudotsuga menziesii) and white fir (Abies concolor) occurred in association with ponderosa pine, primarily on north slopes. Over 95% of the area was privately owned. Human activity was minimal. Use of private lands was limited to cattle grazing, some logging, and recreation. METHODS Trapping, Marking, and Radio-tracking Turkeys were baited with oat hay and corn, and livetrapped with drop nets or cannon nets during February and March 1986, 1988, and 1989. No birds were trapped in 1987. Captured birds were classified to age and sex (Hoffman 1962), and banded with serially numbered aluminum leg bands. Numbered and color-coded (by year) Allflex livestock eartags were attached to the patagium. Ages were recorded as subadult (8-10 months) or adult (>18 months). Body weight and length of primaries, carpal, spur, and beard were measured on each bird. One hundred and forty-seven birds (16 adult males, 10 subadult males, 101 adult females, 20 subadult females) were equipped with lithium battery powered transmitters (Models HLPB 2750 and 2l20-LD, Wildlife Materials, Carbondale, IL) attached with a poncho collar (Amstrup 1980) or tail-clip (Bray and Corner 1972). The poncho radio package weighed < 40 gm, the tailclip package < 35 gm. Tracking was conducted from the ground using a 3element Yagi antenna and Telonics TR-2 receiver. All locations were verified by visual observation and recorded to nearest 50 m as Universal Transverse Mercator coordinates. Two aerial searches were conducted each year between late April and late May for birds not found during ground searches. Birds found during aerial searches were subsequently located from the ground. Nesting Activity Flocks containing radio-marked birds were monitored a minimum of 3 times/week beginning in late February to determine the period of flock dispersal. Locations of radio-marked hens following flock break-up varied depending on how far they moved from wintering to breeding areas. Birds moving longer distances were located less frequently because they required more search time to find. During May, hens were located once every 3-5 days to minimize disturbance. All nests were located after incubation had begun. Suspected nest sites were circled and flagged from >30 m away. Some nests were visually observable from this distance. Others were monitored but not approached for �77 30 days unless the radio-signal indicated the hen was gone. Nest sites were visited almost daily as the anticipated hatch date approached. Onset of incubation was estimated by backdating 28 days (Bailey and Rinell 1967) from date of hatch. Most hens were located often enough just before and during the early stages of incubation to approximate within 3 days of when they started incubating. Onset of incubation was not calculated for unsuccessful hens located on nests during later stages of incubation. Clutch size, fertility, and nest success were determined from egg shell characteristics after eggs hatched or after the nest was abandoned or depredated. Differences among years for mean dates of initiation of incubation were tested using ANOVA. Gobbling Activity Gobbling indices were conducted from 1 April to 15 June and categorized as preseason (-1-15 Apr), season (-16 Apr-IS May), and postseason (-16 May-IS Jun). Opening and closing dates of the hunting season varied by 3 days over the study period. An attempt was made to conduct 3 valid indices/week/time period (AM and PM). A gobbling index was considered valid if (1) positive identification was made of the radio-marked bird that was gobbling, (2) the bird was not disturbed before or during the index, (3) the time the bird left (AM index) or went (PM index) to roost was known, (4) the index included time on and off the roost, and (5) it was known whether the bird was alone, or associated with other males and/or females. An index lasted 1 hour from one-half hour before to one-half hour after sunrise (AM index) or sunset (PM index). The l-hour period was divided into time spent on and off the roost and whether females were present or absent. Roosting times were determined by observing or hearing the birds fly to or from the roost. Presence of hens was ascertained from sightings or calls heard during the index or by locating and observing the birds after the index. In the case of simultaneous gobbling bouts, it was assumed the radio-marked male was participating. Gobbles of subadult males were incomplete and higher pitched than gobbles of adults. Radio-marked males were monitored on a rotating basis with the initial order being randomly selected. If, for example, male A could not be located on the day it was to be monitored, then the next male (B) on the list was indexed. Priority was then given to finding male A and doing an index on male A during the time period male B was supposed to be indexed. This order was adhered to as best possible. However, some males were indexed less than others because they were more difficult to find on a regular basis. For AM indices, the male was located on the roost the evening before. Males selected for a PM index were located at least 1 hour before sunset. Gobbling data were totaled for each radio-marked male for each category of comparison (i.e., on and off the roost, hens present and hens absent, AM and PM, etc.) and converted to gobblesjhour. For example, if 200 gobbles (150 on and 50 off roost; 20 total gobbles during preseason) were recorded during 12 AM indices (720 min) on male A, 4 each during the preseason, season, and postseason totaling 252 minutes on and 468 minutes off the roost, then the preseason AM gobbling rate for male A was calculated as total gobbles recorded during the 4 preseason AM indices (20) divided by total minutes of observation (240) times 60 = 5 gobblesjhour. The on roost AM gobbling rate was calculated as total gobbles on the roost (150) divided by total minutes of observation on the roost (252) times 60 = 36 gobbles/hour. �78 The Wilcoxon signed rank test was used to test the null hypothesis of no difference in gobbling on and off the roost, and in the presence or absence of hens. The same procedure was used to compare gobbling rates during the preseason, season, and postseason, and before, during, and after the peak period of incubation. Control of the overall error rate for these comparisons was maintained by use of the Bonferroni inequality. Gobbling of adults between years (1986 and 1989) was compared using the Mann-Whitney Wilcoxon test. Roosting Activity Roosting sites of males were located in conjunction with conducting gobbling indices. Roosting times were determined by observing or hearing the birds fly to and from the roost. Distances between roost sites were calculated from consecutive locations (same day) and for locations ~ 3 days apart. A roost site was considered reused if the same bird was located within 150 m of a previous roost location. The Kruskal-Wallis test was used to test the null hypothesis of no difference in roosting times, and no difference in distances between roosting sites among males within an age class. Comparisons between age classes were made using the Mann-Whitney-Wilcoxon test. The Wilcoxon signed-rank test was used to compare median distances between roost sites for the same male during preseason, hunting season, and postseason periods. Frequency of use of previously-occupied roost sites by adults and subadults was compared with a chi-square 2 x 2 contingency table. Movements and Home Range Harmonic mean transformation (HMT) (Dixon and Chapman 1980) was used to estimate spring home range size for males that survived from 1 April to 15 June and were located ~ 20 times during this period. Home range size was calculated based on the 90% contour of area using the computer program MCPAAL (Stuwe and Blowhowiak 1986). Minimum convex polygon (MCP) (Mohr 1947) ranges were also computed for comparative purposes using the same computer program. Movements from wintering to breeding areas were measured as the minimum straight line distance between the winter trap site and the arithmetic center of the spring home range (males) or the nest site (females). Mean movements and home range sizes were compared using ~ tests. Permits, Questionnaires, Hunters were seasons: 1. required and Wing to obtain Collections 1 of 2 types of permits during spring and fall Special Unlimited Hunting Permit - Unlimited in number and free of charge, these permits were available to any holder of a valid turkey license. Whereas the license could be purchased from any license agent, permits were only available from CDOW offices either on a walk-in basis or by mail application. The special unlimited permits were valid for one season (spring or fall). Unsuccessful spring hunters could hunt in the fall without purchasing a new license, but before doing so, they were required to obtain another permit valid for the fall season. Successful spring hunters who wanted to hunt the fall season needed to purchase another license in addition to obtaining a special unlimited fall permit. By requiring spring and fall permits, hunters could be categorized as spring only, fall only, or spring-fall hunters and surveyed accordingly. �79 Special unlimited permits for the spring season were available from 1 March through the end of the season; those for the fall were available from 1 August to the end of the season. Special unlimited permits were valid for all areas not requiring a limited permit. 2. Limited Hunting Permits - Limited in number, free of charge, and available by public drawing. Only mail applications were accepted for limited permits. The drawing and issuance of permits was handled through the Denver office. Hunters could apply for a limited permit without first purchasing a license. However, if they succeeded in drawing a limited permit, they needed to purchase a license before going hunting. Drawings were held in late March and late August for the spring and fall seasons, respectively. Limited permits were only valid for the area, season, and time period indicated on the permit. Holders of a limited permit were required to obtain a special unlimited permit if they hunted outside the area for which their limited permit was valid. Consequently, some hunters obtained 2 permits and were subsequently mailed 2 harvest questionnaires. This problem was identified as the questionnaires were processed and duplicates were excluded. Questionnaires were mailed to all permit holders immediately after the spring and fall seasons. Non-respondents were mailed a followup questionnaire approximately 3 weeks later. Mean values calculated from responses to the second mailing were used to project answers for those permit holders not responding to either questionnaire. Every hunter that obtained a limited or unlimited permit was also issued a wing envelope coded by permit number. The instructions on the envelope requested each successful hunter to (1) complete the questionnaire printed on the envelope including their name, address, time of harvest, and location of harvest (county, small game management unit, and nearest town), and (2) to remove the least damaged wing and 3 or 4 breast feathers as depicted by 2 schematic diagrams, place them in the envelope, and mail the postage-paid envelope. Envelopes were addressed to the Wildlife Research Center in Fort Collins. Upon receipt, wings were frozen until processed. Information provided by hunters was transcribed onto a standardized form. The envelope's contents were examined to determine if both a wing and breast feathers were enclosed. Samples were identified to subspecies and classified to age and sex. Whenever possible, the following measurements and feather characteristics were recorded for each wing: length, condition (growing, fully grown, empty, broken, worn, pointed) and status (adult or juvenal) of primaries I through X (numbered proximal to distal), quill diameter of P IX and X at their insertion into the follicle, stage of primary molt, carpal length, and status (adult or juvenal) of the first secondary and tertials. Since the range of the Merriam's and Rio Grande wild turkey does not overlap in Colorado, except possibly along the Arkansas River west of Pueblo, the 2 subspecies were identified from wing samples based on the location of harvest. However, inspection of the wings suggested that subspecies could be distinguished by wing color. Rio Grande's tended to have wider black bars and narrower white bars on the primaries than Merriam's, thus, giving the wing a darker appearance. Sex was ascertained by presence of buffy (female) or black (male) tipping on the breast feathers (Hoffman 1962). Wings were assigned to age classes based on the shape, color, wear, and barring pattern of primaries IX and X (Hoffman �80 1962). Juvenal primaries IX and X were pointed, grayish-brown, and lacked white barring near the tip. Comparatively, adult primaries IX and X were rounded and black or blackish-brown, with white barring extending almost to the tip. Subadults possessed the outer primaries characteristic of juveniles except they were faded and worn as a result of being retained longer. Wings obtained from the spring season were classified as adults (> 22 months) or subadults (10-11 months). Those from fall were separated into 3 age classes: adults (> 26 months), subadults (14-16 months), and juveniles « 6 months). RESULTS Capture and marking Two hundred and thirty-nine turkeys (1986 = 99, 1988 = 86, 1989 = 54), including 18 adult males, 145 adult females, 50 subadult females, and 26 subadult males, were trapped in February and early March using drop nets (129), cannon nets (106), or clover traps (4); 141 were equipped with transmitters and released at the trap sites (Table 1). Six birds died as a result of trapping. Efforts to trap adult males in 1988 were unsuccessful. No adult males were observed at the bait sites and no flocks of adult males consistently remained in one area where they could be attracted to bait. Consequently, only subadult males were trapped and monitored in 1988. Table 1. Colorado, Wild turkeys trapped and equipped with radios in Longs Canyon, 1986, and 1988-89. Females Year 1986 1988 1989 Adult Males Subadult Adult 4 7 3 13 o 9 Subadult o 10 o Totals 50 52 39 aOne additional adult hen radio-marked in 1986 was monitored in 1988. bFive additional adult hens radio-marked in 1988 were monitored in 1989. In 1987, trapping operations were initiated in January and continued through mid-April. Except for a flock of 20-30 birds in Sowbelly Canyon, no other turkeys or sign were found in Longs Canyon or any of its tributary canyons. Efforts to trap birds in Sowbelly Canyon were unsuccessful as they could not be attracted to bait sites. Instead, they preferred to feed on hillsides rather than in canyon bottoms where they could be trapped. Several different baits were offered to no avail. Cattle were also a problem as they consumed the oat hay and trampled the bait site, including the cannon nets and wire leads. The lack of use of bait sites was attributed to an abundant supply of pinyon nuts. �81 Bait sites were placed at other known turkey use areas in Longs, Colorow, Saruche, Martinez, and Little Martinez canyons. No birds were observed at any of these sites. Flocks were found near Lake Dorothey and Spanish Peaks. However, because of access and landowner problems, it was not considered feasible to trap these flocks. In addition, the resulting data would not have been directly comparable with that collected in 1986. It was considered better to skip a year and try to trap at the same sites in 1988 than to trap at another location in 1987. Trapping operations were terminated on 1 April 1987, which was the latest date to trap and still meet the objectives of the study. Physical Characteristics Twenty four percent (35/145) of the adult hens and none of the subadult hens captured had beards. Average beard length for adult hens was 137 ± 49(SD) rom. The shortest beard was 35 rom and was clearly visible through the breast feathers. Comparatively, beard length for adult and subadult males averaged 219 ± 18 rom and 64 ± 17 rom, respectively. Weights of adult males did not differ (f = 0.52) between years (1986 and 1989), whereas adult females weighed less (f < 0.001) in 1986 than in 1988 or 1989 (Table 2). Adults of both sexes weighed more (f < 0.01) than their subadult counterparts. Table 2. Weights (kg) of wild turkeys trapped and radio-marked Canyon, Colorado, 1986, 1988-89. Males Females Adult Subadult Year ~ SD 1986 1988 1989 All 3.87 4.34 4.30 4.14 0.25 0.35 0.33 0.38 Recoveries in Longs ~ 3.02 3.60 3.30 3.32 Subadult Adult SD 0.46 0.26 0.29 0.45 ~ SD 7.64 0.65 7.46 7.54 0.49 0.55 ~ 5.55 SD 0.51 and Mortality Twenty-one radios, all from hens, were recovered between 1 March and 15 July 1986, and radio-contact was lost with an adult male in early May and a subadult female in mid-April. Fourteen recoveries were classified as from birds that died of natural causes, 4 were from birds that slipped their transmitters, and 3 were from birds that suffered from capture stress. Eleven of the 14 suspected natural mortalities occurred between 1 March and 10 April. Two hens died in early May prior to onset of incubation and 1 hen was killed on the nest in late May. Mortality of hens from late winter (1 Mar) through the nesting period (15 Jul) was 26%. No males died during this period. All 4 transmitters thought to have been dropped still attached to the 2 central retricies when or body parts were found near the transmitters killed, nor had the transmitters sustained any were tailmounts that were recovered. No other feathers to indicate the birds were damage (i.e., tooth marks, �82 scratches, ...etc.). It was suspected that the calamus portion of the feather was crushed when the radios were attached causing the bird to molt the damaged feathers prematurely. Three instrumented hens were recovered < 500 m from the trap site within 10 days of when they were trapped. Two had been killed (or possibly scavenged) by mammalian predators and 1 was captured alive by hand. The captured bird could not fly and fell over when attempting to run. The bird was necropsied at the Colorado State University Diagnostic Lab. It weighed 1.1 kg less (4.2 vs. 3.1 kg) than when initially captured. Gross lesions, indicative of capture myopathy (Spraker et al. 1987), were found in the thigh muscles. No lesions were found in the wing muscles. Although the 2 hens killed near the trap site appeared to be natural mortalities, they too may have suffered from capture myopathy or some other form of capture stress predisposing them to predation. Thirteen radios were recovered between 1 March and 15 July 1988, 11 from hens and 2 from males. Eleven recoveries (10 females, 1 male) were from birds that died of natural causes and 2 (1 male, 1 female) were from birds that lost their transmitters, both of which were tailmounts. Five mortalities, including the 1 male, occurred between 7 and 27 March, 2 between 1 and 17 May, and 4 between 1 and 4 June. Only 1 hen was killed on the nest. Mortality during the monitoring period was 24% for hens and 10% for males. Capture myopathy was not a mortality factor in 1988. Radios were recovered from 8 hens and 2 males in 1989. Five birds (4 females, 1 male) died of natural causes, 1 bearded hen was legally shot, 1 unbearded hen was illegally shot, and 1 male lost its tail-mounted transmitter. The remains of 2 other radio-marked hens were found together « 50 m apart) near the trap site 6 days after they were released. The primary cause of death was classified as capture stress. Two natural mortalities (including the male) occurred in late March, 1 in early April, and 2 (nesting hens) in early June. Only 11% of the radio-marked males (1/9) and females (4/35) died of natural causes between 1 March and 15 July. Nesting In 1986, 26 hens survived into the nesting season and only 5 (19%) nested successfully. Ten attempted to nest but failed, including 3 hens that abandoned their nests after being disturbed. Excluding these hens, nesting success was 22%. Two hens were killed during the laying period and 1 hen was killed on the nest. Eight hens either did not attempt to nest or lost or abandoned their clutches prior to onset of incubation. One renesting attempt was documented. Eleven (30%) of 37 hens survLvLng into the 1988 nesting season nested successfully, whereas 11 attempted to nest but failed, 4 were suspected of laying when predated, 1 was killed on the nest, and 10 either did not nest or lost their clutch during laying. Two hens that lost their transmitters during laying were not included in the sample. Only 1 hen attempted to renest. Thirty-one hens survived into the 1989 nesting season, 2 of which were shot, thus, reducing the sample to 29 hens. Only 2 (7%) nested successfully and 9 nested unsuccessfully; none attempted to renest. One hen was killed prior to incubation and 2 were killed on the nest during incubation. Fifteen hens lost their clutches during laying or did not attempt to nest. �83 Of 20 radio-marked subadult hens, 12 survived into the nesting season, but only 1 (8%) attempted to nest and did so successfully. There was no indication the others attempted to nest based on their movement patterns; i.e., they did not localize. In comparison, 50 (62%) of 80 adult hens surviving into the nesting season attempted to nest and 18 (22%) were successful. Clutch size of first nest attempts was ascertained for 7 nests in 1986 [~ 11.7 ± 1.2(SD) eggs], 8 nests in 1988 (9.3 ± 1.9 eggs), and 2 nests of 9 and 11 eggs in 1989. The smallest clutch (ll = 6 eggs) resulted from the only documented nesting attempt of a subadult. The largest clutch was 14 eggs. Clutch sizes of the 2 renesting attempts were 6 and 12 eggs. Combined clutch size of first nest attempts over the 3-year period was 10.4 ± 1.9 eggs; 134 (89%) of 151 eggs in successful nests were known to hatch. No efforts were made to collect hens or to measure nest abandonment rates of incubating females in response to human disturbances because of the small sample of hens that nested successfully each year. Instead, priority was given to documenting hatching dates and onset of incubation. Mean date for onset of incubation differed marginally (f = 0.103) among years, being later in 1989 (14 May, II = 12) than 1986 (18 May, II = 14) and 1988 (21 May, II = 22). Earliest and latest dates for initiation of incubation of first nest attempts were 6 May and 8 June, respectively. Fifty-six percent (27/48) of the hens started incubation after the spring hunting season; another 35% (17/48) started the last week of the hunting season. The peak period for onset of incubation was 16-25 May. Gobbling Adults.--Two hundred and three valid indices were obtained on 12 different males in 1986 (99 indices on 7 males) and 1989 (104 indices on 5 males). Gobbling rates did not differ between years for AM (f = 0.687) or PM (f 0.591) comparisons. Gobbling was first heard on 11 March and continued through 15 June when the gobbling indices were terminated. Radio-marked males were heard gobbling 2,830 times during 120 of 133 (90%) morning indices and 384 times during 43 of 70 (61%) evening indices. On a daily basis, gobbling was extremely sporadic. Even during peaks of gobbling under ideal conditions, there were indices when no gobbling was heard. Some males gobbled consistently more than others. The typical pattern for an adult male was to gobble more (f = 0.007) in the AM than PM, more on than off the roost for both AM (f = 0.003) and PM (f = 0.006) comparisons, more (f = 0.016) in the absence than presence of hens, and more during than before (f = 0.01) or after (f = 0.034) the hunting season (Table 3). Gobbling did not differ (f = 0.60) between preincubation and incubation periods, but occurred less frequently during postincubation than during either preincubation (f = 0.015) or incubation (f = 0.012). Two distinct peaks of gobbling were evident in 1986 (Fig. 1). The peaks were less pronounced in 1989, although the second peak in 1989 occurred at the same time as in 1986. Both second peaks of gobbling approximated the peak of incubation and occurred after the hunting season. Subadu1ts.--Eighty-five valid indices were conducted on 8 different subadult males in 1988. Subadults were first heard gobbling on 15 April and last heard on 3 June. Only 62 gobbles were recorded between 1 April and 15 June, �84 50 PEAK ONSET OF INCUBATION a:: J: <, 40 en _. w OJ OJ 30 0 (!J :E <i. 1989 ---------20 Z <t C 10 W :E o 1-10 11-20 21-30 1-10 Chronologie distribution wild turkeys. 21-30 MAY APRIL Fig. 1. Merriam's 11-20 of gobbling activity 1-10 JUNE of adult male �85 including 60 in the morning and 2 in the evening. No gobbling was recorded during 37 of 50 (74%) morning and 33 of 35 (94%) evening indices. The number of AM gobbles per index (n = 13) when subadults did gobble ranged from 1 to 27 (median = 1 gobblejhr). Subadults gobbled at a slightly higher (f = 0.076) rate on than off the roost (Table 3). Only 2 AM gobbles were heard during 16 indices when hens were absent and 58 during 34 indices when hens were present. Seventy-seven percent (46/60) of all AM gobbling occurred in the absence of adult males. Table 3. Merriam's Gobbling rate wild turkeys. (gobblesjhr) Subadults Median Category Time of day AM PM Roosting status AM on AM off PM on PM off Hensb Present Absent Timing of incubationb Preincubation Incubation Postincubation Timing of hunting Preseason Season Postseason of adult (n 12) and subadult Range (n 8) Adults Median Range 50 35 1.0 0.1 0.4-2.5 0.0-0.2 133 70 23.9 6.3 2.1- 44.8 0.0- 24.7 1,309 1,691 1,012 1,088 1.3 0.9 0.0 0.0 0.6-2.6 0.3-2.5 0.0-0.1 0.0-0.1 3,153 4,827 1,586 2,614 38.7 10.7 12.3 0.4 3.9-114.8 0.8- 29.3 0.0- 43.2 0.0- 23.8 34 16 1.3 0.0 0.4-4.0 0.0-0.7 53 46 12.0 29.2 2.8- 34.0 1.4 -56.0 27 18 0.2 2.0 0.1-5.6 0.0-2.8 C C 74 38 21 29.7 23.2 2.0 1.4- 53.7 0.7- 88.0 0.0- 30.0 0.3 0.8 1.7 0.0-0.6 0.4-4.5 0.0-3.3 38 50 45 ll.5 34.5 10.0 0.0- 52.1 2.0- 77.0 0.0- 38.6 5 seasonb 18 21 11 aTotal indices conducted except for roosting activity, which is expressed as total minutes of observation (i.e., AM min. on + AM min. off + 60 = total AM indices). bBased on AM indices only. CInadequate sample of indices per bird to compute a median and range. Subadults appeared to gobble more during incubation than preincubation (Table 3); however, because of variation in gobbling among individuals, the difference was not significant (f = 0.69). Four of 8 subadult males gobbled more during the preincubation period and 4 gobbled more during incubation. Only 5 indices were conducted during postincubation, precluding comparisons with other periods. Gobbling also appeared to increase progressively from the preseason, through the season, and into the postseason (Table 3), but again, �86 because of individual variation, no comparisons were significant (£ = 0.076, 0.222, and 0.688 for preseason vs. season, preseason vs. postseason, and season vs. postseason, respectively). Four males gobbled most during the postseason. The other 4 males did not gobble in the postseason, but instead gobbled most in the preseason (3) and season (1). Movements and Home Range Movements of 71 birds were documented from winter ranges to spring breeding areas including 11 adult and 8 subadult males, and 51 females (50 adults, 1 subadult) located on nests. Movements from the winter trap site to the area occupied in mid- to late May (nesting period) were documented for an additional 33 hens (22 adults, 11 subadults) not located on nests. Distances moved by females located on nests were similar (£ = 0.27) among years and did not differ (£ - 0.42) from distances traveled by hens not located on nests; consequently, data for females for all years were pooled. In addition. no differences (£ = 0.15) were evident in distances moved by adult males between 1986 and 1989. The data indicated that (1) subadult males moved farther (£ < 0.01) than adult males, (2) there was no difference (f = 0.91) in distances traveled by adult and subadult females, (3) adult females moved farther (f < 0.01) than adult males, and (4) subadult males and females moved similar (f = 0.61) distances (Fig. 2). Home range size locations/male) 8 subadult males were not located sizes. estimates were calculated of 11 adult males and 289 that were monitored from often enough during this from 278 locations (20-28 locations (20-31 locations/male) of 1 April through 15 June. Females time period to calculate home range The HMT spring range size of adult males averaged 5.2 ± 3.3 km2 and did not differ (f = 0.22) between years. Subadults occupied larger (12.3 ± 4.9 km2, £ < 0.01) HMT spring home ranges than adults. Minimum convex polygon home ranges for both age classes were substantially larger (f < 0.01) than HMT estimates, averaging 13.9 ± 8.2 km2 for adult males and 28.7 ± 13.8 km2 for subadult males. All 6 adult males monitored in 1986 shared a portion of their spring home range with at least 1 and up to 4 other radio-marked males. In 1989, 3 of 5 adult males shared spring home ranges, and in 1988, 7 of 8 subadults had overlapping home ranges with 1-3 other radio-marked subadults. Spring and winter home ranges overlapped for 5 of 11 adult males and 3 of 8 subadult males. Three adult males with distinct winter and spring home ranges left the winter range the first week of April; the other 3 adults and 5 subadults did not leave until the third week of April in conjunction with peak departure of hens. For those males with overlapping winter and spring home ranges, it was not clear when in April they moved onto their spring range. Roosting The median distance between morning and evening roosts used on the same day was 996 m (n = 44, range = 0-6,900 m) for subadults and 1,074 m (n = 25, range = 50-2,507 m) for adults. The median distance between roost sites located ~ 3 days apart, but not on the same day, ranged from 798-1,979 m for adult males and 859-3,389 m for subadult males. Differences (£ = 0.04) were apparent �87 25 20 (/) a: 15 W I- w _.0 ~ (72) 10 (12) (8) ~ 5 ( 11) o ADULT SUB ADULT FEMALES SUB ADULT ADULT MALES Fig. 2. Distances traveled by Merriam's wild turkeys from wintering to breeding areas. Vertical line = range, horizontal line = mean, vertical bar = standard deviation. Sample sizes are in parentheses. �88 among males within an age class. Median distances between roost sites increased during the hunting season compared to preseason (f = 0.03 for adults and subadults) and postseason (f ~ 0.03 for adults, f = 0.08 for subadults) periods, but were similar (f = 0.35 for adults, f = 0.71 for subadults) during the pre- and postseason (Table 4). Table 4. Median distance (m) between roosting sites of male Merriam's turkeys in relation to the hunting season.a Periodb n Preseason Season Postseason 39 59 53 Adult Median 1,137 1,765 1,165 Range n 50-1,371 917-2,961 800-1,597 51 50 21 Subadult Median 869 1,823 1,068 wild Range 761-2,055 925-5,162 600-2,055 aBased on roost sites located ~3 days apart but not on the same day. bpreseason -1-15 April, hunting season -16 April to 15 May, postseason -16 May to 15 June. The frequency with which adult males were found at previously-used roost sites was 19%. Subadults returned to the same roosts more often (29%) than adults, although the difference was marginal (f = 0.14). Whereas all subadult males reused at least 1 roost site, 2 adult males did not. No roost site was used more than 4 times and seldom (adults = 4%, subadults = 9%) was a roost site used on the same day; i.e., morning and evening. However, subadults (median 7 days, range = 0-24) returned to previously-occupied roosts sooner (f < 0.01) than adults (median = 12 days, range = 1-78). Departure and roosting times were recorded for males during 183 (adults = 133, subadults = 50) morning gobbling indices and 105 (adults = 70, subadults = 35) evening indices. Departure and roosting times did not differ (f = 0.35 and 0.45 for adults, f = 0.24 and 0.30 for subadults, respectively) among males within an age class. There was approximately a 20-minute period when birds went to roost or left the roost (Table 5). Extreme variations in roosting times were usually associated with overcast conditions accompanied by snow or rain. The tendency was for males to leave roosts later in relation to sunrise as spring progressed (Table 5). Adult and subadult males left the roost 11-13 minutes before sunrise during the preseason and 5-7 minutes after sunrise during the postseason. However, during the hunting season adults continued to leave the roost before sunrise, about 9 minutes earlier (f = 0.04) than subadults which departed at or just after sunrise. �89 Evening observations indicated males roosted earlier in relation to sunset as spring period progressed (Table 5). Preseason roosting times of adults and subadults occurred 6-7 minutes after sunset. During the hunting season, subadults roosted just before sunset and adults slightly after sunset. By the postseason period, adults and subadults were flying to roost before sunset. Table 5. .Spring departure <minutes before or after sunrise) and roosting (minutes before or after sunset) times of male Merriam's wild turkeys in relation to the hunting season. Subadult Adult Oe~arture Roost Perioda n E so Preseason Season Postseason 26 46 37 +10.6 + 8.2 - 5.3 10.8 8.4 8.9 n 20 26 19 E +7.2 +0.5 -7.5 Roost Oe~arture so 8.4 8.2 10.9 E n 18 20 12 +12.7 - 0.7 - 6.5 so n ~ SO 7.1 7.4 7.0 11 13 11 + 5.9 - 3.2 -13.5 8.0 9.2 5.6 apreseason -1 to 15 April, season -16 April to 15 May, postseason -16 May to 15 June. bpositive departure values indicate minutes off the roost before sunrise and negative values minutes off the roost after sunrise. CPositive roost values indicate minutes on the roost after sunset and negative values minutes on the roost before sunset. Hunter Compliance - Permit System Hunter compliance with the spring permit requirement over the 3-year period from 1986 to 1988 was 93% (Table 6). Questionnaires were sent to all permit holders by 1 June with a followup questionnaire mailed to all non-respondents 3 weeks later. The combined return rates (excluding non-deliverable surveys) for the first and second mailings were similar (-73%) among years (Table 7). Table 6. Year 1986 1987 1988 Hunter compliance Licenses 4,788 4,607 5,078 sold with permit requirement, Permits issued 4,698 4.257 4,569 spring season, 1986-88. Compliance (%) 98 92 90 There were more permits issued for the fall season than there were licenses sold. The reason was that unsuccessful spring hunters who wanted to hunt the fall season were not required to purchase another license, but were required to obtain a fall permit. Consequently, total license sales and total permits issued were not directly comparable. Hunter compliance was therefore measured as the proportion of new license buyers who also obtained a permit. The estimates were 94% (759 of 809) in 1986, 87% (776 of 892) in 1987, and 88% (782 of 889) in 1989. Questionnaires were sent to the fall permit holders by 20 October. Approximately 75% of the fall hunters responded to the questionnaire (Table 7). �90 Table 7. Response to spring and fall wild turkey harvest S:Qring 1987 1986 Surveys mailed Surveys returned Non-deliverable Percent returnb 4,698 3,365 119 73 4,257 2,999 121 73 surveys, 1986-88.a 1988 1986 Fall 1987 1988 4,569 3,267 125 74 1,819 1,328 45 75 1,751 1,253 56 74 1,599 1,163 40 75 aCombined responses for first and second mailings. bExcluding non-deliverable surveys. Hunter Activity and Harvest - Questionnaire Projected estimates for the spring season indicated an average of 3,888 hunters harvested 603 turkeys for a success rate of 16% (Table 8). Average total harvest, including a reported crippling loss of 17%, was estimated at 728 turkeys per year. Spring hunters averaged 3.5 days afield over the course of the 30-day season. About 66% fewer hunters participated in the fall season (i - 1,318 hunters) and 48% fewer birds (i = 355 turkeys) were harvested (Table 9). However, hunter success (24%) was higher and crippling loss (15%) was slightly lower. Fall hunters spent 2.7 days afield or about one day less than spring hunters, but fall hunters had only 16 days of hunting which included just 3 weekends. The spring season lasted 14 days longer, allowing for 2 additional weekends of hunting opportunity. Table 8. Spring turkey harvest Descriptive statistic N in sample N hunters % hunters N hunters observing turkeys % hunters observing turkeys a N successful hunters (harvest) % successful huntersa N hunters days Days/huntera Crippling loss % crippling loss Total harvest aBased only on those hunters and hunter activity, 1986 4,698 4,174 89 2,612 63 563 13 14,996 3.6 138 20 701 who actually hunted. 1986-88. 1987 4,257 3,670 86 2,367 64 575 16 13,215 3.6 138 19 713 1988 4,569 3,819 84 2,687 70 671 18 12,849 3.4 100 13 771 �91 Table 9. Fall turkey harvest and hunter activity, Descriptive statistic li in sample li hunters % hunters li hunters observing turkeys % hunters observing turkeys a li successful hunters (harvest) % successful huntersa li hunters days Daysfhuntera Crippling loss % crippling loss Total harvest 1986-88. 1986 1987 1988 1,819 1,414 78 768 54 313 22 3,603 2.5 47 14 360 1,751 1,314 75 715 54 278 21 3,704 2.8 60 18 338 1,599 1,225 77 677 55 355 29 3,378 2.8 60 14 415 aBased only on those hunters who actually hunted. Most hunting pressure and harvest occurred on opening weekend during both seasons (Tables 10, 11). Pressure and harvest did not change substantially over the remainder of the fall season. There was an increase in pressure on weekends during the spring season. Counties and Game Management Units in the Southeast Region accounted for over 70% of the spring and fall harvest (Tables 12, 13). Las Animas County was the leading harvest area with 23 and 30% of the spring and fall harvest, respectively. Public land supported about 54% of the spring and fall hunting pressure, but produced only 43% (spring) and 35% (fall) of the harvest (Tables 14, 15). Over 70% of the birds harvested during spring were taken before noon compared to 60% for the fall season; the primary harvest periods (Mountain Daylight Time) were 0600-0900 hours for spring and 0700-1000 hours for fall (Table 16). Hunter success on spring limited permit areas generally exceeded the statewide success rate of 16% (Table 17). It was not as advantageous in terms of success to hunt on a limited permit area during fall (Table 18). It was a definite advantage to hunt on private (spring = 26% success, fall = 43% success) vs. public land (spring = 12% success, fall = 17% success) in both seasons. Wing Collection Program Compliance.--Of the 1,440 successful spring hunters who responded to the questionnaire, 1,048 (73%) said they returned a wing. However, 1,266 wings were processed, meaning 218 successful hunters who did not respond to the questionnaire still returned a wing. The wing collection program sampled 70% of the estimated spring harvest. A total of 673 wings was collected from the fall wing survey representing 71% of the estimated fall harvest. Seventy-eight percent (579 of 739) of the successful hunters responding to the questionnaire indicated they returned a wing and 94 successful hunters not responding also returned a wing. �92 Table 10. 1986-88.a Chronological distribution of hunting pressure and harvest during the spring turkey season, 1986 Dateb Hunter da~s x !! 1987 Harvest % !! 1988 Harvest x !! Hunter da~s % .!! Harvest Hunter da~s % !! x .!! WE WK WE WK WE WK WE WK WE 3,354 1,215 1,652 650 1,201 550 884 352 619 32 12 16 6 12 5 8 3 6 1n 61 34 26 19 20 23 10 4 47 17 9 7 5 5 6 3 1 2,345 1,117 1,529 657 1,043 571 782 509 814 25 12 16 7 11 6 8 6 9 134 55 72 36 28 31 26 24 17 32 13 17 8 7 7 6 6 4 1,880 979 1,094 772 1,388 812 1,197 572 969 20 10 11 8 14 8 13 6 10 178 60 70 37 48 30 28 27 26 35 12 14 7 10 6 6 5 5 Totals 10,477 100 369 100 9,367 100 423 100 9,663 100 504 100 1st 1st 2nd 2nd 3rd 3rd 4th 4th 5th aBased on hunter days and harvest that could be assigned to specific time periods. bWE = weekend, WK = week (Mon-Fri). Table 11. 88.a Chronological distribution of hunting pressure and harvest during the fall turkey season, 1986- 1986 Dateb 1st 1st 2nd 2nd 3rd Hunter da~s % .!! 1988 1987 Harvest .!! % Hunter da~s % .!! Harvest % .!! Hunter da~s x .!! Harvest .!! x WE WK WE WK WE 1,018 398 547 285 424 38 15 20 11 16 107 35 36 24 34 45 15 15 10 15 883 459 539 284 331 35 19 22 11 13 94 30 33 22 28 45 14 16 11 14 927 400 467 322 242 39 17 20 14 10 109 42 47 36 49 38 15 17 13 17 Totals 2,672 100 236 100 2,496 100 207 100 2,358 100 283 100 aBased on hunter days and harvest that could be assigned to specific time periods. bWE = weekend, WK = week (Mon-Fri). �93 Table 12. ~ild turkey harvest by county, spring and fall 1986-88. 1986 County Las Animas Fremont Huerfano Custer Pueblo Dolores ]a MontezlJIIa Logan a Morgan ] ~ashington YlJIIa ]a Kit Carson Lincoln Mesa El Paso Larimer Baca Archuleta Teller Jefferson Boulder Bent Chaffee Otero Douglas Costi lla Park Garfield Clear Creek Delta Gilpin Adams ~eld La Plata Montrose S(2ring 1987 1988 t! x t! 93 47 38 50 31 23 12 10 13 8 106 49 43 35 31 26 25 11 10 8 7 6 17 4 15 20 5 11 5 12 13 7 7 7 4 4 6 6 3 1 3 3 2 2 2 1 1 2 2 2 1 1 1 1 6 2 1 2 2 1 1 1 1 2 Totals 397 Unknown 36 < < < < < < < < < 1 1 1 1 1 100 x 1986 Fall 1987 1988 x t! x t! x t! x 110 51 41 35 50 39 21 10 8 7 10 8 71 29 21 25 27 31 13 9 11 12 56 29 18 10 24 29 15 10 5 13 81 39 21 22 29 29 14 7 8 10 4 23 4 8 3 11 3 19 4 6 3 8 4 12 4 24 15 11 7 10 6 3 5 3 3 1 8 1 4 6 4 3 2 2 1 1 1 1 1 1 2 1 1 31 18 10 17 12 11 6 3 2 3 2 2 1 1 1 1 1 1 1 1 2 1 1 1 1 1 1 10 3 10 2 6 5 2 1 3 4 1 4 1 3 2 1 1 1 6 2 2 7 10 4 1 < 3 1 1 4 5 2 1 11 8 8 6 7 3 1 4 3 3 2 2 1 1 < 1 < 4 1 2 100 230 1 2 < < < < < < < t! 1 1 5 425 22 100 6 1 4 4 < 6 2 4 3 8 2 2 2 1 1 1 514 7 < < < < < < < < 2 4 1 2 7 1 4 < 9 aHarvest was pooled by area which included more than one county. 4 3 5 3 2 2 1 2 1 100 190 21 100 < 278 9 1 1 2 1 1 < 1 < 1 < 1 100 �94 Table 13. Wild turkey harvest by Game Management Unit (GMU), spring and fall, 1986-88. Fall Sl2ring Sering 1986 Fall 1986 !! % !! % !! x !! x 111 39 8 11 4 6 9 8 48 17 4 5 2 3 4 4 84 85 140 143 64 52 60 19 26 15 12 14 4 6 73 54 43 31 38 14 11 8 6 7 33 32 27 20 17 16 14 10 47 41 27 19 17 15 10 7 13 15 20 3 4 5 26 23 23 5 4 4 5 3 13 7 8 8 19 3 3 7 4 2 11 3 19 4 8 4 12 4 6 3 9 17 8 5 9 6 2 4 2 1 2 1 4 3 3 2 9 2 2 1 5 2 1 1 2 1 1 1 1 3 3 2 2 2 2 2 2 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 < 1 < 1 < 1 < 1 < 1 < 1 < 1 < 1 < 1 < 1 8 8 4 1 7 2 1 1 1 18 13 12 12 9 10 8 8 8 6 5 5 5 4 4 4 4 4 4 4 4 3 3 3 3 3 3 3 3 2 2 2 1 1 1 1 1 1 1 4 2 1 2 1 1 22 3 8 9 2 4 2 4 1 2 4 8 1 3 3 1 1 1 1 1 1 1 1 2 6 1 4 1 514 100 X !! 84 80 68 157 37 34 25 17 16 16 11 11 11 9 8 8 6 6 6 5 3 3 3 3 2 2 2 1 1 1 1 39 9 8 6 4 4 4 3 3 3 2 2 2 1 1 1 1 < 1 Totals 405 Unknown 28 < 1 < 1 < 1 < 1 < 1 < 1 < 1 < 1 < 1 < 1 100 1988 GMU !! 79 1987 X CHUa 81 42 70 32 36 50 44 88 78 6 58 82 56 10 30 52 76 60 83 86 90 62 64 48 1988 1987 7 1 1 2 1 4 4 2 2 230 9 < < < < 3 1 1 1 1 2 2 1 1 100 71 711 421 96 59 103 109 107 69 581 851 86 511 58 42 20 123 144 136 78 135 130 83 771 38 118 147 140 40 77 19 51 501 56 41 142 39 145 128 46 95 99 461 57 751 146 137 141 125 52 30 l" r 2 5 5 8 8 8 5 4 4 3 1 1 5 1 1 2 1 1 1 425 22 < < < < < < < < < < < < < < < 1 1 1 2 2 2 1 1 1 1 1 1 1 1 1 1 1 1 1 100 7 < < < < 1 2 1 1 1 < 1 < 1 < 1 < < 1 1 < 2 1 10 2 6 2 1 3 196 15 aGame management unit numbers and boundaries were changed in 1987. bHarvest was pooled by area which included more than one game management unit. < < 1 5 1 3 1 1 2 3 4 3 3 1 2 100 278 9 100 �95 Table 14. Distribution of hunting pressure and harvest by land status, spring, 1986-88. 1986 Land status Hunters .!l. x 1988 1987 Harvest % .!l. Hunters % .!l. Harvest % .!l. Hunters % .!l. Harvest x .!l. Public Private Both 1,593 953 417 54 32 14 161 194 45 55 1,395 794 370 55 31 14 164 224 42 58 1,521 1,026 288 54 36 10 219 302 45 58 Totals 2,963 100 355 100 2,559 100 388 100 2,835 100 521 100 Table 15. Distribution of hunting pressure and harvest by land status, fall, 1986-88. 1986 Hunters .!l. x .!! Public Private Both 541 410 99 52 39 9 1,050 100 Totals Harvest - 0559 0659 0759 0859 0959 1059 1159 1259 1359 1459 1559 1659 1759 1859 1959 Harvest .!! % .!! x .!! % 54 37 9 74 134 36 64 438 327 50 54 40 6 94 183 34 66 100 208 100 815 100 277 100 % 81 146 36 64 491 336 84 227 100 911 of the wild turkey harvest SI1ring 1987 Hunters Harvest .!! 1986 0500 0600 0700 0800 0900 1000 1100 1200 1300 1400 1500 1600 1700 1800 1900 Hunters % Table 16. Distribution fall seasons, 1986-88. Time pe r i.od" 1988 1987 land status spring and by time period, Fall 1987 1986 1988 N % N % N % N % N % 9 81 60 57 26 27 14 16 7 5 11 14 25 28 6 2 21 15 15 7 7 4 4 2 1 3 4 6 7 2 5 100 63 49 31 19 19 6 7 9 11 10 27 11 5 1 27 17 13 8 5 5 2 2 3 3 3 7 3 1 7 88 107 64 46 33 18 10 18 9 10 21 20 29 15 1 18 22 13 9 6 4 2 4 2 2 4 4 6 3 0 11 38 34 29 18 15 8 10 9 17 15 28 7 0 0 5 16 14 12 8 6 3 4 4 7 6 12 3 0 0 10 31 33 29 12 9 3 3 1 10 11 14 10 0 0 6 17 18 16 7 5 2 2 1 6 6 8 6 0 aMountain daylight time. 1988 N % 0 16 34 22 32 24 12 10 7 6 10 25 27 22 0 0 6 14 9 13 10 5 4 3 2 4 10 11 9 0 �96 Table 17. Hunter success on limited permit areas during spring wild turkey seasons, 1986·88. Permit area .t! Lake Dorothey Beaver Skagway Units 103 & 109 Unit 107 Unit 96 Colo. Spgs. SIIA Unit 42 75 20 40 20 50 1986 Harv 13 2 16 9 17 Succ (%) .t! 1987 Harv Succ .t! 1988 Harv Succ 17 10 40 45 36 75 20 40 20 70 5 16 4 11 1 23 0 21 20 27 5 33 0 75 30 40 20 70 5 10 13 3 15 6 22 1 6 17 10 37 30 31 20 60 aNl.Illber permi ts iSSUed. Table 18. Hunter success on limited permit areas during fall wild turkey seasons, 1986-88. Permit area Lake Dorothey Beaver Skagway Units 103 & 109 Unit 96 Spanish Peaks .t! 1986 Harv Succ (%) .t! 1987 Harv Succ .t! 1988 Harv Succ 65 30 20 20 81 8 8 12 7 23 12 27 60 35 28 75 30 25 15 4 8 20 13 32 75 30 25 7 7 8 9 23 32 90 11 12 90 5 6 aNt.mberpermi ts iSSUed. Subspecies Composition.--The Merriam's wild turkey was the dominant subspecies identified from inspection of wings. It comprised 91 and 96% of the spring (Table 19) and fall (Table 20) samples, respectively, and was harvested in 29 of the state's 63 counties. Rio Grande's were taken in 10 counties: Fremont, EI Paso, Pueblo, Logan, Washington, Morgan, Yuma, Kit Carson, Lincoln, and Weld. Both subspecies were harvested in Fremont, EI Paso, and Pueblo counties. Age and Sex Composition.--Seventy-three percent of the spring harvest (males only) of Merriam's wild turkeys were adults compared to 50% for Rio Grande's (Table 19). Twenty-two (2%) wings examined from the spring sample were from females. All were Merriam's, most (91%) were adults, and all were assumed to be from legally harvested hens (i.e., bearded hens). However, the 2 subadults may have been illegally harvested as subadu1t hens seldom have visible beards. Females (Merriam's only) comprised 32 (excluding juveniles) and 53% (including juveniles) of the fall sample (Table 20). Samples « 20 wings/year) of Rio Grande wings were inadequate for meaningful interpretation. Sex ratios of fall-harvested adults favored females (1.4:1), whereas the sex ratio of juveniles favored males (1.3:1). Sex ratios of subadults, which were poorly represented (4%) in the fall harvest, were strongly skewed (13.5:1) towards females. Production, based on percent juveniles in the harvest (48%) and the ratio of juveniles to adult and subadult females (1.5:1), was only fair (1986 and 1987) to poor (1988). �97 Table 19. Sex, age, and subspecies composition harvest based on wing analyses, 1986-88. Females Adult Subadult % % N N Males Subadult % N Subspeciesa Year 1986 1987 1988 1986-88 Adult % N of the spring wild turkey M RG M & RG 80 20 100 23 57 26 264 15 279 76 43 73 0 0 0 M RG M & RG 60 14 74 18 41 20 269 20 289 79 59 77 2 0 2 M RG M & RG 147 24 171 33 50 34 294 24 318 66 50 65 0 0 0 M RG M & RG 287 58 345 25 50 27 827 59 886 73 50 71 2 0 2 1 1 < 1 < 1 5 0 5 1 8 0 8 2 7 0 7 1 20 0 20 2 Total 349 35 384 1 339 34 373 2 448 48 496 1 1,136b 117 1,253 2 aM = Merriam's, RG = Rio Grande. bThirteen (1986 = 6, 1987 = 4, 1988 = 3) additional Merriam's wings were processed but could not be classified to age or sex, bringing the total wings to 1,149. Table 20. 1986-88.a Age and sex composition of the fall harvest of Merriam's wild turkey based on wing analyses, Adults Females Males Year 1986 1987 1988 1986-88 .!i % .!i 38 32 56 126 42 41 42 42 52 46 76 174 % Totals .!i Males % .!i 58 90 40 59 78 45 58 132 57 58 300 48 2 0 0 2 Subadults Females .!i % 18 6 0 7 0 9 7 25 82 100 100 93 % Totals !i. % 11 7 9 27 aExcludes Rio Grande's due to small sample sizes (1986 5 4 4 4 = Juveniles Males Females % !i. % !i. 75 40 52 167 7, 1987 61 46 58 56 = 47 47 37 131 39 54 42 44 8, 1989 = Totals Sample % size .!i 122 87 89 298 19). 55 51 (39) (48) 223 172 230 625 Poults/ hen 2.0 1.6 1.0 1.5 �98 DISCUSSION Nesting The average clutch size for Merriam's wild turkeys in Colorado was similar to average clutch sizes of 9.3, 10.9, and 9.9 eggs reported from New Mexico (Lockwood and Sutcliffe 1985), Oregon (Lutz and Crawford 1987E), and Washington (Mackey 1982), respectively. None of the unsuccessful hens in the Oregon study attempted to renest. The renesting rate in New Mexico was 29%, with only adult hens attempting to renest. Estimates of hatching success ranged from 87% in New Mexico to 96% in Washington. Yearling Merriam's wild turkeys apparently have a lower propensity to nest than adults. For example, 75% of the adults and only 8% of the yearling hens monitored in New Mexico attempted to nest (Lockwood and Sutcliffe 1985). Nesting rates of yearling hens in South Dakota (Wertz and Flake 1988) and Oregon (Lutz and Crawford 1987E) were 0 and 25%, respectively, compared to 42 (South Dakota) and 75% (Oregon) for adults. The nesting rates in this study were 8% for yearlings and 62% for adults. However, hens classified as nonnesters (especially adults) may have attempted to nest, but lost their clutches during laying. The frequency of non-breeding by yearlings, and possibly some adults, merits further study. Average nesting success of adults in this study was lower than the 37% success reported by Lockwood and Sutcliffe (1985) in New Mexico and the 75% success reported for an introduced population in Oregon (Lutz and Crawford 1987E). In addition, nesting rates (97%), renest attempts (23%), percent yearlings nesting (95%), and nesting success (58%) were all higher in an expanding Rio Grande population in northeast Colorado (Schmutz and Braun 1989) than in the Merriam's population studied in southcentra1 Colorado. These data lend support to the hypothesis that reproductive performance in stabilized turkey populations is lower than that in expanding populations (Vangilder et al. 1987, Porter 1978). No density estimates were available to ascertain whether the Merriam's population studied was being maintained at such low reproductive rates. High density (20-30 birds/km2 of timber) populations were maintained in Missouri when nest success averaged 35-40% (Vangilder et al 1987) and annual hen survival rates averaged 44% (Kurzejeski et al. 1987). Lockwood and Sutcliffe (1985) estimated the median incubation date for Merriam's wild turkeys in southeastern New Mexico to be 20 May. In central Arizona, Scott and Boeker (1972) reported the peak of hatch occurred around 15 June; backdating 28 days places the peak of incubation around 19 May. Nest initiation dates in South Dakota ranged from 20 April to 13 June (Wertz and Flake 1988). The corresponding incubation dates were 15 May and 8 July. No distinction was made between first and second nest attempts. Hatching dates for 15 nests in Oregon occurred over an 8-day period from 28 May to 4 June (Lutz and Crawford 1987E). Incubation dates for these nests ranged from 2 to 8 May and were similar to dates reported by Mackey (1982) in Washington. Unpublished data from northern Arizona (H. G. Shaw, Ariz. Game and Fish Dep.) and southeastern Montana (J. E. Gobielle, Mont. State Univ.) indicated most hens started incubation in early May and late May, respectively. The Montana data may have included second nest attempts. Jonas (1966), in contradiction to Gobielle's data, reported incubation in southeastern Montana started in �99 late April-early May. Jonas (1966) derived incubation dates indirectly from hatching dates assigned to poults visually classified to age in the field or from harvest samples of poults that were classified to age based on primary measurement techniques developed for eastern wild turkeys (M. g. silvestris) (Knoder 1959). This approach produced an inflated estimate of age and consequently an earlier estimate for onset of incubation. The peak of incubation (16-25 May) in southcentral Colorado approximated (± 2 weeks) dates reported from elsewhere within the native and expanded range of the Merriam's wild turkey. This relative uniformity among states suggests that photo-period ultimately controls nesting. Photoperiod best explained the synchrony of turkey nesting in Vermont (Wallin 1983). Spring weather had a secondary influence on timing of nesting as evidenced by earlier nesting in 1989 compared with 1986 and 1988. Vangilder et al. (1987) attributed annual variations in nesting chronology to spring temperatures. However, there appears to be a period, regardless of weather, before which hens will not initiate nesting. In this study and a study in southeastern New Mexico (Lockwood and Sutcliffe 1985), no hens started incubation before 6 May despite annual differences in spring weather conditions. Late April appears to be the earliest Merriam's wild turkeys initiate incubation. The median date for onset of incubation by Rio Grande wild turkeys in northeastern Colorado in 1986 was 6 May (Schmutz and Braun 1989). The median date of incubation for Merriam's turkeys in southcentral Colorado in 1986 was 18 May, almost 2 weeks later. Rio Grande's were incubating as early as 21 April. Phenological differences between study areas were at least partially responsible for advanced nesting in northeastern Colorado, but behavioral or physiological differences between subspecies also may have contributed to earlier nesting by Rio Grande's. Gobbling In Minnesota, the primary peak of gobbling associated with mating was consistent among years and occurred during the third and fourth weeks of April (Porter and Ludwig 1980). A secondary peak of shorter duration occurred in mid-May. Gobbling was heard throughout the monitoring period from 1 April to 17 June. Bevill (1975) documented a similar pattern of gobbling activity in South Carolina, except the primary (mid-Apr) and secondary (late Apr) peaks were earlier. Dates when gobbling was first and last heard ranged from 1 March to 10 July (Bevill 1973). Gobbling activity in Alabama peaked during the first and last weeks of April and ceased by mid-June (Davis 1969). The chronology of gobbling activities in Colorado most closely resembled patterns reported in Minnesota. Methods used to monitor and quantify gobbling have differed among studies, making comparisons difficult. Bevill (1973, 1975) recorded AM gobbling from fixed stations. He monitored both general and individual gobbling behaviors. Most other studies (Donohoe and Martinson 1963, Scott and Boeker 1972, Porter and Ludwig 1980) focused on general gobbling behavior and generated data from morning call-count routes. Davis (1969) used a combination of call-count routes and fixed stations to study general gobbling behavior. In this study, AM and PM gobbling indices were conducted on individuals, with location of the index depending upon where the bird roosted. �100 Wide daily variations in gobbling activity were apparent in all studies. When individuals were studied, it was apparent some males called more prolifically than others. Weather conditions accounted for some of this variation (Davis 1969, Bevill 1973), as did progression of the breeding season and, especially, onset of incubation. Gobbling increased on the roost and in the absence of hens. None of the cited studies specifically assessed gobbling in the presence or absence of hens. If gobbling serves to attract females (Bailey 1967:105), it should intensify in the absence of hens, which it did. Males also gobbled in the presence of hens and did not always gobble in their absence, suggesting gobbling may function in ways besides attracting females. Gobbling was most consistently heard during the morning when males were still on the roost. Bevill (1975) recorded his highest counts during the 20 minutes preceding sunrise. In 89 days of monitoring, he heard ~ 1 gobble on 53 (60%) mornings between 10 and 20 minutes before sunrise. I heard gobbling 73% of the time during the same interval. Bevill's (1973) study further indicated that adult eastern wild turkeys gobbled more than subadults. Converting his data to gobblesfhour revealed that adults gobbled an average of 62 timesfhour and subadults 13 timesfhour. Both age classes of eastern wild turkeys gobbled more than their respective age class of Merriam's. Bevill's (1973, 1975) data were collected from an unhunted population and may not reflect the true gobbling characteristics of eastern wild turkeys. He reported sporadic gobbling patterns for all stations on hunted areas. He excluded these data from the analyses. Davis (1969) also had difficulty interpreting gobbling data collected on hunted areas. His comparisons were complicated by use of different methods of monitoring gobbling on hunted (fixed stations) and unhunted (call-count routes) areas. Although questionable, evidence from his study indicated that gobbling was more sporadic and occurred less frequently on hunted areas. This may be a normal response to hunting pressure. Males monitored in southcentral Colorado were subjected to low hunting pressure and may have gobbled more than males on more heavily hunted areas. Mortality Neither the illegal harvest of non-bearded hens nor the legal harvest of bearded hens were major causes of mortality. Furthermore, only 2 radio-marked hens were reported harvested during the fall either sex season. Poaching accounted for almost 39% of the losses of radio-marked hens in Missouri; 42% of the total losses attributed to poaching occurred as a result of illegal harvest during the spring gobbler season (Kurzejeski et al. 1987). The legal fall harvest accounted for only 3% of the hen mortality. Spring hunting pressure in Missouri greatly exceeds that in Colorado. This factor alone was probably responsible for the high hen losses in Missouri due to illegal harvest. Mortality from late winter (Mar) through the nesting period (mid-Jul) was greater for hens than males. No males were harvested during the year they were being monitored and only 2 were shot in subsequent years. No gobblers were killed by predators during the breeding season (1 Apr-IS Jun). Most hen mortalities occurred prior to the onset of incubation. Nest losses were high, but mortality of hens on the nest was low. Although seasonal losses of hens in Missouri were highest (23%) during spring (14 Mar-31 May), few hens were �101 actually killed while incubating eggs (Kurzejeski et a1. 1987). In contrast, Speake (1980) identified incubation and the first 2 weeks of brood rearing as the time when hens were most vulnerable to predation. Home Range and Movements Although both HMT and MCP estimates indicated male Merriam's wild turkeys occupied large spring home ranges, MCP estimates exceeded HMT estimates by > 100%. MCP home ranges of Merriam's wild turkeys in Oregon (Lutz and Crawford 1989) were also larger than HMT home ranges reported here, averaging 1,655 ha for adult males and 2,345 ha for subadult males. Conversely, MCP home ranges (~ = 150 ha) of subadu1t males in Washington (Mackey 1982) were much smaller. Both investigators studied introduced populations and relied on triangulation techniques. Furthermore, Mackey (1982) collected home range data incidental to other objectives and, as a consequence, had additional sampling biases associated with his data set that were not apparent in the other studies. Brown (1980) found similar, but less extreme disparities in published home range sizes for eastern wild turkeys which he partly explained by differences in sampling and analyses from one study to another. The MCP home ranges included all location points between 1 April and 15 June. The HMT method, using the 90% contour interval, eliminated outlier locations and, thus, excluded some areas that were part of the MCP ranges. There was behavioral evidence that supported using the HMT approach, and suggested the MCP approach included (1) areas that were seldom used, and (2) portions of the winter and transitional ranges that were not part of the spring range. For example, all males used portions of their range more intensively than others. In addition, few males moved directly from winter to spring ranges, but instead, left the winter range over a period of about 3 weeks and used transitional ranges in the interim before moving onto spring ranges. Finally, some males used portions of their winter range throughout the spring period, whereas others had distinct winter and spring ranges. Subadu1ts occupied larger spring home ranges and used more areas within their ranges than adults. This was consistent with findings of Lutz and Crawford (1989), but in contradiction to published findings for eastern wild turkeys (Fleming and Webb 1974). Subadults frequently localized in certain areas within their range, sometimes for as long as 2 weeks, but seldom returned to these areas once they left. In comparison, adults repeatedly used several different areas within their spring ranges. Occasionally, they ventured away from normal areas of activity, but usually returned in < 1 week. These intense use areas varied from < 0.5 to almost 6 km apart. Wigley et a1. (1986) also found widely separated areas of intense activity within home ranges of eastern wild turkeys in Arkansas. Poor quality habitat has been cited as an explanation for large annual home ranges occupied by eastern wild turkeys in Arkansas (Wigley et a1. 1986) and Merriam's wild turkeys in Oregon (Lutz and Crawford 1989). Spring ranges of males, however, may be influenced by other factors in addition to habitat quality. Porter (1977) noted that during the breeding season males moved more and with less predictability than females. He contended the male's reproductive drive was the primary factor controlling spring movements. The period of increased movements of males, as indicated by distances between roosting sites, occurred from mid-April to early May and coincided with when �102 hens were moving to nesting areas. This increase in mobility of males occurred in conjunction with the spring hunting season. However, hunting pressure was minimal and not considered a factor influencing movements. Based on chronology of nesting events, mid-April to early May was also the period when hens were most receptive to males. Since males did not associate with the same hens throughout the breeding season, they had to periodically search for receptive hens. Males were more sedentary both before and after the hunting season when hens were still in flocks on winter range or localized on nesting areas. These observations partially agree with Porter's (1977) contention and suggest that size of a male's spring range was at least partially dependent upon availability of hens. Home range characteristics for wild turkeys differ greatly across their geographic range. Even within the range of individual subspecies, home range sizes vary markedly. Expansion into non-historic habitats has added to this variation and confounded development of management strategies precise enough to be meaningful, but general enough to apply to the range of environments occupied by wild turkeys. This especially applies to the concept of home range because it is rarely defined the same way and is affected by a number of other factors. Brown (1980) recommended that instead of developing management plans based on home range data obtained from several studies conducted over a wide area, it may be more appropriate to develop separate plans, each specific to the region where the data were collected. Brown's (1980) recommendation may be excessive and suggests emphasis should be placed on standardizing sampling methods and data analysis. Care must also be taken to define the time frame over which the home range estimate represents and what areas are included within the home range. For example, spring as defined included the period when some males were in transition between winter and spring ranges, while other males were still on winter range. This had a pronounced influence on home range estimates and, had it not been identified, values representing size of spring home ranges would have been misleading. Home range estimates revealed little about spring movements of males but, in combination with data on spatial relationships of roosting sites, indicated that males moved extensively during spring. Scott and Boeker (1972) did not fully understand the extent of spring movements and inconsistent gobbling patterns of Merriam's wild turkeys when they proposed that gobbling surveys could be used as a population index for this subspecies. Even with the tight control of extrinsic variations that was accomplished by Porter and Ludwig (1980), use of gobbling surveys to monitor relative abundance and trends in population growth of Merriam's wild turkeys is questionable without further research. Roosting Male Merriam's wild turkeys demonstrated low fidelity to spring roosting sites. This was not surprising considering the size of their spring home ranges. Even adults, which repeatedly used the same areas within their spring ranges, did not select the same roosting sites within these areas. In Oregon, Lutz and Crawford (1987h) identified few traditionally-used roost sites during any season. They did not quantify fidelity, but reported all spring roosting sites were only lightly used. I implied this to mean at least some sites were used more than once. In New Mexico, hen turkeys used only I of 29 summer �103 roosts more than once (Schemnitz et al. 1985). roosts were repeatedly used. In the same study, 29 winter Rio Grande wild turkeys in Oklahoma consistently left the roost about 15-20 minutes before sunrise throughout spring (Logan 1973). Roosting times ranged from sunset to 30 minutes after sunset, becoming later as spring progressed into summer. Early spring (analogous to the preseason period) departure times of Merriam's wild turkeys monitored in a previous Colorado study (Hoffman 1968) ranged from 10 to 15 minutes before sunrise; roosting times were more variable, ranging from 3 minutes before to 19 minutes after sunset. I observed a wider range of departure times than reported by Logan (1973) or Hoffman (1968), but a comparably broad range of roosting times. Hens were often present or nearby when I was monitoring roosting behavior. When this occurred, males departed and roosted concurrently with hens. In the absence of hens, adult males remained on the roost longer in the morning, but did not alter evening roosting patterns. Harvest Survey The permit system allowed quick access to names and addresses of hunters so the survey could be conducted immediately following each season. The other option was having to wait for the license agents to return license stubs containing hunter's names and addresses. Although agents have deadlines for returning license stubs, they do not always comply. In some cases, license stubs for the spring season are not received until August. In 1984, an attempt was made to survey spring hunters from names on license stubs. Because of the time required to obtain all the license stubs from the agents, the first mailing was delayed until 23 July. By bypassing the license agents and issuing permits at CDOW offices, the first mailing of the 1986 spring survey was out by 1 June, just 2 weeks after the season closed. Prior to 1984, the CDOW conducted 1 turkey harvest survey annually. Surveys were mailed in November and sampled about 50% of the license buyers with no attempt to stratify by spring only, fall only, or spring and fall hunters. In addition, not all license agents had returned license stubs when the sample was taken. Incomplete addresses recorded on the licenses plus the time lag between when the license was issued and when the survey was mailed resulted in a high percentage of non-deliverable surveys; i.e., 6-10% using license stubs vs. 2% using permits. Surprisingly, response rates did not differ between surveys conducted using names taken from license stubs or from permits. In 1989, the CDOW went to a separate license for spring and fall, thus, eliminating some problems with the old license system. Hunting pressure and harvest of wild turkeys in Colorado increased nearly 400% between 1973 and 1983 (Colo. Div. Wildl. 1984). The primary increases occurred in spring hunting pressure (600%) and spring harvest (1,000%). The harvest survey conducted between 1986 and 1988 suggested that hunting pressure and harvest had stabilized. However, as huntable turkey populations become reestablished on the West Slope, another increase in hunting pressure and harvest is anticipated. These populations have the potential to offer more hunting opportunity than East Slope populations, because they frequently occur on public lands readily accessible to the hunters. �104 Wing Collections The validity of any population index calculated from harvest samples is dependent upon the assumption that different age and sex classes are harvested in proportion to their occurrence in the population. Long-term population and harvest data are often necessary to test this assumption. Such data are not available for the Merriam's wild turkey in Colorado or elsewhere throughout its range. In addition, if wings are collected over a broad geographic area, they may not accurately reflect characteristics of local populations. These limitations do not preclude the use of wing data as a tool in formulating management strategies. Biologists must be aware of them and use caution in interpreting the data. Examination of wings (Merriam's only) originating from the fall season revealed the sex ratio of adults and subadults combined favored females, while the sex ratio of juveniles favored males. The biological implications of these ratios are difficult to interpret without knowledge of the population structure. Hoffman (1962) reported winter counts of 1.6 females:l male along the Front Range in southeastern Colorado. The sex ratio of 171 turkeys harvested on the Uncompahgre Plateau during the fall 1961 and 1963-67 was 1.7 females:l male (Myers 1973). Both ratios are based on combined samples of adults, subadults, and juveniles. A comparable ratio using the 1986-88 data would be 1.1 females:l male. Myers' data examined by age class indicates the sex ratio of adults and subadults (2.5:1) and juveniles alone (1.5:1) still favored females. Subadults were poorly represented in the fall harvest samples in all years. Biologically, this could be interpreted as poor production and subsequent low recruitment. However, another factor contributing to the low number of subadults is associated with the onset of primary molt and whether key feathers (primaries IX and X) for separating adults and subadults are still present at the time wings are collected. If these feathers have already molted, subadu1ts cannot be distinguished from adults. This was the case as 100% of the males and 51% of the females had completed their primary molt by the fall season. The problem of separating adults and subadults was compounded for males because they start and finish their primary molt before females. Misidentification of subadults is not a problem in spring when the birds are just beginning to molt their primary feathers. LITERATURE Amstrup, S. C. 44:214-217. 1980. A radio-collar CITED for game birds. J. Wildl. Manage. Bailey, R. W. 1967. Behavior. Pages 92-112 in O. H. Hewitt, ed. turkey and its management. The Wildl. Soc., Washington, D.C. _____ , and K. T. Rinell. 1967. Events in the turkey year. Pages H. Hewitt, ed. The wild turkey and its management. The Wildl. Washington, D.C. Bevill, W. V., Jr. turkeys. Proc. The wild 73-91 Soc., 1973. Some factors influencing gobbling activity Southeast. Assoc. Game and Fish Comm. 27:62-73. in O. among �105 1975. Setting spring gobbler hunting Proc. Nat1. Wild Turkey Symp. 3:198-204. Bray, O. E., and G. W. Corner. to birds. J. Wildl. Manage. seasons by timing peak 1972. A tail clip for attaching 36:640-642. Brown, E. C. 1980. Home range and movements Proc. Natl. Wit~ Turkey Symp. 4:251-261. of wild Burget, M. L. 1957. The wild turkey in Colorado. Dep. Fed. Aid Proj. W-39-R. 68pp. turkeys Colorado gobbling. transmitters - a review. Game and Fish Colorado Division of Wildlife. 1984. Colorado small game, furbearer varmint harvest. Dep. Nat. Resour., Div. Wildl., Denver. 210pp. Davis, J. R. survival. 52pp. and 1969. Wild turkey investigations - breeding, nesting and brood Job Completion Rep. Alabama Fed. Aid Proj. W-35-R, Job II-E. Dixon, K. R., and J. A. Chapman. 1980. Harmonic activity areas. Ecology 61:1040-1044. mean measure of animal Donohoe, R. W., and R. K. Martinson. 1963. Wild turkey inventory, Pages 43-45 in K. W. Laub, ed. Game research in Ohio. Vol. 2. Nat. Resour., Div. Wildl., Columbus. 1958-61. Ohio Dep. Fleming, W. H., and L. G. Webb. 1974. Home range, dispersal, and habitat utilization of wild turkey gobblers during the breeding season. Proc. Southeast. Assoc. Game and Fish Comm. 28:623-632. Hoffman, D. M. 1962. The wild turkey in eastern Game and Fish Tech. Publ. 12. 49pp. Colorado. Colorado Dep. 1966. Merriam's turkey roost preferences on mountain Colorado Div. Wild1. Game Info. Leafl. 45. 6pp. ranges. 1968. Roosting sites and habits Wildl. Manage. 32:859-866. in Colorado. of Merriam's turkeys J. 1973. Some effects of weather and timber management on Merriam's turkey in Colorado. Pages 263-271 in G. C. Sanderson and H. C. Schultz, eds. Wild turkey management: current problems and programs. Univ. Missouri Press, Columbia. Jonas, R. J. 1966. Merriam's turkeys in southeastern and Game Dep. Tech. Bull. 3. 36pp. Knoder, E. 1959. rate of primary 1:159-176. Montana. Montana Fish An aging technique for juvenal wild turkeys based on the feather moult and growth. Proc. Natl. Wild Turkey Symp. Kurzejeski, E. W., L. D. Vangilder, and J. B. Lewis. 1987. turkey hens in Missouri. J. Wildl. Manage. 51:188-193. Survival of wild �106 Lockwood, D. R., and D. H. Sutcliffe. 1985. Distribution [sic] mortality, and reproduction of Merriam's turkey in New Mexico. Proc. Natl. Wild Turkey Symp. 5:309-316. Logan, T. H. Oklahoma. 1973. Proc. Seasonal behavior of Rio Grande wild turkeys in western Southeast. Assoc. Game and Fish Comm. 26:74-91. Lutz, R. S., and J. A. Crawford. 1987~. Reproductive success and nesting habitat of Merriam's wild turkeys in Oregon. J. Wi1dl. Manage. 51:783787. _____ , and and hen-poult _____ , and Merriam's 1987Q. Seasonal flocks in Oregon. wild 1989. turkey use of roost sites by Merriam's Northwest Sci. 61:174-178. Habitat use and selection and home ranges in Oregon. Great Basin Nat. 49:252-258. Mackey, D. L. 1982. Ecology of Merriam's wild Washington with special reference to habitat Washington State Univ., Pullman. 87pp. Mohr, C. O. mammals. 1947. Table of equivalent Am. Midl. Nat. 37:223-249. 1977. Home range J. Wildl. Manage. of turkey in southcentral utilization. M.S. Thesis, populations of North American Myers, G. T. 1973. The wild turkey on the Uncompahgre Plateau. Colorado Div. Wildl. Fed. Aid. Proj. W-37-R. l53pp. Porter, W. F. Minnesota. turkey hen dynamics of wild 41:434-437. turkeys small Final Rep. in southeastern _____ 1978. The ecology and behavior of the wild turkey (Meleagris gallopavo) in southeastern Minnesota. Ph.D. Thesis, Univ. Minnesota, Paul. l22pp. St. _____ , and J. R. Ludwig. 1980. Use of gobbling counts to monitor the distribution and abundance of wild turkeys. Proc. Natl. Wild Turkey 4:61-68. Symp. Schemnitz, S. D., D. L. Goerndt, and K. H. Jones. management of Merriam's turkey in southcentral Wild Turkey Symp. 5:199-231. 1985. Habitat needs and New Mexico. Proc. Nat1. Schmutz, J. A., and C. E. Braun. 1989. Reproductive Grande wild turkeys. Condor 91:675-680. performance Scott, V. E., and E. L. Boeker. 1972. An evaluation Arizona. J. Wildl. Manage. 36:628-630. of turkey Speake, D. W. 1980. Predation Turkey Symp. 4:86-101. on wild turkeys in Alabama. of Rio call counts in Proc. Natl. Wild Spraker, T. R., W. J. Adrian, and W. R. Lance. 1987. Capture myopathy in wild turkeys (Meleagris gallopavo) following trapping, handling, and transportation in Colorado. J. Wildl. Dis. 23:447-453. �107 Stuwe, M., and C. E. B10whowiak. analysis of animal locations. Smithsonian Inst., Washington, 1986. Micro computer program for the Conserv. and Res. Cent., Natl. Zool. Park, D.C. 18 pp. Vangilder, L. D., E. W. Kurzejeski, V. L. Kimmel-Truitt, and J. B. Lewis. 1987. Reproductive parameters of wild turkey hens in north Missouri. Wildl. Manage. 51:535-540. Wallin, J. A. 1983. Eastern wild turkey tolerance to human disturbance during incubation and its relationship to spring hunting in Vermont. Thesis, Univ. Vermont, Burlington. 43pp. Wertz, T. L., and L. D. Flake. 1988. Wild turkey nesting central South Dakota. Prairie Nat. 20:29-37. ecology J. M.S. in south Wigley, T. B., J. M. Sweeney, M. E. Garner, and M. A. Melchiors. 1986. Wild turkey home ranges in the Ouachita Mountains. J. Wildl. Manage. 50:540544. Prepared by ��109 FINAL REPORT State of: Colorado Project: W-152-R Upland Work Plan: 13 Job Title: Seasonal Habitat County. Colorado Period Covered: Author: Personnel: Anthony Bird Research Job _9_ 01 January Use by Plains through Sharp-tailed 31 December Grouse in Dou~las 1989 W. Hoag C. E. Braun, K. Demarest, D. Prenzlow, Colorado Division of Wildlife; A. W. Hoag, E. Redente, Colorado State University ABSTRACT The status and habitat use by plains sharp-tailed grouse (Tympanuchus phasianel1us jamesii) were investigated during 1986-88 in Douglas County, Colorado. This subspecies once occupied suitable habitats in northeastern Colorado, but its range has been greatly reduced because former habitats have been altered by man. Within Colorado, the estimated population size in 198688 was 175-225 birds existing only in Douglas County in an area with 6 known leks and 2 historic lek sites. Habitat use was studied by following 23 grouse captured on 4 leks that were fitted with solar-powered radios. Seven habitat variables and 7 sharp-tailed grouse activities were studied. Canopy cover, plant height, and distance to clump cover were most important in identifying areas used by grouse. Habitat use varied by season and sex. Shrub communities dominated by Gambel oak (Quercus iambelii) and true mountainmahogany (Cercocarpus montanus) were selected in fall and winter and for escape and nesting. Transition zones between shrub-do~inat~d areas and grasslands were selected in spring and for roosting and nesting. Grasslands ~ere selected in summer and for feedingloafing and mating activities. Home range size varied between 103 and 363 ha for males and 306 and 884 ha for females. Management of private lands inhabited by plains sharp-tailed grouse should focus on maintaining residual vegetation from the previous year to increase cover for nesting, brooding, feeding-loafing, and roosting with a goal of maintaining a minimum of 6 occupied lek sites with documented interchange among sites. A population monitoring system should be established to document number of active leks and birds attending each on an annual basis. ��111 THESIS PLAINS SHARP-TAILED GROUSE USE IN DOUGLAS STATUS AND HABITAT COUNTY, COLORADO Submi tted by Anthony Department In partial W. Hoag of Range fulfillment for the Degree Colorado Science of the requirements of Master of Science State University Fort Collins, Colorado Fall 1989 �112 COLORADO STATE UNIVERSITY November 6, 1989 WE HEREBY RECOMMEND THAT THE THESIS PREPARED UNDER OUR SUPERVISION BY ANTHONY W. HOAG ENTITLED PLAINS SHARP-TAILED GROUSE STATUS AND HABITAT USE IN DOUGL~S COUNTY, COLORADO BE ACCEPTED AS FULFILLI~G IN PART REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE. Committee AdViS~/ on Graduate . @J/ epar;ment Head ii Work �113 ABSTRACT PLAINS SHARP-TAILED OF THESIS GROUSE USE IN DOUGLAS The status and habitat (Tvrnpanuchus phasianellus Douglas County, habitats reduced because Colorado, Colorado. historic lek sites. Habitat that were grouse and distance dominated mountainmahogany and for escape areas nesting. mating males Habitat Grasslands activities. radios. were oak Transition 6 known Canopy Within were selected zones between on 4 leks variables plant Shrub and true in fall and winter shrub-dominated were selected in spring and for roosting were selected in summer and for feeding-loafing Home range size varied and 306 and 884 ha for females. iii between and in identifying and sex. gambelii) birds leks and 2 cover, important by season (Quercus montanus) greatly 175-225 Seven habitat studied. use varied by Gambel by man. was in suitable 23 grouse captured cover were most (Cercocarpus and grasslands altered 1986-88 during its range has been in an area with activities grouse once occupied size in 1986-88 solar-powered and nesting. but by following to clump areas used by grouse. communities County use was studied fitted with sharp-tailed investigated have been population in Douglas 7 sharp-tailed height, Colorado, the estimated only were COLORADO This subspecies former habitats existing COUNTY, use by plains jamesii) in northeastern STATUS AND HABITAT and 103 and 363 ha for and �114 Management should focus to increase with of private on maintaining cover should interchange be established attending inhabited residual for nesting, a goal of maintaining documented lands vegetation brooding, a minimum among sites. to document by plains from the previous feeding-loafing, of 6 occupied A population number sharp-tailed of active grouse year and roosting lek sites with monitoring system leks and birds each on an annual basis. Anthony W. Hoag Department of Range Science Colorado State University Fort Collins, CO 80523 Fall 1989 iv �115 ACKNOWLEDGMENTS This study was supported thank Drs. Ed F. Redente in providing support employees of Wildlife. as I pursued along with their demand my degree. of the Colorado and useful suggestions Division comments. of Wildlife Without to I also thank for their frank Dr. Wayne C. Leininger on the thesis. I efforts for commitment I would not have learned as much as I did. discussions helpful Division and Clait E. Braun for their tireless and direction their open door policy excellence, by the Colorado provided I thank Diane K. Hall for preparing the final copy of the thesis. I am deeply Science indebted and Fishery thank Andy Sipowicz. were invaluable. conduct research to my fellow graduate and Wildlife Biology His critical Tweet Kimball Departments. comments during and Dave Miller on their ranches. students in the Range In particular, I thesis preparation graciously I hope their interest allowed me to in sharp- tailed grouse remains high. Lastly, continued I could not have completed support. I thank my mother degree as enough and for encouraging this project without for not accepting my undergraduate me to finish my masters v my family's degree. �116 Table of Contents ABSTRACT OF THESIS iii ACKNOWLEDGMENTS. LIST OF TABLES . viii LIST OF FIGURES. Chapter 1. ix DISTRIBUTION AND ABUNDANCE OF PLAINS SHARP-TAILED GROUSE IN DOUGLAS AND ELBERT COUNTIES, COLORADO. Introduction. 1 Study Area. 2 Methods Results ') Discussion. Literature Chapter 2. 6 Cited. SEASONAL HABITAT USE BY PLAINS SHARP-TAILED DOUGLAS COUNTY, COLORADO GROUSE IN Introduction. 8 Study Area. 9 Methods 11 Results 14 14 Trapping Movements Habitat Discussion. Literature 14 and Home Range Characteristics. 16 30 . . . 38 Cited. vi �117 Table of Contents Chapter 3. (Continued) MANAGEMENT RECOMMENDATIONS FOR MAINTENANCE OF PLAINS SHARP-TAILED GROUSE IN DOUGLAS COUNTY, COLORADO Analysis. 41 . . . Recommendations 43 Literature 44 Cited. vii �118 List of Tables 1.1 2.1 2.2 2.3 2.4 2.5 2.6 High counts of plains sharp-tailed County, Colorado 1986-88 .... Daily and seasonal time periods grouse behaviors. . . . grouse, Douglas 4. for plains sharp-tailed . . . . . 12 Number of plains sharp-tailed grouse observed and captured in Douglas County, Colorado, 1987-88 15 Distances (km) moved seasonally by plains sharp-tailed grouse in Douglas County, Colorado, 1987-88 ... 16 Dominant plant species (%) at plains sharp-tailed use and random sites. . .. . . 31 Clump speci~s random sites. (%) at plains sharp-tailed . Plant group cover (%) at feeding-loafing of plains sharp-tailed grouse viii grouse grouse use and . 32 and nest sites . 33 �119 List of Figures Figure 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9 2.10 2.11 2.12 Distribution of plains sharp-tailed Douglas County, Colorado, 1986-88 Home range size of 10 male plains Cherokee Lek, spring 1987-88 grouse use areas sharp-tailed Home range size of 6 male plains Cherokee Lek, summer 1987-88 sharp-tailed Home range size of 3 male plains Cherokee Lek, fall 1987-88 sharp-tailed grouse, . 17 18 grouse, 19 . sharp-tailed . sharp-tailed grouse, 20 grouse, 21 . Horne range size of 1 female plains sharp-tailed Dakin Lek, spring-summer 1987 . . . . . . . . Horne range size of 2 male plains sharp-tailed Indian Meadows Lek, spring 1988 . . . . .. 10 grouse, . Home range size of 2 female plains Cherokee Lek, spring 1988 Home range size of 3 male plains Dakin Lek, spring 1987.. in . grouse, 22 grouse, .... 23 Structural variables at plains sharp-tailed grouse use and random sites in spring (Mar-May), Cherokee Lek ... 25 Structural variables at plains sharp-tailed grouse use and random sites in summer (Jun-Aug), Cherokee Lek ... 26 Structural variables at plains sharp-tailed grouse use and random sites in summer (Jun-Aug), Indian Meadows Lek. 28 Structural variables at plains sharp-tailed grouse use and random sites in summer (Jun-Aug), Dakin Lek .... 29 ix �120 Chapter DISTRIBUTION AND ABUNDANCE IN DOUGLAS 1 OF PLAINS AND ELBERT SHARP-TAILED COUNTIES, GROUSE COLORADO INTRODUCTION Plains sharp-tailed historically Mountains occupied from Arapahoe, sharp-tailed in Larimer Douglas, and Niedrach 1914). However, With by Snyder 1965). 1912), Boulder (1897) and Sclater The plains declined along 1909), Douglas (Aiken and Warren (1912) reported Apparently, race of sharp-tailed to all sharp-tailed and the foothills dramatically by Lincoln Lincoln, This race of (Henderson in Colorado. from Colorado name was applied 1912) south into El are in the Denver was most abundant and abundance was described of the Rocky this area, specimens and Yuma counties grouse were not common and 1887 (Cooke 1897). Colorado Within 1965), and El Paso counties distribution subspecies 1912). (Bailey and Niedrach both Cooke jamesii) 1914) and east to Kit Carson, grouse historically (Bailey jamesii) (Cooke 1897, Sclater Elbert, History phasianellus east of the Front Range (Cooke 1897, Sc1ater sharp-tailed (Tympanuchus (Cooke 1897, Sclater of Natural species' County (Aiken and Warren Yuma counties Museum habitats from Larimer Paso County grouse that the between grouse 1877 (I. p. (1917)_ and the grouse in eastern (1939). the exception columbianus) in northwest sharp-tailed grouse of a study of sharp-tailed Colorado in Colorado grouse (Dargen et al. 1942), appear in the literature (I. p. few records of from the early �121 2 1900's until the 1960's. Fish initiated Colorado a survey (Rogers described reported During Elbert, on habitat grouse sharp-tailed Elbert counties grouse Stearns and possibly Rogers (1962-65) in El Paso, grouse Colorado (1964) and greater (1969) reported to occur Phillips, in (1968) and Evans sharp-tailed were known grouse in east-central and El Paso counties) used by a hybrid of Game and of sharp-tailed (1. cupido) in Yuma County. plains Department the same period, used by sharp-tailed Douglas, prairie-chicken Yuma of the distribution 1969). habitats (Arapahoe, In 1962, the Colorado that in Douglas Sedgwick, and Teller, and counties. Numbers of plains Colorado are unknown present. Personnel although of 200-300 Div. Wildl. 1978). tailed Endangered grouse most recent birds being Division than at of Wildlife commonly 1960's in the early as endangered (Kahn 1979) resulted historically higher 1970's Act of 1973 resulted classified inventory that occurred from the early Discussions Species grouse they were obviously of the Colorado an estimate federal sharp-tailed in used to late 1970's (Colo. and enactment of the in the plains in Colorado sharp- in 1976. in an estimate The of 175-200 birds. The objectives distribution Douglas and estimate and Elbert observed males historical of this study were the abundance counties. and females to delineate Included of plains the present sharp-tailed are data on maximum at each known grouse number lek, and the status of of inactive leks. STUDY AREA The study area extended County east to Kiowa from the foothills and from northern Douglas in western County in Douglas to northern El �122 3 Paso County. Elevations in the southwest. by broken The western foothills 2000 m in height eastern region (Stearns separated consisted little changes. was between 35 and 60 em with grass Eragrostis The short-grass by low diurnal over a 30-year occurring prairie gambelii, Cercocarpus and riparian was dominated trichodes, and Yucca (Stearns precipitation The and plains and wide in the area were: pine prairie tectorum, spp. foothills winds, than regions. was characterized the majority types shrubland, The mid-grass comata, into 3 distinct rain, moderate annual greater period in late spring 1974). vegetative prairies, land. Mean foothills rolling The climate to 2,438 m of the area was characterized Two parallel of intermixed temperature The major one-third the study area 1979). sunny days, (U.S. Dep. Agric. from 1,554 m in the northeast and buttes. 1968, Kahn humidity, varied glauca. forests, Shrub lands were montanus, Forbs were common surveys and counts of birds (Evans gracilis, dominated Rhus aromatica, 1968). spp., ~ longifolia by Bouteloua and short- and cultivated by Andropogon and Calamovilfa was dominated mid-grass Bromus by Quercus Poa pratensis, in all habitat 1964). and Carex types. METHODS Roadside from 0430 Roadside scope, private to 0900 hours surveys during for sharp-tailed Landowners interviewed through May with Surveys were 3-5 minute stops microphone, taken along every leks were conducted 1986 and 1988. were aided by use of a parabolic and binoculars. ranches March on active county roads 0.8 km to listen spotting and on and look grouse. with known to provide historical historical leks were contacted data and status and of current populations �123 4 on their land. Where to obtain coverage data were plotted necessary, of ranches field searches with on topographic suitable maps were conducted appearing to establish on foot habitat. priority Location areas to be searched. RESULTS Plains 6 active sharp-tailed grouse were only located leks documented apparently inactive, the vicinity. (Table 1.1). although Fifty-three 4.3 females/lek) were males observed Table 1.1. High counts Colorado 1986-88. male in Douglas Two historical sharp-tailed on active of plains with leks were grouse (R = 8.8 males/lek) County were observed in (~ = and 26 females leks. sharp-tailed grouse, Douglas County, Males Females Cherokee 10 7 Dakin 11 4 Woodho~se 16 11 Greenland 5 0 Indian Meadows 3 0 Lincoln Mountain 8 4 Rancha 2 0 2 0 Lek Highlands Winkler Rancha aInactive although The 6 lek locations topography. non-displaying were The communities isolated of Castle males were observed in 2 distinct in the area. areas by local Rock and Larkspur lie between the �124 5 2 areas decreasing the probability leks were in Elb~rt observed If only one-half high count (Robel the population apparently is doubtful grouse during at the 6 active leks was between leks were grouse was greater No active the and Boag 1974) and the sex ratio was historical sharp-tailed 1986-88 in an area were present sharp-tailed that all active of the plains during of the males of plains to west movement. County. 1970, Rippin inactive of east population and 2 200 and 225 birds. located. Thus, It the minimum in Douglas 1:1, County, size Colorado than 200 birds. DISCUSSION The cumulative effect developments, intensified the available habitat Elbert counties County prior of conversion grazing, for plains (Chapter 2). and wildfire sharp-tailed to 1986, but no birds Rangeland than in Douglas County, The key factors in the apparent in Elbert grouse County County appear conducted shrub suburban of plains and in Elbert has a smaller disappearance has reduced in surveys but there are also fewer to suburban in Douglas existed were observed in Elbert rangeland suppression A small population 1986 and 1988. grouse of native in component developments. sharp-tailed to be the lack of nesting and escape cover. Key factors appear plains to be loss of native conifers as a result grazing. grouse affecting leks Development on 2 factors, rangeland surrounding proceeding for plains development grouse to development, of fire suppression, is currently The future sharp-tailed in Douglas invasion and intensive sharp-tailed or is in preliminary and habitat. grouse stages. in Douglas If development of livestock 5 of the 8 documented sharp-tailed County County continues depends a~ its �125 6 present rate, useful decline. The key needs km) with interspersion If habitat habitat hope habitats management, improvement for survival for sharp-tailed for sharp-tailed grouse of native grasses, including maintaining is not initiated of plains grouse are open forbs, open space, grouse land and shrubs in the near sharp-tailed will markedly future, (1.6 x 3.2 (Chapter 2). or intensive there in Douglas is little County, Colorado. LITERATURE CITED Aiken, C. E. H., and E. R. Warren. 1914. Colorado. Colorado Coll. Sci. Series The birds of El Paso County, 12:455-603. Bailey, A. M., and R. J. Niedrach. 1965. Denver Mus. Nat. Hist., Denver, Colo. Birds of Colorado. 454pp. Vol. 1. Colorado Division of Wildlife. 1978. Essential habitat for threatened and endangered wildlife in Colorado. Colorado Div. Wildl., Denver. 84pp. Cooke, W. W. 1897. 37. l44pp. The birds of Colorado. Colorado Dargan, L. M., H. R. Shepherd, and R. N. Randall. tailed grouse in Moffat and Routt counties. Dep., Sage Grouse Surv., Vol. 4. 28pp. Agric. Coll. Bull. 1942. Data on sharpColorado Game and Fish Evans, K. E. 1964. Habitat evaluation of the greater pra1r1e chicken in Colorado. M.S. Thesis, Colorado State Univ., Fort Collins. 98pp. Henderson, J. Colorado. Kahn, 1909. Univ. An annotated list of birds of Boulder Colorado Studies 6:219-242. R. 1979. Prairie sharp-tailed grouse Colorado Div. Wildl., Denver. 12pp. segment. Lincoln, F. C. 1917. A review of the Pedioecetes Biol. Soc. Washington 30:83-86. Rippin, A. B., and D. A. Boag. male sharp-tailed grouse. Robel, R. J. prairie 1970. chicken 1974. Recruitment J. Wildl. Manage. County, Unpubl. Rep., in Colorado. Proc. to populations 38:616-621. of Possible role of behavior in regulating greater populations. J. Wildl. Manage. 34:306-312. Rogers, G. E. 1969. The sharp-tailed grouse in Colorado. Div. Game, Fish and Parks Tech. Publ. 23. 94pp. Colorado �126 7 Sc1ater, W. H. 1912. A history Co., London, U.K. 576pp. Snyder, L. L. 1939. 56:184-185. Great of the birds Plains races of Colorado. of sharp-tailed Stearns, F. D. 1968. Sharp-tailed grouse density Colorado related to selected habitat factors. Colorado State Univ., Fort Collins. 133pp. U. S. Department of Agriculture. 1974. area, Colorado. u.s. Dep. Agric., D.C. 129pp. Witherby grouse. and Auk in east-central M.S. Thesis, Soil survey of Castle Rock Soil Conserv. Serv., Washington, �127 Chapter SEASONAL HABITAT 2 USE BY PLAINS IN DOUGLAS COUNTY, SHARP-TAILED GROUSE COLORADO INTRODUCTION Plains occupy sharp-tailed suitable into southern grouse habitats Canada from northeastern (Aldrich range of this subspecies periphery, (Aldrich largely favoring and shrubs sharp-tailed In Colorado, suitable habitats Larimer County Boulder along the mixture (Sisson Nebraska The its southern grazing sharp-tailed by man is a dominant grouse. Range of grasses, forbs, 1976). grouse historically 1912) south 1912). occupied (1897) and Sclater grouse were not common Lincoln, in Colorado. and Yuma This race of sharp-tailed 1909), Douglas (Aiken and Warren in Larimer (Bailey grouse (Cooke 1897, and Niedrach 1914). (1912) reported Apparently, 1877 and 1887 (Cooke 1897). from into El Paso County along the foothills (Henderson and El Paso counties between 1965). have been altered 1914) and east to Kit Carson, Sclater Both Cooke and western and Niedrach reduced jamesii) east of the Front Range of the Rocky Mountains most abundant uncommon has altered sharp-tailed was apparently 1965), Colorado grouse use for cover (Cooke 1897, Sclater 1912), phasianellus Livestock range of plains livestock plains 1980). (Cooke 1897, Sclater (Aiken and Warren counties former habitats and Graul land use in the remaining management 1963, Bailey has been greatly because 1963, Miller (Tympanuchus that sharp-tailed the species At present, became �128 9 self-sustaining populations occur only in Douglas of plains sharp-tailed County within Colorado. This study was initiated remnant population Colorado. grouse are known to in 1987 to describe habitats of plains sharp-tailed used by the grouse in Douglas County, Included were data on seasonal habitat use and home ranges. STUDY AREA The study area was in the foothills County, Colorado (Fig. 2.1). and tablelands Elevation ranged from 1,850 to 2200 m. The climate was high inland continental Mountains and the Palmer Divide as modified (U,S. Dep. Agric. climatic characteristics variable winds, and a Ivide temperature included low precipitation, (U.S. Dep. Agric. 1974). period receiving 1974). range. The temperature Upland deposits. gravelly 16-20% of the annual (Dec-Feb) was the driest (U.S. Dep. Agric. 1974). rolling to broken of Plum and Cherry creeks. widely, but were primarily red arkose, sedimentary, calcareous, tablelands dissected Soil parent materials varied arkosic fans and pedisediments, and red sedimentary bedrock, Soils were loam to clay in the western in the eastern low humidity, ranged from -30 to 38°C with an annual mean of areas were mOderately by tributaries Winter General The annual precipitation only 8-9% of the annual moisture 19°C (U.S. Dep. Agric. by the Rocky 1974). was 37-44 cm with May the wettest month receiving moisture of Douglas and Eolian section and sandy to section of the study area (U.S. Dep. Agric. 1974). Gambel oak (Quercus gambelii) (Cercocarpus Fragrant montanus) and true mountainmahogany were the dominant sumac (Rhus aromatica) shrubs in uplands and rabbitbrush areas. (Chrvsothanmus spp.) �129 10 Nt 8 S 6 (J) N at '" 7 -i 0 ~ 8 z (J) J: -0 !:E 9 10 71 70 69 68 RANGE 67 66 (W) Rg. 2.1. Distribution of plains sharp-tailed grouse use areas, Douglas County, Colorado, 1986-88. �130 11 were common Melilotus, included understory shrubs. Descurainia, Poa, Bromus, ranching, housing Understory and Yucca. Koeleria, developments, forbs Dominant included grasses in upland and Agropvron. Primary and small farming. scale Astragalus, areas land uses were METHODS Plains Walk-in sharp-tailed traps with tailed grouse the lek. powered were captured et al. 1982). leg bands, capture, radio and plastic individual birds' time of capture packages. radio failure. element Yagi the birds. Geologic Survey times. sites were to direct sharp- or at the periphery of nets and spotlight were classified and fitted to designate to sex with a solar- serially-numbered 15-25 g and were aluminum lek of less than 3% of at lek, feeding-loafing, (use sites). birds located Locations using interpretation time for birds relocated map by season classified (Table 2.1). until a portable were not taken to adjust to predated or receiver and 3- on a 7.5-minute and activity. during escape, by triangulation was plotted 2-3 times each week were Radiolocations were ascertained Each observation topographic All behaviors facilitate coded for 3 days to allow Grouse antenna. to locate color Birds were periodically flushing made 1944), 1980), weighed were obtained roost, .nest, and brood radio grouse for study. body weight. Radiolocations from (Amstrup Radios cannon (Ammann bandettes sex, and age. with chosen traps in the middle Captured et al. 1967) and age poncho-mounted leks were leads connecting into them were placed (Giesen (Henderson at 4 active chicken-wire Some birds trapping grouse randomly by time periods or by U.S. Attempts were selected use or location to �131 12 Table 2.1. Daily grouse behaviors. and seasonal time periods Behavior for plains sharp-tailed Time period/location Feeding-loafing 0.5 hours after Roosting 0.5 hours before Escape Area Lek 0.5 hours to where sunrise to 0.5 hours before sunset sunset to 0.5 hours before sunrise bird before flushed sunrise to 3 hours after sunrise, 1 Mar to 30 May Nesting 15 May to 15 Jul, hens Brooding Hens with chicks, Radiolocations polygon were pooled was drawn around behavior. Vegetation site and season. obtained using a random 3.2 km2 area defined initial year started with measurements polygon was met. least taken separately for each use with use sites were with all points grouse within locations samples from a randomly to sample generated taken at 25-m intervals were obtained, sample 1 m of the random vegetation were used If the edge of the use polygon Each vegetation residual generator transects polygon within were sites for comparison number (> 2) for each of observations by all sharp-tailed random sampling sufficient concentrations for each use site and a a 1.6 x during the of study. Stratified Vegetation bi-weekly measurements Random 1 Jul to 31 Aug included point, that could 50% cover measured a second canopy distance supply within a use the edge of the was met before transect cover, was mean plant cover a sharp-tailed on a spherical point until random to clump use sites. densiometer), started. height (any live or grouse with height, at width, and �132 13 length of clump clump cover, species. cardinal species were from the sample and random measurements nesting/brooding species were site, sites taken to test for habitat 10 times within sampling derived (121 possible feeding-loafing canopy cover Home estimates Blowhowiak means. Aug), contours to illustrate Movements and fall were Procedure 1 summer, structural spring compared based samples around high use areas categorized on a lOom was used to the nearest using seasonal harmonic were constructed for and 25-m grid for and study population. to calculate 1%. Dixon and and total MCPAAL means for (Stuwe and using a 10 x 10 at 10, 25, 50, and 95% calculated for spring were (Mielke (2 spring Indian Meadows). variables summer between (MRPP) collected habitat versus generator were analyzed mean measure characteristics sets of data were Dakin; Points sites, through (Mar-May), harmonic summer (Jun- (Sep-Nov). Vegetation Permutation sites. A 0.5-m2 quadrat and movements 1985) was used intervals for nest samples. for each bird Isopleth number feeding-loafing and cover was estimated (1980) harmonic radio locations points) and random range analysis Chapman's grid. sample a random selection. 50 m of a general a 1.6 x 3.2 km2 area for random grid in each of the 4 at feeding-loafing, 20 times within using and point. and 20 within were at the random point, taken once cover was estimated nesting/brooding, Random plant All measurements directions Plant dominant habitat, examined using the Multi-Response et al. 1976, Mielke and 2 summer, Tests Cherokee; of data distribution were performed and between sites due to differing 1979). on random versus use categories. grazing intensities. Six 1 summer, for 6 use sites, Data were not �133 14 Plant standing data were pooled dead). MRPP was used into 3 categories Percentages were calculated to test for differences feeding-loafing, nesting, and random (grasses/ forbs, shrubs, for use and random between plant cover sites. groups at sites. RESULTS Trapping Thirty-nine in walk-in spotlight of 42 plains traps. One bird was caught and 2 were recaptures). The most effective Eighteen different at the Dakin (Table Lek, 3 at Lincoln of sharp-tailed of 11 males males and for short periods (Table were grouse in the center adult males. Lek, were of the Thirteen 5 were and 2 at the Indian captured Meadows Lek at lek sites varied at the Cherokee study making observed site. it difficvlt study site to a low of 3 Hens visited to obtain from a leks infrequently reasonable counts 2.2). Movements and Home Range Plains summer. sharp-tailed They used moving movements marked grouse and 4 females at the Indian Meadows while net and sharp-tailed at the Cherokee Mountain, caught (19 of 42 captures lines of traps placed were captured were a long handled to direct captured captured 2.2). Numbers high using set-up of 23 individuals individuals grouse caught with a cannon-net into traps was 3 separate lek. sharp-tailed the same transitional between were grouse used different seldom winter habitat and summer habitats. greater hen was predated habitats in winter in spring Daily than 2 km (Table 2.3), and fall and seasonal although 8 km from the last site at which and one radio- she was �134 15 located. Longer movements between leks probably occurred but were not observed. Table 2.2. Number of plains sharp-tailed in Douglas County, Colorado, 1987-88. grouse observed and captured Captured Observed M Lek F M F 1987 13 2 7 1 7 4 3 2 7 o 3 o Cherokee 9 4 3 2 Indian Meadows 3 o 2 o 39 10 18 5 Cherokee Dakin Lincoln Mountain 1988 Totals Spring movements visitation, dominated grouse habitats. and flushed movements August. areas feeding-loafing fed, loafed, habitat were largest and roosted grassland-forb shrub-dominated winter. season progressed, in grass/forb to areas dominated The longest areas lek to oakbrush- sharp-tailed typical of summer by mountainmahogany. Summer 1.6 km2 at all study sites from late May to middaily movements areas. habitats areas included on ridges near leks, and escape As the breeding were within to escape as daily activity of 500 m were from feeding-loafing In late August through and September transitional (Gambel oak) where grouse moved grass-shrub they remained habitats through the from into �135 16 Table 2.3. Distances (km) moved seasonally grouse in Douglas County, Colorado, 1987-88. by plains sharp-tailed Males = 13) Females (n = 5) (n Period n n Range locations 163 0.15-8.0 0.125-0.50 64 0.0-l.25 131 0.375-2.50 25 0.35-2.5 23 0.125-2.25 5 Spring (Mar-May) 439 0.25-3.25 Summer (Jun-Aug) 304 Fall (Sep-Nov) Winter (Dec-Feb) The area used by marked (diam.), respectively at the Indian Meadows Home females ranges leks. of males were more continuous One hen nested visit throughout 0.10-2.25 3.2 and 4.8 km Sharp-tailed during grouse summer. and smaller than those of 4.8 km from lek of capture the summer. the summer Males and were remained observed and within to the lek. Characteristics Habitat Measurements Cherokee of habitat Lek were based characteristics on 7 males data at Indian Meadows radio-marked sufficient at Lincoln on 3 males were based Mountain data could be collected. for birds in 1987 and 3 males Data at the Dakin Lek were based 1987, while males and Dakin in the same area throughout infrequently ·grouse was site were only followed 500 m of lek of capture 1988. sharp-tailed at Cherokee (Figs. 2.2-2.8). remained .Range locations marked at the and 2 females and 1 female on 2 males during in 1988. in 1987 were predated in before Three �136 17 • •• •• ••I • • .. ~ '" • • •• •• • • • •• • • Rg. 2.2. Home range size of 10 male plains sharp-tailed grouse, Cherokee Lek, spring 1987-88. (Hectares in each contour interval are: 10% = 1.73. 25% = 14.82, 50% = 35.86, 95% 220.13). = �137 18 • • • • • • • • • •• •• • • • Fig. 2.3. Home range size of 6 male plains sharp-tailed grouse, Cherokee Lek, summer 1987-88. (Hectares in each contour interval are: 10% = 1.54, 25% = 6.03,50% = 15.71,95% = 103.84). �138 19 • Fig. 2.4. Home range size of 3 male plains sharp-tailed grouse, Cherokee Lek, fall 1987-88. (Hectares in each contour interval are: 10% = 2.83, 25% = 9.99, 50% 53.28, 95% = 363.28). = �139 20 • • Rg. 2.5. Home range size of 2 female plains sharp-tailed grouse, Cherokee Lek, spring 1988. (Hectares in each contour interval are: 10% 3.55, 25% 25.78, = 50% = 67.95: 95% = 306.96). = �140 21 • • • • Fig. 2.6. Home range size of 3 male plains sharp-tailed grouse, Dakin Lek, spring 1987. (Hectares in each contour interval are: 10% = 2.68, 25% = 7.5, 50% = 19.17, 95% = 125.32). �141 22 ffJ. Rg. 2.7. Home range size of 1 female plains sharp-telleo grouse, Dakin Lek, spring (upper left)· Summer (lower right) 1987. Hectares in each contour interval are: 10% = 6.84, 25% = 10.69, 50% = 48.17, 95% = 139.13, 95% = 884.01). �142 22 • Fig. 2.8. Home range size of 2 male plains sharp-tailed grouse, Indian Meadows Lek, spring 1988. (Hectares in each contour interval are: 10% = 2.62, 25% = 10.97, 50% = 27.32,95% = 128.32). �143 24 Data from the Cherokee were used in analysis pooled. summer of use versus and Indian Meadows random sites. Feeding-loafing differed All variables (f < 0.05) from random (f < differed and roosting 0.05) sites sites during leks were no site and the data during spring and summer with strong height, (f selection width, and (Figs. 2.9-2.10). Lek and nesting from random sites sites during length and greater sites had greater length cover, distance canopy and less distance Escape sites and clump clump height, randomly differed (Figs. 2.9-2.10). sites. (f < Canopy and Nest width, and (Fig. 2.10). 0.05) from random cover, plant (f < 0.05) from random sites, but differed and length had a 1 in 4 probability were examined lek, and nest sites within and 1988 from Cherokee sharp-tailed grouse sites during among feeding-loafing, each study were pooled of occurring selected for similar spring when be similar sites, but differed feeding-loafing, cover escape, Data collected feeding-loafing, the best available Canopy between site. roosting, and clump width escape, during for use site comparisons. (Figs. 2.9-2.10). between width, clump height, sites Lek (Figs. 2.9-2.10). Differences escape than random than random and summer 2.9-2.10). clump height, cover (f < 0.05) differed (Figs. plant height, to clumps distance width, plant height, cover, in spring and summer measured to clump for 7 of 12 variables height, spring for all variables sites had less canopy sites (1988) for 11 of 12 variables. < 0.01) against random sites for clump distance, length There (f > 0.05) between years at the Cherokee differences were (1987-88) 1987 Plains roosting, and cover was shrubs and length appeared to (f < 0.05) for all variables and lek sites, but not for roosting. �f-' -l> ·1> -- 100 ffi 60 o 40 > o 1- 0 ~--.-- A FL AO E L 30 12 25 10 8 ~ ~ 6 4 w 20 14 a: C/) oR 80 a: w w I~ C/) 20 a: ~ 15 w ~ 0 A 12 251----- 0 A I. FL AO E 0 L E L A FL AO E L CLUMP LENGTH I :tlill A FL AO E 25 20 a: ~ 15 w ~ I L 10 5 0 N Vl C/) !::tlill 6 2 AO 30 30 4 FL CLUMP WIDTH 14 10 8 10 5 2 CLUMP HEIGHT C/) CLUMP DISTANCE PLANT HEIGHT CANOPY COVER A FL AO E Rg. 2.9. Structural variables at plains sharp-tailed grouse use and random sites in spring (Mar-May), Cherokee Lek. (R= random, n = 42; FL = feeding-loafing, n = 106; RO= roosting, n = 54; E = escape, n = 54; L = lek, n = 20). L �CLUMP DISTANCE PLANT HEIGHT CANOPY COVER 14 100 #. -a: w > 0 o I 80 60 40 20 12 (j) a: w 10 8 tD 6 ~ 4 (j) a: 2 _._----- 0 A 1 ---- FL AO E 0 NE a: w rw ~ - A FL ._ -AO E ~ w ~ 1 10 5 0 A NE 14 30 30' 12 25 25 10 8 4 2 0 20 (j) 20 ~ 15 ~ 15 ~ 10 ~ 10 w 6 A .11 FL AO E 5 NE 0·-----, A AO E NE IV o- I (j) a: FL CLUMP LENGTH CLUMP WIDTH CLUMP HEIGHT (j) 15 a: w I FL AO 5 E NE 0 A FL AD E NE Rg, 2.10. Structural variables at plains sharp-tailed grouse use and random sites in summer (Jun-Aug), Cherokee Lek. (A = random, n = 101; FL = feeding-loafing, n = 144; AO = roosting, n = 51; E = escape, n = 69; NE = nest, n = 5). r-" -I> \..I, �146 27 Canopy cover and clump width intermediate tailed between grouse width, loafing and roosting comparisons Indian site. Escape sites were aromatica, closer clumps usually clump for feeding- (E < Lek sites had less All other seasonal marked at Cherokee, Dakin, of Stipa dominated feeding-loafing comata and Yucca by Cercocarpus glauca. montanus, than any other use site and were within clumps (3 nests) gambelii at Nest dead vegetation Quercus and communities. or Svrnphoricarpos albus (2 nests) 2.11-2.12). sharp-tailed changed grouse true mountainmahogany forbs. In summer, grouse gambel oak for escape spring and summer categories. vegetation predators. for escape sites selected cover. that could from spring dominated habitat sites with to clumped and length (Figs. 2.9-2.10). as by tall shrubs by grasses true mountainmahogany vegetation and or (E > 0.05) between among grouse were not more provide to summer dominated There was no difference sharp-tailed Clump width from habitat to summer for distance Plains dramatically moved through clumped Sharp- (E < 0.05), except for clump distance or less frequently Use site selection plains distances sites. by birds to large cover and escape. plant height, sites were near or within in areas sites had greater of standing clump sites were < 0.05). Roosting but were (Figs. CE leks differed sites, Rhus cover, than all other use sites. use site selection Meadows the Dakin and greater than at escape differed Summer for lower canopy and length 0.05) vegetation at roosting those used for feeding-loafing selected height, and length at least minimal also did not differ use than 14 m from cover between (50%) from seasons �a: ~ 100 4 w 60 en a: .,_ 0 40 ~ > 80 w w 0 20 0' en a: .,_ 2 1.1 , FL AD 0 E 25 3 1 ~~A 30 _I 5 ~ - CLUMP DISTANCE PLANT HEIGHT CANOPY COVER FL A AD w w ~ 20 15 10 5 0 A E FL AD E N CLUMP HEIGHT 30 - I 25~_ en a: .,_ w w ~ CLUMP WIDTH ::[1 30 30 25 25 en .,_a: 20 w w 15 ~ 10 5 5 0 01--A FL AD E CLUMP LENGTH I. A FL AD en a: .,_ 20 w w 15 ~ 10 ('C 5 I E 0 A FL AD E Rg. 2.11. Structural variables at plains sharp-tailed grouse use and random sites in summer (Jun-Aug), Indian Meadows Lek. (A = random, n = 88; FL = feeding-loafing, n = 46; AD = roosting, n = 39, E = escape, n = 66). t·1> -I �I-' ~ 00 CLUMP DISTANCE PLANT HEIGHT CANOPY COVER 4 -cfffi > o 100 3 80 C/) 60 w tw 40 ~ a: 0 ~ 0 FL AO .1 0 E I I 15 a: w t- w ~ I. FL AO ~ E 10 30 25 25 20 C/) 20 ~ 15 ~ 15 a: W Of FL --I AO E ~ FL AO E CLUMP LENGTH 30 C/) a: 5 10 8 6 4 2 0 CLUMP WIDTH CLUMP HEIGHT 20 w tw 2 20 C/) 16 14 12 N '0 W ~ 10 5 0 10 5 FL AO E 0 FL AO E Rg. 2.12. Structural variables at plains sharp-tailed grouse use and random sites in summer (Jun-Aug), Dakin Lek. (FL = feeding-loafing, n = 43; AO = roosting, n = 34; E = escape, n = 25). �30 Use versus random sites at the ungrazed (f < 0.05) for 17 of 18 structural variables differed Feeding-loafing areas had greater clump height, distance, sites were typically dense, with random sites. standing versus Escape and width, longer, during and plant than random and closer sites had greater for comparison sites feeding-loafing, (Tables canopy and closer height areas, and less Roosting but were vegetation cover, clumped summer. sites. clumped roosting, higher less than at plant and vegetation than did (grass/forb, shrub, by standing areas as escape for feeding-loafing sites. spring Trees were plants at use grouse at the Shrubs grouse selected and roosting, an available were for and shrub- cover group, at any of the 3 sites. sharp-tailed grouse than was available selected at random They also selected sites and for shrubs tailed dead vegetation. In swruner, sharp-tailed by grasses nesting-brooding. or clump sites during as escape but were not selected plants and escape selected areas dominated of dominant Plains-sharp site were dominated sites. into 3 groups 2.4-2.5). Cherokee vegetation and length at use sites were pooled random Plains cover area (Fig. 2.11). dead) dominated canopy on the edge of feeding-loafing and wider, sites Plants width, less height clump height, random Indian Meadows for areas with more sites against standing for feeding-loafing shrubs at the 2 Dakin nesting/brooding and at feeding-loafing sites (Table 2.6). DISCUSSION Plains seasonal sharp-tailed movements cover used grouse have been between in winter grassland (Hamerstrom reported areas used and Hamerstrom to make extensive for breeding 1951). and woody Use of land for �150 31 Table 2.4. Dominant plant species use and random sites. (%) at plains sharp-tailed grouse Site Category Standing Grasses/forbs Shrubs Feeding-loafing 41 18 41 Roosting 42 2 56 Escape 36 27 37 Feeding-loafing 86 3 11 Roosting 83 2 15 Escape 53 42 5 76 19 5 Feeding-loafing 87 10 3 Roosting 69 16 15 Escape 53 37 10 Feeding-loafing 83 2 15 Roosting 83 5 12 0 100 0 71 14 15 Cherokee Spring Sununer Random Dakin Indian Meadows Escape Random dead �lSl 32 Table 2.5. Clump species random sites. (%) at plains sharp-tailed grouse use and Site Category Standing Grasses/forbs Shrubs Feeding-loafing 25 43 32 Roosting 28 16 56 Escape 18 72 10 Feeding-loafing 76 13 11 Roosting 57 28 15 Escape 28 67 5 72 21 7 Feeding-loafing 77 20 3 Roosting 38 46 15 Escape 57 33 10 Cherokee Spring Sununer Random Dakin Indian Meadows Feeding-loafing 84 Roosting 72 16 12 o 100 o 66 16 18 Escape Random 1 15 dead �I---' l./l N Table 2.6. Plant group cover (%) at feeding-loafing Annual Category Site and nest sites of plains sharp-tailed Perennial Forbs Grasses grouse. Shrubs Standing dead Bare ground 5.4 1.2 1.0 5.6 33.0 13.6 10.6 3.0 42.5 44.6 31.1 11.2 0.0 9.5 15.4 10.7 13 .1 0.0 36.2 37.2 15.7 16.9 0.0 3.7 37.0 23.3 0.0 0.0 Grasses Forbs Cherokee Random 12.0 20.2 8.4 6.1 30.7 50.0 Indian Meadows Random 2.9 17.0 7.1 6.2 4.5 7.9 8.3 12.2 Feeding-i.oafing Dakin Random Nest w w Cherokee Cherokee Cherokee Random Dakin 1 Dakin 2 Random 1 2 3 16.8 2.2 3.1 20.2 16.3 9.3 7.7 6.1 28.9 47.7 33.1 50.0 17.4 11.0 24.4 1.2 2.9 6.4 6.4 5.6 19.5 20.1 24.3 13.6 2.0 3.2 4.1 2.9 4.0 9.8 4.5 11.0 3.2 8.3 40.2 30.4 36.2 2l. 7 18.1 15.7 8.0 0.1 0.0 19.0 40.0 37.0 0.0 0.0 0.0 �153 34 agriculture has reduced and reduced their extensive 1968). Spring, size of areas summer, grouse in Montana 1982). Movements were movements generally Three grouse summer habitat Lek was separated grassland without at Lincoln Sharp-tailed summer home ranges through 1980). averaged The daily campestris) with birds Nebraska cruising rarely between ranged The Cherokee, within Dakin, daily use habitats, contiguity populations The Lincoln by intensively grazed sites moved Dakota during with from winter a maximum grouse nests to sharp-tailed Plains meters area throughout and Indian Meadows may be a key factor at each lek site in Douglas woody grouse cover 1976). seasonal The extent for sharp-tailed 1.6 km in (Sisson sites had distinct County. of 3.2 km. was about the year but all sites were contiguous. of 3.2 km (I. £. from heavy sharp-tailed was 0.9 km in Douglas distance fall and winter than a few hundred a 4.8-km habitat. and leks in Saskatchewan of prairie 1951). birds shrub/grassland. sharp-tailed radius and Hamerstrom available. the lek and winter and Dakin nests in Wisconsin more into data were 1.2 km from the lek with a maximum grouse (Hamerstrom habitat between contiguous Five plains habitat varied habitat. This may be why all radio-marked 1972) and was 1.3 km in North CKobriger County from winter and Yde County habitat and summer habitat. at the Cherokee The mean distance (Pepper winter (Nielsen to spring breeding winter and Wood sharp-tailed in Douglas at the 4 leks for which were predated grouse of male grouse grouse (Johnsgard 1.6 km of leks from oakbrush shrub cover. Mountain within habitat moved sites had contiguous Mountain or migrations sharp-tailed from 1.6 to 4.8 km from winter grass/forb to sharp-tailed and fall distributions of plains Sharp-tailed available grouse of and �154 35 Good quality tailed grouse communities, ecotones particularly apparently important and brushy cover are essential (Hillman anc Jackson (Moyles 1981). habitat grassland grassland-shrub provide Interspersion mixtures optimum habitat composition in determining loafing, roosting, all birds marked The average and lek (grassland escape, height of all vegetation plant height In Douglas Swenson areas) sites (40-80%) for of Nebraska use sites sites (70-171 cm) than available cover (Hamerstrom 1981, Nielsen to and structural (1981) reported sharp-tailed 750 m of cropland. grouse in Wisconsin 50% woody canopy cover. in the Sandhills County, plains sharp-tailed 1981, Swenson the distance woody cover within sharp-tailed Cover at feeding- Many authors cite the occurrence grouse near shrubland 1951, Sisson 1976, Moyles cover. grouse in (33-60 cm) for grassland at shrubland at random sites (15-112 cm). few investigated grouse at all lek sites and seasons. grouse used areas with less height sharp-tailed sharp-tailed more areas) sites (13-43%) was less and nest (shrubland was 21.2 cm (Sisson 1976). and greater of Alberta are generally sharp-tailed County varied by type of use and season. than at random, with extensive (West 1961). The range in canopy cover used by plains Douglas of plant in the parklands Height and density of vegetation than species quality 1973). for sharp- Grange probably of plains and Hamerstrom and Yde 1982), but parameters of shrub land grouse used 100% of the (1948) reported prairie could not tolerate more than Other authors have reported minimUm amounts of shrub cover as 1.5% (Janson 1953), 1-4% (Brown 1968), and 5% (Edminster 1954). �155 36 Shrubland necessary cover was used mainly cover within sharp-tailed grouse use sites require themselves? clumped to escape predators not greater than 14 m for all use site measurements clump vegetation was greater about I hypothesized vegetation to flushing. but what approaching where prior for escape, Distance that plains from to clump except than 50 m distant the cover was lek sites (Cherokee Lek). Grassland use sites had a range of clump distance of 0.8 to 5.8 m while shrubland use sites had an average distance of 0 to 1.5 m. Distance o to clumps at random sites clump ranged from 1.3 to 10 m in spring and to 8 m in summer. Plains structural select were closely scattered cover 1976, Moyles herbaceous in early spring Janson habitat over a 6-year Random plant groups cover), tailed of woody to be 10% shrubs 3.5% weedy proportions grass/forbs and standing dead (5-19% cover). County 1982). for nesting (Blus and Walker sharp-tailed Typical grouse habitat consisted and 1.5% woody at 4 study sites demonstrated were in Douglas also types cover. for sharp- and 90% grassland. present grouse and Yde constructed of cover also grouse cover were component cover, grouse and summer, 1981, Nielsen (1955) evaluated for sharp-tailed areas in South Dakota. optimum measurements report spring when most nests were (1981) believed grouse Do sharp-tailed was an important 21% cropland, selected during 1981, Swenson (1953) and Podoll of 74% grassland, tailed grassland and adjacent period County Many authors vegetation 1966). Swenson with shrubs in Douglas at use sites. communities? associated (Sisson Residual grouse characteristics for plant although used sharp-tailed (76-83% selected cover), In spring, that dominant shrubs plains for 41% grass/forb, (16-19% sharp16% shrub, �156 37 and 43% standing dead cover. used feeding-loafing and 12% standing grass/forb, in Douglas County and roosting sites with 82% grass/forb, dead cover. Sites used for escape averaged vegetation. was quite different Feeding-loafing of 27% grass/forbs, and roosting from dominant 10% standing 30% shrubs, and 43% standing grass/forbs, standing plains dead cover. Clump vegetation sites used for feeding-loafing vegetation 35% in summer was available sharp-tailed with clumped in with an average of 68% had an average of 29% grass/forbs, Dominant dead cover, 72% shrubs, and 21% shrubs, and 11% standing dead cover. dead. plant sites during spring had an while escape sites had an average of 18% grass/forbs, grassland 6% shrubs, 60% shrubs, and 5% standing dead vegetation. Clump cover vegetation average In summer, sharptails Escape site clump 70% shrubs, and 1% plant group and clump group data indicate grouse selected or shrub vegetation for both grassland nearby « that and shrub sites 14 m) for initial escape cover. Of the 4 sites studied, closest agreement the literature. the Dakin and Indian Meadows in structural sites had and plant group data to that reported Both were ungrazed or lightly grazed. Birds at these 2 sites also had smaller home ranges than those at the Cherokee Cherokee site had a patchy distribution lightly grazed areas interspersed mountainmahogany. habitat with areas dominated site. moderate, The and by oakbrush and Summer home ranges were larger, and structural and plant group variables had less agreement other states with plains sharp-tailed Data on plains sharp-tailed similar of intensively, in to other studies with data from grouse. grouse in Douglas County, Colorado were in home range size, distance moved between and �157 38 within seasons, canopy cover, plant height, Lincoln Mountain site differed grouse moved between and plant groups. The from other studies in distance that seasons due to the discontiguity Availability of habitat transitional habitat dominated by grass/forbs of habitat. and shrubs with close by may be most critical to plains sharp- tailed grouse in Douglas County, Colorado. There was adequate habitat for selection escape, and leks at the Cherokee, of feeding/loafing, roosting, Dakin, and Indian Meadows sites. identified in habitat Five nesting similar to that Sisson sites were also (1976) reported in Nebraska. Numbers of sharp-tailed grouse lek sites in Douglas County appear to be stable, but lower than in other states. of sharp-tailed have interchange The patchy distribution grouse in Douglas County suggests between the population lek sites or that intervening habitats may not are unsuitable. If true, loss of a lek site could result in loss of plains sharp-tailed grouse using that particular site. LITERATURE CITED Aiken, C. E. H., and E. R. Warren. 1914. The birds of El Paso County, Colorado. Colorado Coll. Sci. Series 12:455-603. Aldrich, J. W. 1963. Geographic orientation of American Tetraonidae. J. Wildl. Manage. 27:529-545. Ammann, G. A. 1944. Determining age of pinnated grouse. J. Wildl. Manage. 8:170-171. Amstrup, S. C. 1980. A radio-collar Manage. 44:214-217. and sharp-tailed for game birds. J. Wildl. Bailey, A. M., and R. J. Niedrach. 1965. Birds of Colorado. Denver Mus. Nat. Hist., Denver, Colo. 454pp. Vol. 1. Blus, L. J., and J. A. Walker. 1966. Progress reports on the prairie grouse nesting study in the Nebraska Sandhills. Nebraska Bird Rev. 34(2) :23-30. �158 39 Brown, R. L. 1968. Effects of land-use practices on sharp-tailed grouse. Montana Dep. Fish and Game, Job Completion Rep., Fed. Aid Proj. W-9l-R-9, Job II-F. llpp. Cooke, W. W. 1897. 37. l44pp. The birds of Colorado. Colorado Dixon, K. R., and J. A. Chapman. 1980. Harmonic activity area. Ecology 61:1040-1044. Edminster, F. C. 1954. Charles Scribner's Agric. Coll. mean measure American game birds of field Sons, New York, N.Y. 490pp. Bull. of animal and forest. Giesen, K. M., C. E. Braun, and T. J. Schoenberg. 1982. Methods for trapping sage grouse in Colorado. Wildl. Soc. Bull. 10:224-231. Grange, W. B. 1948. Wisconsin grouse problems. Conserv. Dep., Madison. Publ. 328. 3l8pp. Hamerstrom, F. N., and F. Hamerstrom. 1951. tailed grouse in relation to its ecology Midl. Nat. 46:174-226. Wisconsin Mobility of the sharpand distribution. Am. Henderson, F. R., F. W. Brooks, R. E. Wood, and R. B. Dahlgren. 1967. Sexing of prairie grouse by crown feather patterns. J. Wildl. Manage. 31:764-769. Henderson, J. Colorado. 1909. Univ. An annotated list of birds of Boulder Colorado Studies 6:219-242. Hillman, C. N., and W. W. Jackson. in South Dakota. South Dakota Tech. Bull. 3. 64pp. County, 1973. The sharp-tailed grouse Div. Game, Fish and Parks. Janson, R. 1953. Prairie grouse habitat survey, 1950-1952. South Dakota Dep. Game, Fish and Parks Job Completion Rep. P-R Proj. W17-R-7. llpp. Johnsgard, P. A., and R. E. Wood. 1968. interaction between prairie chickens midwest. Wilson Bull. 80:173-188. Distributional changes and and sharp-tailed grouse in the Kobriger, G. D. 1980. Habitat use by nesting and brooding sharptailed grouse in southwestern North Dakota. North Dakota Outdoors 43(1):2-6. Mielke, P. W. 1979. On asymptotic non-normality of MRPP statistics. Commun. Statist. Theor. of null distributions Meth. A8: 1541-1550. _____ , K. J. Berry, and E. S. Johnson. 1976. MUlti-response permutation procedures for a priori classifications. Commun. Statist. Theor. Meth. A5: 1409-1424. �159 40 Miller, G. C., and W. D. Graul. 1980. Status of sharp-tailed grouse in North America. Pages 18-28 in P.A. Vohs, and F. L. Knopf, eds. Proc. Prairie Grouse Symp. Oklahoma State Univ., Stillwater. Moyles, D. L. J. 1981. Seasonal and daily use of plant communities by sharp-tailed grouse (Pedioecetes phasianellus) in the parklands of Alberta. Can. Field-Nat. 95:287-291. Nielsen, L. S., and C. A. Yde. 1982. The effects of rest-rotation grazing on the distribution of sharp-tailed grouse. Pages 147-165 in J.M. Peek and P.D. Dalke, eds. Proc. Wildl.-Livestock Relations Symp., For., Wildl., Range Exp. Stn., Univ. Idaho, Moscow. Pepper, G. W. 1972. The ecology of sharp-tailed grouse during spring and summer in the aspen parklands of Saskatchewan. Saskatchewan Dep. Nat. Resour. Wildl. Rep. 1. 56pp. Podoll, E. 1955. Prairie Grouse habitat study, 1953-1955. South Dakota Dep. Game, Fish and Parks Job Completion Rep. P-R Proj. W17-R-10. 10pp. Sclater, W. H. 1912. A history Co., London, U.K. 576pp. Sisson, L. study. of the birds of Colorado. Witherby and 1976. The sharp-tailed grouse in Nebraska: a research Nebraska Game and Parks Comm., Lincoln. 88pp. Stuwe, M., and C. E. Blowhowiak. 1985. Micro-computer analysis animal locations. Conserv. Res. Center, Natl. Zool. Park, Smithsonian Inst., Washington, D.C. 21pp. for Swenson, J. E. 1981. The hardwood draws of southeastern Montana: their importance to wildlife and vulnerability to man's activities. Pages 37-61 in J. H. Ormiston, ed. Management of riparian ecosystems. Proc. 1981 Meeting, Montana Chapter, The Wildl. Soc. U.S. Department Colorado. 129pp. West, D. R. prairie Wagner, of Agriculture. 1974. Soil survey of Castle Rock area, U.S. Dep. Agric., Soil Conserv. Serv., Washington, D.C. 1961. Thoughts grouse habitat. Okla. 10pp. on factors influencing the quality of 12th Annu. Great Plains Habitat Conf., �160 Chapter MANAGEMENT RECOMMENDATIONS 3 FOR MAINTENANCE OF PLAINS SHARP-TAILED GROUSE IN DOUGLAS COUNTY. COLORADO ANALYSIS Results tailed grouse appear of this study indicate (Tympanuchus that populations phasianellus to be stable at presently jamesii) occupied of plains sharp- in Douglas County However. occupied sites. sites are isolated with little, if any. interchange management sites should be to insure preservation of existing occupied of lek sites and to maintain sharp-tailed slightly snowberry grazed pastures sites were in ungrazed areas dominated spp.). dead vegetation or or clumps of Two of 5 nests were in areas used in and roosting. to moderately by shrubs unbrowsed to heavily to ensure (5) found were in ungrazed in either standing (Symphoricarpos habitats Thus. remain stable or increase. grouse nests summer for feeding/loafing ungrazed and improve surrounding grouse populations All sharp-tailed among sites. Feeding. grazed pastures by livestock. grazed pastures loafing. and roosting with escape to Lek sites were in with shrubs within 200 m for escape cover. The need of plains sharp-tailed and density, status within and standing Douglas Ideal management from grazing grouse for adequate dead vegetation County suggests for sharp-tailed plant height coupled with their isolated a complex management problem. grouse would appear to be retirement of several 40-ha blocks near leks to provide nesting sites, �161 42 and implementation High use areas between of light for sharp-tailed among the contiguity average 2.3 km apart km apart 11 km apart. number This of active 1986-88. attain greater suggests lek sites By lightly greater plant standing management height areas with is possible sharp-tailed local population The minimum documented should be to maintain interchange establish leks and birds a population attending of Wildlife should maintaining populations would tailed grouse sites. easements purchase populations. to 6 in and spring. lek site will If grazing (into and brushy into these areas draws), and thus, it the over time. management The Colorado system sharp-tailed of suitable in Douglas lek sites with Division of Wildlife to document basis. landowners on areas critical (Kahn 1979) winter each on an annual of plains Also, and fall with of 6 active monitoring lek sites were summer grouse a minimum also work with be to investigate conservation among County during stability on 1973) and 4.8 of habitat. of grassland goal of sharp-tailed were vegetation grouse will move greater states on areas, the immediate interspersion may have plains from 10 in 1979 during in size dependent and Jackson or by not grazing in areas beyond surrounding (Hillman and density ranged is largely a lack of contiguity grazing use areas. and seasons. leks in other decreased grouse County 1976), but in Douglas dead vegetation occurs should Active (Sisson in Douglas or use sites in South Dakota in Nebraska in high use sites lek sites, of habitat. systems grouse 16 and 68 ha for different Interchange County grazing number The Colorado to generate grouse. properties to maintenance Division interest Other of in approaches or to acquire of plains sharp- �162 43 RECOMMENDATIONS All active during lek sites should be surveyed peak attendance pastures ranch, within pasture winter every 2 years for active Ranch.--Livestock Cattle should be on a 4 pasture grazed. The stocking grazing supplement food resources. should be mapped monitored limiting The distribution and sharp-tailed to understand rest rotation grouse use patterns to investigate Kimball The pasture use, especially winter by hens. of cattle. more residual vegetation Another these areas. to examine and learn if shrubs are estate A buffer pasture suitability southeast during for sharp-tailed the grouse shrubs) between west of the lek should be leased zone around each excluded cattle on a rest rotation for cover. should be should be allowed 65 ha (containing The Kimball Perry Park Road and the Curtis property be surveyed of the shrub community should be monitored should be created by grazing to on the ranch. No grazing and summer use sites immediately for exclusion the lek site. of the 65-ha ungrazed of the lek should be continued. lease period. 3 ha per could be continued The Frank Woodhouse land use options Property.--Leasing basis with one grouse use of the shrub community (too many or too few). contacted from two 65-ha rate should not exceed of small grain crops on bottomlands use on the rest of the animal unit month on one half of the area surrounding Planting Potential leks. should be excluded 1 km of the lek site. if allowed, of birds present (late Apr or early May) each year. areas should be surveyed Bob Woodhouse for number basis property to leave west of the south of the Cherokee for expansion pasture Lek should of grouse use into �163 44 Pine Cliffe Ranch.--The should be reduced number of cattle grazing and ungrazed Dakin Lek Property.--This with grazing to be excluded. be leased and grazing exclosures property (~ 40 ha) created. should be leased or purchased Pastures reduced this property north of the lek site should also to expand the range of the population. area (100 x 100 m) in the center of the site could be mowed spring to create an artificial Greenland-Lincoln of plains through possibility of expanding numbers grouse in this area should be investigated grazing management timing of grazing). in early lek center. Mountain.--The sharp-tailed An (reduction in grazing Grass and shrub communities for large populations intensity, appear changes in to be suitable of grouse, but only a few exist. Indian Meadows.--Leasing should be investigated. of State school land southwest The intent would be to reduce of the lek livestock numbers. Highlands population habitat and Winkler surveys inventory Further conducted selection, Ranches.--Further sharp-tailed of these areas should be conducted for consideration Research.--If on sharp-tailed as possible along with a reintroduction sites can be leased, research grouse habitat among leased pastures selection, and pastures grouse should be especially managed sites. nest site on a rest rotation basis. LITERATURE CITED 1973. The sharp-tailed grouse in Hillman, C. N., and W. W. Jackson. South Dakota Div. Game, Fish and Parks Tech. Bull. South Dakota. 3. 64pp. Kahn, R. 1979. Prairie sharp-tailed grouse segment. Colorado Div. Wildl., Denver. l2pp. Unpubl. Rep., �164 Sisson. L. 1976. The sharp-tailed grouse in Nebraska: a research study. Nebraska Game and Parks Comm., Lincoln. 88pp. Prepared by An~ag~~ Graduate Research Assistant Approved by __ &a.;;.._=~.:;..·_ _;;;Z;;._,:__ --,-~:::___;,.;;;.__ C1ait E. Braun Wildlife Research Leader �165 INTERIM FINAL REPORT ,. .. .; State of: Colorado Project: W-152-R Upland Bird Research Work Plan: 14 Job _3_ Job Title: Seasonal Movement Period Covered: 01 january Author: A. Schroeder Michael Personnel: and Habitat Use by Greater Prairie-Chickens through 31 December 1989 M. A. Schroeder and G. C. White, Colorado Braun, Colorado Division of Wildlife State University; C. E. ABSTRACT Several hypothesis have been proposed to explain the evolution of clumped (lek mating) from dispersed (territorial mating) polygyny; most recent models suggest that increased female home range size leads to increased female choice for males and/or leks. I propose an alternative, the female tolerance hypothesis, in which decreased intrasexual aggression among females during the breeding season results in increased aggregation of males. Predictions of the female preference and hotspot hypotheses were examined in a population of greater prairie-chickens (Tympanuchus cupido) in northeastern Colorado during 1986-88. Nest-1ek distances were used as indirect measures of home range size. Sixty-six of 89 females (74%) nested closer to a lek other than where captured and 67 of 79 females (85%) visited >1 lek during the breeding season. These results contradicted predictions of the female preference hypothesis. Breeding potential was estimated as a measure of proximity between any given location and nest locations of females. Monte Carlo simulations were used to examine breeding potentials under varying conditions of actual and random lek locations and actual and random distribution of nest locations. Distributions of both leks and nests suppo r t ed .pre9:ictionsof the hotspot hypothesis. Examination of seasonal movement between breeding and wintering ranges indicated that spring migration occurred during February-March and autumn migration occurred during June-August. Most of the variability in timing-of autumn migration for females appeared to be due to their brood status; females without broods migrated earlier than those with broods. The average migration distance between breeding and wintering ranges was 10.6 km for females and 2.9 km for males. Examination of seasonal locations for individuals between years indicated that most greater prairie-chickens displayed site fidelity to both breeding and winter areas. A lek attendance rate of 50% for males has been used to estimate population size for greater prairie-chickens; consequently lek attendance was examined in northeastern Colorado during 1986-90. Observations of 21 radio-marked males �166 indicated they were on leks 95.1% of the time during peak display periods (Mar-Apr); consequently use of 50% as a value for 1ek attendance may result in over-estimates of population size. Mean annual turnover rate of leks was 23.8%; 24 leks were active aIlS years. A dissertation incorporating 5 manuscripts for publication has been submitted to fulfill requirements for the Doctor of Philosophy degree in Fishery and Wildlife Biology. Prepared by ~~~i1~~~J,l~U~ Michael A. Schroeder Graduate Research Assistant Approved &~~=.:...< by _ ..•... =2.:;...:.__._~~;_'....;....:...:=--Clait E. Braun Wildlife Research Leader _ �1..l1 JOB PROGRESS REPORT Colorado State of: Project: W-l52-R Upland Bird Research Work Plan: 17 Job Title: Population Dynamics of White-tailed Ptarmigan Period Covered: Author: : Job _7_ 01 January through 31 December 1989 Clait E. Braun and Kenneth M. Giesen Personnel: Clait e. Braun and Kenneth M. Giesen, Colorado Division of Wildlife ABSTRACT Long-term studies of populations of white-tailed ptarmigan (Lagopus leucurus) were continued at hunted (Mt. Evans) and unhunted (Rocky Mountain National Park) areas in Colorado through 1989. Densities of breeding ptarmigan increased at Mt. Evans and at Rocky Mountain National Park. Nesting success at Mt. Evans was average while it was poor at Rocky Mountain National Park. At least 11% (10.7) of the ptarmigan leg-banded in 1989 at Mt. Evans were harvested during the fall hunting season. ��169 POPULATION DYNAMICS OF WHITE-TAILED PTARMIGAN Clait E. Braun and Kenneth M. Giesen Long-term studies of trends in population size and investigation of reasons for fluctuations in size of tetraonid populations are lacking. Studies on the population dynamics of unhunted and hunted populations of white-tailed ptarmigan were initiated in Colorado in 1966 and have continued essentially uninterrupted at 2 sites. Studies of the unhunted population (Rocky Mountain National Park) identified possible short-term cycles of 7-8 years with an amplitude of 25-30% between high and low breeding densities. Conversely, studies of the manipulated population (hunted) at Mt. Evans have not indicated any cyclic pattern and it would appear that controlled hunting may mask any long-term trend that may occur. This study is designed to examine the question whether white-tailed ptarmigan are truly cyclic and whether hunting affects the apparent oscillations. P. N. OBJECTIVES The goals of this investigation are to be able to predict the length and amplitude of cycles in white-tailed ptarmigan in Colorado, to examine the impact of hunting on cycles, and to clarify underlying causes of the apparent cycles. SEGMENT OBJECTIVES 1. Conduct breeding (May-Jun) and brood (Aug-Sep) censuses ptarmigan using tape-recorded calls of males (breeding) (broods) . 2. Censuses will be conducted on previously established, defined study areas at Mt. Evans (hunted) and at Rocky Mountain National Park (unhunted). 3. Capture (noose poles) and band (aluminum and plastic color-coded bands) all unmarked white-tailed ptarmigan encountered on study areas at Mt. Evans and at Rocky Mountain National Park. 4. Individually identify all ptarmigan observed on study areas at Mt. Evans and Rocky Mountain National Park through use of binoculars. 5. Make hunting season and bag limit recommendations for Mt. Evans and collect hunting data through use of volunteer wing barrels and hunter field checks. 6. Compile data, analyze results, and prepare progress of white-tailed and chicks reports. �STUDY AREA AND METHODS Areas investigated were Mt. Goliath-Mt. Evans in Clear Creek County and at Tombstone Ridge-Sundance Mountain to Fall River Pass in Rocky Mountain National Park in Larimer County. The physiography, geology, location, and vegetation of these study areas have been previously described (Braun 1969, 1971; Braun and Rogers 1971; Giesen 1977). Ptarmigan were located through use of tape-recorded calls (Braun et al. 1973), captured through use of telescoping noose poles (Zwickel and Bendell 1967) as described by Braun and Rogers (1971), and classified to age and sex and banded following Braun and Rogers (1971). Age of chicks was estimated following Giesen and Braun (1979). Numbered plastic bandettes were not used as in earlier years (Braun and Rogers 1971) as a color-code system using up to 4 different colored plastic bandettes was instituted in 1977-78. A check station was operated on the Mt. Evans highway during the opening weekend of the ptarmigan season in that area. A volunteer wing collection station was available to hunters in the area when the check station was not in operation until the season closed. RESULTS AND DISCUSSION Breeding Densities Mt. Evans.--Timing of breeding events in the Mt. Evans area was about the same in 1989 as in 1988. During the May-early June. interval, 15 pairs and 2 single males were identified. Thus, breeding densities increased slightly in 1989 (Table 1). During the breeding season,S of 17 males identified were yearlings while 9 of 15 hens were yearlings. Recruitment of yearlings was lower than in 1988. Rocky Mountain National Park.--Timing of breeding events on the Trail Ridge study area was similar to that in 1988 but 1 week earlier than the long-term (1966-88) average. Surveys of ptarmigan on breeding territories along Trail Ridge Road in RMNP in May and June indicated a minimum population of 67 birds and included 29 pairs and 9 unpaired males. This represents a 34% increase from 1988 and the highest population since 1982. However, this population level is only about half that recorded during peak years in 1969 and 1976 (Table 1). The increased breeding density reflected above average survival of banded adult males from 1988 (31 of 45, 68.9%) and excellent production of chicks in 1988 and their recruitment to 1989. Eight of 25 chicks banded in 1988 returned (32.0%) and yearlings comprised 46.1% of all adult ptarmigan identified in 1989. Nesting Success and Brood Size Mt. Evans.--Twenty-three hens were located during mid July-early September 1989 on or immediately adjacent to the study area. Eleven hens (47.8%) were with broods while 12 were apparently unsuccessful nesters (without chicks). Average brood size to 1 September was good (4.0 chicksfhen). Data from 29 chicks that were banded indicated hatch dates from 27 June to 8 August with only 17% hatching after 16 July. �Rocky Mountain National Park.--Nest success was estimated at <30% as only 4 of 14 hens observed in July-August were with chicks. The low nest success may be partly attributed to hen inexperience as yearlings comprised 62.5% of all hens identified in 1989. Hatch dates calculated from primary molt of chicks indicated the median hatch date was 6 July (range 2-11 Jul). Average brood size was 3.0 chicks(hen. Table 1. 1966-89. White-tailed ptarmigan breeding densities (birds/km2), Colorado Study area Year Rocky Mountain National Park (5.5 km2) Mt. Evans (4.0 km2) 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 11.3 9.8 11.5 12.0 9.6 9.1 8.7 7.8 8.0 11.1 13.5 12.9 10.7 8.7 8.4 8.2 7.8 6.7 5.8 6.0 4.5 6.0 5.4 6.2 3.0 2.7 2.7 2.2 2.0 4.2 7.5 6.2 6.2 6.2 6.7 > 6.0 7.5 10.3 9.5 9.0 6.5 6.5 8.0 8.0 6.5 5.0 7.5 8.0 Harvest Mt, !v.n.,--The hunting season at Mt. Evans in 1989 opened on 16 September and closed on 8 October (23 days) with a bag and possession limit of 3 and 6. Thus, the season was delayed 1 week from the statewide opening as it was in 1981 and 1986-88. The season opening was delayed 2 weeks from 1978 to 1980 and 1982 to 1985, Prior to 1978, experimental seasons were in effect (19701976) or the season opened with the statewide grouse seasons (dates from 17 �Aug to 14 Sep). Unlike 1987 and 1988 when the Kt. Evans road was closed due to reconstruction (Lincoln Lake area in 1987, above Ptarmigan Flats in 1988), the Mt. Evans road was open to Summit Lake from 16 to 29 September when it closed because of snow. Only 6 hunters were checked on 16-17 September. These hunters observed 11 ptarmigan and harvested 8 of which all were banded. Two additional wings (1 wing tag) were received on 25 September in the wing barrel and legs from 3 banded ptarmigan from Mt. Evans were received in the Loveland Pass wing barrel on 30 September. Thus, a minimum of 13 ptarmigan were harvested at Mt. Evans of which 12 (2 wing-tagged only) were banded. Six of the banded birds harvested were banded in 1990 (6/56 - 10.7%) of which 4 were chicks (4/28 - 14.3%). The other banded birds harvested were banded in 1984 (1), 1985 (1), 1986 (1), and 1988 (1). LITERATURE CITED Braun, C. E. 1969. Population dynamics, habitat, and movements of whitetailed ptarmigan in Colorado. Ph.D. Thesis, Colorado State Univ., Fort Collins. l89pp. 1971. Habitat requirements of Colorado white-tailed ptarmigan. West. Assoc. State Game and Fish Comm. 51:284-292. Proc. _____ , and G. E. Rogers. 1971. The white-tailed ptarmigan in Colorado. Colorado Div. Game, Fish and Parks Tech. Pub1. 27. 80pp. _____ , R. K. Schmidt, Jr., and G. E. Rogers. 1973. Census of Colorado whitetailed ptarmigan with tape recorded calls. J. Wildl. Manage. 37:90-93. Giesen, K. M. 1977. Mortality and dispersal of juvenile white-tailed ptarmigan. M.S. Thesis, Colorado State Univ., Fort Collins. 55pp. _____ , and C. E. Braun. 1979. A technique for age determination of juvenile white-tailed ptarmigan. J. Wildl. Manage. 43:508-511. Zwickel, F. C., and J. F. Bendell. J. Wildl. Manage. 31:202-204. 1967. Prepared by Clait E. Braun Wildlife Research Leader Ke~iesen Wildlife Research B A snare for capturing blue grouse. �173 JOB PROGRESS State of: Colorado Project: W-152-R Upland Work Plan: 21 Job Title: Sandsage-B1uestem Period Covered: Author: Personnel: REPORT Warren Bird Research : Job _3_ 01 January Prairie through Renovation 31 December 1989 D. Snyder C. Braun, W. Brown, L. Crooks, T. Davis, J. Heinemann, L. Robb, M. Schroeder and W. Snyder, Colorado Division Wildlife S. Posner, of ABSTRACT Extremely dry weather until mid-May 1989 delayed vegetation growth and reduced soil moisture. Above average precipitation in June, August, and September stimulated vegetation recovery, however, it was extremely dry through fall and early winter. Vegetation phenology through spring was similar to that in 1987-88. Due to favorable precipitation in 1988, the height-density index (HDI) of residual grass-forb vegetation increased similarly on both burned and control transects from 1988 to 1989 on all except burn 1-84. Impacts of prescribed burns could no longer be detected. Sandsage (Artemisia filifolia) had not fully recovered from fire by early spring 1989 on any of the burns based on HDI. Sandsage, like grass-forb vegetation, responded favorably to increased precipitation in 1987 and 1988. Crown cover sampling on the 1-84 and 3-84 burns did not reveal any remaining evidence of fire impact. The HDI of residual switchgrass (Panicum virgatum) within 1985 revegetation sites remained high (4.-38 dm) in early spring 1~89. Within the t-illage renovation test site, 1989 HDI's and crown cover were similar to.-that in previous' years indicating 1986 renovation was no longer Lmpac t Lng t alL, warm- season grasses. Little change Ln vegetation. conditions dur Lng ·1989 was noted within the 1985 spray and spray-burn sites. Monitoring of greater'prairie~chickens (Tympanuchus cupido) indicated most hens ne st e.d near leks which were on private land. Nesting hens did not appear to be attracted to greater HDI of grass-forb vegetation within the Tamarack Prairie, however, sample size was small. All nests were associated with sandsage. High mortality and nest predation occurred in May and early June and was primarily attributed to coyotes (Canis latrans). ��SANDSAGE-BLUESTEM PRAIRIE RENOVATION Warren D. Snyder P. N. OBJECTIVES Test renovation and revegetation techniques for increasing standing residual height-density of grasses, the proportion of tall warm-season grasses within the composition, and for reducing the quantity of sand sagebrush to <30% canopy cover in an ungrazed sandsage-bluestem prairie on the South Tamarack, South Platte Wildlife Area in northeastern Colorado. SEGMENT OBJECTIVES Monitoring follows: of environmental and vegetation conditions and changes continued as 1. Precipitation was monitored throughout the year supplementing data from 4 electronic rain gauges with information from nearby weather stations through the winter months. 2. Soil moisture accumulations, plant phenology, and weather were monitored primarily in spring and especially at the time of controlled burns. 3. Visual obstruction (height-density) measurements were obtained on treatments and controls where applicable in late winter and/or early spring prior to green-up. 4. Crown cover, species composition, and frequency of occurrence measurements were obtained from mid-summer to early fall. 5. Six male .and 7 female greater prairie-chickens were trapped adjacent to the Tamarack Prairie in April 1989 and equipped mounted solar transmitters. Monitoring of their survival, of habitat types for nesting, feeding and loafing, and use Tamarack Prairie was conducted. 6. Data compilation and writing the annual job progress conducted during fall and winter 1989-90. on leks with ponchomovements, use of the report were METHODS Approaches used were described by Snyder (1986~, 1986Q, 1987, 1988) and are outlined in the Segment Objectives. Segment Objective #4 was conducted within the 1984 burn sites and their controls after not being conducted in 1988 and was deleted within the 1985 and 1986 burns. Walk-in traps were placed on leks located near the Tamarack Prairie in early April 1989 to trap greater prairie-chickens. Trapped birds were marked with a numbered aluminum leg band and color-marked with 2 or 3 plastic bandettes. �176 Poncho-mounted, solar-powered, single-stage transmitters were placed on 6 males and 6 females; a 7th hen was instrumented later in April after recove~y of a transmitter from a predated male. Additional males trapped on the leks were banded and released. RESULTS AND DISCUSSION Environmental Conditions Three of 4 automatic precipitation recorders provided continuous data from early spring through fall 1989 and the 4th provided intermittent data due to electrical failure and plugging by rodents. These data were supplemented by U.S. Weather Bureau records from nearby weather stations for winter months. Only a few light showers were received and conditions remained extremely dry from March through 14 May 1989, with the first significant rainfall coming on 15 May. However, above average precipitation was received in June, August, and September and the June through September combined total was above that of the long-term average (Fig. 1). As a consequence, considerable vegetation recovery occurred. October through December remained extremely dry with about 1 dm of snow occurring in mid December. Annual precipitation was near the long-term average (Fig. 1). Soil probe sampling showed only the upper 4-5 dm of soil contained moisture throughout spring 1989 (Table 1). This was considerably less than in previous years of study. April-May averages were 9.9 dm in 1985, 15.6 dm in 1986, and 11.1 dm in both 1987 and 1988. Table 1. Soil moisture accumulations (dm)& based on soil probe during spring 1989, Tamarack Prairie, Colorado. ~ Location 12 West North-central South-central East 3.51 4.65 4.52 3.89 2 3.87 4.70 4.89 4.44 &Mean depth (dm) of 4 probes/location precipitation gauges. Ma 15 4.25 5.78 4.64 samples 31 Jun 9 3.87 6.67 4.64 5.78 4.19 6.73 7.18 6.48 taken near each of the 4 Average monthly temperatures were above normal during March, near normal in April and May, and below normal in June 1989 (Fig. 2). Phenology of vegetation was similar to that of the 2 previous years except that a hard freeze on 1 May and below average soil moisture conditions retarded development and flowering of herbaceous plants. Progression of development was summarized for a selected list of species (Table 2). More favorable precipitation after mid-May observably stimulated growth and flowering of many species. �22 '"' (/) W J: U ....., Z 20 OCT-DEC 18 SEP 16 12 I- 10 I- 8 U 6 a: a. 4 W --, ~ 14 Z 0 « -a-. :------; . AUG , I I .':. ··'··i i··.·· ••··..··." I JUN [~ff~~ MAY _APR JAN-MAR 2 0 X 84 85 86 87 88 89 YEAR Fig. 1. Monthly and annual precipitation (in.) from 1984 through 1989 in relation to the long-term mean, Tamarack Prairie, Colorado. �1~3 80 : JUN 75 ~ i~ (J) ,------, ~ ' . MAY W 70 a: o 65 APR C 60 MAR w W ~ ., .. ' 1 , W a: ::J l- <t 55 SO a: w a. 45 ~ 40 W I- 35 30 x 84 85 86 87 88 89 YEAR Fig. 2. Monthly average temperature (F) from March through June, 84-89 in relation to the long-term mean, as an index to vegetation phenology, Sterling, Colorado. �Table 2. Phenological conditions of selected vegetation during spring 1989, Tamarack Prairie, A r Species 12 Artemisia filifolia ~. l udov ic iana AstragaLus sp. Cymopteris montanus lathyrus polymorphus leucocrinum montanum Mentzel ia nuda Penstemon ~stifolius PhLox andicola Psoralea lanceolata Sphaeralcea coccinea Tradescantia occidentalis Tragopogan sp. avegetation height (em). E=early,M=mediun,L Height-Density 5-7.6a Budded E. budded Basal leaf Emerg. L. bud E. bloom 5-10 2.5 Basal IOhorl 5 Leafing 5-7.6 Dormant Dormant Dormant Dormant Dormant 5-10 2.5-5 Agropyron smithii Bouteloua gracilis Calamovilfa longifolia Panicum virgatum Paspalum stramineum Sporobolus eryptandrus Stipa f.Q!!!2ll ~sp. 18 late,F Leafed CoLorado. Ma 24·27 10 22 M. budded Basal leaf 5.0 F. budded 2.5 Leaf 7.6 BLoom l. bloom 5-7.6 Leaf 5 -7.6 BLoom E. bloom F. bloom 5- 7.6 10 Leafed 2.5 l. bloom L. bloom 7.6-10 10·15 F. bLoom 2.5 5-7.6 15-20 F. bloom l. bLoom Headed 15 E. bLoom l. bLoom 7.6 F. bLoom E. bloom Seeded 10-12.7 5 10 10 5 5-7.6 15-20 Headed 15 5-7.6 15-20 15-25 5-7.6 10-15 7.6-10 Emerging Dormant Dormant 7.6-10 2.5 < 7.6 5 Dormant 10-12.7 E. bloom 5-7.6 15 L. bloom 15 E. heading full. Sampling Within Burned Sites Height-density indices (HDI) of residual grass-forb and sandsage indicated that precipitation subsequent to prescribed burns was the primary factor determining how rapidly vegetation recovered from fire within sandsage dominated rangelands in eastern Colorado. Grass-forb recovery following the 1984 burns required 3 growing seasons (Fig. 3) when annual precipitation was below the long-term average through the 1st two post-burn (1984-85) growing seasons and exceeded the average during the 3rd year (Fig. 1). Recovery following the 1985 and 1986 burns occurred within 2 years (Figs. 4, 5). First-year impact of fire in 1986 was also markedly reduced in comparison with previous years leading to more rapid recovery because of above average precipitation that year (Fig. 5). Although regression analysis did not indicate a strong relationship CR2 ~ 0.11), markedly higher HDI indices of residual grass-forb vegetation were recorded in 1988-89 after years of above average precipitation. Grass-forb HDI declined from 1988 to 1989 only within burn 1-84. That site received less precipitation in 1988 than other locations within the Tamarack Prairie. Greater precipitation was especially evident at burn 3-84 in 1988. Blizzards and/or compacting snows, which sometimes mat grass-forb vegetation, probably interfered with the precipitation-HDI relationship. Lack of major snows during winters 1987-88 and 1988-89 may be one reason subsequent HDI's were above those of preceding years. �180 0.6 0.5 """ E ..._, > "'C 0.4 - t- CONTROL '" (J) Z W 0.3 ,' 0----··--(1 ,, C ,, ,, I t- J: -J: e" -. 0.2 W BURN 0.1 o 84 85 86 87 88 89 YEAR Fig. 3. Height-density (dm) of residual grass-forb vegetation from 1984 (pretreatment) to 1985-89 (post-treatment) within the 1-34 burn and control sites, Tamarack Prairie, Colorado. �1.2 1.1 , 1 , , , , , """" 0.9 "t'J ......, 0.8 , , E > I- - (/) , r , 0.7 I , Z .~ C \;._) BURN W , 0.6 I l- ::I: "-::I: I I I 0.5 I W I 0.4 I I -. CONTROL 0.3 0.2 0.1 85 86 87 88 89 YEAR Fig. 4. Height-density (dm) of residual grass-forb vegetation from 1985 (pretreatment) to 1986-89 (post-treatment) within the 1985 combined burn and control sites, Tamarack Prairie, Colorado. �182 1 , 0.9 r , , , r 0.8 '"E , , 0.7 "'C , I """ > I- - I ,""'_' 0.6 \p (/) Z , , I 0.5 W I BURN C I I- 0.4 ::I: -::I: e" W CONTROL 0.3 0.2 0.1 o 85 86 PRE-BURN 87 88 89 POST-BURN Fig. 5. Height-density (dm) of residual grass-forb vegetation from 1985-86 (pretreatment) to 1987-89 (post-treatment) within the 1986 combined burn and control sites, Tamarack Prairie, Colorado. �Height-density samples from 1985 and 1986 burns provided evidence that fire stimulated increased grass-forb growth during the 2nd growing season. However, this increased growth, if it occurred, apparently was only temporary and sampling in subsequent years did not show continued stimulation (Figs. 4, 5). Apparently changes in HDI of grass-forb vegetation, whether ?ositive or negative, are not long-term. Sand sagebrush usually loses most of its woody, above-ground structure in fires, and therefore, recovers more slowly than grass-forb vegetation (Tables 3-6). Regrowth was not fully regained for several years. However, sandsage. like grass-forb vegetation, responded favorably to increased precipitation i~ 1987 and 1988 and increased HDI's were recorded on all controls. Sandsage within burned sites recovered to pretreatment levels within 3-4 years. Crown Cover, Composition, and Frequency of Occurrence Sampling Covariance analyses were conducted on crown cover data derived from metric belt transects within the 1-84 and 3-84 burns and controls (Tables 7, 8). These had been sampled prior to prescribed burns in late winter 1984 and again during mid to late summer 1984, 1985, 1987, and 1989. In general, there was little evidence of any remaining effect of fire in 1989, either positive or negative, on most species or species groups (Tables 9, 10). Blue grama (Bouteloua gracilis) showed little response to fire on either site. In contrast needle-and-thread grass (Stipa comata), a cool season species, was suppressed by fire through the first growing season, but recovered quickly. Crown cover in 1989 remained lower (f < 0.05) within burned transects than within controls in burn 1-84 when tested against pretreatment indices, however, the opposite occurred within burn 3-84 (Fig. 6). It is doubtful tha: site differences were important, but rather, that needle-and-thread recovered to pretreatment levels by 1987, and sampling error, primarily during pretreatment, was responsible for the discrepancies. Other major grasses, sand dropseed (Sporobolus cryptandrus), pralrle sandreed (Calamovilfa longifolia), and sand bluestem (Andropogon halli) previously responded positively to fire, but the response was not sustained. Among these grasses, sand bluestem seemed to respond most favorably to above average precipitation in recent years. Total warm-season grasses seemed to be enhanced by fire for 2-3 years, but the response was not evident in 1989 (Fig. 7). Crown cover sampling of sandsage occurred after it had attained new annual growth and indicated almost complete recovery of sands age within 2-3 years. In contrast HOI sampling of early spring residual indicated a slower regrowth of woody parts. General trends in sandsage crown cover have been do~~ward since studies were initiated (Fig. 8), whereas HOI provides evidence of increased growth (Tables 3-6). Reasons for these differences are not clear. Crown cover of perennial forbs was lowest in 1985 on all 1984 burns and controls. Increases occurred on both sets of controls in 1987 and 1989 in contrast to little change within the 2 burn sites. However, significant differences, even when data were combined, were not detected between pretreatment to 1989 and 1985 to 1989 intervals. Combined annual forbs, in contrast to perennials, peaked in 1985 on the 2 burns and the 3-84 control and have declined since. However, data within the 1-84 controls has not conformed to that in other samples. �Table 3. Mean height-density (dm) within burn 1-84 and controls spring 1984-89, Tamarack Prairie, Colorado. Years Grass/fb Burn Sandsage Combined 1984 1985 1986 1987 1988 1989 0.256 0.134 0.232 0.208 0.589 0.527 0.856 0.313 0.358 0.526 0.836 0.753 0.372 0.162 0.256 0.258 0.627 0.577 Cr ass y fb 0.253 0.295 0.301 0.224 0.587 0.571 during Control Sandsage CombineG 0.814 0.687 0.643 0.847 1.045 1.052 0.334 0.355 0.368 0.309 0.622 0.628 I Values 1984-89 1988-89 0.78 1.10 1.18 0.51 0.79 Table 4. Mean height-density (dm) within Burn 3-84 and controls spring 1984-89, Tamarack Prairie, Colorado. during Years Grass/fb Burn Sandsage Combined Grass/fb Control Sandsage Combined 1984a 1985 1986 1987 1988 1989 0.222 0.021 0.106 0.077 0.466 0.647 0.827 0.121 0.356 0.286 1.215 1.435 0.493 0.047 0.201 0.157 0.831 1.159 0.183 0.191 0.200 0.216 0.625 0.934 0.935 0.797 1.044 0.688 1.494 1.888 0.531 0.503 0.629 0.424 1.059 1.548 I Values 1984-89 1988-89 2.13 0.21 apretreatment. 1.44 0.44 1.21 0.27 �Table 5. Colorado. Mean height-density (dm) within 1985 burns and controls Burn Year 3 2 during spring 1985·89, Tomarack Prairie, Control s 2 3 ~ GRASS' FORB 1985a 1986 1987 1988 1989 0.507 0.138 0.419 0.780 1.081 0.180 0.053 0.244 0.667 0.850 0.381 0.119 0.472 1.028 1.297 0.379 0.110 0.408 0.877 1.142 0.374 0.325 0.322 0.511 0.774 0.346 0.280 0.376 0.775 1.220 0.158 0.180 0.156 0.443 0.771 0.323 0.277 0.325 0.6';" 1.025 0.872 0.828 0.925 1.070 1.242 0.615 0.745 0.560 1.732 1.873 0.537 0.767 0.648 1.253 1.844 0.627 0.771 0.670 1.317 , .765 0.444 0.391 0.383 0.581 0.844 0.369 0.335 0.393 0.859 1.332 0.323 0.465 0.411 0.878 1.609 0.380 0.382 0.394 0.782 1.254 SANDSAGE 1985a 1986 1987 1988 1989 1.008 0.417 0.523 0.938 1.313 0.679 0.342 0.591 1.167 1.582 0.642 0.438 0.769 1.144 1.385 0.744 0.388 0.672 1.125 1.446 COMBINED 1985a 1986 1987 1988 1989 0.565 0.142 0.425 0.786 1.109 0.407 0.130 0.476 1.037 1.340 0.302 0.088 0.330 0.766 1.017 VALUE 1985·89 Grass-forb Cootlined apretreatment. 0.28 1.22 1988·89 0.27 3.15 0.429 0.124 0.426 0.899 1.196 �136 Table 6. Colorado. Mean height·density (drn)within 1986 burns and controls during spring 1985'89, Tamarack Prairie, Burn Year 3 2 Control g 2 3 ~ GRASS, FORB 1985a 1986a 1987 1988 1989 0.326 0.304 0.120 0.543 0.723 0.314 0.337 0.437 1.359 1.159 0.259 0.256 0.202 0.824 1.019 0.292 0.290 0.226 0.855 0.963 0.294 0.277 0.234 0.387 0.794 0.316 0.312 0.404 0.858 1.141 0.233 0.213 0.269 0.644 1.074 0.275 0.265 0.289 0.600 0.983 0.829 0.756 0.885 1.265 1.353 0.417 0.300 0.833 1.667 1.250 0.618 0.690 0.671 1.197 1.810 0.682 0.693 0.750 1.237 1.650 0.345 0.342 0.325 0.494 0.886 0.317 0.331 0.413 0.873 1.145 0.298 0.323 0.357 0.743 1.304 0.320 0.332 0.357 0.679 1.116 SA~DSAGE 1985 1986 1987 1988 1989 0.621 0.838 0.169 0.904 1.047 0.375 0.500 0.250 1.000 0.395 0.650 0.250 1.194 0.857 0.566 0.793 0.189 0.945 1.007 C~BI~ED 1985 1986 1987 1988 1989 0.421 0.470 0.127 0.616 0.815 0.315 0.339 0.434 1.355 1. 159 0.272 0.281 0.204 0.834 1.008 F VALUE 1985·89 Grass'forb Conbined 0.51 6.88b 1988·89 2.~ 10. a1985 and 1986 were both pretreatment years. be. < 0.050. 0.335 0.367 0.224 0.863 0.969 �Table 7. treatment (X) , and frequency of occurrence Crown cover (0.01-m2), species c~sition and control transects on Tamarack Prairie burn 1-84, Surmer 1989. Treatment Vegetation Crown cover COOll· of vegetation wi th in Corit r ~ l Freq. / occur. Compo Freq./ occur. 801.0 1112.5 201.0 582.0 141.0 9.0 1.5 23.83 33.10 5.98 17.32 4.20 0.27 0.04 100.00 100.00 88.89 100.00 57.78 4.1.,4 4.1.1. Crown cover Bare ground Dead vegetation 796.5 1535.5 Bouteloua gracilis Stipa ~ Spgrobolus cr~ptandrus Calamovilfa longifolia Andropogon hallii Paspalum stramineum Agropyron smithii Aristida sp. Panicum virgatum ~ & Carex spp. 906.0 1000.0 261.5 517.5 197.0 25.75 28.43 7.43 14.71 5.60 97.78 100.00 93.33 97.78 48.89 2.5 3.0 65.5 0.07 0.09 1.86 4.44 2.22 40.00 35.0 1.04 28.89 372.0 10.63 82.22 125.0 3.72 57.78 Opuntia spp. Ambrosia ~ilostach~a Artemisia ludoviciana Tradescantia sp. Phlox andicola Evolvulus nuttalianus Lath~rus pollmQrpbus Penstomen spp. Psoralea tenuiflora Ph~salis subglabrata Theles~rma sp. HaploP2~s spinulosus Mentzelia nuda Erigeron spp. 19.5 18.0 2.0 82.0 4.0 9.0 5.0 0.55 0.51 0.06 2.33 0.11 0.26 0.14 28.89 20.00 2.22 51. 11 4.44 17.78 6.67 13.0 91.0 15.0 84.5 5.0 10.0 6.0 4.0 2.0 0.39 2.71 0.45 2.51 0.15 0.30 0.18 0.12 0.06 24.44 37.78 4.44 57.78 8.89 11. 11 2.22 8.89 2.22 3.0 2.5 0.09 0.07 2.22 4.44 1.0 2.0 0.01 0.06 2.22 2.22 14.5 0.5 6.0 1.5 0.43 0.01 0.18 0.04 15.56 4.44 4.44 4.44 Croton texensis ChenoPOdium album Plantago ~ Lepidium densiflorum Cryptantha sp. Eupborbia sp_ Amaranthus sp. Lesguerella sp. Eriogonum ~ ~sp. Con~za canadensis He lianthus sp. 6.5 24.5 0.5 6.5 2.0 0.5 2.0 0.5 0.18 0.70 0.01 0.18 0.06 0.01 0.06 0.01 8.89 55.56 2.22 17.78 6.67 2.22 4.44 2.22 9.5 40.5 3.5 17.5 8.0 3.0 1.0 0.28 1.20 0.10 0.52 0.24 0.09 0.03 17.78 55.56 8.89 13.33 2.22 6.67 2.22 2.0 0.5 14.0 0.5 0.06 0.01 0.42 0.01 4.44 2.22 15.56 2.22 Artemisia filifolia 705.0 1784.0 �188 Table 8. treatment Crown cover (0.01-mZ), species CompoSltlon (X) , and frequency of occurrence and control transects on TamaraCK Prairie burn 3-84, SLITIT1er1989. Treatment Vegetation Crown cover COIlll· of vegetation wi th i n Control Freq./ occur. Crown cover COIlll· Freq. / occur. 357.5 5n.0 147.5 401.0 189.0 150.5 79.0 13.5 8.86 24.08 11.35 9.04 4.74 0.81 70.0 100.0 100.0 100.0 40.0 20.0 13.5 8.0 0.81 0.48 15.0 10.0 Bare ground Dead vegetation Bouteloua gracilis ~ comata Sporobolus cryptandrus Calamovilta longitolia Andropogon hallii Paspalum stramineum Agropyron smithii Panicum virgatum Muhlenbergia sp. i(oleria cristata ~ & Carex spp. 486.0 529.0 190.5 426.0 115.5 169.5 69.0 19.5 1.5 15.0 2.0 3.5 35.0 12.02 26.88 7.29 10.69 4.35 1.23 0.09 0.95 0.13 0.22 2.21 95.0 100.0 100.0 100.0 60.0 50.0 5.0 15.0 5.0 15.0 ao.o 40.0 2.40 65.0 Artemisia tilifolia Opuntia spp. Ambrosia 2§ilostach~a Artemisia ludoviciana Tradescantia sp. Phlox andicola Evolvulus nuttalianus Lath~rus pol~rebus Penstomen spp. 359.0 22.5 12.5 9.0 15.5 2.5 16.5 2.0 22.65 1.42 0.79 0.57 0.98 0.16 1.04 0.13 95.0 50.0 10.0 15.0 35.0 15.0 45.0 10.0 457.5 24.0 25.5 18.0 20.0 27.48 1.44 1.53 1.08 1.20 100.0 45.0 20.0 10.0 30.0 6.0 7.0 3.0 0.36 0.42 0.18 15.0 20.0 10.0 Physalis subglabrata Theles~rma sp. Mentzelia nuda Erigeron spp. ~ leptochylla Asclepias s~ciosa 2.0 1.5 4.0 21.0 20.5 2.0 0.13 0.09 0.25 1.32 1.29 0.13 5.0 5.0 5.0 40.0 15.0 5.0 1.5 25.5 7.5 7.0 0.09 1.53 0.45 0.42 10.0 45.0 25.0 5.0 Croton texensis ChenoPOdium album Plantago ~ Lepidium densitlorum Cryptantha sp. Euchorbia sp. Amaranthus sp. Polanisia trachys~rma Lesguerella ludoviciana 3.0 11.5 2.0 1.0 7.5 15.0 6.0 1.5 0.19 0.73 0.13 0.06 0.47 0.95 0.38 0.09 15.0 45.0 15.0 5.0 30.0 55.0 15.0 5.0 8.0 5.0 1.5 0.5 0.48 0.30 0.09 0.03 35.0 40.0 15.0 5.0 5.0 0.5 0.30 0.03 25.0 5.0 0.5 0.03 5.0 �Table 9. Mean crown cover (0.01-m2) for selected species, species groups, and covers within burn and control samples during pretreatment (1984) and post-treatment (1984-89) intervals, burn 1-84, Tamarack Prairie, Colorado. Pre-tr 1984 Aug 1984 Jul 1985 Aug 1987 Jul 1989 362.4 257.7 44.9 185.7 28.7 213.7 102.9 16.3 3.1 117.3 124.8 310.2 86.5 75.4 197.3 47.7 245.8 75.8 40.8 4.1 386.6 7.5 116.5 103.3 58.5 73.3 16.5 89.9 84.0 12.5 15.9 458.5 301. 9 155.2 127.9 53.0 103.7 31. 9 136.3 76.7 15.4 6.0 260.1 400.2 193.6 213.7 55.9 110.6 42.1 153.3 78.6 27.4 9.2 170.2 328.1 510.3 187.0 34.4 194.8 21. 8 216.9 63.5 27.6 1.6 104.4 91. 7 325.1 183.2 36.9 191. 5 29.2 221.7 67.2 28.8 50.6 91. 3 233.9 92.6 125.7 18.4 95.6 9.7 105.8 49.5 15.9 2.4 231.4 601.4 136.3 151. 2 34.9 123.6 25.1 150.0 32.1 33.0 16.2 198.5 477.8 171. 2 237.7 43.0 124.4 30.1 156.4 23.0 51. 3 21.4 150.6 381. 3 Andropogon, Panicum vigatum, BURNED TRANSECTS Boute1oua gracilis Stipa comata Sporobolus crvptandrus Ca1amovilfa longifolia Andropogon hal1ii Warm season grassesa Artemisia filifolia Perennial forbs Annual forbs Bare ground Dead vegetation CONTROL TRANSECTS Bouteloua gracilis Stipa comata Sporobolus cryptandrus Calamovilfa longifolia Andropogon hallii Warm season grassesa Artemisia filifo1ia Perennial forbs Annual forbs Bare ground Dead vegetation alnc1udes Ca1amovi1fa, and Paspalum spp. �190 Table 10. Mean crown cover (0.01-m2) for selected species, species groups, and covers within burn and control samples during pretreatment (1984) and post-treatment (1984-89) intervals, burn 3-84, Tamarack Prairie, Colorado. Pre-tr 1984 Aug 1984 Jul 1985 Aug 1987 Jul 1989 166.4 132.7 110.0 97.6 17.7 128.0 242.4 102.1 2.6 202.9 121. 6 87.5 23.7 69.4 68.2 9.1 89.4 116.6 85.5 14.3 706.2 31.4 35.1 59.4 53.1 86.1 4.6 95.0 196.4 50.2 63.0 250.2 153.1 67.3 84.1 76.9 92.3 15.2 126.4 154.1 52.0 32.3 359.6 275.0 91.6 204.8 55.5 81. 5 33.1 131. 2 171. 1 52.3 22.8 233.7 131.2 169.9 130.4 120.7 15.6 155.7 224.4 61.4 1.6 301. 2 141. 9 93.6 87.1 71.0 72.0 9.4 90.6 293.4 43.4 9.9 189.2 337.0 25.7 51. 7 41. 8 61. 8 5.5 70.0 250.3 27.1 17.8 273.8 460.8 61. 8 137.7 109.2 63.2 17.8 91.1 193.0 37.3 12.7 265.6 320.0 70.9 192.8 90.9 72 .4 38.0 123.4 207.2 58.2 10.1 171.9 277.7 Andropogon, Panicum vigatum, BURNED TRANSECTS Bouteloua gracilis Stipa comata Sporobolus cryptandrus Calamovilfa longifolia Andropogon hal1ii Warm season grassesa Artemisia filifolia Perennial forbs Annual forbs Bare ground Dead vegetation 254.3 CONTROL TRANSECTS Boute1oua gracilis Stipa comata Sporobo1us cryptandrus Ca1amovi1fa longifo1ia Andropogon hal1ii Warm season grassesa Artemisia filifolia Perennial forbs Annual forbs Bare ground Dead vegetation alnc1udes Ca1amovilfa, and Paspalum spp. �300 250 '" EI 1-84 BURN ~ .~ I C\I or- , , I I 1-84 CONTROL 200 0• 0 W > 0 I , , I ,, ,, \ \ ,, \ 150 \ \ , , \ 0' \ U Z , , , 8--- -----8 ~ '-'" a: v- / \ \ / \ 100 / \ / ~ -, 0 a: U / I / <, " 50 I ..... <, ".' ...1' ...:.. , 3-84 BURN / \: 3- 84 CONTROL '. o PRE-84 POST-84 85 87 89 YEAR Fig. 6. Crown cover (O.Ol-m2) of needle-and-thread grass from early spring 1984 (pretreatment) through summer 1984-89 (posttreatment) within the 1-84 and 3-84 burn and control sites, Tamarack Prairie, Colorado. �192 800 c----, : , I '-" , , 700 E ,....I , , .\ /\ _I, 600 o• o ....., a: 500 > o 400 1-84 BURN 1-84 CONT. w (J z 300 ~ o a: (J 200 3-84 BURN 100 PRE-84 POST-84 85 87 89 YEAR Fig. 7. Crown cover (O.Ol-m2) of total warm season grasses from early spring 1984 (pretreatment) through summer 1984-89 (post-treatment) within the 1-84 and 3-84 burn and control sites, Tamarack Prairie, Colorado. �4.5 j /,/ /'/ GRASS /'/ /< 4 0 . ,/ "..... E "'C '-' 3.5 - 3 > I- Ie SAGE ~ /. // /'" 0 ~ 0 '/: 0: TOTAL ~ ~ (/) Z W ~ 2.5 ~ ~ C I I- J: -WJ: ~ ~ 2 0 ~ (!' ~ 1.5 1 0.5 o PRIV BURNS CONT RENOV REVEG VEGET ATION TYPE Fig. 8." Height-density (dm) of residual grass-forb, sandsage, and total vegetation within selected vegetation types, early spring 1989 tamarack Prairie, Colorado. �194 The amount of bare ground increased dramatically following fire and has since declined to near pretreatment levels. In contrast, dead vegetation, initially reduced by fire, had almost fully recovered by 1987. Minor recovery was noted from 1987 to 1989 although the total amount for all samples was lower in 1989. The current growth status of vegetation, determined by precipitation and time of sampling, can influence year to year variation. Revegetation and Renovation Treatments Ti11age-Reseeding.--Sampling of residual HDI within the 1985 revegetation strips yielded nearly the same HDI (4.38 dm) as recorded in 1988 (4.44 dm). This index of residual (primarily switchgrass) quality was far greater than for native range whether or not it was grazed or burned (Fig. 8). Crown cover within the revegetation 3trips was not sampled in 1989 although switchgrass and its standing residual probably continued to increase in dominance. Renovation of Interseeded Tracts.--Evaluation of vegetation changes continued in 1989 within an interseeded site (1981-82) that was partially renovated (disked, harrowed, treated with atrazine herbicide) in spring 1986. The HDI of residual grass-forb vegetation declined slightly from the previous year in both treatment and control transects (Table 11). Comparisons between treated and control transects from 1986 (pretreatment) to 1989 showed that grass-forb vegetation within the renovated site possessed greater HDI (f < 0.01). Sandsage was too sparse to be considered in evaluations. Table 11. Mean height-density (dm) of grass-forb, sandsage, and combined cover from pre- (1986) to post-treatment (1987-89) intervals between a tillage-herbicide and untreated control within a previously interseeded location, Tamarack Prairie, Colorado. Year Tillage-herbicide Control 1986 1987 1988 1989 Grass-forb 0.382 0.314 1.554 1.388 0.393 0.271 0.551 0.449 1986 1987 1988 1989 Sandsage 1.234 0.227 1.136 1.143 0.696 0.608 1.268 0.781 1986 1987 1988 1989 Combined 0.579 0.306 1.525 1.373 0.544 0.375 0.677 0.520 �Crown cover of combined bluestems and switchgrass (interseeded species) increased steadily on both control and tillage-herbicide renovation transects from 1985 (pretreatment) to 1989. However, rate of increase within the treated sites was markedly greater (f < 0.005, Table 12). The rate of crown cover increase from 1988 to 1989 was approximately the same on both (f > 0.25), providing evidence that most effects of renovation had already occurred. Table 12. Average crown cover/transect of selected species and species groups from pre- (1985) to post-treatment (1986-89) intervals within tillageherbicide renovation of an interseeded site and control, Tamarack Prairie, Colorado. Species/group 1985 1986 1987 1988 1989 13.7 11.3 17.6 1.5 24.2 12.4 5.4 0.6 36.3 13.8 6.4 1.2 39.8 11.2 6.8 1.6 47.1 13.8 9.9 2.6 16.4 10.7 17.1 2.2 16.6 10.6 16.8 1.7 22.1 10.2 19.8 3.4 24.6 8.1 20.4 1.7 33.9 Tilla~e-herbicide Andropo~on/Panicum Calamovilfa longifolia Bouteloua, Stipa, etc. Perennial forb Control Andropogon/Panicum Calamovilfa longifolia Bouteloua, Stipa, etc. Perennial forbs E Andropogon/Panicum Calamovilfa longifolia Bouteloua, Stipa, etc. Values 8.5 27.6 1.2 1985-89 28.1P 3.23 30.964 4f < 0.005. Prairie sandreed, a deep-rooted native, continued to show little change from pretreatment status. Renovation had severely reduced other, more shallowrooted, perennial grasses. These continued to increase within control transects at greater rates than where treated (Table 12). Numerous other previously interseeded (1981-82) tracts on the Tamarack Prairie were renovated by management personnel in 1986-87. Pretreatment data were lacking but the HDI of 10 sites were sampled in early spring 1989. Grass-forb vegetation, primarily switchgrass and bluestems, was the primary cover with an average HDI of 2.71 dm (Fig. 8) and a range among transects of 1.66-4.14 dm. �196 These renovation tracts all possessed better cover than the site used in previous evaluations. Their HDI was less than that for the revegetation seeded in 1985 (Fig. 8). Strip Spraying of Sandsage.--Little change from the previous year occurred in HDI of residual vegetation within the 1985 sandsage spray site (Table 13). The HDI of grass-forb vegetation increased slightly, whereas that of dea~ sagebrush declined. Pretreatment data are lacking so changes in grass-forb HDI caused by spraying remain unknown. A 99+% kill of sandsage had been previously recorded. Although residual sandsage has slowly deteriorated, it still provides considerable cover. Sandsage was the visual obstruction in 37.2% of the 1987 samples, contrasted to 24.1% in 1988, and 22.7% in 1989. A portion of the sprayed site had been burned (prescribed) in May 1986 which removed all residual sagebrush. Grass-forb HDI was suppressed during the 1st post-treatment growing season, but recovered dramatically during the 2nd year under favorable precipitation (Table 13). Early spring 1989 HDI of grass-forb was slightly less than during the previous year indicating less favorable growing conditions in 1988 than in 1987. However, the data showed a high probability that fire enhanced HDI when compared with HDI trends of the sprayed site. Table 13. Mean height-density (dm) within the 1985 sandsage spray site during 1986 and 1987 post-treatment intervals and preburn (1986) to post-burn (1987-89) intervals within a portion burned in 1986, Tamarack Prairie, Colorado. Years li Transects Grass-forb 1985 SANDSAGE 1986 1987 1988 1989 SANDSAGE 1986 1987 1988 1989 4 4 4 4 Combined 0.879 0.800 1.127 0.960 0.468 0.469 0.737 0.740 0.713 0.333 1.250 0.404 0.129 1.133 0.911 SPRAY SITE 0.246 0.272 0.613 0.675 2 11 11 11 Sandsage SPRAY - 1986 BURN 0.279 0.122 1.131 0.911 Vegetation crown cover within the 1985 spray site, sampled in late May 1989, showed little change from previous years that could not be attributed to extremely dry early spring weather (Table 14). Annuals and perennials, that make major growth in early spring (e.g., Bromus, Stipa, Lathyrus, Allium, Cymopterus, Phlox), were less abundant or lacking in 1989 samples whereas the �Table 14. Crown random transects intervals within Tamarack Prairie, cover (point frame) of vegetation and ground cover within during pre-(1985) and post-treatment (1986-89) spring the June 1985 herbicide-treated sandsage spray tract, Colorado. Vegetation/cover Bare ground Dead vegetation Perennial grass Bouteloua gracilis Stipa comata Sporobolus cryptandrus Calamovilfa longifolia Other perennial grasses Annual grass Bromus tectorum Festuca sp. Artemisia filifolia (alive) A. filifolia (dead) Opuntia & Echinocereus spp. Perennial forbs Ambrosia & Artemisia spp. Tradescantia occidentalis Lathyrus polymorphus Psoralea tenuiflora Evolvulus nuttalianus Phlox andicola Allium textile Mentzelia nuda Leucocrinum montanum Penstemon angustifolius Thelesperma megapotimicum Cymopteris montanus Abronia fragrans Erigeron sp. Annual forbs Croton texensis Chenopodium ilhYm Pepidium & Lesguerella sp. Cryptanthia sp. Tragopogan sp. Plantago purshii Salsola kali Conyza canadensis Lactuca sp. Unid. forbs Pretreatment 1985 Post-treatment 1987 1988 1986 11 1989 402 1,425 529 1,324 536 1,083 387 990 472 1,207 142 194 34 51 4 148 510 46 65 15 312 425 106 53 19 307 683 161 120 15 311 563 108 124 15 70 117 4 116 7 38 10 436 112 1 259 o o 4 287 210 201 44 50 62 83 75 22 123 7 17 121 12 5 18 114 55 6 4 2 1 9 77 1 6 1 1 1 6 1 1 1 8 1 6 45 2 8 12 1 2 2 1 1 2 1 1 3 24 3 1 4 1 1 1 4 3 2 2 2 1 2 1 3 1 1 1 1 �198 amount of bare ground and dead vegetation increased (Table 14). Dominant grasses continued as in preceding years. Four live sage plants were tallied in 1989 showing recovery had ~egun for that species. Mentzelia and Thelosperma were also recorded as new post-treatment species. Monitoring of Greater-Prairie Chickens All but one of the poncho-mounted, solar-powered transmitters worked moderately well when placed on 12 greater prairie-chickens (6 males, 6 hens) trapped near the Tamarack Prairie in early April 1989. Eight other male prairie-chickens and one male sharp-tailed grouse (X. phasianellus) were trapped, banded, and released during trapping operations. Monitoring revealed the males remained near their respective leks, but their display activity was markedly impacted. Mike Schroeder (pers. commun.) reported this was typical unless males were radiomarked early in the year and had time to adjust. One male was found dead, [evidence indicated predation by a great-horned owl (Bubo virginianus)] in mid-April, and the transmitter was placed on a hen (total = 7). Hen 860 was trapped at Lek 1 (Fig. 9) as an adult on 2 April 1989. Early to mid-April movements were up to 4 mi. (6.4 km) northeast, however, she reduced movements in mid April and was routinely found in section 21, T10 N, R 49 W northeast of Lek 9 on the Tamarack Prairie (Fig. 9). She nested under a dead sagebrush plant on a hill near the center of Section 21 and began incubation in early May. She had nearly completed incubation on 30 or 31 May when killed by a coyote and buried (uneaten) about 100 m to the south. Size of the consumed clutch could not be determined. Hen 890 was trapped at Lek 1 as an adult on 2 April 1989. She usually remained north of Lek 1 and became stable primarily within the south one-half of Section 20, (T 10 N, R 48 W) on the Tamarack Prairie, usually not far from Lek 9 (Fig. 9). She nested under sagebrush 25 m south of the Tamarack Prairie boundary in the NWl/4, NWl/4 of Section 29 (T 10 N, R 48 W) and began incubation about 16-19 May. She was found away from the nest in early June and her coyote-predated nest was found on 12 June. Clutch size was unknown. She remained within Sections 20-21 on the Tamarack Prairie through the remainder of the summer and early fall. She was last found there on 12 October. Later attempts to find her were unsuccessful, possibly because her transmitter signal had become weak. Hen 921 was trapped, instrumented, and a blood sample was taken at Lek 9 on 3 April. She was flushed near Lek 9 on 6 April and flew well. Her signal was lost for a few days before she was found on 18 April about 4 miles (6.4 km) south-southwest in Section 8 (T 9 N, R 48 W). She remained near newly discovered Lek 14 through the remainder of April (Fig. 9). Her transmitter did not operate properly and contact was lost through May and early June except for 2 brief signals. These provided evidence she was still in the area, but her position could not be triangulated. She was flushed (without a brood) from a wheat strip in Section 8 on 19 June. That was the last time her transmitter signal was heard so her status through summer and fall remained unknown. Hen 132 was trapped and released at Lek 9 on 5 April. She moved 2-3 miles (3.2-4.8 km) east in early to mid April and became more sedentary, primarily �'Q"'1 11 t-'. IllOQ ...... t-'. .• n ..•. <;1 •• - n -.u ..-::::;; -z-: •• n "0 YY - n •• I YY 11 \0 (I) ~ • 13 I-' \Or-' 1-.1 00(1) . \0;0;"" TAMJ. RACK PR AIR IE III E. ~ »> '0 11 III •.... 11 •.... 111 890X n ~ •.... 211 n N CJl >c( o n rt t-'. III 18 23 *9 27 25 83~ X x 132 II) (I) I-' -1 ___ 21 1*2 ;0;"" (I) rt 86 0 X ~ (I) ::s ::s V 13 111 ~ 31 35 X 775 *1 C? 0) a 33 c( ~ 35 31 0 a: :f 2 1110 o ::s CJl 5 1 •.... 1 3 ::s 8 11 z c( (I) I-' III rt •.... o ::s 11 7 • *14 rt Q ~ CJ CJ 11 0 ..J a 12 w U) ROAD 46 o >- tt- Z ::J Z 0 ::J 0 v >- __ =c.=·= rt ffi t-'I ~ 11 III n ;0;"" 13 * X LEK· =-_ 0 NEST· L FGEND 1 SCALE t 3.2KM ==:l2ml. J~ 13 .'-D '.D �200 in the north one-half of Section 32 (T 10 N, R 48 W). Incubation had apparently begun on or before 12 May. The signal was lost from 25 May through 4 June. The transmitter and her remains (coyote predated) were found 4 m from the nest on 7 June. Predation bad apparently occurred at the time of signal loss as the 10 eggs, found under sandsage, were still intact, having been incubated 10-12 days. Hen 298 was trapped at Lek 1 on 3 April and remained near there, primarily to the north, through mid April. She was found in a sandsage swale in the southeast corner of Section 29 (T 10 N, R 48 W) in late April. An inconsistent signal obtained on 1 May gave evidence that she might be alive and nesting. The signal was lost in mid May and remains of an apparent coyote predation were recovered in early June. No evidence of a nest (shell fragments, etc.) was found so her nesting status and exact time of death are unknown. A severely chewed transmitter with a missing antenna was recovered 50 m southwest of the site later in summer. Hen 775 was trapped on 24 April at Lek 1 after the transmitter had been recovered from a predator-killed male. She remained near Lek 1 and began incubation about 20-22 May. She escaped when her nest of 10 eggs, located under sandsage 100 m northeast of Lek 1 in a sagebrush swale, was predated by a coyote on 1 June. The hen remained nearby through early summer, primarily east in Section 32. She was found and flushed on 1 August, but could not be located during subsequent searches there or elsewhere. Hen 835 was trapped on Lek 2 on 5 April. She moved about 2 miles (3.2 km) northeast and remained primarily in the NE1/4 of Section 36 (T 10 N, R 48 W). Incubation began about 8 May in the NWl/4 of the SWl/4 of Section 30 (T 10 N, R 47 W) 300 m into Sedgwick County. Her signal was last heard on 12 May, however, the nest site was not searched until early June in case she had covered the transmitter solar panels while incubating. The nest (6 eggs) was found intact under sandsage, where it had been abandoned. Extensive searches on the ground and with an airplane failed to yield a signal. Therefore, her survival status remained unknown. She was assumed lost to a predator. In summary, radio-marked hens gave no evidence of fidelity to the lek where captured, but nearly all nested near a lek. Hens moved extensively from lek to lek, reducing movement in mid- to late April. Nests and leks were widely dispersed and on private land. Two hens, trapped on Lek 1, moved and spent nearly all of their time near Lek 9 primarily on the Tamarack Prairie. One nested there and the other nested on private land within 25 m of the Tamarack Prairie boundary. Two other hens trapped on Lek 9, moved to nest near other leks. Oata (small sample sizes) failed to provide evidence that greater grass-forb residual cover quality on the Tamarack (Fig. 8) attracted nesting hens from adjacent private lands. All nest placements were associated with sandsage. Events deteriorated rapidly after mid-Mayas several hens and all nests (5, probably 6) were lost. By the end of the nesting season (early summer), 3 hens had been killed, 3 were confirmed alive, and the signal of one was lost. The area was flown in mid July, however, receiver problems hampered efforts to find birds. Visual obstruction sampling (HOI) was conducted at 5 nest sites and 10 proximal (~0:8 km) random locations. Sands age HOI - 2.11 dm) obstructed 56.7% of the readings at nest sites. At random sites sage obstructed 40.5% of the readings and averaged 1.75 dm. Grass-forb HOI averaged 0.33 dm at nest sites and 0.45 at random locations. Combined (grass-forb + sandsage) HOI's (& �201 averaged 1.34 dm at nests and 0.98 at random sites. Small sample sizes make these indices suspect (£ > 0.10), but provide evidence of nest site selection in relation to higher densities of better quality sandsage. Observations of nests sites supported this relationship. After one mortality in April, 3 radio-marked males were predator killed in May or early June and the transmitter from another was recovered at a site where no evidence of predation was found. This transmitter may have slipped off. Thus, one male (possibly 2) survived to mid summer. The male was not found during the fall, but weakening transmitter signals made monitoring difficult. Coyotes were suspected in all but the initial predation and in all nest predations. Primary predation coincided with rainfall and increased humidity apparently allowing them to stalk chickens by smell. During monitoring, males were not approached closely during the interval when they were killed and every effort was made to stay 50-100 m or more away from nesting hens. It is possible the radio-marked population received markedly greater direct and nesting predation than unmarked birds, but this remains unknown. Monitoring of leks in early spring 1990 will provide additional evidence, and assuming major popUlation declines have not occurred, radio-marking may be repeated. Sample sizes were inadequate to confirm a nest site-lek relationship if it occurred, however, most nesting and hen activities were near leks. Most radio-marked birds near Lek 9 resided on the Tamarack Prairie. It is possible that more use of the property would be made if more leks were present, however, few suitable lek sites exist there. Based on the possibility that use is related to presence of leks, several potential leks sites were mowed within the Tamarack Prairie in September 1989. These mowed sites should be maintained in future years to determine if more birds will occupy the Tamarack Prairie. Two additional small leks were found on private lands south of the Tamarack Prairie during monitoring efforts in late May and June. Both were found too late to obtain counts. One lek, contained 2 males based on track evidence after a rain. LITERATURE CITED Snyder, W. D. 1986~. Sandsage-bluestem prairie renovation. Job Progress Rep., Colorado Div. Wild1., Wildl. Res. Rep., Fed. Aid Proj. 01-03-045 (W-37R). Apr:475-498. 19862. Sandsage-bluestem prairie renovation. Job Progress Rep., Colorado Div. Wildl., Wildl. Res. Rep., Fed. Aid Proj. 01-03-045 (W-37R). Apr:499-525. 1987. Sandsage-bluestem prairie renovation. Job Progress Rep., Colorado Div. Wildl., Wi1d1. Res. Rep., Fed. Aid Proj. W-1S2-R. Apr:33l356. _______ . 1988. Sandsage-bluestem prairie renovation. Job Progress Rep., Co1orado'Div. Wildl., Wildl. Res. Rep., Fed. Aid Proj. W-152-R. In Press. �202 Prepared by -:1(1_~6} ~ Wuren D. Snyder Wildlife Researcher C �203 INTERIM FINAL REPORT State of: Colorado Project: W-152-R Work Plan: Job Title: 21 : Job Bird Research 4__ Accipiter Nest Site Selection and Foraging Behavior shinned Hawks in Mature Aspen and Conifer Habitats Period Covered: Author: Upland 01 January through 31 December of Sharp- 1989 Suzanne M. Joy Personnel: C. E. Braun, R. W. Hoffman, Colorado Division of Wildlife; R. T. Reynolds, U.S.D.A. Forest Service; R. L. Knight, S. M. Joy, Colorado State University ABSTRACT Feeding ecology of 11 sharp-shinned hawk (Accipiter striatus) pairs nesting in mature aspen, conifer, and mixed aspen-conifer habitats in western Colorado was investigated during 1988 and 1989. Sizes and taxa of prey were compared among nests and breeding stages, and between prey taxa. Relationships between diet composition and resource availability in habitats surrounding the nests were also compared. Small (~ = 2lg) birds comprised 91% of 513 prey items identified. Small mammals (~ = 4lg) made up the remainder of the diet. Unique plucking and feeding habits of individual pairs accounted for differences in total prey and avian prey numbers among nests. Over twice as many prey items were found during the nestling period at the majority of nests (67%) than during incubation or fledgling stages. Although birds were consumed more often than mammals during all breeding stages, mammalian prey frequency progressive~y increased from incubation through the fledgling stage. Numbers avian prey at the nest sites increased 164% from incubation to the nestling stage and declined 274% from nestling to the fledgling stage. Nearly 60% of the birds eaten during nestling and fledgling stages were young of t~e year. A shift in median bird prey weight from 17.4 g during incubation to . 12.1 g during the nestling period reflected the tendency of hawks to consume more young birds as the breeding season progressed. Hawks at nests dominated by aspen habitat (n - 5) consumed fewer mammals (6.8% of diet) than hawks at 1 nest surrounded by coniferous habitat (28.3% of diet). Conifer habitats appeared to have less foraging value compared to mature aspen types. Avian prey were consumed in proportion to their size availability and species composition in habitats surrounding nests, whereas certain taxa and size classes of mammals were preferentially preyed upon. A draft thesis incorporating the feeding ecology, nest site characteristics, and prey delivery rates of sharp-shinned hawks will be completed by 31 July 1990 in partial fulfillment of the requirements for a Master of Science. of Prepared by: Q;t ~'-' i1A '!tr S~z~ nne M. Joy Graduate Research Assistant Approved by: Wildlife Researcher ��205 JOB PROGRESS REPORT State of: Colorado Project: 'W-lS2-R Work Plan: 21 Job Title: Evaluation of Habitat Quality on Conservation Eastern Colorado Period Covered: Author: : Job Upland Bird Research 01 January S__ through 31 December Reserve Lands in 1989 Warren D. Snyder Personnel: C. E. Braun, J. J. Heinemann, D. D. Johnson, W. D. Snyder, Colorado Division of Wildlife T. E. Remington, and ABSTRACT Visual obstruction readings (VOR) , the primary index to grass-forb quality for wildlife, remained low (0.36 drn) in early spring 1989, but increased to 0.77 drn by mid summer within 140 Conservation Reserve Program (CRP) fields sampled in eastern Colorado. This overall sample included lOS fields randomly selected for sampling in 1988. Early spring VOR indices averaged 0.34 drn contrasted to 0.S6 drn in mid summer for this group (the group included 7 fields cost shared by the Division of Wildlife). The 7 fields were combined with 3S additional fields cost shared by the Division of Wildlife for evaluations initiated in 1989. 'JaR indices averaged O. Sl d.rnin early spring increasing to 1.13 drn in swnrner for this cost-shared sample. Fields containing switchgrass (Panicum virgatum), grass-alfalfa, and grass-sweet clover mixtures had the highest VOR indices in summer. Lack of precipitation hampered grass establishment in spring 1989, but above-average moisture in summer stimulated growth of previously established stands. Treatment of numerous fields by mowing or with herbicides for weed suppression continued. Canopy cover averaged 43% in summer 1989 of which seeded perennials comprised 33%; annual forbs and grasses comprised 67% of che total canopy cover. A relationship between percent land in the CRP and ring-necked pheasant (Phasianus colchicus) crowing count indices could not be detected. Use of a PATREC model to evaluate eastern Colorado pheasant range prior to, and during CRP, predicted increased pheasant density should occur. Lack of winter cover was a primary constraint. Similar models for scaled quail (Callipepla squamata) and northern bobwhite (Colinus virginianus) could not discern benefits of CRP to these species. ��207 EVALUATION OF HABITAT QUALITY ON CONSERVATION RESERVE LANDS IN EASTERN COLORADO Warren D. Snyder INTRODUCTION Personnel of the U.S. Fish and Wildlife Service, National Ecology Center (NEC) in Fort Collins, coordinating the national evaluation of the Conservation Reserve Program (CRP) for wildlife, did not request data for 1989. However, data were collected for evaluation of vegetation changes within CRP lands in Colorado as part of an ongoing study. The Division of Wildlife cost shared with many farmers in eastern Colorado attempting to promote establishment of vegetation that would be of greater value to wildlife within CRP fields. A random selection of these fields was included in the evaluation in 1989 to assess possible benefits. Limited PATREC (pattern recognition) modeling was also conducted to assist in estimating potential benefits of CRP to ringnecked pheasants. P. N. OBJECTIVES Determine distribution and quantity of Conservation Reserve Program land in eastern Colorado in relation to distribution of selected wildlife species, evaluate the quality of vegetation on these lands for selected wildlife species, measure response of selected wildlife species to the Conservation Reserve Program using existing annual surveys, and evaluate the impact of the Colorado Division of Wildlife's cost-share program on cover quality. SEGMENT OBJECTIVES 1. Conduct evaluations of randomly selected CRP fields within eastern Colorado as part of a regional and national assessment of the Conservation Reserve Program coordinated by the National Ecology Center (NEC) of the U.S. Fish and Wildlife Service. 2. Conduct intensive visual obstruction readings (VOR) in a stratified random sample of CRP fields and proximal controls. 3. Conduct intensive visual obstruction readings (VOR) in a sample of fields cost shared by the CDOW (for enhancement of cover quality) for comparison with CRP fields not cost-shared. 4. Conduct pre- and post-treatment assessments of a randomly selected sample of CR fields based on PATREC and HSI models for a selected sample of wildlife species. 5. Estimate response of selected wildlife population surveys. species to the CRP based on annual �208 6. Compare the distribution of CRP lands in eastern Colorado selected wildlife species. 7. Compile and analyze data and prepare annual progress to that of report. METHODS Methods used were described by Snyder (1989). Visual obstruct: ~n readings were based on the Kirsch method of sampling rather than the pr __ddure required by NEC personnel. The Colorado Division of Wildlife (CDOW) cost shared directly with farmers on numerous CRP fields in eastern Colorado to increase use of grass species possessing greater herbaceous cover quality for ring-necked pheasants and other ground-nesting wildlife. Seven of these fields were within the initial NEC sample of 104 fields monitored beginning in 1988. Thirty-six additional fields were randomly selected in February 1989 and landowners were contacted for access permission to monitor vegetation conditions. It was learned later that one of these fields had not been cost shared by the CDOW. Therefore, that field was included within the NEC sample bringing total fields not cost shared to 98 among 105 total NEC fields. Forty-two fields, cost shared by the CDOW, were monitored bringing the overall sample to 140. Efforts were coordinated with T. E. Remington, who conducted a companion study assessing use of CRP fields by avian wildlife. RESULTS AND DISCUSSION Environmental Conditions and Stand Establishment Early 1989 remained extremely dry throughout eastern Colorado and most of the winter wheat crop was either severely stunted or lost. However, considerable rainfall was received, starting in May, in southeastern Colorado and continued through most summer months (Table 1). Limited precipitation was received during mid-May in northeastern Colorado, but it remained dryer than southeastern Colorado until June. July was moderately dry at most locations (Table 1). The last significant rainfall in eastern Colorado :ccurred on 20 September and conditions remained extremely dry through October, November, and December at nearly all locations. A few inches of snow were received in midDecember but added little soil moisture. The distribution pattern of moisture in 1989 was similar to that in 1988 (Table 1), with Springfield in extreme southeastern Colorado receiving the greatest amount both years. Springfield and Greeley were the only stations exceeding long-term averages. Except in the eastern tier of counties, most sites received below average annual precipitation in 1989 and less than in 1988 (Table 1). Conditions for grass establishment on newly seeded CRP tracts were especially poor in early spring, however, some establishment of cool-season species was noted during the early summer. Weather conditions were more conducive to growth of previously established warm-season stands and excellent stands were noted by late summer in Baca County and other eastern tier locations. �Monthly Table 1. and annual precipitation Month Springfield lamar Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec 1989 1988 0.22 0.20 0.20 0.72 3.35 6.01 1.54 2.25 2.63 0.74 0.02 0.24 18.12 17.43 0.27 0.23 0.47 0.53 2.72 3.17 0.56 2.69 2.36 0.12 T 0.33 13.45 11.58 16.65 15.40 longterm R (in.) at 9 u.s. Weather Bureau locations in Eastern Colorado, 1989. limon Burlingtona Akron Greeley Holyoke 0.13 0.26 0.33 0.57 1.56 1.89 0.91 1.89 1.64 0.18 0.05 0.19 9.60 9.42 0.43 0.45 0.39 0.34 2.20 2.42 0.88 3.00 1.13 0.07 0.01 0.39 11.71 14.53 0.17 0.19 0.19 0.44 3.30 6.00 1.20 2.72 1.80 0.07 0.00 0.15 16.23 13.10 0.54 0.47 0.20 0.50 0.95 4.14 2.01 3.30 1.00 0.20 0.04 0.20 13.55 16.24 0.84 0.71 0.53 1.09 1.27 3.26 3.41 1.81 1.85 0.58 0.02 0.34 15.71 12.34 0.92 0.50 0.65 0.56 1.75 6.35 0.99 3.19 1.38 0.09 0.06 0.32 16.76 16.45 0.23 13.57 14.69 12.50 15.09 17.41 17.34 12.45 17.61 14.96 Rocky Ford Sterling 0.46 0.31 0.38 0.69 1.05 3.51 2.51 2.94 1.38 0.11 T -aHay through September data were missing and a weighted average between nearby stations was used. I'J o <o �210 An arbitrary rating of grass stand establishment within the 140 sampled CRP fields during summer 1989 revealed that perennial grass was rapidly gaining dominance (Table 2). The highest proportion of fields classed as none to poor occurred within strata 3 (57%) and 5 (41%). However, 61% of the tot~l fields contained grass stands rated fair to good in mid summer 1989. Table 2. Grass stand establishment summer 1989. Stratuma Fields 1 2 3 4 5 6 Percent 26 21 28 18 32 15 within 140 CRP fields, eastern Colorado, None Poor Fair Good 5 1 5 10 8 11 3 8 3 30.7 9 5 5 5 12 8 31.4 7 8 7 7.9 aSee Fig. 1 for locations 9 4 4 30.0 of strata. Five fields were not seeded to grass in 1988 due to failure of prerequisite sorghum cover crops. Among these, a cover crop and grass stand was attained in 1 field, cover crops were obtained on 2 others, and efforts on the other 2 remained unsuccessful. Extensive stands of field bindweed (Convolvulus arvensis) were a primary hindrance to cover crop establishment. One additional field in Kiowa County was replanted to a cover crop and later to grass apparently because of failure of the initial grass planting. Numerous fields in stratum 1 (primarily Phillips County) were seeded to smooth brome (Bromus inermus) in 1988 and subsequently treated with chlorsulfuron (Glean). However, the herbicide killed bromegrass, so nearly all fields were replanted in spring 1989 when extremely dry weather made grass establishment difficult. Five of these fields, subsequently reseeded in 1989, were included within the random sample under evaluation. Evaluation of Conservation Reserve Vegetation Quality Visual obstruction readings, the primary indices of herbaceous cover quality for nesting within CRP fields, were summarized for pre-greenup (early spring) and nesting (summer) intervals in 1989 (Table 3). During pre-greenup, the VOR index varied widely among strata (Fig. 1, Table 3) with highest indices occurring in northeastern Colorado. One reason was that most fields in strata 1 and 2 were seeded a year later than those in strata 3. Soil quality, grass species selection, and management practices used on the fields were also influential. Average indices were 0.299 dm (N = 98) for fields not cost shared, contrasted to 0.505 dm (N - 42) for CDOW cost-shared fields yielding an overall mean of 0.36 dm (N - 140) during the pre-greenup 1989 interval (Table 3). Most (29 of �211 42) CDOW cost-shared fields occurred within the eastern 1) which usually possessed high VOR indices. Therefore, conducted comparing VOR indices in 42 cost-shared fields CRP fields (tl - 46). There was no difference (f> 0.40) greenup data sets. tier of strata (Fig. an analysis was with non cost-shared between the 2 pre- Table 3. Visual obstruction readings (dm) within sampled Conservation Reserve fields in eastern Colorado ''.lringearly spring and mid-summer, 1989. tl Stratum fields INITIAL NEC RANDOM 1 SAMPLE 17 2 11 3 4 5 6 Total CDOW COST-SHARED COMBINED 1 2 3 4 5 6 COMPARISON NEG CDOW Spring 1989 Mean Summer Sum 1989 Mean 24 14 27 12 7.717 8.365 4.103 1.478 8.382 2.276 0.454 0.760 0.171 0.106 0.310 0.190 13 .052 12.115 10.617 8.740 12.127 7.141 0.768 1.101 0.442 0.624 0.449 0.595 105 32.321 0.342 63.792 0.564 FIELDS ADDED IN 1989 35 17.708 0.521 43.806 1.252 SAMPLES 13.232 15.205 5.391 3.767 9.995 2.439 0.509 0.724 0.193 0.209 0.312 0.174 26.266 26.892 16.405 14.516 14.603 8.816 1.014 1.281 0.586 0.806 0.456 0.588 50.029 0.360 107.598 0.769 60.229 47.432 0.615 1.129 NEC AND COST-SHARED 26 21 28 18 32 15 Total Sum 140 OF NON-COST SHARED AND COST SHARED FIELDS 98 29.320 0.299 42ab 20. 709 0.505 aInc1udes 7 fields from the 105 NEC random sample. b4l fields were sampled in spring and 42 in summer, 1989. Comparison of VOR indices for the NEC sample between 1988 and 1989 pre-greenup intervals revealed the index was slightly lower in 1989 (Table 4). Many CRP fields in 1988 still contained a standing (sorghum) cover crop into which grasses had not yet been seeded. This was probably the reason for the higher index in spring 1988. �212 Fig. 1. Strata for sampling conservation reserve fields in eastern Colorado, 1989. �213 Table 4. Average visual obstruction readings (dm) within NEC Conservation Reserve fields sampled during pre-greenup and nesting season intervals, eastern Colorado 1988-89. Pre-greenup 1988 1989 0.39 0.34 Nesting 0.20 0.56 Although weather conditions were extremely dry during spring 1989, abundant rains beginning in May (southeastern Colorado) and June (northeastern Colorado) dramatically enhanced growth of herbaceous vegetation within CRP fields (Table 1). Most fields, especially in southeastern Colorado, were treated with a contact herbicide to suppress annual forbs, however, frequent rains reduced the effectiveness of these efforts. As a consequence, many fields were mowed, primarily after nesting season sampling had been completed. Indices of VOR increased dramatically from early spring to nesting season in 1989 (Table 3) and were markedly above those obtained in 1988 (Table 4). VOR indices within 42 cost-shared fields averaged 1.129 dm compared with 0.615 dm for the remaining 98 fields in the NEC survey. Comparison of VOR indices from 42 cost-shared fields with those from the remaining 46 fields in the eastern tier of strata revealed that cost-shared fields possessed greater cover quality (f < 0.05, ~ - 2.19, df = 86). An attempt was made to derive a more accurate assessment of cost sharing by using a paired comparison analysis. One severely hailed CRP field was excluded. However, this analysis failed to show a marked difference (f > 0.05, ~ = 1.75. df - 40). Several factors including year seeded, stand establishment, soil type differences, and mowing and/or herbicide treatments, made it difficult to analyze the data by paired comparison. As grasses develop and dominate in future years, a more realistic evaluation can be obtained. Switchgrass and other tall warm-season grasses were the primary species in cost-shared fields. Comparison of nesting season VOR's by dominant cover showed these possessed higher indices than most other cover types (Table 5). Fields where alfalfa had been added to cool season grass stands [wheatgrasses (Agropyron spp.) and smooth brome] also yielded higher VOR indices than those containing single grass species. The CDOW provided free alfalfa seed to farmers. However, the U.S. Department of Agriculture, Soil Conservation Service was responsible for approving seed mixtures in CRP fields. They were reluctant to approve use of alfalfa because it has a relatively short life span in much of eastern Colorado. Other perennials apparently do not readily replace it when it dies and concern for possible erosion had priority over its wildlife value (H. Sprock, Soil Conserv. Serv., Greeley, pers. commun.). One field of tall wheatgrass (a. elongatum) in northwestern Weld County, attained dramatic growth in 1989, primarily because of abundant winter moisture in that location (Table 5). However, pure stands of other wheatgrass species, primarily within strata 5 and 6, possessed lower VOR indices. Four CRP fields �214 contained abundant sweet clover and had high VOR's during nesting season sampling. Since this was the 2nd year for this biennial, cover conditions will decline markedly in the future. Grama grasses (Bouteloua spp.) were dominant in most southeastern Colorado CRP fields and usually possessed poor to fair VOR indices (Table 5). Table 5. Visual obstruction readings (drn) in relation to dominant grass species within Conservation Reserve fields in summer 1989, eastern Colorado. Dominant species/covera Switchgrass Wheatgrass Tall wheatgrass Wheatgrass-alfalfa Smooth brome Smooth brome-alfalfa Sideoats grama Blue grama Sweet clover Warm season mixes No significant perennial ~ VOR Fields 13 43 1.579 0.425 4.625 1.023 0.872 2.413 0.505 0.044 2.372 0.767 0.581 140 0.769 13 25 1 6 7 1 25 2 4 grass Total/average aMany fields dominated annual forbs. by grasses contained large quantities of Cover quality (VOR) for nesting ring-necked pheasants was ranked based on data collected in past studies (Snyder 1984, 1988, Table 6). During pre-greenup 1989, 89 (64%) of 139 fields were rated extremely poor « 0.26 dm), 23 (16.5%) were poor (0.26-0.50 drn), 13 (9.3%) were fair (0.51-1.0 drn), 11 (7.9%) were moderate (1.1-2.0 drn), and 3 (2.2%) were good (> 2.0 drn). In comparison with data from the 1988, fewer fields occurred in extremely poor and good classes and more fields were rated poor to fair (Table 6). Considerable improvement in cover quality occurred by summer 1989. Fields rated extremely poor for pheasant nesting decreased to 32.9%, 24.3% were poor, 19.3% were fair, 14.3% were moderate, and 8.6% were rated as good. This was dramatically better than during the previous summer. Among fields sampled, 31 (22.1%) were outside pheasant range and another 48 (34.3%) were in range classed as remnant (Table 7). Twenty-nine (20.7%) were within moderate to high pheasant range (Snyder 1985~). �215 Table 6. Ranking of Conservation Reserve fields for nesting pheasants by VOR classification (Kirsch-dm) during pre-greenup and nesting intervals, eastern Colorado, 1988-89. N Sample <0.26 fields 0.26-0.50 0.51-1.0 1.1-2.0 >2.0 Total Pre-greenup NEC-1988 NEC-1989 76 69 10 18 7 10 5 5 6 2 104 104 All 89 fields 89 23 13 11 3 139 Nesting Season NEC-1988 NEC-1989 87 42 8 26 4 20 3 12 2 5 104 105 All 89 fields 46 34 27 20 12 140 Table 7. Ring-necked pheasant range site ranking of sampled Conservation Reserve fields, eastern Colorado 1989. Rank Absent Remnant Low Fair Moderate Totals to high NEC 29 (76 present) 36 22 11 7 105 Fields CDOW cost shared Totals 2 12 10 7 4 31 (109 present) 48 32 18 11 35 140 Canopy cover averaged 37.3% during pre-greenup 1989 contrasted to 27.4% during spring 1988. The highest density occurred in stratum 1 and was dominated by annual vegetation (Table 8). Canopy cover increased to 42.6% during the 1989 nesting season (compared to 24% in 1988) and perennial grasses comprised 34% of the total vegetation. Average vegetation height increased from 9.7 cm (3.8 in.) in summer 1988 to 17.7 cm (7.0 in.) in early spring 1989 and averaged 16.3 cm (6.4 in.) during summer 1989 (Table 8). Thus, growth in 1989 resulted in increased density, but not height; most of the increased density was attributed to perennial grasses. �216 Table 8. Canopy cover of total vegetation, perennial grasses, annual vegetation, and mean vegetation height within Conservation Reserve fields among strata during pre-greenup and nesting season intervals, eastern Colorado 1989. N fields Total vegetation 26 21 28 18 32 14 139 46.9 43.8 29.6 32.3 35.4 35.5 37.3 Nesting 1989 1 2 3 4 5 6 Total/mean 26 21 28 18 32 15 140 47.0 50.2 47.0 46.0 34.9 28.8 42.6 Quantity, Quality, and Security Stratum Pre-greenup 1 2 3 4 5 6 Total/mean Canol2l:cover Perennial grass ~%2 Annual vegetation Height (em) 7.2 7.1 6.1 10.7 11.4 9.9 8.7 39.7 36.7 23.5 21.6 24.0 25.6 28.6 20.8 27.2 14.1 11.7 16.5 15.0 17.7 15.6 15.8 11.4 20.9 13.3 11.5 14.5 31.4 34.4 35.6 25.1 21.6 17.3 28.1 18.9 20.5 13.4 16.4 12.7 18.4 16.3 1989 of Land Use Types Percentages of vegetation types were sampled by strata in early spring 1989 (Table 9). Rangeland was most abundant especially in strata 4 and 5. These data reflect land use in areas proximal to CRP fields, which, in turn, were formerly cropland. Percent rangeland would be much higher, especially in strata 4 and 5, if sampling was completely random. When small grains, standing stubble, and mulched stubble were combined, they totaled about 42% of the available land. Comparisons of standing and mulched stubble clearly demonstrate that most wheat stubble is left standing over winter in northeastern Colorado (strata 1 and 6) whereas little remains in southeastern Colorado. Some land classed as fallow was formerly wheat stubble tilled after harvest to provide volunteer wheat pasture for livestock through fall and winter. Corn was a minor crop in most sampled areas except in strata 1 and 2, whereas sorghums and millets were common in all areas. Millets, especially, were planted on lands set aside from other crops under the Federal Farm Program. Land in CRP ranged from 7.7% (stratum 1) to 23.3% (stratum 4). However, when rangeland was deleted, CRP increased from 9% (stratum 1) to 56.2% (stratum 4) of the cropland. It was evident that CRP, like sampling, was not distributed uniformly among croplands since CRP had not been allowed to exceed 25% of the eligible cropland per county by more than 2-3%. Among vegetation types (Table 9), row crops and millet were not considered nesting cover for wildlife although they undoubtedly received minor use by some species. Rangelands provide relatively secure nesting habitat for several passerines including horned larks (Eremophila alpestris), lark buntings (Calamospiza melanocorys), western meadowlarks (Sturnella neglecta), and grasshopper sparrows (Ammodramus savannarum). Because of low VOR's and location, rangelands seldom contribute significantly to pheasant production in eastern Colorado. Where rangelands contain sand sagebrush (Artemisia �217 filifolia) or other shrubby growth forms, they provide nesting cover for mourning doves (Zenaida macroura). marginal to fair Small grains, primarily winter wheat, were more common than CRP in all strata except 4 and 5 (Table 9, Fig. 1), and few pheasants occur within those strata. VOR indices of wheat growth through spring 1989 surpassed those of CRP fields by late April in most parts of eastern Colorado pheasant range (Fig. 2). However, early spring drought conditions throughout eastern Colorado severely stunted wheat growth. Its pheasant nesting value in 1989 was much lower than in most previous years (Snyder 1984). Table 9. Major vegetation spring 1989. Vegetation types 1 (%) among strata, 2 3 eastern Stratum 4 Colorado, 5 early 6 13.6 26.0 2l.l 58.6 5l. 7 22.7 27.8 3.8 4.8 6.1 0.3 l.6 7.3 4.5 Small grains 32.1 26.9 25.2 8.6 10.4 28.2 24.2 Stubble Standing Mulched 23.1 2.5 13.7 6.1 3.8 11.2 1.5 2.9 5.9 2.0 17.3 4.4 12.4 5.3 Conservation Reserve % of total % of cropland 7.7 9.0 9.1 12.3 19.2 24.4 23.3 56.2 20.4 42.2 13.3 17.3 14.0 19.5 Sorghum/millet 3.7 4.9 6.8 4.7 7.9 3.9 5.3 11.4 7.2 2.5 l.7 4.8 Beans/beets 0.8 0.5 Sunflowers 0.2 0.6 Alfalfa 0.7 0.2 Unfarmed 0.4 Rangeland Fallow Corn N samples 1,390 0.3 0.1 3.8 0.3 0.4 0.3 0.9 l.1 0.4 1,498 1,292 408 1,000 1,012 6,600 Nesting interval 1989 sampling of wheat fields in all strata but #6 (where nearly all fields had already been harvested) was done to obtain VOR, canopy cover, and height data for comparison with CRP fields (Table 10). In general, wheat fields were much more abundant than CRP fields in strata 1, 2, and 6 and wheat VOR indices were higher in all strata. VOR indices of maturing wheat were approximately one-half those obtained during June 1988 sampling. �218 3.5 3 2.5 E "'C I a: GREEN WHEAT 2 0 > 1.5 1 WHEA T STUBBLE 0.5 CRP MULCHED STUBBLE o 13 25 APR 7 19 31 MAY Fig. 2. Visual obstruction reading indices of dominant covers in eastern Colorado, spring 1989. nesting �219 Table 10. Nesting cover abundance-quality indices among strataa between maturing wheat and Conservation Reserve, eastern Colorado, summer 1989. Stratum 1 Variable % occurrence VOR index Canopy coverage ~ Ht.(dm) 2 3-4 Wheat Wheat CR Wheat CR 32.1 3.1 51. 7 5.5 7.7 1.0 47.0 1.9 26.9 3.5 55.8 5.7 9.1 1.3 50.2 2.1 22.4 1.7 43.6 3.9 5 CR Wheat CR 20.2 0.6 47.0 1.5 10.4 2.3 41.1 4.6 20.4 0.5 34.9 1.3 in stratum 6 prior to sampling aNearly all wheat fields were harvested and it was not included. Wheat stubble, left standing over winter, provides night roosting, feeding, loafing, and escape cover for pheasants. Its VOR quality is important in determining overwinter survival of pheasants in relation to avian predation and blizzards (Snyder 1985h). In spring it attracts nesting by pheasants and other wildlife, but it is extremely insecure because nearly all stubble is tilled prior to completion of nesting efforts (Snyder 1984). Standing wheat stubble was more prevalent than CRP within strata 1, 2 and 6 and possessed higher early spring VOR indices than CRP in all but stratum 2 (Table 11). An index of its relative importance among strata, obtained by multiplying the 2 variables, showed that it remained especially important in stratum 1 and was of little significance as pheasant nesting cover in southeastern Colorado (Table 11). A similar abundance versus quality index for CRP fields indicated they exceeded wheat stubble in importance in southeastern Colorado (strata 3, 4, and 5) but: we re less important in strata 1, 2, and 6. Table 11. Comparison of wheat stubble occurrence (% of vegetation VOR (dm) by stratum, eastern Colorado, early spring 1989. types) and Stratum Variable Wheat Stubble Occurrence, % VOR index Import. indexa Conservation Reserve Occurrence, % VOR index Import. index 1 2 3 4 5 6 23.1 1. 06 4.5 13.7 0.64 8.8 3.8 0.34 1.3 1.5 0.39 0.6 5.9 0.36 2.1 17.3 12.4 0.31 0.62 7.7 5.4 7.7 0.51 3.9 9.1 0.72 6.6 19.2 0.19 3.6 23.3 0.21 4.9 20.4 0.31 6.3 13.3 14.0 0.17 0.36 2.3 5.0 "DerLved by mul.tLpLyLng percent occurrence x the VOR index , �220 Timing of stubble tillage was monitored through April and May previously sampled route in stratum 1. Among fields standing had been tilled by 24 April, 75% by 8 May, 83% by 18 May, and Thus, as in previous years of study (Snyder 1984, 1988, 1989), were cultivated during the peak interval for establishment of pheasants. Influence 1989 along a on 12 April, 11% 97% by 1 June. most fields first nests by of CRP on Wildlife Pheasant crowing counts, conducted along established routes in eastern Colorado in May, provide one means of monitoring potential influences of the CRP on pheasant populations in eastern Colorado. Crowing counts are a product of previous year production and summer-to-early spring survival. They provide indices of spring breeding population levels. A preliminary review of available data indicated that most routes were within locations containing relatively low percentages of CRP (Table 12). Pheasant breeding populations in most east-central and northeastern Colorado routes remained relatively stable at suppressed levels (based on historic data). Within CRP, the sorghum cover crop and weedy seral stages of establishment, by providing brood and fall-winter survival combinations, would seemingly have the greatest impacts on pheasant populations and subsequent indices. The cover crop did not provide nesting cover to pheasants during the year it was planted (1986-87 in southeastern Colorado and 1987-88 in northeastern Colorado). Grass seeding operations during the second year of treatment left little nesting cover and annuals dominated the first year after seeding. Among 140 CRP fields, only 1 was seeded to perennial grass in 1986, 51 (36.4%) were seeded to begin growth in 1987, 84 (60%) were seeded to begin growth in 1988, and 4 were seeded in 1989 or later. Thus, most CRP fields provided extremely marginal nesting cover until 1989 when better cover conditions existed (Tables 3 and 4). Table 12. Ring-necked pheasant crowin~ count indices (~ calls/station) and the percentage of Conservation Reserve along established census routes , eastern Colorado 1986-89. Census route Stratl.n Alli1erst -Paol i Julesburg-Alli1erst Julesburg-Crook Holyoke-Fleming lJages-Haxtun Fleming-Leroy Eckley-Yl.IIl8 Lonestar-Akron Sterling-Proctor Ft. Morgan-Narrows Platner-Elba Ida lia- Joes Bonny Reservoir Burlington North Smokey Hill Lamar-Bristol Las Animas Two Buttes Konatz-Stonington 1 1 1 1 1 1 1 1 1 6 2 2 2 2 2 3 4 &5 3 3 County Phi IIips Sedgwick Sedgwick-Logan Phi IIips-Logan Phi IIips-Yl.IIl8 Logan Yl.IIl8 lJashington Logan Morgan lJashington Yl.IIl8 Yl.IIl8 Kit Carson Kit Carson Prowers Bent Baca Baca % CR 2.4 0 <1.0 3.3 5.3 7.7 3.3 16.3 <1.0 2.7 3.3 2.5 0.9 3.2 0.5 5.4 1.7 33.3 12.0 1986 8.8 4.1 6.2 1.3 11.5 16.0 7.6 9.3 Crowins count index 1988 1987 16.8 18.3 14.1 13.4 10.4 8.1 16.4 5.8 12.7 9.8 7.1 10.2 3.0 5.6 5.5 11.9 3.0 16.8 12.6 apercent CRP was determined using a 1 or 2 mi. wide strip bisected by the census route. 23.1 13.4 19.1 7.0 16.7 9.0 23.8 10.4 14.7 8.7 2.5 10.5 6.3 12.1 10.6 4.0 31.3 39.0 1989 12.5 15.4 18.3 9.2 7.8 15.7 14.5 8.7 7.2 6.1 2.7 6.3 2.9 2.9 21.0 19.2 �221 Conservation Reserve comprised approximately 12 and 33%, respectively, of the land bordering the Konantz-Stonington and Two Buttes Routes in Baca County (Table 12), Consistent breeding population increases were noted along these routes from 1986 through 1988. These increases may have been at least partially a product of the cover crops and early seral vegetation provided by CRP. If so, the increases were primarily the result of increased survival rather than of markedly enhanced reproduction. Pheasant crowing indices along the Baca County routes declined in spring 1989 (Table 12) even though good populations were present in fall 1988 and winter survival conditions were not severe. Recent data indicate reproduction there was poor in 1989 even though nesting cover conditions had improved. Pheasant crowing counts along much of the eastern edge of Colorado declined from 1988 to 1989 whether appreciable CRP was present or not (Table 12). Poor reproduction influenced by dry spring weather in 1988 was considered a primary factor causing the decline in the northeastern corner although influences to the south were less certain. Severe winter weather has not been a factor in winter survival in recent years although pheasants in many areas are thought to be stressed by lack of suitable winter cover. It is evident that if CRP is benefitting pheasant populations in eastern Colorado, crowing counts have not provided strong evidence of it. More years of data are needed, and even then, it must be recognized that crowing indices, because of inadequate replications, are not highly accurate. Nor are pheasant population trends necessarily consistent throughout eastern Colorado. Numerous mourning doves and passerines were observed nesting and using CRP fields primarily during the nesting season inventory. However, reference is made to companion studies by T. E. Remington, who is monitoring reproduction and populations of several avifauna using CRP fields in eastern Colorado. Evaluation of CRP Using PATREC Models A PATREC (pat~ern recognition) model (Russell et al. 1980) for evaluating pheasant habitat in the High Plains of eastern Colorado was developed earlier (Snyder 1983). Model testing in dryland wheat farming areas had strengthened confidence in its predictive accuracy. Therefore, it was used to predict general carrying capacity of tracts of land prior to and after implementation of CRP. Based on a random sample (19) of eRP fields within strata 1, 2, and 3, predicted pheasant population increases averaged 7.4 birds/l.6l km2 (1 mi2) or 81.6% (Table 13). Predicted carrying capacity during both pre and post-CRP implementation intervals was higher in northeastern than in southeastern Colorado. In 3 instances no marked increase was predicted because these sites already contained considerable ungrazed perennial grass. In other cases, increases of several hundred percent were predicted. The model also indicated that addition of vast areas of nesting cover would not markedly improve pheasant populations in locations where winter survival habitat was deficient. Weaknesses in the model were identified in the evaluation. However, it was believed to be effective in predicting general population changes affected by CRP. Pheasant densities were higher in eastern Colorado when the model was developed than at present. Therefore, the model may predict higher than actual density increases. �222 Table 13. Predicted pheasant carrying capacity (birds/km2) on random sites prior to and following establishment of CRP, eastern Colorado, 1989. Stratum N 1 2 3 4 Total/~ Pheasantsikm2 Prior to CRP During CRP Percent change 6 5 7 1 12.7 8.8 7.3 2.0 21.8 16.4 13.7 5.0 87.9 150.0 19 9.1 16.5 81.6 71.7 86.3 Similar PATREC models had been under development and testing for evaluating scaled quail and northern bobwhite habitats. However, preliminary testing of the scaled quail model, in sites containing both CRP and scaled quail range, indicated no pronounced change would occur, or if it occurred, the change would be temporary. This was because cropland associated with shrubby rangeland provides feeding sites and seral edges. Conversion to CRP can temporarily enhance the food base and provide excellent feeding cover, but as grasses dominate the site, the food base and seral edge areas, so critical to quail, can actually diminish. Nesting habitat, which CRP might provide, is not considered limiting, and essential shrubby life form habitats are not provided by CRP. The same scenario exists for northern bobwhite except that their habitat requirements are even more restrictive than those of scaled quail in eastern Colorado. LITERATURE CITED Russell, K. R., G. L. Williams, B. A. Hughes, and D. S. Walsworth. 1980. WILDMIS-A wildlife mitigation and management planning system demonstrated on oil shale development. Final Adminis. Rep., Colorado Coop. Wi1d1. Res. Unit., Colorado State Univ., Fort. Collins. 152pp. Snyder, W. D. 1983. Pheasant habitat assessment and the PATREC model. Page 146 in R. T. Dumke, R. B. Stiehl, and R. B. Kah1. editors. Perdix III Gray partridge and ring-necked pheasant workshop. Wisconsin Dep. Nat. Resour. and Univ. Wisconsin, Campbellsport. 1984. Ring-necked pheasant nesting ecology and wheat farming on the High Plains. J. Wild1. Manage. 48:878-888. 1985~. Management procedures for ring-necked Colorado Div. Wi1dl. Spec. Rep 59. 53pp. 1985Q. Colorado. pheasants Survival of radio-marked hen ring-necked Wi1dl. Manage. 49:1044-1050. J. in Colorado. pheasants in �1988. Evaluation of no-till wheat farming. Job Final Rep., Colorado Div. Wildl., Wildl. Res. Rep., Fed. Aid Proj. W-1S2-R. Apr:303-32S. 1989. Evaluation of habitat quality on conservation rese rv e lands in eastern Colorado., Colorado Div. Wildl., Wildl. Res. Rep., Fed. Aid Proj. W-1S2-R. Apr:22l-244. Prepared by Warren D. Snyder Wildlife Researcher ��225 JOB PROGRESS REPORT State of: Colorado Project: W-152-R Upland Bird Research Work Plan: 21 Job Title: Evaluation Pro ram Period Covered: Author: : Job _6_ of Wildlife Responses 01 January through 31 December to the Conservation Reserve 1989 Thomas E. Remin~ton Personnel: Clait E. Braun, Thomas E. Remington, Warren D. Snyder, David A. Wilson, Colorado Division of Wildlife ABSTRACT Breeding birds were counted along line transects within 39 Conservation Reserve Program (CRP) Fields throughout eastern Colorado. Densities of breeding birds were computed from counts using a Fourier series estimator. Twenty-four species of birds were observed using CRP fields, but few were common residents. Horned larks (Eremophila alpestris) occurred in 25 fields sampled at an average density of 11 ± 12 birds/10 hectares. Lark buntings (Calamospiza melanocorys) were present in 26 fields at an average density of 4.0 ± 4.7 birds/10 hectares. Meadowlarks (Sturnella neglecta) occupied 32 fields (density 1.4 ± 1.3 birds/10 hectares). Mourning doves (Zenaida macroura) were common but many probably went undetected because of their inconspicuous breeding behavior. They were found in just under half of the fields sampled at an average density of 1.2 ± 2.3 birds/lO hectares. Grasshopper and Cassin's sparrows (Ammodramus savannarum and Aimophila cassinii) were fairly common in fields with heavier cover. They were found in 18 and 10 fields at densities of 3.7 ± 8.4 and 0.8 ± 1.7 birds/lO hectares, respectively. Several breeding birds were found in only 1 or 2 fields including ring-necked pheasant (Phasianus colchicus), northern bobwhite (Cciliriusvir~iri.ianus), d i ckc Ls se L (Spiza americaria), 'killdeer (Charadrius vociferus), upland 'sandpiper (Bartramia lon~icauda), red-winged blackbird (A~elaius phoeniceus), and song sparrow (Melospiza melodia). Other birds observed were migrating (vesper sparrow (Pooecetes gramineusJ, house wren (Tro~lodytes aedonl) or nested in adjacent rangeland or shelterbelts (lark sparrow (Chondestes ~rammacusJ, northern oriole [Icterus galbulaJ, loggerhead shrike [Lanius ludovicianus], American robin [Turdus migratoriusJ, western kingbird [Tyrannus verticalus]). Seventy-nine bird nests and/or broods were found in 29 CRP fields by dragging a rope between 2 observers. Nesting species, in decreasing rank order, were lark bunting, mourning dove, horned lark, meadowlark, Cassin's sparrow, pheasant, grasshopper sparrow, red-winged blackbird, and greater prairie-chick~n (Tympanuchus cupido). Some species were under-represented because peak ~: nesting occurred before we searched (horned larks and meadowlarks), or because of their inconspicuous response to the rope drags (grasshopper sparrow). ��227 EVALUATION OF WILDLIFE RESPONSES TO THE CONSERVATION RESERVE PROGRAM Thomas E. Remington P. N. OBJECTIVES 1. Measure and compare avian use of CRP to use of alternative types (wheat and summer fallow). 2. Measure and compare nesting wheat and summer fallow. 3. Relate patterns characteristics success habitat of birds on CRP to success of avian use and nest success and surrounding land uses. in CRP fields on to cover SEGMENT OBJECTIVES 1. Measure and compare densities stubble, and CRP fields. 2. Measure and compare nesting success wheat stubble, and CRP fields. 3. Relate patterns characteristics 4. Capture and attach transmitters to 40 pheasant hens; periodically relocate to determine extent, timing, and type of use of CRP fields. s. Compile of birds breeding of birds nesting of avian use and nest success and surrounding land uses. and analyze in green wheat, data, and prepare progress wheat in green wheat, in CRP fields to cover reports. METHODS A subset of CRP fields randomly selected by the U.s. Fish and Wildlife Service (from the Universe of all CRP fields in eastern Colorado) was randomly chosen for study. The random sampling procedure was stratified to obtain equal samples of fields in each of 3 land-use settings which were thought important to bird use, abundance, or diversity. These were: 1) CRP fields within primarily agricultural areas; 2) CRP fields within predominately sand sagebush rangeland; and 3) CRP fields within predominately short grass rangeland. Permission to enter and measure bird abundance within these fields was obtained via letter, and usually by telephone, from landowners. CRP, and some green wheat, and wheat stubble fields were sampled within an hour of sunrise to count breeding birds. A 0.8 km long transect was established near the midpoint of each field with plastic flags. An observer walked along the transect line slowly watching and listening for birds. Birds observed were identified to species and the perpendicular distance from each bird to the line transect was measured by tape, or eventually using a ro1otape, to the nearest meter. Breeding bird density was estimated from perpendicular distance measures using a Fourier series estimator (Burnham et al. 1980, 1981). �228 Eight hectares within CRP fields were searched for nesting birds by dragging 30-meter long rope between 2 observers. The area around the point a bird flushed was searched intensively for a nest. It was not possible to repetitively sample individual :ields to determine nest success because of time constraints. a RESULTS AND DISCUSSION Twenty-four species of birds were observed along line transects within CRP fields, but relatively few of these were resident breeders (Table 1). Horned larks were the most abundant breeding bird while meadowlarks were the most ubiquitous, i.e., they occurred in the most fields (32 of 39). Lark buntings were second in abundance and ubiquity (number of fields they were present in). The relative importance of these species is indicative of the generally poor condition of the grass stand and cover within these fields. Grasshopper sparrows were relatively common, especially in fields with earlier sign-ups and better cover. Cassin's sparrows were common in fields, generally in the southeast, that were planted to native bunchgrasses. Mourning doves were fairly common in fields with low cover and open ground, but were probably under represented because of their tendency to display from telephone wires or from elevated perches near field edges. Dickcissels were observed while singing in only 2 fields where sweet clover was abundant. Vesper sparrows were frequently observed in late April and early May, but were not observed later and were presumed to be using CRP fields while migrating. Occasional use of CRP fields by species that nested elsewhere (in trees or shrubs) was noted. These species included lark sparrow, northern oriole, American robin, loggerhead shrike, and western kingbird. Nests and/or broods of 9 species were found during searches of 29 fields. These were lark bunting (29), mourning dove (12), horned lark (10), meadowlark (9), Cassin's sparrow (7), pheasant (7), grasshopper sparrow (2), red-winged blackbird (2), and prairie-chicken (1). The number of nests found for each species was not necessarily indicative of breeding density or effort. Horned larks and meadowlarks were probably under represented because most nest searching was conducted after the peak of nesting for these species. Nests of grasshopper sparrows were difficult to find because of their habit of either running off the nest prior to flushing or not flushing at all. Three pheasant nests and 4 pheasant broods were found in CRP fields. It seems likely that pheasant broods moved into, rather than were produced in, CRP fields since the quality of cover was more suitable as brood habitat than nesting habitat. Nest densities of mourning doves were more indicative of dove use of CRP fields than breeding density estimates from line transects. The birds breeding in CRP fields, at this point in the program, primarily represent habitat generalists rather than grassland species. Few line transects or nest searches were conducted in green wheat or wheat stubble this year to increase sample sizes in CRP fields. However it was apparent that these habitats are used for breeding by lark buntings, horned larks, meadowlarks, mourning doves, and perhaps grasshopper sparrows. These are also the species commonly using CRP fields. Thus, the benefits of this program to breeding birds is unclear. �229 Table 1. Birds observed, breeding density (birds/10 ha), and percentage of sampled Conservation fields in which birds occurred, eastern Colorado, 1989. Species Number Fieldsa Horned lark Lark bunting Grasshopper sp. Meadowlark Cassin's sp. Vesper sp. Mourning dove Red-winged blackbird Pheasant Western kingbird Dickcissel Song sparrow Mallard Northern bobwhite Lark sparrow Field sparrow House wren 236 l30 86 72 39 30 24 9 7 7 64 67 46 82 26 28 46 10 8 10 3 3 8 3 3 3 3 3 3 3 3 3 Ki.Ll.de e r Northern oriole American robin Loggerhead shrike Upland sandpiper apercentage 5 5 4 3 2 1 1 1 1 1 1 1 of sampled Density ~ SD 11.0 4.0 3.7 1.4 0.8 12.0 4.7 8.4 1.3 1.7 1.2 2.3 Conunents Bunch grass fields Seen during migration Probably breeding Tree grove in field Tree grove in field fields in which species occurred. Attempts to capture and radio-mark pheasant hens to quantify pheasant use of CRP fields, versus alternative cover types, for nesting and brood rearing were not successful. Additional trapping techniques will be attempted this winter. LITERATURE CITED Burnham, K. P., D. R. Anderson, and J. L. Laake. 1980. Estimation density from line transect sampling of biological populations. Wildl. Monogr. 72. 202 pp. _____ , , and . 1981. Line transect estimation of bird population density using a fourier series. Studies Avian Biol. 6:466-482. Prepared by~f Thomas E. Remington Wildlife Researcher of ��231 JOB PROGRESS State of: Colorado Project: W-152 -R Work Plan: 21 Job Title: Avifauna Period Covered: Author: Personnel: REPORT Upland Bird Research : Job _7_ Responses 01 January Cynthia to Grazing through 31 December 1989 P. Melcher C.E. Braun, K.M. Giesen, C.E. Poley, Colorado Division of Wildlife; Cynthia P. Melcher, Colorado State University ABSTRACT Long-term population data indicate that white-tailed ptarmigan (Lagopus leucurus) breeding densities have declined in Rocky Mountain National Park (RMNP). Examination of elk (Cervus elaphus) population estimates for the Park reveals a negative relationship between decreas~ng ptarmigan and increasing elk population trends. Willow (Salix spp.), an essential ptarmigan food, may be declining in areas that are more heavily browsed as the elk population increases. A pilot study was conducted from 12 May through 16 August 1989, to ascertain whether local levels of browsing by wild ungulates in alpine/krummholz areas (ptarmigan breeding habitat) were related to local differences in breeding bird densities. Four 17.44 - 22.24 ha study sites were selec~ed ~ear Trail Ridge Road (TRR) in RMNP on the basis of historically similar ptarmigan densities and similar amounts of krummholz canopy cover. Two 100 x SO m plots were selected at each site in willow dominant or codominant communities. Sixty m2 sample plots were randomly chos~n in each plot. Willow charact~ristics,and herbIvore pellet/chip frequencies were measured in the 480 sample plots, '.and breeding bird densities were measured throughout each st~dy, site. IJngula;e pe ll.et Zch Ip f'requency distributions indicated ,that·ungulate presence-was greater in both plots at the Sundance Basin (RCS) and Tombstone Ridge' (TSR) sites, and in one plot at the Gore Range Overlook (GTO) site. Willow characteristics that implicate'high levels of ungulate browsing were: high basal stem densities (including adventitious shoots), low live/unbrowsed terminal leader abundance, low bud abundance, and extremes (low and high) in height. One RCS plot, 2 TSR plots, and 1 GTO plot were characterized by heavily browsed willows. Ptarmigan counts were lowest at RCS and TSR. Avian species richness was greater at RCS and TSR. ��233 AVIFAUNA RESPONSES TO GRAZING Cynthia P. Melcher INTRODUCTION In Colorado, ptarmigan subsist on willow buds and leader tips from fall through spring (Braun 1969, Schmidt 1969, May and Braun 1972, May 1975, Hoffman and Braun 1975, Braun et a1. 1976, Giesen et al. 1980). Willard and Marr (1970) described how limiting resources and extreme climatic constraints can retard recovery of damaged vegetation in alpine tundra ecosystems. It is highly probable, therefore, that severe damage to willows in a1pine/krummho1z regions would result in a long term decline of ptarmigan densities. Ptarmigan in RMNP have been banded and monitored since 1966 (Braun and Giesen 1988). Census data illustrate the ptarmigan population has exhibited a precipitous and persistent population depression since the mid-1980's. Colorado Division of Wildlife researchers observed an inverse relationship between elk and ptarmigan population trends in the Park (Fig. 1). Historically, ptarmigan habitats (Braun 1969, Schmidt 1969, Braun et al. 1976, Giesen et al. 1980) and elk seasonal ranges (Natl. Park Servo 1975, Green 1982, Bear 1989) in RMNP have not overlapped significantly. In addition, researchers have presented evidence that willows were not preferred elk foods (Kufeld 1973, Hobbs 1979, Baker and Hobbs 1982) in RMNP. However, other researchers have shown that heavy browsing by ungulates has altered habitats in the Park (Ratcliff 1941, Packard 1947, Stevens 1971, Olmsted 1979, Stevens 1980), and at least one study (Olmsted 1979) describes the decline of a Salix species as a result of heavy browsing by elk. Loss of elk habitat and migration routes, predator elimination, and heavy hunting pressure adjacent to RMNP may be causing changes in elk habitat use patterns in tna Park. Researchers and observers have noted that increasing numbers of elk are using alpine/krummholz communities year-round (Bear 1989; C.E. Braun and K.M. Giesen, pers. commun.), and it may be that limiting supplies of preferred foods in those areas are encouraging elk to forage on willows. High elk pellet frequencies and heavily browsed or trampled willows appear to characterize areas in which ptarmigan densities have declined. P. N. OBJECTIVES This study is designed to 1) document avian species richness and breeding bird densities in alpine/krummholz areas, 2) compare ungulate pellet/chip frequency distributions among alpine/krummholz areas, 3) compare willow characteristics among alpine/krummholz areas, 4) evaluate the relationships between ptarmigan breeding densities, ptarmigan habitat/resource quality, and levels of ungulate browsing among alpine/krummholz areas, and 5) initiate preliminary monitoring for a long term study. �(v W .j" 4 ~-------------------------------------'0 '68 '72 '76 '80 '84 '88 Fig. 1. Trends in white-tailed ptarmigan breeding densities, elk population (RMNP total), and number of elk using Trail Ridge Road (TRR) areas in Rock y Mountain National Park during 1966 - 89. �235 SEGMENT OBJECTIVES 1. Review pertinent literature applicable to ptarmigan and elk ecology, avian census techniques, herbivore pellet surveys, vegetation measurement techniques, alpine ecosystems, and National Park Service Policy. 2. Establish four study areas near Trail Ridge Road in Rocky Mountain National Park. Select and permanently mark 2 100 x 50 m plots at each site, and randomly select 60 m2 sample plots within each plot. Permanently mark and identify all sample plots. 3. Conduct at least 8 avian census visits at each site: abundance and species richness for each site. 4. Sample herbivore pellet frequency distributions, willow characteristics, and percent canopy cover for all plant species within each plot. 5. Randomly select one half of each plot for elk exclosure exclosures. 6. Analyze data, prepare season. METHODS reports. Submit proposal AND DESCRIPTION calculate relative treatment, for subsequent build field OF STUDY AREAS Four study sites were selected near TRR in RMNP, one each at Medicine Bow Curve (MBC), GTO, RCS, and TSR (Figs. 2a and 2b). Approximate areal measures are: MBC - 22.24 ha, GTO - 17.44 ha, RCS - 19.86 ha, and TSR - 19.33 ha. Criteria used for study site selection were 1) historical presence of whitetailed ptarmigan, 2) presence of willow-dominated or willow co-dominated plant communities,,3) relatively equal proportions of similar habitat types, and 4) minimal visibility from TRR. Site boundaries were determined according to compass bearings and notable landmarks, and were plotted on topographic maps. Time constraints and snow depth precluded the possibility of grid-marking sites. Eight 100 x 50 m plots (2 / site) were selected for sampling willow characteristics and herbivore pellet/chip frequencies. The criteria used for choosing plots were 1) occurrence of a site representative community of willow/krummholz vegetation and 2) minimal visibility from TRR. Plot boundaries were permanently marked at 50-m intervals with 30 cm metal stakes. Stakes were sprayed with fluorescent orange paint for visibility. Thirty m2 sample plots were randomly selected from 2,209 possible grid-coordinates within each 50 x 50 m half of the plots (assuming that one 50 x SO m half would be assigned to elk exclosure treatment at the end of the first field season). Each of the 480 resulting m2 sample plots were permanently marked at their centers with 30-cm metal stakes. Stakes were sprayed with fluorescent orange paint for visibility and marked with tags bearing alpha-numeric ID codes, which identified sample plots by site, plot, and row/column coordinates; e.g., MBC-A-04-2l. �236 SCALE 1:24000 CONTOUR NATIONAL GEODETIC INTERVAL VERTICAL 40 FEET OATU~ OF 1929 Fig. 2a. Locations of study sites and 100 x 50 m plots (plot 1 = solid fill, plot 2 = hollow fill) at East Sundance Basin (ReS) and Tombstone Ridge (TSR), Rocky Mountain National Park. �237 SCALE 1:24000 o Fig. 2b. Locations of study sites and 100 x 50 m plots (plot 1 = sol id fill, plot 2 = hollow fill) at Gore Range Overlook (GTO) and Medicine Bow Curve (MBC), Rocky Mountain National Park. �238 Eight avian census visits were conducted at each of the 4 sites from 27 May through 10 July 1989. Spot mapping methods and standards (IntI. B~.rd Census Comm. 1970) were used to record all species present. I plotted avian locations on graph paper according to compass bearings, landmarks, and counted paces. Sites were visited in sequential order, with at least 3 days separating visits at anyone site. Censuses began between 0600 and 0700 MDT and lasted approximately 3.5 - 5 hours, weather permitting. Poor visibility (dense fog, snow) and extreme wind yielded unreliable results; subsequently, I did not conduct census visits on such days. Ptarmigan were censused, regardless of weather, by using tape-recorded playbacks of ptarmigan vocalizations (Braun et al. 1973) and by tracking birds in fresh snow. Each avian census visit/site was recorded as one sample estimate relative densities of nesting species, species number/species using alpine/krummholz sites for foraging were marked with unique combinations of plastic colored allowed identification of individual birds and absolute (8 samples/site) to richness, and only. All ptarmigan leg bands, which ptarmigan counts. Vegetation and herbivore pellet sampling began 27 May and was completed on 3 August 1989. A l-m2 quadrat frame was used to define sample plot boundaries. The center of the quadrat frame was placed at the sample plot marker, with quadrat frame edges paralleling the 100 x 50 m plot markers. Vegetation sampling included visual estimates of plant species composition (percent canopy cover for each species/m2) within each sample plot and 13 different willow shrub measurements for those sample plots containing willow (Appendix A). Characteristics of prostrate (mat) willow species were not measured, but their canopy cover was recorded. Pellet sampling included identifying and recording occurrence of all herbivore pellets and chips within each sample plot (Appendix B). All pellets were removed from sample plots. Plant species ~anopy cover, willow characteristics, and pellet data were summarized and analyzed with Statistical Analysis Software (SAS). Chi-square frequency analysis was used to test homogeneity of herbivore pellet distributions among plots and sites. Summary statistics were used to describe willow characteristics and canopy cover for all plant species among plots. Because willow data were not normally distributed and have unequal variances among samples, further analysis to test for statistically significant differences in willow characteristics among sites will require data transformations or non-parametric tests. In September, the National Park Service decided to disallow the proposed elk exclosure experiment. Subsequently, all data for each 50 x 50 m plot (1/2 of 100 x 50 m) were pooled so that plots A and B are plot 1, and plots C and D are plot 2, for all study sites. �239 RESULTS Herbivore Pellet Survey Results Ptarmigan pellet and ungulate pellet/chip frequency distributions (Figs. 3 and 4) were significantly different from predicted frequency distributions (chi-square values were 35.428 and 209.636, respectively: (f < 0.0001 for both values). High ptarmigan pellet frequencies in MBG plot 1 and low ptarmigan pellet frequencies in both TSR plots contributed most to the overall chi-square value for ptarmigan pellet frequencies. Low ungulate pellet frequencies in both MBG plots and GTO plot 1 versus high ungulate pellet frequencies in both TSR and RGS plots contributed heavily to the overall chi-square value for ungulate pellet frequencies. Avian Census Results Avian census results varied by site (Table 1). Absolute counts of ptarmigan breeding pairs and single males (territorial and non-territorial) were higher Table 1. Ptarmigan Bird census results, counts are absolute. Spring 1989, Rocky Mountain Mean breeding Speciesa MBG WTPT WGSP WAPI HOLA LISP WIWA YRWA ROW'R MBBId AMROd ROFId PISId NOFLd +lb, 8.75 8.00 4.00 0 0 0.5c 0 2 2 10+ 5+ 0 3.5 +0.5c (3.99) (1.77) (1.60) GTO 3.66 -o.si18.91 (2.98) 8.08 (1.19) 3.81 (1.46) 0 0 0 0 1 2 0 10+ 0 pairs/site National Park. (SD) , n - 6 RGS TSR 0.55 +0.55c 22.20 (5.15) 7.60 (0.83) 3.19 (0.64) 1. 57 (0.53) 2.54 (1.11) 0 0 2 2 1 0 1 0.56 +0.5.6b 25.84 (7.51) 5.40 (0.99) 3.12 (1. 58) 0 0.85 (0.71) 0.68 (0.55) 0.80b (0.49) 3 4 0 0 0 aWTPT - white-tailed ptarmigan, WGSP - white-crowned sparrow, WAPI water pipit, HOLA - horned lark, LISP - Lincoln's sparrow, WIWA - Wilson's warbler, YRWA - yellow-rumped warbler, ROW'R - rock wren, MBBI - mountain bluebird, AMRO - American robin, ROFI - rosy finch, PISI - pine siskin, NOFL ~ northern flicker. bSingle territorial male. CSingle non-territorial male. dSpecies (no. pairs observed simultaneously) using site for foraging only. �240 60 CJ) W ....J 50 a. ~ « 40 CHI-SQUARE 35.43 0 CD P < 0.0001 (/) <, 30 >- o z 20 w ::J 0 w 10 a: u, ,0 1 2 MBC 1 2 1 GTO 2 1 RCS 2 TSR EXP Fig. 3. White-tailed ptarmigan pel let presence/60 l-m2 sample plots, expected frequency of ptarmigan pellet presence/60 l-m2 sample plots, and chi-square test for homogeneity of ptarmigan pel let distributions among plots. 60 CJ) w ...J 50 a. ~ ~~~-. CHI-SQUARE 209.64 P < 0.0001 :"., -, " "' '", ' ',,~ "' . "J « 40 (/) 0 CD <, ~~ 30 >- o ~~ ::J ~~ ~~ z 20 w 0 W 10 0: u, 0 ~~ 1 2 MBC 2 1 GTO 1 2 RCS 1 2 TSR Fig. 4. Ungulate pellet presence/60 l-m2 sample plots, expected frequency of ungulate pellet presence/60 l-m2 sample plots, and chi-square test for homogeneity of ungulate pellet distributions among plots. EXP �241 at MBC and GTO sites, and white-crowned sparrow (Zonotrichia leucophrys) densities were highest at RCS and TSR sites. Water pipit (Anthus ~libescens) and horned lark (Eremophila alpestris) densities were less variable among sites. Other species suspected of using sites for nesting were Wilson's warbler (Wilsonia pusilla), yellow-rumped warbler (Dendroica coronata), and Lincoln's sparrow (Melospiza lincolnii). These species were present at RCS and TSR but not at MBC or GTO. However, Wilson's warblers and Lincoln's sparrows were frequently heard just west of GTO, in willow stands that had more canopy cover (possibly due to greater moisture availability) than those within the study site. Yellow-rumped warblers were occasionally observed at MEC, but there was no evidence they nested at the site. A territorial male rock wren (Salpinctes obsoletus) was also recorded at TSR; neither female nor nest was located, however. More species were observed using MEC and RCS sites for just foraging than were observed using GTO and TSR for just foraging. Overall nesting species richness and breeding bird abundance was greatest at TSR and RCS. Vegetation Survey Results Percent canopy cover estimates were based on 60 l-m2 samples within each plot (Fig. 5). Overall, there was more willow canopy cover in MEC and GTO plots, and more conifer canopy cover in RCS and TSR plots. Mat willow was found in MBC plots, but not in other plots. Significant stands of mat willow were noted at GTO, but plot boundaries did not capture these stands. Mat willow was not observed at either RCS or TSR sites. Mean heights for all sampled willows were greater in both TSR plots, GTO plot 2, and RCS plot 1 (Fig. 6). However, standard errors for willow height were greater at TSR and RCS. When mean willow heights were calculated without including willows that grew among or adjacent to conifer plants, mean heights were reduced by 30 - 50% at TSR, and increased in plot I at RCS, in both plots at MBC, and in plot 2 at GTO. Willow bud abundances / 0.01 m2 were highest in both MEC plots, GTO plot 1, and TSR plot 2 (Fig. 7). When mean bud abundances were calculated without including willows that grew among or adjacent to conifer plants, mean bud abundances were higher in MEC plot 1, RCS plot 2, and both TSR plots: the same plots had lower mean willow heights when not growing in association with conifer plants. Basal stem abundance/0.25 m2 was greatest in RCS plots, and lowest in MBC plots (Table 2). Overall, live/unbrowsed leader abundance/O.Ol m2 was greater at MEC and TSR sites, although abundance was variable among the 4 plots. DISCUSSION Herbivore Pellet Survey Results Ungulate pellet/chip frequencies were higher (f < 0.0001) than predicted frequencies at RCS and TSR. Although occurrence of pellets/chips does not prove that elk have been foraging at those sites, it is an indication indication that elk are using alpine/krummholz habitats at RCS and TSR significantly more than alpine/krummholz habitats at GTO or MEC. �242 o o ,... o CO o U) o C\J C\I t- a: en ,... C\I en o a: ,... C\J o t- ,...G ,... o ~3J\O~ AdON\f~ .lN3~!::I3d en - -c..o ""a.. Q) a.. o - > o o >c.. o c ""c c o Q) a.. Q) ~ Q. t'J') u, �243 50 I ~ o ALL WILLOWS WILLOWS WITHOUT CONIFER ASSCCIA TION 40 - I ; E 30 o I- a- 20 W ::J: 10 1 TSR 2 Fig. 6. Mean heights and standard error bars for all willows and for those willows not associated with conifer species. 80 ~ N fI21J ALL WILLOWS WILLOWS W /0 CONIFER ASSOC. 60 ,...E o o 40 <, (/) c ::J 20 CO 1 MBC 2 1 GTO 2 1 RCS 2 1. TSR Fig. 7. Mea~ bud densities/O.Ol m2 and standard error bars for all willows and for willows not associated with conifer species. 2 �244 Table 2. leaders. Abundance of willow basal stems and live/unbrowsed Character Plot Measurement basis willow terminal ~ SE 8.33 9.58 18.40 13 .46 23.65 23.22 15.33 12.61 2.09 1. 53 2.52 1. 95 2.38 4.06 4.04 1. 86 10.27 6.00 10.53 6.61 5.32 5.85 5.00 9.93 1.24 0.90 1. 38 0.78 0.82 1. 20 0.90 1.11 ,., MBC MBC GTO GTO RCS RCS TSR TSR MBC MBC GTO GTO RCS RCS TSR TSR 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 Vegetation Basal Stems " " " " " " " LivejUnbrowsed Terminal Leaders Abundance/0.25 m- " Abundance/O.Ol " " " " " " " " " " " " ,., m- Survey Results Mean willow heights were similar among sites when means were calculated for all willows sampled i~ each plot. However, omitting willows that grew in association with conifers reduced willow height means by 30 - 50%. I suggest that willows are generally taller when growing in association with conifer because 1) willows may have to grow taller to compete with conifers for light resources, 2) willows may benefit from the increased moisture that collects in the form of snow around tall conifers, and 3) conifers protect willows from browsing ungulates. Basal stem frequencies are important indicators of browsing levels. Salix species reproduce largely through adventitious shoot production, which is stimulated when apical meristems are damaged or removed by herbivores (Bryant 1981). Any adventitious shoot, whose basal stem origin could not be seen or palpated below the leaf litter, was classified as a basal stem and was included with basal stem abundance data. Higher basal stem numbers at ReS were probably related to heavy browsing by large ungulates. TSR did not have high willow basal stem abundance, possibly because most willows were protected by conifers from browsing. Avian Census Results Overall, avian species richness was lowest at GTO and greatest at RCS. Total breeding bird densities were similar at all sites except MBC where density was low. It is likely that differences in site characteristics and locations �245 account for the observed differences. RCS and TSR sites have southeastern exposures while MBC and GTO sites are exposed to the north and northwest, respectively. RCS and TSR sites include wide topographic depressions (saddles) southeast of ridges. These topographic characteristics, combined with prevailing (NW) winds, cause snow to accumulate in these areas. Consequently, deep snow patches and wet areas (solifluction pools and associated drainages) persist well into the breeding/growing seasons at these sites. In contrast, MBC and GTO sites are slightly concave to convex slopes that are exposed to prevailing winds, which makes them drier and cooler than RCS and TSR sites. White-crowned sparrow densities were much lower at MBG than elsewhere. These sparrows have been shown to prefer nesting sites in open stunted woodland and scrub (Harrison 1984), and sites that are proximal to water (Morton et al. 1972); MBG was the driest site, which could explain the lower sparrow density at MBC. Sparrows used high perches in dwarf conifers, almost exclusively, for territorial and mate attraction vocalizations. Their preference for territories that included clumps of krummholz vegetation that were dominated by conifer species could also explain the low densities of sparrows at MBC. The ratio of dwarf conifer vegetation to willow vegetation is lower at MBC than at other sites (Braun 1969). It may be that grazing enhances white-crowned sparrow habitat. Sparrows were observed foraging in small open short-grass patches, which may be maintained or created by heavy grazing. Furthermore, browsing may discourage willows and reduce competition for conifers, cover which sparrows appear to prefer in RMNP krummholz areas. Sparrows consistently fled from forage patches into dense conifer krummholz when alarmed. Both Lincoln's sparrows and Wilson's warblers prefer dense shrub nest sites, especially in boggy areas or along water courses (Harrison 1984). Both species were found at or immediately adjacent to RGS and TSR sites where preferred nest sites were available. Neither species was recorded at GTO, but both were frequently heard just northwest of the site where krummholz was more prevalent and was proximal to water. Yellow-rumped warblers prefer to nest in open coniferous woodland, often at woodland edges (Harrison 1984). Typically, none of the study sites would be appropriate nesting habitat for this species, but MBC and TSR sites were close enough to treeline to explain the appearance of this species at these sites. It may be these birds used the sites for foraging only, although the record for TSR was based on repeated observations of a singing male whose territory partially overlapped the TSR study site. Single ye1low-rumped warbler males were observed on a number of occasions at MBG, but they were not heard singing. The TSR site included part of a rock wren territory. The apparently single male was observed singing from a number of rocky perches. Although rock wrens are known to nest in rocky alpine habitats (to 3350 m), alpine areas are not preferred (Braun 1980). Therefore, greater rock wren densities would not be expected at any study site. Horned lark and water pipit densities were similar among sites. Both species use open areas for nesting (Verbeek 1970, Harrison 1978), and are not as likely to be effected by browsing ungulates. �246 Counts of white-tailed ptarmigan were absolute. Most birds were responsive to playback tapes which revealed their presence. Low ptarmigan densities, sedentary behavior, and individual color-coded leg bands provided the conditions necessary for accurate counts. Furthermore, ptarmigan leave unique and obvious tracks in fresh snow. Since ptarmigan move primarily by walking, censuses conducted just after snow allowed detection of previously undetected non-territorial males and, in one case, a pair which was not responsive to playback tapes (GTO). All sites were censused at least once after fresh snow. SUMMARY AND CONCLUSIONS Ptarmigan densities at MBC and GTO were greater than at RCS and TSR. Historically, densities at RCS and TSR were similar to MBC and GTO (Braun 1969). The overall ptarmigan population decline (Fig. 1) has been a consistent trend for almost 10 years (Braun and Giesen 1988), and is largely accounted for by the depressed densities at RCS and TSR. Ptarmigan numbers began to decline 10 years after the 1968 decision by National Park Service (NPS) policy makers to discontinue elk management in RMNP. In an elk management statement (Natl. Park Servo 1988), the NPS described elk concentrations in RMNP as possibly being in excess of winter range capacities, largely due to elimination of useable habitats and dispersal areas outside the Park. If elk have exceeded traditional winter range capacities, then it is likely they are seeking non-traditional winter ranges and forage, such as alpine/krummholz areas and willows. The willow characteristics, which I labeled as indicators of heavy browsing (high basal stem abundance, extreme height variance, low bud abundance, low live/unbrowsed terminal leader abundance), tended to occur in those areas where ungulate pellet frequencies were highest. It is reasonable to conclude that elk are foraging on willows when they are using alpine/krummhclz habitats. Because ptarmigan densities are lower at RCS and TSR, it is also reasonable to conclude that willow habitat structure and willow food resource quality and/or availability have been altered by browsing elk, which has in turn caused a depression in ptarmigan densities. It is evident from ungulate pellet, willow, canopy cover, and ptarmigan census data that there is an inverse relationship between ptarmigan densities and browsed willows, minimal willow canopy cover, and ungulate pellet presence. Until willow data can be tested for statistical differences, however, the strength of this relationship cannot be ascertained. Therefore, I am unable to reject the null hypothesis that current patterns of habitat use by elk have no effect on ptarmigan densities. However, the decision by NPS staff to disallow the fencing experiment precludes further speculation regarding these relationships. As a result, I have submitted a proposal for another experiment which would allow investigation of responses of ptarmigan to different habitats, including habitats that seem to be unsuitable as a result of browsing by ungulates. �247 LITERATURE CITED Baker, D. L., and N. T. Hobbs. 1982. Composition and quality diets in Colorado. J. Wildl. Manage. 46:694-703. of elk summer Bear, G. D. 1989. Seasonal distribution and population characteristics of elk in Estes Valley, Colorado. Colorado Div. Wildl. Spec. Rep. 65. 14pp. Braun, C. E. 1969. Population dynamics, habitat, and movements of whitetailed ptarmigan in Colorado. Ph.D. Diss., Colorado State Univ., Fort Collins. l89pp. Braun, C. E. 1980. Alpine bird communities in western North America. Pages 280-291 in R.M. DeGraff and N.G. Tilghman, compilers. Workshop Proc. Manage. West. For. and Grasslands for Nongame Birds. U.S. Dep. Agric., For. Servo Gen. Tech. Rep. INT-86. _____ , R. K. Schmidt, Jr. and G. E. Rogers. 1973. Census of Colorado whitetailed ptarmigan with tape-recorded calls. J. Wildl. Manage. 7:90-93. _____ , R. W. Hoffman, and G. E. Rogers. ecology of white-tailed ptarmigan Spec. Rep. 38. 38pp. 1976. Wintering areas and winter in Colorado. Colorado Div. Wildl. _____ , and K. M. Giesen. 1988. Population dynamics of white-tailed ptarmigan, Job Progress Rep., Colorado Federal Aid Project W-152-R. Apr. 1988. Bryant, J. P. 1981. Phytochemical deterrence of snowshoe hare browsing adventitious shoots of four Alaskan trees. Science 213:889-890. by Giesen, K. M., C. E. Braun, and T. A. May. 1980. Reproduction and nest-site selection by white-tailed ptarmigan in Colorado. Wilson Bull. 92:188-199. Green, R. A. 1982. Elk habitat selection and activity patterns in Rocky Mountain National Park, Colorado. M.S. Thesis, Colorado State Univ. , Fort Collins. 165pp. Harrison, C. 1978. A field guide to nests, eggs, and nestlings of North American birds. The Stephen Greene Press, Brattleboro, Vt. 4l6pp. Hobbs, N. T. 1979. Winter diet quality and nutritional status of elk in the upper montane zone, Colorado. Ph.D. Diss., Colorado State Univ., Fort Collins. 131pp. Hoffman, R. W., and C. E. Braun. 1975. Migration of a wintering population of white-tailed ptarmigan in Colorado. J. Wi1d1. Manage. 39:485-490. International Bird Census Committee. 1970. Recommendations for an international standard for a mapping method in bird census work. Audubon Field Notes 24:722-726. �248 Kufeld, R. C. 1973. 26:106-112. Foods eaten by the Rocky Mountain elk. J. Range Manage. May, T. A. 1975. Physiological ecology of white-tailed ptarmiga~ Colorado. Ph.D. Diss., Univ. Colorado, Boulder. 311pp. in , and C. E. Braun. 1972. Seasonal foods of adult white-tailed in Colorado. J. Wildl. Manage. 36:1180-1186. ptarmigan Morton, M. L., J. L. Horstmann, and J. M. Osborn. 1972. Reproductive cycle and nesting success of the mountain white-crowned sparrow (Zonotrichia leucophrys oriantha) in the central Sierra Nevada. Condor 74:152-163. National Park Service. 1975. Resource management plan: Rocky Mountain National Park and Shadow Mountain National Recreation Area. U.S. Dep. Inter., Natl. Park Serv., Estes Park, CO. 88pp. National Park Service. Natl. Park Servo 1988. Statement for management. U.S. Dep. Inter. NPS D-23d. Washington, D.C. 60pp. Olmsted, C. E. 1979. The ecology of aspen with reference to utilization by large herbivores in Rocky Mountain National Park. Pages 87-89 in M.S. Boyce and L.D. Hayden-Wing, eds. North American elk: ecology, behavior, and management. Univ. Wyoming, Laramie. Packard, F. M. 1947. A study of deer and elk herds of Rocky Mountain National Park. J. Mammal. 28:4-12. Ratcliff, H. M. 1941. Winter range conditions in Rocky Mountain Park. No. Am. Wildl. Conf. Trans. 6:132-139. National Schmidt, R. K., Jr. 1969. Behavior of white-tailed ptarmigan in Colorado. M.S. Th~sis, Colorado State Univ., Fort Collins. 174pp. Stevens, D. R. 1971. Plant succession as related to grazing on the upper montane region in Rocky Mountain National Park, Colorado. Proc. Northwest Sect., The Wildl. Soc., Abstract. 1980. The deer and elk of Rocky Mountain National Park: a ten-year study. U.S. Dep. Inter., Natl. Park Servo Rep. ROMO-N-13, Estes Park, CO. l63pp. Verbeek, N. A. M. 451. 1970. Breeding ecology of the water pipit. Auk 87:425- Willard, B. E., and J. W. Marr. 1970. Effects of human activities tundra ecosystems in Rocky Mountain National Park, Colorado. Conserv. 2:257-265. on alpine Biol. �249 APPENDIX A Definitions and Techniques Used for Willow (Salix) Shrub Measurements in Each m2 Sample Plot Canopy Cover - visual estimate of percent canopy cover for each plar.t species within m2, including willow shrubs and mat willow. Percent Dead - visual is dead. estimate Patch Diameter - approximate measured with meter stick. of percent diameter of Salix spp. shrub within of any Salix spp. clump within Patch Distance - distance from Salix spp. shrub within m2 and nearest spp. shrub not within m2, measured with meter stick. m2 that m2, Salix Height - average height of Salix spp. shrub within m2. Meter stick used to average arbitrarily selected heights (ground to tip at selected points); 2 measures taken for each 0.25 m2 containing Salix spp. shrub. No. Basal Stems - absolute count of basal stems within arbitrarily selected 0.25 section of m2. Determination of basal stem made on basis of whether basal stem could be palpated below leaf litter. No. Live Terminal Leaders with Terminal arbitrarily selected 0.01 m2. No. Dead or Browsed selected 0.01 m2. Terminal Leaders Buds - absolute - absolute count within No. Buds per Live Terminal Leader with Live Terminal arbitrarily selected 0.01 m2. No. Buds per Dead or Browsed Terminal arbitrarily selected 0.01 rnZ. Leader =Le~n~~~~~~~~~~~~~~~~w~i~t~h~T~e~rm~i~n~a~l~Bu~d cm ruler. B owsed Terminal Leaders ruler. count within arbitrarily Bud - average within - average within - average within - average within arbitrarily arbitrarily �250 APPENDIX B Definitions and Techniques Used to Measure Abundance of Herbivore Pellets in m2 Sample Plots Ungulate Pellets or Chips - record occurrence each sample plot. White-tailed Ptarmigan each sample plot. Lagomorph plot. Pellets Pellets Prepared by: Approved by: - record occurrence - record occurrence All pellets not collected ungulate pellets or chips in of ptarmigan or lagomorph pellets were removed from sample plots. \\~c_~ :' ,~~~v. Cjn ia P. Melcher Graduate Research Assistant 1/aJ'Z_~J Clait E. Braun Wildlife Research Leader pellets in in each sample �251 JOB PROGRESS REPORT Colorado State of: Project: W-152-R Upland Bird Research Work Plan: 22 Job Title: Avian Research Period Covered: Author: Job _1_ Publications 01 January through 31 December 1989 C1ait E. Braun Personnel: L. A. Benson, C. E. Braun, K. M. Giesen, R. W. Hoffman, J. W. Hupp, O. B. Myers, T. E. Remington, J. W. Schmutz, W. D. Snyder, M. A. Zablan, Colorado Division of Wildlife ABSTRACT Progress was made on preparing and submitting manuscripts to technical journals. The following technical papers and abstracts were published 1989. Benson, L. A. sagebrush Workshop. in 1989. Response of sage grouse to prescribed burning in the big type. Proc. 1989 Western Sage and Sharp-tailed Grouse Abstract. Braun, C. E. 1989. Impacts of reduced bag limits on sage grouse harvest and lek counts. Proc. 1989 Western Sage and Sharp-tailed Grouse Workshop. Abstract. Hoag, A. W. 1989. Plains sharp-tailed Colo.-Wyo. Acad. Sci. 21(1):11 .. Hupp, J. W.. and C. E. Braun. grouse during courtship. , and winter. grouse research in Colorado. J. 1989. Endogenous reserve s- of adult male sage ~ondor 91:266-271. 1989. Topographic distribution J. Wildl. Manage. 53:823-829. of sage grouse foraging Myers, O. B. 1989. Sage grouse use of fertilized rangeland. Western Sage and Sharp-tailed Grouse Workshop. Abstract. Proc. 1989 Remington, T. E. 1989. Why do grouse have ceca? A test of the fiber digestion theory. J. Exp. Zool. S~ppl. 3:87-94. Schmutz, J. A., and C. E. Braun. 1989. Reproductive Grande wild turkeys. Condor 91:675-680. performance of Rio in �252 turkeys. and W. F. Andelt. 1989. Nest habitat Wilson Bull. 191:591-598. use of Rio Grande wild Schroeder, M. A. 1989. Movement and lek visitation by female greater prairie-chickens: a test of the female preference hypothesis for lek evolution. J. Colo.-Wyo. Acad. Sci. 21(1):23. Zablan, M. A. 1989. Survival estimates of sage grouse under changing harvest regulations. Proc. 1989 Western Sage and Sharp-tailed Grouse workshop. Abstract. Prepared by _J.<::!%~U=-' ~Z=-';'---"~~==";:___ __ Clait E. Braun Wildlife Research Leader �253 JOB PROGRESS State of: Project: REPORT Colorado W-152-R Upland Bird Research Work Plan: 25 Job Title: Evaluation Farmin Period Covered: Author: Thomas Personnel: Job 1__ of Wildlife 01 January through Responses 31 December to Pesticides Used in Wheat 1989 E. Remington Todd Abell, Clait E. Braun, Richard W. Hoffman, Carol Mehaffy, Michael W. Miller, Thomas E. Remington, Joan Ritchie, Lyn Stevens, David A. Wilson, Colorado Division of Wildlife; Wendy Meyer, Frank Peairs, Stan Pilcher, Colorado State University; Claude Ross, Ted Warfield, FMC Corporation. ABSTRACT The impacts of two methods using pesticides to control pests in winter wheat on bird communities were investigated. These were aerial spraying of Lorsban (chlorpyrifos) or Di~Syston (disulfoton) to control Russian wheat aphids (Diuraphis noxia) in Spring and microtubule injection of Furadan (carbofuran) at planting to control grasshoppers. No dead or impaired birds were found during searches of 162 ha within 10 fields sprayed with either Lorsban (97 ha) or Di-Syston (65 ha). Fifteen active nests (or broods) of several bird species were ,found within sprayed fields and were not obviously affected. Searchers located 57% of carcasses placed in fields to evaluate search efficiency. Brain cholinesterase activity of horned larks (Eremophila alpestris) collected from a field 2 days after it was sprayed with Di-Syston appeared depressed relative to activity in birds collected from a field sprayed with Lorsban or unsprayed controls (avg. - 5.0, 8.2, and 6.7 ,umoles/min/g, of brain tissue, respectively). Spraying of Di-Syston, but not Lorsban, decreased (f < 0.0.5) survival of pheasant chicks slightly in a controlled experiment. Weight dynamics of chicks in pens within sprayed strips were similar to those in unsprayed strips. Spraying of Di-Syston or Lorsban to control Russian wheat aphids in Spring does not appear to impact birds significantly. Two horned larks and 1 meadowlark (Sturnella neglecta) were found dead during searches of 38 km around the perimeters of 13 fields 510 days after seed planting and microtube injection of Furadan. The meadowlark appeared to have been shot but no wounds or injuries were observed on the ,horned larks. Searcher efficiency averaged 69%. Furadan residues in grasshoppers averaged 0.22 mg/kg, a level unlikely to cause avian mortality. Residue analysis did not include toxic metabolites of Furadan, however. Brain cholinesterase activity of Canada geese (Branta canadensis) was unaffected by feeding (24 hours) within pens where 50 or 100% of wheat had been treated. ��255 EVALUATION OF WILDLIFE RESPONSES TO PESTICIDES USED IN WHEAT FARMING Thomas E. Remington P. N. OBJECTIVES 1. Determine acute mortality of wildlife 1-3 days post-spray prescribed levels of Di-Syston and chlorpyrifos. resulting 2. Recover carcasses of wildlife suspected of dying from pesticide poisoning for analysis of cause of death and pesticide residue levels. 3. Monitor impacts of spraying prescribed chlorpyrifos on avian nesting success. 4. Determirie level of brain cholinesterase depression in songbirds and pheasants 24-48 hours post-spray as an index to both pesticide exposure and potential effects. 5. Measure toxicity of prescribed 7-10 day-old pheasant chicks. levels of Di-Syston levels of Di-Syston from or and chlorpyrifos to SEGMENT OBJECTIVES 1. Measure acetylcholinesterase activity in brain tissue of carcasses dead in, or collected from, wheat fields during spring 1989. 2. Conduct pesticide residue analysis of tissues from birds found dead in wheat fields and exhibiting ~ 50% brain acetylcholinesterase inhibition. 3. Search wheat fields for dead or moribund of Di-Syston, chlorpyrifos, and possibly 4. Determine rates of carcass by search crews. 5. If warranted by results of 1989 spring field season, pesticide in wheat and/or insects in wheat fields will be measured. 6. Compile wildlife following carbofuran. removal by scavengers, and analyze data, and prepare progress found application and carcass detection residues reports. METHODS Wheat Aphid Study Three research approaches were used. Transects were conducted to search for dead or impaired wildlife in sprayed fields. Exposure of breeding birds to pesticides (potentially causing sublethal effects on behavior) was evaluated by collecting horned larks from sprayed and unsprayed fields. Acute toxicity and indirect effects of these pesticides on survival of pheasant chicks was evaluated by experiment. Specific methodology for each approach follows. �256 Transects were conducted by 4-5 people walking abreast, about 20 rows apart 13 days post-spray until 8.1 ha had been searched per field. Most searches were conducted with the rows, although it was found that as wheat grew taller visibility of carcasses was greatest walking against the rows. Search efficiency was measured as the percent recovery of house sparrows (Passer domesticus) placed in sprayed fields prior to the search. To confirm that carcasses would remain 1-3 days post-spray, the disappearance rate of house sparrow carcasses placed in sprayed fields was measured. Horned larks were collected from sprayed and unsprayed fields with a shotgun and immediately frozen. Brain cholinesterase activity was measured by the Ellman procedure as described by Hill and Fleming (1982). The design of the pheasant chick experiment is illustrated in Fig. 1. Nine strips, 121.9 m wide and 381 m long, were marked with flagging and randomly assigned to 1 of 3 treatments; Di-Syston, Lorsban, or unsprayed control. Within each strip, 3 of 36 possible 0.1 ha (1/4-ac) pen locations were randomly selected. Pens were constructed of 61 by 2.54 cm (24" by 1") mesh poultry netting supported by rebar and aluminum rods. Pens were offset 13 m from either edge of the strip to prevent or minimize drift of pesticide across treatments. Within the center of each pen a 9.75 m2 circular pen was constructed of 1.22 m (48") hardware cloth. Ten, 6-day old pheasant chicks were placed in each pen from 0830 to 0950. Di-Syston and Lorsban strips were sprayed with 1.12 kg active ingredient (a.i.) per ha or 0.84 kg a.i./ha, respectively. Spraying began at 1030 and was completed by 1200 hours. Conditions were ideal for spraying; sunny, warm and little or no breeze. To facilitate recovery and prevent escape, chicks were initially placed in the small pens. Chicks were removed at the end of each day and kept indoors at 32.1 C (90 F) and provided pelleted chick feed and water. Thirty (of the original 90) chicks from each treatment were individually marked by wrapping a numbered strip of scotch tape around 1 leg. These chicks were weighed each morning and returned to the larger pens for 7 days to evaluate the suitability of sprayed areas as brood habitat. Remaining chicks were labelled with nail polish which identified the treatment to which they were exposed and kept indoors to monitor subsequent mortality (2 legs painted - Lorsban; 1 leg painted - Di-Syston; neither painted - control). Carbofuran Study The purpose of this research was to gather preliminary information on mortality or morbidity of wildlife thought to be at risk from microtubule injection of Furadan as an at-planting treatment to control grasshoppers. Effects on passerine birds were evaluated by searching borders of treated fields (Furadan 4F at a rate of 0.5 fluid ounces per 305 linear meters) for dead or impaired birds at about 1 and 2 weeks after planting. Searcher efficiency was estimated as recovery of house sparrow carcasses placed within field margins prior to the search. The effect on Canada geese of exposure to furadan-treated wheat was evaluated by penning birds captured by cannon net into 1 of 3 enclosures consisting of 6 rows of treated wheat, 6 rows of treated and 6 rows untreated wheat, or 6 rows of untreated wheat (check). These enclosures were within 12, 305-m long strips of 6 rows of wheat each planted on 19-20 August in a field near �I I 122-m SPRAY STRIPS .- : .. . CIJ i~[:[:::[:: ...... ::::::: :::: E ex) ,.... C? . .......... .. . -- n LJ ·· .. · .... ..... . . - - . . . . n . - ::: : ••• _0 . ~ - - tc ~ - . :z - o o - - - 13-m BUFFER Fig. 1. Experimental design and layout of treatment strips and pens used to assess toxicity of Lorsban and Di-Syston to pheasant chicks. h) l_" -., �258 Briggsdale. Furadan 4F was applied to alternate strips by microtubule injection at a rate of 0.5 fluid ounces per 305 linear meters. The birds left within the pens for a day and a half to feed. They were then killed CO~ asphyxiation. Heads were removed, placed on dry ice, then tra~sported and stored within a freezer at -60 C. Brain acetylcholinesterase activity determined following Hill and Fleming (1982). RESULTS were by to was and DISCUSSION Russian Wheat Aphid Study Mortality Transects.--Ten fields, 162 ha total, (65 ha treated with Di-Syston, 97 ha treated with Lorsban) were searched and no dead or impaired wildlife were found. Fields were searched 3 hours (1), 1 day (6), 2 days (1), or 3 days post-spray (2). Horned larks, lark buntings, meadowlarks, and mourning doves were common in sprayed fields and appeared to act normally. Nests with eggs or chicks and/or fledged baby birds were found in some sprayed fields, usually when the attending parent(s) flushed. We located 7 horned lark nests (2 with eggs, 2 with chicks) or fledged broods (3 broods of 1-3 chicks), 5 lark bunting nests (5 with eggs), and 3 mourning dove nests (2 with eggs, 1 with young). At least in these instances, spraying did not impair normal reproductive behavior such as incubation and feeding young, nor did it kill dependant young in nests. This was confirmed by follow up visits to most nests located within a day of spraying. We recovered 57% (12 of 21) of carcasses placed in fields prior to searches. Carcasses were removed from wheat fields slowly by scavengers (Fig. 2); 75% were still present and visible 3 days after placement and 50% were present after 6 days. These data suggest that if significant mortality had occurred as a result of spraying we would have detected it. Exposure of Breeding Birds.--A pilot collection effort was made this spring, if results indicate significant exposure then more extensive collections are justified. Horned larks were collected from a field 2 days after Di-Syston spraying (5), from another field 1 day after Lorsban had been sprayed (5), and from a rangeland area about 20 km away (3 to serve as controls). Cholinesterase activity appeared depressed in horned larks collected from the field sprayed with Di-Syston relative to those collected from the field sprayed with Lorsban or the unsprayed controls (~± S.E. - 5.0 ± 0.5, 8.2 ± 0.8, and 6.7 ± 1.0 ~moles/min/g of brain tissue, respectively). Sample sizes are too small to make definitive conclusions. Results do suggest 20-30% depression in acetylcholinesterase activity, indicative of exposure to DiSyston. Further collections are warranted next spring. These enzyme rates are only about 50% of levels found in horned lark brain tissue by McEwen et al. (1986). This may be due to the prolonged (4-5 months) time carcasses were frozen. This will be investigated. McEwen et al. (1986) found that aerial spraying of Lorsban on winter wheat at 0.56 and 1.12 kg a.i./ha to control cutworms resulted in significant depression of horned lark cholinesterase activity at 3 and 9 days post-spray. There was no evidence of suppression 1 day after spraying with Lorsban. Pheasant Chick ToxicitY.--Di-Syston spraying lowered (f < 0.05) 48-hour, but not 24-hour survival of pheasant chicks (Fig. 3). An average of 7.3, 8.2, and �259 18 16 14 (/) W (/) (/) <t 0 12 a: <t 0 ZI 10 8 6 ~~--~--~--~--~--~--~~--~--~--~--~--~--o 1 2 3 4 5 6 7 DAYS Fig. 2. Removal rate of house sparrow carcasses from wheat fields in Spring. �IV Q\ 24-HOUR MORTALITY 30 10 SUBSEQUENT SURVIVAL 28 r:«. .. 8 Fig. 3. Survival of pheasant (1.12 kg a.i./ha). 26 chicks 1-5 days following spraying with Lorsban (0.84 kg a.i./ha) or Di-Syston o �261 8.2 (of 10) chicks survived 48 hours per pen sprayed with Di-Syston, Lorsban or unsprayed, respectively. Subsequent survival was similar among treatment groups. A storm cell passing to the south late the afternoon of the day of spraying lowered temperatures dramatically and may have exasperated toxicity of Di-Syston (Rattner et al. 1982). Given that Di-Syston and Lorsban have similar avian toxicities (Smith 1987), it is puzzling that chick mortality differed. This may be due to lower application rates of Lorsban (0.84 kg a.i./ha versus 1.12 kg a.i./ha) or to higher rates of Di-Syston reaching ground (chick) level (F. Peairs, pers. commun.). Chick weight dynamics were similar across all treatments (Fig. 4). Birds generally lost weight when kept outside. This suggests that wheat fields are poor brood habitat for chicks of this age, at least in the absence of a hen. Carbofuran Study Mortality Transects.--Thirty-eight km around the perimeters of 13 fields were searched for dead or moribund birds within 5-10 days of planting. Four of these fields (12 km) were searched again 2-3 weeks after planting. Two horned larks and 1 meadowlark were found dead. None of these carcasses was fresh enough for measurement of acetylcholinesterase activity. The meadowlark had injuries consistent with death resulting from shotgun pellets (radius and ulna of right wing broken, hole in right thigh and right side of pelvis broken). No wounds or injuries were detected in the horned larks. One carcass was little more than a feather pile when found and may have died as a result of predation or it may have been scavenged. The other horned lark had 3 wheat seeds within it's stomach. Consumption of treated seeds is a possible route of exposure. A horned lark was found dead within a treated strip at the Severance study site and also contained 7 wheat seeds in it's stomach. Acetylcholinesterase activity of this bird was 62% depressed relative to 2 control larks shot on the same day. There may have been additional mortality as well, since several feather piles were observed over the course of the pronghorn stl~dy. The experimental design used may have increased potential exposure of birds to treated seeds. The seed drill was in and out of the ground 6 times over 305 m to create 6 treated strips. If the microtubule injector and drill are left on while the drill is pulled out of the ground at the end of each strip, then treated wheat seeds will be exposed on the soil surface. Data should be obtained on furadan and total carbamate residues of treated wheat seeds so risks from this exposure can be evaluated. Searcher efficiency was evaluated in 5 fields and averaged 69% (25 of 36 carcasses recovered). Efficiency within each field was 67, 75, 63, 87, and 60%. These results indicate that had significant mortality occurred we would have detected it, assuming carcasses remained until our searches. This assumption was not tested directly during this study. It appears that mortality of passerines is limited under this application method of furadan to the borders of fields. These results should be interpreted cautiously, however. An inherent weakness in this approach is that birds could feed within treated field margins, flyaway to die elsewhere, and go undetected during our searches. This could be a problem with horned larks who have very large home ranges in fall and winter. Exposure of insectivorous birds to furadan-contaminated grasshoppers was presumably �N or", 40 ,..... -- CONTROL ------ DI-SYSTON 35 ::I: /:~j------I'" <, <, <, ..••. .// .-.._ .._ .._ .._ .._ .._ .._ .._ .._ .._ .._ .. w //------.'/ g 30 UJ -W // ::> - -- ~/ 0 /.7 . ...•......... _1 .. /·· -, -- / // ••••• •..., ..••. /f CJ ~ <, LORSBAN m ~ I- <, // / w g en w g en ••••• ••••• o 0 ::> 25 ::> 20 1 2 3 4 5 6 7 DAYS SINCE SPRAYING Fig. 4. Weight dynamics (1.12 kg a.i./ha). of pheasant chicks sprayed with Lorsban (0.84 kg a.i./ha) or Di-Syston 8 �263 minimal since few insectivorous birds were observed during our searches. Fall migrations may mitigate potential exposure of these birds to furadan, although the extent to which this occurs may vary annually. There may he potential for mortality from treated grasshoppers. A starling (Sturnus vulgaris) was found dead at the Severance study site with no apparent injuries and a partially digested grasshopper in it's stomach (along with several Russian olive [Eleagnus angustifolia] drupes. The premise of death from ingestion of a single grasshopper is not supported by toxicity assessments from furadan residues measured in grasshoppers collected from treated fields. Maximum furadan residue of composite grasshopper samples was 0.22 mg/kg. Ass~~ing a 50 g (0.05 kg) bird and a grasshopper weight ',f 0.89 g (0.00089 kg), ingestion of a single grasshopper represents about U L_ the LD50 for red-winged blackbirds and about 0.33% of the LD50 for house sparrows. Residue analysis did not include toxic carbamate metabolites of furadan. Risks to insectivorous birds are mitigated by timing of migrations and low furadan residues in grasshoppers, but additional experiments with caged birds may be useful. The primary risk appears to be to granivorous birds, principally horned larks and meadowlarks. Data on residue levels of wheat seeds sprayed with furadan and left on the soil surface should be obtained. Experiments with caged horned larks and meadowlarks would establish whether consumption of treated seeds is toxic. If treated seeds are toxic to these birds, a more difficult task would be to establish with what frequency mortality occurs under normal planting conditions. Repetitive transects appear to be the only means to determine this. Effects on Canada geese.--Thirteen geese were captured by cannon net on 7 December and transported to the Briggsdale study site. One goose died enroute. We finished marking, wing clipping, and handling birds about 1330 hours. After some initial pacing along the fence, birds began to feed. Feeding appeared heavy throughout the following day. Geese were removed and sacrificed just before dusk on 8 December. Brain acetylcholinesterase activity was unaffected by this exposure (Table 1). Two wheat samples collected at. the Briggsdale site prior to this study had furadan residues of 0.06 ppm. Residues at the Loveland site were 0.84 and 1.16 ppm while treated wheat at the Severance site had residues of 0.34 and < 0.05 ppm. Variation among sites seems large since all 3 were planted at the same time using the same equipment and applicator. It is not surprising that no effect was observed in geese at the Briggsdale site given these residues. In retrospect the Loveland site would have been a better test. It's possible that furadan at the Briggsdale site had been metabolized in the soil or wheat and was still present as a toxic metabolite. Apparently inconsequential fields with higher residue needed. impacts on Canada geese may not be typical of levels. Additional work on residue decay is �264 Table 1. Brain acetylcholinesterase activity of 12 Canada geese after feeding in pens containing untreated, 50% furadan treated or 100% furadan treated wheat. Treatment Mean ± S.D. Untreated 50% Treated 1007. Treated 1502 967 1007 959 867 751 1296 975 1072 1256 881 964 1108 ± 263 972 LITERATURE ± 234 1043 ± 162 CITED Anonymous. 1988. The Russian wheat aphid: a serious new pest of small grains in the Great Plains. Great Plains Agric. Counc. Pub1. 124. 6pp. Hill, E. F., and W. J. Fleming. 1982. Anticholinesterase field monitoring and diagnosis of acute poisoning. Chem. 1:27-38. poisoning of birds: Environ. Toxico1. McEwen, L. C., L. R. DeWeese, and P. Sch1adwei1er. 1986. Bird predation on cutworms (Lepidoptera: Noctvidae) in wheat fields and chlorpyrifos effects on brain cholinesterase activity. Environ. Entomo1. 15:147-151. Peairs, F. B., K. G. Beck, W. M. Brown, Jr., H. F. Schwartz, and P. Westra. 1988. 1988 Colorado pesticide guide - field crops. Colorado State Univ., Coop. Ext., Agric. Exp. Stn. XCM-45. Rattner, B. A., L. Sileo, and C. G. Scanes. 1982. Hormonal responses tolerance to cold of female quail following parathion ingestion. Biochem. Physiol. 18:132-138. and Pest. Smith, G. J. 1987. Pesticide use and toxicology in relation to wildlife: organophosporous and carbamate compounds. U.S. Dep. Inter., Fish and Wild1. Servo Resour. Publ. 170. 17lpp. Snyder, W. D. 1984. Ring-necked pheasant nesting ecology on the high plains. J. Wi1d1. Manage. 48:878-888. Prepared bY~ Thomas E. Remington Wildlife Researcher and wheat farming � https://cpw.cvlcollections.org/files/original/b4588c857ba1ad76abd3bef193ce974d.pdf b377ee66f3f44046e3220a0171ffc9a7 PDF Text Text 1 Colorado Division Wildlife Research July 1990 of Wildlife Report JOB PROGRESS REPORT State of Colorado Project No. W-lS3-R-3 Work Plan No. Job No. _ Multispecies 7 Period Covered: Author: ~l Mammals Research Research Terrestrial Research Publication, Editing and Library Service July 1, 1989 - June 30, 1990 J. A. Boss, L. H. Carpenter Personnel: R. B. Gill, N. McEwen, L. Lovett, K. Williams, all Mammals Researchers C. Hanna and Abstract During the 1989-1990 Segment, 4 books were purch~sed for permanent reference by DOW personnel. Twenty-six additional publications were located, ordered, and obtained free of charge for use. Six of these were purchased, obtained on Interlibrary Loan, or given to the library. An Additional 431 individual references requested by Mammals Researchers were located by library staff and made available for use. About 10 of these requests were not available locally and were obtained through interlibrary loans. Sixteen manuscripts were published in various journals. ��3 MAMMALS PUBLICATION, EDITING AND LIBRARY SERVICES Jacqueline A. Boss and Len H. Carpenter P. N. OBJECTIVES To provide a centralized support program for manuscript editing and library services to facilitate publishing results of research conducted in projects 01-03-047 - 11700 and 01-03-048 - 11700 and 16700. SEGMENT OBJECTIVES 1. To provide coordinated and efficient editing and library services and publish findings for all Colorado Mammals Research programs. 2. To provide for the centralized support program for Mammals Research editing, library, and publishing services so that Mammals Research scientists can be most efficient in publishing results of their research. SUMMARY OF SERVICES Publications purchased with Mammals Research funds and placed in the Research Center Library Callicott, J. B. and A. Leopold. 1989. In defense of the land ethic: essays in environmental philosophy. State Univ. of New York Press, Albany, NY. 325pp. King, C. 1989. Natural history of weasels and stoats. Assoc., Ithaca, NY. 253pp. Comstock Publishing Phillips, B. G. and A. M. Phillips III. 1987. Annotated checklist of vascular plants of Grand Canyon National Park. Museum of Northern Arizona, Flagstaff, AZ. 79pp. Yates, S. 1988. Adopting a stream: a Northwest handbook. Press, Seattle, WA. l34pp. Univ. of Washington Publications obtained free or at low cost In addition to books purchased with Federal Aid Funds, about 26 free reports and short publications from state or federal agencies or from private sources were located, ordered, and obtained for use by Mammals Research personnel. Theses purchased, obtained on Interlibrary Loan or as gifts for use by Researchers Ackerman, B. B. 1989. Visibility bias of mule deer aerial census procedure in Southeast Idaho. Ph. D. Diss., Univ. of Idaho, Moscow, ID. l16pp. �4 Banks, J. L. 1964. Fish species distribution in Dinosaur National Monument during 1961 and 1962. M. S. Thesis, Colorado State Univ., Fort Collins, CO. 96pp. Bertagnoli, G. G. 1986. Availability and use of foods by black bears in Wisconsin. M. S. Thesis, Univ. of Wisconsin-Stevens Point, Stevens Point, WI. 42pp. Brian, N. J. 1982. A preliminary study of the riparian coyote willow communities along the Colorado River in Grand Canyon National Park, AZ. M. S. Thesis, Northern Arizona Univ., Flagstaff, AZ. 84pp. Gerhardt, D. R. 1989. Population dynamics, movement and spawning habitat of channel catfish in the Powder River System, Wyoming-Montana. M. S. Thesis, Univ. of Wyoming, Laramie, WY. 94pp. Koehler, G. M. 1988. Demography of a low productivity Ph. D. Diss., Univ. of Idaho, Moscow, ID. 32pp. bobcat population. Neve, L. L. 1976. Live history of the roundtail chub, Gila robusta grahami, at Fossil Creek, Arizona. M. S. Thesis, Northern Arizona Univ., Flagstaff, AZ. 46pp. Schreckengast, G. E. 1988. River otter reintroductions M. S. Thesis, West Virginia Univ., Morgantown, WV. in West Virginia. l02pp. Staines, B. W. 1970. Management and dispersion of a red deer population in Glen Dye, Kincardineshire. Ph. D. Diss., Univ. of Aberdeen, Aberdeen, Scotland. pp. Stevens, L. E. 1984. Aspects of invertebrate herbivore community dynamics Tamarix chinensis and Salix exigua in the Grand Canyon. M. S. Thesis, Northern Arizona Univ., Flagstaff, AZ. l62pp. on 1989. Mechanisms of riparian plant community organization and succession in the Grand Canyon, Arizona. Ph. D. Diss., Northern Arizona Univ., Flagstaff, AZ. llSpp. Tango, P. J. 1988. Home range .. of reintroduced river otter in West Virginia. M. S. Thesis, West Virginia Univ., Morgantown, WV. l17pp. Whiting, W. R. 1988. Implementing prohibitive policy: the Endangered Species Act. M. S. Thesis, Univ. of Wisconsin-Green Bay, Green Bay, WI. 80pp. Zabel, C. J. 1986. Reproductive behavior of the red fox (Vulpes vulpes) : longitudinal study of an island population. Ph. D. Diss., Univ. California, Santa Cruz, CA. 98pp. Zeckmeister, M. 1985. Population characteristics of unexp10ited beaver muskrat populations. M. S. Thesis, Univ. Wisconsin-Stevens Point, Stevens Point, WI. 4Spp. Reference document location and and delivery The Research Center Library staff also located and delivered about 431 individual articles on request for Mammals Researchers during this segment; a �5 about 10 were not available locally and were obtained through interlibrary loan procedures. Manuscripts published Job Progress Reports; Federal Aid. All studies. Anderson, A. E., D. C. Bowden, and D. E. Medin. 1990. Indexing the animal fat cycle in a mule deer population. J. Wildl. Manage. 54: (In Press) Bear, G. D., G. C. White, L. H. Carpenter, R. B. Gill, and D. J. Essex. Evaluation of aerial mark-resighting estimates of elk populations. Wildl. Manage. 53:908-915. 1989. J. Carpenter, L. H. 1989. Colorado's liability for big game damage to livestock forage. Pages 96-100 in Ninth Great Plains wildlife damage control workshop proceedings, April 17-20, 1989, Fort Collins, Colorado. USDA Forest Service, Rocky Mtn. Forest and Range Exp. Stn., Fort Collins, CO. Freddy, D. J., E. R. Ryland, and R. M. Hopper. 1990. In L. A. Renecker (ed.) Proceedings of the Second International Wildlife Ranching Symposium, Edmonton, Alberta, June 1990. Colorado's wildlife ranching program: the Forbes Trinchera experience. (In Press) Gill, R; B., and T. D. I. Beck. 1990. Black bear management plan. Div. of Wildlife, Div. Rep. No. 15. 44pp. Colorado Green, R. A., and G. D. Bear. 1990. Seasonal cycles and daily activity patterns of Rocky Mountain elk. J. Wildl. Manage. 54:272-279. Harlow, H. J., T. D. I. Beck, L. M. Walters, and S. S. Greenhouse. 1990. Seasonal serum glucose, progesterone, and cortisol levels of black bears (Ursus americanus). Can. J. Zool. 68:183-187. Hobbs, N. T. 1990. Diet selection by generalist herbivores: a test of the linear programming model. Pages 395-414 in Hughes, R. N. (ed.), Behavioural mechanisms of food selection. NATO ASI Series. Vol. G 20: Ecological Sciences. Springer-Verlag, Heidelberg, NY . . 1990. Interactions of vertebrates with crown fire regimes: chemical, structural, and spatial dynamics. In R. D. Laven and P. N. Omi (eds.), Pattern and process in crown fire ecosystems. Princeton University Press, Princeton, NJ. (In Press) __ , M. W. Miller, J. A. Bailey, D. W. Reed, and R. B. Gill. 1990. Biological criteria for introducing large mammals: using simulation models to predict impacts of competition. Trans. N. Amer. Wildl. and Natur. Resour. Conf. 55: (In Press) _____ , and T. A. Hanley. 1990. Habitat evaluation: do use/availability data reflect carrying capacity? J. Wildl. Manage. 54: (In Press) Kufeld, R. C., D. C. Bowden, and D. L. Schrupp. 1989. Distribution and movements of female mule deer in the Rocky Mountain foothills. J. Wildl. Manage. 53:871-877. �6 _____ , D. C. Bowden, and D. L. Schrupp. 1989. and why. Colo. Outdoors 38(5):36-37. Where the deer are - Miller, M. ~., fi._T. Hobbs, and M. C. Sousa. 1990. Detecting stress responses in Rocky Mountain bighorn sheep (Ovis canadensis): reliability of cortisol concentrations in urine and feces. Can. J. Zool. 68: (In Press) Smith, M. H., K. T. Scribner, L. H. Carpenter, and R. A. Garrott. 1990. Genetic characteristics of Colorado mule deer (Odocoileus hemionus) and comparisons with other cervids. Southwest. Nat. 35:1-8. Trust, K. A., M. ~. Miller, J. K. Ringelman and I. M. Orme. 1990. Effects ingested lead on antibody production in mallards (Anas platyrhynchos). J. ~ildl. Dis. 26: (In Press) of ~alsh, N. E. 1990. Responses of bull elk to simulated elk vocalizations during rut. M. S. Thesis, Colorado State Univ., Fort Collins, CO. 97pp. Prepared by Len H. Carpenter wildlife Research Leader �Colorado Division Wildlife Research July 1990 of Wildlife Report JOB PROGRESS REPeRT State of Colorado Project No. ~W_-~1~5~3_-~R~-~3 _ Mammals Research Work Plan No. __~2 _ Deer Investigations Development of Census Methods for Deer in Plains Riverbottom Habitats Job No. 7 Period Covered: July 1, 1989 - June 30, 1990 Author: R. C. Kufeld Personnel: D. Bowden, D. Younkin ABSTRACT White-tailed and mule deer, radio-collared in the South Platte Riverbottom during January and February, 1987, 1988, and 1989, were located at approximately 2-week intervals throughout the segment. A computer system for analyzing data from tracking radio-collared deer during the course of the study was prepared. Movement data from 29 radio-collared deer which have died or whose transmitters have quit and movement patterns, to date, of 21 radiocollared deer that are still alive and currently being monitored, suggest that deer in the South Platte Riverbottom fall into 4 movement categories. These categories include resident deer which occupy the same general area yearlong, and migratory deer which occupy separate winter and summer areas. These patterns were observed for both mule deer and whitetails and are described in detail. ��9 DEVELOPMENT OF CENSUS METHODS FOR DEER IN PLAINS RIVERBOTTOM HABITATS Roland C. Kufeld P. N. OBJECTIVES 1. To determine seasonal movements and home range size of white-tailed and mule deer in plains riverbottom habitats. 2. To develop and test methods for estimating size of deer populations in plains riverbottom habitats. SEGMENT OBJECTIVE To determine seasonal movements and home range size of white-tailed and mule deer in plains riverbottom habitats. STUDY AREA The study area was described by Kufeld (1989). METHODS AND MATERIALS White-tailed and mule deer, radio-collared in the South Platte Riverbottom during January and February, 1987, 1988, and 1989, (Kufeld 1987, 1988, 1989) were located at approximately 2-week intervals throughout the fiscal year. Deer locations were recorded by UTM coordinates and plotted on USGS 1:50000 scale maps. Vegetation type for each deer location was also recorded. A computer system for analyzing data from tracking deer that were radiocollared during the course of this study was prepared. The system was used to analyze part of the data collected for radio-collared deer that have died or whose transmitters have quit prior to writing of this report for purposes of including that information in this report. Since many of the radio-collared deer are still alive and are still being located at 2-week intervals, a complete analysis of data must await termination of monitoring of those deer. The area which encompassed all observed locations for an individual deer during a specific time period was determined using the minimum convex polygon feature in version 1.2 of program McPaal developed at the Smithsonian Institution, Front Royal, VA 22630. Median activity centers for each such group of locations were computed according to Barry et al. (1984). Distances along the river channel between selected points, herein referred to as "river km", airline distances between selected deer locations or median activity centers and the closest point on the river channel, and airline distances between selected deer locations or median activity centers were computed using version 6.03 of program SAS, SAS Institute Inc., SAS Circle, Cary, NC 27512. During initial planning for this study, which included discussions with Colorado Division of Wildlife Northeast and Southeast Regional personnel, it was decided to concentrate first efforts on a single river system. The South Platte River from Platteville, Colorado to the Nebraska State Line was �10 selected as the initial study area, and is the area where the study has been conducted since 1987. This study area is in the Northeast Region. It was agreed that upon completion of work on the South Platte other river situations would be addressed. The South Platte phase is expected to terminate during the winter of 1990-91. During this project segment meetings were held with Southeast Regional personnel and plans were made to shift emphasis to selected rivers in southeastern Colorado. RESULTS AND DISCUSSION Movement data from 29 radio-collared deer which have died or whose transmitters have quit (Tables 1 through 4), and movement patterns, to date, of 21 radio-collared deer that are still alive and are currently being monitored, suggest that deer in the South Platte Riverbottom fall into 4 movement categories. Since data in Tables 1 through 4 represent only part of the sample of radio-collared deer in the study, because some study deer are still alive, mean areas which encompassed all deer locations by species and sex and comparisons of means between species, sexes, or movement categories are not presented for data in Tables 1 and 2. A much more in-depth analysis of these data will be presented in a subsequent report when monitoring of remaining radio-collared deer has been completed and all data derived from the study become available. This will include information on habitat selection, more on movements, and a discussion of how findings from this study apply to development of a census technique for plains riverbottom habitats. Movement categories are described as follows: Movement Category 1: Deer which remain in or near the South Platte Riverbottom throughout the year and occupy a segment of riverbottom to 20 km in length. from 2 Movement Category 2: Deer which remain in or near the riverbottom throughout the year, but travel relatively long distances (30 to 120 km) up and/or downstream. Movement Category 3: Deer that spend most of the year, including winter, in or near the riverbottom, but during late spring or summer they leave the riverbottom and spend at least several weeks out on the plains. Radiocollared deer that exhibited this behavior returned to the same section of riverbottom sometime before December. During the period they were in the riverbottom they occupied a riverbottom segment from 1 to 12 km in length. Those that survived more than 1 year occupied the same general area each summer (Tables 1 and 2). Departure and return dates, length of time these radio-collared deer spent away from the riverbottom, and distance they traveled from the river varied among individual deer within the 3rd category. Some individual deer were gone from the riverbottom for only a few weeks, while others were gone for months (Tables 3 and 4). Some individuals moved to plains habitats only 1 or 2 km from the riverbottom, while others traveled to plains habitats more than 35 km from the riverbottom (Tables 1 and 2). Movement .Category 4: Deer that live on the plains and rarely visit the riverbottom. Several radio-collared deer, including those which have since died and those still alive were trapped in the riverbottom, but shortly after being tagged they moved out onto the plains and were rarely �11 or never again located in the riverbottom during more than 2.5 years of monitoring. All 4 categories have radio-collared deer of both species (mule deer and whitetails) except that no radio-collared mule deer are in category 2. Most of the radio-collared deer are in category 1 with category 3 ranking second. Categories 2 and 4 contain relatively few animals (Tables 1 and 2). The white-tailed doe in movement category 2 (Tables 2 and 4) had a complex movement pattern which involved several centers of activity, all of which were located in the riverbottom but separated by relatively long distances. The deer was radio-collared 1-18-87, 1.9 river km downstream from Weldona, Colorado. Between 3-6-87 and 8-27-87, 14 locations were in an area of 0.3 km2 centered 22.7 river km upstream from its trapsite and 4.2 river km upstream from Orchard, Colorado. Between 9-12-87 and 10-13-87 it moved downstream 120.6 river km to a point near Iliff, Colorado. Between 10-27-87 and 5-28-89, 42 locations were in an area of 8.6 km2 centered 17.0 river km upstream from its easternmost location near Iliff. This area was in the vicinity of Dune Ridge State Wildlife Area, which is located about 8 km southwest of Sterling, Colorado. During the 10-27-87 to 5-28-89 period, however, it made 2 trips to a point centered 32.6 river km upstream near Messex State Wildlife Area. It was located there 3 times between 11-1 and 11-29-88, and 3 times between 3-28 and 4-24-89. Those 6 locations were in an area of 0.9 km2• After each visit to the Messex area it returned to the vicinity of Dune Ridge and it died there on 5-28-89. White-tailed doe 9791, in category 3 (Tables 2 and 4), spent the winter in the riverbottom in a relatively small area, but during summer it travelled such long distances that it could rarely be found. It was radio-collared 2-3-87 5.9 river km downstream from of Hardin, Colorado. It spent the winters of 1986-87 and 1987-88 in the riverbottom in the vicinity of its trapsite. Between 4-17 and 9-12-87 it could not be located despite an intensive search by aerial telemetry of the area encompassed by a perimeter extending along Interstate 25 from the Wyoming State Line south to the U.S. Air Force Academy, east to Limon, north to the Wyoming State Line, and west to Interstate 25. The South Platte River from Platteville to Julesburg, Colorado, was also searched. On 9-12-87 the deer was back in the vicinity of its trapsite where it spent the winter. It left its winter area the following April and on 4-2988 it was located near Roggen, Colorado, 17.5 km from its winter median activity center and 13.5 km from the nearest point on the South Platte River Channel. On 7-25-88 it was relocated 19.8 km south-southeast of Akron, Colorado, 105.9 km fro. its winter median activity center and 51.6 km from the nearest point on the South Platte River Channel. Two weeks later it could not be found within a 24-ka radius of that location, and it was never again relocated. Study area priorities that have been established by Colorado Division of Wildlife Southeast Region personnel for continuing the study in the Southeast Region are: (1) Arkansas River from John Martin Reservoir to the Kansas State Line, and the lower 16 km of Big Sandy Creek, (2) Arkansas River from Fowler to LaJunta, and Fountain Creek from Colorado Springs to Pueblo, (3) South Republican River from treeline to the Kansas State Line, and (4) Big Sandy Creek from Simla to Hugo. Objectives and procedures will be the same as those employed during the South Platte River phase of the study. Fieldwork for the southeastern Colorado phase will begin in December, 1990. �12 LITERATURE CITED Berry, K. J., P. W. Mielke, and K. L. Kvamme. 1984. Efficient permutation procedures for analysis of artifact distributions. Pages 54-74 in H.J. Hietala, ed. Intrasite spatial analysis in archaeology. Cambridge Univ. Press, Cambridge, U.K. Kufeld, R. C. 1987. Development of census methods for deer in pla~ns riverbottom habitats. Colo. Div. Wildl., Wildl. Res. Rep. July (1):13-19. 1988. habitats. Development of census methods for deer in plains Colo. Div. Wildl. Res. Rep. July (1):11-21. riverbottom 1989. Development of census methods for deer in plains habitats. Colo. Div. Wildl. Res. Rep. July (1):11-17. riverbottom Prepared by --,-~__:._ __ ~_·_C..:::..._..;_%dc--l· Roland C. Kufeld Wildlife Researcher r-W_' _.~ __ �13 Table 1. Movement patterns of radio-collared mule deer tagged in the South Platte Riverbottom that have died or whose transmitters have ~it. Deer location Distance from MAC Distance between winter and area ~Km2l' to river ~knQ2 Movement 10 No_ Sl.JIIlIer Annual Annual Winter Winter Sl.JIIlIer sl.Jllller MACS ~I(ml Sex categorJ! 0.2 9521 4.8 F 9522 2.7 0.2 9551 5.3 0.2 9652 3.1 0.1 9671 5.7 0.2 9732 6.6 0.2 M 3 F 9382 9291 9539 9631 9970 9452 9651 9701 4 F 9859 Area which enc~ased 0.2 3.1 0.5 0.6 11.1 1.6 1.9 3.6 6.43 0.9 0.4 0.2 0.3 0.2 5.6 16.7 26.0 1.7 6.8 17.8 29.1 1.9 3.1 4.8 5.2 11.43 31.4 3.5 0.7 0.3 0.0 1.6 14.7 0.9 1.5 14.8 1.6 28.2 2.3 all deer locations. ~istanc:e frOllMedian Activity Center (MAC) to the nearest point on the river channel (KIll).Median Activity Centers were computed as described by Berry et al. (1984). 3This deer spent 2 s~rs in this area. Table 2. MOVeMent patterns of radio-collared white-tailed deer tagged in the South Platte Riverbottom that have died or whose transmitters have ~it. Deer location Distance from MAC Distance between area (Km2l' to river 0("Q2 winter and Movement SI.Jlllle!" MACS (Km) Sex 10 No. Annual Winter Summer Annual Winter Summer categorJ! F 0.3 15.9 9252 6.1 9262 0.3 9731 0.3 4.2 8_3 9921 0.1 0_3 9971 6.8 M 2 F 3 F M 9351 9361 9391 9770 6.2 7.7 18.5 10.8 0.3 0.3 0.2 0.1 95113 9370 9780 'Area which enc~ssed 13.4 10.4 3.1 0.1 0.2 36.0 34.3 13.7 7.8 0.3 2.1 0.5 0.1 19.2 17.3 23.8 28.6 all deer locations ("",2). ~istance frOllMedian Activity Center (MAC) to the nearest point on the river chamel Activity Centers were ca.puted a. described by Berry et al. (1984). ~ovements of this deer are described in the text. (KIll). Median ""ean distance frOll3 Median Activity Centers (Orchard, Messex, DI.I"eRidge) to the nearest point on the river channel (KIll). �14 Table 3. N~r of locations and dates when radio-collared IIIJledeer in movement category 3 were away from the riverbottom during late spring and/or summer. Movement No. of locations Dates of summer plains ~riod Annual category Sex lJinter Summer 1st year 2nd year 10 No. 49 1 F None 9521 29 None 9522 None 9551 32 None 32 9652 64 None 9671 None 9732 16 9382 3 4 F F None 35 9291 9539 9631 9970 18 8 23 11 10 13 21 4 6-29 5-25 5-30 7-17 9452 9651 9701 32 10 8 20 14 12 5-25 to 8-17 6-30 to 10-28' 5-30 to 10-13 9859 8 12 5-30 to 10-13 to to to to 11-1 10-22' 10-13 6-8 8- 18' to 10-28 6-19 to 11-28 'Deer found dead or was killed on plains as of last date. Table 4. N~r of locations and dates when radio-collared wnite-tai led deer in movement category 3 were away from the riverbottom during late spring and/or summer. Movement Dates of summer plains ~riod No. of locations 1st year 2nd year category Sex 10 No. Annual lJinter Summer 15 None F 9252 24 None 9262 None 9731 15 None 9921 48 None 9971 17 None None None None M 9351 9361 9391 9nO 25 56 22 22 2 F 9511 66' 3 F 9809 9791 27 21 6 M 9370 9780 10 18 5 7 'Movements of this deer are described in the text. 1>eer found dead or was ki lled on plains as of last date. 6-1 1 to, 7-29 7-15 to 9-7 2 6-30 to 9-12 " �15 Colorado Division Wildlife Research July 1990 of Wildlife Report JOB PROGRESS REPORT State of Colorado Project No. W-153-R-3 Work Plan No. ~3 Job No. 2 Period Covered: Author: Mammals _ Research Elk Investigations Trapping, Transporting, of Elk at Livestock-Elk and Maintenance Grazing Study July 1, 1988 - June 30, 1990 G. D. Bear Personnel: B. Rakestraw, C. Woodward, C. Reichert, M. Bauman, B. Dupire, R. Harthan J. Madison, J. Haskins, ABSTRACT A total of 212 elk were trapped near Meeker, Craig, and Hayden, Colorado, to provide cows for stocking the experimental pastures on the Little Snake River Wildlife Area north of Maybell, Colorado. The pastures were stocked with 54 adult cows in late December, which were removed in early April. heavy snow accumulations, cold temperatures, and depleted range conditions contributed to heavy elk losses in the heavy and medium density pastures during the last 6 weeks of the grazing period. Telemetry-collared elk and visual observations indicated elk occupied the Little Snake River winter range from December through April. The collared elk then migrated to the traditional summer ranges north and south of the CraigHayden area. ��17 TRAPPING, TRANSPORTING, AND MAINTENANCE OF ELK AT LIVESTOCK-ELK GRAZING STUDY George D. Bear P. N. OBJECTIVE To provide assistance to the Livestock-Elk Grazing Study near Maybell, Colorado, by capturing and maintaining an experimental elk herd. SEGMENT OBJECTIVES 1. To trap, mark, stock, and maintain 54 adult female elk in the livestockelk grazing pasture complex during the period December 15 to April 20. 2. To determine seasonal movements released from the pastures. of radio-collared elk after they are METHODS AND RESULTS Capturing and Handling of Elk The usual amount of effort was put into repa~r~ng the fences and corrals the elk pens, repairing the elk traps, securing hay and other trapping supplies during the late summer and fall. at In late November and December trapsites were established at the following locations: (1) 5 mi southeast of Meeker near the Environmental Plant Center, (2) 28 mi east of Meeker on the Marvine Ranch, (3) Trapper Mine - 5 mi south of Craig, and (4) Senica Mine - 10 mi southeast of Hayden. These sites were selected because they were not open to hunting during the late elk hunting seasons (December-January). Details for handling elk at the trap sites and in the corral at the livestock-elk grazing pasture complex are presented in previous Federal Aid Reports. A total of 212 elk (101 cows, 84 calves, 12 yearling cows, 11 spikes, and 4 adult bulls) were captured from December 1 to February 14. Fifty-four cows were initially used to stock the experimental pastures (December 18 to December 16). Additional cows were kept in the pole-corral and hub to be used as replacements when needed. Weather conditions were very mild throughout January and the first half of February. There was less than 4 in of snow on the ground at anyone time. In late January the lack of snowcover made it necessary to provide water for the elk in the pastures. Daytime temperatures ranged from the mid-30s to low 50s; nighttime temperatures ranged from 0 to -lO°F. The elk in the experimental pastures did very well with these conditions. Elk in the high density pastures were replaced with new animals in February, as scheduled. A severe winter storm (38 in of as well as, cold stormy weather stressed the elk so weight loss and medium density pastures. A snow and temperatures -29°F) in late February, during the first week of March, severely was very noticeable for the elk in the high total of 27 elk died of malnutrition during �18 the following 6 weeks and left many others very weak and in poor condition. Sixteen of these 27 elk died in the high density pastures, in spite of the February rotation. The remaining 11 deaths occurred in the medium density pastures; none in tte low density pastures. Because of the poor range conditions and high elk mortality, elk were removed from the pastures a week early (April 7-12). Shrubs grew very well (10-12 in) last summer, in spite of the dry conditions; but forbs and grasses did not put on any new growth. Many of these shrubs (Artemisia ~, Chrysothamnus sp., Purshia tridentata, Eurotia sp.) maintained the elk in early winter; however, they were severely overgrazed. With the food supply reduced, elk in the high and medium density pastures were forced to subsist primarily on Artemisia tridentata in late winter. Elk in the low density pastures fared better; here the shrubs were not overgrazed, and some residue grasses remained. Elk Movements Migration patterns this year were similar to those reported in previous years. Although the December hunting season kept the elk moving around in smaller groups, they were commonly observed in the vicinity of the study area after December 1, then throughout the winter. As the snow melted and green grass became apparent in April, the elk started moving eastward toward higher elevations. An aerial survey on April 9 revealed a large concentration of elk north of Craig and only a few scattered groups in the vicinity of the study area. By early May, only an occasional elk was seen on the low winter range. Radio-collared elk migrated to the traditional high summer ranges north of Craig and Hayden and summer ranges at the head of Williams Fork and Maraposa Creek south of Craig and Hayden. Prepared by ~'6'.~~ ceorgeaear Wildlife Researcher �19 Colorado Division of Wildlife Wildlife Research Report July 1990 JOB PROGRESS REPORT State of C~o~l~o~r~a~d~o~ _ Project No. W-lS3-R-3 Work Plan No r ---::.3 _ 6 Job No. Period Covered: Author: Mammals Research Elk Investigations Effect of Elk Harvest Breeding Biology Systems on Elk July 1, 1989 - June 30, 1990 D. J. Freddy Personnel: M. Miller, T. Hobbs, C. Wetherill, M. Cousins, J. Olterman, Colorado Division of Wildlife, G. White and N. Walsh, Colorado State University, and E. Ryland, Forbes Trinchera Ranch. ABSTRACT We evaluated effects of time and ambient temperature on progesterone concentrations in plasma and serum collected from elk. Progesterone concentrations differed by blood type (~= 0.01), elk (~- 0.002), and time since collection (l~ 0.024). Concentrations were greater in serum than in plasma and increased in both serum and plasma with time since collection. These results were surprising, and we plan to repeat the experiment in 1991. Reproductive tracts were collected from female elk in December by hunters on the Forbes Trinchera Ranch. Fetal sex ratios for 1989 were 12 M:24 F and were 49 M:Sl F for years 1986-89, pooled. In 1989, male fetuses were smaller than females in body size and weight. Size of male fetuses has declined since 1987 to a greater extent than size of female fetuses. Abstracts are included from the following publications: elk using p:r~gesterone assays and real-time ultrasound" elk to simulated elk voca.l.fz a t i.ons during rut". . "Pregnancy testing and "Responses of bull ��21 JOB PROGRESS REPORT EFFECT OF ELK HARVEST SYSTEMS ON ELK BREEDING BIOLOGY David J. Freddy P. N. OBJECTIVE To evaluate effects of harvest systems on breeding biology of elk. SEGMENT OBJECTIVES 1. Continue to determine reproductive status of elk and deer on the Forbes Trinchera Ranch using fetal collections and blood assays. 2. Provide administrative and technical assistance to the graduate student project investigating rutting-bugling behavior of bull elk on the Forbes Trinchera Ranch. 3. Prepare manuscript for publication summarizing reproductive measurements. ACKNOWLEDGMENTS We thank T. Thorne and H. Long of the Wyoming Game and Fish Department, Sybille Research Station, for providing blood from adult female elk for progesterone assays. METHODS AND MATERIALS Pregnancy Testing-Blood Assays We evaluated effects of time and ambient temperature on progesterone concentrations in plasma and serum collected from elk. Using a syringe, we collected 30 cc of blood via the jugular vein from each of 2 pregnant and 1 nonpregnant adult female elk on 2 May 1989. Blood from each elk was immediately divided into 20, 1.5 cc subsamples and maintained at ambient temperature during the 1 hr when blood was collected from all 3 elk. These 20 subsamples were allocated as follows: (1) 10 were placed into heparinized glass tubes (plasma) with 5 each to be stored at 24 C and 4 C for up to 24 hrs, and (2) 10 were placed into non-heparinized glass tubes (serum) with 5 each also stored at 24 C and 4 C for up to 24 hrs. At 1 hr post-collection (Time 0), 1 sample from each blood and temperature treatment for each elk (4 samples/elk) was centrifuged and plasma and serum were harvested and frozen at -18 C and all remaining samples were placed in either a refrigerator or container maintained at ambient temperature. At 3, 6, 12, and 24 hrs thereafter, samples were removed from their respective thermal environments, centrifuged, and plasma and serum were harvested and frozen at -18 C. Progesterone concentrations in each sample were determined by radioimmunoassay on 8 May by the Physiology Laboratory, Colorado State University, Ft. Collins, CO. Progesterone concentrations were analyzed using a randomized complete block (elk) for 2 way factorial with repeated measures (serial blood samples for time intervals) (Proc GLM, SAS 1988). �22 Fetal Collections Reproductive tracts of female elk were collected by hunters on the Forbes Trinchera Ranch on 9-11 and 16-18 December 1989. Tracts were deposited at check stations and kept cool until processing. Pregnancy status was determined from the presence of fetuses, embryos, and developed uterine tissue. Questionable uteri were preserved for later examination. Fetal measurements were made on fresh specimens (subsequently preserved) and followed definitions of Armstrong (1950). Fetal age was determined from growth curves of Morrison et al. (1959). Compliance with the request to collect reproductive organs was poorer than during the previous 3 years. RESULTS AND DISCUSSION Blood Assays Progesterone concentrations were most affected by blood type, individual elk, and time and less so by temperature. Overall, progesterone differed among elk (~= 0.002) and between plasma and serum ( ~ = 0.011) (Fig. 1). Progesterone concentrations were not different between ambient and refrigerated samples (~= 0.783), although there was some tendency for higher concentrations of progesterone in refrigerated samples of both plasma and serum (Fig. 1). There were no interactions between temperature and blood type (~~ 0.65). This pattern of results also generally held true within each time period. Progesterone concentrations were affected by time since collection (~ = 0.024), most noticeably by an increase in plasma samples at 24 hrs postcollection (Fig. 1). There were no interactions between time and elk, temperature, or blood type (~> 0.40). we were surprised by the differences in progesterone values between plasma and serum, by the lack of differences between refrigerated and ambient samples, and by the lack of a steady decline in progesterone over time. These results contrast with observations in muskoxen (Rowell and Flood 1987). ~e are cautious about these results and plan to repeat the experiment with captive elk in January and May 1991. ~e also have delayed submitting a completed an internally reviewed manuscript to The Journal of ~ildlife Management (Appendix A) pending the outcome of experiments in 1991. Fetal Pregnancy Collections-Forbes Elk Rates Pregnancy rates for adult cows (~ 1 yr old) averaged 78% with a range of 7782% from 1986-89 (Table 1). Pregnancy rates were 90% for prime-age adults but were 20% for yearlings and 60% for cows aged 11+ years. Pregnancy was not observed in calves. Litter size was 1 in all but 3 (2%) of 151 litters. The 3 sets of twins were: 2 females and 1 male and 1 female in cows 9 and 12 years old, respectively in 1986, and 2 females in a 6 year old cow in 1989. Infected uteri not capable of supporting pregnancy occurred in 5 (3%) of 193 adults examined. Infections were in 1 yearling and 4 cows ~ 13 years old with 3 found in 1986, 1 in 1988, and 1 in 1989. �23 Fetal Sex Ratios In 1989, the fetal sex ratio was 12 M:24 F and the ratio pooled among years was 49 M:5l F (Table 2). Fetal sex ratios continued to alternately favor males and females between years (Table 2, Freddy 1989). Fetal Size Male fetuses length, and been larger declined in Conception were hind than size slightly smaller than females in body weight, crown-rump foot length in 1989 (Table 3). In previous years, males have females (Freddy 1989). Since 1987, male fetuses have to a greater extent than females (Table 3). Dates Conceptions occurred from 18 September to 26 October, a 39 day interval. Median conception date was 2 October which was the latest date observed since 1986 and about 4 days later than the average median date from 1986-88 (Freddy 1989, Fig. 2). Fetal Collections-Forbes Deer Collections of female mule deer for reproductive and physiological measurements did not occur as planned in March, 1989. Scientific collection of deer was not considered prudent at this time. Bugling Behavior of Bull Elk The study "Responses of bull elk to simulated elk vocalizations" was completed in June, 1990, by master of science degree candidate Noreen Walsh, Department of Fishery and Wildlife Biology, Colorado State University. The abstract of this thesis is presented in Appendix B. LITERATURE Armstrong, R. A. 1950. Fetal development Amer. Mid. Nat. 43:650-666. CITED of northern white-tailed deer. Freddy, D. J. 1989. Effect of elk harvest systems on elk breeding Colo. Div. of Wildl. Game Res. Rep. July: 35-60. biology. 1959. Breeding seasons Morrison, J. A., C. E. Trainer, and P. L. Wright. in elk as determined from known-age embryos. J. Wildl. Manage. 23:27-34. Rowell, J., and P. F. Flood. 1987. concentration between collection 51:901-903. Changes in muskox blood progesterone and centrifugation. J. Wildl. Manage. SAS Institute, Inc. 1988. SASjuser's Inc., Cary, N. C. 1028pp. Prepared by guide, version 6.03. SAS Inst. �24 Table 1. Pregnancy rates for age classes of female elk collected on the Forbes Trinchera Ranch in December 1986-89. 1986 n Preg Age" (yrs) 1987 n Preg Year 1988 n Preg 1989 n Preg 1986-88 Preg %Preg n 1 2 3-4 5-6 7-8 9-10 11-12 13-19 6 7 6 3 4 1 2 5 2 7 6 3 4 1 2 3 7 5 15 10 5 3 1 3 0 5 15 9 4 3 0 2 7 6 19 12 8 2 0 7 2 4 17 11 7 2 0 4 5 3 12 12 5 5 3 4 1 2 10 11 5 5 2 2 25 21 52 37 22 11 6 19 5 18 48 34 20 11 4 11 20 86 92 92 91 100 67 58 Totals 34 28 49 38 61 47 49 38 193 151 78 aAge for ~2 from replacement and wear, for 3+ from dental cementum. Table 2. Fetal sex ratios for age classes of female elk collected on the Forbes Trinchera Ranch in Decembe 1986-89. Age" (yrs) M 1 2 3-4 5-6 7-8 9-10 11-12 13-19 1 1 4 3 3 0 2 3 Totals 17 Total Fetuses 1986b F U M 1987 F U M 1988 F U M 1989 F U % 1986-89 F M U Male 0 2 1 0 0 2 1 0 1 4 1 0 1 0 0 0 0 3 3 2 2 1 0 1 0 2 11 7 2 1 0 1 0 0 1 0 0 1 0 0 1 2 8 8 5 1 0 1 0 2 5 3 2 1 0 2 1 0 2 0 0 0 0 0 1 0 4 4 0 0 1 2 0 2 6 7 4 4 1 0 0 0 0 1 1 1 0 0 3 0 6 8 19 23 17 17 10 8 2 8 3 2 7 3 2 4 4 1 2 2 0 0 100 43 45 50 56 20 60 70 6 7 12 24 2 26 15 3 12 24 3 67 69 15 49 30 38 44 39 151 aAge for ~2 from replacement and wear, for 3+ from dental cementum. bFetal sex M=ma1e F=fema1e U=unknoWll. �25 Table 3. 1989. Year/ Statistic Measurements of elk fetuses from Forbes Trinchera Body Weight {g} Female Male Crown-Rump Length {rrun} Male Female Ranch, Hind Foot Length {rrun} Male Female 29 Nov-5 Dec 1986 Mean SD min max n 27.6 13.8 12.2 58.5 17.0 20.9 7.0 11.9 29.0 6.0 90.2 16.7 63.5 115.5 17.0 86.0 9.5 72.0 94.0 5.0 20.5 4.8 13.5 28.5 17.0 19.8 2.7 l6.5 23.0 5.0 60.4 33.2 12.0 135.0 24.0 141.4 21.1 113.0 180.0 12.0 115.4 23.3 66.0 151.0 24.0 42.3 10.9 31.0 68.0 12.0 30.3 8.8 14.0 46.0 24.0 66.7 29.7 23.0 120.0 15.0 120.0 19.9 77 .0 163.0 26.0 121. 7 17.9 87.0 150.0 15.0 32.6 7.7 17.0 47.0 26.0 33.1 7.0 20.0 45.0 15.0 60.0 45.4 4.9 149.0 24.0 102.4 37.5 35.0 169.0 12.0 111. 5 33.4 38.0 163.0 24.0 27.9 12.3 13.0 53.0 11.0 29.9 10.8 10.0 48.0 23.0 12-14 & 19-21 Dec 1987 Mean SD min max n 118.7 54.7 66.0 242.0 12.0 10-12 & 17-19 Dec 1988 Mean SD min max n 74.5 35.2 22.0 162.0 26.0 9-11 & 16-18 Dec 1989 Mean SD min max n 57.8 59.9 3.5 188.0 12.0 1986- �26 5 .•..... PLASMA24C SERUM 24C PLASMA4C -----6---. SERUM 4C • - 4.5 ....................................... ---8-· . __ E ....._ C) c: w 4 Z o a: w Q .... ..•.........• ...............• . •...........:::.. "'"'.:.~.~.. ...••., 3.5 l- ...... ~...... '."<: ............ & ::;''_ :-- - - -£3" ••• t/) W .... ~ o a: ...•...........................................•. ........... _.:0-. 3 e, ......... _ _ _ _:_ •._--~ __.,_---_ ~_-------- ----8- - 2.5 _ -2 1.5 o 3 6 12 24 HOURS POST-COLLECTION Fig. 1. Progesterone concentrations from 0 to 24 hours post-collection. of values from 3 captive elk. in plasma and serum at 24 C and 4 C Each data point represents the average �27 50 I en Z 40 YEARLY MEDIAN DATES ~ I- [I -e. 0 1986 - n = 26 l- w (.) 30 - ~ 1987 ---_ ..__ .' n = 37 Z 0 (.) I- 20 EJ - 1988 --_ ... _------- n = 44 Z W (.) a: w e. I 1989 10 ----- n = 35 f-- r j]i 0 I I 9/8 t I> I> 9/18 9/28 I 10/8 ~ ro v "'l )<1 I 10/18 10/28 I I 11/7 Xi 11/17 MONTH AND DAY (Begin S-day intervals) Fig. 2. Conception dates for elk on the Forbes Trinchera Ranch, 1986-89. �28 APPENDIX PREGNANCY TESTING David J. Freddy, 80459 ELK USING PROGESTERONE Colorado Division A ASSAYS AND REAL-TIME of Wildlife, ULTRASOUND P.O. Box 252, Kremmling, CO Michael W. Miller, Colorado Division of Wildlife, Prospect Street, Fort Collins, CO 80526 Research Center, 317 W. Gary C. White, Department of Fishery and Wildlife University, Fort Collins, CO 80523 Biology, Colorado LaRue Johnson, College Veterinary Collins, CO 80526 Medicine, Colorado State State University, Fort Abstract: We evaluated blood progesterone radioimmunoassays (RIA), a blood progesterone ELISA kit, and real-time rectal ultrasound for assessing pregnancy status of adult (~l yr old) elk (Cervus elaphus nelsoni). Progesterone was a significant (f < 0.001) predictor of pregnancy status and concentrations in December and January >1.8 ng/ml indicated elk were pregnant with 95% probability. Accuracy of the ELISA kit varied with progesterone levels but was 88% accurate when progesterone was >2.0 ng/ml. Real-time rectal ultrasound was accurate in assessing pregnancy status and should also prove beneficial in assessing fetal growth of elk. Pregnancy status of elk can be reliably determined using progesterone assays or real-time rectal ultrasound. ~ WILDL. MANAGE. 00(0):000-000 Key words: Cervus elaphus nelsoni, progesterone, ultrasound Colorado, elk, pregnancy testing, �29 ABSTRACT OF THESIS RESPONSES OF BULL ELK TO SIMULATED ELK VOCALIZATIONS DURING RUT Responses of mature bull elk (Cervus elaphus nelsoni) to repeated hunter bugling were studied on the Forbes-Trinchera Ranch in south-central Colorado during September and October of 1988 and 1989. Individual bulls, identified by radio collars, were repeatedly approached and bugled. Treatment bulls were harassed with the sound of a shot after bugling sessions; control bulls were not harassed. Ability of bulls to learn to avoid hunter vocalizations after repeated harassment was tested using a logistic regression model. Inclusion of previous harassment (f = 0.003, 1988; f - 0.0001, 1989) in a model that fit the observed data suggests that previous harassment had a significant effect on bull movement toward observer in subsequent sampling sessions. A period of peak response to bugling was isolated based on bull movement toward observer and presence of bulging responses during sampling sessions. This period of peak vulnerability fell between 17 September and 2 October both years. In 1988, mean distance moved by harassed bulls within 24 hrs of harassment was found to be no different (f - 0.481) than distances moved by control bulls. Setting hunting seasons before and after a pre-determined window of peak response to bugling may be advisable in areas where younger bulls, who may not have learned to avoid a bugling hunter, are the primary breeders. Noreen E. Walsh Fishery and Wildlife Biology Colorado State University Ft. Collins, CO 80523 ��Colorado Division Wildlife Research July 1990 of Wildlife Report JOB PROGRESS REPORT State of Colorado Project No. W-1S3-R-3 Mammals Research Work Plan No. 3 Elk Investigations Job No. 7 Elk Census Methodology Period Covered: Author: July 1, 1988 - June 30, 1990 D. J. Freddy Personnel: G. White, G. Byrne, and D. Masden, Colorado and E. Ryland, Forbes Trinchera Ranch. Division of Wildlife, ABSTRACT Estimates of population sizes for elk and deer in the Troublesome sub-unit of Middle Park were obtained using systematically spaced aerial line transects. Additionally, an estimate of population size for' deer was obtained using a random quadrat (2.59 km2) sampling system. All flights were conducted using a Bell-Salay helicopter. The exponential polynomial (EXPL) , negative exponential (NEXP) , and the exponential power series (EXPS) sighting models fit the observed data for elk with about equal quality (f ~ 0.7310). Estimates of popUlation size for elk ranged from 3,659-4,263 with the EXPL model providing the best precision of ±43% (95% CI) of the mean. Estimates of popUlation size for deer ranged from 5,021-9,541 with an EXPL sighting model providing the best precision of ±13% of the mean. Estimated popUlation size for deer based upon quadrats was 1,848 and precision was ±32% (95% CI) of the mean. Line transects and quadrats provided ditferent estimates of population size for deer (f < 0.001)., Sighting distributions for elk and deer were spiked at the transect' line which suggested fundamental problems in either detecting'groups of animals or in assigning groups to distance Lntervals. Further evaluation of line transects and quadrats is warranted if large numbers of radio-collared elk and/or deer are used as a "known marked" popu lat.Lon to validate es t Lmace s provided by different techniques. ��33 ELK CENSUS METHODOLOGY David J. Freddy P. N. OBJECTIVE Evaluate methods to estimate numbers of elk during winter. SEGMENT OBJECTIVES 1. Determine efficacy of line transect methodology to estimate and mule deer on winter ranges shared by both species. numbers 2. Compare estimates of mule dEer density based on line transects quadrats in an area where both methods can be employed. of elk and ACKNOWLEDGMENTS Partial funding Foundation. for helicopter services provided by the Rocky Mountain Elk STUDY AREA The area selected for evaluating aerial line transects (Burnham et al. 1980, White et al. 1989) was the Troublesome portion of Game Management Unit (GMU) 18 within Middle Park, northcentral Colorado, between Highway 125 and Troublesome Creek and north of the Colorado River. Significant numbers of elk and mule deer reside in this relatively small area which can be efficiently sampled with aerial surveys. Primary vegetation types include sagebrush (Artemisia tridentata), aspen (Populus tremuloides), conifer forests (Pinus contorta, Pseudotsuga menziesii, Picea engelmanni), and mixed stands of aspenconifer. Terrain varies from rolling hills to steep canyons. METHODS AND MATERIALS Two sets of parallel transects, each having 26 transects systematically spaced at 1,000 m intervals, were flown to estimate numbers of elk and deer. These 2 sets resulted in transects occurring every 500 m across the sampled area. Transects were oriented on true north-south bearings perpendicular to changes in elevation, were delineated on 1:24,000 scale topographic maps, and were not marked with flight markers on the ground. Placement of transects allowed for counting elk and deer simultaneously. Line transects flown to estimate densities of elk totaled 948 km (588 mi) in length, were 1-15 km long, and were distributed within an area of 246 km2 (95 mi2) between elevations of 2,257-2,990 m (7,400-9,800 ft). Transects were limited to those areas where elk generally reside during January and February (pers. comm. R. Thompson, R. Firth, CDOW). The same line transects were flown to estimate densities of deer but only deer found at elevations S 2,590 m (8,500 ft, primarily sagebrush habitat) were �34 counted. With this restriction, line transects for deer totaled 602 km (374 mi) in length, were 1-14 km long, and were distributed within an area of 166 km2 (64 mi2) which was also the area delineated for a quadrat (1 mi2) sampling system used to estimate numbers of deer in the Troublesome area during the last 22 yrs (Gill 1969). One set of transects was flown each morning and afternoon on 9 and 11 January (4 replicate flights) using a Bell-Soloy helicopter flown at 65-80 kmph (40-50 mph) and 35-50 m above the ground or tree canopy. A navigator and observer were responsible for observing elk or deer. The observer, seated on the right, estimated perpendicular distances to and counted those groups of elk or deer located from the transect center line to the right. The navigator, seated in the middle, maintained course bearing, and located groups on or near the transect line. Observers were well experienced in aerial counts of elk and deer and practiced estimating distance intervals from the helicopter using markers placed at 10 m intervals from a practice transect line. We considered sets of transects to be independent and thus pooled results from all flights to derive estimates of elk and deer populations. We had no indication that counting elk or deer disturbed either sufficiently to cause them to move from 1 transect to another while flying one set of transects and thus, we feel duplicated counts within a set of transects did not occur. Estimates of population size derived from line transects followed methods outlined by White et al. (1989) and estimates derived from quadrats followed methods of Gill (1969) and Mendenhall et al. (1971). We wanted to compare estimates of population size for deer derived from line transects and quadrats. We, therefore, expanded an existing quadrat system from 10 to 30 quadrats by drawing at random an additional 20 quadrats to improve precision of the estimate (Freddy 1989). Quadrats were flown on 10 January using the same Bell-Soloy helicopter and pilot and flown at similar speeds and elevations above the ground as during line transects. At least 1 corner of each quadrat was marked on the ground to aid in locating quadrats from the helicopter. The navigator and observer both searched for and counted deer. RESULTS AND DISCUSSION Although we searched for deer and elk simultaneously on line transects, the 2 species were seldom seen in close proximity. We, therefore, believe that searching for both species had little effect on detecting either species. We found little evidence to suggest elk were moving prior to detection and moved little after detection. Elk were often bedded in aspen or conifer habitats. However, deer moved in response to the helicopter more than did elk. Although snow cover was nearly 100%, snow depth was only 4 cm at lower elevations and this did not deter deer from moving. Elk were often found in snow depths of 30-60 cm. Only 2 groups of deer were seen along line transects outside the sample area demarcated for quadrats. Line Transects-Elk Estimated population size for elk was 3,659-4,263 depending upon which sighting probability model was used (Table 1). The EXPL models provided the best precision which was ±43% of the mean (95% CI, Table 1). A minimum number of elk within the sampled area was 419 based upon elk classified to sex and �35 age during flights on 15 January 1990. A previous estimate for the same area based on one set of line transects flown in January, 1988, (Freddy 1988) was 1,276 with a known minimum of 744. The EXPL, NEXP, and EXPS probability models having center intervals of 0-15 m and strip widths of 95 or 155 m fit the observed data with about equal quality (f ~ 0.7310, Table 1). Enlarging or reducing the width of the center interval or the cut point for strip width generally resulted in models that were poorer in fitting the data (f < 0.100, see White et al. 1989). The unexpectedly high estimates of total elk resulted from high numbers of groups observed within the center line interval compared to numbers of groups seen within intervals away from the center line (Fig. 1). Probability of sighting groups decayed rapidly away from the center line which is cause for concern and suggests fundamental problems in detecting groups or assigning groups to proper distance intervals. Average group size was 11.59 ± 3.5 (95% CI) elk. Line Transects-Deer Estimated population size for deer was 5,021-9,541 (Table 2). The EXPL, NEXP, and EXPS sighting models with a cut-point of 95 m fit the data best (0.0908 ~ f ~ 0.2024, Table 2). Enlarging or reducing the width of the center interval or strip width produced poorer fitting models (f < 0.001). As with elk, high numbers of groups were observed within the center line interval compared to intervals away from the center line and probability of sighting groups of deer decayed rapidly away from the center line (Fig. 2). Rapid decline in groups of deer detected away from the center line was also found on line transects flown within the Piceance Basin of Colorado (White et al. 1989), on Forbes Trinchera Ranch in southcentral Colorado (January, 1990), and on Cedar Ridge (February, 1989), also in Middle Park (Fig. 2). Habitats were primarily juniper-pinyon woodland (Juniperus osteosperma-Pinus edulis) in the Piceance Basin, juniper-pinyon woodland and sagebrush parks on Forbes Trinchera Ranch, and predominately sagebrush with scattered clumps of juniper on Cedar Ridge. The EXPL sighting model generally fit these sets of data the best (see White et a1. 1989, for Forbes Trinchera f ~ 0.2579, for Cedar Ridge f ~ 0.7639). The data from Piceance Basin and Forbes were obtained by 2 different sets of observers and a third set of observers obtained the Troublesome and Cedar Ridge data but all data were collected using a Be11Soloy helicopter. The rapid decline in detecting groups away from the center line in all of these efforts suggests that this problem is not necessarily dependent upon observers or habitat but could be related to the type of helicopter used. Size of groups detected on line transects appears to increase in more open habitats. In the Troublesome, where deer were primarily in sagebrush, group size was 6.1 ± 0.64 (95% CI) deer. At Forbes Trinchera in juniper-pinyon and sagebrush habitats, group size was 4.2 ± 0.55 (95% CI); while on Cedar Ridge in sagebrush-juniper habitats, group size was 3.2 ± 0.81 (95% CI) deer. In the continuous juniper-pinyon habitat of the Piceance Basin, average group size ranged from 1.45-2.54 deer (White et a1. 1989). Quadrats-Deer Estimated population size for deer was 1,848 (n - 30 quadrats, Table 3). This estimate represented only 25% of the deer estimated by the best fitting EXPL model for line transects (Table 2). We expected quadrats to underestimate true population size because area-based sampling strategies assume 100% of the �36 animals present are counted and this seldom occurs. In juniper-pinyon, observers detected only 65% of the known deer on quadrats (Bartmann et al. 1986). In the open sagebrush vegetation, we might detect 80% of the deer under optimal conditions (LeResche and Rausch 1974); and if true, the true population of deer based on quadrat data would approach 2,310 deer. Given a worst case scenario of detecting only 65% of the deer on quadrats, we would expect a true population of 2,843 deer. Line transects estimated nearly 2.5x this number of deer and have been found to estimate 90% of the true population (White et al. 1989). I cannot explain the large discrepancy in estimates of population size between quadrats and line transects. Estimated population sizes based upon the original 10 or the additional 20 quadrats were not different (f < 0.01, t-test, Table 3). Frequencies of deer/quadrat were also not different between 10 and 20 quadrats (Chi-Square contingency - 0.34, P > 0.50, Table 4). Precision improved with increasing sample size in part because a finite population correction factor could be used, unlike line transects (pers., comm. G. White, Table 3). Quadrats, therefore, provided a consistent estimate of mean density (consistent bias) across different levels of sampling. Line Transects vs Quadrats-Deer Line transects provided a higher estimate of total deer than did quadrats (f < 0.001, t-test, Tables 2, 3) and provided better precision (± 13% vs ± 32%, 95% CI, Tables 2, 3). However, changing numbers of quadrats (sample units) had virtually no effect on estimates of population size, but changing strip width, and, therefore, the numbers of groups (sampled distances) used in calculations, or sighting models altered population estimates ~1.5x (Tables 2, 3). Massaging line transect data by altering cut points, center interval distances, or sighting models can produce different estimates of population size and is a hazard of the technique (White et al. 1989). Calculating population estimates based on quadrats appears more straight forward and less subject to alteration. Flying Effort Actual flying time used to count elk and deer along line transects was consistent among replicated flights (Table 5). However, flights conducted during morning hours always detected more groups of both elk and deer. Although replicated flights of transect sets 1 and 2 were to be alternately flown during am and pm periods, this did not occur. Set 1 was always flown during am, and set 2 during pm hours. Differences in numbers of groups detected could, therefore, be due to differences in sets of transects and not time of day. We feel time of day was a factor in detecting fewer groups during the pm flights, probably due to a decrease in animal activity or observer fatigue. We replicated flights of line transects to obtain anticipated minimums of 130 groups of elk (Freddy 1988) and 200 groups of deer (White et al. 1989) to achieve precision approaching ± 15%. We did not obtain the expected groups of elk but did achieve our expected precision for deer. Assuming line transects for deer accounted for 64% of the total time spent on transects, line transects for deer took 579 mins to complete compared to 445 mins to complete 30 quadrats (Table 5). �37 CONCLUSIONS Sighting distributions for elk and deer were spiked at the transect line as observed by White et al. (1989). This characteristic of line trans~ct data collected from a helicopter appears to be independent of observer or habitat type, at least for deer. For Middle Park, line transects produced surprisingly high estimates of population sizes for elk and deer. A fundamental question is why sighting distributions are so spiked at the transect line. Trying to answer this question will be difficult. Mounting a video camera inside of the helicopter (if possible) to record groups relatively close to the transect line may provide insight into whether the "heaping" of groups at the center line is due to incorrectly assigning groups to this interval or whether groups to the right of the center interval are not detected. A second question involves the accuracy of population estimates for both elk and deer. In free-ranging populations, the only possible means to estimate true density is to radio-collar up to 100 animals of each species and use these radio-collared animals as a "known marked" population to create a detection function when flying transects or quadrats. To evaluate the accuracy of either transects or quadrats for estimating populations of either elk or deer, this expensive radio-collaring alternative should be strongly considered. LITERATURE CITED Bartmann, R. M., L. H. Carpenter, R. A. Garrott, and D. C. Bowden. 1986. Accuracy of helicopter counts of mule deer in pinyon-juniper woodland. Wildl. Soc. Bull. 14:356-363. Burnham, K. P., D. R. Anderson, from line transect sampling 202pp. Freddy, D. J. 1988. July (1):78-82. and J. L. Laake. 1980. Estimation of density of biological populations. Wildl. Mono. 72. Elk census methodology. 1989. Elk census methodology. Rep. July (1):61-65. Colo. Div. Wildl. Game Res. Rep. Colo. Div. Wild1. Game Res. Gill, R. B. 1969. A quadrat count system for estimating game population. Colo. Game, Fish, & Parks Game Info. Leaflet 76. 2pp. LeResche, R. E., and R. A. Rausch. 1974. Accuracy moose censusing. J. Wildl. Manage. 38:175-182. and precision Mendenhall, W., L. Ott, and R. L. Scheaffer. 1971. Elementary sampling. Duxbury Press, Belmont, Calif. 247pp. of aerial survey White, G. C., R. M. Bartmann, L. H. Carpenter, and R. A. Garrott. 1989. Evaluation of aerial line transects for estimating mule deer densities. J. Wild1. Manage. 53:625-635. Prepared by ~ 0;;. Wildlife •FrdY Researcher �38 Table 1. Population estimates for elk based upon different sighting probability models for line transects, Troublesome sub-unit, Middle Park, January 1990. Modela Center Interval/ Cut Point (m) Groups n Chi-square Prob. (f) POIl· Size & 95% CI N Upper Lower Density Elk/mil EXPLb 0-15 / 155 66 0.8252 3659 5232 2086 38.5 EXPLb 0-15 / 95 57 0.7310 3664 5224 2104 38.6 NEXP 0-15 / 95 57 0.8248 3662 5380 1943 38.6 EXPS 0-15 / 95 57 0.7514 4263 9813 0 44.9 aModels are: EXPL=exponential polynomial, NEXP-negative exponential, EXPS-exponential power series. bThese models provided the best precision: ±43% of the population mean. Table 2. Population estimates for deer based upon different probability models for line transects, Troublesome sub-unit, January 1990. Modela Center Interval/ Cut Point (m) Groups n Chi-square Prob. (f) sighting Middle Park, PO!;!.Size ~ 95% CI Lower N Upper Density Deer/mil EXPL 0-15 / 155 185 0.0001 5021 7144 2898 78.4 EXPLb 0-15 / 95 139 0.0908 7496 8466 6526 116.9 NEXP 0-15 / 95 139 0.1420 7494 10,538 4451 117.0 EXPS 0-15 / 95 139 0.2024 9541 17,327 1746 149.0 aModels are: EXPL - exponential polynomial, NEXP - negative exponential, EXPS - exponential power series. bThis model provided the best precision: ± 13% of population mean. Table 3. Estimated size of the mule deer population in the Troublesome subunit of Middle Park, January, 1990 based upon deer counted on square-mile quadrats. Variances calculated using finite population correction factor. Quadrat System Original Additional All (N - 64) Quadrats n Mean 95% Lower 10 20 30 1891 1827 1848 506 1028 1264 C.1. Upper 3276 2625 2431 Precision % of Mean Density Deer/mil ± 73% ± 44% ± 32% 29.5 28.5 28.8 �39 Table 4. Frequencies of numbers of deer counted per quadrat for varying numbers of quadrats flown, Troublesome sub-unit, Middle Park, January, 1990. . Deer per Ouadrat Quadrats Flown Original 10 Additional 20 All 30 0-9 10-29 ~30 4 7 2 6 4 7 11 8 11 Table 5. Numbers of groups of elk and deer detected on transects and estimates of counting times for transects and quadrats flown in the Troublesome sub-unit, Middle Park, January, 1990. Flight Method Trans. Trans. Trans. Trans. Set Set Set Set 1 2 1 2 Right Rep Rep Rep Rep 1 AM 1 PM 2 AM 2 PM Trans. Totals Quadrats (30) for Deer Only Groups De t ec t.ed" Elk Deer Left Tot. Right 24 17 22 16 4 2 1 4 79 11 28 Courrt Lng'' Time (mins) 23 20 60 44 60 44 229 227 229 220 90 208 905 17 445 8Groups of elk were assigned as to either the left or right of center line. Groups to the left were generally close to the center line. Our searching intensity to the left was limited. Only groups to the right were used in calculations. We searched only to the right for deer. bTime shown is the best estimate of flying time used to count animals. Flying time to/from for refueling not included. �40 20 ~----------------------------------~ ELK TROUBLESOME en Q. ~ 15 o a: JAN 1990 o LL o n = 66 GROUPS 10 a: w r:a :E 5 ~ Z o ~~~~~~~,~~~~~~~~~~~~ LON o t.n t.n t.n M t.n t.n ,.... N lI::t LOCD r--CO t.n t.n t.n LO LO t.n t.n t.n CD r-- CO t.n LO M t.n t.n lI::t LO t.n C)~ en DISTANCE FROM CENTER LINE (m) Fig. 1. Numbers of groups of elk observed at different distance intervals ,from the center line of line transects, Troublesome sub-unit, Middle Park, January 1990. �41 70 60 50 40 30 20 10 t/) a. MULE DEER JAN 1990 . n= 185 groups .. 0 40 MULE DEER ~ 0 30 C!J u, 0 20 w m 10 ~ 0 a: TROUBLESOME 1 .....• -.- -- .. --- - __ n J.::::.. FORBES TRINCHERA cA ..N 1~~.Q _ _._._ = 106 GROUPS a: :i Z 14 MULE DEER 12 1·_··_·_-_··············· 10 __ ····_···················_··... . n 8 CEDAR RIDGE FEB 1989 - - ..- .. - - -.- ..-............. . -. _ . = 56 GROUPS 8 4 2 0 ,... It) o ,... It) It) It) N M It) It) It) It) CD ••••• It) en ,... It) It) co en DISTANCE FROM CENTER LINE (m) Fig. 2. Numbers of groups of mule deer observed at different distance intervals from the center line of line transects: TOP, Troublesome sub-unit, Middle Park, January 1990; MIDDLE, Forbes Trinchera Ranch, southcentral Colorado, January 1990; BOTTOM, Cedar Ridge, Middle Park, February 1989. �0-15 15 - 25 en ~ 3 95 -15 m 85 - 95 -- z en "lJ o c 0 co ::0 co ~ G') 55 - 65 II ::J ...•. (II » C 75 - 85 ::0 ...•. 0 en c.. en (II 45 - 55 35 -45 m 65 -75 z -i m 0 s:: 0 ,..::0 m 0 25 - 35 z C 0 NUMBER OF GROUPS m s:: 0 en m c m r 0 -i "::0 r m 0 N -1-' tv �43 70 60 50 40 30 20 10 0 en MULE DEER TROUBLESOME JAN 1990 .. n - 185 groups 40 MULE DEER FORBES TRINCHERA JAN 1990 D. ~ 0 a: 30 LL 20 a: w 10 e 0 m n = 106 GROUPS :E ~ Z 0 14 MULE DEER CEDAR RIDGE 12 ._._._ 10 _._ _ _ _ _ __ .--- n 8 FEB 1989 _.__ _-_ _ = 56 GROUPS _.__ .•._._ .._ _--_ .._._ .....• _ ----_ _ .._._ 6 - .. _ .._ .. _ __ . _ . . 4 2 0 It) ,.... It) 0') I o It) ,.... It) N It) M It) -.::t It) It) It) CD It) •••••• DISTANCE FROM CENTER LINE (m) ,.... ��Colorado Division Wildlife Research July 1990 of Wildlife Report JOB PROGRESS REPORT State of Colorado Project No. ~W_-~1~5~3_-~R~-~3~ Work Plan Mammals Research lA Multispecies Job No. 1 Animal and Pen Support Facilities for Mammals Research Period Covered: July 1, 1989 - June 30, 1990 Author: No. _ Investigations M. W. Miller Personnel: D. J. T. M. L. Baker, M. A. Wild, P. H. Neil, M. J. McArtor, R. B. Gill, M. R. Szymczak, D. A. Magnuson, R. Ritchie, C. Y. Irvin, S. M. Willis, B. J. Maynard, L. Stevens, N. T. Hobbs, and H. J. Lucking K. Ringelman, ABSTRACT The Colorado Division of Wildlife's Foothills Wildlife Research Facility (FWRF) maintained captive animals (up to 73 wild'and domestic ungulates of 6 species and 95 migratory and upland game birds of 7 species) and facilities supporting 9 different mammalian and avian research projects. Current facility management practices were assessed, and a plan was developed to outline needs and strategies for improving FWRF management. Over the next 5 years, objectives of FWRF operations are twofold: (1) to provide a~d maintain captive wildlifp. populations and facilities supporting CDOW's Terrestrial Wildlife Research Program, as well as programs of CDOW cooperators, and (2) to develop improved animal and facility management practices that will provide maximum research opportunities at minimum cost. Management strategies for meeting these objectives through three broad operational activities (animal maintenance, facility maintenance, and facility improvement/expansion) were detailed in the FWRF Operations Plan. ��I -s 4/ ANIMAL AND SUPPORT FACILITIES FOR MAMMALS RESEARCH Michael W. Miller P. N. OBJECTIVE To provide and maintain populations of captive animals and pen facilities support Mammals and Avian Research Programs. to SEGMENT OBJECTIVES 1. Maintain and improve animal research and holding 2. Coordinate 3. Maintain up to 15 elk, 30 mountain sheep, 30 pronghorn antelope, deer, and 6 domestic cows in suitable health to perform required experiments. 4. Conduct management experiments efficiency and efficacy. all rearing, training, maintenance, facilities. and research activities. 15 mule research to increase feeding and maintenance METHODS AND MATERIALS Routine animal care and facility maintenance activities supporting new and ongoing Terrestrial Wildlife Research Program projects were accomplished as previously described. In addition to these activities, current facility management practices were assessed and a plan was developed to outline needs and strategies for improving FWRF management. RESULTS AND DISCUSSION Based on assessment of current management practices and anticipated needs of the Terrestrial Wildlife Research Program over the next 5 years, the following Operations Plan for CDOW's Foothills Wildlife Research Facility was developed: Prepared hi bYLt~ Michael W. Miller Wildlife Researcher ��DRAFT PROGRAM NARRATIVE State of Colorado Project No. W-153-R-3 Work Plan No. __~l~A Job No. A. I Mammals Research _ Multispecies Investigations Animal and Pen Support Facilities for Terrestrial Wildlife Research NEED Reliable knowledge offers an essential foundation for policy decisions. Faced with needs for more sophisticated management techniques and growing public scrutiny of management practices, wildlife professionals have recognized the value of reliable data to their management decisions. Experimental studies of captive wild animals are an important source of scientific information used by wildlife biologists in developing resource management strategies. Because wildlife management issues are becoming increasingly complex, more, not less, science will be required to resolve those issues in the future. Wild animals tamed for research purposes afford wildlife scientists unique opportunities for studying processes that affect free-ranging individuals and populations. Many experiments conducted using captive wildlife couJ.d not be performed in the field. Moreover, the ability to control experimental conditions for captive wild animals vastly improves the reliability of data gathered from such studies. Over the past 15 years, the Colorado Division of Wildlife's Foothills Wildlife Research Facility (FWRF) has been central to captive wildlife research efforts in Colorado. Results of "laboratory" experiments conducted at FWRF have enhanced knowledge of ecology, physiology, and management of native wild ungulates and, more recently, migratory and resident game birds. There is little doubt that FWRF has contributed substantially to the overall productivity and success of CDOW's Terrestrial Wildlife Research Program. Developing and improving husbandry practices for captive wildlife and facilities for captive wildlife research, as well as ongoing needs for centralizing and coordinating research activities, have guided operation of FWRF over the last decade. In the next 5 years, emphasis needs to be shifted toward improving economy and utility of FWRF operations to fully exploit opportunities for gaining knowledge about wildlife. Maximizing efficacy and efficiency of established husbandry practices and enhancing facilities to reflect a growing diversity of wildlife research pursuits will be necessary for FWRF to serve the anticipated needs of CDOW's Terrestrial Wildlife Research Program. Here, planned strategies for improving economy and utility of FWRF operations during FY90-9l to FY94-95 are outlined. B. OBJECTIVES �50 DRAFT - To provide and maintain captive wildlife populations and facilities supporting CDOW's Terrestrial Wildlife Research Program, as well as programs of CDOW cooperators. - To develop improved animal and facility management practices that will provide maximum research opportunities at minimum cost. Specific objectives include the following: To reduce overall feeding costs by 10% for FY90-91 (based on FY89-90 costs); to establish a refined standard feeding program by FY92-93. To establish a conservation-oriented approach for providing utilities/services to FWRF; to lower utilities/services costs by 10% in FY91-92. To establish a standard program for evaluating health wildlife; to reduce overall animal health maintenance 10% in FY92-93. of captive costs for FWRF by To establish standard procedures for preventive maintenance and repairs at FWRF; to reduce overall facility repair costs by 10% in FY93-94; to reduce overall costs of routine/ preventive maintenance by 10% in FY94- 95. To enhance facilities research needs. C. EXPECTED RESULTS to serve a growing diversity of anticipated AND BENEFITS Availability of captive wildlife and support facilities enhances the productivity of CDOW's Terrestrial Wildlife Research Program by providing opportunities for scientific study of wild animals under controlled conditions. More efficacious and efficient husbandry practices will improve overall health of research subjects and add to existing knowledge about biology and ecology of those species. Enhanced facilities will broaden the diversity of wildlife research problems that can be addressed at FWRF. More economical and utilitarian operation of FWRF will free fiscal resources for use in additional research experiments, and will lower costjbenefit ratios for captive wildlife research. Success in bighorn sheep management is limited by effects of disease on longterm population performance; experiments at FWRF will be directed toward developing management options for bighorn disease problems. Efficacy and effect of various harvest strategies on elk population management is a pivotal issue; experiments at FWRF will focus on developing field techniques to evaluate nutrition, reproduction, and survival in elk populations subjected to alternative harvest regimes. Livestock and agricultural conflicts influence pronghorn management decisions; FWRF will provide pronghorn for pasture experiments examining livestock competition, and experiments at FWRF will explore approaches for improving pronghorn habitat evaluation techniques. Avian cholera causes tremendous annual losses in migratory waterfowl; experiments at FWRF will investigate relationships between environmental toxins and susceptibility to cholera in captive mallards. Use of pesticides �51 DRAFT to control crop damage presents a potential threat to upland game, birds and other species that reside in agricultural areas; experiments at rw~.F will augment field studies measuring the effects of common pesticides on various avian species inhabiting Colorado's farmlands. Through these and other projects, research activities at FWRF will provide scientific information to aid wildlife professionals in addressing management issues in Colorado and elsewhere. D. APPROACH Management of FWRF encompasses three broad operational activities: ANIMAL MAINTENANCE, FACILITY MAINTENANCE, and FACILITY IMPROVEMENT/EXPANSION. Each of these operational categories can be further subdivided into major programs, projects, activities, or emphases (eg. FWRF ANIMAL MAINTENANCE includes both Nutrition and Health Maintenance Programs). This organizational structure forms a foundation for the following overview of strategies for managing FWRF operations during FY90-9l to FY94-95. Annual FWRF Work Plans will be completed by 1 April of the fiscal year prior to enaction (except in FY90-9l, when the annual plan will be completed by 31 July 1990). Study plans, project descriptions, and annual Work Plans will be appended to appropriate subsections as they become available. Ongoing assessment of FWRF operations will be included in annual Job Progress Reports. 1. ANIMAL MAINTENANCE A variety of mammalian and avian species will be maintained at FWRF for use in research experiments. Feeding and care of these captive animals constitute major expenditures in the FWRF operating budget. At present, over 60 head of hoofed stock (elk, pronghorn, bighorn sheep, mountain goat, and domestic cattle) and nearly 50 ducks (several species) are wholly or partially supported by FWRF. Under present space and labor constraints, numbers of wild and domestic ungulates that can be maintained at FWRF are approaching upper limits. Strategies for anticipating species and numbers of individuals required to support ongoing and upcoming research, for controlling growth of our captive herds, and for placing or otherwise disposing of surplus animals will be developed during FY90-9l to FY94-95. Upcoming research needs will be identified largely through CDOW's species management planning process, from which experiments will arise to address specific issues or problems. In addition, management experiments will be conducted over the next 3 years to develop more cost-effective husbandry practices for maintaining research herds at FWRF. Standard operating procedures for humane care and treatment of research animals will be developed by CDOW's Animal Care Committee in compliance with Federal Animal Welfare regulations. a. Nutri tional Maintenance Under current ad libitum feeding regimes for hoofed stock, all individuals are maintained in good to excellent body condition year round. However, in some cases animals are overconditioned, and these feeding regimes are also wasteful. Strategies will be developed for improving efficiency of feeding programs for hoofed stock. Candidate strategies include: a) use of alternative foodstuffs (eg. grass hay, �52 DRAFT cubed alfalfa, complete pelleted rations) to reduce waste; .b) modified feeders (eg. hay racks, self-feeders) to improve access to hay and minimize trampling losses; c) structural protection of feed from rain and snow to minimize weather-related waste; and d) approaches for dispersing feed to reduce competition and increase feeding opportunities for subdominant animals. Improvements on feeding efficiency will target reductions in both direct feed costs and indirect costs of labor associated with feeding and clean-up. Management: Overall, costs for feeding research animals at FWRF total nearly $23,000 annually under current management practices. Objectives of feeding program improvements are to reduce overall feeding costs by 10% for FY90-9l (based on FY89-90 costs), and to establish a refined standard feeding program by FY92-93. Compared to cost estimates based on feeding to meet calculated energy requirements of species currently maintained at FWRF (Appendix A), feed costs for ad libitum feeding appear 10-lS% higher than necessary to maintain healthy research animals. During FY92-93, standard operating procedures for FWRF's feeding program will be established. In addition to reducing expenditures for animal feed, workstudy and YCC employees will be hired for feeding and clean-up beginning in FY90-9l to reduce human resource expenditures. The feeding programs for captive waterfowl and upland birds will be included in overall FWRF management duties beginning in FY90-9l; those programs may also be modified after assessment. Research: Management experiments will be used during FY90-9l and FY9l92 to evaluate and compare aforementioned strategies for improving feeding efficiency. Study plans for these management experiments will be appended as they become available. Standard operating procedures for FWRF's feeding program will be based on results of management experiments. As an adjunct to management experiments and future FWRF management planning, a computerized system for recording data on feed consumption, tracking feed supplies, and projecting feed needs will be developed during late FY90-9l and early FY9l-92. A description of that system will be appended, along with annual evaluations and updates. b. Health Maintenance Maintaining research animals in optimum health represents another significant activity in FWRF operations. The primary emphasis of FWRF's herd health program is preventive management and medicine. Current strategies for optimizing health in ungulate herds include proper nutritional management (see above), routine vaccinations and health care (eg. hoof trimming, parasite control), regular health monitoring (at minimum, daily inspections and weekly weighing), prompt veterinary examination and work-up followed by treatment and/or isolation for sick individuals, necropsy and diagnostic work-up of dead animals, and modifications of husbandry practices and facilities to minimize or prevent health problems. Major health problems persisting in the face of FWRF's current herd health program include necrotic stomatitis in pronghorn and bighorn sheep, overconditioning in pronghorn and bighorn sheep, chronic wasting disease in elk, pasteurellosis in bighorn sheep, �S3 DRAFT and traumatic injuries facility design. in pronghorn that arise from inadequacies in Management: Over the next 5 years, development of FWRF's herd health program will be directed toward early detection of serious illness in research animals and improvement of animal health record keeping. The objective of these modifications is a 10% reduction in overall health maintenance costs by FY92-93. In FY90-9l, a formal system for daily animal health assessment and reporting will be devised, tested, and implemented. In late FY90-9l and FY9l-92 a computerized database for tracking animal health and weight performance will be developed. Additionally, accounting systems for tracking drug and veterinary supply inventories as well as health maintenance expenses will be established in FY9l-92 and FY92-93. These systems will be used in budgeting for veterinary care, and in evaluating husbandry and preventive medicine strategies to improve herd health management at FWRF. Research: Management strategies for specific health problems affecting captive wildlife at FWRF will be evaluated through planned management experiments. Preferred strategies will then be summarized in problemspecific management plans. A study of necrotic stomatitis treatment and management in pronghorn is planned for FY90-9l. Nutritional management of captive ungulates will be addressed as described earlier. The Chronic Wasting Disease Management Plan developed for FWRF in FY86-87 will be revised in FY9l-92. Ongoing pasteurellosis research in bighorns currently precludes plans for more intensive management of that problem. Experiments using alternative visual barriers in pronghorn pens to reduce injuries caused by fence collisions will be completed in late FY90-91. 2. FACILITIES a. MAINTENANCE Basic Utilities i. and Services During FY90-9l, standard operating procedures designed to provide basic utilities and services to FWRF at minimum cost will be developed and implemented. In particular, improved heating efficiency for buildings and water-conserving approaches for irrigating pastures need to be developed. Quarterly expenditures for utilities and services (eg. electricity, water, propane, trash removal) will also be analyzed. During FY9l-92. SOP's will be modified to further reduce utilities/services expenditures; the objective of this project is to lower utilities/services costs by 10% in FY9l-92 and establish a conservation-oriented approach for providing these to FWRF. b. Routine/Preventive Maintenance i. During early FY90-91, a system for identifying, prioritizing, scheduling, and completing routine/preventive maintenance tasks at FWRF will be developed in conjunction with a similar system for facility repairs; descriptions of these systems will be appended later. These systems will be implemented and evaluated during FY90- �54 DRAFT 91. Based on performance of the trial system, a modified system and standard operating procedure will be developed and implemented during FY9l-92. Analyses of expenditures for maintenance and repairs will be made during FY9l-92 and FY92-93. Those data will be used to identify recurrent repair problems that could be addressed through preventive maintenance. Specific objectives for this project are reduction of overall facility repair costs by 10% in FY93-94 and of routine/preventive maintenance costs by 10% in FY9495. c. Repairs i. 3. During late FY89-90, a system for identifying, prioritizing, scheduling, and completing repairs at FWRF will be developed in conjunction with a similar system for routine/preventive maintenance. (See above for further details.) The management goal of both projects is to minimize investment in unscheduled emergency repairs at FWRF by identifying problem areas and addressing needs through preventive maintenance and/or facility modifications. FACILITIES EXPANSION/ENHANCEMENT Major modifications in existing FWRF facilities (Fig. 1) required to accomplish research objectives and/or improve operational efficiency, effectiveness, and/or safety will be planned and budgeted for as individual projects. In general, individual projects requiring significant expenditures of human (> 40 worker-hours) and/or fiscal (> $200) resources will be planned. During FY90-91, a standard operating procedure for anticipating annual expansion/improvement projects will be implemented. Detailed project plans will be included in subsequent annual work plans for FWRF. These project plans will include: brief description of the need, blueprints, sketches, and diagrams where appropriate, scheduled starting and completion dates, and budget including calculated or estimated costs of - personal services or labor (including contracts), operating supplies and services (including materials, supplies, services), added utility costs (for facility maintenance only), reimbursable travel costs (if any), equipment, and total projected cost. A recent evaluation of FWRF operations and needs for facility improvements identified several areas requiring attention in FY90-91. Priority projects included the following: a. Improving drainage for west-side animal paddocks and alleyway. Planned completion -- 31 July 1990. (Details will be included FY90-9l FWRF Work Plan.) in �DRAFT b. Staining east-side alleyway. Planned completion -- 7 A~gust (Details will be included in FY90-91 FWRF Work Plan.) c. Rebuilding west-side alleyway. Planned completion -- Phase I, peripheral alleyways: 31 October 1990; Phase 2, central alleyway/isolation pen complex: 30 June 1990. (Details will be included in FY90-91 FWRF Work Plan.) d. Expansion of animal paddocks (Fig. 2). Planned completion -- 31 October 1990. (Details will be included in FY90-91 FWRF Work Plan.) e. Design and construction be appended.) of elk handling facilities. 1990. (Details will Other facility modifications under consideration for subsequent fiscal years (listed here in no particular order of priority) include redesign of feeding areas to improve accessibility and clean-up, revegetation of overgrazed pastures with grazing-resistant forages, development of a facility-wide landscape plan, expansion of upland and migratory bird facilities, and establishment of on-site laboratory facilities. Ongoing and upcoming research projects will also dictate needs for modifications and improvements at FWRF. Annual projects covered under FWRF's operating budget will be selected on the basis of human and fiscal resources available, and will be prioritized using the following criteria: human safety> animal welfare/safety> facility-wide benefits > project-specific benefits. Other projects may be accomplished with supplemental funding and labor provided by ongoing or planned research projects. All projects will be approved by the facility manager and the Research Facility and Animal Committee (RFAC) prior to initiation. E. SCHEDULE Fiscal Year 1990-95 Activity 1. Animal MaintenAnce a . ltatritianal Maint.c.an.ce i. .DewPlop a sywt_ for estimating animal food requirements o~ animal and type of feed. by specie. ii. Implement and improve above system. iii. Conduct manag_nt experiments iv. Develop computerized v. Develop and implement vi. Maintain vii. Estimete D. Health system projected for tracking feeding FWRF animals to illlprovefeeding feed use and needs. SOP. in reasonable annual efficiency. physical feed requirements condition. and budget. 90 Oct-Dec 90 .Jan 91-Jun Apr-Sep 92 91 Jul 92-Jun 93 Jul 90-Jun 95 Jan-Mar 91-95 MainteDAnCe i. Devise, impl.~t.. 4t\d teat. daily hf>alth ~uess~t 11. Conduct prepare manag_nt management experiments to improve h •• lth program, plans for specific health problema. iii. Develop iv. Jul-Sep Revise computerized Chronic system Wasting for tracking Disease Management health Plan. 151''11._. data. Jul 90-Mar 91 Sap 90-Jun 95 Apr-Sep 91 Jan-Mar 92 �56 DRAFT v. Develop vi. Use preventive medicine and health monitoring maintain FWRF animals in optimum health. vii. 1990-95 2. Facility Prepare computerized system annual health for drug/supply care cost estimates inventory. programs Apr-Sep 92 to Jul 90-Jun and budget. 95 Ja.'1-Mar91-95 Maintenance a. Basic Utilities and Services i. Develop and implement ii. Revise SOP to minimize iii. Provide minimum b. Routine SOP's to lower utilities utilities expenditures. expenditures. necessary basic utilities/services to FWRF at cost emphasizing resource conservation. or Preventive i. Implement ii. Improve 90 Jul 91-Jun 92 Jul 90-Jun 95 Jul 90-Jun 91 Jul 91-Jun 92 Jul 90-Jun 95 Jul 90-Jun 91 Jul 91-Jun 92 Jul 90-Jun 95 Maintenance and evaluate and modify Oct-Dec formal maintenance system and SOP. system. iii. Perform routine maintenance, emphasizing use of preventive maintenance to reduce unscheduled repairs. c. Repairs i. Implement ii. Improve and evaluate and modify formal repair system (see 2bi). system. iii. Perform routine repairs, emphasizing use of preventive maintenance to reduce unscheduled repairs. 1990-95 3. Facility Expansion/Enhancement a. Project 1. Improving drainage for west-side animal alleyway. (Detailed in FY90-91 FWRF Work Plan.) paddocks b. Project 2. Stur::'~g east-side FWRF Work Plan.) in FY90-91 alleyway. (Detailed and Jul 90 Aug 90 c. Project 3. Rebuilding west-side alleyway. (Detailed in FY90-91 FWRF Work Plan.) Phase 1: Peripheral alleyway replacement. Phase 2: Central alleyway/isolation pen reconstruction. Aug-Oct 90 Nov 90-Jun 91 d. Project 4. Expansion of animal paddocks (Detailed in FY90-91 FWRF Work Plan.) Jul-Mar e. Design and conduct additional and handling expanSion/enhancement facilities. ?rojects. 90 Jul 90-Jun 95 Planned improvements in operating efficiency of FWRF are intended to hold annual requirements for fiscal resources constant in the face of inflation :::..:.::-ing FY90-91 through FY94-9S. In the later portion of this S-year planning period, increasing costs for personal services are anticipate _0 be offset by lowered costs for animal maintenance and facility repairs as described above. �.) I DRAFT Budget Category {OIl Personal Services Permanent employee (WRT III) Temporary employees (vacant) Workstudy/YCC employees (vacant) FTE Requirements PFTE TFTE 1.0 0.5 Operating Supplies/Services Animal Maintenance Feed Veterinary supplies and care Facilities Maintenance Routine preventive maintenance Facility expansion/enhancement 7,500 3,500 ~ Total (21o) .Costs 61,000 1.5 $ 10,000 2,000 and repairs 3,000 5,000 20,000 (21u) $ Utilities 3,500 3,500 Travel (28) $ Expenses 500 500 (31) Capital $ Expenditures 2,500 2,500 TOTAL ESTIMATED AVERAGE ANNUAL BUDGET $ 87,500 G. LOCATION The CDOW FWRF is that part of the Colorado State University Foothills Campus known as the Wildlife Research Experimental Area, originally consisting of 44 acres located as follows: 6.6 acres in the SE 1/4 of the SE 1/4 Section 6; 16.4 acres in the SW 1/4 of the SE 1/4 of Section 6; 7 acres in the NE 1/4 of the NE 1/4 of Section 7; and 14 acres in the NW 1/4 of the NE 1/4 of Section 7, all in T7N, R69W, 6PM. Boundaries have been modified slightly by annexation of about 10 acres of the southeast corner of the original lease by the City of Ft. Collins. H. RELATED Federal FEDERAL PROJECTS Aid to Wildlife Restoration Projects W-153-R-4 and W-152-R-2. �58 DRAFT Appendix A Wildlife Researcher Dan Baker estimated feeding requirements for captive ungulates housed at FWRF using species-specific energy requirements and feedspecific energy availabilities. Estimated feeding costs per head for nonreproducing individuals of all 3 species are below estimated costs for FWRRF's current ad libitum feeding program. These estimates will be evaluated using management experiments at FWRF to develop a more efficacious and efficient program for feeding captive wildlife. GENERAL INPUT DATA ENERGY COST: 1. 2. 3. 4. S. Average year-round body weight (kg) Metabolic BW Basal Metabolic Rate (BMR) - 70 kcal/kg BW 0.7S/day Average daily metabolis rate (ADMR) - 140 kcal/kg BW 0.7S/day Gestation Cost: First trimester: 1.35% x BMR 11.89% x BMR Second trimester: Third trimester: 42.32 x BMR 6. Lactation cost: First month: 1.5 x BMR 2.0 x BMR Second month: Third month: 1.5 x BMR 7. Conception date 8. Gestation period 9. Parturition date 10. Weaning date INTAKE 11. 12. (b/day): Grass hay intake - 1.84% x BW Alfalfa hay intake 1.95% x BW ENERGY CONTENT 13. 14. 15. (ME kca1/g): Grass hay - 1.9 Alfalfa hay - 2.2 Supplement - 2.4 - 2.9 depending on species FEED COST: 16. 17. Hay $lOO/ton Supp1 - $2S0/ton It is assumed requirements Requirements supplement. prevent them that all animals will eat to fill on hay diets. If total are met by this intake, then no supplement is consumed. not met by hay intake will be met by the consumption of Animals are supplemented only when fill limitations of hay intake from meeting ME requirements. �59 DRAFT PRONGHORN DATA INPUT: Ave. year-round boy weight - 46 kg MEW - 17.7 BMR - 1239 kcal/day ADMR - 2476 kcal/day Suppl - 2.4 ME kcal/g Ave. conception date - Sept. 15 Ave. gestation period - May 22 Weaning date - Sept. 1 1. Hon-pregnant Adult. ME Requirements ~~~~_~~~~!"_~ __i~~~y_l ___________ ~c~~j_~~y Month _ Grass Suppl ~~6 ~~~ ~~_7 ~~_1 kg/animal/yr 309 120 327 77 lbs/animal/yr 679 264 720 169 $$$/aniaal/yr 34 33 36 21 ADMR ___ ~!~_][~ Total ~~~_1_~~ ~~~_1_~~ Alfalfa Suppl _ TOTAL 2. Reproducing female. _______ ~__ ~':~~=:~-=~~~ __~c:~L~~~L Month Maint Gest Lact !.':.~~~!! __ (_g.L~!YJ Total Grass Suppl _ Alfal Suppl Jan 2476 147 0 2623 846 391 851 384 Feb 2476 147 0 2623 846 391 851 384 Mar 2476 523 0 2999 846 548 851 540 Apr 2476 523 0 2999 846 548 851 540 May 2476 523 0 2999 846 548 851 540 Jun 2476 0 1728 4204 846 1103 851 1043 Jul 2476 0 1728 4204 846 1362 851 1043 Aug 2476 0 1728 4204 846 1103 851 1043 Sep 2476 0 2492 846 330 897 329 (birth) 16.7 (wean fawn. and breed f••• le.) Oct 2476 16.7 0 2492 846 330 851 329 Nov 2476 16.7 0 2492 846 333 851 329 0 2623 ----- ---- ----- ---- ----- ---- ----- ----------- -----846---- -----333---- ----- ---- ----- ----Dec 2476 147 851 384 TOTAL kg/anima1/yr 309 222 311 210 lbs/animal/yr 679 488 683 462 $$$/aniaal/yr 34 61 34 58 Averaqe annual cost per head (nonreproducinq): $ 57.00 �DRAFT 60 Current estimated cost of ad libitum feeding: MOUNTAIH $ 74.00 SHEEP DATA INPUT ,: Ave. year-round body weight 90 kg MBW = 29 BMR = 2030 kcal/day ADMR = 4066 kcal/day Suppl. = 2.8 ME kcall/g Ave. conception date = Nov 15 Ave. gestation epriod 177 days Ave. parturition date = May 10 Weaning date = Sept 1 1. Hon-pregnant Adults ME Requirements !_~=? __~~~~~_~__i~~~X_l ___________ ~_c~~j_~~x _ Month ADMR Total Grass Suppl Alfalfa Suppl All yr 4066 4066 1656 275 1755 604 100 640 27 lbs/animal/yr 1330 220 1409 58 $$$/animal/yr 66 28 80 7 7 TOTAL kg/animal/yr 2. Reproducing females .!.z:t:~~= __(_g..L~~.¥ J______________ _______~ __ ~~~ ~.::;n~~~= __(_~c:.~j_<!.~X.L ____________________ Month ADMR Gest Lact Total Grass Suppl Alfal Suppl Jan 4066 484 0 4550 1656 448 1755 246 Feb 4066 484 0 4550 1656 448 1755 246 Mar 4066 1720 0 5786 1656 890 1755 687 Apr 4066 1720 0 5786 1656 890 1755 687 May 4066 1720 0 5786 1656 890 1755 687 Jun 4066 0 3045 7111 1656 1363 1755 1161 Ju1 4066 0 4060 8126 1656 1725 1755 1523 Aug 4066 0 3045 7111 1656 1363 1755 1161 Sep 4066 0 0 4066 1656 275 1755 73 Oct 4066 0 0 4066 1656 275 1755 73 Nov 4066 0 0 4066 1656 275 1755 73 (birth) (wean) (conception) 83 4066 27 4093 285 1755 0 1656 -----Dec---- -------- ----- ---- ----- ----------- -------- ----- ---- ----- ---- ----- ----TOTAL kg/anima1/yr 604 274 640 201 �61 DRAFT lbs/animal/yr 1330 ~$~$~$~/~a~n~ia~a~1~/~y~r~ 6~6~ 603 1409 ~7~5 70 Average annual cost per head (nonreproducing): $ 87.00 Current estiaated cost of ad libi~um feeding: $ 96.00 442 55 �DRAFT 62 DATA INPUT: Ave. year-round body weight 229 kg MBW = 58.8 BMR = 4136 kca11jday ADMR = 8272 kca1jday Suppl = 2.4 ME kca1jg Ave. conception date = Sept 15 250 days Ave. gestation period Ave. parturition date = June 1 Weaning date = Sept 1 1. Hon-pregnant Adult. ME Requirements ___________ !_c~}j_~~x ~~~ __~~~~~_e i~~~X_l Alfalfa Suppl o 4466 o 1538 0 1630 0 lbsjanimaljyr 3384 0 3586 0 $$$/animal/yr 169 0 179 0 Month ADMR Total Grass All yr 8272 8272 4214 kgjanima1jyr Suppl _ TOTAL 2. Reproducing females _______ ~ __z:~~~E:~_=~!__:J __C~c:~j_~~~L_____________________ !.~t:~~= __(_gl~~XJ _______________ Month Maint Gest Lact Total Grass Suppl Alfal Suppl Jan 8272 490 0 8762 4214 157 4466 0 Feb 8272 490 0 8762 4214 157 4466 0 Mar 8272 1742 0 10014 4214 679 4466 80 Apr 8272 1742 0 10014 4214 679 4466 80 May 8272 1742 0 10014 4214 679 4466 80 Jun 8272 0 6174 14446 4214 2525 4466 1926 Jul 8272 0 8232 16504 4214 3383 4466 2783 Aug 8272 0 6174 14446 4214 2525 4466 1926 Sep 8272 56 0 8326 4214 0 4466 0 (birth) (wean fawns and breed females) Oct 8272 56 0 8328 4214 0 4466 0 Nov 8272 56 0 8328 4214 0 4466 0 8272 0 490 0 8762 4214 157 4466 -----Dec ---- -------- ----- ---- ----- ---- ----- ---- ----- ---- ----- ---- ----- ---- ----- - - --TOTAL kgjanimaljyr 1538 328 1630 206 Ibsjanima1jyr 3384 722 3586 454 $$$/aniaal/yr 169 90 179 57 �63 DRAFT Average annual cost per head Current estiaated (nonreproducing): cost of ad libitum feeding: $179.00 $265.00 ��65 Colorado Division of Wildlife Wildlife Research Report July 1990 JOB PROGRESS REPORT State of ~C~o~l~o~r~a~d~o~_ Project No. w-lS3-R-3 Work Plan No. ~l~A~ Job No. _ 3 Period Covered: Author: Mammals Research Multispecies Investigations Mammals Research Administration July 1, 1989 - June 30, 1990 R. B. Gill Personnel: R. B. Gill, L. E. Lovett, L. H. Carpenter ABSTRACT Progress towards the objectives of this job include: 1. Since June 1, 1989, I assumed temporary superv1s1on of all personnel affiliated with the Mammals 1 and Mammals 2 Research Sections, including 11 Wildlife Researchers, 1 Senior Secretary, and 1 Publications Specialist. I also assisted in the supervision of 1 Wildlife Research Technician. 2. During Fiscal Year 1989-90 a Final Draft of a Statewide Black Bear Management Plan was completed, reviewed, and published. Management activities necessary to implement the plan were initiated. ��67 MAMMALS RESEARCH ADMINISTRATION R. Bruce Gill P. N. OBJECTIVES Administer research within the Mammals productivity at the lowest cost. Research 2 Unit for the highest SEGMENT OBJECTIVES 1. Assign and supervise the research 2. Assign and supervise secretarial 3. Lead the development, management plans. of 6 Wildlife and clerical implementation, Researchers. work of 1 Senior Secretary. and evaluation of statewide species RESULTS AND DISCUSSION Objective 1. From June 1, 1989 through June 30, 1990, I was temporarily assigned responsibility for supervising the research activities of both the Mammals 1 and Mammals 2 Research Sections. Duties included directing and evaluating the work performance of 11 Wildlife Researchers, 1 Senior Secretary, and 1 Publications Specialist. During this period 11 manuscripts either were published or accepted for publication in peerreviewed journals. In addition, 2 manuscripts were accepted for publication in symposia proceedings. Drafts of Standard Operating Procedures were prepared to guide Division activities in management of mule deer, wapiti, black bear, mountain sheep, mountain goat, pronghorn, and puma. State Personnel Department's Performance Evaluation for Colorado (PACE) Performance Evaluation and Performance Planning Documents prepared for each Mammals 1 and Mammals 2 employee. Employees were Objective 2. A manual was prepared and updated continually regarding Standard Operating Procedures for accounting, purchasing, temporary personnel recruitment, and activities reporting. Four consecutively revised drafts of a black bear management plan were prepared and submitted for review and comment by DOW employees and interested organizations and concerned individuals. The Mammals Research Section Secretary received training with the most recent update of WordPerfect word processing software (WP 5.1) and documents were upgraded to WP 5.1 format. Budget status reports were prepared and sent to Wildlife Researchers quarterly for the period July 1, 1989-March 31, 1990. Budget status was reported monthly for the period April 1-June 30, 1990. �68 Objective 3. 1990-1995 necessary A final draft of the Division's Black Bear Management Plan was completed, reviewed, and published. Management activities to implement the plan were initiated. Work was begun on the preparation Plan. of a statewide LITERATURE Furbearer CITED Gill, R. B., and T. D. I. Beck. 1990. Black bear management Div. Wildl. Div. Rep. No. 15. 44p. Prepared S5 by: ~~~ R. Bruce Gill Wildlife Research Leader Management plan. Colorado �Colorado Division Wildlife Research July 1990 of Wildlife Report JOB PROGRESS REPORT State of Colorado Project No. W-lS3-R-3 Work Plan No. J~~. 4 Period Covered: Authors: Personnel: ~l~A~ Mammals Research _ Multispecies Wild Ruminant Investigations Forage Selection Dynamics July 1, 1989 - June 30 1990 B. J. Maynard P. N. Lehner, M. W. Miller, M. A. Wild, S. Roberts, R. B. Gill ABSTRACT Wavelength perception trials continued in the Y-maze. A successful research protocol for testing pronghorn visual perception has been established. A scotopic sensitivity curve for the pronghorn eye was generated using electroretinography. ��~, I.L WILD RUMINANT FORAGE SELECTION DYNAMICS Barbara J. Maynard P. N. OBJECTIVE To evaluate the role of v~s~on in diet selection foraging ruminant--the pronghorn. of a small, selectively SEGMENT OBJECTIVES 1. Conduct experiments concerning light wave-length and behavioral responses to light cues. sensitivity of pronghorns METHODS AND MATERIALS Wavelength perception trials continued in the Y-maze following the protocol outlined by Maynard (1989). A few changes were made in this protocol. The decision was made in December, 1989, to concentrate work on 1, rather than 6, pronghorn at a time; 1 male (ZO) was run daily from December, 1989, through June 1990. This male was fed alfalfa hay ad libi~, but was deprived of pelleted food except when he completed a run through the maze. For each trial, the male was led into the Y-maze and allowed to choose between 2 arms which differed only in the presence or absence of a light cue. A positive response to the light-off cue was rewarded with food and immediate release. A positive response to the light-on cue was punished with no food and 1 minute of confinement in the arm. Learning of this association and, therefore, perception of the light cue, was considered to have been demonstrated by a response level significantly better than would be expected due to chance; that is, entering the light-off arm at least 17 out of 20 trials (p $ 0.05). The number of trials run per day was determined by the animal's behavior. Trials were terminated when the animal was slow to re-enter the Y-maze for the next trial. Experience has shown this hesitancy to be predictive of lack of attention to the light cues. Number of trials run per day varied between 1 and 7. The initial training cue was white light emitted from a 100 W incandescent bulb through a I-inch hole in the end wall of the Y-maze arm. Once ZO demonstrated an ability to discriminate white light from no light, the cue was narrowed in wavelength to 10 nanometer (nm) bandwidths by the insertion of narrow bandpass filters in front of the light source. Again, perception of the light cue was accepted when ZO correctly responded during 17 of 20 trials. RESULTS AND DISCUSSION ZO reached the successful performance criterion of 17 correct choices out of 20 trials with light cues of white, 500 nm, 550 nm, 600 nm, 650 nm, 700 nm, 750 nm, and 850 nm wavelengths. Trials in which no light cue was presented were then run as a check to confirm that ZO was indeed responding to the light and not to some other unknown cue. ZO got 15 of 20 light trials �72 "correct." Such a high level of correct performance suggested the presence of an additional cue which "told" ZO which arm of the Y-maze was the "correct" choice. Odor was suspected as the additional extraneous cue. The introduction of an odor flooding procedure eliminated ZO's high performance on the no light trials. Wavelength trials are now being retested. Electroretinography (ERG) was performed on 4 castrated pronghorn from the Division of Wildlife's Foothills Wildlife Research Facility (FWRF). Dr. Steve Roberts, D.V.M., of Colorado State University's Veterinary Teaching Hospital Ir-----------------------------------------~ - I IJ .~~~~--_.--_.--~--~--~--._--~--~~ 1'0 400 .1 ••••••• It. II. A •• ,. N •• 70 700 710 W••••••• tII (_) Fig. 1. Pronghorn retinal responses to electroretinographic stimulation at varying wavelengths of light. �73 performed the ERG, and Dr. Mike Miller, CDO~ Research D.V.M., administered the anesthesia. The ERGs suggest that the pronghorn eye under scotopic (darkadapted) conditions is sensitive to a range of light wavelengths (Fig. 1). The shape of the pronghorn spectral sensitivity curve is typical for mammalian eyes (Horn and Lehner 1975, Hope and Bhatnagar 1979). Photopic (lightadapted) sensitivity curves are generally the same shape as scotopic curves, but are shifted to the right along the X-axis about 50 nm (Coren et al. 1984). LITERATURE CITED Coren, S., C. Porac, and L. M. ~ard. 1984. Sensation ed.). Academic Press, Inc. New York, NY. Hope, G., and K. P. Bhatnagar. 1979. spectral stimulation: comparison Experientia 35:1189-1191. Horn, S., and P. Lehner. 1975. 1atrans. J. Compo Physio1. Prepared by and perception (2nd Electrical responses of bat retina of four microchiropteran species. Scotopic sensitivity Psych. 89:1070-1076. in coyotes, Canis to ��75 Colorado Division of Wildlife Wildlife Research Report July 1990 JOB PROGRESS REPORT State of Colorado Project No. W-1S3-R-3 Work Plan No. Job No. _ 5 Period Covered: Author: -=lA~ Mammals Research Multispecies Consulting Analysis Investigations Services for Mark-Recapture July 1, 1989 - June 30, 1990 G. C. White Personnel: R. B. Gill, R. M. Bartmann, T. D. I. Beck, A, Neal, D. Reed ABSTRACT Progress towards the objectives of this job includes: 1. A manuscript summarizing the results of the Piceance Basin deer population studies has been developed, with publication intended as a Wildlife Monograph of The Wildlife Society. 2. A study of compensatory effects of harvest on the Piceance Basin mule deer population has been initiated, with the first experimental harvest in December, 1989. Radio collars to monitor over-winter survival of fawns were placed on the animals during November, 1989. 3. A Monte Carlo simulation study to evaluate mark-resighting methods estimate mountain sheep numbers was completed and manuscripts are currently undergoing peer review. . 4. Consultation has been provided in the design and analysis of markrecapture methods to estimate cray' fish population size. 5. A manuscript has been prepared on evaluating various methods to compute confidence intervals for age and sex ratio data from mule deer and elk. The manuscript is currently undergoing peer review, and is intended for submission to the Journal of Wildlife Management. 6. An interactive interface has been developed for the trapping and marking database in dBase 111+. The user is able to enter data, and to interrogate the database for potential radio frequencies or neck collar colors that could be used in a specified geographic area. 7. A model of a Colorado black bear population has been developed and used to simulate various experimental harvest strategies. A manuscript is currently being prepared for submission to the Journal of Wildlife Management. to ��77 CONSULTING SERVICES FOR MARK-RECAPTURE ANALYSES Gary C. White P. N. OBJECTIVES Model and s i mu La t e population estimates of deer, elk, mountain mountain goats with mark-recapture methods. sheep, and SEGMENT OBJECTIVE Evaluate mountain the joint hypergeometric sheep population size. maximum likelihood population estimator for RESULTS AND DISCUSSION Mark-resight is one procedure to estimate the size of mountain sheep (Ovis canadensis canadensis) populations. This method involves first capturing and marking animals. Subsequently, 1 or more sighting surveys identify and count marked and unmarked animals to provide an estimate of the proportion of the population that is marked, and thus the population size (Rice and Harder 1977, McQuivey 1978, Leslie and Douglas 1979). Mark-resight is based on mark-recapture with a Lincoln-Petersen estimate. Several assumptions of this method must be satisfied to estimate abundance accurately (i.e., without bias and with good precision) (Otis et a1. 1978). First, the population must be closed demographically and geographically. Second, animals must not lose their marks. Third, all marked animals must be correctly identified, counted, and recorded. Fourth, all animals (both marked and unmarked) must have the same, independent probability of being captured or of being sighted during an individual sighting occasion. Violations of these assumptions may result in an estimate that is biased, imprecise, or both. In reality, several of these assumptions may be violated during a survey. If animals form more or less stable social associations, the probability of sighting 1 animal is no longer independent of the sighting probability of another animal, and samples are no longer random. Likewise, some animals are more likely to be sighted than others due to behavioral differences (White et al. 1982), and sighting probabilities will be different. After the initial marking session, each sighting occasion results in a separate Lincoln-Petersen estimate. These individual estimates must be combined to obtain 1 overall population estimate, and the estimator chosen affects the accuracy of the estimate. The joint maximum likelihood estimator formed from the combination of 2 or more hypergeometric distributions (Chapman 1951, Seber 1982:130, Seber 1986), or JHE (Bartmann et al. 1987), possesses the optimal properties of any estimators to combine all the individual estimates. This estimator and the associated profile likelihood confidence interval has advantages over the mean or median of the estimates from each sighting occasion: the variance is smaller; the lower confidence bound in never lower than minimum number alive; and the confidence interval width is narrower (Bartmann et al. 1987, White and Garrott 1990:265-6). JHE of mark- �78 resight is the value of maximized: N which maximizes the following likelihood (ll) (1) 5i.(N and the terms are defined for all i = 2 to k+l number of sightings. The estimate of II can be found by iterative numerical methods. Program NOREMARK has been developed by us to run on an IBM PC or compatible to perform these computations. The effectiveness and dependability of mark-resight methods have been questioned (Furlow et al. 1981, DeYoung 1986, DeYoung et al. 1989). Visibility bias can lead to erroneous population estimates (Pollock and Kendall 1987). Violations of the assumptions of mark-resight may affect the estimator. To date, few studies have mathematically assessed the extent of problems inherent in the mark-resight method. we simulated mark-resight population estimation using JHE to evaluate the precision, bias, and coverage of this estimator with known population sizes. we tested the robustness of JHE to violations of the assumptions of independent and equal sighting probabilities (assumption 4). In addition, we tested JHE when sighting probabilities change between sighting occasions. within each simulation, we examined different population sizes, different number of sighting occasions, and different capture and sighting probabilities to observe possible influences on JHE. Given these four variables, we checked the predictability of the precision and bias for a proposed study. Simultaneously, helicopter surveys of a mountain sheep population in south central Colorado were used to estimate heterogeneity sighting probabilities and group information. METHODS Aerial Surveys and Study Area Field research estimated capture probabilities and sighting probabilities of mountain sheep across multiple resighting occasions, probability distributions for these sightings, average group size, and frequency of aggregation. Markresight surveys took place at Trickle Mountain, 38°09'N l06°20'w, Saguache County, Colorado. Using drop nets, net guns, and capture guns, 25 adult ewes were radio collared in January, February, and March, 1988. Only ewes were collared to avoid potential differences in sighting probability or behavior between age or sex classes, so that heterogeneity of sighting probabilities represents within age/sex class variation. Each collar was uniquely marked for visual identification, with unique radio frequencies in the range 172-173 MHz. Intensive helicopter resightings occurred on 15 occasions in January and February, 1989. To minimize differences between flights, the same 2 observers and pilot were used for all flights, the same seating arrangement was maintained, all surveys took place in the morning between 0800hr and l200hr, �79 and the flight plan was identical from flight to flight. An area was chosen where animals had been previously located and where all the animals were thought to remain during the winter; thus geographic closure was met. When a group was spotted, the location was noted, and the animals were followed and counted, carefully scrutinized for marks, and collars visually identified. Afterward, the helicopter returned to the original flight path. In addition, information was collected on ages, sexes, and reactions of the animals, distances at which groups were sighted, habitat types, and percent snow and cloud cover. The flight path was unidirectional and non-overlapping to avoid counting groups twice. Because of logistic and weather restrictions, pairs of flights were flown on consecutive days 6 times, and 3 flights flown on 3 consecutive days 1 time. After each survey, a verification flight with a telemetry receiver documented which marked animals were off the study area during the survey, which marked animals were not seen but were on the study area, and which marked animals had died. During verification, visual sightings were attempted for marked animals off the study area to expose all animals to the same treatment which may have occurred if the animals had been on the study area. Weather variables were fairly constant in these periods and were not considered a confounding factor. Animal locations during the survey and during the verification were plotted on maps following the flights. In the analysis of group information from the aerial surveys, all pair-wise combinations of marked animals in each group were considered; a proportion of the total flights where 1 animal was seen in association with 1 other animal was calculated. The maximum number of times 1 of the 2 animals was seen was used as the denominator. These proportions account for when an animal was seen, but do not account for when an animal was alone, when an animal was the only marked animal in the group, or group size. Maximum likelihood logistic regression tests were used to detect differences in sighting probabilities between individuals, differences in comprehensive sighting probabilities between flights, and possible interactions of the sighting probabilities among the animals and the date of the flight. Maximum likelihood was used to estimate the parameters of the beta distribution that best fit the observed sighting probabilities. The Kolmogorov-Smirnov goodness-of-fit test was used to test the fit of theoretical distributions to the observed sighting probability distribution. Simulations The Statistical Analysis System (SAS) (SAS Institute 1985) was used for Monte Carlo simulations. For the basic set of simulations, population sizes of 50, 100, 200, and 500 were used with 5, 10, 15, and 20 sighting occasions, capture probabilities of 0.1, 0.3, and 0.5, and sighting probabilities of 0.1, 0.3, 0.5, and 0.7. To reduce sampling variation, 1,000 replications were performed for each of these 192 scenarios. Capture and sighting probabilities were assumed constant for each member of the population. For each simulation, the number of marked animals, and the number of marked and unmarked animals seen per flight were generated from a binomial distribution. An estimate was then derived using the JHE. The 95% confidence interval and its length were computed (modified from Hudson 1971) by finding the values of N less than and greater than the estimate for which the log likelihood value was greater than 2 units below the log likelihood value of the maximum; the confidence interval values were the next values above and below each of these values, respectively. The confidence interval length was the difference between the upper and lower confidence interval endpoints. The coverage of the true population size by the confidence interval was then determined (above, below, covering). �80 Several additional sets of simulations were conducted to evaluate the robustness of the estimator. The assumption of independent observations was tested by simulating a range of aggregations among marked animals. Group sizes ranged from 1 to 24; the group size distribution was determined from field data. Two sets were simulated with groups remaining intact for all occasions: sighting probabilities were a function of group size or were independent of group size. Second, the robustness of JHE to the violation of the assumption of equal sighting probabilities was tested by simulating a continuous range of heterogeneity of sighting probabilities of individuals. A symmetrical, bellshaped distribution of sighting probabilities was used. Third, sighting probabilities were changed through time: random changes between 0.05 and 0.95, exponential increase (e.g., sighting probabilities increase as animals become more tolerant of helicopters), and exponential decrease (e.g., animals become more helicopter shy). To evaluate the simulations, length (elL), and confidence as percent relative bias (PRB) , confidence interval interval coverage were used. PRB was estimated PRB (E(N) - N) x 100 (2) N where E(N) is estimated as the mean of the simulated estimates. Overestimation (E(N) > N) results when PRB is positive, and underestimation (E(N) < N) results when PRB is negative. Percent. en, (peIL) is the en, divided by the population size to allow across population size comparisons. Analysis of variance (ANOVA) was used to identify variables affecting the estimate for PRB and pelL. Logistic regression was used to identify variables affecting coverage. These analyses were performed for the basic simulations where all assumptions were met. For the simulations with violations of assumptions, PRB, pelL, and coverage were compared to those of the basic simulations through means, standard errors, and t tests. RESULTS Aerial Surveys Flights occurred in January and February, 1989 (Table 1). The number of marked animals in the area surveyed ranged from 14 to 23 of 25 animals. Time spent searching with 2 observers was approximately 2.6 min/km2 for 33.67 km2. Proportion of marked animals seen ranged from 0.35 to 0.86, with a mean of 0.57 (SO - 0.153). There was no difference (P - 0.208) between the proportion of marked animals seen between the first and the second day of each set of flights. Observed group size ranged from I to 43 (Figure 1) for all flights and verification flights. Mean group size was 6.3 (SO - 6.0) with a median of 4. There was no difference (f - 0.638) in group size between the first and second day of each set of flights. Some animals spent most of their time together, e.g., 3 were together for 10 of 11 flights where they were sighted. �81 Table 1. Statistics Date Jan 14 Jan 17 Jan 18 Jan 19 Jan 27 Jan 28 Feb 1 Feb 2 Feb 8 Feb 9 Feb 15 Feb 16 Feb 22 Feb 23 Mean collected on the Trickle Mountain _1 T· tli !!h Ih 25 25 25 25 25 25 25 25 25 25 25 25 25 25 22 21 21 20 20 20 22 18 23 18 20 17 20 21 9 11 14 11 10 9 19 8 13 8 14 6 13 17 40 63 66 57 52 61 87 46 63 30 69 35 51 57 sheep population. Chapman Estimate 93.3 116.3 97.3 100.5 100.2 129.2 100.2 98.2 108.7 64.4 97.0 9l. 6 77 .0 69.9 79.6 30 25 20 >u ~ OJ ~ 15 0" OJ ,_, n T! ~ 10 5 o 1 3 5 7 9 11 13 Group 15 17 Slze 19 21 Figure 1. Frequency distribution of group sizes observed surveys and the verifications (hatched - survey, unshaded 23 28 43 during both the - verification). However, mostly the marked animals seemed to intermix fairly readily, and some marked individuals were never seen together. The number of times that 1 marked individual was sighted with another marked individual varied from 1 (0.083 of the total flights where an animal was sighted) to 11 (0.917 of the total flights where an animal was sighted). On average, 2 marked animals were �82 sighted together about 26.4% of the flights where they were seen together at least once and where at least 1 of the animals was seen. Because all marked animals were determined to be on or off the study area for each flight, a sighting probability for each marked animal could be calculated. Sighting probabilities for these 25 animals ranged from 0.33 to 0.86, with a mean of 0.58 (SD = 0.16) (Figure 2). A logistic regression model that included flights, animals, and their interactions demonstrated no significant differences in sighting probabilities between animals (P = 0.272) but there were significant differences between dates (P = 0.022). Specifically, a high proportion of marks (0.86) were seen on 1 February (flight #7) and a low proportion (0.33) were seen on 16 February (flight #12). When these dates were removed from the analysis, date no longer affected the sighting probability (P - 0.208). The beta distribution providing the best fit to the observed sighting probabilities has parameters a = 5.21 and = 3.75. Note that this fitted distribution will have a larger variance than the true sheep sighting distribution because the observed sighting probabilities include binomial sampling variation. a Based on a Kolmogorov-Smirnov goodness-of-fit test, the symmetrical, bellshaped beta distribution (a - 3, P - 3) used in the simulations fit the sighting probabilities from the field data (greatest difference - 0.16, reject at 0.38). The uniform distribution also fit, but not as well (greatest difference - 0.25). 7 0.28 ~ ~ 6 ro ~ ~ 5 ~ > ~ ~ 4 0.24 ~ ro ~ ~ ·rl > rl 0.20 ~ c ~ ~ G.16 x ~ ro 3 0.12 ~ ~ w ~ 2 8 ~ 0.08 r w c 'rl ~ 0 8 0 z 1 0 c 0 'rl 0.04 0.2 0.4 Sighting 0.6 D. 8 probabllity 1.0 ~ ~ 0 ~ 0 ~ ~ Figure 2. Frequency distribution of sighting probabilities for all marked animals. Circles indicate exact sighting probabilities; bars indicate groupings of sighting probabilities used in goodness-of-fit tests; the smooth curve is the best fitting beta distribution (a = 5.21, P = 3.75). Simulations A more complete summary of the simulations described here is presented in Neal (1990). Here, we present only general trends observed when the 4 parameters �83 of the simulations are summarized separately. In Neal (1990), the simulations are summarized for each of the individual scenarios. All Assumptions Met.--Precision of the estimator improved and bias decreased as true population size, number of sighting occasions, capture probability, or sighting probability increased. PRB was 1.04% with standard error 0.03 (Table 2), and ranged from -0.2 to 19.9% in the 192 scenarios. Low values of population size, number of flights, or capture or sighting probabilities caused poorer PRB (f < 0.001). Mean PCIL was 34.60% with standard error 0.32, and ranged from 2.0 to 12001.0%. As with PRB, scenarios with low population sizes, few number of flights, or low capture or sighting probabilities had greater PCIL (f < 0.001). Coverage was used to assess overall performance of the estimator. Percent coverage had a mean of 95.34% and ranged from 93.6 to 96.7%; percent of confidence intervals above the true population size ranged from 13 to 3.5%; and percent below ranged from 1.1 to 3.8%. Coverage with JHE is very good even with scenarios having low values of the 4 variables. With logistic regression procedures, neither popUlation size, number of flights, sighting probability, capture probability, nor any of their interactions affected coverage (P ~ 0.32). Apparently, coverage was not affected when all assumptions are met. Aggregations among animals.--With the groups intact for all sighting occasions and sighting probabilities independent of group size, PRB was 1.08% with a standard error 0.03 (Table 3), and ranged from -0.2 to 16.5%. PRB was not different between the basic simulations without groups and the group simulations. �84 Table 2. Percent relative bias (PRB) , mean percent confidence interval length (PClL), and coverage results for basic model simulations where all assumptions of the mark-resight estimator are satisfied. SE Subdivision PRB All variables 1.044 Population size 50 1. 766 100 1. 373 200 0.796 500 0.239 Number of flights 5 2.181 10 0.917 15 0.574 20 0.502 Sighting probability 0.1 3.065 0.3 0.745 0.5 0.264 0.7 0.101 Capture probability 0.1 2.025 0.3 0.754 0.5 0.351 SE PRB 0.029 Mean PClL 34.600 PClL 0.322 Coverage (%) Above (%) 95.345 2.291 0.081 0.068 0.046 0.025 53.008 41.386 27.764 16.242 1.015 0.754 0.204 0.074 95.317 95.352 95.285 95.427 2.273 2.273 2.360 2.258 0.092 0.053 0.040 0.033 60.468 32.126 24.823 20.982 1. 254 0.213 0.123 0.097 95.354 95.179 95.429 95.419 2.310 2.350 2.188 2.317 0.106 0.041 0.026 0.017 78.964 29.651 18.365 11.418 1.260 0.104 0.058 0.036 95.369 95.315 95.421 95.277 2.292 2.429 2.173 2.271 0.073 0.042 0.025 57.091 29.264 17.443 0.886 0.360 0.088 95.339 95.294 95.403 2.306 2.294 2.273 Table 3. Percent relative bias (PRB) , mean percent confidence interval length (PClL), and coverage data for simulations with the same groups remaining intact for all sighting occasions. SE Subdivision PRB All variables 1.097 Population size 1. 803 50 1.444 100 200 0.822 500 0.319 Number of flights 5 2.067 10 1.139 15 0.705 20 0.478 Sighting probability 0.1 3.168 0.3 0.647 0.5 0.224 0.7 0.349 Capture probability 0.1 2.013 0.3 0.905 0.5 0.374 SE PRB 0.031 Mean PClL 36.354 PClL 0.435 Coverage (%) 90.592 Above (%) 6.981 0.080 0.075 0.052 0.026 55.336 44.831 29.055 16.197 1.244 1.094 0.512 0.076 79.110 92.704 95.223 95.331 18.446 4.694 2.448 2.338 0.091 0.066 0.040 0.033 68.773 34.588 23.105 18.982 1.461 0.910 0.154 0.113 92.625 91. 056 89.838 88.850 4.860 6.427 7.862 8.775 0.113 0.043 0.023 0.014 92.712 28.331 15.941 8.433 l. 688 0.290 0.056 0.033 95.083 94.583 92 .146 80.556 2.481 2.881 5.323 17.240 0.076 0.046 0.027 60.240 31.402 17.420 1.127 0.592 0.262 90.739 90.802 90.236 6.739 6.816 7.389 �85 PClL had a mean of 36.35%, a standard error of 0.44, and a range over the 192 scenarios from 0 to 15935.0%. PClL was significantly lower for the basic simulations than for the group simulations. The main differences occurred in the number of flights, lower capture probabilities, and sighting probabilities. Percent coverage had a mean of 90.59% and ranged from 22.7 to 96.8%; percent of intervals above the true population size ranged from 1.1 to 76.5%; and percent below ranged from 0.6 to 3.9%. Coverage appeared to be much poorer when individuals were grouped than when they were not, as in the basic simulations. Percent below is similar to the basic simulations, but percent above was much greater; overestimation is the worse result of this increased variability. For a second set of simulations of aggregations among animals, sighting probabilities were a function of group size. In this set of simulations, mean sighting probability was 0.6 overall so a direct comparison between this and the previous simulations where Z sighting probability - 0.4 is not possible. PRB was 0.39% with a standard error of 0.02 (Table 4) and a range from -0.1 to 1.9%. While these values appear better than other simulations, when compared to the basic simulations with a similar interpolated mean between 0.5 (PRB = 0.26%, SE - 0.03) and 0.7 (PRB = 0.10%, SE = 0.02), these simulations with sighting probability a function of group size are more biased than simulations wi thout groups. Table 4. Percent relative bias (PRB), mean percent confidence interval (PClL), and coverage data for simulations where sighting probabilities by group size (Z sighting probability = 0.6). Subdivision PRB All variables 0.387 Population size 50 0.876 100 0.371 200 0.225 500 o.on Number of flights 5 0.411 10 0.405 15 0.331 20 0.403 Capture probability 0.1 O. rn 0.3 0.258 0.5 0.132 SE PRB 0.025 Mean PCIL 12.032 SE PCIL 0.046 Coverage {%2 78.425 Above {%2 12.529 0.061 0.059 0.046 0.029 14.546 14.511 11.616 7.454 0.116 0.100 0.073 0.041 70.833 80.625 80.900 8l. 342 21.858 9.958 9.150 9.150 0.061 0.051 0.045 0.043 18.635 12.126 9.417 7.950 0.124 0.078 0.060 0.051 88.383 80.325 74.742 70.250 6.833 11.617 14.458 17.208 0.061 0.037 0.024 18.873 10.724 6.499 0.094 0.060 0.036 78.862 78.894 n . 519 12.131 12.144 13.312 length differ PClL had a mean of 12.03%, a standard error of 0.05, and a range from 0 to 160.0%. When comparing these values with the simulations without groups at an interpolated mean of 0.6 between 0.5 (PClL - 18.36%, SE - 0.06) and 0.7 (PClL = 11.42%, SE - 0.04), pelL appears the same. Percent coverage had a mean of 78.42% and ranged from 55.2 to 91.2%; percent of confidence intervals above the true population size ranged from 4.3 to �86 34.2%; and percent below ranged from 3.2 to 15.8%. Coverage is worse here than with no groups or where groups have the same sighting probability regardless of their size. With smaller population sizes, overestimation becomes much worse. Heterogeneity of Individual Sighting Probabilities.--Heterogeneity of sighting probabilities of individuals was simulated from a symmetrical, bell-shaped, beta distribution (a = 3, P - 3). PRB had a mean of 0.60%, a standard error of 0.04 (Table 5), and a range of -0.18 (absolute minimum of 0.01) to 2.73%. These simulations had a mean sighting probability of 0.5. When compared to the basic simulations with constant sighting probabilities (PRB = 0.26%, SE 0.03), PRB with heterogeneity of sighting probabilities was significantly more biased than with homogeneity (f < 0.001). All variables made a difference, especially low capture probabilities. Table 5. Percent relative bias (PRB) , mean percent confidence interval length (PCIL), and coverage data for simulations with heterogeneity of individual sighting probabilities 0.5) . (R sighting probability Subdivision PRB All variables 0.595 Population Size 50 0.890 100 0.817 0.434 200 500 0.239 Number of flights 5 0.907 10 0.595 15 0.521 20 0.357 Capture probability 0.1 1.246 0.3 0.431 0.5 0.108 SE PRB 0.039 Mean PClL 18.701 SE PClL 0.064 Coverage (%} 78.852 Above (%) 10.267 0.102 0.090 0.067 0.041 25.997 22.113 16.429 10.265 0.142 0.139 0.102 0.057 79.267 78.892 78.583 78.667 9.842 10.125 10.417 10.683 0.094 0.076 0.073 0.069 27.998 18.844 15.094 12.868 0.174 0.107 0.085 0.071 88.992 81. 492 74.792 70.133 5.575 9.092 12.308 14.092 0.095 0.059 0.037 28.920 16.738 10.444 0.128 0.083 0.048 78.594 79.006 78.956 10.506 10.138 10.156 PCIL had a mean of 18.70% with a standard error of 0.06, and ranged from 3.0 to 207.0%. When compared to the basic simulations with constant sighting probabilities (PClL - 18.46%, SE - 0.06), the simulations with heterogeneity of sighting probabilities had significantly higher PClL (f < 0.001). The differences were greater with low population sizes, with all number of flights, and with lower capture probabilities. Percent coverage with heterogeneity of sighting probabilities had a mean of 78.8% and ranged from 68.4 to 90.4%; percent of confidence intervals above the true population size ranged from 4.5 to 16.4%; and percent below ranged from 3.9 to 17.9%. The increase in bias and decrease in precision with heterogeneity of sighting probabilities causes problems with confidence interval coverage. Sighting Probabilities Change with Time.--The last set of simulations involved simulating changes with time in sighting probabilities - random changes, exponential increase (helicopter tolerant), and exponential decrease �87 (helicopter shy). For random changes, our simulations had nondirectional changes with each flight with a mean sighting probability of 0.5 and were compared to the basic simulations with mean sighting probability of 0.5. PRB (mean - 0.09%, SE - 0.02) (Table 6) was similar to the basic sim~lations. With increasing sighting probabilities per flight, our simulations had a mean sighting probability of 0.6; PRB (mean = 0.20%, SE - 0.01) (Table 7) increased slightly. With decreasing sighting probabilities per flight, our simulations had a mean sighting probability of 0.4; PRB (mean = 0.44%, SE ~ 0.01) (Table 8) were similar to constant sighting probabilities if initial sighting probabilities were high. In other words, if sighting probabilities change with time, PRB will not be significantly worse as long as some of the flights have fairly high sighting probabilities. PCIL followed similar trends to PRB. With nondirectional changes with each flight, PCIL (mean ~ 12.92%, SE = 0.06) was similar to the basic simulations. With increasing sighting probabilities per flight, PCIL (mean = 14.96%, SE 0.02) also was similar to the basic simulations with constant sighting probabilities. With decreasing sighting probabilities per flight, PCIL (mean = 23.44%, SE - 0.02) could be lower if the initial sighting probabilities are very high. Coverage was also similar with changes in sighting probabilities and some high sighting probabilities than with constant: coverage with nondirectional changes was 95.3%; coverage with increasing sighting probabilities was 95.4%; and coverage with decreasing sighting probabilities was 95.4%. Table 6. Percent relative bias (PRB) , mean percent confidence interval length (PCIL), and coverage data for simulations where sighting probabilities for the population changes with each sighting occasion (R sighting probability - 0.5). Subdivision PRB All variables 0.087 Population size 50 0.070 100 0.156 200 0.100 500 0.022 Number of flights 5 0.276 10 0.094 15 -0.003 20 -0.019 Capture probability 0.1 0.305 0.3 0.021 0.5 -0.065 SE PRB 0.020 Mean PCIL 12.925 SE PCIL 0.055 Coverage (%) 95.252 Above (%) 2.002 0.052 0.046 0.036 0.021 17.426 15.202 11.707 7.365 0.136 0.122 0.092 0.050 96.017 95.267 94.375 95.350 l. 500 2.133 2.358 2.017 0.061 0.037 0.029 0.024 21.288 12.790 9.653 7.970 0.167 0.084 0.059 0.047 95.200 94.792 95.492 95.525 2.150 2.175 l. 950 1.733 0.048 0.031 0.019 20.026 11.606 7.144 0.120 0.077 0.044 95.406 95.269 95.081 2.050 2.000 l. 956 �88 Table 7. Percent relative bias (PRB) , mean percent confidence interval (PClL), and coverage data for simulations with sighting probabilities increasing with each occasion (z sighting probability = 0.6) . Subdivision PRB All variables 0.202 Population size 0.284 50 0.298 100 200 0.160 0.066 500 Number of flights 5 0.486 10 0.166 15 0.086 20 0.071 Capture probability 0.1 0.433 0.3 0.142 0.5 0.031 SE PRB 0.006 Mean pelL 14.956 SE pelL 0.017 Coverage (%} 95.376 Above (%} 2.228 0.016 0.014 O.Oll 0.007 20.409 17.690 13.335 8.392 0.042 0.039 0.028 0.016 95.510 95.340 95.278 95.377 2.059 2.243 2.362 2.247 0.020 O.Oll 0.009 0.007 25.951 14.593 10.666 8.616 0.050 0.025 0.018 0.015 95.351 95.297 95.404 95.454 2.224 2.265 2.233 2.190 0.015 0.010 0.006 23.125 13.417 8.327 0.038 0.024 0.014 95.368 95.356 95.406 2.240 2.241 2.202 Table 8. Percent relative bias (PRB) , mean percent confidence interval (pelL), and coverage data for simulations with sighting probabilities decreasing with each occasion (z sighting probability - 0.4). Subdivision PRB All variables 0.440 Population size 50 0.706 100 0.589 200 0.328 500 0.139 Number of flights 5 0.688 10 0.441 15 0.343 20 0.289 Capture probability 0.1 0.895 0.3 0.325 0.5 0.101 length length Coverage (%} Above (%) 0.024 95.432 2.243 32.868 27.681 20.490 12.726 0.057 0.054 0.039 0.021 95.413 95.390 95.451 95.474 2.213 2.262 2.241 2.256 0.023 0.018 0.016 0.015 30.599 23.852 20.692 18.623 0.065 0.046 0.039 0.034 95.397 95.445 95.466 95.419 2.213 2.232 2.236 2.291 0.022 0.014 0.009 36.174 20.990 13 .161 0.052 0.032 0.018 95.462 95.430 95.403 2.226 2.277 2.225 SE PRB Mean PClL 0.009 23.441 0.024 0.021 0.016 0.009 SE pelL DISCUSSION Mark-resight is one procedure used to estimate population size of mountain sheep. Several assumptions of this method may be violated in practice. Groups of Animals.--The assumption animals group together. Groups of precision but not as much as other simulations, groups of individuals of independent sightings can be violated if individuals affect the bias and the violations of assumptions. From our will not bias the estimate but, with low �89 sighting probabilities, the estimate will be more imprecise and coverage poorer. Rice and Harder (1977) also report an increase in variance of the estimate with increases in clumping. With populations of over 200, this increase in variability is minimal. Overestimation is a problem regardless of population size, number of flights, capture probability, or sighting probability. In reality, sighting probabilities are probably a function of group size, and from simulations with sighting probability differing by group size, the bias increases but the precision remains the same as in a population without groups. Coverage decreases so that overestimation is more of a problem with smaller population sizes, but both underestimation and overestimation are problems with larger population sizes. The Trickle Mountain data indicates that the animals do not tend to stay together for long periods of time. While information could not be obtained on group sighting probabilities, we expect that they vary with group size (Samuel et al. 1987). Consequently, we expect the estimates to be more biased than a population without groups, but not much more imprecise. Overestimation is our primary concern with the herd at Trickle Mountain. Heterogeneity of Sighting Probabilities.--Heterogeneity of sighting probabilities seems to affect the estimates more than violations of the other assumptions. White and Garrott (1990:264) suggest bias occurs with individual sighting heterogeneity. Our simulations show that heterogeneity leads to biased estimates, especially with smaller population sizes. The bias, however, is a maximum of 7.45% for one scenario; if this is a tolerable level, bias may not be a concern. More importantly, coverage tends to be too low and the estimates are more imprecise than the estimated confidence interval would suggest. With mountain sheep at Trickle Mountain, although the individual sighting probabilities range from 0.33 to 0.86, there is no significant difference among animals. Alternatively, the goodness-of-fit test indicates that a symmetrical, bell-shaped distribution fits the field data well. Precision and coverage are much poorer for data generated with a symmetrical, bell-shaped distribution than data generated with equal sighting probabilities; coverage can be as low as 68.4% (mean sighting probability of 0.5). Perhaps more flights or more animals would reveal if there are differences in the sighting probabilities of the animals at Trickle Mountain. For the populations where heterogeneity of sighting probabilities is a concern, greater sighting or capture probabilities would decrease the bias slightly and reduce confidence interval length. Sighting Probabilities Change with Occasion.--In addition to violations of assumptions, other sources of variation may have had an impact on the estimate: comprehensive sighting probabilities changing randomly with each flight, or sighting probabilities increasing or decreasing exponentially over time. When sighting probabilities change with each flight, the precision is greatly improved if sighting probabilities for some days are very high (e.g., ~0.95); if sighting probabilities are not high, then precision may be poorer. The bias is also lower than when sighting probabilities are constant, especially with fewer flights and smaller capture probabilities. Overall, even when some flights have high sighting probabilities, the coverage is lower and some underestimation is a problem. In the surveys at Trickle Mountain, the proportion of marks seen per day was variable but followed no particular pattern of increase or decrease. The population estimate and its associated variance are expected to be lower than actual values. We recommend concentrating effort to maximize the proportion of animals seen by methods such as choosing ideal weather conditions and using experienced observers. �90 The proportion of marks seen can also change with number of flights as animals become more or less habituated to the helicopter. When sighting probabilities increase with time, either the observers' abilities increased, the animals run more and are more visible, or the animals do not try to evade the helicopter and remain in the open. The estimates are slightly more biased but are not more imprecise than when sighting probabilities are constant. The resulting estimate would appear fairly precise but would be slightly biased. Maximizing sighting probability is important, so ideally flights should continue until sighting probabilities are high. When sighting probabilities decrease with time, the bias and variability may be lower than when a constant proportion of marks is seen if the initial sighting probability is very high. More flights and higher capture probabilities can improve the coverage slightly but the initial sighting probability is the main factor. However, the decrease in coverage in both these scenarios is minimal enough that these concerns could be disregarded. In the analysis of the field data, proportion of marks sighted do not follow a trend with time and there is no significant difference between consecutive flights or from the beginning to the end of the survey. Likewise, the group sizes do not change between consecutive flights. These results suggest that the sheep at Trickle Mountain are not becoming more or less habituated to the helicopter with time, nor are they reacting by changing group size. Population Estimate for Trickle Mountain Population The Trickle Mountain study area did not contain all the marked animals for all the sighting occasions, so the assumption of geographic closure was violated. Therefore, the estimator presented in Eq. (1) is not appropriate for this experiment. We have extended the estimator in Eq. (1) to include immigration/emigration through a binomial process. Assume that the total sheep population that has any chance of being observed on the study area is N*, and that at the time of the ith survey, Ni sheep are on the study area. We are interested in estimating the mean number of sheep on the study area, li, and possibly N*. At the time of the ith sighting occasion, a known number of the marked sheep (Mi) are on the study area of a possible Ii sheep with transmitters. The probability that an individual sheep is on the study area on the ith occasion can be estimated as Mi/Ii' Then the likelihood for the model that included immigration and emigration is a product of the binomial distribution for emigration/immigration times the joint hypergeometric likelihood of Eq. (1): The parameters N* and Ni for i-2 to k+l can be estimated by numerical iteration to maximize this likelihood, with the constraints that Ni > (Mi - fii + TIi) and N* > Nt for i-2 to k+l. Profile confidence intervals can be obtained for the k+l parameters. We were not interested in the k population estimates for each sighting occasion, but rather desired the mean of the Ni estimates. Therefore, we re-parameterized the likelihood to estimate the mean population size on the study area directly, and obtain its profile likelihood confidence interval. �91 For the data taken on the Trickle Mountain sheep population (Table 1), the total population (N·) was estimated as 120 (95% CI 109-133), with the average number of sheep on the study area estimated to be 96.1 (95% CI 88-106). These estimates seem reasonable because the minimum number of animals known to exist is 93 (66 unmarked sheep observed on 1 Feb flight, plus 25 marked sheep). Program NOR&~~ is available to perform these calculations. In theory, th~ estimator presented in Eq. (3) should be superior to the estimators presented by Eberhardt (1990) because it is a maximum likelihood estimator and hence is asymptotically minimum variance and consistent. However, Chapman (1951) demonstrated the bias of the maximum likelihood estimator for small sample sizes. Eberhardt's (1990) estimators are based on Chapman's estimator, which corrects the bias for situations where ill + ilz ~ N. Hence, for small sample sizes, his estimators may be appropriate, but the differences should decrease as population size increases. A second consideration is confidence interval construction. The profile likelihood intervals advocated here always generate a lower bound ~ to the minimum number of animals known alive. Other methods for constructing confidence intervals do not explicitly incorporate this logical lower bound. Finally, for poor quality experiments, 1 or more values of illi - O. Eberhardt (1990) refers to this problem as an infinite bias. For the 2 estimators presented here, if at least 1 illi > 0, the estimator will be defined and still produce a maximum likelihood estimate. Occasions with illi - 0 contribute information about population size, but not enough information to produce logical estimates. When these occasions are combined with the remaining occasions where illi > 0, a logical estimate can still be produced. If illi - 0 for all i, no logical estimate of N can be constructed, although an upper and lower probability bound can be constructed. The estimator described by Minta and Mangel (1989) should provide a method robust to heterogeneity of individual sighting probabilities. Although this estimator requires geographic closure, immigration/emigration from the study area can be viewed as a form of sighting heterogeneity and the estimator is still appropriate to estimate the total sheep population for the study area. The number of resightings for the 25 marked sheep were 1 with 2, 4 with 3, 2 with 4, 3 with 5, 4 with 6, 2 with 7, 2 with 8, 3 with 9, 2 with 10, and 2 with 12, giving 162 marks observed. For unmarked sheep, 615 were observed in the 14 flights. The program supplied by Minta estimates the population size as 95, with a 96.9% CI as 87-104. This estimate seems too low, given that the minimum number of animals known to exist is 93. Further, the lower confidence bound is less than this known quantity, and in general this confidence interval seems too narrow. The heterogeneity caused by immigration/emigration may cause this estimator to not perform well with these data. Design of Mark-Resight Experiments Two approaches are available to compute the sample size necessary for the thorough design of a mark-resight experiment. First, if the expected parameters of the experiment fall into the range of parameter values simulated in Neal (1990), linear interpolation can be used to evaluate elL and PRS from the tables presented there. Program NOREMARK has been developed to provide linear interpolations based on these simulations. Second, additional simulations can be conducted with parameters specific to the proposed experiment. Program NOREMARK can also perform these simulations, although the computer time required is considerably greater than for linear interpolation. �92 CONCLUSIONS We recommend mark-resight and the JHE as a population estimation method for mountain sheep where closure and a well-defined study area are available. The bias is low and the precision is good if all assumptions are met. We note that confidence interval length and coverage can be very poor with small population sizes, few flights, or low capture or sighting probabilities (e.g. Bartmann et al. 1987). The violation of the assumption of random samples (e.g. groups of animals) increases the bias and decreases the precision, but not as much as heterogeneity of sighting probabilities. The groups in the Trickle Mountain herd, however, are not generally large enough or dependent enough to be a large consideration in affecting an estimate. The violation of the assumption of equal sighting probabilities worsens the precision and the coverage more than with equal probabilities. From the Trickle Mountain data, heterogeneity of sighting probabilities may exist and needs to be considered as a factor decreasing the precision and the coverage. In all cases, increasing the sighting probabilities and increasing the number of flights have the most influence in improving the estimator; during flights, concentrated effort to achieve high sighting probabilities is most beneficial. LITERATURE CITED Bartmann, R. M., G. C. White, L. H. Carpenter, and R. A. Garrott. 1987. Aerial mark-recapture estimates of confined mule deer in pinyon-juniper woodland. J. Wildl. Manage. 51(1):41-46. Chapman, D. G. 1951. Some properties of the hypergeometric distribution with applications to zoological sample censuses. Univ. Calif. Publ. Stat. 1(7):131-160. DeYoung, C. A. 1986. Accuracy of helicopter Wi1d1. Soc. Bull. 13:146-149. surveys of deer in south Texas. _______ , F. S. Guthery, S. L. Beasom, S. P. Coughlin, and J. R. Heffelfinger. 1989. Improving estimates of white-tailed deer abundance from helicopter surveys. Wi1d1. Soc. Bull. 17:275-279. Furlow, R. C., M. Haderlie, and R. Van den Berge. 1981. Estimating a bighorn sheep population by mark-recapture. Desert Bighorn Council, Trans. 1981:31-33. Eberhardt, L. L. 1990. Using radio-telemetry for mark-recapture edge effects. J. Applied Ecology 27:259-271. Leslie, D. M., Jr., and C. L. Douglas. 1979. Desert bighorn River Mountains. Nevada. Wild1. Monogr. 66:1-56. studies with sheep of the ______ , and C. L. Douglas. 1986. Modeling demographics of bighorn sheep: current abilities and missing links. N. Amer. Wi1d1. Natur. Res. Conf. Trans. 51:62-73. McQuivey, R. P. 1978. Bio1. Bull. No.6. The desert bighorn 81pp. sheep of Nevada. Nevada Fish, Game �93 Minta, S. and M. Mangel. 1989. A simple population estimate based on simulation for capture-recapture and capture-resight data. Ecology 70:1738-1751. Otis, D. L., K. P. Burnham, G. C. White, and D. R. Anderson. 1978. Statistical inference from capture data on closed animal populations. Wildl. Monogr. 62:1-135. Pollock, K. H., and W. L. Kendall. 1987. Visibility bias a review of estimation procedures. J. Wildl. Manage. in aerial surveys: 51(2):502-510. Rice, W. R., and J. D. Harder. 1977. Application of multiple aerial sampling to a mark-recapture census of white-tailed deer. J. Wi1d1. Manage. 41(2):197-206. SAS Institute Inc. 1985. SASR Language Guide for Personal 6 Edition. SAS Institute Inc., Carey, N.C. 429pp. Computers, Version Samuel, M. D., E. O. Garton, M. W. Schlegel, and R. G. Carson. 1987. Visibility bias during aerial surveys of elk in northcentral Idaho. Wildl. Manage. 51(3):622-630. Seber, G. A. F. parameters 654pp. J. 1982. The estimation of animal abundance and related (2nd edition). Macmillan Pub1. Co., Inc., New York, N.Y. 1986. A review 292. for estimating animal abundance. Biometrics 42:267- White, G. C., D. R. Anderson, K. P. Burnham, and D. L. Otis. 1982. Capturerecapture and removal methods for sampling closed populations. Los Alamos National Laboratory. LA-8787-NERP. Los Alamos, N.M. 235pp. , G. C., and R. A. Garrott. 1990. Analysis of wildlife data. Academic Press, New York, N.Y. 383pp. Prepared by &n?J (: (:>j~. Gary C. Wh' e Associate Professor radio-tracking ��Colorado Division of Wildlife Wildlife Research Report July 1990 JOB PROGRESS REPORT State of Colorado Pro J ec t No. _W",--=-15~3~-....:R,",---,3==-_ Mammals Research Work Plan 2A Mountain Job No. 4 Experiments to Identify and Manage Stress in Mountain Sheep Period Covered: July 1, 1989 - June 30, 1990 Author: No. Sheep Investigations M. W. Miller Personnel: M. A. Wild, H. A. Petit, N. T. Hobbs, R. B. Gill, J. A. Bailey ABSTRACT Sampling procedures markedly affected (f < 0.0001) ability to culture Pasteurella spp. from bighorn sheep. Nonhemolytic f. haemolytica was isolated from 18/19 tonsillar swabs or biopsies from adult bighorns, but from only 4/19 nasal swabs. Nonhemolytic f. haemolytica was cultured from 14/19 tonsillar swabs plated directly, but from only 2/19 swabs stored for 24 hr in modified Amies with charcoal; f. haemolytica was not recovered from any of 19 swabs stored for 24 hr in modified Stuart's medium. Based on airect cultures of tonsillar samples, all 26 bighorns sampled were infected with f. haemolytica; 2 lambs developed pneumonia during this study. Tonsillar swabs or biopsies plated directly onto blood agar and incubated immediately offer the greatest probability of recovering nonhemolytic f. haemolytica from healthy bighorn sheep. Simulation modeling was used in preliminary evaluations of alternatives for managing disease cycles in bighorn populations. Epidemic disease caused simulated populations to display episodic increases and declines; disease-driven cycles repeated at 15- to 25-year intervals. Analysis of management scenarios, including several tactics for population and habitat manipulation, revealed that aggressive control of ewe numbers (e.g. harvest, trapping) was the most effective method for ensuring long-term stability and abundance in simulated bighorn populations. Simulations also indicated that short-term monitoring and/or comparison of asynchronous populations may be misleading in evaluating management practices. Modeling results suggested that relationships between density of bighorn populations and their susceptibility to disease, as well as effects of removal techniques (harvest, trapping) on sheep distribution, offer important unresolved questions in bighorn management. Environmental conditions may influence cortisol concentrations measured in bighorn feces. Fecal cortisol levels were affected by simulated environmental treatment (P ~ 0.0001) and exposure time (P ~ 0.0001); environment and time also interacted (P ~ 0.0001). For fecal samples held at 20 C, mean cortisol concentrations were about 2.7 times higher than controls held at -20 C after 2 days; a similar pattern emerged after cyclic freezing and thawing. Freezing at 20 C appeared to stabilize cortisol levels in bighorn feces for ~ 32 days. Influences of environmental conditions on measurable cortisol levels in bighorn feces could lead to misinterpretation of field data, and should be considered in applying these techniques to management studies of stress in bighorns. ��97 PLANE OF NUTRITION AND BIGHORN SHEEP POPULATION PERFORMANCE Michael W. Miller P. N. OBJECTIVE To treat bighorn sheep to control disease where necessary. SEGMENT OBJECTIVES 1. Develop a research strategy and proposal for managing viral diseases in mountain sheep populations. 2. Design population and management level experiments stress levels of mountain sheep populations. bacterial and to detect changes in Management of Bacterial and Viral Diseases in Mountain Sheep Populations Inability to control infectious disease outbreaks and subsequent mortality in mountain sheep populations represents a significant obstacle to long-term success in their management. Although the "bighorn pneumonia complex" has been studied intensively for over 3 decades, little is known about many aspects of its etiology and epizootiology. Moreover, management interventions recommended for preventing or controlling this problem remain untested. A research strategy and proposal for managing bacterial and viral diseases in bighorn populations will be developed over the next year as part of a statewide plan for bighorn management. However, improved and standardized methods for collecting and interpreting diagnostic data, along with innovative approaches for evaluating potential management tactics, appear prerequisite to developing and implementing such a research strategy. To this end, recent research has focused on improving tools available for use in management experiments that will be designed to study etiology, epizootiology, and prevention or control of disease outbreaks in bighorn populations: Epizootiology of Pasteurellosis (M. W. Miller and M. A. Wild) in Mountain Sheep Populations Pasteurellosis appears to be the primary component of most pneumonia outbreaks in bighorn populations. As part of our ongoing research efforts to improve understanding of the etiology and epizootiology of pasteurellosis in bighorns, we conducted and reported on a series of experiments evaluating techniques for detecting Pasteurella spp. infections in healthy bighorn sheep: Wild, M. A., and M. W. Miller. 1991. Detecting nonhemolytic Pasteurella haemolytica infections in healthy Rocky Mountain bighorn sheep (Ovis canadensis canadensis): influences of sample site and handling. J. Wildl. Dis. In Press. �98 ABSTRACT: Effects of sampling procedures on ability to culture Pasteurella spp. from Rocky Mountain bighorn sheep (Ovis canadensis canadensis) were examined experimentally. Sample site influenced (f < 0.0001) recovery of f. haemolytica in adult bighorn sheep. We isolated nonhemolytic f. haemolvtica from 18 of 19 tonsillar swabs and 18 of 19 tonsillar biopsies from adult sheep, yet only 4 of 19 nasal swabs yielded isolates. Sample handling also affected (f < 0.0001) recovery of f. haemolytica. Nonhemolytic f. haemolytica was cultured from 14 of 19 tonsillar swabs plated directly onto blood agar, but from only 2 of 19 swabs stored for 24 hr in modified Amies with charcoal; we failed to recover f. haemolytica from any of 19 swabs stored for 24 hr in modified Stuart's medium. We detected nonhemolytic f. haemolytica at least once in bronchial aspirates from 4 and in nasal swabs from 3 of 6 bighorn lambs. Based on direct cultures of tonsillar swabs and/or biopsies, all 26 bighorn sheep (7 lambs, 19 adults) sampled were infected with nonhemolytic f. haemolytica; only 2 lambs developed pneumonia during the study period. Thirty-four of 37 nonhemolytic f. haemolytica isolates tested were biotype T; 3 were biotype A. Serotypes 3; 4; 3,4 and 3,4,10 were identified in a subsample of 17 isolates. Our data suggest tonsillar swabs or biopsies plated directly onto blood agar and incubated immediately offer the greatest probability of recovering nonhemolytic f. haemolytica from healthy bighorn sheep. Managing Bighorn Population Cycles: A Simulation (M. W. Miller, N. T. Hobbs, and R. B. Gill) Approach As part of an adaptive environmental assessment exercise reported elsewhere (Hobbs et al., 1990), we developed a computer model that simulated disease and other natural processes influencing bighorn population performance. Using this model, we conducted simulation experiments to examine and compare management strategies for controlling disease outbreaks in a simulated bighorn herd. Results of these experiments will contribute to refined modeling approaches, which will be used in formulating research strategies and proposals for managing bacterial and viral diseases in bighorn populations. MATERIALS AND METHODS Simulation modeling was used in preliminary evaluations of alternatives for managing disease cycles in bighorn populations (see Hobbs et al., 1990 for details of the model and simulation experiments). This model was structured as a Leslie matrix with 2 sexes and 12 age classes, and was formulated such that natality and overwinter survival rates responded to changes in net primary production of forage and its availability. In addition, bighorn survival across all age classes was reduced by periodic disease epidemics, and parasitism also contributed to lamb mortality. Epidemic disease was assumed to be a stochastic processes mediated by animal density, and severity of parasitism was affected by both density and ewe condition. RESULTS AND DISCUSSION In the absence of epidemic disease, simulated bighorn populations showed exponential increase and asymptotic stability (Fig. 1); parasitism lowered the realized asymptote (Fig. 1). Epidemic disease caused simulated populations to display episodic increases and declines in bighorn abundance; these disease- �99 driven cycles repeated at 15- to 25-year intervals (Fig. 1). Analysis of alternative management scenarios, including several tactics for population and habitat management, revealed that population control (e.g. harvest, trapping) was the most effective method for ensuring long-term stability and abundance in bighorn populations. Although ram-only harvest preserved epizootic cycles, aggressive removal of both ewes and rams appeared to eliminate those cycles (Fig. 2). Simulations also indicated that short-term monitoring and/or comparison of asynchronous populations may be unreliable in evaluating efficacy of bighorn management practices. Modeling results suggest that relationships between density of bighorn populations and their susceptibility to disease, as well as effects of removal techniques (harvest, trapping) on sheep distribution, offer important unresolved questions in bighorn management. Detecting and Managing Stress in Mountain Sheep Populations Plans for designing and conducting management experiments to detect changes in stress levels of mountain sheep populations will proceed pending identification of suitable treatment and control herds for use in such studies. In the interim, we proceeded with development and evaluation of our techniques for detecting stress responses in bighorns: Effects of Environmental Variables on Cortisol Feces (H. A. Petit and M. W. Miller) MATERIALS Levels Measured in Bighorn AND METHODS We conducted an experiment to examine effects of simulated environmental variables (temperature, moisture, and UV light) over time on cortisol concentrations measured in feces from an ACTH-treated bighorn sheep. A pooled sample of bighorn feces (about 350 g) was equally divided into 93, 3.5 g aliquots. Aliquots were randomly assigned to 1 of 5 simulated environmental treatments (n - 18 aliquots/treatment): - constant freezing (-20 C). cyclic (12 hr) freezing/thawing (20 C/-20 C). constant room temperature (20 C). alternate-day soaking with deionized water (at 20 C). cyclic (12 hr) UV light exposure (at 20 C). Three additional aliquots were frozen immediately after collection and lyophilized for 24 hr to serve as absolute controls. On days 1, 2, 4, 8, 16, and 32 of treatment, we randomly chose 3 aliquots from each treatment group; these were frozen for 2 hr at -20 C, then lyophilized for 24 hr. All lyophilized samples were stored at -20 C until ground in a Wiley mill and prepared for fecal cortisol analysis as previously described (Miller et al. 1990). Cortisol was measured by radioimmunoassay, and cortisol concentrations were expressed as ng cortisol/g dm feces (Miller et al. 1990). We analyzed data by 2-way ANOVA (SAS, PROC GLM), with environmental treatment and time as main effects; we further examined effects of select environmental conditions using orthogonal contrasts. �100 RESULTS AND DISCUSSION Simulated environmental conditions had profound influences on measured fecal cortisol concentrations (Fig. 3). Preliminary analyses revealed that fecal cortisol levels were affected by environmental treatment (P ~ 0.0001), as well as exposure time (P ~ 0.0001); effects of environment and time were interactive (P ~ 0.0001). The most pronounced effects on measured cortisol occurred in fecal samples held at room temperature: within 2 days, mean cortisol concentrations from those samples were about 2.7 times higher than in frozen controls. A similar pattern emerged in samples exposed to cyclic freezing and thawing, although the changes were of smaller magnitude and peak response was delayed about 2 days. Freezing at -20 C appeared to stabilize cartisol levels in feces for at least 32 days. These results resemble those observed in an earlier experiment conducted before a reliable fecal cortisol assay had been developed (Miller and Hobbs, 1987). Patterns of change in measurable fecal cortisol levels observed here support our earlier hypothesis that bacterial and/or protozoal degradation of an undetected metabolite of cortisol, perhaps a sulfate or glucuronide, causes an increase in free cortisol levels in feces for 2-4 days after excretion. During this initial phase, the rate of converting cortisol metabolites to free cortisol may exceed degradation rates for free cortisol, whereas after 4-8 days cortisol degradation takes over. Because none of the samples were completely depleted of cortisol, we presume that bacterial and/or protozoal degradation of cortisol in excreted feces limited by time and is probably modified by temperature and moisture. We plan to conduct HPLCs on fecal extracts to gain further insights into the mechanisms driving this phenomenon. These assays should reveal presence of cortisol metabolites that may be contributing to observed elevations in measurable free cortisol. Ultimately, controlled incubation of fresh feces at room temperatures may provide a more sensitive estimate of total fecal cortisol excretion in stressed bighorns, thereby improving overall reliability of the assay. However, in the short term our results should serve to caution those intent on applying fecal cortisol data to solving bighorn management problems. The marked influences of environmental variation on measurable cortisol levels in bighorn feces could easily lead to misinterpretation of data collected under field conditions. Management decisions based on these data will be equally erroneous. LITERATURE CITED Hobbs, N. T., M. W. Miller, J. A. Bailey, D. A. Reed, and R. B. Gill. 1990. Biological criteria for introductions of large mammals: using simulation models to predict impact of competition. Trans. No. Amer. Wildl. and Nat. Res. Conf. (in press). Miller, M. W., and N. T. Hobbs. 1987. Bighorn sheep investigations: plane of nutrition and bighorn sheep population performances. Colo. Div. Wildl., Wildl. Res. Rep., Fed. Aid Proj. 01-03-048, WP2, J4, Job Prog. Rep., July 1986 - June 1987. Miller, M. W., N. T. Hobbs, and M. C. Sousa. 1990. Detecting stress responses in Rocky Mountain bighorn sheep (avis canadensis canadensis): reliability of cortisol concentrations in urine and feces. Can. J. Zool. In Press. �101 ~ild, M. A., and M. ~. Miller. 1991. Detecting nonhemolytic Pasteurella haemolytica infections in healthy Rocky Mountain bighorn sheep (Ovis canadensis canadensis): influences of sample site and handling. J. ~ildl. Dis. In Press. / I ,I Prepared by'. //"""/1 ,v,,#/7 ({I;' (/ K./ . ~ ,,/'~ ~ ~/ ...._./ --___ Michael ~. Miller ~ildlife Researcher ./ ��Colorado Division Wildlife Research July 1990 103 of wildlife Report JOB PROGRESS REPORT State of Colorado Project No. W-lS3-R-3 Work Plan No. ~2~A~ Job No. 6 Author: N. T. Hobbs Personnel: Mammals Research _ Mountain Sheep Investigations Adaptive Assessment Management-Simulations of CO-Management of Mountain Sheep and Mountain Goat Populations M. W. Miller, J. A. Bailey, D. F. Reed, and R. B. Gill ABSTRACT We constructed a simulation model to evaluate alternatives for' managing populations of Rocky Mountain bighorn sheep (Ovis canadensis canadensis) and Rocky Mountain goats (Oreamnos americanus) and to examine the potential for competition between them. Our model was structured as a Leslie matrix for two species with 2 sexes and 12 age classes. We formulated the model such that natality and overwinter survival rates of both species respond to changes in net primary production of forage and its availability. Bighorn nllIT~~r5were also regulated by effects of epidemic and endemic disease, as well as lamb mortality caused by parasitism. We assumed epidemic disease and parasitism are stochastic processes mediated by animal density. Simulated bighorn populations disFlayed episodic increases and declines in abundance caused by disease epidemics. Disease-driven cycles repeated at 15- to 20-year intervals. Epizootic cycles were preserved when ram-only harvest was added to the simulation, but aggressive harvest of both ewes and rams appeared to eliminate those cycles. Simulated mountain goat populations showed classical exponential increase and asymptotic stability following introduction into new habitat. In the absence of management, mountain goats virtually el~minated bighorns from alpine ranges. Aggressive harvest of both goats and'sheep allowed both species to maintain stable equilibria. Benefit/cost analysis of alternative management scenarios, including several tactics for population and habitat management, revealed that harvest may be the only-cost effective action influencing long-term abundance and distribution of bighorn popUlations. Adding mountain goats to ranges occupied by an established bighorn population did not enhance benefit/cost ratios. We conclude that mountain goats can be introduced to habitats occupied by bighorn sheep without harming bighorn populations, provided both species harvested aggressively. However, we failed to find any economic incentive for such introductions. We suggest that relationships between density of bighorn populations and their susceptibility to disease, as well as effects of harvest on sheep distribution, offer important unresolved questions in bighorn management. ��105 ADAPTIVE ASSESSMENT MANAGEMENT - SIMULATIONS OF CO-MANAGEMENT OF MOUNTAIN SHEEP AND MOUNTAIN GOAT POPULATIONS N. Thompson Hobbs P. N. OBJECTIVES 1. To develop management an adaptive assessment management model to simulate alternative strategies with mountain goat and mountain sheep populations. 2. To develop statewide guidelines mountain goat populations. for co-management 3. Implement through management guidelines of mountain the planning/budgeting sheep and process. INTRODUCTION High mountain landscapes offer unusual recreational opportunity to the people of Colorado. In particular, many people who enjoy the Colorado mountains also value the chance to observe and hunt alpine ungulates, including Rocky Mountain goats (Oreamnos americanus) and Rocky Mountain bighorn sheep (Ovis canadensis canadensis). It is the mission of the Division of Wildlife to preserve and enhance that opportunity. Mountain goats are probably not indigenous to Colorado, but were successfully introduced into several of the state's mountain ranges during 1948 to 1971. Currently, Division of Wildlife policy prevents further transplanting of mountain goats into Colorado because it is widely believed that they may compete with native populations of bighorn sheep for limited areas of suitable high elevation habitat. Bighorn sheep populations have performed poorly in many areas of Colorado, and hence it is wise to avoid actions that might further restrict in their distribution and abundance. However, policy discouraging introductions of mountain goats may unduly limit choices for managing alpine ungulates, and may thereby impair the Division of Wildlife's ability to offer diverse and plentiful recreation to the people of Colorado. It may be possible to successfully manage introduced populations of mountain goats such that they do little or no harm to bighorn sheep. In light of the extraordinary value of alpine ungulates to outdoor recreation, we believe that policy options for managing mountain goats, including a resumption of transplants, should be carefully considered. We analyzed policy alternatives for managing sympatric populations of mountain goats and bighorn sheep. The objectives of our analyses included 1) predicting the probable risks to bighorn populations resulting from reinitiating transplants of mountain goats, 2) developing adaptive management actions that would control those risks, and 3) suggesting research and monitoring that would future management capabilities. Here, we report this analysis. APPROACH We examined the consequences of introducing mountain goats into habitats occupied by bighorn sheep. Although this is clearly not the only policy �106 option for managing sheep and goats (for example, transplants could occur where sheep are not currently found), we believe it is ultimately che choice that must be scrutinized. This is the case because any introduction of mountain goats into Colorado, regardless of its location, increases the likelihood that dispersing, introduced mountain goats will compete with resident, indigenous bighorn sheep. Moreover, such introductions also may limit future locations for bighorn sheep transplants if the question of competition remains unresolved. Thus, although transplanting mountain goats to bighorn sheep ranges represents an extreme management tactic, it also offers the strongest test of the viability of introducing more mountain goats to the state. RESULTS The results of our analyses were reported in a paper presented at the North American Wildlife and Natural Resources Conference. The paper is appended. Prepared by N. Thompson Hobbs Wildlife Researcher �107 BIOLOGICAL USING CRITERIA SIMULATION N. Thompson FOR INTRODUCTIONS MODELS TO PREDICT OF LARGE MAMMALS: IMPACTS OF COMPETITION Hobbs and Michael W. Miller Colorado Division of Wildlife, Wildlife Research Center. 317 W. Prospect Road. Fort Collins, CO 80526 James A. Bailey Department of Fisherv and Wildlife Biology. Colorado State University. Fort Collins CO 80523 Dale F. Reed and R. Bruce Gill Colorado Division of Wildlife, Wildlife Research Center, 317 W. Prospect Road. Fort Collins, CO 80526 INTRODUCTION The current abundance of many important species of wildlife America can be traced to successful reintroduction. Outstanding elaphus nelsoni), programs examples ring-necked of introduction species that collectively recreational to the citizens of this continent. is clear that benefits environmental costs. value gained by transplanting in prospect, In retrospect, the harm done. gone bad, for example European common carp (Cyprinus carpio), and cheatgrass (Bromus tectorum). Remembering carefully examine the potential communities the There are brome our obligation impacts before adding a species of plants and animals. whether starlings (Sturnus vulgaris), these failures underscores it their it is often uncertain animals will outweigh many examples of introductions and rainbow provide enormous offered by these species have exceeded However, elk (Ce.rvus (Phasianus colchicus), trout (Salmo gairdneri), opportunity and/or include Rocky Mountain pheasants in ~~orth to to existing �108 Hobbs Here, we consider introducing animals particular, we describe impacts plan ungulates that information into habitats from which on competitive to influence our approach to a variety models of those In and to Although general, of potential species, interactions. is substantially of birds absent. to examine among 2 on the problem they have been interactions the outcomes as an example, application can be focused the use of simulation of introductions actions provide ways et al. -- we use and should and mammals. PROBLEM DESCRIPTION There are many Introductions of survival may improve of pests species believe resources limiting 1986). 1982, of animal population that adding animal populations, introductions, such actions we must and design while growth those risk. 1979, Connor and Simberloff the likelihood desirable between interventions by fauna may upset of local influences species that are introduced. the tradeoffs Many (reviewed that competition may harm introducing with habitat species to an existing and increase management minimizing 1977, Huston or contribute However, at equilibrium of individual the probability diversity, or competition. exist species introductions foresee increase community To the extent with to an ecosystem. does not occur without populations disadvantage to wildlife, predation communities via competition. competitive benefits through a new species enhance 1983; but also Wiens among extinctions animals, animal It follows equilibria access to a new habitat ecologists Schoener to introduce human of endangered to reduction an animal reasons growth that suffer a In planning benefits and costs to capitalize of on their their costs. EXAMPLE APPLICATION We recently used simulation modeling to evaluate and costs of trans locating mountain goats (Oreamnos Colorado, and to formulate criteria for deciding the biological americanus) when such benefits within translocations �1 ' .•. v, Hobbs would be appropriate. broadly useful Policv Conte::t. We believe approach in planning It is not certain Colorado. as recently disappearance (Hibbard coincided 1980). successfully Over and Taber introduced (Rideout indigenous with native for limited have performed 1987, Risenhoover et al. 1988), that might further discouraging restrict introductions alpine poorly Division goats because they were it seemed wise (Ovis to avoid and abundance. ~ountain (~akelyn actions However, limit choices to successfully has it was widely of Colorado goats may unduly goats goats were of Wildlife habitat. in many areas goats 1977). elevation It may be possible Rocky 1972, when sheep their distribution of mountain 1948-1971, of mountain high and hence, of mountain ungulates. populations populations of mountain (Feltner 1975,Denney of mountain to tundra habitats that mountain of the Colorado areas of suitable sheep populations a into the southern are reports 1900's until policy they compete illustrates and that their certain and Hoffman translocations to mountain There it is fairly additional introduced ranged glaciation, 1967). from the early the last 15 years, canadensis) managing goats ever large scale loss of alpine Regardless, from Colorado believed goats were as late as 1897, but these are questionable Rutherford prevented mountain that mountain with application introductions. as the Wisconsin 1958, Hoffman in Colorado absent whether Some hypothesize Mountains our specific et al. -- policy for manage such that they do little or no harm sheep. Objectives We analyzed mountain goats resulting alternatives and mountain their populations included: policy sheep using in an alpine 1) predicting a model ecosystem. the probable from reinitiating for managing risks trans locations sympatric simulating The objectives to mountain of mountain populations the dynamics of of of our analyses sheep populations goats, 2) developing �110 Hobbs translocation research Model guidelines and monitoring habitats policy enhance the consequences occupied option by mountain for managing occur where choice goats likelihood into Colorado, indigenous future risks, 4 and 2) ~i~ggesting management to mountain sheep capabilities. populations and mountain Mountains. we assumed that modeled types, and meadows. we chose location to enhance from different any introduction increases compete of establishing "worst grasslands the generality geographic goats goat conditions. in the central there were of it also of mountain traversed to simulate also may new mountain case" interactions this area with mountain tactic, of the if the question management populations within it was the transplanting to simulate on alpine intensively. us to draw on data although to represent not the only such introductions transplants an extreme into alpine transplants we believed goats will Moreover, Thus, km2 were used than an actual goat of its location, sheep we wanted goat populations cliffs (for example, test of the viability was constructed this is clearly mountain sheep. represents in the state. The model introduced unresolved. the strongest goats This was the case because for mountain ranges mountain found), regardless mountain locations remains offers Although sheep are not currently competition sheep Rocky 40 km2, of which 2 distinct a hypothetical 20 habitat area rather of our model, and to allow locations. Description Our model cohorts was structured of 2 populations modified at a yearly conditions. computer senior future sheep and goats that dispersing. resident, Model those of introducing sheep. that had to be examined. mountain limit that would minimize -- Context we examined could that would et al. code Model (Fig. 1). matrix Vectors time step in response behavior implementing author). as a Leslie was governed for 2 sexes for survivorship to simulated by several these assumptions and 12 annual and natality were environmental key assumptions. can be obtained by writing (The to the �111 Hobbs The mode.l represented alpine and subalpine habitats populations ~sed different frequently on steep cliffs 1981; Chadwick data), mountain 1971; D. F. Reed, assumption goats assumed spatial 1983; D. F. Reed, Colo. Div. wildl., in density, were not strongly sheep were competition with we further dispersal. assumed to steep terrain a pivotal as their Moreover, It distribution of 1983:83). Thus, (Chadwick in Colorado. seen in northern we assumed to lower elevations that competition reduced the number between we ranges that mountain to avoid mortality offspring For both produced in juveniles 1986, Smith 1986). declining food availability responded to random, than are sheep. annual variation season, which forage rates and mountain sheep in sufficient in spring, and elevated (Nichols because 1980, mountain sheep goats have (Adams and that goats are less sensitive Vegetation growth stochastically. in and undernutrition in precipitation also varied growth species, than those of mountain et al. 1984), we assumed and mountain harvest, goats and adults However, goats Population mortality, mountain status. that are more catholic 1983, Dailey mountain to obtain high quality of surviving and wishart natural simulated their nutritional rates of over-winter between of both populations. by natality, their ability lowered of the growing (Geist into meadows from predators, separation growth regulated amounts food habits unpubl. slope However, goats disperse most 1980; Schoen Div. wildl., data). using goats. Interactions that impaired reside but that sheep will not use cliffs. and did not migrate sheep acts to depress the model were goats of lower of ecological operative sedentary Colo. 5 that sheep and goat 1977; Thompson grasslands unpubl. of pressure is not limited mechanisms mountain -- goat populations we assumed (McFetridge sheep use adjacent increase mountain Bailey terrain: and ridges that in the absence Jorgenson 2,700 m. of the model was that mountain populations appears sheep and mountain above alpine and Kirchoff while mountain et al. as well to in the model as to duration Stocking rates of �11:2 Hobbs mountain sheep individuals and goats on alpine by controlling In addition ranges the amount to nutritional directly of forage effects et al. -- influenced available on recruitment, 6 nutrition of per capita. simulated mountain \ sheep populations assumed were that infectious cause precipitous survival (Demarchi the population population and/or compromise May and Anderson probability above a density herds through infections increases with individuals in the model increased linearly pneumonia epizootics in Rocky 1984, Andryk and Irby 1986, Bailey lungworm lamb mortality (Protostrongylus in mountain infections as ewe density increased and Samson mountain juveniles sheep, (Geist 1984, Festa-Bianchet declined 1987, Robb goat populations unpubl. data). We assumed classes 1-5) reduced sheep 1986, (Stelfox 1974, 1987, Samson that dispersal their survivorship (Hudson and Stelfox 1987, Samson Festa-Bianchet to disease and by juvenile than are by dispersal and Stevens at each with et a1. 1987). tend to be regulated 1983; Houston spp.) associated 1987, Robb are far less subject but their numbers 1982; Stevens rates sheep populations lungworm Mountain mortality et al. 1980, Wishart Lamb mortality Festa-Bianchet 1987). with density Mountain et a1. 1979). 1984, 1979, 1988). that parasitic et al. 1987) and ewe condition and May 1983, Mollison for epidemic increased individuals 1982, May 1972, Feuerstein of increasing (Anderson et a1. 1974, Schmidt Samson the proportion naive (Woodard 1976, Festa-Bianchet pneumonia parameters Festa-Bianchet can cause high assumed and that lamb 1982, Dietz and Wishart We also assumed following of immunologically (Demarchi We spp. can occasionally We derived North America 1986, years diseases. epidemics, We further in protected during et al. 1980, Onderka 1986). recruitment of a die-off documented from pneumonia to pasteurellosis of immunity and parasitic by Pasteurella for several 1972, Bailey threshold. throughout Schwantje resulting 1979, Anderson Thus, from those caused reduced susceptible density by infectious disease die-offs is drastically outbreaks regulated of 1988; J. A. Bailey, mountain time step. goats (age (We do not imply �113 Hobbs that all dispersing goats die, but the model mortality in ~he same way.) for males (Ste·v·ens1983), Harvest was used Achieved harvests between population determined function objective. objective Accuracy and precision by the level of resources in greater accuracy Our model was written execution Modeling on IBM compatible Environment model construction species invested than in the model. The difference population size harvested rate, which of population a~d of coho r t numbers. of animals success 7 for fe~ales objectives. of current The number and hunter 33% lower of both by population and estimates for dispersa: fraction populations influenced objectives of harvest controlled resulted were rates were and were a constant to regulate the harvest stochastically. Dispersal accounts et al. -- varied estimates in census. was a were Higher investment and precision. in FORTRAN 77 (Microsoft microcomputers. (Quaternary We used Software, version 4.01) the TIMEO Fort Collins, for Integrated CO) to facilitate and analyses. Model Analvses We used mountain the model to examine goats and mountain 1) What populations mechanisms 3) are most Do characteristics of mountain important the outcome Can harvest increase the likelihood To address sheep, of alpine habitats goats and mountain goat populations (e.g., of the other, and in absence of any management (i.e., food, disease, dispersal, primary interactions? between sheep? simulated we presumed on impacts? topography, of coexistence 1, we examined simulations, those of competitive question these initial of and what ecological in controlling influence in populations 3 questions: mountain production) mountain the absence of regulation sheep and to answer is the impact of introduced established 2) mechanisms dynamics all regulating parasitism) of each species interventions. factors were operative. except in In harvest �114 Hobbs We then structured mountain sheep and mountain and mountain goats of the simulation behavior simulation assumed added sheep populations, To address question of competition. habitat contained primary production. question trajectories parasitism) "conservative" minimal in census (> 1/2 curl), and either-sex included and ewe mountain RESULTS populations all any difference regulators of mountain in census, harvest of We first We of mountain variables levels on the the percentage of annual of harvest of net on (food, competition, two harvest a males-only in outcomes. the influence The conservative and either-sex species. runs, we altered We also examined activity, goats. of habitat and altered all natural investment portrayed of mountain simulated distinct patterns sheep and mountain mountain sheep populations The series of random for years 51-100. 1 used sheep, effects model harvest at year a the 2 halves as regulators in model both approach harvest goats. specified regimes, included a for mountain sheep The aggressive harvests of mountain of both ram goats. AND DISCUSSION The model goats, heavy of both and disease and meadows activated. Thus, of mountain growth differences and "aggressive." investment strategy regulated 3, we examined with between we reinitialized 51. to effects In separate introduced in the first and second half 2, we examined in cliffs To address of year of parasitism and examined outcome population alone that included comparison vs. 51-100 years), sheep populations effects sheep were To facilitate runs could be attributed subsequently simulations Mountain at the beginning that food supply disease, 50. (0-50 years of mountain of lOa-year goats. at year processes1 stochastic a series 8 et al. -- numbers used of growth goats. In the absence displayed for years in naturally-regulated episodes 1-50 was of mountain of exponential identical to that �UJ Hobbs increase follo~ed sequentially slow recovery (fig. 2A). approximately 20-year 3 females/km2. about mountain sheep Feuerstein absence These Similar throughout of mountain patterns introduced (Caughley goats was about 4.6 females/km2. (r = 0.13, simulated goats resemble those calculated When population depended included of mountain sheep, we formulated the two species mountain 1975, Stevens in simulations mechanisms shifted reached downward Density-dependent growth by reducing and Driver a stable effects natality of mountain 1986). sheep gr owcb . population 2C). (Fig. following on food supply and increasing et al. an established of mountai~ equilibrium goats Fig. 2A). and Wishart regulated sheep for mountain after their population alone in mountain of ungulates 1978, Youds representing regulated In the displayed populations performance of (r) for mountain 1980, Jorgensen so that food supply eventually point goats. population on which the model for growing the resulting data). (r = 0.11, sheep of 1976; density the first 10 years (Haas and Decker goats were strongly equilibrium sheep unpubl. (Fig. 2B) typical those for mountain (Hibbs et al. 1969, Vaughan 1980) and mountain and Stelfox rates of increase at in populations Equilibrium and peak densities goat populations 1970). During instantaneous Fig. 2B) exceeded These values observed (Hudson mountain by equilibrium into new habitat introduction, have been 9 s~asls, repeated attaining 1989; Colo. Div. Wildl., simulated followed temporary sequences with populations the Rocky Mountains sheep, growth decline, disease-induced intervals, et al. 1980; Hass exponential by precipitous et a l.. -- growth, That introduction acted winter When of to regulate mortality in both populations. When mountain effects were of food supply, added stable sheep populations to the simulation, or increasing Fig. 2D). performance parasitism, Thus, following of goats of sheep populations initially and epidemic the mountain for 27 years introduction were regulated disease, and mountain sheep population appeared goat introduction initially by increasing appeared stability by the combined (years to improve goats to be 50-77, the in their numbers. �116 Hobbs However, during epidemic. patterns during years in mountain epidemic disease. prevented occurred 50-77. increase importance the original mountain these conditions from that values sheep mountain onset competitive appeared net primary only slightly sheep numbers the rate of goats ultimately vigorously following emphasizes interactions the between these effects. to have 8 females/km2, in mountain of density-dependent from recovering little impact production (1,700 kgfha) , equilibrium responded natality rate retarded with mountain of time lagged When we increased reduced (Fig. 2D vs. Fig 2C) results in influencing from 4.5 to almost regulates densities by 1,000 kg/ha of mountain but population (Fig. 3A). on the outcome at levels well below goats trajectories This suggests over of that disease the food-based carrying of the environment. In contrast, and meadows changing substantially introduction of goats, sheep populations than changes area. populations because did not respond These patterns To the extent of habitat mountain to changes on probability can expand characteristics influenced responded and hence, the proportion and had moderate in production goat populations rates. did not recover in growth and delayed competition and the importance of competition. competition sheep population Comparing Environmental increased because This reduction However, of disease two species habitat ~y a disease was interrupted sheep population sheep density the mountain the die-off. density, mountain stability 10 (Fig. 2D). These capacity 77, this apparent The simulated die-off sheep year et al. -- thereafter in habitat of the effects depend translocation sites (Fig 3B). Simulated in initial penalty that these assumptions before Mountain extent of area on population on our assumption without by cliffs area to a greater outbreaks. to these changes into meadows of alpine sheep performance effects of disease contributed conditions that mountain in their are correct, in Colorado mountain for goat reproductive then landscape will have little or �1.L, -I Hobbs no impact on the eventual considerations outcome in selecting sites When we added hunting equilibrium propelled between failed allowed mountain to the model, investment to eliminate goats disease sexes of goats were harvested mountain goats can be potent management coexistence of these species (Fig. 3C) because their densities Based potent mountain below migratory release a threshold ability to thrive in meadow interactions populations of mountain habitats will depend after is a commitment investment in inventory It is clear controlling processes population goats extent to direct goat numbers occur over time scales numbers goats on demography difficult of mountain of to allow sustainable densities by maintaining goats can be depress non- mountain goat of the demonstrated and their apparent Preventing ability competitive for managing sites before they occur. goat trans locations regime of We within for both species and that regime. that a "wait and see" approach involves substantial that make changes to compare appears Characteristics on criteria harvest and in the absence in planning sheep. for mountain harvest outbreaks. ranges. than on choosing to an aggressive needed regime is the case because used by mountain criterion sheep that mountain 1983), strategy that although stabilized criteria long-term even when both and can significantly (Stevens to a greater sheep, for disease on alpine from our simulations mountain were unimportant introductions the fundamental Colorado critical This to achieve harvest at maximum we concluded to those ranges. with harvest ranges tr~,slocations. sheep populations, with mountain sheep, sheep Thus, we surmise sheep populations sheep populations dispersal surmise on alpine sites may be relatively translocations in competition 11 impcr:dGt Conservative in mountain an appropriate with mountain mountain (Fig. 3D). cycles competitors on these simulations, competitors an aggressive moderately. (Fig. 2D), applying mountain it was possible in census to prevail but remain for potential sheep and goats using by liberal regimes of competition et al. -- among peril. in mountain locations. sheep populations to Cyclic disease sheep Effects of mountain may be impossible to detect, �Hobbs particularly if sheep herds to "control" herds declining phase misleading. elsewhere. if those observations years!). sheep Such cycles The outcomes mountain sheep assumptions future contribute needed present. regardless on animal offer about (e.g., the future even 27 of mountain on our assumptions and responses. to any policy to hinge sympatric on preventing mountain between as well unresolved and motivate in managing recurrent disease goat populations density to disease, important Success of These recommendations, that uncertainty. that relationships distribution, goats have no effect long time periods security or to be sheep may be dangerous, rely heavily and their susceptibility to be in a static are likely that inferring performances appears are compared understood. scenarios of whether We suggest populations poorly to reduce happen in mountain spurious uncertainty sheep populations epidemics, suggests and goat population research mountain remain of these "controls" over relatively offer goat introductions such comparisons no change occur inferences if disease cycle, the model on observing receiving If those in the disease Moreover, on sheep based in areas 12 et al. are of mountain sheep as effects questions of harvest in mountain sheep management. Both efficacy of population strategies for regulating experiments incorporated experiments should achieving require managers, biologists, treatments. comprehensive ecosystems. design, ability plans of both census reliable conclusions, combined with such endeavors to manage disease sheep. and removal herd long-term commitment to apply and evaluate for a diversity and These methods responses such management are essential outbreaks in controlled for mountain and should monitor and researchers We believe for preventing should be evaluated into management objectives, To provide careful densities test effectiveness population management. control to experiments and cooperation will of management to enhancing of ungulate for our species in alpine �L9 Hobbs et al. -- 13 CONCLUSIONS Our example planning application introductions. initiating Models translocations. characteristics illustrates can be used Although of the habitat where management after mountain depend goats and mountain on the ability Models populations impacts of one species introductions. longer Models focus on it is clear to execute For example, specific co-existence in Colorado of will probably in future performance that relatively must be considered that brief, on effects in tor are planned, "surprises" illustrated suggests models them. local extinction for decisions species. long-term in criteria of mountain sheep empirical studies of introductions may be the only feasible of for that occurred .may offer a on competition way to make infcLences over time scales. Planning program introductions designed fundamentally substantial should in identifying difficulty experimental (Walters those criteria responses and Collie are met. as well Models as a can be in the target population, Our model comparisons underscores for success, demonstrated among populations the need when the they cycle for sophisticat2~ 1988, Walters et al. 1988) if those are to be reliable. A common that the best response to modeling efforts available information is simply is to say, we don't understand with habitats No doubt, criteria performance. in making This difficulty designs involve whether those of population out of phase. comparisons to evaluate useful particularly That on another of sympatry foundation between Our model usually ranges harvest in revealing The predicted after 27 years weak sheep on alpine are also useful criteria on the wherewithal are introduced. to effectively of interest. to evaluate trans locations should be contingent species uses of simulatinn such criteria that introductions actions several well enough to model like the one we have outlined wildlife not up to the task at hand. populations their behavior there are times when we should heed is or their interactions at any level of resolution. such doubts. To the extent that �120 Hobbs it is better appropriate However, to be ignorant ;"'hendata are scarce it can also be argued the potential risks than mislead, benefits involved. offered a modeling and processes that when approach are poorly a species prudence may demand of simulation modeling lies in forcing may ~ot be is that poor, cannot In such cases, 14- -- unde rs t oo d . our understanding by introducing et al. offset then the that we leave well enough alone. The utility recognize Bleloch the " ...consequences 1986:3). us communicate of what we believe This process why we made can improve us to explicitly to be true." the quality (Starfield of our choices and and help them. ACKNOWLEDGMENTS D. J. we thank M. Festa-Bianchet, Samuel, K. G. Smith, W. D. Snyder, reviewing helpful an early version comments editorial wildlife REFERENCES Adams, Freddy, assistance. Restoration central and for of this manuscript. N. McEwen by Colorado offeed p,o'lided Federal Ai.d to w-1S3-R-2. J. A. Bailey. J. Colorado. R. M. disease 1982. agents. Population wildl. Pages winter Manage. forages and control in R. M. Anderson of infectious of mountain goats diseases. of infectious and R. M. May, Dahlem eds. Konferenzen. Berlin. Part I. Nature 1979. in 47: 1237-1243. dynamics 149-176 R. M., and R. M. May. diseases: 1983. Transmission biology Springer-Verlag, Anderson, W. D. wishart W. M. R. C. 'Kufeld and C. E. Braun Our work was supported Project and T. M. Pojar, CITED L. G., Anderson, H. Olterman, D. A. welch, of our model. in their reviews J. Population 280:361-367. biology of infectious �, ), rz; Hobbs T. A., Andryk, and L. R. Irby. 1986. use by rncun t a i.n sheep prior Northe~n J. Bailey, sheep: Caughley, 1986. The increase biology, management Sheep and Goat Counc. G. 1970. Himalayan Chadwick, D. H. 1983. CA. Dailey, T. V., N. T. Hobbs, Denney, Proc. Dietz, K. 1972. Northern R. N. populations Ecology 1986. Canyon bighorn Symp. Northern with emphasis on 51:53-71. Sierra Club, San Competition, scientific method, and Am. Sci. 74:155-162. and T. N. Woodard. 1984. by mountain 1977. goats Experimental and mountain sheep in disease May, Population Konferenzen. G. 1972. Goat Symp. population agents. biology and Parks Division. sheep. goats in Colorado. 87-102 in the transmission dynamics in R. M. Anderson and R. M. diseases. Dahlem Game, of the Colorado Fish and Parks Div. Game, Denver, Fish CO. 64pp. Festa-Bianchet, ewes. M. 1987. Ph.D. Thesis, Individual Univ. of Berlin. A 75 year history Colorado bighorn 1:29-36. of infectious Springer-Verlag, A look back: of mountain patterns Pages of East Kootenay 2:22-28. and management Mountain Overall eds. recovery Wild Sheep Counc. Status Internatl. 1982. Post die-off infectious Feltner, S:;ir.p. J. Wildl. Manage. 48: 799-806. R. A. Trans. Bienn. Bienn. the color of winter. of diet selection Colorado. Demarchi, and habitat 208 pp. in ecology. comparisons 15 5:325-340. E. F., and D. Simberloff. null models of Waterton and dismanagement. of ungulate A beast die-off. -- 5:272-289. and die-off thar in New Zealand. Francisco, Connor, Eruption characteristi~s to a pneumonia Wild Sheep and Goat Counc. A. Wild Population et al. reproductive Calgary, Alberta. success 229 pp. of bighorn sheep �122 Hobbs Festa-Bianchet. M. comments Goat 1988. on preventive Counc. Festa-Bianchet, Northern Wild V. Geist, Sheep V. Mountain Press, 1982. goats. sheep herd. Chicago, Hibbard, C.W. Hibbs, Mich. R.S., Acad. reference Wild Hudson, predation, special Univ. Symp. Northern Wild 1988. Pleistocene 1969. Counc. bighorn 2: 143-167. and mammalian local The mountain Conf. and history dynamics 2: 100-112. faunas. Arctic tundra pp. 143- and alpine in mountain goats: 66: 228-238. Populations a system 34: 409-418. IN. limitation Can. J. Zool. goat in of holarctic (eds.). Bloomington, Resource 1976. Sheep Counc. introduced to their vertebrate Press. Rockies: and Proc. 43:3-32. Origin cropping. sheep 67: 699-705. reference R. J., and J. G. Stelfox. of the Canadian Univ. 3: 23-24. inbreeding, and Nat. Resour. 1967. D. B., and V. Stevens. sheep area, to mountain Sheep and Goat and D. L. Gilbert. N. Am. Wildl. a test by experimental 1980. and evolution. A study of a recently Sci. Arts Lett. Inidana SYTIp. Spec. Rep. 48. 19 pp. 170. In H.E. Wright, Jr., and W.H. Osburn environments. Bienn. Triangle and Goat Counc. of North American and R.D. Taber. with River-Almont Wild Sheep Can. J. Zool. Summary Trans. ecosystems, Houston, and to and W. H. Rutherford. with lamb mortality: L. D., F. A. Glover, Hoffmann, W'il.dSheep in relation sheep. in behavior control" Symp. Northern effects. Colorado. ·.d.th 11. 383 pp. 1980. Bighorn 1958. faunas. Northern bighorn Div. Wildl., a study Symp. Northern Bienn. 1989. population Colo. On "population Bienn. C. C. sheep, 4: 364-371. in the Taylor sheep: Haas, W. J., and E. Decker. Hass, Symp. Lamb survival C. P. Hibler, a case study. 1971. 1984. and Goat Counc. sheep mortality Chicago in bighorn Bienn. load in Rocky Mountain V., R. L. Schmidt, 1978-1979: Geist, management. M., and J. Samson. lungworm Bighorn epizootic 16 6:66-76. maternal Feuerstein, A pneumonia et a l . -- and diseases approach. of bighorn Proc. Bienn. �123 Hobbs Huston, M. 1979. A general hypothesis of species e tal. diversity. -- 17 Am. ~at. 113: 81-101. J. Jorgenson, T., and W. D. Wishart. population Bienn. May, R. M. responses of an expanding Symp. Northern 1983. 1986. Wild Parasitic Preliminary bighorn sheep herd Sheep and Goat Counc. infections observations as regulators on in Alberta. 5:348-365. of animal populations. Am. Sci. 71:36-45. May, R. M, and R. M. Anderson. diseases: McFetridge, Mollison, Part II. Nature 1977. Strategy R. J. groups. D. Proc. 1987. Soc. London Nichols, L. Onderka, of resource Risenhoover, use by mountain Mountain Goat Population dynamics of mammalian Aerial census Symp. goat nursery 1:169-173. and classification Symp. Northern in southern Wild Sheep 1984. Alberta. diseases. Symp. Zool. of mountain goats in and Goat Counc. A major Bienn. bighorn 2: 523-589. sheep Syrnp. Northern Wild die-off Sheep from and 4:356-363. C. B., and R. S. Hoffmann. Species of infectious 280:455-471. D. K., and W. D. Wishart. Goat Coune. Rideout, biology Internat1. Bienn. pneumonia Population 58:329-342. 1980. Alaska. 1979. 1975. Oreamnos americanus. Mammal. 63. 6pp. K. L., J. A. Bailey, Rocky Mountain bighorn and L. A. Wake1yn. sheep management 1988. problem. Addressing Wi1d1. the Soc. Bull. 16:346-352. Robb, L. A. 1987. Gastropod Protostrongy1idae) transmission. Rutherford, Samson, W.H. on a bighorn M.S. Thesis, 1980. with 1ungworms. Univ. Alberta, J. T. Jorgenson, infections hosts range: of 111 pp. 29(6):32-35. and W. D. Wishart. rocky mountain Dis. 23:396-403. (Nematoda: aspects Edmonton. Colo. Outdoors of free-ranging J. Wildl. of lungworm sheep winter Sheep or goats? J., J. C. Holmes, Experimental intermediate 1987. bighorn sheep �124 et al. Hobbs Schmidt, R. L., C. P. Hibler, evaluation Manage. J. Schoen, of drug treatment and W. H. Rutherford. of lungworm in bighorn 18 1979. An sheep .. ;. wildl. 43:461-467. W., and M. D. Kirchoff. southeast Alaska. and w-2l-l, Schoener, T. R. Spraker, -- 2. T. W. Alaska 1981. Habitat use by mountain Dep. Fish and Game. goats in P-R Proj. Rep. w-17-l0 67 pp. 1982. The controversy over interspecific competition. Am. on interspecific competition. Am. Sci. 70:586-594. Schoener, T. w. Nat. 1983. 122:240-285. Schwantje, H. M. 1986. southeastern Goat Smith, Starfield, study of bighorn Columbia. Bienn. Rates Alaska. and causes and wildlife G. J. wildl. Manage. management. 1974. regulation Symp. Northern of mortality A. M., and A. L. B1eloch. J. sheep herds wild in Sheep and 5:231-252. 1986. southeast A comparative British Counc. C. A. Stelfox, Field experiments Range theories. Building Macmillian. Proc. models New York, of bighorn Bienn. goats in 50: 743-746 1986. ecology in mountain sheep for co,1servation NY. 253 pp. in relation Symp. Northern wild to self- Sheep Counc. 1: 67-76. Stevens, V. 1983. population. Stevens, The dynamics Ph.D. Thesis, V., and C. Driver. goat herd Thompson, R. W. impacts 1980. Goat Counc. in an introduced washington, Initial Mountains. Seattle. observations Bienn. mountain goat 201 pp. on a tagged mountain Symp. Northern wild Sheep 1:165-174. Population and trapping Nest wilderness Univ. 1978. in the Olympic and Goat Counc. of dispersal Area, dynamics, of introduced Colorado. 2: 459-464. habitat utilization, Rocky Mountain Bienn. goats Symp. Northern recreational in the Eagles wild Sheep and �)1_J Hobbs Vaughan, M. R., 1975. Oregon. Wake1yn, M. S. Thesis, L. A. 1987. Colorado. Walters, Oregon Changing J. Wi1d1. to fisheries Fish. Aquat. J. A. Wa11ow~ Ll.? Corvallis. conditions on bighorn -- 19 ~ountains, pp . sheep range~ in 51:904-912. 1988. management? transient goat ecology, State Univ., habitat Manage. C. J., J. S. Collie, estimating Wiens, of mountain C. J., and J. S. Collie. useful Walters, Aspects et al. Is research Can J. Fish. Aquat. and T. Webb. responses on environmental 1988. Sci. 45:1848-1854. Experimental to management factors designs disturbances. for Can. J. Sci. 45:530-538. 1977. On competition and variable environments. Am. Sci. 65:590-597. Wishart, W. D., J. Jorgenson, bighorns from pneumonia Sheep and Goat Counc. Woodard, in southern Wildl. survival, Manage. Northern and W.H. Rutherford. and mortality W. K. Hall, data on mountain Wild Alberta. A minor die-off of Bienn. Symp. Northern 1974. Bighorn lamb in south-central Colorado. and R. A. Demarchi. 1980. J. 38(4):771-774. J. A., D. M. Hebert, Preliminary 1980. 2: 229-247. T.N., R.J. Gutierrez, production, Youds, and M. Hinton. goat population Sheep and Goat Counc. growth. 2: 482-519. Bienn. Symp. Wild �126 Hobbs FIGURE CAPTIONS Figure 1. deaths of mountain Age-structured disease, to the upper In both species, the number is controlled 2. alpine habitats sheep epidemic disease, mountain sheep (dashed dispersal mountain sheep regulated (dashed Effects of mountain sheep, assuming and 2,700 kg/ha). lines) epidemic and habitat sheep populations disease, are limited and parasitism. by food supply of mountain and of of mountain by food supply, are regulated habitats. performance by lines) Mountain A- of mountain in alpine and dispersal. of on simulated into alpine (dashed goats C - Effect performance management production of D - Effect are limited on population into by food supply goat populations introduced of net primary Mountain growth of mountain by food supply. Mountain introduction introduced performance goats are regulated populations two levels of mortality for food supply, growth sheep populations disease. of population goat of animals is controlled for food and dispersal. are regulated Mountain of ungulate Effect (solid B - Population (solid line) on population and dispersal. food supply, competition mountain line). by flow by epidemic A - Population line), where food supply trajectories populations (solid line) on population introduction and of winter adds to effects introduction sheep goats is controlled of management. competition and epidemic 3. sheep by intraspecific parasitism, Figure dispersal of ungulate and parasitism. and mountain goat availability in mountain 20 and harvest. trajectories by intraspecific goat Mortality in mountain in the absence mountain by per capita and harvest; starvation, Simulated regulated green-up. Mortality parasitism, one. -- births is indicated dispersal, Figure representing Natality food and date of spring on goat numbers. model goats. to the population by starvation, dynamics sheep and mountain from the lower box added population et al. habitats are limited goat populations B - Effect of (1,700 by �127 Hobbs mountain goat assuming different Numbers introduction by arr0"S that is contributed limited Effect arrangements (solid by meadows. of mountain of mountain sheep to "aggressive" epidemic lines) (thick z:1e0p, in alpine h ab i t.a t s , of the total area (cliffs sheep populations disease, and parasitism. are limited lines), and meadows by food supply (solid lines) assuming lines) and "conservative" Mountain lines) are goat and dispersal. c- performance are harvested (thin lines) 21 + meadows) (dashed on population populations -- of mountain Mountain goat introduction (dashed performance of cliffs give the proportion by food supply, populations on population et al. harvest according regimes. �H ,, ,, en .••. .----t Q) ;:I 'H Ii:.. ,- .... .. ,,, , , , , , ,, \ \ I N AT \ AGE = 1 I I I I , / I / / I " .. '" " ~ __ J"',_, ; ••••••• •• •• \ \ \ .. .. \ \ N AT - ;,, , ,., I ,. I , I HARVEST 00 N .-j , AGE == 2-12 \ .. , \ ,, �'"_-< rol 300 ----------------------- 7.5 N ------r - ~oo .-------- s. '" I (1) H ;:::l 1>0 'r-l rJc, - 6.0 A 6.0 B 200 (/) 200 ~ 4.5 W 4.5 (/) _J « :E , , , 100 W , LL 1 I I / --- \ '" / / ~ / / o ___ -'- __ J. - . I I 3.0 100 ,, ---- 1.5 30 0.0 o --10 0 50 40 0.0 =t:t: 50 40 30 20 W - >- N (/) 300 -------------------------- 7.5 0 « _J 200 /, .- f \ r- ••.• - "\ - •.•. , \ '\ f , 100 ,, - - ,, f / . 3.0 '- , a.. . - , /1 1.5 I . '" 1 100 - , I ,' I I \ - I I I 1 o L... _____ 0 L _____ 20 '_'-- 40 __ 0.0 100 _J 60 <:-r 4.5 ,, ~ I r n, D 200 ' ... / ::> o 6.0 C 80 YEARS AFTER o t- ~j7.S 300 Z I- « ~ W LL /- _J.__ __ W ...J 1.5 / ___t__ ___ ____ I , '" 1 / 20 10 0 / , ,' 3.0 , E sc - L ______ 0 20 I I \ ,, ,. ,' - " INTRODUCTION 6.0 0 4.5 Z 3.0 t- ' 1 I 1.5 ---,_- 0.0 60 80 0 -c ...J I 1 / 40 -, ffi 100 ::> a.. 0 a.. �130 3 Figure 400 ~--------------------------------~ 10.0 A 8.0 300 2700 6.0 200 1700 \ ~; , 100 ~ C/) W ..J < :E ,,4' I 0 L_ __ 0 4.0 1\, , -, 1.:- ••;", « " c \... , I ,,\.... ,,'/ 20 40 60 ~ - 0.6\ .r, ..'.. /' 200 N 0.3\':': I- -c ..J ::::l oa.. a.. 80 0.0 100 7.5 X /' u, 4.5 - 3.0 , 20 1.5 ---_- \-:.. -;. -:.. '- 60 80 0.0 100 . 7 5 300 ,..---------------------, C 6.0 CONSERV A TIVE 200 I 4.5 3.0 100 1.5 o AGRESSSIVE L- o -L 20 YEARS L-~ 40 AFTER __ ~ 60 ~ 80 !:: (/) .' ..J o ~----~------~~~~------~~~-o ~ >- " 40 -c 6.0 I -:. ~ C/) W ..J W B . E ~ I"\:' . ~ _;:"1 .~., / -, ".' i .. . , .... '/r ... , I, <, .7 '\ / , '\ , •• , .,' e ..... _,.".-, / / -~ .•-.; ,. -.". ,, 100 Z o- \ 0.3, 0.6, 0.9 0.9 W C/) 2.0 300 ,..-------------------, u, - CII I -I ___:;:~:::::::::=:J .Ls..:» __ W - _,- ,,_<. _J' ,/" ~~~ t____-=-=----1 _ /'- ~ 0.0 100 INTRODUCTION z w c z o l- -e ..J ::::l o, o a.. �l.3;' Colorado Division of Wildlife Wildlife Research Report July 1990 JOB PROGRESS REPORT State of Colorado Project No. ~W_-=1~5~3_-~R~-~3 Work Plan No. Job No. Period Covered: Author: _ Mammals Research 2A Mountain Sheep Investigations 7 Experimental Evaluation of Mountain Sheep Transplanting Disease Treatment and July 1, 1989 - June 30, 1990 M. W. Miller Personnel: N. T. Hobbs, R. B. Gill, J. H. Olterman, D. A. Reed E. E. Ryland, and ABSTRACT A prospectus was prepared outlining a management experiment to examine the effects of ewe harvest on movements and distribution of bighorn populations. Thirty-four Rocky Mountain bighorn sheep (19 ewes,S rams, 10 lambs) were captured from 2 sites in British Columbia, Canada, and transported to the Forbes Trinchera Ranch in southcentral Colorado to establish an experimental management herd. Translocated bighorns appeared healthy, although some individuals were infected with Pasteurella spp. and/or Protostrongylus spp.; no serologic evidence of exposure to respiratory viruses was detected. Movements of translocated sheep were largely confined to the vicinity of the release site for the first 2.5-3 mo after release; however, by late July individuals from the transplant herd had contacted bighorns from at least 3 other herds. Preliminary observat:i,.onssuggest potential for disease transmission'by translocated bighorns, as well as effects of dispersal on establishment of transplant h¢rds, warrant further investigation'through experimental management. ��133 EXPERIMENTAL EVALUATION OF MOUNTAIN SHEEP TRANSPLANTING AND DISEASE TREATMENT Michael W. Miller P. N. OBJECTIVES Design, conduct, and report on management experiments to evaluate efficacy transplanting and disease treatment practices for managing mountain sheep populations. AGREEMENT 1. Design management 2. Submit research proposals 3. Begin implementing of OBJECTIVES level experiments. for funding. funded research. Colorado's mountain sheep management program is among the most aggressive in North America. Although some combination of management practices appears to have produced a threefold increase in bighorn numbers statewide over the last 20 years, the contributions of individual management strategies to this overall success cannot be discerned (Bailey 1990). Transplanting bighorns to unoccupied ranges and treating resident herds to control parasites are 2 of the most intensive (and costly) practices used to manage sheep in Colorado, yet neither of these programs has ever been evaluated experimentally to assess long-term efficacy. Recent analyses of available herd data suggested that only about half of all transplant efforts succeed in establishing viable bighorn herds, and that parasite treatment does not necessarily prevent disease outbreaks or improve performance of treated bighorn herds (Bailey 1990). Unfortunately, management practices for most of Colorado's bighorn populations are so confounded that even interpretation of the aforementioned evaluation is equivocal. Clearly, management-level experiments are needed to evaluate and compare efficacy and efficiency of bighorn management practices in Colorado. The Colorado Division of Wildlife's Terrestrial Wildlife Research Section is committed to conducting management experiments to evaluate efficacy of transplanting and disease treatment practices for managing mountain sheep populations. Agency-wide, however, using this approach for managing sheep (or any other species) has met with mixed reception. Until a broad base of support for experimental management is established within CDOW, the ability to conduct population-level research will be severely limited. The following are highlights of our initial attempt to establish bighorn herds for use in management experiments: Research Prospectus: Effects of Ewe Harvest on Movements and Distribution of Bighorn Populations - M. W. Miller, N. T. Hobbs, R. B. Gill, J. H. Olterman, and E. E. Ryland We have been working to identify priorities for research on bighorn sheep during the coming decade. As part of that effort, we recently built a �134 computer model depicting the current state of knowledge about bighorn sheep population dynamics. The model serves 2 functions: first, it allows us to examine the outcomes of different management strategies assuming our knowledge of sheep populatior iynamics is correct; second, it highlights whece knowledge is inadequate. One of the central assumptions of the model is that sheep populations tend to increase until high levels of mortality cause sudden and dramatic declines in their numbers. These population "crashes" are caused by epidemic disease. In our model, we assume that population density controls the probability of disease outbreaks, and thus controls the cycles in bighorn abundance. Published accounts of real bighorn sheep population histories as well as management experience strongly support this assumption. When we incorporate this disease-density relationship into modeling exercises, our model realistically mimics the dynamics of many Colorado bighorn populations. Without this relationship, model output fails to resemble what we think we see in the real world. Our modeling efforts suggest 2 conclusions that are centrally important to future management of bighorn populations. First, if we wish to prevent cyclic episodes of peaks and crashes in bighorn numbers we must regulate densities below threshold levels through harvests and/or transplant removals. It is also clear that harvests or transplant removals comprised entirely of mature rams will not prevent episodic disease outbreaks. Removing ewes must be part of the management solution. Transplanting offers only a short-term means of regulating density below threshold levels because all potential transplant sites in Colorado will be occupied in the foreseeable future. Moreover, inaccessibility and/or conflicting land use policies preclude trapping and transplanting as an option for managing some bighorn herds. Thus, ewe harvest will almost certainly emerge as an increasingly important tool for bighorn management in the next decade. The second important outcome of the modeling exercise focuses on the relationship between hunting and animal movements. Our simulations suggested that the efficacy of harvest as means of preventing epidemic disease requires that animal numbers be reduced without reducing animal home ranges. That is, if ewe hunting acts to restrict the area used by bighorn populations, then it may worsen rather than solve the disease problem. Our modeling results motivate a specific research project, which we outline here. The effects of hunting on patterns of habitat use and dispersion by bighorn sheep are poorly understood. It may be that activities of hunters act to limit patterns of range use by bighorn populations by concentrating them in areas of escape cover. Whether this concentration occurs, and how long it may persist, remain unknown. It is also possible, however, that harvest may enhance distribution of bighorns by encouraging animals, particularly juveniles, to use unfamiliar terrain. We propose to answer these questions to enhance our future ability to manage density in bighorn populations. To do so, we wish to participate in a cooperative program to reintroduce bighorns to the Forbes Trinchera Ranch. Such participation will provide us with an unequalled experimental opportunity to resolve the uncertainties we outlined above. We propose that Rocky Mountain bighorns from British Columbia be introduced to the Forbes Trinchera Ranch to provide 2 independent herds, established under similar conditions, from similar stock, in similar, but spatially distinct habitats. Once both herds are well-established (in 3-5 years), we will conduct an experiment to investigate the effects of ewe harvest on their distribution and movements. Specifically, we will test the hypothesis that �, ~- 1..)) hunting ewes will reduce density of bighorn populations dispersal. by enhancing Our experiment would use a design know as "before and after controlled intervention" (BACI) to determine sheep responses to harvest activities. In essence, this design involves observing the responses of both populations before and after imposing a limited ewe season on 1 of the 2 populations. Thus, the "un t rea t ed" population (exposed to limited ram-only ha rve s t ) serves as a control for any serial effects of time that may occur during the study. Operationally, we foresee using radio telemetry to observe patterns of habitat use by bighorns in the 2 Forbes Trinchera herds. We would monitor distribution and movements of adult ewes in both herds for 2-3 years under a limited ram-only harvest system. We would then monitor ewe responses after imposing a ewe hunting season on 1 of those herds during each of 3 subsequent years. A detailed agreement on season lengths and dates, permit numbers, a process for permit distribution, and other particulars will be reached through negotiations among Terrestrial Wildlife Resources, the Southwest Region, and Forbes Trinchera Ranch during a planning phase and approved by the Wildlife Commission before initiating this study. This project represents a unique chance to learn about the consequences of management activities and provide added recreational opportunities for Colorado's bighorn hunters without impacting established bighorn herds and their management programs. Our study will also complement other existing and planned bighorn research and management activities. Moreover, its results should aid in guiding decisions on alternatives for future bighorn sheep management in Colorado. Forbes Trinchera Bighorn Translocation MATERIALS AND METHODS Rocky Mountain bighorn sheep were translocated from British Columbia to Colorado in order to establish 1 (or more) bighorn herd(s) for experimental management purposes. Bighorns from Columbia Lake and Stoddard Creek, British Columbia, Canada, were captured using a corral trap or drop net on 16-18 March 1990. We collected blood from and examined each animal included for translocation, and injected each with ivermectin and long-acting oxytetracycline (LA 200); we also collected pharyngeal swabs from 15 ewes. Feces from adults and lambs were pooled and examined for Protostrongylus spp. larvae. Modified horse trailers were used to transport a total of 34 bighorns to the Forbes Trinchera Ranch near Ft. Garland, Colorado. On 18 and 20 March 1990, 2 groups of bighorns were released on the ranch in the North Fork of Trinchera Creek vicinity; altogether, 19 ewes,S rams, and 10 lambs were introduced (Table 1) at the same location. All sheep were marked with blue eartags (numbered 1-28 and 41-46). In addition, we fitted 17 adults (12 ewes,S rams) with radio transmitters (frequency range 172-173), and marked the remaining adult ewes with blue visual collars (numbered 1-4 and 9-11). Forbes Trinchera Ranch personnel have monitored sheep movements from the ground weekly, and all radioco1lared sheep have been relocated from the air twice since the release. �136 RESULTS AND DISCUSSION All bighorns appeared healthy at capture. No titers to PI3, BRSV and IBR were detected in 21 samples submitted for serology. Nonhemolytic Pasteurella haemolytic~ was recovered from at least 4 of 15 ewes. All fecal samples examined contained Protostrongylus spp. larvae; counts averaged about 200 11/g feces [mean(se) counts for adults: 244(284) 11/6 feces; for lambs: 155(164) ll/g feces]. All sheep survived capture and transport, and thus far no mortalities have been detected. To date, all 17 radios have been relocated, although transmission range appears severely impaired on 4 collars. Most of the remaining individuals have been identified at least once during ground observations made since the release. During the first 2.5-3 mo after release all bighorns remained on the ranch, although some dispersal from (and return to) the release site occurred. (No~e: By late July, translocated bighorns had scattered widely, and individual sheep had contacted at least 3 other established herds occupying nearby ranges (Trinchera Peaks, Spanish Peaks, Mount Blanca). Observations made in late August indicate some of these individuals may be returning to the original ~~Iease site, or to other locations on the ranch.) Preparat: _= a draft study plan outlining short-term objectives and initial monitoring needs for this release has been delayed, and will not be completed until internal problems with a cooperative agreement outlining sectional responsibilities for research and management of this transplant herd are resolved. In the interim, short-term monitoring by ranch personnel should provide useful (although limited) data on initial performance and movements of translocated sheep. Results of disease screening support other recent research (see WP2A,J4) suggesting Pasteurella spp. can be carried by healthy bighorn sheep, and that potential for disease transmission should be a consideration in planning bighorn translocations. (In addition, preliminary observations suggest dispersal of translocated bighorns and joining of these with herds established nearby may account for some apparent failures of smaller bighorn transplants in Colorado.) LITERATURE CITED Bailey, J. A. Colorado. Prepared by: 1990. Management of Rocky Mountain bighorn sheep herds in Colo. Div. Wildl. Spec. Rep. No. 66, Ft. Collins, In Press. Lb~ Michael W. Miller Wildlife Researcher �. .., - 1.) I EAR TAG COLLAR (COLORS!ID)' FREQ2 SEX AGE 3 BLUE 1 R: ORANGE/BLACK .2625 F 4.8 2 R: BLACK/I.IHITE .2375 F 7.8 3 R: BLUE .3625 F 5.8 4 --- --- M 0.8 F 0.8 5 6 R: YELLOIJ .3875 F 4.8 7 R: ORANGE .4375 F 1.8 8 V: BLUE 1 -- F 3.8 9 R: ORANGE/I.IHITE .2875 F 2.8 10 R: RED "O"/I.IHITE .7375 F 7.8 11 R: BLUE/BLACK .538 M 2.8 12 R: RED "9"/I.IHlTE .7125 F 4.8 13 V: BLUE 2 -- F 5.8 14 V: BLUE 3 F 3.8 F 0.8 .6875 F 5.8 -- F 0.8 .462 M 2.8 -- M 0.8 3.0125 F 4.8 .409 M 4.8 --- -- 15 16 R: RED "8"/I.IHITE -- 17 18 R: I.IHITE/BROIJN 19 -- 20 R: BLACK 21 R: GREEN/BLACK "M"/I.IHITE 22 .- -- F 0.8 23 R: REO .314 M 2.8 24 R: YELLOIJ/RED .611 M 1.8 25 R: BLACK .9125 F 7.8 26 V: BLUE 4 -- F 1.8 27 R: BLACK .8875 F 3.8 28 -- -- M 0.8 F 2.8 F 5.8 F 0.8 M 0.8 F 7.8 F 0.8 "F" /I.IH ITE "E" /I.IH ITE 41 V: BLUE 9 42 V: BLUE 10 ------- --- 43 44 45 V: BLUE 11 -- 46 1 R-RADIO; V-VISUAL. 2 FREQ-l72. 3 ESTIMATED BY HORN ANNULI; ASSUMED MAY BIRTH DATES. ��139 Colorado Division of Wildlife Wildlife Research Report July 1990 JOB PROGRESS REPORT State of ~C~o~l~o~r~a~d~o~_ Project No. W-1S3-R-3 Work Plan No. Job No. 2 Period Covered: Author: ~3~A~ July 1, 1989 Mammals _ 2 Research Pron~horn Investigations Habitat Selection and Population Performance of a Pioneering Pronghorn Population June 30, 1990 T. M. Pojar ABSTRACT The population growth and distribution of the Middle Park pronghorn herd is monitored with the use of radio-collared animals. Four years of radio location data are available. Locations of radioed animals biweekly permits calculation of areas of habitation during winter (represented by January locations) and during the time of maximal dispersal (represented by June locations). The distribution and areas of habitation are presented for yearlong data for all 4 years. Growth patterns of this herd have mimicked the early stages of the theoretical sigmoid growth curve. The rate of increase for the past 3 years provides a minimally meager data set upon which to estimate the K-value for this population. ��l4l HABITAT SELECTION AND POPULATION PERFORMANCE OF A PIONEERING PRONGHORN POPULATION Thomas M. Pojar P. N. OBJECTIVE Describe population dynamics pronghorn population. and habitat use of a pioneering, expanding SEGMENT OBJECTIVES 1. Describe seasonal population. and annual distribution of the Middle Park pronghorn 2 Determine necessary 3. Monitor population dynamics of Middle Park pronghorns with: a. Ground counts to describe changes in population size. b. Ground counts to quantify population sex and age composition. sample sizes of radio-collared animals and observations to describe habitat preferences. STUDY AREA The study area is described in Pojar (1988:183-184). METHODS AND MATERIALS Seasonal and Annual Distribution The radioed pronghorns were located biweekly (with very few exceptions) since 1 January 1987. Tracking was done mostly from the ground to increase the probability of observing and identifying animals with numbered plastic collars. Fixed-wing aircraft was used if an animal could not be located after reasonable effort from the ground. Legal descriptions of animal locations were recorded to the nearest quarter mile then converted to UTM (U.S. Army 1973) coordinates for computer processing. All areas of habitation are based on minimum convex polygons, harmonic mean contours, and core areas as calculated by the program HOME (Ackerman et al. 1989). The key map used was a U.S. Department of Agriculture Forest Service Routt National Forest, Colorado, Sixth Principal Meridian, 1975 (reprint 1985). The original scale of 1:126720 was reduced to 1:230400 via photographic techniques at the Colorado State University Photo Lab. The areas plotted by the program HOME are automatically scaled according to the number of points and the area involved. Therefore, the key map (Figure 1) is to be used for general orientation and prominent geophysical features are included on the computer plots. January and June were selected to represent m~n~mum and maximum dispersion, respectively; and the plots include locations of all radioed animals. In �142 addition, plots of yearlong areas of habitation, by year, were made. Each plot has the 100, 95, and 75% minimum convex polygons and the 95, 75, and 50% harmonic mean contours. The 50% harmonic mean contour is considered the "core" area because it contains approximately 50% of the volume of observations (Tables 1, 2, and 3). Population Size and Structure During late summer and fall all animals accompanying the radioed pronghorns were counted and classified by sex and age. Herd structure estimates used in the following analysis were based on the classification with the largest sample size. The exception to this is the herd structure estimate for the fall of 1986, which was based on the sample of 47 animals trapped on 16 December, 1986. The population size estimates were made during winter when the population was concentrated on the wintering area. Counts were made from the ground using a 20X spotting scope and/or lOX binoculars. It was assumed that a count of all animals accompanying the radioed animals at this time of year yielded a total count of the population. It was possible to count the number of bucks that were 1 1/2 years old and older as a check on the earlier herd structure estimate. RESULTS Seasonal and Annual Distribution The plots generated for January and locations of 7 to 9 radioed animals, 1988. After the trapping operation in operation. Therefore, all plots 1990 are the result of locations of June of 1987 and 1988 are based on the as are the yearlong plots for 1987 and in December of 1988, there were 24 radios (January, June, and yearlong) for 1989 and 19 to 24 radioed animals. Direct comparison between areas of habitation for 1987-1988 and 1989-1990 is not possible because of the difference in number of radioed animals. Comparisons between 1987 and 1988, and 1989 and 1990 are legitimate because of a similar number of radios during each of these years. Although, the "yearlong" area for 1990 includes only January through June, the total area very likely encompasses the maximum distribution for the year. Late summer and autumn movements usually are simply a return to the winter/spring areas. Weather and snow depth influence the distribution of animals during winter. The winters of 1986-87 and 1989-90 were mild with light snow accumulation. The winter of 1988-89 was most severe with deep snow accumulation which limited pronghorn movement. The winter of 1987-88 could be classified as moderately severe. The areas (in km2) for the plot of the January distribution of animals (Figures 2 and 3) is found in Table 1, the areas for June distribution (Figures 4 and 5) in Table 2, and the areas for yearlong distribution (Figures 6 and 7) in Table 3. For proper orientation, the reader is referred to the key map (Figure 1) and the legend for the geophysical codes (Table 4). If this population is to survive in Middle Park, it seems that expansion of the wintering area will be seen as the population size increases. The large �l4 J increase in January area of habitation seen between 1989 and 1990 (Table 1) is most likely an artifact of the extremely mild winter conditions in 1990. The movement of the animals was not impeded by snow accumulation as it was in 1989. Continued monitoring of this population through varied climatic conditions will expose trends in the areas of habitation as the population increases. Population Size and Structure Early growth patterns of the MP pronghorn population have mimicked the early stages of the theoretical sigmoid growth curve (Figure 8). It took from 1979 until 1984 for the population to double the first time. It doubled again over the next 2 years. Population estimates since deployment of radio transmitter collars (1986-87) are considered to be accurate and have been used to calculate the population's rate of increase. The rate of increase (N2-N1/N1) for the past 3 years provides a minimally meager data set upon which to estimate a regression (Caughley 1977) on population size and project a K-value for this population (Figure 9). Figure 9 is based on the following observed values: Year Pop. Size Rate of Increase 1986-87 1987-88 1988-89 1989-90 80 122 160 223 .52 .31 .39 Using the regression equation from Figure 9, an estimated K-value for the population is 624 animals. This is key information for formulating long range management plans but must be used with caution since the regression is based on only 3 points. There are 2 major situations that could drastically alter the above relationship. It is possible that as a wider range of climatic (i.e. winter) conditions are manifest in the data set, the populations' response to density may be overshadowed by its response to climatic conditions. In this case, the rate of increase relationship with density would become scrambled and would not provide us with a good projection for the K-value. At this point, management objectives would have to be tied to some arbitrary decision regarding projected population density potential. A second contingency that could affect the rate of increase and, therefore, the K-value projection, is continued pioneering of the population into other parts of MP. Fragmenting of the wintering herd could release much more potential for growth than current projections indicate. Due to the gregarious nature of this species, it was possible to classify onehalf to three-fourth of the population by classifying all animals that accompanied radioed animals. Even with only 5.7% of the population radioed (1987), the herd structure sample was 52% of the population (Table 5). �144 LITERATURE CITED Ackerman, B. B F. A. Leban, M. D. Samuel, and E. O. Garton. 1989. User's manual for program HOME RANGE. Forest, Wildl. and Range Experiment Sta., Tech. Rep. 15. University of Idaho, Moscow. 79pp. Caughley, G. 1977. Analysis of Vertebrate Populations. John Wiley & Sons, NY. 234 pp. Pojar, T. M. 1988. Habitat selection pioneering pronghorn population. 181-192. and population performance of a Colo. Div. Wildl. Res. Rep. July, pp U.S. Army. 1973. Technical Manual: Universal transverse mercator grid. Headquarters, Dep. of the Army, Washington D.C. No. 5-241-8, 64 pp. ™ Prepared b~af;Jrfw Wildlife Researcher �145 Table month 1. Areas (in K(2) of habitation of January. MCP-Minimum Convex 95% MCP for Middle Park pronghorns Polygon. HM=Harmonic Mean. 75% MCP 95% HM during the 50% HM 75% HM Year 100% Mep 1987 3.2 3.2 1.1 3.9 3.2 0.24 1988 6.9 6.9 2.3 10.1 4.5 3.0 1989 3.4 2.7 2.7 6.3 3.4 1.3 1990 52.1 44.1 8.2 44.5 26.5 15.8 Table month for Middle Park pronghorns 2. Areas (in K(2) of habitation Mean. MCP=Minimum Convex Polygon. HM=Harmonic of June. during the 95% HM 75% HM 41.0 225.6 159.6 82.1 149.4 36.1 192.6 118.6 50.7 507.6 369.6 225.3 602.7 344.4 194.9 451. 5 438.1 173.1 757.2 425.3 253.4 Year 100% MCP 95% MCP 1987 170.1 150.7 1988 149.4 1989 1990 75% MCP Table 3. Areas (in K(2) of habitation for Middle MCP-Minimum Convex Polygon. HM-Harmonic Mean. Year 100% MCP 95% MCP 1987 204.6 165.6 1988 217.7 174.2 1989 603.8 522.2 1990 471. 6 331.6 Park pronghorns 50% HM yearlong. 95% HM 75% HM 50% HM 54.6 561.4 236.0 154.1 31. 5 458.2 186.5 121. 9 192.8 1554. 665.3 417.6 133.2 1007. 443.9 304.5 75% MCP �146 Table 4. Description and UTM coordinates of geophysical reference points the key map (Figure 1) and the computer plots of pronghorn distribution. REFERENCE UTM Ycord for CODE Xcord Kremmling K 381500 4435000 Town Wolford Mt. WM 381500 4442500 4.5 mi. N of K Gunsite Pass GSP 387000 4451000 12 mi. NNE of K Whitley Peak WP 373000 4464000 20 mi. NNW of K Parshall P 400000 4434500 Town Red Mountain RM 385500 4434500 2.5 mi. E of K Gravel GP 386500 4438500 4 mi. NE of K JH 389500 4437200 6 mi. ENE of K POINT Pit Jody Hill Ranch Table 5. Herd structure YEAR POP. SIZE 19861 80 1987 122 7 1988 160 1989 223 lThis year's 16 December 1986. NO. RADIO based on sample obtained X DESCRIPTION by locating radioed animals. B:D RATIO F:D RATIO SAMPLE X OF 36 77 47 59 5.7 54 77 63 52 24 15.0 40 32 108 68 22 10.2 56 50 161 72 RADIO data based on the sample of the population trapped POP. �147 Figure 1. Key map of Middle Park with geophysical features designated to assist in orientation and perspective of the animal distribution plots by the program HOME, Figures 2-7. Also see Table 4 for codes. �148 MIDDLE PARK PRONGHORN - JANUARY 1987 4442278 JH 4437588 3869313 nETERS 389948 nIDDLE PARK PROHGHORN - JANUARV 1988 4442288 ..•... tJ. JH 4434358 3839813 nETERS 39888B Figure 2. Plots from program HOME for pronghorn distribution during January 1987 and January 1988. See Table 4 for key to codes of geophysical features on plot. �MIDDLE PARK PRONGHORN - JAHUARY 1989 44481813 tJo GP JH tJo 44366513 3894113 "ETERS 3868113 "IDDLE PARK PRONGHORN - JANUARY 19913 444741BO~--------~=---------------------------------------' ...... =¥' "" .... .... ..... tJoJH 44337613 381351313 Figure 3. Plots from program HOME for pronghorn distribution during January 1989 and January 1990. See Table 4 for key to codes of geophysical features on plot. 3914613 �150 MIDDLE PARK PRONGHORN - JUNE 1987 4457810 .+.. t. GSP t. K 4429430 tlETERS 37696B HIDDLE PARK PRONGHORN 397338 - JUNE 1988 4452188 /). GSP 4431668 37985B tlETERS 399928 Figure 4. Plots from program HOME for pronghorn distribution during June 1987 and June 1988. See Table 4 for key to codes of geophysical features on plot. �151 447341B MIDDLE PARK PRONGHORN - JUNE 1989 + .. t, p 442884B 365998 "ETERS 482888 "IDDLE PARK PRONGHORN - JUME 1998 4469858 .... t,p 442686B 365118 "ETERS 48386B Figure 5. Plots from program HOME for pronghorn distribution during June 1989 and June 1990. See Table 4 for key to codes of geophysical features on plot. �152 MIDDLE PARK PROHGHORN - VEARLOHG 1987 4457818 -+- WM +++ .. ..-++ +-+i- +++ .... GP#+ IJH~.··. + .. ++-1+ .:-5- -+-. + -+ ++- I 1 I , .6+ +..,... .-++ r -+ + t::.-.~+ + .. .++ 4426290 375480 tlETERS tllDDLE PARK PHOHGHORN 488460 - YEARLOHG 1988 4454450 + .. Gt· ...t++.. +++ -*+ +.6-44 ... + +~~+..~ :tt ++.. . . .... t::.++'+ RM 4426270 376020 tlETERS 488690 Figure 6. Plots from program HOME for pronghorn distribution yearlong 1987 and yearlong 1988. See Table 4 for key to codes of geophysical features on plot. �153 MIDDLE PARK PROHGHORN - ~EARLOHG 1989 4476818 ++ F ~.... "". ~, '~ .... .. /).p Rl"i 4419620 tlETERS 365088 tlIDDLE PARK PROHGHOHN 4469810 487580 - THRU JUHE 1998 -ru~""'" ~" .... ~p 4422810 366428 tlETERS 484540 Figure 7. Plots from program HOME for pronghorn distribution yearlong 1989 and January through June 1990. See Table 4 for key to codes of geophysical features on plot. �154 300 PREDICTED - .••••• 250 200 z c: CIJ r ~ 0.998 0 :::I: (!J z 150 0 c: a.. u.. 0 c: 100 w CD ~ ::::J Z 50 o 1974 1976 1978 1980 1982 1984 1986 1988 1990 Figure 8. Growth pattern of the Middle Park pronghorn population from 1974 thru 1990. �1)5 1 0.8 y=0.555-0.00089x Q) ~ 0.6 ~ (.J c:: '0 Q) 1ti 0.4 a: K 0.2 100 = 624 • 200 700 Population Size Figure 9. Least squares regression of observed rate of increase for the Middle Park pronghorn population, 1987-89. ��Colorado Division Wildlife Research July 1990 157 of Wildlife Report JOB PROGRESS State of ~C~o~l~o=r~a~d~o~ _ Project No. Mammals W-lS3-R-3 Work Plan No. ~3~A~ Job No. _ 4 Period Covered: Author: REPORT Research Pronghorn Investigations Statewide Pronghorn Management Plan July 1, 1988 - June 30, 1990 T. M. Pojar ABSTRACT Some very preliminary efforts were made toward development of a statewide pronghorn management plan. These efforts involved meetings with the Mammals 2 Research Leader and appropriate State Wildlife Managers for direction and format of the management plan. A memo as drafted requesting information from Regional Wildlife Biologists on the current status, biological potential, and biopolitical constraints on pronghorn populations' from all regions of the state. Further progress on this project is contingent on organizational commitment and direction from top administrators. Prepared bYY/k;;J ~ (})~~ Thomas M. P jar Wildlife Researcher ��159 Colorado Division of Wildlife Wildlife Research Report July 1990 JOB PROGRESS REPORT State of Colorado Project No. W-IS3-R-3 Mammals Research Work Plan No. 6A Mountain Lion Investigations Job No. 1 Mountain Lion Population Period Covered: Author: Dynamics July 1, 1989 - June 30, 1990 A. E. Anderson Personnel: See acknowledgments Abstract Progress on the manuscript, "The puma on Uncompahgre Plateau, Colorado", is briefly summarized. Statistical analyses and preparation of figures are nearly complete. Work has begun on writing the first draft. ��161 MOUNTAIN LION POPULATION DYNAMICS Allen E. Anderson P. N. OBJECTIVE To assess the effects of sport hunting on mountain lion populations. SEGMENT OBJECTIVE To prepare final reports for publication. ACKNOWLEDGMENTS Most statistical analyses were performed by Dr. D. C. Bowden, Department of Statistics, Colorado State University, Fort Collins. N. McEwen produced or supervised CSU Graphics Department personnel in the production of most of the figures mentioned. D. Masden, Inventory Biologist, Southwest Region, assisted with mapping relative average densities of mule deer and elk on winter ranges of Game Management Units 61 and 62. M. Roelke and her colleagues from the Florida Panther Biomedical Investigation and the National Institute of Health kindly provided their information on Uncompahgre Plateau puma hematology, genetics and spermatozoa. METHODS AND MATERIALS Between April 16, 1981, and April IS, 1988, 57 puma were captured and 49 were fitted with radio collars and located from the air at 6-8 day intervals using the Universal Transverse Mercator (UTM) system. General methodology was described in Anderson (1983). Following a computer-aided validation of 3,117 aerial locations, we used the HOME RANGE personal computer program (Ackerman et al. 1989) to test the assumption of the independence of aerial locations, select 2 specific estimator(s) of home range size and utilization, and calculate home range size and utilization with those estimators. Additional programs and graphs were written in SAS. These included graphic exploration of (1) overlap of home range where the proportion of overlap was measured with an electronic planimeter and (2) seasonal shifts in home ranges by individual puma using the total number of locations. A nongraphic analysis of the spatial patterns of puma locations was made with the multi-response permutation procedure (Mielke and Berry 1982) which tests for temporal differences in the distribution of locations. Differences in mean monthly elevations of the locations were compared graphically with 95% confidence intervals of individual puma about the means. Published estimates of survival in our sample of Uncompahgre Plateau puma (Anderson et al. 1989) were being elaborated as the fiscal year ended. An index to puma activity was obtained by calculating mean distances consecutive distances over 6-8 day intervals by sex and season. between �162 Possible interactions between puma pairs was indicated by the mean (± SD) distance between their near-simultaneous aerial locations. Body weights and measurements were described statistically and linear mUltiple regression equations developed to predict body weight from body measurements for each sex. For comparative purposes, blood serum and cellular constituents and some morphological characteristics of the spermatozoa were described statistically. Some genetic characteristics were tabulated. Based on capture-recapture measurements, I calculated approximate growth rates in body weight, body length, and chest girth for 4 male and 3 female puma over growth periods 2-27 months in length. Dispersion of 13 young puma from their known or presumed natal areas were mapped and described statistically. Sightability of puma by humans was inferred from questionnaires returned by 18 of 40 homeowners residing in Log Hill Village, an upscale home development encompassing about 8 km2. There were, at various times, 2-6 radio-collared puma whose inferred home ranges impinged upon or enveloped the development. I summarized damage claims paid and sex and age class of livestock killed by puma from 1 year prior to the beginning of the study to about 1 year beyond the study. I mapped the site of each paid damage claim. I mapped the spatial relationships between the relative numbers of deer and elk and the total number of aerial locations of puma on Uncompahgre Plateau winter range, 1982-88. We are currently considering using the population model by Barrett (1987) to gain some insight into the population dynamics of Uncompahgre Plateau puma. The narrative description of the study area environment and its puma population was augmented by maps of its vegetation, topography, land status, Game Management Units and drainages and tabulations of puma, deer, bobcat, and coyote kills. RESULTS AND DISCUSSION Status of Manuscript First drafts of the "Introduction" and "Description of the Population and Its Environment" sections of the manuscript entitled, "The puma on Uncompahgre Plateau, Colorado" are underway. The aim is to produce a CDOW Technical Publication to appear in 1991. About SO tables and most of 30 figures are complete. The number of tables and figures are approximate because some deletion, additions, or combinations are likely. Except for additional work on analysis of puma survival rates and possible simulations with a population model (Barrett 1987), all planned data analyses have been completed. I abandoned the attempt to assess habitat selection within puma home ranges because of lack of time. In any case, use of home range to determine what �163 habitat is available 1990:201). is now considered a source of bias (White and Garrott Home Range, a Comment It is likely that neither of the nonparametric estimators used (minimum convex polygon, harmonic mean) actually measures size of home ranges but rather provide useful indices to size (Beier and McCullough 1990:17). There are some instances to be described in the CDOW Technical Publication which suggest that such indices are useful. In one instance, the harmonic mean (0.95) boundaries of 2 mature males essentially coincided for over 2 years. Yet the mean elevation of the younger male's locations were lower (P < 0.05). Proportionally more locations of the young male occurred on the lower slopes of large canyons. LITERATURE CITED Ackerman, B. B., F. A. Leban, M. D. Samuel, and E. O. Garton. 1989. User's manual for Program HOME RANGE. Second Ed. Tech. Rept. 15, Forestry, Wildlife and Range Exp. Sta., Univ. Idaho, Moscow. 79pp. Anderson, A. E. 1983. Program Narrative Proj. 45-01-503-15050, Work Plan 6, Job 1. Mountain lion population dynamics. 7pp. (+3 tables and Appendix lA). Anderson, A. E., D. C. Bowden, and D. M. Kattner. 1989. Survival in an unhunted mountain lion (Felis concolor hippolestes) population in southwestern Colorado. pp. 57 in R. H. Smith, Ed., Proc. 3rd Mtn. Lion Workshop, Dec. 6-8, 1988, Prescott, AZ, Arizona Game and Fish Dept., Phoenix. 88pp. (Abstract) Barrett, R. H. 1987. Population model for mountain lion in California. Dept. Forestry and Resource Manage. Univ. California, Berkeley. 10pp + Appendix. Beier, P., and D. R. McCullough. 1990. Factors influencing white-tailed activity patterns and habitat use. Wildl. Monogr. 109. 51pp. deer Mielke, P. W., Jr., and K. J. Berry. 1981. An extended class of permutation techniques for matched pairs. Commun. Stat. 11:1197-1207. White, G. C., and R. A. Garrott. 1990. Analysis of wildlife data. Academic Press, Inc., New York. 383pp. Prepared by /dLt.."-e f(j 'C=" Ci?7vfeWtr1(./ Allen E. Anderson Wildlife Researcher radio-tracking ��165 Colorado Division of Wildlife Wildlife Research Report July 1990 JOB PROGRESS State of ~C~o~l~o~r~a~d~o~ _ Project No. Mammals Research W-lS3-R-3 Work Plan No. Job No. ~8~A~ _ 1 Period Covered: Author: REPORT Small Carnivorous Development Procedures Mammal Investigations of River Otter Reintroduction July 1, 1989 - June 30, 1990 T. D. I. Beck Personnel: K. Buege, L. Malville, J. White, G. White ABSTRACT Twenty river otters (Lutra canadensis) were received from Oregon in 1989. Fourteen were eventually released into the Dolores River, bringing the total released in 2 yrs to 19. Two animals died soon after release. A new prototype radio transmitter was used with variable results. Modifications were made on the system during the year, both in electronics and surgical procedure. Dispersion of river otters in the Dolores and San Miguel Rivers is widespread. Increased dispersal appeared stimulated by the reduction of river flows from 78 to 20 cfs, although definitive support for this position is lacking. No reproduction was documented in 1990. Crayfish (Orconectes virilis group) studies involved a removal test and a mark-recapture test. Results suggest further efforts on mark-recapture are warranted. Minimum adult crayfish densities were 2.75/sq m for the removal study area and an estimated 3.l/sq m for the mark-recapture area. These densities would produce 490 kgjha wet weight of adult crayfish. ��167 DEVELOPMENT OF RIVER OTTER REINTRODUCTION Thomas PROCEDURES D. I. Beck P. N. OBJECTIVE Develop procedures for river otter reintroductions in Colorado and establish a self-sustaining population of river otters from which to collect river otters for future translocations. SEGMENT up to 20 river otters OBJECTIVES 1. Introduce into the Dolores 2. Monitor all river otter release past reintroductions. 3. Develop techniques to monitor survival, reproduction, dispersal of river otters after reintroduction. sites in Colorado METHODS Dolores River drainage. to evaluate success dispersion, of and AND MATERIALS River Map Our river map developed in 1987 was found to be in error. The cause of the error was that we started measuring from the confluence with the San Miguel and used RM 58 for the confluence based on an earlier report from the U.S. Fish & Wildlife Service. Biologists for BioWest were working below the San Miguel and obtained new topographic maps and remeasured the river from the Colorado River to Slickrock. We combined their map with ours and altered the measurements accordingly. Dolores River Release River otters were captured in Oregon by Oregon Dept. of Fish & Wildlife personnel under the direction of John Thiebes. The traps used were a new design cage-type trap developed by Oregon personnel. River otters were placed in transport cages (41 X 41 X 76 cm) for shipment to Durango, CO, via commercial airlines. All river otters were shipped at 0700 (PDT) and arrived at destination at 1400 (MDT). When possible the river otters were taken directly to Cortez Animal Clinic for radio transmitter surgery and subsequently released into the Dolores River the same evening. At times it was not possible to schedule surgery on such short notice and then river otters were kept in holding pens (2.6 X 5.2 m) with access to food and water until a surgery time could be arranged. A new radio transmitter system was tried in 1989. In 1988, 5 river otters were released with Telonics Model 400 intraperitoneal implants. These implants were 95 mm long and 32 mm in diameter, weighing 110 gms. Because of concerns with the potential physiological impacts of such a large radio transmitter we worked with AVM Equipment Co. to develop a smaller unit that could be placed subcutaneous along the back. The prototype transmitter was �168 52 mm long and 16 mm in diameter with a 195 mm long whip antenna which also was placed subcutaneously. The AVM system weighed 22 gms. The initial electronic performance required was 50 bpm receivable at 250 m for 24 months. All transmitters were gas sterilized at least 10 days prior to surgery. River otters were transferred from transport boxes to a squeeze box at the clinic. Immobilization was accomplished with intramuscular injection of ketamine (100 mg/ml) and xylazine (20 mg/ml) at a dose rate of 15 mg ketamine per kg of animal. The transmitter was placed along the dorsal line between the shoulders. A 2-3 cm incision was made at that point and then a pocket was pushed in to make space for the transmitter. A 1 cm incision was made along the dorsal line about 200 mm behind the transmitter pocket. Blunt sponge forceps were pushed up under the skin to the anterior incision where the antenna was grasped and pulled along the back. Both incisions were closed with subcuticular stitches and super glue. Each river otter was tagged with a Monel #3 metal tag in the web between 2 front toes and measured for total length; tail length; neck, chest, and head girth; and weight. All injuries were recorded. Laboratory checks were made of hematocrit, white blood cell count, and fecal parasites. Most river otters were released into the Dolores River within 3 hours of surgery. A few were kept in large holding pens for observation and/or antibiotic treatment. One juvenile female river otter was kept for 2 weeks for a feeding study. Radio tracking was sporadic during this segment because of personnel problems and extremely low water. Radio tracking was primarily conducted from a canoe and on foot. McPhee Dam releases were reduced to 20 cfs on 1 March, and this precluded canoeing the upper 65 miles of the study area. All locations of river otters were recorded to the nearest 0.1 mi. Crayfish Studies Two separate studies were conducted in efforts to estimate adult crayfish densities in the Dolores River. A removal study was conducted at RM 144.7 between 25 and 28 July. The purpose of the study was to look for obvious trap biases by crayfish size and/or sex over time as well as assess the utility of a removal study for density estimates. The study stretch was 30 m wide (the river width) and 30 m long. Wire crayfish traps were placed in a 5 X 5 grid and baited with a commercial crayfish bait. Traps were checked once each day for 4 days. The following data were collected for each crayfish: sex, weight, carapace length, trap location, trap day. All captured crayfish were removed to an area 2 mi upriver and released. Based on the results from the removal study, a mark-recapture procedure was attempted. It was decided to mark each crayfish with a mark to designate each of 3 days rather than apply a unique mark to each of approximately 2,000 crayfish. This procedure is less satisfactory than a unique number for each animal but is much faster and cheaper. So it was decided to try the simpler procedure before the more complicated one. We again used a 5 X 5 grid with 25 baited wire traps; this time at RM 149.4. Traps were checked each of 4 days with the following data recorded: sex, carapace length, trap location, trap day, and capture history. All crayfish caught on Day I were marked by punching out a quarter-circle section from the outer ramus of the uropod on the right side of the animal when viewed dorsally. A paper punch was used to clip the ramus. All crayfish caught on Day 2 had a quarter-circle section removed from the outer ramus of the uropod on the left side. All crayfish �169 caught on Day 3 had a 5 mm scratch mark etched into the carapace with a triangle file. Population estimates were derived from several procedures through use of Program Capture. All estimates were conducted by G. White, Colorado State University. RESULTS AND DISCUSSION Dolores River Map The new map work indicated a 6.5-mi error in the old map. The confluence of the Dolores and San Miguel Rivers occurs at RM 64.5 rather than RM 58. All maps above the confluence have been corrected and all data collected in previous years has been corrected to the new river mile locations. Copies of the new maps were sent to all cooperators. All readers who have reports dated prior to 1 August 1990 should be aware of the inaccuracies in river miles. Merely add 6.5 to the old mileages and they will be corrected. Dolores River Release We received 20 river otters from Oregon between 21 August and 22 October 1989. Two juvenile males were dead on arrival at Durango. An adult male suffered from a heart murmur and high WBC and died 3 days after arrival while in the holding pens. A juvenile male was extremely lethargic upon arrival and never recovered from the anesthesia. An adult female, lethargic on arrival, was kept in the pens for 3 days before she died of unknown causes. Fourteen river otters were released into the Dolores River at RM 146.7 (Table 1). Nine were females; thus bringing our cumulative release total to 10 females and 9 males. One female, RO-28, died within 2 weeks of release. She had been diagnosed with anemia, elevated WBC, and elevated temperature upon arrival in mid-October. Necropsy indicated no fat reserves and a completely empty stomach and intestines. An adult male, RO-32, was found dead of starvation along Colo. Hwy. 666 in mid-December. He was at least 6 airline miles from the river and nearly 250 m higher in elevation. After release at RM 146.7 he moved upriver to RM 152.5 the first day; then downriver to RM 132.5 during the next week. He probably died around late Nov. as all the ponds froze up then. The closest point to the river was RM 169.5 so he apparently went upriver 37 mi before climbing out of the canyon. River otter scat found near the ponds along Hwy. 666 indicated he had been feeding on crayfish. Necropsy indicated severe emaciation. RO-2l expelled her radio transmitter, probably because of a suture failure. Performance of the prototype transmitters was variable. A weakness discovered early in the season was that the antenna wire was too thin and was either breaking or coiling up around the transmitter. However, some of the thin-wire units performed adequately. Only 4 thin-wire units were released; subsequently all transmitters were rebuilt with a heavier wire antenna. It became necessary to place a tear-drop bead of epoxy on the end of the heavywire antenna to prevent the antenna from puncturing the skin. Performance of the prototype transmitters was variable, with above ground transmission distances of 250-400 m. While acceptable, we discovered that transmission from underground (bank beaver dens) was severely restricted, varying from 30-70 m. Since the released river otters spend so much time in the beaver dens, it became critical to increase transmission distance. A new �170 prototype was developed with a slightly larger battery and greater output. We still anticipate a battery longevity of 23 months. The new unit measures 66 mm long X 15 mm diameter with a 195 mm whip antenna and weighs 23 gm. Field tests indicate the underground transmission distance should be at least 200 m, which is quite acceptable, with a few units transmitting to 400 m. Personnel problems, combined with conflicting new assignments, resulted in little radio tracking during the winter season. Unfortunately, this year was also a very low snowfall year, resulting in no spring runoff below McPhee Dam. The operators of McPhee Dam lowered the releases to 20 cfs on 1 March 1990, which made canoeing of the upper 65 mi of the study area nearly impossible. Surveys in the lower canyons in March, April, and May indicated river otters were spread throughout the Dolores River at least to the San Miguel River. It is reasonable that the sudden reduction in river flows from 78 to 20 cfs triggered movement downstream in anticipation of the river drying up. Numerous sightings of river otters have been reported from the Colorado River downstream to Lake Powell, but it is unknown if these are from the Dolores River release or possibly dispersal from other release sites. Unfortunately, the only female released in 1988, RO-5, has not been located in 1990 so her reproductive status is unknown. At least one male, RO-7, has moved upriver into McPhee Reservoir. Another male, RO-4, missing since March 1989, was located on 31 July 1990 in the San Miguel River approximately 43.1 mi upriver from the Dolores River and 125.3 mi from the release site. An adult female, RO-25, had moved downstream into Slickrock Canyon. She was found dead of natural causes in late July 1990 at RM 107.7. She was not present at that site 3 weeks previously and the radio transmitter was performing well. The widespread dispersion of the river otters strongly supports the need to release at least 30 animals in one area. Since fish and crayfish appear to be abundant and bank cover, both vegetation and beaver dens, is abundant it is doubtful if the cause of dispersal is habitat, although there is no definitive data. RO-9 was kept in captivity for 2 weeks in an attempt to assess food habits estimates from scats. The data collected was of limited utility. Her intake varied from a low of 50 gms of crayfish (tails only) to 1350 gms of trout (Oncorhynchus gairdneri and Salmo trutta). Soon after the beginning of the feeding trials we were forced to use the adjacent pen for holding other river otters. The greatly increased human activity may have caused a significant reduction in food consumption. RO-9 then bit the employee feeding her and had to be euthanized in order to check for rabies. The feeding trial is worthwhile but will only be repeated with one of the last animals to be handled in 1990. During the first week of the trial, RO-9 was ingesting food amounting to 35-40% of her live weight daily. �171 Table 1. River otters released in the Dolores River, co. 1989. ID Length (em) Total Tail Circumference (em) Head Neck Chest Tag if Origin Sex Wt(kg) RO-9 F 3.9 91 36 22.5 21.0 29.0 9 Columbia R. , OR RO-12 M 5.2 95 38 24.0 23.0 33.0 12 Tenmile Lake, OR RO-15 M 8.3 116 46 26.0 27.0 41.0 15 Tenmile Lake, OR RO-17 M 10.9 125.5 47 25.5 30.5 47.0 17 Columbia R. , OR RO-21 F 5.9 104 37 24.0 26.0 40.0 21 Tenmile Lake, OR RO-22 F 5.4 98 nd nd nd nd 22 Willamette R. , OR RO-23 F 8.3 113 41 25.5 29.5 44.5 23 Willamette R. , OR RO-24 M 4.8 94 33 23.0 23.0 35.5 24 Columbia R. , OR RO-25 F 7.4 114 44 24.0 25.5 39.5 25 Tenmile Lake, OR RO-26 F 7.3 108 41 24.0 27.0 42.0 26 Willamette R. , OR RO-27 F 4.1 93 nd 22.5 21.0 33.5 27 Marion Fks Hatch, OR RO-28 F 4.6 99 38 21.5 21.5 32.0 28 N Umpqua R. , OR RO-29 F 6.1 104 38 23.5 24.0 32.0 29 N Umpqua R. , OR RO-31 F 7.5 117 43 24.0 26.0 37.5 31 Willamette R. , OR RO-32 F 8.3 115.5 42 27.0 28.0 40.0 32 Deschutes R. , OR Crayfish Studies During the last 20 years there have been many studies throughout the world attempting to estimate crayfish numbers by means of trapping programs. The variability in study sites is matched by the variability of results. When using any mark-recapture program the assumptions on equal catchability are critical so many workers have attempted to quantify capture biases by sex, season, and capture method. Unfortunately, there are no general patterns. While most authors suggest a trap bias for large males (often exceeding 75% of total catch) (Momot and Gowing 1983, Brown and Brewis 1978, and many others) there are instances of trapping being more selective for females (Brown and Bowler 1977). Seasonal differences can be quite strong because of reproductive requirements. This is why we did our studies over a short period �172 (5 days) during late summer after the major molting periods. A complication with short studies is that all crayfish are not active every night and that perhaps certain groups of crayfish have coinciding activity periods (Hazlett et al. 1979). Thus, one could capture and mark a large number of animals on Day 1, then rarely see these for 2 days only to see large numbers of recaptures of Day 1 animals on Day 4. The removal study was conducted first to check for obvious biases in capture rates. The initial assumption was that large males would dominate the early catch but once removed from the system, females and small males would show up later. Based on earlier work on Orconectes virilis (Hazlett et al. 1979), we were confident that all crayfish within the study reach were within easy reach of a baited trap. Since we were concerned with immigration into the study reach, we planned the study for 5 days as we optimistically figured to catch nearly all the crayfish within that time. After 4 trap nights we had captured and removed 2,470 crayfish (1,283 male, 1,187 female). No diminution in catch occurred through time: 633 on day 1, 532 on day 2, 678 on day 3, 627 on day 4. There was no strong trend based on sex, as percentage of males for the respective days was 55.5, 47.2, 51.3, 53.1. There was no change in length-frequency curves for either sex over the 4 days, thus no identifiable trap bias based on size. There was no evidence that significant immigration into the study reach occurred. Number of captures in Row 1 (upstream row) was 115, 158, 153, and 115 for days 1 through 4. Number of captures in Row 5 (downstream row) was 80, 47, 174, and 128 for days 1 through 4. There was surprising uniformity in catch by row (541, 464, 501, 535, 429 for 1-5, respectively). There did appear to be a tendency for the mid-stream column of traps to fish less effectively than the more shoreline traps, but the traps closest to the shoreline did not differ from the next column in toward midstream. More detailed descriptions of possible trap bias are being prepared by G. White. Our tentative conclusions from this study were: 1) no observable trap bias based on size or sex, 2) we certainly did not trap out the adult crayfish population, 3) the minimum known density of adult crayfish in the study reach was 2.75/sq m. This density exceeds all reports that I have located for rivers in N. America and Europe (range 0.131.8) but is not directly comparable because other workers are estimating density over a much larger area with heterogenous habitat. Our study reach was relatively uniform in habitat. The mark-recapture study was conducted upriver in a similar habitat but with slightly less boulder on the bottom. The same grid system was employed. During 4 nights of trapping 1,390 individuals were caught, of which 343 were recaptured at least once. Detailed analyses are still being conducted by G. White, but initial runs through Program Capture resulted in promising estimates. More work needs to be done on capture heterogeneity, and in 1990 a study will be conducted using uniquely marked individuals. The Null, Jacknife, and Darroch estimators produced similar results with surprisingly small 95% confidence intervals. The General Removal Method produced an estimate 3 times greater than the others. For comparative purposes, G. White (pers. comm.) suggested the Jacknife method as a reasonable estimate. The estimate was 2,795 with the 95% C.l. being 2,673-2,917. This would provide a minimum adult crayfish density of 3.1/ sq. m. The average size of the 2,470 crayfish weighed in the removal study was 15.75 gms. At a density of 3.1 adults/sq. m., the standing stock biomass of adult crayfish would be 490 kgfha (437 lb/ac). That amounts to 31,098 adultsfha or �173 12,585 adults/ac. Thus,it appears that crayfish can support numerous predators in the portion of the Dolores River between RM 173 and 130. LITERATURE CITED Brown, D. J., and K. Bowler. 1977. A pcpu l.a t Lon study of the British freshwater crayfish Austropotamobius pallipes (Lereboullet). Freshwater Crayfish 3:33-49. Brown, D. J., and J. M. Brewis. 1978. A critical look at trapping as a method of sampling a population of Austropotamobius pallipes (Lereboullet) in a mark and recapture study. Freshwater Crayfish 4:159164. Hazlett, B., D. Rittschof, and C. Ameyaw-Akumfi. 1979. Factors affecting daily movements of the crayfish Orconectes virilis (Hagen, 1870) (Decapoda, Cambaridae). Crustaceana, Suppl. (Leiden) 5:121-130. the Momot, W. T., and H. Gowing. 1983. Some factors regulating cohort production of the crayfish, Orconectes virilis. Freshwater BioI. 13:1-12 . .f/' () f A~ n Prepared by ~7?p? ,i'/.:/' Thomas D. I. Beck Wildlife Researcher ��Colorado Division Wildlife Research July 1990 175 of Wildlife Report JOB PROGRESS State of Project REPORT Colorado No. W-IS3-R-3 Mammals Research Work Plan No. 9A Elk Investigations Job No. I Impact of Elk Winter Production Grazing on Livestock and Work Plan No. 3 Job No. S Period Covered: Authors: Personnel: July 1, 1989 - June 30, 1990 N. T. Hobbs, D. L. Baker, G. Bear M. Miller, J. Miller, R. Reid, L. Carpenter, C. Woodward, B. Seely, H. Seely, L. Lovett B. Gill, B. Petch, ABSTRACT Elk grazing during winter influenced cattle performance on sagebrush-grassland range during spring. Weights of cows and calves at the end of the spring grazing season declined in response to increasing intensity of winter and spring grazing by elk. Reductions in spring weights resulted from reduced rates of gain. Cattle performance declined annually during 4 study years. We attribute annual effects to the combined influence of cattle .grazing and, drought on the productivity of the site we studied. However; 'the magnitude of the effect, of el~ grazing Aid not; ,de-eend on year, sugge st Lng that no cumulative effects of elk grazing- were prese.nt;' Analysis of data' from year 4 on vegetation responses is in progress. ��177 IMPACTS OF YINTER GRAZING BY ELK ON CATTLE PRODUCTION N. Thompson Hobbs Dan L. Baker P. N. OBJECTIVES 1. To test the hypothesis that elk grazing during winter influences the productivity and botanical composition of herbage on sagebrush grassland ranges during spring. 2. To test the hypothesis that elk grazing during winter influences the body weights and rates of gain of cows and calves using sagebrush grassland ranges during spring. METHODS AND MATERIALS Study Area We conducted experiments on the Little Snake Wildlife Management northwestern Colorado (township 9 north, range 95 west, sections area is about 35 km (19 mi) north of Maybell, Colorado on County Although this area does not typically contain high concentrations during winter, it is representative of areas that do have those densities. Area in 9, 10). The Road 19. of elk high Topography of the area includes level ridge tops, rolling hills, and deep gullies, ranging in elevation from 1,800 to 2,000 m (5,900 to 6,600 ft). Aspects are southern and southwesterly with an average slope of 15 degrees. Soils are generally sandy and sandy loam. Climate of the area is dry and cold. The growing season averages only 81 days. Annual mean temperature is 6.06 C (42.9 F). Annual precipitation averages 27.5 cm (12.5 in). Vegetation is dominated by big sagebrush (Artemisia tridentata) with an understory predominated by needle and thread (Stipa comata), western wheatgrass (Agropyron smithii), Indian ricegrass (Oryzopis hyrnenoides), Junegrass (Koleria cristata), and cheatgrass (Bromus tectorum). Important forbs include wallflower (Erysimum asperum), peppergrass (Lepidium perfoliatum), silver lupine (Lupinus argenteus), and scarlet globe mallow (Sphaeralcea coccinea). Experimental Design We observed effects of elk grazing on forage and cattle responses in a randomized complete block design with four levels of elk density (0 elk/km2, 8 elk/km2, 15 elk/km2, and 31 elk/km2) and three replications per level. There were three blocks, each consisting of four pastures. Each pasture within a block was stocked with one level of elk density such that each block contained all levels. During year 1, the twelve available pastures were blocked by pretreatment biomass of perennial grasses with the four lowest grass biomass pastures forming one block, and the four highest grass biomass pastures forming a second block, and the remaining four pastures serving as the third block. The four levels of elk density were randomly assigned to pastures within each block during year 1. �178 Procedures We stocked pastures with elk in December and January 1987-89. Average date of release into pastures was January 3. All elk were removed from pastures during April 10-20. We introduced 7 cow-calf pairs and one dry heifer into each pasture on May 9, 1990, and removed them 5 weeks later. This represents a departure from the first two study years when animals remained in the pastures for 6 weeks. However, a marked reduction in forage production during the 1988-1989 growing seasons compelled us to reduce the stocking rate during 1990 in order to preserve consistent utilization rates in control pastures. With the exception of addition of new heifers and other replacements required by death losses, etc., cows during year 4 were the same animals we observed during year 3 and were assigned to the same pastures they were in previously. We observed the birth dates of all calves and weighed them to the nearest 0.05 kg immediately after birth. Cows and calves were weighed to the nearest 1 kg when they were introduced to pastures and were reweighed 5 weeks later when they were removed. We estimated canopy cover of herbs shrubs immediately after removing cattle from pastures. Cover was estimated from the summed length of interception by each plant along 30, l2-m transects randomly placed in each pasture during year 1. We estimated standing crop, productivity, and utilization of forbs, perennial grass, and annual grasses by harvesting samples from 40 pairs of 0.70 m2 plots in each pasture on each of three sample dates. Pastures were sampled immediately after the elk were removed (April 27-29), at the midpoint of the spring grazing season (May 30-June 1), and at its end (June 30-July 2). Samples were dried at 60 C for 48 hrs, separated by hand into live and dead, and weighed to the nearest 0.01 g. We used a repeated measures analysis of covariance for a randomized complete block design to analyze weight responses of cattle. Calf birth weight was used as a concomitant observation for calf weights out of pastures and for rates of gain. Cow weights into the pastures were used similarly for adult cattle and heifers. We believe that this analysis offers conservative tests of hypotheses because any cumulative, depressing effect of treatment on the covariate would tend to make it more difficult to detect significant differences among treatments. However, we are in the process of consulting with Dave Bowden (Statistics Department, Colorado State University) on the details of this analysis to assure our statistical approach is reliable. Future analyses will reflect these consultations, and may offer somewhat different results than we report here. RESULTS All results described in this report are preliminary and are subject to revision as analyses are checked for accuracy. Here, we report our findings to provide a timely summary of progress, but we are obliged to thoroughly revisit them later. �179 Effects of Elk on Cattle Performance Elk grazing during winter influenced the performance of cattle during spring. Averaged over 4 study years, weights of calves at the end of the spring grazing season declined in direct proportion to the winter/spring density of elk (Figure 1, linear effect P - 0.023). End of spring weights of calves in the high density treatment averaged about 8 kg less than those in the control (Figure 1, control vs 31 elk/kmz P - 0.021). The average weights of calves (adjusted for differences in birth weight) in the control were greater than the average weights of calves in pastures stocked with elk (Figure 1, control vs. treatment P - 0.042). However, differences in calf weight between the low density elk treatment and the control were small (about 1 kg) and were not significant (P - 0.20). Averaged across treatments and controls, calf weights tended to decline during years 3 and 4 relative to the first 2 study years (Year effect P - 0.0001). However, the magnitude of the effect of elk grazing was consistent across all 4 years of the study (Year x Level P - 0.479). That is, the effect of elk grazing did not intensify as a result of "cumulative" effects as the study progressed. Effects of elk grazing on end of spring weights of calves reflected differences in average daily gain (Figure 2, linear effects P - 0.01). However, the rates of gain in the low and medium density pastures were quite similar to those in the control, suggesting threshold effects of elk grazing between the 15 and 31 elk/kmz levels. We will compare threshold with linear models in future analyses. We observed significant differences among years in calf rates of gain (year effect P - 0.0001), but saw no year by treatment interaction (year x level P - 0.81). This once again suggests that the magnitude of the effect of elk grazing remained constant over time and did not accumulate over study years. Calf birth dates were delayed in response to all levels of elk grazing (Figure 3, control vs. 8 elk/km2 P - 0.03, control vs. 15 elk/km2 P - 0.01, control vs 31 elkjkm2 P - 0.06). However, this effect was not linear and tended to reach a maximum at the 15 e1k/kmz level (Figure 3, quadratic effect P - 0.02). Responses of cows to elk grazing resembled those of calves, but higher interanimal variability tended to obscure the significance of treatment. Average weights of cows at the end of the spring grazing season (adjusted for differences in pre-season weights) declined in proportion to elk density, but this effect only approached statistical significance (linear effects P 0.07). Similarly, rates of gain declined as elk density increased, but this response was highly variable (linear effects P - 0.07). Cow weights at the end of spring as well as spring rates of gain were strongly influenced by year (P < 0.0001). However, the size of the effect of elk density on cattle performance did not change with year for either response (P > 0.89) suggesting no cumulative effects were operative. Effects of Elk on Vegetation Effects of elk on vegetation production, years are currently being analyzed. condition, and trend after 4 study �180 DISCUSSION A tendency for elk grazing during winter to reduce the performance of cattle during spring has clearly emerged in our studies after four years. We can be quite confident that intense elk grazing during winter and spring harms the performance of calves during spring and can be reasonably certain of similar effects on cows. Prepared by N.r:::h-tt vLVWildlife Researcher q_~,f'_(~ Dan L. Baker Wildlife Researcher �181 1.0 >: jg 0.9 0, /\ " 0.9 r: 6 al •.. 0.8 :::J 1i) as a... .S 1.0 = :; (1] 32 ._a u .::t::. * 0.8 Q) •... ::l Ui &: 0.7 0.7 .S c c '113 '(i3 o o -o 0.6 0.6 ~ 0.5 ~ 0.5 C/) - co o <U __J 0.4 0.4 0 5 10 15 20 25 30 35 Elk Density (animals/km2) iYear ! '* *::.::1 ::;:4 Means adjusted to common birth weight. OJ!" A,11.. Years control vs c::ont.ro~ va c::ont.ro~ va other. cont.ra_t. ss po Va:Lu_ Pr 3.50 1.34 0.00 11.29 10.28 1.76 0.1106 0.2905 0.9514 0.0152 0.01B5 0.2324 > F ~~~~~~:l.e~;_~~a-:l.k/k:m! quadrat.ic:: _~~_c::t._ 1 1 1 0.03534255 0.01356235 0.00004078 0.11400822 0.1038:01473 0.01781855 Year 1 contro~ VB other_ c::ont.ro:l. va 8 _~k/kn121 c::ont.ro~ va 15 e~k/k:m c::ont.ro~ 'V's 31 e~k/k:m :Lin_ar e~~_c::t._ quadrat.ic:: _~~_c::t._ 1 1 1 1 1 1 0.00539948 0.00011359 0.000625:018 0.02083060 0.02384608 0.00297932 0.77 0.02 0.09 2.96 3.39 0.42 0.4147 0.9030 0.7757 0.1361 0.1152 0.5393 1 0.00685662 0.00556753 0.00061201 0.01070700 0.00765803 0.00000573 1.25 1.02 0.1.1 1..96 1.40 0.00 0.3059 0.3522 0.7495 0.2115 0.2817 0.9752 1 1 1 1 0.00863837 0.01124589 0.00:0185982 0.03065762 0.02071642 0.01003653 1.46 1.90 0.4B 5.19 3.50 1.70 0.2722 0.21.70 0.5127 0.0630 0.1104 0.2403 1 1 1 1 1 1 0.01606900 0.00173012 0.00027242 0.06370886 0.06685997 0.01205902 1..13 0.12 0.02 4.48 4.70 0.85 0.3287 0.7392 0.8945 0.0787 0.0733 0.39:017 1 a _~k/k:m2[ 15 _ :l.k/k:m! ~O~{rg1 V_ oont.ro:l. ~_ oth_r_ 8 _~k/k:m2 c::ont.ro~ ~_ 15 -:l.k/k:ml oont.ro:l. ~_ 31 _~k/k:m 1in __ r _~~~ct._ , quadrat.ic _~~_ct._ ! x.or contro1 :3 va oth_r_ c::ont.ro1 va 8 _:l.k/k:m2j c::ont.ro~ va 15 _:Lk/k:ml c::ont.ro1 v_ 3~ _~k/km, :Linear e~~_c::t._ ' quadrat.ic::_~~_ct.a Year 4 c::ont.ro:l. v_ ot.h.r_ \ c::ont.ro~ ~a 8 _1k/km2\ c::ont.ro~ v_ 15 .1k/k:m c::ont.ro~ v_ 31 e:l.k/k:m 1in.ar .rr_ct_ quadrat.ic:: .~~_ct.a I a, 1 1. 1 1 1 1 a, 1 Figu~e2~ Eff,ectsof ~lk,~ip~~r,-gr~ngJ?~,Jates>ot.g~i~ of"caIY~:during the spring gr~ll:!9,s.ea~ .. Dat~,PQln~.areleast sqy8(ep;rl1~$.:(LSM};,fQ(i~h pastureduring each'year~ line Show$ I;.;SM.-vallies acrcss.all years. V.erti~ :pars-= 2 standard errorsotme mean. ,. ;.,'----:-'. -..-" -- �182 - 100 100 a~ 0) 6 ~ 90 ::J en e (5 0) 80 - :-.:: 1 80 i . ~ 70 (5 "S 0 -'"'- ==fzz -; 70 1: 0 ~ - <ti (ij ::: ::J (Ij "< ;,,; ~ o ~ en o, "S 0 E 90 ~ 60 60 o ~ C/) ~ _J 50 50 0 10 5 Elk Density Year ~ _, ::-::1 20 15 OF Y:!!ars c::ontroOi vs ot.hers c:.ontro1 va 8 e1k/krn2 oC)ntro~ va 15 eJ..k/krn va c:ont.ro~ 31 eJ..k/k:m e:rr_c'ts l...i..near qu ••• clr ••• t:1.c effect.s 30 25 (animals/km2) .::_2 Means adjusted to common Contrast 35 =4 =3 birth weight. Cont.rast 55 F Va~ue Pr >- F :!!k1~ 1 1 1 :1. 1 1 326.82828197 103.66999879 153.96600527 470.57267042 448.04755462 12.91123368 6.67 2.11 3.14 9.60 9.14 0.26 0.0417 0.1961 0.1267 0.0212 0.0233 0.6261 1 1 1 1 1 1 12.28671445 3.33767567 27.54509577 26.67350990 46.75838244 1.01851203 0.42 0.11 0.94 0.91 1.59 0.03 0.5422 0.7478 0.3708 0.3780 0.2544 0.8586 1 1 1 1 1 1 16.30259696 2.40639274 2.75382664 44.61511961 46.61121821 2.51519165 0.82 0.12 0.14 2.25 2.35 0.13 0.3999 0.7397 0.7224 0.1846 0.1764 0.7341 ot.h_r_ va v_ cont.rol.. 8 eJ..k/km2 c:ontro.1. 15 _1k/k:m cont.rol. va 31 e1k/km art_ct._ l..,in_ar _rr_ct._ qu. ••• clr ••• t.1.c 1 1 1 1 1 1 94.48924391 72.27683100 7.58981980 157.59954439 116.31941128 0.58705130 7.87 6.02 0.63 13.13 9.69 0.05 0.0309 0.0495 0.4568 0.0111 0.0208 0.B323 4 c:ont.:ro.l.. ot.hers c:::ont.ro.1. B _1k/krn2 control.. 15 e1k/krn contro1 31 e1k./km _ 1!:r_c::t._ .1.i.n_ar quadrat.ic: er1!ec::t.s 1 1 1 1 1 1 356.95826295 147.32595506 229.63877726 360.52185069 319.83276786 72.74990977 5.02 2.07 3.23 5.07 4.49 1.02 0.0664 0.2002 0.1226 0.0654 0.0783 0.3510 v_ v. XS2ar • other_ con.t.:ro~ va c:ont.ro~ 8 eJ..k/km2 c::ontro.1 VEl 15 eJ..k/krn cont.:ro~ 31 eJ..k/k:m 1.in_ar e~~ect_ ef-r_c::'ta quadr ••• t.1.c ". 20~{rg1 v. oontrol. ot.hers 8 e1k/km2 'Va_iS contro.1. eJ..k/kDl control. 31 eJ..k/k:m _~r_c::t._ 1.1.,n_ar quadrat.1.c_f~_c::t._ ,,- ;;o~Erg~ ,,_ Xe!!E ". "._ ,,_ ,, Figure 1. Effects of elk winter grazing on weights of calves at the end of the spring grazing season. Data points are least squared means (LSM) for each pasture during each year. Line shows LSM values across all years. Vertical bars = 2 standard errors of the mean. �183 05/00F===============:::_:_- 05/08 ~ <i] ~ 04/28 +-- ---:=::-- __ 04/28 -==- __ ___c=-.:._ C g = ~ 04/1a-'--=-~=---=-' Ii: a:J t , u. 04/08 f5 ,------lI=--_ T 7_- __ Y 04/18 x --:" -r---:""":;::-' ----~t~: _____:_ -.~ ---=-=__ --=-"-- 04/08 Z 03/29-,===================03/29 10 15 o 5 20 25 30 35 Elk Density (animals/km2) Year -s; Contrast x1 2 OF ,"",,11 Years control vs others control vs s e~k/k~2 control vs 1S elk/km contro~ ~B 31 e1k/k~ 1.i.near effect.s quadratic effects ~~n{r~1 ~_ others contro~ va e e~k/k~2 contro1 va 15 e1k/km contro1 v_ 31 81k/k_ ~in_ar _~~_c:::t_ quadratic _~~_ct_ Xear »< :3 contro1 vs others contro~ va a _lk/km2 control ~_ 15 elk/km contro1 VB 31 e1k/k_ linear ef~_ct_ quadratic: _f~ect_ V_ Year 4control other_ control va e 81k/km2 contro1 v_ 15 e1k/km contro1 v_ 31 e1k/km l.i.nEusr 4!t~~_ct_ quadratic _~fects 1 1 1 1 1 1 cont.rast -3 ~4 SS 249.03703704 162.00000000 242.00000000 107.55555556 71.42857143 198.10678211 F Va.1ue Pr 13.09 8.51 12.72 5.65 3.75 10.41 0.0111 0.0267 0.0118 0.0550 0 ..1.008 0.0180 :>- F 10.02777778 24.00000000 1.50000000 2.66666667 0.02142857 !5.8!528138!5 0.34 0.81 0.05 0.09 0.00 0.20 0.5821 0.4031 0.8296 0.7744 0.9794 0.672!5 1 1 1 1 96.69444444 48.16666667 66.66666667 80.66666667 66.40238095 30.13852B14 6.32 3.15 4.36 !5.27 4.34 1.97 0.04!57 0.1264 0.0819 0.0615 0.OB24 0.2101 1 1 1 1 1 1 205.44444444 104.16666667 308.16666667 54.00000000 40.23809524 271.24675325 4.B5 2.46 7.27 1.27 0.'95 6.40 0.0699 0.1679 0.0357 0.3020 0.3674 0.0447 1 1 1 1 1 :1. 1 :1. Figure 3. Effects of elk winter grazing on calf birth dates. Data points are means foreachpasture during ~achyear. Line shows mean values across all years. V~rtfC$fba~¥;~'_~ndar~ errors of the mean. - ....- '." ;~<; . _-~ .. ~ <7: ::-~. .:"_~ : -""~_:;i:-_~:~"~:~-.~:~ ~?:.-~~~f::,:'-:' .- �184 460 ~~:================================~ cn450~~--------------------------------------------450 e CD .._ '< x: 440 44OTI~~--------~C-------------------------------~'- ::::l ;.t( 430 ~ 430--=o--------~------~------------------~------ cf'+-0 5410 t f -, 42O~: -+t---------~-F-" ---------'<---- K 420 410 4OO~. --===----±---.lr------=~o---~------=~---400 ~ 390: -= 390 380 ~ 380: o . Um m ~B ~ ~ 350 350 o 5 15 10 20 25 3S 30 Elk Density (animals/km2) Year ::oK X x 1 t: L .;i 2 =:___! = 3 Means adjusted to common weight into pasture. Cont.ra_t l!o~;!' OF Cont.rast. 55 F Va~u_ Pr :> F :;(eara 'VB ot.her. contro1 c::on:t.ro a, 'VB 8 _~l<./l<.rn2 'Va cont.rol.. 1..5el..l<./l<.:m 'Va c::ont..ro1 :31 _~l<./l<.m 1inear effect. quadrat.i.e::_~~_c:t._ 1 1.. 1 1 1 1 812.03046807 478.93347940 114.6032:3541 1384.67221..728 1119.6184:3059 2.1..5293548 3.44 2.0:3 0.49 5.86 4.74 0.01.. 0.1131.. 0_2043 0.5121.. 0.0518 0.0723 0.9270 1.. V!!!or others c:ont.ro~ 'Va cont.rol. va 8 e~l<./l<.:m2 v_ cont.rol.. 1..5_1l<./l<.:m c:ont.ro1 va :31 _~l<./l<.:m l..i. •..• ear e~r_c:t.1!I quadrat.i.e::e:l!:I!ect._ 1.. 1 1 1.. 1 1 0.21785646 27.0533751..4 107.23 .••. 58082 39.65761..081 1:3.47957560 1..13.59717956 0.00 0.14 0.56 0.21.. 0.07 0.59 0.9743 0.7209 0.4841 0.6662 0.8003 0.4719 5.14 2.03 1.88 7.59 6.84 0.11 0.0640 0.2036 0.2194 0.0331 0.0:398 0.7522 0.:3486 0.8:309 0.5554 0.1516 0.1287 0.7999 0.4515 0.:3978 0.7291 0.5109 0.6559 0.7399 g~~€r~1 v_ other_ _~l<./l<.rn2 c:ont.ro1 vl!Il_8 control.. va _~lc/lc:m 15 c::ont.,ro.1 'Va :31 _~lc/l<.Dl ~i. •..• _ar e-rr_c:t_ _~r.c:t._ quadrat. i.e:: X-AE 1 1 1 1 1 1 733.:39748071 290.46453972 268.:32726608 108:3.1785936:3 976.5661:3814 15.60084540 1 166.78463458 8.034449:32 ~ v_ ot.h_r_ control.. c:on:t:.rol. 'V_ 8 _1lc/l<.:m2 c:ontro.J.. 'V_ .1.5_:l.l<./lc:m 'V_ cont.ro1 _~lc/l<.Dl 31 efr_ct._ l..i.near quadrat.i.e::err_ct._ Vear ~ ot.h_r_ control. 'VB cont.rol.. 'Va 8 _~lc/l<.m2 e::o •..• t.ro~ va _~l<./lc:m ~5 cont.rol.. va _l.l<./l<.:m 31 l...i.n_ar ef~eot._ quadrat.i.e::e:l!:I!_e::t._ .1. 1.. 1 1 435.42384833 500.78900985 11.:32868318 1.0:3 0.05 0.39 2.70 3 •.1.0 0.07 1 1.. 1 l. :1. :1. 273.2:3166622 :349.:32660952 55.5420l.859 205.82058047 92.57464015 50.97447616 0.65 0.8:3 0.l.3 0.49 0.22 0.12 a, 62 .. 9268:3282 Figure 4. Effects of elk winter grazing on cow weights at the end of the spring grazing season. Data points are least squared means (LSM) for each pasture during each year. Line shows LSM values across all years. Vertical bars = 2 standard errors of the mean. �185 2 2 ,\ W 'i x x 3iE ~ >K LJ - ~ ~ ~ < ,''. '* e- 0 " - i~ CD •... :::::l 1' t~ W (/) CIl c, ,S; c: 'n; H CJ 1 ~ ,£ c 'n; 0 CJ ~ 0 ~ - (.) ::E 0 o ::E .;::: CJ) ::::J (ii :::: ;- -=CD..._ CJ) ....J ...J -1' -1 0 10 5 15 20 25 30 35 Elk Density (animals/km2) Year c< x ~ 1 2.2 ,--' =3 Means adjusted to common weight into pasture. OF contraat ~;;*€~~!FVI!& other. v_ 1 C:On.t.FO~ 8 e:1.1<:./1<:.:m2 1 c::ont.ro.1.. va :1.5el..1<:./1<:.:m 1 c:ont.ro1 va 31 e11<:./1<:.:m 1 err_ct.. l.inear 1 e~~_cta quadratic 1 ,,~~~(r!.1.. ot.h_r_ 1 v_ c:::ont.ro1 8 _l..1<:./1<:.:m2 1 control.. "a _l..1<:./1<:.m1 15 c;ontro1 _l..1<:./1<:. ••• va 31 1 l..inear e~~ec:ta 1 _~~ec:t_ quadratic 1 g~~(rg1 va other_ 1 control.. "a a el..1<:.,1<:._2 1 control.. l..!5 el../kDl 1 contra:&. va 3J..el..1<:./1<:._J.. l.in•••• r e~~_c:ta J.. quadratic J.. e~~ec:ta ,,- ~~:ttrg1 va ,,c::oftt.ro1 oth_ra 1 8 e11<:./1<:._2 1 c::ont.:ro1 va el..1<:./1<:._l.. 15 c:ont.ro.1. va el..1<:./1<:.:m 31 1 l..in_ar e~~_cta 1 quadratic: _~~_cta :1. ~g~[:r@1 "a othera :1. va a el..1<:./1<:.:m2 1 va J..5_l..1<:./1<:._1 va e11<:./1<:.:m :1. 31 1in •••• r _~~ecta 1 _~~_cta quadratic 1 cont.ro.]. contro:&. cont.ro.]. Cont.ra •• t 55 po Va1u_ Pr > 0.70495984 0.35636241 0.0994:1.263 1.30960362 1.11228462 0.00855596 3.08 :1..56 0.43 5.72 4.86 0.04 0.:1.298 0.2586 0.5343 0.0539 0.0697 0.8531 0.00006598 0.0l..803876 0.09624188 0.02434338 0.00682232 0.0964:1.649 0.00 0.:1.5 0.81 0.21 0.06 0.81 0.9819 0.7097 0.4019 0.6661 0.8182 0.4015 0.37847991 0.16770179 0.J..2706827 0.54902660 0.48000712 0.00800233 4.35 1.93 1.46 6.31 5.52 0.09 0.0820 0.2143 0.2723 0.0458 0.0571 0.77J..9 0.36275296 0.01386065 0.l..4087886 0.96478172 1.12J..99042 0.02683770 J...OJ.. 0.04 0.39 2.69 3.:1.3 0.07 0.3533 0.8506 0.5539 0.1521 0.1273 0.7936 0.22077174 0.28341587 0.04368912 0.16772162 0.07542428 0.03995186 0.64 0.82 0.:1.3 0.48 0.22 0.12 0.4548 0.4002 0.7344 0.5:1.23 0.6570 0.7455 f:;';;:/~:-'~:'~:~i~~~~ri_ar1!ec;g:_':'_'~ffi6& clciUZ~~I~iit~'~1,n (atiJ~ia]gafn.Of:Cmv$ dt,jring-~e-,spring '~:;'~~~tr grazin_gseason. Om ~~.. :.,---,,-each yew. U-ne: errors~ :mean. 1" - -squared means_(~§Mrfgr. e~Chpasture during YaIUee across alt y ars. \lertlcatJ)ats ==:~ <standard ��187 Colorado Division- of Wildlife Wildlife Research Report July 1990 JOB PROGRESS State of Project Work Colorado No. ~W_-~1~5~3~-~R~ Plan No. __~2~ Job No. Period Author: REPORT 15 Covered: July _ Mammals _ Deer Investigations 1 Research Compensatory Effects Mule Deer Population of Harvest in a 1, 1989 - June 30, 1990 R. M. Bartmann Personnel: T. A. Abbott, H. E. Burdick, L. H. Carpenter, K. M. Chociej, B. L. Dupire, J. H. Ellengerger, D. J. Freddy, J. P. Gray, R. Harthan, E. K. Jones, J. D. Madison, J. E. Morris, R. E. Thomas, C. L. Teter, G. C. White, and numerous others from the Colorado Division of Wildlife and Colorado State University. ABSTRACT Eighty-four fawns and 49 adult female mule deer were radio collared on the control and treatment units of the Ridge study area. Preliminary estimates of over-winter fawn survival for 1989-90 were 0.667 (SE 0.096) and 0.583 (SE 0.082) on the control and treatment units, respectively. Adult survival was 0.905 (SE 0.064) and 0.667 (SE 0.091) on the same respective units. All adult mortalities except 1 were hunting related. Sixty deer were checked during the 3 regular seasons: 22 bucks, 36 does, and 2 fawns. Twenty-_one percent of the does were yearlings and 15% were aged as ~8 years old. The harvest estimate for the antlerless late season in December 1989 was 451 de~r. Twenty-eight percent of the deer recorded at check stations were fawns. Of the does, 10% wereyeCirlings and 37% were aged as ;::8 years old, about re";erse of the composition for the regular seasons. Estimates of p~pulation size-from aerial line transects were 1,227 (95% CI 746 - 1,708) and 1,420 (95% CI 1,173 1,667) on the control and treatment units, respectively. Fawn:doe ratios on the same respective units were 63.2:100 (95% CI 55.2 - 72.1) and 105.9:100 (95% CI 89.8 - 125.0). ��189 COMPENSATORY EFFECTS OF HARVEST IN A MULE DEER POPLATION Richard M. Bartmann P. N. OBJECTIVES 1. Increase the winter survival rate of mule deer fawns by lowering total deer density to reduce competition for forage during winter. 2. Increase the harvest rate of deer through increased productivity of adult does and decreased natural mortality of fawns resulting from closer alignment of population size with carrying capacity. 3. Evaluate deer harvest rates and population response on a DAD basis resulting from bucks-only hunting and from buck hunting with additional antlerless permits designed for annual removal of 20% of the adult female population. SEGMENT OBJECTIVES 1. To reduce the winter popUlation of mule deer on the Ridge treatment unit to a density <40/km2 and maintain that density for at least 3 years. 2. To estimate units. winter survival rates of fawns on control 3. To estimate annual survival rates of adult 4. To estimate annual survival rates of adult males. 5. To estimate productivity 6. To estimate harvest treatment units. 7. To estimate 8. To estimate age structure of adult females on the treatment unit condition of yearling females on control and treatment units. 9. To estimate age structure of adult males males on control and treatment units. rates condition females. of deer on control of bucks, does, of fawns on control METHODS and treatment and treatment units. and fawns on control and treatment and condition and units. and of yearling AND MATERIALS Deer Trapping Mule deer were captured with dropnets on 12 sites 13 sites on the treatment unit of the Ridge study on the control unit and area during November �190 1989·. Fawns and -adu Lt; females were radio collared and released to estimate over-winter and annual survival rates, respectively (White et al. 1987). Adult males were not radio collared. Experience last year revealed it is not possible to achieve a reasonable sample size and the extra effort to monitor them detracts from monitoring fawns and does. All transmitters contained a mortality mode set with a 3-4-hour delay. Male fawn collars were a drop-off type so they could be retrieved after -8 months and reused. Female fawn collars were expandable so survivors could continue to be monitored as adults. Adult female collars were of fixed size and remained permanently attached. In addition, all fawns and yearling adults, whether collared or not, were tagged in both ears to identify them as known-age animals to help refine aging techniques. All deer were weighed (kg) and total body length (cm) and left hind foot length (cm) measured (White et al. 1987). Antler measurements taken on bucks included circumference (mm) 2.S cm above the burr, main beam length (cm), and number of points and brow tines ~2.S-cm long. During the middle of January, February, and March, all radio-collared deer were located by triangulation to detect permanent movement between the 2 study units. Survival Radio-collared fawns were monitored for survival -S days/week beginning immediately after trapping until they migrated in late April (White et al. 1987). From then until IS June, they were monitored from a fixedwing aircraft approximately every 2 weeks. Adult females were monitored once weekly throughout the winter and monthly through the summer. All radio-collared deer were monitored daily during the regular and late hunting seasons. Each mortality signal was checked as soon as possible to determine cause of death, using criteria described by (White et al. 1987), or to retrieve a drop-off collar. Harvest Check stations were found inefficient during the 3 regular fall deer hunting seasons. Instead, visits were made to hunting camps or hunters were contacted in the field to obtain data for deer killed on the 2 study units. Each deer was weighed (kg) if the carcass was intact except for normal field dressing and had been dead <3 days. The same antler measurements described for live bucks were also taken. Each deer was field aged by tooth replacement and wear if permission was obtained to slit the jaw, although yearlings could still be identified without doing this. The 2 middle incisors were also taken from animals field aged as ~2 years for aging with the dental cementum technique (Erickson and Seliger 1969). If a known-age deer was killed, effort was made to collect the lower jaw. A late season for antlerless deer was held on the treatment unit 1-31 December 1989 to begin reducing the population. There was a concern with hunter crowding because of the small size of the hunt area (-28 km2) and �191 the large number of-deer to be removed. to take 2 deer on 1 license. Therefore, hunters were allowed For ease of description, the hunt area was expanded to include -6 km to the west of the treatment unit and a l-km band along the south side of the Ridge road was excluded. Hunt area boundaries were the powerline road on the east, the Ridge road on the south, Piceance Creek road on the west and the White River on the north. All boundaries except the river were signed at -1/2-km intervals. In addition, signs indicating direction and distance to the hunt area were placed at strategic locations along access roads. Each successful applicant was mailed a packet with a brochure describing the background and goals of the study and other information about camping, access, weather, etc.. A map was also enclosed along with a survey/tooth mailer to get harvest data and deer teeth from hunters not contacted at a check station. Hunters could either mail them or deposit them in collection boxes placed around the area. A check station was operated daily during the late season from 8AM to 6PM at the Little Hills headquarters. A temporary station was operated on weekends near the mouth of Hay Gulch. Information taken for each deer was the same as during the regular season. To encourage hunters to stop at a check station or return their survey/tooth mailers, a souvenir patch designed for the late season was given to all that did so. No harvest estimate was possible for the regular seasons. Instead, the number of deer checked in the field was used as the minimum number of deer removed. An estimate of total harvest for the late season was taken from the random survey conducted by the Division's Limited License Group. This survey also provided estimates of the number of people that did not hunt, that were unsuccessful, that killed 1 deer, and that killed 2 deer. Check station data were used to estimate the fawn sex ratio and age composition of the harvest. Final age assignments were based on all available data with jaw aging as the default for small or unreconcilab1e differences between techniques. Population Size An aerial line transect survey was flown in early January after the late season to estimate size of deer populations on the 2 study units. The transects were established in 1985 and had been flown every winter since then. Twenty-five transects were oriented north and south and systematically spaced at 0.4 km across both study units. There were 10 transects on the control unit, 12 on the treatment unit, and 3 transects crossed both units. These latter 3 transects were split where they crossed unit boundaries and the segments considered separate transects for their respective units. Lengths of transects ranged from 1.5 - 5.6 km and averaged 4.0 km. �192 Line transects were flown by 2 obervers and a pilot in a Bell Soloy helicopter. The survey was flown again the next day, in the opposite direction to the first flight, to increase the sample size of deer groups. Observer calibration and counting methods and analysis procedures were as given by White et al. (1989). Sex and Age Ratios Aerial sex and age classification were made on the 2 study units between the 2 line transect surveys. Classifications were made along the same lines used for line transects except that deer groups within 0.2 km on either side of the flight line were classified to achieve complete coverage. Buck:doe and fawn:doe ratio estimates were based on groups rather than individuals to improve precision. RESULTS Deer Trapping Eighty-four fawns and 49 adult females were radio collared on the control and treatment units from 13-30 November 1989 (Table 1). The last 4 female fawns captured received drop-off collars. Table 1. Number of mule deer trapped and radio and treatment units, 13-30 November 1989. Unit Control Treatment Radio collared collared on the control Adult Fawn Male Female Yes No 19 21 Yes No 18 3 Male Female 21 3 26 3 28 18 Body Size Measurements.--Based on 3 measures of body size, male fawns averaged larger than females on their respective units (Table 2). Also, male and female fawns and adult females on the control unit were larger than their counterparts on the treatment unit. There were not enough yearling and adult males and yearling females to make similar comparisons. No significance tests were performed. Movements.--No permanent movements of deer between the control and treatment units was observed. Seven deer, 2 from the control unit and 5 from the treatment unit, were each located 1 or 2 times on the opposite unit, but all were also located at least once on their original units. The 7 deer were captured on 3 trapsites located 0.5-2 km from the boundary between units. �193 Table 2. Body size measurements of mule deer captured with dropnets on the control and treatment units, November 1989. Unit Age Control Fawn Fawn 19 21 34.1 3l.4 0.42 0.37 131. 6 130.5 0.56 0.52 42.2 4l.4 0.32 0.33 2 2 48.5 46.0 l.63 l.79 151. 8 148.4 2.02 l.06 46.7 47.9 0.59 l. 23 2 19 73.5 67.7 2.88 0.52 167.9 168.6 2.32 0.52 50.7 48.1 0.38 0.24 21 26 3l. 5 29.9 0.41 0.34 127.0 126.3 0.51 0.45 4l. 3 40.3 0.23 0.24 1 7 60.8 53.3 1.17 168.5 l58.7a l. 36 50.0 46.7 0.33 1 39 93.4 65.0 0.39 196.0 165.9 0.42 5l.0 47.8 0.18 M M M F Yrlg M F Adult M F a n = Left hind foot length {cm) x SE M F Treatment Total body length {cm) x SE n F Adult {kg) SE Sex F Yrlg Weight x 6. Survival The previous high over-winter fawn survival was 0.520 (SE 0.071) on the treatment unit in 1985-86. The 1989-90 winter was extremely mild and this rate was surpassed on both units in 1989-90 (Table 3). These high rates occurred even with hunting mortality occurring for the first time since survival monitoring began on the Ridge in 1982-83. Table 3. Fate and preliminary survival estimates of radio-collared mule deer on control and treatment units from time of trapping in November 1989 to 15 June 1990. Age n Control Fawn Adult 40 21 Treatment Fawn Adult 44 28 Unit Hunting related Undeter. 3 2 1 3 1 6 5 9 3 Starve Accid. Trap related 2 Still alive Sa SE( S) 14 16 19 0.667 0.905 0.096 0.064 8 21 18 0.583 0.667 0.082 0.091 Collar drop-off a Trap-related mortalities and pre-mature collar drop-offs excluded from calculations of survival rates. are There was a problem with pre-mature collar drop-offs on male fawns and the 4 female fawns fitted with this type of collar. The release mechanism, latex tubing, was the same as used last year when no similar problems were encountered. This year, there were 22 pre-mature dropoffs, 64% of them with fawns from the control unit. The last fawn mortality occurred on 17 April 1990. If the assumption is made that all �194 fawns with collars dropping off after that date lived survival rates increase to 0.758 (SE 0.075) and 0.643 control and treatment units, respectively. to 15 June, (SE 0.074) on the Adult female survival to 15 June 1990 on the control unit was slightly above the previous 6-year average of 0.874 for the entire Ridge area. However, this could still change as adult survival is measured to 1 December. In contrast, survival on the treatment unit was already below the lowest rate recorded on the Ridge (0.764, SE 0.065, in 1986-87)and was due to hunting, the sole mortality cause, during the late season Harvest Sixty deer killed on and around the 2 study units were checked during the 3 regular hunting seasons (Table 4). One, 21, and 38 deer were checked during the 1st, 2nd, and 3rd seasons, respectively. Twenty-two of the 60 were bucks, 36 were does, and 2 were fawns. All deer except 1 were killed on the North Ridge, but not necessarily on the 2 study units. The exception was an eartagged buck killed on the ridge south of the Little Hills headquarters. Since the regular season harvest cannot be estimated, 60 is the minimum number assumed killed. Table 4. Mule deer checked on and around the control and treatment units during 3 regular hunting seasons and a late season in December, 1989. Season Regular 1st 2nd 3rd Control unit Buck Doe Fawn Treatment unit Buck Doe Fawn 1 1 7 8 4 146 Late a Antlers antlerless. 5 were <5-inches long and were 1 Not recorded Buck Doe Fawn 2 18 7 5 1 57 legal under the definition of The harvest estimate for the late season was 451. A breakdown of survey results shows 21.7% of the 375 license holders did not hunt, 10.0% were unsuccessful, 16.7% killed 1 deer, and 51.7% killed 2 deer. thus, slightly more than 500 deer were killed during all seasons, but this does not include any adjustment for wounding loss or illegal kill. Two radiocollared deer from the control unit, a fawn and a doe, were killed during the late season, but it is not known which unit they were on at the time. Two yearling bucks were killed, but their antlers were <5-inches long and they qualified as antlerless deer. Yearlings comprised 50% of the bucks checked during the 3 regular seasons (Table 5). Of the does, 21% were yearlings and 15% were estimated as ~8 years old. During the late season, 28% of the deer checked were fawns. Ten percent of the does were yearlings and 37% were aged as ~8 years old; reverse of the composition during the regular seasons. The hypothesis of constant age ratios across seasons was rejected (f = 0.050) and implies 1 �195 or both of these ratios do not reliably estimate the true population structure. Because of the intensity of hunting and the favorable snow conditions during the late season, age ratios for this season were assumed closer to truth. Table 5. treatment Season Regular Late a Estimated ages of mule deer checked on or near the control units during the regular and late hunting seasons, 1989. Sex Male Female 2 Male Female 29 28 Antlers were <5-inches of ant1er1ess. A e 4 5 6 7 ~8 3 1 5 1 2 2 5 13 21 7 9 54 1 2 3 11 7 4 6 3 3 2 9 17 Fawn and 2a 15 long and so were We compared age ratios derived from jaw aging and dental cementum when both were available for the same deer. Yearlings were excluded as incisors were usually not collected from j aw- aged deer and there were no corresponding jaw ages for mai1in teeth. The 2 age distributions in Fig. 1 were independent (f = 0.007). Compared to jaw aging, dental cementum placed a higher percentage of animals into younger age classes (2-4 years) and fewer into older age classes (~8 years) •... :z UJ (,) :: 1 JAW AGE the definition DENTAL CEWENTUW AGE 30 a: 20 UJ Q.. legal under i ._----------------- 10 AGE Fig. 1. Age composition of the doe harvest on the control and treatment units during all 1989 seasons as indicated by jaw aging and dental cementum. During the late season, most deer were killed early in the month and on weekends (Fig. 2). Opening day was on a Friday and helped spread the opening weekend pressure over a 3-day period. Fig. deer unit late marks 2. Distribution of the harvest on the treatment during the December 1989 season. The heavy tick indicate weekends. ""_, '"_, -'"' "" •.... '" X I:> 1 0 '" DAY �196 There was a significant difference in field-dressed weights of does between the regular and late seasons (f = 0.044), ages (f < 0.001), and the season x age interaction (f = 0.026). Yearling females killed during the late season were smaller than those killed during the regular seasons. Weights of mature females were nearly the same, but were confounded by the different age compositions between seasons. Population Size Population estimates from aerial line transects were 1,420 (95% CI 1,173 - 1,667) on the treatment unit and 1,227 (95% CI 746 - 1,708) on the control unit. The survey was made immediately after the late season and, therefore, reflects harvest removals from the treatment unit. For an unknown reason, the estimate on the control unit is much lower than the 1,800 - 2,00 deer estimated the past few years. We were unable to refly these transects to see whether or not this might have been a sampling artifact. Sex and Age Ratios Buck:doe ratios on the control and treatment units were 9.8:100 and 6.7:100, respectively. Both were lower than the 12.4:100 estimated for the entire Piceance Basin (Northwest Region, unpubl. data), but all confidence intervals overlapped. Fawn:doe ratios on the control and treatment units were 63.2:100 (95% CI 55.2 - 72.1) and 105.9:100 (95% CI 89.8 - 125.0), respectively. The ratio on the control unit is similar to the estimate for the entire Piceance Basin (65.7:100, 95% CI 61.3 - 70.3) (Northwest Region, unpubl. data), but both were lower than the ratio on the treatment unit. The reason for the high ratio on the treatment unit is unknown. The differential harvest of does and fawns would have increased the ratio slightly above that on the control, but not to this extent. LITERATURE CITED Erickson, J. A., and W. G. Seliger. 1969. Efficient sectioning of incisors for estimating ages of mule deer. J. Wildl. Manage. 33:384388. White, G. C., R. M. Bartmann, L. H. Carpenter, and R. A. Garrott. Evaluation of aerial line transects for estimating mule deer densities. J. Wildl. Manage. 53:625-635. _____ , R. A. Garrott, R. M. Bartmann, L. H. Carpenter, and A. W. Alldredge. 1987. Survival of mule deer in northwest Colorado. Wildl. Manage. 51:582-589. Prepared by 1989. J. � https://cpw.cvlcollections.org/files/original/f2cec1f683159cb7831b16a5b25d465e.pdf 9d2e1faa1745ec56d41d025b4c576640 PDF Text Text I Colorado Division of Wildlife Wildlife Research Report September 1990 JOB PROGRESS REPORT State of Colorado Project 01-03-212 Work Plan 1__ Job Migratory 17 Job Title: Habitat use by wintering Period Covered: Authors: Game Bird Research mallards along the Front Range of Colorado 1 April 1989 through 31 March 1990 James K. Ringelman and Michael R. Szymczak Personnel: R. Falise, Colorado and Cattle Company State University; L. Roberts, Front Range Land ABSTRACT No additional field work was conducted this segment. Project objectives can be attained with existing data on wetlands classification and availability, weather, habitat selection by radio-marked mallards, bird movements and behavior,. and effects of hunter disturbance. Detailed data summaries of the 1986-87, 1987-88, and 1988-89 field seasons are presented in October 1988 and October 1989 annual reports. These data will be analyzed in greater detail for inclusion in technical journal publications to be completed next segment. ��3 Colorado Division of Wildlife Wildlife Research Report September 1990 JOB PROGRESS REPORT State of Colorado Project 01-03-212 Work Plan 1__ : Job Migratory 18 Job Title: Winter survival and reproductive Period Covered: Authors: Game Bird Research success of female mallards 01 April 1989 through 31 March 1990 James K. Ringelman and Michael R. Szymczak Personnel: D. Anderson, Colorado Cooperative Fish and Wildlife Researc~ Unit; C. Jeske, Colorado State University; R. Hopper, J. Ringelman, M. Szymczak, Colorado Division of Wildlife; M. Nail, R. Schnaderbeck, U.S. Fish and Wildlife Service. ABSTRACT Authorization to bait-trap mallards on the Monte Vista National Wildlife Refuge was denied, consequently no new field work was conducted this segment. However, comprehensive analyses were completed on three years of field data. Logistic regression analysis was used to investigate relationships between body condition of mallards and winter through spring survival rates. Potential effects of radio-marking were also evaluated. Body mass and condition indices were compared to other mallard populations, among age and sex classes of mallards in the San Luis Valley, and between years for individuals captured during successive field seasons. Relative recovery rates of wing-banded mallards were examined in relation to body condition, age, and sex of the bird. Wing remains were used to partition mortality attributable to starvation, avian cholera, and lead poisoning. Basal energetic requirements were deduced from oxygen respirometry experiments. "Energy acqui~ition was quantified using field-feeding trials coupled with data on metabolizable energy of common foods. Trace element analysis of down feathers was performed using feathers of known origin, but did not prove effective for distinguishing between migrant birds and those that were resident in the San Luis Valley during winter. Graduate Research Assistant Clint Jeske is preparing a dissertation that will serve as the final report on this project. The dissertation will be finalized in December, 1990, and reported next segment. ��5 Colorado Division of Wildlife Wildlife Research Report September 1990 JOB PROGRESS REPORT State of Colorado Project 01-03-212 Work Plan 1__ : Job Job Title: Evaluation Period Covered: Authors: Migratory Game Bird Research 19 of Nesting Habitat Management for Ducks 01 April 1989 through 31 March 1990 David W. Gilbert. James K. Ringelman. Personnel: D. Anderson, Colorado Cooperative Fish and Wildlife Research Unit; J. Ringelman, M. Szymczak, Colorado Division of Wildlife; D. Gilbert, . Colorado State University; S. Brock, S. Burlinger, M. Nail, R. Schnaderbeck, U.S. Fish and Wildlife Service. Abstract Personnel on the Monte Vista National Wildlife Refuge (MVNWR) have used transects to collect information on nesting ducks every year since 1964 (except 1977). Data on species composition and nesting success of 4,074 duck nests located over 24 years provide a basis to evaluate the benefits of habitat management practices including grazing, water application, burning and predator control. Work performed this segment focused on identifying and obtaining information on management practices, and preliminary exploration of duck nesting data. Mallards composed 54% of the nesting duck population on the MVNWR, followed in abundance by teal species and pintail. Units 3, 9, and 18 produced more ducks than other areas, but units 9 and 18 increased in relative importance when nests were expressed on a density basis. Most nest failures were caused by predators (21.7%) or desertion (12.3%). Baltic rush was the vegetation most commonly associated with nest sites; whereas few nests were found in sedge, spikerush, or cattail. Because detection rates of nests decreased with increasing distance from the center line of the transect, line transect theory is being used to derive detection functions that will enable unbiased estimates of nest numbers. Two different cattle grazing practices were employed on 57 pastures during 1964-88. Hay was also cut on 14.4% of the refuge beginning in 1964, but was phased out by 1976. Predator populations 'and control efforts were conducted throughout the period, but few quantitative data are available. Records of water application are more complete, but because of the varied water sources and methods of transport and application, data will have to be treated on a refuge-wide basis. Prescribed burning was initiated in 1981, with 12 burns conducted on -5,200 acres during the 1980's. Aerial photographs of the MVNWR have been obtained for 1969, 1978, and 1985. Vegetative associations will be delineated from these photographs and digitized in preparation for analyses of vegetation change and relationships to duck nesting biology. ��7 EVALUATION OF NESTING HABITAT MANAGEMENT FOR DUCKS David W. Gil bert James K. Ringelman P. N. OBJECTIVES 1. Relate nesting duck species composition, density to habitat management practices. nest success rate, and nest 2. Assess changes in wetland and upland vegetation between 1962 and 1985 by contrasting digitized habitat information derived from aerial photographs. 3. Determine nesting habitat preferences of duck species by comparing of nesting habitat with relative habitat availability. usage SEGMENT OBJECTIVES 1. Conduct an extensive literature review on the effects of grazing, burning, herbicides, predator control, early water application, and vegetative characteristics on nesting waterfowl. 2. Verify the correctness and completeness of nesting transect data already entered on computer. Enter data into spreadsheet software, then conduct preliminary analyses on species composition, nest success rates, nest density, etc. by year and wetland unit. 3. Locate records on grazing, burning, and if possible water application and predator control for individual wetland units on the Monte Vista National Wildlife Refuge (MVNWR). Computerize these data for entry as independent variables for categorical data analyses. 4. Obtain aerial photographs (appx. 1:6,000 scale) of the MVNWR. vegetative associations, and verify accuracy of classifications truth surveys. Determine the area of vegetative types. 5. Complete annual Delineate by ground report. STUDY AREA Nesting data used in this study originate from the MVNWR, located 6 miles south of the town of Monte Vista (e1ev. 7,600 ft.) in the San Luis Valley (SLV) of south-central Colorado. The SLV is best described as a cool desert, receiving less than 12 in. of precipitation annually. Therefore the agricultural communities largely rely, as does the refuge, on the irrigation water supply. The vegetation on the refuge is predominantly baltic rush (Juncus balticus), greasewood (Sarcobatus vermiculatis), and rabbitbrush (Chrysothamnus spp.). Szymczak (1986) provides additional descriptions of the MVNWR. �8 The surrounding mountains (San Juan and Sangre de Christo ranges) collect snow and subsequently supply water to the SLV via the Rio Grande River and a network of irrigation canals. Underground aquifers are also an important supply of water for the SLV and the MVNWR. The refuge began to use these water sources to produce wetlands in the early 1950's. Today a network of pumps, ditches, and dikes keeps 22 of 24 (total 14,189 ac.) individual nesting management units flooded for waterfowl use (units 12 and 13 are never surveyed for nesting waterfowl and will not be mentioned in reporting). The MVNWR offers a wide variety of nesting habitat. Generally, the western side of the refuge supports an upland greasewood/rabbitbrush association interspersed with numerous impoundments. A band of agricultural fields exist along a north-south line with the refuge headquarters (Fig. 1). Grasslands and wetlands continue eastward until near the eastern edge of the refuge, where greasewood/rabbitbrush again prevails. INTRODUCTION TO METHODS Beginning in 1964, an annual waterfowl production the MVNWR. This survey was to provide the following: 1. 2. 3. survey was initiated on A representative, statistically adequate sample of the duck hatch as a basis to project duck production. Information on the location, cover types used, and success of nesting waterfowl throughout the refuge for use in evaluation of ecological succession of vegetation and related habitat management practices. A consistent record of long-term population trends of other wildlife species important to the refuge. The survey design resulted from pre-sampling in 1961-63. It was calculated that 640 acres of transect would be required to estimate production within 10%-15% with a 95% level of confidence (MVNWR survey instructions). Originally (1964-68) 142 "strip" transects were used to estimate the parameters of interest. At 16.5 feet wide (8.25 ft./side), 5.5% of the refuge was sampled (640 acres of 11,570 available for nesting, or 320.5 miles of transect). No correction for missed nests was used. Nesting surveys were scheduled for mid-May, mid-June and July. June surveys were used to verify the fate of nests initially located in May, as well as to locate any new nests. The July survey was included to determine fate of nests found in June. Nesting parameters were collected at the time of the initial survey and included date, unit number, transect number, species, incubation stage, number of eggs (also number of parasitic eggs) and vegetative cover type in which the nest was located. Fate was determined and causes for failure were recorded on subsequent surveys. Other information was collected sporadically, such as nests of species other than ducks, distance to standing water, overhead cover, and distance to the centerline of the transect. These data were valuable but rarely complete or consistent. Beginning in 1969 the number of transects were halved due to manpower limitations. In order to not substantially decrease the area sampled, transect width was increased to 24 feet (12 ft./side) and all nests located were included for use in density estimates. The area covered then was 4.0% of total nesting habitat. The 1969 Annual Narrative Report indicated that the new method would be evaluated (i.e. estimation of the percent of nests �:x: 0 I-0:: z •• w (.!). =:J L1.. W 0:: L1.. W •..... _J 0 _J :::= ...... _J e:( z 0 •..... I-- z e:( e:( I-I/} •..... >w I-Z 0 :::i: I/} I-- •.....• Z =:J I-- z w =:: w (.!) e:( z e:( ::::: LI.. .,... 9 �10 actually found out to the distance of 12 ft.) at a future date. No evaluation of the 24 foot transect been found nor has there been any change in estimation procedures. Production was calculated from the number of nests found in the strip, mUltiplied by a factor to compensate for fractional sampling. Data were then corrected for species in the breeding population, species composition of the nests found, percent hatching success, and brood size for each species (see Appendix 1 for weighting calculations). However, no correction for missed nests at the increased transect width was used. Data used in sections of this preliminary report, such as species composition and success, were derived only from the nests actually found and recorded. When appropriate detection function curves are derived from data of the recorded distance from centerline, a correction factor will be used to accurately determine nest density by correcting for nests missed. At present, the approximate nest densities were calculated using actual data collected. No use was made of the annual production figures derived by MVNWR, which take into account brood size and breeding pair composition. METHODS Because this report contains analysis of data not derived from a controlled experiment, it is important to point out what records will be used and the source of information. Effort has gone into the collection of appropriate records as opposed to the traditional collection of new field data, as is standard in most research. Records Search Nesting Data. - Nesting and associated attribute data were entered on computer directly from actual field forms on file at MVNWR. Often the original field forms were recopied to summary sheets so most comments (for example, distance to centerline of the transect for nest relocation) were eliminated. Nesting information comes from the following nests: 2,180 222 571 104 354 90 553 Mallard nests, 53.5% Gadwall nests, 5.4% 14.0% Teal nests, Shoveler nests, 2.6% Pintail nests, 8.7% Redhead nests, 2.2% Unknown nests, 13.6% This totals 4,074 nests for the period 1964-88 (excluding 1977 when data was not collected); 1989-90 will be analyzed at a later time. One canvasback, one ring-necked duck and nine ruddy ducks were recorded over the period but were not included in the tabulation. Fourteen green-winged teal were reported but were grouped with all "teal", as teal species were not easily distinguished during nesting surveys. Management Data. - Records of management practices, (grazing, water application, burning, and predator control), elevated platform photography, vegetative cover mapping, and changes in transect methods were derived form Annual Narrative Reports kept at the MVNWR office in Alamosa. Water records �11 used to determine wet/dry cycles were obtained from gauging station records at Del Norte, Colorado (Colorado Water Resources, 1964-88). This station records flows prior to major diversions and is a good indicator of general SLV water conditions during a given year. SLV breeding pair information was provided by the Division of Wildlife (M. Szymczak, pers. comm.) These data were obtained to allow for detection of years where nesting densities may have been related to the overall SLV duck population as well as refuge conditions. Similarly, North American breeding pair survey records are included for additional comparisons (USFWS 1989). Ve~etation. - Original study plans called for comparative analysis of elevated platform photographs of unit 6 taken during the late 1960's (Fig. 1), with new photographs to be taken at the same location in 1990. A file was found that explicitly outlined the procedure and some locations of photo points, but no photographs were found. A pilot project to develop a standardized system to map vegetative cover types on the MVNWR preceeded the elevated platform photo work in the 1960's. Information found pertaining to this proposed cover mapping is similar to that described by Graham (1945). These files contained procedures outlining how to map the refuge. Preliminary work using this system was conducted in units 6 and 19 prior to wetland development by student trainee Leslie Beaty. This initial survey provides a 1962 "snapshot" of those units. Correspondence associated with those early files emphasizes the importance of future evaluations of cover conditions compared with baseline data collected in 1962. However, no further evaluations were located and it is doubtful that follow-up mapping occurred. Aerial Photographs. - A new methodology was employed to fulfill the objectives of the original photo/cover work. A search for a series of refuge-wide aerial photographs that could provide information on temporal change in refuge vegetation was undertaken. Suitable photographs were located for 1969, 1978, and 1985. These photographs will be ground truthed and features and cover attribute information will be digitized into the Geographic Information System (GIS) ARC/INFO. This methodology will allow accurate area determinations of each cover type by unit, as well as areas of ponds, dike roads, parking areas, fields, wetlands, and uplands. Such data are essential for accurate determination of nesting densities used for comparison between different management regimes. Initial cover mapping and polygon delineation will be performed using the 1985 photography. Later, plotter-produced mylar overlays will be superimposed over 1969 and 1978 imagery and polygons will be re-drawn were necessary to identify vegetation changes. These changes will then be digitized into separate databases to enable quantification of temporal change in refuge vegetation. Transect Data Correction For the purpose of this project, nest projection totals must be corrected for missed nests during the nest search procedure. Because all nests were not found at a transect width of 16 ft. (Anderson and Pospahala 1969), projection of total nests would be low if the number found were not adjusted. This error is multiplied when distance sampled was increased to 24 feet. Data on nest distance from centerline (measurements were available �12 1969-74, 1986, and 1987) will be used in program "Transect" (Burnham et. al. 1980) to develop a model that will fit the curve of falling detectability. A correction value will be determined from the appropriate model. Detectability differences between species and years will be evaluated and appropriate corrections will be made to account for changes in data collection over time. RESULTS Nesting Data The species composition for the 24-year period 1964-88 (1977 omitted) depicts relative increases in gadwall (Anas strepera) and teal species, and decreases in northern pintail (~ acuta) and "unknown" ducks (Fig. 2). The percent composition made up of mallards (Anas platyrhynchos) has remained relatively steady, except for a decrease in the late 1970's (Fig. 3). These data represent all nests found regardless of changes in effort among years. The highest duck nesting densities were found in units 3, 9, and 18, all units located directly east of the Empire Canal (Fig. 1). These units produced more ducks than eastern or western units (Fig. 4). Enright (1971), reported a two ft. drop in elevation in unit 18 from west to east. This compares to a 40 ft. drop in unit 19. Other units, such as 6 and 15, have moderate drops (approximately 20 ft.) but have extensive wetland development and are moderately productive. Shallow standing water is known for production of lush vegetation, a possible reason for the large number of nests in units with slight elevation changes. When data were corrected for the length of each transect (Fig. 5) or area of each unit (Fig. 6), units 1, 2, 6, and 8 increased in relative importance. Units 9 and 18 had the highest nest densities on the MVNWR. These data will be further refined to better reflect true nesting densities as detection function correction values are derived. Data may also change slightly when acreages are calculated from the GIS database. It should be noted that unit 24 was not searched until 1981; therefore caution must be exercised when comparing unit 24 to the other units. According to nests found, redheads (Aythya americana) showed a preference for units 3, 6, 7, 9, and 13, gadwall for units 3 and 6, and teal for units 3 and 18 (Fig. 7). Most nests were located in baltic rush, the dominant herbaceous vegetation on the MVNWR (Table 1), although greasewood often provided shrub cover in the vicinity of the nest when they were located on hummocks above the high water line. Few nests were found in sedge, spikerush, or cattail. Most nest failures were caused by predators or were deserted for unknown reasons (Table 2). The overall success rate of 4,074 nests, uncorrected for exposure, was 52.8%. Records Search Grazing. - Records of cattle grazing on the refuge have been located being computerized. Two grazing strategies were employed. The first for grazing after the nesting season (mostly during fall) on virtually entire refuge annually. Later, there was a reduction in total grazing and are called the and an �Z 100 o - 100 U5 o a.. 80 80 ~ o o 60 60 J- 40 40 20 20 r-t0_"r-r-l J- Z W () a: W a.. o o mallard li!l!ili!ll!:illl unknown 1°:°::::::1 teal 1::,::::1 pintail '85 '80 '75 '70 '65 1»1 gadwall ~ shoveler D redhead Fig. 2. Species composition of duck nests found on the Monte Vista National Wildlife Refuge, 1964-88. Data for 1977 not included. Composition is uncorrected for missed nests. I-' W �14 ~r-------------------------------------~~ Malard 110 . r---------------------------------------, 14 .............•...•.......•.... 12 10 ..••••..•.•••••............... 10 8 ~ 14 12 - •••................ e -...... ..- ................•..... . 4 Yerl Yerl ,-------------------------------------~~ Teal Rect>ead 8 30 .....••....••....•....................•••••••......•. 20 ....•.•••••••••.........•••...•.......•••.•...•...•••••• •••••.•.....•.••••••.••••••••••• ..........•••••.....................................................•...•....••••••....•......... 30 " 3 . . . . 10 Yerl ~,-------------------------------------~~ 20 ...•.........••.....•.••.•.........•••.••.........•.....••.....•••••........•••••.•.......••.... 20 15 ...........................••••................................ 15 10 ........•........... . Shoveler 5 •. -----.-- ...•. --.- 4 .•....••••••••••••••••••. 3 - .••.••.•. ------. ... - .• - .. ------.-.- •. --.-.- . .........•••..........••........•.•••.....•••••••••••• 10 2 . Yest: 30,--------------------------------------30 Urknown ..........•..•. ~ ... _............. •..•.•.............•..... . ...•.•....••.•....•.....• -............. 20 15 ..........................••••..•........ 15 10 .••••••.................................. 10 Fig. 3. Temporal change in species composition for principal nesting ducks on the Monte Vista National Wildlife Refuge, 1964-88. .•. ---- .. -...... .-. 5 �15 1 2 3 4 5 6 7 8 9 10 11 14 15 16 17 18 19 20 21 22 23 24 Unit Number Fig. 4. Total duck nests found by unit, 1964 through 1988, uncorrected for differences in effort. Data from 1977 not included. 70 eo 50 ~ c: Q) 40 ::::I c:r e 30 U. 20 10 1 2 3 4 5 6 7 8 9 10 11 14 15 16 17 18 19 20 21 22 23 24 Unit Number Fig. 5. Total nests found by unit, 1964-88, corrected by the distance of each transect. Data from 1977 not included. 0.8 ~ c Q) o.e ::::I c:r •.. Q) U. 0.' 0.2 1 2 3 4 5 6 7 8 9 10 11 14 15 16 17 18 19 20 21 22 23 24 Unit Number Fig. 6. Total nests found by unit, 1964-88, corrected for the area of each unit. Data for 1977 not included. �I-' o- Mallard D Gadwall Teal ~ Pintail ~ Shoveler ~ Redhead 30 +oJ C Q) o 20 ~ Q) D... 10 o I 1lt111l1JIlIdi 11m-.I 1 2 3 4 5 IIUl 11m•• n "HI 6 7 8 9 IIlil IIIN 11m "Ib! "IJ "11 "rN "" 10 11 14 15 16 17 18 IIIdI 19 20 lin "'I 21 22 -b uu 23 24 Unit Number Fig. 7. Species composition of nests found by unit, 1964-88 (1977 excluded). , �17 every-third-year rotation plan was implemented. The rotation system was intended to remove most of the dense residual cover every third year (Mel Nail, MVNWR, pers. comm.) A total of 57 pastures were grazed during 1964-88, but not all were completely grazed and total ADM's differed considerably among years (Fig. 8). The relation of grazing treatment to nesting densities will be evaluated in the future. Table 1. Vegetation associated with 4,074 nests located National Wildlife Refuge, Colorado, 1964-88. Dominant plant species Number Sedge (Carex spp. ) on the Monte Vista Percent of total 55 l.4 Saltgrass 130 3.2 Spike rush 61 l.5 146 3.6 2,761 67.8 619 15.2 78 l.9 3,850 94.6 Grasses Baltic rush Greasewood Cattail Totals Table 2. Colorado, Sources of nest failure on the Monte Vista National during 1964-88. N = 4,074 nests. Cause of failure Number Wildlife Percent of total Deserted 502 12.3 Predator/Eaten 886 2l. 7 Flooded 126 3.0 Unknown 480 11.7 1,994 48.7 To t a.I s" Totals differ from the summation more than one cause recorded. a of individual causes because Refuge, some nests had �18 .....................................................................................................•............................... 6,000 5,000 ~ :J « co +-' r--0 4,000 3,000 2,000 1,000 0 , 70 '65 '75 '85 '80 Year Fig. 8. Total cow-calf units (AUM) grazed on the Monte Vista National Wildlife Refuge, 1964-89. 2,000 "0 1,500 Q) ~ I * = no records Q) OJ co .._ 1,000 Q) o « 500 o .................................................•.•.. * '65 , 70 * ,75 Year Fig. 9. Total hay produced on the Monte Vista National Wildlife Refuge, 1964-76 ·········· ·.. 1 �19 Haying. - At onset of this project we were unaware that haying ever occurred on the MVNWR. Haying can have detrimental effects on nesting success depending upon chronology of flood irrigation and hay cutting. In 1964 hay was cut on 14.4% of the refuge (1,665 acres). Subsequently the acreage was gradually reduced and finally eliminated (Fig. 9). Haying was restricted to units 4, 7, 14, 20, 22 and 23 (Bill McDermith, MVNWR pers. comm.). We could not locate records of the amount and acreage hayed by unit or information on whether a particular unit was hayed in a given year. Units in which haying was know to occur will be treated differently when appropriate in future analyses. Predator control. - Predator control data has been assembled but have not been analyzed. From the annual narrative reports it is apparent that control varied widely, as did predator populations. Poisons were used in the 1960's, and were reportedly effective. However, no quantitative data were available. Trapping effort is also missing for most years. Water application. - The application of water is of utmost importance to waterfowl on the MVNWR. Anderson (1967), Pospahala (1969), and Robinson (1971) demonstrated the relationship between early water application, vegetation characteristics and nesting densities on the MVNWR. Schroeder (1973) demonstrated> 2-fold increases in densities of aquatic invertebrates in areas with early water application. Summer water is equally important for brood rearing. Since the work of the above mentioned authors, early water application has been a major objective at the refuge. (Bill McDermith MVNWR, pers. comm.). Total water use on the MVNWR was stable during 1964-73, decreased during 1974-77, then increased markedly thereafter (Fig. 10). The increase reflected increased wetland habitat developed during the 1980's. Unfortunately, potential benefits of water distribution are easier to document than the actual water distribution over a large area over time. Available water records indicate the actual amount of pumped and canal water available by year. Total water in the Rio Grande at Del Norte, Colorado is a general indicator of SLV water conditions (Fig. 11). However, there are a myriad of ways water can be applied to different areas of the refuge through a complex network of ditches. Winds, winter precipitation, winter ice buildup, and timing of the spring thaw all play a role in how successful the refuge is at water application. Further complicating this picture are the> 200 artesian wells located throughout the refuge that vary in size, depth, and pressure, and flow unmetered at different amounts in different years depending on local water table conditions. It will therefore be impossible to determine the exact timing or amount of water applied on a per unit basis. Therefore, subsequent analyses will examine refuge-wide duck recruitment parameters in relation to overall water application. Relationships may exist among SLV water conditions, MVNWR water conditions, SLV breeding pair numbers and MVNWR nest densities. For instance, in a dry year, a larger proportion of the breeding ducks available in the SLV may be attracted to MVNWR, where water is artificially applied than would be the case in a wet year when SLV wetlands are full. Yearly changes in numbers of breeding pairs on the MVNWR generally reflected Valley-wide changes for all �20 35,000 1- 30,000 "'0 c . 25,000 0 o Q) en -;:,. 20,000 Q) Q) LL o 15,000 ..0 :::J o 10,000 5,000 o , 65 '70 , 80 '75 '85 Year Fig. 10. Total water used on the Monte Vista National Wildlife Refuge from all sources, 1964-89. 600,000 500,000 "'0 c 0 o Q) 400,000 en -;:,. Q) Q) LL U ..0 :::J o 300,000 200,000 100,000 0 '65 '70 '75 , 80 '85 Year Fig. 11. Rio Grande River water totals as an index of San Luis Valley water conditions. �21 species combined (Fig. 12) and mallards (Fig. 13). Mallards comprised a l~wer percent of the total SLV breeding population according to aerial surveys (z 34.3%) than nests found on MVNWR (K = 53.5%) over time. There was a low correlation between the number of SLV mallards and the number of North American mallards, total ducks, and total ducks other than mallards (r-square = 0.149, 0.187 and 0.099 respectively). The SLV is apparently very important to nesting ducks and the MVNWR is becoming an increasingly important portion of the SLV. Habitat changes Valley-wide have reduced the number of breeding pairs off refuge, yet refuge numbers appear relatively stable. Further analyses will focus only on changes within the MVNWR in relation to habitat conditions and management. Prescribed burning. - Prescribed burning (ditches excepted) was not used at the MVNWR as a habitat manipulation procedure until 1981, according to records located to date (Table 3). The 1980 Annual Narrative discusses initiation of training procedures, acquisition of fire control equipment, and beginning of a fire use strategy. Up to this time, the only references to burning were passages in annual reports that stated: "no prescribed burns occurred in calendar year 19 ,nor do we feel the need for such management". Burn~ng Table 3. Prescribed Refuge, Colorado. burns conducted on the Monte Vista National Wildlife Size (acres) Year Unites) 1982 4 250 1983 9 unknown 1983 19 180 1984 14/15 560 1984 17 450 1984 3/4 600 1984 18 400 1986 unknown 320 1988 unknown 25 1989 6 640 1989 8/17 880 1989 9/10 900 �22 MVNWR 1,;,;,;,:-:-;;·;·1 30,000 - SLY .............................................. = _ ··············· · .._ o, .._ ··· ·· ·..·..1 CJ) "tlj Q) ..Q 20,000 ········ * no ·••········..······ count · ··..·····..·······1 E ::::J Z 10,000 '65 '70 , 80 '75 '85 Year Fig. 12. Duck breeding pair estimates for the San Luis Valley and Monte Vista National Wildlife Refuge, 1964-89. I 15,000 I MVNWR - SLY ...................••.......................................................................................................................................................................... .._ CJ) "tlj n, .._ Q) * no 10,000 ..Q E ::::J Z 5,000 o count •...••...........•..................................................................................................................... ~LULULilLU~~-=~.u~~~UL~~~~~~u.~u_ '65 '70 __ '75 '80 , 85 Year Fig. 13. Mallard breeding pair estimates for the San Luis Vallay and the Monte Vista National Wildlife Refuge, 1964-89. �23 200 ~---------------------------------- 150 ~ ............................................................................................................ ·······································1 'Q) ...Q E ::::J 100 Z 50 o o 2 4 6 8 10 12 Distance from Center (ft) Fig. 14. Detectability function of 1,465 duck nests found on the Monte Vista National Wildlife Refuge. �24 began at neighboring Alamosa NWR for control of common reed (Phragmities australis) and was increasingly used thereafter at both refuges into the 1980's. Typically, burning followed the third year of grazing and was described to completely clean a given unit of dense residual cover. Eleven prescribed burns have been conducted on the MVNWR since 1982 (Table 3.) Sampling Correction Measurements of the distance from the transect line were obtained for 1,465 nests. Neither duck species or personnel changes (i.e., year-to-year variation) contributed to variation in the detectability function. Visual inspection of the nest detection data (Fig. 14) suggests a nearly linear decrease in the detectability of nests as a function of distance. However, no particular model has yet been selected as the "best fit" model to use in detection function equations necessary to correct for missed nests. DISCUSSION The species composition and nesting densities of ducks on the MVNWR has fluctuated dramatically since 1964. Variation of comparable magnitude was also apparent in the number of nests found over time within individual management units. Detailed analyses of these relationships will have to wait until sampling corrections and vegetation analysis is performed. On an individual unit basis, rarely are there enough nests to consistently detect statistical differences in densities within years (typically N ~ 10 for any unit within a year). The actual number of nests found was also dependent on weather conditions, the size of the breeding population, the species composition of the breeding population, and the availability of water in the SLV and on the MVNWR. Thus, to determine cause and effect relationships between nesting and management practices it will be necessary to pool data from units with common treatments. These analyses will be the primary focus of work to be conducted next segment. Literature Anderson, D. R. and habitat. Cited 1967. Effects of water manipulation on waterfowl M.S. Thesis, Colorado State Univ., Ft. Collins. Anderson, D. R., R. S. Pospahala. studies of immotile objects. production 60pp. 1970. Correction of bias in belt transect J. Wildl. Manage. 34:141-146. Burnham, K. P., D. R. Anderson and J. L. Laake. 1980. Estimation of density from line transect sampling of biological populations. Wildl. Monogr. 72:1-202. Colorado Water Resources. 1964-1987. Colorado stream flow information. �25 Enright, C. A. 1971. An analysis of mallard nesting habitat on the Monte Vista National Wildlife Refuge. M.S. Thesis, Colorado State Univ., Fort Collins. l13pp. Graham, S. A. 1945. Ecological Manage. 3:182-190. classification of cover types. J. Wildl. Monte Vista National Wildlife Refuge. 1964-1989. Annual Narrative Reports and miscellaneous files. Alamosa National Wildlife Refuge Files. Alamosa, Colo. Pospahala, R. S. 1969. Waterfowl production and vegetative water. M.S. Thesis, Colorado State Univ., Ft. Collins. response 96pp. to early Robinson, G. W. production. 1971. Vegetation and physical factors influencing waterfowl M.S. Thesis, Colorado State Univ., Fort Collins. l48pp. Schroder, L. J. production. 1973. Effects of invertebrate utilization on waterfowl M.S. Thesis, Colorado State Univ., Ft. Collins. 44pp. Szymczak, M. R. 1986. Characteristics of duck populations in the intermountain parks of Colorado. Colorado Div. Wildl., Div. Rep. 6. 13pp. u.s. Fish and Wildlife Service and Canadian Wildlife of waterfowl and fall flight forecast. 39pp. Service. 1989. Status U.S.D.A. Soil Conservation Service. 1980. Common Colorado Range Plants. Soil Conservation Service., Denver Co. 55pp. Prepared by: ~) David W. Gil bert Graduate Research Assistant C!fo-.2<~ James K. Ringelman Wildlife Researcher C �26 Append; x 1 Forms outlining estimation of production, MVNWR -:.--nTw::;~'!;:r;:,-___, riW"J~~o~~~1J No. ne3t. Frail. Spa COIlP_ Known flea tIS I forll !P..e:ks __ .... !_c;-,V_ .J-.n. :;eoi~bt <Col. e_~ 3 .. Z No. Bre.dins Duck~ (C) __ ~~,,'r.'::l .\~?t;:_--ll~ Sp. COQP. 1n .~ . St'_ W.is~t Coail_ <Col. 6 _1~_;;'_.__!,. 1)_ ·J.l~ht Sp.Ccop. 11'\ see , (Gol. ~ (~) (Col. 9 ..:.£.2.!. •.2l!-_) ___. .!'~l._ -. -- ~}~~- nn • ----- r~ntall 1-, Gad""3-11 I - - I ShoVf!ll~t" ; - I Clnn.T~al ~:.:t. Te~l f -- - .- !\~dh~a.d i !lU~~T ' -(Total) -I---, Iii _L_ ~-,,!!,wn ~i- 1 - I-- Coot feria t I TCT"L - -'-t--. r---~-- I· --._.----r---·· --t--+-·_-j--_·r-I I ...~.- ..-tl . ~ _L__L__l __~L_ D-6. (10) Sp. Coap. in tlos. Specie.:t _ ( 12) (11) Jatching ne , Ne e t e Hatched on Tr"an8~c Success ta 1n ~ (Col. 9 (for_"~e) for" x~t~o~t.~l~~~-'~ __ (££!::l!) +-~Col. 2 fluddy '1---1-1--1 --L--+---t---r--+-~Un:<now~n ~-=.l=~ ± L- ..--·~1:~_~~_L_~J__ I _ ..~'."'" =+===l====i====1==-= AL I TO:r =c.~ot,------,-_-"-- .. __ -_I 4= ~ = -I .~ �27 Colorado Division of Wildlife Wildlife Research Report September 1990 JOB PROGRESS REPORT State of Colorado Project 01-03-212 Work Plan 1__ : Job Job Title: Development in Colorado Period Covered: Authors: Migratory Game Bird Research 20 and Evaluation of Moist-soil Management Techniques 01 April 1989 through 31 March 1990 James K. Ringelman and Michael R. Szymczak Personnel: R. Brown, R. Hopper, S. Hoover, T. Ostertag, J. Ringelman, M. Szymczak - Colorado Division of Wildlife; L. Roberts - Front Range Land and Cattle Company; C. Crosby - Big Meadow Farm, Inc. ABSTRACT Technical and popular literature was reviewed for information on moistsoil management techniques applicable to Colorado. Wildlife researchers in Missouri and Texas responded to letters of inquiry and provided details of ongoing moist-soil studies in those states. Personnel in Kansas and adjacent states were consulted during a waterfowl habitat management workshop hosted by the State of Kansas. Progress on the "expert system" computer software for moist-soil management was reviewed with personnel of the U.S. Fish and Wildlife Service. Based on this information, eight criteria were developed for screening potential sites to construct experimental moist-soil impoundments. Four candidate sites were considered, and field surveys were conducted at each site. One site, the Wellington Wildlife Management Area northeast of Fort Collins, was deemed acceptable for impoundment .construction. Barring unforeseen complications, impoundmeni: ':designand c'onst ruc t i.on will begin at Wellington ih fall 1990, with a sprihg 1991 completion date~ High soil permeability, timing of water availability, amount of water available, and cost of water are potential limitations to moist-soil management in Colorado. Most soils near eastern plains riparian areas are sand or sandy-loam, which make it difficult to maintain water and moist-soil conditions. Ditchwater, the most cost-effective source of water for impoundments, is generally unavailable until mid-May. Conventional moist-soil techniques would require that water remain impounded over winter "or be applied in March or April in preparation for May drawdown. However, these and other limitations are common throughout eastern Colorado, therefore any suitable moist-soil management techniques must work within these cons t ra lnt s if such management is to be widely adopted. Work next segment will evaluate the feasibility of moist-soil management given limitations on water availability. ��29 Development and Evaluation of Moist-soil Management Techniques in Colorado James K. Ringelman Michael R. Szymczak P. N. OBJECTIVES 1. Formulate a list of desirable moist-soil plants likely to benefit waterfowl in Colorado, as well as a list of potential problem species techniques for their management. and 2. Develop prescriptions for moist-soil management in eastern Colorado, detailing time of water application, duration of inundation, and water quality and substrate characteristics necessary for establishment of selected plant species and control of nuisance species. 3. Evaluate waterfowl responses to moist-soil units, including rate of food acqu~s~t~on, plant species preference, differential usage by waterf~wl species, and seasonal use patterns. SEGMENT OBJECTIVES 1. Review literature on moist-soil management practices, and consult with experts to gain insight into management techniques applicable to Colorado. Determine the status of the prototype expert system for moist-soil management being developed by the U.S. Fish and Wildlife Service, and evaluate the feasibility of applying this system to moist-soil units in Colorado. 2. Design moist-soil management experiments based on information obtained in approach 1, including plant species objectives and number of replicates/experimental treatments needed to reliably define management prescriptions. 3. Define essential criteria for experimental moist-soil management units, including location, soil characteristics, water source and quality, and minimal size of each unit. Consult with CDOW personnel and private individuals to locate potential sites. Work with engineering and property management personnel to develop guidelines and specifications for construction of experimental moist-soil units. METHODS Information Review Technical journals and popular literature were reviewed for information on moist-soil management techniques. Correspondence was sent to wildlife researchers active in the area of moist-soil management. During 13-16 February 1990, J. Ringelman traveled to a waterfowl management workshop hosted by the State of Kansas and consulted with experts on moist-soil management. Personnel at the Office of Information Transfer and the National Ecology �30 Research Center (both offices of the U.S. Fish and Wildlife Service) were consulted for information on moist-soil management techniques. Site Selection Criteria for suitable moist-soil experimental wetlands were developed based on data gathered during the information review. Correspondence outlining study objectives was forwarded to District Wildlife Managers and Area Biologists employed by the Colorado Division of Wildlife. In addition, private landowners with suitable sites were contacted and field visits were made to several sites. RESULTS Information Review One of the most efficient wetland management techniques available is the collection of practices known as moist-soil management: the management of plant species that grow on exposed mud flats (Fredrickson and Taylor 198~). Seeds of these species regularly survive flooding and can exist for years in areas subjected to prolonged drying (Van der Valk and Davis 1978). Under favorable conditions, usually at or slightly below field capacity, rapid germination allows re-establishment of these plants (Van der Valk 1981). Production of vegetation in naturally flooded or impounded wetlands is related to the timing of water removal in the spring. Annual changes in weather and seed availability, and succession in the wetland plant community, affect the resultant diversity and growth of plants, as does the time it takes to drain and re-flood an area (Fredrickson and Taylor 1982). Vegetation density and composition is dramatically affected by the degree of soil drying that follows water removal (Kirby 1988). Waterfowl food production is enhanced and maintained through proper timing of water removal and careful control of water levels throughout the year. The foods of greatest value are the seeds of native aquatic and semiaquatic plants and the invertebrates that associate with these plants (Fredrickson and Reid 1986), but vegetative growth and tubers are also consumed (Kirby 1988). Invertebrates provide protein used by female ducks during egg laying (Swanson et al. 1979, Drobney 1982), food for ducklings during their growth stage (Krapu and Swanson 1977), and nutritional supplements for waterfowl during fall and winter (Heitmeyer 1988). Moist-soil techniques allow management schemes targeted at important species by manipulating food availability to coincide with migration and breeding phenology, and providing preferred vegetative foods. Shorebirds are especially responsive to moist-soil management practices (Rundle and Fredrickson 1981). Wetlands are dynamic systems, and therefore cannot be continuously maintained at a desired high level of productivity. Periodic low levels of productivity must be accepted as part of the cost of maintaining a wetland of overall high quality. The effects on the waterfowl community are minimized by providing a complex of wetlands, managed to provide continuously available, high quality habitat. Initial development of moist-soil impoundments is expensive if heavy equipment is required for dike construction and elaborate water control structures are needed. A reliable source of water, precise control of water �31 levels, and the ability to completely drawdown an impoundment are all vital characteristics of moist-soil units (Fredrickson and Taylor 1982). Once constructed, moist-soil impoundments are often less expensive to manage than row crops, and have several advantages over row crops that are traditionally cultivated for waterfowl. Nutritional qualities of natural seeds are more balanced than those of cereal grains, and are often preferred by waterfowl. The total energy in moist-soil foods often is as high or higher than that in corn, milo, or soybeans (Fredrickson and Taylor 1982). Row crops, by virtue of their large seed size and physical structure, are suitable only for a select group of the larger waterfowl: geese, mallard (Anas platyrhynchos), and Northern pintail (A. acuta). These crops also do not provide shelter for waterfowl, another desirable trait of moist-soil plants. Unlike row crops, which may fail or have reduced yields as a result of unfavorable weather, weather conditions have a lesser effect on a naturally occurring flora which includes species that reproduce well under a variety of conditions. In addition to this plant diversity, aquatic invertebrates, reptiles, and amphibians flourish in moist-soil impoundments. This variety of organisms, while desirable in itself, attracts a wide range of larger wildlife species to an impoundment (Burgess 1969). Over 80% more species are found in moist-soil units than on adjacent row crops (Fredrickson and Taylor 1982). The distribution of plant and animal species differs with latitude. Consequently, wetland management techniques that work well in some areas may not work as well or at all in other regions. Colorado has environmental conditions that differ from those regions of the country where moist-soil management is traditionally practiced. The growing season is relatively short, precipitation is low, and evapotranspiration is high. Moreover, water is scarce, expensive, and seasonal abundance varies considerably. Traditional water management practices often run counter to regimes beneficial to moistsoil plants. For example, most irrigation reservoirs are filled during September-November, remain full during winter, then are drained rapidly in June-August to meet irrigation demands (Ringelman et al. 1989). Moist-soil impoundments generally benefit from low pool levels during winter, a gradual filling during March-April, high water levels through June, then an abrupt lowering of water levels in early July to allow seed germination (Nelson et al. 1978). Despite these apparent handicaps, moist-soil management has been successfully adapted for use in arid environments (Kirby 1988). At Bosque del Apache National Wildlife Refuge in New Mexico, techniques have been developed to encourage desirable waterfowl foods including alkali bulrush (Scirpus acutus), horned pondweed (Zannichellia palustris), millets (Echinocholoa spp.), sago pondweed (Potamogeton pectinatus), spike rush (Eliocharis spp.), and smartweeds (Polygonun spp.), while controlling undesirable plants such as cottonwoods (Populus spp.), willows (Salix spp.), saltcedar (Tamarix chinensis), Russian olive (Elaeagnus angustifolia), phragmites (Phragmites communis), cocklebur (Xanthium strumarium), and cattail (Typha spp.). Several of these species are present on Colorado's eastern plains, and it is reasonable to expect that many of the techniques applicable in New Mexico would succeed in Colorado. Ongoing research is focusing on substrate cultivation techniques as a technique to enhance germination rates of moist-soil plants and provide a more desirable moist-soil plant community (L. H. Fredrickson, Gaylord Memorial Laboratory, pers. comm.). These investigations of "advanced" management �32 methods are being conducted in Missouri, where moist-soil techniques have been practiced for a decade and plant responses in relation to drawdown date and moisture regimes are well documented. In Texas, investigators are initiating research on managing playa lakes for moist-soil species (L. M. Smith, Texas Tech University, pers. comm.). Their research is intended to develop management prescriptions similar to those identified in this study. Wildlife managers in Kansas and elsewhere in the Mississippi alluvial valley have developed moist-soil waterfowl impoundments using technology adapted from the rice cultivation industry. Levee plows are used to confine water in small, shallow impoundments. Such levees can be built and removed quickly in response to water availability and waterfowl management objectives. Lightweight, portable water control structures are used to precisely regulate water levels inside impoundments. The timing and duration of drawdowns, as well as control of undesirable plant species, are complex considerations in moist-soil management. The complexity is increased if water delivery systems do not allow independent flooding and dewatering of impoundments. To aid the manager in decision making, the U.S. Fish and Wildlife Service is developing an "expert computer system" that prescribes management actions based on responses to a series of questions (D. Hamilton, National Ecology Research Center, pers. comm.). The system currently being developed has applicability to Mississippi Valley habitats, but basic relationships may be adapted to conditions in Colorado. The expert system is schedules for completion by January, 1991. Criteria for Site Selection Sites for experimental moist-soil management units in Colorado are most limited by four factors: (1) soil permeability, (2) the timing of water availability, (3) the amount of water available, and (4) the cost of the water. Many state wildlife areas and private duck clubs are located in the South Platte River Valley. Soil types are primarily sand or sandy-loam, underlain with occasional and often non-contiguous clay layers. Consequently, seepage rates are high, and continual water application might be necessary to maintain moist-soil conditions during plant germination and growth. Rivers and water storage reservoirs are the sources of irrigation ditchwater in eastern Colorado. Often, such water is unavailable until several agricultural users place a "call" for irrigation water in mid-May. Unfortunately, this is the period when water should be removed from an impoundment to encourage many desirable moist-soil plants. Thus, either an early source of water must be used to fill impoundments during April, or impoundments must be able to hold water through fall and winter into spring. During most of the year, water is not free anywhere in Colorado. Water users must own shares of water (which entitle them to a portion of the total water available from an irrigation company) or must be willing to pay pumping costs of well water. In either case, the quantity of water must be sufficient to fill impoundments to the required depths while compensating for seepage and evaporative loss. The cost of ditch water ($4 -$20 / 1,000 mJ) is usually lower than water pumped from wells ($20 - $31 / 1,000 m3), but under some circumstances both may be cost-prohibitive for moist-soil impoundments. Moreover, Colorado water law states that water must be put to beneficial use, and therefore water rights are adjudicated for specific purposes. Most ditch �33 and well water is designated for agricultural applied to moist-soil impoundments. Experimental Criteria were developed for screening The area must be able to accommodate in size. 2. Each moist-soil unit must have flooding and de-watering. Elevation legally be Area Specifications 1. 3. use, and cannot change along potential 6 impoundments, independent the longest moist-soil water dimension study sites: each at least control 2 acres structures of the impoundment for should be < I ft. 4. Water sources develop. and delivery 5. Water 6. Water rights research. 7. The area must be under government control or, if privately owned, the landowner(s) must be willing to enter into a 5-year agreement to allow moist-soil research and agree to finance the cost of capital construction as needed. 8. Soil surveys must verify that seepage rates are sufficiently is feasible to develop wetlands on the site. must be available must be in place in sufficient must not preclude Based on these criteria, suitability. Candidate systems several quantity the use of water candidate or cost-effective at the appropriate for moist-soil areas were evaluated to times. management low that it for Sites Brush Prairie Ponds State Wildlife Area. - This newly-acquired state wildlife area is slated for major wetland developments as the first Ducks Unlimited MARSH (tlatching aid to Restore ~tate's Habitats) project in Colorado. The area consists of undulating sand hills with "pothole" type wetlands in the lower basins. Smartweeds and other moist-soil plants are present in some wetlands. Water is owned by the City of Brush, 2 miles to the north, and is applied via a series of ditches that transport water from the nearby South Platte River. The municipal water supply for Brush is drawn from deep wells located near the wildlife area. Surface water is applied to recharge the aquifer, therefore newly-designed wetlands are designed to have high seepage rates. This site was judged unsuitable for experimental moist-soil units. The volume and amount of ditchwater available for flooding impoundments is dependent upon river flows and calls on the ditch, which make the water source unpredictable. Additionally, seepage rates will average 1 foot/day for new �34 impoundments. Thus, shallow moist-soil overnight if deprived of water. impoundments would literally dry up Chestnut Slough (Front Range Land and Cattle Company). - This area provides a complex of warmwater and coldwater wetlands used primarily for waterfowl hunting. The area is < 600 acres in size, and contains little relief in the area potentially available for moist-soil impoundments. Water for impoundments would need to be supplied from wells. Because of the small area available for impoundment construction, high cost of well water, and legal constraints on capital construction (levees, control structures, etc.) on private land, this area was judged unsuitable for experimental moist-soil units. Big Meadow Farm. Inc. - Another private property, this hunting club had several desirable qualities that made it a good prospect for moist-soil units. Half of its 470 acres was set aside as a waterfowl refuge. Water was available from several sources, and a ditch delivery system was in place. A clay layer underlaid most of the flat bottomland, which bordered on the South Platte River. Several thousand waterfowl use the property during fall and winter. Club members were eager to cooperate on the project and were willing to make arrangements to help finance capital construction. Preliminary engineering surveys in fall 1989 indicated that elevation gradients were unfavorable for constructing moist-soil units. Ditches were in need of repair and cleaning, and water, although plentiful most of the year, was scarce during March and April when it was most needed for moist-soil work. This area was dropped from consideration as an experimental site because of these and other considerations. Wellington State Wildlife Area. - This wildlife area is located within 30 minutes of the Fort Collins Research Center. A full-time property technician oversees development and management of waterfowl impoundments and associated upland nesting habitat. The area has abundant water supplied by the Cowan Ditch, which flows from about May 10 through 20 September during average years. Several acres of flat ground located northeast of the property headquarters have been designated for future waterfowl impoundments. The property technician and associated management personnel are eager to cooperate on an experimental moist-soil venture. Heavy equipment and personnel are available to construct levees and install water control structures. A turnout and lateral ditch already exist in the area designated for potential development. The only known limitation to this site is the timing of ditchwater availability in spring. This appears to be a common limitation throughout eastern Colorado, and it is now believed that any practical moist-soil management system must be operable within the constraint of ditchwater availability. Consequently, this site has been selected for construction of experimental moist-soil management units beginning in late fall and winter 1990, barring any unforeseen complications. Target date for completion of the units and initial moist-soil work is spring 1991. �35 DISCUSSION Highly permeable soils and the timing of ditchwater availability are major constraints to moist-soil management in Colorado. Any moist-soil management system must overcome these obstacles before it can be widely applied in the eastern regions of the state. Experimental moist-soil units will be constructed at the Wellington Wildlife Management Area to evaluate the practicality of moist-soil management under these constraints and to derive management prescriptions to achieve desired plant responses. LITERATURE CITED Burgess, H. H. 1969. Habitat management on a mid-continent refuge. J. Wildl. Manage. 33:843-847. Drobney, R. D. adaptations waterfowl 1982. Body weight and composition changes and for breeding in wood ducks. Condor 84:300-305. Fredrickson, L. H., and F. A. Reid. 1986. Wetland and riparian habitats:' A nongame management overview. Pages 59-96 in J. B. Hale, L. B. Best, and R. L. Clawson eds., Management of nongame wildlife in the midwest: a developing art. The North Central Section of the Wildlife Society. 1982. Management of seasonally Fredrickson, L. H., and T. S. Taylor. U.S. Dep. Inter., Fish and flooded impoundments for wildlife. Wi1dl. Servo Resour. Pub. 148. 27pp. Heitmeyer, M. E. 1988. Protein costs of the prebasic molt of female mallards. Condor 90:263-266. Kirby, R. E. 1988. Highlights of the moist-soil management workshop. Fish and Wildlife Service, Office of Information Transfer, Fort Collins, Colorado. 57pp. U.S. Krapu, G. L., and G. A. Swanson. 1977. Foods of juvenile, brood hen, and post-breeding pintai1s in North Dakota. Condor 79:504-507. Ringelman, J. K., W. R. Eddleman, and H. W. Miller. 1989. High plains reservoirs and sloughs. Pages in L. M. Smith, R. L. Pederson, and R. M. Kaminski, eds. Habitat management for migrating and wintering waterfowl in North America. Texas Tech Press, Lubbock. Rundle, W. D., and L. H. Fredrickson. 1981. Managing seasonally flooded impoundments for migrant rails and shorebirds. Wildl. Soc. Bull. 9:80-87. Swanson, G. A., G. L. Krapu, and J. R. Serie. 1979. Foods of laying female dabbling ducks on the breeding grounds. Pages 47-57 in T. A. Bookhout, ed. Waterfowl and wetlands - an integrated review. The Wildlife Society, Washington, D. C. �36 Van der Va1k, A. G. 1981. Succession Ecology 62:688-696. in wetlands: a G1easonian approach. Van der Va1k, A. G., and C. B. Davis. 1981. The role of seed banks in the vegetation dynamics of prairie glacial marshes. Ecology 59:322-335. Prepared by: q...,. ~~ James K. Ringelman Wildlife Researcher C �37 Colorado Division of Wildlife Wildlife Research Report September 1990 JOB PROGRESS State of Colorado Project 01-03-212 Work Plan 10 Job Title: Cooperative Period Covered: Author: : Job REPORT Migratory Game Bird Research _1_ Management Programs 01 April 1989 through 31 March 1990 Michael R. Szymczak Personnel: James K. Ringelman Wildlife and Michael R. Szymczak, Colorado Division of ABSTRACT The statewide waterfowl management plan was completed and submitted for administrative approval. Recommendations for wetland habitat improvements on present holdings or proposed acquisitions were provided for Russell Lakes, Spinney Mountain Ranch, Brush Prairie Ponds, South Republican and Union Slough and Stagecoach Reservoir SWA's; for Hebron Ponds and Walden Reservoir (BLM); and for the Routt and Arapaho National Forests and Pawnee National Grasslands (USFS). An extensive st.udy of duck production was conducted in montane habitat study areas on the Rout-t; National Forest. At least a dozen educational presentations concerning duck recruitment in montane habitats or general w'aterfowl ecology were prepared· and given. Responsibiliti~s as Colorado's representative on Pacific Flyway Study Committees and Council were fulfilled. Methodology for conducting duck, Canada goose and sandhill crane production surveys were prescribed for some specific areas. ��39 Cooperative Migratory Bird Management Programs Michael R. Szymczak James K. Ringelman In 1988, the Colorado Division of Wildlife (CDOW) created the Migratory Game Bird Program Unit (MBPU) within the Terrestrial Wildlife Section. This administrative change combined all individuals having statewide responsibilities for research and management of migratory game birds. Members of the MBPU work in concert to improve migratory bird management in Colorado. This job was created to allow team members to participate in these management programs. P. N. OBJECTIVES 1. Participate in developing and implementing habitat-based waterfowl management plans on a statewide, habitat region, and project basis. 2. Advise state and federal land managers on beneficial habitat acqu~s~tions and/or developments and provide expertise in preparation of development and/or management plans. Advise private land managers in developing habitat management plans and assessing impacts on waterbird populations. 3. Present information on the principles of waterfowl management to workshop attendees, educational classes, and conservation organizations. 4. Participate levels. 5. Cooperate in developing surveys and techniques of migratory bird management programs. in migratory bird management meetings at the state and flyway that will assess the impact SEGMENT OBJECTIVES 1. Complete final draft on waterfowl habitat of statewide waterfowl region plans. 2. Provide biological expertise for wetland development programs on the Brush Prairie Ponds State Wildlife Area (SWA), Russell Lakes SWA, South Republican SWA, Hebron Ponds (BLM jurisdiction), Walden Reservoir (BLM jurisdiction), Routt National Forest (U.S. Forest Service jurisdiction, and other areas when requested. 3. Prepare and present when requested. 4. Compile appropriate population status information and represent Colorado at Pacific Flyway Technical Committee and Council meetings. Attend migratory bird management program meetings in Colorado when requested. informational programs management on migratory plan. Begin work bird'management �40 5. Provide surveys methodology for wetland when requested. habitat and migratory bird population RESULTS Waterfowl Management Plans Early drafts of the statewide waterfowl management plan were reviewed and corrections and changes made based on comments from CDOW personnel and outside reviewers. Tables and figures were improved and prepared for publication. The final draft of the statewide waterfowl management plan was submitted to the CDOW Director and the Colorado Wildlife Commission for approval. The plan was published as: Colo. Div. of Wildl. 1989. Colorado Statewide Waterfowl Management Plan 1989 -2003. Colo. Div. of Wildl., Terrestrial Wildl. Sect., Migr. Game Bird Prog. Unit. Ft. Collins, co. 98pp. The Russell Lakes SWA management plan, prepared by CDOW Southwest Region personnel, was reviewed and changes were recommended. This plan will be rewritten in a more detailed form following the accumulation of water rights data and the construction of contour maps. Proposed developments at the South Republican, Brush Prairie Ponds and Spinney Mountain, and Stagecoach Reservoir SWA's were recommended. Suggestions for wetland alterations have been included in the management plan for the South Republican SWA. Water flow as through the. Brush Prairie Ponds SWA was altered as a result of our recommendations. The South Park waterfowl management plan, including habitat alterations suggested for the Spinney Mountain SWA, is being developed by CDOW Southeast Region personnel. Proposed wetland developments on public lands at Hebron Ponds and Walden Reservoir (BLM) , the Pawnee National Grasslands (USFS) and on the Arapaho National Forest (USFS) in Middle Park were visited and comments made to the appropriate agencies. In addition, a proposal for aquatic vegetation control on Shadow Mountain Reservoir was reviewed in reference to possible effects on waterfowl populations. Construction of l8-Island Reservoir in the Hebron Pond complex was completed by the BLM and negotiations for a stable water supply continue. Wetland development sites north of Walden Reservoir were proposed in conjunction with the BLM. A delivery system that will enable additional water to be placed in Walden Reservoir was cooperatively financed and constructed. An agreement to use a portion of the CDOW water in Walden for additional wetland developments is being negotiated. Roosevelt National Forest personnel developed plans and a priority ranking for wetland enhancement on the Pawnee National Grassland. Proposals were submitted for funding and some were approved. An extensive, cooperative study of nesting waterfowl on a forested study site just south of Big Creek Lakes (Routt National Forest) was undertaken to identify factors limiting production on this site and to set protocol for future waterfowl research and management on Forest Service lands. Recommendations were developed for habitat manipulations to address those factors. Results of this cooperative study are presented as an appendix to this report. �41 Proposed Wetland Acquisitions A proposed addition to the Union Slough SWA was visited, and potential developments discussed. The addition to the Union Slough SWA has been included in acquisition plans of the CDOW Northeast Region. We served on a committee to develop a rating criteria for wetland acquisitions and development for the Northeast Region of the CDOW. These criteria have been approved for use by the Regional Manager. Population Survey Methodology Spring duck count surveys were planned for Russell Lakes and Spinney Mountain SWA's. Project and Southeast regional personnel will cooperate conducting surveys on Spinney Mountain SWA. Informational in Programs At the invitation of the State of Kansas, lectures were presented on current waterfowl research and management in Colorado. Formal presentations on waterfowl recruitment in montane habitats were presented to the national meeting of the Society of Wetland Scientists and the Colorado Chapter of The Wildlife Society's annual meeting. Other formal presentations were made to U.S. Fish and Wildlife Service personnel at the National Ecology Research Center, the Wildlife Management Shortcourse and Wildlife Biology Workshop at Colorado State University, and to U.S. Forest Service personnel at several locations throughout the region. Waterfowl Technical Committee and Council Meetings Colorado was represented at the July 1989 Pacific Flyway Study Committee and Council meetings and March 1990 Study Committee meeting. Waterfowl population status was reviewed in July and hunting season recommendations forwarded to the U. S. Fish and Wildlife Service's Regulation Committee. Populations of specific interest to Colorado whose status was reviewed in July were (1) breeding and wintering mallards inhabiting the Pacific Flyway portion of Colorado and (2) the Rocky Mountain Canada goose population. General information on Pacific Flyway migratory game bird populations was exchanged by committee members in March. Regulatory recommendations made to the Pacific Flyway Council included hunting season regulations for Rocky Mountain Greater Sandhill Cranes and Four-corners Band-tailed Pigeon populations. Population Survey Methodology Methods for conducting duck brood counts were outlined for the Russell Lakes SWA. A reconnaissance of the Spinney Mountain SWA was conducted in fall of 1989 and methods of measuring waterfowl production and defining limiting factors relayed to managers. After consulting historical survey information, a time schedule for duck breeding pair and brood surveys was recommended. Methods were outlined for conducting Canada goose production surveys on Walden and MacFarlane reservoirs and Lake John Annex in North Park during spring 1990 in order to document suspected poor productivity on Walden �42 Reservoir. All goose nests on the 3 areas will be visited, marked and mapped in early May just prior to hatch. The number of eggs in each nest will be recorded. Periodic brood counts will be conducted after hatch with brood size and age recorded. The nests will be revisited after all have hatched, and the fate of each nest will be recorded. "Nearest nesting neighbor" distances will also be recorded for all nests located on islands. Comments were made on the draft recovery plan for nesting Greater Sandhill Cranes in Colorado that was prepared by the CDOW Northwest Region. In addition, information on minimum viable populations, south-central Wyoming nesting populations and general behavior during chick rearing of greater sandhill cranes was obtained from a number of sources and conveyed to Northwest Region personnel. DISCUSSION Project personnel proved to be a useful resource in planning and evaluating waterfowl management and habitat enhancement programs in Colorado and educating land management agency personnel about the habitat requirements of waterfowl. We anticipate that with increased emphasis on habitat enhancement in Colorado as outlined in the statewide Waterfowl Management Plan that our services will be more in demand. Continued participation on Flyway committees ensures that Colorado will remain informed on migratory bird matters and have input in migratory bird hunting regulations. Prepared by: ?Y/,"J{!..IJ{? ~....e Michael R. Szymczak Wildlife Researcher C �43 APPENDIX A Report on the cooperative habitats investigation in the Routt Field of waterfowl National Forest, recruitment Colorado Personnel R. Langley, Colorado State University M. Willms, Colorado Division of Wildlife J. Ringelman, Colorado Division of Wildlife Cooperators J. Ficke, U.S. Forest Service, Steamboat Springs C. Neelan, U. S. Forest Service, Steamboat Springs S. Kozlowski, U.S. Forest Service, Walden C. Yancey, U.S. Forest Service, Walden G. Hetzel, U.S. Forest Service, Denver L. Mullen, U.S. Forest Service, Denver S. Porter, Colorado Division of Wildlife, Walden K. Snyder, Colorado Division of Wildlife, Walden M. Szymczak, Colorado Division of Wildlife, Fort Collins R. Hopper, Colorado Division of Wildlife, Fort Collins S. Steinert, Colorado Division of Wildlife, Fort Collins B. Neeley, The Nature Conservancy, Boulder in montane �44 WATERFOWL ABUNDANCE, PRODUCTION, AND HABITAT ROUTT NATIONAL FOREST, COLORADO USE ON THE James K. Ringelman, Mark A. Willms, and Richard S. Langley Colorado Division of Wildlife, Ft. Collins, CO 80526 Abstract: We quantified abundance, habitat use, and production of waterfowl in a 9.6 km2 study area south of Big Creek Lakes on the Routt National Forest. In 1989, we enumerated 59 breeding pairs of ducks (6.1 pairs/km2) representing 6 species. Mallard (Anas platyrhynchos), ring-necked duck (Aythya collaris) and green-winged teal (A. creeca carolinenis) were the most common species. Pair and brood use were strongly correlated with pond size, with larger wetlands receiving greater than expected use. Pond origin was a secondary factor affecting wetland use by waterfowl. Based on the high pair use, species diversity, general lack of re-pairing, and high proportion of pairs observed with broods, such factors as pond numbers, upland nesting cover availability, nest success, and early season invertebrate numbers appear not to be limiting duck production. However, suitable nest cavities may be in short supply for bufflehead (Bucephala albeola). Mean brood size at fledging was lower than expected (4.1 young/brood), suggesting that duckling survival is the factor most limiting duck recruitment. The most likely causes of poor duckling survival were the general absence of flooded emergent cover after 1 July, and heavy growth of cowlily (Nuphar polysepalum) in late summer, which impeded brood movements and caused separation of ducklings. Management actions recommended to improve duckling survival include hard-stem bulrush (Scirpus acutus) introduction and herbicide spraying to create travel lanes in ponds with dense cowlily. Bufflehead nest boxes are also recommended to improve nest use and success for this species. Recent declines in North American duck populations have prompted coordinated action by state and federal natural resource agencies. Under the umbrella of the North American Waterfowl Management Plan (NAIJMP), the U.S. Forest Service (USFS) has initiated a 5-year, $18 million program to survey and improve waterfowl habitat on USFS lands. This program, called "Taking Wing", addresses the needs of breeding waterfowl through habitat preservation and development. Significant waterfowl production occurs in mountainous areas of Colorado (Colorado Division of Wildlife 1989). Reliable water sources create a wetland community that is less subject to annual perturbations than grassland habitats. Stable habitat is particularly important during times of widespread drought on the prairie breeding grounds, such has occurred in recent years. Despite the importance of mountain habitat for waterfowl, information on breeding duck densities and habitat use is scarce. Previous studies are largely restricted to those by Rutherford and Hayes (1976) for the upper Rio Grande drainage and Frary (1954) for the White River Plateau in Colorado. Over half of the ducks produced in these studies were mallards, a species of primary concern in the NAIJMP because of recent population declines. Habitat development techniques proposed for the "Taking Wing" program include (1) wetland vegetation management and construction of new wetlands, (2) upland nesting cover enhancement and creation of secure nesting habitat in the form of islands, nest structures, or predator-proof enclosures, and (3) �45 Ko,,'tt tl<\tiWQI Fcce.s+ j I I K'P\ I I I .I I Co lo.-o..elo Ar,'J. MeSas �W I-I-- « _J . a. z u.. o ~ * cr: o u, * z * 46 ~ 8: -z _... o- ~ . 0<) ..•.. \)"'0 J .Jtn 8~ 3 ~ V :t~«l ./ ./ / / d~ 0 ~ * .. .v. * �~ w W 0:: o * co o M N t'( ::)- aM '...., .,.. c:J ..., cf - 0 :l- ~~ D 0.:' 47 �48 ~ w w a: o WM en o o C!' ,~ .. " * * / I I , I ,I I J \ I ~ I ;' * ._..-_- * * �~ w w a: o W 0: LL « :c en M * * * ~ -:£. \ !J~ \ ~ t:.. "'0 11'1 <, 6.:s- <V ,... ~ 0 * 49 / I I\ \ _ ______.. /-\-/ \ J I I .•. c:1\ -t'\ I ,) \ I \ / I I \ I i I I �50 prov~s~on of additional permanent water areas and establishment of animal and vegetative foods. These techniques, which have been developed and tested primarily in prairie habitat, may attract and hold breeding pairs, increase nest success, and enhance brood survival, respectively. Unfortunately, which of these factors may limit waterfowl production in Rocky Mountain breeding habitat is unknown. Applying any management technique without first identifying the limiting factor is a waste of valuable resources. In order to determine which of the habitat management options might be most cost/effective in terms of increasing waterfowl production on National Forest lands, we decided to take a management experiment approach (Walters 1986). This paper describes the first phase of this project, that of setting up the experimental units, obtaining pre-treatment data, determining the factor(s) most limiting waterfowl production, and recommending specific management actions aimed at addressing these shortcomings. The second phase of this experiment would be to implement the management proposals on the designated treatment units, repeat the same data collection procedures for 12 years, and by comparing the results on the treatments and control units with the pre-treatment data, assess the actual effectiveness of the management actions. Our specific objectives were to (1) determine the species composition, abundance, and habitat use by breeding waterfowl pairs, (2) use the ratio of lone drakes to pairs as an indication of nesting chronology and nest loss/renesting activity, (3) measure duckling survival rates to fledging, and habitat use by brood-rearing hens, and (4) recommend management actions directed at factors limiting waterfowl production. STUDY AREA We investigated 186 wetland basins of either beaver or glacial origin within a 9.6 km2 section of the Routt National Forest just south of Big Creek Lakes, approximately 33 km northwest of Walden, Colorado (Fig 1). Elevation of the study area ranged from 2573-2774 m (8520-9100 ft). Dominant forest vegetation in the area consisted of lodgepole pine (Pinus contorta), Engelmann spruce (Picea engelmannii), subalpine fir (Abies lasiocarpa) and quaking aspen (Populus tremuloides). Riparian zones were dominated by willows (Salix spp.), with smaller amounts of water birch (Betula occidentalis) and mountain alder (Alnus incana). Common aquatic vegetation included cowlily, grassleaf pondweed (Potamogeton gramineus), fineleaf pondweed (E. filiformis), spiked watermilfoil (Myriophyllum spicatum), northern mannagrass (Glvceria borealis), and sedges (Carex vescaria and Carex utriculata). Less common aquatic plants included Richardson pondweed (E. richardsonii), common mares tail (Hippuris vulgaris), hemlock waterparsnip (Sium ~), yellow water buttercup (Ranunculus flabellaris), common arrowhead (Sagittaria latifolia), common spikerush (Eleocharis palustris), narrowleaf burreed (Sparganium angustifolium), water sedge (Carex aguatilis) and bluejoint reedgrass (Calamagrotis canadensis). All beaver ponds were located on or adjacent to either the North Fork of the North Platte (NFN Platte) or 3 of it's primary tributaries: Forester, Goose and Shafer Creeks. Boundaries partitioning the study area into 4 study units were established along their watershed divides (Figs. 2-5). Without prior knowledge of the study area, the NFN Platte and Goose Creek sites were designated as the treatment units and Forester and Shafer Creek as control �51 Table 1. Comparative attributes of the 4 study units National Forest, Colorado, 1989. Unit Size (k~) No. wetlands Beaver located in the Routt available Moraine Wetland density N. Fork N. Platte 2.00 66 9 37.5 Forester 2.56 22 21 16.8 2.38 39 5 18.5 2.67 1 21 8.2 9.61 128 56 19.1 Goose Creek Creek Shafer Creek Totals units. Although attempts were made to keep these units comparable, large differences in both the number and type of wetlands in each study block occurred (Table 1). Because of stream gradient differences, both the number and complexity of beaver ponds present along each of these 4 creeks varied. Thus the NFN Platte and Goose Creek, both low gradient streams, had the highest number and diversity of beaver ponds, while Shafer Creek, a high gradient stream, had the lowest. Forester Creek, with a high gradient headwater and a low gradient lower reach, was intermediate. Since the principal water source for these creeks was high mountain snowfields, water levels on all main-stern beaver ponds remained relatively stable throughout the breeding season. In contrast, beaver ponds located on secondary drainages lacking mountain water sources tended to be more unstable. Glacial moraine ponds were randomly scattered throughout the forested inter-creek areas. Hydrologically, these moraine ponds were isolated by thick ridges of glacial till. Since their water supplies are entirely dependent on local snow melt and summer thunderstorms, water levels gradually decline as the breeding season progresses. The long term stability of these ponds most closely resemble those in the prairie parklands. Most of the larger moraine ponds either contained active beaver or signs of recent activity. METHODS We conducted breeding pair surveys on all wetlands present in each of the 4 study units at 7 day intervals from 22 May to 7 July, 1989. Data recorded included species, number of pairs or lone individuals present, general activity (i.e. sitting, flying), wetland location, and time of day. We used quiet observation techniques to avoid flushing ducks and causing count duplications. For analysis of habitat selection, only pair data recorded between 26 May and 15 June were considered, since both pairs and lone drakes began dispersing from the study area after mid-June. Nests were located opportunisticly in conjunction with other activities. Nest information �52 recorded included species, number of eggs present, egg fertility, incubation stage, nest fate, and general habitat characteristics. Early morning and late evening brood surveys were initiated on 10 July and continued through 16 August, again using quiet observation. Unlike pair surveys however, we sampled only a selected subset of the wetlands present in each study unit. This was done primarily to increase the probability of obtaining repeated counts on individual duck broods. For each brood observed we recorded the species, brood size, duckling age (Gallop and Marshall 1954), presence or absence of the hen, wetland location, and time of day. An attempt was made to observe each brood at least once a week. Wetland habitats were categorized using a classification system (Cowardin et al. 1979) modified specifically for montane wetlands. Information included in this system included wetland origin (moraine or beaver created), dominant vegetative class (submergent, floating leaf, emergent, moist soil, shrub, or organic bottom), water permanency (percent of water surface lost from basin during the summer) and beaver activity. Other supplementary information recorded for each wetland included the relative impacts from human disturbance and cattle grazing as well as the percentage of the pond basin dominated by submergent, floating leaf, emergent, woody, and moist soil vegetation. A cover map of each wetland was drawn to scale for later use in calculating pond size with a hand planimeter. From these data, we calculated total surface water available in late May within 200 m of each wetland. The mean distance from each wetland to its 6 nearest neighbors and the total number of ponds within 200 m were calculated from study unit maps and aerial photographs. Habitat selection analyses were performed using all these variables in a stepwise regression procedure for predicting wetland pair and brood use days. RESULTS Species Composition and Breeding Chronology Nine waterfowl species were observed in the study area (Table 2). Mallard was the most abundant nesting species (40.6%), followed by ring-necked duck (23.7%), green-winged teal (13.5%), cinnamon teal (Anas cyanoptera; 8.5%), bufflehead (8.5%), and American wigeon (Anas americana; 5.1%). Other nonbreeding species included Canada geese (Branta canadensis), wood duck (Aix sponsa) and common merganser (Hergus merganser). Nesting chronologies were estimated from known age broods using literature values for average clutch size, incubation periods (Erskine 1972, Klett et al. 1986), and fledging dates (Gallop and Marshall 1954). Peak nest initiation for waterfowl occurred between mid-May and early June, with mallards and buffleheads the earliest nesting species and ring-necked ducks the latest. Consequently, most ducklings hatched between mid-June and mid-July and fledged sometime in August. Based on estimated hatching dates, the length of the nesting season varied among species (Fig. 6). Mallards had the longest breeding period of any species, extending over 8 weeks, followed' closely by green-winged teal at 7 weeks. Although sample sizes for cinnamon teal, ringnecked duck, and bufflehead are small, their nesting seasons appeared shorter, with none of them exceeding 3 weeks in length. Since the extended breeding seasons for mallards and green-winged teal lacked a definitive peak, some renesting may have occurred. Except for a few mallards, most drakes left the study area by early July to undergo post-breeding molt. �53 en 7 o o 6 rn 5 ••••• o 4 10- 3 "'C - Mallard - Green-winged teal CJ American wigeon cs::Sl Cinnamon teal 1",,1 Ring-necked duck [ZJ Bufflehead Io- Q) ..c 2 ::J 1 E Z o , 1 I 2 I 3 I 4 I 5 , I 6 7 9 Weeks Figure 6. Number of duck broods hatching each week on the Routt National Forest, Colorado, 1989. Week 1 corresponds to week of 25 - 31 May. 10 �54 Table 2. waterfowl Species composition, abundance, and nesting chronologya breeding in the Routt National Forest, Colorado, 1989. for Mean date of Incubation Hatching commencement Species Number of pairs Nest initiation Mallard 24 20 May 28 May 23 June 18 August 8 29 May 7 June 30 June 4 August Green-winged teal Fledging Cinnamon teal 5 2 June 11 June 4 July 12 August American wigeon 3 1 June 9 June 3 July 21 August 14 7 June 16 June 11 July 31 August 5 15 May 27 May 25 June 16 .August Ring-necked Bufflehead duck a Based on backdating known age broods. Sample sizes are in parentheses: mallard (13), green-winged teal (6), cinnamon teal (2), American wigeon (1), ring-necked duck (4), and bufflehead (2). Because individual birds were not marked, m~n~mum home range sizes of pairs and broods were determined based on conservative assumptions using species identification, age class, brood size, and behavioral associations with other ducks. Our confidence in assigning individual identities was higher for broods than for pairs. In general, dabbling duck pairs moved more frequently than diving duck pairs, and lone males moved more frequently than pairs. However, few individuals traveled more than 500-600 m from their core area of use, and most simply appeared to move between the closest set of available, preferred wetlands. Brood movements were variable and highly dependent upon wetland isolation. Broods using moraine ponds rarely moved during the entire brood rearing period, whereas those using strings of beaver ponds used several ponds within a 300-400 m area. Pair Density A total of 59 pairs of breeding ducks were observed on the 4 study units. Overall pair densities for this area averaged 6.1 pairs/km2 (Table 3), but was not uniform among units. The NFN Platte and Goose Creek had the highest densities with 9.5 and 9.2 pairs/km2 respectively, whereas Shafer and Forester Creeks had the lowest with 3.7 and 3.1 pairs/km2. Species were distributed proportionally among all 4 units, except for cinnamon teal, which were only observed in the Goose Creek unit. �55 Table 3. Estimated number National Forest, Colorado, Species of pairs using Study Unit Forester N. Fork N. Platte Mallard Green-winged each study unit in the Routt 1989. teal Goose Shafer Total 11 3 8 2 24 2 1 2 3 8 Cinnamon teal 0 0 5 0 5 American wigeon 2 0 1 0 3 3 2 6 3 14 1 1 1 2 5 19 7 23 10 59 Ring-necked Duck Bufflehead All species Table 4. Weekly changes Forest, Colorado, 1989. in lone male to pair ratios on the Routt National Week May June June June June July Species 26-29 5-8 12-15 19-22 25-28 4-7 Mallard 0.9:1 1. 6: 1 5.5:1 3:1 5:1 3:0 7:1 5:0 2:1 3:0 1:1 0:0 Green-winged teal Cinnamon teal 3:1 4:1 4:1 2:0 0:0 0:0 American wigeon 0:1 0.3:1 2:0 0:0 0:0 0:0 0.2:1 0.3:1 2.2:1 1. 7: 1 2.3:1 0:1 0.3:1 2:1 2:0 0:0 0:0 0:0 0.7:1 1. 2: 1 3.3:1 2.4:1 3:1 3:1 Ring-necked Duck Bufflehead All species The ratio of lone males:pairs increased for all species as the breeding season progressed (Table 4). Once lone males began appearing in large numbers, only a few pairs of mallard, green-winged teal and ring-necked ducks �56 re-appeared in the population. Dates reflecting the lone male:pair ratio reversal corresponds closely with the estimated start of incubation for wigeon, ring-necked duck, and bufflehead based upon brood ages (Table 2). For mallards, the sharp reduction in pair numbers occurred approximately 2 weeks after the estimated peak in incubation (based upon brood ages), but it occurred at least 2 weeks prior to the estimated start of incubation for both teal species. Not all females disappeared during June, however. At least 34 lone female buffleheads were observed during all times of the day after midJune, a period when they should have been incubating. Nest Success Due to the abundance of available nesting cover, the strong site tenacity of incubating females, and the lack of time in which to conduct a thorough nest searching operation, only 2 mallard nests were discovered. Both of these nests were found during early egg laying while conducting pair counts along wetland margins. One nest was abandoned because of our disturbance, but all 8 eggs in the other nest eventually hatched. Although several other hens were flushed from vegetation adjacent to wetlands or observed returning from incubation breaks, none of their nests could be located. Without a larger sample size, a definitive estimate of nest success is not possible. However, inferences based on the low re-pairing rates and the high brood sighting:pair ratios suggest that overall hen success (percent of hens that nest successfully) in this part of the Routt National Forest is at least 50% and more likely around 70%. Brood Survival A minimum of 28 broods were observed on the 54 wetlands we surveyed regularly. Of these, at least 15 (53.6%) fledged or were class lIb or older when last sighted. Assuming that all of these broods fledged young with no further mortality, mean duckling survival was 4.1 young/brood fledged (Table 5). The largest fledging brood had 9 ducklings, the smallest only 1. Overall production was at least 6.3 ducklings/km2, or 1.1 fledged young/pair. However, since only a subset of wetlands surveyed for pairs were checked for broods, it is likely that some broods were unrecorded. Also, some broods (i.e. mallards and teal) were very difficult to relocate due to visibility factors, and may well have fledged young without our knowledge. Assuming 60 pairs of ducks in the study area, 70% hatching success, 65% brood survival, and 4.1 ducklings/brood fledged; waterfowl production in the area could have been as high as 10.7 ducklings/km2 or 1.7 fledged young/pair. Minimum/maximum production figures for each species ranged from 3.6-4.5 ducklings/km2 for mallards to 0.0-0.3 ducklings/km2 for wigeon. As with pair data, individual study units were not equivalent in their waterfowl production (Table 6). The NFN Platte unit fledged at least 10.0 ducklings/km2, compared to only 4.5-6.3 ducklings/km2 for the other 3 units. Despite having the largest number of fledging broods, Goose Creek also had the lowest mean brood size at fledging, thus countering its overall impact on production. �57 Table 5. Observed and best estimates National Forest, Colorado, 1989. for duckling Mallard Green-winged teal on the Routt No. of young fledged 2 km er 2er breeding 2air 2 observed estimated observed estimated -No. of fledgedbroods young Species production 7 35 3.6 4.5 1.5 1.7 2 6 0.6 1.4 0.8 1.6 Cinnamon teal 1 4 0.4 0.9 0.8 1.8 American wigeon 0 0 0.0 0.3 0.0 1.0 3 9 0.9 2.5 0.6 1.7 2 7 0.7 0.9 1.4 1.8 15 61 6.3 10.5 1.1 1.7 Ring-necked duck Bufflehead All species Table 6. Observed duckling production National Forest, Colorado, 1989. Unit No. of fledged broods young within each study unit on the Routt No. of voung fledged per breeding pair per km2 N. Fork N. Platte 4 20 10.0 1.1 Forester 3 16 6.3 2.3 5 13 5.5 0.6 3 12 4.5 1.2 Goose Shafer Creek Creek Creek Duckling survival rates using 15 broods with known fates suggested a dichotomy between date of hatch and pond type. Nests that hatched prior to 26 June had an average of 5.4 ducklings/fledged brood, compared to only 2.8 ducklings/fledged brood for those hatching later (E = 0.07, Van der Waerden Test). Although not statistically significant, broods using beaver ponds fledged an average of 4.5 ducklings/brood compared to just 2.8 for moraine ponds (E = 0.38). Duckling survival rates declined sharply for the first 2 to 3 weeks following hatch, then stabilized at about 4.0 ducklings/brood (Fig. 7). The deflated values calculated for class II broods are primarily the �58 "'C 0 0 7 .c <, 6 ~ C) C 5 0 4 :::J ~ •••• 3 ~ Q) 2 E 1 Z 0 0 .c :::J la lib Age Class Figure 7. Mean brood size by age class for all duck species combined, Routt National Forest, Colorado, 1989. �59 result of 3 large broods (R = 8.3 ducklings) which were not observed while members of these age classes. Inflated values for class III broods reflect the absence of 5 small, late hatching broods (R = 2.1 ducklings) which were too young to qualify for this class when brood observations were terminated. Habitat Use Examination of the correlation matrices for the 17 habitat variables indicates that they basically reflect 3 separate wetland factors: pond size, origin and dominant covertype (Table 7). The remaining 14 variables are Variables correlated with pond sizea, pond or~g~n, and dominant in the pond, Routt National Forest, Colorado, 1989. Table 7. covertype Correlated Basin variables size 0.98** Permanency Wetland Habitat Pond origin -O.ll Factors Dominant covertype -0.09 -0.00 -0.71** 0.10 0.07 0.53** -0.02 intensity -0.00 0.53** 0.03 disturbance 0.03 0.61** 0.01 Beaver activity Grazing Human Pond size % Water 0.17* % Submergent cover % Floating leave % Emergent cover cover -0.01 -0.04 -0.19* 0.42** -0.66** 0.41** -0.46** -0.22* 0.10 -0.51** -0.02 % Woody cover -0.28* 0.70** 0.06 % Moist soil cover -0.23* 0.16* 0.12 -0.21* 0.73** -0.10 0.15 -0.71** 0.03 Number Mean ponds/200-m distance/6 Total water nearest surface/220-m 0.10 0.04 -0.06 a pond size = portion of pond conta~n~ng water in May, pond or~g~n = beaver or moraine created, dominant covertypes in pond are ordered from deepest to shallowest. b.r. < 0.05" .r. < 0.0001 "* �60 related to these 3 as follows. Larger ponds tended to have more floating vegetation and less woody cover than did smaller ponds. Ponds dominated by submergent or floating leaf vegetation had a greater percentage of their basins filled with water. And finally, beaver created ponds tended to (1) be closer together, (2) more permanent, (3) more likely to contain active beavers, (4) have greater impacts from grazing and human disturbance, and (5) contain a higher percentage of submergent, woody, and moist soil vegetation, but less floating leaf vegetation than ponds of glacial origin. Use of wetlands by waterfowl pairs was strongly related to pond size (Fig. 8), with pond origin a minor secondary factor (Table 8). The best single predictor of pair use among related pond origin variables was the amount of human disturbance around the wetland. Pond size was the most influential variable affecting pair use for mallard, ring-necked duck and bufflehead. Pond size was less important for green-winged teal, cinnamon teal, and wigeon. Pond origin, whether predicted best by human disturbance, grazing intensity, or beaver activity, was a secondary factor for all species, with preference towards beaver created ponds. Table 8. Wetland characteristics best explaining on the Routt National Forest, Colorado, 1989. Dependent variable Species Independent observed variablesa + Disturb Surwtr (0.65)""'''b Mallard Surwtr (0.52) *"" + Disturb (0.02)* Surwtr (0.11) (0.03)* Surwtr (0.10)*** Su rw t r (0.43) ""'* + Disturb Su rw t r (0.50)., ....• Cinnamon teal Ring-necked Bufflehead teal Duck *"* + Disturb + Graz + Beav (N (partial R2) 160) (0.06)""''' Pair Use Days Green-winged pair use + Class (-0.01)* (0.09)*** (0.05)"" (0.01)" a Surwtr = pond water surface area in late May, Disturb = relative human disturbance, Class = dominant vegetation class of wetland basin, Graz = cattle grazing intensity around wetland, and Beav = beaver activity in wetland. b f < 0.0001 ***, f < 0.01 "*, f < 0.05 " Because not all wetlands were surveyed for duck broods, habitat selection within this reproductive phase may be biased. Nevertheless, like pairs, pond size was the principal factor explaining brood use for all species but teal (Table 9). Significant secondary factors included ponds dominated by floating leaf vegetation for all broods combined and pond origin (active beaver ponds) for mallards. �61 en :10.. -- Wetland size clas ••• 120 0- 0-200, 201-400, 401-800,801-1600,1601-3200,3201-6400, 6401-12,800,12801-25,120, CO <, (ha)' ascending and' 25,121 ord.r. 100 en "'C c 0 £:::,. 80 o 0•••••• 0 :10.. Q) ..c E ::J Number of ponds Pair use 60 o 40 £:::,. 0 o 20 Z c: o - 0 o o 3 4 o o c: 0 I 0 1 2 Wetland 5 6 7 8 9 Size Class Figure 8. Total pair use days and number of ponds available within selected wetland size classes, Routt National Forest, Colorado, 1989. in �62 Table 9. Wetland characteristics best explaining on the Routt National Forest, Colorado, 1989. Dependent variable Species Brood Use Days Mallard Green-winged Cinnamon teal Ring-necked Bufflehead teal Duck Independent Surwtr (0.58)**~ Surwtr (0.53)*** Surwtr (0.13)** Graz observed variablesa + % Water brood use (N 56) (partial R2) (0.03)* + Beav (0.09)** (0.08)* Surwtr (0.22)** Surwtr (0.32) *** a Surwtr = pond water surface area in late May, % Water = % of basin covered by water, Graz = cattle grazing intensity around wetland, and Beav beaver activity in wetland. b f < 0.0001 "", f < 0.01'" f < 0.05 * Since larger ponds provide more surface area/pond for duck pairs to occupy, a strong relationship between pond size and the number of waterfowl use-days should be expected. Thus the mere existence of such a relationship does not necessarily indicate actual preference by duck pairs for big ponds. In order to separate these 2 relationships, we divided the ponds into 4 size classes and calculated pair density/ha of surface water available (Table 10). Once again pair use increased with pond size, with large ponds having nearly 3 times the pair density of small ponds. Pair densities for American wigeon, ring-necked duck, and bufflehead all increased as pond size increased, whereas mallard and the teal species showed no consistent trend. As a result, dabbling ducks were the dominate group in the smaller wetland classes while diving ducks dominated in the largest size class. A similar analysis for brood was performed, but since it has to be corrected for observational effort, is thought to contain some bias. An analysis of nest site selection was not possible because of insufficient sample size. However, both nests which were discovered had similar cover and site characteristics. Each was located in a thick island clump of Carex along the edge of a wetland about 1 to 2 m from shore. Other unconfirmed sites based on hens flushing or observing hens returning from incubation breaks were in similar habitats. However, at least one mallard hen returning from an incubation break headed straight into the forest. A mallard brood was later seen on this pond, thus it seems likely her nest was in the nearby forest. �63 Table 10. Duck pair densities within selected Routt national forest, Colorado, 1989. wetland size classes Wetland Size Classes {ha}a 0.1 - 0.2 0.2 - 0.8 on the Species < 0.1 Mallard 0.8 2.4 1.4 1.6 0.6 0.2 0.9 0.3 Green-winged teal > 0.8 Cinnamon teal 0.4 0.4 0.3 0.6 American wigeon 0.0 0.0 0.0 0.3 0.2 0.4 0.9 2.3 0.0 0.0 0.3 0.9 2.1 3.5 3.9 6.2 90 88 67 45 10 12 33 55 Ring-necked duck Bufflehead All species % Dabblers % Divers (N = 88) (N = 59) a Total amount of water (ha) available 0.2 = 4.6, 0.2-0.8 = 11.7, > 0.8 = 12.1. in each size are: < 0.1 4.8, 0.1- Several other important wetland habitat variables were noted qualitatively during the course of the investigation. Invertebrate populations were sampled in emergent vegetation along several wetland margins in late May and early June, and large numbers of snails, leeches, fresh water shrimp, and other small crustaceans were found. Most ponds appeared to have a sufficient number of invertebrates to support local waterfowl pairs. Although we did not sample during the summer months, invertebrate densities likely dropped in many of the moraine ponds when most emergent vegetation and several entire ponds dried up as the summer progressed. As a result, invertebrate densities for broods were probably lower and more patchy in their distribution than they were for pairs. Upland nesting cover was readily available, especially around beaver ponds where dense thickets of willows and sedges were common. Moraine ponds generally had less available cover, since most were surrounded by lodgepole pines with little understory. Even so, there were still numerous sites with tall brush, windfall timber, and thick wetland vegetation for hiding nests. Potential nest cavities for bufflehead may have been more limiting because (1) large aspens were generally absent from many suitable ponds due to past beaver activity and (2) northern flickers (Colaptes auratus), which are the main cavity excavators, were uncommon on the study area. Water quality in the study wetlands was excellent; no major upstream impacts such as forest fires, clear-cutting, mining or other commercial �64 development activities have taken place for at least for several decades. Water stability in the beaver ponds (except those in the Cinnamon drainage) is dependable throughout the summer. In contrast, the moraine ponds appear to be in a current drought cycle and, based on historical aerial photographs, have likely been in decline since 1984. Several moraine ponds went dry before the end of the brood rearing period this year. Without significant snowfall this winter, most of them will probably be in extremely poor condition next spring. DISCUSSION Breeding pair densities in this portion of the Routt National Forest were much higher than either the 1.4 pairs/km2 reported by Rutherford and Hayes (1976) or the 0.6 pairs/km2 found by Frary (1954) in forested habitats elsewhere in Colorado. Compared to grassland associations (Bellrose 1979), our observed dabbling duck density of 4 pairs/km2 is almost equal to that found on the short and tall grass prairies (6-8 pairs/km2), but is much lower than that on the mixed grass prairies and parklands (24-30 pairs/km2). Species compositions were also different from other montane studies. Although mallards and green-winged teal were commonly found in all 3 studies, we had no nesting gadwall (Anas strepera; Rutherford and Hayes 1976), and several species considered uncommon elsewhere were much more abundant in our study area. These included cinnamon teal, ring-necked duck and bufflehead. In fact prior to this study, no breeding records of bufflehead had existed in the state of Colorado (Ringelman and Kehmeier 1989). Wetland use by waterfowl pairs was strongly related to pond size. Large ponds may be more attractive to ducks because they provide a greater degree of water permanency, a more dependable food supply, better access in a forested environment, and more available space for inclusion within a pair's horne range, thus reducing the need to divide time among several smaller, more scattered wetlands. Large wetlands may provide enough space for 2 or more pairs of the same species. Also, because of their greater range of water depths, large ponds may offer a more diverse habitat which can be exploited by several different duck species simultaneously. Wetland habitat diversity leads to greater species diversity, which translates to greater use days/pond. A secondary factor influencing habitat use by waterfowl pairs was pond origin. Wetlands created by beavers had greater use than did those formed by glacial action. Of the 14 related pond origin variables, human disturbance was the one most correlated with pair use. Beaver ponds which experienced the heaviest fishing activity tended to be those located close together in large complexes or series. Such complexes were most attractive to fishermen if they also contained a few large, deep ponds with good trout potential scattered among the group. Duck pairs attracted to these larger wetlands also readily used smaller, neighboring ponds, which in some cases were only 3 to 5 m away over a beaver darn. As a result, small wetlands that might not have been used had they been located in a more isolated situation experienced greater than expected pair use. This heterogeneous nature of beaver pond complexes likely increased the effective size of each wetland member, attracting pairs in densities similar to those of larger individual wetlands. In summary, because of the large number and diversity of breeding ducks attracted to this part of the Routt National Forest, it appears that there is an adequate amount of wetland habitat and a sufficient mixture of different wetland types available. If one also considers the large early season �65 invertebrate populations, the abundance of suitable upland nesting cover, and the reliable sources of May pond water, this area probably provides as high a quality wetland base for nesting waterfowl as one will find at this elevation. The only limiting factor in this phase of the reproductive cycle may be the availability of nesting cavities for buffleheads. Based on the general lack of re-pairing and the high proportion of pairs for whom broods were observed, nest success appears to be very high in the Routt National Forest. Our estimated rate of 50% to 70% nest success is much higher than the 10% to 15% commonly found in the northern prairies (Cowardin 1983). These high rates are most comparable with those obtained from intensively managed lands (i.e. nesting islands, peninsula cut-offs, electric fenced fields) throughout the prairie pothole region (Lokemoen et al. 1982, Duebbert et a1. 1983). Potential nest predators in our study area included striped skunk (Mephitis mephitis), coyote (Canius 1atrans), mink (Mustela vison), badger (Taxidea taxus), long-tailed weasel (Mustela frenata), and common raven (Corvus corax). However, few of these species were abundant and widespread. Other possible sources of nest loss were from flooding (in early July all irrigation ditches previously diverting water for hay in North Park were shut off to prepare fields for harvesting) and inclement weather (i.. e. snow, hail, falling trees). Overall, these other events probably effected only a small portion of the total clutches initiated. Consequently, this portion of the Routt National Forest appears to be as safe a nesting area as those managed with much greater intensity and financial resources. Brood survival rates varying from 27% to 81% have been reported (Ball et al. 1975, Ringelman and Longcore 1982, Talent et al. 1983, Duncan 1986). Since our observed rate of 54% and estimated rate of 65% is in the upper range of these values, total brood loss is probably not excessive. Mean brood sizes at fledging for all these previously cited studies was around 5.0 - 5.5 ducklings. Similarly, Frary (1956) found average brood sizes of 5.6 ducklings for mallard and green-winged teal in the montane habitat of northwest Colorado. Thus the mean brood size of 4.1 ducklings estimated for the Routt National Forest is slightly below that expected, indicating that partial brood loss may be a problem. Reasons for this increased rate of duckling attrition are largely speculative. On numerous occasions we observed that brood cohesion was relatively weak, even for class I ducklings. Downy young were often seen scattered about the pond feeding alone or in small groups, becoming strung out behind females while traveling through thick floating vegetation, or being left behind when females moved to neighboring beaver ponds 10 to 300 m away. Orphaned ducklings are likely to die from predation or exposure during frequent mountain thunderstorms and chilly nights. In addition, many broods were forced to feed away from cover, since water levels in beaver ponds were too stable for an emergent cover zone to develop and moraine ponds were too unstable for this zone to contain water past early July. This exposure increases their likelihood of succumbing to predators or environmental factors. By mid-summer, only the deeper floating leave, submergent and open water zones were available for broods. Because the emergent zone is shallower and has correspondingly greater light intensities, it has greater primary production and stem density than does the deeper submergent zone (Westlake 1965, Moss 1980). This productivity and habitat complexity in turn supports an abundant and rich diversity of invertebrates (Moss 1980), a critical food �66 source for ducklings during their first 2 weeks of life (Sudgen 1973, Reinecke 1977). The shallow water of the emergent zone may help concentrate these invertebrates so they can be more easily reached. Without emergent areas, ducklings are forced to search for food in deeper water zones where invertebrate densities may are generally lower, more unevenly distributed and more difficult to reach. Invertebrate biomass also tends to decrease with the age of beaver ponds (Reinecke 1977). Based on aerial photographs, most beaver ponds in our study area appeared to be more than 15 years old. Thus, while emergent vegetation zones are available in spring, reduced availability and productivity during summer may create local food shortages for ducklings. If such a food shortage hypothesis is correct, one might expect ducklings to (1) spend more time searching for food, (2) spread out more while feeding to reduce competition with siblings, (3) have increased chances of being separated from the rest of the brood, and (4) be in poorer body condition and hence more suspectable to cold, disease, and predators if separated. One might also expect brood females to frequently switch ponds in search of better feeding areas, particularly if alternatives wetlands were close by. Larger wetlands might be more attractive to brood hens because of their greater degree of water permanency, greater likelihood of containing water lilies and perhaps other submergents, and higher probability of constant beaver activity maintaining a dynamic wetland community. Earlier hatching broods might be expected to have higher survival rates than later hatching broods due to the better water conditions present and more abundant food resources available earlier in the year. Beaver ponds might fledge more ducklings than moraine ponds due to their more diverse submergent zones, the possibility of beavers raising their dams and flooding new upland cover, and the presence of small beaver channels offering some protected shallow water feeding areas. Thus the ultimate factor most likely limiting waterfowl production in this section of the Routt National Forest is the lack of cover and reduced food supplies caused by the gradual loss (or lack of) flooded emergent vegetation as the season progresses. MANAGEMENT RECOMMENDATIONS Duckling survival appears to be the phase of the reproductive cycle most limiting waterfowl production in the Big Creek Lakes area of the Routt National Forest. Buffleheads may lack enough suitable nest cavities. Other potential bottlenecks such as pond numbers, wetland quality, nesting cover, and rates of nest loss appear to be of little or no concern. Thus all management efforts are best directed at improving duckling survival rates and bufflehead nest sites, two actions which will return the largest benefit per dollar invested. Brood separation and subsequent duckling losses due to exposure or predation, resulting from insufficient amounts of flooded emergent cover, appears to be the factor most likely causing smaller than expected broods at fledging. Dense concentrations of cowlilies on many ponds probably contribute to brood separation and impede efforts by the hen to regroup ducklings. The dominant emergent plants currently found on the study area are 3 species of sedges, all typically found growing in < 0.3 m of water (Windell et al. 1986). Consequently, they are very susceptible to drying out in moraine ponds and unlikely to become established in beaver ponds, since the latter �67 usually have deep waterline edges caused by beavers collecting mud for their lodges and darns. For improving waterfowl production, we suggest that an emergent species having a wider range of water depth tolerances be introduced into the area. The most promising candidate is hard-stern bulrush, which grows in acidic, fresh, or slightly alkaline water up to 1 m deep and is found in dense stands at Lake John (2553 m) and another site in the Routt National Forest (2530 m), both less than 30 krn away. Establishing this species is relatively easy; individual root stocks should be firmly shoved into a soft bottom in either spring or fall at rates of 1000 plants/acre (Lemberger 1981). Since the study area is at the upper elevational limits for this species, it might be best to arrange the plantings in several large clumps in order to assure some individuals survive and establish a strong network of rootstocks. Attempts to establish hardstem bulrush stands from seeds may be more difficult because of the need to scarify seeds (spring seeding) or control water levels (fall seeding) (Harris and Marshall 1960), and because the colder sediment temperatures at this elevation may cause poor germination. In order to improve duckling mobility on ponds with dense cowlily growths, we suggest using an aquatic herbicide (such as Rodeo, a trademarked product from Monsanto) to create a 3 m network of travel lanes throughout a pond. Since beavers readily feed on cowlilies, once such channels are established these rodents may continue to maintain them for their own use. Beaver activity on several ponds is currently maintaining an attractive mix of open water and lily beds. However ponds that are very large or have not had beavers occupy then for several years are completely choked. Thus this management technique should be regarded as more of a supplemental method of improving waterfowl production than as a permanent large scale necessity. Buffleheads will readily use artificial nesting boxes (Gauthier 1988). These should be constructed with maximum dimensions of 15 x 15 x 40 cm with a 6.5 cm hole in order to minimize competition with other species. Since buffleheads seem to prefer choosing from several sites, 3 or 4 boxes should be placed around each wetland. Higher use and success rates will most likely occur from boxes placed 3 to 7 m high in heavy coniferous forest, close to the wetland edge with few obstructions in front of them. Since bufflehead pairs were only observed on the larger wetlands or beaver pond complexes, nesting box efforts should be concentrated on such wetlands, particularly those currently used by buffleheads. For each of these 3 management recommendations, we suggest selecting 2 wetlands in each experimental unit (NFN Platte and Goose Creek) for purposes of the evaluation. In order to avoid cross experimental effects, a separate wetland should be selected for each treatment. Thus a total of 12 wetlands will be used (Table 11). A similar set of 12 wetlands should be identified in the 2 control units (Forester and Shafer Creeks). This will allow for both wetland and unit comparisons during the 2 year evaluation, an important consideration since there will be multiple treatments occurring simultaneously in each experimental unit. The effectiveness of each management technique will be determined by the difference in mean brood size at fledging between the 1989 pre-treatment data and the 1990-91 treatment data, adjusted for any such changes occurring in the control units during this same time (nonexperimental effects). A fourth management recommendation, which is separate from the other 3 because there is no available replicate, is that of rejuvenating a series of 19 beaver ponds in the so-called "cinnamon drainage" within the Goose Creek �68 Table 11. Wetlandsa recommended for specific management action in the experimental units along with those wetlands designated as being their experimental controls. Management Treatments wetlands N. Fork N. Platte Goose action Hard-stern bulrush Herbicide Bufflehead a planting spraying nest boxes See Figures 2-5 for the exact Control Forester wetlands. Shafer 67,86 35,53 1,41 20,23 80,85 18,33 26,42 14,18 13,83 22,37 15,50 22,28 location of each wetland listed. unit. This drainage system, which had 5 or 6 breeding pairs of ducks in May, fledged only 1 brood totaling 4 ducklings. A major reason for this poor production is that little beaver activity has occurred in this area since 1980, based on annual growth rings of previously damaged trees (Lawrence 1952). As a result, many of the beaver darns are deteriorating and no longer hold water. Although the area has extensive groves of aspen saplings that should be attractive to beaver, very little water flows through the complex after 31 May. An upstream beaver pond complex has changed flow patterns so that most of the water is currently being diverted away from the cinnamon drainage. To rejuvenate this series of ponds, we recommend using 2 sets of beaver tubes (Laramie 1963) to divert a small amount of water from a beaver pond located on Goose Creek into a second beaver pond complex, then out into the cinnamon drainage (Fig. 9). This will likely attract beavers to recolonize the area, repair the present darns, and re-flood the ponds. These new ponds should provide an abundant source of invertebrates for ducklings as well as additional habitat for breeding pairs and broods. This portion of the Routt National Forest is very high quality waterfowl habitat, with pair densities that rival some prairie breeding areas and nest success rates that exceed all but the most intensively managed waterfowl habitats. If one or more of our proposed recommendations prove effective, the Forest Service will have an outstanding example of the waterfowl production potential that exists in some forested habitats. By using the methodologies and results from this study to improve areas of similar or lower quality, enhanced waterfowl production with continental significance could be achieved. In addition, habitat information indicating that waterfowl pair use is highly correlated with pond size and origin can be readily incorporated- into predictive models. Such models, when combined with remote sensing technology, could prove invaluable for estimating total waterfowl production and identifying areas where waterfowl management projects exist on other National Forest land in Colorado. �~ iu W 0: o WM C/) o o o \ \ ~, " - . " "I (.._.r.-.. -.~..--. ---__./'/ . I \..- : I 69 ~, ,,- * �70 LITERATURE CITED Ball, I. J., D. S. Gilmer, L. M. Cowardin, and J. H. Reichman. 1975. Survival of wood duck and mallard broods in north-central Minnesota. Wildl. Manage. 39:776-780. J. Bellrose, F. C. 1979. Species distribution, habitats, and characteristics breeding dabbling ducks in North America, Pages 1-15 in T. A. Bookhout (ed.), Waterfowl and wetlands-an integrated review. La Crosse Printing Co., Inc., Wisc. Colorado Division of Wildlife. 1989. Colorado state waterfowl plan. Colorado Division of Wildlife, Fort Collins, CO. management Cowardin, L. M., V. Carter, F. C. Golet, and E. T. LaRoe. 1979. Classification of wetlands and deepwater habitats of the United U.S. Dept. Interior FWS/OBS-79/3l. 103pp. Cowardin, L. M., A. B. Sargeant, potentials for prairie ducks. and H. F. Duebbert. Naturalist 34:4-11. Duebbert, H. F., J. T. Lokemoen, and D. E. Sharp. of mallards and gadwall on Miller Lake Island, Mange. 47:729-740. 1983. 1983. North A. J. Frary, L. G. Colorado. 1972. Buffleheads. Can. Wildl. States. Problems anp Concentrated nesting Dakota. J. Wildl. Duncan, D. C. 1986. Survival of dabbling duck broods on pra~r~e in southeastern Alberta. Can. Field-Nat. 100:110-113. Erskine, of Servo Monogr. impoundments Ser. 4. 240 pp. 1954. Waterfowl production on the White River Plateau, M. S. Thesis, Colorado State University, Ft. Collins. Gauthier, G. 1988. Factors affecting nest-box use by buffleheads cavity-nesting birds. Wildl. Soc. Bull. 16:132-141. 93 pp. and other Gollop, J. B. and W. H. Marshall. 1954. A guide for aging duck broods the field. Unpublished report, Mississippi Flyway Council Tech. Sec., 14 pp. (mimeo). in Harris, S. W. and W. H. Marshall. 1960. Germination and planting experiments on soft-stem and hard-stem bulrush. J. Wildl Manage. 24:134-139. Klett, A. T., H. F. Duebbert, C. A. Faanes, and K. F. Higgins. 1986. Techniques for studying nest success of ducks in upland habitats in the prairie pothole region. U.S. Fish Wildl. Servo Resour. Publ. 158. 24 pp. Laramie, H. A. Jr. 1963. Manage. 27:471-476. Lawrence, W. H. 1952. Manage. 16:69-79. A device Evidence for control of problem of the age of beaver beavers. ponds. J. Wildl. J. Wildl. �71 Lemberger, Wise. J. J. 1981. 32 pp. What brings them in? Wildlife Nurseries, Lokemoen, J. T., H. A. Doty, D. E. Sharp, and J. E. Neaville. fences to reduce mammalian predation on waterfowl nests. Bull. 10:318-323. Moss, B. 1980. Ecology of freshwaters. Inc., New York. 332 pp. Halsted Press, Oshkosh, 1982. Wildl. Electric Soc. & John Wiley Sons, Reinecke, K. J. 1977. The importance of freshwater invertebrates and female energy reserves for black ducks breeding in Maine. Ph.D. Dissertation, Univ. of Maine, Orono. 113 pp. Ringelman, J. K., and J. R. Longcore. 1982. Survival of juvenile during brood rearing. J. Wildl. Manage. 46:622-628. Ringelman, J. K., and K. J. Kehmeier. Colo. Field. Ornith. (in press). Buffleheads breeding black in Colorado. Rutherford, W. H. and C. R. Hayes. 1976. Stratification as a means improving waterfowl surveys. Wildl. Soc. Bull. 4:74-78. Sudgen, L. G. and lesser 1973. scaup for Feeding ecology of pintail, gadwall, American wigeon, ducklings. Can. Wildl. Servo Rep. Ser. No. 24. 45 pp. Talent, L. G., R. L. Jarvis, and G. L. Krapu. 1983. Survival broods in south-central North Dakota. Condor 85:74-78. Walters, C. J. 1986. Press, New York. ducks Adaptive 374pp. management of renewable of mallard resources. Macmillan Westlake, D. F. 1965. Some basic data for investigators of the productivity of aquatic macrophytes, p. 231-248. In C. R. Goldman (ed.), Primary productivity in aquatic environments. Mem. 1st. Ital. Idrobiol., 18 Suppl., Univ. of California Press, Berkley. Windell, J. T., B. E. Willard, D. J. Cooper, S. Q. Foster, C. F. Knud-Hansen, L. P. Rink, and G. N. Kiladis. 1986. An ecological characterization of Rocky Mountain montane and subalpine wetlands. U. S. Fish Wildl. Servo Biol. Rep. 86(11). 298 pp. �72 APPENDIX A. CLASS VARIABLE CODING SHEET Dominant wetland covertype 1 submergent 2 floating 3 emergent leave 4 moist 5 woody 6 organic 2 beaver soil Pond origin 1 = moraine Water Permanency Beaver 1 0-10% 4 51-70% 2 11-30% 5 71-100% 3 31-50% I active activity o Grazing = inactive intensitv by cattle 1 none 4 extensive 2 slight 5 severe 3 moderate Relative degree of human disturbance 1 none 4 extensive 2 slight 5 severe 3 moderate �73 APPENDIX B. DUCK PAIRS CENSUS ED ON THE ROUTT NATIONAL FOREST, COLORADO. Week of 29 May, 1989. Unit S:Qecies Mallard Lm ." If. b pro C North Fork of North Platte 3 0 8 Forester Creek Goose Creek 0 6 0 3 1 1 Shafer Creek 2 0 0 Total 11 1 12 Lm . Greenwinged teal lf. pro 3 0 0 0 0 0 1 0 1 3 0 0 0 1 Cinnamon teal 1m. lf. pro 0 0 0 0 0 0 3 0 1 0 0 0 3 0 1 Ring-necked duck Lm . 1f. pro 1 0 2 0 0 2 1 0 5 0 0 3 2 0 12 American wigeon 1m. 1f. pro 0 0 1 0 0 0 0 0 1 0 0 0 0 0 2 Bufflehead 1m. if. pro 0 0 1 0 0 1 0 0 1 1 0 1 1 0 4 Common merganser 1m. If. pro 0 0 1 0 0 0 0 0 0 0 0 0 0 0 1 All species 1m. if. pro 7 0 0 6 11 1 10 6 0 13 0 4 24 1 33 Appx. no. pairs 20 6 22 10 58 alone males b10ne females "pa i r s 7 �74 Appendix B. (Continued) 1989. Week of 5 June, Unit SI2ecies Mallard 1m. a If.b pr. C Greenwinged North Fork of North Platte 3 1 4 Forester Creek 4 1 1 Shafer Creek Total 13 3 8 4 2 0 3 1 0 1 0 0 5 0 0 1m. teal If. pro 2 0 0 1 0 0 1 0 0 1m. 1f. pro 0 0 4 0 4 0 0 0 0 0 0 1 0 1 1m. 1f. pro 1 0 2 0 0 1 Cinnamon teal Ring-necked duck American wigeon 1m. If. pro 1 0 0 1 0 0 1m. 1f. pro 1 0 0 0 1 0 1m. If. pro 0 0 0 1m. 1f. pro Appx. no. pairs Bufflehead Common merganser All species alone males b10ne females "pa i r s Goose Creek 0 1 0 0 4 3 0 0 1 0 3 0 10 1 0 3 2 0 0 1 0 0 2 0 1 3 1 0 2 0 2 0 0 0 0 0 0 0 0 8 1 7 5 2 11 4 28 0 3 6 2 10 4 24 16 7 21 10 54 2 �75 B. (Continued) Appendix Week of June 12, 1989. hens with broods. Bold face type denotes nesting hens; italics indicates Unit SIlecies Mallard 1m. a If.b pro C Greenwinged 1m. teal If. pro North Fork of North Platte 3 1,1,1 0 Forester Creek 1 1,1 1 Goose Creek 1 0 Shafer Creek 2 1 0 5 Total 11 2,3,2 1 2 0 0 0 0 2 1 0 2 1 0 0 4 0 4 Cinnamon teal 1m. 1f. pro 1 0 0 0 0 0 3 1 1 0 0 0 4 Ring-necked duck 1m. 1f. pro 1 0 0 2 0 0 9 0 2 1 0 13 0 4 6 American wigeon 1m. 1f. pro 0 0 0 0 0 0 2 0 0 0 0 0 2 0 0 Bufflehead 1m. If. pro 1 0 0 0 0 0 1 0 0 0 1 0 2 1 0 Common merganser 1m. If. pro 0 0 0 0 1 0 0 0 0 0 0 0 0 1 0 All species 1m. 1f. pro 9 1,1,1 0 3 1,1,1 3 21 4 1,1 2 5 4 37 3,3,2 12 1 1 alone males "Lone females "pa i.r s .: �76 Appendix Week of June 19, 1989. hens with broods. B. (Continued) Bold face type denotes nesting hens; italics indicates Unit North North S12ecies Mallard Greenwinged 1m. a If .b pr. C 1m. teal If. pro Fork of Platte 0 Shafer Creek 1 0 0 0 0 0 0 1 0 3 0 0 1 0 0 2 0 Forester Creek 2 2,2,1 0 0 0 0 Goose Creek 0 3 Total 3 5,2,3 0 3 Cinnamon teal 1m. If. pro 0 0 0 0 0 0 2 1 0 0 0 0 2 1 0 Ring-necked duck 1m. lf. pro 1 0 0 1 0 0 7 0 2 1 0 4 10 0 6 American wigeon 1m. If. pro 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Bufflehead 1m. If. pro 0 0 0 0 0 0 0 1 0 0 1 0 0 2 0 Common merganser 1m. If. pr. 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 All species 1m. lf. pro 1 2 0 3 2,2,2 9 5,1 18 8,2,4 3 3 5 1 4 alone males blone females cpairs 7 �77 Appendix Week of June 26, 1989. hens with broods. B. (Continued) Bold face type denotes nesting hens; italics indicates Unit North North S2ecies Mallard Greenwinged 1m. a 1f. b pr. C 1m. teal lf. pro Fork of Platte 1 0 0 Forester Creek 4 4,1 2 Goose Creek 2 2 0 Shafer Creek 3 1,1 0 0 1 0 1 0 0 0 2 1 0 0 0 3 Total 10 3,4,2 2 1 1 Cinnamon teal 1m. lf. pro 0 0 0 0 0 0 0 1 0 0 0 0 0 1 0 Ring-necked duck 1m. If. pro 1 0 1 1 0 0 3 0 1 2 0 1 7 0 3 American wigeon 1m. If. pro 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Bufflehead 1m. lf. pro 0 2 0 0 0 0 0 1 0 0 3 0 0 6 0 Common merganser 1m. 1f. pro 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 All species 1m. If. pro 2 6 4,1 2 5 5 2,1 4,2 4,1 1 18 10,4,5 6 alone males b10ne females cpairs 1 2 �78 Appendix W'eek of July 3, 1989. hens with broods. B. (Continued) Bold face type denotes nesting hens; italics indicates Unit S:gecies Mallard North Fork of North Platte Lm .a If.b pr. C Greenwinged 1m. teal If. pro 0 1 Shafer Creek 3 1,1 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Forester Creek Goose Creek 0 1 1 4 0 0 0 0 0 Total 7 2,2 Cinnamon teal 1m. If. pro 0 0 0 0 0 0 0 0 0 0 0 0 Ring-necked duck 1m. If. pro 0 0 0 0 2 0 0 0 2 1 1 1,1 1 1m. 1f. pro 0 0 0 0 0 0 0 0 0 0 0 1m. If. pro 0 2 0 0 0 0 1 3,1 0 0 0 0 0 0 Cornmon merganser 1m. If. pro 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 All species 1m. If. pro 0 4 2 1 1 9 6,5 0 0 3 3,1 1 American wigeon Bufflehead alone males blone females cpairs 1 0 1 0 2,3 1 1 1 0 2 �79 APPENDIX OBSERVATION C. LOG FOR EACH INDIVIDUAL BROOD Brood Species 2) Green-winged 3) Green-winged teal teal 4) Mallard 5) Mallard size age" F-43 6-6 8 Ia Yes F-41 6-20 5+ >Ia Y F-43 6-27 8 Ie Y G-21 6-19 6 Ia Y G-24 6-26 7 Ib Y G-24 6-26 3+ Ia Y P-6 6-14 7+ Ia Y P-35 6-21 7 Ib Y P-16 6-28 6 Ic Y P-6/15 7-10 5 lIb Y P-19 7-14 2+ lIb y P-14 7-25 5 II No P-13 8-15 5 Juv N P-37 6-21 8 Ia Y 8-2 8 III N P-6/7 6) Green-winged 7) Mallard teal Hen present Date pond 1) Mallard No.a P-6 6-28 2+ Ia Y S-18 6-29 2 Ib y S-18 7-7 2 Ic Y S-18 7-13 2 IIa Y .: �80 Appendix c. (continued) age Hen present Brood Species pond No. Date size S-18 7-19 2 lIb Y S-18 7-26 2 IIc Y S-18 8-16 2 juv N 8) Mallard S-18 6-29 5 Ic Y 9) Bufflehead F-SO 7-5 7 Ia Y F-50 7-14 7 Ib Y F-50 7-19 6 Ib y F-50 7-21 6 Ic y F-50 7-26 6 IIa Y F-50 8-1 7 lIb Y F-SO 8-4 6 lIb Y F-50 8-8 5 IIc Y F-SO 8-11 5 IIc Y F-SO 8-16 5 III ~ P-13 7-6 6 Ia Y P-79 7-6 6 Ia Y P-83 7-6 6 Ia Y P-83 7 -13 4 1b y P-83 7-18 1 Ie Y P-83 8-8 2 IIc Y 7) Mallard 10) Ring-necked 11) American 12) Mallard duck wigeon .: �81 Appendix C. (continued) Brood Species 13) Cinnamon 14) Cinnamon pond No. teal teal 15) Green-winged teal 16) Bufflehead 17) Ring-necked duck Date size age Hen present G-45 7-11 5 Ia Y G-45 7-16 2 Ib N G-45 7-21 6 Ie N G-50 7-24 5 IIa N G-45/50 8-14 4 juv N G-18 7-10 4 Ie Y G-18 7-24 3 IIb Y F-50 7-12 7 Ia Y F-50 7-19 1 Ib Y F-50 7-26 3 Ie Y F-50 8-1 1 IIb Y F-50 8-11 2 lIe N $-18 7 -13 2 IIb Y S-18 7-19 2 IIb Y $-18 7-26 2 lIe Y $-18 8-3 2 III y S-18 8-10 2 III N $-18 8-16 2 Juv N P-83 7-13 6 Ia Y P-83 7-18 6 Ib Y P-83 7-25 6 IIa y �82 Appendix C. (continued) Brood Species pond No. Date size age Hen present P-83 8-1 5 IIb Y P-83 8-8 4 IIc Y P-83 8-14 5 IIc Y F-SO 7-14 7 III Y F-SO 7-19 3+ III Y F-SO 7-26 9 III Y F-SO 8-1 6 Juv N F-SO 8-4 6 Juv N F-SO 8-8 8 Juv N F-SO 8-16 7 Juv N F-IS 7-14 8 Ia Y F-21 8-9 1 lIb Y 20) Mallard G-21 7-17 1 IIa Y 21) Mallard G-30 7-19 2 Ic Y G-18 7-20 3 Ic Y G-19 8-1 2 lIb Y G-19 8-4 2 IIc Y G-24 8-14 1 IIc y G-IS 7-17 6 Ia Y G-IS 7-20 3+ Ia Y G-IS 7-21 6 Ia Y 17) Ring-necked duck 18) Mallard 19) Mallard 22) Ring-necked duck �83 Appendix C. (continued) Hen present Brood Species 22) Ring-necked 23) Green-winged pond No. duck teal 24) Mallard 25 ) Ring-necked duck 26) Mallard 27) Green-winged 28) Mallard teal Date size age G-18 7-24 6 Ib Y G-18 7-27 5 Ib Y G-19 8-1 1 Ic Y G-21 8-14 1 lIb N G-2l 8-29 1 III N G-18 7-17 4 IIc Y G-2l 8-4 4 III N P-68 7-18 5 Ia Y G-18 7-24 3 Ic Y G-18 7-27 3 Ic Y G-2l 8-1 2 IIa Y G-18 8-7 3 lIb Y G-18 8-14 3 IIc Y S-28 8-3 8 III N S-28 8-3 2 Ia Y S-28 8-11 2 Ib Y P-54 8-8 2 lIb Y a Wetland letter codes are as follows: P = N. Fork N. Platte, Creek, G = Goose Creek, and S = Shafer Creek. b Age classes flying young. are after Gallop and Marshall (1954). Juv F Forester juvenile or ��85 Colorado Division of Wildlife Wildlife Research Report September 1990 JOB PROGRESS REPORT State of Colorado Project 01-03-212 Work Plan 22 Job Title: Author: : Job Migratory Period Covered: Migratory Game Bird Research _2_ Bird Publications 01 April 1989 through 31 March 1990 Michael R. Szymczak Personnel: James K. Ringelman Wildlife and Michael R. Szymczak, Colorado Division of ABSTRACT The following list contains those articles that were prepared and/or submitted for publication or published during this segment: Ringelman, J. K., W. R. Eddleman, and H. W. Miller. 1989. High plains reservoirs and sloughs. Pages 311-340 in L. M. Smith, R. L. Pederson, and R. M. Kaminski, eds .. Habitat management for migrating and wintering waterfowl in North America. Texas Tech Press, Lubbock. Ringelman, Colorado. J. K., and K. J. Kehmeier. 1990. Colo. Field Ornithol. 24:46-48. Ringelman, J. K. 1990. Buffleheads. Buffleheads Colo. Outdoors. breeding in 39:8-9. Ringelrnan, J. K. 1991. Differential harvest of male and female mallards under conventional and point system regulations. Wildl. Soc. Bull. (submitted for review) Ringelman, J. K. 1991. Evaluating and managing waterfowl Colorado. Colo. Div. Wildl. Spec. Rep. (In prep.) habitat in Rusch, D. H., C. D. Ankney, H. Boyd, J. R. Longcore, F. Montalbano III, J. K. Ringelman, and V. D. Stotts. 1989. Population ecology and harvest of the American black duck: a review. Wildl. Soc. Bull. 17: 379-406. � https://cpw.cvlcollections.org/files/original/11b497b1e4d395bd13b75c856f99532e.pdf 534ccfcb3cdee989760316c875b0b6c9 PDF Text Text 1 JOB PROGRESS REPORT State of ~C~o~l~o~r~a~d~o _ Project: (W-150-R-4) Period Covered: 1 July, 1990 - 30 June, 1991 : Peregrine Falcon Restoration Program Personnel: G.R. Craig, Colorado Division of Wildlife and J.H. Enderson, The Colorado College. ABSTRACT In the 1991 peregrine breeding season, 58 territories were occupied by 49 breeding pairs that fledged 91 young. The West Slope subpopulation consisted of 42 occupied sites while East Slope occupancy increased to 16 sites. Eggshell thicknesses in 1990 averaged 10.2% thin. Contents of 12 nonviable eggs that were analyzed for organochlorines yielded relatively low levels. Due to population numbers, fostering and hacking efforts were discontinued. This Job Progress Report represents a preliminary analysis and is subject to change. For this reason, information presented herein MAY NOT BE PUBLISHED OR QUOTED without permission of the author. ��3 PEREGRINE FALCON RESTORATION PROGRAM Gerald R. Craig SEGMENT OBJECTIVES 1. Annually monitor the number of breeding pairs of peregrines and their reproductive success in Colorado. 2. Annually monitor organochlorine peregrines in Colorado. 3. Monitor breeding population turnover through band recoveries, presence of color markers, and telephotographic identification of individual breeding adults. 4. Augment poor wild production by placement of captive hatched wild young and captive produced young into occupied wild nests. 5. Release captive hatched and captive produced young at potential and vacant wild territories. 6. Monitor recruitment of reintroduced peregrines into the wi ld breeding population of Colorado. pesticide levels in wild breeding METHODS AND MATERIALS 1. Visit all known peregrine breeding territories throughout Colorado and observe them from a distance to establish the presence of breeding adults. Breeding pairs will be kept under surveillance to determine initiation of egg laying. Depending upon the individual female's reproductive history and eggshell condition (obtained through measurement of previous year's eggshell thicknesses) and availability of captive hatched young for release, breeding pairs either will monitored or manipulated as outlined in approach 4. Those pairs not designated to be manipulated will be revisited periodically throughout the nesting season to document reproductive success. When a pair's behavior indicates that egg laying has occurred and incubation is underway, the eyrie wi 11 be visited to document the number of eggs produced. The eggs will be candled to ascertain viabi 1ity and approximate age. All nonviable eggs will be collected for chemical analysis. A second visit will be made to determine productivity, band nestlings, and collect eggshell fragments and unhatched eggs for thickness measurement and analysis under 2a and 2b. 2a. Eggshell fragments encountered during eyrie visits described in approaches 1 and 4a will be measured for index to thickness following standardized procedures. 2b. Whole, nonviable eggs which are encountered during eyrie visits will be collected, preserved and submitted to the appropriate Fish and Wildlife Service approved -laboratory for pesticide analysis. Eggs collected from the wild in the course of Approaches 4a, 4b and 4c that are artificially �4 at the Peregrine Fund's Boise, Idaho facility also will be submitted for shell thickness measurement and chemical analysis. 3. Peregrines present at breeding territories will be examined to determine the presence of bands or color markers. Band confi rmation will be accomplished through observation from a distance with telescopes and concealed remote controlled cameras. When banded falcons are encountered, every effort wi 11 be made to read band numbers without trapping or hand1 ing the bi rds. It is possible this can be accompl ished in most situations with a Questar field model telescope (80-130x). When band numbers cannot be discerned, attempts will be made to trap and examine the falcon at a time when capture will have least impact upon breeding activities. 4a. In accordance with an annual release plan developed and approved by the state, U.s. Fish and Wildlife Service, Bureau of Land Management, National Park Service, and the Forest Service, a predetermined number of wi ld breeding pairs will be manipulated to augment natural productivity. Pairs with a history of reduced clutch size, cracked eggs, or infertile or dead eggs will be candidates for fostering efforts. 4b. On occasion, it may be necessary to recycle several early breeding pairs in order to delay them until captive hatched young of the proper age are available for placement into wild sites. No later than 10 days after the last egg has been deposited, the eyrie will be visited and the entire clutch removed without replacement. Approximately 14 days after removal of the clutch, the pair will recyc 1e, se1ect anothe r nest 1edge, and deposit a second clutch of eggs. If the eggs are thin shelled, they may be replaced with plastic replicas and treated as outlined in approach 4a. This technique also works well to augment captive production with wild produced eggs. 4c. At times, pairs will select inferior eyrie ledges that may compromise nest success such as ledges that are too narrow to support a brood of large nestlings, the site may be vulnerable to predators, or it may be exposed to the elements. If the ledge cannot be mechanically improved, pairs can be relocated to other ledges through the recycling method described in approach 4b since they invariably relocate and select a new ledge when recycled. 5. In accordance with an annual release plan developed and approved by the State, U.S. Fish and Wildlife Service, Bureau of Land Management, National Park Service, and the Forest Service, a predetermined number of captive produced falcons will be released at unoccupied or potential sites through the technique of hacking. This technique is employed at locations that do not have the benefit of protection or care from adults. Young falcons of about 35 days of age will be placed in a hack box on a suitable cliff ledge at the reintroduction site. They wi 11 be fed and cared for by attendants until they are flying and capable of fending for themselves. This technique assures that the young become familiar with their surroundings and hopefully will return to the site as adults and take up residency. Hacking requi res constant attendance and observation to protect the vulnerable young and assure they have sufficient food while they are dependent upon the hack site. While the hack sites wi 11 be �5 operated by the State, actual costs to operate the sites will be borne by the appropriate land administering agency (Forest Service, Bureau of Land Management, and National Park Service). 6. Confirmed breeding territories and selected potential breeding sites will be surveyed annually to document the presence of released falcons and ultimately determine the success of recovery efforts. RESULTS AND DISCUSSION Territory Occupancy Breeding territory occupancy increased from 44 sites in 1990 to 58 in 1991 (Table 1). The expansion was due to reoccupancy of two historical sites (sites 21 and 23), discovery of 10 previously unreported sites (sites 65 through 74) and reoccupancy of five sites (25, 33, 41, 51 and 54) that were vacant in 1990. Eight of the ten sites that were discovered in 1991 were situated on cliffs that had been surveyed and found to be vacant in the past. West Slope nesting pairs continued to expand (Table 2) with 42 sites occupied by 40 breeding pairs. East Slope site occupancy increased to 16 sites with reoccupancyof 2 historical sites (Sites 21 and 23) and addition of 5 new territories (Sites 65, 68, 69, 70, and 72). Reproduction Due to the population increase, no remedial efforts such as fostering or hacking were undertaken in 1991. Peregrine productivity in 1991 averaged 2.28 young fledged per successful pair (40 pairs fledged 91 young) and 1.72 young fledged per total pair (53 pairs fledged 91 young)(Table 3). The 42 occupied sites on the West Slope fledged 75 young (1.79 young per occupied territory) while the 16 East Slope sites produced 16 fledglings (1.00 young per occupied territory) (Table 2.). The lower productivity of the East Slope may be due in part to presence of inexperienced breeders. Although site 39 started the fourth year with an the adult male peregrine paired with a female prairie falcon, she was replaced by a first year female peregrine. They did not produce eggs. Eggshell Condition First clutch eggshell fragments representing 19 nests were collected in 1991. The shell thickness (with membrane) from that sample averaged 12.0% thin (0.316 mm). Since fragments were encountered during nest visits, their origin from each egg could not be ascertained. Since thickness varies as much as 5% from the egg waist to the poles, random fragments can be expected to deviate by that amount. More accurate thickness measurements from the waist of whole eggs encountered at 9 sites yielded 0.322 mm (10.3% thin) which is similar to the 1990 average. �6 Individual eggshell thicknesses continued to vary wildly. As an example, Site 18 produced eggshells that were normal thickness (0.353mm) while an egg from Site 31 was 22.8% thin (0.277mm). Organochlorine Residue in Eggs Results of chlorinated hydrocarbon scans of 12 nonviable whole eggs encountered at eight wild nests in 1990 have been received. All the residue levels were low. DOE values ranged from 2.4ppm to 7.5ppm (fresh wet weight) and total PCB's ranged from 0.54ppm to 7.5ppm. The 9 whole eggs that were collected in 1991 will be submitted for analysis. Prepared By:_ ____:Ce-=::;....:......:...;;?;...·__;:6=...;;..;;.CV-_,.;.~~.__ Gerald R. CraigJ Wildlife Researcher C �Table 1. OCCUPANCY OF PEREGRINE BREEDING TERRITORIES IN COLORADO SITE PRE NO.1964 1964 1965 - 1966 1967 -- 1968 1969 1970 P P P 1 P 2 P P + P A P 3 + P A A 4 + M P 5 6 7 8 V 9 + 11 A P P 12 A M 13 + 14 F P + 15 + 16 + 17 + 18 V + 19 + 20 V + V 21 + A 22 23 V + -------------------------~-------------------------------(Htstortcal 24 25 26 27 28 29 30 31 32 33 34 35 37 + 38 39 40 41 - -- - - -- - -- = -- - --- --- ----- -- -- - - --- - - - -- --- - --- --- -- -- - - - --- --- --- -- = - -- - ----- - - -- --- P P P P P -- V V V V -- V - 1971 1972 1973 1974 1975 1976 - P P P P p P p P P P A P p p p p V P P P P P p p p p V A V V V V V V V M V P P P V p P V V P V V V V V V V V V V V M V P P p1 V V V V p1 M P - ----- V - - -- -- - --- - -- --- - - - - - - - V V V V ---- -- = V V A F4 - --- V A --- - V V V V - - - - - ---- - --- -- -- -- - -- --- - P P V M V V V p1 V P p5 V V P V V V V V V V V V V V V V V V V V V stt8S p p p p --- - ---- --- - - - = 1977 1978 V V V V V V V V M V p1 P P P V V P P V V V V V V P P V V V V V V V V V V V V V V V V V V abov8 thts p V p V p3 P p p P M p P p p p1 --- M -- F Lone adult female. M Lone adult male. A Lone adult. P Adult patr. 2:Imm. male & ad. female. 3:Ad. female replaced by tmm. 1 Ad. male & imm. female. 6:Ad. male dead tn vtcintty. 5 Ad. male replaced by tmm. male mtdway. 7:Imm. male 9 Ad. male paired with female prairie falcon. 1979 1980 V V V V P p1 p2 V V V V p1 1981 1982 1983 1984 1985 V V V V P V V V V V p1 p1 V V P V V V V P p1 M V V V F V V V V M P F V V V V V V V P P P P P P P P P1 V V V V P F V P V V V V V V V V V V V V V V V p1 V P M V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V ltn8)--------------------------------------------------------V V V M M V V V P P P P P P M V V V V V V M8 P F P P P P V V V P V V V A V V V V V V p p p p p p p P M P P P P P M V V V V V V p p V V P V V p1 P P P P P P P P P P P - --- - - - -- - - - p6 p P P p p1 - - P p1 V P 1986 1987 1988 P V P P P V V P V P V V V V P V V V V V V V P V P P P V P V P P V V V V P V F V V V V V P P P p1 V P V P V V p 1989 1990 1991 P V P V P P P V V V P V P V V V V V P P P P P V P V P P V V V V P V P V V V V F P P P P P V P V P P p1 V P V P V P p V P V P V P p V P V P V P p V V V P V P P V V P P P V M P P p1 P p1 P P V P P P P P P P P P P P P p9 P P V P P P P p9 P P P P - P P P p1 V V V P V P V V V V V p P~ P V P P P P P V P V P P P V V' V P V P V V P V P V P V P V P P P P P P P P P p1 P F = V Vacant site. female midway. 4:0ead & female. 8:Imm. male. ad. female found in victnity. '"'-J �Table 1 (cont.). OCCUPANCY OF PEREGRINE BREEDING TERRITORIES IN COLORADO SITE PRE NO.1964 1964 1965 1966 19117 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 p 42 43 44 45 46 48 1985 1986 1987 1988 1989 1990 1991 P p p p p p p p p p P P p P P p p p1 P p2 V P M p p p p p p p P P p p p p p p p p p1 49 50 51 52 V V V P p9 p9 V V P p P 53 54 p F 55 F p p p 56 57 58 P V P V p P p p 59 60 61 P p p p 62 63 64 66 v 67 68 v v 69 p p V A P P M p p P A P p p p p 70 V V V V V V V V V V v V V V V V V V V V v V 72 73 74 p p P 65 71 p p p A = Lone adult. F = Lone adult female. M = Lone adult male. P = Adult pair. V = Vacant site. 1:Ad. male & imm. female. 2:Imm. male & ad. female. 3:Ad. female replaced by imm. female midway. 4:Dead ad. female found in vicinity. 5:Ad. male replaced by imm. male midway. 6:Ad. male dead in vicinity. 7:Imm. male & female. 8:Imm. male. 9:Ad. male paired with female prairie falcon. P M F P �9 Table 2. COMPARISON OF EAST AND WEST SLOPE SITE OCCUPANCY West SloQe Occupied Sites Breeding Pairs Young Produced Young Hacked 73 74 75 76 77 78 79 80 81 82 83 84 85 11 9 7 7 10 9 11 11 11 11 13 13 13 1 4 4 3 8 6 4 5 6 6 8 11 12 a 10 5 4 11 16 12 16 15 20 21 26 26 a a a a a a 5 7 10 8 15 18 12 East SloQe Occupied Sites Breeding Pairs Youn£, Produced Young Hacked 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 CUM 3 2 1 1 2 2 2 2 a a 1 1 1 2 5 8 7 9 16 3 2 1 1 1 a a a a a a 1 1 1 3 2 4 6 9 2 3 a 3 0 a a a a a a 3 2 2 10 5 6 11 16 63 a a a a a 4 5 4 3 5 8 9 12 10 19 26 25 12 a 110 86 19 17 32 4 87 24 20 45 88 24 22 44 89 32 27 60 90 35 33 51 a a a a 91 CUM 42 40 75 435 a 79 �10 Table 3. Site 1 2 3 4 5 7 9 11 12 16 18 21 23 25 27 29 30 31 32 33 34 35 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 Male A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A Age Female Sumary of 1991 Peregrine Production Eggs 1st 2nd Clutch Clutch Young Hatched Young Fostered Young Fledged A A A A A A A A A A A A A A A + 3+ 3+ 3+ 2+ 2+ 4+ 0 2+ 4 2+ 1+ 4 4 2 1 4 + 3+ 3+ 0 3+ 2+ 0 3 2 3 3+ 4 3 1 0 2 1 0 A A A A A A A 2+ + 2+ + 2+ 3+ 2+ 2 2+ 2+ 0 1 3+ 2+ 2 I I 0 2 2 2 1 2 0 0 2 2 A A A A A A A A A A A A A A A A A A A 2+ 2+ 2 4+ 2+ 3+ 4+ 2+ 3+ 3+ 3+ 2+ 3 2+ 3+ 2+ + + 3+ 3+ 4 2+ 3 4 2+ 3 3+ 3+ 2+ 3 2 3 2+ 2 2 3 4 2 2 3 3 2 1 1 3 2 1+ 3+ 3+ 1 2 3 A A A + 3+ 3 2 3+ 2 2 3 2 A A A A A A A 2+ 3+ 3+ + + 3+ 2+ 2+ 3+ 3+ 0 0 2+ 2+ 2 3 3 0 0 2 2 A 2+ 2+ 2 Total Sites Occupied: Total Breeding Pairs: Total Young Produced: Average Fledged Brood Total Fledglings + 58 49 100+ Size: 2.28' Total Total Total Young Adult Pairs: 51 Successful Pairs: 40 Young Fledged: 91 Fledged Per Total Pair:1.72 Divided by Total Successful Pairs �..l.J.. JOB PROGRESS REPORT State of P roject: __ ---"'C=o....:..lo;:::_r:....:a=d=o:..._ _ Bald Eagle Nest Site Protection and Enhance__l.(..!.!W_-..!.,;15~1!,_-....:..R~-=3.L.) _ ment Program Period Covered: 1 July, 1990 - 30 June, 1991 Personnel: G.R. Craig, Colorado Division of Wildlife, R.L. Knight, Colorado State University, and R. McLean, Center for Disease Control. ABSTRACT Breeding bald eagles occupied 13 Colorado nesting territories in 1989. Two new territories was located in each of Mesa and Weld counties. Nine pairs successfully fledged 19 young. Samples of feather pulp and blood sera to determine genetic similarity of the Colorado population to adjacent populations were collected from 13 nestlings representing 6 sites. This Job Progress Report represents a preliminary analysis and is subject to change. For this reason, information presented herein MAY NOT BE PUBLISHED OR QUOTED without permission of the author. ��13 BALD EAGLE NEST SITE PROTECTION AND ENHANCEMENT PROGRAM Gerald R. Craig SEGMENT OBJECTIVES 1. Monitor nest site occupancy and reproductive success. 2. Document survival rates and mortality factors. 3. Determine migration and wintering areas. 4. Determine if philopatry occurs in breeding eagles. 5. Determine nest site tenacity by individual breeding eagles. 6. Quantify nesting habitats and associated foraging areas in an effort to document nest site parameters conducive to improved reproduction. 7. Document pesticide contamination through eggshell measurement and chemical analysis of nonviable eggs. 8. Where necessary, occupancy. implement actions to stabilize nests and maintain METHODS AND MATERIALS This work will be a cooperative endeavor between the Division and Dr. Richard Knight of Colorado State University. 1. Annually visit all documented breeding sites to determine the presence of bald eagles. Pairs at territories will be documented by DWMs and other field personnel. Previously unrecorded pairs will probably be revealed in the course of aerial eagle and waterfowl flights. DWMs will confirm actual incubation from ground visits. 2. Occupied territories will be visited by DWMs periodically throughout the breeding season to determine hatch of young, nesting failures, etc. 3. In May and June, a Utility Worker will observe breeding eagles from a distance and endeavor to follow thei r movements to locate important foraging areas. Responses of eagles to various human activities and land uses will be recorded. 4. In June, when the young are determined to be old enough to band, sites will be visited by Craig and Knight to place a federal band on one leg and a colored, alpha numeric marker on the other. The color markers will permit identification if the young return in subsequent years. During the same nest visit the following will be recorded: Physical parameters such as tree species, height, DBH, condition, and dominance. Nest condition, size, and location. �14 Vegetative community and land use practices. In addition, collect prey remains, nonviable eggs and eggshell fragments. 5. Approximately 5cc's of blood will be collected from each blood will be analyzed at the Savannah River Ecology Lab Carolina. Electrophoretic examination will permit genetic samples collected from other populations in Saskatchewan, and Arizona, as well as determine the heterogeneity of the nestling. The in Aiken, South comparison with the Lake States Colorado birds. 6. When necessary, remedial actions will be taken to stabilize nests that are threatened by wind throw. Should the tree be decadent and in danger of falling, an artificial nest base may be placed in a suitable, adjacent tree. Action will be taken only after it has been deemed desirable to encourage the eagles to nest at the same location. RESULTS AND DISCUSSION Territory Occupancy Bald eagle nesting activities for Colorado are summarized in Table 1. In 1991, 13 territories (Adams, Archuleta, La Plata #1 and #3, Mesa, Moffat #1, #2 and #3, Montezuma #2 and #3, Rio Blanco #1 and #3, and Weld #3) were occupied. Two new territories (Mesa and Weld #3) were added in 1991. Although there is the slight possibility that the pair at Weld #3 may have relocated from Weld #1, it is unlikely, given the distance. According to the landowner, Weld #3 was occupied in the late 1950's and had 2 young, but failed when the adults were shot. Land Status The new territory at Weld #3 is on private land. The landowner is aware of the site and is protective of the pair. Land use in the area is livestock ranching. The pair at Moffat #2 relocated from private land across the river to a golf course. The pai r adjusted to golfers immediately below the nest and were successful. The golf course management and many of the patrons are aware and protective of their eagles. The new site at La Plata #3 is of Forest Service land. Its remote location insures minimal, if any disturbance. The Mesa site is also new. It is situated on an island in the river on land administered by the Colorado Division of Wildlife. This site has the potential for moderate to high disturbance. . Reproduction Reproductive efforts of the 13 pairs are summarized in Table 2. In all, 19 young were hatched and fledged by 9 pairs (2.11 young per successful pair) which yielded an overall productivity of 1.46 young per nesting attempt. Two broods of 3 young were encountered at Adams and Rio Blanco #3. Two sites (Archuleta and Mesa and Montezuma #2) were not visited because they failed to produce young. Montezuma #3 was not visited because the young had already fledged, and Moffat #1 was inaccessible because of mud. Because of �15 landowner hesitance, Rio Blanco #1 was not climbed, but 2 large young were observed from a distance. Thirteen young were banded at 6 locations (Adams, La Plata #1 & #3, Moffat #2, Rio Blanco #3, and Weld #3). Fish and Wildlife bands were affixed to the nestlings' right legs and red alpha-numeric bands with yellow flags were affixed to the ir 1eft 1egs. Blood samp 1es , cu 1men 1ength and foot pad 1ength measurements were obtained from the eaglets that were banded. Eggshell Condition No eggshell fragments or nonviable eggs were encountered in 1991. Banded Adults The adult female at Moffat #2 was banded with on the right leg with a rivet style USF&WS band. Most, but not all of the digits were read with a spotting scope. The prefix was obscured by a tar-like substance, but the last 5 digit were 3(?)4402. It was also possible that the adult male may have a olive green or brown band. Nest Stabilization Efforts In the course of the visit to the Adams site in 1989, one of the support wires to the artificial nest structure was found to be broken allowing the nest to list slightly. The site was revisited in the fall of 1989 and it was determined that additional nest material had caused the nest to drop approximately 6 inches where the support wire had broken. The structure was raised and stabilized by fastening one end of an 8 foot redwood 2X4 to the tree trunk below the nest and attaching the other end to the rim of the nest basket. In conjunction with two support limbs, the 2X4 formed and inverted tripod that should provide a secure nest support for many years. The eagles were present at the time of the nest modification and continued to occupy the nest in the 1991 breeding season. Climber access to the nests has been a problem for several years. The placement of nests in dead limbs at the top of decadent cottonwoods along with their bulk combine to make it almost impossible to climb into some nests. In past years, climbing efforts were aborted because of danger to the climber and the nest. In 1991, 1/4 inch cable loops were affixed to the most difficult nests to aid in gaining access on future visits. The cables were affixed in such a fashion that as additional nesting material is added, the loops will become incorporated into the nests and yet offer secure hand and footholds at critical points. Due to its cantilevered placement in the tree, the artificial nest structure at the Adams nest required a cable ladder that was affixed to an adjacent limb and was folded and stored on the underside of the nest. In all, 6 nests (Adams La Plata #3, Moffat #2 and #3, Rio Blanco #3 and Weld #3) were modified using this technique. Prepared by: G/(. (lfl<:u_~ G~raid R: Crai Wildlife Researcher C �•..... 0\ Table 1. Bald Eagle Nesting Efforts in Colorado 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 Site La Plata Co. #1 Moffat Co. #1 La Plata Co. #2 Grand Co. Montezuma Co. #1 Rio Blanco Co. #1 Rio Blanco Co. #3 Weld Co. #1 Montezuma Co. #2 Moffat Co. #2 Moffat Co. #3 Adams Co. Archuleta Co. Montezuma Co. #3 Weld Co. #2 La Plata Co. #3 Weld Co. #3 Mesa Co. 1egg IA IA IA IA 1yng 2yng 2yng 2yng -- 2yng 2yng 2yng -- -- -- Oyng ----- -- --.- ----- ------- --- -------- -- -- o Total Young Total Pairs Young/Occ. Terr. IA = Inactive ---- ----------- ----- ----- --- -- ------- -- ---- --- ------ ---- -- IA 1yng Oyng Oyng ---- ---- -- --- ---- --- ? ? ? ? -IA IA A 2yng IA IA A 1yng 2yng IA IA A 1yng 3yng 2yng IA IA IA IA A IA ? eggs 2yng 2yng -- 2yng -- -- ---- --- --- ----- ----- ----- ---- -- ---- ---- -- ---- -- -- -- ------ --- -- -- ? -- --- ---- -- ---- ? Oyng IA IA IA Oyng 2yng 2yng 2yng ---- -- -- ---- -- eggs 1yng IA IA IA 2yng Oyng eggs 1yng 1yng 1egg eggs eggs -- --- --- 2yng 2yng A IA IA 2yng 1yng IA 1yng Oyng IA 1egg 2yng --- ---- 1yng 1yng IA IA IA 2yng A IA 1yng 2yng 2yng 3yng IA IA IA 2yng 2yng IA 1yng 3yng eggs ? eggs 2yng IA IA 1yng 1yng -- eggs --- -- -- --- Oyng 2yng IA IA IA 2yng 1yng Oyng 1yng 2yng IA 2yng IA 1yng IA 2yng --- 1yng 2yng IA IA IA 2yng 3yng IA Oyng 2yng Oyng 3yng Oyng 2yng IA 2yng 2yng Oyng 1 4 4 4 1 0 3 6 2 6 6 5 10 8 16 13 19 1 1 2 2 3 3 1 3 4 2 4 5 10 9 8 10 10 13 0.00 1.00 2.00 2.00 1.33 0.33 0.00 1.00 1.50 1.00 1.50 1.20 0.50 1.11 1.00 1.60 1.30 1.46 A = Active �17 Table 2. Colorado Bald Eagle Nesting Efforts - 1991 Site Age of Birds Male Female Young Produced Young Fledged Comments Adams Co. Adult Adult 3 3 Archuleta Co. Adult Adult o o La Plata Co. #1 Adult Adult La Plata Co. #3 Adult Adult 2 2 Mesa County Adult Adult o o Moffat Co. #1 Adult Adult 2 2 Moffat Co. #2 Adult Adult 2 2 Moffat Co. #3 Adult Adult o o Montezuma Co. #2 Adult Adult o o Montezuma Co. #3 Adult Adult 2 2 Adults incubating, no eggshells found. Adults incubating, site not visited. Site not visited. Rio Blanco Co. #1 Adult Adult 2 2 Nest not climbed. Rio Blanco Co. #3 Adult Adult 3 3 Weld Co. #3 Adult 2 2 19 19 Total Adult 13 13 ~ disappeared part way through incubation. Eggs failed to hatch. not visited. Site not visited. Site ��JOB PROGRESS REPORT State of ~C~o~l~o~r=a=d=o _ Project: (W-156-R-2) Period Covered: 1 July, 1990 - 30 June, 1991 : Pawnee Grasslands Raptors Personnel: G.R. Craig, Colorado Division of Wildlife and Douglas G. Leslie, Colorado State University Data gathered for the latter half of 1989-90 field season was reported in the previous progress report. This segment was devoted to nest site analysis, digitizing spatial data, coursework, and comprehensive exams. Manuscript preparation and publication has been delayed until 1992. � https://cpw.cvlcollections.org/files/original/73dd7c9ecfbdd40e17d87d2450de8bad.pdf a549edc8a699b278a22534f9280f9b29 PDF Text Text 1 INTERIM FINAL REPORT Colorado State of: Proj ect: 3 Work Plan: Job Title: Period Author: Upland W-152-R 13a Sage Grouse Survival Covered: Marilet Personnel: : Job 01 January Bird Research Estimation, North Park, Colorado through 31 December 1990 A, Zablan C, E. Braun, Colorado Division of Wildlife; Zablan, Colorado State University G. C, White and M, A. ABSTRACT Sage grouse (Centrocercus urophasianus) band recovery data from a l6-year period in North Park, Colorado were summarized by age and sex. Computer programs BROWNIE and ESTIMATE were used to test for differences in survival rates between males and females and to generate annual survival rate estimates by sex and age class (subadult and adult). The null hypothesis that adult males and females had the same recovery and survival rates was rejected (f = 0.013). Survival rates for males differed by year (f = 0.007) and goodness-of-fit and likelihood ratio tests indicated that Model HI fit the male recovery data. Model HI assumes yearspecific and age-specific annual survival and recovery rates. Program ESTIMATE was used to demonstrate that the simplest model (Model 3), with constant survival and recovery rates across years, fit the female recovery data (f = 0,336), Survival rate estimates for males banded and released as adults ranged from 0% (0 - 19.9%) in 1985 to 84,6% (33.8 - 100%) in 1979 and averaged 38.4% (32,9 43,8%). Survival rate estimates for males banded and released as subadults ranged from 15,0% (0 - 37.5%) in 1986 to 100% (27.3 - 100%) in 1974 and averaged 51. 7% (40'.0 - 63.3%). Survival rate estimate for both age classes of females over years (not an average value) was 54.7% (50,1 - 59.2%). Program SURVIV was used to test hypotheses that weather variables had no effect on annual survival of male sage grouse, as annual differences in survival were not found for females. Annual weather differences in sprLng and winter precipitation and temperature were" incorporated in a logistic model for sur-vf.va l, estimation and were assumed to be responsible for yearly variation in recovery and survival rates. Incorporation of four weather covariates did not significantly improve model fit (f = 0.194 to 0.787). Recommended banding sample sizes for future studies were calculated with Program BAND2. These results and those from a study of grouped recovery effects on survival estimation will be included in a thesis to be completed by 01 December 1991 in partial fulfillment of requirements for the Master of Science degree from Colorado State University, Prepared by: e)f/gAl&h.xf~. Marilet A. Zablan Graduate Research Asst. Approved by: t/atY2.~ C'LaLt; E, Braun Wildlife Research Leader ��3 INTERIM REPORT Colorado State of: Project: W-152-R Work Plan: _3_: Job Title: Upland Bird Research Job ~ Responses Period Covered: Author: FINAL of Sage Grouse to Vegetation 01 January through 31 December Fertilization 1990 Orrin B. Myers Personnel: C. E. Braun, Colorado Division of Wildlife; White, Colorado State University O. B. Myers, G. C. ABSTRACT A dissertation is being prepared for partial fulfillment of a Ph.D. degree in wildlife biology, Department of Fishery and Wildlife Biology, Colorado State University. Completion of all degree requirements should be accomplished no later than December 1991. Four chapters will discuss directly the responses of sage grouse (Centrocercus urophasianus) to vegetation fertilization or will develop methodology used in evaluating sag~ grouse responses. Chapter 1 . describes field studies on the response of sagebrush (Artemisia tridentata) and sage grouse to nitrogen fertilization in North Park, Jackson County, Colorado. Sagebrush plants responded to application of 112 kg-Nfha in Fall 1985 with increased growth and increased levels of foliar crude protein. Foliar crude protein levels remained significantly elevated over controls in samples collected in Spring 1989. A second application of fertilizer in 1987 produced a less dramatic response by sagebrush. Overall, soil and foliar nitrogen decreased exponentially during the study, but gross morphological differences were evident for 3 seasons after treatment. Sage grouse used fertilized study plots for feeding significantly more often than adjacent unfertilized plots. The magnitude of the effect declined gradually for 3 years and then returned to control levels in the 4th season after treatment. Sage grouse use of A. 1. vaseyana plants was not affected by fertilization. I used captive-reared sage grouse to test whether sagebrush subspecies-specific feeding preferences of grouse were due to environmental learning and the relative availability of each subspecies. In chapter 2 I demonstrate the utility of Bradley-Terry models for estimating preference from pairedcomparison experiments, and show how the model can be reparameterized to account for factorial treatment structures. In Chapter 3 I use Bradley-Terry models and analysis of variance on balanced incomplete blocks to test whether birds raised in captivity and having equal access to each sagebrush treatment show any preference for or avoidance of sagebrush treatments. Initial forage selection was a random process for all but fertilized A. 1. vaseyana, which was avoided, but I detected preference for fertilized and unfertilized A. 1. wyomingensis over A. 1. vaseyana when cumulative intakes over 60 minutes were compared. Fertilization increased the intake of both subspecies, but �4 consumption of fertilized A. !. vaseyana was far below that of unfertilized A. ~. wyomingensis. In the final chapter I summarize results from single treatment digestibility trials. Wild-caught sage grouse performed poorly on all treatments, but birds were better able to maintain body mass on A. ~. wyomingensis treatments than on A. !. vasevana treatments. Analyses to evaluate the effects of fertilization on food quality are ongoing. Prepared by Approved by Clait E. Braun Wildlife Research Leader �5 INTERIM FINAL REPORT State of: Colorado Project: __ W~-~1~5~2_-~R~ _ Work Plan: 3__ : Job Job Title: Response of Selec.ted Avifauna Sagebrush Type Period Covered: Author: Upland Bird Research 17 01 January to Prescribed through 31 December Burning in the Big 1990 Lee A. Benson Personnel: Lee A. Benson and Wayne C. Leininger, Department of Range Science, Colorado State University; Clait E. Braun, Colorado Division of Wildlife ABSTRACT Sage grouse (Centrocercus urophasianus) numbers and spring habitat use were monitored at 3 sites b~rned in 1987 in Jackson and Moffat counties, Colorado. Burns ranged in size from 38 to 1200 ha and were within 1 kID of sage grouse leks. Two burns were wildfires while 1 site was a prescribed burn. The prescribed fire site burned in a mosaic pattern and resulted in 50% reduction in big sagebrush (Artemisia tridentata) canopy cover within a 38-ha area. The wildfires consumed nearly all sagebrush within the 120 and l200-ha sites. There was no sage grouse use of the smaller wildfire site. Sage grouse use of the 2 other sites was primarily restricted to areas with remnant patches of sagebrush. Sage grouse use of burns increased each year post fire, but numbers of male sage grouse counted on leks decreased in either the first or second year following fires at all sites. Numbers of males counted on the prescribed burn site and one of the wildfire sites increased in 1990 and returned to or exceeded pre-burn lek numbers. However, fluctuation in sage grouse numbers on leks post-fire did not differ from preburn levels. Prepared by aeQJ~J Lee A. Benson Graduate Research Assistant Approved by Clait E. Braun Wildlife Research Leader ��JOB PROGRESS REPORT Colorado State of: Project: W-152-R Work Plan: Job Title: ·8 Upland Bird Research Job _5_ Population Inventory and Habitat Use by Lesser Prairie-chickens in Southeast Colorado Period Covered: 01 July 1985 through 30 June 1991 Author: M. Giesen Kenneth Personnel: C. E. Braun, D. Clarkson, M. B. Dillon, S. J. Fellinger, K. M. Giesen, R. W. Hoffman, G. M. Kittel, T. B. Lundt, D. M. Picken, J. S. Slater, C. H. Wagner, and B. D. Will, Colorado Division of Wildlife ABSTRACT Lek density and. spacing, male breeding density, movements and home ranges, habitat characteristics of occupied rangeland, lek sites, nest sites, and flush sites of lesser .prairie-chickens (Tympanuchus pallidicinctus) were investigated in Baca County in southeastern Colorado from April to September 1986 through 1990. The number of active leks on the 4l.4-km2 study area ranged from 6 to 9 annually, with a mean distance between nearest leks varying annually from 1.24 to 1.44 km. There was a strong positive correlation between male breeding density and number of leks (£ - 0.92) but npt between male breeding density and average lek size (£ - 0.15). Average lek size ranged from 6.5 males/lek in 1990 to 10.2 males/lek in 1989. Helicopter surveys during the peak of hen attendance detected 13 of 15 active leks (86.7%) but only 7 of 15 leks (46.7%) 2 weeks after the peak of hen attendance. Roadside listening transects detected 20 to 100% of active leks within 1.6 km and varied with weather conditions, observer, and date. Home ranges (minimum convex polygon) of females (5.96 ± 8.18 km2) were larger than those of males (2 ..11 ± 1. 54 'km2). Combined home ranges of all birds from individual leks was estimated at 34.67 12.78 km2 (95% ellipse) to 46.30 ± 2 12.69 km (minimum convex polygon). Movements of hens from lek of capture to nest sites averaged 1.80 ± 1.04 km and ranged from 0.22 to 4.85 km. Habitat characteristics of the study area indicated bare ground comprised 75.7% of the area with grasses averaging 19.9% ground cover. Sand sagebrush (Artemisia filifolia) density averaged 2497 plantsjha and had 3.1% canopy cover. Residual vegetation in April and May had an average height-density of 0.98 dm. Lek sites were dominated by buffalograss (Buchloe dactyloides) and blue grama (Bouteloua gracilis). Most nests were adjacent to sand sagebrush or soapweed (Yucca glauca). Heights of shrubs, forbs, and grasses were taller at nest sites than at adjacent areas. Vegetative height above nest bowls averaged 50.7 cm. Vegetative parameters measured at grouse flush sites were similar to those measured at dependent sites for both males and females indicating little apparent selection for vegetation structure at summer feeding-loafing sites. ± �8 RECOMMENDATIONS 1. All known active leks should be visited during March - May each spring to document activity status (active, inactive, abandoned). Searches of known occupied habitats should be conducted annually from the ground or air to locate additional leks or leks that moved. 2. When possible, aerial searches of historic range within Colorado should be made in mid-April to document occupied rangeland and locate leks. 3. Population trend analysis should use lek density or total numbers of active leks rather than average lek size (number of males/lek). 4. Management of occupied rangelands should include maintenance of sand sagebrush and forbs, and bunchgrass height-density above 1.0 dm with adequate bunchgrass between 3.0 and 5.0 dm for nesting. 5. Suitability of rangeland for lesser prairie-chickens should be ascertained on the basis of sand sagebrush density and canopy cover, and canopy cover and residual height-density of native bunchgrasses in spring. �9 POPULATION INVENTORY AND HABITAT USE BY LESSER PRAIRIE-CHICKENS IN SOUTHEAST COLORADO Kenneth M. Giesen INTRODUCTION Both distribution and populations of lesser prairie-chickens in North America have decreased by > 90% from historic levels of the 1800's (Taylor and Guthery 1980). The exact historic distribution of lesser prairie-chickens is unknown, although early reports suggested they were abundant and widely distributed throughout their range (Bendire 1892, Judd 1905, Bent 1932, Baker 1953, Sands 1978). Aldrich (1963) indicated lesser prairie-chickens historically inhabited about 360,000 km2 in 5 states. Recent estimates suggest the current range of lesser prairie-chickens is restricted to 125,000 km2 (Taylor and Guthery 1980, Johnsgard 1973). Historic population estimates were crude but lesser prairie-chickens were reported to be common to abundant throughout their range at the beginning of the century (Bent 1932, Baker 1953, Bailey and Niedrach 1965, Oberho1ser 1974). Litton (1978) reported estimates of 2 million prairie-chickens in Texas prior to 1900. Density estimates for different habitats ranged from 1.0 to 14.6 males/km2 (Davison 1940, Jones 1963, Copelin 1963). If the historic range was 360,000 km2, then historic populations may have been in the millions of birds. Current North American breeding populations are estimated at 50,000 birds (Crawford 1980, Taylor and Guthery 1980). Historical evidence suggests that 'while lesser prairie-chickens were peripheral in Colorado, they were common to abundant in 6 southeastern counties (Baca, Prowers, Bent, Kiowa, Lincoln, and Cheyenne) and peripheral in adjacent counties (Loeffler 1983). Surveys for lesser prairie-chickens in Colorado since 1960 have documented breeding populations in Baca, Prowers, and Kiowa counties with an estimated population of < 700 birds (Hoffman 1963, Loeffler 1983, Rash 1985). The lesser prairie-chicken is currently classified as a threatened species in Colorado. The objective of the Colorado Division of Wildlife is to improve the status of lesser prairie-chickens through habitat manipulations, reintroductions or transplants, and other management approaches (Colo. Div. Wi1d1. 1983). However, basic information on habitat characteristics of rangelands currently occupied by lesser prairie-chickens in Colorado is lacking. Methods for documenting annual population changes or population responses to management programs also need to be evaluated. P. N. OBJECTIVES 1. Ascertain the relationship between the average number of males attending leks (lek size), the total number of active leks (lek density), and the number of lek-attending males (male breeding density) on the study area. �10 2. Measure the accuracy and pre~1s1on in measuring lek densities. of aerial and ground quadrat surveys 3. Document seasonal movement patterns of lesser prairie-chickens in relation to lek sites including lek-to-nest dispersal of hens and home range size of males and females. 4. Ascertain nest success of radio-marked hens in relation to vegetative cover at nest sites, and document timing and causes of nest failure. 5. Measure vegetative characteristics of lesser prairie-chicken occupied rangeland, leks sites, nest sites, brood-rearing sites, and feedingloafing sites. STUDY AREA Approximately 41.4 km2 (16 mi2) in and adjacent to pasture 1AE of the Comanche National Grasslands in eastern Baca County was selected as the primary study area for documenting habitat use, movements, and home range. This area was also used to examine the relationship between lek density, lek size, and male breeding density. Included were sections 21-28, and 33-36, T34S, R44W, and sections 1-4, T35S, R44W. Movements of radio-marked birds resulted in some vegetation sampling in adjacent pastures. Aerial surveys for leks were conducted on three areas in Baca County southwest of Campo (48 km2), southeast of Campo (46 km2), and east of Campo (40 km2). The latter area included most of the primary study area. Four roadside transects (17.6 - 22.5 km in length) to locate leks were in southeastern Baca County and adjacent to the primary study area and quadrats used in aerial surveys. The primary study area was comprised of approximately 90% rangeland and 10% dryland agriculture (fallow or grain sorghum). The U. S. Forest Service., Comanche National Grasslands managed the majority (72%) of the area with the remainder being privately owned (22%) or controlled by the State Board of Land Commissioners (6%). Livestock grazing was the predominate range use with cattle being present from May through November of each year. The habitat was characterized as sand sagebrush rangeland with mixed bunchgrasses dominated by sand dropseed (Sporobo1us cryptandrus), red threeawn (Aristida longiseta), and sideoats grama (Bouteloua curtipendula). Soapweed and broom snakeweed (Gutierrezia sarothrae) were common in some pastures. Dominant forbs included western ragweed (Ambrosia psilostachya) and Russian thistle (Salsola iberica). Plant nomenclature follows Harrington (1964). METHODS Lek Counts and Surveys Historic lek locations in Baca County were checked within 2 hours of sunrise from mid-February through May each year from 1986 to 1990 to document activity and obtain counts of males and females. Leks were defined as displaying aggregations of ~ 2 males (Martin and Knopf 1981), although lek sites with a history of activity were counted when only 1 male was regularly observed �11 displaying on the traditional arena. Leks on the study area were surveyed at least weekly in April and May and searches for additional leks conducted opportunistically·each year by listening from all accessible roads and trails in the area. Lesser prairie-chickens attending leks were classified to sex by display behavior and plumage characteristics. In a few instances when sex of birds could not be ascertained it was estimated that 90% of the birds on leks were males. The peak of hen attendance on leks was estimated from lek counts and trapping and summarized by 10-day intervals. The peak number of males at each lek within the study area was used to estimate male breeding density, as no radio-marked or leg-banded males were known to attend more than one lek. Lek density was calculated as the number of active leks within the 4l.4-km2 study area. Mean lek size was calculated as the peak number of males attending leks within the study area divided by the number of active leks. Information on numbers of leks and males on leks from· 1980 to 1985 was obtained from Colorado Division of Wildlife files and included for calculation of correlations between male breeding density and number of leks, and between male breeding density and average lek size. Lek surveys on the study area in these years were judged sufficient to locate all leks and count males on leks. Aerial surveys of the 3 quadrats were conducted using a Bell Soloy helicopter at 50-100 m elevation and flying parallel north-south transects separated by approximately 400 m (1/4 mi). Orange and white flagging was placed on fences (where available) to mark transects. Topographic features and a map also assisted the pilot and observer in· following transects. Flight speed was approximately 80 km/hr. The observer was naive to lek locations, at least during initial flights. One quadrat was surveyed each day between 0530 and 0730 on 14-16 April and again during 28-30 April 1987. Active leks were located prior to the initial survey by intensive searches on foot using a trained pointing dog and by listening for displaying birds with a parabolic microphone. Roadside transects were surveyed by personnel naive to lek locations on 14-15 April and 30 April-l May in 1987 and on 19-20 April 1988. Transects were initiated 30 minutes prior to sunrise and consisted of 3-minute stops every 0.8 km (1/2 mi) to listen for displaying lesser prairie-chickens. If birds were heard the compass direction to the source was drawn on a topographic map to estimate lek locations. In 1988 the observers retraced their routes after completion and attempted to visually locate lekking birds heard during the transect. The transects on 14-15 April 1987 and on 19-20 April 1988 were chosen to coincide with the peak of hen attendance on leks when display activity was most vigorous. Trapping. Banding. and Radio-tracking Lesser prairie-chickens were trapped using walk-in funnel traps connected with chicken-wire leads placed across leks (Toepfer et al. 1988) and with cannon nets set up near dominant males on leks. Usually 4-6 traps were placed on leks in 2 arrays surrounding activity centers and designed to intercept females moving across leks. The chicken-wire leads varied in length from < 4 m to > 10 m and connected adjacent traps. Traps were moved regularly in response to movements of males and females. The number of leks trapped varied with personnel available but was usually 4-6 each week during hen attendance. All but 2 active leks in the study area were trapped during one or more years of the study (1 lek was in a pasture used by cattle each spring, another lek �12 was active only during a short period in 1 year). Cannon nets were used occasionally on some leks in an attempt to capture more hens but the effort in setting up and moving nets in response to hen locations proved excessive. Persistent winds and rainfall were also detrimental to efforts in using the cannon nets. Captured birds were banded with a numbered aluminum band and a unique combination of 3 colored plastic bandettes. Age (adult or yearling) was ascertained from examination of distal primaries (Ammann 1944). Selected birds (2-3 males from each lek annually, most captured hens) were fitted with solar or battery-powered transmitters attached to ponchos (Amstrup 1980) (1986-90) or necklaces (1990). Solar transmitters weighed 22-26 gms and battery-powered transmitters weighed 13-20 gms (transmitters were 2.0-3.5% of the birds weight). Radio tracking was conducted on foot using Telonics TR-2 or Advanced Telemetry Systems (ATS) portable receivers and hand-held 3-element yagi antennas. Attempts were made to visually locate each bird weekly but this was not always attained during April and May when lek counts and trapping consumed most available time. Radio tracking usually lasted into AugustSeptember depending on survival of birds and battery life of transmitters. All locations were plotted on U. S. Geological Survey topographic maps (scale 1:24,000) and later recorded to the nearest 100 m as UTM coordinates. Home range sizes were calculated using the McPAAL software package (Stuwe and Blohowiak 1985) with 3 estimators; the minimum convex polygon (Mohr 1947), Koeppl et al.'s (1975) 95% ellipse, and Dixon and Chapman's (1980) harmonic mean. With the harmonic mean, a 10 x 10 grid was used with a 90% estimator. Individual·home ranges were calculated only for birds located a minimum of 10 times. Combined home ranges for all birds associated with individual leks were estimated using the minimum convex polygon and 95% ellipse, and calculated for 4 leks having at least 5 males and 5 (emales radio-marked. The combined home range for all birds associated with these leks was converted to a circular plot and a radius calculated for estimating the area around leks used by these birds. Vegetation Sampling Vegetation characteristics were measured at 10 randomly-selected transects in each cadastral section of the primary study area in April and May 1986-87 (8 sections/year) to ascertain vegetative composition and to measure residual cover. Along each 10-m transect, line-intercept of canopy cover (Canfield 1941) was recorded for all shrubs, grasses, forbs. and bare ground. Height of the nearest shrub, forb, tall grass, and short grass was measured every 2 m (6/transect). Sand sagebrush density was measured using 0.001 circular plots at the beginning, middle, and endpoint of each 10-m transect. Height-density readings (Robel et al. 1970) were obtained at 2-m intervals along the transect (6/transect). Vegetation was sampled at grouse use sites and nest sites by centering a lO-m transect along a north-south axis on the flush site (or nest site) and measuring vegetative parameters. For each grouse flush site a paired vegetation sample site was selected by randomly walking either north or south for a random distance of 1 to 400 m, and then randomly walking east or west for a random distance of 1 to 400 m. These paired (dependent) random sites were an attempt to measure vegetation characteristics within the flush radius �13· (home range) of lesser prairie-chickens. Selection for habitat variables was ascertained using the Wilcoxon signed rank test procedure for paired samples. Vegetative composition at lekking sites (arenas) was visually estimated to the nearest 5% and recorded in June after completion of male display and after vegetation was growing. RESULTS Lek Counts and Surveys Forty-two lesser prairie-chicken leks in southeastern Baca County were active for one or more years during 1986-90 (Table 1.) The number of leks counted each year ranged from 22 to 30. Thirteen leks were initially located during the study. Fourteen of 29 leks active in 1986 were occupied during aIlS years, and 8 leks were known to become inactive for 1 or more years. The proportion of males counted was 82.8% of the total high counts for all years combined and ranged from 79.9 to 86.1%. Female lek attendance patterns (Fig. 1) derived from 375 hens (range 49 - 114 annually) observed during daily lek surveys and trapping revealed a peak of attendance in early April in 1986 and in mid-April for 1987-90. I analyzed lek survey and lek count data collected from 1980 to 1990 on the 4l.4-km2 study area to ascertain relationships between male breeding density, average lek size, and lek density (Table 2). There was a strong correlation between male density and number of leks (~- 0.92, f - 0.0001) but not between male breeding density and average lek size (~- 0.15, f - 0.66). Correlations· using data from 1986 to 1990 showed similar patterns but were less significant (~- 0.85, f - 0.07; and ~ - 0.74, f - 0.15, respectively). Although lek density on the study area varied annually, there was little change in mean distance between neighboring leks. The mean distance to the nearest lek ranged from 1.24 to 1.44 km annually and from 0.96 to 2.28 km for individual leks. Aerial Quadrat Surveys Intensive ground searches resulted in detection of 15 lesser prairie-chicken leks on the 3 quadrats selected for aerial surveys (3, 3, and 9 leks, respectively). Numbers of males/lek ranged from 2 to 18 and averaged 6.7. During the initial helicopter flight during the peak of hen attendance 13 leks (86.7%) were detected. All leks were detected on the first 2 quadrats and 7 of 9 on the third quadrat (Table 3). When survey flights were repeated 2 weeks later, only 7 of 15 leks (46.7%) were detected with 3r 0, and 4 leks being detected on the 3 quadrats, respectively. Only 1 lek missed during the initial survey was detected during the second flight, but 7 leks detected initially were not observed during the second flight. �14 Numbers of lesser prairie-chickens counted, Baca County, Colorado, Table 1. 1986-90. Lek 2 3 4 5 6 7 12 14 15 17 18 23 25 27 28 30 31 33 35 36 37 38 39 40 41 42 43 44 ·46 86-1 86-2 86-3 87-1 87-2 87-3 87-4 87-5 88-1 88-2 88-3 90-1 90-2 1986 Males Totala 16 2 1 17 6 9 5 10 NC 4 10 2 2 0 12 1 2 7 10 1 12 2 11 8 3 10 5 0 1 15 6 7 19 3 1 19 8 10 5 10 NC 8 14 2 4 0 14 1 8 7 13 1 14 2 11 8 3 10 5 0 1 19 8 8 -- 236 . Totals 197 8.1 Avg./lek 6.8 1987 Males Total 9 3 3 18 3 8 4 8 NC 7 13 0 6 0 17 NC 9 8 18 2 18 7 6 9 0 10 NC 5 0 11 5 8 4 7 4 7 5 10 6 6 19 5 8 5 9 NC 9 15 0 6 0 24 NC 12 9 21 2 18 7 6 9 0 10 NC 8 0 14 6 8 4 7 4 9 5 242 281 8.1 9.4 1988 Males Total 12 11 6 23 6 12 2 6 NC 11 11 NC 5 7 20 NC NC 11 16 NC NC 0 7 -6 NC NC NC NC NC 12 0 6 3 9 NC 9 4 3 5 8 15 13 9 25 6· 16 4 13 NC 14 13 NC 8 7 25 NC NC 15+ 16 NC NC 0 14 6 NC NC NC NC NC 15 0 6 5 9 NC 14 4 3 6 8 231 289 8.9 11.1 1989 Males Total 13 9 0 21 7 16 3 5 17 11 11 0 6 3 15 NC NC 10 22 NC 11 0 0 13 NC 6 NC NC NC 10 0 11 0 9 NC 9 3 0 12 2 12 10 0 21 9 20 3 8 21 13 17 0 7 4 19 NC NC 14 25 NC 15 0 0 16 NC 9 NC NC NC 12 0 11 0 to NC 10 3 0 13 9 255 311 10.2 12.4 aIncludes males, females, and birds not classified to sex. 1990 Males Total 9 9 0 15 3 7 2 5 NC 4 5 NC 6 10 NC NC 2 NC NC NC 9 0 8 NC NC NC NC NC 7 0 0 0 8 NC 6 0 NC 6 8 5 5 11 12 0 17 4 9 2 9 NC 4 6 NC 7 5 13 NC NC 2 NC NC NC 9 0 10 NC NC NC NC NC 8 0 0 0 8 NC 6 0 NC 6 11 7 5 143 6.5 171 7.8 4 �15 100 C 1986 90 W > a: W en D 1987 80 ~ Dl 0 en 70 ~ _. W « ~ 60 1988 1989 1990 ~ W u, _. 50 0 40 u, 0 30 « I- l- IZ W 20 a: a. 10 0 UJ o MAR 1-10 APR 11-20 APR 21-30 APR MAY DATE Fig. 1. Timing of attendance of female lesser prairie-chickens in southeastern Colorado, 1986-90. on leks �16 Table 2. Lek count data for lesser prairie-chickens on the 4l.4-km2 study area, Baca County, Colorado, 1980-90. n n ~ Year Leks Males Ma1es/1ek 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 6 6 6 6 4 5 8 8 9 7 6 59 55 59 65 46 46 76 74 100 82 54 9.8 9.2 9.8 10.8 11.5 9.2 9.5 9.2 11.1 11.7 9.0 Ma1es/km2 1.42 1.33 1.42 1.57 1.11 1.11 1.84 1.79 2.42 1.98 1.30 Leks/km2 0.14 0.14 0.14 0.14 0.10 0.12 0.19, 0.19 0.22 0.17 0.14 Table 3. Lesser prairie-chicken leks detected during experimental helicopter surveys, Baca County, Colorado, 1987. Quadrat Quadrat size (km2) 47.9 46.0 40.1 1 2 '3 n Leks" 3 3 9 n Leks detected Flight Ib Flight 2c 3 3 7 3 0 4 aNumber of active leks within quadrat boundaries. b14-16 April. c28-30 April. Roadside Transects Roadside transects were evaluated along 4 transects by 2 observers in 1987 and along 2 transects by 2 observers in 1988. Overall, 59.3% of known active leks within 1.6 km (1 mi) of transects were detected (Table 4). In 1987 the observers detected 20.0 - 66.7%, with a higher proportion of leks being detected during the peak of hen attendance than 2 weeks later (62.5% vs 53.8%) under similar conditions. Observer bias was also documented with one obserVer detecting 63.2% of known leks and the 2nd observer detecting 46.7% of the known leks under similar weather conditions. In 1988, when all transects were covered during the peak of hen attendance, lek detection ranged from 76.9 to 100%. �17 Table 4. Leks detected during experimental County, Colorado, 1987-88. roadside transects in Baca n Leks detected Observer Surve:;l Transect Year Length(km) 1 2 1987 1987 1987 1987 1988 1988 -17.6 22.5 19.2 19.2 19.2 19.2 3 4 5 6 n 3 7 11 5 1 2 1 2 2 1b 1b 1 3 2 1 1 1 7 6 4 10 4 6 4 3 3 12 4 13 4 1.6 km of transects. aNumber of known active leks within "winds > 30 km/hr. Trapping Leks'' and Banding One hundred and forty-nine lesser prairie-chickens (75 males, 74 -females) from 10 leks on or adjacent to the study area were trapped and banded from 1986 to 1990 (Table 5). Most (n - 126, 84.6%) were trapped from 4 leks on the study area which were occupied aIlS years. In addition there was 1 trap mortality (heat stress), and 11 recaptures (all males) of birds trapped earlier within the year (n - 7) or from previous years (n - 4). Locations of male and female lesser prairie-chickens Table 5. Baca County, Colorado, 1986-90. trapped in F Totals M F Year 1987 1986 1988 Lek M F M F 2 3 4 5 18 28 8 5 3 1 6 6 2 4 2 4 Totals 3 1 2 3 4 3 7 1 2 6 4 4 3 5 1 2 4 4 1 4 5 5 M 13 17 M 14 3 3 3 3 12 3 9 19 19 3 24 2 14 14 4 1 4 19 1990 F F 33 86-1 86-2 87-4 1989 M 2 14 15 6 18 . 75 6 4 2 32 13 1 9 1 6 74 �18 Movements and Home Range Lek to nest distances were calculated for lek-of-capture and nearest lek-tonest for all radio-marked hens for which nests were located (n - 31). The mean distance from lek-of-capture to nest site was 1.80 ± 1.04 km (range 0.22 - 4.85 km) and farther (f - 0.0009) than the mean distance from the nest site to the nearest lek (1.04 ± 0.60 km, range 0.22 - 2.50 km). Radio transmitters were placed on 108 lesser prairie-chickens (37 males, 71 females) from 1986 to 1990 to facilitate location for estimation of movements and home range. Transmitter loss or failure, dispersal from the area, and high mortality from depredation resulted in loss of data from 75 birds within 2 months. The remaining 33 birds (19 males, 14 hens) provided an adequate sample of relocations for estimates of summer home range (Table 6). Because of unequal variances, non-parametric Wilcoxon tests were used to compare home range size between sexes. Home range size estimated using the minimum convex polygon did not appear to be affected by sample size for males (~ - 0.13, f 0.60) or females (~- -0.14, f - 0.63). Average home range'size of females (5.96 ± 8.18 km2) was larger (f - 0.03) than average home range size of males (2.11 ± 1.54). Home range estimates using the 95% ellipse also suggested females had larger (f - 0.009) home ranges than males (28.52 vs. 8.53 km2). The Harmonic Mean estimate of home range did not detect a significant difference between sexes (f - 0.07) although the trend.was for females to have larger home ranges (4.12 vs 2.30 km2). . . A combined home range estimate for all prairie-chickens associated w~th individual leks was calculated using the minimum convex polygon and the 95% ellipse. This "lek home range" estimated the minimum habitat used by a population of lesser prairie-chickens associated with individual leks. Four of 10 leks from which I trapped lesser prairie-chickens had a sufficient sample of home ranges from both males and females (Table 7). The average home range size (minimum convex polygon) was 46.30 ± 12.69 km2 while that calculated using the 95% ellipse averaged 34.67 ± 12.78 km2. If a circular range around leks is assumed, these values represent a radius around leks of 3.84 km and 3.32 km, respectively. Habitat Characteristics 'Study area Cursory measurements of vegetation in April and May, 1986-87 indicated a diversity of vegetative structure and species composition among and within the pastures which comprised the primary study area (Table 8). Height and canopy cover of residual grass varied widely with grass height being weakly correlated to height-density (~- 0.44, f - 0.08). Grass cover averaged 19.9% and bare ground accounted for 75.7% of the basal area. Forb cover (4.4%) was low because many of the annual forbs had not begun their growth when transects were measured. Sand sagebrush averaged 3.1% canopy cover with a mean density of 2,497 plantsfha. �19 Table 6. Home range estimates of lesser prairie-chickens in Baca County, Colorado, 1986-90. Band Mean 306 308 315 317 318 319 324 327 335 344 349 351 356 362 366 373 403 404 405. 406 408 411 418 423 424 425 426 427 432 438 442 446 447 Year Sex Lek 1986 1986 1986 1986 1986 1986 1986 1986 1987 1987 1987 1987 1987 1988 1988 1988 1989 1989 1989 1989 1989 1989 1989 1989 1989 1989 1989 1990 1990 1990 1990 1990 1990 M M 86-1 5 5 2 5 5 86-2 28 2 28 5 28 5 4 86-1 4 86-1 86-1 ·87-4 86-1 5 86-1 86-1 2 2 28 2 28 28 5 28 3 3 F F M F M F M F F M M M M F M F F M M F M M M M M M F F M F F Horne range size estimate (krn2) Min. Convex Polygon 95% ellipse 90% Harmonic 1.16 2.07 5.67 32.32 2.44 12.70 1.04 2.26 2.25 3,45 2.32 3.95 6.77 2.65 1.55 4.45 1.69 2.30 5.83 1.36 0.08 5.24 3.17 0.94 0.55 0.66 1.24 3.77 4.12 0.23 2.72 1.52 1.01 7.61 4.93 22.27 199.12 8.34 45.89 3.21 9.68 11.65 10.47 10.28 15.11 39.38 8.40 5.34 15.62 6.42 11.24 29.70 6.08 0.39 19.08 13.22 4.00 2.94 2.32 6.36 9.13 12.29 1.01 7.43 7.68 4.96 1.03 1.68 6.65 10.79 4.34 8.64 1.33 1.49 2.45 4.14 2.08 3.11 3.94 3.44 2.07 4.11 2.53 1.67 2.80 2.75 0.07 5.60 3.04 1.32 0.63 1.06 0.63 5.19 3.70 0.43 3.10 1.02 4.52 �20 Combined home range estimates (lan2) of lesser prairie-chickens Table 7. Baca County, Colorado, 1986-90. n Birds in Home range estimator Observations MCP 95% ellipse 90% HM Lek 2 Males Females Totals 8 5 13 68 24 92 5.30 51.30 51.30 6.51 163.27 51.30 4.87 22.24 33.35 Lek 5 Males Females Totals 10 23 33 88 123 211 14.59 58.03 61.92 11.46 56.08 38.09 9.12 28.25 31.38 Lek 28 Males Females Totals 9 11 20 77 69 146 15.88 29.30 25.28 11.53 40.01 24.75 6.76 24.65 19.66 Lek 86-1 Males Females Totals 8 6 14 77 37 114 16.11 21.60 36.69 16.47 28.92 24.52 8.82 .17.71 21.57. Lek sites Vegetative composition was, measured at 9 active leks on and adjacent to the primary study area (Table 9). Shrubs, including sand sagebrush were relatively scarce, and comprised approximately 10% of the cover. Short grasses were relatively common, and taller grasses were usually grazed to a low height. �Table 8. Habitat characteristics of the 41.4-km2 lesser prairie-chicken study area, Baca County, Co Lo r ado", Height (em) Sand sagebrush Canopy Density cover (X) Shrub Forb Grass Heightdensity Sec.21,T34S,R44W Sec.22,T34S,R44W Sec.23,T34S,R44W Sec.24,T34S,R44W Sec.25,T34S,R44W Sec.26,T34S,R44W Sec.27,T34S,R44W Sec.28,T34S,R44W Sec.33,T34S,R44W Sec.34,T34S,R44W Sec.35,T34S,R44W Sec.36,T34S,R44W Sec.i,T35S,R44W Sec.2,T35S,R44W Sec.3,T35S,R44W S·ec.4,T35S, R44W 40.5 44.6 40.0 33.6 31.5 31.0 29.2 29.8 35.5 36.6 32.2 33.8 28.1 31.1 42.9 32.8 3.3 3.8 3.6 3.1 4.6 4.5 1.5 1.7 3.2 4.7 4.0 1.9 2.0 4.4 3.6 6.0 6.0 27.1 22.5 16.9 13.9 17.2 6.6 7.4 11.0 15.5 12.0 4.4 9.7 9.9 17.4 11.2 0.28 0.90 1.16 1.46 0.87 1.58 0.32 0.79 1.67 1.18 0.97 0.78 1.33 0.57 1.07 0.78 0 2,360 4,160 7,480 . 2,467 2,840 2,067 4,266 2,800 1,560 2,880 533 667 1,520 1,267 3,080 Average 34.6 3.5 13.0 0.98 2,497 Location Bare ground Forbs Grass 0.0 7.1 10.0 5.4 3.1 4.6 0.6 3.0 5.1 2.6 3.5 0.2 0.1 0.9 1.9 1.5 81. 2 62.4 65.6 77 .1 86.1 70.4 55.2 84.5 81.9 72.3 77 .8 79.8 89.8 74.7 73.4 78.4 0.9 8.4 11.6 11.5 5.6 11.0 1.4 5.5 1.4 2.3 2.8 1.3 2.9 1.7 1.3 1.1 18.0 29.1 22.7 11.5 8.9 18.6 43.6 10.0 16.7 24.0 19.5 18.9 7.3 23.4 25.3 20.5 3.1 75.7 4.4 19.9 aVegetation measured in April-May, 1986-87. N I-' �22 Table 9. Vegetative characteristics at lesser prairie-chicken leks in Baca County, Colorado. Lek Sand sagebrush DensityC ~ hgt. (cm) 2 3 4 5 27 28 86-1 86-2 87-4 0 889' 0 111 111 0 1,556 111 111 Avg. 310 Buda Bogr n/a 37.4 n/a 37.0 47.0 n/a 27.3 31.0 64.0 45 10 25 30 50 20 10 65 40.6 20.0 Vegetative com:Qositiona,b Bocu Spcr Sihy Arlo FoSh 25 20 20 5 5 75 10 50 5 5 10 5 10 5 5 5 10 5 5 40 40 7.2 16.7 55 40 5 90 19.4 15.6 10.5 apercent of vegetative cover bBuda - Buch10a dacty1oides, Bogr - Boute1oua gracilis, Bocu - ~. curte:Qendula, Spcr - Sporobo1us cIYPtandrus, Sihy - Sitanion hystrix, Aristida longiseta, FoSh - forbs and shrubs. CP1ants/ha 10 10 5 15 15 5 15 10 10 10.6 Ar1o- Nest sites Vegetative characteristics were measured at 29 nest sites from 1986 to 1990. All but 2 nests were located following radio-marked hens. Measurements reflect site characteristics after hatch or nest loss as I attempted to minimize disturbance to nests during egg laying or incubation. Most nests (n - 20, 69.0%) were adjacent to shrubs, primarily sand sagebrush (n - 12, 41.4%) and soapweed (H - 6, 20.7%), with the remainder being within clumps of bunchgrasses. The tallest vegetation over nest bowls averaged 50.7 ± 14.7 cm (range 29-81 cm). Nests were typically in sites having greater (f < 0.001) height of shrubs, forbs, and grasses and higher height-density readings than areas immediately adjacent to them (Table 10). Vegetative cover was relatively sparse with a line intercept average of 69.5% bare ground, 29.4% grasses, and 1.4% forbs. Density and canopy cover of sand sagebrush varied with the site and averaged 3,410 shrubs/ha and 7.2% canopy cover. Feeding-loafing sites Vegetation at lesser prairie-chicken flush sites and paired dependent sites were analyzed separately for males and females for each month data were collected (May - Sep). Dependent sites averaged 280.5 ± 109.5 m (range 55 541 m) from flush sites and were measured the same day as corresponding flush sites. �23 Vegetative Table 10. Baca County, Colorado, cm Height-density, Nest Transect density Canopy cover, % Sandsage Grasses FQrbs Bare ground, % SD Range 47.6 37.6 14.9 9.7 29 2.7 - 81 57.2 21.2 15.7 11.0 5.8 5 6.3 - 54 24.7 36.1 27.4 15.0 12.5 9 9.3 - 60 59.0 3.2 2.0 1.5 0.7 1.0 1.0 - - 6.5 3.4 0 - 12,667 26a cm Grass height, Nest Transect ~ nest sites in 26a cm Shrub height, Nest Transect Sandsage at lesser prairie-chicken n Characteristic Forb height, Nest Transect characteristics 1986-90. 29 dm 29 29 3471 3439 29 29 aShrubs and forbs were not present 7.2 29.4 1.4 9.4 14.9 1.5 0 9.3 0 - 36.1 61. 8 6.5 69.5 14.3 38.2 - 87.7 along 4 nest transects. When comparing heights of shrubs, forbs, and grasses, and height density at flush sites of male lesser prairie-chickens with the dependent site for each month, shrub height, grass height, and height-density were greater (f < 0.05) at flush sites than at dependent sites in July only (Table 11). No differences were detected in sandsage density or canopy cover between male flush sites and dependent sites (Table 12) within any month. There were no differences in percent bare ground or canopy cover of grass or forbs between flush sites and dependent sites within any month (Table 13). Vegetative parameters measured at flush sites of female lesser prairiechickens were similar to those at dependent sites. No differences were detected in heights of shrubs, forbs, or grasses within any month and heightdensity was greater (f < 0.05) at flush sites only in July (Table 14). No differences in sandsage density or canopy cover were detected between flush sites and dependent sites within any month (Table 15). There was no difference detected in percent bare ground or canopy cover of grass or forbs between f1ush.sites and dependent sites within any month (Table 16). �24 Table 11. Shrub, forb, and grass height (cm) and height-density (dm) at flush sites and dependent sites of male lesser prairie-chickens in southeastern Colorado, 1986-90. May Jun SD Shrub height FSa DSb Jul SD Aug SD Sep SD SD 35.8 34.3 14.2 8.9 37.2 36.6 13.7 9.5 46.9 40.3 8.3C 41.0 11.1 8.6 34.3 18.5 Forb height FS DS 9.2 7.9 2.6 1.7 9.6 8.7 2.7 3.8 23.7 21.2 6.0 8.0 23.9 20.7 7.7 11. 5 29.6 26.6 24.6 10.0 Grass height FS DS 11.0 10.0 2.5 1.9 10.2 10.3 2.5 3.0 27.8 20.2 7.9C 4.5 28.4 27.9 8.5 15.3 23.4 18.9 0.6 26.7 Height-Density 1.25 FS 1.13 DS 0.73 0.83 2.01 1. 96 0.87 0.67 2.17 1.44 0.55C 0.56 2.04 1.24 1. 50 1. 75 1. 30 0.92 0.71 1.05 33.6 11.7 29.8 18.4 aF1ush site. bDependent site . .CSignificant (f < 0.05) difference between flush site and dependent site. Table 12. Sandsage density (plants(ha) and canopy cover at flush sites and dependent sites of male lesser prairie-chickens in southeastern Colorado, 1986-90. Month Flush site SD ~ May Jun Ju1 Aug Sep 2970 4077 3111 3905 1834 2558 3873 34i2 2242 2595 Densit::i Dependent site SD ~ 3303 5205 1444 3000 500 1946 4198 1951 2465 707 Gano~::icover {%) Flush site Dependent site SD SD ~ ~ 4.0 6.4 5.6 9.7 0.0 4.8 9.6 5.2 9.4 4.1 6.3 1.7 9.1 3.2 5.6 6.8 2.3 12.7 4.5 �25 Table 13. Bare ground (%) and canopy cover (%) of grasses and forbs at flush sites and dependent sites of male lesser prairie-chickens in southeastern Colorado, 1986-90. Cano12:£cover (X) Month May Jun Jul Aug Sep Bare ground (X) DS6 FSa SD SD ~ ~ 86.5 90.4 56.3 72.6 46.3 14.6 5.1 22.3 15.5 29.4 89.9 91.0 59.8 57.4 41. 7 Forbs Grass FS 8.5 5.4 21.5 25.4 21.1 DS FS ~ SD ~ SD ~ 10.7 7.9 42.7 26.5 53.7 14.9 5.8 23.3 15.7 29.4 8.3 7.4 39.4 42.2 58.0 8.7 6.0 21.8 25.7 21.5 1.3 1.8 0.9 0.8 0.0 DS SD 1.2 2.0 1.0 1.7 SD ~ 1.0 1.7 0.8 0.5 0.2 1.1 1.5 0.7 0.4 0.3 aF1ush site. bDependent site. Table 14. Shrub, forb, and grass height (cm) and height-density flush sites and dependent sites of female lesser prairie-chickens southeastern Colorado, 1986-90. May Jul Jun ~ SD 13.2 7.5 35.2 33.7 13.6 10.7 11.7 5.0 Grass height FS 17.6 DS 12.5 11.9 7.9 SD Shrub height 42.8 FSa DSb 37.2 Forb height FS DS ~ Height-density 2.49 FS 1.09 DS 1.01 0.50 ~ SD ~ SD 14.8 16.5 41.0 41.1 7.9 18.7 37.0 36.3 10.6 21.5 19.8 20.7 5.5 6.6 22.5 22.5 7.7 7.7 21.2 23.0 9.1 10.5 23.5 21.4 8.6 8.6 25.9 27.0 7.6 10.7 28.2 29.3 11.6 5.9 ~ 10.3 13.2 40.9 35.8 14.3 11.6 3.8 3.9 12.3 12.2 6.4 3.9 1.22 1.54 Sep Aug SD 2.15 1.56 (dm) of in 2.03 1.74 0.8F 0.74 1.25 0.47 1.32 0.38 1.33 1.52 1.52 1.18 aF1ush site. bDependent site. CSignificant difference (f < 0.05) between flush site and dependent site. �26 Table IS. Sand sagebrush density (p lant.syha) and canopy cover at.flush sites and dependent sites of female lesser prairie-chickens in southeastern Colorado, 1986-90. Month Sand sagebrush densit~ Flush site Dependent site SD SD ~ ~ Cano:Q~ cover (%} Flush site Dependent site SD SD ~ ~ May Jun Jul Aug Sep 5778 6857 5000 2833 6334 12.2 11.4 13.0 4.0 11.8 6883 4857 3957 3590 6471 3000 5667 3030 2916 2000 3283 3991 3497 3458 1660 8.3 13.9 13.5 7.4 11.0 2.2 5.8 6.7 4.1 2.1 2.0 6.8 10.0 5.8 3.9 Table 16. Bare ground and canopy cover of grasses and·forbs at flush sites and dependent sites of female lesser prairie-chickens in southeastern Colorado, 1986-90. Month May Jun Jul Aug Sep s 88.8 85.6 73.7 69.1 68.2 Bare ground ~Xl b FS8 OS SO SO .& 5.6 7.4 16.4 15.9 23.7 78.5 88.6 68.5 66.3 64.7 15.1 8.2 20.0 20.4 24.2 Canoex cover ~*l Grass FS OS SO SO .& .& 9.9 12.0 24.8 30.1 31.2 5.9 8.8 17.2 16.5 23.7 19.4 9.5 30.4 32.4 35.0 16.3 8.6 21.0 21.6 24.0 Forbs s 1.3 2.4 1.5 0.8 0.6 FS OS SO 0.5 2.0 1.3 0.9 0.5 .& 2.0 2.5 1.2 1.2 0.3 SO 0.5 2.1 1.3 1.4 0.3 aFlush site. boependent ·s.i te~ DISCUSSION Lek counts and Surveys Counts of male lesser prairie-chickens on leks have been used to estimate population size or monitor population changes (Davison 1940, Copelin 1963, Hoffman 1963, Jones 1963, Crawford 1974, Crawford and Bolen 1976). An implicit assumption in using lek counts for population estimates is that either all males or at least a constant portion of the male population attend arenas each year and that changes in counts are correlated to actual population changes. Studies of black grouse (Lyrurus tetrix) (Robel 1969) and sharp-tailed grouse (TyID:Qanuchus:Qhasianellus) (Rippin and Boag 1974) suggest that less than onehalf of the male population may attend a lek on any given day, and that for sharp-tailed grouse, there is a segment of the adult male population which fails to become territorial in spring (Moyles and Boag 1981). Beck and Braun (1980) observed that there was no proven relationship between counts of males �27 sharp-tailed grouse, there is a segment of the adult male population which fails to become territorial in spring (Moy1es and Boag 1981). Beck and Braun (1980) observed that there was no proven relationship between counts of males on leks and sage grouse (Centrocercus urophasianus) populations. Despite these shortcomings, 1ek counts continue to be used as a management tool for population estimates of sage grouse (Patterson 1952, Stanton 1958, Dalke et a1. 1963, Eng 1963, Hartzler 1972, Jenni and Hartzler 1978, Rothenmaier 1979, Beck and Braun 1980), greater prairie-chickens (Tympanuchus cupido) (Ammann 1957, Robel 1970, Westemeier 1971, Hamerstrom and Hamerstrom 1973, Kirsch et a1. 1973) and sharp-tai1ed,grouse (Ammann 1957, Kobriger 1965, Pepper 1972, Hillman and Jackson 1973, Kirsch et a1. 1973, Yde 1977, Mattise 1978, Nielson 1978, Ziegler 1979). In recent years there have been efforts to count pra1r1e grouse leks instead of individual males on leks as an index to annual fluctuations in population size (Cannon and Knopf 1981, Martin and Knopf 1981). Some states have traditional "booming ground" routes they survey each year to detect changes in number of leks (Horak 1985). Because these routes were not randomly selected, population inferences could not be made to areas other than the routes. The value of this technique assumes a high degree of precision (repeatability) if not accuracy, and little or no observer bias. This study indicated both poor precision and accuracy as well as individual differences among observers, even under similar weather conditions. '. Aerial transects appeared to give accurate estimates of lek density when flown during the peak of hen attendance, probably because the aircraft caused hens to f1ush'where they were more easily detected. The number of males attending leks several ·weeks later was similar but less than half the 'leks were· detected, presumably because males were less active and less likely to flush. Too few flights were made to estimate the precision of this technique, and because of the high cost, it cannot be recommended as a standard management practice. However, it does have the advantage in surveying large areas quickly and will likely provide information on distribution of prairiechickens if conducted during the peak of hen attendance. Data from lek counts and intensive and extensive lek surveys in Baca County suggest that populations of lesser prairie-chickens increased from 1986 to 1990, and'were likely increasing since at least 1980. More leks were located both on and adjacent to the study area and some historic leks which had been inactive previously became active for one or more years. Since lek activity and 1ek locations have changed annually, the best method for documenting incremental changes in prairie chicken population remains intensive ground surveys. Survey procedures should incorporate monitoring of leks active in the previous year, surveys of,historic and abandoned lek sites, and listening routes to detect newly established leks. Trapping and banding Walk-in funnel traps were relatively effective for capturing lesser pra1r1echickens on leks. Trapping success for males was highest during the initial trapping period before they became habituated to traps and altered territory boundaries with respect to traps and chicken-wire leads. Moving traps to different locations on leks resulted in additional males being trapped. Capturing hens was most successful when wire leads bisected leks and caused �28 hens to enter traps as they moved across leks. When several hens were on a lek at the same time they often moved as a flock, although aggression was commonly observed. The chasing of hens by males often resulted in captures of both males and females. Multiple captures of females was more common than multiple captures of males. More than one male in a trap resulted in constant fighting and injury (scrapes to head and wings) to the subordinate bird, unless they were removed quickly. In contrast, when> 1 hen was in the same trap there was little aggression and few injuries, thus they could be left in traps longer. Most birds (n - 139, 93.3%) were trapped during April when hens visited leks and male display activity was highest. During May, few hen were observed on leks and male activity declined in intensity and duration, and few males were captured. Movements to Nest Sites Mean distances moved by radio-marked hens from lek-of-capture to nest sites has ranged from 1.2 km for 8 hens in Texas (Sell 1979) to 3.4 km for 37 hens in New Mexico, with individual hens moving up to 14 km (Davis "et al. 1979). Distances moved from leks to nests and hen home ranges might be attributed to habitat characteristics or weather, with shorter movements and smaller home ranges in better quality habitats and during seasons of average or above rainfall (Sell 1979, Merchant 1982). There was no indication that radio packages affected movements of birds. Other studies of nesting lesser prairie-chickens (Copelin 1963, Donaldson 1969, Haukos 1988) failed to mention movements of hens to nests or distances from nearest leks to nests. Identification of nesting habitat based on movements between lek-of-capture and nest site may be misleading if hens visit several leks before mating as shown for greater prairie-chickens (Schroeder 1990)~ Although no radio-marked hen was known to visit other leks, such movements could have easily occurred without my knowledge or occurred prior to being marked. Since the inter-lek distance in this study (1.32 km) was less than the distance from lek-ofcapture to nest site (1.80 km) it is likely that hens visited> 1 lek or traveled past other leks during the breeding season. Movement of hens from leks to nests and lek-nest distances have implications for identifying critical nesting and brood-rearing habitats. If females tend to cluster nests around leks, or if leks are focal to female home ranges (Bradbury 1981), then nesting areas can be protected from disturbance or enhanced through management without documenting nesting preferences for each population. Furthermore, leks are easier to locate than nests, especially when nests are widely dispersed, well camouflaged, and occur at low densities. Home Range Although home range is defined as the area used by an individual during its normal activities (Burt 194~), home range size has proven difficult to quantify or estimate. The minimum convex polygon (Mohr 1947) is the oldest and most common estimate of home range size but is a function of the number of observations used in the estimate (Boulanger and White 1990). By eliminating individual birds having < 10 locations I was able to minimize this error and correlations between home range size and number of locations were not significant for either males or females. �29 Males had smaller summer home ranges than hens (all 3 home range estimators) which may reflect their attachment to leks in spring and fall and the importance of maintaining territories on leks to assure a chance of mating during the breeding season. Home ranges of males often overlapped and radiomarked males were often observed with other birds (presumed males) during summer. This association among males may be important for maintaining dominance relationships among males attending the same lek. The value of larger home ranges to hens is unknown but may be related to nesting and rearing young. Too few radio-marked hens were successful in nesting to test for differences in home range size between hens with and without broods. Vegetation on the Study Site The diversity of vegetative structure and species composition reflects different soil and topographic conditions as well as differences in past use and revegetation efforts. Draws were typically vegetated with sand sagebrush and a diversity of annual forbs (many of which hadn't begun growth when transects were measured). Individual cadastral sections were dominated by different grass species, usually sand dropseed or sideoats grama, which reflected previous revegetation efforts by the U. S. Forest Service. Height of residual grasses and height density were influenced by livestock grazing the previous season, especially grazing from October to December when grasses were dormant. The amount of bare ground was attributed to lack of soil moisture which reflected low annual rainfall, high summer ambient temperatures and evaporation, and sandy soils which retained subsurface moisture poorly. Lek Habitat The use of habitats with short vegetation for lekking activity of lesser prairie-chickens has been well documented (Copelin 1963, Jones 1963, Crawford and Bolen 1973, Davis et al. 1979). The advantage to sites with high visibility inclu~e enhancement of visual displays for territorial defense as well as for attracting hens (Kermott and Oring 1974, Sparling 1981). Also, open habitats allow better detection of predators which may be attracted to displaying grouse (Hamerstrom et al. 1965, Anderson 1969). Nesting Habitat Preferred nest sites of lesser prairie-chickens usually occur where height of residual cover is greatest «Copelin 1963, Haukos 1988) or in ungrazed or lightly grazed rangeland (Davis et al. 1979). Height of vegetation above nest bowls usually exceeded 50 cm (Donaldson 1969, Suminski 1977, Riley 1978) and was similar regardless of plant height (shrub or grass). In areas where livestock grazing was heavy, hens tended to nest under shrubs (Davis et al. 1979, Sell 1979, Wisdom 1980, Merchant 1982) even though nests under shrubs tended to be less successful than nests under bunchgrasses (Sell 1979). Furthermore, regardless of habitat type, height and density of vegetation at successful nests were greater that at unsuccessful nests (Davis et al. 1979). Because of these relationships, nesting success was inversely related to grazing pressure. Grazing has previously been implicated in causing declines of lesser prairie-chickens (Jackson and DeArment 1963). �30 Feeding-Loafing Sites With few exceptions, both male and female lesser prairie-chickens used habitats that were similar to those at dependent sites. Since dependent sites were within flight distance of the grouse (avg. - 280.5 m), any differences in habitat variables would have indicated possible habitat selection. Few studies addressed summer habitat preferences of lesser prairie-chickens in sand sagebrush habitats, primarily because most occupied ranges are characterized as shinnery oak (Quercus havardii)" grasslands. Summer habitat use by lesser prairie-chickens within sand sagebrush rangelands showed higher relative use of shrubby cover to grassy cover by broods and flocks of males (Riley 1978, Davis et a1. 1979, Cannon 1980). This is thought to reflect the higher abundance of insect food in shrub and forb substrates compared to grass (Davis et al. 1980). Insects comprise a large portion of the summer diet of lesser prairie-chickens (Jones 1964, Davis et a1. 1980) and may be more visible or accessible in shrub and grass habitats. Shrub and forb dominated habitats may also have more bare ground which may facilitate movement by young broods when foraging. Microclimate may also be a factor in selection of shrubby habitats as lesser prairie-chickens are known to seek shade during hot weather (Copelin 1963) and sand sagebrush may provide more shade than grass. This may be critical when livestock grazing results in reduction of grass height but has little impact on shrub height or structure. The lack of selection of habitat structure by lesser prairie-chickens documented in this study may be. the result of relatively uniform habitat conditions on the study area compared to other rangeland. The study area was selected on the basis of reported high densities of lesser prairie-chickens which suggests that habitat conditions were generally favorable. Furthermore, the U.S. Forest Service has managed pasture lAE for wildlife values including lesser prairie-chickens which resulted in less use of forage in that pasture by livestock than on adjacent rangelands. In more marginal habitat, selection may have been easier to detect. LITERATURE CITED Aldrich, J. 1963. Geographic orientation of American Tetraonidae. Manage. 27:529-545. J. Wildl. Ammann, G. A. 1944. Determining the age of pinnated and sharp-tailed grouse. J. Wild1. Manage. 8:170-171. 1957. The prairie grouse of Michigan. Bull. 1. Lansing. 200 pp. Amstrup, S. C. 1980. 44:214-217. Michigan Dep. Conserv. Tech. A radio-collar for game birds. J. Wildl. Manage. Anderson, R. K. 1969. Prairie chicken responses to booming-ground cover type and height. J. Wildl. Manage. 33:636-643, �Bailey, A. M., and R. J. Niedrach. 1965. Birds of Colorado. Mus. Nat. History, Denver, Colo. 454 pp. Baker, M. F. 1953. Prairie chickens of Kansas. History, Misc. Pub1. 5. 68 pp. Vol. 1. Denver Univ. Kansas Mus. Nat. Beck, T. D. l., and C. E. Braun. 1980. The strutting ground count: variation, traditionalism, management needs. Proc. West. Assoc. Fish and Wildl. Agencies 60:558-566. Bendire, C. E. 1892. Life histories of North American birds with special reference to their breeding habits and eggs. U.S. Nat1. Mus. Spec. Bull. 1. 446 pp. Bent, A. C. 1932. Life histories of North American gallinaceous birds. S. Nat1. Mus. Bull. 162. 490 pp. U. Boulanger, J. G., and G. C. White. 1990. A comparison of home range estimators using Monte Carlo simulation. J. Wi1d1. Manage. 54:310-315. Bradbury, J. W. 1981. The evolution of leks. Pages 138-169 in R. D. Alexander and D. W. Tinkle, eds. Natural selection and social behavior: recent research and new theory. Chiron Press, New York, N.Y. Burt, W. H. 1943. Territoriality and home range concepts as applied to mammals. J. Mammal. 24:346-352. Canfield, R. H. 1941. range vegetation. Application of the line intercept method in sampling J. For. 39:388-394. Cannon, R. W. 1980. Current status and approaches to monitoring populations and habitats of lesser prairie chickens in Oklahoma. M.S. Thesis, Univ. Oklahoma, Stillwater. 62 pp. ______ , and F. L. Knopf. grouse populations. 1981. Lek numbers as a trend index to prairie J. Wildl. Manage. 45:776-778. Colorado Division of Wildlife. 1983. Todays Strategy ...tomorrows wildlife. A comprehensive management plan for Colorado's wildlife. Colorado Div. Wild1., Denver. 96 pp. Copelin, F. F. 1963. The lesser pra~r~e chicken in Oklahoma. Wildl. Conserv. Dep., Tech. Bull. 6. 58 pp. Oklahoma Crawford, J. A. 1974. The effects of land use on lesser pra~r~e chicken populations in west Texas. Ph.D. Diss., Texas Tech. Univ., Lubbock. pp, 63 1980. Status, problems, and research needs of the lesser prairie chicken. Pages 1-7 in P. A. Vohs and F. A. Knopf, eds. Proc. Prairie Grouse Symp. Oklahoma State Univ., Stillwater. ______ , and E. G. Bolen. 1973. Spring use of stock ponds by lesser prairie chickens. Wilson Bull. 85:471-472. �32 ______ , and Texas. 1976. Effects of land use on lesser prairie chickens in J. Wi1d1. Manage. 40:96-104. Dalke, P. D., D. B. Pyrah, D. C. Stanton, J. E. Crawford, and E. F. Sch1atterer. 1963. Ecology, productivity, and management of sage grouse in Idaho. J. Wi1d1. Manage. 27:811-841. Davis, C. A., T. Z. Riley, R. A. Smith, and M. J. Wisdom. 1980. Springsummer foods of lesser prairie chickens in New Mexico. Pages 75-80 in P. A. Vohs and F. L. Knopf, eds. Proc. Prairie Grouse Symp., Oklahoma State Univ., Stillwater. ______ , H. R. Suminski, and M. J. Wisdom. 1979. Habitat evaluation of lesser prairie chickens in eastern Chaves County, New Mexico. New Mexico State Univ. Agric. Exp. Sta., Las Cruces. 141 pp. Davison, V. E. 1940. Manage. 4:55-62. An 8-year census of lesser prairie chickens. J. Wildl. Dixon, K. R., and J. A. Chapman. 1980. Harmonic mean measure of animal activity areas. Ecology 61:1040-1044. Donaldson, D. D. 1969. Effect on lesser prairie chickens of brush control in western Oklahoma. Ph.D. Thesis, Oklahoma State Univ., Stillwater. 80 pp. Eng, R. L. 1963. Observations on the breeding biology of male sage grouse. J. Wildl. Manage. 27:841-846. Hamerstrom, F., D. D. Berger, and F. N. Hamerstroni, Jr. 1965. The effect of mammals on prairie chickens on booming grounds. J. Wi1dl. Manage. 29:536-542. Hamerstrom, F. N., and F. Hamerstrom. 1973. The prairie chicken in Wisconsin. Wisconsin Dep. Nat. Res. Tech. Bull. 64. 52 pp. Harrington, H. D. 1964. Manual of the plants of Colorado. Swallow Press Inc., Chicago. 666 pp. Second ed. The Hartzler, J. E. 1972. An analysis of sage grouse behavior. Univ. Montana, Missoula. 234 pp. Ph.D. Thesis., Haukos, D. A. 1988.' Reproductive ecology of lesser prairie-chickens in west Texas. M.S. Thesis, Texas Tech. Univ., Lubbock. 81 pp. Hillman, C. N., and W. W. Jackson. 1973. The sharp-tailed grouse in South Dakota. South Dakota Dep. Game, Fish and Parks Tech. Bull. 3. 64 pp. Hoffman, D. M. 1963. The lesser prairie chicken in Colorado. Manage. 27:726-732. Horak, G. J. 1985. Kansas prairie chickens. Wildl. Bull. 5. 65 pp. J. Wildl. Kansas Fish and Game Comm., �33 Jackson, A. S., and R. DeArment. 1963. The lesser pra~r~e chicken in the Texas panhandle. J. Wildl. Manage. 27:733-737. Jenni, D. A., and J. E. Hartzler. 1978. Attendance at a sage grouse lek: implications for spring census. J. Wildl. Manage; 42:46-52. Johnsgard, P. A. 1973. Grouse and quails of North America. Press, Lincoln. 553 pp. Univ. Nebraska Jones, R. E. 1963. Identification and analysis of lesser and greater prairie chicken habitat. J. Wildl. Manage. 27:757-778. 1964. Habitat used by lesser prairie chickens for feeding related to seasonal behavior of plants in Beaver County, Oklahoma. Southwest. Nat. 9:111-117. Judd, S. D. 1905. The grouse and wild turkeys of the United States, and their economic value. U. S. Dep. Agric., BioI. Surv. Bull. 24. 55 pp. Kermott, L. H., and L. W. Oring. Acoustical communication of male sharptailed grouse (Pedioecetes phasianellus) on a North Dakota dancing ground. Anim. Behav. 23:315-386. Kirsch, L. M., A. T. Klett, and H. W. Miller. 1973. Land use and pra~r~e grouse population relationships in North Dakota. J. Wildl. Manage. 37:449-453. Kobriger, G. D. 1965. Status, movements, habitats, and foods of prairie grouse on a sandhills refuge. J. Wildl. Manage. 29:788-800. Koeppl, J. W., N. A. Slade, and R. S. Hoffmann. 1975. A bivariate home range model with possible application to ethological data analysis. J. Mammal. 56:81-90. Litton, G. W. 1978. The lesser prairie chicken and its management in Texas. Texas Parks and Wildl. Dep. Booklet .7000-25. 23 pp. Loeffler, C .. 1983. The status and management of the lesser pra~r~e chicken in Colorado. Unpubl. rep. Colorado Div. Wildl., Colorado Springs. 9 pp. Martin, S. A., and F. L. Knopf. 1981. Aerial survey of greater prairie chicken leks. Wildl. Soc. Bull. 9:219-221. Mattise, S. N. 1978. Effects of grazing systems on sharp-tailed grouse habitat. M.S. Thesis, South Dakota State Univ., Brookings. 46 pp. Merchant, S. S. 1982. Habitat-use, reproductive success, and survival of female lesser prairie chickens in two years of changing weather. M.S. Thesis, New Mexico State Univ., Las Cruces. 62 pp. Mohr, C. O. 1947. Table of equivalent populations of North American small mammals. Am. Midl. Nat. 37:223-249. �34 Moyles, D. J. L., and D. A. Boag. 1981. Where, when, and how male sharptailed grouse establish territories on arenas. Can. J. Zool. 59:15761581. Nielsen, L. S. 1978. The effects of rest-rotation grazing on the distribution of sharp-tailed grouse. M.S. Thesis, Montana State Univ. , Bozeman. 52 pp. Oberholser, H. C. 1974. Austin. 503 pp. Patterson, R. L. 1952. Colo. 341 pp. The bird life of Texas. Vol. 1. Univ. Texas Press, The sage grouse in Wyoming. Sage Books, Inc. Denver, Pepper, G. W. 1972. The ecology of sharp-tailed grouse during spring and summer in the aspen parklands of Saskatchewan. Saskatchewan Dep. Nat. Res. Wildl. Rep. 1. 55 pp. Rash, M. T. 1985. 25 May, 1985. pp. Survey of the lesser prairie-chicken in Colorado, 3 AprilUnpubl. rep. Colorado Div. Wildl., Colorado Springs. 22 Riley, T. Z. 1978. Nesting and brood-rearing habitat of lesser pra~r~e chickens in southeastern New Mexico. M.S. Thesis, New Mexico State Univ., Las Cruces. 59 pp. Rippin, A. B., and D. A. Boag. 1974. Recruitment to populations of male sharp-tailed grouse. J. Wild1. Manage'. 38:616-621. Robel, R. J. 1969. Movements and flock stratification within a population of blackcocks in Scotland. J. Anim. Ecol. 38:755-763. 1970. Possible role of behavior in regulating greater prairiechicken populations. J. Wildl. Manage. 34:306-312. ______ , J. N. Briggs, J. J. Cebula, A. D. Dayton, and L. C. Hulbert. 1970. Relationships between visual obstruction measurements and weight of grassland vegetation. J. Range Manage. 23:295-297. Rothenmaier, D. 1979. Sage grouse reproductive ecology: breeding season movements, strutting ground attendance and site characteristics, and nesting. M.S. Thesis, Univ. Wyoming, Laramie. 97 pp. Sands, J. L. 1978. Game bird studies'. New Mexico Dep. Game and Fish. Fed. Aid Proj. W-l04-R-19. Sante Fe. 5 pp. Schroeder, M. A. 1990. Movement and dispersion of greater prairie-chickens in northeastern Colorado. Ph.D. Diss., Colorado State Univ., Fort Collins. 135 pp. Sell, D. L. 1979. Spring and summer movements and habitat use by lesser pra~r~e chicken females in Yoakum County, Texas .. M.S. Thesis, Texas Tech. Univ., Lubbock. 42 pp. �35 Sparling, D. W. 1981. Communication in pra~r~e grouse. I. Information content and intraspecific functions of principal vocalizations. Behav. Neuro~. BioI. 32:463-486. Stanton, D. C. 1958. A study of breeding and reproduction in a sage grouse population in southeastern Idaho. M.S. Thesis, Univ. Idaho, Moscow. 87 pp. Stuwe, M., and C. E. Blohowiak. 1985. McPAAL software package for calculating home ranges. Conserv. Res. Cent., Natl. Zool. Park, Smithsonian Inst., Front Royal, VA. Suminski, H. R. 1977. Habitat evaluation for lesser prairie-chickens in eastern Chaves County, New Mexico. M.S ..Thesis, New Mexico State Univ. , Las Cruces. 49 pp. Taylor, M. A., and F. S. Guthery. 1980. Fall-winter movements, ranges and habitat use of lesser prairie-chickens. J. Wildl. Manage. 44:521-524. Toepfer, J. E., J. A. Newell, and J. Monarch. 1988. A method for trapping prairie grouse hens on display grounds. Pages 21-23 in A. D. Bjugstad, Tech. Coord. Prairie chickens on the Sheyenne National Grasslands. U.S. Dep. Agric., For. Servo Gen. Tech. Rep. RM-159. Westemeier, R. L. 1971. The history and ecology of pra~r~e chickens in central wisconsin. Univ. Wisconsin, Res. bull. 281. 66 pp. Wilson, D. L. 1987. Nesting habitat of lesser prairie chickens in Roosevelt and Lea counties, New Mexico. M.S. Thesis, New Mexico State Univ., Las Cruces. 36 pp. Wisdom, M. J. 1980. Nesting habitat of lesser pra~r~e chickens in eastern New Mexico. M.S. Thesis, New Mexico State Univ., Las Cruces. 38 pp. Yde, C. A. 1977. Effects of rest-rotation grazing on the abundance and distribution of sharp-tailed grouse. M.S. Thesis, Montana State Univ., Bozeman. 70 pp. Ziegler, D. L. 1979. Distribution and status of the Columbian sharp-tailed grouse in eastern Washington. Upland Game Invest., Washington Dep. Game Fed. Aid. Wildl. Rest. Compl. Rep. Proj. W-70-R-18. 26 pp. Prepared by Kenneth M. Giesen Wildlife Researcher C ��37 JOB FINAL REPORT Colorado State of: Project: W-152-R Work Plan: Job Title: 12 Job _l2g_ Chronology of Breeding and Nesting Activities Relation to Timing of Spring Hunting Seasons Period Covered: Author: Upland Bird Research of Wild Turkeys in 01 July 1985 through 30 June 1991 Richard W. Hoffman Personnel: T. D. Abell, J. L. Aragon, C. E. Braun, R. W. Hoffman, R. L. Holder, B. S. Linkhart, T. B. Lundt, R. K. Mueckler, and T. J. Spezze, Colorado Division of Wildlife ABSTRACT An interim final report was prepared ~n 1989-90. The following manuscripts incorporating data presented in this report were prepared or published in 1990-91: Hoffman, R. W. 1990. Chronology of gobbling and nesting activities Merriam's wild turkeys. Natl. Wild Turkey Symp. 6:25-31. of 1991. Spring movements, roosting activities, and home range characteristics of male Merriam's wild turkeys. Southwest. Nat. press) . Schmutz, J. A., and R. W. Hoffman. 1991. Variable first prebasic molt in Rio Grande and Merriam's wild turkeys. Wilson Bull. 300. primary 103:295- Western States Wild Turkey Commi,ttee. 1991. Management guidelines for Merriam's wild turkeys. Natl. Wild Turkey Fed., Edgefield, South Carolina. 55pp. (Submitted for review). Prepared by Mtv IItL,..___--: Richard W. Hoffman Wildlife Researcher C 36: (in ��39 JOB FINAL REPORT State of: Colorado Project: Work W-ls2-R Plan: Job Title: Period Author: 14 Upland : Job _3_ Seasonal Covered: Michael Bird Research Movement 01 January and Habitat 1986 through Use by Greater 30 June Prairie-Chickens 1991 A. Schroeder Personnel: M. A. Schroeder and G. C. White, Colorado State University; C. E. Braun, J. F. Corey, F. M. Pusateri, M. E. Rasmussen, and L. A. Robb, Colorado Division of Wildlife ABSTRACT Several hypotheses have been proposed to explain the evolution of clumped (lek mating) from dispersed (territorial mating) polygyny; most recent models suggest that increased female home range size leads to increased female choice for males and/or leks. I propose an alternative, the female tolerance hypothesis, in which decreased intrasexual aggression among females during the breeding season results in increased aggregation of males. Predictions of the female preference and hotspot hypotheses were examined in a population of greater prai~ie-chickens (Tympanuchus cupido) in northeastern Colorado during 1986-88. Nest-lek distances were used as indirect measures of home range size. Sixty-six of 89 females (74%) nested closer to a lek other than where captured and 67 of 79 females (85%) visited> 1 lek during the breeding season. These results contradicted predictions of the female preference hypothesis. Breeding potential was estimated as a measure of proximity between any given location and nest locations of females. Monte Carlo simulations were used to examine breeding potentials under. varying conditions of actual and random lek locations and actual and random distribution of nest locations. Distributions of both leks and nests supported predictions of the hotspot hypothesis. Examination of seasonal movement between breeding and wintering ranges indicated that spring migration occurred during February-March and autumn migration occurred during June-August. Most of the variability in timing of autumn migration for females appeared to be due to their brood status; females without broods migrated earlier than those with broods. The average migration distance between breeding and wintering ranges was 10.6 km for females and 2.9 km for males. Examination of seasonal locations for individuals between years indicated that most greater prairie-chickens displayed site fidelity to both breeding and winter areas. �A lek attendance rate of 50% for males has been used to estimate population size for greater prairie-chickens; consequently lek attendance was examined in northeastern Colorado during 1986-90. Observations of 21 radio-marked males indicated they were on leks 95.1% of the time during peak display periods (Mar-Apr); consequently use of 50% as a value for lek attendance may result in over-estimates of population size. Mean annual turnover rate of leks was 23.8%; 24 leks were active aIlS years. �41 DISSERTATION MOVEMENT AND DISPERSION OF GREATER PRAIRIE-CHICKENS IN NORTHEASTERN COLORADO Submitted by Michael A. Schroeder Department of Fishery and Wildl ife Biology In partial fulfillment of the requirements for the Degree of Doctor of Philosophy Colorado State University Fort Collins, Colorado Fall 1990 �42 COLORADO STATE UNIVERSITY 20 July 1990 WE HEREBY RECOMMEND THAT THE DISSERTATION PREPARED UNDER OUR SUPERVISION BY MICHAEL A. SCHROEDER ENTITLED MOVEMENT AND DISPERSION OF GREATER PRAIRIE-CHICKENS IN NORTHEASTERN COLORADO BE ACCEPTED AS FULFILLING IN PART REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY. Adviser /' ~:EhA-0.J~.~ Department Head �43 ABSTRACT OF DISSERTATION MOVEMENT AND DISPERSION OF GREATER PRAIRIE-CHICKENS IN NORTHEASTERN COLORADO Several hypotheses have been proposed to explain the evolution of .clumped (lek mating) from dispersed (territorial mating) polygyny; most recent mod~ls suggest that increased female home range size leads to increased female choice for males and/or leks. I propose an alternative, the female tolerance hypothesis, in which decreased intrasexual aggression among females during the breeding season results in increased aggregation of males •. Predictions of the female preference and hotspot hypotheses were examined in a population of greater prairie..: chickens (Tympanuchus cupido) in northeastern Colorado during 1986-88. Nest-lek distances were used as indirect measures of home range size. Sixty-six of 89 females (74%) nested closer to a lek other than where captured and 67 of 79 females (85%) visited> season. 1 lek during the breeding These results contradicted predictions of the female preference hypothesis. Breeding potential was estimated as a measure of proximity between any given location and nest locations of females. Monte Carlo simulations were used to examine breeding potentials under varying conditions of actual and random lek locations and actual and random distribution of nest locations. Distributions of both leks and nests supported predictions of the hotspot hypothesis. �...•...• Examination of seasonal movement between breeding and wintering ranges indicated that spring migration occurred during February-March and autumn migration occurred during June-August. variability Most of the in timing of autumn migration for females appeared to be due to their brood status; females without broods migrated earlier than those with broods. The average migration distance between breeding and wintering ranges was 10.6 km for females and 2.9 km for males. Examination of seasonal locations for individuals between years indicated that most greater prairie-chickens displayed site fidelity to both breeding and winter areas. A 1ek attendance rate of 50% for males has been used to estimate population size for greater prairie-chickens; consequently 1ek attendance was examined in northeastern Colorado during 1986-90. Observations of 21 radio-marked males indicated they were on leks 95.1% of the time during peak display periods (Mar-Apr); consequsnt ly use of 50% as a value for 1ek attendance may result in over-estimates population size. of Mean annual turnover rate of leks was 23.8%; 24 leks were active all 5 years. Michael A. Schroeder Department of Fishery and Wildlife Biology Colorado State University Fort Collins, CO 80523 Fall 1990 �45 TABLE OF CONTENTS TITLE PAGE ........................................................ i SIGNATURE PAGE ...•................................................ ABSTRACT ii .•...••.•••.••••......•.••...••.•......•......•......••..• ii; TABLE OF CONTENTS ACKNOWLEDGMENTS . vi viii •••••••••••••••.•••...•••••••••••.••.••••••..•.•••• CHAPTER 1. TOLERANCE AMONG FEMALES AND EVOLUTION OF LEKS . 1 . 1 Male fitness and evolution of l~ks •..........•.•.•.•.• Female fitness and evolutlon of leks ••.••...•......... Least costly male hypothesis ...•....•......•.•.. Female preference hypothesis ..•••••••••...••••.. Hotspot'hypothesis .•.••••.••.•••...•..•••.• ~•••. Hotshot hypothesis ..••••.....•..••••.•..•....••• Female tolerance ...•••..••..•.•••.••••.•.•...•.. 8 10 Rev; ew of hypotheses Proposed tests literature cited . .o •••••••••••• ~ ••••••••••••••••••••••••••••• CHAPTER 2. MOVEMENT AND LEK VISITATION BY FEMALE .GREATER PRAIRIE-CHICKENS: A TEST OF THE FEMALE PREFERENCE HYPOTHESIS OF LEK EVOLUTION ....•..•.........••.•............ I nt roduct ion Methods· Resul ts Home range size Lek visitation Di scussion " ~ . . . . . Literature cited ....•.•........••........•...•.......•...... CHAPTER 3. MOVEMENT BY FEMALE GREATER PRAIRIE-CHICKENS IN RELATION TO LEK LOCATION: EVALUATION OF THE HOTSPOT HYPOTHESIS OF LEK EVOLUTION 12 13 14 17 19 20 29 . ' 11 29 30 34 36 41 44 50 . 52 Introduction . 52 Methods . 54 Analysis Analysis Analysis Analysis 1 2 3 4 57 58 59 59 �46 TABLE OF CONTENTS (CONTINUED) Results..................................................... 60 Nest - capture lek movements ••..•••..•.••.•••..••..•.. 60 -1 •••••••••••••••••••••••••••••••••••••••••••• Analysi s 2 Analysis 3 ~.......................... 62 62 65 An a 1y sis Analysis 4 •••••••••••.•••••••.••.••••.•••••••.•...••.. Discussion •.••••••••••••••••••••••••••••••••••••••••.•.•.••• 65 68 Literature 69 ci ted CHAPTER 4. GREATER PRAIRIE-CHICKEN MIGRATION IN NORTHEASTERN COLORADO ••••••••••••••••••••••••••••••••••••••••••.••••••••• 70 Introduction Methods Results 70 ..•..........................•....................... 72 75 e;......... 75 Timing of movement .••.•..•....•••...•.•.•.•.••........ Distance of movement •••..••..••.••.•... :.............. Site fidelity 79 0.............. 89 Observat ion of movements •..•....•...•.•.••...•......... Di scussi on Literature cited ,. ' 0 • ,. • • • • • • • • • • • • • • • • • • • • • • • • 10................................ APPENDIX A. WALK-IN TRAPS FOR CAPTURING GREATER PRAIRIECHICKENS· IN NORTHEASTERN COLORADO 0 •••••••••••••••••••••••••• Examination of technique •••••.•••••••••••.•.••••••••.•.•.... Literature cited •••••••.••••••••••••••••••••••.•••.••.••••.• APPENDIX B. LEK PERSISTENCE AND ATTENDANCE OF MALE GREATER PRAIRIE-CHICKENS IN NORTHEASTERN .COLORADO ....•••............ Introduction Methods ..•.•..••.....•..•..•.. c...................... 0.0°............................ Results •••••••••••••••••.••••••••••••••••.••••••.•.........• Discussion................................... Literature cited ••.•••••••••••••.••..••••••.•.•••........... 93 • 95 101 105 105 112 114 114 115 117 131 134 �, 47 ACKNOWLEDGMENTS Financial support for this research was generously provided by the Colorado Division of Wildlife through Clait E. Braun. Academic supervision was provided by John T. Ratti and Gary C. White of the Department of Fishery and Wildlife Biology, Colorado State University. I am particularly grateful to Gary C. White for assuming the supervisory role for my project under unusual circumstances. Numerous people helped make this research possible. Rasmussen and Leslie A. Robb assisted with the fieldwork. Mary E. Robert E. Bennetts, Jack W. Bradbury, Cl ait E. Braun, Robert M. 'Gibson, Susan J. Hannon, Jerry W. Hupp, Richard L Knight, David B. Lank, Kathy Martin~ Orrin B. Myers, Leslie A. Robb, Frank A. Pitelka, Joel A. Schmutz, Beatrice Van Horne, Gary C. White, Marilet A. Zablan, and other graduate stUdents provided verbal input at different stages of the research. George Bear, C1ait E. Braun, Jack Dumbacher, Peter o. Dunn, Kenneth M. Giesen, N. Thompson Hobbs, Richard L.'Knight, Hans Landel,. John T. Ratti, Leslie A. Robb, Beatrice Van Horne, and Gary C. White reviewed drafts of proposals and chapters. I am particularly thankful for the continuous and unquestioning emotional support of Leslie A. Robb. ��49 CHAPTER 1. TOLERANCE AMONG FEMALES AND EVOLUTION OF LEKS REVIEW OF HYPOTHESES A common focus of research on mating systems has been the evolution of clumped (mating on leks) from dispersed (mating on territories) polygyny (Emlen and Oring 1977, Bradbury 1981). Clumped polygyny is characterized by the congregation of males at display sites, or leks, for breeding purposes; males typically provide no resources for females (Bradbury 1981, Payne 1984). Although observation of lek behavior in various species has been partly a function of intensity of research, examples have been documented within numerous families of animals (Table I-I). Early research on the evolution of behavior associated with leks included hypotheses of male emancipation 8(Alexander correlations (Snow 1963), display cycles 1974, Otte 1974, Buck and Buck 1978}, predation/habitat (Berger et al. 1963, Koivisto 1965, Boag and Sumanik 1969, Hjorth 1970, Wiley 1974, Wittenberger 1978), flocking (Kruijt et al. 1972), and density (Snow 1973h, Pitelka et al. 1974, Emlen and Oring 1977). These hypotheses generally focused on the importance of male behavior, since male fitness was considered to be more important than female fitness for understanding the ultimate causes for evolution of leks. The difficulty with hypotheses stressing the importance of male fitness has been with explaining tendencies of males to form leks when �I.J C Table 1-1. Descriptions of lek behavior by phylogeny. Phylum Class Order Family Common name Arthropoda Insecta Odonata Anlsoptera . Dragonfly Lepidoptera Papil ionidae Swa11owta il Diptera Gasterophil idae Botfly Drosophilidae Fly Hymenoptera Pompilidae Tarantula hawk wasp Vertebrata Osteichthyes Microcyprlni Poec 11iidae Gila topminnow Amphibia Salientia Continued. Scientific name Citations Plathemis l..mi! Campanella and Wolf 1973 Papilio zelicaon Shields 1967, Sims 1979 Gasterophilys intestinalis Catts 1979 Ringo 1976 Hodosh et ale 1979 Drosophila grimshawi Hemipepsis ustulata Alcock 1981 Orians 1969 Pocciliopsis occidentalis Constantz 1975 �Table 1-1. Continued. Phylum Class Order Family COl1l1lon .name Bufonidae Toad Ranidae Bullfrog Mammalia Chiroptera Pteropodidae .Hal1l1lerheaded bat Emba 11 onuridae Artiodactyla Bovidae Uganda kob Lechwe Topi Aves Scientific name Citation$ Bufo bufo Davies and Halliday 1978 Rana catesbaiana Emlen 1976 Hvpsignathus monstrosus Bradbury 19771, 1977R, Wrangham 1980 Bradbury 1977R, Bradbury and Vehrencamp 1976 Kobus kob Kobus leche Damiliscus korrigum Jarmen 1974 Buechner and Schloeth 1965, Buechner and Roth 1974, Wrangham 1980 Jarmen 1974 Gosling 1987 Oring 1982, Avery 1984, Payne 1984, Loffredo and Borgia 1986, Hoglund 1989 Galliformes Phasianidae Hjorth 1970, Wiley 1974, Wittenberger 1978, De Vos 1979, Sigurjonsdottir 1981 Continued. VI t-' �VI N Table 1-1. Continued. Phylum Class Order Family Common name Scientific name Citations Sage grouse Centrocercys urophasianus Capercalll1e Tetrao uroga11ys Black-billed capercai11ie Black grouse ht.rQ tetrix Wiley 1973, Wrangham 1980, Bradbury and Gibson 1983, Bradbury et al. 1986, 19891, h . Andreev 1979, ,Huller 1979, Moss 1980, Jones 1981 Andreev 1979 Koivisto 1965, Kruijt et a1. 1972, Kruijt and Hogan 1967, De Vos 1979, 1983 Sharp-tat led grouse Greater prairie-chicken Lesser prairie-chicken Great argus pheasant Lumsden 1968, Rippen and Boag 1974 Hamerstrom and Hamerstrom 1960, Robel 1966, Robel and Ballard 1974 Tympanuchus pa11idicinctus Hoffman 1963 .Gilliard 1963, Herton 1975, Davison Argusianus J.r!I!!1 Tympanychys cupjdo 1981 Gruiformes Otididae Great bustard Charadriiformes Sco10pacidae Ruff Pectoral sandpiper Continued. Otis tarda Cramp 1980 Philomachus pugnax Pitelka et a1. 1974, Hyers 1981 Hogan-Warburg 1966, Van Rhijn 1973, Pitelka et al. 1974 . Pite1ka 1959 Calidris melanotos �Table 1-1. Continued. Phylum Class Order Family Common name Buff-breasted sandpiper Painted snipe Great snipe Forest snipe Psittaciformes Psittacidae Kakapo Apodiformes Trochi lidae long-tailed hermit Reddish hermit Anna's hummingbird little hermit Hairy hermit Guy's hermit hummingbird Violet-headed hummingbird Crimson topaz Piciformes Indicatoridae Greater honeyguide lesser honeyguide Continued. Scientific name Citations Tryngites subruficollis Galljnago stenyra Gall inago media Gallinago megala MyerS 1979, Cartar and lyon 1988 Tuck 1972, Sutton 1981 Tuck 1972, Avery and Sherwood 1982, Hoglund 1987, Hoglund and lundberg 1987, Hoglund et al~ 1990 Tuck 1972, Sutton 1981 Strigops habroptilus Forshaw 1978, Merton et al. 1984 Phaethornis superciliosus phaethornis ryber Calypte "anna Phaethornis longuemareus Glaucis hirsuta phaethornis m Klais guimeti Jopaza pella Snow 1973h, Stiles and Wolf 1979 Stiles and Wolf 1979 Nicholson 1931, Snow 19731 Stiles 1973 Snow 1968, Wiley 1971 Snow 1973a Snow 1974~ 1977 Payne 1984 Payne 1984 Indicator indicator Indicator minor Friedmann 1955 Ranger 1955 Ranger 1955 VI W �VI ~ Table 1-1. Continued. Phylum Class Order Family Common name Scaly-throated honeyguide Passeriformes . Cotingidae Red-ruffed fruitcrow Rufous piha Red cot1nga Plum-throated cotinga Pihas Capuchinbird Cock-of-the-rocks Black-and-gold cotinga Umbrellab1rds Guianan cock-of-the-rock Pipridae White-bearded manakin Golden-headed manakin Striped manakin Fiery-capped manakin Wire-tailed manakin Swallow-tailed manakin White-throated manakin White-ruffed manakin Club-winged manakin Continued. Scientific name Citations Indicator variegatys Ranger 1955 Snow 1982 Snow 1982 Pyroderus scutatus Willis and Eisenmann 1979 lipaugys unirufus Snow 1982 phoenicircus carnifex Snow 1982 Cotinga maynana Snow 1982 Lipaugus sp. Snow 1982 perissocephalus tricolor Snow 1982 Rupicola sp. Snow 1982 TUuca atra Snow 1982 CephaJopterus sp. Trail 1985 Rupicola rupicola Sick 1967, Snow 1963 Manacus manacus lill 1974i, h Pipra erythrocephala lill 1976 Machaeropterus regylus Sick 1967 Machaeropterus pyrocephalus Sick 1967 finri fjlicauda Schwartz and Snow 1978 Chiroxiphia caudata Foster 1981 Corapipo guttyralis Snow 1963 Corapipo leucorrhoa Snow 1963 Machaeropterus deliciosus Willis 1966 �Table 1-1. Continued. Phylum Class Order Family Common name Pin-tailed manakin Band-tailed manakin Tyrannidae Ochre-bellied flycatcher McConnell's flycatcher Oxyruncidae Sharpbil1 pycnonotidae Yellow-whiskered greenbu1 Paradisaeidae Goldie's bird of paradise Count Raggi's bird of paradise Lesser bird of paradise Buff-tailed sicklebill Pti10norhynchidae Golden bowerbird Tooth-billed catbird Satin bowerbird Macgregor's bowerbird Ploceidae Jackson's whydah Pin-tailed whydah Scientific. name Citations 11icura mi1itaris fiDrj fasciicauda Mionectes oleagineus Mionectes macconne11i Payne 1984 Robbins 1985 Skutch 1960 Snow and Snow 1979 Willis et a1. 1978 Oxvruncys cristatys Stiles and Whitney 1983 Andropadys Jatirostris paradisaea decora Paradisaea raggjana Paradisaea minor Epimachus albertisi Brossett 1982 Schodde 1976, Cooper and Forshaw 1979 LeCroy 1981, LeCroy et a1. 1980 Payne 1984 Beehler 1983 Beehler 1987 Prionodura newtoni Scenopoeetes dentirostris Ptilonorhvnchus violaceus Amblvornis macgregoriae Gilliard 1969 Cooper and Forshaw 1979 Ve11enga 1970, Borgia et a1. 1985 Pruett-Jones and Pruett-Jones 1982 Eup1ectes jacksoni Vidua macroura Van Someren 1947 Shaw 1984 Vl Vl �56 most males have lower breeding success in a lek system than in a territorial system (Alexander 1974, Kirkpatrick 1982, Avery 1984). Recent work has focused on the importance of female fitness, including the female preference (Bradbury 1981), least costly male (Wrangham 1980), hotspot (Bradbury and Gibson 1983)~ and hotshot (Beehler and Foster 1988) hypotheses. Although each differs with respect to relative importance of male-male competition and unanimity of female choice, they are similar in that female choice drives the development of leks. However, none of these hypotheses provides a clear mechanism explaining the evolutio~ of lek behavior in a species. Emlen and Oring (1977) suggested that lekking tendencies were correlated with a decreasing ability of fem~les to defend their home ranges. Similarly, Bradbury (1981) indicated that- increased female home range size would lead to increased tendencies for males to form leks. However, little justification was given for possible causation in either case. Here I examine current theories on evolution and maintenance of lek behavior. I then propose the female tolerance hypothesis as a mechanism to explain the continuum between dispersed and clumped polygyny. I suggest that decreased intrasexual aggression among females during the breeding season permits increased female choice. The subsequent change in distribution of mating success among males leads to increased encroachment by 'unselected' males on 'selected' males. MALE FITNESS AND EVOLUTION OF LEKS Many hypotheses explaining evolution of lek behavior have suggested that the ultimate advantage to forming leks can be interpreted �57 in terms of male fitness. Hypothetically, whatever mating system males develop. females can adjust to One suggestion is that emancipation of males from parental duties permits development of leks (Snow 1963). However, male emancipation is common among both clumped and dispersed polygyny systems. Hence, while male emancipation may be a prerequisite for lek formation, emancipation may not result in lek behavior. Maximization of effectiveness of male displays has been suggested to lead to lek formation (Alexander 1974, Otte 1974). By concentrating in localized areas, males theoretically optimize the vocal, olfactory, and/or visual potential of their displays, thereby attracting more females. Consequently, clustered males should attract more females/male than single males. on intermittency Furthermore, leks should have an optimum size, based (periodicity) of each male"s display cycle (Bradbury 1977, Buck and Buck 1978). When the latter predictions were exami~ed (Hamerstrom and Hamerstrom 1955, Bradbury 1981, Dunn and Braun 1985), leks were found to be too large for their display potential to be adequately improved (i.e., there were fewer females/male on larger leks than smaller leks). Likewise, few males do most of the mating on leks, indicating that most males do not benefit solely from the lek's potential as a display site. Among the Tetraoninae, species mating on leks frequently occur in open areas, while species mating on dispersed territories occur primarily in forests or arctic/alpine areas (Hjorth 1970, Kruijt et al. 1972, Wiley 1974). surveillance One proposed reason is that cooperative predator is more effective in open country, hence males are more likely to form leks (Berger et ale 1963, Koivisto 1965, Boag and Sumanik �58 1969, Hjorth 1970, Wiley 1974, Wittenberger documentation been shown. 1978). However, little on the effects of predation on groups of various size has Kruijt et ale (1972) suggested that birds in open habitats may form social bonds while feeding in flocks with these bonds subsequently continuing into the breeding season. Unfortunately, these patterns are difficult to predict; forest species display a full range of dispersions, social behaviors, and mating systems (Snow 1963; Sick 1967; Snow 1973, 1974; Bradbury 1977). Snow (1973) suggested that increased male density, without increased resources, could force males to switch strategies from defense of resources to advertisement of breeding potential; lek species should have relatively high densities. While the importance of resou~ce . . abundance and male density on mating systems has been discussed by Emlen and Oring (1977) and Pitelka et ai. (1974), these ideas are not consistent with data on lek species (Bradbury 1981). FEMALE FITNESS AND EVOLUTION OF LEKS Most of the mating success on leks is obtained by relatively few males (Alexander 1974). The difficulty in understanding why males would form leks when most would have lower mating success than on territories is the principal reason for explaining the evolution of lek mating behavior in terms of a process initiated by females. Most current hypotheses that suggest females initiate the evolution of lek mating systems are directly or indirectly based on the premise that increased female home range size leads to increased opportunities for females to select males and/or leks (Bradbury 1981). Bradbury's hypothesis was ·similar to an earlier proposal by Emlen and �59 Oring (1977) which suggested that a necessary prerequisite for evolution of leks was the inability of females to defend their home ranges. Variability in home range size may explain the apparent continuum in level of dispersion among males of polygynous species (such as with 'exploded leks'; Gilliard 1963). Mating systems that are intermediate between clumped and dispersed polygyny may be common (Gilliard 1963, 1969; Lack 1968; Hjorth 1970; Ellison 1971, 1973; Schodde and McKean 1973; Gullion 1976; Payne and Payne 1977; Cooper and Forshaw 1979; Lill 1979; Bradbury 1981). Least Costly Male Hypothesis In the least costly male hypothesis, Wrangham (1980) suggested that females avoid leks close to nesting areas and select leks in nonnesting habitat. By selecting leks relatively far from nesting areas, females potentially reduce predation pressure around their nest sites, a factor that may be important during the nesting season when predation pressure on females is relatively high. The least costly male hypothesis may have less to do with lek evolution than with lek location (Wrangham 1980). However, since habitat may vary between lek and nest sites, a clear effect of nest-lek distance on predation pressure has not been detected. is no evidence that nest sites ~re concentrated lower predation. Likewise, there in specific areas of Since females presumably would make no special effort to reduce predation on their neighbors, males near nest sites of females may have the opportunity to breed with females which are not nesting nearby. Moreover, since most males have little opportunity to breed on leks, they should attempt alternative mating strategies, such as �ou displaying closer to nesting areas, even if they risk rejection by some females in the process. Female Preference Hypothesis Similar to Wrangham (1980), Bradbury (1981) suggested that female choice drives the process of lek evolution. preference' In Bradbury's 'female hypothesis females prefer and select large clusters of males, thus improving their opportunities for mate choice. The most important aspect of this hypothesis is the implied tie between female home range size and tendencies for males to form leks, because it provides insight into why some species may form leks and others may not. Unfortunately, Bradbury did not provide or suggest an adequate test of the hypothesis that increased home range size among females results in an increased tendency of males to form leks. Bradbury (1981) assumed that unanimity of female choice for certain males and/or leks would lead to a characteristic females and leks. distribution of He suggested that males would not congregate in areas . where females had the opportunity to visit other nearby leks. Subsequently, he predicted the distance between the 'active space' of adjacent leks should equal spring home range diameters of females (including the distance at which females can detect males) and that most females should visit only 1 lek during the breeding season. predictions These of female home range size and lek visitation are testable (despite problems with measuring home range size; Beehler and Foster 1988); examinations of published literature have lead to both support (Bradbury 1981) and rejection (Oring 1982) of the hypothesis that females are unanimous in their selection of males and/or leks. �61 Hotspot Hypothesis The hotspot hypothesis was proposed by Bradbury and Gibson (1983) to suggest that leks should form in areas of maximum overlap among home ranges of females (hotspots). This hypothesis suggested that clusters of males should form in areas of high female traffic; areas of female traffic should be influenced by home range.size of females. One difference between the female preference and hotspot hypotheses is that the latter accounts for leks that are more abundant (closer together), and females that may visit more than one lek during a breeding season. In other words, distance between leks could be used to reject the female preference hypothesis, but not the hotspot hypothesis. Bradbury and Gibson (1983) labeled the hotspot hypothesis of lek evolution as a process based on male fitness. distribution force. Nevertheless, size and of home ranges for females is the hypothesized driving For example, even though males theoretically clump at hotspots, these hotspots are determined by location, density, and movement of females. The major problem with the hotspot hypothesis is the difficulty of testing it (Bradbury et al. 1986). Beehler and Foster (1988) noted the circular nature of using female home ranges to determine hotspots, particularly when home ranges may be a function of lek location. predictions While about female lek visitation and lek dispersion may be tested, the results will not clearly support or reject the hotspot hypothesis. Nevertheless, some research has provided support for the hotspot hypothesis (Bradbury et al. 1989Q, Chapter 3). However, in certain species with specific resources associated with their breeding �62 season home ranges, examination of lek locations as a response to experimental manipulations of resources may provide an adequate test. Hotshot Hypothesis The 'hotshot' hypothesis suggests that male~male interactions determine both size and location of leks (Beehler and Foster 1988). Theoretically, a few 'hotshot' males should do most of the mating while other unsuccessful males should display near hotshots, resulting in clusters of males or leks (Foster 1983, Beehler and Foster 1988). The more a particular hotshot is selected by females, the larger his lek should be. Hence, removal of a hotshot should change the attendance of both males and females at a lek. Beehler and Foster's hypothesis is marginally different from the hotspot hypothesis which suggests that subordinate males determine whether to settle alone or join m~les in preferred locations. In the later case, males in preferred locations are assumed to obtain most of the copulations with females in an area. Whether these males are called 'preferred' or 'hotshots' seems irrelevant, as the result appears to be the same. Although Beehler and Foster (1988) treated the hotshot hypothesis as a process of lek evolution ultimately driven by considerations of male fitness (Lewin 1988), they suggested that female choice should lead to an intensification formation of leks. of male-male interactions, a precursor to They also suggested that the female preference and hotspot hypotheses would apply only if females were able to freely choose mates from among males on a lek, a possibility they considered improbable because of the importance of male-male interactions. �63 However, Bradbury (1981) suggested that the female preference hypothesis was dependent on unanimous choice by females for certain males and/or leks, not on unrestricted choice. Presumably, unanimity of choice would not be dependent on lack of male restrictions on female mate choice. In the case of the hotspot hypothesis, neither unrestrict~d or unanimous selection of females appeared to be essential (Bradbury and Gibson 1983). Beehler and Foster (1988) suggested several ways to test and differentiate hypotheses. among the hotshot, hotspot, and female preference First, they suggested that removal of hotshots and subsequent evaluations of lek disruptions and/or changes in female and male lek attendance could be used to differentiate between hotshot and other hypotheses (reduction in attendance predicted by .the hotshot model). This test would be difficult due to seasonal variability both female and male attendance4 in In addition, the predictions are not exclusive to the hotshot hypothesis. Disruptions in lek attendance would not constitute a rejection of either the female preference or hotspot hypotheses. If mate choice by females is important, then reductions in lek attendance (particularly by removal of a hotshot) may be predicted by both female preference and hotspot hypotheses. In addition, a decrease in lek size may cause a predicted decline in lek attendance in the female preference hypothesis. Beehler and Foster suggested a second test in which subordinate males could be removed on some leks, with males on other leks left as controls. If the hotshots remained on leks, female attendance should remain the same, regardless of the total lek size. suggested that this would differentiate Beehler and Foster among the female preference and �04 the hotshot - hotspot hypotheses, since the former predicts that females should start visiting larger leks. However, according to Bradbury (1981) and Bradbury and Gibson (1983) females may only have 1 lek within their home range, thus making additional lek visitation improbable, and at least, unpredictable. A third test proposed by Beehler and Foster is that dominance status be est.imated using more direct measurements, such as male-male interactions and mating disruptions, instead of male morphology (as with Gibson and Bradbury 1985). The hotshot model predicted that mating success should be more closely correlated with within-lek male dominance than with male morphology. Since this test is based on correlations, it does little to tease apart the 3 hypotheses. The final test proposed by Beehler and Foster involves an examination of seasonal and yearly variability in lek size. They suggest the hotshot hypothesis predicts more ('more' is undefined) variation than either the hotspot or female preference hypotheses. of a lek is hypothetically Size related to the length of tenure of the hotshot male(s); hence, variations in lek size would be related to differences in lengths of tenure. Although the hotspot hypothesis does not make a prediction about variability in lek size, the female preference hypothesis predicts that leks will have an optimum size and, therefore, will be less variable (Beehler and Foster 1988). The amount of variability needed to support or reject the hypotheses is difficult to quantify ('more' and 'less' are not defined). Therefore, like the previous tests, this argument does little to separate the 3 hypotheses. �65 Female Tolerance The least- costly male, female preference, hotspot, and hotshot hypotheses all suggest directly or indirectly that increased tendency for males to form leks is a consequence of increased ability by females to select males (Bradbury 1981). The primary differences between the hypotheses are that lek size and distribution are caused by unanimity of female choice for males and/or leks in the least costly male and female preference hypotheses, by maximization of male access to females in the hotspot hypothesis, and by male-male interactions in the hotshot hypothesis. Rarely do the theories present a mechanism which could lead to evolution of leks from a dispersed mating system. Even Bradbury (1981) interpreted the increase in female home range size as a correlative factor with lek formation and gave little justification possible causation. for In the hotspot and hotshot hypotheses, little information is given to explain a possible mechanism for lek evolution. Female home range size during the breeding season has obvious implications for a female's ability to select either leks or males (Bradbury 1981). a corresponding A result of increased female home range size, without decrease in female density, would be an increased probability of encountering conspecifics. Therefore, if territorial females increase the size of their home ranges, the response should be either a decrease in density or a decrease in territoriality Oring 1977). (Emlen and • While large female home ranges during the breeding season would permit increased female choice by permitting females to examine many males, the amount of freedom of female choice for males would be inversely correlated with levels of intrasexual aggression among �66 . females. If females are not able to tolerate other females within close proximity, as on a lek, the resulting female dispersion should create additional mating opportunities aw~ from normal display areas; this would tend to decrease rather than increase the aggregation of males. Hence, lek formation could result from increased sociality or tolerance among females, but not directly from large female home ranges in which females still maintain some conspecific intolerance or territoriality. Large home ranges do not, by themselves, explain why some males give up territories to pursue females into the territory of another male (a likely scenario resulting in territory breakdown). The originally- selected male would undoubtedly chase any intruding males out, as is typical with most territorial animals. from territoriality Therefore, it seems the change to lek mating would come at the point where defense of mates becomes more important than defense of territories, a point which is only reached when more than one female is present within the range of a male. If more than 1 female is with a male at the same time, other males may be able to encroach on the selected male. Likewise, the number of unselected males will be proportional to the unanimity of female selection (Hammerstrin and Parker 1987). The presence of more than 1 female would require female tolerance for other females •. Whether females arrive on the lek together or alone may not matter. The important consideration is that they tolerate each other within the territorial boundaries of the same male; even if the boundaries are on a lek. Relative differences in tolerance among females may explain why leks vary in dispersion from tight clusters to expanded formations ('exploded leks')(Bradbury 1981). Even on large leks (those with many males), evidence suggests that males maintain �67 relatively constant boundaries suggestive of territoriality. Furthermore, males also display site fidelity between years. Tolerance among females may be the key to understanding the evolution of lek mating systems; it may provide a testable alternative to the idea that home range size led to evolution of Iek mating systems. Possible reasons for conspecific tolerance may be found in feeding behavi or, food requi rements (causes for a 1arge range) '"or predation risks (possibly redu~ed by flocking)(Bradbury 1981). Lek mating systems are relatively common among avian species and harem mating systems are relatively common among mammalian species. Both systems are characterized by relatively large groups of females mating with relatively few males. If intrasexual tolerance among females can increase the likelihood of breeding on leks, an additional consideration would be to understand the difference between lek and harem mating systems. Differences in female mobility may explain some of these differences, exceptions to the generalities include antelope (Jarmen 1974) and bats (Bradbury and Vehrencamp 1976) among mammals and bustards (Cramp 1980), bar-headed geese (Anser indicus)(Lamprecht Buhrow 1987), and turkeys (Meleagris gallooavo)(Bent These considerations and 1932) among birds. raise the possibility that leks and harems may be identical with respect to the female tolerance hyp~thesis. PROPOSED TESTS The least costly male, female preference, hotspot, and hotshot hypotheses appear to pertain more to location and maintenance than to the evolution of leks. of leks Hence, I suggest that only 2 hypotheses explaining the evolution of lek behavior can be tested: 1) Bradbury's �(1981) model that increased female home range size leads to an increased tendency to mate on leks; and 2) the female tolerance hypothesis (breakdown in territory version proposed by Emlen and Dring [1977]). Test 1: Resources could be manipulated in species (preferably a species with behavioral plasticity) that mate either on leks or on territories (change dispersion of resources). A decrease in congregations of males in response to a decrease in home range size of females, or an increase in male congregations in response to an increase in home range size of females, would support Bradbury's (1981) hypothesis and the female tolerance hypotheses. If increased female home range size leads to increased tolerance of conspecifics, the predictions of Bradbury's (1981) home range model and the female tolerance hypothesis are not m~tually exclusive. Test 2: manipulated ~evels of aggression in females could be artificially (hormone administration, Wingfield [1984]) and the subsequent pattern of male and female spacing analyzed. If female aggression could be artificially increased, males should respond by increasing their dispersion and/or decreasing their tendency to display on leks. effect. A decrease in female aggression should have the opposite If home range size remained constant, this test would be used for rejection of Bradbury's (1981) hypothesis and support of the female tolerance hypothesis. LITERATURE CITED . AlcoCK, J. 1981. Lek territoriality in the tarantula hawk wasp Hemipepsis ustulata (Hymenoptera: Pompilidae). Behav. Ecol. Sociobiol.8:309-317. Alexander, R. D. 1974. The evolution of social behavior. Ecol. Syst. 5:325-383. Annu. Rev. �69 Andreev, A. V. 1979. Reproductive behaviour in black-billed capercaillie compared to capercaillie. Pages 134-139 in T. W. I. Lovel, ed. Behaviour of woodland grouse, woodland grouse symposium. World Pheasant Assn., Bures, Suffolk, U.K. Avery, M. I. 1984. Lekking in birds: choice, competition and reproductive constraints. Ibis 126:177-187. _____ , and G. Sherwood .. 1982. The lekking behaviour of great snipe. Ornis Scand. 13:72-78. Beehler, B. M. 1983. 100:992-995. Lek behavior of the lesser bird of paradise. Auk _____ . 1987. Ecology and behavior of the buff-tailed sicklebill. 104:48-55. Auk _____ , and M. S. Foster. 1988. Hotshots, hotspots, and female preference in the organization of lek mating systems. Am. Nat. 131:203-219. Bent, A. C. 1932. Life histories of North American gallinaceous birds. Dover Publications, Inc., New York, N.Y. 490pp. Berger, D. D., F. Hamerstrom, and F. N. Hamerstrom, Jr. 1963. The effect of raptors on prairie chickens on the booming grounds. 91:75-86. Auk Boag, D. A., and K. M. Sumanik. 1969. Characteristics of drumming sites selected by ruffed grouse in Alberta. J. Wildl. Manage. 33:621-628. Borgia, G., S. G. Pruett-Jones, and M. A. Pruett-Jones. 1985. The evolution of bower-building and the assessment of male quality. Z. Tierpsychol. 67:225-236. Bradbury, J. W. 1977~. Lek mating behavior in the hammerheaded bat. Z. Tierpsychol. 45:225-255. ______ • 1977h. Social organization and communication. Pages 1-72 in W. Wimsatt, ed. The biology of bats, Vol. 3. Academic Press, New York, N.Y. _____ . 1981. The evolution of leks. Pages 138-169 in R. D. Alexander and D. W. Tinkle, eds. Natural selection and social behavior: recent research and new theory. Chiron Press, New York, N.Y. ______ , and R. M. Gibson. 1983. Leks and mate choice. Pages 109-138 in P. Bateson, ed. Mate choice. Cambridge Univ. Press, Cambridge, U.K. _____ , and S. L. Vehrencamp. 1976. Social organization and foraging in emballonurid bats. II. A model for the determination of group size. Behav. Ecol. Sociobiol. 1:383-404. �70 _______, R. M. Gibson, and I. M. Tsai. 1986. Hotspots and the dispersion of leks. Anim. Behav. 34:1694-1709. _______, S. L. Vehrencamp, and R. M. Gibson. 1989~. Dispersion of displaying male sage grouse. I. Patterns of temporal variation. Behav. Eco1. Sociobiol. 24:1-14. ______ , R. M. Gibson, C. E. McCarthy, and S. L. Vehrencamp. 1989Q. Dispersion of displaying male sage grouse. II. The role of female dispersion. Behav. Ecol. Sociobio1. 24:15-24. Brosset, A. 1982. The social life of the African forest yel10wwhiskered greenbu1 Andropadus latirostris. Z. Tierpsycho1. 60:239-255. Buck, J., and E. Buck. 1978. Toward a functional interpretation of synchronous flashing by fireflies. Am. Nat. 112:471-492. Buechner, H. K., and H. D. Roth. 1974. antelope. Am. Zool. 14:145-162. The lek system in Uganda kob _______, and R. Sch1oeth. 1965. Ceremonial mating behavior in Uganda kob (Adenota kob thomasi). Z. Tierphysio1. 22:209-225. Campanella, P. J., and L. L. Wolf. 1973. Temporal leks as a mating system in a temperate zone dragonfly (Odonata:Anisoptera). I: P1athemis lydia (Drury). Behaviour 60:49-87. Cartar, R. V., and B. E. Lyon. 1988. The mating system of the buffbreasted sandpiper: lekking and resource defense polygyny. Ornis Scand. 19:74-76. Catts, E. P. 1979. Hilltop aggregation and mating behavior by Gasterophilus intestina1is (Diptera: Cuterebridae)~ J. Med. Entomo1. 16:461-646. Constantz, G. D. 1975. Behavioral ecology of mating in the male Gila topminnow, Pocci1iopsis occidenta1is. Ecology 56:966-973. Cooper, W. T., and J. M. Forshaw. 1979. The birds of paradise and bowerbirds. Collins, Sydney. 412pp. Cramp, S., ed. 1980. Handbook of the birds of Europe, the Middle East and North Africa. The birds of the western Palearctic, Vol. 2. Oxford Univ. Press, Oxford, U.K. 695pp. Davies, N. B., and T. R. Halliday. 1978. Deep croaks and fighting assessment in toads, Bufo bufo. Nature 274:683-685. Davison, G. W. H. 1981. Sexual selection and the mating system of Argusianus argus (Aves:Phasianidae). Bio1. J. Linn. Soc. 15:91104. �71 0 De Vos, G. J. 1979. Adaptedness of arena behaviour in black grouse (Tetrao tetrix) and other grouse species (Tetraonidae). Behaviour 68:277-314. ______ . 1983. Social behaviour of black grouse; an observational experimental field study. Ardea 71:1-103. and Dunn, P.O., and C. E. Braun. 1985. Natal dispersal and lek fidelity of sage grouse. Auk 102:621-627. Ellison, L. N. 1971. 88:652-664. Territoriality in Alaskan spruce grouse. Auk _____ . 1973. Seasonal social organization and movements of spruce grouse. Condor 75:375-385. Emlen, S. T. 1976. Lek organization and mating strategies in the bullfrog. Behav. Ecol. Sociobiol. 1:283-313. _____ , and L. W. Oring. 1977. Ecology, sexual selection, and the evolution of mating systems. Science 197:215-223. Forshaw, J. H. 1978. Parrots of the world. Abbot, U.K. 616pp. David and Charles, Newton Foster, H. S. 1981. Cooperative behavior and social organization of the swallow-tailed manakin (Chiroxiphia caudata). Behav. Ecol. Sociobiol. 9:167-177. _____ 0. 1983. Disruption, dispersion, and dominance in lek-breeding birds. Am. Nat. 122:53-72. Friedmann, H. 292pp. 1955. The honey-guides. U. S. Natl. Hus. Bull. 208. Gibson, R. H., and J. W. Bradbury. 1985. Sexual selection in lekking sage grouse: phenotypic correlates of male mating success. Behav. Ecol. Sociobiol. 18:117-123. Gilliard, E. T. 46. _____ 1963. The evolution of bowerbirds. 1969. Birds of paradise and bower birds. Nicholson, London, U.K. 485pp. 0. Sci. Am. 209:38Weidenfeld and Gosling, L. H. 1987. Lekking in ~opi: a high cost, specialist strategy. Anim. Behav. 35:616-618. Gullion, G. W. 1976. Reevaluation of "activity clustering" by male grouse. Auk 93:192-193. Hamerstrom, F. N., Jr., and F. Hamerstrom. 1955. Population density and behavior in Wisconsin prairie chickens (Tympanuchus cupido pinnatus). Trans. Int. Ornith. Congr. 11:459-466. �1'1. ___ , and _-::-_. 1960. Comparability of some social displays of grouse. Proc. Int. Ornith. Congr. 12:274-293. Hammerstein, P., and G. A. Parker. 1987. Sexual selection: games between the sexes. Pages 119-142 in J. W. Bradbury and M. B. Andersson, eds. Sexual selection: testing the alternatives. John Wiley and Sons, Chichester, U.K. Hjorth, I. 1970. 7:184-"596. Reproductive behavior in Tetraonidae. Viltrevy Hodosh, R. J., J. M. Ringo, and F. T. McAndrew. 1979. Density and lek displays in Drosophila grimshawi (Diptera:Drosophilidae). Z. Tierpsychol.49:164-172. Hoffman, D. M. 1963. The lesser prairie chicken in Colorado. Wildl. Manage. 27:726-732. J. Hogan-Warburg, A. J. 1966. Social behavior of the ruff, Philomachus Dugnax (L.). Ardea 54:109-229. Hoglund, J. 1987. Sexual selection in a monomorphic lek-breeding bird: correlates of male mating success in the great snipe Gallinago media. Behav. Ecol. Sociobiol. 21:211-216. ___ . 1989. Size and plumage dimorphism in lek-breeding birds: ~ comparative analysis. Am. Nat. 134:72-87. _____ , and A. Lundberg. 1987. Sexual selection in a monomorphic lekbreeding bird: correlates of male mating success in the great snipe Gallinago media. Behav. Ecol. Sociobiol. 21:211-216. _____ , J. A. Kalas, and L. Lofaldli. 1990. Sexual dimorphism in the lekking great snipe. Ornis Scand. 21:1-6. Jarmen, P. J. 1974. their ecology. .Jones, A. M. 1981. Ibis 123:579. The social organization of antelope in relation to Behaviour 48:215-267 • A field study of capercaillie Tetrao urogallus. Kirkpatrick, M. 1982. Sexual selection and the evolution of female choice. Evolution 36:1-12. Koivisto, I. 1966. Behavior of the black grouse, lyrurus tetrix, during the spring ~isplay. Finn. Game Res. 26. 60pp. Kruijt, J. P., and J. A. Hogan. 1967. Social behavior on the lek in black grouse, lyrurus tetrix tetrix (l.). Ardea 55:203-240. ___ , G. J. De Vos, and 1. Bossema. 1972. The arena system of black grouse. Proc. Int. Ornith. Congr. 15:399-423. �73 lack, D. 1968. Ecological adaptations for breeding in birds. and Co., ltd., london. 409pp. lamprecht, J., and H. Buhrow. 1987. geese. Ardea 75:285-292. Methuen Harem polygyny in bar-headed leCroy, M. K. 1981. The genus Paradisaea: display and evolution. Mus. Novit. 2714. 52pp. ___ Am. , A. Kulupi, and W. S. Peckover. 1980. Goldie's bird of paradise: display, natural history and traditional relationships of people to the bird. Wilson Bull. 92:289-301. lewin, R. 1988. Hotshots, hotspots, and female preference. 240:1277-1278. Science . lill, A. 1974A. Sexual behavior of the lek-forming white-bearded manakin (M. manacus trinitatis Hartert). Z. Tierpsychol. 36:1-36. _____ . 1974h. Social organization and space utilization in the lekforming white-bearded manakin, M. manacus trinitatis (Hartert). Z. Tierpsychol. 36:513-530. ____ . 1976. lek behavior in the golden-headed manakin, Pipra erythroceohala in Trinidad (West Indies). Z. Tierpsychol. Supple 18:1-84. . loffredo, C. A., and·G. Borgia. 1986. Sexual selection, mating systems, and the evolution of avian acoustical displays. Am. Nat. 128:773-794. lumsden, H. G. 1968. Displays of the sharp-tailed grouse. Dep. lands and Forests Res. Rep. 83. 68pp. Merton, D. V. 1975. Serv. 6:39-51. Kakapo. Wildlife - a review. Ontario New Zealand Wildl. _____ , R. V. Morris, and I. A. E. Atkinson. 1984. lek behaviour in a parrot: the kakapo, Strigops habroptilus, of New Zealand. Ibis 126:277-283. Moss, R. 1980. 447. Why are capercaillie cocks so big? Br. Birds 73:440- Muller, F. 1979. A IS-year study of a capercaillie lek in the western Rhon-Mountains (W. Germany). Pages 120-130 in T. W. I. Lovel, ed. Behaviour of woodland grouse, woodland grouse symposium. World Pheasant Assn., Bures, Suffolk, U.K. Myers, J. P. 1979. 33:823-825. ____ Leks, sex, and buff-breasted sandpipers. Am. Birds . 1981. .Cross-seasonal interactions in the evolution of sandpiper social systems. Behav. Ecol. Sociobiol. 8:195-202. �74 Nicholson, E. M. 83. 1931. Communal display in hummingbirds. Ibis 13:74- Orians, G. H. 1969. On the evolution of mating systems 1n birds and mammals. Am. Nat. 103:589-603. Oring, l. W. 1982. Avian mating systems. Pages 1-92 in D. Farner, J. King, and K. Parkes, eds. Avian biology. Vol. VI. Academic Press, New York,_ N.Y. Otte, D. 1974. Effects and functions in the evolution of signalling systems. Annu. Rev. Ecol. Syst. 5:385-416. Payne, R. B. 1984. Sexual selection, lek and arena behavior and sexual size dimorphism in birds. Ornithol. Monogr. 33. 52pp. _______, and K. Payne. 1977. Social organization and mating success in local song populations of village indigobirds, Vidua chalybeata. Z. Tierpsychol. 45:113-173. Pitelka, F. A. 1959. Numbers, breeding schedule, and territoriality in pectoral sandpipers of northern Alaska. Condor 61:233-264. ________ , R. T. Holmes, and S. F. Maclean. 1974. Ecology and evolution of social organization in arctic sandpipers. Am. Zool. 14:185204. Pruett-Jones, M., and S. G. Pruett-Jones. 1982. Spacing and distribution of bowers in Macgregor's bowerbird (Amblyornis macgregoriae). Behav. Ecol. Sociobiol. 11:25-3Z. Ranger, G. A. 1955. On three species of honeyguide; the greater (Indicator indicator), the lesser (Indicator minor) and the scalythroated (Indicator variegatus). Ostrich 26:70-87. Ringo, J. M. 1976. A communal display in Hawaiian Drosophila (Diptera: Drosophilidae). Annu. Entomol. Soc. Am. 69:209-214. Rippin, A. B., and D. A. Boag. 1974. Recruitment to populations of male sharp-tailed grouse. J. Wildl. Manage. 38:616-621. Robbins, M. B. 1985. Social organization ~f the band-tailed manakin (PiDra fascficauda). Condor 87:449-456. Robel, R. J. 1966. Booming territory size and mating success of the greater prairie chicken ·(Tympanuchus cupido Dinnatus). Anim. Behav. 14:328-331. ______ , and W. B. Ballard. 1974. Lek social organization and reproductive success in the greater prairie chicken. Am. Zool. 14:121-128. Schodde, R. 1976. resynthesis. Evolution in the birds-of-paradise and bowerbirds, a Proc. Int. Ornithol. Congr. 16:137-149. �75 ______ , and J. L. McKean. 1973. The species of the genus Parotia (paradiseidae) and their relationships. Emu 73:145-156. Schwartz, P., and D. W. Snow. 1978. Display and related behavior of the wire-tailed manakin. Living Bird 17:51-78. Shaw, P. 1984. The social behavior of the pin-tailed manakin (Pipra fascilicauda). Wilson Bull. 95:321-504. Shields, O. 1967. Hilltopping. J. Res. Lepid. 6:69-178. Sick, H. 1967. Courtship behavior in the manakins (Pipridae): a review. Living Bird 6:5-22. Sigurjonsdottir, H. 1981. The evolution of sexual size dimorphism in gamebirds, waterfowl and raptors. Ornis Scand. 12:249-260. Sims, S. R. 1979. Aspects of mating frequency and reproductive maturity in Papilio zelicaon. Am. Midl. Nat. 102:36-50. Skutch, A. F. 1960. Life histories of Central American birds, Pt. 2. Pacific Coast Avifauna 34. Snow, B. K. 1973i. Social organization of the hairy hermit Glaucis hirsuta. Ardea 61:94-105. ______ . 1973h. The behavior and ecology of hermit hummingbirds in the Kanuku Mountains, Guyana. Wilson Bull. 85:163-177. ______ . 1974. Lek behavior and breeding of Guy's hermit hummingbird Phaethornjs gyy. Ibis 116:278-297. _______. 1977. Comparison of the leks of Guy's hermit hummingbird Phaethornis gyy in Costa Rica and Trinidad. Ibis 119:211-214. _______, and D. W. Snow. 1979. The ochre-bellied flycatcher and the evolution of lek behavior. Condor 81:286-292. Snow, D. W. 1963. The evolution of manakin displays. Ornith. Congr. 13:553-561. Proc. Int. ______ . 1968. 7:47-55. The singing assemblies of little hermits. _______. 1982. 203pp. The cotingas. Living Bird British Museum (Nat. Hist.), London, U.K. Stiles, F. G. 1973. Food supply and the annual cycle of the Anna hummingbird. Univ. Calif. Publ. Zool. 97. 109pp. _____ , and l. l. Wolf. 1979. Ecology and evolution of lek mating behavior in the long-tailed hermit hummingbird. Ornithol. Monogr. 27. 78pp. �10 _______, and B. Whitney. 1983. Notes on the behavior of the Costa Rican sharpbill (Oxyruncus cristatys frater). Auk 100:117-125. Sutton, G. M. snipes. 1981. On aerial and ground displays of the world's Wilson Bull. 93:457-477. Trail, P. 1985. Courtship disruption modifies mate choice in a lekbreeding bird. Science 227:778-780. Tuck, L. M. 1972. The snipes: a study of the genus Capella. Wildl. Servo Monogr. 5. 429pp. Can. Van Rhijn, J. G. 1973. Behavioural dimorphism in male ruffs, Philomachus Dugnax (L.). Behaviour 47:153-229. Van Someren, V. D. 1947. The dancing display and courtship of Jackson's whydah (Coliuspasser jackson; Sharpe). E. Afr. Nat. Hist. Soc. J. 18:131-141. Vellenga, R. E. 1970. Behaviour of the male satin bowerbird at the bower. Austr. Bird Bander 8:3-11. Wiley, R. H. 1971. Song groups in a singing assembly of little hermits. Condor 72:28-35. ______ • 1973. Territoriality and non-random mating in sage grouse, Centrocercus yrophasianys. Anim. Behav. Monogr. 6:85-169. _____ . 1974. Evolution of social organization and 1ife history patterns among grouse (Aves:Tetraonidae). Q. Rev. Biol. 49:201227. Willis, E. O. 1966. Notes on a display and nest of the club-winged manakin. Auk 83:475-476. ______ , and E. Eisenmann. 1979. A revised 1ist of birds of Barro Colorado Island, Panama. Smithson. Contrib. Zool. 291. 31pp. ______ , D. Wechsler, and Y. Oniki. 1978. On behavior and nesting of McConnell's flycatcher (Pipromorpha macconnelli): does female· rejection lead to male promiscuity? Auk 95:1-8. Wingfield, J. C. 1984. Androgens and mating systems: testosteroneinduced polygyny in normally monogamous birds. Auk 101:665-671. Wittenberger, J. F. 1978. Condor 80:126-137. The evolution of mating systems in grouse. Wrangham, R. W. 1980. Female choice of least costly males; a possible factor in the evolution of leks. Z. Tierpsychol. 54:357-367. �77 CHAPTER 2. MOVEMENT AND LEK VISITATION BY FEMALE GREATER PRAIRIECHICKENS: A TEST OF THE FEMALE PREFERENCE HYPOTHESIS OF LEK EVOLUTION INTRODUCTION Current hypotheses explaining evolution of lek behavior from territoriality, its hypothesized precursor, suggest that females initiate the process of lek evolution (Emlen and Dring 1977, Wrangham 1980, Bradbury 1981, Bradbury and Gibson 1983, Beehler and Foster 1988). Increased home range size theoretically leads to an increased ability to select males and/or leks. Increased female home range size may also explain the continuum between lek and territorial species (Bradbury 1981). Whether large home ranges are caused by clumped food sources, predation, or other unexplained factors is unresolved; Bradbury (1981) suggested that female choice drives lek evolution. In Bradbury's 'female preference' hypothesis, females prefer and select large clusters of males, thus improving their opportunities choice. for mate Bradbury assumed that unanimity of female choice for certain leks would lead to a characteristic distribution of females and leks; females should have spring home range diameters equal to inter-lek distances and most females should visit only 1 lek during the breeding season. In this paper I will test the female preference hypothesis by examining lek visitation and home range size of female greater prairiechickens. If most females visit more than 1 lek or have home ranges �78 that include more than 1 lek, the female preference hypothesis can be rejected. METHODS Greater prairie-chickens were studied in 1986-88 on a 301-km2 study area centered (40 11' N, 102 22' W) 10 km northeast of Eckl ey, 0 Colorado (Fig. 2-1). 0 The area consisted of grassland, sand sagebrush (Artemisia filifolia), and small soapweed (Yucca glauca) intermixed with irrigated fields of agriculture, primarily corn. Locations of leks (2 or more displaying males) and maximum male attendance at each lek each year were based on at least 2 observations. Trapping was concentrated on a core area of approximately 100 km2• Female greater prairie-chickens were captured at winter feeding sites using walk-in traps baited with corn and on leks using walk-fn traps and cannon nets. Captured females were banded with a numbered aluminum band and a unique combination of 3 colored plastic bands, and fitted with battery- or solar-powered markers (Amstrup 1980). each birds' body weight. radio transmitters attached to poncho-type Radio weights ranged between 1.8 and 2.3% of Birds were classed as yearlings (first breeding year) or adults (> first breeding year) (Ammann 1944). Observations of birds were made during early (IS Feb-31 Mar) and late spring (1 Apr-IS May); designation of seasons was based on aspects of breeding behavior and movement (Robel et ale 1970). Radio-marked greater prairie-chickens were located using a portable receiver and 3-element Yagi antenna. Daily observations for each bird were obtained by triangulation; transmitters 3 or more azimuths were obtained ~ 1.5 km of target and at angles-of-incidence greater than 35 and less than 0 �79 MEAN YEARLY LEK ATTENDANCE 0·1 >1·3 >3·6 • o 0 4460 e- : , ,:: :. z ::! 0 : :: 0 0 : : :....: ~o' ~ ~ 0 ,: 1 o - BAITSITES >15 .. : "', ~ 0 .. ~ ~ ~ 0 : .! 4450 -; 5 LEKS * • ~ ~;--·r::-~i··;J:=:~*....... !..... t;: r,' , ~ >10·15 000 J~7····~-··r~~~""":~~&"c,O<~. .. N : [4455 >6·10 CAPTURE SITES ~ + '~O· I0 , l 0 . I: ,~. I ""f": : :'" :' • ~: ~L·········f··········.•• ···························I·.o············~:.:·~··················· .~ ~ r::::.· ~ • i.: : ~ ·.·.··.. I '''''.. : ~ <e-.......•.......~ : : '- - ~ -, ~ 0 : ~ o?J o· 0 4445 r-············t······...•·~·~···'························l································~·········t··········· (d Q ' .. I0 "", : b 0 : ~ r-- 0 EC~EY ".. ...:0 1 1 1 : : i ~..--+----------i----------1------J 720 725 !.. o - 0: 715 , , 730 , ,,' 735 UTM (KM EAST) Fig. 2-1. Locations of leks and capture sites on greater prairie- chicken study area in 1986-1988 (301 km2 outlined by dashed boundary) in northeastern Colorado. �80 145°. All locations were recorded using Universal Transverse Mercator coordinates (nearest 10-m interval). Home range size was estimated as the area within a 75% probability contour generated with harmonic means (grid size of 25 X 25) for each radio-marked female for each season (Dixon and Chapman 1980). The accuracy of the triangulation technique was examined with 13 test transmitters placed on the ground (by another observer) in a relatively hilly 4-~ area, 1-2 km from the nearest access points. The area was chosen because it was inaccessible relative to the majority of the study area. Five azimuths were obtained for each transmitter from each of 5 locations (4 locations for 3 transmitters). The standard .deviation between observed and average azimuths was 2.10° (n - 310). In contrast, estimated azimuths deviated an average of 6.97° from actual azimuths (the overall bias was -1.66°). Most of the variation· apparently was due to the location where the azimuth was obtained and the transmitter that was being located. Estimated locations were compared with actual locations of nests, dead birds, and test transmitters (Fig. 2-2). About 90% of the locations derived by triangulation were within 250 m of actual locations. Although this sample was not affected by problems associated with bird movement, other factors may have adversely affected accuracy; transmitters usually were close to, or on, the ground. Transmitters from dead birds frequently were buried, upside down, and/or damaged (chewed); transmitters on nesting fem~les were usually in dense vegetation, close to the ground, and shielded by the female. In the female preference hypothesis, Bradbury (1981) specifically predicted that: 1) the diameter of female home ranges should be less �81 2S,_----~~------------------------------------------, [] TEST TRANSMITTERS E1 SITES OF NESTS &1 SITES OF DEAD BIRDS 20 a: 1S w c::l :::E :::::3 Z 10 S o 0·25 50·75 100·125 150·175200·225 250·275 300-325350-375 400·425 450·475500·525 DISTANCE BElWEEN Fig. 2-2. ACTUAL AND ESTIMATED LOCATIONS (M) Distribution of distances between estimated and actual locations of 13 test transmitters, prairie-chickens 79 nests, and 44 dead greater in northeastern Colorado, 1986-88. �than inter-1ek distances, 2) most females should have only 1 1ek within their home range, and 3) each 1ek should have an exclusive population of females. Thus, most females should visit' only 1 1ek. was not defined by Bradbury, I conservatively Although 'most' defined most as > 50%. The validity of the specific predictions were examined by Beehler and Foster (1988). I examined Bradbury's predictions for the female preference hypothesis by identifying lek - nest distances (an indirect measure of home range size) and 1ek visitation by female greater prairie-chickens. Home range size for females was estimated for both early (prior to most lek visitation) and late spring (during lek visitation and nest establishment). probability Home range size was estimated as the area within a 75% contour generated with harmonic means (grid size of 25 X . 25); this was a conservative estimate of home range size (Dixon and Chapman 1980). Additionally, home range diameter was calculated as though each home range was a circle (conservative estimate of diameter). RESULTS Radios were placed on 36 adult and 56 yearling female greater prairie-chickens during 1986-88 breeding seasons. Since spring capture techniques potentially disturb females, data from an 'undisturbed' sample of 12 adults and 7 yearlings captured at feeding sites during winter 1987-88 were examined as a 'control'. Sixty-five different leks were located on the study .area during 1986-88 (Fig. 2-3). All leks were not active from year-to-year; and 47 were located during each breeding season, respectively. 41, 42, The mean distance between a lek and its nearest neighbor was 1.31 ± 0.56 (~ ± �83 N SIZE OF LEKS . 0 0 ·0 0 00 2-5 MALES 6-10 MALES 0 .• 0 0 0 0 • 0 o 0 11-15 MALES >16 MALES 0 0 <e 0 0 0 0 2 4KM , • I , , 0 0 0 '0. 0 0° • o· ° "'0 0 ! 0 0 °0 00 ~a 0 0 0 0 cc 0 0 0° . e • 0 0 0 0 0 Fig. 2-3. • Distribution and size of greater prairie-chicken northeastern Colorado, 1986-88. leks in �u"t SO), 1.20 ± 0.70, and 1.18 ± 0.62 km each year, respectively (medians of 1.42, 1.32, 1.15 km). Home Range Size Distances between nest locations and both the nearest 1ek and the 1ek where each female was first observed were determined separately for Females (n - 89) nested an average of 3.62 km each year (Fig. 2-4). from leks where they were first observed (Table 2-1); some females moved relatively long distances (Fig. 2-5). Hens (n s 81) nested an average of 1.00 km from the nearest 1ek (Table 2-1) with distances ranging from 0.23 to 2.39 km (Fig. 2-6). Eight of 89 females nested off the study area; hence, distances to the nearest 1ek could not be determined. There was no difference in distance between the 1ek where each female was first observed and her nest that could be attributed to year (f - 1.35, f - 0.265). However~ yearling females tended to move further between the 1ek where first observed and their nest site than adults (f - 3.89, f - 0.052); the difference may be partially attributable to yearling dispersal. distances For instance, the 5 females with the longest (18-29 km) between nests and the 1ek where first observed or captured, were yearlings. In addition, 12 of 16 movements> 5 km were f - 0.611) or by yearlings. There was no difference by year (f - 0.50, age (f - 0.02, f - 0.901) for nest-nearest 1ek distance (Table 2-1). Disturbance may also have been a factor. Females captured during the breeding season (disturbed) had longer distances between 1ek first visited and nest site (1 = 2.352, f - 0.022) than females captured prior to the breeding season (undisturbed). Undisturbed females may have been marked prior to normal movements between seasonal home ranges. �85 1986· N o LEKS * NESTS 1987 1988 o Fig. 2-4. Distribution of female greater prairie-chicken 2 4 KM nests in relation to lek where each female was first observed or captured in northeastern Colorado, 1986-88. �00 Table 2-1. Distances (m) between greater prairie-chicken nests and leks in northeastern Colorado, 1986-88. Lek nearest to nest Lek visited first Category 11 so Median 11 so Median Year 1986 22 2.09 2.86 2.11 21 0.84 0.96 0.53 1987 31 1.65 2.91 4.34 29 0.78 0.94 0.53 1988 36 1.86 4.71 6.84 31 1.14 1.08 0.47 Adult 42 1.79 2.28 2.25 38 0.97 0.98 0.51 Yearling Status.1I 47 2.04 4.82 6.62 43 0.96 1.01 0.52 Undisturbed 14 1.87 2.01 1.87 14 1.05 1.16 0.51 Disturbed 75 1.98 3.92 5.55 67 0.84 0.96 0.50 89 1.93 3.62 5.19 81 0.96 1.00 0.50 Age- Totals -One female of unknown-age was excluded from the sample for nearest lek - nest distances. IIFemalescaptured prior to lek visitation were considered undisturbed and those captured on leks during lek visitation were considered disturbed. �87 ~~----------------------------------------------, YEARLINGS 15 • en ADULTS w ..J ~10 W LL 5 o 5 10 15 ~ 25 30 DISTANCE (KM) Fig. 2-5. Distribution of distances between 89 female greater prairie- chicken nest sites and leks where each female was first observed or captured in northeastern Colorado, 1986-88. �88 10 YEARLINGS 8 ffi _J II ADULTS 8 < ~ W u.. 4 2 o 0.0 0.5 1.0 1.5 2.0 2.5 DISTANCE (KM) Fig. 2-6. Distribution of distances between 82 female greater prairie- chicken nest sites and nearest lek in northeastern Colorado, 1986-88. �89 Consequently, 4 undisturbed females moved 6-18 km prior to being observed on leks during the breeding season. In contrast, all disturbed females were observed on leks, often prior to making similar long movements. Hence, distances between lek first visited and nest site were biased toward shorter distances for undisturbed females. There was no difference for nest-nearest lek distance attributable to disturbance (1 - 1.660, f - 0.113)(Table 2-1). Sixty-six of 89 females (74%) nested closer to a lek different from the lek where they were first observed or captured. When compared with a maximum predicted value of 50%, the observed value was higher (~ - 20.775, f < 0.001). Disturbed females (76.0%, n • 75), did not differ from undisturbed females (64.3%; n - 14)(~ • 0.845, f - 0.431). If disturbance was a factor, disturbed females would be expected to avoid leks where they were captured. There were no differences in home range size associated with sex, age, or disturbance status (f > 0.10)(Table 2-2). However, home range size during late spring (X • 624 hal was larger (1 - 2.098, f than in early spring (X - 213 ha)(Table 2-2). a 0.038) The average diameter for home ranges was 1.51 km (SO - 0.66) during early spring and 2.24 km (SO - 1.71) in late spring. These estimates are greater than distances between neighboring leks (mean yearly range of 1.18-1.31, median range of 1.15-1.42). Lek Visitation Seventy-nine radio-marked females were observed on leks at least twice during the breeding season; 84~8% of these females were observed on more than 1 lek (Fig. 2-7), which was greater (~ • 38.291, f < �90 Table 2-2. Home range size (ha) for female greater prairie-chickens in northeastern Colorado, 1986-88. Early spring Category n Median Late spring ! SO n Median ! SO Year 1986 3 165.0 149.3 115.6 21 252.5 526.9 749.8 1987 2 328.7 328.7 247.9 36 264.4 349.4 287.8 1988 25 156.8 210.9 252.7 40 310.9 921.6 2861.1 Adult 22 165.0 175.9 97.2 49 266.0 369.2 304.7 8 134.9 313.6 439.2 48 271.0 883.7 2645.6 24 160.8 231.3 260.1 19 461.3 625.1 959.1 6 113.1 137.7 111.6 78 260.4 623.5 2049.0 30 160.8 212.6 239.3 97 266.0 623.8 1881.5 Age Yearling Status· Undisturbed Disturbed Totals ·Females captured prior to lek visitation were considered undisturbed and those captured on leks during lek visitation were considered disturbed. �91 - 0=15 ~ ~ ~ ~ ~ ,~ ,~ ~ --b-· OISTURBEO ~ ~ 0=22,, 0=34 Z CiS 5 90 ... . 0-79 -- - .. fo .... . 85 8..... W LL. ,, 0=56 •• •• "Z:1 ••••••••••• 'E:l .•. ........ .0" ,, , , --8' : , .Q A .•.• , , , ,, I A J= : I ····8···· 95 n=6 ,~A-------- TOTAL 0 UNOISTURBEO 100 .0. . o 80 2 345 6 MINIMUM NUMBER OF LEK VISITS Fig. 2-7. Proportion of female greater prairie-chickens visiting more than 1 lek in relation to number of times observed on a lek in northeastern Colorado, 1986-88. 7 �92 0.001) than the hypothesized upper-limit of 50%. Undisturbed females (88.2~, n • 17) did not differ from disturbed females (83.9%, n - 62) in their likelihood to visit more than 1 lek ()( • 0.197, f - 0.741). If disturbance was a factor, disturbed females should have visited more leks than undisturbed females. As number of visits to leks increased, number of visits to different leks also increased (Fig. 2-8). One yearling female was observed on 6 different leks in 8 1ek visits. Another yearling female visited the same lek on 9 consecutive observations. Capture as a possible disturbance factor was examined by compiling data on consecutive lek visits (Fig. 2-9). If females were adversely affected by being captured they should be expected to avoid the capture lek on their next lek visit. However, there was no difference ()( - 0.192, f· 0.740) in the likelihood to visit a different lek on the second 1ek visit between distyrbed (75.8%) and undisturbed females (70.6%). The probability of visiting a different lek on lek visit 2 (n - 79, 74.7~) was less ()( • 7.889, f • 0.005) than in lek visits ~ 3 (n - 137 combined, 55.5%). The tendency to visit the same lek(s) in later visits may represent selection for leks by females, as opposed to effects of disturbance. DISCUSSION Bradbury predicted that most females (presumably more than 50%) should have home ranges that include only 1 1ek and that they should visit only 1 lek. However, most female greater prairie-chickens (74%) nested closer to a lek other than that where captured and most (85%) visited> 1 lek during the breeding season. Similarly, Svedarsky (1988) �93 7,--------------------------------------------------, TOTAL o UNDISTURBED --b.-- DISTURBED .... ;:J .... MAXIMUM NUMBER OF DIFFERENT LEK VISITS POSSIBLE 3 4 5 >5 LEKVISIT Fig. 2-8. Number of different leks visited by female greater prairie- chickens in relation to number of times they were observed on leks in northeastern Colorado, 1986-88. �94 ~ 80,-----------------------------------------------~ t: - TOTAL >. UNOISTURBEO en ~ ~ o --b-· 70 .• 0-79.... o .. OISTURBEO ····8···· .... .. .•. ... 0,.,56 .... .• .•. •. .•. •••• •t, ~ e, ~ 60 ,, , ffi ffi H: 0=22 .. ~....- .. .... 50 .. •., ,&,.... " •, "'" ..·8········ , D=25 , -El C t: ·en s ~ 40 ~~--------~~--------~----------~----------~~ 2 3 4 5 >5 ·LEKVlSIT Fig. 2-9. Likelihood of a female greater prairie-chicken visiting a lek different than the previous lek visited in relation to the consecutive number of her lek visit in northeastern Colorado, 1986-88. �95 found that 11 of 18 (61%) females in Minnesota nested closer to a lek other than that where captured. Although direct examinations of home range size are difficult, and possibly not interpretable (Beehler and Foster 1988), home range sizes were estimated to provide an additional evaluation of the relevancy of Bradbury's (1981} prediction that diameters of home ranges should be less than distances between neighboring leks. The average home range diameter (based on only 75% contours) was larger than predicted, despite the fact that a detection distance (distance at which males can be detected read i1y by females) was not added to the home range diameter (Bradbury 1981). Home ranges for 20 female sage grouse in California were also larger than predicted; they each included an average of 2.2 leks (Bradbury et al. 1989. It was not logistically feasible to detect all lek visits and movements by females. Therefore, measures of home range size and lek visitation are probably conservative. These results do not support Bradbury's (1981) predictions that females should visit only one lek and have home ranges that include only one lek. Thus, Bradbury's preference' hypothesis is rejected for greater prairie-chickens northeastern 'female in Colorado. One possible reason for rejecting the female preference hypothesis may be that distribution of resources has changed since the lek mating system evolved among greater prairie-chickens. observations For example, of large home ranges in this study partly appeared to be a function of the distribution of important resources such as corn (in center-pivot irrigated fields). If resources were dispersed more evenly throughout the range of greater prairie-chickens, lek visitation and �home range size of females might approximate the predictions of Bradbury (1981). Although historical changes in habitat of greater prairie-chickens have been dramatic, their original North American habitat has been difficult to document adequately. that greater prairie-chickens Most available evidence indicates survived on acorns (Quercus spp.) prior to introduction of cereal grains (Bogardus 1874, Brewster 1890, Judd 1905, Leopold 1931, Hamerstrom'et al. 1941, Sharpe 1968). historical distribution of greater prairie-chickens Consequently, was altered by both movement of agriculture north and west, and transformation savanah and oak woodland/grassland the of oak- into cropland (Bogardus 1874, Cooke 1888, Gross 1930, Leopold 1931, Schmidt 1936, Hamerstrom and Hamerstrom 1949). Since oak trees were hypothesized to exist in scattered clumps and groves throughout the original' range of the greater prairie-chicken (Kuchler 1964), it is doubtful that resources were ever evenly dispersed. The most important aspect of Bradbury's (1981) theory is the relationship form leks. between female home range size and tendencies for males to Sizes and distributions of home ranges may provide insight into why some species form leks and others do not. Unfortunately, Bradbury (1981) did not clearly explain how increased home range size of females could result in,the development of lek behavior by males. Other hypotheses, 'least costly male' (Wrangham 1980), 'hotspot' (Bradbury and Gibson 1983), and 'hotshot' (Beehler and Foster 1988), have incorporated the idea that increased opportunities for mate choice by females can result in the evolution of lek mating systems. While Bradbury (1981) suggested that female preference could result in a �97 predicted distribution of leks and female home ranges, his proposed predictions did not provide an adequate test for the theory that increased home range size ultimately leads to increased concentrations of males (Chapter 1). Similarly, recent hypothesis have done less to examine the importance of female choice than to interpret factors such as male-male competition (Beehler and Foster 1988), variable male breeding potential (Bradbury and Gibson 1983), and selection of males in non-nesting habitat (Wrangham 1980). Movements of females also have implications for management of greater prairie-chickens and other lek species. land use practices in areas critical to 1ek species may be modified to conform to specifications relating to 1ek locations. One way this may be done is to delineate an arbitrary circumference of habitat around all known leks~ Th.se ar~as are generally corisidered the most critical for nesting, a factor considered when mitigating for habitat disturbance or loss. However, little information is available describing female movements and nest sites in relation to 1ek location. For example, Beck (1977) and Connelly et al. (1988) showed that Wal1estad and Pyrah's (1974) arbitrary radius of habitat (3.2 km) that should be protected around sage grouse (Centrocercus urophasianus) leks would not be effective for populations because of larger than expected movements. The problem of delineating habitat may be of particular importance for greater prairie-chickens establishment in northeastern Colorado. of greater prairie-chickens The re- into formerly occupied range may necessitate additional information on both the critical features of nest sites and on the proximity of nests to lek sites. �98 Management of greater prairie-chickens in northeastern Colorado would be difficult if based on selection of arbitrary regions of habitat around known lek locations. Although all female greater prairie- chickens nested within 2.5 kIDof a lek, arbitrary selection of habitat within 2.5 kIDof .leks results in delineation of approximately 99% of the 300 ~ study area. Hense, if the management objective is to adequately protect nesting habitat for greater prairie-chickens (a logical objective), a procedure based on lek locations does little to delineate nesting habitat from non-nesting habitat. LITERATURE CITED Ammann, G. A. 1944. Determining age of pinnated and sharp-tailed grouses. J. Wildl. Manage. 8:170-171. Amstrup, S. C. 1980. 44:214-217. A radio-collar for game birds •. J. Wildl. Manage. Beck, T. D. I. 1977. Sage grouse flock characteristics selection in winter. J. Wildl. Manage. 41:18-26. and habitat Beehler, B. M., and M. S. Foster. 1988. Hotshots, hotspots, and female pref~rence in the organization of lek mating systems. Am. Nat. 131:203-219. Bogardus, A. H. 1874. Field, cover, and trap shooting. Co., New York, N.Y. 343pp. J. B. Ford and Bradbury, J. W. 1981. The evolution of leks. Pages 138-169 in R. D. Alexander and D. W. Tinkle. eds. Natural selection and social behavior: recent research and new theory. Chiron Press, New York, N.Y. ___ , and R. M. Gibson. 1983. Leks and mate choice. Pages 109-138 in P. Bateson. ed. Mate choice. Cambridge Univ. Press, Cambridge, U.K. ___ , , C. E. McCarthy, and S. L. Vehrencamp. 1989. Dispersion of displaying male sage grouse. II. The role of female dispersion. Behav. Ecol. Sociobiol. 24:15-24. Brewster, W. 1890. The heath hen. Forest and Stream 35:188. �99 Connelly, J. W.,· H. W. Browers, and R. J. Gates. 1988. Seasonal movements of sage grouse in southeastern Idaho. J. Wi1d1. Manage. 52: 116-122. Cooke, W. W. 1888. Report on bird migration in the Mississippi Valley in the years 1884 and 1885. U.S. Dep. Agric., Div. Econ. Ornith., Bull. 2. 313pp. Dixon, K. R., and J. A. Chapman. 1980. Harmonic mean measure of animal activity areas. Ecology 61:1040-1044. Emlen, S. T., and L. W. Oring. 1977. Ecology, sexual selection, and the evolution of mating systems. Science 197:215-223. Gross, A. O. 1930. Progress report of the Wisconsin prairie chicken investigation. Wisconsin Conserv. Comm., Madison. 112pp. Hamerstrom, F. N., Jr., and F. Hamerstrom. 1949. Daily and seasonal movements of Wisconsin prairie chickens. Auk 66:313-337. ___ , F. Hopkins, and A. J. Rinzel. 1941. An experimental study of browse as a winter diet for prairie chickens. Wilson Bull. 53:185-195. Judd, S. D. 1905. The grouse and wild turkeys of the United States and their economic value. U.S. Bur. Bio1. Survey, Bull. 24. 55pp. Kuchler, A. W. 1964. Potential natural vegetation of the conterminous . United States. Am. Geogr. Soc. Spec. Pub1. 36. Leopold, A. 1931. Report on a game survey of the North Central States. Am. Game Assoc., Madison, Wisc. 299pp. Robel, R. J., J. N. Briggs, J •.J. Cebula, N. J. Silvy, C. E. Viers, and P. G. Watt. 1970. Greater prairie chicken ranges, movements, and habitat usage in Kansas. J. Wi1d1. Manage. 34:286-306. Schmidt, F. J. W. 1936. Winter food of the sharp-tailed grouse and pinnated grouse in Wisconsin. Wilson Bull. 48:186-203. Sharpe, R. S. 1968. The evolutionary relationships and comparative behavior of prairie chickens. Ph.D. Thesis, Univ. Nebraska, Lincoln. 187pp. Svedarsky, W. D. 1988. Reproductive ecology of female greater prairie chickens in Minnesota. Pages 193-239 in A. T. Bergerud and M. W. Gratson. eds. Adaptive strategies and population ecology of northern grouse. Vol. I. Univ. Minnesota Press, Minneapolis. Wallestad, R., and D. B. Pyrah. 1974. Movement and nesting of sage grouse hens in central Montana. J. Wildl~ Manage. 38:630-633. Wrangham, R" W. 1980. Female choice of least costly males; a possible factor in the evolution of leks. Z. Tierpsychol. 54:357-367. �100 CHAPTER 3. MOVEMENT BY FEMALE GREATER PRAIRIE-CHICKENS IN RELATION TO LEK LOCATION: EVALUATION OF THE HOTSPOT HYPOTHESIS OF LEK EVOLUTION INTRODUCTION Bradbury (1981) was the first to suggest that increased female home range size could result in an increased ability by females to select males and/or leks. In the 'female preference' hypothesis, he suggested that female choice drives the process of lek evolution. Subsequently, Bradbury and Gibson (1983) proposed the 'hotspot' hypothesis in which they suggested that males form clusters in areas of high female traffic. The major objective of this research was to evaluate the feasibility of the hotspot hypothesis by examining associated predictions with respect to lek locations and patterns of movement by female greater prairie-chickens in northeastern Colorado. The basic prediction of the hotspot hypothesis is that female traffic determines lek location. Although this prediction is difficult to test (Bradbury et al. 1986, Beehler and Foster 1988), Bradbury et al. (1986) suggested that overlap in female home ranges could be used to determine optimum lek locations, thus permitting comparison of optimum and actual lek locations. Unfortunately, even if home ranges can be determined accurately, possible biases associated with female movements to leks cannot be avoided. Beehler and Foster (1988) noted the circular nature of using �101 female home ranges to determine hotspots, particularly when home ranges may be a function of lek location. Also, there are no A prior; ways to determine how close or far apart optimum and actual lek locations should be to support or reject the hotspot hypothesis. differences Finally, inevitable in trapping intensity and absence of a completely marked female population must be dealt with. I used the actual distances between each female's nest site and lek where she was captured as indirect measures of home range size. Use of this procedure avaoids problems with estimation of female home ranges (Beehler and Foster 1988) and the absence of a completely monitored population and takes into account the importance of female ability to move between their home ranges and leks. Complexity of potential biases is the major reason why analysis of the hotspot hypothesis is complicated. Consequently, 4 separate analyses were used to examine the relationship between lek locations and areas of female traffic; all analyses were designed to test the prediction that leks were in areas of high female traffic. Since each analysis could have strengths and weaknesses, the overall results for the 4 analyses should provide a better examination of the hypothesis. An initial analysis was conducted to provide the suggested comp~rison between breed~ng potential at optimum and actual lek locations (Bradbury et al. 1986). Since this analysis was potentially affected by numerous biases, including methods used for trapping females, additional analyses were conducted. A second analysis was designed to examine breeding potential at actual display sites and determine the relative importance of lek size, lek stability (on a yearto-year basis), and trapping intensity. Although the second analysis �. 102 enabled an examination of all leks, most leks were on portions of the study area where no females were trapped. Consequently, they were in locations with few or no marked nesting females. . A third analysis was conducted to analyze breeding potential at leks on the core portion of the study area only. One objective was to compare breeding potential at actual and random lek locations. characterized Unfortunately, the previous 3 analyses were by bi~ses that made direct comparisons of breeding potential at actual and random lek locations difficult to analyze and interpret. Even though random locations could be selected within a study area that was large enough to include all female nest sites, the density of radio-marked females would be greatest in the middle of the study area. Hence, proportionally more random locations would be on the edge of the study area where density of marked females was lower. A fourth analysis was conducted to compare breeding potential at actual and random lek locations, while controlling such biases. METHODS Greater prairie-chickens were studied in 1986-88 on a 301-knf study area centered 10 km northeast of Eckley, Colorado (40 II' N, 102 0 22' W). The area consisted of grassland, sand sagebrush (Artemisia filifolia), and small soapweed (Yucca glauca) intermixed with irrigated fields of agriculture, primarily corn. Locations of leks (2 or more displaying males) and maximum male attendance at each lek were determined each year. Estimates of male attendance were based on a minimum of 2 visits to each lek each year. Leks active during each year were considered 'permanent' and those 0 �103 active during only 1 or 2 years were considered 'temporary'. Trapping was concentrated on 16 leks within a core area of approximately Female greater prairie-chickens traps and cannon nets. 100 km2• were captured on leks using walk-in One to 15 females were captured on each lek. Captured females were banded with a numbered aluminum band and a unique combination of 3 colored plastic bands, and fitted with batteryor solar-powered radio transmitters attached to poncho-type markers (Arnstrup 1980). Radio weights ranged between 1.8 and 2.3% of each birds' body weight. Nests for radio-marked greater prairie-chickens were located using a portable receiver and 3-element Vagi antenna. All locations were recorded using Universal Transverse Mercator coordinates (nearest 10-m interval). Although home range size was not estimated directly, distance between each female's nest site and the lek where captured was used as an indirect measure of home range size. An estimate of breeding potential was used to evaluate the relative quality of given locations; breeding potential was defined as the proportion of distances between a given location and all. nest locations that were closer than observed distances between nests and lek-of-capture locations. Bradbury et al. (1986) suggested that pattern of male settlement in combination with overlap of female home ranges could be used to estimate the breeding potential at any given location. Breeding potential was defined as the proportion of distances between a given location and all nest locations that were closer than observed distances between nests and lek-~f-capture locations (Fig. 3-1). Breeding potential for a given location was calculated as: �104 OBSERVED DISTANCES * NEST o LEK BREEDING POTENTIAL 0/9 0/9 Fig. 3-1. Breeding potential for an example with 3 leks and 3 nesting females. Heavier shading represents higher breeding potentials (sample values are shown in some areas). �105 Breeding potential - n n I I Oq i-I j-l ,where n - number of females with known nest and lek-of-capture locations, . ~ - distance between nest location and lek-of-capture location for i-I, ••• , n females, ~ - distance between given location and j=l, ..., n nest locations, 0'1 - 0 if A, < 01, Oq - 1 if ~ > ~, and Oq - 0.5 if ~ • ~. The estimate of breeding potential resulted in values between 0 and 1. In the example (Fig. 3-1), breeding potential for males at the 3 leks shown (left to right) would be 2.5/9, 1.5/9, and 4.5/9. Larger breeding potentials reflect proximity to nesting females, whereas breeding potentials of about 0 should be expected for locations far from the study area. All simulations were done with Statistical Analysis System (SAS Institute 1985). Analysis 1 The initial analysis was conducted to estimate optimum lek locations based on 'peaks' in breeding potential, thus permitting comparison of optimum and actual lek locations (Bradbury et ale 1986). Optimum lek locations were identified by examining the breeding potential at grid nodes on the study area gridded by horizontal and �iUb vertical axes. Axes initially were separated by 0.1 km and subsequently refined to identify 'peaks' of breeding potential. 'Peaks' represented topographic rises in breeding potential of 0.001 from surrounding breeding potentials. Distances between locations of peaks in breeding potential and nearest actual lek locations were compared with distances between 100 random locations and nearest actual lek locations. Random locations were selected within a rectangular core of the study area bounded on all 4 sides by lek-of-capture locati~ns. The analysis was conducted with MRPP (Zimmerman et al. 1985, Biondini et al. 1988) for all 3 years combined. Analysis 2 The second analysis was conducted to compare breeding potential at .actual display sites with actual and simulated dispersions of females. In addition, the analysis permitted examination of the importance of breeding potential relative to lek size, lek stability (on a year-toyear basis), and trapping intensity. Breeding potential was first estimated for display sites on the study area in relation to nest locations of each female. Second, breeding potential was identified at the same display sites with 1000 randomly generated sets of nests. Each set of nests was generated by randomly selecting a distance from the actual set of nest-capture lek distances and a random direction for each female. Subsequently, simulated nest locations were identified with random directions and distances relative to each female's actual lek-ofcapture location. �107 In addition to comparison of breeding potential at leks with actual and simulated distributions of females (If tests), breeding potential at lek sites was examined (correlation) in relation to number of males attending the lek, number of females trapped at the lek, and year-to-year lek stability. Leks active in all 3 years were considered permanent, while leks active 1 or 2 years were considered temporary. All comparisons were conducted for each year separately. Analysis 3 The third analysis was designed for analyzing breeding potential at leks on the core portion of the study area only; most leks were in portions of the study area where no females were trapped. The breeding potential at leks where females were captured was estimated and compared with the breeding potential for the same leks with 1000 randomly generated sets of nests for each year (If tests). The mean breeding potential value for each set of data was weighted by the number of females captured at each 1ek. Each random set of nests was generated using the procedure from Analysis 2. Analysis 4 The fourth analysis was designed to compare breeding potential at actual and random 1ek locations with actual distributions of nests. Instead of eliminating biases associated with random number selection, this analysis used random dispersions of nests (as with Analysis 2) for estimating the size of the biases. For control, the breeding potential was examined at actual and random lek locations with random dispersions of nests. If there were no biases associated with selection of �locations, the breeding potential should not differ for actual and random lek locations with random distributions of nests. Experimental simulations consisted of actual and random lek locations with actual dispersions of nests. Comparison of experimental and control simulations was with a ~ contingency table. Random lek locations were chosen within the minimum rectangular core of the study area that . included all leks where females were captured. Analyses were conducted for each year separately. RESULTS Radios were placed on 36 adult and 56 yearling female greater prairie-chickens during the 1986-88 breeding seasons. different leks were on the study area during 1986-88. active from year-to-year; Sixty-five All leks were not 41, 42, and 47 were active during each breeding season, respectively. The distance between nearest neighboring leks was 1.31 ± 0.56 (!± SO), 1.20 ± 0.70, and 1.18 ± 0.62 km each year, respectively (medians of 1.42, 1.32, 1.15 km). Nest-Capture Lek Movements Females (n - 75) nested 0.2 to 29.1 km from leks where captured (Fig. 3-2). There was no apparent difference (f - 0.184) in movement attributable to year (General Linear Model with year and age as class variables). transformed Similar results were obtained when distances were log- (f - 0.111). Directions of movement for females between capture leks and nest sites were random ()( z 9.160; [8 45 categories]). 0 f = 0.241, OF = 7 �109 N 2 Fig. 3-2. 4 KM Distribution of distances and directions of movement for 75 female greater prairie-chicken nest sites and leks where each female was captured in northeastern Colorado, 1986-88. �110 Analysis 1 Seventeen possible lek sites were pinpointed at 'peaks' of breeding potential in the core of the study area (Fig. 3-3). Sites of peak breeding potential were surrounded by capture leks indicating that predicted lek locations may have been biased by trapping intensity. Overall, predicted leks were closer than random locations to actual leks. The mean distanc~ between predicted and nearest actual lek was 0.63 kID (SO - 0.40, n - l7). Only 1 of 100 distances between random locations and actual leks was smaller than 0.63 kID (X - 0.92; SO = 0.12); the 2 distributions were different (MRPP, f < 0.001). Since 11 of the actual leks closest to predicted leks were not permanent (active during only 1 breeding season), distances were also measured to permanent leks (X - 1.12, SO - 0.69). Only 7 of 100 distances between random locations and permanent leks were smaller than 1.12 kID (X - 1.31, SO ill 0.14); the 2 distributions were different (MRPP, f < 0.001). Analysis 2 Breeding potential was compared at all display sites with actual and random distributions of nests (Fig. 3-4). Most estimated lek breeding potentials were less for random nest dispersions than with actual nest dispersion (68.3% for 1986, 48.8% for 1987, 67.3% for 1988). If nest and lek dispersion were independent of each other, 50% of the lek breeding potentials with random nest dispersions should have been less than the breeding potentials with actual nest dispersions ()( = 12.075, f = 0.003). Breeding potential at display sites did not appear to be affected by male attendance (maximum number of males observed on each lek) for �III LEKS CAPTURE LEKS HYPOTHETICAL LEKS 60 :E" 55 l:t o Z ~50 ~ •••• ::l 45 15 20 25 30 UTM (KM EAST) Fig. 3-3. Contour lines for levels of breeding potential on greater prairie-chicken study. area in northeastern Colorado, 1986-88. �·~==========~· · · · ·· 1 301····· II 1986 (0=41) D [J 1987 (D=43) 1988 (D=55) ·~===========..· - ~20~ ·j ~ ~ 10 0-10 10-20 20-30 30-40 40-50 50·S0 SO-70 70-80 PROPORTION OF HIGH RANDOM VALUES (%) Fig. 3-4. Examination of breeding potential at actual display sites (n - 41 in 1986, n - 43 in 1987, n - 55 in 1988) with random distributions of females in relation to the breeding potential at the same display. sites with the actual distribution of females for greater prairie-chickens in northeastern Colorado, 1986-88. Proportion of high random values refers to the percentage of each set of random nests (n = 1000) resulting in higher breeding potentials at a given lek location than the actual distribution of nests. �113 1986 (~ < 0.001, f • 0.912), 1987 (~ • 0.003, f -0.752), and 1988 a (~2 • 0.004, f • 0.666). However, trapping success (number of females caught) appeared to improve breeding potential for 1986 (~2 • 0.014, f = 0.469), 1987 (~ • 0.002, f • 0.760), and 1988 (~ - 0.007, f - 0.534). There were no differences in breeding potential associated with lek stability (permanent or temporary) for 1986 (~ • 1.835, f - 0.176), 1987 (~ • 2.931, f • 0.087), and 1988 (~ - 0.231, f = 0.631). Analysis 3 The breeding potential at leks where females were captured was estimated and compared with the breeding potential for the same leks with 1000 randomly generated sets of nests for each year (Fig. 3-5). Although most simulated values were smaller than actual values, the results were not consistent for 1986 (~ • -495.616,f < 0.001), 1987 (~ • 3.364, f • 0.067), and 1988 (~ - 0.324, f • 0.569). Analysis 4 Breeding potential was compared for actual and random lek locations. The biases associated with selection of random locations on a fixed study area were obvious when the difference in breeding potential for random nest dispersions with actual and random lek dispersions was examined (Table 3-1). Despite biases, the proportion of breeding potentials smaller for random than actual leks with actual distributions of nests was smaller than control values (random vs. actual leks with random distributions of females) for 1986 (~ = 79.360, f < 0.001), 1987 (~ = 40.751, f < 0.001), and 1988 (~ = 69.929, f < 0.001). �114 30 _ . fit -en ~ 25 - o . [ill 20 - . Z . '.::::: ... o r-. t ::t !;;r 15 ..J ~ 1986 1987 1988 . ::~~~:: 10 ~- - i·· •• 1·-1········.·1········_··- --·--··---: 5 _._ •. _. __ ...- 1·· :••1·· •• •• 1J J-ll-"~i--.. :.::: .... ::. ':' ..,,::: ::.~.::~::.~.::~ .:.:. .:.:. l .t« .: :::: :. :::: .: :::: .: :::: :-::~ .: :::: r-r::J md ... ~.ua~ ....:~~~~.~.·~~m:~<~·:~~~.-~:·~~:I~:~·~ul~·: -a, O~~~~~· I I -I.s TO-1.0 . :: I I I I I -as TO-2.0 -1.5TO-1.0 ~.s TO0.0 :.:. I • :.:. :.:. I I 0.5TO1.0 • :.:• ITT 1.5TO2.0 I 2.5 TO3.0 I 3.5 TO4.0 DIFFERENCE IN BREEDING POTENTIAL Fig. 3-5. Examination of breeding potential at capture leks with 1000 random distributions of nests in relation to the breeding potential at the same leks with the actual distribution of nests for greater prairiechickens in northeastern Colorado, 1986-88. Difference in breeding potential refers to simulated breeding potential minus actual breeding potential. �115 Table 3-1. Simulated breeding potential for combinations of actual and random leks and nests (n • 1000 for each combination) for greater prairie-chickens in northeastern Colorado, 1986-88. Actual leks Random leks Random leks> Year Median ! SO actual leks (%) -• 0.243 0.247 0.059 9.60 0.284 0.030 0.222 0.223 0.044 24.60 0.185 0.185 -• 0.171 0.171 0.007 2.50 Random 0.182 0.184 0.034 0.169 0.170 0.039 9.20 Actual 0.330· 0.330 -• 0.325 0.325 0.006 .21.50 Random 0.332 0.043 0.330 0.328 0.047 38.65 Nests Median ! Actual 0.315 0.315 Random 0.283 Actual SO 1986 1987 1988 0.331 -Only 1 value was calculated for actual 1ek - actual nest combinations, hence no standard deviation was given. �DISCUSSION The results do not provide conclusive evidence the hotspot hypothesis is applicable to lek behavior in greater prairie-chickens. However, several results indicate that leks are in areas of relatively high female traffic. First, observations of visits to leks by females indicate that females usually fly to a lek from their approximate nesting area and return to their nesting area after leaving the lek. In addition, when the random nature of female movements from leks where they were captured to nest sites is considered, leks essentially are visited from all directions. Second, the breeding potential at most lek sites was higher with the actual dispersion of females than it was with random dispersions of females. Likewise, breeding potential at lek sites where females were captured gener~lly was higher with the actual dispersion of females than with a random dispersion of females. Finally, actual leks had a higher breeding potential than random leks. The selection of hotspots based on female nest sites and movements appeared to indicate that actual leks were not in the best areas. However, these results may have been biased by trapping intensity. Furthermore, distances between predicted and actual leks could not be compared with expected distances. These results fail to reject the hotspot ~ypothesis. One possible reason for the failure to reject may be the lack of adequate tests; in part, this may be caused by the circular nature of predictions associated with the hotspot hypothesis (Beehler and Foster 1988). Alternate procedures for examination of the hotspot hypothesis might include manipulation availability, of female traffic by alteration of food induced changes of intrasexual aggression among females, �117 or removals of males and/or leks. Furthermore, if local populations of males could be removed, settlement of replacement males on previously occupied sites could be used as support for the ability of males to assess the quality of a site (Warner 1988); considerations of quality could include the likelihood of en.countering females. LITERATURE CITED Amstrup, s. C. 1980~ 44:214-217. A radio-collar for game birds. J. Wildl. Manage. Beehler, B. M., and M. S. Foster. 1988. Hotshots, hotspots, and female preference in the organization of lek mating systems. Am. Nat. 131:203-219. Biondini, M. E., P. W. Mielke, Jr., and E. F. Redente. 1988. Permutation techniques based on euclidean analysis spaces: a new and powerful statistical method for ecological research. Coenoses 3:155-174. Bradbury, J',W. 1981. The evolution of leks. Pages 138-169 in R. D. Alexander and D. W. Tinkle. eds. Natural selection and social behavior: recent research and new theory. Chlron Press, New York, N.Y. ___ , and R. H. Gibson. in P. Bateson, ed. Cambridge, U.K. ___ , leks. 1983. Leks and mate choice. Pages 109-138 Hate choice. Cambridge Univ. Press, , and I. H. Tsai. 1986. Hotspots and the dispersion of Anim. Behav. 34:1694-1709. SAS Institute. 1985. SAS User's guide: statistics. SAS Inst., Inc., Cary, N.C. 956pp. Version 5 edition. Warner, R. R. 1988. Traditionality of mating-site preferences in a coral reef fish. Nature 335:719-721. Zimmerman, G. H., H. Goetz, and P. W. Mielke, Jr. 1985. Use of an improved statistical method for group comparisons to study effects of prairie fire. Ecology 66:606-611. �CHAPTER 4. GREATER PRAIRIE-CHICKEN MIGRATION IN NORTHEASTERN COLORADO INTRODUCTION Migration can be defined as na regular round trip within a lifespan of the individual" (Sinclair 1983:241). Tetraoninae patterns. lagoDus and Members of the subfamily are particularly variable with respect to migration While some species such as willow and rock ptarmigan (Lagopus l. mutus) are clearly migratory (Weeden 1964, Irving et ale 1967), most species are considered either partial- or non-migrants. Bl ue grouse (Dendragaoy.s obscyrys)(Wing .1. 1947, Mussehl 1960, Zwi ckel et 1968), spruce grouse (Dendr~gapys canadensis)(Herzog and Keppie 1980, Schroeder 1985), and white-tailed ptarmigan (Lagopus leucurys)(Hoffman and Braun 1975) display a partial migration in which some individuals migrate and others do not. Although blue grouse migration may be caused by elevational differences in habitat (Wing 1947, Mussehl 1960; Zwickel et al. 1968), spruce grouse migration has been attributed to fidelity to specific wintering and breeding areas and variabiliti in dispersal movements during their first spring (Herzog and Keppie 1980, Schroeder 1985). Greater prairie-chickens Tetraoninae, Hamerstrom are among the most mobile of the with seasonal movements to 170 km (Hamerstrom and 1949). Early reports indicated that greater prairie-chickens migrated between winter and breeding ranges; movements were suggested to be particularly extensive among females and typically related to habitat �119 (Gross 1930, Leopold 1931, Schmidt 1936, Hamerstrom and Hamerstrom 1949). Observations of large flocks moving north or south, skewed sex ratios among wlnter flocks, and seasonal increases or decreases in prairie-chicken movements. abundance have been cited as evidence of these migratory It has been hypothe~ized that introduction of corn· throughout the range of greater prairie-chickens resulted in the elimination of their migratory tendencies (Gross 1930). Patterns of movement may have important implications for the management of prairie-chickens. The appropriate timing of transplants to re-establish populations and management of blocks of habitat of restrictive sizes (Kirsch et al. 1973) both require an understanding of such movements. Additionally, extensive movements may have important effects on effective population size and/or genetic diversity of a population (Shields 1982). I examined timing and distance of seasonal movements by greater prairie-chickens in northeastern Colorado during 1986-89 with respect to sex, age, and reproductive status. The study area had both Wintering and breeding populations of prairie-chickens. Hence, there was no A priori reason to expect large movements between winter and breeding ranges by members of the population. Seasonal movements were described with respect to basic questions: 1) Are greater prairie-chickens migratory?, 2) Do females migrate farther and more frequently than males?, and 3) Do movements occur in response to factors such as habitat, natal dispersal, weather, and nesi failure? �120 METHODS A study area of 301 ~ 0 centered 10 km northeast of Eckley, Colo.rado (40 11' N, 102 22' W) was chosen for research on greater prairie-chickens. 0 The area consisted of grassland, sand sagebrush (Artemisia filifolia), and small soapweed (Yucca glauca) intermixed with irrigated fields of agriculture, primarily corn. The area contained 41-47 active leks (> 2 males on each) during each breeding season, 1986-89 (Chapter 2). Trapping was concentrated on a core area of approximately 75~. Trapping at winter feeding sites with walk-in traps baited with corn and at leks with walk-in traps and cannon nets resulted in capture of 243 different greater prairiechickens (Table 4-1). Since cannon nets were believed to cause more disturbance than other trapping methods, they were used sparingly and only in 1986. All captured birds were banded with a numbered aluminum band and a unique combination of 3 colored plastic bands. Age was ascertained by examining patterns of feather wear (Ammann 1944). Birds were classified as yearlings, 5-17 months of age (1 Nov of 1st year to 31 Oct of 2nd year), and adults, older than 17 months of age (after 31 Oct of 2nd year). Battery- and solar-powered radio transmitters were attached to poncho-type markers (Amstrup 1980) and placed on 145 greater prairiechickens (Table 4-1). Radio weights were between 1.8 and 2.3% of each birds' body weight. Seasons were defined j priori as breeding (15 Feb - 30 Jun), late summer (1 Jul - 15 Aug), autumn (16 Aug - 31 Oct), and winter (1 Nov 14 Feb); designation of seasons was based on aspects of breeding behavior and movement (Robel et al. 1970). Radio-marked greater �121 Table 4-1. Number of greater prairie-chickens captured, banded, and fitted with radio transmitters in 1986-89 in northeastern Colorado. Captures (including recaptures) Walk-in traps Category Bait sites Lek sites Cannon nets Totals Radios 76 86 21 183 34 Adults 32 56 14 102 20 Yearlings 44 30 7 81 14 52 132 6 190 III Adults 30 61 2 93 48 Yearlings 22 71 4 97 63 128 218 27 373 145 Males Females Totals �122 prairie-chickens Vagi antenna. were located using a portable receiver and 3-element Observations typically were obtained daily between 1 Mar and 15 Aug (sightings were up to 4 days apart during unusual circumstances). Occasionally, birds could not be found for periods up to 3 months; inevitably, when these birds were found, they had moved relatively long distances from locations where previously observed. The only exceptions were birds that were dead; some transmitters did not work for relatively long periods when they were partially buried in soil and/or vegetation. Aerial searches were conducted 2-3 times each year to locate missing transmitters. Locations for each sighting were obtained by direct visual observation (including nest sites) or by triangulation; 3 or more azimuths were obtained ~ 1.5 km of target transmitters and at angles-of-incidence greater than 35° and less than 145°. All locations were recorded using Universal Transverse Mercator coordinates (nearest lO-m interval). Examinations of accurac~ indicated that 90% of the locations derived by triangulation were within 250 m of actual locations (Chapter 2). Spring migration refers to winter - breeding season movements and autumn migration to breeding season - winter movements, regardless of the specific timing. Migration was estimated as the distance between the lek (males) or nest (females) location and the mean location (mean x and mean y coordinate) during the winter season. Additional movements were estimated using mean locations of late summer and autumn seasons as intermediate pOints between breeding and wintering areas; this procedure was used to help identify the specific timing of migratory movements. �123 Dates of movements were only estimated for distinct movements of ~ 5 km (clearly not part of the normal home range). Although arbitrary, the purpose of this procedure was to estimate the timing of migration, not the likelihood of migration. Furthermore, most movements meeting this criteria were distinct; birds often moved from the study area (~ 5 km) within 36 hours of being observed within their normal home range (females were often on or near their nest locations). In these cases, timing of movement was estimated based on searching intensity (time between searches) and/or the typical duration of a long migratory movement (2-4 days). For example, if a bird was located on 12 June, but not on or after 14 June, it was assumed to have migrated on 13 June. Statistical comparison of di~tributions of distances and dates was with Multi-Response Permutation Procedures (MRPP)(Zimmerman et a1. 1985, Biondini et a1. 1988). Comparisons of variances were conducted with f tests and comparisons of means with 1 tests, assuming unequal variances. RESULTS Timing of Movement The timing of distinct seasonal movements was recorded for greater prairie-chickens throughout the year (Fig. 4-1). The average date of spring migration was 27 March (SE - 6, n - 13) for females and 20 February (SE - 4, n - 3) for males; differences in averages (1 - 4.518, f - 0.001) and distributions (f - 0.002) were significant. The average date of aut~mn migration was 4 July (SE - 10; n - 33) for females and 28 July (SE 0.792, = 30, n= 5) for males; difference between the averages (1 = f = 0.464) and distributions (f = 0.520) were not significant. �124 7-----------6 ::::sc·Rrk~·1I1~CAT10"l·:::::. .... 1:' .•:t.•••. ",I. ~rnTJ 'T . ~s a: asu.4 o a: .......................... ........................... al ....................... ....................... W3 ::E ::>2 [] BROOD FEMALES E NON-BROOD FEMALES ~ MALES AUTUMN MIGRATION Z 1 1~ JAN 11 FEB II JAN 11 MAR lll'EII I "'ft II MAR 1 MAY .112APR 3.IUN .110MAY 1 JUL 17 JUN 21 JUL 15 JUL 21 AUG 12 AUG 25 11 SEP sat 23 OCT I OCT DATE (BEGINNING OF 2-WEEK PERIOD) Fig. 4-1. Distribution of movement dates for greater prairie-chickens in northeastern Colorado, 1986-89. �125 The tendency for males to move early to their respective 1ek areas was expected since leks were frequently active by mid-February. movements by females appeared to be bimodal. Much of this variability apparently was associated with the female's brood status. of migration was 10 June (SE • 9, n - 23) Autumn The mean date for females without broods and 26 August (SE - 19, n - 10) for females with broods (1 - 3.693, f 0.003). Similarly, differences associated with brood status were detected ()( - 9.989, f - 0.002) when the proportions of females migrating early and late were compared (separated by period between midJuly and mid-September when no migratory movements were observed, Fig. 4-1) • Many birds that apparently migrated were not observed during the winter or breeding seasons. This appeared to be due to mortality prior to the primary season of interest (winter or breeding season). Consequently, ~reeding season - winter distances were subdivided into breeding season - late summer and late summer - winter components. These distances were estimated only for females tracked throughout all seasons to provide additional information on the specific timing of movements (Fig. 4-2). These distances supported the finding that most long migratory movements occurred during May and June. For example, if many birds had movements for the 2 periods that were equal (along the 45 line), it would indicate that movements of many birds occurred 0 either throughout either season or during both seasons. -Most exceptions to the trend of early movements (May and Jun) were brood females; they tended to move toward winter ranges later than females without broods. The only brood female which moved a long distance early in the summer lost her brood within 2 days of hatch �::E 9 0 ~ BROOD FEMALES W 0 o NON·BROOD FEMALES D. ~ -a: en e C 0 0 D. W !z ;: 0 .~ I a: W s, 3 ::E ::E D. D. ~ :::> en- 0 c: 0 D. D. W 5 0 o 5 10 15 20 25 NEST - LATE SUMMER DISTANCE (KM) Fig. 4-2. Relationship between nest - late summer distances and late summer - .winter distances for female greater prairie-chickens northeastern ~olorado, 1986-89. in �127 (prior to movement). In general, movements between late summer and autumn ranges were relatively small for all females (Fig. 4-3). However, the 4 longest movements were by females with broods. Further support for timing of migration was obtained by examining dates of disappearance for radio-marked individuals that subsequently were not found (Fig. 4-4). Although some of these radios may have become disabled following mortality of the bird, the synchrony of disappearance dates with observed migration dates indicates that many radios may have been on birds that migrated but were not found. Hence, it is likely that estimates of migration tendencies and distances were low. Distance of Movement Fifty distances were recorded between winter and breeding areas for radio-marked greater prairie-chickens. Distances ranged between 0.64 and 39.98 km for females and 0.95 and 6.05 km for males (Fig. 4-5). These movements were treated as a continuum rather than categories in which migrants could be distinguished from non-migrants. Adult females moved an average of 8.20 km (SE - 1.73, n - 32), yearling females moved 14.39 km (SE - 5.90, 2.48 km (SE • 0.52, n • 3). n - 9), adult males moved and yearling males moved 3.52 km (SE = 0.75, Although yearlings of both sexes tended to move farther than adults, differences in distributions for females, £ • 0.289 for males). movements n - 6), (MRPP test) were slight (£ - 0.501 For both ages combined, female (!a 9.18 km, SE • 1.73) differed (MRPP test, male movements (!a 2.74 km, SE = 0.44). £= 0.019) from Distances for females typically were larger (1 - 3.600, £ - 0.001) and more variable (£ z �128 e,---------------~~--------------------------~ 5 • BROOD FEMALES [ill NON-BROOD FEMALES 1 o 0.0-0.5 0.5-1.0 1.0-1.5 1.5-2.0 2.0-2.5 2.5-3.0 3.0-3.5 3.5-4.0 DISTANCE (KM) Fig. 4-3. . Distribution of late summer - autumn distances for female greater prairie-chickens in northeastern Colorado, 1986-89. �129 o 5.--------------------------------------------------, [ill) Z BROOD FEMALES 11 NON-BROOD ~4 ~ FEMALES MALES Q. SPRING MIGRATION ~ ~3 AUTUMN MIGRATION BROOD FEMALES C en a: C -mu, 2 o a: ~1 :E ::::> z o 2SFEB 2SMAR 22APR 20 MAY 17JUN 16JUL 12AUG 11 SEP 1a0CT 11 MAR • APR 8 MAY JUN 1 JUL 21aJUL 28 AUG 25 SEP 23 OCT a DATE (BEGINNING OF 2-WEEK PERIOD) Fig. 4-4. Distribution of disappearance dates in relation to average date of migration (arrows) for greater prairie-chickens Colorado, 1986-89. in northeastern �12 en C a:: 9 . til FEMALES IT]. MALES [8J YEARLINGS al U. 0 a:: 6 W al ::E ::> z 3 o 0-2 4-6 2-4 8-10 12.•14 16-18 20-22 24-26 28-30 32-34 36-38 6-8 10-12 14-16 18-20' 22-24 26-28 30-32 34-36 ·38.40 DISTANCE (KM) Fig. 4-5. Distribution of distances between breeding and winter areas for 38 female and 12 male greater prairie-chickens in northeastern Colorado, 1986-89. �131 49.51, f < 0.001) than for males. Results for comparisons of distributions were similar when age groups were not combined (f = 0.048 for adults, f - 0.168 for yearlings). In addition, fewer male (0%, Fig. 4-6) than female (31.6%, Fig. 4-7) migrations crossed the study area boundary (~ - 4.986, f - 0.026). In addition to differences associated with migration distance and sex, movements were unusual with respect to their variability. Among females, both small and large distances were common, suggesting a bimodality in tendency to migrate. Furthermore, many individuals occupied winter or breeding areas on the main study area as the result of relatively large migratory movements (Fig. 4-7). Many females bypassed occupied habitats during the course of their movements. Analysis of migration distances (Figs. 4-5, 4-6, and 4-7) was conservati~e, only birds with known winter and breeding areas were included; this was despite the fact that migratory movements typically occurred during May and June. Consequently, many birds which died prior to a season of interest were not included in the analysis, even though many had obviously made long migratory movements. The analysis of movement data was expanded to additional birds, based on locations between breeding and late summer - winter areas (Fig. 4-8 for males, Fig. 4-9 for females). When the expanded analysis was conducted, females had an average migration distance of 10.58 km (n • 60) and males a distance of 2.86 km (n = 22, MRPP test, f < O.OOl)(Fig. 4-10). The distribution of distances supports the overall bimodality in migration tendency. Twenty-one of 60 females (35.0%) and 1 of 22 (4.5%) males migrated farther than 5 km; the difference between females and males was �132 4460 .. , N .. , .. , ::I: , : " ~ ,: 4455 ~ o i ...•, ~ z ::E ~4450 ::E ·S ?---i :'T'--"- _...,-_._.._' : ···-·--···-I··············'<:·-·-~~: i ~: -··l'~·:<~~:····--r·····---· '-T' - - rI ___ :_ 4445 . 715 : - •. I ---------------------~------. ,-720 725 730 735 UTM (KM EAST) Fig. 4-6. Straight-line distances between winter ranges (circles) and lek sites (asterisks) for 12 male greater prairie-chickens in northeastern Colorado, 1986-89. �133 N 4470 4460 ~ ~ 4450 a: o z - r ---. •• _~_ :E 4440 ~ - - - - - - ~ - - <I' 730 •••••••• 740 ,-------------- , ~ ~ 4430 ·························r··· . .. .... .. .. .. .. INSET •. 1 4420 4410 ---·r·----· I --..-- ··T-·.... 710 ....• .. ..• •. 1 1 .. .. .. ..• 1 .... 1- _ •• .. .. .. ,--------------------_. , ,, 720 UTM (KM EAST) Fig. 4-7. Straight~line distances between winter ranges (circles) and nest sites (asterisks) for 38 female greater prairie-chickens northeastern Colorado, 1986-89. .. in The inset shows females with both winter range and nest site on the study area. �4460 ''''... t .. , - N 'I :~ . ......... ··..·;..·.;··;;..··· ~;···..·····..~·r··· ···· ··· . ~4455 ct o Z ~ ~4450 ~ !5 4445 , I , , 705 710 715 720 725 730 735 UTM (KM EAST) Fig. 4-8. Straight-line distances between estimated winter ranges (circles) and lek sites (asterisks) for 22 male greater prairie-chickens in northeastern Colorado, 1986-89. �135 4470 4460 ~ ~ 4450 a: o z ~ 4440 730 .._.. ~ ,'-------------" , , ~ ~ 4430 :J 740 . .•. , ... INSET .•. , .•. I ...•. I ... , ... 4420 ..• I I .•. , .•.•. 4410 .• I I 1- _, ... .•. ... , .•. 710 ~-------------------_# 720 UTM (KM EAST) Fig. 4-9. Straight-line distances between estimated winter ranges (circles) and nest sites (asterisks) for 60 female greater prairiechickens in northeastern Colorado, 1986-89. The inset shows females with both winter range and nest site on the study area. , ,, �136 12 E FEMALES [J. MALES . en C a: 9 as LJ.. 0 a: 6 W £Xl :E ~ Z 3 0-2 . 4-6 8-10 12-14 16-18 20-22 24-26 28-30 32-34 36-38 40-42 44-46 OISTANCE (KM) Fig. 4-10. Distribijtion of migration distances for 60 female and 22 male greater prairie-chickens in northeastern Colorado, 1986-89. �137 significant ()( - 8.230, f - 0.004). The pattern of movement indicated there was no obvious relationship between direction of movement and tendency to move north or south prior to particular seasons (Fig. 4-11). Site Fidelity Site fidelity is important for distinguishing between regular migrations and nomadic movements (nomadic movements could be associated with unsuccessful ne~ting attempts). Of 6 radio-marked males monitored during consecutive years, all attended the same lek. An additional 10 banded males without radios were also observed on the same lek in consecutive years. Nest locations between consecutive years were an average of 0.S3 km (SE - 0.19, n - S) apart. The distributions of these distances were not different (f - 0.599) from distances (!- O.Sl km, SE - 0.11, n - 10) between first nests and re-nests within the same year (Fig. 4-12). There were no differences detected (MRPP) between males and females for fidelity between consecutive late summer (f - 0.440), autumn (f - 0.196), and winter (f - 0.070) ranges; the trend detected for the winter season was primarily caused by the late summer movements of a single yearling male. Sample sizes were too small to detect differences in fidelity associated with sex. differences With sexes combined, there were no between seasons (f • 0.629). Distances between consecutive winter ranges (based on mean location for each winter range) were relatively small (!- 1.0S km, SE - 0.33, consecutive late summer (!~ 1.30 km, SE (!- 1.99 km, SE - 0.79, n z n a 9) and distances between = 0.39, n = 7) and fall ranges 6)'were variable (Fig. 4-13). This �138 FEMALES N o MALES * o, o 48KM , , Fig. 4-11. Distribution of migratory movements by distance and direction for 60 female and 22 male greater prairie-chickens northeastern Colorado, 1986-89. movement of 4 km. in Each concentric circle represents a �139 5-,-----------.,. ·4 " CONSECUTIVE YEARS EJ FIRST AND RE-NESTS en c a: m3 u. o a: w CO2 :::E ::) z 1 o 0.0-0.2 0.2-0.4 0.4-0.S O.S-O.S 0.S-1.0 1.0-1.2 1.2-1.4 1.4-1.S 1.S-1.8 DISTANCE (KM) Fig. 4-12. Distribution of distances between nests during consecutive years and distances between first nests and re-nests within the same year for female greater prairie-chickens 89. in northeastern Colorado, 1986- �140 6,-----~~--------------------------------------~ Ii! LATE SUMMER [ill WINTER o AUTUMN 5 (J) ~4 CD U. 03 a: W CD :::e2 :::> Z 1 o 0-1 1-2 2·3 3-4 4·5 5·6 DISTANCE (KM) Fig. 4-13. Distribution of distances between late summer, autumn, and winter ranges during consecutive years for greater prairie-chickens northeastern Colorado, 1986-89. in �141 variability may have been caused by differences in timing of migration and brood status, rather than changes in site fidelity between years. Observations of Movements Due to speed and distance of many migratory movements, observations of actual movements were difficult to obtain (5 examples, Fig. 4-14). Three larger migrations (Fi%. 4-14A, B, and C) were typical of June migratory movements to winter ranges following nest predation. Female C made the movement within a 71-hour period and female B moved between her winter and breeding range within a 137-hour period. Female C was of interest for 3 additional reasons: 1) her first movement between her winter area and nesting area may have been a dispersal movement (she was a yearling), 2) her subsequent return to the same general winter area may indicate that migration movements may mirror dispersal movements, and 3) she wintered on the same general area in which females A and B spent the breeding season. While many of the characteristics of movement and/or disappearance patterns were consistent with the hypothesis that migrations were relatively fast and direct, 2 movements were notable because of their unpredictability. One female (Fig. 4-140) was observed during an early June migration of approximately 8 days and 13.4 km. It is doubtful that this movement was typical, since most migratory birds disappeared the study area within a 36-hour period. from Another female (Fig. 4-14E) appeared to change winter areas during winter (late December). Her new area was closer to her breeding area, indicating the possibility that some females may not display site fidelity to their winter areas. �142 N B 4470 \ \ 4460 ,------- --, , , ,, , I •. • I I ,, I ----------------' 4430 * NEST LOCATION o WINTER RANGE o TRANSITIONAL LOCATION 4420 710 720 730 740 750 UTM (KM EAST) Fig. 4-14. northeastern Seasonal movements of 5 female greater prairie-chickens Colorado, 1986-89. in �143 DISCUSSION Results from this study clearly indicated that greater prairiechickens in northeastern Colorado are partially migratory; some individuals move sUbstantial distances (up to 40 km) between seasonal ranges while others remain resident on relatively small areas. In addition, 95.5% of 22 males and 65.0% of 60 females migrated < 5 km indicating that females are more likely to migrate than males. Last, observations of direction, timing, and behavior associated with movement indicated that movements apparently were not caused by differences habitat, weather, and nest failure. in The possibility that migratory movements may retrace a bird's dispersal movement between its first wintering and breeding area was not rejected. Askins (1913) suggested that greater prairie-chickens. were -basically non-migratory. In contrast, several lines of evidence have been used to suggest that greater prairie-chickens were historically migratory in portions of their original and acquired range (Gross 1930, Leopold 1931). Grange (1948) considered evidence for regular migrations prior to 1850 as 'indisputable'. First, seasonal increases and decreases in population size were noted in both southern and northern parts of the range. Second, evidence of biased sex-ratios in resident and/or migratory populations has been used to demonstrate migratory tendencies. Finally, large flocks of greater prairie-chickens have been observed making directional movements and/or in areas with no acceptable habitat. Data from banded and radio-marked individuals also has provided support for seasonal movements. Seasonal changes in regional populations of greater prairiechickens have often been used as evidence of migrations. For example, �winter increases have been observed in several locations where greater prairie-chickens were either absent (Bogardus [1874] for central Illinois; Leopold [1931] for Ames, Iowa; Webster [1912] for Charles City, Iowa, and Manitowoc County, Wisconsin; Stempel and Rodgers [1961] and Spengler [1984] for southwestern Iowa) or less abundant (Cooke [1888] for southern Iowa, northern Missouri, and southern Illinois; Judd [1905] for Kentucky; Leopold [1931] for Trempealeau County, Wisconsin; and Schmidt [1936] f~r Oklahoma, Texas, Arkansas, Missouri, southern Illinois, and eastern Kansas) during the breeding season. Similar population changes have been noted during the breeding season in areas where birds were either absent (Leopold [1931] for Forest County, Wisconsin and North Dakota and Banta [1892] for northeastern Minnesota) or less abundant (Cooke [1888] and Leopold [1931] for northern Iowa and. southern Minnesota) during the winter. A bias in sex ratio during winter has been used as evidence confirming migratory tendencies of greater prairie-chickens, in particular. - and females For example, of the few birds remaining in North Dakota during winter, all were reportedly males (Leopold 1931). Likewise, Schmidt (l936) found a male:female ratio of 121:16 in northern Wisconsin and 49:272 in southern Wisconsin. Cooke (1888) suggested that females migrate farther and more frequently than males. Unfortunately, problems associated with accurate identification of sex ratios in winter flocks have made adequate examinations of regional movements difficult. Stronger evidence for a sex bias in movement tendencies has been gathered with banding and radio-telemetry Hamerstrom studies. Hamerstrom and (1949) found that females migrated farther and more frequently than males in Wisconsin. The longest migration distance �145 recorded for Wisconsin was about 170 km for a female (Hamerstrom and Hamerstrom 1949). In addition, 41.8% of 134 males and 4.7% of 107 females banded during winter were observed on the same area in spring. When distances between winter and breeding areas were examined, 32.2% of 298 distances for females and 4.9% of 834 distances for males were> km. 8 In North Dakota, 15 females moved 4.3 km and 6 males moved 2.6 km on average between winter and breeding season (Toepfer and Eng 1988). Evidence from this study demonstrated similar patterns of movement; 95.5% of 22 males and 65.0% of 60 females migrated < 5 km. Larg~r distances and frequency of movement by females was consistent with patterns of movement documented for other grouse including spruce grouse (Herzog and Keppie 1980, Schroeder 1985, 1986), blue grouse (Hines 1986), sharp-tailed grouse (Tympanuchus phasianellus) .(Kobriger 1965), sage gr9use (Centrocercus urophasianus)(Beck white-tailed 1977), ptarmigan (Hoffman and Braun 1975), rock ptarmigan (Weeden 1964), and willow ptarmigan (Weeden 1964, Irving et al. 1967). Although reasons for more extensive movement by female than by male Tetraoninae are not known, possible factors may include sex differences intraspecific competition in (Herzog and Keppie 1980), habitat selection (Gross 1930, Schmidt 1936), and inheritance (Keppie 1980, Schroeder 1988). Gross (1930) and Leopold (1931) examined anecdotal evidence for several hypothesized migrations including observations for South Dakota; Burlington, Iowa; northeastern Minnesota; Door County, Wisconsin; Ripley County, Missouri; and Peoria County, Illinois. in Burlington, The migratory movements Iowa apparently were predictable enough that hunters were able to prepare for them. �146 Migratory tendencies possibly were reduced with the introduction of corn throughout northern parts of greater prairie-chicken (Hamerstrom and Hamerstrom 1949). range Even so, planted food patches in Wisconsin failed to prevent birds from migrating (Grange 1948). Unfortunately, migrations. little information is available on pre-settlement Cooke (1888) suggested that migratory movements were as long as 1,000 km before corn was introduced. Stempel and Rodgers (1961) suggested that birds wintering in southwestern to have migrated at least 160 km. Iowa into the 1960's had Neither example can be verified. Data from Hamerstrom and Hamerstrom·(1949) for banded individuals in Wisconsin has shown that movements can be substantial movements were 46 and 170 km for females). in Minnesota (2 largest Migratory movements of 12 km (Svedarsky 1988) and 30 km in Michigan (Ammann 1957) have also been observed for females. Movements up to 40 km were observed in this study; whereas many individuals were essentially resident on one area, others moved relative long distances between breeding and wintering areas. It was impossible to detect any differences between patterns of migration observed in this study and patterns described in previous research. Most migratory movements of greater prairie-chickens occur during October-December and Hamerstrom March-April Occasionally, in autumn (Leopold 1931, Lehmann 1939, Hamerstrom 1949, Mohler 1952, Ammann 1957) and in spring during (Leopold 1931, Hamerstrom and Hamerstrom 1949). movements may be as early as August (Leopold 1931). However, none of the examples of migratory movements was documented with observations \ of banded individuals. �147 Data collected on radio-marked birds in this study indicated the timing of migration may be relatively complicated. movements occurred in February-April Although spring (males earlier than females), autumn movements occurred as early as late-May. Examination of movements by females indicated that timing of autumn movements may be bimodal; brood females moved during October-November and unsuccessful females typically moved during June. The early timing of most autumn migrations by greater prairiechickens appears to be unusual for grouse. Most species appear to move during late summer or autumn (Irving et a1. 1967; Zwicke1 et a1. 1968; Hoffman and Braun 1975; Schroeder 1985, 1986). possibility that greater prairie-chickens, This raises the particularly females without broods, are migrating prior to the normal summer molt period. supported by bimodality in movement timing. Females with broods may wait until after the molt before undertaking long movements. period may be important for prairie-chickens often long, necessitating This is The molt since their movements are at least some flying. In contrast, spruce grouse have been observed walking on their migratory routes (Schroeder 1985). Variability in migration timing, sex ratios, and distance has resulted in several possible explanations for movements. First, movement in response to regional food availability has been a common explanation for migration of greater prairie-chickens (Cooke 1888, Gross 1930, Leopold 1931, Lehmann 1939, Hamerstrom and Hamerstrom 1949, Mohler 1952, Ammann 1957, Toepfer and Eng 1988). Second, Banta (1892) suggested that large flocks of greater prairie-chickens 1875-8 may have been fleeing from hunters. in Duluth in Third, Leopold (1931) and �148 Hamerstrom and Hamerstrom (1949) suggested that tendencies to migrate varied with weather. Data from radio-marked birds in this study both contradicted and supported many of the previous suggestions: 1) the prevalence of early movements (most in mid-summer) indicated that weather was not a major driving force, 2) movements were typically in random directions, birds were just as likely to. move north as south prior to winter, 3) many birds moved onto the same areas that other birds moved from, 4) some individuals remained on the same areas during winter and summer while others migrated, and 5) females migrated more frequently and farther than males. When these observations were considered in combination with the general availability of corn throughout the region (typical winter food), habitat was ruled out as an adequate explanation for the observed movements. Svedarsky (1988) suggested that many of the long movements by females may be due to disturbances by predators at nest sites; this explanation Colorado. could be considered for greater prairie-chickens in However, although female migrations frequently were made following nest failure (Fig. 4-1), females regularly returned to their previous nesting areas, despite nest failure. Likewise, females producing broods also migrated, even though their movements were later during autumn or early winter. Some exceptions may be explained by either late re-nesting attempts (late migration by a female without a brood) or early brood loss (early migration by a female with a brood). Although sample sizes for males were small, males apparently approached their winter ranges following the breeding season (late May - Jun). The �149 timing of male movements appeared to be closely tied to their attendance at leks. Males often displayed at leks until mid-June. One possible explanation for migration is that migratory movements mirror dispersal movements; birds may disperse between their place of hatch and breeding area while displaying fidelity to both their first wintering and breeding areas. A similar type of dispersal phenomenon has been noted for blue grouse (Hines' 1986) and spruce grouse (Herzog and Keppie 1980, Schroeder 1985, 1988). Henc~ it would be difficult, if not impossible, to differentiate between the causes of dispersal and migration~ Similarity between patterns of dispersal and migration may be especially important, given the context of most observations of largescale migrations. , recorded,during Consequently, Many .observations of migratory movements were periods of rapid range expansion (late 1800's). it may be possible that migrations of greater prairie- chickens have not consisted of more than local movements in response to habitat availability in combination with dispersal tendencies and fidelity to wintering/breeding sites. The similarity between adult and yearling patterns of movement for females indicated that migratory movements of adults may mirror the dispersal movements of yearlings ,during their first spring (as noted with spruce grouse; Herzog and Keppie 1980, Schroeder 1985). This may help explain the· female bias in movement that is consistent with most patterns of dispersal in birds (Greenwood and Harvey 1982). LITERATURE CITED Ammann, G. A. 1944. Determining the age of pinnated and sharptailed grouses. J. Wildl. Manage. 8:170-171. . �___ • 1957. Lansing. The prairie grouse of Michigan. 200pp. Michigan Oep. Conserv., Amstrup, S. C. 1980. A radio-collar for game birds. Manage. 44:214-217. Askins, C. 1913. The prairie chicken. Banta, D. D. 1892. 39:443. J. Wildl. Field and Stream 18:634-641. A prairie chicken migration. For. and Stream Beck, T. D. I. 1977. Sage grouse flock characteristics and habitat selection in winter .. J. Wildl. Manage. 41:18-26. Biondini, M. E., P. W. Mielke, Jr., and E. F. Redente. 1988. Permutation techniques based on Euclidean analysis spaces: a new and powerful statistical method for ecological research. Coenoses 3:155-174. Bogardus, A. H. 1874. Field, cover, and trap shooting. Co., New York, N.Y. 343pp. J. B. Ford and Cooke, W. W. 1888. Report on bird migration in the Mississippi Valley in the years 1884 and 1885. U.S. Dep. Agric., Div. Econ. Ornith., Bull. 2. 313pp. Grange, W. B. 1948. Oep., Madison. Wisconsin grouse problems. 318pp. .. Wisconsin Conserv. Greenwood, P. J., and P. H. Harvey. 1982. Natal and breeding dispersal of birds. Annu. Rev. Ecol. Syst •.13:1-21. .Gross, A. O. 1930. Progress report of the Wisconsin prairie chicken investigation. Wisconsin Conserv. Comm., Madison. 112pp. Hamerstrom, F. N., Jr, and F. Hamerstrom. 1949'- Daily and seasonal movements of Wisconsin prairie chickens. Auk 66:313-337. Herzog, P. W., and D. M. Keppie. 1980. Migration in a local population of spruce grouse. Condor 82:366-372. Hines, J. E. 1986. Recruitment of young in a declining population of blue grouse. Ph.D. Oiss., Univ. Alberta, Edmonton. 256pp. Hoffman, R. W., and C. E. Braun. 1975. Migration of a wintering population of white-tailed ptarmigan in Colorado. J. Wi1dl. Manage. 39:485-490. Irving, L., G. C. West, L. J. Peyton, and S. Paneak. 1967. Migration of willow ptarmigan in arctic Alaska. Arctic 20:77-85. Judd, S. D. 1905. The grouse and wild turkeys of the United States and their economic value. U.S. Bur. Bio1. Survey Bull. 24. 55pp. �151 Keppie, D. H. grouse. 1980. Similarity of dispersal among sibling male spruce Can. J. Zool. 58:2102-2104. Kirsch, L. H., A. T. Kleff, and H. W. Hiller. 1973. Land use and prairie grouse population relationships in North Dakota. J. Wildl. Hanage. 37:449-453. Kobriger, G. D. 1965. Status, movements, habitats, and foods of prairie grouse on a sandhills refuge. J. Wildl. Manage. 29:788800. Lehmann, V. W. 1939. The heath hen of the south. Oyster Comm., Bull. 16. Ilpp. Texas Game, Fish and Leopold, A. 1931. Report on a game survey of the North Central States. Am. Game Assoc., Madison, Wisc. 299pp. Hohler, L. L. 1952. Fall and winter habits of prairie chickens in southwest Nebraska. J. Wildl. Manage. 16:9-23. Hussehl, T. W. 1960. Blue grouse production, movements, and populations in the Bridger Mountains, Montana. J. Wildl. Manage. 24:60-68. Robel, R. J., J. N. Briggs, J. J. Cebula, N. J. Silvy, C. E. Viers, and P. G. Watt. 1970. Greater prairie chicken ranges, movements, and habitat usage in Kansas. J. Wildl. Manage. 34:286-306~ Schmidt, F. J. W. 1936. Winter feed of the sharp-tailed grouse and pinnated grouse in Wisconsin. Wilson Bull. 48:186-203. Schroeder, H. A. 1985. Behavioral differences of female spruce grouse undertaking short and long migrations. Condor 87:281~286 • .' ______ • 1986. The fall phase of dispersal in juvenile spruce grouse. Can. J. Zool. 64:16-20. ______ • 1988. Dispersal in spruce grouse: is inheritance involved? Anim. Behav. 36:305-308. Shields, W. M. 1982. Philopatry, inbreeding, and the evolution of sex. State Univ. New York Press, Albany. 245pp. Sinclair, A. R. E. 1983. The function of distance movements in vertebrates. Pages 240-258 in I. R. Swingland and P. J. Greenwood, eds. The ecology of animal movement. Clarendon Press, Oxford, U.K. Spengler, R. 1984. Greater prairie chicken in Osceola County. Bird Life 54:21. Iowa Stempel, M. E., and S. Rodgers, Jr. 1961. History of prairie chickens in Iowa. Proc. Iowa Acad. Sci. 68:314-322. �152 Svedarsky, W. chickens Gratson, northern O. 1988. Reproductive ecology of female greater prairie in Minnesota. Pages 193-239 in A. T. Bergerud and M. W. eds. Adaptive strategies and population ecology of grouse. Vol. I. Univ. Minnesota Press, Minneapolis. Toepfer, J. E., and R. l. Eng. 1988. Winter ecology of the greater prairie chicken on the Sheyenne National Grasslands, North Dakota. Pages 32-48 in A. J. Bjugstad, Tech. Coord. Prairie chickens on the Sheyenne National Grasslands. u.S. Oep. Agric. For. Servo . Gen. Tech. Rep. RM-159. Webster, C. l. 1912. Winter migration of the prairie chicken. and Stream 78(15):471. For. Weeden, R. B. 1964. Spatial segregation of the sexes in rock and willow ptarmigan. Auk 81:534-541. Wing, l. 1947. Seasonal movements of the blue grouse. Wildl. Conf. 12:504-510. Trans. No. Am. Zimmerman, G. M., H. Goetz, and P. W. Mielke, Jr. 1985. Use of an improved statistical method for group comparisons to study effects of prairie fire. Ecology 66:606-611. Zwickel, F. C., I. o. Buss, and J. H. Brigham. 1968. Autumn movements .of blue grouse and their relevance to populations and management. J. Wildl. Manage. 32:456-468. �153 APPENDIX A. WALK-IN TRAPS FOR CAPTURING GREATER PRAIRIE-CHICKENS IN NORTHEASTERN COLORADO EXAMINATION OF TECHNIQUE Numerous techniques have been used for capturing greater prairiechickens (Tympanuchus cupido) including cannon nets'(Silvy and Robel 1967, 1968, Arthaud 1968, Robel et al. 1970), mist nets (Silvy and Robel 1968, Robel et al. 1970), helicopters (McCune 1970, Brown 1981), drop nets (Jacobs 1958), bownets (Anderson and Hamerstrom 1967, Robel et al. 1970), noose carpets (Berger and Hamerstrom 1962), drop-nets (Jacobs 1958, 1959, 1964, Arthaud 1968), dip-nets (Arthaud 1968), night-lfghting w~th hand-nets (Robel et ai. 1970), and walk-in traps (Berger and Hamerstrom 1962, Hamerstrom and Truax 1938, Robel et al. 1970, Hamerstrom and Hamerstrom 1973, Horak 1985, Toepfer et al. 1988). of the techniques include modifications Many such as trap placement (leks, bait sites, and travel corridors), trap design, use of hen decoys (Anderson and Hamerstrom 1967), and male vocal izations (Sllvy and Robel 1967). Other than Toepfer et al. (1988), little has been documented about specific use of walk-in traps at lek sites. The objective of this paper is to document the design and effectiveness northeastern of walk-in traps used at greater prairie-chicken Colorado. leks in Although Toepfer et al. (1988) documented a similar trap design for greater prairie-chicken different type of funnel. leks, they used a �154 Walk-in traps were made with 6' X IS" (15.2 X 45.7 cm) sections of 2" X 4" (5.1 X 10.2 cm) welded wire. Thus, approximately 32 traps could be made from a 100' (30.5 m) roll of 3' (91.4 cm) wide wire. By cutting the roll in half lengthwise, each trap had a series of 2n points that could be imbedded in the'ground. 1" chicken-wire (Fig. A-I). A funnel was made for each trap with Funnels were specifically designed so that most exposed points (places where the wire was cut) were on the bottom of the funnel. Other exposed wire points should be bent into positions that reduce the probability of captured birds contacting them. Funnels were fitted between the ends of each cut section welded wire. Each trap was covered with nylon net that could be opened near the trap edge to retrieve captured birds. Each walk-in trap was arranged with chicken-wire leads to maximize the potential for capturing birds. wUh All traps and leads were secured metal stakes pushed into the ground. Traps generally were placed in areas of highest prairie-chicken activity; these locations were determined through observation or by examination of concentrations droppings and feathers. used (Fig. A-2). of Several basic designs of traps and leads were Some traps were oriented such that birds were likely to encounter the trap when walking toward the center of the lek and others were positioned so that birds would encounter traps when walking away from the center of the lek. orientation Although definitions of trap (inward, neutral, and outward) were arbitrary, they enabled the relative effectiveness of different trap orientations to be compared. A total of 1,970 trap days (each funnel counted as a trap) were used to capture 231 birds (Table A-I). Females were captured more �155 c D A B 32 INCHES (81.3 MM) ,15 INCHES (38.1 MM), T e INCHES (15.2 MM)l Fig. A-I. 11 INCHES ENTRANCE EXIT (27.9 MM) BOlTOM Funnel design for greater prairie-chicken walk-in traps in northeastern Colorado, 1986-89. To construct, similar letters should be conected, dashed lines should be folded, and exposed points should be blunted. �156 '* •••• I , '----, .•. "-~- .•. .. , , ,, ,," .. , .. ' .-:::-"-~ ": '__ - -- . ... .. " , , o .. .•. ...•.. ~ ' ,, " (3:' .•. .. .•. .•. '_---- ..•." .•. .•. ... " ,, '" , .._-~ , ., '. @'".~' , ' I \ I \ tr -,•._... _.,tp.. \ .. ,.. " ,- _ ~ Gl I... I .- ' ••• "* •• , .... containing -- \ I I , ,, ,r " " .. 6'- ...•. ..._-£) ,, I , ,, , .. ~ ••• I Typical arrangements of walk-in traps and 18" high chicken- wire leads (dashed lines) for capturing greater prairie-chickens northeastern '' . e,.-,--' '..' ..._"NL._ \ e- - - _, ,, ,I r I t9 , ,, , ,, I ,, I , I , Fig. A-2. ...• 20M , .•. I -, ~__--8 , 10 I .•. ,, ,, , - .•. I I ", , .',' •. \ ,, I ...•:9-- ... _ .• -.. \ ,•, 1 ,'" Colorado, 1986-89. Two types of traps were used: 1 a Single funnel and the other containing 4 funnels. in .•. ,, �157 Table A-I. Trap success of male and female greater prairie-chickens relation to trap orientation in northeastern Colorado, 1986-89. in Funnel openings were pOinted toward the center of the lek (inward), away from the center of the lek (outward), or with no particular direction (neutral) . Category Trap days Males Females Total Number Percent Number Percent Number Percent Inward 753 25 3.32 35 4.65 60 7.97 Outward 425 26 6.12 36 8.47 62 14.59 Neutral 792 48 6.06 61 7.70 109 13.76 1970 99 5.03 132 6.70 231 11.73 Total �frequently than males in these traps ()( - 5.008, f - 0.025). In addition, the 3 orientations of traps were variable with respect to proportion of males ()( - 7.055, 0.048) trapped. f - 0.070) and females ()( - 7.907, f • Both sexes apparently were more difficult to catch in traps facing inward. Finally, time of year appeared to be more important for capturing females than for males (Fig. A-3). Females easily could be caught during 5-20 April, while males could be captured throughout the breed~ng season. The methods proposed herein can be modified. Observations at leks can be used to determine the effectiveness of the trap design and arrangement of traps on leks. Funnels were modified several times to increase either the likelihood of a bird entering and/or the probability of the bird remaining in the trap. In addition, effectiveness of traps for capturing males and/or females apparently was improved with movement of traps and chicken-wire leads on leks. The effectiveness of these techniques for capturing males and females could not be compared since the abundance of either sex was not known. Walk-in traps have several advantages over other trapping methods. Unlike most other trapping techniques, walk-in traps can be set at several locations at the same time. In addition, the lack of continuous attendance by researchers minimizes disturbance, especially if traps are . positioned in a manner that permits them to be monitored from relatively long distances. modify. Finally, walk-in traps are inexpensive to make and/or. Traps of this basic design also have been used to capture lesser prairie-chickens, Colorado. sharp-tailed grouse, and sage grouse in �159 15 Ii MALES Ifill -fa FEMALES 12 ~ Cf) 9 8 :J Cf) 8 ~ 3 ""7 1-11 12·16 11-11 20023 24-27 28-31 104 MAR Fig. A-3. loa 1012 13011 17·20 21·24 25028 >28 APR Distribution of greater prairie-chicken trapping success during the breeding season in northeastern Colorado, 1986-89. �160 LITERATURE CITED Anderson, R. K., and F. Hamerstrom. 1967. Hen decoys aid in trapping cock prairie chickens with bownets and noose carpets. J. Wildl. Manage. 31:829-832. Arthaud, F. l. 1968. Populations of the prairie chicken related to land use in southwestern Missouri. M.A. Thesis, Univ. ,Missouri, Columbia. 134pp. Berger, D. D., and F. Hamerstrom. 1962. Protecting a trapping station from raptor predation. J. Wildl. Manage. 26:203-206. Brown, D. l. 1981. The helinet: a device for capturing prairie chickens and ring-necked pheasants. Proc. S. E. Assoc. Fish Wildl. Agencies 35:92-96. Hamerstrom, F. N., Jr., and M. Truax. 1938. Traps for pinnated and sharp-tailed grouse. Bird-Banding 9:177-183. Hamerstorm, F. N., Jr., and F. Hamerstrom. 1973. The prairie chicken in Wisconsin. Highlights of a 22-year study of counts, behavior, movements, turnover and habitat. Dep. Nat. Resour., Tech. Bull. 64. 52pp. Horak, G. J. 1985. Kansas prairie chickens. Comm., Wildl. Bull. No.3. 65pp. Kansas Fish and Game ' Jacobs, K. F. 1958. A drop net trapp~ng technique for greater prairie _ chickens. Proc. Oklahoma Acad. Sci. 38:154-157. ______ ,. 1959. Restoration of the greater prairie chicken. Oklahoma Dep. Wildl. Cons., Div. Fed. Aid Wildl. Restoration Proj. W-65-R. 42pp. ______ • 1964. An electrically actuated release mechanism for dropnets. Proc. Annu. Conf. Southeastern Assoc. Game Fish Comm. 18:30-34. McCune, R. A. 1970. Prairie chicken moving days. Wildl. 28(7):12-15. Texas Parks and Robel, R. J., J. N. Briggs, J. J. Cebula, N. J. Silvy, C. E. Viers, and P. G. Watt. 1970. Greater prairie chicken ranges, movements, and habitat usage in Kansas. J. Wildl. Manage. 34:286-306. Silvy, 'N. J., and R. J. Robel. 1967. Recordings used to help trap booming greater prairie chickens. J. Wildl. Manage. 31:370-373. ______ , and . 1968. Mist nets and cannon nets compared for capturing prairie chickens on booming grounds. J. Wildl. Manage. 32:175-178. �161 Toepfer, J. E., J. A. Newell, and J. Monarch. 1988. A trapping prairie grouse he~s on display grounds. A. J. Bjugstad, Tech. Coord. Prairie chickens on National Grasslands. U.S. Oep. Agric. For. Servo RM-159. method for Pages 21-23 in the Sheyenne Gen. Tech. Rep. �162 APPENDIX B. LEK PERSISTENCE AND ATTENDANCE OF MALE GREATER PRAIRIECHICKENS IN NORTHEASTERN COLORADO INTRODUCTION Many species of Tetraoninae congregate on leks for breeding purposes including sage grouse (Centrocercus urophasianus), grouse (Tymoanuchus ohasianellus), pallidicinctus), sharp-tailed lesser prairie-chickens and greater prairie-chickens. (Tympanuchus Density of leks and attendance of males at leks have both been used as indices to overall status of their populations (Cannon and Knopf 1981). Despite frequent use of such indices, little. is known about male attendance at leks and the stability, consistency, and/or distribution of leks. Previous research on greater prairie-chickens suggested that only about 50% of males attend leks on any given day (Robel 1970). Robel's (1970) conclusions were based on capture of males on 3 leks and subsequent observation of marked males on the same leks. His data did not include observations of marked males captured off leks. Also, Robel (1970) did not track radio-marked males during peak breeding periods eliminating the possibility of observing males at locations other than the 3 lek sites. Similar research on sage grouse suggested that although visitation rates may be relatively high (approximately 90%), vistation may depend on time of year (Emmons and Braun 1984). Miller (1984) suggested that 41.5% of greater prairie-chicken disappeared between any 2 years. leks However, these conclusions were based �163 on 3 years of research on the edge of occupied range eliminating the possible differentiation between permanent and temporary leks (referred to as dominant and satellite leks, respectively, by Hamerstrom and Hamerstrom [I973]). This latter factor is particularly important because of Cannon and Knopf's (I98I) suggestion that lek densities adequately represent population trends because of the formation of satelite leks by subordinate males. Greater prairie-chickens are a useful subject for research on the dynamics of males and/or leks; (I) they can be captured during the winter to reduce biases associated with lek captures, (2) they can be monitored with radio-telemetry, monitored. and (3) all leks can be located and This paper documents the attendance at leks of radio-marked male greater prairie-chickens and the stability and distribution of leks. METHODS A study· area of 301 knf, centered 10 km northeast of Eckley, Colorado (40 II' N, 102 22' W) was monitored in 1986-90 to estimate 0 0 lek density and male lek attendance for greater prairie-chickens. additional 631 knf peripheral area was added in 1988. displaying males were considered a lek. An Two or more Attendance for each lek (within a year) was estimated by using the maximum of at least 2 counts of males at each site. Males, at leks were counted during March, prior to visitation by females. Leks were defined as 'persistent' (active during all previous years of research, 'permanent' (active during all 5 years of research), 'additional' 'temporary' (inactive at least 1 year), and/or (not described in previous years). �164 Trapping was concentrated on a core area of approximately during 1986-88. Male greater prairie-chickens 100 km2 were captured on winter feeding sites using walk-in traps baited with corn and on leks using walk-in traps and cannon nets. Captured males were banded with a numbered aluminum band and a unique combination of 3 colored plastic bands, and fitted with battery- or solar-powered radio transmitters attached to poncho-type markers (Amstrup 1980). Radio weights' ranged between 1.8 and 2.3% of each birds' body weight. yearlings Birds were classed as (first breeding year) or adults (> first breeding year) (Ammann 1944). Observations of 20 adult and 14 yearling radio-marked males were gathered between 0.5 hours before t9 2.0 hours after sunrise during 1 March-30 April, 1986-88; these dates were chosen to adequately represent the normal display season for males of 15 FebruarY-IS May. Radio-marked greater prairie-chickens were located using a portable receiver and 3-element yagi antenna. Observations of birds consisted of visual signtings of radio-marked males on leks (usually from 0.2 to 1.0 km away) and/or triangulation target transmitters (3 or more azimuths obtained ~ 0.5 km of and at angles-of-incidence greater than 35° and less than 145°. All locations were recorded using Universal Transverse Mercator coordinates Examinations (nearest 10-m interval). of accuracy indicated that 90% of the locations derived by triangulation 2). were within 250 m of actual locations (Chapter Even so, differentiation leks were relatively distinct. between observations of males on and off Radio-telemetry was used to estimate locations of radio-marked males before leks were approached (when traps �165 were checked); inevitably males that were estimated to be on the lek flushed with the other males on the lek. RESULTS Seventy-six leks were documented on the study area during 1986-90; 35 leks were found on the surrounding 631 ~ Fig. B-1). area in 1988 (Table B-1, Fifteen locations (10 on main study area) were also documented that were used as display sites for single males only. The density of leks was relatively stable at 0.14 leks/km2 (0.36 leks/mile2)(Table B-2). Although attendance of males at leks fluctuated between years, attendance at permanent leks (n - 19 leks with counts of displaying males all years) was positively correlated with yearly lek f - 0.006)(Fig. B-2). Attendance at all leks showed a similar relationship (~ - 0.787, f - 0.072). density (~ - 0.960, Yearly changes in attendance of males at leks throughout the study were examined for consecutive years (Fig. B-3). that fluctuations independent; These results indicated in attendance of males at neighboring leks may not be for example, attendance of males at leks generally increased on the western side of the study area in 1988 and decreased 1989. in However, when yearly changes in attendance of males (number) were compared between each lek and the nearest neighboring lek, the yearly changes were not positively correlated for 1986-87 (~ - 0.062, f 0.663), 1987-88 (~ - -0.069, f - 0.582), 1988-89 (~ - -0.035, f 0.791), and 1989-90 (~ - -0.230, f • 0.028). relationship was not significant periods combined (Fig. B-4). Similarly, the (r • 0.028, f • 0.375) with all time �166 Table 8-1. Maximum attendance of male greater prairie-chickens at display sites in northeastern Colorado, 1986-90 (/+'.- active lek without a count of males; ,_, - activity at site not known). Univeral Transverse Male Mercators (Zone 13) Lek 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 attendance Legal description East (m) North (m) Township Range Section Quarter 707970 709090 709830 710230 710480 710650 711180 711380 711750 712900 713270 714390 714540 714880 715420 715820 715830 716370 716610 716670 717050 717100 717150 717180 717250 717730 717750 717780 717930 717890 718130 718350 718690 718860 719030 719040 Continued. 4448270 4449010 4447340 4454050 4449370 4454850 4449000 4453280 4450530 4458300 4462440 4464320 4460620 4453420 4454310 4453760 4448010 4453120 4456480 4453840 4455600 4457710 4448870 4454700 4454290 4441630 4451570 4447490 4464480 4448140 4450100 4449920 4453530 4457680 4453800 4440690 3N 2N 2N 3N 2N 3N 2N 3N 3N 3N 4N 4N 4N 3N 3N 3N 2N 3N 3N 3N 3N 3N 2N 3N 3N 2N 3N 2N 4N 2N 2N 2N 3N 3N 3N IN 46W 46W 46W 46W 46W 46W 46W 46W 46W 46W 46W 46W 46W 46W 46W 46W 46W 46W 46W 46W 45W 45W 45W 45W 45W 45W 45W 45W 45W 45W 45W 45W 45W 45W 45W 45W 7 5 8 20 5 21 4 28 33 10 27 23 35 25 24 24 12 25 13 24 18 7 7 19 19 31 31 . 7 19 7 6 6 29 8 20 5 NW SW SE SE NE NW SW NE SE NE NE NE -SE NW SW SW NW NE NE SE SW SW NW NW SW NW NE SE NE SE NE NE NW SW SW NW 86 87 88 89 90 - - - - - + 14 3 + - - 3- - - 3 0 0 0 0 + 9 3 0 5 9 6 9 3 2 6 4 7 15 0 0 14 12 + - - 0 0 0 2 - 0 0 8 8 0 0 - 0 2 9 8 3 3 5 1 2 17 10 3 2 7 4 10 0 2 5 19 3 10 17 0 12 17 1 12 6 2 0 9 4 15 12 26 - - - - - - - - - 0 0 0 12 15 0 8 12 0 11 13 0 7 - 8 - 0 0 0 0 7 0 11 6 o 2 8 6 0 8 0 - - 0 6 14 0 6 9 0 14 11 0 10 - 6 0 10 0 9 0 - �167 Table 8-1. Continued. Male Univeral Transverse Mercators (Zone 13) attendance Legal description Lek East (m) North (m) Township Range,Section Quarter 86 87 88 89 90 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 719100 719340 719590 719630 719890 719920 720060 720470 720620 720980 721230 721250 721620 721750 721900 722100 722370 722440 722510 722850 723020 723300 723320 723440 723450 723780 724110 724120 724180 724270 724420 724450 724830 724960 725280 725750 725870 726290 726410 726520 Continued. 4443610 4451710 4464240 4453920 4464190 4449630 4448250 4456320 4455520 4458240 4455970 4450580 4445270 4461370 4455230 4453700 4463250 4453940 4456230 4456700 4456460 4458750 4463170 4450180 4455900 4444340 4453510 4454510 4463360 4459920 4449180 4455370 4448130 4447950 4450980 4447080 4445950 4447050 4449940 4455730 2N 3N 4N 3N 4N 2N 2N 3N 3N 3N 3N 2N 2N 4N 3N 3N 4N 3N 3N 3N 3N 3N 4N 2N 3N 2N 3N 3N 4N 3N 2N 3N 2N 2N 3N 2N 2N 2N 2N 3N 45W 45W 45W 45W 45W 45W 45W 45W 45W 45W 45W 45W 45W 45W 45W 45W 45W 45W 45W 45W 45W 45W 45W 45W 45W 45W 45W 45W 45W 45W 45W 45W 45W 45W 45W 45W 45W 45W 45W 44W 29 32 20 16 21 5 8 16 16 9 16 4 21 34 22 27 27 22 15 15 15 2 26 2 14 23 26 23 26 2 2 23 '11 12 36 13 13 13 1 18 NW NE NE SW NW SE NE NW SW NE SE NE NE NW . NW NW NE SE NE NE NE SW NW NW SW SW NE SE NE NE SE NE SE SW SW NW SE NE SE SW + + 19 16 11 5 0 0 - - 4 0 2 0 - - 3 14 18 24 0 2 0 0 0 8 0 2 0 6 6 16 0 0 8 0 0 8 + + 3 - - 6 18 10 12 11 14 14 6 0 5 0 0 1 0 4 5 4 4 6 6 5 0 0 2 0 0 4 0 4 1 0 0 5 2 0 0 0 0 1 1 5 6 13 10 8 10 0 0 1 12 6 14 7 8 0 7 8 8 0 o 10 2 5 0 2 0 0 0 2 1 12 9 9 - - - 4 0 - - 0 2 - - 14 0 6 0 5 2 0 10 3 3 0 2 1 0 + 9 9 10 12 9 - - 0 0 11 0 0 0 0 2 1 0 0 0 0 0 0 0 0 0 0 0 7 9 0 6 8 10 11 0 0 0 11 6 7 0 3 11 6 10 0 0 0 10 �168 Table 8-1. Continued. Male Univeral Transverse Mercators (Zone 13) attendance Legal description Lek East (m) North (m) Tow~ship Range Section Quarter 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 III 112 113 114 115 116 726700 727160 727280. 727410 727450 727590 727760 728090 728280 728300 728590 728630 728730 728740 728830 728940 729030 729820 730150 730360 730390 730430 730550 730970 731000 731320 732060 732420 732470 733080 734040 734150 734620 734700 735240 735690 725230 732740 733210 729960 Continued. 4450140 4444880 4444760 4461430 4454630 4456700 4462440 4458320 4447390 4460090 4448100 4448610 4445330 4448260 . 4464820 4448110 4457150 4466220 4458700 4449870 4446390 4450070 4464090 4455770 4447120 4445470 4462010 4445100 4454660 4446330 4459030 4451990 4459970 4453650 4448720 4456600 4455640 4445270 4444830 4452240 2N 2N 2N 4N 3N 3N 4N 3N 2N 3N 2N 2N 2N 2N 4N 2N 3N 4N 3N 2N 2N 2N 4N 3N 2N 2N 4N 2N 3N 2N 3N 3N 3N . 3N 2N 3N 3N . 2N 2N 3N 44W 44W 44W 44W 44W 44W 44W 44W 44W 44W 44W 44W 44W 44W 44W 44W 44W 44W 44W 44W 44W 44W 44W 44W 44W 44W 44W 44W 44W 44W 44W 44W 44W 44W 44W 44W 45W 44W 44W 44W 6 19 19 31 19 18 30 8 17 5 8 8 20 8 20 8 8 16 9 4 16 4 21 16 16 . 21 34 22 22 14 2 35 1 25 12 13 13 22 23 33 NW SE SE NE SE NE SE NW NW NW SW NW NW NW NE SE SE NW NW SW SW NW SE SE NE NE NE SE SE SW SE NE NW NW NW NE SW NE SW NW 86 87 88 89 90 9 9 8 9 8 9 10 0 0 0 7 17 - 6 + 12 17 14 + + 9 4 - - 1 - - 8 4 0 0 0 1 0 0 1 0 8 7 0 0 9 10 10 8 0 o 10 9 8 0 0 1 0 5 + 2 2 3 0 0 1 2 7 18 26 12 0 0 4 0 6 - 9 8 10 14 10 0 0 4 0 6 6 4 5 0 4 4 0 0 0 2 3 + 8 - 10 3 1 0 0 0 3 0 0 0 7 0 0 o 11 0 0 0 3 - 1 0 9 - 25 4 - 0- 0- - - - - - - - - - - - 0 6 17 0 - 0 11 0 - 6 0 0 0 0 0 17 a �169 Table 8-1. Continued. Male Univeral Transverse Mercators (Zone 13) attendance Legal description Lek East (m) North (m) Township Range Section QUarter 86 87 88 89 90 117 118 119 120 121 122 123 -124 125 126 727720 716660 717290 719070 728830 716840 717280 719840 721780 716610 4452770 4452040 4451310 4445590 4453620 4451610 4450980 4452720 4457720 4458130 3N 3N 3N 2N 3N 3N 3N 3N 3N 3N 44W 46W 45W 45W 44W 45W 4SW 4SW 45W 46W 30 36 31 20 29 31 31 29 10 12 SE NE NW NW NE NW SW SE SW NE 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 o 0 0 0 0 0 0 0 0 0 12 0 0 0 0 0 0 0 9 1 4 8 1 1 1 4 o 14 0 2 �.J../V N DISPLAY SITES o R47W! R4SW R45W ! R44W , I I T4N T3N I ~ ! 1 ~ I ; , : :: q 2( 100 :]::TJ;~~iC:;;,t!lo;~~~~J~it~;~ 1;:~;~J T3N ~ ~ I ; . , i :2 : I . : : 0 : ~'Q :.. : : 0.: : 52 54 ~80 i20 :0 i 111: 0: : 0: ;185 : ; ·······it~~~~~·:l24·t·······t···~t71···t····9.+·····1bS j +.., r······I·······l······I········I······i,······· I o~S1~13 !78~ ~047 07: : l' ~ . . CO : ! 9 80081 :0 ~ I ~ 75:)0: : : ~~': : :: o. o· ! i -r----r-- 0:: T2N. T1 N ! . r·······j······+······~····-+·····-+······ ·····.··i·······"'6'····82····a~··+······+···o·i······-+·······~~··i·Q.······· : I ~ ! ~~ 0 i i . 170! O~! ~ ~ S~70 i QOoJ2: i 1 10, :~::l~·:r=I:::~r:::I=;-T::t~~:r::J=T:::rtJ~Ivti;J~ 1 T2N :::r~t::::tTK¢Vi··r::l~rt··1°.1-:";._; l,otF @' , , , . :·+I+··-I-!·······H·····+·-r·l·ri@H, i· I : i I . . , : : " , i : '0' :48: : 82 : , i ~,o ,. ,02 ~ 0:011 i··' .;-_. Y fRP- - -!.•'_~-._••-'_ ••.,'_••'_••-_-•.'_.••. '•.•.•.: iii i iii R47W R4SW R45W R44W e , , T1N 2 I , 4KM I Fig. 8-1. Display sites of male greater prairie-chickens in northeastern Colorado during early - late spring (15 Feb-IS May) 198690. Numbers refer to the lek numbers on Table 1. �171 Table 8-2.· Indices of male greater prairie-chicken abundance in northeastern Colorado, 1986-88. 1986 1987 1988 1989 1990 Active 41 42 47 42 41 Persistent 41 33 28 25 24 Additional 41 9 15 6 5 33 30 34 34 Index Number of leks Still active from previous year Number of displaying mal es/l ek All leks .& 7.11 7.26 10.66 9.02 8.41 SO 3.72 4.35 6.46 3.80 4.72 .& 8.90 9.79 13.37 9.79 7.79 SO 3.64 3.93 5.05 2.40 3.19 Leks/krrf 0.14 0.14 0.15 0.14 0.14 Leks/mile2 0.35 0.36 0.40 0.36 0.35 Mal es/krrf 0.98 1.02 1.69 1.23 1.17 Ma 1es/mi 1e2 2.55 2.66 4.38 3.20 3.03 Permanent leks Lek density Displaying male density- ~ales displaying alone were used in the analysis. �172 16,-------------------------------~ 1988 6 ~--------~--------~--------~------~ 40 42 44 46 48 NUMBER OF LEKS Fig. 8-2. Relationship between maximum attendance of male greater prairie-chickens at 19 permanent leks (active every year with complete information) and total number of leks on the study area in northeastern Colorado, 1986-90 (vertical lines are 95% confidence intervals). �173 Fig. 8-3. 1986-87 1987-88 1988-89 1989-90 Change in male attendance at 86 display sites between years on a study area in northeastern Colorado, 1986-90. Lines represent contours of equal change in attendance of males at leks. Darkened shading indicates areas with an increase and lighter shading indicates areas with a decrease in lek attendance. �1./4 ~ * W ...J , , ~ W 20 - • •• ._._. ::: __ • 1. __ - : .•__ .• .•.• _~ .•.•.• , c::: L5 * * z 1() -~ * ~ ..••.... ~ .. ~~ , r-, o z o -- ~ ~ -10 W ~ , : 2 * *4 :*3** * t : 2* * ~* :* : 2 *: * : : * 2*:32 * : 2433: 2*'" ,* : * ** * ~4 *3* *.* :** ***5*2~62 4* ** • ·············-······f·········~-·~-·~~·~~~··~-···~···~* • 2442643*** * :* * 2 * * *3'**22 **4 : ** 24+4 3 : * * 2 * *3:* 3* : * *** *"1'* 2* 'I' *2*:** *: . 2: *2 2*: 2 2: , '" ' ••••••••••••••••••••••••••••••• _••••••• J••••••••••••••••••• _L._ ~ __ ., : . * '" Z « J: I . T (!') o :: .. " ... w C§ z w ** * -1() ~ . ~ . I . ,, , ** I I I o 1() 2() CHANGE IN ATTENDANCE AT EACH LEK Fig. 8-4. Comparison of yearly changes in attendance of males (number) at each lek and its nearest neighboring lek for the study area in northeastern Colorado, 1986-90. �175 Twenty percent (8 of 41) of leks active in 1986 were inactive in 1987, 29% (12 of 42) of leks active in 1987 were inactive in 1988, 28% (13 of 47) of leks active in 1988 were inactive in 1989, and 19% (8 of 42) of leks active in 1989 were inactive in 1990 (Fig. 8-5). lek disappearance rate was 23.8%. The annual Twenty-four leks were active all 5 years, 3 for 4 years, 10 for 3 years, 12 for 2 years, and 27 for 1 year (Fig. 8-6). Over the limited duration of this study, the number of permanent leks (active every year) appeared to approach an asymptote of over 20 leks. Likewise, the number of additional leks (leks not previously described) appeared to be approaching an asymptote indicating a finite number of possible lek locations on the study area. Male greater prairie-chickens ~ere on leks an average of 94.5% of the time they were located during normal periods of lek activity (n - 13 males with an average of >20.observations/bird periods)(Fig. B-7). during normal display Values ranged from 86.7% to 100.0%. Although 5 males captured during winter (X - 96.8%, n - 5) tended to have slightly higher lek attendance rates than males captured on leks (X - 93.1%, n = 8), differences in rates of lek attendance by individual males were not detected (~ - 10.647; f - 0.559). When lek attendance for radio-marked males was examined independent of individual differences, the attendance rate was 95.1% for 286 observations of 21 males. Males captured during winter (96.9%) tended to have slightly higher (~ - 2.719, f - 0.099) lek attendance rates than males captured on leks (92.7%). Although, there were no apparent differences associated with age (1 = 0.910, f • 0.382), yearlings apparently visited more leks during the breeding season than adults. Three of 4 yearling males visited> 1 �176 II4IW 14N N ---.---.-.~. ~:--.. --~: SIZE OF LEKS .9 TaN 2-5 MALES o 6-10 MALES o 11-15 MALES o >15 MALES o 2 II4IW I UN ~.~.~:~ TaN TIN 41<M , II•• W T4N T4N 9 UN TaN TIN 9 1987 WllAY 1988 WllAY T4N T4N ----------r.-""-'!"""""-""",,, ~ TIN .TIN 1989 II4IW Fig. 8-5. II•• W II4IW II•• W Distribution of greater prairie-chicken leks on the study area in northeastern Colorado~ 1986-90. �177 -•••o g;!1 ..• <2 -- ...... - en TEMPORARY L5 >3 .. .•. a: ,, LL .•. 1986 - ... - --- --- 1987 -< . ,, 1988 ~ :::c 0 ffi4 r::c PERSISTENT 1989 :E ::> Z5 1990 o 10 20 30 40 NUMBER OF LEKS Fig. 8-6. Stability of leks (histogram) and number of persistent (active all previous years of study) and additional leks (not previously described) on the study area in northeastern Colorado during 1986-90. �.lHS 4 rum CAPTURED ON LEKS [J CAPTURED IN WINTER C1J3 W ...J -c ::is LL °2 a: W to ::is ::> Z1 o >86-88 >88-90 >90-92 >92-94 >94-96 >96-98 >98-100 LEK ATTENDANCE (%) Fig. 8-7. Frequency of lek attendance by 8 male greater prairie- chickens captured on leks and 5 males captured at winter feeding sites in northeastern Colorado during 1986-90. �179 lek while only 1 of 17 adults visited> 1 lek ()( - 10.032, f - 0.002). In addition, 1 yearling was observed on 4 leks while another was observed on 6 leks. same morning. The latter male also was observed on 2 leks on the In only 1 case was a radio-marked yearling male not observed on a lek during the normal morning display period. DISCUSSION Examinations of lek stability indicate the possibility that a finite number of permanent leks may exist on the study area. there may be a fin ite number of potent iall ek 1ocat ions. information desirable is useful for numerous reasons. Likewise, Th is Permanent leks may be more for territory acquisition than temporary leks. Lewis and Zwickel (1980, 1981i, h) determined that yearling male blue grouse (Dend~agapus obscurus) spend more time near permanent territories. Whether selection of leks is due to the status of dominant males (Beehler and Foster 1988), accessibility to females (Lewis and Zwickel 1981a, Bradbury and Gibson 1983), or habitat quality (Wrangham 1980) has not been adequately examined. territory desirability, In addition to problems associated with permanent leks may be more consistent with respect to male attendance than temporary leks. Differences in attendance may be important if permanent leks are easier to find than temporary leks. This latter consideration could affect estimates of both lek density and male attendance. Yearling males appeared to have visited more leks than adults. addition to the unusual pattern of lek visitation by yearling males, yearlings also visited many temporary leks. specific evidence of age composition Although there was no at temporary leks, it was possible In �they were generally composed of males unable to obtain territories on permanent leks (Cannon and Knopf 1981). more likely to be yearlings. Presumably these birds were Although multiple lek visitations were observed in other populations, ages of males were rarely known ·(Hamerstrom and Hamerstrom 1949, Arthaud 1968, Silvy 1968, Robel et ale 1970, Robel and Ballard 1974). Rippin and Boag (1974) found that removals of territorial male sharp-tailed grouse from leks resulted in an influx of new males; they suggested that these newly established males were more likely yearlings that had previcrusly been non-territorial lek. Whether these males visited> in areas associated with the 1 lek during the breeding season or occupied territories on temporary leks was not known. Although numbers of males on leks may increase at the same time numbers of leks are increasing, no spatial correlation in yearly changes in male attendance at 1eks was found. The 1ack of a corre 1at ion may be . caused by interactions among possible factors associated with lek dynamics including: 1) formation of smaller temporary leks near larger permanent leks, 2) relative consistency in numbers of males at permanent versus temporary leks, 3) survival, mortality~ and/or movement of 'hotshot' males (Beehler and Foster 1988), 4) regional fluctuations in abundance of males, 5) compensatory changes in abundance of males at neighboring leks (if 1 lek increases, an adjacent lek may decline), and/or 6) maintenance of genetic heterozygosity. Lek visitation have showed similar variability. visitation of greater prairie-chickens and 50% (Robel 1970). Estimates of lek ranged between 95% (this study) Most variation in lek attendance my be due to time of year (Bradbury et al. 1989). Visitation rates of sage grouse �181 were> 90% during the period of peak male attendance at leks and between 67% (yearlings) and 100% (adults) during the peak of hen attendance (Emmons and S'raun 1984). These results for male lek attendance pertain to serious issues concerning the census of greater prairie-chickens grouse. and other prairie Robel's (1970) paper has been cited as evidence that the actual number of males may be twice as high as the number counted on leks. However, the data in this study indicated that most males (about 95%), may be attending leks every day. This discrepancy indicates that population size would be over-estimated when lek counts and visitation rates of 50% are used. Robel (1970) estimated a value for lek attendance by male greater prairie-chickens of 50% by dividing 33 late-April territory-holders into 60 males thought to be on the area during early March (actually 55%). There are several problems which may account for the differences between Robel's estimate and the lek attendance rate observed in this study. First, sUbstantial mortality could have occurred among the estimated 27 males (60 minus 33) that disappeared between early March and late April. Second, many males that disappeared may have dispersed from the study area, which was limited to 3 leks. Likewise, the presence and status of leks around the periphery of the study area was not documented making it difficult to quantify the status of the males that disappeared. Finally, radio telemetry would have made it possible to determine whether males had dispersed. from the study area, were on or off leks, or were dead. In addition to lek visitation, methods of surveying lek attendance of prairie grouse should be evaluated. Accurate counts of males may be �· .LO,," easier to obtain when birds are flushed than when they are observed from a distance. Some males may be off the edges of leks, and hence, difficult to observe. The methodology for estimating numbers of males could be of critical importance if rates of lek visitation are used to estimate population size. LITERATURE CITED Ammann, G. A. 1944. Determining age of pinnated and sharp-tailed grouses. J. Wildl. Manage. 8:170-171. Amstrup, S. C. 1980. 44:214-217. A radio-collar for game birds. J. Wildl. Manage. Arthaud, F. L.· 1968. Populations of the prairie chicken related to land use in southwestern Missouri. M.A. TheSiS, Univ. Missouri, Columbia. 134pp. Beehler, B. M., and M. S. Foster. 1988. Hotshots, hotspots, and female preference in the .organization of lek mating systems. Am. Nat. 131:203-219. Bradbury, J. W., and R. M. Gibson. 1983. Leks and mate choice. Pages 109-138 in P. Bateson, ed. Mate choice. Cambridge Univ. Press, Cambridge, U.K. ___ , S. L. Vehrencamp, and R. M. Gibson. 1989. Dispersion of displaying male sage grouse. I. Patterns of temporal variation. Behav. Ecol. Sociobiol. 24:1-14. Cannon, R. W., and F. L. Knopf. 1981. Lek numbers as a trend index to prairie grouse populations. J. Wildl. Manage. 45:776-778. Emmons, S. R., and C. E. Braun. 1984. Lek attendance of male sage grousew J. Wildl. Manage. 48:1023-1028. Hamerstrom, F. N., and F. Hamerstrom. 1949. Daily and seasonal movements of Wisconsin prairie chickens. Auk 66:313-337. ___ , and . 1973. The prairie chicken in Wisconsin. Nat. Resour. Tech. Bull. 64. 52pp. Wis. Dep. Lewis, R. A., and F. C. Zwickel. 1980. Removal and replacement of male blue grouse on persistent and transient territorial sites. Can. J. Zool. 58:1417-1423. , and . 1981A. Differential use of territorial sites by ---male blue grouse. Condor 83:171-176. �183 ___ , and _~_. 1981.12. Survival and delayed breeding in male blue grouse. Can. J. Zool. 60:1881-1884. Miller, G. C. 1984. Development of a preservation program for insular populations of prairie grouse. Colorado Div. Wildl. Fed. Aid Rep. N-I-R. Jan. Pages 129-170. Rippin, A. B., and D. A. Boag. 1974. Recruitment to populations of male sharp-tailed grouse. J. Wildl. Manage. 38:616-621. Robel, R. J. 1970. Possible role of behavior in regulating greater prair)e chicken populations. J. Wildl. Manage. 34:306-312. ___ .'J. N. Briggs, J. J. Cebula, N. J. Silvy, C. E. Viers, and P. G. Watt. 1970. Greater prairie chicken ranges, movements, and habitat usage in Kansas. J. Wildl. Manage. 34:286-306. ___ , and W. B. Ballard, Jr. 1974. lek social organization and reproductive success in the greater prairie chicken. Amer. Zool. 14:121-128. Silvy, N. J. 1968. Movements, monthly ranges, reproductive behavior, and mortality of radio-tagged greater prairie chickens (Tympanuchus cupido pinnatus). M.Sc. Thesis, Kansas State. Univ., Manhattan. 135pp. Wrangham, R. W. 1980. Female choice of least costly males; a possible factor in the evolution of leks. Z. Tierpsychol. 54:357-367. Prepared by _'l1LJJ~~':"=;..Jo..._~~_":_'..::;:~::;;;';;,_:;_--=:.._;_=-__ Michael A. Schroeder Graduate Research Assistant Approved by _..;;;;..a.....;;~;..;..;....__. _Z_. ....I.~-=--.;;____;;;_ Clait E. Braun Wildlife Research Leader _ �\j �185 JOB PROGRESS REPORT Colorado State of: Project: W-ls2-R Upland Bird Research Work Plan: 17 Job Title: Population Dynamics of White-tailed Ptarmigan Period Covered: Author: Job 7 _ 01 January through 31 December 1990 Clait E Braun and Kenneth M. Giesen Personnel: Kathy Martin, University of Toronto; Clait E. Braun and Kenneth M. Giesen, Colorado Division of Wildlife ABSTRACT Long-term studies of populations of white-tailed ptarmigan (Lagopus leucurus) were continued at hunted (Mt. Evans) and_unhunted (Rocky Mountain National Park) areas in Colorado through 1990. Densities .of breeding ptarmigan increased at Rocky Mountain National Park and were stable at.Mt. Evans. Nest success at Mt. Eyans was average (50%) and was below avera~e (40%) at Rocky Mountain National Park. At least 7% of the ptarmigan banded in 1990 at Mt. Evans were harvested during the 1990 hunting season. Nonreporting of bands may be an important problem. ��187 POPULATION DYNAMICS OF WHITE-TAILED PTARMIGAN C1ait E. Braun and Kenneth M. Giesen Long-term studies of trends in population size and investigation of reasons for fluctuations in size of tetraonid populations are lacking. Studies on the population dynamics of unhunted and hunted populations of white-tailed ptarmigan were initiated in Colorado in 1966 and have continued essentially uninterrupted at 2 sites. Studies of the unhunted population (Rocky Mountain National Park) identified possible short-term cycles of 7-8 years with an amplitude of 25-30% between high and low breeding densities. Conversely, studies of the manipulated population (hunted) at Mt. Evans have not indicated any cyclic pattern and it would appear that controlled hunting may mask any long-term trend that may occur. This study is designed to examine the question whether white-tailed ptarmigan are truly cyclic and whether hunting affects the apparent oscillations. P. N. OBJECTIVES The goals of this investigation are to be able to predict the length and amplitude of cycles in white-tailed ptarmigan in Colorado, to examine the impact of hunting on cycles, and to clarify underlying causes of the apparent cycles. SEGMENT OBJECTIVES 1. Conduct breeding (May-Jun) and brood (Aug-Sep) censuses of white~tailed ptarmigan using tape-recorded calls of males (breeding) and chicks (broods). 2. Censuses will be conducted on previously established, defined study areas at Mt. Evans (hunted) and at Rocky Mountain National Park (unhunted). 3. Capture (noose poles) and band (aluminum and plastic color-coded bands) all unmarked white-tailed ptarmigan encountered on study areas at Mt. Evans and at Rocky Mountain National Park. 4. Individually identify all ptarmigan observed on study areas at Mt. Evans and Rocky Mountain National Park through use of binoculars. 5. Make hunting season and bag limit recommendations for Mt. Evans and collect hunting data through use of volunteer wing barrels and hunter field checks. 6. Compile data, analyze results, and prepare progress reports. �188 STUDY AREA AND METHODS Areas investigated were Mt. Goliath-Mt. Evans in Clear Creek County and at Tombstone Ridge-Sundance Mountain to Fall River Pass in Rocky Mountain National Park in Larimer County. The physiography, geology, location, and vegetation of these study areas have been previously described (Braun 1969, 1971; Braun and Rogers 1971; Giesen 1977). Ptarmigan were located through use of tape-recorded calls (Braun et al. 1973), captured through use of telescoping noose poles (Zwickel and Bendell 1967) as described by Braun and Rogers (1971), and classified to age and sex and banded following Braun and Rogers (1971). Age of chicks was estimated following Giesen and Braun (1979). Numbered plastic bandettes were not used as in earlier years (Braun and Rogers 1971) as a color-code system using up to 4 different colored plastic bandettes was instituted in 1977-78. A check station was operated on the Mt. Evans highway during the opening weekend of the ptarmigan season in that area. A volunteer wing collection station WaS available to hunters in the area when the check station was not in operation until the season closed. RESULTS AND DISCUSSION Breeding Densities Mt. Evans.--Timing of breeding events in the Mt. Evans area was about the same in 1990 as in 1989. During the May-early June interval, 16 pairs were identified. Thus, breeding densities remained essentially unchanged from 1989 (Table 1). During the breeding season, 2 of 16 males identified were yearlings while 5 of 16 hens were yearlings. Recruitment of yearlings was lower than in 1989. Rocky Mountain National Park.--Timing of breeding events on the Trail Ridge study area was similar to the long term average and about 1 week later than in 1989. Surveys of ptarmigan on breeding territories along Trail Ridge Road in May and June indicated a minimum population of 75 birds and included 31 pairs and 13 unpaired males. This represents a 10% increase over 1989 and is the highest population since 1981. However, this population level is less than 70% of the breeding densities recorded during peak years in 1969 and 1976. The increased breeding density reflected above average survival of banded adult males (37 of 60, 61.7%) and females (17 of 35, 48.6%) from 1989. Survival of banded chicks was 20% (2 of 10) but yearlings comprised 26% of all adult ptarmigan identified in·1990. Mt. Evans.--Thirty hens were located during mid July-early September 1990 on or immediately adjacent to the study area. Fifteen hens (50.0%) were with broods while 15 were apparently unsuccessful nesters (without chicks). Average brood size to 1 September was poor (2.1 chicksjhen). Hatch dates varied from early to late July. Rocky Mountain National Park.--Nest success was estimated from the proportion of hens with broods and broodless hens observed in July and August. Seven of 18 hens observed during summer surveys were with broods for an estimated nest success rate of 40%. However, 4 hens (radio-marked) observed without chicks were known to have hatched clutches but lost all chicks soon after hatch. The �189 peak of hatch occurred during the 2nd and 3rd weeks of July. size in August was 4.2 chicksjhen. Table 1. 1966-90. Average brood White-tailed ptarmigan breeding densities (birdsjkm2), Colorado Study area Year 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 .1982 1983 1984 1985 1986 1987 1988 1989 1990 Rocky Mountain National Park (5.5 km2) Mt. Evans (4.0 km2) 11.3 9.8 11.5 12.0 9.6 9.1 8.7 7.8 8.0 11.1 13.5 12.9 10.7 8.7 8.4 8.2 7.8 6.7 5.8 6.0 4.5 6.0 5.4 6.2 7.6 3.0 2.7 2.7 2.2 2.0 4.2 7.5 6.2 6.2 6.2 6.7 > 6.0 7.5 10.3 9.5 9.0 6.5 6.5 8.0 8.0 6.5 5.0 7.5 8.0 8.0 Harvest Mt. Evans.--The hunting season at Mt. Evans in 1990 opened on 15 September and closed on 7 October (23 days) with a bag and possession limit of 3 and 6. Thus, the season was delayed 1 week from the statewide opening as it was in 1981 and 1986-89. The season opening was delayed 2 weeks from 1978 to 1980 and 1982 to 1985. Prior to 1978, experimental seasons were in effect (19701976) or the season opened with the statewide grouse seasons (dates from 17 Aug to 14 Sep). Unlike 1987 and 1988 when the Mt. Evans road was closed due to reconstruction (Lincoln Lake area in 1987, above Ptarmigan Flats in 1988), the Mt. Evans road was open tb Summit Lake from 15 to 28 September when it �190 closed because of snow. Twenty hunters were checked on 15-16 September. These hunters observed 50 ptarmigan and harvested 13 of which 12 were banded. Seven additional wings (no bands) were received on 21 and 23 September in the wing barrel. Thus, a minimum of 20 ptarmigan were harvested at Mt. Evans of which at least 12 were banded. Five of the banded birds harvested were banded in 1990 (5/69 - 7.2%) of which 4 were chicks (4/24 - 16.7%). The other banded birds harvested were banded in 1986 (1), 1987 (1), 1988 (1), and 1989 (4). LITERATURE CITED Braun, C. E. 1969. Population dynamics, habitat, and movements of whitetailed ptarmigan in Colorado. Ph.D. Thesis, Colorado State Univ., Fort Collins. l89pp. 1971. Habitat requirements of Colorado white-tailed ptarmigan. West. Assoc. State Game and Fish Comm. 51:284-292. Proc. _____ , and G. E. Rogers. 1971. The white-tailed ptarmigan in Colorado. Colorado Div. Game, Fish and Parks Tech. Publ. 27. 80pp. _______ , R. K. Schmidt, Jr., and G. E. Rogers. 1973. Census of Colorado whitetailed ptarmigan with tape recorded calls. J. Wild1. Manage. 37:90-93. Giesen, K. M. 1977. Mortality and dispersal of juvenile white-tailed ptarmigan. M.S. Thesis, Colorado State Univ., Fort Collins. 5Spp. _______ , and C. E. Braun. 1979. A technique for age determination of juvenile white-tailed ptarmigan. J. Wild1. Manage. 43:508-511. Zwickel, F. C., and J. F. Bendell. J. Wildl. Manage. 31:202-204. Prepared by Clait E.'Braun Wildlife Research Leader Kenneth M. Giesen Wildlife Researcher C 1967. A snare for capturing blue grouse. �JOB PROGRESS REPORT Colorado State of: Project: W-152-R Upland Bird Research Work Plan: 21 Job Title: Sandsage-B1uestem Period Covered: Author: Job _3_ 01 January Prairie Renovation through 31 December 1990 Warren D. Snyder Personnel: C. Braun, L. Crooks, T. Davis, T. Remington, M. Schroeder, Wilson, and W. Snyder, Colorado Division of Wildlife D. ABSTRACT Except for June, which was hot and dry, precipitation in 1990 was above average and conducive to growth of rangeland vegetation. Below average precipitation in 1989, however, was reflected in lower height-density indices (HDI) of residual grass-forb vegetation in early spring 1990. The data confirm that impacts of fire on grass-forb vegetation (either positive or negative) were relatively short-term. Effects of treatments conducted from 1984 through 1986 were not sustained intq 1990. Sand sagebrush (Artemisia filifo1ia) ~ecovered more slowly from fire and differences between treatments and controls were still noted in 1990. Prairie sandreed (Calamovilfa longifolia), especially, and other warm-season grasses, to some extent, were suppressed by hot, dry weather in June, 1990. There was no evidence that sand dropseed (Sporobolus cryptandrus) was impacted by fire. Limited data indicated that sand b1uestem (Andropogon halli) was enhanced by fire and consistently increased under the ungrazed management of the Tamarack Prairie. Neither perennial forbs nor annual forbs showed evidence of being impacted by fire. The height-density of switchgrass (Panicum virgatum) dominated strips, revegetated in 1985, declined slightly in 1990 due to decreased precipitation in 1989. However, the me,an HD;r of 3.6 dm was much greater than that of other vegetation within the TairiarackPrairie. Within the t~l1age-herbicide renovation tract, treated in 1986, sustained enhancement of switchgrass': b1uestem vegetation continued. Standing residual of sand sagebrush, killed by herbicide in 1985 within an east Tamarack Prairie site, deteriorated markedly from the previous year. Monitoring of greater prairie-chickens (Tympanuchus cupido) revealed deterioration of 2 established leks'on private lands near the southern edge of the Tamarack Prairie; establishment of a small lek at a mowed site on the property; and discovery of a new 1ek about 4.5 kID south of the Tamarack Prairie. Increased numbers of sharp-tailed grouse (I. phasianellus) were evident in spring 1990 and several hybrid crosses of the 2 species were observed at the new lek. Management personnel trapped 23 greater prairie chickens in Yuma County in April 1990 and released them on the east part of the Tamarack Prairie. Among 4 radio-marked hens that were released, 2 nested �and reared young on the property. Two of 4 radio-marked transplanted males survived and became active respectively at leks land 2. One-third of the 12 radio-marked hens (resident and transplanted) were lost to predation, however, 7 of 8 surviving hens nested successfully. This sample included 1 female sharp-tailed grouse. Brood survival through summer was not accurately assessed. �193 SANDSAGE-BLUESTEM PRAIRIE RENOVATION Warren D. Snyder P. N. OBJECTIVES Test renovation and revegetation techniques for increasing standing residual height-density of grasses, the proportion of tall warm-season grasses within the composition, and for reducing the quantity of sand sagebrush to <30% canopy cover in an ungrazed sandsage-bluestem prairie on the South Tamarack, South Platte Wildlife Area in northeastern Colorado. SEGMENT OBJECTIVES Monitoring follows: of environmental and vegetation conditions and changes continued as 1. Precipitation was monitored throughout the year supplementing data from 4 electronic rain gauges with information from nearby weather stations through the winter months. 2. Soil moisture accumulations, plant phenology, and weather were monitored primarily in spring and especially at the time of controlled burns. 3. Visual obstruction (height-density) measurements were obtained on treatments and controls where applicable in late winter and/or early spring prior to green-up. 4. Crown cover, species composition, and frequency of occurrence measurements were obtained from mid-summer to early fall. 5. Male and female greater prairie-chickens were trapped on leks adjacent to the Tamarack Prairie in April 1990 and equipped with battery-powered (7 g) transmitters. Their survival, movements, use of habitat types for nesting, feeding and loafing, and use of the Tamarack Prairie was monitored through spring, summer, and fall. 6. Data compilation and writing the annual job progress conducted during winter 1990-91. report were METHODS Approaches used were described by Snyder (1986~, 1986Q, 1987, 1988,' 1990) and are outlined in the Segment Objectives. Segment Objective #4 was conducted within the 1985 and 1986 burn sites and their controls after not being conducted in 1989, and was deleted within the 1984 burns. Walk-in traps were placed on leks near the Tamarack Prairie in early April 1990 to trap greater prairie-chickens, plains sharp-tailed grouse, and hybrids of the 2 species. Trapped birds were marked with numbered aluminum leg bands �and color-marked with plastic bandettes. Poncho-mounted, battery-powered transmitters were placed on 3 males (2 were sharp-tailed grouse) and 8 females (1 sharp-tailed grouse) resident in the area. Transmitters were also placed on 4 males -and 4 females among 23 greater prairie-chickens released on the Tamarack Prairie in April 1990. Additional males trapped on the leks were banded and released. RESULTS AND DISCUSSION Environmental Conditions Precipitation during the first few months of 1990 was slightly above average (Fig. 1) which increased soil moisture in comparison to spring 1989. Soil probe indices in early and late April and on 24 May remained nearly constant (Table 1) averaging 9.04 dm. This was almost twice the depth (4.67 dm) of probes recorded in 1989. • .JAN-MAR illIIll em X API=! MAY - 84 JUN 85 JUL U1 86 0: « W >- • AUG 87 SEP 88 • OCT-DEC 89 gO a 2 4 6 8 10 12 PRECIPITATION 14 16 18 20 22 (INCHES) Fig. 1. Monthly and annual precipitation (in.) from 1984 through 1990 in relation to the long-term mean, Tamarack Prairie, Colorado. �195· Table 1. Soil moisture accumulations (dm)a based on soil probe samples during spring 1990, Tamarack Prairie, Colorado. AJ2r Location West North-central South-central East ~ 24 30 2 8.95 10.35 12.13 7.62 8.51 8.95 12.32 6.82 aMean depth (dm) of 4 probes/location precipitation gauges. 8.06 8.32 10.22 6.22 taken near each of the 4 Precipitation during the first 3 months of 1990 was above average whereas it was below average in April, near average in May, and markedly below average in June (Fig. 1). Mean-temperatures in April and May remained below average, but June was exceptionally hot, especially during the last 2 weeks of the month (Fig. 2). Temperatures moderated in July, which was one of the wettest months recorded during the study. Precipitation continued to remain slightly above average through the late summer, fall, and early winter. Total precipitation in 1990 was above average and comparable to that received in 1987. Phenological conditions during April and May 1990 were monitored as in previous years (Table 2). Vegetation growth and development appeared similar to that recorded in 1989 although temperatures averaged slightly cooler (Fig. 2). A hard freez~ on 1 May 1989 accompanied by dry soils possibly retarded vegetation development that spring in contrast to 1990. ~ e4 85 0: -<: 111111111 A~~ [::::::::::::::1 MAY t::~:;:::;:::;:::;:::;::1 ..JUN 8S L.1..J >- 87 ee 89 90 30 35 40 45 SO 55 60 TEMPERATUI=lE 65 70 75 eo C F) Fig. 2. Monthly average temperature (F) from March through June, 1984-90 in relation to the long-term mean, as an index to vegetation phenology, Sterling, Colorado. �Table 2. Phenological conditions of selected vegetation during spring 1990, Tamarack Prairie, Colorado. Species Alliun textile Artemisia filifolia A. ludoviciana Astragalus sp. Cymopteris mont anus Erigeron bellidiastrun Lath:z:ruspolymorphus Leucocrinun montanun Mentzelia nuda Penstemon angustifolius Phlox andicola PSOralea lanceolata Sphaeralcea coccinea Tradescantia occidentalis Tragopogan sp. Agrop:z:ronsmithii Bouteloua gracilis Calamovilfa longifolia Panicun virgatun Paspalum stramineum Sporobolus cr:z:ptandrus Stipa ~ C:z:perussp. avegetation height (em). E = early, M = medium, L A r 19 2 5a budded E bloom F bloom emerging 5 emerging 5-7.6 leafing 5 late, F E bloom bloom leafed bloom 2.5 7.6-10 5 E bloom 15-18 emerging 10-13 bloom = full, 8 F blooni 5-7 leaf 5 P 2.5 15-18 = 21 4 L bloom 5-7.6 bloom L seedhead F bloom L bloom L bloom 5-10 7.6 20-30 F bloom 15-18 5 10-15 10 L bloom 5-7.6 E bloom 30-38 F bloom 20-25 5-7.6 15-20 15 P bloom 10-15 F bloom bloom 5 20-25 15 E head E bloom F bloom 5-7.6 7.6-10 emerging 5-7.6 2.5 = budded 2.5 leaf leafing 5 5-7.6 Jun Ma:z: 27 F bloom 12.7 25-30 7.6-10 15 headed past. Height-Density Sampling Within Burned Sites Below average precipitation in 1989 was the apparent reason that heightdensity indices (HDI) of residual grass-forb vegetation, in early spring 1990, were below those of the previous year. Within burns 1-84 and 3-84, treatment and control HDI transects have shown similar trends from 1987 through 1990 (Figs. 3 and 4, Tables 3 and 4) indicating there was no enhancement of vegetation growth after the initial (1984-86) suppression by fire. The HDI of grass-forb vegetation within the 3-84 burn remained suppressed based on pretreatment H~I levels (Fig. 4). Two possible explanations exist. Initial HDI samples may have been erroneous as sample size was small within the 3-84 burn. It is also possible that fire within burn 3-84, which was hotter than fire in burn 1-84, may have damaged vegetation to the extent that it was not able to completely recover. Sand sagebrush within the 1-84 and 3-84 burns recovered more slowly than grasses. Height-density indices were fully recovered by 1990 within burn 1-84 (Fig. 5), but within burn 3-84, treatment indices remained below those in controls (Fig. 6). Possibly, sage was more severely impacted by fire in burn 3-84. Grass-forb vegetation, based on HDI, recovered by the season after treatment within the 3 1985 burns (Table More rapid recovery was noted within the 3 1986 burns Their regrowth, in contrast to slower recovery within end of the 2nd growing 5, Fig. 7). (Table 6, Fig. 8). the 1984 burns, was �197 0.7 0.6 1\ E U u O.S >I(f) 0.4 CONTROL Z W 0 I 0.3 lI 19 0.2 W I 0.1 BURN a ~----~--~--------~--~----._ 84 85 86 87 88 89 90 YEAR Fig. 3. Height-density (dm) of residual grass-forb vegetation from 1984 (pretreatment) to 1985-90.(post-treatment) within the 1-84 burn and control sites, Tamarack Prairie, Colorado. / / \ \ \ > 85 86 87 88 89 90 YEAP Fig. 4. Height-density (dm) of residual grass-forb vegetation from 1984 (pretreatment) to 1985-90 (post-treatment) within the 3-84 burn and control sites, Tamarack Prairie, Colorado. �Table 3. Mean height-density (dm) within burn 1-84 and controls during spring 1984-90, Tamarack Prairie, Colorado. Years Grass/fb Burn Sandsage Combined 1984a .1985 1986 1987 1988 1989 1990 0.256 0.134 0.232 0.208 0.589 0.527 0.434 0.856 0.313 0.358 0.526 0.836 0.753 1.011 0.372 0.162 0.256 0.258 0.627 0.577 0.540 E 1984-90 1989-90 0.253 0.295 0.301 0.224 0.587 0.571 0.430 0.814 0.687 0.643 0.847 1.045 1.052 0.893 Combined 0.334 0.355 0.368 0.309 0.622 0.628 0.475 Values 1.12 0.54 0.01 0.65 Control Grass/fb Sandsage 0.79 3.25 apretreatment samples. Table 4. Mean height-density (dm) within Burn 3-84 and controls during spring 1984-90, Tamarack Prairie, Colorado. Years Grass/fb Burn Sandsage 1984a 1985 1986 1987 1988 1989 1990 0.222 0.021 0.106 0.077 0.466 0.647 0.378 0.827 0.121 0.356 0.286 1.215 1.435 1.068 E 1984-90 1989-90 2.54 0.04 apretreatment. Combined " 0.493 0.047 0.201 0.157 0.831 1.159 0.680 Values 0.88 0.58 2.26 0.86 Grass/fb Control Sands age Combined 0.183 0.191 0.200 0.216 0.625 0.934 0.446 0.935 0.797 1.044 0.688 1.494 1.888 1.579 0.531 0.503 0.629 0.424 1.059 1.548 0.984 �199 1.1 1.0 1\ E 0.9 U CONTROL U >- 0.8 I/) 0.7 " ,,l· ,, l•... )(,- Z W 0 , , , , , , -_.J,/" 0.6 I lI 0.5 W o.~ 19 I 8URN 0.3 0.2 84 8S 86 87 88 89 90 YEAR Fig. 5. Height-density (drn) of sand sagebrush from 1984 (pretreatment) to 1985-90 (post-treatment) within the 1-84 burn and control sites, Tamarack Prairie; Colorado. 2.0 1. B 'E' -0 '--.J >- I- 1.6 1.4 1.2 U1 :z: L.l.J 1.0 Cl I I- :c t.:J L.l.J :c 0.8 0.6 0.4 0.:2 0 S""I 8:5 86 87 88 89 90 YEAR Fig. 6. Height-density (drn) of sand sagebrush from 1984 (pretreatment) to 1985-90 (post-treatment) within the 3-84 burn and control sites, Tamarack Prairie, Colorado. �200 Table 5. Colorado. Mean height-density (dm) within 1985 burns and controls during spring 1985-90, Tamarack Prairie, ControL Burn Year 2 3 2 3 ! 0.374 0.325 0.322 0.511 0.774 0.423 0.346 0.280 0.376 0.775 1.220 0.667 0.158 0.180 0.156 0.443 0.771 0.316 0.323 0.277 0.325 0.644 1.025 0.529 0.8n 0.828 0.925 1.070 1.242 1.443 0.615 0.745 0.560 1.732 1.873 1.730 0.537 0.767 0.648 1.253 1.844 1.267 0.627 0.771 0.670 1.317 1.765 1.414 0.444 0.391 0.383 0.581 0.844 0.535 0.369 0.335 0.393 0.859 1.332 0.750 0.323 0.465 0.411 0.878 1.609 0.661 0.380 0.382 0.394 0.782 1.254 0.666 ! GRASS-FORB 1985a 1986 1987 1988 1989 1990 0.507 0.138 0.419 0.780 1.081 0.542 0.381 0.119 0.4n 1.028 1.297 0.552 0.180 0.053 0.244 0.667 0.850 0.315 0.379 0.110 0.408 0.877 1.142 0.495 SANDSAGE 1985a 1.008 0.417 0.523 0.938 1.313 0.688 1986 1987 1988 1989 1990 0.679 0.342 0.591 1.167 1.582 1.050 0.642 0.438 0.769 1.144 1.385 0.955 0.744 0.388 0.6n 1.125 1.446 0.914 COMBINED 1985a 1986 1987 1988 1989 1990 0.565 0.142 0.425 0.786 1.109 0.548 0.407 0.130 0.476 1.037 1.340 0.567 0.302 0.088 0.330 0.766 1.017 0.359 F VALUE 1989-90 1985-90 Grass-forb Conbined apretreatment. 2.56 10.95 1.36 8.75 0.429 0.124 0.426 0.899 1.196 0.513 �201 Table 6. Colorado. Mean height-density (dm) within 1986 burns and controls during spring 1985-90, Tamarack Prairie, Burn Year 2 Control 3 ! 2 3 ! GRASS-FORB 1985a 0.326 0.314 0.259 0.292 0.294 0.316 0.233 0.275 1986a 1987 1988 1989 1990 0.304 0.120 0.543 0.n3 0.386 0.337 0.437 1.359 1.159 0.666 0.256 0.202 0.824 1.019 0.361 0.290 0.226 0.855 0.963 0.438 0.277 0.234 0.387 0.794 0.360 0.312 0.404 0.858 1.141 0.561 0.213 0.269 0.644 1.074 0.432 0.265 0.289 0.600 0.983 0.435 0.829 0.756 0.885 1.265 1.353 1.557 0.417 0.300 0.833 1.667 1.250 1.000 0.618 0.690 0.671 1.197 1.810 1.343 0.682 0.693 0.750 1.237 1.650 1.430 0.345 0.342 0.325 0.494 0.886 0.454 0.317 0.331 0.413 0.873 1.145 0.564 0.298 0.323 0.357 0.743 1.304 0.509 0.320 0.332 0.357 0.679 1.116 0.500 SANDSAGE 1985 1986 1987 1988 1989 1990 0.621 0.838 0.169 0.904 1.047 0.860 0.375 0.500 0.250 1.000 1.250 0.395 0.650 0.250 1.194 0.857 0.600 0.566 0.793 0.189 0.945 1.007 0.852 COMBINED 1985 1986 1987 1988 1989 1990 0.421 0.470 0.127 0.616 0.815 0.486 Grass-forb Combined 0.315 0.339 0.434 1.355 1.159 0.473 0.2n 0.281 0.204 0.834 1.008 0.365 1985-90 F VALUE 1989-90 0.41 0.57 0.09 0.07 1985 and 1986 were both pretreatment years. 3 0.335 0.367 0.224 0.863 0.969 0.474 �202 1.2 1.1 1'"\ e 1.0 "0 0.9 >I- 0.8 (/l Z W 0.7 .,!. D.S o 0." v 0 0.6 :::t W :::t D.l 0.2 0.1 0 85 86 87 S8 89 90 YEAR Fig. 7. Height-density (dm) of residual grass-forb vegetation from 1985 (pretreatment) to 1986-90 (post-treatment) within the 1985 burn and control sites, Tamarack Prairie, Colorado. 1.0 D.9 O.B 'E' '0 V 0.7 ~ 0.6 VI Z W 0 O.S I I- ::I: 0.4 W 0.3 l!) ::I: 0.:2 0.1 0 IS 88 8B 87 89 90 YEAR Fig. 8. Height-density (dm) of residual grass-forb vegetation from 1985-86 (pretreatment) to 1987-90 (post-treatment) within the 1986 combined burn and control sites, Tamarack Prairie, Colorado. �203 attributed to greater precipitation subsequent to treatment. All grass-forb HDI's declined from 1989 to 1990 in direct relation to respective precipitation amounts (Fig. 8). The impacts of fire did not persist. Comparison of HDI's among the 4 controls showed similar trends for all except the 1-84 site in 1989 (Fig. 9). This was attributed to a local precipitation deficiency there in 1988 (Snyder 1990). 1.2 1.1 1985 1.0 1\ E lJ 0.9 U >I:- 0.8 0.7 (f) Z W 0 0.6 1986 I lI o 0.4 W I 0.3 0.2 "- 0.1 3-84 a 84 85 86 87 88 89 90 YEAR Fig. 9. Height-density (dm) of residual grass-forb vegetation among 1984 through 1986 controls from 1984 through 1990, Tamarack Prairie, Colorado .. Sand sagebrush within the 1985 and 1986 burns, like that in the 1984 burns, recovered more slowly than other herbaceous vegetation (Tables 5 and 6). However, comp arLs on among years showed sage recovered more rapidly in the 1985 and 1986 burns than in 1984 burns, apparently in response to increased precipitation. An initial rapid recovery was noted, but subsequent growth was slower. Three of the 4 control groups showed similar sage growth patterns from 1987 through 1990 (Fig. 10). However, growth of sage within burn 1-84, like that of grass-forb vegetation, was markedly reduced in 1989 due to deficient precipitation in 1988. Sandsage, within the 1986 burns, showed little recovery in 1990 from the previous year (Table 6). �ZU4 2.0 1\ 'l'-. '\ 'f~. '\ r I (\ E 1.6 I I 'o , >- I I 1.2 o I I 3-84 I l- ./ I l!) W I •.. •.. 0.8 '\ ~./ •.. .•.. I \ ./ .•.. ./ j \ ,~ \. \ /I"""'__ 1/ iJ _,....-- \ \\ // , i;1/ \ \\ 1/ I \ \\ .:i , tt V) Z // I l- w I \ 1/ I u ,'I \ \ '- \ , /1 /1 /1 1/ 1986 ••.•.. I \ ./ \ 1-84 ,7 /~ 1985 0.4 84 85 86 87 88 89 .90 YEAR Fig. 10. Height-density (dm) of sandsage among 1984 through 1986 controls from 1984 through 1990, Tamarack Prairie, Colorado. Crown Cover Analysis Point frame sampling of vegetati~n crown cover was completed in August 1990 within both the 1985 and 1986 burns ~nd their respective controls. Tables 710) .. Trends among major species and combined groups of species from 1984 (pretreatment) through 1990 for the 1985 burns and controls varied (Table 11). Both 1984 and 1985 samples were pretreatment within the 1986 burns (Table 12). Bouteloua gracilis.--Fire enhancement of blue grama occurred within both the 1985 and 1986 burns until 1988. No change occurred after that other than blue grama remained greater on burns. This may have resulted from removal of dead vegetation, or an actual sustained increase may have occurred. Stipa comata.--Needle-and-thread grass was severely depressed by the 1985 burn for several years, but full recovery was evident by late summer 1990. Trends were less evident within the 1986 burns. Pretreatment indices in 1984 and 1985 did not agree. Fire apparently depressed needle-and-thread and the treatment and control transects showed similar trends from 1988 to 1990. �205 Table 7. Crown cover (point frame total), species composition (X), and frequency of occurrence of vegetation within treatment transects on 3 sites burned in 1985, Tamarack Prairie, August 1990. Crown cover 2 3 Category/species Totals Bare ground Dead vegetation 262 557 289 767 263 213 814 1,537 Bouteloua gracilis ~ comata Sporobolus cryptandrus Calamovilfa longifola Andropogon hallii Agropyron smithii ~ virgatum Oryzopsis sp. Cyperus & Carex spp. Artemisia filifolia Opuntia sp. 262 399 125 225 24 9 308 858 366 308 54 894 1,509 633 575 4 47 3 204 13 324 252 142 42 94 8 109 1 9 64 23 Ambrosia psilostachya Artemisia ludoviciana Tradescantia occidentalis Evolvulus nuttalianus Erigeron sp. Psoralea tenuiflora Physalis subglabrata Thelesperma megapotimicum ~ leptophylla Mentzelia nuda Liatris sp-.- 54 20 17 108 Salsola kali Helianthus'petiolaris Croton texensis Chenopodium album Conyza canadensis Pepidium densiflorum Cirsium sp. Eriogonum ~ Ameranthus sp. Lactuca sp. Polygonum sp. 82 51 rr 5 5 6 9 2 3 16 2 6 10 1 87 9 11 12 9 5 3 47 9 1 12 23 2 2 5 C~. Freq. Occur. 94 109 1 16 315 36 17.91 30.22 12.68 11.52 3.44 1.88 2.18 0.02 0.32 6.31 0.72 0.94 1.00 1.00 0.94 0.65 0.76 0.12 0.06 0.29 0.76 0.35 182 17 5 11 9 2 4 17 2 16 1 3.65 0.34 0.10 0.22 0.18 0.04 0.08 0.34 0.04 0.32 0.02 0.59 0.18 0.06 0.29 0.06 0.06 0.18 0.29 0.06 0.18 0.06 181 74 11 16 59 23 1 1 1 1 5 3.63 1.48 0.22 0.32 1.18 .0.46 0.02 0.02 0.02 0.02 0.10 0.65 0.29 0.29 0.53 0.53 0.29 0.06 0.06 0.06 0.06 0.06 1n �Crown cover (point frame total), species composition (X), and frequency of occurrence of Table 8. vegetation within 3 1985 control sites, Tamarack Prairie, August 1990. Crown cover Category/species 2 3 Totals C~. Freq. Occur. Bare ground Dead vegetation 238 589 358 653 415 246 1,011 1,488 Bouteloua gracilis Stiea £2!!!!!.!! Sporobolus cryptandrus Calamovilfa longifola Andropogon hallii Paspalum stramineum Agropyron smithii Muhlenbergia sp. Aristida sp. Koleria cristata Panicum virgatum Amual spp. Cyperus & ~ spp. Artemisia filifolia Opuntia sp. 257 490 166 142 20 5 9 267 883 180 157 23 146 95 161 54 5 3 7 1 2 670 1,468 507 362 150 5 68 3 7 1 2 1 14 495 62 13.83 30.30 10.46 7.47 3.10 0.10 1.40 0.06 0.14 0.02 0.04 0.02 0.29 10.22 1.28 1.00 1.00 1.00 0.88 0.41 0.06 0.71 0.12 0.12 0.06 0.06 0.06 0.41 0.82 0.41 99 3 6 24 5 1 21 15 2 2 2 2.04 0.06 0.12 0.50 0.10 0.02 0.43 0.31 0.04 0.04 0.04 0.59 0.18 0.12 0.35 0.06 0.06 0.24 0.29 0.12 0.06 0.06 747 33 25 2 8 3 9 5 1 6 9 1 15.42 0.68 0.52 0.04 0.17 0.06 0.19 0.10 0.02 0.12 0.19 0.02 0.65 0.59 0.24 0.06 0.12 0.12 0.12 0.12 0.06 0.12 0.12 0.06 63 107 1 140 7 1 6 64 10 7 291 45 Ambrosia psilostachya Artemisia ludoviciana Tradescantia occidentalis Evolvulus nuttalianus Erigeron sp. Psoralea tenuiflora Physalis subglabrata Thelesperma megapotimicum Sphaeralcea coccinea lygodesmia ;uncea Mentzelia nuda 39 14 46 3 Salsola kali Chenopodium album Pepidium densiflorum Helianthus petiolaris Croton texensis Eriogonum !Q!l!:!!! Conyza canadensis Euphorbia sp. Cirsium sp. Polygonum sp. Ameranthus sp. ~sp. 10 13 5 6 11 2 13 5 1 19 6 2 1 n7 18 2 2 9 1 2 10 2 18 1 3 6 5 7 3 1 6 9 �207 Table 9. Crown cover (point frame total), species composition (X), and frequency of occurrence of vegetation within treatment transects on 3 sites burned in 1986, Tamarack Prairie, August 1990. Crown cover Category/species 2 3 Totals Compo Freq. Occur •. 0.95 1.00 1.00 0.74 0.53 0.74 0.26 0.05 0.05 0.05 0.37 Bare ground Dead vegetation 535 614 176 410 427 705 1,138 1,729 Bouteloua gracilis Stipa ~ Sporobolus cryptandrus Calamovilfa longifola Andropogon hallii Agropyron smithii Aristida sp. Panicl.J1l virgatl.J1l Muhlenbergia sp. Oryzopsis sp. Cyperus & ~ spp. 5n 384 219 147 148 9 8 128 344 151 278 33 24 633 517 131 301 15 37 2 1,338 1,245 501 726 196 70 10 23 23 8 1 6 3 6 8 1 15 25.05 23.31 9.38 13.59 3.69 1.31 0.19 0.43 0.15 0.02 0.28 Artemisia filifolia ~sp. 233 5 10 1 62 37 305 43 5.71 0.81 0.79 0.53 161 47 1 15 2 22 3 8 1 13 33 3.01 0.88 0.02 0.28 0.04 0·.41 0.06 0.15 0.02 0.24 0.62 0.47 0.16 0.05 0.26 0.05 0.32 0.16 0.05 0.05 0.05 0.32 343 45 72 34 31 13 5 4· 1 2 3 1 6.42 0.84 1.35 0.64 0.58 0.24 0.09 0.07 0.02 0.04 0.06 0.02 0.84 0.16 0.75 0.32 0.26 0.42 0.11 0.11 0.05 0.05 0.05 0.05 Ambrosia psilostachya Artemisia ludoviciana Tradescantia occidentalis Evolvulus nuttalianus Lathyrus polymorphus Erigeron sp. Thelesperma megapotimicum Haplopappus spinulosus Ipomoea leptophylla Mentzelia nuda SphaeralceS-COccinea 42 119 47 10 2 5 1 5 Salsola kali Helianthus petiolaris Chenopodium album Conyza canadensis Croton texensis Pepidil.J1l densiflorum Euphorbia sp. Eriogonum.!!!l!:!!! ~sp. Cirsium sp. Kochia scoparia Amaranthus sp •. 22 12 13 2 8 8 3 1 6 1 17 93 11 14 18 1 11 1 8 1 13 16 228 33 48 18 5 4 2 3 2 3 �208 Table 10. Crown cover (point frame tota'L), species composition (X), and frequency of occurrence of vegetation within 3 1986 control sites, Tamarack Prairie, August 1990. Crown cover 2 3 Category/species Totals C~. Freq. Occur. Bare ground Dead vegetation 340 814 125 400 355 882 820 2,096 Bouteloua gracilis 604 484 101 350 19 2 127 500 93 182 27 5 6 2 359 690 192 365 47 21 5 21 7 1 15 39 1,090 1,674 386 897 93 28 11 21 7 1 15 41 20.60 31.63 7.29 16.95 1.76 0.53 0.21 0.40 0.13 0.02 0.19 0.77 1.00 1.00 1.00 0.95 0.53 0.37 0.16 0.11 0.05 0.05 0.11 0.16 6 3 300 37 462 63 8.73 1.19 0.68 0.68 2 5 68 6 1 6 6 2 3 1 98 8 4 11 1 3 5 7 3 3 1.85 0.15 0.08 0.21 0.02 0.06 0.09 0.13 0.06 0.06 0.26 0.21 0.11 0.16 0.05 0.11 0.21 0.11 0.05 0.05 269 6 5 32 15 13 2 3 10 2 3 5.08 0.11 0.09 0.60 0.28 0.25 0.04 0.06 0.19 0.04 0.06 0.42 0.11 0.16 0.32 0.16 0.26 0.11 0.05 0.11 0.05 0.05 !!le~ Sporobolus cryptandrus Calamovilfa longifola Andropogon h!l!ii Agropyron smithii Paspalum stramineum Muhlenbergia sp,: Orvzoes issp. Koleria cristata Annual grass spp. Cyperus & ~ spp. Artemisia filifolia Opuntia sp. 156 23 Ambrosia Dsilostachya Artemisia ludoviciana Tradescantia occidentalis Evolvulus nuttalianus lathyrus polymorphus Physalis subglabrata Thelesperma megapotimicum Sphaeralcea coccinea lygodesmia juncea Mentzelia nuda 30 Salsola kal i Helianthus petiolaris Croton texensis Chenopodium album Eriogonum .!!!:!m!!! Conyza canadensis Euphorbia sp. Pepidium densiftorum Polygonum sp. Gaura A'iiiaranthus sp. 61 4 3 5 15 203 2 5 16 2 11 1 5 7 1 3 1 2 3 3 1 3 10 2 3 �209 Table 11. Crown cover (point frame mean/transect) for selected species and species groups among pre- and post-treatment samples within combined 1985 burns and their controls, Tamarack Prairie, Colorado, 1984-90. Species/category Bouteloua gracilis Stipa comata Sporobolus cryptandrus Calamovilfa .longifolia Andropogon hallii Agropyron smithii Selected warm season Total warm season Total cool season Artemisia filifolia Comb. peren. forbs Combined annual forbs Bare ground Dead vegetation 1985 28.5 31.4 4.8 36.8 1.7 0.6 39.4 72.7 32.1 22.5 4.4 4.1 30.9 260.1 21.2 27.7 21.1 54.5 5.1 1.8 61.8 104.0 29.5 8.2 3.4 2.8 110.3 170.1 Treatment 1986 20.6 48.2 21.1 58.8 9.6 10.4 70.5 112.2 58.6 11.9 5.7 3.5 90.7 145.1 1988 1990 26.6 36.2 19.8 47.0 11.2 11.6 63.1 109.5 47.9 18.0 17.5 15.5 31.1 189.6 52.6 88.8 37.2 33.8 10.1 5.5 50.4 140.2 94.3 18.5 15.6 22.1 47.9 90.4 20.2 43.1 23.5 35.4 9.1 8.9 44.6 88.2 52.0 34.6 12.7 20.9 42.7 176.0 39.4 86.4 29.8 2l.3 8.8 4.0 28.5 99.5 90.4 29.1 10.6 49.9 59.5 87.5 Control Bouteloua gracilis Stipa comata Sporobolus cryptandrus Ca1amovilfa longifolia Andropogon hallii Agropyron smithii Selected warm season Total warm season Total cool season Artemisia filifolia Comb. peren. forbs Combined annual forbs Bare ground Dead vegetation apretreatment. 30.8 31.3 8.1 33.1 2.1 2.2 35.9 74.8 33.5 24:7 2.9 3.6 42.2 249.3 21.9 55.2 22.7 28.5 4.1 3.5 32.8 77 .4 58.7 22.6 2.1 3.5 47.1 217.9 10.5 .67.2 22.1 26.4 .4.6 11.0 31.1 63.8 78.2 23.8 4.2 4.2 67.7 187.9 �210 Table 12. Crown cover (point frame mean/transect) for selected species and species groups among pre- and post-treatment samples within combined 1986 burns and their controls, Tamarack Prairie, Colorado, 1984-90. Species/category Boute1oua gracilis Stipa comata Sporobo1us cryptandrus Ca1amovi1fa longifo1ia Andropogon ha11ii Agropyron smithii Selected warm season Total warm season Total cool season Artemisia fi1ifo1ia Comb. peren. forbs Combined annual forbs Bare ground Dead vegetation 1985 37.3 42.5 10.3 26.9 1.9 28.9 76.5 46.1 18.1 7.2 5.2 39.5 234.2 36.1 40.0 13.7 25.3 2.4 4.3 27.6 77 .4 44.2 16.2 4.7 1.6 41.1 241.4 Treatment 1986 38.1 38.5 24.5 44.9 6.8 9.5 51.7 114.3 47.9 7.9 15.4 3.2 182.4 57.2 1988 1990 42.5 38.4 19.0 45.9 8.4 10.1 54.4 115.9 48.5 14.8 25.9 12.9 46.9 163.6 70.4 65.5 26.4 38.2 10.6 3.7 49.7 146.5 69.2 16.1 16.1 29.2 59.9 91.0 31.0 48.7 19.4 54.4 2.8 4.2 57.6 108.0 52.9 29.7 11.7 12.2 34.7 177 .0 57.4 &8.1 20.3 47.2 4.9 1.5 52.1 129.8 89.2 24.3 7.5 18.9 43.2 110.3 Control Bouteloua gracilis Stipa comata Sporobo1us cryptandrus Ca1amovi1fa longifo1ia Andropogon ha11ii Agropyron smithii Selected warm season Total warm season Total cool season Artemisia fi1ifo1ia Comb. peren. forbs Combined annual forbs Bare ground Dead vegetation apretreatment. 40.6 28.1 4.6 39.5 0.6 40.1 85.3 29.1 14.4 5.1 3.1 28.9 263.9 34.6 52.1 14.8 33.8 1.1 1.8 34.9 84.3 53.7 19.3 3.0 1.3 32.6 226.6 18.0 61.5 15.3 37.2 1.2 5.3 38.3 71.6 66.8 22.6 6.3 2.3 46.6 204.8 �211 Sporobo1us cryptandrus.--Sand fire within the 1985 burns. the 1986 burns. dropseed showed no evidence of being impacted by Short-term enhancement may have occurred within Ca1amovi1fa 10ngifo1ia.--The 1985 burns markedly stimulated growth of pra~r~e sandreed for 2 years. A similar trend was evident within the 1986 burns, however, data for the 2nd post-treatment year were lacking and enhancement did not continue to the third year (1988). There was no evidence that impacts of fire were more than temporary. Hot, dry weather in June was suspected as the reason reduced growth of prairie sandreed occurred in 1990 (Tables 11 and 12). Andropogon hallii.--Although sample sizes were small, sand b1uestem increased following fire within both the 1985 and 1986 burns. It also increased within the controls indicating lack of grazing may be enhancing the species. Warm-season Combinations.--Sand bluestem crown cover was combined with pra~r~e sandreed and switchgrass as selected warm-season grasses (Tables 11 and 12). A strong fire enhancement was noted, however, sandreed dominated this combined group so trends followed that species (Fig. 11). When total warm-season grasses were combined (blue grama and sand drop seed were added), the data show fire stimulated a pronounced short-term increase with retention of a lower rate of enhancement through 1990 (Tables 11 and 12). 80 8S BURN IU W (/) z « 0: 60 86 I- CONTROL "(f) / ..............•. l86 BURN I « IT ...• -/ 40 w > « ...•.....••. , - -.... / ..... ...•••...•.. _,...,.-:/•.... -'-' --- ........._._ ..•. ./ .~ ;;.t-........ "- .•... / . / .......•../ W 19 ---..." / / / \. / ./. / \. " \. " / " ~.~..7··- - -._ - - " .•••...... 85 CONTROL " "" " 20 84 85 86 88 90 YEAR Fig. 11. Crown cover (point frame) of selected warm-season grasses within the 1985 and 1986 burn and control ·sites from 1984 through 1990, Tamarack Prairie, Colorado. Pretreatment samples were obtained in 1984 (1985 burns) and 1984-1985 (1986 burns). �Sand Sagebrush.--Sand sagebrush was markedly reduced by 1985 and 1986 burns. Some recovery was noted in 1988 and 1990, but recovery to pretreatment levels was not attained (Tables 11 and 12). Perennial Forbs.--Combined perennial forbs showed no immediate effect of fire within the 1985 burns, but evidence they were stimulated by fire was indicated within the 1986 burns (Tables 11 and 12, Fig. 12). Increases in perennial forbs during both 1988 and 1990 were noted within the 1985 burns. Western ragweed (Ambrosia psi1ostachya), a warm-season perennial, showed evidence of fire enhancement within the 1985 and 1986 burns in 1988 (Snyder 1989). This increase was only partially sustained in 1990. Evidence of enhancement was not noted within the 1984 burns. 30 ...........•.... 25 I- / o W « ... . ..... (f) z r .! 20 85 BURN ......•. 0:: ..:"' I- . .•.. ". . . . <, (f)' l- 15 ,;<----_" I B6 BU'N W o « 0:: 10 W > « 5 /" "-~--""""// ~<.i>. ••••• •. l· -------.-" 85 CONTroL o 84 85 86 88 90 YEAR Fig. 12. Crown cover (point frame) of perennial forbs within the 1985 and 1986 burn and control sites from 1984 through 1990, Tamarack Prairie, Colorado. Pretreatment samples were obtained in 1984 (1985 burns) and 1984-1985 (1986 burns). Annual Forbs.--Annual forbs showed no post-tr~atment effects of fire based on crown cover sampling within the 1985 and 1986 burns (Tables 11 and 12). I~ 1990 annual forbs within the 1985 and 1986 controls increased much more than within the burns. Russian thistle (Saisola kali) was primarily responsible, increasing dramatically within some transects, especially within thin stands of need1e-and-thread. �213 Bare Ground and Dead Vegetation.--The amount of bare ground was near pretreatment levels in 1988 within the 1985 and 1986 burns and remained there through 1990 (Tables 11 and 12). Dead vegetation and litter declined for 2 years following the 1985 burns and remained in unison with that in the controls during both 1988 and 1990 sampling periods. Sampling was not conducted in 1987 within the 1986 burns and it is not known if a 2nd year of reduction occurred, but near full recovery was noted by 1988. Trends toward less dead vegetation were apparent within both the burned and control transects. Possible responsible factors include sampling when vegetation was greener in later years, slight changes in sampling procedure, or a combination. Revegetation and Renovation Treatments Tillage-Reseeding.--The height-density of residual vegetation, primarily switchgrass, was sampled beginning in early spring 1987 after the 2nd (1986) growing season. The average HDI was 1.8~ dm among 12 transects. Visual obstruction readings increased to 4.44 dm in 1988 and 4.38 dm in 1989, but declined to 3.55 dm in 1990. Although a downward trend was noted, HDI's continued to average above other vegetation within the Tamarack Prairie. Point-frame sampling of vegetation types within the 1985 revegetation strips in late summer 1990 continued to show dominance by tall, warm-season grasses (Table 13). Standing residual (primarily switchgrass) was the most frequently recorded item during the last 3 years. Increases in annual forbs were noted in 1990 but ot~er native grasses remained suppressed by taller grasses. The amount of bare ground continued to decline. Table 13. Crown cover (point frame total) among vegetation groups from 1985 through 1990 within 12 1985 revegetation transects, Tamarack Prairie, Colorado. Vegetation/category Bare ground Dead vegetation/litter Dominant native grasses Lesser grasses and sedges Seeded tall grasses Sandsage and cactii Perennial forbs Annual forbs 1985 1986 1987 1988 1990 804 212 217 79 311 6 22 76 812 156 226 37 475 10 12 0 278 372 183 21 833 8 2 31 120 620 98 13 829 15 20 13 107 394 138 14 850 14 14 197 Renovation of Interseeded Tracts.--The HDI of residual grass-forb vegetation declined from the 1988 peak in both the tillage-herbicide treatment and control transects during early spring 1990 (Table 14). However, the HDI within the treated area remained far greater than that within the control (f < 0.01), indicating treatment effect continued ~hrough the 1989 growing season. �L..14 Table 14. Mean height-density (dm) of grass-forb, sandsage, and combined cover from pre- (1986) to post-treatment (1987-90) intervals between a tillage-herbicide and untreated control within a previously interseeded location, Tamarack Prairie, Colorado. Year Tillage-herbicide Control Grass-forb 1986 1987 1988 1989 1990 0.382 0.314 1.554 1.388 1.043 0.393 0.271 0.551 0.449 0.375 Sandsage 1986 1987 1988 1989 1990 1.234 0.227 1.136 1.143 0.750 0.696 0.608 1.268 0.781 1.154 Combined 1986 1987 1988 1989 1990 0.579 0.306 1.525 1.373 1.038 0.544 0.375 0.677 0.520 0.544 Crown cover sampling in late summer 1990 also showed that bluestems and switchgrass continued to be enhanced by the early spring 1986 tillageherbicide treatment (f < 0.005), whereas blue grama, needle-and-thread, and other native grasses remained suppressed (Table 15). Prairie sandreed was apparently neither enhanced nor hindered by treatment. This tall, warm-season native grass survived the tillage-herbicide treatment because it was deeprooted like the bluestems and switchgrass. Perennial forbs, likewise, have shown little long-term effect of treatment (Table 15). In 1990, tall, warmseason species showed little change from the preceding year. Tillage-herbicide renovation was applied to numerous other interseeded tracts by management personnel in 1986-87. Pretreatment HOI data were lacking, but within a sample of 10 sites, the average HOI of standing residual was 2.71 dm in 1989. Their HOI increased to 3.41 dm in 1990 and was comparable with that (3.55 dm) within the strips revegetated in 1985. �215 Table 15. Average crown cover/transect (0.01-m2) of selected species and species groups from pre- (1985) to post-treatment (1986-90) intervals within tillage-herbicide renovation of an interseeded site and control, Tamarack Prairie, Colorado. 1985 1986 1987 1988 1989 1990 Andropogon/Panicum Calamovilfa longifolia 13.7 11.3 24.2 12.4 36.3 13.8 39.8 11.2 47.1 13.8 48.3 12.4 Bouteloua, !!ie!, etc. Perennial forb 17.6 1.5 5.4 0.6 6.4 1.2 6.8 1.6 9.9 2.6 12.5 4.9 16.4 10.7 17.1 2.2 16.6 10.6 16.8 1.7 22.1 10.2 19.8 3.4 24.6 8.1 20.4 1.7 33.9 8.5 27.6 1.2 32.2 9.0 35.0 2.3 Species/group Tillage-herbicide Control Andropogon/Panicum Calamovilfa longifolia Bouteloua, Stipa, etc. Perennial forbs F Values Andropo90n/Panicum Calamovilfa longifolia Bouteloua, !!ie!, etc. 8e < 1985-90 1989-90 42.198 1.44 34.828 4.08 1.42 0.04 0.005. Strip Spraying of Sandsage.--Much of the residual sagebrush that had remained standing after being killed in 1985 deteriorated in 1989, contributing little to the HOI in early spring 1990. Sandsage was the primary obstruction, 37.2% of the time in 1987, 24.1% in 1988, 22.7% in 1989, and 8.2% in 1990. The HOI of grass-forb vegetation declined to 0.49 dm in 1990 in a pattern similar to that within the 1985 and 1986 burns (Figs. 7 and 8). Grass-forb vegetation within the site sprayed in 1985 and burned in 1986 (Table 16) also showed markedly lower HOI's in spring 1990. Point frame sampling of crown cover, conducted in late May 1990 (Table 17) was compared with pretreatment and other preceding years. Grasses gradually increased with needle-and-thread being among the main benefactors. Increases in 1990 probably resulted from favorable spring moisture in contrast to 1989 (Table 17). Sandsage was beginning to show minor recovery and prickly pear cactus, already abundant, steadily increased. Combined perennial forbs also increased, except in 1989, when dry weather persisted in spring. Increases in total species contacts were observed. Monitoring of Prairie Grouse Radio-marked Birds from 1989.--0ata obtained in 1990 supplemented that obtained from 6 males and 7 female greater prairie-chickens radio-marked with solar-powered transmitters in spring 1989 in the vicinity of the Tamarack Prairie (Fig. 13). Among that group, 2 nested on the Tamarack Prairie, 1 nested a few meters to the south, and the others nested at more distant locations. However, none of the hens was confirmed to have successfully nested. Among the 7 hens, 3 (probably 4) were killed by predators while �216 Table 16. Mean height-density (dm) within the 1985 sandsage spray site during 1986 to 1990 post-treatment intervals and preburn (1986) to post-burn (1987-89) intervals within a portion burned in 1986, Tamarack Prairie, Colorado. Year N Transects Grass-forb Sands age Combined 1985 SANDSAGE SPRAY SITE 1986 1987 1988 1989 1990 2 11 11 11 11 0.246 0.272 0.613 0.675 0.494 0.879 0.800 1.127 0.960 1.021 0.468 0.469 0.737 0.740 0.538 0.713 0.333 1.250 0.404 0.129 1.133 0.911 0.63~ SANDSAGE SPRAY - 1986 BURN 1986 1987 1988 1989 1990 4 4 4 4 4 0.279 0.122 1.131 0.911 0.639 nesting and 3 survived until spring 1990. Among the latter, 2 survived through summer 1990. One nested successfully in 1990 but the other was unsuccessful. The 3rd hen was monitored as she visited Lek 1 in April, but her weak signal was not detected again and her survival and nesting success remain unknown. Among 6 males radio-marked in spring 1989, 4 were killed by predators in spring, 1 apparently slipped its harness, and 1 survived to mid summer after which it could not be located (probably due to a weak signal). Monitoring of Transplanted Prairie-Chickens.--Management personnel trapped 23 greater prairie-chickens in Yuma County in April 1990 and released them on the east half of section 15 (Township 10 N, Range 48 W) in the northeast part of the Tamarack Prairie (Fig. 13). Releases were made at a proposed 1ek site (R 1ek) and tape-recorded calls were used in an unsuccessful attempt to retain the males at the 1ek. Eight (4 of each sex) were equipped with 10-g batterypowered transmitters which had a 1-year life expectancy. These transmitters, attached with an elastic loop around the lower neck, worked well and seemed to cause little stress to the grouse. The transmitters, previously tested as operating, were not tested at time of release. The signal from 1 hen was not picked up in subsequent monitoring. Either the transmitter quit working while the hen was captive or she left the area immediately after release. The other 3 hens moved about 1.6 km south of the release site (near M lek, Fig. 13) where 2 (possibly 3) nested on the Tamarack Prairie. One was killed by a raptor and an abandoned nest (possibly hers) was later found in the vicinity. The other 2 hens successfully nested and raised their broods in the southeast part of the Tamarack Prairie. They shifted to the south in fall and their signals were eventually lost prior to early winter. �.217 Table 17. Crown cover (point frame total) of vegetation and ground cover within 11 random transects during pre-(1985) and post-treatment (1986-90) spring intervals in the June 1985 herbicide-treated sandsage spray tract, Tamarack Prairie, Colorado. Vegetation/cover Bare ground Dead vegetation Perennial grass Bouteloua gracilis Stipa ~ Sporobolus cryptandrus Calamovilfa longifolia Other perennial grasses Annual grass BrOlllJstectorun Festuca sp. Artemisia filifolia (alive) A. filifolia (dead) ~ & Echinocereus spp. Perennial forbs Ambrosia & Artemisia spp. Tradescantia occidentalis Lathyrus polymorphus Psoralea tenuiflora .Evolvulus nuttalianus Phlox andicola All iun textile MeiitZelia nuda Leucocrinum-montanun Penstemon angustifolius Thelesperma megapotimicun Cymopteris mont anus Abronia fragrans Erigeron sp. Liatris sp. Total perennial forbs Annual forbs Croton texensis Chenopodiun albun Pepidiun & Lesguerella sp. Miscellaneous annual forbs Pretreatment 1985 1986 1987 402 1,425 529 1,324 536 1,083 142 194 34 51 4 148 510 46 65 15 70 Post-treatment 1988 1989 1990 387 990 472 1,207 454 1,028 312 425 106 53 19 307 683 161 120 15 311 563 108 124 15 327 722 103 87 23 117 4 116 7 38 10 436 112 1 259 0 287 0 210 4 201 5 124 44 50 62 83 75 97 5 18 114 55 6 4 2 12 1 2 2 2 1 9 77 7 17 121 1 1 8 1 6 12 45 2 29 22 123 1 1 6 1 107 8 1 1 19 1 1 6 1 6 3 11 1 223 3 24 3 1 96 3 154 156 76 4 1 1 6 1 4 4 6 178 2 1 9 8 �"'d"1 •... N Ii 1-" IIIOQ 1-" , Ii 1-" 00 t-' R 49 W - ~ po- H ""1:1 W (I) W R 48 'II - R 47 'II ~p t-' \Or-' 13 \0(1) o~ ~ !;;. ~ III 8"d Ii III 1-" Ii 1-" CD I g. CrAMARACK PRAIRIE ~. .> 19 ~ M 7< X X '"""W ~---- o X * 25 12 :>;"' CD ::l 27 29 ::l III rt >- ~ () i 25 tTl If) N t-' o 19 23 * 9 1-" (I) 13 ~7< * 31 35 en 1, 33 0 ~ III 31 35 ,. 2 X rt i 1-" o ::l III >- 5 1 1-" 7< ::l 16 I 3 1 ~ X t5 'z Ii (I) t-' III rt 7 11 9 X *14 1-" o §N 11 ROAD 46 ::l rt o rt ::T (I) t-:l ~ III Ii III o :>;"' LI '13 7< X LEK GEND t 3,2 KM # NEST Il 0 1 SCALE 2 mi . j~ l 13 , I lJ 12 �219 Two of 4 radio-marked males died soon after release, probably to predation. One was recovered a short distance east of the release site and the other was found to the south near the Tamarack Prairie boundary. One surviving male became an active resident at lek 1 and the other was active at lek 2 through the remainder of the spring. They remained near «3 km) their respective leks through the remainder of summer and fall. Radio-marked Birds From 1990.--Trapping, banding, and instrumenting of resident prairie grouse using leks closely asso~iated with the Tamarack Prairie continued in spring 1990. Lek monitoring and searches for new leks also continued and 2 new leks were found and monitored. One new lek (lek 16, Fig. 13) contained 1 greater prairie-chicken, 2 or more plains sharp-tailed grouse and several male hybrid crosses of the 2 species. Several males were observed at lek 9 near the Tamarack Prairie border in early spring 1990. However, use suddenly diminished prior to initiation of trapping and only 1 male remained through spring. Two males began using a mowed lek (M lek) on the Tamarack Prairie near lek 9 (Fig. 13). Only 1 male prairie-chicken and 1 sharp-tailed grouse remained at lek 12 through spring 1990. Trapping was not conducted on either lek in 1990. Thus, prairie-chicken activity associated· with the Tamarack Prairie seemed diminished for unknown reasons in 1990. Thirteen resident greater prairie-chickens (7 hens and 6 males) were trapped and banded at leks 1 and 14 (Fig. 13). All hens and 1 male were instrumented with 7-g battery-powered transmitters (attached by a small elastic cord around the lower neck) with a life expectancy of 6 months. These transmitters worked well, seemed to have little impact on the grouse, and most transmitters operated for 7 to 8 months. One hen, whose solar transmitter no longer operated properly, was a retrap from the previous year. She was radio-marked with a new transmitter. Two hybrid males were trapped, banded and released. and 1 female sharp-tailed grouse were radio-marked. Two of 3 trapped males Total radio-marked grouse at the end of the April 1990 trapping effort included 10 female and 5 male prairie-chickens and 1 female and 2 male sharptailed grouse (Table 18). These data do not include 2 hens, 1 transplanted and 1 radio-marked in 1989, whose signals were not subsequently located. One hen (hen 921), instrumented in 1989, but not trapped in 1990, was subsequently located on her nest in late May 1990 adding to the monitored population through late spring and summer 1990. Most hens (7 of 8), that survived through summer 1990, successfully nested (Table 18). One-third of the 12 hens were killed by predators. Three of 7 males died including 2 transplants that died soon after release to predation or other causes. The height-density of grass-forb and sandsage vegetation was sampled at, and proximal to, nest sites and compared with samples taken at 2 random locations within a 65-ha area surrounding each nest. The average HOI of grass-forb vegetation at 10 nest sites was 0.622 dm, whereas that at 20 random sites averaged 0.567 (f> 0.10). The HOI of sandsage averaged 2.53 dm at nest sites compared to 2.14 dm at random locations (f> 0.10). There was no difference for either grass-forb or sandsage vegetation and similar findings were obtained when 1989 data were included. Among the 10 nests found in 1990, all but except 1 were partially concealed by sandsage. �220 Table 18. Transmitter I"IUItIer Radio-marked prairie grouse data, northeastern Logan County, Colorado, 1990. Species Age Survival days Resident! transplant 25a 200+ 200+ 15a R R R R R R R T T T R R Successful nest Nunber of nests Totals Hatched 2 6 6 2 12 12 8 0 11 8 ? 8 12 8 13 Eggs FEMALES 906 918 949#1 968 977 987 998 057 121 141 959b 921 PC PC PC PC PC PC PC PC PC PC STG PC 928 937 949#2 019 037 099 178 PC STG STG PC PC PC PC J A A A J A A J rr 200+ 200+ 25a A 152+ 200+ 163+ 100+ J J J 200+ 185+ 30a A A A No Yes No No No Yes Yes No Yes Yes Yes Yes 1 2 12 14 13 ~ A A J A r 185+ 14a 200+ R R R T T T T aKnown mortality. bHen 921 was radio-marked in 1989 and found incubating in May 1990. Radio-marked hens, that had successfully nested, were flushed in late August in an attempt to estimate brood size. However, by that date some broods had merged with others and some hens apparently no longer were with broods. Therefore, meaningful brood size data were not obtained. Although surviving radio-marked transplanted hens remained on the Tamarack Prairie and raised their broods there, little use of the property was noted by prairie grouse resident to the south. One hen, from 1ek 1, moved onto the Tamarack Prairie but was killed prior to nest initiation. Two other hens temporarily shifted to the .Tamarack Prairie after their nests were predated, but moved back to previously established home ranges within a few days. All hens with broods raised them within several kilometers of their respective nest sites. Several hens used grain field-rangeland interspersions near lek 14, but primarily resided in rangeland. Nearly all birds shifted south in fall to the vicinity of 1ek 14 where grain fields existed. The transplanted hens showed evidence of similar moves although their wintering residence was not found. �221 LITERATURE CITED . Snyder, W. D. 1986~.· Sandsage-b1uestem prairie renovation. Job Progress Rep., Colorado Div. Wi1d1., Wi1d1. Res. Rep., Fed. Aid Proj. 01-03-045 (W37-R). Apr.:475-498. 1986Q. Sandsage-b1uestem prairie renovation. Job Progress Rep., Colorado Div. Wi1d1., Wi1d1. Res. Rep., Fed. Aid Proj. 01-03-045 (W-37-R). Apr. :499-525. 1987. Sandsage-b1uestem prairie renovation. Job Progress Rep., Colorado Div. Wi1d1., Wildl. Res. Rep., Fed. Aid Proj. W-152-R. Apr. :331356. __________ 1988. Sandsage-bluestem prairie renovation. Job Progress Rep., Colorado Div. Wildl., Wi1d1. Res. Rep., Fed. Aid Proj. W-152-R. Apr. 3:395 -421. 1990. Sandsage-bluestem prairie renovation. Job Progress Rep., Colorado Div. Wildl., Wildl. Res. Rep., Fed. Aid Proj. W-152-R. Apr. :173202. Prepared by '%£&.)z!J.t!J) IJ¥ Warren D. Snyder Wildlife Researcher C �.~ './ �JOB FINAL REPORT . State of: Colorado Project: W-152-R Work Plan: Job Title: 21 : Job _4_ Feeding ecology of Sharp-shinned Hawks and nest site characteristics of accipiters in Colorado Period Covered: Author: Upland Bird Research 01 January 1987 through 31 December 1990 Suzanne M. Joy Personnel: C. E. Braun, R. W. Hoffman, Colo~ado Division of Wildlife; R. T. Reynolds, U.S.D.A. Forest Service; R. L. Knight, S. M. Joy, Colorado State University ABSTRACT Feeding ecology of 11 sharp-shinned hawk (Accipiter striatus) pairs and nest site characteristics of 14 sharp-shinned hawk, 4 Cooper's hawk (A. cooperii), and·2 northern goshawk (A. genti1is) nests in mature aspen, conifer, and mixed aspen-conifer habitats in southwest Colorado were investigated during 1988 and 1989. Small birds (i - 21g) and mammals (i- 4lg) comprised 91 and 9% of 513 prEy items identified, respectively. Unique plucking and feeding habits of individual pairs accounted for differences in total prey and avian prey numbers among nests. More (>50%) prey items were found during the nestling than incubation or fledgling stages. Although birds were consumed more often than mammals during all breeding stages, mammalian prey frequency progressively increased from incubation through fledgling. Numbers of avian prey at the nest sites increased 164% from incubation to the nestling stage. Nearly 60% of the birds eaten during nestling and fledgling stages were young of the year. This was reflected in a shift in median prey (bird) weight from 17.4 g during incubation to 12.1 g during the nestling stage. Hawks .at nests dominated by aspen habitat (n - 5) consumed fewer mammals (6.8% of diet) than hawks at 1 nest surrounded by coniferous habitat (28.3% of diet). Conifer habitats appeared to have less foraging value compared to mature aspen types. Avian prey were consumed in proportion to their size availability and species composition in habitats surrounding nests, whereas certain taxa and size classes of mammals were preferentially preyed upon. The data suggest that sharp-shinned hawks are vulnerable to predation and require the protective cover provided by conifers at the nest. Foraging success in aspen habitats may also influence nest site location for this species. Whereas northern goshawks appear to require relatively open, old-growth coniferous forests for foraging, as well as large aspen trees for nesting, Cooper's hawks primarily used aspen dominated habitats for foraging and nesting. Because accipiters require specific stand characteristics for nesting, active and potential nest sites should not be modified by timber harvests. �RECOMMENDATIONS 1. Insular patches (1-14 ha) of mature conifers (preferably species with a dense crown such as Engelmann spruce or subalpine-fir) on northerly exposures ·should be maintained within continuous stands of aspen to provide nesting habitat for sharp-shinned hawks. - 2. The insular patches of conifers should occur within 400 m of water and should be surrounded by 1-2 km2 of mature and or mixed conifer-aspen types with no more than 20% of the area in openings of < 10 ha. 3. The maximum distance between conifer patches within aspen types should be < 2 km. No roads should occur within 400 m of these insular conifer stands and no timber harvest should occur within 1 km. 4.. Active accipiter nest sites should be marked and protected. 5. Because sharp-shinned hawks forage opportunistically in a variety of habitats, limited timber harvest may occur in these habitats provided the density of small birds is not significantly altered. Small « 10 ha), irregular shaped cuts are preferred over large, block shaped cuts. Openings created by timber harvest should be no more than 60 m wide. �225 ACKNOWLEDGMENTS The Colorado Division of Wildlife (CDOW) and the U.S. Department of Agriculture Forest Service (USFS) provided partial funding for this project through a cooperative agreement. I received additional support from the Hawk Mountain Sanctuary Association (1989 Hawk Mountain Research Award) and Colorado State University (Colorado Graduate Fellowship). The financial support of all was appreciated greatly. I thank R. L. Knight, my major adviser, for advice, critiques, many open discussions, and his willingness to take on a microbiologist for a .wildlife study. For encouragement, constructive comments, and sharing his knowledge of and enthusiasm for accipiters, I thank.R. T. Reynolds. I thank G. C. White for statistical advice (without which Chapter 1 would have been a statistical nightmare) and lessons in hypothesis testing. B. Van Horne opened my mind to new ideas and provided critical comments on both my study design and thesis for which I am grateful. I thank R. W. Hoffman for many candid discussions, and for helpful comments on CDOW reports, study plan, and thesis. C. E. Braun also provided constructive comments on CDOW reports and the thesis. I am indebted to my field assistants for their hard work, unwavering enthusiasm, and friendship. S. Bedard, M. Kralovek, B. Rusk, and S. Severs greatly enriched my time in the field. For use of museum collections, I thank M. Bogan and C. Ramotnik of the U.S. Fish and Wildlife Service (USFWS) in Fort Collins, CO, and R. �c. Banks of the USFWS in Washington, D.C. I also thank C. Dove and R. Laybourne for invaluable help in improving my skills as a feather detective. I thank W. D. Shepperd of the USFS for permitting me to use his core counter and for many discussions on aspen ecology. I thank B. Cade of the USFWS for help in running MRPP. I am grateful to my fellow graduate students for the many coffee breaks, long talks, and parties although I didn't make it to all. I thank my family for their love and encouragement during the good and bad times, and for believing in me. Finally, I am especially grateful to S. C. Frye, my husband, for his moral support, love. and friendship . • �227 TABLE OF CONTENTS Page ABSTRACTOF THESIS ....•............................................ iii ACKNOWLEDGMENTS v LIST OF TABLES ix LIST OF FIGURES Chapter 1. x FEEDING ECOLOGYOF BREEDINGSHARP-SHINNEDHAWKSIN COLORADO Study 1 Area 2 Methods ........•..•..•........................................... Prey Collection Diet Composit;ion .......•....................................... 4 5 Foraging Zone Resource Use and Availability Diet 4 7 with Breadth Respect " to Habitat , 7 10 Results 10 Prey Collection 10 Diet Compos Led en 11 Fc'r'ag l ng Zone Resource Diet 18 Use and Availability with Respect Breadth to Habitat 18 20 Discussion 25 Prey Collection and Possible Data Bias Diet Composition .•........................................... 25 25 �TABLEOF CONTENTS(Continued) Foraging Zone .....•..................•....................... Resource Use and Availability Diet Chapter with 28 Respect to Habitat Breadth 29 30 Management Implications 31 Literature 31 2. Cited •............................................... NESTINGHABITATOF ACCIPITERS IN SOUTHWEST COLORADO Study Area 38 Methods .....•.............. Results ' ...•................................. 39 ...........•............................ Discussion 36 " 41 .........•............................................ 47 Management Implications 52 Literature 53 Cited .•.....•......•................................. APPENDICES...............•....... " Appendix 1 .................•.•............... Appendix 2 ...•.................................................. ~........•.... " 57 58 62 �229 LIST OF TABLES Table Page 1.1 Maximum likelihood log-linear model for numbers of total (avian and mammalian) prey and avian prey in the diet of breeding sharp-shinned hawks in southwest Colorado, 1988-89 12 1.2 General linear model analysis of variance for total (avian and mammalian) prey and avian prey biomass in the diet of breeding sharp-shinned hawks in southwest Colorado, 1988-89 16 1.3 Dominant, secondary, and limited habitats surrounding nests of sharp-shinned hawks in southwest Colorado during 1988-89 breeding seasons , 19 1.4 Taxonomic ranking of mammalian prey consumed by sharp-shinned hawks in dominant, secondary, and limited habitats surrounding nests in southwest Colorado during 1988-89 21 1.5 Size ranking of mammalian prey consumed by sharp-shinned hawks in dominant, secondary, and limited habitats surrounding nests in southwest Colorado during 1988-89 22 1.6 Food-niche breadth values of breeding sharp-shinned hawks in Alaska, Colorado, Oregon, and Utah based on numbers of prey species, genera, and size-class categories used in each study ...24 2.1 Characteristics of accipiter nest sites in southwest Colorado, 1988-89 0 •••••••••••••••••••••••••••••••• 45 �230 LIST OF FIGURES Figure Page 1.1 Gunnison and Grand Mesa National forests in Colorado 3 1.2 Proportion of total (birds and mammals) and avian prey consumed during incubation, nestling, and fledgling stages at nests of sharp-shinned hawks in southwest Colorado, 1988-89 13 1.3 Proportion of total and avian prey consumed during incubation, nestling, and fledgling stages by sharp-shinned hawks in southwest Colorado, 1988-89 14 1.4 Mean biomass (vertical bar) and SE (vertical line) of avian and mammalian prey during incubation, nestling, and fledgling stages at nests of sharp-shinned hawks in southwest Colorado, 1988-89 ..17 1.S Prey (avian and mammalian) availability and use curves of sharp-shinned hawks in southwest Colorado, 1988-89 23 2.1 Relative location of goshawk (NOGO), Cooper's hawk (CORA), and sharp-shinned hawk (SSRA) nests in 1 region of Gunnison and Grand Mesa National forests, Colorado, 1987-89 43 2.2 Aspect of sharp-shinned hawk, Cooper's hawk, and goshawk nest sites in southwest Colorado, 1988-89 46 Directional exposure of sharp-shinned hawk, Cooper's hawk, and goshawk nests in southwest Colorado, 1988-89 48 2.3 �231 Chapter 1 FEEDING ECOLOGY OF BREEDING SHARP-SHINNED HAWKS IN COLORADO In Colorado, sharp-shinned hawks (Accipiter striatus) occur in quaking aspen (Populus tremuloides) and conifer (Abies, Picea, Pseudotsuga)-dominated forests. Aspen stands comprise over 25% of the state's 4.58 million ha of commercial timberland (Green and Van Hooser 1983) and most (> 70%) of these stands are in mature (70-120 yr) and old-growth (120+ yr) age classes (Shepperd 1990). Management practices such as clear-cutting, thinning, and removal of invading conifers have been prescribed to encourage regrowth of aspen stands, especially in areas where aspen is seral to conifers (Jones et al. 1985). Reynolds (1989) suggested that differences in foraging habitats used by sharp-shinned hawks may be more closely related to prey availability than to tree species composition or habitat structure. Alterations in forest structure and/or composition may affect bird and mammal populations, which may in turn affect reproductive performance of sharp-shinned hawks. Small birds (10-30 g) associated with forest canopies comprise approximately 95% of the hawk's diet (Craighead and Craighead 1956, Storer 1966, Reynolds and Meslow 1984). Flack (1976) demonstrated that aspen canopies in 6 western states, including Colorado, contained 41% (818 of 1,975 individuals counted) of all breeding birds in the 4 nesting guilds he defined (canopy, shrub, hole, �232 and ground). Birds nesting in the canopies of mature aspen forests in western Colorado were 53% more abundant than in the other nesting guilds (Flack 1976). The upper-most stratum of Colorado's aspen forests thus appears to contain abundant food resources for foraging hawks. To understand the potential effects of aspen harvests on sharp-shinned hawks, pre-harvest data on hawk/forest relationships and hawk use of prey relative to prey availability in aspen and conifer forests are needed. During 1988-89r I examined the food habits of breeding sharp-shinned hawks in mature aspen, conifer, and mixed aspen-conifer forest types. Sizes and numbers of prey delivered to nests were compared among nests, nesting stages, and between prey taxa. Relationships between diet composition and resource availability were examined. Diet breadths of sharp-shinned hawks breeding in montane and subalpine regions of 4 states, including Colorado, were reviewed to learn if hawks in Colorado used prey similarly to sharp-shinned hawks elsewhere in the U.S. STUDY AREA The study was conducted in Gunnison and Grand Mesa National forests (Fig. 1.1) in Gunnison, Delta, and Mesa counties, Colorado. Large (> 500 ha) and small (100-200 ha) climax aspen stands and conifer communities existed in these forests between 2,750 and 3,200 m elevation. Principal conifers were subalpine fir (~. lasiocarpa), Engelmann spruce (l. engelmannii), and blue spruce (l. pungens). precipitation (primarily snow) averaged 50.7 cm (Natl. Oceanic and Annual �,, ,, , ,, .. _.....-- GRAND MESA N. F. , ,, , , ,, , , N .,. ,, ,, ,, .. , , ,, , ,, ,, , ,, ,, , ,, , , ,, GUNNISON N. F. ,, ,, .. , ,, , ,, ,, , • DENVER ...--------1 .. .. .' o .. ' 50 100 -» ' ' .. .' .' ' .' -» COLORADO Fig. 1.1. Gunnison and Grand Mesa National forests in .Colorado. Slash marks delineate approximate area searched. N W W �geranium (Geranium richardsonii) , Barbey larkspur (Delphinium barbeyi), white-flowered peavine (Lathyrus leucanthus), and monkshood (Aconitum columbianum). Prominent low-growing plants included elk sedge (Carex geyeri), wild strawberry (Fragaria ovalis), yellow prairie violet (Viola nuttallii), and fringed brome (Bromus ciliatus). The shrub component consisted primarily of snowberry (Symphoricarpos spp.), chokecherry (Prunus virginiana), and true mountain-mahogany (Cercocarpus montanus). Conifer understories were dominated by low-growing herbs including heart-leafed arnica (Arnica cordifolia), field horsetail (Eguisetum (Vaccinium myrtillus) and kinnikinnik (Arctostaphylos uva-ursi). Plant names follow Weber (1976). Portions of the forests were interspersed with privately-owned lands used for livestock grazing. In addition, sheep and cattle foraged freely in the study area during July and August. Consequently, most of the study area had been grazed by the end of each field season. METHODS Prey Collection Prey remains (flight and contour feathers, bills, and feet of birds; tails, fur, skull fragments, and feet of mammals) were collected �235 from nesting areas once a week. As changes in nesting stage (incubation, nestling, or fledgling) approached, prey remains were collected every 2 days. Nesting-stage changes were often indicated by changes in behavior of breeding hawks (e.g., females became increasingly aggressive prior to and during egg-hatch). Boundaries of nesting areas were delineated based on observations of prey exchanges, plucking, feeding, and roosting behavior of pairs around nests. were recorded from blinds placed near nests. Pair activities The area of activity near the nest tree, exclusive of foraging areas, was defined as the nest site. Nest sites were searched completely for prey remains during each nest visit. Prey remains were collected from nest sites until young dispersed or nests failed. Diet Composition Parts of uneaten prey were sorted by nest, collection date, size, color, and texture. Avian remains were reconstructed and compared with National Museum of Natural History (Washington, D.C.) specimens, identified, and counted following Reynolds and Meslow (1984). Muddy or matted feathers were washed with Ivory Snow and dried with compressed air before being identified (R. Laybourne, U.S. Fish and Wildlife Service, pers. commun.). Single prey feathers were not included, as they may have been from molting birds. Mammalian prey were processed similarly at the U.S. Fish and Wildlife Service museum in Fort Collins, Colorado. Adult prey weights were obtained from Hall (1946), Armstrong (1972), and Dunning (1984), or from museum specimens (APPENDIX 1). Prey were identified to species where possible. Weights of prey �236 identified to genera level were estimated by averaging weights of all species within the genus that occurred in the study area (Reynolds and Mes10w 1984). "Unknown birds" and "unknown sparrows" were assigned weights based on the mean weight of birds and mean.weight of sparrows in the study area, respectively. Sex-specific weights were used when sex of prey was known; otherwise, the mean weight of both sexes was assigned. Avian prey were classified to age (adult, subadu1t, or nestling) by plumage and amount of sheathing on flight feathers (Reynolds and Meslow 1984). Adults were defined as individuals with unsheathed or complete feathers, whereas, partly sheathed feathers were classified as from subadu1ts. Remiges on subadult birds emerged faster than retrices, which allowed further age classification. Subadults with unsheathed remiges and partly sheathed retrices were assigned the adult weight because they were nearly grown; subadu1ts with partly sheathed remiges and retrices were assigned thre£-quarters of their adult weight. Feathers from nestlings were defined as completely sheathed and assigned one-half the adult weight. Only adult weights could be assigned to mammals. Prey frequencies were compared among nests and nesting stages, between adult and young (subadult and nestling) avian prey, and between avian and mammalian taxa using a maximum likelihood log-linear model (MLLM) (CATMOD; SAS Inst. Inc. 1987). General linear models (GLM) (GLM; SAS Inst. Inc. 1987) were used to test for differences in total prey (avian and mammalian) and avian prey biomass among nests and nesting stages, between adult and young avian prey, and between avian and mammalian taxa. Data collected from unsuccessful nests (n - 5) were �237 excluded from the analyses. I tested the null hypothesis of no difference among the independent effects. All analyses were evaluated at the a - 0.05 significance level. Multi-response permutation procedure (MRPP; Mielke and Berry 1982, Biondini et al. 1988) was used to test for differences in distribution of avian and mammalian prey weights among nesting stages. from all nests (n - 11) were used in this procedure. Prey remains Distributional differences were examined further in pairwise tests of equal variances using Moses' rank-like procedure (Hollander and Yolfe 1973). In addition, nonparametric tests of equal means (analysis of variance) and medians (Yilcoxon rank-sum test) were performed with SAS (NPARlYAY; 1987). In all analyses, I tested the null hypothesis of no difference among independent effects. the a All pairwise comparisons were evaluated at 0.05/3 - 0.017 (Bonferroni inequality) level. Foraging Zone I compared prey numbers taken from each of 5 foraging zones [ground-shrub, shrub-canopy, canopy, aerial, and generalist (Reynolds and Meslow 1984)] (APPENDIX 1) using K2 goodness of fit tests (Steel and Torrie 1980) to examine if hawks in Colorado foraged equally in all zones. Birds were assigned to the zone in which they were observed most often. Birds not observed, but found in remains at nests, were assigned zones following Reynolds and Meslow (1984:APPENDIX 1). Resource Use and Availabili~y wi~h Respect to Habitat Prey use was compared with prey availability in 3 forest cover levels (dominant, secondary, and limited) surrounding nests to learn if �prey were used in proportion to their occurrence in these habitats. Forest composition within a 2-km radius of each nest was ascertained from aerial photographs (1:24,000). Sharp-shinned hawks forage primarily within an area of 1.5 km from the nest (Reynolds 1983); therefore, a 2-km radius was presumed to enclose most foraging areas used by the hawks. Cover types comprising ~ 50% of the area surrounding nests were termed "dominant". forest types occupying> "Secondary" habitats were defined as 5% to ~ 49% of £oraging areas. Cover types making up ~ 5% of surrounding areas were termed "limited". Because hawks were presumed to not forage extensively in limited cover types, tests of prey use versus availability in these forests were designed to assess differences in prey availability between restricted and more "common" (dominant and secondary) cover types. That is, were there differences in prey use between these habitat levels that might explain the low occurrence of limited types within the hawks' foraging radius? The null hypothesis that prey were equally used in each habitat level was tested using program PREFER (Johnson 1980); prey items were classified by size and taxon. each habitat level. A separate analysis was performed for PREFER examined the relationship between ranks of prey use and availability for each nest to identify if breeding hawks used prey in proportion to their availability in the 3 habitat levels. Use and availability data were paired on individual years to eliminate possible differences in prey abundance between years. Waller and Duncan's (1969) mUltiple comparison procedure was used in PREFER to assess relative component "preferencen• Analyses in PREFER were limited to sizes and taxa of prey likely to be consumed by sharp-shinned hawks so that number of nests would exceed �239 the number of size or taxon classes and produce a more robust analysis. Program PREFER was dimensioned for I (number of nests) - 1 (number of prey sizes or families) + 1 degree of freedom, and J ~ 1. Avian prey were assigned weights (Dunning 1984) and partitioned into 6 size classes following Storer (1966) (APPENDIX 2). Mammalian prey were assigned weights (Armstrong 1972, 1975) and partitioned into 9 size classes (Storer 1966). Mammalian prey in classes 9 (166-216 g), 6 (64-'91.1g), and 2 (8-15.6 g) were excluded from analyses because they were not found in remains nor counted during surveys (APPENDIX 2). Ten .avian and 5 mammalian taxonomic categories (families) were also used in the analyses (APPENDIX 2). To ensure that number of families did not exceed number of nests, families of avian prey in surveys, but containing S 2 consumed individuals at all nests were excluded from analyses. To further reduce the number of families, Sittidae and Certhiidae were combined based on their similar behavior. Weasels (Mustelidae) were excluded from analyses because they are unlikely to be consumed by sharp-shinned hawks. Finally, prey use was compared with the total range of "available" bird and mammal sizes surveyed in the area to identify overall patterns in prey size use by the hawks. Quantitative estimates of avian and small mammal availability during the 1988 and 1989 breeding seasons were provided by R. T. ,Reynolds (U.S. For. Serv., unpubl. data). Bird abundances (raw counts) were estimated between mid-June and mid-July using variable circular-plots on 15 40-ha plots (D. M. Finch and R. T. Reynolds, U.S. For. Serv., Study Plan RM-420l.l-l). Mammal abundances (raw counts) were estimated between late-July and late-August with baited live and pitfall traps on 15 IOxlO grids (15 m trap spacing) (total trap nights - �~4U 27,OOO/yr) (R. T. Reynolds and D. M. Finch, U.S. For. Serv., Study Plan RM-420l.l-2). The 15 plots (9 in aspen, 3 in conifer, 3 in mixed aspen-conifer cover types) were systematically sampled each year. Survey data were assumed to reflect (1) approximate abundances of bird and mammal populations present throughout the hawks' breeding season in each forest type and (2) approximate prey availabilities perceived by the hawks. Diet Breadth Food-niche breadths were calculated for breeding sharp-shinned hawks in Alaska (Clarke 1984), Colorado (this study), Oregon (Reynolds and Meslow 1984), and Utah (D. L. Fischer, unpubl. data) to examine geographical variation in resource use (Levins 1968:43). Calculations were based on the use coefficient, ~i1' which represents the proportion of food resource i used by consumer 1. Prey were grouped by species, genus, and size classes (Storer 1966). RESULTS • Prey Collection Eleven sharp-shinned hawk nests were studied during May 1988 (4) and 1989 (7). Most (91%) nest sites (exclusive of foraging areas) were within small (1-14 ha), insular conifer stands surrounded by aspen or mixed aspen-conifer stands. forest. One nest was in a contiguous conifer No nests were depredated in 1988 allowing collection of prey from 4 nest sites during all nesting stages. Five of 7 nests were depredated in 1989; thus, prey remains were only collected from 2 nest sites during all nesting stages, 7 nest sites during incubation, and 4 �241 nest sites during fledgling stages. A total of 686 prey items was identified, including 52 species and 39 genera of birds and 11 species and 10 genera of mammals (APPENDIX 1); 513 of these.prey (i.e, those from successful nests) were used in MLLM and GLM procedures. Mammals were found at 10 of 11 nests and birds were found at all nests. Diet Composition Prey Frequency. of the diet. Small birds comprised the largest (91.1%) portion The proportion of mammals in the diet ranged from 1.2 to 11.8% for nests dominated by aspen (n - 5) to 28.3% for 1 nest z, do~inated by conifers (indicated in a nest by taxon interaction; Table 1.1). Prey'numbers varied interactively with nest and nesting stage; more prey were consumed during the nestling period at the majority of nests (67%) than during other stages (Table 1.1, Fig. 1.2). Numbers of prey also differed interactively with prey taxon and nesting stage (Table 1.1). During all breeding stages, hawks. fed primarily on birds (Fig. 1.3); however, the proportion of mammals in the diet increased from 7.8 to 16.5% between incubation and fledgling stages. Voles (Clethrionomys, Microtus, and Phenacomys) comprised over 60% of the mammals eaten. Voles were also counted more often (n - 1,077) during surveys than other diurnal or mostly diurnal small mammals (Sorex, n 69.2;Tamias, n - 478; Thomomys, n - 18; and Zapus, n 37). Numbers of birds consumed differed with independent effects: nests, nesting stages, and prey ages (Table 1.1). Of the 466 birds used in MLLM analyses, 86 (18%) were found during the fledgling period,' 144 (31%) were during incubation, and 236 (51%) were during the nestling period. Overall, adults comprised a larger proportion (61%) of the �242 Table 1.1. Maximum likelihood log-linear model for numbers of total (avian and mammalian) prey and avian prey in the diet of breeding sharp-shinned hawks in southwest Colorado, 1988-89. Interactions are indicated by an "x". Source of variation J.2 df f. Total Prey Nest "5 74.07 < 0.001 Breeding stage- 2 4.50 0.105 Nest x Breeding stage 10 37.07 < 0.001 Taxon (Birds, Mammals) 1 89.83 < 0.001 Nest x Taxon 5 25.02 < 0.001 Breeding stage x Taxon 2 7.39 0.025 10 8.41 0.589 GOODNESS OF FIT Avian Prey Nest 5 183.14 < 0.001 Breeding stage 2 49.59 < 0.001 10 23.93 0.008 Age (Adult, Youngb) 1 32.81 < 0.001 Nest x Age 5 0.31 0.379 Breeding stage x Age 2 54.96 < 0.001 Nest x Breeding stage GOODNESS OF FIT 10 .i!lJ 8.41 -Incubation, nestling, and fledgling stages. bSubadults and nestlings. ,,'_., 0.838 �243 I OINCUBATION [ill NESTLING IFLEDGLINGI TOTAL PREY 224 53 85 61 34 56 100 80 I- 60 ::l 40 Z o o > m > w a:: c, I- 20 o AVIAN PREY 204 55 84 38 30 55 1 2 3 4 5 6 1..00 Z w 80 o a:: w 60 Q. 40 20 o NEST Fig. 1.2. consumed Proportion during sharp-shinned columns of total (birds and mammals) incubation, hawks nestling, in southwest are prey counted/nest. and fledgling Colorado, 1988-89. and avian prey stages at nests of Numbers above �244 TOTAL PREY 100% I• 100 BIRDS 0 MAMMALS 236 86 eo I- Z ::l o o 60 40 > 20 m o > W100 a: n. 36 12 AVIAN PREY 139 I OADUL T {8JYOUNG I I- eo Z W 60 o a:: w a. 48 130 40 20 5 o INCUBA TION NESTLING FLEDGLING BREEDING STAGE Fig. 1.3. Proportion of total and avian prey consumed during incubation, nestling, and fledgling stages by sharp-shinned hawks in southwest Colorado, 1988-89. Prey were partitioned by taxon (birds, mammals); avian prey were partitioned by age (adult, young). numbers are above each column. Prey I �245 avian diet than young (39%). The number of avian prey at individual nests varied interactively with nesting stage (Table 1.1). As with total prey numbers, more avian prey were found during the nestling than incubation or fledgling stages at the majority (67%) of nests. Age differences in avian numbers also occurred among nesting stages (Table 1.1). During incubation, sharp-shinned hawks fed primarily on adult birds (Fig. 1.3). Young birds (subadults and nestlings) became increasingly more important in the diet as the breeding season progressed. Yellow-rumped warblers (Dendroica coronata) appeared most often in the avian prey remains (n - 78), followed by American robins (Turdus migratorius) (n - 52), white-crowned sparrows (Zonotrichia leucophrys) (n - 44), and dark-eyed juncos (~hyemalis) Prey Biomass. The mean weight of all prey items differed between taxa (Table 1.2); mammals comprised almost twice the biomass ~ ± SE - 41.1 ± 3.3 g) of birds hawks' diet. voles. (n - 41). (n - (n - 47, 466, ~ ± SE - 20.9 ± 0.8 g) in the Over 54% of the prey biomass of mammals was contributed by Fifty percent.of the avian biomass was comprised of yellow-rumped warblers, American robins, white-crowned sparrows, and dark-eyed juncos. Interactions in total prey biomass existed among prey taxa (birds, mammals), nests, and nesting stages (Table 1.2, Fig. 1.4). The null hypothesis of no difference in avian prey biomass among main effects or interactions was not rejected (Table 1.2). Prey-weight Distribution. The null hypothesis that avian prey weights came from a common distribution among nesting stages was rejected (E - 0.051). showed significant (E stages. Pairwise MRPP comparisons of the distributions 0.011) variation between incubation and nestling Differences in measures of distributional variance �Table 1.2. General linear model analysis of variance for total (avian and mammalian) prey and avian prey biomass in the diet of breeding sharp-shinned hawks in southwest Colorado, 1988-89. Interactions are indicated by an "x". Source of variation df SS Total Prey Nest 5 1,342.00 0.87 0.500 Breeding stage- 2 1,169.25 1. 90 0.151 10 4,199.25 1.36 0.195 Taxon (Birds, Mammals) 1 4,343.62 14.09 < 0.001 Nest x Taxon 5 1,328.34 0.86 0.506 Breeding stage x Taxon 2 1,054.51 1.71 0.182 8 3,993.77 2.59 0.025 3,·033.24 2.12 0.062 Nest x Breeding stage Nest x Taxon x Breeding stage Avian Prey Nest 5 Breeding stage 2 151.46 0.26 0.768 10 4,828.30 1.69 0.082 Age (Adult, Youngb) 1 171.18 0.60 0.440 Nest x Age 5 1,022.71 0.71 0.613 Breeding stage x Age 2 341.03 0.60 0.552 Nest x Breeding stage x Age 8 4,356.00 1. 90 0.158 Nest x Breeding stage -Incubation, nestling, and fledgling stages. bSubadults and nestlings. �o 60 '"' m INCUBA TION 1m NESTLING ~FLEDGLING '"'" >w a: o, LL 40 BIRDS 20 o C/) C/) « s oJ o 20 z « w 40 -m s ,...., r::: MAMMALS 60 80 1 2 4 3 5 6 NEST Fig. 1.4. Mean biomass (vertical bar) and SE (vertical line) of avian and mammalian prey during incubation, nestling, and fledgling stages at nests of sharp-shinned hawks in southwest Colorado, 1988-89. N .p. "-I �(dispersion), as well as mean weight between the 2 stages, were not detected [l ~ 0.017 (Bonferroni inequality) for all tests]. However, median avian weights differed (l< 0.001) between incubation and nestling stages. Median weight of birds taken during incubation was 17.4 g compared to 12.1 g during the nestling period. Median bird weight increased to 14.7 g during the fledgling period. Density distributions of mammalian prey weights did not differ (l - 0.323) among nesting stages. Foraging Zone Sharp-shinned hawks captured food with equal frequency from ground-shrub, shrub-canopy, and canopy foraging zones (Xl - 0.121, df - 2, f - 0.942). differentially (X2 - Other zones (aerial, generalist) were used 62.3, df - 1, l< 0.005), with more prey .taken from the generalist (124 items or 18% of total prey) than aerial (27 items or 4% of total prey) zone. Zone use appeared to reflect relative prey availabilities in each foraging layer (28, 24, 30, 16, and 2%, respectively). Resource Use and Availability with Respect to Habitat Habitat designations for each nest varied (Table 1.3). Avian prey sizes were used in proportion to their availability in dominant (l ~ 0.25) and secondary (l - 0.172) habitat types. Avian size use differed (l - 0.03?) from availability in limited cover types. Bird taxa were consumed in proportion to their occurrence in all habitats (0.15 ~ l ~ 0.25). �249 Table 1.3. Dominant, secondary, and limited habitats surrounding nests· of sharp-shinned hawks in southwest Colorado during 1988-89 breeding seasons. Cover typeh Nest Year Dominant Secondary Limited 1 1988 A M C 2 1988 A M C 3 1988 A M C 4 1988 C M A 5 1989 A M C 6 1989 A M C 7 1989 A M C 8 1989 A M C 9 1989 A M C 10 1989 M A C 11 1989 M C A ·Dominant - cover types comprising ~ 50% of foraging areas; Secondary - cover types comprising> 5% to ~ 49% of foraging areas; Limited - cover types comprising ~ 5% of foraging areas. hA (aspen), C (conifer), M (mixed aspen-conifer). �Sharp-shinned hawks appeared to select similarly among size classes and taxa of mammals (l~ 0.03 for all habitat designations) (Tables 1.4, 1.5). Dipodidae and Geomyidae were most used among taxa, and Sciuridae and Sorcidae were least used in all habitat categories. Muridae, which contained all vole species and 1 deer mouse, was used most closely in proportion to its availability in all habitats, i.e., it had the lowest absolute mean difference between ranks of use and availability among taxa. Highest ranked among weight categories of mammals were size classes 8 (125.0-166.0 g), 7 (91.1-125 g), and 4 (27.0-42.9 g); size class 1 (3.6-8.0 g) was least used. Among the total range of bird and mammal sizes occurring in the study area, sharp-shinned hawks used only smaller (logloweight 0.7-2.1 g) size classes (Fig. 1.5). Within these smaller classes, the hawks consumed avian prey in nearly equal proportion to their overall occurrence. A lack of congruence between availability and use curves for m£mma1ian prey confirmed that hawks concentrated foraging efforts on only a few sizes of mammals. Diet Breadth Sharp-shinned hawks in Colorado preyed more uniformly over a wider range of prey species than hawks in either Oregon or Alaska (Table 1.6). Genera of prey were selected more uniformly by hawks in Colorado and Oregon than in Alaska (Table 1.6). Niche-breadth values with respect to biomass did not appear to differ among hawks breeding in Alaska, Colorado, Oregon, and Utah (Table 1.6). �251 Table 1.4. Taxonomic ranking of mammalian prey consumed by sharp-shinned hawks in dominant, secondary, and limited habitats surrounding nests in southwest Colorado during 1988-89. Taxon classes and average difference between ranks (MRD) of use and availability are Taxonomic use differed (f $ 0.01) in listed from most to least used. all habitats. Cover typeDominant Taxon Secondary MRDb Taxon MRD Limited Taxon MRD Dipodidae -1.50 Dipodidae -1.40 Dipodidae -1.70 Geomyidae -0.90 Geomyidae -1.00 Geomyidae -0.70 Muridae -0.35 Muridae -0.35 Muridae -0.35 Sciuridae 0.80 Sorcidae 1.05 Sorcidae 1.15 Sorcidae 1.95 Sciuridae 1.70 Sciuridae 1.60 -Dominant - cover types comprising ~ 50% of foraging areas; Secondary - cover types comprising> 5% to $ 49% of foraging areas; Limited - cover types comprising $ 5% of foraging areas. bMean rank difference. Negative values indicate proportional use was greater than availability; positive values indicate proportional use was less than availability. �252 Table 1.5. Size ranking of mammalian prey consumed by sharp-shinned hawks in dominant, secondary, and limited habitats surrounding nests in southwest Colorado during 1988-89. Taxa and average difference between ranks (MRD) of use and availability are listed from most to least used. Taxonomic usage differed (l~ 0.03) in all habitats. Cover typeDominant Weight (g) (Size c Las s") Secondary Limited Weight (g) (Size class) MRDc Weight (g) (Size class) (8) -1.30 125.0-166.0 (8) -1.15 91.1-125.0 (7) -1.20 91.1-125.0 (7) -0.75 27.0-42.9 (4) -0.70 27.0-42.9 {4) -1.15 42.9-64.0 (5) -0.55 91.1-125.0 (7) -0.70 42.9-64.0 (5) -0.S5 27.0-42.9 (4) -0.65 42.9-64.0 (5) -0.55 125.0-166.0 (8) -0.20 lS.6-27.0 (3) 0.8S lS.6-27.0 (3) 1.1S 15.6-27.0 (3) 1.15 3.4-8.0 (1) 2.25 3.4-8.0 (1) 1.95 3.4-8.0 (1) 1.95 125.0-166.0 MRD MRD -Dominant - cover types comprising ~ 50% of foraging areas; Secondary - cover types comprising> 5% to ~ 49% of foraging areas; Limited - cover types comprising ~ 5% of foraging areas. bFo1lowing Storer (1966). during surveys were excluded. Sizes classes of mammals not counted CMean rank difference. Negative values indicate proportional use was greater than availability; positive values indicate proportional use was less than availability. �253 40 AVAILABLE o AVIAN PREY USED 30 ---8--- 20 > 0 Z W 10 :J o W a: 0 W 35 l<t 30 0.9 0.5 1.3 2.1 1.7 2.5 2.9 3.3 3.7 LL - > ..J ~4 , , , , PREY I I , , 25 I I W a: MAMMALIAN I I I I I 20 . , I 15 • , 10 I. 5 0 0.5 0.9 1.3 1.7 2.1 2.5 2.9 3.3 3.7 4.1 4.5 4.9 5.3 5.7 LOG PREY WEIGHT (g) Fig. 1.5. Prey (avian and mammalian) availability and use curves of sharp-shinned hawks in southwest Colorado, 1988-89. �254 Table 1.6. Food-niche breadth values of breeding sharp-shinned hawks in Alaska, Colorado, Oregon, and Utah based on numbers of prey species, genera, and size-class categories used in each study. Food-niche breadtha Location H nests Prey species (n) Prey genera (n) Size classesb (n) Source Alaska 14 12.95 (41) 8.37 (32) 3.28 (6) Clarke (1984) Colorado 11 21.73 (63) 18.11 (49) 3.37 (8) This study Oregon 14 12.67 (44) 15.52 (37) 3.53 (9) Reynolds and Meslow (1984) 3 C Utah C 4:49 (6) D. L. Fischer (Unpubl. data) ·Ca1culated per Levins (1968). bFo1lowing Storer (1966). CData unavailable. .. �255 DISCUSSION Prey Collection Schipper determination biased and Possible Data Bias (1973) and Snyder and Wiley based on prey remains and casts collected estimates of diet composition. were underepresented Reptiles, in diets of the birds (1973) also found that in harrier also underestimated. amphibians) (1976) showed that diet Because and invertebrates diet must be detected they studied. lower vertebrates possibility of detecting difficult nesting to find. arthropods, of stomach amphibians, Another consumed Schipper contents casts were difficult and reptiles of the sharp-shinned and in the or casts 1984). excluded Casts were small of stomach contents and birds were their presence and Meslow and therefore I could not determine Analysis 5% by number) these prey. Furthermore, stage because regurgitated. « or cast contents and insects (e.g., reptiles are rarely plucked, (Snyder and Wiley 1976, Clarke 1984, Reynolds stomach amphibians, (Circus) diets, mammals through analyses not analyze at nests gave I did the (0.2-0.7 cm) and to assign to a when they had been was impractical. However, appear to make up a small portion hawk diet (Snyder and Wiley 1976). possible. bias in the data includes prey that were plucked and/or away from the nest site. may not depict all possible Thus, diet compositions prey consumed reported here by the hawks. Diet Composition Although hawk's geographic patterns birds. prey-species composition varies range, most investigators in prey-size Small birds use. the sharp-shinned have observed Hawks in Colorado (10-30 g) also comprised across primarily similar consumed a large portion small (> 95%) of �256 the hawks' diet in Wyoming, Craighead 1956, Storer addition, sharp-shinned 1966, Clarke hawks sizes and taxa of mammals reflected Michigan, present and maintaining foraging young, efforts vulnerability density nest defense, or understory use were reflected avian prey ages, vegetation by changes nest-taxon-nesting incubation increased energy both parents adults or fledgling demands stage. (Snyder and Wiley nest leaving fewer remains prey numbers responded consumed in foliage in prey taxa, and only in a the nestling stages, which probably During 1976). reflected the fledgling Also, after young at the nest site. adept at predator more often than young stage, to declining fledged, to forage prey (1) and their own food and eaten it away from the corresponded to seasonal period fewer prey in response Temporal to the abundance in terms of when young birds hatched, became prey differences fewer prey to encourage may have caught and/or require with nests, in prey biomass delivered (2) females avian prey changes Temporal in prey numbers of producing In addition, with seasonal of growing young. may have delivered overall demands stage interaction. probably populations of sharp-shinned and self-maintenance More prey were found at nests during during on certain This probably patterns The energetic height. and by changes In to predation. that vary with nesting may vary temporally and 1984). preyed in the study area. stages. (Craighead and Mes10w successfully (Clarke 1984) have examined hawk prey use across nesting and Oregon 1984, Reynolds in Colorado size or taxon vulnerability Few studies Alaska, changes escape. during in and vulnerability fledged, Consequently, in prey availability. than mammals changes all nesting of and migrated hawks may have Although stages, birds were mammalian �257 prey frequency reduction increased between in bird numbers incubation in mammalian Sharp-shinned stages. of mammals, was responsible hawks fed primarily season; however, nestling birds consumed and nestling the proportion through of adult, subadult, changed with nesting almost entirely stages, on adult birds the ~roportion avian prey consisted stage. of adult birds. of young birds incubation consume more young prey during the nestling the tendency period. (17.4 g) and of hawks prey .. Others (Newton 1979, Geer 1982, Clarke 1984, Newton stage prey biomass 1.4.). mammals period differences fully open. vegetated in use of prey biomass a pattern At 2 nests dominated were consumed corresponded During mid-summer and small mammals foraging sharp-shinned hawks. with nest, was apparent by aspen of young and Marquiss taxa, and regarding (nests 3 and 6, Fig. only during the incubation to early summer ground-shrub with the appearance period. (Jul) , the understory (especially Aspen understory all remaining available throughout nests became those associated remained (Fig. 1.4), the nesting This (Jun) when aspen understory season. was not densely with the zone) may no longer have been available At nearly of in accipiters. (Table 1. 2. Fig. 1.4). use. to Thus, nesting hawks similar patterns coincides In addition, sharp-shinned nesting in Colorado incubation in the diet increased there was a shift toward lighter prey between (12.1 g) stages, which reflected incubation, Between period. nestling the and During 5-fold and remained high during the fledgling Despite for the prey frequency. breeding 1982) reported A found at nests during the latter nesting stage, rather than elevated numbers increase and fledgling to dense until late August. mammals were apparently �Several investigators (Snyder and Wiley 1976, Clarke 1984, Reynolds and Meslow 1984) reported that during incubation and nestling stages, male sharp-shinned hawks provide .most of the prey. Females begin to forage during the late-nestling stage and continue to make prey deliveries through the fledgling period. Although females weigh about 69% more than males (Dunning 1984), and thus have the potential to capture larger prey (Storer 1966), they do not appear to do so. I detected no difference in mean prey biomass among nesting stages (l - 0.151). One explanation for the absence of larger prey during late-nesting stages may be that males catch a large portion of the prey delivered by females. During late-nesting scages , males were occasionally observed transferring food to females away from the nest; females then flew toward their nests with the prey. Behavioral differences among sharp-shinned hawks may exist as some hawks plucked more prey at nests than other pairs. Alternatively, variation in prey detectability and vulnerability among foraging sites may have caused differences in prey numbers among nests. It appeared that hawks at nests dominated by aspen habitat consumed fewer mammals than at 1 nest surrQunded by mostly coniferous habitat; however, insufficient numbers of nests in conifer habitat precluded statistical comparison between habitat types. In the relatively open understory of a conifer forest, small mammals may be more vulnerable to sharp-shinned hawks than in densely vegetated aspen understories. Foraging Zone Sharp-shinned hawks in Colorado foraged in nearly equal proportion in the canopy, shrub-canopy, and ground-shrub layers of the forest. �259 Conversely, sharp-shinned hawks in Oregon concentrated foraging efforts in upper portions of the forest canopy (Reynolds and Meslow 1984). An interdependence between prey size and foraging-zone occurrence (Reynolds and Meslow 1984) may explain these patterns of zone use. Small birds in Oregon tended to occur in upper portions of the canopy, whereas in Colorado they were distributed evenly among ground-shrub to canopy layers of the forest. A higher proportion of mammals in the diets of hawks in Colorado (9%) versus Oregon « 5%) may have account~d for different levels of ground-shrub use between studies. Most (92%) mammals consumed by hawks in Colorado occurred in the ground-shrub layer of the forest; most (64%) mammals taken by hawks in Oregon were zone generalists or canopy dwellers (Reynolds and Meslow 1984). Habitat differences between Oregon and Colorado study areas may have caused differences in prey vulnerability, and hence prey availability, among foraging zones. Study sites in Oregon consisted primarily of dense, contiguous conifer or conifer-dominated mixed stands (Reynolds et al. 1982). These forests allow little light to penetrate to lower vegetation levels, which may result in sparse undergrowth. Conversely, aspen stands permit sufficient light to penetrate the canopy to support abundant shrub, herb, and forb growth (Mueggler 1985). Due to insufficient cover, fewer small birds and mammals may exist in lower vegetation levels of Oregon's conifer forests than in the well vegetated understory of Colorado's aspen forests. Resource Use and Availability with Respect to Habitat An absence of preference for specific bird sizes and taxa in dominant and secondary habitats indicated that hawks foraged �260 opportunistically in each habitat. In limited habitats, avian prey use differed from availability for size classifications but not for taxa, suggesting that prey-size abundance or vulnerability varied between limited and ncommonn (dominant and secondary) habitats. Sharp-shinned hawks nested in conifers, yet coniferous habitat comprised 82% of the limited cover types; mature aspen and mixed aspen-conifer comprised 73 and 82% of dominant and secondary habitats, respectively. Perhaps hawks selectively foraged in aspen or aspen-dominated forests because avian prey of suitable sizes were more available in these forests. Similar taxa of avian prey were apparently available in each habitat. The importance of small avian prey in the diet of sharp-shinned hawks is probably a function of accipiter body size and evolutionary adaptations that permit energetically efficient hunting (Reynolds 1972). A sharp-shinned hawk's small body size limits the size of prey that it can capture and subdue without injuring itself. Within this .prey-size limit, the hawk exploits the entire range of available avian prey sizes. Diet Breadth Niche-breadth values provide comparable estimates of the extent of specialization in food selectivity among populations of breeding sharp-shinned hawks. A wider range and greater equitability of prey species were used by sharp-shinned hawks in Colorado than in Oregon or Alaska. The same pattern was true for classes of prey genera for hawks in Colorado and Oregon versus Alaska. However, hawks in Alaska, Colorado, Oregon, and Utah used size classes of prey similarly. Sharp-shinned hawks thus appear capable of exploiting variable numbers of prey taxa, while they are restricted to a relatively limited r~nge of �261 prey sizes. reflect Hawks Taxon niche-breadth the relative in Colorado differences availability consumed among these regions may of prey in each habitat twice as many species studied. of mammals and 1.3 times as many species of birds as hawks in Oregon or Alaska. habitats species used by hawks richness MANAGEMENT and evenness than habitats hawks conifers; appeared however, Manipulative needed forests. In addition, layers of the forest canopy while foraging. habitats integrity conifers appeared to be important of mature aspen, conifer, the nesting sharp-shinned be performed in mosaics hawks. experimentally, Mature of these forests areas for sharp-shinned with mature for nesting. and mixed cover types are Until controlled to perpetuate hawks aspen and foraging value of these habitats aspen management forest perturbation should maintain suitable nesting D. M. Univ. Kansas 1975. Assoc., 1972. Printing of mammals Serv., Lawrence. Rocky Mountain mammals. Inc., Estes Park, CO. 174 pp. in Colorado. 415 pp. Rocky Mountain can and foraging in Colorado. Distribution to the LITERATURE CITED Armstrong, of they exploited to have more foraging value when compared studies to examine breeding greater prey in Oregon or Alaska. in Colorado used prey resources aspen and aspen-associated several contain IMPLICATIONS Sharp-shinned mature in Colorado may therefore Foraging Nature �Banks, R. C., R. W. McDiarmid, and A. L. Gardner. 1987. Checklist of vertebrates of the United States, the U. S. Territories, and Canada. U. S. Dep. Inter., Fish and Wi1d1. Servo Resour. Pub1. 166, Washington, D.C. 79 pp. Biondini, M. E., P. W. Mielke, and E. F. Redente. 1988. Permutation techniques based on euclidean analysis spaces: a new and powerful statistical method for ecological research. Coenoses 3:155·174. Clarke, R. G. 1984. The sharp·shinned hawk (Accipiter striatus viei11ot) in interior Alaska. Fairbanks. M.S. thesis, Univ. Alaska, 130 pp. Craighead, J. J., and F. C. Craighead, Jr. and wildlife. Hawks, owls, Stackpole Co. and Wi1dl. Manage. Inst., Harrisburg, PA and Washington, D. C. Dunning. J. B. 1956. 1984. American birds. 443 pp. Body weights of 686 species of North Western'Bird Banding Assoc. Monogr. 1. 38 pp. Flack, J. A. 1976. North America. Geer, T. A. 1982. Bird populations of aspen forests in western Ornitho1. Monogr. 19. 97 pp. The selection of tits Parus spp. by sparrowhawks Accipiter nisus. Green, A. W., and D. D. Van Hooser. the Rocky Mountain states. Resour. Bull. INT-33. Hall, E. R. 1946. Los Angeles. Ibis 124:159-167. 1983. U. S. Dep. Agric., For. Servo 127 pp. Mammals of Nevada. 710 pp. Forest resources of Univ. California Press, �263 Hollander, M., and D. A. Wolfe. methods. Johnson, D. H. 1973. 'Nonparametric statistical John Wiley & Sons, Inc., New York, NY. 1980. 503 pp. The comparison of usage and availability measurements for evaluating resource preference. Ecology 6:65-71. Jones, R. J., ,R. P. Winokur, and W. D. Shepperd. Management overview. P. Winokur, eds. Pages 193-195 in N. V. DeBy1e and R. Aspen: western United States. 1985. ecology and management in the U. S. Dep. Agric., For. Servo Gen. Tech. Rep. RM-119. Levins, R. 1968. Evolution in changing environments: theoretical explorations. Monogr. Pop. Biol. 2. Univ. Press, Princeton, N. J. Mielke, P. W., and K. J. Berry. some Princeton 120 pp. 1982. An extended class of permutation techniques for matched pai~s. Commun. Statist. 11:1197-1207. Mueggler, W.' F. 1985. Vegetation associations. N. V. DeByle and R. P. Winokur, eds. Pages 45-55 in Aspen: management in the western United States. ecology and U. S. Dep. Agric., For. Servo Gen. Tech. Rep. RM-119. National Oceanic and Atmospheric Administration. 1988. Climatological data, Colorado. U. S. Dep. Comm., Nat1. Oceanic and Atmospheric Admin. 93(13):5, 11. Newton, I. 1979. Population ecology of raptors. Vermillion, S. D. 399 pp. Buteo Books, �__________ , and M. Marquiss. 1982. Food, predation, and breeding season in sparrowhawks (Accipiter nisus). J. Zool., Lond. 197:221-240. Reynolds, R. T. 1972. new hypothesis. 1983. Sexual dimorphism in accipiter hawks: Condor 74:191-197. Management of western coniferous forest habitat for nesting accipiter hawks. Servo Gen. Tech. Rep. RM-107. 1989. a U.S. Dep. Agric., For. 7 pp. The status of accipiter populations in the western United States. Pages 92-101 in B. Pendleton, K. Steenhof, and M. N. Kockert, eds. Proc. western raptor management symposium and workshop. Nat1. Wildl. Fed., Washington, D. C. __________ , and E. C. Mes1ow. 1984. Partitioning of food and niche characteristics of co~xisting Accipiter during breeding. ~ Auk 101:761-779. , E. C. Meslow, and H. M. Wight. 1982. Nesting habitat of coexisting accipiter hawks breeding in Oregon. J. Wildl. Manage. 46:124-138. SAS Institute Inc. Version 6 Ed. Schipper, W. J. A. 1987. SAS/STAT Guide for Personal Computers, SAS Institute Inc., Cary, NC. 1973. 1028 pp. A comparison of prey selection in sympatric harriers Circus in western Europe. Gerfaut 63:17-120. Shepperd, W. D. 1990. A classification of quaking aspen in the central Rocky Mountains based on growth and stand characteristics. West. J. Appl. For. 5:In press. �Snyder, N. F. R., and J. W. Wiley. 1976. in hawks and owls of North America. Sexual size dimorphism Ornitho1. Monogr. 20:96 pp. Steel, R. G. D., and J. H. Torrie. 1980. Principles and procedures of statistics -- a biometric approach. McGraw-Hill, Inc., New York, NY. Storer, R. W. 1966. 633 pp. Sexual dimorphism and food habits of three North American accipiters. Waller, R. A., and D. B. Duncan. Auk 83:423-436. 1969. A Bayes rule for the symmetric multiple comparisons problem. J. Am. Statist. Assoc. 64:1484-1503. Weber, W. A. 1976. Rocky mountain flora. Press, Boulder, CO. 479 pp. Colorado Assoc. Univ. �Chapter 2 NESTING HABITAT OF ACCIPITERS IN SOUTHWEST COLORADO North American accipiters are sympatric throughout much of the western United States (Reynolds et al. 1982; Moore and Henny 1983; Fischer 1986). During the nesting period, northern goshawks (Accipiter genti1is), Cooper's hawks (a. cooperii), and sharp-shinned hawks (a. striatus) exploit different habitats for nesting (Reynolds et a1. 1982, Moore and Henny 1983) and forage for birds and mammals of different sizes and taxa -(Storer 1966, Reynolds and Meslow 1984). In Colorado, accipiters occur on the east and west slopes of the Continental Divide where they nest and forage in mature quaking aspen (Populus tremu1oides) and conifer (abies, Picea, Pseudotsuga) forests. In western North America, accipiter nest-site characteristics have been quantified in conifer-dominated forests (Reynolds et al. 1982, Moore and Henny 1983), but few data exist on nests in or proximate to extensive deciduous stands. Shuster (1980) reported northern goshawks in predominantly pine (Pinus) forests of western Colorado nested in mature aspen trees (n - 10 nests) more often than in lodgepole pine (l. contorta) (n 5 nests) or ponderosa pine (l. ponderosa) (n - 5 nests); however, nest-site characteristics of other accipiters were not investigated. Fischer (1986) examined relative abundance of nesting accipiters in 6 habitat classes [including an aspen-bigtooth maple (Acer �267 grandidentatum) forest class], but did not identify habitat variables associated with nests in aspen forest. Accipiters appear to use foraging habitats opportunistically (Reynolds 1989). Sharp-shinned hawks, the smallest species, should therefore be capable of using a wider range of habitat than goshawks, the largest species. However, accipiters appear to have specific nesting habitat requirements (Reynolds 1983). In Oregon, sharp-shinned hawks nested in young, dense, even-aged conifer forests; Cooper's hawks used older, less dense, even-aged stands; and goshawks used mature to old-growth, multilayered stands (Reynolds et al. 1982). Aspen stands in Colorado comprise over 60% of the 1.78 million ha of commercial aspen areas in the western United States (Green and Van Hooser 1983). Currently, over 70% (1,484 of 2,004 stands surveyed) of Colorado~s aspen trees are in 70-120+ age classes (Shepperd 1990). Proposed aspen management by commercial thinning, clear-cutting, and . removal of conifers to encourage growth of replacement aspen stands (Jones et a1. 1985) emphasizes the need for data on accipiter use of aspen and aspen-associated forests. This study describes the vegetation and topography at accipiter nest sites in mature aspen, mature conifer, and mixed aspen-conifer forest types. Factors responsible for nesting-habitat use are discussed under the assumption that accipiters attempt to maximize nest productivity. are examined. Sharp-shinned hawk nesting chronology and productivity �STUDY AREA The study area included portions of Gunnison and Grand Mesa National forests in Gunnison, Delta, and Mesa counties, Colorado and was bordered on the north by Colorado Highway 65, northwest by Colorado Highway 330 extending east to the border between Pitkin and Gunnison counties, east by the Ruby Range, and south by the southern border of Grand Mesa National Forest. Climax aspen stands and conifer communities occur in these forests between 2,750 and 3,200 m in elevation. Large (> 500 ha) and small (100-200 ha) stands of mature aspen, conifer, and mixed aspen-conifer were searched for accipiter nests. conifer species were subalpine fir (a. lasiocarpa), engelmannii), and blue spruce (E. pungens). Principal Engelmann spruce (E. At lower elevations, forests contained large (100-500 ha) clearings with low shrub, forb, and grass cover. South-facing slopes were dominated by Gambel's oak (Quercus gambelii). Aspen forests were even-aged, single storied, and possessed narrow, dome-like crowns. Stands were relatively open due to sparse canopy cover and an absence of dead branches to within -1 m below the crown. Mixed aspen-conifer forests appeared more dense than pure aspen forests, and often possessed multi-layered canopies. Understory vegetation of both pure and mixed aspen stands was dense. Conifer forests were single- or multi-storied depending on the level of past disturbance (fire, cutting, disease) or deterioration, especially in old-growth (120+ yr) stands. All conifer forests had dense overstories and sparse understories. Understory vegetation types varied depending on elevation, slope, aspect, and forest cover type. Aspen and mixed aspen-conifer �lb9 understories were mostly herbaceous with a minor shrub component. Dominant forbs included butterweed groundsel (Senecio ~), white geranium (Geranium richardsonii) , Barbey larkspur (Delphinium barbeyi), white-flowered peavine (Lathyrus leucanthus), and monkshood (Aconitum co1umbianum). Prominent low-growing plants included elk sedge (Carex geyeri), wild strawberry (Fragaria ovalis), yellow prairie violet (Viola nuttallii), and fringed brome (Bromus ci1iatus). The shrub component consisted primarily of snowberry (Symphoricarpos), chokecherry (Prunus virginiana), and true mountain-mahogany (Cercocarpus montanus). Conifer understories were dominated by low-growing herbs including heart-leafed arnica (Arnica cordifo1ia), field horsetail (Eguisetum arvense), and elk sedge, and low shrubs including myrtle blueberry (Vaccinium myrti11us) and kinnikinnik (Arctostaphylos uva~ursi). Plant names follow Weber (1976). Private lands within the national forests were used primarily for livestock grazing. Sheep and cattle foraged freely in the study area during July and August. Most of the study area had been grazed by the end of the field seasons. METHODS Ground searches for accipiter nests began in May 1988 and 1989, and continued through the breeding season (incubation, nestling, and fledgling stages). Stands with characteristics of accipiter nest sites in western montane forests (Platt 1973, Shuster 1976, Hennessy 1978, Reynolds et al. 1982, Moore and Henny 1983, Fischer 1986) were selectively searched, although "unsuitablen stands were also explored en route to more nsuitablen stands. Stands for intensive search efforts �were selected from aerial photographs (1:24,000) and topographic maps (7.S-min series) and included mature aspen, conifer, and mixed aspen-conifer habitats near ephemeral (drainages, marshes) and permanent (streams, creeks, ponds) sources of water. In addition, known sharp-shinned hawk nest sites occupied in previous years (1 used in 1986, 1 used in 1986 and 1987, and 2 used in 1987) (R. T. Reynolds, U.S. For. Serv., unpubl. data) were examined. Active nest sites were identified by the presence of a nest, breeding pair, and plucking areas with recent signs of activity (e.g., feathers, fur, casts, feces, and molted accipiter. feathers). Nest site is defined as "the area (immediately) surrounding the tree ...used by a nesting pair during an entire nesting season, exclusive of foraging areas" (Reynolds et al. 1982:126). Vegetation and physiographic characteristics of active accipiter nest sites were measured after young dispersed or nests failed. variables measured included: Habitat 1) species, age, height, and diameter at breast height (dbh) of the nest tree; 2) nest-site elevation, aspect, slope, and canopy cover (the latter measured by point-centered quarter method) (Cottam and Curtis 1956); 3) "nest height and directional exposure; 4) distance of nest tree to water; 5) dominant understory vegetation (estimated visually as the species covering the largest proportion of ground); and 6) average stand age. Heights and slopes were measured with a clinometer, aspects and exposures with a compass, dbh with a logger's cm tape, canopy cover with a densiometer, distances with a meter tape, and elevations from topographic maps. Stand age was estimated by coring the nest tree and 3-5 randomly selected overstory trees in the nest site, then counting and averaging the core rings. �· Z /l. Twenty years were added to ring-counts of conifers for tree growth to breast height (Alexander 1974), whereas 0 years were added to aspen ring-counts (Y. D. Shepperd, U.S. For. Serv., unpubl. data). Differences in nest-tree height, dbh, and age, nest-site slope, and nest height between sharp-shinned and Cooper's hawks were tested with Mann-Whitney tests (NPAR1YAY; SAS Inst. Inc. 1987) and evaluated at the a - 0.10 significance level. Rayleigh's R statistic (Zar 1984) was used to test the null hypotheses that accipiter (1) nest-site aspect and (2) nest directional exposure were randomly distributed. were evaluated at the a - 0.05 significance level. Rayleigh's tests All aspects and exposures (0-359°) were assumed to be equally available. In addition, mean directional aspects and exposures and their angular deviations (AD) were determined for each species (Rayleigh's procedure, Zar 1984). Sharp-shinned hawk nests were visited 2-3 times each week and their status recorded until fledglings dispersed. Nesting chronology of goshawks and Cooper's hawks was not monitored. RESULTS Approximately 28,750 ha of national forest lands were searched for nesting accipiters. Teams of 2-4 individuals spent 160, 100, and 100 hours searching for nests in contiguous conifer, aspen, and mixed aspen-conifer stands, respectively. Twenty accipiter nests were found including 7 sharp-shinned hawk, 2 Cooper's hawk, and 1 northern goshawk nest in 1988 and 7 sharp-shinned hawk. 2 Cooper's hawk, and 1 northern goshawk nest in 1989. Of 4 sharp-shinned hawk nest sites used in 1987, 1 (25%) was reused in 1988. Two (29%) of the 7 sharp-shinned hawk nest sites found during 1988 were reused in 1989. At reoccupied sites, all �272 new nests (n - 3) were built < 40 m from former nests. Cooper's hawks did not reuse nest sites from previous years; however, the goshawk site was reoccupied. The 1988 and 1989 goshawk nests were 150 m apart.. Mean clutch size for 6 sharp-shinned hawk nests was 4.7 ± 0.2 (SE) eggs. Hatching occurred between 6 and 13 July in 1988 (n - 4), and on 28 June (n - 1) and between 4-6 July (n - 2) in 1989. nests were found post-hatch. The remaining Assuming a 30-32 day incubation period (Platt 1973, Hennessy 1978, Reynolds and Wight 1978), egg laying was initiated in late May and incubation in early June. Fledging occurred between 28 July and 8 August in 1988 and between 2 and 6 August in 1989. Thus, duration of the nestling period varied between 20 and 33 days. An average of 3 fledglings was observed at nest sites during 1988 (n - 8 nests, SE - 0.2) and 1989 (n - 2 nests, SE - 1). Fledgling dispersal occurred from mid- to late-August during both years. All (n - 7) sharp-shinned hawk nests in 1988 were successful; however, only 2 of 7 nests fledged young during 1989. A female and male disappeared at 2 nests and nestlings at 3 nests were depredated. One Cooper's hawk nest was depredated in 1988; young successfully fledged from the remaining nests (n - 3). Goshawk nests were successful both years. Most (70%) accipiter nests found were in 700-2,000-ha sections of forest separated by 3-6 km of forest containing no nests (Fig. 2.1). Average distance between active nests of sharp-shinned hawks (nearest neighbor) was 2.2 km and ranged from 1.0 to 2.5 km. Two sharp-shinned hawk nests were approximately 0.5 and 1.5 km from active Cooper's hawk nests, and another sharp-shinned hawk nest was within 1.2 km of a �NATIONAL FOREST BOUNDARY \ \/ ./ • ~ ,-.",. / . I J • • GRAND MESA NATIONAL FOREST . Km I I -1 -DENVER I 0 __ 1 /' • GUNNISON NATIONAL F~REST COLORADO, N .. .. .. I .. .' ...... ~i ..____ .. . ... J. ... :. .. .·1 o SSHA 1988 • SSHA 1989 ~COHA 1987 6COHA 1988 ONOGO 1988 .. Fig. 2.1. .. . . .. .NOGO 1989 Relative location of goshawk (NOGO), Cooper's hawk (COHA), and sharp-shinned hawk (SSHA) nests in 1 region of Gunnison and Grand Mesa National forests, Colorado, 1987-89. state map represents the expanded area. Slash marks delineate areas searched. Shaded portion of N .....• w �274 goshawk nest. Nests of goshawks and Cooper's hawks were not found in the same section of forest. Accipiter nest sites ranged from 2,354 (Cooper's hawk) to 2,935 m (goshawk) in elevation and were located 3 (goshawk) to 400 m (sharp-shinned hawk) from water (Table 2.1). All sharp-shinned hawk nests were in small (1-14 ha), insular conifer or mixed aspen-conifer stands surrounded by either aspen forest (10), mixed forests (3), or conifer forests (1). sparse. Understory vegetation at all nest sites was Surrounding forests varied from pure aspen (3) to mixed (1) forests for Cooper's hawk nests, whereas contiguous conifer forests surrounded goshawk nests. Carex, Arnica cordifolia, and species of moss dominated understories at sharp-shinned hawk nests. at goshawk nest sites. Carex was dominant Other prominent species included Vaccinium, Eguisetum, moss, and young 6. lasiocarpa. Dominant understory species were variable at nest sites of Cooper's hawks. Eleven of 14 sharp-shinned hawk nests were in mature (70-120 yr) stands; 1 nest was in an old-growth (120+) stand, and 2 nests were in young (60 yr) stands. Goshawks and Cooper's hawks nested in old-growth abd mature stands, respectively. Nest-site slope gradient did not differ (l - 0.14) between sharp-shinned and Cooper's hawks, but appeared steeper for sharp-shinned hawks than goshawks (Table 2.1). aspect for all species was northerly (n - 20, l ~0.002) (Fig. 2.2). n- & - 11°, AD - 54°; Mean aspects ranged from northwest (311°, 2, AD - 68°) for goshawks to northeast (51°, Cooper's hawks. Nest-site n- 4, AD - 56°) for Site canopy cover was high (~ 94%) for all 3 species. The majority (3 of 4) of Cooper's hawk nests were in aspen trees. Sharp-shinned hawks nested only in conifers (Abies, Picea) and goshawks �275· Table 2.1. Characteristics of accipiter nest sites in southwest Colorado, 1988-89. Accipiter Sharp-shinned hawk (n - 14) Characteristic a Cooper's hawk (n - 4) i SE Goshawk (n - 2) SE SE Nest Site Elevation, m 2,682 55 2,590 80 2,931 5 130 32 157 60 25 4 Slope, % 38 5 21 3 14 4 Stand age, yrs 86 5 84 6 Canopy cover, % 94 2 95 3 Distance to water, m 148 22 95 2 Ne.st Tree Species· AL (9), PE (4), PP (1) AL (1), PT (3) PT Height, m 19 1 23 3 26 0 Dbhb 29 2 36 7 44 4 Age, yrs 93 8 102 7 167 5 Height, m 11 1 16 1 19 o Canopy cover, % 99 1 97 2 94 2 Nest aAL - Abies lasiocarpa, PE - Picea engelmannii, PP - f. pungens, PT - Populus tremuloides; sample sizes are in parentheses. bDiameter of tree measured at breast height. �SHARP-SHIN'NED HAWK. COOPER'S HAWK .•• GOSHAWK. , • ~ ~ ~. I .,:() -, 1.. :::::::: :HHHn: • • • Fig. 2.2. 1988-89. Aspect of sharp-shinned hawk, Cooper's hawk, and goshawk nest sites in southwest Colorado, �'L only in aspen trees. II Nests of Cooper's and sharp-shinned hawks were in i 102 yrs, SE - 7 and n - 11, R - 93 yrs, trees of similar age (n 4, SE - 8, respectively; l 0.47), whereas goshawk nests were in old-growth aspen trees (n - 2, R - 167 yrs, SE - 5). Tree height and dbh were similar between sharp-shinned and Cooper's hawks (l - 0.22 for both tests), and appeared to be somewhat taller and larger for goshawks. Nest height was lower for sharp-shinned than Cooper's hawks (l - 0.06), and appeared to be lower than goshawk nests. On average, nests of goshawks and Cooper's hawks were at 73 and 70% of nest-tree height, respectively; sharp-shinned hawk nests were at 58% of total tree height. Whereas nests of sharp-shinned hawks were within the tree crown, Cooper's hawk nests were generally in the lower portion of the crown, and goshawk nests were below the crown. .Mean accipiter nest exposure in nest trees was southeasterly (n - 20, R- 135°, AD - 55°, l S 0.05) and ranged from 127° (n ~ 14, AD - 47°) for sharp-shinned hawks to 208° (n - 2, AD - 45°) for goshawks (Fig. 2.3). Nest-tree canopy cover did not appear to differ among species. DISCUSSION Despite the greater effort searching for nests in contiguous conifer habitat, more accipiter nests were found within large stands of aspen or aspen-dcmfnared forests. The northern goshawk was the only species that nested within conifer habitat. Small, insular conifer or mixed aspen-conifer stands were important to sharp-shinned hawks, as indicated by their exclusive use of these stands. Nest concealment and protection may have affected sharp-shinned hawk preference for these forests. This accipiter appears less �~- SHARP-SHINNED HAWK ~ I COOPER'S HAWK ~ GOSHAWK. t( \L (ll ~ El1illD ~ ~ Fig. 2.3. Directional exposure Colorado, 1988-89. 0 sharp-shinned (j) _,. hawk, Cooper's hawk, and goshawk nests in �279 effective in defending associated nests against predators. with deep, conical crowns of conifers not provided (Reynolds by the high, relatively 1989). The requirement explain why sharp-shinned conifers. leaf-out Layered Because offers protective open crowns of mature for nest concealment sharp-shinned hawks initiated conceal~ent of nest-building cover aspen trees would also hawks placed their nests within crowns of nest building of aspen trees, conifer or densely-stocked stands provided foliage prior to mixed conifer-aspen activities that pure aspen stands did not. The sharp-shinned or aspen-dominated Scott and Crouch conifer forests composition hawks' preference stands possibly (1988) reported of west-central between habitats to nest within stems from foraging that bird densities Colorado did. Sharp-shinned and Meslow birds occurring western Colorado in surveys) Flack's 1984, Chapter in the upper-most 1). than in other strata Flack (1976) demonstrated (shrub, hole, ground). hawks hawks. that nested of counted Although abundances between aspen contain In this study, the single pair of in conifer habitat consumed more mammals by aspen-dominated (7% of diet). The goshawk, because on that aspen forests (53%; 178 of 338 individuals (28% of diet) than hawks at 5 nests surrounded forests feed primarily forest types, his data showed that aspen canopies ample food for sharp-shinned sharp-shinned but species with the forest canopy (1976) study did not compare nest-guild and conifer in aspen and hawks stratum of mature were more abundant aspen requirements. did not differ, small ~irds. many of which are associated (Reynolds extensive of its larger size and ability nest, may have little need for nest concealment. Rather, to defend prey its �280 detectability and vulnerability, and foraging habitat structure probably affected goshawk nest-site choice and placement within the forest landscape. Goshawks feed largely on vertebrates associated with lower portions of the vegetative column (Reynolds and Meslow 1984). The openness of the conifer understory probably enhanced prey vulnerability and goshawk maneuverability more so than dense herbaceous understory in the aspen forests. Cooper's hawks, intermediate in size between goshawks and sharp-shinned hawks, require more open understories for unrestricted flight than sharp-shinned hawks and perhaps more nest concealment than goshawks. Despite nestling depredation at 1 nest, 75% of Cooper's hawk nests were clearly visible in mature aspen trees. Nest concealment, therefore, may not be as important to Cooper's hawks in southwest Colorado as to Cooper's hawks elsewhere in the U.S. (Reynolds et al. 1982, Moore .and Henny 1983). Nest shading and/or accessibility may be more critical in nest placement (Reynolds et al. 1982, Kennedy 1988). Cooper's hawks' nests in this study were surrounded by mature aspen (3 nests) or mixed aspen and conifer (1 nest) forests. Factors affecting Cooper's hawk nest proximity to mature aspen forests may be similar to those influencing sharp-shinned hawk nest-site choice (e.g., prey composition and foraging success). Overall accipiter nest-site aspect was northerly and nest exposure southeasterly. Reynolds et al. (1982) attributed the tendency of accipiters in Oregon to nest on sites with northerly aspects to a preference for cool, shaded sites. Northerly sites receive less direct sunlight during summer months and thus minimize the possibility of heat stress to nestlings. Northerly sites also experience less moisture loss �281 due to evaporation, southern-exposed exposure which results nests may be a function well as thermoregulation. Branch density side of the trunk, especially densities tree density sites (Reynolds et al. 1982). of accipiter southern in greater provided improved than dryer, The southeasterly of tree conformation tended to be greater in conifers. shade, nest support, Greater and perhaps as on the branch nest concealment. Nest-site goshawks. nested gradient was steep for sharp-shinned Shuster sharp-shinned (1980) also found that goshawks (& - on gentle 12.5%) slopes. hawks appeared power. Apparent in western Slope gradient to be steeper differences at nests of perhaps in nest-site due to lack of slope among species may result from the hawks' prefer.ence for specific are associated Distance to watEr varied greatly among species, was higher required Previous observed from the greater availability trees than Cooper's to support investigators that nest-site different that Finally, water Nest-site or Cooper's hawks, of old-growth goshawks or sharp-shinned appeared to use hawks, which were their large nests. (Reynolds et al. 1982, Moore and Henny characteristics forest successional among stands used by accipiters with young although for nesting. than sharp-shinned stands at higher elevations. larger-diameter pr?bably to all accipiters for goshawks and may have resulted conifer habitats with slope rather than choice of slope directly. seemed to be important elevation Colorado than at nests of Cooper's hawks, but they·did not differ statistically, statistical hawks and gentle for stages. of accipiters Associated [sharp-shinned (40-60 yr) stands, Cooper's corresponded 1983) to with age differences hawks were identified hawks with older (50-80 yr) �stands, and goshawks with mature (80+ yr) stands] were structural differences among nest sites. Conclusions about habitat partitioning among Colorado accipiters are speCUlative due to the small number of nests found. However, habitat partitioning among Colorado birds appears to be related less to stand age differences (particularly between Cooper's and sharp-shinned hawks), and more to structural differences, among forest stands. Sharp-shinned hawks require dense stands of conifers; Cooper's hawks use less dense stands of aspen or mixed aspen-conifer cover types; and goshawks use relatively open coniferous forests interspersed· with large aspen trees. In southwest Colorado, the apparent absence of habitat partitioning on tree age was perhaps due to a dearth of younger-aged stands. Aspen-associated forests also may exhibit less stand-age variability than pine- and fir-dominated forests surveyed in previous studies .. ~AGEMENT IMPLICATIONS Few studies of nesting accipiters (but see Shuster 1976, 1980) have focused on use of aspen forests. Aspen management, primarily clear-cutting, has been proposed for much of Colorado's aspen timberland and emphasizes the need for information on accipiterfhabitat relationships. The data suggest that sharp-shinned hawks are vulnerable to predation and require protective cover provided by conifers at the nest. Foraging success in surrounding aspen habitats may also influence nest-site location for this species. Whereas northern goshawks appear to require relatively open, old-growth coniferous forests for foraging, as well as large aspen trees for nesting, Cooper's hawks primarily use aspen-dominated habitats for foraging and nesting. �283 Because accipiters require specific stand characteristics for nesting (Reynolds 1983), active and potential nest sites should not be modified by timber harvests. Tree cutting or thinning would alter stand structure and vegetative composition, and perhaps make nest sites unsuitable for accipiters. Alternatively, because accipiters use forested lands opportunistically for foraging (Reynolds 1989), and include fragmented (whether natural or man-made) as well as continuous forests within their home ranges (Reynolds 1983), limited timber management may occur in habitats used for foraging, but only in years during which forests are not used by accipiters. Guidelines for aspen management should be developed on based recommendations obtained from this and future studies on accipiter use of these forests. LITERATURE CITED Alexander, R. A. 1974. Silviculture of subalpine forests in the central and southern Rocky llountains: the status of our knowledge. u.S. Dep. Agric., For. Servo Res. Paper RM-121. 88 pp. Cottam, G., and J. T. Curtis. 1956. in phytosociological sampling. Fischer, D. L. 1986. hawks in Utah. Provo, UT. Flack, J. A. The use of distance measure Ecology 37:451-460. Foraging and nesting habitat of Accipiter Ph.D. dissertation, Brigham Young Univ. , 33 pp. 1976. North America. Bird populations of aspen forests in western Ornithol. Monogr. 19. 97 pp. �Green, A. W., and D. D. Van Hooser. the Rocky Mountain states. Resour. Bull. INT-33. Hennessy, S. P. 1978. 1983. Forest resources of U. S. Dep. Agric., For. Servo 127 pp. Ecological relationships of accipiters in northern Utah -- with special emphasis on the effects of human disturbance. M.S. thesis, Utah State Univ., Logan. 66 pp. Jones, R. J., R. P. Winokur, and W. D. Shepperd. Management overview. P. Winokur, eds. Pages 193-195 in N. V. DeByle and R. Aspen: western United States. 1985. ecology and management in the U. S. Dep. Agric., For. Servo Gen. Tech. Rep. RM-119. Kennedy, P. L. 1988. Habitat characteristics of Cooper's hawks and northern goshawks nesting in New Mexico. in Pages 218-227 R. L. Glinski, B. G. Pendleton, M. B. Moss, M. N. LeFranc, Jr., B. A. Millsap, and S. W. Hoffman, eds. southwest raptor management symposium and workshop. Proc. Natl. Wildl. Fed., Washington, D. C. M oore, K . R ., an d C ..J Henny . 1983 . Nest site characteristics of three coexisting accipiter hawks in northeastern Oregon. Raptor Res. 17:65-76. Pl at,t J .'B 1973 . Habitat and time utilization of a pair of nesting sharp-shinned hawks (Accipiter striatus velox). M. S. thesis, Brigham Young Univ., Provo, UTe 41 pp. R eyno lds , R . T . 1983. Management of western coniferous forest habitat for nesting accipiter hawks. Servo Gen. Tech. Rep. RM-107. 7 pp. U.s. Dep. Agric., For. �285 1989. The status of accipiter populations in the western United States. Pages 92-101 in B. Pendleton, K. Steenhof, and M. N. Kockert, eds. Proc. western raptor management symposium and workshop. Natl. Wildl. Fed., Washington, D. C. __________________ , and E. C. Meslow. 1984. Partitioning of food and niche characteristics of coexisting Accipiter during breeding. Auk 101:761-779. and H. M. Wight. 1978. Distribution, density, and productivity of accipiter hawks breeding in Oregon. Bull. Wilson 90:182-196. _______________ , E. C. Mes1ow, and H. M. Wight. 1982. Nesting habitat of coexisting accipiter hawks breeding in Oregon. J. Wildl. Manage. 46:124-138. SAS Institute Inc. 1987. Version 6 Ed. SAS/STAT Guide for Personal Computers. SAS Instit~te Inc., Cary, NC. Scott, V. E., and G. L. Crouch. 1988. 1028 pp. Summer birds aridmammals of aspen-conifer forests in west-central Colorado. Dep. Agric., For. Servo Res. Paper. Shepperd, W. D. 1990. RM-280. U.S. 6 pp. A classification of quakirigaspen in the central Rocky Mountains based on growth and stand characteristics. Shuster, W. C. 1976. montane Colorado. 1980. West. J. Appl. For. S:In press. Northern goshawk nesting densities in West. Birds 7:108-110. Northern goshawk nest site requirements in the Colorado Rockies. West. Birds 11:89-96. �1989. The status of accipiter populations in the western United States. Pages 92-101 in B. Pendleton, K. Steenhof, and M. N. Kockert, eds. Proc. western raptor management symposium and workshop. Nat1. Wi1dl. Fed., Washington, D. C. _________________ , and E. C. Meslow. 1984. Partitioning of food and niche characteristics of coexisting Accipiter during breeding. Auk 101:761-779. and H. M. Wight. 1978. Distribution, density, and productivity of accipiter hawks breeding in Oregon. Bull. Wilson 90:182-196. _________________ , E. C. Meslow, and H. M. Wight. 1982. Nesting habitat of coexisting accipiter hawks breeding in Oregon. J. Wildl. Manage. 46:124-138. SAS Institute Inc. 1987. Version 6 Ed. SAS/STAT Guide for Personal Computers. SAS Instit~te Inc., Cary, NC. Scott, V. E., and G. L. Crouch. 1988. 1028 pp. Summer birds aridmammals of aspen-conifer forests in west-central Colorado. Dep. Agric., For. Servo Res. Paper. Shepperd, W. D. 1990. RM-280. U.S. 6 pp. A classification of quakirigaspen in the central Rocky Mountains based on growth and stand characteristics. Shuster, W. C. 1976. montane Colorado. 1980. West. J. App1. For. 5:In press. Northerrigoshawk nesting densities in West. Birds 7:108-110. Northern goshawk nest site requirements in the Colorado Rockies. West. Birds 11:89-96. �Storer, R. W. 1966. Sexual dimorphism and food habits of three North American accipiters. Weber, W. A. 1976. Rocky mountain flora. Press, Boulder, CO. Zar, J. H. 1984. Auk 83:423-436. 479 pp. Biostatistical analysis. Englewood Cliffs, NJ. Colo. Assoc. Univ. 718 pp. Prentice-Hall, Inc., �APPENDICES • �APPENDIX 1. Prey species at 11 sharp-shinned hawk nest sites in southwest Colorado during 1988-89 breeding seasons. Common names are in parentheses. Adult Foraging weighta•b Prey Actitis macularia (spotted sandpiper) Selasphorus platycercus (broad-tailed hummingbird) Sphyrapicus nuchalis (red-naped sapsucker) Picoides pubescens (downy woodpecker) g. villosus (hairy woodpecker) Picoides spp. (woodpecker) Contopus sordidulus (western wood-pewee) Empidonax hammondii (Hammond's flycatcher) E. oberholseri (dusky flycatcher) E. difficilis (western flycatcher) ·Empidonax spp. (flycatcher) Progne subll (purple martin) Tachycineta bicolor (tree swallow) I. thalassina (violet-green swallow) Parus atricapillus (black-capped chickadee) g. gambeli (mountain chickadee) Parus spp. (chickadee) Sitta canadensis (red-breasted nuthatch) Certhia americana (brown creeper) Salpinctes obsoletus (rock wren) Troglodytes aedon (house wren) Regulus satrapa (golden-crowned kinglet) zcne" (g) Adult (n) Prey age SubadultC Nestlingd (n) 1 40.4 2 2 3.6 1 3 50.3 3 3 27.0 6 3 66.3 2 3 46.7 1 3 12.8 13 3 10.1 10 3 10.4 6 3 10.9 4 2 3 10.3! 2 1 4 49.4 1 4 20.1 16 4 14.2 8 3 1 (l)g 2 -10.8 5 9 (3) 2 10.8 9 18 (2) 2 10.8 2 3 9.8 1 3 8.4 3 1 .16.5 1 1 10.9 23 3 6.2 3 1 5 (n) �APPENDIX 1. Continued. Prey R. calendula (ruby-crowned kinglet) Regulus spp. (kinglet) Sia1a mexicana (western bluebird) ~. curro ides (mountain bluebird) M~adestes townsendii .(Townsend's solitaire) Catharus ustulatus (Swainson's thrush) g,yttatus (hermit thrush) Catharus spp. (thrush) Turdus migratorius (American Robin) Dumate11a caro1inensis (gray catbird) Vireo solitarius (solitary vireo) y. gilvus (warbling vireo) y. olivaceus (red-eyed vireo) ~spp. (vireo) Vermivora celata (orange-crowned warbler) Dendroica petechia (yellow warbler) Q. coronata (yellow-rumped warbler) Dendroica spp. (warbler) Oporornis to1miei (MacGillivray's warbler) Yi1sonia pusi11a (Yilson's warbler) Piranga ludoviciana (western tanager) Pipilo chlorurus (green-tailed towhee) Aimophila spp. (sparrow) ~. Adult Foraging weight zone (g) Adult (n) Pre~ age Subadult Nestling .(n) (n) 6 (1) 3 6.7 11 3 6.4 2 5 28.1 2 5 29.6 1 5 34.0 1 1 30.8 1 1 31.0 10 1 30.9 5 5 77 .3 27 2 36.9 1 2 16.6 6 2 12.0 14 2 16.7 2 15.1 1 2 9.0 2 2 9.8d'd' 3 9.2~~ 12.1 3 10.8 1 1 10.4 5 2 6.9 8 3 28.1 6 5 2 29.4 5 1 1 19.7 1 1 2 4 (8) 2 15 (9) 14 (3) 1 2 (1) 7 3 39 4 1 38 (1) 1 �APPENDIX 1. Continued. Prey Adult Foraging weight zone (g) Sgizella arborea (American tree sparrow) .§.. gasserina (chipping sparrow) Pooecetes gramineus (vesper sparrow) Passerculus sandwichensis (savannah sparrow) Melospiza melodia (song sparrow) t1. lincolnii (Lincoln's sparrow) Zonotrichia leucophrys (white-crowned sparrow) Junco hyemalis (dark-eyed junco) Unknown juvenile sparrow Agelaius phoeniceus (red-winged blackbird) Molothrus ater (brown-headed cowbird) Pinico1a enucleator (pine grosbeak) Carpodacus cassinii (Cassin's finch) s:!. mexicanus (house finch) Cardue1is pinus (pine siskin) Coccothraustes vesgertinus (evening grosbeak) Unknown juvenile bird Sorex montico1us (montane shrew) Tamias minimus (least chipmunk) Tamaisciurus hudsonicus (Pine squirrel) Thomomys talpoides (northern pocket gopher) Peromyscus maniculatus (deer mouse) C1ethrionom;:!sgapgeri (Gapper's red-backed vole) Adult (n) Pre;:!age Subadult Nestling (n) (n) (1) 2 20.1 5 12.3 1 25.7 1 20.1 6 2 20.8 5 2 17.4 26 1 1 25.5 35 6 (3) 5 19.6 25 12 (4) 2 21.1 2 1 41.5 (1) 2 49.0rU 1 3 38.8~~ 56.4 8 1 3 26.5 2 3 21.4 2 5 14.6 10 3 59.4 1 1 1 1 (1) 22.1 1 1 (2) 1 10 2 1 6.8h 2 5 46.1 3 5 225.3 2 1 94.61 7 1 24.6 1 1 26.8 17 4 �APPENDIX 1. Continued. Prey Phenacom::lsintermedius (heather vole) Microtus montanus (montane vole) M. longicaudus (long-tailed vole) Microtus spp. (vole) Zapus princeps .(western jumping mouse) Unknown mouse Adult Foraging weight zone (g) Adult (n) 1 36.8 3 1 45.3 2 1 46.8 5 1 46~1 10 1 23.8 7 1 24.2 1 Pre::la~e Subadult Nestling (n) (n) -Bird weights are from Dunning (1984) and mammal weights are from Armstrong (1972) unless otherwise specified. bSubadults with unsheathed remiges and partly sheathed retrices were assigned the adult weight; subadults with partly sheathed remiges and retrices were assigned 3/4 (adult weight); nestlings received 1/2 (adult weight). CBirds with partially sheathed feathers. dBirds with fully sheathed retrices and remiges . •(1) ground-shrub; (2) shrub-canopy; (3) canopy; (4) aerial; (5) generalist. fWeight - 1/2(E. hammondii weight + E. oberholseri weight) &Numbers in parentheses are subadults with unsheathed remiges (adult weight). hHall (1946). iU.S. Fish and Wildlife Service museum specime~ labels, Fort Collins, CO. �APPENDIX 2. Size and taxonomic classification of prey used in preference analyses. Familyb Size class (.g). Avian prey Mammalian prey Avian prey Mammalian prey 3.4-8.0 (1) 3.4-8.0 (1) Picidae Sorcidae 8.0-15.6 (2) 15.6-27.0 (3) Tyrannidae Sciuridae 15.6-27.0 (3) 27.0-42.9 (4) Hirundinidae Geomyidae 27.0-42.9 (4) 42.9-64.0 (5) Paridae Muridae 42.9-64.0 (5) 91.1-125.0 (7) Sittidae/Certhiidae Dipodidae 64.0-91.1 (6) 125.0-166.0 (8) Trog1odytidae Muscicapidae Vireonidae Emberizidae Fringillidae ·Size classes according to Storer (1966). Size classes of mammals not found in remains nor counted during surveys were excluded. bFami1y names follow Banks et al. 1987. Prepared-by Approved by �JOB PROGRESS REPORT State of: Project: Colorado W-152-R Upland Bird Research Work Plan: 21 Job Title: Evaluation of Habitat Quality on Conservation Reserve Lands in Eastern Colorado Period Covered: Author: Job 5 _ 01 January through 31 December 1990 Warren D. Snyder Personnel: David C. Wilson and Warren D. Snyder, Colorado Division of Wildlife ABSTRACT Visual obstruction readings (VQR) , the primary index to grass-forb quality for wildlife, were 0.47 in early spring 1990 and 0.83 dm in mid summer within 140 Conservation Reserve Program (CRP) fields sampled in eastern Colorado. This overall sample included 105 fields randomly selected for.sampling in 1988. Early spring VOR indices averaged 0.39 dm contrasted to 0.67 dm in mid summer for this group (including 7 fields cost shared by the Division of Wildlife). These 7 fields, combined with 35 additional fields cost-shared by the Division of Wildlife, yielded average VQR indices of 0.66 dm (early spring) and 1.09 dm (summer). F.ields containing switchgrass (Panicum virgatum), grass-alfalfa, and grass-sweet clover mixtures had the highest summer VQR indices. Precipitation stimulated early spring and late summer vegetation growth, however, June was extremely hot and dry. Canopy cover averaged 52.3% in spring and 54.2% in summer 1990 of which seeded perennials respectively comprised 40 and 51%. A relationship between percent land in the CRP and ring-necked pheasant (Phasianus colchicus) crowing count indices could not be detected in spring 1990. �EVALUATION OF HABITAT QUALITY ON CONSERVATION RESERVE LANDS IN EASTERN COLORADO Warren D. Snyder INTRODUCTION Data collection continued for the third year as part of an ongoing study evaluating vegetation quality on Conservation Reserve Program (CRP) fields in eastern Colorado. Two data sets were under assessment. The first was a random sample of 105 fields for which data were supplied to the u.S. Fish and Wildlife Service, National Ecology Center (NEC) at Fort Collins. Seven of those fields were included within a second data set of 42 CRP fields which the Division of Wildlife cost shared to obtain better cover quality. This is the second year of evaluation of CDOW cost-shared fields. P. N. OBJECTIVES Determine distribution and quantity of Conservation Reserve Program land in eastern Colorado in relation to distribution of selected wildlife species, evaluate the quality of vegetation on these lands for selected wildlife species, measure response of selected wildlife species to the Conservation Reserve Program using existing annual surveys, and evaluate the impact of the Colorado Division of Wildlife's cost-share program on cover quality. SEGMENT OBJECTIVES 1. Conduct evaluations of randomly selected CRP fields within eastern Colorado as part of a regional and national assessment of the Conservation Reserve Program coordinated by the National Ecology Center (NEC) of the U.S. Fish and Wildlife Service. 2 .. Conduct intensive visual obstruction readings (VOR).in a stratified random sample of CRP fields and proximal controls; 3. Conduct intensive visual obstruction readings (VOR) in a sample of fields cost shared by the CDOW (for enhancement of cover quality) for comparison with CRP'fields not cost-shared. 4. Estimate response of selected wildlife species to the CRP based on annual population surveys. 5. Compile and analyze data and prepare the annual progress report. METHODS Methods used were described by Snyder (1989 and 1990). Visual obstruction readings were based on the Kirsch method of sampling rather than the procedure required by NEC personnel. However, data supplied to the NEC were converted to their index procedure. �L~O RESULTS AND DISCUSSION Environmental Conditions Precipitation averaged among 9 eastern Colorado weather stations through the first 4 months of 1990 was greater than that received during the same interval in 1989 (Tables 1, 2). Growth of cool-season grasses responded favorably (Table 3). May precipitation was near average, however, June was extremely hot and dry throughout eastern Colorado (Table 2). Above average rainfall was received in July. Increased precipitation continued through fall and early winter 1990 when compared to 1989. Extreme southeastern Colorado (represented by Springfield) consistently received above average precipitation in recent years and more moisture than most other eastern Colorado areas in 1988, 1989, and 1990 (Table 1). However, rainfall was more evenly ·distributed through eastern Colorado in 1990 except for Greeley. Although numerous fields had been mowed or sprayed during summer 1989 (23X), such efforts decreased in 1990 primarily because of advanced grass establishment. Herbicide treatments occurred in late June and July on several fields (16, llX) primarily in Baca County although most fields contained good stands of sideoats g~ama (Bouteloua curtipendu1a) and/or other species. One field in Prowers County was severely hailed in mid-June, 1 field in Pueblo County was redrilled, and a field in Yuma County was disked, apparently for replanting. Observations subsequent to the summer survey revealed that several fields, dominated by annual weeds, were mowed in mid to late summer 1990. Table 1. Annual precipitation eastern Colorado, 1987-90. (in.) at 9 U.S. Weather Bureau locations in Long.:.term Location Springfield Lamar Rocky Ford Limon Burlington Akron Greeley Holyoke Sterling Combined average ~ 16.65 15.40 12.50 15.09 17.41 .17.34 12.45 17.61 14.96 15.49 1987 1988 1989 1990 21.84 15.10 11.51 19.84. 18.22· 19.85 .15.00 20.06 14.92 17.43 11.58 9.42 14.53 13.'10 16.24 12.34 16.45 14.69 18.12 13.45 9.60 11. 71 16.23 13.55 15.71 16.76 13.57 18.75 15.71 17.87 19.60 16.10 20.88 13.70 16.41 18.28 17.37 13.98 14.30 17 .. 48 c ~. \ \ �Table 2. 1990. Monthly and annual precipitation (in.) at 9 U. Springfield Lamar Rocky Ford Limon Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec 1.32 1.38 0.65 1.10 1.93 0.54 5.58 1.57 2.67 0.44 1.16 0.41 1.00 1.87 0.73 0.92 2.11 2.47 2.30 0.64 2.12 0.23 0.80 0.52 0.80 0.79 0.96 0.28 3.63 0.38 4.91 1.22 2.37 1.07 1.07 0.39 1.04 0.42 1.50 2.19 1.91 0.90 4.34 3.00 2.15 1.14 0.96 0.05 1990 1989 1988 18.75 18.12 17.43 15.71 13.45 11.58 17.87 9.60 9.42 Longterm .! 16.65 15.40 12.50 Month s. Weather Burlington Bureau locations in Eastern Colorado, Akron Greeley 1.01 0.51 1.74 0.77 3.88 1.18 2.53 1.54 1.18 0.81 0.95 0.00 0.75 0.17 1.70 1.50 4.09 0.93 4.71 4.41 0.70 1.06 0.78 0.08 0.96 0.36 4.13 0.94 1.38 0.21 1.14 1".46 1.37 0.45 1.04 0.26 19.60 11.71 14.53 16.10 16.23 13.10 20.88 13.55 16.24 15.09 17.41 17.34 Holyoke Sterling 1.38 0.04 1.11 1.61 2.94 0.36 4.21 1.42 0.86 0.96 1.39 0.13 0.09 0.05 2.40 1.55 1.86 0.92 4.69 3.47 0.99 1.04 1.13 0.09 13.70 15.71 12.34 16.41 16.76 16.45 18.28 13.57 14.69 12.45 17.61 14.96 Table 3. Average visual obstruction readings (dm) within Conservation Reserve fields sampled during pre-greenup and nesting season intervals, eastern Colorado 1988-90. Pre-greenup Nesting Fields Within the NEC Sample 1988 1989 1990 Fields Within the cnaw 1989 1990 as even of 42 cnaw 0.39 0.34 0.39 0.20 0.56 0.67 0.51 0.66 1.13 1.09 Cost Sharea fields were included within the NEC sample. �Evaluation of Conservation Reserve Vegetation Quality Pre-greenup sampling in March 1990 yielded an average VOR of 0.39 dm; similar to early spring indices of previous years (Table 3). Among NEC samples, highest 1990 indices were obtained within the eastern tier of strata (Table 4, Fig. 1). Within stratum 4, 3 CDOW cost-share fields, which contained good cover, increased the average there. In spring 1990, fewer fields (63 vs. 89) were ranked as very poor «0.26 dm) compared to 1989 (Table 5). However, poor to very poor VOR categories «0.51 dm) still dominated (72%) the total sample. By summer (late Jun - early Jul) , VOR indices had increased to 0.67 dm among 105 NEC fields and 1.09 dm among 42 CDOW fields (Table 4), and were above indices from previous years (Table 3). Among all fields, 58 were rated very poor to poor «0.51 dm), 45 were fair (0.51-1.0 dm), and 37 were good to excellent in summer 1990 with respect to quality for pheasant nesting (Table 5). However, about 56% of the fields were in locations where pheasants were either absent or existed only in remnant numbers. Summer VOR's markedly improved over previous years. Summer VOR indices averaged higher in the eastern 3 strata as they had during pre-greenup sampling (Table 4, Fig. 1). VOR indices within Division cost-shared fields averaged >60% higher than fields that were not cost shared (Table 4). Cost-shared fields averaged 0.66 dm in early spring, or fair for pheasant nesting, and 1.09 dm in summer, or good for pheasant nesting. This was partially because most CDOW fields.were in the eastern tier of strata where growing condd t Lons were usually better. Many of the cost-shared fields also contained switchgrass and other grasses that possessed better growth characteristics. Several CDOW fields continued to possess poor to very poor cover through summer 1990. The pre-greenup VOR of cost-shared fields increased from 1989 to 1990 (Table 3). However, there was no marked difference between 1989 and 1990 samples during the summer nesting season. Paired comparison of cost-shared fields with nearby NEC sample fields indicated the cost-shared fields contained higher VOR indices '(f> 0.025, ~ - 2.77, df - 39). Conservation Reserve fields dominated by switchgrass continued, as in 1989, to possess' the highest VOR indices during early summer (Table 6). Fields containing alfalfa also possessed above average VOR indices, although the sample size was low and most possessed poor stands of grass (herbicides could not be used in establishment). Fields with no significant grass varied draJIlat.ically in cover quality from tall weeds to sparse stands of annual gxaases;: thu~:,the rang~:of VOR indices was also wide~ .However,' on average these fields (45) provided above average cover ·quality (Table ,6).' Sandour (Cenchrus spp.), an annual grass ,that was prevalent-in previous years, decreased markedly by 1990. Several fields of smooth brome (Bromus'inermis) were reseeded in spring 1989 (after previously being impacted by Glean herbicide) and the others were younger than average yielding more vigorous growth. In addition, favorable late winter and early spring precipitation stimulated growth of this cool season species, which was primarily planted in stratum 1. Several stands, classed as warm-season mixes, contained considerable sideoats grama and several were in areas marginal for farming in eastern Pueblo and Crowley counties. These factors reduced their average VOR �WI!! L D s ... • LAS ANIMAS ~.~.- . • .1_ -: .•~.L. :__ -,-, .. 1 Stratification Fig. 1. eastern Colorado. .... for random sampling ~ _ conservation Reserve fields in �jUU Visual obstruction readings (dm) within Conservation Reserve fields Table 4. in eastern Colorado during early spring and mid-summer, 1990. - N Stratum fields S12ring 1990 Sum Mean Summer 1990 Sum Mean INITIAL NEC RANDOM SAMPLE 1 2 3 4 5 6 Total 17 11 24 14 278 128 105a 7.192 6.678 9.841 4.091 10.501 1.903 40.206 0.423 0.607 0.410 0.292 0.404 0.190 12.817 8.766 14.795 7.855 18;504 7.329 0.754 0.797 0.616 0.561 0.685 0.611 0.394 70.066 0.667 CDOY COST-SHARED FIELDS ADDED IN 1989 35 24.626 0.704 40.671 l.162 25.269 18.644 28.122 13.093 23.017 8.592 0.972 0.888 1.004 0.727 0.719 0.573 116.737 0.834 65.176 45.661 0.664 1.087 COMBINED NEC AND COST-SHARED SAMPLES 1 2 3 4 5 6 Total 26 21 28 18 328 15a 13.850 12.993 12.754 9.004 13.052 3.179 0.5_33 0.619 0.456 0.500 0.421 0.245 1408 64.832 0.473 COMPARISON OF NON-COST SHARED AND COST SHARED FIELDS NEC CDOY·· 98a 42b .. 37.205 27.627 0.392 0.658 aOne field in stratum 5 and 2 fields within stratum 6 were omitted during spring 1990 samples reducing the NEC total to 102 and the combined total to 137. bInc1udes 7 CDOY fields sampled within the total 105 field NEG sample. �Ranking of Conservation Reserve fields for nesting pheasants by VOR Table 5. classification (Kirsch-dm) during pre-greenup and nesting intervals, eastern Colorado, 1988-90. N Sample <0.26 0.26-0.50 fields 0.51-1.0 1.1-2.0 >2.0 Totals Pre-greenup NEC-1988 NEC-1989 NEC-1990 All 89 fields All 90 fields 76 69 50 89 63 10 18 27 23 35 7 10 17 13 22 5 5 6 11 12 6 2 2 3 5 104 104 102 139 137 87 42 28 46 33 8 26 22 34 25 4 20 35 27 45 3 12 17 20 29 2 5 3 12 8 104 105 105 140 140 Nesting Season NEC-1988 NEC-1989 NEC-1990 All 89 fields All 90 fields Table 6. Visual obstruction readings (dm) in relation to dominant grass species within Conservation Reserve fields in summer 1990, eastern Colorado. Dominant species/cover SwitchgrassAlfalfa-grass inixes· Sweet clover-grass mixes No significant perennial grass Smooth br ome" Warm-season grass mixes Sideoats grama dominantb Wheatgrasses Fields 8 5 5 45 8 11 2826 ~ VOR 2.14 1. 70 1.41 0.95 0.80 0.77 0.69 0.52 aMost fields of smooth brome were planted more recently than others. bMany stands contained considerable blue grama and western wheatgrass. �302 quality. Stands dominated by sideoats grama, primarily in southeastern Colorado, often contained considerable blue grama (Bouteloua gracilis) and western wheatgrass (Agropyron smithii). Stand establishment there was above average because of favorable precipitation (Table 1) and earlier sign-up in ~he CRP resulting in earlier seeding. Wheatgrasses (Agropyron spp.) dominated in stratum 6, and several fields contained dense wheatgrass monocultures, usually with low VOR indices. Percent canopy cover of pere~nial grass increased from 1989 to 1990 during pre-greenup and early summer nesting intervals (Table 7). Percent annual vegetation remained nearly the same, ranging between 25 and 30%. Total canopy cover increased, in part because CRP fields contained more perennial vegetation. However, favorable moisture conditions and anticipated increases with time were also factors. Height remained nearly the same although increases from pre-greenup to summer 1990 were noted. Favorable spring moisture for growth and decreased mowing may have been factors. Both total canopy cover and height tended to be greater in the eastern tier of stra~a (Table 8). Among the 105 NEC fields, 59% contained >25% perennial grass canopy cover in summer 1990 and only 18% contai~ed ~5% (Table 9). Among the additional 35 CDOW fields, 37% contained >25% perennial grass. Winter cover on CRP fields, rated for its over-winter 1989-90 value for pheasants, was predominately poor quality. Among the 140 field sample, over 74% were rated poor, 13% were fair, and <9% were good to excellent. Anticipated quality during winter 1990-91 was estimated during summer sampling and showed similar results. However, some fields containing tall annuals, were mowed subsequent to sampling. Total NEC and additional cost-shared fields were compared with those within pheasant range (Table 10). Winter cover conditions during 1989-90 and 1990-91 were rated little better within pheasant range than within total samples. A higher proportion of the CRP fields within pheasant range contained food within 0.25 mile and were associated with good wintering sites «2 miles) (Table 10). A higher percentage of the fields within pheasant range also contained or were associated with pheasant brood habitat. Table 7. Canopy cover (%) of total vegetation, perennial grass, and annual vegetation, and mean vegetation height (em) within Conservation Reserve fields during pre-g~eenu~ and nesting intervals in 1989 and 1990, eastern Colorado. Pre-greenup Nesting Variable 1989 1990 1989 1990 Perennial grass Annual vegetation Total canopy cover Vegetation height 8.7 28.6 37.3 17.7 20.8 31.5 52.3 16.4 14.5 28.1 42.6 16.3 27.8 26.4 54.2 19.7 �Table 8. Canopy cover (%) of total vegetation, perennial grasses, annual vegetation, and mean vegetation height (cm) within Conservation Reserve fields among strata during pre-greenup and nesting season intervals, eastern Colorado 1990. fields 26 21 28 18 31 13 63.0 57.2 50.7 50.6 45.4 45.2 22.6 17.8 19.3 24.6 19.1 24.6 40.5 39.5 31.4 26.1 26.3 20.6 18.3 17.3 17.6 17.3 13.6 13.7 137 52.3 20.8 31.5 16.4 61.1 53.5 57.3 . 59.8 50.4 39.0 34.4 24.2 25.8 33.5 25.8 22.3 26.7 29.4 31.5 26.3 24.5 16.7 25.0 22.2 16.9 16.7 17.7 20.4 54.2 27.8 26.4 19.7 within NEC and CDOW N Stratum CanoI!~ cover Perennial Annual vegetation grass Total vege.tation Height Pre-greenuI! 1 2 3 4 5 6 Total/mean _-. --. .. Nesting 1 2 3 4 5 6 Total/mean 26 21 213 18 32 15 140 Table 9. Ranking of perennial grass stand establishment fields in eastern Colorado, summer 1990. Percent Variable <5.1 5.1-25 CanoI!~ Cover 25.1-50 >50 NEC SamI!1e 19 18.1 Additional CDOW Cost-share 5 14.3 24 22.9 40.0 42 17 48.6 17.1 20 19.0 SamI!le 6 7 20.0 �Table 10. Conditions on, and associated with, CRP fields rated for pheasants, eastern Colorado, 1990. Variable Sample size, fields Total NEC 105 Percent of CRP Fields within NEC within Additional CDOW within pheasant CDOW pheasant range cost share range 76 35 33 Winter Cover Conditions 1989-90 cover on CRP Poor Fair Good Excellent 80.0 15.2 2.9 1.9 77 .6 15.8 3.9 2.6 57.1 22.9 17.1 2.9 54.5 24.2 18.2 3.0 1990-91 cover on CRP Poor Fair Good Excellent 76.2 20.0 3.8 0.0 72.4 25.0 2.6 0.0 54.3 28.6 14.3 2.9 54.5 27.3 15.2 3.0 Food <0.25 mL 58.1 72.4 80.0 84.8 Good wintering area within 2 mi. 46.7 63.2 54.3 54.4 46.7 78.1 53.9 88.2 91.4 97.1 87.9 97.0 Summer Brood Conditions Brood cover on CRP Brood cover <0.5 mi. Wheat Stubble, Green Wheat and CRP Comparisons Drought conditions during early spring 1989 severely stunted the growth of green wheat throughout eastern Colorado so that residual stubble, left standing over winter (primarily in northeastern Colorado), was also of poor quality. The average VOR of standing wheat stubble in spring 1990 was 0.30 dm; much less than the 0.47 dm average for CRP vegetation during the pregreenup interval (Table 11). In 1988 wheat stubble VOR was 0.40 versus 0.39 dm for CRP. In 1989 wheat stubble VOR averaged 0.62 dm whereas that of CRP was 0.36. The VOR was only 0.13 dm for wheat stubble that was undercut or mulched with a sweep plow (similar to that of previous years). �Table 11. Respective comparisons (VOR) of wheat stubble and green wheat with CRP among strata during pre-greenup and summer intervals in 1990, eastern Colorado. Stratum 1 1 3 & 4 5 6 Average Pre-greenuJ;! Wheat stubble CRP Summer nesting Green wheat CRP 0.39 0.30 0.21 0.22 ·0.28 0.53 0.62 0.47 0.42 0.25 6.63 5.16 4.64 5.34 5.64 0.97 0.79 0.90 0.72 0.57 0.30 0.47 5.40 0.83 Wheat stubble fields were not secure nesting cover because nearly all were tilled between late April and early June. In 1990, 28% of the stubble had been tilled by 4 May, 52% by 15 May, 79% by 25 May and 94% by 8 June. This progression is about average for northeastern Colorado where the survey was taken (Snyder 1984). The VOR index of green wheat was several times greater than CRP during the early summer pre-harvest nesting interval (Table 11). However, VOR indices greater than 2 - 3 dm probably were of little additional value to nesting pheasants and other wildlife .. Taller covers usually were more open near the ground, and therefore, less attractive to ground-nesting avifauna. Green wheat often grows rapidly to become availab~e for nesting in early May. During spring 1990 in northeastern Colorado, the VOR of green wheat was 0.12 dm on 9 April, 1.47 dm on 28 April, 2.96 dm on 15 May and 5.08 dm on 25 May. Wheat growth was slightly earlier than average during 1990 and increased rapidly (Snyder 1984, 1990). Because of extremely hot, dry weather during June 1990, wheat harves.t started in late June, whereas the average harvest initiation date in northeastern Colorado is about 7 July. Thus, pheasant nests placed in green wheat had an increased probability of being destroyed or covered by straw and abandoned during wheat harvest in 1990 (Snyder 1984). Percent CRP and Pheasant Density Attempts to correlate pheasant crowing indices with the percentage CRP along pheasant crowing routes were initiated in 1989 (Snyder 1990). However, in spring 1990 data were. inadequate (3 routes completed) in northeastern Colorado preventing detection of any relationship, if it existed. Crowing count census was more complete in southeastern Colorado, but again, no relationship with CRP was evident (Table 12). Crowing indices and pheasant numbers remained low thr~ughout eastern Colorado in 1990 and there was no observed increase either in areas where CRP was common or sparse. Reproduction was poor in 1990 possibly because of June weather conditions, the early wheat harvest, or other unknown factors. �Table 12. Ring-necked pheasant crowing count indices (!calls/station) and the percentage of Conservation Reserve along established census routes, eastern Colorado 1986-90. Census route Stratun County Anilerst-Paoli Julesburg-Amherst Julesburg-Crook Holyoke-Fleming Wages-Haxtun Fleming-Leroy Eckley-Yuna Lonestar-Akron Sterling-Proctor Ft. Morgan-Narrows Platner-Elba Idal ia-Joes Bomy Reservoir Burlington North Smokey Hill Lamar-Bristol Las Animas Two Buttes Konatz-Stonington 1 1 1 1 1 1 1 1 1 6 2 2 2 2 2 3 Phi IIips Sedgwick Sedgwick-Logan Phill ips-Logan Phill ips-Yuna Logan Yuna Washington Logan Morgan Washington Yuna Yuna Kit Carson Kit Carson Prowers Bent Baca Baca 4 &5 3 3 % CR 1986 2.4 o <1.0 3.3 5.3 7.7 3.3 16.3 <1.0 2.7 3.3 2.5 0.9 3.2 0.5 5.4 1.7 33.3 12.0 8.8 4.1 6.2 1.3 11.5 16.0 7.6 9.3 1987 Crowing count index 1988 1989 1990 12.5 15.4 18.3 17.7 16.8 18.3 14.1 13.4 10.4 8.1 16.4 5.8 12.7 9.8 7.1 10.2 3.0 5.6 5.5 11.9 3.0 16.8 12.6 23.1 13.4 19.1 7.0 16.7 9.0 23.8 10.4 14.7 8.7 2.5 10.5 6.3 12.1 10.6 4.0 31.3 39.0 9.2 7.8 15.7 14.5 8.7 7.2 6.1 2.7 6.3 2.9 2.9 21.0 19.2 8.1 12.6 14.7 2.0 10.6 2.7 20.9 9.6 ·Percent CRP was determined using a 1 or 2 mi.-wide strip bisected by the census route. LITERATURE CITED Snyder,_ W. D. 1984. Ring-necked pheasant nesting ecology and wheat farming on the High Plains. J. Wi1d1. Manage. 48:878-888 _________1989. Evaluation of habitat quality on conservation reserve lands in eastern Colorado. Colorado Div. Wi1d1., Wi1d1. Res. Rep., Fed. Aid Proj. W-1S2-R. Apr. 221-244. _________1990. Evaluation of habitat quality on conservation reserve lands in eastern Colorado. Colorado Div. Wildl., Wildl. Res. Rep., Fed. Aid Proj. W-1S2-R. Apr. 205-223. Prepared by -;1tJdJp,) ¥ Warren D. Snyder �JOB PROGRESS REPORT Colorado State of: Project: W-152-R Upland Bird Research Work Plan: 21 Job Title: Evaluation of Wildlife Responses to the Conservation Reserve Program Period Covered: Author: Job _6_ 01 January through 31 December 1990 Thomas E. Remington Personnel: Clait E. Braun, Thomas E. Remington, Warren D. Snyder, David A. Wilson, Colorado Division of Wildlife ABSTRACT Breeding birds were counted along line transects within 27 Conservation Reserve Program (CRP) fields throughout northeastern Colorado. Densities of breeding birds were computed from counts using a Fourier series estimator. Horned larks (Eremophila alpestris) occurred in 19 (of 27) fields sampled at an average density of 6.9 ± 1.3 birds/lO ha. Lark buntings (Caiamospiza melanocorys) were present in 24 fields.at an average density of 6.1 ± 7.9 birds/lO ha. Meadowlarks (Sturnella neglecta) occupied all 27 fields (density 1.9 ± 3.2 birds/lO ha). Mourning doves (Zenaida macroura) were common but many probably went undetected because of their inconspicuous breeding behavior. They were found in just over half of the fields sampled at an average density of 3.0 ± 8.3 birds/lO ha. Grasshopper sparrows (Ammodramus savannarum) were common and abundant (25 of 27 fields; density 4.0 ± 8.7 birds/lO ha), probably because of maturing cover in fields. Breeding birds found in a few fields included ring-necked pheasant (Phasianus colchicus), Cassin's sparrow (Aimophi1a cassinii), dickcissel (Spiza americana), upland sandpiper (Bartramia longicauda), red-winged blackbird (Agelaius phoeniceus) and lark sparro:w (Chondestes grammacus). Dickcisse1s and red-winged blackbirds bred and pested wlthin fields only if tall; ,structurally ~obust plants such as sweet cLover-or alfalfa were present;. ,Other birds observed' were migrating (vesper sparrow [Pooecetes.. gramineus], white-crowned sparrow [Zonotrichia leucophrys], song sparrow [Melospiza melodia]), or nested in adjacent rangeland or shelterbelts (loggerhead shrike [Lanius ludovicianusJ, eastern and western kingbird [Tyrannus tyrannus and T. verticalus]). Rope drags were used to locate 194 nests of 8 species. Common nesting species were lark bunting (106 nests), mourning dove (60), meadowlark (9), grasshopper sparrow (6), horned lark (5), and red-winged blackbird (4). Some species were under-represented because peak of nesting occurre~ before we searched (horned larks and meadowlarks), or because of their inconspicuous response to the rope drags (grasshopper sparrows). ��Evaluation of Wildlife Responses to the Conservation Reserve Program Thomas E. Remington P. N. OBJECTIVES 1. Measure and compare avian use of CRP to use of alternative .(wheat and summer fallow). 2. Measure and compare nesting wheat and summer fallow. 3. Relate patterns characteristics land uses success of birds on CRP to success of avian use and nest success and surrounding land uses. on in CRP fields to cover SEGMENT OBJECTIVES 1. Measure and compare densities of birds breeding in green wheat, wheat stubble, and Conservation Reserve Program fields. 2. Measure and compare nesting success of birds nesting wheat stubble, and CRP fields; 3. Relate patterns characteristics 4. Capture and attach transmitters to 40 pheasant hens; periodi~ally relocate to determine extent, timing and type of use of CRP fields. 5. Compile in green wheat, of avian use and nest .success in CRP fields to cover and surrounding land uses. and analyze data, and prepare progress reports. METHODS A subset of CRP fields randomly selected by the U.S. Fish and Wildlife Service (from the universe of all CRP fields in eastern Colorado) was randomly chosen for study. A sample of fields in which the Colorado Division of Wildlife cost-shared on a wildlife grass seed mixture was added for comparison. Pe rmi.s.sLonto e?ter and measure bird abundance within these fields was obtained via letter, and usually by· telephone, from landowners. CRP, and some green wheat and wheat stubble· fields were sampled within an hour of sunrise to count breeding birds. A 0.8-km long transect was established .near the midpoint of each field with plastic flags. An observer walked along the transect line slowly watching and listening for birds. Birds observed were identified to species and. the perpend{cular distance from each bird to the line transect was measured to the nearest meter using a rolotape. Each field was censused once between 20 April and 20 May and again between 21 May and 20 June to ensure both early (horned larks, meadowlarks) and late (lark buntings, grasshopper sparrows) breeding species were counted. Breeding bird density was estimated from perpendicular distance measures (from the highe;r of the 2 counts) using a Fourier series estimator (Burnham et al. 1980, 1981). �Eight hectares within CRP fields were searched for nesting birds by dragging a 30-meter long rope between 2 ~bservers. The area around the point a bird flushed was searched intensively for a nest. Each field was searched 2-3 times throughout the breeding season in an attempt to document nesting by all birds using the field. We attempted to revisit each nest at 7-day intervals to evaluate nest success; this was not possible for all nests because of time constraints. RESULTS AND DISCUSSION Fifteen species of birds were observed along line transects within CRP fields, but only 6 species were common resident breeders (Table 1). Horned larks were the most abundant breeding bird while meadowlarks were the most ubiquitous, -i.e., they occurred in the most fields (27 of 27). Lark buntings were second in abundance and ubiquity (number of fields in which present). The relative importance of these sp~cies was indicative of the generally poor condition of the grass stand and cover within these fields. Grasshopper sparrows were relatively common, especially in fields with better cover. Cassin's sparrows were commonly observed in 1989 in fields that were planted to native bunchgrasses. We observed only 5 in 3 fields in 1990, primarily because fields -in northeast Colorado were planted to smooth brome or because cover had yet to develop in fields planted to bunch grasses. Mourning doves were fairly common in fields with low, sparse cover, but were probably under represented because of their tendency to display on telephone wires or from elevated perches near field edges. Dickcissels were observed while singing in only 1 field where sweet clover was abundant. Vesper sparrows were frequently observed in late April and early May, but were not observed later and were presumed to be using CRP fields while migrating. Occasional use of CRP fields by species that nested elsewhere (in trees or shrubs) was noted. These species included lark sparrow), northern oriole (Icterus ga1bula), American robin (Turdus migratorius), loggerhead shrike, and eastern and western kingbirds. Nests and/or broods of 8 species were found during searches of 29 fields. These were lark bunting (106), mourning dove (60), horned lark (5), meadowlark (9), pheasant (1), grasshopper sparrow (6), red-winged blackbird (4), and lark sparrows (3). Cassin's sparrow nests were detected in 1989, but not in 1990. Nest success estimates have not yet been calculated. The number of nests found for each species was not necessarily indicative of breeding density or effort. Horned larks and meadowlarks were probably .under represented because most nest searching was conducted after the peak of nesting for these species. The peak of nesting effort by horned larks in 1989 was late May. In 1990, despite intensive searching beginning in late April, we found only 5 horned lark nests, 4 of these were apparently second nest attempts found on 14 June. Either 1989 was a relatively late year or 1990 was a relatively early year for horned lark nesting. Grasshopper sparrow n~sts were difficult to find because birds either ran off nests prior to flushing or didn't flush in response to the rope. Most of the 6 grasshopper sparrow nests were found when a bird flushed off the nest because we walked within 1 meter. Only one pheasant nest was located in 1990. While the quality of cover for nesting within these fields was generally poor, it was also pos~ible that pheasant hens did not flush in response to the rope. Nest densities of mourning doves were more indicative of dove use of CRP fields than breeding density estimates from the line transects. �Table 1. Number observed, breeding density (birds/10 ha), and percentage of sampled Conservation fields in which birds were observed, 1990. Species Lark bunting Grasshopper sparrow Horned lark Yes tern Meadowlark Mourning dove Red-winged blackbird Vesper sparrow Cassin's sparrow Ring-necked pheasant Yestern kingbird Dickcissel Song sparrow Upland sandpiper White-crowned sparrow Eastern kingbird N Observed Frequency 226 153 140 108 35 24 8 5 4 4 4 4 3 1 1 89 93 70 100 56 41 22 11 15 11 4 7 11 4 4 % Density ! so 6.1 4.0 6.9 1.9 3.0 Conments 7.9 8.7 1.3 3.2 8.3 Breeding/migration Seen during migration Bunchgrass fields 1 field only Migrating Migrating The birds.breeding in CRP fields, at this point in the program, primarily represent habitat generalists rather than grassland or grassland/shrub species. Relatively few line transects or nest searches were conducted in green wheat or wheat stubble in 1990 to increase sample sizes in CRP f~elds. However, it was apparent that these habitats were used for breeding by lark buntings, horned larks, mourning doves, pheasants, and perhaps grasshopper sparrows .. These are also the species commonly using CRP fields. Thus, the benefits of this program to breeding birds are unclear. Woody plantings within CRP fields shou14 enhance diversity of breeding birds. Lark sparrows, loggerhead shrikes, American robins, mourning doves, and red-winged blackbirds would be expected to use CRP (or use it more extensively) for breeding if woody plantings were present. Attempts to capture and radio-mark pheasant hens to quantify pheasant use of CRP fields, versus alternative cover types, for nesting and brood rearing were not successful., Additional trapping techniques will be attempted in winter 1990-91. LITERATURE CITED Burnham, K. P., D. R. Anderson, and J. L. Laake. 1980. Estimation of . de·risityfrom line transect sampling of biological populations; Wildl. Monogr. 72. 202 pp. ___ , , and 1981. Li.ne transect estimation of bird population density using a fourier series. Studies Avian BioI. 6:466-482. Thomas E. Remington Wildlife Researcher B ��313 INTERIM FINAL REPORT State of: Colorado Project: W-152-R Work Plan: 21 Job Title: Author: : Job _7_ Avifauna Period Covered: Upland Bird Research Responses to Grazing 01 January through 31 December 1990 Cynthia P. Melcher Personnel: C. E. Braun, K. M. Giesen, T. E. Remington, Colorado Wildlife; C. P. Melcher, Colorado State University Division of ABSTRACT The hypothesis that high densities of elk (Cervus elaphus) in Rocky Mountain National Park had contributed to observed declines in breeding densities of white-tailed ptarmigan (Lagopus leucurus) was investigated. Willows (Salix spp.) are essential as food and cover for ptarmigan, and may have been impacted in areas heavily used by elk. Relationships between ptarmigan, elk, and willow were assessed by measuring occurrence of ungulate feces, willow characteristics, and ptarmigan foraging ecology/reproductive success during 1989 and 1990. Four study areas (17 - 22 ha) were selected in alpine/ krummholz habitat near Trail Ridge Road. Numbers of breeding Ptarmigan had declined at two sites, but not at the other two sites. Density and species richness of alpine/willow passerines were measured to establish baseline information against which future trends may be compared. Ungulate feces were more abundant (f < 0.01) in areas where ptarmigan had 4eclined. Willow characteristics hypothesized as important to foraging ptarmigan (willow patch perimeter/diameter, height, density of buds, and unbrowsed terminal leaders), did not differ in 'a consistent .and predicted pattern between high and low elk density sites. However, extreme high and low values were generally observed at sites where ptarmigan densities had declined and moderate values were observed at sites where densities had remained relatively stable. Lengths of unbrowsed terminal leaders were greater (f < 0.01) at sites where frequencies of ungulate feces were greater. Differences were not detected among other willow characteristics thought to be important indicators of heavy browsing. Avian species richness was greater in sites where ptarmigan had declined. Species suspected of using sites for feeding only, but which were observed at sites regularly during censuses, contributed to species richness and density. Twelve ptarmigan hens were marked with radios. One disappeared, one fledged 4 chicks, and the remaining 10 were unsuccessful. Hens spent 48.8% of their time foraging (23.4 to 95.9%), and an average of 26% of their time sitting/resting/sleeping. Hen foraging rates averaged 32.6 bites/minute (range 24.3 to 48.0) when foraging on shrub willow. No salient patterns emerged for hens among sites, possibly due to small sample size. A thesis �314 incorporating two manuscripts for publication and two appendices will be submitted by 31 December 1991 in partial fulfillment of the requirements for a Master of Science degree in Biology at Colorado State University. Prepared by C~;;'~:ic~· 1(VLc0..C/~ Graduate Research Assistant Approved bY~(. Thomas E. Remington Wildlife Researcher �315 JOB PROGRESS REPORT Colorado State of: Project: W-152-R Upland Bird Research Work Plan: 22 Job Title: Upland Bird Research Publications Period Covered: Author: Job _l_ 01 January through 31 December 1990 Clait E. Braun Personnel: Clait E. Braun and Richard W. Hoffman, Colorado Division of Wildlife ABSTRACT The following articles were published in 1990: Braun, C. E. 1990. Effects of regulations on sage grouse harvest and population size. Annu. Mtg., Central Mtns. and Plains Sect., The Wildlife Soc. 35:Abstract. 1990. Review of "Adaptive strategies and population ecology of northern grouse". A. T. Bergerud and M. W. Gratson, eds. Univ. Minnesota Press, Minneapolis, 1988:809pp. Wilson Bull. 102:191-192. 1990. Summary remarks of "Introductions and reintroductions of wildlife populations. Trans. No. Am. Wi1dl. and Nat. Resour. Conf. 55:633-634. Cade, B. S., and R. W. Hoffman. 1990. Winter Use of Douglas-fir forests by blue grouse in Colorado. J..Wildl. Manage. 54:471-479. Hoag, A. W., and C. E. Braun. 1990. Status and distribution of plains sharptailed grouse in Colorado. Prairie Nat. 22:97-102. Hoffman, R. W. 1990. Chronology of gobbling and nesting activities of Merriam's wild turkeys. Proc. Nat1. Wild Turkey Symp. 6:25-31. Schmutz, J. A., C. E. Braun, and W. F. Ande1t. 1990. Brood habitat use of Rio Grande wild turkeys. Prairie Nat. 22:177-184. VanSant, B. F., and C. E. Braun. 1990. Distribution and status of greater prairie-chickens in Co1orad~. Prairie Nat. 22:225-230. White, J. A., and C. E. Braun. 1990. Growth of young band-tailed pigeons in captivity. Southwest. Nat. 35:82-84. Prepared by _~&_V_<_~____.. ~u::::.._,--....:..;:_ __ C1ait E. Braun ��':)J.I JOB PROGRESS REPORT State of: Project: Colorado W-152-R Upland Work Plan: 25 Job Title: Evaluation Fannin Period Covered: Bird Research Job _1_ of Wildlife 01 January Responses through 31 December to Pesticides Used in Wheat 1990 Author: Thomas E. Remington Personnel: Clait E. Braun, Carol A. Mehaffy, Michael W. Miller, Thomas E. Remington, Joan D. Ritchie, Lyn Stevens, David A. Wilson, Colorado Division of Wildlife; James Echols, Wendy Meyer, Frank Peairs, Stan Pilcher, Colorado State University; John Pearson, Claude Ross, Ted Warfield, FMC Corporation. ABSTRACT The impacts of two methods using pesticides to control insect pests in winter wheat on bird communities were investigated. These were aerial spraying of Lorsban (chlorpyrifos)'or Di-Syston (disulfoton) to control Russian wheat aphids (Diuraphis noxia) in spring and microtubule injection of Furadan (carbofuran) at planting to control grasshoppers. No dead or impaired birds were found during a search of 16 ha within a wheat field sprayed with Lorsban. Searchers located 5 of 8 carcasses placed in fields to evaluate search efficiency. Spraying of Di-Syston or Lorsban had no effect (f < 0.05) on 24 or 48-hour survival of pheasant chicks in a controlled experiment. Spraying of Di-Syston or Lorsban to control Russian wheat aphids in spring does not appear to impact birds significantly. Two horned larks (Eremophila alpestris) were found dead during searches of 27 km around the perimeters of 8 fields·710 and 14-21 days after seed planting and microtube injection of Furadan. One horned lark appeared to have been hit by a car. Carbamate resi~ues in grasshoppers averaged 1.31 mg /kg , a level unlikely t e .cause av i an mo rt aLL ty. Canada geese (Branta canadensis) largely rej ected 'Furad.im-treated' wheat in a controlled experiment. About 34 g of wheat remained in 0.785 m2 plots in the treatment pen vs. 2 grams in the untreated pen. Brain cholinesterase activity of Canada geese exposed to Furadan-treated wheat was unaffected (f> 0.05), although plasma cholinesterase activity was depressed about 38%. ��EVALUATION OF WILDLIFE RESPONSES TO PESTICIDES USED IN WHEAT FARMING Thomas E. Remington INTRODUCTION Wheat is the predominant cultivated crop in Colorado with upwards of 3 million acres·planted each year. Wheat fields are used extensively for foraging. and or breeding by a diverse assemblage of wildlife including ring-necked pheasant (Phasianus colchicus), horned larks, lark buntings (Calamospiza melanocoIYs), mourning doves (Zenaida macroura), Canada geese, and pronghorn (Antilocapra americana). Russian wheat aphids have caused extensive damage to wheat in Colorado since arriving in 1985. Consequently, massive aerial spraying has been conducted to control aphids; as much as 45% of planted wheat has been treated annually. Aerial applications of Di-Syston or Lorsban (chlorpyrifos) have emerged as the predominant treatment methods. Both are highly toxic organophosphate pesticides with the potential for wildlife mortality (Smith 1987). Microtubule injection of Furadan is a relatively new technique for controlling grasshopper depredation on immature winter wheat in fall. This technique has also been shown to be.effective in controlling aphids in wheat (Guerra-Sobrevilla 1988). Microtubule injection has several advantages (from a wildlife perspective) compared to more conventional methods of pestici~e application .. The pesticide is applied only as a narrow border treatment (typically 12-14 rows) and is injected as.a liquid 5 cm deep in the soil around the seed; consequently wildlife exposu~e is minimized. Furadan is a systemic insecticide which is highly toxic to wildlife (Flickinger et al. 1980, James 1987, Smith 1987, Littrell 1988). Previous bioassays of immature winter wheat. treated by microtubule injection of Furadan have shown significant insecticidal activity into April. While this has positive benefits for control of grasshoppers and/or aphids, it suggests that Furadan residues or metabolites persist overwinter and indicates a potential for significant wildlife exposure. Wildlife potentially at risk of significant exposure to this pesticide would be those that feed directly on exposed seeds or immature wheat,. or on grasshoppers (or other insects) that feed on wheat. Horned larks and meadowlarks (Sturnella magna) feed on exposed wheat seeds. Pronghorn and Canada geese fe,ed on winter wheat extensively from late'fall until early' spring. Several insectivorous birds, primarily eastern and western kingbirds (Tyrannus tyrannus and I. verticalis), American kestrels (Falco sparverius), burrowing owls (Speotyto cunicularia), and loggerhead shrikes (Lanius ludovicianus) ,.could be expected to feed on contaminated grasshoppers and may be at some risk. Risks may be mitigated for some species by timing of fall migrations. However Swainson's hawks (Buteo swainsoni) assemble in large premigratory and migratory aggregations within farmland and rangeland of eastern Colorado and feed extensively, if not exclusively, on grasshoppers (Bailey and Niedrach 1965, Woffinden 1986, Johnson et al. 1987). The purpose of this study was to evaluate the impacts on avian wildlife of aerial spraying to control aphids and microtubule injection of Furadan to �320 control grasshoppers or aphids. Emphasis was on ascertaining direct mortality but indirect effects such as cholinesterase depression were also investigated. P. N. OBJECTIVES 1. Determine acute mortality of wildlife 1-3 days post-spray resulting from prescribed levels of Di-Syston and chlorpyrifos. 2. Recover carcasses of wildlife suspected of dying from pesticide poisoning for analysis of cause of death and pesticide residue levels. 3.' Monitor impacts of spraying prescribed levels of Di-Syston or ch1orpyrifos on avian nesting success. 4. Determine level of brain cholinesterase depression in songbirds and pheasants 24-48 hours post-spray as an index to both pesticide exposure and potential effects. 5. Measure toxicity of prescribed levels of Di-Syston and chlorpyrifos to 7-10 day-old pheasant chicks. SEGMENT OBJECTIVES 1. Measure acetylcholinesterase activity in brain tissue of carcasses found dead in, or collected from, wheat fields during fall 1999. 2. Conduct pesticide residue analysis of tissues from birds found dead in wheat fields and exhibiting ~ sox brain acetylcholinesterase inhibition. 3. Search wheat fields for dead or moribund wildlife following application of Di-Syston, chlorpyrifos, and possibly carbofuran. 4. Determine rates of carcass removal by scavengers, and carcass detection by search crews. 5. Compile and analyze data, and prepare progress reports. METHODS Wheae Aphid Study Two research approaches were used during this period. Transects were conducted to search for dead or impaired wildlife in sprayed fields. Acute toxicity and indirect effects of these pesticides on survival of pheasant chicks was evaluated by experiment. Specific methodology for each approach follows. Transects were conducted by 4-5 people walking abreast, about 20 rows apart 13 days post-spray until 8.1 ha had been searched per field. Most searches were conducted with the rows, although it was found that as wheat grew taller visibility of carcasses was greatest walking against the rows. Search efficiency was measured as the percent recovery of house sparrows (Passer dornesticus) placed in sprayed fields prior to the search. '_ �A randomized complete b~ock design was used in the pheasant chick experiment (Fig. 1). Twelve strips, 244 m long by 55 m wide, were marked with flagging within a 24-ha wheat field about 3 km south and 3 km east of Akron, Colorado. A 27-m buffer strip was marked between strips to prevent drift of pesticide across treatments. Treatments, Di-Sys-ton, Lorsban, or unsprayed control, were randomly assigned to 1 of 3 strips within each of 4 btocks. 'Within each strip, 2 of 16 possible 0.1 ha (1/4 acre) pen locations were randomly selected. Pens were constructed of 61 by 2.54 cm (24" by 1") mesh poultry nettin~ supported by rebar and aluminum rods. Within the center of each pen a 9.75-m circular pen was constructed of 1.22 m (48") hardware cloth. Hot boxes were constructed of 46 or 51 cm (18 or 20") diameter galvanized duct pipe, covered with fiberglass insulation and clear plastic and placed in each circular hardware-cloth pen. Hot boxes were heated to 32 C (90 F) by l25-watt heat; lamps suspended at the top and powered by 3 generators. Line voltage thermostats were used to maintain a temperature of approximately 32 C. A door was cut into the bottom of each hot box to allow chicks access. Ten, 6-day old pheasant-chicks were placed in each pen from 0645 to 0730 hours. DiSyston and Lorsban strips were sprayed on 22 May with 1.12 kg active ingredient (a.i.) per ha or 0.84 kg a.i./ha, respectively. Two Pawnee PA 25 aircraft calibrated to apply treatments in 3.1 liters of spray volume per ha (2 gallons/acre) were used. Spraying began at 0740 and was completed by 0830 hours. Conditions were ideal for spraying; sunny, warm and little or no breeze. To facilitate recovery and prevent escape, pheasant chicks were initially placed in the small pens. Chicks were placed within the hot boxes near sunset each day and during inclement weather periods. They were provided pelleted chick feed and water ad libitum. Chicks were released from the hot boxes each morning around 1000 when air temperature had warmed. Carbofuran Study The purpose of this research was to gather preliminary information on mortality or morbidity of wildlife thought to be at risk from microtubule injection of Furadan as an at-planting treatment to control grasshoppers. Effects on passerine birds were evaluated by searching borders of treated fields (Furadan 4F at a rate of 14.8 ml per 305 linear meters) for dead or impaired birds at about 1 and 2 weeks after planting. Searcher efficiency was estimated as recovery of house sparrow carcasses placed within field margins prio,r to the search._ The effect on Canada geese of consumption of furadan-treated wheat was evaluated by placing birds into pens containing furadan-treated wheat or untreated wheat (control), allowing them to eat for 48 hours, and then sacrificing them and measuring brain cholinesterase activity. Geese (29) were captured while flightless on 4 June from the Loveland Golf course and maintained at the Foothills Wildlife Research Facility until the feeding trial. Birds were fed a commercial pelle ted chow (Purina Turkey Ration) , although they supplemented this by feeding on grass and weeds within their pen. Geese were wing clipped in late July to minimize injuries within the pen and to prevent escape from the welded wire pens in the wheat field. A similar experiment in 1989 was hampered by extremely low residues of Furadan in the treated wheat to which geese were exposed, either because of poor �BLOCK 1 BLOCK 2 BLOCK 3 BLOCK 4 meter buffer 55 meter spray Fig. 1 Design used in pheasant chick experiment. uptake, metabolism, or possibly erroneous measurement of Furadan residues by the contract laboratory. To insure high residue levels, representative of a worst case scenario, we planted wheat into sub-irrigated ground and exposed the geese 25-26 days after planting rather than 98-99 days as in 1989. Agri Pro, Inc., provided an irrigated field at their research farm about 4 km southwest of Berthoud, Colorado. Winter wheat (Sandy variety) was planted on 17 September 1990, at a rate of 16.2 kg per ha and a 0.3 m (12-inch) row spacing. Furadan 4F (14.8 ml per 305 linear meters) and liquid fertilizer (7.68 kg N, 26.1 kg P20S per ha) were applied in-furrow by micro-tubule injection. Two blocks of untreated wheat were planted slmilarly, although without}'uradan Lnjec t Lon (Fig. 2). Geese were transported,to the site on 10 October and released into a 47.2 m by 14.6 m (48 row) welded wire enclosure around untreated wheat. Geese were allowed to feed within this enclosure-for 2 days to acclimate them to a wheat diet and to conditions at the site. They were they moved into 1 of 2 adjacent 47.2 m x 7.3 m. (24 row) enclosures containing either Furadan treated (15 geese) or untreated wheat (14 geese). The birds were left within the pens for two days to feed. They were then killed by CO2 asphyxiation. Heads were removed, placed on dry ice, then transported to and stored within a freezer at -60 C. Brain and plasma acetylcholinesterase activity was determined following Hill and Fleming (1982). All remaining wheat above ground level was clipped and weighed from 10, 0.785-m2 circular plots randomly located within treated and untreated enclosures to measure feeding intensity. �J.t:.J Two 100 g samples of treated and untreated wheat were collected from outside the enclosures on the first trial day, kept frozen, and sent to A&L Mid West Laboratories for analysis of Furadan and Furadan metabolite residues. Analysis was by HPLC with a ultraviolet detector (no other details available) with a detection. limit of 0.1 ppm. Because of questions about the accuracy of the residue levels found by A&L, the samples were sent to the ACG Developmental Chemistry Department laboratory of FMC Corporation in Princeton, New Jersey for additional analysis. RESULTS and DISCUSSION (f) o w IT W I- I- ~ w '\ IZ « w IT LO ~ Russian Wheat Aphid Study Mortality Transects.--Sixteen hectares within one field sprayed with Lorsban on 2 May were searched on 3 May for dead or impaired wildlife. No dead or impaired birds or other animals were detected, although 5 of 8 birds placed to evaluate searcher efficiency were located. Aphid infestations were relatively light. Treatable infestations occurred near Pietz and north into.southern Wyoming for the .first' time. o W I(f) IT W IW « W IT I- o W I- « Spraying of Di-Syston apparently resulted in ,W z the deaths of 9 jack rabbits (Lepus ~ « IT Ica1ifornicus) about 19 kilometers southwest of o '\ Z « Pritchett in Southeast Colorado. The dead LO ~ IT jackrabbits were found in a shortgrass pasture :::) LL on 10 May, 1990, 1-30 m from a wheat field which had been sprayed late the day before. No dead jackrabbits were found within the wheat field. Jackrabbits are extremely sensitive to a metabolite of Di-Syston produced when the parent ·compound is 48 ROWS metabi;>1izedby .growing vegetation. A likely scenario would be that vegetati,on in 'the pasture was contaminated with Di-Syston by Fig. 2 Field design of overspray or drift. Jackrabbits feeding on Furadan-wheat, goose study. vegetation within the pasture would be exposed to a large dose of the metabolite of Di-Syston because the spray would be concentrated on the relatively short vegetation in the pasture. Jackrabbits have been observed within wheat fields sprayed wit:h Di-Syston and were apparently unaffected. They appeared to be using maturing wheat fields for cover while foraging in nearby pastures. Jackrabbits are probably not at significant risk if spraying is confined to wheat. �Pheasant Chick Toxicity.--Spraying of Di-Syston or Lorsban did not decrease (f An average of 9.7, 9.7, and 9.9 (of 10) chicks survived for 24 hours post-spray in pens exposed to Di-Syston, Lorsban, or unsprayed, respectively. Survival to 48-hours post-spray for these groups was 9.4, 9.1, and 9.5 chicks per pen, respectively. Survival of chicks was markedly improved from the experiment conducted in 1989. This was attributed to use of the hot boxes which allowed maintainenance of chicks at or near their preferred temperature and eliminated the need to capture chicks and transport them indoors. Di-Syston significantly reduced survival of chicks in the previous experiment but had no effect in this experiment. This supports the argument (Remington 1990) that the cold temperatures associated with a thunderstorm the afternoon of the day spraying occurred resulted in disproportionately more chicks sprayed with DiSyston dying. Cold stress can make birds more susceptible to carbamate pesticides (Rattner et a1. 1982). > 0.05) 24 or 48-hour survival of pheasant chicks relative to controls. A male lark bunting was found apparently sick and unable to fly about 7 hours after spraying near strip 12 which had been sprayed with Lorsban. This bird was placed inside one of the pens and apparently recovered as it was not found later. It may have been exposed to Di-Syston or Lorsban. Two mallard (Anas p1atyrhynchos) nests were discovered in the wheat field while we were building pens. These hens were each incubating 10 eggs and both were in strips sprayed, -one with Lorsban and one with Di-Syston. Both hens continued to incubate their clutches for at least a week post-spray. Data from the mortality transects and pheasant chick experiments indicate that significant avian mortality is unlikely when labelled rates of these two pesticides are sprayed to control aphids. Pheasant chicks younger than 5-6 days of age may be more susceptible to these pesticides. However, most spraying occurs before most pheasant eggs have hatched, particularly nests in green wheat. Most sp~aying occurs early in the morning before temperatures have warmed up and winds begin to increase. Chicks younger than 5-6 days are probably brooded by hens at this time and not directly exposed to pesticides. Carbofuran Study, Mortality Transects.--Twenty-seven kilometers around the perimeters of 8 fields were searched for dead or moribund birds within 7-10 days of planting and again 2-3 weeks after planting. Two dead horned larks were found. Upon examination, 1 carcass found adjacent to a road had injuries consistent with being; struck by an automobile. The other carcass was too degraded to yield useful fnformation to cause of death.' In 1989 two horned lark carcasses were recov~red in '~e,archesof 3,8'Ian around 13'treat~d fields '. S~arch~r efficiency was not evaluated this year,'but averaged 69% (25 of 36 carcasses recovered) in 1989. Efficiency within each field was 67, 75, 63, 87, and 60%. These results indicate that had significant mortality occurred, we would have detected it, assuming carcasses remained until our searches. This assumption was not tested directly during this study. Fifty percent of house sparrow carcasses were still present after 7 days during previous research in wheat fields in eastern Colorado during spring (Remington 1990). as It appears that mortality of passerines is negligible under this application method of furadan to the borders of fields. These 'results should be '-' �interpreted cautiously. An inherent weakness in this approach is that birds could feed within treated field margins, flyaway to die elsewhere, and go undetected during our searches. This could be a problem with horned larks that have large home ranges in fall and winter. Swainson's hawks were plentiful in and around treated fields during the period we conducted mortality transects. Examination of hawk scats indicated almost complete feeding on insects, predominately grasshoppers. We did not detect any dead, moribund, or abnormally behaving Swainson's hawks. Exposure of other insectivorous birds to furadan-contaminated grasshoppers was presumably minimal since few insectivorous birds were observed during our searches. Fall migrations may mitigate potential exposure of these birds to furadan, although the extent to which this occurs may vary annually. Furadan residues of composite grasshopper samples measured by A&L laboratories were below detection limits from 1 treated field and up to 30 times levels measured in 1989 from another. Residues measured by FMC Corporation's lab were repeatable, much less variable (Table 1), and verified by spiking check samples. Thus-residues measured by FMC were used for hazard assessments. Assuming a SO g (0.05 kg) bird, a grasshopper weight of 0.89 g (0.00089 kg), and total carbamate residues of 1.31 ppm (from Table 1), ingestion of a single grasshopper represented about 5.6% of the LDSO for·red-winged blackbirds (Agelaius phoeniceus) and about 1.9% of the LDSO for house ·sparrows. Swainson's hawks may eat 100 grasshoppers per day (Johnson et al. 1987). Assuming an intake of 100 grasshoppers with an ·average ~arbamate residue of 1.31 ppm, and a bird weight of 0.908 kg, Swainson's hawks would be exposed to 3.2% of an LDSO. These hazard assessments suggest little avian risk from consuming grasshoppers killed or injured by Furadan, and support the complete lack-of insectivorous bird carcasses recovered during mortality transects. Analysis of scats found at several treated fields indicated that striped skunks (Mephitis mephitis) were feeding extensively on grasshoppers along field margins. It is difficult to assess the potential impact to skunks of this exposure since information on how many grasshoppers they consume in a day and their sensitiv·ity to furadan is not available. LDSO's of dogs and rats are 19 and 11 mg/kg, respectively (Smith 1987). Using 10 mg/kg as an LDSO, and assuming a body weight of 3 kg and consumption of 100 or 500 grasshoppers per night, skunks would be exposed to 0.39 - 1.94% of an LDSO. If these residues are accurate, and these assumptions reasonable, mortality of skunks seems unlikely, Canada Goose Eneriment.--Geese adapted·well to experimental conditions and grazed the untreated wheat heavily during the 2 days of acclimation. Almost no wheat remained in this pen after two days. Geese within the control (untreated) pen continued to graze heavily following their transfer to this pen, but feeding was minimal in the pen with Furadan-treated wheat. Avoidance of Furadan-treated wheat continued through the second day, despite the fact that there was no other food source. Wheat (above ground) within the control pen was essentially eliminated by the conclusion of the trial. When wheat samples were clipped and weighed fr~m each pen the next morning, some recovery of grazed wheat was apparent. Still, only 2;03 ± 0.86 g remained in 0.785 m2 'plots in the control pen, significantly less (l < 0.05) than the 33.95 ± 9.33 g remaining in plots within the treatment pen. �Avoidance of Furadan-treated wheat is significant, because large die-offs of geese and other waterfowl have occurred when alfalfa treated with Furadan 4F f1owab1e was (apparently) eaten (Flickinger et a1. 1980, reviewed in Smith 1987). Canada geese have been successfully deterred from feeding on rye and turf by methiocarb (another carbamate pesticide) application (Conover 1985, 1989), but methiocarb is substantially less toxic to birds than Furadan. Carbamate residues in treated wheat samples, adjusted for percent recovery, were 5.81 and 6.08 ppm. Dietary LC50 for mallards is 190 ppm (Smith 1987), over 30 times higher than levels measured in wheat. Mallard LD50's are substantially lower, in the range of 0.4-0.6 mgfkg (Smith 1987). A 2.5-kg Canada goose eating 0.35 kg of green wheat with a total carbamate level of 5.7 mgfkg would be exposed to a dose of 0.8 mgfkg body weight, or in excess of a mallard LD50. Thus, if geese had eaten large quantities of green wheat in this trial some mortality might have occurred. It is interesting that geese apparently eat enough Furadan-treated alfalfa to poison themselves yet avoid wheat containing potentially lethal levels. A possible explanation is that geese don't reject Furadan-treated vegetation on the basis of taste, but rather because of post-ingestive consequences such as feeling ill. Geese may sample alfalfa sprayed with Furadan 4F, be exposed to a lethal dose and die. Sampling wheat in this study would have exposed them to a sub-lethal dose that may have made them sick enough to avoid additional feeding. Table 1. Furadan and Furadan metabolite residue levels (ppm) in winter wheat from a Canada goose study treatment field (treated with an at-planting microtubule injection of Furadan of 14.8 m1/305 meters) and in grasshopper samples collected (dead or dying) from similarly treated fields (percent recovery of spiked samples from each matrix indicated in parentheses). Sample Wheat 102 103 Carbofuran 2.86 (94.2) 2.48 Grasshoppers 927 919 0.16 (87.5) 0.35 aNot detected. 3-Keto Carbofuran NDa (95.7) ND 0.01 (89.0) 0.03 3-Hydroxy Carbofuran 2.62 (94.2) 3.25 0.71 (85.1) 0.73 �Acetylcholinesterase (ChE) activity of brain tissue from geese in treatment and control pens did not differ (l < 0.05, Table 2). Plasma ChE was depressed (l < 0.05) an average of about 38% in birds exposed to Furadan treated wheat. Plasma ChE inhibition is a sensitive indicator of exposure to antiChE agents but is not in itself indicative of negative health consequences because the function of plasma ChE is not known (Kutty 1980). Plasma ChE may be inhibited by as much as 75% when brain ChE reaches a detectable and biologically significant 10-20% (Ludke et al. 1975). These results support the premise that initial feeding on treated wheat exposed geese to Furadan, but feeding ceased prior to significant brain ChE depression. . Table 2. Brain and plasma acetylcholinesterase activity (mean ± SD) of Canada geese from pens containing Furadan-treated (n - 15) or untreated (n - 14) winter wheat. Plasmaa Untreated Treated 672.1 ± 201.9 419.9 ± 124.3 5.57 ± 0.83 5.58 ± 1.48 ~U acetylthiocholine iodide hydrolyzed per min·per ml of plasma at 25 C. b Jlmol acetylthiocholine iodide hydrolized/g (wet weight) of brain tissue at 25 C. Wildlife risks from microtubule injection of furadan appear slight. Residue levels in green wheat are low relative to LC50 or LDSO values of mallards or geese. Pronghorn tested in 1989 showed no aversion to treated wheat nor consequences from ingesting it. Geese avoided eating treated wheat and kept intakes below levels that would reduce brain ChE. Risks to insectivorous birds are mitigated by timing of migrations and low furadan residues in grasshoppers. The primary risk appears to be to granivorous birds, principally horned larks and meadowlarks. Few unexplained mortalities were found, suggesting that losses were slight or nonexistent. LITERATURE CITED Bailey, A. M., and R. J. Niedrach. 1965. Nat. Hist. Denver, Colo. 895 pp. Birds of Colorado. Denver Mus. Conover, M. R. 1985. Alleviating nuisance Canada goose problems through methiocarb-induced aversive conditioning. J. Wildl. Manage. 49:631-636. 1989. Can goose damage to grain fields be prevented through methiocarb-induced aversive conditioning? Wildl. Soc. Bull. 17:172-175. Flickinger, E. L., K. A. King, W. F. Stout, and M. M. Mohn. 1980. Wildlife hazards from Furadan 3G applications to rice in Texas. J. Wildl. Manage. 44:190-197. �Guerra-Sobrevilla, L. 1988. Effectiveness of carbofuran applied to the soil and as a seed treatment for the control of aphids on wheat in northwestern Mexico. Crop Prot. 7:336-337. Hill, E. F., and W. J. Fleming. 1982. Anticholinesterase poisoning of birds: field monitoring and diagnosis of acute poisoning. Environ. Toxicol. Chem. 1:27-38. James, P. C., and G. A. Fox. 1987. Effects of some insecticides on productivity of burrowing owls. Blue Jay 45:65-71. Johnson, C. G., L. A. Nickerson, and M. J. Bechard. 1987. Grasshopper consumption and summer flocks of nonbreeding Swainson's hawks. Condor 89:676-678. Kutty, K. M. 1980. 13:239-243. Biological function of cholinesterase. Clin. Biochem. Littrell, E. E. 1988. Waterfowl mortality in rice fields treated carbamate, carbofuran. Calif. Fish and Game 74:226-231. with the Ludke, J. L., E. F. Hill, and M. P. Dieter. 1975. Cholinesterase (ChE) response and related mortality among birds fed ChE inhibitors. Arch. Environ. Contam. and Toxicol. 3:1-21. Rattner, B. A.., L. Sileo, and C. G. Scanes. 1982. Hormonal responses and tolerance to .cold of female quail following parathion ingestion. Pest. Biochem. Physiol. 18:132-138. Remington, T. E. 1990. Evaluation of wildlife responses to pesticides used in wheat farming. Colorado Div. Wild!. Job Prog. Rep., Fed. Aid Proj. W-152-R. Apr:253-264. Smith, G. J. 1987. Pesticide use and toxicology in relation to wildlife: organophosporous and carbamate compounds. U.S. Dep. Inter., Fish and Wildl. Servo Resour. Publ. 170. l7lpp. Woffinden, N. D. 1986. Notes on the Swainson's hawk in central Utah: insectivory, premigratory aggregations, and kleptoparasitism. Great Basin Nat. 46:302-304. Thomas E. Remington Wildlife Researcher �329 JOB PROGRESS REPORT Colorado State of: Project: W-152-R Upland Bird Research Work Plan: 26 Job Title: Analysis Period Covered: Author: Job _1_ of Upland Bird Population 01 January through 31 December Trends 1990 Clait E. Braun Personnel: C1ait E. Braun, Ken M. Giesen, Thomas E. Remington, Snyder, Colorado Division of Wildlife Warren D. ABSTRACT The following were prepared Braun, C. E. 1990. 1990. 1990. counties, Blue Mountain Cold Spring Mountain Sage grouse harvest Colorado, 1976-90. sage grouse harvest harvest data, Eastern Moffat and Western Eagle County sage grouse harvest 1990. Gunnison 1990. Middle 1990. Northcentral 1990. North Park sage grouse harvest 1990. Piceance 1990. Yampa Area sage grouse harvest R. W. 1990. data, 1976-90. data, 1976-90. 1990. Giesen, K. M. Colorado, 1990. .Hoffman, in 1990. data, 1990. Basin sage grouse harvest Park sage grouse harvest Routt data, 1990. data, 1990. Moffat County sage grouse harvest data, 1976-90. data, 1990. Basin sage grouse harvest data, 1990. data, 1990. 1990. Columbian sharp-tailed grouse harvest data, northwest 1976-90. Lesser prairie-chicken breeding survey, 1990 . 1990. Blue grouse wing analyses, Blue grouse wing analyses, Gunnison Northeast Region. Basin. �330 1990. Blue grouse wing analyses, Northwest Region. 1990. Disease monitoring guidelines for Merriam's wild turkey in Colorado. 1990. Results of tests for Mycoplasma in wild turkeys trapped in the Northwest Region. 1990. Rifle versus shotguns only for hunting wild turkey. Remington, T. E. Mountain. 1990. Analysis of blue grouse harvest trends on Blue Prepared by Clait E. Braun Wildlife Research Leader � https://cpw.cvlcollections.org/files/original/231b7e34bd7af6897431f00757ef261b.pdf 1c66cb0e132bc9434714d9cdd1176e1f PDF Text Text 1 Colorado Division of Wildlife Wildlife Research Report July 1991 JOB PROGRESS REPORT State of ~C~o~l~o~r~a~d~o~ _ Project No. W-1S3-R-4 Mammals Research Work Plan No. Multispecies Investigations Job No. Period Covered: Author: Mammals Publication, Editing, and Libraty Services 7 July 1, 1990 - June 30, 1991 J. A. Boss Personnel: J. A. Boss, N. W. McEwen . Abstract • 'Dul;:ingthe i990-91, Segment· 3·pub1~cations were purchased wi~h .Mamma1s -~esearch funds for' the Research Library, "13 publi-cations were acqui red ac BO cost, -and 20 theses were obtained via Interlibrary Loan or as gifts. _-Mammals Research staff had 13 manuscripts published in scientific journals or symposia proceedings. An additional 6 manuscripts were submitted for peer review in anticipation of publication. .r: ��3 MAMMALS PUBLICATION. EDITING AND LIBRARY SERVICES Jacqueline A. Boss P. N. OBJECTIVE To provide a centralized support program for manuscript editing and library services to facilitate publishing results of research conducted by staff of Federal Aid Project W-153-R. SEGMENT OBJECTIVE To provide a centralized support program for Mammals Research editing, library, and publishing services so that Mammals Research personnel can be most efficient in publishing results of their research. SUMMARY OF SERVICES Publications Purchased with Mammals Research Funds and Placed in the Research Center Library Bromley, M. 1989. Bear-people conflicts: proceedings of a symposium on management strategies. Northwest Territories Dept. of Renewable Resources. Yellowknit:~, Northwest "TerritorIes, Cattada. 246pp. Hoage, R.-J. 1989. Perceptions of animals in American culture. Institution Press. Washington, DC. l51pp. Forest, L. R. 1988. Field guide to tracking animals in snow. " Books. Harrisburg, PA. 185pp. Smithsonian Stackpole Publications Obtained Free or at Low Cost In addition to books purchased with Federal Aid Funds, about 13 free reports and short publications from state or federal agencies or from private sources were located, ordered, and obtained for use by Mammals Research personnel. Theses Purchased. Obtained on Interlibrary Loan or as Gifts for Use by Researchers Allen, R. B. 1985. Research and management implications of the pursuit of black bears with trained dogs. M.S. Thesis, University of Montana, Missoula. 51pp. Allred, W. S. 1989. The effects of Abert squirrel herbivory on ponderosa pines. Ph.D. Dissertation, Northern Arizona University, Flagstaff. 107pp. "Alt, G~_".L. 1989._ Reprodu~ti:v~"bio10gy .of female_black bears and early gro~th " _ and development of cubs in nortnwestern Pennsylvania. Ph.D. Dissertation, West Virginia University, Morgantown. 127pp. �4 Ball, L. C. 1989. Efficiency of nutrient and energy assimilation from mice and himalaya berries by gray foxes. M.S. Thesis, Humboldt State University, Arcata, CA. 59pp. Borden, D. L. 1990. Behavioral response of raccoons (Procyon lotor) to experimental manipulation of density. M.S. Thesis, Tennessee Technological University, Cookeville. 38pp. Bown, R. R. 1988. Beaver habitat along rivers and reservoirs in central Montana. M.S. Thesis, University of Montana, Missoula. Cassirer, E. F. 1990. Responses of elk to disturbance by cross-country skiers in northern Yellowstone National Park. M.S. Thesis, University of Idaho, Moscow. 101pp. De Marais, B. D. 1986. Morphologic variation in Qils (Pisces: Cyprinidae) and geologic history: lower Colorado River basin. M.S. Thesis, Arizona State University, Tempe. 85pp. Easter-Pilcher, A. L. 1987. Forage utilization; habitat selection, and population indices of beaver in northwestern Montana: M.S. Thesis, University of Montana, Missoula. 88pp. Evenden, A. G.- 1989, Ecology and distribution of riparian vegetation in the Trout Creek Mountains of southeastern Oregon. Ph.D. Dissertation, Oregon Sta~e University, Corvallis. 156pp. Gentile'- J. R. - 1983. The evolution and geographic aspects of the' antitrapping movement: -a classic resource cbnf1icc. -Ph.D. Dissert~tion, --Oregon State University, Corvallis. l45pp. Herriges, J. D. 1986. Movement, activity and habitat use of white-tailed deer along the lower Yellowstone River. M.S. Thesis, Montana State University, Bozeman. l32pp. Hewitt, D. G. habits. 1989. Correcting grizzly bear fecal analysis to observed food M.S~ Thesis, Washington State University, Pullman. 20pp. Keay, J. A. 1990. Black bear population dynamics in Yosemite National Park. Ph-.D. Dissertation, University of Idaho, Moscow. 126pp. Loy, R. R. 1981. An ecological investigation of the swift fox, (Vulpes velox) on the Pawnee National Grasslands, Colorado. M.A. Thesis, University of Northern Colorado, Greeley. 64pp. Neal, A. K. 1990. Evaluation of mark-resight population estimates using simulations and field data from mountain sheep. M.S. Thesis, Colorado State University, Fort Collins. 198pp. Reeves, K. A. 1989. Summer diet and status of river otters on Redwood Creek. M.S. Thesis, Humboldt State University, Arcata, CA. Rohlman, J. A. 1989. Black bear ecology near Priest Lake, Idaho. .'rhesis,_~iv!_~sity of Idaho, Mos.cow. 76pp. M.S . �5 Travis, S. E. 1990. Resources and sociality: effects of food resource variations on the mating system of Gunnison's prairie dog (Cynomys iUDDisoni). M.S. Thesis, Northern Arizona Univers~ty, Flagstaff. 63pp. Yeager, B. 1984. Use of cover by mule deer on a western Montana winter range. M.S. Thesis, University of Montana, Missoula. 76pp. Reference Document Location and Delivery The Research Center Library staff also located and delivered about 878 individual articles on request for Mammals Researchers during this segment; about 19 were not available locally and were obtained through Interlibrary Loan procedures. Manuscripts Published in FY 1990-91 Job Progress Reports; Federal Aid. All studies. Anderson, A. E. 1991. Frequency of mountain lion sightings by residents and employees of a housing development. (Abstract). Mountain Lion-Human Interactions Workshop and Symposium. Colorado Division of Wildlife, Fort Collins. (in press). Bear, G. B. 1990. A century of Colorado elk. Co~o. Outdoors 39(5):5-7. ·Carpenter,.L.·H.'. 19!H. Elk hunting regulations: 'the Colorado experience. Elk Vulnerability Symposium. Montana State University, Bozeman£ (in press). Gill, R. B .• and T. D. 1. Beck. 1991. ·W. Black Bear Workshop, Proc. 7. Black bear status report: Colorado. (in press) . Hobbs, N. T., and M. W. Miller. 1991. Interactions between pathogens and hosts: simulations of population regulation in bighorn sheep. in D. R. McCullough, ed. Wildlife 2000: populations. Elesevier Scientific. (in press). _____ , D. S. Schimel, C. E. Owensby, and D. J. Ljima. 1991. Fire and grazing in the tall grass prairie: contingent effects on nitrogen budgets. Ecology. (in press). Miller, M. W., N. T. Hobbs, and E. S. Williams. 1991. Spontaneous pasteurellosis in captive Rocky Mountain bighorn sheep (Ovis canadensis): clinical, laboratory, and epizootiological observations. J. Wildl. Dis. 27. (in press). O'Brien, S. J., M. E. Roelke, M. Yuhki, K. W. Richards, W. E. Johnson, W. L. Franklin, A. E. Anderson, O. L. Bass, Jr., F. C. Belden, and J. S. Martenson. 1990. Genetic introgression within the Florida panther (Felis concolor coryi). National Geographic Res. 6(4):485-494. Pojar,_ T. M., an4 R. B. Gill. _ 1990. Harvest management options for a pioneering pronghorn popUlation. Pronghorn Antelope Workshop, Proc. 14:112-122. . �6 Reed, D. F. 1990. Mirror-image stimu1ation--what can we ascertain about mountain goat social structure? Bienn. Symp. North Wild Sheep and Goat Counc. 7:274-281. Scribner, K. T., M. H. Smith, R. A. Garrott, and L. C. Carpenter. 1991. Temporal, spatial, and age-specific changes in genotypic composition of mule deer. J. Mammal. 72(1):126-137. Wild, M. A., and M. W. Miller. 1991. Detecting nonhemo1ytic Pasteurella haemglytica infections in healthy Rocky Mountain bighorn sheep (Ovis canadensis): influences of sample site and handling. J. Wildl. Dis. 27:53-60. ______ , and M. W. Miller. 1981. Bottle-raising wild ruminants in captivity. Outdoor Facts, No. 114. Colorado Division of Wildlife, Fort Collins. 6pp. Manuscripts in Review FY 1990-91 Andelt, W. F., D. L. Baker, and K. P. Burnham. 1991. Relative effectiveness of repellents for reducing elk damage. J. Wild1. Manage. (in review). Bartmann, R. M.·, G. C" White, and L. H. Carpenter·. 1991. Compensatory mortality in a Colorado mule deer population. Wildl. Monogr. (in review) • Pojar:. T.·M. 1991. 'CensUs technique~:' in R. McCabe, ed. Pronghorn management. Wildlife Management Institute, Washington, DC. (s~bmitted 6 Mar. 1991). 1991. The use of population models. in R. McCabe, ed. Pronghorn management. Wildlife Management Institute, Washington, DC. (submitted 6 Mar. 1991). ____ , D. C. Bowden, and R ..B. Gill. 1991. Evaluation of census techniques for pronghorn~ J. Wildl. Manage. (in review). Wild, M. A., M. W. Miller', D. L. Baker, R. B. Gill, N. T. Hobbs, and B. J. Maynard. 1991. Comparison of growth rate and milk intake of ·bott1e raised and dam-raised bighorn sheep, pronghorn antelope, and elk neonates. Can. J. Zool. (in review). o Prepared by ,~'~ (J ~ .; ABoJ.L ~~~ ac~lEline A. Boss ibUian �7 Colorado Division of Wildlife Wildlife Research Report July 1991 JOB PROGRESS REPORT State of ~C~o~l~o~r~a~d~o~ _ Project No. W-lS3-R-4 Mammals Research Work Plan No. 1 Multispecies.lnvestigations Job No. 9 Mammals 1 Research Administration Period Covered: Author: July 1, 1990 - June 30, 1991 R. B. Gill Personnel: R. B. Gill, L. E. Lovett Abstract Accomplishments ~f this Job are summarized in the Job Progress Report for W-lS3-R-4 Work Plan ~·Jo~ 3. R. Bruce Gill Wildlife Research Leader ��Colorado Division of Wildlife Wildlife Research Report July 1991 JOB PROGRESS REPORT State of ~C~o~l~o~r~a~d~o _ Project No. W-153-R-4 Work Plan No. ~D~e~e~r~I~n~v~e~s~t~iQg~a~t~i~o~n~s Job No. Personnel: _ Development of Census Methods for Deer in Plains Riverbottom Habitats Period Covered: Author: Mammals Research July 1, 1990 ~,June 30, 1991 R. C. Kufeld D. Bowden, D. Younkin, J. Slater, B. Will, G. Ausmus, G. Everhart Abstract During January 3-9, 1991, 68 white-tailed deer were marked along the Arkansas River between Prowers, Colorado, and the Kansas State Line. Radiocollars and 2 numbered eartags were placed on 7 adult buc~, 32 adult does, and 5 doe fawns. Eleven ·adult bucks and 13 buck fawns r~ceived only eartags. ; These radJo-cq11ared:deer'vere located at about 2-week intervals· through Ju~y, '1991. Deer radio-corlared in·the South Platte Riverbottom during ,1987-89 weFe located at about 2-week intervals from July 1, 1990, through December 31, 1990. Deer in the Arkansas River between Prowers and Carlton, and the lower 12 Km of Big Sandy Creek, were counted and classified by sex and age 6 times by helicopter during January 21-24, 1991. Observers also recorded the number of radio-collared deer they saw during each count. Radio-collared deer were located by fixed-wing aircraft while. counts were in progress, so the number of radio-collared deer in the census area was known. A mean of 17% of radiocollared bucks and 43% of radio-collared does and fawns were seen. The population for the area flown 6 times was estimated within 38% with 95% confidence for bucks, within 12% for does and fawns, and within 14% for all deer. One count was made of the Arkansas River from John Martin Dam to the Kansas line plus the lower 12 km of Big Sandy Creek. Data suggest an estimated popUlation in that area of 806 white-tailed deer including 266 bucks and 540 does and fawns. Only 8 mule deer were counted in that river segment, which suggests that the Lower Arkansas Riverbottom habitat contains nearly all white-tailed deer. Buck/doe and fawn/doe ratios recorded during counts were believed to be too lo~, because (1) radio-collared does and fawns were found to be more observable than radio-collared bucks; and (2) the counts were made relatively late in the season when some bucks may have lost their antlers resulting in their being classified as does, and large size of fawns may have resulted in some fawns being classified as does. It is recommended that counts to determine population trends and sex and age ratios be made in December, when bucks still have their antlers and fawns are smaller, so both can be more easily distinguished from does. It is also recommended that the difference in obse!yability of bucks versus does an9.fawns be t~~en into consideration whe~ analyzing sex and age data, and when·adjusting aerial·deer count- data for unseen animals to project total population size~ ��11 DEVELOPMENT OF CENSUS METHODS FOR DEER IN PLAINS lUVEllBOTTOH HABITATS Roland C. Kufe1d P. N. OBJECTIVES 1. To determine seasonal movements and home range size of white-tailed and mule deer in plains riverbottom habitats. 2. To develop and test methods for estimating size of deer populations in plains riverbottom habitats. SEGMENT OBJECTIVE To determine seasonal movements and home range size of white-tailed and mule deer in plains riverbottom habitats. STUDY AREA The South Platte study area is the South Platte River extending from Platteville, Colorado, to the Nebraska State Line. It is dominated primarily by Cottonwoods (Populus sargentii) and ~~llows (Salix spp.). Th~s river .is. bordered mostly by agricultural lands, mainly ccznf'Le Lds'.. Some _stretches are bordered by rangelands dominated by mixed prairie ·or.sand sagebrush .(Artemisia £ili£olia) (Costello 1954). Streamflow in the South Platte is greatest of th~ plains rivers in eastern Colorado. The Arkansas study area is the Arkansas River extending from John Martin Dam to the Kansas State Line. The riverbottom supports some stands of cottonwood, but these are rapidly deteriorating due to lower river flows caused by upstream dam construction. The dominant vegetation along lower portions of the Arkansas River in Colorado is .sa1t cedar (Tamarix pentandra), which is replacing the cottonwoods (Snyder, 1988). Most of the riverbottom along the lower Arkansas River is bordered by agricultural lands. Deer populations along the South Platte River consist of both mule deer and-whitetails, but whitetails are the predominant species. The deer population within the lower Arkansas Riverbottom itself consists of almost all whitetails, but mule deer are common on the plains near the river. METHODS AND MATERIALS During January 3 through 8, 1991, white-tailed deer were captured and marked in the Arkansas Riverbottom between Prowers Bridge and the Kansas State Line, and along the lower 3 km of Big Sandy Creek. Some adult bucks, all adult does, and all female fawns received a radio-collar and 2 orange, numbered eartags. Radio collars placed on adult bucks were designed to expand and contract to allow for growth and expansion during the rut. Due to the limited number of expandable collars available, not all adult bucks captured were radio collared. Some adult bucks and all male fawns received only 2 orange, numbered eartags. Deer were captured in Clover traps (Clover 1956) baited with eazed eorn and alfalfa hay<--- �12 White-tailed and mule deer, radio-collared in the South Platte Riverbottom during January and February, 1987, 1988, and 1989 (Kufeld 1987, 1988, 1989), were located, primarily by aerial telemetry, at approximately 2-week intervals through December, 1990. White-tailed deer radio-collared in the Arkansas Riverbottom during January, 1991 were located, in the same manner, at approximately 2-week intervals through June, 1991. Deer locations were recorded by UTM coordinates and plotted on USGS 1:50000 scale maps. Vegetation type for each deer location was also recorded. During January 21 through 24, 1991, deer were counted 6 times by 2 observers (in addition to the pilot) in a Bell Soloy helicopter in the segment of the Arkansas between Prowers bridge and Carlton bridge, a distance of approximately 41 river km. The count area also included the lower 12 Km of Big Sandy Creek. This census area contained 98% of the radio-collared deer. Deer were counted once in the entire river segment from John Martin Dam to the Kansas State Line, a distance of approximately 99 km. plus the Big Sandy Creek section. Between John Martin Dam and Carlton bridge main roads, including state highways SO and 196 and Prowers County Road HR, parallel the river averaging about 1.5 km to the north and south o~ the river. The census area included the riparian habitat along the river, which averaged about 0.4 to 0.8 km in width, as well as agricultural lands between those roads. From Carlton bridge to the Kansas State Line other boundaries were used which included agricultural lands about 1 Km north and south of the riparian zone bordering the river. As the helicopter moved along the river it flew at treetop altitude and followed a zig-zag course from the outer edge of the riparian zone on one side of the river to the outer edge of t~e riparian zone on the other side. The adjacent agricultura~ lands in the censuS.area were ve~ open to vieli. Thus~ flights.·into ·the-ag~icultuJ:'alareas were made onl)i· to count and classify deer which were seen· from the hel~copter as it ~urned at the edge of the riparian zone. Each deer was counted and classified 'by species, sex and age, and a notation made if a deer was wearing a radiocollar. During each day that helicopter counts were conducted, a fixed-wing airplane flight was also made to locate all of the radio-collared deer. Thus, the number of radio-collared deer available to be seen in the census area while counts were in progress was known, and the proportion of radio-collared deer seen by. observers in the helicopter provided an estimate of the accuracy of their count. Unpaired-t tests were used to analyze sex and age ratio differences between counts. Differences were considered significant if P S 0.10. RESULTS Deer trapping and monitoring Sixty-eight white-tailed deer were marked along the Arkansas River between Prowers, Colorado, and the Kansas State Line and the lower 3 Km of Big Sandy Creek during the January 3 through January 9, 1991, period (Table 1). This included 18 adult bucks (7 radio collared and eartagged, and 11 only eartagged), 32 adult does (all radio collared and eartagged), 5 female fawns (all radio-collared and eart~gged), ~Ild 13 male fawns (all eartaggecl). No mule deer were c~ught'-auring the trapping operation. AnalysiS of data for �13 Tabla 1. Whit.a-t.ailad daar t. •••ed alema t.haArk__ aa lUvar bat._en Prowar., Colorado IIDdt.haKan.a. St.at.a Lin. Japuary 3. t.hroushJapuary 8, 1991.' Eart... As. whan Dat.a Radio-collar attachad 'numb.r Sax capturad capt.ur.d Captura locat.ion 148,140 Doa Adult. 1-3-91 Ro.ar. farm, O.S mi. E. of Prowar. Co. Rd. 13 149.210 Doa Yrl, 1-3-91 Rosar. farm, O.S mi. E. of Prewar. Co. Rd. 13 149.190 Doa Yrl, 1-3-91 Rosar. farm, O.S mi. E, of Prewar. Co. Rd. 13 149,422 Buck Yrl, 1-3-91 Rosar. farm, O.S mi, E. of Prewar. Co, Rd. 13 149,351 # Doa Yrl. 1-3-91 Bridsa farm, BiS Sandy Craak, 0.7S mi. H. of Hwy, 196 149,291 # Doa Adult. 1-3-91 Bridsa farm, BiS Sandy Craak, 0.7S mi. H. of Hwy. 196 148.060 Doa Yrl. 1-3-91 Bridsa farm, BiS Sandy Craak, 0.7S mi. H. of Bwy. 196 148.390 Doa Yrl. 1-3-91 Bridsa farm, BiS Sandy Craak, 0.7S mi. H. of Bwy. 196 148.040 Doa Adult. 1-3-91 Fayco.b farm, O.S mi. E. of Prowar. Co. Rd. 4 148.330 Doa Adult. 1-3-91 Fayco.b farm, O.S mi. E. of Prewar. Co. Rd. 4 Buck Fawn 1-3-91 Fayco.b farm, 0.2S mi. E. of Prewar. Co, Rd. S 148,1~0 Buck Adu-lt. 1-3-91 Cantar farm; 0.3 mi. E. of Prowar. Co. Rd. lS 149,431 Buck Adult. 1-3-91 Cantar farm, 0.3 mi. E. of Prewar. Co. Rd, lS 148,250 Doa Adult. 1-3-91 Cantar farm, 0.3 mi. E, of Prewar. Co. Rd. 16 Buck Fawn 1-3-91 Cant.arfarm, O.S mi. E. of Prewar. Co. Rd. 16 149,320 Doa Fawn 1-3-91 Cantar farm, 0.6 mi. E, of Prewar. Co, Rd. 16 Buck Fawn 1-3-91 C.nt.arfarm, 0.6 mi. E. of Prowar. Co, Rd, 16 Buck Fawn 1-3-91 McMillllDfarm, 0.4 mi. E. of Prowar. Co. Rd. 17 i 148,430 II Doa Adult. 1-3-91 McMillan farm, 0.4 mi. E. of Prewar. Co. Rd. 17 149,461 Buck Adult. 1-3-91 HcHillllDfarm, 0.4 mi. E. of Prewar. Co. Rd. 17 148,100 Buck Adult. 1-3-91 McMillllDfarm, 0.4 mi. E. of Prewar. Co. Rd. 17 148.130 Buck Adult. 1-3-91 Doranc.mp farm, O.S mi. W. of Kan.a. Stat.aLina Buck Yrl. 1-3-91 Dorancamp farm, O.S mi. W. of Kan.a. St.ataLina 148.290 $ Doa Fawn 1-4-91 Rosar. farm, O.S mi. E. of Prewar. Co. Rd. 13 Buck Fawn 1-3-91 Duvall farm, at.Prewa~. Co. Rd. 32 Buck Fawn 1-4-91 Rosar. farm, O.S mi. E. of Prewar. Co. Rd. 13 $ 149,770'" Doa Adult. 1-4-91 Rosar_ farm, O.S mi. E. of Prewar. Co. Rd., 13 Buck Fawn 1-4-91 Rosar_ farm, O.S mi. E. of Prewar. Co. Rd. 13 148,090 Buck ~ult 1-4-91 Bridsa farm, BiS SlIDdyCraak, 0.7S mi. H, of Bwy, 196 148,300 Doa Adult. 1-4-91 Bridsa farm, BiS SlIDdyCraak, 0.7S mi. H. of Bwy. 196 Doa, Yrl. 1-4-91 Fayco_b farm, O.S mi. E. of Prewar_ eo. Rd. 4 149.180 148,420 + 'Doa AdUlt 1-4-91 fayco_b f,!lrm, 0.2S mi. E., of Prewar. Co. Rd. S 149:200 + Doa Yrl. 1-4-91 Fayco_h farm, 0.2S mi.'E. of Prowar. Co. Rd. S 149,140 Doa Adult. 1-4-91 Cantar farm, 0.3 mi. E. of Prewer. Co. Rd,15' 148,400 ., Ooa Fawn 1-4-91 Cant.ar'farm,0.3 mi. E. of Prewar_ Co, Rd. 16 149,520 Doa Adult. 1-4-91 McMillan farm, 0.4 mi. E. of Prewar_ Co. Rd. 17 Buck Adult. 1-4-91 McMillllDfarm, 0.4 mi. E. of Prewar. Co. Rd. 17 Buck Yrl. 1-4-91 Dorancamp farm, O.S mi. ~. of Kan.a_ Stata Lina Buck Fawn 1-4-91 Dorancamp farm, O.S mi. W. of Kan.a. Stat.aLina Doa Adult 1-S-91 Rosar_ farm, O.S mi. E. of Prewar_ Co. Rd. 13 148,200 149,020 Doa Adult. 1-S-91 Rosar_ farm. 0.5 mi. E. of Prewar. Co. Rd. 13 149,370 Doa Adult 1-S-91 Bridaa farm, BiS Sandy Craak, 0.7S mi. H, of Bwy, 196 Buck Fawn 1-S-91 Fayco_h farm, O.S mi. E. of Prewar. Co. Rd. 4 & Buck Fawn 1-S-91 Fayco_h farm, O.S mi. E. of Prewar_ Co. Rd. 4 & Doa Adult 1-S-91 Fayco_h farm, O.S mi. E. of Prewar. Co. Rd. 4 148.3S0 Doa Adult. 1-S-91 Fayco_h farm, O.S mi. E. of Prewar_ Co. Rd. 4 148,410 148,270 Doa Adult. 1-S-91 Fayco_h farm, 0.2S mi. E. of Prewar. Co. Rd. 5 Doa Adult. 1-S-91 Grasmick farm, 0.7S mi. E. of Prewar. Co. Rd. 10 149.150 Buck Adult. 1-S-91 Cantar farm. O.S mi. E. of Prewar. Co. Rd. lS Buck Fawn 1-5-91 Cant.arfarm, 0.6 mi. E. of Prewar_ Co. Rd. 16 Doa Adult. 1-5-91 McMillan farm, 0.4 mi. E. of Prewar. Co. Rd. 17 148.240 148,340 Doa Fawn 1-6-91 Fayco_h farm, O.S mi, E. of Prewar. Co, Rd. 4 Doa Adult 1-6-91 Fayco_h farm, 0.2S mi. ,E. of Prowar. Co. Rd. 5 149.230 Buck Adult 1-6-91 Cant.arfarm, O.S mi. E. of Prewar_ Co. Rd, lS Doa Adult. 1-6-91 Cant.arfarm, O,S mi. E. of Prewar. Co. Rd. 16 149,380 * Buck Fawn 1-7-91 Rosar. farm, O.S mi. E. of Prewar. Co. Rd. 13 148.450 Doa Adult 1-7-91 Rosar. farm, O.S mi. E. of Prewar. Co. Rd. 13 149.110 Doa Fawn 1-7-91 Bridaa farm, BiS Sandy Craak, 0.7S mi. H. of Bwy. 196 Doa Yrl. 1-7-91 Fayco_h farm, O.S mi. E. of Prewar. Co. Rd. 4 149.240 Buck Yrl. 1-7-91 Fayco_b farm, 0.2S mi. E. of Prewars Co. Rd. 5 Buck Adult 1-7-91 McHillllDfarm, 0.4 mi. E. of Prewar. Co. Rd. 17 Buck Adult 1-7-91 Cantar farm, O.S mi. E. of Prewars Co. Rd. lS Buck Adult. 1-8-91 Dorancamp farm, O.S mi. W. of Kan.a. Stat.aLina 149.170 Doa Adult. 1-8-91 Fayco.h farm, O.S mi. E. of Prowars Co. Rd. 4 Buck Fawn 1-8-91 Fayco.b farm, 0.2S mi, E, of Prewar. Co, Rd. S Buck Yrl. 1-8-91 Bridsa farm, BiS Sandy Craak. 0.7S mi. H. of Bwy. 196 149,220 Doa Adult. 1-8-91 Bridaa farm, BiS Sandy Craak. 0.7S mi. H, of Bwy. 196 Buck YrL 1-8-91-" Bridsa farm, Bis Sandy Creak, 'O,7S m1. fL'-OfBwy, 196 in tba ._a 'Pair_ of daar da.ianat.adwit.ht.hafollowins .ymbols wara t.rappador ratrappad t.osat.bar trap and may ba ralatad a. mot.har- off.prins, or a_ twin.: #, ., $, ",'+, *, &, 69 101 128 129 130 131 132 133 134 13S 136 137 138 139 140 141 142 143 144 14S 146 148 149 lS0 lS4 lSS 179 180 181 182 183 184 18S 186 187 188 189 190 191 192 193 194 19S 196 197 198 199 200 201 202 203 204 20S 206 207 208 209 210 211 212 213 214 21S 216 217 ,218 219 220 ---r- �14 movements, home range size, and habitat use will be presented in a future report when periodic monitoring of deer is completed. Deer Counts Presence of snow cover is usually considered a prerequisite to conducting aerial counts of deer. Adequate snow cover in Southeastern Colorado is relatively rare and may occur only once in several years. Thus, deer counts can rarely be scheduled when snow cover is adequate. Therefore, the decision was made to fly during average winter conditions which likely meant inadequate or no snow cover. Du~ing the counts, snow cover varied from almost none most of the time to as much as 100% during part of one day. The increased snow cover did not appear to affect the observabi1ity of deer. Percentages of radio-collared deer in the count area were 98% on 3 counts, 95% on 2 counts, and 93% on 1 count. The proportion of radio-collared deer seen was relatively low, averaging less than 40% on the 6 counts (Table 2). Kufe1d (1989) reported that an average of 69% of the radio-collared, white-tailed deer were seen on 3 similar counts of deer in the South Platte Riverbottom between Platteville, Colorado, and the Nebraska State Line. Does and fawns were much more observable than bucks (Table 2). The number of bucks and does and fawns recorded during a count was adjusted upward, based on the percent of radio-collared deer seen during' that count, to provide an estimat~ of the total deer population in the area censused (Table 2). Relative precision of population estimates is based on adjust~d total deer seen during the 6 counts . .~e popu~atio"!lwas estimated wi~hin 38% with 95% confidence for bucks, within 12% for does and fawns and within 14% for all deer. Table 2. White-tailed deer seen on January 21-24, 1991, helicopter counts of deer in the Arkansas Riverbottom between Prowers Bridge and Carlton Bridge. % Q{ I~dio-coll~I§ Does & fawn§ Fl.ight ~ucks 33% 56% 1 44 0 2 3 17 33 46 4 17 17 43 5 6 12 II x SE seen All deer 52% 38 31 41 39 aa Iotal de~[ seen Does & Bucks fawns 30 150 23 107 83 23 110 22 18 103 111 22 Adju§ted total Does & Bucks· fawns 268 91 230 243 135 252 239 129 240 106 129 ill deer1 deer 359 473 387 368 346 446 17% 43% 39% 137 260 397 4% 4% 3% 20 12 21 :Adjusted upward for the percent of radio-collars seen.' Example: 17% of total buck radio-collars seen and 23 total bucks seen: 23/.17 - 135 adjusted bucks seen. On flight 3, which provided a count of deer in all river segments from John ·Martin Dam to the Kansas Line plus-the lower 12 km of-Big Sandy Creek, 218 white-tailed deer (38 bucks and 180 does and fawns) and 14.3% of the.radio- ., �15 c~llared bucks and 33.3% of the radio-collared does and fawns were observed. The estimated adjusted total population of white-tailed deer for this area is 806 deer including 266 bucks and 540 does and fawns. Only 8 mule deer were counted during flight 3 and none were counted during the 6 repetitive flights, so no population estimate for mule deer is available. However, it appears safe to suggest that the mule deer population in the riverbottom downstream from John.Martin Dam to Kansas and in the lower 12 km of Big Sandy Creek is very low. Although mule deer were not seen in the riverbottom itself they were frequently seen in upland agricultural lands and rangelands within 2 Km or so of the study area. Deer Sex and A&e Classification ~ite-tai1ed deer sex and age ratios in all river segments (Flight 3) were 32 bucks/lOO does and 54 fawns/lOO does. These are similar to mean sex and age ratios for the 6 counts between Prowers Bridge and Carlton Bridge (Table 3). No significant differences in buck/doe ratios occurred between any of the 6 counts. No significant differences in fawn/doe ratios occurred between counts 1,2,3,4, or 6, but fawn/doe ratios on count 5 were significantly higher than those on the other counts (Table 3). Estimating precision for buck/doe ratios was better than for fawn/doe ratios. Based on 6 counts the mean buck/doe' ratio was estimated within 15% of the mean with 95% confidence compared to-28% of the mean for fawn/doe ratios. Table 3. Sex and age ratios for white-tailed deer seen on January 21-24, 1991, helicopter-counts of deer in the Arkansas-Riverbottom between Prow:ers Brigge and Carlton Bridge. Flight Bucks/laO Does Fawns/lOO- Does 1 2 3 4 5 6 x SE 29 31 41 30 31 42 II 43 322 49 37 48 49 74 5 DISCUSSION The Arkansas Riverbottom and lower Big Sandy Creek in eastern Colorado present a rather unique situation for aerial censusing of deer in that the area is very long, narrow, and winding. Because of its shape the area is not suited for implementation of a randomized helicopter quadrat census system (Kufeld et al. 1980), or an aerial line transect system (Burnham et al. 1980, ~ite et al. 1989). Thus, it was deemed more efficient to census the entire area. DeYoung (1985), Beasom et a1. (1986), Bartmann et al. (1986), and Kufeld (1989) have shown that during aerial counts of white-tailed and mule deer somewhat fewer than 100% of the deer present are actually seen and recorded. Beasom et al. (1986) reported accuracy of aerial counts for whitetails was unaffected by sample intensity, but precision increas~d with percent of the area sampled. Since we sampled- the--ent1re-area of interest, ftndings of· - Beasom et a1. (1986) suggest the only way to increase our precision is with �16 replicate counts. constraints. The number of replicate counts is subject to economic Sightabi1ity of animals during aerial counts is influenced by factors such as speed, height above ground, transect width, observers, group size and vegetation cover (Caughley et al. 1976, Barnes et al. 1986, Samuel et al. 19.87). Aerial deer counts in plains riverbottom habitats should be designed and conducted with these factors in mind. For example, the helicopter should fly at a speed of from 56-72 km/hr, and altitude of about 30 m. Zig-zag paths taken by the helicopter across the census area should be sufficiently narrow that there is good visibility of the area between subsequent passes. Observers should be experienced and familiar with the area. Counts should be conducted only in winter when leaves are gone from deciduous trees and shrubs. Kufeld (1989) suggested that, when censusing deer in the South Platte Riverbottom in northeastern Colorado, counts should be conducted only when there is 100% snow cover. That should not be a restriction when censusing deer in the lower Arkansas Riverbottom, because the relatively rare occurrence of adequate snow cover in that area could preclude deer counting for several years at a time. Even though snow cover might increase the number of deer observed, a realistic population estimate can be obtained without snow cover as long as data are available (such as provided by this study) to show the mean proportion of deer observed when snow is absent. The relatively low proportion of radio-collared deer seen on flights of the Arkansas River compared to flights of the South Platte may have been influenced to some extent by the presence of snow cover on the South Platte, but was more likely due to a difference in vegetation. It is more difficult to see deer in the thick salt cedar vegetation along-the lower Arkansas River than it is in the more open Cotto~wood typ~ along the South :Platte: Accuracy of deer sex and age ratio data is dependent upon validity of~the . assumption that deer in all sex and age categories are equally observable. Leon et al. (1987) found no sex or age bias in the composition of white-tailed deer encountered during helicopter counts in Texas. This is contrary, however, to our finding that radio-collared does and fawns were more observable than instrumented bucks. Lower observability of radio-collared bucks casts doubt on the accuracy of our recorded buck/doe ratios even though differences among the 6.counts in buck/doe ratios were not significant. A higher ratio than 32 bucks/lOO does should have resulted if bucks and does were equally observable. Fur~her evidence that the buck/doe ratio observed on counts may be too low is that 56 adult bucks/lOO does were caught during the trapping operation, and it has been the author's experience that adult whitetail bucks are usually much harder to catch than does.· The suspected relatively low buck/doe ratios on counts could have been due to a tendency of bucks to be more wary, but could also have been at least partly due to the time of year counts were made. Since counts could not begin until trapping was completed, these counts were made rather late in the year for accurate classification by sex and age. Some bucks may have lost their antlers which would have made it difficult to distinguish bucks from adult does. This was not a problem with radio-collared bucks, however, because they had collar configurations which identified them as bucks. Higher variability in buck/doe ratios compared to fawn/doe ratios were reported by Bowden et a1. (1984), and Kufeld (1989). However, during our 6 flights of the lower Arkansas River we observed lower variability and higher estimating preci"sion .fo'r -buck/do_eratios. The higher variability of -fawn/doe ratios was largely due--to the significantly higher ratio on the 5th count. If that count is deleted standard errors for fawn/doe ratios on the remaining 5 �17 counts are slightly lower than for buck doe ratios. Because the counts were made relatively late in the year it was difficult to distinguish fawns from adult does due to their size. Thus, reported fawn/doe ratios, like buck/doe ratios, may also be too low. These findings give rise to the following recommendations. RECOMMENDATIONS It is recommended that counts to determine population trends" and sex and age ratios be made in December when bucks still have their antlers and fawns are smaller, so both can be more easily distinguished from does. It is also recommended that the difference in observability of bucks versus does and fawns be taken into consideration when analyzing sex and age data, and when adjusting aerial deer count data for unseen animals to project total population size. LITERATURE CITED Barnes, A. G.", G. T. E. Hill, and G. R. Wilson. 1986. Correcting for incomplete sightability in aerial surveys of Kangaroos. Aust. Wildl. Res. 13:339-348. Bartmann, R. M., L. H. Carpenter,"R. A. Garrott, and D. C. Bowden. 1986. Accuracy "of he.licopter ccuncs of mule" deer in pinyon-j.uniper woodland. Wildl. Soc. Bull. 14:356-363.' Beasom, S. L., F. G. Leon III, and D. R. Synatzske. 1986. Accuracy and precision of counting white-tailed deer with helicppters at different sampling intensities. Wildl. Soc. Bull. 14:364-368. Bowden, D. C., A. E. Anderson, and D. E. Medin. 1984. Sampling plans for mule deer sex and age ratios. J. Wildl. Manage. 48:500-509. Burnham, K. P., D. R. Anderson, and J. L. Laake. 1980. Estimation of density from line transect sampling of biological populations. Wildl. Monogr. 72. 202pp. Caughley, G., R." Sinclair, and D. Scott-Kemmis. survey. J. Wild1. Manage. 40: 290- 300. Clover, M. R. 201. 1956. Single-gate deer trap. 1976, Experiments in aerial Calif. Fish and Game 42:199- Costello, D. F. 1954. Vegetation zones in Colorado. Pages iii-x in H, D. Harrington, ed. Manual of the plants of Colorado. Sage Books, Denver, CO. DeYoung, C. A. 1985. Accuracy of helicopter surveys of deer in south Texas. Wildl. Soc. Bull. 13:146-149. Kufeld, R. C. 1987. Development of census methods for deer in plains . riverbottom habitats-:- C-olo: Div. Wild!., Wildl Res. Rep. July: ]:-1".;;21. �18 Kufe1d, R. C. 1988. Development of census methods for deer in plains riverbottom habitats. Colo. Div. Wi1d1., Wildl. Res. Rep. Ju1y:11-2l. Kufeld, R. C. 1989. Development of census methods for deer in plains riverbottom habitats. Colo. Div. Wi1d1., Wi1d1. Res. Rep. Ju1y:11-19. Kufe1d, R. C., J. H. 01terman, and D. C. Bowden. 1980. A helicopter quadrat census for mule deer on Uncompahgre Plateau, Colorado. J. Wi1d1. Manage. 44:632-639. Leon, F. G., III, C. A. DeYoung, and S. L. Beasom. 1987. Bias in age and sex composition of white-tailed deer observed from helicopters. Wi1d1. Soc. Bull. 15:426-429. Samuel, M. D., E. o. Garton, M. W. Schlegel, and R. G. Carson. 1987. Visibility bias during aerial surveys of elk in northcentral Idaho. Wi1d1. Manage. 51:622-630. Snyder, W. D. 1988. Dynamics of cottonwood regeneration. Wi1d1. Res. Rep. Apri1:327:393. J. Colo. Div. Wi1d1., White, G. C., R. M. Bartmann, L. H. Carpenter, and R. A. Garrott. 1989. Evaluation of aerial line transects for estimating mule deer densities. J. Wi1dl. Manage. 53:625-635. Prepared by ·~c%/dd Roland C. Kufe1d Wildlife Researcher C. �19 . Colorado Division of Wildlife Wildlife Research Report July 1991 JOB FINAL REPORT State of ~C~o~l~o~r~a~d~o _ Project No. W-153-R-4 Deer Investigations Work Plan No. Job No. 8 Period Covered: Author: Mammals Research Accuracy of mule and white-tailed deer harvest reported by deer hunters on mailed questionnaire surveys. July I, 1990 - June 30, 1991 R. C. Kufeld Personnel: R. C. Kufeld, H. D. Riffel, D. C. Bowden, Numerous Northeast and Southeast Region field personnel . Abstract Successful dee-rhUnt~-rs who -were checked in the field with their deer during the 1990'plains deer seasons were sent mailed questionnaires, and thejr .answers compared with their known kill to determine the degree of accuracy with which hunters report the species, sex, and age of their deer ~n the annual Colorado Division of Wildlife hunter questionnaire survey. Species was reported correctly by 96% of hunters who killed whitetails compared with 85% of hunters who killed mule deer. All hunters reported sex of adult deer correctly, but only 58% of fawns killed were reported as fawns. The remaining 42% were reported as adult does. It is recommended that harvest projections based on questionnaire data be adjusted to allow for reporting error before using them in population models or in making management decisions. During the plains deer seasons (October and December), hunter questionnaire data (adjusted for hunter error in reporting deer by species) show that mule and white-tailed deer made up 83% and 17% of the harvest in'1989, and 81% and 19% in 1990 in all plainS game management units. ��21 ACCURACY OF MULE AND WHITE-TAILED DEER HARVEST REPOR.TED BY DEER HUNTERS ON HAILED QUESTIONnAIRE SURVEYS R.oland C. Kufeld P. N. OBJECTIVES To determine the degree of accuracy with which successful plains deer hunters report the species, sex and age of their kill on the annual Colorado Division of Wildlife hunter questionnaire survey. SEGMENT OBJECTIVE Same as P. N. Objective. STUDY .AREA. All game management units in eastern Colorado where plains deer seasons were scheduled ~uring October and December of 1990. QUODS AND MATERIALS Colorado: Division of Wildlife District Wildlife M~nagers and other DOW f'LeLd persopnel stationed in areas where plains deer seasons were scheduled~were _ asked to participate in collection of field data. Data collection was duringthe course.of·their regular law enforcement activities when patrolling the October and December, 1990, plains deer seasons. When a successful deer hunter was encountered the officer recorded the hunter's name, date of birth, and hunting license number on his regular hunter contact form, and the species, sex and age (whether fawn or older) of the deer. If the hunter asked the officer what species of deer they had killed, the officer, before answering, asked the hunter what species they thought they had killed. Then that information plus the actual species was recorded on the contact form. At the end of the plains deer seasons, wildlife research personnel reviewed the contact forms and extracted the pertinent contact information. Since all plains deer licenses are specified licenses, names and addresses of all license holders are available prior to beginning of the season, and the annual, random, questionnaire survey sample is drawn before the season starts. A questionnaire was sent to each successful deer hunter who was contacted in the field. Only about 55% of deer hunters are sampled during the annual questionnaire survey. Thus, successful, field contacted deer hunters not selected in the annual questionnaire survey were sampled in a separate mailing using identical questionnaire forms. Questionnaire responses from all field contacted plains deer hunters were compared with field contact forms to·determine the proportion of questionnaires on which species, sex, and age (fawn or older) of deer was .reported c;:orr~ctlyand incorrectly. �22 Deer hunters in plains game management units have been surveyed, annually, by mailed questionnaire for a number of years. During 1989 and 1990, hunters were asked to report their deer kill by species. To date, the annual Colorado Division of Wildlife B~g Game Harvest Report has reported the number of deer killed, but the harvest was not separated by species. The 1989 and 1990 plains deer hunter questionnaire data were further analyzed a~ part of this study to provide an estimate of plains deer harvest of mule and white-tailed deer in each game management unit. Harvest estimates were adjusted for hunter error (error data acquired during this study) in reporting species of . deer harvested on mailed questionnaires. RESULTS AND DISCUSSION Only 85 usable questionnaire forms were received from hunters who were checked in the field with their deer by DOW personnel. There are at least 2 reasons for the low sample size.. (1) Most plains deer hunters return home each day following their hunt. Thus, unlike deer hunters in the mountains, relatively few stay in camps for several days at a time where they are available for contact by law enforcement personnel. Their tendency to return home after each day's hunt and the extensive road system on the plains 'reduces the likelihood that a successful deer hunter, 'accompanied by his or her kill, will be encountered by an officer. (2) Not all successful deer hunters contacted -returned their questionnaires even though up to 2 fo110wup questionnaires were' sent to nonrespondents. The proportion, of-questionnaires returned by hunters, known to have been 'suc;cessfu1,was 78%. Hunters who harvested a white-tailed deer (Odocoileus virg~nianus) had a higher rate of repoiting the speci~s correctly tha~ did hunters who harvested a mule deer (Odocoileus hemionus). More mule deer ,hunters tended to report their deer as a white-tail (Table 1). All hunters who harvested an adult buck or an adult doe reported the age and sex correctly (Table 1). Rasmussen and Palmateer (1984) also observed a low error rate of only 1.3% in reporting sex of adults, but their questionnaires on which sex was reported erroneously showed a tendency of hunters to report adult females as adult males. In our study, although the sample size is low, a large proportion of fawns of both sexes harvested were reported as adult does (Table 1). These findings have some important implications for management uses of harvest data derived from the hunter questionnaire. They suggest that the harvest of white-tailed deer projected from'questionnaire data may be too high and projected harvest of mule deer may be too low. However, harvest projections of each species could be adjusted according to the percentages reported in Table 1. The questionnaire data also suggest that, since fawns are frequently reported as adults, estimates of adult doe harvest from questionnaire surveys, may be too high by a substantial percentage, and estimates of fawn harvest may be too low. If fawn/doe ratios in the harvest, based on questionnaire data, are used in population models without adjustment for incorrect reporting of age of antler1ess animals, substantial error could occur in projected population estimates. . �23 Table 1. Percentage of deer, checked by Division of Wildlife personnel during October and December, 1990 plains deer season, that were reported correctly and incorrectly by species, sex, and age on mailed hunter survey questionnaires. % Reported Actual No. of Reported Correctly Incorrectly Deer Deer Killed Deer Killed Whitetail Mule Deer 50 96 _2. 4 52 Mule Deer Mule Deer Whitetail 28 85 _2 15 33 Adult Buck Adult Buck 36 100 0 Adult Doe Adult Doe 30 100 0 Fawn Buck Fawn Buck Adult Doe 5 63 37 .1 8 Fawn Doe Fawn Doe Adult Doe 2 50 50 .2. 4 Fawn (all) Fawn Adult 7 58 _2 42 -12 Sinc~mailed questionnaire responses from hunters with known 15% of hunters who killed a mule deer tend to report it as a of hunters who killed a whitetail tend to report it as a ~le the following procedure was used to adjust the 1989 and 1990 hunter questionnaire data for hunter error in reporting deer kills show that whitetail and 4% deer (Table 1), plains deer kill by species: Where: Reported harvest of -mule deer x 0.15 - A And: Reported harvest of mule deer x 0.04 - B Then: Reported harvest of mule deer - B + A - Adjusted harvest of mule deer. And: Reported harvest of whiteotailed deer - A + B - Adjusted harvest of white-tailed deer. Total adjusted harvest of both species must equal total reported harvest. Thus, if adjusted harvest of one species is more than the total reported harvest of both, the adjusted harvest of that species becomes equal to the total reported harvest of both, and the adjusted harvest of the other species is O. For example, this could happen if the ~eported harvest of mule deer was very high and reported harvest of whitetails very low. Adjustments were made only in certain game management units which had relatively large portions of white-tailed dee'r in t_lleharvest during 1989:-~Q (Table 2). �24 Tmb 2. Barve.t of mul. and whit.-t.iled d••r durins 1989 and 1990 Octob.r and D.cemb.r 9 9 llDi~ t2!all st•• , 87 88 89 90 91* 92* 93* 94* 47 41 9S* 78 56 20 41 84 92 51 47 71 15 141 527 39 i78 52 75 65 951* 96* 97 98 99 100 101 102* 103* 104 lOS 106 107* 109" 110 111 112 113* 114 115* 116 117 118"' 119 120 121 122" 123* 124* 125* 126* 127* 128* 129* 130'" 13Z'" 133 134 13S 136 137 138* 139'" 141 142 143 144 145"' 146* 147 Total 3S 21 lS 16 6S 4S 2S 26 71 ·29 18 9 13. 30 ·14 6 64 4 40 83 14 49 39 12 . 13 41 7 2 8 23 11 7 39 S 29 SO 0 2 31 ___a 2733 !:!!!UI~aU.1 4 4 2 1 47 26 0 23 S 21 82 0 S S 3 1 21 22 2 44 0 12 lS 6 2 3 2 S 8 4 0 . 14 0 1 0 S 11 4 25 11. 17 7 1!1 11 31 0 1 1 1 0 3 9 0 0 0 0 6 10 _0 558 pl.ina rifl. d•• r 990 Iotal ~1. 12•• , Wh!t!taH!I! Total Sl 4S 39 42 34 29 7 17 43 1 0 1 3 75 36 1 23 0 29 89 4 9 1 7 3 21 29 7 40 42 37 22 62 42 6S 68 83 77 102 41 89 97 54 48 92 37 143 571 39 190 67 81 67 28 28 76 37 22 9 27 .30 15 . S6 78 S7 17 36 81 104 48 37 82 7 141 521 66 186 42 104 49 23 26 '78 39 31 9 19 62 lS 6 11 69 15 44 108 S9 2S 66 46 27 24 72 7 '3 9 24 11 10 48 12 19 72 32 S6 39 9 31 S6 0 8 41 __ 2 9 2 3 21 4 27 32 0 31 84 3 12 27 __ 0 3291 2876 S 29 SO S2 0 6 7 5 3 0 2 10 0 0 0 8 4 0 1 6 7 8 41 20 31 9 9 6 28 2 0 0 0 0 S 12 0 0 7 0 12 14 _0 644 3S 32 82 53 44 79 78 86 106 40 90 105 S5 40 103 36 148 573 66 192 59 109 S2 23 28 88 39 31 9 ., 27 .66 lS 1"2 65 19 27 113 52 87 48 18 37 84 11 2 3 21 4 32 44 0 31 91 3 24 41 __ 0 3540 * Th ••• sam. manas.aent unit. had r.lativ.ly lars. portion. of whit.-tail.d d ••r in the harv ••t durins 198990. Barv •• t ••timate. in the •• unit. have bean adju.ted fo~ e~ror in reportina .peci.. of deer harve.t.d on mailed qu •• tionair ••• _.DUring the plains deer seasons (October and'December)., hunter questionnaire d&ta show that mule and white~tailed deer made up 83% and 17% of the harvest �25 in 1989, and 81% and 19% in 1990 (Table 2). Both species of deer occur throughout the eastern Colorado plains. Whitetails occur mostly in or near riparian h~bitat along rivers and creeks, and around lakes and ponds. Game management units where high harvest of whitetails occurred (Table 2) were those along the South Platte, Arkansas, and South Republican Rivers, Fountain Creek, and Big Sandy Creek, as well as tributaries of all of those streams. Mule deer occur in those same habitats as well as being found far out on the plains away from riparian habitat. Mule deer have occurred on the plains of eastern Colorado since before the first white men came to the state. According to Hunter (1949) there were no white-tailed deer in Colorado in the late 1940s. There is evidence that the whitetail population in eastern Colorado has grown at a relatively rapid rate (Kufeld 1989). If white-tailed deer popUlation trends continue upward_on the plains of eastern Colorado, competition between whitetails and mule deer could possibly result in a negative impact on mule deer, at least in riparian habitats where both species occur. A need may arise for separate management of white-tailed and mule deer populations in plains game management units in order to keep the 2 populations in balance. Also, as a result of the increasing whitetail population, deer damage to irrigated agricultural fields, which are commonly located near riparian habitat favored by whitetails, may become more of a problem. Deer popUlation control for purposes of alleviating damage to such fields may be more efficiently accomplished if white-tailed and mule deer populations are managed separately. Colorado deer hunting licenses have traditionally been issued with no specification as to species of deer that can be taken. " Manipulation of the population level of one deer species through hunting without changing the population level of the other species would be possible only by issuing deer licenses which specify speci-es of deer to be taken. This "would require "tnat hunee rs be "able to idei}.tify deer by species in the field. _ -" Data in Table 1 suggests that most hunters are able to identify mule and white-tailed deer. -The minority of hunters who cannot identify them correctly could probably be educated by including pictures of identifying characteristics in the license application brochure and with the specified license when issued. LITERATURE CITED Hunter, G. N. 1948. History of white-tail"ed deer of Colorado. and Fish Dept., Denver. 10pp. Colorado Game Kufeld, R. C. 1989. Development of census metho~s for deer in plains riverbottom habitats. Colo. Div. Wildl. Wildl. Res. Rep. July 1:11-17. Rasmussen, G. P., and J. R. Palmateer. 1984. Hunter error in reporting the sex of deer. New York Fish and Game Jour. 31:138-145. Prep!,"ed by ~ .. - . C .Ro1a 9~ C. K fela Wi1d1i£e Researcher ��Colorado Division of Wildlife Wildlife Research Report July 1991 JOB PROGRESS REPORT State of ~C~o~l~o~r~a~d~o~ _ Project No. W-153-R-4 Mammals Research Deer Investigations Work Plan No. Job No. Period Covered: Author: Compensatory Effects of Harvest in a Mul~ Deer Population 5 July 1, 1990 - June 30, 1991 R. M. Bartmann, G. C. White Personnel: A. W. Alldredge, L. H. Carpenter, K. C. Crane, W. I. Dean, W. Devergie, B. L. Dupire, D. J. Freddy, J. Frothingham, J. L. Godbey, V. K. Graham, J. P. Gray; R. Harthan, E. K. Jones,- J. D. Madison, J. E. Morris, D. G. Saltz, C. L. Vardaman, M. A. Warner, and numerous others from the Colorado Division of Wildlife and Colorado State University. Abstract. Two aerial line transect surveys were flown on the Ridge study area to try and detect the extent of deer reduction on the treatment unit during the December hunting season. The estimated decline from 57 to 32 deerfkm2 was about 40% greater than what could be accounted for by the estimated harvest. Based on other data, the deer population on the ·treatment unit is still considered above the target level of 40fkm2. Poor trapping conditions allowed only 71 fawns to be radio collared--far below the goal of 120 fawns. Estimated fawn survival rates were 0.320 (SE - 0.090) and 0.274 (SE - 0.078) on control and treatment units, respectively. Twenty-seven does were radio collared bringing the totals to 39 and 42 on the control and treatment units, respectively. Estimated adult survival to IS June was 0.949 (SE - 0.035) and 0.853 (SE - 0.(61) on the same respective units. All adult deaths except for 1 on each unit were hunting related. Fawn:doe ratios were estimated before (control 79:100, treatment 81:100) and after (control 67:100, treatment 64:100) the late season. Neither the differences between units for each survey, nor within units between the 2 surveys were significant (~> 0.050). The harvest estimate for the late season was 392 of which 32% were fawns. Male fawns continued to be larger than females in weight, total body length, and left hind foot length (~< 0.001), but there were no differences between study units (~~ 0.239). The estimated age structures of does harvested during the 1990 regular and late seasons were different (~- 0.050). As in 1989, more yearlings and fewer old does (>7 years old) were harvested during the regular seasons. To date, no meaningful differences have been found in body size parameters of yearling and adult does between the 2 study units. Som~ ~if.fere11:~es J,n_antler and body _size c;haracteristics.were. 4!!t;ec.ted with year+ing bucks, but sample sizes were small (1-9). ��29 COMPENSATORY EFFECTS OF HARVEST IN A MULE DEER POPULATION Richard M. Bartmann and Gary C. White P. N. OBJECTIVES 1. Increase the winter survival rate of mule deer fawns by lowering total deer density to reduce competition for for~ge during winter. 2. Increase the harvest rate of deer througn increased productivity of adult does and decreased natural mortality of fawns resulting from closer alignment of population size with carrying capacity. 3. Evaluate deer harvest rates and population response on a DAU basis resulting from bucks-only hunting and from buck hunting with additional antlerless permits designed for annual removal of 20% of the adult female population. SEGMENT OBJECTIVES 1. Reduce the winter population of mule geer on the Ridge treatment area to a density <40jkm2 and maintain' t?e density. for at least 3 y~ars~ 2. Estimate winter survival rates of fawns on control ~nd treatment.a.reas. 3. Estimate annual survival rates of adult females. 4. Estimate annual survival rates of adult ·males. 5. Estimate productivity 'of deer on control and treatment areas. 6. Estimate harvest rates of bucks, does, and fawns on control and treatment areas. 7. Estimate condition of fawns on control and treatment areas. 8. Estimate age structure of adult females on the treatment area and condition of yearling females on control and treatment areas. 9. Estimate age structure of adult males and condition of yearling males on control and treatment areas. METHODS Most methods remain the same as previously reported (Bartmann 1990). The main exception is the use of the staggered-entry Kaplan-Meier estimator (Kaplan and Meier 1958, Pollock et al. 1989) to estimate fawn survival rates instead of .the ..binomial e.s.timator. Unlike ..the binomial.es1:_imator, the Kaplan..-Me.ier._ estimator uses censored animals (those with failed radios or with collars that �30 dropped off prematurely) to calculate annual survival rates. For data with no censored observations, Kaplan-Meier and binomial estimates are identical. The past few years, the number of censored animals has increased. During the 1989-90 winter, a problem occurred with premature collar drop-offs beginning in mid-March. This past winter, there was a greater than expected number of radio failures. When such problems occur later in the winter, a considerable amount of information can be lost unless these animals are included in survival rate estimates. The exception to the use of the Kaplan-Meier estimator is when multifactor analyses are used for hypothesis testing. These situations require that fawn survival data be treated as a binomial process. RESULTS AND DISCUSSION Population Reduction From 1985 to 1988, deer density estimates from helicopter line transect surveys on control and trea~ent areas fluctuated around 70~ (Fig. 1). The removal of an estimated 451 deer by hunting in December 1989 was not reflected in line transect results that year as deer density on the control area, 47jkm2, was lower than on the treatment area, 66~. In 1990, 2 line transect surveys were flown, 1 immediately before and the other immediately after· the experimental December late season, to see if ~e could detect an imm.ediate·effect·of·the harvest on deer. density. Pre-season ..dens icy estimates seemed- reasonable.·.67 and 57 deerjkm2 on the control and treatment areas, respectively. An estimated_392 deer were harvested during the 1990 late season, but the post-season density estimate of 32 deerjkm2 indicated a population reduction about 40% greater. The predicted density was still well within the 95% confidence interval because of high variability. We presume this was due, in part, to observing fewer groups as a result of the lowered density. In the future, additional transects will have to be flown to increase the number of groups seen. Unexpectedly, density on the control area increased 27% to 85 deer~ and was significantly higher than on the treatment area (l < 0.05) despite wide confidence intervals. This increase was not explained by movement of animals because radio-collared deer remained on .their respective areas. The use of different observers for the 2 surveys was a potential source of bias, but we had no way to evaluate it. Based on other data, we believe true deer density on the treatment area is still above the target level of 40jkm2. Accordingly, the number of late season licenses for the December, 1991, season (350) will only be slightly less than in 1990 (400). Fawn Suryi val Deer were trapped from November 12-30, 1990. Extremely poor trapping conditions resulted in only 71 fawns being radio collared; slightly more than half the l2a fawns desired. Contributing to the problem was the limited trapping period that was constrained by the end of the third regular deer season and the start of the late season. Trapping may be mo~e_difficult in 1991-,-even with more··conducive weathe-r conditions, because. the number. of fawns �31 to be radio collared increases to 160, and there will be fewer deer available on the treatment area. Estimated fawn survival on control and treatment areas in 1989-90 was greater than 50% for only the second time (survival was 0.537 on the control unit in 1985-86) since collection of survival data began on the Ridge study area in 1982-83 (Table 1). That winter was quite mild, but there haye been similar winters without such high survival. These higher survival rates will decrease the chance of detecting differences in survival between the 2 areas over time because of the increased variation added to the survival time series. Table 1. Survival rate estimates (i) for radio-collared mule deer fawns on control and treatment units of the Ridge study area in Piceance Basin, Colorado, from time of collaring in November and December until the following 15 June 1982-83 through 1990-91. Winter 1982-83 1983-84 1984-85 1985-86 1986-87 1987-88 1988-89 1989-90 1990~-9l n r 29 28 34 59 60 32 34 38 34 Control unit SE(.~) i 0.321 0.071 0.196 0.537 0.431 _0.241 'o. 270 0.7580.320 0.088 0.049 0.078 0.070 0.064 .0-.077 0.083 0.078 0.090 n 31 32 26 58 58 28 28 44 36 TIeatment unit SE(i) i 0.387 0.033 0.431 0.439 0.471 0.107 0.445 0-.657 0.274 0.087 0.033 0.105 0.070 0.067 0.058 0.096 0.072 0.078 .f of equal i(.~) 0.578 0.774 0.075 0.157 0.565 _ 0.006 0.509 0.250 0.487 Winter conditions in 1990-91 were more typ~ca1 and estimated fawn survival rates on the control (0.320) and treatment (0.274) areas declined to near past levels and did not differ significantly from each other (l- 0.487). -Although deer density has been lowered on the treatment area, our previous efforts at lowering density by trapping in 1985 and ~986 indicated a considerable reduction will be required before there is a significant impact on survival. Further, stress from increased harassment during the montholong late hunting season may tend to enhance mortality on the treatment area. A higher than expected number of radio failures compounded the problem of poor trapping success (Table 2). Nine fawn radios were assumed to have failed; however, it is possible that at least some were on fawns killed during the late season and not reported. Most failures occurred with units purchased from 2 new vendors. As expected, hunting and starvation were major causes of fawn mortality on the treatment area. Nine fawns (27%) were known to have been shot and 8 (24%) were determined to have starved. In contrast, predation was the major mortality cause on the control area accounting for 10 fawn deaths' (37%) with .__.another 6 (22%}_dead of sta~ation ..'These numbers and survival rates may s.till change because effor_ts to account for all fawns will continue until fall, 1991. �32 Table 2. Cause of mortality treatment units of the Ridge of collaring in November and through 1990-91. Percentages Unit Winter No. of fawns Censored 1982-83 1983-84 1984-85 1985-86 1986-87 1987-88 1988-89 1989-90 1990-91 Treatment 1982-83 1983-84 1984-85 1985-86 1986-87 1987-88 1988-89 1989-9{} ,. 1990 .• 91 . Control for radio-collared mule deer fawns on control and study area in Piceance Basin, Colorado, from time December until the following 15 June 1982-83 are of total uncensored· fawns. 28 28 34 59 60 32 34. 38 34 31 32 26 58 58 28 28 44 ·36 1 7 11 6 2 4 14 7 1 1 4 9 5 3 9 3 Starvation % No. 15 22 16 10 14 22 10 3 6 15 27 8 17 16 19 9 6 8 56 79 59 21 26 73 33 13 22 50 87 36 35 30 68 36 17 24 cause Mortalit;)!: PIedation Hunting % % No. No. 4 4 5 7 17 15 14 19 15 31 5 17 Other No .. % 1 7 3 7 3 1 2 3 1 1 5 5 -4 3 14 2 2 18 20 11 9 2 2 9 2 3 2 11 13 1 1 33 7 10 9 22 25 4 4 3 9 .... 5 . 9 8 14 27 4 15 6 7 23 13 4 7 ._.Uncensored fawns are those with nonfai1ing radios or with collars ehac; did not drop off prematurely. Adult Doe Survival Twenty-seven does were radio collared, which in~reased numbers to 39 on the control area and 42 on the treatment area. Another 14 mature and 4 yearling does and 16 year1i~g bucks were measured and released. All yearlings were ear tagged for future identification as known-age. Compared to fawns, adult doe survival has been relatively high from year to year with little annual variation (Table 3) .. Through 15 June 1991, doe survival was 0.949 on the control area and 0.853 on the treatment area; the difference being mostly due to hunting mortality as only 1 deer on each area died from non-hunting causes. As with fawn radios, failures also were higher than expected with adult radios. There were 10 assumed failures but, again, some may have been for does killed during the late season and not reported. Seven does were killed during the late season. One had been trapped on the control area. The trapsite is close to the boundary between the 2 areas, so it is possible the deer was on the treatment area when killed. Of the 2 does not killed by hunters, 1 was killed by a mountain lion (Felis concolor) and the other ~ied from an undetermined cause during late February. These·numbers and survival rates-ean still change because adult survival is·calculated on an annual basis from 1 December to 30 November. �33 Table 3. Annual (1 Dec-30 Nov) survival rate estimates (i) for radiocollared adult female mule deer on control and treatment units of the Ridge study area in Piceance Basin, Colorado, 1982-83 through 1990-91. 'Winter 1982-83 1983-84 1984-85 1985-86 1986-87 1987-88 1988-89 1989-90 1990-91· • n 10 15 9 23 27 137 23 40 ContIol unit SE(i) i 0.800 0.779 1.000 0.955 0.709 0.818 0.857 0.765 0.949 0.126 0.113 0.044 0.093 0.116 0.132 0.105 0.035 lI!i:atmentunit SE(i) i n 11 15 10 21 17 11 5 28 35 0.909 0.929 1.000 0.900 0.811 1.000 0.800 0.672 0.853 0.087 0.069 0.067 0.099 0.179 0.089 0.061 .f of ' equal i 0.448 0.271 1.000 0.486 0.544 0.329 0.854 0.179 0.175 Survival rate estimates for 1990-91 are only to 15 June 1991. All radio-collared deer were located 3 times (January, February, and ,March) to check for movements between the treatment and control units. No fawns were located off the unit on which they-were trapped during any of the 3 periods. Only 4 does from the treatment Unit were located 1 or more times on the control unit, and 1 doe from the control unit was located 2 tf~es on the trea~ent unit. Thus, no major:movements ~ave been detecte4 that might be attriouted to the ~educed density on the treatment unit. Productivity 'We estimated post-season fawn:doe ratios to serve as a crude index to possible changes in productivity, i.e., recruitment to December or January, between the control and treatment ,areas. In 1988, the year before the first late season, ratios were not significantly different (f> 0.050), 73 and 70 fawns/100 does on the control and treatment areas, respectively. In 1989, the'ratio on the treatment area (106:100) was significantly higher (l< 0.050) than on the control (63:100), but was too high to be explained by the number and composition of deer removed. In 1990, surveys were flown before and again after ·the late season to see if we could detect an effect of the harvest. Pre-season ratios (control 79:100, treatment 81:100) and post-season ratios (control 67:100, treatment 64:100) were not significantly different (l> 0.050) between units, and neither were the decreases within units between the 2 periods (.f>0.050). As with the pre- and post-season line transect surveys, there were different observers for the 2 classification surveys. Age classification surveys will not be conducted in the future. Although we would expect an increase in fawn production as density declines, it would be difficult to detect such a change for several reasons. First, fawn:doe ratio estimates have been imprecise, usually no better than ±1520% of the mean at the 95% confidence level, which makes it d~fficult to ~~tect any except qui~e large changes. Second, ~he true fawn_production is confounded by t~e unkriown extent ot'mortality from birth to 6 months of age. TJhile it can be argUed that age ratios measure recruitment to 6 months of age, this information has little practical value and can also �34 be derived from census data. Therefore, the costs are better spent to expand line transect surveys because density estimates are more crucial to decisions on harvest levels. Harvest Several assumptions are made concerning harvest. One is that there is no harvest on the control area. We know this is not true, but assume the effect is negligible compared to the treatment area. Another is that there are no bucks killed. Again, we know this is not true. Several bucks killed in 1989 had antlers less than 5 inches long, which made them legal antlerless deer under the hunting regulations. Other bucks have either been checked or found abandoned in the field. Except for being illegal, these bucks have little effect on the study because they comprise a small proportion of the total population. Late season permits were increased by 25 to 400 in 1990 (Table 4). However, the estimated number of deer killed declined by 59 deer from the previous year to 392. The proportion of fawns in the'harvest (32%) increased slightly over that in 1989 (28%). Compared to 1989, more people reported hunting in 1990, but more were unsuccessful and more killed only 1 deer. The overall lower hunter success rate probably reflects a combination of lower deer density and snow conditions. that were worse than in 1989. Even with good snow conditions in 1991, we expect to see a further drop in hunter success because of still lower deer density. Table 4. Hunter participation and success and deer harvest estimates for the_December late·season on the treatment unit of the Ridge study area, 1989-90. 1989 No. 375 Number of Licenses· Number·of survey respondents 60 Hunter estimates - Did not hunt 81 Unsuccessful 37 Harvested 1 deer 63 Harvested 2 deer 194 Harvest estimates - Does Fawns Total 324 127 451 1990 % No. % 400 92 21.6 10.0 16.7 51.7 48 104 104 144 12.0 26.1 26.1 35.8 71.8 28.2 261 126 66.7 32.2 392b • Hunters were allowed to take 2 antlerless deer on a license. b One respondent reported harvesting a male fawn, but the deer checked in the field was a yearling (illegal) buck. This projects to 5 bucks and accounts for the difference between the estimated number of does and fawns harvested and the total. __ �35 Body Condition of Fawns Mean body weights of fawns on control and treatment units were within the range of means for the respective units during the previous 8 years (Table 5). Males were larger than females with respect to weight, total body length, and left hind foot length (l < 0.001) but, for all fawns, no differences were detected between units (l ~ 0.239). Table 5. Weights (kg) and body measurements (cm) of mule deer fawns on control and treatment units of the Ridge study area in Piceance Basin, Colorado, 1982-91. Weight SD Unit Year n Control 1982 1983 1984 1985 1986 1987 1988 1989 28 28 34 60 58 33 34 40 34.6 31. 7 32.3 32.6 31. 9 29.9 29.5 32.7 1990 1982 1983 1984_ 1985 1986 1987 1988· 1989 1990 35 30 32-. 26 60 61 28 30 47 36 30.8 4.29 32.8 4.18 3.1232.3 32.3 . 5.07 32-.3 4.62 31. 7 4.13 30.2 5.34 28.8 4.13 30.6 3.33 30.7 4.41 Treatment • n- h c nn- X 3.10 4.40 4.65 4.02 3.89 3.60 3.10 3.31 Total body l~ngth g SD Left hind foot length g SD 124.0 124.2 123.9 124.4 128.1· 127.3 123.8 131.0 4.64 5.65 7.25 6.26 6.53 6.12 7.83 5.81 41.1 40.6 40.8 41.1 41.0· 40.8 41.0 41.8 126.5 121. 7h 123.6 124.7 124.4 126.0· 127.5 124.9 126.6 127.6c 9.02 5.45 5.53' -7.25 6._28· 6.62 8.86 7.07 5.33 6.57 40.8 1.75 41.1h 1.65 40.6 1.34 89 40.8 _1 .. 40.8 1.77 41.0lL 2.11-41.2 1.66 40.6 1.88 40.7 1.41 41.1c 1.69 1.08 1.73 1.53 1.48 .95 1.72 1.37 2.16 60 31 32 Body condition was also indexed by the ratio of body weight to total body length. As with the separate measurements, there was no difference in mean ratios between units (l- 0.297). Age Structure and Body Condition of Does As in 1989, age estimates from dental cementum and from examination of lower jaws were inconsistent, particularly with the younger (yearling and 2 years old) and the older (>7 years old) age classes. Compared to Jaw aging, dental cementum tended to overestimate the age of younger animals and underestimate the age of older animals (Fig. 2). We are quite confident that we can accurately age yearlings and most 2-year-01ds from jaws. On the other hand, limited tests of the dental cementum technique with known-age teeth suggest errors for these 2 age classes may approach SOX, all biased high. Therefore, all analyses are based on age estimates derived from jaws. . �36 The estimated age structure of does harvested differed between the regular and late seasons (l- 0.050). During the regular seasons, more yearlings and fewer old (>7 years) deer were killed suggesting a differential vulnerability (Fig. 3). We do not know the true age structure of the population, but assume the late season harvest data are more representative because of the greater intensity of hunting pressure and the better hunting conditions. Age structures almost differed significantly between years (l- 0.087). The trend was towards more yearlings and fewer old animals in 1990. However, if the various ages of deer are being removed in similar· proportions to their presence in the population, we would not expect large changes in structure over time. Rather, the greatest effect would be from annual changes in fawn survival as it affects the number of yearlings recruited into the population. the increased proportion of yearlings killed during the 1990 late season did follow a winter of exceptionally high fawn survival. If this relationship holds true, we would expect the proportion of yearlings in the December harvest to drop in 1991 because fawn survival in 1990-91 was below 30%. As expected, carcass weights of adult does killed during the regular and late seasons in 1990 (44.4 kg, SD - 5.82) were greater than for yearlings (37.2 kg, SD - 4.12) (l< 0.001). Also, all deer harvested during the regular seasons averaged about 2 kg heavier than those taken during the December late season (l < 0.001). No differences in mean weights were found between deer killed in 1989 and 1990 (l- 0.618). There were no differences in live weights or total body lengths of adult does trapped in 1989 and 1990 (l~ 0.426), or between study units (l0.318) ·within years (Table '6)~ Likewise, li~e weights and total body leng,ths of yearling d~es did not differ between'years (f ~ 0.308), but ye,arlings on the treatment Unit averaged t?ear_ly6 cm Longez chan those 'on the control unit (l- 0.054). . Table 6. Weights (kg) and body measureme.nts (cm) of yearl~ng and adult female mule deer on control and treatment units of the Ridge s~udy area in Piceance Basin, Colorado, 1988-90. Unit Year n X Total bod:t:length SD X We1&bt SD Left hind foot length SD X IeAIl1n'gs Control Treatment 1988 1989 1990 1988 1989 1990 2 2 10· 2 7 8 46.2 46.0 49.4 50.2 53.3 50.8 1.91 6.43 2.38 0.35 9.61 3.38 146.1 148.4 147.9 151.2 158.7 150.1 10.75 2.26 2.49 3.18 11.03 4.94 45.8 47.8 45.8 46.0 46.7 45.6 0.99 3.04 1.08 168.6 166.7 165.9 166.8 5.20 6.98 6.76 5.81 48.1 47.3 47.8 48.1 1.10 1.09 1. 32 1.44 0.78 1.26 Adults Control Treatment 1989 1990 1989 1990 19 21 39 25 67.7 66.9 65.0 67.6 • Sample size for weight is 9. -'. 5.22 5.51 5.78 5.14 �Body Condition of Yearling Males There were no diffe~ences in mean antler lengths of yearling males between study units (l~ 0.782) or years (l~ 0.609) (Table 7). Live weights were nearly significantly different among years (l- 0.086). Yearling males on the treatment unit averaged nearly 5 kg heavier than those on the control unit (l- 0.004) and were nearly 8 cm longer in total body length (l- 0.02l). There were also differences in body length among years (l- 0.014) with the mean length in 1988 (148.5 cm) smaller than in 1989 (157.4 cm) and 1990 (157.1 cm). There were no differences in left hind foot length between areas (l- 0.093) or years (l- 0.836). Table 7. Weights (kg), body measurements (cm), and main antler beam measurements (cm) of yearling male mule deer on control and treatment units of the Ridge study area in Piceance Basin, Colorado, 1988-90. Weight Total bodv 19th Left h1l'ld foot 19th Unit Year n ! SO ! SO ! SO Control 1988 1989 1990 1988 1989 1990 6 2 2 6 1 9 54.0 48.5 52.2 51.8 60.8 60.6 4.91 5.30 3.99 4.01 145.1 151.8 154.8 151.9 168.5 157.6 11.53 8.13 6.67 6.07 47.4 46.7 48.0 48.5 50.0 48.7 1.39 0.71 3.52 1.19 Treatment 3.20 5.75 1.08 Main antler beam length Left Right SO SO t ! 16.9 15.8 16.6 17.3 14.5 19.3 3.49 1.77 4.47 3.47 5.52 16.4 18.4 15.6 16.9 13.5 20.1 2.30 0.49 4.03 4.11 6.07 SUMMARY The questionable accuracy and imprecision of recent deer density estimates based on aerial line transects mandate a cautious approa~h to evaluating the extent of the reduction on the treatment unit. The 1990 post-season density estimate indicates we have achieved the desired reduction, :-but co'rolLary data suggest the population is stil·l above our goal. Therefore, we consider 1 additional year of reduction to be necessary. Thereafter, the number of licenses will be reduced to a level that will maintain the desired density. Adjustments in the timing and length of the late season are also contemplated to allow a greater period to complete trapping, to reduce the period of stress on the deer,' and to reduce the cost of collecting harvest data. Although significant differences were found between many of the parameters we estimated, it would be surprising if they were influenced to any great extent by the reduction in deer density that has occurred thus far . .Rather, they probably are more a result of annual variation that makes it extremely difficult to separate treatment effects from environmental "noise". LITERATURE Bartmann, R. M. population. CITED 1990. Compensatory effects of harvest in a mule deer Colo. Div. Wildl., Wild1. Res. Rep. July:187-196. Kaplan, E. L., and P. Meier. incomplete observations. 1958.' Nonparametric estimation from J. Am. Stat. Assoc. 53:457-481. �Pollock, K. H., S. R. Winterstein, C. M. Bunck, and P. D. Curtis. 1989. Survival analysis in telemetry studies: the staggered entry design. J. Wi1d1. Manage. 53:7-15. Prepared by &/_~ IRichard M. Bartmann Wildlife Researcher 90r---------~------------------------------~~ CONTROL - 80 a: 70 C ..._, 60 - .•• ~ TREATMENT -- N ::E ~ ..._ w w \ \ ~ CiS 50 1------·-----------------------------------------,···--·---\----\ Z \ \ W C 40 __ ._ __ _ _--._-_ ..--. .. .. .. \ \ .... --... -.- ...----.-- ..... ---.~.-.-----. \ \ , '. 30~------~------~--------~------~------~~ 1986 1987 1988 1989 1990 1985 YEAR Fig. 1. Deer density estimates from aerial line transects during early winter. on control and treatment units �39 +5 6 : +4 (J'J •••• Z W +3 5 +2 7 8 2 W AGREE 6 16 17 17 9 a:: o Z a:: US >;,.... Z -c -, ~ 0:E 2 I 11 2 4 17 3 6 2 2 12 -2 3 8 4 2 11 -3 1 4 3 8 1 5 2 1 -1 -4 3 1989 -5 3 -6 +7 +6 ; ; ; +5 , 1 i 1 ·1' j . -_ J ! 1 I 3 ; +1 I 19 1, 13 I 11 26 2 ~ ~ , , +4 , , ; I , -~ AGREE - -1 > W 0 I , ,I +2 0 ! 1 i o -_ +3 a: u, Z 17 : W C!l 8 1 +1 :E 2 6 . - 0-0 4 5 , 2 1 , 5 2 I 6 , ! ! 7 I 4 2 -2 2 2 -3 I : 8 3 9 I : 2 : 1 2 I 5 6 1 3 2 1 -4 2 17 I 15 I 4 3 3 , 2 1 1990 , -5 1 ; -6 1 2 1 345 6 7 >7 JAW AGE (YEARS) Fig.- 2. Comparisons of deviations between dental cementum and jaw/incisor-based age estimates of does harvested during the 1989 and 1990 hunting seasons on the Ridge study area. Values are numbers of animals. �40 • - REGULAR SEASONS .• - LATE SEASON 50 ~----------------------------------------------~ 1989 40 ---- -.---- ..- --- - - - -- - 30 ------_._--_._ .. __ , - - -..- . I ._ __ .._ --_._ _ _ I·· _._ ...•. I I 20 I- Z W 0 ~------~----~------~----~------~----~------~ ~ W c.. 50 ~------------------------------------------------~ 1990· . 40 ~-----------~---.-------- .. ..-.-------.-------~. _ _-_ _ _ _ ..•......... _ _--_ _. .. - ...... -.- .--f..-I --_._--_ _. __ ._---_ _._ _ _ - - -. .................. _/I I l 10 --._-- 1 2 3 4 5 6 7 >7 ESTIMATED AGE (YEARS) Fig. 3. Estimated study area, age structure 1969 and 1990. of does harvest during regular and late seasons on the Ridge �Colorado Division of Yildlife Yildlife Research Report July 1991 JOB FINAL REPORT State of ~C~o~l~o~r~a~d~o _ Project No. Y-IS3-R-4 Elk Investigations York Plan No. Job No. 2 Period Covered: Author: Mammals Research Trapping, Transporting, and Maintenance of Elk at Livestock-Elk Grazing Study July I, 1990 - June 30, 1991 G. D. Bear Personnel: Bert Rakestraw, Chuck Yoodward Abstract Four years of field work have been comple"ted and tech~ical publicati~-!_lS are being prepared under Federal Aid Reports York Plan 3, Job 5; and York Plan 9A, Job I (Impact of Elk Yinter Grazing on Livestock Production). ��43 TRAPPING TRANSPORTING AND MAINTENANCE OF ELK AT LIVESTOCK-ELK GRAZING STUDY J J G. D. Bear P. N. OBJECTIVE To provide assistance to the Livestock-Elk Grazing Study near Maybell, Colorado, by capturing and maintaining an eXperimental elk herd. SEGMENT OBJECTIVES Same as P. N. Objective. METHODS AND RESULTS The primary purpose of this job was to assure the experimental pastures near Maybell in northwestern Colorado were properly stocked with elk and cattle during the designated grazing periods, which is a satellite of a more detailed job reported on by D. L. Baker (Federal Aid Report Work Plan 3, Job 5) and N. T. Hobbs (Federal Aid Report Work Plan 9A, Job 1): Impact of Elk Winter G~azing on Livestock Production . • 'Four years o-f field research has been completed and the final results ar~ :being prepared.~oi·technical pub~ication. . Prepared by _~ -i',--·__ ceorge:D.iear SJ. ~_"__e_e_,,--::~~ Wildlife Researcher ��45 Colorado Division of Wildlife Wildlife Research Report July 1991 JOB PROGRESS REPORT State of ~C~o~l~o~r~a~d~o _ Project No. W-153-R-4 Work Plan No. Elk Investigations Job No. 6 Period Covered: Author: Mammals Research Effect of Elk Harvest Systems on Elk Breeding Biology July I, 1990 - June 30, 1991 D. J. Freddy Personnel: M. Miller, M. Cousins, C. Wetherill, Colorado Division of Wildlife; D. Bowden and G. White, Colorado State University; E. Ryland, Forbes Trinchera Ranch. Abstract . We· evaluated' effects of blood type. (plasma or se~), time 'sinc~'collection,· and storage temperature on progesterone concentrations in whole bloo~ collected fro~ 4 captive pregnant elk in December 1990 and May 1991~ '-Effects~ of blood. type, time, and temperature varied between months. In December, progeste~one was higher in serum than plasma (l_ 0.015) and higher in refrigerated samples (f.'_ 0.045), but not affected by time since collection (l_ 0.687). Plasma stored at ambient temperature provided the most stable values. In May, progesterone was not affected by blood type or temperature (l > 0.635) but markedly Lncreased in all samples, except; refrigerated serum, with time since colleetion (l _ 0.. 0001). Refrigerated serum provided the most stable values. In both months, progesterone did not decline with time since collection. These data suggest that progesterone values in samples collected by hunters would have a tendency to be higher, not lower, than actual progesterone levels. We continued to collect reproductive tracts from female elk harvested by hunters on the Forbes Trinchera Ranch. Pregnancy rates for adult cows (~ 1 yr old) were 61% and were the lowest recorded since 1986. This low rate was associated with low pregnancy rates in cows 1-3 years old which may reflect effects of low moisture during the last 2 years. Fetal sex ratios approached unity for the first time since 1986 and were 46 M: 54 F. One set of female twins was observed. Sizes of male and female fetuses increased from 1989. Conceptions occurred from 13 September.to 7 October, a 25 day interval, with a median date of 23 September. This was the shortest conception interval and earliest median date observed in this population. Brucellosis Was not detected in 45 blood samples from predominately female elk. Average age of bull eLk har{ested on_Forbes TriJlchera has st.~adily increased while ages of harVested'b~ck deer and female elk have remained,relatively stable. Average age of female deer harvested abruptly increased' in 1990' to 4.8 years. ��47 - EFFECT JOB PROGRESS REPORT OF ELK HARVEST SYSTEMS ON ELK BREEDING BIOLOGY David J. Freddy P. N. OBJECTIVE To evaluate effects of harvest systems on breeding biology of elk. _ SEGMENT OBJECTIVES 1. Continue to determine reproductive status of elk on the Forbes Trinchera Ranch using fetal collections and blood assays. 2. Continue to determine physical condition of elk and deer on the Forbes Trinchera Ranch by collecting body weights, antler weights and measurements, and ages of animals harvested. 3. Complete, by June 30, 1991, a Program Narrative (approved study plan) to assess reproductive, body weight, and survival responses of elk to large scale habitat improvements on transition and winter ranges. Study would begin in FY 91-92. METHODS Blood Assays - Pregnancy Testing We evaluated effects of time and ambient temperature on progesterone concentrations in plasma and serum collected from elk (Cervus elaphus). Using a syringe, we collected 60 cc of whole blood via the jugular vein from ~ yearold captive female elk on 26 December 1990 and again on 14 May 1991. In December, we collected blood from 4 pregnant and 1 non-pregnant elk and in May, we sampled the same 4 pregnant elk (about 35 days prior to term) and 1 different non-pregnant elk. We used xylazine (0.3-0.4 mgfkg) or medetomidine (0.05 mgfkg) to calm elk prior to collecting blood. Upon collection, blood from each elk was immediately divided into 20, 3.0 cc .subsamples and maintained at ambient temperature (13-16 C) during the 1 hr when blood was collected from all elk. These 20 subsamples were allocated as follows: 1) 10 were placed into heparinized glass tubes (plasma) with 5 each to be stored at 22-24 C and 2-4 C for up to 24 hours, and 2) 10 were placed into non-heparinized glass tubes (serum) with 5 each also stored at 22-24 C and 2-4 C for up to 24 hrs. At 1 hr post-collection (Time 0), 1 sample from each blood and temperature treatment· for each elk (4 samples/elk) was centrifuged, and plasma and serum were harvested and frozen at -18 C, and all remaining samples were placed in either a refrigerator or container maintained at ambient temperature. At 3, 6, 12, and 24 hours thereafter, samples were removed from their respective thermal environments, centrifuged, and plasma and serum were harvested and frozen at -18 C. _.Prostesterone ~0!lcent~~tions in each sample were ~~termined by radioi~unoassay on 7 JanuAry o~ 27 May 1991 by the Physiology Laboratory, Colorado. State University, Ft. Collins, CO. ~~Qeesterone concentrations transformed to log; �48 were analyzed using analysis of variance with repeated measures (serial blood samples for time intervals) (Proc GLM, SAS 1988). We also continued to monitor progesterone levels in ant1er1ess elk harvested by hunters in December, 1990, on the Forbes Trinch~ra Ranch. Hunters were instructed to obtain blood from the thoracic cavity of elk immediately after harvesting their elk and keep the non-heparinized vials of blood cool until depositing vials at ~heck stations. Blood was received from hunters usuallywithin a few hours of when animals were harvested and then refrigerated until centrifuging (usually within 12 hrs) at which time serum was frozen and stored at -18 C until processing. Hunters also collected the reproductive organs from their elk which allowed us to determine the pregnancy status of each elk. Blood ASSAYS - Disease Monitorin, Serum from elk harvested by hunters during December, 1990, on the Forbes Trinchera Ranch was tested for the presence of brucellosis in April, 1991, by the USDA Laboratory, Denver, CO. Serum was stored at -18C from centrifuging until tested for brucellosis. Fetal Collections Reproductive tracts of female elk were collected by hunters on the Forbes Trinchera Ranch from 8-17 December 1990. Tracts were deposited at check stations and kept cool until processing. Pregnancy status was determined from the presence of fetUses, embryos, and developed uterine tissue. Questionable uteri were preserved for later examination. Feta~ measurements w.ere made. on fresh ~pec::imens(subsequently pre~erVed) and .followed-definitions of·Aimstrong (1950). Fetal age was estimated from growth curves of-Morrison et a1. (1959). Inst~ctions showing hunters how to collect reproductive organs from elk wereimproved by providing better step-by-step illustrations: Body Measurements Eviscerated body weight, antler weight (including frontal bone), antler score (Nesbitt and Reneau 1986), and age were obtained for male elk and mule deer (Odocoileus hemionus) harVested on Forbes Trinchera Ranch. Eviscerated body weight, hind leg length, and age were also obtained for female elk and deer that were harvested. Age for all animals was estimate4 using both replacement and wear (Quimby and Gaab 1957, Robinette et al. 1957) and dental cementum (Keiss 1969, Stevens 1987). RESULTS AND DISCUSSION Blood Assays Progesterone-Temperature and Time Experiment Analysis of the effects of blood type, temperature, and time since collection on progesterone concentrations was limited to the 4 pregnant elk. In these animals, progesterone at time 0 hours ranged from 1.31 to 1.52 ng/ml in December and 2.15 to 2.29 ng/ml in May (Table 1). Progesterone in the nonpregnant elk was S 0.16 ng/ml and at these low concentrations effects of _~l_oodtyp~_, time, and _~emperature wer~_ undetectable. �49 Table 1. Average progesterone concentrations in blood collected from 4 captive pregnant elk in December 1990 and May 1991. Progesterone determined at 0, 3, 6, 12, and 24 hours post-collection for plasma and serum stored at 22-24 C or 2-4 C. . Time Since Plasma Progesterone (ng/ml) Serum Progesterone (ng/ml) Collection (hr) 22-24 C 2-4 C 22-24 C 2-4 C December o 3 6 12 24 o 3 6 12 24 1.31 1.23 1.32 1.40 1.30 1.49 1.40 1.68 1.66 1.57 1.52 1.65 1.47 1.47 1.50 1.48 1.61 1.77 1.84 1.88 2.15 2.12 1.99 3.12 2.66 2.29 2.22 1.97 3.08 2.92 2.22 2.25 2.16 3.18 2.72 2.26 2.53 2.40 2.60 2.49 December Samples .' Progesterone was higher in se~ tha~ plasma (~ ~ 0.015) and'higher in refri"gerated t~an ambient samples (f - 0.045). H<;Iwever,progesterone ~as not 'affected by time since collection (l - 0.687; Table 2, Fig. 1) .. In-general, steady declines in progesterone did not occur with any of the treatments, but~ a steady increase in concentration occurred in refrigerated serum and the most stable values were in ambient plasma (Fig. 1). At 12 hours, refrigerated serum. concentrations had increased 20% (0.30 ng/ml) over values at 0 hours. May Samples Progesterone increased markedly with time since collection (f - 0.0001) but was unaffected by blood type (l - 0.615) and temperature (l - 0.635). All samples, except refrigerated serum, decreased slightly through 6 hours then increased at 12 hours post-collettion (Fig. 1). Refrigerated serum provided the most consistent values increasing only 10% (0.25 ng/ml) over values at 0 hours. Unlike December, time since collection was involved with all significant effects (Table 2). December and May Samples Pooled Progesterone was higher in May than December and, overall, increased with time since collection (Tables I, 2). Effects of blood type, temperature, and time on progesterone. were highly dependent upon the month of collection (Table 2). �.IV Table 2. Results of univariate tests from analysis of variance of progesterone values in plasma and serum stored at either 22-24 C or 2-4 C and centrifuged at 0, 3, 6, 12, and 24 hours post-collection. Whole blood obtained from 4 captive preiDant elk in December, 1990, and May. 1991, Effect Tested DF Prob > F December Blood Type Temperature Time Blood Type X Temp Blood Type X Time Temp X Time Blood Type X Temp X Time 1 1 4 1 4 4 4 0.0149** 0,0450** 0,6868 0.9429 0.5169 0.0159** 0,1152 Blood Type Temperature Time Blood Type X Temp Blood Type X Time Temp X Time Blood Type X Temp X Time 1 0,6149 0.6349 0.0001** 0,1324 0.0248** 0.1215 0.0433** 1 4 1 4 4 4 December and May Pooled Month Blood Type Temperature Time Month X Blood Type Month X Temp Month X Time Blood Type X Temp Blood Type X Time Temp X Time Month X Blood Type X TempMonth X Blood Type X'Time Month X Temp X Time Blood Type X Temp X T~me Month X Blood Type X Temp X Time " 1 1 1 4 1 1 4 1 4 4 1 4 4 4 4 0.0061** 0.0814 0.1097 0.0019** 0.0100** 0.0077** 0.0583** 0.3957 0.0573** 0.6289 0.1519 0.4550 0.0042** 0.1331 0.0363** The inconsistency of our results may indicate an inadequate number of test animals, highly variable RIA analyses, or possibly the need to validate RIA analyses specifically for progesterone in elk. Our data suggest that progesterone in either plasma or serum has a greater tendency to increase than decrease in concentration with time since collection. These results contrast with observed decreases in both plasma and serum progesterone in muskox (Rowell and Flood 1987). Our data also suggest that it is prudent to consistently control the storage conditions for whole blood to reduce var~ationJn progesterone assay~ and that yalues d~rived from plasma and serum should not be treated as equiva~ent data. �51 Our low threshold value of 1.8 ng/ml for pregnant elk in December/January (Freddy 1990) does not appear to be low because progesterone decayed in samples collected by hunters. We were concerned that samples obtai~ed by hunters might be misleading because of the potential time delay between obtaining whole blood from the harvested elk and centrifuging the blood. Progesterone-Forbes Elk Average serum progesterone concentrations in known pregnant elk have been ~ 2.13 ng/m1 and for known nonpregnant elk, S 0.34 ng/ml in all years since 1987 (Table 3). These data will be used to refine our predictive curve for determining pregnancy status of elk (Freddy 1990). Table 3. Serum progesterone (ng/ml) in known pregnant and nonpregnant adult elk (~ 1 year old) harvested by hunters on Forbes Trinchera Ranch 1987-1990. ~erum ~rogesterone Pregnant Year/Statistic (ngLml) Non~regnant llll Mean SD SE CV% n 3.53 2.42 0.41 68.5 35 0.30 0.26 0.09 86.0 8 2.79 1.38 0.21 49.6 -42 0.29 0.59 0.17 201.0 12 4.40 2.97 0.53 67.4 31 --0.34 0.27 0.11 80.1 6 ill.§. Mean SD SE CV% n1989 -_Mean SD SE CV% n 1990 Mean SD SE CV% 2.13 0.10 0.19 46.9 27 n Disease Surveys-Forbes 0.14 0.22 -0.05 155.7 16 Elk All serum samples collected during December, 1990, from elk on the Forbes Trinchera Ranch tested negative for brucellosis. The 45 samples were from 1 male calf, 7 female calves, and 37 adult females. Previous tests for brucellosis on serum collected in 1987 and 1988 were also negative (Freddy 1989). Fetal Collections-Forbes Elk Hunters provided 52 ~eproductive tracks from 54 female elk (including calves) harves ted.---- �52 Precnanc;y Rates Pregnancy rates in 1990 for adult cows (~ 1 yr old) were 61% and were the lowest recorded since 1986 (Table 4). This low rate was associated with pregnancy rates of 0, 50, and 67% in yearlings, 2-year olds, and 3-year olds, respectively. For all years, pregnancy rates for prime-aged adults (3-10 yr old) were about 90% but were 15% for yearlings and < 60% for cows aged ~ 11 years (Table 4). Pregnancy in calves (age 6 mos) was not observed. Table 4. Pr.snancy 9 6-90 r.t.. for as. cl..... 1986 4&'':IE.1 1987 la-, D f,·. D of famal •• lk coll.cted • 1988 1989 fE·. l:E!1 D D 2 7 6 3 4 1 2 3 7 0 S S lS 15 10 9 5 4 3 3 1 0 3 2 7 2 6 4 19 17 12 11 7 8 2 2 0 0 7 4 S 3 12· 12 5 5 3 4 Tot..l. 34 28 49 38 61 47 % Pr., 82 78 77 1 2 3-4 5-6 7-8 9-10 11-12 13-19 6 7 6 3 4 1 2 5 "4&. for ~2 frOID r.pl.c_ent and •• ~; on the Forb.. Trinch.r. 1990 D fE·. D 0 3 6 3 8 4 1 3 34 27 61 42 30 15 7 23 49 38 46 28 239 78 61 1 2 10 11 5 5 2 2 9 6 9 5 8 4 1 4 1986-90 fre. Ranch in D.cemb.r %EIi:·. S 21 54 37 28 15 5· 14 15 78 89 88 93 100 71 61 179 75 75 for 3+, frOID dent.l c_entUID. Litter size was 1 in all but 4 (2%) of 179 litters. The 4 sets of twins were: 2 females and 1 male and 1 female in cows 9 and 12 years old, respectively in 1986; 2 ff!ma1e~ in a 6-year-01d cow in 1989; and, 2 females in a 4 year. old ~ow in 1990. infected uteri not ~apable of supporting pregnancy occu~red in 5 (2%) of 239 adu1t:s examined. Infec:;tionswe're in ;I., yearling and 4 cows ~ .i3 years old with 3 found in 1986, 1 iri 1988, and 1 in 1989. Fetal Sex Ratios In 1990, the fetal sex ratio was 46 M:54 F and the ratio pooled among all years was 49 M:50 F (Table 5). Fetal sex ratios approached unity in 1990 for the first time since 1986 when collections began, and this occurred with the lowest. observed pregnancy rate (Tables 4, 5). Tabl. S. 9 -90 4&''lAl~ 1 2 3-4 5-6 7-8 9-10 11-12 13-19 Tot.la F.t.l .ex r.t.io. for ),9!W .s. cl..... 1987 H E 1l 1 1 4 3 3 0 2 3 17 0 2 1 0 0 2 1 0 6 1 4 1 0 1 0 0 0 7 of famal. .lk coll.ct.d 19§9 12§8 E !l H E !l H 0 0 3 2 3 11 2 7 2 2 1 1 0 0 1 1 12 24 0 0 1 .1 2 8 8 S 1 0 1 26 0 2 S 3 2 1 0 1 0 2 0 0 0 0 0 3 1 0 4 4 0 0 1 H o . 0 1 0 0 2 2 15 E !l H 0 0 0 0 1 1 1 0 0 3 0 1 3 0 6 2 2 6 7 4 4 1 2 0 12 24 0 1 13 on th. Forb.. Trinch.ra 1990 F 0 2 4 3 2 1 1 2 is 1986-90 !l H l 0 3 7 22 17 16 4 3 8 80 0 10 27 20 10 9 3 5 84 0 0 0 0 1 0 0 1 30 38 44 74 63 % H!la ~3 .' "4&a for ~2 frOID r.plac_ant. and .aar: "Fat.l .ax H-mala F-f-.1. U-unlmown. - 39 29 33 46 for 3+, frOID dent.l c_antUID. % !l Tot.l Fatu ••• Ranch in D.cemb.r 180 49 2 4 4 1 2 3 0 0 16 Hal· 100 41 45 46 62 31 SO 62 49 �, 53 Fetal Size Body weight, crown-rump length, and hind foot length for male and female fetuses were larger than in any previous year except for male fetuses in 1987 (Table 6). These larger fetuses occurred during what was considered to be the second consecutive year of below average moisture. Table 6, Year/ Statistic HeasuIements of e~ IIS!!!I Weisbt ~151 Ma.e [ema.e fetuses from Forbes Trinchera Ranch, Crown-Rump Bind Foot Lenl5th ~aml ~aml l£enl5th Hale Femal. [ema •• Hlle 1986-1990, 29 Nov-5 Dec 1986 Hean SO lllin III!X CV% n 20,9 7.0 11.9 29.0 33.5 6.0 90,2 16.7 63.5 115.5 18,5 17,0 86,0 9.5 72.0 94.0 11.1 5.0 60.4 33.2 12.0 135.0 55,0 24.0 141.4 21.1 113.0 180.0 14.9 12.0 66.7 ~9.7 23.0 120.0 44.5 15.0 60.0 45.4 4.9 149.0 27,6 13.8 12.2 58.5 50.0 17.0 20.5 4.8 . 13.5 28.5 23.4 17.0 19.8 2.7 16.5 23.0 13.6 5.0 115.4 23.3 66.0 151.0 20,2 24.0 42.3 10.9 31.0 68.0 25.8 12.0 30.3 8.8 14.0 46.0 29.0 24.0 120.0 19.9 77,0 163.0 16.6 26.0 121.7 17.9 87,.0 150.0 14,·7 15.0 32.6 7.7 17,0 47.0 23.6 26,0 33.1 7.0 20.0 45.0 21.2 15,0 24.0 102.4 37.5 35.0 169.0 36.6 12.0 111.5 33.4 38.0 163.0 30,0 24.0 27.9 12.3 13.0 53.0 44.1 11.0 29.9 10.8 10.0 48.0 36.1 23.0 78,4 43.0 31.0 166.0 54.9 14,0 128.0 25.1 84.0 170.0 19.6 13.0 127.6 22.6 95.• 0 161.0 17,7 15,0 35.1 9,7 19,0 51.0 27.7 13,0 35,7 10.5 22.0 58.0 29.5 15.0 12-14 & 19-21 Dec 1987 Heail SD lllin III!X CV% n 118.7 54.7 66.0 242.0 46.1 12.0 10-12 & 17-1·9Dec 1988 Hean SO IlliD III!X CV% D 74.5 35.2 22.0 162.0' 47.3 26,0 ;.- 9-11 & 16-18 Dec 1989 Hean SD IlliD III!X CV% D ·57.8 59.9 3.5 188.0 103·.6 12.0 7S.7 8-17 Dec 1990 Hean SD IlliD III!X CV% n 92.2 52.5 26.0 190.0 56.9 13,0 Concention Dates Conceptions occurred from 13 September to 7 October, a 25-day interval, with a median conception date of 23 September. This is the shortest conception interval and the earliest median date for breeding among all years (Fig. 2). One hypothesis that might explain this compressed'breeding period is that breeding was more restricted, than in previous years, to those prime-aged.cows that achieved adequate body condition to breed during a prolonged period of drought. Larger and relatively less variable sized fetuses may also reflect -_fewer fetuses fr~m young' cows (Table 4)." �forbes Elk and Deer Harvests Average ages of buck deer and female elk harvested from 1986 to 1990 have remained relatively stable with generally increasing, but variable, harvests (Table 7). Average age of bull elk harvested has steadily increased along with increasing harvests, and this most likely reflects greater selectivity for and improved efficiency in finding older bulls for harvest (Table 7). Average age of female deer harvested remained relatively stable from 1986-1989 but abrUptly increased in 1990 (Table 7). This increase may reflect poor recruitment of young deer in 1989. Table 7. Average age of adult (~ 1 year old) deer and elk harvested on the Forbes Trinebera Ranch, 1986-90, Species-Season/ Statistic· 1986 1987 1988 1989 1990 4.8 0.2 72 4.9 0.2 93 4.8 0.2 87 5.2 0.2 102 5.7 0.2 117 5.6 0.2 87 6.6 0.2 109 6.6 0.2 lOS 6.0 0.2 11S 6,0 0.2 lOS ·Bull Elk - PrivateMean SE. n Buck Deer - Privateb Mean SE n Female Elk Publ~cc ~ 5,-5 O.S 34 -Mean SE n 5.2 0.6 51 5.4 0.5 64 5.7 0.6 51 - 5.4 -0,6 46 Female Deer - Publicd Mean SE_ n -Age bAge CAge dAge 3.S 3.5 3.5 3.6 4.S 0.3 0.2 0.2 0.2 0:2 .50 . 113 115 153 134 based on tooth replacement and wear. based on tooth replacement and wear. for 1-2 from replacement and wear; for 3+, from dental cementum. for 1 from replacement and wear; for 2+, from dental cementum. Eviscerated body weight, antler weight, and antler score were obtained for nearly all male elk and deer harvested. Eviscerated body weight and hind leg length were also obtained for nearly all female elk and deer harvested. These data have not as yet been completely summarized. Program Narrative A program narrative regarding elk survival rates was not completed this segment because long-term research needs for elk were being evaluated. z; �55 LITERATURE CITED Armstrong, R. A. 1950. Fetal development of northern white-tailed Amer. Mid. Nat. 43:650-666. deer. Freddy, D.~. 1989. Effect of elk harvest systems on elk breeding biology. Colo. Div. of Wildl. Game Res. Rep. July:35-60. 1990. Effect of elk harvest systems on elk breeding biology. Colo. Div. of Wildl. Game Res. Rep. July:19-29. Keiss, R. E. 1969. Comparison of eruption-wear patterns and cementum annuli as age criteria in elk. J. Wildl. Manage. 33:175-180, Morrison, J. A., C. E. Trainer, and P. L. Wright. 1959. Breeding seasons in elk as determined from known-age embryos. J. Wildl. Manage. 23:27-34 . .Nesbitt, W. H., ana J. Reneau (eds.). 1986. Boone and Crockett Club's 19th big game awards. Boone and Crockett Club, Dumfries, Vermont. Quimby, D. C., and J. E. Gaab. 1957. Mandibular dentition as an age indicator in Rocky Mountain elk. J. Wildl. Manage. 21:134-153. Robinette, W. L., D. L. Jones, G. Rogers, and J. S. Gashwei1er. 1957. Notes on tooth development and wear for Rocky Mountain mule deer. J. Wi1dl. Manage. 21:134-153. 'Rowell, J..,-and P. F. Flood. 1987.. Changes in muskox 'blood progesterone' concentration between' collection and 'ce·ntrifugation. J. Wildl. Manage. =-51:901-903. SAS Institute, Inc. 1028pp. 1988. SAS/user's guide. SAS Inst. Inc., Cary. N. C. Stevens, M ..L. 1987. Apparent accuracy of cementum annuli for estimating ages of mule deer. M. S. Thesis, Colorado State University, Fort Collins. Prepared by -4~"""~WW;id'_J""'~-"'1,,--~_"'~,-=·:!f;=..I.~..o:;.;:~-· _ Wildlife Researcher �56 1.4 . PLASMA 22-24 C SERUM 22-24 C ----.---. 1.2 -t--------------PLASMA 2-4 C SERUM 2-4 C o - ---fi--· c::- E 1 C) .s ...-- w z o a: ...8-- _ ---------------0 MAY w til w ~ ~ .--~ . 06 . -+--------------~--------------~~--~~-~--~--~---------l c.. ~. C( G) 3 0.4 0.2 DECEMBER o o 3 6 12 24 HOURS POST-COLLECTION Fig. 1. Average progesterone concentrations (log. ng/ml) in plasma and serum collected in December 1990 and May 1991 from 4 captive elk. Whole blood samples stored at 22-24 C or 2-4 C and centrifuged at 0, 3, 6, 12, and 24 hours post-collection. �57 so :0 :0 :0 I I :! I U) z 0 i= ~ 40: e, 0 0 :- 30 !Z w : : 0 a: w e, - 1986 1987 .......•.••..__ .. ~ ®1~~411988 ------_ .•. w Z 0 YEARLY MEDIAN DATES n =26 n =37 n =44 0 1989 n =35 I22J 1990 -- n =27 20 10 I I o 9/8 In ~ ~ 9/18 9/28 10/8 II 10/18 B T 10/28 11/7 11/17 MONTH AND DAY (BEGIN 5-DAY INTERVALS) ~------------------------------------------~ ~ •. so _ - MEDIAN DATE ~-----------~------------------~----------~~ ~ 40 -1------- ~ o z o o !Z w ~ w a, _ 1986-1990 n = 169 30 -4------- 20 -4-.-----~----- o 9/8 9/18 9/28 10/8 10/18 10/28 11/7 11/17 MONTH AND DAY (BEGIN 5-DAY INTERVALS) Fig. 2. Conception dates for elk on the Forbes Trinchera Ranch, 1986-1990. ��59 Colorado Division of Wildlife Wildlife Research Report July 1991 JOB PROGRESS REPORT State of ~C~o~l~o~r~a~d~o _ Project No. W-153-R-4 Work Plan No. 3 Job No. Mammals Research Elk Investigations Elk Census Methodology Period Covered: Author: July I, 1990 - June 30, 1991 D. J. Freddy Personnel: R. Bartmann, G. Byrne, G. White Abstract .. Estimates of population sizes for elk and deer in the Troublesome sub-unit of ~iddle Park were obtained using systematically spaced aerial line transects. Additionally, an estimate~of popuiation size for deer was obtained using a random quadrat -(2.59 km2) sY$tem .. All flights were cortducted in Jan~ry, 1990,- and March, 1991, using a Bell-Soloy helicopter. The exponentialpolynomial (EXPL) sighting model provided the best combination of precision and model fit for both elk and deer. Precision (± 95% CI) ranged from ± 4350% and 13-38% for elk and deer, respectively. Estimates of population size varied with each sighting model and ranged from 2,364-4,263 for elk and from 3,908-7,496 for deer. Line transects provided higher estimates of total deer and better precision than quadrats in both years (l< 0:10). Differences between technique~ were greater than expected with quadrats estimating only 26-37% of the deer estimated by line transects. High estimates prov~ded by line transects were related to the-high proportion of groups of elk or deer seen ~i_thin-. the,center line interval compared to intervals away from the center line." A video .camera ,system with further. improyeni~nt~_may. provide a possible means.'to docunient groups along the center line interval. ± ��6l JOB PllOGllESS UPOB.T EIJC CENSUS METHODOLOGY David J. Freddy P. N. OBJECTIVE Evaluate methods to estimate numbers of elk during winter. SEGMENT OBJECTIVES 1. Determine efficacy of line transect methodology to estimate numbers of elk and mule deer sharing sagebrush-aspen-conifer winter ranges. Specifically, attempt to ascertain apparent problems with assigning observed groups to the ·center line" interval by using a video camera. 2. Compare estimates of mule deer density based on Lf.ne transects and quadrats in an area where both methods can be emp~oyed. 3. Prepare manuscript for publication. INTRODUCTION Methods to reliably esti~te numbers of ~lk (Cervus elap1;1us)have yet to be -developed. Bear et al. (1989):evaluated mark-res~ghting-to estimate population size, and they -found the method: provided-precise estimatj!s of popUlation size that were likely negatively biased; was sensitive to-colors of markers placed on elk; and was costly. Samuel et a1. (1987) and Unsworth (1990) developed and used sightability factors to correct for negatively biased counts of elk during surveys attempting to completely count all elk. This approach has merit although not having a fixed sampling design to determine those areas searched may lead to yearly variability in estimates due to changes in observers and patterns of search. These methods for elk are all subject to underestimating true population size as has been found in many efforts to estimate true densities of ungulates (LeResche and Rausch 1974; Bartmann et al. 1986; Pollock and Kendall 1987; Samuel et al. 1987). Underestimates occur because all animals are not observed on defined sample units such as quadrats or strip-transects. Line transect theory allows observers to miss animals, except those occurring directly on the center line of the transect and, therefore, could provide less negatively biased estimates of population size (Burnham et al. 1980; White et al. 1989). We chose to assess the field logistics and precision of aerial line transects as a sampling method to estimate numbers of elk. Secondarily, we also wanted to further evaluate line transects for estimating numbers of mule deer (Odocoileus hemionus) (White et al. 1989). The area selected for evaluating line transects was the Troublesome portion of Game Management Unit (GMU) 18 within Middle Park, northcentral Colorado. During winter, significant numbers of elk and mule deer concentrate into a 250 Iqn2 area.which can_be efficiently sampled with aerial surveys, Vegetatio~ ---within winter -ranges frequented by elk is primarily sagebrush (Artemisia tridentata), aspen (Populus tremuloides), conifer forests (Pinus conCorta, �62 Pseudotsugs menziesii, Pice. engelmanni), and mixed stands of aspen-conifer. For deer, winter ranges are dominated by sagebrush and aspen. from rolling hills to steep canyons. Terrain varied METHODS Two sets of parallel line transects, each having 26 transects systematically_ spaced at 1,000-m intervals, were flown to estimate numbers of elk and deer. These 2 sets resulted in transects occurring every 500 m across the sampled winter range. Transects were oriented on true north-south bearings perpendicular to changes in elevation and expected gradients in elk and deer densities (White et al. 1989), were delineated on 1:24,000 scale topographic maps, and were not marked with flight markers on the ground. Placement of transects allowed for counting elk and deer simultaneously. Line transects flown to estimate densities of elk totaled 948 km (588 mi) in length, were 1-15 km long, and were distributed within an area of 246 km2 (95 mi2) between elevations of 2,257-2,990 m (7-,400-9,800 ft). Transects were limited to those areas where elk generally reside during January and February. The same line transects were flown to estimate densities of deer but only deer found at elevations S 2,590 m (8,500 ft, primarily sagebrush habitat) were counted. With this restriction, line transects for deer totaled 602 km (374 mi) in length, were 1-14 km long, and were distributed within an area of 166 ~ (64 mi2) , which was also the area delineated for a quadrat (2.59 km2, 1 2 mi ) sampling system used to estimate numbers of deer in the Troublesome area during the last 22 yrs rcrn -1969).Trans~cts were conducted with a Bell-Soloy helicopter flown at 65-80 kmph (40: 50 mph) and 35-50 m above the ground or tree canopy. We flew each set of transects twice (4 replicate flights) and attempted to fly each set once in the morning and once in the afternoon but weather altered this scheduling of transects. Transects were flown in 1990 on 9-and 11 January and in 1991 on 1, 3, 4, and 8 March. A navigator and observer were responsible for observing elk or deer. The observer, seated on the right, estimated perpendicular distances to and counted those groups of elk or deer located from the transect center line to the right. The navigator, seated in the middle, maintained course bearing, and aided in locating groups on or near the transect line. Observers were experienced in aerial counts of elk and deer and practiced estimating distance intervals from the helicopter using markers placed at 10-m intervals along a practice transect line. Primary observers were different each year. We considered sets of transects to be independent and thus pooled results from the 4 replicate flights to derive estimates of elk and deer populations. We had no indication that counting elk or deer disturbed either sufficiently to cause them to move from 1 transect to another while flying one set of transects and thus we feel duplicated counts within a set of transects did not occur. Estimates of population size derived from line transects followed methods outlined by White et a1. (1989) and estimates derived from quadrats followed methods of Gill (1969) and Mendenhall et al. (1971). We _wanted--to compare estimates of population siZe -for deer derived from -lrite transects and quadrats. We, therefore, expanded an existing quadrat system �63 from 10 to 30 quadrats by drawing at random an additional 20 quadrats to improve precision of the estimate. Quadrats were flown on 10 January 1990 and 3-4 March 1991 using the same helicopter and pilot used each year for line transects. Quadrats and transects were flown at similar flight speeds and elevations above the ground. At least 1 corner of each quadrat was marked on the ground to aid in locating quadrats from the helicopter. The navigator and observer both searched for and counted deer. Estimates of population density were compared using a z-test. RESULTS AND DISCUSSION . Although we searched for deer and elk simultaneously on line transects, the 2 species were seldom seen in close proximity. Ye, therefore, believe that searching for both species had little effect on detecting either species. Ye found little evidence to suggest elk were moving prior to detection. Elk were often bedded or standing in aspen or conifer habitats where snow depths were 25-60 cm. However, deer moved in response to the helicopter more than did elk because snow depths did not deter deer from moving. In 1990, snow cover was nearly 100% but snow depth was only 4 cm at lower elevations while in 1991, many south-facing slopes were free of snow. Even with the mild snow conditions both years, deer were almost exclusively within the sampled area demarcated for deer. There were 2 and 1 group(s) of deer in 1990 and 1991, respectively, that were seen along line transects outside the sample area demarcated for deer quadrats and line transects. Line Transects-Elk . Estimated population size for elk varied with which· sighting probability- model was used. In 1990, estimated size ranged from 3,404 to 4,263 (Table 1). A minimum number of elk in the sampled area based on helicopter flights was 419 in January 1990 and 744 in January 1~88 (pers. ·comm. R. Thompson). Our estimates in 1991 were more variable and of lower precision because fewer groups of elk were encountered on transects. In 1991, we found elk were widely distributed outside our defined sample area because low snowfall had allowed elk to reside at higher elevations~ Ye, therefore, discontinued flights for elk after completing one replicate of each set of transects (Table 4). The EXPL models provided the best precision during both years which was ± 43% and ± 50% of the mean (95% CI, Table 1). Sighting curves were spiked at the center line both years indicating that probability of detecting elk away from the center line fell rapidly and effectively ended at 95 m (Fig. 1). This spiked distribution was largely responsible for the unexpectedly high estimates of population size. Enlarging or reducing the width of the center interval or the cut-point for strip width generally resulted in models that fit the data poorly. All models fit reasonably well with about equal quality (good fit - Chi-square l> 0.10; Table 1). �Table 1. Population estimates for elk based upon different sighting probability models for line transects, Troublesome sub-unit, Middle Park, January 1990 and March 1991, Pop. Size & 95% CI Density Center Interval/ Groups Chi-square Prob, (P) Elk/mi2 Cut Point (m) Mean Upper Lower Modela n 0-15 0-15 0-15 0-15 0-15 / / / / / 155 95 95 95 95 66 57 57 57 57 0.8252 0.7310 0.8248 0.7514 0.7309 3659 3664 3662 4263 3404 5232 5224 5380 9813 5053 2086 2104 1943 o 1754 38,5 38.6 38.6 44.9 35.a 24,9 4738 2364 EXPLb 0-15 / 155 27 0.6397 o 3247 6136 0.6470 358 34.2 EXPL 0-15 / 95 23 6194 3247 34.2 0.8163 302 NEXP 0-15 / 95 23 4163 12,506 43.8 0.8010 o EXPS 0-15 / 95 23 0,6470 2743 5184 301 28.9 EXfLc 0-15 / 95 23 aModels are: EXPL-exponential polynomial, NEXP-negative exponential, EXPS-exponentia1 power series. ~ese models provided the best precision: ±43% of the population mean in 1990; ±50% of the population mean in 1991. C!n these models, average group size was.based only upon the groups seen •within the cut-point truncat~on distanc~; other models used all groups seen to .calculate average group's ize .. Average number of elk per group tended to increase with distance away-from the ceriter interval (X - 0.30, slope positive). This was pronounced in groups observed at ~ 95 m which suggests a bias in detecting large groups beyond this distance (Fig. 2), In calculating population size, program TRANSECT uses all groups observed to calculate an average group size, regardless of the cutpoint for the sighting curve (White et al. 1989). We altered the program so that only groups seen within the truncation distance were used to calculate average group size which reduced estimates of population size for the EXPL . model 7-16% in 1990 and 1991 (Table 1). Groups were generally larger and more frequently observed in aspen and sagebrush than in conifer habitats but differences in average group size among major habitat types were not significant (ANOVA, E> 0.50, Fig. 3). Line Transects-Deer Estimated population size for deer also varied with which sighting probability model was used. In 1990, estimated size ranged from 5,021 to 9,541 and in 1991 from 3,908 to 5,158 (Table 2). The EXPL models provided the best precision during both years which was ± 13% and ± 38% of the mean (95% CI, Table 2). Precision declined in 1991 because average group size increased (6.08 to 7.55) and was more variable and total groups observed decreased (208 to 132). These changes reflected the more concentrated distribution of deer in 1991. Sighting models for deer were more variable in their goodness-of-fit than for elk, and in 1990 especially, models often fit poorly (Chi-square ... l < O._lO;_._Table2)'--· �65 Table 2. Population estimates for deer based up~n different sighting probability models for line transects, Troublesome sub-unit, Middle Park, January 1990. Pop. Size & 95% Cl Density Center Interva1/ Groups Chi-square Model· Cut Point (m) n Prob. (P) Mean Upper Lower Deer/mi2 1990 EXPL EXPLb NEXP EXPS EXPLbc 0-15 0-15 0-15 0-15 0-15 / 155 / 95 / 95 / 95 / 95 185 139 139 139 139 0.0001 0.0908 0.1420 0.2024 0.0908 5021 7496 7494 9541 7005 7144 8466 10,538 17,327 7992 2898 6526 4451 1746 6018 78.4 116.9 117.0 149.0 109.3 ll2l EXPLb EXPL NEXP EXPS EXfLc 0-15 / 155 123 0.1797 4483 6190 2776 69.9 0-15 / 95 101 0.1877 5158 8177 2139 80.5 0-15 / 95 101 0.2711 5156 7314 2998 80.4 0-15/95 101 0.1964 5771 10,783 757 90.0 0-15 I 95 101 0.1872 3908 6207 1610 61.0 -Models are: -EXP-L- exponential polynomial, NEXP - negative exponential, EXPS - exponential power series. bThese models provided the best precision: ±13% of population mean in 1990; ± 38% in 1991. cIn these models, average group size was based only upon the groups seen within the cut-point truncation distance; other models used all groups seen to calculate averag~ group -size. . As with elk, sighting curves were spiked at -the center line indicating that probability of detecting deer away from the center line fell rapidly and effectively ended at 95 m (Fig. 2). This spiked sighting curve was observed both years even ~hough primary observers were different each year. Similar sighting distributions for deer occurred with other observers and habitats, suggesting that these sighting distributions are a function of technique and not observer (Freddy 1990). As with elk, _this spiked sighting curve was largely responsible for the unexpectedly high estimates of population size. Average number of deer per group tended to increase with distance away from the center interval (X - 0.45, slope positive) and was again pronounced in groups observed at ~ 95 m (Fig. 4). Using only groups within the truncation distance to calculate average group size, reduced estimates of population size 7-24% in 1990 and 1991 (Table 2). Quadrats-Deer Estimated population size for deer was 1,848 in 1990 and 1,647 in 1991 (Table 3). Precision declined from ± 32% to ± 53% of the mean (95% CI) from 1990 to 1991. Deer were more concentrated in 1991 which resulted in more variablity in numbers of deer per quadrat. Estimates of population size were not different between years (~0.50). . �66 Table 3~ Estimated size of the mule deer population in the Troublesome subunit of Middle Park, January, 1990 and March, 1991 based upon deer counted on square-mile quadrats. Variances calculated using finite population correction factor and N - 64.1 mi2. 95% C.l, Precision Density Quadrats Deer/mi2 Lower Upper % of Mean n Mean Year 1848 1264 2431 28.8 1990 30 ± 32% 1647 25.7 1991 30 771 2524 + 53% Line Transects vs. Quadrats-Deer Line transects provided a higher estimate of total deer than did quadrats in 1990 (l< 0.001) and 1991 (l< 0.10) and generally provided better precision (Tables 2, 3). Population estimates derived from quadrats in 1990 and 1991 represented only 26 to 37%, respectively, of the deer estimated by the most precise EXPL models for line transects (Tables 2, 3; Fig. 5). We expected quadrats to underestimate true population size because area based sampling strategies assume 100% of the animals present are counted and this seldom occurs. In juniper-pinyon (Juniperus oseeosperma-Pinus edulis) habitats, observers detected only 65% of the known deer on quadrats (Bartmann et al. 1986).· In open sagebrush habitats frequented by our deer, we might detect 80% of the deer under optimal conditions (LeResche and Rausch 1974), and if true, the corrected estimated population of deer based on quadrats would be 2,310 and 2,050 in 1990 and 1991, respectively. Given a worst case scenario of detecting only 65% of the deer on quadrats, we would expect a true population of 2,843 and_2,534 in these-same years, respectively. Because line transects ~a.ve been found .to estimate 90% .of the true population ("Whiteet a1.,·-1989)we ·might_have-reasoned that line transects wo~ld provide an ~stimate of about 3,OOO-de~r instead of an est1mat~d 4,000-7,000. The data leave us in-doubt as to which method is most biased. Fiyin, Effort Actual flying time used to count elk and deer along line transects was consistent among replicated flights in 1990 (Table 4), In 1991, flying time was also consistent when transects were flown for elk and deer (replicates 1) and for deer only (replicates 2) (Table 4). Groups detected per replicate were somewhat consistent; although in 1990, flights conducted in the morning always detected more groups of either elk or deer. Line transects for deer accounted for about 65% of the total flying time devoted to transects that sampled both deer and elk. Compared to quadrats, line transects for deer used 1.3-l.8x the amount of flying time to sample the same geographic area. The amount of flying time devoted to replicating transects was dictated by our objective to detect 130 groups of elk and 200 groups of deer to achieve precision approaching ± 15% ("Whiteet al, 1989). In general, we did not achieve this desired precision suggesting that densities of groups were insufficient. �67 Table 4. Numbers of groups of 'elk and deer detected on transects and estimates of counting times for transects and quadrats flown in the Troublesome sub-unit, Middle Park, January, 1990 and March 1991. Groups Detected· Elk Deer Right Right Flight Method Countingb Time (mins) l2.2Q Trans. Trans. Trans. Trans. Set Set Set Set 1 2 1 2 Rep Rep Rep Rep 1 1 2 2 AM PM AM PM Trans. Totals 24 22 60 44 60 16 44 229 227 229 220 79 208' 90S 17 445 Quadrats (30) for Deer Only illl Trans. Trans. Trans. Trans. Set Set Set Set 1 Rep 2 Rep 1 Rep 2 Rep 1 1 2 2 PM AM AM PM Trans. Totals 21 12 n/a n/a 33 29 257 262 46 27 30 174 132 855 162 366 Quadrats (30) for Deer Only -Groups of elk and deer to the right of center line. . botime shown is the best estimate of flying time used to count anima~s. . ,. Flying. time to/frC?m for refueling not included. . Evaluation of Video Camera System We explored the potential to use a video camera system to document the accuracy of assigning groups of deer or elk to the center line interval of line transects. We mounted a Canon CI-20 miniature video camera (Sx8x10 cm) along the center console of a Bell-Soloy helicopter. The. camera was'wired to a Sony Video Walkman 8 mm recorder using Sony metallic particle color tape. Both wide-angle and telephoto lenses, ranging from 4.8 mm to 2S mm focal length, were used. To determine angle and width of coverage, we flew test flights down an airport runway where runway markers could be used to . accurately determine camera ~overage. All lenses provided coverage to either side of a hypothetical center line with the 16 mm providing the most acceptable coverage. We then flew both sagebrush and juniper-pinyon habitats where deer were present. Although deer were detectable on the video tape, primarily in open habitats, they generally could not be seen until the helicopter was close, thereby reducing the ability to detect groups at reasonable distances in front of the aircraft. Aircraft vibration, slow electronic shutter speeds, and rapid changes in exposure contributed to reduced quality of the video image. A better aircraft mounting system combined with newer cameras having faster shutter speeds and higher resolution may allow the camera to provide an independent estimate of groups within the center interval. �68 CONCLUSIONS Sighting distributions for elk and deer were spiked at the transect line during both 1990 and 1991. This characteristic of line transect data collected from a helicopter appears to be independent of species, observer, or habitat type. For Middle Park, line transects produced surprisingly high estimates of popUlation sizes for elk and deer, and this is attributed to the high proportion of groups observed in center interval of the transect. Estimates of deer population size derived from transects were significantly higher than estimates based on quadrats and were more precise. We expected differences but not to the degree observed and, therefore, are left with the difficult objective of determining which method is most biased. We found that intense efforts by the pilot, navigator, and observer were required to effectively fly line transects. Whether this effort is more or less intensive than quadrats, is open to debate; but, subjectively, line transects were more likely to increase observer fatigue. At the suspected moderate densities of elk and deer on our study area, we had difficulty acquiring the required numbers of groups to achieve desired precision for line transect estimates even though we replicated transects. Our data suggest that transects may be more costly to fly than quadrats depending upon desired precision. LITEBATQJE CITED Ba~tmann, R. M., L. H. Carpenter, R. A. Garrott. and D. C. Bo~den. 1986. A~curacy of helicopter counts of mule"~eer in pinyon-juniper woodland. vriai. Soc. Bull. 14:356-363. Bear, G. D., G. C. White, L. H. Carpenter, R. B. Gill, and D. J. Essex. Evaluation of aerial mark-resighting estimates of elk populations. Wildl. Manage. 53: 908-915. 1989-. J. Burnham, K. P., D. R. Anderson, and J. L. Laake. 1980. Estimation of density from line transect sampling of biological populations. W~ldl. Mono. 72. 202pp. Freddy, D. J. 1988. Elk census methodology. Rep. July (1):78-82. __ '. 1990. Elk census methodology. July: 31-43. Colo. Div. Wildl. Game Res. Colo. Div. Wildl. Game Res. Rep. Gill, R. B. 1969. A quadrat count system for estimating game population. Colo. Game, Fish, & Parks Game Info. Leaflet 76. 2pp. LeResche, R. E., and R. A. Rausch. 1974. Accuracy and precision of aerial .moose censusing. J. Wildi. Manage. 38:175-182. Mendenhall, Y., L. Ott, and R. L. Scheaffer. 1971. Elementary survey. sampling'. DUxbury Press, Belmont, Calif. 247pp . Pollock, K. H., and W. L. Kendall. 1987. Visibility bias in aerial surveys: a-review of estimation procequres. J. Yildl. Manage. 51:502-510. �69 Samuel, M. D., E. O. Garton, M. W. Schlegel, and R. G. Carson. 1987. Visibility bias during aerial surveys of elk in northcentral Idaho. Wildl. Manage. 51:622-630. J. Unsworth, J. W., L. Kuck, and E. O. Garton. 1990. Elk sightability model validation at the National Bison Range, Montana. Wildl. Soc. Bull. 18:113-115. White, G. C., R. M. Bartmann,' L. H. Carpenter, and R. A. Garrott~ 1989. Evaluation of aerial line transects for estimating mule deer densities. J. Wildl. Manage. 53:625-635. �IU 20 ELK TROUBLESOME ff:::l UNIT 15 o a:: e ~ JAN 1990 n=66 It:::t/:J MAR 1991 n=27 u, o a:: w m ::E :::l 10 Z 5 0 .•..• U) I <:) U) t\I ..It) U) (') U) ~ I U) U) U) U) It) U) ••••• CD U) Q) It) It) U) It) U) U) .•.. I It) It) t\I (') It) •.... U) U) ~ DISTANCE FROM CENTER UNE (m) U) en CD 70 . MULE DEER TROUBLESOME til UNIT -so Q. :::l ~ 0 50 a:: n e JAN 1990 = 185 mFi:~:1 MAR 1991 n = 123" LL 0 40 a: w m ::E 30 :::l Z 20 ......... ~.---.. - .. 10 0 ..- It) <:) U)" It) t\I (') ~ U) U) (') It) I ..It) t\I It) It) I() U) It) • ~ <0 U) It) •.... U) It) CD Q) It) U) It) It) •.... CD It) It) ..It) Q) DISTANCE FROM CENTER UNE (m) Fiq. 1. Numbers of qroups of elk and deer observed at different distance intervals from the center line of line transects, Troublesome sub-unit, MIddle Park; -January -1990 and-March 1991. - �11 35 8J( 6.97 ~ -b--------------- TOTAL GROUP SIZE GROUPS 1990-1991 n =- 95 ~ 1-5 • 6-10 • 11-15 [] >16 25 fJ) 0.. ~ ·-·_-·--·---------------------------------f 20 CJ o Z 14.27 15 7.56 10 12.00 7.38 10.43 22.29 4.33 -------.---------.1><1----1 5 o 45-55 15-25 0-15 65-75 75-85 85-95 95-155 INTERVAL DISTANCE (M) Fia. 2. - Frequacy of .U: &rOup -a elistanc •. int.rvw for lin. trans.cts, Troubl.som. suh-uD1t, Middl. Parlt, . J_ary 1990 and March 1991. A_ras •. ala. of_all &rOUpsoba.rved within an int.rvaf -is shown above _th. to~al aroups· s.a in •• ch int.rval. . . 50 11.94 ELK GROUPS 1990-1991.. 40 , "~" ._.__ ._._. I- Z w 30 0 a: w 7.32 11.91 a.. 20 8.63 10 o CONIFER Fia. 3. -Fr.ciuac:y of .U: sroups -a January 1990 and March 1991. ASPEN habitat Av.ras. SAGEBRUSH OR OPENING MIXED CONIFER! ASPEN typea for lin. trans.cts, Troubl.som. suh-unit, Middl. Parlt, croup ala. for .ach habitat type is shown at th. top. �72 100 5.07 • TOTAL 1990-1991 GROUP SIZE GROUPS n •• 307 80 ~ 1-5 • 6-10 111 11-15 [J >16 ~ - -. _ .. "_"-_._--"---._._-"_ .. .." DEER 60 ~--.;><J.-.-.--------- ~- ... .. 9.06 ---------- 40 20 o 0-15 45-55 15-25 65-75 75-85 85-95 95-155 INTERVALDISTANCE (M) Fia. 4. FJ:eqa~ of d•• r aroups __ diatclc. iIlt.rvala for liD. trmaect., Troubl•• OIIIe .ub-un1t.. Middlti Pad:, J_ary 1"0 mel March1991. A_r •••• iz. of all aroup. ob•• rved within an int.rval b .hown move th. tot.al aroups ••••. iIl .ach iIlt.rval. ._ 10,000 MEANS +/- 95%CI ------_._-_ 8,000 ffiw ... __ .-r------ .-r- 6,000 C - o Z r- 4,000 2,000 _. , -=t=- o 1990 QUADS. Fia. S. 1990 TRANS. Populat.ion siz •• of d.er b•• ed upon quadrat.s and lin. . Yad:, January 1990 and March1991. - 1991 1991 QUADS. TRANS. t.rans.ct.s in th. Troublesom.sub-unit., Middle �73 Colorado Division of Wildlife Wildlife Research Report July 1991 JOB PROGRESS REPORT State of ~C~o~l~o~r~a~d~o _ Project No. W-153-R-3 Work Plan No. Mammals Research· Multispecies Investigations 1A Job No. Animal and Pen Support Facilities for Mammals Research Period Covered: Authors: Personnel: July 1, 1990 - June 30, 1991. M. W. Miller, M. A. Wild, B. J. Maynard, and G. A. Stout. D. D. K. C. L. R. W. L. Baker, A. L. Case, R. B. Gill, C. Y. Irvin, B. Krabel, Magnuson, M. J. McArtor, J. K. Ringelman, T. R. Ritchie, Scott,· R. B. Snyder, M. R. Szymczak, A. N. Torres, Wagner, J. Whipperman, and S. M. Willis. Abstract . . The"·.coloradoDivision of Wildlife's Foothtlls Wildlife Research Facility (FWRF) maintained captive animals (up to"95 wild and domestic ungula~~s of 6 species and 103 migratory and upland game birds of 6 species) and facilities supporting 14 different mammalian and avian research projects. Routine animal care a~d facility maintenance programs were conducted as previously described. In accomplishing these ongoing activities, however, we attempted to develop ways to improve current facility management practices, particularly those that would increase efficiency and efficacy of feeding and maintenance activities related to research facility operations. All daily feeding and weekly cleanup duties associated with research animal maintenance were accomplished using about.25-30 hrs/wk of work-study labor; consequent~y, costs associated with these activities have been reduced by about 75%. A variety of modifications in feeding regimes and practices were also evaluated; elimi~ating supplement from adult bighorn maintenance diets, feeding cub~d alfalfa to elk, and using hay racks for feeding alfalfa hay appeared viable approaches for further reducing animal maintenance costs. Thirty neonatal ungulates were bottleraised Using established procedures; we also began evaluating modifications to improve our rearing protocol. Surgical treatment appeared to offer limited success in curing lumpy jaw in pronghorn based on preliminary evaluations; a Bacteroides nodosus vaccine produced antibody responses but failed to protect adult pronghorn from developing jaw abscesses. ��75 ANIMAL AND SUPPORT FACILITIES MAMMALS RESEARCH FOR Michael V. Miller P. N. OBJECTIVE To provide and ~intain popula~ions of captive animals and pen facilities to support Mammals and Avian Research Programs. SEGMENT OBJECTIVES 1. Main~ain and improve animal research and holding facilities. 2. Coordinate all rearing, training, maintenance, and research activities. 3. Maintain up to 15 elk, 40 mountain sheep, 30 pronghorn antelope, 15 mul~ deer, and 10 domestic cows in suitable health to perform required research experiments. 4. Conduct management experiments to increase efficiency and efficacy of feeding and maintenance activities related to research facility operations. .METHODS ANn MATERrALS Routine animal care and facility maintenance programs supporting new-~nd ongoing Terrestrial Wildlife Research Program projects were conducted ·as previously described. In accomplishing these ongoing activities, emphasis was placed on developing ways to improve current facility management practices. This year, we emphasized approaches for increasing efficiency and efficacy of feeding and maintenance activities related to research. facility operations. To this end, we worked toward improving management and operation of CDOW"s Foothills Wildlife Research Facility through the following: ANIMAL MAINTENANCE General: We established a labor force for daily feeding and care of research animals that relies wholly on workstudy students for routine feeding, cleaning, training, weighing, and other related duties. In order to estimate costs of maintaining captive wildlife at FWRF, a log of time spent feeding, stocking feed storage. areas, and cleaning pens was recorded. The time spent .in these activities was estimated each day by the person assigned to feed over a l7-day period. Approximations of time spent stocking and cleaning were also made during that period. NUTRITIONAL MAINTENANCE Feeding protocols: We evaluated utility of a feeding regime for elk based on calculated energy requirements for various age, sex, and reproductive classes. In addition,._we conducted pil_ot ~tudies to evaluate effects _9f dietary supplementation qn performanc~ of dam-ra~sed calves, as well as -efficacy of �76 high-energy wafers and alfalfa cubes as alternatives base diet for adult elk. to long-stem hay as a As a means of improving access to and reducing waste of alfalfa hay, we installed hay racks in bighorn, pronghorn, and mule deer pens. In addition, we began evaluating alternative feeding regimes for maintaining captive bighorn and pronghorn herds; modifications revolved largely around adjusting the amount of high-energy feed offered to adult animals to supplement their base. diets of longastem alfalfa hay. Bottle-raisin, neonates: Neonatal elk, mule deer, pronghorn antelope, and bighorn sheep wer~ hand-raised in 1990 and/or 1991 for use in various ongoing research projects. All neonates were bottle-raised using methods described elsewhere (Wild"and Miller 1991). In 1991, we began evaluating a modification of our established feeding protocol. This modification involved feeding evaporated milk at ad libitum rates only through 6-8 weeks of-age, and limiting daily milk consumption after that time to the average daily intake for the last 1-2 weeks of ad libitum consumption. Milk volumes offered were then gradually decreased until weaning at 90 days of age. Copper, in the form of CuS04 (100 mg/ml, 0.4 ml/day), was added to pronghorn milk in an attempt to prevent health problems that were attributed to copper deficiency in previous years. HEALTH MAINTENANCE General: We developed and implemented use of a system for recording daily health status_ for, research animals. That ~ystem wa~ evaluated and modified .throughout the year. •. Methods for treating and/or preventing lumpy Jaw in captive pronghorn~ We began studying the pathogenesis of pronghorn lumpy jaw using evaluations on 9 does (consisting of diagnostic skull radiographs and documentation of size and site of current lesions) to identify, describe and monitor progress of the disease. To identify the etiologic agent(s), we collected samples from peripheral and deep lesions from four does £or bacterial culture. To develop and evaluate a surgical protocol for treatment of lumpy jaw lesions, six does were paired based on radiographic changes and randomly assigned to treatment group. Treatment animals were anesthetized for surgical approach to apical tooth root abscesses on upper and/or lower premolars and molars. Apical abscesses were treated by trephining through the bone into the root abscess, then curetting to debride the site of infected bone and necrotic tissue. Extraction was used only in one case of extensive bone resorption around the tooth root with tooth instability. Sites were flushed with betadine solution, then saline, and left open to drain and heal by second intention. Ceftiofur (150 mg subcutaneously) was administered intraoperatively and again two days postsurgery. Control animals received no treatment. Animals that became moribUnd during the study period were humanely euthanized using intravenous injection of T6la. Gross necropsies and tissue collection for histopathology and bacterial culture were performed on these animals. We obtained 4-mon~~ postoperative skull radiographs from treatment and control .. pronghorn· to .. aid in prelimi~ry evaluation of -surgical treatment. �77 In a related study, we began testing the efficacy of a killed Bacteroides nodosus vaccine (Footvaxa) in prevention of lumpy jaw in unaffected pronghorn. Prevaccination, we collected blood from all pronghorn for determination of titers to Bacteroides sp. One affected doe (SF) was initially vaccinated subcutaneously in the neck with 1 m1 vaccine using aseptic technique to test safety of the vaccine in pronghorn. Subsequently, four does and two bucks were vaccinated and three control castrates received 1 m1 sterile saline using the same procedure. Two vaccinated animals (236 and MI) were housed with affected does; 4 others (JJ, LA, ~I and BD) remained unexposed to affected animals. We collected blood from all pronghorn again about 10 weeks postvaccination. Plasma and sera were stored at -40 C until assayed. Serum samples from clinically affected, unaffected vaccinated and unaffected, assumed unexposed free-ranging pronghorn were submitted to Greg Ferrier at the Regional Veterinary Laboratory, Victoria, Australia. Titers to B. nodosus were determined using an enzyme-linked immunosorbent assay (ELISA) previously developed for sheep (Ferrier et al. 1988); pilot studies had previously revealed that anti-sheep/goat IgG could be used as a conjugate for detecting pronghorn antibody. In addition to serologic studies, vaccinated pronghorn were observed daily for development of jaw abscesses. FACILITY MAINTENANCEjREPAIRS/IMPROVEMENTS A vaxiety ~f scheduled and unscheduled maintenance and repair activities were necessary to.support faci1ity·ope~ation and ongoing re~earch.programs. As with appr'oaches to modifying our ANIMAL MAINtENANCE program, emphasis was placed on improving current facility man~gement practices through preventive maintenance projects and high-quaiity repairs to existing facilities~~ A "project board" system was implemented for identifying and assigning repair and maintenance tasks at FWRF. We continued to modify and improve that system throughout the year. RESEARCH PROJECTS .Facility operations supported offered support for pilot studies, special studies, and research experiments that were initiated, conducted, or continued using FWRF animals and facilities throughout the year. RESULTS AND DISCUSSION ANIMAL MAINTENANCE Estimated time required for accomplishing routine animal maintenance activities per individual animal was generally less when the number of animals per pen was large, but the total time required for accomplishing these activities increased with increasing numbers of animals. Estimated time spent in these activities was quite consistent among individual caretakers on a dayto-day basis. Time required for feeding, stocking feed storage areas, and cleaning pens averaged about 2.2 hrs/day or 15.5 hrs/week (Table 1). Estimates of time required for the specific duties described are given in Table J.. �78 Table l. Estimated time (minutes) required for accomplishing care activities at FWRF. daily animal A!I:!;;1v1~ Number of Individuals Feeding Stocking Cleaning Total Time Elk 15 25.7 4.3 6.6 36.6 2.4 Sheep 28 22.9 4.3 8.6 35.8 1.3 Goat 1 3.9 0 2.1 6.0 6.0 Pronghorn 14 16.3 4.3 4.3 24.9 1.8 .Deer 10 11.3 2.1 2.9 16.3 1.6 Cow 1 3.8 0.7 1.4 5.9 5.9 43 4,5 0 2.9 7.4 0.2 112 88.4 15.7 28.8 132.9 1.2 Species Waterfowl Totals All daily feeding and weekly clean-up accomplished using about 25-30 hrs/wk costs associated with these activities $207/wk for Utility Worker I labor to Time/ An./day duties, as well as most repairs, were of work-study labor. Consequently, have been reduced by about 75% (from $47/wk for student labor). -NUTRITIONAL MAINTENANGE Feeding protocol's: Preliminary analysis of performance data suggested the origi!tai feeding regime developed for elk underestimated maintenance-=. requirements for both nonlactating and lactating (particularly lactating) cows. Based on this assessment, we recalculated energy requirements andadjusted the feeding regime accordingly. Its evaluation is in progress. Supplemented elk calves showed somewhat higher average daily gains than calves fed alfalfa alone (0,40 and 0.35 kg/day for supplemented vs. 0.35 and 0.30 kg/day for unsupp1emented), Feeding high-energy wafers ad libitum caused digestive upset (as indicated by anorexia, cribbing, and changes in'defecation patterns) in adult elk; changing their diets to commercial alfalfa cubes eliminated these problems. Alfalfa cubes offered an alternative experimental foodstuff that was readily consumed, easily quantified, and without apparent digestive complications in short-term feeding trials using adult elk. We plan further evaluations in both elk and bighorn sheep. Hay racks installed in bighorn, pronghorn, and mule deer pens appeared to improve access to and reduce waste of alfalfa hay; this approach was particularly successful for feeding bighorns: subjectively, modification of hay feeders, combined with restriction of supplemental feed (see below) appeared to reduce their hay waste by about 90%. For both bighorn sheep and pronghorn, our feeding regimes have called for providing ad libitum quantities of alfalfa and supplement in recent years to assure all individuals were in optimum condition. This approach has proven both wasteful and detrimental. We eliminated supplement from daily diets of adult bighorns in mid-November; after 1 month of restricted feeding, their average weights were unchanged, and feeding efficiency had improved dramatically (see above). We continued monitoring responses and adjusted �79 diets as energy requirements increased during late gestation. This feeding regime had no apparent detrimental effects on lamb birth weights or survival. Our results suggest supplementation is unnecessary for adult bighorn sheep through most of the year. Using a similar approach, we restricted the amount of supplement offered to pronghorn. In late October, we began feeding them quantities of high-energy pellets just sufficient to meet average daily maintenance requirements. Our goals with this modified regime were to increase alfalfa consumption and reduce fatness in adult does. The new feeding program appears to have stabilized average doe weights since November; in contrast to this year's weight dynamics, the same individuals gained weight during November-December 1989. As with the bighorn trial, we continued monitoring responses and adjusting diets to meet seasonal changes in maintenance requirements. Several older does failed to maintain base body weights late in gestation, as reflected in postparturient weights, suggesting the calculated ration of supplement was insufficient to sustain both maintenance and fetal growth requirements. However, we observed no apparent adverse effects on birth weights or survival of pronghorn fawns. A reevaluation of this feeding regime is planned for next year. Bottle-raising neonates: We successfully hand-raised 2 deer fawns and 2 elk calves in 1990. In 1991, a total of 26 neonates were being bottle-raised by late June; these included bighorn lambs (4 male, 3 female), elk calves (2 male, 2 female), mule deer fawns (2 male, 2 female), and pronghorn fawns (5 male, 6 female) were born at the facility. Aside from 2 elk calves, 2 deer fawn.s.,and 1 pronghorn fawn, all necnaces were· born in capt·ivity at FWR¥. Birth dates ranged,from April·28 (bighorn) through .June 28 (pronghorn) (Table 2),-and varied by ·species. Average daily intakes and average weights on a weekly·basis for each spe~ies are shown in Figs. 1-2. Table 2. Birth dates and mean weights for neonatal ungulates born at FWRF dUIi.ng Ha:£-J:une},991. BIRTH DATES BODY WEIGHT (kg) SPECIES N bighorn sheep 7 28 Apr elk 4 17 May - 28 Jun (10 Jun) 21.0 pronghorn 11 4 Jun - 26 Jun (16 Jun) 2.9 mule deer 4 15 Jun - 19 Jun (19 Jun) 3.1 RANGE (MEAN) 16 May (9 May) 5.0 Minor neonatal health problems varied among the four different species. Mild viral enteritis attributed to rotavirus occurred in elk calves in both 1990 and 1991. Clinical signs included diarrhea, slight depression, and anorexia (0 - 50% of normal milk intake). In elk calves, this disease appeared to be self-limiting without treatment -- milk intake returned to normal within 1-7 days, and diarrhea resolved within 1-2 weeks. Captive-born mule deer (n-2) and pronghorn (n-2) fawns developed diarrhea within 2-7 days of birth. Affected animals with fecal pH >7 and/or body temperature >103 F were treated with injectable or oral trimethoprim-sulfa; mule deer fawns were also provided with electrolyte solution in bo~ls. These ·cases resolved within 1 week. Six �80 of 7 hand-raised bighorn lambs developed pneumonia by 6 weeks of age. (See WP2A,J4 for additional details.) Affected lambs decreased milk intake and weight gains as pneumonia developed, but these rates returned to near-normal within 1-1.5 weeks of initiating antibiotic therapy (Wild and Miller 1991). To date, no mortalities have occurred in association with any of these health problems. Limiting milk consumption at 6-8 weeks was implemented in an attempt to more. closely approximate intakes and weight gains observed in dam-raised neonates.· Milk consumption of dam-raised ungulates peaks at about 3 weeks of age (Robbins et al. 1981), at which time the young begin to consume more dry matter. Our existing protocol does not limit the neonates' milk intake, and it is conceivable that excessive milk intake could delay the surge in dry matter intake for our bottle-raised neonates, consequently depressing their growth rates below those of dam-raised individuals. Under our approach, limitation was delayed until 6-8 weeks of age to allow neonates to fully accept bottle feeding .. Only bighorn lambs had reached 6-8 weeks of age by June 30. and they had not been on the limited intake schedule long enough to evaluate its effect on performance. We plan to continue monitoring weights of all neonates both before and after intake limitation to evaluate this modification of our rearing protocol. A new technique was used for starting reluctant calves on the bottle. On separate occasions, 2 elk calves that resisted bottle-raising were put in a pen with· a socialized calf at feeding time. The recalcitrant calves would try to suckle from the socialized calves; the feeder could then place a bottle under the socialized calf for the ~ther to take .. This "surrpgate mother" technique was used only at scheduled feeding times. and only after the feeder had tried to promote suckling on her own, Because this technique could hamper imprinting on the human feeder, it was used only to get calves startea on the~ bottle. After-2-3 days, the calves were no longer allowed contact with their "surrogate mothers," forcing them to take the bottle from their human feeder. This approach appeared to be an effective means of preventing 7-14 day delays in initiating bottleQnursing in elk calves previously endured at FWRF. HEALTH MAINTENANCE General: Our system for recording daily health status for research animals markedly improved communications, detection, and treatment of animal health problems at FWRF. Overall, captive wildlife maintained at FWRF remained healthy throughout the year. In addition to ongoing problems with lumpy jaw -inpronghorn and pasteurellosis in bighorn described elsewhere, chronic wasting disease was confirmed in a SQyear-old elk cow (B86) in May; this represents the second case since our 1985 attempt to eradicate this disease from captive cervids at FRWF. Methods for treating and/or preventing lumpy jaw in captive pronghorn: Examination of captive pronghorn does revealed that 3 does (203, 236. NI) had no visible soft tissue abscesses, and 6 had 1-4 soft tissue abscesses each. Skull radiographs from all animals showed mild to moderate dental changes, including periodontitis, widening of apical germinal buds (suggestive of early abscesses), apical tooth root abscesses, osteomyelitis, tooth loss, malapposition of teeth and retained deciduous teeth. Of 27 root abscesses detected radiographically, 18 were associated with mandibular teeth. The second molar was mos~ frequently-involved on the lower arcade; however, the �81 fourth-premolar and first and third molar were all commonly abscessed. On the upper arcade, the fourth premolar was most frequently abscessed; the third premolar and first and second molar were involved less frequently. Two 27month-old does and one 39-month-01d doe had retained deciduous premolars permanent premolars usually erupt at 27-29 months of age in pronghorn. Radiographic abscessation pronghorn. attributable findings suggested that a progression from periodontitis to root to osteomyelitis may occur in the pathogenesis of lumpy jaw in Initiating factors for periodontitis remain unclear, but may be to abnormal tooth eruption and/or wear or gingivitis. Preliminary culture results revealed presence of Actinomyces pyogenes, Pasteurella multocida, Proteus mirabilis, Bacteroides sp. and Fusobacterium necrophorum in lesions associated with abscesses. Although these bacteria could be part of normal oral flora, presence in high proportions could lead to colonization and infection. Sanitation, high animal density and/or contact with infected animals may contribute to increased bacterial exposure, while diet form and composition, oral lesions and/or immunocompromise may increase the possibility of bacterial colonization. We performed surgery on 1 pilot case (SF) and 3 treatment cases (202, 231 and JO). We treated 2 maxillary and 8 mandibular root abscesses identified on radiographs, and 1 maxillary lesion identified by a draining tract. In addition, 1 lower molar was extracted (202) due to extensive bone resorption around the root, exposing it to the skin surface. All does recovered uneventfully postoperatively, with the exception of 202: she developed myiasis -at th~ site of molar extraction and ~emained ~epressed for about 10 day~. Although treatment with betadihe solution flush, KRS~ fly spraY'and 3 additional doses of ceftiofur appeared to resolve this problem, 'she subs~quent1y developed bronchopneumonia and died about 10 days postsurgery. In addition, 1 control animal (SB) developed severe necrotizing aspiration pneumonia refractory to antibiotic therapy; she was euthanized about 2 weeks after study initiation. . Preliminary evaluation of radiographs taken 4 months after surgery revealed that 1 or more lesions appeared to be healing in each of 3 treatment animals; however, 1 animal (231) still had 1 active tooth root abscess and another (SF) had 2. Healing was evidenced by formation of bony callous where lytic bone was surgically removed. These observations suggest that surgical treatment of tooth root abscesses can resolve the lesions, especially if lesions are acute. However, lumpy jaw itself may not be resolved in these animals because other tooth roots may subsequently be independently infected and require additional surgery. Skull radiographs and possibly surgery may be indicated if repeated attempts to manage abscesses by draining and flushing are not effective or result in chronically draining tracts. No adverse reaction to the killed B. nodosus vaccine was observed in the first doe vaccinated (SF). However, al16 study animals vaccinated subsequently developed firm swellings at vaccination sites; 2 of these drained pus, but subsequent culture results revealed these were sterile abscesses. Lesions resolved without treatment in 6-12 weeks. No reactions occurred in 3 controls injected with sterile saline. Despite these localized reactions, Footvax~ appeared to be safe in pronghorn. �Estimated antibody titers to B. nodosus did not appear to differ between clinically affected and unaffected animals; however, postvaccination titers were markedly higher than those in any unvaccinated animal (Table 3). We may not have detected titers in affected animals for any of several reasons: 1) pronghorn may harbor a different antigenic strain or species of Bacteroides than that used for the vaccine or ELISA antigen; 2) Bacteroides sp. organisms may be present in low numbers or "walled off" and protected from host immune response; 3) natural titers may be occur upon initial infection but be short~ lived. Table 3. Serologic affected pronghorn. History Unaffected Unexposed· Vaccinated Pre titers to B. nodosus in unaffected, Animal ID 1 2 3 236 LA NI JJ BD ;MI vaccinated Titer History Animal ID 150 150 330 Exposed 233 234 238 246 180 360 310 560 190 190. Postb 236 LA NI JJ. BD MI and Titer 220 350 190 290 3300 11760 7020 20000 80000 133000. Affected JO SL SF 202 SB 231 203 270 ·180 150 310 250 400 190 ·Freemranging pronghorn bSerum collected about 10 weeks after initial vaccination Jaw abscesses occurred in both vaccinated bucks (BD, MI) in February, 1991. Actinomyces sp., Fusobacterium necrophorum, Fusobacterium sp. and Bacteroides sp. were cultured from these abscesses. Soft tissue lesions resolved within 1 week of draining and flushing the abscesses; however, bony proliferatIon is apparent on BD's mandible. Four does remained clinically unaffected until June, when #236 showed signs of a small maxillary abscess that may have developed and ruptUred. We plan to continue monitoring vaccinated animals for development and progression of lesions to further evaluate protective ability of the measured antibodies. �H3 FACILITY MAINTENANCE/REPAIRS/IMPROVEMENTS Significant maintenance/repair/improvement year included: projects completed at FWRF this - installing drainage ditches in the west alleyway and upper bighorn pens; - cleaning, sterilizing, and regraveling the common area east of the isolation pens, installing new gates to improve access to that area, and installing cutoff gates to improve animal handling capabilities in this· area; . - staining alleyways and shelters in the east facility to reduce weather damage; - completing a facility-wide clean-up -- numerous unused items were removed to increase storage space for research materials at FWRF; - installing new feeding structures in elk paddocks, and constructing and installing wooden or steel pipe hay racks in bighorn, pronghorn, and mule deer pens; - building a feed storage sheds and shelters in mule deer and pronghorn pens; - weather-proofing the hay storage shed; - adding new latches and completing other improvements for both east- and west-side alleyways; - preparing a SOP listing all procedures undertaken in our annual facility-wide winterization process; - preparing a plan for expansion of hoofed-stock paddocks; - rebuilding and fortifying the elevated ramps entering the digestion cages; .".. buildirig and" ins.ta:Uing new· dcczs- in the west side· scale room, freezer room: and primary feed storage shed;· -. _.:_ installing plywood flooring in the east side primary feed storage ahed ;> - installing wing fences and visual barriers in pronghorn and mule deer pens; . " - rebuilding the lumber storage rack; - constructing and installing a new facility for weighing and handling pronghorn and mule dee~; - planting windbreaks along the main road and creek bottom. RESEARCH PROJECTS In addition to ongoing facility management experiments and improvements described above. the following (listed in no particular order of importance) pilot studies, special studies, and research experiments were initiated, conducted, or continued using FWRF animals and facilities this year: - New models of the functional response in vertebrate herbivores: the role of plant availability and animal morphology -- Spalinger, Hobbs, Wunder, and Gross; - Relative effectiveness of repellents for reducing feeding by elk -Andelt, Baker, and Burnham; - Use of behavioral observations to estimate conception and recurrent estrus in captive elk -- Ritchie, Baker, and Lehner; �84 - Effects of plant and free water on voluntary feed intake in bighorn sheep -- Anderson, Hobbs, and Baker; - Development of an experimental model to reproduce clinical carfentanil renarcotization in elk -- Miller, Lance, Wild, Dunn, and Tipton; - Experimental evaluation of a sustained-release formulation of naloxone hydrochloride for preventing carfentanil renarcotization in elk -Miller, Lance, Wild, and Dunn; - Effects of sample handling on measured blood progesterone captive elk -- Freddy, Miller, and White; - Pronghorn spectral sensitivity levels from -- Maynard, Lehner, Gill, and Roberts; - Epizootiology of pasteurellosis in captive Rocky Mountain bighorn sheep -- Miller, Wild, Mills, and Snipes; - Seasonal changes in fecal cortisol excretion in captive Rocky Mountain bighorn sheep -- Miller and Ritchie. - Field dose determination of medetomidine-ketamine sedation of cervidae with reversal by atipamezole Wild; - Nutritional combination for -- Miller, Lance, and ecology of sage grouse during winter -- Remington; E~feet~ of'pregruincy on intake and digest~Dility and MagnUson .. LITERATURE in captive elk :"-.Baker CITED Wild, M. A., and M. W. Miller. 1991. Bottle~raising wild ruminants in captivity. Outdoor Facts 1/114, Colo. Div. Wildl., Denver. 6pp. Robbins, C. T., R. S. Podbielancik-N~rman, D. L. Wilson, and E. D. Mould. 1981. Growth and nutrient consumption of elk calves compared to other ungulate species. J. Wildl. Manage. 45:172-186. �85 Colorado Division of Wildlife Wildlife Research Report July 1991 JOB PROGRESS REPORT State of ~C~o~l~o~r~a~d~o_ Proj ect No. 0 __ Mammals Research ..:;W:..;;..• 1.&5"",3.;;.,i-R!;l:,-...;:4~ __ Work Plan No. lA Multispecies Job No. 3 Mammals Research Administration Period Covered: Author: Investigations July 1, 1990 - June 30, 1991 0 R. B. Gill Personnel: R. B. Gill, L. E. Lovett Abstract Duri~g the 1990-91 segment members of the Mammals Research staff produced the following: 0 ;dr~f~s of 5 management analysis repprts; ~oarticres in professional journalsfbooks; -:7 articles in professional sympos La ; 1 agency technical publication; and 2 non-tachnical 0 publications. Staff members served on a variety of management task forces and steering committees and produced drafts of 5 species management analysis reports. The Mammals Research staff exceeded all of the production objectives for Federal Aid Prooject W-153-R-4 and expended only 95% of the resources allocated to accomplish those objectives. ��87 HAHKALS RESEARCH ADMINISTRATION a. Bruce Gill P.N. OBJECTIVE Administer research within the Mammals 1 and·2 Research Units for the highest productivity at the lowest cost. SEGMENT OBJECTIVES 1. Assign and supervise the research of 6 Wildlife Researchers. 2. Assign and supervise secretarial and clerical work of 1 Senior Secretary. 3. Lead the development and implementation of statewide species management plans. 4. Assist in planning of Rocky Mountain Arsenal Watchable Wildlife Showcase. RESULTS ~ DISCUSSION During the FY 1990·-91 segment; I prepar~~_'performance plans, supervised, and· During -.the evaluated the technical performance 11 Wildlife Researchers. segment these employees produced the following: drafts of 5 management analysis reports; 6 articles in professional journalsfbooks; 7 articles in professional symposia; 1 agency technical publication; and 2 non-technical publications. In addition, I co-authored 4 manuscripts which have been accepted for ·publication in 1991-92 pending revision and co-authored 3 manuscripts which currently are in internal peer review and will be submitted to professional journals for acceptance in FY 91-92. Members of the Mammals Research staff served on the Division's Habitat Partnership Steering Committee; the Commercial Wildlife Parks Working Group; the Rocky Mountain Arsenal Watchable Wildlife Steering Committee; the Mountain Sheep and Mountain Goat Auction Project Selection Committee. Mammals Research staff produced a draft Problem Bear Management Administrative Directive; policy level objectives for the Long Range Plan update on black bears; worked closely with the Habitat Section and the Public Services Section to develop solutions to the San Juan grizzly .bear controversy; developed a Mission Statement and alternative Program Structures for DNR; worked closely with the mountain lion symposium committee to organize ana evaluate the-Mountain Lion Symposium and develop implementation recommendations for the Director's consideration.· �88 The Mammals Research staff produced drafts of 5 management analysis reports: deer, elk, pronghorn, mountain sheep, and trapping; developed a second survey of public attitudes concerning Colorado's black bear management program which would include only those likely to vote; produced a draft of a Black Bear Technical Bulletin on schedule for release at the September, 1991, Wildlife Commission Meeting; developed successful proposals for "decision item" -funding concerning lynx status survey, kit fox status survey, and development of a black bear inventory process; developed 5 research proposals: a) ef!ects of early hunting on the distribution of elk in Northwestern Colorado; b) the effects of disturbance by recreationists on the Arkansas River mountain sheep population; c) development of fertility control techniques to manage unhuntable deer popUlations; d) experimental evaluation of mountain sheep transplanting and disease management; e) effects of density on dispersal of pronghorns in Middle Park Colorado. During FY 1990-91, we were allocated $1,194,056 to accomplish the objectives of Federal Aid Project W-153-R-4. We were unable to match the allocation against real expenditures because the Division's COFRS system was unable to produce budget status reports for all of FY 199091. However, the Mammals Research in-house computerized "checkbook" accounting system indicated we exceeded the production objectives for FA Project W-153-R-4 at a total cost of $1,131,324 or at 5% under budget. Prepared by ~~.~.~w9 R~aGiil -.- Wildlife Research Leader �0;;1 Colorado Division of Wildlife Wildlife Research Report July 1991 JOB FINAL REPORT State of ~C~o~l~o~r~a~d~o~ _ Project No. W-153-R-4 Mammals Research Work Plan No. lA Multispecies Investigations Job No. 4 Wild Ruminant Forage Selection Dynamics Period Covered: Author: July 1, 1990 - June 30, 1991 B. J. Maynard Personnel: B. J. Maynard, P. N. Lehner, R. B. Gill Abstract A Master of Science Thesis entitled: Spectral Sensitivity of Pronghorn (An~ilocapra americana) was completed-under the supervision of Dr. P. N. Lehner of the Department of Biology at Colorado State University, Fort Collins, Colorado-80523. The abstract of that thesis follows: The spectral sensitivity- of pronghorn, Antilocapra americana, was - examined by the use of corneal ~lec~roretinography and with a two-choic~ behavioral discrimination procedure. Electroretinography revealed a ~rk-adapted curve of the typical shape with peak sensitivity around 550 nm. _Behavioral discrimination tests confirmed that spectral sensitivity decreases rapidly between 700 and 750 nm. Pronghorn failed to respond to stimuli of 750 nm and longer wavelengths under both procedures. Prepared by _._A.-;;, ...•.••• A ~""",",N J~A,/Jn",,+1--' •...••. ~~ Lr.=.;. A=-->o' ~~ Master's Candidate IJ •,'--•... _ ��Colorado Division of Wildlife Wildlife Research Report July 1991 JOB PROGRESS REPORT State of ~C~o~l~o~r~a~d~o~ _ Project No. W-153-R-2 Mammals Research Work Plan No. lA Multispecies Investigations Job No. 5 Consulting Services for Mark-Recapture Analysis Period Covered: Author: July 1, 1990 - June 30, 1991 G. C. White Personnel: R. M. Bartmann, D. Saltz, L. H. Carpenter, R. B. Gill, T. D. I. Beck Abstract Progress towards the objectives of this job include: 1. A manuscript s~arizing ~he results of the Piceance Basin deer populatiQn studies has been accepted for publication as a WildlifeMonograph of The Wildlif~ Society. 2. A study of compensatory effects of harvest on the Piceance Basin mule deer population has been designed, with experimental harv~sts in December, 1989 and 1990. Radio collars to monitor over-winter survival of fawns were placed on the animals during November, 1989 and 1990. 3. The TRAPMARK database has been developed to store records of marked animals. Users are able to enter data, and to interrogat,e the database for potential radio frequencies or neck collar colors that could be used in a specifi~!igeographic area. 4. Consultation has been provided in the design and analysis of evaluating the validity of harvest estimates of elk and deer in North Park. 5. Consultation has been provided on sampling and design and analysis procedures to estimate crayfish biomass to aid in river otter recovery research. ��93 CONSULTING SERVICES FOR MARK-RECAPTURE ANALYSES G. C. White . P. N. OBJECTIVES Model and simulate population estimates of deer, elk, mountain sheep, and mountain goats with mark-recapture methods. SEGMENT OBJECTIVES Evaluate correlations of over-winter survival rates of wild mule deer (Odocoileus hemionus) fawns with levels of urinary cortisol:creatinine and urea nitrogen:creatinine ratios in 14 tame yearlings. RESULTS AND DISCUSSION Determining population condition is critical to the sound management of wild populations. Although never clearly defined, 'population condition' generally refers to the ability of a population to function in a deteriorating environment in terms of survival and birth rates. Population condition has been assessed by a variety of indices. Most fall under one of 3 categories .(Hanks 1978): (1) physiological condition of individual animals determined by .fat deposits, bl-ood parameters,' body' growth, etc., (2) demographic' parameters such ~s survival and fecundity of each sex and age group in the populC!tion, and (3) behavioral attributes such as feeding strategy and the rate and quality of social interactions. Physiological indices of popUlation condition are the most commonly investigate~ (Franzmann 1985). Nonetheless, as Nisbet et al. (1989) point out, these indices have a fundamental problem - they are measured at the level of the individual animal, while the manager's primary interest is at the population level. It is often assumed 'that, taken collectively, individual animal mea$urements can depict the condition of the population, although there are few data to support this hypothesis (Hanks 1978, Nisbet et al. 1989). Cortisol and urea nitrogen are physiological parameters recently used as animal condition indices (Saltz and White 1991a, DelGiudice et al. 1989). Cortisol is a glucocorticoid that mobilizes energy reserves in animals experiencing chronic stress (Stephens 1980). As such, it is assumed to reflect the net instantaneous energy loss by an animal. By contrast, urea nitrogen is a metabolite that is positively correlated with protein intake (Franzmann 1985) and is, therefore, indicative of diet quality_ However, this relation is reversed under extreme conditions (Brown 1984). Both cortisol and urea nitrogen can be measured in serum or as ratios to creatinine in urine (Franzmann 1985, Miller 1991). In a recent study, Saltz and White (1991a) investigated the effects of different environmental conditions' on the levels of urinary cortisol:creatinine and urea nitrogen:creatinine in tame mule deer. The tame deer wintered in two fenced pastures stocked with wild deer at high and low densities. Cortisol:creatinine ratios increased with population density and as winter progressed, and were lower in animals entering the winter with greater body mass. Urea nitrogen:creatinine ratios decreased as winter �94 progressed and increased with cortisol levels. In a parallel study conducted in the same pastures (Bartmann et al. 1991), fawn survival decreased with population density. Thus, we hypothesize that both cortiso1:creatinine and urea nitrogen:creatinine are reliable indices of mule deer population condition. The purpose of this paper is to test these hypotheses. We do so by correlating cortisol:creatinine and urea nitrogen:creatinine ratios described by Saltz and White (1991a) with the fawn survival rates described by Bartmann et al. (1991). METHODS The study was conducted in fenced pastures just east of the Colorado Division of Wildlife Little Hills Experiment station in northwestern Colorado. Two pastures (66 and 169 ha) were stocked with wild mule deer during NovemberDecember 1987, at high (1.33 deerjha) and low (0.44 deerjha) densities (Bartmann et al. 1991). The stocked populations consisted of 50 radiocollared fawns in each pasture and the number of adults needed to achieve the desired densities. Radios were equipped with mortality sensors. The fawns were radio monitored on a daily basis between December 1987 and mid-April· 1988. Fawns with collars that emitted mortality signals were located, and the cause of mortality determined. A detailed description of the pastures, trapping, and collaring techniques, can be found in Bartmann et a1. (1991). Urine samples were obtained from 14 hand reared mule deer yearlings. Yearlings were paired based on sex and body'mass, and individuals from each pair ~ere assigned randomly to the high- and low-density pastures·. They were place~ in the pastures November 18, 1987, and except tor token amount~ of .pellet feed (Saltz and White 1991a) survived by feeding on existing native vegetation. Urine was collected by placing the yearlings in specially built cages located in the pastures. An effort was made to collect one sample per yearling per week. Levels of urine cortisol and urea nitrogen were determined in relation to urine creatinine levels (Berman et al. 1980). Analyses for cortisol were radioimmunoassay procedures '(Olson et al. 1981), and were validated by .para1lelism and quantitative recovery (Saltz and White 1991b). Analysis for urea nitrogen was a colorimetric procedure using a Gilford spectrophotometer. Creatinine levels were determined by the Jaffe (1886) procedure with an Hitachi analyzer. A detailed description of the urine collecting methods, urine analysis and assay validation can be found in Saltz ~nd White (1991a,b). All statistics were performed and are presented using the ratio of micrograms cortisol per milligram creatinine (cortisol:creatinine ratio), or milligrams urea nitrogen to milligrams creatinine (urea nitrogen:creatinine ratio). Mean cortisol:creatinine ratio and urea nitrogen:creatinine ratio were determined weekly for each pasture. Weekly survival rates of radio-collared fawns were calculated as the number of fawns surviving at the end of the week divided by the number alive at the beginning of the week. Data were analyzed using logistic regression procedures (PROC CATMOD, SAS 1987), with weekly survival rate as the dependent variable. We ran separate analyses with weekly means of cortisol:creatinine as the predictor, and weekly means of urea nitrogen:creatinine as the predictor. Survival was weighted by number of animals alive at the start of the week. and data were blocked by pasture (density). Because higher cortisol:creatinine and urea nitrogen:creatinine ratios, as well as lower survival rates, were associated with the higher density pasture (Saltz and White 1991a), a co1inearity exists between the �95 predictors and blocking factor. Furthermore, when running the regressions on the combined data from both pastures, significance may be due to pasture effects alone and not to fluctuations within each pasture. To determine if weekly survival rates were correlated to weekly means of cortisol:creatinine or urea nitrogen:creatinine, we ran additional regressions using data from each pasture separately. Results were considered significant at f < O.OS. RESULTS A total of 19 weeks of survival and urinary data were collected between December 7, 1987, and April 15, 1988. Data from the two smallest yearlings were excluded from our analyses after the high density yearling developed an abscess and was subsequently artificially fed. By the 17th week all fawns in the high density pasture died. Consequently, we ran logistic regressions on the weekly survival from both pastures for the first 17 weeks only. Cortisol:creatinine in the high density pasture increased throughout the winter while survival decreased (Fig. 1). By contrast, ·survival and cortisol:creatinine response patterns in the low density pasture were more ambiguous (Fig. 2). There was a mild decline in survival throughout the winter with a possible increase towards the end of the winter. Cortisol:creatinine levels increased at the beginning of the winter and remained roughly constant until the end of the winter when they declined to their lowest levels. Initially, we ran a full model logistic regression on suryival with cortisol:creatinine and pasture density as·predictors. We found no significant interaction (f - 0.1007) between the predictors and therefore removed the interaction term. In the reduced model, the effect of �96 mGH DENSITY PASTURE 1.0 1.0 0.11 ,I• -= ---=.- 0.9 <lie - = IE •• 0.6 &l 0.5 8 0.4 ~: --I ---_. ---W === =~ == = === = I i 0.9 0.8 0.7 til 0.6 ~ 0.5 ~ 0.4 ~ 0.3 0.2 0.1 ~ O.~ o Figure 1. Fawnsurvival (_-..-) and yearling urinary cortisol: creatinine - -) and urea nitrogen:creatiniue (••••• ) levels in the high density pasture. -(- - �97 LOW DENSITY PASTURE 0.11 1.0 1.0 0.9 0.10 0.8 0.09 0.7 fIl U 0.6 ~ 0.5 Z o 0.4 0.08 ~ 6 0.6 ~ 0.07 0.5 ~ 0.06 0.4 ~ ~:~8 0.3 O.OS 0.2 ~ 0.1 ~ 0.04 0.1 o o I DEC 87 II JAN 88 II FEB 88 IIMAR 88 II APR I Figure 2.· Fawn survival (---) and yearling ur1nary cortisol:creatinine .- -) and urea nitrogen:creatinine (••••• )-~evels in the low density pastur-e. (- cortisol:creatinine was highly significant (f <0.0001), but pasture density was not (f - 0.5087). When regressions were run separately for each density we found a highly significant effect (f < 0.0001) of cortisol:cr~atinine on survival in the high density pasture, but no significance in the low density pasture (f - 0.8830 for 17 weeks, and f < 0.6793 for 19 weeks). Urea nitrogen responses were similar in both pasture, declining at the onset of winter and increasing at the end. In the full regression model the interaction between the two predictors (urea nitrogen:creatinine and population density) was not significant (f - 0.4914). In the reduced model, urea nitrogen:creatinine was significant (f - 0.0074), as was population density (f - 0.0020). Regressions run separately for each pasture produced a significant effect (f - 0.0204) of urea nitrogen:creatinine in the high density pasture, but no effect in the low (f - 0.2832 for 17 weeks, and f 0.7835 for 19 weeks). DISCUSSION AND MANAGEMENT IMPLICATIONS If an index of population condition is to be of any practical use, it must be linked to the survival and/or birth rates of the population. In this study we had no data on adult survival and reproquctive success, and compared our indices with fawn survival only" Neve rrheLess , juvenile survdva'l,is the major factor in the dynamics of mule deer (Connolly 1981), .and other mammalian populations (Hanks 1978). To our knowledge, this is the first time a �98 physiological index has been successfully correlated to a component of population condition. Our study was carried out in only 2 pastures with different densities. Thus, because there are no pasture replicates, results are limited to these pastures alone. Although the differences in survival, cortisol: creatinine, and urea nitrogen:creatinine ratios between the 2 pastures are real, their applicability to other areas should be viewed with caution. Both cortisol:creatinine and urea nitrogen:creatinine were correlated to fawn survival. In the urea nitrogen:creatinine model both predictors (urea nitrogen:creatinine and pasture density) were significant. By contrast, we found no effect of pasture density in the cortisol:creatinine model. This suggests cortisol:creatinine accounts for density effects while urea nitrogen does not. We therefore conclude that cortisol:creatinine is superior to urea nitrogen:creatinine as a predictor of population condition. This is in agreement with results presented by Saltz and White (199la). The lack of significance in regression of cortisol:creatinine on fawn survival in the low density pasture may be the result of 2 factors: (a) Minor changes in mean survival and cortisol: creatinine 'in relation to the variance. This suggests that under low stress levels, the noise in the data overshadows the actual effects, rendering this technique questionable u~der such conditions. (b) A drop in cortisol:creatinine towards the ,end of the winter that was not associated with_a rise in survivorship. This may be due to correlating a physiological response of yearlings with the survival of fawns. Yearlings may be able to recover from winter starvation more quickly than fawns, causing the above discrepancy. Regressions of just the first 15 weeks of data in the low density pasture did produce a significant (f - 0.0446) effect of cortisol: creatinine, but not of urea nitrogen:creatinine (f 0.3525). Cortisol is a mobilizer of energy reserves needed for survival in a deteriorating environment. Thus, cortisol is not a direct indicator of condition, but rather an index to the rate at which an animal's condition is deteriorating. Physiologically, elevated blood levels of cortisol are defined as stress (Stephens 1980). Ecologically, stress may be described as any environmental change that acts to reduced the fitness of an individual Koehn and Bayne (1989). Although widely used in ecological studies, the term still draws substantial criticism. The debate over the term is not over its definition, but rather whether the term should be used at all in ecology (Grime 1989). Data presented in this paper demonstrate that physiological stress is in fact correlated to fitness; and, therefore, its use in ecology is justified. Urinary cortisol offers an approach for managers and agencies to assess deer population condition and fawn survival. Urine collection methods, and the logistic problems involved in them, have been previously discussed by Kreeger et al. (1986) and Saltz and White (199la). Although further research is needed, urinary cortisol app~ar~ __ to be a reliable and easy technique to asses popUlation condition: �99 LITERATURE CITED Bartmann, R. M., G. C. White, and L. H. Carpenter. 1991. Compensatory mortality in a Colorado mule deer population. Wildl. Monogr. In press. Brown, R. D. 1984. The use of physical and physiological indices to predict the nutritional condition of deer - a review. Pages 52-63 in Proc. deer in the southwest: a workshop. New Mexico State Univ., Las Cruces. Connolly, G. E. 1981. Limiting factors and population regulation. Pages 245-286 in Wallmo, O. C., ed. Mule and black-tailed deer of North America. University of Nebraska Press, Lincoln. DelGiudice, G. D., ,L. D. Mech, and U. S. Seal. 1989. Physiological assessment of deer populations by analysis of urine in snow. J. Wildl. Manage. 53:284-291. Franzmann, A. W. 1985. Assessment of nutritional status. Pages 239-259 in R. J. Hudson and R. G. White, eds. Bioenergetics of wild herbivores. CRC Press, Inc. Boca Raton, FL. Grime, J. P. 1989. The stress debate: J. Linn. Soc. 37:3-17. symptom of impending synthesis? BioI. Hanks, J. 1978. Characterization of population condition. Pages 47-7~ in C. W. Fowler, and J.:D. Smith, eds. Dynamics of large mammal "popul.at Lons . Wiley, New York. Jaffe,-M. 1886. Veber den Reidersch1ag, we1chen Pikrinsaeure in normalen Harn erzeugt and veber eine neue Reuktion des Kreatinins. Z. Physiol. Chem. 10:391-400. Koehn, R. K., and B. L. Bayne. 1989. Towards a physiological and genetical understanding of the energetics of the stress response. Bio1. J. Linn. Soc. 37:i57 -171. Kreeger, T. J., G. D. DelGiudice, U. S. Seal, and P. D. Karns. 1986. Methods of urine collection for male white-tailed deer. J. Wild1. Dis. 22:442444. Miller, M. W., N. T. Hobbs, and M. C. Sousa. 1991. Detecting stress responses in Rocky Mountain bighorn Sheep (Ovis canadensis canadensis): reliability of cortisol concentrations in urine and feces. Can. J. Zoo1. 69:15-24. Nisbet, R. M., W. S. C. Gurney, W. W. Murdoch, and E. MCCauley. 1989. Structured population models: a tool for linking effects at individual and population level. BioI. J. Linn. Soc. 37:79-99. Olson, P. N., R. A. Bowen, P. W. Husted, and T. M. Nett. 1981. Effects of storage on concentration of hydrocortisone (cortisol) in canine serum and plasma. Am. J. Vet. Res. 42:1618-1620. Saltz, D.;"and G. C:-White. 1991a. Urinary cortisol and urea nitrogen responses to winter stress in mule deer. J. vrrai . Manage. 55-: 1-16. �100 ______ , and G. C. White. 1991b. Urinary cortisol and urea nitrogen responses in irreversibly undernourished mule deer fawns. J. Wi1d1. Dis. 27:41-46. SAS Institute, Inc. 1987. SAS/STAT guide for personal computers, version 6 ed. SAS Inst. Inc., Cary, NC. 1028pp. Stephens, D. B. 1980. Stress and its measurement in domestic animals: review of behavioral and physiological studies under field and laboratory situations. Adv. Vet. Sci. Compo Med. 24:179-210. Prepared by ~C0t;;b;;~te .. Colorado Sta~e University a �101 Colorado Division of Wildlife Wildlife Research Report July 1991 JOB PROGRESS REPORT State of ~C~o~l~o~r~a~d~o~ _ Project No. W-153-R-4 Mammals Research Work Plan No. 2A Mountain Sheep Investigations Job No. 4 Experiments to Identify and Manage Stress in Mountain Sheep Populations Period Covered: Authors: Personnel: July 1, 1990 - June 30, 1991 M. W. Miller, N. T. Hobbs, and M. A. Wild. K. W. Mills, K. P. Snipes, E. S. Williams, A. Boeger-Fields, and T. R. Spraker. Abstract We propose using an adaptive environmental assessment approach for studying and managing bacterial and viral diseases in free-ranging bighorn populations. Our .;tpproachwill involve simultaneous development'of_.a.computer model simulating bighorn population dynamic~ (including e~izootiology of . pasteurellosis) and of tools for identifying strains of Pasceurella haemolycica and quantifying immunological responses of bighorns to infection by these pathogens. We will use the model to identify variables sensitive to management perturbations in altering the dynamics of disease in bighorn populations; results will serve as the basis for designing management level experiments in the future. Diagnostic tools under development will be key components of laboratory and field experiments designed to evaluate potential tactics (including vaccination and/or treatment) for managing pasteurellosis in wild sheep, and appear prerequisite to initiating management level experiments. In conjunction with numerous cooperators, we summarized data gathered to date from a series of studies conducted in Colorado and elsewhere to develop a simulation model and serologic and bacteriologic tools for use in studying and managing pasteurellosis in free-ranging bighorn populations. We examined Pasceurella spp. isolates (n-5l) from wild bighorns (n-142) representing 5 indigenous herds during January-March 1991. Four of 5 herds sampled (Taylor River, Avalanche Creek, Tarryall Mountains, Chalk Creek) yielded Pasceurella spp. isolates. Prevalence estimates based on cultures of pharyngeal swabs ranged from 17-83% (mean 49%) among sampled herds. At least 6 distinct phenotypes (T3; 4; 3,4; 3,4,10; A7; and untypable) of Pasceurella spp. were identified among the 4 herds; we recovered T3 isolates from all 4 herds, but prevalence varied. Representative isolates from each herd have been submitted for genomic fingerprinting, and sera will be submitted for ELISA to exami.neexposure to a battery of P .. beemo l.ytz i.ce ..strains. �102 A pneumonia epizootic in the Taylor River-Almont Triangle bighorn herd killed at least 17 sheep across all age/sex classes; necropsies of 5 dead sheep (2 rams, 3 ewes) revealed subacute to chronic bronchopneumonia. Four distinct serotypes (T3; 4; 10; and 3,4,10) of nonhemo1ytic P. haemolytica were isolated from 4 sheep; Horaxella spp. and Actinomyces pyogenes were also isolated from some affected individuals. Immunoperoxidase staining of lung tissue revealed presence of bovine respiratory syncytial virus in 1 ewe; other viral or chlamydial agents were not detected. Gross, histologic, and fecal examinations revealed that lungworm did not appear to play a role in this epizootic. Preliminary evaluation of data from monthly collections of feces from captive bighorns suggested reproductive status may influence fecal cortisol excretion in bighorn ewes (Fig. 2); by May, fecal cortisol concentrations from pregnant ewes (mean±SE-39.S±6.9 ng/g dm) were about 77% higher than those from open ewes (22.2±3.0 ng/g dm)(P<O.OS). Based on these findings, further evaluation of seasonal and other influences on fecal cortisol measurements is warranted before these techniques are applied to studies of stress responses in freeranging bighorns. �103 EXPERIMENTS TO IDENTIFY AND MANAGE STRESS IN MOUNTAIN SHEEP POPULATIONS Michael W. Miller P. N. OBJECTIVE To treat bighorn sheep to control disease where necessary. SEGMENT OBJECTIVES 1. Develop a research strategy and proposal for managing bacterial and viral diseases in mountain sheep populations; begin conducting approved and funded research. 2. Design a population-level experiment to evaluate techniques for detecting stress in free-ranging bighorn sheep; begin conducting approved and funded research. MANAGEMENT OF BACTERIAL AND VIRAL DISEASES IN MOUNTAIN SHEEP POPULATIONS Inability to control infectious disease ou~breaks and subsequen~ mortality in mountain sheep populations. represents a significant obstacle to long-term success in their management. Although the "bighorn pn~umonia complex~_has been studied intensively for over 3 decades, little is known about many aspects of its etiology and epizootiology. Moreover, management interventions recommended for preventing or controlling this problem remain untested. Most previous efforts to improve understanding and management of the epizootiology of pneumonia in bighorns involved post hoc investigations .of dieoffs occurring in free-ranging sheep herds. These studies identified various etiological agents associated with known mortalities and attempted to determine predisposing causes and population consequences of individual outbreaks. From these investigations, comparisons of real or perceived patterns became the basis for hypotheses on the epizootiology of pneumonia in bighorns. Recognition of similar patterns in other outbreaks served as evidence supporting these as unifying hypotheses. Unfortunately, several of these hypotheses have failed to withstand rigorous experimental testing (Miller 1988, Miller et al. 1990, 1991). And, despite our best management efforts, bighorns continue to die. Our strategy for developing a better understanding of the epizootiology and management of bacterial and viral diseases in bighorn populations differs -generally, we propose to take an adaptive environmental assessment approach for studying the bighorn pneumonia complex. As a foundation for our research strategy, we have initially attempted to assimilate existing knowledge on bighorn population dynamics (including the epizootiology and consequences of infectious disease) into computer simulation models (Hobbs et al. 1990, Hobbs and Miller 1991). Because pasteurellosis appears to underlie virtually all r~s·piratory disease problems reported for bighorns: our modeling efforts have focused on the epizootiology of pasteurellosis in sheep populations .. We have constructed models that reflect dynamics of bighorn populations seen in nature �104 using the simplest assumptions necessary to reproduce those behaviors. Once we have a reliable working model, we plan to conduct simulation experiments to identify variables that might be particularly sensitive to management perturbations in altering the dynamics of disease in bighorn populations. Those results will serve as the basis for designing management level experiments in the future. In parallel with our modeling efforts,-we plan to conduct a series of experiments to develop, improve and standardize methods for collecting and interpreting diagnostic data to provide better estimates of key parameters driving our models. In particular, _we have been developing tools for identifying strains of Pasteurella haemolytica and quantifying immunological responses of bighorns to infection by these pathogens. These tools will be key components of laboratory and field experiments designed to evaluate potential tactics (including vaccination and/or treatment) for managing pasteurellosis in wild sheep, and appear prerequisite to initiating management level experiments. To this end, our recent efforts have focused on simulation modeling and on improving tools available for use in future management experiments that will be designed to study etiology, epizootiology, and prevention or control of disease outbreaks in bighorn populations: METHODS AND MATERIALS In conjunction with numerous cQoperators, we summarized data gathered to ~ate from a series of cooperative studies conducted in-Colorado and elsewhere to :devel-ppa simulation model (see details in WP2A,J8 Progress Report) and -- serol~gic and bacteriologic tools for use i~ studying and managing pasteurellosis in free-ranging bighorn popUlations; specific methods used in these projects have been described in detail elsewhere. Based on preliminary tindings generated from these efforts, we prepared a Program Narrative (Appendix A) and a Study Proposal (Appendix B) outlining additional experiments designed to provide data and to further develop diagnostic tools for eventual application in statewide bighorn management programs. Both plans were peer-reviewed and studies are in progress; one of these will be funded largely through Colorado Division of Wildlife Special Bighorn Sheep and Mountain Goat Auction and Raffle Funds available in FY91-92. Epizootiology of pasteurellosis in Colorado's indigenous Rocky Mountain bighorn sheep populations: We collected blood samples and nasal and pharyngeal swabs from indigenous bighorn herds trapped in January-March, 1991. Methods used to collect and handle field samples were as described in Appendix A, except that nasal swabs from 3 herds (Avalanche Creek, Tarrya11, Chalk Creek) were transported in modified Amies medium rather than Port-a-cul* tubes. _ Laboratory methods were as described in Appendix A. In a related activity, we also investigated an all-age pneumonia epizootic that occurred in the Taylor River-Almont Triangle bighorn herd. We performed postmortem examinations on 5 sheep carcasses and collected representative samples for diagnostic evaluation using established laboratory techniques. We also assisted in field operations to assess the extent and severity of this epizootic. ... .- Immunity to pasteurellosis in Rocky Mountain bighorn sheep: assessingspecificity of antibody responses to Pasteurelia haemolytica: We collected sera from free-ranging bighorns in conjunction with the preceding study. �105 Preparation of antigens from distinct phenotypic strains of P. haemolytica is currently underway, and serologic assays will be completed in FY9l-92 once funding becomes available. EXPERIMENTS ON MEASURING AND MANAGING STRESS IN BIGHORNS A proposal to evaluate stress responses of acclimated and wild free-ranging bighorn sheep to human disturbances in the Arkansas River corridor was developed as part of a comprehensive 3-year study to assess the impacts of recreational activities on bighorn population performance and habitat use in that area. Briefly, we plan to use both heart rate telemetry and fecal cortisol concentrations to compare stress responses between wild and bottleraised bighorns exposed to spontaneous and controlled human disturbances, and to evaluate management strategies for mitigating adverse responses (should they occur). The experimental approach proposed for use in this study is described in more detail in the 1990-91 Program Narrative for WP2A,J9. In planning that study, it became apparent that further development and understanding of fecal cortisol measurements were necessary in order to credibly apply these techniques to the Arkansas River study. In order to anticipate potential confounding influences on field applications of fecal cortisol measurements, we collected feces from captive bighorns at monthly intervals to examine seasonal influences on cortisol excretion. Samples were collected over a 3-5 day period at the-beginning of each month from individual captive bighorns of varied age/sex classes .. Feces were stored at -20 C until proce~sed for cortisol determination. Methods f~r processing feces and measuring cortisol were as described by MIller et al. (1991). RESULTS AND DISCUSSION The following abstracts summarize significant findings from a series of cooperative studies conducted in Colorado and elsewhere to develop tools for use in studying and managing pasteurellosis in free-ranging bighorn populations: REGULATION OF POPULATIONS BY INFECTIOUS DISEASE: SIMULATIONS OF PASTEURELLOSIS OUTBREAKS IN BIGHORN SHEEP (OVIS CANADENSIS). N. Thompso~ Hobbs and Michael W. Miller. Abstract: Many populations of bighorn sheep (Ovis canadensis) appear to be limited by their interactions with pathogens, notably Pasteurella spp., rather than by food supplies or predators. We developed a simulation model to represent interactions ~etween Pasteurella spp. and bighorn sheep populations. The model is an adaptation of the classical SIR system of coupled differential equations that has been widely applied to human diseases. Our model differs, however, in that its compartments (susceptible, infected, recovered) are represented as states within individual animals. Flows among compartments are translated into probabilities of transitions among states. We used this model to examine the idea that prevalent patterns in bighorn population dynamics can be explained by__intrinsic features of host-pathogen interactions. Averaged over several simulated populations, our individual-based model predicted an equilibrium virtually identical to that of the SIR model. However, because our model is inherently �106 stochastic, single populations showed a broad range of dynamical behavior including stasis, exponential growth, and episodic crashes in abundance. Thus, our model is faithful to 2 overriding patterns in bighorn population dynamics: regional suppression of animal numbers, and nonequilibrium behavior at local scales. We argue that both of these patterns can be explained by internal properties of host-pathogen interactions. External expla~ations, including habitat degradation and· fragmentation, environmental stress, parasitism, and contact with domestic animals are not necessary to explain epidemics and the ensuing population cycles frequently observed in bighorns. EPIZOOTIOLOGY OF PASTEURELLOSIS IN CAPTIVE ROCKY MOUNTAIN BIGHORN (OVIS CANADENSIS) LAMBS. Michael W. Miller, Margaret A. Wild, Kenneth W. Mills, Elizabeth S. Williams, and Amy Boerger-Fields. CANADENSIS Abstract: Poor lamb survival is often observed in Rocky Mountain bighorn sheep (Ovis canadensis canadensis) populations, particularly in the years following pneumonia epizootics. Causes for this reduced lamb recruitment are incompletely understood. Following an outbreak of pasteurellosis in a captive bighorn herd in 1984, 19 of 29 lambs born into that herd between 1985 and 1990 developed bronchopneumonia before 3 mo of age. We attributed all cases of pneumonia to nonhemolytic Pas~eurella haemoly~ica infections uncomplicated by protostrongylosis. Infection with P. haemoly~ica occurred shortly after birth. Three of 6 ewes sampled annually between 1988 and 1991 consistently -~hed Pa.~~eurella spp. in nasal ·secret;ionsat p~rttirition. Nonhemolytic P. haemoly~ica was not isola-ted from nasal or pharyngeal swabs of·n~wbo~ ~ambs (524 hrs old; n-22) , but was recovered from both sites in lambs ~ 5 days old. Lamb-to-lamb transmission also appeared to contribute to epizootic spread of Pas~eurella spp. Overall, 30 of 30 lambs sampled over a 4-yr period were infected by mid-summer. Biotype T. serotype 4, and untypable catalase positive isolates of nonhemolytic P. haemolytica were predominantly recovered from both sick and healthy lambs. An enzyme-linked immunosorbent assay measuring levels of antibodies to nonhemolytic P. haemoly~ica. biotype T, serotype 4, in colostrum from ewes and serum from lambs revealed the half-life of passive antibody levels in bighorn lambs was about 3 wks. Passive antibodies may have afforded lambs temporary protection from disease; pneumonia tended to occur in lambs >3 wks old. Our observations suggest that neonatal pasteurellosis may have profound effects on health and survival of b Lgho rn lambs. AN ENZYME-LINKED IMMUNOSORBENT ASSAY FOR DETECTING ANTIBODIES TO PASTEURELLA HAEHOLYTICA IN BIGHORN SHEEP (OVIS CANADENSIS). Ken Mills, Amy Boerger-Fields, Michael Miller, Margaret Wild, and Beth Williams. Abstract: Periodic pasteurellosis outbreaks cause extensive mortality in bighorn sheep (Ovis canadensis) populations, and may limit bighorn abundance throughout North America. It follows that tools for detecting presence of or exposure to Pas~eurella spp. infections could enhance the efficacy of management efforts for bighorns. We developed an enzymelinked immunosorberit assay (ELISA) to detect antibody to P. haemoly~ica in bighorn sheep. Antigen was initially prepared from a 20-hr culture of nonhemolytic P. haemoly~ica, biotype T, serotype 4, isolated from a �107 captive bighorn. Blocking assays using the original antigen strain, other P. haemolytica serotypes, and other bacterial species demonstrated high type-specificity for this ELISA. Based on these results, we developed a multivalent ELISA measuring antibody levels to distinct strains of P. haemolytica isolated from bighorn sheep or domestic livestock. Antibody levels measured with this multivalent ELISA appeared to differ between two captive bighorn herds differing in histories of exposure to P. haemolytica. Our results suggest measuring antibodies to P. haemolytica by ELISA may have applications in studying the epizootiology of pasteurellosis in bighorn populations, as well as i~ guiding relocation and other bighorn management activities. USING RIBOSOMAL RNA GENE RESTRICTION PATTERNS IN DISTINGUISHING STRAINS OF PASTEURELLA HAEMOLYTICA FROM BIGHORN SHEEP (OVIS CANADENSIS). K. P. Snipes, K. P., R. W. Kasten, M. A. Wild, M. W. Miller, D. A. Jessup, R. L. Silflow, and T. E. Carpenter. Abstract: Pasteurella haemolytica isolates (n - 31) from 2 captive herds of Rocky Mountain bighorn sheep (Ovis canadensis canadensis) were characterized using established phenotyping and newer genomic fingerprinting methods. Phenotypes of isolates varied both within and between source herds. Nine different pheno cype.swere identified; biotype T and serotypes 4 and 3,4,10 predominated. Most isolates (23 of 31) were nonhemolytic, and all hemolytic isolates were from one herd. Three phenotypes (nonhemolytic T4; T3;4; and T3,4,10) were common to 'both herds. Evaluation of 14 restriction endonucleases revealed that· EcoRI, HincII, and PstI yielded optimal DNA fragmentation patterns for fibotyping P. haemolytica isolates using a 32P-labeled Escherichia coli rRNA probe. Ribotype pattern appeared to be a relatively stable trait: an identical ribotype was conserved within and across 19 passages of a stock P. haemolytica strain. Among the 31 bighorn isolates, ribotyping produced 6 subjectively distinct patterns; probing after digestion with EcoRI, HincII, or PstI consistently assigned isolates to respective ribotype groups. In contrast to phenotype~, ribotypes appeared unique to each herd. Comparing phenotypic and genotypic traits among isolates revealed associations of multiple serotypes with a single ribotype and mUltiple ribotypes with a single serotype, as well as presence and absence of hemolysis among otherwise identical isolates. These findings suggest ribotyping may be a useful adjunct to other bacteriology methods in.studying the epizootiology of pasteurellosis in bighorn sheep. EXPERIMENTAL AND FIELD EVALUATION OF PORT-A-CUL® TRANSPORT TUBES FOR RECOVERY OF PASTEURELLA HAEMOLYTICA FROM BIGHORN SHEEP (OVIS CANADENSIS) PHARYNGEAL SWABS. Margaret A. Wild, Michael W. Miller, Terry R. Spraker, and William J. Adrian. Abstract: Three pharyngeal swabs were collected from each of 25 healthy captive bighorn sheep (Ovis canadensis). We recovered nonhemolytic Pasteurella haemolytica from 23 of 25 (92%) swabs streaked onto blood agar plates and incubated immediately, from 16 of 25 (64%) swabs held in Port-A-Cu1® transport tubes for 24 hr, and from 1 of 25 (4%) swabs held in Port-A-Gul® tubes for 48 hr. Although the recovery rate from swabs held in Port-A-Cu1® tubes for 24 hr was only about 70% of that from direct swabs, rates were markedly higher than those for other transport �108 media. We subsequently used Port-A-Cul® tubes to transport pharyngeal swabs from healthy, free-ranging bighorns trapped throughout Colorado in winter 1990-91. Nonhemolytic P. haemolytica was isolated from 7 of 8 herds sampled; within those herds, isolation rates ranged from l7-87X of individuals sampled. Isolation rates for P. haemolytica exceeded rates in previous years when pharyngeal (1989-90) or nasal (1988-89) swabs were transported in modified Amies with charcoal, suggesting our previous sampling efforts may have underestimated prevalence of P. haemolytica infections in free-ranging herds. Using Port-a-cul® tubes and minimizing time between sample collection and processing (~24 hr) appears to be a practical way to optimize recovery of P. haemolytica from bighorn pharyngeal swabs. Complete manuscripts describing these studies and their results are in prep (Miller et al., Mills et al., Wild et al.) or are in review (Hobbs and Miller, Snipes et al.) for publication in peer-reviewed professional journals or proceedings. Epizootiology of pasteurellosis in Colorado'S indigenous Rocky Mountain bighorn sheep populations: One hundred forty-two wild bighorns from 5 indigenous herds were sampled in conjunction with capture operations during January-March, 1991. We examined Pasteurella spp. isolates (n-5l) from 4 of 5 bighorn herds sampled (Taylor River, Avalanche Creek, Tarryall Mountains, Chalk Creek); failure to recover Pasteurella spp. from 19 samples collected at Cottonwood Creek was likely due to delayed laboratory processing. Prevalence estimates based on cultures of pharyngeal swabs ranged from l7-83X· (mean 49X) among sampled herds (Fig. 1). Based on serotyping data from isolates we . examined,·there appeared to be differences in strains of P. haemolyti~a _ endemIc to these herds. In all, we encountered at least 6 distinct phenotypes (T3; 4; 3,4; 3,4,10; A7; and untypable). among the 4 herds. We recovered T3 isolates from all of these, but prevalence of this strain appeared to vary among herds (Fig. 1). Representative isolates from each herd have been submitted for genomic fingerprinting, and sera will be submitted for ELISA to examine exposure to a battery of P. haemolytica strains. We plan to use comparisons of genotypic and phenotYJ1ic traits of Pasteurella spp. among indigenous herds, along with serological evidence of exposure, to improve our understanding of the epizootiology and management of pasteurellosis in Colorado'S bighorn populations. The pneumonia epizootic in the Taylor River-Almont Triangle population occurred in the face of one of the more intensive bighorn herd management programs in Colorado; that program includes ongoing lungworm treatment and popUlation control programs, as well as extensive habitat improvement and protection efforts. Coughing rams were first observed in mid-March; further field observations revealed signs of respiratory disease in bighorns of all age/sex classes scattered throughout the winter range. To date, about 17 bighorns are known to have died; necropsies of 5 dead sheep (2 rams, 3 ewes) revealed subacute to chronic bronchopneumonia. Nonhemolytic P. haemolytica was isolated from 4 of these (the fifth was too decomposed to attempt bacteriology). At least 4 distinct serotypes (T3; 4; 10; and 3,4,10) were identified; 3 of these were from lungs of different individuals. Horaxella spp. and Actinomyces pyogenes were also isolated from some affected individuals. Immunoperoxidase staining of lung tissue revealed presence of bovine respiratory syncytial virus in 1 ewe; other viral or chlamydial agents have not been detected to date. Based on gross, histologic, and fecal examinations, lungworm did not appear to playa role in this epizootic. The duration of this outbreak prior to its detection remains unknown: although �109 prevalence of Pasteurella spp. was lowest among herds sampled this year, at least 1 yearling ram transplanted from the Taylor River-Almont Triangle herd to Glenwood Canyon in January died after developing pasteurellosis in late March; the fate of 39 other sheep removed from the Taylor River-Almont Triangle area remains unknown. We will continue working with Southwest Region Personnel to study this outbreak and its aftermath. EXPERIMENTS ON MEASURING AND MANAGING STRESS IN BIGHORNS Preliminary evaluation of data from monthly collections of feces from captive bighorns suggested "reproductive status may influence fecal cortisol excretion in bighorn ewes (Fig. 2); by May, fecal cortisol concentrations from pregnant ewes (mean±SE-39.S±6.9 ng/g dm) were about 77% higher than those from open ewes (22.2±3.0 ng/g dm)(P<O.OS). Based on these findings, further evaluation of seasonal and other influences on fecal cortisol measurements is warranted before these techniques are applied to studies of stress responses in freeranging bighorns. LITERATURE CITED Miller, M. W. 1988. Experiments toward detecting and managing stress in ROCky Mountain bighorn sheep (Ovis canadensis canadensis). Ph.D. Thesis, Colorado State University, Fort Collins. 106pp. Miller, M. W., "N. T. Hobbs, and M. C. -Sous a , 1991. Detecting stress responses in Rocky Mountain bighorn sheep (Ovis canadensis canadensis): reliability of cortisol conceritrations in urine and feces. Can.-J. Zoo1. 69:15-24. > Miller, M. W., N. T. Hobbs, and E. S. Williams. 1991. Spontaneous pasteurellosis in captive rocky mountain bighorn sheep (Ovis canadensis canadensis): clinical, laboratory, and epizootio10gica1 observations. J. Wi1d1. Dis. 27: in press. Wildlife Researcher ��111 Appendix A PROGRAM NARRATIVE State of Colorado Project No. W-1S3-R-4 Mammals Research 2A Mountain 4 Experiments to Identify and Manage Stress in Mountain Sheep Work Plan Job No. No. Sheep Investigations EPIZOOTIOLOGY OF PASTEURELLOSIS IN COLORADO'S INDIGENOUS ROCKY MOUNTAIN BIGHORN SHEEP POPULA~IONS A. NEED Successful bighorn sheep management appears depends on mitigating or eliminating pneumonia epizootics in otherwise thriving herds. Periodic pneumonia outbreaks cause extensive mortality in bighorn sheep populations throughout North America (Buechner 1960, Forrester 1971). Viral, bacterial, and parasitic agents have all been incriminated in these outbreaks. spp. are perhaps the most common pathogens associated with bronchopneumonia in bighorns. Two species (P. haemolytica and P. multocida), and several biotypes and serotypes within those species, have been isolated from bighorns during die-offs in the last decade (Feuerstein et ale 1980, Wishart et ale 1980, Foreyt and Jessup 1982, Onderka and Wishart 1984, Spraker et ale 1984, Schwantje 1986, Coggins 1988, Festa-Bianchet 1988, Onderka and Wil;Jhart1988., Onderka et al. 1988, Foreyt 1989)." A nonhemolytic variant of P. haemolytica biotype"T that may be .endemic in some bighorn herds (Onderka and Wishart 1988, Onderka et ale 1988) has been linked to a pandemic that affected bighorn populations across southwestern Canada in the early 1980's (Onderka and Wishart 1984, 1988). Experimental infections with other strains of P. haemolytica isolated from healthy domestic sheep have shown marked pathogenicity in bighorns (Onderka and Wishart 1988, Onderka et ale 1988, Foreyt 1989), supporting field reports linking bighorn and domestic sheep interactions with pneumonia in bighorns (Goodson 1982, Foreyt and Jessup 1982, Coggins 1988). Pasteurella Epizootiology of pasteurellosis in bighQrn populations is poorly understood. A sound unifying hypothesis explaining the ecological basis of pneumonia in bighorn sheep has not emerged (Miller 1988, Miller et ale 1991). The relative importance of endemic and introduced disease agents as factors limiting bighorn abundance remains unclear (Miller et ale 1991, Wild and Miller 1991). In the absence of knowledge about the epizootiology of pasteurellosis, effective strategies for managing pneumonia in bighorn populations have not emerged. Wildlife managers can neither predict nor prevent outbreaks in bighorns. consequently, long-term attempts to manage bighorns often fail. In order to effectively manage pneumonia in bighorn sheep, we must begin to unravel and understand complicated ecological relationships involving disease agents, wild and domestic hosts, and environmental factors affecting both. To begin doing this, we need to compare distribution, prevalence, and transmission of specific pathogens among and within bighorn populations, compare individual and population responses to specific pathogens, and assess exposure history and status of bighorn herds with respect to various disease agents. Because pasteurellosis prevails as an important component of pneumonia outbreaks in bighorn sheep," we propose to sample indigenous bighorn herds to meaeuze and compare distribution, prevalence, and immunity to Pasteurella spp. infections. �112 B. OBJE~IVES Specific objectives of this study are to: (1) compare rates for tonsillar carriage and nasal shedding of Pasteurella spp. among 10 indigenous bighorn herds, as well as among ewe, ram, and lamb subpopulations within those herds, (2) compare isolated strains of P. haemolytica among and within sampled herds using phenotypic and genotypic characterizations of isolates, (3) compare serum antibody titers to Pasteurella within sampled herds, and (4) examine correlations between infection rates, P. haemolytica strains, herd immunity, and population performance (historic and current) in sampled bighorn populations. C. spp. among and EXPECTED RESULTS AND BENEFITS Managing the bighorn pneumonia complex is essential to improving success of management programs for Rocky Mountain bighorn sheep. By understanding how Pasteurella spp. persist in, spread through, and are distributed among bighorn populations, specific management practices can be directed toward preventing pneumonia outbreaks. Knowledge of distribution and prevalence of various Pasteurella spp. strains will aid in detecting and/or preventing introduction of novel strains to susceptible bighorn populations. Determining relationships between antibody levels and susceptibility or resistance to pasteurellosis will be useful in developing approaches for predicting pneumonia outbreaks ~hrough herd moni~oring. Incorporating these estimates for pa~ameters describing bighorn pasteurellosis-into simulation models (eg. Hobbs et al. 1990) will allow managers to explore potential consequences of management actions with respect to disease outbreaks, and to examine -alternatives for controlling pneumonia outbreaks. Moreover, analyzing strain and immunity data could lead to development of vaccines preventing pasteurellosis in bighorns. Ultimately, we believe developing approaches for estimating these parameters will provide a foundation for herd health monitoring and management as part of comprehensive management programs for bighorn sheep in Colorado and elsewhere. D. APPROACH 1. Hypotheses Rates for tonsillar carriage and nasal shedding of Pasteurella spp. will not differ among or within indigenous bighorn herds sampled. Phenotypic and genotypic traits of P. haemolytica will not differ among-or within herds. Serum antibody titers to Pasteurella among or within herds. isolates spp. will not differ 2. Methods a. Study herds - We will sample 10 indigenous herds of Rocky Mountain bighorn sheep (Ovis canadensis canadensis) over a 2-year period. Five of these (Alamosa Canyon, Chalk Creek, Ouray, Taylor River, and Waterton Canyon) have experienced documented pneumonia outbreaks within the last 10 years; 5 others (Avalanche Creek, Cottonwood Creek, Sangre De Cristo Mountains, Tarryall MQ~ntains, and Trickle _ Mountain) have -been free -of epizootic- pneumonia for at lea-at 20 years. Estimates of population performance and composition for each herd will be obtained from existing inventory data generated by regional management programs. �113 E. F. b. Sample collection - Based on current population estimates, we will calculate target sample sizes for each herd necessary to provide prevalence estimates within ± 10% of true prevalence at 95% confidence level (DiGiacomo and Koepsell 1986). Samples for bacteriology and serology will be collected from each herd in conjunction with live capture operations and/or hunter surveys. Nasal and tonsillar swabs from individually identified bighorns will be plated directly onto bighorn sheep blood agar and MacConkey's agar and incubated immediately (Wild and Miller 1991, and unpubl. data). We will also collect 12 ml of blood from each sheep sampled; serum will be separated and stored at -20 C. c. Characterization of P. haemolytica isolates - Pasteurella spp. colonies will be isolated and identified from primary culture plates using established methods (Carter 1984). All Pasteurella spp. isolates will be lyophilized. Those biochemically identified as P. haemolytica (regardless of hemolysis reactions) will be serotyped using both rapid plate agglutination (Frank and Wessman 1978) and indirect hemagglutination (Biberstein et ale 1960). We will further characterize P. haemolytica isolates using genomic fingerprinting (Snipes et ale in review). d. Serology - Serum antibodies to Pasteurella spp. will be measured using an enzyme-linked immunosorbent assay (Gillette et ale 1989, Mills et ale in prep). We will use western blot analysis to examine strain specificity of bighorn antibodies to Pasteurella spp. among and within sampled herds. ANALYSIS We will calculate prevalence rates for tonsillar carriage and nasal shedding of Pasteurella spp. and compare these among herds and among ewe, ram, and lamb subpopulations within herds using'categorical m9deling. We .will use Cochran's Q test to compare distributions of phenotypically and/or genotypically distinct strains of P. haemolytica among and-within herds. Serum antibody levels to Pasteurella spp. will be compared by analysis of variance with herd, subpopulation, and health histories as main effects. SCHEDULE February 1991 - April 1993 May 1991 - July 1993 Collect samples from indigenous herds, culture and identify Pasteurella spp. isolates. Perform genomic fingerprinting on P. isolates, measure serum antiboby levels and examine strain specificities. haemolytica August - December January 1993 - May 1994 G. PERSONNEL Michael W. Miller' Margaret A. Wild' Kurt P. snipes2 Colorado Prospect 2 Analyze data. Prepare manuscripts for publication in peer-reviewed journals and offer management recommendations based on data generated. Principal Investigator co-principal Investigator Co-Principal Investigator Division of Wildlife, Wildlife Research Road, Fort Collins, ·Colorado 80526. ' center,- 317 West Department of Epidemiology and Preventive Medicine, School of Veterinary Medicine, University of California, Davis, California 95616. �114 Kenneth W. Mills3 H. Co-Principal BUDGE~ Personal Services Operating Supplies and Services Bacteriology Serotyping DNA Typing Serology Total Operating Travel Expenses Capital Expenditures TOTAL I. Investigator LI~RATtJRE $ 40,000 $ $ $ S $ 5,000 2,500 5,000 2,500 15,000 3,000 2,000 60,000 CI~D Biberstein, E. L., M. G. Gills, and H. D. Knight. 1960. Serologic types of haemolytica. Cornell Vet. 50:283-300. Buechner, H. K. 1960. The bighorn sheep in the United States, its past, present, and future. Wildl. Mono. 4:74. Carter, G. R. 1984. Diagnostic procedures in veterinary bacteriology and mycology. C. C. Thomas, Springfield, Ill., 515 pp. Coggins, V. L. 1988. The Lostine Rocky Mountain bighorn sheep die-off and domestic sheep. Bienn. Symp. Northern Wild Sheep and Goat Counc. 6:57-64. DiGiacomo, R. F., and T. D. Koepsell. 1986. Sampling for detection of infection or disease in animal populations. J. Am. Vet. Med. Assoc. 189: Pasteurella 22-23. Festa-Bianchet, M. 1988. A pneumonia epizootic in bighorn sheep, with comments on preventive management. Bienn. Symp. Northern Wild Sheep and Goat Counc. 6:66-76. . Feuerstein. V., R. L. Schmidt, C. P. Hibler, and W. H. Rutherford. 1980. BighQrn sheep mortality in the Taylor River-Almont Triangle area, 1978-1979: a case study. Colo. Div. of Wild., Spec. Rep. No. 48. l~-pp. Foreyt, W. J. 1989. Fatal Pasteurella haemolytica pneumonia in bighorn sheep after direct contact with clinically normal domestic sheep. Am. J. Vet. Res. 50:341-343. Foreyt, W. J., and D. A. Jessup. 1982. Fatal pneumonia of bighorn sheep following association with domestic sheep. J. Wildl. Dis. 18:163-169. Forrester, D. J. 1971. Bighorn sheep lungworm-pneumonia complex. pp. 158-173 in J. W. Davis; and R. C. Anderson, eds. Parasitic Diseases of Wildlife. Iowa State Univ. Press, Ames. Frank, G. -H., and G. E. Wessman. 1978. Rapid plate agglutination procedure for serotyping Pasteurella haemolytica. J. Clin. Microbiol. 7: 142-145. Gillette, K. G., G. H. Frank, and J. M. Sacks. 1989. Development and evaluation of an enzyme-linked immunosorbent assay for bovine antibody (IgG) to Pasteurella haemolytica. Am. J. Vet. Res. 50:106-110. Goodson, N. J. 1982. Effects of domestic sheep grazing on bighorn sheep populations: a review. Proc. Biennial Symp. Northern Wild sheep and Goat Counc. 3:287-313. Hobbs, N. T., M. W. Miller, J. A. Bailey, D. A. Reed, and R. B. Gill.. 1990. Biological criteria for introductions of large mammals: using simulation models to predict impacts of competition. Trans. N. Am. Wildl. Nat. Res. Conf. 55: 620-632. Miller, M. W. 1988. Experiments toward detecting and managing stress in Rocky Mountain bighorn sheep (Ovis canadensis canadensis). Ph.D. Thesis, Colorado State Univ, Ft. Collins. l06pp. Miller, M. W., No T. Hobbs, and E. So Williams. 1991. Spontaneous pasteurellosis in captive Rocky Mountain bighorn sheep (Ovis canadensis canadensis): clinical, laboratory, and epidemiological observations. J. Wildl. Dis. 27: in press. - 3 Wyoming state Veterinary Laboratory, Street, Laramie, Wyoming 820700 University of Wyoming, 1190 Jackson �115 Mills, K. M., M. W. Miller, E. S. Williams, and M. A. Wild. 1991. Development and evaluation of an enzyme-linked immunosorbent assay for measuring antibody" titers to Pasteurella haemolytica in Rocky Mountain bighorn sheep (Ovis canadensis canadensis). in prep. Onderka, D. K., and W. D. Wishart. 1984. A major bighorn sheep dieoff from pneumonia in southern Alberta. Bienn. Symp. Northern Wild Sheep and Goat Counc. 4: 356-363. Onderka, D. K., and W. D. Wishart. 1988. Experimental" contact transmission of Pasteurella haemolytica from clinically normal domestic sheep causing . pneumonia in Rocky Mountain bighorn sheep. J. Wildl. Dis. 24: 663-667. Onderka, D. K., S. A. Rawluk, and W. D. Wishart. 1988. Susceptibility of Rocky Mountain bighorn sheep and domestic sheep to pneumonia induced by bighorn and domestic livestock strains of Pasteurella haemolytica. Can. J. Vet. Res. 52: 439-444. " schwantje, H. M. 1986. A comparative study of bighorn sheep herds in southeastern British Columbia. Bienn. Symp. Northern Wild Sheep and Goat Counc. 5: 231-252. Snipes, K. P., R. W. Kasten, M. A. Wild, M. W. Miller, D. A. Jessup, R. L. Silflow, and T. E. Carpenter. " Using ribosomal RNA gene restriction patterns in distinguishing strains of Pasteurella haemolytica from bighorn sheep (Ovis canadensis). J. Wildl. Dis., in review. Spraker, T. R., C. P. Hibler, G. G. Schoonveld, and W. S. Adney. 1984. Pathologic changes and microorganisms found in bighorn sheep during a stress-related die-off. J. Wildl. Dis. 20:319-327. Wild, M. A., and M. W. Miller. 1991. Detecting nonhemolytic Pasteurella haemolytica infections in healthy Rocky Mountain bighorn sheep (Ovis canadensis canadensis): influences of sample site and handling. J. Wildl. Dis. 27: in press. " Wishart, W. D., J. Jorgenson, and M. Hinton. 1980. A minor die-off of bighorns from pneumonia in southern Alberta. Bienn. Symp. Northern Wild Sheep and Goat Counc. 2: 229-247. ��117 Appendix B STUDY PROPOSAL IMMUNITY TO PASTEURELLOSIS IN ROCKY MOUNTAIN BIGHORN SHEEP: ASSESSING SPECIFICITY OF ANTIBODY RESPONSES TO PASTEURELLA HAEHOLYTICA Michael W. Miller1, Kenneth W. Mills2, Amy Boerger-Fields2, Margaret A. Wild1, and Elizabeth S. Williams2, Principal Investigators 1 Colorado Division of Wildlife, Wildlife Research Center, 317 West Prospect, Fort Collins, Colorado 80526. 2 Wyoming State Veterinary Laboratory, University of Wyoming, 1174 Snowy Range Road, Laramie, Wyoming 82070. _ A. NEED Long-term success of bighorn management depends on our ability to manage or eliminate pneumonia epizootics in otherwise thriving herds. Periodic pneumonia outbreaks cause extensive mortality in bighorn- sheep populations, and limit bighorn abundance throughout North America (Buechner 1960, Forrester 1971). Various viral, bacterial, and parasitic agents have been incriminated in these out1::lreaks historically (Spraker and,Hibler 1982), but Past:eurella spp_ are perh~ps the most common pathogens associated with bronchopneumonia in bighorns. Two species (P. haemolyt:ica and P. mult:ocida) and several '~strains" (biotjpes and serotypes) of Past:eurella spp. have been isolated from bighorns during dieoffs in the last decade (Feuerstein et al. 1980, Wishart et al. 1980, Foreyt.and Jessup 1982, Onderka and Wishart 1984, Spraker et al. 1984, Schwantje 1986, Coggins 1988, Festa-Bianchet 1988, Onderka and Wishart 1988, Onderka et al. 1988, Foreyt 1989). Among these, a non-hemolytic vari~nt of P. haemolyt:ica biotype T may be endemic in bighorn herds in Colorado and elsewhere (Onderka and Wishart 1988, Onderka et al. 1988, Miller et al. 1991, M. W. Miller, unpubl. data); this agent appears responsible for the pandemic that affected bighorn populations from southern British Columbia an~ Alberta to Montana and Idaho 5-6 years ago (Onderka and Wishart 1984, 1988). Other strains of P. haemolyt:ica isolated from healthy domestic sheep have shown marked pathogenicity in experiments using captive bighorns (Onderka and Wishart 1988, Onderka et al. 1988, Foreyt 1989), supporting field reports linking bighorn and domestic sheep interactions with pneumonia in bighorns (Goodson 1982, Foreyt and Jessup 1982, Coggins 1988). Despite compelling data from experiments and retrospective field studies, the relative importance of endemic and introduced disease agents as factors limiting bighorn abundance remains unclear. Because many epidemiological .aspects of the bighorn pneumonia complex are poorly understood, wildlife managers can neither predict nor prevent pneumonia in bighorns. Consequently, long-term attempts to manage bighorns often fail. In order to effectively manage pneumonia in bighorn sheep, we must begin to unravel and understand complicated ecological relationships involving disease agents, wild and dC?_mestichos~s, and environmental factors affecting both. To do this, we need tools for measuring individual and population responses to specific pathogens, �118 as well as for assessing exposure history and status of bighorn herds with respect to various disease agents. Serological tests provide this type of information by detecting or measuring antibodies induced by host/pathogen interactions. Antibody levels indicate exposure to specific pathogens, and may also be correlated with disease . resistance in some cases (reviewed by Confer 1988 for bovine pasteurellosis). Data of this nature could be invaluable in efforts to clarify many aspects of the bighorn pneumonia complex. Over the past two years we have developed an enzyme-linked immunosorbent assay (ELISA) that measures antibodies to Pasteurella spp. in bighorn sheep (Mills et al., 1991). Preliminary data generated using this ELISA have already provided insights into the epizootiology of pasteurellosis in bighorn lambs (Mills et al. 1991, M. W. Miller, unpubl. data). We propose to use-sodium dodecyl sulfatepolyacrylamide gel electrophoroesis (SDS-PAGE) and western blot analyses to further develop our assay. This approach will allow ELSIA_measurement of antibody levels to specific strains of P. haemolytica isolated from freeranging bighorn herds throughout Colorado. Data from these analyses will be used in comparing performance of bighorn herds exposed to different strains of P. haemolytica. Such data will also be used in evaluating health of potential donor herds as a means of minimizing introduction of novel pathogens to recipient herds through the statewide bighorn transplanting program. B. OBJECTIVES Spec Lf Lc objectives of this study ,are to: (i) use SDS-PAGE and western blot ana~yses to separate and identify strain-specific antigenic proteins from 5 strains of P. haemolytica isolated from Rocky Mountain bighorn sheep, (2) modify the Pasteurella spp. ELISA using a battery of strainspecific antigens, and (3) compare levels of antibody to Pasteurella spp. in sera from 5 geographically distinct bighorn herds using the modified ELISA. C. EXPECTED RESULTS AND BENEFITS Managing the bighorn pneumonia complex is essential to improving success of management programs for Rocky Mountain bighorn sheep. By understanding how respiratory pathogens persist in and spread through bighorn populations, management practices can be directed toward preventing pneumonia outbreaks -developing assays that measure antibodies to specific disease agents in bighorns appears pivotal in gaining this knowledge. Techniques for measuring bighorn antibodies to specific strains of Pasteurella spp. will allow biologists and researchers to document, study, and perhaps differentiate between pneumonia epizootics in bighorns caused by endemic and introduced strains. Data from studies of relationships between antibody levels and susceptibility or resistance to pasteurellosis will be useful in developing approaches for predicting pneumonia outbreaks through herd monitoring and/or simulation modeling; moreover, these data could lead to development of vaccines preventing pasteurellosis in bighorns. Ultimately, we believe these and other similar tools will provide a foundation for herd health monitoring �119 and management as part of comprehensive management programs for bighorn sheep in Colorado and elsewhere. D. APPROACH 1. Hypothesis Ho: 2. Antibody levels to strain-specific P. haemolytica antigens will not differ among or within geographically-distinct sheep herds. Methods Techniques described elsewhere (Ellis et al., 1991) will be used in SDSPAGE and western blot analyses to separate and identify strain-specific antigenic proteins from P. haemolytica isolates recovered from Rocky Mountain bighorn sheep. Generally, we will use SDS-PAGE to separate bacterial proteins from 5 phenotypically-distinct strains of P. haemolytica isolated from free-ranging Rocky Mountain bighorn sheep. Separated proteins will be transferred directly to nitrocellulose filters. We will then react bighorn sheep sera to these proteins to determine antigen specificity and activity of antibodies to the various P. haemolytica strains. Once unique strain-specific antigens have been identified and produced, we will incorporate them into the existing Pasteurella spp. ELISA (Mills et al., 1991). We will then measure antibody leve-ls to these 5 P". haemolytica strains 'Ln sera from 5 _ geographically:distinct bighorn-herds using ELISA techniques described-by Mills et al. (1991). All antibody levels will be reported as optical density (OD) readings. E. ANALYSIS Antibody levels measured in bighorn sera will be compared using two-way analysis of variance, with P. haemolytica strain and herd as main effects. All analyses will be performed using the SAS program for general linear models (SAS 1987). F. SCHEDULE January-March 1991 July-August 1991 Collect sera and P. haemolytica isolates from freeranging bighorns in conjunction with ongoing studies. Perform SDS-PAGE and western blot analyses on P. isolates, identify and produce unique antigens. haemolytica September-October 1991 Modify, test, and revalidate ELISA. November-December January-March 1992 1991 Measure antibody levels to sera from free-ranging bighorn herds using modified ELISA. Analyze and summarize data, prepare manuscripts for publication in peer-reviewed journals. �120 G. BUDGET Personal Services Laboratory Technician, 1 mo Operating Expenses SDS-PAGE and western blot analyses Antigen purification and production ELISA determinations TOTAL· * $ 1000 2000 1000 1000 $ 5000 (Amount requested from Special Bighorn Sheep and Mountain Goat Auction and Raffle Funds. In addition, CDOW will provide support for collecting and preparing serum and bacterial samples, and will pay publication charges; WSVL will provide miscellaneous supplies and capital equipment needed to run assays, and will waive administrative charges. Investigators' salaries will be paid by their respective agencies.) H. LITERATURE CITED Buechner, H. K. 1960. The bighorn sheep in the United States, its past, present, and future. Wi1dl. Mono. 4:74. Coggins, V. L. 1988. The Lostine Rocky Mountain bighorn she~p die-off and domestic sheep. Bienn. Symp. Northern Wild Sheep and Goat Counc. 6:57-64. Confer ,.A. W. 1988.· Bovine pneumonic pasteurellosis': Immunity to . Pasteurella liaemolyiica. J. Am. Vet. Med. Assoc. 193: 1.308~13l6. Ellis, J. A., D. A. Hawk, K. W. Mills, and D. L. Pratt. 1991". Ant Lgen specificity and activity of ovine antibodies induced by immunization with Corynebacterium pseudotuberculosis culture filtrate. Vet. Immuno1. Immunopath. in press. Festa-Bianchet, M. 1988. A pneumonia epizootic in bighorn sheep, with comments on preventive management. Bienn. Symp. Northern Wild Sheep and Goat Counc. 6:66-76. Feuerstein, V., R. L. Schmidt, C. P. Hibler, and W. H. Rutherford. 1980. Bighorn sheep mortality in the Taylor River-Almont Triangle area, 1978-1979: a case study. Colo. Div. of Wild.·, Spec. Rep. No. 48. 19 pp. Foreyt, W. J. 1989. Fatal Pasteurella haemolytica pneumonia in bighorn sheep after direct contact with clinically normal domestic sheep. Am. J. Vet. Res. 50:341-343. Foreyt, W. J., and D. A. Jessup. 1982. Fatal pneumonia of bighorn sheep following association with domestic sheep. J. Wi1dl. Dis. 18:163-169. Forrester, D. J. 1971. Bighorn sheep lungworm-pneumonia complex. pp. 158-173 in J. W. Davis, and R. C. Anderson, eds. Parasitic Diseases of Wildlife. Iowa State Univ. Press, Ames. Goodson, N. J. 1982. Effects of domestic sheep grazing on bighorn sheep populations: a review. Proc. Biennial Symp. Northern Wild sheep and Goat Counc. 3:287-313. Mills, K. M., M. W. Miller, E. S. Williams, and M. A. Wild. 1991. Development and evaluation of an enzyme-linked immunosorbent assay for measuring levels of antibody to Pasteurella haemolytica in Rocky Mountain bighorn sheep (Ovis canadensis canadensis). in prep. Onderka, D. K., and W. D. Wishart. 1984. A major bighorn sheep dieoff from pneumonia in southern Alberta. Bf erm Symp. Northern Wild Sheep and Goat Counc. 4: 356-363. �121 Onderka, D. K., and W. D. Wishart. 1988. Experimental contact transmission of Pasteurella haemolytica from clinically normal domestic sheep causing pneumonia in Rocky Mountain bighorn sheep. J. Wildl. Dis. 24: 663-667. Onderka, D. K., S. A. Rawluk, and W. D. Wishart. 1988. Susceptibility of Rocky Mountain bighorn sheep and domestic sheep to pneumonia induced by bighorn and domestic livestock strains of Pasteurella haemolytica. Can. J. Vet. Res. 52: 439-444. SAS. 1987. SAS/STAT Guide for personal computers, version 6 edition. SAS Institute, Cary, North Carolina. 1028pp. Schwantje, H. M. 1986. A comparative study of bighorn sheep herds in southeastern British Columbia. Bienn. Symp. Northern Wild Sheep and Goat Counc. 5: 231-252. Spraker, T. R., and C. P. Hibler. 1982. An overview of the clinical signs, gross and histological lesions of the pneumonia complex of bighorn sheep. Bienn. Symp. Northern Wild Sheep and Goat Counc. 3:163-172. Spraker, T. R., C. P. Hibler, G. G. Schoonveld, and W. S. Adney. 1984. Pathologic changes and microorganisms found during a stress-related die-off. J. Wildl. Dis. 20:119-327. Wishart, W. D., J. Jorgenson, and M. Hinton. 1980. A minor die-off of bighorns from pneumonia in southern Alberta. Bienn. Symp. Northern Wild Sheep and Goat Counc. 2: 229-247. �122 .---. ~ 0 - 100 _ ..._... 80 W T3 DT4 f2SCI 0 Z 60 ...I 40 > W 20 OTHER T oOTHER W <C x"49 H A I:C Q. 0 AC TR TA CC HERD .-..E "0 -- 50 C') C') c ....,_, ..J 0 40 30 en -III: -.- PREGNANT -.- OPEN I RUT I /i 'I T 20 ~~_-_,t' __ 0 o ...I <C 10 W 0 c LL ---.•.. - --- rOCT NOV DEC JAN /,1 T -' FEB MAR MONTH T I ,,'''1 I l.AMBING I APR MAY �123 Colorado.Division of Wildlife Wildlife Research Report July 1991 JOB PROGRESS REPORT State of ~C~o~l~o~r~a~d~o~ _ Project No. W-153-R-4 Mammals Research Work Plan No. 2A Mountain Sheep Investigations Job No. 7 Experimental Evaluation of Mountain Sheep Transplanting and Disease Treatment Period Covered: Authors: Personnel: July I, 1990 - June 30, 1991 M. W. Miller, C. Anderson, Jr., and J. Vayhinger D. A. Reed, S. Ogilvie, C. Anderson, Sr. Abstract To foster the concept of approaching statewide transplanting and disease treatment programs for mountain sheep populations as management experiments, we prepared a research·prospectus outlining a management experiment to·evaluate ·and compare management strategies for reducing lungworm burdens a~d improving lamb ~urvival in selected mountain sheep herds. We proposed to evaluate alternative lungworm treatment regimes in a 4-year management experiment with l~tin square design, using mountain sheep populations in the Tarryall Mountains and Collegiate Peaks to evaluate the following treatment·regimes: a) baiting with alfalfa hay and apple pulp treated with fenbendazole (bait/treat); b) baiting with alfalfa hay and apple pulp without fenbendazole (bait/no treat); c) placing fenbendazole-treated salt blocks on bait stations (no bait/treat); ·and d) withholding bait and fenbendazole (no bait/no treat). Treatments will be rotated among herds ,such that at the end of 4 years each site will have received all 4 treatments. We pLan to compare lainbsurvival among ewe/treatment groups as a measure of treatment efficacy using individual radioco1lared ewes (n-20/site) for our'observations. . During February-March, 1991, 50 adult ewes were captured and radiocollared at 4 sites, 2 in the Tarryall (Cotton Gordon's, Sugarloaf Mountain) and 2 in the Collegiate Mountains (Chalk Creek, Cottonwood Creek); both dropnets and chemical immobilization were used in captures. At least 2 ewes died as a direct result of capture, and 2 others have died since capture. Preliminary pretreatment observations revealed that 31 of 46 surviving ewes (67%) were observed either lactating (n-ll) or nursing lambs (n-20) during May and/or June 1991; 6 ewes showed no evidence of lambing, and the status of 9 others remained undetermined as of late June. Only 1 known lamb loss, presumed a mortality, has been detected to date. Movements have varied widely both among and within monitored herds. ��125 EXPERIMENTAL EVALUATION OF MOUNTAIN SHEEP TRANSPLANTING AND DISEASE TREATMENT Michael W. Miller P. N. OBJECTIVES Design, conduct, and report on management experiments to evaluate efficacy of transplanting and disease treatment practices for managing mountain sheep populations. AGREEMENT OBJECTIVES Prepare a prospectus for management level experiments evaluating Colorado's mountain sheep transplanting and parasite control programs. Colorado's mountain sheep management program is recognized as among the most aggressive in North America. A combination of management practices, including transplanting to establish new herds and anthelmintic treatment to improve herd health, appears to have produced a threefold increase in bighorn numbers statewide over the last 20 years. However, the contributions of individual management strategies to the overall success of this program cannot be discerned (Bailey 1990). Transplanting bighorns to unoccupied ranges and treating resident herds to control lungworms are 2 of the most intensive (and costly) practices used to manage sheep in Colorado, yet neither of these vrograms has ever been evaluated experimentally to assess long-term efficacy. Recent" analyses of available herd data suggested that only about half of all transplant efforts succeed in establishing viable bighorn herds, and that parasite treatment does not necessarily prevent disease outbreaks or improve performance of treated bighorn herds (Bailey 1990). Unfortunately, management practices for most of Colorado's bighorn populations are so confounded that even interpretation of the aforementioned evaluation is equivocal. Clearly, management-level experiments are needed to evaluate and compare efficacy and efficiency o~ bighorn management practices in Colorado. The Colorado Division of Wildlife's Terrestrial Wildlife Research Section is committed to conducting management experiments to evaluate- efficacy of transplanting and disease treatment practices for managing mountain sheep populations. In conjunction with the Division's Southeast Regional staff, we have proposed the following management experiment to evaluate and compare management strategies for reducing lungworm burdens and improving lamb survival in selected mountain sheep herds: RESEARCH PROSPECTUS Improving Mountain Sheep Population Performance With Anthelmintic Treatment: An Experimental Evaluation of Management Alternatives Michael W. Miller, Jack Vayhinger, and Stan Ogilvie PRDBIEH Mountain sheep (Ovis canadensis canadensis) popula-tions throughout North America are plagued by pneumonia outbreaks caused by a complex of bacterial, �126 parasitic, and/or viral agents (reviewed by Spraker and Hibler 1982). These epizootics and subsequent population declines represent a significant obstacle to long-term success in mountain sheep management (Buechner 1960, Forrester 1971). Treating with anthe1mintics to reduce parasitic lungworm (Protostrongylus spp.) burdens has become an integral part an aggressive management program to reduce disease outbreaks and enhance productivity of mountain sheep herds in Colorado (Schmidt et al. 1979). Although some combination of treating, trapping and other management activities has increased sheep numbers statewide over the last 2 decades, it is unclear which strategies were most influential in this achievement (Bailey 1990). In light of the growing number of significant mountain sheep herds throughout Colorado, the apparent intensity of management required to perpetuate individual populations, and the costs associated with applying intensive management, it is becoming increasingly important to explore alternatives for efficient but efficacious management practices. Here, we propose a management experiment to examine effects of alternative lungworm treatment strategies on mountain sheep lamb survival and population performance. APPROACH We plan to evaluate alternative lungworm treatment regimes in a 4-year management experiment with latin square design. We will use mountain sheep populations in the Tarryall Mountains and Collegiate Peaks to evaluate treatment regimes. Four existing bait stations are already established within occupied ranges of the 3 distinct populations (Tarryall, Cottonwood Creek, Chalk Cliffs) that will be included in our study; 2 sites with distinct group fidelity (Sugarloaf Mountain, Twin Eagles) are established for the Tarryal1 herd.· Sites will be paired by location (Tarryall ~r ColIegiate) and average annual site attendance. At each site, we will initially capture, radiocollar, and release 15-20 adult ewes that will be used to measure treatment effects. Beginning in the following winter, we will impose one of the following management treatments at each site: a) baiting with alfalfa hay and apple pulp treated with fenbendazole (bait/treat); b) baiting with alfalfa hay and apple pulp without fenbendazole (bait/no treat); c) placing fenbendazole-treated salt blocks on bait stations (no bait/treat); and d) withholding bait and fenbendazole (no bait/no treat). In each succeeding winter, these treatments will be rotated such that at the end of 4 years each site will have received all 4 treatments. We will compare lamb survival among ewe/treatment groups as a measure of treatment efficacy. Radio-collared ewes with known treatment histories will be relocated every 2 weeks from 1 May to 30 September. We will observe ewes for presence or absence of a nursing lamb, and will also note condition, behavior, and apparent health of all lambs that are observed. We will also collect fecal pellet groups from lambs nursing marked ewes at monthly intervals. In addition to data on lamb survival and fecal lungworm levels, we will record group sizes, composition, and locations for all sheep observed. Differences in lamb survival and fecal lungworm levels within and among- ewe/treatment groups will be examined using rep~ate(Cmeasures analysis of variance. �127 SCHEDULE Activity Capture and radiocollar ewes 1991 Impose treatments and monitor responses 1992-96 Analyze data, publish results, and recommend management alternatives 1996-97 MATERIALS AND METHODS We began capturing bighorn ewes in preparation for our management-level experiment evaluating Colorado's mountain sheep parasite control program. During February and March, 1991, we used dropnets and/or chemical immobilization (about 2.6 mg carfentanil HCl and about 15 mg xylazine HCl/animal) to capture adult and yearling ewes at 4 sites in the Tarryall and Collegiate Mountains. Captured ewes were fitted with radiocollars (148.0-149.9 MHz) marked with color/symbol combinations unique to each capture site to allow visual identification of individuals; ewes were also eartagged using color/number combinations unique to each capture site. Radiocollared ewes were monitored irregularly to detect mortality and movements between February and May. To aid in planning further captures and in refining design of the field monitoring phase of this management experiment, pretreatment data were gathered from May to'present. Telemetered ewes were ·located and observed every 2-4 weeks. Location (by UTM coordinate) and .presence o~ absence of a -mrr s Lng lamb was recorded for each observa t Lon, All field data were transcribed into a computerized database to aid in mapping seasonal range movements and determining preliminary lamb survival rates. RESULTS AND DISCUSSION Fifty adult ewes were captured at 4 sites, 2 (Cotton Gordon's, Sugarloaf Mountain) in.the Tarryall and 2 (Chalk Creek, Cottonwood Creek) in the Collegiate Mountains (Table 1). At least 2 ewes died as a direct result of capture, and 2 others have died since capture. Further attempts to collar ewes were thwarted by unfavorable weather that precluded helicopter flights and by impending lambing seasons. Additional capture attempts are planned for late summer and/or for ~id-November after big game seasons close. Overall, 31 of 46 surviving ewes (67%) were observed either lactating (n-ll) or nursing lambs (n-20) during May and/or June; 6 ewes showed no evidence of lambing, and the status of 9 others remains undetermined (Table 1). Only 1 known lamb loss, presumed a mortality, has been detected to date. Movements have varied widely both among and within monitored herds. A summary of field observations follows: COLLEGIATE - Chalk Creek Ewe P was found dead about 200-300 m northeast of the Chalk Creek bait site on 15 May; she may have fallen from a cliff, and appeared to have died >2 weeks earlier. All 10 remaining collared ewes in this area have been seen with lambs and·appear to he lactatiD:g ("circle";· "square", "tria~gle", 1, 2, 4, 5, 6, 7, and E), and 6 of these ("square", 2, 4, 5, 6, and 7) have been observed nursing lambs. Ewe 2's lamb apparently disappeared in late �128 June -- she was observed on 10 July for 2.5 hours and was alone for the entire time. Six of these ewes ("square", "triangle", I, 5, 6, and E) currently inhabit Cascade Canyon. The other 4 ("circle", 2, 4, and 7) seem to travel back and forth from the Brown's Creek area to private salt licks near the Love Ranch Meadow. - Cottonwood Creek Five of the 13 collared ewes have been seen with lambs and appear to be _ lactating ("square", 1, 4, 9, and E), and 3 of these ("square", 4, and 9) have been observed nursing lambs. Four ewes (2, 5, 6, and 7) are not lactating and do not appear to have lambs (ewe 7 showed teats but no udder development in June, suggesting she may have lost her lamb before monitoring began). The status of 4 ewes ("circle", "triangle", 8, and H) is unknown; 3 of these ("circle", "triangle", and 8) have been observed in ewes/lambs groups, but their individual status co~ld not be determined. Ewe H has not been seen. The sheep in this area are quite mobile and travel in large groups of >30 individuals. They have been moving back and forth between Mount Yale and a natural salt lick near Cottonwood Hot Springs. TARRYALL - Cotton Gordon's Nine of 15 collared ewes in this group have been seen with lambs and appear to be lactating ("circle", "square", "triangle", I, 6, 7, K, U, and X); 5 of these ("square", I, 6, U, .and X) have been observed nursing lambs. None of these ewes appear to have Los t,iambs. to date. The other 6 collared ewes in this area have not been seen -with lambs (2, 4; 8, 9, M, and T). Ewes 8 and 9 are not lactating and do not appear to have lambs. Ewes M and T didnot have lambs with them when seen on 5 June, and ewe 4 did not have a lamb with her when seen on 7 June, but lactation status for these ewes remains undetermined. Ewe 2 (a 2-yr-old) has been seen with other ewes and lambs, but has not been observed lactating or nursing a lamb. It appears that most of these ewes lambed on south facing slopes northeast of the Tarryall River and Carpenter Ranches (Hay Creek area) in early June. After lambing, they moved into the Goose Creek drainage (approx. 1-3 km northeast of Goose Creek Trailhead) in mid-June, then moved back into the Hay Creek area at the end of June. - Sugarloaf Mountain Ewe "square" was found dead on 22 May about 1.2 km northeast of the Sugarloaf bait site. She was found under a tree, had died without apparent struggle, and appeared to have died about 1 to 2 weeks earlier. Six of the 8 remaining collared ewes in this group appeared to be lactating and have been seen nursing lambs ("circle", ~triangleW, I, 4, 5, and R). Ewe T did not have a lamb when observed on 2 June, but appeared to be lactating. The status of ewe 2 is unknown -- she was seen only once on 23 May. This ewe group was observed in the area between "X-rock" and Sugarloaf Mountain near the Tarryall River Road in late May and early June, and then moved over McCurdy Mountain into an area around the intersection of Lost and--McCurdy Creeks. They--have remained in this area ·since early .June-Although it appears to be a lambing area, only 1 lambing (ewe "circle") has been confirmed at this site -- the other ewes had already lambed when first observed in this area in late May. �129 Intensive monitoring will continue through September with emphasis on documenting movements and survival of known lambs. Additional monitoring is also planned for November-December in preparation for initiation of experimental treatments in late December. LITERATURE CITED Bailey, J. A. 1990. Management of Rocky Mountain bighorn sheep herds in Colorado. Colo. Div. Wildl. Spec. Rep. 66, Colo. Div. Wildl., Denver. 24pp. Buechner, H. K. 1960. The bighorn sheep in the United States, its past, present, and future. Wildl. Mono. 4:74. Forrester, D. J. 1971. Bighorn sheep lungworm-pneumonia complex. Pages 158-173 in J. W. Davis, and R. C. Anderson, eds. Parasitic Diseases of Wildlife. Iowa State Univ. Press, Ames. Schmidt, R. L.,·C. P. Hibler, T. R. Spraker, and W. H. Rutherford. 1979. An evaluation of drug treatment of lungworm in bighorn sheep. J. Wildl. Manage. 43:461-467. Spraker, T. R., and C. P. Hibler. 1982. An overview of the clinical signs, gross and histological lesions of the pneumonia complex of bighorn sheep. in J. and G. G. Schoonveld, eds. Proc. Bienn. Symp. N. Wild Sheep 3:163-172. Wildlife Researcher �130 Table 1. Summary data for individual bighorn ewes captured for use in experimental evaluation of alternatives strastegies for lungworm treatment. COLLEGIATE - CHALK CREEK1 FREQUENCY COLLAR -SEX EARTAG AGE LAMBING STATUS 149.040 BLUE • YEL 12 F 5+ Lactating. .080 BLUE 13 F 5+ Nursed lamb. .100 • BLUE 14 F 5+ Lactating. .120 BLUE 1 2 F 4+ Lactating. .148 BLUE 2 3 F 4+ Nursed lamb; lamb lost. .180 BLUE 4 15 F 4+ Nursed lamb. .202 BLUE 5 5 F 4+ Nursed lamb. .220 BLUE 6 6 F 5+ Nursed lamb. .240 BLUE 7 16 F 6+ Nursed lamb. .318 BLUE E F 4+ Lactating. 1 1 Blue/white collars; freq. 149.040 - 149.492; black/yellow eartags; trapped 28 February 1991 (7 E; 2 L; 2 R); additional ewes darted 11-20 March 1991. COTTONWOOD CREEK2 COLLEGIATE FREQUENCY COLLAR 149.520 BLACK • .540 BLACK • .560 BLACK .•. .582 BLACK .620 EARTAG RED 8 SEX AGE LAMBING STATUS F 4+ Undetermined. 13 F 4+ Nursed lamb. 9 F 5+ Undetermined. 1 14 F 3+ Lactating . .BLACK 2 20 F 3+ Not lactating; no lamb . .642 BLACK 4 10 F 4+ Nursed lamb. .660 BLACK 5 17 F 5 Not lactating; no lamb. .697 BLACK 6 15 F 2 Not lactating; no lamb. . 722 BLACK 7 11 F 3 Not lactating; no lamb . .740 BLACK 8 6 F 5+ Undetermined. .782 BLACK 9 19 F 2+ Nursed lamb. .800 BLACK E 12 F 6 Lactating . .820 BLACK H F 2 Undetermined; not seen . 2 Black/yellow collars; freq. March 1991 (13 E; 7 L; 1 R). 1 148.520 - 148.980; wh1teJred eartags; trapped 16 �131 TARRYALL - COTTON GORDON' S3 FREQUENCY COLLAR EARTAG • • SEX AGE LAMBING STATUS F 3.5 Lactating. 9 F 3.5+ Nursed lamb. 10 F 5.5+ Lactating. 1 29 F 2.5 Nursed lamb . BlACK -2 5 F 1.5 Undetermined . .120 BlACK 4 31 F 2.5 Undetermined. .160 BlACK 6 25 F 4.5 Nursed lamb . .180 BlACK 7 38 F 3.5+ Lactating. .200 BlACK 8 23 F 4.5 Not lactating; no lamb. .220 BlACK 9 12 F 1.5 Not lactating; no lamb. .302 BlACK K 6 F 3.5 Lactating. .340, BlACK M 3 F 6.5+ Undetermined. .418 BlACK T 48 F 3.5+ Undetermined. .440 BlACK U 32 F 3.5 Nursed lamb . .460 BlACK X· 27 F 3.5 Nursed lamb. 148.020 BlACK -.040 BlACK .060 BlACK ••• .080 BlACK .100 YEL 43 3 Black/white collars; freq. 148.020-148.460; black/yellow eartags; trapped 21 February 1991 (16 E; 12 L; 6 R) TARRYALL - SUGARLOAF MOUNTAIN4 FREQUENCY COLLAR AGE LAMBING STATUS EARTAG SEX OR 35 F 2+ Nursed lamb. F 5 Nursed lamb. 149.530 BlACK .580 • BLACK ••• .600 BLACK 1 6 F' 5 Nursed lamb. .620 BLACK 2 7 F 3+ Undetermined. .642 BlACK 4 4 F 4 Nursed lamb. .660 BlACK 5 10 F 4 Nursed lamb. .880 BlACK R 49 F 5+ Nursed lamb. .940 BlACK T 28 F 2 Lactating. 38 4 Black/b1ue collars; freq. 149.500-149.999; black/orange eartags; trapped 9 March 1991-(9 E; 13 L· .' 4 R). ��133 Colorado Division of Wildlife Wildlife Research Report July 1991 JOB PROGRESS REPORT State of ~C~o~l~o~r~a~d~o~ _ Project No. W-1S3-R-4 Mammals Research Work Plan No. 2A Mountain Sheep Investigations Job No. 8 Statewide Mountain Sheep Management Plan Period Covered: Author: July 1, 1990 - June 30, 1991 N. Thompson Hobbs, Michael W. Miller Personnel: D. Reed Abstract A survey of attitudes on management of mountain sheep populations revealed differences in opinion among the general public, mountain sheep hunters, and D·ivision of .Wildlife employees. The public at'large tended to disapprove of hunting of mountain sheep,-particularly hunting .~or·trophies and huriting ewes. Hunting rams is strongly supported by mountain sheep hunters, but they expressed some disfavor for ewe hunting, even if they knew it would be inexpensive and would help prevent disease outbreaks in bighorn populations. Division of Wildlife (DOW) employees held beliefs intermediate to those of the general public and mountain sheep hunters. Hunters and the general public expressed strong approval for current management programs for mountain sheep. However, in the case of the general public, it was clear that their approval was not based on knowledge of DOW actions. We developed a discrete-time, stochastic simulation model to represent interactions between Pasteurella spp. and mountain sheep (Ovis canadensis) populations. Our model is derived from a system .~'f.differential equatri.ons describing dynamics ~f susceptible,· immune, and recovered (S~R) compartments in a popUlation. Our model differs from the SIR·approach, however, in that we represent compartments as states within individual animals and represent flows among compartments as probabilities of transitions among states. Results of simulations averaged over several populations showed Pasteurella spp. regulating bighorn numbers at a steady state approached via damped oscillations. This result closely resembled the outcome of the SIR model. However, because our model is stochastic, single populations showed a broad range of dynamical behavior including exponential growth, stasis, and episodic crashes in abundance. Equilibrium never prevailed. We surmise from our simulations that fluctuations in bighorn abundance can be plausibly explained by internal properties of host-pathogen interactions. External forces, including habitat degradation and fragmentation, environmental stress, and contact wi~h domestic animals axe.not necessary to p~oduce periodic epizootic. ��135 STATEWIDE MOUNTAIN SHEEP MANAGEMENT PLAN N. Thompson Hobbs Michael W. Miller P. N. OBJECTIVES 1. Develop a model of disease transmission in mountain sheep populations simulating alternatives for disease management. 2. Develop a Statewide Mountain Sheep Management Plan incorporating modeling results, information from the scientific literature, and input from organizational units. 3. Present Management Report to organizational units and interested publics. METHODS AND MATERIALS Opinion Survey We surveyed the general public, mountain sheep hunters, and Division of Wildlife (DOW) employees to identify issues likely to affe~t mountain sheep management during the coming decade. We chose to sample these three groups based on the following reasoning: The DOW is mand~ted to serve the public in .Colorado. Understanding the beliefs and values of those we serve will facilitate setting meaningful objectives for bighorn management and will help identify political and sociological constraints on meeting those objectives. Mountain sheep hunters take a somewhat greater interest in how mountain sheep are managed than does the public at large. Because of this interest, and because these hunters represent clients we sere directly, their views are also critical to guiding our future management. Finally, we chose to sample DOW employees because our own beliefs are likely to influence how we manage. In formulating strategies and tactics for managing Colorado's wildlife, we need to be particularly mindful of how our own beliefs differ from those of the people we serve. Surveys were conducted by Standage Accureach using telephone interviews .. Five hundred members of the general public were chosen at random from telephone directories; 125 mountain sheep hunters were selected from records of license holders, and 100 DOW employees were chosen from a personnel roster. Survey questions were written to obtain responses in 4 general areas: 1) quality of service provided by the DOW, 2) values and beliefs about mountain sheep populations, 3) attitudes on bighorn management, and 5) willingness to pay for mountain sheep management via alternative funding. The survey included 12 questions and took about 15":20 minutes to deliver. Simulation Model The classical model of transmission of microparasites- (viruses, bacteria, . etc.) in animal populations represents interactions of 3 pools within a host �136 population (Anderson and May 1979, May 1983, May 1986, Hassell and Anderson 1989, Anderson 1991). These pools include animals that are susceptible to infection (S), those that are infected and infectious (I), and those that have recovered from the disease and are immune to it (R) (hence, the acronym for the model, SIR). At any given time, the sum of these 3 compartments gives the total number of animals in the population, i.e., (1) Changes in the sizes of these compartments over time can be specified in three, coupled differential equations: tiS dt = a (S+I+R) ~~ = -bS-~SI ~SI- (cub+l) dR dt = I (2) (3) (4) lI-bR where a is the per capita rate of birth, b is the per capita rate of death from causes other than disease, B is a transmission coefficient, a is the disease induced mortality rate, and 1 is the ra~e of recovery from disease. Productipn of new infections is given by the average number o~ infective contacts per infected-(i.e. clinically ill) animal per unit time (p) multiplied by the total number of contacts between infected and susceptible individuals (SI). The transmission coefficient, p, is roughly analogous to the searching efficiency of a predator in simple predator-prey models (Hassell and Anderson 1989, Anderson 1991); its ~eciprocal is proportional to the time between infectious encounters (Anderson et al. 1981). In the case where immune animals are also infective (i.e. carriers), a second transmission term (e) can be applied to produce infections as a result of contacts between recovered and susceptible individuals (eSR) (Cooke 1982). The SIR model depends on assumptions that are inappropriate for mountain sheep. For example, we must assume that populations mix homogeneously and instantaneously (Bailey 1975). This assumption fails when social structures and movement patterns restrict contacts among individuals (May 1986, Mollison 1987, Andreasen and Christiansen 1989). More importantly, like all compartment models, the SIR model offers highly misleading results when applied to small, local populations because it allows pool sizes and flows among pools to fall to levels of less than one animal. This allows steady states to occur in models that would not occur in nature. These problems motivated us to develop an individual-based (Ackerman et al. 1984, Huston et al. 1988), discrete-time formulation that applies to small populations (e.g. 10-1,000 animals) characteristic of mountain sheep; and that relaxes assumptions on mixing. There are other discrete alternatives to the SIR model (e.g., Olsen et al. 1988), but these assume a constant flow of new infectives into the population, an assumption we wished to avoid. Our model is based on the following rules.- Infected animals t~ansmit infection to'susceptible animals at a high rate; and recovered animals transmit infection to susceptible animals at a relatively low rate. Once �137 infected, an animal can recover or die. Recovered animals cannot become infected again. These rules follow from observations of pasteurellosis in both mountain sheep and domestic ruminants. Mountain sheep with clinical pasteurellosis consistently shed bacteria in nasal secretions (Onderka and Wishart 1984, Miller and Hobbs 1989, Miller et al. 1991, Wild and Miller 1991), but Pasteurella spp. are recovered from nasal secretions of healthy mountain sheep relatively infrequently (Dunbar et al. 1990, Miller et al. 1991, Wild and Miller 1991). Infection with Pasteurella spp. may convey lasting immunity to recovered animals (Confer et al. 1984, Cho and Jericho 1986, Donachie et al. 1986, Confer 1988, Miller et al. 1991), but apparentlyimmune mountain sheep remain weakly infective after they recover from the disease -- bighorn lambs ,and subadults have developed clinical pasteurellosis when exposed to symptomless ewes who had recovered from this disease several years before (Miller and Hobbs 1989, Miller et al. 1991, Wild and Miller 1991). Such lifelong infectivity is not unusual for bacterial diseases (Smith 1982). Our model represents changes in the disease state of individuals. At any given time, an individual can be susceptible or infected or recovered. The model tracks the state of each animal and represents dynamics of the population by summing over all individuals at the end af each time step. Individual states are analogous to pools in the SIR model, and changes in state are analogous to flows. Transitions in state are regulated by probabilities derived from the turnover rates used in the SIR model. Deriving probabilities in this way allows state transitions of large populations to be directly comparable,to flow rates in the SIR model (Ackerman et al. 1984, King 1991). ' Hence, although our model is stochastic and the SIR model is deterministic, both models can be based on the same parameter values. For example, Equation 3 represents deaths of infected animals as a first order process. It follows from this representation that the probability that an animal remains in the infected pool (i.e., remains alive) after t units of time following infection is P{infected) t = (5) e-ct• where lla is the average life expectancy of an animal that eventually will die of 'the disease, as in the SIR model. If time periods are discrete and independent, Equation 6 can be used. to calculate the probability that an individual animal dies during each time step: PUnfected - dead) = (6) l-e-C, Following the same logic, we can derive the probability of recovery and of transmission. Given that an animal is infected at the beginning of a time step, then the probability that it recovers during that time step is PUnfected - recovered) = (7) l-e-1 �138 where 1/1 is the mean duration of infection of animals that will eventually recover. Similarly, given that an animal is susceptible at the beginning of a time step, the probability that it is infected at the end of the time step is P(susceptible~infected) = 1- (e-1I1 • e-er) (8) where P is the average number of infections produced per infected animal per unit time, e is average number of infections produced per recovered animal per unit time, i is the number of infected animals encountered by a susceptible individual and r is the number of recovered animals encountered by a susceptible individual. In our model, the number of encounters during each time step is determined by the average group size (gs). For each infected and recovered animal, a random draw of gs individuals from the current population determines which animals are encountered. In addition to its disease state, each individual in the population has states for sex (male or female) and for age (lamb, yearling, adult). Births and deaths due to causes other than disease (Jorgensen.and Wishart 1986), as well as the sex of offspring, are determined using the probabalistic approach outlined above. Changes in state are determined for each animal by comparing the transition probabilities to random deviates drawn from a uniform distribution with endpoints. 0 and 1 (Ackerman et al. 1984). Thus, our algor~thm requires cy~ling through all "animals in the populatj.on at each time step. Disease states of individuals are updated weekly; births and deaths due to causes other than disease are updated annually. We estimated parameter values for p, e, a, and 1 based on observed duration of infections and life expectancy of captive mountain sheep infected with Pasteurella spp. (Miller and Hobbs, unpubl.), and on time series data for pasteurellosis outbreaks in wild bighorn popUlations (Simmons 1982, Festabianchet 1988). Estimates of P and e are particularly coarse, and may not be reliable for quantitative predictions. However, we believe they allow an adequate starting point for examining qualitative behavior. Values for life history parameters of mountain sheep were set to produce an average intrinsic rate of increase of 0.12 (Haas and Decker 1980, Jorgensen and Wishart 1986). For simplicity. we did not represent density-dependent effects on birth and death rates. This means that in th~ absence of disease, simulations would produce exponential growth. RESULTS AND DISCUSSION Opinion Survey As in past surveys, the DOW received high levels of public approval. Sixtythree percent of the public at large and 74% of mountain sheep hunters reported a generally favorable impression of the DOW (Appendix, Table 1). SJmilar approval ratings were given for managing big game hunting, fishing opportunity, and mountain sheep populations (Appendix, Table 3). Both of our constituent groups expressed a high level of trust in bur ability to manage mountain sheep "effectively and correctly" (Appendix, Table 3). However, this �139 was despite the fact that 40% of the general pubic reported that they did not know much about us (Appendix, Table 4). These responses suggest that for much of the public, we do a good job of projecting an "image" to the public without communicating information on precisely what we do. On the other hand, 81% of mountain sheep hunters said they knew a great deal or quite a bit about the DOW. Thus, bighorn hunters believe they are quite knowledgeable about our activities, the general public appears to know substantially less about us, but both groups hold us in high regard. The survey revealed a marked discrepancy between the general public and mountain sheep hunters on why they value mountain sheep (Appendeix, Table 5). Over 80% of the public believed that "...the value of mountain sheep as a part of a camping and backpacking experience exceeds the animal's value for sport hunters.", while only 20% of bighorn hunters believed this to be tue. DOW employees were evenly split on this issue. No other question produced such a polarization among the 3 populations sampled. In contrast, there was strong agreement among the 3 populations on the statement "I may never see a mountain sheep in the wild, but it-is very important for me to know that they exist in Colorado." (Appendix, Table 5). Over 80% of all 3 groups expressed some level of agreement with that statement. We interpret this to mean that all of the people we serve want to see viable bighorn populations in Colorado for their cultural and heritage values. On the other hand, the people we serve disagree about the relative importance of consumptive vs. non-consumptive values. Knowledge of mountain sheep biology differed among sample groups. A sub~tantial proportion of all 3 groups sampled knew that the status of mountain sheep had improved in Colorado since 1970 (Appendix, Table 6). However, over 60% of the public believed that "Mountain sheep are among the endangered species in Colorado", while in excess of 75% of bighorn hunters and DOW employees realized this was not the case (Appendix, Table 7). This may reflect some misunderstanding about the meaning of "endangered species", because only 45% of the public surveyed believed that "...mountain sheep populations will disappear in many parts of Colorado in the next ten years." It was also not widely understood by the public that, "Mountain sheep are likely to have more diseases than most other wildlife." Over 80% of bighorn hunters and about half of DOW employees recognized that disease was a particularly important problem for mountain sheep. These responses indicate that the general public and mountain sheep hunters have strong opinions on the value of mountain sheep populations. However, the public is not particularly knowledgeable about biological problems confronting mountain sheep, while hunters appear to have somewhat greater understanding of those problems. It should be emphasized that a significant portion of the public seems to believe that survival of mountain sheep as species is one of the important issues in bighorn management. Mountain sheep hunters and DOW employees believed that this is not a particularly important problem. The schism in consumptive and non-consumptive values between the general public and mountain sheep hunters was also reflected in specific attitudes about hunting. Eighty-five percent of the general public opposed hunting bighorn rams as trophies, and 78% opposed hunting rams for recreation. We asked the question about recreational val--ueof hunting rams in--two forms (Appendix, Table 9 and 10). Regardless of how we asked the question, 70% or more of the general public opposed hunting rams as "sport" or "recreation" (Appendix, Table 9 and 10). However, opposition was less strong to hunting �140 rams as a source of food. About half of the public was opposed to " ...hunting adult male mountain sheep as a source of food"; about half were in favor. There was no clear trend in opinions on hunting relative to age of respondents in the general public. The proportion of the population opposed to hunting rams was quite uniform across all age groups. It was not surprising that mountain sheep hunters strongly supported recreational hunting and hunting for food. DOW employees expressed similar support for these types of hunting. However, a surprisingly large proportion of bighorn hunters (27%) and DOW employees (34%) opposed hunting rams as trophies. With respect to hunting females, we asked bighorn hunters whether they would buy a ewe license for $25. Sixty-two percent of bighorn hunters said it was not at all likely they would do so (Appendix, Table 12). Knowledge that ewe hunting might be important to reducing disease outbreaks did not appreciably change this ambivalence (Appendix, Table 13). The lack of interest in ewe hunting was attributable to a competing interest in hunting for trophies (40% of those who wouldn't buy a ewe license) as well as to the belief that ewe hunting reduces the population (35% of those who would not buy a license) (Appendix, Table 14). Twenty-three percent of bighorn hunters who would not buy a ewe license opposed ewe hunting as a matter of principle (Appendix, Table 14). We focused several questions on specific approaches to managing disease in bighorn populations (Appendix, Table 15): Over 80% of all 3 populations approved of "Placing some mountain sheep in captivity for disease resear~h purposes." Among mountain sheep hunters and DOW employees~ there was also relatively high approval for "Increasing habitat for mountain sheep, even though it could be harmful to some habitat used for other wildlife, such as deer." (Appendix, Table 15). The general public was less supportive of such actions. Seventy-three percent of bighorn hunters approved of using ewe hunting to control disease (Appendix, Table 15). However, remember that despite this view, only 26% of bighorn hunters were likely to buy a ewe license (Appendix, Table 13). DOW employees expressed similar approval of ewe hunting (62%). Forty-two percent of the general public was opposed to ewe hunting (even as part of an effort to reduce disease); 22% were srrongly opposed to ewe hunting (Appendix, Table 15). We pointed out to all persons surveyed that the preponderance of financial support for bighorn management in Colorado came from hunting revenues. We asked whether they felt bighorn management should depend on license sales. Sixty-one percent of the general public and over 80% of bighorn hunters and DOW employees felt that mountain sheep management should be supported by" ... some sort of contribution" from funding sources other than sale of hunting licenses. However, this left 39% of the public believing that bighorn management should depend entirely on hunting revenue. This is problematic in light of the opposition to hunting expressed by most of the public. Of those who supported alternative funding sources, there was strong support for using general fund revenues, unredeemed bottle return deposits and tax check-offs to pay for mountain sheep management. There was not strong support for generating additional revenue by taxes on sporting goods or for users' fees on public land. Several political issues are likely to impact the ability of the DOW to manage mountain sheep populations in Colorado during the coming decade. The general public and bighorn hunters hold sharply different opinions on the objectives �141 population density. However, declines in the number of recovered animals resulting from natural mortality opposed these trends, and occasionally allowed the population to escape the regulating effects of disease (Fig. 2A). The magnitude of simulated disease outbreaks was strongly dependent on time since the previous epizootic. This dependence resulted because the frequency of disease-induced deaths during a simulated epizootic increased in proportion to the number of susceptible animals available for infection. The susceptible pool increased as long as the population remained disease-free. Consequently, long intervals between epizootic were associated with high levels of diseaseinduced mortality during previous epizootic. During the last century, bighorn populations have fluctuated unpredictably, largely as a result of episodic outbreaks of disease and parasitism. Many workers believe that fluctuations in bighorn numbers are ultimately caused by changes in their habitats. Disease is seen as a proximate symptom of deficiencies in habitat resources; these deficiencies include poor forage conditions (Stelfox 1976), lack of migration corridors (Risenhoover et al. 1988), and excessive disturbance from man (Sprake~ et al. 1984). This view holds that the dynamics of diseases in mountain sheep populations are ultimately driven by extrinsic forces like plant succession (Risenhoover and Bailey 1985, Wakelyn 1987, Risenhoover et al. 1988), human encroachment on traditional ranges (King and Workman 1986) and recurrent interactions with domestic livestock (Goodson 1982, Foreyt and Jessup 1982, Onderka and Wishart 1988,'Foreyt 1989, C~llan et al. 1991). Our view differs. Onc~.Pasteurella spp. are introduced'into a bighorn' population by domestic livestock our other means, our model illustrates that 2 simple assumptions on disease transmission are sufficient to produce local epizootic at irregular intervals. We show that disease-induced fluctuations can result whenever 1) infected animals transmft disease at high rates and 2) immune animals transmit disease at low rates. High rates of transmission between infected and susceptible animals ~plify the effect of an initial infection and allow disease to spread geometrically among susceptible individuals once a single animal becomes infected. A low rate of transmission between immune and susceptible animals following an epizootic delays the onset of another outbreak and allows large numbers of susceptible individuals to accumulate in a population. The tension between these two routes of transmission produces unpredictable cycles of population growth and decline (Fig. 2A-C). Our assumptions appear quite plausible for Pasteurella spp. infections in mountain sheep. Many animals that show no signs of disease harbor Pasteurella spp. in their tonsils (Rosen 1981, Onderka and Wishart 1988, Dunbar et al. 1990, Wild and Miller 1991). When this is the case, the probability of transmission of Pasteurella spp. to another animal is extremely low. However, when animals are clinically ill, then Pasteurella spp. are abundant in nasal exudates (Rosen 1981, Onderka and Wishart 1984, Onderka and Wishart 1988, Miller and Hobbs 1989, Miller et al. 1991, Wild and Miller 1991). As a result, coughing and nasal discharge associated with illness dramatically increase the likelihood of transmission (Rosen 1981). We believe this simple, physical mechanism can explain many of the dynamics of disease outbreaks in mountain sheep. We invoke stochast Icf.cy to explain these"dynamics, but it Ls--also possible that disease outbreaks are deterministically chaotic (Aron 1984, Schwartz 1985, Olsen et al. 1988, O'Lseri and Schaffer 1990). Although data on mountain sheep are inadequate to distinguish between chaos and noise, we emphasize that �142 for bighorn management and disagree on the role of hunting in meeting those objectives. Although there was a broad consensus among all groups surveyed that mountain sheep were highly valuable in their own right, bighorn hunters favored management that includes hunting as a primary use of mountain sheep, while the public at large seemed to favor managing mountain sheep to enhance their value as part of other recreational pursuits, like camping and backpacking. As a result, the appropriate goals for bighorn management are not particularly clear. There is also disagreement about what management tactics are appropriate in achieving those goals. Harvest is a important tool in managing mountain sheep as well as other wild ungulates. However, the general public does not strongly support hunting bighorn males or females. Bighorn hunters failed to show strong support for ewe hunting, even if it were needed to prevent disease outbreaks. These responses emphasize that using harvest to regulate the distribution and abundance of bighorn populations may not be feasible in the foreseeable future. It is certain, however, that continued harvest of rams and expanded harvest of ewes will depend on documenting benefits of hunting relative to specific objectives for bighorn populations. For example, if we believe that hunting is needed to reduce the frequency and magnitude of disease outbreaks, then we will need to convince the public and sport hunters that this is the case. In the absence of convincing arguments we may face stout political opposition from the public and lack of meaningful participation from bighorn hunters. Simulation Model Individual-based simulations produced episodic, self-limiting epizootic (Fig. 1). Outbreaks of disease were initiated by conversion of a single animal from the susceptible state to the infected state after contact with a weakly infective, recovered animal. Because susceptible animals were much more likely to become infected after contact with an infected animal than after contact with a recovered one, the addition of a single infected animal into the population caused a rapid increase in the likelihood of new infections. If the infected animal transmitted the disease before it died, then a geometric rise in the number of infected animals usually ensued, resulting in -a full-blown epizootic with high mortality. As time passed, the decline in the number of available susceptible animals and the recovery or death of infected animals acted collectively to reduce disease transmission. As the number of new infections approached 0, the population stabilized at a new, reduced level. These dynamics depended only on the probabilities of changes in state and were not driven by any external processes. Epizootics caused dynamic, varied behavior in simulated mountain sheep populations. We ran simulations to represent 20 separate populations over 50 years. All populations had the same transition probabilities and differed only in the series of random numbers used to drive state transitions. Trajectories of these populations included exponential growth following stasis (Fig. 2A), cycles (Fig. 2B), and erratic fluctuations following exponential growth (Fig. 2C). Final populations sizes varied from 100 to 1,500 animals. However, when we averaged results over all populations, our model approached an equilibrium of about 200 animals via damped oscillations (Fig. 2D). This behavior resembled the approach to steady state predicted by the SIR model. �143 t/J ...J -e ~ Z 400 -- ---" 300 < II. 0 ffim . 100 SUSCEPT. .6 , \ \ o6o6 ~ 2 RECOV. " ,, ,, ,, 200 .• ~-- ,' .•. ! __ -~- 0 0 5 TOTAL 10 __ INFECT • ---------- •....t...•...•... 15 20 25 WEEKS Figure 1. Changes in numbers of susceptible. infected. and recovered mountain sheep during the course of a simulated epizootic initiated by infection of a single susceptible animal following contact with a recovered animal in year O. Final steady state included 6 susceptible animals; all other survivors were immune and weakly infectious. Figure 2. A-C. Simulation output for individual populations given identical parameter values and different random number seeds. Three infected animals were introduced to each population at year O. Parameter values for disease dynamics were ~ - 0.10 animal weeks", e - 0.0001 animal weeks", 1 = 0.125 weeks", & - 0.5 weeks". ~ - 10 animals. D. Simulation output averaged over 20 individual populations. Steady state conditions illustrate the ability of disease to regulate population growth. In the absence of disease, growth would have been exponential. The occurrence of simulated disease outbreaks was weakly dependent on time since the previous epizootic. This was the case because the number of susceptible animals relative to the number of recovered animals fell to a local min:i,mumimmediately aft~:ran epizootic and .i.ncreasedthereafter. As the number of susceptibles increas~d. the ratio of susceptible to recovered animals within a group also increased. As a result, the contact rate between individuals harboring disease and those susceptible to it tended to rise with �144 both processes are internally driven and do not require external driving forces. Thus, if the behavior of our model is faithful to processes in nature, then external forces like environmental stress and habitat loss are not necessary to sustain episodic outbreaks of disease in wild bighorn populations. This does not mean that changes in habitat cannot influence the frequency and intensity of epizootic by influencing transmission and pathogenicity. But it does mean that restoration of pristine habitat conditions, elimination of ~stress", and segregation of mountain sheep from domestic livestock would not necessarily eliminate disease cycles from many established bighorn populations. Pasteurella spp. appear to be thoroughly endemic in many mountain sheep populations in North America -- if our assumptions on the mechanism of transmission are correct, then we should probably expect that even the best-managed herds will experience epizootic (e.g. Bailey 1990). Our simulations offer several important cautions for revealing cause and effect in bighorn population dynamics. Internally driven cycles offer tremendous opportunity for spurious correlation with external conditions and events. Consider a population regulated as in Fig. 2A. If we began treating the population for parasites during year 25 and observed its subsequent _ growth, we might be severely tempted to assume that treatment improved the population's performance. Similar problems can arise in evaluating bighorn popUlations established by relocation. Given the diversity of trajectories that we might expect from populations regulated by the mechanisms we propose (Fig. 2A-C) , we should anticipate high variability in performance among transplanted ~heep herds. Such variability is commonly observed (e.g. Bailey 1990). However, differences in performance among transplanted h~rds m~y be only weakly related to habitat conditions, environmental s trre ssoxs , genetic heterozygosity, or other factors that are often invoked to explain the success (or failure) of transplants. Our simulations emphasize that separating the effects of disease from the effects of other influences on bighorn population dynamics will require careful experimental design and large samples. We are likely to be misled by anything less. It is becoming increasingly clear that population models based on the behavior of individuals offer qualitatively different results than models based on molar, average behavior (Lomnicki 1988, May and Southwood 1990). At regional scales, our model suggests that disease can suppress the growth of metapopulations of mountain sheep and can maintain steady state conditions (Fig. 2d). At local scales, however, disease appears to produce dramatic fluctuations in animal numbers (Fig. 2A-C). Contrast between regional and local behavior has emerged as a pivotal feature of populations of plants and animals in nature (Wiens and Rotenberry 1981, Allen and Starr 1982, Shugart and Urban 1988, Hanski 1991, Verboom et al. 1991). Many management decisions on ungulates are inherently local in scope, but the inferences of molar population models may not be appropriate on local scales. Such models offer inferences on "average behavior", while managers must often respond to the variation that creates the average. We believe that the concept o~ equilibrium derived from molar models, a concept that dominates thinking on regulation of animal numbers (Sinclair 1989), may rarely apply to the local dynamics of mountain sheep popUlations affected by disease. LITERATURE CITED Ackerman, E., L. R. Elveback, and J. P. Fox. 1984. Simulations of Infectious Disease Epidemics. Charles Thomas, Springfield, IL. 202pp. �145 Allen, T. F. H., and T. B. Starr. 1982. Hierarchy: perspectives for ecological complexity. Univ. Chicago. Press, Chicago. 310pp. Anderson, R. M. 1991. Populations and infectious diseases: ecology or epidemiology? The eighth Tansley lecture. J. An. Eco1. 60:1-50. Anderson, R. M., and R. M. May. 1979. Population biology of infectious diseases: Part I Nature 280:361-367. Anderson, R. M., H. C. Jackson, R. M. May, and A. M. Smith. 1981. dynamics of fox rabies in Europe. Nature 289:765-771. Population Andreasen, V., and F. B. Christiansen. 1989. Persistence of an infectious disease in a subdivided population. Math. Biosci. 96:239-253. Aron, J. L. 1984. Seasonality and period-doubling bifurcations in an epidemic model. J. Thero. BioI. 110:665-679. Bailey, N. T. J. 1975. its Applications. Bailey, J. A. 1990. Colorado Colo. The Mathematical Theory of Infectious Diseases and Charles Griffin and Co., London. Management of Rocky Mountain bighorn sheep herds in Div. Wild1. Spec. Rep. 66. 24pp. Buechner, H. K. 1960. The bighorn sheep in the United States, its past, present, and future. Wi1d1. Mono. 4:74. Callan, R. J., T. p. Bunch, G. W. Workman, and R. E. Mock. 19-91. Development of pneumonia .in desert bighorn sheep after exposure to a flock of exotic· wild and domestic sheep. J. Am. Vet. Med. Assn. 198:1052-1056. Caugh1ey, G. 1970. Eruption of ungulate populations with emphasis on Himalayan thar in New Zealand. Ecology 51:53-71. Cooke, K. L. 1982. Models for epidemic infections with asymptotic cases l: one group. Math. Modelling 3:1-15.. Dunbar. M. R., A. C S. Ward, and G. Power. 1990. Isolation of Pasteurella-haemolytica from tonsillar biopsies of Rocky Mountain bighorn sheep. J. Wi1dl. Dis. 26:210-213. Foreyt, W. J., and D. A. Jessup. 1982. Fatal pneumonia of bighorn sheep following association with domestic sheep. J. Wild1. Dis. 18:163-168. Forrester, D. J. 1971. Bighorn sheep lungworm pneumonia complex. Pages 158-173 in J. W. Davis and R. C. Anderson, eds. Parasitic Diseases Of Wild Mammals. Iowa State University Press, Ames. Goodson, N. J. 1982. Effects of domestic sheep grazing on bighorn sheep: review. Biennial Symp. N. Am. Wild Sheep and Goat Counc. 3:287-313. a Hanski, I. 1991. Single-species metapopu1ation dynamics - concepts, models and observations. BioI. J. Linn. Soc. 42:17-38 . .._ HasseLL, M. P .• and R. M. Anderson , 1989. Predator-prey and host pathogen interactions. Pages 147-196·in J. M. Cherret, ed. Ecological Concepts. Blackwell Scientific, Oxf'ord , �146 Houston, D. B. 474pp. 1982. The Northern Yellowstone Elk. Macmillan, New York. King, M. M., and G. W. Workman. 1986. Response of desert bighorn sheep to human harassment: management implications. Trans. N. A. Wildl. Nat Res. Conf. 51:74-85. Klein, D. R. 1968. The introduction, increase, and crash of reindeer on St. Matthew Island. J. Wildl. Manage. 32:350-367. Lomnicki, A. 1988. Population ecology of individuals. Princeton Univ. Press. 223pp. Monog. Pop. Ecol. 25. Lutz, L. L. 1970; The bighorns: wild sheep of California, part 1. Forests. 76:28-31. . Amer. May, R. M. 1983. Parasitic infections as regulators of animal populations. Am. Sci. 71:36-45. May, R. M. 1986. Population biology of microparasitic infections. Pages 405-442 in T. G. Hallam and S. A. Levin, eds. Mathematical Ecology. Biomathimatics Vol. 17. Springer-Verlag, Berlin. _____ , and T. R. E. Southwood. 1990. Introduction. Pages 1-19 in B. Shorrocks and I. R. Swingland, eds. Living in a Patchy environment. Oxford Univ. Press, Oxford. McCullo~gh, D. R. 1979. The George Res~rve deer herd: population ecology of a K- selected species. Univ. Michigan Press, Ann Ar~or. 27lpp. Mollison, D. 1987. Population dynamics of mammalian diseases. Pages 329-342 in S. Harris, ed. Mammal Population Studies. Symp. ·Zool. Soc. Lond. 58. Olsen, L. F., and W. M. Schaffer. 1990. Chaos versus noisy periodicity: alternative hypotheses for childhood epidemics. Science 249:499-504. Olsen, L. F., G. L. Truty, and W. M. Schaffer. 1988. Oscillations and chaos in epidemics: a nonlinear dynamics study of six childhood diseases in Copenhagen, Denmark. Theor. Pop. BioI. 33:244-370. Onderka, D. K., and W. D. Wishart. 1984. A major bighorn sheep dieoff from pneumonia in southern Alberta. Pages 356-363 in Proc. N. Wild Sheep and Goat Counc. 4. _____ , and 1988. Experimental contact transmission of Pasteurella-haemolytica from clinically normal domestic sheep causing pneumonia in Rocky Mountain bighorn sheep. J. Wildl. Dis. 24:663-667. Risenhoover, K. L .• and J. A. Bailey. 1985. Foraging ecology of mountain sheep ovis-canadensis implications for habitat management. J. Wildl. Manage. 49:797-804. _____ • • and L. A. Wakelyn. 1988. Assessing the Rocky Mountain bighorn sheep management problem. Wildl. Soc. Bull. 16:346-352. Rosen, M. N. 1981. Pasteurellosis. Pages 244-252 in J. W. Davis, L. H. Karstad, and D. O. Trainer. Infectious Diseases of Wild Mammals. Iowa State Univ. Press, Ames, Iowa: �147 Schwartz, I. B. 1985. Multiple stable recurrent outbreaks and predictability in seasonally forced nonlinear epidemic models. J. Math. Biol. 21:347-362. Shugart, H. H., and D. L. Urban. 1988. Scale, synthesis, and ecosystem dynamics. Pages 279-290 in L. R. Pomeroy and J. J. Alberts, eds. Concepts of Ecosystem Ecology. Springer-Verlag. New York. Sinclair, A. R. E. 1977. The African Buffalo: A Study of Resource Limitation of Populations. Univ. Chicago Press. 355pp . . 1989. Population regulation in animals. J. M. Cherrett, ed. Ecological Concepts: to an Understanding of the Natural World. Oxford. Pages 197-241 in The Contributions of Ecology Blackwell Scinentific, Smith, C. E. G. 1982. Major factors in the spread of infections. Pages 207-235 in M. A. Edwards and U. McDonnell, eds. Animal Disease in Relation to Animal Conservation. Symp. Zool Soc. Lond. 50. Spraker, T. R., C. P. Hibler, G. G. Schoonveld, and W. S. Adney. 1984. Pathologic changes and microorganisms found in bighorn sheep ovis-canadensis-canadensis during a stress-related die-off. J. Wildl. Dis. 20:319-327. Stelfox, J. G. 1971. Bighorn sheep in the canadian rockies a history 1800-1970. Can. Field. Nat. 85:101-122. 1976. Range eco'Logyof rocky mountain bighorn sheep in Canadian national parks. Pages 1-50 in Can Wildl Serv Rep Ser. (39). 1976. Turner, J. C., and J. B. Payson. 1982. Occurrence of selected infectious diseases in the desert bighorn sheep herds of the Santa-Rosa Mountains, California. Calif. Fish. Game 68:235-243. Verboom, J., K. Lankester, and J. A. J. Metz. 1991. Linking local and regional dynamics in stochastic metapopulation models. BioI. J. Linn. Soc. 42:39-55. Wiens, J. A., and J. T. Rotenberry. 1981. Habitat associations and community ·structure of birds in shrubsteppe environments. Ecol. Monogr. 51:21-41. Wild, M. A., and M. W. Miller. 1991. Detecting nonhemolytic Pasteurella-haemolytica infections in healthy Rocky Mountain bighorn sheep (Ovis-canadensis) - influences of sample site and handling. J. Wildlife. Dis. 27:53-60. Prepared by __;7\.}=-\~_.;._-=-;..;,;~~---!.]b~ N. Thompson Hobbs Wildlife Researcher _ Wildlife Researcher �148 APPENDIX Table 1. Responses of the general public and bighorn sheep hunters to the question: "In general, are your impressions of the Colorado Division of Wildlife very positive, somewhat positive, neutral, somewhat negative, or very negative?" Population Sampled Random Sample Bighorn Sheep Hunters DCN PersOfYlel Nl.Ililer Percent Nl.Ililer Percent Nl.Ililer Percent Responding Responding Responding Responding Responding Responding Impression of Division of Wi ldl ife not asked 92 18 28 22 · · Somwhat positive 225 45 65 52 134 27 18 14 · · · Neutral 30 6 10 8 · Very ne98tive 6 1 3 2 · DK/no answer 11 2 1 1 · Very positive Somewhat negative · · · · Table 2. Responses of the general public and bighorn sheep hunters to the question: "Tell me if you feel the Colorado Division of Wildlife is doing an excellent-job, good job, fair job, or poor job in: . . Managing big game hunting. Managing the bighorn sheep population in Colorado. Providing adequate fishing opportunities throughou~ the state of Colorado. Population Random Sample Sampled Bighorn Sheep Hunters DCN Personnel Nl.Ililer Percent Nl.Ililer Percent Responding Responding Responding Responding Nl.Ililer Percent Responding Responding How well does DCN manage bi 9 game? 52 10 32 26 · Good 239 48 61 49 Fair 127 26 29 23 Poor 25 5 3 2 Do not know 55 11 . . · · · · Excellent · · · · · �149 Table 2. cont. Population SiIq)led Bighorn Sheep Hunters Random S8q)le . DOW Personnel NUJi)er NUJi)er Percent NUJi)er Percent Percent Responding Responding Responding Responding Responding Responding How well does DOW manage bighorns? 60 12 56 45 · · Good 215 43 48 38 · Fair 98 20 11 9 · Poor 10 2 7 6 115 23 3 2 · · · · · · Excellent Do not know Population S8q)led Random S8q)le Bighorn Sheep Hunters DOW Personnel NUJi)er NUJi)er Percent NUJi)er Percent Percent Responding Responding Responding Responding Responding Responding How well does DOW manage fishing? Excellent · 98 20 17 14 Good 256 51 59 48 Fair 85 17 29 23 · Poor 25 5 8 6 Do not know 34 7 11 9 · · · -- · · · Table 3. Responses of the general public and bighorn sheep hunters to the question: "Do you strongly agree, somewhat agree, somewhat disagree or strongly disagree with the following: I trust the Colorado Division of Wildlife to manage bighorn sheep effectively and correctly in Colorado." Population S8q)led Random SiIq)le Bighorn Sheep Hunters DOW Personnel NUJi)er NUJi)er NUJi)er Percent Percent Percent Responding Responding Responding Responding Responding Responding I trust DOW to manage bighorns well. Strongly agree 253 51 88 70 Somewhat agree 215 43 24 19 Somewhat disagree 12 2 6 5 Strongly disagree 13 3 6 5 5 1 1 1 Do not know/No opinion · · · · · · · · · · �150 Table 4. Responses of the general public and bighorn sheep hunters to the question: "Would you say you know a great deal about the Colorado Division of Wildlife, quite a bit, some or not too nuch?" Population S~led Random S~le Bighorn Sheep Hunters DOW Personnel Nurber Percent Nurber Percent Nurber Percent Responding Responding Responding Responding Responding Responding Knowledge of Division of Wi ldl ife · · · · 24 2 100 14 11 . . · · · A great deal 22 4 26 21 Quite a bit 77 15 55 44 Some 195 39 30 Not too nuch 197 40 6 1 None (volunteered) · Table 5. Responses of the general public, bighorn sheep hunters, and Division of Wildlife (DOW) employees to the question: "Do you strongly agree, somewhat agree, somewhat disagree or strongly disagree with the following: I think the value of bighorn sheep as a part of a c~ing exceeds the animal's value for sport hunters. and backpacking experience I may never .see a bighorn sheep in the w.ild, but it is very ilJl)Ortantfor me to know that they exist in Co~orado." . " Population S8q)led. Random S8q)le Bighorn Sheep Hunters DOW Personnel Nurber Percent Percent Nurber Percent Nurber Responding Responding Responding Responding Responding Responding Nonconsumptive value exceeds hunting value. Strongly agree 284 57 10 8 20 20 Somewhat agree 127 26 15 12 31 30 Somewhat disagree 58 12 65 52 20 20 Strongly disagree 22 4 27 22 25 25 7 1 8 6 6 6 Do not know/No opinion �151 Table 4 cont. Population S~led DOW Persomel Bighorn Sheep Hunters Random SlIq)le ; Percent Nl.II1ber Percent Nunber Percent Nunber Responding Responding Responding Responding Responding Responding I value "knowing bighorns are there." Strongly agree 380 76 87 70 89 87 Somewhat agree 96 19 14 11 9 9 Somewhat disagree 11 2 12 10 2 2 Strongly disagree 8 2 6 5 1 1 Do not know/No opinion 3 1 6 5 1 1 Table 6. Responses of the general public, bighorn sheep hunters, and DOW employees to the question: "As far as you know, have the nunber of bighorn sheep in Colorado decreased, increased, or remained about the same since 19701" Population SlIq)led Bighorn Sheep Hunters Random S~le DOW Persomel Percent Nl.I1ber Percent Nunber Percent Nunber Responding Responding Responding Responding Responding Responding Status of bighorns since 19701 . .: Decreased __ Increased Remained the same Do not know 61 12 1 1 10 10 208 42 102 82 67 66 90 18 15 12 12 12 138 28 7 6 13 13 Table 7. Responses of the general public, bighorn sheep hunters, and Division of Wildlife (DOW) employees to the question: "00 you strongly agree, somewhat agree, somewhat disagree or strongly disagree with the following: Bighorn sheep are among the endangered species in Colorado. At the current rate of mortality, I think the bighorn sheep population will disappear in many parts of Colorado in the next ten years. Bighorn sheep are likely to have more diseases than most other wildlife." Population S...,led Random S...,le Bighorn Sheep Hunters DOW Persomel Nunber Nunber Percent Nunber Percent Percent Responding Responding Responding Responding Responding Responding Bighorn sheep are an endangered species. Strongly agree 106 21 4 3 9 9 Somewhat agree 201 40 7 6 7 7 Somewhat disagree 117 24 82 66 75 74 Strongly disagree 44 9 32 26 11 11 Do not know/No opinion 29 6 . . . . _-- �152 Table 7 cont. Population SiIq)led Random SlIq)le Bighorn Sheep Hunters DOW Personnel NUlber Percent NUlber Percent NUlber Percent Responding Responding Responding Responding Responding Responding Bighorn populations likely to disappear in future. Strongly agree 74 15 6 5 6 6 Somewhat agree 151 30 4 3 14 14 Somewhat di sagree 160 32 87 70 60 59 Strongly disagree 75 15 26 21 17 17 Do not know/No opinion 37 7 2 2 4 4 Population S8q)led Random SiIq)le Bighorn Sheep Hunters DOW Personnel NUlber Percent NUlber Percent Percent NUlber Responding Responding Responding Responding Responding Responding Bighorn sheep are prone to diseases. Strongly agree 51 10 68 54 Somewhat agree 123 25 35 28 . 28 27 32 31 19 19 Somewhat dis~ree 186 38 7 6 StrOngly disagree 76 15 12 10 14 14 Do not know/No opinion 58 12 3 2 9 9 Table 8. Responses of the general public and bighorn sheep hunters to the question: "Dg you favor or oppose hunting adult male bighorns to obtain a trophy?" Population Sampled Bighorn Sheep Hunters Random Sample DOW Persomel NUlber Percent Nl.Ilb!r Percent Nl.Ilb!r Percent Responding Responding Responding Responding Responding Responding Opinion hunting rams as trophies? Favor 52 10 90 72 53 52 Oppose 422 85 34 27 35 34 22 4 1 1 14 14 No opinion �:1.53 Table 9. Responses of the general public, bighorn sheep hunters, and Division of Wildlife (DOW) employees to the question: "00 you favor or oppose hunting adult male bighorns as recreation?" Population SBq)led Random S8q)le Bighorn Sheep Hunters DOW Persomel Nl.IIber Percent Nl.IIber Percent Responding Responding Responding Responding Nl.IIber Percent Responding Responding Opinion of hunting rams as recreation? Favor 86 17 105 84 79 78 Oppose 386 78 17 14 18 18 No opinion ·23 5 3 2 4 4 Table 10. Responses of the general public, bighorn sheep hunters, and Division of Wildlife to the question: (DOW) employees liDo you strongly agree, somewha.t agree, somewhat disagree or strongly disagree with the following: Hunting male bighorn sheep is a form of sport and recreation, them should be allowed to do SO." and people who want to hunt Population SBq)led Randon! SBq)le Bighorn Sheep Hunters DOW Personnel Nl.IIber· Percent Nl.IIber Percent Responding Responding Responding Responding Nl.IIber Percent Responding Responding -, People should be allowed to hunt rams. Strongly agree 52 11 111 92 70 rz Somewhat agree 95 19 8 7 14 14 Somewhat disagree 117 24 . . 5 5 Strongly disagree 227 46 1 1 7 7 2 0 1 1 1 1 Do not know/No opinion Table 11. Responses of the general public, bighorn sheep hunters, and Division of Wildlife to the question: (DOW) employees "Do you favor or oppose hunting ac:lul t male bighorns as a source of food?- Population Random SlIq)le SlIfI1lled Bighorn Sheep Hunters DOW Personnel Nl.IIber Percent Nl.IIber Percent Responding Responding Responding Responding Nl.IIber Percent Responding Responding Opinion of hunting rams as food? Favor 250 50 104 83 84 82 Oppose 221 45 18 14 12 12 25 5 3 2 6 No opinion .. _'_" 6 �154 Table 12. Responses of bighorn sheep hunters to the question: If the Colorado Division of Yildlife were to allow female bighorn sheep to be hunted in Colorado and the license fee to hunt bighorn sheep was $25, how likely would you be to buy a license to hunt female bighorns? Very likely, somewhat likely, or not at all likely? Population Random Sample Sampled Bighorn Sheep Hunters DOIJ Personnel NUJi:ler Percent NUJi:ler Percent NUJi:ler Percent Responding Responding Responding Responding Responding Responding Yould you buy a ewe license for S25? · · · Very likely Somewhat likely Not at all likely Table 13. Responses · · 33 26 1 9 15 12 3 27 · 77 62 7 64 of bighorn sheep hunters to the question: UDoes knowing that hunting female bighorn sheep would help control population size and help prevent disease outbreaks among bighorn sheep make you change your mind about how likely you would be to buy a license to hunt adult female bighorns in Colorado?" Population Random Sampl e Sampled Bighorn Sheep Hunters DOIJ Personnel Nl,IIj)er Percent Nl,IIj)er Percent NUJi:ler Percent Responding Responding Responding Responding Responding Responding Buy a eWe license (ifit reduces disease). -- Yes . No 1 Do not know 1 . 17 22 . 50 58 74 6 50 3 4 . . 100 . �155 Table 14. Responses of people who said they would not buy a ewe license to the question: "\lhich of the following statements best describes why you would not be likely to buy a license to hunt female bighorn sheep in Colorado? I am not interested in hunting female bighorn sheep because I only hunt bighorn for the trophy. I don't think adult female bighorns should be hunted because it will cause a decrease in the bighorn population. I think it is OK to hunt male bighorns. but I oppose hunting female bighorns just as a matter of principle.Population S8q)led 00\1 Persomel Bighorn Sheep Hunters Random S8q)le Percent NUlber Percent NUlber Percent NUlber Responding Responding Responding Responding Responding Responding \lhy would you not buy a ewe license? Interested in trophy . . 30 40 6 86 Reduces population 1 50 26 35 . . Matter of principle 1 50 17 23 1 14 None of "these (volunteered) . . 2 3 . . Table 15. Responses of the general public. bighorn sheep hunters. and Division of Wildlife (00\1) employees to the question: "Do you strongly approve. somewhat approve. somewhat disapprove, or strongly disapprove of the following methods for minimizing disease outbreaks among bighorn sheep in Colorado. Placing some bighorn sheep in captivity for disease research purposes. Increasing"habitat for bighorn sheep. even though it could be harmful to some habitat used for other wildlife. such as deer: " Allowing sport hunting of adult female bighorns to control the sheep population size in Colorado and reduce the incidence of disease outbreaks." Population S8q)led Random S8q)le Bighorn Sheep Hunters 00\1 Persomel Percent NUlber Percent NUlber Percent NUlber Responding Responding Responding Responding Responding Responding OK to use captive bighorns for research? Strongly approve 218 44 85 68 61 60 Somewhat approve 203 41 32 26 26 25 Somewhat disappr~ve 40 8 4 3 7 7 Strongly disapprove 36 7 2 2 4 4 0 2 2 4 4 No opinion 1 �156 Table 15 cont. Population Sampled Random Sample B ighom Sheep Hunters DOW Persomel Nunber Percent Nunber Percent Nunber Percent Responding Responding Responding Responding Responding Responding Increase bighom harms deer? habitat if Strongly approve 88 18 T7 62 43 42 Somewhat approve 198 40 31 25 34 33 Somewhat disapprove 146 29 7 6 10 10 Strongly disapprove 58 12 8 6 8 8 8 2 2 2 7 7 No opinion Population Sampled Random Sample DOW Persomel Bighorn Sheep Hunters Nunber Percent Nunber Percent Nl.II1ber Percent Responding Responding Responding Responding Responding Responding Allow ewe harvest to control disease? Strongly approve 88 18 73 58 63 62 Somewhat approve 196 39 28 22 22 22 Somewhat disapprove 99 20 16 13 12 12 Strongly disapprove 108 22 7 6 2 2 7 1 1 1 3 3 No opinion �157 Colorado Division of Wildlife Wildlife Research Report July 1991 JOB PROGRESS REPORT State of ~C~o~l~o~r~a~d~o~ _ Project No. W-153-R-4 Mammals Research Work Plan No. 2A Mountain Sheep Investigations Job No. 9 Quantity and Quality of Mountain Sheep Habitat with Regard to Minimum Viable Populations and Response of Mountain Sheep to Human Activity Period Covered: Author: July I, 1990 - June 30, 1991 D. F. Reed Abstract Increasing recreation in the recently designated Arkansas Headwaters Recreation_Area has resulted in a concern that increa$ed human activity might be detrimental to resident mountain sheep populations. Cooperative efforts have developed between the Bureau of Land Management (BLM) ,'the Colorado Division of Parks and Outdoor Recreation, and the Division of Wildlife (DOW). BLM became committed to evaluating bighorn sheep habitat. It was judged that a DOW ManagementjResearch joint effort could do the habitat evaluation in cooperation with BLM as well as test a habitat model and the effects of disturbance on bighorn sheep. A Research Proposal, "Impacts of human activity on mountain sheep distribution and abundance" was prepared and presented to BLM. Secondly, a Program Narrative (PN) was prepared. Thirdly, a Cooperative Agreement was signed between the BLM and DOW where BLM committed $30,000 to the work described in the PN for the Segment beginning 1 July 91 and potentially $30,000 each of 3 following years, for a total of $120,000. Contract work has been planned with the University of Colorado Colorado Springs Geography and Environmental Studies Department to map and evaluate potential mountain sheep habitats. This will be done by using existing geographic information systems (GIS), scanning in BLM IR photos, and inputing bighorn sheep distribution and habitat field data. ��159 QUANTITY AND QUALITY OF MOUNTAIN SHEEP HABITAT WITH REGARD TO MINIMUM VIABLE POPULATIONS AND RESPONSE OF MOUNTAIN SHEEP TO HUMAN ACTIVITY Dale F. Reed P. N. OBJECTIVE Evaluate the quantity and quality of mountain sheep habitat with regard to minimum viable populations and test the response of mountain sheep to human activity. SEGMENT OBJECTIVES 1. Prepare Research Proposal. 2. Prepare Program Narrative. 3. Complete Cooperative Agreement between BLM and DOY for joint funding. ACKNOWLEDGMENTS I thank co-principal investig~tors M. Y. Mi-Her. R. B. Gill, J. Vayhinger, and S. Ogilvie for their ideas and suppor t, Division personneL D. Finch, Y. Travnicek, and J. Backstrand were most helpful in establishing the study areas. Division personnel D. L. Schrupp and D. C. Lovell were most helpful for their ideas and in coordinating GIS. BLM personnel E. Brekke provided base-line information, digital elevation maps (DEMS) , IR photos, and important coordination; and T. R. Grette assisted with plant identification. T. P. Huber (University of Colorado, Colorado Springs) scanned in 1R photos and assisted in establishing habitat training sites or "ground truthing". DESCRIPTION OF AREA Two areas of the Arkansas River Canyon between Nathrop and Parkdale were included in the study: Area A - about 10 mi2 north of the Arkansas River from about 1.0 mi east of Texas Creek to 1.5 mi west of Parkdale, and Area B about 20 sq mi north and east of the Arkansas River from east of Nathrop to east of Salida. Areas A and B include approximately the lower half of Segment 4 and Segment 2, respectively, as delineated by BLM (Bureau of Land Management 1988:1-3). The areas include steep to very steep mountain slopes both north and east of the Arkansas River Canyon. Rock outcrops are mainly of igneous rock, intermingled with very shallow soils, terraces, and gravelly alluvia. The vegetation is typically Pinus-Juniperus Spp. and/or mountain shrub interspersed with graminoid and Cholla communities. �160 METHODS AND RESULTS Methods for evaluating both the quantity and quality of bighorn sheep habitat follows Smith et al. (in prep). Methods for testing the response of bighorn sheep to human activity as covered in the PN will need to be developed further. RESULTS AND DISCUSSION A Research Proposal and a Program Narrative were prepared. Agreement between BLM and DOW was completed. A Cooperative LITERATURE CITED Bureau of Land Management. 1988. Final Arkansas River Recreation Management Plan and Environmental Analysis. U.S. Dep. of Int. Smith, T. S., J. T. Flinders, and D. S. Winn. 19 A habitat evaluation procedure for Rocky Mountain bighorn sheep in the intermountain west. (in prep) prepar9~eoQ Dale F. Reed Wildlife Researcher �161 Colorado Division of Wildlife Wildlife Research Report July 1991 JOB PROGRESS REPORT State of ~C~o~l~o~r~a~d~o~_ Project No. W-153-R-4 Mammals Research Work Plan No. 3A Pronghorn Investigations Job No. 2 Habitat Selection and Population Performance of a Pioneering Pronghorn Population Period Covered: Author: July 1, 1990 - June 30, 1991 T. M. Pojar Abstract The distribution of the Middle Park pronghorn (Antilocapra americana) herd has been monitored ~ince January 1, 198.1. During this. time, the population has increased from 80 to approximately 280 aruma.Ls.. (summer 1991). Based on 546 relocations, the"area of the minimum convex polygon (Mep) was 936 km2 (361 mi2) for "the season of maximum dispersal, i.e. summer. During winter when the popUlation distribution is most compressed, the area of the MCP was 81 km2 (31 mi2) based on 745 relocation points of marked animals. The populations rate of increase ranged from 0.17 to 0.52 and averaged 0.35 over the 4 years of available data. Based on a regression of popUlation size and rate of increase, the K-value for the population is in the range of 400 to 625 animals. Pronghorn hunting was permitted in Mid~le Park" during the 1990 season for the first time since the season was closed statewide in 1899. Fifteen permits were issued (10 buck and 5 doe/fawn) with 13 of the 15 permittees participating in the hunt. Nine bucks and 4 does were harvested and 1 buck was found unretrieved." Horn le~gths of 10 bucks were measured and averaged 32.3 cm (12.7 in)·'and.ranged from '26.2 em (10.3 in) '"to' 36-.'6cm (14.4 in) (Sherwick Min, Pers. Comm.). ��163 HABITAT SELECTION AND POPULATION PERFORMANCE OF A PIONEERING PRONGHORN POPULATION Thomas M. Pojar P.N. OBJECTIVE Describe population dynamics and habitat use of a pioneering, expanding pronghorn population. .SEGMENT OBJECTIVES 1. Describe seasonal and annual distribution of the Middle Park pronghorn population. 2. Determine sample sizes of radio-collared animals and observations necessary to describe habitat preferences. 3. Monitor population dynamics of Middle Park pronghorn with: a. Ground counts to describe changes in population size. b. Ground counts to quantify population sex and age composition. . STUDY AREA The study area is described in Pojar (1988:183-184). For orientation of Middle Park in relation to the state of·Colorado see Figure 1. The approximate area of sagebrush steppe habitat in Middle Park is outlined in Figure 1. This is considered to be the potential area of distribution for pronghorn in Middle Park. METHODS AND MATERIALS Seasonal and Annual Distribution Tracking was done mostly from the ground to increase the probability of observing and identifying animals with numbered plastic collars. A fixed-wing aircraft was used if an animal could not be located after a reasonable effort from the ground. Legal descriptions of animal locations were recorded to the nearest quarter mile then converted to UTM (U.S. Army 1973) coordinates for computer processing. All radioed animals have been located biweekly· (with very few exceptions) since January 1, 1987. The program MCPAAL (Stwue NO) was used to calculate the area of the m~n~mum convex polygon and the 95% ellipse for summer and winter areas of habitation. The relocation points for May, June and July define the "summer" area and points for December, January and February define the "winter" area. Population Size and Structure .... Herd structure ·estimates are obcaf.ned in late summe r or early autumn by locating all radioed animals and·classifying all anima_ls that accompany them. The herd structure estimate that is used in population projections is the one with the largest sample size obtained in August or September. Classification �164 after October 1 is not used because the probability of mistaking early fawns for does increases. In all winters since this population has been monitored, they have wintered in essentially the same relatively limited area. At some time during the winter they are all in a single group, which makes it possible to locate the entire population and obtain a complete count. In addition, it is possible to count all mature bucks (age 1~5 years and older) and verify or correct the buck:doe ratio from the earlier herd structure estimates. The total population count and number of mature bucks counted during winter are used in population projections. Population projections are based on the following assumptions: 1. Winter counts represent the total population and the total number of mature bucks in Middle Park. 2. Late summer age ratio estimates represent "recruitment" into the population. 3. Annual survival of mature bucks and does and female fawns is 92.5%. 4. Annual survival of males in their first year (after weaning) is 50%. (This severe mortality on male fawns is arbitrary, however, it allows the number of mature males in subsequent years of match fairly well with winter counts.) RESULTS Seasonal and Annual Distribution The last deployment of radios in Middle Park was in December, 1988. Attrition of the radios and mortality of radioed animals have resulted in a steady decline in the number of working radios. From a maximum of 24 functional radios in December, 1988, that number has decreased to 17 as of June, 1991. There are very few of the numbered yellow plastic collars that were put on animals during the December, 1986, trapping operation still remaining on animals. Many of the numbered blue plastic collars installed in December 1988 are still providing relocation data. Since December, 1986, there have been 2,659 relocations of radioed and plastic collared pronghorn in Middle Park. The total number of plastic collar relocations is 868 and the total number of radio relocations is 1,791. The relocations in these data bases have been sorted by "winter" which includes the months of December, January and February, and "summer" which includes May, June and July (Table 1). Winter relocations define the critical area inhabited during the time the population distribution is most compressed. The area of habitation during summer represents the maximum dispersion of the population. For this population the wintering area relocations are located within the area circumscribed by the summer area relocations (Figures 2,3,4 and 5). The wintering area comprises 2.3% of the summer area for 1991 and 8.6% of the summer area for all years. In Alberta Canada, the northern most range for this species, Barrett (1982) suggests that wintering areas typically repr~sent about 8% of th~ summer range. �165 Table 1. Area of habitation for Middle Park pronghorn in square kilometers (divide by 2.59 to convert to square miles) for 2 estimators of area--minimum convex polygon (MCP) and 95% ellipse as calculated by the program MCPAAL (Stuwe ND}. WINTERz SUMMERl 95% 95% N3 Ellipse MCP Ellipse MCP N3 1991 Radios Plastic Collars 85 437 1128 131 19 33 29 799 1981 49 7 15 114 817 1329 180 19 35 Radios Plastic Collars 426 507 954 571 75 67 120 821 1131 174 38 71 BothS 546 936 996 745 81 71 Both4 All Years 1 Includes relocations from May, June and July. z Includes re·locations from December, January and February. 3 Total- number of relocations. 4 See Figures 2 and 3 fQr a graphic representation of these locations. S See Figures 4 and 5 for a graphic representation of these locations. Population Size and Structure Accurate counts of the total population size can be obtained during winter and herd structure estimates are obtained in late summer (Table 2). The annual changes in population size are used to calculate the rate of increase (Nz - Nl / N1) (Table 3) which is then regressed on population size to estimate the populations K-value (Figure 6) (Caughley 1977). The rate of increase for 1990-91 of 0.17 is well below the 4-year average of 0.35. For this reason, the regression predicting the K-value has a steeper negative slope than the regression from the previous year and predicts a K-value of 377 animals (Figure 6). This is probably an unrealistic regression because it projects the intrinsic growth rate (the Y-intercept) at 0.70 which is most likely well above what is biologically possible for·this species given the observed herd structure of this population. Since there are only 4 points upon which to base the regression, the best estimate of carrying capacity for the Middle Park pronghorn, at this point in time, is a range of between 400 and 625 animals. As more data points are available, the reliability of the projected K-value should increase. Hunting was permitted for pronghorn in Middle Park during the 1990 season for the first time since the season was closed statewide in 1899 (Hoover et. al 1959) . Fifteen permits were issued (10-buck and 5 doe/fawn) with 13 of ·the 15 permittees participating. Nine bucks and 4 does were harvested-with 1 confirmed buck wounding loss and 1 additional suspected buck wounding' loss or �166 poaching (Sherwick Min, Pers. Comm.). Therefore, a total of 11 bucks and 4 does were removed from the population during the hunt. Having been an unhunted population, it would be expected that the harvested animals would represent older age classes and that the males would have relatively large horns. Min (1990) measured the horns of the 9 harvested males and 1 unretrieved buck. Horn length ranged from 26.2 cm (10.3 in) to 36.6 cm (14.4 in) and averaged 32.3 cm (12.7 in). The Boone and Crockett scores for these 10 animals averaged 69.58 and ranged from 58.0 to 79~25. The highest score in Boone and Crockett records is 93 for an animal taken in Arizona. Colorado's highest scoring animal is from Weld county with a score of 914/8, The minimum score to be included in the Boone and Crockett record book is 82 (Nesbitt and Wright 1981). The population projection for late summer 1991 is presented in Table 4. The projection is based on the age ratio obtained in August, 1990, and the total count of the herd and the number of bucks obtained during the winter of 19901991. In this projection, it is assumed that the age ratio will be the same as in 1990--47 fawns:lOO does. This is a critical assumption and the observed ~ge ratio for 1991 will have a significant impact on this projection. Table 2. Herd structure of Middle Park pronghorn based on a sample obtained by locating radioed animals in late summer. POP. NO. 'X B:D F:D SAM'X OF .YEAR RATIO· SIZE RADIO RADIO RATIO PLE POP. 1986' 1987 1988 1989 1990 80 122 160 223 261 7 24 22 17 5.7 15.0 10.2 6.5 36 54 40 56 22 77 77 -32 50 47 47 63 108 161 148 59 52 68 72 66 , This year's data based on the sample of the population trapped 16 December 1986. Table 3. Population size of the Middle Park pronghorn herd during winter and the calculated rate of increase. YEAR POP. SIZE RATE OF INCREASE 1986-87 1987-88 1988-89 1989-90 1990-91 80 122 160 223 261 .52 .31 .39 .17 �167 Table 4. Population projection for the Middle Park pronghorn population. text for the assumptions. POPULATION BUCKS DOES FAWNS TOTAL 62 124 60 246 WINTER MORTALITY 62 X .075 - 5 MORT 124 X.075 - 9 MORT 30X.5-l5B 30X.075-2D 31 PRE-FAWNING 1991 62 - 5 57 MATURES + 15 YRLS TOTAL - 72 124 - 9 115 MATURES + 28 YRLS TOTAL - 143 MATURE 57 YRLS 15 TOTAL 72 MATURE 115 YRLS 28 TOTAL 143 See WINTER '90-91 LATE SUMMER 1991 215 @ 47F:lOOD 143 X .47 67 FAWNS 282 LITERATURE CITED Barrett, M. W. 1982. Ranges, habitat and mortality of pronghorns at the northern limits of their range. Ph.D. Thesis. University of Alberta, Edmonton, Alberta. 223pp. Caughley, G. 1977. NY. 234pp. Analysis of Vertebrate Populations. John Wiley ~ Sons, Hoover, R. L., C. E. Till, and S. Ogilvie. 1959. The Antelope of Colorado. Colorado Game and Fish Dep. Tech·. Bull. 4. 1l0pp. Min, S. 1990. Effects of female choice on male reproductive success in male pronghorn: 1990 field season report. Dep. Systematics & Ecology, University of Kansas, Lawre~ce. 5pp xerox. Nesbitt, W. H -. , and P. L. Wright. 1981. Records of North American big game. Eighth Edition. The Boone and Crockett Club, Alexandria, VA. 409pp. Pojar, T. M. 1988. Habitat selection and population performance of a pioneering pronghorn population. Colo. Div. Wildl. Res. Rep. July, 181-l92pp. Stuwe, M. (ND). MCPAAL micro-computer programs for the analysis of animal locations. Program Documentation. Cons. and Res. Center, National Zoological Park, Smithsonian Institution, Front Royal, Virginia, USA. 18 printout pages. U.S. Army. 1973. Technical Manual: Universal Transverse Mercator Grid. Headquarters, Dep. of the Army, Washington D.C. TM No. 5-241-8, 64 pp. ��169 Colorado Division of Wildlife Wildlife Research Report July 1991 JOB PROGRESS REPORT State of ~C~o~l~o~r~a~d~o~ _ Project No. W-lS3-R-4 Mammals Research Work Plan No. 3A Pronghorn Investigations Job No. 4 Statewide Pronghorn Management Guidelines Period Covered: Author: July 1, 1990 - June 30, 1991 T. M. Pojar Abstract Discussions were held with Division of Wildlife biologists and administrators to help identify prominent issues regarding pronghorn management in Colorado. Using these issues as a general guide, a questionnaire was designed to identify other issues as perceived by the general public, pronghorn hunters and Division of Wildlife employees. An Lndependent; coinpany ; Standage Accureach,. Inc., conducted. the survey and the results '.areavailable from the author or the Research Center Library in Fort Collins. A document titled Pronghorn Management Analysis Guide, 1992-1994 (third draft) is available from the auchor . This document outlines the history of pronghorn management in Colorado and provides an analysis of the management environment. The issues are examined in terms of the social, political, economic and biotic environment. Public meetings will be held to discuss currently identified issues and attempt to identify other issues that the public might raise. (_ ~ Prepared by ~~ . . omas M. POJa Wildlife Researcher oyw . ��171 Colorado Division of Wildlife Wildlife Research Report July 1991 JOB PROGRESS REPORT State of ~C~o~l~o~r~a~d~o~ _ Project No. W-1S3-R-4 Mammals Research Work Plan No. 6A Mountain Lion Investigations Job No. 1 Mountain Lion Population Dynamics Period Covered: Author: July 1, 1990 - June 30, 1991 A. E. Anderson Personnel: See Acknowledgments Abstract A'working outline for the CDOW Tech. B~ll. nThe puma (Felis concolor hippolestes) on Uncompahgre Plateau Colorado" by A. E. Anderson, D. C. Bowden, and D. M. Kattner is presented with percentages of each major section completed. Essentially, all the narrative is complete except for the RESULTS .AND DISCUSSION section (30%) and the SYNTHESIS AND CONCLUSIONS section (0%). ��173 MOUNTAIN LION POPULATION DYNAMICS Allen E. Anderson P. N. OBJECTIVE To assess the effects of sport hunting on mountain lion populations. SEGMENT OBJECTIVE To prepare final reports for publication. ACKNOWLEDGMENTS Most statistical analyses and computer graphics were performed by Dr. D. C. Bowden, Dept. Statistics, Colorado State University, Fort Collins. D. Masden, Inventory Biologist, Southwest Region, CDOY, assisted with mapping relative average densities" of·deer and elk on winter ranges of Game Management Units 61 and 62. N. McEwen produced or supervised CSU Graphics Dept. personnel prepare most of the figures for the puma publication. M. Roelke, D.V.M., and her colleagues from the Florida Panther Biomedical Investigation and the National Institute of Cancer kindly provided their information on the hematology genetics, spermatozoa, and drug immobilization of Uncompahgre Plateau puma. METHODS AND MATERIALS Between 16 April 1981 and 15 April 1988, 57 puma were captured and 49 were fitted with radio collars. I obtained 3,117 aerial locations of 38 surviving radio-collared puma on 349 fixed-wing flights totaling 1,081 hours (3."1 hours/flight). These locations, recorded as UTM coordinates, permitted calculation of indices of puma home range size, movement, intraspecific interactions and survival. General methodology is described in Anderson (1983). Additional information on some aspects of morphology and physiology came from opportunistic acquisition and necropsy of puma carcasses. RESULTS AND DISCUSSION The manuscript for CDOY Technical Publication, "The puma (Felis concolor hippolestes) on Uncompahgre Plateau, Colorado" by A. E. Anderson, D. C. Bowden, and D. M. Kattner was not completed because of severe arthritis coupled with the side effects of several new, mostly ineffectual prescription drugs for arthritis during the fall, winter, and early spring of 1990-91. Here, I present the outline of that manuscript with the estimated percentages of each section completed in handwritten form. This does not include the "final handwritten copy ready for typing. �174 Percent Complete INTRODUCTION 100 DESCRIPTION OF STUDY AREA AND VICINITY Physiography and Geology Climate Vegetation Livestock and Range Condition Timber Harvest Mammals and Birds Puma.-Mule deer.-Elk. -Pronghorn ante1ope.-Desert bighorn sheep.-Black bear.-Bobcat.-Coyote.-- 100 100 METHODS General Approach Capture and Drug Immobilization Processing Radio Telemetry Home Range Movements Intraspecific Interactions Interspecific Interactions Puma and wild prey.-Puma and 1ivestock.-Puma and humanso-Population Dynamics Sex and age structure.-Litter size.-Numbers and density.-Mortality.-Estimates of survival.-Morphological and Physiological Body mass and measurements.-Growth rateso-Indices of carcass fat.-Organ-gland masses and measurements.-Genetics.-Cellular and chemical constituents of b1ood.-Semen characteristics.-Parasites. diseases and injuries.-Population Dynamics Sex and age structure.-Litter size.-Calculation of numbers and density.-Morta1ity.-- RESULTS AND DISCUSSION Capture Drug Immobilization Aerial Telemetry Indices of Home Range Size. Indices of Movements Intraspecific Interactions �175 Percent Complete Interspecific Interactions Puma and wild prey.-Puma and livestock.-Puma and humans.-Morphological and Physiological Body mass and measurements.-Growth rates.-Indices of carcass fat.-Organ-gland masses and measurements.-Genetics.-Cellular and chemical constituents of blood.-Semen characteristics.-Parasites. diseases and injuries.-Population Dynamics Sex and age structures .--Litter size.-Numbers and density.-Mortali ty. -Estimates of survival.-Population model-potential harvest rates.--a o SYNTHESIS AND CONCLUSIONS LITERATURE CITED 80 _aTo be completed by Gary C. White, if he deems the_available data are relevant and adequate for that exercise. As previously stated, about 50 tables, 30 figures, and 7 appendices will be included. LITERATURE C.ITED Anderson, A. E. 1983. Program Narrative Proj. 45-01-503-15050, Work Plan 6, Job 1. Mountain lion population dynamics. 7pp. (+3 tables and 1 appendix) . Prepared by atLt;-t<) £_' {f_'1l iJ'/f''l.t T Allen E. Anderson Wildlife Researcher ¢;ti ) ��Colorado Division of Wildlife Wildlife Research Report July 1991 JOB PROGRESS REPORT State of ~C~o~l~o~r~a~d~o~ _ Project No. -W-lS3-R-4 Work Plan No. Mammals Research 8A Small Carnivorous Mammals Investigations Job No. Development of River Otter Reintroduction Procedures Period Covered: Author: July 1, 1990 - June 30, 1991 T. D. I. Beck Personnel: T. Benjamin, J. White, G. White Abstract .. . Three juvenile male river otters-(Lutra canadensis) were received· in good health from Oregon in 1990 for release into the Dolores River. At least 2 died prior to spring, and the fate of the third is unknown. Radio telemetry on the animals released in 1989 has been limited because of technical deficiencies in equipment. Limited results from the 3 otters this year suggest telemetry problems may be solved. Dispersion of river otters is widespread throughout the Dolores and San Miguel Rivers. Survival of reintroduced juvenile river otters seems. to be less than for adults. A third population estimation procedure was tested for estimating crayfish (Orconectes virilis) density. Number captured indicate a minimum density of 6.4 adults/m2• Estimation procedure analyses have not been completed. Sightings of river otter at previous release sites suggest the reintroductions have been successful in the short-term and dispersal is occurring. ��179 DEVELOPMENT OF RIVER OTTER REINTRODUCTION PROCEDURES Thomas D. I. Beck P. N. OBJECTIVE Develop procedures for river otter reintroductions in Colorado and establish a self-sustaining population of river otters from which to collect river otters for future translocations. SEGMENT OBJECTIVES 1. Introduce up to 10 river otters-into the Dolores River drainage. 2. Develop techniques to monitor survival, reproduction, dispersion, and dispersal of river otters after introduction. 3. Monitor other river otter release sites in Colorado to evaluate success of past reintroductions. METHODS AND MATERIALS Dolores River Release River otters were captured ~hroughout Oregon by personnel of the-Or~gon Dept. of-Fish & Wildlife under the direction of Meg Eden. The cage-type trap' developed by-Oregon staff was used. River otters were shipped in_transpor.t cages (41 X 41 X 76 cm) via commercial airline to Durango, Colorado. Animals were shipped at 0700 (PDT) and arrived in Durango at 1630 (MDT). River otters were transported to Dolores and released in holding pens (2.6 X 5.2 m) with running water. Food (crayfish and suckers) was provided. All prototype radio transmitters from AVM Equipment Co. were reconstructed to provide greater power to the transmitter in order to increase range from subterranean beaver dens. The larger battery resulted in a slightly heavier and longer trarismitter package; the unit weighs 26 gms and is 70 mm long with a diameter of 17 mm. The 185 mm long whip antenna has an epoxy bead on the distal point to provide a blunt surface. The desired specifications for the upgraded radio transmitter required reception at 300 m when the transmitter was buried in wet soil to a depth of 20 cm and-battery life of 24 months. River otters were removed ~rom the holding pens and transported to Cortez Veterinary Clinic for surgical implantation of the radio transmitters on the morning following their arrival. The river otters were transferred to a squeeze box at the clinic, where immobilization was obtained from intramuscular injection of ketamine (100 mg/ml) and xylazine (20 mg/ml) at a dose rate of 15 mg ketamine per kg of animal. A paste made from betadine and petroleum jelly was thoroughly massaged into the incision areas. This allowed the hair to be combed away from the incision line, omitting the need to shave the area. Two incisions were made along the dorsal mid-line. The larger (2-3 cm) was made at a point between the shoulders while the smaller (1 cm) was made 190 mm posterior to the first. A pocket large enough to hold the -t-ransmitte-r was made under the skin at the anterior LncLsLcn , Blunt sponge -forceps were used to push under the skin from the posterior to anterior incision. The transmitter antenna was pulled under the skin from the anterior �180 incision. Both incisions were closed with subcuticular sutures and Super Glue. Each river otter was tagged with a Monel #3 metal tag in the web between front toes and was measured for total length; tail length; neck, chest, and head girth; and weight. All injuries were recorded. Fecal parasites, white blood cell count, and hematocrit were analyzed. River otters were placed in transport boxes in a cool room until recovered from anesthesia. They were transported to the Dolores River for release. Radio tracking was conducted on foot, by canoe, and from the air. All locations were recorded to the nearest 0.1 mi. Two aerial searches .of the San Miguel, Dolores, and Colorado Rivers downstream to Lake Powell were conducted after coordination with Utah Div. of Wildlife Resources. Crayfish Studies A 5-day mark-recapture study was conducted at RM 151.3 during late July. Baited wire crayfish traps were set in a 5 X 5 grid equally spaced throughout an area 21-m square (the width of the river). A pe1leted, commercially available bait was used with a standard amount in each trap. Traps were checked once daily for 4 days. Each captured crayfish was uniquely marked by writing a number on the carapace with an ultra fine point Sharpie permanent marker. The carapace length, sex, identification number, trap location, and day were recorded for each capture. Marked crayfish were released at the site of capture prior to returning the baited trap to the river. Data were stored in a dBase file and forwarded to Gary White, Colorado State University, for analyses of population estimates and possible bias. Release Site Monitoring .. A.standardized river otter-sighting-form was developed and distributed- to District Wildlife Managers in areas of river otter release. All reported sightings are stored by river drainage. An l8=km stretch of the Piedra River was surveyed by canoe during low water (July) to search for river otter sign. The selected area was chosen because of consistent reports by fishermen of river otter presence. Periodic contact with Utah Division of Wildlife Resources personnel was continued to monitor movement of radio-transmittered river otters in the Green River along the state borders. Contacts were made with commercial boating companies and BLM river rangers to obtain river otter sightings in the Colorado River in the Utah-Colorado border area. RESULTS AND DISCUSSION Dolores River Release Four male river·otters were received from Oregon. Three juveniles were received in good health in late October and early November (Table 1). All animals ate promptly upon release in pens. Recapture the following mo~ing was accomplished by leaving the transport boxes in the pen for a den, with the sliding door in the up position. By quietly entering the pen at first light, the door was closed with the unstressed animal within. The modified surgical procedure which omitted shaving of hair appeared to work well although more time was needed to close the incision cleanly. �l8l Table 1. River otters released in the Dolores River, CO, 1990. Length (cm) Iotal IaU ~lI~YmfeIence (cm) Che§t H~ad He~k III S~X Yt(kg) RO-35 M 3.9 95 35 23.0 21.5 32.5 35 Tenmile Lake, OR RO-36 M 5.2 106 41 24.5 25.0 34.5 36 Tenmile Lake, OR RO-37 -M 6.0 106 44 24.5 26.0 34.5 37 Yillamette Slough, OR lag ~ Qrigin An adult was dead on arrival. He had chewed a large hole through the metal trap upon capture and was considerably stressed. A necropsy indicated massive hemorrhaging in the lungs but no metal scraps in the gastrointestinal or pulmonary tracts. One juvenile male was found dead along the shoreline 3 km upstream from the release site 2 weeks following release. A necropsy indicated an absence of subcutaneous and mesenteric fat with accompanying signs of severe emaciation. The stomach and intestines were empty. As no other wounds were discovered, starvation was considered the likely cause of death. Another juvenile male remained in an abandoned beaver den near the release site. The river iced over the first week of December. Fresh river otter scat was regularly observed near the den area throughout December. This male also died during the winter. The decomposed body was discovered at the entrance to the beaver den in 0.7 m of water after ice break up in the spring. The third juvenile male- expelle-d the radio transmitter in a beaver den _soon after release. Fresh rive~ otter scat, tracks, prey:remairts, and 2 sightirigs indicated a river otter used the beaver den regularly through December and January. The transmitter was recovered at a depth of 1.7 m and 16 m from the shoreline. The improved radio transmitters did allow for location when river otters wer·e in subterranean beaver dens although the short periods of tracking provided limited data. ~adio transmission exceeded 300 m at all times, even when the transmitter was. deep within a beaver den. The initial prototype radio transmitter performed inadequately to allow consistent monitoring of river otter movements. If the animal was on shore or swimming, the transmitters could be detected at distances of 250-400 m. Ho~ever. once the river otter entered a beaver den, reception usually went to zero and rarely exceeded 30 m. An adult male (RO-4) from the November, 1988, release was located in the San Miguel River near the mouth of Horsefly Creek in August, 1990, 43 river miles upstream from the Dolores River confluence. This site is 124 river miles from the release site. He had not been located since April, 1989, until this time. Other river otters from the 1988 release were last located within the Dolores River. Cessation of radio transmission on all was likely normal battery exhaustion. RO-l, an adult male, was last located at RM 130 in May 1990. He had been remaining between RM 130 and 140 during March, April, and May. RO-7, an adult male, was-found in McPhee Reservoir through most of the-summer of 1990 and was last located in September at RM 170, about 12 miles downstream of the dam. An adult male, RO-3, and an adult female, RO-5, were last located in December, 1989, using the area between RM 165 and 170. �182 Location of the river otters released in 1989 remains largely unknown because of poor equipment performance. Two transmitters were recovered from beaver dens after being expelled from the animal, likely because of suture failure. An adult female, RO-2s, was recovered dead at RM 107 on 26 July 1990. She was lying on a sandbar where we camped 3 weeks previous. Her whereabouts between release at RM 146 on 9/12/89 and death are unknown. An adult female, RO-26, moved downstream into the Colorado River where she was located several times in August and September, 1990, in the Castle Creek area about 10 miles downstream of the Dewey Bridge (Utah). Distance moved was at least 158 miles since release on 10/5/89. Personnel of Utah Div. Wildlife Resources were notified of her location and we ceased monitoring in September, 1990. Two juvenile females, RO-27 and 28, died within 2 weeks of release within 2 miles of the release site. An adult female, RO-31, was located at the confluence of the San Miguel and Dolores River,-RM 64.5, in September, 1990. She had been at RM 135 in late March. RO-32, an adult male, left the river and was found dead 6 miles from the river. Another transmitter, from adult male RO-12, was located in a draw amidst wheat fields approximately 4 miles from the Dolores River. Transmission was weak at the time (July 1990). By the time permission was obtained to trespass, no signal was received and aerial searches have not located the transmitter. An unmarked, nursing female was killed in a conibear trap set for beaver on 10 June, 1991, about 5 miles above the town of Dolores. No young were found after 1 day of ground searching. The source of this animal is unknown, but she may be the offspring of-RO-S. Although presence of young was never confirmed for RO-S in 1989, her restricted movement in April and May, 1989, suggested she may have young. The strong affinity for large subterranean beaver-~ens ~imited_our success at visual sightings. Although RO-s moved -about ls--miles upstream in late summer, such a move would- not be difficult for jUvenile river ~tters at this time. Juvenile river otter received from Oregon in early September have ranged from 90 to 105 -cm in total length. A dead juvenile river otter was recovered from the Piedra River on 6/30/91 and had a total length of 49 cm. _Although quite limited, it appears that juveniles have attained nearly SOX of their adult length by 1 July. This suggests that accurate identification of juvenile age status will become increasingly difficult after 1 July. The loss of released animals to death and dispersal further suggests the need to release a lot of animals in each release site. The relatively poor survival of juvenile river otters after translocation suggests this age group should be avoided for reintroductions. Another 10 river otters are anticipated for release in 1991. With the apparently improved radio transmitter, prospects of learning" appear to be better in 1991-92. Crayfish Studies Total crayfish marked during the 4 capture periods was 3,101. Over 800 individuals were recaptured and only 2 individuals had unreadable markings. Both of these were initially marked on day 1 and recaptured on day 4. Analyses of the data are not completed. Complete reporting and discussion of the 3 crayfish population estimation tests (removal, unique mark for day of capture only, unique numbered individual) will be completed in the coming .segment. Using the same capture procedure at 3 sites for the 3·tests has produced 4-day capture totals of 1,390, 2,470, and 3,101. Minimal densities of adult . crayfish based on captures range from 2.7 to 6.4/m2. Since captures likely do �183 not represent the total population density can be presumed to be higher. A density of 6.4/m2 provides a standing stock biomass of crayfish of 1,012 kg/ha (903lb/ac) (avg. adult crayfish - 15.75 gms, n - 2,470). Crayfish provide an abundant food source for river otters in the Dolores River. There is reason to believe that crayfish numbers are increasing in the lower canyon where crayfish were rare in 1988. The relative abundance survey conducted from RM 173 to RM 74 will be repeated in 1991. An observation of a river otter foraging at RM IlIon 8 December 1990 was enlightening. This stretch of river had few crayfish present in 1988, yet the river otter was successful at catching a crayfish on each of 21 consecutive dives during an hour period during a season when crayfish are inactive. Release Site Monitoring Observations by outdoorsmen strongly indicate that river otters are present in Navajo Reservoir and all the major tributaries of this large reservoir. While the majority of sightings occur in the reservoir and the Piedra River, reputable sightings also come from the San Juan, Navajo, and Pine Rivers. River otters are consistently sighted throughout the Piedra River drainage and the small tributaries. At least 3 river otters released in the Green River in Utah have ventured into Colorado, traveling in both the Green and Yampa Rivers. Since only a small pr'oportion of the animals released in Utah were radio transmitter equipped, there could be others. It appears likely that continued releases in Utah will result in some river otters taking permanent .residence in .Colorado'.sportion '.o-f'the Green River. Sightings of river otter are consistently made on the Colorado River between Westwater Ranger Station and Dewey Bridge in Utah. The regularity of sightings is probably a response to the high recreational use of this river segment. While at least 1 river otter from the Dolores release has joined this population, the appearance of this population predates the initial Dolores release. It is speculated that this population is the result of dispersal from either or both the North Fork Colorado River or the Gunnison River releases. There was a reputable sighting of a river otter in the White River during 1990. Either overland travel from the Colorado River or a long dispersal from the Green River are possible sources for this animal. By .---.L~· -"",=~....;...;;....;j)J:;.;;......_~___;;;;;;~_ Thomas D. I. Beck Wildlife Researcher ��185 Colorado Division of Wildlife Wildli~e Research Report July 1991 JOB PROGRESS REPORT State of Colorado Project No. W-lS3-R-4 Work Plan No. 9A job No. 1 Mammals Research Elk Investigations Impact of Elk Winter Grazing on Livestock Production and Work Plan No. 3 Job No. S Period Covered: Author: July 1, 1990 - June 30, 1991 N. T. Hobbs, D. L. Baker, and G. Bear Personnel: D. Bowden Abstract We continued analysis of cattle weight responses. Calf weights taken at the end of the spring grazing season showed weak effects of treatment (30 elkfkm2 vs. control, ~ - 0.091) when effects of calf weights into the pasture were removed. using covariance. Analysis of covariance was needed to increase precision in the analysis and to remove significant tendency for calf weigh,ts i:qtothe pasture to be heaviest in the contrro Ls . This cendency was not rela.'tedto treatment and resulted froJ;ll random assignment of replacement animals. Calf weights in fall were also weakly ·influenced by elk grazing (linear effects ~ - 0.09). Cow weight performance was not affected by elk grazing. a ��187 IMPACTS OF WINTER GRAZING BY ELK ON CATTLE PRODUCTION P. N. OBJECTIVES 1. To test the hypothesis that elk grazing during winter influences the productivity and botanical composition of herbage on sagebrush grassland ranges during spring. 2. To test the hypothesis that elk grazing during winter influences the body weights and rates of gain of cows and calves using sagebrush grassland ranges during spring. METHODS AND MATERIALS Study Area We conducted experiments on the Little Snake Wildlife Management Area in northwestern Colorado (township 9 north, range 95 west, sections 9, 10). The area is about 35 km (19 mi) north of Maybell, Colorado, on County Road 19. Although this area does not typically contain high concentrations of elk during winter, it is representative of areas that do have those high densities. Topography of the area includes level ridge tops, rolling hills, and deep gullies, ranging in elevation from 1800 to'2000 m (5900 to 6600 ft). Aspects are southern and southwesterly with'an, average slope of 15 degrees. Soils care generai.lY sandy and sandy loam .. Climate of the area is dry and cold. The growing season averages only 81 days. Annual mean temperature is 6.06 C (42.9 F). Annual precipitation averages 27.5 cm (12.5 in). Vegetation is dominated by big sagebrush (Artemisia tridentata) with 'an understory predominated by needle and thread (Stipa comata). western wheatgrass (Agropyron smithii), Indian ricegrass (Oryzopis hymenoides), Junegrass (Koleria cristata), and cheatgrass (Bromus tectorum). Important forbs include wallflower (Erysimum asperum), peppergrass (Lepidium perfoliatum), silver lupine (Lupinus argenteus), and, scarlet globe mallow (Sphaeralcea coccinea). Experimental Design We observed effects of elk grazing on forage and cattle responses in a randomized'comp1ete block design with 4 levels of elk density (0 elkjkm2, 8 e1kfkm2, 15 e1kfkm2, and 31 elkjkm2) and 3 replications per level. There were 3 blocks, each consisting of 4 pastures. Each pasture within a block was stocked with one level of elk density such that each block contained all levels. During year 1, the 12 available pastures were blocked by pretreatment biomass of perennial grasses with the 4 lowest grass biomass pastures forming I block, and the 4 highest grass biomass pastures forming a second block, and the remaining 4 pastures serving as the third block. The 4 levels of elk density were randomly assigned ,to pastures within each block during year 1. Procedures We stocked pastures with elk in December and January, 1987-89. Average date of release int:opastu~es.was Jan~ry~. All elk were removed from pastures during April 10-20. We introduced 7 cow-calf pairs and one dry heifer into each pasture on May 9; 1990, and removed them 5 weeks later. This represents a departure from the first 2 study years when animals remained in the pastures for 6 weeks. �.... ,.,., .. "'.. ,~"..... , .... ,...•.•••••.• " ,.•""u~. 188 However, a marked reduction in forage production during the 1988-1989 growing seasons compelled us to reduce the stocking rate during 1990 in order to preserve consistent utilization rates in control pastures. With the exception of addition of new heifers and other replacements required by death losses etc., cows during year 4 were the same animals we observed during year 3 and were assigned to the same pastures they were in previously. We observed the birth dates of all calves and weighed them to the nearest 0.05 kg immediately after birth. Cows and calves were weighed to the nearest 1 kg when they were introduced to pastures and were reweighed 5 weeks later when they were removed. We estimated canopy cover of herbs shrubs immediately after removing cattle from pastures. Cover was estimated from the summed length of interception by each plant along 30, l2-m transects randomly placed in each pasture during year 1. We estimated standing crop, productivity, and utilization of forbs, perennial grass, and annual grasses by harvesting samples from 40 pairs of 0.70-m2 plots in each pasture on each of 3 sample dates. Pastures were sampled immediately after the elk were removed (April 27-29), at the midpoint of the spring grazing season (May 30-J'une 1), and at its end (June 30-July 2). Samples were dried at 60 C for 48 hrs, separated by hand into live and dead, and weighed to the nearest 0.01 g. We used an analysis of covariance for a randomized complete block design to analyze weight responses of cattle. Calf weight into the pasture was used as a conco~itant observation for calf weights out of pastures and for rates of gain. Cow weights into the.pastures were used similariy for adult cattle and .heifers. We are in·the process of consUlting with Dave Bowden (Statistics Department, Colorado State University) on the details of. this analysis to assure our statistical approach is reliable. Future analyses will reflect these consultations, and may offer somewhat different results than we report here. RESULTS AND DISCUSSION Averaged across the 4 study years, elk grazing during winter and early spring caused small reductions in calf performance during spring on sagebrush grassland range (Fig. 1). These reductions were shown in statistically weak effects on calf weights at the end of the spring grazing season (control vs. 30 elkfkm2, ~ - 0.091, linear effect ~ - 0.11, Appendix Table 1) and·in the fall (control vs 30 elkfkm2, ~ - 0.12, linear effect ~ - 0.11, Appendix Table 2). Calf Weights at Weaning Covariate-Calf Weight Into Pasture Calf Weights out of Pasture Covariate-Calf Weight Into Pasture 130 270 120 250 ~ 110 --;; 100 ~230 --;; 210 o ~ 90 ~ ~ 80 70 1/1 .. 60 ~ 50 ~ 190 ] 170 I---..1 I -tI----~·TI----------__J~ ~ 10 20 Elk Oensity (animols/km2) 30 I ~ 150 ~ 130 ~ 110 40 o I--____~I_-'+l------tI 40 90 70~ -r o ~ 10 20 ~ 30 --r 40 Elk Density (animols/km2) Fisure 1. Effects of elk srazins on calf ~eishts at the end of the sprins srazing season and in the fall averased over 4 study years. Means adjusted to common weight into pasture. Vertical bars - ! 2 standard errors of the mean. �·r.o·" ...•.•_.,., .... ,',.' ..•..•~:,_•.•'~.'.'.'.:.', ' •..••. ", .•.•• ":' ..• .'.: ..' •... ,'., ', ,',_". , ...•. , ... . "< ., ....•.•. 189 Although there appeared to be a statistically significant effect of treatment on the covariate (linear effect ~ - 0.02, Appendix Table 3), this appearance was created by effects of initial randomization and replacement of cattle· over the 4 study years. Weights of animals taken into the pasture in the animal's first study year tended to be heaviest in the control pastures (linear effect ~ - 0.13. level effect - 0.10, Appendix Table 4). It is clear that there was no opportunity for treatment to influence these weights, since these animals were not in the experiment the previous year. Consequently, it was important to remove the effect of weight into the pasture from the analysis of weight out of the pasture to remove this spurious influence. Effects of elk grazing on cow weights were less strong than effects on calf weights (Fig. 2). Cow weights at the end of the spring grazing season were not influenced by treatment (level effect ~ - 0.53, Appendix Table 5), nor were weights in the fall (level effect ~ - 0.81 Appendix Table 6). Cow Weights at Weaning Covariate-Cow Weight Into Pasture Cow Weights Out of Pasture Covariate-Cow Weight Into Posture soo 700 'Clsoo 'Clsoo Co c ~400 Co I e I I I '":! ~ 500 ..., I :! ~300 I I I ~400 -.. -.. III III ~300 ~200 ..J 100 200 0 . 10 20 30 40· 0 "0 20 30 Elk Density (animols/km2) Elk Density (animols/km2) Fi&ure 2. Effects of elk &raziD&on cow wei&hts at the end of the sprins &razins season and in the fall averqed over 4 study years. Meena adjusted to coamon wei&ht into pasture •. Vertical bars - :t 2 standard errors of the .ean. AJ! Prepared by JI. ~~~~~~~ __,~~~-------N. Thomp~obbs Wildlife Researcher Dan L. Baker Wildlife Researcher &-:;9.~ George D. Bear ,!ildlife.R~searcher 40 �190 APPENDIX Table 1. Analysis of effects of elk grazing on calf weights out of the pasture. Weight into the pasture used as a covariate. Source DF Sum of Squares F Value Pr > F Model 27 70258.8345613 Error 269 8677.1499520 Corrected Total 296 78935.9845133 R-5quare C.V. 0.890074 7.317328 Tests of Hypotheses using the Type III MS for LEVEL*BLOCK as an error term Source LEVEL 80.67 0.0001 CALFWT3 Mean 77.61754376 DF Type III SS F Value 3 653.84565390 1.83 Pr > F 0.2429 Tests of Hypotheses using the Type III MS for LEVEL*BLOCK as an error term Contrast linear in level > DF Contrast S5 F Value 1 411.35624562 3.45 0.1128 Pr F_. Tests of Hypotheses using the Type III MS for LEVEL*BLOCK .as an error term Contrast control vs others DF Contrast SS F Value Pr > F 1 150.47337906 1.26 0.3045 Tests of Hypotheses using the Type III MS for LEVEL*BLOCK as an error term Contrast control vs 30 DF Contrast S5 F Value Pr > F 1 479.59535164 4.02 0.0919 �191 Table 2. Analysis of effects of elk grazing on calf weights at weaning. Weight into the pasture used as a covariate. Source DF Sum of Squares F Value Pr > F 8.91 0.0001 Model 27 90144.7329757 Error 194 72660.5311826 Corrected Total 221 162805.2641582 R-Square C.V. CALFWT4 Mean 0.553697 10.26008 188.62436648 Tests of Hypotheses using the Type III MS for LEVEL*BLOCK as an error term Source LEVEL DF Type III SS F Value Pr > F 3 1905.21474849 1.56 0.2948 Tests of Hypotheses using the Type III MS for LEVEL*BLOCK as an error term Contrast linear in level DF Contrast SS 1 1642~36756617 F Value· 4.02 Pr > F 0.0917 Tests of Hypotheses using the Type III MS for LEVEL*BLOCK as an error term Contrast control vs others DF Contrast SS F Value Pr > F 1 355.89964895 0.87 0.3865 Tests of Hypotheses using the Type III MS for LEVEL*BLOCK as an error term Contrast control vs 30 DF Contrast SS F Value Pr > F 1 1347.22823490 3.30 0.1192 \" �192 Table 3. Analysis of effects of elk grazing on calf weights into the pasture. Source DF Sum of Sguares F Value Pr > F Model 23 2911.23904230 Error 282 31535.07620304 Corrected Total 305 34446.31524534 1.13 0.3097 R-Square C.V. CALFWT2 Mean 0.084515 20.50860 51. 56277357 Tests of Hypotheses using the Type III MS for LEVEL*BLOCK as an error term Source DF Type III 55 F Value Pr > F LEVEL 3 911.82076041 3.30 0.0993 Tests of Hypotheses using the Type III MS for LEVEL*BLOCK as an error term Contrast linear in level DF Contrast·5S F Value Pr > F 1 788.27400753 8.56 0.0264 Tests of Hypotheses using the Type III M5 for LEVEL*BLOCK as an error term Contrast control vs othe.rs DF Contrast 5S F Value Pr > F 1 681.64425126 7.41 0.0346 Tests of Hypotheses using the Type III MS for LEVEL*BLOCK as an error term Contrast control vs 30 DF Contrast S5 F Value Pr > F 1 902.62088763 9.81 0.0203 �193 Table 4. Analysis of level effects on calf weights into of the pasture for animals that had not been exposed to treatment during the previous year. Dependent Variable: CALFWT2 Calf Wt. into Pasture (kg> Source DF Sum of Squares F Value Pr > F Model 23 2582.60168367 1.03 0.4327 Error 111 12072.78073069 Corrected Total 134 14655.38241435 R-Square C.V. CALFWT2 Mean 0.176222 20.95520 49.76802617 Tests of Hypoth~ses using the Type III MS for LEVEL*BLOCK as an error term Source LEVEL DF Type III SS F Value Pr > F 3 320.17215152 2.98 0.1183 Tests of Hypotheses using th~ Type III MS for LEVEL*BLOCK as an error term Contrast linear in level DF 1 Contrast SS F Value Pr > F 108.66582810 3.03 0.1322 Tests of Hypotheses using the Type III KS for LEVEL*BLOCK as .an error term Contrast control vs others DF 1 ss F Value Pr > F 26.31300408 0.73 0.4243 F Value Pr > F 3.52 0.1098 Contrast Tests of Hypotheses using the Type III MS for LEVEL*BLOCK as an error termContrast control vs 30 DF 1 Contrast SS 125.98786452 �194 Table 5. Analysis of treatment effects on cow weights out of the pasture. Weight into the pasture used as a covariate. Source DF Sum of Squares F Value Pr > F 259.20 0.0001 Model 27 2506670.67384 Error 321 114976.27262 Corrected Total 348 2621646.94646 R-Square -C.V. COWWT3 Mean 0.956143 4.611057 410.44136229 Tests of Hypotheses using the Type III MS for LEVEL*BLOCK as an error term Source LEVEL DF Type III SS F Value Pr > F 3 10098.2755001 0.82 0.5280 Tests of Hypotheses using the Type III MS for LEVEL*BLOCK as an error term Contrast linear in level DF Contrast SS F Value Pr > F 1 7688.19114~03 1.88 0.2199 Te_sts of Hypotheses using the Type III MS for LEVEL*BLOCK as an error term Contrast control vs others DF Contrast SS F Value Pr > F 1 5297.15925425 1.29 0.2990 Pr > F Tests of Hypotheses using the Type III MS for LEVEL*BLOCK as an error term Contrast control vs 30 DF Contrast SS F Value 1 9345.20106116 2.28 0.1'818 �195 Table 6. Analysis of level effects on calf weights out of the pasture. into the pasture used as a covariate. Dependent Variable: COWWT4 Cow Wt. in Nov. (kg) Source DF Sum of Squares F Value Pr > F Model 27 1387620.27444 76.55 0.0001 Error 295 198063.05616 Corrected Total 322 1585683.33060 R-Square C.V. COWWT4 Mean 0.875093 5.630769 460.17495866 . Dependent Variable: COWWT4 Cow Wt. in Nov. (kg) Tests of Hypotheses using the Type III MS for LEVEL*BLOCK. as an error term Source LEVEL DF Type III SS F Value Pr > F 3 1827.83167023 1.26 0.3704 Contrast 55 F Value Pr > F 31.43407666 0.06 0.8076 Tests of Hypotheses using the Type III MS for LEV.EL*BLOCK as an error term Contrast linear in level DF 1 Tests of Hypotheses using the Type III MS for LEVEL*BLOCK as an error term Contrast control vs others DF Contrast 5S F Value Pr > F 1 1149.45664340 2.37 0.1748 Contrast 55 F Value Pr > F 162.37424777 0.33 0.5841 Tests of Hypotheses using the Type III MS for LEVEL*BLOCK as an error term Contrast control vs 30 DF 1 Weight � https://cpw.cvlcollections.org/files/original/d8e3d34335d9f1fbe05e530e5782cb36.pdf e43924251705d46137dd522c25f6e5e7 PDF Text Text 1 Colorado Division of Wildlife Wildlife Research Report October 1991 JOB FINAL REPORT State of _......:::C~o~l~o~r~a~d~o:..._ _ Project Work Plan Avian Research - Migratory Game Birds W-152-R-4 _1_ Job 17 Job Title: Habitat use by wintering mallards near the Front Range of Colorado Period Covered: Author: 1 April 1990 through 31 March 1991 James K. Ringelman Personnel: R. Falise, Colorado State University; L. Roberts, Front Range Land and Cattle Company. ABSTRACT All data from this project have been collected and partially analyzed. Analyses continued early in this segment, but were later temporarily suspended while the principal investigators explored the feasibility of using a Geographical Information System (GIS) as a more desirable tool for data analysis and reduction. Training on appropriate GIS systems will take place next segment, with data analysis and manuscript preparation scheduled for completion next segment. Final analyses and publication preparation will be conducted under Work Plan 22, Job 2, Migratory Game Bird Publications. Prepared by: James K. Ringelman Wildlife Researcher C ��3 Colorado Division Wildlife Research September 1991 of Wildlife Report JOB FINAL REPORT State of __~C~o~l~o~r~a~d~o~ Project 1__ : Job Job Title: Winter Author: Avian Research W-152-R-4 Work Plan Period _ Covered: Clint W. Jeske, Game Birds 18 Survival 1 April - Migratory and reproductive 1990 through success 31 March James K. Ringelman of female mallards 1991 and Michael R. Szymczak Personnel: J. Ringelman, M. Szymczak, Colorado Division of Wildlife; G. White and B. Wunder, Colorado State University; D Anderson, J. Armstrong, J. Boulanger, K. Frye, M. Gilbert, D. Gilbert, Colorado Cooperative Wildlife Research Unit; M. Nail and R. Schnaderbeck, U.S. Fish and Wildlife Service, Monte Vista National Wildlife Refuge. ABSTRACT Mallards (Anas platyrhynchos) wintering in the San Luis Valley, Colorado were trapped and banded on the Monte Vista National Wildlife Refuge from December through March 1986-87,1987-88, and 1988-89. Body mass and a condition index for each bird were recorded to compare mass and condition among the age-sex classes; to identify :trends in.body mass through the winter; and to compare consistency of condition, relative to the population, of birds captured in subsequent years. Generally, .adults lla.,rl greater mass than Lmmatiuxe birds, and males had greater mass than .f'ema Les. .Females tended .to be in .be t ce r condition than males., and adults in better condition than immatures. Birds recaptured 'Lri subsequent years tended to be in similar relative condition classes. The low body masses recorded for mallards in the San Luis Valley may reflect responses to local environmental conditions (high elevation, predictable food resources, and avian predation) rather than poor nutrition. Carcass searches were conducted on the Monte Vista National Wildlife Refuge from January through April 1988 and January through April 1989. In 1988 and 1989, we recovered 4,604 and 1,738 carcasses, respectively. Avian cholera, which was a minor source of mortality in 1989, was prevalent in 1988. The sex ratio of carcasses recovered differed from the sex ratio of the population determined from aerial photographs in 1988 (X2 ~ 80.8, f < 0.0001), but not in 1989 (X2 ~ 0.3386, f - 0.5607). Starvation, determined from lipids in the ulna, was more prevalent among mallards which had been banded than unmarked birds. Logistic regression was used to determine relations of condition at time of banding and ��5 INTRODUCTION This final report is written primarily in dissertation format. The dissertation is composed of chapters that will be submitted for publication. All future work on this job by project personnel will be covered under Work Plan 22, Job 2 - Migratory Bird Publications. CHAPTER 1 BODY MASS AND CONDITION OF MALLARDS WINTERING IN THE SAN LUIS VALLEY, COLORADO Evans and Smith (1975) defined condition as "a measure of the chances of survival of an individual at a particular time of year, and/or of its potential for breeding successfully." Condition can be quantified by comparing body reserves of biologically important compounds, such as lipid and protein, in relation to environmental demands at any particular time. For waterfowl, body mass is often used in conjunction with external measurements to estimate lipid reserves and calculate a condition index (Bailey 1979, Chappell and Titman 1983, Gauthier and Bedard 1985, Ringelman and Szymczak 1985). These condition indices are then used to quantify changes in condition. Several important events (e.g., prebasic molt and pairing) occur during winter for migratory waterfowl (Heitmeyer 1985). Studies of cross-seasonal interactions among body condition and timing of biological events suggest body condition and hunting mortality are related (Bain 1980, Greenwood et al. 1986, Hepp et al. 1986, Reinecke and Schaiffer 1988), as are body mass and annual survival rate estimates (Haramis et al. 1986). Hepp (1986) found body mass and age to be related to time of pairing in American black ducks (~ rubripes). Pattenden and Boag (1989) found pairing and l~ying date to be related to winter body mass for captive mallards (~ p1atyrhynchos), but no relation between clutch size and body mass was found. Because body reserves used in waterfowl reproduction often accumulate before arrival on the breeding grounds (Heitmeyer and Fredrickson 1981), numerous studies have documented when and where reserves are accumulated (i.e., Krapu 1981, Heitmeyer 1985, Delnicki and Reinecke 1986, Hohman et al. 1988, laGrange and Dinsmore 1988). Here, we report on variations of body mass and condition for a semi-resident mallard population wintering under severe environmental conditions. Our objectives were to (1) identify trends in body mass and condition indices of marked birds from January through March; (2) compare body mass and condition of marked and unmarked mallards within and between winters; and (3) identify environmental components that may influence condition dynamics. STUDY AREA This study was conducted in the San Luis Valley (SLV), Colorado. The SLV is a l2,960-km~ intermountain basin in south-central Colorado, bounded by the San Juan Mountains to the west and the Sangre de Cristo Mountains to the east. Soils within the SLV are primarily fine sandy clay with little humus (Fenneman 1931). As a result of the surrounding mountains and a valley �6 elevation of 2,286 to 2,438 m, the SLV has an arid climate characterized by short, cool summers and cold, dry winters (Lantis 1942). The growing season is 100 to 120 days beginning in mid-May, but frost and snow may occur any day of the year (Rama1ey 1942). The cool, arid climate and sandy soils have resulted in a 'cold desert' plant association (Oosting 1956). Greasewood (Sarcobatus vermiculatus) and rabbit brush (Chrysothamous spp.) are the dominant vegetative species on the SLV floor (Lantis 1942, Enright 1971, u.s. Department of Interior 1979). In very alkaline soils, saltgrass (Distichlis stricta) is the only vegetative cover (U.S. Department of Interior 1979). Sedges (Carex spp.), rushes (Juncus balticus), spikerush (E1eocharis spp.), and saltgrass are common plants along wetland edges (Enright 1971, Szymczak 1986). Cattail (~ latifolia) and bulrush (Scirpus validus and~. paludosis) are the dominant emergent species (Enright 1971). Although precipitation within the SLV is low, spring run-off from the surrounding mountains provides ample water for irrigation (Hopper et al. 1975). Streams entering the northern part of the SLV percolate under the ground and continue flowing toward the center of the SLV under a layer of clay. This water emerges in central portions of the SLV as artesian wells (Carhart 1932). During the 1970's, irrigation techniques changed from flood irrigation to center pivot sprinklers. Water removed by center pivot sprinklers and reduced surface flows lowered the water table and subsequently reduced the number of artesian wells (U.S. Department of Interior 1979). The small human population in the SLV « 38,000) is concentrated in towns along the Rio Grande and Conejos Rivers (Mason 1974). The local economy is agriculturally based (Mason 1974). Major crops are irrigated potatoes, barley, vegetables, and hay (Hopper et al. 1975). Livestock production occurs throughout the SLV (Mason 1974). From 1964 through 1980, breeding duck estimates averaged 25,371 pairs, but declined after 1980 primarily because of decreasing habitat (Szymczak 1986). Mallards are the most abundant nester (Szymczak 1986). An average of 18,734 ducks winter in the SLV (1984-90 average, Colo. Div. Wi1d1., unpubl. data); most are mallards on the Monte Vista National Wildlife Refuge (MVNWR). Wintering mallards rest on open water maintained by pumping or artesian flow, and forage in nearby grain fields. METHODS Body Mass and Condition Mallards were captured on the MVNWR with Salt Plains bait traps (Szymczak and Corey 1976) and cannon nets (Dill and Thornberry 1950) in December 1986 and 1988, January 1987, 1988, and 1989, and February-March 1987, 1988, and 1989. Age and sex of all birds captured were determined using plumage characteristics (Carney 1964), and body mass (± 2 g), and wing length (to nearest mm) were recorded, and a condition index (estimated body fat / estimated fat-free body mass; Ringelman and Szymczak 1985) calculated for each. All mallards were banded with standard U.S. Fish and Wildlife Service leg bands. Mean body mass and condition indices were calculated for each trapping period. The Shapiro-Wilk statistic (Shapiro and Wilk 1965) was calculated using the UNIVARIATE Procedure (SAS Institute 1985) to test for normality of body mass and condition index distributions by age-sex class for each trapping period. Duncan's multiple range test was used to compare body mass and �7 condition index means by trapping period and age-sex class. To examine relations among body condition and trapping method, body mass of mallards captured in bait traps and cannon nets were compared for birds captured 11-18 January 1988 and 9-10 February 1989. . Body mass and condition index for mallards banded in 1987 or 1988 and recaptured in subsequent years were examined for similarity between years using linear regression. Body masses and condition indices were adjusted by subtracting the mean body mass and condition index for the corresponding agesex class and banding period to develop a relative condition index for additional between-year comparisons. Linear regression was used to examine relations between relative body mass and·condition in the banding year and subsequent years. Energy Intake Apparent metabolizable energy (AME) was measured for 3 grains grown in the SLV and fed upon by wintering mallards. Ten captive adult mallards (5 males and 5 females) were fed test food gg libitum for a minimum of 21 days before testing to assure gut morphology had adjusted to the diet (D. Johnson, Colo. State Univ., pers. comm.). Birds were placed in isolation cages and trials commenced when all birds were eating and had returned to their preacclimation body mass. If birds had not returned to pre-acclimation body mass, they would have been excluded from the trails, but all did return to pre-trial body mass. Before the 7-day trial, 8 meals exceeding gg libitum intakes were weighed to the nearest gram and divided into 2 portions, 1 to be fed in the morning and the other in the evening. Birds were weighed the day before trial initiation and again at its completion. Ducks were denied food for 8 h before the first day of the trial to allow food in the intestinal tract to be passed. Throughout the trial, birds were maintained at thermoneutrality (26°C, Smith and Prince 1973). Dry mass of meals fed, feces produced, and food not eaten were recorded. A composite fecal sample from days 3-6 was used for gross energy and nitrogen content measurements. After drying and grinding, composite grain and fecal samples were frozen and later analyzed for gross energy with a bomb calorimeter. Nitrogen content was determined by the Kjeldahl method, then multiplied by 6.25 to estimate protein content (Robbins 1983). The difference between mass of food fed and grain spilled was considered Lnt.ake.;differences between energy: intake and energy excre t ed was termed AME (Mhi'~~ ·~tici Reinecke 1984). T-tests were used to cOInpare'AME among foods. Determination of Foraging Rates Twelve captive birds (6 of each sex) were transported daily to an 64-ml enclosure erected in a field of standing barley on the MVNWR. After a 4 week acclimation period, birds appeared to be foraging in a manner similar to wild ducks. The enclosure was then moved to a site with abundant standing grain, and the birds were denied food for 18 h. Six mallards, 3 of each sex, were weighed to the nearest gram immediately before the trial. All captives were released ~ ~ into the enclosure, allowed to forage for precisely 5 min, then captured and reweighed. The amount of barley ingested was assumed to equal the difference between starting and ending body mass. �8 Estimation of Energy Expenditure Standard metabolic rates (SMR) were measured using the open circuit oxygen consumption method. Captive adult mallards, held in a pen with indoor and outdoor access and subject to natural photoperiod, were supplied with food (commercial poultry food and barley) and water ~ libitum. Individual test birds were denied access to food and water a minimum of 8 h before taking measurements. Birds were placed in a darkened respirometer (35.5 x 17.8 x 27.3 cm) between 2000 and 0400 h and room air pumped through the chamber at the rate of 1.6 l/min. Flowmeters measured flow into and out-of the chamber. For analysis, air from the respirometer was passed through columns of soda lime and drierite to remove COl and HlO, respectively. Oxygen content of the dried effluent air was measured with a Beckman F3 oxygen analyzer. Room air and a standard gas with 20.27% oxygen were used for calibration. All measures were adjusted to standard temperature and pressure. Lower critical temperature was estimated by taking mean SMR measures at -10°C and a line developed through the mean SMR measure and the intersection with the x axis at 42°C (mallard body temperature, unpubl. data) to estimate thermal conductance (McNab 1980). The point at which this line intersected the mean value of SMR measures made within the thermoneutral zone (Smith and Prince 1973) was identified as the lower critical temperature. Mean lipid reserves for each age-sex class were calculated by year to estimate distances the birds could travel if they migrated in response to severe weather. Flight energy was assumed to be 14 x SMR (Berger et al. 1971). Mean lipid levels were multiplied by the lipid energy content (assumed to be 9 kcal/g; Ricklefs 1974) to estimate energy reserves. Potential flight range was estimated by dividing energy reserves by flight energy (flight hours) and multiplying flight hours by the flight speed of mallards (assumed to be 64 kmjh; Whyte and Bolen 1988). Survival times were calculated using the same SMR and energy content estimates. Daily energy expenditure of free-living mallards in autumn was estimated as 3.0 x SMR (Prince 1979). Year-specific energy reserves were divided by daily energy expenditure to provide estimates of survival time, in the absence of food (fasting endurance), for the average mallard age-sex class. Mean monthly temperatures and departures from normal, total precipitation, snowfall, and departure of total precipitation from normal were obtained from monthly National Oceanic and Atmospheric Administration summaries for the Alamosa Airport, located 30 km east of the MVNWR. Generally, temperatures at the Alamosa Airport tended to be < 2°C lower than at the MVNWR. Snowfall in the SLV was variable, with the MVNWR usually receiving more snow than the Alamosa Airport. RESULTS Condition Indices Population ~.-Distributions of body masses and condition indices of mallards captured during each trapping period were not different from normal distributions (£ > 0.1) for each age-sex class. Body mass and condition indices were greatest in 1987-88 and least in 1988-89 (Tables 1.1 and 1.2). Adult females were in better condition than the other age-sex classes. Adult males and immature females were in similar condition, whereas �9 Table 1.1. Yearly and overall mean mass (SD) of mallards of different age-sex classes captured Dec-Mar, 1986-89, in the San Luis Valley, Colorado. Comparisons with letters a-c denote differences in the means among years (a 0.05, Duncan's Multiple Range Test). Letters w-z denote differences among age-sex classes within a year (a - 0.05, Duncan's Multiple Range Test). Yearly overall means are grand means rather than weighted means. Age Sex Adult Female Adult Male Immature Female 1986-87 933(87)by 948(78)ay 1988-89 909(77)cy Overall 931(82)y (N - 192) (N - 411) (N - 324) 1,043(100)bw (N - 573) 1,088(93)aw (N - 884) 1,028(82)cw 1,054(95)w (N - 602) (N - 2,059) (N - 927) 904(82)az 554) 896(81)az 262) (N - 844(74)bz 887(83)z (N - 201) (N - 1,017) 978(90)bx 594) 1,015(89)ax (N - 595) (N - (N - Immature Male 1987-88 (N - 959(81)cx 987(90)x 244) (N - 1,473) Table 1.2. Yearly and overall mean (~) condition indices of mallards of different age-sex classes captured Dec-Mar, 1986-89, in the San Luis Valley, Colorado. Comparisons with letters a-c denote differences in the means among years (a - 0.05, Duncan's Multiple Range Test). Letters w-z denote differences among age-sex classes within a year (a - 0.05, Duncan's Multiple Range Test). Yearly overall means are grand means rather than weighted means. Age Sex 1986-87 1987-88 1988-89 Overall Adult Female 14.5(4.9)bw (N - 192) l5.3(4.5)aw (N - 411) 13.2(4.6)cw (N - 324) Adult Male 12.7(4.7)bx (N - 573) 14.9(4.2)aw G! - 884) 13.5(4.5)x 12.7(4.1)bw (N - 1,059) (N - 602) Immature Female l4.0(4.5)aw (N - 262) 14.3(4.5)ax (N - 554) l2.0(4.6)bwx 13.5(4.7)x (N - 201) (N - 1,017) Immature Male 11.6(4.4)by (N - 594) 13.2(4.2)ay (N - 595) 10.9(4.2)bx 12.0(4.4)y (N - 1,473) (N - 244) 14.4(4.7)w (N - 927) �10 immature males were in the poorest condition. Body mass and condition of adult males and females and immature males were greatest in 1987-88 (Tables 1.1 and 1.2). Males had greater condition indices in 1987-88 than the other 2 years, whereas condition indices of adult females were different in all three years. Immature females had similar body mass and condition in 1986-87 and 1987-88, and both parameters were less in 1988-89. Each year, adult females were in the best and immature males the poorest condition. Adult males and immature females were in similar condition in 1987-88 and 1988-89, but immature females were in better condition in 198687 (Table 1.2). Condition indices of adult females generally reached lowest levels in January before beginning to increase through early March (Table 1.3). Condition of immature females decreased through February, but increased thereafter. Condition indices of adult and immature males were greatest in the first trapping period each year, whether in December or January, and rapidly dropped to a relatively constant level for the rest of the trapping season. Table 1.3. Mean (SQ) condition index, by trapping period, of mallards captured in the San Luis Valley, Colorado. Letters a-d compare means by agesex class within a banding period (a - 0.05, Duncan's Multiple Range Test). Letters t-z compare means by banding period, within an age-sex class (a 0.05, Duncan's Multiple Range Test). Year 1986-87 1988 1988-89 Trapping Period Adult Female 19.9(4.4)at (N - 18) Jan 13.6(4.7)avw (N - 121) Feb 14.8(3.7)auv (N - 56) Jan 15.9(4.4)au (N - 274) Jan-Feb 13.0(5.2)avw (N - 43) Feb 12.1(3.1)aw (N - 18) Feb-Mar 14.6(4.0)auv (N - 91) 13.9(6.0)auvw Dec (N - 21) 12.3(4.6)aw Jan (N - 219) Feb 14.0(3.3)auvw (N - 25) Feb-Mar 15.3(3.8)auv (N - 77) Dec Adult Male 18.0(3.6)abt (N - 78) 12.5(3.9)bwx (N - 343) 10.4(3.5)cx (N - 165) 15.4(4.0)au (N - 783) 13.3(3.4)avw (N - 34) 10.9(3.0)ax (N - 46) 12.1(3.5)bwx (N - 152) 14.6(4.3)auv (N - 42) 12.4(4.2)awx (N - 553) 11.7(3.1)bcwx (N - 37) 12.0(3.8)bwx (N - 134) Immature Female Immature Male 17.1(4.7)bt 16.5(3.8)bt (N - 30) (N - 63) 13.7(4.0)atuv 11.8(4.0)buv (N - 169) (N - 330) 13.3(3.8)btuv 9.7(3.6)cv (N - 69) (N - 216) 14.8(4.3)btu 13.9(3.8)ctu (N - 390) (N - 495) 14.0(3.3)atu 10.6(4.6)bv (N - 45) (N - 16) 9.7(4.9)av 9.7(2.9)av (N - 22) (N - 11) 12.6(4.5)buv 10.0(3.8)cv (N - 122) (N - 144) 13.9(4.3)atuv 12.6(3.6)auv (N - 19) (N - 18) 10.9(4.4)auv 10.6(4.4)av (N - 163) (N - 225) 13.1(4.5)abtuv 10.5(2.7)cv (N - 15) (N - 20) 11.9(4.4)buv 10.5(3.7)cv (N - 47) (N - 57) �11 Effect Qf Capture Method. - In January 1988, mallards captured in cannon nets were in better condition than those captured in bait traps (Table 1.4). Differences were significant for all age-sex classes except adult females (~- -4.17, 623 df, £ < 0.0001; ~ - -1.83, 416 df, £ - 0.0679; ~ - 2.47, 348 df, f - 0.0139; and ~ - -1.16, 130.8 df, £ - 0.2499 for adult males, immature males, immature females,and adult females respectively). On 9-10 February 1989, condition indices were greater for mallards captured with cannon nets (Table 1.4), but only differences for immature males approached significance (~ - -1.64, 30 df, £ - 0.1108; f > 0.15 for other age-sex class comparisons). Table 1.4. Mean condition indices (SE) of mallards captured 11-18 January 1988 and 9-10 February 1989, in the San Luis Valley, Colorado, in cannon nets and Salt Plains bait traps. Year Trap Method Male b,gylt Female Immat1U:~ Male Female 1988 Cannon Net 16.8(3.8) n - 88 16.4(3.8) n - 63 14.9(3.6) n - 38 16.7(3.6) n - 32 Bait Trap 14.9(4.1) n - 537 15.8(4.7) n - 191 13.7(3.8) n - 380 14.7(4.3) n - 318 Cannon Net 12.5(3.1) n - 56 12.4(3.8) n - 37 11.2(2.5) n - 18 12.3(4.2) n - 28 Bait Trap 11.7(3.0) n - 22 12.4(3.9) n - 12 9.5(3.2) n - 14 9.8(3.3) n-6 1989 Recaptureg Indivigyals. - Mallards banded and then recaptured in 1988 had lower mass and condition indices at recapture than at the time of initial capture (l - 16.96, 1,204 df, f - 0.0001 and l - 5.83, 1,204 df, f - 0.0166, for body mass and condition indices, respectively). At recapture, immature birds had lower body mass than unmarked birds trapped at the same time, whereas adult males recaptured were similar to unmarked birds. In 1989, recaptures did not differ in body mass (l - 2.67, 1,500 df, £ - 0.1026) or condition index (l - 0.51, 1,499 df, f - 0.4753) between initial capture and recapture. All age-sex classes of unmarked and marked birds showed similar body mass each banding period. One hundred and twenty mallards banded in 1986-87 were recaptured in 1987-88. Only 4 adult females from 1986-87 were recaptured in 1988, and no relation of condition or relative condition (condition index - mean condition index) between the 2 years was detected (f 0.00, 1,3 df, £ - 0.9998 and l -0.00, 1,2 df, £ - 0.9892 for condition and relative condition, respectively). Condition in 1986-87 was related to condition in 1987-88 for adult males (f - 13.39, 1,50 df, £ - 0.0006), and immature males (f - 8.66, 1,36 df, £ - 0.0057), but not for immature females (f - 0.44, 1,5 df, £ - 0.5362). Relative condition classes showed the same relations (Table 1.5; Fig. 1.1 for males) �12 co CO I 15 I"CO 12 (J) ~ 9 C 6 C 3 a 0 C (3) a U CD > ......., ~ --...- Immature ·····0······ ......, ."0 Adult o o , 00 O. .:" (6) o • ..'. . -~0 ~• 0 ~ , .' 0_,," • - 8.,-0;. _ .,0 • -:'0 ",'. • • " , • -• 0 •• " (9) . • o " (12) CD a: (15)(15) (12) (9) (6) (3) 0 3 6 9 12 15 Relative Condition in 1986-87 Figure 1.1. Relative condition indices of adult and immature males captured in 1986-87 and recaptured in 1988 in the San Luis Valley, Colorado. The solid line represents the least squares linear regression line for adult males and the dotted line the regression line for immature males. �Table 1.5. Linear regression relations between condition index of mallards at the time of initial capture and recapture in subsequent years in the San Luis Va1ley, Colorado. l:;aIlt1U;:!il Initial X!ilSlI': Recapture a& bb r2 i'" Error df £ Age Sex Adult Male 1986-87 1986-87 1988 1988 1988-89 1988-89 0.46 -0.71 0.16 0.40 -0.34 0.23 0.82 1.33 0.05 8.93 9.13 4.92 49 32 91 0.0044 0.0050 0.0290 Female 1988 1988-89 0.42 -1.08 0.21 4.30. 17 0.0546 Male 1986-87 1986-87 1988 1988 1988-89 1988-89 0.97 -0.81 0.38 0.97 0.01 0.18 0.44 0.720.27 18.32 3.43 13.68 30 17 38 0.0002 0.0827 0.0007 Female 1986-87 1986-87 1988 1988 0.34 0.38 0.24 1988-89 -0.82 2.39 0.31 1988-89 0.96 -0.30 0.58 1.59 1.40 16.05 6 4 13 0.2634 0.3224 0.0017 Immature • Slope of the regression function. b Intercept of the regression function. C Two-tailed test of the hypothesis of the slope ~ o. Fifty-seven mallards banded in 1986-87, including only 1 adult female, were recaptured in 1988-89. Condition indices in 1986-87 were positively related to condition indices in 1988-89 for adult males (£ - 9.24, 1,31 df, f - 0.0048), as were relative condition indices (Table 1.5; Fig. 1.2). Neither condition indices nor relative condition indices were related for immature mallards (~ - 2.57, 1,16 df, f - 0.1283 and ~ - 1.56, 1,3 df, £ - 0.3001, for male and female condition indices, respectively; Table 1.5 for relative condition indices; and Fig. 1.2 for immature males). One hundred and sixty-three mallards captured in 1987-88 were recaptured in 1988-89. Condition indices of adult females captured in 1987-88 were not related to indices in 1988-89 (~ - 2.22, 1,16 df, f - 0.1553), but other agesex classes were positively related (~- 5.44, 1,90 df, f - 0.0219; ~ - 17.14, 1,37 df, f - 0.0002; and ~ - 15.18, 1,12 df, f - 0.0021 for adult males, immature males, and immature females, respectively). Relative condition indices in 1987-88 were positively related for all age-sex classes (Table 1.5; Figs. 1.3 and 1.4). Metabolizable Energy Three grains grown in the SLV were measured for metabolizable energy: triumph and steptoe barley, and field peas. Gross energy of steptoe barley was 18.85 (~ - 0.24) kj/g, of triumph barley l7.90(~ - 0.19) kj/g, and of field peas 18.37 (~ - 0.05) kj/g. Overall, birds maintained on steptoe barley gained 13.2 (~ - 21.6) g, those on triumph barley gained 20 (~ 33.3) g, whereas those fed field peas lost an average of 19.3 (~ - 19.8) g. Apparent metabolizable energy was similar for steptoe barley and field peas (~- 0.27,1 df, f - 0.4161), while the AME of triumph barley was less than the �14 . I i 0') CO I CO CO 15 -.- Adult 0') "r- iO c: Immature ······0······ X CD 5 "0 c: c: a • c: a 0 ,~" 0 0cl' ~ " . 0,'· •o· 0 • -+-' "0 • •° °• • .. . (5) .: ,, ._~ (10) -+-' -C'UCD (15) a: (15) ,, ,, • ' • • •• 0 o (10) (5) 0 5 10 15 Relative Condition Index in 1986-87 Figure 1.2. Relative condition indices of adult and immature males captured in 1986-87 and recaptured in 1988-89 in the San Luis Valley, Colorado. The solid line represents the least squares linear regression line for adult males and the dotted line the regression line for immature males. :.~'~'. .. �15 15 -..- (j) COI CO CO (j) .,.... .e e _ a ...._ "'C e a 0 CD Adult 12 9 6 0 • • •• • I· 3 •• •• 0 (3) .,' 0 •• ••• 0 o • (6) ._ ...-> (9) CD (12) •• • (15\15) ., • • • 0 • • ~ a:: •• Immature .....-0 ...... • (12) (9) (6) (3) 0 3 6 9 12 15 Relative Condition in 1987-88 Figure 1.3. Relative condition indices of adult and immature males captured in 1987-88 and recaptured in 1988-89 in the San Luis Valley, Colorado. The solid line represents the least squares linear regression line for adult males and the dotted line the regression line for immature males. �16 (j') COI 15 CO 12 CO (j') .,.... 9 C 6 C a .+-' .- -.-- Adult Immature ······0······ • 0:> o..~·· 3 o 0 • "'0 C (3) a 0 (6) (].) > .- (9) - m (12) ~ (15)(15) • • 0 • ..- .' • • o • (12) (9) (6) (3) 0 3 6 9 12 15 Relative Condition in 1987-88 Figure 1.4. Relative condition indices of adult and immature females captured in 1987-88 and recaptured in 1988-89 in the San Luis Valley, Colorado. The solid line represents the least squares linear regression line for adult females and the dotted line the regression line for immature females. �17 others (~ ~ 6.15,1 df, f ~ 0.0513 and ~ - 4.61, 1 df, f ~ 0.0680 for comparisons of triumph and steptoe barley and field peas, respectively; Table 1. 6). Table 1.6. Mean (£Q) daily intake, fecal energy (Kj/day), and apparent metabolizable energy (Kj/day) of field peas and barley grown in the San Luis Valley, Colorado. Food Intake energy (Kj/day) Fecal energy (Kj/day) Apparent metabolizable energy (Kj/g intake) Peas 787.9(232.8) 193.6(67.0) 13.91(0.42) Steptoe 984.0(213.6) 258.2(51.9) 13.85(0.56) Triumph 764.5(185.4) 210.1(49.8) 12.97(0.26) Feeding Trials Drift in the electronic balance used in the field-feeding trials resulted in some erroneous body mass measurements; these measurements were excluded from analyses. Females ingested barley at the rate of 0.92 g/min (~ - 0.77, range 0.0-2.4 g/min, n - 27) and males consumed 1.50 g/min (SQ - 0.99, range 0.0-4.0, n - 31) from fields on MVNWR. There were differences in the rate of consumption between sexes (F - 6.05, 1,54 df, ~ - 0.0170) and among individuals (F - 8.69, 1,54 df, ~ - 0.0047), but no individual*sex interaction was detected (F - 0.89, 1,54 df, ~ - 0.3507). Males consumed standing grain heads, whereas females were only observed foraging on shelled grain. Standard Metabolic Rate SMR trials require that the birds be inactive long enough for a stable rate of oxygen consumption to be identified (approximately 45 min). Only 11 SMR measurements were obtained under these conditions (Fig. 1.5). The basal rate was estimated as 368 (~ - 75, n - 5) kj/day, the mean of the values obtained at 20-30 °c. At -9 to -10°C, SMR was 675 (~ - 189, n - 6) kj/day. Based on these measures, the lower critical temperature is about 13 to 14°C. DISCUSSION Although age and sex variations in body mass of SLV mallards are similar to those found in other wintering mallard populations (Sugden et al. 1974, Owen and Cook 1977, White 1982, Jorde et al. 1984, Delnicki and Reinecke 1986, Whyte et al. 1986), mean body mass appears to be lower (Table 1.7). However, structural size does not seem to account for the lower body masses of SLV wintering mallards, because wing lengths, a common index of structural size (Ringelman and Szymczak 1985), were no smaller for SLV mallards than for other wintering populations (Table 1.8). Environmental conditions could, however, account for low mean body masses either directly by resulting in a negative energy balance (the forced hypothesis), or indirectly where low body mass is an adaptive response to conditions (the adaptive hypothesis). I used �18 1,100 • 1,000 900 • 800 700 > CO --o_ 600 0 500 ~ 400 300 200 100 0 (20) Figure 1.5. (10) 0 10 20 30 40 Ambient Temperature Standard metabolic rate measurements of captive, post- absorptive, adult mallards measured using the open circuit oxygen consumption method at temperatures above and below the lower critical temperature in the San Luis Valley, Colorado. 50 �19 energetic data to explore these alternative hypotheses. Table 1.7. Mean body mass (g) for wintering mallard populations. Location Reference· Alberta Male agYlt Female Male Imm51tyte Female Male Qo!Ut!ined Female 1,221 (1) N - 1,479 England New Jersey (2) 1,183 N. - 594 1,068 N. - 774 1,111 N - 135 1,106 N - 96 1,400 106 (3) N Missouri (4) Missouri (5) Missouri (6) Mississippi (7) Texas (8) Oklahoma (9) SLV (10) 1,020 N. - 103 1,102 N - 143 1,220 N - 215 1,246 1,095 N - 1,308 N - 453 1,237 1,088 N - 87 N. - 42 1,196 1,052 N - 92 N - 71 1,053 930 N - 2,059 N - 927 1,059 H - 703 1,264 N - 97 990 N - 88 1,133 N - 66 1,040 1,181 N - 169 N - 188 1,214 996 N - 19 N - 20 1,123 976 1,160 H - 105 N - 119 984 881 1,019 N - 1,473 N - 1,017 1,008 906 • (1) Sugden et a1. (1974); (2) Owen and Cook (1977); (3) Figley and VanDruff (1982); (4) White (1982); (5) Heitmeyer (1985); (6) calculated from Combs (1987); (7) De1nicki and Reinecke (1986); (8) Whyte et al. (1986); (9) Gordon (1981); (10) this study. �20 S Table 1.8. Mean wing lengths populations. (mm) reported for wintering mallard Reference Adult Male Owen and Cook (1977) 274 338 N- Female W'hite (1982) N Delnicki and Reinecke (1986) 293 1,308 297 N - 2,059 N This study a 259 465 275 - 103 N- 276 N - 453 281 N - 927 Male N N N Immature Female 273 - 121 286 - 169 284 - 1,473 257 N - 94 272 N - 88 N N 270 - 188 271 - 1,017 Techniques for measuring wing length may have varied between studies. The Forced Hypothesis Severe environmental conditions, such as low ambient temperatures or food limitations, could regulate boodymass and condition of mallards wintering in the SLV. If ambient temperature directly controls body condition by increasing mallard thermoregulatory demands, then their condition should have been related to ambient temperatures before and during our trapping efforts. Yet, mallards in the SLV were in better condition in 1987-88 than the other two trapping years, even though December and January temperatures were lowest in 1987-88 (Table 1.9). Therefore, ambient temperature does not appear to directly control body condition. Similarly, no relations are apparent among condition and December and January snowfall (Table 1.9). These results were not unexpected, because SMR measurements suggest that mallards acclimate well to cold temperatures. Like many other avian species, such acclimatization may include the ability to restrict carbohydrate use (Marsh and Dawson 1982), and vasoconstriction in the bill and legs to reduce heat loss (King and Farner 1961, Hagan and Heath 1980). �21 Table 1.9. Mean and departures from normal for monthly temperature (OC), total and departure from normal for precipitation (em), and snowfall (em) recorded at the Alamosa Airport, Alamosa, Colorado, December - March 1986-87, 1987-88, and 1988-89. Month Year Dec Jan Feb &r Dec Jan Feb Mar Dec Jan Feb Mar Temperature Precipitation Mean Departure from Normal Total Snow Departure from Normal 1986 -6.4 1987 -10.3 1987 -4.8 1987 -0.6 1987 -9.6 1988 -15.3 1988 -8.2 1988 -0.3 1988 -7.9 1989 -9.4 1989 -5.8 1989 2.9 1.3 -1.4 0.3 -0.3 -1.8 -6.4 -3.1 -0.1 -0.1 -0.6 -0.7 3.1 0.3 1.7 1.2 0.7 1.3 0.7 0.6 O.~ 0.3 0.8 0.7 0.3 4.8 32.5 17.8 9.9 19.1 15.2 7.4 8.6 3.6 13.2 7.6 2.5 -0.6 1.0 0.6 -0.2 0.4 -0.0 -0.0 -0.5 -0.6 0.1 0.1 -0.7 Using mean ambient temperatures to determine daily energy requirements of SLV mallards may cause an overestimation of these energy requirements. SLV mallards roost on open water areas where water temperatures may be several degrees above freezing. In addition, the steam that rises from the water surface may produce a warmer microclimate than general SLV ambient temperatures as recorded at the Alamosa Airport. Also, mallards may use daily solar radiation (Lustick 1969, Heppner 1970) and may alter their behavior (Lustick 1973, Brodsky and Weatherhead 1985) to reduce thermoregulatory energy costs. Proximate factors, such as snow cover and waste grain density, are important in determining energy intake by field-feeding ducks. Jorde et al. (1983) reported that mallards in Nebraska use fields in which grazing cattle trampled the snow, thereby making grain available to ducks. Likewise, waterfowl wintering on the MVNWR must field-feed because little food remains in wetlands. Captive mallards chose durum wheat and barley over hard spring wheat (Clark et al. 1986), a preference that was attributed to the difference in handling time (i.e., the time required for the bird to remove the kernel from the head and ingest it). Clark et al. suggested the rate of energy intake was more important than energy content of the food. Our observations of females not using standing grain heads suggest different handling times for males and females. Prince (1979) suggested free-living mallards required 3.0 x SMR for daily energy expenditure (DEE). Based on monthly mean SLV temperatures during 1964-88, if the lower critical temperature is 13°C and SMR at -10°C is 675 kj/day (SMR measurements), then a mallard would have to consume 129.9 g of barley/day in November, 150.3 g of barley/day in December, and 154.5 g of barley/day in January to meet DEE. Therefore, a female mallard would have to field-feed for 141 min/day in November, 162 min/day in December, and 168 min/day in January. A male would need to forage 87 min/day in November, 99 �22 min/day in December, and 102 min/day in January to meet DEE. A SLV mallard would need to consume 13.35 kg of triumph barley from 1 November through 31 January to meet DEE. Thus, a wintering population of 20,000 mallards would require 267 metric tons of triumph barley from November through January. If triumph barley yields 51.4 bujha and bushel mass is 24.5 kg/bu (Dillion 1989), triumph barley would produce 1,259 kgjha. Consequently, 212 ha would be needed to support 20,000 wintering mallards from November through January assuming the ducks completely utilized the grain. During the three years of our study, from 96 to 167 ha were cultivated on MVNWR (S. Brock, pers. comm.), suggesting that the low body masses observed may result from insufficient food supplies if, in fact, 212 ha are needed to support the SLV population and mallards only foraged on the refuge. Although food may be a factor limiting body condition of SLV mallards, the year with greatest grain production (167 ha produced 38,500 bu in 1989; S. Brock, pers. comm.) was also the year that mallards had the lowest condition indices. Conversely, in 1986 and 1987, substantially less grain was produced on MVNWR (23,600 and 26,300 bu in 1986 and 1987, respectively; S. Brock, pers. comm.), but mallards were in better condition. Presently, grain availability in fields adjacent to MVNWR appears to have little effect on mallard body condition from December through February. Mallards may have foraged in grain fields adjacent to MVNWR to supplement intake, although our observations suggest few mallards forage off MVNWR from December through February. In summary, low body masses ("poor" condition) of mallards wintering in the SLV are not entirely attributable to food availability because the winter with the most food available corresponded to the year mallards were in poorest condition. Additionally, ambient temperatures and snow fall do not account for yearly differences in winter body condition. The Adaptive Hypothesis There may be selective advantages for low winter body mass in the SLV. Consistency of relative condition indices between years suggests that body condition is controlled endogenously. In support of this, captive mallards in the SLV were maintained with ~ libitum barley supplemented with high-protein poultry pellets; these birds maintained body masses similar to wild birds captured in the SLV (mean mass for 8 adult males was 1,178 g (SQ - 79) and for 9 adult females was 1,114 g (SQ - 86) on 5 January 1989; unpub1. data). Selective forces that may determine optimum winter body mass are the risks of starvation and predation (Lima 1986): body mass should decrease as the predictability of food resources increases or as predation risk rises. In Lima's model, optimal body mass decreases with increases in the frequency of predator attacks or probability that attacks will be successful. Advantages of reduced body mass would be reduced foraging time and an increased probability of evading predator attacks. Conversely, heavier body mass reduces the risk of starvation in unpredictable environments. Predictable, but not plentiful, food resources for wintering SLV mallards are barley grown and left standing in fields on the MVNWR and waste grain in fields adjacent to the refuge. Mean lipid reserves would allow birds to survive 2.5-4 days (Table 1.10) at 3-4 times SMR (Prince 1979) or to make the 400 km flight to the Rio Grande River Valley in New Mexico or the 600 km Table 1.10. Mean body mass, estimated basal metabolic rate, mean estimated lipid reserves, energy content of lipid reserves, and estimated flight range �23 and survival time for San Luis Valley mallards. Age Sex Body mass SMR Fat (g) (kcal/h) (g) Fat energy (kcal) Flight range (km) 933 948 909 3.48 3.52 3.41 120 127 107 1,080 1,143 963 1,421 1,485 1,293 4.3 4.5 3.9 1986-87 1,043 1987-88 1,088 1988-89 1,028 3.78 3.90 3.74 119 143 112 1,071 1,287 1,008 1,293 1,510 1,235 3.9 4.6 3.7 896 904 844 3.38 3.40 3.24 112 114 84 1,008 1,026 756 1,363 1,382 1,069 4.1 4.2 3.2 978 1986-87 1987-88 1,015 1988-89 959 3.60 3.70 3.55 101 120 91 909 1,080 819 1,152 1,331 1,056 3.5 4.1 3.2 Year Adult Female 1986-87 1987-88 1988-89 Adult Male Immature Female 1986-87 1987-88 1988-89 Immature Male Survival time (d) flight to the Southern High Plains in Texas (Table 1.10). During 3 field seasons, snow sufficient to cover the barley left standing in refuge fields occurred only for about 1 week in January 1987 (Table 1.9); during that period, many birds succumbed to starvation (unpubl. data), but I did not observe any obvious migratory flights of mallards leaving the SLV. Few data have been presented to substantiate that predators remove weakened individuals from small mammal or avian populations (Temple 1987). Temple suggested that the degree to which predators removed weakened individuals was influenced by how difficult the prey was to kill. Difficultto-kill species would show the greatest deviations (e.g. numbers of parasites, body condition, age, number of injuries) between the individuals killed and the population. Factors that help determine how difficult prey is to kill include environmental complexity, escape maneuvers of the prey, and once caught, how difficult the prey is to dispatch. In the SLV, raptors account for most mallard predation. Mallards may increase their probability of detecting predators (Moriarty 1976), and the complexity of their environment by flocking (Rubenstein 1978, Pulliam and Millikan 1982). When attacked by a raptor, escape maneuvers are a mallard's last defense. Most avian predators' easily kill mallards once the birds are caught. Bald eagles (Ha1iaeetus 1eucocepha1us) were the primary predator of waterfowl wintering in the SLV, although we also observed golden eagles (AQuila chhYsaetos), peregrine falcons (~ pere~rinus), and prairie falcons (~ mexicanus) capturing wintering waterfowl. Attacks usually occurred when waterfowl were roosting. When a raptor approached, the flock flushed and circled until the raptor either made a kill or ended the attack. Birds separated from the main flock were usually targeted for attack. When singled out by a raptor, a flying mallard's principle escape maneuver was to attempt to outfly the predator until the mallard could land on open water and dive, or land in vegetation and hide. Lower body mass of mallards wintering in the SLV may be advantageous in escaping avian predators, because reductions in wing �24 loading (g/cm2 wing area) increase maneuverability (Greenwalt 1975). In 198788, when mallard body condition was greatest, a severe outbreak of avian cholera resulted in numerous carcasses; eagles scavenged these carcasses, thereby reducing their predation on live birds. Low body mass of mallards wintering in the SLV may also reduce flight energetics at the high elevation (2,300 m). Most interest in altitude physiology has focused on the effects of hypoxia (e.g., Monge and Whittembury 1976, Cerretelli 1987). Little interest has been shown on other effects (such as reduced air density) of high-altitude environments. At the 2,300 m elevation of the SLV, air density is about 79.8% of the air density at sea level (Ava~lone and Baumeister 1987). Energy per unit distance flown at 2,300 m altitude would be 25.4% greater than a similar flight at sea level when all variables other than air density are held constant (Greenwalt 1975). Alternatively, if variables other than body mass are constant, a bird flying a given distance at sea level could weigh 12.0% more and expend the same energy as a bird flying the same distance at 2,300 m (Greenwalt 1975). However, mallards wintering in Mississippi (26-33 m elevation) weighed 17-20% more than mallards wintering in the SLV (Delnicki and Reinecke 1986; Table 1.7); which is more than the estimated body mass differences expected if flight energetics alone limited SLV mallard body mass in the winter. I conclude that low body mass of mallards wintering in the SLV does not directly reflect stressful winter conditions, but instead may be a response to an unusual wintering environment. Although ambient temperatures in ~he SLV appear severe, microclimates of winter roosts probably are not severe. Grain production appears to be insufficient for this wintering population, but years with greatest grain production on MVNWR did not result in greater condition indices. Altitude may be an important factor controlling body mass by influencing flight energetics and the ability to elude avian predators. While aspects of both the forced and adaptive hypotheses probably influence body mass of mallards wintering in the SLV, my evidence suggests low body masses are more adaptive than forced. My results emphasize the need to include environmental parameters other than temperature and precipitation when evaluating an animal's adaptive responses to environmental conditions. LITERATURE CITED Avallone, E. A., and T. Baumeister, III. (eds.) 1987. Mark's standard handbook for mechanical engineers. McGraw-Hill Book Co., New York. 1,239pp. Bailey, R. O. 1979. Methods of estimating total lipid content in the redhead duck (Aythya americana) and an evaluation of condition indices. Can. J. Zool. 57:1830-1833. Bain, G. A. C. 1980. The relationship between preferred habitat, physica~ndition and hunting mortality of canvasbacks (Aythya valisineria) and redheads (Aythya americana) at Long Point, Ontario. M. S. Thesis, Univ. Western Ontario, London. 38pp. Berger, M., J. S. Hart, and O. Z. Roy. 1971. Respiratory water and heat loss of the black duck during flight at different temperatures. Can. J. Zool. 49:767-774. �25 Brodsky, L. M., and P. J. Weatherhead. 1985. Variability in behavioural response of wintering black ducks to increased energy demands. Can. J. Zool. 63:1657-1662. Carhart, A. H. 1932. Colorado. Coward - McCann, Inc., New York, NY. 322pp. Carney, S. M. 1964. Preliminary keys to waterfowl age and sex identification by means of wing plumage. U. S. Fish Wildl. Serv., Spec. Sci. Rep. -- Wildl. 82. 47pp. Cerretelli, P. 1987. Extreme hypoxia in air breathers: some problems. Pages 137-150 in P. Dejours, ed. Comparative physiology of environmental adaptations, Vol. 2. 8th Conf. European Soc. Camp. Physiol. Biochem. Strasbourg 1986. S. Karger AG, Basel, Switzerland. 224pp. Chappell, W. A., and R. D. Titman. 1983. Estimating reserve lipids in greater scaup (Aythya marila) and lesser scaup (a. affinis). Can. J. Zool. 61:35-38. Clark, R. G., H. Greenwood, and L. G. Sugden. 1986. Influence of grain characteristics in optimal diet of field-feeding mallards Anas platyrhynchos. J. Appl. Ecol. 23:763-771. Combs, D. L. 1987. Ecology of male mallards during winter in the Upper Mississippi Alluvial Valley. Ph.D. Thesis, Univ. of Missouri, Columbia. 22lpp. Delnicki, D., and K. J. Reinecke. 1986. Mid-winter food use and body weights of mallard and wood ducks in Mississippi. J. Wildl. Manage. 50:43-51. Dill, H. H., and W. H. Thornberry. 1950. A cannon projected net trap for capturing waterfowl. J. Wildl. Manage. 14:132-137. Dillion, M. A. 1989. 1989 small grain research report. Colorado State University Cooperative Extension, Center, Colo. 14 pp. Enright, C. A. 1971. An analysis of mallard nesting habitat on the Monte Vista National Wildlife Refuge, Colorado. M.S. Thesis, Colorado State University, Fort Collins. l13pp. Evans, P. R., and P. C. Smith. 1975. Studies of shorebirds at Lindisfarne, Northumberland. 2. Fat and pectoral muscle as indicators of body condition in the bar-tailed godwit. Wildfowl 26:64-76. Fenneman, N. M. 1931. Physiography of western United States. McGraw-Hill Co., Inc., New York, NY. 534pp. Figley, W. K., and L. W. VanDruff. 1982. The ecology of urban mallards. Wildl. Monogr. 81:1-39. �26 Gauthier, G., and J. Bedard. 1985. Fat reserves and condition indices in greater snow geese. Can. J. Zool. 63:331-333. Gordon, D. H. 1981. Condition, feeding ecology, and behavior of mallards wintering in northcentral Oklahoma. M. S. Thesis, Oklahoma State Univ, Stillwater, OK. 66pp. Greenwalt, C. H. 1975. The flight of birds. Trans. Amer. Phil. Soc. 65:1-67. Greenwood, H. R., R. G. Clar~, and P. J. Weatherhead. 1986. Condition bias of hunter-shot mallards (Anas platyrhynchos). Can. J. Zool. 64:599-601. Hagan, A. A., and J. E. Heath. 1980. Regulation of heat loss in the duck by vasomotion in the bill. J. Therm. BioI. 5:95-101. Haramis, G. M., J. D. Nichols, K. H. Pollock, and J. E. Hines. 1986. The relationship between body mass and survival of wintering canvasbacks. Auk 103:506-514. Heitmeyer, M. E. 1985. Wintering strategies of female mallards related to dynamics of lowland hardwood wetlands in the upper Mississippi delta. Ph.D. Thesis, Univ. Missouri, Columbia. 378pp. _____ , and L. H. Fredrickson. 1981. Do wetland conditions in the Mississippi Delta hardwoods influence mallard recruitment? Trans. N. Am. Wi1dl. Nat. Resour. Conf. 46:44-57. Hepp, G. R. 1986. Effects of body weight and age on the time of pairing of American black ducks. Auk 103:477-484. ----- , R. J. Blohm, R. E. Reynolds, J. E. Hines, and J. D. Nichols. 1986. Physiological condition of autumn-banded mallards and its relationship to hunting vulnerability. J. Wildl. Manage. 50:177183. Heppner, F. 1970. The metabolic significance of differential absorption of radiant energy by black and white birds. Condor 72:50-59. Hohman, W. L., T. S. Taylor, and M. W. Weller. 1988. Annual body weight change in ring-necked ducks (Aythya collaris). Pages 257-269 in M. W. Weller, ed. Waterfowl in winter. Univ. Minnesota Press, Minneapolis. Hopper, R. M., A. D. Geis, J. R. Grieb, and L. Nelson, Jr. 1975. Experimental duck hunting seasons, San Luis Valley, Colorado, 1963-1970. Wildl. Monog. 46:1-68. Jorde, D. G., G. L. Krapu, and R. D. Crawford. 1983. Feeding ecology of mallards wintering in Nebraska. J. Wildl. Manage. 47:1044-1053. �27 King, J. R., and D. S. Farner. 1961. Energy metabolism, thermoregulation and body temperature. pp. 215-288 in A. J. Marshall, ed. Biology and comparative physiology of birds. Vol. II. Academic Press, New York, NY. Krapu, G. L. 1981. The role of nutrient reserves in mallard reproduction. Auk 98:29-38. LaGrange, T. G., and J. J. Dinsmore. 1988. Nutrient reserve dynamics of female mallards during spring migration through central Iowa. Pages 287-297 in M. W. Weller, ed. Waterfowl in winter. Univ. Minnesota Press, Minneapolis. Lantis, D. W. 1942. Sage of the San Luis Valley in Colorado. Folks and Fortunes 1:38-40. Lima, S. L. 1986. Predation risk and unpredictable feeding conditions: determinants of body mass in birds. Ecology 67:377-385. Lustick, S. I. 1969. Bird energetics: effects of artificial radiation. Science 163:387-390. _____ . 1973. Cost-benefit of thermoregulation in birds: influences of posture, microhabitat selection, and color. Pages 265-294 in W. P. Aspey and S. I. Lustick, eds. Behavioral energetics: the cost of survival in vertebrates. Ohio State Univ. Press, Columbus, OH. Marsh, R. L., and W. R. Dawson. 1982. Substrate metabolism in seasonally acclimatized American goldfinches. Am. J. Physiol. 242:R563-R569. Mason, J. 1974. San Luis Valley resource area -- social -- economic profile. Western Interstate Commission for Higher Education, Boulder, CO. 2l9pp. McNab, B. K. 1980. On estimating thermal conductance in endotherms. Physiol. Zool. 53:145-156. Miller, M. R., and K. J. Reinecke. 1984. Proper expression of metabolizable energy in avian energetics. Condor 86:396-400. Monge, C., and J. Whittembury. 1976. High altitude adaptations in the whole animal. Pages 289-308 in J. Bligh, J. L. Cloudsley-Thompson, and A. G. MacDonald, eds. Environmental physiology of animals. John Wiley & Sons, New York. 456pp. Moriarty, D. J. 1976. The adaptive nature of bird flocks: a review. The Biologist 58:67-79. Oosting, H. J. 1956. The study of plant communities. W. H. Freeman and Co., San Francisco, CA. 440pp. Owen, M., and W. A. Cook. 1977. Variations in body weight, wing length and condition of mallard Anas platyrhynchos and their relationship to environmental changes. J. Zool. (Lond.) 183:377-395. �28 Pattenden, R. K., and D. A. Boag. 1989. Effects of body mass on courtship, pairing, and reproduction in captive mallards. Can. J. Zool. 67:495-501. Prince, H. H. 1979. Bioenergetics of postbreeding dabbling ducks. Pages 103-117 in T. A. Bookhout, ed. Waterfowl and wet1ands--an integrated review. Proc. 1977 Symp. Northcent. Sect., Wild1. Soc., Madison, Wis. 152pp. Pulliam, H. R., and G. C. Millikan. 1982. Social organization in the nonreproductive season. Pages 169-197 in.D. S. Farner, J. R. King, and K. C. Parks, eds. Avian Biology, Vol. VI. Academic Press, New York, NY. Ramaley, F. 1942. Vegetation of the San Luis Valley in southern Colorado. Univ. Colo. Studies. Series D. Physical and Biological Sciences. 1:237-279. Reinecke, K. J., and C. W. Shaiffer. 1988. A field test for differences in condition among trapped and shot mallards. J. Wi1d1. Manage. 52:227-232. Ricklefs, R. E. 1974. Energetics of reproduction in birds. Pages 152-297 in R. A. Paynter, Jr., ed. Avian energetics. Nuttall Ornith. Club, Cambridge, Mass. Ringelman, J. K., and M. R. Szymczak. 1985. A physiological condition index for wintering mallards. J. Wildl. Manage. 49:564-568. Robbins, C. T. 1983. Wildlife feeding and nutrition. Academic Press, New York, NY. 343pp. Rubenstein, D. I. 1978. On predation, competition, and the advantages of group living. Pages 205-231 in P. P. G. Bateson and P. H. Klopfer, eds. Perspectives in ethology, vol. 3. Plenum Press, New York, NY. SAS Institute, Inc. 1985. SAS procedures guide for personal computers, version 6 edition. Cary, N.C. 373pp. Shapiro, S. S., and M. B. Wilko 1965. An analysis of variance test for normality (complete samples). Biometrika 52:591-611. Smith, K. G., and H. H. Prince. 1973. The fasting metabolism of subadult mallards acclimatized to low ambient temperatures. Condor 75:330335. Sugden, L. G., W. J. Thurlow, R. D. Harris, and K. Vermeer. 1974. Investigations of mallards overwintering at Calgary, Alberta. Can. Field-Nat. 88:303-311. Szymczak, M. R. 1986. Characteristics of duck populations in the intermountain parks of Colorado. Tech. Publ. 35. 88pp. �29 _____ , and J. F. Corey. 1976. Construction and use of the Salt Plains duck trap in Colorado. Colorado Div. Wildl., Div. Rep. 6. l3pp. Temple, S. A. 1987. Do predators always capture disproportionately from prey populations? substandard individuals Ecology 68:669-674. U.S. Department of Interior. 1979. Final environmental Basin Division, San Luis Valley Project, Alamosa counties. Colorado Bur. Rec. FED 79-37. v.p. statement, Closed and Saguache White, D. C. 1982. Leaf decomposition, macro invertebrate production wintering ecology of mallards in Missouri lowland hardwood wetlands. M. S. Thesis, Univ. Missouri, Columbia. 264pp. Whyte, R. J., G. A. Baldassarre, and E. G. Bolen. 1986. Winter condition of mallards on the southern high plains of Texas. J. Wildl. Manage. 50:52-57. _____ , and E. G. Bolen. 1988. Flight ranges and lipid dynamics mallards wintering on the Southern High Plains of Texas. Ornithol. 59:143-148. and of J. Field �30 CHAPTER RELATIONS 2 AMONG WINTER BODY CONDITION AND SURVIVAL OF MALLARDS WINTERING IN THE SAN LUIS VALLEY, COLORADO Nonhunting mortality of waterfowl is an important component of the compensatory versus additive mortality issue (e.g. Anderson and Burnham 1976; Conroy and Eberhardt 1983; Nichols et al. 1984). Most reports of nonhunting mortality are related to extreme environmental conditions that result in starvation or disease outbreaks (Stout and Cornwell 1976). Other estimates of nonhunting mortality have focused on lead poisoning losses (e.g., Bellrose 1959, Zwank et al. 1985). Regardless of the cause, estimates of nonhunting mortality are difficult to obtain because scavengers remove carcasses quickly and birds often hide before they die, making the carcass more difficult to find (Humburg et al. 1983, Stutzenbaker et al. 1986). Many reported deaths probably reflect mortality that swamped the ability of local scavengers to remove carcasses. Factors which may decrease an individual's survival probability are difficult to quantify. Many waterfowl studies have quantified protein and lipid reserves through the nonbreeding season as a measure of condition (e.g., Baldassarre et al. 1986, Miller 1986, Whyte et al. 1986, Heitmeyer 1988, Whyte and Bolen 1988, Hohman et al. 1988, Serie and Sharp 1989), but few have quantified how body reserves are related to survival or reproduction. Most studies that have attempted to quantify the relation between body condition and survival have focused on relations among condition and hunting mortality. Hepp et a1. (1986) found a negative relationship between body condition and direct recovery probability of autumn-banded mallards (Anas platyrhynchos). Greenwood et a1. (1986) also found hunter-killed birds weighed less than their collected sample. Haramis et al. (1986) reported relations between earlywinter body mass and overwinter and annual survival rates for canvasback (Aythya va1isineria). Krementz et a1. (1989) did not find a relation between late-winter body mass and annual survival rates for American black ducks (Anas rubripes), but the power of their tests was low. Here, we report on winter mortality of marked mallards in a severe winter climate. Our objectives were to (1) document and identify causes of nonhunting mortality, (2) estimate the percent recovery of carcasses, and (3) determine if survival was related to body condition. METHODS Mallard Handling and Marking Mallards on the Monte Vista National Wildlife Refuge (MVNWR) were captured with Salt Plains bait traps (Szymczak and Corey 1976) and cannon nets (Dill and Thornberry 1950) in December 1986 and 1988, January 1987, 1988, and 1989, and February-March 1987, 1988, and 1989. After removal from the trap, mallards were placed in cages lined with straw and transported back to a �31 central processing area. Upon arrival, birds were unloaded from the cages and placed in holding pens lined with straw in a darkened building to dry and to digest grain remaining in their crops. Birds in the holding pens were disturbed when birds were captured for processing or more birds were brought from the traps. This disturbance caused birds to thrash against the fabric sides of the holding pens and the ground, which often frayed their primary tips. Birds remained in the holding pens varying periods of time, from about 2-24 hr. For processing, mallards were removed from the holding cages in groups of 10-20 and carried to another building where data on age, sex, body mass, and wing length were obtained, and a condition index (estimated body fat / estimated fat-free body mass; Ringelman and Szymczak 1985) calculated for each. All mallards were banded with standard U.S. Fish and Wildlife Service leg bands. During 11-18 January 1988 and 12-17 January 1989, most mallards also had monel bands affixed to the shaft of primary feather VIII of both wings. Only leg bands were applied during the other trapping periods. After processing, birds were placed in poultry crates until all crates were full (approximately 120 birds), then birds were transported to within 200 m of open water and released simultaneously. Releases generally occurred during the daylight hours, but some birds (generally around 100 per day) were released during twilight or after dark during the January trapping sessions. Handling Effects To determine possible handling effects, mallards recaptured within 7 days of their initial capture were plotted as a percent of the bird's initial weight. Linear regression was used to test for any relation between duration of time between recapture and differences in initial and recapture body masses. Carcass Collection and Processing In 1987, carcass remains were collected incidentally while conducting other field work. After marking in 1988 and 1989, waterfowl concentration areas and raptor perches on the MVNWR were regularly searched to collect remains of dead waterfowl. Waterfowl concentration areas and eagle perches were searched by walking and/or wading the periphery of the wetland or perch while collecting all carcass remains. Each wing was assumed to represent 1 carcass. When 2 wings were found attached, 1 wing was retained for analyses, the other had the primaries clipped and was transported with the carcasses for disposal by burning. Remains were picked up by the bill, bagged to avoid disease transmission (Friend 1987a), and the location noted. Carcasses were transported to a central location for processing. When possible, species, age, sex, location recovered, and band numbers (if bands were present) were recorded for each carcass. In 1988, every fifth mallard wing, excluding those from banded carcasses, was frozen in airtight bags for later analyses of lipids in the ulna and Pasturella in the radius marrow. One wing from each marked (wing or leg banded) carcass and one wing from each carcass in 1989 was retained for the same analyses. Hutchinson and Owen (1984) suggested that lipids within the ulna are exhausted immediately before a bird succumbs to starvation. To further clarify the use of ulna lipids as an index of malnutrition, I captured 12 mallards, sacrificed the birds after they had been denied food for varying time periods and extracted lipids from the ulna and the carcass (Appendix A). �32 A relation between body lipids and ulna lipids was found where ulna lipids appeared to be depleted when carcass lipids were around 4-5%. Refinement of the technique led to a simple blotter test, wherein an ulna is broken and the contents touched to blotter paper to reveal lipid residue. If a visible lipid· spot remained after being allowed to dry, then the bird had not exhausted its lipid stores (Appendix A). No tests were made to determine if lipids leached from the ulna over time. Wings from 1987, 1988, and 1989, were tested and classified as to the status of lipid reserves. From mid-December 1988 through mid-March 1989, up to 10 complete mallard carcasses were collected weekly and frozen for subsequent necropsies. Carcasses were necropsied and lesions noted (Wobeser 1981). Ulna lipids were used to assess the status of lipid reserves. Impression smears were made of the marrow of a longitudinal slice of the radius and the hepatic parenchymal surface, stained with methylene blue and examined for bipolar-staining bacteria. Cultures were made from the same tissues and examined for growth of Pasturella multocida. Subsequent diagnoses were used to estimate relative importance of the different mortality factors. Percent Carcass Recovery Carcasses with monel bands attached to primary VIII of both wings were planted around wetland roost concentration areas to obtain estimates of the percentage of dead birds being found. Carcasses were placed in locations where dead birds had been previously found. In 1988, 25 carcasses were placed around 3 roost areas searched every·third day by field technicians. In 1989, 65 carcasses were placed around 5 roost areas searched every third day. Only the person planting the carcasses knew when carcasses were planted and their locations. At the end of the field season, percent of the planted carcasses recovered was calculated each year as an estimate of the percent of carcasses recovered. No estimates were made of the number of leg-banded birds recovered. Sex Ratios Aerial photographs of roosting waterfowl were taken 4 February 1988 and 18 January 1989 from an elevation of 150 m. Male and female mallards were counted from selected transparencies in which sex identification was possible. These sex ratio estimates were assumed to represent the population. Sex ratios of carcasses collected were compared with those from the photographs using a Chi-square test to determine if nonhunting mortality was sex-specific. Relations of Body Condition and Survival Body condition of birds found dead within the January-April period of the banding year, reported as recovered during the subsequent hunting season, and recaptures the year after banding were analyzed to identify relations among recovery status and condition using logistic regression in the CATMOD Procedure (SAS Institute Inc. 1987). We used forward, stepwise, univariate selection procedures with a - 0.25 as criteria for inclusion in multivariate models (Hosmer and Lemeshow 1989). Age and sex were bivariate variables included in the screening process because different age/sex mean condition indices may lead to incorrect inferences of relations among body condition and survival. Condition index is treated as a continuous variable. Because some cells contained zero frequencies, maximum-likelihood estimates were obtained for those cells (SAS Institute Inc. 1987). Differences in the likelihood ratio Chi-square of the model with and without the variable of interest are �33 Chi-squares with 1 df, and were used to determine significance of a variable (Hosmer and Lemeshow 1989). RESULTS Handling Effects In 1988, mean body mass of birds recaptured within 7 days of initial capture was less than at initial capture (Fig. 2.1). The difference in mass approached significance while that for condition was not significant (r 3.026, 1,114 df, f - 0.0846 and r - 0.917, 1,114 df, f - 0.3402, for body mass and condition indices, respectively). In 1989, mass and condition were less when recaptu~ed within 7 days, than at initial capture (Fig. 1.2; r - 4.70, 1,307 df, f - 0.0309 and r - 4.91, 1,307 df, f - 0.0274, for body mass and condition indices, respectively). Table 2.1. Numbers of mallard carcasses found and bands collected each field season in the San Luis Valley, Colorado. Year Number of Mallard Carcasses Found Number of Bands Found Percent 1987 229 30 13.1 1988 4,193 201 4.8 1989 1,663 169 10.2 Carcass Collections In 1987, carcasses were collected incidental to other field work, resulting in limited search time; only 229 remains, which included 30 leg bands, were recovered (Table 2.1). In 1988 and 1989, variable effort was devoted to searching for carcasses (Table 2.2). Between 4 January and 1 April 1988 and between 17 January and 1 April 1989, 6,281 carcasses were collected (5,856 mallards), which included 370 banded birds (Table 2.1). Other carcasses were primarily Canada geese (Branta canadensis; 140 carcasses) and northern pintails (Anas acuta; 128 carcasses). �34 CIJ115 CIJ C'U ~ Adult • 110 ~ "0 o 105 c: .- 100 o 95 c: Q) o ~ 90 co C'U 0) ~ ..•....• (23) i (12) o (24) I (6) (5) ~ (19) .J.. (4) Q) CL Immature 850~----~----~2------~3------4~-----5~----~6 Days to Recapture Figure 2.1. Body masses (as % of initial capture mass) of adult and immature male mallards recaptured within 7 days of initial capture in the San Luis Valley, Colorado, January through March 1988. Sample sizes are shown in parentheses with adults above and immatures below the standard deviations. �35 00 115 ~--------------------------------------~ 00 C'O ~ Adult • 110 ~ '"C a CO Immature o 105 ~ -- en l... o 100 (3) 95 ..•...•. c CD o l... (4) 90 f CD CL 85 0 1 234 - (4) 5 6 Days to Recapture Figure 2.2. Body masses (as % of initial capture mass) of adult and immature female mallards recaptured within 7 days of initial capture in the San Luis Valley, Colorado, January through March 1988. Sample sizes are shown in parentheses with adults and immatures below the standard deviations. �36 CIJ 115 CIJ Adult Clj ~ (10) u a co - 105 ~ 100 -en (72) (~5) (34) 95 (1-7) I ~ (26) (10) 1 T I 1 I ~ o • 110 ~ 2 3 l) Immature (2) T 1 • 456 o (1 ) • 7 8 Days to Recapture Figure 2.3. Body masses (as % of initial capture mass) of adult and immature male mallards recaptured within 7 days of initial capture in the San Luis Valley, Colorado, December 1988 through March 1989. Sample sizes are shown in parentheses with adults above and immatures below the standard deviations. �37 en en 115 CO ~ Adult (1) ~ "0 a co - 105 ~ 100 -- (11 ) (7) - 01 \0- 0 ...-c CD o • 110 ~ I 95 I (7) (14) (10) 90 Immature (9) 0 •(7) 0 I (1 ) \0- CD n, 85 0 1 2 3 4 5 6 7 8 Days to Recapture Figure 2.4. Body masses (as % of initial capture mass) of adult and immature female mallards recaptured within 7 days of initial capture in the San Luis Valley, Colorado, December 1988 through March 1989. sizes are shown in parentheses with adults and immatures below the standard deviations. Sample �38 Table 2.2. Hours per week spent searching for waterfowl carcasses and number of carcasses found on MVNWR in 1988 and 1989. Weeks are encoded to start after the January banding effort (11-18 January 1988, and 12-17 January 1989). Week Hours of Search Time 1988 1989 1 2 3 4 5 6 7 8 9 10 11 12 52.0 39.0 27.8 29.8 33.2 33.8 33.5 18.0 12.5 21.5 43.2 20.5 17.0 24.0 3.5 20.0 7.0 0.0 19.5 7.0 11.0 4.5 Number of Carcasses Collected 1989 1988 544 591 383 469 543 399 391 265 169 172 184 93 237 157 45 197 182 3 169 101 95 22 In December 1987, an outbreak of avian cholera on MVNWR prompted premature closure of the refuge to waterfowl hunting. Throughout January and early February 1988, more than 400 carcasses per week were collected (Table 2.2). Although the number of carcasses collected per week declined through late February and early March, this was more a function of reduced search effort and dispersal of birds as wetlands thawed than reduced mortality. Avian cholera was the suspected cause of death in most recorded cases, but some waterfowl died from lead poisoning, starvation, predation, and collisions with wires. Ninety-three mallard carcasses were necropsied. Forty-five of the carcasses had petechiae on the epicard;.um, lesions associated with avian cholera; 8 had penetrating trauma wounds; 4 had exhausted lipid reserves; 2 appeared to have succumbed to lead poisoning; 1 had lesions of both avian cholera and lead poisoning; 1 appeared to have succumbed to aspergillosis; and 20 had no visible lesions. Only 14 cultures of the marrow and 6 hepatic cultures were positive for avian cholera. Four hepatic smears were positive for avian cholera. The strain of avian cholera in the SLV has proven difficult to culture at the National Wildlife Health Center (M. Miller, Colo. Div. Wildl., pers. comm.). In 1988, age and sex was determined from 735 mallard remains; two hundred and five (27.9%) were adult females, 307 (41.8%) were adult males, 106 (14.4%) were immature females, and 117 (15.9%) were immature males. Percentage of males in the carcass collection (57.7%) differed (X2 - 80.8, 1 df, f < 0.0001) from the percentage of males counted in photographs taken 4 February 1988 (79% of 803 mallards counted). In 1989, avian cholera resulted in high waterfowl mortality (Fig. 2.2, and Table 2.2). Of 591 carcasses aged and sexed, 238 (40.3%) were adult males, 96 (16.2%) were immature males, 162 (27.4%) were adult females, and 95 (16.1%) were immature females. Males, which were 56.5% of the carcasses, did not differ from that counted in the aerial photographs (58% of 1,135 mallards counted 18 January 1989; X2 - 0.3, 1 df, f - 0.5607). In 1987, 11 of 85 wings (12.9%) sampled for ulna lipids came from birds �39 that had depleted their lipid reserves. In 1988, more banded than unmarked birds had starved (29 of 201 [14.4%] marked and 31 of 699 [4.4%] unmarked mallards; X2 - 25.1, 1 df, f < 0.0001). Similar analyses in 1989 determined greater percentage of banded than unmarked birds starved (26 of 106 [24.5%] .banded and 42 of 911 [4.6%] unmarked birds; X2 - 60.4, 1 df, P < 0.0001). Percent Carcass a Recovery Bands were recovered from 17 (68%) of the planted carcasses in 1988. (31.5% of 130 wings), representing 23 birds (35.4%) of the birds were recovered. In 1989, 41 wings Relations of Body Condition and Survival Mortality of birds found dead within 3 months of banding was not related to body condition at banding in any year (Table 2.3). Logistic regression identified a relation between condition and recaptures in 1987 (X2 - 4.8, 1 df, f - 0.0293), with adult males in poorer condition and the other age-sex classes being in better condition then most other classification categories (Table 2.3). No condition relation among birds found dead or shot in 1987 was detected (X2 - 0.1, 1 df, f - 0.7913 and X2 - 0.9, 1 df, f - 0.3349, for birds found dead or shot by hunters, respectively). �40 Table 2.3. Mean (SO) condition indices of mallards banded in the San Luis Valley, Colorado in 1987-89 according to status: found dead January -- April of the banding year, direct hunting recovery, recaptured one year later, or not contacted again. Year Age 1987 Adult Sex Not Contacted Found Dead Male 12.8(4.3) N - 495 Female 14.7(5.0) N - 75 Immature Male 11.3(4.3) N - 535 Female 14.0(4.0) N - 248 1988 Adult Male 14.9(4.2) N - 717 Female 15.3(3.9) N - 207 Immature Male 13.2(4.1) N - 498 Female 14.4(4.2) N - 500 1989 Adult Male 12.1(4.2) N - 556 Female 13.0(4.6) N - 292 Immature Male 10.4(4.4) N - 218 Female 10.7(4.7) N - 186 12.7(4.5) N - 10 15.8 N - 1 11.1(6.2) N - 9 14.3(8.4) N - 3 15.7(4.0) N - 70 14.5(5.8) N - 14 13.4(4.2) N - 50 13.8(5.4) N - 35 11.8(4.7) N - 36 11.0(4.6) N - 6 10.0(2.8) N - 22 10.7(4.2 N - 13 Direct Recovery 12.8(5.0) N - 15 19.1(4.2) N - 2 13.3(3.8) N - 12 11.2(9.5) N - 4 15.6(4.6) N - 13 14.1(4.6) N - 3 12.5(6.2) N - 8 12.8(4.3) N - 5 Recapture 11.9(4.2) N - 53 16.7(3.3) N - 5 11.8(3.3) N - 37 15.4(4.9) N - 7 14.5(3.9) N - 83 16.3(2.7) N - 12 13.2(4.0) N - 39 13.7(4.4) N - 14 11.0(3.6) N - 7 13.6(2.3) N - 2 7.6(4.4) N - 3 In 1988 univariate analyses, condition did not meet the a - 0.25 screening criteria for inclusion in a multivariate model, so no multivariate model was tested. Only birds found dead January-April and direct hunting recoveries were available for analysis in 1989, and condition approached significance (X2 - 1.8, 1 df, f - 0.1859 and X2 - 0.3, 1 df, f - 0.5839 for found dead and direct recoveries, respectively). DISCUSSION Differences between percent of marked birds and unmarked birds which starved suggests either unmarked birds included many migrants that did not winter in the SLV or that handling affected survival probabilities of banded birds. Many mallards wintering in the SLV are thought to be residents (Szymczak 1986), but we are unable to estimate numbers of mallards migrating through the SLV to determine if our sample of dead mallards included many migrants. Similarities in condition indices among unmarked birds and shortterm recaptures trapped January through March each year (Chapter 1) suggests condition of marked birds through this period does not differ substantially �4~ from unmarked birds. Handling probably contributed to mortality of some marked birds. Several components of our handling could have affected the survival of marked birds. Some birds flew into fences when released and injured themselves. Other birds were released after dark and they may have been unable to find suitable roosting habitat, and predators may have killed some birds or others succumbed to the low ambient temperatures. Frayed primaries resulting from the birds attempting to escape holding pens may reduce flight efficiency and alter the bird's energy budget such that it may be more likely to succumb to starvation. Logistic regression did not detect a condition effect for banded birds found dead, suggesting that condition and any handling effects are not related. Recovery rates of marked carcasses placed around roost areas were similar to other studies (Humberg et al. 1983, Stutzenbaker et al. 1986). Humberg et al. (1983) reported a recovery rate of 61 (73%) of 79 planted carcasses. Their observations indicated half the carcasses were scavenged within 1 day and the remaining half removed from the area within 4 days. They found recovery of planted carcasses was greatest for visible species, such as snow geese (Chen caerulescens), and least for female dabbling ducks. Stutzenbaker et al. (1986) reported that 50% of the duck carcasses were removed within 1 day. They noted that carcasses restrained in open water persisted longer (11.2 days) than those placed along open (3.8 days) or in vegetative cover (2.6 days). Humberg et al. (1983) suggested shoreline searches cannot be used to provide reliable estimates of total nonhunting mortality. We had more of a controlled situation than just a shoreline search and used our estimates of carcass recovery and the total number of carcasses collected to estimate January through April losses. If carcass recovery in 1988 was 70%, then approximately 6,577 birds died. However, because search effort was variable and carcasses were planted only around roost areas searched, 70% probably overestimates the true recovery rate. The 1989 recovery rate (35%) was probably less than in 1988 because of reduced search effort (Table 2.1). The 35% recovery rate suggests that 4,791 birds died. Because an unknown number of migrants died in the SLV, these estimates cannot be used to directly estimate mortality rates of the wintering population. Mortality in 1988, a year with severe cholera, appeared to have been greater for females than males (57.7% of carcasses collected but 79% of the birds counted in the photographs in 1988 were males). In 1989, no differential mortality related to sex was identified. The 1988 results may have resulted from poor quality photographs which biased counts towards the more easily detectable males. Alternatively, environmental conditions that resulted in the severe avian cholera outbreak may have affected female susceptibility to avian cholera. ~cLandress (1983) found more male than female Ross' geese (Chen rossii) and lesser snow geese died during cholera epizootics, but no differences in species vulnerability was observed. Likewise, Friend (1987b) reported species losses during avian cholera outbreaks were similar to species composition in the area of the outbreak. The lack of any relations among condition at banding and probability of being found dead in January-April may be a result of the large number of avian cholera mortalities. Because susceptibility to avian cholera is not related to body condition (Wobeser 1981, McLandress 1983), and it rarely affects body condition before the bird dies (Wobeser 1981), any relations among body condition and other causes of mortality are relatively insignificant. �42 A threshold response may exist, whereby some minimum body condition is needed for survival, but additional lipid reserves do not increase survival probability substantially. Mallards in the SLV, on average, have lipid reserves to survive at least 2.5 - 4 days without food during the winter (Chapter 1). Food resources in the SLV may be predictable for wintering .ma11ards (Chapter 1), and a mallard may gain little in survival probability by increasing lipid reserves. If lipid reserves are insufficient for a bird to survive 2 days without food, survival probability may decrease substantially. A condition index of 5 - 6 represents about 50 g of lipids for both sexes, which equates to 1.8 days survival time (using a standard metabolic rate (SMR) of 85 kcal / day, energy content of lipids of 9 kca1 / g, and a free-living existence of 3.0 x SMR). In 1987 and 1988, few birds with condition indices of < 6.were reported as hunting mortality or recaptured. Most banded birds that starved had low condition indices (Table 2.4), so many low condition birds may die before the hunting season or trapping periods. Table 2.4. Mean (SD) condition indices at time of capture for banded mallards that died from malnutrition in the San Luis Valley, Colorado. Year Age Adult Sex Female Male Immature Female Male 1988 6.8(12.4) N - 2 12.9(5.0) N - 12 12.2(4.5) N - 9 9.4(12.6) N - 3 1989 4.7(4.5) N - 3 12.5(5.5) N - 10 8.0(0.9) N - 3 10.0(6.9) N - 2 Because annual survival rates of mallards banded preseason in the SLV (63%, 59%, 58%, and 65% for adult and immature males, and adult and immature females, respectively; Szymczak 1986) are similar to or greater than the continental mallard population (63%, 56% and 50% for adult males and females and immature of both sexes combined, respectively; Anderson 1975:22), avian cholera may be compensatory in this population. Low hunting mortality may result in winter mallard populations that exceed carrying capacity for MVNWR, resulting in dense winter concentrations that facilitate cholera transmission. The failure to identify relations among winter body condition and survival may be the result of small sample sizes and mortality resulting from avian cholera. These results do not support or refute the hypothesis that winter or annual survival is directly related to winter body condition. �43 LITERATURE CITED Anderson, D. R. 1975. Population ecology of the mallard. V. Temporal and geographic estimates of survival, recovery, and harvest rates. U.S. Fish Wi1d1. Serv., Resour. Pub1. 125. 110 pp. _____ , and K. P. Burnham. 1976. Population ecology of the mallard: VI. The effect of exploitation on survival. U.S. Fish Wi1d1. Serv., Resour. Pub1. 128. 66pp. Baldassarre, G. A., R. J. Whyte, and E. G. Bolen. 1986. Body weight and carcass composition of nonbreeding green-winged teal on the Southern High Plains of Texas. J. Wi1d1. Manage. 50:420-426. Be11rose, F. C. 1959. Lead poisoning as a mortality factor in waterfowl populations. Illinois Nat. Hist. Surv. Bull. 27:235-288. Conroy, M. J., and R. T. Eberhardt. 1983. Variation in survival and recovery rates on ring-necked ducks. J. Wi1dl. Manage. 47:127-137. Dill, H. H., and W. H. Thornberry. 1950. A cannon projected net trap for capturing waterfowl. J. Wi1dl. Manage. 14:132-137. Friend, M., ed. 1987a. Field guide to wildlife diseases. U.S. Fish Wi1dl. Serv., Resour. Publ. 167. 225 pp . . 1987b. Avian cholera. Pages 69-82 in M. Friend, ed. Field guide to wildlife diseases. Vol. 1. General field procedures and diseases of migratory birds. U.S. Fish Wildl. Servo Resour. Publ. 176. 225pp. Greenwood, H., R. G. Clark, and P. J. Weatherhead. 1986. Condition bias of hunter-shot mallards (Anas platyrhynchos). Can. J. Zool. 64:599-601. Haramis, G. M., J. D. Nichols, K. H. Pollock, and J. E. Hines. 1986. The relationship between body mass and survival of wintering canvasbacks. Auk 103:506-514. Heitmeyer, M. E. 1988. Body composition of female mallards in winter in relation to annual cycle events. Condor 90:669-680. Hepp, G. R., R. J. Blohm, R. E. Reynolds, J. E. Hines, and J. D. Nichols. 1986. Physiological condition of autumn-banded mallards and its relationship to hunting vulnerability. 'J. Wildl. Manage. 50:177-183. Hohman, W. L., T. S. Taylor, and M. W. Weller. 1988. Annual body weight change in ring-necked ducks (Aythya co11aris). Pages 257-269 in M. W. Weller, ed. Waterfowl in winter. Univ. Minnesota Press, Minneapolis. Hosmer, D. W., Jr., and S. Lemeshow. 1989. Applied logistic regression. John Wiley & Sons, New York, NY. 307pp. �44 Humburg, D. D., D. Graber, S. Sheriff, and T. Miller. 1983. Estimating autumn-spring waterfowl nonhunting mortality in north Missouri. Trans. N. Am. Wildl. Nat Res. Conf. 48:241-256. Hutchinson, A. E., and R. B. Owen. 1984. Bone marrow fat in waterfowl. J. Wildl. Manage. 48:585-591. Krementz, D. G., J. E. Hines, P. O. Corr, and R. B. Owen, Jr. 1989. The relationship between body mass and annual survival in American black ducks. Ornis Scand. 20:81-85. McLandress, M. R. 1983. Sex, age, and species differences in disease mortality of ROSS'· and lesser snow geese in California: implications for avian cholera research. Calif. Fish and Game 69:196-206. Miller, M. R. 1986. Northern pintail body condition during wet and dry winters in the Sacramento Valley, California. J. Wildl. Manage. 50:189-198. Nichols, J. D., M. Compensatory evidence and Amer. Wildl. J. Conroy, D. R. Anderson, and K. P. Burnham. 1984. mortality in waterfowl populations: a review of the implications for research and management. Trans. N. and Natur. Resour. Conf. 49:535-554. Ringelman, J. K., and M. R. Szymczak. 1985. A physiological condition index for wintering mallards. J. Wildl. Manage. 49:564-568. SAS Institute Inc. 1987. SAS/STAT guide for personal computers, Version 6 edition. Cary, NC. 1028pp. Serie, J. R., and D. E. Sharp. 1989. Body weight and composition dynamics of fall migrating canvasbacks. J. Wildl. Manage. 53:431441. Stout, I. J., and G. W. Cornwell. 1976. Nonhunting mortality of fledged North American waterfowl. J. Wildl. Manage. 40:681-693. Stutzenbaker, C. D., K. Brown, and D. Lobpries. 1986. Special report: an assessment of the accuracy of documenting waterfowl die-offs in a Texas coastal marsh. pages 88-95 in J. S. Feierabend and A. B. Russell (eds). Lead poisoning in wild waterfowl- a workshop. Proc. of the symposium held 3-4 March 1984, Wichita, KS. Szymczak, M. R. 1986. Characteristics of duck populations in the intermountain parks of Colorado. Colorado Div. Wild1., Tech. Publ. 35. 88pp. �45 _____ , and J. F. Corey. 1976. Construction and use of the Salt Plains duck trap in Colorado. Colorado Div. Wildl., Div. Rep. 6. l3pp. Whyte, R. J., G. A. Baldasssarre, and E. G. Bolen. 1986. Winter condition of mallards of the southern high plains of Texas. J. Wildl. Manage. 50:52-57. _____ , and E. G. Bolen. 1984. Impact of winter stress on mallard body composition. Condor 86:477-482. _____ , and . 1988. Flight ranges and lipid dynamics of mallards wintering on the southern high plains of Texas. J. Field Ornithol. 59:143-148. Wobeser, G. A. 1981. Diseases of wild waterfowl. Plenum Press, New York, NY. 300pp. Zwank, P. J., V. L. Wright, P. M. Shealy, and J. D. Newsom. 1985. Lead toxicosis in waterfowl on two major wintering areas in Louisiana. Wild1. Soc. Bull. 13:17-26. �46 CHAPTER 3 SURVIVAL OF ADULT FEMALE MALLARDS INSTRUMENTED IN THE SAN LUIS VALLEY, COLORADO Declines in the continental mallard (Anas platyrhynchos) population have greatly concerned waterfowl managers in re~ent years. Independent indices indicate a genuine decline in mallard numbers and an uncoupling of the relation between May pond conditions and mallard population size (Johnson and Shaffer 1987). Because survival estimates using band recovery data do not partition mortality into different periods, researchers have often used radiotelemetry to estimate waterfowl survival rates for different periods (e.g., Ringelman and Longcore 1983; Kirby and Cowardin 1986; Bowman and Longcore 1989; Conroy et al. 1989). Telemetry allows a cohort of instrumented birds to be intensively monitored and fates of most birds to be determined. My objectives were to determine January through April survival rates of adult female mallards in relatively "good" and "poor" body condition using radio telemetry. STUDY AREA This study was conducted in the San Luis Valley (SLV) , Colorado. The SLV is a l2,960-km2 intermountain basin in south-central Colorado, bounded by the San Juan Mountains to the west and the Sangre de Cristo Mountains to the east. Soils within the SLV are primarily fine sandy clay with little humus (Enright 1971). As a result of the surrounding mountains and a valley elevation of 2,286 to 2,438 m, the SLV has an arid climate characterized by sort, cool summers and cold dry winters (Lantis 1942). The growing season is 100 to 120 days beginning in mid May, but frost may occur any day of the year (Ramaley 1942). The cool, arid climate and sandy soils have resulted in a 'cold desert' plant association (Oosting 1956). Greasewood (Sarcobatus vermiculatus) and rabbit brush (Ch£Ysothamnus spp.) are the dominant vegetative species on the SLV floor (Enright 1971, Szymczak 1986). In soils that are extremely alkaline, saltgrass (Distichlis stricta) is the only vegetative cover. Sedges (Carex spp.), rushes (Juncus balticus), spikerush (Eleocharis spp.), and saltgrass are common plants along wetland edges (Enright 1971, Szymczak 1986). Cattail (Typha latifolia) and bulrush (Scirpus validus and~. paludosis) are the dominant emergent species (Enright 1971). Although precipitation within the SLV is low, spring run-off from the surrounding mountains provides ample water for irrigation (Hopper et al. 1975). Streams entering the northern part of the SLV percolate under the ground and continue flowing toward the center of the SLV under a layer of clay. This water emerges in central portions of the SLV as artesian wells (Szymczak 1986). Recently, irrigation techniques have changed from flood irrigation to center pivot sprinklers. Water removal by center pivot �47 sprinklers and reduced surface flows have lower the water table and reduced the number of artesian wells recently (Szymczak 1986). The human population in the SLV is low (less than 38,000) and has declined throughout this century; prospects for growth are considered to be low (Mason 1974). Most of the population is concentrated in towns along the Rio Grande and Conejos Rivers (Mason 1974). The local economy is agriculturally based (Mason 1974), with the major crops being irrigated crops, primarily potatoes, barley, vegetables, and hay (Hopper et al. 1975). Livestock production occurs throughout the SLV (Mason 1974). From 1964 through 1980 breeding pair estimates averaged 25,371 pairs, but have declined since primarily because of decreasing habitat (Szymczak 1986). Mallards are the most abundant nester (Szymczak 1986). An average of 18,734 ducks wintered in the SLV (1984-90 average, Colo. Div. Wildl., unpubl. data), with most being mallards wintering on the Monte Vista National Wildlife Refuge (MVNWR). Wintering mallards rest on open water maintained by continuous pumping or artesian flow and forage in nearby grain fields. METHODS We captured mallards wintering on MVNWR with Salt Plains bait.traps (Szymczak and Corey 1976), baited with barley, or cannon-projected nets (Dill and Thornberry 1950), from 11-18 January and 20-22 February 1987 and 11 January through 4 March 1988. Data on age, sex, weight, and wing length were obtained, and a condition index (estimated body fat / estimated fat-free body mass; Ringe1man and Szymczak 1985) calculated for each (See Chapter 1 for Additional Data). We summarized the condition indices of adult females captured the first trapping day of each trapping period, then assigned adult females to 1 of 3 condition classes based upon their condition index relative to the population distribution of the females captured. Subsequently, birds with a condition index in the upper 33% of the range were classified as in "good" condition, the middle 33% were considered "average," and the lower 33% were considered in "poor" condition. In January 1987, 52 adult females in good condition and 51 in poor condition were instrumented with 26 g, back-mounted transmitter packages attached with a harness similar to that described by Dwyer (1972). In February 1987, 18 adult females in good condition and 17 in poor condition were instrumented. In January 1988, 52 adult females in good condition and 51 in poor condition were instrumented. High mortality of instrumented females in 1987 prompted us to anticipate high mortality again in 1988. Initially, we planned to trap weekly and reuse radios collected from dead females by re-applying radios to adult females based upon their condition index relative to the condition index classes used during the 11-18 January trapping period. This protocol resulted in transmitters being re-applied to 23 adult females, 6 in good condition, 1 in average condition, and 16 in poor condition. However, by early February extremely high mortality, difficulties with trapping, and changes in the condition indices of the population made it impractical to reuse radios weekly. Consequently, all radios subsequently collected were reapplied during trapping on 29 February through 4 March. Procedures for determining the relative condition of each bird compared to the population were similar to those used in early January. We instrumented 22 females in good condition, 17 in average, and 24 in poor condition. Standard techniques were used to obtain weekly locations of instrumented ducks from the ground (Cochran 1980:517-518) using vehicle-mounted and hand- �48 1 Good • CO > .> ':::J en Poor 0.8 --{)--. 1 1 0.6 CD > +-' ~ :::J .. 0.4 E :::J o 0.2 a Week Figure 3.1. Cumulative weekly survival rates and standard errors of adult females instrumented in January 1987. �49 held antenna-receiver systems. Birds not found early in the week were located from the air (Gilmer et al. 1981) using a strut-mounted antenna attached to a Cessna 182. All birds located from the air were ground-checked the next day to determine if they were alive or dead, because the radio transmitters did not have mortality sensors. Any birds contacted from the ground that were in unusual locations, did not have attenuating signals, or were in the same location for 2 consecutive weeks were checked visually to determine whether they were alive or dead. The date, estimated date of death, likely cause of death, site description, Universal Transverse Mercator coordinates, and other pertinent comments were recorded for each dead instrumented bird. Carcass remains found with the radio were frozen for further analyses. In 1988, wings from carcasses of instrumented females were analyzed for ulna lipids to determine if the females died from malnutrition, as indicated by ulna lipids of 4-5% (Appendix A). Logistic regression (Hosmer and Lemeshow 1989) was used to determine if the probability of an instrumented females succumbing to malnutrition was related to body condition at instrumentation in 1988. Survival rates of females instrumented in January through 1 April 1987 and January through 26 March 1988 were estimated using the Kaplan-Meier method (Lee 1980, White and Garrott 1990). When an individual cannot be located and classified as alive or dead, the animal is classified as censored. Censored individuals were delete from further analyses. Survival functions were estimated by plotting Kaplan-Meier survival estimates. Comparisons of weekly survival between condition classes were made using 2 x 2 Chi-square tests. All Chi-square tests are one-sided tests assuming survival of birds marked in good condition was greater than survival of poor condition birds. Comparisons between condition class survival of females instrumented in January 1987 and January 1988 were made using logistic regression (Hosmer and Lemeshow 1989). Survival functions of females instrumented in January 1987 and January 1988 were compared using log rank tests (White and Garrott 1990). Because nearly all the mallards wintering in the SLV roost on the MVNWR, counts of mallards on aerial photographs taken on 28 January and 4 February 1988, were used to estimate the survival rate for the period between the 2 photgraphic surveys. This estimate was then compared with the corresponding survival estimate for all instrumented females using a z test to determine if survival rates of instrumented birds differed from the population. RESULTS In 1987, weekly survival rates identified better survival by good condition birds between weeks 3-4 and 7-8 (X2 - 3.2, 1 df, f - 0.0369 and X2 3.7, 1 df, f - 0.0272 for weeks 3-4 and 7-8, respectively). Between weeks 5-6 and 9-10, greater survival rates of good condition birds approached significance (X2 - 1.8, 1 df, f - 0.0890 and X2 - 1.9, 1 df, f - 0.0834 for weeks 5-6 and 9-10, respectively). Logistic regression also identified a relation between condition indices and status (f - 0.0025) and the log rank test identified different survival functions, with better survival for good condition birds (X2 - 4.1, 1 df, f - 0.0422). Kaplan-Meier survival estimates also suggest a relation between condition and survival (Fig. 3.1). In 1988, survival of good condition females was greater between weeks 56 (X2 - 3.3, 1 df, f - 0.0338) and approached significance between weeks 0-1 and 6-7 (X2 - 2.4, 1 df, f - 0.0626 and X2 - 1.9, 1 df, f ~ 0.0838 for weeks 0-1 and 6-7, respectively). Logistic regression (f - 0.2143) and the log rank �50 test (Xl - 1.4, 1 df, f - 0.2432) failed to detect a difference in survival rates or survival functions between the two groups. Kaplan-Meier survival estimates assuming censored birds lived or died does not suggest any relations between condition and survival (Fig. 3.2). Because the two groups did not show differences in survival rates, the samples were pooled for further analyses. The 1 week population survival rate estimated from the aerial photographs, 0.885, did not differ from the pooled estimate for the instrumented females during that week (0.841 ± 0.043, z - 1.023, f - 0.3063) in 1988. This suggests that the instrumented sample was representative of the population during this week. Twelve of 52 wings (23.1%) collected from instrumented females that died in 1988 had exhausted (0-4%) lipid reserves in the ulna (Fig. 3.3). Seven more had reduced (4-8%) lipid reserves in the ulna (Fig. 3.3). Most birds (14 of 18, 78%) with reduced lipid reserves were collected within 3 weeks of instrumentation (Fig. 3.4). Similar percentages of females instrumented in good and poor condition succumbed to malnutrition (Xl - 0.14, 1 df, f 0.7085). Equal numbers of birds instrumented in good and poor condition (6 for each class) were recovered with ulna lipids in the 0-4% category. All birds collected with reduced ulna lipid reserves, 4-8%, were instrumented in poor condition. Recovery patterns for each condition-ulna lipid class were similar (Fig. 3.5) except for the lack of birds instrumented in good condition with 4-8% ulna lipids. Percent of instrumented females dying with exhausted ulna lipids (0-4%) did not differ (Xl - 1.6, 1 df, f - 0.21) from the percent of banded birds found dead with exhausted ulna lipids, (29 of 201, 14.4%, Chapter 2), but the power of this comparison was low (lambda - 2.28, 1 df, B - 0.33, a - 0.05). DISCUSSION Survival was related to condition in 1987 but not in 1988, even though data from instrumented females collected in 1988 suggested that the transmitter package affected weight dynamics of most females (Chapter 4). Analyses of birds banded but not instrumented also revealed a relation between survival and condition in 1987, but not in 1988 (Chapter 2). Carcass collections (Chapter 2) and survival rate estimates of adult females instrumented in the SLV suggest high winter mortality from January through March. Analysis of band recoveries of adult females banded preseason in the SLV produced an survival rate estimate of 0.55 (95% C. I. 0.50 - 0.60; Szymczak 1986:46), substantially higher than 70-day winter survival estimates for our samples of instrumented females (0.3 to 0.4; Figs. 3.1 and 3.2). Both years, survival the rest of the year would have to approach 1.00 to produce an annual survival rate similar to that of females banded in late summer. The preseason mallard population in the SLV may be substantially different from the wintering population, but the distribution of band recoveries suggests many mallards are residents in the SLV. Likewise, these 70-day survival rates are less than mean annual estimates presented in Nichols and Hines (1987:126-151) and Hyland and Gabig (1980) for mallards banded postseason, although annual and geographic variation in survival rate estimates can be high. Weekly mortality was not consistently greater for females instrumented in poor condition. A couple of factors may explain these results. Weekly survival estimates are a function of the number of birds alive at the start of the week. If more good condition birds are at risk, such as in 1987, it is �51 Good ~ 0.8 ._ • Poor \> ,,_ - - B-- :::l (f) CD > ._ ......., -:::l 0.4 (\j --_ E :::l o 0.2 Week Figure 3.2. Cumulative weekly survival rates and standard errors of adult females instrumented in January 1988. �52 15 ~ Q) +-' o Q) 0 o 10 (/J 0), c: .~ "t- O 5 ~ CD .0 E :::l Z 0 0-4 4-8 8-1212_1616-202Q_~424:-2828_3232-3636_4040-44 Percent Llpld in the Ulna Figure 3.3. Distribution of percent lipids in the ulna of 50 wings recovered from instrumented adult female mallards that died in 1988. Ulnas with less than 4-5% lipids are associated with exhaustion of body lipid reserves (starvation). �53 5 "'0 CD +-' as o o •• 0-4 % Ulna Lipid LJ 4-8 % Ulna Lipid 4 en 3 0') ._C ~ *-2 o \.,. CD .0 E 1 ~ z o .• 2. 3 4 5 6 7 week After Instrumentation Figure 3.4. « Chronology of recovery of instrumented birds with low 8%) ulna lipid levels. Figure includes 2 females instrumented in average condition in March, 1988 that were recovered 2 and 3 weeks after instrumentation. �54 5 "'0 "Gooo" Condition 0-4 % Ulna Lipids CD ......, (.) 4 CD 1771 "Good" Condition o o ~ ~ ~ 3 C/) D 0') C ._ 4-8 % Ulna Lipids IIPOOr Condition 0-4 % Ulna Lipids IIPOOrl Condition 4-8 % Ulna Li ids l ~ ,+-2 o ~ CD .c E1 :J Z o 1 2 345 6 7 8 Week of Collection Figure 3.5. Recovery week of wings co1lected with < 8 percent ulna lipids and condition class of the female at instrumentation. �55 possible that more good condition birds may die. Also, if the transmitter package affected survival, poor condition birds may not die at greater rates than good condition females. Although avian cholera was an important cause of mortality of unmarked birds, we saw no instrumented females displaying cholera symptoms. All carcasses recovered that had not been scavenged exhibited a degree of malnutrition (mass loss) or effects of the transmitter package (e.g., the bill being caught in the harness). In both years, most deaths of instrumented females occurred within 4 weeks of instrumentation (Figs. 3.6 and 3.7). Less aerial search time in 1987 resulted in the recovery of some transmitters several weeks after the bird had died. Because most birds that died from malnutrition in 1988 died within 4 weeks of instrumentation, I believe that birds found dead within that period in both years probably died of transmitter-package related reasons (Chapter 4). Although direct causes of death usually could not be identified, the high percent of deaths attributable to malnutrition suggests that many birds may have succumbed to proximate causes of death (e.g., predation), but the ultimate cause of death was associated with the transmitter package. My results suggest that body mass losses identified for instrumented birds (Chapter 4) affected survival of our samples. Although the transmitter package may have affected results, the relation between survival and condition in 1987 but not in 1988 suggests that mallards in better condition in January do have greater survival under certain environmental conditions. Generally, January and February 1987 had a lower prevalence of avian cholera, greater levels of avian predation, milder temperatures and greater snowfall than January and February 1988. The greater snowfall made food unavailable and many instrumented and unmarked mallards were found which had starved. Although no instrumented birds were observed with symptoms of avian cholera, the greater prevalence of avian cholera in 1988 probably did result in some instrumented birds succumbing to avian cholera. Because susceptibility to avian cholera is not related to body condition (Wobeser 1981, McLandress 1983), relations between body condition and other mortality factors may be masked by more females instrumented in good condition dying from avian cholera. Relations between winter body condition and survival need to be examined as the relations among body condition and specific causes of mortality. My results suggest survival is related to winter body condition when starvation and predation are important mortality factors, but not when avian cholera is the prevalent mortality cause. LITERATURE CITED Bowman, T. D., and J. R. Longcore. 1989. Survival and movements of molting male black ducks in Labrador. J. Wildl. Manage. 53:10571061. Cochran, W. W. 1980. Wildlife telemetry. pp. 507-520 in S. P. Schemnitz (ed.). Wildlife management techniques. Fourth ed. Wildlife Soc. Inc., Washington, D. C. Conroy, M. J., G. R. Costanzo, and D. B. Stotts. 1989. Winter survival of female American black ducks on the Atlantic Coast. J. Wildl. Manage. 53:99-109. �56 Dill, H. H., and W. H. Thornberry. 1950. A cannon-projected net trap for capturing waterfowl. J. Wildl. Manage. 14:132-137. Dwyer, T. J. 1972. An adjustable radio package for ducks. Bird-Banding 43:282-284. Enright, C. A. 1971. An analysis of mallard nesting habitat on the Monte Vista National Wildlife Refuge, Colorado. M.S. Thesis, Colorado State University, Fort Collins, co. l13pp. Gilmer, D. S., L. M. Cowardin, R. L. Duval, C. W. Schaiffer, and V. B. Kuech1e. 1981. Procedures for the use of aircraft in wildlife biotelemetry studies. U.S. Fish and Wildl. Serv., Resour. Publ. 140. 19pp. Hopper, R. M., A. D. Geis, J. R. Grieb, and L. Nelson, Jr. 1975. Experimental duck hunting seasons, San Luis Valley, Colorado, 1963-1970. Wildl. Monog. 46:1-68. Hosmer, D. W., and S. Lemeshow. 1989. Applied logistic regression. John Wiley & Sons, New York. 307pp. Hyland, J. M., and P. J. Gabig. 1980. Survival and recovery distribution of central and western Mississippi flyway winter-banded mallards. Nebraska Game Parks Comm., Tech. Ser. 6. l32pp. Johnson, D. H., and T. L. Shaffer. 1987. Are mallards declining in North America? Wildl. Soc. Bull. 15:340-345. Kirby, R. E., and L. M. Cowardin. 1986. Spring and summer survival of female mallards from northcentral Minnesota. J. Wildl. Manage. 50:38-43. Krapu, G. L. 1981. The role of nutrient reserves in mallard reproduction. Auk 98:29-38. Lantis, D. W. 1942. Sage of the San Luis Valley in Colorado. Folks and Fortunes 1:38-40. Lee, E. T. 1980. Statistical methods for survival data analysis. Lifetime Learning Publ., Belmont, CA. 557pp. Mason, J. 1974. San Luis Valley resource area -- social economic profile. Western Interstate Commission for Higher Education, Boulder, CO. 2l9pp. McLandress, M. R. 1983. Sex, age, and species differences in disease mortality of Ross' and lesser snow geese in California: implications for avian cholera research. Calif. Fish and Game 69:196-206. �57 Nichols, J. D., and J. E. Hines. 1987. Population ecology of the mallard. VIII. Winter distribution patterns and survival rates of winter-banded mallards. U.S. Fish Wildl. Serv., Resour. Publ. 162. l54pp. Oosting, H. J. 1956. The study of plant communities. W. H. Freeman and Co., San Francisco, CA. 440pp. Ramaley, F. 1942. Vegetation of the San Luis Valley in southern Colorado. Univ. Colo. Studies. Series D. Physical and Biological Sciences. 1:237-279. Ringelman, J. K., and J. R. Longcore. 1983. Survival of female black ducks, Anas rubripes, during the breeding season. Can. Field-Nat. 97:62-65. _____ , and M. R. Szymczak. 1985. A physiological condition index for wintering mallards. J. Wildl. Manage. 49:564-568. Szymczak, M. R. 1986. Characteristics of duck populations in the intermountain parks of Colorado. Colo. Div. Wildl. Tech. Publ. 35. 88pp. _____ , and J. F. Corey. 1976. Construction and use of the Salt Plains duck trap in Colorado. Colorado Div. Wildl., Div. Rep. 6. 13pp. White, G. C., and R. A. Garrott. 1990. Analysis of wildlife radiotracking data. Academic press, New York, NY. 383pp. Wobeser, G. A. 1981. Diseases of wild waterfowl. Plenum Press, New York, NY. 300pp. �58 10 _ "GOOD" CONDITION h----J "POOR" CONDITION 9 8 (J) U •.... Co 7 6 U m Q) 0 a •.... 5 4 '.. Q) .!J E :J Z 3 ~. 2 .,.. ." " 1 0 1 2 3 4 5 6 7 8 9 10 ·11 12 Weeks After Instrumentation Figure 3.6. Week of recovery of females instrumented in 1987. I considered the week after the last radio was applied during each banding session as the first week. �59 30 25 (/) 0 a: CO 20 0 « w 0 u, 15 0 a: w CO 10 ~ ::J Z 5 o 1 2 3 4 5 6 7 8 9 Weeks After Instrumentation Figure 3.7. Week of recovery of females instrumented in 1988. I considered the week after the last radio was applied during each banding session as the first week. Because we detected no survival differences related to condition, condition classes are combined. �60 CHAPTER 4 EFFECTS OF INSTRUMENTATION ON BODY MASS OF ADULT FEMALE MALLARDS WINTERING IN THE SAN LUIS VALLEY, COLORADO Biotelemetry can provide insights often unattainable using other methods. Numerous studies of waterfowl have used radio telemetry to determine habitat use (e. g. Gilmer et al. 1977, Ringelman et al. 1982, Costanzo et al. 1983, Jorde et al. 1984, Kirby et al. 1985, Thompson and Baldassarre 1988, Morton et al. 1989); survival rates (e. g. Ringelman and Longcore 1983, Kirby and Cowardin 1986); and field-feeding behavior (Jorde et al. 1983). Ideally, behavior of instrumented animals should not be altered by a transmitter or its harness. Yet only a few studies have evaluated potential effects of radio-markers on birds. Greenwood and Sargeant (1973) found that captive mallards (Anas platyrhynchos) and blue-winged teal (~ discors) equipped with back-mounted radiopacks lost more mass and showed skin irritation and behavioral differences when compared with control groups. Perry (1981) noted similar physical and behavioral effects, but no mass loss in canvasbacks (Aythya valisineria). Conversely, Gilmer et al. (1974) was unable to detect effects of breast-mounted transmitters on breeding and social behavior of wild mallards and wood ducks (Aix sponsa). Similarly, Cowardin et al. (1985) and Ringelman et al. (1982) were unable to detect any effects on reproduction attributable to back-mounted transmitters. As part of a study on the relations among winter body condition, survival and reproduction, I applied radios to adult female mallards (Anas platyrhynchos) in January 1987, then noted high mortality of instrumented birds. To investigate sources of this mortality, adult female mallards were again instrumented in 1988 and a sample of these females collected 2-5 weeks after instrumentation. I hypothesized that the high mortality was related to depletion of body reserves caused by behavioral changes associated with instrumentation. The objective of these collections were to determine changes in body mass after instrumentation. STUDY AREA This study was conducted in the San Luis Valley (SLV), in south-central Colorado. The SLV is a 12,960 km2 intermountain basin bounded by the San Juan Mountains to the west and the Sangre de Cristo Mountains to the east. As a result of the surrounding mountains and an elevation of 2,286 to 2,438 m, January temperatures average -4 C, and snow and ice fog are common. An average of 14,700 mallards winter in the SLV (1982-84 average, Colo. Div. Wildl., unpubl. data), with most residing on the Monte Vista National Wildlife Refuge (MVNWR). Wintering mallards rest on open-water areas maintained by continuous pumping or artesian flow, and forage in nearby grain fields. Temperatures below the lower critical temperature for mallards (Chapter 1) and snow cover that reduces food availability make the SLV an �61 energetically stressful environment for wintering mallards. METHODS Instrumentation We captured mallards wintering on MVNWR with Salt Plains bait traps (Szymczak and Corey 1976) or cannon-projected nets (Dill and Thornberry 1950) from 11 January through 4 March 1988. Data on age, sex, weight, and wing length were obtained, and a condition index (Ringelman and Szymczak 1985) calculated for each bird. We summarized the condition indices of adult females captured the first trapping day of each trapping period, then assigned adult females to 1 of 3 condition classes based upon their condition index relative to the population distribution of the adult females captured. Subsequently, birds with a condition index in the upper 33% of the range were classified as in "good" condition, the middle 33% were considered "average," and the lower 33% were considered in "poor" condition. In January 1988, 52 adult females in good condition and 51 in poor condition were instrumented with 26 g back-mounted transmitter packages attached with a harness similar to that described by Dwyer (1972). From 29 February through 4 March, 22 females in good condition, 17 in average, and 24 in poor condition were instrumented. From 18 January through 4 March 1988, mallards were captured periodically and processed similarly to the 11-18 January trapping period. Some mallards captured during earlier trapping efforts were recaptured and differences in mass and condition were calculated and compared with condition indices of the collected samples. Collection In 1988, 10 instrumented females (5 in good and 5 in poor condition at instrumentation) were collected 2-3 and 4-5 weeks after the 11-18 January trapping period, and 2-3 weeks after the March trapping effort. Females were collected opportunistically, because time limitations prevented randomly selecting a particular female to be collected. Additional instrumented females were captured during other field activities during the collection periods and data on body mass recorded and included in analyses of changes in condition. Birds were collected by hand-held dip nets, cannon nets, shotguns or rifles. Collected birds were immediately placed in plastic bags. As soon as possible, body mass, wing length, condition index, and any pertinent comments were recorded before the carcasses were frozen. Carcasses were later thawed and necropsied. Continuous Monitoring In 1989, 5 male and 5 female mallards were instrumented 13 December 1988. Beginning 20 December 1988, 1 male and 1 female per week were continuously monitored from 45 min before sunrise until the bird settled on the roost after sunset. Time of movement initiation and termination were recorded for the instrumented bird relative to important events involving unmarked birds (e.g., movements of large numbers of birds to and away from the site the instrumented bird was using, predator activity, and initiation of field-feeding flights). �62 RESULTS I collected 8 instrumented females in the good and 7 in the poor condition classes during the first collection period, and 5 in each condition class during the second period. Three good, 4 average, and 3 poor condition females instrumented in March were collected. Mean body mass declined in most birds collected 2-3 weeks after instrumentation, but birds appeared to have regained lost mass by weeks 4-5 (Fig. 4.1). During the first collection period, females were collected an average of 15.0 (±2.2) days after instrumentation and mean mass change was -238 g (range 59 to -410 g). Body mass change resulted in an average decline in condition of 15.4 (range 4.4 to -27.8). Body mass and condition of collected females did not differ by condition class at the time of collection (Tables 4.1 and 4.2). In the first collection period, unmarked adult females weighed more than the collected birds (.t - -6.45, 51 df, £ - 0.0001) and were in better condition (.t - -6.51, 51 df, £ - 0.0001). Table 4.1. Mean (SE) body mass (g) and condition index at the time of collection, and changes since initial capture for instrumented female mallards collected 2-3 and 4-5 weeks after instrumentation in 1988 in the San Luis Valley, Colorado. Collection period Januar:£ 2-3 weeks 4-5 weeks Initial condition class Body mass Condition index mass Change in condition Good Poor 752(117) 2.3(8.0) 720(115) 1.8(6.8) -298(86) -18.9(6.8) -170(106) -11.5(7.2) Good Poor 920(63) 13.1(3.3) 799(30) 6.1(1.5) -151(27) -112(23) - 8.3(2.1) - 7.4(1.3) 805(31) 784(29) 817(59) -186(35) -152(30) - 54(31) -12.0(2.3) -10.2(2.6) - 3.2(1.7) Februa:O:-March 2-3 weeks Good Average Poor 7.0(2.3) 5.4(3.3) 7.6(4.0) �63 1,200 1,150 Good 1,100 Average .... ~ .... 0 (1 1,050 Poor - - 8-- en1,000 en Unmarked • ctl 950 ~ 900 ~ "'C a co 850 800 750 700 650 600~~~~'_~~~~~~~~~~~'_'_~~~~ Jan 1 Jan 21 March 1 March 21 Date Figure 4.1. Mean body mass and standard error for females collected 2-3 and 4-5 weeks after instrumentation. Also given are mean body masses at capture for the collected birds and unmarked adult feamles captured during similar time intervals. near the mean. Sample sizes are shown in parenthesis �64 Table 4.2. Comparisons of body masses and condition indices at collection and mass and condition index changes for adult female mallards instrumented in good and poor condition and collected 2-5 weeks after instrumentation in January 1988. Collection period Comparison ,t value df f 2-3 weeks Body mass Condition Mass change Condition change 0.53 0.11 -2.05 -l. 61 13 13 13 13 0.6069 0.9154 0.0611 0.1323 4-5 weeks Body mass Condition Mass change Condition change 3.81 4.31 -2.16 -0.74 8 8 8 8 0.0052 0.0026 0.0632 0.4799 In the second period, females collected 28.3 (±2.6) days after instrumentation lost an average of 132 g (range -71 to -177 g) and declined an average of 8.3 (range -4.4 to -10.1) in condition. Mass and condition declines were similar between the two groups (Tables 4.1 and 4.2). Four unmarked females were captured during the second period. Neither their mean mass of 805 g nor their condition was different from the mean body mass or condition of the collected birds (,t 0.96, 12 df, f - 0.3574 and,t - 0.78, 12 df, f - 0.4491, respectively). Females instrumented in March were collected 18.1 (±4.5) days after instrumentation, lost an average of 133 g (range -32 to -209 g), and condition indices declined an average of 8.6 (range -1.4 to -14.4). Body mass and condition changes differed between good and poor condition females, and between average and poor condition females, but not between good and average condition females (Tables 4.1 and 4.3). �65 Table 4.3. Comparisons of body masses and condition indices at collection and mass and condition index changes for adult female mallards instrumented in good, average and poor condition 29 February-4 March 1988 and collected 2-3 weeks after instrumentation. Condition classes compared Comparison .t value df f Good and Poor Body mass Condition index Mass change Condition change -0.31 -0.24 -4.87 -4.35 4 4 4 4 0.7702 0.8230 0.0082 0.0122 Good and Average Body mass Condition index Mass change Condition change -0.91 -0.73 1.38 0.81 5 5 5 5 0.4028 0.5001 0.2257 0.4541 Average and Poor Body mass Condition index Mass change Condition change -0.98 -0.83 -4.18 -3.47 5 5 5 5 0.3708 0.4455 0.0087 0.0179 High mortality rates among mallards instrumented in December 1988 only allowed for 4 continuous monitoring sessions. Most deaths occurred 2-3 weeks after instrumentation. Instrumented birds were among the last birds to leave roosts on field-feeding flights in the evening, and among the first to return to roost areas. Consequently, instrumented birds were away from the roost for only 48.5 (± 15.2) min whereas the entire flock field-feeding flight lasted 133.3 (±40.0) min. We recorded no instances of large groups of mallards returning to roost areas after having initiating field-feeding and then leaving again for another flight. DISCUSSION Without a larger sample of females that were banded concurrently during the time radios were applied and then recaptured when instrumented birds were collected, it is impossible to determine if measured mass loss indicates: (1) a handling effect from the banding process; (2) a response to the radio package; or (3) a general reduction in mass of adult females during this period. Comparative mass losses of recaptured banded birds suggest that handling effects were not alone responsible for the mass losses observed among instrumented females (Chapter 1). Also, we cannot determine the relation of observed mass loss to high mortality rates of instrumented females (Chapter 3), because other winter survival estimates (Chapter 2) suggest high winter mortality within this mallard population. Effects of transmitter packages on bird behavior and survival range from little to no effect (e.g., Gilmer et al. 1974, Herzog 1979, Hines and Zwickel 1985, Wanless et al. 1988, Sedinger et al. 1990) to reduced survival (Warner and Etter 1983, Marks and Marks 1987), and include body mass loss and behavioral effects (Greenwood and Sargeant 1973), abnormal behavior (Perry 1981), and increased energy expenditure during flight (Gessaman and Nagy �66 1988). This wide range of reported avian responses to transmitter packages may occur because: (1) sampling has been insufficient to detect effects; (2) instrumentation may affect certain species more than others; or (3) species may be more sensitive to instrumentation at certain times of the year. Although instrumented mallards participated in field-feeding flights in 1989, foraging time was less than the time required to meet estimated standard metabolic rate needs (Chapter 1). Because standard metabolic rate does not include energy required for activity, instrumented females were in a negative energy balance that resulted in the observed mass loss. Morton et al. (1989) found that an unexpected number of American black duck (Anas rubripes) females instrumented in the winter remained on their study area after they should have migrated. They could not rule out the possibility of the radio package ~ffecting behavior, but suggested that because flight behavior seemed normal, it was more likely that resident birds had been instrumented. These birds may also have been affected in ways that could not ·be detected simply by evaluating flight behavior. Based only on this criterion, we would also have concluded that radio packages had no effect because instrumented mallards participated in field-feeding bouts. Sedinger et al. (1990) found no increase in energy expenditure of Pacific black brant (Branta bernicla ni~ricans) affixed with radio transmitters housed in cages that allowed little activity, whereas Gessaman and Nagy (1988) found homing pigeons (Columbia livia) flying with transmitter packages increased CO2 production from 41 to 52%. Results from these studies suggests transmitter packages affect flight, but not maintenance energetics. Transmitter packages probably increase mallard energy expenditure during flight in a way similar to pigeons. SLV mallards already have low body masses (Chapter 1), and may be reluctant to increase energy expenditure in response to radio transmitters. Most studies have not been designed to identify subtle behavioral effects that may influence results, or tracking intensity has not been sufficient to determine if mortality and migration (or lack of normal migration) occurs normally. One approach (i.e., Derleth and Sepik 1990) is to exclude birds obviously affected by the transmitter packages (i.e. birds which die soon after instrumentation) but assume surviving birds behaved normally, which may not be the case. Experimental design should allow for evaluation of possible adverse effects of the transmitter package itself. Results from my study suggest that application of transmitters during energetically stressful times may alter body mass dynamics of the instrumented sample of birds. �67 LITERATURE CITED Costanzo, G. R., T. T. Fendley, and J. R. Sweeney. 1983. Winter movements and habitat use by wood ducks in South Carolina. Proc. Annu. Conf. Southeast. Assoc. Fish and Wildl. Agencies 37:67-78. Cowardin, L. M., D. S. Gilmer, and C. W. Shaiffer. 1985. Mallard recruitment in the agricultural environment of North Dakota. Wildl. Monog. 92. 37pp. Derleth, E. L., and G. F. Sepik. 1990. Summer-fall survival of American woodcock in Maine. J. Wildl. Manage. 54:97-106. Dill, H. H., and W. H. Thornberry. 1950. A cannon projected net trap for capturing waterfowl. J. Wildl. Manage. 14:132-137. Dwyer, T. J. 1972. An adjustable radio package for ducks. Bird-Banding 43:282-284. Gessaman, J. A., and K. A. Nagy. 1988. Transmitter loads affect the flight speed and metabolism of homing pigeons. Condor 90:662-668. _Gilmer, D. S., I. J. Ball, L. M. Cowardin, and J. H. Riechmann. 1974. Effects of radio packages on wild ducks. J. Wildl. Manage. 38:243252. _____ , R. E. Kirby, I. J. Ball, and J. H. Riechmann. 1977. Post-breeding activities of mallards and wood ducks in north-central Minnesota. J. Wildl. Manage. 41:345-359. Greenwood, R. J., and A. B. Sargeant. 1973. Influence of radio packs on captive mallards and blue-winged teal. J. Wildl. Manage. 37:3-9. Herzog, P. W. 1979. Effects of radio-marking on behavior, movements, and survival of spruce grouse. J. Wildl. Manage. 43:316-323. Hines, J. E., and F. C. Zwickel. 1985. Influence of radio packages on young blue grouse. J. Wildl. Manage. 49:1050-1054. Jorde, D. G., G. L. Krapu, and R. D. Crawford. 1983. Feeding ecology of mallards wintering in Nebraska. J. Wildl. Manage. 47:1044-1053. Jorde, D. G., G. L. Krapu, R. D. Crawford, and M. A. Hay. 1984. Effects of weather on habitat selection and behavior of mallards wintering in Nebraska. Condor 86:258-265. Kirby, R. E., J. H. Riechmann, and L. M. Cowardin. 1985. Home range and habitat use of forest-dwelling mallards in Minnesota. Wilson Bull. 97:215-219. �68 Kirby, R. E., and L. M. Cowardin. 1986. Spring and summer survival of female mallards from northcentral Minnesota. J. Wildl. Manage. 50:38-43. Marks, J. S., and V. S. Marks. 1987. Influence of radio collars on survival of sharp-tailed grouse. J. Wi1dl. Manage. 51:468-471. Morton, J. M., R. L. Kirkpatrick, M. R. Vaughan, and D. F. Stauffer. 1989. Habitat use and movements of American black ducks in winter. J. Wildl. Manage. 53:390-400. Perry, M. C. 1981. Abnormal behavior of canvasbacks equipped with radio transmitters: J. Wild1. Manage. 45:786-789. Ringe1man, J. K., J. R. Longcore, and R. B. Owen, Jr. 1982. Breeding habitat selection and home range of radio-marked black ducks (Anas rubripes) in Maine. Can. J. Zool. 60:241-248. _____ , and . 1983. Survival of female black ducks, Anas rubripes, during the breeding season. Can. Field-Nat. 97:62-65. _____ , and M. R. Szymczak. 1985. A physiological condition index for wintering mallards. J. Wild1. Manage. 49:564-568. Sedinger, J. S., R. G. White, and W. E. Hauer. 1990. Effects of carrying radio transmitters on energy expenditure of Pacific black brant. J. Wi1dl. Manage. 54:42-45. Szymczak, M. R., and J. F. Corey. 1976. Construction and use of the Salt Plains duck trap in Colorado. Colorado Div. Wild1., Div. Rep. 6. 13pp. Thompson, J. D., and G. A. Baldassarre. 1988. Postbreeding habitat preference of wood ducks in northern Alabama. J. Wildl. Manage. 52:80-85. Wanless, S., M. P. Harris, and J. A. Morris. 1988. The effect of radio transmitters on the behavior of common murres and razorbills during chick rearing. Condor 90:816-823. Warner, R. E., and S. L. Etter. 1983. Reproduction and survival of radio-marked hen ring-necked pheasants in Illinois. J. Wildl. Manage. 47:369-375. Prepared by: ~.{'i..!l::f>7F~ Wildlife Researcher C �69 APPENDIX A USE OF ULNA LIPIDS TO IDENTIFY STARVATION IN MALLARDS Condition indices are important tools for monitoring nutrient reserves in wild animals (Kirkpatrick 1980). Cheatum (1949) first described the use of bone marrow fat to indicate starvation in deer. Hutchinson and Owen (1984) examined lipid storage in medullary bone of American black ducks (Anas rubripes), common eiders (Somateria mo11issima), and brant (Branta bernic1a) and the possibility that medullary lipid levels could be used as a condition indicator. They reported that medullary lipids were mobilized disproportionately to other lipid depots, and were one of the last depots depleted before death. Their results suggest that medullary lipid levels may serve as an indicator of death resulting from malnutrition. The objectives of this study were to confirm that depletion of medullary lipid levels in the ulna of mallards (Anas p1atyrhynchos) is indicative of death resulting from malnutrition, and to develop a test that could identify individuals dying·of malnutrition without extracting medullary lipids. METHODS Twelve mallards (1 adult female, 3 adult males, 7 immature males, and 1 immature female) were trapped 18 March 1987. Three birds were euthanized immediately using CO2 in an airtight box. The remaining birds were held in captivity with water, but no food, for varying periods before being euthanized. Carcasses were plucked and frozen in plastic bags for storage. Later, carcasses were thawed and the bill, feet below the proximal end of the tarsus, and gut contents were removed. Before grinding, both ulnas were excised from the wings, cleaned of adhering tissue, and frozen in plastic bags for later lipid extraction. The partly refrozen carcasses were weighed and ground three times through the 9-mm plate of a commercial meat grinder. The grinder was thoroughly cleaned in hot water between each samp1e.After partial refreezing, the homogenate was passed three more times through a 3.5-mm grinding plate. During the final grinding, two subsamp1es of about 20 g were selected and placed into preweighed 25- x 100-mm cellulose extraction thimbles. Samples in thimbles were weighed (0.0001 g) and then dried in a convection oven at 100 C to a constant weight (48 h) to determine percent carcass water. After drying, fat was extracted for 6-8 h using petroleum ether in a Soxh1et extraction apparatus. Ulnas were cracked open, placed in preweighed extraction thimbles and processed similarly to the carcass homogenate samples. Percent fat in the ulna and percent total body fat, on a dry weight basis, were plotted. To reduce time and cost of exam~n~ng a large sample of wings collected from dead mallards, a simpler test was examined. A random sample of 50 ulnas were cracked open, then the marrow touched to a clean piece of blotter paper. The resulting spot was allowed to dry and then was examined for residual fat. �70 The resulting blots indicated if a bird had starved. The cracked ulnas were placed in preweighed extraction thimbles; lipids extracted in a manner similar to the extraction procedure for the ulnas from the complete carcasses. Three observers repeated classification of the resulting blots three times. Classifications were then compared with lipid extraction results to determine classification error rate. RESULTS Body masses ranged from 565 to 1022 g, and percent body fat from 0.7 to 26.6% dry weight. Relatively constant ulna lipid levels were maintained to a critical range, around 5% total body fat, after which ulna lipid levels declined (Fig. A.l). To simplify examination of a larger sample, 2 of the 50 ulnas used in the blotting paper spot tests came from birds that had very low lipid levels in the ulna and were classified as having succumbed to depletion of body lipids. No classification errors were made in any trial; birds classified as starved had depleted body lipids, and those classified as not starved did not deplete reserves. DISCUSSION Determining whether a mallard had died with depleted lipid reserves allows mortality to be partitioned into causes that affect physical condition and those that do not affect condition. However, this technique does not allow the ultimate cause of mortality to be determined. Several mortality causes, such as malnutrition and plumbism, result in depletion of lipid reserves. The technique of cracking an ulna and touching it to blotter paper provides a quick and simple test to determine if a bird has exhausted its lipid reserves. Using this technique to determine percentage of birds dying from starvation, we were able to sample more than 1,000 wings in less than a week. Although ulna lipids have been shown to be related to total body lipids in other waterfowl species, we suggest that before sampling wings to determine starvation percentages in other species, these relations be verified for the species in question. Further work may be able to combine other measures, such as lead levels in the humerus or staining for Pasturella mu1tocida in the radius, to allow ultimate causes of mortality to be identified. By identifying ultimate mortality causes, managers will be able to manipulate conditions that are indirect causes. For example, although identification of predation as a proximate cause of death may be easy, individuals vulnerable to predators may have died because some ultimate factor altered their physiology, behavior, or habitat such that their survival probability is decreased. �71 30 * 25 - * ~ C"U * LL C"U 20 c: :J ~ * :I< - * ;5 - c: CD o ~ ;0 - CD (L * 5 0 0 " ;0 5 ;5 20 25 Percent Carcass Fat Figure A.I. Relation between total body lipids and lipids extracted from the ulna of 12 mallards. 30 �72 LITERATURE CITED Cheatum, E. L. 1949. Bone marrow as an index of malnutrition in deer. New York State Conserv. 3:19-22. Hutchinson, A. E., and R. B. Owen. 1984. Bone marrow fat in waterfowl. J. Wildl. Manage. 48:585-591. Kirkpatrick, R. L. 1980. Physiological indices in wildlife management. pp 99-112 IN S. D. Schemnitz (ed.). Wildlife management techniques manual, 4th ed. The Wildlife Soc., Washington, D.C. �73 APPENDIX B USE OF BANDS ON PRIMARY FEATHERS TO IDENTIFY MORTALITY OF MARKED MALLARDS Avian researchers have marked birds with leg bands since the turn of the century (Marion and Shamis 1977). Unfortunately, leg bands are often not recovered from birds killed by predators or from scavenged carcasses. Such carcasses are often represented just by the wings, which mayor may not be attached by the coracoids. In our study of wintering mallards (Anas p1atyrhynchos), loss of leg bands to predators or scavengers probably reduced the information we obtained from our banded sample the first year. To increase our sample of marked carcasses recovered, we developed a technique to mark a primary on each wing with a numbered monel band. METHODS Mallards captured in the San Luis Valley, Colorado, in January 1988 and January 1989 were captured in Salt Plains bait traps (Szymczak and Corey 1976) and cannon nets (Dill and Thornberry 1950). Data on age sex, body mass, and wing length were obtained for all birds, and a condition index (estimated body fat / estimated fat-free body mass; Ringe1man and Szymczak 1985) calculated for each. All birds were marked with standard USFWS leg bands, and most marked with similarly numbered size 4 monel bands (National Band and Tag Co. , Newport, Kentucky) placed on the rachis of primary VIII on each wing. To determine rates of wing band loss, mallards subsequently recaptured were examined for retention of both wing bands. To restrain the bird for attachment of the bands, a 50 cm x 50 cm piece of 0.8 mm p1exig1ass had 4 pairs of holes drilled as shown in Figure B.1. Cord was looped through the holes with 1 end secured by a knot and the other by a sliding catch. A mallard was secured by placing the head through the top loop and securing the loop snugly around the neck. The wings were individually pulled through a loop and the cord secured next to the body. The bottom loop could be placed over the abdomen for additional restraint if necessary. Epoxy was attached to the bottom of the p1exig1ass, or edges were trimmed, to assure all restraining boards were equal mass. The bird was then weighed on the restraining board to reduce handling. RESULTS In 1988 and 1989, 1,700 and 2,061 wing bands were placed on wintering mallards (Table B.1). Of the 161 mortalities of banded birds recorded in 1988, 96 (59.6%) were identified only by the recovery of at least 1 wing band from the bird. In 1989, 136 mortalities of banded birds were recorded, of which 107 (78.7%) were identified from wing band recovery. In both years, use of wing bands increased the recovery of marked birds. �74 30CM 3CM c5c9 0 ~$.SCM -, 13 eM 3OCM 0 7CM ~3CM 0 SCM 9CM Figure B.l. Diagram showing hole locations in the plexiglass panel used for restraining mallards. �75 Table B.l. Numbers of wing bands placed on wintering mallards in 1988 and 1989 in the San Luis Valley, Colorado. Number Win& Banded Age Adult Adult Immature Immature Sex Female Male Female Male 1988 1989 236 683 320 461 414 988 274 385 Forty-six mallards were recaptured and examined for wing band retention in 1989. Only 2 (4%) of the wing banded birds (1 adult and 1 immature male) had lost both wing bands (Table B.2). Six (13%) had lost 1 wing band (Table B.2). Although sample sizes are small, band loss appears to be greatest for immature females and least for adult males (Table B.2). Table B.2. Number of wing bands retained by 46 mallards recaptured in the San Luis Valley, Colorado, in 1989. All birds initially had 1 band attached to primary VIII on each wing. Age Sex Number of Wing Bands on Recaptured Mallards 0 2 1 Adult Adult Immature Immature Females Male Female Male 0 1 0 1 1 2 2 1 4 24 2 8 DISCUSSION We found the use of monel bands on primary VIII to be an effective technique to increase the recovery of banded mallards. Wing bands provided an individual marker that did not present any visible cues that might affect survival rates. Scavengers often remove leg bands from a carcass, which results in small recovery samples in studies focusing on non-hunting waterfowl mortality. By using wing bands, we were able to more than double our recovery samples. Additional uses included using wing bands to measure the number of mallard carcasses that were planted by one researcher and recovered by the people searching for waterfowl carcasses. Use of wing bands allowed the number of carcasses represented by single wings recovered to be estimated. Loss of wing bands probably reflects the interaction of 2 factors: diameter of the primary rachis and variability in who applied the wing band. Diameter of the rachis of primary VIII tends to be less in females than males, and smaller in immatures than adults. Consequently, loss in immature females should be greatest, as we observed. Additionally, the person applying the bands may not test the band by attempting to move the band on the rachis. If the band moved, it could be tightened before the bird was released. A potential problem with the use of wing bands involved applying too �76 much pressure during application such that the band crushed the rachis of the primary. When this occurs, the primary may be broken soon after release. We believe this rarely happened in our sample since no birds were recaptured with primary VIII broken. Use of wing bands provides researchers studying non-hunting mortality in waterfowl a technique to increase sample sizes. Application time is only a couple of minutes, and the individualized markers appear to have little effect upon the bird. LITERATURE CITED Dill, H. H., and W. H. Thornberry. 1950. A cannon projected net trap for capturing waterfowl. J. Wildl. Manage. 14:132-137. Marion, W. R., and J. D. Shamis. 1977. An annotated bibliography of bird marking techniques. Bird-Banding 48:42-61. Szymczak, M. R., and J. F. Corey. 1976. Construction and use of the Salt Plains duck trap in Colorado. Colorado Div. Wildl., Div. Rep. 6. l3pp. �77 APPENDIX C FACTORS INFLUENCING SIZE OF THE MALLARD POPULATION WINTERING IN THE SAN LUIS VALLEY, COLORADO Nichols et al. (1983) reported that mallards (Anas platyrhynchos) wintering in the Mississippi Alluvial Valley were recovered further south during cold winters, and more immatures and adult females were recovered in the Mississippi Alluvial Valley in wet winters than dry winter. Adult males did not show any response in wet and dry years. They concluded that temperatures, precipitation, and population size did affect winter distributions. Our objective in this study was to identify whether winter temperatures, precipitation and size of the breeding population are important factors determining the number of mallards wintering in the San Luis Valley, Colorado. STUDY AREA This study was conducted in the San Luis Valley (SLV), Colorado. The SLV is a 12,960-km2 intermountain basin in south-central Colorado, bounded by the San Juan Mountains to the west and the Sangre de Cristo Mountains to the east. As a result of the surrounding mountains and a valley elevation of 2,286 to 2,438 m, the SLV has an arid climate characterized by short, cool summers and cold dry winters (Lantis 1942). From 1964 through 1980, breeding pair estimates averaged 25,371 pairs, but have declined since then primarily because of decreasing habitat (Szymczak 1986). Mallards are the most abundant nester (Szymczak 1986). An average of 18,734 ducks wintered in the SLV (1984-90 average, Colo. Div. Wildl., unpubl. data); most are mallards wintering on the Monte Vista National Wildlife Refuge (MVNWR). Wintering mallards rest on open water maintained by pumping or artesian flow, and forage in nearby grain fields. METHODS To identify factors determining the number of mallards wintering in the SLV, winter survey and breeding mallard popUlation estimates in the SLV from 1964 through 1988 were obtained from the Colorado Division of Wildlife (unpubl. data). National Oceanic and Atmospheric Administration weather summaries for the Alamosa Airport reporting station were examined for mean monthly temperatures and total precipitation for November, December, and January for years corresponding to those that had winter inventory data available. Linear regression was used to identify relations among winter population size and the other variables. Linear regression was used to compare estimated breeding populations with the estimated wintering population the previous January to examine a relation between wintering population size and the subsequent breeding population. Relations between dependent variables were e~amined with the Pearson product-moment correlation. �78 RESULTS The number of mallards wintering in the SLV was influenced by size of the breeding population (F - 22.98, 1,18 df, f - 0.0002), December temperature (F - 9.12, 1,18 df, f - 0.0077), total precipitation in December (F - 8.33, 1,18 df, f - 0.0102), and an interaction between breeding population size and December temperature (F - 5.73, 1,18 df, f - 0.0284). Because year and breeding population size were strongly correlated (r - -0.828, f - 0.0001), the relation between year and winter population size that was first identified (F - 6.02, 1,18 df, f - 0.0397) reflects the change in breeding population rather than a relation between winter population and year. December temperature and precipitation were negatively correlated (r - -0.652, f - 0.0004). Winter population size was positively correlated with December temperature (r - 0.128, f - 0.5428) and negatively correlated with December precipitation (r - -0.224, f 0.2822). The breeding population in a year is related to size of the population the previous January (F - 14.08, 18 df, f - 0.0016). DISCUSSION The positive relation between breeding and wintering population sizes is not surprising considering many mallards are thought to be residents of Colorado (Szymczak 1986). The relations between winter population estimates and December temperatures and precipitation suggest more mallards winter in the SLV in winters that begin with warm, dry conditions. Mallards may be assessing food availability through December, and if environmental conditions reduce food availability, may then migrate from the SLV. Because counts are made in January, the lack of any relation with January weather variables may reflect that counts generally occurred before many birds could respond to severe weather. Although systematic counts were not taken during our study, our belief is chat no noticeable exodus of mallards from the SLV occurred during periods of severe weather in January. We believe that mallards will respond to severe environmental conditions, probably by responding to food availability by migrating in December, but severe conditions later in the winter do not result in migration of many mallards from the SLV. LITERATURE CITED Lantis, D. W. 1942. Sage of the San Luis Valley in Colorado. Folks and Fortunes 1:38-40. Nichols, J. D., K. J. Reinecke, and J. E. Hines. 1983. Factors affecting the distribution of mallards wintering in the Mississippi Alluvial Valley. Auk 100:932-946. Szymczak, M. R. 1986. Characteristics of duck populations in the intermountain parks of Colorado. Colorado Div. Wildl., Tech. Publ. 35. 88pp. �79 APPENDIX D NESTING CHRONOLOGY OF MALLARDS IN THE SAN LUIS VALLEY, COLORADO Several factors have been related to waterfowl nest initiation dates. Sowls (1955) reported a relation between environmental temperatures and nest initiation dates, but noted several inconsistencies in the data. Other factors that influence nest initiation dates are spring water conditions (Krapu et al. 1983), geographic origin (Batt and Prince 1978, 1979), diet quality (Eldridge and Krapu 1988), and winter body mass (Pattenden and Boag 1989). Our objective in this study was to determine the date of hatch for mallards nesting in the San Luis Valley, Colorado. STUDY AREA This study was conducted in the San Luis Valley (SLV), in south-central Colorado. The SLV is a 12,960 km2 intermountain basin, bounded by the San Juan Mountains to the west and the Sangre de Cristo Mountains to the east. As a result of the surrounding mountains and a valley elevation of 2,286 to 2,438 m, the SLV has an arid climate characterized by short, cool summers and cold dry winters (Lantis 1942). From 1964 through 1980, breeding pair estimates averaged 25,371 pairs, but have declined since then, primarily because of decreasing habitat (Szymczak 1986). Mallards are the most abundant nester (Szymczak 1986). An average of 18,734 ducks wintered in the SLV (1984-90 average, Colo. Div. Wildl., unpubl. data); most are mallards wintering on the Monte Vista National Wildlife Refuge (MVNWR). Wintering mallards rest on open water maintained by pumping or artesian flow, and forage in nearby grain fields. METHODS Nesting chronology of mallards was monitored using pair and brood counts. In 1987, brood counts were made whenever broods were encountered. A survey route was established on MVNWR and driven weekly from 19 April through 15 June 1988, and 22 April through 14 June 1989. Male mallards encountered were classified as paired (with female), lone male (single isolated drake without a visible associated female), or grouped (2-5 associated males in 1987 and 1988, 2-4 associated males in 1989). In 1989, groups of 5 males were not considered grouped to conform with standard operating procedures for breeding pair counts (H. Funk, pers. comm.). Broods encountered along the route or at any other time were recorded as to species, number, and age (Gollop and Marshall 1954). Brood ages were used to back-date to the estimated hatching date. �80 RESULTS Throughout the nesting period, many males appear paired (Figs. 0.1 and 0.2). In two of the three years, estimated hatch dates of mallard broods reflected the drawn-out nesting period suggested by the pair counts (Figs. 0.3, 0.4, and 0.5). In 1988, the hatching curve showed a peak in hatch the week of 1 June (Fig. 0.4). In 1987 and 1989, the hatch peak covered several weeks from mid-May through mid-June (Figs. 0.3 and 0.5). DISCUSSION The pair counts and hatch curves suggest that nest initiation is prolonged in the SLV. Whether this pattern results from the low winter body mass or other environmental factors can not be determined from these data. Pattenden and Boag (1989) reported that low body mass in winter had a greater influence on laying date than low body mass in spring. The extended period of nest initiation may reflect variability in the low body masses and the time required for a female to achieve the body condition necessary for a reproductive attempt. We do not know how rapidly a female can increase her body condition when environmental conditions improve. Heitmeyer (1988) suggested timing of events (e.g., prebasic molt and pairing) in the annual cycle of female mallards are influenced by a female's body condition, and the females that complete these events early may realize reproductive advantages. Unlike mallards wintering in Missouri, mallards wintering in the SLV had not initiated prebasic molt in January 1988 (C. W. Jeske, unpubl. data). This suggests mallards wintering in the SLV are not as advanced in their reproductive cycle as mallards in Missouri. Because most mallards wintering in the SLV do not migrate, they delay molt until after severe winter conditions have passed and molt while other wintering populations are migrating. This delay also may influence the initiation of nesting. More information on the reproductive chronology of this population is needed. We were not able to determine if winter body condition is related to subsequent reproduction. Consequently, this remains an important question in determining if the low body conditions measured in SLV mallards are related to the unusual nesting chronology in the SLV. The prolonged nesting chronology noted may reflect different reproductive efforts by immature and adult females, or it may reflect different efforts by residents and migrants. How age and residency influence a female's reproductive effort need to be explored. �81 260 ~--~--------------------------------------. 240 220 U 200 CD 180 +-' C 160 ::J a (J en 140 120 CD 100 C\j 80 ~ 60 40 20 o 4/19 4/26 5/10 5/17 5/24 6/3 6/8 6/15 Date of Count Figure 0.1. Number of male mallards counted from a census route on MVNWR and classified as lone or paired from mid-April through mid-June 1988. �82 260 240 _ Lone Male 220 ~ Pair Grouped Males r2Z3 '"C 200 Q) 180 ~ C 160 ::J a (J 140 120 CIJ 100 C'U 80 Q) ~ 60 40 20 o 4/22 4/30 5/6 5/14 5/22 5/30 6/8 6/14 Date of Count Figure 0.2. Number of male mallards counted from a census route on MVNWR and classified as paired, lone, or grouped from mid-April through mid-June, 1989. �83 40 CIJ -0 30 a a ~ CO '+- 020 MAY 15 JUNE 1 JUNE 15 JULY 1 JULY 15 Week Figure 0.3. Estimated hatch dates of mallard broods encountered in the SLV, Colorado, 1987. �84 40 CIJ "'C 30 a a \"". co lot- a 20 o MAY 15 JUNE 1 JUNE 15 JULY 1 JULY 15 Week Figure 0.4. Estimated hatch dates of mallard broods encountered in the SLV, Colorado, 1988. �85 CIJ U o 020 1- CO '+- o 1- CD ..c E 10 :::J Z o MAY 1 MAY is JUNE 1 JUNEiS JULY 1 JULY1S Week Figure 0.5. Estimated hatch dates of mallard broods encountered in the SLV, Colorado, 1989. �86 LITERATURE CITED Batt, B. D. J., and H. H. Prince. 1978. Some reproductive parameters of mallards in relation to age, captivity, and geographic origin. J. Wildl. Manage. 42:834-842. , and . 1979. Laying dates, clutch size and egg weight of --------- captive mallards. Condor 81:35-41. Eldridge, J. L., and G. L. Krapu. 1988. The influence of diet quality on clutch size and laying pattern in mallards. Auk 105:102-110. Gollop, J. B., and W. H. Marshall. 1954. A guide for aging duck broods in the field. Mississippi Flyway Tech. Sect. 14 pp. Heitmeyer, M. E. 1988. Body composition of female mallards in winter in relation to annual cycle events. Condor 90:669-680. Krapu, G. L., A. T. Klett, and D. G. Jorde. 1983. The effect of variable spring water conditions on mallard reproduction. Auk 100:689-698. Lantis, D. W. 1942. Sage of the San Luis Valley in Colorado. Folks and Fortunes 1:38-40. Pattenden, R. K., and D. A. Boag. 1989. Effects of body mass on courtship, pairing, and reproduction in captive mallards. Can. J. Zool. 67:495-501. Sowls, L. K. 1955. Prairie ducks. Stackpole Co., Harrisburg, PA. 193pp. Szymczak, M. R. 1986. Characteristics of duck populations in the intermountain parks of Colorado. Colorado Div. Wildl., Tech. Publ. 35. 88pp. �87 APPENDIX E NESTING OF INSTRUMENTED ADULT FEMALE MALLARDS IN THE SAN LUIS VALLEY, COLORADO Cross-seasonal interactions between winter body condition and reproductive success of waterfowl have been suggested (Heitmeyer and Fredrickson 1981, Heitmeyer 1985, Pattenden and Boag 1989), but no data are available on the winter body condition of individual females and subsequent reproductive success. Our objective in this study was to measure reproductive success of adult female mallards (Anas p1atyrhynchos) whose January body condition was known. STUDY AREA This study was conducted in the San Luis Valley (SLV) , in south-central Colorado. The SLV is a 12,960 km2 intermountain basin, bounded by the San Juan Mountains to the west and the Sangre de Cristo Mountains to the east. As a result of the surrounding mountains and a valley elevation of 2,286 to 2,438 m, the SLV has an arid climate characterized by short, cool summers and cold dry winters (Lantis 1942). From 1964 through 1980, breeding pair estimates averaged 25,371 pairs, but have declined since then, primarily because of decreasing habitat (Szymczak 1986). Mallards are the most abundant nester (Szymczak 1986). An average of 18,734 ducks wintered in the SLV (1984-90 average, Colo. Div. Wi1d1., unpub1. data); most are mallards wintering on the Monte Vista National Wildlife Refuge (MVNWR). Wintering mallards rest on open water maintained by pumping or artesian flow, and forage in nearby grain fields. METHODS We captured mallards wintering on MVNWR with Salt Plains bait traps (Szymczak and Corey 1976) or cannon-projected nets (Dill and Thornberry 1950) from 11-18 January and 20-22 February 1987 and 11 January through 4 March 1988. Data on age, sex, weight, and wing length were obtained, and a condition index (Ringe1man and Szymczak 1985) calculated for each. We summarized the condition indices of adult females captured the first trapping day of each trapping period, then assigned adult females to 1 of 3 condition classes based upon their condition index relative to the population distribution of the females captured. Subsequently, birds with a condition index in the upper 33% of the range were classified as in "good" condition, the middle 33% were considered "average," and the lower 33% were considered in "poor" condition. In January 1987, 52 adult females in good condition and 51 in poor condition were instrumented with 26 g back-mounted transmitter packages attached with a harness similar to that described by Dwyer (1972). In February 1987, 18 adult females in good condition and 17 in poor condition �88 were instrumented. In January 1988, 52 adult females in good condition and 51 in poor condition were instrumented. High mortality of instrumented females in 1987 prompted us to anticipate high mortality again in 1988. Initially, we planned to trap weekly and reuse radios collected from dead females, by applying radios to adult females based upon their condition index relative to the condition index classes used during the 11-18 January trapping period. This protocol resulted in transmitters being reapplied to 23 adult females, 6 in good condition, 1 in average condition, and 16 in poor condition. However, by early February extremely high mortality, difficulties with trapping, and changes in the condition indices of the population made it impractical to reuse radios weekly. Consequently, all radios subsequently collected were reapplied during trapping on 29 February through 4 March. Procedures for determining the relative condition of each bird compared to the population were similar to that used in early January. We instrumented 22 females in good condition, 17 in average, and 24 in poor condition in late-February and early-March. Standard techniques were used to obtain weekly locations of instrumented ducks from the ground (Cochran 1980:517-518) using vehicle-mounted and handheld antenna-receiver systems. Birds not found early in the week were located from the air (Gilmer et al. 1981) using a strut-mounted antenna attached to a Cessna 182. All birds located from the air were ground-checked the next day to determine if they were alive or dead. Any birds contacted from the ground that were in unusual locations, did not have attenuating signals, or were in the same location for 2 consecutive weeks were checked visually to determine whether they were alive or dead. In both years, instrumented females were monitored through June to locate their nests. When a nest was found, it was checked after the estimated hatching date to determine success. RESULTS In both years, 71 females were possibly alive during the nesting period. Each year, only 2 instrumented females were found with nests. In 1987, a female instrumented in poor condition in January and another female in good condition in February nested. In 1988, both nesting females were in poor condition when instrumented in January. None of these four females successfully hatched any ducklings. DISCUSSION The low number of instrumented females nesting each year may reflect effects of the transmitter package on weight dynamics of these instrumented females, or females wintering in the SLV may have low reproductive rates. Instrumented females may not be able to recover body mass lost during the instrumentation adjustment period, and never obtain the body reserves important for mallard reproduction (Krapu 1981). Alternatively, the generally low body condition of mallards wintering in the SLV may result in low reproductive rates. Our results suggest that body mass losses identified for instrumented birds (Chapter 4) affected survival of our samples. We caution other researchers to carefully evaluate any effects their marking technique may have on the behavior being studied. �89 LITERATURE CITED Cochran, W. W. 1980. Wildlife telemetry. pp. 507-520 in S. P. Schemnitz (ed.). Wildlife management techniques. Fourth ed. Wildlife Soc. Inc., Washington, D. C. Dill, H. H., and W. H. Thornberry. 1950. A cannon projected net trap for capturing waterfowl. J. Wildl. Manage. 14:132-137. Dwyer, T. J. 1972. An adjustable radio package for ducks. Bird-Banding 43:282-284. Gilmer, D. S., L. M. Cowardin, R. L. Duval, C. W. Schaiffer, and V. B. Kuech1e. 1981. Procedures for the use of aircraft in wildlife biotelemetry studies. u.s. Fish and Wildl. Serv., Resour. Pub1. 140. 19pp. Heitmeyer, M. E. 1985. Wintering strategies of female mallards related to dynamics of lowland hardwood wetlands in the Upper Mississippi Delta. Ph. D. Thesis, Univ. Missouri, Columbia. 376pp. _____ , and L. H. Fredrickson. 1981. Do wetland conditions in the Mississippi delta hardwoods influence mallard recruitment? Trans. North Am. Wildl. Nat. Resour. Conf. 46:44-51. Krapu, G. L. 1981. The role of nutrient reserves in mallard reproduction. Auk 91:28-38. Lantis, D. W. 1942. Sage of the San Luis Valley in Colorado. Folks and Fortunes 1:38-40. Pattenden, R. K., and D. A. Boag. 1989. Effects of body mass on courtship, pairing, and reproduction in captive mallards. Can. J. Zool. 67:495-501. Ringelman, J. K., and M. R. Szymczak. 1985. A physiological condition index for wintering mallards. J. Wild1. Manage. 49:564-568. Szymczak, M. R. 1986. Characteristics of duck populations in the intermountain parks of Colorado. Colorado Div. Wildl., Tech. Publ. 35. 88pp. _____ , and J. F. Corey. 1976. Construction and use of the Salt Plains duck trap in Colorado. Colorado Div. Wildl., Div. Rep. 6. l3pp. Prepared by: Michael R. Szymczak Wildlife Researcher C ��91 Colorado Division Wildlife Research October 1991 of Wildlife Report JOB PROGRESS REPORT State of Colorado Project W-152-R-4 Work Plan Job Title: 1__ : Job Evaluation Period Covered: Authors: Avian Research - Migratory Game Birds 19 of nesting habitat 1 April 1990 through David W. Gilbert. management 31 March James K. Ringelman for ducks 1991 and David R. Anderson Personnel: D. Anderson, Colorado Cooperative Fish and Wildlife Research Unit; J. Ringe1man, M. Szymczak, Colorado Division of Wildlife; D. Gilbert, Colorado State University; S. Brock, S. Berlinger, A. Morkill, R. Schnaderbeck, U.S. Fish and Wildlife Service. ABSTRACT Personnel at the Monte Vista National Wildlife Refuge (MVNWR) used transect surveys to collect information on duck nests annually since 1964 (except 1977). Total transect distance was originally (1964-1968) 323.6 mi per year (one walk), but was latter reduced by half. The 141 transects (71 later years) were run twice annually (plus additional nest fate checks). Transect surveys sampled 4-5.5% of the 14,189 acre refuge. Mallards composed 54.5% of the duck nests on MVNWR, followed in abundance by teal species and pintail. Units 6, 9, and 18 had highest nest densities (1164, 1050, and 1366 nests/mil respectively) and units 5, 20, and 23 had the lowest nest densities (27, 156 and 140 nests/mil respectively). Refuge-wide nest densities ranged from a low of 92.7 nests/mi2 in 1979 to a high of 417.9 nests/mi2 in 1973. Mean annual nest success was 49.4%, (52.4% for the 4,156 nests included over the entire period) but ranged from 27% in 1978 to 72% in 1966. Available habitat was quantified using color-infrared aerial photography, ground truthing and geographic information system software. Ground cover on the MVNWR was 48.4% baltic rush, 29% greasewood and rabbitbrush, 4.8% saltgrass, 3.4% cattail and 4.6% grass species. The remainder of the refuge has varying small amounts of 6 other vegetation types. Ten percent of the refuge was non-surveyed, water was 4.25% and roads were <1%. The 2 most predominant cover types, baltic rush and greasewood, accounted for 68% and 16% of total nests, respectively. Species composition and nesting success of 4,156 duck nests located over 26 years provide a basis to evaluate the benefits of habitat management practices including grazing, water application, burning and predator control. Three different grazing schedules were utilized. One, the rest rotation schedule, will allow grazing effects to be evaluated. Grazing ranged from 0.0 to 1.8 AUM/acre. Prescribed burning treatments were conducted 14 times. Eight such burns were evaluated and �92 ABSTRACT (cont.) indicate a decrease in nest density the year of the burn. Predator control was conducted at 3 intensities. The first period saw widespread use of poison and was significantly effective in decreasing nest failure due to predation as compared to a second, limited predator control effort period. A third period again had intensive predator control, but no poison and did not significantly reduce predation on .duck nests. Evaluation of a haying program (1964-1976) indicated that there was a slight impact on nesting density during cutting. Some confounding was found between treatments and is reported. The interim completion report for this project is presented this segment. Additional analyses will be performed next segment, and a comprehensive monograph will be prepared for submission to a scientific journal. �93 EVALUATION OF NESTING HABITAT MANAGEMENT FOR DUCKS David W. Gilbert James K. Ringelman David R. Anderson P. N. OBJECTIVES 1. Relate nesting duck species composition, density to habitat management practices. nest success rate, and nest 2. Assess changes in wetland and upland vegetation between 1962 and 1985 by contrasting digitized habitat information derived from aerial photographs. 3. Determine nesting habitat preferences of duck species by comparing of nesting habitat with relative habitat availability. usage SEGMENT OBJECTIVES 1. Locate records on grazing, burning, water application and predator control (habitat treatments) for individual wetland units on the Monte Vista National Wildlife Refuge (MVNWR). Computerize these data for entry as independent variables for categorical data analyses. 2. Conduct categorical analyses of nest transect data, employing log-linear models, logistic regression and contingency table methods to test hypotheses concerning the effects of habitat treatments on the species composition, nest success and nest density of breeding waterfowl. 3. Obtain aerial photography of the MVNWR for three time periods (1960's, 1970's, and 1980's). Using ARC/INFO and other remote sensing software, digitize vegetation habitat maps for the MVNWR. Calculate areal coverage of habitat types, and relate changes over time to habitat management practices and natural vegetation succession. 4. Relate changes in waterfowl nesting densities, distribution success to changes in vegetative associations over time. 5. Write interim and final reports, and submit findings to relevant scientific publications and interested natural resource agencies. and nesting STUDY AREA Duck nesting and management records evaluated were collected on the MVNWR in the San Luis Valley (SLV), 9.6 km (6 mi) south of Monte Vista, Colorado. The MVNWR is at 2,316 m (7,600 ft) elevation and is considered a cool desert, high elevation mountain park. Surrounding mountain ranges, including the San Juan range on the west and the Sangre de Christo mountains on the east, collect snow pack and subsequently supply water to the SLV via the Rio Grande River and a network of irrigation canals. Underground sources of water from �94 vast aquifers also are an important source of water for both the SLV and the MVNWR. The refuge began to use these water sources to produce and enhance wetlands in 1953, when it was established. A network of artesian wells, pumped wells, ditches and dikes have since been developed to keep many of MVNWR 24 management units (5,742 ha, 14,189 acres) flooded to varying depths for waterfowl and other wildlife to use. The SLV receives less than 30 cm (12 in) of precipitation annually. Therefore, the refuge and surrounding agricultural communities rely on irrigation and pumped water from underground sources to supplement the low precipitation. Vegetation on the MVNWR is predominantly baltic rush (Juncus balticus), greasewood (Sarcobatus vermiculatis), and rabbitbrush (Chrysothamnus spp.). General topography can be described as extremely flat with an overall minor downward slope from west to east. Szymczak (1986) further describes characteristics of the MVNWR. The physical boundaries of the refuge stretch 12.9 km (8 mi) east from the base of the San Juan mountains, and is 6.4 km (4 mi) wide from north to south. The 24 individual management units are separated by fences, roads, and/or ditches (Fig. 1). Details on size and other characteristics for each unit are given in Appendix A. Most units have the ability to be managed separately from other units with respect to grazing, burning and water application. The MVNWR has recorded 194 different species of migratory and nonmigratory birds; about one-third (66 species) are listed as common or abundant during some time of the year (USDI, USFWS 1986). Principal duck species nesting include mallards (Anas platyrhynchos), gadwall (Anas strepera), pintail (Anas acuta) blue-winged teal, (Anas discors), cinnamon teal (Anas cyanoptera), shoveler (Spatula clypeata), and redhead (Aythya americana). Other duck species nest at MVNWR, but are less numerous and were not examined in this study. Banding studies (Szymczak 1986) reveal that the mallard population is largely resident year round. The other species are mostly migratory, and few winter in the SLV. METHODS The following sections describe the many sources of data collected and how they are used in analysis. Nest data originate from actual or summarized survey forms provided by the refuge. These data were entered into a general computerized spreadsheet for analyses using SAS (SAS Institute 1987) and Quattro Pro (Borland International 1990). Area, size, length of transects and other similar measurement information used to derive density estimates and cover area were computed from GIS (Arc-Info) analysis of aerial photographs. Refuge features were corrected for scale using U.S. Geological Survey 1:24000 topographic maps. Nesting Data Nesting survey data used in this analysis extended from 1964-1990 excluding 1977, a drought year when no survey was conducted. The duck nests detected and used in all evaluations were as follows: �GJ 3: - .•.. "•.... "0 •.... C ftJ •.... ftJ o .•.. ~ ftJ z: ~ en ..-> CIJ o C ~ X - i ; 95 �96 Species No. Detected Mallard Percent 2,264 54.5 Teal Pintail Gadwall Shoveler 563 359 251 116 13.5 8.6 6.0 2.8 Redhead Unknown 83 529 2.0 12.6 4,156 ·100.0 Total Other minor species were removed from analysis. Teal species include bluewinged teal, cinnamon teal and green-winged teal (Anas crecca), but nearly all (550 of 563) were blue-winged or cinnamon teal. Nest data were used for all nests located during mid-May and mid-June searches. June walks were used to determine the fate of May nests as well as to locate new nests. A July walk was included only for the purpose of determining the fate of nests located in June. In some years, nests located in July were included because nesting in those years was thought to be delayed. These nests, however, were removed from analysis to maintain consistency. Analysis of May to June nest ratios in years with July nests do not reveal exceptionally later than normal nesting chronology when the entire period is examined. Production and density estimates from years where late nests were recorded are not comparable to other years. Appendix B gives the year and number of late nests eliminated from analysis and chronology of nesting between years, thus the rational for their removal. Information collected at each nest included: date, unit number, transect number, species, incubation stage, number of eggs, and vegetation cover surrounding the nest. If nest failure was encountered, probable cause was recorded. Distance to the transect centerline was often collected as an aid in nest relocation, but in some years distance data were discarded nest records were summarized. The 529 nests recorded as "unknown" were reapportioned among the other species based upon species composition values for each year. Unknown nests typically were labeled as such if the hen was absent from the nest or the eggs destroyed. The percentage of unknown nests in the early years was greater than in later years (Appendix C), probably because a reference egg collection used in later years facilitated identification. Information on refuge management programs was taken from Annual Narrative Reports (MVNWR 1964-1990), and was supplemented by detailed files on specific activities such as grazing and water application. Data in this report can be considered final for the period of record described. Values were derived from actual nest data forms collected, not from summarized reports. Transect Survey The transect survey design resulted designed such that 640 acres of transect from pre-sampling in 1961-63. sampling would result in nest It was �97 estimates within 10-15% of the true value 95% of the time. In early years (1964-68), 142, 16.5-ft. wide "strip" transects, 300 feet apart (west to east) were used. At w - 8.25 ft., (8.25 ft./side of transect centerline), 5.5% of the refuge was sampled. The transects covered approximately 320 mi. each walk (described below), sampling 640 acres of the 11,570 acres available for nesting. The survey technique called for searchers to "line sight" 2 or more numbered, transect marker poles with the aid of binoculars. The searcher then walked the transect centerline and observed for nest signs or ducks flushing from a nest. Upon location of a nest, surveyors marked nests and recorded distance measurements from transect centerline so the nest could be relocated during the subsequent nest fate search. Nesting information was also recorded at this time. Beginning in 1969, the number of transects were halved by removing all even-numbered transects. In order not to decrease the area sampled by half, transect width was increased to 24 ft. (w - 12 ft.), resulting in coverage of 4.0% of the total nesting habitat. Sampling in 1971 was further reduced (that year only) to walking every other odd-numbered transect (1 of every 4 original transects) because of a labor shortage and a very dry year when nesting was thought to be poor. Since 1972, all odd-numbered transects have been walked twice each year. Transect widths were kept at 24 ft from 1971 through 1979, changed back to 16.5 ft during 1980-1986, again increased to 24 ft for 198687, then changed back to 16.5 ft wide 1988-1990 (Table 1). All data are corrected to adjust for these variations, although these changes cause the precision of the nest density estimates to vary. Nest abundance was estimated from the number of nests found in the strip divided by the fraction of the area sampled. Production estimates were derived from the number of total nests corrected for the percent of each species in the breeding population, species composition of the nests found, percent hatch success, and brood size for each species (see Appendix D for weighting calculations). Our interest was with nest densities and variability in nest density through time in relation to a variety of natural conditions and management treatments. Brood data and production estimates were not considered here. All values are reported as nests/mi2 and are corrected for nests undetected in the strip. Estimation of Nest Density An assumption of a strip transect require that all nests are found within the strip. This assumption was shown to be unrealistic by Anderson and Pospahala (1970), who estimated that approximately 11% of the nests at MVNWR were not detected, even with the narrow transect width of 16.5 ft. Thus nest density was estimated using contemporary theory for line transect sampling (Burnham et al. 1980 and Buckland et al. 1992) and the distance from detected nest to the transect centerline. The percent of nests undetected varies with �98 Table 1. Summary of changes in transect survey statistics due to changes in search effort, 1964-1990.~ Years shown in bold italic indicate that measurements from each nest to the centerline of the transects are available. Width Year Total Acreage Year Width Total Acreage transect sampled ft. transect sampled ft. (2w) length, (2w) length, one walk one walk 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 16.5 16.5 16.5 16.5 16.5 24 24 24 24 24 24 24 24 323.6 323.6 323.6 323.6 323.6 170.6C l70.6c 80.7 161.6 161.6 161.6 161.6 161.6 647 647 647 647 647 496 496 253 469 469 469 469 469 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 24 24 16.5 16.5 16.5 16.5 16.5 16.5 24 24 16.5 16.5 16.5 161.6 161.6 161. 6 168.8 168.8 168.8 168.8 168.8 168.8 168.8 168.8 168.8 168.8 469 469 323 337d. 337 337 337 337 337 491 337 337 337 aValues may differ slightly «1%) from that reported by MVNWR but were determined using GIS and are assumed to be more accurate. bNo survey was conducted in 1977. CUnit 6 had extra effort during 1969 & 1970. dUnit 24 added to survey in 1981-90. changing strip width over the years (see results). Therefore, unbiased estimates of nest density (D) were computed with program DISTANCE (Laake 1991) using the following general estimator: 15 - nr(o) /2L where n - the number of nests det~cted within the strip, the estimated probability density function of the distances, L the length of transect line. £(0) The central issue in estimation of density from line transects is the selection of a proper model for £ (x) , the probability density function for the distance data, x. Here a half-normal key function was used with a Hermite polynomial adjustment function, however estimates of nest density were similar for other models of £(x). The estimator of D can be used for individual �99 species, management units, and years, or can be pooled over some of these categories. Standard procedures were used to estimate standard error, confidence intervals and other measures of precision. Data were also evaluated for differences in detectability among species and between years using Chi-square tests (whenever nest distance data from transect centerline were available). Years were pooled to increase sample size when evaluating differences in detectability between species. Final estimates for each year were derived using data from all duck species and all years with distance measurements at both w - 8.25 ft and w 12 ft. This approach fully utilized available distance data (n - 1,476 nests with distances for w - 12 ft, and n - 1,788 for w - 8.25 ft distance years). Values used to correct any w - 12 ft year (refer to Table 1) were obtained from these data. However, data were truncated at 11 ft-1l in. to remove heaping near the transect boundary. Values to correct derisity for any year in which w - 8.25 ft were calculated the same way. Here again, the last 0.25 ft were truncated to eliminate heaping near the transect boundary. Density estimates for individual species were calculated using the same methods, but because of smaller sample size had less precision. All estimated nest densities reported here are corrected for undetected nests using methods outlined above. Vegetation Mapping, Transect Length and Area Calculations The Arc-Info Geographical Information System (GIS) was used to derive the cover attributes of area, patch perimeter, total transect distance, and transect distance within cover patches. These measurements are important to determine use/availability values. The original polygon delineation of refuge features used in GIS were derived from color infra-red photography of MVNWR, at a scale of 1:7,920 (8 in on photo equaled 1 mi). Polygon size was limited to approximately 1 acre as the minimum mapping unit (Campbell 1987). Some smaller polygons occurred where a road or other feature bisects a given polygon. Polygons of vegetation associations were delineated on mylar overlays wherever distinct contrast differences could be detected between patches from the photograph. A baltic rush patch, for example, could easily be delineated form a greasewood upland community directly from the aerial photographs. However, some patches that had even interspersion of several plant species initially had only the external perimeter delineated, then a ranking of species dominance was later assigned during field inspection ("ground truthing"). Although ground truthing was done 5 years after initial photography, it is unlikely that any patch changed significantly with respect to composition. It is more likely that patch boundaries advanced or receded for a given patch but composition remained constant. Most fields had to be walked thoroughly to determine actual plant composition of plants for a dominance ranking. Some species appeared dominant when in bloom and viewed from a distance, but actually make up a small fraction of the vegetation complex. Most patches were single species, but if not, patches were ranked for the 3 most predominant species in descending order. Patches with >4 dominant plant species were labeled as "transition". A typical example would be a seasonally-flooded greasewood-rabbitbrush community in transition to a spikerush-baltic rush, sedge-wet meadow (scientific names given in a later table). Only a relatively small fraction «25% of vegetation polygons) of the refuge could not be classified into patches containing 1 dominant plant �100 species. Transect locations were noted on the mylar overlays whenever a transect marker pole could physically be located on the photograph by association to a recognizable feature. Upon completion of ground truthing, unit boundaries, polygons, and known transect locations were digitized into Arc-Info GIS. Remaining transect lines were generated easily by Arc-Info because of the known spacing. Attribute information were then entered into the GIS, yielding summary information. Digitizing and polygon summary concerning area, perimeter, and individual transect length was performed by TGS Inc., Computer Mapping Division, Fort Collins, Colorado. Use of photographs was restricted to the approximate effective areas. Mylar overlays were computer-registered to U.S. Geological Survey, 1:24,000 topographic maps to correct for any photographic distortion error. The polygon "stretching" required for correction was minor, but enhanced the accuracy of polygon boundaries. Original study plans called for comparison of the 1985 analysis with earlier years aerial photographs. Two sets of photographs were found (1969 and 1978 black and white photographs) and were enlarged for the comparison. It was hoped to that changes in polygon boundaries could be detected, thus tracking succession and wetland development. Unfortunately, delineation problems associated with the black and white photography prohibited this analysis. Only major wetland development changes were obvious and some general trends were observed and are reported. This technique offers promise if additional infrared photography is obtained at a future date. Additional Data Evaluated Data were also available on water conditions in the SLV for determination of relative nesting conditions off of the refuge. Data for this purpose came from gauging station records collected above Del Norte, Colorado. This station serves as a monitor of water within the Rio Grande drainage prior to extensive diversion for irrigation. Finally, also included are estimates of breeding populations from surveys performed in the SLV by the Colorado Division of Wildlife for the same period as the nesting survey. This survey was an aerial transect survey which enumerated breeding pairs and included most of the SLV. Breeding pair populations within the SLV are probably related to the nesting density on the refuge. RESULTS Species Composition Adjustments A total of 4,156 duck nests was found during 1964-1988 (Table 2). Annual differences in number of nests found are attributable both to real changes in nest density and variable search effort. Mallards were the most common nesters, followed in abundance by teal species, pintail, and gadwall (Table 3). Species abundance changed over time (Fig. 2). �100 , ,< i it <. K:SI5 iJ 5 i , < i ts:SJ '''" "'" '"" "'11 1968 1970 1972 1974 I< <I LL:1Ii [ <• p::::;;i4 m r7 ] i," <l , < i 1< \j [ '" i K <I i '" q iC '" 10 ; , m P' 7 j c:s::::s r 7] V 71 I eo ••••• coo Q) ~ Q)4O D.. 20 0' "II "'" 1964 1966 1111' 1976 11"' 11"' 1979 1981 11'11 1983 11"' 11"' 1985 1987 1111, " 1989 Year D Mallard • Teal III Pintail II] Gadwall ~ Shoveler ~ Redhead Fig. 2. Species Composition of duck nests on MVNWR, 1964-1990. t-' o t-' �102 Table 2. Number of duck nests found on transects during 1964-1990.8 Year Number Year Number 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 261 212 260 244 315 230 229 114 127 239 109 65 95 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 Total 55 53 89 104 110 114 150 152 197 262 150 130 90 4,156 Data represent real changes in nest density caused by differential search effort (see Table 1). Nest data were not collected in 1977. 8 Estimation of Nest Density Nest detectability did not differ among species or years (P > 0.05). A histogram of distance data available for all duck species in all years when w - 12 ft illustrates the decreasing probability of nest detection with increasing distance from transect centerline (Fig. 3). Heaping of nests at 0 and 3 ft and apparent "missing" nests at 1 ft are shown. Heaping at 3 ft probably results from rounding to "1 yard" rather than measuring nest distances. A similar situation also occurs at 10 ft. The low number of nests at 1 ft is caused by observers categorizing nests close to centerline as being on "center1ine". The 12th ft interval actually was only 1 in. wide (at 144 in.) because nests were only included out to 12 ft. This also is the case for the last 0.25 ft in the w - 8.25 ft years. Histograms for most species and years show little deviation from this pooled histogram, as reflected in the Chi-square tests. Histograms for pintail and teal, however, did differ slightly from than that shown for all ducks. Teal nests were generally less detectable than other species and the pintail data were problematic because fewer than expected were found near centerline due to their substantial preference for greasewood. Greasewood is sturdy, full of thorns, and may be avoided by nest searchers, thus causing nests near the centerline to be missed. Detectability difference of nests in the 2 major vegetation groups, greasewood and baltic rush, were insignificant (X2 - 7.12, P - 0.789), but the greasewood histogram resembled that of the pintails. Estimates of annual nest density varied markedly during 1964-90 (Fig. 4, Appendix E). The sampling variance of n, (var(n)), was about 1.7 n, and the coefficient of variation (cv) of n ranged from 7.3 to 17.9%. The cv(f(O)) �200 ~--------------------------------------------------~ 150 •.. Q) .0 E 100 ::l Z 50 o centerline 1 2 3 7 8 4 5 6 Distance from transect centerline 10 9 {Ft 11 12 1 Fig. 3. Nest detection at distances from transect centerline, data from all species during all years with distance data available. (n =1,476 nests) f--' o w �500 ~-----------------------------------------------. --~cen 400 Q) -0 t5 Q) 300 c "'C Q) 10 -_E t5 w 200 100 o 1964 1966 1968 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 Year Fig. 4. Estimated nest density, nests/sq. mi. MVNWR 1964-1990, 19n excluded • ..:t a rl �105 varied from 6.1 to 10.1% during 1964-1990. Analysis of the relative contribution of var(n) vs. var(£(O)) indicated the 2 sources to be of similar magnitude. Mean estimated density for all survey years and cv values for the w - 12 ft years for individual species are: CV Species Mean estimated density Mallard 164.5 8.2 Teal 49.5 10.1 Pintail 35.6 8.5 Gadwall 19.8 11.6 Shoveler 10.4 17.2 The above values indicate greater uncertainty in the estimates of nest density for the minor species due to smaller sample sizes. Some units of the refuge consistently have more nests than do other units (Fig. 5). Annual importance to nesting of each unit fluctuates and is not shown, but clearly there are large differences between the best and worst units. The 4 best units (3, 8, 9 and 18) compose 22.9% of available nesting habitat and account for 49.3% of nests detected. Conversely, the worst 4 units (5, 20, 22 and 23) account for 12.4% of available nesting area but only 3.2% of the nests found. Unit 24 is a better nesting unit than figure 5 would indicate; it was added to the survey in 1981 and only nests found from 1981 to 1990 are reflected in the figure. Appendix F gives yearly and mean nest density for all units. Breeding Pair Survey A positive relationship with low correlation (r - 0.0896) existed between mallard nesting density on MVNWR and the mallard breeding population surveyed in the SLV (Fig. 6). Variability in the data are probably caused by relative differences in water conditions between the MVNWR and the SLV. Most mallards are resident year round in the SLV, (Szymczak 1986) and are therefore partially dependent on wintering conditions in just 1 region. This same relationship would not be expected for the more migratory species because many migrate out of the SLV after the survey and do not winter in the SLV. Nevertheless, refuge nest density, heavily influenced by mallards, is only a weak indicator of the annual breeding population in the SLV. Total numbers of breeding pairs have declined in the SLV but remained relatively stable at the MVNWR (Fig. 7); particularly for mallards (Fig. 8). Thus, the refuge is important in maintaining local duck populations. There is no apparent shift to MVNWR as habitat was reduced elsewhere in the SLV, merely a reduction in numbers valley-wide. Szymczak (1986) believed that reduced habitat in the SLV was caused by the reduction of flood irrigation, which create wetlands from runoff. The relationship between North American breeding pair surveys (U.S. Fish and Wildlife Service 1989) and the SLV breeding pair survey show low correlation for either mallards, total ducks, or total species other than mallards (r - 0.149, 0.187, and 0.099 respectively). Thus, the �t-' o 0\ 25 ~------------------------------------------------------------~ ~ I '::::: It@rl " of available area contributed by unit Q) rn 15 t-------------I - CO ....., "of all nests found {:::::::';:':In each unit Note:Units12 & 13 arenotsurveyed. C Q) ::;::: .;:::~~: U 'Q) n, 10 I m',f--. --------1 ~:}i ..::: :;~~~f 5 t---- ::.:-. o 1 2 3 :~~~~t .:}: ..;:.: :::~;: 4 5 :'i:: 1 6 7 8 9 10 11 14 15 16 17 18 19 20 21 22 23 24 UnH Fig. 5. Percent available area and nests for each unit, MVNWR 1964-1990, 1977 excluded. See Appendix A for % nests I % area values. �300 ~----------------------------------------------~ 13 .. o o o o 50 0 o o ~----~----~~----~----~------~----~------~ o 2,000 4,000 6,000 8,000 10,000 12,000 Estimated mallard breeding pairs, SLV 14,000 Fig. 6. Relationship between mallard nests MVNWR & mallard breeding pairs SLV, 1964-1990. No survey done In 19n & 1985. f-' a -...J �I-' o co 35,000 CJ) l.- [ 30,000 II CJ) c .- II SLVtotal pair. MVNWR excluded MVNWR total pairs - - 25,000 r- 20,000 ~ - 15,000 i- - - - - - - - - - U Q) Q) l- .e 0 - - - _- - - - - - - - - - - - - - - - - - - - _- - - - - - -- - - - - .- I- Q) .e E 10,000 ~ - - ,-- 5,000 >--III--~lf~ -~~~-~tt~J~11 -- 68 68 70 72 74 76 78 80 82 Year Fig. 7. SLV total breeding pairs and total breeding pairs on the MVNWR. No survey was conducted In 19n & 1985. - - - ~ ~ - - ..- ; .' 0 64 - ._I ::J Z -- 84 8S I'll ~ 88 J -90 �14,000 .----------------------------. 12,000 1----- ~ ·cuCo 10,000 t- 8,000 I- 6,000 I- 4,000 I- 2,000 I- O) ----.-.--.--.-- .. • II SLY mallard pairs MVNWR excluded MVNWR total pairs .5 "0 m... .0 '0 ... ~ -.- E ::l Z ~I~. o 64 66 sa 70 72 74 76 78 eo e2 84 e6 ee 90 Year Fig. 8. SLV mallard breeding pairs and mallard breeding pairs on the MVNWR. No surveys were conducted in 1977 & 1988. I--' o \.0 �110 SLV breeding populations evidently fluctuates relatively independent of the North American population surveyed. Water Early water, defined as water applied to wetland units before 1 April, is needed to attract nesting ducks to the MVNWR (Schroeder et al. 1976). The MVNWR staff secured increasing amounts of water for early water applications (Fig. 9). Valley conditions (as indexed by Rio Grande River water), however, are more susceptible to the natural variations in precipitation (Fig. 10). Figure 11 shows the water by source for the totals given in figure 9. Vegetation Composition at MVNWR The MVNWR, which totals 13,912 acres (not included are unit 12, or units 13 and 19 west of the Monte Vista Canal), was classified into 697 polygons for GIS analysis. Nine polygons (1,506 acres) were not surveyed (generally agricultural fields). Five additional polygons (88 acres) were county roads between units and the Empire Canal. These 14 polygons were removed from consideration. Of the remaining 683 polygons important to nesting ducks, 513 (75%) were single-feature or single-vegetation patches. The remainder were patches consisting of 2 or more species in a patch. All areas (except the 591 acres of open water and the non-surveyed areas) were considered as habitat available for nesting, because ducks are known to nest on dikes, ditch banks, and near buildings. Accordingly, total area available to nesting ducks was 11,727 acres (18.3 mi2, excluding unit 12, units 13 and 19 west of Monte Vista Canal, all non-surveyed areas and open water bodies). There were 11 types of vegetation and transition type associations (4 or more dominant species in 1 patch Table 4). Even though there were 71 combinations of these types, the majority of polygons could be delineated into single-species patches unless they were extremely heterogenous. A total of 772 acres (146 polygons, 6% of the surveyed refuge) did not contain vegetation and were classified as: No. polygons No. acres Attribute 81 Open Water 591 28 134 Dike roads 7 9 Parking area& 30 38 Ditches 146 772 &Includes research facility. Individual unit characteristics varied (Appendix G). Addition printouts describing individual polygons will be filed at MVNWR and are not reported here. All digital data and computer-generated maps derived from the GIS work �600,000 ,.----------------, 35,000 ,.---------------, - 30,000 •.. ca:: 500,000 r---t 25,000 ;S Q) _ 1U 3: - 20,000 ! 15,000 e--I-I •.. 0 -c 1::: 11111111 _ lu III;IIJIIIIIIIIIIII ~ 300 000 , 1----------. 200,000 ~ ••••••••• - •• - I111I1I111111I 0'···················--------1 90 75 80 85 70 65 Year fig. 9. MVNWR water total, all sources for 1964·1990. ' '0 1D .e - 400 000-- 100,000 01---------------------------. 65 70 75 80 85 90 Year Fig. 10. Rio Grande total water measured at the Del Norte gauging station, 1964·1990. (An Index of water conditions In the SLV) I--' I--' I--' �•..... •..... N 35,000 ,---------------------------. Monte Vista Canal 30,000 II Spring Creek HtM Artesian Water Pumped Water Empire Canal Other Decreed Water 25,000 t-----------------------l ~ Q) 'ta ~ 20,000 (5 Q) .e e 15,000 ~ 10,000 5,000 o 64 66 68 70 72 74 Year 76 78 82 84 86 "rot. 1981, pump & artesian omitted separately brol Fig. 11. Water by source MVNWR 1964-1990 80 1982, pump & artesian omitted separately 88 90 �113 are also on file with the Migratory Division of Wildlife. Nesting Game Bird Program Unit of the Colorado Cover Preference The MVNWR is dominated by 2 plant species: baltic rush and greasewood (some rabbitbrush also in association with the greasewood). During ground truthing, greasewood communities were classified as either having an understory or devoid of understory (e.g. greasewood on otherwise alkali field). Ducks nested more commonly in greasewood bushes with some understory (typically rabbitbrush, baltic rush, saltgrass or weeds). Nest records do not provide comparable detail, but this information was quantified during ground truthing and may be useful to refuge managers in deciding where to develop wetlands. Preference for nest vegetation is evident from nest records (Table 5). Of the 4,156 nests evaluated, 2,582 (62%) were found in baltic rush (48.4% of the refuge). An additional 226 nests were found in baltic rush combined with another species, but with baltic rush the dominant type. Thus, 68% of all nests were associated with baltic rush. Conversely, greasewood covers 29% of the refuge yet accounted for only 16% of nests (619 nest sites). Less than half (295 of 619) of all nests were found in pure stands of greasewood; the remainder were in mixtures of greasewood and some other vegetation. Of the 226 baltic rush/combination nests, 66 were in cattail, 65 were in spikerush, 47 were in sedge, and 15 in greasewood; the few remaining nests were distributed among the other types (Table 5). The greasewood/combinations consisted of mixtures of baltic rush (140 nests), saltgrass (121 nests), and other grass species (31 nests). Many nests located in other nesting cover types also had baltic rush as a vegetative component. Saltgrass, cattail and weeds also serve as important plant species when nests are located in combination with other vegetation. Duck species preference varied slightly among cover types (Table 6). Pintail preferred greasewood and saltgrass, whereas shoveler appear to avoid greasewood. Redhead nest in cattail more than other species, whereas gadwall avoid this covertype. However, gadwall appear to be the most general nester, nesting at a higher rate in the "all other" category more often than the other duck species. �114 Table 3. Species composition of duck nests found (i.) for the period 19641990, adjusted for unknown nests. Year Mallard Teal spp. Pintail 1964 71 8 11 1965 73 8 1966 73 1967 Gadwall Shoveler Redhead 6 2 3 8 5 2 5 8 10 5 3 1 70 8 12 4 3 2 1968 72 8 13 4 2 1 1969 66 10 12 6 4 2 1970 62 14 16 2 5 1 1971 67 11 11 3 3 4 1972 59 8 16 10 5 2 1973 50 22 16 3 4 5 1974 75 5 11 3 4 1 1975 63 14 22 2 0 0 1976 57 20 6 14 4 0 1978 49 24 22 0 2 2 1979 27 40 17 8 2 6 1980 40 36 15 4 1 4 1981 49 29 10 7 1 5 1982 55 25 8 6 1 6 1983 70 16 4 4 0 6 1984 75 9 4 8 0 3 1985 48 29 4 13 5 1 �115 Table 3. Continued Year Mallard Teal 1986 60 23 6 1987 52 25 1988 62 1989 Pintail Gadwall Shoveler Redhead 9 2 2 9 8 6 0 20 0 11 5 2 57 6 9 18 9 0 1990 64 9 3 20 3 0 Means 60 17 11 7 3 2 �116 between mallard nesting density on MVNWR and the mallard breeding population surveyed in the SLV (Fig. 6). Variability in the data are probably caused by relative differences .inwater conditions between the MVNWR and the SLV. Most mallards are resident year round in the SLV, (Szymczak 1986) and are therefore partially dependent on wintering conditions in just 1 region. This same relationship would not be expected for the more migratory species because many migrate out of the SLV after the survey and do not winter in the SLV. Nevertheless, refuge nest density, heavily influenced by mallards, is only a weak indicator of the annual breeding population in the SLV. Total numbers of breeding pairs have declined in the SLV but remained relatively stable at the MVNWR (Fig. 7); particularly for mallards (Fig. 8). Thus, the refuge is important in maintaining local duck populations. There is no apparent shift to MVNWR as habitat was reduced elsewhere in the SLV, merely a reduction in numbers valley-wide. Szymczak (1986) believed that reduced habitat in the SLV was caused by the reduction of flood irrigation, which create wetlands from runoff. The relationship between North American breeding pair surveys (U.S. Fish and Wildlife Service 1989) and the SLV breeding pair survey show low correlation for either mallards, total ducks, or total species other than mallards (rZ - 0.149, 0.187, and 0.099 respectively). Thus, the SLV breeding populations evidently fluctuates relatively independent of the North American population surveyed. Water Early water, defined as water applied to wetland units before 1 April, is needed to attract nesting ducks to the MVNWR (Schroeder et al. 1976). The MVNWR staff secured increasing amounts of water for early water applications (Fig. 9). Valley conditions (as indexed by Rio Grande River water), however, are more susceptible to the natural variations in precipitation (Fig. 10). Figure 11 shows the water by source for the totals given in figure 9. Vegetation Composition at MVNWR The MVNWR, which totals 13,912 acres (not included are unit 12, or units 13 and 19 west of the Monte Vista Canal), was classified into 697 polygons for GIS analysis. Nine polygons (1,506 acres) were not surveyed (generally agricultural fields). Five additional polygons (88 acres) were county roads between units and the Empire Canal. These 14 polygons were removed from consideration. Of the remaining 683 polygons important to nesting ducks, 513 (75%) were single-feature or single-vegetation patches. The remainder were patches consisting of 2 or more species in a patch. All areas (except the 591 acres of open water and the non-surveyed areas) were considered as habitat available for nesting, because ducks are known to nest on dikes, ditch banks, and near buildings. Accordingly, total area available to nesting ducks was 11,727 acres (18.3 mi2, excluding unit 12, units 13 and 19 west of Monte Vista Canal, all non-surveyed areas and open water bodies). There were 11 types of vegetation and transition type associations (4 or more dominant species in 1 patch Table 4). Even though there were 71 combinations of these types, the majority of polygons could be delineated into single-species patches unless they were extremely heterogenous. A total of 772 acres (146 polygons, 6% of the surveyed refuge) did not contain vegetation and were classified as: �117 Table 4. Vegetation categories Common name Scientific Baltic rush Juncus GreasewoodRabbitbrush Sarcobatus vermiculatus, Chrysothamnus and acreage name balticus on the MVNWR. a,b Acreage % of available 5,967 48.4 3,577 29.0 595 4.8 578 4.6b spp. Saltgrass Distichlis stricta Grasses Several Cattail Typha latifolia 424 3.4b Spikerush Eleocharis macrosachya 149 1. 2b Bulrush Scirpus 59 0.4b Sedge Carex spp. 39 0.3b Weeds Several spp. 36 0.3 Transition 4+ spp/patch 44 0.3 spp. validus Foxtail Hordeum jubatum 3 0.0 barley arable does not show water, roads, and park~ng areas. Values are rounded the nearest acre. bExtremely heterogenous patches cause less dominant species to be under represented compared to dominant species (mainly bullrush and cattail). to Nest Success Nest success, defined as the percentage of nests that hatch at least 1 egg, has fluctuated widely at MVNWR over time but has always been relatively high (~ - 49.4%, SE - 9.5, n - 26) for all years averaged over all species. The lowest recorded success was 27% in 1978 (following the drought years); the highest was 72% in 1966. The mid- and late-1960's saw widespread use of strychnine as a means of predator control at MVNWR (a yearly success table is given later when success is compared with management treatments). Causes of nest failure were: Cause of failure Percent Range (%) lPredation 25.8 9.2 to 43.6 Flooding 4.9 0.0 to 18.0 Desertion 10.6 3.3 to 27.0 Unknown 10.8 0.0 to 32.0 �us Table 5. Availability and nest site selection for all duck nests recorded at MVNWR,. 1964-1990. Values are heavily influenced by mallards. Non-surveyed areas excluded. Area of Total when Cover 11 when II when II when available sole combined dominant and subdominate habitat species but the (%) (%) dominant Baltic-Rush Greasewood Saltgrass Cattail Grass spp. Weeds Spikerush Sedge Unidentified All other 48.4 29.0 4.8 3.4 4.6 0.3 l.2 0.3 2,582 295 31 84 78 74 57 51 62 226 324 103 10 68 55 6 6 0 0.3 2,808 619 134 94 146 129 63 57 62 44 (67.6) (14.9) (3.2) (2.3) (3.5) (3.1) (l.5) (l.4) (l.4) (l.1) Table 6. Species use of nest cover at MVNWR 1964-1990a. percent of nests found in each major cover. 341 17 128 74 48 37 74 60 10 Values represent Mallard Gadwall Teal Species Pintail 66.3 59.8 73.9 58.2 15.6 19.5 6.9 24.2 3.1 0.0 l.2 0.6 Sa1tgrass Grasses 2.3 l.6 4.8 6.7 3.1 4.8 6.6 2.8 8.6 Spikerush Weeds l.5 0.4 * 3.2 * l.1 0.9 2.4 3.5 6.0 * 0.7 * 3.6 2.6 2.4 Cover type Baltic rush Greasewood Cattail * Shoveler Redhead * 70.7 74.7 * 2.6 * * * 9.6 5.1 6.0 * 5.1 0.0 * 2.4 * * All 4.5 4.3 2.6 2.7 8.0 * 2.4 * Others 4Aster1sks represent range ot h1gh and low use spec1es tor each vegetat1on. Only primary vegetation is represented. �119 The number of nest failures from unknown causes fluctuated widely each year. This annual variability reflects partly on the nest searcher's willingness to carefully examine a destroyed nest. Because of fluctuations in "unknown" failures, variation in annual success, and small sample sizes I have not reported annual rates of nest success. Success for all species across all nests and years was 52.4% (2,179 successful nests/4,156 total) Individual species' success varied little (mallard were 54.2% successful, gadwall 63.7%, teal spp. 54.5%, pintail 57.6%, shoveler 56.0% and redhead 55% successful). Success of 520 unknown nests was lower (~- 31.9%), probably because destroyed nests could not always be identified to species. The result, however, is that individual species' success are biased upwards. When the different species have "unknown" nests reapportioned based upon overall species composition, and a 31.9% success rate applied to those "unknown" apportionments, success declines slightly (mallard 51.4%, gadwall 59.8%, teal spp. 51.7, pintail 54.4%, shoveler 52.9%, and redhead 52.5%). Chi-square tests for significant differences between species or between species and overall success showed no difference. Gadwall, the species with the largest difference in nest success, had a X2 P value of 0.082. Management activities Grazing. - Grazing by domestic cattle was used to manipulate nesting habitat at MVNWR throughout the study period under three grazing schedules. The objective of the grazing program was to open dense vegetation mats and increase stand vigor (Melvin T. Nail, Pers. Comm.). First, from 1964-1976, dormant season grazing was allowed after the nesting season (usually after September 15) through early winter (usually before December 31), with only minor variability in these dates. Almost all units were grazed annually. Second, from 1976-1987, dormant season grazing was continued but individual units were placed on a 3-year, rest-rotation cycle. There was a reduction in total ADM during the rest rotation period, however, there was an increase in grazing intensity on average over the area actually grazed (Fig. 12) in a particular year. A third period began in 1988 with most units grazed as described for 1976-1987, but others were changed to short-term, high-intensity, growingseason grazing requiring units be subdivided into smaller cells. Short-term, high-intensity grazing has a variety of objectives; primarily this program is used at MVNWR for weed control and to reduce residual cover. This program was expanded and by 1990 included all or a portion of 7 units. Appendix G reports grazing intensity (ADM/acre) for each year and unit, however grazing density is difficult to access in the multi-cell units because only portions were grazed and cattle rotated; those values are omitted in the Appendix. This third period was in a state of constant change during 1988-1990 as the new grazing program was expanding. During 1964-1976, an evaluation of grazing effects on nest densities was impossible because the units very rarely went ungrazed (most were always grazed), thus no control exists for comparison. The second rest-rotation period, however, allows for a critical evaluation of nest density response immediately after, and 2 nesting seasons following grazing. The rest-rotation schedule (1976-1987 and partially continued to 1990) was used on 19 units included in the transect nest survey (Units 5, 12, 13, 23, 24 and year 1976 are excluded because they were not grazed, partially surveyed for nests but grazed completely, or data were unavailable). No nest survey was done in 1977 �I-' N 7,000 o 1----------------------- 6,000 1 5,000 :E ::> -c 4,000 3,000 2,000 1,000 o 64 66 68 70 72 74 76 78 Year Fig. 12. Annual grazing totals (AUM) MVNWR 1964-1990. 80 82 84. 86 88 90 �121 thus nest densities are unavailable, and units changed to the third grazing schedule are also omitted from the analysis after they were managed under the short-term, high-intensity grazing system. It was known by the mid-1960's that the number of duck nests varied substantially among units and years. Thus for the analysis, units and years were treated as class variables and estimated nest density (D) was regressed on grazing intensity (I) recorded as AUM/acre, and sub-period (5). Sub-period denoted either the 1st, 2nd, or 3rd nesting season following a graze cycle in the 3-year rest-rotation grazing schedule (i.e., the values of the variable sub-period were 1, 2, or 3). The analysis of the effect of grazing on the density of duck nests was based on the following regression model: E(D~)- b~ + b1(i) + b~(j) where: + bl(1) + bi(5~) + €~ i-year - 1977-1990 (grazing occurred in 1977, but no nesting survey was conducted that year). j - unit 1,2,3,4,6,7,8,9,10,11,14,15,16,17,18,19,20,21,22. bo - intercept b; and b4 - partial regression coefficients. b~ - vector of partial regression coefficient for the 14 units. b~ - vector of partial regression coefficient for the 19 years. €i1 - error terms, assumed to have mean of zero, independent and constant variance. Grazing intensity (Appendix H) was a continuous independent variable. Clearly, if grazing increased nest density the following year, then b3 would be positive; however, if grazing intensity decreased nest density thefollowing year, then b3 would be negative. The null hypothesis of no effect of grazing is b3 - O. Our ~ priori belief was that grazing intensity tends to decrease nest density and we treated the test of the null hypothesis as onesided. We explored a log-transformation of D, but found no improvement and, since it made interpretation more difficult, these results are not presented. Sub-period was treated as a continuous independent variable. The null hypothesis of interest was that b4 - 0 while the one-sided alternative hypothesis is that nest density increased each sub-period (year) following grazing (thus b4 > 0). The response of nest density to years of non-grazing was slightly nonlinear, thus 5~ was used in the model. There was significant variation in nest density across years and units (Table 7). In particular, the variation among units is very substantial. Grazing intensity significantly decreases nest density. Nest density increases each year after grazing is terminated. The regression model can be expressed for an average year and average unit by taking the mean of the partial regression coefficients (bl and b2, respectively). Then estimated density of nests per square mile (D) is expressed by the following equation: 5- 590 - 263.7(1) + 20.25(Sub2). This equation combines the average of the partial regression coefficients for years (b1) and units (b2) into the intercept term, thus bo - 590. This model, with 32 parameters, fit the data well (R2 - 0.60, residual MSE 406.5). We explored a log-transformation of D, but found no improvement and, since it made interpretation more difficult, these results are not presented. �122 Table 7. Analysis of variance results for nest density and the significance of four main effects. Source df Sum of s uares Mean s uare F value Pr > F Model 32 44,206,980 1,381,468 8.36 0.0001 year 12 6,535,762 544,646 3.30 0.0003 unit 18 36,120,180 2,006,677 12.14 0.0001 intensity 1 696,968 696,968 4.22 0.0415" Sub-period 1 854,070 854,069 5.17 0.0242" 175 28,920,182 165,258 Error 73,127,161 207 Corrected total "See text for one-tailed significance levels. The partial regression coefficient for grazing intensity (-263.7) indicates that nest density declined (t - 1.79, one-tailed P - 0.0367) as grazing intensity increased (Fig. 13). Nest density increased (t - 2.27, onetailed P - 0.01) each of the 3 years following termination of grazing. Examination of residuals from the model (Draper and Smith 1981) revealed that the 8 most extreme points (all positive) were associated with the 3 best units (units 3, 9, 18) in good years. Thus, the model failed to predict the density of nests in the best units in generally good years. We explored several squared and cubic terms in the model but found no significant improvement. We conclude that the model represents the variability in the data quite well. The regression model allows predictions of the decline in nest density caused by moderate grazing and how nest density increases each year following the termination of grazing. Expected nest densities 1-, 2- and 3-years following grazing for two grazing intensities (I - 0.5 and 1.0 ADM/acre) are: Expected nest densities (nest/mi2) by year following cessation of grazing Grazing intensity 1st year 2nd year 3rd year 0.5 478 539 640 1.0 347 407 509 It is difficult to predict the nest density for an average year and unit because no controls (units that were not grazed for 7-10 years) are available. However, without extrapolating beyond the available data, one can set I - 0 and predict nest density at the third year after grazing had been terminated, using: 5 - 590 + 20.25(32) - 772. This figure is probably too low, as the recovery in ne-stdensity probably would take >3 years. Using this estimate of 772 nests per square mile, nest �123 1000 >- -+- 900 (J) c <D 800 0 -+- 700 (J) <D Z 600 lJ <D 500 -+- 0 E 400 -+(J) w 300 1 200 0 0.5 1 1.5 Subperiod Grazing Intensity Fig. 13. Estimated duck nest density as a function of grazing intensity and the number of years following the termination of grazing (subperiod). �124 density is decreased 38% ([772 - 478]/772) the first nesting season after being grazed at 0.5 ADM/acre, and 55% the first nesting season after being grazed at 1 ADM per acre. Does the vegetation (primarily baltic rush) get too thick and matted for good nesting habitat? The extensive data at MVNWR suggest quite the opposite; ducks require heavy residual cover for nesting. Moderate, winter grazing on a 3 year, rest-rotation cycle has had a significant detrimental effect on the density of duck nests on the MVNWR. Even 3 years following the termination of grazing at 0.5 AUM/acre, the expected nest density is depressed at least 17% ([772 - 640]/772). It.is not clear if moderate grazing is beneficial if applied, for example, every 8 or 10 years. Nest density would decrease for at least 3 years following grazing, but perhaps there would be a long-term benefit in occasional grazing. Prescribed Burning. - Burning was not used as a habitat manipulation procedure until 1981. Often, prescribed burning followed grazing during years when a unit was grazed during the rest-rotation grazing regime. . Burns which were conducted in early spring prior to nesting were intended to clean a unit of residual cover. Caution should be exercised in evaluation of these data because they are confounded by grazing, were often incomplete, and because of small sample size. Ten prescribed burns have occurred at MVNWR, (some burns covered part of 2 units for a total of 14 treatments). Two more burns, 1 <25 acres, and 1 in a unrecorded unit, are not reported. Only 8 of the 14 treatments burned >50% of a unit, making an evaluation of treatment difficult (Table 8). No effort was made to evaluate treatments with <50% coverage. Evaluation of the 8 burns with >50% coverage showed that in 7 out of 8 burn treatments (88%), nesting density decreased from the previous year while during 5 of those 8 cases, refuge-wide nest density (burn units excluded) went up. This small sample of 8 burns appears to indicate that nest density is reduced the year of a burn. MVNWR has high densities of early nesting mallards which, combined with a short growing season would support the data suggesting that burning inhibits nesting directly after a burn. One year after burn treatments, 6 of 8 units had increases in nest density and 1 remained unchanged. On a refuge-wide basis, nest densities 1 year after burning increased in only 3 of 8 cases. These data suggest that a burn treatment has a negative effect the year of the burn but may be beneficial to nesting 1 year thereafter. �125 Table 8. Prescribed burning % burned Year Unit 4 9 19 14 15 17 3 treatments on the MVNWR, 1964-1990. ab % burned Year Unit 39 lOOC 1982 1983 4 18 47 60 1984 1984 28 1983 1984 6 100 100 1989 1984 1984 17 9 1984 10 42 48 75 33 .SAcreage burned was generally bIncomplete burning. CReported unit burns usually 8 estimated resulted as: "burned unit 9", assumed 73 1989 1989 50 72 1989 1989 by refuge staff. from greasewood to be a complete vegetation not burn. Predator control. - Predator management records revealed 3 separate periods of predator control effort during 1964-1990 (Table 9). Unfortunately, qualitative measures of predator control (person-days trapping, number of traps, number of bait stations) are essentially unavailable. In addition, the effect of these programs is likewise essentially unknown or unrecorded, because exact numbers of dead avian predators, skunk etc. were not always accurately recorded. Thus, only a very qualitative index to predator control can be deduced from the data presented. All available records were used in these analyses, however total numbers are not reported here. From 1964-1970, (period 1) nest predators were poisoned and trapped. Poison control made it difficult to quantify effect because animals may wander before death. This is especially true with avian predators, which are the predominate nest predators at MVNWR. Numeric records of trapped mammals were found for this period. Period 1 was probably the most effective of the 3 time periods because of the effectiveness of poison baits. From 1971-1980, (period 2) there were few records found concerning numbers of trapped mammals. Data were found that indicated extensive aerial gunning for coyote (Canus 1atrans) during this period. Poison control ceased at the beginning of period 2, but magpie bait traps were reported to be used frequently. Magpie nest destruction was also employed, but no numbers were recorded. The 1971-1980 period had a less aggressive predator control program compared to earlier years. In the third period, 1981-1990, predator management intensified. Predator removal was quantified and data were retained on file. Mammals and avian predators were trapped and shot at every opportunity. Predator control appears related with nest success (Fig. 14). Nest .success, however, is dependent on other variables in addition to predation. When only the nests predated is expressed as a percentage of total nests, predator removal has a noticeable effect (Fig. 15). The mean percent of predated nests of all failed nests during the three periods was 16.6, 33.7 and 25.1 respectively. Nest success varied among predator management periods �t-' N 0\ 80 Posion & active trapping .: \ \ Active trapping, avain & mammal 60 End of poison, decreased trapping / / ---... period 2 -, \ -. .•....• C ID () 40 LID n, 20 o 1964 1966 1968 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 Fig. 14. Nest Success MVNWR all species 1964-90 Year �100 End of pOison, decreased tra~ // mean = 33.7% period 2 Active trapping, avain & mammal \ I~ / 80 = \ mean 25.1% • period 3 Poison & active trapping / 60 mean = 16.6% period 1 ~ c Q) o \... 40 Q) o, 20 o 1964 1966 1968 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 Year Fig. 15. Percent of failed nests due to predators, all species, MVNWR 1964-1990 •..... N '-l �128 Table 9. Summary of predator control efforts on MVNWR, 1964-1990. specific records are on file at the MVNWR Control Period 1 Period 2 Period 3 Poison bait Magpie trapping Destroy magpie nests Mammal trapping Aerial coyote gunning Overall control effort yes no no yes yes High no yes yes limited yes Low no yes some yes yes Hf.gh" aProbably not as effective as period 1 when poison was used (F - 6.89, 2 df, P - .0045) with nest success during period 1 significantly greater than during period 2. Period 3 was not significantly different from either period but had greater success than period 2. Three important factors must be considered when evaluating these data. First, there was no survey of predator levels in conjunction with predator control. Second, the "unknown" nest failures fluctuated greatly throughout the survey period. Most certainly, some predated nests were recorded as "unknown" cause of failure. For example, if all eggs were removed from the nest area by a predator and the nest bowl could not be located, then the nest would probably have been recorded as unknown resulting in a negative bias of predation rates. Third, a regression of nest success on nest density (Fig. 16) yields a positive relationship (rZ of 0.309, P - <0.01), indicating possible that nest density may increase faster than predator numbers. It is possible that the "surplus" nests are predated at a lower rate. Either scenario complicates evaluations of the effects of predator control. A more detailed examination is necessary to address this issue. Data provided by the nest survey, and the qualitative data collected on predator control (especially during the second period), are insensitive measures of this treatment effect. Haying Program. - In 1964, hay was cut on units 4, 7, 14, 20, 22, and 23 (Bill McDermith, MVNWR Pers. Comm.) The amount hayed was 14.4% of the refuge (1,665 acres). Only portions of each units were cut and specific areas actually cut within these units is not available. The program was reduced, and finally eliminated by 1977 (Fig. 17). Nest density' in units that were hayed was compared to refuge-wide nest density (hayed units removed). Mean nest density for the units hayed during the first 7 years (1964-1970) of the hay program was 257 nests/miz. During this same period, refuge-wide nest density, less hayed units, was 599 nests/mi2. Additional analyses of these same hayed units for the period 1984-1990 (a later 7-year period with no haying) show a mean density of 361 nests/mi2• Refuge wide nest density, less the 6 units, for the same period was 705 nests/mi2. Chi-square tests for difference between units hayed during active cutting vs. non-cut units for the 2 time periods are not appropriate because the density values represent means, not frequency data. However, analysis of �80 ~------------------------------------------------~ o ro U) I 0 60 . ~ 0 ~ it 0 Z 40 0 0 u UU~ 50 o 0 0 0 O~ :.._;..;.;;0 0 30 20 t 0 0 . 100 200 Nest density 300 400 500 (nests I sq.rnl.) Fig. 18. Success In relatIon to nest density, MVNWR 1964·1990. •.... N \.0 �•..... w o 2,000 '"0 r---~~-----------------___;_' 1,500 Q) iO' .c: Q) en 1,000 CO Q) 'o « 500 o 64 65 66 67 68 69. 70 71 Year 72 73 74 75 76 rt No acreage hayed data available 1965,1973-74. Fig. 17. Acreage and percent hayed on the MVNWR, 1964-1977 �131 tall dense cover represents a behavioral deterrent as well as physical barrier to crows hunting on foot, thus predator management may be partially controlled by vegetation manipulation. Schranck (1972) and Dwernychuk and Boag (1972) also noted that predation was less in dense cover. Duebbert and Kantrud (1974) determined that predator control efforts made a significant contribution in areas with high nest density. However, predator removal in areas which do not have an abundance of prime nesting cover will not result in greatly increased duck production. Predators quickly reoccupied habitats when control programs ceased. Lastly, Picozzi (1975) demonstrated that marked nests were predated at a higher rates than unmarked nests, thus predation at MVNWR may be somewhat inflated. This was not studied at MVNWR but demonstrated elsewhere (MacInnes and Misra 1972). Clearly, predator management can be effective in some circumstances. However the amount of benefit and therefore the cost-effectiveness, varies among habitats and has not been quantified at MVNWR. Kirsch (1969) stated "that grazing should be discontinued as a regular land-use practice on areas managed primarily for upland nesting ducks and upland game birds". Despite conflicting opinion to the contrary (Kantrud 1986), results of our analyses of grazing on the MVNWR support Kirsh's view of the detrimental effects of grazing. Grazing under the dormant season, 3-year rest-rotation system employed in the past had dramatic, negative effects on nesting ducks. Continued use of this grazing system is not recommended. Burning treatments at MVNWR seem to indicate a decrease in nest density the year of a burn followed by an increase the following year. Bouffard et al. (1987) discussed burn treatments and concluded they should be used sparingly, only in "small blocks", and that nesting may be affected longer than just the year of the initial burn for overwater-nesting species. Burn treatments, however, were again never experimentally controlled or in isolation from other treatments. The small sample of large-scale burn treatments also limit evaluation. The MVNWR is subject to many external factors that may affect nest density. It has been speculated that good SLV water conditions could possibly redistribute ducks from the refuge and reduce density regardless of management activities. Johnson and Grier (1988) pointed out that by virtue of their mobility, ducks have many options for selecting habitat in which to breed. They proposed 3 settling patterns which determine distribution of nesting ducks. The species evaluated at MVNWR (with the lessor exception for pintail and teal spp.) were shown to have homing-inherent settling patterns. Therefore these species may move on if nesting conditions are inadequate. Johnson and Grier also demonstrated that homing to a successful area has reproductive advantage over pioneering. This implies that changing habitat conditions which force ducks to new areas may result in lower success rates. The composition of nesting duck species fluctuated over time. The SLV breeding pair survey shows a decrease in the number of ducks valley-wide. However, the refuge counts remain more consistent over the same time period. This may suggest that duck populations and hence nesting densities in the valley are related to relative water conditions. Species composition data show general trends of an overall increase in gadwall nests but a decrease in pintail. Mallard nest numbers declined during a dry years in the mid-1970's whereas teal species increased. Neither group has changed much when viewed over the entire period. Shoveler and redhead nest densities show no trend and remain relatively constant over time, however sample sizes are low for these species (131 and 121 annually, respectively). Duck nests on MVNWR have a high success rate (52.4% for all nests �132 in high densities (26 year ~ _ 422 nests/mi2 for all units, range - 27 to 1366 nests/mi2). Higgins (1977) evaluated success in untilled and tilled agricultural land in the Dakotas. He found success rates of 25% and 17%, respectively, and reported other work with 39% and 33% nest success. He found a mean of 8.0 nests/mi2 (3.1 nests/km2) for the 2 types of habitat. Martz (1967), comparing mowed vs. unmowed lands, reported a high of 268 nests/mi2 (4.2 nests/lO acre) in the best year in the unmowed areas. Duebbert and Lokemoen (1976) studied duck nesting in undisturbed grass-legume cover in north central South Dakota and found a three-year density of 121 nest/mi2 (47 nests/km2). Duebbert and Kantrud (also from study in northcentral South Dakota) found density as high as 775 nests/mi2 (299 nests/km2) in predator-controlled idle lands. Nest success in their work ranged from 51% to 92% for agricultural areas without predator control and idle lands with predator control respectively. Kirsch (1969) reported density in grazed vs. ungrazed areas of a North Dakota nest study to be 108 and 179 nests/mi2 (0.17 and 0.28 nests/ A.), respectively. Percent hatched in that study was 14% for grazed and 28% for the ungrazed areas. The MVNWR has a history as an area on which to apply unproven management practices in an experimental framework. The refuge continues to have this experimental potential because of the transect survey system and individual management units. Future management experiments that evaluate habitat treatments should be encouraged on the MVNWR. studied) and also occur MANAGEMENT RECOMMENDATIONS Because nesting success and nest density are already very high, additional management aimed at improving either parameter will likely result in limited success. One exception is the elimination of grazing, which would result in an immediate overall increase in duck nesting densities. In addition, managers should focus on increasing nest density in low density units through habitat enhancement. Monitoring, through proper evaluation, should precede and follow any management treatment to manipulate nesting qualities or vegetation characteristics at MVNWR. There are surfacing issues facing conservation agencies dealing with maintenance of biodiversity. MVNWR has the opportunity to become a hub for the evaluation of management activities on all species. Many miles of transect have been surveyed for evaluating duck production. This system, if maintained and or expanded, could also develop baseline information for other species. Management for optimal duck production mayor may not favor other wildlife. MVNWR has both the potential and the responsibility to fully understand the benefits and/or consequences of management practices. The Colorado Statewide Waterfowl Management Plan (1988) identifies limited funds for new land purchases for conservation as a problem facing ducks in Colorado. Steps have been taken to raise additional funds, such as the new state waterfowl stamp and the MARSH program. MVNWR has the potential to serve as an example for future land development and management. The transect system at MVNWR should be included in all future management activities to aid in the evaluation of implemented goals and objectives. The quality of the data was diminished when even-numbered transects were removed. However the survey is still good and worth pursuing. With minimal efforts such as including nests at increased distances from transect centerline~ incorporating program DISTANCE, and including other variables or species, the survey could be improved upon. Better analysis of management practices on �133 vegetation should also into refuge management evaluated. Additional conducted as part of a be included. Experimental design could be incorporated planning so that management activities could be recommendations will follow through work being separate contract. LITERATURE CITED Anderson, D. R. and R. S. Pospahala. 1970. transect studies of immotile objects. 146. Correction of bias in belt J. Wildl. Manage. 34:141- Balser, D. S., Dill, H. H. and H. K. Nelson. 1968. Effect of predator reduction on waterfowl nesting success. J. Wildl. Manage. 32:669682. Borland International. 1990. Quattro Pro Version 2.0 Users Guide. Borland International, Scotts Valley CA. 800pp. Bouffard S. H., Sharp, D. E. and C.C. Evans. 1987. Overwater nesting by ducks: A review and management implications. Paper presented at 8th Great Plains wildlife damage control workshop. Rapid City, S.D., April 27-30., 1987. Burnham, K. P., D. R. Anderson and J. L. Laake. 1980. Estimation of density from line transect sampling of biological populations, Wildl. Monogr. 72. 202pp. Campbell, J. B. 1987. Introduction to remote sensing. New York, NY. 55lpp. Gilford Press. Colorado Division of Wildlife. 1989. Statewide Waterfowl Management Plan. Colo. Div. Wildl., Migratory Game Bird Program Unit. Ft. Collins. 99pp. Doty H. A. and A.J. Rondeau. 1987. Predator management to increase duck nest success. Paper presented at 8th Great Plains wildlife damage control workshop. Rapid City, S.D. April 27-30., 1987. Draper, N. R. and H. Smith. 1981. Applied Regression Analysis. 2nd. ed. John Wiley & Sons. New York. 709pp. Duebbert, H. F. and H. A. Kantrud. 1974. to land use and predator reduction. Upland duck nesting related J. Wildl. Manage. 38:257-265. _________ , J. T. Lokemoen. 1976. Duck nesting in fields of undisturbed grass-legume cover. 40:39-49. ____________ , and D.E. Sharp. 1983. Concentrated nesting of mallards and gadwalls on Miller Lake Island, North Dakota. J. Wildl. Manage. 47:729-740. Dwernychuk, L. W. and D. A. Boag. 1972. ducks nests from egg-eating birds. How vegetative cover protects J. Wild. Manage. 36:955-958. �134 Hammond, M. C. and D. H. Johnson. 1984. Effects of weather on breeding ducks in North Dakota. U.S. Fish Wildl. Serv., Tech. Rep. 1. l7pp. Higgins K. F. Dakota. 1977. Duck nesting in intensively farmed areas of North J. Wildl. Manage. 41:232-242. Johnson, D. H. and J. W. Grier. 1988. Determinants of breeding distributions of ducks. Wildl Monogr. 100:1-37. Kantrud, H. A. 1986. Effects of vegetation manipulation of breeding waterfowl in prairie wetlands-a literature review. U.S. Fish and Wildl. Serv., Fish Wildl. Tech. Rep. 3. l5pp. Kirsch, L. M. 1969. Waterfowl production in relation to grazing. J. Wildl. Manage. 33:821-828. Monte Vista National Wildlife Refuge. 1964-1990. Annual Narrative Reports and miscellaneous files. Alamosa National Wildlife Refuge Files. Alamosa, Co. Enright, C. A. 1971. An analysis of mallard nesting habitat on the Monte Vista National Wildlife refuge. M.S. Thesis, Colorado State Univ., Ft. Collins. ll3pp. Labisky, R. F. 1957. Relation of hay harvesting to duck nesting under a refuge-permittee system. J. Wi1d1. Manage. 21:194-200. MacInnes, C. D. and R. K. Misra. 1972. Predation on Canada goose nests at McConnell River, Northwest Territories. J. Wild1. Manage. 36:414-422. Martz, G. F. 1967. Effects of nesting cover removal on breeding puddle ducks. J. Wildl. Manage. 31(2):236-247. Picozzi, N. 1975. 39:151-155. Crow predation on marked nests. J. Wildl. Manage. Robinson, G. w. 1971. Vegetation and physical factors influencing waterfowl production. M.S. Thesis, Colorado State Univ., Ft. Collins. l48pp. SAS Institute Inc. 1987. SAS/STAT user's guide release 6.0. SAS Inst., Inc., Cary, NC. 1028pp. Schranck, B. w. 1972. Waterfowl nest cover and some predation relationships. J. Wildl. Manage. 36:182-186. Schroeder, L. J. 1973. Effects of invertebrate utilization on waterfowl production. M.S. Thesis, Colorado State Univ., Ft. Collins. 44pp. Schroeder, L. J. et al. 1973. Effects of early water application on waterfowl production. J. Wildl Manage. 40(2):227-232. �135 Sugden, L. G. and G. W. Beyersbergen. 1987. Effect of nesting cover density on American crow predation of simulated duck nests. J. Wildl. Manage. 51:481-485. Szymczak, M. R. 1986. Characteristics of duck populations in the intermountain parks of Colorado. Colorado Div. Wi1d1., Div. Rep. 6. 13pp. U.S. Fish and Wildlife Service, and Canadian Wildlife Service. Status of waterfowl and fall flight forecast. 39pp. 1989. U.S.D.l. Fish and Wildlife Service. 1986. Birds of the Alamosa Monte Vista National Wildlife Complex. GPO 848-805., Alamosa, Colo. 8pp. Prepared by: ~ £ ~ g;man Jame~.R Wildlife Researcher C �136 Appendix A. Individual unit summaries of acreage, and nesting importance, MVNWR for the period 1964-1990.~ Unit Tot. A. Non surveyed Open water Available acerage % of tot. avai1able (A) % of nests found (B) Nest density 0 (B)/(A) 1 291 0 34 257 2.1 2.5 569 1.165 2 136 0 10 126 1.1 1.0 495 0.946 3 901 0 0 901 7.6 12.6 1000 1.647 4 641 0 1 640 5.4 3.0 366 0.545 5 625 322 0 303 2.6 0.1 27 0.055 6 625 0 44 581 4.9 6.8 1164 1.379 7 1052 0 38 1013 8.6 8.6 330 0.999 8 394 0 27 367 3.1 4.5 704 1.438 9 887 0 2 865 7.4 19.8 1540 2.677 10 629 0 18 611 5.2 3.5 412 0.670 11 611 0 3 608 5.1 2.1 194 0.399 14 655 132 51 473 4.0 2.4 301 0.582 15 599 0 94 504 4.3 5.1 519 1.185 16 625 0 40 585 4.9 2.3 235 0.458 17 602 45 83 474 4.0 5.3 570 1.303 18 661 91 19 550 4.7 12.3 1366 2.627 19 626 0 52 573 4.9 1.8 200 0.369 20 637 186 3 448 3.8 1.1 156 0.296 21 641 0 1 640 5.5 2.3 222 0.414 22 620 31 1 589 5.0 1.7 172 0.335 23 321 210 1 110 0.9 0.2 140 0.257 24 644 66 68 510 4.3 0.9 142 0.210c aUnits 12 and 13 are not surveyed. All acreage values are rounded. bDensity value is the mean of annual densities for the unit over the period of record. See Appendix E for annual refuge density. CUnit 24 was not sampled until 1981, no. nests is biased low. �137 Appendix B. Numbers of nests found and nest chronology MVNWR 1964-1990. year 1977 is excluded. Year No. July nests removed from survey statistics 1964 0 261 41 1965 4 212 40 1966 7 260 61 1967 57 244 43 1968 2 315 42 1969 4 230 57 1970 6 229 45 1971 3 114 60 1972 1 127 52 1973 6 239 44 1974 0 109 40 1975 3 65 15 1976 0 95 50 1978 0 55 33 1979 0 53 55 1980 0 89 75 1981 0 104 47 1982 4 110 54 1983 22 114 23 1984 10 150 37 1985 0 152 55 1986 1 197 39 1987 0 262 45 1988 0 150 40 1989 0 130 53 1990 0 90 39 Average "Does No. found, May & Percent found June searches in May not reflect effort changes 160a 45.6 The �l38 Appendix C. Annual number of unidentified nests found during nest searches on MVNWR, 1964-1990 that were reapportioned to other species. Year No. unknown Percent unknown Year No. unknown Percent unknown 1964 62 24 1978 14 25 1965 19 9 1979 5 9 1966 54 21 1980 4 4 1967 31 13 1981 1 1 1968 70 22 1982 6 5 1969 60 26 1983 1 1 1970 62 27 1984 1 1 1971 16 14 1985 6 4 1972 20 16 1986 6 3 1973 35 15 1987 14 5 1974 17 16 1988 0 0 1975 6 9 1989 1 1 1976 9 9 1990 0 0 �139 Appendix Forms used by the refuge to calculate D. ... - production on the MVNWR. , --·l''O'=-,........fzf'i'!~ .••. a,.) a.', UI C_tt. ••••1. r... r... 1u.!W__ 1_ ~41_~" ...t. -- ... )1 . ·l.'.••.' ':~, c.,. ". ...,•.., • Z 0. • _If..o:''' ~.11_ ~.i.!l.(C.l., i- .---F!;"'~m' .,.~ ". ~fcl·:.,~~:·E..••••. ,n Ultn-- '" ·-m!'!.. ..+n (.j C. c_,. _. (wl.' Ib,UUL I,.C.,. :£ ".l4IIt' la (~'.~~~: lR ••••• 5) (;'1.' f-l'. toU,l_ __ tlR'.U c:a.tw.11 I . '~e••l1n CIR •• T•• 1 CI.~ I T•• I ...~. _. _. ~.d~ •• ' ._- - , I I I I - TeT~1. I ce.t (Cll. • v T•• I Tr•••• ca. 1•. , (~~~ (~I ~C~!i. ~~, (~rea r.,., - , Ul' :Ia'•••l •• ~ IU' ....•...• ,. U~I U:J (U' :Ie. I'ooc:&llltca t •••• r",!., U" DTt:~ .• ''''C.a~ ~ u•• C:.'••l, .. (C.~, '!'~'.\;.' : I I ! T.ol ~j_ i ~ !dI' I I T~It. Cas Cr..... I T•• l •••~tI"'" h.UM •••••• 4 ... ..• .,,, I ....••'c•••• ,. ....r... ,...,... ......•. ••• .._.i.. .•.. '_~l ~~~,." . t.~. ,... .. f---._. ~-1-----,-, - -_. .-.. ±= _._" . I - ..._t. ••••• 11M f.'11 a.•••• ,. c.•• IIJ:'l.1l_ l'l!:s1.lU •......• - ••••••• (U) I,. !:IlJM.. ••;;:;;Cl';'.~-':~ ·iir-rr;:~·M.· .•~ 1••••••• , ".t. (SOl .JWIl... _LI I ,,'.uac imW._ I-- .t ! I .JIet." v,..,,, ••• - I I -r- __L___L =t' I- - i i-- I. ; .1- -..~ ! _j__ . i-- -4- ._1-- • I 1- I =ti---1._" �140 Appendix E. Summary of survey data and estimated nest density for MVNWR from 1964-1990, 1977 excluded. Transect width (2w) Length of transects (L) No. of nests Estimated (n) value Estimated nest density 1964 16.5 323.6 261 0.144 306.7 1965 16.5 323.6 212 0.144 249.0 1966 16.5 323.6 260 0.144 305.4 1967 16.5 323.6 244 0.144 286.7 1968 16.5 323.6 315 0.144 370.1 1969 24 170.6 230 0.107 380.9 1970 24 170.6 229 0.107 379.3 1971 24 80.7 114 0.107 398.7 1972 24 161.6 127 0.107 222.1 1973 24 161.6 239 0.107 417.9 1974 24 161.6 109 0.107 190.6 1975 24 161.6 65 0.107 113.7 1976 24 161.6 95 0.107 166.1 1978 24 161.6 55 0.107 96.2 1979 24 161.6 53 0.107 92.7 1980 16.5 161.6 89 0.144 209.4 1981 16.5 168.8 104 0.144 234.2 1982 16.5 168.8 110 0.144 247.7 1983 16.5 168.8 114 0.144 256.7 1984 16.5 168.8 150 0.144 337.8 1985 16.5 168.8 152 0.144 342.3 1986 24 168.8 197 0.107 329.6 1987 24 168.8 262 0.107 438.4 1988 16.5 168.8 150 0.144 337.8 1989 16.5 168.8 130 0.144 292.8 1990 16.5 168.8 90 0.144 202.7 Year Average f(O) 277 .1 �141 Appendix E. Summary of survey data and estimated nest density for MVNWR from 1964-1990, 1977 excluded. Year 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 Average Transect width (2w) Length of transects (L) No. of nests (n) value 16.5 16.5 16.5 16.5 l6.5 24 24 24 24 24 24 24 24 24 24 16.5 16.5 16.5 16.5 16.5 16.5 24 24 16.5 16.5 16.5 323.6 323.6 323.6 323.6 323.6 170.6 170.6 80.7 16l.6 16l.6 16l.6 16l.6 16l.6 16l.6 16l.6 16l.6 168.8 168.8 168.8 168.8 168.8 168.8 168.8 168.8 168.8 168.8 261 212 260 244 315 230 229 ll4 127 239 109 65 95 55 53 89 104 llO ll4 150 152 197 262 150 130 90 0.144 0.144 0.144 0.144 0.144 0.107 0.107 0.107 0.107 0.107 0.107 0.107 0.107 0.107 0.107 0.144 0.144 0.144 0.144 0.144 0.144 0.107 0.107 0.144 0.144 0.144 Estimated f(O) Estimated nest density 306.7 249.0 305.4 286.7 370.1 380.9 379.3 398.7 222.1 417.9 190.6 ll3.7 166.1 96.2 92.7 209.4 234.2 247.7 256.7 337.8 342.3 329.6 438.4 337.8 292.8 202.7 277 .1 �146 AppendixH. Densityof cattle,AUM/acrefor grazedunits MVNWR 1964-l990ab• Unit Year 1 2 3 4 567 1964 0.7 1.2 0.6 0.3 0.3 0.4 1965 0.0 0.0 0.4 0.2 1966 1.2 0.8 0.7 1967 0.0 0.8 1968 0.8 1969 8 9 10 0.3 0.6 0.5 0.2 0.6 0.2 0.4 0.4 0.4 0.0 0.0 0.0 0.5 0.0 0.4 0.7 0.0 0.7 0.1 0.6 0.5 0.2 0.4 0.7 0.0 0.8 0.4 0.1 0.0 0.5 0.3 0.5 0.3 0.0 0.9 1.0 0.5 0.6 0.0 0.6 0.5 0.4 0.2 0.0 1970 0.4 0.8 0.8 0.2 0.0 0.4 0.2 0.3 0.9 0.0 1971 0.4 0.6 0.7 0.2 0.0 0.5 0.2 0.3 0.7 0.0 1972 0.6 0.5 0.6 0.6 0.0 0.3 0.2 0.2 0.6 0.8 1976 0.6 0.5 0.3 0.5 0.4 0.7 0.1 0.2 0.3 0.4 1977 0.2 0.1 0.1 0.0 0.1 0.1 0.1 0.1 0.0 0.0 1978 0.0 0.1 0.0 0.0 0.1 0.1 0.1 0.1 0.0 0.0 1979 0.8 0.0 0.0 0.0 0.0 0.4 0.0 0.0 0.5 0.0 1980 0.0 0.0 0.6 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1981 0.0 0.8 0.0 0.0 0.4 0.0 0.4 0.4 0.0 0.6 1982 1.2 0.0 0.0 0.0 0.0 0.8 0.0 0.0 0.8 0.0 1983 0.0 0.0 0.5 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1984 0.0 0.0 0.0 0.3 0.3 0.0 0.3 0.0 0.0 0.3 1985 1.6 1.0 0.0 0.0 0.0 0.5 0.0 0.5 0.8 0.0 1986 0.0 0.0 0.4 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1987 0.0 0.0 0.0 0.3 0.4 0.0 0.3 0.0 0.0 0.3 1988 1.8 0.9 0.0 0.0 0.0 0.6 0.0 0.4 0.7 0.0 1989 0.0 0.0 0.7 0.0 0.1 0.0 0.0 0.0 0.0 0.0 1990 1.1 0.0 0.4 0.8 0.2 0.3 0.0 0.0 0.2 (continued) c �147 Appendix H. (continued) Unit year 11 14 15 16 17 18 19 20 21 22 23 64 0.4 0.4 0.2 0.7 0.6 0.7 0.3 0.3 0.6 0.9 0.2 65 0.4 0.5 0.0 0.5 0.4 1.3 0.2 0.9 0.7 0.1 0.0 66 0.5 0.5 0.2 0.6 0.4 1.0 0.2 0.5 0.8 0.5 0.4 67 0.5 0.5 0.2 0.6 0.4 1.0 0.1 0.6 0.8 0.7 0.0 68 0.6 0.5 0.2 0.5 0.5 0.3 0.2 0.5 0.7 0.4 0.0 69 0.8 0.5 0.3 0.5 0.4 0.0 0.3 0.6 0.8 0.2 0.1 70 0.8 0.5 0.2 0.6 0.3 1.1 0.6 0.6 0.8 0.6 0.3 72 0.8 0.5 0.2 0.6 0.2 0.9 0.3 0.2 0.8 0.3 0.2 72 0.8 0.5 0.2 0.6 0.2 0.8 0.2 0.5 0.8 0.5 0.* 76 0.3 0.3 0.2 0.5 0.2 0.4 0.0 0.4 0.2 0.2 0.5 77 0.1 0.2 0.0 0.0 0.1 0.0 0.0 0.2 0.0 0.1 0.0 78 0.1 0.0 0.0 0.0 0.1 0.0 0.* 0.2 0.1 0.1 0.0 79 0.0 0.0 0.0 0.0 0.0 0.0 0.4 0.5 0.0 0.0 0.0 80 0.0 0.3 0.3 0.5 0.0 0.8 0.0 0.5 0.0 0.0 0.0 81 0.4 0.0 0.0 0.0 0.4 0.0 0.0 0.0 0.0 0.8 0.0 82 0.0 0.0 0.0 0.0 0.0 0.0 0.8 0.0 0.9 0.0 0.0 83 0.0 0.4 0.4 1.0 0.0 0.8 0.0 0.8 0.0 0.0 0.0 84 0.3 0.0 0.0 0.0 0.0 0.0 0.7 0.0 0.0 0.9 0.0 85 0.0 0.0 0.0 0.0 0.5 0.0 0.0 0.0 1.2 0.0 0.0 86 0.0 0.3 0.3 0.9 0.0 0.5 0.0 1.2 0.0 0.0 0.0 87 0.4 0.0 0.0 0.0 0.0 0.0 0.9 0.0 0.0 1.1 0.0 88 0.0 0.0 0.0 0.0 0.4 0.0 89 0.1 0.2 0.2 1.0 0.0 1.2 90 0.2 0.0 0.0 0.0 0.0 0.0 aNo unit hvalues as 0.* data available 1973-75. rounded to nearest 0.1, values <0.05 ADM/acre but not 0 are recorded C---Reveals units where high intensity grazing paddock arrangement and cattle were rotated. was utilized in a multi- ��149 Colorado Division of Wildlife Wildlife Research Report September 1991 JOB PROGRESS REPORT State of __~C~o~l~o~r~a~d~o~ Project Work Plan _ Avian Research - Migratory Game Birds W-152-R-4 __l__: Job 20 Job Title: Development and evaluation of moist-soil management techniques in Colorado Period Covered: Author: 1 April 1990 through 31 March 1991 James K. Ringe1man Personnel: R. Brown, D. Ferrin, S. Hoover, T. Ostertag, J. Ringe1man, M. Szymczak, Colorado Division of Wildlife. ABSTRACT Wellington Wildlife Management Area, a 6,669 ha property owned and operated by the Colorado Division of Wildlife, was selected as the site for moist-soil experiments. A levee plow and water control structures were purchased in late winter, and the proposed site for impoundment construction was surveyed for grade by a professional engineer on 26 February 1991. Anticipated water depths in each impoundment were expected to range from 40 cm at the distal end of the impoundment to near 0 cm at the proximal end. Based on this engineering survey,S impoundments, each $0.5 ha in area, were constructed in March and April using 0.5 m high levees. A filling and drawdown schedule was established on 28 May, but ditchwater was unavailable until 1 June because of wet spring weather and resultant Lnsuf'f Lc Lent; call for irrigation water. Ditchwater was again unavailable in mid-June as early summer rains continued. After initially filling all impoundments on 1 June, it became apparent that the initial engineering of impoundment grades was in error, because water only accumulated for 3-6 m behind the distal levee at maximum capacity. Consequently, impoundments were effectively unusable for experimental purposes and no experiments were performed. The impoundment site will be re-examined for grade in fall and winter 1991 and a decision will be made whether to scrape and level the impoundments to enable continuation of the experiment, or whether to abandon the study. Prepared by: James K. Ringelman Wildlife Researcher C ��151 JOB PROGRESS Colorado State of Project No. Work Plan No. Job Title: Period REPORT Covered: W-152-R-4 Avian Research 1 Job No. Analysis of Mallard 1 September, Winter Game Birds 21 Banding 1990 - 31 March, - Migratory Data Collected in Colorado 1991 Personnel: M. Wotawa, D. Anderson, K. Burnham, Colorado Cooperative Fish and "Wildlife Research Unit; B. Wunder, Colorado State University; H. Funk, D. Hopper, M. Szymczak, J. Ringelman, Colorado Division of Wildlife. ABSTRACT During the winter of 1963-64 personnel from the state of Colorado began banding mallards in Eastern Colorado to assess the feasibility of managing mallards on a flock basis. They attempted to band 1000 mallards, 250 in each age and sex class, within 7 major winterfharvest concentrations in eastern Colorado. Banding continued until the early to mid 1980' s. Additionally, banding occurred in 1 major wintering area in western Colorado from 1973-82. A total of 128,211 banding and 11,263 recovery records representing 10 management units in eastern Colorado and 2 management units in Western Colorado are available to evaluate the possibility of distinct winter populations which potentially could be managed individually. Estimated survival and recovery rates and recovery distributions will be the criterion used to distinguish populations. I present in this initial report preliminary sorting and tabulation of data in preparation for various tests of age, sex, geographic, and temporal specificity of survival and recovery rates, and recovery distributions. ��153 ANALYSIS OF MALLARD WINTER BANDING DATA COLLECTED IN COLORADO Mark A. Wotawa P. N. OBJECTIVES 1. Estimate population parameters of mallards wintering in (1) the eastern or Central Flyway portion of Colorado, and (2) the western or Pacific Flyway portion of Colorado by age and sex class, and by management unit where possible. Specific parameters include: recovery rates, survival rates, mean life span, and geographic distribution of harvest. 2. Prepare a final report upon which to base management recommendations. SEGMENT OBJECTIVES 1. Obtain banding and recovery tapes from the u.S. Fish and Wildlife Service Bird Banding Laboratory for: a. Normal, wild mallards banded postseason in eastern Colorado during the period 1963-64 through'1986-87 and recovered during the 1964-65 through 1989-90 hunting seasons. b. Normal, wild mallards banded postseason in western Colorado during the period 1973-74 through 1982-83 and recovered during the 1974-75 through 1989-90 hunting seasons. 2. In all analyses use only recoveries resulting from the above bandings, recovered anywhere, and only those shot and/or found dead during the period September 1 through January 31. 3. Employ standard computer programs to produce desired printouts of number banded, recovery tables, and geographic distribution of recoveries by age and sex class and by banding location. 4. Estimate recovery rates, survival rates, and mean life spans from computer programs using two analysis methods developed by (1) Brownie and Robson (1974), and (2) Seber (1970), and Robson and Youngs (1971). Use other analysis methods when appropriate. 5. Analyze recovery data by banding location to compare population parameters among locations and to identify any distinct population sub-units. 6. All birds were aged (adult vs. subadu1t) during all years of banding in both eastern and western Colorado. However, the accuracy of such aging is questionable because of changes in personnel beginning with the 1976-77' banding period in eastern Colorado and with the 1981-82 banding period in western Colorado. Analyze data in each area separately for the two periods to check for differences before pooling. 7. Prepare progress report. �154 BACKGROUND A new era in managing waterfowl through harvest regulations began in 1952 with the acceptance of 4 flyways as the principle management units within the United States. Although these flyways were considered administrative units, they were generally based upon different major migrational pathways of birds, especially anatids, across the U.S. This was the first major step to manage waterfowl at a population level. Subsequent divisions and refinements of flyways were imminent as states became involved in the regulation process via flyway councils and the knowledge to better define unique waterfowl populations increased. Indeed, in 1961 a special mallard management unit was established at the Columbia Basin of the Pacific Flyway, and a second, the High Plains Mallard Management Unit (HPMMU) , was established in 1972 for the western portion of the Central Flyway. In the winter of 1963-64, Colorado independently began a long term winterbanding effort to study the possibility of managing their eastern Colorado mallard population on a flock basis. The resulting data set may well be the largest for any state in the country for mallards. Winter-banding of mallards east of the Continental Divide continued uninterrupted through the winter of 1984-85, except at Bonny Reservoir where banding continued through 1986-87. Efforts targeted 1,000 bandings, 250 of each age and sex class, for each of 7 management units. A similar effort was initiated in western Colorado in the winter of 1973-74 and ended in 1982-83. Approximately 114,000 and 14,000 mallards were banded in eastern and western Colorado respectively, of which approximately 10,000 of the eastern Colorado bandings and 1,400 of the western Colorado bandings were recovered during subsequent hunting seasons. The High Plains Mallard Management Unit (HPMMU) was proposed after analysis of banding data prior to 19~6 indicated differences in population and harvest parameters for the western and eastern Central Flyway (Grieb et al. 1966; Grieb and Funk 1966). In addition, the Central Flyway states initiated a cooperative mid-winter banding study in 1965-66 to gain further knowledge of major mallard wintering areas. The results of this study provided justification for such a unit (Funk et al. 1971), and in 1972 the HPMMU was established as that portion of the Central Flyway lying to the west of the 1000 W. longitude. Differences in distribution of harvest, recovery and survival rates, and hunting pressure were the main factors supporting 2 management units. Along with the split came differential hunting regulations favoring the HPMMU, specifically, those targeting drake mallards (Funk et a1. 1968; Grieb et al. 1970). More importantly, the creation of the HPMMU further established the trend of waterfowl managers to base regulations on biologically distinct populations. Preliminary analyses of the Colorado banding data are presented in previous federal aid game research reports but are limited to only partial summaries. Additionally, Hopper et al. (1978) examined age-specificity for mallards banded in eastern Colorado through the 1976-77 hunting season. They found no difference in recovery and survival rates between sub-adults and adults for either sex. However, they did find that sub-adults of both sexes tended to wander more than adults during the first year after banding. They did not examine sex specific differences by banding area. Nichols and Hines (1987) included Colorado bandings while analyzing winter-bandings nationwide. They also found no difference between survival rates by age. However, when Colorado was broken down into 4 �155 geographic regions, they found sex-specific differences in survival and recovery rates, and recovery distributions within some regions were age specific. The last analysis of data from western Colorado was conducted in 1980 for recoveries through the 1979-80 hunting season (Hopper 1981). Western Colorado is located in the Pacific Flyway and mallards were exposed to a generally more liberal set of regulations characteristic of that flyway. Several major events important to management of mallards have occurred since the onset of Colorado's banding studies. Foremost is the drought condition the prairie-pothole region has experienced in the 80's and the subsequent decline of the continental mallard population (Johnson and Shaffer 1987). The vast majority of Central Flyway and thus Colorado's wintering mallards are produced in this region (Anderson and Henny 1972). Because of declining duck numbers and the poor breeding ground conditions, the U.S. Fish and Wildlife Service has called for a reevaluation of the HPMMU to determine if it should continue in its present form. Analyses of Colorado's banding data will contribute strongly to this reevaluation. The views of how harvest affects overall mortality of mallard populations (additive vs. compensatory) are largely contradictory (Nichols et al. 1984), but have leaned toward the compensatory hypotheses in recent years (Anderson and Burnham 1976; Nichols et al. 1984). Inherent within the compensatory hypotheses is a threshold harvest level above which harvest is no longer compensatory, but becomes additive (Anderson and Burnham 1976). Continent wide, mallard harvest is probably below this threshold, but on a local basis this may not be necessarily true. The implications of the above is that proper management may rely heavily upon the ability to more closely identify individual populations and regulate harvest below the threshold levels via carefully selected sex-specific regulations. The development of proper methods to analyze banding data has increased rapidly during the past 20 years. A summary of these methods which derive parameter estimates and their standard errors, and tests among strata (e.g. sex, age and years) are given by Brownie et al. (1985). Several studies to evaluate the failure of assumptions associated with the models of Brownie et al. (1985) have also been conducted in recent years. Anderson and Burnham (1980) examined the effect of delayed reporting of bands and Nelson et al. (1980) examined the effect of band loss on survival estimates. The assumption of homogeneous individual survival probabilities was considered by Pollock and Raveling (1982), Nichols et al. (1982), and Rexstad (1990). Extensions of banding analysis methods include the simultaneous use of live recapture data (Anderson pers. comm.) and stratification after bands are recovered (Schwarz 1988; Schwarz et al. 1988). This proj ect is the culmination of the banding effort in Colorado since the winter of 1963-64. This report presents the initial efforts of data acquisition and tabulation by banding area, age, sex, and time. I also present summaries of distribution of recoveries within North America and Colorado. METHODS Data acquisition All records were obtained from the BBL and included only normal wild �156 mallards (code 3-00) banded in Colorado during Dec. - Mar., 1963-64 through 198687, and recoveries from these bandings which were shot (code 01) or found dead (code 00) during the hunting season (Sep. - Jan.) through 1989-90. Changes in age codes were implemented by the BBL in July, 1967 resulting in the following age classifications for this analysis: Birds banded prior to 1968 with code 1 are adults and code 2 or 5 are subadults. Birds banded in 1968 and later, during December and coded 1, 5, or 6 are adults and coded 2 are subadults. Birds banded in 1968 and later during Jan. -Mar. and coded 6 are adults and coded 2 are subadults. These files were reduced to include bandings from 20 Dec. -29 Feb. to minimize the time interval over which banding occurred (Brownie et a1. 1985). Consequently, estimation of annual survival rates will refer to the approximate period of 24 January to 23 January the following year. The few bandings which occurred outside the main study areas and any recoveries obtained the same winter as they were banded were dropped from the analysis. Unit and Region Designations One objective of this analysis was to identify biologically distinct wintering populations of mallards based on available banding data. Banding data are coded by 10' latitude/longitude blocks at the BBL, but in most cases, banding did not occur sufficiently or consistently within 10' blocks to conduct meaningful analyses. Consequently, adjacent 10' blocks will initially be grouped into larger areas for analysis representing Colorado's original management units. I had to make a compromise when grouping individual 10' blocks of data between a-priori knowledge of primary mallard winter/harvest concentrations and availability and location of data. My objective was to group the data into the smallest areas possible and still maintain sufficient data to produce meaningful results. When banding began in 1963, Colorado was divided into 16 "waterfowl management units" representing major winter/harvest concentrations (see ego Grieb and Funk 1966). Szymczak (1986) used similar "harvest units" in his analysis of waterfowl banded in Colorado's mountain parks, but divided western Colorado into 5 units and grouped the 6 units of southeast Colorado. I adopted unit boundaries and their respective names from Szymczak (1986) for this analysis. The 20 unit boundaries (Fig. 1) differ slightly from the original and Szymczak's to facilitate grouping of adequate data in each unit. In many cases, preliminary testing suggested data were insufficient at the unit level. For these cases, after appropriate tests to justify pooling, some adjacent units may be combined into what I define as regions (Fig. 1). Banding and recoveries for units of southeast Colorado were especially scarce (see appendix A). Although data are summarized separately for units in the southeast, parameter estimates and hypotheses testing will most likely require these units to be pooled into 1 region. In a few instances pooling of regions may be required; this is especially true for female mallards, particularly at Bonny Reservoir and Southeast Colorado where recovery rates were low. Mallards populations in eastern and western Colorado are recognized as completely different populations. Consequently, these 2 areas were kept separate during all phases of the analysis. Recovery Distribution Distribution of recoveries are presented (as a percent) for each banding �I Ll ~~~~~;~ I soun- PLAJ"TE R. I ~ Yampa \ River' \ \ FOOTHILLS 41 Sterling- Ft. Mor~a~~' \ Julesburg Greeley-" ...., Sterling Ft. Morgan \ Southern Foothills Colorado River ESTERN COLORADO ······················r······r·································l li L.···· . BONNY RESERVOIR MOUNTAIN PARKS Gunnison River Unit 10 Unit 13 ....J Unit 12 San Luis Valley Southwest f..· ' ····································l : 38 l····· ····················r····· Unit 11 Unit 14 ----, 109 108 Figure 1. 107 Unit and region 106 designations 39 SOUTHEAST South Park !.........- ~------------T\ Unit 8 ................................................................ , \ ~ ' 40 105 used in analyses of Colorado's winter banded 37 102 103 104 r mallards. I-' V1 -....J �158 region by age, sex, and year after banding (direct and indirect recoveries). Appendix B tables recoveries by state and flyway, while appendix C tables Colorado recoveries by unit and region. One should be aware when examining the recovery distribution tables in this report that true movement patterns are confounded with variable harvest and band reporting rates among age, sex, geographic and temporal classes. Variable bag limits, season timing and lengths, hunter preference, susceptibility to hunters, and band reporting rates across the continent and among years result in recovery distributions which reflect harvest and band reporting rates rather than true movement patterns. The tables presented in this report are only the first steps in analyzing movement patterns and are presented only as data summaries. RESULTS AND DISCUSSION Data Samples Banding occurred in 51 10' latitude/longitude blocks representing 10 management units in eastern Colorado and 2 management units in western Colorado. A total of 128,211 banded mallards and 11,263 recoveries were used in this analysis (table 1). Adult and subadult males were the most frequently banded in each region, comprising 31 and 26 percent of the total respectively, while adult and subadult females comprised 20 and 22 percent. The number of bandings varied annually, but age-sex ratios within years were fairly consistent. Percent recovered was highest for males which were 2-3 times that of females. Subadults tended to have a slightly higher percentage recovered than adults in both sex classes. Bonny Reservoir and the units of southeast Colorado tended to have a lower percent recovered especially for female mallards. Recovery Distribution Recovery distributions of subadults tended to be more scattered than adults; adults banded in eastern Colorado regions had a greater percentage recovered (from 7.5 - 37.5) in the High Plains Unit (tables 12-13). Greater differences occurred for direct recoveries and for those banded at Bonny Reservoir and Southeast Colorado. The difference was generally divided among the Low Plains Unit, Canada, and the Pacific and Mississippi Flyways. The higher proportion of subadults recovered in Canada may reflect differences in nesting chronology and timing of migration, although no obvious latitudinal patterns emerged between adults and subadults within the High Plains Unit. The greater proportions recovered elsewhere most likely reflect a true difference in movement patterns. The apparently low proportion of recoveries for mallards banded at Bonny Reservoir and Southeast Colorado (tables 2-3) may be misleading. These regions exhibited the lowest recoverY rates in Colorado which are lower than that of the Low Plains portion of the Central Flyway and the Mississippi Flyway. Thus the low recovery proportions in these areas may reflect differences in harvest patterns rather than true differences in movement. The distribution of Colorado recoveries (tables 4-5) indicate high fidelity to the winter location of banding. The South Platte and Foothills had the most exchange between any 2 regions. There were few recoveries south of their respective banding location for those recovered in Colorado indicating little �159 tendency for mallards to migrate further than where they spent their first winter. F1JTURERESEARCH During the next segment, analyses will include evaluation and parameter estimates of age, sex, geographic, and temporal variation of survival and recovery rates. Additionally, more objective methods to evaluate differential recovery distributions by age, sex, banding area, banding year, and year after banding will be implemented. LITERATURE CITED Anderson, D. R. and C. J. Henny. 1972. Population ecology of the mallard: I. A review of previous studies and the distribution and migration from breeding areas. Resour. Publ. 105. 94 pp. Anderson, D. R. and K. P. Burnham. 1976. Population ecology of the mallard: VI. The effect of exploitation on survival. Resour. Publ. 128. U.S. Fish and Wildl. Servo 66 pp. Anderson, D. R. and K. P. Burnham. 1980. Effect of delayed reporting of band recoveries on survival estimates. J. Field Ornithol. 51:244-247. Brownie, C., D. R. Anderson, K. P. Burnham, and D. S. Robson. 1985. Statistical inference from band recovery data: a handbook. 2nd ed. U.S. Fish Wildl. Servo Resour. Publ. 156. 216 pp. Brownie, C. and D. S. Robson. 1974. Models allowing for age-dependent survival rates for band-return data. Cornell Univ. Biometrics Unit. (mimeo). Funk, H. D., J. R. Grieb, G. F. Wrakestraw, and D. Witt. 1968. A proposed mallard. drake season in Central Flyway Montana, Wyoming, and Colorado. Central Flyway Technical Committee Report. 12 pp. Funk, H. D., J. R. Grieb, D. Witt, G. F. Wrakestraw, G. W. Merrill, T. Kuck, D. Timm, T. Logan, and C. D. Stutzenbaker. 1971. Justification of the Central Flyway High Plains Mallard Management Uni t. Central Flyway Technical Committee Report. 48 pp. Grieb, J. R. and H. D. Funk. 1966. Analysis of mallard recoveries from birds banded in eastern Colorado prior to 1962. Colo. Div. Wildl. Game Res. Rept., Oct., 109-126. Grieb, J. R., H. Funk, D. Witt, G. Wrakestraw, andL. Serdiuk. 1966. Aproposed Mallard Management Unit for the Central Flyway. Central Flyway Technical Committee Report. 32 pp. Grieb, J. R., H. D. Funk, R. M. Hopper, G. W. Wrakestraw, and D. Witt. 1970. Evaluation of the 1968-69 experimental mallard drake season in Montana, Wyoming and Colorado. 22 pp. �160 Hopper, R. M., H. D. Funk, and D. R. Anderson. 1978. Age specificity in mallards banded postseason in eastern Colorado. J. Wildl. Manage.42:263270. Hopper, R. M. 1981 Population characteristics of mallards wintering in westcentral Colorado. Colo. Div. of Wildl. Wildl. Res. Rept., Oct. p.37-50. Johnson J. H. and T. L. Shaffer. 1987. Are mallards declining in North America. Wildl. Soc. Bull. 15:340-345. Nelson, L. J., D. R. Anderson, and K. P. Burnham. 1980. The effect of band loss on estimates of annual survival. J. Field Ornitho1. 51:30-38. Nichols, J. D. and J. E. Hines. 1987. Population ecology of the mallard VIII: Winter distribution patterns and survival rates of winter-banded mallards. u.S. Fish Wildl., Resour. Publ. 162. 154 pp. Nichols, J. D., S. L. Stokes, J. E. Hines, and M. J. Conroy. 1982. Additional comments on the assumptions of homogeneous survival rates in modern bird banding estimation models. J. Wildl. Manage. 46:953-962. Nichols, J. D., M. J. Conroy, D. R. Anderson, and K. P. Burnham. 1984. Compensatory mortality in waterfowl populations: a review of the evidence and implications for research and management. Trans. N. Amer. Wildl. and Natur. Reour. Conf. 49:535-554. Pollock, K. H. and D. G. Raveling. 1982. Assumptions of modern band-recovery models, with emphasis on heterogeneous survival rates. J. Wildl. Manage. 46:88-98. Rexstad, E. A. 1990. Models of exploitation and heterogeneity in survival probabilities of banded waterfowl populations. PhD Thesis, Colorado St. Uni., Ft. Collins, CO. Robson, D. S., and W. D. Youngs. 1971. Statistical analysis of reported tagrecapture in the harvest from an exploited population. Cornell Univ. Biometrics Unit Mimeo. Series, Paper No. BU-369-M. Schwarz, C. J. 1988. Post Release Stratification and migration models in bandrecovery and capture-recapture models. PhD. Dissertation. U. Manitoba, Winnipeg, Manitoba, Canada. 384 pp. 1988 Post-release Schwarz, C. J., K. P. Burnham, and A. N. Arnason. stratification in band-recovery models. Biometrics. 44:765-785. Seber, G. A. F. 1970. Estimating time-specific survival and reporting rates for adult birds from band returns. Biometricka. 57(2):313-318. Szymczak, M. R. 1986. Characteristics of duck populations in the intermountain parks of Colorado. CO Div. Wildl. Tech. Publ. 35. 88 pp. Prepared by: ~C)~celJajL-0t.-~ Howard D. Funk Wildlife Program Specialist �161 Table A-1. Colorado. Recovery matrix for male mallards banded dur i ng winter within the Sterling-Julesburg management unit, Adults Year Recovered Nunber Year Banded 64 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 19n 1978 1979 1980 1981 1982 1983 60 288 249 286 192 97 0 161 183 240 251 308 234 151 158 126 119 124 125 125 143 65 0 66 67 68 69 70 71 72 73 74 75 76 rr 78 79 80 81 82 83 84 85 86 87 88 89 0 6 5 0 8 8 14 0 9 7 5 5 0 4 5 5 8 3 2 1 0 7 6 3 0 0 4 1 3 2 3 0 11 0 0 2 6 4 7 0 11 13 0 0 2 2 1 0 0 2 9 15 0 0 0 1 1 0 0 1 1 4 6 0 0 1 1 1 0 0 2 4 5 11 16 0 0 0 0 0 1 0 2 1 7 3 7 8 0 0 0 0 1 0 0 3 0 1 4 4 2 3 0 0 1 0 0 0 0 0 1 5 2 2 6 3 5 0 0 0 0 0 0 0 0 0 1 5 1 3 1 4 5 0 0 0 0 0 0 0 0 2 0 2 2 3 0 3 1 1 0 0 0 0 0 0 0 0 0 0 0 3 4 0 3 2 2 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 3 1 2 0 0 0 0 0 0 0 0 0 0 1 0 3 0 1 1 3 2 5 4 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 2 0 3 3 1 7 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 1 1 7 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 2 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 Subadults Year Recovered Nunber Year Banded 64 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 19n 1978 1979 1980 1981 1982 1983 81 82 274 328 201 114 0 143 175 207 257 299 222 210 162 127 121 125 124 125 146 5 65 66 67 68 69 70 71 72 1 3 1 0 8 1 2 5 8 2 2 7 6 14 3 2 4 9 2 3 0 0 4 7 3 8 0 1 1 2 6 0 1 0 3 1 0 3 3 2 1 0 6 11 74 75 76 rr 78 79 80 81 82 83 84 85 86 87 88 89 0 0 0 0 . 1 3 2 2 3 1 1 0 0 0 4 0 9 2 7 4 5 0 0 0 0 0 1 0 2 6 5 12 11 0 0 1 0 0 0 0 1 1 1 3 10 3 0 0 1 2 0 0 0 0 1 0 4 8 4 7 0 0 0 0 0 0 0 4 0 3 3 2 5 3 7 0 0 0 0 0 0 0 0 1 2 5 1 5 4 0 3 0 0 0 0 0 0 0 0 0 0 1 2 2 3 0 2 4 0 0 0 0 0 0 0 0 0 3 0 1 3 3 4 5 1 5 0 0 0 0 0 0 0 0 0 0 1 4 1 0 0 2 4 6 4 0 0 0 0 0 0 0 0 0 0 1 0 1 0 2 2 5 5 2 8 0 0 0 0 0 0 0 0 0 0 0 1 0 1 0 0 1 0 1 4 4 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 2 0 0 4 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 2 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 73 1 1 1 �162 Table A-2_ Colorado. Recovery matrix for male mallards banded during winter within the Ft. Morgan-Sterling management unit, Adults Year Recovered Number Year Banded 64 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 365 777 329 437 407 207 141 303 266 247 206 247 250 150 229 125 125 128 125126 137 128 8 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 7 12 8 12 6 7 16 7 13 4 15 3 5 19 3 13 3 8 11 10 5 6 6 13 20 14 3 5 3 3 6 7 7 5 14 0 6 1 6 10 1 8 17 14 0 1 0 1 3 3 2 9 14 8 0 0 0 0 0 0 1 3 3 3 1 0 0 0 1 3 0 1 2 5 4 7 6 0 0 0 0 0 0 1 1 1 3 4 7 8 0 0 0 0 0 0 0 2 0 3 1 4 5 4 0 0 0 0 0 0 0 0 0 1 3 5 11 3 5 0 0 1 0 1 0 0 0 0 2 1 0 0 6 4 4 0 0 0 0 0 0 0 1 1 1 0 1 1 1 6 2 1 0 0 0 0 0 0 0 0 0 0 0 1 0 1 1 2 2 3 0 0 0 0 0 0 0 0 0 0 1 0 2 1 2 2 3 7 4 0 0 0 0 0 0 0 0 0 0 0 0 1 3 2 2 1 1 3 9 0 0 0 0 0 0 0 0 0 0 1 0 0 0 1 2 2 3 2 2 3 0 0 0 0 0 0 0 1 0 0 0 0 0 0 1 0 0 1 0 1 6 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 2 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 1 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 Subadults Year Recovered Number Year Banded 64 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 227 209 410 283 314 238 148 318 272 295 205 289 242 196 147 117 166 129 127 127 132 74 7 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 5 4 5 1 7 2 5 11 10 5 1 10 7 17 2 2 11 3 10 16 1 3 5 8 8 7 6 2 1 7 3 4 9 3 10 1 2 4 0 8 6 3 17 14 1 3 3 3 0 2 2 9 10 9 0 0 2 1 0 1 0 2 6 8 7 0 0 0 0 1 1 0 4 5 10 6 10 0 0 0 0 2 0 1 1 2 4 7 5 7 0 0 1 1 1 2 2 0 1 3 4 5 9 8 0 0 0 0 0 1 1 1 2 1 2 1 4 7 3 0 0 0 0 0 1 0 0 1 1 4 4 7 0 4 2 0 0 0 0 0 0 0 0 1 1 1 3 2 0 1 0 1 0 0 0 0 0 0 0 0 0 0 0 1 2 0 2 2 7 7 0 0 0 0 0 0 0 0 0 1 0 1 0 2 1 0 1 3 5 0 0 0 0 0 0 0 0 0 3 0 1 1 0 1 1 4 3 6 7 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 0 3 2 1 1 0 0 0 0 0 0 0 0 0 0 1 0 1 1 0 0 3 1 1 3 5 5 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 1 4 1 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 �163 Table A-3. Colorado. Recovery matrix for male mallards banded during winter within the Greele~-Ft. Morgan management unit, Adults Year Recovered Number Year Banded 64 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 99 410 287 250 252 162 150 198 204 241 197 248 354 103 302 139 125 125 146 125 134 5 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 1 12 0 10 7 1 9 10 14 0 6 6 8 11 0 5 8 2 7 4 2 5 3 4 10 6 5 0 2 4 5 5 1 7 12 0 5 1 1 3 4 2 5 10 0 2 1 1 1 4 0 4 5 14 0 1 0 0 0 1 2 3 3 6 2 0 0 0 0 2 0 0 1 1 2 4 11 0 0 0 1 0 0 0 1 3 2 2 3 11 0 0 0 0 1 2 0 2 2 4 4 3 13 3 0 0 0 0 0 1 0 0 1 0 6 1 5 6 7 0 0 0 0 0 0 0 1 0 0 0 2 4 1 4 2 0 0 0 0 0 0 0 1 0 0 1 1 1 0 4 3 1 0 0 0 0 0 0 0 0 0 1 0 0 4 0 4 2 3 7 0 0 0 0 0 0 1 0 0 0 0 0 3 0 3 4 1 1 4 0 0 0 0 0 0 0 0 0 0 1 0 1 1 3 2 1 7 7 7 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 3 1 2 1 2 2 0 0 0 0 0 0 0 0 0 0 0 0 1 0 2 1 0 1 2 0 4 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 1 1 0 1 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 4 2 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 Subadults Year Recovered Number Year Banded 64 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 67 320 322 320 252 152 159 199 209 226 209 289 289 102 79 129 132 129 149 129 132 4 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 2 5 0 8 9 0 4 5 9 0 4 6 10 10 0 3 9 11 5 4 2 6 4 7 8 9 4 1 4 3 3 2 4 3 11 0 4 4 3 4 5 8 7 9 0 0 2 4 1 1 3 5 5 7 0 0 0 2 1 0 1 3 1 8 3 0 1 1 1 1 1 0 4 2 1 4 5 0 0 0 0 2 0 2 2 4 3 3 8 7 0 1 0 0 0 0 1 1 1 0 2 6 3 4 0 0 0 0 0 0 1 1 1 2 0 7 9 6 4 0 0 0 1 0 0 0 0 3 2 0 2 6 2 0 6 0 0 0 0 0 0 0 0 1 1 2 0 0 1 4 0 6 0 0 0 0 0 0 0 1 0 0 0 0 4 1 1 2 3 6 0 0 0 0 0 0 0 0 1 1 2 0 1 0 0 2 5 2 3 0 0 0 0 0 0 0 0 0 0 1 0 1 0 1 1 3 5 2 5 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 1 1 2 3 4 4 0 0 0 0 0 0 0 0 0 1 0 0 0 0 1 0 1 0 0 0 5 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 3 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 2 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 1 1 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 �164 Table A-4. Colorado. Recovery matrix for male mallards banded during winter within the Northern Foothills management unit, Adults Year Recovered Number Year Banded 64 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 493 472 284 290 252 150 150 160 197 201 153 245 176 213 196 185 78 127 124 129 143 160 7 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 10 12 5 10 5 7 12 12 12 12 8 8 10 5 3 4 2 7 9 5 3 6 4 4 7 7 6 2 3 2 1 9 2 2 10 1 1 1 6 2 3 5 6 12 0 1 1 3 0 0 6 3 5 7 1 0 1 2 0 0 0 2 2 2 4 0 1 0 0 2 1 1 3 3 4 6 6 0 0 0 0 0 1 1 0 0 1 3 7 0 0 0 0 0 0 0 0 0 1 2 3 6 7 6 0 0 0 0 0 0 0 0 1 1 5 9 10 9 7 0 0 1 1 0 1 0 0 0 1 0 2 4 6 5 10 0 0 0 0 0 0 0 0 1 0 0 1 0 4 3 1 3 0 0 0 0 0 0 0 0 0 0 1 2 3 3 1 4 2 8 0 0 0 0 0 0 0 0 0 0 0 1 0 0 2 1 1 0 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 3 1 2 6 3 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 1 0 1 2 3 4 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 1 1 1 2 1 3 8 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 2 1 1 1 4 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 1 2 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 5 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 1 Subadults Year Recovered Number Year Banded 64 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 333 238 299 289 245 160 152 152 226 89 158 178 149 135 214 75 76 119 135 137 130 52 9 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 10 6 11 2 11 2 4 6 13 2 7 7 5 5 5 2 5 9 7 5 2 1 9 4 12 5 5 1 2 4 5 2 5 8 7 1 1 4 1 4 8 3 3 12 0 0 1 2 5 4 3 5 9 6 1 0 0 1 0 0 0 1 2 2 3 0 0 0 0 0 1 0 0 4 0 3 9 0 0 1 0 0 0 1 1 5 1 1 3 0 0 0 0 0 1 0 2 2 1 2 2 0 1 0 1 0 0 0 0 0 0 0 2 0 0 0 0 0 0 1 0 0 0 1 1 0 0 0 0 0 0 0 0 0 0 2 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 6 0 0 0 0 0 0 0 0 0 0 0 2 0 2 2 2 2 3 2 0 0 0 0 0 0 0 0 0 0 0 2 2 0 0 0 0 0 1 0 2 1 1 5 1 0 1 6 1 2 1 0 1 1 1 5 7 8 1 0 2 0 1 1 6 6 0 0 2 0 1 1 2 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 3 3 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 3 2 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 2 0 0 6 1 1 2 3 1 2 5 2 2 1 4 7 0 2 9 �165 Table A-5. Colorado. Recovery matrix for male mallards banded dur i ng winter within the Southern Foothills management unit, Adults Year Recovered Number Year Banded 64 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 547 593 479 249 295 149 29 151 222 316 200 249 419 232 291 216 174 132 128 172 129 242 13 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 14 9 14 19 17 9 16 19 9 4 10 10 8 16 6 8 10 2 7 9 4 2 9 7 10 6 1 2 4 2 6 4 3 1 6 2 3 4 3 7 4 0 9 17 0 1 1 1 2 1 0 2 1 23 0 1 0 0 1 0 1 1 2 3 9 0 0 0 0 1 0 0 2 3 7 13 11 0 0 0 0 1 0 0 0 3 4 7 9 13 0 0 0 0 0 0 0 0 1 2 1 0 12 6 0 0 0 0 0 1 0 0 0 1 0 7 14 13 9 0 0 0 0 0 0 0 0 1 2 2 1 9 1 7 10 0 0 0 0 0 0 0 1 0 1 0 2 1 1 3 2 7 0 0 0 0 0 0 0 0 0 1 0 2 6 2 5 3 4 4 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 6 6 14 0 0 1 3 1 1 0 4 0 0 1 0 2 5 4 0 0 1 0 0 1 1 4 0 1 4 85 86 87 88 89 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 a 1 3 3 4 4 a 3 a 4 3 4 10 a a 1 a 2 4 4 10 a a a a a a a a a a a a 0 a a Subadul ts Year Recovered Number Year Banded 64 1963 1964 1065 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 196 240 300 265 305 130 25 205 232 311 228 266 418 231 256 236 173 149 127 136 126 108 9 65 66 67 68 4 6 5 1 12 3 12 4 17 4 5 5 7 14 69 70 71 72 73 74 75 76 77 78 2 1 6 6 11 16 5 3 0 2 3 5 4 3 1 9 0 2 2 4 7 5 0 3 13 1 1 1 2 1 5 2 6 6 15 0 0 0 0 0 0 0 2 8 6 7 0 0 1 0 0 a 0 a 0 0 0 0 0 0 1 0 0 a 7 2 10 2 a a 1 7 5 9 10 2 0 a 1 4 6 4 9 15 a 0 0 a a 2 1 3 1 5 11 10 a 0 1 5 2 3 12 9 10 79 80 81 82 83 84 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 2 0 6 5 1 2 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a 0 0 0 0 a 0 1 a 4 3 4 3 11 6 a 0 0 0 0 0 0 3 2 4 5 2 a 0 1 0 2 5 1 4 12 a 0 a 0 0 1 2 2 5 5 5 2 6 11 a 1 0 a 0 1 0 0 1 1 1 3 9 a a a a a a 0 a a a a 0 1 a 3 1 1 1 5 1 3 a a 0 a 0 a a 1 0 1 a a a 1 a 0 1 3 3 a a a a a a a a a a a a 1 0 0 a a a a a a a 0 a 1 1 1 2 1 0 0 0 0 0 a a 0 a a a a 0 0 0 a a a a 0 a 0 a 1 1 �166 Table A-6. Recovery matrix for male mallards banded during winter within the Bonn~ Reservoir management unit, Colorado. Adults Year Recovered Number Year Banded 64 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 86 595 331 258 248 151 153 196 256 202 263 259 250 528 716 401 363 250 519 151 133 248 154 83 2 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 1 8 1 13 8 1 7 6 2 2 3 5 12 3 0 2 6 3 6 5 1 5 3 4 7 1 2 0 2 3 4 1 2 1 2 0 3 2 1 4 1 9 4 10 0 3 1 1 1 2 0 4 3 4 0 1 0 2 1 0 1 2 3 6 11 0 0 0 2 0 0 1 2 2 2 7 11 0 0 0 0 0 1 1 2 5 2 5 5 7 0 0 0 1 1 0 1 1 2 1 2 5 9 12 0 0 0 0 0 0 1 0 3 2 2 3 1 19 12 0 0 0 0 0 0 0 0 0 0 2 2 3 4 17 10 0 0 0 0 0 0 0 1 0 0 3 1 0 1 7 5 2 0 0 0 0 0 0 0 0 0 1 0 1 3 11 7 11 6 3 0 0 0 0 0 0 0 0 0 1 0 2 0 3 10 4 4 9 22 0 0 0 0 0 0 0 0 0 0 1 0 0 4 4 5 4 10 15 4 0 0 0 0 0 1 0 0 0 0 0 0 1 2 1 0 3 t 8 3 4 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3 3 0 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 1 3 1 1 1 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 1 1 8 1 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 1 1 0 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 6 Subadults Year Recovered Number Year Banded 64 65 66 67 68 69 70 71 5 0 1 1 2 9 3 1 6 7 2 1 5 6 8 1 2 1 7 9 6 0 1 6 2 7 4 4 0 0 4 5 2 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 98 148 288 277 267 173 150 210 246 152 224 250 250 498 380 399 306 272 150 150 124 59 177 75 2 0 9 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 0 1 0 2 1 2 2 2 3 4 7 0 1 0 0 0 1 0 1 1 1 6 0 0 3 1 0 0 1 0 2 0 7 10 0 0 0 0 0 1 2 1 0 1 4 2 9 0 0 1 1 0 0 1 1 0 3 2 3 7 12 0 0 0 0 1 0 0 1 2 1 3 3 4 7 9 0 0 0 0 0 0 2 0 1 1 3 0 2 5 8 12 0 0 0 1 0 0 0 0 1 0 2 0 2 4 7 6 11 0 0 0 0 0 0 0 1 0 0 1 0 2 5 3 5 6 4 0 0 0 0 0 0 0 1 1 0 0 2 0 4 2 8 0 5 4 0 0 0 0 0 0 0 0 1 0 0 0 0 0 3 5 2 7 4 7 0 0 0 0 0 0 0 0 0 0 0 0 0 2 0 2 3 5 1 4 3 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 1 2 0 1 0 0 0 4 2 0 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 1 0 0 6 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 2 1 1 2 1 0 0 1 0 2 4 3 3 2 3 6 7 1 0 0 0 0 2 1 1 1 0 6 0 1 �167 Table A-7. Recovery matrix for male mallards banded during winter within management unit 10, Colorado. Adults Year Recovered Number Year Banded 64 1963 1964 1965 1966 1967 1968 1969 1970 1971 19n 1973 1974 1975 1976 19n 1978 1979 1980 1981 1982 1983 135 0 138 0 0 125 0 0 0 53 0 0 0 0 0 0 0 0 0 3 32 2 65 66 67 68 69 70 71 72 73 74 75 76 rr 78 79 80 81 82 83 84 85 86 87 88 89 1 0 2 0 0 1 0 2 0 2 0 3 0 0 0 0 3 0 0 3 1 0 4 0 0 3 0 1 0 0 0 0 5 0 0 1 0 5 0 0 3 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 2 0 0 0 1 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Subadu[ts Year Recovered Number Year Banded 64 1963 1964 1965 1966 1967 1968 1969 1970 1971 19n 1973 1974 1975 1976 19n 1978 1979 1980 1981 1982 1983 76 0 169 0 38 161 0 0 0 46 0 0 0 0 0 0 0 0 0 2 33 3 65 66 67 68 69 70 71 72 73 74 75 76 rr 78 79 80 81 82 83 84 85 86 87 88 89 0 0 0 0 3 0 0 2 0 1 0 4 0 1 0 0 3 0 0 3 0 0 0 0 0 6 0 0 0 2 0 0 4 0 0 0 0 0 0 1 5 0 0 0 0 0 1 0 0 2 0 0 0 2 0 0 0 0 0 1 0 0 0 3 0 0 0 2 0 0 1 0 0 0 2 0 0 0 0 1 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 �168 Table A-8. Recovery matrix for male mallards banded during winter within management Unit 11, Colorado. Adults Year Recovered NUilber Year Banded 67 1966 1967 1968 1969 1970 1971 19n 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 0 0 0 77 11 100 87 154 119 217 241 38 250 194 271 0 0 0 0 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 0 0 0 0 0 0 0 0 1 0 0 0 1 0 0 0 0 0 0 2 0 0 0 3 0 3 4 0 0 0 1 0 0 0 0 0 0 1 2 2 0 0 0 0 0 1 1 6 1 3 0 0 0 0 0 2 0 1 2 2 6 0 0 0 0 0 0 1 0 1 1 6 2 0 0 0 0 0 0 0 2 1 2 2 0 6 0 0 0 0 0 0 0 1 1 1 3 0 6 3 0 0 0 0 0 0 0 1 1 0 4 0 1 3 7 0 0 0 0 0 0 0 0 0 0 0 1 2 2 .6 0 0 0 0 0 0 0 1 0 0 0 1 0 7 4 4 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 O. 0 0 0 Subadults Year Recovered Number Year Banded 67 1966 1967 1968 1969 1970 1971 19n 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 8 0 0 60 15 68 24 47 171 84 200 23 175 113 28 0 0 1 0 68 69 70 71 rz 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 1 0 0 0 0 0 0 0 3 0 0 0 0 0 0 0 0 0 0 1 0 0 0 1 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 1 1 1 2 0 0 0 0 0 1 0 0 1 2 0 0 0 1 0 0 0 1 3 0 4 0 0 0 0 0 0 0 0 2 0 3 1 0 0 0 0 0 1 1 0 0 1 3 0 4 0 0 0 0 0 0 0 2 0 1 0 0 3 3 0 0 0 0 0 1 0 0 0 0 3 0 2 1 0 0 0 0 0 1 0 0 0 0 0 0 0 1 4 0 0 0 0 0 0 0 0 0 0 0 0 2 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 �169 Table A-9_ Recovery matrix for male mallards banded during winter within management unit 12, Colorado. Adults Year Recovered Number Year Banded 64 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 380 0 196 . 160 143 25 75 246 216 192 20 49 150 12 249 11 73 0 0 0 217 8 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 7 0 5 0 6 4 0 6 5 2 0 2 4 0 2 0 7 3 5 0 2 0 4 0 1 1 3 2 0 2 1 1 3 0 15 1 0 1 0 3 0 2 4 7 0 0 1 1 1 0 1 2 3 3 0 0 0 1 1 0 1 2 2 4 2 0 0 0 0 0 0 0 4 2 3 1 1 0 0 0 0 0 0 0 0 0 1 0 1 3 0 0 0 0 0 0 1 1 0 2 1 0 3 1 0 0 0 0 0 0 0 0 2 3 0 0 0 1 7 0 0 0 0 0 0 0 0 0 1 0 0 0 0 2 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 2 0 0 0 0 0 0 0 1 0 0 0 0 2 0 3 0 1 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 1 0 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 4 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 4 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Subadul ts Year Recovered Number Year Banded 64 1963 1964 1965 1966 1967 1968 1969 1970 1971 19n 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 157 0 102 152 147 30 77 186 40 154 8 17 150 7 185 4 26 0 0 0 217 3 65 66 67 68 69 70 71 2 0 3 0 0 1 0 2 4 1 0 0 3 3 1 0 2 1 6 0 1 0 0 2 1 0 3 0 0 1 1 6 2 2 4 rz 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 1 0 5 0 6 0 1 7 2 0 0 0 2 0 0 2 2 1 7 0 0 1 0 1 0 0 0 0 5 0 0 0 0 0 2 0 0 2 0 2 0 0 0 0 0 0 0 0 1 0 0 2 1 0 0 0 0 0 0 1 0 1 0 2 0 0 3 0 0 0 0 0 2 0 0 0 0 2 0 0 2 0 7 0 0 0 0 0 0 0 1 0 0 0 1 2 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 2 0 1 0 0 1 0 1 0 1 0 0 0 0 0 2 0 2 0 1 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 0 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 1 0 0 0 0 0 0 10 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 5 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 4 �170 Table A-10. Recovery matrix for male mallards banded during winter within management unit 13, Colorado. Adults Year Recovered Number Year Banded 67 1966 1967 1968 1969 1970 1971 1972 1973 1974 160 124 144 150 25 179 83 124 143 2 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 6 4 2 3 7 3 4 5 4 2 2 3 2 0 0 1 4 2 0 4 1 0 1 2 0 6 8 0 0 0 3 0 2 3 3 0 1 0 1 0 5 2 1 3 1 0 0 0 0 4 0 6 1 0 0 0 0 0 2 0 2 2 0 0 0 0 0 2 1 0 0 0 0 0 0 1 0 0 1 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 0 Subadults Year Recovered Number Year Banded 67 1966 1967 1968 1969 1970 1971 1972 1973 1974 147 75 61 29 4 97 19 37 53 6 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 6 0 4 0 2 5 2 1 2 4 0 0 2 2 0 1 2 1 0 0 0 0 2 2 0 0 2 0 0 3 1 1 1 0 0 0 0 3 0 0 2 1 0 0 0 0 1 0 2 1 0 1 0 0 0 1 0 2 1 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 1 1 �171 Table A-11. Recovery matrix for male mallards banded during winter within the Gunnison River management unit, Colorado. Adults Year Recovered Number Year Banded 74 75 76 1973 1974 1975 1976 19n 1978 1979 1980 1981 1982 140 157 241 165 245 295 64 153 100 99 3 5 6 1 10 10 rr 2 4 3 4 78 79 80 81 82 83 84 85 86 87 88 89 0 2 6 2 12 0 2 1 .3 5 13 3 1 4 3 5 7 4 0 0 3 3 3 2 2 4 0 0 5 1 1 6 1 5 4 0 0 0 0 0 1 1 3 3 4 0 0 0 0 2 1 1 2 1 2 0 0 0 1 0 0 0 1 1 2 0 0 0 0 0 0 0 0 0 2 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 0 Subadults Year Recovered Number Year Banded 74 75 76 1973 1974 1975 1976 19n 1978 1979 1980 1981 1982 161 271 274 313 307 232 42 159 45 103 7 8 13 rr 78 79 80 81 82 83 84 85 86 87 88 89 1 3 1 9 5 4 10 14 10 13 7 9 1 0 3 0 2 0 6 3 8 6 16 10 2 0 0 0 2 1 1 2 11 0 1 0 2 4 2 2 5 2 0 0 0 1 2 3 2 3 3 5 0 0 0 1 0 1 0 2 1 6 0 0 0 0 1 1 0 1 1 1 0 0 0 0 0 0 0 0 0 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 �172 Table A-12. Recovery matrix for male mallards banded during winter within the Colorado Adults Year Recovered Number Year Banded 74 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 167 236 155 253 484 209 240 236 118 100 2 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 1 4 2 6 9 0 6 3 11 0 1 5 7 21 1 2 1 3 19 10 0 3 2 4 8 4 17 1 1 1 4 7 3 4 5 0 0 1 4 5 3 5 8 5 0 0 0 0 2 2 2 6 1 6 1 0 0 0 2 3 2 7 4 3 0 0 0 1 0 0 0 4 2 2 0 0 0 0 0 0 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 1 Subadults Year Recovered Number Year Banded 74 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 113 267 92 168 408 339 258 245 82 100 4 75 76 77 3 7 2 6 3 0 9 1 4 78 79 80 81 82 83 84 85 86 87 88 89 1 3 5 7 19 . 0 3 0 3 15 15 0 2 0 3 6 11 18 2 0 0 4 9 5 8 12 0 1 0 0 3 7 5 5 4 0 1 0 0 3 6 3 9 6 4 0 0 0 0 1 4 1 7 2 4 0 0 0 1 2 2 1 1 2 1 0 0 0 1 2 1 1 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 River management unit, Colorado. �173 Table A-13_ Colorado. Recovery matrix for female mallards banded during winter within the Sterling-Julesburg management unit, Adults Year Recovered . Number Year Banded 64 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 24 75 71 68 121 83 0 150 143 204 176 195 143 128 175 119 58 125 125 88 75 0 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 1 0 0 0 1 2 1 0 0 0 0 0 0 0 0 2 0 0 1 0 1 2 0 2 4 0 0 0 1 0 1 0 0 3 2 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 2 1 3 3 0 0 0 0 0 0 0 1 0 0 1 0 3 1 0 0 0 0 0 0 0 0 0 0 2 0 1 0 4 0 0 0 0 0 0 0 0 0 0 1 0 0 1 3 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 3 2 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 2 3 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 2 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 O. 0 0 0 0 0 1 3 1 3 Subadults Year Recovered Number Year Banded 64 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 47 64 221 293 264 59 0 93 191 96 262 219 127 158 104 125 107 124 125 139 111 2 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 0 3 0 0 4 0 0 4 7 0 0 1 4 5 0 0 1 2 5 1 0 0 1 4 5 0 0 0 0 1 3 1 0 0 3 1 0 1 0 2 2 0 3 5 0 0 0 2 1 0 0 3 3 2 0 0 0 0 0 0 0 0 1 2 3 0 0 0 0 0 0 0 0 0 1 2 4 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 1 1 1 2 1 5 0 0 0 0 0 0 0 0 0 0 0 1 1 2 1 0 0 0 0 0 0 0 0 0 0 1 0 0 2 1 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 1 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 3 2 3 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 0 1 1 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 2 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 �174 Table A-14. Colorado. Recovery matrix for female mallards banded during winter within the Ft. Morgan-Sterling management unit, Adults Year Recovered Number Year Banded 64 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 72 143 213 109 253 237 72 238 265 246 212 136 216 125 63 86 116 157 155 137 120 81 0 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 0 4 0 1 4 0 2 2 2 1 0 0 1 2 0 1 1 0 3 7 0 0 1 0 4 2 2 0 0 2 1 2 2 1 4 0 0 0 0 2 1 0 3 6 0 0 0 0 1 0 0 2 2 4 0 0 0 0 0 0 0 1 1 2 6 0 0 0 0 0 0 0 0 1 1 4 2 0 0 0 0 0 0 1 0 0 2 3 1 3 0 0 0 0 0 0 0 0 0 1 0 0 2 2 0 0 0 0 0 0 0 0 0 1 1 0 2 1 1 0 0 0 0 0 0 0 0 1 1 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 2 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 4 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Subadults Year Recovered Number Year Banded 64 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 91· 61 285 118 260 165 239 191 200 206 173 222 234 179 60 120 93 84 93 108 111 116 2 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 0 0 1 2 2 0 1 5 6 0 0 0 3 3 0 0 0 1 8 1 0 0 3 0 1 4 3 0 0 0 0 0 2 6 6 0 0 1 0 3 1 2 2 2 0 0 1 0 0 0 2 2 3 2 0 0 0 0 0 0 1 1 0 5 1 0 0 0 0 1 0 1 1 1 4 3 4 0 0 0 0 1 0 0 0 0 1 3 0 5 0 0 0 0 0 0 1 0 0 1 0 0 1 2 0 0 0 0 0 0 0 1 1 0 0 0 1 1 0 0 0 0 0 1 0 0 0 0 1 0 2 4 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 1 1 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 2 1 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 1 0 2 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 1 0 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 1 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 �175 Table A-15. Colorado. Recovery matrix for female mallards banded during winter within the Greeley-Ft. Morgan management unit, Adults Year Recovered Number Year Banded 64 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 19n 1978 1979 1980 1981 1982 1983 43 103 170 152 174 138 104 192 160 273 146 155 198 48 63 154 126 111 95 69 143 65 66 67 68 69 70 71 72 73 74 75 76 rr 78 79 80 81 82 83 84 85 86 87 88 89 0 2 0 2 4 1 1 2 4 0 0 0 2 3 1 0 1 4 2 2 0 1 2 1 4 2 1 0 0 1 0 1 1 1 3 0 0 0 0 2 0 1 3 6 0 0 0 0 0 1 2 2 4 3 0 0 0 0 0 0 0 1 1 3 3 0 0 0 1 0 0 0 1 0 1 2 2 0 2 0 0 0 0 0 0 0 2 0 0 4 0 0 0 0 0 0 0 0 0 1 0 0 0 1 0 0 0 0 0 0 0 1 0 0 0 1 2 2 0 0 0 0 0 0 0 0 0 0 1 0 0 1 1 0 5 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 3 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Q 0 0 0 0 0 0 0 0 1 0 0 2 1 1 2 0 Subadults Year Recovered Number Year Banded 64 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 19n 1978 1979 1980 1981 1982 1983 71 124 218 278 321 147 186 210 223 148 241 156 205 65 50 133 117 134 108 62 88 2 65 66 67 68 69 70 71 72 73 74 75 76 rr 78 79 80 81 82 83 84 85 86 87 88 89 0 1 0 1 7 2 2 1 6 1 2 6 6 7 0 1 2 3 5 3 0 0 2 1 5 3 3 0 0 0 3 1 2 2 4 0 0 0 0 1 0 0 2 6 0 0 0 0 1 0 0 3 3 2 0 0 0 0 0 0 0 0 1 1 3 0 0 0 0 0 1 0 2 3 2 1 5 0 0 0 0 0 0 1 0 3 1 0 1 5 0 0 0 0 0 0 1 0 0 0 2 2 2 1 0 0 0 0 0 0 0 0 1 0 1 1 1 1 1 0 0 0 0 0 0 1 0 0 0 0 1 1 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 1 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 2 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 2 2 0 0 0 0 0 0 0 0 0 0 0 0 1 0 1 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 1 0 0 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 �176 Table A-16. Colorado. Recovery matrix for female mallards banded during winter within the Northern Foothills management unit, Adults Year Recovered Number Year Banded 64 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 70 117 188 210 253 203 111 145 160 260 109 252 169 155 233 111 46 105 139 136 133 121 2 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 0 6 0 1 1 0 1 4 3 0 1 0 0 6 0 0 2 0 4 4 0 1 1 3 1 4 5 0 0 0 2 4 2 0 8 0 0 0 1 0 4 1 2 2 0 0 0 1 2 0 0 2 2 8 0 0 0 0 0 0 0 1 0 3 2 0 0 0 0 0 1 0 0 1 4 2 4 0 0 0 0 0 0 0 0 0 3 1 4 1 0 0 0 0 0 0 1 0 0 3 0 2 1 0 0 0 0 0 0 0 0 0 0 1 0 0 1 3 3 0 1 0 0 1 0 0 0 0 1 1 2 0 3 2 3 0 0 0 0 0 0 0 0 0 0 0 0 0 2 0 1 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 2 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 3 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Subadults Year Recovered Number Year Banded 64 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 122 156 244 224 344 87 177 136 255 83 174 149 132 124 136 86 46 76 100 98 106 65 3 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 0 1 0 0 4 1 1 3 6 0 1 1 5 7 0 1 2 3 5 1 0 0 2 1 4 1 3 0 1 1 0 0 0 2 3 0 0 0 1 0 1 1 1 12 0 0 0 0 0 0 2 1 4 3 0 0 0 0 0 1 3 0 1 2 4 0 0 0 0 0 0 1 1 1 0 2 9 0 0 0 0 0 0 0 1 2 0 2 1 2 0 0 0 0 0 0 0 0 0 0 2 1 2 5 0 0 0 0 0 0 0 0 0 0 1 1 2 2 1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 2 0 2 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 1 1 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 1 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 2 �177 Table A-17. Colorado. Recovery matrix for female mallards banded during winter within the Southern Foothills management unit, Adults Year Recovered Number Year Banded 64 1963 1964 1965 1966 1967 1968 1969 1970 1971 19n 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 143 127 129 179 180 171 3 111 133 322 204 242 305 188 231 148 200 144 107 84 113 167 4 65 66 67 68 69 70 71 rz 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 2 0 0 1 1 0 1 4 2 1 0 0 2 4 0 0 0 0 1 3 0 0 0 1 1 3 0 0 0 0 2 2 1 0 1 0 0 0 0 1 2 0 4 2 1 0 0 0 0 1 0 0 2 5 0 0 0 1 0 0 0 0 0 3 7 0 0 0 0 0 0 0 0 0 2 3 3 0 0 0 1 0 0 0 1 0 3 1 3 4 0 0 0 0 0 0 0 0 0 2 0 0 0 4 0 0 0 0 0 0 0 0 1 2 0 0 3 3 7 0 0 0 0 0 0 0 0 0 0 1 0 5 1 1 2 0 0 0 0 0 0 0 0 0 0 0 1 1 1 2 0 1 0 0 0 0 0 0 0 0 0 0 1 0 1 1 1 1 2 3 0 0 0 0 0 0 0 0 0 0 0 1 0 0 3 0 1 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 2 2 1 5 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 2 0 2 2 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 Subadults Year Recovered Number Year Banded 64 1963 1964 1965 1966 1967 1968 1969 1970 1971 19n 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 132 142 182 218 281 106 7 169 308 158 165 264 3n 196 205 246 202 145 136 107 130 182 2 65 66 67 68 69 70 71 rz 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 2 2 1 2 2 0 0 1 9 1 1 0 0 3 1 0 1 3 7 4 0 1 0 3 4 3 0 0 0 0 1 3 1 0 4 0 0 0 1 0 1 0 4 6 0 0 0 0 2 1 0 0 8 0 0 0 1 0 0 0 0 3 1 2 4 0 0 0 1 1 0 0 0 4 1 1 9 0 0 0 0 0 0 0 1 0 1 0 3 9 0 0 0 0 0 0 0 1 0 0 2 2 5 5 0 0 0 0 0 0 0 0 0 0 2 2 3 0 5 0 0 0 0 0 0 0 0 0 0 0 1 0 2 2 5 0 0 0 0 0 0 0 0 0 0 0 0 0 4 1 1 4 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 6 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 2 1 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 0 2 2 0 1 4 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 1 0 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 �178 Table A-18. Colorado. Recovery matrix for female banded mallards during winter within the Bonnl! Reservoir management unit, Adults Year Recovered Number Year Banded 64 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 24 160 162 148 186 153 155 129 246 148 235 219 148 203 348 247 428 325 178 92 197 123 74 65 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 0 1 0 2 1 1 2 0 2 0 1 1 2 3 0 3 1 1 2 0 0 0 0 0 2 2 5 0 0 0 0 0 5 2 4 0 0 0 0 3 1 1 3 5 0 0 0 0 0 1 0 1 4 2 0 0 0 0 1 0 1 0 0 1 0 0 0 0 0 0 1 0 1 1 2 1 2 0 0 0 0 0 0 0 0 0 0 0 3 2 0 0 0 0 0 0 0 0 0 0 2 1 1 3 0 0 0 0 0 0 0 0 0 0 1 1 1 2 6 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 4 0 0 0 0 0 0 0 0 0 0 0 0 1 1 2 1 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3 0 4 2 0 0 0 0 0 0 0 1 0 0 0 1 0 0 2 0 5 4 5 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 1 0 1 0 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 Subadults Year Recovered Number Year Banded 64 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 48 87 215 315 299 120 184 216 249 116 147 264 197 315 149 232 356 233 144 136 93 113 80 54 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 1 0 0 0 7 0 1 2 7 0 0 0 6 4 0 0 2 6 2 4 0 0 1 1 3 1 1 0 0 0 2 1 0 2 6 0 1 0 1 2 1 1 4 3 0 0 0 0 1 0 1 0 1 1 0 0 0 0 2 0 0 0 2 0 2 0 0 0 0 1 0 2 0 2 1 2 6 0 0 0 0 0 0 0 0 3 1 0 1 0 0 0 0 0 0 0 0 1 0 1 5 0 0 0 0 0 0 0 0 0 0 0 3 1 5 2 0 0 0 0 0 0 0 0 0 0 0 2 0 1 0 4 0 0 0 0 0 0 0 0 0 0 0 1 0 2 0 3 6 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 6 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 3 5 2 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 1 2 0 2 4 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 1 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 1 4 3 4 o . 0 0 2 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 �179 Table A-19. Recovery matrix for female mallards banded during winter within management unit 10, Colorado. Adults Year Recovered Number Year Banded 64 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 44 0 52 0 34 62 0 0 0 13 0 0 0 0 0 0 0 0 0 0 32 0 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 0 0 1 0 1 1 0 2 0 1 0 0 0 1 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 83 84 85 86 87 88 89 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 81 " 82 Subadults Year Recovered Number Year Banded 64 1963 1964 1965" 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 91 0 129 0 37 122 0 0 0 29 0 0 0 0 0 0 0 0 0 0 56 2 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 0 0 0 0 3 0 0 2 0 0 0 0 0 0 0 0 2 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 2 0 0 0 0 2 0 1 0 0 0 0 0 0 0 0 0 1 0 0 0 1 0 0 1 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 �180 Table A-20. Recovery matrix for female mallards banded during winter within management unit 11, Colorado. Adults Year Recovered Number Year Banded 64 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 19n 1978 1979 1980 1981 1982 1983 65 66 17 0 0 108 59 191 88 214 82 73 122 6 30 44 153 0 0 26 67 68 69 70 71 72 73 74 75 76 rr 78 79 80 81 82 83 84 85 86 87 88 89 0 1 0 0 0 0 0 0 0 1 0 0 0 0 1 0 0 0 0 0 2 0 0 0 0 0 0 0 1 0 0 0 1 1 0 4 0 0 0 0 0 1 0 1 1 0 0 0 0 0 0 0 1 0 1 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 1 1 1 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Subadults Year Recovered Number Year Banded 64 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 19n 1978 1979 1980 1981 1982 1983 10 0 0 139 36 145 47 83 228 98 249 25 93 157 55 0 0 10 65 66 67 68 69 70 71 72 73 74 75 76 rr 78 79 80 81 82 83 84 85 86 87 88 89 0 1 0 0 0 0 0 0 0 1 1 0 0 2 1 0 0 0 1 0 6 0 0 0 0 0 2 1 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 3 0 0 0 0 1 0 1 1 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 2 0 1 4 0 0 0 0 0 0 0 0 0 1 0 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 0 0 0 0 0 0 0 0 0 0 1 0 0 0 1 0 0 0 0 0 0 0 0 0 0 1 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 (J 0 4 �181 Table A-21. Recovery matrix for female mallards banded during winter within management unit 12, Colorado. Adults Year Recovered Number Year Banded 64 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976. 19n 1978 1979 1980 1981 1982 1983 61 0 81 138 132 16 55 188 0 65 66 67 68 69 70 71 72 73 74 75 76 rr 78 79 80 81 82 83 84 85 86 87 88 89 2 0 1 0 1 1 0 1 2 0 0 0 1 3 0 0 1 1 1 0 0 0 0 1 2 0 0 0 0 0 0 0 0 0 1 1 0 1 0 0 0 2 2 1 0 0 0 0 0 0 0 2 1 2 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 rr 96 6 12 73 3 73 4 27 0 0 0 132 Subadul ts Year Recovered Number Year Banded 64 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 19n 1978 1979 1980 1981 1982 1983 rr 0 54 103 144 31 95 152 42 102 4 11 127 10 80 3 7 0 0 0 224 65 66 67 68 69 70 71 72 73 74 75 76 rr 78 79 80 81 82 83 84 85 86 87 88 89 0 0 1 0 1 0 0 0 4 0 0 O· 0 5 0 0 0 2 3 0 0 0 0 0 0 0 2 0 0 0 0 1 0 1 2 0 0 0 0 2 0 1 0 1 0 0 0 0 0 0 0 0 1 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 �182 Table A-22. Recovery matrix for female mallards banded during winter within management unit 13, Colorado. Adults Year Recovered Number Year Banded 67 1966 1967 1968 1969 1970 1971 19n 1973 1974 115 71 82 39 7 74 27 91 56 2 68 69 70 71 72 73 0 1 1 2 1 0 0 0 0 1 0 1 0 0 0 0 0 0 0 1 0 1 0 0 0 4 1 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 0 0 0 0 0 0 0 . 0 0 0 0 0 0 0 2 0 1 0 0 0 0 0 0 0 3 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Subadults Year Recovered Number Year Banded 67 1966 1967 1968 1969 1970 1971 19n 1973 1974 163 41 35 11 2 96 4 29 48 4 68 69 70 71 rz 2 2 0 0 0 1 1 0 0 0 0 1 0 0 1 0 1 0 0 0 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 0 0 0 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 �183 Table A-23. Colorado. Recovery matrix for female mallards during banded winter within Adults Year Recovered Number Year Banded 74 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 133 63 263 62 130 134 68 14547 92 5 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 0 1 0 0 6 2 1 4 2 1 0 3 2 0 0 0 0 0 2 2 1 1 1 1 2 2 4 0 1 0 0 1 0 0 3 0 0 1 0 0 0 0 1 2 0 0 0 0 0 0 0 1 2 4 0 0 0 0 0 0 0 2 0 1 0 0 0 0 0 0 0 0 0 2 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Subadults Year Recovered Number Year Banded 74 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 162 277 287 177 239 148 98 142 71 105 3 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 1 8 1 5 2 2 3 5 1 3 3 4 6 0 0 1 1 4 3 1 0 4 2 2 4 1 1 0 1 0 2 1 2 5 1 2 1 0 1 0 0 2 2 0 0 0 0 1 0 1 0 0 1 0 0 0 0 1 0 0 0 0 1 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 1 2 1 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 '6 the Gunni son River management uni t I �184 Table A-24. Colorado. Recovery matrix for female mallards banded during winter within Adults Year Recovered Number Year Banded 74 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 141 140 98 180 301 86 135 222 99 67 3 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 2 1 2 4 3 0 1 2 9 0 0 1 2 8 0 0 1 3 1 6 0 0 0 0 4 0 4 0 0 0 1 3 0 3 4 0 1 0 0 3 1 0 3 1 0 0 0 0 0 1 0 2 1 0 0 0 0 0 0 1 1 3 1 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 O. Subadults Year Recovered Number Year Banded 74 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 82 446 90 180 259 346 294 189 101 132 2 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 1 9 2 7 2 1 6 1 3 0 3 2 4 10 1 2 0 0 5 4 0 1 0 1 3 9 7 0 1 0 1 0 4 4 2 0 0 1 0 0 1 2 3 1 0 0 0 1 3 2 1 2 0 7 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 , 0 1 3 1 4 the Colorado River management unit, �185 Table B-1. Distribution of direct recoveries (expressed as a percent) for mallards banded during winter within the Sterling-Julesburg management unit, Colorado from 1963-83. Adult Male Georgia Atlantic Flyway IIIinois Iowa Louisiana Mississippi Missouri Tennessee Mississippi Flyway Total (percent) Total Recovered Female 1.6 1.6 0.8 3.3 1.8 1.8 3.6 3.3 3.3 0.8 7.1 3.6 7.1 14.3 1.6 2.4 11.4 60.3 57.1 33.3 1.6 13.5 14.3 75.4 Total birds 0.3 0.3 1.8 Colorado Kansas Montana Nebraska North Dakota South Dakota Texas Wyoming High Plains - Central Flyway Alberta Manitoba Ontario Quebec Saskatchewan Canada and Alaska Male 0.8 0.8 0.8 3.2 3.2 0.8 Wyoming Pacific Flyway Female 0.8 0.8 Kansas Nebraska North Dakota Oklahoma South Dakota Texas Low Plains - Central Flyway Arizona Cal ifornia Colorado Idaho Montana Oregon Utah Washington Subadult 71.4 3.3 18.7 0.8 4.9 0.8 0.8 62.6 5.4 3.6 1.8 3.6 8.9 21.4 1.8 1.8 10.7 1.8 3.6 41.1 1.8 43.5 0.3 2.1 15.0 0.3 1.8 0.6 0.9 64.6 1.6 1.8 1.8 3.6 1.8 1.8 1.8 5.7 1.8 16.1 0.3 5.7 0.8 2.4 0.8 2.4 3.0 2.7 0.6 0.3 1.5 1.5 9.6 0.3 0.3 0.6 0.6 1.8 0.6 0.3 0.9 0.8 0.8 0.8 0.3 0.3 1.2 0.3 0.3 0.3 2.7 7.9 7.1 8.9 0.8 0.8 21.4 1.8 0.8 4.0 12.7 7.1 14.3 6.5 17.1 5.4 28.6 10.5 0.6 0.3 0.3 5.4 17.1 100.0 126 100.0 28 100.0 123 100.0 56 100.0 333 �186 Table B-2_ Distribution of indirect recoveries (expressed as a percent) for mallards banded during winter within the Sterling-Julesburg management unit, Colorado from 1963-83. Adult Subadult Male Female Male Female Arkansas IIIinois Iowa Kentucky louisiana Minnesota Mississippi Missouri Tennessee Mississippi Flyway 1.2 0.3 1.6 1.9 0.3 0.3 1.1 2.4 4.7 Kansas Nebraska North Dakota Oklahoma South Dakota Texas low Plains - Central Flyway 1.8 0.9 1.2 0.6 1.6 Colorado Kansas Montana Nebraska New Mexico North Dakota South Dakota Texas Wyoming High Plains - Central Flyway 55.7 0.9 0.6 14.7 1_6 1.6 0.9 0.3 0.3 1.2 5.8 0.9 0.6 0.6 3.1 77.1 Colorado Idaho Montana Oregon Utah lJashington Wyoming Pacific Flyway 0.6 0.3 0.6 Alberta British Columbia Manitoba Ontario Saskatchewan Canada and Alaska Total (percent) Total Recovered 1.1 0.3 1.1 3.1 4.7 59.4 9.4 3.1 1.6 . 73.4 0.3 3.7 1.9 1.6 0.9 0.6 0.6 2.2 7.8 45.6 0.3 1.9 16.3 0.3 0.3 0.6 1.3 1.6 68.1 3.3 1.1 3.3 2.2 6.6 33.0 3.3 17.6 3.3 3.3 4.4 64.8 Total birds 1.5 0.2 0.2 0.1 0.5 0.1 0.1 0.2 0.1 3.2 1.6 1.1 1.5 0.5 0.2 1.6 6.6 49.4 0.5 1.4 15.2 0.1 0.9 1.1 0.9 2.4 71.8 1.6 1.6 0.3 0.3 2.1 0.9 0.9 0.3 0.3 0.3 0.9 3.1 3.7 7.0 14.1 0.9 9.7 0.3 0.3 4.6 12.5 14.1 6.2 16.6 1.1 5.5 25.3 10.0 0.1 0.5 0.1 5.0 15.7 100.0 327 100.0 64 100.0 320 100.0 91 100.0 802 0.6 0.5 0.5 0.2 0.1 0.5 0.1 2.6 18.7 �187 Table B-3. Distribution of direct recoveries (expressed as a percent) for mallards banded during winter within the Ft. Morgan-Sterling management unit, Colorado from 1963-84. Adult Male Arkansas IIIinois Iowa Louisiana Mississippi Missouri Mississippi Flyway 0.6 0.6 0.6 · Subadult Female Male Female 1.7 0.6 6.0 1.7 0.6 0.6 1.9 3.4 Kansas Nebraska North Dakota Oklahoma South Dakota Texas Low Plains - Central Flyway 1.3 1.3 2.6 0.6 3.4 3.4 5.8 12.1 Colorado Kansas Montana Nebraska New Mexico North Dakota South Dakota Wyoming High Plains - Central Flyway 59.6 0.6 1.3 5.1 53.4 0.6 0.6 3.2 71.2 1.7 Arizona Cal Hornia Colorado Idaho Montana Oregon Utah Washington Pacific Flyway Alberta British Columbia Manitoba Saskatchewan Canada and Alaska Total (percent) Total Recovered · 3.4 1.7 0.6 2.5 2.5 2.5 1.2 0.6 2.5 9.2 2.0 2.0 10.0 2.0 2.0 42.9 1.2 24.0 13.5 0.6 1.2 3.1 3.1 65.6 16.0 4.0 4.0 4.0 52.0 1.3· 0.6 2.0 0.6 0.6 1.8 1.8 1.2 0.6 1.2 7.4 4.0 2.0 1.7 8.6 1.7 67.2 0.6 · 3.2 2.0 10.0 Total Birds 1.2 0.2 0.7 0.7 0.2 0.2 3.3 2.1 1.9 1.4 0.9 0.2 0.9 7.5 48.2 0.7 0.7 10.1 0.2 1.4 1.9 3.0 66.3 0.2 0.5 0.5 1.4 1.2 0.5 0.2 0.7 5.2 17.2 9.2 7.1 17.9 17.2 1.2 4.9 15.3 10.0 26.0 11.5 0.2 0.5 5.6 17.8 100.0 156 100.0 58 100.0 163 100.0 50 100.0 427 10.9 14.0 2.0 �188 Table B-4_ Distribution of indirect recoveries (expressed as a percent) for mallards banded during winter within the Ft. Morgan-Sterling management unit, Colorado from 1963-84. Adult Male Subadult Female North Carol ina Atlantic Flyway 0.2 0.2 Arkansas Illinois Iowa Louisiana Minnesota Mississippi Missouri Wisconsin Mississippi Flyway 0.8 1.2 0.4 1.2 1.2 0.2 0.2 0.2 1.8 Kansas Nebraska North Dakota Oklahoma South Dakota Texas Low Plains - Central Flyway 0.6 0.8 0.4 Colorado Kansas Montana Nebraska New Mexico North Dakota South Dakota Texas Wyoming High Plains - Central Flyway 62.8 0.6 0.8 10.1 0.6 0.6 0.4 1.2 1.2 78.4 Colorado Idaho Montana Oregon Utah Washington Wyoming Pacific Flyway Alaska Alberta British Columbia Manitoba Saskatchewan Canada and Alaska 0.6 0.8 3.2 0.2 0.6 0.8 9.5 0.2 0.4 4.4 14.5 1.1 0.2 0.2 0.5 0.2 0.5 0.5 6.4 3.7 3.2 8.3 2.4 1.2 0.7 2.3 0.7 0.7 0.5 2.3 7.0 1.8 1.8 0.9 0.9 53.6 0.2 1.1 11.3 0.5 0.2 0.5 0.7 1.4 69.4 47.7 2.4 6.1 62.2 1.2 7.3 1.2 n.o 1.2 0.9 1.6 0.7 0.9 0.9 1.8 7.3 0.9 4.6 0.9 1.8 0.9 1.8 58.7 1.8 0.9 0.9 0.9 6.1 0.5 0.5 0.2 4.3 4.6 6.1 7.9 0.9 15.6 6.1 12.2 0.5 7.7 16.0 4.6 21.1 Mexico Total (percent) Total Recovered Female 0.2 100.0 495 Total birds 0.1 0.1 1.2 3.7 0.2 1.8 Male 100.0 82 100.0 444 1.5 0.1 0.1 0.5 0.3 0.3 0.3 0.1 3.1 0.7 1.4 0.7 0.4 0.4 1.6 5.3 57.7 0.4 1.0 9.8 0.4 0.4 0.6 0.9 1.2 n.5 0.4 1.2 0.7 0.1 0.4 0.4 0.2 3.4 0.1 9.2 0.1 0.4 5.8 15_6 0.1 100.0 109 100.0 1130 �189 Table B-5. Distribution of direct recoveries (expressed as a percent) for mallards banded during winter within the Greele~-Ft. Morgan management unit, Colorado from 1963-83. Adult Male Arkansas Iowa Louisiana Mississippi Missouri Mississippi Flyway Kansas Nebraska North Dakota Oklahoma South Dakota Texas Low Plains - CentraL Flyway Colorado Kansas Montana Nebraska New Mexico South Dakota Texas Wyoming High Plains - Central Flyway Cal Hornia CoLorado Idaho Montana Oregon Utah Washington gyoming Pacific FLyway Alberta British Columbia Manitoba Ontario Saskatchewan Canada and Alaska Total (percent) Total Recovered Subadult Female Male 0.7 1.3 2.0 2.4 0.8 0.8 0.8 2.0 2.0 4.8 0.8 0.8 3.2 1.3 Female 1.6 1.6 3.2 4.8 1.6 2.0 Total birds 0.8 0.5 1.3 0.3 0.3 3.1 1.0 0.3 1.3 0.8 0.5 0.3 4.1 1.3 2.0 4.8 3.2 1.6 11.1 69.5 56.9 0.7 2.6 11.8 42.4 1.6 1.6 9.6 46.0 1.6 1.6 3.2 55.4 0.8 1.0 6.2 3.2 55.6 0.3 0.8 0.3 2.3 66.9 0.8 1.6 0.7 1.3 74.8 0.7 1.3 2.0 2.0 2.0 n.5 . 2.0 2.0 0.7 3.2 60.8 · 1.6 1.6 2.4 2.4 0.8 1.6 0.8 3.2 1.6 1.6 0.3 0.5 1.5 2.1 0.3 1.3 0.5 0.3 6.7 0.7 5.3 3.9 8.0 · 9.5 10.6 17.6 12.0 0.8 12.7 1.6 5.3 16.6 2.0 19.6 8.8 21.6 · 1.6 4.8 20.6 12.3 0.5 0.3 0.3 5.9 19.2 100.0 151 100.0 51 100.0 125 100.0 63 100.0 390 0.7 �190 Table B-6. Distribution of indirect recoveries (expressed as a percent) for mallards banded during winter within the Greele:x:-Ft.Morgan management unit, Colorado from 1963-83. Adult Male Subadult Female Male New York Atlantic Flyway Arkansas Iowa Louisiana Michigan Minnesota Mississippi Mississippi Flyway Kansas Nebraska North Dakota Oklahoma South Dakota Texas Low Plains - Central Flyway 0.5 0.5 1.1 0.3 0.3 0.5 3.3 Colorado Kansas Montana Nebraska New Mexico North Dakota Oklahoma South Dakota Texas Wyoming High Plains - Central Flyway 61.5 0.5 0.8 9.1 0.3 0.3 0.3 0.5 0.8 1.1 75.3 Arizona Cal itornia Colorado Idaho Montana Oregon Utah Washington Pacific Flyway Alaska Alberta British Columbia Manitoba Ontario Saskatchewan Canada and Alaska Total (percent) Total Recovered 0.1 0.1 0.8 0.9 0.9 2.7 0.3 0.3 2.1 4.5 0.4 0.2 0.7 0.1 0.2 0.2 1.9 1.1 0.3 0.3 1.1 0.3 0.5 0.3 1.1 1.4 3.6 . 1.1 2.2 1•. 1 1.1 1.1 5.6 56.2 4.5 9.0 0.3 1.1 1.6 0.8 0.3 1.3 5.3 56.6 0.3 0.8 8.5 0.3 1.1 2.2 73.0 1.1 1.1 1.1 3.4 . Total birds 0.9 0.9 0.8 0.3 0.3 Female 0.9 1.8 2.7 1.8 3.6 10.8 42.3 0.9 2.7 11.7 0.9 0.5 1.3 2.7 71.0 1.8 1.8 3.6 65.8 0.3 0.8 1.9 0.3 0.8 1.3 0.3 5.6 0.9 . 0.9 0.9 2.7 0.9 9.9 0.6 1.0 1.4 0.7 0.2 1.3 5.2 56.8 0.4 1.4 9.1 0.2 0.3 0.1 0.6 1.1 2.1 n.2 0.1 0.4 0.4 1.3 0.9 0.4 0.5 0.2 4.3 8.8 0.3 0.3 13.5 9.6 1.1 7.4 16.8 2.2 16.9 0.3 0.3 5.9 16.0 3.6 15.3 0.1 9.7 0.1 0.4 0.1 5.9 16.3 100.0 364 100.0 89 100.0 376 100.0 111 100.0 940 0.9 �191 Table B-7. Distribution of direct recoveries (expressed as a percent) for mallards banded during winter within the Northern Foothills management unit, Colorado from 1963-1984. Adult Male Arkansas Illinois Iowa Louisiana Minnesota Missouri Mississippi Flyway Kansas Nebraska North Dakota Oklahoma South Dakota Texas Low Plains - Central Flyway Colorado Kansas Montana Nebraska New Mexico North Dakota South Dakota Texas lJyoming High Plains - Central Flyway Arizona California Colorado Idaho Montana Utah Washington Pacific Flyway Alberta British Columbia Manitoba Saskatchewan Canada and Alaska Total (percent) Total Recovered Subadult Female Male 1.5 1.5 0.7 3.0 3.0 . 0.7 1.5 1.5 3.0 0.7 0.7 0.7 0.7 2.2 2.2 7.5 n.7 65.7 51.5 0.7 4.9 0.7 0.7 0.7 0.7 2.1 83.2 1.5 4.5 0.7 5.2 0.7 0.7 2.2 0.7 Female 3.0 74.6 3.7 64.9 1.4 2.7 40.5 1.4 1.4 9.5 66.7 2.7 1.4 33.3 5.4 62.2 100.0 4.5 9.1 . 0.7 6.0 12.7 13.5 1.4 3.5 13.3 9.0 14.9 0.7 1.5 14.9 6.8 21.6 100.0 143 100.0 67 100.0 134 100.0 74 1.5 . 0.5 0.2 0.7 0.2 1.0 1.0 3.6 1.4 1.4 3.0 Total birds 0.5 0.2 0.2 0.5 0.5 0.2 2.1 1.4 1.4 1.5 0.7 0.7 2.2 1.5 1.5 1.5 9.7 0.7 0.7 Male 0.7 0.7 0.7 0.7 0.7 1.4 Unknown 59.1 0.2 1.0 5.7 0.5 1.2 1.2 0.2 3.3 72.4 0.5 0.2 1.2 1.7 1.2 0.7 1.0 6.4 1.4 4.1 2.7 1.4 2.7 12.2 10.5 0.2 0.5 4.3 15.4 100.0 3 100.0 421 �192 Table B-8. Distribution of indirect recoveries (expressed as a percent) for mallards banded during winter within the Northern Foothills management unit, Colorado from 1963-84. Adult Male Subadult Female Male South Carolina Atlantic Flyway Arkansas Illinois Iowa louisiana Minnesota Missouri Mississippi Flyway 0.2 Kansas Nebraska North Dakota Oklahoma south Dakota Texas low Plains - Central Flyway 0.5 0.2 0.5 0.2 0.2 0.2 1.9 Colorado Montana Nebraska New Mexico North Dakota South Dakota Texas Wyoming High Plains - Central Flyway 79.8 1.0 2.6 0.2 Cal ifornia Colorado Idaho Montana Nevada Oregon Utah lJashington Wyoming Pacific Flyway Alaska Alberta Manitoba Saskatchewan Canada and Alaska Total (percent) Total Recovered 0.2 1.8 1.8 0.9 0.9 0.9 2.8 65.1 2.8 2.8 0.9 0.2 1.4 85.2 3.7 75.2 . 0.2 0.2 0.5 1.8 0.9 0.9 1.6 0.3 0.3 1.1 0.8 0.5 4.6 0.5 1.4 0.3 0.3 0.3 0.8 3.5 63.4 1.9 4.9 0.8 0.3 0.3 2.2 73.7 Unknown Female 0.1 0.1 1.0 2.1 0.7 0.4 0.3 0.4 0.3 0.2 2.3 3.1 1.0 1.0 16.7 2.1 4.2 16.7 49.0 3.1 4.2 2.1 1.0 3.1 62.5 0.5 0.8 1.4 1.6 1.0 1.0 4.2 1.0 1.0 1.0 9.4 0.9 1.9 4.6 0.3 1.1 1.1 0.3 7.0 11.9 7.0 16.7 3.7 15.6 4.1 11.1 3.1 19.8 100.0 109 100.0 369 100.0 96 100.0 420 Total birds 1.0 1.0 0.5 0.5 0.2 6.0 0.2 4.3 10.7 Male 83.3 83.3 0.5 0.7 0.6 0.2 0.2 0.7 2.9 69.2 1.7 3.6 0.5 0.1 0.4 0.1 2.1 77.7 0.3 0.5 1.2 0.9 0.1 0.2 0.7 0.7 0.2 4.8 0.1 8.0 0.1 4.0 12.2 100.0 6 100.0 1000 �193 Table B-9. Distribution of direct recoveries (expressed as a percent) for mallards banded during winter withing the Southern .Foothills management unit, Colorado from 1963-84. Adult Male Subadult Female Male Female Arkansas Louisiana Mimesota Missouri Mississippi Flyway 0.9 0.4 1.3 0.5 2.2 Kansas Nebraska North Dakota South Dakota Texas Low Plains - Central Flyway 0.4 0.4 1.8 0.4 0.5 1.5 0.5 3.4 4.5 1.1 3.1 3.3 3.3 0.5 3.0 2.2 11.2 Colorado Montana Nebraska New Mexico North Dakota South Dakota Texas Wyoming High Plains - Central Flyway 75.8 2.2 1.8 0.4 0.4 0.4 0.9 2.2 84.1 77.0 1.6 1.6 1.6 68.2 2.0 4.5 0.5 0.5 0.5 51.7 1.1 2.2 3.3 85.2 2.0 78.3 3.4 60.7 Arizona California Colorado Idaho Montana Nevada Utah Washington Wyoming Pacific Flyway Alberta British Columbia District of Mackenzie Manitoba Saskatchewan Canada and Alaska Total (percent) Total Recovered 0.5 . 1.1 1.1 1.1 1.1 0.5 0.4 1.6 1.6 1.6 0.4 0.4 . . 1.3 4.9 6.2 4.9 2.0 2.5 1.1 5.6 1.1 1.1 1.1 Total birds 0.5 0.2 0.2 0.2 1.0 0.9 1.4 1.0 0.2 0.9 4.3 69.6 1.7 2.8 0.5 0.7 0.5 0.3 2.4 78.6 0.2 0.2 1.7 1.4 0.3 0.2 0.5 0.2 0.2 4.9 0.5 0.5 0.5 6.6 10.1 6.6 0.5 9.0 1.1 1.1 4.5 15.7 6.6 0.3 0.2 0.2 3.8 11.1 100.0 89 100.0 575 . 1.6 4.0 10.1 6.6 4.5 11.6 100.0 227 100.0 61 100.0 198 �194 Table B-10. Distribution of indirect recoveries (expressed as a percent) for mallards banded during winter within the Southern Foothills management unit, Colorado from 1963-84.· Adult Male Arkansas Illinois Kentucky louisiana Minnesota Mississippi Missouri Wisconsin Mississippi Flyway Subadult Female 0.2 1.2 3.7 Kansas Nebraska North Dakota Oklahoma South Dakota Texas Low Plains - Central Flyway 0.4 0.2 1.0 0.9 Colorado Kansas Montana Nebraska New Mexico North Dakota South Dakota Texas Wyoming High Plains - Central Flyway 73.4 0.2 1.0 5.3 0.2 0.4 0.4 0.2 2.1 83.2 Total (percent) Total Recovered Male 0.8 1.9 0.9 0.9 Alaska Alberta British Columbia District of Mackenzie Manitoba Saskatchewan Canada and Alaska Female 0.2 1.8 1.9 0.9 3.7 59.3 0.9 0.9 3.7 0.9 3.7 69.4 0.5 2.4 0.2 0.2 1.4 0.8 4.0 0.2 1.1 0.7 0.8 0.8 0.7 1.4 4.1 0.8 0.8 3.2 64.3 54.0 2.0 6.1 0.5 0.2 0.2 1.4 1.4 76.0 0.8 2.4 0.2 0.4 1.2 0.8 0.2 0.2 0.4 0.6 0.2 0.2 4.1 0.5 2.0 2.0 0.9 0.9 . 6.5 0.5 1.4 0.2 4.0 6.8 11.1 13.0 0.9 7.0 0.9 0.8 12.7 0.8 3.5 9.8 2.8 16.7 0.2 3.6 11.8 5.6 19.8 100.0 512 100.0 108 100.0 442 100.0 126 6.1 100.0 100.0 66.7 0.2 1.3 5.1 0.3 0.3 0.3 0.8 2.1 77.0 0.1 0.3 1.1 1.4 1.1 0.1 0.4 0.8 0.7 0.1 6.1 . 2.4 0.9 1.9 0.9 0.9 . 0.3 0.6 0.9 0.1 0.3 0.7 2.9 0.8 0.8 3.2 61.9 Total birds 0.2 0.2 0.1 0.7 0.2 0.2 0.3 0.1 1.8 0.5 0.2 0.2 0.2 0.2 0.2 Arizona Cal ifornia Colorado Idaho Montana Nevada Oregon Utah Washington Wyoming Pacific Flyway Male Unknown 2.4 1.6 0.8 0.1 7.7 0.5 0.1 0.1 3.7 12.2 0.2 100.0 1 100.0 1189 �195 Table B-11. Distribution of direct recoveries (expressed as a percent) for mallards banded during winter within the Bonn~ Reservoir management unit, Colorado from 1963-86. Adult Male Alabama Arkansas Illinois Iowa Louisiana Minnesota Mississippi Missouri Tennessee Mississippi Flyway Kansas Nebraska North Dakota Oklahoma South Dakota Texas Low Plains - Central Flyway Colorado Kansas Montana Nebraska New Mexico North Dakota Oklahoma South Dakota Texas Wyoming High Plains - Central Flyway Subadult Female Male 1.8 1.8 0.6 4.5 0.6 1.9 1.3 0.6 0.6 0.6 0.6 1.3 2.6 1.3 1.3 1.3 2.6 9.0 3.6 0.6 11.0 3.6 3.6 1.8 3.6 9.1 21.8 6.5 6.5 1.9 1.9 3.2 3.9 23.9 54.2 3.9 3.2 6.5 30.9 5.5 1.3 3.6 7.3 1.3 0.6 71.0 47.3 California Colorado Idaho Montana Oregon Utah Pacific Flyway 0.6 Alberta British Columbia Manitoba Saskatchewan Canada and Alaska 9.7 12.7 7.7 17.4 100.0 155 Total (percent) Total Recovered 14.8 1.9 1.3 11.0 1.3 0.6 1.3 1.3 33.5 Unknown Female Male Female 0.2 2.5 0.2 0.9 2.0 0.2 0.7 0.2 0.4 7.4 3.9 7.9 1.3 1.3 1.3 15.8 7.9 6.6 5.3 1.3 33.3 . 7.9 28.9 33.3 66.7 14.5 50.0 2.6 50.0 100.0 0.2 0.2 0.9 1.3 0.4 0.2 3.4 1.3 2.6 1.9 0.6 0.6 5.8 0.6 1.3 4.5 4.5 2.5 1.6 1.6 4.9 19.5 30.5 2.7 1.6 7.6 0.4 1.6 0.2 1.1 0.9 0.4 47.1 3.9 1.3 2.6 1.3 26.3 Total birds 2.6 1.3 5.3 13.2 33.3 1.8 12.7 27.3 13.5 0.6 0.6 11.0 25.8 1.3 9.2 23.7 33.3 12.1 0.2 0.7 9.6 22.6 100.0 55 100.0 155 100.0 76 100.0 3 100.0 446 100.0 2 �196 Table B-12. Distribution of indirect recoveries (expressed as a percent) for mallards banded during winterwithin the Bonn~ Reservoir management unit, Colorado from 1963-86. Adult Male Subadult Female Male Pennsylvania Atlantic Flyway Arkansas Illinois Iowa Louisiana Minnesota Mississippi Missouri Mississippi Flyway 0.6 0.2 0.2 0.6 3.2 Kansas Nebraska North Dakota Oklahoma South Dakota Texas Low Plains - Central Flyway 1.3 2.8 1.7 0.9 1.3 1.3 9.2 Colorado Kansas Montana Nebraska New Mexico North Dakota Oklahoma South Dakota Texas Wyoming High Plains - Central Flyway 45.4 2.2 1.1 10.5 0.9 1.7 0.4 0.6 1.5 1.1 65.4 California Colorado Idaho Montana Oregon Utah Ilashington Wyoming Pacific Flyway Alaska Alberta British Columbia Manitoba Ontario Saskatchewan Canada and Alaska Total (percent) Total Recovered 1.5 2.0 1.0 1.0 4.0 2.0 2.0 5.0 9.0 28.0 6.0 1.0 13.0 2.0 2.0 1.0 1.0 54.0 0.4 Female 9.5 0.2 0.2 Total birds 0.1 0.1 3.6 0.3 1.1 1.4 0.8 0.6 0.8 8.5 3.4 0.9 2.3 0.2 0.4 1.1 0.6 0.3 0.8 5.6 3.6 2.5 2.8 1.1 1.1 4.4 15.4 2.6 1.7 2.6 4.3 28.7 2.8 1.4 15.7 0.3 1.4 0.3 1.1 1.9 0.8 54.3 . 1.7 0.9 0.9 7.8 2.1 2.5 2.0 1.2 1.1 2.8 11.7 1.7 12.9 19.0 3.4 1.7 12.1 0.9 0.9 55.6 22.2 . 3.4 41.4 1.7 0.9 1.0 0.4 0.2 1.3 Male 0.9 0.9 0_3 0.3 2.5 0.3 0.3 0.2 Unknown 77.8 11.1 11.1 1.4 1.7 1.0 5.0 4.3 19.0 1.0 8.0 0.9 19.8 . 2.6 11.0 20.9 12.0 32.0 0.3 8_5 16.8 100.0 465 100.0 100 100.0 363 9.5 32.8 100.0 116 22.2 35.1 2.8 1.2 12.8 0.6 1.5 0.3 0.9 1.8 0.9 57.9 0.1 0.2 1.0 0.5 0.1 0.2 0.9 0.1 3.0 0.1 10.9 0.2 0.4 0.1 10.0 21.7 100.0 9 100.0 1053 �197 Table B-13. Distribution of direct recoveries (expressed as a percent) for mallards banded during winter in Southeast Colorado from 1963-83. Adult Male Arkansas Illinois Iowa Louisiana Missouri Mississippi Flyway 1.4 0.7 3.4 Kansas Nebraska North Dakota Oklahoma South Dakota Texas low Plains - Central Flyway 1.4 0.7 1.4 1.4 0.7 5.5 Colorado Kansas Montana Nebraska New Mexico North Dakota Oklahoma South Dakota Texas Wyoming High Plains - Central Flyway 44.8 1.4 0.7 6.9 0.7 1.4 0.7 0.7 4.1 2.1 63.4 Cal ifornia Colorado Idaho Montana Oregon Washington Wyoming Pacific Flyway Alberta Manitoba Saskatchewan Canada and Alaska Total (percent) Total Recovered 1.4 Subadult Female Male 2.5 2.0 1.0 . Female 2.5 . . 0.7 0.7 0.7 0.7 0.7 0.7 4.1 3.0 5.0 5.0 6.0 5.0 2.5 2.5 1.0 5.0 2.0 1.0 2.0 2.0 13.0 3.3 6.7 1.7 1.7 5.0 18.3 30.0 5.0 3.0 9.0 11.7 3.3 3.3 8.3 3.0 6.0 3.0 59.0 1.7 2.5 2.5 10.0 32.5 2.5 2.5 2.5 40.0 2.5 2.5 2.5 28.3 1.4 0.3 0.3 2.3 0.3 4.6 1.2 3.5 1.2 1.2 2.3 1.2 10.4 33.3 2.6 1.7 7.2 0.6 0.6 0.6 1.2 3.8 1.7 53.3 7.5 7.0 1.7 16.7 0.6 2.3 1.4 1.4 0.9 0.3 0.6 7.5 25.0 21.7 1.7 8.3 31.7 14.2 1.4 8.4 24.1 100.0 60 100.0 345 2.0 1.0 2.0 2.0 13.8 1.4 8.3 23.4 12.5 37.5 6.0 2.0 7.0 15.0 100.0 145 100.0 40 100.0 100 1.7 6.7 3.3 3.3 Total birds �198 Table 8-14. Distribution of indirect recoveries (expressed as a percent) for mallards banded during winter in Southeast Colorado from 1963-83. Adult Male Alabama Arkansas Iowa Louisiana Minnesota Mississippi Missouri Mississippi Flyway Subadult Female Male 0.3 2.1 0.3 1.0 0.3 Female 2.9 1.5 0.3 0.3 0.5 1.5 4.2 2.9 5.8 Kansas Nebraska North Dakota Oklahoma South Dakota Texas Low Plains - Central Flyway 0.8 1.5 2.6 1.5 0.5 1.8 8.7 1.5 1.5 1.7 2.1 1.4 1.4 1.5 0.3 0.7 4.2 9.0 2.9 2.9 2.9 11.6 Colorado Kansas Montana Nebraska New Mexico North Dakota Oklahoma South Dakota Texas Wyoming High Plains - Central Flyway 51.3 1.3 2.0 7.4 1.0 0.5 52.2 27.5 0.5 4.6 1.0 69.6 1.5 3.0 36.8 0.7 1.0 8.3 1.0 0.7 0.3 2.1 3.8 3.1 58.0 California Colorado Idaho Montana Nevada Oregon Utah Washington Pacific Flyway Alberta British Columbia Manitoba Saskatchewan Canada and Alaska Total (percent) Total Recovered 4.5 9.0 1.5 9.0 67.2 0.3 1.0 0.5 1.8 1.5 0.8 0.3 4.6 1.5 7.1 14.9 0.3 9.2 16.6 100.0 392 2.1 1.7 1.0 0.3 1.4 0.7 1.4 8.7 7.2 2.9 2.9 2.9 4.3 47.8 1.4 1.4 23.2 Total birds 0.1 0.7 0.1 0.7 0.1 0.2 0.2 2.3 1.2 1.7 1.2 1.2 0.7 2.9 9.1 44.2 0.9 1.5 7.8 0.9 0.7 0.1 1.3 4.0 2.0 63.5 0.1 1.2 1.0 1.2 0.1 0.5 0.7 0.6 5.5 1.5 4.5 20.9 9.7 0.3 2.1 8.0 20.1 10.1 33.3 10.0 0.1 1.0 8.5 19.6 100.0 67 100.0 288 100.0 69 100.0 816 �199 Table B-15. Distribution of direct recoveries (expressed as a percent) for mallards banded during winter within the Gunnison River management unit, Colorado from 1973-82. Subadult Adult Male Female Kansas Low Plains - Central Flyway Colorado Kansas Montana Nebraska New Mexico Wyoming High Plains - Central Flyway Arizona Colorado Idaho Montana New Mexico Utah Wyoming Pacific Flyway Alberta Saskatchewan total Total (percent) Total Recovered Male Female 1.1 1.1 5.7 3.1 4.7 7.8 1.6 62.5 4.7 1.6 1.6 7.8 4.7 84.4 0.5 0.5 2.5 2.5 6.9 6.9 79.3 3.4 3.4 . 86.2 1.1 2.3 3.4 12.5 1.1 65.9 3.4 2.3 1.1 5.7 1.1 80.7 4.7 3.1 7.8 6.9 6.9 4.5 1.1 5.7 100.0 64 100.0 29 100.0 88 Total birds 5.0 10.0 3.6 0.5 2.3 0.5 0.9 2.3 10.0 75.0 0.9 66.1 4.1 1.8 0.9 5.9 1.8 81.4 5.0 10.0 15.0 5.0 3.2 8.1 100.0 40 100.0 221 62.5 5.0 7.5 �200 Table S-16. Distribution of indirect recoveries (expressed as a percent) for mallards banded during winter within the Gunnison River management unit, Colorado from 1973-82. Adult Arkansas louisiana Minnesota Missouri Wisconsin Mississippi Flyway Nebraska North Dakota Texas Low Plains - Central Flyway Colorado Montana Nebraska New Mexico North Dakota South Dakota Texas Wyoming High Plains - Central Flyway Colorado Idaho Montana Nevada New Mexico Oregon Utah Washington Wyoming Pacific Flyway Alberta Saskatchewan Canada and Alaska Total (percent) Total Recovered Subadult Male Female 0.8 0.8 2.7 Male 1.4 . 1.4 2.7 0.6 0.6 1.3 2.7 0.6 1.6 2.7 0.6 1.3 5.4 2.7 1.6 Female 2.9 0.8 0.8 0.8 9.3 . 2.7 0.5 0.5 0.3 0.3 0.3 1.8 0.5 0.5 0.3 1.3 1.6 0.8 0.8 Total birds 3.2 0.6 2.5 0.6 1.3 10.1 1.4 2.9 1.4 2.5 10.8 1.4 17.4 53.6 1.4 1.4 5.1 0.5 1.8 0.8 0.5 0.3 0.3 1.5 10.7 60.5 9.3 0.8 62.2. 2.7 0.8 0.8 81.4 2.7 2.7 5.4 2.7 5.4 83.8 60.5 5.7 1.3 0.6 1.3 0.6 7.6 1.3 2.5 81.5 3.9 2.3 6.2 2.7 5.4 8.1 3.8 1.3 5.1 7.2 7.2 4.3 1.8 6.1 100.0 129 100.0 37 100.0 157 100.0 69 100.0 392 . 9.3 1.4 10.1 1.4 2.9 n.5 59.4 5.9 1.0 0.3 1.3 0.5 8.4 1.0 2.3 80.1 �201 Table B-17. Distribution of direct recoveries (expressed as a percent) for mallards banded during winter within the Colorado River management unit, Colorado from 1973-82. Subadult Adult Male Louisiana Mississippi Flyway Female Arizona Colorado Idaho Montana Oregon Utah Washington Wyoming Pacific Flyway Alberta British Columbia Manitoba Saskatchewan Canada and Alaska Total (percent) Total Recovered Female 1.1 1.1 2.2 2.2 1.1 1.1 1.1 0.4 0.4 0.8 4.4 4.3 3.3 2.1 2.6 1.1 1.1 6.7 63.3 5.6 4.4 Total birds 0.4 0.4 1.1 1.1 Nebraska Texas Low Plains - Central Flyway Colorado Kansas Montana Nebraska New Mexico South Dakota Wyoming High Plains - Central Flyway Male 2.6 8.9 4.3 10.6 61.5 7.7 2.2 50.0 13.3 2.2 2.1 55.3 6.4 2.1 8.5 2.1 2.1 78.7 3.0 0.4 1.9 0.4 0.4 0.4 1.1 7.5 1.1 57.1 8.6 2.6 0.4 10.5 0.4 1.9 82.7 11.1 2.6 15.4 8.9 1.1 85.6 87.2 3.3 80.0 4.4 5.1 6.7 2.2 6.7 5.1 10.3 1.1 1.1 8.9 2.1 10.6 5.6 0.4 0.4 2.3 8.6 100.0 90 100.0 39 100.0 90 100.0 47 100.0 266 6.4 2.1 �202 Table 8-18. Distribution of indirect recoveries (expressed as a percent) for mallards banded during winter within the Colorado River management unit, Colorado from 1973-82. Adult Male Arkansas Louisiana Mississippi Flyway Kansas Nebraska Oklahoma .South Dakota Texas Low Plains - Central Flyway Colorado Montana Nebraska New Mexico Texas Wyoming High Plains - Central Flyway Arizona California Colorado Idaho Montana New Mexico Oregon Utah Washington Wyoming Pacific Flyway Alberta British Columbia Manitoba Saskatchewan Canada and Alaska Total (percent) Total Recovered Subadult Female Male Unknown Female Male 1.0 2.0 2.0 0.4 0.2 0.6 1.0 1.0 0.5 0.5 0.4 0.4 0.2 0.2 0.2 1.4 1.1 1.1 0.6 0.6 4.0 1.7 0.6 2.0 1.1 7.5 2.0 64.9 5.7 0.6 2.0 2.0 64.7 3.9 2.0 0.6 7.5 2.0 9.8 1.1 80.5 7.8 94.1 8.6 0.6 0.6 1.7 11.5 2.0 100.0 174 2.1 2.2 4.6 1.5 2.2 1.1 1.1 1.1 2.6 0.5 1.5 10.8 0.5 1.0 60.8 5.2 1.0 1.0 0.5 6.7 1.0 1.5 79.4 1.1 6.7 58.9 3.3 1.1 50.0 50.0 50.0 1.1 13.3 3.3 81.1 4.6 0.5 7.8 2.0 1.5 6.7 2.2 10.0 100.0 51 100.0 194 100.0 90 . Total birds 50.0 3.9 1.4 0.4 1.2 0.2 1.2 8.2 0.4 0.6 62.2 4.9 1.0 0.4 0.8 8.4 0.4 2.3 81.4 6.3 0.4 0.2 1.6 8.4 100.0 2 100.0 511 �Table C-1. Distribution of Colorado recoveries (expressed as a percent for mallards banded during winter within the Sterling-Julesburg Colorado from 1963-83. Indirect recoveries Direct recoveries Subadult Adult Adult Subadult Male Female Male Female Total Male Female Male Female Total All birds Sterling-Julesburg Ft. Morgan-Sterling Greeley-Ft. Morgan South Platte River 60.8 20.3 5.4 86.5 75.0 18.8 30.8 23.1 23.1 76.9 55.2 21.7 6.3 83.2 56.3 20.2 9.8 86.3 71.1 18.4 · 45.0 25.0 5.0 75.0 89.5 35.1 26.4 13.5 75.0 40.0 20.0 16.7 76.7 48.6 22.3 10.8 81.7 50.4 22.1 9.6 82.1 Northern Foothills Southern Foothills Foothills 1.4 6.8 8.1 · · 2.5 17.5 20.0 7.7 7.7 15.4 2.1 9.1 11.2 4.4 6.0 10.4 10.5 10.5 · 4.7 16.2 20.9 3.3 16.7 20.0 4.0 11.0 15.0 3.5 10.5 14.0 Bonny Reservoir 1.4 · · 0.7 1.1 3.3 1.5 1.3 · · 6.2 · · · · · . 0.5 · · · · ·· 2.0 0.3 1.4 · · · · ·· · · · 0.3 0.2 0.2 0.2 0.2 0.7 · 0.7 0.7 1.4 0.5 0.3 0.8 0.7 0.4 1.1 · 0.5 0.3 0.8 0.4 0.4 0.7 100.0 399 100.0 542 Unit Unit Unit Uni t Southeast 8 10 11 12 Colorado San Luis Valley Middle Park Mountain Parks Gunnison River Colorado River Western Colorado Total (percent) Total recovered · 1.4 2.7 93.7 · · 6.2 2.5 · ·· · · 100.0 74 100.0 16 100.0 40 1.4 1.4· 2.5 2.5 2.5 management unit, · 7.7 7.7 · 100.0 13 0.7 0.7 0.7 2.1 · 0.5· 1.4 0.7 2.1 0.5 0.5 . 1.1 · 0.7 0.7 · 1.1 · · · · 100.0 143 100.0 183 100.0 38 · · · · · 0.7 0.7 100.0 148 . 100.0 30 N o w �Table C-2. Distribution of Colorado recoveries (expressed as a percent for mallards banded during winter within the Ft. Morgan·Sterling management unit, Colorado from 1963-84. N 0 ~ Direct recoveries Adult Indirect recoveries Subadult Adult Subadult All birds Male Female Male Female Total Male Female Male Female Sterling·Julesburg Ft. Morgan·Sterling Greeley·Ft. Morgan South Platte River 12.8 57.4 12.8 83.0 10.0 60.0 16.7 86.7 10.1 36.2 20.3 66.7 16.7 50.0 25.0 91.7 11.7 50.2 16.6 78.5 8.9 51.8 20.0 80.7 9.8 54.9 23.5 88.2 10.1 30.3 23.9 64.3 5.8 38.5 19.2 63.5 9.1 43.0 21.7 73.8 . 9.8 44.8 20.4 75.0 Northern Foothills Southern Foothills Foothills 1.1 11.7 12.8 3.3 6.7 10.0 11.6 17.4 29.0 8.3 5.4 12.2 17.6 2.6 13.1 15.7 3.9 5.9 9.8 6.3 24.8 31.1 15.4 17.3 32.7 5.1 17.2 22.3 5.2 16.0 21.2 1.1 · 2.9 1.5 1.3 · 0.8 1.9 1.1 1.2 0.5 0.7 0.7 ·· 0.4 0.5 0.5 0.2 0.8 1.9 0.5 0.4 0.2 0.6 1.6 0.4 0.4 0.5 0.5 0.7 0.7 0.4 0.8 1.3 0.2 0.3 0.5 0.1 0.2 0.4 100.0 646 100.0 851 Bonny Reservoir Unit Uni t Uni t Unit Southeast 8 10 12 13 Colorado San luis Valley Mountain Parks .. 1.1 · · · 3.3 2.1 2.1 1.1 Gunnison River Colorado River Western Colorado Total (percent) Total recovered 3.3 100.0 94 · · · 8.3 · ·· 0.5 · · · · 1.0 · 0.3 1.6 · 1.4 1.4 ·· 1.5 1.5 0.7 0.7 · · · · · · ·· 100.0 30 100.0 69 100.0 12 100.0 205 . 100.0 305 · 2.0 2.0 · · 100.0 51 . 1.9 0.4 1.3 2.1 1.9 100.0 238 100.0 52 Total �Table C-3. Distribution of Colorado recoveries (expressed as a percent for mallards banded during winter within the Greeley-Ft. Morgan management unit, Colorado from 1963-83. Direct recoveries Indirect recoveries Subadult Adult Adult Subadult -Male Female Male Female Total Male Female Male Female Total All birds Sterling-Julesburg Ft. Morgan Sterling Greeley-Ft. Morgan South Platte River 1.0 8.6 60.0 69.5 6.9 17.2 41.4 65.5 7.5 15.1 34.0 56.6 3.3 13.3 46.7 63.3 3.7 12.0 49.3 65.0 4.1 13.6 48.6 66.4 4.0 12.0 36.0 52.0 2.4 10.9 38.9 52.1 6.5 37.0 43.5 · 3.0 11.8 42.5 57.3 3.2 11.8 44.5 59.5 Northern Foothills Southern Foothills Foothills 5.7 22.9 28.6 6.9 20.7 27.6 5.7 34.0 39.6 10.0 23.3 33.3 6.5 25.3 31.8 8.2 22.7 30.9 18.0 28.0 46.0 12.3 30.8 43.1 21.7 30.4 52.2 12.0 27.1 39.1 10.3 26.6 37.0 1.0 3.4 1.9 1.4 0.9 · 0.9 2.2 0.9 1.1 · · · · · 3.4· · ·· 2.0 · 0.9 0.5 0.5 1.9 2.2 3.4 0.5 0.5 0.5 1.4 · 2.2· 0.8 0.6 0.4 1.7 0.5 0.4 0.4 1.3 · ·· . . ·· 0.4 Bonny Reservoir Unit Unit Unit Southeast 8 10 13 Colorado North ParI<: San Luis Valley Middle ParI<: Mountain Parl<:s Gunnison River Colorado River Ya~a River Western Colorado Total (percent) Total recovered 1.0 · ·· · · 1.9 · · · 0.5 0.5 0.5 0.5 0.5 1.4 0.5 · · 2.0 · · · · · · · · · · · · · · · · · 100.0 105 100.0 29 100.0 53 100.0 30 100.0 217 100.0 220 100.0 50 1.0 1.9 · 3.3 3.3 ·· 0.5 ·· · 0.5 0.5 0.5 0.5 0.5 1.4 100.0 211 · ·· · 100.0 46 0.4 0.1 0.4 0.1 0.7 0.2 0.2 0.2 0.6 0.1 0.1 0.1 0.4 100.0 527 100.0 744 N o VI �Table C-4_ Distribution of Colorado recoveries (expressed as a percent) for mallards banded during winter within the Northern Foothills management unit, Colorado from 1963-84. N 0 (J\ Indirect recoveries Direct recoveries Adult Subadult Male Female Male Female Sterling-Julesburg Ft. Morgan-Sterling Greeley-Ft. Morgan South Platte River 1.9 4.8 8.7 15.4 6.7· · 3.3· 6.7 1.4 8.5 7.0 16.9 Northern Foothills Southern Foothills Foothills 44.2 38.5 82.7 64.4 20.0 84.4 32.4 47.9 80.3 Bonny Reservoir Unit Unit Uni t Uni t Southeast 8 10 12 13 Colorado North Park South Park San Luis Valley Mountain Parks Southwest Gunnison River Colorado River Yafl1'8River Western Colorado Total (percent) Total recovered · 1.0 1.0 · 1.0 1.0 · · · 2.2· 2.2 2.2 · · 2.2 · · · ·· 2.2 2.2 4.4 100.0 104 100.0 45 · · ·· · · · · · 100.0 71 -- --- Male Female Male Female Male 50.0 1.6 4.4 7.1 13.1 0.9 1.5 6.3 8.7 1.4 7.0 8.5 · 0.8 5.9 17.2 23.9 2.1 8.3 10.4 44.8 38.5 83.3 52.6 36.6 89.2 56.3 31.0 87.3 28.6 44.5 73.1 50.0 35.4 85.4 0.4 2.1 0.8 0.4 · · · 50.0 43.3 93.3 50.0 50.0 3.3 3.3 · · · · · · · · · · · · · · 1.2 1.2 0.4 0.4 0.8 1.6 0.3 · 0.3 0.3 0.3 0.9 · 0.6· · · ·· · 0.4 0.8 0.3 100.0 2 100.0 252 100.0 333 · 100.0 30 Unknown Total 50.0 · · · Subadult Male 3.3 · 1.4 1.4 2.8 Adult Unknown · 0.4· 2.8 1.4 4.2 . . 1.3 · 0.6 · 0.3 · · · 100.0 71 0.4 0.4 0.4 · · ·· 2.1· 2.1· Total All birds 0.7 3.0 10.2 13.9 0.9 3.4 9.4 13.7 44.6 38.6 83.2 44.7 38.6 83.2 0.4 0.3 · · 0.6 0.3 0.1 0.3 1.3 0.4 0.2 0.1 0.5 1.3 · · 0.1 0.3 0.4 0.1 0.2 0.4 0.7 14.3 0.3 0.3 0.1 · · · 57.1 28.6 85.7 1.3 · 14.3 0.7 0.2 0.2 0.2 0.1 0.7 100.0 238 100.0 48 100.0 7 100.0 697 100.0 949 · · · · �Table C-5_ Distribution of Colorado recoveries (expressed as Colorado from 1963-84. 8 percent for mallards banded during winter within the Southern Foothills management unit, Direct recoveries Adult Indirect recoveries Subadult Adult Subadult Unknown Total All birds · ·· 1.5 3.5 11.3 16.4 1.4 3.0 10.9 15.3 · 19.0 62.2 81.2 17.5 63.6 81.1 0.4 0.4 0.1 0.1 0.1 0.2 0.3 0.3 0.4 1.1 Male Female Male Female Total Male Female Male Female Sterling-Julesburg Ft. Morgan-Sterling Greeley-Ft. Morgan South Platte River 1.2 3.0 10.9 15.2 · 17.0 17.0 · 1.4 1.4 7.9 10.8 2.0 2.0 5.9 9.8 1.2 2.0 10.0 13.2 1.6 3.2 9.3 14.1 1.5 3.1 4.6 1.8 4.2 14.1 20.1 1.4 4.3 18.8 24.6 Northern Foothills Southern Foothills Foothills 13.3 68.5 81.8 6.4 72.3 78.7 15.8 66.2 82.0 21.6 54.9 76.5 14.4 66.4 80.8 22.3 61.7 84.0 16.9 73.8 90.8 15.5 62.5 78.1 17.4 52.2 69.6 100.0 100.0 · · 0.7 2.0 0.5 0.5 · 0.4 · 2.1 · · 0.4 1.4 1.4 2.2 0.2 0.5 0.5 1.2 2.5 · · 2.1 · · 2.0 2.0· · · · 2.2· 0.4 2.9 · 2.1· 1.4 · · 3.9· 1.4 3.9 0.5 0.2 0.5 1.2 · 0.8· 2.0 3.9 5.9 . 0.3 1.4 1.4 2.9 0.7 1.0 1.7 0.3 0.5 3.1 3.1 0.7 2.9 · 100.0 139 100.0 51 100.0 402 100.0 376 100.0 65 100.0 283 100.0 69 100.0 1 Bonny Reservoir Unit Unit Unit Uni t Southeast 8 10 12 13 Colorado 0.6 0.6 0.6 1.2 3.0 South Park San luis Valley Middle Park Mountain Parks · · Southwest Gunnison River Colorado River Western Colorado · · · · Total (percent) Total recovered · 100.0 165 2.1 · ·· 100.0 47 · · · · · · · 0.8 · 0.4 1.5 · 0.4· · 0.7 1.5 · ·· · · 2.9 Male · · · 0.4 · · · · · 0.1 0.5 0.6 0.3 0.4 0.2 0.8 0.1 0.3 0.6 1.0 0.1 0.4 0.8 1.3 100.0 794 100.0 1196 . N o -...J �Table C-6. Distribution of Colorado recoveries (expressed as a percent) for mallards banded during winter at Bonny Reservoir, Colorado from 1983-86. N 0 co Direct recoveries Adult Subadult Indirect recoveries Adult Unknown Subadul t Unknown --Male · Female Male Female · 4.3 8.7 · · Sterling-Julesburg Ft. Morgan Sterling Greeley-Ft. Morgan South Platte River 7.1 2.4 9.5 5.9 5.9 11.8 Northern Foothills Southern Foothills Foothills · 5.9 3.6 3.6 Bonny Reservoir Uni t Uni t Uni t Uni t Southeast 10 12 13 14 Colorado San Lui s Valley Mountain Parks Gunnison River Colorado River Western Colorado Total (percent) Total recovered 85.7 5.9· 64.7 · · 17.6 1.2 1.2 · · · 13.0 · 16.7 16.7 · 34.8 34.8 16.7 16.7 47.8 41.7 · · 4.3· 4.3 8.3 Male Total Male Female Hale Female 0.7 6.6 3.6 10.9 4.3 3.8 5.2 13.3 3.6 13.6 3.6 7.7 7.7 7.7 23.1 0.7 9.5 10.2 1.0 2.9 3.8 3.6 7.1 10.7 4.8 11.5 16.3 13.6 73.0 80.5 78.6 50.0 68.2 ·· 2.2 1.5 3.6 . 0.5 0.5 1.0 3.6 · 2.9 3.8 1.9 0.7 4.4 1.9 7.1 8.7 · 1.0 1.0 0.7 0.7 0.5 100.0 137 100.0 210 · ·· · · · · 100.0 · · · · · 8.3 8.3 ·· · · · · 8.3 8.3 ·· · · ·· 100.0 84 100.0 17 100.0 23 100.0 12 100.0 1 · · 17.6 8.3 · --- · 0.7 0.7 . . . 0.5 · 4.5 18.2 13.6· Total All birds · · 5.7 4.3 6.2 16.3 4.3 4.9 5.5 14.8 · ·· 3.0 5.4 8.4 2.4 6.5 8.9 70.5 71.1 1.4 1.4 1.4 · 4.1 1.6 1.4 1.0 0.2 4.2 Male 60.0 60.0 40.0 · ·· · · · · · 0.3 0.3 0.4 0.4 1.0 · · · ·· · 0.5 · . 0.5 0.4 0.2 0.6 100.0 28 100.0 104 100.0 22 100.0 5 100.0 369 100.0 506 · · . 1.0 · . . �Table C-7. Distribution of Colorado recoveries (expressed as a percent) for mallards banded during winter in Southeast Colorado from 1963-83. Indirect recoveries Direct recoveries Adult Subadult Adult Subadul t Male Female Male Female Total Male Female Male Female Total All birds Sterling-Julesburg Ft. Morgan-Sterling Greeley-Ft. Morgan South Platte River 1.6 6.3 7.9 15.9 3.3 3.3 0.8 5.1 4.2 10.2 3.0 7.5 4.0 14.6 6.2 12.5 3.6 5.4 3.6 12.6 · · ··· 6.2 7.1 5.3 5.3 10.5 3.3 6.1 4.2 13.6 2.7 5.8 4.2 12.7 Northern Foothills Southern Foothills Foothills 3.2 1.6 4.8 7.1 7.1· 23.3 23.3 9.1 9.1 · 2.5 7.6 10.2 2.5 8.5 11.1 12.5 12.5 · 6.3 17.1 23.4 · · 3.3 11.1 14.4 3.1 10.2 13.4 Bonny Reservoir 1.6 · 6.7 9.1 3.4 2.0 3.1 2.7 5.3 2.5 2.7 0.8 2.5 · 0.9 5.3 1.9 1.7 13.6 5.9 24.6 22.9 0.8 68.6 10.6 13.1 18.1 25.1 15.6 9.4 25.0 18.8 10.5 10.5 26.3 26.3 69.3 · 68.7 · 10.8 0.9 18.0 22.5 0.9 54.1 11.1 8.9 19.1 23.8 0.3 65.1 11.7 8.1 20.5 23.6 0.4 66.0 0.5 2.5 3.1 · 3.1· 0.9 0.9 0.9 2.7 0.6 2.2 0.3 3.0 1.0 2.1 0.6 3.8 0.6 0.8 Unit 8 Uni t Uni t Unit Unit Uni t Southeast 10 11 12 13 14 Colorado South Park San Luis Valley Middle Park Mountain Parks Gunnison River Colorado River Yampa River Western Colorado Total (percent) Total recovered 7.1 · · · 1.6 12.7 6.3 31.7 22.2 1.6 76.2 21.4 21.4 13.3 6.7 16.7 23.3 71.4 60.0 · 1.6· 7.1 7.1 1.6 14.3 28.6· ·· · · 100.0 63 100.0 14 · · 3.3 3.3 6.7 · · 9.1 9.1 · 45.5· 27.3 18.2 · 18.2· 2.5 1.7 1.7 5.9 9.1 0.8 · 9.1 18.2 100.0 30 100.0 11 3.0 1.7 · · 100.0 118 100.0 199 . 0.8 · · · 78.9 · 5.3 5.3 1.8 2.7 · · 4.5 · 1.4 0.6 0.6 0.2 1.5 100.0 32 100.0 111 100.0 19 100.0 361 100.0 479 . Table C-8. Distribution of Colorado recoveries (expressed as a percent for mallards banded during winter within the Gunnison River management unit, Colorado from 1973-82. N 0 I.D �Table C-8. Distribution of Colorado recoveries (expressed as a percent for mallards banded during winter within the Gunnison River management unit, Colorado from 1973-82. N I-' 0 Indirect recoveries Direct recoveries Subadult Adult Male Sterling-Julesburg Ft. Morgan-Sterling Greeley-Ft. Morgan South Platte River · 2.4 Northern Foothills Southern Foothills Footh ills 2.4 Bonny Reservoir 2.4 Unit Unit Unit Southeast 8 13 14 Colorado San Luis Valley Mountain Parks Southwest Gunnison River Colorado River Ya~a River Western Colorado Total (percent) Total recovered · · · 2.4 2.4 · Female Male · ·· 1.6 Female · · · · 1.6· · ·· 1.6 3.2 4.8 · · 1.6· · · · · · · 1.6 · · · 92.9 · 4.3 73.9 17.4 4.3 100.0 100.0 42 100.0 23 76.2 16.7 Adult Total 0.7 · 1.2 1.2 · · · 1.3 1.3 2.6 2.4 2.4 · 0.7 · · · · 0.7· · 0.7 0.7 · 4.2 4.2 8.0 80.0 12.0 1.2 78.8 11.8 79.2 12.5 1.0 66.3 26.5 91.8 · 91.7 · 93.9 100.0 85 100.0 24 100.0 98 0.7 100.0 · 100.0 63 100.0 25 100.0 153 · Male · 92.1 · Female · 0.7 2.0 71.2 20.9 0.7 94.8 63.5 28.6 Male Subadult ·· Female 2.0 2.0 2.4 1.2 1.6 · 1.0 2.0 3.1 7.1 2.4 9.5 1.6 2.4 4.0 1.5 2.0 3.5 2.4 · · · 0.8 0.7 1.2 1.2 · · · · · · · 2.4· · 2.4 0.4 0.4 . 0.4 1.2 0.2 0.5 0.2 1.0 0.8 0.8 0.7 0.7 69.0 16.7 0.8 72.3 18.5 85.7 · 91.6 1.2 71.9 19.4 0.2 92.8 100.0 42 100.0 249 100.0 402 · 2.4 · 4.2 4.2 · · 1.0 1.0 0.4 All birds 0.2 0.2 0.7 1.2 · 2.4 Total . �Table C-9. Distribution of Colorado recoveries (expressed as a percent for mallards banded during winter within the Colorado River management unit, Colorado from 1973-82. Direct recoveries Adult Indirect recoveries Subadult Adult Subadult Unknown --Male Sterling-Julesburg Ft. Morgan Sterling Greeley-Ft. Morgan South Platte River 1.7 Northern Foothills Southern Foothills Foothills 1.7 1.7 1.7 Unit 8 Uni t 13 Southeast Colorado South Park San Luis Valley Middle Park Mountain Parks Southwest Gunnison River Colorado River Yampa River Western Colorado Total (percent) Total recovered Female · · · · · · · · · · · ·· · · · · 29.2· 1.7 10.2 84.7 70.8 96.6 100.0 59 Male · · · · · 4.1 4.1 Female · ·· · 7.1· 7.1 2.0 2.0 4.1 · · 2.0· · 2.0 100.0 · 2.0 16.3 69.4 2.0 89.8 100.0 24 100.0 49 · · 3.6 3.6 Total Male Female Male · 0.8 0.6 · 0.6 · 2.9 0.8 1.7 3.1 3.1 1.7 1.7 0.6 0.6 1.3 1.7 1.7 · 0.6 0.6 1.3 3.6 · 0.8 0.8 · · · 0.8 2.4 3.1 2.9 · 0.8 0.8 · 3.6 · · · · Total All birds · · · 0.3 0.9 0.6 1.8 0.2 0.8 0.4 1.4 · 0.3 1.8 2.1 0.2 2.2 2.4 Male 50.0 50.0 ·· · ·· . 0.9 0.9 0.2 0.8 1.0 0.6 1.5 0.3 2.4 0.4 1.2 0.4 2.0 1.0 15.1 76.1 1.0 93.2 100.0 498 · · 0.8 3.9 · 0.8 · 4.7 1.8 1.8 14.2 80.0 · 2.9 14.7 79.4 1.8 18.2 70.9 3.6 94.5 50.0 · 50.0· 0.9 15.4 75.4 1.2 92.9 100.0 55 100.0 2 100.0 338 0.8 · 89.3 1.3 14.4 77.5 0.6 93.7 94.2 · 97.1 0.8 15.7 72.4 1.6 90.6 100.0 28 100.0 160 100.0 120 100.0 34 100.0 127 7.1 82.1 Female · · · N •..... •..... ��Colorado Division Wildlife Research September 1991 of Wildlife Report JOB PROGRESS REPORT State of __~C~o~l~o~r~a~d~o~ Project _ Avian Research W-152-R-4 Work Plan 10 Job Title: : Job Cooperative Management Programs 01 April 1990 through 31 March 1991 Author: R. Szymczak Personnel: Game Birds _1_ Period Covered: Michael - Migratory James K. Ringelman Wildlife and Michael R. Szymczak, Colorado Division of ABSTRACT The final draft of the San Luis Valley Waterbird Plan was completed and submitted for administrative approval. Recommendations concerning waterfowl and wetland issues were made to committees formulating the North Park waterfowl plan and San Luis Valley Water Plan. Recommendations for wetland habitat improvements on present holdings or proposed acquisitions were provided for Russell Lakes, South Republican and Riggs SWA's and the Ham property and Pole Mountain Reservoir. Presentations on various aspects of waterfowl ecology were given at training and educational schools, workshops, and short courses. Waterfowl counts designed to help identify population limiting factors were conducted on ducks and/or geese in portions of North Park and South Park. An extensive study of duck production was conducted in montane habitat study areas on the Routt National Forest. Responsibilities as Colorado's representative on Pacific Flyway Study Committees and Council were fulfilled. ��215 Cooperative Migratory Bird Management Programs Michael R. Szymczak James K. Ringelman In 1988, the Colorado Division of Wildlife (CDOW) created the Migratory Game Bird Program Unit (MBPU) within the Terrestrial Wildlife Section. This administrative change combined all individuals having statewide responsibilities for research and management of migratory game birds. Members of the MBPU work in concert to improve migratory bird management in Colorado, This job was created to allow team members to participate in these management programs. P. N. OBJECTIVES 1. Participate in developing and implementing habitat-based waterfowl management plans on a statewide, habitat region, and project basis. 2. Advise state and federal land managers on beneficial habitat acquisitions and/or developments and provide expertise in preparation of development and/or management plans. Advise private land managers in developing habitat management plans and assessing impacts on waterbird populations. 3. Present information on the principles of waterfowl management to workshop attendees, educational classes, and conservation organizations. 4. Participate in migratory bird management meetings at the state and flyway levels. 5. Cooperate in developing surveys and techniques that will assess the impact of migratory bird management programs. SEGMENT OBJECTIVES 1. In conjunction with the statewide waterfowl management plan, begin work on waterfowl habitat region plans. 2. Provide biological expertise for wetland development programs on the Brush, Russell Lakes, South Republican State Wildlife Areas, Hebron Ponds and Walden Reservoir (BLM-CDOW), Routt National Forest (U.S. Forest Service - TAKING WING), and other areas when requested. 3. Prepare and present informational programs on migratory bird management when requested. 4. Compile appropriate population status information and represent Colorado at Pacific Flyway Technical Committee and Council meetings. Attend migratory bird management program meetings in Colorado when requested. 5. Provide methodology for wetland habitat and migratory bird population surveys when requested. �216 RESULTS Waterfowl Management Plans The San Luis Valley Waterbird Plan was completed by a multi-agency plan committee and awaits submission to CDOW administrators. Migratory Game Bird personnel compiled sections of the plan dealing with historical levels of game bird resources and formulated objectives, issues, and strategies dealing with game bird populations and their habitats. The entire plan was reviewed a number of times by project personnel. The current draft of the CDOW Water Management Plan for the San Luis Valley was reviewed. Input on current years water allocation was provided in a meeting of plan committee members. Current and future management plans for the Monte Vista National Wildlife Refuge were discussed with refuge personnel. Current on-the-ground changes were visited and unofficial comments were made to refuge personnel on the direction of current management. Some preliminary information on waterfowl populations was forwarded to CDOW personnel responsible for developing a Waterfowl Management Plan for North Park. In conjunction with the North Park Plan, project personnel participated on a multi-agency committee formulating a proposal for a major expansion of the Arapaho NWR. The proposal encompasses the purchase of MacFarlane and Pole Mountain reservoirs and a major portion of the Soap Creek Ranch. Completion of planned purchases would result in a reliable water source for the Hebron Sloughs development. Wetland Developments and Acquisitions Potential sites for wetland developments and existing wetlands were visited on the Russell Lakes SWA in the San Luis Valley, the South Republican SWA at Bonny Reservoir and the Rigg SWA near Rocky Ford. Recommendations for wetland construction, water management, and management of existing wetlands were given to CDOW management personnel. Possible solutions for impending water availability problems at Eads Lakes were discussed with CDOW Southeast Regional personnel. Proposed areas for wetland enhancement in the Denver Metropolitan area were toured and suggestions made for development to CDOW Central Region personnel. Waterfowl count information for Pole Mountain Reservoir in North Park was compiled and forwarded to the Northeast Region in support of acquisition of the Reservoir. Wetlands on the Ham property located along the lower Arkansas River in Southeast Colorado were evaluated for CDOW acquisition. Project personnel, as members of the multi-agency Waterfowl Habitat Project Review Committee, reviewed and rated wetland enhancement and acquisition proposals from land management agencies for funding with Colorado State Duck Stamp monies. Ten enhancement proposals were recommended for funding and forwarded for CDOW Terrestrial staff and appropriate Regional approval. Informational Programs Formal presentations on waterfowl recruitment in montane habitats were presented to U.S. Forest Service personnel in Steamboat Springs, the Colorado Natural Areas Committee, and the national meeting of the Society of Wetland Scientists. Presentations on waterfowl ecology and management were made to CDOW District Wildlife Management Trainees, the Platte River Symposium, the Wildlife Management Shortcourse and Wildlife Biology Workshop at Colorado State University and the Avian Management course at Colorado State University. Presentations on �217 wetland development for waterfowl and duck identification were given at the Southeast Region training school. cnaw Waterfowl Technical Committee and Council Meetings The Rocky Mountain Greater Sandhill Crane and Rocky Mountain Canada Goose Management Plans were reviewed and revisions proposed by respective population committee members, including Colorado, at a special meeting in January 1991. Crane plan changes were forwarded to the Pacific Flyway Council and approved in March 1991. All RMP goose plan changes have yet to be finalized. Colorado was represented at the July 1990 and March 1991 Pacific Flyway Study Committee meetings by project personnel. Waterfowl population status was reviewed in July and hunting season recommendations forwarded to the U. S. Fish and Wildlife Service's Regulation Committee. Populations of specific interest to Colorado whose status was reviewed in July were (1) breeding and wintering mallards inhabiting the Pacific Flyway portion of Colorado and (2) the Rocky Mountain Canada goose population. General information on Pacific Flyway migratory game bird populations was exchanged by committee members in March. Regulatory recommendations made to the Pacific Flyway Council included hunting season regulations for Rocky Mountain Greater Sandhill Cranes and Four-corners Band-tailed Pigeon populations. Population Survey Methodology Spring duck count surveys were conducted on Spinney Mountain SWA in South Park and in kettle lake-beaver pond habitat immediately south of Big Creek Lake in North Park. Spinney Mountain counts were designed to obtain pre-wetland development baseline information on breeding pairs on the SWA. Counts in North Park were designed as a test of a predictor equation for estimating the number of duck breeding pairs in montane habitats using aerial photographs, with pond size the critical variable. The equation was developed using information obtained during a cooperative study of waterfowl recruitment in montane habitats (Langley et al. 1990). An extensive, cooperative study of nesting waterfowl on forested study sites immediately west of North Park (Routt National Forest) continued in 1990. New study sites were selected in 1990 with the primary intent of establishing protocol for future waterfowl research and management on Forest Service lands. Results of this cooperative study are presented as an appendix to this report. Surveys of nesting and brood rearing Canada geese on Walden and MacFarlane reservoirs in North Park during spring 1990 documented poor productivity at both sites. Goose nests visited, marked, and mapped in early May just prior to hatch, and revisited after all had hatched showed good hatchability. Periodic brood counts made post-hatch indicated gosling mortality probably occurred shortly after hatch. A detailed report was submitted to Northeast Regional and Bureau of Land Management personnel. DISCUSSION Project personnel proved to be a useful resource in planning and evaluating waterfowl management and habitat enhancement programs in Colorado and educating land management agency personnel about the habitat requirements of waterfowl. We experienced increased involvement in long-range planning for waterfowl on specific areas. We anticipate that with increased emphasis on habitat enhancement in Colorado as outlined in the statewide Waterfowl Management Plan that our services will be more in demand. �218 Continued participation on Flyway committees ensures that Colorado will remain informed on migratory bird matters and have input in migratory bird hunting regulations. Prepared by: Michael R. Szymczak Wildlife Researcher C �219 APPENDIX A Report on the cooperative investigation of waterfowl recruitment habitats in the Routt National Forest, Colorado in montane Field Personnel R. Langley, Colorado Division of Wildlife M. Wotawa, Colorado Division of Wildlife J. Ringelman, Colorado Division of Wildlife Cooperators C. Neelan, u.s. Forest Service, Steamboat Springs S. Kozlowski, U.S. Forest Service, Walden G. Hetzel, U.S. Forest Service, Denver L. Mullen, U.S. Forest Service, Denver S. Porter, Colorado Division of Wildlife, Walden K. Snyder, Colorado Division of Wildlife, Walden M. Szymczak, Colorado Division of Wildlife, Ft. Collins R. Hopper, Colorado Division of Wildlife, Ft. Collins ��221 WATERFOWL ABUNDANCE AND PRODUCTION ON THE ROUTT NATIONAL FOREST, COLORADO, 1990 James K. Ringelman, Mark A. Wotawa, and Richard S. Langley Colorado Division of Wildlife, Ft. Collins, CO 80526 Abstract: We enumerated duck breeding pairs and broods, derived indices of nesting success, and documented duckling survival in 6 montane wetland communities in the Routt National Forest near the western edge of North Park, Colorado. Of the estimated 92 resident pairs, the majority were mallards (51%) and ring-necked ducks (21%), followed by bufflehead (13%) and green-winged teal (10%). Overall pair density averaged 11.2 patrs/krn". Pair densities ranged from 6.0 pairs/krrr' at Newcomb Creek to 25.7 palrs/krn" at Upper Beaver Creek. Newcomb Creek and Lone Pine\Bear Creeks had fewer than 10.0 pairs/krn", whereas Colorado Creek, Livingston Park, and Upper Beaver Creek had > 20.0 palrs/krrr'. Ratios of lone males to pairs indicated that nest success was relatively high (> 50%) for all areas and species. A minimum of 56 broods were produced, but duckling survival was low for ring-necked ducks and bufflehead. Consequently, the 99 ducklings that were observed on the areas near the time of fledging was far short of the potential recruitment of 230 birds. Compared with the 1989 Routt Forest study area, the 6 areas surveyed in 1990 had higher pair densities on a unit area basis (11.2 versus 6.1 km2) but similar densities of pairs per pond (0.30 versus 0.32). Nest success appeared high during both years. Overall estimates of minimum brood survival were also similar (54% in 1989 versus 48% in 1990). However, high duckling mortality rates among ring-necked ducks and buffleheads limited recruitment both years. Recent declines in North American duck populations have prompted coordinated action by state and federal natural resource agencies. Under the umbrella of the North American Waterfowl Management Plan, the U.S. Forest Service (USFS) has initiated a 5-year, $18 million program to survey and improve waterfowl habitat on USFS lands. This program, called "Taking Wing", addresses the needs of breeding waterfowl through habitat preservation and development. Based on USFS landholdings in 1988, forested montane habitat covers at least 56,000 km2 in the central Rocky Mountain region. Although comparable in size to the United States portion of the prairie pothole region (280,000 km2), Rocky Mountain montane waterfowl habitats have attributes that set them apart from their grassland counterparts. First, montane wetland communities are relatively intact compared to the widespread wetland degradation typical of the northern Great Plains. Second, except for relatively small, confined, impacts such as mining operations, roads, and ski areas, adjacent upland plant communities still possess a compliment of native plant species despite some grazing and timber harvest. Third, snowfall provides a reliable source of spring water, creating waterfowl breeding habitat that is relatively stable compared to prairie breeding grounds. Such stability may be particularly important during times of widespread drought on the prairies, which has occurred in recent years. Habitat development techniques proposed for the ''Taking Wing" program include (1) wetland vegetation management and construction of new wetlands, (2) upland nesting cover enhancement and creation of secure nesting habitat in the form of islands, nest structures, or predator-proof enclosures, and (3) provision of additional permanent water areas and establishment of animal and vegetative foods. These techniques, which have been developed and tested primarily in prairie habitat, may attract and hold breeding pairs, increase nest success, and enhance brood survival, respectively. Unfortunately, little is known about which of these factors may limit waterfowl production in Rocky Mountain breeding habitat. Applying any management technique without first identifying the limiting factor is a waste of valuable resources. In 1989, 4 wetland complexes in Routt National Forest were investigated to determine which phase of the waterfowl reproductive life cycle inhibited production (Ringelman et aI., 1989). Their �222 results indicate that low brood survival was the major cause of low duck recruitment. In 1990, we used the same survey techniques to determine inhibiting factors on 6 new wetland complexes within Routt National Forest. Our specific objectives were to (1) determine the species composition, abundance, and habitat use by waterfowl breeding pairs in 6 wetland communities on Routt National Forest, (2) use the ratio of lone males to pairs as an indication of nesting chronology and nest loss/renestlnq activity, (3) measure duckling survival rates, hatch to fledging, and habitat use by brood-rearing hens, and (4) identify management options to improve waterfowl recruitment. STUDY AREAS We surveyed 6 wetland communities within Routt National Forest along the western edge of a dry intermountain area known as North Park, Colorado (Figs. 1-7). Number, density, origin, and elevation of ponds varied for each study unit (Table 1). Other general characteristics unique to each study unit are described below. Dominant forest vegetation consisted of lodgepole pine (Pinus contorta), Engelmann spruce (~ engelmannii), subalpine fir (~ lasiocarpa), and quaking aspen (Populus tremuloides). Riparian zones were dominated by willows (Salix spp.), with smaller amounts of mountain alder (~~). Common aquatic vegetation included cowlily (Nuphar polysepalum), northern mannagrass (Glyceria borealis), grassleaf pondweed (Potamogeton gramineus), fineleaf pondweed (.E. filiformis), spiked watermilfoil (Myriophyllum spicatum), and sedges (~ vescaria and Q. utriculata). Predators which could prey upon waterfowl and their nests in the area include coyote (~ latrans), badger (Taxidea ~), mink (Mustela ~), pine marten (Martes americana), striped skunk (Mephitis mephitis) and northern goshawks (Accipiter gentilis). Newcomb Creek The largest, deepest ponds in this unit (Teal Lake, Tiago Lake, and Burns Reservoir) were accessible by a maintained road and were used extensively for recreational purposes. Camping in developed and undeveloped sites, fishing, and use of small boats at these lakes began in early June (Memorial Day), increased substantially by early JUly (Independence Day), and continued throughout the summer. Other ponds had various degrees of accessibility (main road, logging roads, and no roads) but were not used for recreational purposes. Clearcuts of pine (north of Newcomb Creek) and aspen (south of Newcomb Creek) were common and bordered several ponds. Runoff from snowmelt and the low summer precipitation were the only source of water for ponds in this unit. Thus, ponds were shallow, water level declined during the summer, and aquatic vegetation was prominent by late summer. Colorado Creek This unit is composed of two different segments, east and west. The western ponds were higher in elevation and were ice-free at least 2 weeks later than the eastern ponds. A high percentage of snow cover persisted in the western segment until early June. The lower elevation, eastern segment had been heavily clearcut adjacent to most ponds, and revegetation of both forest and herbaceous vegetation was minimal. Clearcuts skirted the western segment and some of the ponds along the perimeter, but the segment as a whole was uncut. Large amounts of forest debris from blowdown and cut trees that weren't removed was common in both segments. Cattle grazing began in late August in both segments, but was most pronounced in the clearcut areas of the eastern segment. Other than the main road, use of logging roads by recreationists was rare during the summer. As with Newcomb Creek, the water source of these ponds was limited to snowmelt and summer precipitation, resulting in decltninq water levels as the summer progressed. �223 Table 1. Compartative attributes of 6 study units located in Routt National Forest, Colorado, 1990. Size (km2) Study Unit Newcomb Creek Number of Wetlands Beaver Glacial Wetland Density Elevation (m) 5.16 3 47 9.7 2667 (west) 0.37 8 16 64.9 2804 (east) 0.54 5 28 89.2 2728 0.42 33 2 83.3 2658 (LPC) 0.35 27 4 88.6 2697 (BC) 0.35 26 0 74.3 2743 Upper Beaver Creek 0.35 35 13 137.1 2819 Lower Beaver Creek 0.68 55 1 82.3 2652 Total 8.22 192 111 36.9 Colorado Creek Livingston Park Lone Pine! Bear Creeks �224 . ," ~~~ .. u P f~ r t Lo •••.•cw- I , 41\11\& ~C\'IeC'" (.~c\< . " ,• \. .. - #., . . ", ' •..• ~". ·•·. .. , I I I ...•• •.. tt <I ~ , ... . " d C D .; " :z. ., . ( I ,, r . ~ I f~ ~U'( 5.) z ;,.Ioe\ \.u; \ At'ww~" I ~ :' . , , ., .". ---. -_ ....~ , •\ LC)\'\.( rih~ to 8'lcu.. ~ l.)e c\... C1"ee. \<.5 mo",,~ .'" rr, j . I ~ "' .••., o �:;.0 c-'l 0- o ~ 0 \) '0 ~c;; ~ c;?~ \j :r 225 a- 0 :r' (-L 0 . ~, II CJ �226 o -ee l- o - U 2t.e~ l-L �- ·2 -...J ,/ \ ....••. \ \ 227 \\ 1\ / \ I \ \ \ - \ \ \ \,. '\ "'laI �228 =z ~ v. -...._,/ 1 ~ cJ 6: ~ \ -.(. '-s C!) �,~ ~ ..1l ;r ~ r-~ ~ ~ :;:) <:» 0 :7" 0 ~ J ~ c1 ~ ':.. '::) C- C- cY ~ q~ ..•0 ;;r <, 229 �230 j f / / / / / rv <,.. J - a: u. �231 .MIBd 15 en "'C o o l.... 10 rn "t- O l.... CD .o E ::s Z 5 o 2 3 4 5 6 7 8 9 Week Figure 8. Week of hatch for duck broods found on Routt National Forest, Colorado, 1990. Week 1 corresponds to 28 May - 3 June. 10 11 �232 Livingston Park This unit is a beaver complex located in the back of an broad, open, steep sided boxed valley which opens to North Park proper in the east. In addition to snowmelt and surface ground water from precipitation, these ponds are fed by a small creek draining the mountains to the west. The bottom of the valley to the east of the main beaver complex was intermittently saturated with runoff and surface ground water, but contained few ponds. A portion of the Grizzly-Helena trail system bordered the main beaver complex, but the unit was generally isolated and undisturbed. Cattle were present on the unit in August, but numbers were unknown. Lone Pine/Bear Creeks This unit could also best be described as 2 segments separated by a large ridge: Lone Pine Creek to the south and Bear Creek to the north. Both segments had a constant source of runoff , water available to most ponds. The Lone Pine segment is a beaver complex that is accessible by a road along its north side. Recreationists in this segment were common but not necessarily abundant. It appears that beaver activity in Lone Pine is minimal. The complex is a series of ponds with a braided creek flowing through them, rather than a series of well maintained beaver ponds. Beaver dams are degenerating and willows are revegetating the complex. The Bear Creek segment consists of a series of small beaver complexes located in a narrow, forested, steep-sided valley. Access is limited to rough trails, resulting in little or no disturbance during the early summer months. Several of the ponds that were not fed directly by Bear Creek were dry. Upper and Lower Beaver Creeks The 2 units fed by Beaver Creek have similar wetland and upland habitats, and were considered separate units only because of a small separation between them and ease of sampling. Each is characterized by a large complex of stable beaver ponds, although the upper unit does contain some distinct, relatively large glacial ponds. Good roads border each unit, and logging roads and trails exist in the upper unit. The road along the north border of the lower unit is directly adjacent to ponds and is used consistently by logging trucks. Except for traffic on the main roads, disturbance to both units is probably minimal. Clearcuts are common near both units (especially the glacial ponds of the upper unit), but are not directly adjacent to the main beaver complexes of either unit. Cattle were present near the west border of the upper unit in mid August but did not enter the unit during this study. METHODS Abundance and Success of Nesting Pairs We surveyed Newcomb Creek, Colorado Creek, Livingston Park, and Lone Pine/Bear Creeks weekly from 16 May - 21 June, and Upper and Lower Beaver Creek weekly from 6-21 June. For each pond visited we recorded time, species present, sex, number of lone males, lone females, pairs, grouped birds, and whether birds were flushed. Ponds were approached silently to avoid flushing birds and duplicating counts. Surveys began during late morning (0800-1000) because incubating female otten take incubation breaks during early morning; surveys that consistently counted these females would result in lone male to pair ratios that were not indicative of nesting chronology or success. Changes in lone male to pair ratios were used as an index to nest success. We located nests opportunistically and recorded species, number of eggs, egg fertility, incubation stage, nest fate, and general habitat characteristics. �233 Brood Survival Surveys to locate and follow broods through fledging began on 25 June. Survey times included the first and last 4 hours of daylight (approximately 0500-0900 and 1700-2100). Only wetlands that were consistently used by breeding pairs were surveyed for broods. High priority ponds (those receiving highest pair use) were checked weekly, whereas lower priority ponds were checked intermittently. Lone Pine and Bear Creeks were dropped from brood surveys due to insignificant use by breeding pairs. We again approached ponds quietly, then remained until we observed broods or believed that the observation time was adequate to have detected broods had they been present (usually 15-90 minutes). For each brood, we recorded species, brood size, brood age (Gollop and Marshall, 1954), presence or absence of the hen, and pond number. Since broods were not marked, re-identification was based on records of previous locations, and brood size and age. We attempted to relocate all broods once a week. RESULTS Species Composition and Abundance We observed 8 species of ducks during the nesting season (Table 2). Estimated numbers of pairs decreased from 104 to 68 (35%) after the first week, remained relatively stable for the next 3 weeks, then dropped again during the final 2 weeks. The progressive decrease in pair numbers was similar among species, but was most pronounced at Newcomb Creek and Livingston Park but not Colorado Creek or Lone Pine/Bear Creeks (Table 3). Because ducks were probably migrating through the area during the first week of pair surveys, we averaged the estimated number of pairs for weeks 2-4 to obtain estimates of number of resident breeding pairs for each species (Table 4). Of the 92 estimated pairs of breeding ducks, the majority were mallards (Anas platyrhynchos: 51%) and ring-necked ducks (Aythya collaris; 21%), followed by bufflehead (Bucephala albeola; 13%) and green-winged teal (Anas crecca; 10%). This method underestimated pairs known to have hatched broods for ring-necked ducks at Newcomb Creek and buffleheads at Colorado Creek, Livingston Park, and Upper Beaver Creek. Breeding pair surveys for ring-necked ducks may be biased, because this species did not begin incubation until after surveys ended. Additional waterfowl species observed on the study unit were a pair of sandhill cranes (~ canadensis) at Colorado Creek, a molting male wood duck (Aix sponsa) at Lower Beaver Creek, and several pairs and non-breeding groups of Canada geese (Branta canadensis). The largest group of Canada geese were 14 birds observed at Livingston Park on 25 May. No geese were seen after 7 June. Breeding Chronology Estimated date of hatching was estimated by backdating duckling age using published information on duckling development (Gollop and Marshall 1954). Date of nest initiation was predicted by subtracting mean incubation period (Klett et at. 1986) from the date of hatch (Table 5). Fledging chronology was estimated using values from Bellrose (1976). Nest initiation had already begun upon our arrival on 15 May. Mean date of nest initiation was earliest for mallards and latest for ring-necked ducks, with 5 weeks separating them. Peak of hatch occurred during the last half of June for mallards and green-winged teal, and during mid-July for ring-necked ducks and buffleheads. The resulting hatching curve had a bi-modal distribution of hatching dates (Fig. 8). Mallard hatch spanned 8 weeks, whereas hatching of ring-necked ducks, bufflehead, and greenwinged teal nests spanned only 4, 4, and 3 weeks respectively. Mean date of hatch was latest at Colorado Creek for mallards and ring-necked ducks, and nearly so for bufflehead. Mallards at Upper and Lower Beaver Creeks hatched earlier than in the other 4 units. �234 Table 2. Weekly changes in estimated number of pairs for 8 species of ducks found on Routt National Forest, Colorado, 1990. Numbers with parentheses include Upper and Lower Beaver Creeks to facilitate comparisons between weeks. Species May 16-19 May 22-25 May 28-31 June 1-7 June 11-14 June 18-21 Mallard 42 34 40 30 20 7 (57) Ring-Necked Duck 32 15 16 (48) 20 (19) Bufflehead 16 11 10 8 4 4 1 0 2 2 1 0 0 Total 104 68 76 (108) 0 ( 0) ( 1) 0 ( 0) ( 0) 0 ( 0) 50 (73) 33 (39) 1 ( 1) 0 1 ( 1) ( 0) 0 1 1 1 ( 1) ( 0) ( 2) American Coot ( 2) 2 0 2 6 ( 8) ( 3) ( 0) Common Merganser (7) 3 3 0 5 ( 5) (11) ( 3) Northern Pintail (14) 11 7 3 14 (17) (13) (11) American Wigeon (19) 13 9 4 (8) (24) (15) Green-Winged Teal (30) ( 1) 71 (101) �235 Table 3. Weekly changes in estimated number of pairs found on 6 study units in Routt National Forest, Colorado, 1990. Numbers in parenthesis exclude Upper and Lower Beaver Creek to facilitate comparisons between weeks. May 16-19 May 22-25 May 28-31 June 1-7 Newcomb Creek 54 29 32 32 17 7 Colorado Creek 25 22 21 29 20 15 Livingston Park 17 11 15 9 8 9 8 6 8 1 5 2 Upper Beaver Creek 9 17 14 4 Lower Beaver Creek 23 13 9 2 108 (76) 101 (71) 73 (50) 39 (33) Study Unit Lone Pine/Bear Creeks Total 104 (104) 68 (68) June 11-14 June 18-21 �236 Table 4. Estimated number of breeding pairs for 7 species of ducks found on 6 study units in Routt National Forest, Colorado, 1990. Species Mallard Newcomb Colorado Livingston Creek Creek Park Lone Pine\ Bear Creeks Upper Beaver Creek Lower Beaver Total & Creek (Percent) 14 13 5 3 5 7 47 (51) Ring-Necked Duck 6 6 3 2 1 1 19 (21) Bufflehead 7 1 1 0 2 1 12 (13) Green-Winged Teal 1 2 1 1 1 3 9 (10) American Wigeon 2 0 1 0 0 0 3 (3) Common Merganser 0 1 0 0 0 0 1 (1) American Coot 1 0 0 0 0 0 1 (1) 31 23 11 6 (34) (25) (12) (6) Total Percent 9 12 (10) (13) 92 �237 Table 5. Nesting chronolgy of 5 species of ducks found on Routt National Forest, Colorado, 1990. Species Nest Initiation Mean Date of Incubation Commencment Hatching Fledging Number of Broods Ma11ard 24 May 31 May 25 June 20 August 21 Ring-Necked Duck 21 June 25 June 21 July 10 September 18 Bufflehead 31 May 4 June 3 July 24 August 10 9 June 2 July 6 August 4 Green-Winged Teal 2 June American Wigeon 31 May 1 June 26 June 14 August 1 Common Merganser 10 June 15 June 12 July 15 September 1 American Coot 11 June 17 June 9 July 3 September 1 �238 Pair Density Overall pair density averaged 11.2 pairs/krn", but ranged from 6.0 pairs/krn" at Newcomb Creek to 25.7 palrs/krrr' at Upper Beaver Creek. Newcomb Creek and Lone PlnexBear Creeks had < 10.0 pairs/krn", whereas Colorado Creek, Livingston Park, and Upper Beaver Creek had > 20.0 palrs/km". Because these wetland complexes tend be isolated areas with variable pond densities, pairs per pond is a more meaningful statistic than pairs/krrr' to compare units. Newcomb Creek had the highest pairs per pond (0.62), followed by Colorado Creek (0.40), Livingston Park (0.31), Lower Beaver Creek (0.21), Upper Beaver Creek (0.19), and Lone PinevBear Creeks (0.10). Nest Success .. We did not conduct nest searches, and therefore needed to rely on indices of nest success. Lone male to pair ratios increased throughout the survey period for all species except mallards (Table 6). Interpretation of lone male to pair ratios for ring-necked ducks is questionable because they had just begun nesting at the time surveys were completed. A peak in mallard lone male to pair ratios occurred just before peak incubation commencement, indicating some nest failure occurred during early incubation. In addition, high brood sighting/pair ratios occurred for all species except mallards, further indicating some nest failure occurred for mallards. A consistent increase then steady decline in lone male to pair ratios suggests that nest failure and subsequent repairing of lone males was uncommon among dabbling ducks. Nest 'success appears to be high (6080%) for all species except mallards (40-60%). We only found 3 duck nests incidental to our surveys. A green-winged teal at Colorado Creek hatched all 8 fertile eggs of a 9 egg clutch on 26 June. We found a freshly depredated mallard nest at Lower Beaver Creek on 8 June, which had at least 5 incubated eggs. Another mallard nest bowl with no down or eggs was also found at Lower Beaver Creek on 8 June. Other waterfowl nests included 2 Canada goose nests on beaver lodges at Newcomb Creek and a sandhill crane nest on an island at Colorado Creek. We believe all 3 of these nests hatched, but the sandhill crane lost its young during brooding-rearing. Duckling Survival We found a minimum of 56 broods of 7 species of ducks (Table 7). Number of broods in each area generally reflected estimated breeding pairs (Table 4). However, 51% of the breeding pairs were mallards whereas only 37% of the broods were mallards. In turn, the proportion of ring-necked duck and bufflehead broods was greater than their respective breeding pairs. Also, the proportion of broods at Newcomb Creek (41%) was greater than the proportion of breeding pairs (34%) using this unit. Mean brood size decreased between la and Ib age classes for all species, and continued at a steady decline through class III for bufflehead (Table 8). After the initial decrease, mallards and ring-necked ducks remained relatively constant. Mean brood size for ring-necked ducks again declined at the IIc age class, and average brood size even of young broods was low for this species. The notable increase in size of class IIc and III mallards is a result of relocating several large broods. Data on mean brood size were less reliable for mallards than for other species, because we rarely located mallard broods in 2 consecutive age classes. However, when mallard broods were sighted on 2 consecutive age classes after classes la-Ib, loss of young was recorded on only 1 of 11 occasions. Total Production We estimated waterfowl production using observed and estimated number of young fledged (Table 9). Numbers from observed values represent a minimum because both pair and brood surveys probably underestimated true numbers. To derive our total production estimates we used �239 Table 6. Weekly changes in lone male to pair rat~os for 7 species of ducks found on Routt National Forest, Colorado, 1990. Species May 16-19 May 22-25 May 28-31 June 1-7 June 11-14 June 18-21 Mallard 0.8 :1 1.3 :1 3.7 : 1 3.0 :1 3.2 :1 3.0 :1 0.4 : 1 0.9 : 1 0.3 : 1 0.5 : 1 1.0 : 1 1.8 : 1 0.2 : 1 0.4 :1 0.3 :1 1.0 :1 3.7 :1 4.0 :1 0.4 : 1 1.0 :1 4.5 :1 10.0 : 1 6.0 :1 8.0 :1 American Wigeon 1.0 : 1 0.0 : 1 0.5 :1 0.0 :3 0.0 :2 1.0 :1 Northern Pinta i1 0.0 :0 1.0 : 1 0.0 :0 0.0 :0 0.0 :0 0.0 :0 Common Merganser 0.0 :2 1.0 : 1 1.0 :1 1.0 :1 1.0 :1 0.0 :0 Total 0.5 : 1 0.9 :1 1.6 : 1 1.5 :1 2.3 : 1 3.3 :1 Ring-Necked Duck Bufflehead Green-Winged Teal �242 Table 9. Minimum production estimates for 7 duck species using observed and estimated number of young fledged. Species No. of Fledged Young Broods Estimated Young Number of Young Fledged ger km2 Per Pair Obs. Est. Obs Est. 10 51 140 6.2 17.0 1.1 3.0 Ring-Necked Duck 7 21 55 2.5 6.7 1.1 2.9 Bufflehead 7 18 25 2.2 3.0 2.1 2.1 Green-Winged Teal 1 1 3 0.1 0.4 0.1 0.3 American Wigeon 1 1 1 0.1 0.1 0.3 0.3 Common Merganser 0 a 2 0.0 0 .. 2 0.0 0.2 American Coot 1 7 4 0.8 0.5 7.0 4.0 27 99 230 12.0 28.0 1.1 2.5 Mall.ard Total �243 the following species-specific values: a. b. c. Breeding Pairs = Breeding pair estimates (Table 4) Nest Success = 50% for mallards, 70% for other species Duckling Survival = Mean size of class liB broods (Table 8) therefore: Total Production = axbxc We observed 27 broods age lib or older, totaling 99 young. The estimated potential recruitment from all areas (230 young) was 2.3 times greater than that observed. Observed and estimated production differed substantially for mallards and ring-necked ducks but not for buffleheads. Minimum/Maximum production for each species varied from 0.0-0.2 younq/krrr' for common merganser to 6.2-17.0 younq/krn" for mallards. Observed number of young fledged per pair was greatest for Upper Beaver Creek and Newcomb Creek (Table 10). DISCUSSION Species composition was similar to that found by Ringelman et at. (1989) on Routt National Forest in 1989, except we found no cinnamon teal (Anas cyanoptera), and mallards composed > 10% of the breeding pairs during 1990. Gadwall (Anas strepera) are a common nesting species in Colorado but were not found during this study. In contrast, ring-necked ducks and bufflehead are uncommon breeders elsewhere in Colorado, but were common in Routt National Forest. Gadwall and cinnamon teal prefer more open areas for nesting, whereas bufflehead are commonly found in forested montane wetlands throughout it's range. Ringelman and Kehmeier (1989) first documented a successful nesting pair of bufflehead in Colorado just a year ago. Our results show a range extension for nesting bufflehead, however the range of breeding buffleheads in Colorado is unknown. Mean date of nest initiation was 3-4 days later than that observed in 1989 (Ringelman et al. 1989) for early nesting species (green-winged teal and mallard), but 14-15 days later for late nesting species (bufflehead and ring-necked duck). A similar pattern was evident in the mean date of hatch, but to a lesser degree (2-3 days and 8-10 days, respectively). Wetlands became available to waterfowl 1-2 weeks later in 1990 than in 1989, which could account for the relatively later nest initiation dates. Clutch sizes were less than those reported for prairie-nesting habitats (Klett et al. 1986). Our estimate of 11.2 pairs/km'' was almost double that of 1989 (6.1 pairs/km"), and also exceeded the 1.4 pairs/km" reported by Rutherford and Hayes (1976) and the 0.6 palrs/km" observed by Frary (1954) in forested habitats elsewhere in Colorado. However, number of pairs/pond was similar on the Routt National Forest during 1989 (0.32 pairs/pond) and 1990 (0.30 pairs/pond). Wetland availability could account for the differences in pairs/pond among study units. Newcomb Creek was one of the first areas available (ice-free) to breeding ducks, whereas Colorado Creek was the last. In addition, Bear Creek not only was ice free late, but its location in a steep valley may have made it unattractive to waterfowl. Based on the generally increasing lone male to pair ratios and the high proportion of pairs for which broods were observed, nest success appears to be relatively high in Routt National Forest. Similar results were recorded in 1989. Predators were common in the area, but appeared not to have affected nest success to a great degree. Availability of alternate prey or habitat preferences of foraging predators could account for the lower nest success of mallards than later nesting species. Our minimum estimate of 48% duckling survival is slightly less than the 54% found in 1989 on Routt National Forest. This estimate is only a minimum because of several factors: (1) mallards and green-winged teal tended to be secretive and commonly moved their broods to unknown locations. �244 Table 10. Minimum production estimates for 5 study units using observed number of young fledged. Study Unit Number of Fledged Young Broods Number of Young Fledged per km2 per Pair Newcomb Creek 12 45 5.5 1.4 Colorado Creek 5 21 2.5 0.9 Livingston Park 3 9 1.1 0.8 Upper Beaver Creek 5 14 1.7 1.5 Lower Beaver Creek 2 10 1.2 0.8 27 99 12.0 1.1 Total �245 (2) ring-necked ducks and buffleheads broods tended to disperse and merge by class liB. and (3) we only sampled a portion of the available wetlands. Our estimate fell within the large range of 27-81% duckling survival rates found by others (Ball et at. 1975, Ringelman and Longcore 1982, Talent et at. 1983, Duncan 1986). However, our estimate of 3.7 young fledged per brood was less than the 5.05.5 found in the above studies and the 5.6 ducklings per brood for mallard and green-winged teal found by Frary (1954) in montane habitat of northwest Colorado. Compared to last year, we fledged 10% fewer young per brood, but also appeared to have hatched 22% fewer young per clutch. Thus, although total production of young per brood was less in 1989, percent duckling survival was greater. Reasons for relatively low duckling survival rates are unknown. Mortality rates may have increased when emergent cover became dry, thus making ducklings more susceptible to mammalian predators. Another more likely possibility is that ducklings became separated from adults and died from exposure during the frequent cold or stormy nights. Lone ducklings were commonly observed during brood surveys. To what extent this is a natural occurrence or due to vegetation hindering brood cohesiveness is unknown. Why buffleheads experienced a continuous decline in mean brood size while other species did not is also unknown. Buffleheads usually remained on the same pond throughout brood rearing while other species commonly used several ponds. Predator search images may have focused on ponds in which they were previously successful. In addition, bufflehead ducklings are dependent upon aquatic invertebrates for food. Competition between fish and ducklings for this limited food has been implicated in brood studies of the closely related goldeneye (Bucephala spp.). MANAGEMENT OPTIONS The phase of the reproductive cycle that most inhibited recruitment varied by species and area. Livingston Park, Lone Pine Creek, and Bear Creek did not appear to attract and hold breeding pairs. Nest success appeared to be low for mallards and good for other species. Duckling survival was high for mallards but generally low for ring-necked ducks and buffleheads. Management techniques to attract and hold breeding pairs are generally associated with creating new wetlands, providing additional nest structures or areas, and improving the overall wetland environment. Techniques to improve nest success include habitat predator control and management to improve or provide secure nesting cover. Predator management is not considered a management option for 2 major reasons: (1) it is usually costly, logistically difficult, and largely ineffective in the long term unless continued indefinitely, (2) techniques we used in this study weren't designed to quantify predator effects, thus implementation of predator management without further research could be a waste of time and money. Habitat management to improve food and cover for ducklings are the management techniques most often used to increase duckling survival. The management options discussed below are tailored to the portion of the Routt National Forest covered in this study. Although they may vary in economic and logistic efficiency, all should result in increased waterfowl production. General Treatments The following 3 management options could be applied to all study units. If one or more of these options is selected, we urge that (1) no more than one treatment be applied to a wetland complex, (2) portions of the treated unit(s) be set aside as a control area(s) with no treatment, and (3) that a follow-up evaluation be made to assess the results of the treatment. 1. To reduce separation of ducklings on ponds with dense cowlily growth, an aquatic herbicide (such as Rodeo, a trademarked product from Monsanto) could be applied to produce open travel lanes through some ponds. Most ponds with active beaver colonies naturally maintain such travel lanes. However, most ponds currently choked with cowlily �246 are probably too shallow or lack food to maintain beaver populations throughout winter months. 2. The dominant emergent plants in the area typically grow in less than 0.3 m of water (Windell et aI., 1986). Consequently, they are very susceptible to drying in glacial ponds by the time brood rearing commences and are unlikely to become established in beaver ponds, which tend to have deeper shorelines. To increase cover and food during brood rearing, one might introduce an emergent that has a wider range of water depth tolerances. A promising candidate is hard-stem bulrush. It is currently growing in at least 3 areas in North Park and Routt National Forest. If this treatment is implemented, we suggest planting root stocks during spring or fall at rates of 1,000 plants/acre (Lemberger, 1981). Planting is most easily accomplished by firmly shoving rootstocks into the pond substrate. Establishing hard-stem bulrush by seeding may be more difficult because of the need to scarify seeds (spring seeding) or control water levels (fall seeding) (Harris and Marshall 1960), and because the colder sediment temperatures at this elevation may cause poor germination. 3. Buffleheads are present throughout the study areas. No natural nest sites were located during casual searches this season. Nest boxes erected in early spring also received no documented use. However, surveys conducted near larger wetlands typically used by buffleheads revealed very few natural nesting cavities. Given the age of structure and species composition of the surrounding forest, it is likely that a lack of suitable nest sites limits the breeding population of buffleheads. Rejection of nest boxes this year may be related to the strong homing traditions of buffleheads and/or the relatively few nest boxes installed. Consequently, installation of additional nest boxes is a reasonable strategy to augment breeding populations of this species. Gauthier (1988) found that buffleheads readily use artificial nesting boxes. Maximum dimensions of 15 x 15 x 40 cm with a 6.5 cm hole are recommended (Fig. 9). Several boxes around a pond, placed 3-7 m high and close to the wetland edge, will probably bring best results. Boxes should be placed at larger ponds and ponds with records of pairs that did not produce broods. Colorado Creek The Colorado Creek study unit is virtually devoid of upland nesting cover in clearcut areas. We recommend seeding grasses in clearcut areas adjacent to wetlands to stabilize soils and provide nesting cover for waterfowl. Cattle exclosures should be erected to allow proper development of newly seeded areas and assure residual vegetation is present in early spring. Newcomb Creek This complex has generally high waterfowl production, so care must be taken not to overdo management treatments. Some ponds on Newcomb Creek are in the later stage of wetland succession, and consequently fill with emergents by mid-spring and dry out thereafter. This may attract nesting ducks to the wetland, but then strand broods without suitable rearing habitat. Pond #44 in this unit is an example of a wetland that could be deepened in one or two locations to provide brood habitat and a refugia for aquatic invertebrates during dry periods. Other open water areas could be created using heavy equipment or explosives in select wetlands. �247 Lone Pine Creek Most of the Lone Pine Creek wetland complex consists of old, inactive beaver ponds with failing dams. One approach to rejuvenate this area would require several steps. First, the water levels could be reduced by breaching old dams. Next, the area would be allowed to revegetate into willow and aspen stands. Later, beaver could be reintroduced and maintained by a limited harvest strategy. New beaver complexes are more productive of waterfowl foods and have greater habitat diversity, thus this management strategy would increase the potential for waterfowl production. Livingston Park Livingston Park has high potential for an increased number of wetlands. Currently, the main wetland complex covers only about 30% of the open valley floor. Norris Creek along the south edge of the valley and runoff and surface ground water from a large watershed could supply water for additional, stable wetlands. After a mid-summer rainstorm, six previously dry beaver ponds filled virtually overnight, and the valley floor became saturated with ankle deep water from runoff which persisted beyond the end of the study. If this undependable source of water could be supplemented with water from Norris Creek in the early spring, a stable complex of wetlands could persist throughout the summer months. Beaver are already plentiful in the area and would probably create wetlands if a water source were available, thus keeping the number of man-made impoundments, and cost, to a minimum. Once stable beaver impoundments became established, timing and amount of water from Norris Creek would be targeted to assure persistence of ponds through the waterfowl production season. The exact methods to create this wetland complex are beyond the scope of this study. Among others, consideration of the drainage patterns of the valley and amount of water to maintain additional wetlands would have to be addressed. This management treatment is a large undertaking, but would result in the creation of significant new waterfowl habitat. LITERATURECITED Ball, I. J., D. S. Gilmer, L. M. Cowardin, and J. H. Reichman. 1975. Survival of wood duck and mallard broods in north-central Minnesota. J. Wildl. Manage. 39:776-780. Bellrose, F. C., (ed.). 1976. Ducks, geese, and swans of North America. Stackpole Books, Harrisburg, Pa. 543pp. Coward in, L. M., A. B. Sargeant, and H. F. Duebbert. 1983. Problems and potentials for prairie ducks. Naturalist 34:4-11. Duncan, D. C. 1986. Survival of dabbling duck broods on prairie impoundments in southeastern Alberta. Can. Field-Nat. 100:110-113. Frary, L. G. 1954. Waterfowl production on the White River Plateau, Colorado. M.S. Thesis, Colorado State University, Ft. Collins. 93 pp. Gollop, J. B. and W. H. Marshall. 1954. A guide for aging duck broods in the field. Unpublished report, Mississippi Flyway Council Tech. Sec., 14 pp. Gauthier, G. 1988. Factors affecting nest-box use by buffleheads and other cavity-nesting birds. Wildl. Soc. Bull. 16:132-141. Harris, S. W. and W. H. Marshall. 1960. Germination and planting experiments on soft-stem and hard-stem bulrush. J. Wildl. Manage. 24: 134-139. �248 Klett, A. T., H. F. Duebbert, C. A. Faanes, and K. F. Higgins. 1986. Techniques for studying nest success of ducks in upland habitats in the prairie pothole region. U. S. Fish Wildl. ServoResour. Publ. 158. 24 pp. Lemberger, J. J. 1981. What brings them in? Wildlife Nurseries, Oshkosh Wisc. 32 pp. Ringelman, J. K., and J. R. Longcore. 1982. Survival of juvenile black ducks during brood rearing. J. Wildl. Manage. 46:622-628. Ringelman, J. K., M. A. Willms, and R. S. Langley. 1989. Waterfowl abundance, production, and habitat use on the Routt National Forest, Colorado. Unpublished field report, Colorado Division of Wildlife, Fort Collins, CO. Ringelman, J. K., and K. J. Kehmeier. 1990. Buffleheads breeding in Colorado. Colo. Field Ornith. 24(2):46-47. Rutherford, W. H. and C. R. Hayes. 1976 Stratification as a means for improving waterfowl surveys. Wildl. Soc. Bull. 4:74-78. Talent, L. G., R. L. Jarvis, and G. L. Krapu. 1983. Survival of mallard broods in south-central North Dakota. Condor 85:74-78. Windell, J. T., B. E. Willard, D. J. Cooper, S. Q. Foster, C. F. Knud-Hansen, L. P. Rink, and G. N. Kiladis. 1986. An ecological characterization of Rocky Mountain montane and subalpine wetlands. U.S. Fish Wildl. ServoBioI. Rep. 86( 11). 298 pp. �249 Colorado Division of Wildlife Wildlife Research Report September 1991 JOB PROGRESS REPORT State of Colorado Project W-152-R-4 : Job 22 Work Plan Job Title: _2_ Migratory Game Bird Publications Period Covered: Author: Avian Research - Migratory Game Birds 01 April 1990 through 31 March 1991 Michael R. Szymczak Personnel: James K. Ringe1man and Michael R. Szymczak, Wildlife Colorado Division of ABSTRACT The following list contains those articles that were prepared submitted for publication or published during this segment: and/or Gilbert, D. W., D. R. Anderson, J. K. Ringelman, and M. R. Szymczak. Analysis of duck nesting in relation to vegetation on the Monte Vista National Wildlife Refuge, 1964-90. Wildlife Monograph (In prep.) Ringe1man, J. K. 1991. Differential harvest of mallards under conventional bag and point system regulations. Wi1dl. Soc. Bull. 19:258267. Ringe1man, J. K. 1991. Evaluating and managing waterfowl habitat CoLorado . Co10. Div ..WUdl.,. Div. Rep. 16. 46pp. . . . in . Ringe1man, J. K., M: R. Szymczak, C. W. Jeske, and.K. E. Ragotzkie. Ulnar lipid as an indicator of depleted fat reserves in mal,lards..J. Wi1dl. Manage. (In review). Szymczak, M. R., and E ..A. Rexstad. 1991. Harvest distribution and survival of a gadwall population. J. Wild1. Manage. 55: (In press) . Willms, M., and J. K. Ringe1man. mallards. (In revision) Prepared by: ~ R, ~~ Michael R. Szymczak Wildlife Reasercher C Use of cattle feedlots by wintering � https://cpw.cvlcollections.org/files/original/00180de9a148002f0e3977d2e5105446.pdf 33fd921801c322366e51dd0985b3252c PDF Text Text .1 JOB PROGRESS REPORT State of: Colorado Upland Bird Research Project: 'W-167-R 'Work Plan: _1_: Job Title: Evaluation of Habitat Development for Ring-necked Pheasants in eastern Colorado Period Covered: Author: Job _1!L 01 January through 31 December, 1991 'Warren D. Snyder Personnel: C.E. Braun, T.E. Remington, and 'W.D. Snyder ABSTRACT This study remained in a planning phase through 1991. Some contractual efforts with farmers in extreme northeastern Colorado were conducted on a test basis during spring 1991. Farmers were paid to establish sorghum survival plantings for pheasants on small areas using a modified CHIP (Cooperative Habitat Improvement Program) contract. Quality of the plantings varied widely, but response of farmers was generally favorable. Time and manpower required in preparation of voluminous contracts and slow payment made the approach impractical. Alternate approaches were under review and development. Personnel also reviewed, analyzed, and developed recommendations concerning CHIP, CRP cost share, and other habitat development programs. A report was submitted to Division admi~istrators for review and approval. Prepared by: 1t;1.u.u,) I) ¥ 'Warren D. Snyder 'Wildlife Researcher ��3 JOB PROGRESS REPORT State of: Project: Colorado Upland Bird Research Y-167-R York Plan: 1__ Job Title: Farmin~ for Rin~-necked Pheasants - Pro~ram Development and Evaluation Period Covered: Job __~2=5 __ 01 January through 31 December 1991 Author: Thomas E. Remin~ton Personnel: C1ait E. Braun, Thomas E. Remington, Yarren D. Snyder, Colorado Division of Yi1d1ife. ABSTRACT Ring-necked pheasant (Phasianus co1chicus) populations in Colorado have declined markedly because of habitat loss and degradation resulting from intensive farming. Current cost-share programs do not provide incentives for landowners to develop habitat or reimburse them for income lost from land developed into wildlife habitat. Pheasant hunters have indicated a willingness to pay for hunting opportunity in numerous surveys and by their support of existing fee hunting areas. Ye developed and proposed a conceptual model for a community-based fee hunting program that increases funding for and development of, habitat, controls access, and compensates landowners at minimal cost to the Division of Yi1d1ife in dollars or FTE commitments. ��5 FARMING FOR RING-NECKED PHEASANTS - PROGRAM DEVELOPMENT AND EVALUATION ,, Thomas E. Remington INTRODUCTION Pheasant populations have steadily declined in Colorado over the last 30-40 years. While severe winters have been responsible for occasional die-offs, habitat deterioration caused by intensive farming has been implicated as the cause of long term declines. Preservation or development of pheasant habitat excludes that land from crop production and, thus, represents an economic loss to farmers. While Division of Wildlife (and some Federal) programs have shared the costs of habitat development, no programs have provided economic incentives to replace revenue lost on developed lands, or loss in yields from altering farming practices to benefit pheasants. Pheasant hunters provide significant economic benefits to farming communities by renting motel rooms" eating in restaurants, and by buying gas, groceries, etc. A~though most economic models assume this money "turns over" 2-3 times within these communities, it is unlikely that much, if any, money gravitates to farmers who actually provide the resources which are hunted. On the contrary, hunters (at their worst) represent a source of harassment by trespassing, leaving gates open, damaging fences, fields, etc. Habitat development programs funded by the Division of Wildlife, such as the Cooperative Habitat Improvement Program (CHIP), or the Pheasant Habitat Improvement Program (PHIP; Work Plan 1, Job 24), can begin to replace pheasant habitat losses. However, it is unlikely these programs can restore pheasant populations over a large area for two reasons. First, habitat losses are so extensive that it is unlikely the Division could allocate either enough money, or for a long enough period to completely replace them. Second, some landowners are unlikely to participate in habitat developments on their land or alter farming practices unless they derive an economic benefit from increased pheasant populations, or at least are compensated for lost income. A program to provide additional (non-CDOW) funding for habitat development through economic incentives to landowners to participate could complement existing programs and contribute substantially to improving pheasant hunting in Colorado. Nationally, both participation in, and acceptance of, fee hunting programs is increasing (Smith et al. 1992). Colorado pheasant hunters have indicated a willingness to pay for the opportunity to hunt pheasants through their support of community-based fee hunting programs (i.e., Burlington and Yuma), shooting preserves, and by their responses in the Upland Bird Questionnaire recently conducted by the Upland Bird Program (Upland Bird Management Guide, Colorado Division of Wildlife, 1992). Unfortunately, this payment has not resulted in significant habitat improvements. Revenue generated by the Burlington access fee program, sponsored by the Rotary Club, has been used in part to purchase game farm pheasants for stocking rather than for habitat development. �6 P. N. OBJECTIVES ,, Develop a program to link hunters willing to pay for pheasant hunting opportunity with landowners willing to 1) provide access for a fee, 2) develop habitat~for pheasants within the program area, and 3) amend farming practices to make them more compatible with production and survival of pheasants. SEGMENT OBJECTIVES 1. Meet with farm groups, Chambers of Commerce, sportsman groups (Colorado Wildlife Federation, Pheasants Forever, etc.), Cooperative Extension Service, Agricultural Stabilization and Conservation Service, and Division of Wildlife personnel to identify interest in, and willingness to support, a fee hunting/habitat program for pheasants. 2. If interest level is sufficient, develop draft guidelines for a prototype program in consultation with interested parties from these groups. 3. Disseminate draft guidelines to interested parties for review and amend as appropriate. 4. Present prototype program to appropriate Division personnel for approval to proceed on a trial basis. 5. Select area(s) for implementation and evaluation of prototype program. 6. Prepare detailed study plan for implementation and evaluation of program. RESULTS and DISCUSSION Segment objectives 1-4 were completed. The conceptual framework for a community-based fee hunting program and draft guidelines were presented to the Director's staff in November. Following that meeting the decision was made to defer implementing a prototype program for at least a year to avoid impeding progress on redirecting the CHIP program and implementation of the new PHIP program. In the interim, surveys will be conducted to ascertain hunter interest in a program of this type, the potential market for fee hunting under different fee rates and hunt qualities, and landowner interest and needs. PHEASANT COOPERATIVE PROGRAM The Pheasant Cooperative Program is designed to develop habitat and increase pheasant populations by overcoming several significant barriers to habitat improvement for pheasants and other farm wildlife in eastern Colorado; namely providing financial and other incentives for landowners to develop habitat, offering technical assistance, and generating funding. The Division, landowners, and sportsman will be equal partners in improving pheasant populations and hunting in Colorado. In addition, this program has �educational benefits, increases access for hunting, and should significantly ,, increase pheasant harvest and decrease landowner-sportsman conflicts over hunting. The program minimizes CDOY FIE commitments because sportsman do much of the legwork and habitat developments are completed by landowners and/or sportsmen. At least two communities in eastern Colorado have organized programs to match hunters willing to pay to hunt pheasants with landowners willing to provide access for a fee. These are Ringneck Raiders in Yuma, run by the Future Farmers of America, and Rooster Roundup, run by the Burlington Rotary Club. These programs provide and control access but do not improve habitat and consequently have not increased pheasant populations. Landowners do not receive any proceeds from these programs because they are used as fund-raisers for the sponsoring groups. Hunter satisfaction is poor and organizers may be tempted to stock pen-raised pheasants to the gun to retain hunter interest and dollars. Ye propose that the Division assist in the formation of Community Pheasant Cooperatives consisting of individual farmers and sportsman groups. Hunters would pay a fee to hunt lands within the Cooperative. Some of this money would be paid to the landowners, some would pay for habitat developments within the Cooperative, and a small portion would pay expenses of Cooperative administration. The Division of Yildlife would match hunter contributions for habitat development. Community groups, particularly those which benefit from hunting revenue such as gas stations, restaurants, motels, Chambers of Commerce, etc., would be encouraged to join the Cooperative as sponsors by donating $100, $250, $500, etc. for habitat development. The Division would match these contributions as well. The Division would contract with the Cooperatives while the Cooperatives would contract with the individual farmers. This greatly reduces CDOY FIE requirements to process contracts. Habitat developments would follow prescriptions described in the Pheasant Habitat Improvement Program (PHIP) and would be detailed in individual management plans prepared for each farm in the Cooperative at the outset. It is anticipated that this habitat development money can be used to leverage significant Federal Farm Bill money by paying the farmer's cost share on developments through the Conservation Reserve Program, Forest Stewardship Program, and other Farm Bill programs. Thus, the Division contribution may represent only 25% of total costs. The Cooperative would contract with individual farmers to provide required amounts and quality of habitat, and agree to alter specified farming practices to benefit pheasants (such as timing of stubble tillage in spring to prevent pheasant nest destruction). The Cooperative would also serve as a forum for educational efforts to increase landowner and sportsman knowledge about farming and habitat development practices beneficial to pheasants and other farm wildlife. Many details about how this program would be administered remain to be developed. We believe most details should be left to individual cooperatives since it will be their program. For the Division to participate and cost share developments, habitat plans would have to be developed and habitat quantity and quality goals would have to be met. Cooperatives would be run by a Board consisting of landowner representatives, sportsman representatives, and a CDOY representative. This board would make decisions about program implementation, monitor compliance, and make cash distributions (of hunter fee money). Examples of questions/answers about program implementation details are illustrated in Table 1. Actual programs may differ from this example. �8 Table 1. Questions/answers about Pheasant Cooperative Program implementation. ,, Q. What are significant advantages of this program? A. Provides monetary return to the farmer for access for hunting and burden of hunter control is shifted from the landowner. Farmers are compensated for developing habitat on productive farm ground and this compensation is from groups directly benefitting, i.e., hunters, motels, restaurants, etc. Will significantly increase pheasant populations and harvest at minimal cost to the Division in dollars and FTEs. Access and hunter participation should increase, albeit at a fee. Q. Where will money come from to develop habitat? A. From hunter fees (25-50% to habitat, remainder to farmers after costs). and community sponsors (motels, restaurants, Chambers of Commerce. Pheasants Forever. etc.). The Division of Wildlife would match these contributions. and the entire pool would be used to leverage Federal Farm Bill program $. Q. How much will hunters be charged? A. Rates will be set by individual cooperatives and influenced by supply and demand, but $15/day, $25/weekend. $75/season seem reasonable. Possibilities exist for embellishments such as reduced fees or a sliding scale for youngsters, discounts for mid-week, or surcharges for opening weekend. Q. What about enforcement? A. Additional enforcement to prevent trespass will likely be necessary, at least initially. It is anticipated this program will be self-policing to some extent, because both landowners and hunter participants have a vested interest to prevent trespass. Participating hunters will be identifiable by windshield/dashboard cards and by some form of identification worn on their backs or on their head to facilitate enforcement. Acreage enrolled will be intensively signed (provided by CDOW and erected by cooperative) and participants will be furnished maps. Q. What about liability to landowners and other cooperative members? A. They may be covered under state's general liability immunity (true for participants in South Dakota's pheasant program). If not, liability insurance would be obtained by the Cooperative and paid from fees generated. Q. How will money be paid to landowners equitably. and yet in a manner that encourages habitat and farming practices to benefit pheasants? A. We suggest half of the money to be paid to farmers be allocated based on the amount of land enrolled. The other half would be allocated based on what proportion of pheasants harvested came from that farmers' land, or what proportion of hunters hunted on his land. Evaluation The type, quantity, and distribution of habitat developments will parallel those in PHIP; thus, intensive evaluation of pheasant population responses will not be necessary as they can be inferred from evaluation of PHIP. Evaluation will focus on hunter participation, harvest per unit area and effort, and hunter and landowner satisfaction relative to pre-Cooperative levels and/or similarly farmed areas not in the Program. Acreage of habitat improvements and income generated for farmers will also be documented. �LITERATURE CITED ,, Smith, J. L. D., A. H. Berner, F. J. Cuthbert, and J. A. Kitts. 1992. Interest in fee hunting by Minnesota small game hunters. Wildl. Soc. Bull. 20:20-26 Thomas E. Remington Wildlife Researcher ��1 )~ JOB FINAL REPORT State of: Colorado Project: Y-37-RIW-152-R 'Work Plan: Job Title: 3 Job l3b Responses of Sage Grouse to Vegetation Fertilization Period Covered: Author: Upland Bird Research 1 July 1985 through 31 December 1991 Orrin B. Myers Personnel: C. E. Braun, Colorado Division of 'Wildlife; O. B. Myers, Colorado State University. ABSTRACT I integrated field and laboratory studies to estimate effects of nitrogen fertilization on nutritional quality of big sagebrush (Artemisia tridentata) for sage grouse (Centrocercus urophasianus). By manipulating sagebrush chemistry and morphology on 33 10-ha plots, I tested whether grouse use of 2 sagebrush subspecies was affected. I also used captive sage grouse to experimentally estimate forage preference rankings and nutritional quality. Nitrogen fertilization had large effects on sagebrush morphology and foliar crude protein. Protein content of foliage increased 30 - 52% following treatment but declined rapidly. Morphological changes were more persistent. Sage grouse increased their use of fertilized 'Wyoming big sagebrush (~. ~. xyomingensis, [AT'W])relative to unfertilized AT'Wbut not of fertilized mountain big sagebrush (~. ~. vaseyana, [ATV]). Elevated grouse herbivory rates on fertilized AT'Wwere more persistent than foliar chemistry effects and were correlated with effects on plant morphology. I allowed captive-reared sage grouse to develop feeding preferences for fertilized and unfertilized AT'W and ATV without the confounding influences of variable resource availability or predictability common in field studies. I conducted paired comparison preference trails with these birds and used one-way and factorial BradleyTerry models to estimate preference rankings and the effects of experimental treatments on rankings. Cap.tivebirds developed strong .. preferences for AT'W over ATV, like wild birds, but also increased their consumption of ATV when it was fertilized. I used birds captured from a population that specialized on AT'W during fall - spring and tested their ability to assimilate nutrients from fertilized and unfertilized ATV and AT'W. Assimilation coefficients for drY matter and energy were strongly influenced by the subspecies tested but not by fertilization. Nitrogen balance was not affect by fertilization or subspecies, despite substantial between treatment variation in nitrogen intake. These birds did not fully acclimate to captive conditions and mos.t replications were at sub-maintenance intakes. Although I used analysis models that are less sensitive to level of intake than traditional approaches; potential problems due to captive birds may have prevented detection of fertilization effects. ��,' DISSERTATION SAGE GROUSE HABITAT ENHANCEMENT: EFFECTS OF SAGEBRUSH FERTILIZA nON Submitted by Orrin Bruce Myers Department of Fishery and Wildlife Biology & Program for Ecological Studies In partial fulfillment of the requirements for the Degree of Doctor of Philosophy Colorado State University Fort Collins, Colorado Spring 1992 �14 COLORADO STATE UNIVERSITY March 10, 1992 WE HEREBY RECOMMEND THAT THE DISSERTATION PREPARED UNDER OUR SUPERVISION B:' ORRIN BRUCE MYERS ENTITLED SAGE GROUSE HABITAT FERTn.IZATION ENHANCEMENT: EFFECTS BE ACCEPTED AS FULFILLING FOR THE DEGREE OF DOCTOR OF PHILOSOPHY. Department Head OF SAGEBRUSH IN PART REQUIREMENTS �15 ABSTRACT OF DISSERTA nON SAGE GROUSE HABITAT ENHANCEMENT: EFFECTS OF SAGEBRUSH FERTILIZATION I integrated field and laboratory studies to estimate effects of nitrogen fertilization on nutritional quality of big sagebrush (Artemisia tridentata) for Sage Grouse (Centrocercus urOjlhasianus). By manipulating sagebrush chemistry and morphology on 33 lO-ha plots, I tested whether grouse use of 2 sagebrush subspecies was affected. I also used captive Sage Grouse to experimentally estimate forage preference rankings and nutritional quality. Nitrogen fertilization had large effects on sagebrush morphology and foliar crude protein. Protein content of foliage increased 30 - 52 % following treatment but declined rapidly. Morphological changes were more persistent. Sage Grouse increased their use of fertilized Wyoming big sagebrush (d. 1. wyomineensis, [ATW]) relative to unfertilized ATW but not of fertilized mountain big sagebrush (d. 1. vaseyana, [ATV]). Elevated grouse herbivory rates on fertilized ATW were more persistent than foliar chemistry effects and were correlated with effects on plant morphology. I allowed captive-reared Sage Grouse to develop feeding preferences for fertilized and unfertilized ATW and ATV without the confounding influences of �16 variable resource availability or predictability common in field studies. I conducted paired comparison preference trials with these birds and used one-way and factorial Bradley-Terry models to estimate preference rankings and the effects of experimental treatments on rankings. Captive birds developed strong preferences for ATW over ATV, like wild birds, but also increased their consumption of ATV when it was fertilized. I used birds captured from a population that specialized on ATW during fall spring and tested their ability to assimilate nutrients from fertilized and unfertilized ATV and ATW. Assimilation coefficients for dry matter and energy were strongly influenced by the subspecies tested but not by fertilization. Nitrogen balance was not affected by fertilization or subspecies, despite substantial between treatment variation in nitrogen intake. These birds did not fully acclimate to captive conditions and most replications were at sub-maintenance intakes. Although I used analysis models that are less sensitive to level of intake than traditional approaches, potential problems due to captive birds may have prevented detection of fertilization effects. Orrin B. Myers Department of Fishery and Wildlife Biology & Program for Ecological Studies Colorado State University Fort Collins, CO 80523 Spring 1992 �1./ TABLE OF CONTENfS TITLE PAGE - .. w • • • • • • • • • • • • • • • • • • • • • • • • • •• SIGNATURE PAGE 1 11 iii ABSTRACT OF DISSERTATION . . . . . . . . . . . . . . . . . . . . . . . . . . . .. v TABLE OF CONTENTS ACKNOWLEDG~S Vlll CHAPTER 1 EFFICACY OF NITROGEN FERTILIZATION FOR ENHANCING SAGE GROUSE HABITAT . 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . •. . 1 ME'I"H ODS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 INTRODUCTION Study area .•...•................•..•............. 3 Experimental design and treatment of plots •..........•......• 5 Vegetation and soil sampling and analyses Data analysis . .•.................................... RES'ULTS ni . S o il nitrate-nitrogen 7 . ........•........................ 5 9 9 Sagebrush response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 11 Sage Grouse response 15 DISCUSSION 20 �18 LI1"ERATURE CI1'ED . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 26 < APPENDIX 1 31 CHAPTER 2 BRADLEY-TERRY MODEL ESTIMATION OF PREFERENCE RANKlNGS FROM PAIRED COMPARISON EXPERIMENTS . . . . . . • . . . . . . . . . . .. INTRODUCTION ~ODS 35 . . . . . . . . . . . . • . . . . • . . . . . • . . . . . . . . . . • .. ..•.•.........•.........••..•....•.••... 35 0 37 Paired comparison preference trials . • . . . . . . . . . . . . • . . . . . . . .. 37 Statistical analysis 39 Bradley-Terry models . . . . . . . . . . . . . . . . . . . . . . . . • . . . .. 39 Parameter estimation and hypothesis testing 41 ..•.......•..•.. RESULTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 43 DISCUSSION 46 ...........................•.•..•...... LI1"ERATURE CI fED . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 50 APPEND IX 2 •.........•......•.....••..•.••..••.••. S2 CHAPTER 3 EXPERIMENTAL ESTIMATION OF SAGE GROUSE FEE.DIN'G PREFERENCES ......•.•..•...•.•••. INTRODUCTION CI • • • • • • • • • •• 55 . . . . . . . . . . . . . . . . . . . . . . . . . . . • . . . • . . . .. 55 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 56 Preference trials . . . . . . . . . . . . . . . . . . . . . • . . . . . . . . • . . . .. 56 Statistical analysis 58 ME11iODS .....•..•..........•..••••.....•.. RESULTS Bradley-Terry model analysis General linear model analysis eo.. ..........•.........•...•.. • • • • • • • • •• 60 60 62 < �DISCUSSION 64 LrrERA TURE CI'TED . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 67 CHAPTER 4 EFFECT OF NITROGEN FERTILIZATION ON THE NUTRITIONAL QUALITY OF BIG SAGEBRUSH FOR SAGE GROUSE 69 INTRODUcrI'ON . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 69 ME'l'HODS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . • . .. 70 Feeding trials 70 Statistical analysis 73 RESULTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 75 DISCUSSION 80 LrrERA TURE CI'TED . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 87 �20 ACKNOWLEDGMENTS Support for this work carne from the Colorado Division of Wildlife, the U. S. Bureau of Land Management, the Department of Fishery and Wildlife Biology and its Alumni, the Program for Ecological Studies, and a Colorado Mined Land Reclamation Fellowship. The Arapahoe National Wildlife Refuge, U. S. Fish and Wildlife Service provided facilities to house captive Sage Grouse. I am especially thankful to refuge staff Eugene C. Patten, Raymond N. Varney, and Terri L. Follett for keeping my coffee mug filled. Colorado Division of Wildlife District Wildlife Managers, Kirk F. Snyder and Stephen H. Porter, came through numerous times to help me out of sticky situations and provided hours of welcome entertainment. Drs. Gary C. White and Clait E. Braun served on my graduate committee as co-advisers. I also benefitted from the service of Drs. Richard L. Knight and David M. Swift on my graduate committee. My friends and fellow graduate students gave me much support and stimulation: Robert E. Bennetts, Ada C. Fowler, Frank P. Howe, Leslie A. Robb, Michael A. Schroeder, and last but not least Marilet A. Zablan were the best at this. �< CHAPI'ER EFFICACY OF NITROGEN ENHANCING 1 FERTILIZATION FOR SAGE GROUSE HABITAT INTRODUCfION Sage Grouse, Centrocercus urgphasianus, are sagebrush (Artemisia spp.) obligates (patterson 1952). Over the last half century sagebrush control programs have impacted > 10% of the 20 million ha of this vegetation type on public lands (Vale 1974) and an unknown proportion of the estimated historical total of 38 million ha (Blaisdell et al. 1982). During the same period Sage Grouse also apparently have declined, at least in overall distribution (Braun 1987). Sagebrush stands have been physically removed by plowing, burning, chaining, and herbicide use (Vale 1974, Braun 1987). Although declining numbers of Sage Grouse have been associated primarily with decreasing landscape coverage of sagebrush, reduced quality of remaining sagebrush stands for maintaining grouse populations may have exacerbated the areal loss. Livestock at high stocking density have been cited as having deleterious effects on quality of Sage Grouse and other wildlife habitats (Cooperrider 1990). Habitat fragmentation that occurs concomitant to sagebrush removal operations also has likely reduced the capacity of remaining sagebrush areas to support grouse. Another human-mediated process capable of lowering quality of Sage < �22 Grouse habitats is disturbance caused by human activities, particularly those caused by extraction of petroleum or mineral reserves (Vale 1974, Braun 1987). In many areas inhabited by Sage Grouse, open pit mining operations and other habitat alteration activities can remove or fragment Sage Grouse habitats. Many of these areas are critical habitats for Sage Grouse, especially when sage-covered ridges and benches are affected. available forage in winter. These wind-swept, exposed sites often offer the only Forage quality may not be uniform among or within sites however, as variation in plant chemistry between those plants used and not used as forage by grouse has been reported (Remington and Braun 1985). Part of this observed selectivity was attributable to subspecies of big sagebrush used by grouse. Birds fed mainly on Wyoming big sagebrush [ATW]) and little on mountain big sagebrush CA. t. vaseyana CA. tridentata) CA. t. wyomingensis [ATV]). The population studied by Remington and Braun (1985) also showed significant feeding preferences for plants having higher foliar crude protein and lower monoterpene levels. Although preference studies are not reliable for inferences about herbivore nutritional requirements, one could speculate that grouse select plants on the basis of protein and that increased foliar crude protein could increase their reproductive success (Beckerton and Middleton 1982). Alternatively, foliar monoterpenes or other secondary plant compounds could act as anti-herbivory compounds, thereby lowering the nutritional quality of ATV relative to ATW and causing some plants to be avoided. In either case, plant primary and secondary chemistry may be influenced by resources available to plants (Bryant et al. 1984). Thus, this could represent < �opportunities for improving nutritional quality of sagebrush if resources available to plants can be manipulated. I tested whether nitrogen fertilization of sagebrush altered forage selection patterns of free-ranging Sage Grouse. I examined the effects of nitrogen fertilization on soil nitrate and foliar crude protein dynamics under field conditions. I also examined the responses of free-ranging Sage Grouse to habitat manipulation with fertilizer as a first step in evaluating efficacy of fertilization as a habitat enhancement, mitigation, or remediation tool in sagebrush shrubsteppe. METHODS Study area I conducted my study in North Park, Jackson County, Colorado (40· 50' _. 106· 20'), where Sage Grouse and coal, oil, and natural gas extraction activities cooccur. North Park is an intermountain basin at an elevation of about 2,500 m. It is drained to the north by the North Platte River, which is fed by many smaller streams. Topography is flat to rolling with numerous ridges and benches. Climate is cold and dry with an average annual frost-free period of 46 days. Sagebrush-dominated grasslands cover upland sites, and grasses and sedges occur in native and irrigated meadows that border drainages. Artemisia tridentata is the dominant sagebrush species and includes 2 subspecies, A. 1. wyomin~ensis and A. 1. vaseyana. Other species of sagebrush occurring with limited distribution in North Park are A. lon~iloba, A. sana. and A. auil10sa (Beetle 1960). �24 MODERATE HIGH -., LOW JACKSON COUNTY • FERTILIZED D CONTROL 1 km FIg. 1-1. Study plot locations in North Park, Jackson County, Colorado. Locations were stratified into areas expected to have low, moderate, and high levels of grouse use. �25 Experimental desien and treatment of plots In October-November < < 1985 granular ammonium nitrate fertilizer (33.5% N) was applied at a rate of 112 kg-N/ha to 33 fertilizer treatment plots using conventional agricultural spreaders. Locations for 33 experimental blocks were chosen by restricted randomization among 3 strata (Fig. 1-1) predicted as having low, moderate, and high levels of winter use by Sage Grouse (C. E. Braun, personal communication) within the McCallum anticline area of North Park (Beekly 1915) where petroleum development has occurred. The 11 blocks in each stratum were 20 ha in size. Their dimensions in the low and high-use strata were about 183 x 1094 m. In the moderate-use stratum, blocks were about 366 x 547 m due to constraints of terrain. The long axis of each block was east-west. Each block was divided in half along the long axis and fertilization treatments (fertilized or control) were assigned at random to each half. In October 1987 fertilizer was applied to 6 additional randomly-located l-ha treatment plots. Unfertilized plots were not true controls, because fertilizer spreaders did not traverse the plots. Vegetation and soil sampling and analyses Focus for the study was on early-spring forage selection by grouse to coincide with the period when maternal nutrition might most likely affect Sage Grouse reproduction. I sampled 12 randomly-selected blocks in the 4 years subsequent to the first growing season response to fertilization; the springs of 1987-90 correspond to years 1 - 4 of the experiment. May. Sagebrush foliage was sampled in mid-April to early Equal numbers of each subspecies were sampled along transects in control and fertilized plots of randomly-selected study blocks. Sampling points were at random �26 distances along transects. Stems and leaves from 5 plants of each subspecies at each level of fertilization were clipped, bagged, and placed on dry ice or snow, and then stored in a freezer. Once leaves were separated from stems, equal amounts from each plant were pooled and ground in a mortar and pestle under liquid nitrogen and frozen until chemical assays were performed. Kjeldahl nitrogen (Horwitz 1980) was measured to estimate foliar crude protein (% nitrogen x 6.25) in addition to neutral detergent fiber (NDF) and acid detergent fiber (ADF) (Van Soest 1963,a,h; Mould and Robbins 1981; Van Soest 1982). Gross energy estimates were obtained by bomb calorimetry. Coumarin content of sagebrush leaves was indexed by percent transmittance following Welch and McArthur (1986). Sage Grouse use of study blocks was estimated by examining individual plants for evidence of feeding activity prior to removing stems for analysis of foliar chemistry. The characteristic appearance of leaves that have been fed upon by grouse was described by Remington and Braun (1985). I scored plants as 1 if fed upon by grouse or 0 if not. I described sagebrush growth responses by measuring new growth of stems on plants. Three stems were measured from plants along 20-m transects in 5-m segments. Density within l-m wide strips and canopy cover along transect lines also were measured in the 3 blocks I sampled in each stratum. I scored each transect segment as 1 if ~ 1 plant was fed upon by grouse or 0 if no plants had been used by grouse. �27 I used the 6 I-ha blocks fertilized in 1987 to estimate the response of soil nitrogen to fertilizer treatment. Two experimental blocks located on Bosler loam and 4 were on Morset loam (Fletcher 1981). In June of each year 6 subsamples from the A and B horizons were pooled from each subplot. Soil chemistry analyses were performed by the Colorado State University Soil Testing Laboratory, Fort Collins (Workman et al. 1988). Data analysis Foliage samples came from random subsets of the blocks available, thus experimental units were not completely independent between years nor were repeated measurements made on the same experimental units in each year. I used nonlinear least-squares regression models to describe responses of sagebrush foliage to fertilization following Potvin et al. (1990). My global model for jointly describing the data and for testing hypotheses about the responses of the 4 sagebrush treatments over time consisted of each treatment level being estimated by its own exponential decay model of the form: (1.1) I systematically reduced the parameter space of the global model to test hypotheses about the responses of sagebrush and to obtain more parsimonious models. I assessed the significance of the reductions in parameter space by computing E-ratio tests between full and reduced models (Zar 1984): �28 << (1.2) ResSSlt is the residual sums of squares obtained by fitting the reduced model having q parameters, ResSSpis the residual sums of squares obtained by fitting the full model having p parameters (p > q), and n is sample size. I reduced the exponential models to linear least-squares models to test for linear trends in time. Parameter estimates from curve-fitting exercises can have ecological interpretations beyond that available from inferential methods (potvin et ale 1990). I constructed tests about parameter estimates and computed E-ratio tests or Student's 1 tests for the equality of 2 parameters (Zar 1984). My times series data consisted of winter-spring foliage chemistry for years 1-4 after the initial growth response by sagebrush. I coded year as (year-I) before analyses to aid in interpretation of parameter estimates. My observations of Sage Grouse herbivory on plants formed binary responses suitable for maximum likelihood logistic regression of the probability of grouse herbivory on subspecies and fertilization status of plants (pROC CATMOD, SAS Institute, Inc. 1988; Hosmer and Lemeshow 1989). I also scored each pooled foliage sample as 1 if ~ 1 plants were fed upon by grouse or as 0 if no plants in the pooled foliage sample had been fed upon by grouse. Pooled foliage samples assayed for crude protein represented 5 plants for each combination of block, subspecies, and fertilizer treatment level. Standard logits of binary responses were regressed on subspecies and foliar crude protein to estimate their effects on probability of grouse herbivory using maximum likelihood logistic regression where �L':1 (1.3) and where fJ is the probability of grouse herbivory on at least 1 plant in a pooled sample of 5, given the crude protein content (x) of the foliage. I collected soil samples from the same experimental units in each year and used repeated measures analysis of variance (ANDV AR, PRDC GLM, SAS Institute, Inc. 1988) when analyzing soil nitrate-nitrogen data. I used univariate ANDV AR hypothesis tests and multivariate ANDV AR profile analysis to examine the response of soil nitrate-nitrogen to fertilization treatments during years 1-3. Mauchly's criterion was used to test the assumption of compound symmetry for the variancecovariance matrix, and Greenhouse-Geisser (G-G) corrected probability values were used to establish the significance of tests if the variance-covariance asymmetrical. matrix was The G-G tests were the most conservative of the adjusted E tests (SAS Institute, Inc. 1988). RESULTS Soil nitrate-nitrocen Prior to fertilization in October 1987 (Year 0) control and treatment plots contained similar nitrate-nitrogen levels (f = .363, Fig. 1-2). The following June treatment plots contained about 9x the nitrate-nitrogen levels of controls. By year 2 nitrate-nitrogen on fertilized plots decreased to 170% of controls and decreased to 122% of controls in year 3. Averaged over 3 years, soil nitrate-nitrogen was �30 increased by fertilization (f = .043, Table 1-1), with the size of this effect changing over time (f = .031). Maucbly's criterion test for sphericity of the variance- covariance matrix rejected (f = .023) due to high variance in year 1, so probabilities for within subject tests were based on Greenhouse-Geisser adjusted F tests. The magnitude of the treatment effect declined (f = .038) between the first and second years after fertilization, however, fertilized plots still retained greater (f = .025) soil nitrate-nitrogen levels than controls in year 2. By the third year after treatment soil nitrate-nitrogen returned to control levels (f = .587). Other soil parameters showed little variation in response to fertilization (APPENDIX 1). ,,--.... E 00....._.... 10 CONTROL 8 FERTILIZED Z w C) 0 6 ~ ~ Z 4 I w ~ « ~ 2 ---...-~ ~ Z 0 0 2 1 __ ..._.. •••••• 3 YEAR Fig. 1-2. Nitrate-nitrogen (::r ± SE) dynamics of soils in control and fertilized plots in North Park, Jackson County, Colorado. << �31 Table 1-1. Univariate repeated measures analysis of variance of soil nitrate-nitrogen measured from f-ha study plots. Probabilities for within subjects effects were based on Greenhouse-Geisser adjusted E tests. Source of variation df F MS Between subjects Blocks 5 4.27 0.59 .709 Fertilizer 1 52.56 7.30 .043 Error 5 7.19 2 20.84 5.24 .066 10 4.57 1.15 .439 2 33.15 8.33 .031 10 3.98 Within subjects Years Years*blocks Years*fertilizer Error Sa,eebrush response Fertilization produced dramatic changes in plant structure so that treated plots were readily identifiable without the aid of plot markers when I was on foot, in vehicles or in fixed-wing aircraft. Fertilized plots were easily recognizable during the first and second years after treatment, less so in the third year, and were difficult to identify without plot markers and a compass in the fourth year. The annual growth increment of sagebrush in the first year more than doubled, and number of reproductive structures increased many times (personal observation), which apparently contributed to the visual differences between fertilized and control plots. for fertilized ATV (FATV) was 7.1 em + 0.2 New growth (SE), which was about 240% of < ' �32 controls, compared to 7.3 em ± 0.3 for fertilized ATW (FATW) and 2.9 em ± 0.2 for ATW. Density of sagebrush plants (! = 3.3 plants/rrr', f = .639) and canopy cover were not altered by fertilization (! = 11%, ~ = .815). Total plant height of ATV, excluding reproductive structures, increased from 26.8 to 31.8 em after fertilization ~ = .005), but height of ATW was not affected (! = 23.4 em .f = .560) Total big sagebrush cover differed between strata ~ = ± 0.9, .031) with the expected low-use stratum having 15% canopy cover compared to 10% in the moderate-use stratum and 7% in the high-use stratum. Cover values for each subspecies differed between strata as density for each subspecies varied among strata ~ < .001). Density of ATV in the low-use stratum averaged 3.1 plants/m" compared to 1.1 and 0.6 plants/nf in the high and moderate-use strata, respectively. The low-use stratum contained the lowest density of ATW (! = 0.5 plants/m~ = 2.6, compared to the other strata ~ Licb = 2.0). Total density of big sagebrush did not vary among strata (.f = .734). Crude protein in the foliage of unfertilized ATV and ATW varied between r = 0.10, Fig. years with a weak linear trend (.f = .004, more crude protein (.f < .001, I = 12.45%, SE SE = 1-3), but ATW contained 0.26) than ATV (X = 10.14%, = 0.22). Gross energy content of ATV (! = 23.80 kJ/g) was greater than ATW (! = 23.04 kJ/g), but NDF re = .003) ex = 0.266, f = .125) and ADF (! = 0.181, .f = .124) were similar for each subspecies. Transmittance at 364 nm was higher for ATW (23.6%) than in ATV (10.9%) indicating higher ~ < .001) coumarin levels in ATV relative to ATW . < < �33 _.... 24 ATV ~ 0 o ~20 ...__ ATW A FATV FATW -.- z A w 16 r0 a::: CL w 0 :::> 0:: 8 .------_. ---------'-".~~+ ~-----~ ------------~ -~$ U a::: « 4 _J 0 u, o 1 2 4 3 YEAR Fig. 1-3. Foliar crude protein levels (x % DM spring in years subsequent to fertilization. ± SE) of big sagebrush during early Foliar crude protein in fertilized sagebrush subspecies increased and then declined exponentially in the years subsequent to fertilization (Fig. 1-3). Fertilizer treatments increased foliar crude protein by 52% for ATV and 30% for ATW in the first year of response (Fig. 1-3). In each year, but year 2, fertilized sagebrush contained more crude protein than unfertilized sagebrush ~ < .03). In year 2 FATW contained more protein that ATW ~ = .038). Two jointly estimated functions for FATV and FATW gave a better fit to the data ~ = 0.80) than a single function ~ < .001, Table 1-2). Exponential decay �34 Table 1-2. Parameter estimates for nonlinear models describing responses of sagebrush to fertilization. < < Model Full Parameters al (Ji A; Reduced a· I Estimate 95% CI r FATV 10.03 8.67 - 11.39 0.80 FATW 13.64 12.49 - 14.80 FATV 7.24 5.43 - 9.06 FATW 8.48 6.79 - 10.17 FATV 1.59 0.41 - 2.76 FATW 2.36 0.34 - 4.38 FATV 10.02 9.03 - 11.01 FATW 13.69 12.71 - 14.68 Treatment level (J FATV &.FATW 7.84 6.61 - 9.07 x FATV &.FATW 1.95 0.89 - 3.00 models not only gave a good summary of the data cr - also can have meaningful ecological interpretations. 0.79 0.80), but the parameters The asymptotic level of crude protein in foliage was estimated by a. Thus, a test of the hypothesis that aj = a was equivalent to testing LTV = L1W. Asymptotic protein levels in foliage were not the same for FATV and FATW (Table 1-2, :e < .(01). These levels did not differ from the average level for ATV (f = .561) but did for ATW (f = .032). The initial effect of fertilization on each subspecies was estimated by {J. Under the null hypothesis {Ji = 0, and if {JFATV ¢ {JPA1Wthere subspecies x fertilizer interaction. was evidence for a I found no evidence of significant subspecies x fertilizer interaction effects (Table 1-2, :e = .327), and fertilization increased each subspecies by 7.84 percentage points over unfertilized controls <:e < .001). Hypothesis tests about a and {J resemble between-subjects tests in an ANOV AR, while A described the rate of decline in foliar crude protein and was �analogous to ANOV AR within-subjects tests. The 2 subspecies I tested had the same exponential decay constant (Table 1-2, f = A71). Precision of A was low (cv = 37 - 43%), however, relative to the other parameters (cv = 7 - 16%), and power to detect a difference between ~ of this size (0.77) was only about 11 %. I formed a reduced model for predicting the functional responses of FA TV and FA TW by estimating common a and P and separate A for each subspecies (Table 1-2). The reduced model did not differ from the full model in the amount of explained variation ~ = .263, r = 0.79). The form of the response by FATV and FATW, therefore, differed only in the predicted asymptotic protein levels. Foliar crude protein in FA TW was a predictable function of ATW from adjacent control plots and time since fertilization (f < .001, r = 0.753). This relationship was (1.4) FATW - ATW + 9.066e-1.547(ycar-l). Fertilization did not influence estimated gross energy ~ = .844), or ADF ~ = .370) of sagebrush. <e = .468), NDF Although I detected significant site effects (Blocks, f = .014) on coumarin levels, these differences were not related to sampling strata ~ = .880). Sage Grouse response During my 4 years of monitoring rates of Sage Grouse herbivory on sagebrush plants the proportion of plants used by grouse in a season varied between 4 and 8 % • Total herbivory rates on ATW plants (0.120 ± 0.016) were higher than on ATV < < �36 < <. 0.35 ... ,.-, ATV ATW ---0--FATV ~FATW -6-- 0.30 >- 0::: 0.25 0 > co 0.20 • 0:::: W I L-.J 0.15 rn 0 0::: 0.10 0.. 0.05 __.. .. 0.00 1 ::::===:::--. 3 2 4 YEAR Fig. 1-4. Probability of sage grouse herbivory (J> years subsequent to fertilization. (0.012 ± 0.005, :f ± SE) on big sagebrush plants in < .001). No trend was evident in total herbivory rates for either subspecies during the study ~ = .315). Although total herbivory did not change, shifts in foraging locations of grouse were inferred from the observed rates on plants in fertilized and control plots (Fig. 1-4). Herbivory rates were greater on FATW than on ATW in year 1 ~ = .004) and year 2 (f < .001) but not in year 3 ~ = .063) or year 4 (f = .508). Sampling effort for estimating herbivory rates was not intensive, so only differences >0.25 could be detected with 95% probability at ex = 0.05. Increased sampling effort of n = 80, 120, and 250 plants in each treatment group would have �.)1 been necessary to detect differences of 0.2, 0.15, and 0.10, respectively with the same probability. For ATW and FATW, I partitioned the 3 df from year effects into single df contrasts. The average herbivory rate for FATW during years 1 to 3 differed from the herbivory rate in year 4 ~ = .039). The same contrast for ATW was not significant ~ = .492). The average herbivory rates in year 1 and 2 were not different from the average rate for year 3 and 4 for FATW ~ ~ = .075) or ATW .186). Herbivory rates on ATV and FATV were low and no annual or treatment effects could be detected (f > .1). (1 - = e c Power to detect these effects was low 0.30). I summarized annual patterns of Sage Grouse herbivory rates on ATW and FATW using linear and quadratic logistic regression models. The model containing terms for intercept, year, and treatment x year, where year = to, 1, 4, 9}, = produced the best fit to the data (likelihood ratio goodness-of-fit X:4 1.53, f = .822). The joint model for predicting probability of herbivory on ATW and FATW was: 1Pr (herbivory I year, treatment) - 1- 1 1 + e -1.Gl2 • -O.IlS)'etr 1 1 + e -3.176 • O.079)'etr ' 1.42, £ = I observed no strata x treatment interaction (x22 = , FATW (1.5) ATW .491) effects on herbivory rates, but total herbivory on ATW and FATW did vary among strata (x21 = 19.46, £ < .001). Total herbivory on the ATW subspecies was not related to �38 <- 1.00r--------------ATV ,...., ••••••••• 0.80 .. .. ... """": ..... :o---..., ' ' ....•. ....• ....••. ....'. . ATW ' .........,... 0.60 ......' 0.40 ~ >o > .........~ 0.20 ,;' •...•~ ,--... 0:: ....... ..•... . ,. »" ___ ........• --. ........• ~. ........... .. .•....~ ,. • Ii•• ••• , co O.OO~~---~---~--~---~--~ W 1.00r----------------~-___, 0:: I "-" rn o g: 0.80 .'.' .. ..... .' ' 0.60 .. .' .. ." 0.40 ..••.' ' ...... .. .' .' ........ ,-: ' ' ............. ,.... 0.20 •.•..•. ........... 0.00~-8~~~10----~12-----1~4-----1~6----~18 FOLIAR CRUDE PROTEIN (% DM) Fig. 1-5. Predicted herbivory rates for sagebrush samples from experimental plots. Dotted lines are 95 % confidence intervals. The approximate probability for a single plant is P(O.2). ' �Table 1-3. Parameters for logistic prediction of sage grouse herbivory rates from foliar crude protein (Fig. 1-5). The approximate probability for a single plant is p15. Likelihood ratio goodness of fit Parameters Model 111 110 2 Subspecies: ATV -4.658 0.218 ATW -4.879 0.353 -6.873 0.496 Combined f>~ .701 .685 plant density when density of ATW was substituted for strata (likelihood ratio goodness of fit X-2 = 5.43, f = .066). The high-use level of strata had an observed probability of grouse herbivory of 0.063 (SE = 0.020), significantly lower than what I observed in the moderate-use (p = 0.188 + 0.033, f = (p = 0.118 ± 0.027, £ = .(03) and low-use strata .002). I used crude protein estimates obtained from pooled foliage samples and Sage Grouse herbivory scores to estimate the probability of herbivory in relation to foliar crude protein (Fig. 1-5). The 2 subspecies model was adequate for joint prediction of subspecies-specific herbivory rates on ATV and ATW plants (likelihood ratio goodness of fit, f = .700, Fig. 1-5). However, this model was indistinguishable from a reduced model that predicted herbivory rates for both subspecies with a single function ~ = .281, Fig. 1-5, Table 1-3). For both models, prediction of the probability of herbivory on individual plants was approximated by P(0.2) since analyses were conducted on pooled foliage samples. �40 DISCUSSION There is little dispute that sagebrush shrubsteppe has been altered by human activities and that further impacts will continue (Vale 1974, Braun 1987). One potential means for manipulating habitat quality for Sage Grouse is use of nitrogen fertilizer. Fertilization treatment used in this study produced changes in soil nitrate- nitrogen and sagebrush chemistry and structure that caused an increase in the use of treated sites by grouse. The rapid decline in soil nitrate-nitrogen likely was due in part to increased growth increments and increases in foliar crude protein. Movement of nitrogen deeper into the soil profile and uptake of soil nitrogen by grasses, forbs, and soil microbes also contributed to loss of nitrate-nitrogen (Schimel and Parton 1986). Mineralization of organic nitrogen should have continued to provide inorganic nitrogen for sagebrush in later years, which probably accounts for significant elevations of crude protein in sagebrush foliage 4 years after treatment. Variation among sites occupied by ATV and ATW could account for different asymptotic FA TW levels from ATW but not for FATV and ATV. Nitrogen mineralization varies spatially and temporally with soil moisture, temperature, and the availability of other nutrients controlling nitrogen mineralization at small spatial and temporal scales (Burke 1989). Although ATV and ATW plants sometimes were separated by < 1 m, they tend to segregate by soil depth and topography (Burke et al. 1989). ATW tends to occupy sites with warmer soil temperatures and shallower soils relative to ATV (West 1983) and could be limited more by soil moisture than ATV. Fertilized plots tended to capture and hold more �"'t.J.. snow than control plots, which could delay onset of spring growth but provide additional soil moisture. Total herbivory rates on sagebrush did not change during the study. One interpretation of this observed constancy is that numbers of grouse on the study area were not affected by fertilization. My estimates of herbivory rates did not include an estimate of the amount of biomass removed; thus increased rates of tissue removal from individual plants could not be detected. I cannot evaluate, therefore, whether grouse population size on the study area changed in any way as a result of fertilization. Birds shifted foraging activities at least within scales of S 300 m, as shown by the higher herbivory rates on fertilized plots relative to controls. Fertilization may have shifted grouse foraging activities at greater spatial scales, but I did not have data on Sage Grouse behavior before and after treatment upon which to base an evaluation. The probability of grouse herbivory on ATW plants was a predictable function of foliar crude protein. Remington and Braun (1985) used crude protein and plant vigor in discriminant function analysis on pooled plant samples to discriminate between browsed, unbrowsed, and random plants, but discrimination was poor. By substituting predicted foliar crude protein levels in fertilized sagebrush based on unfertilized plant chemistry and time since treatment (1.4) into the model that predicts herbivory rates from foliar crude protein (Table 1-3), I summarized expected results obtained by fertilizing sagebrush having different levels of crude protein (Fig. 1-6). The increase in the probability that a sagebrush plant was used as forage by grouse may be greatest on sites where crude protein is lowest prior to fertilization, but the �42 .,--.... w (f) 0.15 ~ .....__.., 0::: CL. o W 0.10 0.05 (f) « w 0::: o U 1 Z 10 4 Fag. 1-6. Predicted increase in herbivory rates on sagebrush plants in relation to prefertilization foliar crude protein (% DM) and time after treatment. benefits may also decrease most rapidly at these sites. However, effects of fertilization may still be > 0 after 4 years, regardless of initial conditions. Although the probability of herbivory on plants was related to crude protein, patterns of herbivory I observed over time had a qualitatively different response: the response of foliage and herbivory rates predicted on the basis of foliar crude protein was concave, whereas the measured herbivory response was convex. The general appearance of fertilized sagebrush over the 4 years varied in a way that was consistent with the observed herbivory response, which suggests that plant morphology is an important determinant of forage selection. Remington and Braun (1985) previously detected a correlation between their subjective assessment of plant vigor and grouse �43 use. Red Grouse lLagopus la~2Pus scoticus) also increased their use of fertilized sites ' over unfertilized sites that differed both in plant appearance and chemistry (Miller 1968). Fertilization altered the appearance and chemistry of both subspecies, but only ATW was used regularly by Sage Grouse. Remington and Braun (1985) suggested that ATV was unsuitable forage due to plant secondary chemistry. One may speculate that increased dietary nitrogen produced by fertilization could overcome certain negative effects of plant secondary compounds (Remington 1990). Why then were FATV plants not used more by grouse? Variation in sagebrush chemistry among plants (Remington and Braun 1985, Welch et al. 1988), populations and subspecies (Welch and McArthur 1979, 1981), and species (Kelsey and Shafizadeh 1979) is not unusual, but plant chemistry has not always correlated well with feeding preferences (White et al. 1982, Hupp 1987). Variation exists also in the sagebrush taxa used by Sage Grouse as forage. For example, Sage Grouse studied by Hupp (1987) used only ATV and Sage Grouse studied by Welch et al. (1988) fed upon ATV, ATW, and A. 1. tridentata. Forage use efficiencies have been correlated with grouse feeding preferences (Remington 1990). These observed preferences could be a result of localized adaptations to variation in forage quality ranldngs or could be due to more dynamic physiological adaptations to forage quality rankings that are conditional upon forage availability. For the North Park Sage Grouse I studied, ATW was, on average, more available to grouse than ATV. ATW tends to occupy sites more likely to be exposed during winter. If there is a transient cost associated with dietary switching (Levey and Karasov 1989), foraging preferences for ATW should have �44 evolved and my failure to detect a fertilization effect in FA TV would be caused by confounding influences of availability and not because forage quality was not adequate. Even if availability was even, time-lagged responses of grouse due to experience and learning could have been important (Wiens et al. 1986). Prediction of grouse use areas from earlier studies with other objectives were not adequate for predicting intensity of grouse use. Designation of grouse use areas apparently was a function of sampling intensity, although spatial shifts in grouse-use areas could have occurred. Pretreatment description of grouse use area and grouse herbivory rates probably would have expedited more effective placement of experimental units and would have provided a background upon which to judge the effectiveness of fertilization for increasing the total amount of foraging on the study area. The predicted low-use stratum may provide the best opportunity of the strata I used for continued testing of Sage Grouse-sagebrush foraging relationships. The stratum had the highest density of ATV, yet received the highest degree of herbivory by grouse. Numerous small patches of ATW that were a few metres in diameter occurred among the ATV. These ATW patches were completely covered by snow in some years, thereby reversing the typical availability ranking for the 2 subspecies. Additional study of grouse foraging dynamics on fertilized and unfertilized sites in this strata could reveal increased use of FA TV with time. Estimation of the relationships between Sage Grouse herbivory and plant chemistry in this and other studies (Remington and Braun 1985, Hupp 1987) were limited by pooling plant samples before chemical assays. Nested sampling designs �would better allow assessment of plant chemistry and grouse foraging at the scales of plants, sites, subspecies, and species and would aid identification of the appropriate units for management actions. Fertilization of sagebrush shrub steppe with 112 kg-N/ha sufficiently altered soil and plant chemistry to the extent that Sage Grouse exhibited localized shifts in foraging sites to treated areas. Whether or not fertilization caused the total number of birds using the study area to increase could not be assessed due to insufficient information on pretreatment conditions. Treated sagebrush foliage contained increased levels of crude protein 4 years after treatment. Sage Grouse herbivory rates on ATW were increased by fertilization but seemed to have no effect on ATV, although time lags in Sage Grouse responses could have prevented my detection of this effect. Whether use of fertilized sagebrush will have substantive nutritional benefits for Sage Grouse requires information on use efficiencies of forages and nutritional requirements of grouse. Fertilization was predicted to have the largest effect on Sage Grouse herbivory rates on sagebrush in areas where foliar crude protein was lowest. Regardless of initial conditions, fertilization should increase the probability of use of preferred subspecies for forage with benefits potentially lasting 4 seasons or more. Nitrogen fertilization therefore has potential for locally increasing the probability of feeding by grouse within areas already used by grouse. By fertilizing within Sage Grouse habitats, land managers might shift grouse-use areas away from impact areas. The increase in growth of fertilized sagebrush might also speed remediation of degraded sagebrush shrub steppe. < < �46 LITERATURE CITED Beckerton, P. R., and A. L. A. Middleton. 1982. Effects of dietary protein levels on Ruffed Grouse reproduction. Journal of Wildlife Management 46:569-579. Beekly, A. L. 1915; Geology and coal reserves of North Park, Colorado. U. S. Geological Survey Bulletin 596, 121pp. Beetle, A. A. 1960. A study of sagebrush-section Tridentatae of Artemisia. University of WYQmingAgricultural Experiment Station Bulletin 368. 83pp. Blaisdell, J. P., R. B. Murray, and E. D. McArthur. 1982. Managing intermountain rangelands-sagebrush-grass ranges. U. S. Department of Agriculture, Forest Service, General Technical Report INT-I34. 41pp. Braun, C. E. 1987. Current issues in Sage Grouse management. Proceedings of the Western Association of Fish and Wildlife Agencies 67:134-144. Bryant, J. P., F. S. Chapin, m, P. Reichardt, and T. Clausen. 1984. Adaptation to resource availability as a determinant of chemical defense strategies in woody plants. Recent Advances in Phytochemistry 19:219-237. Burke, 1. C. 1989. Control of nitrogen mineralization in a sagebrush steppe landscape. Ecology 70: 1115-1126. Burke, I. C., W. A. Reiners, and R. K. Olson. 1989. Topographic control of vegetation in a sagebrush steppe landscape. Vegetatio 84:77-86. Cooperrider, A. Y. 1990. Conservation of biological diversity on western rangelands. Transactions North American Wildlife and Natural Resources Conference 55:451-461. �'+1 Fletcher, L. A. 1981. Soil survey of Jackson County area, Colorado. U. S. Deparrnent of Agriculture, Soil Conservation Service, U. S. Government Printing Office 237-568/69, Washington, D.C, USA. 159pp. Horwitz, W., editor. 1980. Official methods of analysis of the Association of Official Analytical Chemists. Association Official Analytical Chemists., Washington, D.C., USA. 1018pp. Hosmer, D. W., and S. Lemeshow. 1989. Applied logistic regression. John Wiley and Sons. New York, N. Y. USA. 307pp. Hupp, J. W. 1987. Sage Grouse resource exploitation and endogenous reserves in Colorado. Ph.D. Thesis, Colorado State University, Fort Collins. 73pp. Kelsey, R. G., and F. Shafizadeh. 1979. Sesquiterpene lactones and systematics of the genus Artemisia. Phytochemistry 18:1591-1611. Levey, D. J., and W. H. Karasov. 1989. Digestive responses of temperate birds switched to fruit or insect diets. Auk 106:675-686. Miller, G. R. 1968. Evidence for selective feeding on fertilized plots by Red Grouse, hares and rabbits. Journal of Wildlife Management 32:849-853. Mould, E. D., and C. T. Robbins. 1981. Evaluation of detergent analysis in estimating nutritional value of browse. Journal of Wildlife Management 45:937-947. Patterson, R. L. 1952. The Sage Grouse in Wyoming. Sage Books, Denver, Colorado, USA. 341pp. �48 Potvin, C., M. J. Lechowicz, and S. Tardif. 1990. The statistical analysis of ecophysiological response curves obtained from experiments involving repeated measures. Ecology 71:1389-1400. Remington, T. E. 1990. Food selection and nutritional ecology of blue grouse during winter. Ph.D. dissertation. University of Wisconsin, Madison. 116pp. Remington, T. E., and C. E. Braun. 1985. Sage Grouse food selection in winter, North Park, Colorado. Journal of Wildlife Management 49:1055-1061. SAS Institute, Inc. 1988. SAS/STAT'" user's guide, release 6.03 edition. SAS Institute, Cary, North Carolina, USA. 1028pp. Schimel, D. S., and W. J. Parton. 1986. Microclimatic controls of nitrogen mineralization and nitrification in shortgrass steppe soils. Plant and Soil 93:347-357. Vale, T. R. 1974. Sagebrush conversion projects: an element of contemporary environmental change in the western United States. Biological Conservation 6:274-284. Van Soest, P. J. 1963.a,.Use of detergents in the analysis of fibrous feeds, I. Preparation of fiber residues of low nitrogen content. Journal of Association Official Agricultural Chemists 46:825-829. ___ . 1963}2. Use of detergents in the analysis of fibrous feeds. n. A rapid method for the determination of fiber and lignin. Journal of Association Official Agricultural Chemists 46:829-835. ___ . 1982. Nutritional ecology of the ruminant. 0 and B Books, Inc., Corvallis, Oregon, USA. 374pp. �Welch, B. L., and E. D. McArthur. 1979. Variation in winter levels of crude protein among Artemisia tridentata subspecies grown in a uniform garden. Journal of Range Management 32:467-469. Welch, B. L., and E. D. McArthur. 1981. Variation of monoterpenoid content among subspecies and accessions of Artemisia tridentata grown in a uniform garden. Journal of Range Management 34:380-384. Welch, B. L., and E. D. McArthur. 1986. Wintering mule deer preference for 21 accessions of big sagebrush. Great Basin Naturalist 46:281-286. Welch, B. L., J. C. Pederson, and R. L. Rodriguez. 1988. Selection of big sagebrush by Sage Grouse. Great Basin Naturalist 48:274-279. West, N. E. 1983. Western intermountain sagebrush steppe. Pages 351-374 in N. E. West, editor. Temperate deserts and semideserts. Ecosystems of the World Series. Volume 5. Elsevier, New York, New York, USA. White, S. M., J. T. Flinders, and B. L. Welch. 1982. Preference of pygmy rabbits (Brachyla~us idahoensis) for various populations of big sagebrush (Artemisia tridentata). Journal of Range Management 35:724-726. Wiens, J. A., J. T. Rotenberry, and B. Van Horne. 1986. A lesson in the limitations of field experiments: shrub steppe birds and habitat alteration. Ecology 67:365-376. Workman, S. M., P. N. Soltanpour, and R. H. Follett. 1988. Soil testing methods used at Colorado State University for the evaluation of fertility, salinity and trace element toxicity. Technical Bulletin LTB88-2. Colorado State University Cooperative Extension, Fort Collins, USA 29pp. ,' �50 Zar, 1. H. 1984. Biostatistical analysis. Second edition. Prentice-Hall, Englewood Cliffs, N. 1., USA 62Opp. �J.L < < APPENDIX 1 APPENDIX 1-1. Soil chemistry estimates from study plots in North Park, Jackson County, Colorado, at time of fertilization (Year 0) and in subsequent growing seasons (Years 1-3). Within subjects and between subjects effects tested were Years x Fertilizer and Fertilizer, respectively. Control Soil parameter Conductivity Cu Fe K MD NO~ Year l>f Fertilized I SE I SE 0 0.2 0.03 0.1 0.03 1 0.2 0.03 0.3 0.03 2 0.2 0.02 0.2 0.02 3 0.5 0.03 0.5 0.03 0 2.1 0.06 2.1 0.06 1 1.9 0.09 1.7 0.09 2 8.5 0.21 8.7 0.21 3 2.5 0.27 2.8 0.27 0 23.9 1.99 24.0 1.99 1 14.9 2.56 19.6 2.56 2 22.1 1.54 21.0 1.54 3 23.5 1.90 20.6 1.90 0 231 10.5 205 10.5 1 194 7.9 203 7.9 2 218 11.0 208 11.0 3 235 24.3 241 24.3 0 3.0 0.14 2.9 0.14 1 2.0 0.14 2.4 0.14 2 2.2 0.11 2.1 0.11 3 3.5 0.59 3.7 0.59 0 0.6 0.06 0.7 0.06 1 0.8 1.53 7.0 1.53 Between subjects Within subjects .441 .120 .552 .528 .879 .236 .847 .732 .424 .697 .043 .031 �52 APPENDIX 1-1. Continued. Control Soil parameter OM P pH Zn Year :1 SE Fertilized :1 SE 2 1.0 0.15 1.7 0.15 3 1.8 0.39 2.2 0.39 0 3.1 0.15 2.9 0.15 1 1.8 0.19 1.8 0.19 2 2.6 0.09 2.3 0.09 3 1.8 0.16 1.5 0.16 0 4.2 0.35 3.7 0.35 1 2.6 0.19 2.8 0.19 2 2.0 0.17 2.3 0.17 3 3.0 0.62 3.6 0.62 0 6.8 0.13 6.7 0.13 1 7.5 0.21 7.2 0.21 2 7.1 0.02 7.0 0.02 3 6.8 O.OS 6.9 0.05 0 0.8 0.07 0.8 0.07 1 0.4 0.06 0.5 0.06 2 0.5 0.01 0.5 0.01 3 0.6 0.08 0.5 0.08 '. < P>F Between subjects Within subjects .132 .627 .280 .762 .471 .388 .844 .188 �)j APPENDIX 1-2. Crude protein dynamics (% DM) in sagebrush foliage subsequent to fertilization in October-November 1985 from study plots in Jackson County, Colorado. ANOV A summary Year Month 1986 May ATV ATV ATV ATV ATV ATW 1988 April 17.27 0.70 Subspecies <.001 No 13.00 0.52 Fertilizer <.001 Yes 22.12 0.65 SxF .016 No 8.83 0.17 Blocks .501 Yes 10.74 0.27 Subspecies <.001 No 10.64 0.37 Fertilizer <.001 Yes 11.59 0.22 SxF .167 No 8.42 0.20 Blocks .017 Yes 9.09 0.22 Subspecies No 10.75 0.29 Fertilizer .026 Yes 11.82 0.56 SxF .166 No 8.54 0.14 Blocks .001 Yes 9.93 0.37 Subspecies <.001 No 11.47 0.29 Fertilizer <.001 Yes 12.82 0.53 SxF .939 No 11.40 0.16 Blocks .564 Yes 11.52 0.46 Subspecies No 12.86 0.55 Fertilizer .118 Yes 14.45 0.75 SxF .177 No 8.72 0.38 Blocks .049 Yes 10.17 0.46 Subspecies No 12.27 0.34 Fertilizer .153 Yes 13.95 1.30 SxF .811 No 8.56 0.23 Blocks .006 Yes 10.29 0.20 Subspecies No 11.68 0.38 Fertilizer .002 Yes 13.54 0.89 SxF .858 Yes ATW 1988 January <.001 10.73 ATW 1987 May Blocks No ATW 1987 February 0.36 ATV ATW 1986 November f>E Fertilizer ATW 1986 August Source Subspecies ATV ATW ! SE <.001 <.001 <.001 <.001 c, e, �54 APPENDIX 1-2. Continued. ANOV A summary Year Month 1988 June 0.24 Blocks .296 7.81 0.17 Subspecies .001 No 9.71 0.68 Fertilizer .461 Yes 10.41 0.24 SxF .084 No 7.91 0.32 Blocks .003 Yes 7.86 0.35 Subspecies No 10.29 0.21 Fertilizer .914 Yes 10.99 0.48 SxF .135 No 9.40 0.21 Blocks .003 Yes 10.11 0.37 Subspecies No 12.05 0.50 Fertilizer .026 Yes 13.76 0.80 SxF .202 ! ATV No 8.37 Yes ATV ATW 1989 April f>f .Fertilizer ATW 1988 October Source Subspecies ATV ATW SE <.001 <.001 �< < CHAPTER 2 BRADLEY-TERRY MODEL ESTIMATION OF PREFERENCE RANKINGS FROM PAIRED COMPARISON EXPERIMENTS INTRODUCTION Ecologists often use animals in experimental situations to establish preference or avoidance rankings for objects with different properties. These objects or subjects may differ naturally as in coloration or size or may differ because their attributes were manipulated by the experimenter. Preference rankings of objects or experimental treatments often are estimated from cafeteria-style trials where each test animal is permitted to choose among several objects or alternatives simultaneously. The response of interest is whether or not a particular object is selected in each trial. Examples of how selection may be registered are the mass of each test ration consumed, the total leaf area removed, the number of bites taken of each forage, or the number of other behaviors observed in reference to each object during the trial. Thus, each object or treatment accumulates some sort of preference score during the test interval, be they grams, bites, or number of courtship displays. Analyses of these data generally follow analysis of variance procedures or contingency table analyses. When one or more test objects are strongly preferred or avoided certain conditions or assumptions relevant to statistical procedures can be violated, especially �56 those of homogeneity of variances and the minimum number of observations per cell. Cafeteria-style trials also may have low ability to discriminate rankings of the less preferred objects. For example, consider the alternatives offered to a human community regarding the fate of a local stream: a) allow a stretch of highly-prized trout stream to remain undeveloped and undegraded, b) construct a moderately-sized dam and reservoir to help meet several community needs and, at the same time, inundate 50% of the stream, c) construct a larger reservoir that floods 75% of the stream but provide mitigation elsewhere at a rate of 3 miles of mitigation to every mile of inundated stream, and d) construct a large dam that floods the entire stream reach but provide no mitigation. If an angling club devoted to stream fishing was asked "Which option do you prefer: A, B, C, or D?", then one might expect results to highly favor response A. The preferred alternative for this population would be identified without providing much information about less desirable but possibly acceptable alternatives. cafeteria-style. A similar situation could arise when testing animal forages One highly preferred forage with high nutritional value that generally is in short supply could be used with little or no use of other more abundant but nutritionally adequate alternatives. An alternative to cafeteria trials for estimating preference rankings is to use paired comparison trials where each subject chooses among test objects or treatments offered in all possible pair-wise combinations (David 1988). For the dam example where responses were multinomial, responses are binomial when alternatives are evaluated with a paired comparison design. When responses are continuous each pairing of treatments can be treated as separate experiments where treatments are �57 paired and analyzed as paired 1 tests. For each pairing one can compute multiple tests of the hypothesis that each alternative in a pair has the same probability of being preferred. An acceptable overall experimentwise error rate can be achieved by applying the Bonferroni inequality to the a-level for each pairing. The continuous data situation also may be handled as a balanced incomplete block design, which should be more efficient and would enable analysis of factorial as well as one-way treatment structures. A combined analysis of qualitative responses from paired comparison experiments is possible in a way that is analogous to the normal-theory balanced incomplete block design. I use results of paired comparison feeding preference trials to demonstrate the utility of Bradley-Terry models (Bradley and Terry 1952) for estimating preference rankings. I also show the model can be reparameterized to include factorial treatment structures of ecological experiments. METHODS Paired comparison preference trials I used 4 sage grouse in paired comparison preference trials to estimate the preference rankings of 4 sagebrush treatments formed by factorial combinations of 2 levels of subspecies (Artemisia tridentata vaseyana [ATV] and A. 1. wyomingensis [ATW]) and 2 levels of nitrogen fertilization (fertilized [PATV and FATW] and unfertilized [ATV and ATW]). Ammonium nitrate (33% nitrogen) was applied to sagebrush rangeland in North Park, Jackson County, Colorado at a rate of 112 kg- N/ha: in a randomized complete block design (Chapter 1) to manipulate foliar < < �58 chemistry. In winters after the first year of growth response, I removed leaves from plants for estimating feeding preferences of grouse. The 4 treatment levels of the 2 x 2 factorial treatment structure were grouped into the 6 possible pairs and assigned an identification code at random. I presented each pairing to birds in paired comparison preference trials. Preference trials involved simultaneous presentation of sagebrush leaves in 400-ml beakers to each bird, allowing it to feed for 60 min, and estimating the fresh mass intake to the nearest 0.1 g of each treatment in the paired comparison. Position of treatments on the left or right side of each cage was randomized, and positions were switched on the subsequent replication of the experiment. Mass changes of control sagebrush leaves were used to estimate moisture loss during each trial. Birds were fasted overnight prior to each trial, and trials commenced in the first hour after sunrise. Simultaneous to intake-based preference trials, I observed onset of feeding by birds. At the start of each replication, I retreated from view of the birds and recorded the first beaker into which each bird approached and initiated feeding. I defined the initiation of feeding as whenever a bird broke the plane formed by the top of the beaker with its beak. I scored these treatments as preferred (1) and the beakers not used first as not preferred (0). If a bird did not initiate feeding within 15 min both treatments were scored as not preferred. I also scored the intake assays as 1 or o on the basis of relative intakes so that feeding initiation patterns could be compared with feeding preferences over longer periods. << �Statistical analysis Bradley-Teny models.-I used a series of models based on the Bradley-Terry (BT) model (Bradley and Terry 1952) to estimate preference ranks of each treatment from paired comparison preference trials. The behavioral assay for forage preference consisted of independent Bernoulli trials where a bird was scored as preferring (1) or not preferring (0) treatments in paired comparisons. If only 2 objects or treatments existed and were compared by several judges (or birds), the probability of choosing object i over objectj in a paired comparison is estimated by the number of times object i is selected divided by the total number of choices made (i + J). Likewise, the probability of choosing object j over object i is the number of wins by object j divided by the total number of paired comparisons. These probability estimates are simple binomial probabilities that can be represented as Tij' the probability of object i being chosen over object j in a paired comparison of the 2 objects. The BT model can be applied to situations where I objects, 1'" 1'2' ••• , 1'" are compared by n judges, with the 'Ythjudge making r~ replications of all possible (~) combinations. When an object i is selected over object j in a paired trial by judge 'Y making the oth comparison of i and j it can be scored as a random variable, a, such that 1 indicates preference for the object in the trial and 0 indicates the object was not preferred. Let this random variable be represented as: ~, i,j = 1,2, ... , I, i=Fj; 'Y = 1,2, ... , n; 0 It is assumed that all comparisons are independent so that independent, except for ~ + aj~ = = aij-,~ 1,2, ... , r.,. are mutually 1. The probability of object i being chosen over objectj in a particular paired comparison is represented as Pr(a~ = 1) = TiirO �60 with the restriction that 0 ~ 1 and that the sum of all preference probabilities Tij,& ~ within a paired comparison is 1. Assuming no effects due to replication or due to judge then the BT model simplifies to its basic form: T, Tij where Tij • (0 :S ~--~ (T.I +. T.) J T; :S 1, r, T, • 1, i ¢j). (2.1) is the probability of object i being preferred over objectj in a paired comparison. Parameters -,rj and Tj are the preference probability rankings of particular objects over the paired-comparison experiment to be estimated from measurements from the pair-wise preference probabilities, Tij. Thus, the Tj estimate the probability of each treatment being selected when t objects are compared. Under Ho: Tj = T = for estimating = lit Tj 0.25 for the model in this experiment. The likelihood equation has the form (Bradley 1976): n T, il, !£(T, la,) • --'--- n<.J (T, + T,)"" (aj• r , '" , a..), (2.2) fI I where aj are the row sums of the preference scores in a preference matrix when preferences are scored as 1 or 0 by each judge in each paired comparison. The a; also rank preferences when designs are completely balanced. Independent maximum likelihood estimates of (t - 1) preference probabilities are achieved via solution of the likelihood equations (Bradley 1976), �61 , I (i • 1, 2, ... , t and r it - 1) . (2.3) i E j "i I reparameterized the basic B-T model to test for main effects of subspecies (a) and fertilization (/1) in a 2 x 2 factorial framework (Abelson and Bradley 1954, Table 2-1). The multiplicative parameterization of subspecies and fertilizer effects reproduce the Ti when the interaction term (0) allows the effects of fertilizer to vary among levels of a. Parameter values for a and ~ were constrained to the interval o - 1 and 0 must lie between -1 and 1. The null hypothesis for the factorial model thus was Ho: a = 0.5, ~ = 0.5, and 0 = O. In my paired comparison experiments I tallied the number of times sagebrush treatment i was chosen over treatment j in the 6 possible pairings of 4 treatments. Treatment positions in cages were randomized and reversed in 2 replications for each bird. I assumed there were no nonrandom effects due to individual birds. Parameter estimation and hypothesis testing.-lterative numerical estimation of parameters was accomplished using Program SURVIV (White 1983). The program constructs the likelihood functions from the BT model (2.1), uses the observed cell values to produce maximum-likelihood estimates of the unknown parameters by solving the likelihood equations (2.2), and constructs likelihood ratio tests between nested sub models that are formed by constraining one or more of the parameters to their values under H,', The software also can perform Monte Carlo simulations for evaluating power associated with pre-planned hypothesis tests. Overall adequacy of models used in tests and for estimating parameters was assessed with likelihood ratio < �62 Table 2-1. Parameterization of Bradley-Terry models for one-way and 2 x 2 factorial classifications of treatments. Factorial parameterization One-way parameterization 1- Tl - T2 - T3 Subspecies (a) Fertilizer (1 - a) (1 - (J) (/3) a T2 (1 - a) T3 a ATV + 0)] [1 - (/3 Tl Sagebrush treatment ATW {J (/3 FATV + 0) FATW goodness-of-fit tests and using Akaike's Information Criterion (AlC) (Akaike 1973), defined as AlC = -2ln(~ + 2(number of parameters estimated). The AlC is useful in selecting the most parsimonious model upon which to base inferences (Lebreton et al. 1992). I used the basic BT model to obtain summary statistics for each of the 4 sagebrush treatments. I obtained estimates of rATW, TpATV' and TpATW and their associated standard errors directly from Program SURVIV using the basic BT model. Preference probability for ATV was estimated as that ETi = 1 (Table 2-1). I estimated variance of T' ~, u! + 2T' u.. ~ i «] 'r,,'ffJ TATV = TATV 1- rATW - TpATV - TpATW as (Steel and Torrie 1980:109). The factorial model parameterization was equally general, compared to the so �OJ basic model, when interaction effects were included and was used to test hypotheses about the effects of subspecies and fertilization on feeding preferences of grouse. By constraining one or more parameters in the factorial BT model to their expected values under Ho: a sequence of nested submodels was constructed that, when tested .- against their more general alternatives, were analogous to tests about effects in a 2 x 2 factorial analysis of variance. with df = Tests were likelihood ratio x- tests calculated as dfRcduced - dfOcacra.l. RESULTS I compiled preference matrices from scores (0,1) for each pairing of treatments in the behavior and intake assays (Table 2-2). Rows contain counts of the number of times the row treatment was preferred over the treatments heading up each column. The row sums (aJ were used in solving the likelihood equations (2.2), but the preference matrix was used as data input (APPENDIX 2). The basic 3-parameter BT model predicted that treatment selection based on initiation behavior was a random process (,(-3 Behavioral scores). not cover 0.25. = 4.746, f = .191, Table 2-2: Only FATV had an asymptotic 95% confidence interval that did Several trials could not be used in the analysis because some birds did not initiate feeding within 15 min. I used 1000 Monte Carlo simulations to estimate statistical power to reject the null hypothesis that at least 1 Power was about 0.4 for this example. Tj ¢ 0.25. , < �0'\ -'" Table 2-2•. Preference matrices and estimated preference probabilities (i-) from intake and behavioral paired comparison preference trials for sagebrush. Parameter estimates Preference matrix Sagebrush treatment ATV ATW FATV FATW al i- SH 95% CI Behavioral scores 2 ATV ATW 3 FATV 1 2 FATW 3 3 4 2 8 0.257 0.103 0.054 - 0.458 6 4 13 0.383 0.113 0.161 - 0.603 3 6 0.116 0.050 0.018 - 0.215 11 0.245 0.086 0.077 - 0.414 5 Intake scores 0 ATV ATW 8 FATV 5 0 FATW 8 5 3 0 3 <0.001 <0.001 -0.001 - 0.001 8 3 19 0.374 0.009 0.357 - 0.392 0 5 <0.001 <0.001 -0.002 - 0.002 21 0.626 0.008 0.609 - 0.642 8 �65 I was able to score all intake based trials (Table 2-2: Intake scores). Preference probabilities were not equal for all treatments .f = < .(01). ~3 < = 45.372, Two treatments did not receive non-zero scores when paired against the other 2 treatments and caused their preference probabilities to approach zero. The other 2 treatments were used significantly more than expected by chance (Table 2-2). The presence of this many zeros in the preference matrix also caused rounding errors to dominate the numerical solution algorithm after parameters were estimated to 3 significant digits. The strong preference for ATW and FATW in the one-way BT model analysis was expressed in the factorial BT model analysis as a strong effect by the a (subspecies) main effect (Table 2-3). Tests for factorial effects correspond to model a_ft_o vs. a_ft for the interaction effect (0), model a_ft vs. f3 for the subspecies effect, and a _ft vs. a for the fertilization effect. Each model also can be tested against the null model (In[;e:] -25.2204, df = 6, Ale = 50.4407), where all parameters were constrained to their assumed values under Ho. These results in an analysis of variance table format indicated non-significant effects for subspecies x fertilization interaction ~ > .999), and fertilization (f = .315) and a strong effect on selection probability due to subspecies ~ < .001). The small Ale for model a indicates it was the best-fitting, parsimonious summary of these data. Factorial model analysis of the behavioral scores revealed no significant main effects (Table 2-4). Power to detect differences of this size was not good null model (In[~ (f = -9.4352, df = 6, Ale = « 0.30). The 18.8704) gave an adequate fit to the data .489), but its Ale was larger than for model a {3. Under this model an ATW < �66 Table 2-3. Factorial Bradley-Terry model analysis of paired-comparison preference trials for sagebrush. Data are scored intake values and parameterization follows (Table 2-1). Model parameters estimated afJa afJ fJ a In(~ df AIC G-O-P- -2.5343 3 11.0686 >.999 -2.5343 4 9.0686 >.999 -24.9697 5 51.9394 <.001 -3.0397 5 8.0793 .962 Likelihood ratio tests between models General submodel Reduced sub model Main effect tested afJa afJa afJa afJ afJ afJ fJ a fJ a Interaction Subspecies had a 0.356 + 0.964 df 0.001 1 >.999 44.871 2 <.001 1.011 2 .603 44.871 1 <.001 1.011 1 .315 Fertilizer • Likelihood ratio goodness of fit test f > subspecies had a 0.644 )(2 ± f>)(2 t. 0.096 probability of being selected and a fertilized plant probability of being selected. DISCUSSION A strongly preferred object in a cafeteria-style preference trial should be chosen in accordance with its preference probability, leaving few if any observations < < �01 Table 2-4. Factorial Bradley-Terry model analysis of paired-comparison preference trials for sagebrush. Data are scored feeding behaviors and parameterization follows (Table 2-1). Model parameters estimated In(!:e) df Ale G-O-F" apo ap -7.0624 3 20.1248 .8760 -7.1489 4 18.2977 .9301 P -8.1843 5 18.3685 .7105 a ":8.1843 5 18.3685 .7105 Likelihood ratio tests between models General sub model Reduced sub model Main effect tested expo expo apo ap exP exP p ex p ex Interaction df ~>-K 0.173 1 .6775 2.244 2 .3257 2.244 2 .3257 Subspecies 2.071 1 .1501 Fertilizer 2.071 1 .1501 • Likelihood ratio goodness of fit test f -K > -K. for addressing the relative preferences of the remaining objects. Thus, paired comparison experiments potentially are more powerful for discriminating preferences among several objects (David 1988). When preferences for objects can only be scored subjectively as 1 or 0, the BT model can be used to estimate the individual probabilities of the objects tested in paired comparison experiments. be generalized and reparameterized This model can to account for replication effects, within-pair e, < �68 1.0 0.8 ~ 0.6 ~ 00.4 ~ 0.2 '1t' UNDER Ha: N Fig. 2-1. Empirical estimates of power to detect differences from T = 0.50 with a factorial Bradley-Terry model for several sample sizes and alternative hypotheses. Power was estimated from 1000 Monte Carlo simulations of each scenario. ordering effects (Bradley 1976) and, as I have done, to test for factorial effects (Abelson and Bradley 1954). If a judge evaluates every possible paired combination of objects, the resultant situation is equivalent to a Round-Robin tournament in sport where every player competes against every other player. The problem of obtaining rankings of multiple objects from paired comparisons was first given attention by Zermlo (1929) (cf. David 1988) in attempting to rank chess players and again by Bradley and Terry (1952) while developing statistical methods for sensory difference tests (Bradley �1976). The BT model was formulated to deal with the situation where t objects (or treatments) are compared and scored by n judges in ~~) paired comparisons. Paired comparisons have an advantage over cafeteria-style comparisons when objects can only be judged subjectively and when fine-scale judgements are otherwise difficult by extraneous influences from the presence of > 2 objects (David 1988). The influence of multiple objects may confound preference scoring in cafeteria trials especially when different object attributes or combinations of attributes drive choices in different pairings of objects. For example, color might drive choices of fruits chosen by foraging birds, but other cues related to ripeness or nutrient content may influence choices within or between color classes. The approach is to formulate a general model using the parameters of interest and to fit increasingly reduced models to the data until the "best" fitting model is selected, given the data. Tests of hypotheses can be formulated based on sets of sub models. By constraining a parameter equal to some nominal value or equal to another parameter, the parameter space is reduced. The remaining parameters then are estimated under an imposed constraint that corresponds to an assumption about the nature of the data. This assumption can be tested by performing a likelihood ratio test against the more general model. I tested the hypotheses that effects of a (subspecies), {3(fertilization), and (a x (3) did not influence the preference probabilities for sagebrush. ° The full factorial BT model (model a_ft_o) produced individual T identical to the basic BT model (2.1). To form the test for a significant interaction effect I constrained ° = O. Tests of main effects, a and {3,were estimated by constraining a = 0.5 and {3 = 0.5 in model , ' �70 a_fi to produce models f3 and a, respectively, and calculating the likelihood ratio x: relative to a_fi. Bradley-Terry models suffer from reduced efficiency relative to normal theory analogs under optimal conditions (Bradley 1955). These models are useful, however, for data sets that do not meet assumptions for analysis of variance. Numerical solution of likelihood equations is no longer a problem; the software I used has capabilities for modeling many conditions (White 1983). The method of paired comparison experiments to estimate preference and avoidance rankings should have broad applicability in studies of sexual selection and mate choice, forage selection, or evaluating the efficacy of feeding deterrents. LITERATURE Abelson, R. M., and R. A. Bradley. comparisons. Akaike, H. CITED 1954. A 2 x 2 factorial with paired Biometrics 10:487-502. 1973. Information theory and an extension of the maximum likelihood principle. Pages 267-281 in B. N. Petran and F. Csaki, editors. symposium on information theory. Second edition. International Akademiai Kiadi, Budapest, Hungary. Bradley, R. A. 1955. Rank analysis of incomplete block designs. ID. Some large- sample results on estimation and power for a method of paired comparisons. Biometrika 42:450-470. -_. 1976. Science, statistics, and paired comparisons. Biometrics 32:213-232. �11. Bradley, R. A., and M. E. Terry. 1952. Rank analysis of incomplete block designs. I. The method of paired comparisons. Biometrics 16: 178-188. David, H. A. 1988. The method of paired comparisons. Charles Griffin, London, England. 188pp. Lebreton, J. D., K. P. Burnham, J. Clobert, and D. R. Anderson. 1992. Modeling survival and testing biological hypotheses using marked animals: a unified approach with case studies. Ecological Monographs. 62:67-118. Steel, R. G. D., and J. H. Torrie. 1980. Principles and procedures of statistics. McGraw-HilI, New York, New York, USA. 633pp. White, G. C. 1983. Numerical estimation of survival rates from band-recovery and biotelemetry data. Journal of Wildlife Management 47:716-728. Zermlo, E. 1929. Die Berechnung der Turnier-Ergebnisse als ein Maximumproblem der Wahrscheinlichkeitsrechnung. Mathematische Zeitschrift 29:436-460. < < �72 '. APPENDIX 2 APPENDIX 2-1. Example of SURVW (White 1983) input file used to analyze paired comparison preference trial data with a one-way Bradley-Terry model. PRoe TITLE BRADLEY-TERRY MODEL ANALYSIS: ONE-WAY MODEL: PRoe MODEL NPAR=3: eOHORT=5 1* Total Dumber paired comparisons: pairing 1 */: 2:(I.DO-S(1 )-S(2)-S(3»/«I.00-S(1 )-S(2)-S(3» + S(1» 1* Wins by ATV in pairing 1 *1; 3:S(I)/«1.00-S(I)-S(2)-S(3»+S(I» 1* Wins by ATW in pairing 1 *1; COHORT = 5 1* Total Dumber paired comparisons: pairing 2 *1; 4:(I.00-S(1 )-S(2)-S(3»/«1.00-S(1 )-S(2)-S(3» + 5(2» 1* Wins by ATV in pairing 2 *1; 1:S(2)/«1.00-S(I)-S(2)-S(3»+ S(2» 1* Wins by FATV in pairing 2 *1; eOHORT=S 1* Total Dumber paired comparisons: pairing 3 *1; 2:(I.00-S(I)-5(2)-S(3»/«I.DO-S(1)-5(2)-S(3» + 5(3» 1* Wins by ATV in pairing 3 *1; 3:S(3)/«I.00-S(1)-S(2)-S(3»+ S(3» 1* Wins by FATW in pairing 3 */: COHORT =8 ,. Total Dumber paired comparisons: pairing 4 *1; 6:S(I)/(S(I) + 5(2» 1* Wins by ATW in pairing 4 */: 2:5(2)/(S(I)+ S(2» /* Wins by FATV in pairing 4 *'; eOHORT=8 /* Total Dumber paired comparisons: pairing 5 */; 5:5(1)/(5(1)+5(3» /* Wins by ATW in pairing 5 */; 3:5(3)/(5(1)+5(3» /* Wins by FATW in pairing 5 */: COHORT = 8 /* Total Dumber paired comparisons: pairing 6 */; 3:5(2)/(5(2) + 5(3» /* Wins by FATV in pairing 6 *'; 5:S(3)/(5(2) + 5(3» /* Wins by FATW in pairing 6 *1; LABELS; 5(1)=PROB(ATW) SELECTED; S(2)=PROB(FATV) SELECTED; S(3)=PROB(FATW) SELECTED: 1* (1.DO-S(I)-5(2)-S(3»=PROB(ATV) SELECTED *1 PRoe ESTIMATE NAME = GENERAL; INITIAL; 5(1)=0.2; 5(2)=0.01; S(3)=0.4; PROe ESTIMATE NAME=NULL; INITIAL; 5(1)=0.25; CONSTRAINTS: 5(1)=5(2); S(I)=S(2): S(I)=S(3); PROe TEST; PROe STOP; ' �1.5 APPENDIX 2-2. Example of SURVIV (White 1983) input me used to analyze paired comparison preference trial data with a factorial Bradley-Terry model. PROC TITLE BRADLEY-TERRY MODEL ANALYSIS: FACTORIAL MODEL; PROC MOOEL NPAR=3; 1* Total number paired comparisons: pairing 1 *1; COHORT= 5 2:(1.00-S(I»*(1.00-(S(2)+S(3)))1 «1.00-S(I»*(I.00-(S(2)+S(3)))+S(I)*(1.00-S(2))) 1* Wins by ATV in pairing 1 *1; 3:S(I)*(I.00-S(2»/«I.DO-S(1»*(I.00-(S(2)+S(3»)+S(I)*(1.00-S(2») 1* Wins by ATW in pairing 1 *1; 1* Total number paired comparisons: pairing 2 *1; COHORT=5 4:(I.00-S(I»*(I.DO-(S(2)+ S(3»}1 «I.00-S(I»*(I.00-(S(2)+S(3)))+(I.00-S(I))*(S(2)+S(3)}) 1* Wins by ATV pairing 2 *1; 1:(1.00-S(I»*(S(2)+S(3»1 «I.00-S(1»*(I.00-(S(2) +S(3))) + (1.00-S(I»*(S(2) + S(3») 1* Wins by FATV in pairing 2 *1; 1* Total number paired comparisons: pairing 3 *1; COHORT=5 2:(I.00-S(I»*(I.00-(S(2)+S(3»)/«I.00-S(I»*(1.00-(S(2)+S(3»)+ S(I)*S(2» 1* Wins by ATV in pairing 3 *1; 3:S(I)*S(2)/«I.00-S(1»*(I.00-(S(2)+S(3)))+ S(I)*S(2» 1* Wins by FATW in pairing 3 *1; COHORT=8 1* Total number paired comparisons: pairing 4 *1; 6:S(I)*(I.00-S(2»/(S(I)*(I.00-S(2»+(I.00-S(I»*(S(2)+ S(3») 1* Wins by ATW in pairing 4 *1; 2:(I.00-S(I»*(S(2)+S(3»/(S(I)*(I.DO-S(2»+(I.00-S(1»*(S(2)+ S(3)}) 1* Wins by FATV in pairing 4 *1; COHORT=8 1* Total number paired comparisons: pairing 5 *1; 5:S(I)*(I.00-S(2»/(S(I)*(I.00-S(2»+S(I)*S(2» 1* Wins by ATW in pairing 5 *1; 3:S(I)*S(2)/(S(I)*(I.00-S(2»+S(1)*S(2» 1* Wins by FATW in pairing 5 *1; COHORT= 8 1* Total number paired comparisons: pairing 6 *1; 3:(I.00-S(I)}*(S(2)+S(3)}/«(1.00-S(I»*(S(2)+S(3»+S(I)*S(2» 1* Wins by FATV in pairing 6 *1; 5:S(I)*S(2)/«I.00-S(1»*(S(2)+S(3»+S(I)*S(2» 1* Wins by FATW in pairing 6 *1; LABELS; S(I)=ATW LEVEL OF SUBSPECIES; S(2)=TREATMENT LEVEL OF FERTn.IZATION; S(3)=INTERACTION TERM; PROC ESTIMATE NAME=A_B_X; INITIAL; S(I)=0.4; S(2)=0.6; S(3)=0.1; CONSTRAINTS; S(3) > -1.0; S(3) < 1.0; PROC ESTIMATE NAME=A_B; INITIAL; S(I)=0.3; S(2)=O.4; CONSTRAINTS; S(3)=0.0; �74 APPENDIX 2-2. Continued. PROC ESTIMATE NAME=B; INITIAL; S(2) =0. 1; CONSTRAINTS; S(I)=0.5; S(3)=0.0; PROC ESTIMATE NAME=A; INITIAL; S(I)=O.25; CONSTRAINTS; S(2)=0.5; S(3)=0.0; PROC ESTIMATE NAME=NULL; INITIAL; S(I)=0.5; CONSTRAINTS; S(I)=S(2); S(I)=S(2); S(3)=O.O; PROCTEST; PROC STOP; < < �7) CHAYI'ER 3 EXPERIMENTAL ESTIMATION OF SAGE GROUSE. FEEDING PREFERENCES INTRODUCTION Sage Grouse forage selectively with respect to topography (Hupp and Braun 1989), sagebrush subspecies (Remington and Braun 1985, Welch et 31. 1991), and chemical characteristics of individual plants (Remington and Braun 1985, Welch et 31. 1988). Forage selection patterns, however, have not been consistent. Sage Grouse in north-central Colorado forage selectively on Wyoming big sagebrush (Artemisia tridentata wyomineensis) and avoid mountain big sagebrush CA. 1. vaseyana) (Remington and Braun 1985, Chapter 1), but Sage Grouse in southern Colorado (Hupp and Braun 1989) and Utah (Welch et 31. 1991) use mountain big sagebrush. Plant chemistry was correlated with forage selection in some cases (Remington and Braun 1985, Welch et 31. 1988) but not others (Hupp 1987, Welch et 31. 1988). The availability of different forages (Welch et 31. 1991) and Sage Grouse physiological adaptation to particular diets (Levey and Karasov 1989) likely are strong components of any explanation for observed inconsistencies in forage selection patterns. I increased the use of Wyoming big sagebrush by Sage Grouse by applying nitrogen fertilizer to plants (Chapter 1), presumably by improving forage quality. �76 Use rates of mountain big sagebrush were not affected, which does not preclude an increase in forage quality when nitrogen availability is increased. < < If Sage Grouse experience with forages of variable availability or predictability is an important determinant of forage selection, then unequal predictability or availability of forages can confound evaluation of forage selectivity. To remove the potential confounding effects of forage availability on feeding preferences, I used captive-reared Sage Grouse to experimentally estimate foraging preferences for 2 sagebrush subspecies. I also tested whether feeding preferences for these subspecies were affected by nitrogen fertilization. METHODS Preference trials I used 8 Sage Grouse in preference trials to estimate the effects of 2 subspecies and 2 levels of nitrogen fertilization (0 or 112 kg-N/ha) on feeding preferences of grouse. Ammonium nitrate (33 % nitrogen) was applied to sagebrush rangeland in North Park, Jackson County, Colorado in a randomized complete block design (Chapter 1) to manipulate foliar chemistry. In winters after the first and second year of growth response by each subspecies, I removed leaves from plants for estimating feeding preferences of grouse. The 4 treatment levels formed from the 2 x 2 factorial were randomly grouped into the 6 possible pairs and presented to each bird in paired comparison (David 1988) preference trials . I used 4 hand-reared birds from eggs obtained from Sage Grouse in North Park to estimate preferences of birds for sagebrush treatments in the first year of �I I fertilizer response. Birds were about 8 months old when testing began. They had been maintained primarily on commercial game bird rations but were also given all 4 sagebrush types to become familiar with experimental treatments. equally available to birds. Treatments were In the second year of fertilizer response I used 2 birds obtained on loan from the University of Wyoming, Laramie captive flock and 2 birds obtained from the wild population in North Park. reared from eggs collected s The Laramie birds were captive- 150 km NE of North Park. Maintenance conditions were similar to those I used except that no attempt was made to provide access to all treatments until several weeks preceding the trials. The 2 North Park birds were captured during February and trained to consume the same ration used to maintain the Laramie birds. All birds were held in 77 x 62 x 161 em (width x height x length) or 74 x 62 x 145 em metabolism cages in an airy, unheated garage. The posterior 1/3 of each cage was covered so that birds could be isolated from disturbance at the front of the cage caused by placement and removal of treatments. Snow was placed in each cage to provide a water source as temperatures were mostly below freezing. Trials conducted in year 1 were run under a controlled photoperiod of 14: 10, dark:light hrs: I did not control photoperiod in year 2. Preference trials involved simultaneous presentation of sagebrush leaves in 2 400-ml beakers to each bird, allowing it to feed for 60 min, and estimating the fresh mass intake to the nearest 0.1 g of each treatment in the paired comparison. Positions of treatments on the left or right side of each cage were randomized, and positions were switched on the subsequent replications of each experiment. Mass changes of �78 control sagebrush leaves were used to estimate moisture loss from sagebrush during each trial. Birds were fasted overnight prior to each trial, and trials commenced in the first hour after sunrise. Simultaneous to the intake-based preference trials, I observed onset of feeding by birds. At the start of each replication, I retreated from view of the birds and recorded the first beaker into which each bird approached and initiated feeding. I defined the initiation of feeding as whenever a bird's beak:broke the plane formed by the top of the beaker. I scored treatments used first as preferred (1) and the beakers not used first as not preferred (0) in the paired comparison. If a bird did not initiate feeding within 15 min both treatments were scored as not preferred. I also scored the intake assays as 1 or 0 on the basis of relative intakes so that feeding initiation patterns could be compared with feeding preferences over longer periods. I used the behavioral assay only with sagebrush in the first year of response to fertilization. Statistical analyses Paired comparison experiments to estimate feeding preference rankings and to evaluate factors governing preferences were analyzed as balanced incomplete blocks with main and subplot structure. Birds (ID) and pairings of sagebrush treatments (pairing) formed main plot effects that were tested against the main plot error term formed by the ID x Pairing effect. Each pairing formed an incomplete block having 2 of 4 possible treatments. Sagebrush treatment was the subplot effect that I partitioned into 3 single degree of freedom contrasts for subspecies, fertilizer, and subspecies x fertilizer interaction., These data were subjected to a rank-transformation (Conover and Iman 1981) due to heterogeneous variances of sagebrush treatment response �I ';j groups and analyzed in a generalized linear model framework (SAS Institute, Inc. 1988:549-640). For each of the 6 pairings of treatments I also tested the hypothesis that intakes were equal by computing Bonferroni paired 1 tests. The behavioral assay for forage preference consisted of independent Bernoulli trials where a bird was scored as preferring (1) or not preferring (0) treatments in paired comparisons. In my paired comparison experiments I tallied the number of times sagebrush treatment i was chosen over treatment j in the 6 possible pairings of 4 treatments. Treatment positions in cages were randomized and reversed in 2 replications for each bird, so the effect of replication was removed by design. assumed there were no nonrandom effects due to individual birds. I I used Bradley-Terry (BT) models to estimate the preference probability ('J') ranking for each treatment from paired comparison preference trials (Bradley and Terry 1952), and tested for factorial treatment effects with a reparameterized BT model (Chapter 2). Preference probabilities estimate the probability of treatments being selected in a simultaneous test of all treatments. o - 1 and I:Tj = Thus, their parameter values fall into the interval 1. I obtained numerical solutions of BT model parameters (1t) and constructed likelihood ratio x: tests of nested submodels using program SURVIV (White 1983). Program SURVIV computed asymptotic 95 % confidence intervals (CI) for preference rankings as T ± 1.96sE. �80 RESULTS < < Bradley-Terry model analysis The initial treatment selection by grouse making paired comparisons among sagebrush treatments was a random process Behavior scores). ~3 = 4.746, f = .191, Fig. 3-1: The expected value for .•. under Ho: was 0.25, and only FATV had a T with an asymptotic 95% CI that did not cover 0.25. 60 min for ATV and FATV were not different from 0 ~ Preference probabilities after < .05), and ATW and FATW were selected significantly more than expected by chance (Fig. 3-1). Estimates for ATV and FA TV were 0 because these treatments did not receive nonzero scores when paired against ATW or FA TW. Fertilizer had no effect on preference ~ = .315, Fig. 3-1), despite FATW having a significantly higher individual preference probability compared to ATW (Fig. 3-1). Although Fig. 3-1 indicates a subspecies x fertilizer interaction effect among treatments, FA TV was favored over ATV at the same rate as FA TW over ATW. The factorial BT model has a multiplicative combination of subspecies and fertilizer effects rather than an additive one (Chapter 2), thus the apparent interaction was caused by the estimate of the fertilizer effect being multiplied by the near-zero value of the estimate of the effect of ATV subspecies status on treatment preference. Power to detect a significant subspecies x fertilizer effect of the observed magnitude was 7%. Behavioral assay estimates differed from preference probabilities estimated from the intake assays ~3 = 19.688, f < .001). Replication had no effect on preference probability estimates ~3 = 1.081, f = .782), so use of the basic BT model without replication effects was justified. Power to detect replication effects was low (12 %) �81 0::: 0 ~ I w CD ATV ATW FATV FATW • I HO: 1f = SOURCE SSP FERT S*F 0.150 0.150 0.678 SOURCE SSP FERT S*F p <0.001 0.315 1.000 P 0.25 ATV ~ « ATW rz FATV FATW w t-e-l 0.0 0.4 0.2 0.6 0.8 1.0 PROBABILITY PREFERENCE Fig. 3-1. Sage grouse preference probabilities (7 and 95% CI) for sagebrush in the first year after treatment. model analysis. (/) w 0:: 0 ATV I-e--I <..> ATW w ~ ~ FATV (/) Main effects summary is from a factorial Bradley-Terry P SOURCE SSP <0.001 FERT 0.215 S*F 0.918 • • z FATW HO: 1f 0.0 0.2 = 0.4 PREFERENCE 0.25 0.6 0.8 1.0 PROBABILITY Fig. 3-2. Sage grouse preference probabilities (?f- and 95 % CI) for sagebrush in the second year after treatment. model analysis. Main effects summary is from a factorial Bradley-Terry �82 for effects of this size. < < In the second year of foliar response to fertilization the preference probabilities for ATV and FATV increased over the first year, but both were significantly lower than 0.25 (Fig. 3-2). The preference probability for FATW was significantly greater than 0.25, and overall, a large subspecies effect persisted (Fig. 3-2, £ < .001). Fertilizer did not cause an overall increase in preferences for sagebrush ~ = .215). General linear model analysis Consumption (g fresh matter) of treatments by birds varied significantly in the paired comparison intake trials (Table 3-1). Subspecies status had a large effect on intake ~ < .(01), as birds consumed ATW and FATW at about 5 - 10 times the ATV subspecies (Fig. 3-3). No subspecies x fertilization interaction was detected, thus fertilizer affected the intake of both subspecies ~ = .043). Consumption patterns by grouse in the second year of fertilizer response were similar to those observed in the first year (Fig. 3-3, Table 3-2) except that fertilizer no longer influenced intake ~ = .152). Power to detect this effect was about 0.29 and would not have been improved substantially by increasing the number of birds used by a factor of 5. Experimental error of this experiment (CV = 33.8%) was larger than in the first year (CV = 22.4%), however, if experimental errors in the second year trials were of the same magnitude as in the first year power to detect this effect would only increase to 57%. I partitioned the 3 df associated with birds to include a contrast of Colorado vs. Wyoming birds and found that bird source did not affect the total amount consumed during trials ~ = .644). �Table 3-1. Factors affecting consumption of sagebrush during preference trials in first year of sagebrush response to fertilization. Source of variation df £>1: 1: MS Pairing 5 292.94 6.90 ID 3 466.17 10.99 <.001 15 42.43 1.40 .232 3 2106.06 69.67 <.001 Subspecies 1 6132.78 202.88 <.001 Fertilizer 1 140.28 4.64 .043 Subspecies x fertilizer 1 45.13 1.49 .252 21 30.23 Pairing x ID (main plot error) Treatment Subplot error YEAR 1 .002 TREATMENT YEAR 2 MEAN (SE) QTI 0.151 3 0.8 (004) ) '" FATV QIJ 9 [J 1 ) -,~ <y ,7 r-, 0 0 0 2.3 (0.7) ) ? r") -, 0.144 FATV 0.640 0 0 I.() ci 0 ~ ~ N 0 0 ~ FATW 3.9 (0.7) D:J 1 ) 0.558 FATW 8.3 ( 1.1) Fag. 3-3. Average intakes of sagebrush (g fresh mass) during paired comparison preference trials. Values along arrows are probabilities of treatments within paired comparisons being the same. < < �84 Table 3-2. Factors affecting consumption of sagebrush during preference trials in second year of sagebrush response to fertilization. Source of variation df F MS £>F Pairing 5 254.62 2.82 .055 ID 3 192.88 2.14 .138 1 20.02 0.22 .644 15 90.24 1.32 .274 3 1336.36 19.54 <.001 Subspecies 1 3773.63 55.18 <.001 Fertilizer 1 150.95 2.21 .152 Subspecies x fertilizer 1 84.50 1.24 .279 21 68.39 Bird source Pairing x ID (main plot error) Treatment Subplot error Paired Bonferroni 1 test analyses were considerably less efficient than the balanced incomplete block analysis (Fig. 3-3). Subspecies effects were detectable in year 1 but not in year 2 when experiment-wise Type I error rate was controlled with the Bonferroni inequality. No effect of fertilizer could be detected with this analysis. DISCUSSION I gave captive-reared Sage Grouse equal access to all 4 treatments tested and allowed them to establish feeding preferences without the potentially confounding influences of variable availability or predictability. These birds showed the same, strong feeding preference for Wyoming big sagebrush and against mountain big sagebrush as observed in field studies in North Park, Colorado (Remington and Braun , < �1985, Chapter 1). Preference for Wyoming big sagebrush does not, however, mean mountain big sagebrush was nutritionally inadequate for maintaining Sage Grouse. I also make no claim that Wyoming big sagebrush is universally preferred over mountain big sagebrush and agree with Welch et ale (1988) that previous field studies of Sage Grouse foraging ecology have not adequately accounted for availability differences (Remington and Braun 1985). I compared equal numbers of Wyoming and mountain big sagebrush plants (Chapter 1) and experimentally estimated feeding preferences where availability was equal (this study). In each case, Wyoming big sagebrush was favored over mountain big sagebrush. Sagebrush chemistry can vary according to species (Kelsey and Shafizadeh 1979), subspecies and accessions (Welch and McArthur 1979,1981), and among plants within a feeding site (Remington and Braun 1985, Welch et ale 1988) making formulation of a general explanation of forage selectivity patterns difficult without experimental evidence across taxonomic units and geography. Nitrogen availability for sagebrush had a positive influence on my measures of forage preference. Birds consumed more sagebrush foliage of both subspecies when it was fertilized than when it was not. An effect of fertilization of forage use was not detected in field studies of mountain big sagebrush (Chapter 1), which may have be due to lag effects between habitat manipulations and population responses (Wiens et ale 1986). Birds also may avoid switching among sagebrush taxa if physiological adaptation to a different diet imposes transitory metabolic costs (Levey and Karasov 1989). These costs would discourage use of less available or predictable forages. < v �86 The initial selection of sagebrush by grouse was a random process, but after sampling, birds strongly preferred Wyoming big sagebrush over mountain big sagebrush. In my experiments, several potential edaphic and morphological cues were eliminated by presenting sagebrush as leaves in beakers. Without these cues birds could not correctly choose the "best" food without random sampling. Even with all morphological and edaphic cues available to birds it is possible that grouse may still sample plants within a site to locate individual plants of acceptable quality. This aspect of Sage Grouse foraging ecology has not received rigorous attention as variation among plants within plots or sites have been pooled into between-plot variation for analyses (Remington and Braun 1985, Hupp 1987) or replication of feeding sites has been inadequate (Welch et al. 1988). Better understanding of the dynamics of plant characteristics and Sage Grouse herbivory could be obtained with studies that have nested sampling structures so that hierarchial analyses of plant chemistry and herbivory can be incorporated into models of forage selection. Preference studies can be made more powerful relative to paired 1 test analyses by using paired comparison experiments that have a balanced incomplete block design. Qualitative responses in preference bioassays also should be considered when planning preference/avoidance studies. Bradley-Terry model analysis allows one-way and factorial effects analysis of treatments tested in paired comparison experiments (Chapter 2). <' �87 LITERATURE CITED Bradley, R. A., and M. E. Terry. 1952. Rank analysis of incomplete block designs. 1. The method of paired comparisons. Biometrics 16:178-188. Conover, W. J., and R. L. Iman. 1981. Rank transformations as a bridge between parametric and nonparametric statistics. The American Statistician 35: 124-128. David, H. A. 1988. The method of paired comparisons. Charles Griffm, London, England. 188pp. Hupp, J. W. 1987. Sage Grouse resource exploitation and endogenous reserves in Colorado. Ph.D. Thesis, Colorado State University, Fort Collins. 73pp. Hupp, J. W., and C. E. Braun. 1989. Topographic distribution of Sage Grouse foraging in winter. Journal of Wildlife Management 53:823-829. Kelsey, R. G., and F. Shafizadeh. 1979. Sesquiterpene lactones and systematics of the genus Artemisia. Phytochemistry 18:1591-1611. Levey, D. J., and W. H. Karasov. 1989. Digestive responses of temperate birds switched to fruit or insect diets. Auk 106:675-686. Remington, T. E., and C. E. Braun. 1985. Sage Grouse food selection in winter, North Park, Colorado. Journal of Wildlife Management 49: 1055-1061. SAS Institute, Inc. 1988. SASISTA'J"1' user's guide, release 6.03 edition. SAS Institute, Cary, North Carolina, USA. 1028pp. Welch, B. L., and E. D. McArthur. 1979. Variation in winter levels of crude protein among Artemisia tridentata subspecies grown in a uniform garden. Journal of Range Management 32:467-469. �88 Welch, B. L., and E. D. McArthur. 1981. Variation of monoterpenoid content among subspecies and accessions of Artemisia tridentata grown in a uniform garden. Journal of Range Management 34:380-384. Welch, B. L., J. C. Pederson, and R. L. Rodriguez. 1988. Selection of big sagebrush by Sage Grouse. Great Basin Naturalist 48:274-279. Welch, B. L., F. J. Wagstaff, and J. A. Roberson. 1991. Preference of wintering Sage Grouse for big sagebrush. Journal of Range Management 44:462-465. White, G. C. 1983. Numerical estimation of survival rates from band-recovery and biotelemetry data. Journal of Wildlife Management 47:716-728. Wiens, J. A., J. T. Rotenberry, and B. Van Home. 1986. A lesson in the limitations of field experiments: shrubsteppe birds and habitat alteration. Ecology 67:365-376. �CHAPTER 4 EFFECT OF NITROGEN FERTILIZATION ON mE NUTRITIONAL QUALITY OF BIG SAGEBRUSH FOR SAGE GROUSE INTRODUCTION The Sage Grouse, Centrocercus urwhasianus, is a seasonal dietary specialist on sagebrush foliage, Artemisia spp. (patterson 1952). Sage Grouse subsist as generalist omnivores when insects and forbs are available but switch to a diet consisting almost totally of sagebrush foliage during other months. The specific sagebrush taxa exploited varies at landscape scales as sagebrush species composition changes across the range of the Sage Grouse. At smaller scales, Sage Grouse sometimes make selective use of sagebrush subspecies and plants (Remington and Braun 1985, Welch et al. 1991). One explanation advanced to explain observed forage selectivity of Sage Grouse in north-central Colorado was an interplay between plant primary and secondary chemistry (Remington and Braun 1985). If Sage Grouse populations are limited by metabolizable energy or nitrogen of sagebrush, the potential exists to manipulate plant chemistry and thus improve habitat quality for Sage Grouse. Remington and Braun (1985) proposed applying nitrogen fertilization to sagebrush to alter resources available to sagebrush (Bryant et al. 1983) so that �90 nutritional quality is increased. They assumed that unused plants and subspecies were not nutritionally adequate and that fertilization would increase plant quality. With improved quality, use of formerly palatable plants and subspecies should be increased. Increased dietary nitrogen has produced an increase in several fitness correlates of Ruffed Grouse @onasa umbellus) (Beckerton and Middleton 1982) and Red Grouse lLa&<wus la&wus scoticus) (Miller et ale 1970). In this way the amount and quality of foraging habitat could be increased and provide a tool for habitat remediation, enhancement, or mitigation. In other work I showed that nitrogen fertilization increased the use of Wyoming big sagebrush, Artemisia tridentata ~omingensis (ATW) under field conditions (Chapter 1). On that study area, ATW was the winter forage most likely to be selected, whereas mountain big sagebrush, A. 1. vaseyana (ATV) was avoided (Remington and Braun 1985). In laboratory trials I also demonstrated that consumption of both ATW and ATV could be increased with fertilization (Chapter 3). Here I test whether nitrogen fertilization affects the availability of energy and nitrogen in these forages for Sage Grouse. METHODS Feedin& trials I used 8 female Sage Grouse captured from a wild population in North Park, Jackson County, Colorado to experimentally estimate forage digestibility. I captured birds during February and March and held them in metabolism cages (77 x 62 x 161 cm or 74 x 62 x 145 em) in an airy, unheated garage. The posterior �91 1A of each cage was covered so that birds could be isolated from disturbance at the front of the cage caused by placement and removal of treatments. Snow was placed in each cage to provide a water source as temperatures were mostly below freezing. Photoperiod was set at 14:10, dark:light hrs. I manipulated foliar chemistry of ATV and ATW, the 2 sagebrush subspecies that predominate in the areas I captured grouse. By applying 112 kg-N/ha to ATV and ATW plants I produced 4 treatments (ATV, FATV, ATW, and FA TW) having a 2 x 2 factorial structure for the main effects subspecies (SSP) and fertilization (FERT). Foliage was collected in the winter following the first growing season after nitrogen application and promptly tested in bioassay. I collected foliage from sites on or 'near Sage Grouse feeding sites. Fertilization produced increased nitrogen levels in foliage (fable 4-1), but not as much as in more extensive field testing (Chapter 1). I used a modification of Sibbald's (1979) true metabolizable energy (TME) assay rather than classical digestion trials, because preliminary work indicated that wild birds would not consume diets at maintenance levels. The TME assay has been previosuly used on wild birds (Jorde and Owen 1988). Trials were conducted by presenting to birds a single treatment at a time for 2 d and collecting all excreta. After 2 d, birds were fasted for another 1 d to collect any food remaining in the gut (Sibbald 1979). Rather than controlling intake to 3-4 pre-selected levels I allowed birds to feed aQ libitum and used the observed variation in intake to explore relationships among ingested and retained nutrients. At the same time I estimated palatability of different forages from amounts consumed. I weighed each bird at the start and finish of each trial to monitor body mass dynamics. �92 Table 4-1. Characteristics of sagebrush foliage fed to captive sage grouse. Nitrogen (% DM) Gross energy (kJ/g-DM) NDF (% DM) ADF (% DM) Subspecies Fertilization ATV No 1.27 23.04 27.1 17.3 Yes 1.40 23.21 26.6 16.1 No 1.67 22.89 31.1 19.1 Yes 2.03 23.03 31.1 19.1 ATW The experimental design was 2-4 x 4 latin squares with assignments randomized by row and column separately for each latin square. I presented treatments as whole stems, because most wild birds could not be conditioned to feed upon picked leaves. sunrise. Sagebrush stems and foliage were added to cages at about Sagebrush also was placed on top of cages in the morning and weighed at night to estimate water loss during the day. After the evening feeding period sagebrush was removed from the cages and weighed. collected, weighed, and discarded. Spilled sagebrush was Diurnal fecal droppings were collected and stored overnight in resealable plastic bags. The following morning fecal droppings were collected and added to those already stored. Cecal droppings were placed into separate plastic bags. Droppings were frozen until processed by freeze-drying and grinding in a mortar and pestle. Dry matter content of foliage and grouse droppings was determined after drying overnight at 100°C. Samples were ashed overnight at 500°C to determine organic matter and ash content. energy (Horwitz 1980) of samples were measured. Kjeldahl nitrogen and gross �Statistical analysis I estimated effects of subspecies (SSP) and fertilization (FERT) on daily intakes of dry matter, energy, and nitrogen with analysis of variance (pRDC GLM; SAS Institute, Inc. 1988). I used analysis of covariance (ANCDV A) procedures to test hypotheses about effects of subspecies and fertilization on relationships between dry matter intake and grouse body mass dynamics (SAS Institute, Inc. 1988). I report precision estimates for parameters as ± ISE. Apparent metabolizable energy (AME) intake (kJ/g dry matter intake) and assimilation coefficients (AMEC) are traditional measures for summarizing nutritional quality of avian forages. They are the appropriate measures when test animals are fed at maintenance levels and when predicting forage intake requirements from existence metabolism. TME is appropriate for estimating forage intake requirements to meet energy requirements calculated as a multiple of basal metabolism (Miller and Reinecke 1986). AMEC is estimated by (4.1) where Qi = g DM intake, Qe = g DM excreted, GEi = kJ/g DM of food, and G~ = kJ/g DM of all excrement. AME and AMEC are not independent of level of intake, because they do not include excretion of fecal metabolic energy (FF,J or urinary endogenous energy ('lTEe)as excretory energy that is not immediately of food origin (Miller and Reinecke 1986). AMEC values are related to true metabolizable energy coefficients (TMEC) by -£FEm + UEJ/(Q(GEJ (Sibbald 1979, Miller and Reinecke 1986). The value for �94 £FEm + UEJ is constant under standard conditions and has been estimated from ( Qc'GEe measured from fasting birds (Sibbald 1976, 1979). Although the size of this bias is only about -0.03 when birds are fed at maintenance (Karasov 1990), AMEC estimates become progressively more biased at low intake rates. Any tests for differences among AMEC for different forages therefore are also biased by level of intake. TMEC and TME are unbiased by level of intake (Sibbald 1979, Miller and Reinecke 1986). TMEC as a fraction of intake energy is estimated by TMEC • GE.oQ. - (GE oQ - [FE + UE]) J e J GEj e -a, me. (4.2) Rearrangement of (4.2) gives a linear model, for estimating TMEC from data normally collected during digestion trials. This model is more general than simple ratio estimators, emphasizes the relationship among the intake and retention of nutrients, and provides a familiar framework for displaying relationships and testing underlying assumptions for valid inference (Sokal and Rohlf 1981). By combining terms, a general model for any nutrient (v) having an endogenous or urinary loss component, e.g., energy, nitrogen and dry matter, has the form . a..-t :«,, AR • where: AR. = apparent retained amount of nutrient, v, a. = true metabolizability or digestibility coefficient for v J. = amount of v ingested, and (4.4) ( �'.:1:) a, = fecal metabolic plus urinary endogenous contribution to AR~. < This model can be coupled with dummy variables (ANCOVA) to test hypotheses about treatment effects on 0 and a, e.g., TMEC and [FEm + UEJ, respectively. I used model (4.4) in an ANCOVA to estimate 0, and a, for fertilized and unfertilized ATV and ATW foliage. Dummy variables in ANCOVA models were SSP, FERT, and individual birds used in the trials (ID). I examined the residuals from all models to ascertain their compliance with assumed normality and homogeneity of error variances. No information is lost by using (4.4) since apparent use efficiency coefficients and 0, are related by Apparent use efficiency of J! •• (4.5) 0, - ;' . , RESULTS I used 8 birds during the experiment, but only 6 completed testing on all treatments thereby reducing the design to randomized complete blocks. I estimated treatment effects on palatability of sagebrush by measuring daily foliage ingestion rates. Daily intakes of dry matter and energy were not strongly affected by sagebrush subspecies or fertilization status (Table 4-2). Birds consumed more fertilized ATW than unfertilized ATW ~ = .059), which also caused nitrogen intake on this diet to be nearly 3 x greater than for other diets (Table 4-2). Nutrient intakes for fertilized and unfertilized ATV were similar ~ > .6). < �\0 0\ Table 4-2. Effects of big sagebrush subspecies (SSP) and fertilization (FERl) status on intake of dry matter, . energy and nitrogen by 6 captive sage grouse. Intake Dry matter (g/d): Subspecies Fertilization ATV No ATW Energy (kJ/d) Nitrogen (g/d) SB X SB X 22.7 5.2 524 119 0.29 0.07 Yes 17.1 8.0 397 186 0.24 0.11 No 14.9 6.6 341 151 0.25 0.11 Yes 36.7 9.4 846 215 0.75 0.19 E f>E E X SB ANOVA f>E f>E df E ID 5 0.91 .502 0.91 .502 1.03 .436 SSP 1 0.61 .446 0.59 .455 3.36 .087 FERT 1 1.15 .301 1.18 .295 3.12 .098 SSP"'FERT 1 3.31 .089 3.29 .090 4.62 .048 Source Error 15 �'71 Table 4-3. Effects of sagebrush subspecies (SSP) and fertilization (FERT) on sage grouse body mass dynamics (g fresh mass lost or gained per day). £>E E Source df Intake 1 963.9 7.29 .013 SSP 1 442.2 3.35 .078 Intake*SSP 1 1342.9 10.16 .004 FERT 1 <0.1 <0.01 .988 Intake*FERT 1 7.6 0.06 .812 SSP*FERT 1 130.9 0.99 .324 Intake*SSP*FERT 1 38.0 0.29 .597 24 132.1 Error MS Intakes of sagebrush foliage were below maintenance levels, and Sage Grouse lost an average of about 2% of body mass per d. Dry matter intake ~ sagebrush subspecies ~ = = .013) and .004) both affected body mass dynamics of birds (Table 4-3). Conversion efficiency of dry matter intake to body mass for ATV-FATV was not different from 0 ~ > 0.8, Fig. 4-1), and birds lost an average of 31.4 g/d while on ATV-FATI. treatments regardless of intake (Fig. 4-1). When only ATW-FATW treatments were considered, the relationship between mass balance and intake was Y = -49.6 + 0.70·?, ~ < .001, r = 0.597). Predicted maintenance intake was 70.8 g-DM/d. Fertilization did not affect the rate of metabolizable energy capture ~ = .948) from intake energy, but TMECs differed ~ The TMECs were 0.414 + 0.063 and 0.561 = .013) between subspecies (Table 4-4). + 0.047 for ATV and ATW subspecies, �98 '. < 20 ATV -e-,--.... ATW • -0 <, 01 0 Y - -O.03x Y - O.70x - 49.6 • "-"" w - 31.4 • • • u ~ -20 _.J « CD (f) (f) « ~ -40 • • • - 60 '--"-- o "'-...I.._,____''' 20 40 60 80 DRY MATTER INTAKE (g/d) o Fig. 4-1. Sage grouse body mass dynamics in relation to daily dry matter intake of 2 big sagebrush subspecies. respectively. Metabolizable energy was extracted from dry matter at 9.6 for ATV and 12.9 ± + 1.4 kJ/g 1.1 kJ/g for ATW. For birds consuming sagebrush foliage at the predicted maintenance rate, estimated TME intakes were 609, 769, 1050, and 888 kJ/d for ATV, FATV, ATW, and FATW, respectively. I estimated £FEm+ UEJ to be 71.6 fertilization (f CEs.ll = ± 30.8 kJ/d for all levels of subspecies (f = .118) and .390). Estimates of £FEm+ UEJ also did not differ among birds = 1.34, ~ = .316). Predicted AME intakes were calculated as TME - £FEm+ UEJ. �Table 4-4. Effects of big sagebrush subspecies (SSP) and fertilization (FERT) status on digestibility (6) of dry matter, energy and nitrogen by 6 captive sage grouse. Digestibility coefficients Dry matter A Nitrogen A 6 SE 6 SE 0.077 0.37 0.080 -1.49 0.994 0.47 0.074 0.47 0.076 0.14 0.858 No 0.62 0.088 0.65 0.091 -0.62 0.858 Yes 0.55 0.046 0.54 0.048 0.08 0.371 E f>E E 6 Subspecies Fertilization ATV No 0.35 Yes ATW Energy A SE ANCOVA E f>E Source df Intake 1 102.85 <.001 99.44 Intake"'SSP 1 7.85 .Oll 7.52 Intake"'FERT 1 0.15 .707 Intake"'SSP"'FERT 1 2.37 .140 Error <0.01 2.39 <.001 f>E 0.77 .391 .013 0.33 .574 .948 2.76 .113 .139 0.46 .505 19 <.C \.( �100 Birds did not consume enough forage to maintain a positive nitrogen balance (intake nitrogen - excretory nitrogen) on any diet. Nitrogen balance (g/d) was -1.2 ± 0.3 for ATV, -0.7 for FATW. ± 0.1 for FATV, -1.1 + 0.8 for ATW, and -0.8 ± 0.4 Nitrogen balance was not influenced by subspecies ~ = .999) or fertilization ~ = .183). Although total nitrogen intake was 3 x greater for FATW than for ATW, nitrogen balance was not affected ~ = .430). nitrogen (g/d) was a function of intake nitrogen ~ = .018, Total excretory f = 0.228), but amounts of metabolized nitrogen were not related to intake nitrogen (Table 4-4, £ = .391). Dry matter digestibility coefficients mirrored TMECs (Table 4-4). Digestibility was affected by subspecies ~ = .011) but not fertilization ~ = .707). Combined fecal metabolic plus endogenous dry matter excretion averaged 4.9 1.3 gld and did not differ among birds <Es.11 = 1.26, f = ± .346). DISCUSSION This study demonstrated that Sage Grouse forage selectivity observed in field studies (Remington and Braun 1985, Chapter 1) was correlated with several measures of food quality. Although fertilization of ATW in the field produced significantly increased use by grouse (Chapter 1) and increased consumption rates for both subspecies in laboratory trials, fertilization had no measureable effect on forage quality. The rate of metabolizable energy capture from foliage was greater for Wyoming big sagebrush, the forage selected in the field by the population of birds I sampled, than for mountain big sagebrush. Predicted metabolizable energy intake of ATW at maintenance was 913 kJ/d and provided 84 - 94% of predicted field �lUJ.. metabolic rate for non-passerines the size of female Sage Grouse (971 - 1084 kJ/d, Nagy 1987). Mountain big sagebrush at the same level of dry matter intake provided 63 - 70 % of predicted field metabolic rate. Predicted existence metabolism for the birds I tested should have been 719 - 793 kJ/d (Andreev 1988), which was slightly larger than predicted daily metabolizble energy intake of ATV (681 kJ/d). Birds on ATW diets converted dry matter intake to body mass at an efficiency of 70%, which was significantly greater than for ATV, to further support assertions that quality differences existed among subspecies. My regression of body mass dynamics on dry matter intake explained about the same amount of variation in mass balance (i = 0.597) Red Grouse (i = as did the relationship between ME intake and mass balance for 0.53, Moss and Trenholm 1987). Remington (1990), however, found no relationship between forage intake and body mass dynamics. Sage Grouse have meager endogenous nutrient reserves to exploit during reproduction (Remington and Braun 1988, Hupp and Braun 1989) making dietary sources critical. Variation in nitrogen intake of the magnitude I produced in Sage Grouse significantly enhanced fitness correlates in Ruffed Grouse, e.g., for every 1 % increase in dietary crude protein content, clutch size was increased by 0.7 (Beckerton and Middleton 1982). Nitrogen balance should be achieved at intakes of between 0.5 - 0.6 g-N/d (Robbins 1981) or 0.4 - 0.9 g-N/d (Andreev 1988). All treatments contained enough nitrogen to match predicted maintenance nitrogen requirements if consumed at predicted maintenance dry matter intake. Reproductive costs are about 175% of maintenance (Robbins 1981), however, and only ATW and FATW would meet this level. Actual maintenance needs may be higher than predicted, because ,' �102 detoxification metabolism of sagebrush anti-herbivore chemistry may impose energetic and nitrogen costs (Remington 1990). Fertilization effects on sagebrush chemistry should influence metabolizable energy and nitrogen by mitigating detoxification costs, but these effects have been difficult to demonstrate with typical environmental variation in plant secondary chemistry. Remington (1990) measured increased ornithuric acid and ammonium ion excretion by Blue Grouse IDendra~apus obscurus) when benzoic acid was added to Douglas-fir (Psuedotsu~a menziesii) needles but detected no effects on ME, MN, or nitrogen balance. By quadrupling the monoterpene load of Douglas-fir he caused a significant reduction in ME and nitrogen balance, but these parameters were not affected by a doubling of dietary monoterpenes. ME showed no linear trend with level of monoterpene intake. The measured reduction of ME and nitrogen balance was attributed to increased contributions of ammonium and glucuronic acid nitrogen to excretory nitrogen at the highest levels of monoterpene intake. Although total nitrogen excretion may have been increased due to detoxification metabolism, this was not reported. By introducing additional explanatory variables into linear models additional variation in dependent variables can be explained and added insight gained if these terms represent ecological processes. The energy excreted for each g of nitrogen retention or loss is potentially an important determinant of forage quality (Remington 1990), so rather than assuming a constant value, this parameter can be estimated from digestion trial data. The linear model, ,< �lUj estimated 0 (AMECN) for ATV and ATW as 0.42 and 0.56, respectively (f = 0.88). The coefficient, "I, was 37.6 kJ/g N balance did not differ by subspecies even though the ATV diet should have more anti-herbivore compounds (Remington and Braun 1985). Karasov (1990) presented a non-linear, mechanistic model for predicting apparent MEC that has an equivalent linear form similar to (4.6). from (4.6) in the adjustment to zero nitrogen retention. The difference The amount of retained nitrogen is multiplied by a constant, 34.5 kJ/g N, that is essentially the same as values used in conventional estimates of nitrogen-corrected energy (~, apparent metabolizable Sibbald 1979) and the one I estimated. I used linear models to estimate the rates of nutrient capture by Sage Grouse, rather than estimated ratios of retained and ingested nutrients, e.g., AMB. Had I constructed hypothesis tests about apparent assimilation coefficients the tests would have been biased by level of intake. The difference between dry matter digestibility of ATW and ATV at observed dry matter intakes was 36 % larger than if both estimates had been made at maintenance (Fig. 4-2). By using a linear model I tested whether endogenous dry matter excretion and digestibility were affected by treatments and transformed the primary test of interest from one of a difference in the elevation of 2 curves to a test of slopes (Fig. 4-2). The latter test is independent of intake. Other models can be formulated to aid understanding and analysis in forage digestibility studies. For example, the effect of ingesting putative anti-herbivory compounds on metabolizable energy could be estimated by substituting g of the �104 ,, __.- . 0.60 ,....-----------------, ATV ATW 0.40 (:) OBSERVED ?7~~ /~ ::E (:) t- Z 0.20 w a::: t ~/ -c MAINTENANCE INTAKE ," , Q.. Q.. I I I I -c 0.00 .,__-+-;.-: ----------------4 -0.20 50 ---. "'0 40 <, 0\ t- '-" ::E w (:) Z »: 30 ,;' 0::: -c Q.. Q.. -c (:) w 20 OBSERVED INTAKES 10 ~ Z ~ w a::: ;,' """"""".. /" .,1/(" t " MAINTENANCE INTAKE ... ,..,../' ' " 0 0 20 40 60 80 100 INTAKE (g/ d) Fig. 4-2. Relationships among apparent dry matter digestibility (DMD), apparent retained dry matter (DM), and DM intake under linear and nonlinear model formulations. �lU) compound ingested for nitrogen balance. If anti-herbivory compounds exert a net negative effect on bioenergetics, the rate coefficient would be significantly < 0 and should estimate detoxification costs. There are several advantages to analyzing digestion trial data with mechanistic models over simple ratio estimators: 1) because they are mechanistic, they promote rigorous estimation and testing of effects on ecological or physiological processes, 2) the estimation and testing theory is well-developed with numerous supporting software packages, and 3) plots provide useful summaries of relationships among dependent and independent variables and may provide insight into whether assumptions are being met. Inference from any model is limited by the validity of attendant assumptions regarding their use. For the linear models I described the assumptions are essentially those of linear least-squares regression. I observed few problems with homogeneity or normality of errors in residual plots from models having reasonable fits (f > 0.80) to the data. Measurement errors in quantities of nutrients. ingested are controllable and should be small relative to experimental errors in dependent variables. When other explanatory variables, such as nitrogen balance, are added measurement errors could cause precision estimates to be biased. processes that are modeled may in fact be nonlinear. Finally, ecological Adequate nonlinear models may be formulated that include dummy variables for hypothesis testing. Fertilization significantly increased the potential for Sage Grouse to ingest nitrogen at rates shown to have positive effects on fitness. Measurement of an effect on the availability of nutrients was problematic. The birds I tested were adapted to ATW, which is qualitatively and quantitatively different from ATV. Use efficiency < < �106 of nutrients in ATV may have improved over time (Levey and Karasov 1989) as Sage Grouse physiology adapted to a qualitatively different diet. The effects of diet switching on use efficiency could be tested by estimating use efficiencies with birds from populations that are physiologically adapted to different diets and then switching diets in a cross-over experiment. Digestibility trials would need to be conducted for longer than usual for avian studies to measure rate of physiological adaptation. One may argue that birds in negative energy and nitrogen balance do not provide reasonable models for free-ranging birds consuming the same forage, yet dry matter digestibility and metabolizability energy estimates were not unusual for grouse (Karasov 1990). Variation in bird performance certainly contributed to my failure to detect fertilization effects on nitrogen balance when inter-treatment variation in nitrogen intake varied 3-fold. Starvation causes accelerated sloughing of gut mucosa (Bayer et al. 1981) and by-products of protein catabolism likely confounded treatment effects. Both processes add to fecal metabolic and urinary nitrogen, part of which is recycled to the ceca and rapidly decomposed by cecal microbes (Mortensen and Tindall 1981). Cecal recycling may spare nitrogen from excretory processes (Mortensen and Tindall 1984) and may explain why I found only a weak relationship between nitrogen intake and excretory nitrogen and no relationship between intake itrogen and nitrogen balance. <' �10i LITERATURE CITED Andreev, A. V. 1988. Ecological energetics of Palaearctic Tetraonidae in relation to chemical composition and digestibility of their diets. Canadian Journal of Zoology 66: 1382-1388. Bayer, R. C., J. H. Rittenberg, F. H. Bird, and C. B. Chawa. 1981. Influences of short term fasting on chicken alimentary canal mucosa. Poultry Science 60: 1293-1302. Beckerton, P. R., and A. L. A. Middleton. 1982. Effects of dietary protein levels on Ruffed Grouse reproduction. Journal of Wildlife Management 46:569-579. Bryant, J. P., F. S. Chapin, m, and D. R. Klein. 1983. Carbon/nutrient balance of boreal plants in relation to vertebrate herbivory. Oikos 40:357-368. Hupp, J. W., and C. E. Braun. 1989. Endogenous reserves of adult male Sage Grouse during courtship. Condor 91:266-271. Jorde, D. G., and R. B. Owen. 1988. Efficiency of nutrient use by American Black Ducks. Journal of Wildlife Management 52:209-214. Karasov, W. H. 1990. Digestion in birds: chemical and physiological determinants and ecological implications. Studies in Avian Biology 13:391-415. Levey, D. J., and W. H. Karasov. 1989. Digestive responses of temperate birds switched to fruit or insect diets. Auk 106:675-686. Miller, G. R., A. Watson, and D. Jenkins. 1970. Response of Red Grouse populations to experimental improvement of their food. Pages 323-335 in A. Watson, editor. Animal populations in relation to their food resources. British �108 Ecological Society Symposium 10. Blackwell Scientific Publ., Oxford and Edinburgh, England. Miller, M. R., and K. J. Reinecke. 1986. Proper expression of metabolizable energy in avian energetics. Condor 86:396-400. Mortensen, A., and A. R. Tindall. 1981. Caecal decomposition of uric acid in captive and free ranging Willow Ptarmigan (Lae;qpuslae;Oj)uslae;opu~). Acta Physiologica Scandinavica 111:129-133. Moss, R., and 1. B. Trenholm. 1987. Food intake, digestibility and gut size in Red Grouse. British Poultry Science 28:81-89. Mortensen, A., and A. R. Tindall. 1984. The role of urea in nitrogen excretion and caecal nitrogen metabolism in Willow Ptarmigan. Journal of Comparative Physiology B 155:71-74. Nagy, K. A. 1987. Field metabolic rate and food requirement scaling in mammals and birds. Ecological Monographs 57:111-128. Patterson, R. L. 1952. The Sage Grouse in Wyoming. Sage Books, Denver, Colorado, USA. 341pp. Remington, T. E. 1990. Food selection and nutritional ecology of Blue Grouse during winter. Ph.D. dissertation. University of Wisconsin, Madison, USA 116pp. Remington, T. E., and C. E. Braun. 1985. Sage Grouse food selection in winter, North Park, Colorado. Journal of Wildlife Management 49:1055-1061. Remington, T. E., and C. E. Braun. 1988. Carcass composition and energy reserves of Sage Grouse during winter. Condor 90:15-19. �Robbins, C. T. 1981. Estimation of the relative protein cost of reproduction in birds. Condor 83:177-179. SAS Institute, Inc. 1988. SASISTAT'" user's guide, release 6.03 edition. SAS Institute, Cary, North Carolina, USA. 1028pp. Sibbald, I. R. 1976. A bioassay for true metabolizble energy in feeding stuffs. Poultry Science 55:303-308. _. 1979. Metabolizable energy evaluation of poultry diets. Recent Advances in Animal Nutrition 13:35-49. Sokal, R. R., and F. J. Rohlf. 1981. Biometry. 2nd ed. W. H. Freeman and Co., San Francisco, California, USA. 859pp. Welch, B. L., J. C. Pederson, and R. L. Rodriguez. 1988. Selection of big sagebrush by Sage Grouse. Great Basin Naturalist 48:274-279. Welch, B. L., F. J. Wagstaff, and J. A. Roberson. 1991. Preference of wintering Sage Grouse for big sagebrush. Journal of Range Management 44:462-465. Prepared by ~r----. Orrin B. M ers Graduate Research Assistant Approved by t1~2. ~ Clait E. Braun Wildlife Research Leader ��JOB PROGRESS REPORT State of: Colorado Project: Y-167-R York Plan: Job Title: 13 Upland Bird Research Job: 10 Movements. Reproductive Success. and Habitat Use by Introduced Plains Sharp-tailed Grouse Period Covered: 01 July through 31 December 1991 Author: Personnel: Yildlife Kenneth M. Giesen Clait E. Braun and Kenneth M. Giesen, Colorado Division of ABSTRACT A study plan for evaluating success of transplants of plains sharp-tailed grouse (Tympanuchus phasianellus jamesi) into historic range along the Front Range of Colorado was completed. A trip to sharp-tailed grouse range in southeastern Yyoming was completed in October to evaluate potential trapping sites and coordinate transplant activities with Yyoming Game and Fish personnel and local landowners. ��MOVEMENTS, REPRODUCTIVE SUCCESS, AND HABITAT USE BY INTRODUCED PLAINS SHARP-TAILED GROUSE Kenneth M. Giesen INTRODUCTION Plains sharp-tailed grouse (Tympanuchus phasianellus jamesi) historically occurred along the Front Range of Colorado. Sharp-tailed grouse populations declined with human settlement and were extirpated from most of their range in eastern Colorado by the late 1800's. In recent years breeding populations were documented only in Douglas County, although winter migrants or transients have been reported from Yuma, Logan, and Weld counties (Hoag and Braun 1990). Plans to increase distribution and populations of plains sharp-tailed grouse in Colorado will rely primarily on transplants (Braun et al. 1992). While numerous transplants of prairie grouse have occurred, few have been successful (Toepfer et al. 1990, Rogers 1992, Hoffman et al. unpubl. ms.). Thus, it is desirable to document responses of sharp-tailed grouse to experimental transplants and evaluate parameters potentially affecting success including movements, habitat use, mortality, and reproduction. P. N. OBJECTIVES The objectives of this project are to assist with trapping and transplanting plains sharp-tailed grouse into selected sites along the Front Range of Colorado and evaluate transplant success. Population characteristics of the transplanted population including movements and home range size, mortality, and production will be compared to a naturally pioneering sharp-tailed grouse population in Logan County near the Tamarack State Wildlife Area. SEGMENT OBJECTIVES 1. Review literature on prairie grouse introductions, movements, and habitat use. 2. Coordinate efforts with Wyoming Game and Fish personnel and affected landowners in southeastern Wyoming to locate potential trapping sites for plains sharp-tailed grouse. 3. Prepare annual progress report. RESULTS AND DISCUSSION After review of the draft recovery plan for plains sharp-tailed grouse in Colorado (Braun et al. 1992) and appropriate literature, a study plan was prepared to evaluate transplants of plains sharp-tailed grouse into the Front Range of Colorado. A work trip was taken to southeastern Wyoming in October to meet with Wyoming Fish and Game personnel and explore habitats occupied by sharp-tailed grouse in Laramie and Goshen counties. Wyoming personnel had little information on sharp-tailed grouse populations (lek surveys, hunter harvest) in these areas, but observations indicated adequate populations existed for trapping and transplanting. �I ]1.+ LITERATURE CITED Braun, C. E., R. B. Davies, J. R. Dennis, K. A. Green, and J. L. Sheppard. 1992. Plains sharp-tailed grouse recovery plan. Colorado Div. Wi1d1., Denver. 33 pp. Hoag, A. W., and C. E. Braun. 1990. Status and distribution of plains sharptailed grouse in Colorado. Prairie Nat. 22:97-102. Hoffman, R. W., W. D. Snyder, G. C. Miller, and C. E. Braun. (in prep.). Reintroduction of greater prairie-chickens in northeastern Colorado. Rogers, R. D. 1992. A technique for establishing sharp-tailed grouse in unoccupied range. Wi1d1. Soc. Bull. 20:101-106. Toepfer, J. E., R. L. Eng, and R. K. Anderson. 1990. Transplanting prairie grouse: what have we learned? Trans. North Am. Wi1d1. and Nat. Resour. Conf. 55:569-579. Prepared by _ .• ~~....;.....;._;..:...:.'#I._..:;.L~~·....;.... __ Kenneth M. Giesen Wildlife Researcher C �JOB PROGRESS REPORT < <. State of: Project: ~C~o~l~o~r~a~d~o ~W_-~1~67~-~R~ 14 Work Plan: Job Title: _ Job Upland Bird Research 4 Movements. Reproductive Success. and Habitat Use by Introduced Greater Prairie-chickens Period Covered: Author: _ 01 January through 31 December 1991 Grant M. Beauprez Personnel: Clait E. Braun, Shane Briggs, Larry Rogstad, Mike Schroeder, Colorado Division of Wildlife; Grant M. Beauprez, Jennifer Clarke, University of Northern Colorado ABSTRACT Greater prairie-chickens (Tympanuchus cupido pinnatys) were captured on leks in Colorado and Kansas with walk-in traps during March-April 1991. The birds were released between 2 and 9 April at Pinneo and Yells Ranch, Colorado. Forty-three birds (23 hens, 20 males) were released at Pinneo, while 50 (23 hens, 27 males) were released at Yells Ranch. Twenty-four birds (6 hens and 6 males at each site) were equipped with battery-powered, necklace-attached radio transmitters, and were periodically relocated to obtain data on survival, movements, lek establishment, and nesting. Birds were tracked from 2 April through 31 December. Eight birds were depredated (3 males at Pinneo; 2 males, 3 females at Wells Ranch), 7 birds lost their radio transmitters (3 males, 1 female at Pinneo; 2 males, 1 female at Wells Ranch), and radio contact was lost with 7 birds (4 females at Pinneo; 2 males, 1 female at Wells Ranch). Eight leks were formed (5 at Pinneo; 3 at Yells Ranch). The maximum number of males on any lek was. 7 at Pinneo. Seven hens were known to have nested (5 at Pinneo, 2 at Yells Ranch), and 6 were successful (4 at Pinneo, 2 at Wells Ranch). Mean distance of nests from the nearest lek was 2.9 km. Average clutch size was 11.8 eggs (range 10-16). The maximum documented dispersal distance was 79 km by a hen released at Yells Ranch and found at the Pinneo study area. Birds from Yells Ranch had a mean dispersal distance of 20.1 km compared to 9.0 km for Pinneo birds. ��111 MOVEMENTS, REPRODUCTIVE SUCCESS, AND HABITAT USE BY INTRODUCED GREATER PRAIRIE-CHICKENS Grant M. Beauprez The greater prairie-chicken originally ranged over much of the central and eastern Great Plains of North America (Aldrich 1963); it's distribution has been greatly reduced due to destruction of native grassland habitats (Aldrich 1963, Jones 1963, Christisen 1969, Johnsgard 1983). The first record of greater prairie-chickens in Colorado occurred in 1897 in the extreme northeast part of the state (Cooke 1898). Evidence suggests that greater prairie-chickens markedly expanded their range west and north with human settlements and advent of grain farming during the late 1800's and early 1900's (Sclater 1912, Schorger 1944, Beck 1957, Stempel and Rodgers 1961, Christisen 1969, Horak 1985). However, with intensified farming, overgrazing, and a series of drought years in the 1930's, their distribution and numbers in Colorado markedly decreased (Evans and Gilbert 1963). By 1963, the reported number of greater prairie-chickens in Colorado had dropped to 700-800 (Evans 1964). In 1973, population estimates were as low as 600 birds (Graul 1975) and the greater prairie-chicken was declared endangered under Colorado's Nongame, Endangered, or Threatened Species Conservation Act (Title 33, Article 8, Colorado Revised Statutes) (Pusateri 1990). The most current population estimates (1981-83) indicate that at least 3000 and possibly 6000 birds were present in Yuma, Washington, and Phillips counties (VanSant and Braun 1990). In 1990, the Colorado Division of Wildlife prepared a greater prairie-chicken recovery plan (Pusateri 1990). The main objective of this plan is to remove the greater prairie-chicken from the state's endangered and threatened list by 1995. An eventual goal is to classify the greater prairie-chicken as a game species. Attainment of these goals involves increasing -the greater pra1r1echicken's distribution in the state by means of transplanting birds into previously occupied and unoccupied range. P. N. OBJECTIVES The objectives of this study are to successfully transplant greater pra1r1echickens to two sites in northeastern Colorado, evaluate the success or failure of greater prairie-chickens to establish breeding populations at the two sites and in surrounding areas, and define guidelines for introduction of greater prairie-chickens into new or previously occupied habitats. < < �118 SEGMENT OBJECTIVES 1. Review literature on prairie grouse introductions, movements, reproduction, and habitat use. 2. Periodically relocate all radio-marked greater prairie-chickens at the Pinneo and Wells Ranch release sites until the radios fail or are lost. 3. Ascertain seasonal movements of all radio-marked greater prairiechickens. 4. Document establishment and attendance at leks. 5. Determine nest establishment and nest success of radio-marked greater prairie-chickens. 6. Document juvenile recruitment. 7. Identify habitats used by radio-marked greater prairie-chickens. 8. Compile data, analyze results, and prepare reports/thesis and publish findings in appropriate technical journals. STUDY AREA Greater prairie-chickens were studied at two separate areas. One area (Wells Ranch) was 26 km east of Greeley, in Weld County, and the other area (Pinneo) was19 km southeast of Brush, in Washington County. Transplant stock for the Wells Ranch was obtained from southcentral Kansas near Winfield, and stock for Pinneo was obtained in Yuma County, Colorado near Wray. The Yuma County area is characterized by rolling sandhills with a mix of sand sagebrush, grasses, and nearby agricultural fields. The Kansas site was more mesic with tall grass pra1r1e in a landscape of rolling hills with exposed rock on side hills and wooded draws or agricultural fields at lowest elevations. METHODS Greater prairie-chickens were captured on leks with walk-in traps (Schroeder and Braun 1991). Captured birds were weighed, classified to age and sex, .and marked with a serially-numbered aluminum band on the right leg and colored plastic bandettes on both legs (red at Pinneo, green at Wells Ranch). The two releases were completed between 2 and 9 April 1991. Forty-three birds (23 hens, 20 males) were released at Pinneo, while 50 (23 hens, 27 males) were released at Wells Ranch. Twenty-four birds (6 hens and 6 males at each site) were equipped with battery-powered, necklace-attached radio transmitters (Holohil Inc., Woodlawn, Ont.). Radio-marked prairie-chickens were relocated using a Telonics TR-2 portable receiver and a hand-held, 3-element yagi antenna. Each bird was visually observed once every 7-14 days. Additional observations (at least 2 every 7 days) were obtained by triangulation. Three �119 or more azimuths were obtained <1.5 km of target transmitters and at anglesof-incidence >35 degrees and <145 degrees (Schroeder 1991). Locations of radio-marked birds, leks, and nests were recorded to.the nearest 20 m as Universal Transverse Mercator (UTM) grid coordinates (Grubb and Eakle 1988), and plotted on U.S. Geological Survey 7.5-minute topographic maps. Leks were located by systematically searching suitable habitats using a spotting scope and a parabolic microphone listening device during morning display periods. Counts of male and female greater prairie-chickens on all known leks were made 1-2 times weekly during the morning display period. Greater prairie-chicken nests were located by tracking radio-marked hens. Clutch size, incubation period, number of eggs hatched, nest fate, distance from release site, and distance from nearest lek were recorded. Three flights using a Cessna 185 with strut-mounted antennae and a Telonics receiver/scanner were made (23 Apr, 29 May, 14 Nov)· to locate missing radio-marked birds. Habitat types used and selected by radio-marked greater prairie-chickens will be described in detail during the 1992 field season. RESULTS AND DISCUSSION Lek Surveys Eight leks were located (Figs. 1 and 2) and surveyed (Tables 1 and 2). The mean distance of leks from release sites was 7.0 km for Pinneo and 10.9 km for Wells Ranch. The maximum number of males on any lek was 7 on lek #4 at Pinneo. No females were ever observed on a lek, however, 0906 was near lek #1 at Pinneo on several occasions during the morning display period. Also, female 0369 was 2.8 km west of lek #1 at Wells Ranch with a non-radio-marked male on one occasion. Lek #3 at Pinneo was discovered on 14 May in an alfalfa field, 1.7 km southwest of the release site with 4 displaying males in attendance. However, subsequent visits to this site showed no males in attendance. Nesting Seven radio-marked hens definitely nested (5 at Pinneo, 2 at Wells Ranch) (Figs. 1 and 2, Table 3). At Pinneo, 4 hens successfully nested and a fifth hen was on a clutch of ~l eggs until 21 May. She was no longer on the nest after this date, but was still in the area until the end of October. It is unknown if she was successful as her nest was not relocated after she left it. The 2 hens that nested at Wells Ranch were both successful. Average clutch size for 5 nests was 11.8 eggs (range 10-16). It is unknown how many eggs were in the clutches of two of the hens as no opportunity to count their eggs occurred because they were not located away from their nests. Three eggs failed to hatch (10 of 11 from 0896 and 9 of 11 from 0378). The mean distance of nests from leks was 2.9 km suggesting that hens were attracted to the leks. Five of seven nests were in areas with heavy sand sagebrush (ArCemisia filifolia), and were well hidden within the vegetation. Another nest was in dense, tall grass while the final nest was in a wheat field. Brood counts conducted on 21-22 July revealed that 3 hens at Pinneo (0906, 0896, 0554) had broods consisting of at least 10, 3, and 1 chicks respectively. Females 1096, ~ « �120 PINNEO STUDY AREA ;R56~ T6 N I : ---If---I-R1r~-~---l---- N 'T" -Rl4\J--- ---- ________8>.3~/ ;-_ I I i1 ii- / • BRUSH ~ /' V / COLORADO I I I I iI , I I I I I I I 1 ° /' , ./ V V • i:J !--- LEKS ° NESTS ° fa .--V 1- • , ,, ,, ,, ,, ,,, ,, T 4N ,, ,, , ,, ,, ,, , X RELEASE SITE L. .O_C iBH US ~---~ :, I ./ I J '/ V I I V [L -~ ,/ I r-.. I I I I !'- r-. ~ ,, ,, I. ·,1' I I I T3 N , ,, , ,T ,,, 5N , ,, ../ I I I I T4 N : T6 N ,,, , ,,, /' I T5 N --- -R52Y{, <, I I I <, r-......... I ~ = "" I I I \. ';!I I 1 , 3. ' \ ~' ,. .. I I I I I I T2 N i I I I I I I ! , ,, ,, ,, , I I I I ! T1 N ,, , I I ,---- ---- .•-_ •. R56W I •.. :" ,, , ,T ,, 3N ,, , ,, ,, ,, ••••• ,,, ,, ,, : :T , 2N ,, , ,, --- •.•..._ --- R55W --- --- --- --- --- :T1 N __ 1- R54W --- _- ---- _ •.•. -.. --- .__ . •.-•.•. _.- •._-- R53W , ---..,~ R52W Fig. 1. Pinneo study area, Five leks and nests of five radio-marked hens were located. �WELLS R65W •••~ ••• r'" II •• ••••••• );tQh<'ii. ••••••• ... ... Elfi .~ ... .... ..8QQ : ~ I r-'B ~Rt ES JIL I E H vy- 26 - l- __.fI"!_ I ,...... """'" . 2. ~ Wy -3 l~ <, III... ~ i. ~. ••••• "\.... ~ I L~ TE GREELEY , ! 1\ ~RI VEf S I [ E ~ RE' ERYOI 'l1li C '-., r---.."'0111~ 0 I'--. ~ ~V ~ -L ~ r- ]_ COLORADO M I I I I I • I I I I I I I • LEKS 0 NESTS X .. • ,..,. :..... I R65W ---- ---~------- Fig. 2. Wells ........ - I RELEASE I I Ranch study area. Three R62W leks and nests I I SITE .... R63W I I R ,~ I ~ ~ -, :T 4N I I J '\ I ~ : '-,1,: : 't : --- ....t... Jl.,.J,."l, ..Jl.,. --- --- -_ ~------- -- R64W ,--~ ~ C:_T il:'Dci I I I I I : I ,......_ I ! : : T 5N ~ ~ r-, ~I I I T3 N ! I ~ r- I I I R. I I I T4 N ~ I I - "•..•.... -.__ •.......•••.. • l---- I T 6N i I I I I I I ! 3. " I I I I 0 I ~ 'f" I «:~ V If I I I : (RE EL ""y ., AREA / I I I I I I I T5 N: STUDY R91W." ... ." "" ~9 .W ••"." " I I I T6 N RANCH .•.. of two radio-marked 'T : 3N : ".1..,. ... ...J.... ... R61W I I : ... _I .• , 1 1: •••.••.••.•• R60W hens were located. I-' N I-' ~~ �122 Table 1. Location of and attendance at greater prairie-chicken leks formed by birds released at Pinneo, < < Colorado, 1991. VTM's Lek Legal location Range (West) Section Observation date .ti males NW 17 Apr 19 Apr 21 Apr 23 Apr 24 Apr 28 Apr 04 May 10 May 01 Jun 3 2 3 2 4 0 1 1 0 1 NE 24 Apr 26 Apr 28 Apr 04 May 10 May 11 Jun 4 4 2 0 0 2 54 5 SE 14 May 21 May 4 0 4 54 19 SW 03 Jun 05 Jun 08 Jun 11 Jun 20 Jun 7 7 6 3 1 4 54 19 SW 03 Jun 05 Jun 08 Jun 3 1 0 Township (North) Meters East Meters North 1 633700 4447300 2 54 4 2 629540 4447880 2 55 3 632960 4446800 2 4 630120 4461200 5 630680 4461280 Quarter observed �l.L') Table 2. Location of and attendance at greater prairie-chickens leks formed by birds released at Wells Ranch, '.'. Colorado, 1991. UTM's Lek Township (North) Legal Location Range (West) Section Observation Date N males Meters East Meters North 1 551721 4471620 5 63 13 SW 01 03 05 09 12 17 23 24 30 07 May May May May May May May May May Jun 6 5 5 4 4 3 3 0 3 0 2 565260 4476940 6 61 32 SE 04 Jun 06 Jun 07 Jun 10 Jun 12 Jun 19 Jun 4 1 1 6 5 0 3 568260 4477200 6 61 34 SE 04 06 10 12 19 4 3 3 1 0 Quarter Jun Jun Jun Jun Jun observed �Table 3. Nesting activity of greater prairie-chickens released at Pinneo and Wells Ranch, Colorado, 1991. I-' N Freq. Age Date of Date of nesting8 hatch Distance from Nearest lek{km) Clutch release area{km) size ~ !! eggs hatched Fate of nest pinneo 0976 y 31May 0906 A 0896 Unknownb 11 Jun #5 (5.5) 17.1 All Lost contact with hen after 11 Jun. 10 May 8 Jun #1 (0.6) 1.4 16 16 Flushed with 10 chicks on 21 Jul. y 31 May 27 Jun #4 (1.0) 14.8 11 10 Flushed with 3 chicks on 21 Jul. 0554 A 31 May 27 Jun #5 (0.7) 14.0 All Observed with 1 chick on 21 Jul. 1096 y 13 May Unknown #5 (1.2) 12.8 11 Unknownb Unknown? Was no longer on nest after 31 May. Not flushed with any chicks. Wells Ranch 0.~40 y 17 May 19 Jun #1 (9.8) 11.7 10 10 Not flushed with any chicks. 0378 A 04 Jun 6 Jul #2 (1.2) 14.6 11 9 Not flushed with any chicks. aDate when first observed on a nest. bClutch size was unknown because hen was not observed away from nest. cFailed to relocate nest after hen left. �125 0440, and 0378 were not flushed with a brood. However, this may not have been a true indicator of their brood status. It is possible that not all broods flushed when approached, that broods were separated from hens when the hens flushed or there may have been significant mortality among broods. Brood hens stayed in areas with good sand sagebrush and grass cover, and high concentrations of insects (grasshoppers). Telemetry Investigations Two hundred and fifty-five radio locations were obtained from April through December. Twelve birds (6 males, 6 females) accounted for 90% of the radio locations (Tables 4 and 5). At Pinneo, radio contact was lost with 7 birds (5 females, 2 males) within 5 days post-release while at Wells Ranch contact was lost with 8 birds (5 females, 3 males) within 7 days post-release. A 3-hour aerial search on 23 April was successful in locating 5 birds (3 females, 2 males) from Pinneo and 7 birds (4 females, 3 males) from Wells Ranch. However, ground contact was not made with 2 females from Pinneo nor with 3 females and 3 males from Wells Ranch. A second aerial search on 29 May was successful in locating 2 femaies from Pinneo and 2 males and 3 females from Wells Ranch. Radio contact was confirmed from the ground for all of these birds. Eight mortalities were documented; 3 males from Pinneo and 2 males and 3 females from Wells Ranch. Five mortalities were caused by mammalian predators and two by raptors. One female from Wells Ranch was shot on 27 November southwest of Riverside Reservoir on the Eagle's Nest Ranch. Radio transmitters were lost from 7 birds (3 males, 1 female at Pinneo; 2 males, 1 female at Wells Ranch). No evidence of predation could be found for any of these birds. As of 31 December, locations of 2 birds were known (one female at each site) and location/fate of 7 birds were unknown (Tables 4 and 5). The maximum distance any bird was located from the release site was 79 km. This bird was an adult female (0288) released at the Wells Ranch. She was located by aerial search on 29 May in the same area as several birds from Pinneo between Interstate 76 and Highway 34. She was found dead on 20 June. It is unknown if she was nesting. The mean dispersal distance for all birds at both sites was 14.5 km (Tables 4 and 5). Birds from Wells Ranch had a mean dispersal distance of 20.1 km compared to 9.0 km for Pinneo birds. This may indicate the Pinneo site has more suitable habitat within a smaller range than the area around Wells Ranch. Future habitat analysis will test this hypothesis. Females at both sites (Wells Ranch-27.0 km; Pinneo-12.l km) dispersed farther than males (Wells Ranch-13.1 km; Pinneo-3.8 km). This may indicate that females range farther when searching for suitable nesting habitat. Dispersal from the release site at Pinneo was oriented west and northnorthwest (Fig. 1). Four hens (1096, 0976, 0896, 0554) moved approximately 14 km north-northwest to an area between Highway 34 and Interstate 76. Two males (0079, 0018) localized approximately 5 km west of the release site where they were observed displaying on lek #2. One female (0906) localized within 1.5 km of the release site while the final hen (0968) was not located by aerial or ground search after the release. One male (0301) localized within 1 km of the < < �126 Table 4. Movements of greater prairie-chickens released at Pinneo, Colorado, 1991. N radio Freq. Sex locations Distance from release site where bird localized (km) Fate 0976 Y F 3 0968 A F 0 0906 A F 16 1.5 0896 y F 14 14.8 Nested successfully. Found transmitter on 31 Aug. 0554 A F 21 14.0 Nested successfully. Lost contact after 20 Sep. 1096 y F 20 13.2 Nest success unknown. Still alive on 31 Dec. 0301 y M 12 2.1 0280 A M 1 Found dead on 6 Apr. 0109 y M 3 Observed displaying on lek #1 in association with 0301 on 17 Apr. Found dead on 28 Apr. 0079 A M 20 0048 A M 4 0018 y M 21 17.0 Nested successfully. Lost contact after 11 Jun. Not located after release. 4.7 Nested successfully. Lost contact after 20 Sep. Observed displaying on lek #1. Found transmitter on 10 Jul. Observed displaying on lek #2 in association with 0018. Found transmitter on 27 Aug. Found transmitter on 23 Apr. 4.7 Observed displaying on lek #2 in association with 0079. Found dead on 20 Sep. <"' �127 Table 5. Movements of greater prairie-chickens released at Wells Ranch, Colorado, 1991. !! radio Freq. Sex locations Distance from release site where bird localized (km) Fate 0449 Y F 0 0440 y F 17 11.7 Nested successfully. Was shot on 27 Nov. 0378 A F 27 14.6 Nested successfully. 31 Dec. 0369 A F 4 2.8 0310 y F 2 0288 A F 5 78.9 Found dead on 20 Jun 79.1 km from release site. 0419 y M 14 16.4 Observed displaying on lek #3 in association with 0349 and 0339. Found transmitter on 13 Jul. 0410 y M 3 3.8 0349 A M 24 16.2 Observed displaying on lek #3 in association with 0339 and 0419. Found dead on 25 Oct 0339 y M 23 16.1 Observed displaying on lek #2 on 4 Jun. Observed displaying on lek #3 on 6 Jun in association with 0349 and 0419. Lost contact after 3 Oct. 0319 A M 0 Located by aerial search on 23 Apr. Lost contact after 23 Apr. 0089 A M 1 Found dead on 18 Apr. Not located after release. Still alive on Lost contact after 29 May. Found transmitter on 21 Nov. Found dead on 18 Apr. Found transmitter on 25 Apr. �128 release site and was observed displaying on lek #1 but later moved 5 km west of the release site near 0079 and 0018. Two males (0280, 0109) were found dead within 1.5 km of the release site within 30 days post-release. The final male (0048) lost his radio 0.8 km from the release site. The radio was found on 23 April. At Wells Ranch, dispersal was oriented east, south, and southeast (Fig. 2). Four birds (3 males, 1 female) localized 16 km east of the release site in an agricultural area of wheat and summer fallow. All 3 males were observed displaying on leks. Bird 0089, an adult male, was found dead on 18 April, 3.7 km west of the release site. On 25 April, the radio collar of 0410, a yearling male, was found 3.8 km west of the release site. Bird 0319, an adult male, was located east of the National Hog Farm Complex by aerial search on 23 April, however, ground contact was not made with this bird. The second aerial search on 29 May also failed to locate this bird. Bird 0449, a yearling female, was not located after release by air or ground search. Bird 0310, a yearling female, was found dead on 18 April, 1 km west of the release site. Bird 0369, an adult female, was located by aerial search on 23 April, 26 km east of the release site. On 30 April, this same bird was located 0.7 km south of the release site. Contact'was lost with this bird after 3 May. Her radio transmitter was found on 21 November approximately 7 km southeast of the release site on the National Hog Farm property. Bird 0440, a yearling female, localized 11.7 km southeast of the release site. A snowstorm in late October induced some migratory movements by the remaining radio-marked birds. One female (0378) from Wells Ranch was found approximately one mile south of Jackson Reservoir in a corn stubble field. One female (1096) from Pinneo wintered along the South Platte River in the corn stubble fields approximately 2-3 miles southwest of Prewitt Reservoir. Both of these birds were still in these areas as of 31 December. I also flushed one flock of 8 birds near the release site at Pinneo following the snowstorm in October. There were several other reports of flocks of prairiechickens being seen along the South Platte River between Hillrose and Prewitt Reservoir from October through December. LITERATURE CITED Aldrich, J.W. 1963. Geographic orientation of North American Tetraonidae. J. Wildl. Manage. 27:529-545. Beck, J.V. 1957. 25:8-12. The greater prairie chicken in history. Nebraska Bird Rev. Christisen, D.M. 1969. National status and management of the greater prairie chicken. Trans. North Am. Wildl. and Nat. Resour. Conf. 34:207-217. Cooke, W.W. 1898. Bulletin 37. Further notes on the birds of Colorado. An appendix to Colorado Agric. ColI. Bull. 44. Tech. Ser. 4:147-176. Evans, K.E. 1964. Inventory of greater prairie chickens. Colorado Dep. Game and Fish, Fed. Aid Rep. W-37-R-17. April; pp. 343-367. « �, and D.L. Gilbert. 1963. Grouse of the grasslands: the greater prairie chicken in Colorado. Colorado Outdoors. 12(6):15-18. Graul, W. 1975. Grassland boomers. Colorado Outdoors. 24(6):24-28. Grubb, T.G., and W.L. Eakle. 1988. Recording wildlife locations with the universal transverse mercator (UTM) grid system. U.S. Dep. Agric., For. Serv., Res. Note RM-483. 3 pp. Horak, G.J. 1985. Kansas prairie chickens. Wildl. Bull. 3. 65 pp. Kansas Fish and Game Comm. Johnsgard, P.A. 1983. The grouse of the world. Lincoln. 413 pp. Univ. Nebraska Press. Jones, R.E. 1963. Identification and analysis of lesser and greater prairie chicken habitat. J. Wildl. Manage. 27:757-758. Pusateri, F.M. 1990. Greater prairie-chicken recovery plan. Wildl. Fort Collins. 25 pp. Colorado Div. Schorger, A.W. 1944. The prairie chicken and sharp-tailed grouse in early Wisconsin. Trans. Wisconsin Acad. Sci., Arts, and Letters. 35:1-59. Schroeder, M.A. 1991. Movement and lek visitation by female greater prairiechickens in relation to predictions of Bradbury's female preference hypothesis of 1ek evolution. Auk. 108:896-903. _____ , and C.E. Braun. 1991. Walk-in traps for capturing greater prairiechickens on leks. J. Field Ornitho1. 62:378-385. Sc1ater, W.L. 1912. A history of the birds of Colorado. London, U.K. 576 pp. Stempel, M.E., and S. Rodgers, Jr. Iowa. Proc. Iowa Acad. Sci. Witherby and Co., 1961. History of prairie chickens in 68:314-322. Van Sant, B.F., and C.E. Braun. 1990. Distribution and status of greater prairie-chickens in Colorado. Prairie Nat. 22:225-230. Prepared by: « ��131 JOB PROGRESS REPORT State of: Colorado Project: W-167-R Upland Bird Research 17 Job Title: Population Dynamics of White-tailed Ptarmigan Period Covered: Author: Job __z__ Work Plan: 01 January through 31 December 1991 Clait E Braun and Kenneth M. Giesen Personnel: Kathy Martin, University of Toronto; Clait E. Braun and Kenneth M. Giesen, Colorado Division of Wildlife ABSTRACT Long-term studies of populations of white-tailed ptarmigan (Lagopus leucurus) were continued at hunted (Mt. Evans) and unhunted (Rocky Mountain National Park) areas in Colorado through 1991. Densities of breeding ptarmigan at Rocky Mountain National Park decreased while those at Mt. Evans were essentially stable. Nest success at Mt. Evans was poor in 1991 while that in Rocky Mountain National Park increased over that recorded in 1990. At least 10% of the banded birds alive at Mt. Evans were reported shot in 1991. Nine of the 10 banded birds reported harvested were banded in 1989 and 1990. Nonreporting of bands may be an important problem. ��lJJ POPULATION DYNAMICS OF WHITE-TAILED PTARMIGAN Clait E. Braun and Kenneth M. Giesen Long-term studies of trends in population size and investigation of reasons for fluctuations in size of tetraonid populations are lacking. Studies on the population dynamics of unhunted and hunted populations of white-tailed ptarmigan were initiated in Colorado in 1966 and have continued essentially uninterrupted at 2 sites. Studies of the unhunted population (Rocky Mountain National Park) identified possible short-term cycles of 7-8 years with an amplitude of 25-30% between high and low breeding densities. Conversely, studies of the manipulated population (hunted) at Mt. Evans have not indicated any cyclic pattern and it would appear that controlled hunting may mask any long-term trend that may occur. This study is designed to examine the question whether white-tailed ptarmigan are truly cyclic and whether hunting affects the apparent oscillations. P. N. OBJECTIVES The goals of this investigation are to be able to predict the length and amplitude of cycles in white-tailed ptarmigan in Colorado, to examine the impact of hunting on cycles, and to clarify underlying causes of the apparent cycles. SEGMENT OBJECTIVES 1. Conduct breeding (May-Jun) and brood (Aug-Sep) censuses ptarmigan using tape-recorded calls of males (breeding) (broods). 2. Censuses will be conducted on previously established, defined study areas at Mt. Evans (hunted) and at Rocky Mountain National Park (unhunted). 3. Capture (noose poles) and band (aluminum and plastic color-coded bands) all unmarked white-tailed ptarmigan encountered on study areas at Mt. Evans and at Rocky Mountain National Park. 4. Individually identify all ptarmigan observed on study areas at Mt. Evans and Rocky Mountain National Park through use of binoculars. 5. Make hunting season and bag limit recommendations for Mt. Evans and collect hunting data through use of volunteer wing barrels and hunter field checks. 6. Compile data, analyze results, and prepare progress of white-tailed and chicks reports. �134 STUDY AREA AND METHODS Areas investigated were Mt. Goliath-Mt. Evans in Clear Creek County and at Tombstone Ridge-Sundance Mountain to Fall River Pass in Rocky Mountain National Park in Larimer County. The physiography, geology, location, and vegetation of these study areas have been previously described (Braun 1969, 1971; Braun and Rogers 1971; Giesen 1977). Ptarmigan were located through use of tape-recorded calls (Braun et al. 1973), captured through use of telescoping noose poles (Zwickel and Bendell 1967) as described by Braun and Rogers (1971), and classified to age and sex and banded following Braun and Rogers (1971). Age of chicks was estimated following Giesen and Braun (1979). Numbered plastic bandettes were not used as in earlier years (Braun and Rogers 1971) as a color-code system using up to 4 different colored plastic bandettes was instituted in 1977-78. A check station was operated on the Mt. Evans highway during the opening weekend of the ptarmigan season in that area. A volunteer wing collection station was available to hunters in the area when the check station was not in operation until the season closed. RESULTS AND DISCUSSION Breeding Densities Mt. Evans.--Timing of breeding events in the Mt. Evans area was about the same in 1991 as in 1990. During the May-early June interval, 13 pairs were identified, 3 less than in 1990. Only 5 of 13 males identified were yearlings while 8 of 20 hens were yearlings. Thus, recruitment of yearlings was higher than in 1990. Despite fewer males in 1991, the breeding density increased slightly (Table 1). Rocky Mountain National Park.--Timing of breeding events on the Trail Ridge study area was similar to the long term average and about 1 week later than in 1990. Surveys of ptarmigan on breeding territories along Trail Ridge Road in May and June indicated a minimum population of 61 birds and included 26 pairs and 9 unpaired males. This represents a 19% decline over 1990 and is less than 50% of the breeding densities recorded during peak years in 1969 and 1976. The decreased breeding density reflected low survival of banded adult males (34 of 57, 59.6%) and females (17 of 38, 44.7%) from 1990. There was no recruitment of chicks banded in 1990, although yearlings comprised 42% of all adult ptarmigan identified in 1991. Nesting Success and Brood Size Mt. Evans.--Twenty hens were located during August-early September 1991 on or immediately adjacent to the study area. Sixteen (80%) were without broods while only 4 had broods. Average brood size to 1 September was poor (2.7 chicks/hen). Hatch dates for 13 chicks varied from 13 July to 3 August. Rocky Mountain National of hens with broods and 16 hens observed during success rate of 50%, an Park.--Nest success was estimated from the proportion without broods observed in July and August. Eight of summer surveys were with broods for an estimated nest 85% increase from 1990. The median hatch date �1,)::> calculated from wing molt of 28 juveniles was 12 July (range 4-24 July) and was similar to the 1966-90 average. Brood size in August averaged 3.4 chicksjhen (range 2-6). Table 1. White-tailed ptarmigan breeding densities (birdsjkm2), Colorado 1966-91. Study area Year 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 Rocky Mountain National Park (5.5 km2) 11.3 9.8 11.5 12.0 9.6 9.1 8.7 7.8 8.0 11.1 13.5 12.9 10.7 8.7 8.4 8.2 7.8 6.7 5.8 6.0 4.5 6.0 5.4 6.2 7.6 6.7 Mt. Evans (4.0 km2) 3.0 2.7 2.7 2.2 2.0 4.2 7.5 6.2 6.2 6.2 6.7 > 6.0 7.5 10.3 9.5 9.0 6.5 6.5 8.0 8.0 6.5 5.0 7.5 8.0 8.0 8.2 Harvest Mt. Evans.--The hunting season at Mt. Evans in 1991 opened on 14 September and closed on 6 October (23 days) with a bag and possession limit of 3 and 6. Thus, the season was delayed 1 week from the statewide opening as it was in 1981 and 1986-90. The season opening was delayed 2 weeks from 1978 to 1980 and 1982 to 1985. Prior to 1978, experimental seasons were in effect (19701976) or the season opened with the statewide grouse seasons (dates from 17 Aug to 14 Sep). Unlike 1987 and 1988 when the Mt. Evans road was closed due to reconstruction (Lincoln Lake area in 1987, above Ptarmigan Flats in 1988), �136 the Mt. Evans road was open to Summit Lake throughout the season. Twenty-one hunters were checked on 14-15 September. These hunters observed 50 ptarmigan and harvested 4 of which all were banded. Nine additional wings (7 on 20 Sep 1991, 2 on 29 Sep 1991) were received in the Mt. Evans wing collection barrel (but no bands) and bands from 6 birds were received through the mail. Thus, at least 10 banded birds were harvested in 1991 (4 of 13 possible from 1989, 31%; 5 of 25 possible from 1990, 20%; 1 of 53 possible from 1991, 2%). This represented at least lOX of the birds believed alive in September 1991 on the study area. However, it is possible (probable) that not all banded birds harvested were reported. LITERATURE CITED Braun, C. E. 1969. Population dynamics, habitat, and movements of whitetailed ptarmigan in Colorado. Ph.D. Thesis, Colorado State Univ., Fort Collins. l89pp. 1971. Habitat requirements of Colorado white-tailed ptarmigan. West. Assoc. State Game and Fish Comm. 51:284-292. Proc. _______ , and G. E. Rogers. 1971. The white-tailed ptarmigan in Colorado. Colorado Div. Game, Fish and Parks Tech. Publ. 27. 80pp. _______ , R. K. Schmidt, Jr., and G. E. Rogers. 1973. Census of Colorado whitetailed ptarmigan with tape recorded calls. J. Wildl. Manage. 37:90-93. Giesen, K. M. "1977. Mortality and dispersal of juvenile white-tailed ptarmigan. M.S. Thesis, Colorado State Univ., Fort Collins. 55pp. _______ , and C. E. Braun. 1979. A technique for age determination of juvenile white-tailed ptarmigan. J. Wildl. Manage. 43:508-511. Zwickel, F. C., and J. F. Bendell. J. Wildl. Manage. 31:202-204. 1967. __;;;&,--V...;;_·_.z._.--+~-==""""",,_ Prepared by _ Clait E. Braun Wildlife Research Leader Kenneth M. Giesen Wildlife Researcher C A snare for capturing blue grouse. � https://cpw.cvlcollections.org/files/original/12ebd80b9cd347febbcd8e280db11ae4.pdf 654db8da94918a4428c03d42d3ae415d PDF Text Text lJl JOB FINAL REPORT State of ~C~o~l~o~r~a~d~o~ _ Project: W-152-R(W-165-R Upland Bird Research Work Plan: -=2..=1 : Job _3_ Job Title: Sandsage-Bluestem Prairie Renovation Period Covered: 01 January 1984 through 31 December 1991 Author: Warren D. Snyder ABSTRACT Prescribed burns, tillage~herbicide and herbicide renovation, and revegetation treatments were applied to sites within the Tamarack Prairie during 1984-86 to increase the proportion of tall, warm-season grasses and height-density of standing residual, and reduce the composition of sand sagebrush (Artemisia filifolia). Evaluations through spring 1991 demonstrated that prescribed burns that coincided with emergence of warm-season grasses reduce4 heightdensity for: 1 year post treatment on all sites, several years if precipitation was below average, and increased it for one or more if precipitation was above average. Precipitation received prior to and after a fire was the primary factor influencing vegetation" response. The heightdensity value of sand sagebrush was severely impacted for several years by fire. Burns conducted in late spring more severely impacted sage, however, it was seldom killed. Composition of blue grama (Bouteloua gracilis), needleand-thread (Stipa comata), prairie sandreed (Calamovilfa longifolia), and sand dropseed (Sporobolus cIYtandrus) was not markedly altered by a single spring burn. Needle-and-thread, a cool-season species, was most severely impacted during the first growing season, but recovered with dramatic seed production during the 2nd year. Crown cover of blue grama, prairie sandreed, and sand bluestem (Andropogan hallii) increased when burned under favorable moisture conditions, whereas sand dropseed was little affected by fire. Sample sizes for perennial and annual forbs were too small to provide meaningful data and no consistent trends were evident. Western ragweed (Ambrosia psilostachya) increased after the 1985-86 burns but total forb composition was low and number of species did not change after fire. Revegetation (tillage to remove shallow-rooted species, application of atrazine herbicide, and seeding' to tall warm-season grasses) was effective in rapid co~version of rangeland to taller, more lodge-resistant cover with greater r'esLduaL height-density. Drilling tall grasses into undisturbed sod was ineffective as between-row competition severely suppressed seeded grasses where an interseeder opened furrows for planting prior to this study (1982-83). Tillage-herbicide renovation (1986) was highly effective in removing between-row competition and increasing composition and height-density of deep-rooted seeded species. Tillage and tillagefherbicide treatments effectively renovated deep-rooted, warm-season tallgrass species. Aerial application of 2,4-D herbicide to thin stands of sand sagebrush resulted in near total kill of sage and severe repression of forbs. Sand sagebrush was used extensively for nesting and loafing by prairie grouse. In spite of greater grass height-density within the ungrazed Tamarack Prairie than on adjacent grazed rangelands, the latter received more use by pra1r1e grouse. Thus, factors in addition to height-density must be considered in management for prairie grouse. However, prairie grouse use of the Tamarack Prairie increased steadily during the study~ "Management options to enhance the quality of the Tamarack Prairie are presented with direct application to other rangelan~s. .._._-_.._._- 111~11111~)II~)jil~imlli~~mii~1 ~mlli~11 BDOW026417 �138 RECOMMENDATIONS 1. Rangelands in eastern Colorado, such as the Tamarack Prairie, should be retained in a vigorous subclimax condition for prairie grouse. Fire, tillage; and partial revegetation are recommended in preference to grazing where management· can be directed toward prairie grouse. Such sites should not be left completely idle and unmanaged, nor should they receive annual grazing. 2. The Tamarack Prairie lacks a strong base of forbs, legumes, and foodproducing vegetation essential to prairie grouse and the site remains dominated by mid and shortgrass species of marginal value. Most other eastern Colorado rangelands have similar composition characteristics. Management procedures should strive to provide a mosaic of diverse habitat components, including food, that will provide year-long needs of prairie grouse in close proximity. 3. If ungrazed, sites (~ 130 ha) should be burned on a 8-10 year rotation to remove residual. More frequent burns, at 5-10 years, should be used where tall, warm-season grasses occur within revegetated or interseeded sites. At least 0.75 m of soil moisture should be available (based on soil probe samples) before burning. Burns should be timed to coincide with emergence of warm-season grasses following procedures summarized by Wright and Bailey (1980). Burns, conducted where sand sagebrush exceeds 15% crown cover, should be less frequent and relatively smali « 16 ha) because of slow sandsage recovery. 4. Aerial application of herbicides should not be used to thin sand sagebrush. Where needed, spraying with a tractor-mounted sprayer should be conducted in narrow, linear strips, 10-30 m wide, to create mosaics of openings that impact < 25% of the site. Less than 50% of the openings should be revegetated to tall, warm-season grasses « 1 ha). Other openings should be cultivated and seeded to seed-yielding annual forbs/grasses and to perennial legumes to provide green leafy vegetation. Periodic tillage should be used to renovate these sites and annual disturbance tillage may be needed on some sites. 5. Tillage destruction of shallow-rooted grasses, especially needle-andthread, and seeding of tall, warm-season mixtures into small « 1 ha) linear tracts is r~ommended in preference to int~rseeding. Su~h sites should be distributed throughout the Tamarack Prairie with a revegetation goal of 33% of the area. Where interseeded strips already exist, tillage-herbicide renovation combinations can used to release interseeded species. Seed rates should be reduced below those normally recommended in attempts to obtain moderate (as opposed to dense) stands of seeded grasses. 6. Sites dominated by indigenous pra1r1e sandreed, sand bluestem, and other deep-rooted vegetation should be renovated using a prescribed burn followed by shallow disturbance tillage to remove competing shallowrooted species such as needle-and-thread and blue grama. Direct seeding, using a range drill, into existing undisturbed range should not be used in this semiarid climate as competition from existing vegetation is too severe. �7. Application of atrazine herbicide (at low rates) (Snyder 1990~) effectively suppressed annual forbs within revegetated ~nd renovated sites (previously interseeded) and allowed switchgrass and bluestems to quickly establish. even during years when spring moisture was marginal. Other newly registered applicable herbicides may be substituted for atrazine in future plantings. 8. Two or more, 2-ha winter food plots dominated by grain sorghums are recommended on a 2 to 3-year trial basis within the Tamarack Prairie. Their use by, and retention of, prairie grouse through winter should be monitored. Patches « I ha) of alfalfa, vetches, and other legumes should be tested and monitored. Fire lines; prepared during the summer prior to burning, should be planted in fall to winter annuals such as rye, triticale, or winter wheat to provide green succulents through winter. Fire lines prepared in early spring should be planted to oatslegume (alfalfa) combinations. 9. Several large unused irrigation wells exist north of 1-76. Potentially, one of these could be changed from off-season to growing season use and water could be pumped under 1-76 to irrigate up to 65 ha within the Tamarack Prairie to raise alfalfa, corn, sorghums, etc. to benefit prairie grouse. 10. At several windmills, overflow basins, previously underlain with sheet plastic, should be tilled, seeded to legumes, and managed to retain wetmeadow vegetation. Change-of-use permits for the windmills should be obtained and drip systems (equipped with filters) should be installed to distribute water directly from the windmills. Water in eXisting stock tanks should be lowered to shallow «1 dm) depths to reduce drowning by passerines and still provide water for mourning doves, deer, and other resident wildlife. 11. Management procedures, previously described, are not compatible with intensive livestock grazing management. If grazing is used as a substitute for burning on part of the Tamarack Prairie, it should not be replicated more frequently than once ever 4 - 5 years (> 75% would remain idle each year) on sites.no larger than 40 - 50 ha using electric fences.to facilitate rotations among years .. Moderately·intense.fallwinter grazing to remove residual 'and disturb the soil while not impacting p Lant; vigor is t~e preferred procedure. 12. Vehicle noise along Interstate 76 is hypothesized as a deterrent to lek establishment closer than 1 - 1.5 km to the highway, which in turn may limit other use of the Tamarack Prairie by prairie grouse. If prairie grouse use of portions of the Tamarack Prairie proximal to 1-76 does not increase within the next 3-5 years, even with provision of food plots, new management alternatives should be considered. Among these could be exploration of exchange-af-land, or exchange-of-use for other sites away from the Interstate. For example, the narrow west 3.2 - 4.8 km (2-3 miles) of the Tamarack Prairie could be exchanged for lands south of the property that are managed by the Colorado State Land Board. Portions of rangelands north of 1-76 might also be traded while retaining hunter access rights.' Neither receives significant use by prairie grouse nor is it managed for other wildlife at the present. 13. Predator calling and trapping activities should be encouraged on and adjacent to the Tamarack Prairie to manage numbers of coyotes. �141 SANDSAGE-BLUESTEM PRAIRIE RENOVATION Warren D. Snyder INTRODUCTION Greater prairie-chickens (Typanuchus cupido) were declared endangered in 1973 under Colorado's Nongame, Endangered, or Threatened Species Conservation Act (Title 33, Article 8, Colorado Revised Statutes) (Pusateri 1990). Through the 1960's and early 1970's only a few hundred were present in Yuma County based on lek inventories. Prairie-chickens increased dramatically, and extensive range expansions occurred in the late 1970's and early 1980's, with several thousand birds present by 1982 (Van Sant and Braun 1990). These higher numbers have been sustained since. Attempts to restore the Division of Wildlife's Tamarack Prairie in eastern Logan County for greater prairie-chickens were begun in 1978. A habitat management plan, prepared by W. Busby for W. D. Graul, was submitted to R. Desilet in December 1978. Recommendations of a State Wildlife Areas Planning Committee were also submitted to R. Desilet by R. M. Hopper in October 1978. Nongame research personnel began vegetation sampling on the site, and, based on Busby's recommendation, an interseeder was acquired and native grass seeding was initiated in 1982 and 1983. Following appraisal of initial results of the interseeding effort, a consultant (L. M. Kirsch) was contracted to provide habitat renovation recommendations (L. M. Kirsch to R. M. Hopper and W. D. Graul, Sep. 1983). This study was designed to measure the impact of manipulations and to provide recommendations for future prairie restoration efforts. In 1984, the year this study was initiated, 36 prairie-chickens, trapped in Yuma County, were released on the Tamarack Prairie. Prior to the release, 1 male had been found by L. R. Crooks (District Wildlife Manager) on private rangeland during a survey of the Tamarack Prairie and adjacent lands. An additional 40 birds from Yuma County were released on the Tamarack Prairie in 1985. That spring, 5 leks (1 to 7 males/lek) were found on lands primarily south of the Tamarack Prairie. The resident population has slowly increased and expanded since then. In recent years (1989-91) a small, but expanding, population of plains sharp-tailed grouse (~ phasianellus jamesi) has become established in the area. These birds are presumed transients from Nebraska. STUDY AREA The Tamarack Prairie is the rangeland portion of the South Platte Wildlife Area (Tamarack Ranch) south of Interstate Highway 1-76 (Fig. 1). It contains approximately 1,820 ha bounded on the east by Logan County Road 93, on the west by Colorado Highway 25, and by private land and state school lands containing sandsage-bluestem prairie on the south. The elevation approximates 1,160 m progressively rising in sandy ridges and hills southeast from the South Platte River Valley. Soils are Dailey and Valent loamy sands (Amen et al. 1977). Precipitation averages approximately 40 cm (16 in.) per year. D~minant vegetation includes blue grama, needle-and-thread, sand dropseed, and prairie sandreed along with sand sagebrush. Sand b1uestem and other warmseason grasses, which are still present, were believed more prevalent prior to years of intensive grazing. �142 The Tamarack Prairie was purchased by the Colorado Department of Game and Fish (Division of Wildlife) in 1949 as part of the Tamarack Ranch, The property was primarily acquired for waterfowl hunting along the South Platte River. The rangeland portions of the Ranch continued to be grazed by cattle under lease agreements through 1977. Prior to that time it is assumed they were annually grazed by livestock following settlement of the locality. Portions of the Prairie were homesteaded and farmed, and then reverted back to rangeland. Interstate Highway 1-76 was constructed through the rangelands and completed in the late 1960's. �I T I I i I I i i i J .. l Ir III ~ Ir \ '-" --___j _ CIt N , __ •• 4 ,• ,, ,__ , 0 N -----------;~~---------N~! I I ! I I I ...I • t• , f•• Q 1&1 - ••~~ N >< E z- : ,~ ,,~ , Z 1&1 1&1 .J e o Fig. 1. Prescribed burns, renovation treatment, and revegetation treatments conducted within the Tamarack Prairie, eastern Logan County, Colorado. �144 P. N. OBJECTIVES Test renovation and revegetation techniques for increasing standing residual height-density of grasses, the proportion of tall warm-season grasses within the composition, and for reducing the quantity of sand sagebrush to <30% canopy cover in an ungrazed sandsage-bluestem prairie on the South Tamarack, South Platte Wildlife Area in northeastern Colorado. METHODS Following coordination with management personnel concerning site selection and fire management procedures, and design coordination with D. C. Bowden (statistician), two burn (treatment) sites were selected in 1984 (burns 1-84 [50 hal and 3-84 [20 hal) for evaluation (Fig. 1). The sites were mapped, gridded at 100-m intervals, and random vegetation sampling sites were positioned using a measuring wheel and compass. Approximately 5-m wide disked fire guards were placed around the burns and control areas. Burn 1 contained 9 treatment and 9 control vegetation sampling transects and burn 3 contained 4 of each. Each transect point was permanently marked using a steel post; metal pins were used to mark vegetation lines extending away from the point. In preparation for fires in 1985, 3 areas were selected, mapped, gridded at 100-m intervals, and subdivided into equal units. Random selection was used to determine burned and control portions. Due to dry conditions in spring 1985, only part of each of the 3 treatments was burned. The remainder of each was burned in spring 1986 (Fig. 1, Table 1). Each of the 3 areas contained 24 transect points (12 in each treatment and control area), or 72 transects among all areas. The 1985 burns encompassed 17 transects; 19 were contained in the 1986 burns with identical numbers in their control areas. The size and dates of burns conducted in 1984, 1985, and 1986 varied (Table 1). Management personnel tilled a fire guard around each burn, back fired along the leeward edge of the burn, and then used a progressive series of headfires. Propane torches were used in fire ignition and vehicle-mounted pumper units were used to control fires. Table 1. Characteristics of prescribed burns conducted during 1984-86, Tamarack Prairie Colorado. Phenolo~y (hei~ht cm) Stipa comata Calamovi1fa spp. ac Transects Date burned Year Site ha 1984 1 3 50.0 20.0 1985 1 2 3 17.8 26.7 14.5 124 49 44 66. 36 9 4 4 May 4 May 5 8 4 16 May 16 May 16 May 59.0 146 17 46.0 19.0 27.1 114 47 67 7 4 8 Subtotal 108.2 268 19 Subtotal 1986 1 2 3 30 Apr 23 Apr 23 Apr 15 Not emerged 50 (heading) 25 15-25 15-25 30 15 7.5-15 �14J Height-density indices (HDI) of residual vegetation within all treatments and controls were taken following Robel et al. (1970) as modifie~ by L. M. Kirsch (unpubl. rep., U. S. Fish and Wildl. Serv., Jamestown, N.D., 1977; the highest obstructed 0.5 dm was recorded using a 0.5-dm diameter square Robel pole). Sampling was conducted in March and early April subsequent to major snowfall and prior to significant green-up of vegetation. The first 4 readings (1 from each side of the Robel pole) were taken starting 5 paces to the right of the transect post, and subsequent readings were taken at 8-pace intervals in a rectangular direction around each transect within all burns and their controls. Ten stops (40 readings) were completed per transect. Sample size was reduced to 24 readings/transect in 1991. Grass-forb or sand sagebrush were recorded separately based on the primary visual obstruction when reading the Robel pole. Height-density indices of grass-forb, sandsage, and total (combined) vegetation were tested by analysis of covariance, analysis of variance, ~ tests, and regression analysis. Crown cover, species composition, and frequency of vegetation within the 1-84 and 3-84 burns were measured using metric-belt transects (Schmutz et al. 1982) during pretreatment (spring 1984) and post-treatment (mid to late summer) intervals in 1984, 1985, 1987, and 1989. A 50-m line extended from the steel post in a randomly selected direction. It was divided at 0, 12.5, 25.0, 37.5, and 50.0 intervals, and 5 7.9-m transect lines were extended 90· to the left, right, left, etc. at 5 points along the line. A 0.1-m2 frame (31.6 cm/side) was divided in half and one-half was further divided into 5 0.01-m2 segments. This frame was positioned 25 times along each 7.9-m line yielding a 2.5-m2 sample/line or 12.5 m2 /transect. Crown cover occupied by each species within each sample frame was visually estimated. Beginning in 1985, monitoring intensity was reduced by sampling alternate frames (13 instead of 25; 52%) along each 7.9-m line or 6.5 m2 /transect. Data were adjusted for analysis to directly compare with that obtained in 1984. A point-frame vegetation composition sampling procedure (Floyd and Anderson 1983) was used in mid to late summer to sample crown cover within the 1985 and 1986 burns and controls during mid to late summer from pre- through posttreatment years. A framework 0.5 x l-m in size, on telescoping legs, contained a grid of fluorescent-yellow, nylon fishing line spaced at 1 dm intervals to yield 36 sample points. A second grid was positioned 1 dm directly below the first, providing 2 pairs of cross-hairs on a vertical plane to sight past in recording vegetation. Each transect line was in a randomly selected direction, with 3 lines (7-m long) running perpendicular (left, right, left) to the main line at randomly selected points. The 36-point frame was positioned 4 times (at every other m) along each of the 3 lines yielding 144 points/line and 432 points/transect. Analysis of covariance was use in all crown cover evaluations. Within sites where prescribed burns and wildfires occurred in 1984, 1985, and 1986, individual sand sagebrush plants in clusters of 5 or 10 in random locations were marked with metal pins and monitored for survival in comparison with controls in proximal unburned locations. Precipitation gauges were placed within or near burns 1-84 and 3-84 in 1984, but could not be read immediately following each rain. Rainfall data were obtained from M. J. Gardner at the Tamarack State Wildlife Area headquarters, from H. F. Hamilton (Crook), and the U.S. Weather Service.recording station �146 (south of Sedgwick). Two automatic rain gauges that recorded each 0.025 cm (0.01 in) of rainfall were installed near burn 1-84 and east ,of burn 3-85 in spring 1985 to monitor rainfall through the growing season. Four automatic gauges were used in 1986 and subsequent years. Information was supplemented with U.S. Weather Service data during winter months as the gauges did not accurately measure snowfall. Attempts to use a soil moisture meter to determine percent moisture at 5, 15, and 30-cm depths in 1984 were discontinued because measurements were neither accurate nor meaningful. A soil probe (steel rod), approximately 1.9 m long and 1.4 cm in diameter was used to measure accumulated soil moisture starting in 1985. This device, used by agronomists for determining subsoil moisture in croplands (D. E. Smika, pers. commun.), provided reliable indices of spring subsoil moisture accumulations in the fine sands. Tests showed that when the probe was inserted by hand into the ground it consistently penetrated the moist soil, but stopped in the transition zone to dry soil. Sandy soils retain about 2.5-3.8 cm moisture/0.3 m (1 to 1.5 in./ft.) (D. E. Smika, pers. commun.). The probe was used in spring before top soils dried, hindering penetration. Leafing, budding, etc. status of plants resident on the Tamarack Prairie' was monitored at approximate 10-day intervals each spring to determine differences in plant phenology among years. Approximately 21 ha of rangeland dominated by needle-and-thread and blue grama within the northeast to middle part of the Tamarack Prairie were selected for revegetation in late winter 1985. Nineteen strips of varying length, each 26m wide, were placed perpendicular to prevailing winds. The strips were double disked with a tandem disk by management personnel after the ground thawed in late winter. Supplemental treatments included harrowing with a spike-toothed harrow and application of atrazine herbicide at 0.75 kg/ha (0.67 lb/ac); Seeding of a switchgrass-dominated tall, warm-season grass mixture followed within 1 or 2 days of herbicide application in late April 1985. Heightdensity and point-frame sampling techniques were used in subsequent evaluations. In 1986, management personnel shallow disked, applied atrazine herbicide at 0.75 kg/ha (0.67 Ib/ac), and harrowed the majority of 30 small tracts that had been interseeded to tall, warm-season grasses in 1982-83. Most contained fair to good stands which were suppressed by competition with native vegetation and annual forbs. Some, containing poor stands of grass, were reseeded. Heightdensity and metric-belt pretreatment sampling had been conducted on 1 location, so renovation treatment was applied to one-half (randomly selected) of the site. Each half contained 15 random transects (1 was subsequently lost in the treated portion), each with a single 7.9-m metric-belt line. Posttreatment evaluations of the treatment were continued. Height-density sampling of residual vegetation was also conducted on 10 addition (randomly selected) interseeded, renovated tracts during post-treatment intervals from 1987 through 1990. A herbicide application to reduce sandsage canopy cover was implemented by management personnel in spring 1985 within a 83.3-ha site in the east central portion of the Tamarack Prairie (Fig. 1). An aerial applicator was contracted to apply 0.84 kg/ha of 2,4-D low vol with 18.7 £/ha of water (0.75 Ib/ac of �herbicide in 2 gal of water) plus a wetting agent in 18.3-m wide strips at 36.6-m wide intervals (50% cover) during the bloom state of ~andsage (early Jun). The site, containing varying moderate to high densities of sandsage, was gridded and 12 random points under the center of the spray plane path were selected for vegetation sampling: At each of the 12 points, a transect was marked to obtain 8 36-point frame samples at I-m intervals . .In summer 1986, 38 transects, that contained nearly pure stands of prairie sandreed and sand b1uestem, distributed among 1984, 1985, and 1986 burns, their controls, 1985 revegetation strips, and 1986 renovation (interseeded) strips, were sampled. At each transect, percent seeded status was evaluated by ocular estimate. In addition, 12 height measurements and 12 HDI's were obtained within stands of both grass species. Residual grass-forb and sand sagebrush were sampled along 27 transects within grazed pastures primarily south of the Tamarack Prairie during early spring 1989 using height-density procedures. Stratification among pastures was used to insure that all pastures were included. Starting points of transects along accessible vehicle trails within, or adjacent to pastures were selected based on odometer readings derived from a random numbers table. All transects began at least 50 m from fencelines. Direction of transects was away from fencelines, if present, or randomly selected if a fence was not present. Walk-in traps (Schroeder and Braun 1991) were placed on leks near the Tamarack Prairie in early April 1989 and 1990 to trap greater prairie-chickens, plains sharp-tailed grouse, and hybrids of the 2 species. Trapped birds were marked with numbered aluminum leg bands and color-marked with plastic bandettes. Poncho-mounted, solar powered (1989) and battery-powered (1990) transmitters were placed primarily on females, but also on males, resident in the area. Additional males trapped on leks were banded and released. Four males and 4 females were radiomarked among 23 prairie chickens trapped in Yuma County and released on the Tamarack Prairie in April 1990. RESULTS Environmental Conditions Preciuitation and Phenology -- Above average precipitation in 1983 influenced the quality of residual vegetation in spring 1984. Annual precipitation was below the long-term average (approximately 41 cm) during 1984, 1985, and 1988 and was above average during 1986, 1987, 1989, and 1990 (Fig. 2). Wide monthly fluctuations and varied amounts within different parts of the Tamarack Prairie occurred during the study. The automatic recorders did not effectively document differences due to occasional problems with rodents, deer, batteries, etc. Most precipitation was received during the growing season with peak moisture occurring during May and June (Table 2). �148 _ x JAN-MAR ~APR 84 ~MAY ~JUN 85 a: w t2SJ JUL L2J AUG CSJ SEP l1li OCT-DEC 86 <C >- 87 88 89 90 o 5 10 15 20 25 30 35 40 45 50 55 PRECIPITATION (CM) Fig. 2. Monthly and annual precipitation (cm) from 1984 through 1990 in relation to the long-term mean, Tamarack Prairie, Colorado. �Table 2 •. Annual precipitation (em) from 1983 to 1990 and the long-term mean (IT Colorado. K) on Month the Tamarack Prairie, IT-K" 1983 1984 1985 1986 1987 1988 1989 1990 Jan 0_81 0.13 1.19 1.19 0.20 0.41 3.00 1.37 2_79 Feb 0.89 0.15 2.84 0.33 2_36 5.26 0.30 1.27 0.13 Mar 1.98 9.25 1.68 0.97 1.37 3.05 2.67 0.76 3.43 Apr 5.13 3.96 8.69 5.69 10.01 1.60 3.28 1.83 3.89 May 6.48 6.35 5.11 6.32 6.45 12.55 13.77 4.06 5.99 Jun 6.40 14.78 7.11 2.08 7.62 7.75 4.22 8.99 2.97 Jul 5.31 5.59 1.24 9.60 2.90 6.32 9.88 3.71 13.28 Aug 5.18 3.28 1.68 2.11 3.89 5.84 1.75 7.90 6.53 Sep 3.48 0.08 1.12 3.61 5.33 2.41 2.87 3.66 5.21 Oct 2.51 1.55 3.53 0.94 5.16' 1.17 0.10 0.56 2.67 Nov 1.12 5.50 0.03 2.69 1.09 3.12 0.33 0.33 3.30 Dec 1.25 1.01 2.03 3.05 1.12 2.11 0.32 0.59 0_00 40.54 51.61 36.25 38.58 47.50 51.59 42.49 35.03 50.19 Totals "tong-term (44 year) means, 1983, and data for winter months were obtained from U. S. Weather Service records. Spring temperatures, which influenced vegetation growth and conditions at time of prescribed burns, averaged above the long-term mean each all spring except 1984 (Fig. 3). Average temperatures .in March and/or April especially, were frequently above normal. Plant phenology on 1 May reflected accumulated (MarApr) temperature departures from the long-term average (Fig. 3). Warm-season grasses had not yet emerged at the time of the 4 May 1984 burns during an especially cool spring (Fig. 3, Table 1). In contrast, needle-andthread was already heading and prairie sandreed was about 30 cm tall when burns were conducted on 16 May 1985. Plant phenology was approximately 3 weeks earlier in 1985 than during the previous spring. Spring 1986 phenology was slightly in advance of that in 1985 and warm-season grasses had started growth when burns were initiated on 23 April. Weather factors prevented completion that day and most of burn 1 in 1986 was completed on 30 April. Plant phenology in spring 1987 was slightly behind that of the previous spring primarily because of below average temperatures in March (Fig. 3). Several hard freezes in mid to late April 1988 and on 1 May 1989 (accompanied by prolonged dry weather) retarded vegetation growth in relation to average temperatures. Phenology in spring 1990 was similar to that of preceding years. �1)U _MAR X ~ APR 84 ~MAY 85 ~JUN c: 86 -c w >- 87 88 89 90 o 5 10 15 20 25 30 35 40 45 50 55 TEMPERATURE (F) Fi~. 3. Monthly average temperatures (OF) from March through June, 1984-90 in relation to the long-term mean, as an index to vegetation phenology, Sterling, Colorado. �1.'::>1. Soil Moisture -- Little soil moisture was stored from one year to the next within rangelands. Native range vegetation used available moisture as late as October. Most winters were relatively dry (Fig. 2) and not conducive to major soil moisture accumulation. Therefore, precipitation received during the spring - early summer growing season was the primary determinant of annual growth. Cool-season grasses (needle-and-thread and western wheat [Agropyron smithiil) grew well during years with favorable winter and/or early spring moisture, whereas warm-season species were most benefitted by above average May through July rainfall. Soil probe samples (Table 3), obtained primarily in April and May (1985-90), revealed a close relationship with winter-early spring precipitation patterns (Fig. 2) among years. Winter and early spring moisture accumulated in the soil. However, as spring growth advanced, depletion of soil moisture accelerated. Most favorable accumulations occurred in spring 1986 through 1988; whereas, deficient precipitation in spring 1989 was evident (Table 3). Dry soils at or near the surface usually prevented sampling during summer. Table 3. Soil probe depths (dm) as an index to available soil moisture within the Tamarack Prairie, northeastern Colorado, 1985-90. ·SE Year Date Sample size X depth 1985 22 Apr 17 May 5 Jun 86 26 20 6.66 10.92 11.91 0.19 0.34 0.73 1986 4 7 19 30 Apr May May May 21 11 8 13 12.19 17.17 18.16 16.20 0.65 0.38 0.08 0.65 1987 1 Apr 29 Apr 6 May 2 Jun 2 Jul 32 28 16 16 17 12.04 9.86 12.86 11.21 7.32 0.23 0.32 0.47 0.42 0.49 1988 4 18 27 4 22 Apr Apr Apr May May 16 16 16 16 16 9.76 8.99 10.37 12.73 13.43 1.02 0.63 0.74 0.49 0.46 1989 12 2 15 31 9 Apr May May May Jun 16 16 12 16 16 4.13 4.48 4.89 5.24 6.30 0.16 0.22 0.51 0.40 0.45 1990 2 Apr 30 Apr 24 May 16 16 16 9.16 9.76 8.21 0.69 0.64 0.52 �152 Height-Density Sampling Within Burned Sites Height-density indices (HDI) sampled standing residual vegetation in late winter or early spring prior to significant growth of new vegetation. Therefore, they reflected the previous year's growth and the vegetation's ability to remain standing during the subsequent fall and winter. The 2 burns conducted in 1984 (burn 1-84 and burn 3-84, Fig. 1) were separated for analysis because of apparent site differences in soil, vegetation, and precipitation that affected HDI. The 1985 and 1986 sites possessed similar soils, vegetation, and precipitation, and trends were similar. Therefore, initial analyses were conducted separately, but data were combined for additional analysis (ignoring year; paired ~ of differences between treatments and controls) using the 6 site means. Grass-forb HDI -- Prescribed burns conducted in May 1984 on sites 1-84 and 384 reduced the height-density quantity of residual grass-forb vegetation in spring 1985 (f < 0.001) and subsequent years (Fig. 4, Tables 4 and 5). Grassforb residual within burn 1-84 recovered by 1987; that within burn 3-84 by 1991. Possible reasons for slower recovery in burn 3-84 include greater density of sand sagebrush (which resprouted rapidly and may have suppressed grasses), a fire that was more intense and potentially more damaging, poorer quality soils, and less rainfall during 1984. Grass-forb HDI's increased dramatically in 1988 from previous years within the burned and control transects on both burns (Fig. 4). Prescribed burns in 1985 also reduced grass-forb HDI through the first growing season (Tables 6 and 7, Fig. 4, f < 0.001). However, in contrast to the 1984 burns, grass-forb HDI recovered by the end of the 2nd growing season within the 1985 burns and exceeded the controls by the end of the 3rd year (f < 0.001). It remained high through the 4th year before fire enhancement was negated (Fig. 4). Departures of HDI means in 1991 were believed a result of error due to reduced sample size. Within the 1986 burns, grass-forb HDIs were not markedly suppressed following the first post-treatment growing season (f > 0.05) and exceeded controls by the end of the 2nd post-treatment growing season (Fig. 4, £ < 0.001). However, the enhanced growth caused by fire lasted only 1 year and was not evident in 1989 and subsequent years. �1::>,) Table 4. Height-density (dm) within burn 1-84 and controls during spring 1984-90, Tamarack Prairie, Colorado. Years Sandsage GrassLfb x x SE x SE Combined SE Burn 19848 1985 1986 1987 1988 1989 1990 0.256 0.134 0.232 0.208 0.589 0.527 0.434 0.022 0.019 0.028 0.026 0.061 0.048 0.062 0.856 0.313 0.358 0.526 0.836 0.753 1.011 0.082 0.040 0.022 0.082 0.078 0.066 0.109 0.372 0.162 0.256 0.258 0.627 0.577 0.540 0.041 0.017 0.025 0.024 0.051 0.041 0.051 0.123 0.043 0.083 0.171 0.179 0.118 0.122 0.334 0.355 0.368 0.309 0.622 0.628 0.475 0.038 0.025 0.024 0.035 0.622 0.047 0.035 Control 1984 1985 1986 1987 1988 1989 1990 0.253 0.295 0.301 0.224 0.587 0.571 0.430 0.021 0.023 0.026 0.012 0.051 0.041 0.044 0.814 0.687 0.643 0.847 1.045 1.052 0.893 F Values Pre- to post-treatment 1984-85 1984-86 1984-87 1984-88 1984-89 1984-90 8 b c d Grass-fb Sandsage Combined 47.50d 4.24 0.52 0.00 0.78 0.01 40.10d 14.98d 1.28 0.60 1.18 1.12 66.10d 10.51d 2.05 0.01 0.51 0.79 Pretreatment samples. f. < 0.050. f. < 0.025. f. < 0.001. Post-treatment Grass-fb 1985-86 1985-87 1985-88 1986-87 1986-88 1987-88 1988-89 1989-90 8.19c 0.64 . 20.39d 0.04 7.39c 0.15 1.10 0.65 Sands age Combined 8.06c 0.41 1.33 1.40 0.22 0.32 0.02 1.48 9.32d 0.01 6.73c 0.10 0.79 3.25 0.54 �154 Table 5. Mean height-density (dm) within burn 3-84 and controls during spring 1984-91, Tamarack Prairie, Colorado. Years 19848 1985 1986 1987 1988 1989 1990 1991 0.222 0.021 0.106 0.077 0.466 0.647 0.378 0.488 Combined Sandsage Grass/fb X SE 0.033 0.005 0.021 0.023 0.052 0.138 0.058 0.120 X SE 0.827 0.121 0.356 0.286 1.215 1.435 1.068 1.575 X SE 0.059 0.037 0.026 0.054 0.086 0.110 0.089 0.211 0.493 0.047 0.201 0.157 0.831 1.159 0.680 1.023 0.053 0.014 0.015 0.027 0.052 0.119 0.029 0.167 0.088 0.040 0.094 0.062 0.161 0.226 0.309 0.280 0.531 0.503 0.629 0.424 1.059 1.548 0.984 0.900 0.028 0.028 0.050 0.069 0.051 0.252 0.165 0.224 Control 1984 1985 1986 1987 1988 1989 1990 1991 0.183 0.191 0.200 0.216 0.625 0.934 0.446 0.417 0.021 0.028 0.020 0.046 0.064 0.264 0.058 0.119 0.935 0.797 1.044 0.688 1.494 1.888 1.579 1.369 F Values Pre- to post-treatment 1984-85 1984-86 1984-87 1984-88 1984-89 1984-90 1984-91 8 b C d Grass-fb Sandsage Combined 22.50d 5.96 6.62b 5.68 2.13 2.54 0.01 114.10d 35.31d 13.55c 0.85 1.44 0.88 1.18 187.00d 53.18d 9.22b 3.77 1.21 2.26 0.58 Pretreatment samples. f < 0.050. f < 0.025. f < 0.001. Post-treatment Grass-fb 1985-86 1985-87 1985-88 1986-87 1986-88 1987-88 1988-89 1989-90 2.32 2.77 2.89 4.34 1.73 0.70 0.21 0.04 Sandsage Combined 0.91 4.45 0.13 4.68 1.27 0.19 0.44 0.58 0.14 0.00 0.02 1.46 0.89 0.00 0.27 0.86 �Table 6. Mean height-density (dm) within 1985 burns and controls during spring 1985-91, Tamarack Prairie, Colorado. Burn Year 1 2 3 All Control 3 1 2 Al1 0.374 0.325 0.322 0.511 0.774 0.423 0.301 0.346 0.280 0.376 0.775 1.220 0.667 0.376 0.158 0.180 0.156 0.443 0.771 0.316 0.357 0.323 0.277 0.325 0.644 1.025 0.529 0.345 0.872 0.828 0.925 1.070 1.242 1.443 1.068 0.615 0.745 0.560 1.732 1.873 1.730 1.815 0.537 0.767 0.648 1.253 1.844 1.267 1.226 0.627 0.771 0.670 1.317 1.765 1.414 1.408 0.444 0.391 0.383 0.581 0.844 0.535 0.385 0.369 0.335 0.393 0.859 1.332 0.750 0.583 0.323 0.465 0.411 0.878 1.609 0.661 0.694 0.380 0.382 0.394 0.782 1.254 0.666 0.551 Grass-Forb 19858 1986 1987 1988 1989 1990 1991 0.507 0.138 0.419 0.780 1.081 0.542 0.604 0.381 0.119 0.472 1.028 1.297 0.552 0.526 0.180 0.053 0.244 0.667 0.850 0.315 0.663 0.379 0.110 0.~08 0.877 1.142 0.495 0.579 Sandsage 19858 1986 1987 1988 1989 1990 1991 1.008 0.417 0.523 0.938 1.313 0.688 1.500 0.679 0.342 0.591 1.167 1.582 1.050 1.000 0._642 0.438 0.769 1.144 1.385 0.955 1.853 0.744 0.388 0.672 1.125 1.446 0.914 1.438 Combined 19858 1986 1987 1988 1989 1990 1991 8 0.565 0.142 0.425 0.786 1.109 0.548 0.675 Pretreatment. 0.407 0.130 0.476 1.037 1.340 0.567 0.577 0.302 0.088 0.330 0.766 1.017 0.359 0.916 0.429 0.124 0.426 0.899 1.196 0.513 0.680 �.L)O Table 7. Height-density (dm) within 1985 combined burns and combined controls during spring 1985-91, Tamarack Prairie, Colorado. Years Sandsage GrassLfb SE X. x SE X Combined SE Burn 19858 1986 1987 1988 1989 1990 1991 0.379 0.110 0.408 0.877 1.142 0.495 0.579 0.036 0.011 0.038 0.058 0.076 0.034 0.074 0.065 0.095 0.050 0.103 0.112 0.149 0.168 0.429 0.124 0.426 0.899 1.196 0.513 0.680 0.031 0.009 0.031 0.046 0.069 0.031 0.077 0.062 0.059 0.106 0.191 0.136 0.145 0.157 0.380 0.382 0.394 0.782 1.254 0.666 0.551 0.026 0.024 0.029 0.048 0.096 0.046 0.069 0.744 0.388 0.672 1.125 1.446 0.914 1.438 Control 1985 1986 1987 1988 1989 1990 1991 0.323 0.277 0.325 0.644 1.025 0.529 0.345 0.027 0.019 0.034 0.051 0.086 0.047 0.021 0.627 0.771 0.670 1.317 1.765 1.414 1.408 F Values Pre- to post-treatment 1985-86 1985-87 1985-88 1985-89 1985-90 1985-91 Grass-fb Sandsage Combined 96.62d 1.59 7.78d 0.28 2.56 10.12d 95.85d 0.38 1.31 115.2sd 0.18 3.53 1.22 10.95d l.S9 Pretreatment samples. l < 0.050. e l < 0.025. d l < 0.001. 8 b Post-treatment Grass-fb 1986-87 1986-88 1987-88 1988-89 1989-90 1990-91 21.60d 17.29d 6.80e 0.27 1.36 9.41d Sandsage 2.71 1.47 1.53 Combined 2.02 3.04 2.43 3.15 8.7Sd 1.67 �0.7 1 E :5!.. 0.5 enz 0.4 W 0.3 :Eo07 co~~~, £; . CONTROL ~ en 0.6 ~,' z t-!- J: 0.4 C!J 0.3 jjj 0.2 J: .•....... ---I._'"""-_1.84 85 86 87 88 89 90 91 YEAR OI.....l.._.l:I;;....---'---'-_ YEAR 1-84 BURN enz w a t-!I C!J iii I 3-84 BURN '.2 1.2 1.1 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 "0 • E >o.a l- e;; Z w aU I l- ::I: 52 •.• w ::I: 0.2 85 86 87 88 89 YEAR 1985 BURN Fig. 4. , ,, 0.1 o~---~--------~--------~----84 85 88 87 88 8g go E __~_--b--~ 0.2 0.1 ~ ~ , ~ 0.5 ,, ,, ,, (!) W :J: " E 0.8 ~ ~ :J: ,~ , , ,' ,, ,,, , , , 0.9 0.6 90 91 •• •• ., •• •• YEAR 10 "' 1986 BURN Response of grasses and forbs to burning, Tamarack Prairie, Colorado. �158 Height-density was analyzed (ignoring year) using differences between site means for the 6 sites burned during 1985 and 1986. There was significant 1styear post-treatment reduction in grass-forb HOI but a significant increase in HOI by the 2nd year (Table 10). This 2nd year increase above pre-treatment levels was not evident within the 1984 burns. Sandsage HOI -- Sand sagebrush recovered more slowly than grass-forb vegetation and did not recover completely until 1990 (burn 1-84) and 1991 (burn 3-84) based on HOI (Fig. 5, Tables 4 and 5). Sage was potentially more severely impacted by the more intense fire within burn 3-84. Sage recovered more rapidly within the 1985 burns (Fig. 5, Tables 6 and 7) than within either the 1984 or 1986 burns (Fig. 5, Tables 8 and 9). Reasons for more rapid recovery subsequent to the 1985 burns are uncertain (sample sizes were small) but more favorable precipitation and cooler, less damaging headfires occurred in 1985. Fires were phenologically later than in 1984. There was no evidence that fire enhanced growth. Height-density indices derived from sagebrush obstructions were consistently higher than those for grass. Thus, it contributed to visual obstruction (HOI) in greater proportion than its composition within the study sites. Transect sample sizes for sandsage within the 1-84 burn, and the 1985 and 1986 burns were often small which increased variance. Therefore, HOI means were not as precise as those for grass-forb vegetation. �1.':>9 ~---4 ' , , , , , ,,~ 1.0 E :E. ~ enz 0.8 ~ o.e W J: e jjj J: ,, 'A_ 0.7 , ---~ ' ,, ,, , , E 1.6 CONTROL , o.s . 2.0 1.8 1.1 '0 •... CONTROL', ~1.4 en 1.2 Z w c 1.0 ,, ,, ,," ~ o m::I: 0.4 0.3 0.2 84 .•, 85 86 87 8g 88 0 go 84 85 86 87 88 89 90 91 YEAR 3-84 BURN • 1.8 1.8 ,, , ,, , , , 1.6 , ,, ,, ,, " ,, 1.6 \ 1.4 1.4 E ::E. 1.2 1.2 CONTROL I 0 ,.:.. 0.8 l: Cl W 0.6 l: 0.6 BURN 0.4 0.4 0.2 0.2..__---85 86 87 ---~-88 89 90 91 o •.....•. _-"-_-"-_-'-_-'-_-'-_-'-85 86 88 89 90 91 1986 BURN 1985 BURN of sand sagebrush 87 YEAR YEAR Response " ',/ , z w am 0.8 5. - ... _- I (i) t-!- Fig. , , ~ w 1.0 c I \ BURN 1-84 BURN enz \ 0.6 0.4 0.2 YEAR ~ ,' ::I: 0.8 0.5 -.-g , ,, to burning, Tamarack Prairie, Colorado. �160 Table 8. Mean height-density (dm) within 1986 burns and controls during spring 1985-91, Tamarack Prairie, Colorado. Burn Year 1 2 3 All 1 Control 2 3 All Grass-Forb 19858 19868 1987 1988 1989 1990 1991 0.326 0.304 0.120 0.543 0.723 0.386 0.398 0.314 0.337 0.437 1.359 1.159 0.666 0.470 0.259 0.256 0.202 0.824 1.019 0.361 0.418 0.292 0.290 0.226 0.855 0.963 0.438 0.425 0.294 0.277 0.234 0.387 0.794 0.360 0.385 0.316 0.312 0.404 0.858 1.141 0.561 0.440 0.233 0.213 0.269 0.644 1.074 0.432 0.376 0.275 0.265 0.289 0.600 0.983 0.435 0.395 0.829 0.756 0.885 1.265 1.353 1.557 1.258 0.417 0.300 0.833 1.667 1.250 1.000 1.167 0.618 0.690 0.671 1.197 1.810 1.343 1.463 0.682 0.693 0.750 1.237 1.650 1.430 1.387 0.345 0.342 0.325 0.494 0.886 0.454 0.502 0.317 0.331 0.413 0.873 1.145 0.564 0.457 0.298 0.323 0.357 0.743 1.304 0.509 0.631 0.320 0.332 0.357 0.679 1.116 0.500 0.547 Sandsage 19858 19868 1987 1988 1989 1990 1991 0.621 0.838 0.169 0.904 1.047 0.860 0.988 0.375 0.500 0.250 1.000 1.250 1.667 0.395 0.650 0.250 1.194 0.857 0.600 0.794 0.566 0.793 0.189 0.945 1.007 0.852 0.973 Combined 19858 19868 1987 1988 1989 1990 1991 a 0.421 0.470 0.127 0.616 0.815 0.486 0.567 0.315 0.339 0.434 1.355 1.159 0.473 0.498 Pretreatment years. 0.272 0.281 0.204 0.834 1.008 0.365 0.443 0.335 0.367 0.224 0.863 0.969 0.474 0.500 �Lo L Table 9. Height-density (dm) within 1986 combined burns and combined during spring 1985-91, Tamarack Prairie, Colorado. Years Sands age GrassLfb x x SE controls Combined x SE SE Burn 1985a 1986a 1987 1988 1989 1990 1991 0.018 0.023 0.039 0.092 0.067 0.052 0.027 0.292 0.290 0.226 0.855 0.963 0.438 0.425 0.566 0.793 0.189 0.945 1.007 0.852 0.973 0.066 0.098 0.046 0.122 0.086 0.085 0.111 0.335 0.367 0.224 0.863 0.969 0.474 0.500 0.020 0.028 0.063 0.087 0.058 0.050 0.029 0.064 0.054 0.080 0.210 0.204 0.147 0.115 0.320 0.332 0.357 0.679 1.116 0.500 0.5470 0.016 0.013 0.024 0.049 0.063 0.031 0.048 Control 1985 1986 1987 1988 1989 1990 1991 0.275 0.265 0.289 0.600 0.983 0.435 0.395 0.017 0.018 0.025 0.054 0.054 0.029 0.021 0.682 0.693 0.750 1.237 1.650 1.430 1.387 F Values Pre- to post-treatment 1985-87 1985-88 1985-89 1985-90 1985-91 1986-87 1986-88 8 Grass-fb Sandsage Combined 3.13 4.69b 0.51 0.41 0.19 5.26b 4.58b 28.22d 11.15d 8.45d 3.31 6.88c 0.57 1.11 8.71d 3.60 Pretreatment f < 0.050. c P < 0.025. d f < 0.001. b 26.42d 5.19b samples. Post-treatment 1987-88 1988-89 1989-90 1990-91 Grass-fb Sandsage Combined 33.88d 2.67 0.09 0.77 0.19 5.69c 10.99d 0.07 0.55 �162 Table 10. Height-density (dm) differences (1 test) among pre- and post-treatment intervals (ignoring year) within 6 sites burned in 1985 and 1986, Tamarack Prairie, Colorado. Post-treatment Interval 2 1 3 Year 4 5 Grass-forb Pretreatment 1st year post-treatment 2nd year post-treatment 3rd year post-treatment 4th year post-treatment 2.968 l. 91 7.34c l.06 2.25 l.11 2.758 0.72 0.40 l.19 3.42b 1.55 Sand Sagebrush Pretreatment 1st year post-treatment 2nd year post-treatment 3rd year post-treatment 4th year post-treatment 8.50c 8 b c 3.82b 0.25 2.738 0.96 l.32 0.07 2.07 0.50 0.28 Combined Pretreatment 1st year post-treatment 2nd year post-treatment 3rd year post-treatment 4th year post-treatment l. 76 2.728 3.108 Vegetation l.43 10.36c 0.34 l.82 0.79 l.06 l.3l 0.65 1.39 l. 91 0.12 f < 0.05. f < 0.02. f < 0.0l. Moderate fluctuation in percent obstruction occurred from year to year among the control transects (Table 11). Sampling differences among personnel, primarily in 1990, were partially responsible. Combined vegetation -- Sandsage obstructed 10 - 20% of the HDI readings within burn 1 in 1984 and within all burns in 1985 and 1986 (Table 11). Therefore, trends for combined vegetation (total HDI) were heavily weighted by grassforb, and followed grass-forb HDI trends (Figs. 6, Tables 4, 6 - 9). In contrast, about 50% of the HDI obstructions within burn 3-84 were by sandsage (Table 5), thus sage and grass-forb residual contributed equally to combined HDI (Fig. 6). �IbJ Grass-forb - Weather Relationships -- Early spring HDI's of residual grassforb (primarily grass) vegetation were impacted by previous-year growing conditions and over-winter snow accumulations. Correlations of grass-forb HDI's within controls, averaged for the 1985-86 burns, with precipitation received during (1) the previous year, (2) the previous spring and summer, and (3) the previous May-June interval showed the latter to be most closely related to subsequent HDI's (K - 0.73, l < 0.05, 5 df). When an index of average maximum snow depth was subtracted from the May-June precipitation index, the correlation coefficient increased (K - 0.82, l < 0.05, 5 df). Although both sample sizes and coefficients were relatively weak, they supported empirical observations that the relationship existed. Table 11. Percent obstruction of height-density readings by sand sagebrush within the burns and their controls, Tamarack Prairie, 1984-91. Burn 3-84 1-84 1985 1986 Treatment Control Treatment Control 51.5 13.8 18.7 15.5 11.0 38.1 50.8 4.8 21.3 15.2 15.7 13.6 38.3 44.0 7.1 20.1 6.4 14.7 15.3 7.8 48.8 50.0 9.1 20.4 9.0 12.4 1989 22.2 11.9 65.0 64.4 17.7 30.9 13.4 20.0 1990 18.3 9.7 43.8 47.5 4.3 15.4 8.7 6.6 49.2 50.8 11.8 19.1 13.8 15.3 Year Treatment Treatment Control 1984 19.4 16.7 44.8 46.3 1985 13.3 15.9 26.3 1986 12.7 12.9 1987 15.8 1988 Control 1991 Average' (controls) Overall average' among controls • All years conbined 50.1 13.7 19.4 20.9 13.5 �164 0.7 1.6 i b 0.5 A _1.4 - O.6 ! 12 ~ . ~ 1.0 Q 0.8 ffi o 0.4 !2 ~ 0.6 ~ 0.3 0 ~ 0.4 0.2 0.1 0.2 0 84 85 86 87 88 89 90 84 85 86 87 88 89 90 91 YEAR YEAR 5-84 BURN 1-84 BURN i ~ ffi Q ~ 1.4 1.2 1.2 i 1.0 - 1 ~ ffi 0.8 0.8 Q 0.6 ~ 0 0.4 0.6 ~ ~ 0.4 I, 0.2 O~----------'_~--85 86 87 88 89 90 91 '-------=---------85 86 87 88 89 90 91 0.2 YEAR 1985 BURN F.ig. 6. Response of all Colorado. vegetation YEAR 1986 BURN (combined) to burning, .- Tamarack Prairie, �16~ The HOI of grasses was suppressed by fire within all 4 burns through the 1st post-treatment growing season. However, in 1986, when May-June precipitation totaled 1.40 dm (contrasted to 1.15 dm [1984] and 0.84 dm [1985]) fire suppression seemed less severe (Fig. 7). Ouring the 2nd post-treatment growing season consistent HOI recoveries were noted. Grass-forb HOI increased 0.10 dm (burn 1-84) and 0.09 dm (burn 3-84) during 2nd post-treatment growing seasons when May-June rainfall was 0.84 dm. Grass-forb HOI increased 0.30 dm (1985 burns) when May-June rainfall was 1.41 dm; and increased 0.71 dm (1986 burns) when May-June rainfall was 2.03 dm. Post-treatment changes in grassforb HOI from the previous year within 1984-1986 burns were tested against rainfall departures from the 8-year (1983-90) mean for (1) May-June and (2) April-July intervals. The latter showed the highest correlation (I - 0.69, ~ < 0.01, 13 df). Rate of grass-forb recovery was directly related to rainfall. The same relationship was evident for sand sagebrush (Fig. 7). These data support the precipitation-snow compaction relationship with grass-forb HOI within the controls. Increased precipitation through spring intervals from 1986 through 1988 (Fig. 2) was the most logical reason for fire enhancement of grass-forb vegetation subsequent to the 1985 and 1986 burns. In contrast, below average precipitation was received for 2 years following the 1984 burns and enhanced HOI's did not occur. Trends of grass-forb HOI's within the 4 controls were similar through most of the study (Fig.8). All HOI's increased following the 1987 growing season as a result of above average precipitation. Less precipitation was received within burn 1-84 (40 dm) than within site 3-84 (49 dm) in 1988; burn 1-84 was the only site departing from the trend (Fig. 8). All HOI's declined following dry weather in 1989. Trends for sandsage HOI's within controls followed a similar pattern (Fig. 8). Similar HOI trends among controls strengthens confidence that departures among the treatments were caused by fire. Crown Cover and Composition Pretreatment sampling of crown cover was conducted in early spring 1984, prior to initiation of the spring 1984 burns. Vegetation that had grown the previous year was not easily distinguished from older residual and some forbs had deteriorated beyond recognition. This probably weakened pre- to posttreatment comparisons (the latter were conducted in mid to late summer) among 1984 burn analyses. Pretreatment samples for 1985 and 1986 burns were conducted in mid to late summer, and therefore, were more directly comparable with post-treatment data. Blue Grama (Bouteloua gracilis) -- Prescribed burning within site 1-84 provided evidence of enhancing blue grama from pre- to 1984-87 post-treatment intervals (f < 0.01, Tables 12 and 13). In contrast, no evidence of crown cover increase due to fire for the species occurred within the 3-84 burn (Fig. 9, Tables 14 and 15). Fewer transects and less blue grama/transect reduced sample sizes in the 3-84 burn. Some of the increase in blue grama may be attributed to fire removal of overs tory residual which made blue grama more visible. B~ue grama responded positively to fire following both the 1985 and 1986 burns (Fig. 9, Tables 16 through 19). Major recovery occurred during the second post-treatment growing season and extended through the 4th year within the 1985 burns (f < 0.05). Recovery occurred during the Ist>growing season within the 1986 burns, apparently because of increased precipitation that year f< 0.01). This recovery was not sustained during subsequent years. Fire �166 1.2 E 1.0 ~ 0.8 - 1985 ~ ./"' "'C en z w c 1986 f-!. C} W 0.4 ::c 0.2 ... ;"\ \ ,./ '-.,. ,,, \ \ I 0.6 -'.~.....,, ,, ,, , ..".. .... I ::c .. \ .. , o ~--~"-'--~--~--~--~--_.YR-1 YR·2 YR-3 YR-4 YR-5 YR-6 YR-7 YEAR 1.6 3-84 1.4 E -c - 1.2 ~ 1 z w c 0.8 ~ 0.6 en CJ m ::c , ~ . ... __ ..,., 1986 '...... .•.... /'....... PRE ,, , ,, <: ~/ - _, , '~ ,,, "" , 0.2 ,, , ,, .•.......... '\"\ .....:r / 0.4 0 ." ,,. 1985 ,., ~ ,, ,, 2 3 4 5 6 7 YEAR Fig. 7. burning, Recovery Tamarack of grasses and forbs (upper) and sand sagebrush Prairie, Colorado. (lower) to �167 1.2 E ~ ~ .~ ~ ._!. J: Cl w 1.1 1.0 0.9 0.8 0.7 0.6 0.5 0.4 J: 0.3 0.2 0.1 ------~-.---3-84 O~--~--~~--~--~~~ 84 85 86 87 88 89 90 YEAR GRASS/FORB 2.0 " 10, /j\ ".. .!/ \ .• - I· •• II ..-1 E 1.6 , \ • i/ ," .,:i ~ ,, .,h ~ , '1 UJ w 1.2 3·84 ~ J: o " 0.8 ., '.. '\ ~ .~\ . . ;.~~ 1986 . , fJ Z mJ: • ...•.. " .-.....,' " ..... . ' ""' .••..•. • ., ,. ,'l--- . , ,:... ........._198S 84 85 86 87 88 89 90 YEAR SAND SAGEBRUSH Fig. 8. Response of grasses and forbs (upper) and sand ~agebrush control plots, Tamarack Prairie, Colorado. (lower) in �168 Table 12. Mean crown cover (proportion/l-m2 x 100) for selected species, species groups, and covers within burn and control samples dtiring pretreatment (1984) and post-treatment (1984-89) intervals, burn 1-84, Tamarack Prairie, Colorado. Burn/ control Pre-tr 1984 Aug 1984 Jul 1985 Aug 1987 Jul 1989 Bouteloua gracilis B C 29.0 40.8 24.8 26.0 9.3 7.4 12.4 10.9 15.5 13.7 Stipa comata B C 20.6 15:0 6.9 14_7 8.3 10.1 10.2 12.1 17.1 19.0 Sporobolus cryptandrus B C 3.6 2.8 6.0 3.0 4.7 1.5 4.2 2.8 4.5 3.4 Calamovilfa longifo1ia B C 14.9 15.6 15.8 15.3 5.9 7.6 8.3 9.9 8.8 10.0 Andropogon hallii B C 2.3 1.7 3.8 2.3 1.3 0.8 2.6 2.0 3.4 2.4 Tall warm-season grass8 B C 17.1 17.4 19.7 17.7 7.2 8.5 10.9 12.0 12.3 12.5 Total warm-season grass B C 49.7 60.9 50.5 46.7 21.2 17.3 27.6 25.7 32.2 29.6 Artemisia filifolia B C 8.2 5.1 6.1 5.4 6.7 4.0 6.1 2.6 6.3 1.8 Perennial forbs B C 1.3 2.2 3.3 2.3 1.0 1.3 1.2 2.6 2.2 4.1 Annual forbs B C 0.2 0.1 0.3 4.0 1.3 .0.2 0.5 1.3 0.7 1.7 Bare ground B C 9.4 8.4 30.9 7.3 36.7 18.5 20.8 15.9 13.6 12.0 Dead vegetation B C 10.0 7.3 0.6 18.7 24.2 48.1 32.0 38.2 26.2 30.5 Species/category 8 Includes Ca1amovilfa, Andro:gogon, Panicum virgatum, and Pas:galum spp. �16Y Table 13. Crown cover relationships for selected species among years, burn 1-84, Tamarack Prairie, Colorado. and species groups Pre- to Post-treatment E. Value Species/catagory 1984-Aug Bouteloua gracilis Stipa comata Sporobolus cryptandrus Calamovilfa longifolia Andropogon hallii Tall warm-season grasses Total warm-season grasses Artemisia filifolia Combined perennial forbs Combined annual forbs Bare ground Dead vegetation Post-treatment 84 21. 64c 38.75c 3.61 0.32 1.46 1.06 45.54c l7.32c 8.06b 9.80c 132.70c 211. 01 c Bouteloua gracilis Stipa comata Sporobolus cryptandrus Calamovilfa longifolia Andropogon hallii Tall warm-season grasses Total warm-season grasses Artemisia filifolia Combined perennial forbs Combined annual forbs Bare ground Dead vegetation < 0.05. < 0.025. c f < 0.010. b P l 14.89c 19.26c 9.82c 5.728 1.54 3.16 10.18c 0.00 0.11 l7.22c l66.60c 87.05c 1984-87 1984-89 16.42c 22.36c 2.50 1. 32 0.02 0.64 8.48b 1. 30 2.01 4.04 10.75c 1l.16c 0.01 5.088 0.66 0.22 0.04 0.00 7.86b 3.09 1.12 2.64 0.82 5.498 to Post-treatment Aug 84-85 a 1984-85 4.24 3.37 l5.56c 7.2lb 0.02 5.558 2.58 l4.l0c 5.998 7.92b 0.82 1.54 Aug 84-87 2.72 5.178 0.16 2.56 1. 95 2.08 0.07 l3.00c 10.l9c 2.33 4.37 0.37 1985-87 0.03 0.05 1.09 0.43 0.74 0.94 0.48 3.75 6.36b 5.928 6.86b 1l.18c Aug 84-89 1.25 2.24 0.12 0.82 0.09 0.62 0.03 l6.86c 4.00 0.93 l7.23c 1. 69 �170 50 40 a:: 80 ,, , !!;! o o , 86 BURL_ / co " ,~ , 30 § 5 70 1-84 CONTROL ~ 50 z 20 Ixl § 40 Ixl 30 ...... 10 , 20 o L- PRE-84 ~ POST-84 1985 1987 10 ._. 84 1989 .•..•••• __ 85 YEAR 1984 Fig. 9. Trends ,,• , , ,,~ " , , ~_ ...... ...... ' , --"L.' 85 CONTROL "-- __ 86 .....•. _ 88 YEAR 1985-86 BURNS in cover of blue grama, Tamarack 25 Prairie, BURNS Colorado. 100 go , , , 20 ,' ,-", 1-84 CONTROL ,,~ ....• ,'.-- , 80 c: ~ 70 ;: co 8z 85 CONTROL 0 c: o / 50 Ixl 40 5 30 o ~------~------~----~------._ PRE·84 POST-84 1985 1987 1989 20 .__---~--------------~ 84 85 86 YEAR 1984 BURNS Fig. 10. Trends in cover of needle-and-thread, 88 gO YEAR 1985-86 Tamarack BURNS Prairie, Colorado. �171 Table 14. Mean crown cover (proportion/l-m2 x 100) for selected species, species groups, and covers within burn and control samples during pretreatment (1984) and post-treatment (1984-89) intervals, burn 3-84, Tamarack Prairie, Colorado. Species/category Burn/ control Bouteloua gracilis B C Stipa comata B C Sporobolus cryptandrus B C Calamovilfa longifolia B C Andropogon hallii B C Tall warm-season grass8 B C Total warm-season grass B C Artemisia filifolia B C Perennial forbs B C Annual forbs B C Bare ground B C Dead vegetation B C 8 Pre-tr 1984 Aug 1984 Jul 1985 Aug 1987 Jul 1989 13.3 10.5 7.0 7.5 2.8 2.1 5.4 4.9 7.6 5.7 10.6 13.6 l.9 7.0 4.7 4.1 6.7 1l.0 16.4 15.8 8.8 10.4 5.6 5.7 4.2 3.3 6.2 8.7 4.4 7.3 7.8 9.7 5.5 5.8 6.9 4.9 7.4 5.1 6.5 5.8 l.4 l.2 0.7 0.8 0.4 0.4 l.2 l.4 2.6 3.0 10.2 12.5 7.2 7.2 7.6 5.6 10.1 7.3 10.5 9.9 32.4 33.4 19.7 20.4 14.7 1l.0 2l. 6 2l.0 22.3 22.8 19.4 18.0 9.3 23.5 15.7 20.0 12.3 15.4 13.7 16.6 8.2 4.9 6.8 3.5 4.0 2.2 4.2 3.0 4.2 4.7 0.2 0.1 l.1 0.8 5.0 l.4 2.6 l.0 l.8 0.8 16.2 24.1 56.5 15.1 20.0 2l.9 28.8 2l.2 18.7 13.8 9.7 11.4 2.5 27.0 12.2 36.9 22.0 25.6 20.3 22.2 Includes Calamovilfa, Andropogon, Panicum virgatum, and Paspalurnspp. �172 Table 15. Crown cover relationships for selected species and species groups among years, burn 3-84, Tamarack Prairie, Colorado Pre- to Post-treatment I Value Species/catagory Bouteloua gracilis Stipa comata Sporobolus cryptandrus Calamovilfa longifolia Andropogon hallii Tall warm-season grasses Total warm-season grasses Artemisia filifolia . Combined perennial forbs Combined annual forbs Bare ground Dead vegetation 1984-Aug 84 2.29 6.78a 0.43 0.32 0.10 0.30 0.00 23.90c 2.60 0.68 280.20c 90.00c 1984-85 0.43 0.95 0.71 4.15 0.40 3.18 l.89 0.07 0.24 18.00c 94.49c 3l.81c 1984-87 l.39 2.42 9.00a 3.72 0.49 4.43 0.42 0.23 0.03 2.23 4.99 l.05 1984-89 0.01 6.22 l.65 0.52 0.36 0.44 0.02 0.18 0.39 3.31 10.70c 0.02 Post-treatment to Post-treatment Aug 84-85 Bouteloua gracilis Stipa comata Sporobolus cryptandrus Calamovilfa longifo1ia Andropogon hal1ii Tall warm-season grasses Total warm-season grasses Artemisia fi1ifo1ia Combined perennial forbs Combined annual forbs Bare ground Dead vegetation a P < 0.05. b E < 0.025. c f < 0.010. 3.23 l.53 0.30 3.55 0.07 l.95 l.81 12.10b 0.02 16.78c l.18 0.00 Aug 84-87 0.36 0.88 5.79 9.45b 0.32 7.898 0.50 14.04b 0.04 2.40 0.39 l.15 1985-87 3.16 7.088 10.42b 0.04 0.00 0.08 0.40 0.47 0.08 0.46 0.08 0.52 Aug 84-89 0.68 0.65 2.07 0.23 0.18 0.06 0.00 3.58 2.16 3.58 2.87 0.83 �173 Table 16. Crown cover (mean point-frame ta11ies/432-point transect) for selected species and species groups among pre- and post-treatment samples within combined 1985 burns and their controls, Tamarack Prairie, Colorado, 1984-90. Species/category Burn/control .1985 1986 1988 1990 Bouteloua gracilis B C 28.5 30.8 21.2 21.9 20.6 10.5 26.6 20.2 52.6 39.4 Stipa comata B C 31.4 31.3 27.7 55.2 48.2 67.2 36.2 43.1 88.8 86.4 Sporobolus cryptandrus B C 4.8 8.1 21.1 22.7 21.1 22.1 19.8 23.5 37.2 29.8 Calamovilfa longifolia B C 36.8 33.1 54.5 28.5 58.8 26.4 47.0 35.4 33.8 21.3 Andropogon hallii B C 1.7 2.1 5.1 4.1 9.6 4.6 11.2 9.1 10.1 8.8 Agropyron smithii B C 0.6 2.2 1.8 3.5 10.4 11.0 11.6 8.9 5.5 4.0 Selected warm-season B C 39.4 35.9 61.8 32.8 70.5 31.1 63.1 44.6 50.4 28.5 Total warm-season B C 72.7 74.8 104.0 77 .4 112.2 63.8 109.5 88.2 140.2 99.5 Total cool-season B C 32.1 33.5 29.5 58.7 58.6 78.2 47.9 52.0 94.3 90.4 Artemisia filifolia B C 22.5 24.7 8.2 22.6 11.9 23.8 18.0 34.6 18.5 29.1 Combined perennial forbs B C 4.4 2.9 3.4 2.1 5.7 4.2 17.5 12.7 15.6 10.6 Combined annual forbs B C 4.1 3.6 2.8 3.5 3.5 4.2 15.5 20.9 22.1 49.9 Bare ground B C 30.9 42.2 110.3 47.1 90.7 67.7 31.1 42.7 47.9 59.5 Dead vegetation B C 260.1 249.3 170.1 217.9 145.1 187.9 189.6 176.0 90.4 87.5 a Pretreatment. �174 Table 17. Crown cover relationships for selected species among years, 1985 burns, Tamarack Prairie, Colorado and species groups Pre- to Post-treatment .E Value Species/catagory Boute1oua gracilis Stipa comata Sporobo1us cryptandrus Calamovi1fa longifolia Andropogon hallii Tall warm-season grasses Total warm-season grasses Total cool-season grasses Artemisia filifolia Combined perennial forbs Combined annual forbs Bare ground Dead vegetation Post-treatment 1984-86 1984-88 1984-90 0.11 15.99c 0.46 17.11c 0.68 19.02c 18.19c 16.99c 30.01c 0.18 0.10 92.20c 32.35c 5.91b 4.708 0.64 20.77c 5.65 37.20c 41.44c 2.98 15.96c 0.19 0.34 30.10c 14.74c 4.278 0.06 1.74 0.00· 3.78 8.98c 0.14 11.35c 0.00 0.70 0.44 1.09 2.70 0.02 4.298 3.65 0.00 5.028 9.99c 0.11 3.23 0.06 3.52 0.20 0.08 1985-86 1985-88 1986-88 1988-90 8.00c 6.98b 0.07 6.02b 5.348 0.35 3.41 0.17 4.678 0.22 0.50 9.65c 5.098 0.19 1.08 13.62c 4.658 0.04 1.00 0.60 11.51c 2.98 2.14 0.00 1.41 4.178 1. 70 0.00 0.04 1. 98 4.00 0.91 2.62 0.39 2.89 0.00 0.25 1.12 to Post-treatment Bouteloua gracilis Stipa comata Sporobolus cryptandrus Ca1amovi1fa longifolia Andropogon hallii Agropyron smithii Tall warm-season grasses Total warm-season grasses Total cool-season grasses Artemisia fi1ifolia Combined perennial forbs Combined annual forbs Bare ground Dead vegetation 8 P < 0.05. b f < 0.025. c f < 0.010. 1984-85 1.69 4.938 6.96b 22.70c 9. nc 3.06 0.06 0.09 3.80 3.68 5.87b 0.06 3.60 0.06 0.00 0.68 23.2Sc 10. nc �175 Table 18. Crown cover (mean point-frame tallies/432-point transect for selected species and species groups among pre- and post-treatment samples within combined 1986 burns and their controls, Tamarack Prairie, Colorado, 1984-90. Species/category Burn/control 19848 19858 1986 1988 1990 Boute1oua gracilis B C 37.3 40.6 36.1 34.6 38.1 18.0 42.5 31.0 70.4 57.4 Stipa comata B C 42.5 28.1 40.0 52.1 38.5 61.5 38.4 48.7 65.5 88.1 Sporobolus cryptandrus B C 10.3 4.6 13.7 14.8 24.5 15.3 19.0 19.4 26.4 20.3 Calamovi1fa longifo1ia B C 26.9 39.5 25.3 33.8 44.9 37.2 45.9 54.4 38.2 47.2 Andropogon ha11ii B C 1.9 0.6 2.4 1.1 6.8 1.2 8.4 2.8 10.6 4.9 Agropyron smithii B C 4.3 1.8 9.5 5.3 10.1 4.2 3.7 1.5 Selected warm-season B C 28.9 40.1 27.6 34.9 51.7 38.3 54.4 57.6 49.7 52.1 Total warm-season B C 76.5 85.3 77.4 84.3 114.3 71.6 115.9 108.0 146.5 129.8 Total cool-season B C 46.1 29.1 44.2 53.7 47.9 66.8 48.5 52.9 69.2 89.2 Artemisia filifolia B C 18.1 14~4 16.2 19.3 7.9 22.6 14.8 29.7 16.1 24.3 Combined perennial forbs B C 7.2 5.1 4.7 3.0 15.4 6.3 25.9 11.7 6.1 7.5 Combined annual forbs B C 5.2 3.1 1.6 1.3 3.2 2.3 12.9 12.2 29.2 18.9 Bare ground B C 39.5 28.9 41.1 32.6 182.4 46.6 46.9 34.7 59.9 43.2 Dead vegetation B C 234..2 263.9 241.4 226.6 57.2 204.8 163.6 177 .0 91.0 110.3 8 Pretreatment. �176 Table 19. Crown cover relationships for selected species among years, 1986 burns, Tamarack Prairie, Colorado. and species groups E Value Pre- to Post Treatment 1985-86 1985-88 8.93c 2.59 1.14 3.01 0.28 Stipa comata 6.l4b 0.34 1.47 1.07 1.34 Sporobolus cryptandrus 8.11b 0.01 3.52 3.41 3.66 longifolia l7.00c 0.27 0.21 6.38b 0.08 7.67b 2.30 3.91 0.61 0.21 0.79 0.50 1. 74 0.48 0.00 Species/group Bouteloua gracilis Calamovilfa Andropogon hallii Agropyron smithii Tall warm-season grasses 1985-90 Post-treatment 1986-88 20.83c 1.04 0.90 8.44b 1988-90 Total warm-season grasses 28.62c 2.43 7.16c 11. 77c 1.22 Total cool-season grasses 2.33 2l.16c 0.05 1.34 2.18 3.27 6.46b 2.19 5.398 2.21 8 3.19 1. 79 0.40 0.09 0.09 0.14 0.17 1.77 Bare ground 0.15 l80.89c 1.04 2.76 13 .42c 1.33 Dead vegetation l37.04c 2.61 4.80b 3.32 2.31 Artemisia filifolia Combined perennial Combined annual a b c r. < r. < r. < forbs forbs 4.26 0.05. 0.025. 0.010. apparently increased blue grama crown cover within the 3 burns where it was abundant and increases were partially sustained through 1989 and 1990. Blue grama was one of the dominant species on the Tamarack Prairie with composition ranging from 10 to 30% among the 4 control groups (Table 20). Needle-and-thread (Stipa comata) -- Crown cover of needle-and-thread was reduced during the first growing season following fire within all sites (r. < 0.05, Fig. 10, Tables 12 - 19). This cool-season species had attained considerable growth (heading was noted in 1985) by the time burns were conducted (Table 1). First-year growth was severely repressed but extensive seed production occurred consistently during the following spring. Crown cover of needle-and-thread recovered slowly in subsequent years. The phenologically earlier burns in 1984 seemingly had a less severe impact on the species (Fig.lO), but enhancement due to fire was not evident. Precipitation and soil moisture during spring markedly influenced growth of this cool-season species. Composition within the 4 control groups averaged between 17 and 30% for this species (Table 20). It was far more abundant than other cool-season grasses, primarily western wheatgrass (Agropyron smithii, 0.1 - 3.1% composition) and sandhill muhly (Muhlenbergia pungens., 0 - 0.5% composition). �177 Table 20. Species composition (%) averaged over the interval of sampling years) within burned site control groups, Tamarack Prairie, Colorado. Catagory/species Grasses & Sedges Bouteloua gracilis Stipa comata Sporobolus cryptandrus Calamovilfa longifo1a Andropogon ha11ii Agropyron smithii Panicum virgatum Aristida sp. Paspa1um stramineum Muh1enbergia sp. Minor perennial grassesa Total perennial 3-84 1985 1986 30.67 25.43 4.74 20.19 3.23 0.35 10.11 17.12 12.02 10.76 2.28 0.01 0.39 12.86 28.79 10.67 15.94 2.79 3.10 0.04 0.11 0.09 0.06 0.01 16.51 30.20 8.46 20.91 1.10 1. 65 0.04 0.07 0.05 0.21 0.03 0.09 0.15 1.03 0.53 0.01 84.86 54.25 74.46 79.23 Annual grasses Cyperus & Carex spp. 0.37 0.50 0.01 2.13 0.16 0.31 0.09 0.75 Sage Artemisia 6.89 34.08 14.22 11.74 Cactus Opuntia grasses 1-84 filifolia sp. Perennial 0.50 1.23- 0.84 1.22 1.11 0.27 1.68 0.25 0.24 0.07 0.09 0.07 1.59 0.43 1.24 0.12 0.35 0.88 0.36 0.02 0.38 0.07 0.10 0.01 0.01 0.12 0.04 0.02 0.04 1.15 0.51 0.42 0.14 0.22 0.04 0.07 0.03 0.07 0.15 0.54 0.02 0.03 0.01 0.82 0.05 0.04 0.10 0.01 0.01 0.02 0.55 0.16 0.99 0.03 0.46 0.03 0.49 0.12 0.14 0.05 0.29 0.27 0.10 0.03 0.01 0.03 4.51 6.91 Forbs Ambrosia psilostachya Artemisia 1udoviciana Tradescantia occidenta1is Phlox andico1a Evolvu1us nutta1ianus Lathyrus po1yrnorphus Erigeron sp. Psora1ea tenuif10ra Physalis subg1abrata The1esperrna megapotimicum Sphaera1cea coccinea Mentzelia nuda Lygodesmia juncea Hap10pappus spinu10sis Penstemon angustifo1ius Allium testi1e Cichorium intybus Lesguerel1a 1udoviciana Ipomoea 1eptophy11a Minor perennial forbsb Total perennial forbs 0.01 0.01 0.01 0.03 0.02 3.07 3.44 (5 �178 Table 20 cont. Catagory/species 1-84 Annual Forbs Helianthus petiolaris Croton texensis Chenopodium album Euphorbia sp. Conyza canadensis Cirsium sp. Cryptantha sp. Eriogonum annum Amaranthus sp. Lactuca sp. Polygonum sp. Salsola kali Lepidium densiflorum Plantago purshia Other minor speciesc Total annual 0.01 0.25 1.16 0.18 0.34 8 Includes Koleria cristata, b Includes Asclepias, C Includes Froelichia, 1985 1986 0.03 0.28 0.64 0.07 0.74 0.02 0.02 0.16 0.02 0.16 0.05 1.27 0.07 3.55 0.01 0.12 0.07 0.01 0.01 0.11 0.17 0.01 0.08 0.28 1.21 0.20 1.22 0.09 0.02 0.07 0.10 0.17 0.05 3.23 0.16 0.01 0.05 2.37 1.66 6.94 0.09 0.12 0.01 0.01 forbs 3-84 Erogrostis Astragalus, 0.34 0.54 0.20 0.01 0.01 0.07 0.02 0.01 cilianensis, Cymopteris Guara, Argemone, 0.02 and Oryzopsis montanus, and Liatris spp. punctata. and other species. Sand Dropseed (Sporobolus cryptandrus) -- Fire impact on sand dropseed was not consistent among burns and it appeared least responsive among the grasses to fire (Tables 12 -.19). Some evidence of enhancement was noted within burn 184 but not within burn 3-84 (Fig. 11). No enhancement occurred within the 1985 burns but a temporary 1st-year increase was noted within the 1986 burns (f < 0.025, Fig. 11). First-year stimulation of seed production was evident. This species comprised between 5 and 12% of the total composition among controls for the 4 burns during the study interval. It ranked 4th among the dominant grasses. Prairie Sandreed (Calamovilfa longifolia) -- Comparisons from the 1984 burns are uncertain with little evidence of fire impact (Fig. 12). Crown cover was suppressed from pretreatment to 1985 (f < 0.05). Growth within 3-84 burned transects was enhanced from summer 1984 through 1987. Rainfall following the 1984 burns was not conducive to dramatic growth of this warm-season species. In contrast, phenologically later burns in 1985 and 1986, followed by increased rainfall, stimulated growth of prairie sandreed (Fig. 12, f < 0.025). Increased growth was sustained for 2 years following the 1985 burns. Second-year sampling, following the 1986 burns, was not conducted, but . increased growth occurred for at least 1 year. Fire impact was not evident by the 3rd year. Composition of prairie sandreed ranged between 11 and 21% within the burn site controls (Table 20). It ranked 3rd'among dominant grasses within the Tamarack Prairie. �12 40 10 - \ - ... \~84 ~ 8 ~ 8 - CONTROL I -. \ •... •... ···· -.~3-84 I / / 6 - ~ ..... / / ... ------~ Ixl .:fI' ~~ 2 - 1-84 CONTROL ~ •• ' " i ...... ,--- _. 20 ~ POST-84 1985 1987 .. 10 1989 ... '. ..../.~. 861:0NTROL / // ~------~----~------~----~-84 85 86 90 88 YEAR 1984 BURNS Trends . ... YEAR Fig. 11. ,\, ..•. o Ixl o~~----~------~----~------~ o PRE-84 ,----_ ..-," ,'... -----~ ,, 86 BURN , I BURN ....~ 5 85 CONTROL 30 I 1985-86 in cover of sand dropseed, Tamarack BURNS Prairie, Colorado. 20 60 16 / .•.. / .•...•.. 50 1-84 CONTROL " -, -, ~ " .~ ..~.~.':."_'~ ...: ---------- ~ 4 § 3-84 BURN ~ --- / ~ 86 BURN 40 (3 Ixl .... =r: . / I I ~~.<»> 30 3-84 CONTROL ..... .. - .. ," " ,', ~,, .•... 85 CONTROL o ~------~----~----~------~ PRE-84 POST-84 1985 1987 1989 20 ._...------~----__..------~----~-84 85 86 YEAR Trends in cover of prairie , 88 YEAR 1985-86 1984 BURNS Fig. 12. .... " .... II: sandreed, Tamarack BURNS Prairie, Colorado. , 90 �180 Sand Bluestem (Andropogon hallii) -- This tall, warm-season species, although widely distributed, comprised only I - 3% of the total composition within the Tamarack Prairie. Thus, sample sizes were small. A one-year growth increase was noted in burn 1-84, however, no increase was evident within burn 3-84 (Fig. 13). Data from the 1985 burns provided limited evidence that the species was enhanced by fire. Evidence was more positive within the 1986 burns (f < 0.025). Increased crown cover was sustained through 1990 within both the 1985 and 1986 burns (Fig. 13). Crown cover of sand bluestem declined from 1984 to 1985 following a dry growing season, but sustained increases within the controls were noted on all burns through the remainder of the study (Table 20). Selected Warm-season Grasses -- This group included prairie sandreed, sand bluestem. and minor amounts of switchgrass (Panicum virgatum) and sand paspalum (Paspalum stramineum). Trends were similar to prairie sandreed which dominated the group (Tables 12 - 19). Total Warm-Season Grasses -- This group included most of the major grasses within the Tamarack Prairie (needle-and-thread, western wheatgrass. and sandhill muhly were excluded) and comprised 37 - 59% of the total composition. Fire increased crown cover of these combined species within burn 1-84 and the increase was sustained through 1989 (f < 0.05. Fig. 14. Tables 12 and 13). In contrast. little change was noted within burn 3-84 although sample sizes were much smaller (Tables 14 and 15). Sustained crown cover increases occurred within the 1985 burns (f < 0.01. Fig. 14. Tables 16 and 17). and were also apparent within the 1986 burns (Tables 18 and 19). These data provide strong evidence that prescribed burns, especially in 1985 and 1986 benefitted combined warm-season grasses. One site. previously interseeded to 1-84. was sampled pre-treatment and growing season (1985). Analysis of switchgrass crown cover was greater. controls from pre- to post-treatment big bluestem and switchgrass within burn at the end of the 2nd post-treatment covariance revealed that bluestemwhere impacted by fire, than within intervals (F1,27 = 7.27. f < 0.05). Total Cool-season Grasses -- This group. which included western wheatgrass and sandhill muhly. was dominated by needle-and-thread and trends closely followed those for this species (Tables 16 - 19). Sandhill muhly primarily occurred within the more hilly. sandy conditions in burn 3-84. Observations indicated it was suppressed by fire. Western wheatgrass was often approaching dormancy. and therefore. was probably underrated in crown cover samples. It was least evident within burn 3-84 and most abundant within the 1985 burns and their controls (Table 20). Prickly Pear Cactus (Opuntia spp.) -- Most above-ground parts of prickly pear cactus were severely burned during the 1984 burns. It, like sand sagebrush. quickly resprouted. Based on crown cover samples. the species was suppressed and gradually recovered. Samples within the controls showed no marked trends. Fires in subsequent years were cooler and less destructive to prickly pear. Trends within burned sites were not clearly indicated. Crown cover gradually increased within the combined 1985 and 1986 controls, but not within the 1984 c9ntrols. �181 5 12 4 1.········ ,;, 10 .... 1-64 BURN ~ 8 3 ...... / / ./ 1 1 85CONTROL 1 1 ...~ --~,,/~ / ~ 52 .. Ixl ,~, - -.....••• .•••.. .~, ,, . ·.1-84 CONTRO~ '. ~ "":"-::_ ,,~. ..... ~.~ 3-84 CONTROL .~., , , /~ / ' 1... .: I·' ' .; / .' ,~ "'~ 86 BURN ..... 3-84 BURN . 2 ..•••.-./ PRE-64 POST-84 1985 1987 .;/~ 84 1984 BURNS , ~ " 90 160 , 140 cr w ffi 40 - Z 30 -.I. 8z 120 > o (J "~ a: \ ..~ 1~120 .~ "~ ~ 10 88 ~1-84CONTROL 60 o 86 Trends in cover of sand bluestem, Tamarack Prairie, Colorado. 80 ;:: 85 YEAR 1985-96 BURNS YEAR Fig. 13. 86 CONTROL o ~---~---'----~---~ 1989 '" '" '" ",""/ --- o'--'----...&.-----'----'-----'-- '" .•• S-84BURN o '-- __ PRE·84 .•..•.. \ :;:: 0 / 5100 \ , \ \ \ ..., ..•...-,-.~ ....•. ,, Ixl ... / " .... ~~·••. ·4 , '~ .•. / •• ~ --'- __ ---'~_~__ POST-84 85 YEAR 1984 BURNS / / 89 / ,~,. , , """"* , / /, 80 , , 3_.8_,~_C_O_N_T_R_'O,'_L_ 87 / ,'85 " """""- T CONTROL " 60 84 85 86 88 YEAR 1985-86 BURNS Fig. 14. Trends in cover of all warm-season grasses, Ta~arack Prairie, Colorado. 90 �182 Combined Perennial Forbs -- Trends within burns, in relation to those within controls, were variable. A 1st-year increase (f < 0.025) oc'curred in burn 184, but no similar trend for total forbs occurred within 3-84 (Fig. 15). Perennial forbs within the 1986 burns showed 1st-year enhancement (f < 0.05) which was sustained through 1990. In contrast, burns conducted in 1985 did not increase perennial forbs through 1986, but increases,were noted during 1988 and 1990 (Fig. 15). Sample sizes for individual species were generally too small to provide meaningful data. Combined perennial forbs comprised 3 7% of the composition among controls and included about 20 primary species (Table 20). There were no distinct pre- to post-treatment changes in numbers of sampled species. Several early species, including sweet pea (Lathyrus polymorphus), Psoralea spp., and Phlox andicola, had matured and nearly disappeared by mid to late summer sampling. Grasshoppers and other insects reduced the crown cover of some species. Western ragweed (Ambrosia psilostachya) and spiderwort (Tradescantia occidentalis) were the most common species (Table 20). Western ragweed, a warm-season species, showed evidence of fire enhancement within the 1985 and 1986 burns in 1988. This increase was only partially sustained in 1990. Fire impact on western ragweed was not evident within the 1984 burns. Annual Forbs -- Combined annual forbs were highly variable as to species and occurrence within both the treatments and controls among all burns. Pretreatment data for the 1984 burns were especially weak because much of the residual of annual forbs had deteriorated when sampled in early spring 1984. Fire suppressed annual forbs through the 1st growing season within burn 1-84 (Fig. 16). This trend was not evident among other burns. Annual forbs increased more within burn 3-84 than within its controls. This contrasts with trends on other burns. Annual forbs increased markedly within controls in the 1985 burns during 1988 and 1990 samples (Fig. 16). Most of the increase was attributed to 2 species; Russian thistle (Salsoli kali) and horseweed (Conyza canadensis). Similar trends were evident during 1988 and 1990 sampling within the adjacent 1986 burns; however, annual forbs in the treated sites exceeded those in the controls. When data from the 1985 and 1986 burns were combined, no evidence of fire impact on annual forbs was evident. Fire neither increased nor suppressed numbers of species. Mean composition within controls for combined annual forbs averaged < 2% (burn 3-84) to almost 7% (1985 controls, Table 20). Bare Ground and Dead Vegetation -- Removal of dead vegetation (f < 0.01) and sand sagebrush with fire dramatically increased the amount of bare ground within all burns (f < 0.01, Tables 12 - 19). More residual was removed during the earlier, more intense fires conducted in 1984 than in subsequent years. Recovery was nearly complete by late summer 1987 within the 1984 burns. Nearly full recovery occurred by late summer 1988 within sites burned in 1985 and 1986. �183 so 10 25 8 , ex: 20 § ex: ~CI o .: / 85 BURN ~ 15 ia Ixl 86 BURN ~./ 10 Ixl 2 PRE-84 POST-84 1985 YEAR 1987 - - __ - in cover -, ~', , 86 CONTROL ' --~ 85 CONTROL o ~---~---~--~---~ 1989 84 85 88 86 90 YEAR 1984 BURNS Trends - . #,-, '" Y,/' .. 5 O •......•.. ------'-----..___--_._--~ Fig. 15. ,.... •••• 5 54 1985-86 of perennial forbs, 5 50 4 40 ~ o o Tamarack BURNS Prairie, Colorado. ....,,' , , ,, 85CONTROL I 30 z :s: 86BURN...,~ , , ,, ..... ""· I ~ 20 Ixl 85 BURN.......' . -> _,. .~... _,. .';;' 10 .,./ ,'/ , o ~---_'_------'---~-~~~ PRE-84 POST-84 1985 YEAR 84 1987 Trends in cover 85 86 88 YEAR 1985-86 1984 BURNS Fig. 16. 86 CONTROL --- of annual forbs, Tamarack Prairie, BURNS Colorado. 90 �184 Sand Sagebrush (Artemisia filifolia) -- All fires reduced the crown cover of sand sagebrush (Fig. 17, Tables 12 - 19, f < 0.01). Nearly' all above-ground woody portions were destroyed by fire in 1984; most were destroyed during 1985 and 1986 burns. The 1985 and 1986 burns were cooler than those conducted in 1984 and were phenologically later in spring. Therefore, they potentially stressed sage plants more, since leafing and use of energy reserves were more advanced. Rapid first-year regrowth following burns in 1984 did not occur in 1985 and 1986. However, nearly all plants survived and had made considerable recovery by 1988. Crown cover sampling, when compared to HDI data, indicated more rapid recovery because the latter sampled the species when it was dormant. Prescribed spring burns severely impacted sandsage but did not increase mortality. Monitoring of wildfires that occurred on the Tamarack Prairie on 24 June 1985 (lightning) and 5 July 1986 (highway traffic) showed that mortality was not markedly reduced, but plants were severely stunted and recovery was extremely slow. Late August sampling indicated low survival but sampling during the subsequent spring showed 96% of the marked plants survived the 24 June 1985 burn. Late August 1986 inventory of sandsage survival within the 5 July burn showed only 14 of 50 plants (28%) had resprouted and 42 of 50 (84%) had resprouted in a location where the fire was less intense. However, as during the previous year, some plants were just beginning to sprout indicating slow recovery especially where burns were intense. Herbicide Treatment to Reduce Sand Sagebrush Aerial application of 2,4-D herbicide, applied on 6 June 1985 to alternate strips within an 80.8-ha site in the east-central part of the Tamarack Prairie, caused dramatic mortality of sandsage. Winds, which should have been nearly calm, were blowing from the south at approximately 8-13 kmjhr causing herbicide drift. The hilly terrain also prevented the spray plane from remaining close to the ground in all locations increasing drift. Precipitation was not received for several weeks following the treatment. As a consequence, nearly all of the sand sagebrush and broad-leafed vegetation within the site was killed rather than the partial strip reduction that was planned. Mortality of sandsage extended 100 m to the north of the tract and stunting of broad-leafed plants was noted for an additional 300 m. All sand sagebrush plants within established transects were killed and only 17 of 689 plants (2.5%) were alive within a walking traverse of the site in June 1986. Dead sage remained standing for several years deteriorating primarily in 1989. During early spring post-treatment HDI samples (time constraints prevented pretreatment sampling), sandsage was the primary obstruction 37.2% of the time in 1987, 24.1% in 1988, 22.7% in 1989 and 8.2% in 1990. Height-density trends within the sprayed site were similar to those within nearby site 2 and 3 controls for burns conducted in 1985 and 1986 (Tables 6 and 8). Grasses growing within dead sands~ge contributed to its HDI. Dramatic pre- to post-treatment changes in vegetation occurred within the herbicide-treated site based on crown cover sampling (Table 21). Blue grama, �25 1984 BURNS ", ..•••. ..•••. ", ", a: w > 0 ", 20 ". ",.~" o Z ~ 0 --- -. 15 ..•••. ......•.••. s-84 CONTROL ••••• ..... ..... s-84BURN ••••••• .............-........ a: w o ••••••••••_----- . . .. . . . .... ....... .-.. .... ..... CJ < a: . ",_ 10 ., w > < ~1.a4BURN -----_ .. ----_ --- 5 - 1.a4CONTROL --------III --- _ ------_ o PRE-84 POST-84 1985 1987 1989 YEAR 40 a: w > o 32 85CONTROL~ 0 Z ~ 28 0 a: o w 24 --- - - - _ _ 20 < a: ---."...""" w .""" 16 " ,,,,, ", CJ > < ,.. ,, .... .... ,,'''..... " ..... 1985-86 BURNS 36 "" . ". ". _ - - ."...""" ."...""" -. -,-, 12 ,'''' ~~ " ,. ..... •••••• ..... ~ Be CONTROL .... . ....•.•••.••... ........ .' ..' ....... .. •••••• .' .....:...~ Be BURN ' 8 84 85 86 88 90 YEAR Fig. 17. Trends in cover of sand sagebrush, Tamarack Prairie, Colorado. �186 needle-and-thread, and sand dropseed increased markedly; pra~r~e sandreed increased at a slower rate. Prickly pear (abundant in parts,of the treated area) increased steadily from 1985 through 1990. Perennial forbs were impacted both in quantity and number of species. Gayflower (Liatris sp.), a late-season species, was the only perennial forb observed on the site in summer 1985. Sweet pea was among the first to recover. Increased numbers and species were noted from 1986 through 1990, but full recovery of forbs had not occurred by 1990 (Table 21). Deficient precipitation in spring 1989 (Fig. 2) probably caused declines that year. Annual forbs were uncommon and showed trends similar to those for perennial f'orbs, Prescribed fire was applied to a 16-ha segment of the herbicide-treated site in spring 1986. Residual (grass) vegetation was reduced in early spring 1987 (HDI - 0.13 dm, f <0.05), but recovered under favorable precipitation in 1987 yielding increased HDI (1.13 dm) in early spring 1988 when compared with nearby controls (f < 0.001). Height-density indices in 1989 and 1990 gradually declined; apparently because of low amounts of precipitation. Renovation and Revegetation Treatments Renovation of Interseeded Tracts -- A site that had been interseeded in 198283 and partially burned in May 1984 was selected for renovation treatment (disked, treated with atrazine, and harrowed) in April 1986. Interseeded bluestems and switcbgrass were present within rows spaced 1.1 m apart. However, they were suppressed by between-row competition from native and annual vegetation. The objective of renovation was to stimulate growth of interseeded grasses. Sampling was conducted to document vegetation changes occurring following the 1986 renovation effort. Height-density of grass-forb residual grass-forb vegetation was suppressed during the first growing season following treatment (Fig. 18). HDI increased within the treated site during the 2nd growing season (1.55 dm); this increase was sustained through the end of the study (f < 0.05). Renovation increased the HDI quality of grass-forb vegetation. Sandsage (sparsely present) was severely suppressed by renovation. Height-density sampling of other interseeded sites that received the same renovation treatment in 1986 or 1987 confirmed that interseeded grasses were released from competition yielding markedly higher height-density indices. Pretreatment data were not obtained, but within 10 sites, the average HDI of standing residual was 2.71 dm in 1989 and 3.41 in 1990 (greater than within the initial test site). Observations indicated pretreatment HDI was similar to that within the previously renovated site « 1.0 dm). Crown cover sampling of selected species and species groups within the initial study site in late summer showed renovation increased growth of interseeded bluestems and switchgrass, even during the first season after treatment. Increases were sustained for several years (Table 22). Prairie sandreed, a deep-rooted indigenous grass, was neither enhanced nor decreased by treatment. Short- and midgrasses were suppressed and recovered slowly. Perennial forbs were suppressed but sustained limited recovery in subsequent years. �Hi / Table 21. Crown cover (mean point frame tallies/288-point transect) of vegetation and ground cover within 11 transects during pre- .(1985) and posttreatment (1986-90) spring intervals in the June 1985 herbicide-treated sandsage spray tract, Tamarack Prairie, Colorado. Vegetation/cover Pretreatment 1985 1986 Bare ground Dead vegetation Perennial grasses Boute1oua gracilis Stipa comata Sporobo1us cryptandrus Ca1amovi1fa longifo1ia Other perennial grasses Annual grasses Bromus tectorum Festuca sp. Artemisia filifo1ia (live) A. filifolia (dead) Opuntia & Echinocereus spp. Perennial forbs Ambrosia & Artemisia spp. Tradescantia occidental is Lathyrus po1ymorphus Psora1ea tenuiflora Evo1vu1us nutta1ianus Phlox andico1a Allium textile Mentze1ia nuda Leucocrinum montanum Penstemon angustifo1ius The1esperrna megapotimicum Cymopteris montanus Abronia fragrans Erigeron sp. Liatris sp. Total perennial forbs Annual forbs Croton texensis Chenopodium album Pepidium & Lesguere11a sp. Miscellaneous annual forbs Post-treatment 1987 1988 1989 1990 402 1,425 529 1,324 536 1,083 387 990 472 1,207 454 1,028 142 194 34 51 4 148 510 46 65 15 312 425 106 53 19 307 683 161 120 15 311 563 108 124 15 327 722 103 87 23 70 117 116 4 7 38 10 436 112 1 o o 4 259 287 210 201 5 124 44 50 62 83 75 97 5 1 9 77 7 6 29 22 123 17 121 12 45 107 18 114 55 6 4 2 1 6 1 1 1 6 1 1 1 6 2 8 1 1 12 8 1 1 19 1 2 2 1 3 2 1 1 11 1 223 96 154 156 76 3 24 2 1 4 1 1 4 1 9 4 1 6 6 8 3 1 178 3 �188 1.6 1.4 -E 1.2 "C ~ en Z w 0 t-!- :I: G W ~ 1 RENOVATION 0.8 0.6 :I: 0.4 0.2 0 86 87 ·89 88 90 91 YEAR Fig. 18. Response of all vegetation following grass) areas, Tamarack Prairie, Colorado. renovation of interseeded (with �189 Table 22. Average crown cover (proportion/1-m2 x 100) of selected species and species groups from pre- (1985) to post-treatment (1986-90) .intervals within tillage-herbicide renovated and control portions of a previously interseeded site, Tamarack Prairie, Colorado. Species/group 1985 1986 1988 1989 1990 36.3 13.8 6.4 1.2 39.8 11.2 6.8 l.6 47.1 13.8 9.9 2.6 48.3 12.4 12.5 4.9 22.1 10.2 19.8 3.4 24.6 8.1 20.4 1.7 33.9 8.5 27.6 1.2 32.2 9.0 35.0 2.3 1987 Tillage-Herbicide Andro~ogon/Panicum Ca1amovi1fa longifo1ia Boute1oua, Sti~a, etc. Perennial forb 13.7 11.3 17.6 1.5 24.2 12.4 5.4 0.6 Control Andro~ogon/Panicum Calamovi1fa longifo1ia Boute1oua, Sti~a, etc. Perennial forbs 16.4 10.7 17.1 2.2 16.6 10.6 16.8 1.7 E Values Andro~ogonl Panicum Calamovilfa longifo1ia Boute1oua, Perennial forbs Sti~a, etc. Pre-to Post-treatment 1985-86 1985-87 1985-88 1985-89 1985-90 34.688 61. 948 51. 938 28.118 42.198 1.04 2.76 2.03 3.23 1.44 62.118 66.708 38.838 30.968 34.828 1. 76 2.92 0.03 3.62 3.54 4.08 1. 78 2.87 0.34 0.04 l. 56 Post-treatment 1986-87 1986-88 1989-90 8 .r. < 0.05. l.42 �190 Revegetated Strips -- Nineteen strips were double disked (to destroy existing vegetation), treated with atrazine, harrowed, and seeded to tall, warm-season grasses in 1985. Stand establishment, primarily switchgrass, ranged from sparse to moderate. Height-density sampling of residual within 12 of 19 strips was not conducted in spring 1986, but HOI would have approximated 0.5 dm. Favorable precipitation and herbicide suppression of annual forbs prompted rapid growth of seeded grasses. The HOI in spring 1987 (after the 2nd growing season) was 1.83 dm and standing residual peaked in 1988 and 1989 at over 4 dm (Fig. 19). Comparisons with grass-forb HOI within nearby controls (for 1985 and 1986 burns) show marked contrast from 1987 through 1991. Late summer point-frame sampling documented gradual decreases in bare ground and dramatic increases in tall, warm-season grasses (Table 23). Prairie sandreed was not destroyed by treatment. Seeded species seldom established where it occurred. Few other grasses survived the treatment. If they did, they were suppressed by seeded tallgrasses. Little pronounced change occurred among other vegetation. Revegetation in subsequent years often yielded dense stands of switchgrass that did not obtain the tall, vigorous growth occurring in sparse to moderate stands.' However, height-density values remained greater than those in adjacent native range. Sampling in late summer 1986 of indigenous pra1r1e sandreed and sand bluestem within revegetated strips showed tillage released them from competition and stimulated growth. Extensive production of seed heads was noted and mean height for both averaged about 1 m, over twice that on untreated controls. �1'::11 5 -- 4 E "C ~ en Z w 3 SWITCHGRASS ~ C I I-- ::I: C!J 2 w ::I: 1 o ~------~----~----~----~----~86 87 88 89 90 91 YEAR Fig. 19. Response of all vegetation to revegetation, Tamarack Prairie, Colorado. �192 Table 23. Crown cover (mean point tallies/144-point transect) among vegetation groups from 1985 through 1990 within 12 1985 revegetation transects, Tamarack Prairie, Colorado. 1985 1986 1987 1988 1990 Bare ground 805 812 278 120 107 Dead vegetation/litter 212 156 372 620 394 Dominant 217 226 183 98 138 79 37 21 13 14 311 475 833 829 850 6 10 8 15 14 22 12 2 20 14 76 0 31 13 197 Vegetation/category native grasses Lesser grasses Seeded tall grasses Sandsage Perennial Annual and sedges and cactii forbs forbs Comparison of HDI's Among Grazed, Ungrazed, and Treated The mean HDI for residual grass-forb vegetation in transects in grazed pastures, south of the Tamarack = 0.028, n - 719). Sandsage obstructed 361 (33.4%) a mean HDI of 1.030 dm (SE = 0.063). The mean for (SE - 0.053). Sites early spring 1989 among 27 Prairie, was 0.160 dm (SE of 1,080 samples yielding combined data was 0.451 dm Within grazed sites, residual grass-forb vegetation was much lower than within either burned sites or ungrazed controls (Fig. 20). There was less contrast among sand sagebrush samples, however, that within grazed pastures remained lower. The dramatically greater HDI's for renovated and revegetated grasses is evident. Wildlife Use of the Tamarack Prairie Greater Prairie-chickens -- Releases of greater prairie-chickens (transplanted from Yuma County) were conducted on the Tamarack Prairie in 1984 (36), 1985 (40), and 1990 (23). Only 1 male was found near the Tamarack Prairie during a pretransplant search for leks. Five leks with 1 to 7 males were found in grazed rangelands adjacent to the Tamarack Prairie in spring 1985. Numbers of leks and males have slowly, but steadily, increased since then, primarily south of the Tamarack Prairie (Fig. 21). An active lek, containing several males, established about 150 m south of the Tamarack Prairie in 1987, was trapped during spring 1989, but was abandoned by all males but 1 for unknown reasons in spring 1990 and was not used in 1991. Two males were subsequently observed using a mowed lek about 1 km north within the Tamarack Prairie. The �4.5 GRASS 4 "E -c '-' > Ien Z W 3.5 3 III SAGE TOTAL ~ 2.5 C I I- J: G 2 J: 1.5 W 1 0.5 o PRIV BURNS CONT RENOV REVEG VEGET A TION TYPE Fig. 20. Vegetation response by treatment (and control), Tamarack Prairie, Colorado, 1984-90. �194 R49W-R48W ~ .skW V -"'" 13 ~* rrAMAl ~CKP J{AlRIE ~ 19 ~ ~ ~ 13 R48W-R47W ~ M* '--- 19 23 *9 ~25 ~ Z ~ 35 -. :c 0 E-< 31 ~ 25 27 29 , ~ 33 ~ 35 31 .. 2 z 0\ E-< 5 1 1 3 *16 11 7 8 9 * 14 ~ 0 ...J 11 ROAD 46 13 13 LEGEND *LEK# ~~~I~~~3·1KM o 1 SCALE Fig. 21. Colorado. Location ~ ~ of greater prairie-chicken 2rni. leks, Tamarack Prairie, 0 u l't ~ 812 Ul CI) �mowed lek was not used in 1991 but a new lek with several males was found 200300 m away, within a site previously burned by wildfire. At. least 2 hens, among 1990 transplants, nested and reared broods within the Tamarack Prairie. Nesting efforts there, by wild radio-marked hens, were unsuccessful. Monitoring of radio-marked greater prairie-chickens, trapped on leks south of the Tamarack Prairie in 1989 and 1990, indicated that hens were not strongly attracted to the ungrazed Tamarack Prairie for nesting possibly because of distances involved. Grass-forb HDI's were significantly greater there than in grazed rangelands to the south (Fig. 20), however, those for sandsage were not. All but 1 of 15 nests found during 1989-90 were associated with sandsage and most were within 1 km of established leks. Observations of greater prairie-chickens have gradually increased within the Tamarack Prairie; most occurring in the southeast part of the property (Fig. 21). Little use has been noted within the more narrow, western part of the property based on both monitoring and observations. Use of ungrazed rangelands between 1-76 and the South Platte River bottom to the north (Fig. 21) was not noted during the study. Only an occasional sighting in that area, primarily during winter, has been reported. Limited monitoring of radio-marked grouse, primarily in 1990, showed most shifted south from the Tamarack Prairie toward grain fields in fall. Birds present at lek 1 during spring 1990 were observed flying out of sight to the south, apparently to grain fields. Limited monitoring and observations within the Tamarack Prairie also suggested that hens were concentrating along mowed trails and other disturbed sites to feed. More detailed accounts of radiomarked greater prairie-chicken activities and use of the Tamarack Prairie were previously reported (Snyder 1990Q, 1991). Plains Sharp-tailed Grouse -- This species has been observed sporadically during winter months along the Tamarack Prairie or along the nearby South Platte River in recent years. Three males were observed displaying on greater prairie-chicken leks in spring 1989. Several males and 2 hens were associated with several leks in 1990. One hen, radiomarked and monitored, nested successfully and reared a brood. Surveys in 1991 showed continued increases and several males were found on a newly established lek within the Tamarack Prairie. In addition, the Tamarack up to 4 hybrids of the 2 species were observed Prairie in 1990 and 1991. on a lek south of Other Wildlife -- Horned larks (Eremophila alpestris), lark buntings (Calamospiza rnelanocorys), western meadowlarks (Sturnella neglecta), mourning doves (Zenaida rnacroura), and grasshopper sparrows (Arnmodramus savannarum) were common summer avifauna. Cassin's sparrow (Airnophila cassinii) were present but not abundant. One or more pairs of upland plovers (Bartramia longicauda) were inconsistent occupants, primarily during years when burns were conducted. Ring-necked pheasants (Phasianus colchicus) were present in low numbers «12) and mallards (Anas platyrhynchos) occasionally nested on the site. Swainson's hawks (Buteo swainsoni), red-tailed hawks, (~. jamaicensis), kestrels (Falco sparverius), and great-horned owls (Bubo virginianus) were among raptors observed on the property but nest locations were not available for these species. Only 2 jack rabbits (Lepus spp.) were .observed and there was little evidence of desert cottontails (Sylvilagus audubonii). Pocket gophers, (Geornys sp.) thirteen-lined ground squirrels (Sperrnophilus tridecernlineatus), and kangaroo rats (Dipodomys sp.) were common rodents, and �196 undoubtedly the primary food for resident coyotes (Canis latrans). Only occasional evidence of badgers (Taxidea taxus) and striped skunks (Mephitis) was noted. Mule deer (Odocoileus hemionus) and white-tailed deer (Q. virginianus) were common at all times of year, whereas pronghorn (Antilocapra americana) used the site inconsistently. A few other wildlife species occasionally passed through the property. DISCUSSION This study examined several aspects of range manipulation and management with enhancement of habitats for greater prairie-chickens as the underlying theme. Kirsch (1974) stated that quality of nesting and brood-rearing habitats appears to be the universal limiting factor for prairie-chickens through their range. Management and evaluations were based on that assUmption. The primary questions asked at study initiation were: will prescribed burning increased the quality of standing residual grass to enhance nesting cover for prairiechickens? Will burning increase the quantity of tall, warm-season grasses, which usually possess better ability to remain standing over winter (Duebbert et al. 1981)? Will burning thin dense stands of sand sagebrush which are present only within this western edge of the greater prairie-chicken's range? Little information concerning the impact of fire on sand sagebrush was found within literature reviewed prior to this study. Grasses Prescribed burns have been used successfully to increase the composition of tall, warm-season grasses primarily within tallgrass prairies to the east where cool-season grasses commonly invade decadent stands of warm-season species, or in locations with lower precipitation/evaporation ratios than those in eastern Colorado (Kucera and Ehrenreich 1962, Anderson et al. 1970, Wright and Bailey 1980). In areas of higher precipitation, mulch increases rapidly on ungrazed range and suppresses growth of grasses so that periodic renovation is needed. Prescribed burns have been used successfully to restore grassland for prairie grouse and several species of waterfowl (Westemeier 1972, Kirsch et al. 1978) and to manage habitats for northern bobwhite (Colinus virginianus) and cottontails (McWorter and Lange 1973, Queal 1973, George et al. 1978). This study was initiated to evaluate the effect of prescribed burn management within a semiarid site considered marginal for prairie grouse. We did not know if prescribed burns could be used effectively to enhance vegetation quality within the Tamarack Prairie. A consultant (L. Kirsch) was contracted in 1983 to assess the potential. He recommended prescribed spring burning as the primary management tool (memo from L. Kirsch to R. Hopper and W. Graul, Sep. 1983). Wright and Bailey (1980) stated that evaluations of prescribed burns had not been conducted within the Central High Plains including eastern Colorado. Previous evaluations of burning were conducted on wildfires which generally occurred during dry conditions and usually were harmful. Continued review of the literature indicates little additional pertinent information has been published. A study in southcentra1 Nebraska within loam soils (Schacht and Stubbendiech 1985) is partially applicable. Other data are available from the Nebraska Sandhills (Wolfe 1972, Bragg 1978, Morrison et ai. 1986), but only Bragg's data involved prescribed burns. Soils within Nebraska sandhill studies contained higher percentages of sand than those within the Tamarack �Prairie. Interpretation of findings conducted within semiarid regions from Texas (Wright 1974h) through the Dakota's (White and Currie ~983, Engle and Bultsma 1984, Whisenant and Uresk 1989) all suggest using managed fire with caution. Findings of this study within a mixed-prairie sandhill setting of northeastern Colorado support that contention. Wright and Bailey (1980) reported that prescribed burning should not be used on shortgrass prairie except in unusual situations. Precipitation within the Tamarack Prairie is the same as that within nearby shortgrass ranges. Soil types (loamy sands vs. loams) are the primary differences. About one-fourth to one-third of the grass composition within the Tamarack Prairie is shortgrass. Grass-forb Vegetation Findings obtained on the Tamarack Prairie show prescribed burns reduced the height-density of residual grass-forb vegetation for 1 or more years on all sites. Part of that reduction resulted from removal of standing residual and was expected. Actual changes in phytomass were not sampled although crown cover sampling provided insight (Robel et al. 1970). Increased production of seed heads was noted for warm-season grasses during the first year and for needle-and-thread grass during the second growing season. If below average precipitation was received following burns, as in 1984, recovery was slow and fire did not increase the height-density quality of standing residual (Fig. 4). If above average precipitation occurred prior to and following the burn as in 1985 and 1986, grass-forb vegetation growth was more positive and height-density quality, exceeding that on the controls, was attained by the end of the 2nd post-treatment growing season. Enhancement due to fire lasted only 1 to 3 years (Fig. 4). Results indicate burns should be conducted during years of above average rainfall. This agrees with findings and recommendations of Wright and Bailey (1980), White and Currie (1983), Engle and Bultsma (1984), and Whisenant and Uresk (1989) for the semiarid High Plains. We can select years for burning when soil moisture conditions in early spring are above average. However, we can not predict post-treatment precipitation. Crown cover sampling provided insight into vegetation changes but left many unanswered questions. The 1984 burns, which were followed by below average precipitation, did not markedly enhance warm-season grasses. In contrast, burns during 1985 and 1986 stimulated increased growth of blue grama, prairie sandreed, and sand bluestem, and this growth began during the first post-burn growing season based on crown cover data. Prairie sandreed was probably the primary species contributing to increased height-density following the 1985-86 burns. Sand dropseed appeared least affected. Blue grama and sand dropseed were neither harmed nor benefitted by fire in southern mixed prairies (Wright 1974E). Needle-and-thread, which had attained considerable growth at the time it was burned, was negatively impacted during the first year, but consistently responded with dramatic seed production during the 2nd post-treatment growing season although crown cover did not increase. Kirsch and Kruse (1972) listed this species as an increaser under burning management. Time of burning undoubtedly is an important factor. Sample sizes for other grasses including western wheatgrass, sandhill muhly, sand paspalum, and June grass (Koleria cristata) were too small to detect changes. Observations indicated western wheatgrass and sandhill muhly, both cool-season species, were retarded by prescribed burns. These single-treatment burns did not change composition toward switchgrass, Indiangrass, and bluestems to any detectable extent. Sand bluestem was the only species present in more than trace amounts (except within interseeded or �198 revegetated sites) and it showed evidence of modest increase. It was increasing within the ungrazed control sites as well, especially during years of above-average precipitation. Its height-density cover value, when occurring naturally, was generally low within the Tamarack Prairie. Limited data indicate tall, warm-season species used in interseeding and revegetation efforts responded more favorably than indigenous species to prescribed burns. Thus, if these species were more prevalent within the Tamarack Prairie, prescribed burning might more effectively increase height-density of standing residual and the quality of the site for prairie grouse. Further evaluation is needed. Schacht and Stubbendieck (1985) noted that a single burn did not revert warmseason shortgrass to tallgrass prairie in southcentral Nebraska where annual precipitation averages about 30% greater than on the Tamarack Prairie. Fire increased shortgrasses. The same was evident on the Tamarack Prairie. Kirsch et al. (1978) noted that a period of restoration management may be needed before maintenance management can be practiced. Thus, revegetation, at least in small scattered tracts, may be needed to reduce blue grama and other native grasses in preference for taller species before applying prescribed burns to increase height-density within the Tamarack Prairie. Sand Sagebrush All prescribed burns severely impacted sand sagebrush but did not kill it. Height-density sampling, which occurred when the species was dormant, showed slower recovery than crown cover sampling, which occurred when the species was fully leafed. The 1984 burns were phenologically earlier than those in subsequent years, vegetation was drier, and nearly all above-ground twigs and stems were burned. Burns in 1985 and 1986 were conducted under more humid conditions, but after considerable leafing of sand sagebrush had already occurred. Sand sagebrush rapidly produced new growth after the 1984 burns. Within burn 3-84, where high densities of sage existed, it appeared to outcompete and suppress adjacent grasses following fire. Emergence of sand sagebrush was much slower following the 1985-86 burns, apparently because fire had depleted the energy reserves of sandsage. However, survival of the species approached 100% on all sites. Monitoring within sites burned by late June and early July wildfires revealed that emergence of new growth was often extremely slow, but nearly all plants survived. An initial study objective was to reduce sand sagebrush canopy cover to below 30%. It approximated such densities only within burn 3-84 and adjacent, more sandy and hilly rangelands (Fig. 1). Fire markedly reduced both the visual obstruction (height-density) and crown cover of the species and complete recovery had not been achieved by the end of the study. Percent visual obstruction of HDI's for sand sagebrush was approximately twice the percent composition obtained during crown cover sampling. This illustrates the importance sand sagebrush plays in providing concealment to pra1r1e grouse and other wildlife because of its taller lodge-resistent growth form. Aerial spraying, designed to remove up to 50% of the sand sagebrush within a . test area resulted in a 97% kill. Based on accumulated knowledge obtained during this study and previous studies (G. C. Miller pers. commun., Schroeder and Braun 1992), sage is used extensively by greater prairie-chickens for nesting and loafing protection in northeastern Colorado. It is a valuable component of their habitat. Similar findings have been ~eported where sandsage occurs within the range of lesser prairie-chickens (~ pallidicintus) �lYY (Jones 1963) and positive Cannon and Knopf (1981). Prickly correlations were found with percent sagebrush by Pear Cactus Wright (1974Q) reported that fire did not directly kill prickly pear cactus but that it injured the species making it vulnerable to attack by insects resulting in indirect kill. Prickly pear was severely burned during 1984 fires and less severely impacted during subsequent burns. Based on crown cover samples, it was suppressed by fire but gradually recovered. Spring burns did not markedly reduce the species in northeastern Colorado. It increased where competing sand sagebrush was killed by herbicides. It also increased steadily (140% from 1984 through 1990) within the 1985-86 control transects but did not increase within controls monitored for the 1984 burns. Perennial and Annual Forbs Native perennial forbs and grasses reacted similarly to fire. Cool-season species, such as wild onion (Allium textile), may be retarded by mid- to latespring burns whereas warm-season species, such as western ragweed, may be enhanced (Anderson et al. 1970, Kirsch and Kruse 1972, Mueggler 1976, Wright and Bailey 1980). Bragg (1978) found that sunflowers (Helianthus spp.) and Mentzelia nuda were adversely affected by spring fires. Early spring burns are preferable, however, wildfires under dry conditions can be injurious (Launchbaugh 1972, Wright 1974Q) and may favor annuals (Mueggler 1976). Bidwell et al. (1990) found that forbs increased when backfires, rather than headfires were used in spring and reported that backfiring small areas in late spring could be used to increase wildlife habitat. Impacts of fires on perennial forbs during this study were not clearly evident. Sample sizes were small and cool-season forbs had often matured and dried or been consumed by grasshoppers by the time crown cover samples were taken. Western ragweed benefitted from the 1985-86 burns but increases were temporary. Trends concerning fire impacts on annuals were even less apparent than for perennials. Jones (1963) noted the importance of subclimax vegetation containing forbs and higher densities of insects as brood habitat for both greater and lesser prairie-chickens. Kirsch et al. (1973) reported that prairie-chickens, like some other upland game birds, have a marked affinity for subclimax vegetation. Nesting and brood rearing habitats were considered limiting for prairiechickens throughout their range (Kirsch 1974). It was not a study objective to increase perennial and annual forbs for prairie grouse within the Tamarack Prairie, however, hindsight indicates it probably should have been. Forbs represented only a small proportion of the total composition (Table 20). Late spring head fires were not the proper management tool for enhancing most forbs. Other techniques, primarily disturbance tillage and seeding, would be needed to increase their abundance. Some species including leadplant (Amorpha canescens) still occur as remnants within nearby roadsides, but past grazing has depleted them within the Tamarack Prairie and adjacent rangelands. Such species could potentially be restored by seeding. Renovation and Revegetation Treatments Both treatments increased height-densities of standing residual. Conversions to tall, warm-season grasses, known to respond to fire in higher rainfall climates, were also achieved. These decreasers are more lodge resistent and �200 provide better residual cover for early spring nesting by pra~r~e grouse (Duebbert et al. 1981). Removal of most shallow-rooted perennial vegetation seems essential in such efforts. Direct seeding of grasses into existing stands in spring 1984 by management personnel yielded no detectable composition change. Competition of existing grasses was too great to allow new seedlings to establish and survive. Even where stands were established within interseeded furrows, adjacent vegetation quickly invaded and suppressed growth. Single treatments of atrazine herbicide were effective in suppressing annual weeds and allowing seedlings or established plants to attain vigorous growth. A 2nd application during early spring of the following year continued weed suppression and aided grass establishment. As with burning, best results occurred during years of above average rainfall. Costs/unit were greater than for burning but much greater results were obtained quickly and sustained for several more years. Based on preliminary data, future prescribed burns will be more effective, when applied to these sites, than to native grass or sagedominated sites. Findings showed tillage or sites containing indigenous effectively stimulate major renovate revegetated stands densities occur. tillage-herbicide treatments can also'be applied to stands of prairie sandreed and sand bluestem to growth. Such treatments may also be needed to of switchgrass, etc. especially when high Monitoring failed to show that interseeded tracts were used by prairie grouse within the Tamarack Prairie. Possibly, height-density qualities were too great. However, data within the literature (Miller 1963, Westemeier 1972, Kirsch 1974) tend to discount this suggestion. Monitored hens continued to use sand sagebrush for nesting and reared broods in either grazed or ungrazed native range. Prairie Grouse Occurrence on the Tamarack Prairie Prairie grouse, monitored through fall 1990 shifted toward grain fields south of the Tamarack Prairie. There they spent much of their loafing time in rangeland which was the greatest distance from windmills and least intensively grazed. It is hypothesized that if food sources had been associated with the tillage-herbicide renovated and revegetated tracts, they may have been used by grouse through fall and winter. Reasons that the Tamarack Prairie has not received greater use by pra~r~e grouse remain uncertain. Available data suggest potential benefits of providing both tillage disturbance and supplemental food sources within the Tamarack Prairie. Greater prairie-chickens visit established leks during almost all months of the year (G. C. Miller, pers. commun.). It is possible that if food sources are not available within flying distance of potential lek' sites, establishment of new leks and range expansion is less likely to occur. If leks are not present, use by hens for nesting and brood rearing is also less likely to occur. This is one possible explanation for the limited use of the Tamarack Prairie to date. Attempts to attract males to establish new leks by mowing knolls, resulted in use of one location in 1990. Truck traffic along Interstate 76 was suspected a~ another possible factor in limited grouse use of the Tamarack Prairie (Fig. 21), Traffic, while not heavy, was fairly constant. It was difficult to listen for displaying prairie grouse within 2 km of the highway due to �201 persistent truck traffic which could be heard for up to 3.2 km. Possibly, prairie grouse found this near-constant noise too competitive with their own auditory displays. Grouse use of Grazed Rangeland Monitoring of radio-marked prairie-chickens, which were trapped on private or lands managed by the Colorado State Land Board showed they usually remained there. Height-density samples taken at nest locations and at random locations within 0.8 km. failed to detect any increased vegetation quality at nest sites. Monitoring of hens showed most use occurred in lands that were not intensively grazed and contained considerable residual grass as well as sand sagebrush. Observations of rangelands in western Nebraska, where warm-season grasses are intensively grazed and sand sagebrush is lacking, show greater prairie-chickens are fairly abundant there. Height-density there was considerably lower than within sand sagebrush rangelands on and near the Tamarack Prairie. Thus, factors other than vegetation height-density appear to at least partially determine greater prairie-chicken presence and density. Annual grazing and long-term (~ 10 year) idled habitats are undesirable for prairie grouse (Kirsch 1974). Much of the Tamarack Prairie has been idled for > 10 years whereas most adjacent rangelands are grazed annually. Management for prairie grouse should be directed toward developing and maintaining vigorous subclimax grass-forb communities (Miller 1963, Kirsch et al. 1973). Thus, the future direction of management for prairie grouse on the Tamarack Prairie seems apparent. Prescribed burns on a 7 - 10 year rotation accompanied by revegetation of small tracts to native tall, warm-season grasses should be used in preference to grazing. Provision of legumes and seed producing annuals' should be evaluated. �202 Literature Cited Amen, A. E., D.L. Anderson, T. J. Hughes, and T. J. Weber. 1977. Soil Survey of Logan County, Colorado. U.S. Dep. Agric., Soil Conserv. Serv., Washington. D.C. 252pp. Anderson, K. L., E. F. Smith, and C. E. Owensby. 1970. Burning bluestem range. J. Range Manage. 23:81-92. Bidwell, T. G., D. M. Engle, and P. L. Claypool. 1990. Effects of spring headfires and backfires on ta11grass prairie. J. Range Manage. 43:209-212. Bragg, T. B. 1978. Effects of burning, cattle grazing, and topography on vegetation of a choppy sands range site in the Nebraska sandhi1ls pra~r~e. Proc. Int. Rangeland Congr. 1:248-253. Cannon, R. W., and F. L. Knopf. 1981. Lesser prairie chicken densities on shinnery oak and sand sagebrush rangelands in Oklahoma. J. Wildl. Manage. 45:521-524. Duebbert, H. F., E. T. Jacobson, K. F. Higgins, and E. B. Podoll. 1981. Establishment of seeded grasslands for wildlife habitat in pra~r~e pothole region. U.S. Dep. Inter., Fish and Wi1dl. Servo Spec. Sci. Rep. - Wi1dl. 234. 21pp. Engle, D. A., and P. M. Bu1tsma. 1984. Burning of northern mixed prairie during drought. J. Range Manage. 37:398-401. Floyd, D. A., and J. E. Anderson. 1983. A new point intercept frame for estimating cover of vegetation. Pages 107-113 in Idaho Natl. Eng. Lab. Radioecology and Ecology Programs. U.S. Dep. Energy DOE/ID 12098. George, R. R., A. L. Farris, C. C. Schwartz, D. D. Humburg, and J. M. Kienzler. 1978. Effects of controlled burning on selected upland habitats in southern Iowa. Iowa Conserv. Cornm.,Wildl. Res. Bull. 25. 38pp. Jones, R. E. 1963. Identification and analysis of lesser and greater prairie chicken habitat. J. Wildl. Manage. 27:757-778. Kirsch, L. M. 1974. Habitat management considerations for prairie chickens. Wi1d1. Soc. Bull. 2:124-129. ____________ , and A. D. Kruse. 1972. Prairie fires and wildfire. Proc. Tall Timbers Fire Eco1. Conf. 12:289-303. __________ , H. F. Duebbert, and A. D. Kruse. 1978. Grazing and haying effects of habitats of upland nesting birds. Trans. North Am. Wildl. and Nat. Resour. Conf. 43:486-497. __________ , A. T. Klett, and H. W. Miller. 1973. Land use and prairie grouse population relationships in North Dakota. J. Wildl. Manage. 37:449453. Kucera, C. L., and J. H. Ehrenreich. 1962. Some effects of annual burning on central Missouri prairie. Ecology 43:334-336. Launchbaugh, J. L. 1972. Effect of fire on shortgrass and mixed prairie species. Proc. Tall Timbers Fire Ecol. Conf. 12:129-151. McWhorter, R. E., and C. Lange. 1973. Ecology of bobwhite quail management. Unpubl. rep., Annu. Meeting Central Mtns. and Plains Sect., The Wildl. Soc. 8pp. Miller, H. A. 1963. Use of fire in wildlife management. Proc. Tall Timbers Fire Eco1. Conf. 2:19-30. Morrison, L. C., J. D. DuBois, and L. A. Kapustka. 1986. The vegetational response of a Nebraska sandhi1ls grassland to a naturally occurring fall burn. Prairie Nat. 18:179-184. �203 Mueggler, W. F. 1976. Ecological role of fire in western woodland and range ecosystems. Pp 1-9 in Use of prescribed burning in western woodland and range ecosystems, a symposium. Utah State Univ., Logan. Pusateri, F. M. 1990. Greater prairie-chicken recovery plan. Colorado Div. Wildl., Fort Collins. 24pp. Queal, L. 1973. Fire -- Tool or tyrant. Kansas Game and Fish. 30(1):7-9. Robel, R. J., J. N. Briggs, A. D. Dayton, and L. C. Hulbert. 1970. Relationships between visual obstruction measurements and weight of grassland vegetation. J. Range Manage. 23:295-297. Schacht, W., and J. Stubbendieck. 1985. Prescribed burning in the loess hills mixed prairie of southern Nebraska. J. Range Manage. 38:47-51. Schmutz, E. M., M. E. Reese, B. N. Freeman, and L. C. Weaver. 1982. Metric belt transect system for measuring cover, composition, and production of plants. Rangelands 4:162-164. Schroeder, M. A. and C. E. Braun. 1991. Walk-in traps for capturing greater prairie-chickens on leks. J. Field Ornithol. 62:378-385. 1992. Seasonal movement and habitat use by greater prairie-chickens in northeastern Colorado. Colo. Div. Wildl. Spec. Rep. 68. 44pp. Snyder, W. D. 1990g. Establishing switchgrass for wildlife in eastern Colorado. Colo. Div. Wild1., Outdoor Facts. Game Infor. Leaf. 113. 4pp. 1990Q. Sandsage-b1uestem prairie renovation. Job Progress Rep., Colorado Div. Wildl., Wi1dl. Res. Rep., Fed. Aid Proj. W-152-R. Apr.:173-202. 1991. Sandsage-b1uestem prairie renovation. Job Progress Rep., Colorado Div. Wi1dl., Wi1d1. Res. Rep., Fed. Aid Proj. W-152-R. Apr. :191-221. Vance, R. 1977. Interactions of pheasants and prairie chickens in Illinois. Pp 9 (abstract) in Proceedings Prairie Grouse Technical Council Biannual Meeting. 12: Pierre, SD. Van Sant, B. F., and C. E. Braun. 1990. Distribution and status of greater prairie-chickens in Colorado. Prairie Nat. 22:225-230 Westemeier, R. L. 1972. Prescribed burning in grassland management for prairie chickens in Illinois. Proc. Tall Timbers Fire Ecol. Conf. 12:317-338. Whisenant, S. G., and D. W. Uresk. 1989. Burning upland, mixed prairie in Badlands National Park. Prairie Nat. 21:221-227. White, R. S., and P. O. Currie. 1983. Prescribed burning in the northern Great Plains: yield and cover responses of 3 forage species in the mixed grass prairie. J. Range Manage. 36:179-183. Wolfe, C. W. 1972. Effects of fire on a sandhills grassland environment. Proc. Tall Timbers Fire Ecol. Conf. 12:241-255. Wright, H. A. 1974g. Effect of fire on southern mixed prairie grasses. J. Range Manage. 27:417-419. 1974Q. Range burning. J. Range Manage. 27:5-11. and A. W. Bailey. 1980. Fire ecology and prescribed burning in the Great Plains -- a research review. U.S. Dep. Agric., For. Servo Gen. Tech. Rep. INT-77. Prepared by: 1tJoJd:t.J~ Warren D. Snyder Wildlife Researcher ��205 JOB PROGRESS REPORT State of: ~C~o~l~o~r~a~d~o~ Proj ect: 'Work Plan: _ 'W-167-R 21 Job: Upland Bird Research 5 Job Title: Evaluation of Habitat Quality on Conservation Reserve Lands in Eastern Colorado Period Covered: 1 January through 31 December 1991. Author: 'Warren D. Snyder ABSTRACT Monitoring of vegetation quality as nesting cover was limited to 41 fields in 1991 of the 140 Conservation Reserve Program fields normally sampled. Sampling was primarily confined to the eastern tier of strata. Vegetation height density remained the same as that obtained in 1990 during pre-greenup sampling but increased dramatically during nesting season samples, primarily _ within northeastern Colorado. Above average rainfall in May and June resulted in the increased growth. Canopy cover of perennial grass continued to ( -v . increase within sampled fields. -, ��207 EVALUATION OF HABITAT QUALITY ON CONSERVATION RESERVE LANDS IN EASTERN COLORADO Warren D. Snyder P. N. OBJECTIVES Determine distribution and quantity of Conservation Reserve Program (CRP) land in eastern Colorado in relation to distribution of selected wildlife species, evaluate the quality of vegetation on these lands for selected wildlife species, measure response of selected wildlife species to the Conservation Reserve Program using existing annual surveys, and evaluate the impact of the Colorado Division of Wildlife's cost-share program on cover quality. SEGMENT OBJECTIVES 1. Conduct evaluations of randomly selected CRP fields within eastern Colorado as part of a regional and national assessment of the Conservation Reserve Program coordinated by the National Ecology Center (NEC) of the U.S. Fish and Wildlife Service. 2. Conduct intensive visual obstruction readings (VOR) in a stratified random sample of CRP fields and proximal controls. 3. Conduct intensive visual obstruction readings (VOR) in a sample of fields cost shared by the CDOW (for enhancement of cover quality) for comparison with CRP fields not cost-shared. 4. Estimate response of selected wildlife species to the CRP based on annual population surveys. 5. Compile and analyze data and prepare annual progress report. METHODS Methods used were described by Snyder (1989, 1990). Visual obstruction readings were based on the Kirsch method of sampling rather than the procedure required by NEC personnel. The number of CRP fields sampled in 1991 was reduced to 41 (contrasted to 140 in preceding years). These were concentrated within strata 1 through 4 (within the eastern part of Colorado). It is anticipated that all fields will be sampled in 1992, the last year for field surveys. �208 RESULTS AND DISCUSSION Environmental Conditions Precipitation recorded at 9 U. S. Weather Bureau stations, distributed within eastern Colorado, fluctuated extensively among months and locations in 1991 (Table 1). Large areas within strata 1 and 2 (Fig. 1) received considerable rainfall from early May through mid June, 1991 as indicated by data from Holyoke, Akron, and Burlington. Most stations recorded average or better precipitation in July, however, it was below that received the previous year. Overall, spring"'summer rainfall averaged near or above normal in 1991 and was conducive to considerable growth of vegetation within CRP fields, especially those within the eastern tier of strata. Springfield (Baca County) did not lead in total annual rainfall for th~ first time,in ~ev~ral years based on data through the 'fiq;t 10'monthsof.199l; howeve r , it ,iike most stations in the eastern part of Colorado, received above average precipitation (Table 1). Only 3 stations, Sterling, Akron, and Rocky Ford received below average amounts. Table 1. Monthly (Jan - Oct) precipitation (in.) at 9 U. S. ~eather Bureau locations in eastern Colorado, 1991. Springfield Lamar Rocky Ford Limon Burlington Akron Greeley Holyoke Sterling Jan 0.48 0.08 0.42 0.40 0.06 0.07 0.40 0.18 0.29 Feb 0.05 0.01 0.10 0.03 0.08 0.14 T 0.28 0.11 Mar 0.84 0.93 1.45 1.79 1.27 1.11 0.34 1.54 0.66 Apr 1.13 0.99 0.80 1.40 0.41 0.84 1.21 0.87 0.17 May 1.28 1.36 0.71 1.37 2.44 4.10 2.25 6.45 2.87 Jun 2.48 3.04 1.44 1.63 4.17 2.07 2.38 3.39 2.04 Jul 3.84 1.30 2.08 3.85 5.30 3.15 2.97 1.89 1.30 Aug 1.95 6.39 0.91 6.02 1.77 1.01 0.85 0.33 2.02 Sep 2.73 1.22 1.20 0.27 1.04 0.12 1.00 1.89 2.28 Oct 0.47 0.47 0.73 0.60 1.13 0.40 0.67 1.51 0.58 Sun 15.25 15.79 9.84 17.36 17.67 13.01 12.07 18.33 12.32 14.80 13.58 10.30 12.85 14.32 14.70 11.75 16.76 14.23 Month Longterm x ,"" �209 we L 0 1 ... Fig. 1. Strata for sampling conservation reserve program fields in eastern Colorado. ~ or �210 Evaluation of Conservation Reserve Vegetation Quality Monitoring of vegetation quality was confined to the eastern tier of strata (plus 4 fields in southern stratum 4) in 1991. It included 41 fields; 23 of which had been cost shared for cover quality enhancement by the Division of Wildlife. During pregreenup sampling, height-density within stratum 1 increased over that obtained the previous year (Table 2). However, changes within other strata were less evident and the overall average changed little from 1990 (0.74 dm) to 1991 (0.73 dm). Heightdensity changes from 1990 to 1991 during summer (nesting) were much more dramatic (Table 2). Abundant rainfall received in strata 1 and 2 was reflected in markedly higher height-density indices. Changes in strata 3 and 4 were less pronounced. Table 2. Average height-density (~) indi.ces obtained from a partial sample of Conservation Reserve Program fields during pre-greenup 'and nesting intervals during 1991 compared with indices from the same fields in 1990, eastern Colorado. 11 1990 Pre-greenuE 1991 1990 1991 1 14 0.54 0.80 1.01 2.51 2 10 0.95 0.62 1.06 1.76 3 13 0.58 0.65 1.08 1.28 4 4 1.42 0.99 1.64 1.24 41 0.74 0.73 1.14 1.81 Stratum Total Nesting Average height-density during pre-greenup sampling was 0.59 dm within 18 NEC random fields in comparison to 0.73 dm within 23 CDOW cost shared fields. During the nesting season these respective indices increased to 1.14 and 1.81 dm. Perennial grass approximated 45% canopy cover during both the pre-nesting and nesting season samples (Table 3). Perennial grass canopy cover during pre-greenup increased from 24.2% (1990) to 44.2% (1991), and from 30.6% (1990) to 45.2% (1991) during nesting season. Annual vegetation declined from 1990 to 1991 within both the pregreenup and nesting seasons. During nesting season sampling, vegetation height was tallest within stratum 1 (37.7 cm) and averaged 30.1 cm among all samples. �Table 3. Canopy cover (%) of total vegetation, perennial grass, and annual vegetation, and mean vegetation height (em) within Conservatio~ Reserve Program fields during pre-greenup and nesting intervals in 1990 and 199i, eastern Colorado. Nesting Pre-greenup Variable 1990 1991 1990 1991 Perennial grass 24.2 44.2 30.6 45.2 Annual vegetation 38.2 31.4 34.7 17.2 Total canopy cover 62.4 75.6 65.3 62.4 Vegetation height 16.4 11.9 LITERATURE CITED Snyder, Y. D .. 1989. Evaluation of habitat quality on conservation reserve lands in eastern Colorado., Colorado Div. Yildl., Yildl. Res. Rep., Fed. Aid Proj. Y-152-R. Apr. 221-244. 1990. Evaluation of habitat quality on conservation reserve lands in eastern Colorado. Colorado Div. Yildl., Yildl. Res. Rep., Fed. Aid Proj. Yl52-R. Apr. 205-223. Prepared by 11l4llrAt...2 ¥ Yarren D. Snyder Yildlife Researcher or ��213 JOB FINAL REPORT State of: Colorado Project: 'W-lS2/167-R Upland Bird Research 'Work Plan: 21 Job Title: Evaluation of 'Wildlife Responses to the Conservation Reserve Program Period Covered: Author: : Job _6_ 01 January 1989 through 31 December 1991 Thomas E. Remington Personnel: Clait E. Braun, Bradley J. Parks, Thomas E. Remington, 'Warren D. Snyder, David A. 'Wilson, Colorado Division of 'Wildlife ABSTRACT Breeding birds were counted a~ong line transects within 27-39 Conservation Reserve Program (CRP) fields throughout northeastern Colorado during spring 1989-91. Densities of breeding birds were computed from counts using a Fourier series estimator. Initially, horned larks (Eremophila alpestris) were the most common and abundant breeding bird, but the percentage of fields occupied and breeding densities declined over time (11.0, 6.9, 4.0 birds/10 ha, 1989-91, respectively). Conversely, frequency of occurrence and density of lark buntings (Calamospiza melanocotys), western meadowlarks (Sturnella neglecta), and grasshopper sparrows (Ammodramus savannarum) increased over time as cover quality increased. Breeding density of lark buntings increased from 4'.0birds/lO ha in 1989 to 6.1 and 13.1 in 1990 and 1991, respectively. Meadowlarks were common in all years, but at relatively low density (1.4, 1.9 and 3.3 birds/10 ha from 1989 to 1991, respectively. Grasshopper sparrows were common and reasonably abundant (3.7, 4.0, and 6.0 birds/10 ha from 1989 to 1991, respectively). Mourning doves (Zenaida macroura) were common breeders but many probably went undetected because of their inconspicuous breeding behavior. They were found in about half the fields sampled at a density of about 1-3 birds/lO ha. Breeding birds found in a few ffelds included ring-necked pheasant (Phasianus colchicus), Cassin's sparrow (Aimophila cassinii), dickcissel (Spiza americana), upland sandpiper (Bartramia longicauda), red-winged blackbird (Agelaius phoeniceus), northern bobwhite (Colinus virginianus), and lark sparrow (Chondestes grammacus). Other birds observed were migrating (vesper sparrow [Pooecetes gramineus], white-crowned sparrow [Zonotrichia leucophtys], song sparrow [Melospiza melodia]), or nested in adjacent rangeland or shelterbelts (loggerhead shrike [Lanius ludovicianus], eastern and western kingbirds [Iyrannus tyrannus and T. verticalus]). Common breeding birds in CRP fields.were also common breeders in green wheat and/or wheat stubble. Rope drags were used to locate nests in CRP fields, green wheat, and wheat stubble. Nests of.9 species were found in CRP fields, but nests of lark buntings, mourning doves, and meadowlarks comprised 90 and 93% of all nests found in 1990 and 1991. ��215 Evaluation of Wildlife Responses Thomas .to the Conservation Reserve Program E. Remington P. N. OBJECTIVES 1. Measure and compare avian use of CRP to use of alternative (wheat and summer fallow). 2. Measure and compare nesting wheat and summer fallow. 3. Relate patterns characteristics success of birds of avian use and nest success and surrounding land uses. land uses on CRP to success in CRP fields on to cover Title XII of the Food Security Act of 1985 established a Conservation Reserve Program (CRP). The goal of the CRP program is to retire 45 million acres of highly erodible cropland from production for 10 years. Benefits of the program are primarily to reduce erosion and surplus commodities, but other benefits, such as wildlife habitat improvement, are likely (Bartlett 1987). Response by Colorado landowners to the CRP has been enthusiastic, with almost 2 million acres enrolled. The response of many wildlife agencies and personnel has been no less enthusiastic. The Colorado Division of Wildlife (CDOW) established a program to cost-share up to 70% of the landowners share of costs to establish vegetative cover under CP 1, CP 2, or CP 4. The goal of this program is to improve habitat for primarily pheasants and prairiechickens (Tympanuchus spp.). The Division's investment in this program was about $385,000. CRP tracts may benefit rap tors by providing undisturbed areas where prey bases may increase. Other birds are expected to be affected primarily through changes in quantity and quality of nesting habitat. Bird populations may change either because the quantity of nesting habitat is altered so that some segment of the population is or is not non-breeding, or because nest success changes concurrent with changes in habitat quality. Wildlife response to the CRP will depend on land-use patterns in the surrounding area, seed mixture and techniques used to establish perennial cover, vegetative response within CRP fields, species composition and densities of wildlife in the region, and many other factors. The USDA and state wildlife agencies must have information on the benefits of CRP to wildlife to plan intelligently for the future of CRP tracts after contracts expire. Without knowledgeable planning, it is likely that most CRP land will, as in previous Federal set-aside programs, return to agricultural production and any wildlife values will be negated (Bedenbaugh 1987, Leitch 1987, Cacek 1988). The objective of this study was to measure and compare avian use of CRP to use in alternative land uses (wheat and summer fallow). �216 METHODS A subset (n-68) of CRP fields randomly selected by the U.S. Fish and Wildlife Service (from the universe of all CRP fields in eastern Colorado) was randomly chosen for study. A sample (12) of fields in which the Colorado Division of Wildlife cost-shared on a wildlife grass seed mixture was added for comparison. Permission to enter and measure bird abundance within these fields was obtained via letter, and usually by telephone, from landowners. Green wheat (n-17) and wheat stubble fields (n-13) near CRP fields were also sampled. CRP, and wheat and wheat stubble fields, were sampled within an hour of sunrise to count breeding birds. A 0.8-km long transect was established near the midpoint of each field with plastic flags. An observer walked along the transect line slowly watching and listening for birds. Birds observed were identified to species and the perpendicular distance from each bird to the line transect was measured to the nearest meter using a rolotape. In 1990, each field was censused once between 20 April and 20 May and again between 21 May and 20 June to ensure both early and late breeding species were counted. Breeding bird density was estimated from perpendicular distance measures (from the higher of the 2 counts) using a Fourier series estimator (Burnham et al. 1980, 1981). Eight hectares within each CRP and green wheat or wheat stubble fields were searched for nesting birds by dragging a 30-meter long rope between 2 observers. The area around the point from which a bird flushed was searched intensively for a nest. In 1990 each field was searched 2-3 times throughout the breeding season in an attempt to document nesting by all birds using the field. Hen pheasants were captured and fitted with radio transmitters between 6 March and 23 April 1991, to evaluate extent of use and relative success of pheasant hens nesting in CRP fields. Radio-marked hens were relocated periodically through the nesting season to ascertain the cover types in which they nested and measure nest success. Hens were captured at night with long-handled nets (Snyder 1977) on or adjacent to the Turk's Pond State Wildlife Area in Baca County. Captured hens were weighed, banded, primary 1 was removed for age classification, and radio transmitters were attached. Transmitters were manufactured by either Holohil Systems (model RI-2B; 9 gms) or by Telemetry Systems Inc. (model LT1800-lSM-LD, 19 grams with poncho). Holohil transmitters were equipped with 10 cm elastic loops which were used to attach them around the neck. Telemetry Systems radios were mounted on naugehyde ponchos and fitted to each birds neck. RESULTS AND DISCUSSION Breeding Bird Study Twenty-four species of birds were observed along line transects within CRP fields, but only 6 species were common resident breeders (Table 1). Frequency of occurrence and density of birds in CRP fields varied over time as grass became established and cover quality increased. Most fields sampled entered the program in 1986 or 1987, a cover crop (sorghum) was planted in 1987 or 1988, and grass was planted in 1988 or 1989. Thus, initial censusing was of �Table 1. Density of breeding breeds observed in Conservation Reserve Program (CRP) fields, wheat fields, and wheat stubble, 1989-91, and percentage of sampled CRP fields in which birds were observed. CRP Fields Freauencv1 Species Densitv 1989 1989 1990 1991 x Horned lark Lark bunting Grasshopper sp. Meadowlark Cassin's sp. Mourning dove Vesper sp. Red-winged b1ackb. Pheasant Western kingbird Mallard Dickcissel Song sparrow Northern bobwhite "Lark sparrow Field sparrow House wren Killdeer Northern oriole American robin Loggerhead shrike Upland sandpiper ~Percentage 64 67 46 82 26 46 28 10 8 10 8 3 3 3 3 3 3 3 3 3 3 3 70 89 93 100 11 56 22 41 15 11 42 95 84 82 13 63 0 21 26 5 4 7 0 0 0 0 0 0 0 0 11 8 0 5 5 0 0 3 0 0 0 0 11.0 4.0 3.7 1.4 0.8 1.2 1990 Wheat Stubble Density Density 1991 S.D. x S.D. x S.D. x 12.0 4.7 8.4 1.3 1.7 2.3 6.9 6.1 4.0 1.9 1.3 7.9 8.7 3.2 4.0 13.1 6.0 3.3 1.3 1.9 1.1 8.3 9.9 6.0 2.3 3.5 6.4 2.8 S.D. x 16.1 1.6 1.7 S.D. 3.8 9.9 7.0 of sampled fields species was found in. N •..... -.._J �218 fields either been planted. stands. during, 1, or in a few cases, 2 growing seasons after, grass had Censusing in 1990 and 1991 was of progressively maturing grass In 1989, when cover was sparse, horned larks were the most abundant breeding bird while meadowlarks were the most ubiquitous, i.e., they occurred in the most fields (27 of 27; Table 1). Density of horned larks in CRP declined steadily and fairly dramatically as grass cover increased. Lark bunting frequency and density increased markedly from 1989 to 1991 (Table 1). In 1991 they were the most common (observed in the most fields) and abundant bird in CRP fields. Grasshopper sparrows were relatively common, especially in CRP fields with better grass and/or weedy cover. Grasshopper sparrow densities also increased markedly from 1989 to 1991 when they were the second most common and abundant bird. Mourning doves were fairly common in fields with low, sparse cover, but were probably under represented because of their tendency to display from telephone wires or elevated perches near field edges. Cassin's sparrows were commonly observed in 1989 in fields that were planted to native bunch grasses. We observed only 5 in 3 fields in 1990, and 13 in 5 fields in 1991. Fields sampled in 1990 and 1991 were primarily in northeastern Colorado where most fields were planted to smooth brome. Fields planted to bunch grasses in this area had yet to develop good stands of grass. Dickcissels were observed while singing in 1, 2, and 3 fields in 1989-91, respectively, where sweet clover was abundant. Northern bobwhite were flushed from 3 fields that either had a tree grove within the field (1 field) or bordered wooded or brushy areas,.but breeding was not confirmed. Vesper sparrows were frequently observed in late April and early May, but were not observed later and were presumed to be using CRP fields while migrating. Occasional use of CRP fields by species that nested elsewhere (in trees or shrubs) was noted. These species included lark sparrow, northern oriole (Icterus galbula), American robin (Turdus migratorius), loggerhead shrike, and eastern and western kingbirds. Green wheat and wheat stubble were also used as breeding and nesting habitat for several of these birds. Horned larks were the most common and abundant bird in green wheat and wheat stubble fields (Table 1), although estimated breeding densities may have been inflated by inclusion of some birds foraging for waste grain or insects. Lark buntings were also observed, presumably breeding, in green wheat, and to a much lesser extent in wheat stubble. Meadowlarks were commonly observed in green wheat and wheat stubble, at densities comparable to (stubble) or higher (green wheat) than in CRP fields. Grasshopper sparrows were observed in wheat stubble, but not in green wheat. Estimated densities were low relative to densities in CRP fields (Table 1). Nests of 9 species were found during searches of CRP fields (Table 2). Lark bunting nests were most frequently encountered, followed by mourning dove and meadowlark nests. In 1990 and 1991 these 3 species accounted for 90 and 93% of nests encountered in CRP fields, respectively. Horned lark, grasshopper sparrow, red-winged blackbird, Cassin's sparrow, and pheasant nests were encountered at low densities. Green wheat and wheat stubble fields were searched for nests to confirm use by breeding birds. Relatively little effort was put into these habitats when few nests were found. Ten nests were found in 12 green wheat fields searched; 2 lark bunting nests, 5 red-winged blackbird nests, and 1 nest each of horned lark, mallard CAnas platyrhynchos) and pheasant. Four nests were found in 6 wheat stubble fields searched; 3 lark bunting nests and 1 horned lark nest. �219 Table 2. Number of bird nests located during rope-drag searches wheat, and wheat stubble fields in eastern Colorado, April-June, CRP Species (N fields) Lark bunting Mourning dove Meadowlark Red-winged blackbird Grasshopper sparrow Horned lark Lark sparrow Mallard Pheasant Cassins's sparrow of CRP, green 1989-l99l. Green wheat 'Wheat Stubble 1989 (29) 1990 (29) 1991 (36) 1989-1991 (12) 1989-1991 29 12 9 2 2 10 106 60 9 4 6 5 3 177 70 23 6 3 1 2 3 3 7 1 (6) 5 1 1 1 1 The relative lack of nests in green wheat fields was surpr1s1ng. During searches for carcasses in 11 green wheat fields (sprayed with insecticides to control wheat aphids) in spring 1989 we located 6 lark bunting nests, 5 horned lark nests, and 4 mourning dove nests. Although twice as much area was searched, search methodology presumably was less likely to detect nests since, rather than drag a rope between two observers, several people walked abreast about 20 rows apart. None of these fields bordered, or were within about 10 km of any CRP fields. It seems likely that increased density of birds and nests in CRP fields over time resulted, at least in part, from birds shifting breeding activity from wheat and/or wheat stubble fields into CRP fields. The relative lack of nests in green wheat or wheat stubble in this study probably occurred because we selected fields close to, or in some cases adjacent to, CRP fields. The number of nests found for each species was not necessarily indicative of breeding density or effort. Horned larks and meadowlarks were probably under represented because, in some years, nest searching may have been conducted after the peak of nesting for these species. The peak of nesting effort by horned larks in 1989 was late May. In 1990, despite intensive searching beginning in late April, we found only 5 horned lark nests, 4 of these were apparently second nest attempts found on 14 June. Either 1989 was a relatively late year or 1990 and 1991 were relatively early years for horned lark nesting. Grasshopper sparrow nests were difficult to find because birds either ran off nests prior to flushing or didn't flush in response to the rope. Most of the 6 grasshopper sparrow nests were found when a bird flushed off the nest because we walked within 1 m. Only one pheasant nest was located in 1990, and none were located in 1991. While the quality of cover for nesting pheasants within these fields was generally poor, it was also possible that pheasant hens did not flush in response to the rope. Nest densities of mourning doves were more indicative of dove use of CRP fields than breeding density estimates from the line transects. �220 Pheasant Nest Success Fifty-one hens were captured and fitted with radio-transmitters, 29 adults (57%), 21 juveniles (41%) and 1 bird for which age could not be ascertained. Of these 51, 8 birds were judged to have lost their transmitters (6 Ho1ohi1, 2 TSI), apparently by pulling them over their heads. This was confirmed in one instance when a bird was recaptured without it's radio (TSI). Some of these may have been predated by coyotes (Canis latrans), which seldom left carcasses or evidence of predation other than chewed radios in some cases. Contact was lost with 5 radios between 2 April and 15 May. Of these, 1 juvenile disappeared at a time period when other juvenal birds were dispersing from the core winter area (mid-Apr), and may have dispersed to a location where we were unable to get a signal. The other 4 birds were likely predated, although it is possible their radios failed. Two birds disappeared 2 weeks before any known dispersal off the area, and two disappeared several weeks after dispersal had apparently ended. Radios from predated birds were frequently difficult to recover, either because they were carried long distances to the nest by a great horned owl (Bubo virginianus) (1 known radio), or more frequently because coyotes had chewed off the antenna or buried the radio. Relatively little useful nesting information was obtained because of extensive mortality prior to, and during, laying. Of 43 birds known to have retained radios, only 8 (19%) were known to have survived long enough to complete incubation. Two of these 8 hens successfully brought off young. Nesting attempts by the other 6 hens failed because of nest predation (6) or abandonment (3). Mortality was attributed almost exclusively to predation by great horned owls and coyotes, although 1 hen was struck and killed by a vehicle along the county road which transected the study area. Owls were frequently observed along Horse Creek to the north, in the trees bordering Turk's Pond on the west, and on an abandoned trailer just south of the intensive trapping area. An active coyote den was located on the north end of the CRP field where most birds were trapped. This den was within the home ranges of most pheasants wintering on the study area. Another active den was located about 1.25 miles south. Predation was extensive despite the fact that cover in this area was relatively good. About half the CRP field (that served as a focal point for most pheasant home ranges) had been planted to a wildlife mixture consisting of bunch grasses, and about 40 acres was planted with the Division cost-shared seed mixture of primarily switchgrass. The wildlife area had ample cover, including strips of unharvested wheat, blocks of unharvested alfalfa, weedy cane along the pond, and extensive willow/weed complexes along Horse Creek. Why was mortality so extensive? I believe two factors were primarily involved. Because this region is a mosaic of cropland, rangeland, and wooded creek bottoms, the density of coyotes and owls appears quite high. The imposition of CRP fields into this mosaic may have created an ecological trap for pheasants, by concentrating them in an area where coyotes denned and hunted, and an area where owls could effectively hunt. Prior to CRP, pheasant hens would likely have been dispersed throughout the area, feeding, nesting, and roosting in primarily cropland. Coyotes likely denned and hunted in rangelands, while owls nested and hunted from creek bottoms. Thus, the net effect of CRP in mixed rangeland, cropland, and creek bottom areas such as this might have been to concentrate pheasants· and predat~rs in the same fields, with inevitable consequences. �221 MANAGEMENT IMPLICATIONS The net effect of the Conservation Reserve Program on birds in eastern Colorado is not clear cut or clearly positive. This study has demonstrated that about 8 species of birds utilize CRP for breeding to at least some degree. Other species use it for foraging or hunting. The birds breeding in CRP fields, at this point in the program, primarily represent habitat generalists rather than grassland or grassland/shrub species. Green wheat fields and wheat stubble fields were used for nesting by lark buntings, horned larks, mourning doves, red-winged blackbirds, pheasants, mallards, and probably grasshopper sparrows and meadowlarks. These are also the species commonly using CRP fields. Thus, the benefits of this program, which largely replaced a wheat/wheat stubble rotation, to breeding birds are unclear. If nesting success in CRP fields is superior to success in wheat and/or stubble, then populations may have increased. It is possible however, that birds would have bred at the same density and with the same success in those fields without the CRP. Species diversity within CRP fields was not enhanced to any significant extent, primarily because vegetative structure was similar to that provided by wheat and/or wheat stubble. The utility of the CRP on pheasants can be increased. Bunch grasses, particularly switchgrass, which provide ample residual cover in spring, should be encouraged where pheasant production is a priority. Smooth brome, the predominant grass planted in northeastern Colorado, was not used by pheasants for nesting. In future signups, or if revegetation is desired, switchgrass should be encouraged to increase nest success and survival of pheasants. Pheasant survival within CRP fields can probably be increased in arid areas such as Colorado by planting food-cover patches of tall forage sorghum, particularly in fields planted to smooth brome. Woody plantings within CRP fields would enhance diversity of breeding birds. Bobwhite quail, lark sparrows, loggerhead shrikes, American robins, mourning doves, and red-winged blackbirds would be expected to use CRP (or use it more extensively) for breeding if woody plantings were present. Bunch grasses seem to provide acceptable habitat structure for the broadest range of bird species, and provide significantly better nest protection for pheasants, mallards, and other large birds. Smooth brome was an adequate cover for horned larks, lark buntings, and mourning doves, but supported only low densities of other nesting birds. �222 LITERATURE CITED Bartlett, E. T. 1987. Social and economic impacts of the Conservation Reserve Program. Pages 52-54 in J. E. Mitchell, ed. Impacts of the Conservation Reserve Program in the Great Plains. U.S. Dep. Agric., For. Serv., Gen. Tech. Rep. RM-158. Bedenbaugh, E. J. 1987. History of cropland set aside programs in the Great Plains. Pages 14-17 in J. E. Mitchell, ed. Impacts of the Conservation Reserve Program in the Great Plains. U.S. Dep. Agric., For. Serv., Gen. Tech. Rep. RM-158. Burnham, K. P., D. R. Anderson, and J. L. Laake. 1980. Estimation of density from line transect sampling of biological populations. Wildl. Monogr. 72. 202 pp. _________ , and . 1981. Line transect estimation of bird population density using a fourier series. Studies Avian BioI. 6:466-482. Cacek, T. 1988. After the CRP contract expires. Conserv. 43:291-293. J. Soil and Water Leitch, J. 1987. Policy questions from CRP in the Midwest. Pages 91-94 in J. E. Mitchell, ed. Impacts of the Conservation Reserve Program in the Great Plains. U.S. Dep. Agric., For. Serv., Gen. Tech. Rep. RM-158. Snyder, W. D. 1977. Pheasant mortality investigation. Job Final Rep., Colorado Div. Wildl., Wildl. Res. Rep., Fed. Aid Proj. W-37-R-30. Apr. :1-33. Thomas E. Remington Wildlife Researcher �223 JOB FINAL REPOR.T State of: Colorado Project: W-152-R/W-167-R Work Plan: 21 Job Title: Avifauna Responses to Grazing Period Covered: Author: Personnel: Upland Bird Research : Job _7_ 01 January 1989 through 31 December 1991 Cynthia P. Melcher C. P. Melcher and Beatrice Van Horne, Colorado State University; C. E. Braun, K. M. Giesen, C. E. Poley, T. E. Remington, and S. M. Swift, Colorado Division of Wildlife Researchers from the Colorado Division of Wildlife monitored densities of white-tailed ptarmigan (Lagopus leucurus) on Trail Ridge in Rocky Mountain National Park (RMNP) from 1966-1988. A decline in the population was detected in the late 1970s when a population increase had been expected. The decline was deep and prolonged, with no sign of recovery by spring 1989. Willow (Salix spp.), an essential food resource for ptarmigan, appeared damaged and decadent in areas where breeding densities of ptarmigan had declined most. Researchers with the Colorado Division of Wildlife hypothesized that breeding densities of ptarmigan had declined in response to habitat (willow) degradation caused by elk (Cervus elaphus) in alpinejkrummholz habitat. Densities of elk in RMNP had been controlled artificially until 1968 when the National Park Service discontinued control programs; the elk population subsequently increased significantly. Fragmentation and loss of elk habitat, extirpation of predators, and heavy hunting pressure immediately outside RMNP's boundaries may have encouraged elk to remain within the Park and increase their use of some habitat types, including alpinejkrummholz. The objectives of this study were to evaluate relationships between avian abunda,nce/species richne~s, presence of ungulates, characteristics of willow, and activity budgets/reproductive success of ptarmigan hens ·in alpine/krummholz habitats of Rocky Mountain National Park. All bird species were censused at 4 17-22 ha sites using spot-mapping techniques from 27 May to 10 July 1989, and from June 7.to 1 August 1990. Activity budgets, foraging rates, and reproductive success of 12 radio-marked hens (three at each site) were studied during the breeding season of 1990. Percent willow cover (shrub and mat forms), bud densities and densities of live unbrowsed terminal leaders (shrub forms) were measured as available willow-food resources for ptarmigan among sites. Patch perimeter and diameter, height, densities of live browsed and/or dead terminal leaders, and lengths of terminal leaders of shrub willows were measured as indicators of browsing by ungulates among sites. Frequencies of ungulate feces were measured as indicators of ungulate presence among sites. Willow characteristics and ungulate feces were measured in both 1989 and 1990. Ungulate feces were significantly (f <0.01) more frequent at the two southeast (SE) sites on Trail Ridge than they were at.the two northwest �224 (NV) sites. Overall avian species richness and densities were greater at SE sites, primarily due to inherent site differences. However, .. ptarmigan densities, including pairs and unpaired males, were lowest at SE sites, and declines in ptarmigan densities since 1966 were significant (l <0.01). In 100 x 50 m sampling plots and on 10-m transects, willow characteristics indicative of food-resources available to ptarmigan tended to be greater at NV sites, although not significantly (l >0.05). On lO-m transects, however, willow cover was significantly greater (l <0.05) at locations used by ptarmigan hens than it was at random locations within the territories of the hen's mates. All but two willow characteristics indicative of browsing by ungulates were similar among sites (l >0.05). Lengths of live unbrowsed terminal leaders were significantly greater (l >0.05) at SE sites (where ungulate presence had been greater), perhaps in response to browsing. Also, willow-patch diameters (on transects) were greater (l <0.05) at SE sites; patch diameters were smaller (l <0.05) at locations used by ptarmigan hens than they were in random locations within the territories of the hen's mates; and patch diameters at locations used by hens more closely matched patch diameters found at NV sites. Time budgeted to feeding by, and bite rates of, ptarmigan hens were not different among regions of Trail Ridge. However, bite rates on shrub willow were significantly lower (l <0.01) than bite rates on other foods, including mat willow. Reproductive success was poor for all hens, which precluded detection of site-related differences in reproductive success. �225 THESIS AVIFAUNA RESPONSES TO INTENSIVE BROWSING BY ELK IN ROCKY MOUNTAIN NATIONAL PARK Submitted by Cynthia Porter Melcher Department of Biology In partial fulfillment of the requirements for the Degree of Master of Science Colorado State University Fort Collins, Colorado Spring 1992 �226 COLORADO STATE UNIVERSITY April 6. 1992 VE HEREBY RECOMMEND SUPERVISION INTENSIVE ACCEPTED THAT THE THESIS PREPARED UNDER OUR BY CYNTHIA P. MELCHER ENTITLED AVIFAUNA BROWSING BY ELK.IN ROCKY MOUNTAIN NATIONAL AS FULFILLING IN PART REQUIREMENTS RESPONSES TO PARK BE FOR THE DEGREE OF MASTER OF SCIENCE. .. Committee on Graduate Work Adviser ./"' I'~ Department Head �227 ABSTRACT OF THESIS AVIFAUNA RESPONSES TO INTENSIVE BROWSING BY ELK IN ROCKY MOUNTAIN NATIONAL PARK Researchers from the Colorado Division of Wildlife monitored densities of white-tailed ptarmigan (Lagopus leucurus) on Trail Ridge in Rocky Mountain National Park (RMNP) from 1966-1988. A decline in the population was detected in the late 1970s when a population increase had been expected. The decline was deep and prolonged, with no sign of recovery by spring 1989. Willow (Salix spp.), an essential " food resource for ptarmigan, appeared damaged and decadent in areas where breeding densities of ptarmigan had declined most. Researchers with the Colorado Division of Wildlife hypothesized that breeding densities of ptarmigan had declined in response to habitat (willow) degradation caused by elk (Cervus elaphus) in alpinefkrummholz habitat. Densities of elk in RMNP had been controlled artificially until 1968 when the National Park Service discontinued control programs; the elk population subsequently increased significantly. Fragmentation and loss of elk habitat, extirpation of predators, and heavy hunting pressure immediately outside RMNP's boundaries may have encouraged elk to remain within the Park and increase their use of some habitat types, including alpine/krummholz. �228 The objectives of this study were to evaluate relationships between avian abundance/species richness, presence of ungulates, characteristics of willow, and activity budgets/reproductive success of ptarmigan hens in alpine/krummholz habitats of Rocky Mountain National Park. All bird species were censused at 4 17-22 ha sites using spot-mapping techniques from 27 Hay to 10 July 1989, and from June 7 to 1 August 1990. Activity budgets, foraging rates, and reproductive success of 12 radio-marked hens (three at each site) were studied during the breeding season of 1990. Percent willow cover (shrub and mat forms), bud densities and densities of live unbrowsed terminal leaders (shrub forms) were measured as available willow-food resources for ptarmigan among sites. Patch perimeter and diameter, height, densities of live browsed and/or dead terminal leaders, and lengths of terminal leaders of shrub willows were measured as indicators of browsing by ungulates among sites. Frequencies of ungulate feces were measured as indicators of ungulate presence among sites. Willow characteristics and ungulate feces were measured in both 1989 and 1990. Ungulate feces were significantly (f <0.01) more frequent at the two southeast (SE) sites on Trail Ridge than they were at the two northwest (NV) sites. Overall avian species richness and densities were greater at SE sites, primarily due to inherent site differences. However, ptarmigan densities, including pairs and unpaired males, were lowest at SE sites, and declines in ptarmigan densities since 1966 were significant (f <0.01). In 100 x 50 m sampling plots and on 10-m transects, willow characteristics indicative of food-resources available to ptarmigan tended to be greater at NV sites, although not sigrtificantly (f .. �22~ >0.05). On 10-m transects, however, willow cover was significantly greater (f <0.05) at locations used by ptarmigan hens than it was at random locations within the territories of the hen's mates. All but two willow characteristics indicative of browsing by ungulates were similar among sites (f >0.05). Lengths of live unbrowsed terminal leaders were significantly greater (f >0.05) at SE sites (where ungulate presence had been greater), perhaps in response to browsing. Also, willow-patch diameters (on transects) were greater (f <0.05) at SE sites; patch diameters were smaller (f <0.05) at locations used by ptarmigan hens than they were in random locations within the territories of the hen's mates; and patch diameters at locations used by hens more closely matched patch diameters found at NV sites. Time budgeted to feeding by, and bite rates of, ptarmigan hens were not different among regions of Trail Ridge. However, bite rates on shrub willow were significantly lower (f <0.01) than bite rates on other foods, including mat willow. Reproductive success was poor for all hens, which precluded detection of site-related differences in reproductive success. Cynthia P. Melcher Department of Biology Colorado State University Fort Collins, CO 80523 Spring 1992 '. �230 TABLE OF CONTENTS TITLE PAGE.................................................... Page i SIGNATURE PAGE................................................ ii ABSTRACT. ..................................................... iii TABLE OF CONTENTS............................................. LIST OF FIGURES vi viii LIST OF TABLES................................................ ACKNOWLEDGEMENTS. ............................................. CHAPTER 1. ix x EXTENSIVE USE OF ALPINE/KRUHMHOLZ HABITATS BY WILD UNGULATES: CONSEQUENCES FOR WHITE-TAILED PTARMIGAN? Introduction . Ptarmigan-Elk-Willow Relationships . Hypotheses . Study Sites and Hethods . Plots and Transects: Selection and Sampling . Occurrence of Ungulate Feces . Willow Characteristics Indicative of Available Food Resources . Willow Characteristics Indicative of Browsing by Ungulates . White-tailed Ptarmigan: Sampling . Plots and Transects: Data Transformation and Analysis Occurrence of Ungulate Feces . Willow Characteristics . Ptarmigan Densities: Data Transformation and Analysis . Results Frequencies of Ungulates Feces . Willow Characteristics Indicative of Available Food Resources . Willow Characteristics Indicative of Browsing by Ungulates . White-tailed Ptarmigan Densities Among Sites . Discussion . Literature Cited ; / . .. 1 2 4 4 8 12 13 13 15 15 16 17 18 21 23 26 26 36 �231 CHAPTER2. AVIANSPECIES RICHNESSANDABUNDANCE IN ALPINEjKRUMMHOLZ HABITATSUSED EXTENSIVELYBY llILD UNGULATES IN COLORADO Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. Study Sites....................................... Methods........................................... Results........................................... Discussion. . . . . . . . . . . . . . . . . . . • . . • • . . . . . . . . . . . . . . .. Literature Cited.................................. CHAPTER3. 42 44 45 47 53 59 PRE-NESTINGACTIVITY BUDGETSANDREPRODUCTIVE SUCCESSOF FEMALEWHITE-TAILEDPTARMIGAN USING HEAVILYBROllSEDALPINEjKRUMMHOLZ HABITATS Introduction. . . . . . . . . . . . . .•. . . . . . . . . . . . . . . . . . . . . . . Hypotheses. . . . . . . . . . . . . . . . . . . . • . . . . . . . . . . . . . . . . . Study Sites...... ................................. Methods Selection of Hens.. . . • . • • • • • • • • • • • . • . . . . . • . • • • . . Pre-Nesting Activity Budgets of Hens .••.••..•••. Data Analysis. . . . . . . . . . • • . . . • • . • . . . . . . . . . • . . . • . . Results and Discussion Activity Budgets........................... . • . .. Bite Rates...................................... Measures of Fitness............................. Literature Cited.................................. 63 65 66 68 69 71 72 74 76 83 .. �232 LIST OF FIGURES Figure 1-1 1-2 1-3 1-4 1-5 2-1 Locations of study sites and plots on Trail Ridge in Rocky Mountain National Park, 1989 and 1990 7 Schematic representation of sampling protocol and sample hierarchies used in plots and on transects in Rocky Mountain National Park, 1989 and 1990 10 Observed and average frequencies of ungulate feces in sample plots at four sites in alpinejkrummholz habitats, Rocky Mountain National Park, 1989 and 1990 19 Observed and average frequencies of ungulate feces on transects at four sites in alpinejkrummholz habitats, Rocky Mountain National Park, 1990 20 Densities of white-tailed ptarmigan at four sites in alpinejkrummholz habitats, Rocky Mountain National Park, 1989 and 1990............................................. 27 Avian species richness and average pairs/20 ha at four sites in a1pinejkrummholz habitats, Rocky Mountain National Park, 1989 and 1990 "...•.......................... 48 �233 LIST 1-1 1-2 OF TABLES ANOVA results testing for site (or region) effects on willow characterisitics important to white-tailed ptarmigan as food at four sites in Rocky Mountain Park, 1989 and 1990 22 ANOVA results testing for site (or region) effects on willow characterisitics indicative of browsing by wild ungulates at four sites in Rocky Mountain Park, 1989 and 1990 24 2-1 Average densities of avian species breeding or regularly foraging at four sites in alpinejkrummholz habitats in Rocky Mountain National Park, 1989 and 1990 .. 49 2-2 All avian species observed using four sites in alpine/ krummholz habitats in Rocky Mountain National Park, 1989 and 1990 '. 51 3-1 Summary of activity budgets of pre-nesting white-tailed ptarmigan hens in Rocky Mountain National Park, 1990 ..... 73 3-2 Summary of age, weight, and reproductive success of white-tailed ptarmigan hens in Rocky Mountain National Park, 1990............................................... 77 �234 ACKNOVLKDGEHERTS The financial support, equipment, and field housing for this study was provided by the Colorado Division of Wildlife. Funding sources included the Colorado non-game and endangered species program, Welder Wildlife Foundation, and Federal Aid Restoration Project W-l52-R. My graduate committee went through considerable evolution over time, and I am grateful to every member, past and present, for their respective contributions to this project. Dr. Beatrice Van Horne, Department of Biology, Colorado State University, served as my adviser and gave consistent advice, support, and encouragement; Dr. Thomas Remington, Colorado Division of Wildlife, gave my thesis and manuscripts very thorough review and patiently listened to my counter arguments on every point he made; Dr. David Steingraeber, Department of Biology, Colorado State University, gave many helpful suggestions concerning willow-sampling and some "Good Work" kudos when I needed them most; Dr. Myron Baker, Department of Biology, Colorado State University, challenged me with lively discussions and shared his ideas concerning approaches to this project; Dr. Clait Braun, Colorado Division of Wildlife, initiated this project and secured the funding and resources. Dr. Braun also provided me with opportunities to give presentations, both professional and popular, for which I am grateful. .. �235 I wish to acknowledge the generous cooperation, assistance, and support provided by the staff of Rocky Mountain National Park, National Park Service. In particular, David Stevens, biologist, ensured my procurement of necessary passes and keys to operate within the Park. My field work could not have been completed without the help of Sue Swift and Craig Poley. her companionship was a joy. Sue was an excellent field assistant, and Craig helped with several heavy tasks. I am very grateful to both of them. Many researchers and support personnel at the Colorado Division of Wildlife were very generous with their time and assistance. Kenneth Giesen, who "got" me my "life" ptarmigan, Richard Hoffman, and Dr. Thomas Remington willingly brainstormed with me and provided many exciting research ideas; Diane Hall helped me deal with computer crashes, solved paperwork puzzles, and supported me through several other traumas; Dr. N. Thompson Hobbs tolerated and answered endless questions concerning SAS and data analysis. Numerous other personnel with the Colorado Division of Wildlife also provided assistance or advice. My sincerest thanks go to everyone. At Colorado State University, advice was given generously by numerous faculty and graduate students, including Dr. William Alldredge, Dr. Phillip Lehner, Dr. Ronald Ryder, and Dr. Orrin Myers. Most invaluable was the patience of Dr. David Bowden, who saw me through the statistical analyses. Most of all, I give my love and very special thanks to Ken Giesen, Jean Melcher, Pete Melcher, Maggi Melcher, and Jean Boylan. Without their unfailing love, support, and belief in me, I could not have completed my graduate career. ~ �236 CHAPTER 1 USE OF ALPINE/KRUMMHOLZ HABITATS BY WILD UNGULATES: CONSEQUENCES FOR WHITE-TAILED PTARMIGAN? EXTENSIVE INTRODUCTION White-tailed use habitat sub-alpine in alpine drainage or co-dominated brachycarpa spring Nutt.; basins. almost communities, and drainage in Colorado and alpine basins Pursh, and dominated S. 1976) are used from fall through 1975, Braun et al. 1976, Schmidt on willow Willows Ptarmigan S. brachycarpa, prior early 1988) when buds and terminal in alpine(krummholz of breeding 1988). and catkins altipetens) (Salix planifolia exclusively components (S. planifolia, leaves, Krummholz 1972). et a1. 1980, Schmidt tips, krummholz after Weber (May and Braun willow tundra, and Braun forage also essential (Lagopus leucurus by shrub willows (Hoffman the birds tips ptarmigan territories habitats are (Braun 1969, Giesen hens forage S. nivalis leader extensively Hook.) to and during nesting on buds, leader (May and Braun 1972, May 1975). Elk (Cervus elaphus canadensis) Colorado by the late l800s transplants Colorado. received RMNP initiated residents (Swift 1940). extirpated from through a series However, in 1913, elk were successfully The Estes ParkfRocky 49 elk from northwest in 1915. were nearly The population Mountain Wyoming National just prior grew rapidly, and RMNP staff were complaining reintroduced Park of to (RMNP) area to establishment of and by 1941 local that browsing elk had caused �L~/ significant damage to vegetation (Ratcliff 1941, Packard 1947). As a result, prescribed numbers of RMNP elk were periodically shot or removed until 1968 when the National Park Service (NPS) adopted a new resource-management policy that discontinued manipulative approaches to wildlife management (Natl. Park Servo 1975, 1988). Subsequent surveys indicated exponential growth of RMNP's elk population: the estimated population grew from 1,000 in 1968 to over 3,000 by the late 1980s (Braun et al. 1991), and winter-use of alpine/krummholz regions increased from 29 elk in 1933 to over 300 in 1976 (Stevens 1980b; et al. 1989). Bear Hobbs et al. (1982) estimated that carrying capacity for elk in RMNP was between 1,000-1,700, well below densities observed by the mid 1980s. Ptarmigan-Elk-Willow Relationships Historical accounts (pre-1900s) indicated that elk in Colorado used alpine regions infrequently or intermittently (Swift 1940, Packard 1947, Murie 1951, Thomas and Toweil1 1982). However, accounts of habitat-use by ptarmigan (Braun 1969, Schmidt 1969, Braun and Rogers 1971, Braun et al. 1976, Giesen et al. 1980) and recent accounts of habitat-use by elk in RMNP (Nat1. Park Servo 1975, Stevens 1980b, Green 1982, Bear 1989) indicate that significant overlap of habitat-use by ptarmigan and elk may be occurring. As elk habitat was lost to development and became more fragmented, and as hunting pressure outside Park boundaries increased, elk may have been encouraged to concentrate in the Park, especially in areas where they could find preferred foods. Generally, willows do not comprise significant portions of elk diets, although elk commonly feed on willows in RMNP during all seasons (Kufeld 1973, Hobbs 1979, Hobbs et �238 al. 1981, Baker and Hobbs 1982), and at least two studies (Olmsted 1979, Stevens 1980b) implicated heavy browsing by elk in the decline of several Salicaceae species (S. planifolia, tremuloides S. brachycarpa, Populus Michx.) in RMNP. Structures of avian communities have been altered when habitats were subjected to intensive use by ungulates (Zwickel 1972, Butler 1979, Watson and O'Hare 1979, Ryder 1980, Taylor 1986, Myrberget 1987, Tucker 1987, Knopf et al. 1988, Jackson 1991). may have occurred in RMNP. A similar phenomenon Willard and Marr (1970) showed that heavy trampling by people caused severe damage to alpine vegetation in RMNP, the recovery of which has been exceptionally slow due to severe climatic constraints. While alpine vegetation has evolved to withstand severe climate (Billings 1979), it isn't likely to tolerate processes with which it did not evolve, including intensive and prolonged trampling or browsing by ungulates. Competition for food at high densities may have pressured elk within a1pine/krummho1z areas to incorporate more willow into their diets. As a result, willow-food resources may have declined, causing the decline in breeding densities of ptarmigan. Ptarmigan using areas near Trail Ridge Road (TRR) in RMNP have been censused and banded since 1966 (Braun and Giesen 1990). Prior to the late 1970s, population dynamics of TRR ptarmigan were characterized by 7 to 10-year cycles (Braun et al. 1991). However, during the late 1970s, when a population increase was expected, densities of ptarmigan at the southeast end of Trail Ridge began declining, and eventually fell well below historically-low densities. Braun et a1. (1991) suggested that the deep and prolonged decline .. �indicated that factors besides population were influencing dynamics. This study was designed presence (frequency potentially feces), to ptarmigan, of use by browsing of ungulates densities to quantify of ungulate important indicative presence cyclic fluctuations and ungulates. Hypotheses of ungulate 2) willow characteristics 3) willow characteristics My objective and characteristics of ptarmigan. 1) an indicator of willow was to relate to breeding tested were: Ho There are no differences (P >0.05) in frequencies of ungulate feces between areas where ptarmigan had and had not declined. Ho: Willow characteristics do not differ (P >0.05) between areas with high versus low frequencies of ungulate feces. I predicted that 1) current ptarmigan lower where frequencies cover and densities unbrowsed terminal of ungulate of willow-foods breeding densities feces were higher, used by ptarmigan of ungulate feces, and frequencies of ungulate feces would be characterized variations 2) willow (buds, live leaders) would be lower in areas with high frequencies of live browsed would be and/or dead terminal in willow-patch 3) willows leaders, perimeters in areas with high by high densities low shrub heights, or diameters, wide and short terminal leaders. STUDY SITES AND METHODS Two study sites were selected TRR [Sundance breeding Basin ptarmigan the northwest in the southeast (SB) and Tombstone had declined, Ridge (SE) region of (TR)] , where densities and two study areas were selected (NW) region of TRR [Medicine Bow Curve (MBC) and Gore of in �240 Turnout (GT)], where densities of breeding ptarmigan had not declined (Fig. 1-1). including Two additional criteria were used for site selection, 1) presence of alpine/ krummholz communities, including willows, and 2) minimal visibility from TRR (requested by NPS). Site boundaries were determined by using notable landmarks, compass bearings, and counted paces. Sites were plotted on U.S. Geological Survey topographic maps (scale 1:24000), and site areas were determined using a planimeter. Sites were the areal bases for comparisons of local ptarmigan densities. Approximate site areas were: MBC - 22 ha, GT - 17 ha, SB - 20 ha, TR - 19 ha. Marr (1967) described alpine communities and climates typical of Colorado's Front Range (data collected from four sites in Niwot Ridge area, approximately 32 km south of TRR study sites). typical throughout all TRR study sites were: Community types Trifolium cushion fellfield; Dryas octopetala hookeriana Juz. and/or Dryas-Salix (S. nivalis Hook.) turf; alpine meadow comprised of Kobresia, Carex, Trifolium, Poa, and/or Acomastylis rossii R. Br. (formerly Geum rossii R. Br.); snow accumulation areas comprised of Artemisia, Carex, Poa, and/or Acomastylis; and patches of krummholz comprised of Salix spp. (S. planifolia Pursh, S. brachycarpa Nutt.), Picea engelmanii Parry, and/or Abies lasiocarpa Hook. Communities more typical of dry windswept areas (Dryas and Kobresia turf) dominated NY sites, and communities more typical of moist snow-accumulation areas (Carex and Artemisia) dominated SE sites. Community types were described in detail and mapped by Braun (1969). �Fig. 1-1. Locations of four study areas on Trail Ridge in Rocky Mountain National Park, 1989-90. Northwest (NW) region sites: Medicine Bow Curve (MBC) and Gore Turnout (GT). Southeast (SE) region sites: Sundance Basin (SB) and Tombstone Ridge (TR). Sites are outlined, and were areal bases for censusing white-tailed ptarmigan. Relative locations of 100 x 50 m plots used for sampling frequencies of ungulate feces and willow characteristics are indicated with small solid rectangles. For each site, plot 1 is the most northern plot, plot 2 is the most southern. N ~ .. t-' �I',,) .l:"I',,) ROCKY MOUNTAIN NATIONAL PARK ·":~::::::::::::::1:::::::::. DENVER COLORADO MBC 3505·3658 m ~ W TRAIL RIDGE ALPINE VISITOR CENTER »: N w o > o ...J SUNDANCE MOUNTAIN ~ Z W RAINBOW CURVE RCS z i= z I o o MILES o o 1 1 2 3 2 3 FOREST CANYON '3385.3566 4 KILOMETERS SCALE • APPROXIMATELY 1:73500 .. TSR 0 m~ - =<:» ~ TO ESTES PARK �243 Slopes at each site ranged from near 0° to approximately NW sites were steepest, intermittently. level benches narrow benches SE sites were dominated by relatively and saddles, with steep areas occurring SB was the wettest driest. with relatively Aspect 45°. occurring wide nearlyintermittently. site, TR and MBC were intermediate, and GT was the W at GT, NE to SE at ranged from NE to NW at MBC, N to SB, and NE to S at TR. Plots and Transects: Selection and Sampling In May 1989, two permanent established ungulate 100 x 50 m (rectangular) at each site (Fig. 1-1) for sampling feces and characteristics each plot was chosen subjectively representative bearing willow. of willows. plots were frequencies A starting to ensure presence for compass the first 100-m plot side; subsequent plot sides were laid out in a clockwise marked with permanent point of locally- From the starting point, a random was used to delineate of pattern. 75-cm metal rebar stakes. Plot boundaries were Plots were overlaid with a l-m line grid (Canfield 1941) (Fig. 1-2); 30 line-intercept coordinates were selected plot, ensuring 50 m plot. plots from each 50 x 50 m half of each a fairly even distribution Line-intercepts (Fig. 1-2). plot) were established placement 3 August over each 100 x of l-m2 sample- centers were marked with permanent Thus, 120 l-m2 sample-plots of quadrats, 75- (60 per 100 x 50 m per study site, for a total of 480. edges of sample-plots so that they parallel led plot edges. May through of samples were used as the centers Sample-plot cm metal rebar stakes. standardize randomly To were oriented Sampling was conducted from 27 1989 and from 6 July through 22 July 1990. �I'V ~ ~ Fig. 1-2. Two 100 x 50 m rectangular plots were established within or adjacent to each of 4 sites to sample locally-representative examples of willow. A line-intercept grid overlay (Canfield 1941) was used to define 1-m line intercepts in each plot. Thirty line-intercepts were chosen randomly from each 50 x 50 m half of the plots, ensuring relatively even spacing of samples across plots, and were used to determine the centers of 1-m2 sample-plots. Line transects (10-m) were centered on locations (visual) of radio-marked hens (HL transects) observed during a concurrent study. A random compass bearing was taken from each hen location, and followed for 100 m, where the center of another 10-m transect (RL transect) was established. Sampled areas were delineated during sampling with quadrat frames. To standardize placement of quadrats, quadrat edges paralleled plot edges and quadrat center lines were placed on transect lines. " .' �PLOTS TRANSECTS 30 line-intercepts selected randomly from 1 x 1 m grid overlay hen location tsQ) / CJ) s:: ~ +-' E I 0 50x50 m half-plot ~ E 0 LO ~ 100 along random compass bearing I HL i m ,,' transect 1 tsQ) m 2 sample-plot CJ) c quadrats ~ +-' ~o.25m2 E o I ~O.~~·~2 T"" ........... quadrats Itt o::lm~2 (not drawn to scale) RL transect N .j:'- \J1 ~ �246 In 1990, an additional the frequencies sampling of ungulate occurring at specific nesting. Between feces and characteristics places selected 3, this volume). season). The total number period, divided by 4 to determine 1st relocation was selected, sequences on the 36 locations season), between to another location used and distances additional and known season was relocations - HL transects) locations and walking (i.e., the moved by ptarmigan (Fig. 1-2). (Fig. 1-2). their mates' .. To territories, (random locations a 2nd - RL were determined by selecting 100 m along that compass bearing areas hens during the breeding yielding Two hens without radios to be paired with males holding territories season; were centered - RL transects), to sites) were also located several per selected by hens and random 1969): 10-m line transects (random locations transects per were centered (100 m was chosen on the basis of typical see Braun 1969 and Schmidt those locations were chosen systematically. Ten-m line transects of these transects compass bearing locations to yield a sample of 4 relocations was established Locations from 21 May on the number of relocations they could have used within set of 10-m transects a random repeatedly (see then every other 2nd, 3rd, or 4th (hen locations differences transects). facet of this study of selected depending hen (a total of 36 relocations). locations hens from each study by hen, during pre-nesting was always selected, hen during pre-nesting evaluate hens prior to The place of every relocation relocations of relocations, to assess of willow To obtain a sample of these hen-selected across the pre-nesting relocation for another Hens were relocated and 26 June 1990 (pre-nesting was recorded. by ptarmigan 23 and 29 May 1990, 3 ptarmigan site (12 hens) were radio-marked Chapter scheme was initiated on 36 (both banded in or adjacent times during the pre-nesting �247 season. Four transects (2 HL, 2 RL) for each of these hens were established in the same manner. To minimize observer Thus, 80 transects bias, all transects were established. were oriented north Each 10-m transect was defined with a meter tape during standardize placed placement on transect August, of quadrats, lines. sampling. center lines of quadrats Sampling was conducted to south. were 23 July to 10 1990. Occurrence of Ungulate Feces Elk fecal material is generally easy to identify, although and age class cause form, shape, and size of fecal material precluding positive sympatrically observed using identifications measured identification (Neff 1968). and Rocky Mountain bighorn where ungulate Mule deer (Odocoileus sheep (Ovis canadensis the sites on several occasions, questionable occurrences regard for species in some instances. of ungulate of origin. the 480 sample-plots fecal material Occurrences and within species occur hemionus hemionus) 'canadensis) were rendering fecal Therefore, were recorded of ungulate .. all without feces within the first and every other l-m2 along (five l-m2 samples per transect) were scored as present or absent (0). This method was used instead of pellet-group (Bennett et al. 1940) because breakup of pellet-groups. immediately diet to vary, transects plots To several alpine factors after inspection. counts contribute In 1989, feces were removed (1) to from samp~e· �248 Willow Characteristics Indicative In sample-plots, Nutt., S. planifolia was estimated intercepted measured percent On transects, a meter tape placed densities (S. brachycarpa (S. nivalis where willow on transect Hook.) (shrub or mat) lines, percent cover was live unbrowsed in a 0.01 m2 sub-sample. were estimated by mUltiplying terminal leaders the number and/or dead) by the average buds per leader type, and adding the products. If willow plot, but not in the lower right corner, (clockwise) transect, were examined (where willow Willow occurred, Characteristics With a cm ruler, live browsed Indicative lengths node when lengths samples, and the nearest adjoining in sample-plots, and/or dead terminal leaders were counted as described above). of in the samp1ecorners For each at 0.5 m intervals edges). (live unbrowsed; as the distance An average number by Ungulates leaders (in 0.01 m2 sub-samples, were measured Live browsed live browsed of Browsing of terminal leaders were missing). each type of leader Leader consecutive 5-cm from willow and/or dead) were measured tip of the woody portion occurred in a 0.01 m2 sub-sample beginning type The lower right 0.01 m2 until willow was encountered. buds were counted Bud of each leader vs. live browsed of the quadrat was used. and buds Leaders were counted. (live unbrowsed corner per 1-m2 transect. For each sample-plot, were measured Food Resources cover of shrub willow Pursh) and mat willow visually. as m/10-m of Available between leader the (or leader length was calculated as described above). only. leaders were counted on transects) (only (in 0.01 m2 sub- for .. �Height of shrub willow was measured-by placing a meter stick vertically into the shrub: height was measured as distance from substrate to tip of highest woody portion that was closest to the meter stick. For plots, 2-8 measurements were made per l-m2 sample- plot (2 measurements were made in corners of each 0.25 m2 quadrat section occupied by shrub willow) and then averaged. Average heights were weighted by percent cover of shrub willow in each sample-plot to minimize biases possibly caused by differences in percent of willow cover. For transects, heights of shrub willow were measured at 0.25 m intervals where willow intercepted the meter tape (beginning 5 cm and ending at least 5 cm from willow edges) and then averaged for each transect. A willow patch was defined by canopy separations of ~ 0.1 m (if canopies arose from different basal stems). Patch perimeter was estimated for plots (1990, only), and patch diameter was measured for transects. The perimeter (m) of any shrub-willow patch that fell completely or partially within a l-m2 sample-plot was approximated with a meter stick, and if more than 1 willow patch fell in the quadrat an average of patch perimeters was used. On transects, intercept (m) of shrub willow was measured for each patch and an average diameter per transect was calculated. It appeared that close association with conifers (Picea engelmanii Parry or Abies lasiocarpa Hook.) had major effects on willow morphology. To control for conifer effects, occurrence of conifer was recorded during sampling. -Sampled willows growing within or leeward (within I m) of conifers were scored as "1", those not growing within or leeward to conifers were scored as "0". �250 White-tailed Ptarmi~an: Samplin~ Ptarmigan were censused at each site throughout the breeding seasons of 1989 and 1990. Absolute counts of all ptarmigan, including pairs, single territorial males, and non-territorial males were attempted. Methods used to locate birds were: 1) methodically traversing sites while playing taped recordings of territorial male vocalizations (Braun et a1. 1973), which elicited responses from most territorial males, 2) locating ptarmigan during spot-mapping census visits (International Bird Census Committee 1970), and 3) following ptarmigan tracks in fresh snow, which allowed detection of territorial (paired or single) and non-territorial males that were unresponsive to the tape. Most birds were identified by their unique bandette color combination (Braun and Giesen 1990), minimizing the possibility of inflated counts. If birds appeared to be using an area fully encompassed by site boundaries, they were counted as 1 pair (or 1 bird if single). If, however, the birds appeared to use additional space outside the site boundaries, they were counted as 0.5 pairs (or 0.5 single birds), according to International Bird Census Committee standards. Plots and Tansects: Data Transformations and Analysis Occurrence of Ungulate Feces Frequency of occurrence of ungulate feces in sample-plots was calculated by summing the 60 presence/absence scores within each 100 x 50 m plot, by year; frequency of occurrence of ungulate feces on transects was calculated by summing the 20 (or 10 for the hens without radios) scores within each hen, by transect type (HL or RL). Plots and transects (hens) were subsequently assigned to a feces frequency .. �L51 category (FFC): an FFC of "B" (below) was assigned if the observed frequency was below average frequency (across plots or hens) and an FFC of "A" (above) was assigned if observed frequency was above average frequency. (Average frequencies were adjusted for uneven sample sizes; Le.; hens with 4 vs. 8 transects.) Frequencies of ungulate feces in plots were converted to percent values (frequency/60) for each plot (within site), by year. Frequencies of ungulate feces on transects were converted to percent values (frequency/20 or 10) for each hen (within site), by transect type. Square-roots of percent values were taken, arcsine-transformed (Snedecor and Cochran 1980), and then data sets were subjected to ANOVAs to test for differences among sites (PROC GLM, SAS 1988). Nested error terms for main effects were used to generate appropriate F-va1ues according to the scheme as follows: For PLOT data I used plot(site) for site effect, [site x year(site)] for year and for site x year effects; For TRANSECT data set I used hen(site) for site effect, [hen x transect type(site)] for transect type and for hen x transect type effects. Least square differences were used to determine whether effects were significant (P < 0.05). If site effect was significant, a contrast (NY vs. SE sites) was used to ascertain regional differences. Subsequently, appropriate FFCs were assigned to sites. Willow Characteristics Zero values resulting from no occurrence of willow in a sampleplot were treated as missing values (except for percent willow covers). Within plots, measurements of willow characteristics were averaged (percent cover was totaled) by site, year (not used for patch .. �252 perimeter), and conifer score, and were log-transformed [In(y+1) for percent cover of mat willow, 1n(y) for all other variables]. Within transects, by FFC and transect type (site was not used, due to missing data cells that would have resulted), willow measurements were averaged and log-transformed when necessary: [In(y+1) for percent willow covers and patch diameters; 1n(y) for heights]. Plot and transect data sets were subjected to ANOVAs to test for differences among sites or FFCs (PROC GLM, SAS 1988). Nested error terms for main effects were used to generate appropriate F-va1ues according to the scheme as follows: For PLOT data set I used p1ot(site) for site effect, [plot x conifer(site)] for conifer and site x conifer effects, [plot x year(site)] for year and for site x year effects; For TRANSECT data set I used hen(FFC) for FFC effect, [hen x transect type(FFC)] for transect type and for FFC x transect type effects. Least square differences were used to determine whether effects were significant (P < 0.05). of freedom were> If main effects were significant (and degrees 1), contrasts (NW vs. SE sites) were used to ascertain regional differences. Conifer was not used as a main effect in ANOVAs testing for differences in transect (hen) willow characteristics because transects had few occurrences of conifer (resulting in numerous missing data cells). Instead, willow characteristics (except percent cover) on transects were analyzed in 2 ways: 1) all measures combined, and 2) only those measures taken when conifer was absent. Ptarmigan Densities: Data Transformation and Analysis Site densities were converted to birds (pairs or single males)/20 ha site and to birds/40 ha region (within.year), and then '. �253 averaged over years. densities To ascertain of ptarmigan were statistically significant, combined, Separate I applied trends Cox-Stuart in breeding males) in RMNP tests for trends 1990, Braun et data (Braun and Giesen tests were conducted for MBC/GT observed (including pairs and territorial (Daniel 1978) to the population al. 1991). whether on data for all sites sites only, and for SB/TR sites only. A one- tailed test was used in each case. RESULTS Freguencies of Ungulate Frequencies Feces of ungulate feces were not homogenous for 1989 or 1990, or among transects In plots, average for 60 l-m2 samples was 37.5 in 1989 and 41.5 in 1990 (Fig. frequency 1-3). (hens). among plots Ungulate feces were more frequent SB and TR; observed 57 in 1990. than average in both plots at values ranged from 51 to 58 in 1989 and from 53 to Ungulate feces were less frequent than average in both plots at MBC and GT; observed values ranged from 13 to 37 in 1989 and from 18 to 38 in 1990. On transects, average samples was 25.2 (12.6 for 20 l-m2 samples) were more frequent values ranged frequent than average observed values along transects Ungulate feces at SB and TR; observed from 32 to 35 for hens with 8 transects Ungulate for 40 l-m2 (Fig. 1-4). than average along transects 20 for hens with 4 transects. transects). •frequency fecal material and from 18 to was less for all hens at MBC and GT; ranged from 14 to 25 (all MBC and GT hens had 8 An FFC value of "A" was, therefore, and an FFC of "B" was assigned to NW sites. assigned to SE sites, �254 MBC PLOT 1 MBC PLOT 2 iii 1989 [] 1990 ~ AVERAGE. ALL PLOTS GT PLOT 1 S8 PLOT 1 TR PLOT 1 . o 10 20 30 40 50 60 FREQUENCY Fig. 1-3. Observed frequencies of ungulate feces in 2 plots (60 possible per plot) at each of 4 sites in Rocky Mountain National Park, 1989-90. Average frequencies for all plots, by year, are also shown. �MBC/1 MBC/2 MBC/3 GT/4 z GT/5 w J: ..._ w e-n GT/6 I- SB/7 SB/8 .. TR/10 TA/001 I OBSERVED TA/002 ~ AVERAGE, ALL HENS o 10 20 30 40 FREQUENCY Fig. 1-4. Observed and average frequencies of ungulate feces on 8 (4 for hens 001 and 002) 10-m transects per hen (40 possible per hen; 20 possible for hens 001 and 002) at locations used by 11 pre-nesting white-tailed ptarmigan hens in Rocky Mountain National Park, 1990. �256 Site was a significant factor explaining variation in 3·, frequencies of ungulate feces (PLOTS: F ~ 27.44, P ~ 0.004, df TRANSECTS: F ~ 25.72, P 0.0004, df - 3). SE frequencies were significantly higher than NW frequencies (PLOTS: F - 78.02, P 0.0009, df - 1; TRANSECTS: F - 68.33, P 0.0001, df - 1). Although a year effect was not evident (P > 0.05), 6 of 8 plots had higher frequencies of ungulate feces in 1990 than in 1989, which may be biologically significant (see discussion). Willow Characteristics Indicative of Available Food Resources ANOVA results regarding analyses of all willow characteristics indicative of food resources are summarized in Table 1-1. In plots, neither shrub nor mat willow covers were different among sites, although year (F - 33.59, P - 0.0044, df - 1) and site x year (F 21.75, P 0.0061, df - 3) effects were significant for shrub willow (SE vs. NW contrast was not significant). On transects, percent cover of shrub and mat willow were greater at NW sites than at SE sites, but not significantly (P - 0.2091 and P - 0.2811, respectively). However, more shrub willow occurred on HL transects than on RL transects (F 7.22, P - 0.0312, df - 1). In plots, densities of live unbrowsed terminal leaders were greater at NW sites, although significance was not detected (P 0.2463). Densities were greater when associated with conifer (F 21.69, P - 0.0096, df - 1). There was also a significant site x conifer interaction (F ~ 22.42, P - 0.0058, df - 3): densities of live unbrowsed terminal leaders were greater at SE sites when associated with conifer and greater at NW sites when not associated with conifer (F - 47.15, P ~ 0.0024, df - 1). On transects, no main effects �L:JI Table 1-1. ANOVA results testing for site or FFC [above (A) versus below (B) average frequency of ungulate feces] effects on willow characteristics indicative important as food resources (percent shrub willow or mat willow cover, densities of live unbrowsed terminal leaders, bud densities) to white-tailed ptarmigan on Trail Ridge in Rocky Mountain National Park, 1989 and 1990. Medicine Bow Curve (MBC) and Gore Turnout (GT) were northwest (NY) sites (FFC-B), and Sundance Basin (SB) and Tombstone Ridge (TR) were southeast (SE) sites (FFC-A). VARIABLE Shrub Cover Mat Cover Live Unbrowsed Terminal Leaders Buds SOURCE df F P T Site Site Typeb 3 3 1 1.78 1.96 7.22 NS NS 0.031 P Site 3 4.78 0.083 T Site 3 1.56 NS P Site Conifer 3 1 2.07 21.69 NS 0.010 Site x Conifer 3 22.42 Contrast 0.006 0.002 pa T FFC FFC x Type 1 1 0.06 7.92 NS 0.048 P Site Site x Conifer 3 2.17 3 7.12 Contrast NS 0.044 0.026 FFC Type 1 1 0.30 0.01 NS NS FFC x Type 1 0.29 NS T LS MEANS higher at NY HL-O.ll RL-0.05 GT-O.OO MBC-5.17 SB-0.69 TR-1.36 higher at NY higher NY 0-1.56 1 -1.82 NYx 0-1.97 NYx 1-1.85 SE x 0-1.17 SE x 1-1.80 higher at B B x HL-3.12 B x RL-2.83 A x HL-2.30 A x RL-3.46 higher at NY NYx NYx SE x SE x 0-3.44 1-3.27 0-3.11 1-3.40 HLRLB x B x A x A x 25.67 49.00 HL-36.74 RL-32.25 HL-17.28 RL-61.84 P _ plots, T - transects bType - transect type [Hen location (HL) versus Random location (RL)] cPresence of Conifer - 1, Absence of Conifer - 0 a .' . �258 influenced measures densities of live unbrowsed combined. conifer However, (because missing significant - 0.0481, data cells would have resulted, in which case effects measurements FFC x transect densities In plots, bud densities SE sites, although an interaction densities densities occurring 0.0255). At SE sites, bud densities conifer, at NW sites bud densities present. Effects On transects, of transect transect influenced measurements a 7.92, P (F at at NW sites. at NW sites than at However, there was df - 3), with with conifer with conifer and (F - 12.06, P - in the presence of were higher when conifer was not were not different type (F - 8.90, P 0.0406, (F - 10.48, but the effects could on RL transects on HL transects were higher bud densities type interactions significant, occurring at SE sites when associated at NW sites when not associated by of conifer had to (F - 7.12, P - 0.0441, greater conifer was detected was not detected. of site x conifer influenced by conifer), tended to be greater significance being greater influenced type interaction df - 1), with greater SE sites and greater leaders with all after removing measurements not be used as a main effect, be removed by deleting terminal df - 1) and FFC x P - 0.0318, df - 1) were disappeared were removed among FFCs. completely (P - when conifer- 0.9266 and 0.6174, respectively). Willow Characteristics ANOVA indicative results sites. However, of conifer regarding of browsing plots, perimeters Indicative analyses by ungulates of shrub-willow perimeters (F ~ 10.21, of Browsing P = of all willow are summarized characteristics in Table 1-2. In patches were not different were significantly 0.0331, by Ungulates df ~ 1). greater among in the presence On transects, all main .. �Table 1-2. ANOVA results testing for site or FFC [above (A) versus below (B) average frequency of ungulate feces] effects on willow characteristics indicative of browsing by ungulates (shrub willow patch perimeter or diameter, shrub willow heights, densities of live browsed and/or dead terminal leaders, lengths of terminal leaders) on Trail Ridge in Rocky Mountain National Park, 1989 and 1990. Medicine Bow Curve (MBC) and Gore Turnout (GT) were northwest (NW) sites (FFC-L), and Sundance Basin (SB) and Tombstone Ridge (TR) were southeast (SE) sites (FFC-H). VARIABLE SOURCE df Patch Perimeter pa Site Coniferb 3 1 0.85 10.21 NS 0.033 FFC 1 9.59 0.017 Typec 1 43.73 0.003 FFC x Type 1 56.62 0.002 Site Conifer 3 1 0.27 15.73 NS 0.017 FFC FFC x Type 1 1 0.06 7.92 NS 0.048 Length Live P Unbrowsed Terminal Leaders Site 3 7.67 Contrast 0.039 0.010 Live Browsed/ Dead Terminal Leaders P Site 3 0.77 NS Live Browsed T Terminal Leaders Site 3 0.30 NS Length Live Browsed and/or Dead Terminal Leaders Site 3 3.30 NS Diameter Height T P T P F P LS MEANS 0-2.77 1-3.39 B-0.67 A-l.32 HL-0.48 RL-1.52 B x HL-0.75 B x RL-0.60 Ax HL-O.21 Ax RL-2.43 0-3.09 1-3.73 higher at B B x HL-3.12 B x RL-2.83 Ax HL-2.30 Ax RL-3.46 NW-0.87 SE-1.50 P ~ plots, T - transects Presence of Conifer - 1, Absence of Conifer - 0 cType - transect type [Hen location (HL) versus Random location (RL)] a b ~ .. �260 effects influenced patch diameters. Patch diameters were greater at SE sites (F - 9.59, P - 0.0174, df 1) and greater on RL transect types (F - 43.73, P - 0.0027, df 1); FFC x transect type interactions were significant (F 56.62, P - 0.0017, df - 1): diameters were greatest on RL transects at SE sites and greatest on HL transects at NW sites. Also, extreme high and low values occurred on RL and HL transects (respectively) at SE sites; values observed for HL and RL transects at NW sites were intermediate and similar to each other. In plots, willow heights were similar among sites. However, willow height was taller- in the presence of conifer (F - 15.73, P 0.0166, df - 1). On transects, willow heights were not different among FFCs, although they tended to be shorter at SE sites. However, an FFC x transect type interaction was detected when measurements influenced by conifer were removed (F - 7.92, P - 0.0481), with height being greater on HL transects at SE sites and greater on RL transects at .NW sites. Densities of live browsed and/or dead terminal leaders (plots) and densities of live browsed terminal leaders (transects) did not differ among sites or between FFCs. Also, effects of conifer, transect type, and interactions of main effects were not significant. In plots, lengths of live unbrowsed terminal leaders were different among sites (F - 7.67, P - 0.0391, df - 3), with lengths being greater at SE sites than at NW sites [1.50 versus 0.87 em, (In-transformed) respectively; F - 21.89, P - 0.0095, df - 1]. Lengths of live browsed and/or dead terminal leaders were not different among sites. �261 White-tailed Ptarmigan Densities Among Sites The ptarmigan census was believed to be complete. were highest at NW sites where FFC was "B". Densities What contributed to higher densities at NW sites were not only more pairs, but more single territorial and non-territorial males. The 2-year average number of pairs/20 ha for MBC and GT sites was 2.50 and 2.90, respectively, versus 1.59 and 0.79 at SB and TR sites (Fig. 1-5). Single territorial males/20 ha numbered 0.68 and 0.29 at MBC and GT sites, respectively, versus 0.00 and 0.26 at SB and TR sites. Single non- territorial males/20 ha numbered 0.23 and 0.29 at MBC and GT sites, respectively, and 0.00 at both SE sites. Ptarmigan density/40 ha during the breeding seasons (averaged over 1989 and 1990) was 12.29 at NW sites and 4.84 at SE sites. The Cox-Stuart test revealed a highly significant downward trend in ptarmigan breeding densities across all sites [P (K ~ 0 1-12, 0.50) - 0.0002] between 1966 and 1989 (1990 densities not published yet). Interestingly, downward trends were significant at MBC/GT sites (P {K ~ 1 I 12, 0.50} - 0.0031) and at SB/TR sites [P (K ~ 0 I 11, 0.50) 0.0005] (1966 through 1990) (data extrapolated from Braun et a1. 1991). DISCUSSION Frequencies of ungulate feces were unquestionably greater at SE sites where ptarmigan densities had declined most consistently (Braun et al. 1991 explained that annual variations in breeding densities of ptarmigan were lower at SE sites than they were at NW sites, indicating a more steady downward trend at SE sites). As a relative index of elk presence and habitat use among sites, feces frequency �262 3.5 II PAIRS 3 ILl NON-TERRITORIAL MALES D SINGLE TERRITORIAL MALES 2.5 co .c: 0 2 C\I ~ en z w 1.5 Cl .. 1 0.5 o MBC GT 5B TR Fig. 1-5. Densities of white-tailed ptarmigan pairs, single territorial males, and single non-territorial males at Medicine Bow Curve (MBC), Gore Turnout (GT), Sundance Basin (SB), and Tombstone Ridge (TR) in Rocky Mountain National Park, 1989 and 1990 breeding seasons. Densities were calculated from a two-year average of complete counts. �263 data were useful: either more adult elk used SE sites, or similar numbers of elk used all sites but spent more time at SE sites. Transect data confirmed that frequency results in plots were not artifacts of bias resulting from observer choice of plots. It was surprising to find no differences in frequencies of ungulate feces on HL versus RL transects. Assuming that HL transects represented areas preferred by hens on a small spatial scale (within territory), ptarmigan did not seem to avoid areas used heavily by ungulates. However, on a larger scale, (sites) it may be that ptarmigan did avoid areas (SE sites) used heavily by ungulates. Frequencies of ungulate feces also represented presence of mule deer or bighorn sheep. Size, shape, and form generally allowed positive identification of fecal material originating from elk, but there were exceptions. In general, however, ungulate feces represented presence of elk. Although a significant year effect was not detected, frequencies of ungulate feces tended to be higher in 1990 than in 1989. Freezing and desiccation, 2 prevalent conditions in Colorado alpine systems, retard decomposition rates of biotic material (Marr 1967). As a consequence, ungulate feces counted in 1989 probably represented many years of deposition. However, because feces were removed from plots in 1989, feces observed in plots during 1990 were less than 1 year old (except where older feces may have blown into sample-plots between years). Therefore, higher frequencies of ungulate feces in 1990 may be biologically significant, even if not statistically significant. Ungulates presence in the alpine may be increasing. Interestingly, higher frequencies were mostly attributable to increases at NY sites. If presence of ungulates is increasing, it may be increasing at NY �264 sites faster than at SE sites, which could be predicted if food resources at SE sites have been diminished. Feces data from transects, which represented cushion fellfields, turf, meadows, snow accumulation areas, and areas below treeline, as well as willow(krummholz, indicated that ungulates occurred in all habitat types. Neff (1968) pointed out that ungulates tend to defecate where they rest and ruminate. In other words, high frequencies of ungulate feces in willow/krummholz habitat do not necessarily indicate that ungulates foraged in that habitat. However, elk were observed using willows for foraging in alpine/krummholz habitat throughout both field seasons. Elk cows with calves often were observed foraging or resting in or near GT plots where small ungulate feces, probably originating from calves, contributed to high frequencies of ungulate feces. While feces data did not distinguish among age classes of elk, it was apparent that elk of all ages used TRR sites. Holmgren and Hutchings (1972) and Clements (1990) found that shrub production and vigor declined when shrubs were browsed during periods of vegetative growth as opposed to being browsed during periods of dormancy (Clements examined a willow spp. in a montaneriparian habitat in northern Colorado). In alpine regions of RMNP, where low moisture availability and brief growing seasons constrain net primary production, it is likely that willow vigor and production would be especially poor under heavy browsing pressure during periods of growth. occurred. forms. Fecal form (pellets vs. chips) can indicate when browsing Presence/absence data did not distinguish between fecal It was clear, however, that pellets were considerably more prevalent than chips (after accounting for total volume/defecation) .. �throughout all sites, indicating that elk used sites primarily during periods of vegetation dormancy. In addition, 6 elk antlers were found at MBC in early spring 1990, confirming that bull elk were present during winter 1989-1990. Elk chips were encountered at all sites, indicating elk used sites during periods of vegetation growth, as well. I observed adult elk frequently foraging on willow at NY sites throughout spring and summer during 1989 and 1990; to be used similarly, but to a lesser extent. SE sites appeared Because browsing x season interactions can have profoundly different effects on shrub vigor, future studies of elk/willow relationships should be designed to evaluate levels of browsing among sites during each season. If breeding densities of ptarmigan along TRR in RMNP were limited by quantities of willow-food resources, then densities should have been lowest where willow foods were lowest. Also, significantly more shrub-willow cover should have occurred in locations chosen by hens than in random locations within their territories, which was observed. While no variables indicative of available willow-food resources differed among sites, cover of mat and shrub willow (on transects), densities of live unbrowsed terminal leaders, and densities of buds all tended to be greater at NY sites where ptarmigan densities were greater. Several factors may have precluded finding statistically significant differences. First, plots were chosen specifically to include willow, and it may be that plots did not adequately represent coverage of willow (scale too small). Second, transects were chosen on the basis of hen choices, which presumably would include some willow, and it may be that there is a lower threshold of willow cover with which they associate (see discussion in Chapter 3, this volume). Again, a larger-scale approach (measuring �266 willow cover over a whole site) may have yielded significant site effects. Third, it is possible that percent willow cover had been greater at SE sites prior to increases in RMNP elk populations, but was diminished over time so that differences between SE and NW sites no longer existed. Lack of historical data precludes making inferences concerning declines in percent cover of willow, although Braun et al. (1991) did find long-term declines of willow cover on transects in various sub-alpine and alpine regions of RMNP. Effects of year and site x year interactions on willow cover in plots were probably a result of observer differences between years, only. Fourth, if differences in willow cover did exist, then largescale differences in absolute numbers of buds and live unbrowsed leaders probably existed. Densities may be relatively similar from shrub to shrub (small scale). Fifth, it may be that there were no real differences in food resources among sites. Elk may have browsed equally at all sites. Significant downward trends in breeding densities of ptarmigan were detected for both SE and NW regions, and I observed groups of elk foraging on willow at NW sites frequently. Therefore, further studies should be designed to compare willows along TRR with willows in regions of Colorado that are relatively unused by browsing ungulates. Interestingly, live unbrowsed leader densities and bud densities were greater at SE sites (in plots) when associated closely with conifers. Conifers may have provided refuge for willows by diminishing willow "apparency" to browsing elk. The fact that leader and bud densities at SE sites were greater when associated with conifer supports this possibility. The FFC x transect type .. �267 interaction indicated that appeared when effects that SE hens selected lower densities selected This could happen unavailable see Chapter Willow browsing on willow shrubs while willow patches and heights patches and height. of leaders were as indicators to foraging When pre-nesting on the substrate of willow that patch perimeters have suggested dynamics that animals choose (Kamil et al. 1987). in terms of perceived are similar, but patch size at site B is larger, causing the other hand, numerous ptarmigan ptarmigan. reach. between efficiency. and foraging could contribute to ptarmigan sites A and B would be lower at site B. small and scattered and movement foraging food abundance abundance. ratios to be smaller at site B than at site A, then food abundance more searching foraging However, or "apparent" if areas of willow cover at hypothetical "perceived" or diameters of ptarmigan. For example, perimeter:area hens (i.e., they did not enter to foraging food abundance of ptarmigan, ptarmigan they fed along the perimeters theorists food must be measured leader densities). important could also be important could be important with greater (whereas NW hens densities potentially to feed), indicating Optimal-foraging leaders that had (too tall for hens to reach from the shrubs, standing shrubs 3, this volume). such as patch dimension foraged with higher characteristics by ungulates terminal shrubs that had higher if willows to the birds substrate; areas with willow of live unbrowsed areas with willow of conifer were removed patches, The same principle possibly require reducing If patch size does affect perceived efficiency, to localized patches might declines then sub-optimal in breeding patch size densities would apply to willow of too tall to On �268 Variations may indicate ptarmigan assumed in patch size were more extreme heavy browsing foraging at those sites. to be atypical. However, willows moderately-sized willow patches possibly see Bryant 1981) or die-back because territories However, ( in response appeared diameters to heavy browsing. to be "mowed" of willow patches diameters FFC x transect pre-nesting It appears diameters, not the extreme indicated of browsing diameters transformed), were on HL of willow patches at choice by of moderate at SE sites. did not differ among sites, they may be by ungulates, although variations I suggest that real [heights on HL at NW sites were 3.12 and 2.83 cm (In- respectively, and heights on HL versus RL transects SE sites were 2.30 and 3.46 cm (In-transformed), sites, willows on transects that willow-patch observed were masked by site-specific versus RL transects finding for all HL transects. that hens prefer patches willow heights poor indicators did not differ hens may depend on local average patch dimensions. differences patch diameters type interactions ptarmigan Because (severely hedged) were smaller across all sites, and average diameters NW sites more closely matched tended to > 30 m or < 0.05 m choice may have precluded Interestingly, in many (e.g., 2-4 m diameter). In plots, willow patch perimeter at SE sites. transects of willows were (via growth of new basal shoots; occurred among sites, and again observer greater patches for in patch dimensions in ptarmigan spreading In some cases, willow patches differences. Extremes at SE sites were extreme diameter), into tiny patches. and have consequences Based on my observations alpine areas of Colorado, grow in discrete by ungulates at SE sites, which were barely respectively]. taller than cushion plants areas, but many grew quite tall in snow accumulation at At SE in exposed areas where they .. �would receive abundant moisture, and where they would be protected from browsing and weather a large part of the year. The significant FFC x transect type interaction (with effects of conifer removed) indicated that hens chose sites with shorter willows at SE sites and sites with taller willows at NY sites. There is probably a maximum willow height suitable for ptarmigan, which may have been exceeded at SE sites more than at NY sites (willows on RL transects at SE sites were considerably taller than willows on HL transects at SE and NY sites or willows on RL transects at NY sites). Of the remaining willow characteristics believed to be important indicators of browsing by ungulates, only lengths of live unbrowsed terminal leaders were different among sites. SE sites. This result was unexpected. Lengths were greater at Assuming that more browsing on willows by elk had occurred at SE sites, and assuming that willow vigor at those sites would be poor, I expected all leaders to be shorter at SE sites. Interestingly, a large portion of live unbrowsed terminal leaders at SE sites were primary (new basal shoots), while most at NY sites were secondary, tertiary, etc. Bryant (1981) found that browsing by snowshoe hares (Lepus americanus) stimulated development of adventitious basal shoots in several tree/shrub species, including willows. If more browsing at SE sites resulted in removal of most terminal leaders (and terminal buds), then significant reductions in levels of auxin hormone (normally released by terminal meristems and which supresses development of new shoots) could have allowed development of unusually long leaders from any remaining buds. There are numerous alternative hypotheses that could explain declines in ptarmigan densities. One possibility is that browsing by elk elevated levels of secondary plant compounds (phenolic glycosides �270 and other compounds) in willows. Ptarmigan may avoid willows with elevated levels of toxins or exhibit lowered fitness by feeding on more toxic willows. Numerous studies have shown that herbivores avoid plants containing high levels of toxic compounds (Freeland and Janzen 1974, Bryant and Kuropat 1980, Bryant 1981, Coley et al. 1985, Clausen et al. 1989, and Remington 1990), and Bryant (1981) found that browsing stimulated growth of adventitious shoots that were heavily defended with elevated levels of secondary compounds. When blue grouse (Dendragapus obscurus) were fed unpalatable plants (assumed to have higher toxin levels), excretions of nitrogen per gram of ingested food were higher (Remington 1990), indicating that detoxifying secondary plant compounds was metabolically expensive. If severe and chronic browsing stimulated elevated levels of secondary compounds in willows, and if those compounds required ptarmigan to use significantly more nutrients to detoxify the compounds, then decreased ptarmigan survival and/or reproductive fitness could occur as a result. The combined effect of decreased food resources and greater toxicity could have synergistic effects. Breeding densities of ptarmigan fluctuated at fairly regular intervals until about 1979, indicating that ptarmigan densities may have cycled by virtue of intrinsic mechanisms (Wynne-Edwards 1964, Chitty 1967, Moss 1971, Moss et al. 1984). If densities were low in the late 1970s'due to such a cycle, but were "slammed" down further by other factors such as severe climatic events, increases in predator populations that followed increases in elk populations, and/or diseases, then a population low could have been deeper and more prolonged. Ptarmigan densities have increased slightly during the �LIL past two years, increase. indicating Time may be all that is needed In conclusion, Rocky Mountain restricted greater are positively In addition, of Trail Ridge. correlated significant abundance are prevalent However, of willows long-term establish corroborations that warrant my recommendation On the premise in eastern I believe further to the National regarding ungulate with greatest overlap LITERATURE occurrence resources of used by have declined most, pressure most. results are necessary to that my results present strong In addition, it is that current management be reviewed very critically. (Odocoileus virginianus) densities Parks are causing ecological densities and have declined to those I have studied here, Warren that ungulate is not densities Park Service in of elk and ptarmigan investigation. populations that high ungulate National on Trail Ridge that indicate heavy browsing data and experimental cause and effect. this possibility. More consistent of willow-food at sites where ptarmigan to an and the decline tend to be lower at sites where numbers and characteristics similar densities Park are significant, feces, and indicate ptarmigan may be yielding to support in ptarmigan to just one portion habitat-use. policies declines National declines ungulate that the decrease be controlled problems, problems (1991) has recommended artificially in National Parks. CITED Baker, D. L., and N. T. Hobbs. 1982. of elk summer diets in Colorado. 46:694-703. Composition and quality J. Wildl. Manage. Bear, G. D. 1989. Seasonal distribution and population characteristics of elk in Estes Valley, Colorado. Wi1d1. Spec. Rep. 65. l4pp. Colo. Div. .. �272 Bear, G. D., G. C. White, L. H. Carpenter, R. B. Gill, and D. J. Essex. 1989. Evaluation of aerial mark-resighting estimates of elk populations. J. Wildl. Manage. 53:908915. Bennett, L. J., P. F. English, and R. McCain. 1940. A study of deer populations by use of pellet-group counts. J. Wildl. Manage. 4:398-403. Billings, W. D. 1979. Alpine ecosystems of western North America. Pages 7-21 in D.A. Johnson, ed. Special management needs of alpine ecosystems: Range Science Series No.5. Society for Range Management, Denver, Co. Braun, C. E. 1969. Population dynamics, habitat, and movements of white-tailed ptarmigan in Colorado. Ph.D. Diss., Colorado State Univ., Fort Collins. 189pp. _____ , and G. E. Rogers. 1971. White-tailed ptarmigan in Colorado. Colorado Div. Game, Fish and Parks Tech. Publ. 27. 80pp. _____ , R. K. Schmidt, Jr., and G. E. Rogers. 1973. Census of Colorado white-tailed ptarmigan with tape-recorded calls. J. Wild1. Manage. 7:90-93. _____ , R. W. Hoffman, and G. E. Rogers. 1976. Wintering areas and winter ecology of white-tailed ptarmigan in Colorado. Colorado Div. Wildl. Spec. Rep. 38. 38pp. _____ , and K. M. Giesen. 1990. Population dynamics of whitetailed ptarmigan, Job Progress Rep., Colorado Fed. Aid Project W-152-R. Apr. 1990. 264pp. _____ , D. R. Stevens, K. M. Giesen, and C. P. Melcher. 1991. Elk, white-tailed ptarmigan, and willow relationships: a management dilemma in Rocky Mountain National Park. Trans. 56rh N. Am. Wi1d1. and Nat. Res. Conf., Wildl. Manage. Institute. Bryant, J. P. 1981. Phytochemical deterrence of snowshoe hare browsing by adventitious shoots of four Alaskan trees. Science 213:889-890. _____ , and P. J. Kuropat. 1980. Selection of winter forage by subarctic browsing vertebrates: the role of plant secondary chemistry. Ann. Rev. Ecol. Syst. 11:261-285. Butler, D. C. 1979. Effects of a high-density population of ungulates on breeding bird communities in deciduous forest. M.S. Thesis, Colorado State Univ., Fort Collins. 67pp. Canfield, R. H. 1941. Application of the line interception method in sampling range vegetation. J. For. 39:388-394. .. �L/J Chitty, D. 1967. The natural selection of self-regulating behaviour in animal populations. Proc. Ecol. Soc. Australia 2:51-78. Clausen, T. P., P. B. Reichardt, J. P. Bryant, R. A. Werner, K. Post, and K. Frisby. 1989. Chemical model for short-term induction in quaking aspen (Populus ~remuloides) foliage against herbivores. J. Chem. Eco1. 15:2335-2346. Clements, A. 1990. Seasonal defoliation effects on two important montane riparian species. M.S. Thesis, Colorado State University, Fort Collins. 100pp. Coley, P. D., J. P. Bryant, and F. S. Chapin. 1985. availability and plant anti-herbivore defense. 230:895-899. Resource Science Daniel, W. W. 1978. Applied nonparametric statistics. Houghton Mifflin Co., Boston, MA. 503pp. Freeland, W. J., and D. H. Janzen. 1974. Strategies in herbivory by mammals: the role of plant secondary compounds. Am. Nat. 108:269-289. Giesen, K. M., C. E. Braun, and T. A. May. 1980. Reproduction and nest-site selection by white-tailed ptarmigan in Colorado. Wilson Bull. 92:188-199. Green, R. A. 1982. Elk habitat selection and activity patterns in Rocky Mountain National Park, Colorado. M.S. Thesis, Colorado State Univ., Fort Collins. 165pp. Hobbs, N. T. 1979. Winter diet quality and nutritional status of elk in the upper montane zone, Colorado. Ph.D. Diss., Colorado State Univ., Fort Collins. l3lpp. __ , D. L. Baker, J. E. Ellis, and D. M. Swift. 1981. Composition and quality of elk winter diets in Colorado. J. Wildl. Manage. 45:156-171. _____ , , , , and R. A. Green. 1982. Energy and nitrogen-based estimates of elk winter-range carrying capacity. J. Wildl. Manage. 46:12-21. Holmgren, R. C., and S. S. Hutchings. 1972. Salt desert shrub response to grazing use. Pages 152-164 in C.M. McKell, J.P. Blaisdell, and J.R. Goodin (eds.). Wildland shrubs: their biology and utilization. USDA For. Servo Tech. Rep. INT-1. 494pp. International Bird Census Committee. 1970. Recommendations for an international standard for a mapping method in bird census work. Audubon Field Notes 24:722-726 .. �274 Jackson, S. 1991. Changes to bird communities from grazing by wild ungulates. Proc. Ecol. Soc. Am. Kamil, A. C., J. R. Krebs, and H. R. Pulliam, eds. 1987. Foraging Behavior. Plenum Press, New York. 676pp. Knopf, F. L., J. A. Sedgewick, and R. W. Cannon. 1988. Guild structure of a riparian avifauna relative to seasonal cattle grazing. J~ Wildl. Manage. 52:280-290. Kufeld, R. C. 1973. Foods eaten by the Rocky Mountain elk. J. Range Manage. 26:106-112. Marr, J. W. 1967. Ecosystems of the east slope of the Front Range in Colorado. Univ. Colorado Ser. BioI. 8. l34pp. May, T. A. 1975. Physiological ecology of white-tailed ptarmigan in Colorado. Ph.D. Diss., Univ. Colorado, Boulder. 3llpp. _____ , and C. E. Braun. 1972. Seasonal foods of adult white-tailed ptarmigan in Colorado. J. Wildl. Manage. 36:1180-1186. Moss, R., A. Watson, and P. Rothery. 1984. Inherent changes in the body size, viability, and behavior of a fluctuating red grouse (Lagopus lsgopus scoticus) population. J. Anim. Ecol. 53:171-189. Mossop, H. D. 1971. A relation between aggressive behavior and population dynamics in blue grouse. M.S. Thesis, Univ. British Columbia, Canada. l18pp. Murie, O. J. 1951. The elk of North America. Harrisburg, PA. 376pp. Stackpole Co., Myrberget, S. 1987. The effects of sheep grazing on a willow grouse breeding habitat. Fauna 40:144-149. National Park Service. 1975. Resource management plan: Rocky Mountain National Park and Shadow Mountain National Recreation Area. U.S. Dept. Inter., Natl. Park Serv., Estes Park, CO. 88pp. 1988. Statement for management. U.S. Dep. Inter. Natl. Park Servo NPS D-23d. Washington, D.C. 60pp. Neff, D. J. 1968. The pellet-group count technique for big game trend, census, and distribution: a review. J. Wildl. Manage. 32:597-614. Olmsted, C. E. 1979. The ecology of aspen with reference to utilization by large herbivores in Rocky Mountain National Park. Pages 87-89 in M. S. Boyce, and L. D. Hayden-Wing, eds. North American elk: ecology, behavior, and management. Univ. Wyoming, Laramie. .. �L/~ Packard, F. M. 1947. A study of deer and elk herds of Rocky Mountain National Park. J. Mammal. 28:4-12. Ratcliff, H. M. 1941. Winter range conditions in Rocky Mountain National Park. Trans. No. Am. Wildl. Conf. 6:132-139. Remington, T. E. 1990. Food selection and nutritional ecology of blue grouse during winter. Ph.D. Diss., Univ. of Wisconsin, Madison. l16pp. Ryder, R. A. 1980. Effects of grazing on bird habitats. Pages 51-64 in Management of western forests and grasslands for non-game birds. U.S. Dep. Agric., For. Servo Gen. Tech. Rep. INT-86. SAS Institute, Inc .. 1988. SAS/STATe Procedures guide: release 6.03 edition. Cary, NC. 1029pp. Schmidt, R. K., Jr. 1969. Behavior of white-tailed ptarmigan in Colorado. M.S. Thesis, Colorado State Univ., Fort Collins. l74pp. 1988. Behavior of white-tailed ptarmigan. Pages 270299 in Bergerud, A. T., and M. W. Gratson, eds. Adaptive strategies and population ecology of northern grouse. University of Minnesota Press, Minneapolis. Snedecor, G. W., and W. G. Cochran. 1980. Statistical methods, 7th ed. Iowa State University Press, Ames. 507pp. Stevens, D. R. 1980a. The deer and elk of Rocky Mountain National Park: a ten-year study. U.S. Dep. Inter., Nat1. Park Servo Rep. ROMO-N-13, Estes Park, CO. 163pp. 1980b. Effect of elk on alpine tundra in Rocky Mountain National Park. Pages 228-241 in C.L. Jackson and M.N. Schuster, eds. Proc. High Altitude Revegetation Workshop 4. Colorado State Univ. Infor. Series 42. Swift, L. W. 1940. A partial history of the elk herds of Colorado. J. Mammal. 26:114-119. Taylor, D. M. 1986. Effects of cattle grazing on passerine birds nesting in riparian habitat. J. Range Manage. 39:254-258. Thomas, J. W. and D. E. Toweill. 1982. Elk of North America: Ecology and Management. Stackpole Books, Harrisburg, PA. 698pp. Tucker, T. L. 1987. Cattle grazing affects nongame wildlife populations and fish habitat in a montane riparian area. M.S. Thesis, Colorado State Univ., Fort Collins. 84pp. �276 Warren, R. J. 1991. Ecological justification for controlling deer populations in eastern National Parks. Trans. 56th N. Am. Wildl. and Nat. Res. Conf., Wildl. Manage. Institute. Watson, A., and P. J. O'Hare. 1979. Red grouse populations on experimentally treated and untreated Irish bog. J. Appl. Ecol. 16:433-542. Weber, W. A. 1976. Rocky Mountain Flora. Univ. Press, Boulder. 479pp. Colorado Assoc. Willard, B. E., and J. W. Marr. 1970. Effects of human activities on alpine tundra ecosystems in Rocky Mountain National Park, Colorado. Biol. Conserve 2:257-265. Wynne-Edwards, V. C. 1964. Sci. Am. 211:68-75. Population control in animals. Zwickel, F. C. 1972. Some effects of grazing on blue grouse during summer. J. Wildl~ Manage. 36:631-634. �L. I I CHAPTER 2 AVIAN SPECIES RICHNESS AND ABUNDANCE IN ALPINEjKRUMMHOLZ HABITATS USED EXTENSIVELY BY WILD UNGULATES INTRODUCTION Avian diversity can be altered and/or diminished when vegetation is severely and persistently damaged by grazing/browsing ungulates (Zwickel 1972, Butler 1979, Ryder 1980, Taylor 1986, Myberget 1987, Tucker 1987, Knopf et al. 1988, Jackson 1991). Although domestic livestock have been the focus of most ungulate-grazing/avifaunaresponse studies, high densities of wild ungulates have been implicated in at least two studies. Butler (1979) found that high densities of white-tailed deer (Odocoileus virginianus) using a small wildlife preserve in an eastern U.S. deciduous forest had eliminated the forest understory and its associated guild of ground-nesting birds. Jackson (1991) found that highest levels of browsing by elk (Cervus elaphus) and moose (Alces alces) were correlated with lowest densities of passerine birds in montane willow (Salix spp.) riparian systems in Wyoming. In Rocky Mountain National Park (RMNP) , situated within Colorado's Front Range, elk populations were controlled artificially from the mid-1940s until the late 1960s. However, in 1968, the National Park Service adopted a new policy that discontinued management of wild-animal populations. Discontinued management and .. �278 several other increase factors probably of the elk population concentrate in RMNP see Chapter 1, this volume). surrounding areas from browsing Olmsted (1979) concluded impaired Damage densities (Braun and Giesen alpine/krummholz other species 1966, Pattie a period when elk (pers. comm.) observed severely damaged that breeding to declines avifaunal communities habitats; 4-5 breeding occur less frequently and Verbeek that and densities in willow of cover (see vertical richness (1970) demonstrated in part due to low moisture are common growing by elk could significantly 1965, Johnson 1988). structure in Colorado's (human foot-traffic) as the alpine and a few Severe of plant alpine tundra. that growth of alpine vegetation availability, and very brief climate-constrained species in Rocky Mountain (Bailey and Niedrach composition/limited from perturbations debris, are simple 1966, Braun 1980, Kingery limit avian species and Marr and recovery browsing 1980b) (Lagopus leucurus altipitens) had declined 1988) in response and simple communities driven (1980a, 1. this volume). Typically. Willard Stevens during in RMNP were becoming ptarmigan of aspen (Salix planifolia Pursh, S. shrubs Braun and Giesen shrubs by the late 1970s. 1976) in RMNP was seriously by elk. by the early 1980s, and suspected white-tailed climate after Weber et al. 1991; in RMNP and and regeneration in RMNP had declined increased. alpine willow to vegetation that vigor rapid of those elk to by elk was observed cover of willow brachycarpa Nutt.) Chapter and the tendency by high levels of browsing found that canopy to a subsequent (Bear 1989, Bear et al. 1989, Braun (Populous tremuloides Michx.; decadent contributed shearing seasons. tundra, severe in RMNP effects In a system is slow, of windas and persistent alter associations and .. �279 morphologies of vegetation and, consequently, the structures of avian communities. Long-term monitoring can be useful for detecting community-level changes. First, however, baseline data must be established. The objective of this study was to establish baseline data by documenting current species richness and abundance of breeding birds in alpinejkrummholz habitats in RMNP. In addition, I recorded avian behaviors and habitat use that could be relevant with respect to changes in vegetation, and discuss the possible consequences of those changes for species using alpinejkrummholz habitats. STUDY SITES Four study sites were selected near Trail Ridge Road (TRR) in RMNP for comparisons of local avian communities: Medicine Bow Curve (MBC) , Gore Turnout (GT) , Sundance Basin (SB), and Tombstone Ridge (TR). The criterium for site selection was presence of relatively similar proportions of alpinejkrummholz-habitat types. Study area boundaries were determined by using compass bearings taken from distinctive geomorphic and topographic features. Boundaries were plotted on U.S. Geological Survey maps (scale 1:24000), and site areas were determined by using a planimeter. Approximate areas were: MBC 22 ha, GT - 17 ha, SB - 20 ha, TR ~ 19 ha. Elevations ranged from 3475 to 3660 m. Climate and community associations typical of Front Range alpinejkrummholz communities were described by Marr (1967). Vegetation associations found within my study sites were mapped and described by Braun (1969). Briefly, community types (after Braun 1969) and plant species (after Weber 1976) common throughout all sites �280 were: Trifolium and/or Dryas-Salix Kobresia, cushion fellfield; Dryas octopetala (S. nivalis Carex, Trifolium, hookeriana Juz. Hook.) turf; alpine meadow comprised of Poa, and/or Acomastylis rossii R. Br. (formerly Geum rossii R. Br.); snow accumulation areas comprised of Artemisia, Carex, Poa, and/or Acomastylis; comprised of Salix spp. (S. planifolia Picea engelmanii and patches of krummholz Pursh, S. brachycarpa Parry, and/or Abies lasiocarpa Hook. Nutt.), Communities more typical of dry wind-swept areas (Dryas and Kobresia turf) dominated sites at MBC and GT, and communities more typical of moist snow-accumulation areas (Carex and Artemisia) and TR. dominated sites at SB Sites were described in greater detail in Chapter 1 (this volume). METHODS Birds were censused at each site throughout the breeding seasons of 1989 and 1990. Census methods for ptarmigan included: 1) methodically traversing sites while playing tape recordings of territorial-male vocalizations (Braun et al. 1973), which elicited responses from most territorial males, 2) locating ptarmigan during spot-mapping census visits (International Bird Census Committee 1970), and 3) following ptarmigan tracks in fresh snow, which allowed detection of unresponsive males. Most ptarmigan were identified by their unique color-bandette combinations (Braun and Giesen 1990), minimizing the possibility of inflated counts. If a ptarmigan pair appeared to be using an area fully encompassed by site boundaries, it was counted as 1 pair. If, however, the pair appeared to use additional space outside site boundaries, it was counted as 0.5 pair .. �LOL (0.5 bird for single males). Single males, territorial and non- territorial, were counted similarly. Spot-mapping procedures were used to census all other species. Deep snow early in the breeding season and only sporadic occurrence of vegetation tall enough to mark rendered grid-marking impractical. Instead, census routes were established by counting paces along compass bearings taken from distinct geomorphic features. Locations of singing males were mapped on graph paper while traversing the unmarked grid. An attempt was made to use different routes for each census visit to minimize bias related to time of day. Censuses began between 6:00 and 7:00 a.m., and generally took 3-4 hours to complete. Unless a species was observed at least 3 times within a site it was not included in calculations of species richness or densities; however, all species observed during census visits were recorded and I attempted to determine how each species was using each site. Species known to nest in alpine habitat were recorded as nesting within sites if nests with eggs or young, unfledged young, adults carrying food, and/or adults performing distraction displays during observer presence were observed. Also, species observed using sites repeatedly for foraging, but typically not known to nest in alpine habitat (and with no evidence of nesting observed), were included in the results. In 1989, each site was censused 8 times, with at least 3 days between visits at each site. weather. Censuses were conducted regardless of Fog, precipitation, and strong wind severely impeded visibility and/or hearing during censuses; also, defense of avian territories broke down during poor weather. Therefore, the first 3 censuses from 1989 (all conducted during severe weather) were dropped before results were calculated; the 5th 1989 census .at GT was dropped .. �282 for the same reason. through 10 July. Censuses in 1989 were conducted from 27 May After the 1989 census, it was determined that only 4 census visits were required to capture the highest counts, as long as they were conducted when weather was relatively clear and calm. Four census visits in 1990 were conducted at each site between 7 June and 1 August 1990, with at least 3 days between visits at each site. However, the last 1990 census from each site was dropped because breeding territories appeared to have broken down (birds were not singing or were not present) by then. For each census and species, densities were converted to pairs (and single male ptarmigan) per 20 ha site. Species densities within sites were averaged across census visits, for both breeding seasons. To determine species richness, numbers of species using sites for nesting and/or foraging were summed. .. RESULTS Overall avian species richness and density were greater at and TR than at MBC or GT, both years (Fig. 2-1). sa Species richness in 1989 and 1990 was 6.0 at MBC, 5.0 at GT, and 7.5 at both sa and TR. The average number of pairs (not including single ptarmigan males) across all sites was 32.5 in 1989 and 25.6 in 1990. Total pairs/20 ha, averaged across both years, was 23.2 at MBC, 29.4 at GT, 30.5 at sa, and 33.2 at TR. In 1989, densities of ptarmigan pairs at MBC and GT exceeded densities at either sa or TR (Table 2-1). numbers of single males and pairs. Also, MBC and GT had higher The total 1989 ptarmigan density for each site was 7.8 at MBC, 7.0 at GT, 2.0 at sa, and 1.5 at TR. �~ White-crowned Sparrow D White-tailed Ptarmigan Yellow-rumped Warbler (Z3 Wilson's Warbler 40 o Uncoln'sS trmmmmi • D Rock Wren Homed Lark American Pipit tSSl American Robin Mountain Bluebird 30 as s: 0 C\I -..... en 20 a:: « a. .. 10 o ~~~~~~~~~~~~~~~~~ 1989 1990 1989 1990 1989 1990 1989 1990 M8C GT S8 TR Fig. 2-1. Avian species composition and relative abundance at four sites (MBC=Medicine Bow Curve, GT~Gore Range Overlook, SB~ Sundance Basin, TR=Tombstone Ridge) in Rocky Mountain National Park, 1989 and 1990. Species using sites for nesting and/or foraging were included. �284 Table 2-1. Average densities (SE) of avian species breeding or foraging in alpine/krummholz habitat at 4 sites in Rocky Mountain National Park. Densities were ascertained during spot-map census visits during 1989 (n-5; n-4 at GT) and during 1990 (n-3). Pairs/20 ha Site GT SB MBC TR Species AMPIa Year 1989 1990 6.2 (0.68) 6.1 (0.53) 7.1 (0.96) 5.5 (1.80) 6.6 (0.55) 5.0 (1.00) 4.4 (0.88) 4.6 (0.61) AMROb 1989 1990 1.5 (0.50) 0.9 (0.00) 2.1 (0.60) 1.2 (0.00) 1.2 (0.84) 1.0 (0.00) 3.2 (0.74) 1.8 (0.61) HOlA 1989 1990 4.5 (1.29) 3.6 (0.91) 2.4 (0.96) 1.6 (0.70) 3.2 (0.84) 0.3 (0.58) 3.6 (2.18) 2.1 (0.94) LISP 1989 1990 1989 1990 ROWRC 1989 1990 WCSP 1989 1990 WIWA 1989 1990 WTPT 1989 1990 1.6 (0.55) 2.3 (0.58) 1.2 (0.84) 1.1 (0.76) 0.9 (0.00) 1.2 (1.64) 0.8 (0.47) 1.4 (1.22) 9.5 (1.65) 7.0 (0.53) 17.7 (3.96) 15.3 (4.24) 3.2 0.9 d 0.5 e 1.8 0.5 d 1989 1990 2.9 0.6 d 0.6 e 2.9 16.2 (1.64) 14.3 (2.89) 22.5 (2.64) 17.2 (1.61) 2.8 (1.30) 2.3 (1.53) 1.3 (0.88) 1.0 0.5 0.5 d 2.0 1.1 0.6 (0.58) aAMPI-American pipit, AMRO=American robin, HOlA=horned lark, LISPLincoln's sparrow, MOBL-mountain bluebird, RO'WR-rock wren, WCSPwhite-crowned sparrow, WIWA-Wi1son's warbler, WTPT-white-tai1ed ptarmigan, YRWA-ye11ow-rumped warbler. bObserved singing male and female, but no nest: species may have used sites for foraging, only. cObserved territorial male, but no female or nest. dSing1e territorial males. eSingle non-territorial males. '. �285 The total density for each site in 1990 was 4.1 at MBC, 5.8 at GT, 4.0 at SB, and 2.2 at TR. Interestingly, between 1989 and 1990 densities dropped at MBC/GT sites and increased at SB/TR sites. White-crowned sparrows (Zonorrichia leucophrys) were the most abundant birds at all sites (Fig. 2-1). Mean densities of sparrow pairs/20 ha ranged from 7.0 at MBC in 1990 to 22.5 at TR in 1989 (Table 2-1). While nests were not observed at all sites (1 nest was found at MBC and 1 at TR), frequent observations of adult sparrows carrying food into krummholz cover confirmed nesting at GT and SB. American pipits (Anrhus rubescens) were the second most abundant birds at all sites (Fig. 2-1). Mean densities of pipit pairs/20 ha ranged from 4.4 in 1989 at TR to 7.1 in 1989 at GT (Table 2-1). Three nests were observed at MBC, and 1 was observed at TR,. Adults carrying food were observed at GT and SB, confirming nesting of this species at all sites. Horned larks (Eremophila alpesrris) used all sites for nesting (Fig. 2-1). Mean densities of lark pairs/20 ha site ranged from 0.3 in 1990 at SB to 4.5 in 1989 at MBC (Table 2-1). An unfledged lark was observed being fed by an adult at MBC (the nestling had probably been frightened away from the nest by my disturbance or by elk that had just moved through the area), and adult larks were observed performing territorial and distraction displays throughout the breeding season at all other sites. American robins (Turdus migrarorius) used all sites for foraging, and possibly for nesting (Fig. 2-1, Tables 2-1 and 2-2). Territorial males were observed at SB and TR, although nesting was not confirmed. A robin nest was observed in a conifer between SB and TR �286 Table 2-2. Avian species observed using study sites during breeding season c~nsuses along Trail Ridge in Rocky Mountain National Park, 1989 and 1990. Sites were Medicine Bow Curve (MBC; 22 ha), Gore Range Overlook (GT; 17 ha), Sundance Basin (SB; 20 ha), and Tombstone Ridge (TR; 19 ha). Species American Kestrel (Falco sparverius) American Pipit (Anthus rubescens) American Robin (Turdus migratorius) Barn Swallow (Hirundo rustica) Brown-capped Rosy Finch (Leucosticte arctoa) Brewer's Blackbird (Euphagus cyanocephalus) Broad-tailed Hummingbird (Selasphorus platycercus) Cassin's Finch (Carpodacus cassinii) Chipping Sparrow (Spizella pallida) Clark's Nutcracker (Nucifraga columbiana) Common Raven (Corvus corvax) Dark-eyed Junco (Junco hyemalis) Gray Jay (Perisoreus canadensis) Hermit Thrush (Catharus ustulatus) Horned Lark (Eremophila alpestris) Lark Sparrow (Chondestes grammacus) Lincoln's Sparrow (Helospiza lincolnii) Mountain Chickadee (Parus gambeli) Mountain Bluebird (Sialia currucoides) Northern Flicker (Colaptes auratus) Pine Grosbeak (Pinicola enucleator) Pine Siskin (Carduelis pinus) Red Crossbill (Loxia curvirostra) Rock Wren (Salpinctes obsoletus) Ruby-crowned Kinglet (Regulus calendula) Swainson's Hawk (Buteo swainsoni) . Townsend's Solitaire (Hyadestes townsendi) Unidentified Flycatcher (Empidonax spp.) Unidentified Sparrows Western Meadowlark (Sturnella neglecta) Wilson's Warbler (Wilsonia pusilla) White-crowned Sparrow (Zonotrichia leucophrys) White-tailed Ptarmigan (Lagopus leucurus) Yellow-rumped Warbler (Dendroica coronata) MBC GT SB TR N F F F N N N F Fn Fn F F F F ~ f f f f f x F F F f f N x f f F f F F f N F f f N N N f F f F f F F F f f F f f f x f n f x f f x N N F N N F x x F x N N N N N N f n aN-nested within site (nest, adults with food, unfledged young, and/or distraction displays observed), n-nesting within site possible (territorial male observed), F-used site for feeding (observed taking food), f-foraging probable (bird flushed from vegetation or rocks), x-present as migrant or wandering beyond typical habitat (may have foraged but no evidence to substantiate). .. �LOI sites, however. Mean densities of robin pairs/20 ha ranged from 0.9 in 1990 at MBC to 3.2 in 1989 at TR (Table 2-1). Mountain bluebirds (Sialia currucoides) foraging (Fig. 2-1, Tables 2-1 and 2-2). used 3 sites for In 1990, a pair nested under the eaves of the Alpine Visitor's Center (pers. obs.), which was approximately 350 m from the southern boundary of the MBC study site. Mean densities of bluebird pairs/20 ha site ranged from 0.9 in 1990 at MBC to 1.2 in 1989 at SB and TR (Table 2-1). Lincoln's sparrows (Helospiza lincolnii) used SB only (Fig. 2-1). One nest was observed in 1990, and adults carrying food were observed at SB in 1989. Lincoln's sparrows were heard singing just below GT in dense willow carr (no willow carr was present within MBC or GT study areas), but none were recorded within the study sites either year. Mean density of pairs at SB was 1.6 in 1989 and 2.3 in 1990 (Table 2-1). Wilson's warblers (Wilsonia pusilla) used SB in both years and used TR in 1989 (Fig. 2-1). Mean densities of warbler pairs/20 ha ranged from 1.3 in 1989 at TR to 2.8 in 1989 at SB (Table 2-1). No Wilson's warblers were observed using MBC or GT sites, although they could be heard singing just below GT in dense willow carr. TR was the only site where yellow-rumped warblers (Dendroica coronata) were observed consistently (Fig. 2-1). At least one male incorporated a portion of TR into his territory during 1989. Mean density of warbler pairs/20 ha was 0.6 at TR in 1989 (Table 2-1). Yellow-rumped warblers were observed using all sites for foraging (Table 2-2), but not enough observations occurred during census visits to include them in species richness or density results. . �288 The rock wren (Salpinctes obsoletus) was the only other species observed consistently during census visits (Fig. 2-1). At least one male incorporated part of TR into his territory during 1989, and two males used TR during 1990. were ever observed. However, no females or evidence of nesting Mean density at TR was 0.8 in 1989 and 1.4 in 1990 (Table 2-1). Table 2-3 lists all species, including occasionals (observed < 3 times per site), using study sites during census visits. Although, occasionals were not included in species richness or density estimates, their activities were documented. DISCUSSION Species richness and total abundance were greater at SB and TR sites than at MBC and GT sites. Habitat differences among sites probably accounted for most of the observed differences among sites. Patches of tall conifer-dominated krummholz were more extensive at SB and TR than they were at MBC and GT (Melcher 1990); this could account for large differences in densities of white-crowned sparrows, which tended to prefer dense conifer-dominated krummholz. Moderately-sized discrete patches of shrub willows were more prevalent at GT and MBC, while extensive willow carr or tiny hedged willow patches were more prevalent at SB and TR (see Chapter 1, this volume); this might have influenced densities of ptarmigan, which preferred moderately-sized discrete patches of willow (see Chapter 3, this volume). Lincoln's sparrows, Wilson's warblers, and yellow-rumped warblers were found only at SB and TR in willow carr and tall conifer-dominated krummholz. Rock wrens are known to use alpine areas where appropriate rocky habitats occur (Collister 1970, Pattie and Verbeek 1966, Medin 1987), �and only SB and TR had such habitat. Most species observed only occasionally were probably nesting in near-by cliff or subalpine habitats, and used alpinejkrummholz habitats only for foraging or while travelling to other areas. Some occasional species were observed very early or late in the breeding season and were probably migrants, non-breeding adults, or dispersing juveniles. Fewer species and lower bird densities were observed in 1990 than in 1989. The 1989 sampling period may have concentrated observations during peak territorial activities when birds were more vocal and were more likely to be counted. In 1990 the sampling period was extended further into the summer than it was in 1989, and may have been more representative of post-dispersal densities rather than breeding densities. While the 4th 1990 census was dropped to correct for dispersal effects, the 3rd census may have been somewhat effected as well. Also, I had determined previously that 4 counts were necessary to capture highest counts; lower density estimates in 1990 may have resulted from the smaller sample size. Little research has been done to determine typical densities of alpine species in Rocky Mountain alpinejkrummholz habitats, so comparisons between RMNP study sites and other alpine areas in the southern Rocky Mountains are few. Pattie and Verbeek (1966) described species observed on Beartooth plateaus (in the Beartooth Mountains, Montana and Wyoming), but densities of birds were not included. Densities of American pipits were lower on Trail Ridge than those found in the Beartooth Mountains, which were estimated at 10 pairs/20 ha by Verbeek (1970); Verbeek (1967) found 15 lark territories on 116 ha, approximately 2.59 pairs/20 ha, similar to lark densities on Trail Ridge. However, our results may not be very comparable. Verbeek's �290 study areas included alpine meadow habitat, only, while my study areas included alpine meadow, krummholz, carr, rocky outcrops, and windswept ridges. Therefore, densities of larks and pipits on Trail Ridge may have been similar to, or even higher than, those on Beartooth Plateau. Densities of American pipits were somewhat higher on Trail Ridge than in alpine regions of the Sierra Nevada (Medin 1987); Medin found 3.5-3.7 pairs of (water) pipits/20 ha. Species richness on Trail Ridge was considerably higher than it was in the Sierra Nevada; I found at least 7 nesting species while Medin found from 1-2 nesting species. Also, total pairs (all species)/20 ha were greater on Trail Ridge (23.2-33.2 versus 3.7-5.2) than in the Sierra Nevada. The differences between my results and those of Medin were probably due to large differences in associations/vertical structures of vegetation. Also, Medin's study area was over 40% rock-covered, whereas my study areas were considerably less rock-covered. He also stated that little, if any, grazing/browsing by domestic livestock had ever occurred in his study area (although he makes no mention of wild ungulates). If intensive browsing or grazing alters vegetation (and subsequently avian species diversity and/or abundance), then some of the differences cited above may have been attributable to effects of ungulates activities in RMNP. It is possible that grazing and browsing by wild ungulates enhances habitat for some birds, while degrading habitat for other species. Wild ungulates, especially elk, grazed and browsed extensively throughout the study sites during all seasons (see Chapter 1, this volume), and they undoubtedly shortened or perpetuated shortness of vegetation to some extent. Ungulates also trampled vegetation in willow and/or conifer-dominated krummholz, creating .. �L~l and/or perpetuating open areas, with very short or sparse vegetation surrounded by tall vertical cover. Most ptarmigan stayed in or near krummholz dominated by shrub willow. May and Braun (1972) found that ptarmigan forage on buds and twig tips of willows throughout the year. I observed ptarmigan foraging frequently on willows along willow-patch perimeters. Where trampling by elk perpetuated or left unchanged the "patchiness" of willow, ptarmigan foraging habitat probably was unaffected. However, where activities of elk altered patch dimensions (height and/or perimeter:area ratio) of willows, and/or where elk diminished canopy cover of shrub willow, then available willow-food resources for ptarmigan probably were diminished (see Chapter I, this volume). It is unlikely that changes to vertical structure affected ptarmigan cover. When alarmed, ptarmigan generally used "freeze" behavior to enhance their best cover: cryptic plumage. Also, vertical cover is not necessary for ptarmigan-nesting sites (Giesen et al. 1980), although a majority of hens nest where there is some vertical cover (geomorphic, topographic, or vegetative) to shield them from prevailing winds. White-crowned sparrows used conifer-dominated cover more than any other, although observed nests were located in willow and under an overhang of sod on the ground in an opening between patches of conifer. Generally, tallest branches of dwarf-conifers were used as perches for singing and for surveying territories. Elk never were observed foraging on conifers, so it is unlikely that perches used by sparrows were diminished by elk activities. Sparrows spent much time feeding in open areas between krummholz patches, where vegetation was shorter than sparrow height and sometimes quite sparse. ~ Sparrows fled .. �292 to adjacent conifer (and sometimes willow) cover when alarmed. While more conifer-dominated krummholz occurred at SB and TR sites, possibly explaining greater densities of white-crowned sparrows at those sites, it is possible that preferred sparrow habitat was maintained or increased as a result of elk activities. There is evidence indicating that elk activities have been greater at SB and TR sites (see Chapter 1, this volume). Robins used habitat much as white-crowned sparrows did, and changes in vegetation caused by browsing would be expected to have similar affects on robins and white-crowned sparrows. American pipit territories were generally located in open turf or meadow, although some used areas broken up by small patches of krummholz and/or large rocks from which they often surveyed their territories. Grazing and browsing by ungulates could perpetuate or create habitat preferred by pipits. Observed nests were situated under small overhangs of sod in fairly short dense turf (sedges, grasses, and cushion plants). In most cases, pipits were observed foraging in areas where vegetation was dense but roughly pipit height or shorter. However, if ungulate activities were to become so intensive that bare ground became prevalent, pipit habitat would be degraded. When alarmed, pipits almost never used vegetation cover, so it's unlikely that altered krummholz vegetation would affect pipits. Generally, larks used high wind-swept areas where vegetation (primarily cushion plants) was shorter than lark height and where little patches of bare, gravelly soil were common. It isn't likely that vegetation in lark habitat is suitable for foraging ungulates. Ungulates used lark habitat during travel, though, and the resulting trampling, coupled with wind effects, may perpetuate lark habitat by keeping succession in check. Larks never were observed using ~ .. �vegetation cover when alarmed. Therefore, intensive foraging activities of elk are not likely to affect larks. Pairs of mountain bluebirds foraged together, usually selecting areas where vegetation was shorter than bluebird height. Use of alpine tundra for foraging by bluebirds may have been preferred early in the breeding season when alpine areas were more snow-free than adjacent sub-alpine areas. Bluebirds were not seen as frequen~ly later in the breeding season. Hovering behavior was frequently observed while bluebirds foraged, probably because suitable dead snags from which to watch for invertebrate prey were few at most study sites. When alarmed, bluebirds used various escapes: flying short distances to open ground, flying short distances to tall rocks, or flying off sites altogether. were used. More rarely, dead snags in vegetation As long as invertebrate populations are not affected significantly by activities of ungulates, it isn't likely that .. bluebirds would be affected by intensive use of alpine tundra by ungulates. Lincoln's sparrows nested and were assumed to forage in willow carr (usually < I m high). willow carr (usually> high). heights. Wilson's warblers (Wilsonia pusilla) used I m high) and dense krummholz (usually> I m Yellow-rumped warblers used dense krummholz of various When alarmed, Lincoln's sparrows and both warbler species generally skulked within or flew to adjacent patches of dense krummholz cover; eventually, sparrows and yellow-rumped warblers would "tee-up" on tall perches to survey their territories, while Wilson's warblers generally remained concealed. If activities of ungulates were to fragment and/or diminish (heights or diameters) stands of carr �294 or krummholz, densities of Lincoln's sparrows and both warbler species could decline. In conclusion, there are numerous changes that could be inflicted on vegetation by ungulates in alpine/krummholz habitat. The possible changes could subsequently alter avian-community structure. Long-term monitoring would be a non-destructive method for evaluating effects of wild-ungulate activities in National Parks. questions to ask would be: Some important 1) are ungulates maintaining, retarding, or changing community associations and/or vegetation structure; 2) if structural changes are occurring in plant communities, how does that affect snow cover, what are the consequences of altered snow covers, and on what level of landscape would those consequences be realized, 3) how do avian communities respond over time to changes in habitat, and 4) how do similar communities outside Park boundaries compare to those inside Park boundaries? Answers to these questions could steer future decisions regarding management of wildlife on public lands and acquisitions of adjacent lands that could preserve or restore the integrity of Park ecosystems. As National Parks become more like islands of habitat in a fragmented landscape, high densities of wild ungulates within Park boundaries will continue to pose a threat to the very communities that Parks are supposed to preserve. LITERATURE CITED Bailey, A. M., and R. J. Niedrach. 1965. Birds of Colorado. Vols. I and II. Denver Mus. Nat. Hist., Denver, CO. 895pp. Bear, G. D. 1989. Seasonal distribution and population characteristics of elk in Estes Valley, Colorado. Colo. Div. Wildl. Spec. Rep. 65. l4pp. .. �_____ , G. C. White, L. H. Carpenter, R. B. Gill, and D. J. Essex. 1989. Evaluation of aerial mark-resighting estimates of elk populations. J. Wildl. Manage. 53:908915. Braun, C. E. 1969. Population dynamics, habitat, and movements of white-tailed ptarmigan in Colorado. Ph.D. Diss., Colorado State Univ., Fort Collins. l89pp. 1980. Alpine bird communities of western North America: Implications for management and research. Pp. 280-291 in Workshop Proc. Management of western forests and grasslands for nongame birds. (R.M. Degraff and N.G. Tilghman, compilers). U.S. Dept. Agric., For. Servo Gen. Tech. Rep. INT-86. _____ , R. K. Schmidt, Jr., and G. E. Rogers. 1973. Census of Colorado white-tailed ptarmigan with tape-recorded calls. J. Wildl. Manage. 7:90-93. _____ , and K. M. Giesen. 1989. Population dynamics of whitetailed ptarmigan, Job Progress Rep., Colorado Fed. Aid Project W-152-R. Apr. 1989. 245pp. _____ , and 1990. Population dynamics of white-tailed ptarmigan, Job Progress Rep., Colorado Fed. Aid Project Wl52-R. Apr. 1990. 264pp. _____ , D. R. Stevens, K. M. Giesen, and C. P. Melcher. 1991. Elk, white-tailed ptarmigan, and willow relationships: a management dilemma in Rocky Mountain National Park. Trans. 56ch N. Am. Wildl. and Nat. Res. Conf .. Wildl. Manage. Institute. Butler, D. C. 1979. Effects of a high-density population of ungulates on breeding bird communities in deciduous forest. M.S. Thesis, Colorado State Univ., Fort Collins. 67pp. Collister, A. 1970. An annotated checklist of birds of Rocky Mountain National Park. Museum Pictorial No. 18, Denver Museum of Natural History. 64pp. Giesen, K. M., C. E. Braun, T. A. May. 1980. Reproduction and nest-site selection by white-tailed ptarmigan in Colorado. Wilson Bull. 92:188-199. International Bird Census Committee. 1970. Recommendations for an international standard for a mapping method in bird census work. Audubon Field Notes 24:722-726 Jackson, S. 1991. Changes to bird communities from grazing by wild ungulates. Proc. Ecol. Soc. Am. Johnson, R. E. 1966. Alpine birds of the Little Belt Mountains, Montana. Wilson Bull. 78:225-227. �296 Kingery, H. E., (ed.). 1988. Colorado bird distribution latilong study, 3rd edition. Colorado Division of Wildlife, Denver. 81pp. Knopf, F. L., J. A. Sedgewick, and R. W. Cannon. 1988. Guild structure of a riparian avifauna relative to seasonal cattle grazing. J. Wildl. Manage. 52:280-290. Marr, J. W. 1967. Ecosystems of the east slope of the Front Range Colorado. Univ. Colorado Ser. BioI. 8. 134pp. May, T. A., and C. E. Braun. white-tailed ptarmigan 36:1180-1186. Medin, in 1972. Seasonal foods of adult in Colorado. J. Wildl. Manage. D. E. 1987. Breeding birds southern Snake Range, Nevada. of an alpine habitat in the Western Birds 18:163-168. Melcher, C. M. 1990. Avifauna responses to grazing, Job Progress Rep., Colorado Fed. Aid Project W-152-R. Apr. 1990. 264pp. Myrberget, S. 1987. grouse breeding The effects of sheep grazing habitat. Fauna 40:144-149. on a willow Olmsted, C. E. 1979. The ecology of aspen with reference to utilization by large herbivores in Rocky Mountain National Park. Pp. 87-89 in M. S. Boyce, and L. D. Hayden-Wing, eds. North American elk: ecology, behavior, and management. Univ. Wyoming, Laramie. Pattie, D. L., and N. A. M. Verbeek. 1966. Alpine Beartooth Mountains. Condor 68:167-176. Ryder, birds of the R. A. 1980. Effects of grazing on bird habitats. Pages 51-64 in Management of western forests and grasslands for non-game birds. U.S. Dep. Agric., For. Servo Gen. Tech. Rep. INT-86. Stevens, D. R. 1980a. The deer and elk of Rocky Mountain National Park: a ten-year study. U.S. Dep. Inter., Natl. Park Servo Rep. ROMO-N-13, Estes Park, CO. 163pp. 1980b. Effect of elk on alpine tundra in Rocky Mountain National Park. Pp. 228-241 in C. L. Jackson and M. N. Schuster, eds. Proc. High Altitude Revegetation Workshop 4. Colorado State Univ. Infor. Series 42. Taylor, D. M. 1986. Effects of cattle grazing on passerine birds nesting in riparian habitat. J. Range Manage. 39:254-258. Tucker, T. L. 1987. Cattle grazing affects nongame wildlife populations and fish habitat in a,montane riparian area. M.S. Thesis, Colorado State Univ., Fort Collins. 84pp. .. �297 Verbeek, N. A. M. 1967. Breeding biology of the horned lark in alpine tundra. Wilson Bull. 79:208-217. 1970. 451. Breeding ecology of the water pipit. Weber, W. A. 1976. Rocky Mountain Flora. Univ. Press, Boulder. 479pp. Auk 87:425- Colorado Assoc. Willard, B. E., and J. W. Marr. 1970. Effects of human activities on alpine tundra ecosystems in Rocky Mountain National Park, Colorado. BioI. Conserv. 2:257-265. Zwickel, F. C. 1972. Some effects of grazing on blue grouse during summer. J. Wildl. Manage. 36:631-634. . �298 CHAPTER :3 ACTIVITY BUDGETS AND REPRODUCTIVE SUCCESS OF FEMALE WHITE-TAILED PTARMIGAN IN ELK-IMPACTED HABITAT INTRODUCTION In Colorado, white-tailed ptarmigan (Lagopus leucurus altipitens) occupy alpinejkrummholz and subalpine habitats that contain willow (Salix spp.) (Bailey and Niedrach 1965, Braun and Rogers 1971). Wintering areas must include willows that remain at least partially emergent above the snow (Braun 1971, Braun et al. 1976) because the birds forage almost exclusively on buds and terminal leaders of willow from late autumn to early spring (May and Braun 1972). Breeding territories also include shrub and/or mat-forming willows (Braun 1971, Giesen et al. 1980, Schmidt 1988). Ptarmigan continue to forage on willow buds and terminal leaders through the breeding season; in addition, leaves and catkins are eaten once they have emerged (May and Braun 1972). Breeding season in the alpine is compressed by climate (Giesen et al. 1980). A ptarmigan hen must delay nesting until she has molted into breeding plumage and suitable nest sites have become snow-free. However, nesting must begin soon enough to ensure an ample growth period for juveniles prior to onset of winter. Furthermore, early nesting increases a hen's potential for renesting if the first clutch is destroyed (Giesen and Braun 1979b, Parker 1981). Therefore, between mid-May and late June, when a full molt and egg production .. �occur, a ptarmigan hen must meet significant nutritional demands. To a large extent, willows provide the nutrients necessary to meet those demands (May 1975). Researchers have suggested that a ptarmigan hen bases her selection of a mate on how much willow his territory contains (Choate 1963, Watson and O'Hare 1979, Giesen et a1. 1980, Schmidt 1988). Also, Schmidt (1988) found evidence that ptarmigan hens, especially yearling hens, may "sample" potential breeding territories prior to settlement, possibly to avoid settling where food resources are inadequate. Because ptarmigan heavily depend on willow for food, declines in breeding densities of ptarmigan probably would occur in areas where willow also had declined. Heavy browsing and/or grazing by large ungulates has been shown to cause changes in habitat structure and vegetative communities, which in turn causes lower densities (or complete elimination) of certain avian species (Zwicke1 1972, Butler 1979, Ryder 1980, Taylor 1986, Myberget 1987, Tucker 1987, Knopf et al. 1988, Jackson 1991). A similar phenomenon may have occurred on Trail Ridge in Rocky Mountain National Park (RMNP), Colorado, where breeding densities of white-tailed ptarmigan have declined (Braun and Giesen 1990). Population levels of elk (Cervus elaphus) in RMNP were controlled until 1968 when the National Park Service discontinued control efforts (National Park Service 1975, 1988). elk population increased rapidly. Subsequently, the Estimates of elk densities in the RMNP-Estes Valley area (Stevens 1980a, Bear et al. 1989) indicated that the population had exceeded the winter carrying capacity (Hobbs et al. 1982) by the early 1980s, and was at least twice the carrying capacity by 1990 (Braun et al. 1991). Elk-inflicted damage to vegetation, especially species belonging to the Sa1icaceae family, in . �300 and around RMNP was reported by the late 1970s (Olmsted 1979; Stevens 1980a, 1980b). Ptarmigan habitats were among those being used by increasing numbers of elk for greater periods of time (Stevens 1980b, Bear 1989). Braun et al. (1991) hypothesized that heavy damage to willow in ptarmigan habitats had been caused by elk, which led to declines in breeding densities of ptarmigan on Trail Ridge in RMNP. Braun et a1. (1991) reported long-term declines in canopy cover of willow and in breeding densities of ptarmigan as elk densities increased. Melcher (Chapter 1, this volume) reported that areas in which breeding densities of ptarmigan had declined most were characterized by significantly higher densities of elk feces (pellets and chips), more browsing damage to willows, and lower abundance of willow-food resources. In response to diminished or degraded food resources, ptarmigan densities could decline through several mechanisms: HI Males experience poor survivorship, resulting in vacant territories. H2 Females select against resource-poor territories, resulting in reduced fitness among territory-holding males and leaving no male recruits to replace aged/deceased territory-holders, and/or males may avoid recruiting to resource-poor territories. H3 Females settle on resource-poor territories where they must spend more time searching for and traveling to food patches, which could result in poor production and/or poor rates of return in subsequent breeding seasons. Because manipulative experiments are not allowed in National Parks, I was unable to conduct experimental tests of these hypotheses. Thus, I conducted a pilot study to evaluate hypotheses relevant to the third hypothesis. Caraco (1980) and Weathers and Sullivan (1989) have shown that the amount of time allocated to foraging by birds can be �301 explained by their nutritional with reproduction. demands status and by energy costs associated Because ptarmigan associated hens must meet high nutritional with molt and egg production within 1968, May 1975, Thomas and Popko 1980, Moss and Watson assumed that difficulties through their foraging behavior. prey intake in procuring (bite rate) depends, Holling demonstrate responses and live unbrowsed of mitigating terminal ptarmigan I assumed to their "prey" hens settling supplies food supplies, of willow. in areas where elk-use as well as would (willow buds If hens are incapable fitness would be poor as compared with adequate that that ptarmigan densities. the effects of inadequate that their inclusive territories leaders) I (1959) demonstrated in part, on prey density, time. functional 1982), then I assumed to hens using I predicted that and elk-inflicted damage to willow was greatest would spend more time foraging search and travel time), have lower bite rates items), and/or experience poor reproductive hens using sites with less elk-inflicted that inclusive fitness values chicks fledged) they budgeted and distances (West enough food would be manifested search time, and food-handling similar 3-5 weeks (more (lower density success when compared damage. .. of food to I also predicted (pre and post-nesting weights, moved by hens would be related number of to time to foraging. STUDY SITES Several historical breeding and/or current season, communities, southeast criteria were used for site selection, and occupancy 2) presence including willows. by white-tailed including ptarmigan of alpine and krummholz 1) during vegetation- Two study sites were selected (SE) region of Trail Ridge, and two study ·sites were in the �302 selected in the northwest (NW) region of Trail Ridge. At SE sites breeding densities of ptarmigan had declined most, elk-use was significantly greater, characteristics of willow indicative of heavy browsing by ungulates were significantly greater, densities of willow buds and terminal leaders were lower, and percent covers of willows (both shrub and prostrate) were lower than they were at NW sites (see Chapter I, this volume). SE study sites were Sundance Basin (SB) and Tombstone Ridge (TR), NW study sites were Medicine Bow Curve (MBC) and Gore Turnout (GT) (approximately 105°45' W long and 40°26' N 1at). Site boundaries were determined by taking compass bearings from distinctive geomorphic and topographic features. Boundaries were plotted on U.S. Geological Survey maps (scale 1:24000), and site areas were determined by using a planimeter. Approximate areas were: MBC 22 ha, GT - 17 ha, SB - 20 ha, TR - 19 ha. Elevations ranged from 3475 to 3660 m. Climate and community associations typical of Front Range a1pine/krummho1z communities were described by Marr (1967). Community types (after Braun 1969) and plant species (after Weber 1976) common throughout all sites were: Trifolium cushion fe11fie1d; Dryas oc~ope~ala hookeriana Juz. and/or Dryas-Salix (S. nivalis Hook.) turf; alpine meadow comprised of Kobresia, Carex, Trifolium, Poa, and/or Acomas~ylis rossii R. Br. (formerly Geum rossii R. Br.); snow accumulation areas comprised of Ar~emisia, Carex, Poa, and/or: Acomas~ylis; and patches of krummho1z comprised of Salix spp. (S. planifolia Pursh, S. brachycarpa Nutt.), Picea engelmanii Parry, and/or Abies lasiocarpa Hook. Communities more typical of dry wind- swept areas (Dryas and Kobresia turf) dominated sites at MBC and GT, and communities more typical of moist snow-accumulation areas (Carex .. �303 and Artemisia) greater dominated sites at SB and TR. detail in Chapter Sites were described in 1 (this volume). METHODS Selection of Hens Two methods were employed sites: 1) methodically recordings traversing of male challenge 1973), which elicited 2) tracking birds of a long-term to locate all ptarmigan study sites while playing calls and flight screams responses from territorial study (Braun and Giesen using study sites had been banded by Colorado personnel. Use of color-band identifications of individuals From 15 through sites were monitored near those areas. breeding from unobtrusive to ascertain If known pairs season) were observed, ptarmigan unique of Wildlife to each bird allowed for breeding attempts they would maintain were made to observe their pair-bond status. separated by at least five days·, in mid to late May were assumed establishing Schmidt on at least two occasions, a pair bond for the 1990 breeding 1969). established together in or as a pair during the 1989 (unbanded prior to spring 1990) to ascertain Pairs observed As part found using the study it was assumed Repeated birds. distances. their potential (observed and 1990), all ptarmigan Division 22 May, 1990, ptarmigan site and mate fidelities. unknown combinations taped (Braun et al. males, in fresh snow to detect unresponsive population using study By late May, three established season of the 12 pairs were captured fitted with radio 4-g transmitters, their capture sites. Transmitters weighed (based on and/or potentially pairs at each site had been selected Females to be for intensive study. on 23 or 29 May, 1990, (± 5 g), then released at were fitted just below ~ the hens' �304 crops with 7 em black-elastic match breeding plumages; and transmitter neck bands, and were painted the batteries had 4 month brown to life expectancies, ranges were 400 m to 5.0 km, depending on topography and weather. Pre-Nesting Activity Budgets From 24 May through (Altmann budgets of Hens 30 June continuous 1974) was conducted of radio-marked to quantify pre-incubation hens. Attempts at least once every other day. Behavior one hour long, except when terminated weather or intolerance stratified focal-animal of observers by period of day: sampling activity were made to observe samples (observations) prematurely by hens. each hen were due to severe Observations were dawn to 1029, 1030 to 1459, and 1500 to dusk. Initially, sites were visited sites were selected been observed considerably of the behavior number selecting so that activities Hen behaviors (interpreted observer, By mid June, however, more often than others. some hens had For the remainder and locating hens, observers (unless another the intimidation (i.e., displacement behavior) factor clearly data were not collected them. disturbance displays, away from rapid feeding and/or constant caused the disturbance). until the hen appeared distances varied themselves disturbing 1969) were running or creeping vocalizations, Observer-hen situated without to indicate observer-induced from Schmidt while on alert to equalize made per hen per time period. of hens could be discerned assumed warning the observer. and hens within study, hens were chosen systematically of observations After budget randomly. systematically, vigilance Activity undisturbed from approximately by �JVJ 5-100 m, depending on tolerance by hens, concealment afforded by proximal geomorphic/vegetational structures, and weather. Digital lap-timer stop watches, with cumulative-time capability, were used to measure the amount of time hens spent in each subsequent activity during I-hour sample periods. Activities and times budgeted to them, weather, location, general activities of males, foods eaten (when discernable), bite rates, and distances hens moved were recorded during each sample observation. Behaviors recorded were: feed, sit (including rest, sleep), alert, maintenance (preen, bathe, scratch, yawn), nest (build nest, lay/cover eggs), court (including copulation), aggression (towards other ptarmigan), locomotion (when not associated with an ultimate behavior), and out of view. When a "proximate" behavior appeared to be associated with an "ultimate" behavior (e.g., walking or running associated with foraging or courtship, respectively), only the ultimate behavior was recorded. However, to eliminate ambiguities protocols were used as follows: walking at the start of an observation period (i.e., possibly no frame of reference to determine ultimate behavior) was recorded as walking; walking between unrelated activities (e.g., dust bathing, resting, or feeding) was recorded as walking; walking preceded by feeding was recorded as feeding for up to 15 additional seconds after feeding had stopped, and if by that time feeding had not resumed and/or was followed by an unrelated behavior, then walking was the behavior recorded. However, to account for the possibility that continued walking ( >15 seconds) actually represented long search times for food or great distances between food patches, walking was recategorized as feeding if feeding resumed within five minutes and no other behavior was apparent during the five minutes. Occasional bites or a series of .. �306 rapid bites assumed taken during disturbance to be "displacement" recorded as alert. of view. of visible to feed then behavior allowed. Behavioral incubating behavior parts of shrubs was recorded indicated as feeding, June through hatching were recorded of further success, for each hen. nesting not out if 1) shrub willow through the success (late Dates of clutch initiation, fledging that on each hen until she began to assess her reproductive late August). success, as Then, each hen was monitored and brood seasons being also. studies were conducted her clutch. was shrubs, hens were out of Food types eaten, categorized 2) other, were recorded nesting with ultimate When hens were feeding, bite rates were recorded visibility or behavior, While feeding on willow view often, but if motions hens continued and/or extreme vigilance attempts, and predation clutch type/failure After nesting failure as determined by subsequent non-breeding areas) or fledging transmitters were removed, sizes, dates (and no possibility movement of chicks, hens were recaptured, then hens were weighed to radio and released. Data Analysis It was reasoned experience effects of disturbance observer-influenced minutes that a recently-disturbed disturbance of each observation hen might even after outward had ceased; to signs of therefore, period were dropped continue the first 10 to further reduce the l possibility (recommended of including observer-influenced by B. Van Horne, pers. comm.). spent out of view was dropped, activity category in view during behavior Also, time that hens and total time budgeted was calculated as a percentage each sample period. data to each of total time spent .. �~VI Early breeding hens, chicks regressed feeding season weights (when radios were applied) fledged per hen, and distances (GLM Procedure, moved by hens were SAS 1988) against percentages (by observation). of time spent Date, hen age, date of clutch number of eggs laid, period of day (and the corresponding term), weather Percentages subjected between factors, and observer were all treated of time spent feeding were averaged, to a one-tailed NW and SE regions activities t-test to ascertain of Trail Ridge. were summarized. to ascertain differences initiation, quadratic as covariates. by hen, then differences Percentages Also, one-tailed in weights of (P <0.05) of other t-tests were conducted and ages between hens using SE versus NW regions. Bite rates observation, (by food type) were averaged within then averaged by hen and subjected (P for differences each sample to one-tailed <0.05) between regions of Trail Ridge. for average bite rates per hen were treated as missing t-tests Zero values data. RESULTS and DISCUSSION Activity Budgets Reported otherwise data represent indicated), (unless otherwise but represent indicated) from the study sites before them. Subsequent reproductive relocations success activity, (Table 3-1). followed because (unless only three hens at SE sites three SE hens moved 4-13 km away any behavior data had been collected of hens allowed assessment from of for all but one hen. Most hens budgeted activity all six hens at NW sites more time to feeding Sitting was generally by maintenance activities. than to any other the second most common Mean.time spent feeding �w o CO Table 3-1. Mean activity budgets (percent of time in view during a sample period) for 9 female white-tailed ptarmigan during pre-nesting season in Rocky Mountain National Park, 1990. Hens 9, 11, and 12 disappeared after being radio-marked. Number of sample periods (n) and total minutes (min) observed (in view) are given. ActivitvA Site Hen (bands) n (min) Feed Sit Maint Nest Alert Locom Court A££ress MBC 1 2 3 (RYYR) (RBBW) (RBYG) 15 9 14 (630.9) (374.8) (492.3) 5l.8 79.5 59.3 24.4 11.6 15.8 13.1 4.5 7.9 5.0 0.0 6.9 2.7 2.9 4.7 2.9 l.2 6.3 0.0 0.0 0.0 0.0 0.0 0.0 GT 4 5 6 (RYWG) (RWRW) (RGGW) 16 15 13 (646.6) (637.1) (571.9) 53.8 4l. 2 39.1 25.7 30.2 33.1 4.9 12.2 14.9 12.5 6.7 9.2 l.0 6.9 l.8 l.7 2.7 l.8 0.0 0.0 0.0 0.0 0.2 0.0 SB 7 8 (RBBR) (RGRY) 12 7 (487.9) (280.7) 58.0 23.7 30.7 43.8 8.1 2.7 0.0 14.1 l.1 5.4 0.5 0.4 0.0 14.3 0.0 0.0 TR 10 (RBRG) 8 (385.5) 39.6 27.6 16.0 12.4 2.6 l.7 0.0 0.0 Feed-search for and handle food. Sit-sit/rest/sleep. " Maint-preen/bathe. Nest-build nest/lay egg/cover eggs. Alert-vigilance/head up/frequent looking up or behind/hiding/creeping/running. Locom-walk/run/f1y. Court- court/chase/copulate. Aggress-aggression (towards any ptarmigan besides own mate) . a .. �by NW and SE hens was 54.1% respectively. ± Time hens budgeted 17.2 (n - 3), to feeding did not differ between 7). NW hens tended to budget more time to feeding hens did, which was opposite original core hypothesis, declined in response necessarily falsified indicator Although to predicted that breeding to declines densities in willow-food territory of food supplies breeding no previously-used season (and subsequent territories early in the breeding selection by yearling than time hens budget significant departures) to feeding. hens would have territories), hens on 1988). If yearling (and subsequently hens do avoid then it may be biologically hens found using SE sites early in the breeding season subsequently left those sites. settling at SE sites due to inadequate (and if hens don't make mistakes Therefore, season could be, in part, attempts prior to settlement that three yearling then differences hens may be a that they used in of yearling (Choate 1963, Schmidt on inferior had was not to which they could return. appearances sample territories the of ptarmigan (Schmidt 1988), yearling territories to assess resources However, resources, adult hens tend to return to territories the preceding settling results. than SE by these results. As stated earlier, better of Trail Ridge (0.10 < P < 0.15; df - significantly However, regions ± 14.6 (n - 6) and 40.4% They may have avoided supplies of willow. in their assessments in time spent foraging If true of resources) among hens would not have been expected. Bite Rates Mean bite rates of hens on shrub willow sites (within regions), did not differ between so hen was used as the error term when testing .. �310 for regional differences in bite rates on shrub willow. However, between-site differences (within regions) were found for bite rates on other foods, in which case site was used as the error term. Hens 1 (MBC) and 10 (TR) were not observed eating shrub willow, and hen 3 (MBC) was not observed eating other foods, which are reflected in sample sizes reported. I had expected that hens would exhibit functional responses to lower densities of food (buds and terminal leaders of willow and/or other food items) through lower bite rates. However, bite rates of hens feeding on shrub willow did not differ (P >0.10) between regions: bite rates/min were 33 f 8.4 (n - 5) and 31 SE hens, respectively. ± 13.7 (n - 2) for NY and Bite rates of hens on other foods (including mat willow) were similar (P >0.25) between regions as well: bites/min were 58 ± 7.3 (n - 2) and 51 ± 12.1 (n - 2) for NY and SE hens, respectively. Bite rates may have been similar for the same reason that foraging budgets were similar: hens may have avoided settling where food densities were inadequate, in which case differences in bite rates would not have been expected. SB hens had the lowest bite rates on other foods (39 and 45 bites/min). NY hens, however, had bite rates of 53-64 bites/min (average was 59) on other foods. Interestingly, the bite rate of hen 10 at TR also was 59 bites/min on other foods. Occurrence of mat willow may have been a main factor influencing these enigmatic results. Mat willow occurred on almost every transect examined in territories used by NY hens and the TR hen, but no mat willow occurred in territories used by SB hens (see Chapter 1, this volume). Although I was never sure when ptarmigan were eating mat willow, subsequent investigations of areas used for feeding on other foods, especially > .. �where bite rates had been high, often revealed presence of mat willow. May and Braun (1972) indicated that mat willow was eaten by ptarmigan, but they did not differentiate proportions of various willow species eaten. Mat willow may have constituted a greater portion of ptarmigan diets than previously believed. If true, and if mat willow escaped damage by foraging elk (too prostrate for elk to bite), then occurrence of mat willow could have accounted for highest bite rates for NY and TR hens, and may have mitigated declines in shrub-willow. This would explain, in part, why ptarmigan densities at NW sites have declined less than those at SE sites. Although elk impacts appeared to be lighter at NW sites, it did appear that shrub willows were damaged there as well. Bite rates on shrub willow (using data for all hens) were significantly greater than bite rates on other foods: 55 ± 9.2 bites/min on other foods versus 33 ± 8.9 bites/min on shrub willow (P - 0.0004; df - 13). The difference was probably due to slower rates of "attack" (longer handling times) when hens fed on shrub willow. When pecking at other foods on the ground, hens did not appear to hesitate between bites, However, when feeding on shrub willow, a relatively long time was spent "aiming at" and "twisting off" buds and tips of terminal leaders. Measures of Fitness Weights and reproductive success of hens are summarized in Table 3-2. 7 chicks. Only 1 hen (hen 2 at MBC) was successful: she fledged 4 of All nest failures were the result of predation, either avian (strongly suspected when hen 4 lost her 1st clutch) or mammalian (strongly suspected when hen 10 lost her clutch). Causes of other �w ~ N Table 3-2. Reproductive success and weights for 3 white-tailed ptarmigan hens/4 sites in Rocky Mountain National Park, 1990. MBC hens 1-3; GT hens 4-6, SB hens 7-9, and TR hens 10-12. Hens 9, 11, and 12 left sites shortly after release; hen 9 was not seen again, hens 11 and 12 were relocated after nesting season to remove transmitters and ascertain reproductive success. Hen 6 moved out of transmitter range, died, or her radio failed, 2 days after loosing her brood. Known-age (as of 15 July 1990) hens denoted with u_u; unknownage hens denoted with u+u; all unknown-age hens were ~ 3 years old by 15 July 1990. Dates are Julian. HEN AgeR 1 RYYR 3- 144 212 390 325 -65 165 200 2 RBBW 1- 144 241 420 370 -50 157 3 RBYG 1- 144 216 400 350 -50 165 4 RYWG 3- 150 212 420 375 -45 171 5 6 RWRW RGGW 3+ 3+ 150 150 212 410 420 380 -30 172 169 201 7 RBBR 1- 144 212 390 380 -10 171 196 "8 RGRY 2- 144 212 405 365 -40 162 193 RBYY RBRG RBRW RBYR 1111- 144 144 144 144 215 215 219 385 385 360 345 370 355 360 -15 -05 +15 165 9 10 11 12 Cap Rcp WtB WtA Chg ? ? XHc D1e D2C 100 2 6 191 100 7 57 194 100 5 o o 6 o 6 871 8 o 1001 6 o 100 6 o o 6 1 1 1 ? CIn ? ? Hch o 4 1 1 o o XFl Notes 1st clutch disappeared before completion. All chicks from 2nd clutch disappeared w/in 5 d of hatching. 3 chicks disappeared w/in 14 d of hatching. All chicks disappeared w/in 7 d of hatching. 1st clutch destroyed @ 14 d (avian). 2nd clutch destroyed @ 2 d. Only clutch destroyed @ 18 d. 1 egg or chick disappeared at or shortly after hatch. All chicks disappeared w/in 3 d of hatching. All eggs or chicks disappeared w/in 24 hrs of hatching. All chicks disappeared w/in 7 d of hatching. Hen disappeared; never observed again. Nest destroyed @ approx. 22 d (mammal), Apparently unsuccessful. Apparently unsuccessful. RCap-capture date, Rcp-recapture date, WtB-weight (g) before nesting, WtA-weight (g) after nesting, Chg=net weight (g) change, CIn-c1utch initiation date, Hch-hatch date, XHc-percent hatched, H1C-number of eggs in 1st clutch, H2C-number of eggs in 2nd clutch, XF1-percent fledged . .. �nest predations complete and of all brood failures brood losses occurred within chick mortality hatching. to chick mortality. inclement weather hens brooded their chicks than they did during good weather. albus) chicks contained Erikstad that drops in ambient air temperature with dense vegetation) reduced feeding critical densities of invertebrate during wet weather. time and avoiding proportions similar of nests depredated brood failures to those reported No measures due to stochastic events, probably reproductive of Giesen that were to be high compared by time budgeted to failure among hens, possibly related to period of day (quadratic any of the variation but they were not significant respectively; effects during this study. swamped any effects Only two covariates, and the date, explained foraging, of nests depredated of fitness were explained Near-total budgets. prey associated (1977). feeding by hens. foraging the dense in starvation. during this study appeared by Giesen they found areas with more prey at such a et al. (1980) reported However, during a warm The synergistic life stage easily could have resulted to proportions (1982) with rain reduced time of chicks by 601, and that chicks avoided (and the highest for longer (Lagopus lagopus In addition, coincident (1988) noted and Spids0 twice as much food (dry weight) dry summer than during a cold wet summer. vegetation all other 14 days of Schmidt found that crops of 0-27 day-old willow ptarmigan feeding All tended to be cold and rainy during peak of hatch, which may have contributed periods 7 days of hatching; (the 3 lost by hen 2) occurred within Weather that during remain undetermined. df - 1 for each covariate). (P among times spent 0.16 and P - 0.20, Less feeding occurred term) �314 during period 2 (mid morning to mid afternoon) than during periods 1 or 3, which agrees with Schmidt's (1988) observations that ptarmigan do most of their feeding early and late in the day. Hens spent a considerable amount of time (25 to 70 minutes, with 45 - 55 minutes being typical) meticulously covering their clutches with bits of vegetation after laying each egg. Between laying eggs, covering clutches, and walking or running to/from nest areas for each egglaying, hens budgeted considerable amounts of time to nesting activities prior to onset of incubation, accounting for the decline in time spent foraging as pre-incubation season progressed. Net weight changes of hens were not directly comparable because dates of recapture varied widely among hens in relation to their nesting schedules. Therefore, only pre-nesting weights of hens were regressed against time budgeted by hens to foraging; the variation in weights was not explained by foraging budgets. However, comparisons of pre-nesting weights and ages of hens using SE versus NY sites were considered later to be worth investigating. Mean pre-nesting weights of NY and SE hens differed (P - 0.01; df - 10), with NY hens weighing 410 ± 12.7 g and SE hens weighing 378 ± 21.8 g. However, GT hens were captured 6 days later than the other hens, and may have gained some weight in that time. hens, mean weight of NY hens was 403.3 ± 15.3 g. After dropping GT The weight difference was no longer statistically significant (P - 0.1232; df 7), although a tendency was still apparent, which may have important implications for hens returning to breed in subsequent years. At two different study areas, Robb et al. (1992) found that adult female willow ptarmigan that were heavier prior to nesting had significantly '. �greater rates of return in subsequent hens with lower body mass not true for yearling rates of return. breeding seasons than did adult (65% vs. 33% and 61% vs. 15%). hens, whose weight Robb et al. suggested for female willow ptarmigan differences The same was did not affect that the costs of reproduction may be cumulative over several breeding seasons, and that female willow ptarmigan survival as they age if they have greater body mass prior to nesting. Therefore, on the basis of weight rates of return for NY hens. yearlings (see below), reproductive weights weight have better However, because and hadn't undergone hens was 12 g lighter of NY yearling hens. age on Trail Ridge, and, ultimately, their survivorship in subsequent breeding could be affected. Mean age of hens at NY sites was 2.3 least 3 years), and rates of return records regions was significant 6.5), (and may have been a partial of hens at SE sites). may be an indirect indicator .. ± 1.0 years (which may be ages of 2 hens, although between the than May's show they were at and mean age of hens at SE sites was 1.2 The age difference weights (1975) study; that food is limiting their weights low due to unknown weight less than other hens of comparable it could be a good indicator seasons of the average than the average hens was 26 g heavier If SE hens weigh effects their lower However, hens at the same time of year in May's average weight yearling most SE hens were cumulative could affect their rates of return. of yearling of alone, one would expect greater cost, there would be a lag-time before of SE yearlings chances (P - factor influencing Age structure 0.0394; df pre-nesting of hens in a given area of food resources adult hens tend to return to previously-used ± 0.4 years. in that area. territories, Because and because �316 reproductive return failure in a prior breeding rates of adult hens generally (K. M. Giesen, should not dominate However, poor survivorship assessing territory quality in which case age structure to lower weights, for selecting inferior territories). reproductive been indicative observed distances (and subsequently settle effects habitat, towards to capture However, gross movements, moved by hens within periods. time spent foraging However, observed using a feeding instead, long- periods may have which could be more larger than those For example, their territories during several most gross ranged periods sample periods from 100-200 were generally hen 8 at SB was area at least 500 m away from her eventual nest site, yet I never captured Schmidt sampling during single observation 5-40 m. of stress samples were not on a spatial-scale I-hour sampling on would be affected. moved by hens during behavior during the 500 m movement during (1969) found that hens generally nests 46-91 m from feeding areas. Braun sampling. located their (1969) also measured daily of hens, with results being similar Interestingly, are and low rates skewed success would not have to differ; m, and those captured movements could become in resource-poor of food quantity In contrast, when food supplies If there were time-lagged by time spent foraging. inadequate at any location. poor survivorship, for both adults and recruits Distances explained territories with breeding term survival hens The same result would occur if hens use other criteria and/or pers. corom.), yearling (i.e., settle where of return), associated to affect could occur if hens make mistakes leading classes. is not known female age classes inadequate, younger season to those of Schmidt. hen 8 spent the least amount of time foraging and the �)1 feeding area hen 8 was travelling willow available to her. in a large drainage feeding; see Chapter wandering with inadequate farther If hens will settle on food supplies, distances too tall to use for then they could be expected to find food. Not only would hens in search of food spend more energy on locomotion, they would be exposed to a greater anecdotal, the movements inadequate food resources ptarmigan and was probably 1, this volume). to range over greater the nearest (Some willow was closer than that, but it occurred territories 500 m to use contained territory risk of predation. While of hen 8 may have been indicative at SB. Zablan of (1986) found evidence that sizes at SE sites were larger than they were in 1969 (Schmidt 1969). A final possibility impacted willow differences to consider throughout all Trail Ridge sites. in any behaviors would not have been expected. damage to willow was observed foraging on and adjacent 1989 and 1990. or fitness values As mentioned supported decline the original is supported predictions by data collected, densities (willow or otherwise) elk-inflicted by evidence densities hypotheses of ptarmigan resources, elk that elk use of ptarmigan (Melcher 1992). of this pilot study were not alternative on territories hens spring and summer, significantly On the basis of my results, may avoid settling earlier, at NY sites and that breeding in breeding supported. among ptarmigan to NY sites throughout at both SE and NY sites have declined While If true, then no at all sites, and I observed This possibility may be increasing is that elk have negatively the on Trail Ridge were I believe with obvious however, to explain that ptarmigan inadequacies hens of food they may not be able to make . I �318 perfect distinctions between marginally-suitable and moderatelysuitable territories. Therefore, some hens may select resource-poor habitat, undergo low gains in pre-nesting weight, and experience poor rates of return and/or survivorship. As a consequence, habitat with marginal food supplies should support low densities of ptarmigan and the population of hens should be dominated by younger age classes. 'LITERATURE CITED Altmann, J. 1974. Observational study of behavior: sampling methods. Behavior 49:227-365. Bailey, A. M., and R. J. Niedrach. 1965. Birds of Colorado. Vol. I. Denver Mus. Nat. Hist., Denver. 895 pp. Bear, G. D. 1989. Seasonal distribution and population characteristics of elk in Estes Valley, Colorado. Colo. Div. Wildl. Spec. Rep. 65. 14pp. Braun, C. E. 1969. Population dynamics, habitat, and movements of white-tailed ptarmigan in Colorado. Ph.D. Diss., Colorado State Univ., Fort Collins. 189pp. 1971. Habitat requirements of Colorado white-tailed ptarmigan. Proc. West. Assoc. State Game and Fish Comm. 51:284-292. _________ , and G. E. Rogers. 1971. White-tailed ptarmigan in Colorado. Colorado Div. Game, Fish and Parks Tech. Publ. 27. 80pp. _________ , R. K. Schmidt, Jr., and G. E. Rogers. 1973. Census of Colorado white-tailed ptarmigan with tape-recorded calls. J. Wildl. Manage. 7:90-93. _______ , R. W. Hoffman, and G. E. Rogers. 1976. Wintering areas and winter ecology of white-tailed ptarmigan in Colorado. Colorado Div. Wildl. Spec. Rep. 38. 38pp. _____ , and K. M. Giesen. 1990. Population dynamics of whitetailed ptarmigan, Job Progress Rep., Colorado Fed. Aid Project W-152-R. Apr. 1990. 264pp. _______ , D. R. Stevens, K. M. Giesen, and C. P. Melcher. 1991. Elk, white-tailed ptarmigan, and willow relationships: a management dilemma in Rocky Mountain National Park. Pages 74-85 in McCabe, R.E., (ed.). Trans. 56th N. Am. Wildl. and Nat. Res. Conf., Wild1. Manage. Institute. �Butler, D. C. 1979. Effects of a high-density population of ungulates on breeding bird communities in deciduous forest. M.S. Thesis, Colorado State Univ., Fort Collins. 67pp. Caraco, T. 1980. environment. On foraging time allocation Ecology 61:119-128. in a stochastic Choate, T. S. 1963. Habitat and population dynamics tailed ptarmigan in Montana. J. Wildl. Manage. 699. of white27:684- Erikstad, K. E., and T. K. Spids0. 1982. The influence of weather on food intake, insect prey selection and feeding behaviour in willow grouse chicks in northern Norway. Ornis Scand. 13:176-182. Giesen, K. M. 1977. Mortality and dispersal of juvenile whitetailed ptarmigan. M.S. Thesis, Colorado State Univ., Ft. Collins. 55pp. _____ , and C. E. Braun. 1979a. Nesting behavior ptarmigan in Colorado. Condor 81:215-217. _____ , and Colorado. _ 1979b. Renesting Condor 81:217-218. of white-tailed of white-tailed ptarmigan in _____ , and T. A. May. 1980. Reproduction and nest-site selection by white-tailed ptarmigan in Colorado. Wilson Bull. 92:188-199. .. Hobbs, N. T., D. L. Baker, J. E. Ellis, D. M. Swift, and R. A. Green. 1982. Energy and nitrogen-based estimates of elk winter-range carrying capacity. J. Wildl. Manage. 46:12-21. Holling, C. S. predation 398. 1959. Some characteristics of simple types of and parasitism. Canadian Entomologist 91:385- Jackson, S. 1991. Changes to bird communities wild ungulates. Proc. Ecol. Soc. Am. from grazing by Knopf, F. L., J. A. Sedgewick, and R. W. Cannon. 1988. Guild structure of a riparian avifauna relative to seasonal cattle grazing. J. Wildl. Manage. 52:280-290. Marr, J. W. 1967. Ecosystems of the east slope of the Front Range Colorado. Univ. Colorado Ser. BioI. 8. 134pp. May, T. A. 1975. Physiological ecology of white-tailed ptarmigan Colorado. Ph.D. Diss., Univ. Colorado, Boulder. 311pp. _____ , and C. E. Braun. 1972. Seasonal foods of adult white-tailed ptarmigan in Colorado. J. Wildl. Manage. 36:1180-1186. in in �320 Moss, R., and A. Watson. 1984. Maternal nutrition, egg quality, and breeding success of Scottish ptarmigan Lagopus mutus. Ibis 126:212-220. Myrberget, S. 1987. The effects of sheep grazing on a willow grouse breeding habitat. Fauna 40:144-149. National Park Service. 1975. Resource management plan: Rocky Mountain National Park and Shadow Mountain National Recreation Area. U.S. Dept. Inter., Natl. Park Serv., Estes Park, CO. 88pp. National Park Service. 1988. Statement for management. U.S. Dep. Inter. Natl. Park Servo NPS D-23d. Washington, D.C. 60pp. Olmsted, C. E. 1979. The ecology of aspen with reference to utilization by large herbivores in Rocky Mountain National Park. Pages 87-89 in M.S. Boyce, and L.D. Hayden-Wing, (eds.). North American elk: ecology, behavior, and management. Univ. Wyoming, Laramie. Parker, H. 1981. Renesting biology of Norwegian willow ptarmigan. J. Wildl. Manage. 45:858-864. Robb, L. A., K. Martin, and S. J. Hannon. 1992. Spring body condition, fecundity, and survival in female willow ptarmigan. J. Anim. Ecol. 61. (In Press). Ryder, R. A. 1980. Effects of grazing on bird habitats. Pages 51-64 in Management of western forests and grasslands for non-game birds. U.S. Dep. Agric., For. Servo Gen. Tech. Rep. INT-86. SAS Institute, Inc .. 1988. SAS/STATe Procedures guide: release 6.03 edition. Cary, NC. 1029pp. Schmidt, R. K., Jr. 1969. Behavior of white-tailed ptarmigan in Colorado. M.S. Thesis, Colorado State Univ., Fort Collins. 174pp. 1988. Behavior of white-tailed ptarmigan. Pages 270Adaptive strategies and population ecology of northern grouse. Minnesota Press, Minneapolis. 299 in Bergerud, A.T., and M.W. Gratson, (eds.). Stevens, D. R. 1980a. The deer and elk of Rocky Mountain National Park: a ten-year study. U.S. Dep. Inter., Natl. Park Servo Rep. ROMO-N-13, Estes Park, CO. 163pp. 1980b. Effect of elk on alpine tundra in Rocky Mountain National Park. Pages 228-241 in C.L. Jackson and M.N. Schuster, (eds.). Proc. High Altitude Revegetation Workshop 4. Colorado State Univ. Infor. Series 42. �)LI Taylor, D. M. 1986. Effects of cattle grazing on passerine birds nesting in riparian habitat. J. Range Manage. 39:254-258. Thomas, V. G., and R. Popko. 1980. Fat and protein reserves of wintering and prebredding rock ptarmigan from south Hudson Bay. Can. J. Zool. 59:1205-1211. Tucker, T. L. 1987. Cattle grazing affects nongame wildlife populations and fish habitat in a montane riparian area. M.S. Thesis, Colorado State Univ., Fort Collins. 84pp. Watson, A., and P. J. O'Hare. 1979. Red grouse populations on experimentally treated and untreated Irish bog. J. Appl. Ecol. 16:433-542. Weathers, W. A., and K. A. Sullivan. 1989. Juvenile foraging proficiency, parental effort, and avian reproductive success. Ecol. Monogr. 59:223-246. Weber, W. A. 1976. Rocky Mountain flora. Univ. Press, Boulder. 479pp. Colorado Assoc. West, G. C. 1968.· Bioenergetics of captive willow ptarmigan under natural conditions. Ecology 49:1035-1045. Zablan, M. A. 1968. Breeding territory size of white-tailed ptarmigan in Rocky Mountain National Park, Colorado. Colorado Div. Wildl. unpubl. rpt., Fort Collins. l5pp. Zwickel, F. C. 1972. Some effects of grazing on blue grouse during summer. J. Wi1dl. Manage. 36:631-634. Prepared by Coa\~ ( Approved 0- ~. \~\ Cynthia P. Melcher by Clait E. Braun Wildlife Research Leader •• .v.. ��)L) JOB PROGRESS REPORT State of: Colorado Project: W-167 -R Upland Bird Research I Work Plan: 22 Job Title: Upland Bird Research Publications Period Covered: Author: Job 01 January through 31 December 1991 Clait E. Braun Personnel: Clait E. Braun, K. M. Giesen, R. W. Hoffman, T. E. Remington, and W. D. Snyder, Colorado Division of Wildlife ABSTRACT The following articles were published in 1991: Benson, L. A., C. E. Braun, and W. C. Leininger. 1991. Sage grouse response to burning in the big sagebrush type. Proc. Issues and Technology in the Management of Impacted Western Wildlife. Thorne Ecol. Inst. 5:97-104. Braun, C. E. 1991. Distribution and status of sage grouse in Colorado. Trans. West. States Sage and Columbian Sharp-tailed Grouse Workshop. 17:12 (Abstract). _____ , D. R. Stevens, K. M. Giesen, and C. P. Melcher. 1991. Elk, whitetailed ptarmigan and willow relationships: a management dilemma in Rocky Mountain National Park. Trans. North Am. Wildl. and Nat. Resour. Conf. 56:74-85. George, J. L., C. E. Braun, R. A. Ryder, and E. Decker. 1991. Response of waterbirds ~o experimental disturbances. Proc. Issues and Technology in the Management of Impacted Western'Wild1ife. Thorne Eco1. Inst. 5:52-59. Giesen, K. M. _____ 1991. Broadcasting birds. Colorado Outdoors 40(3):28-29. 1991. Movements and nest site selection by lesser prairie-chicken hens in Colorado. Proc. Prairie Grouse Tech. Council 19:14 (Abstract). _____ , and C. E. Braun. 1991. Demography of white-tailed ptarmigan in Rocky Mountain National Park, Colorado. Central Mtns. and Plains Sect., The Wild1. Soc. 36:Abstract. Hupp, J. W., and C. E. Braun. 1991. Geographic variation among sage grouse in Colorado. Wilson Bull. 103:255-261. Olson, T. E., C. E. Braun, and R. A. Ryder. 1991. Agricultural land use and mourning doves in eastern Colorado: implications fbr nesting and production in the Great Plains. Prairie Nat. 23:1-10. : �324 Remington, T. E. 1991. Non-target effects of aerial spraying to control Russian wheat aphids. Central Mtns. and Plains Sect., The Wildl. Soc. 36:Abstract. _____ , and C. E. Braun. 1991. How surface coal m1n1ng affects sage grouse, North Park, Colorado. ·Proc. Issues and Technology in the Management of Impacted Western Wildlife. Thorne Ecol. Inst. 5:128-132. _____ , and W. D. Snyder. 1991. Conservation Reserve Program - panacea or ecological trap? Central Mtns. and Plains Sect., The Wi1dl. Soc. 36:Abstract. _____ , and W. D. Snyder. 1991. Outdoors 40(6):11-13. Past and present pheasants. Colorado Schmutz, J. A., and R. W. Hoffman. 1991. Variable first prebasic molt in Rio Grande and Merriam's wild turkeys. Wilson Bull. 103:295-300. Schroeder, M. A. 1991. Movement and 1ek visitation by female greater prairie-chickens in relation to predictions of Bradbury's female preference hypothesis of 1ek evolution. Auk 108:896-903. _____ , and C. E. Braun. 1991. Walk-in traps for capturing greater prairiechickens on leks. J. Field Ornitho1. 62:378-385. Snyder, W. D. 1991. Wheat stubble as nesting cover for ring-necked pheasants in northeastern Colorado. Wildl. Soc. Bull. 19:469-474. _____ , and G. C. Miller, 1991. Changes in plains cottonwoods along the Arkansas and South Platte rivers - Eastern Colorado. Prairie Nat. 23:165-176. Prepared by _tla __v_'_z._.-I.~:"="'_ __;;___ Clait E. Braun Wildlife Research Leader ~ �)L7> JOB FINAL REPORT State of: Project: Colorado Upland Bird Research 'W-152/l67-R 'Work Plan: 25 Job Title: Evaluation of 'Wildlife Responses to Pesticides Used in 'Wheat Farming Period Covered: : Job _1_ 01 January 1989 through 31 December 1991 Author: Thomas E. Remington Personnel: Clait E. Braun, Carol A. Mehaffy, Michael 'W.Miller, Bradley J. Parks, Thomas E. Remington, Joan D. Ritchie, Lyn Stevens, David A. 'Wilson, Colorado Division of 'Wildlife; James Echols, Wendy Meyer, Frank Peairs, Stan Pilcher, Colorado State University; John Pearson, Claude Ross, Ted Warfield, FMC Corporation .. ABSTRACT The impacts of two methods using pesticides to control insect pests in winter wheat on bird communities were investigated. These were aerial spraying of Lorsban (ch1orpyrifos) or Di-Syston (disulfoton) to control Russian wheat ", aphids (Diuraphis noxia) in spring and microtubule injection of Furadan (carbofuran) at planting to control grasshoppers. No dead or impaired birds were found during searches of 178 ha within 11 wheat fields sprayed with either Lorsban (113 ha) or Di-Syston (65 ha). Searchers located 61% of carcasses placed in fields to evaluate search efficiency. Fifteen active nests (or broods) of several bird species were found within sprayed fields and were not obviously affected. Spraying of Di-Syston, but not Lorsban, slightly decreased (l < 0.05) 48-hour survival of pheasant chicks in a controlled experiment. 'When this experiment was repeated, spraying of Di-Syston or Lorsban had no impact on.24 or 48~hour survival of pheasant chicks .. Spraying of Di-Sy!;ton or Lorsban to control Russian wheat aphids in spring does riot appear to impact birds significantly. Carcasses of·4 horned larks (Eremophila alpestris) and 1 meadowlark (Sturnella neglecta) were found dead during searches of 65 km around the perimeters of 21 fields, 7-10 and 14-21 days; after seed planting and microtube injection of Furadan. Two birds were apparently killed by collision with a vehicle or gunshot, causes of death for the other carcasses could not be ascertained. Carbamate residues in grasshoppers averaged 1.31 mgfkg, a level unlikely to cause avian mortality. Canada geese (Branta canadensis) largely rejected Furadan-treated wheat in a controlled experiment. An average of 34 g of wheat remained per 0.785-m2 . sampling plot in the treatment pen vs. 2 grams in the untreated pen. Brain cholinesterase activity of Canada geese exposed to Furadan-treated wheat was unaffected (f> 0.05), although plasma cholinesterase activity was depressed about 38%. ��327 EVALUATION OF WILDLIFE RESPONSES TO PESTICIDES USED IN WHEAT FARMING Thomas E. Remington INTRODUCTION Wheat is the predominant cultivated crop in Colorado with upwards of 3 million acres (1.2 million hectares) planted each year. Wheat fields are used extensively for foraging and or breeding by a diverse assemblage of wildlife including ring-necked pheasant (Phasianus colchicus), horned larks, lark buntings (Calamospiza melanocorys), mourning doves (Zenaida macroura), Canada geese, and pronghorn (Antilocapra americana). Russian wheat aphids have caused extensive damage to wheat in Colorado since arriving in 1985. Consequently, massive aerial spraying has been conducted to control aphids; as much as 45% of planted wheat has been treated annually. Aerial applications of Di-Syston or Lorsban (chlorpyrifos) have emerged as the predominant treatment methods. Both are highly toxic organophosphate pesticides with the potential for wildlife mortality (Smith 1987). Microtubule injection of Furadan is a relatively new technique for controlling grasshopper depredation on immature winter wheat in fall. This technique has also been shown to be effective in controlling aphids in wheat (Guerra-Sobrevilla 1988). Microtubule injection has several advantages (from a wildlife perspective) compared to more conventional methods of pesticide application. The pesticide is applied only as a narrow border treatment (typically 12-14 rows) and is injected as a liquid 5 cm deep in the soil around the seed; consequently wildlife exposure is minimized. Furadan is a systemic insecticide which is highly toxic to wildlife (Flickinger et al. 1980, James 1987, Smith 1987, Littrell 1988). Previous bioassays of immature winter wheat treated by microtubule injection of Furadan have shown significant insecticidal activity into April. While this has positive benefits for control of grasshoppers and/or aphids, it suggests that Furadan residues or metabolites persist overwinter and indicates a potential for significant wildlife exposure. Wildlife potentially at risk of significant exposure to this pesticide would be those that feed directly on exposed seeds or immature wheat, or on grasshoppers (or other insects) that feed on wheat. Horned larks and western meadowlarks feed on exposed wheat seeds. Pronghorn and Canada geese feed on winter wheat extensively from late fall until early spring. Several insectivorous birds, primarily eastern and western kingbirds (Tyrannus tyrannus and I. verticalis), American kestrels (Falco sparverius), burrowing owls (Speotyto cunicularia), and loggerhead shrikes (Lanius ludovicianus), could be expected to feed on contaminated grasshoppers and may be at some risk. Risks may be mitigated for some species by timing of fall migrations. However Swainson's hawks (Buteo swainsoni) assemble in large pre-migratory and migratory aggregations within farmland and rangeland of eastern Colorado and feed extensively, if not exclusively, on grasshoppers (Bailey and Niedrach 1965, Woffinden 1986, Johnson et al. 1987). The purpose of this study was to evaluate the impacts of aerial spraying to control aphids and of microtubule injection of Furadan to control grasshoppers '. �328 or aphids on avian wildlife. Emphasis was on ascertaining direct mortality but indirect effects such as cholinesterase depression were also investigated. P. N. OBJECTIVES 1. Determine acute mortality of wildlife 1-3 days post-spray resulting from prescribed levels of Di-Syston or ch10rpyrifos. 2. Recover carcasses of wildlife suspected of dying from pesticide poisoning for analysis of cause of death and pesticide residue levels. 3. Monitor impacts of spraying prescribed levels of Di-Syston or ch10rpyrifos on avian nesting success. 4. Determine level of brain cholinesterase depression in songbirds and pheasants 24-48 hours post-spray as an index to both pesticide exposure and potential effects. 5. Measure toxicity of prescribed levels of Di-Syston and chlorpyrifos to 7-10 day-old pheasant chicks. METHODS Wheat Aphid Study Three research approaches were used. Transects were conducted to search for dead or impaired wildlife in sprayed fields. Exposure of breeding birds to pesticides (potentially causing sublethal effects on behavior) was evaluated by collecting horned larks from sprayed and unsprayed fields and measuring • brain acetylcholinesterase activity. Acute toxicity and indirect effects of these pesticides on survival of pheasant chicks was evaluated by 2 experiments, 1 each in spring of 1989 and 1990. Specific methodology for each approach follows. Transects were conducted by 4-5 people walking abreast, about 20 rows apart 13 days post-spray until 8.1 ha had been searched per field. Most searches were conducted with the rows, although it was found that as wheat grew taller visibility of carcasses was greatest walking against the rows. Search efficiency was measured as the percent recovery of house sparrows (Passer domesticus) placed in sprayed fields prior to the search. A randomized complete block design was used in the pheasant chick experiment conducted in 1989 (Fig 1). Nine strips, 121.9 m wide and 381 m long, were marked with flagging and randomly assigned to 1 of 3 treatments; Di-Syston, Lorsban, or unsprayed control. Within each strip, 3 of 36 possible 0.1 ha (1/4-ac) pen locations were randomly selected. Pens were constructed of 61 by 2.54 cm (24" by 1") mesh poultry netting supported by rebar and aluminum rods. Pens were offset 13 m from either edge of the strip to prevent or minimize drift of pesticide across treatments. Within the center of each pen a 9.75 m2 circular pen was constructed of 1.22 m (48") hardware cloth. Ten, 6-day old pheasant chicks were placed in each pen from 0830 to 0950. Di-Syston and Lorsban strips were sprayed with 1.12 kg active ingredient (a.i.) per ha or 0.84 kg a.i./ha, respectively. Two Pawnee PA 25 airplanes calibrated to apply �329 treatments in 3.1 liters of spray volume per ha (2 gallons/acre) were used. Spraying began at 1030 and was completed by 1200 hours. Conditions were ideal for spraying; sunny, warm and little or no breeze. 122 m SPRAY STRIPS '. Fig. 1. Experimental design and layout of treatment assess toxicity of Lorsban and Di-Syston to pheasant 1990 (bottom). strips and pens used to chicks in 1989 (top) and �330 To facilitate recovery and prevent escape, chicks were initially placed in the small pens. Chicks were removed at the end of each day and kept indoors at 32.1 C (90 F) and provided pelleted chick feed and water. Thirty (of the original 90) chicks from each treatment were individually marked by wrapping a numbered strip of scotch tape around 1 leg. These chicks were weighed each morning and returned to the larger pens for 7 days to evaluate the suitability of sprayed areas as brood habitat. Remaining chicks were labelled with nail polish which identified the treatment to which they were exposed and kept indoors to monitor subsequent mortality (2 legs painted - Lorsban; 1 leg painted - Di-Syston; neither painted - control). A randomized complete block design was also used in the 1990 pheasant chick experiment (Fig. 1). Twelve strips, 244 m long by 55 m wide, were marked with flagging within a 24-ha wheat field about 3 km south and 3 km east of Akron, Colorado. A 27-m buffer strip was marked between strips to prevent drift of pesticide across treatments. Treatments, Di-Syston, Lorsban, or unsprayed control, were randomly assigned to 1 of 3 strips within each of 4 blocks. Within each strip, 2 of 16 possible 0.1 ha (1/4 acre) pen locations were randomly selected. Pens were constructed as in 1989. Within the center of each pen a 9.75-ml circular pen was constructed of 1.22 m (48") hardware cloth. Hot boxes were constructed of 46 or 51 cm (18 or 20") diameter galvanized duct pipe, covered with fiberglass insulation and clear plastic and placed in each circular hardware-cloth pen. Temperature within each hot box was maintained at 32 C (90 F) by 125-watt heat lamps (suspended at the top) controlled by line voltage thermostats. Three generators furnished electricity to heat lamps. A door was cut into the bottom of each hot box to allow chicks access. Ten, 6-day old pheasant chicks were placed in each pen from 0645 to 0730 hours. Di-Syston and Lorsban strips were sprayed on 22 May using equipment and procedures described for the 1989 experiment. Spraying " began at 0740 and was completed by 0830 hours. Conditions were ideal for spraying; sunny, warm and little or no breeze. To facilitate recovery and prevent escape, pheasant chicks were initially placed in the small pens. Chicks were placed within the hot boxes near sunset each day and during inclement weather periods. They were provided pelleted chick feed and water ad libitum. Chicks were released from the hot boxes each morning around 1000 when air temperature had warmed. Carbofuran Study The purpose of this research was to gather preliminary information on mortality or morbidity of wildlife thought to be at risk from microtubule injection of Furadan as an at-planting treatment to control grasshoppers. Effects on passerine birds were evaluated by searching borders of treated fields (Furadan 4F at a rate of 14.8 ml per 305 linear m) for dead or impaired birds at about 1 and 2 weeks after planting. Searcher efficiency was estimated as recovery of house sparrow carcasses placed within field margins prior to the search. The effect on Canada geese of exposure to furadan-treated wheat was evaluated in two experiments in 1989 and 1990 using slightly different approaches and methods. In 1989 12 wild geese were captured on corn bait piles using a cannon net, and immediately transported to, and placed within, 1 of 3 welded wire enclosures (4 in each). Enclosures consisted of either 6 rows of treated �331 wheat, 6 rows of treated and 6 rows untreated wheat, or 6 rows of untreated wheat (check). These enclosures were within 12, 305-m long strips of 6 rows of wheat each planted on 19-20 August in a field near Briggsdale. Furadan 4F was applied to alternate strips by microtubule injection at a rate of 14.8 ml per 305 linear m. The birds were left within the pens for a day and a half to feed. They were then killed by CO2 asphyxiation. Heads were removed, placed on dry ice, then transported to and stored within a freezer at -60 C. Brain acetylcholinesterase activity was determined following Hill and Fleming (1982). The 1989 experiment was hampered by extremely low residues of Furadan in the treated wheat to which geese were exposed, either because of poor uptake, metabolism, or possibly erroneous measurement of Furadan residues by the contract laboratory. The experiment was repeated in 1990 with a modified design to insure high residue levels, representative of a worst case scenario. We planted wheat into sub-irrigated ground and exposed the geese 25-26 days after planting rather than 98-99 days as in 1989. Agri Pro, Inc., provided an irrigated field at their research farm about 4 km southwest of Berthoud, Colorado. Winter wheat (Sandy variety) was planted on 17 September 1990, at a rate of 16.2 kg per ha with a 0.3 m (12-inch) row spacing. Furadan 4F (14.8 ml per 305 linear m) and liquid fertilizer (7.68 kg N, 26.1 kg P20S per ha) were applied in-furrow by micro-tubule injection. Two blocks of untreated wheat were planted similarly, although without Furadan injection (Fig. 2). Canada geese (29) were captured while flightless on 4 June from the Loveland Golf course and maintained at the Foothills Wildlife Research Facility until the feeding trial. Birds were fed a commercial pelleted chow (Purina Turkey Ration), although they supplemented this by feeding on grass and weeds within their pen. Geese were wing clipped in late July to minimize injuries within the pen and to prevent escape from the welded wire pens in the wheat field. Geese were transported to the site on 10 October and released into a 47.2 m by 14.6 m (48 row) welded wire enclosure around untreated wheat. Geese were allowed to feed within "this enclosure for 2 days to acclimate them to a wheat diet and to conditions at the site. They were then moved into 1 of 2 adjacent 47.2 m x 7.3 m (24 row) enclosures containing either Furadan treated (15 geese) or untreated wheat (14 geese). The birds were left within the pens for two days to feed. They were then killed by CO2 asphyxiation. Heads were removed, placed on dry ice, then transported to, and stored within, a freezer at -60 C. Brain and plasma acetylcholinesterase activity was determined following Hill and Fleming (1982). All remaining wheat above ground level was clipped and weighed from 10, 0.785-m2 circular plots randomly located within treated and untreated enclosures to measure feeding intensity. Two 100 g samples of treated and untreated wheat were collected from outside the enclosures on the first trial day, kept frozen, and sent to A&L Mid West Laboratories for analysis of Furadan and Furadan metabolite residues. Analysis was by HPLC with a ultra-violet detector (no other details available) with a detection limit of 0.1 ppm. Because of questions about the accuracy of the residue levels found by A&L, duplicate samples were sent to the ACG Developmental Chemistry Department laboratory of FMC Corporation in Princeton, New Jersey for additional analysis. ~ �332 61 m (/) co v Fig. 2. m FURADAN TREATED ~ o a: 61 UNTREATED UNTREATED Field design of Furadan-wheat Canada goose study. RESULTS and DISCUSSION Russian Wheat Aphid Study Mortality Transects.--Portions of 11 fields, 178 ha total, (65 ha treated with Di-Syston, 113 ha treated with Lorsban) were searched and no dead or impaired wildlife were found. Fields were searched 3 hours (1), 1 day (7), 2 days (1), or 3 days post-spray (2). Horned larks, lark buntings, meadowlarks, and mourning doves were commonly observed in sprayed fields and appeared to act normally. Nests with eggs or chicks and/or fledged baby birds were found in some sprayed fields, usually when the attending parent(s) flushed. We located 7 horned lark nests (2 with eggs, 2 with chicks) or fledged broods (3 broods of 1-3 chicks), 5 lark bunting nests (5 with eggs), and 3 mourning dove nests (2 with eggs, 1 with young). At least in these instances, spraying did not impair normal reproductive behavior such as incubation and feeding young, nor did it kill dependant young in nests. This was confirmed by follow up visits to most nests located within a day of spraying. We recovered 61% (17 of 28) of carcasses placed in fields prior to searches. Carcasses were removed from wheat fields slowly by scavengers (Fig. 3); 75% were still present and visible 3 days after placement and 50% were present after 6 days. These data suggest that if significant mortality had occurred as a result of spraying we would have detected it. Spraying of Di-Syston apparently resulted in the deaths of 9 black-tailed jack rabbits (Lepus californicus) about 19 kilometers southwest of Pritchett in Southeast Colorado. The dead jackrabbits were found in a shortgrass pasture on 10 May, 1990, 1-30 m from a wheat field which had been sprayed late the day before. No dead jackrabbits were found within the wheat field. Another jackrabbit die-off was reported following spraying of Di-Syston north of Eaton, but we were unable to locate any carcasses. Jackrabbits are extremely sensitive to a metabolite of Di-Syston produced when the parent compound is " �333 metabolized by growing vegetation. A likely scenario would be that vegetation in the pasture was contaminated with Di-Syston by overspray or drift. Jackrabbits feeding on vegetation within the pasture would be exposed to a large dose of the metabolite of Di-Syston because the spray would be concentrated on the relatively short vegetation in the pasture. Jackrabbits have been observed within wheat fields sprayed with Di-Syston and were apparently unaffected. They appeared to be using maturing wheat fields for cover while foraging in nearby pastures. Jackrabbits are probably not at significant risk if spraying is confined to wheat. 18 18 en w en en c:( o a: 14 12 c:( o Z 10 I eoL_--~~-----L2----~3------~4----~5~----~8------:7 DAYS Fig. 3. spring. Removal rate of house sparrow carcasses placed in wheat fields in Exposure of Breeding Birds.--Horned larks were collected from a field 2 days after Di-Syston spraying (5), from another field 1 day after Lorsban had been sprayed (5), and from a rangeland area about 20 km away (3 to serve as controls). Cholinesterase activity appeared depressed in horned larks collected from the field sprayed with Di-Syston relative to those collected from the field sprayed with Lorsban or the unsprayed controls (~± S.E. - 5.0 ± 0.5, 8.2 ± 0.8, and 6.7 ± 1.0 ~moles/min/g of brain tissue, respectively). Sample sizes are too small to make definitive conclusions, but results do suggest 20-30% depression in acetylcholinesterase activity, indicative of exposure to Di-Syston. The lack of detection of carcasses or moribund birds during mortality transects suggests this exposure was sublethal and effects relatively short term. Enzyme activity rates measured were only about 50% of levels found in horned lark brain tissue by McEwen et al. (1986). This may have been due to the prolonged (4-5 months) time carcasses were frozen. McEwen et al. (1986) found that aerial spraying of Lorsban on winter wheat at 0.56 and 1.12 kg a.i.jha to control cutworms resulted in significant depression of horned lark cholinesterase activity at 3 and 9 days post-spray. There was some evidence of suppression I day after spraying with Lorsban in this study, although sample sizes were small. '. �334 Pheasant Chick Mortality.--In the 1989 experiment, Di-Syston spraying lowered (f < 0.05) 48-hour, but not 24-hour survival of pheasant chicks (Fig. 4). An average of 7.3, 8.2, and 8.2 (of 10) chicks survived 48 hours per pen sprayed with Di-Syston, Lorsban or unsprayed, respectively. Subsequent survival was similar among treatment groups. 24-HOUR MORTALITY 10~----------------------------, SUBSEQUENT SURVIVAL 30~------------------------~ 28 8 --.--.-.---.-.-.----.--. ... --..... - ... -- .... --- 26 ---------~-.""..."...----. .... ..... . ' .. .;.~ ...• •• ~ (.) 61-------- 24~----------------------~' :c (.) u, o =It::41----r-----4 ~ CHECK LORSBAN DISYSTON 18 (.) 2 W _.-:r:(.) .. 16 - 14U---~---~--~--__U 1 2 3 4 5 DAYS SINCE SPRAYING Fig. 4. Survival of pheasant chicks 1-5 days following spraying with Lorsban (0.84 kg a.i./ha) or Di-Syston (1.12 kg a.i./ha). Chick weight dynamics were similar across all treatments (Fig. 5). Birds generally lost weight when kept outside. This suggests that wheat fields are poor brood habitat for chicks of this age, at least in the absence of a hen. In the 1990 experiment, spraying of Di-Syston or Lorsban did not decrease (f > 0.05) 24 or 48-hour survival of pheasant chicks relative to controls. An average of 9.7, 9.7, and 9.9 (of 10) chicks survived for 24 hours post-spray in pens exposed to Di-Syston, Lorsban, or unsprayed, respectively. Survival to 48-hours post-spray for these groups was 9.4, 9.1, and 9.5 chicks per pen, respectively. Survival of chicks was markedly improved from the experiment conducted in 1989. This was attributed to use of the hot boxes which allowed maintainenance of chicks at or near their preferred temperature and eliminated the need to capture chicks and transport them indoors. Di-Syston �335 significantly reduced survival of chicks in the previous experiment but had no effect in this experiment. In the 1989 experiment, a storm cell passing to the south late the afternoon of the day of spraying lowered temperatures dramatically and may have exasperated toxicity of Di-Syston. Cold stress can make birds more susceptible to carbamate pesticides (Rattner et al. 1982). Given that Di-Syston and Lorsban have similar avian toxicities (Smith 1987), it is puzzling that chick mortality differed in 1989. This may be due to lower application rates of Lorsban (0.84 kg a.i./ha versus·l.12 kg a.i./ha) or to higher rates of Di-Syston reaching ground (chick) level (F. Peairs, pers. commun.), or perhaps to differential effects of cold stress. 40 - -35 0> - CONTROL ---- DI-SYSTON LORSBAN r-ill o 1- L_ r~ ::> , 25 20 -~ 1 ---- ,~--- r- o --"r!./ /' J ,/ "'. ill Cl en ~ r- ::>::> 0 =«:..... / 3 4 5 ill Cl r- ~ 2 -, -,'., " " ~ ..--_"'_._'..".- en - _--- I: - 6 0 7 8 DAYS SINCE SPRAYING Fig. 5. Yeight dynamics of pheasant chicks sprayed with Lorsban (0.84 kg a.i./ha.) or Di-Syston (1.12 kg a.i./ha). In 1990, a male lark bunting was found apparently sick and unable to fly about 7 hours after spraying near strip 12 which had been sprayed with Lorsban. This bird was placed inside one of the pens and apparently recovered as it was not found later. It may have been exposed to Di-Syston or Lorsban. Two mallard (Anas platyrhynchos) nests were discovered in the wheat field while we were building pens. These hens were each incubating 10 eggs and both were in strips sprayed, one with Lorsban and one with Di-Syston. Both hens continued to incubate their clutches for at least a week post-spray. Data from the mortality transects and pheasant chick experiments indicate that significant avian mortality is unlikely when labelled rates of these two pesticides are sprayed to control aphids. Pheasant chicks younger than 5-6 days of age may be more susceptible to these pesticides. However, most spraying occurs before most pheasant eggs have hatched, particularly nests in green wheat. Spraying usually occurs early in the morning before temperatures have warmed and winds begin to increase. Chicks younger than 5-6 days are probably brooded by hens at this time and not directly exposed to pesticides. '. �336 Carbofuran Study Mortality Transects.--Sixty-five kilometers around the perimeters of 21 fields were searched for dead or moribund birds in 1989 and 1990 within 7-10 days of planting, and again 2-3 weeks after planting (12 fields). In 1989 two horned larks and 1 meadowlark were found dead, while 2 dead horned larks were found in 1990. None of these carcasses were fresh enough for measurement of acetylcholinesterase activity. The meadowlark had injuries consistent with death resulting from shotgun pellets (radius and ulna of right wing broken, hole in right thigh and right side of pelvis broken). One horned lark carcass found adjacent to a road had injuries consistent with being struck by an automobile. The other carcasses were too degraded to yield useful information as to cause of death, and may have been predated. One horned lark had 3 wheat seeds within it's stomach. Consumption of treated seeds is a possible route of exposure. Searcher efficiency averaged 69% (25 of 36 carcasses recovered). Efficiency within each field was 67, 75, 63, 87, and 60%. These results indicate that had significant mortality occurred, we would have detected it, assuming carcasses remained until our searches. This assumption was not tested directly in the fall. Fifty percent of house sparrow carcasses placed in wheat fields in eastern Colorado during spring were still present after 7 days (Fig. 3). It appears that mortality of passerines is negligible under this application method of furadan to the borders of fields. These results should be interpreted cautiously. An inherent weakness in this approach is that birds could feed within treated field margins, flyaway to die elsewhere, and go undetected during our searches. This could be a problem with horned larks that have large home ranges in fall and winter. Swainson's hawks were plentiful in and around treated fields during the period we conducted mortality transects. Examination of hawk scats indicated almost complete feeding on insects, predominately grasshoppers. We did not detect any dead, moribund, or abnormally behaving Swainson's hawks. Exposure of other insectivorous birds to furadan-contaminated grasshoppers was presumably minimal since few insectivorous birds were observed during our searches. Fall migrations may mitigate potential exposure of these birds to furadan, although the extent to which this occurs may vary annually. Furadan residues of composite grasshopper samples measured by A&L laboratories were below detection limits from 1 treated field and up to 30 times levels measured in 1989 from another. Residues measured by FMC Corporation's lab were repeatable, much less variable (Table 1), and verified by spiking check samples. Thus, residues measured by FMC were used for hazard assessments. Assuming a 50 g (0.05 kg) bird, a grasshopper weight of 0.89 g (0.00089 kg), and total carbamate residues of 1.31 ppm (from Table 1), ingestion of a single grasshopper represented about 5.6% of the LD50 for red-winged blackbirds (Agelaius phoeniceus) and about 1.9% of the LD50 for house sparrows. Swainson's hawks may eat 100 grasshoppers per day (Johnson et al. 1987). Assuming an intake of 100 grasshoppers with an average carbamate residue of 1.31 ppm, and a bird weight of 0.908 kg, Swainson's hawks would be exposed to 3.2% of an LD50. These hazard assessments suggest little avian risk from consuming grasshoppers killed or injured by Furadan, and support the lack of insectivorous bird carcasses recovered during mortality transects. �337 Table 1. Furadan and Furadan metabolite residue levels (ppm) in winter wheat from a Canada goose study treatment field (treated with an at-planting microtubule injection of Furadan of 14.8 m1/305 m) and in grasshopper samples collected (dead or dying) from similarly treated fields (percent recovery of spiked samples from each matrix indicated in parentheses). Sample Carbofuran Wheat 102 103 2.86 (94.2) 2.48 ND8 (95.7) ND 2.62 (94.2) 3.25 0.16 (87.5) 0.35 0.01 (89.0) 0.03 0.71 (85.1) 0.73 Grasshoppers 927 919 3-Keto Carbofuran 3-Hydroxy Carbofuran 8Not detected. Analysis of scats found at several treated fields indicated that striped skunks (Mephitis mephitis) were feeding extensively on grasshoppers along field margins. It is difficult to assess the potential impact to skunks of this exposure since information on how many grasshoppers they consume in a day and their sensitivity to furadan is not available. LD50's of dogs and rats are 19 and 11 mg/kg, respectively (Smith 1987). Using 10 mg/kg as an LD50,. and assuming a body weight of 3 kg and consumption of 100 or 500 grasshoppers per night, skunks would be exposed to 0.39 - 1.94% of an LD50. If these residues are accurate, and these assumptions reasonable, mortality of skunks seems unlikely. Canada Goose Experiment.-- In the 1989 experiment, brain acetylcholinesterase activity was unaffected by exposure to furadan treated wheat (Table 2). Two wheat samples collected at the Briggsdale site prior to this study had furadan residues of 0.06 ppm. Residues at the Loveland site were 0.84 and 1.16 ppm while treated wheat at the Severance site had residues of 0.34 and < 0.05 ppm. Variation among sites seems large since all 3 were planted at the same time, using the same equipment and applicator. It is not surprising that no effect was observed in geese at the Briggsdale site given these residues. In retrospect the Loveland site may have been a better test. It's possible that furadan at the Briggsdale site had been metabolized in the soil or wheat and was still present as a toxic metabolite. •• �338 Table 2. Brain acetylcholinesterase activity of 12 Canada geese after feeding in pens containing untreated, 50% furadan treated or 100% furadan treated wheat. Treatment Mean ± S.D. Untreated 50% Treated 100% Treated 1502 967 1007 959 867 751 1296 975 1072 1256 881 964 1108 ± 263 972 ± 234 1043 ± 162 In the 1990 experiment, geese adapted well to experimental conditions and grazed the untreated wheat heavily during the 2 days of acclimation. Almost no wheat remained in this pen after two days. Geese within the control (untreated) pen continued to graze heavily following their transfer to this pen, but feeding was minimal in the pen with Furadan-treated wheat. Avoidance of Furadan-treated wheat continued through the second day, despite the fact that there was no other food source. Wheat (above ground) within the control pen was essentially eliminated by the conclusion of the trial. When wheat samples were clipped and weighed from each pen the next morning, some recovery of grazed wheat was apparent. Still, only 2.03 ± 0.86 g remained in 0.785 m2 ~ plots in the control pen, significantly less (f < 0.05) than the 33.95 ± 9.33 g remaining in plots within the treatment pen. Avoidance of Furadan-treated wheat is significant, because large die-offs of geese and other waterfowl have occurred when alfalfa treated with Furadan 4F flowable was (apparently) eaten (Flickinger et al. 1980, reviewed in Smith 1987). Canada geese have been successfully deterred from feeding on rye and turf by methiocarb (another carbamate pesticide) application (Conover 1985, 1989), but methiocarb is substantially less toxic to birds than Furadan. Carbamate residues in treated wheat samples, adjusted for percent recovery, were 5.81 and 6.08 ppm. Dietary LC50 for mallards is 190 ppm (Smith 1987), over 30 times higher than levels measured in wheat. Mallard LD50's are substantially lower, in the range of 0.4-0.6 mgjkg (Smith 1987). A 2.5-kg Canada goose eating 0.35 kg of green wheat with a total carbamate level of 5.7 mg/kg would be exposed to a dose of 0.8 mgjkg body weight, or in excess of a mallard LD50. Thus, if geese had eaten large quantities of green wheat in this trial some mortality might have occurred. It is interesting that geese apparently eat enough Furadan-treated alfalfa to poison themselves yet avoid wheat containing potentially lethal levels. A possible explanation is that geese don't reject Furadan-treated vegetation on the basis of taste, but rather because of post-ingestive consequences such as feeling ill. Geese may sample alfalfa sprayed with Furadan 4F, be exposed to a lethal dose and die. Sampling wheat in this study would have exposed them to a sub-lethal dose that may have made them sick enough to avoid additional feedi~g. �339 Acetylcholinesterase (ChE) activity of brain tissue from geese in treatment and control pens did not differ (f < 0.05, Table 3). Plasma ChE was depressed (f < 0.05) an average of about 38% in birds exposed to Furadan treated wheat. Plasma ChE inhibition is a sensitive indicator of exposure to antiChE agents but is not in itself indicative of negative health consequences because the function of plasma ChE is not known (Kutty 1980). Plasma ChE may be inhibited by as much as 75% when brain ChE reaches a detectable and biologically significant 10-20% (Ludke et ale 1975). These results support the premise that initial feeding on treated wheat exposed geese to Furadan, but feeding ceased prior to significant brain ChE depression. Table 3. Brain and plasma acetylcholinesterase activity (mean ± SD) of Canada geese from pens containing Furadan-treated (n - 15) or untreated (n - 14) winter wheat, 1990. Plasmaa Brainb S.D. Untreated Treated 672.1 419.9 201.9 124.3 S.D. 5.57 5.58 0.83 1.48 &wu acetylthiocholine iodide hydrolyzed per min per ml of plasma at 25 C. b ~mol acetylthiocholine iodide hydrolized/g (wet weight) of brain tissue at 25 C. " Wildlife risks from microtubule injection of furadan appear slight. Residue levels in green wheat are low relative to LC50 or LD50 values of mallards or geese. Geese avoided eating treated wheat and kept intakes below levels that would reduce brain ChE. Risks to insectivorous birds are mitigated by timing of migrations and low furadan residues in grasshoppers. The primary risk appears to be to granivorous birds, principally horned larks and meadowlarks. Few unexplained mortalities were found, suggesting that losses were slight or nonexistent. LITERATURE CITED Bailey, A. M., and R. J. Niedrach. 1965. Birds of Colorado. Denver Mus. Nat. Rist. Denver, Colo. 895 pp. Vol. II. Conover, M. R. 1985. Alleviating nuisance Canada goose problems through methiocarb-induced aversive conditioning. J. Wildl. Manage. 49:631-636. 1989. Can goose damage to grain fields be prevented through methiocarb-induced aversive conditioning? Wildl. Soc. Bull. 17:172-175. Flickinger, E. L., K. A. King, W. F. Stout, and M. M. Mohn. 1980. Wildlife hazards from Furadan 3G applications to rice in Texas. J. Wildl. Manage. 44:190-197. �340 Guerra-Sobrevilla, L. 1988. Effectiveness of carbofuran applied to the soil and as a seed treatment for the control of aphids on wheat in northwestern Mexico. Crop Prot. 7:336-337. Hill, E. F., and W. J. Fleming. 1982. Anticholinesterase poisoning of birds: field monitoring and diagnosis of acute poisoning. Environ. Toxicol. Chem. 1:27-38. James, P. C., and G. A. Fox. 1987. Effects of some insecticides on productivity of burrowing owls. Blue Jay 45:65-71. Johnson, C. G., L. A. Nickerson, and M. J. Bechard. 1987. Grasshopper consumption and summer flocks of nonbreeding Swainson's hawks. Condor 89:676-678. Kutty, K. M. 1980. 13:239-243. Biological function of cholinesterase. Clin. Biochem. Littrell, E. E. 1988. Waterfowl mortality in rice fields treated carbamate, carbofuran. Calif. Fish and Game 74:226-231. with the Ludke, J. L., E. F. Hill, and M. P. Dieter. 1975. Cholinesterase (ChE) response and related mortality among birds fed ChE inhibitors. Arch. Environ. Contam. and Toxicol. 3:1-21. Rattner, B. A., L. Sileo, and C. G. Scanes. 1982. Hormonal responses and tolerance to cold of female quail following parathion ingestion. Pest. Biochem. Physiol. 18:132-138. Remington, T. E. 1990. Evaluation of wildlife responses to pesticides used in wheat farming. Colorado Div. Wildl. Job Prog. Rep., Fed. Aid Proj. W-152-R. Apr:253-264. Smith, G. J. 1987. Pesticide use and toxicology in relation to wildlife: organophosporous and carbamate compounds. U.S. Dep. Inter., Fish and Wildl. Servo Resour. Publ. 170. l7lpp. Woffinden, N. D. 1986. Notes on the Swainson's hawk in central Utah: insectivory, premigratory aggregations, and kleptoparasitism. Great Basin Nat. 46:302-304. Prepared by t. Thomas E. Remington Wildlife Researcher C �)41 JOB PROGRESS REPORT Colorado State of: Project: W-167-R Upland Bird Research Work Plan: 26 Job Title: Analysis of Upland Bird Population Trends Period Covered: Author: Job _1_ 01 January through 31 December 1991 Clait E. Braun Personnel: Clait E. Braun, Kenneth M. Giesen, Richard W. Hoffman, Thomas E. Remington, and Warren D. Snyder, Colorado Division of Wildlife ABSTRACT The following were prepared in 1991. Braun, C. E. 1991. 1991. Blue Mountain sage grouse harvest data, 1976-91. Cold Spring Mountain harvest data, 1976-91. 1991. Sage grouse harvest data, Eastern Moffat and Western Routt counties, Colorado, 1976-91. 1991. Eagle County sage grouse harvest data, 1991. 1991. Gunnison Basin sage grouse harvest data, 1991. 1991. Middle Park sage grouse harvest data, 1991. 1991. Northcentral Moffat County sage grouse harvest data, 1976-91. 1991. North Park sage grouse harvest data, 1991. 1991. Piceance Basin sage grouse harvest data, 1991. 1991. Yampa Area sage grouse harvest data, 1991. _____ , T. E. Remington, and W. D. Snyder. 1991. Habitat development programs: analysis and recommendations. Colorado Div. Wild1. Unpub1. rep. Fort Collins. 23pp. Giesen, K. M. 1991. Columbian sharp-tailed grouse harvest data, northwest Colorado, 1976-91. '. � https://cpw.cvlcollections.org/files/original/5c482de6ffb17ce3b1d181fcb81e4186.pdf dc06c892556d660e969b90e8c9ddaca6 PDF Text Text Colorado Division of Wildlife Wildlife Research Report July 1992 JOB PROGRESS REPORT State of Colorado Mammals Research Proj ect No. _W~-.::!:.15::::.;3~-...IR~--=5~ _ • Work Plan No. Mu1tispecies 1 Job No. Mammals Publication, Editing, and Library Services Period Covered: Author: Investigations July 1, 1991 - June 30, 1992 J. A. Boss Personnel: J. A. Boss, N. W. McEwen ABSTRACT During the Segment the following were accomplished: * 24 publications were purchased at the request of Wildlife Researchers and placed into the Colorado Division of Wildlife Research Library Collection. * 26 Theses or books were obtained on Interlibrary Loan or as gifts for use by Mammals Researchers. * * 15 manuscripts were published or accepted for publication. 6 manuscripts were prepared and submitted for peer review. ��3 MAMMALS PUBLICATION. EDITING AND LIBRARY SERVICES Jacqueline A. Boss P. N. OBJECTIVE To provide a centralized support program for manuscript editing and library services to facilitate publishing results of research conducted by staff of Federal Aid Project W-153-R . . SEGMENT OBJECTIVE To provide a centralized support program for Mammals Research editing, library, and publishing services so that Mammals Research personnel can be most efficient in publishing results of their research. ··SUMMARY OF SERVICES Publications Purchased with Mammals Research funds and Placed in the Research Center Library Agresti, A. 558pp. 1990. Categorical data analysis. New York; John Yiley & Sons. Bell, S. S., E. D. McCoy, and H. R. Mushinsky, eds. 1991. Habitat structure: the physical arrangement of objects in space. New York; Chapman & Hall. 438pp. Chatfield, C. 1989. Analysis of time series: an introduction. 4th edition. New York; Chapman & Hall. 241pp. Colgan, P. 1989. Animal motivation. Cressie, N. A. C. 1991. Sons. 900pp. New York; Chapman & Hall. Statistics for spatial data. l59pp. New York; John Wiley & Darling, L. M. 1990. Bears - their biology and management. International conference on bear research and management (8th: 1989: Victoria, British Columbia). Victoria, BC; T. D. Mock & Associates. 448pp. Dwight, H. J. 1971. The application of an antifertility agent in the control of a white-tail deer population. Ann Arbor, MI; University Microfilms. l62pp. Ehrenfeld, D. 1981. Press. 286 pp. The arrogance of humanism. New York; Oxford University Fan, L. T., D. Neogi, and M. Yashima. 1991. Elementary introduction to spatial and temporal fractals. New York: Springer-Verlag. l68pp. �4 G1adders, D., and N. Jotham, Coordinators. (1988). Proceedings: International Symposium on Trapping Wild Furbearers. Toronto, Ont.; Fur Institute of Canada. up. Gould, S. J. 1989. Wonderful life: the Burgess Shale and the nature of history. New York: W. W. Norton & Company. 347pp. Halfpenny, J. 1986. A field guide to mammal tracking in North America. Boulder, CO; Johnson Books. 16lpp. Hilfiker: E. L. 1991. Beavers: water, wildlife and History. Windswept Press. 197pp. Hoffmeister, D. F. 1986. Mammals of Arizona. Arizona Press. 602pp. Hosmer, D. W., and S. Lemeshow. 1989. York; John Wiley & Sons. 307pp. Interlaken, NY; [Tucson, AZ]; University of Applied logistic regression. New Hudson, R. J., K. R. Drew, and L. M. Baskin, eds. 1989. Wildlife production systems: economic utilization of wild ungulates. New York; Cambridge University Press. 469 pp. Knobil, E., and J. D. Neill, eds. York; Raven Press. 2 vo1s. 1988. The physiology of reproduction. New Li, H. 1989. Spatio-temporal pattern analysis of managed forest landscapes: a simulation approach. Ann Arbor, MI; University Microfilms International. l66pp. Palo, R. T., and C. T. Robbins. 1991. Plant defenses against mammalian herbivory. Boca Raton, FL; CRC Press. 192pp. Peterle, T. J. 1991. Wildlife toxicology. New York; Van Nostrand. 322pp. Renecker, L. A., and R. J. Hudson, eds. 1991. Wildlife production: conservation and sustainable development. AFES misc. pub. 91-6. Fairbanks, AK; University of Alaska-Fairbanks. 60lpp. Roberts, E. A. 1992. Sequential data in biological experiments: an introduction for research workers. New York; Chapman & Hall. 240pp. Southeastern Association of Fish and Wildlife Agencies. [1991]. Proceedings of the forty-second annual conference; Southeastern Association of Fish and Wildlife Agencies, November 6-9, 1988, Hilton Head Island, SC. 605pp. Servheen, C. 1990. Status and conservation of the bears of the world. Victoria, BC; T. D. Mock & Associates. 32pp. In addition to books purchased with Federal Aid Funds, about 7 free reports and short publications from state or federal agencies or from private sources were located, ordered, and obtained for use by Mammals Research personnel. �5 Theses and Books Obtained on Interlibrary Loan or as Gifts for Use by Researchers Allen, R. B. 1985. Research and management implications of the pursuit of black bears with trained bear dogs. M.S. Thesis, University of Montana, Missoula, MI. 5lpp. Alt, G. L. 1989. Reproductive biology of female black bears and early growth and development of cubs in northeastern Pennsylvania. Ph.D. Dissertation, West Virginia University, Morgantown, WV. l16pp . . Anderson, R. M., B. D. Turner, and L. R. Taylor, eds. 1979. Population dynamics: the 20th symposium of the British Ecological Society. Oxford: Blackwell Scientific Publications. 434pp. Beecham, J. J. 1980. Population characteristics, denning, and growth patterns of black bears in Idaho. Ph. D. Diss., University of Montana, Missoula, MI. 101pp. Berg, H. H. 1983. Random walks in biology. Princeton, NJ. l42pp. Princeton University Press, Davis, L. R. and R. E. Marsh, eds. 1990. Proceedings: vertebrate pest conference (14th: 1990: Sacramento, CA). University of California, Davis. 372pp. Feynman, R. P. 1988. "What do you care what other people think?" further adventures of a curious character. New York: W. W. Norton & Company. 255pp. Frost, H. C. 1990. Population and reproductive characteristics of black bears on an isolated mountain in southeastern Utah. M.S. Thesis, Brigham Young University, Provo, UT. 45pp. Graber, D. M. 1982. Ecology and management of black bears in Yosemite National Park: final report to the National Park Service. University of California, Davis, CA. 205pp. Keay, J. A. 1990. Black bear population dynamics in Yosemite National Park. Ph. D. Diss., University of Idaho, Moscow, 10. l26pp. Knowles, W. R., Jr. 1989. A national comparison of structural factors affecting participation in selected wildlife-related activities. M.S. Thesis, Texas A & M University, College Station, TX. 87pp. Koch, D. B. 1983. Population, home range, and denning characteristics of black bears in.Placer County, California. M.S. Thesis, California State University, Sacramento, CA. 7lpp. Levi-Montalcini, R. York. 220pp. 1988. In praise of imperfection. Basic Books, Inc., New �6 Li, H. 1989. Spatio-tempora1 pattern analysis of managed forest landscapes: a simulation approach. M.S. Thesis, Oregon State University, Corvallis, OR. 166pp. Malvil1e, L. E. 1990. Movements, distribution, and habitat selection of river otters reintroduced into the Dolores River, southwestern Colorado. M.A. Thesis, University of Colorado, Boulder, CO. 67pp. Organ, J. F. 1989. Mercury and PCB residues in Massachusetts river otters: comparisons on a watershed basis. Ph. D. Diss., University of Massachusetts, Amherst, MA. 69pp. Risenhoover, K. L. 1987. Winter foraging strategies of moose in subarctic and boreal forest habitats. Ph. D. Diss., Michigan Technological Univ., Houghton, MI. l20pp. Rohlman, J. A. 1989. Black bear ecology near Priest Lake, Idaho. Thesis, University of Idaho, Moscow, ID. 76pp. M.S. Ropek, M. M. 1990. Mercury levels in Michigan River otters in relation to population characteristics and habitat. M.S. Thesis, Eastern Michigan University, Ypsilanti, MI. 6lpp. Rose, C. L. 1990. Application of the carbon/nitrogen balance concept to predicting the nutritional quality of blueberry foliage to deer in southeastern Alaska. Ph. D. Diss., Oregon State University, Corvallis, OR. 16Opp. Ruther, S. A. 1988. Urban wildlife conservation in Arizona: public opinion and agency involvement. M.S. Thesis, University of Arizona, Tempe, AZ. 9pp. Seber, G. A. F. 1982. The estimation of animal abundance and related parameters. 2nd ed. Macmillan Publications, New York. 654pp. Sweanor, L. L. 1990. Mountain lion social organization in a desert environment. M.S. Thesis, University of Idaho, Moscow, ID. Smith-Swanson, S. A. 1990. Activities and habitat use of black bears in north-central Minnesota. M.S. Thesis, University of Minnesota, St. Paul, MN. 75pp. United States. Bureau of Outdoor Recreation. 1975. Proposed Bruneau Wild and Scenic River, Idaho. U.S. Bureau of Outdoor Recreation, Northwest Regional Office, Seattle, WA. 56pp. White, P. J. 1990. The pathological response of red foxes to unpadded foothold traps. M.S. Thesis, University of Minnesota, St. Paul, MN. 55pp. The Research Center Library staff also located and delivered about 545 individual articles on request for Mammals Researchers during this segment; about 15 were not available locally and were obtained through interlibrary loan procedures. �7 Manuscripts Published FY 1991-92 Job Progress Reports; Federal Aid. All studies. Andelt, W. F., D. L. Baker, and K. P. Burnham. 1992. Relative preference of captive cow elk for repellent-treated diets. J. Wildl. Manage. 56(1):164-173. Anderson, A. E. 1992. The puma on Uncompahgre Plateau, Ccolorado. Div. Wildl. Tech. Publ. No.40. Colo. I Bartmann, R. M., G. C. White, and L. H. Carpenter. mortality in a Colorado mule deer population. 1992. Compensatory Wildl. Monogr. 121. Beck, T. D. I. 1991. Black bears of west-central Colorado. Wild1. Tech. Publ. No. 39. 86pp. 37pp. Colo. Div. Freddy, D. J., E. R. Ryland, and R. M. Hopper. 1991 Colorado's wildlife ranching program: the Forbes Trinchera experience. In: Wildlife production: conservation and sustainable development. eds. L. A. Renecker, and R. J. Hudson, p.336-343. AFES misc. pub. 91-6. University of Alaska Fairbanks. Fairbanks, Alaska. Gill, R. B., and T. D. I. Beck. 1991. Black bear status report: Colorado. W. Black Bear Workshop, Proc. 7 (in press) Gross, J. E., L. A. Shipley, N. T. Hobbs, D. E. Spalinger, and B. A. Wunder. 1992. Foraging by herbivores in food-concentrated patches: tests of a mechanistic model of functional response. Ecoloy. (in press) Hobbs, N. T., and M. W. Miller. 1992. Interactions between pathogens and hosts: simulation of pasteurellosis epizootics in bighorn sheep popUlations. In: Wildlife 2001: populations. D. R. McCullough and R. H. Barrett, eds., Elsevier Science Publishers, Ltd., London, England, (in press) Miller, M. W., N. T. Hobbs, and E. S. Williams. 1991. Spontaneous pasteurellosis in captive Rocky Mountain bighorn ship (Ovis canadensis canadensis): clinical, laboratory, and epizootiological observations. J. Wildl. Dis. 27: 534-542. Miller, M. W., J. M. Williams, T. J. Schiefer, and J. W. Seidel. 1991. Bovine tuberculosis in a captive elk herd in Colorado: epizootiology, diagnosis, and management. Proc. U.S. Anim. Health Assoc. 95: 533-542 Rbyan, J. C., D. A. Saari, E. S. Williams, M. J. Wilson. 1992. Gross and microscopic tuberculosis in a captive herd of wapiti Colorado. J. Vet. Diagn. Invest. 4: (in W. Miller, A. J. Davis, and A. lesions of naturally occurring (Cervus elaphus nelsoni) in press) Snipes, K. P., R. W. Kasten, M. A. Wild, M. W. Miller, D. A. Jessup, R. L. Silflow, W. J. Foreyt, and T. E., Carpenter. 1991. Using ribosomal RNA gene restriction patterns in distinguishing isolates of Pasteurella haemolytica from bighorn sheep (Ovis canadensis). J. Wildl. Dis. 28: 347-354. �8 Spa1inger, D. E., and N. T. Hobbs. 1992. Mechanisms of foraging in vertebrate herbivores: new models of functional response. Am. Nat. 140:325-348. Thorne, E. T., M. W. Miller, D. L. Hunter, and E. s. Williams. 1992. Wildlife management agency concerns about bovine tuberculosis in captive cervidae. Proc. Bovine Tuberculosis in Cervidae Symposium. USDA/APHIS. (in press) Williams, E. 5., M. W. Miller, S. Young, and E. T. Thorne. 1992. Chronic wasting disease: a spongiform encephalopathy of mule deer (Odocoileus heminonus) and Rocky Mountain elk (Cervus elaphus nelsoni) in Colorado and Wyoming, USA. Proc. World Assoc. Vet. Microbio1. Immuno1. 12: (in press) Manuscripts in Review FY 1991-92 Pojar, T. M. 1991. Census techniques. In: Pronghorn management. ed. R. McCabe. Wildlife Management Institute, Washington D.C. Submitted March 6, 1991. 1991. ed. R. McCabe. March 6, 1991. The use of population models. In: Pronghorn management. Yi1d1ife Management Institute, Washington D.C. Submitted ___________ , D. C. Bowden, and R. B. Gill. 1991. techniques for pronghorn. J. Wild1. Manage. Evaluation of census (in review) Ringe1man, J. K., M. W. Miller, and W. F. Ande1t. 1992. Effects of ingested tungsten-bismuth-tin shot on mallards. J. Wildl. Manage. Torbit, S. C., R. B. Gill, A. W. Alldredge, and J. A. Liewer. 1993. Impacts of pronghorn grazing on winter wheat in Colorado. J. Wildl. Manage. (in press) Wild, M. A., M. W. Miller, D. L. Baker, R. B. Gill, N. T. Hobbs, and B. J. Maynard. 1992. Hand-raising wild neonates using an evaporated milk diet fed ad libitum: evaluation in bighorn sheep, pronghorn antelope and elk. Can. J. Zool. (in review) �9 Colorado Division of Wildlife Wildlife Research Report July 1992 JOB PROGRESS REPORT State of C~o~l:!:.:o~r!:.!a~d~o~ _ Project No. W-lS3-R- 5 Mammals Research Work Plan No. 1 Multispecies Job No. 9 Mammals 1 Research Administration Period Covered: Author: Investigations July I, 1991-June 30, 1992 R. B. Gill Personnel: R. B. Gill, L. E. Lovett, and D. K. Hall Abstract Accomplishments of this job are summarized in the Job Progress Report for W -lS3R-S Work Plan LA Job 3. . . '.' . . Wildlife .ResearchLeader ��11 Colorado Division of Wildlife Wildlife Research Report July, 1992 JOB PROGRESS REPORT State of Colorado Project No. ~W_-~1~5~3~~R~-5~ _ Mammals Research Work Plan No. =2 _ Deer Investigations Job No. 7 Period Covered: Author: Development of Census Methods for Deer in Plains Riverbottom habitats July 1, 1991 - June 30, 1992 R. C. Kufeld Personnel: D. Bowden, D. Younkin ABSTRACT White-tailed and mule deer, radio-collared in the Arkansas Riverbottom during January, 1991, were located at approximately 2-week intervals through December, 1991. Movement data from 50 deer, radio-collared as adults, (26 mule deer and 24 whitetails) suggest that adult deer in the South Platte Riverbottom fall into 4 movement categories. These categories include resident deer which occupy the same general area yearlong, and migratory deer which occupy separate winter and summer areas. These patterns were observed for both mule deer and whitetails and are described in detail including sizes of seasonal areas used. Dispersal of deer, eartagged as fawns, of up to 149.3 km was recorded for mule deer and up to 130 km for whitetails. Among deer radio-collared in the South Platte Riverbottom as adults 73% of the mule deer and 67% of the whitetails were nonmigratory. Most of the nonmigratory deer of both species lived in rhe riverbottom yearlong, but some, after being tagged in the riverbottom, established yearlong home ranges on the plains away from the river. All of the migratory mule deer'and most of the migratory whitetails followed a pattern of migration,'in late spring or summer, from the riverbottom to the plains with return l~ter in the summer or in the fall, but some whitetails migrated ,atvarious 'times of the year to distant sites within the riverbottom. Migratory deer of both species exhibited strong fidelity to their riverbottom and plains home ranges, returning to the same areas each winter and summer. Adult radio-collared white-tailed deer in this study were observed to occupy larger minimum convex polygon areas and to have longer movements in general than did the adult instrumented mule deer. Our findings also suggest that white-tailed deer in plains riverbottom habitats in eastern Colorado are considerably more mobile than whitetails in most other parts of the u.S. ��13 .DEVELOPMENT OF CENSUS METHODS FOR DEER IN PLAINS RIVERBOTTOM HABITATS Roland C. Kufeld P. N. OBJECTIVES 1. To determine seasonal movements and home range size of white-tailed and mule deer in plains riverbottom habitats. 2. To develop and test methods for estimating size of deer populations in plains riverbottom habitats. SEGMENT OBJECTIVE To determine seasonal movements and home range size of white-tailed and mule deer in plains riverbottom habitats. STUDY AREA The South Platte River study area was described by Kufeld (1989), and the Arkansas River Study area by Kufeld (1991). METHODS AND MATERIALS White-tailed and mule deer, radio-collared in the South Platte Riverbottom during January and February, 1987, 1988, and 1989, (Kufeld 1987, 1988, 1989, 1990,), were located at approximately 2-week intervals from January, 1987 through December, 1990. White-tailed deer, radio-collared in the Lower Arkansas Riverbottom during January, 1991 (Kufeld 1991), were located at approximately 2-week intervals from January, 1991, through December, 1991. Most locations were made by aerial telemetry using a Cessna 185 aircraft with a 2 element, "H" configuration receiving antenna mounted on each strut. A switchbox permitted the telemetry operator, a passenger in the aircraft, to operate antennas jointly or separately. When the airplane was not available locations were made by tracking on the ground until the animal was observed Deer locations were plotted on USGS 1:50,000 scale maps and recorded by UTM coordinates. Vegetation type was also recorded for each deer location. Analysis of data collected during the South Platte River phase of this study was continued during FY 91-92. Previous progress reports have described movements of those instrumented deer on the South Platte which had died or whose transmitters had quit up to the time the report was written. Deer movement data contained in this report reflect all radio-collared study deer on the South Platte now that monitoring of radio-collared deer there is finished, and movement data for all South Platte study deer are available. Thus, some of the movement data presented in previous reports (Kufeld 1990) are repeated herein. Analysis of certain other aspects of the South Platte and Arkansas River deer data is not yet complete, and those results will be included in future reports. �14 The area which encompassed all observed locations for an individual deer during a specific time period was determined using the minimum convex polygon (MCP) feature in version 1.2 of program McPaa1 developed at the Smithsonian Institution, Front Royal, VA 22630. Median activity centers for each such group of locations were computed according to Berry et a1. (1984). Distances along the river channel between selected points, herein referred to as "river km", airline distances between selected deer locations or median activity centers (MACs) and the closest point on the river channel, and airline distances between selected deer locations or between median activity centers were computed using version 6.03 of program SAS, SAS Institute Inc., SAS Circle, Cary, NC 27512. Differences between species or sexes of deer in size of seasonal areas occupied, or in distances of median activity centers from the river were compared using non-parametric, Mann-Whitney-Wi1coxon procedures (Gibbons 1985) with probabilities of s 0.0500 considered significant. RESULTS Movements of Deer Tagged as Fawns Since fawns were not radio-collared, but only eartagged, movement data from each fawn, in most cases, consists of only 1 location when the animal was reported seen once alive or after being killed. Several deer, eartagged as fawns, however, were seen alive more than once. One or more fawn recoveries are available for 4 mule deer females, 8 mule deer males, 5 whitetail females, and 11 whitetail males. Two of the 4 female mule deer tagged as fawns were recovered in the South Platte Riverbottom only 1.0 and 1.3 km from their capture locations 678 and 441 days, respectively, after capture. The other 2 were recovered 54.5 and 26.3 km from their capture locations after 695 and 296 days. These 2 were recovered on the plains many km from the river. All 8 mule deer males tagged as fawns were recovered relatively long distances from where tagged (range - 17.5 to 149.3 km, median - 51.2 km). The 2 which moved the shortest and longest distances were in the South Platte Riverbottom when recovered. They were recovered after 332 and 310 days. Five others were seen alive and/or killed on the plains, many km from the river, varying in time from 158 to 878 days after capture. One other male mule deer, tagged as a fawn, left its capture site on the south edge of Greeley, and was observed alive 7 times along the Cache la Poudre River in Fort Collins, from 35 to 48 km from its capture site, between 587 and 1136 days after capture. Two of the 5 female white-tailed deer tagged as fawns were recovered in the South Platte Riverbottom 1.6 and 3.5 km from their capture locations 315 and 277 days respectively after capture. Another was seen alive in the riverbottom 88 days after capture 14.7 km from the capture site. The other 2 were killed on the plains 32.2 and 136.0 km from their capture sites after 296 and 1333 days respectively. Ten of the 11 male white-tailed deer tagged as fawns were recovered in the South Platte Riverbottom, although most were recovered long distances from their capture sites. Five of the 10 were recovered between 58.6 and 103.5 km from their capture sites. One of the 5 was recovered 145 days after capture, another after 245 days, and recoveries for the other 3 ranged from 1,013 to �1,380 days after capture. Four of the 10 were recovered between 14.0 and 33.3 km from their capture sites. Two of the 4 were recovered 286 and 304 days after capture and 2 were recovered after 1,412 and 1,776 days, respectively. One of the 10 recovered in the riverbottom was killed 80 days after capture 6.7 km from the capture site. The whitetail male tagged as a fawn that was not recovered in the riverbottom was killed on the plains after 291 days, 130.0 km from its capture site in a direction almost perpendicular to the course of the river. Movements of Deer Radio-Collared as Adults Movement data from 26 radio-collared mule deer and 24 radio-collared whitetailed deer (Tables 1 through 4), instrumented at 18+ months of age, suggest that adult deer in the South Platte riverbottom fall into 4 movement categories. Although the number of radio-collared deer sampled was relatively large (50 total), when deer were categorized by sex, species, and movement category, the number of instances where sample sizes were large enough to permit statistical comparisons by sex or species within a movement category was limited (Tables 1 and 2). Table 1. Movement patterns of radio-collared Mule deer tagged in the South Platte riverbottom. Movement category MCP deer location area ~Km22" Amual Winter SlmIIer Distance from MAC to river ~Kmt Amual Winter SlmIIer Distance between winter and sUllllerMACS (1CIn) Sex ID No. F 9253 9263 9321 9332 9341 9521 9522 9531 9540 9551 9652 9671 9732 9839 9889 3.9 2.9 2.6 3.3 2.3 4.8 2.7 17.2 3.1 5.3 3.1 5.7 6.6 7.7 9.4 0.2 0.1 0.0 0.0 0.1 0.2 0.2 0.2 0.2 0.2 0.1 0.2 0.2 0.4 0.3 M 9382 9642 3.1 16.2 0.2 0.3 F 9291 9539 9631 9970 0.5 0.6 11.1 1.6 1.9 3.6 6.4" 0.9 0.4 0.2 0.3 0.2 5.6 16.7 26.0 1.7 6.8 17.8 29.1 1.9 M 9452 9651 9701 3.1 4.8 5.2 11.4" 31.4 3.5 0.7 0.3 0.0 1.6 14.7 0.9 1.5 14.8 1.6 4 F 9859 9869 • Mini_ convex polygon area which enc~ssed 3 28.2 138.5 2.3 8.3 all deer locations (Km.2). b Distance from median activity center (MAC) to the nearest point on the river channel (Km). Median activity centers were computed as described by Berry et al. (1984). " This deer spent 2 summers in this area. �16 Movement Category l.--This category includes deer which remain in or near the South Platte Riverbottom throughout the year, and move about, fairly regularly, but with certain spots they frequented more than others, within a segment of riverbottom varying from 2 to 20 km in length. Most of the radiocollared deer of both species, 65% of the mule deer and 63% of the whitetails, were in this category. Minimum convex polygon (MCP) areas used by whitetail females and males in movement category 1 were not significantly different in size (P - 0.112). Likewise there was no difference (P - 0.240) between category 1 whitetail males and females in distance of their median activity centers (MACs) from the river channel. This same comparison was not possible for mule deer because Table 2. Movement category Movement patterns of radio-collared white-tailed deer tagged in the South Platte riverbottom. MCP deer location area ~1Cm2t Amual Winter SURlIer Distance from MAC to river ~lCmt Amual Winter Sunner Sex ID No. F 9252 9262 9271 9282 9301 9731 9741 9921 9971 15.9 6.1 5.9 2.0 7.2 4.2 26.9 8.3 6.8 0.3 0.3 0.3 0.4 0.2 0.3 0.2 0.1 0.3 M 9351 9361 9391 9411 9660 9nO 6.2 7.7 18.5 15.3 46.8 10.8 0.3 0.3 0.2 0.1 0.2 0.1 2 F 9311d 9511d 3 F 9781 9791d 9809 9951 6.0· 10.4 13.4 M 9370 9780 13.7 7.8 F 9800 4 Distance between winter and sURlIer MACS (Iem) 0.5" 0.4" 3.r 2.2' 12.9 19.3 3.1 0.9' 0.3 0.2 0.1 0.4 36.0 6.1 34.3 6.4 0.3 2.1 0.5 0.1 19.2 17.3 23.8 28.6 31.9 • Mininun convex polygon area which enc~ssed 40.8 all deer locations (1Cm2). b Distance from median activity center (MAC) to the nearest point on the river channel (Km). activity centers were computed as described by Berry et al. (1984). Median " Mean distance from 2 or more median activity centers to the nearest point on the river channel (Km). Median activity centers for ID No. 9311 are Masters and Orchard, and for ID No. 9511, Orchard, Messex State Wildlife Area, and Dune Ridge State Wildlife Area. d Movements of this deer are described in the text. • Although both of these deer usually moved to a different area during sURlIer there was 1 sURlIer they did not IIigrate. All locations whi le the deer were in the area were used to determine area size. , Deer ID No. 9781 spent 2 sumers in this area. Deer ID No. 9951 spent 3 sURllers in this area. Deer ID No. 9809 was radio-tracked to this area during the first sumner after instrumentation. The transmitter failed the next winter after the deer had returned to the river. It was observed on the same sumner area during the 4th and 5th sumers after instrumentation. �17 only 2 males were in movement category 1, but MCP sizes and MAC distances from the river channel shown for those 2 males (Table 1) are within the ranges observed for females. Thus, for further analysis males and females of each species in movement category 1 are combined. Mule deer in movement category 1 (Table 1) used median MCP areas of 3.9 km2, and median MACs were 0.2 km from the river channel. White-tailed deer in movement category 1 (Table 2) used median MCP areas of 7.7 km2, and median MACs were 0.3 km from the river channel. Thus, median MCPs for white-tailed deer were significantly larger than for mule deer (P - 0.012), but differences between the species in distance of MACs from the river channel were not significant (P = 0.091). Table 3. NUlber of locations for radio-collared aile deer, and dates when radio-collared IUle deer in movement category 3 were away from the riverbottom during late spring or summer or early fall. Movement Category 3 4 No. of locations Winter Surmer Dates of sunner elains I2!riod Year 1 Year 2 Sex 10 No. Amual. F 9253 9263 9321 9332 9341 9521 9522 9531 9540 9551 9652 9671 9732 9839 9889 54 55 69 79 80 49 29 103 45 32 32 64 16 105 98 M 9382 9641 35 98 F 9291 9539 9631 9970 18 8 23 11 10 13 21 4 6-29 5-25 5-30 7-17 M 9452 9651 9701 32 10 8 20 14 12 5-28 to 8-17 6-30 to 10-2S5-30 to 10-13 F 9859 9869 to 11-1 to 10-22'" to 10-13 to 8-1S- 6-10 to 10-28 6-19 to 11-28 69 102 • Deer was found dead or was killed on the plains away from the riverbottom as of the last date. Movement Category 2.--This category includes deer which remain in or near the riverbottom throughout the year, but tend to utilize 2 or more localities which are separated by distances of more than 15 km. In this study several such localized use areas for one deer were spaced over a section of riverbottom 121 km in length. Weeks or months are spent in each locality. Upon leaving one locality the trip to the next one is made rapidly. Individual deer were observed to return to the same locality several times during the monitoring period. This category contained 8% of the radiocollared white-tailed deer, but no mule deer were in this category. �18 White-tailed doe ID no. 9311 in movement category 2 (Tables 2 and 4) occupied 2 localities in the South Platte Riverbottom separated by 17 river km. One area, (1.7 km2) was near Masters, Colorado, and the other (2.3 km2) was east of the town of Orchard. The deer was radio-collared 1-7-88 near Masters and was monitored until 12-31-90. It spent most of its time, including winters, near Masters, but each summer or fall for 3 years it made 1 or 2 trips to the Orchard area. After each visit to Orchard it quickly returned to Masters. Periods during which it was located east of Orchard were 8-17-88 to 8-29-88, 9-22-88 to 10-18-88, 8-4-89 to 10-5-89, and 7-25-90. White-tailed doe, ID no. 9511, in movement category 2 (Tables 2 and 4) had a complex movement pattern which involved several centers of activity, all of which were located in the riverbottom but separated by relatively long distances. The deer was radio-collared 1-18-87, 1.9 river km downstream from Weldona, Colorado. Between 3-6-87 and 10-13-87, 14 locations were in an area of 0.3 km2 centered 22.7 river km upstream from its trapsite and 4.2 river km Table 4. Numer of locations for all deer, and dates when radio-collared white-tailed deer in IIOvement category 3 were away from the riverbottom during late spring or summer or early fall. Movement category Sex ID No. No. of locations Amual Winter Summer Year 1 F 9252 9262 9271 9282 9301 9731 9741 9921 9971 15 24 80 78 82 15 105 48 17 M 9351 9361 9391 9411 9660 9770 25 56 22 81 97 22 2 F 9311 9511 81· 66" 3 F 9781 9791 9809 9951 63 21 27 81 15 6 13 6-1 to 7-29 5-30 to 7-29 M 9370 9780 10 18 5 7 7-15 to 9-~ 6-30 to 9-12 F 9800 81 4 --" b -• Dates of summer elains ~riod Year 2 Year 3 6-27 to 10-5 -" b Year 4 6-7 to 9-24 - 5-31 to 7-12 e 6-7 to 7-25d " Movements of this deer are described in the text. b ID No. 9781 did not migrate from its riverbottom location area during its 1st summer and ID No. 9951 did not leave its riverbottom location area during its 2nd summer. C Although its transmitter was dead by then ID No. 9809 was seen in its summer area on 6-30 of its 4th summer after instrumentation and again on 8-15 of its 5th summer. • Deer was fCUld dead or killed or its transmitter died while it was on the plains away from the riverbottom as of the last date. �19 upstream from Orchard, Colorado. Between 9-12-87 and 10-13-87 it moved downstream 120.6 river km to a point near Iliff, Colorado. Between 10-27-87 and 5-28-98, 42 locations were in an area of 8.6 km2 centered 17.0 river km upstream from its easternmost location near Iliff. This area was in the vicinity of Dune Ridge State Wildlife Area, which is located about 8 km southwest of Sterling, Colorado. During the 10-27-87 to 5-28-89 period, however, it made 2 trips to a point centered 32.6 river km upstream near Messex State Wildlife area. It was located there 3 times between 11-1 and 1129-88, and 3 times between 3-28 and 4-24-89. Those 6 locations were in an area of 0.9 km2• After each visit to the Messex area it returned to the vicinity of Dune Ridge and it died there on 5-28-89. Movement Catesory 3.--This category contained the 2nd highest number of radiocollared deer with 27% of the mule deer and 25% of the whitetails. It includes deer that spend most of the year, including winter, in or near the riverbottom, but during late spring or summer they leave the riverbottom and spend at least several weeks out on the plains. During the period such deer were in the riverbottom they tended to move about, fairly regularly, within a segment of riverbottom varying from 2 to 16 km in length. Radio-collared deer that exhibited this behavior returned to the same section of riverbottom sometime before December. Those that survived more than 1 year also occupied the same general area each summer (Tables 1 and 2). Departure and return dates, length of time these radio-collared deer spent away from the riverbottom, and distance they travelled from the river varied among individual deer within the 3rd category. Some individual deer were gone from the riverbottom for only a few weeks, while others were gone for months (Tables 3 and 4). Some individuals moved to plains habitats only 1 or 2 km from the riverbottom, while others travelled to plains habitats more than 35 km from the riverbottom (Tables 1 and 2). An exception to the above described pattern involved two whitetail does in category 3, monitored for 3 and 4 years, respectively, which had 1 summer when they stayed in the riverbottom and did not visit the plains. Females of both species in category 3 gave birth to their fawns after reaching their plains destination. The fawns accompanied the mothers on the return trip to the riverbottom. Median MCP areas containing locations for deer in movement category 3 while they were in the riverbottom during winter (Tables 1 and 2) were significantly larger for white-tails than for mule deer (white-tailed deer - 9.1 km2, mule deer - 3.1 km2, P - 0.0383). However, HACs for both white-tailed and mule deer were a median distance of 0.3 km from the river channel. During the period deer in movement category 3 were away from the riverbottom in late spring, summer or fall, the size of mule deer MCPs varied widely (Table 1). MCPs for white-tails listed in Table 2 were less variable in size when away from the riverbottom than MCPs exhibited by mule deer (Table 1). However, when white-tailed deer ID number 9791, whose movements are described in detail in the following paragraph, is considered it becomes apparent that white-tailed deer summer HACs were also quite variable in size. When deer in movement category 3 left the riverbottom for a trip to the plains whitetails generally moved farther from the river than did mule deer. Median distance of white-tailed deer HACs from the river channel while deer were away from the riverbottom, not including deer 1D 9791, was 17.3 km compared with 5.6 km for mule deer .. The median distance for white-tails is conservative because deer �20 ID 9791 moved very long distances from the river during summer, as in the following paragraph, and roamed so widely during its summer the plains that its MAC could not be determined. Likewise, median between winter and summer MACs was farther for whitetails (23.8 km including deer 10 9791) than for mule deer (6.8 km). described forays to distance not White-tailed doe 10 9791, in movement category 3, (Tables 2 and 4), spent the winter in the riverbottom in a relatively small area, but during summer it travelled such long distances that it could rarely be found. It was radiocollared 2-3-87, 5.9 river km downstream from Hardin, Colorado. It spent the winters of 1986-87 and 1987-88 in the riverbottom in the vicinity of its trapsite. Between 4-17 and 9-12-87 it could not be located despite an intensive search by aerial telemetry of the area encompassed by a perimeter extending along Interstate 25 from the Wyoming State Line south to the u.S. Air Force Academy, east to Limon, north to the Wyoming State Line, and west to Interstate 25. The South Platte River from Platteville to Julesburg, Colorado, was also searched. On 9-12-87 the deer was back in the vicinity of its trapsite where it spent the winter. It left its winter area the following April and on 4-29-88 it was located near Roggen, Colorado, 17.5 km from its winter MAC and 13.5 km from the nearest point on the South Platte River Channel. On 7-25-88 it was relocated 19.8 km south-southeast of Akron, Colorado, 105.9 km from its winter median activity center and 51.6 km from the nearest point on the South Platte River channel. Two weeks later it could not be found within a 24 km radius of that location, and it was never again relocated. Movement Cate20rv 4.-- This category includes deer that live on the plains and rarely visit the riverbottom. Several radio-collared deer were trapped in the riverbottom, but shortly after being tagged they moved out onto the plains and were rarely or never again located in the riverbottom during up to 4 years of monitoring (Tables 1 and 2). This category contained 8% of the radiocollared mule deer and 4% of the radio-collared whitetails. Three deer, 2 mule deer (Tables 1 and 3) and 1 whitetail (Tables 2 and 4), were in movement category 4. These deer had relatively large annual MCPs compared to deer in movement category 1 (Tables 1 and 2). One of the mule deer, 10 9869, roamed throughout its 138.5 km2 annual MCP from approximately early November until early April during each of 4 winters, but each summer it limited its movements to an agricultural area where its locations were within a MCP of 11.8 km2• Each year from mid August until corn was harvested around early November this deer stayed in one particular cornfield where its locations were within a MCP of 1.6 km2• This deer was tagged in the riverbottom but never again located there during the 4 years it was monitored. The other mule deer in movement category 4 returned to the riverbottom 3 times during the 2.6 years it was monitored, but each visit lasted 1 month or less. The white-tailed deer in movement category 4 left the riverbottom on 5-4-87 after being instrumented on 2-4-87 and had never returned during the next 3.1 years that it was monitored. The MAC for the white-tailed deer in movement category 4 was several times farther from the river channel than MACs for the 2 mule deer (Tables 1 and 2). �21 DISCUSSION Deer Movements Dispersal by Young Deer--Instances of dispersal from maternal home ranges to form new home ranges have been reported for young mule deer, black-tailed deer (Robinette 1966, Severson and Carter 1978, Bunnell and Harestad 1983, Eberhardt et al ..1984), and white-tailed deer (Nelson and Mech 1984, Tierson et al. 1985, Dusek et al. 1989, Nixon et al. 1991). Some white-tailed deer populations are relatively sedentary and apparently dispersal does not occur or is uncommon (Thomas et al. 1964, Larson et al. 1978, Inglis et al. 1986, Mooty et al. 1987,). Even in those deer populations where dispersal has been reported varying portions of the young animals did not disperse. Most fawns of both species (mule and white-tailed deer) and both sexes, tagged in the South Platte Riverbottom, tended to disperse and they dispersed relatively long distances from where captured during their first winter. Both species dispersed to plains and riverbottom habitats, but mule deer exhibited more tendency to disperse to the plains than did white-tails. Whitetails more often dispersed to other parts of the riverbottom. Long range dispersal of white-tailed deer fawns captured in plains riverbottom habitats in eastern Colorado was also documented by Colorado Division of Wildlife, Southeast Region (Unpublished data), in a study conducted at Bonny Reservoir near the Colorado and Kansas State Line. Four white-tailed deer tagged as fawns, including 2 males and 2 females, were subsequently recovered 66, 145, 230, and 233 km from the capture site. Not all whitetails tagged as fawns dispersed, however, as 2 others, 1 male and 1 female, were subsequently recovered near the capture site. Recoveries in that study and in our study occurred in all directions from the capture sites. Although dates of recoveries were sometimes several years after capture, data from both studies suggests that dispersal for both species of deer, if it occurs, tends to occur late in the first year or during the second year of life. The tendency for dispersal to occur at the age of 1 to 2.5 years has been reported by others (Robinette 1966, Kammermeyer and Marchinton 1976, Severson and Carter 1978, Bunnell and Harestad 1983, Nelson and Mech 1984, Tierson et al. 1985, Dusek et al. 1989, Nixon et al. 1991). Nonmigratory Behavior by Adult Deer--Most of the adult radio-collared deer of both species were non-migratory. This includes all deer in movement categories 1 and 4, which together contain 73% of the radio-collared mule deer and 67% of the white-tails. When the deer in movement category 4 were in their plains home ranges they exhibited no migratory tendencies. The radiocollared whitetail doe in movement category 4 was tagged at 1.5 years of age and moved, at about 2.0 years of age, 40.8 km to its plains home range where it lived in a MCP area of 31.9 km2 for the next 3 years. That movement was considered to represent dispersal rather than migration. Initial movement to the plains from the riverbottom, where captured, by the 2 mule deer does in movement category 4 could have represented dispersal or simply a return to the plains from trips of short .duration to the riverbottom. Both deer were at least 2 years old when captured, so it may be more likely that their presence in the riverbottom represented short duration trips. �22 Migratory Behavior by Adult Deer--A minority of the adult radio-collared deer of both species were migratory. This includes all deer in movement categories 2 and 3, which together contain 27% of the radio-collared mule deer and 33% of the whitetails. All of the migratory mule deer wintered in the riverbottom and spent a period of time on the plains during late spring, summer, and sometimes into fall. Most of the migratory white-tails (those in movement category 3) also followed a pattern of migration, in late spring, or summer, to the plains with return later in summer or in the fall, but others (those in movement category 2) migrated at various times of the year to distant sites within the .riverbottom. Migratory deer of both species in movement category 3 exhibited strong fidelity to their riverbottom and plains home ranges, returning to the same area each winter and summer. Migratory white-tails in movement category 2 also exhibited strong fidelity to their riverbottom home ranges by moving back and forth among their respective areas. Strong fidelity of migratory mule and white-tailed deer to respective winter and summer home ranges are consistent with findings of Gruell and Papez (1963), Zalunardo (1965), Robinette (1966), Verme (1973), Bertram and Rempel (1977), Carpenter et al. (1979), Nelson and Mech (1981), Tierson et al. (1985), Garrott et al. (1987), Loft et al. (1989), Thomas and Irby (1990), and Nixon et a1. (1991). Migratory mule and white-tailed deer females in movement category 3 usually gave birth to fawns during the period they were away from the riverbottom. Extent of Movement. Mule Deer Vs Whitetails--Adult radio-collared white-tailed deer in this study were observed to occupy larger MCP areas and to have longer movements in general than did the adult instrumented mule deer. This trend was exhibited by whitetails in movement categories 1, 3, and 4. Movement category 2 contained only whitetails and those deer moved relatively long distances. Our findings also suggest that white-tailed deer in plains riverbottom habitats in eastern Colorado are considerably more mobile than whitetails in most other parts of the U. S. This includes more mobility in terms of length of fawn dispersal, and home range size, and distance of migration of adult animals. Whitetails in south Texas brushland ranges and pine-oak habitats of Alabama and Florida are reported to be nonmigratory and occupy home ranges averaging under 1.0 km2 (Thomas et al. 1964, Inglis et al. 1986, Marchinton and Jeter 1966). White-tails in conifer and deciduous forests of New York, and Northern Minnesota are reported to be migratory, but winter and summer home ranges averaged less than 2.5 km2, and the longest instance of migration by an individual deer in those studies was 40 km (Tierson et al. 1985, Hoskinson and Mech 1976, Nelson and Mech 1981). Average distances between winter and summer ranges were much shorter. Rongstad and Tester (1969) reported winter home ranges 1.6 to 4.9 km2 in size for migratory whitetails in a cedar swamp in east central Minnesota. Populations of nonmigratory whitetails, living in average home ranges of less than 2 km2• in fir-aspen-birch habitat in northern Minnesota and woodland and swamp habitat in Wisconsin, were observed by Mooty et al. 1987, and Larson et al. 1978. In a cedar-firfir-spruce forest in northern Michigan and a hardwood forest in northeast Missouri white-tailed deer migrations averaging 13.8 and 17.1 km, respectively, have been recorded (Verme 1973, Root et al. 1990). In east central Illinois where 64% of the area was cropland and 36% hardwood forest Nixon et al. 1991) found that 80.4% of whitetails were nonmigratory and occupied home ranges from 3.0 to 4.9 km2 in size. The other 19.6% of that �population, which was migratory, occupied winter and summer ranges of less than 5.0 km2 and separated by an average of 13.0 km. In prairie agricultural habitat of eastern Montana Dusek et al. (1989) reported a nonmigratory whitetail population which occupied average home ranges of 4.0 km2. In the lower Yellowstone Riverbottom of eastern Montana, which supported cottonwoodwillow habitat very similar to plains riverbottom habitats in eastern Colorado, Dusek et al. (1989) found that white-tailed deer were, with very few exceptions, nonmigratory, and occupied relatively small home ranges with average activity radii of 0.47 km for females and 0.91 km for males. They also reported that, of those young deer that dispersed, females moved an average of 18.5 km and males moved 19.5 km. These are considerably shorter distances than observed for dispersing whitetails in eastern Colorado. More similar movements of white-tailed deer to those in our study were reported by Sparrowe and Springer (1970) in cottonwood-willow riverbottom habitat adjacent to agricultural lands in eastern South Dakota. There, whitetails migrated between winter and summer home ranges, an average of 23 km within the riverbottom, in a pattern similar to deer in our movement category 2. They occupied winter home ranges averaging 7.0 km2 and summer home ranges averaging 2.6 km2. Wood et al. (1989) also reported movements of white-tailed deer in prairie habitat in southeastern Montana that were similar to those in our study. In their study female white-tailed deer exhibited individual patterns of movement that ranged from relatively small, stable home ranges to erratic shifts within very large home ranges. Their study long MCPs ranged from 3.5 to 80.3 km2, and dispersal of whitetails, tagged as fawns, of up to 93 km was recorded. Mule deer in our study exhibited movement patterns that were somewhat unique compared to those reported in other studies of mule deer on the Great Plains. Wood et al. (1989) also reported that the mule deer population in plains habitats of eastern Montana contained both nonmigratory and migratory individuals. However, a higher portion of those deer were migratory (71%) than in our study (27%). Severson and Carter (1978) reported that in plains habitats of South Dakota nonmigratory populations of mule deer living in rougher terrain and more extensive cover types had smaller home ranges than those living in open, level areas, and suggested that the former habitat conditions allowed those deer to fulfill requirements in smaller areas. Kufeld et al. (1989) also suggested that most deer living along the Rocky Mountain foothills in northcentral Colorado were nonmigratory and occupy small home ranges because of good quality habitat. Thus, the higher proportion of nonmigratory mule deer in our study may have been a result of better quality habitat in the riparian zone along the South Platte River compared to the open plains habitats reported by Wood et al. (1989). Accordingly, Gerlach (1987) observed a nonmigratory mule deer population in open grassland and pinyonjuniper woodland habitat of southeastern Colorado which occupied average home ranges of 12.2 km2• Those home ranges are 3 times larger than the median MCP areas occupied by movement category 1 mule deer along the South Platte River in our study. �24 LITERATURE CITED Berry, K. J., P. W. Mielke, and K. L. Kvamme. 1984. Efficient permutation procedures for analysis of artifact distributions. Pages 54-74 in H. J. Hietala, ed. Intrasite spatial analysis in achaeology. Cambridge Univ. Press. Cambridge, U. K. Bertram, R. C., and R. D. Rempel. 1977. Migration of the North Kings deer herd. California Fish and Game. 63-157-179. Bunnell, F. L., and A. S. Harestad. 1983. tailed deer: models and observations. Dispersal and dispersion of blackJ. Mammal. 64:201-209. Carpenter, L. H., R. B. Gill, D. J. Freddy, and L. E. Sanders. 1979. Distribution and movements of mule deer in Middle Park, Colorado. Div. Wildl. Spec. Rep. 46:1-32. Colo. Dusek, G. L., R. J. Mackie, J. D. Herriges, Jr., and B. B. Compton. 1989. Population ecology of white-tailed deer along the lower Yellowshone River. Wildl. Monogr. No. 104. 68pp. Eberhardt, L. E., E. E. Hanson, and L. L. Cadwell. 1984. Movement and activity patterns of mule deer in the sagebrush-steppe region. J. Mammal. 65:404-409. Garrott, R. A., G. C. White, R. M. Bartmann, L. H. Carpenter, and A. W. Alldredge. 1987. Movements of female mule deer in northwest Colorado. J. Wildl. Manage. 51:634-643. Gruell, G. E., and N. J. Papez. 1963. Movements of mule deer in northeastern Nevada. J. Wildl. Manage. 27:414-422. Inglis, J. M., B. A. Brown, C. A. McMahan, and R. E. Hood. 1986. Deer-brush relationships on the Rio-Grande Plain, Texas. The Caesar K1eberg Research program in Wildlife Ecology. The Texas A. Exp. Sta., Texas A & M Univ. 80pp. Kammermeyer, K. E., and R. L. Marchinton. 1976. white-tailed deer. J. Mammal. 57:776-778. Notes on dispersal of male Kufe1d, R. C. 1987. Development of census methods for deer in plains riverbottom habitats. Colo. Div. Wildl.; Wildl. Res. Rep. July (1):1319. Kufe1d, R. C. 1988. Development of census methods for deer in plains riverbottom habitats. Colo. Div. Wildl., Wi1dl. Res. Rep. July (1):1121. Kufe1d, R. C. 1989. Development of census methods for deer in plains riverbottom habitats. Colo. Div. Wi1d1., Wi1d1. Res. Rep. July (1):1117. �Kufeld, R. C. 1990. Development of census methods for deer in plains riverbottom habitats. Colo. Div. Wildl., Wildl. Res. Rep. July:7-l4. Kufeld, R. C. 1991. Development of census methods for deer in plains riverbottom habitats. Colo. Div. Wildl., Wildl. Res. Rep. July:9-l8. Kufeld, R. C., D. C. Bowden, and D. L. Schrupp. 1989. Distribution and movements of female mule deer in the Rocky Mountain foothills. J. Wildl. Manage. 53:871-877. Larson, T. J., O. J. Rongstad, and F. W. Terbilcox. 1978. Movement and habitat use of white-tailed deer in southcentral Wisconsin. J. Wildl. Manage. 42:113-117. Loft, E. R., R. C. Bertram, and D. L. Bowman. 1989. Migration patterns of mule deer in the central Sierra Nevada. Calif. Fish and Game 75:11-19. Mooty, J. J., P. D. Karns, and T. K. Fuller. 1987. Habitat use and seasonal range size of white-tailed deer in northcentral Minnesota. J. Wild1. Manage. 51:644-648. Nelson, M. E., and L. D. Mech. 1981. Deer social organization and wolf predation in northeastern Minnesota. Wildl. Monogr. No. 77. 53pp. Nelson, M. E., and L. D. Mech. 1984. Home-range formation and dispersal of deer in northeastern Minnesota. J. Mammal. 65:567-575. Nixon, C. M., L. P. Hansen, P. A. Brewer, and J. E. Chelsvig. 1991. Ecology of white-tailed deer in an intensively farmed region of Illinois. Wildl. Monogr. No. 118. 77pp. Robinette, W. L. 1966. Mule deer home range and dispersal in Utah. Wildl. Manage. 30:335-349. J. Severson, K. E., and A. V. Carter. 1978. Movements and habitat use by mule deer in the northern Great Plains, South Dakota. Proc. Internatl. Rangeland Congr. 1:466-468. Sparrowe, R. D., and P. F. Springer. 1970. Seasonal activity patterns of white-tailed deer in eastern South Dakota. J. Wildl. Manage. 34:420431. Thomas, J. W., J. G. Teer, and E. A. Walker. 1964. Mobility and home range of white-tailed deer on the Edwards Plateau in Texas. J. Wildl. Manage. 28:463-472. Thomas, T. R., and L. R. Irby. 1990. Habitat use and movement patterns by migrating mule deer in southeastern Idaho. Northwest Sci. 64:19-27. Tierson, W. C., G. F. Mattfeld, R. W. Sage, Jr., and D. F. Behrend. 1985. Seasonal movements and home ranges of white-tailed deer in the Adirondacks. J. Wildl. Manage. 49:760-769. �26 Verme, L. J. 1973. Movements of white-tailed deer in upper Michigan. Wi1d1. Manage. 37:545-552. J. Wood, A. K., R. J. Mackie, and K. L. Hamlin. 1989. Ecology of sympatric populations of mule deer and white-tailed deer in a prairie environment. Wildlife Division, Montana Dept. of Fish, Wildlife and Parks. 97pp. Za1unardo, R. A. 1965. The seasonal distribution of a migratory mule deer herd. J. Wi1d1. Manage. 29:345-351. Prepared by Ual<C</ c_ 2,{jrttl Roland C. Kufe1d Wildlife Researcher C �27 Colorado Division of Wildlife Wildlife Research Report July 1992 JOB PROGRESS REPORT State of ~C~o~l~o~r~a~d~o~_ Project No. Work Plan No. Job No. Deer Investigations 2 Compensatory Effects of Harvest in a Mule Deer Population 15 Period Covered: Author: Mammals Research W-153-R-5 July 1, 1991··-June 30, 1992. R. M. Bartmann, G. C. White. Personnel: A. W. Alldredge, G. D. Bear, K. Boden, R. W. Bumgardner, W. Devergie, B. L. Dupire, J. Fair; D. J. Freddy, J. Frothingham, V. K. Graham, J. P. Gray, R. Harthan, E. Hein, M. J. Hooker, S. M. Joy, D. G. Leslie, A. A. Lippacher, B. C. Lubow, C. G. McDonald, Jr., J. D. Madison, J. E. Morris, J. R. Morrison, K. W. Robinette, S. L. Schulz, M. W. Sherman, T. Shenk, M. A. Stremel, D. G. Saltz, C. L. Vardaman, E. White, M. Wotawa, M. A. Zablan. and numerous others from the Colorado Division of Wildlife and Colorado State University. ABSTRACT Estimates of deer density from aerial line transects were 47.2 and 23.6 deerfkm2 on the control and treatment units, respectively.·. Only 28 fawns were radio-collared on the control unit and 42 on the treatment unit wi·th estimated survival rates of 0.456 (SE 0.107) and 0.425 (SE 0.084) on the same respective units. Adult doe survival rates from 1 December 1991 to 1 July 1992 were 0.790 (SE 0.066) and 0.684 (SE 0.077) on the control and treatment units, respectively. Twelve deer were located 1-3 times during the winter off the unit where they were trapped. However most were from trapsites close to unit boundaries and most movement was from the treatment to the control unit. The late season harvest by 350 license holders was estimated as 345 deer of which 34.5% were fawns. Fawns captured on the treatment unit were heavier and had higher weight/length ratios than those on the control l S 0.023), but no differences were found for total body length or left hind foot length (l ~ 0.249). No body condition indices for adult does or yearling bucks were found to differ between units (l ~ 0.178). ��29 COMPENSATORY EFFECTS OF HARVEST IN A HOLE DEER POPULATION Richard M. Bartmann and Gary C. White P. N. OBJECTIVES 1. Increase the winter survival rate of mule deer fawns by lowering deer density to reduce competition for forage during winter. total 2. Increase the harvest rate of deer through increased productivity of adult does and decreased natural mortality of fawns resulting from closer alignment of population size with carrying capacity. SEGMENT OBJECTIVES 1. Reduce the winter population of mule deer on the Ridge treatment density <40jkm2 and maintain the density for at least 3 years. area to a 2. Estimate winter areas. 3. Estimate annual survival rates of adult females. 4. Estimate annual survival rates of adult males. 5. Estimate productivity 6. Estimate areas. harvest 7. Estimate condition 8. Estimate age structure condition of yearling 9. Estimate age structure of adult males and condition control and treatment areas. survival rates of fawns on control and treatment of deer on control and treatment areas rates of bucks, does, and fawns on control of fawns on control and treatment and treatment areas. of adult females on the treatment area and females on control and treatment areas. of yearling males on METHODS Methods remained the same as previously reported (Bartmann 1990; Bartmann and White 1991). Several Objectives have been deemed unattainable. Objectives 4 and 9, relating to survival and age structure of adult bucks, cannot be met because sample sizes are inadequate and there is no practical means to increase them. The rationale for dropping Objective 5, estimating productivity, was given in the report for the previous Segment (Bartmann and White 1991). �30 RESULTS AND DISCUSSION Population Reduction Aerial line transects were flown on the Ridge study area 8-10 January 1992. One additional set of flights was made on the treatment unit to increase the sample of deer groups. Estimated deer density on the treatment unit (23.6 deerfkm2) was not significantly lower (f - 0.271) than in 1990 (36.3 deerfkm2), although it was lower (f - 0.007) than on the control (47.2 deerfkm2) (Fig. 1) and the downward trend that began in 1989 is continuing. I _;"Ht'" Estimated deer density on CONTROL TREATHENT the control unit in 1991 was ,.... 120 r------=:,....----------------, .... _._-_ .. _--•..........•... _._ ..............•••.•.•.•.•.•.....•...................•...•...••••••••.•....•...............•...•... N much lower (f - 0.003) than * 100 in 1990 (85.6 deerfkm2). The * lC: >-: ...•.• reason for the erratic 60 I:t: 1>'1 changes in deer density 1!4 60 :::::::: i;;!~~:~····· ~ estimates on the control 40 ::::::::::1:::::::::::: unit the past 3 years has ........... . yet to be discovered. 20 Survival and harvest data o L-~ __ ~ ~ __ ~ -L ~ __ ~~~ provide no insight to these 1985 1986 1987 1968 1969 1990 1991 large fluctuations and, YEAR based on radio-collared animals, there is no Fig. 1. Deer density estimates (w/95% evidence of large-scale confidence intervals) from aerial line transects movements between the 2 on control and treatment units during early units. winter. , Fawn Survival Deer were trapped 10-26 November 1991. The time-constrained trapping period (between the end of the 3rd regular season and the start of the late season on the treatment unit) along with poor trapping conditions again caused us to fall short of our objective of 160 fawns. Only 70 fawns were captured with more trapped on the treatment unit (42) than on the control unit (28) despite the much lower density on the former area (Table 1). Fawn survival tended to reflect the relatively mild winter conditions with similar rates estimated for both units (control - 0.456, SE 0.107; treatment 0.425, SE 0.084). Although estimated survival rates did not differ significantly (f - 0.514) between the 2 units, 9 of 22 fawn mortalities on the treatment unit were hunting related as opposed to none on the control unit (Table 2). Adult Doe Survival Sixty-six does were radio collared during November 1991 of which 48 were new deer and 18 had their radio collars replaced. This brought the total radiocollared adult population as of 1 December 1991 to 44 on the treatment unit and 41 on the control unit. An additional 12 does and 9 yearling bucks were processed and released without radio collars. �31 Table 1. Survival rate estimates (~) for radio-collared mule deer fawns on control and treatment units of the Ridge study area in Piceance Basin, Colorado, from time of collaring in November and December until the following 15 June 1982-83 through 1991-92. Winter 1982-83 1983-84 1984-85 1985-86 1986-87 1987-88 1988-89 1989-90 1990-91 1991-92 n 29 28 34 59 60 32 34 38 34 28 Contlol unit SE(~) ~ 0.321 0.071 0.196 0.537 0.431 0.241 0.270 0.758 0.320 0.456 0.088 0.049 0.078 0.070 0.064 0.077 0.083 0.078 0.090 o .107 f of equal ~(.t) Treatment !:!:nit ~ SE(~) 31 32 26 58 58 28 28 44 36 42 0.387 0.033 0.431 0.439 0.471 0.107 0.445 0.657 0.243 0.425 0.087 0.033 0.105 0.070 0.067 0.058 0.096 0.072 0.081 0.084 0.578 0.774 0.075 0.157 0.565 0.006 0.509 0.250 0.341 0.514 Table 2. Cause of mortality for radio-collared mule deer fawns on control and treatment units of the Ridge study area in Piceance Basin, Colorado, from time of collaring in November and December until the following 15 June 1982-83 through 1991-92. Percentages are of total uncensoredB fawns. No. of fawns Censored Unit Winter Control 1982-83 1983-84 1984-85 1985-86 1986-87 1987-88 1988-89 1989-90 1990-91 1991-92 29 28 34 59 60 32 34 38 34 28 Treatment 1982-83 1983-84 1984-85 1985-86 1986-87 1987-88 1988-89 1989-90 1990-91 1991-92 31 32 26 58 58 28 28 44 36 42 1 7 11 6 2 5 14 9 7 1 1 4 9 5 3 9 7 9 Starvation % No. 15 22 16 10 14 22 10 3 6 7 56 79 59 21 26 73 34 13 24 33 15 27 8 17 16 19 9 6 8 7 50 87 36 35 30 68 36 17 28 21 Mortality: cause Predation Hunting % % No. No. 4 4 5 7 17 15 14 19 15 31 5 17 10 3 40 14 2 3 2 11 13 1 1 7 10 9 22 25 4 4 3 2 10 6 2 5 9 9 8 14 31 27 Other % No. 1 7 3 2 7 3 3 3 4 15 6 7 24 13 12 14 2 7 3 1 1 5 5 4 4 4 14 2 2 18 20 11 14 12 B Uncensored fawns are those with nonfai1ing radios or with collars that did not drop off prematurely. �32 As during the previous 2 years, hunting was an important mortality cause influencing adult doe survival on the treatment unit (Table 3). Six does died from this cause compared to none on the control unit. Estimated survival on the treatment unit to 1 July 1992 (0.684, SE 0.077) was lower than on the control (0.790, SE 0.066), but these estimates are preliminary because adult survival is based on the interval 1 December-30 November. Table 3. Annual (1 Dec-30 Nov) survival rate estimates (~) for radio-collared adult female mule deer on control and treatment units of the Ridge study area in Piceance Basin, Colorado, 1982-83 through 1991-92. 'Winter 1982-83 1983-84 1984-85 1985-86 1986-87 1987-88 1988-89 1989-90 1990-91 1991-92a a !1 10 15 9 25 27 14 7 23 39 41 Survival Control unit SE(~) ~ 0.800 0.779 1.000 0.917 0.720 0.818 0.857 0.829 0.918 0.790 0.126 0.113 0.056 0.090 0.116 0.132 0.092 0.045 0.066 rate estimates !1 11 15 10 21 18 10 5 28 41 42 Treatment unit SE(~) ~ 0.909 0.929 1.000 0.900 0.811 1.000 0.800 0.672 0.630 0.684 0.087 0.069 0.067 0.099 0.179 0.089 0.088 0.077 f of equal ~ 0.448 0.271 1.000 0.821 0.590 0.329 0.854 0.087 0.008 0.228 for 1991-92 are only to 1 July 1992. Radio failures continued to occur more frequently than desired. Some were with radios due to expire from depleted batteries while others were carry-over problems from the past 2 years. Acquisition of new units should resolve this problem in the future. All radio-collared deer were located 3 times during winter to check for movements between the treatment and control units. Six, 8, and 4 deer were located off the units on which they were trapped during January, February, and March, respectively. Twelve different deer, 6 fawns and 6 adults, were involved and all but 2 moved from the treatment to the control unit. Most movements were from trapsites close to unit boundaries and only 1 deer was located during all 3 periods off the unit where it was trapped. Harvest The random surveys of harvest for the 3 late seasons on the treatment unit were rechecked and harvest estimates recalculated. This resulted in an increase of 5 deer and a decrease of 9 deer in the original estimates for 1989 and 1990, respectively (Table 4). In 1991, 350 license holders took an estimated 345 deer. The percent success of people that actually hunted increased from 69% in 1990 to 75% in 1991 but, over the 3 seasons, smaller percentages of successful hunters killed 2 deer. �33 Table 4. Hunter participation and success and deer harvest estimates for the December late season on the treatment unit of the Ridge study area, 1989-91. 1989 1990 % No. Licenses· Survey respondents Did not hunt Unsuccessful Harvested 1 deer Harvested 2 deer Harvest • - Does Fawns Total 375 60 81 38 56 200 (SE) Hunters were allowed Condition 326 130 456 21.6 10.1 14.9 53.4 71.6 28.4 (47.5) to take 2 antlerless No. 400 92 48 109 104 139 278 105 383 1991 % 12.0 27.2 26.0 34.8 No. 350 69 40 76 122 112 72.7 27.3 (40.0) 226 119 345 % 11.4 21. 7 34.9 32.0 65.5 34.5 (39.4) deer on a license . of Fawns Body condition of fawns captured with dropnets was indexed by weight, total body length, left hind foot length, and a weight/length ratio (Table 5). Fawns on the treatment unit had higher weights and weight/length ratios than those on the control (l ~ 0.023) with no differences found for body length and left hind foot length (l ~ 0.249). Age Structure and Condition of Does Age structures of does in the regular and late season harvests were estimated independently each year. To try and achieve some consistency, ages of all lower jaws collected each year were reviewed after the 1991 seasons using the same base criteria for year classes. Whenever age estimates for jaws and from counts of dental cementum annuli disagreed, a "best guess" estimate was made with preference given to jaw ages. The scarcity of known-age jaws limited the confidence placed on age estimates and errors were presumed to increase with increasing age. The significant difference in doe age structure between the regular and late seasons in 1989 (l - 0.025) did not prevail the following 2 years (l ~ 0.418) (Fig. 2). This was primarily due to smaller differences in the percentages of deer >7-years old the last 2 years as well as a smaller difference in the percentages of yearlings in 1991. There was a significant difference (f < 0.001) across years in age structures during the late season due, in part, to the relatively high survival of fawns in 1989-90 that caused increases in the percentage of yearlings in 1990 and of 2-year-old deer in 1991. Neither weight, total body length, left hind foot length, nor weight/length ratio differed between the treatment and control units for yearling or adult females (f ~ 0.178) (Table 6). �34 ••••• REG. SEASONS LATE SEASON 50 1989 40 ••••• n = 33 n = 223 30 20 ..1.•.•.•.•.•..•..•.................................................................... _.............................................................................. . . ~ 111111 10 ~ 0 50 z ~ U 0::: ~ ~ 30 .I.~_~~~.~.~~~ •., _...... •••••• .. " 1·990 40 ~ . ••••• n = 53 n = 201 ~ .....••.~...••••.......................................................•...........•...............................•.•.................................................................•.....•.... III~ 20 10 0 50 1991 40 ••••• n = 22 n = 164 30 20 10 0 1 2 3 - 4 5 6 7 >7 ESTltlATED AGE (YEARS) Fig. 2. Estimated age structures of does harvested seasons on the Ridge study area, 1989-91. during regular and late �35 Table 5. Weights (kg) and body measurements (cm) of mule deer fawns trapped on control and treatment units of the Ridge study area in Piceance Basin, Colorado, 1982-91. Weight Total bodI length SD Left hind foot length SD Unit Year n X SD Control 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 28 28 34 60 58 33 34 40 35 28 34.6 31.7 32.3 32.6 31.9 29.9 29.5 32.7 30.8 30:7 3.10 4.40 4.65 4.02 3.89 3.60 3.10 3.31 4.29 3.58 124.0 124.2 123.9 124.4 128.1· 127.3 123.8 131.0 126.5 128.2 4.64 5.65 7.25 6.26 6.53 6.12 7.83 5.81 9.02 5.47 41.1 40.6 40.8 41.1 41.0· 40.8 41.0 41.8 40.8 40.6b 1.08 1.73 1.53 1.48 1.95 1.72 1.37 2.16 1.75 1.32 Treatment 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 30 32 26 60 61 28 30 47 36 43 32.8 32.3 32.3 32.3 31.7 30.2 28.8 30.6 30.7 32.6 4.18 3.12 5.07 4.62 4.13 5.34 4.13 3.33 4.41 3.54 121.7c 123.6 124.7 124.4 126.0· 127.5 124.9 126.6 127.6d 129.8 5.45 5.53 7.25 6.28 6.62 8.86 7.07 5.33 6.57 6.11 41.1c 40.6 40.8 40.8 41.0· 41.2 40.6 40.7 41.1d 40.8 1.65 1.34 1.89 1.77 2.11 1.66 1.88 1.41 1.69 1.29 • nnc nd n- b Condition X X 60 27 31 32 of Yearling Males No yearling males were trapped on the control unit in 1991, so comparisons of body measurements between units cannot be .made (Table 7). However, comparisons of antler measurements are possible because they include data from deer harvested during the hunting seasons. However, no significant differences were detected between units for total points, left and right main beam lengths, and left and right burr circumferences (£ ~ 0.342) (Table 8). �36 Table 6. Weights (kg) and body measurements (cm) of yearling and adult female mule deer trapped on control and treatment units of the Ridge study area in Piceance Basin, Colorado, 1988-9l. Weight SD Total bod:;llength SD X Left hind foot length SD X Unit Year n Control 1988 1989 1990 1991 1988 1989 1990 1991 2 2 10 8 2 7 8 10 46.2 46.0 49.48 50.9 50.2 53.3 50.8 49.4 Yearlings l.91 6.43 2.38 2.50 0.35 9.61 3.38 3.45 146.1 148.4 147.9 149.2 15l.2 158.7 150.1 149.8 10.75 2.26 2.49 3.38 3.18 1l.03 4.94 5.31 45.8 47.8 45.8 45.7 46.0 46.7 45.6 45.3 0.99 3.04 l.08 l.02 0.00 0.78 l.26 l.51 1989 1990 1991 1989 1990 1991 19 21 30 39 25 32 67.7 66.9 62.9 65.0 67.6 62.8 Adults 5.22 5.51 5.18 5.78 5.14 6.24 168.6 166.7 162.3 165.9 166.8 162.2 5.20 6.98 6.40 6.76 5.81 7.07 48.1 47.3 47.3 47.8 48.1 47.0 l.10 l.09 2.19 l.32 l.44 l.06 Treatment Control Treatment a X n - 9. Table 7. Weights (kg) and body measurements (cm) of yearling male mule deer trapped on control and treatment units of the Ridge study area in Piceance Basin, Colorado, 1988-91. Weight SD Total bod:;llength SD X Left hind foot length SD X Unit Year n Control 1988 1989 1990 1991 6 2 2 0 54.0 48.5 52.2 4.91 5.30 3.99 145.1 151.8 154.8 11.53 8.13 6.67 47.4 46.7 48.0 l.39 0.71 3.52 Treatment 1988 1989 1990 1991 6 1 9 7 5l.8 60.8 60.6 54.9 4.01 151.9 168.5 157.6 154.7 6.07 48.5 50.0 48.7 47.6 l.19 X 3.20 5.16 5.75 7.67 l.08 0.69 �37 Table 8. Antler measurements (cm) of yearling male mule deer harvested and trapped on control and treatment units of the Ridge study area in Piceance Basin, Colorado, 1988-91. Hain antler beam len&th Right Left SD SD X X na ~UII ciIcYmfelence Left Right SD SD X X Unit Year Control 1988 1989 1990 1991 6 8 19 5 16.9 20.2 19.2 22.2 3.49 5.60 5.29 1.42 16.4 21.0 18.7 22.2 2.30 5.79 5.58 1.81 52.7 51.0 57.5 56.2 10.05 20.02 8.86 6.57 54.0 49.0 56.6 57.6 6.93 18.73 10.05 6.84 Treatment 1988 1989 1990 1991 6 4 24 20 17.3 17 .5 21.1 21.1 3.47 6.30 5.61 5.34 16.9 17.0 19.5 20.1 4.11 6.84 5.04 6.04 53.2 40.0 56.5 54.6 7.60 24.38 9.80 7.90 54.8 38.0 54.9 53.8 6.91 23.19 11.43 8.15 LITERATURE CITED Bartmann, R. M. 1990. Compensatory effects of harvest in a mule deer popUlation. Colo. Div. Wildl., Wildl. Res. Rep. July:187-196. ______ , and G. C. White. 1991. Compensatory effects of harvest in a mule deer population. Colo. Div. Wild1., Wildl. Res. Rep. Ju1y:27-40. Prepared by !UL~ Richard M. Bartmann Wildlife Researcher \):j{U)j./ U /j c. .? LC{,+<-tu Dr. clary C. White Professor ��39 Colorado Division of Wildlife Wildlife Research Report July 1992 JOB PROGRESS REPORT State of ~C~o~l~o~r~a~d~o~ _ Project No. W-lS3-R-4 Mammals Research Work Plan No. 2 Deer Investigations Job No. 16 Period Covered: Authors: Personnel: Regulation of Mule peer Population Growth by Fertility Control: Laboratory. Field. and Simulation Experiments July 1, 1991 - June 30, 1992 Dan L. Baker, N. T. Hobbs, T. M. Nett, M. W. Miller A. C. Case, J. R. Ritchie, M. A. Wild Abstract We propose research to develop a practical and economical method of fertility control in mamma1ianwi1dlife that overcomes many of the shortcomings of current technology, particularly problems of treatment duration and environmental safety. We propose to use conjugates of gonadotropin-releasing hormone (GnRH) and cellular toxins to selectively destroy gonadotropin producing cells Ln the anterior pituitary gland, thereby preventing gamete production by the ovaries and testes. This research will consist of laboratory, field, and modeling phases, each of which is designed to address questions that must be answered before widespread application of hormonal-toxin conjugates is possible. In order to provide experimental animals for these studies, we successfully bo trt Ie-Yafsed and trained 16 mule deer fawns from June 12 to June 30 1992. Fawns were acquired from two sources; seven from captive wild does and nine received as orphans. Hand-reared neonates born to wild does in captivity experienced a high incidence of diarrhea compared to orphans. The relatively high prevalence of infectious diseases suggest stress induced immunosuppression in these neonates. ��41 REGULATION OF MULE DEER POPULATION GROWTH BY FERTILITY LABORATORY, FIELD, AND SIMULATION EXPERIMENTS CONTROL: Dan L. Baker P. N. OBJECTIVES 1. To develop a practical and acceptable method deer populations using GnRH conjugates. for controlling mule 2. To demonstrate the feasibility of such control in a field application at the Rocky Mountain Arsenal. 3. To predict population impacts of alternative using simulation modeling. contraceptive regimes . SEGMENT OBJECTIVES 1. Prepare an approved Program Narrative to study the regulation deer population growth by fertility control. 2. Bottle raise and train approximately in fertility control experiments. of mule 20 mule deer fawns to be used METHODS AND MATERIALS Mule Deer Fertility Control Study We prepared a detailed study plan describing research to develop an efficacious method of fertility control in mule deer (Appendix A). This research will be conducted in three phases. The first phase will consist of laboratory experiments. We will conduct controlled experiments with tame, penned mule deer to determine the most effective dose of GnRH-toxin conjugates. We will also determine the stage of the reproductive cycle when fertility control is most effective and evaluate the duration of effectiveness. Finally, we will use these studies to evaluate the safety and side effects (if any) of treatments. All studies in the laboratory phase will be conducted at the Colorado Division of Wildlife's Foothills Wildlife Research Facility, and the Department of Physiology, Colorado State University, Ft. Collins, Colorado. The second phase of research will be a field test of hormonal-toxin conjugates. Our laboratory experiments will provide strong inferences on the potential utility of the GnRH-toxin conjugates as a method of fertility control and will offer background information needed to apply the technique in the field. A test of its success under field conditions will evaluate whether the technique is truly feasible and practical. In phase three, we will develop a simulation model of the population dynamics of the Rocky Mountain Arsenal mule deer herd. Applying GnRH-toxin conjugates to control growth of mule deer populations will require that wildlife managers choose specific tactics for treating animals. Choices must be made on the number �42 and age to treat, the frequency of treatment, and so on. Decisions on the best tactics will depend on comparing the effects of alternative action on population behavior. We will provide support for these decisions by developing an interactive model of deer population dynamics. This model will combine knowledge of mule deer biology with and understanding of the constraints intrinsic in the GnRH-toxin conjugate technique. Using this model, managers will be able to evaluate the probable consequences of their decisions on implementing fertility control regimes. Mule Deer Fawns We successfully hand-raised 16 female mule deer fawns for use in deer fertility experiments. Fawns were acquired from two sources; seven from wild trapped does maintained at the Colorado Division of Wildlife's Little Hills Experiment Station near Meeker, Colorado and nine received at the Foothills Wildlife Research Facility in Ft. Collins as "orphans". All neonates were bottle-raised following the methods described by Wild and Miller 1991. Seventeen wild mule deer does'were captured and transported to a 0.20 ha holding pen at Little Hills Experiment Station approximately 4 months prepartum. Deer were provided ad libitum quantities of third-cutting alfalfa hay, water, and supplement (Baker and Hobbs 1985). Three does died prior to parturition from injuries incurred while in the enclosure. Remaining females gave birth to twenty seven fawns (15 males; 12 females) during the period June 11 to June 28. Average parturition date was June 19. Newborn female fawns were removed from the dam 2448 hours after identification; male fawns were left with the dam. Body weight of female fawns at 24 hrs of age was 3.16 (SE - 0.14)kg. Two weeks following the birth of the last neonate, all does and male fawns were released without incident. Hand-reared neonates born to wild does in captivity experienced a high incidence of diarrhea. A variety of pathogenic organisms including salmonella, coronavirus, rotavirus and E.coli were isolated from affected fawns. The high prevalence of infectious diseases suggest a decreased immunity in these fawns. Previous studies of immunologic response of fawns born to captive wild and tame does support this hypothesis (Parkinson et a1. 1982, Trindle et a1. 1978). These studies indicate that stress induced immunosuppression of neonatal fawns born to captive wild does is largely due to the failure of fawns to nurse sufficient quantities of colostrum because of poor maternal care or because of prolonged separation of fawn from the mother. Failure to absorb antibodies from ingested colostrum occurs most commonly beCause of delayed nursing (Tizard 1977). The ability to absorb colostral antibodies begins to decrease immediately following birth (Stott et al. 1979). By 10 to 14 h of age, the ruminant's ability to absorb antibodies is decreased by 50 percent and by 24 h of age significant levels of antibodies can no longer be absorbed. Stress of confinement of captive wild does held in a relatively small enclosure, lack of concealment barriers from human activities and the daily disturbance of intensive searching for neonates may have predisposed fawns to infectious diseases. �43 LITERATURE CITED Baker, D. L., and N. T. Hobbs. 1985. Emergency feeding of mule deer during winter: tests of a supplemental ration. J. Wild. Manage. 49:934942. Parkinson, D. E., R. P. Phillips, and L. D. Lewis. 1982. Colostrum deficiency in mule deer fawns: identification, treatment and influence on 'neonatal mortality. J. Wi1d1. Dis. 18:17-28. Stott, G. H., D. B. Marx, B. E. Menefee, and G. T. Nightengale. 1979. immunoglobulin transfer in calves. II. The rate of absorption. Dairy Sci 62:1766-1773. Tizard, I. R. 1977. An introduction to veterinary Co., Philadelphia, Pa. 367 pp. immunology. Colostral J. W. B. Saunders Trindle, B. D., L. D. Lewis, and L. H. Lauerman. 1978. Evaluation of stress and its effect on the immune system of hand-reared mule deer fawns. J. Wi1dl. Dis. 14:523-537. Wild, M. A., and M. W. Miller. 1991. Bottle-raising wild captivity. Outdoor Facts #114, Colo. Div. Wildl., Denver. Prepared by ~/~ Dan L. Baker Wildlife Researcher C ruminants 6pp. in ��45 Colorado Division of Wildlife Wildlife Research Report July 1992 JOB PROGRESS REPORT State of ~C~o~l~o~r~a~d~o~_ Project No. W-153-R-5 Mammals Research Work Plan No. 3 Elk Investigations Job No. 6 Effect of Elk Harvest Systems on Elk Breeding Biology Period Covered: Author: July I, 1991 - June 30, 1992 D. J. Freddy Personnel: M. Miller, CDOY, D. Bowden and G. White, CSU, and E. Ryland of Forbes Trinchera Ranch. Abstract Harvests of deer and elk on the Forbes Trinchera Ranch increased from 1986 to 1991 by 40% for buck deer, 72% for bull elk, 170% for female deer and 230% for female elk. Ages of bucks harvested declined across years (P -0.002) while changes in antler scores were inconclusive but possibly declining. Ages of bulls harvested increased across years (P -0.005) while antler scores remained stable or increased (P ~0.2l2). Ages of female deer harvested increased across years (P -0.000) while ages of female elk harvested remained stable (P -0.276). Current harvests of bulls and female deer and elk may be sustainable but harvest of bucks may need to be reduced. Pregnancy rates for elk averaged 80% and ranged from 62-92%. Pregnancy was not detected in calves, rates for yearling~ averaged 17%, and adult ,rates were lowest (78%) for cows ~ II-years old. Fe ca], sex ratios, deviated from 50 M: 50 F,in 4 of 6 years and Lncons Lscent.Ly favored males or females (P 5;();05)but the sex ratio of 55 M:45 F pooled across years was not; different from so M:'50 F (P >0.05). Mean and.' median conception dates ranged from 23 September to 4 October. Earlier conception dates were associated with increasing harvests of male and female elk. All serum samples collected from elk during December 1991 tested negative for brucellosis. ��47 JOB PROGRESS REPORT EFFECT OF ELK HARVEST SYSTEMS ON ELK BREEDING BIOLOGY David J. Freddy P. N. OBJECTIVE Evaluate effects of harvest systems on breeding biology of elk. SEGMENT OBJECTIVES 1. Determine reproductive status of elk on the Forbes Trinchera Ranch using fetal collections from hunter-killed elk. 2. Determine physical condition of elk and deer on the Forbes Trinchera Ranch by collecting body weights, antler weights and measurements, and ages of animals harvested. 3. Submit for publication a manuscript entitled "Pregnancy eeseing,elk using progeseerone assays". INTRODUCTION During the 1980's there was concern throughout Colorado that low numbers of bull elk (Cervus elaphus nelsoni) during the breeding season were reducing conception rates in adult cows and survival of calves. This concern was based primarily on correlations between declining postseason bull:cow ratios and calf:cow ratios in some herds in Colorado and other states. An auxiliary concern was that rising hunting pressure on bull elk during the rut in September could interfere with breeding and subsequently reduce conception rates or delay breeding of cows until their second or third estrus cycle later in the Fall. Part of this concern stemmed from insufficient knowledge about changes over time, if any, in the in vivo reproductive status of elk. Our knowledge of pregnancy rates and conceptions dates for elk in Colorado is limited primarily to one survey conducted in 1965-1966 in southwestern Colorado (Boyd and Ryland 1971) when total hunting pressure and hunting of elk during the rut were comparatively low. As a participant in Colorado's Wildlife Ranching program, the Forbes Trinchera Ranch in south-central Colorado instituted controlled fee-hunting for bull elk and buck deer (Odocoileus hemionus) from September through early December and initiated controlled public hunting for ant1er1ess elk and deer in December beginning in 1986 (Freddy et al. 1991). This hunting system afforded the opportunity to: 1) intensely monitor the harvesting of male and female elk and deer in a "trophy" hunting system, and 2) monitor reproduction in an elk population subjected to chronic hunting of mature bulls during the rut in September by obtaining reproductive organs from female elk harvested by hunters in December when in vivo pregnancy status of elk could be easily determined. �48 Our objectives were to monitor the number and age composition of male and female harvests of deer and elk and measure in vivo pregnancy rates, fetal rates, and fetal sex and weights of elk and to seek correlative relationships between these reproductive variables and numbers and ages of bulls harvested, age and body condition of cows harvested, and trends in precipitation by month and year. Ancillary objectives were to use blood samples obtained by hunters from harvested female elk to monitor disease status of elk and to evaluate accuracy of serum progesterone hormonal assays to predict pregnancy in elk. METHODS Aies of Animals Harvested We estimated age of deer and elk harvested from 1986 to 1991 using replacement and wear (RW) (Robinette et a1. 1957, Quimby and Gaab 1957) and dental cementum (DC) (Stevens 1987, Keiss 1969). We used both techniques because each estimates true age with an unknown degree of bias. Two observers (E. Ryland, primary; D. Freddy, secondary) independently aged lower jaws removed from harvested buck deer and bull elk; heads and antlers of these animals were not present when jaws were examined and observers examined jaws at different times. These same observers examined intact jaws from antler1ess deer and elk harvested and pooled their estimates of age. A median incisor was removed and age determined via dental cementum by the same technician each year at the CDOW research laboratory in Ft. Collins. For most analyses, ages were pooled into classes often representing young, mature, and old individuals. Antler Measurements We measured antlers from harvested bucks and bulls according to Boone and Crockett criteria (Nesbitt and Reneau 1986). Three individuals measured antlers from 1987-1991 with one person common to all years. Antler weights, including the frontal bone, were measured to the nearest 0.1 kg. Body Measurements and Fetal Collections Eviscerated body weights (nearest 0.5 kg) were obtained on deer and elk harvested from 1987 to 1991 using scales located at mandatory check stations. Total body length and hind foot length were also obtained for ant1er1ess animals. Reproductive tracts from female elk, including calves, harvested primarily in December were collected by hunters who were provided step-by-step illustrated instructions prior to their hunt. Tracts were deposited at check stations and kept cool until processing. Pregnancy status was determined from the presence of fetuses, embryos, and developed uterine tissue. Questionable uteri were preserved for later examination. Fetal measurements were made on fresh specimens (subsequently preserved) and followed definitions of Armstrong (1950). Fetal age was estimated from growth curves of Morrison et al. (1959). �49 Blood Assays We monitored progesterone levels of antlerless elk harvested by hunters in December 1987-1991. Hunters were instructed to obtain blood from the thoracic cavity of elk immediately after harvesting their elk and keep the nonheparinized vials of blood cool until depositing vials at check stations. Blood was received from hunters usually within a few hours of when animals were harvested and then refrigerated until centrifuging (usually within 12 hrs) at which time serum was frozen and stored at -18 C until processing. Because hunters also collected the reproductive organs from their elk we could associate progesterone levels with known pregnancy status. Progesterone levels were determined by radioimmunoassay (RIA) at the Physiology Laboratory, Colorado State University, Ft. Collins. Serum from elk harvested by hunters was tested for the presence of brucellosis by the USDA Laboratory, Denver, CO. Serum was stored at -18C from centrifuging until tested for brucellosis. Hunting Seasons Separate hunting seasons were established yearly for private fee-only hunters and public hunters. Private seasons for bucks and bulls were: 13 Sep-9 Nov 1986, 5 Sep-9 Oct and 14 Nov-II Dec 1987, 10 Sep-9 Dec 1988, 9 Sep-8 Dec 1989, 9 Sep-5 Dec 1990, and 7 Sep-6 Dec 1991. Public seasons primarily for female deer and elk were: 29 Nov-5 Dec 1986, 12-14 and 19-21 Dec 1987, 10-12 and 1719 Dec 1988, 9-11 and 16-18 Dec 1989, 8-17 Dec 1990, and 7-15 Dec 1991. Statistical Analysis We used SAS (SAS Institute, Inc. 1988) to evaluate trends in age composition of harvests (PROC FREQ - chi-square, PROC GLM - anova and linear regression), antler measurements (PROC GLM), body measurements (PROC GLM), elk conception dates (PROC UNIVARIATE), elk fetal sex ratios (PROC FREQ), and relationships between fetal sex ratios and ages and body sizes of pregnant cows (PROC LOGISTIC). We generally used P ~0.05 as the critical level for statistical significance. RESULTS AND DISCUSSION Harvests Harvests of both male and female deer and elk increased from 1986 through 1991 (Table 1). Compared to 1986 harvests, maximum increases were: 40% for buck deer, 72% for bull elk, 170% for female deer and 230% for female elk. About 55% of the bucks were harvested in November-December and 85% of the bulls were harvested in September-October. Proportions of the buck, bull, and female deer and elk harvests occurring within the "Blanca" portion of the Ranch were: 70%, 57%, and 80-85%, respectively. Bucks-- Age (RW) frequencies of bucks differed among years due to a declining proportion ~ 7 years old (P -0.001). Mean age declined across years (ANOVA P -0.0001; R2 -0.014 P -0.002; Table 2). This decline in yearly RW age was �50 Table 1. Harvests of male and female deer and elk on Forbes Trinchera Ranch durin~ private fee-only and public hunting seasons. 1986-1991. 1986 Seas./Spec. Private/Deer Private/Elk Pub1ic/Deer Public/Elk M 1987 F M 1988 F M 1989 F M 1990 F M 1991 F 84 68 0 0 107 93 0 0 109 87 0 0 118 103 0 0 108 117 0 0 116 104 0 0 5 3 69 33 6 5 137 63 5 3 133 79 5 3 186 61 17 9 160 56 16 9 168 109 -Includes only males ~ 1 year old. bIncludes adult females and male and female fawns or calves. Table 2. Ages of buck deer harvested on Forbes Trinchera Ranch during private fee-only hunting seasons, 1986-1991. Age based on replacement and wear as estimated by E. Ryland. Age Category 1986 1987 Year !Yearly Percent) 1988 1989 1990 1991 1-3 11 (13.3) 5 ( 4.8) 3 ( 2.8) 16 (13.6) 17 (15.7) 18 70 (15.5) 4-6 40 (48.2) 42 (40.0) 49 (45.4) 50 (42.4) 43 (39.8) 288 64 (55.2) ~ 7 32 (38.6) 58 (55.2) 56 (51.8) 52 (44.1) 48 (44.4) 34 280 (29.3) Avg. Age (SE) 5.84 (0.19) 83 6.62 (0.18) 105 6.56 (0.17) 108 6.02 (0.18) 118 5.98 (0.21) 108 5.52 (0.18) 116 638 n Total corroborated by the second field observer (~ -0.054, P -0.0001). Changes in DC ages among years were less definitive. Age frequencies (DC) of bucks differed among years primarily due to erratic proportions of 1-3 year old deer that were harvested (P -0.017). Yearly mean DC ages were always younger than RW means, differed without pattern among years (ANOVA P - 0.003), and did not exhibit a consistent decline across years (R2 -0.006 P -0.07, Table 3). Caution would argue for accepting a decline in ages of bucks harvested from 1986 to 1991. Age frequencies of bucks differed between RW and DC ages each year with DC indicating fewer bucks ~ 7 years old were harvested (P <0.04, Table 4). �51 Table 3. Ages of buck deer harvested on Forbes Trinchera Ranch during private fee-only hunting seasons. 1986-1991. Age based on dental cementum. 1986 1987 Year (Yearly Percent} 1990 1988 1989 1-3 17 (23.6) 19 (18.1) 19 (18.6) 34 (30.4) 10 ( 9.8) 118 19 (17.9) 4-6 51 (70.8) 69 (65.7) 66 (64.7) 62 (55.4) 71 (69.6) 391 72 (67.9) ~ 7 4 ( 5.6) 17 (16.2) 17 (16.7) 16 (14.3) 21 (20.6) 90 15 (14.2) Avg. Age 4.49 (0.17) 72 5.03 (0.16) 105· 5.05 (0.16) 102 4.63 (0.17) 112 5.40 (0.17) 102 4.97 (0.17) 599 106 Age Category (SE) n 1991 Total Table 4. Yearly comparisons between replacement and wear (RW) and dental cementum (DC) ages of buck deer harvested on Forbes Trinchera Ranch during private fee-only hunting seasons, 1986-1991. Data represent only those deer aged by both techniques. RW age by E. Ryland. Year Aging Technique 1-3 Age Categoo 4-6 ~ 7 n Chi-square Yearly (P) 1986 RW DC 11 17 33 50 27 4 71 71 0.000 1987 RW DC 5 19 42 66 54 16 101 101 0.000 1988 RW DC 3 19 48 66 51 17 102 102 0.000 1989 RW DC 16 33 44 60 48 15 108 108 0.000 1990 RW DC 15 10 41 71 46 21 102 102 0.000 1991 RW DC 16 17 57 71 30 15 103 103 0.038 �52 Differences in techniques cannot be solely attributed to the field observer. The 2 independent estimates of RW age frequencies differed only in 1986 (Table 6). However, a paired t-test indicated that observers differed each year in RWages except 1986 and 1988 (P <0.007, Table 6) but such differences (+0.34 -0.41 years) were smaller than differences in average RW and DC ages (0.55 1.59 years, Tables 2, 3). Bulls-- Age (RW) frequencies of bulls differed among years due to an increasing proportion ~ 7 years old (P -0.000). Mean age increased across years (ANOVA P -0.002; ~ -0.014 P -0.005; Table 8). This increase in RW age across years was not corroborated by the second field observer (R2 -0.0005 P -0.584). Changes in DC ages among years also indicated an increasingly older aged harvest. Age frequencies (DC) of bulls differed among years due to increasing proportion ~ 7 years old (P -0.000, Table 9). Yearly mean DC ages were always younger than RW means but exhibited an increase across years (ANOVA P - 0.000; R2 -0.080 P -0.000). We suggest that ages of bulls harvested increased, or at least remained stable, from 1986 to 1991. Age frequencies of bulls differed between RW and DC ages each year except 1991 with DC indicating fewer bulls ~ 7 years old were harvested (P <0.056, Table 5). Again, differences in techniques cannot be solely attributed to the field observer. The 2 independent estimates of RW age frequencies differed only in 1990 (Table 7). However, a paired t-test again indicated that observers differed each year in RW ages except 1986 and 1989 (P <0.010, Table 7) but such differences (+0.33 - -0.72 years) were generally smaller than differences in average RW and DC ages (0.29 - 1.21 years, Tables 8, 9). Females-Deer-- Age frequencies of female deer differed among years for all age classes including fawns (P -0.000) and for adults only (P -0.000) due to an increasing proportion ~ 7 years old and a declining proportion 2-3 years old. Mean age of adults increased across years (ANOVA P -0.000; R2 -0.031 P -0.000; Table 10). We conclude that ages of female deer harvested increased from 1986 to 1991 based on changes in ages of adult animals. We believe basing our conclusions on changes in adult ages is more appropriate than analyses including fawns because hunters likely select against fawns. Age frequencies of female deer were similar between RW and DC ages in most years (P >0.184) but differed when all years were pooled (P -0.017, Table 11). This difference was primarily caused by a lower proportion of yearlings in the DC sample. We assumed that our RW criteria for yearlings were correct and that yearlings could be considered "known" age. The "error" with DC involving yearlings was caused by the presence of a "cementum line" resulting in some yearlings being classed as 2-years old. We therefore recommend that female deer be aged by RW for fawn and yearling classes and by DC or RW for older ages. Females-Elk-- Age frequencies of female elk were stable among years for all age classes when including calves (P -0.261) and for adults only (P -0.276). Mean age of adults were stable across years (ANOVA P -0.801; R2 -0.004 P -0.232; Table 12). We conclude that ages of female elk harvested remained stable from 1986 to 1991 based on ages adult animals; again, because hunters likely select against calves. �53 Table 5. Yearly comparisons between replacement and wear (RW) and dental cementum (DC) ages of bull elk harvested on Forbes Trinchera Ranch during private fee-only hunting seasons, 1986-1991. Data represent only those elk aged by both techniques. RW age by E. Ryland. Year Aging Technique 1-3 Age Catego;o: 4-6 ~ 7 n Chi-square Yearly (P) 1986 RW DC 10 25 32 26 10 1 52 52 0.001 1987 RW DC 16 46 59 38 15 6 90 90 0.000 1988 RW DC 15 27 63 56 9 4 87 87 0.056 1989 RW DC 15 42 61 53 24 5 100 100 0.000 1990 RW DC 19 21 48 64 45 27 112 112 0.032 1991 RW DC 18 18 52 60 22 14 92 92 0.309 Table 6. observers Trinchera represent Year Ages of deer based on replacement and wear (RW) compared between E. Ryland (ER) and D. Freddy (DF) for bucks harvested on Forbes Ranch during private fee-only hunting seasons, 1986-1991. Data only those deer aged by both observers. Aging Technique 1-3 Age CategoD: 4-6 ~ 7 n Chi-square Yearly (P) Age (Years) Avg. (SE) 1986 RW-ER RW-DF 11 4 38 53 29 21 78 78 0.030 5.78(0.20) 5.88(0.21) 1987 RW-ER RW-DF 5 4 41 32 58 68 104 104 0.365 6.63(0.18) 6.96(0.18) 1988 RW-ER RW-DF 3 4 49 42 56 62 108 108 0.611 6.56(0.17) 6.64(0.17) 1989 RW-ER RW-DF 16 14 49 61 52 42 117 117 0.286 6.02(0.18) 5.80(0.18) 1990 RW-ER RW-DF 17 17 43 52 48 39 108 108 0.410 5.98(0.21) 5.71(0.19) 1991 RW-ER RW-DF 18 23 64 65 34 28 116 116 0.549 5.52(0.18) 5.11(0.17) �54 Table 7. Ages of elk based on replacement .and wear (RY) compared between observers E. Ryland (ER) and D. Freddy (DF) for bulls harvested on Forbes Trinchera Ranch during private fee-only hunting seasons, 1986-1991. Data represent only those elk aged by both observers. Year Aging Technique 1-3 Age Categoo 4-6 ~ 7 n Chi-square Yearly (P) Age (Years) Avg. (SE) 1986 RY-ER RY-DF 12 12 42 44 11 91 65 ··65 0.884 4.88(0.21) 4.89(0.23) 1987 RY-ER RY-DF 13 10 54 52 14 19 81 81 0.552 4.99(0.18) 5.27(0.20) 1988 RY-ER RY-DF 15 9 63 63 96 15 87 87 0.223 4.80(0.16) 5.14(0.17) 1989 RY-ER RY-DF 16 16 61 64 25 22 102 102 0.877 5.18(0.18) 5.24(0.18) 1990 RY-ER RW-DF 19 24 49 66 48 26 116 116 0.008 5.75(0.19) 5.03(0.17) 1991 RY-ER RW-DF 22 25 54 57 27 21 103 103 0.600 5.16(0.20) 4.92(0.18) Table 8. Ages of bull elk harvested on Forbes Trinchera Ranch during private fee-only hunting seasons, 1986-1991. Age based on replacement and wear as estimated by E. Ryland. 1986 1987 Year ,Yearly Percent) 1988 1989 1990 1-3 13 (19.4) 16 (17.4) 15 (17.2) 16 (15.5) 19 (16.4) ·22 101 (21.2) 4-6 43 (64.2) 61 (66.3) 63 (72.4) 61 (59.2) 49 (42.2) 5S 332 (52.9) z 11 (16.4) 15 (16.3) 9 (10.35 26 (25.2) 48 (41.4) 27 136 (26.0) 4.85 (0.20) 67 4.92 (0.17) 92 4.80 (0.16) 87 5.19 (0.18) 103 5.75 (0.19) 116 5.15 (0.20) 104 569 Age Category 7 Avg. Age (SE) n 1991 Total �55 Table 9. Ages of bull elk harvested on Forbes Trinchera Ranch during private fee-only hunting seasona, 1986-1991. Age based on dental cementum. Age Category 1986 1987 Year (Yearly Percent) 1988 1989 1990 1991 1-3 28 (48.3) 46 (50.6) 27 (31.0) 42 (42.0) 21 (18.6) 183 19 (19.4) 4-6 29 (50.0) 39 (42.9) 56 (64.4) 53 (53.0) 65 (57.5) 64 306 (65.3) z 1 ( 1.7) 6 ( 6.6) 4 ( 4.6) 5 ( 5.0) 27 (23.9) 15 58 (15.3) 3.64 (0.17) 58 3.79 (0.15) 91 4.17 (0.15) 87 4.07 (0.15) 100 5.05 (0.18) 113 4.86 (0.18) 547 98 7 Avg. Age (SE) n Total Table 10. Ages of antlerless deer harvested on Forbes Trinchera Ranch during public hunting seasona, 1986-1991. Age based on replacement and wear for tawns An~ yeA~lings and dental cementum for ages > ~ years. Age Category Year (Yearly Percent) 1988 1989 1990 1986 1987 Fawna 13 (20.6) 22 (16.5) 14 (11.8) 32 (17.9) 22 (14.3) 32 135 (20.1) Yearling 4 ( 6.4) 19 (14.3) 10 ( 8.4) 24 (13.3) 16 (10.4) 15 88 ( 9.4) 2-3 23 (36.5) 49 (36.8) 58 (48.7) 67 (37.2) 29 (18.8) 42 268 (26.4) 4-6 18 (28.6) 34 (25.6) 28 (23.5) 40 (22.2) 53 (34.4) 42 215 (26.4) z 5 (7.9) 9 ( 6.8) 9 ( 7.6) 17 ( 9.4) 34 (22.1) 28 102 (17.6) n 3.80 (0.31) 50 3.33 (0.19) 111 3.35 (0.19) 105 3.59 (0.20) 148 4.82 (0•. 26) 132 4.42 (0.23) 673 127 Total Aged 63 133 119 180 154 159 7 Avg. Ageb (SE) aIncludes male and female fawns. bAverage age excluding fawns. 1991 Total 808 �56 Table 11. Yearly comparisons between replacement and wear (RW) and dental cementum (DC) ages of female deer harvested on Forbes Trinchera Ranch during public hunting 8eaaons, 1987-1991. Data represent only those deer aged by both techniques. Aging Technique Fawn· 1987 RW DC 1988 Chi-square Yearly (P) Age Categor~ 2-3 4-6 1 ~ 7 12 12 18 9 31 38 19 24 10 7 90 90 0.306 RW DC 10 10 !J 97 57 35 26 16 7 105 105 0.051 5 1989 RW DC 9 9 19 16 49 64 37 36 28 17 142 142 0.292 1990 RW DC 9 9 15 5 36 34 41 51 29 31 130 130 0.184 1991 RW DC 20 20 13 12 43 41 33 40 32 28 141 141 0.906 ALL RW DC 60 60 196 234 165 177 115 90 608 608 0.017 Year 72 47 n ·Includes male and female fawns. Table 12. Ages of antlerless elk harvested on Forbes Trinchera Ranch during public hunting 8easona, 1986-1991. Age based on replacement and wear for calves. ~earlinqs. and 2-~r olds and dental cementum for ages> 3 ~ears. Age Year (Yearl~ Percent) Category 1986 1987 1988 1989 1990 1991 Total Yearling 2 3-4 5-7 8-10 ~ 11 Avg. Age" (SE) n Total Aged 2 ( 5.7) 14 (22.2) 14 (18.9) 9 (15.3) 10 (18.2) 22 71 (21.0) 6 7 7 6 9 (17.1) (11.1) ( 9.5) (10.2) (16.4) 7 42 ( 6.7) 7 (20.0) 10 (15.9) 8 4 ( 6.8) 7 (10.8) 6 8 (12.7) 18 (24.3) 12 (20.3) 8 (17.1) 5 15 (23.8) 16 (21.6) 18 (30.5) 8 (14.3) (14.6) 14 76 (13.3) 2 ( 5.7) 4 ( 6.4) 5 ( 6.8) 4 ( 6.9) 8 (14.6) 14 37 (13.3) 7 5 6 6 5 (20.0) ( 7.9) ( 8.1) (10.2) ( 9.1) 15 44 (14.3) 5.39 (0.79) 33 35 5.14 (0.59) 49 63 5.13 (0.49) 60 74 5.60 (0.57) 50 59 5.36 (0.65) 45 ·Includes male and female calves. ~verage age excluding calves. (12.7) (14.6) 55 11 47 (10.5) 22 74 (21.0) 6.03 (0.47) 83 320 105 391 �57 Age frequencies of female elk were similar between RW and DC ages each year (P >0.091) and in all years pooled (P <0.069, Table 13). When all years were pooled only marginal differences occurred for age classes ~ 8 years (Table 13). We conclude that DC or RW provide acceptable estimates of age for female elk of all ages. We therefore found that the relative performance of RW and DC aging techniques was dependent on sex of animal. For adult bucks and bulls, DC consistently provided younger ages than RW but the techniques provided similar ages for adult female deer and elk. We believe our subjective criteria for wear on teeth was the same for both sexes. If so, younger ages provided by DC would suggest that wear patterns on male teeth were accelerated compared to females. Another explanation is that male teeth often had crowded cementum lines which were harder to read than the more orderly cementum lines found in females thus raising the possibility of missing lines in teeth of males. Antler Measurements Deer-- Antler scores and weights increased until plateauing at age 6 (Fig. I). Trends were similar based on either RW or DC ages except that scores and weights were slightly higher at younger ages for DC. Gross antler scores near 200 occurred primarily at ag~s ~ 5 but net scores near 200 occurred at age 6. Trends in scores and weights suggested ages 5, 6, and 7 were the most efficient for producing high scoring antlers. Yearly trends in average antler scores and weights were inconclusive. Gross scores declined among and across years (ANOVA P =0.022; R2 =0.006 P =0.06) but net scores did not decline (P -0.090; Table 14). Antler weights declined but this was possibly due to a different scale used in 1987 which was the year contributing to the decline (Table 14). Within age classes ~ 7 and 4-6, yearly differences in gross and net scores along with regressions to detect trends across years were erratically significant depending upon whether RW or DC ages were used. The only consistency was the negative slopes of marginally significant regressions (Table 16). Animal age and year were 2 potential variables affecting antler measurements and age, whether based on RW or DC, was the primary contributor to yearly differences (P -0.000, Table 18). Antler scores and weights increased through age.8 (Fig. 2). Trends were similar based on either RW or DC ages except that scores and weights were again slightly higher at younger ages for DC. Gross antler scores near 375 occurred primarily at ages ~ 6 but highest net scores near 350 occurred at ages ~ 5. Trends in scores and weights suggested ages 6, 7, and 8 were equally likely to produce the highest scoring antlers. Blk-- Yearly average antler scores and weights were stable or slightly increasing among and across years (ANOVA P ~0.212; R2 ~0.0076 P ~0.06; Table 15). A similar pattern occurred within age classes ~ 7 and 4-6 (Table 17). As with bucks, age composition, whether based on RW or DC, was a primary contributor to yearly fluctuations in antler measurements (P =0.000, Table 18). Body Measurements Eviscerated weights of bucks differed among years and tended to decrease from 1987-1991 (ANOVA P =0.000; ~ =0.009 P =0.02; Table 19). However, yearly fluctuations resulted primarily from effects of age (RW or DC) and date of harvest (P <0.004) on body weight. We would expect bucks to lose body weight between September and the rut in December because of declines in body fat. . Bucks-- Eviscerated weights of bulls differed among years and decreased from 1987-1991 (ANOVA P =0.023; R2 =0.018 P =0.000; Table 19). As with bucks, yearly fluctuations resulted primarily from effects of age (RW or DC) and date of harvest (P <O.OOO) on body weight. We would expect bulls to lose body weight from before the rut in early September to after the rut in mid-October because of rapid declines in body fat. Bull.-- �58 Table 13. Yearly compari8ons between replacement and wear (RW) and dental cementum (DC) age. of female elk harvested on Forbes Trinchera Ranch during public hunting seasons, 1986-1991- Data represent only those elk aged by both technigyes. Chi-square Aging Age Categoa Calf5-7 8-10 n Yearly (P) Year Technique 1 2 3-4 ~11 RW DC 2 2 4 1 5 5 3 5 4 2 1 1 3 20 20 0.509 2 1987 RW DC 12 12 7 6 6 9 14 10 8 15 13 4 1 5 61 61 0.091 1988 RW DC 11 12 6 5 8 8 12 17 16 16 9 5 7 6 69 69 0.899 1989 RW DC 5 5 4 4 4 3 9 11 13 16 9 4 5 6 49 49 0.849 1990 RW DC 8 8 7 8 6 4 8 10 8 7 8 7 4 5 49 49 0.988 1991 RW DC 7 7 3 3 9 10 29 22 11 14 18 13 7 15 84 84 0.532 ALL RW DC 45 46 31 28 34 39 77 73 61 72 59 34 25 40 332 332 0.069 1986 'Includes male and female calves. Table 14. Gross and net Boone and Crockett antler scores and antler weights for buck deer harvested during private fee-only hunting seasons on Forbes Trinchera Ranch, 1987-1991. Results of 1-way AHOVA (Type III 55) and linear regression of antler variables among years are given. Antler Variable 1987 1988 Year 1989 '-Test 1990 Gross Score Net Score n wt. n (kg) 153106 158108 2.66103 2.0710 102 15010 108 1.9310 107 1991 P Linear Regression P R2 Slope 163· 0.022 0.06 153'" 116 0.090 0.19 2.0910 115 0.000 0.00 0.006 Neg. 0.080 Neg. -Oifferent letters denote yearly means that were significantly different based on Tukey's HSD test P - 0.05. �59 220 w 0:200 o U CJ) 180 0: w ...J •••• Z 160 « 140 CJ) CJ) 0120 0: o 100 2 3 4 5 6 7 8 9+ 6 7 8 9+ 6 7 8 9+ 220 RW AGE I2SZS2J DC AGE W200 0: o U180 CJ) 0: ~160 •••• ~140 tuZ120 100 2 - 3 4 5 5 CJ) ~4 •••• ~3 m 3:2 0: W ...J •.•• 1 Z « o 2 3 4 5 AGE (Years) Fig. 1. Average gross and net Boone and Crockett antler scores and weights for buck deer harvested during private fee-only hunting seasons on Forbes Trinchera ranch, 1987-1991. Ages determined by replacement and wear (RW) and dental cementum (DC). Vertical lines represent maximums. �60 ~ o ~ a: w ~ Z -c 400 375 350 325 300 275 250 225 CJ) CJ) 200 o 150 125 100 o a: ____ - _RWAGE._~_DC 175 3 4 5 6 7 8 9+ 3 4 5 6 7 8 9+ 5 6 7 8 9+ 2 ~ o ~ 400 375 350 325 300 275 a: 250 ~ 225 ~ 200 -c ~ 175 150 Z 125 100 2 en 14 (!) ~ 12 I- :c 10 jjj 8 a: 6 ~ IZ 4 (!) :: < 2 o 2 3 4 AGE (Years) Fig. 2. Average gross and net Boone and Crockett antler scores and antler weights for bull elk harvested during private fee-only hunting seasons on Forbes Trinchera ranch, 1987-1991. and dental cementum Ages determined by replacement and wear (RW) (DC). Vertical lines represent maximums. �61 30 -.----------------------------------------------~ :I f/) z 25 o h: 20 -+------ ~:! i ...• ~---- MEDIAN DATES 1986 _ 1987._.._.._ 1988 . 1989 _ 1990_ 1991 ------ W U z 8 15 IZ W o a: 10 ·_------------------------1 w a. ~ 1986-1991 n = 242 5 o 9/8 9/18 9/28 10/8 10/18 10/28 11/7 MONTH AND DAY (BEGIN 5-DAY INTERVALS) Fig. 3. Conception dates for elk on the Forbes Trinchera Ranch, 1986-1991. 11/17 �62 Table 15. Gross and net Boone and Crockett antler scores and antler weights for bull elk harvested during private fee-only hunting seasons on Forbes Trinchera Ranch, 1987-1991. Results of I-way ANOVA (Type III SS) and linear regression of antler variables among years are given. Antler Variable 1987 1988 1989 1990 1991 Gross Score 266· 276· 278· 283· 280· 0.214 0.04 0.008 Pos. Het Score n 256· 89 266· 87 267· 103 273· 115 268· 101 0.212 0.06 0.007 Pos. 6.34· 88 6.22· 85 5.9594 6.57· 110 6.22· 96 0.403 0.83 wt. (kg) n Year F-Test P Linear Regression R~ Slope P "'DIfferent letters denote yearly means that were significantly different based on Tukey's HSD test p = 0.05 •. �63 Table 16. Gross and net Boone and Crockett antler scores and antler weights for buck deer aged ~ 7 years and 4-6 years old by replacement and wear (RW) and dental cementum (DC). Animals harvested during private fee-only hunting seasons on Forbes Trinchera Ranch, 1987-1991. Results of 1-way ANOVA (Type III 55) and linear regression of antler variables among years are given. RW age based on E. Ryland. Antler Variable 1987 RW AGE >7 Years Gross Score 168"" Net Score wt. (kg) Year 1989 1990 173- 16Sab 161" 162· 56 n 2.8358 F-Test 1988 1991 P Linear Regression P Rl Slope 0.004 0.06 0.014 Neg. 152" 48 156ab 34 0.035 0.11 0.011 Neg. 0.151 Neg. 2.25" 54 2.05" 46. 2.06" 48 2.20" 34 0.000 0.00 DC AGE >7 Years Gross Score 167· 175· 163· 160· 168· 0.058 0.22 Net Score n 15615 162· 17 146· 15 150· 21 157· 15 0.161 0.41 2.81· 16 2.38ab 17 2.07" 14 1.97" 21 2.23" 15 0.000 0.00 165· 164- 162· 163- 0.513 0.47 146· 42 154· 48 155· 48 151· 43 154· 64 0.287 0.18 2.4039 1.90" 45 1.93" 47 1.97" 42 2.11"" 0.000 63 0.18 0.015 0.01 0.018 Neg. 0.050 0.06 0.010 Neg. 0.000 0.00 0.101 Neg. n wt. (kg) n RW AGE 4-6 Years Gross Score 157Net Score n wt. (kg) n DC AGE 4-6 Years Gross Score 17r 161" Net Score n 67 16065 151" 71 71 2.76· 65 2.06" 63 1.97" 69 2.12" 70 Wt. (kg) n 157"" 2.01" 54 155"" 0.149 Neg. -oifferent letters denote yearly means that were significantly different based on Tukey's HSD test P - 0.05. �64 Table 17. Gross and net Boone and Crockett antler scores and antler weights for bull elk aged ~ 7 years and 4-6 years old by replacement and wear (RW) and dental cementum (DC). Animals harvested during private fee-only hunting seasons on Forbes Trinchera Ranch, 1987-1991. Results of 1-way ANOVA (Type III SS) and linear regression of antler variables among years are given. RW age based on E. Ryland. Antler Variable 1987 1988 Year 1989 RW AGE >7 Years Gross Score 303· 325· 287· 15 312· 8.51· 14 8.65· Net Score n wt. (kg) n 32P 32P 326· 0.158 0.04 0.033 Pos. 309· 26 310· 48 313· 26 0.100 0.03 0.040 Pos. 8.04· 22. 8.40· 46 8.62· 26 0.793 0.79 9 319· 342· 324· 326· 0.724 0.63 299· 17 326· 15 312· 21 313· 15 0.475 0.21 8.70· 17 8.96· 14 8.26· 21 8.66· 15 0.715 0.38 286· 280· 277· 0.258 0.50 26P 58 275· 63 268· 61 268· 49 269· 55 0.382 0.62 6.31· 58 6.55· 61 5.88· 57 5.93· 47 5.94· 51 0.155 0.06 293· 303· 290· 288· 0.183 0.80 277· 38 283· 56 291· 53 282· 64 277· 60 0.275 0.74 7.22· 37 6.85· 55 6.95· 47 6.72· 62 6.3P 55 0.189 0.02 Net Score n n 9.20· 16 RW AGE 4-6 Years Gross Score 27P Net Score n Wt. (kg) n DC AGE 4-6 Years Gross Score 286· Net Score n Wt. (kg) n Linear Regression P Rl Slope 1991 DC AGE >7 Years Gross 315· Score Wt. (kg) F-Test 1990 9 P 0.013 Neg. 0.020 Neg. -Different letters denote yearly means that were significantly different based on Tukey's HSD test P 0.05. = �65 Table 18. Results of analysis of variance (Type III SS) for effects of animal age and year on gross and net Boone and Crockett antler scores and antler weights (kg) for buck deer and bull elk aged by replacement and wear (RW) and dental cementum (DC). Animals harvested during private fee-only hunting seasons on Forbes Trinchera Ranch, 1987-1991. RW age based on E. Ryland. F-Test P Net Score Weight Variation Source Gross Score RW Age Year Year*Age Model 0.000 0.928 0.518 0.000 0.000 0.834 0.552 0.000 RW Age Year Year*Age Model 0.000 0.330 0.109 0.000 0.000 0.46~ 0.183 0.000 P F-Test Net Score Variation Source Gross Score 0.000 0.084 0.692 0.000 DC Age Year Year*Age Model 0.000 0.936 0.612 0.000 0.000 0.526 0.296 0.002 0.000 0.137 0.801 0.000 0.000 0.386 0.321 0.000 DC Age Year Year*Age Model 0.000 0.378 0.295 0.000 0.000 0.220 0.210 0.000 0.000 0.247 0.785 0.000 Weight Table 19. Eviscerated body weights (kg) for buck deer and bull elk harvested during private fee-only hunting seasons on Forbes Trinchera Ranch, 1987-1991. Results of I-way ANOVA (Type III SS) and linear regression of body weight among years are given. Antler Variable Body Weight n 1987 1988 84.8ob 107 87.7· 108 225.9· 87 223.0ob 85 Year 1989 F-Test P Linear P Regression Slope 1990 1991 82.311 108 84.111 O.000 116 0 •02 0.009 Neg. 220.2112 211.4110.023 100 0.00 0.018 Neg. R2 Elk Body Weight n "DIfferent on Tukey's 224.0102 letters denote yearly HSD test p 0.05. F..al••-Dee~ = means that were signifIcantly different based Eviscerated weight, total body length, and hind foot length were stable among and across years for male and female fawns and when fawns were pooled (P >0.08). Male fawns were heavier than females (P 80.006) but hind foot lengths were similar between sexes (P >0.07, Table 20). Although there was no increase in fawn weight across years, there was no evidence of declining weights. Yearling females were also stable in body weight and size among years (P >0.15). Adult weights stabilized after age 4 (Table 20). �66 Table 20. Eviscerated deer and elk harvested 1991. body weights (kg) and hind foot lengths (cm) for female during public seasons on Forbes Trinchera Ranch, 1987- Age-IeSlrs Deer Weight All Avg. (SE) n Hind Foot Fawns M Avg. (SE) n 1 F 2-3 4-6 ~ 7 22.3 0.28 122 23.30.47 44 21. 7b 0.33 77 36.6 0.42 84 42.8 0.28 245 44.5 0.33 197 44.1 0.49 97 41.1 0.23 122 41.6c 0.53 44 40.8c 0.20 78 46.2 0.39 84 47.6 0.43 245 47.9 0.24 197 47.6 0.18 97 Age-Years nk Weight All Avg. (SE) n Hind Foot Avg. (SE) n ~ilv~i M F 1 65.5 1.56 69 66.74 2.44 27 64.54 2.05 34 56.0 0.30 69 56.1c 2.44 28 56.0c 60.1 2.05 0.46 38 36 2 108.1 134.0 2.75 2.00 36 40 ~~ Different letters within rows denote male and female fawns or calves. 62.2 0.43 40 means 3-4 5-7 8-10 ~11 140.8 1.78 68 148.3 1.47 71 155.0 1. 79 35 151.1 2.49 37 62.4 0.33 68 62.4 0.28 71 63.0 0.36 35 63.0 0.44 37 that were different between '..al••-B~k-- Eviscerated .table (ANOVA across weight weight through weight, total body length, and hind foot length were among years for male and female calves and when calves were pooled P >0.06). However, body weight of calves for sexes pooled increased years (R2 =0.081, P -0.027) Male and female calves were similar in and size (P >0.018, Table 20). Yearling females were stable in body and size among years (P >0.15). Adult weights continued to increase age 10 (Table 20). Elk Reproduction Pregnancy Ra~e.-- Pregnancy rates for adult cows (~ 1 yr old) averaged 80% and ranged from 62-92% during 1986 to 1991. Rates differed among years for all adults pooled and for adults ~ 2-years old (P ~0.002). Declining pregnancy rates from 1986 through 1990 may reflect the drought conditions of those years. Pregnancies were not detected in calves (6 mos old) and rates for yearlings were variable and averaged 17%. For mature adults, rates were lowest (78%) in cows ~ II-years old (Table 21) •. Litter size was 1 except for 4 sets of twins (2% of 244 litters). Twins consisted of 3 sets of 2 females (1986, 1989, 1990) and 1 with 1 male and 1 female (1986). Cows with twins were 4, 6, 9, and 12 years old. Infected uteri not capable of supporting pregnancy occurred in 5 (1.5%) of 356 adults examined; these occurred in 1 yearling 4 adults ~ 13 years old with 3 in 1986 and 1 in 1989. �67 Table 21. Pregnancy rates in December, 1986-1991. Age(yrs) 1 2 3-4 5-7 8-10 >11 Uk-Ad" Totals 1986 PG n 5 7 6 5 2 2 7 6 5 2 5 5 1 1 31 \PG 28 90 1987 n PG 6 10 (PG) for female 1988 PG n 1989 n PG elk on the Forbes 1990 PG n 0 7 2 5 1 8 0 10 6 4 2 10 15 7 3 15 14 4 4 3 4 12 16 8 8 8 6 7 8 46 8 8 14 12 '4 4 4 2 2 2 17 16 4 6 4 48 38 60 79 'Age for ~ 2 from replacement "Adults of unknown age. 4 6 1 48 77 and wear; 4 3 1 37 5 4 1 0 45 77 for ~ 28 62 3 Trinchera 1991 n PG n 4 10 22 12 14 14 3 1 21 12. 14 13 3 79 73 9 92 from dental Ranch 1986-1991 PG %PG 35 44 73 71 36 40 12 6 36 66 65 36 31 10 17 82 90 92 100 78 83 311 250 80 80 cementum. Fetal Sez Ratios and Body si.e-- Fetal sex ratios deviated from 50 M:50 F in 4 of 6 years and inconsistently favored males or females (P ~0.05). However, the pooled ratio among years of 55 M:45 F was not different from 50:50 (P >0.05). Fetal sex did not deviate from unity within age classes of cows (P -0.05) (Table 22). Changes in fetal sex among years could not be explained by covariables cow body weight or age (P >0.05, logistic regression, n =173). We tentatively conclude that yearly variation in fetal sex ratios reflects .ampling error. During 1987~1991, fetal weights and crown-rump lengths differed by year and sex with male. being heavier and slightly longer than females (P <0.011, Table 23). Yearly differences were attributable to males whose weights were heavier in 1987 and 1990 (P =0.003) while female weights were stable among years (P =0.471). Unlike females, male weights declined from 1987 to 1991 (R2 =0.051, P -0.012). We are cautious about interpreting differences in fetal weights until fetal body measurements are adjusted for slight yearly differences in fetal ages (Table 24). Conception Dates-Mean and median conception dates ranged from 23 september to 4 October with the earliest breeding dates occurring in 1990 and 1991 (Table 24). The earliest and latest breeding dates were 6 September and 17 November, respectively. Increasing harvests of mature bulls during the rut have been associated with a trend in earlier, not later conception dates. We su.pect that yearly weather variables may explain .orne of the changes in conception date •• Disease Surveys-- All serum samples collected during December 1991 from elk on Forbes Trinchera Ranch tested negative for brucellosis. The 84 samples were from 8 male and 9 female calves and 67 adult females. Progesterone Assays-- Final assay validation trials for elk progesterone RIA are still in process at Colorado state University. When this data is received we will finalize our manuscript on using progesterone assays to determine pregnancy status in elk. �68 Table 22. Petal sex ratios observed in famale elk on the Forbes Trinchera Ranch in December, 1986-1991. Age" (yrs) M 1 2 3-4 5-7 8-10 11 Uk-Ad· 1 1 4 5, 1 5 0 1987 F U 1989 F U M 1 1 3 4 0 3 0 0 2 7 11 3 0 1 0 1 3 2 6 1 0 M 0 2 1 0 0 1 2 1 4 1 0 1 0 0 0 4 2 2 3 1 0 Totals 17 6 7 12 24 2 26 15 5 12 24 2 13 15 1 48 22 3 'Male 74 33 63 33 46 69 0 6 5 9 1 1 2 M 0 0 1 1 0 0 0 1988 F U M 1986 F U 1 2 8 11 1 1 2 0 2 5 3 3 2 0 1 0 2 0 0 1 1 0 0 0 1 1 0 0 1990 F U 0 2 4 5 1 3 0 M 0 0 0 0 1 0 0 1991 F U 0 5 13 9 10 9 2 1 3 7 3 4 3 1 1986-91 M F U 3 14 33 33 21 20 4 0 1 1 0 0 1 0 1 17 29 31 12 10 6 , Male 2 5 5 2 3 2 1 128 106 20 75 45 53 52 64 67 40 55 55 IAge for ~ 2 from replacement and wear; for ~ 3 from dental cementum. "Adults of unknown age. Table 23. Measurements of elk fetuses collected from Forbes Trinchera Ranch in December, 1986-1991. Datel statistic Bod:ll!: Weight {g} Male Female Crown-rum:e {mm} Male Female Hind Foot {mm} Male Female 29 Nov-5 Dec 1986 Mean BE n 27.5 3.32 17 20.8 2.82 6 89.9 4.06 17 85.8 4.22 5 20.4 1.14 17 19.5 1.20 6 12-14 & 19-21 Dec 1987 Mean 118.7 BE 15.78 n 12 59.7 7.04 23 141.4 6.10 12 115.4 4.76 24 42.3 3.15 12 30.3 1.81 24 10-12 & 17-19 Dec 1988 Mean 74.5 BE 6.91 n 26 66.7 7.68 15 120.0 3.90 26 121.7 4.63 15 32.6 1.51 26 33.1 1.80 15 9-11 & 16-18 Dec 1989 Mean 57.6 BE 17.31 n 12 60.4 9.27 24 102.4 10.82 12 111.5 6.81 24 27.9 3.70 11 29.9 2.26 23 8-17 Dec 1990 Mean BE n 92.2 14.55 13 78.4 11.50 14 128.0 6.96 13 127.6 5~85 15 35.1 2.69 13 35.7 2.72 15 7-15 Dec 1991 Mean BE n 65.3 5.95 47 77.7 11.93 22 115.6 3.92 48 123.0 5.87 22 30.4 1.41 48 34.1 2.35 22 �69 Table 24. Estimated conception Trinchera Ranch, 1986-199l. Year Mean dates and fetal ages for elk on Forbes ConceEtion Median Mode Date Min. Max. n Average Fetil Age (l21:ll:s l Female Unk. Male 1986 1 oct 30 sep 2 oct 19 sep 23 Oct 26 67 65 46 1987 3 Oct 2 Oct 29 Sep 18 Sep 17 Nov 37 82 74 34 1988 30 Sep 29 Sep 27 Sep 15 Sep 25 Oct 44 75 76 51 1989 3 Oct 4 Oct 4 Oct 18 Sep 24 Oct 35 70 73 0 1990 24 Sep 23 Sep 21 Sep 14 Sep 7 Oct 27 78 78 0 1991 26 Sep 25 Sep 26 Sep 6 Sep 2 Nov 73 74 76 44 COHCLUSIOHS Effects of increasing harvests of deer and elk should be viewed in terms of initial harvest objectives. OUr strategy was to increment male harvests to levels where changes in age structure or antler quality signaled a need to reduce harvest to provide desired yields of trophy male animals. For females, we hypothesized that harvest could be increased until age of harvested animals declined. Declining age of bucks harvested and inconclusive but possible declining antler quality suggest caution in either maintaining or increasing harvests to detect more conclusive changes. Increasing or stable ages of bulls harvested and stability in antler quality suggest maintaining harvest at current levels to detect declines in age and antler quality. Increasing ages of female deer harvested .uggest current levels may be .ustainable and beneficial to producing a more thrifty female deer population. Alternatively, harvests composed of increasingly older females could signal inadequate recruitment of younger adults. Stable ages of harvested female elk suggest current levels of harvest can be sustained. We hypothesized that increasing harvests of female deer and elk would lower densities of females and positively impact net recruitment by improving nutritional status of young animals. We have not yet detected an increase in fawn or yearling deer body weights, but neither has a decline occurred, while weights of elk calves increased. Factors unrelated to increasing harvests have likely contributed to variable pregnancy rates and fetal sex ratios and earlier conception dates for elk. LITERATURE Armstrong, R. A. 1950. Fetal development Amer. Mid. Nat. 43:650-666. CITED of northern white-tailed deer. Boyd, R. J., and E. E. Ryland. 1971. Breeding dates of Colorado elk as estimated by fetal growth curves. Colo. Game, Fish and Parks Div. Game Info. Leaflet 88. 2pp. �70 Freddy, Keiss, D. J., E. E. Ryland, and R. M. Hopper. 1991. Colorado'a wildlife ranching program: the Forbes Trinchera experience. In: Wildlife Production: conservation and sustainable development. eds. L. A. Renecker and R. J. Hudson, pp. 336-343. AFES misc. pub. 91-6. Univ. Alaska Fairbanks, Fairbanks. R. E. 1969. comparison of eruption-wear patterns and cementum as age criteria in elk. J. Wildl. Manage. 33:175-180. Morrison, J. A., C. E. Trainer, and P. L. Wright. elk as determined from known-age embryos. 34. Nesbitt, W. H., and J. Reneau (eds.). 1986. big game awards. Boone and Crockett Quimby, Boone Club, annuli 1959. Breeding seasons in J. Wildl. Manage. 23:27- and Crockett Club's Dumfries, Vermont. 19th D. C., and J. E. Gaab. 1957. Mandibular dentition as an age indicator in Rocky Mountain elk. J. Wildl. Manage. 21:134-153. Robinette, W. L., D. L. Jones, G. Rogers, and J. S. Gashweiler. 1957. Notes on tooth development and wear.for Rocky Mountain mule deer. J. Wildl. Manage. 21:134-153. SAS Institute, Inc. 1028pp. 1988. SAS user's guide. SAS Inat. Inc., Cary, N. C. stevens, M. L. 1987. Apparent accuracy of cementum annuli for estimating ages of mule deer. M. S. Thesis, Colorado State University, Fort Collins. Prepared by �71 Colorado Division of Wildlife Wildlife Research Report July 1992 JOB PROGRESS REPORT State of ~C~o~l~o~r~a~d~o~ _ Project No. W-153-R-5 Mammals Research Work Plan No. 3 Elk Investigations Job No. 7 Elk Census Methodology Period Covered: Author: July I, 1991 - June 30, 1992 D. J. Freddy Personnel: R. Bartmann, G. White Abstract Estimates of population sizes for elk (Cervus elaphus) and mule deer (Odocoileus hemionus) in the Troublesome subunit of Middle Park, Colorado were obtained using systematically spaced aerial line transects. Additionally, an estimate of population size for deer was obtained using a random quadrat (2.59 km2) system. All flights were conducted in January or March 1990-1992 using a Bell-Soloy helicopter. The exponential polynomial (EXPL) or negative exponential (NEXP) sighting models provided the best combination of precision and model fit for both elk and deer. Precision (95% CI as % mean) depended on sighting model and year and ranged from ± 43-48% for elk and ± 23-67% for deer. Estimates of mean population size varied with sighting model and year and ranged from 3,404-3,500 for elk and 2,759-7,012 for deer. We recommend using the EXPL model with an effective strip width of 95 or 115 in. Line transects provided higher estimates of total·deer during 2 years (P ~0.06) and potentially better precision than quadrats depending on which sighting model was selected but quadrats were less costly to impl~m.ent.· Differences .. in ' population estimates for deer between sampling methods were greater than. expected with the unexpectedly high estimates derived from line transects potentially influenced by "heaping" of deer groups into the center line interval. We are hesitant to recommend line transects instead of quadrats to estimate populations of deer in sagebrush (Artemisia tridentata sp.) habitats until further evaluations can assess the problems of assigning deer to the center line interval of transects. We suspect that line transects for estimating numbers of elk will be unacceptably imprecise because of difficulties in sampling adequate numbers of elk groups. ��73 JOB PROGRESS REPORT ELK CENSUS METHODOLOGY David J. Freddy P. N. OBJECTIVE Evaluate methods to estimate numbers of elk during winter. SEGMENT OBJECTIVES 1. Determine efficacy of line transect methodology elk and mule deer sharing sagebrush-aspen-conifer 2. Compare estimates of mule deer density based mi2 quadrats in sagebrush winter range. 3. Prepare rates. 4. Prepare an approved Program Narrative to evaluate sampling methodologies for estimating numbers of deer or elk using radiocollared animals as a known population. an approved Program Narrative to estimate numbers winter ranges. on 1 mi2 quadrats to estimate elk/deer of and 1/2 survival INTRODUCTION Reliably estimating numbers of elk (Cervus elaphus) has been difficult and expensive. Bear et al. (1989) evaluated mark-resighting to estimate population size and they found this method was sensitive to colors of markers placed on elk and costly but provided precise estimates of population size that were likely negatively biased. Samuel et al. (1987) and Unsworth (1990) used radiocollared elk to develop sightability factors for correcting negatively biased counts of elk during surveys attempting to completely count all elk. Most methods underestimate true densities of ungulates (LeResche and Rausch Underestimates occur because all animals are not observed on defined sample units such as quadrats or strip-transects. Line transect theory allows observers to miss animals, except those occurring directly on the center line of the transect, and therefore, could provide less negatively biased estimates of population density (Burnham et al. 1980; White et al. 1989). We chose to assess the field logistics and precision of aerial line transects as a sampling method to estimate densities of elk. We also evaluated line transects for estimating densities of mule deer (Odocoileus hemionus) (White et al. 1989) and compared these estimates of density with estimates derived from a quadrat sampling system employed over the same geographic area (Gill 1969). We conducted northcentral our project in the Troublesome subunit of Middle Park, Colorado. During winter, significant numbers of elk and mule deer �74 usually migrate and concentrate into a 250 km2 area which can be efficiently sampled with aerial surveys. Vegetation within winter ranges frequented by elk is primarily sagebrush (Artemisia tridentata wyomingensis), aspen (Populus tremuloides), conifer forests (~contorta, Pseudotsuia menziesii, Picea en&elmanni), and mixed stands of aspen-conifer. For deer, winter ranges are dominated by sagebrush and aspen. Terrain varied from rolling hills to steep canyons. METHODS Line Transects Two sets of parallel line transects, each having 26 transects systematically spaced at 1,000 m intervals, were flown to estimate numbers of elk and deer. These 2 sets resulted in transects occurring every 500 m across the sampled winter range. Transects were oriented on true north-south bearings perpendicular to changes in elevation and expected gradients in elk and deer densities (White et ale 1989), were delineated on 1:24,000 scale topographic maps, and were not marked with flight markers on the ground. Line transects flowri to estimate densities of elk totaled 948 km (588 mi) in length, were 1-15 km long, and were distributed within an area of 246 km2 (95 mi2) between elevations of 2,257-2,990 m (7,400-9,800 ft). Transects were limited to those areas where elk generally reside during January and February. The same line transects were flown to estimate densities of deer but only deer found at elevations ~ 2,590 m (8,500 ft, primarily sagebrush habitat) were counted. With this restriction, line transects for deer totaled 602 km (374 mi) in length, were 1-14 km long, and were distributed within an area of 166 km2 (64 mi2) which was also the area delineated for a quadrat (2.59 km2, 1 mi2) sampling system (Gill 1969). Transects were flown with a Bell-Soloy helicopter at 65-80 kmph (40-50 mph) and 35-50 m (115-164 ft) above the ground or tree canopy. We flew each set of transects twice (4 replicate flights) and attempted to fly each set once in the morning and once in the afternoon but in some years weather altered this scheduling. Transects were flown on 9-11 January 1990, on 1, 3, 4, and 8 Karch 1991, and on 14-17 January 1992. A navigator and observer were responsible for detecting elk or deer. The observer, seated on the right, estimated perpendicular distances to the geometric center of each group of elk or deer and counted animals in each group located from the transect center line to the right. The navigator, seated in the middle, maintained course bearing using topographic maps, and aided in locating groups on or near the transect line. The pilot, seated on the left, occassionally spotted animals that may not have been seen by observers. The center line interval was defined to include all elk or deer observed from 0-15 m to the right of course bearing; beyond 15 m, elk or deer were grouped into 10 m intervals out to 155 m (White et al. 1989). Observers were experienced in aerial counts of elk and deer and practiced estimating distance intervals each day that flights were conducted by flying the helicopter along a practice transect line and observing markers placed at 10 m �75 intervals from the line. The navigator was the same person in all years but the primary observer and pilot were both different in 1990. Quadrats We compared estimates of population size for deer derived from line transects and quadrats by sampling 30-2.59 km2 quadrats (47% random sample) within the same 166 km2 delineated for deer transects. Quadrats were flown with the same helicopter, pilot, and observers used each year for line transects. Quadrats and transects were flown at similar flight speeds and heights above the ground. At least 1 corner of each quadrat was marked On the ground to aid in locating quadrats from the helicopter. The navigator and observer both searched for and counted deer by first flying the perimeter boundary and then systematically flying the interior of each quadrat (Kufe1d et a1. 1980). Quadrats were flown on 10 January 1990, 3-4 March 1991, and 15-16 January 1992 after completing replicates 1 and 2 of the transects but prior to completing re1icates 3 and 4. Statistical Analysis Estimates of population size derived from line transects followed methods outlined by White et a1. (1989) using program TRANSECT. We considered replicate sets of transects to be independent and thus pooled results from the 4 replicate flights to derive population estimates for elk and deer. We had no indication that counting elk or deer disturbed either sufficiently to cause them to move from 1 transect to another while flying one set of transects and thus we feel duplicated counts within a set of transects did not occur. Estimates of population size derived from quadrats followed methods of Gill (1969) and Mendenhall et a1. (1971) using finite population correction factors for calculating variances. A z-test was used to compare densities of deer estimated by line transects and quadrats. All other analyses were performed using SAS (PROe REG, PRoe MEANS, PRoe GLM, PROC FREQ) (SAS 1985). RESULTS AND DISCUSSION Countin~ Conditions Although we searched for deer and elk simultaneously on line transects, the 2 species were seldom seen in close proximity. We therefore believe that searching for both species had little effect on detecting either species. We found little evidence to suggest elk were moving prior to detection. Elk were often bedded or standing in snow in aspen or conifer habitats. However, deer moved in response to the helicopter much more than did elk presumably due, in part, to shallower snow depths. We compensated for this movement of deer by attempting to estimate perpendicular distances to those points where movement was initiated. Amount of snow cover was variable among years and affected our ability to detect animals, especially deer. In 1990, snow cover was nearly 100% and provided good visual background for detecting both species while during the mild winter of 1991, south-facing slopes were devoid of snow and hindered detecting deer even though we delayed flying until early March so that snow �76 might accumulate. In March, deer are typically less dispersed, more concentrated, and in larger groups than in Janaury, regardless of snow conditions. Counting conditions for deer were excellent in 1992 when nearly 100% of the sample area was covered by ~ 25 cm of snow. However, snow depths at elevations frequented by elk were insufficient in both 1991 and 1992 to concentrate elk within the sample area so we flew only 2 replicates of transects for elk in 1991 and canceled transects for elk in 1992. Even with the generally mild snow conditions during all years, deer were adequately concentrated within the sample area demarcated for deer. There were only 2 and 1 group(s) of deer seen in 1990 and 1991, respectively, along line transects at elevations above the sample area demarcated for deer and in 1992 cursory observations indicated a high percentage of the deer were within the sample area. Line Transects-Elk The EXPL and NEXP models for effective strip widths of 155 or 95 m provided the best combination of model fit (P >0.73), parsimony, and precision for estimates of population size. However, precision (CI as % of mean) was only ± 43-48% in 1990 and much poorer in 1991 when grossly inadequate numbers of elk groups were detected (Table 1). Estimated population sizes of 3404-3500 elk in 1990 were much higher than known minimums in the sample area of 419 in 1990 and 744 in 1988 based on independent helicopter flights. Sighting distributions were strongly spiked at the center line during both years which indicated that the probability of detecting elk away from the center line fell rapidly and effectively ended at 95 m (Fig. 1). This rapid decline in numbers of groups observed with increasing distance was at least partially acceptable because 74% of the elk groups were observed in aspenconifer forested habitats that hindered detecting elk (Freddy 1991). Number of elk per group increased with distance away from the center interval but correlations were weak (~ < 0.041) and regression slopes became nonsignificant at truncation distances ~115 m (P >0.28). Significant positive slopes were dependent on groups having ~30 animals which accounted for 10% of the total groups detected (Table 3, Fig. 3). Bias associated with detecting larger groups (Drummer and McDonald 1987) could therefore be minimized by restricting searching to distances ~115 m. Group size was not different between years (P >0.32) and averaged 10.7 ± 1.45(SE) elk. Model fit, precision, and minimizing bias of group size on sighting probabilities argue for truncating data at 115 or 95 m. Using these truncation widths instead of 155 m may result in reducing numbers of groups for analyses by 8-15% (Fig. 1). Even though groups may be observed beyond 115 m, searching time should be devoted to areas closer to the center line to improve model fit and precision. Line Transects-Deer The NEXP and EXPL models for effective strip widths of 155, 115, or 95 m provided the best combination of model fit (P >0.14), parsimony, and precision for estimates of population size although neither model fit adequately in 1990 (P <0.20). Precision (CI as % of mean) was marginally better at a truncation width of 155 m during all years and was best (± 23-31%) in 1992 when number of �77 deer groups detected was highest and poorest (± 39-67%) in 1991 when the fewest groups were detected (Table 2). In 1992 when groups observed approached the recommended 200, precision approached ± 20% as predicted by White et a1. (1989). Table 1. Population estimates for elk based upon different sighting probability models for line transects, Troublesome subunit, Middle Park, 19901991. Center Interva1/ Truncation (m)/ Groups Mode1a No. Interva1sb n Chisquare (P)C fQ~, Mean Sl~~~ 2~% ~I CI % Elk I!~ndt:£ Upper Lower Mean km2 mi2 1990 January EXPL EXPL EXPL 0-15/155/10 0-15/95/9 0-15/115/6 66 57 62 0.8252 0.7309 0.3317 3500 3404 3266 ·5123 5053 4787 1878 1754 1744 46 48 47 14.3 13.8 13.3 36.9 35.8 34.4 NEXP NEXP NEXP 0-15/155/10 0-15/95/9 0-15/115/6 66 57 62 0.8920 0.8247 0.4907 3500 3404 3265 5009 5000 4742 1989 1806 1788 43 47 45 14.2 13.8 13.3 36.8 35.8 34.4 EXPS EXPS EXPS 0-15/155/10 0-15/95/9 0-15/115/6 66 57 62 0.8210 0.7514 0.3639 3869 3962 3861 7317 9120 8423 421 0 0 89 130 118 15.7 16.1 15.7 40.7 41. 7 40.6 FSERd 0-15/115/6 62 0.3173 2460 3324 1597 35 10.0 25.9 1991 March EXPL EXPL EXPL 0-15/155/10 0-15/95/9 0-15/115/6 28 23 25 0.5863 0.7205 0.3006 2614 2743 2446 5414 5184 4531 0 301 361 107 89 85 10.6 11.2 9.9 27.5 28.9 25.7 NEXP NEXP NEXP 0-15/155/10 0-15/95/9 0-15/115/6 28 23 25 0.6911 0.7786 0.4822 2614 2742 2446 4723 5231 4588 505 255 303 81 91 88 10.6 11.2 9.9 27.5 28.9 25.7 EXPS EXPS EXPS 0-15/155/10 0-15/95/9 0-15/115/6 28 23 25 0.7465 0.7305 0.4237 3393 4367 3160 8962 13119 8970 0 0 0 164 200 183 13.8 17.7 12.9 35.7 45.9 33.3 FSER· 0-15/115/6 25 0.0812 1438 2534 341 76 5.8 15.1 aMode1s are: EXPL -exponential polynomial, NEXP - negative exponential, EXPS - exponential power series, FSER - Fourier Series. bFor truncation distances of 155 or 95 m, cut points for distance intervals beyond the center interval were 10 m; for truncation at 115 m cut points were at 20 m. cMode1 adequately fits data when ~.20. d'·l parameter model in 1990 and 1991. �78 20 ELK MIDDLE PARK TROUBLESOME ~ 15 -I._.-~, ..--_ - --.-.- __ ._ . -.-- SUBUNIT - _- __ ._.- __.___ .__ ._ __ ._._ _. :;) o a: ~ o MAR 1991 n=66 C!J u. JAN 1990 n=28 10 a: w ttl :!: :;) 5 _j _... LA •• Z o 0·15 15-25 25·35 35-45 45-55 55·65 65·75 75-85 85-95 95-155 DISTANCE FROM CENTER LINE (m) 35 ELK MIDDLE PARK TROUBLESOME en D... _ 1990-91 SUBUNIT n=94 5a: 25 C!J u. 20 o ffittl % Truncated 15% 8% 15 :!: :;) 10 Z 5 o 0·15 25·35 45-55 65-75 85-95 105-115 125·135 145-155 15-25 35-45 55-65 75-85 95-105 115·125 135-145 DISTANCE FROM CENTER LINE (m) Fig. 1. Groups of elk observed yearly (TOP) and all years (BOTIOM) within distance intervals for line transects, 1990-1992. Vertical dashed lines represent points for truncating data. �79 Table 2. Population estimates for mule deer based probability models for line transects, Troublesome upon different 8ighting subunit, Middle Park, 1990- 1992. Model· Center Interval! Truncation (mV No. Intervals n chisquare (P)C fo~. Size & 95% CI Mean Upper Lower CI % peer Mean Jan2 Groups Density mi2 1990 January EXPL EXPL EXPL 0-15/155/10 0-15/95/9 0-15/115/6 185 139 155 0.0000 0.0907 0.0000 4833 7012 5854 6379 9621 7909 3288 4404 3799 32 37 35 29.1 42.2 35.3 75.4 109.4 91.3 NEXP NEXP NEXP 0-15/155/10 0-15/95/9 0-15/115/6 185 139 155 0.0000 0.1420 0.0001 4833 7012 5854 6145 10161 8124 3119 3862 3584 27 45 39 14.6 21.1 17.6 37.7 54.7 45.7 EXPS EXPS EXPS 0-15/155/10 0-15/95/9 0-15/115/6 185 139 155 0.0002 0.2024 0.0006 6102 8925 7462 10226 16419 13081 1978 1431 1843 68 84 75 18.4 26.9 22.5 47.6 69.6 58.2 FSEr 0-15/115/6 155 0.8207 7250 9965 4534 37 21.9 56.6 1991 March EXPL EXPL EXPL 0-15/155/10 0-15/95/9 0-15/115/6 123 101 109 0.1797 0.1877 0.1438 4358 3912 4209 6067 6049 7022 2648 1774 1395 39 55 67 26.3 23.6 25.4 68.0 61.0 65.7 NEXP NEXP NEXP 0-15/155/10 0-15/95/9 0-15/115/6 123 101 109 0.2540 0.2706 0.2474 4357 3910 4208 5995 5547 5875 2720 2272 2541 38 42 40 13.1 11.9 12.7 34.0 30.5 32.8 EXPS EXPS EXPS 0-15/155/10 0-15/95/9 0-15/115/6 123 101 109 0.2565 0.1964 0.1777 5363 4376 5096 9609 8182 9631 1118 570 561 79 87 89 16.1 13.2 15.3 41.8 34.1 39.7 FSEJtO 0-15/115/6 109 0.1510 3600 4905 2295 36 10.9 28.1 1992 January EXPL EXPL EXPL 0-15/155/10 0-15/95/9 0-15/115/6 254 186 212 0.4887 0.3765 0.1845 3006 2759 2937 3686 3465 3650 2327 2052 2223 23 26 24 18.1 16.6 17.9 46.9 43.0 45.8 NEXP NEXP NEXP 0-15/155/10 0-15/95/9 0-15/115/6 254 186 212 0.5974 0.4902 0.3050 3006 2758 2941 3819 3622 3805 2194 1895 2078 27 31 29 9.1 8.3 8.8 23.5 21.5 22.9 EXPS EXPS EXPS 0-15/155/10 0-15/95/9 0-15/115/6 254 186 212 0.5671 0.0209 0.2489 3347 1999 3231 5474 2471 5594 1220 1528 869 64 24 73 10.1 6.0 9.7 26.1 15.6 25.2 FSERe 0"':15/115/6 212 0.1648 2614 3273 1956 25 7.9 20.4 ~odels are: EXPL =exponent1al polynom1al, NEXP = negative exponent1al, EXPS exponential power series, FSER Fourier Series. ~or truncation distances of 155 or 95 m, cut points for distance intervals beyond the center interval were 10 m; for truncation at 115 m cut points were at 20 m. ~odel adequately fits data when P~0.20. ~4 parameter model in 1990, 2 parameter model 1991, 1 parameter model 1992. = = �80 Table 3. Summary of linear regre.aion .tatiatica 8ize ver8U8 8ighting di8tance for line tran8ect8, Park, 1990-1991 for elk and 1990-1992 for deer. Group Size Sighting Distance R2 EIk Probe (P) B = 0 of elk and mule deer group Troublesome 8ubunit, Middle Mule Deer Probe (P) B = 0 All All All All All ~155m ~115m ~95m 0.0289 0.0412 0.0136 0.0280 0.0743 0.0498 0.2816 0.9381 0.0325 0.0193 0.0108 0.0005 0.0001 0.0010 0.0235 0.6331 <30 <30 <30 ~155m ~115m ~95m 0.0012 0.0013 0.0021 0.7536 0.7551 0.7000 0.0110 0.0053 0.0000 0.0132 0.1151 0.9842 Sighting distributions for deer were also strongly spiked at the center line during all years which indicated that the probability of detecting deer away from the center line fell rapidly (Fig. 2). The rapid decline in numbers of group8 observed at distances >15 m is of concern and cannot be explained by dense vegetation hampering observation of deer as may have been the case in pinyon pine (Pinus edulis)-Utah juniper (Juniperus osteosperma) vegetation (White et ale 1989). We were concerned that "heaping" of groups into the center interval occurred because of incorrectly including groups from the left of the center line or from distance intervals near and to the right of the center interval (Burnham et ale 1980, Buckland 1985). Such "heaping" would inflate estimates of deer density (White et ale 1989). In the open and predominately non-forested habitats that we encountered, deer near the center line could be seen ahead of the helicopter at distances of 200 m. To avoid mi8sing deer on the center interval, decisions as to whether 8uch animals were either left or right of the transect were made in relation to predetermined lines of flight; but perception of their location in relation to an imaginary transect line can be influenced by helicopter yaw (White et ale 1989). Secondly, such groups were placed into a distance category when the helicopter passed perpendicular to the observed location of the group. Any slight tendency for the helicopter to "8teer" towards the group would result in underestimating distances to groups. Number of deer per group increased with distance away from the center interval but correlations were weak (R2 < 0.033) and regression slopes became nonsignificant at a truncation distance of 95 m (P >0.60). Significant slopes were dependent on groups having ~30 animals which accounted for 1\ of the total groups detected (Table 3, Fig. 3). Bias associated with detecting larger groups (Drummer and McDonald 1987) could therefore be minimized by restricting 8earching to distancea ~95 m. Group size was smallest in 1992 (P <0.0001) and largest in 1991 (Table 4). These relative sizes of groups corresponded to the most and least precise population estimates (Table 2). Estimated population size decreased from 1990 to 1992 for models based on truncation widths of 95, 115, or 155 m, although the decline was most pronounced for data truncated at 95 m (7,012 to 2,759 deer; P <0.01, Fig. 4, Table 2). The decline waa contrary to perceived atable or increasing trends in the population. We suggest the decline may reflect increasing experience of observers and their improving ability to assign groups of deer to proper distance categories, especially the center interval. The proportion of groups assigned to the center interval declined for models based on either 95 or 155 m truncation distances; for 95 m, proportions declined from 0.44 to 0.25 from 1990 through 1992 (Fig. 2). �81 80 ~--------------------------------------------~ MULE DEER MIDDLE PARK TROUBLESOME ~ 60 ::::> o a::: ~ MAR 1991 JAN 1990 ~ 1992 n=254 n=123 n=185 o IZj JAN SUBUNIT 40 0:. W CO ~ ::::> 20 Z o 0-15 25-35 15-25 45-55 35-45 55-65 65-75 75-85 85-95 95-155 DISTANCE FROM CENTER LINE (m) 160 MULE DEER MIDDLE PARK TROUBLESOME SUBUNIT 140 CJ) _ Q.. 1990-92 n=562 ::::>120 o 0: (!)100 LL. 080 a::: w % Truncated CO60 ~ 24% 15% ::::> z 40 20 o 0-15 25-35 15-25 45-55 35-45 65-75 55-65 85-95 75-85 105-115 125-135 145-155 95-105 115-125 135-145 DISTANCE FROM CENTER LINE (m) Fig. 2. Groups of mule deer observed yearly (TOP) and all years (BOnOM) within distance intervals for line transects, 1990-1992. Vertical dashed lines represent points for truncating data. �82 35 (33) w o a: :::> _--_ _.1990-1991 __ .... _._ ...... --_._-_._.. IIiI Z W ELK MIDDLE PARK TROUBLESOME ... 30 . _ .... .•.. SUBUNIT - -- ..... n=111 25 o o 0 20 w ~ ~ Z 15 c a: 10 W W c.. 5 0 1-2 3-4 5-6 7-9 10-14 15-19 20-29 30+ GROUP SIZE 50 MULE DEER MIDDLE PARK TROUBLESOME SUBUNIT ~ 40 ~-------------------------------------------------------I Z 1990-1992 n=616 w a: (189) G 30 o o w C) 20 ~ Z W ~ 10 W c.. (8) o 1-2 3-4 5-6 7-9 10-14 15-19 20-29 30+ GROUP SIZE Fig. 3. Sizes of groups for elk and mule deer observed on line transects, 1990-1992. Sample sizes in parentheses; includes 11 groups of elk and 54 groups of deer observed beyond 155m from center line. �83 12,000 MULE DEER MIDDLE PARK TROUBLESOME -+ _-_._-_._-_ _ .. _ _ --.-.-----.---.----... - -- -.-.__.. _ .. _- QUADRATS MEAN _ .. _. __ .__ SUBUNIT _-_.__._ _ __ .___. . __ ._._ ____ .. _---I QUADRATS +/- 95% CI I:::::::::::::::::::::::::::::::j TRANSECTS +/- 95% CI o 1990 1991 YEAR Fig. 4. Estimates of population size and 95% confidence intervals for deer based on quadrats and line transects, 1990-1992. Darkly shaded area denotes overlap of confidence intervals. Estimates for transects based on the EXPL model with truncation at 95m. 1992 �84 Table 4. Average group sizes of mule deer observed on line transects at all distances from the center line and on 2.59 km2 quadrats, Troublesome subunit, Middle Park, 1990-1992. Average Group Size n Year 95% Conf. Int. Lower Upper Minimum Maximum Line Transects 1990 1991 1992 All 208 132 276 616 6.087.554.3510 5.62 5.43 5.79 3.88 5.13 6.72 7.55 4.82 6.11 1 1 33 68 32 68 not available 61 13.05c 6.40" 172 233 8.14 9.03 5.50 6.87 17.07 7.30 9.41 1 1 1 97 34 97 1 1 Qyad£ats 1990 1991 1992 All '~Ifferent P<O.OOOI. ~ifferent letters denote different yearly averages for line transects, letters denote different yearly averages for quadrats, P<O.OOOI. Model fit, precision, and minimizing bias of group size on sighting probabilities argue for truncating data at 115 or 95 m. Using these truncation widths instead of 155 m may result in reducing numbers of groups for analyses by 15-24% (Fig. 2). Even though groups may be routinely observed beyond 115 m in open sagebrush vegetation, searching time should be devoted to areas closer to the center line to improve model fit and precision. Qyadrats vs Line Transects-Dee£ Estimated population size for deer based on quadrats was stable or slightly increasing from 1990 through 1992 compared to the declining trend indicated by line transects (Fig 4). Quadrat estimates of 1,848 in 1990 and 2,361 deer in 1992 were not different (P >0.20) (Table 5). However, estimates for quadrats were lower than transects in 1990 (P <0.01) and 1991 (P <0.06) but not in 1992 (P >0.20, Fig. 4). Table 5. Estimated population size for mule deer based on deer counted on 2.59 km2 quadrats, Troublesome subunit, Middle Park, 1990-1992. Variances calculated using finite population correction factor and N = 64.1-2.59 km2 units. ·Quadrats Year Month 1990 1991 1992 January March January fQI!· n Mean 30 30 30 1848 1647 2361 ~.u:~Ii Lower 1251 751 1655 2~% ~I Upper 2445 2544 3067 CI ,Mean 32 54 30 12~~f l2~ml.~~ )em m1. 11.2 9.9 14.2 28.8 25.7 36.8 Yearly precision (CI as \ of mean) of quadrats was similar in magnitude (± 3054\) to precision of line transects although transects provided better precision depending on which model and truncation distance were selected (Tables 2, 5). The most and least precise estimates occurred for both �85 sampling systems in the same years, namely 1992 and 1991, respectively when group sizes of deer were smallest and largest for both systems (Tables 2, 4, 5). Both systems, therefore, performed poorest when deer were concentrated into fewer and larger groups (March, late winter) and when snow cover was not uniform. We expected quadrats to underestimate true population size because area based sampling strategies assume 100% of the animals present are counted which seldom occurs. In pinyon pine-Utah juniper habitats, observers detected only 66\ of the known deer on quadrats (Bartmann et ale 1986) while on the same area, line transects estimated 90\ of the known deer population (White et ale 1989). Based on these results, we expected quadrats to provide density estimates ~ 0.73 of estimates derived from transects assuming deer were as or more likely to be observed in the more open sagebrush vegetation. This expected ratio was achieved or exceeded only in 1992 (Table 6). We suggest that large discrepancies between methods in 1990 and 1991 caused by higher estimates from line transects may reflect observers initially less experienced with line transects than quadrats; alternatively, line transects when used in open habitats may be prone to "heaping" in the center line interval and therefore provide inflated estimates of density. Table 6. Ratios of mean population estimates for mule deer based on quadrats (NQ) and line transects (NT)' Troublesome subunit, Middle Park, 1990-1992. Year 1990 1991 1992 "Transect l~ransect 1848/7012 1647/3912 2361/2759 0.264 ~ 0.421 ~ 0.856 K 1848/4833 1647/4358 2361/3006 ~ 0.382 ~ 0.378 = 0.785 estimate based on truncation distance of 95 m and EXPL function. estimate based on truncation distance of 155 m and EXPL function. Counting time was 1.5x greater for line transects than for quadrats indicating that transects would be a more costly system to estimate densities of deer in our sample area (Table 7). OUr results contrasted with White et ale (1989) who found transects used only 0.40x the time used on quadrats. Differences in studies may reflect more time spent flying between quadrats by White et ale (1989) because they .ampled a amaller proportion of a much larger aample area. OUr increased cost of transects was somewhat offset by improved precision but this was highly dependent of which sighting model was used. When selecting the best of several EXPL or NEXP models each year, precision was 16-30\ better for line transects. However, precision for the recommended EXPL model truncated at 95 m was generally less than precision based on quadrats (Table 7). White at ale (1989) also found quadrats provided slightly better preciaion than tranaects. However, in their case, they concluded transects were more coat effective than quadrats primarily because transects provided less negatively biased estimates of density. Relative efficiencies of transects and quadrats are presumably affected by deer densities. At low densities, quadrats are likely less costly than transects because of anticipated difficulty in detecting adequate numbers of groups along transects to generate acceptable precision (White et ale 1989). At our densities of deer, we achieved the recommended 200 groups of deer per sample only in 1992 even with intense replication of transects; suggesting that densities of deer might be marginal for efficiently using transects (Table 2). We surmised that actual densities of deer in our sample area were considerably lower than densities in the sample area of White et ale (1989) �86 Table 7. Eatimated time devoted to counting deer on line tranaects and quadrat a flown in the Troubleaome subunit of Middle Park, 1990-1992. Flight Method/ Performance Ratios counting 1990 TIme 1991 (mins}" 1992 Average All Years 149 147 149 143 167 170 174 162 167 153 159 144 161 157 161 150 588 673 623 628 445 366 441 417 1.32 1.84 1.41 1.50 0.84 1.16 0.70 1.02 0.77 0.87 0.77 1.02 H~tb2g Tranaecta Transects Transects Transects Set Set Set Set Yearly 1 2 1 2 Rep Rep Rep Rep 1 1 2 2 Total Quadrats RatiQs - lransectsiQyadrats Time Precision Best EXPL or NEXP model" EXPL model @ 95 m "Tim8 shown is estimated flying tIme used to count animals include flying to/from for refueling during flights. "See Tables 2 and 5. and does not and thus were not surprised by the higher cost of transects. However, observed densities of deer on quadrats were comparable: 12.6 (White et al.1989) and 9.9-14.2 deer/km2 (this study). We therefore suggest that negative sighting bias on quadrats is greater in habitats of pinyon pinejuniper than in sagebrush. CORCLUSIORS Three assumptiona critical to achieving reliable estimates of animal density from line transects are: 1) groups of animals on the line will never be missed, 2) groups are fixed at the initial sighting position; they do not move before being detected and none are counted twice, and 3) distances to groups are measured without error (Burnham et al 1980). OUr spiked sighting distributions for elk and deer suggested that groups on the line were not commonly missed, although some probably were (White et ale 1989), but as importantly, groups may have been "heaped" into the center line interval because of errors in estimating distances. Detecting groups of elk and estimating distances to groups was not commonly affected by animal movements. However, deer were commonly moving, at least partly in response to being disturbed by the helicopter, which undoubtedly affected the detection of groups and complicated the estimation of perpendicular diatancea to locations where deer initiated movement. These movements of deer bordered on violating asaumption 2 (Burnham et ale 1980). We violated the third aaaumption because we only estimated distances to groups, although grouping distances into intervals somewhat ameliorates this problem. Errors near the center line were more likely influenced by helicopter orientation to the transect line after detecting a group than actual estimation of distances. Intervals of 10 m beyond the center interval may be stretching our ability to estimate distances with intervals of 20 m more plausible. However, Sighting models based on 20 m intervals did not fit data as well as with 10 m intervals. We therefore experienced problems in meeting the fundamental assumptions of line transect theory not unlike White et ale (1989). �87 We agree with White et ale (1989) that EXPL or NEXP models best fit the sighting distributions for elk and deer although selecting the best model is somewhat arbitrary. OUr concern is that -heaping- may be occurring in the center interval causing the lack of a desirable -shoulder- in the sighting distribution (Burnham et ale 1980). As a check on this problem, we recommend that groups in 2 distance intervals to the left of the center interval be tallied in future surveys to assess the likelihood of erroneously placing groups into the center interval. We agree with White et ale (1989) that 200 groups of animals will be needed to achieve acceptable precision; a task that may be overwhelmingly difficult for elk and marginally obtainable for deer existing at densities near 12/km2• Precision for estimates of deer density derived from quadrats was comparable to precision of line transects but quadrats were 30% less costly than transects at densities of deer common to our study. We undoubtedly violated at le.at the firat of the 2 primary aasumptions of quadrat sampling: 1) all deer on the sample quadrat were counted, and 2) deer were not counted twice. We therefore accept that our quadrat estimateB were negatively biased. However, we cannot at this time recommend abandoning quadrats in favor of line transects for estimating densities of deer in open sagebrush habitats until problems of assigning deer to the center interval of transects can be further assessed. LITERATURE CITED Bartmann, R. M., L. H. Carpenter, R. A. Garrott, and D. C. Bowden. 1986. Accuracy of helicopter counts of mule deer in pinyon-juniper woodland. Wildl. Soc. Bull. 14:356-363. Bear, G. D., G. C. White, L. H. Carpenter, R. B. Gill, and D. J. Essex. Evaluation of aerial mark-resighting estimates of elk popu1ationa. Wildl. Manage. 53:908-915. 1989. J. Buckland, S. T. 1985. Perpendicular distance models for line transect sampling. Biometrics 41:177-195. Burnham, K. P., D. R. Anderson, and J. L. Laake. 1980. Estimation of density from line transect sampling of biological populations. Wildl. Mono. 72. 202pp. Drummer, T. D., and L. L. McDonald. sampling. Biometrics 43:13-21. Freddy, D. J. 1991. Rep. July:59-69. 1987. Elk census methodology. Size bias in line transect Colo. Div. Wildl. Game Res. Gill, R. B. 1969. A quadrat count system for estimating game population. Colo. Game, Fish, & Parks Game Info. Leaflet 76. 2pp. Kufeld, R. C., J. H. Olterman, and D. C. Bowden. 1980. A helicopter quadrat census for mule deer on Uncompahgre Plateau, Colorado. J. Wildl. Manage. 44:632-639. LeResche, R. E., and R. A. Rausch. 1974. Accuracy and precision mooae censusing. J. Wildl. Manage. 38:175-182. Mendenhall, W., L. ott, and R. L. Scheaffer. 1971. Elementary aampling. Duxbury Press, Belmont, Calif. 247pp. of aerial survey Pollock, K. H., and W. L. Kendall. 1987. Visibility bias in aerial surveys: a review of estimation procedures. J. Wildl. Manage. 51:502-510. �88 Samuel, M. D., E. O. Garton, M. W. schlegel, and R. G. Carson. 1987. Visibility bias during aerial surveys of elk in northcentral Idaho. Wildl. Manaqe. 51:622-630. J. SAS Institute, Inc. 1985. SAS language guide for personal computers, version 6 edition. SAS Inst. Inc., Cary, N. C. 429pp. Unsworth, J. W., L.Kuck, and E. o. Garton. 1990. Elk sightability model validation at the National Bison Range, Montana. Wildl. Soc. Bull. 18:113-115. White, G. C., R. M. Bartmann, L. H. Carpenter, and R. A. Garrott. 1989. Evaluation of aerial line transects for estimating mule deer densities. J. Wildl. Manage. 53:625-635. Note, Segment objective #2 was not completed because further comparisons of line transects and 1-mi2 quadrats was later deemed more beneficial than assessing the effects of smaller quadrats on estimates of deer density. Segment objectives #3 and #4 were completed via a Federal Aid Enhancement Proposal for expanded funding to monitor survival rates of elk resulting in a new Federal Aid project for FY 1992-93. �������95 Colorado Division of Wildlife Wildlife Research Report July, 1992 JOB PROGRESS REPORT State of ~C~o~l~o~r~a~d~o~ _ Project No.__ ~W~-~ly5~3~-~R~-~5 _ Mammals Research Work Plan No _ Moose Investigations 1 Job No. Period Covered: Author: ~4~ Development of census methods and determination of movements, habitat selection, and degree of calf mortality of moose in North Central Colorado July 1, 1991 - June 30, 1992 R. C. Kufeld Personnel: D. Bowden, J. Bredehoft, T. Ffo1liott, B. Gill, S. Hoover, S. Kerr, B. Kraabel, S. Kozlowski, B. Lance, M. Miller, S. Porter, D. Reed, D. Rodriguez, G. Schoonveld, K. Snyder, T. Spraker, S. Steinert, D. Younkin, C. Wagner, J. Wenum, C. Wood. ABSTRACT Forty moose were captured, radio-collared and eartagged in North Park, Colorado, during December, 1991, January, 1992, and March, 1992. Five replicate counts of moose in the eastern and southern section of North Park (Game Management Unit 6 south of a line from Walden to the East Sand Dunes, Unit 171, and the portion of Unit 17 south of Rand and east of Arapahoe Ridge; an area of about 1,400 km2) were conducted by helicopter during January, February, and March, 1992. Relocation of radio-collared moose by fixed-wing aircraft during replicate helicopter counts showed that, due to deep snow and cold temperatures, nearly all radio-collared animals were in or adjacent to willow bottoms during the first 4 counts, and most were there during the 5th count. Thus, counts were limited to willow bottoms along drainages, and every drainage within the census area was covered. Moose were classified according to sex and age. During the first 4 counts a mean of 60.0% ± 3.9% (90% confidence) of groups with at least 1 radio-collared moose were seen. Differences in visibility rate between adult radio-collared bulls and cows were not significant (P ~ .10). The moose population in the census area was estimated at 441 ± 55 (90% confidence) animals based on the mean of the 5 counts adjusted for unseen moose. Mean sex and age ratios (90% confidence) for the 5 counts were 64 ± 10 bulls per 100 cows, and 68 ± 7 calves per 100 cows. Twenty two percent of cows with calves had twins. The largest number of moose occurred in Game Management Unit 171, with Unit 6 ranking 2nd, and Unit 17 ranking 3rd. The mean group encountered during the 5 counts contained 1.9 moose. Most moose were seen in groups of 2 animals, with single animals �96 ranking 2nd and groups of 3 ranking 3rd in frequency of occurrence. Relatively few groups were observed which contained more than 3 animals. The largest group contained 7 moose, but only 1 such group was observed. Future moose counts in North Park should include 100% coverage of all drainages during the period of deepest possible snow and coldest possible weather, because that is when the largest portion of the population will be in willow bottoms where they are most visible. The total count should be adjusted for unseen animals by mUltiplying the number of moose counted by 1.667; a factor derived by dividing 1 by 0.600, the mean percent of radio-collared moose seen on the first 4 counts. �97 DEVELOPMENT OF CENSUS METHODS AND DETERMINATION OF MOVEMENTS, HABITAT SELECTION, AND DEGREE OF CALF MORTALITY OF MOOSE IN NORTH CENTRAL COLORADO Roland C. Kufeld P. N. OBJECTIVES 1. To determine the proportion of moose actually observed when aerially counting moose in North Park. 2. To determine the extent of moose calf mortality in late winter. 3. To determine the degree of dispersal of young animals, and seasonal movements, home range size, and habitat selection of North Park moose. SEGMENT OBJECTIVES Same as P. N. objectives. INTRODUCTION Prior to 1978, moose (~~ shirasi) were rarely seen in Colorado. After much negotiation between Colorado Division of Wildlife, Federal land management agencies, and landowners, moose were transplanted into North Park, Colorado beginning in 1978. Twelve moose, obtained from Utah, were released into the Big Bottoms area of the Upper Illinois River drainage southeast of Rand, Colorado in 1978. Twelve more, obtained from Wyoming, were released into the same area in 1979. A third group of 12 moose, obtained from Wyoming, were released into the Upper Laramie River drainage in 1987 (Duvall and Schoonveld (1988). When moose reintroductions were being considered and negotiated landowners in the North Park area expressed concern that moose might compete for livestock forage and cause damage to fences and haystacks. To date that has not happened. However, landowners and Federal land management agencies agreed to the transplants contingent upon an evaluation by Division of Wildlife when the moose population approached 300 animals. A moose management plan was prepared by Colorado Division of Wildlife and approved by the Wildlife Commission on May 14, 1987 (Duvall and Porter, 1987). At that time the moose population in North Park was estimated at 108 animals. The management plan called for an evaluation of the status of the moose population and its relation to other land uses un North Park when the population reached 300 animals. The public was asked to comment on the plan. Subsequent letters from the U.S. Forest Service, U.S. Bureau of Land Management, U.S. Fish and Wildlife Service, Colorado Division of Parks and Recreation, Jackson County Extension Office, Jackson County Board of Commissioners, and the North park Stock Growers Association strongly support and request an evaluation when the moose population approaches 300. Moose counts and population simulations indicate that the moose population in North Park was nearing 300 animals in 1991. �98 Effective moose management is dependent upon an accurate estimate of moose numbers. Helicopter counts of moose have been conducted periodically in North Park, but moose are often difficult to see in the dense willows and coniferous timber types they inhabit. Thus, the proportion of moose present in the count area that are seen and counted is not known, and until that proportion is determined an accurate estimate of the moose population cannot be obtained. Experienced observers flying replicate fixed wing counts of a known number of moose in an enclosure in Alaska saw only 68 percent, on average, of the animals (LeResche and Rausch, 1974). An average of 73 percent of the moose present were observed on helicopter counts in southern Quebec (Crete et al., 1986). One objective of this study is to determine the proportion of observed and missed moose during aerial counts in the North Park habitat conplex. During 1990, a total of 149 moose were seen during aerial counts in North Park. Assuming this represents approximately 67 percent of the moose in North Park, a 1990 population of 222 animals was, projected for the Park. Moose counts in North Park showed cow/calf ratios averaging 66 calves per 100 cows in 1989 and 1990, and 61 calves per 100 cows in 1988. These are comparable to cow/calf ratios reported for Shiras moose in other states. Moose cow/calf ratios averaging 62 calves per 100 cows have been reported in scueheas eern Idaho (Ritchie, 1978), and ranging from 40 to 69 calves per 100 cows in the Gallatin River, Madison River, Centennial, Big Hole, and Gravelly-Snowcrest areas of southwestern Montana (Schladweiler, 1974). A moose population with cow/calf ratios, such as observed in North Park, should have grown to more than twice the 222 animals projected from 1990 counts during the years since moose were first introduced to Colorado. This raises the question of what is happening to the numerous calves that ate being seen during aerial counts? Is there a high rate of calf mortality in late winter after moose classification counts are conducted? Or, since this is still a growing population, are many of the calves, upon leaving their mothers, dispersing out of the area where counts are made and, therefore, no longer part of the counted population? If so, where and how far do they go? Such movement information can indicate whether the moose population is expanding into other areas, and can provide an indication of the rate and extent of that expansion. Knowledge of moose habitat selection is needed also, so that adequate measur.es can be implemented to provide for their habitat needs. STUDY AREA The study area includes game management units 6, 16, 17, and 171, in North Park, Colorado. North park encompasses of all of Jackson County. Elevations range from approximately 2400 to about 4000 meters. North Park is a large open valley almost completely ringed by high peaks. The park is approximately 72 km long by 64 km wide. The Park Range forms part of the northern, and all of the western boundary. The, southern boundary is the Arapahoe Range, and the Never Summer Range forms part of the eastern boundary. The summit of those ranges is the continental divide. The rest of the eastern boundary and part of the northern boundary is formed by the Medicine Bow Range. The central part of North Park consists of relatively flat terrain to rolling hills. North Park contains numerous streams and rivers which support extensive stands of willow (Salix spp.). In the higher elevations these willow bottoms are surrounded by forests of lodgepole pine (Pinus contorta), �99 aspen (Populus tremu10ides) or spruce (Picea spp.) and subalpine fir (~ lasiocarpa). The highest elevations are alpine tundra. Willow bottoms at lower elevations are bordered by large expanses of big sagebrush (Artemisia tridentata). Large tracts of land in North Park are administered by each of the following agencies: U.S. Forest Service, U.S. Bureau of Land Management, U.S. Fish and Wildlife Service, Colorado State Forest Service, and Colorado Division of Parks and Outdoor Recreation. Numerous large, privately owned, cattle ranches also occur in North Park. Principal industries include logging, cattle ranching, hunting, fishing, wildlife watching, and some coal mining and oil production. In addition to moose, other large ungulates include elk (Cervus elaphus), mule deer (Odocoileus hemionus), antelope (Anti10capra americana), and a few white-tailed deer (Ococoi1eus virginiana). METHODS AND MATERIALS Moose were captured throughout the eastern and southern part of North Park during December, 1991, and January, and March, 1992, by tranquilizing them with dart guns from the ground or from a helicopter. The immobilizing drug for adult animals was 2.7 mg (0.9 cc) of Carfentani1 and 40 mg (0.1 cc) of Xy1azine, and for calf moose 1.35 mg (0.45 cc) of Carfentani1 and 20 mg (0.05 cc) of Xylazine. This was loaded into a 1 cc dart. Tranquilized adult moose were reversed with 500 mg of Naloxone, of which 150 mg is administered interveniously (IV) and 350 mg administered subcutaneously (SQ). For calves, the Naloxone dosage was 300 mg of which 100 mg was administered IV and 200 mg SQ. Peneci11in at a dosage of 30 cc for adults and 20 cc for calves was given to tranquilized moose to counteract any adverse affects of immobilization. Sex and age composition of the darted animals was proportional to sex and age categories occurring in the population. Captured moose were fitted with 2 numbered eartags and a numbered radiocollar. Yellow eartags were used for moose captured on the east side of North Park and north of Highway 14, whereas orange eartags were used for moose captured on the east side of North Park but south of Highway 14. Each radio-collar was equipped with a mortality switch which will acitvate if the animal has not moved for 5 hours, suggesting that the animal has died. Instrumented moose were located at approximately 2-week intervals from January, 1992, through June, 1992, and plans call for such monitoring to continue for at least 3 years. Most locations were made by aerial telemetry using a Cessna 185 aircraft with a 2 element, "H" configuration receiving antenna mounted on each strut. A switchbox permitted the telemetry operator, a passenger in the aircraft, to operate antennas jointly or separately. Some locations were made by tracking on the ground until the animal was observed when the airplane was not available. Moose locations were plotted on USGS 1:50,000 scale maps and recorded by UTM coordinates. Vegetation type was also recorded for each moose location. Five counts of moose in the eastern and southern section of North Park were conducted by helicopter on the following dates in 1992: January 27-28, February 6-7, February 20-21, March, 5-6, and March 12-13. The count area included Game Management Unit 6 south of a line from Walden to the East Sand Dunes, Unit 171, and the portion of Unit 17 south of Rand and east of Arapahoe Ridge; an area of about 1,400 km2• Every drainage within that area was �100 covered. All radio-collared moose seen were recorded by the number on the collar. Moose were classified according to sex and age. Sex determination was based on presence or absence of antlers on adult animals during earlier counts before antler drop, and on presence or absence of the vulva patch (Mitchell, 1970) on adults after antler drop and on calves. Vegetation type was recorded for each group of moose observed. A Bell Jet Ranger helicopter with the pilot and 3 observers was used. Each count required approximately 10 hours of flying time. During the period when the 5 helicopter counts were made 4 fixed wing flights were also made to locate all of the radio-collared moose. Thus, the number of radio-collared moose available to be seen in the census area while counts were made was known, and the proportion of radiocollared moose seen by observers in the helicopter provided an estimate of the accuracy of their count. RESULTS Moose capturing and monitoring Forty-one moose were marked in North Park during December, 1991, January, 1992, and March, 1992 (Table 1). Forty received numbered eartags and a radiocollar. One, a bull calf, received only eartags. The radio-collared animals included 8 adult bulls, 13 adult cows, 10 bull calves, and 9 cow calves. Analysis of data for movements, home range size, and habitat use will be presented in a future report when periodic monitoring of moose is completed. Moose counts Snow cover was 100% on all counts except that on the 4th and 5th counts some patches of bare ground appeared in the Michigan River bottom downstream from the 3 Rivers Ranch Headquarters, and in the Illinois River bottom from Walden to Rand. Above those elevations snow was generally knee to belly deep on a moose. During the 4th and 5th counts much of the lower sagebrush covered part of North Park was clear of snow. However, moose were not present in that area. During the period the 5 replicate helicopter counts were made 37 radiocollared moose were available. All ~ere in the count area during each of the 5 counts. Sightability was based on the percent of groups, which contained at least 1 radio-collared animal, that were actually observed. These percentages were used to project moose count data to estimate the population allowing for animals that were missed during counts by helicopter. Our 37 radio-collared moose consisted of 8 radio-collared cow/calf pairs and 1 radio-collared pair of twin calves. They were always together during count I, so on count 1 there were 28 radio-collared groups of moose available to be seen by the observers in the helicopter. If a marked pair separated they were considered to represent 2 separate groups. During counts 2, 3, 4, and 5, one of these pairs had split up, so the number of available groups of moose with at least 1 radio-collared animal was increased to 29. During the first 4 counts 60.0% ± 3.9% (90% confidence) of groups with at least 1 radio-collared moose were seen (Table 2). During the 5th count fewer total animals and fewer radio-collared �Table 1. Moose radio-collared and eartagged In North Park during Decenber, 1991 and January, 1992, and March, 1992. Collar 110. No. lIone 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 •, I, gartas Color 42 221 222 226 236 237 225 228 239 238 7 &8 240 29 9 41 3 &4 19 31 28 13 33 229 5 &6 30 12 27 14 21 40 18 43 227 223 224 26 36 32 251 35 34 50 a, x, *, +, S, Org. Org. Org. Org. Org. Org. Org. Org. Org. Org. Yel. Org. Org. Yel. Org. Yel. Yel. Org. Org. Yel. Org. Org. Yel. Org. Yel. Org. Yel. Yel. Org. Yel. Org. Org. Org. Org. Org. Org. Org. Org. Org. Org. Org. Sex M M F F F F F F F F F F M M M F M M M M F F F F F F F F F F F M M M M M M M M M M Age Date Captured Capture Location Calf 2.5 4.5+ 5.5+ 3.5+ 3.5+ 1.5 3.5+ 3.5+ 1.5 5.5+ 6.5+ 3.5+ 3.5+ 3.5+ 3.5+ 2.5 1.5 3.5+ 3.5+ 3.5+ Calf Calf Calf cilf Calf 2.5+ Calf Calf Calf Calf Calf Calf Calf Calf Calf Calf Calf Calf Calf Calf 03-12-92 12-03-91 12-03-91 12-03-91 12-03-91 12-03-91 12-04-91 12-04-91 12-04-91 12-04-91 12-06-91 12-04-91 12-12-91 12-09-91 03-12-92 12-05-91 12-12-91 12-09-91 12-11-91 01-14-92 12-10-91 12-04-91 12-05-91 12-13-91 01-14-92 12-11-91 01-15-92 01-15-92 03-12-92 12- 12-91 03-12-92 12-03-91 12-03-91 12-03-91 12-10-91 01-15-92 12-10-91 12-06-91 01-15-92 01-15-92 01-15-92 Illinois R. 4.5 ~I. S. of Rand Snyder Creek Old Homestead Lodge Old Homestead Lodge Illinois R. 1 MI. E. of Jack Cr. Rd. Jct. Jack Cr. Rd. Between Whiskey Cr. & Parkvlew Cr. Illinois R. at Whiskey Cr. Willow Cr. Rd. 0.5 mi. W. of Spring Cr. Willow Cr. Rd. 0.75 MI. V. of East Branch S. Fork Canadian R. 1 ~I. N. of Jewell Rd. Illinois R. at Nat. Forest Bdry. 0.5 mi. NV of Aspen Campground S. Fort Canadian R. 1.5 mi. II. of Jewell Rd. Willow Cr. 0.5 mi. S. of Hwy. 125 crossing N. Fork Canadian R. 0.4 MI. V. of II.F. Yurt S. Fork Canadian R. 1 mi. II. of Jewell Rd. Hwy. 125 at Old Homestead Lodge Jack Cr. 1 mi. W. of Nat. Forest Bdry. S. Fork Canadian R. 0.2 mi. E. of Rd. Clos. Willow Cr. 6 mi. W. of Hwy 125 Illinois R. at Whiskey Cr. N •.Fork Canadian R. 0.4 mi. V. of II.F. Yurt Jack Cr. 0.25 mi. W. of Jack Cr. Ranch State For. Turnoff to Michigan L. Campgrnd. Jack Cr. 200 yds W. of Illinois R. Rd. Jct. State For. 1 mi. E. of Turnoff to Park Hq. 3 Rivers Ranch, Michigan R. Illinois R. at Deer Cr. S. Fork Canadian R. 1 mi. II. of Jewell Rd. Illinois R. 2 MI. S. of Deer Cr. Old Homestead Lodge Illinois R. 1 mi. E. of Jack Cr. Rd. Jct. Jack Cr. Rd. Jack Cr. 0.25 mi. V. of Jack Cr. Ranch Ballar Ranch, Michigan R. Willow Cr. 6 mi. W. of, Hwy 125 FS Bdry 0.25 mi. NW of Aspen Campground Ballar Ranch, Michigan R. 0.5 mi. V. of Gould Ballar Ranch, Michigan R. Locat Ion UTM ~oorm E-W N-S 397.75 407.72 403.10 403.10 410.16 475.55 407.40 407.33 400.35 403.15 415.68 408.35 412.13 415.25 397.60 414.40 415.68 403.10 407.03 413.82 398.90 407.33 414.40 410.95 413.20 409.18 412.95 406.45 396.15 415.68 395.25 403.10 410.16 475.55 410.80 408.80 398.90 412.20 408.95 412.20 411.40 485.35 469.7'9 472.20 • 472.20 474.00 , 474.58 a 472.00 473.94 x 471.50 471.90 493.00 * 474.85 485.88 493.30 483.40 496.04 + 493.00 476.85 476.15 496.15 470.30 S 473.94 x 496.04 + 475.00 488.88 & 474.50 488.95 & 497.80 488.85 493.00 * 491.95 477.20 • 474.00 , 474.58 a 474.95 494.75 xx 470.30 S 485.78 494.55 xx 486.14 491.05 &, xx: A pair of these symbols Indicates another - offspring relationship • I-' 0 I-' �102 Table 2. Moose population estimate from 5 helicopter moose counts in North Park, January 27 - March 13, 1992. The estimate was adjusted to allow for unseen moose based on the percentage of groups of moose seen which contained at least 1 radio-collared moose. Count Total moose seen 1 2 3 4 5 280 293 266 240 193 Groups with at least 1 radio-collared moose1 Avail. No. seen % seen Bulls / 100 cows Calves / 100 cows Pop. est.2 60.7 55.2 62.1 62.1 48.3 80 53 59 65 64 II 461 530 428 386 399 60.03 64 68 441 3.9 10 7 55 Lower 90% confidence limit 56.1 54 61 386 Upper 90% confidence limit 63.9 74 75 496 28 29 29 29 29 x t :t: .10 SE 17 16 18 18 14 77 65 58 65 1 Total radio-collared moose - 37 including 9 radio-collared cow-calf or sibling pairs. All pairs were together on count 1. On counts 2 - 5 1 pair had separated. When a pair separated they were considered to be 2 separate groups. 2 Population estimate - Total Moose counted x 3 This is the mean of counts 1-4 during which a higher percentage of moose were in or near willow bottoms where they were more visible. On the 5th count fewer moose and fewer radio-collared moose were observed. 1 % of groups seen with at least 1 radiocollared animal. moose were seen because they appeared to have moved into the timber where they were less visible, perhaps due to relative lateness of the season. Visibility rate, during the 5 counts, of groups containing at least 1 adult (1.5+ years old) bull was not different (P ~ .10) from visibility rate of groups containing at least 1 adult cow. Visibility rate for both was 54%. Since calves were almost always with their mothers during counts their visibility rate was tied to that of adult cows. A moose population estimate based on each of the 5 counts was computed using a formula which adjusts for unseen animals, shown in footnote 2 of Table 2. Apparent movement of some radio-collared moose into the timber where they were less visible during the 5th count had no effect on validity of the population estimate for that count, because although fewer radio-collared animals were �103 seen the total number of moose seen was also lower. The result is an estimated mean population of 441 ± 55 (90 percent confidence). Based on the 5 counts the population was estimated within 12% of the mean with 90% confidence. Mean sex and age ratios (90% confidence) for the 5 counts were 64 ± 10 bulls per 100 cows, and 68 ± 7 calves per 100 cows (Table 2). Estimating precision for cow/calf ratios was better than for bull/cow ratios. Based on 5 counts the mean cow/calf ratio was estimated within 10% of the mean, with 90% confidence, compared to 16% of the mean for bull/cow ratios. Twenty two percent of cows with calves had twins. The largest number of moose occurred in Game Management Unit 171, with Unit 6 ranking 2nd, and Unit 17 ranking 3rd (Table 3). The mean group encountered during the 5 counts contained 1.9 moose. Most moose were seen in groups of 2 animals, with single animals ranking 2nd and groups of 3 ranking 3rd in frequency of occurrence (Fig. 1). Relatively few groups were observed which contained more than 3 animals. The largest group contained 7 moose, but only 1 such group was observed. Adult bulls were largely segregated from cows and calves, in that 82% of adult bulls seen were either alone or with other adult bulls. Most bulls (42%) were alone, 31% were in a group of 2 bulls, 3% were in a group of 3 bulls, and 6% were in a group of 5 bulls. The other 18% were with cows or cows and calves. Forty nine percent of adult cows were in a group consisting of 1 cow and 1 or 2 calves. Seventeen percent of adult cows were solitary, 9% were in a group of 2 adult cows, 12% were in a group of 2+ cows and calves, and 13% were in a group with another cow or calves and 1 or more adult bulls. Table 3. Estimated 1992 winter moose population by Game Management Unit in Unit 6 south of a line from Walden to the east sand dunes, in Unit 171, and in Unit 17 south of Rand and east of Arapahoe Ridge1 Game Management Unit 171 17 Sex and Age 6 Bull, Adults 33 67 22 122 Cow, Adults 51 105 34 190 Bull, Calves 17 34 12 63 Cow, Calves 17 n 12 66 Total Moose 118 243 80 441 1 Total moose (441) is the mean counts made during the period adult bulls, adult cows, bull on the percentage of each sex Management Unit. Total population estimate for the 5 helicopter January 27, - March 13, 1992. Numbers of calves, and cow calves were computed based and age class counted in each Game �104 50 + . ................................................................................................................................. 40 ............................................................................................... " ............................................................................................... " · " ·..·,,· ........................................................................................................... .................... 20 , " ·· · ·..·1 " " , , " ................................................................................................... ·..·..1 ...... 10 . " o 2 3 4 5 6 Number of Moose in Group Fig. 1. Mean number and percent of groups of moose in each size category during 5 helicopter counts of moose in North Park during January through March. 1992. 7 �105 DISCUSSION In most of the North American moose range animals are widely distributed over vast areas. Rivest and Crepeau (1990) have proposed a moose census system for large areas involving 2 phases. First a stratified random sample of parcels is searched by fixed-wing airplane. Then, a subsample of the first-phase sample is selected with probabilities proportional to the size of the track newtorks in the parcel. These parcels are searched by helicopter, and regression of the helicopter count of moose on variables measured during the airplane search is used to estimate the size of the moose population. Aerial counting systems involving helicopter or fixed wing aircraft, which employ quadrat, transect or large plot sampling systems, have been used in large areas throughout North America to inventory moose (Evans et al. 1966, Bergerud and Manuel 1969, Hauge and Keith 1981, Novak, 1981, Karns 1982, Larsen 1982, Gasaway and Dubois 1987, Elliott 1988). In Colorado, randomized helicopter quadrat systems (Kufeld et al. 1980) or aerial line transect systems (Burnham et al. 1980, White et al. 1989) are used to census mule deer in large areas. Moose habitat in North Park is characterized by large expanses of coniferous forest separated by willow bottoms that occur along drainages. Deep snow and very cold temperatures during winter tend to concentrate moose in and adjacent to these willow bottoms. Fortunately the willow type is where they are most visible and can be most easily counted. When moose are scattered throughout the coniferous forests they are very difficult to find and see, however, it is not necessary to attempt to count moose there if the snow is deep enough and temperatures are cold enough to concentrate them in the willow bottoms. Relocation of radio-collared moose by fixed wing aircraft during replicate helicopter counts showed that nearly all radio-collared animals were in or adjacent to willow bottoms during the first 4 counts and most were there during the 5th count. Since the willow bottoms are long and narrow in shape (mostly less than 0.4 km wide) the census area is not suited for implementation of a randomized helicopter quadrat census, or an aerial line transect system. Thus, it is deemed more efficient to census the entire area of willow bottom habitat. Since moose frequently rest under conifers adjacent to the willow bottoms a narrow band of coniferous forest along the edges of the willows was also included in the search area. Beasom et al. (1986) reported accuracy of aerial counts for white-tailed deer was unaffected by sample intensity, but precision increased with percent of the area sampled. Since we sampled the entire area of interest, findings of Beasom et al. (1986) suggest the only way to increase our preCision is with replicate counts. The number of replicate counts is subject to economic constraints. Sightability of animals during aerial counts is influenced by factors such as speed, height above ground, transect width, obsservers, group size and vegetation cover (Caughley et al. 1976, Samuel 1981, Barnes et al. 1986, Samuel et al. 1987). Aerial counts for moose should be designed with these factors in mind. For example, the helicopter should fly at an altitude and speed which gives observers maximum visibility and time to classify animals, yet provides for adequate safety under existing terrain and wind conditions. Counts should be made during December or January. Snow must be deep enough and temperatures cold enough to force moose to concentrate in willow bottoms at lower elevations. These snow depths will be near the maximum, and �106 temperatures will be near the minimum usually encountered for the area during and average winter. Gasaway et al. (1985) reported that sightability of moose generally increased with group size. We did not measure the effects of group size on sightability. However, since the mean group size encountered during our 5 counts was only 1.9 moose the population estimating error caused by not adjusting data for visibility bias toward larger groups should be minimal compared to visibility bias in an elk population, for example, (Samuel et al. 1987) where animals may occur singly or in very large groups. A much more important factor affecting visibility bias for moose, in North Park, than group size is whether the moose are concentrated in willow bottoms or scattered in the coniferous forest. RECOMMENDATIONS Because of budget constraints future moose counts for population estimation in North Park will likely involve only 1 count per winter. On future moose counts in North Park I recommend adjusting the total count to allow for unseen animals by mUltiplying the total number of moose counted by 1.667. This correction factor is derived as follows: 1 / 0.600. The value 0.600 is used because a mean of 60.0% ± 3.9% of groups of moose with at least 1 radiocollared animal were seen during the first 4 helicopter counts. I also recommend that future moose counts be conducted along all drainages during the period of deepest possible snow and coldest possible weather, because that is when the largest portion of the moose population will be in the willow bottoms where they are most visible. LITERATURE CITED Barnes, A. G., G. T. E. Hill, and G. R. Wilson. 1986. Correcting for incomplete sightability in aerial surveys of Kangaroos. Aust. Wild. Res. 13:339-348. Beason, S. L., F. G. Leon III, and D. R. Synatzske. 1986. Accuracy and precision of counting white-tailed deer with helicopters at different sampling intensities. Wildl. Soc. bull. 14:364-368. Bergerud, A. T., and F. Manuel. 1969. Aerial census of moose in central Newfoundland. J. Wildl. Manage. 33:910-916. Burnham, K. P., D. R. Anderson, J. L. Laake. 1980. Estimation of density from line transect sampling of biological populations. Wildl. Monogr. 72. 202pp. Caughley, G., R. Sinclair, and D. Scott-Kemmis. survey. J. Wildl. Manage. 40:290-300. 1976. Experiments in aerial Crete, M., L. P. Rivest, H. Jolicoeur, J. M. Brassard, and F. Messier. 1986. Predicting and correcting helicopter counts of moose with observations made from fixed-wing aircraft in southern Quebec. J. Appl. Ecol. 23:751-761. �107 Duvall, A. C., and S. H. Porter. 1987. Moose management plan, North Park Data Analysis Unit (M-l). Colo. Div. Wildl. Unpubl. Rep. 10pp. Duvall, A. C., and G. S. Schoonveld. 1988. and Management. Alces 24:188-194. Elliott, D. C. M. 24:48-55. 1988. Colorado Moose: Reintroduction Large area moose census in northern Manitoba. Alces Evans, C. D., W. A. Troyer, and C. J. Lensink. 1966. Aerial census of moose by quadrat sampling units. J. Wildl. Manage, 30:767-776. Gasaway W. C., S. D. Dubois, and S. J. Harbo. 1985. transect surveys for moose during May and June. 49:777-784. Biases in aerial J. Wildl. Manage. Gasaway, W. C., and S. D. Dubois. 1987. Estimating moose population parameters. Swed. Wildl. Res. Viltrevy Suppl. 1. p. 603-617. Hauge, T. M., and L. L. Keith. 1981. Dynamics of moose populations in northeastern Alberta. J. Wildl. Manage. 45:573-597. Karns, P. D. 1982. Twenty-plus years of aerial moose census in Minnesota. Alces 18:186-196. Larsen, D. G. 167. 1982. Moose inventory in the southwest Yukon. Alces 18:142- Mitchell, H. B. 1970. Rapid aerial sexing of antlerless moose in British Columbia. J. Wildl. Manage. 34:645-648. Novak, M. 1981. The value of aerial inventories in managing moose populations. Alces 17:282-315. Kufeld, R. C., J. H. Olterman, and D. C. Bowden. 1980. A helicopter quadrat census for mule deer on Uncompahgre Plateau, Colorado J. Wildl. Manage. 44:632-639. LeResche, R. E., and R. A. Rausch. 1974. Accuracy and precision of aerial moose censusing. J. Wildl. Manage. 38:175-182. Ritchie, R. W. 1978. Ecology of moose in Fremont County, Idaho. Wildl. Bull. No.7. 33pp Idaho Rivest, L. P. and H. Crepeau. 1990. A two-phase sampling plan for the estimation of the size of a moose population. Biometrics 46:163-176. Samuel, M. D. 1981. Correction of visibility bias in aerial surveys where animals occur in groups. J. Wildl. Manage. 45:993-997. Samuel, M. D., E. O. Garton, M. W. Schlegel, and R. G. Carson. 1987. Visibility bias during aerial surveys of elk in northcentral Idaho. Wildl. Manage. 51:622-630. J. �108 Schladweiler, P. 1974. Ecology of Shiras moose in Montana. Montana Dept. of Fish and Game Final Rep. Proj. W-98-R and W-120-R. lOOpp. White, G. C., R. M. Bartmann, L. H. Carpenter, and R. A. Garrott. Evaluation of aerial line transects for estimating mule deer densities. J. Wildl. Manage. 53:625-635. ~<t/C~'-L~~'~~~~~~~~ __ Prepared by __ ~ . ~{/'~~. Roland C. Kufeld Wildlife Researcher C 1989. �Colorado Division of Wildlife Wildlife Research Report July 1992 JOB PROGRESS REPORT State of Colorado Proj ect No. _W..,__=-15::..3 ••..-.••R~..• 3=-- _ Mammals Research 1A Multispecies Investigations Job No. 1 Animal and Pen Support Facilities for Mammals Research Period Covered: July 1, 1991 - June 30, 1992. Work Plan Authors: Personnel: No. M. A. Wild, M. W. Miller, B. J. Maynard, and D. R. Magnuson. D. L. Baker, P. Bleicher, A. L. Case, C. Fowler, R. B. Gill, D. Hill, B. J. Kraabel, M. J. McArtor, C. McCarty, J. K. Ringelman, T. R. Ritchie, K. W. Scott, R. B. Snyder, A. N. Torres, and K. D. Williams. ABSTRACT The Colorado Division of Wildlife's Foothills Wildlife Research Facility (FWRF) maintained captive animals (up to 104 wild and domestic ungulates of 6 species and 108 migratory and upland game birds of 6 species) and facilities supporting 10 different mammalian and avian research projects. Routine animal care and facility maintenance programs were conducted as previously described. However, we emphasized a quality and conservationoriented approach in this work by striving to increase efficiency and longterm benefit from the programs and projects undertaken. To alleviate high animal density in some enclosures, we constructed 6 new animal pastures and adopted a policy that places an upper limit on the population size that FWRF will support. A maximum of 88 research animals representing elk, bighorn sheep, mule deer; pronghorn antelope. and a" mountain goat will be maintained; animal numbers greater "than these levels will require funding from specific projects. Routine feeding and care of research animals, including health observation, training, weighing and clean-up, was performed by work-study and summer temporary employees. Communication between caretakers and managers/veterinarians and adequacy of feeding protocols was facilitated through several newly adopted or revised data and recordkeeping forms for animal observation, feeding instructions and feed consumption. Non-pregnant adult elk maintained on a primary diet of hay cubes gained on average 16.5 kg over a 6 month period. Hay cubes are more cost effective to feed to el~ based on less waste and reduced time for pen cleaning as opposed to longstem hay. No significant differences (P~.05) were observed in apparent digestibility or quantity of intake of hay cubes between pregnant and �110 non-pregnant cows in their last trimester of pregnancy. Vaccination of pronghorn with Bacteroides and bighorn with Pasteurella toxoid did not protect against the specified infection. The most common cause of mortality at FWRF this year was trauma. We recruited 25 animals into our captive population in 1991. Most of these were hand-raised at FWRF using evaporated milk fed ad libitum. We attempted to limit-fed milk to pronghorn fawns >42 days of age to stimulate dry feed intake. but we failed to adequately restrict intake to levels markedly lower than those of the ad libitum group. Body weights between the two groups were similar. �III ANIMAL AND SUPPORT FACILITIES MAMMALS RESEARCH FOR Margaret A. Wild et a1. P. N. OBJECTIVES 1. To provide and maintain captive wildlife populations and facilities supporting CDOW's Terrestrial Wildlife Research Program, as well as programs of CDOW cooperators. 2. To develop improved animal and facility management practices that will provide maximum research opportunities at minimum cost. 3. To enhance facilities to serve a growing diversity of anticipated research needs. SEGMENT OBJECTIVES 1. Maintain, improve and expand animal research and holding facilities. 2. Coordinate all rearing, training, maintenance, and research activities. 3. Maintain up to 20 elk, 35 mountain sheep, I mountain goat, 30 pronghorn antelope, 50 mule deer, 1 domestic cow, 80 ducks and 20 upland birds in suitable health to perform required research experiments. 4. Conduct management experiments to increase efficiency aridefficacy of feeding and maintenance activities related to research facility operations. 5. Establish a conservation-oriented approach for providing utilities and services to operate research facilities. 6. Establish a standard program for evaluating and documenting health status of captive wildlife. METHODS AND MATERIALS Routine animal care and facility maintenance programs supporting new and ongoing Terrestrial Wildlife Research Program projects were conducted as previously described. We emphasized a quality and conservation-oriented approach in this work by striving to increase efficiency and longterm benefit from the programs and projects undertaken. Specifically, we performed the following tasks: �112 ANIMAL MAINTENANCE General: Again this year, routine feeding and care of research animals, including health observations, training, weighing and clean-up, was performed wholly by well trained work-study and temporary employees. To aid in alleviating somewhat overcrowded conditions, and to house increased numbers of mule deer for the Fertility Control project, we constructed 6 new animal paddocks at FWRF using high tensile New Zealand electric fencing (Fig 1). Three of the pastures will be used to house mule deer, 2 for elk and 1 for bighorn sheep. Also to avoid overcrowding and to stabilize the cost associated with maintaining FWRF, we (RFAC) adopted a new policy concerning herd size limits for the future. Bighorn sheep lambs and elk calves born to captive females at FWRF in 1992 are being dam-raised. We are collecting samples from bighorn lambs as part of the ongoing Pasteurella monitoring studies (See WP2, J4 for details) . NUTRITIONAL MAINTENANCE Feeding protocols: Pronghorn, mule deer, and bighorn sheep continued on a maintenance diet of long stem hay. Supplementation with high energy deer/elk wafers (Baker and Hobbs 1985) was limited again this year to combat adverse health effects associated with oversupplementation (e.g. fatty liver, reproductive difficulties and possibly lumpy jaw). Fine adjustments were made to more accurately define minimal levels of supplementation required to maintain desirable body condition. In addition, to provide feed in optimal quantities with minimal waste, we implemented a feed record system that asks feeders to estimate daily feed remaining and amount added per pen. This allows the Facility Manager to make timely changes in feeding instructions concerning feed amounts to be offered to each pen in the future. We evaluated the utility of feeding elk a diet of hay cubes supplemented with grass hay and high energy deer/elk wafers. Feed amounts were based on calculated energy requirements for various age, sex and reproductive classes. Elk were slowly acclimated to hay cubes starting about 1 January and by 27 January were on cubes as their primary ration. Effects of late pre~ancv on intake and apparent digestibility of hay cubes by elk: We compared voluntary intake and apparent digestibility of hay cubes by non-pregnant (N - 5) and pregnant (N - 3) cow elk in their last trimester. Quantity of hay cubes consumed ad libitum were determined for individual non-pregnant and pregnant elk for a 12 day study period and the daily mean for each group calculated. Apparent digestibility of hay cubes was determined by collecting fecal output of the elk while housed for 7 days in metabolic cages. Feed and feces were analyzed for crude protein (Kjeldahl N x 6.25), ash and 100% dry matter by procedures described by A.O.A.C. Neutral detergent fiber (NDF) , acid detergent fiber (ADF) and acid detergent lignin (ADL) were determined using sequential analysis without sodium sulfite (Goering and Van Soest 1970, Van Soest and Robertson 1980). Statistical analyses were performed using a student's t-test. �113 A N • • J • I 11 . I I B A ! ,-- __ I ! .L..-_....:..._--,, Fig. 1. Diagram of Foothills Wildlife Research Facility's new animal paddocks labeled A-G. �114 Bottle-raising neonates: We bottle raised bighorn sheep, mule deer, elk and pronghorn in 1991 for ongoing research projects and mule deer and pronghorn in 1992 for use in the Fertility Control studies (See WP2, J16). All bottle-raised neonates were fed according to Wild and Miller (1991), except that we attempted to limit fed milk to older pronghorn fawns in 1991. Eleven pronghorn fawns, six females and five males, were fed milk ad libitum through 42 days of age, then divided into two groups, those receiving limited and ad libitum quantities of milk. An attempt was made to pair individuals: four sets of twins were separated so that one twin of each set was in each group; the remaining three animals were randomly assigned to groups. The group receiving ad libitum quantities continued to receive unlimited evaporated milk through September 3 (age range 69-91 days). At this time, the group was restricted to 720 m1 of milk per day to begin the weaning process. Further milk restrictions were as follows: September 8 -- 480 m1 per day; September 15 -- 240 m1 per day; September 25 -- weaned (age range 91-113 days). The group receiving limited quantities of milk was weaned more gradually. At age 43 days, each of these fawns was restricted to the daily average of its milk intake for the previous week. Fawns whose previous week's average daily intake fell below 780 m1 were assigned a limit of 780 m1. The averages were rounded up for easier measuring of milk volume in baby bottles (range 21-44 m1). On August 25 (age range 63-80 days), daily intakes over 800 m1 were reduced to 800 m1. On August 28, intakes were reduced to 720 m1. On September 8, the limit-fed group began following the same intake restriction schedule as the ad libitum group, and was weaned on September 25 (age range 91-111 days). HEALTH MAINTENANCE General: We continued to use and revise the Daily Animal Observation Form to facilitate communication of health problems from caretakers to veterinarians, and to provide a written health history for each animal. ProfessionaA\AUA]AU* nary care was provided by CDOW employees or a wildlife veterinarian in private practice. Due to confirmation of bovine tuberculosis in a game ranched elk herd in Colorado and the associated question of TB status of all captive elk in the western United States and Canada, we TB tested all individuals in our elk population in February 1992. The single cervical test was performed using the USDA-APHIS protocol for cervids. We evaluated a Pasteurella haemolytica subcomponent vaccine (Presponse-) in bighorn 'lambs in 1991. Lambs were vaccinated at 7-10 days of age and monitored through the summer. Three elk cows and two mule deer does were inadvertently exposed to males of their species during the breeding season. To limit growth of our captive populations, we attempted to terminate the pregnancies. Elk and mule deer each received 5 mg Luta1yse- (dinoprost tromethamine, PGF2 �115 alpha) intramuscularly (1M) as an abortifacient. early January and mule deer in mid-May. Elk were treated in Methods for treating andlor preventing lumpy jaw in captive pronghorn: We completed data collections on pronghorn involved in the project studying methods for treating and/or preventing lumpy jaw. Treatment animals were given boosters of Footvax* in September 1991. Blood was also collected from treatment and control animals at this time and about 8 months later. Titers to Bacteroides nodosus are to be determined as previously described (Miller 1991). Clinical condition of treatment and control animals were also evaluated at this time. FACILITY MAINTENANCEjREPAIRS/IMPROVEMENTS A variety of scheduled and unscheduled maintenance and repair activities were necessary to support facility operation and ongoing research programs. We worked toward a conservation-oriented approach for facility care by undertaking preventive maintenance projects, and performing highquality new construction and repairs to existing facilities. In addition to our "project board" system that identifies and prioritizes repair and maintenance tasks at FWRF, we implemented a Weekly Work Schedule that assigned tasks to qualified individuals with or without the help of assistants. We continued to modify and improve the system throughout the year. RESEARCH PROJECTS Facility operations offered support for pilot studies, student special studies, and CDOW and cooperative research experiments that were initiated, conducted, or continued using FYRF animals and facilities throughout the year. RESULTS AND DISCUSSION ANIMAL MAINTENANCE Detailed feeding instructions, feed consumption forms and health records facilitated communication between caretakers and managers/veterinarians, thus optimizing overall quality of animal care. New animal paddocks added needed space for research animals at FWRF. Mule deer and bighorn sheep were moved into 2 of the 6 new animal paddocks. Moving the mule deer in turn created more space for pronghorn in the east side pastures. We anticipate filling the remaining new paddocks with animals as soon as shelters and gates are completed early in FY 1993. No net gain in animal numbers should occur in the near future according to a new RFAC policy. The policy places the upper limits of herd size at current levels (Table 1) unless a specific research project requires greater numbers. In this case, the project would be responsible for financial support of animals over the baseline level. We also �116 Table 1. Upper population limits of animals supported by FWRF. Species Elk Bighorn sheep Rocky Mtn. goat Mule deer Pronghorn 20 29 1 13 Z2 Total 88 recommended that a core group of individuals, composed of 12 adult females and 1 adult male, of all major species (elk, bighorn sheep, mule deer and pronghorn) be.maintained at all times at FWRF. These popUlations could provide an adequate number of animals to encourage and facilitate many research projects, or if more animals were required, could serve as a source to increase the population size through reproduction. Current population sizes used to formulate upper limits included neonates born thus far in 1992 to captive females. Lambing dates in bighorn were later than recent years and ranged from 6 May to 14 June (mean 27 May). Eight bighorn lambs were born, but the 2 lambs with birth weighs ~2.5 kg died within 24 hr. All surviving lambs weighed 4-5 kg when <24 hr old (mean 4.2 kg). Two elk calves were born, one on 27 May weighing 16.0 kg and the other on 7 June weighing 20.0 kg. NUTRITIONAL MAINTENANCE Feeding protocols: Use of feed records that required caretakers to estimated daily feed remaining and feed added to each pen markedly increased the timeliness of changes made to feeding instructions. This was especially true in the spring and summer months when availability of green forage in the pastures varied greatly. When green forage was available, supplementation with hay and pelleted ration was decreased to avoid feed waste. Supplement was again increased as green forage availability declined. We initially fed all species at levels of supplementation based on average daily metabolic rates as calculated by Baker (Miller 1990). This feeding protocol appeared to be adequate for bighorn sheep and elk, but may be insufficient for maintaining optimal weight in our captive pronghorn and mule deer. We have increased supplement offered to adult pronghorn does and castrates to more than 5 times that calculated (about 1250 g/head/day vs. calculated 211 g/head/day) in order to maintain body weights at good levels. This discrepancy may be due to insufficient understanding of pronghorns' nutritional requirements or may be due to habituation by the pronghorn to a pelleted diet and unwillingness to change over to hay as their primary feed. �117 Non-pregnant adult elk cows fed hay cubes at 5500 g/head/day and pelleted supplement at 500 g/head/day gained on average 16.5 kg over the 6 month period. Their mean weight in June 1992 was 305 kg (SE - 5.6); over 30 kg per animal greater than June 1991 weights (mean 274.6, SE - 8.7). Although hay cubes are consumed in greater amounts than long stem hay (as percent of body weight, see below), we realized cost savings by feeding hay cubes rather than long stem hay to elk. On a per ton basis, hay cubes and long stem hay are similar in price ($117 and $120, respectively). On a weight basis, we offered about 30% more long stem hay than our current quantity of hay cubes because waste is markedly greater for long stem hay than for hay cubes. Additionally, personnel time spent cleaning elk feed areas has been reduced from about 46 minutes/week (with 15 elk; Miller 1991) to about 30 minutes/week (with 20 elk). We observed some elk aged <13 months having difficulty (or unwillingness) in consuming adequate quantities of hay cubes to grow at optimal rates. We supplemented long stem alfalfa hay in addition to hay cubes to these animals and will continue to make long stem hay available to elk <1 year of age in the future. Effects of late pregnancy on intake and apparent digestibility of hay cubes by elk: No significant differences were observed in means from any category of apparent digestibility of hay cubes or quantity of intake by non-pregnant and pregnant elk in their last trimester (P~0.05; Table 2). Non-pregnant cows consumed an average 6609.3 g/day, while pregnant cows averaged 6155.9 g/day (Table 2). Intakes by non-pregnant cows closely approximated the 6807 g/day intake estimate required to achieve daily metabolic energy requirements; however, energy intakes by pregnant cows were markedly lower than the calculated requirement of 8547 g/day. Lower intakes are probably due to insufficient rumen capacity that precludes adequate intake leaving the elk to rely instead on energy stores. To help meet increased energy demands during the last trimester of pregnancy (and lactation) we routinely supplement cows with high energy pelleted ration. On average, cows consumed 23.8 g hay cubes/kg body weight per day. This is markedly higher than reported daily intakes of either poor or good quality hay (14 g/kg BW and 14.8 g/kg BW, respectively; Baker and Hansen 1985, Baker and Hobbs 1987). Bottle-raising neonates: Twenty-three bottle-raised neonates, including 6 elk, 1 bighorn sheep, 12 pronghorn and 8 mule deer, in addition to 2 dam-raised mule deer, were recruited into our captive population in 1991. Four other hand-raised bighorn sheep were sent to Idaho as part of an interagency cooperative program, and two other lambs died from chronic pneumonia in fall of 1991. We imposed forced weaning on some pronghorn fawns in an effort to increase consumption of solid feed earlier in life. Pronghorn in the ad libitum and limit-fed groups remained healthy through the study, except for one male fawn in the limit-fed group that was euthanatized at 63 days of age due to a limb fracture. Daily milk intakes of limited fawns ranged from 780-960 ml, while intakes of fawns fed ad libitum ranged from 487-1060 mI. Intakes of limit-fed fawns were actually greater than ad �118 libitum intakes during 3 of the first 4 weeks of limit-feeding (Fig. 2). This suggests that to truly limit-feed fawns and to attempt to stimulate dry feed intake, we need to further decrease quantity of milk offered by at least 100 ml daily. Body weights of unlimited and limit-fed fawns were not markedly different (Fig. 3). One pronghorn fawn was acquired as an "orphan" in 1992. On 23 June 1992, we received 6 of our goal of 20 female mule deer fawns to hand-raised at FWRF for use in upcoming studies for the Fertility Control project. Three fawns were taken from captive does trapped from the wild and maintained at Little Hills Research Facility, the 3 other fawns were received from various sources as "orphans" (See WP2, J16 for further details). The fawns were diarrhetic when they arrived from Little Hills and Salmonella ~ was cultured by Colorado State University Diagnostic Laboratory. We treated affected fawns with about 33 mg/kg trimethoprim sulfa orally, and provided supportive care (Wild and Miller 1991). More severely affected fawns were placed on injectable gentamicin (2 mg/kg SQ or IV, TID) and ampicillin (6 mg/kg SQ, BID) therapy. Intravenous fluid/electrolyte supplementation was also attempted. HEALTH MAINTENANCE General: Our system for recording daily health status for research animals continued to improve communications, detection and treatment of animal health problems at FWRF. Overall, captive wildlife maintained at FWRF remained healthy throughout the year. Pasteurellosis continued to be a significant health problem in bi-ghorn and is described elsewhere. Severe lumpy jaw lesions and associated mortality declined this year: no lumpy jaw associated mortalities were recorded. Chronic wasting disease (CWO) was confirmed in a 6-year-old elk cow (K86) in June; this represents the third case since our 1985 attempt to eradicate this disease from captive cervids at FWRF. Th.e cow had given birth to a 20 kg calf days before her scheduled euthanasia date. Because of possible vertical transmission of CWO, the calf was euthanatized as well. Trauma was a significant mortality factor in mule deer and pronghorn, primarily due to a dog attack that resulted in 3 mule deer and 3 pronghorn deaths: two dogs penetrated both the perimeter and pen fences to reach the animals. All individuals in our elk herd were negative on the single cervical TB skin test. The threat of TB exposure to our herd is remote since we maintain an almost closed herd, only occasionally adding "orphan" calves from wild populations. We will continue to monitor TB status of our herd in future years however. Vaccination of bighorn lambs failed to prevent development of clinical pneumonia. Antibiotic therapy (Wild and Miller 1991) reduced but failed to completely alleviate clinical signs in these animals. Pregnancy was successfully terminated in 2 mule deer does (as confirmed by significant drops in serum progesterone levels in June), but not in 3 cow elk. Two of 3 cows treated with abortifacient gave birth to healthy calves in May and June, and the other cow is currently pregnant. Likely explanations for.unsuccessful pregnancy termination include insufficient �PROf\JGHORN 1 991 WEIGHTS VS. AGE 24.0 22.0 20.0 18.0 ,_ Q) 16.0 14.0 .:.x. <:» ....., .c 12.0 Q) Q) ~ 10.0 8.0 6.0 4.0 2.0 0.0 0 1 o 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 age (weeks) limited group + unlimited group t--' t--' \0 Fig. 2. Mean daily milk intakes of ad libitum and limit-fed groups of pronghorn fawns. �PRONGHORN 1991 AVERAGE DAILY MILK INTAKES VS. ,A,GE •.... N o 900 ~-------------------------------------------------------- 800 700 /""0.. E 600 <;» <J) .::L o ""'c 500 400 300 200 o 1 2 3 4 5 6 7 8 9 10 11 12 13 14 age (weeks) o Fig. 3. limited group + unlimited group Mean body weigh~s of ad libitum and limit-fed groups of pronghorn fawns. 15 16 �121 dose of Lutalyse· or inability of Lutalyse. to cause abortion in elk (at least at this stage of gestation). Back-calculations from parturition dates indicate that cows were <16 weeks of gestation at time abortifacient injection. Methods for treatins and/or preventins lumpy jaw in captive pron&horn: Bacteroides titers from sera this year have not yet been determined. Clinically, all vaccinated and unvaccinated pronghorn have had at least mild lumpy jaw lesions. This indicates that although high titers were present (Miller 1991), they were not adequately protective. FACILITY MAINTENANCE/REPAIRS/IMPROVEMENTS In addition to numerous daily repairs and maintenance projects, we performed several major improvements. Significant maintenance/repair/improvement projects completed at FWRF this year included: Construction of a feeding and feed storage building in elk pens Wl-2. Construction of a hay cube storage building. Evaluation and replacement of wooden snow fence and Tensare as visual barriers in pronghorn and mule deer pens. Construction of a feed storage area adjacent to the feed shed in E5. Began replacement of the south section of the west alleyway. Construction of 6 new animal paddocks using high tensile New Zealand fencing. Construction of an animal shelter in new paddock A. Construction and modification of fawn rearing pens for mule deer and pronghorn. Planting of windbreaks along the south perimeter fence. Modification of east scale to improve ease of use. Drainage system on west side continued to protect scaleroom from excessive runoff. Construction of a new section of perimeter fence and gate that isolated FWRF from the CSU Radiation Biology section and improved facility security. RESEARCH PROJECTS In addition to ongoing facility management experiments and improvements described above, the following (listed in no particular order of importance) pilot studies, special studies, and research experiments were initiated, conducted, or continued using FWRF animals and facilities this year: - Regulation of mule deer population growth by fertility control: laboratory, field, and simulation experiments (initial GnRH studies on mule deer and elk) -- Baker, Nett, Miller, Hobbs, and Gill. �122 - Effects of ingestion of tungsten-bismuth-tin Ringe1man, Miller, and Ande1t. shot on mallards -- - Preliminary evaluation of antibody responses to Pasteurella haemolytica toxoid in captive bighorn sheep -- Miller and Kraabel. - Use of a medetomidine-ketamine combination for immobilization of bighorn sheep with reversal by atipamezole -- Chaffee, Miller, and Lance. - Effects of pregnancy on intake and digestibility in captive elk -Baker and Magnuson. - New models of the functional response in vertebrate herbivores: the role of plant availability and animal morphology -- Spalinger, Hobbs, Wunder, and Gross. - Relative effectiveness of repellents for reducing feeding by elk -Andelt, Baker, and Burnham. - Relative effectiveness of repellents for reducing feeding by mule deer -- Andelt, Baker, and Burnham. - Epizootiology of pasteurellosis in captive Rocky Mountain bighorn sheep -- Miller, Wild, Mills, and Snipes. - Seasonal changes in fecal cortisol excretion in captive Rocky Mountain bighorn sheep -- Miller and Ritchie. LITERATURE CITED BAKER, D. L., AND D. R. HANSON. 1985. Comparative digestion of grass in mule deer and elk. J. Wi1d1. Manage. 46:807-812. BAKER, D. L., AND N. T. HOBBS. 1985. Emergency feeding of mule deer during winter: tests of a supplemental ration. J. Wildl. Manage. 49: 934-942. BAKER, D. L., AND N. T. HOBBS. 1987. Strategies of digestion: digestive efficiency and retention time of forage diets in montane ungulates. Can. J. Zool. 65:1978-1984. MILLER, M. W. 1990. Animal and pen support facilities for terrestrial wildlife research. Colo. Div. Wildl. Res. Rep. Fed. Aid Proj. W-153R3, WP1A J1, Job Progr. Rep., July 1989-June 1990, Fort Collins. MILLER, M. W. 1991. Animal and pen support facilities for terrestrial wildlife research. Colo. Div. Wildl. Res. Rep. Fed. Aid Proj. W-153R3, WPlA J1, Job Progr. Rep., July 1990-June 1991, Fort Collins. WILD, M. A., and M. W. MILLER. 1991. Bottle-raising wild ruminants in captivity. Outdoor Facts #114, Colo. Div. Wildl., Denver, 6pp. Prepared by ro~;Q Marg et A. Wild Wildlife Research Tech II �123 Colorado Division of Wildlife Wildlife Research Report July 1992 JOB PROGRESS REPORT State of ~C~o~l~o~r~a~d~o~ _ Project No. W-153-R-5 Mammals Research Work Plan No. lA Multispecies Investigations Job No. 3 Mammals 2 Research Administration Period Covered: Author: July I, 1991-June 30, 1992 R. B. Gill Personnel: R. B. Gill, L.E. Lovett, and D. K. Hall Abstract * All objectives for 25 Mammals Research Projects were accomplished within budget allocations. * Job Progress Reports were received for all Mammals Research Projects by the August 31, 1992 deadline. * 9 manuscripts journalsfbooks proceedings. * 32 manuscripts are targeted for preparation in FY 1992-93. were published and 4 articles or are are in in press in professional press· professional society ��125 MAMMALS 2 RESEARCH ADMINISTRATION R. Bruce Gill P. N. OBJECTIVE Administer research studies within the Mammals 2 Research Unit for the highest productivity at the lowest cost. Agreement Ob1ectives 1. Assign and supervise the research of 11 Wildlife Researchers. 2. Assign and supervise secretarial and clerical work of 1 Senior Secretary. 3. Assing and supervise editing and publication work Specialist B. 4. Lead the development and implementation of statewide mammals species management analyses/guides. 5. Assist in the planning and implementation of the Rocky Mountain Arsenal Watchable Wildlife Program. of 1 Publication RESULTS * All objectives for 25 Mammals Research Projects were accomplished within budget allocations. * Program Narratives were prepared and approved for 3 new studies. * Decision item funding was acquired to begin planning efforts for 5 additional studies. * Analyses of issues and alternative management strategies were prepared for deer, elk, and black bear. * 9 manuscripts were journalsfbooks. * 4 articles were published or are in proceedings. * 1 agency technical bulletin was published. * published or are in press press in professional in professional society 6 manuscripts have been submitted to professional journals for peer review. �126 * Staff members served as advisors to: a. b. c. d. e. f. the the the the the the Rocky Mountain Arsenal Watchable Wildlife Advisory Task Force; Habitat Partnership Program; Biodiversity ad hoc planning committee; Northeast Region's moose hunting regulation review process; exotic wildlife and commercial wildlife programs; Terrestrial Wildlife Reorganization Committee. ~~~~j)_ Prepared by ___:_:_~~~ •. R. Bruce Gill Wildlife Research Leader �127 Colorado Division of Wildlife Wildlife Research Report July 1992 JOB PROGRESS REPORT State of Colorado Project No. _W~-1~5~3~-~R~-~2~ _ Mammals Research Work Plan No. _=l!..!A _ Mu1tispecies Investigations Job No. 5 Period Covered: Consulting Services for Mark-Recapture Analysis July 1, 1991 - June 30, 1992 Author: G. C. White Personnel: R. M. Bartmann, R. B. Gill, T. D. I. Beck ABSTRACT Progress towards the objectives of this job include: 1. Simulations were conducted to evaluate the effects of various hunting scenarios on black bear populations. A manuscript was produced, and is undergoing review by the other authors. 2. A study of compensatory effects of harvest on the Piceance Basin mule deer population was continued as part of Federal Aid Project W-1S3-R Work Plan 2 Job 15, entitled Compensatory Effects of Harvest in a Mule Deer Population. Experimental harvests have been conducted in December, 1989, 1990 and 1991. Radio collars to monitor over-winter survival of fawns were placed on the animals during November, 1989, 1990 and 1991. 3. The RADIOS software was expanded to handle elk movement .data recorded as UTM coordinates to assist with the ana.lysis of data from F~dera1 Aid Project W-153-R Work Plan 3 Job 8 entitled Effects of Early Hunting Seasons on Elk Distribution. 4. Consultation has been provided in the design and analysis of a statistical protocol for estimation of the statewide black bear population. As part of this work, the Minta and Mangel (1989) estimator was compared to the joint hypergeometric estimator of Bartmann et a1. (1987) and White and Garrott (1990). 5. Consultation has been provided on sampling and design and analysis procedures to estimate crayfish biomass to aid in river otter recovery research. ��129 CONSULTING SERVICES FOR MARK-RECAPTURE ANALYSES G. C. White P. N. OBJECTIVES Develop a statistical methodology for the estimation of the statewide black bear population. SEGMENT OBJECTIVES 1. Evaluate the Minta and Mangel (1989) estimator of population size for resighting of radio-tagged animals, and compare it to the joint hypergeometric estimator of Bartmann et al. (1987) and White and Garrott (1990). RESULTS AND DISCUSSION Introduction. The estimation of population size using radio tags as marks offers considerable advantages over standard mark-recapture techniques. Amason et a1. (1991) have suggested the phrase -marking and sighting experiments" to identify population estimation techniques where a sample of marked animals are placed in the population, and sightings of marked and unmarked animals are used to estimate population size. When the marked animals carry radio tags, the number of marked animals in the population at any time is known, assuming no censoring from unknown radio failures. Rice and Harder (1977) first reported using simple Lincoln-Petersen estimates corrected for bias (Chapman 1951) for each of 5 sighting surveys of white-tailed deer (Odocolleus virginianus) from a helicopter. The mean of the 5 estimates was used as the overall population estimate, although they also proposed the median of the 5 estimates as an alternative. Marking and sighting also has been used to estimate population size of other ungulates, including mountain sheep (Ovis canadensis canadensis) (Furlow et al. 1981; Leslie and Douglas 1979, 1986; Neal et al. Submitted) and mule deer (Q. hemionus) (Bartmann et al. 1987). Eberhardt (1990) further investigated the Petersen estimator with the Chapman correction (Chapman 1951) for small population sizes where animals could immigrate/emigrate from the study area. His work was motivated by the attempts of Miller et a1. (1987) to estimate black (Ursus americanus) and brown <y. arctos) bear populations in central Alaska. He concluded that both the mean of the individual estimates and the Petersen estimate constructed from the means of the number of marked animals, unmarked animals, and marked animals sighted across sighting occasions provided satisfactory estimates of the average population size on the area of interest. White and Garrott (1990) compared 5 methods of estimating the population size from repeated surveys on a closed population. They compared the mean of the estimates, the mean weighted by the reciprocal of :variance of each of the estimates, the geometric mean of the estimates, the geometric mean weighted by the reciprocal of variance of each of the estimates, the median of the estimates, �130 and the joint hypergeometric maximum likelihood estimator (JHE), first proposed by Bartmann et al. (1987). White and Garrott (1990) concluded that the JHE provided the least biased estimates with approximately 95% confidence interval coverage, although they did not simulate the very small population sizes of Eberhardt (1990). Neal et al. (Submitted) and Neal (1990) performed extensive simulations of the JHE, evaluating the robustness of the estimator to heterogeneity of sighting probabilities, and to lack of independence of sightings (aggregation of animals in groups). Aggregation did not bias the estimate but did decrease precision and confidence interval (CI) coverage. Violation of the assumption of equal sighting probabilities among individuals biased the estimates upwards, especially with smaller populations (~50), but the bias was not large (= 8% maximum). The JHE demonstrated increased bias for the smaller population size simulated, typical of the small sample bias of maximum likelihood estimators. The objective of this paper is to make further comparisons by Monte Carlo simulations of population estimators for mark and sighting surveys. Results from Neal et al. (Submitted) and Neal (1990) will be compared to 2 new estimators. I emphasize the bias of the estimator and its confidence interval coverage as the most important criteria for comparing the estimators simulated. Notation. Ii Number of marked (telemetered) animals in the population at the time of the 1th survey, 1-1, ... ,.t. Hi Number of marked animals in the population that are on the area surveyed at the time of the 1th sighting survey. For all Hi constant, define H E tho Di Number of animals seen during the 1th sighting survey, consisting of mi marked animals and Yi unmarked animals, so that ni - mi + Yi' Ii Number of marked animals seen 1 times during the .t surveys (sighting frequencies). Note that for radio tags, the number of marked animals never sighted, ~, is known. m. Total number of sightings of marked animals seen, so that m. - :Emi - :tili. y. Total number of sightings of unmarked animals seen, so that y. - :tYi' Estimators. Three estimators of population size for marking and sighting experiments will be compared. First is the joint hypergeometric maximum likelihood estimator (JHE) (Bartmann et al. 1987, White and Garrott 1990, Neal 1990, Neal et al. Submitted). JHE is the value of If which maximizes the following likelihood (N): and the terms are defined for all i - 1 to k sighting occasions. The estimate N can be found by iterative numerical methods. Confidence intervals are determined with the profile likelihood method (Hudson 1971, Venzon and Moolgavkar 1988). This estimator assumes that all the marked animals are on the area surveyed for each survey, i.e., that the population is geographically closed. Hence, the number of marked animals (M) is constant for each survey, although the sighting probability is not assumed to be constant for each survey. �131 I< :l(N I N, n., mi) =11 N N - N mi ni - mi (1) 1-1 [:.] Second, Minta and Mangel (1989) suggested a bootstrap estimator (MM) of population size based on the sighting frequencies of the marked animals, Ii. For unmarked animals, sighting frequencies are drawn at random from the observed sighting frequencies of the marked animals until the total number of captures equals y. . The number of animals sampled is then an estimate of the number of unmarked animals in the population, so that H plus the number sampled is an estimator for H. Only samplings where the number of sightings was equal to y. were used, i.e., cases where the cumulative sightings exceeded y. were rejected. Minta and Mangel (1989) accepted the first value where the cumulative sightings equalled or exceeded y. . The stopping rule I used results in less bias than the rule used by Minta and Mangel (1989). One-thousand replications were performed to generate the distribution of Minta and Mangel (1989) suggested the mode of the 1000 replicates as the population estimate. Confidence intervals were computed as probability intervals with the 25~ and 975~ order statistics from this sample of 1000 estimates. H. Third, the JHE estimator has been extended to accommodate immigration and emigration (Neal et al. Submitted) through a binomial process. This estimator is labeled IEJHE, and does not assume that the population is geographically closed. Assume that the total population with any chance of being observed on the study area is H*, and that at the time of the 1th sighting survey, Hi animals occur on the study area. I am interested in estimating the mean number of animals on the study area, ft, and possibly H*. At the time of the ith sighting occasion, a known number of the marked animals (Hi) are on the study area of the possible Ii animals with transmitters. The probability that an individual is on the study area on the !th occasion can be estimated as HilIi, or in terms of the parameters of interest as Hi/H*. Then the likelihood for the model that includes immigration and emigration is a product of the binomial distribution for emigration/immigration times the joint hypergeometric likelihood of Eq. (1): (2) The parameters H* and Hi for 1-1 to ~ can be estimated by numerical iteration to maximize this likelihood, with the constraints that Hi > (Hi + Yi) and H* > Hi for 1-1 to~. Profile confidence intervals can be obtained for the ~+l parameters. I was not interested in the ~ population estimates for each sighting occasion, but rather desired the mean of the Hi estimates. Therefore, Ire-parameterized the likelihood to estimate the total population and mean population size on the �132 study area directly, and their profile likelihood confidence intervals. re-parameterized likelihood, I used Hi - H + ai' where E ai - O. In the Simulations. Two sets of simulations were used to compare JHE and MM. The basic simulations assumed constant probability of capture and sighting across animals, and thus met the assumptions of the JHE. Four population sizes (50, 100, 200, 500), 4 sets of sighting occasions (5, 10, 15. 20), 4 sighting probabilities (0.1,0.3,0.5,0.7), and 3 capture probabilities (0.1, 0.3,0.5) were simulated, yielding 192 scenarios. Heterogeneity simulations assumed variable sighting probabilities across animals, with the sighting probability for each animal drawn from a beta distribution with a - 3 and ~ - 3. Four population sizes (50, 100, 200, 500), 4 sets of sighting occasions (5, 10, 15, 20), and 3 capture probabilities (0.1. 0.3, 0.5) were simulated, yielding 48 scenarios. Each scenario was simulated 1000 times. For each scenario, number of marked animals was determined as a random binomial variable for population size and capture probability prior to sighting. Number of marked animals seen on the .!.th sighting survey was dete'rmined as a Bernoulli trial for each animal given its sighing probability, and likewise for unmarked animals. lEJHE was evaluated with a subset of the basic simulations: 4 population sizes (67, 133, 266, 667), 2 sets of sighting occasions (5, 10), 1 capture probability (0.3) and 2 sighting probabilities (0.5, 0.7). Probability of being on the area of interest at the time of a survey was 0.75, reSUlting in average densities of 50, 100, 200, and 500. One restriction for all simulations was that the number of marked animals in the population had to be at least 10, to represent a reasonable radio-tagging effort. Cases with Ii < 10 were rejected. This restriction is most important in interpreting scenarios with H - 50 and capture probabilities of 0.1, as at least 10 marked animals were always in the population. Ii was simulated given an expected capture probability so that results would be useful for designing studies prior to radio tagging. i.e., a trade-off between radio-tagging (capture) effort and sighting effort could be evaluated. To evaluate the simulations, percent relative bias (PRB) , Cl length, percent Cl coverage, and percent of confidence intervals lying above the true parameter were used. PRB was estimated as PRB - N) = (E(iL) x 100 (3) N where E(H) is estimated as the mean of the simulated estimates. Overestimation (E(H) > H) results when PRB is positive, and underestimation (E(H) < H) results when PRB is negative. Percent Cl length (PClL) is the length of the Cl divided by the population size x 100 and allows comparisons across population sizes, although this statistic is generally only useful when 95% coverage has been achieved. Confidence intervals were computed as integers for JHE, MM, and for total population size of lEJHE. Confidence interval endpoints were defined to be within the 95% confidence interval, i.e., for true H equal to the lower bound, coverage was assumed. Hypothetically, confidence interval length can be zero. Such a situation might occur when H - H and both the upper and lower bound equal �133 i, for instance, when the population estimate is equal to the number of marked animals plus the maximum number of unmarked animals sighted. Results. Results for the basic simulations for JHE (Table 1) and MM (Table 2) where all assumptions of both estimators are met suggest that neither estimator shows serious bias, although both show significant positive bias (£ < 0.001). JHE shows greater positive bias than does MM (f < 0.001). Bias for JHE ranged from -0.2 to 19.9% for the 192 scenarios, and from -0.1 to 15.6% for MM. Smaller population sizes (f < 0.001), lower sighting probabilities (l < 0.001), lower capture probabilities (f < 0.001), and fewer sighting occasions (f < 0.001) lead to increased positive bias for both estimators. For the basic simulations, coverage for JHE is close to the expected 95% for all cases (mean 95.3%, range 93.6 to 96.7% for 192 scenarios), whereas coverage for MM is far below 95% (mean 70.3%, range 44.0 to 90.3% for 192 scenarios). Coverage was not significantly different among the 192 scenarios (f - 0.319) for JHE. For MM, larger population sizes (f < 0.001), lower sighting probabilities (£ < 0.001), lower capture probabilities (f < 0.001), and fewer sighting occasions (f < 0.001) lead to decreased confidence interval coverage. Lower capture probabilities had the most striking effect on coverage, with the coverage approaching 50% for capture probabilities of 0.1 . .Results for the heterogeneity simulations for JHE (Table 3) and MM (Table 4) again suggest that both estimators show slight positive bias (~ < 0.001), but the bias is not biologically serious (mean 0.6%, range for 48 scenarios of -0.2 to 2.7% for JHE; mean 0.6%, range -0.2 to 2.3% for MM). Smaller population sizes (f < 0.001), fewer sighting occasions (f < 0.046), and lower capture probabilities (£ < 0.001) lead to increased bias for both estimators. For the heterogeneity simulations, confidence interval coverage for both estimators was considerably less than the expected 95% (f < 0.001). JHE had better average coverage with 78.8% (range 68.4 to 90.4% for 48 scenarios) compared to 70.1% for MM (range 45.2 to 86.6% for 48 scenarios). For JHE, population size (£ - 0.505) and capture probabilities (l- 0.818) showed little effect, whereas increasing the number of sighting occasions decreased coverage (f - 0.001). For MM, larger population sizes (f < 0.001), and lower capture probabilities (f < 0.001) most affected coverage, with no effect from number of sighting occasions (f < 0.867). PClL for MM was considerably less than for JHE in all cases. This effect is because this estimator conditions on the number of unmarked animals sighted, and the variance of this term (var u.) is not accounted for in the estimation of confidence intervals (K. P. Burnham Pers. Comm.). As a result, even with no serious bias, confidence interval coverage is below the expected 95%. Discussion. One interesting result for MM is the closeness 0: ,:IL for Table 1 (no individual heterogeneity) and Table 3 (heterogeneity). I ~vuld expect wider confidence intervals for heterogeneity. However, the mean sighting probability is 0.5 in Table 3, and so the heterogeneity simulations do not provide as great a range in sighting probabilities across occasions as do the simulated data in Table 1. Hence, the closeness of the observed PClL values may be partially an artifact of the parameters simulated. �134 Coverage of MM can be improved by accounting for the variation of the number of unknown animals sighted. I also incorporated the Robbins-Monro procedure (Buckland and Garthwaite 1990) to provide a more sophisticated approach to computing the MM confidence interval. The population estimate N was computed as recommended by Minta and Mangel (1989). Then, I used algorithm AS 259 (Buckland and Garthwaite 1990) to compute the upper and lower confidence bounds, where the number of unmarked animals sighted (y.) was simulated based on the observed frequencies (in contrast to Minta and Mangel where the observed value was used). Coverage greatly improved for MM with the above procedure (Tables 7 and 8), with the average coverage for the basic simulations 82.8%. However, coverage was still below the expected 95% because the confidence interval is conditional on the observed sighting frequencies of the marked animals. Coverage was 95.0% for the cases where capture probability was 0.5, but was considerably less for capture probabilities of 0.3 (86.8%) and 0.1 (66.6%). I attribute this lack of coverage to the small sample properties of the bootstrap procedure. Specifically, the variation inherent in the sampling process is not show exhibited in the sample when less than 50% of the population is marked. The poor coverage characteristics of MM were not discussed in the original paper (Minta and Mangel 1989), as they did not present coverage results for Monte Carlo simulations. Examples presented by Minta and Mangel (1989) had small numbers (~16) of marked animals, and hence I would expect poor coverage for these cases if the true population size were known. As shown in Table 8, 50% of the population must have marks for 95% coverage to be achieved with MM. Because most radio-tagging studies do not mark this large of proportion of the population being studied, I would expect that MM would typically produce a relatively unbiased estimate, but the confidence interval would be too narrow. Likewise, for typical radio -tagging studies, JHE would also produce too narrow of confidence interval because typically animals do not all exhibit the same sighting probability on a particular occasion. The 3 estimators simulated are all relatively complex to compute, compared to, say, the arithmetic mean of the Lincoln-Petersen estimates. However, White and Garrott (1990) showed the additional effort to compute the JHE was worthwhile. The arithmetic mean showed significantly lower coverage than the expected 95% for Monte Carlo simulations, while JHE was very close to 95%. For situations where the arithmetic mean did exhibit 95% coverage, the confidence interval was wider than that of JHE. Program NOREMARK for PC computers is available from the author to compute the 3 estimators discussed in this manuscript. Acknowledpents. A. K. Neal worked diligently to develop the initial evaluations of the JHE estimator. K. P. Burnham, D. R. Anderson, R. E. Kenward, and R. M. Cormack provided helpful comments on the manuscript. Literature Cited. Amason, A.N., C.J .. Schwarz, and J.M. Gerrard. 1991. Estimating closed population size and number of marked animals from sighting data. J. Wi1dl. Manage. 55:716-730. Bartmann, R.M., G.C. White, L.H. Carpenter, and R.A. Garrott. 1987. Aerial markrecapture estimates of confined mule deer in pinyon-juniper woodland. J. Wi1dl. Manage. 51:41-46. �135 Buckland, S.T., and P.H. Garthwaite. 1990. Algorithm AS 259 -- Estimation confidence intervals by the Robbins-Monro search process. Applied Stat. 39:413-424. Chapman, D.G. 1951. Some properties of the hypergeometric distribution with applications to zoological sample censuses. Univ. Calif. Pub. Stat. 1:131160. Eberhardt, L.L. 1990. Using radio-telemetry for mark-recapture studies with edge effects. J. Appl. Ecol. 27:259-271. Furlow, R.C., M. Haderlie, and R. Van den Berge. 1981. Estimating a bighorn sheep population by mark-recapture. Desert Bighorn Council, Trans. 1981:3133. Hudson, D.J. 1971. Interval estimation from the likelihood function. J. Royal Stat. Soc. Series B 33:256-262. Leslie, D.M., Jr. and C.L. Douglas. 1979. Desert bighorn sheep of the River Mountains, Nevada. Wildl. Monogr. 66:1-56. __ ---:'D.M., Jr. and C.L. Douglas. 1986. Modeling demographics of bighorn sheep: current abilities and missing links. N. Amer. Wildl. Natur. Res. Conf., Trans. 51:62-73. Miller, S.D., E.F. Becker, and W.H. Ballard. 1987. Black and brown bear density estimates using modified capture-recapture techniques in Alaska. International Conf. on Bear Research and Management 7:23-35. Minta, S. and M. Mangel. 1989. A simple population estimate based on simulation for capture-recapture and capture-resight data. Ecology 70:1738-1751. Neal, A.K. 1990. Evaluation of mark-resight population estimates using simulations and field data from mountain sheep. M. S. Thesis, Colorado State Univ., Fort Collins. 198pp. Neal, A.K., G.C. White, R.B. Gill, D.F. Reed, and J.H. alterman. Submitted. Evaluation of mark-resight methods to estimate mountain sheep numbers. J. Wildl. Manage. Rice, W.R. and J.D. Harder. 1977. Application of multiple aerial sampling to a mark-recapture census of white-tailed deer. J. Wildl. Manage. 41:197-206. Venzon, D.J. and S .H. Moolgavkar. 1988. A method for computing profi1elikelihood based confidence intervals. Applied Statistics 37:87-94. White, G.C. and R.A. Garrott. 1990. Analysis of wildlife radio-tracking data. prep.:::d:i~~1e~k';:, ~' C U.~ Dr. Gary C. Whi tfl) Professor �136 Table 1. Percent relative bias (PRS), mean percent confidence interval length (PClL), and coverage results for JHE with basic simulations, where all assumptions of the mark-resight estimator are satisfied. Above gives the percent of confidence intervals where the lower bound exceeded N. Subdivision PRS SE PRS 0.029 All variables 1.044 Population size 50 1.766 0.081 0.068 100 1.373 200 0.046 0.796 0.025 500 0.239 Number of sighting occasions 0.092 5 2.181 0.053 10 0.917 15 0.574 0.040 20 0.033 0.502 Sighting probability 0.1 0.106 3.065 0.3 0.041 0.745 0.5 0.026 0.264 0.7 0.017 0.101 Capture probability 0.1 2.025 0.073 0.3 0.754 0.042 0.5 0.351 0.025 Mean PClL 34.600 SE PClL 0.322 53.008 41.386 27.764 16.242 Coverage Above (Xl (Xl 95.34 2.29 1.015 0.754 0.204 0.074 95.32 95.35 95.28 95.43 2.27 2.27 2.36 2.26 60.468 32.126 24.823 20.982 1.254 0.213 0.123 0.097 95.35 95.18 95.43 95.42 2.31 2.35 2.19 2.32 78.964 29.651 18.365 11.418 1.260 0.104 0.058 0.036 95.37 95.32 95.42 95.28 2.29 2.43 2.17 2.27 57.091 29.264 17.443 0.886 0.360 0.088 95.34 95.29 95.40 2.31 2.29 2.27 Table 2. Percent relative bias (PRS), mean percent confidence interval length (PClL), and coverage results for Minta and Mangel estimator (1989) with basic simulations, where all assumptions of the mark-resight estimator are satisfied. Above gives the percent of confidence intervals where the lower bound exceeded N. Subdivision PRS SE PRS 0.028 All variables 0.785 Population size 50 1.379 0.077 100 0.952 0.063 200 0.629 0.046 500 0.180 0.025 Number of sighting occasions 5 1.528 0.086 10 0.761 0.051 15 0.414 0.038 20 0.437 0.033 Sighting probability 0.1 0.100 2.218 0.3 0.621 0.041 0.5 0.234 0.026 0.7 0.067 0.016 Capture probability 0.1 1.743 0.071 0.3 0.514 0.039 0.024 0.5 0.098 Mean PClL 13.382 SE PClL 0.026 Coverage (X) 70.26 Above (X) 15.18 20.604 15.075 10.874 6.974 0.070 0.049 0.034 0.022 74.07 69.81 68.82 68.33 13.48 15.36 15.85 16.04 19.503 13.571 10.980 9.472 0.070 0.048 0.039 0.033 69.36 69.93 70.65 71.08 15.65 15.29 14.82 14.97 26.331 13.139 8.551 5.505 0.066 0.032 0.021 0.013 68.74 69.19 70.68 72.41 16.40 16.08 14.68 13.57 14.774 13.692 11.679 0.050 0.046 0.039 51.89 73.31 85.57 24.44 13.73 7.38 �J.._Jj Table 3. Percent relative bias (PRB), mean percent confidence interval length (PClL), and coverage results for JHE estimator from simulations with heterogeneity of individual sighting probabilities (X sighting probability - 0.5). Above gives the percent of confidence intervals where the lower bound exceeded N. Subdivision PRB SE PRB All variables 0.595 0.039 Population Size 50 0.890 0.102 100 0.817 0.090 200 0.434 0.067 500 0.239 0.041 Number of sighting occasions 0.907 0.094 5 10 0.595 0.076 0.521 0.073 15 0.357 0.069 20 Capture probability 0.1 1.246 0.095 0.3 0.43i 0.059 0.5 0.108 0.037 Mean PClL 18.701 SE PClL 0.064 25.997 22.113 16.429 10.265 Coverage Above (X) (X) 78.85 10.27 0.142 0.139 0.102 0.057 79.27 78.89 78.58 78.67 9.84 10.12 10.42 10.68 27.998 18.844 15.094 12.868 0.174 0.107 0.085 0.071 88.99 81.49 74.79 70.13 5.58 9.09 12.31 14.09 28.920 16.738 10.444 0.128 0.083 0.048 78.59 79.01 78.96 10.51 10.14 10.16 Table 4. Percent relative bias (PRB), mean percent confidence interval length (PClL), and coverage results for the Minta and Mangel (1989) estimator from simulations with heterogeneity of individual sighting probabilities (X sighting probability - 0.5). Above gives the percent of confidence intervals where the lower bound exceeded N. Subdivision PRB SE PlU\ 0.039 All variables 0.562 Population Size 50 0.950 0.102 100 0.824 0.088 200 0.309 0.066 500 0.164 0.041 Number of sighting occasions 5 0.727 0.091 10 0.546 0.076 15 0.556 0.073 20 0.418 0.070 Capture probability 0.1 1.180 0.093 0.3 (L058 0.384 0.5 0.122 0.038 Mean PClL 13.377 SE fClL 0.027 20.658 15.079 10.842 6.929 Coverage Above (Xl (Xl 70.08 15.23 0.047 0.030 0.019 0.011 73.87 69.80 68.88 67.79 13.45 15.52 15.59 16.37 15.740 13.320 12.469 11.979 0.063 0.053 0.049 0.047 69.96 70.23 70.33 69.81 15.48 15.17 15.22 15.07 14.767 13.660 11.705 0.051 0.047 0.041 52.07 73.24 84.93 24.29 13.66 7.75 �138 Table 5. Percent relative bias (PRB). mean percent confidence interval length (PClL). and coverage results for total population (N*) estimator of lEJHE. where all assumptions of the estimator are satisfied. Above gives the percent of confidence intervals where the lower bound exceeded N*. Subdivision PRB SE PRB 0.047 All variables -0.033 Total Population size 67 0.104 0.138 133 -0.034 0.097 267 -0.073 0.067 0.041 667 -0.008 Number of sighting occasions -0.092 0.077 5 0.053 10 0.087 Sighting probability 0.075 0.5 -0.151 0.7 0.146 0.055 Coverage (X) 97.09 Above (X) 1.26 Mean PClL 22.297 SE PClL 0.097 36.424 24.855 17.219 10.690 0.199 0.111 0.068 0.039 97.80 96.97 96.85 96.72 0.80 1.52 1.32 1.40 26.200 18.394 0.153 0.101 96.87 97.30 1.40 1.12 25.440 19.154 0.152 0.108 96.80 97.37 1.49 1.04 Table 6. Percent relative bias (PRB). mean percent confidence interval length (PClL). and coverage results for average population <in estimator of lEJHE. where all assumptions of the estimator are satisfied. Above gives the percent of confidence intervals where the lower bound exceeded H. Subdivision PRB SE PM 0.043 All variables 0.187 Average Population size 50 0.456 0.128 100 0.251 0.089 200 0.032 0.060 500 0.007 0.038 Number of sighting occasions 5 0.134 0.071 10 0.239 0.049 Sighting probability 0.5 0.053 0.071 0.7 0.320 0.049 Mean Coverage (X) 94.88 Above (X) 2.07 18.364 SE PClL 0.085 30.150 20.497 14.101 8.708 0.183 0.106 0.066 0.038 95.37 94.78 95.12 94.25 1.47 2.17 2.00 2.62 21.500 15.229 0.134 0.090 94.30 95.46 2.42 1. 71 22.176 14.553 0.134 0.084 95.09 94.67 2.17 1.96 PerL �Table 7. Percent relative bias (PRB), mean percent confidence interval length (PClL), and coverage results for Minta and Mangel estimator (1989) using RobbinsMonro confidence intervals (Buckland and Garthwaite 1990) for the basic simulations. Above gives the percent of confidence intervals where the lower bound exceeded N. Subdivision PRB All variables 0.815 Population size 50 100 200 500 1.358 1.082 0.622 0.199 SE PRB Mean PClL 0.028 20.018 0.075 0.066 0.045 0.026 SE PClL Coverage Above 0.038 (X) 82.81 7.49 31.399 22.537 16.012 10.123 0.098 0.069 0.049 0.031 86.04 82.40 81.11 81.68 5.32 7.55 8.61 8.48 0.085 0.053 0.040 0.033 28.558 20.256 16.692 14.564 0.100 0.070 0.057 0.049 82.26 82.82 82.87 83.28 7.86 7.55 7.44 7 12 0.100 0.040 0.026 0.017 38.154 19.720 13.231 8.967 0.094 0.048 0.032 0.022 81.60 82.57 83.04 84.03 8.94 8.01 7.13 5.88 0.070 0.041 0.024 22.012 20.467 17.574 0.071 0.067 0.057 66.55 86.83 95.04 15.42 5.34 1.71 (X) Number of sighting occasions 5 10 15 20 1.554 0.716 0.561 0.430 Sighting probability 0.1 0.3 0.5 0.7 2.340 0.594 0.231 0.095 Capture probability 0.1 0.3 0.5 1.799 0.554 0.091 Table 8. Percent relative bias (PRB), mean percent confidence interval length (PClL). and coverage results for Minta and Mangel (1989) estimator using RobbinsMonro confidence intervals (Buckland and Garthwaite 1990) for simulations with heterogeneity of individual sighting probabilities (X sighting probability - 0.5). Above gives the percent of confidence intervals where the lower bound exceeded N. Subdivision PRB All variables 0.559 Population Size 50 100 200 500 0.968 0.745 0.343 0.178 SE fRB Mean fClL SE PClL 0.040 19.966 0.105 0.089 0.068 0.040 Coverage Above (Xl (Xl 0.042 82.53 7.46 31.317 22.492 15.972 10.083 0.066 0.044 0.028 0.016 85.25 82.12 80.75 82.00 5.53 7.60 8.43 8.28 0.095 0.079 0.072 0.070 23.310 19.932 18.634 17.990 0.095 0.081 0.075 0.073 82.33 82.72 82.79 82.28 7.58 7.41 7.18 7.67 0.095 0.060 0.039 21.974 20.391 17.534 0.077 0.073 0.062 66.32 86.66 94.61 15.17 5.36 1.86 Number of sighting occasions 5 10 15 20 0.695 0.599 0.421 0.518 Capture probability 0.1 0.3 0.5 1.126 0.443 0.106 ��141 Colorado Division Wildlife Research July 1992 of Wildlife Report JOB PROGRESS State of Project REPORT Colorado No. W-153-R-4 Mammals Research Work Plan No. 1A Multispecies Job No. 6 Monitoring and Managing in Colorado Period Covered: Investigations Wildlife Health July 1, 1991 - June 30, 1992 Authors: M. W. Miller, S. Williams Personnel: C. W. McCarty, M. L. Stevens, C. A. Mehaffy, W. J. Adrian, J. Ritchie, T. R. Spraker, and E. T. Fulk, R. Forde Abstract We monitored wildiife populations throughout Colorado for occurrence of disease using a combination of extensive and intensive approaches. A statewide surveillance program was developed for acquiring, examining, reporting on, and summarizing sporadic wildlife disease cases occurring throughout Colorado. At least 21 wildlife cases were submitted for diagnostic examination during July-September. Most cases appeared to represent isolated incidents of trauma or disease. Of these, possible mercury intoxication in Front Range raptors and pneumonia in a bighorn from Glenwood Canyon may warrant further investigation. We continued the annual statewide survey of deer, elk and pronghorn hunters to collect sera for brucellosis screening, but also began modifying and evaluating our survey program in order to improve its efficiency. Of 9,595 hunters surveyed, 1,105 (12%) returned blood samples from antlerless elk harvested throughout Colorado during October-December 1991 for brucellosis scree.ning. Only 461 (42%) of returned samples were usable; marked hemo LysLa and/or contamination precluded evaluation of tpe remaining 644 samples. All sera tested were negative for antibodies to Brucella spp. on the standard card test. Overall, about 5% of the survey kits distributed to hunters provided samples usable in this year's brucellosis survey. The 12% sample return.rate was somewhat lower than anticipated (about 33% f6r elk), but may have reflected relatively low success in harvesting ant1erless elk during regular rifle seasons in several parts of Colorado during 1991. However, nearly half of all returned samples yielded usable sera -- this trend, if sustained, represented a small but notable improvement in the proportion of usable samples returned (also about 33% for elk). These data, combined with those collected during FY 92/93, will be used in assessing strategies for improving the efficiency of statewide serologic surveys that depend on blood samples submitted from harvested animals. At least 12 cases of chronic wasting disease (CWO), a spongiform encephalopathy, have been confirmed in free-ranging deer and elk in Larimer County since 1985 -- all of these are from GMUs 9, 191, 19, or 20. Because �142 reliable estimates for distribution and prevalence of CWO in wild cervids are lacking, we initiated the first in a series of surveys for CWO on select deer and elk populations throughout Colorado to be conducted over the next 3-5 years. Brains from 51 harvested mule deer and 33 harvested elk from the Forbes Trinchera Ranch near Ft. Garland (DAU D31/E33) and 4 brains from harvested elk near Estes Park (DAU E9) were examined for evidence of chronic wasting disease. All 88 brains examined were negative for spongiform encephalopathy. However, 15 (29%) mule deer and 15 (45%) elk from DAU D31/E33 showed mild to moderate encephalitis/meningoencephalitis composed primarily of perivascular lymphocytic cuffs. The cause of these nonsuppurative lesions was not apparent, but it is unlikely that they were clinically important. Past viral infection or reaction to circulating trypanosomes were deemed the most likely probable causes for the lesions observed, and additional tests to determine the cause of these lesions are planned. Bovine tuberculosis was diagnosed in captive elk residing on a game ranch near Powderhorn, CO in June 1991. The infected herd was subsequently destroyed, but the severity and duration of this outbreak prompted investigations into the possibility that tuberculosis might have spread to free-ranging wildlife outside the infected premises. Consequently, 10 mule deer does and 1 cow elk were collected on 19 August 1991 in the immediate vicinity of the infected premises. All of these animals appeared healthy, and we did not observe lesions suggestive of tuberculosis in any of them; histologic examinations and mycobacterial culture also failed to demonstrate tuberculosis infections. Similarly, tuberculous lesions were not seen in hunter-killed deer or elk from GMUs near Powderhorn examined in November 1991. Although initial examinations of wild deer and elk collected in the immediate vicinity of the tuberculosisinfected game ranch have revealed no indication of infection in free-ranging animals, sample sizes to date are small «25 mule deer and elk) and the probability of failing to detect infection is relatively high (e.g., 0.778 if prevalence is 1%). Consequently, we plan to develop and implement a more structured pilot program for monitoring wildlife populations in the Powderhorn area. This program will be used to continue tuberculosis surveillance in this area over the next 5 years, and may also be used in other locations where tuberculosis or other serious disease problems are diagnosed on game ranches. �143 MONITORING AND MANAGING WILDLIFE HEALTH IN COLORADO M. M. W. T. W. Miller L. Stevens J. Adrian R. Spraker and E. S. Williams P. N. OBJECTIVES Develop and implement a program for enhancing statewide efforts manage health of Colorado's terrestrial wildlife populations. AGREEMENT to monitor OBJECTIVES 1. Modify and improve systems for submitting, diagnosing and reporting sporadic disease cases in wild animals throughout Colorado. 2. Develop and use databases for assimilating and analyzing problems identified through surveillance and surveys. and managing wildlife on data on disease 3. Design, conduct, and report results of surveys for brucellosis, tuberculosis, and chronic wasting disease in specific deer and/or populations. 4. Provide assistance in investigating outbreaks in Colorado. and elk disease 5. Design experiments to develop and/or improve techniques for investigating wildlife diseases; begin conducting approved and funded research. Maintaining healthy wildlife populations is a fundamental component of sound wildlife management practices. Habitat degradation, high animal density, extreme weather, and disease can act singly or in combination to compromise the overall health of a wildlife population. As Colorado's wildlife managers, we have developed a variety of tools for monitoring and assessing the effects of habitat loss, animal numbers, and weather on wildlife populations. We have also invested considerably in developing tools to manage these factors to optimize performance of the wildlife populations in our stewardship. In contrast, monitoring and managing the effects of disease on wildlife population performance have received relatively little attention (with a few notable exceptions). This lack of attention may be rooted to some extent in a widely-held belief that wildlife diseases are symptoms of larger underlying population problems that will be resolved if those larger problems are managed properly. Despite this belief, disease can be a significant obstacle to effective and efficient wildlife management in Colorado. Disease outbreaks account for substantial mortality in some wildlife populations. Introduced pathogens have potential to decimate local wildlife populations. Some diseases depress wildlife population performance to levels below resource-based carrying capacity. Many wildlife diseases are shared with domestic animals and/or humans, and in some cases wildlife populations serve as reservoirs for these �144 agents. Disease also detracts from the aesthetic value of wild animals, and may convey a perception of mismanagement to uninformed publics. For these reasons, diseases should be regarded as an integral part of wildlife population dynamics and wildlife management. Select wildlife health problems have been monitored in Colorado for more than 30 years. These longstanding efforts have provided useful information on the diseases studied. However, because these efforts have not always been coordinated on a statewide basis, and because some findings have not been widely available to managers and policy makers, applications to overall management programs have been limited. In order to improve our collective ability to manage wildlife health in Colorado, we need a more coordinated and systematic approach for monitoring, investigating, and reporting on health problems in free-ranging wildlife. A more complete understanding of wildlife diseases and their effects on population performance is fundamental to comprehensive wildlife management. Enhanced surveillance efforts will provide a mechanism for detecting health problems throughout the state before serious impacts to wildlife populations occur. Assimilating diagnostic data will aid in assessing trends suggestive of population-level disease problems. Programs for conducting extensive and intensive surveys for potential and realized wildlife diseases will provide reliable prevalence and distribution data for managers and administrators to use in decision making. Expertise in investigating and managing epizootics and epornitics will ameliorate efficacy and efficiency of efforts to control outbreaks. Improved techniques for diagnosing and studying wildlife diseases will provide a firm foundation for health management programs designed to enhance the quality of Colorado's wildlife populations. MATERIALS Disease AND METHODS Surveillance We monitored wildlife populations throughout Colorado for occurrence of disease using a combination of extensive and intensive approaches. These were organized and conducted as follows: Statewide Surveillance We developed a program for acquiring, examining, reporting on, and summarizing sporadic wildlife disease cases occurring throughout Colorado. A formalized process for submitting cases was developed; this process included criteria for acceptable submissions, submission forms, handling instructions, and a system for networking submissions within CDOW and among the 3 Colorado State Veterinary Diagnostic Laboratories and the Wyoming State Veterinary Laboratory. Information on this system was disseminated through CDOW's internal newsletter ("Tracking Wildlife"), a presentation to CDOW's Terrestrial Biologists, and direct communications with other field personnel. All submissions were subjected to necropsy. Ancillary diagnostics, including histopathology, bacteriology, virology, serology, parasitology, and toxicology were performed at the discretion of the attending pathologist. Preliminary examination and/or test results were telephoned to CDOW's Wildlife Research Center Laboratory within 3 days of completion, and a final report was due within 15 business days of submission. Pertinent data from preliminary and final reports were entered into a permanent database (described below), and copies of reports were filed as well as sent to appropriate field personnel. �145 In conjunction with developing a process for case submissions, we created a computerized database to use in recording, analyzing, and reporting on disease problems occurring throughout Colorado. Both current and archived cases were entered; however, data entry problems precluded summarizing archived entries for this report. Information recorded included species, age, sex, location, number affected, diagnosis, and other pertinent data as available. This database was used to generate quarterly and annual wildlife morbidity and mortality reports. In addition, data are available for analysis of long-term trends in select wildlife disease problems. Surveys Brucellosis Survev: We continued the statewide survey of deer, elk and pronghorn hunters to collect sera for brucellosis screening. Over the next 5 years, however, we plan to develop and implement strategies for expanding utility and improving efficacy and efficiency of this survey. In particular, we will focus on improving return rates on sampling kits, quality of samples returned, and ability to target specific areas or populations for surveillance. Initially, we began examining performance of the existing survey to determine average return rates and sample usability by species and season. We modified mailer kits as described below, and will evaluate those changes by comparing survey returns and sample usability after modifications with means from previous surveys. Based on these assessments and results of intensive survey efforts, we will make additional recommendations for improving the survey in annual Job Progress Reports. Survey results will be reported annually in the Job Progress Report. In addition to the statewide survey, we began developing a process for intensively sampling specific geographic areas or populations using a modified hunter survey. We designed and conducted pilot surveys focused on sampling in select DAUs and/or GMUs to be conducted over a 2-year sampling period. Our survey was constructed such that the probability of failure to detect at least 1 case of brucellosis in the selected population will be ~0.1 even if herd prevalence is 1%. We will compare return rates and sample usability among seasons and collection methods, and use these comparisons to guide future survey efforts. Results for intensive surveys will be reported in the annual Job Progress Report. To initiate our evaluation of the statewide survey system, we mailed about 9,595 blood sampling kits to elk hunters to gather samples for CDOW's annual brucellosis surveillance program conducted in cooperation with the Colorado Department of Agriculture's State/Federal Brucellosis Laboratory in Denver. Over 7000 kits went to sportsmen with antlerless elk permits for the second regular season. In addition, over 1400 second regular season and all 525 late season hunters with antlerless permits in DAU E2 received kits as part of an intensive sampling effort in the vicinity of a recent bovine brucellosis outbreak. We also mailed kits to over 400 second season cow elk hunters in DAU E25 to augment ongoing tuberculosis surveillance activities. Based on recommendations made by participants in a meeting held in March 1991 to review this ongoing survey program, we modified our sampling approaches in the following ways: 1. The letter to hunters included in the kit was rewritten to read: Dear Elk Hunter: The Colorado Division of Wildlife, in cooperation with the Department of Agriculture, is continuing to survey the health status of Colorado's big game populations. In order to carry out a successful �146 survey, it is necessary to sample large numbers of animals. The only way to obtain these samples is with your cooperation. To provide the most useful samples for this survey, please follow the instructions below: 1. IMMEDIATELY UPON KILL, FILL THE BLOOD TUBE COMPLETELY WITH BLOOD FROM THE CHEST CAVITY OR NECK AS YOU BLEED YOUR ANIMAL OUT. 2. CLOSE THE TUBE WITH THE STOPPER SUPPLIED, PLACE THE BLOOD TUBE IN THE SMALL PLASTIC BAG, SEAL THE BAG, AND PLACE IN THE CARDBOARD MAILER. 3. PLEASE 4. DROP YOUR MAILER AT THE NEAREST POST OFFICE AFTER LEAVING THE FIELD -- IT WILL BE SENT DIRECTLY TO A DENVER LAB FOR PROCESSING. POSTAGE KEEP YOUR SAMPLE AS COOL AS POSSIBLE IS WITHOUT FREEZING. PREPAID. Your cooperation is appreciated and will result in better game management through a better understanding of the health status of our wildlife populations. THANK YOU! 2. A condensed mailing box. version of sampling instructions was stamped 3. Kits were distributed in a manner allowing evaluation improving efficacy of intensive sampling efforts. onto the of approaches for Returned samples were identified by GMU of harvest. Usable samples were centrifuged, and sera were tested for antibodies to Brucella spp. using a standard card test. Unused sera were banked and stored at -20 C for future use. Data from this year's survey will be analyzed in combination with additional data to be gathered during FY92/93 to compare sampling strategies. We plan to meet again with program participants after those analyses are completed to discuss further modifications needed to continue improving our survey. Chronic Wasting Disease Survey: At least 5 cases of chronic wasting disease (CWO), a spongiform encephalopathy, have been confirmed in free-ranging deer and elk in Larimer County since February 1990. In "all, 12 cases have been diagnosed in free-ranging deer and elk in Colorado since 1985 -- all of these are from GMUs 9, 191, 19, or 20. Numerous cases of CWO have also been documented in captive wildlife in Colorado and Wyoming since the late 1970's. CWO has not been detected in free-ranging deer or elk outside of Colorado and Wyoming. The historic source of CWO is unknown. There is no evidence for transmission of CWO to either domestic animals or humans. At present, reliable estimates for distribution and prevalence of CWO in wild cervids are lacking. It follows that methodical surveillance is needed to provide data as a basis for any potential management action. In order to assess the extent of this potential problem and obtain reliable estimates of prevalence to use in making management decisions, we will conduct a series of surveys for CWO on select deer and elk populations throughout Colorado over the next 3-5 years. We initially planned to collect fresh brain tissue from hunter-killed mule deer and elk harvested from 2 geographically distinct populations (DAUs D4, 010, D31, E9 and E33) over a 2-3-year period; CWO has previously been confirmed in DAUs D4, D10, and E9. This survey was designed such that the probability of failure to detect at least 1 case of CWO in apparentlyunaffected mule deer and elk population will be ~0.1 even if herd prevalence �147 is 1\. This sampling effort should also result in an estimated prevalence within ±5\ at a 95\ level of confidence for affected populations. All brains will be submitted for histological lesions suggestive of spongiform encephalopathy. In addition to these formal surveys, we increased surveillance efforts by field personnel statewide to encourage submission of carcasses from deer or elk showing clinical signs resembling cwo. We collected brains from hunter-killed elk during December and/or January late seasons near Estes Park, Colorado (DAU E9), and from hunter killed mule deer and elk during special public hunts on the Forbes Trinchera Ranch near Ft. Garland in December (DAU D31/E33). Brains from hunter harvested mule deer and elk were collected within 12 hours of death and fixed in 10\ buffered formalin contained in 4 L plastic bags for at least 3 months. sections of medulla at the obex and frontal portion of the brain including basal ganglia, olfactory cortex and tract, and some frontal cortex were processed routinely for paraffin embedment. Histologic sections were cut at 5-6 pm, stained with hematoxylin and eosin, and examined under a light microscope. Lack of antlerless licenses and unavailability of regional field personnel precluded sampling deer populations along the northern Front Range (DAUs D4 and 010) for chronic wasting disease during the Fall 1991 hunting seasons. However, field personnel have agreed to collect suspect deer during the next year to help improve understanding of the prevalence and distribution of this problem. Disease Investigations Bovine Tuberculosis Investigations: Bovine tuberculosis was diagnosed in captive elk residing on a game ranch near Powderhorn, CO in June 1991. Although the infected herd was destroyed within a few weeks of its detection, the apparent severity and duration of this '~utbreak warranted investigating the possibility that tuberculosis might have spread to freeranging wildlife outside the infected premises. We followed up on last June's depopulation of a tuberculosis-infected captive elk herd near Powderhorn with collections of free-ranging deer and elk in the vicinity of the affected premises. Carcasses were necropsied and examined for gross lesions suggestive of tuberculosis. In addition, samples of parotid, mandibular, retropharyngeal, mediastinal, tracheobronchial, hepatic, and mesenteric lymph nodes were preserved in 10% buffered formalin and sodium borate and submitted to the National Veterinary Services Laboratory in Ames, lA, for histologic examination and culturing. We also established a hunter check station at Blue Mesa Reservoir to examine hunter-killed mule deer and elk harvested in the Powderhorn vicinity (DAU E25). Eviscerated carcasses were examined for gross lesions suggestive of tuberculosis, and samples of parotid, mandibular, and retropharyngeal lymph nodes were preserved in 10\ buffered formalin and sodium borate and submitted to the National Veterinary services Laboratory in Ames, lA, for histologic examination and culturing. Experimental Approaches In addition to disease surveillance and diagnostics, we plan to use controlled research experiments to develop and/or improve techniques for investigating wildlife disease problems. No original projects were initiated during FY 91/92. However, we cooperated with USDA/APHIS/VS and the Department of Environmental Health at Colorado state University in evaluating the reliability of an enzyme-linked immunosorbent assay (ELISA) for detecting bovine tuberculosis infections in elk. Sera from wild hunterkilled elk harvested on the Forbes-Trinchera Ranch and from captive elk slaughtered during depopulation of the infected game ranch near Powderhorn �148 were examined. Results of these Thesis currently in preparation. RESULTS Disease evaluations will be described in a Master's AND DISCUSSION Surveillance Statewide Surveillance At least 21 wildlife cases were submitted for diagnostic examination during July-September (Table 1); each region contributed at least 1 case. Most cases appeared to represent isolated incidents of trauma or disease. Of these, possible mercury intoxication in Front Range raptors and pneumonia in a bighorn from Glenwood Canyon (the second from the Grizzly Creek transplant) may warrant further investigation. We established a computerized database for recording information, and will continue adding new accessions fiscal year. In a~dition, we plan to begin entering reports next year. diagnostic case throughout the coming data from archived Surveys Brucellosis Survey: Of 9,595 hunters surveyed, 1,105 (12%) returned blood samples from antler less elk harvested throughout Colorado during OctoberDecember 1991 for brucellosis screening. Only 461 (42%) of returned samples were usable; marked hemolysis and/or contamination precluded evaluation of the remaining 644 samples. All sera tested were negative for antibodies to Brucella spp. on the standard card test. Overall, about 5% of the survey kits distributed to hunters provided samples usable in this year's brucellosis survey; this is about half of the expected 10% overall annual yield used in survey planning. The 12% sample return rate was somewhat lower than anticipated (about 33% for elk), but may reflect relatively low success in harvesting antlerless elk during regular rifle seasons in several parts of Colorado during 1991. However, nearly half of all returned samples yielded usable sera -- this trend, if sustained, represents a small but notable improvement in the proportion of usable samples returned (also about 33% for elk). These data, combined with those collected during FY 92/93, should provide a quantitative basis for assessing strategies for improving the efficiency of statewide serologic surveys that depend on blood samples submitted from harvested animals. Chronic Wasting Disease Survey: Brains from 51 harvested mule deer and 33 harvested elk from the Forbes Trinchera Ranch (DAU D31/E33) and 4 brains from harvested elk near Estes Park (DAU E9) were examined for evidence of chronic wasting disease. Details of microscopic findings were filed with completed reports and summarized elsewhere. All 88 brains examined were negative for spongiform encephalopathy. Fifteen (29%) mule deer and 15 (45%) elk from Forbes Ranch showed mild to moderate encephalitis/meningoencephalitis composed primarily of perivascular lymphocytic cuffs. A few brains had focal gliosis, and in one elk plasma cells were admixed with lymphocytes suggesting continuing antigenic stimulation. The cause of these nonsuppurative lesions was not apparent from these sections, but it is unlikely that they were clinically important. Past viral infection or reaction to circulating trypanosomes were deemed the most likely probable causes for the lesions observed. Additions tests to determine the cause of these lesions will be pursued in cases where archived samples (serum, tissues) from affected individuals are available for additional testing. �149 Disease Investigations Bovine Tuberculosis Investigations: Ten mule deer does and 1 cow elk were collected on 19 August 1991. All of these animals appeared healthy, and we did not observe lesions suggestive of tuberculosis in any of them; histologic examinations and mycobacterial culture also failed to demonstrate tuberculosis infections. Similarly, tuberculous lesions were not seen in hunter-killed deer or elk from GMUs near Powderhorn. Findings made during the investigation and depopulation of the infected game ranch herd were summarized in 2 manuscripts submitted for publication. Initial examinations of wild deer and elk collected in the immediate vicinity of the tuberculosis-infected game ranch near Powderhorn have revealed no indication of infection in free-ranging animals. However, sample sizes to date are small, and the probability of failing to detect infection is relatively high (e.g., 0.778 if prevalence is 1%). Consequently, during FY92-93 we plan to develop and implement a more structured pilot program for monitoring wildlife populations in the Powderhorn area. This program will be used to continue tuberculosis surveillance in this area over the next 5 years, and may also be used in other locations where tuberculosis or other serious disease problems are diagnosed on game ranches. �150 Table 1. Summary of wildlife diagnostic cases submitted during July 1991 June 1992. REGION SPECIES DIAGNOSIS ACCESSION # NE Elk Hybrid Trauma-induced fractures NE Mule Deer Malignant catarrhal fever NE Mule Deer Peritonitis 912-5251 NE Silverhaired Bat Bite wounds 912-5716 NE Mallard (2) Botulism 912-2143 NE Mallard Botulism 912-2821 NE Osprey Drowning 912-4785 NE Golden Eagle Fractured rib; mercury poisoning 912-5395 CE Red Tail Hawk Possible mercury poisoning 912-5179 CE Mountain Lion Gunshot; posterior paralysis 912-5178 CE Mountain Lion Undetermined 912-5177 CE Mountain Lion Pulmonary hemorrhage 912-4852 CE Mountain Lion Possible carbamate poisoning 912-6218 CE Beaver Undetermined 912-6206 CE Beaver Tularemia 912-7023 SE Elk Possible Se toxicity 912-2358 SW-cwp Elk Hyperthermia 912-6626 SW-CWP Elk Trauma 912-6627 SW-CWP Elk Trauma 912-6628 NW Mule Deer Demodectic mange 912W1060 NW Bighorn Sheep Suppurative bronchopneumonia 912W1062 912-2660 �151 Colorado Division of Wildlife Wildlife Research Report July 1992 JOB PROGRESS REPORT State of ~C~o~l~o~r~a~d~o~ _ Project No. W-153-R-4 Mammals Researeb Work Plan No. 2A Mountain Sheep Investigations Job No. 4 Experiments to Identify and Manage Stress in Mountain Sheep Populations Period Covered: July I, 1991 - June 30, 1992 Author.: K. W. Killer, K. P. Snipes, T. R. Ritchie, B. J. Kraabel, and K. W. Mills. Personnel: E. S. Williams, A. Boeger-Fields, and T. R. Spraker. ABSTRACT We continued examining archived Pasteurella haemolytica isolates from wild bighorns representing 5 indigenous herds; no additional isolates were collected from target herds during 1991-1992. Using ribosomal RNA gene restriction patterns, at least 11 distinct genotypes (A, B, I, J, K, L, M, N, 0, P, and T) of P. haemolytica were identified among isolates (n- 17) from 3 herds (Avalanche Creek, Tarryall, Taylor River) analyzed to date. Ribotypes varied both within and among the 5 distinct phenotypes (T3; 4; 3,4; 3,4,10; and untypable) of P. haemolytica identified among these 3 herds using rapid plate agglutination. In contrast to phenotypes, ribotypes of P. h~emolytica appeared unique for each of these geographically isolated herds. Sera have been scre~ned for antibodies to a battery of P. haemolytica strains using an enzyme-linked immunosorbent assay (ELISA), but those data have n9t been analyzed. We conducted a pilot experiment to evaluate antibody responses of adult bighorns to a commercially available P. haemolytica vaccine (Presponse·). Ten captive adult sheep were paired by age and sex, randomly assigned to vaccinated or saline-injected control groups, injected on days 0 and 14, and bled on days 0, 5, 10, 14, 17, 20, 25, 30, 40, 50, and 60. Antibody levels were measured using a modified ELISA. We detect no differences (P > 0.05) in antibody levels between vaccinated and control groups before or after vaccination. Attempts to modify the ELISA to include vaccine-specific antigens failed. Based on available data, we were unable to determine whether vaccinated bighorns did not produce antibody to Presponse. vaccine or whether those responses were simply unmeasurable using the unmodified ELISA. Plans �152 for further laboratory and field experiments to evaluate the Presponse· vaccine in bighorn sheep are on hold pending further evaluation of data from this pilot study. Preliminary evaluation of additional data from ongoing monthly collections of feces from captive bighorns supported previous observations that reproductive status may influence fecal cortisol excretion in bighorn ewes; by May. fecal cortisol concentrations from pregnant ewes (mean±SE-39.S±6.9 ng/g dm) were about 77% higher than those from open ewes (22.2±3.0 ng/g dm)(P<O.OS). Evaluation of seasonal influences on fecal cortisol measurements will continue through September 1992 to provide 2 complete years of data. �153 EXPERIMENTS TO IDENTIFY AND MANAGE STRESS IN MOUNTAIN SHEEP POPULATIONS M. W. Miller et al. P. N. OBJECTIVE To treat bighorn sheep to control disease where necessary. SEGMENT OBJECTIVES 1. Develop a research strategy and proposal for managing bacterial and viral diseases in mountain sheep populations; begin conducting approved and funded research. 2. Design a population-level experiment to evaluate techniques for detecting stress in free-ranging bighorn sheep; begin conducting approved and funded research. MANAGEMENT OF BACTERIAL AND VIRAL DISEASES IN MOUNTAIN SHEEP POPULATIONS Inability to control infectious disease outbreaks and subsequent mortality in mountain sheep populations represents a significant obstacle to long-term success in their management. Although the "bighorn pneumonia complex" has been studied intensively for over 3 decades, little is known about many aspects of its etiology and epizootiology. Moreover, management interventions recommended for preventing or controlling this problem remain untested. Most previous efforts to improve understanding and management of the epizootiology of pneumonia in bighorns involved post hoc investigations of dieoffs occurring in free-ranging sheep herds. These studies identified various etiological agents associated with known mortalities and attempted to determine predisposing causes and population consequences of individual outbreaks. From these investigations, comparisons of real or perceived patterns became the basis for hypotheses on the epizootiology of pneumonia in bighorns. Recognition of similar patterns in other outbreaks served as evidence supporting these as unifying hypotheses. Unfortunately, several of these hypotheses have failed to withstand rigorous experimental testing (Miller 1988, Miller et al. 1990, 1991). And, despite our best management efforts, bighorns continue to die. Our strategy for developing a better understanding of the epizootiology and management of bacterial and viral diseases in bighorn populations differs -generally, we propose to take an adaptive environmental assessment approach for studying the bighorn pneumonia complex. As a foundation for our research strategy, we have initially attempted to assimilate existing knowledge on bighorn population dynamics (including the epizootiology and consequences of infectious disease) into computer simulation models (Hobbs et al. 1990, Hobbs and Miller 1991). Because pasteurellosis appears to underlie virtually all �154 respiratory disease problems reported for bighorns, our modeling efforts have focused on the epizootiology of pasteurellosis in sheep populations. We have constructed models that reflect dynamics of bighorn populations seen in nature using the simplest assumptions necessary to reproduce those behaviors. Once we have a reliable working model, we plan to conduct simulation experiments to identify variables that might be particularly sensitive to management perturbations in altering the dynamics of disease in.bighorn populations. Those results will serve as the basis for designing management level experiments in the future. In parallel with our modeling efforts, we plan to conduct a series of experiments to develop, improve and standardize methods for collecting and interpreting diagnostic data to provide better estimates of key parameters driving our models. In particular, we have been developing tools for identifying strains of Pasteurella haemolytica and quantifying immunological responses of bighorns to infection by these pathogens. These tools will be key components of laboratory and field experiments designed to evaluate potential tactics (including vaccination and/or treatment) for managing pasteurellosis in wild sheep, and appear prerequisite to initiating management level experiments. To this end, our recent efforts have focused on simulation modeling and on improving tools available for use in future management experiments that will be designed to study etiology, epizootiology, and prevention or control of disease outbreaks in bighorn popUlations: METHODS AND MATERIALS In conjunction with numerous cooperators, we summarized data gathered to date from a series of cooperative studies conducted in Colorado and elsewhere to develop a simulation model (see details in WP2aJ8 Progress Report) and serologic and bacteriologic tools for use in studying and managing pasteurellosis in free-ranging bighorn popUlations; specific methods used in these projects have been described in detail elsewhere. Based on preliminary findings generated from these efforts, we prepared a Program Narrative (Appendix A) and a Study Proposal (Appendix B) outlining additional experiments designed to provide data and to further develop diagnostic tools for eventual application in statewide bighorn management programs. Both plans were peer-reviewed and studies are in progress; one of these will be funded largely through Colorado Division of Wildlife Special Bighorn Sheep and Mountain Goat Auction and Raffle Funds available in FY91-92. Epizootiolo~ of pasteurellosis in Colorado's indisenous Rocky Mountain biShom sheep populations: We collected blood samples and nasal and pharyngeal swabs from indigenous bighorn herds trapped in January-Karch 1991. Methods used to collect and handle field samples were as described in Appendix A, except that nasal swabs from 3 herds (Avalanche Creek, Tarryall, Chalk Creek) were transported in modified Amies medium rather than Port-a-cul· tubes. Laboratory methods were as described in Appendix A. In a related activity, we also investigated an all-age pneumonia epizootic that occurred in the Taylor River-Almont Triangle bighorn herd. We performed postmortem examinations on 5 sheep carcasses and collected representative samples for diagnostic evaluation using established laboratory techniques. We �155 also assisted in field operations to assess the extent and severity of this epizootic. Immunity to pasteurellosis in Rocky Mountain bighorn sheep; assessing specificity of antiboQy responses to Pasteurella haemolytica: We collected sera from free-ranging bighorns in conjunction with the preceding study. Preparation of antigens from distinct phenotypic strains of P. haemolytica is currently underway, and serologic assays will be completed in FY91-92 once funding becomes available. Experimental Evaluation of Antibody Responses to a P. haemplytica Toxoid in Bi&horn Sheep. (Miller and Kraabel) We initiated a pilot study to evaluate antibody responses of adult bighorns to a commercially available Pasteurella haemolytica vaccine (Presponse·). Ten adult sheep (4 rams, I castrated ram, and 5 ewes) were paired by age and sex. We randomly assigned 1 bighorn from each pair to the vaccinated group, and the other to the saline-injected control group. All sheep were injected on day 0, and will be injected again on day 14. Each bighorn was bled prior to injection on day 0, and will be bled on day 5, 10, 14, 17, 20, 25, 30, 40, 50, and 60. Anitbody levels will be measured using a modified ELISA developed by Ken Mills at WSVL. . We were unable to detect serologic responses of adult bighorns to a commercially available Pasteurella haemolytica vaccine (Presponse.) using an enzyme-linked immunosorbent assay described previously. Coworkers at WSVL are currently modifying the ELISA to include vaccine-specific antigens, and will reexamine serum samples for antibody responses. Plans for further laboratory and field experiments to evaluate the Presponse· vaccine in bighorn sheep are on hold pending completion of this pilot study. Results will ultimately be used in planning additional laboratory and field experiments to evaluate vaccination as an approach for managing pasteurellosis in·bighorn sheep. EXPERIMENTS ON MEASURING AND MANAGING STRESS IN BIGHORNS (Miller and Ritchie) In planning field studies of stress responses in wild bighorns, it has become apparent that further development and understanding of fecal cortisol measurements is necessary in order to credibly apply these techniques to freeranging populations. In order to anticipate potential confounding influences on field applications of fecal cortisol measurements, we collected feces from captive bighorns at monthly intervals to examine seasonal influences on cortisol excretion. Samples were collected over a 3-5 day period at the beginning of each month from individual captive bighorns of varied age/sex classes. Feces were stored at -20 C until processed for cortisol determination. Methods for processing feces and measuring cortisol were as described by Miller et a1. (1991). �156 We continued collecting feces from captive bighorns at monthly intervals to examine seasonal influences on cortisol excretion. Samples collected to date since May are in prep and will be submitted for cortisol assay next quarter. Further development and understanding of fecal cortisol measurements are necessary in order to credibly apply these techniques to field studies of stress in bighorn sheep. We continued collecting feces from captive bighorns at monthly intervals to examine seasonal influences on cortisol excretion. Preliminary examination of data for samples collected from October 1990 to December 1991 date suggested significant interassay variation may be influencing fecal cortisol data. We are currently attempting to identify and correct the source of this variability. Monthly sample collection will continue. RESULTS AND DISCUSSION The following abstracts summarize significant findings from a series of cooperative studies conducted in Colorado and elsewhere to develop tools for use in studying and managing pasteurellosis in free-ranging bighorn populations: REGULATION OF POPULATIONS BY INFECTIOUS DISEASE: SIMULATIONS OF PASTEURELLOSIS OUTBREAKS IN BIGHORN SHEEP (OVIS CANADENSIS). N. Thompson Hobbs and Michael W. Miller. Abatract: Many popUlations of bighorn sheep (Ovis canadensis) appear to be limited by their interactions with pathogens, notably Pasteurella spp., rather than by food supplies or predators. We developed a simulation model to represent interactions between Pasteurella spp. and bighorn sheep populations. The model is an adaptation of the classical SIR system of coupled differential equations that has been widely applied to human diseases. Our model differs, however, in that its compartments (susceptible, infected, recovered) are represented as states within individual animals. Flows among compartments are translated into probabilities of transitions among states. We used this model to examine the idea that prevalent patterns in bighorn population dynamics can be explained by intrinsic features of host-pathogen interactions. Averaged over several simulated populations, our individual-based model predicted an equilibrium virtually identical to that of the SIR model. However, because our model is inherently stochastic, single populations showed a broad range of dynamical behavior including stasis, exponential growth, and episodic crashes in abundance. Thus, our model is faithful to two overriding patterns in bighorn population dynamics: regional suppression of animal numbers, and nonequilibrium behavior at local scales. We argue that both of these patterns can be explained by internal properties of host-pathogen interactions. External explanations, including habitat degradation and fragmentation, environmental stress, parasitism, and contact with domestic animals are not necessary to explain epidemics and the ensuing population cycles frequently observed in bighorns. �157 EPIZOOTIOLOGY OF PASTEURELLOSIS IN CAPTIVE ROCKY MOUNTAIN BIGHORN (OVIS CANADENSIS CANADENSIS) LAMBS. Michael W. Miller, Margaret A. Wild, Kenneth W. Mills, Elizabeth S. Williams, and Amy Boerger-Fields. Abstract: Poor lamb survival is often observed in Rocky Mountain bighorn sheep (Ovis canadensis canadensis) populations, particularly in the years following pneumonia epizootics. Causes for this reduced lamb recruitment are incompletely understood. Following an outbreak of pasteurellosis in a captive bighorn herd in 1984, 19 of 29 lambs born into that herd between 1985 and 1990 developed bronchopneumonia before 3 mo of age. We attributed all cases of pneumonia to nonhemolytic Pasteurella haemolytica infections uncomplicated by protostrongylosis. Infection with P. haemolytica occurred shortly after birth. Three of 6 ewes sampled annually between 1988 and 1991 consistently shed Pasteurella spp. in nasal secretions at parturition. Nonhemolytic P. haemolytica was not isolated from nasal or pharyngeal swabs of newborn lambs ($24 hrs old; n-22) , but was recovered from both sites in lambs ~ 5 days old. Lamb-to-lamb transmission also appeared to contribute to epizootic spread of Pasteurella spp. Overall, 30 of 30 lambs sampled over a 4-yr period were infected by mid-summer. Biotype T, serotype 4, and untypable catalase positive isolates of nonhemolytic P. haemolytica were predominantly recovered from both sick and healthy lambs. An enzyme-linked immunosorbent assay measuring levels of antibodies to nonhemolytic P. haemolytica, biotype T, serotype 4, in colostrum from ewes and serum from lambs revealed the half-life of passive antibody levels in bighorn lambs was about 3 wks. Passive antibodies may· have afforded lambs temporary protection from disease; pneumonia tended to occur in lambs >3 wks old. Our observations suggest that neonatal pasteurellosis may have profound effects on health and survival of bighorn lambs. AN ENlYME-LINKED IMMUNOSORBENT ASSAY FOR DETECTING ANTIBODIES TO PASTEURELLA HAEHOLYTICA IN BIGHORN SHEEP (OVIS CANADENSIS). Ken Mills, Amy Boerger-Fields, Michael Miller, Margaret Wild, and Beth Williams. Abstract: Periodic pasteurellosis outbreaks cause extensive mortality in bighorn sheep (Ovis canadensis) populations, and may limit bighorn abundance throughout North America. It follows that tools for detecting presence of or exposure to Pasteurella spp. infections could enhance the efficacy of management efforts for bighorns. We developed an enzymelinked immunosorbent assay (ELISA) to detect antibody to P. haemolytica in bighorn sheep. Antigen was initially prepared from a 20-hr culture of nonhemolytic P. haemolytica, biotype T, serotype 4, isolated from a captive bighorn. Blocking assays using the original antigen strain, other P. haemolytica serotypes, and other bacterial species demonstrated high typespecificity for this ELISA. Based on these results, we developed a multivalent ELISA measuring antibody levels to distinct strains of P. haemolytica isolated from bighorn sheep or domestic livestock. Antibody. levels measured with this multivalent ELISA appeared to differ between two captive bighorn herds differing in histories of exposure to P. haemolytica. Our results suggest measuring antibodies to P. haemolytica by ELISA may have applications in studying the epizootiology of pasteurellosis in bighorn populations, as well as in guiding relocation and other bighorn management activities. �158 USING RIBOSOMAL RNA GENE RESTRICTION PATTERNS IN DISTINGUISHING STRAINS OF PASTEURELLA HAEHOLYTICA FROM BIGHORN SHEEP (OVIS CANADENSIS). K. P. Snipes, K. P., R. W. Kasten, M. A. Wild, M. W. Miller, D. A. Jessup, R. L. Silflow, and T. E. Carpenter. Abstract: Pasteurella haemolytica isolates (n - 31) from two captive herds of Rocky Mountain bighorn sheep (Ovis canadensis canadensis) were characterized using established phenotyping and newer genomic fingerprinting methods. Phenotypes of isolates varied both within and between source herds. Nine different phenotypes were identified; biotype T and serotypes 4 and 3,4,10 predominated. Most isolates (23 of 31) were nonhemolytic, and all hemolytic isolates were from one herd. Three phenotypes (nonhemolytic T4; T3,4; and T3,4,10) were common to both herds. Evaluation of 14 restriction endonucleases revealed that EcoRI, HincII, and ~I yielded optimal DNA fragmentation patterns for ribotyping P. haemolytica isolates using a 32P-labeled Escherichia coli rRNA probe. Ribotype pattern appeared to be a relatively stable trait: an identical ribotype was conserved within and across 19 passages of a stock P. haemolytica strain. Among the 31 bighorn isolates, ribotyping produced six subjectively distinct patterns; probing after digestion with EcoRI, HincII, or ~I consistently assigned isolates to respective ribotype groups. In contrast to phenotypes, ribotypes appeared unique to each herd. Comparing phenotypic and genotypic traits among isolates revealed associations of multiple serotypes with a single ribotype and multiple ribotypes with a single serotype, as well as presence and absence of hemolysis among otherwise identical isolates. These findings suggest ribotyping may be a useful adjunct to other bacteriology methods in studying the epizootiology of pasteurellosis in bighorn sheep. EXPERIMENTAL AND FIELD EVALUATION OF PORT-A-CUU8 TRANSPORT TUBES FOR RECOVERY OF PASTEURELLA HAEHOLYTICA FROM BIGHORN SHEEP (OVIS CANADENSIS) PHARYNGEAL SWABS. Margaret A. Wild, 'Michael W. Miller, Terry R. Spraker, and William J. Adrian. Abatract: Three pharyngeal swabs were collected from each of 25 healthy captive bighorn sheep (Ovis canadensis). We recovered nonhemolytic Pasteurella haemolytica from 23 of 25 (92%) swabs streaked onto blood agar plates and incubated immediately, from 16 of 25 (64%) swabs held in Port-ACuI. transport tubes for 24 hr, and from 1 of 25 (4%) swabs held in Port-ACuI. tubes for 48 hr. Although the recovery rate from swabs held in PortA-CuI. tubes for 24 hr was only about 70% of that from direct swabs, rates were markedly higher than those for other transport media. We subsequently used Port-A-Cul. tubes to transport pharyngeal swabs from healthy, freeranging bighorns trapped throughout Colorado in winter 1990-91. Nonhemolytic P. haemolytica was isolated from 7 of 8 herds sampled; within those herds, isolation rates ranged from 17-87% of individuals sampled. Isolation rates for P. haemolytica exceeded rates in previous years when pharyngeal (1989-90) or nasal (1988-89) swabs were transported in modified Amies with charcoal, suggesting our previous sampling efforts may have underestimated prevalence of P. haemolytica infections in free-ranging herds. Using Port-a-cul. tubes and minimizing time between sample �159 collection and processing (~24 hr) appears to be a practical way to optimize recovery of P. haemolytica from bighorn pharyngeal swabs. Complete manuscripts describing these studies and their results are in prep (Miller et al., Mills et al., Wild et al.) or are in review (Hobbs and Miller, Snipes et al.) for publication in peer-reviewed professional journals or proceedings. Epizootiolo&y of pasteurellosis in Colorado's indieenous Rocky Mountain biehorn sheep populations: One hundred forty-two wild bighorns from 5 indigenous herds were sampled in conjunction with capture operations during January-March 1991. We examined Pasteurella spp. isolates (n-5l) from 4 of 5 bighorn herds sampled (Taylor River, Avalanche Creek, Tarryall Mountains, Chalk Creek); failure to recover Pasteurella spp. from 19 samples collected at Cottonwood Creek was likely due to delayed laboratory processing. Prevalence estimates based on cultures of pharyngeal swabs ranged from 17-83% (mean 49%) among sampled herds (Fig. 1). Based on serotyping data from isolates we examined, there appeared to be differences in strains of P. haemolytica endemic to these herds. In all, we encountered at least 6 distinct phenotypes (T3; 4; 3,4; 3,4,10; A7; and untypable) among the 4 herds. We recovered T3 isolates from all of these, but prevalence of this strain appeared to vary among herds (Fig. 1). Representative isolates from each herd have been submitted for genomic fingerprinting, and sera will be submitted for ELISA to examine exposure to a battery of P. haemolytica strains. We plan to use comparisons of genotypic and phenotypic traits of Pasteurella spp. among indigenous herds, along with serological evidence of exposure, to improve our understanding of the epizootiology and management of pasteurellosis in Colorado'S bighorn populations. The pneumonia epizootic in the Taylor River-Almont Triangle population occurred in the face of one of the more intensive bighorn herd management programs in Colorado; that program includes ongoing lungworm treatment and population control programs, as well as extensive habitat improvement and protection efforts. Coughing rams were first observed in mid-March; further field observations revealed signs of respiratory disease in bighorns of all age/sex classes scattered throughout the winter range. To date, about 17 bighorns are known to have died; necropsies of 5 dead sheep (2 rams, 3 ewes) revealed subacute to chronic bronchopneumonia. Nonhemolytic P. haemolytica was isolated from 4 of these (the fifth was too decomposed to attempt bacteriology). At least 4 distinct serotypes (T3; 4; 10; and 3,4,10) were identified; 3 of these were from lungs of different individuals. Horaxella spp. and Actinomyces pyogenes were also isolated from some affected individuals. Immunoperoxidase staining of lung tissue revealed presence of bovine respiratory syncytial virus in I ewe; other viral or chlamydial agents have not been detected to date. Based on gross, histologic, and fecal examinations, lungworm did not appear to playa role in this epizootic. The duration of this outbreak prior to its detection remains unknown: although prevalence of Pasteurella spp. was lowest among herds sampled this year, at least I yearling ram transplanted from the Taylor River-Almont Triangle herd to Glenwood Canyon in January died after developing pasteurellosis in late March; the fate of 39 other sheep removed from the Taylor River-Almont Triangle area remains unknown. We will continue working with Southwest Region Personnel to study this outbreak and its aftermath. �160 EXPERIMENTS ON MEASURING AND MANAGING STRESS IN BIGHORNS Preliminary evaluation of data from monthly collections of feces from captive bighorns suggested reproductive status may influence fecal cortisol excretion in bighorn ewes (Fig. 2); by May, fecal cortisol concentrations from pregnant ewes (mean±SE-39.5±6.9 ng/g dm) were about 77% higher than those from open ewes (22.2±3.0 ng/g dm)(P<0.05). Based on these findings, further evaluation of seasonal and other influences on fecal cortisol measurements is warranted before these techniques are applied to studies of stress responses in freeranging bighorns. LITERATURE CITED Miller, M. W. 1988. Experiments toward detecting and managing stress in Rocky Mountain bighorn sheep (Ovis canadensis canadensis). Ph.D. Thesis, Colorado State University, Fort Collins, Colorado, 106 pp. Miller, M. W., N. T. Hobbs, and M. C. Sousa. 1991. Detecting stress responses in Rocky Mountain bighorn sheep (Ovis canadensis canadensis): reliability of cortisol concentrations in urine and feces. Can. J. Zool. 69: 15-24. Miller, M. W., N. T. Hobbs, and E. S. Williams. 1991. Spontaneous pasteurellosis in captive rocky mountain bighorn sheep (Ovis canadensis canadensis): clinical, laboratory, and epizootio1ogica1 observations. J. Wi1d1. Dis. 27: in press. Snipes, K. P., R. W. Kasten, M. A. Wild, M. W. Miller, D. A. Jessup, R. L. Si1f1ow, W. J. Foreyt, and T. E. Carpenter. 1991. Using ribosomal RNA gene restriction patterns in distinguishing isolates of Pasteurella haemolytica from bighorn sheep (Ovis canadensis). J. Wild1. Dis. 28: 347-354. 1 Wildlife Researcher C �161 Colorado Division of Wildlife Wildlife Research Report July 1992 JOB PROGRESS REPORT State of ~C~o~l~o~r~a~d~o~ _ Project No. W-lS3-R-4 Work Plan No. · Job No. Personnel: _ Mountain Sheep Investigations Experimental Evaluation of Mountain Sheep Transplanting and Disease Treatment Period Covered: Authors: ~2~A~ Mammals Research July I, 1991 - June 30, 1992 M. W. Miller and J. Vayhinger K. Alderman, C. Anderson, Jr., C. Anderson, Sr., J. Backstrand, R. Green, V. Jurgens, K. McLaughlin, S. Ogilvie, G. Roberts, S. Roush, A. Torres, T. Verry. Abstract We initiated a management experiment to evaluate and compare management strategies for reducing lungworm burdens and improving lamb survival in selected mountain sheep populations. Four mountain sheep herds, 2 in the Tarryall Mountains [Cotton Gordon's (CG) and Sugarloaf Mountain (SL)] and 2 in the Collegiate Peaks [Chalk Creek (CC) and Cottonwood Creek (CW)] , were each managed under 1 of 4 alternative lungworm treatment regimes: CG - baiting with alfalfa hay and apple pulp treated with fenbendazole (bait/treat); CC baiting with alfalfa hay and apple pulp without fenbendazole (bait/no treat); SL - placing fenbendazole-treated salt blocks on bait stations (no bait/treat); and CW - withholding bait and fenbendazole (no bait/no treat). Thirteen additional ewes (5 at CG, 4 at SL, 3 at CC, and 1 at CW) were c.ptured and radiocollared during February and March; another ewe darted at CW died during immobilization. Marked ewes in all 4 groups have appeared healthy. Nonetheless, to date 9 radiocollared ewes have died since they were captured and marked. One of these losses (at CG) was almost certainly capture-related; three more (1 each at CC, CW, and SL) may also have been indirectly related to capture, although these ewes all survived at least 2 months after being captured and radiocollared. The other 5 losses (3 at CG and 1 each at CW and SL) were likely not influenced by previous capture; 1 ewe lost in February 1992 at CG appeared to be a lion kill, and predation may have also been involved in 2 earlier losses in the Tarryalls (1 at CG and 1 at SL); causes of death for 2 other ewes (1 each at CG and CW) were inapparent. �162 In mid-May, we began comparing lamb production and survival among radiocollared ewes (n - 54) as a measure of management treatment efficacies. Numbers of new lambs observed peaked in the Tarryall herds in late May, about 2-3 weeks earlier than in the Collegiate herds. Lamb production, as measured by the number of lambs observed closely accompanying and/or nursing radiocollared ewes, appeared to vary somewhat among herds (lambs/ewes: CC 12/13; CG - 13/17; OW - 10/14; SL - 7/11). Overall, more lambs were observed in baited herds (25 lambs/30 ewes) than in unbaited herds (17 lambs/26 ewes); about an equal number of lambs were observed in anthelmintic-treated (20 lambs/28 ewes) and untreated (22 lambs/27 ewes) herds. All lambs observed with radiocollared ewes, regardless of treatment, survived to 30 June. Intensive monitoring of lamb survival will continue through October with emphasis on documenting survival of known lambs. Additional monitoring of radiocollared ewes and lambs will also be conducted on winter ranges during December-February in conjunction with experimental treatment periods. �163 EXPERIMENTAL EVALUATION OF MOUNTAIN SHEEP TRANSPLANTING AND DISEASE TREATMENT M. W. Miller and J. Vayhinger P. N. OBJECTIVES Design, conduct, and report on management experiments to evaluate efficacy of transplanting and disease treatment practices for managing mountain sheep popUlations. AGREEMENT OBJECTIVES Initiate a management level experiment evaluating Colorado's mountain sheep parasite control program. Colorado's mountain sheep management program is recognized as among the most aggressive in North America. A combination of management practices, including transplanting to establish new herds and anthelmintic treatment to improve herd health, appears to have produced a threefold increase in bighorn numbers statewide over the last 20 years. However, the contributions of individual management strategies to the overall success of this program cannot be discerned (Bailey 1990). Transplanting bighorns to unoccupied ranges and treating resident herds to control lungworms are 2 of the most intensive (and costly) practices used to manage sheep in Colorado, yet neither of these programs has ever been evaluated experimentally to assess long-term efficacy. Recent analyses of available herd data suggested that only about half of all transplant efforts succeed in establishing viable bighorn herds, and that parasite treatment does not necessarily prevent disease outbreaks or improve performance of treated bighorn herds (Bailey 1990). Unfortunately, management practices for most of Colorado's bighorn populations are so confounded that even interpretation of the aforementioned evaluation is equivocal. Clearly, management-level experiments are needed to evaluate and compare efficacy and efficiency of bighorn management practices in Colorado. The Colorado Division of Wildlife's Terrestrial Wildlife Research Section is committed to conducting management experiments to evaluate efficacy of transplanting and disease treatment practices for managing mountain sheep populations. In conjunction with the Division's Southeast Regional staff, we have proposed the following management experiment to evaluate and compare management strategies for reducing lungworm burdens and improving lamb survival in selected mountain sheep herds: MATERIALS AND METHODS We began capturing bighorn ewes in preparation for our management-level experiment evaluating Colorado's mountain sheep parasite control program. During February and March, 1991, we used dropnets and/or chemical �164 immobilization (about 2.6 mg carfentani1 HC1 and about 15 mg xylazine HCl/animal) to capture adult and yearling ewes at 4 sites in the Tarryall and Collegiate Mountains. Captured ewes were fitted with radiocollars (148.0149.9 MHz) marked with color/symbol combinations unique to each capture site to allow visual identification of individuals; ewes were also eartagged using color/number combinations unique to each capture site. Radiocollared ewes were monitored irregularly to detect mortality and movements between February and May. To aid in planning further captures and in refining design of the field monitoring phase of this management experiment, pretreatment data were gathered from May to present. Telemetered ewes were located and observed every 2-4 weeks. Location (by UTM coordinate) and presence or absence of a nursing lamb was recorded for each observation. All field data were transcribed into a computerized database to aid in mapping seasonal range movements and determining preliminary lamb survival rates. We continued pretreatment monitoring of bighorn movements and lamb survival in the Collegiate and Tarryall Mountains in preparation for initiating a field experiment to compare alternatives for lungworm treatment. Movement trends generally continued as described in our FY90-9l Progress Report. Lamb survival apparently remains high (>70%) in all 4 herds being monitored, although some coughing was observed among lambs in the Chalk Creek herd during late September. We continued experimental treatments of bighorn herds in the Tarryall and Collegiate Mountains. Baiting initiated in mid-December continued through 21 February at the Cotton Gordon (CG) and Chalk Creek (CC) sites; attendance was somewhat sporadic at both sites, but most ewes in both groups fed on bait sites throughout most of the baiting period. Sheep at the CG site were also treated with fenbendazole on 31 January and 6 February; all ewes in that herd were treated at least once. Sheep at the Sugarloaf (SL) site showed some use of fenbendazole-treated salt blocks during January-March. Blocks remained on the winter range until ewes moved to transitional ranges and/or lambing areas in mid-May. Marked ewes in the untreated Cottonwood (CW) herd showed greater mobility during the winter than ewes in other herds; most marked ewes spent at least some time on alpine winter ranges. This mobility may be at least partially attributable to absence of bait stations, although these sheep tend to be more mobile than those in other herds included in our study. Marked ewes in all 4 groups have appeared healthy. Nonetheless, to date 9 radiocollared ewes have died since they were captured and marked. One of these losses (at CG) was almost certainly capture-related; three more (1 each at CC, CW, and SL) may also have been indirectly related to capture, although these ewes all survived at least 2 months after being captured and radiocollared. The other 5 losses (3 at CG and 1 each at CW and SL) were likely not influenced by previous capture; 1 ewe lost in February 1992 at CG appeared to be a lion kill, and predation may have also been involved in 2 earlier losses in the Tarryalls (1 at CG and 1 at SL); causes of death for 2 other ewes (1 each at CG and CW) were inapparent. Thirteen additional ewes (5 at CG, 4 at SL, 3 at CC, and 1 at CW) were captured and radio collared during February and March; another ewe darted at CW died during immobilization. At present, all 14 adult ewes (~ 2 yrs) fed at CC and 17 of 18 adult ewes treated at CG are radiocollared. Thirteen SL and 14 �165 CW ewes are also radioco11ared, and further capture attempts will be made in both areas during early April (weather permitting!). Monitoring of lambing and lamb survival to evaluate the first year of experimental treatments will begin in early May. RESULTS AND DISCUSSION Fifty adult ewes were captured at 4 sites, 2 (Cotton Gordon's, Sugarloaf Mountain) in the Tarrya1l and 2 (Chalk Creek, Cottonwood Creek) in the Collegiate Mountains (Table 1). At least 2 ewes died as a direct result of capture, and 2 others have died since capture. Further attempts to collar ewes were thwarted by unfavorable weather that precluded helicopter flights and by impending lambing seasons. Additional capture attempts are planned for late summer and/or for mid-November after big game seasons close. Overall, 31 of 46 surviving ewes (67%) were observed either lactating (n-11) or nursing lambs (n-20) during May and/or June; 6 ewes showed no evidence of lambing, and the status of 9 others remains undetermined (Table 1). Only 1 known lamb loss, presumed a mortality, has been detected to date. Movements have varied widely both among and within monitored herds. A summary of field observations follows: COLLEGIATE - Chalk Creek Ewe P was found dead about 200-300 m northeast of the Chalk Creek bait site on 15 May; she may have fallen from a cliff, and appeared to have died >2 weeks earlier. All 10 remaining collared ewes in this area have been seen with lambs and appear to be lactating ("circle", "square", "triangle", 1, 2, 4, 5, 6, 7, and E), and 6 of these ("square", 2, 4, 5, 6, and 7) have been observed nursing lambs. Ewe 2's lamb apparently disappeared in late June -- she was observed on 10 July for 2.5 hours and was alone for the entire time. Six of these ewes ("square", "triangle", 1, 5, 6, and E) currently inhabit Cascade Canyon. The other 4 ("circle", 2, 4, and 7) seem to travel back and forth from the Brown's Creek area to private salt licks near the Love Ranch Meadow. - Cottonwood Creek Five of the 13 collared ewes have been seen with lambs and appear to be lactating ("square", 1, 4, 9, and E), and 3 of these ("square", 4, and 9) have been observed nursing lambs. Four ewes (2, 5, 6, and 7) are not lactating and do not appear to have lambs (ewe 7 showed teats but no udder development in June, suggesting she may have lost her lamb before monitoring began). The status of 4 ewes ("circle", "triangle", 8, and H) is unknown; 3 of these ("circle", "triangle", and 8) have been observed in ewes/lambs groups, but their individual status could not be determined. Ewe H has not been seen. �166 The sheep in this area are quite mobile and travel in large groups of >30 individuals. They have been moving back and forth between Mount Yale and a natural salt lick near Cottonwood Hot Springs. TARRYALL - Cotton Gordon's Nine of 15 collared ewes in this group have been seen with lambs and appear to be lactating ("circle", "square", "triangle", 1, 6, 7, K, U, and X); 5 of these ("square", 1, 6, U, and X) have been observed nursing lambs. None of these ewes appear to have lost lambs to date. The other 6 collared ewes in this area have not been seen with lambs (2, 4, 8, 9, M, and T). Ewes 8 and 9 are not lactating and do not appear to have lambs. Ewes M and T did not have lambs with them when seen on 5 June, and ewe 4 did not have a lamb with her when seen on 7 June, but lactation status for these ewes remains undetermined. Ewe 2 (a 2-yr-old) has been seen with other ewes and lambs, but has not been observed lactating or nursing a lamb. It appears that most of these ewes lambed on south facing slopes northeast of the Tarrya11 River and Carpenter Ranches (Hay Creek area) in early June. After lambing, they moved into the Goose Creek drainage (approx. 13 km northeast of Goose Creek Trailhead) in mid-June, then moved back into the Hay Creek area at the end of June. - Sugarloaf Mountain Ewe "square" was found dead on 22 May about 1.2 km northeast of the Sugarloaf bait site. She was found under a tree, had died without apparent struggle, and appeared to have died about 1 to 2 weeks earlier. Six of the 8 remaining collared ewes in this group appeared to be lactating and have been seen nursing lambs ("circle", "triangle", 1, 4, 5, and R). Ewe T did not have a lamb when observed on 2 June, but appeared to be lactating. The status of ewe 2 is unknown -- she was seen only once on 23 May. This ewe group was observed in the area between "X-rock" and Sugarloaf Mountain near the Tarrya11 River Road in late May and early June, and then. moved over McCurdy Mountain into an area around the intersection of Lost and McCurdy Creeks. They have remained in this area since early June. Although it appears to be a lambing area, only 1 lambing (ewe "circle") has been confirmed at this site -- the other ewes had already lambed when first observed in this area in late May. Intensive monitoring will continue through October with emphasis on documenting movements and survival of known lambs. Additional monitoring is also planned for December-February in conjunction with the second year's experimental treatments. Researcher �167 Table 1. Summary data for individual bighorn ewes captured for use in experimental evaluation of alternatives strastegies for lungworm treatment. - CHALK CREEK1 COLLEGIATE FllEQUENCY COLLAR • • LAMBING STATUS SEX AGE YEL 12 F 5+ Lactating . 13 F 5+ Nursed lamb . 14 F 5+ Lactating. EARTAG 149.040 BLUE .080 BLUE .100 BLUE .120 BLUE 1 2 F 4+ Lactating. .148 BLUE 2 3 F 4+ Nursed lamb; lamb lost . .180 BLUE 4 15 F 4+ Nursed lamb . .202 BLUE 5 5 F 4+ Nursed lamb. .220 BLUE 6 6 F 5+ Nursed lamb. .240 BLUE 7 16 F 6+ Nursed lamb. .318 BLUE E 1 F 4+ Lactating. 1 Blue/white collars; freq. 149.040 - 149.492; black/yellow eartags; trapped 28 February 1991 (7 E; 2 L; 2 R); additional ewes darted 11-20 March 1991. - COTTONWOOD CREEK2 COLLEGIATE FllEQUENCY COLLAR EARTAG SEX AGE LAMBING STATUS F 4+ Undetermined. 149.520 BlACK • .540 BlACK • 13 F 4+ Nursed lamb. .560 BlACK ~ 9 F 5+ Undetermined. .582 BlACK 1 14 F 3+ Lactating. .620 BlACK 2 20 F 3+ Not lactating; no lamb. .642 BlACK 4 10 F 4+ Nursed lamb. .660 BlACK 5 17 F 5 Not lactating; no lamb. .697 BlACK 6 15 F 2 Not lactating; no lamb. .722 BlACK 7 11 F 3 Not lactating; no lamb. .740 BlACK 8 6 F 5+ Undetermined. .782 BlACK 9 19 F 2+ Nursed lamb. .800 BlACK E 12 F 6 Lactating. .820 BlACK H 1 F 2 Undetermined; not seen. RED 8 2 Alack/yellow collars; freq. 148.520 - 148.980; whiteJred eartags; trapped 16 March 1991 (13 E; 7 L; 1 R). �168 TARRYALL FREQUENCY COLLAR 148.020 BIACK .040 BIACK .060 BIACK .080 • • - COTTON GORDON'S3 EARTAG YEL 43 SEX AGE LAMBING STATUS F 3.5 Lactating. 9 F 3.5+ Nursed lamb. ••• 10 F 5.5+ Lactating. BIACK 1 29 F 2.5 Nursed lamb. .100 BIACK 2 5 F 1.5 Undetermined. .120 BIACK 4 31 F 2.5 Undetermined. .160 BIACK 6 25 F 4.5 Nursed lamb. .180 BIACK 7 38 F 3.5+ Lactating. .200 BIACK 8 23 F 4.5 Not lactating; no lamb. .220 BIACK 9 12 F 1.5 Not lactating; no lamb . .302 BIACK K 6 F 3.5 Lactating. .340 BIACK M 3 F 6.5+ Undetermined. .418 BIACK T 48 F 3.5+ Undetermined. .440 BIACK U 32 F 3.5 Nursed lamb. .460 BIACK X 27 F 3.5 Nursed lamb. 3 Black/white collars; freq. 148.020-148.460; black/yellow February 1991 (16 E; 12 L; 6 R) eartags; trapped 21 TARRYALL - SUGARLOAF HOUNTAIN4 FREQUENCY COLLAR EARTAG SEX AGE OR 35 F 2+ Nursed lamb. LAMBING STATUS 149.530 BIACK .580 • BIACK ••• 38 F 5 Nursed lamb. .600 BIACK 1 6 F 5 Nursed lamb. .620 BIACK 2 7 F 3+ Undetermined. .642 BIACK 4 4 F 4 Nursed lamb. .660 BIACK 5 10 F 4 Nursed lamb. .880 BIACK R 49 F 5+ Nursed lamb. .940 BIACK T 28 F 2 Lactating. 4 Blackjblue collars; freq. 149.500-149.999; black/orange March 1991 (9 E; 13 L; 4 R). eartags; trapped 9 �169 Colorado Division of Wildlife Wildlife Research Report July 1991 JOB FINAL REPORT State of Project No. ~C~o~l~o~r~a~d~o _ W-153-R-4 Mammals Research Work Plan No. 2A Mountain Sheep Investigations Job No. 8 Statewide Mountain Sheep Management Plan Period Covered: July I, 1991 - June 30, 1992 Author: N. Thompson Hobbs, Michael W. Miller Personnel: D. Reed Abstract A survey of attitudes on management of mountain sheep populations revealed differences in opinion among the general public, mountain sheep hunters, and Division of Wildlife employees. The public at large tended to disapprove of hunting of mountain sheep, particularly hunting for trophies and hunting ewes. Hunting rams is strongly supported by mountain sheep hunters, but they expressed some disfavor for ewe hunting, even if they knew it would be inexpensive and would hep prevent disease outbreaks in bighorn populations. Division of Wildlife (DOW) employees held beliefs intermediate to those of the general public and mountain sheep hunters. Hunters and the general public expressed strong approval for current management programs for mountain sheep. However, in the case of the general public, it was clear that their approval was not based on knowledge of-DOW actions. We deve Lopad .a discrete-,time, stochastic _simulation model to represent intera~tiorts-between iasteure1l8 spp • .!i.ndmountainsheep ·(Ovis canadensis) populations. Our model is derived from a system of differential equations describing dynamics of susceptible, immune, and recovered (SIR) compartments in a population. Our model differs from the SIR approach, however, in that we represent compartments as states within individual animals and represent flows among compartments as probabilities of transitions among states. Results of simulations averaged over several populations showed Pasteurella spp. regulating bighorn numbers at a steady state approached via damped oscillations. This result closely resembled the outcome of the SIR model. However, because our model is stochastic, single populations showed a broad range of dynamical behavior including exponential growth, stasis, and episodic crashes in abundance. _Equilibrium never prevailed. We surmise from our simulations that fluctuations in bighorn abundance can be plausibly explained by internal properties of host-pathogen interactions. External forces, including habitat degradation and fragmentation, environmental stress, and contact with domestic animals are not necessary to produce periodic epizootic. ��171 STATEWIDE MOUNTAIN SHEEP MANAGEMENT PlAN N. Thompson Hobbs Michael W. Miller P. 1. Develop simulating N. OBJECTIVES a model of disease transmission in mountain alternatives for disease management. sheep populations 2. Develop a Statewide Mountain Sheep Management Plan incorporating modeling results, information from the scientific literature, and input from organizational units. 3. Present Management Report to organizational units and interested publics. METHODS AND MATERIALS We prepared a Management Analysis for Bighorn A copy of the draft document is attached. Prepared by N. Thompson Hobbs Wildlife Researcher C Sheep populations in Colorado. ��173 Colorado Division of Wildlife Wildlife Research Report July 1992 JOB PROGRESS REPORT State of ~C~o~l~o~r~a~d~o~ _ Project No. W-153-R-5 Mammals Research Work Plan No. 2A Mountain Sheep Investigations Job No. 9 Quantity and Quality of Mountain Sheep Habitat with Regard to Minimum Viable Populations and Response of Mountain Sheep to Human Activity Period Covered: July 1, 1991 - June 30, 1992 Author: D. F. Reed ABSTRACT Contract work was initiated with the University of Colorado, Colorado Springs, Geography and Environmental Studies Department, in mapping and evaluating mountain sheep habitats. This was done by using MIPS (Map and Image Processing System) for GIS (Geographic Information Systems) integration of mountain sheep distribution and movements, vegetation, viewshed, and water sources. Data collected on mountain sheep included group size, location, movements, marks (collars and/or eartags), and sex and age classification. Vegetation classification was done using an algorithm which was "trained" to recognize the statistical characteristics of each class in each band of satellite imagery. Field work was completed for Area A, Echo Canyon to Parkdale railroad siding, and a draft report was prepared. Field work continued for Area B, Nathrop through Brown's Canyon. Based on information from sheep sightings throughout 1991, it was hypothesized that the estimated 50-60 ewes north of the river in Area A had failed to breed in 2 successive mating seasons. In an attempt to solve this problem, three 2-year old rams were transplanted into the area. Two of the 3 rams initially socialized with the ewe groups north of the river, but by 28 and 38 days post-release, they apparently crossed the river and highway, and joined other ewes south of the river. Subsequently, the third ram was estimated to be south of the river as well. ��175 QUANTITY AND QUALITY OF MOUNTAIN SHEEP HABITAT WITH REGARD TO MINIMUM VIABLE POPULATIONS AND RESPONSE OF MOUNTAIN SHEEP TO HUMAN ACTIVITY Dale F. Reed P. N. OBJECTIVE Evaluate the quantity and quality of mountain sheep habitat with regard to minimum viable populations and test the response of mountain sheep to human activity. SEGMENT OBJECTIVES 1. Complete contracts between University of Colorado, Colorado Springs, and the Colorado Divsion of Wildlife for establishing a Geographical Information Sytem (GIS) data-base for Areas A and B . 2. Prepare a draft report for Area A. ACKNOWLEDGMENTS I thank co-principal investigators M. W. Miller. R. B. Gill, J. Vayhinger, and S. Ogilvie for their ideas and support. Division personnel D. Finch, W. Travnicek, J. Backstrand, and L. Spicer were helpful in collecting field data. Division personnel D. L. Schrupp and D. C. Lovell were helpful for their ideas and in coordinating GIS. BLM E. Brekke made field observations and provided important coordination. University of Colorado, Colorado Springs, T. P. Huber developed the GIS procedures and products. DESCRIPTION OF AREAS The study areas have been described by Reed (1991). METHODS AND MATERIALS Methods and materials are described in the appended report. RESULTS AND DISCUSSION Results and discussion for Area A are described in the appended report. Work for Area B is in progress. �176 LITERATURE CITED Reed, D. F. 1991. Quantity and quality of mountain sheep habitat with regard to minimum viable populations and response of mountain sheep to human activity. Colo. Div. of Wildl. GameRes. Rep. July:lS7-l60. ~ Prepared by $-:-~ r A4LJ7 ~~~ Dale F. Reed Wildlife Researcher �177 Draft 20Aug92 (w/o Figs) MOUNTAIN SHEEP HABITAT USE IN THE ARKANSAS RIVER CANYON - PART A: ECHO CANYON TO PARKDALE RAILROAD SIDING Dale F. Reed, Colorado Division of Wildlife, Researcher Center, 317 W. Prospect, Fort Collins, CO 80526 Jack Vayhinger, Colorado Division of Wildlife, 498 Old Wagon Trail, Woodland Park, CO 80863 Erik B. Brekke, Bureau of Land Management, Canon City District, Canon City, CO 81212 Thomas P. Huber, Department of Geography and Environmental Studies, University of Colorado, Colorado Springs, CO 80933-7150 The Arkansas River Canyon from west of Parkdale to Echo Canyon is part of the recently established Arkansas Headwaters Recreation Area (AHRA). A Cooperative Management Agreement was signed between the Bureau of Land Management (BLM) and the Colorado Division of Parks and Outdoor Recreation (DPOR) on 27 October 1989 to establish a partnership for management of recreation resources in the AHRA. The AHRA has experienced increased recreational use, and further increases in recreation and facility development are projected. One of the more publicly visible, and perhaps sensitive, resources in the canyon is mountain sheep (Ovis canadensis canadensis). As recreational use (white water boating in particular) increases and recreational sites are developed, their impact on mountain sheep remains unknown. Specifically, impacts of human disturbance on mountain sheep habitat use and behavior had become BLM and Colorado Division of Wildlife (CDOW) management concerns. This generated needs for information on 1) mountain sheep habitat use, 2) quantity and quality of mountain sheep habitats available, 3) possible habitat enhancement for mountain sheep, and 4) mountain sheep response to human disturbance. To address these needs BLM, Canon City District, committed to a mountain sheep habitat study and subsequently engaged the CDOW in a cooperative effort. Further, the CDOW engaged the Department of Geography and Environmental Studies, University of Colorado, Colorado Springs (UCCS) in implementing Geographical Information System (GIS) technologies to expedite and enhance data capture and display. The purpose of this report is to delineate mountain sheep habitat use, habitats available, and possible habitat enhancements in Area A located along the Arkansas River between Parkdale and Echo Canyon. �178 STUDY AREA Area A includes approximately 13 km (8 mi) of the Arkansas River Canyon from the Denver and Rio Grande Western railroad's Parkdale siding to the eastern side of Echo Canyon. The study emphasized the area north of the Arkansas River but included some aspects south of the river because of the proximity and the relatedness of events. Geological features of the canyon at this location include Precambrian gneiss and schist cut by black and white dikes and some injections of pink granite (Chronic 1980). Side drainages are numerous, often with steep sides, and occasionally have alluvial fans at their terminus. U.S. Highway SO, the Arkansas River, and the Denver and Rio Grande Western railroad generally parallel one another creating a relatively narrow corridor throughout the area. Common vegetation in the lower part of the canyon includes pinyon pine (Pinus edulis)-juniper (Juniperus spp.)(P-J); shrubs littleleaf mockorange (Philadelphus microphyllus), wafer-ash (Ptelea trifoliate), currants (Ribes spp.), and skunkbush (~ trilobata); grasses sand dropseed (Sporobolus cryptandrus), several Stipa and Boutelua spp.; and cholla and pricklypear cactus (Opuntia spp.). Mountain mahogany (Cercocarpus montanus) is often common about a third of the way up the slope north of the river. Gambels oak (Ouercus gambelii) is common at higher elevations north of the river. Climate is semi-arid and during winter snow melts off the southern exposed slopes quickly. METHODS Mountain Sheep Distribution and Movements Ground counts were made from the highway roughly 2-3 times per week during the summer of 1990. These were incorporated into the data eventhough the study had not begun. Bi-monthly counts were then made during 1991 except during the summer when again roughly 2-3 counts were made per week. Counts were made roughly 3-4 times per month January-April 1992. Searches were made from pulloffs along the highway with binoculars. All individuals were classified to sex and age. Group size, location, movements, marks (collars and/or eartags), interactions with boaters and other major stimuli, date, and time were recorded. Locations were plotted on 7.5 minute quad maps and universal transverse mercator (UTM) coordinates determined. Most of these data were entered into dBase IV and then into MIPS (Map and Image Processing System; MicroImages Corp., Lincoln, NE) for GIS integration. MIPS has raster, vector, or CAD file capabilities. Sightings, movements, escape terrain, lambing areas, and home range were done using vector capabilities. Home range was calculated using the minimum polygon method. The 3-dimension (3-D) was done using 3-D raster display function and digital elevation model (OEM; U.S.G.S. survey 3-arcsec 1:250,000) data. Calibration of images was done in UTM - Zone 13 which allowed superimposing of images. MIPS printouts were run on a HP Paint jet XL printer and a Lasergraphics personal film recorder. Vegetation Vegetation classification was done using a maximum likehood classification algorithm. This means that training sites were chosen to represent each vegetation class. The algorithm was "trained" to recognize the statistical characteristics of each vegetation class in each band of the satellite imagery �179 (Landsat TM scene YS120ll702lXO, 15 Jun 87). Training sites were done by conducting line-pace transects (vegetation sampled every other step at the tip of one's boot) through 100 m diameter circle plots (0-360 degrees, direction chosen randomly, n - 50 data points) to verifiy the predominant plant species for given vegetation classes. These were transferred to the computer system and the algorithm was programed to classify all other pixels in the study area according to the "trained" classes. The actual classification was done using the Gaussian curves and standard deviations of each class to assign each pixel to that class where it was statistically most likely to fit. Initially we used 12 classes, but later consolidated these to 5, namely, P-J, mountain shrub, grassland, riparian, and bare rock. Visibility Sheep habitat visibility (converse of percent vegetation and topographic obstruction) was sampled at most of the selected training sites for vegetation. Five directions (0-360 degrees) and 5 distances (10-50 m) were chosen randomly for each of the 100 m diameter. circle plots. The observer estimated by eye the percent of obstruction of aIm square white target at each of the selected directions and distances from the center of the circle plot (n - 5 data points). These percentages were then averaged for an overall percent visual obstruction for the given site. Viewshed Topography visible from the Arkansas River was estimated by calculating a viewshed (MIPS using DEMs, U.S.G.S. survey -arcsec 1:24,000 data). A 1 m height for sheep was assumed and readings were taken from the river elevation every 500 m. Water Ground searches for water were made up the numerous side drainages that fed into the Arkansas River during the summer when dry conditions were prevalent. RESULTS Mountain Sheep Distribution and Movements A total of 297 sheep sightings were made from May 1990 through March 1992. Most of these sightings (261 vs 36) were made north of the river because 1) most the searching was done north of the river, and 2) north of the river was more visible from U.S. Highway 50. Additionally, most of the sightings were made in 1991 (Table 1). Based on the highest number of sightings on given days, an estimated 50-60 sheep were distributed in 2-8 subgroups along the north side of the river. During the summer of 1990, an incremental increase occurred with new lambs (mean lamb:ewe ratio - 74:100, n - 9 groups), but by late summer and fall few lambs were sighted. During late winter in 1991 only 1-2 yearlings were sighted with these 50-60 sheep. During spring, summer, and fall of 1991, no signs of parturiency, nor any lambs, were sighted north of the river. Also, no yearling or older rams had been sighted �180 Table 1. Number of sightings of mountain sheep in Area A from May 1990 to March 1992. Number of sightings Year North of Arkansas River South of Arkansas River Total 1990 58 3 61 1991 128 20 148 1992 75 13 88 261 36 297 Total north of the river since the 1990 hunting sea~on when two 1/2 curl rams were harvested. Thus, by summer of 1991, these 50-60 sheep were ~ ewes. This contrasted markedly with an estimated 30-40 sheep south of the river where a more "normal" sex/age class (mean 1amb:ewe ratio - 67:100, n - 6 groups; mean ram:ewe ratio - 37:100, n - 6 groups) was observed during the same period. Supporting the contention that the sheep on the north were all ewes were 1) observations of 5 hunting permitees who reported seeing only ewes (see 1991 S7 hunter questionnaires) and 2) our sightings during the rut (Nov-Dec) when no rams or breeding activity was observed. Thus, with this finding seemingly confirmed, plans were made to transplant 3 rams into the area. It was judged that young rams would best socialize and possibly stay with the ewes until the mating season in the fall and that a release date should be late enough so that the ewes would not breed and produce late lambs (late lambs probably would have poor winter survival and initiating late parturition patterns could negatively affect recruitment for years to come). On 21 Jan 92 three 2-year old rams were trapped during the drop-net trapping operation on the Rampart Range (Colorado Springs), collared with telemetry tansmitters (Black triangle collar 172.238 Mhz, Black collar (Bk) 172.313 Mhz, and Blue collar (Bl) 172.463 Mhz; Te10nics, 932 E. Impala Ave., Mesa, AZ 85204), and released in near proximity to about 30 ewes north of the river near Floodplain Recreation Site. Two of the rams (Bk and B1) initially socialized with ewe groups north of the river and some late breeding may have occurred (mounting of 2 ewes was observed 31 Jan 92). By 28 and 38 days post-release, however, these 2 rams had left the ewes north of the river, apparently crossed the river and highway, and joined other ewes south of the river. Distribution of sheep did not appear to be importantly different across seasons except that they may have spent more time in the southwest quarter of the study area during summer and apparently avoided it in fall and winter. There was a substantially greater number of sightings in summer (143 vs 17, 80, and 57 in the fall, winter, and spring, respectively) due to our increased �181 sampling activity during the boating season. A potential bias may have occurred if observers spent more time in the southwest half of the study area where either 1) a greater number of interactions between sheep and boaters occurred (specifically at Pinnacle boat launching area, 3-rock rapids, and Spikebuck railroad siding) and/or 2) such interactions were more observable. More specific locations of the sightings are shown for each km (n - 12) from Echo Canyon to Parkdale siding. Most of the sightings north the river occurred in km 2, km 3, and km 5. These include areas north of Pinnacle to 3rocks rapids and north of 5-points picnic site to 3/4 of a km to the east. Most of the sightings south of the river occurred in km 3 and km 5. These include areas south of the curve east of Pinnacle and southeast of 5-points picnic site on the south where the steep northwest exposed slopes were used for lambing. Although there may have been others, these latter 2 areas were the only lambing areas documented south of the river. There is less certainty about the lambing areas north of the river because they are based only on . sightings from May and June of 1990 (there were no lambs sighted north of the river in 1991 as mentioned earlier). Movements of groups of sheep were recorded on given days when such movements were observed from apparent beginning to end. Admittedly, these movements were minimum distances moved on given days because the sheep may have moved before observations began and after observations ended. Also, movements of sheep were estimated by connecting the sighting locations of marked animals over time. The most pausible travel routes, based on where non-marked sheep were sighted, were used to connect the sighting locations. Only one mark, yellow collar 29 (yl 29, an adult female), was present north of the river throughout the study. Later, as previously mentioned, 3 marked rams were transplanted into the area north of the river. Sightings of y1 29 (n - 29) yielded apparent movements along the entire study area corridor. However, she and her associates apparently concentrated their activity in the Pinnacle to 3-rocks rapids area. Sightings of only one of the 3 rams yielded apparent movements of any consequence. Sightings of Bk (n - 9) yielded movements north of the river until he apparently crossed the river and highway between 14 Feb and 26 Mar and joined sheep south of the river. Where he crossed the river was just a guess. By 3 Apr Bk and Bl were sighted together south of the river about 6.3 km east of 5-points where Bk was first sighted south of the river. Hence, they appear to have made east and west movements parallel to and south of the river (n - 3, sightings) with sheep who probably reside south of the river. Of sightings of marked sheep south of the river, yellow 26 (y126, an adult female) had the most (n - 14) and yielded movements along most of the study area corridor south of the river. She was often with other marked animals and her movements were judged to be indicative of most others south of the river. Other than the 2 rams, there was little evidence that sheep crossed the river during the period of study. Two sheep were killed on the highway, red eartag 43 on 14 Ju1 91 and an unclassified sheep (possibly a ram because the head was missing) on 1 Aug 91. These sheep and any others with them may have crossed the highway to drink from the river. The size of sheep groups north and south of the river were significantly different (range 1-44, mean - 10.7, ±SE 0.5, and range 1-18, mean - 6.9, ±SE 0.6), respectively; t - 3.09, df - 330, P < 0.01). No detectable pattern is evident from the distribution by group size. Similarly, no unexpected finding resulted from the calculation of a "corridor-shaped" home range. Assuming a sheep population of 90 (50-60 north and 30-40 south of the river; using �182 medians of 55 and 35, i.e. 55 + 35 - 90) and an area of sq km delineated by the minimum polygon method, the average area per sheep - sq km. These estimates may be biased because of the lack of information on any movements of sheep south of the river to areas further south and possibly east. Sheep habitat use was not investigated south of the river as it was on the north. Interactions between sheep and boaters and other selected stimuli suggest that sheep response was variable and that they probably habituated to boaters, trains, and people separated from them by the river (Table 2). Although the number of interactions noted between sheep and rafts were few (n - 8, Table 2), most (4 of 5) walked or trotted away from the raft when they were </- 30 m from the rafts. Hovever, they moved only about 10 m in response (Table 2). Conversely, when they were> 30 m from the raft(s) there apparently was no overt response. Sheep responded to trains when they were between the river and the railroad tracks, or when they were on the railroad tracks, by moving about 15-30 m upslope to the north. They typically responded at distances safe from the train, however, this was not always the case. Reportedly, there were 2 sheep killed by the train in 1990 and one in 1992 (7 Apr). A near-miss was observed on 2 Jun 1992 when the train traveling at an estimated 30-40 mph blew the train whistle twice at markedly different distances and a group of 16 sheep on the railroad tracks and right-of-way were slow to respond. When people were across the river from the sheep, the only overt response observed was "sheep looking at people". However, when people were north of the river (unusual because of limited access) or when people tried to approach sheep south of the highway for such reasons as archery hunting or throwing rocks, the response was predictably greater (Table 2). Vegetation, Visibility, Viewshed, and Water Ten separate classes of vegetation plus bare rock and water were classified initially. Later these were consolidated into PJ, mountain shrub, grassland, riparian, and rock in order simplify and be consistent with classes typically used by BLM. Variability in the density of dominant vegetation occasionally complicated judging the classification. For example, some grassland was initially classed as low density PJ. Percent of dominant vegetation (percent of PJ hits in the PJ class, etc.) ranged from 22-60% for PJ, 26-56% for mountain shrub, and 28-80% for grassland (Table 3). Most of the sheep north of the river were sighted in mountain shrub (43%) or in PJ (30%). Fewer sightings occurred either on the railroad or in the railroad right-of-way (14%) and in grassland (13%). Some grasses appeared to green-up along the railroad as early as February and March and frequent sheep use was observed. Similarly, sheep were observed feeding on young shoots of Russian thistle (Salsola iberica) on the railroad right-of-way. Hence, the railroad may have provided a somewhat unique microhabitat. Sheep use in Spring was observed along the east side of Echo Canyon grassland (meadow). Other minor uses of grassland included the Harvey Ranch field adjacent to the Parkdale railroad siding and a meadow east of the east end of Spikebuck railroad siding. Also during spring, some shrubs (littleleaf mockorange, in particular) were used extensively. Mountain mahogany was still used during mid-summer. �183 Table 2. Interactions between mountain sheep and rafts, kayaks, trains, people, and mountain lions from June 1990 to June 1992. Distance (m) 1 Year Stimulus Date Reaction 2 Moved To 3 Group stimulus size 4 Marks People/animal activity 1990 raft 17Jun sfr 10 15 18 none ? kayak 4Ju1 4Ju1 ifr ifr 40 100 30 25 15 18 none none ? ? people 22Ju1 sfr 50 35 5 approach w/ camera 12Aug sfr 120 40 2 y130, y132 none 29May 30May 30May sfr sfr mfr 10 10 10 30 20 30 21 9 9 y129 none none ? ? 13Jun 13Jun 14Ju1 28Jul none none none none 0 0 0 0 55 55 10 35 19 19 5 12 none none none none none none 29May 30May 13Jun 28Jul 14Sep sfr mfr mfr mfr none 20 30 20 50 ? 0 210 35 21 10 13 12 10 y129 none none none y129 none none none none none 28Ju1 sfr 5 150 3 none look @ sheep 12Aug 13Sep none mfr 0 180 50 30 26 9 none shoot arrow 13Sep mfr 25 40 9 13Sep mfr 180 45 9 13Sep mfr 35 45 9 13Sep ifr 300 45 9 -rd43 y127, y130 y127, y130 y127, y130 y127, y130 y127, y130 none ifr 0 40 180 40 7 14 none none not looking stalking ? 1991 raft train people mountain lion 22May 18Sep ? ? look @ sheep ? look @ sheep shoot arrow shoot arrow throw rock throw rock �184 Table 2. (continued) Distance (m) 1 Year Stimulus Date Reaction 2 Moved To 3 Group stimulus size 4 Marks People/animal activity 1992 train 20Jan 3lJan l2Feb l2Feb l4Feb l4Feb l4Feb 21Feb 21Feb 27Feb 28Feb 12Mar 13Mar 19Mar 3Apr 2Jun ifr mfr mfr none sfr none ·mfr look sfr none ifr mfr none mfr ifr mfr 30 25 25 0 15 0 20 0 20 0 20 15 0 5 15 15 280 600 100 100 180 50 230 80 20 100 120 900 25 30 5 65 15 21 22 6 22 16 22 10 10 5 12 23 24 8 16 16 people 3lJan 12Feb 2lFeb 27Feb 2Apr 22Jul look sfr sfr look look mfr 0 50 40 0 0 25 100 50 80 40 20 25 21 15 4 5 4 11 none none Bk Bk Bk none Bk none none none y129 y129 y129 none y129 y129 none blow whistle none none none none none none none none slowed none none none blow whistle blow whistle, near-miss none yell whistle Bk none none none none Bk, B1 none y126, approach y130, w/ camera Bk, Bl 1 Reaction includes: sfr - slight flight reaction where animals ~ away from stimulus mfr - moderate flight reaction where animals trot away from stimulus ifr - intense flight reaction where animals run away from stimulus. 2 Distance moved away from and in apparent response to stimulus. 3 Distance between animals and stimulus when initial response detected. 4 yl - yellow collar, Bk - Black telemetry collar, BI - Blue telemetry collar, and -rd43 - red eartag 43 w/o collar. �HI5 Table 3. Vegetation type, number of transects or sites, percent dominant vegetation (Pinus edulis and Juniperus spp. for PJ, shrub spp. for mountain shrub, grass spp. for grassland, and cottonwood spp. and Tamarix ~allica for riparian), percent obscured (how much of sq m target was obscured by vegetation and/or topography), and general location of site. No. Sites Type PJ 5 % Dominant Vegetation 22 44 60 34 30 % Obscured 63 68 97 81 76 General Location 5 Pts Echo Bootlegger Sulphur E Cedar Comment Hi density Mean - 77 .0 ( SE 5.9) Mtn Shrub 6 34 50 26 42 56 44 74 90 83 100 93 66 Twin Gulches Echo 3 Rocks Sulphur E Sulphur 5 Pts Mean - 84.3 ( SE 5.2) Grassland 6 72 28 80 34 70 80 2 57 41 57 60 15 Echo 3 Rocks Sulphur E Sulphur Spikebuck Spikebuck E side lrg meadow Rocky alluvial fan Cholla 20% Old burn Mean - 38.7 ( SE 10.1) 1 Riparian 3 Total 20 60 82 86 na na na Spikebuck Spikebuck Sulphur Cottonwood Cottonwood Tamarisk 1 Transects were run in and parallel to the drainages rather than at randomly chosen directions. Visibility was considered not applicable (na) because there was no evidence that sheep used this vegetation type. �186 Visibility was estimated at 17 of the 20 vegetation transect sites (Table 3). The lack of visibility or the percent of vegetation and topographic obstruction, as estimated by viewing the target, ranged from 63-97% (mean 77.0, ±SE 5.9) for PJ, 66-100% (mean - 84.3, ±SE 5.2) for mountain shrub, and 2-60% (mean - 38.7, ±SE 10.1) for grassland (Table 3). As sampled, there was no significant difference in visual obstruction between PJ and mountain shrub (t - 0.941, df - 9, P > 0.20). More importantly, however, is that these 2 vegetation classes present quite different visual obstructions (PJ more obstruction in canopy and mountain shrub more obstruction at the base) and that topographic features may have accounted for as much of the variability as did differing densities (8 of 17 sites had at least 1 of 5 data points totally out-of-view because of topography). . Based on viewshed calculations, there was little land along the canyon walls and surrounding slopes north of the river that were not in sight from at least some angle. This yielded little area where sheep and people were incapable of sighting each other. A smaller area along the canyon could be described as a -noise-shed- where the railroad, vehicle traffic, and human voices from boating (yelling when in rapids, etc.) could be heard. However, no data in this regard was collected. Sources of water were searched for in most drainages north of the river including two about 1 km northeast of Pinnacle (nicknamed -Twin Gulches-), one north of 3-rocks rapids, Sulphur Gulch, Spike Buck Gulch, Bootlegger Gulch, one north of Shark's Tooth rapids, and Cedar Gulch. Of the few sources of open water found, namely, about 1.4 km of Spike Buck starting about 1.5 km up the drainage from the river as observed on 17 Aug 91, a few seeps and small pools up Bootlegger, and some up Cedar, DQ evidence of sheep use was detected. Contrary to the lack of sheep use of these few sources of water, the river was almost always readily available and regularly used. The only times river water was not readily available were periods in winter when the river froze over or formed an ice shelf along the shore. On several occasions sheep were observed eating hoarfrost or snow at the ice edge and attemping to reach open water just beyond the ice shelf (a potentially dangerous venture). Of several sources of water south of the river, sheep use was observed in a marsh located next to the highway at 5 Points Picnic Area. This source had open water throughout the spring, summer, and fall. Temporal sources of water were also located in the drainage south of 5 Points and up Baker Gulch. Running water and pools were plentiful in the drainage south of 5 Points in early summer 1991 and 1992, but by August no open water was found. DISCUSSION The information on sheep numbers and sex/age composition indicates that north of the river a recruitment problem and a reproductive failure occurred in 1990 and 1991, respectively. Furthermore, the estmated lack of rams and breeding during the rut in late 1991 portended similar results for production in 1992. This is a problem of no small dimension unless some rams or other sex and age classes eventually move into the area or we successfully intervene. The first attempt to solve the problem, that of introducing three 2-year old rams, may not have been successful because they left the ewes north of the river after relatively short periods. Whether they will rejoin the ewes north of the river remains equivocal (Reed et al. 1992). The estimate that rams from south of the river did not cross and breed the ewes on the north, led us to question the hypothesis that the river does not act as a barrier to sheep movement. The 3 introduced rams, however, apparently crossed the river to the south. �187 Alternatively, there may be behavioral and temporal aspects in motivating animals to cross the river. For example, did these introduced rams see other rams south of the river and cross to join a bachelor group despite the fact that when they were observed south of the river, they were usually with ewes? Other hypotheses, especially those involving changes over time that are important to sheep, need to be examined. Some ideas include: - habitat "sinks" where "source" areas export to other areas (sinks) (Lewin 1989, Odum 1992), but where in our case the exporting has temporally or permanently ceased - disturbance as related to the increase in rafting and a greater sensitivity in rams versus ewes - disease where male lambs may have been more susceptible than female lambs and/or where reproductive capabilities of ewes or rams were affected - transplants where sheep from the Grape Creek (1983), Texas Creek (1984), and Bear Gulch (1985) releases may have eventually established or increased the importance of ram areas south of the river - "ugly" ewe syndrome, a development without biological precedence in wild sheep. The information on interactions suggest that sheep in this study generally responded to familiar and predictable stimuli including boating and people-onfoot activities by exhibiting only slight to no apparent overt reactions. When they were on or south of the railroad tracks and suprised by trains, suprised by humans, or when human activity was less predictable, the response was heightened. During late spring and summer, boating became daily and relatively predictable phenomena (typical hours, etc.) and habituation would have been expected. Despite the seemingly slight overt responses, there is a question of whether the interactions with boaters deterred sheep from getting water and how important that was to sheep. The timely use of water during early to mid summer would be especially important to lactating females and new lambs. Furthermore, there may have been physiological impacts and energy costs that occurred in the absence of overt behavioral reactions (i.e. elevated heart rate and excitement levels) (MacArthur et al. 1982). For example, MacArthur et al. (1979) reported that the continued presence of a human within 50 m of sheep resulted in a 20 percent rise in mean heart rate. Additionally, King and Workman (1986) suggested that responses of desert bighorn (~ ~ nelsoni) to encounters with humans were more severe and energy costly for animals that had been previously exposed to relatively high levels of human disturbance. Hence, for animals in this study, the effects of increased boating acitivity may not diminish, but rather increase over time. The habitats that were used by sheep in this study can generally be described as either PJ, mountain shrub, grassland, or bare rock located in.a relatively narrow river canyon with adjacent steep terrain. Some of the components of these habitats that were likely important to sheep were not unlike those often reported by other workers (Boyd et al. 1986): - water, - steep, rocky, escape terrain, �188 - visibility for predator detection and visual communication, - mineral licks, - suitable thermal environments, and - forage of adequate quantity and quality. Favorable combinations of these components existed at best in patchy distributions. For habitats north of the river, the river served as the primary source of water and possibly as some kind of southern boundary or temporal barrier. Similarly, the distances that sheep venture away from water « 3.2 km [Smith et a1. 1991:212.]), in this case the river, may have contributed to a northern boundary, and the distances that sheep venture away from escape terrain (0.5 km [Van Dyke et a1. 1983]) may have established the eastern and western boundaries of this sub-population. Additionally, the steep terrain used for escape and lambing, the openness or visibility used for predator detection and visual communication, the southern exposed slopes used for suitable thermal environments, and the forage quantity (number of species and density) and quality (palatability and nutritional levels), probably contributed to a mosaic of habitats within the delineated home range. These components also apply to the sheep south of the river, except that the river is judged to be used for water only infrequently and that any limits to use of habitats further south and west are unknown. A missing component for habitats both north and south of the river appear to be regularly used mineral licks. Only one area, north of the "Sa1t1ick" boat access area, was estimated to be used by sheep for mineral and this was infrequent. Escape terrain and lambing areas for both north and south of the river were judged to be adequate, although specific measurements were not made. Escape terrain has been variously described as cliffs with at least 8 x 200 m (ht x length) dimensions (McCollough et a1. 1980) and areas of at least 0.16 ha (Van Dyke et a1. 1983). Van Dyke et a1. (1983) suggested a relationship between distance from escape terrain and extent of habitat use. Wake1yn (1987) indicated that ranges supporting greater numbers of sheep had more habitat on or within 0.25 km of escape terrain. A number of smaller cliffs could probably be substituted for a large one and terraced cliffs that allow sheep a variety of escape routes are likely superior. Lambing areas have been described as cliffs with 8 x 260 m (ht x length) dimensions and areas of at least 2 ha (Van Dyke et a1. 1983). However, it may take a computer model to adequately describe the array of variables associated with escape terrain and lambing areas. Good visibility or the lack of visual obstructions is generally espoused as an important component for predator detection, visual communication, and efficient foraging in sheep (Boyd et a1. 1986, Wake1yn 1987), but how it has been measured has varied between workers. Risenhoover and Bailey (1985) estimated the percent of each quarter of the compass over which an object 90 cm in height could be seen at 40 m from the observer. Thomson (1975) used a more exact measure of the degree of obstruction by using a screen behind the vegetation, photographing it, and later projecting the image on a quadrat matrix for analysis. The technique used in this study was considered sufficiently accurate for estimating the visibility of approaching predators �189 (of which would likely be < 1 m in ht) and other sheep, but the measurements were not taken in habitats where sheep were necessarily observed. Hence, considering that these training sites were not representative of habitats used by sheep (possibly having> vegetation and topographic relief), it may not be suprising that measurements of their visibility were relatively low (i.e. for PJ and mountain shrub). Habitat evaluation procedures have been developed for desert bighorn sheep (Q. (Grunigen 1980, Hansen 1980, Holl 1982, Armentrout and Brigham 1988, Cunningham 1989) and more recently for Rocky Mountain bighorn sheep (Smith et al. 1991). However, only the procedure by Smith et al. (1991) critially examines the minimum viable population (MVP; defined as the smallest isolated population having at least a 95% probability of surviving at least 100 years [Shaffer 1983]) in sheep. Using information from Berger (1990) and suggestions of others (e.g. Van Dyke et al. 1983), Smith et a1. (1991) suggested that ll2 individuals represent a "best estimate" MVP. In the present study, sheep south of the river likely interact with other sub-groups, thereby relaxing the MVP requirement, but for those north of the river of which appear to be isolated, it is another matter. £. mexicana) Important considerations in any habitat evaluation are the relationships between habitat use/availability data, habitat quality, and population density. Van Horne (1983) suggested that our understanding of individual species-habitat relationships is still rudimentary. She indicated that without sufficient data it would be difficult to distinguish "source" and "sink" habitats and that habitat quality is inversely related to MVP size. She gave an extreme example where one could imagine a habitat in which all the animals were immigrants and none emigrated or reproduced, and thus, where the quality of the habitat would be zero. Does some of this example apply to the habitat north of the river in our study? Additionally, Hobbs and Hanley (1990) suggested that we do not yet have a scientifically sound and complete method of evaluating habitat quality. They indicated that habitat evaluation systems capable of predicting, for example the effects of human impacts, require understanding the mechanisms by which the environment influences the distribution and abundance of animals. They further suggested that an understanding of the cause-and-effect relations linking the performance of populations to resources in their habitats is fundamentally important. Predictability could potentially be greatly increased by the use of habitat models. However, one of the problems in attempting to validate habitat models is the difficulty of obtaining an independent measure that can serve as a suitable comparison (Schamberger and O'Neil 1986). MANAGEMENT IMPLICATIONS AND RECOMMENDATIONS The implications of habitat use by bighorn sheep north of the Arkansas River suggest different and even larger problems than anticipated before this study began. Specifically, the poor recruitment in 1990, the reproductive failure in 1991, and the 1-2 late lambs in 1992 (likely reSUlting from the temporal presence of the 3 introduced rams) in the 50-60 female sheep north of the river, have become immediate concerns. Unfortunately, potential causes are only hypotheses at this point and it is unknown whether or to what extent the habitat can be implicated. Therefore, the first priority should be to seek ways of solving the reproductive failure by insuring breeding in the females �190 north of the river. Assuming the first attempt (i.e. release of the 3 rams 21 Jan 92) was unsuccessful, another release of younger males such as 10 male lambs from regular trapping operations January-February 1993 should be made. Admittedly, this is a trial-and-error approach, but in the absence of adequate controls, it is one of the few practical options available. If this should fail, new and innovative approaches will need to be developed. A second priority level should include - the question of the effects of disturbance on sheep in the narrowly confined corridor north of the river, - the applicability of validating, refining, or developing habitat models, - and more specific recommendations for given areas within the study area. Specific recommentations for given areas (kms 1-12, Echo Canyon to railroad siding at Parkdale) are as follows: km 1 - diminish or eliminate cattle gazing in the Echo Canyon meadow where sheep use has been shown in order to reduce or eliminate any competition for forage and space which may have limited movements and expansion of sheep use to the west along the north side of the river. km 2 - increase watchable wildlife information on bighorn sheep in the Pinnacle rest and boating access area. km 3 - develop a water source south of the highway about half way up the slope where considerab~e sheep use has been shown in order to reduce any need to use water from the river (this is the site of one of the sheep-vehicle accidents). - develop a boating "rest" area with a walk and stairs so that boaters can "check out" the 3-rocks rapids from the south side of the river vs the north side where ~ frequent sheep use has been shown. km 4 - insure that the water source south of the highway and up the drainage southwest of 5-points picnic site has free-flowing water throughout the summer and fall. - provide a saltlick north of the river and railroad tracks in view of the Bighorn Sheep Canyon Wildlife Viewing Site north of the highway. km 5 - provide a saltlick south of the 5-points picnic site south of the highway and increase watchable wildlife information at the S-points picnic site south of the highway. km 6 - no specific recommendations. �191 km 7 - provide a saltlick north of the river and railroad tracks in view of the east end of the Spikebuck siding pull-off. km 8 - no specific recommendations. km 9 - develop a water source south of the highway about 1/3 the way up the slope where sheep use has been shown in order to reduce any need to use water from the river (this is the site of one of the sheep-vehicle accidents). km 10 - Km 12 - no specific recommendations. SUMMARY Mountain sheep habitat use was evaluated in the Arkansas River Canyon from Echo Canyon to the Parkdale railroad siding. GIS technologies (MIPS) were used to summarize and produce color and 3-dimensional plots. An estimated 5060 sheep in 2-8 subgroups were distributed along the north side of the river. During 1991, these sheep were all females and no reproduction was detected. This contrasted markedly with a "normal" sex/age class of the estimated 30-40 sheep south of the river. Introduction of 3 rams to the sheep north of the river may not have solved the estimated problem of "failing to breed" because they left the sheep north of the river and joined those south of the river. New and innovative approaches may be needed to help solve this unexpected problem. Management recommendations are provided for 1) lessening the potential of disturbance to sheep from boating, 2) improving availability of saltlicks and water in selected habitats, and 3) increasing watchable wildlife opportunities. Acknowledgments. We thank T. Grette and L. M. Berta of the BLM for technical advice and management and budget support, and S. Ogilvie, W. Travnicek, D. C. Finch, D. C., Lovell, D. L. Schrupp, M. W. Miller, and R. B. Gill of the CDOW for field support, advice, and management and budget support. J. Backstrand collected bighorn sheep field data during the summers of 1990 and 1991. L. Spicer collected bighorn sheep field data during late winter and spring of 1992. D. Chess of the Denver and Rio Grande Western railroad and J. A. Rodriguez, W. H. Thornton, E. Lundberg, and C. Schulze of the CDOW assisted in releasing the 3 rams. LITERATURE CITED Armentrout, D. J. and W. R. Brigham. 1988. Habitat suitability rating system for desert bighorn sheep in the Basin and Range Province. USDA Bureau of Land Management Technical Note 384. l8pp. Berger, J. 1990. Persistence of different-sized populations: an empirical assessment of rapid extinctions in bighorn sheep. Conserv. BioI. 4:91-98. Boyd, R. J., A. Y. Cooperrider, P. C. Lent, and J. A. Bailey. 1986. Ungulates. Pages 519-564 in Inventory and Monitoring of Wildlife Habitat. A. Y. Cooperrider, R. J. Boyd, and H. R. Stuart, eds. U.S. Dep. Int., Bur. Land Manage., Denver, CO. 858pp. �192 Chronic, H. 1980. Roadside geology of Colorado. Mountain Press, Missoula, MT. 322pp. Cunningham, S. 1989. Evaluation of bighorn sheep habitat. Pages 135-160 in The desert biighorn sheep in Arizona, R. M. Lee, ed., Arizona Game and Fish Dep. 265pp. Grunigen, R. E. 1980. A system for evaluating potential bighorn sheep transplant sites in northern New Mexico. Bienn. Symp. North. Wild Sheep and Goat Counc. 2:211-228. Hansen, C. G. 1980. Habitat evaluation. Pages 320-335 in Monson, G. and L. Sumner. eds. The desert bighorn -- its life history, ecology, and management. Univ. Arizona Press, Tucson. 370pp. Hobbs, N. T. and T. A. Hanley. 1990. Habitat evaluation: do use/availability data reflect carrying capacity? J. Wild1. Manage. 54:515-522. Holl, S. A. 1982. Evaluation of bighorn sheep habitat. Desert Bighorn Sheep Counc. Trans. 26:47-49. King, M. M. and G. W. Workman. 1986. Response of desert bighorn sheep to human harassment: management implications. Trans. N. Am. Wi1d1. Nat. Resour. Conf. 51:74-85. Lewin, R. 1989. Sources and sinks complicate ecology. Sci. 243:477-478. MacArthur, R. A., R. H. Johnston, and V. Geist. 1979. Factors influencing heart rate in free-ranging bighorn sheep: a physiological approach to the study of wildlife harassment. Can. J. Zool. 57:2010-2021. MacArthur, R. A., V. Geist, and R. H. Johnston. 1982. Cardiac and behavioral responses of mountain sheep to human disturbance. J. Wild1. Manage. 46:351-358. McCollough, S. A., A. Y. Cooperrider, and J. A. Bailey. 1980. Impact of cattle grazing on bighorn sheep at Trickle Mountain, Colorado. Bienn. Symp. Northern Wildl. Sheep and Goat Counc. 2:42-58. Odum, E. P. 1992. Great ideas in ecology for the 1990s. BioSci. 42:542-545. Reed, D. F., M. W. Miller, and J. Vayhinger. 1992. Spinster bighorn ewe groups in the Arkansas River canyon, Canon City, Colorado. Bienn. Symp. North. Wild Sheep and Goat Counc. Abst. (In press) Risenhoover, K. L. and J. A. Bailey. 1985. Foraging ecology of mountain sheep: implications for habitat management. J. Wild1. Manage. 49:797-804. Schamberger, M. L. and L. J. O'Neil. 1986. Concepts and constraints of habitat-model testing. Pages 5-10 in J. Verner, M. L. Morrison, and C. J. Ralph, eds. Wildlife 2000: modeling wildlife-habitat relationships of terrestrial vertebrates. Univ. Wisconsin Press, Madison. �193 Shaffer, M. L. 1983. Determining minimum viable population sizes for the grizzly bear. International Conf. Bear Research and Manage. 5:133-139. Smith, T. S., J. T. Flinders, and D. S. Winn. 1991. A habitat evaluation procedure for Rocky Mountain bighorn sheep in the intermountain west. Great Basin Nat. 51:205-225. Thomson, W. R. 1975. A photographic technique to quantify lateral cover density. J. South Afr. Wi1dl. Manage. Assoc. 5:75-78. Van Dyke, W. A., A. Sands, J. Yoakum, A. Po1entz, and J. Blaisdell. 1983. Wildlife habitat in managed range1ands--the Great Basin of southeastern Oregon: bighorn sheep. USDA Forest Service General Technical Report PNW159. Pacific Northwest Forest and Range Experiment Station, Portland, OR. 37pp. Van Horn, B. 1983. Density as a misleading indicator of habitat qua1ity~ J. Wi1d1. Manage. 47:893-901. Wake1yn, L. A. 1987. Changing habitat conditions on bighorn sheep ranges in Colorado. J. Wi1d1. Manage. 51:904-912. ~_Q_'-__ Prepared by -f;(; ..•• ~~ .)•..• d"",-,,-P_( _£ •.... -----='=) '..:;.." =r- \-//_'"" Dale F. Ree~ Wildlife Researcher C __ ��195 Colorado Division of Wildlife Wildlife Research Report July 1992 JOB PROGRESS REPORT State of -¥C~o .•.. lo~r~a~d=:::oll::.._ _ Project No. W-153-R-4 Work Plan No. 3A Job No. Pron&horn Investigations Habitat Selection and Population Performance of a Pioneering Pronghorn Population Period Covered: Author: Mammals Research July 1, 1991 - June 30, 1992 T. M. Pojar Abstract The distribution of the Middle Park pronghorn (Antl1ocapra americana) herd has been monitored since January 1, 1987. During this time, the population has increased from 80 to approximately 309 animals (summer 1992). There have been 2,614 locations recorded for radioed animals and 841 locations for animals marked with either plastic neck collars and/or ear tags. During summer the distribution ranges from north of Granby (Coffee Divide) west to Kremmling, then north to Muddy Pass mostly north of the Colorado River (Fig. 1). During winter, they inhabit about a 65 km2 (25 mi2) area north and east of Kremmling on ridgetops west of the Troublesome Creek drainage. During a trapping operation on December 19, 1991, 39 new or refurbished radios were deployed. Five solar powered ear tag radios were put on male fawns and thus far have performed comparable to battery powered neck collar radios. The populations rate of increase bas ranged from 0.5~ in 1986-87 to 0.18· in 199·1-92 but is projected to be near zero for 1992-93 based on mid summer fawn:doe estimates~ Huriting·was permitted again (for the·second year) in 1991 with 10·~uck permits and 5 doe/fawn pe~its. There was 100 percent hunter success with no ' documented evidence of wounding loss. ��197 HABITAT SELECTION AND POPULATION PERFORMANCE OF A PIONEERING PRONGHORN POPULATION Thomas M. Pojar P.N. OBJECTIVE Describe population dynamics and habitat use of a pioneering, expanding pronghorn population. SEGMENT OBJECTIVES 1. Describe seasonal and annual distribution of the Middle Park pronghorn population. 2. Determine sample sizes of radio-collared animals and observations necessary to describe habitat preferences. 3. Monitor population dynamics of Middle Park pronghorn with: a. Ground counts to describe changes in population size. b. Ground counts to quantify population sex and age composition. STUDY AREA The study area is described in Pojar (1988:183-184). For orientation of Middle Park in relation to the state of Colorado see Figure 1. The approximate area of sagebrush steppe habitat in Middle Park is outlined in Figure 1. This is considered to be the potential area of distribution for pronghorn in Middle Park. METHODS AND MATERIALS SEASONAL AND ANNUAL DISTllIBUTION Tracking was done mostly from the ground to increase the probability of observing and identifying animals with numbered plastic collars. Fixed wing aircraft was used if an animal could not be located after a reasonable effort from the ground. Legal descriptions of animal locations were recorded to the nearest quarter mile then converted to UTM (U.S. Army 1973) coordinates for computer processing. All radioed animals have been located biweekly (with very few exceptions) since January 1, 1987. POPULATION SIZE AND STllUCTURE Herd structure estimates are obtained in late summer or early autumn by locating all radioed animals and classifying all animals that accompany them. The herd structure estimate that is used in population projections is the one with the largest sample size obtained in August or September. Classification after October 1 is not used because the probability of mistaking early fawns for does increases. �198 In all winters since this population has been monitored, they have wintered in essentially the same relatively limited area. At some time during the winter they are all in a single group which makes it possible to locate the entire population and obtain a complete count. In addition, it is possible to count all mature bucks (age 1.5 years and older) and verify or correct the buck:doe ratio from the earlier herd structure estimates. The total population count and number of mature bucks counted during winter are used in population projections. Population projections are based on the following assumptions: 1. Winter counts represent the total population and the total number of mature bucks in Middle Park. 2. Late summer age ratio estimates represent "recruitment" into the population. 3. Annual survival of mature bucks and does and female fawns is 92.5X. 4. Annual survival of males in their first year (after weaning) is SOX. (This severe mortality on male fawns is arbitrary, however, it allows the number of mature males in subsequent years to match fairly well with winter counts.) RESULTS SEASONAL AND ANNUAL DISTRIBUTION Fifty-one pronghorn were trapped on 19 December 1991 using a corral trap (with curtains) and wing fences. The animals were driven into the trap by helicopter. Thirty-nine of the animals were fitted with radios and the remainder with either plastic neck collars and/or ear tags. Solar powered ear tags were put on 5 male fawns and expandable neck collars were put on 4 female fawns. There were no mortalities during the trapping operation, however, a mature buck suffered a broken left front leg. He was not seen during subsequent surveys of the area and is presumed to have died from this injury. With the deployment of the 39 new radios in December of 1991, the total number of working radios was 46. (See Appendix I for the status of all radios deployed on pronghorn in Middle Park since December 1986) However, by July 1992, some of the radios have failed and some radioed animals have died leaving 40 working radios. The solar powered ear tag radios (weighing 28g vs. 450g for neck collar radios) have performed comparable to battery powered radios. On sunny days, the signal is equal to battery powered radios in volume and distance; under heavy overcast, however, the signal is considerably weaker but still audible. The ear tags were applied with the solar panels to the forward side (inside) of the ear. Better solar exposure would have been possible by putting the solar panels on the back side of the ear but this application would have made the radio more vulnerable to damage or removal by the animal from scratching its ear with its hind foot. Apparently there is ample reflective solar energy to operate the radio with the panels facing somewhat forward and downward. �199 Other than the weight differential between the ear tag radios and neck collar radios, an additional advantage is that the expansion of bucks' necks during rut is of no consequence with ear tags. The distribution of the population, as determined by re-location of radioed and marked animals, has not changed substantially from previous years (see tables and graphs in Pojar 1991). In general, summer distribution is from Granby west to Kremmling and north to Muddy Pass. All observations in the past years have been north of the Colorado River (Figure 1). The winter distribution includes about 65 km2 (25 mi2) to the north and east of Kremmling; mostly on the ridges west of Troublesome Creek. In past years it has been documented that some of the wintering population from the Kremmling area summered near Granby, however, consistent tracking was not possible because no radioed animals were involved. Fortunately, the recent trapping operation resulted in 5-6 animals of the summering Granby herd being radioed. The furthest north and east that they have been located thus far is in the Coffee Divide and Table Mountain area near Grand Lake. There is a total of 30-40 animals that summer in this area. The route they take from the wintering area near Kremmling, parallels the Colorado River to the north. It passes Corral Creek, Parshall Divide and Kinney Creek enroute to the sagebrush covered hillsides north of the Horn ranch and on to Coffee Divide. Several movements of individual animals have been documented between the Kremmling area and the Granby area within a few days time during summer. Consistent evidence of occupation of the sagebrush habitat type south of the Colorado River has been lacking until now. One radioed doe, accompanied by a buck and 2 unmarked does, have summered (1992) about 3-4 km (2+ miles) east and south of Junction Buttes. During the spring dispersal from winter range, 2 radioed does moved into North Park near Coalmont (16 km (10 mi) northeast of Muddy Pass). One of the does returned to Middle Park by mid-summer but the other one (radio 148.672) remains in North Park. Her wintering area will be of particular interest. POPULATION SIZE AND STRUCTURE Accurate counts of the total population size can be obtained during winter and herd structure estimates are obtained in late summer (Table 1). The annual changes in population size are used to calculate the rate of increase (N2 - Nl / N1) (Table 2). The rate of increase for 1991-92 is near zero (Table 2) and if taken at face value, would indicate that the K-value for the population has been reached. With a projected population of 309 animals for 1992 (Table 3) and the addition of the 15 animals removed during the 1991 hunt, the K-value for this population is 324 animals. The projected K-value from past data (Pojar 1991) was 340 animals. These numbers, of course, reflect not only population density but also other biotic and climatic conditions in Middle Park. The population projection for late summer 1992 is presented in Table 3. The projection is based on the age ratio obtained during mid-summer 1992 and the total count of the herd and the number of bucks obtained during the winter of 1991-1992. In this projection, it is assumed that the age ratio for the 1992 production is 51 fawns:lOO does. This is a critical assumption and subsequent ratios based on larger sample sizes may have a significant impact on this projection. �200 ..,,--.•,,- , , I I 1 N ~ ••..•DENVER MIDDLE PARK Colorado Figure 1. Location of Middle Park in relation to the state of Colorado. The dashed line circumscribes the general area of sagebrush steppe habitat. �201 Table 1. Herd structure of Middle Park pronghorn based on a sample obtained by locating radioed animals in late summer. 1986' 1987 1988 1989 1990 1991 OF POP. POP. SIZE NO. JW>IO B:D RATIO F:D RATIO SAlIPIZ % JW>IO 80 122 160 223 261 307 7 24 22 17 13 5.7 15.0 10.2 6.5 4.2 36 54 40 56 22 23 77 77 32 50 47 65 47 63 108 161 148 148 59 52 68 72 66 48 % , This year's data based on the sample of the population trapped 16 December 1986 . .Tab1e 2. Population size of the Middle Park pronghorn herd during winter and the calculated rate of increase. POP. SIZE 1986-87 1987-88 1988-89 1989-90 1990-91 1991-92 1992-93 (projected) 80 122 160 223 261 307 309 RATE OF INCREASE .52 .31 .39 .17 .18 < .01 �202 Table 3. Population projection for the Middle Park pronghorn population. text for the assumptions. I POPULATION BUCKS DOES FAWNS TOTAL 62 138 92 292 WINTER MORTALITY 62 X .075 - 5 MORT 138 X.075 - 10 MORT 3IX.5-16B 3IX.075-2D 32 PRE-FAWNING 1992 62 - 5 57 MATURES + 15 YRLs TOTAL - 72 138 - 10128 MATURES + 29 YRLS TOTAL - 157 MATURE 57 YRLS 15 TOTAL 72 MATURE 128 YRLS 29 TOTAL 157 See WINTER '91-92 LATE SUMMER 1992 229 @ 51F:100D 157 X .51 80 FAWNS 309 REFERENCES CITED Pojar, T.M. 1988. Habitat selection and population performance of a pioneering pronghorn population. Colo. Div. Wi1d1. Res. Rep. July, pp 181-192. Pojar, T.M. 1991. Habitat selection and population performance of a pioneering pronghorn population. Colo. Div. Wi1d1. Res. Rep. July, pp 161-167. U.S. Army. 1973. Technical Manual: Universal Transverse Mercator Grid. Headquarters, Dep. of the Army, Washington D.C. TM No. 5-241-8, 64 pp. ( Prepared ~, /) , Thomas M. Pojar Wildlife Researcher C �203 APPENDIX I �204 Table 1. SEX M M F F F F F F F F F F F F F F F F F F F F F F F F F F F M M M M M M M M F F F F M F F F M M M M M M &IE 3+ F 3+ 2+ 3+ 3+ 3+ 3+ 2 3+ 2+ 3+ 3+ 3+ 3+ 3+ Y Y Y 3+ Y 3+ 3+ F F F F 3+ 3+ F F F 3+ 3+ F F 2+ 3+ 3+ 3+ 2+ 2 Y 3+ 3+ 3 3+ 3+ 2+ 3+ 2 Middle Park pronghorn marked December 19, 1991. EAR IAQ QLD NEW gQllARS QLD NEW W16R? SUEF (DAM - YC17) B29R B12 W5R W22R B23R 148.630 Y4R 148.400 W11R W8R Y1R 148.530 W17R W19R W14R W27R B11R 148.580 W20R W13R W2R W37R B42R B23 W36R W38R W7R W6R W12R W18R SUEF (DAM - YC17) W35R W26R W15L W28L W23L W10R W25R W21L W33L W4R Y11R B1R B34R. B17 W30R W9R W32R W39R Y47R W34L SUEF W24R W42R W29R W31R W40R 148.110 148.130 148.200 148.210 148.220 148.230 148.240 148.250 148.260 148.270 148.280 148.290 148.310 148.320 148.330 148.340 148.350 148.360 148.380 148.420 148.660 148.67 148.680 148.690 148.710 148.720 148.738 148.780 148.790 148.800 148.810 148.820 148.830 148.840 148.950 149.210 149.250 149.330 149.370 B10 B31 B32 B33 B50 B51 B54 B52 B53 B55 gQMMENTS BROKE L FRONT LEG G19L AND PURP23R, FAWN IN '91 YR.L IN '88; AGE - 4.5 YRS YR.L IN '88; AGE - 4.5 YRS 2+ IN '86; AGE - 7+ YRS AGE 7+ YRS; WAS RAD 148.420 YR.L IN '88; AGE - 4.5 YRS YR.L IN '88; AGE - 4.5 YRS TELONICS REF1JRB B14L & G10R; FAWN IN '91 FACTORY REF1JRBISHED SOLAR EAR TAG SOLAR EAR TAG SOLAR EAR TAG SOLAR EAR TAG SOLAR EAR TAG TELONICS TELONICS TELONICS TELONICS REF1JRB; AGE - 7+ REF1JRB; AGE - 3.5 YRS REF1JRB; AGE - 5.5 YRS REF1JRB REMOVED OLD ET; AGE - 5.5 YRS Y5L W5R DAM-148.580 6/6/89 �205 Table 2. Status of radios deployed on Middle Park pronghorn age signifies the animals age as of December 1991. SEX AGE EAR TAG RADIO F 6+ B25R 148.410 F 6+ Y12R 148.450 F M 6+ 4 B26R B2R 148.460 148.520 F 6+ Y1R 148.530 F 4 B6R 148.540 F F 5 F F 4 4 B8R B27R BSR B11R 148.550 148.560 148.570 148.580 M 6+ M 5 F 6+ B3R Y3R B14R 148.590 148.600 148.620 F 4 B23R 148.630 F 6+ B24R 148.640 M F 4 4 B16R B17R 148.650 148.700 M 6+ Y19R 148.750 F 6+ Y10R 148.850 F 4 B10R 148.950 5 **In the field and functional 12/15/88. The STATUS CEASED WORKING 7/7/90, ANIMAL lAST SEEN 11/28/91. **FUNCTIONAL AS OF 12/31/91. REPLACED 149.250 WHICH WAS REFURBISHED AND PUT ON DOE 3+ ET B1R 12/19/91. CEASED WORKING 11/15/91. CEASED WORKING 4/5/90, ANIMAL lAST SEEN 11/28/91. CEASED WORKING 10/1/91, LAST SEEN ON THAT DATE. THIS RADIO REPLACED 148.420 WHICH WAS REFURBISHED AND PUT ON DOE 3+ ET W38R 12/19/91. CEASED WORKING 10/1/91, lAST SEEN ON THAT DATE. **FUNCTIONAL AS OF 12/31/91. **FUNCTIONAL AS OF 12/31/91. **FUNCTIONAL AS OF 12/31/91. REPlACED WITH 148.310, WAS STILL FUNCTIONAL AS OF 12/19/91. **FUNCTIONAL AS OF 12/31/91. **FUNCTIONAL AS OF 12/31/91. CEASED TO FUNCTION AS OF 3/29/91, ANIMAL LAST SEEN 11/27/91. REPlACED WITH 148.210, WAS STILL FUNCTIONAL AS OF 12/19/91. CEASED TO FUNCTION 5/16/91, ANIMAL LAST SEEN 11/27/91. **FUNCTIONAL AS OF 12/31/91. CEASED TO FUNCTION AS OF 9/2/91, ANIMAL LAST SEEN 11/27/91. FORMERLY YELLOW COllAR 118, WAS KIllED ON HlJY 40 E. OF KREMMLING 9/9/90 --REPlACED BY ATS WITH RADIO 148.738 WHICH WAS PUT ON DOE AGE 3+ ET W26R ON 12/19/91. FOUND DEAD 7/27/89, PREVIOUSLY HAD RADIO 149.210. RADIO 149.210 WAS REFURBISHED AND PUT ON DOE AGE 7+ ET Y11R. FOUND DEAD 4/5/90--AGE 3. RADIO REFURBISHED BY ATS AND PUT ON MALE AGE 2+ ET W4R ON 12/19/91. as of 12/31/91. �206 Table 3. Status of radios deployed on Middle Park pronghorn 12/16/86, were females age 2+ at the time which makes them 7+ in December 1991. SEX EAR TAG RADIO F Y2R 148.390 F F Y4R Y1R 148.400 148.420 F F Y9R Y7R 148.430 148.440 F Y10R 149.210 F Y12R 149.250 F Y20R Y13R 149.290 149.370 F all STATUS CEASED WORKING 8/17/90, ANIMAL IAST SEEN 7/17/91. REPIACED W/ 148.220 12/19/91 REPIACED W/ 148.530 12/15/88 AND THEN REPIACED W/ 148.250 ON 12/19/91. RADIO 148.420 WAS REFURBISHED AND PUT ON FEMALE AGE 3+ W/ ET W38R12/19/91. CEASED WORKING 1/7/91, IAST SEEN 1/7/91. CEASED WORKING SHORTLY AFTER 12/12/91, IAST SEEN 12/12/91. REPIACED W/ 148.850 12/15/88, THE ANIMAL WAS FOUND DEAD 7/27/89. RADIO 149.210 WAS REFURBISHED AND PUT ON FEMALE AGE 3+ ET YIIR 12/19/91. REPIACED W/ 148.450 12/15/88, RADIO 148.450 STILL IN FIELD & WORKING 12/31/91, RADIO 149.250 REFURBISHED AND PUT ON DOE 3+, ET B1R 12/19/91. CEASED WORKING 4/7/88, IAST SEEN 4/26/90. KILLED BY AUTO 7/14/87, RADIO REFURBISHED AND PUT ON DOE 2+, ET W30R 12/19/91. Note: In tables 1, 2 and 3 the designation for ear tags are as follows: Y-yellow, B-b1ue, W-white and R-right ear, L-1eft ear. SUEF - ear tags put on animal when it was a fawn by Sue Fairbanks. �207 Colorado Division of Wildlife Wildlife Research Report July 1992 JOB FINAL REPORT State of Colorado Project No. W-lS3-R-4 Mammals Research Work Plan No. 3A Pronghorn Investigations Job No. 4 Statewide Pronghorn Management Guidelines Period Covered: July 1, 1991 - June 30, 1992 Author:· T.M. Pojar Abstract The general public, pronghorn hunters and Division of Wildlife employees were surveyed to get their opinions on pronghorn management issues. The survey was conducted by Standage Accureach. Issues that surfaced during general public and Division of Wildlife focus groups are: 1. Hunting pronghorn during the breeding season. 2. Season structure for reducing hunter density. 3. Limiting archery licenses on a statewide basis. The results from the survey strongly indicate that, although these may be issues, they are not issues that demand any immediate corrective action. In general, all groups are satisfied with the present management and season structure for pronghorn in the state. Based on these'resu1ts, the Wildlife Commission has opted to maintain the current hunting aaasonts trructnrrefor the next 3 years. A document titled Pronghorn Management Analysis Guide, 19921994 (third draft) is available from the author. This document outlines the history of pronghorn management ,~n Colorado and provides an analysis of the management environment. The issues are examined in terms of the social, political, economic and biotic environment. Prepared b~?f!d Wildlife Researcher C ��209 Colorado Division of Wildlife Wildlife Research Report July 1992 JOB PROGRESS REPORT State: Colorado Project Number: W-l53-R-5 Mammals Research Work Plan No. : Pronghorn Research 3A Job No.: Effects of Density on Dispersal Rates Period Covered: Author: July I, 1991-June 30, 1992 Tanya M. Schenk Personnel: Gary C. White, R. Bruce Gill DRAFT PROGRAM NAlUlATlVE A. NEED An animal population is said to be density dependent if its rate of growth is dependent on its size. Thus, density dependence is a population process. The mechanism of density dependence is intraspecific competition (Nicholson 1957). However, the patterns of expression of intraspecific competition within a population are numerous. For example, an increase in intraspecific competition as a result of increased density may result in decreased birth rate or increased death rate. Decreased birth rate may be from an overall decrease in female fecundity, or may result from only a decrease in yearling fecundity. Similarly, death rate may increase because of differential survival based on sex or age structure, or may affect the population randomly. However, regardless of the specific pattern, the ultimate response to density caused intraspecific competition will be a change in birth. death,' i~igration, and/or, emigration r,ates. Therefore, detection of density dependence in natura1populatforts would provide bio1bgists with critical information to be used for management or to provide -insight into the patterns of population demographics. For example, the concept. of compensatory mortality has been one of the .. most studied density-dependent processes (Errington 1945, Anderson and Burnham 1976, Burnham and Anderson 1984, Burnham et a1. 1984, Bartmann et a1. 1992). Verification of compensatory mortality operating within a population supports the idea of a harvestab1e surplus and maximum sustained yield, both of which have direct management implications for hunted populations. Harvest limits and management for desired population age structures can then be based on predicted responses to density dependence. ��211 Density-dependent caused intraspecific competition could increase the possibility of differential survival of individuals. For natural selection to be occurring this must also translate into differential survival of selected genotypes. Thus, a study to verify active density dependence occurring in a population could provide evolutionary insights into questions such as whether natural selection will favor efficiency or speed of resource acquisition. Conservation biologists must incorporate the potential effects of densitydependent feedback processes when estimating required refuge sizes needed to sustain minimum viable populations. That is, they must predict at what density the minimum number of individuals can exist and reproduce sufficient offspring to maintain a viable population. Ginzburg et a1. (1990) modeled the demography of Gadus morhua (cod) to show the relationship between density dependence and extinction risks and found the probability of extinction to depend rather sensitively on the density-dependence relationship. Persistence time for the species was increased when density dependence was active in the model because the growth rate increased as the population size decreased. Thus, the population was more likely to recover and return to equilibrium. Conversely, management of habitat fragmentation also involves densitydependent considerations because of its influence on dispersal and population size through reduced immigration and survival rate (Wilox and Murphy 1985). Fragmentation also indirectly influences survival and recruitment through increased pressure from the surroundings, biotic factors such as predators, competitors and parasites, or abiotic edge-effects such as changed wind and light conditions (Ro1stad 1991). Metapopu1ations may arise from fragmentation as long as the habitat fragments are large enough to persist as local subpopulations connected by dispersal. Numerous studies have attempted to relate any number of habitat variables with habitat use. Fretwell and Lucus (1976) developed one of the earlier models of habitat selection that incorporated the idea of density as a factor in habitat selection. The model predicts that as density increases habitat selection is altered because the "preferred" habitat becomes saturated. The ability to detect density dependence in a population would then allow the biologist to interpret the results of their habitat selection studies in light of density stress. For example, Wiens (1976) manipulated the habitat of birds associated with sagebrush, attempting to detect a difference in territory size following a reduction in sagebrush. Unable to detect a difference, he suggested the study was conducted at the wrong scale. Another possible explanation for detecting no difference in territory size was that sagebrush was not a limiting factor, thus the birds were not under density stress even after the removal of a proportion of the sagebrush. It is essential to have an understanding of the natural density-dependent regulating factors operating in a pest species before any attempt is made at control. Displacement of the natural regulating factor in a pest species with an introduced regulating factor has the potential to result in a serious outbreak of the pest which is likely to reach a density far above that which existed before the control was applied (Nicholson 1954). If the pest species is growing logistically, the mean per capita growth rate (r) will be exponential below a particular threshold (the inflection point)~ Unless the threshold could be artificially lowered, control of the pest species may be ineffective until the threshold density is reached. Understanding what sets �212 the threshold may enable an artificial lowering of the threshold and control may be possible at lower densities. Cyclic behavior in population densities may arise from time delays in the reaction of negative feedback processes when the delay is longer than the natural periodicity of the system (Berryman 1978). Since time-delays feature extensively in biological growth and control mechanisms, studying how changing the size and type of time-lag affects the dynamics of population growth yields valuable insight into the underlying biological process. As numerous and varied as the reasons are for wanting to detect density dependence in a natural population an equal number of difficulties arise to deflect our ability to do so. These difficulties can be classified into 5 major masking effects: (1) confounding population parameters, (2) environmental covariates, (3) innate heterogeneity of the population growth parameters, (4) sampling variance of population parameters, and (5) time delay between the density stress and the density response. The first masking effect, confounding population parameters, arises because density-dependent responses may act differentially by age or sex. Thus, data collected on an entire population may swamp out the effect of a densitydependent response in a specific age class. Vickery and Nudds (1991) detected density dependence in 8 populations, but did not detect it in all data sets coming from each of the 8 populations. They suggest that population estimates taken at certain life stages may be far from showing any effects while at other stages the density-dependent effect is detected. A second possible masking effect, environmental covariates, occurs when field data on demographic parameters are confounded with density-independent factors such as weather .. Thus, density-dependent births and deaths would be confounded with births and deaths unrelated to density. Royama (1977) points out that before attempting to test for density dependence these two types of factors need to be separated. Dispersal (immigration and emigration) is also subject to a combination of a density-independent as well as a densitydependent responses. Instincts and adaptations for dispersal are universal among animals and dispersal is going on all the time from populations both dense and sparse (Andrewartha and Birch 1954). Adjusting for these environmental covariates could clarify the underlying density-dependent feedback processes. Thirdly, individual heterogeneity within even a given sex or age class exists in the 4 population growth parameters (birth, death, immigration, and emigration rates). Explicit density dependence is defined as per capita birth, survivorship, and dispersal responding deterministically with increasing density and implies regulation toward a central equilibrium point (Strong 1986). Density-vagueness implies population change away from the two extremes of extinction and immense densities (Strong 1986). Vague density relationships therefore, show a decrease in per-capita performance at high densities, but emphasize a lack of change in performance over intermediate densities, where many population spend most time. Statistically, density vagueness emphasizes the innate heterogeneity associated with demographic parameters affected by density. This innate variance will further obscure the function between demographic parameters and density. �213 Spatial heterogeneity of subpopulations, as exists for Detapopulations also confounds our ability to detect density dependence (De Jong 1979, Hassell 1985, Hassell and May 1985, Hassell et al. 1987, Mountford 1988, Stewart-Oaten and Murdoch 1990). Hassell et al. (1987) were unable to detect temporal density dependence for 16 generations of viburnam whitefly (Aleurotrachelus Jelineki1). However, when 9 of the sixteen generations were studied in a more detailed analysis (leaf by leaf), density dependence was detected in 8 of the 9 generations. The heterogeneity of the subpopulations masked the effect of density dependence on the population as a whole. The fourth masking effect is the sampling variance associated with the estimates of the parameters used in the tests to detect density dependence. If density dependence is to be detected in natural populations we are forced to work with estimates of our parameters as few natural populations are able to be censused. Also, logistics often demand the parameters of birth and immigration be inseparable as recruitment and emigration and death lumped as survival. Pollock et al. (1992) and Nichols and Pollock (1990) however, have recently developed a method for separating in-situ reproduction numbers and immigration numbers. Detailed information on individual animals yields the most precise estimate of the 4 growth parameters. However, few natural populations could be studied in such detail as to yield complete knowledge of the fates of each individual. Thus, the population parameters would be estimates of birth, death, immigration, and emigration rates. The final masking effect to be addressed is time delay between the density stress and the population response to density. Thus, feedback processes do not affect the next immediate generation but will affect future generations. Regardless of the functional response, the ultimate population response to density stress is change in birth, death, immigration, and/or emigration rates. To overcome the difficulties inherent in attempting to detect density dependence in natural populations we must be able to identify the functional relationship between population size and birth, death, immigratipn, and emigration rates. Models have been developed to describe both density-dependent and densityindependent population growth. A model of density-independent population growth is: N~ - NQ(r + l)~ where Nt is population size at time t, No is population size at time zero, and r is the per capita population growth rate (Pielou 1974). In contrast, an animal population is said to be density-dependent if its growth rate is negatively correlated with its size. Pielou (1974) describes density-dependent population growth as: NI' (r-1)N N,+1= 1+ K ' �214 where Nt+1 is population size at time t+l; Nt is population size at time t; r is the per capita population growth rate as Nt approaches 0, and K is the carrying capacity. It has been generally accepted for several decades that density dependence can be detected by the analysis of long-term life-table data that provide estimates of temporal trends in the density of different life-history stages (e.g. Varley and Gradwell 1960; Eberhardt 1970; Bulmer 1975; Royama 1977; Slade 1977; Berryman 1978; Vickery and Nudds 1984; Pollard et al. 1987; Reddinguis and den Boer 1989; Turchin 1990). However, temporal trends in population density may not be sensitive enough to allow for detection of density dependence (e.g. Strong 1986; Gaston and Lawton 1987; Hassell et al. 1987; Lomnicki 1987; Mountford 1988; Solow 1990; Bartmann et al. 1992). As discussed above, the influence of confounding population parameters such as age and sex, environmental factors such as weather, natural heterogeneity of the population parameters influencing density such as birth, death, immigration, and emigration rates, and sampling variance all serve to mask the effects of density within a population. Therefore, the detection of density dependence would be enhanced by using models sensitive to these potential masking effects. Also, if the response of a population to its density is expressed in changes in birth, death, immigration, and/or emigration rates, it follows that a more sensitive test for the detection of density dependence could be constructed using these more specific parameters of population growth. This follows because it is not population size we are interested in per say but rather the per capita rate of growth of the population. Thus, population size at time t+l (Nt+1) would be better represented as a function of the number of births (Bt), deaths (Dt), immigrants (It), and emigrants (Et) in a population at time ~ given the population size at time ~ (Nt). Thus If density-dependent growth is operating within a population then at least 1 of the growth parameters (birth, death, immigration, and emigration rates) will be a function of Nt. For example, if the number of births (Bt) is density-dependent then Bl - b1(N1)N1 where bt is the mean per capita birth rate at time t and is a function of Nt. Then Nt+1 would be estimated as However, if: where b is the mean per capita birth rate, is a constant for all Nt then the mean per capita birth rate is independent of density. Similar arguments could be made for dt, it, and et. Thus, density dependence is operating if mean per capita birth, death, immigration, and/or emigration rates at time ~ are functions of Nt and the following hypotheses can be tested to detect density �215 dependence in a population: H1:bt H2:dt b(Nt) d(Nt) H3:it i(Nt) H4:et e(Nt) In other words, Hl describes a density-independent birth process if bt is constant for all Nt because the mean per capita birth rate remains constant for all population densities. If we define r, the mean per capita population growth rate, as: r - b - d + i - e and substituting this into equations land 2 for describing density independence and density dependence respectively then eq. 1 can be rewritten as: and eq. 2 can be rewritten as: N,(b-d+i-e) N'+1------~---------~-(b-d+i-e-1)N 1+ K ' These equations, although more explicit by breaking r into its 4 component parts, are still too simple to describe natural populations. More realistic models can be developed to mimic natural populations by incorporating the possible masking effects to density dependence as described above. For example, incorporating a population parameter (P), such as sex or age, as a covariate of the components of r would yield the following functions: s, - b(P)N'flll D~ - d(P)Nlill. I~ - i (P)NE,ill. e(P)N'flll E; - where P is any appropriate population parameter that covaries with birth, death, immigration, and/or emigration rates and Np(t) is the number of individuals in the population at time t with the population parameter of interest. For example, mean per capita birth rate (b) may vary with age of the female. If b~ represents the mean per capita birth rate of subadu1t females and bA represents the mean per capita birth rate of adult females, then �216 where NMCt) is the number of adult females in the population at time t and N~(t) is the number of subadult females in the population at time t. Then Nt+1 would be estimated as Density dependence may be expressed in one, the other, or both age classes. If density dependence is expressed in only the subadult age class, blocking by age would eliminate the swamping effect of collecting data on a mean per capita growth rate of adults and subadults combined. Thus, age can be used as a block to construct a more sensitive test for density dependence and the following hypotheses would be tested: H1:bACt)- bA(Nt) H2:b~Ct) - b~(Nt) d(Nt) H3:dt i(Nt) H4:it e(N Hs:et t) Therefore, if bACt) is constant for all Nt then the population would have a density-independent birth process for adult females. Environmental covariates, such as weather, may act in a density-dependent manner on the population growth parameters by altering carrying capacity (K). The response may be a change in mean per capita birth, death, immigration, and/or emigration rates. If a population is susceptible to density stress then the introduction of an environmental covariate (E) in year t would alter the value of Kt. The population response would be a change in b, d, i, and/or e and the following hypotheses could be tested H1:bt H2:dt H3:it H4:et b(E) d(E) i(E) e(E) If, for example, bt remains constant for all values of E then the population birth process would be density-independent. Innate heterogeneity of population growth parameters among individuals within a population was the third potential masking factor discussed above. The per capita growth parameters b, d, i, and e have been expressed as means, however, none of the models described above have accounted for the variability around these means. To account for individual heterogeneity in the population growth parameters an error term should be included when calculating the number of births (Bt) , deaths (Dt), immigrants (It), and emigrants (Et). Therefore, Bt - D; - I; - E; - bNt + E~ dN; + E~ iN~ + E,!!! eN; + E!I!. where, for example, b is the mean per capita birth rate of the population and EbB is the stochasticity of the mean per capita birth rate attributed to heterogeneity among individuals. �217 To incorporate sampling variation of the parameters used in the models an error term must be included and the equations will be as follows: Bt D; It E~ + dN + iNt + eN~ + bNt 1 E~ E~ Eit E.!!;_ Where Ebt is the sampling variance associated with measuring the mean per capita growth rate (b) at time t. The final masking effect discussed was time delay in the population response to density stress. To model time delay the population growth rates at time t would be functions of density at time t-k where k is the number of generations needed for the density response to be exhibited in the population. The following hypotheses would then test for density dependence in b, d, i, and e: H1:Nt H2:bt - H3: c4. H4: it Hs:et - r(Nt-k) b(Nt-k) d(Nt-k) i(Nt-k) e(Nt-k) Any or all combinations of these masking factors (population parameters, environmental covariates, heterogeneity, and sampling variance of population growth responses) could be operating in a natural population. Thus, an appropriate model may be: where the number of births at time t (Bt) is calculated as the mean per capita birth rate (b) as a function of density (Nt) at time t, population parameter P (e.g. age), and an environmental covariate (e.g. weather) plus the innate and sampling error associated with measuring the mean per capita birth rate at time t. Models developed to m1m1C natural population response to density would select either population size or any combination of the remaining population responses as the predictor of density dependence and incorporate any combination of confounding effects (Table 1). Table 1. Summary of possible population responses to density dependence and potential masking effects. Population response Masking effect size population parameters birth rate environmental covariates death rate heterogeneity immigration rate sampling variance �218 time delay emigration rate A review of the literature reveals that most techniques developed to test the null hypothesis of density independence versus the alternative hypothesis of density dependence rely on series of population size estimates as the population response and no masking effects. Thus, these techniques have focused on the simplest of the models, testing only if Nt+1 is a function of Nt. However, there have been a few models developed to address 1 or more of the potential masking effects. Below is a summary of published tests categorized by the parameters included in the models to detect density dependence. Single response variable models with no masking effects: The most common approach used to detect density dependence in natural populations has been to regress the log population size at time e+l (~+1) on the log population size at time t (~) (Morris 1959). If populations increase or decrease independent of population density, their change through time should follow the simple equation for exponential growth: Xlli - r + PX1 Density independence exists if the plot yields a straight line with a slope (P) of 1, indicating per capita population growth rate (r ) is constant. If density dependence is occurring in the population the plot should result in a line with , < 1. Morris's method has however, been shown to become unreliable as random variation increases (Eberhardt 1970, St. Amant 1970, Maelzer 1970, Bulmer 1975). Although the logic of Morris's technique appears correct, the estimation of P by the regression coefficient b is flawed because estimates of population size are used. Thus, .~ ft~ ,-1 i-1 b=:E (x,-m1)(X'+1-"'2)/:E (x,-m1)2 where 11-1 m1=Lx/(n-1) i-1 and 11-1 ~=:Ex'+1/(n-1) '-1 �219 and even if the true slope p were to be 1, the estimate b (of because of random variation in data (Pollard et al. 1987). P) may be < 1 Slade (1977) proposed calculating the principal axis coefficient as an alternative technique to Morris's (1959) regression analysis when 2 variables have equal variance. Slade (1977) proposed the use of the slope of the principal axis (Pp) of a bivariate scatter plot of xt and xt+l should be less sensitive to random variation in the data and thus be a better test of detecting density dependence. As in regression analysis, if pp - 1 then there is no density dependence, if pp < 1 then there is evidence of density dependence in the population. Bulmer (1975) approached Morris's linear regression problem as a time series analysis and proposed two autoregression methods for the detection of density dependence describing density independence as: and density dependence as: X£tl - ~ - a(X~- ~) + E~ where xt and xt+l are the natural logarithms of population size at time t and t+l respectively, Et'S are independent normal random variables with mean zero and variance 02, ~ is an equilibrium value, and lal < 1. Thus, for a densityindependent population there will be no deterministic trend in population size, nor will there be any tendency for it to return to an equilibrium value. Alternatively, in the density dependent model population size will constantly return to the equilibrium value (Bulmer 1975). The models are then used to test the null hypothesis of density independence against the alternative model of density dependence based on observations N1, ••• Nn• The test statistic is: R-V;U where: .-1 u=E (X'+1-XJ2 i-1 y=E• (X;_X·)2 ,·1 The null hypothesis is rejected for small values of R. �220 Solow (1990) tested the robustness of this model to departures from the null and alternative models. Solow (1990) simulated data from Bulmer's (1975) density-independent model modifying the E'S to have serial correlation: where lei < 1 and the ut are independent normal random variables with mean zero and variance (1-e2)o2. This departure from the null model caused a significant decrease in power of the test. He also simulated data from Bulmer's (1975) density-dependent model. however a was made dependent on Ixt - pi: a - exp(-a(X~ - p)!) and found as a increased the test lost power. Bulmer's autoregression tests have two test criteria: the first. R. which does not allow for measurement error and the second. R! which is more robust to measurement error. Bulmer (1975) states that these methods are only suitable for long runs of data and that R is to be preferred over R! as the test criterion unless significant measurement errors are suspected in the data. Reddinguis (1971) developed a permutation test that may be used as a nonparametric alternative to Bulmer's test. The permutation test uses the rank of the log-range [LR - log(highest density) - log(lowest density)] of the field data within the collection of log-ranges obtained from all possible permutations of the coefficients of net reproduction that were derived from the field data (Reddinguis and den Boer 1989). Thus. in the case of density dependence the actual sequence of net reproduction values can be expected to have a lower log-range than in a significant majority of the permuted series. Let Ni• N2 ••••• Nn denote population densities at equidistant times t - 1. 2 •...•n. Let xt - lnNt• t-l. 2 •...•n. and let Yt - xt+i - xt - InRt for t1.2 •...n-l. The null hypothesis of the permutation test states that the order in which the Y's occur is random. the alternative hypothesis is that the Yvalues occur in such an order that the resulting fluctuations of the Xsequences are restricted (Reddinguis and den Boer 1989). To test the null hypothesis a random sample of the possible permutations of the observed Y's is taken. and the X's and LR's computed. If the original LR is a random drawing from the same population as the sample of LR's obtained by permutations then we fail to reject the null hypothesis. This is a special case of the Mann-Whitney test where one of the two samples to compare has size 1. Let the observed LR be denoted as Lo. and the first. second •...• k+lat order statistic of the combined set of LR's be Li• Lz •...• Lt+i. Let r be the rank of Lo in the sequence of Li's (i-l.2 •...k+l); then if r/(k+l) Sa. where a is the type I error rate. the null hypothesis is rejected (Reddinguis and den Boer 1989). Pollard et al. (1987) also proposed a randomization method to test if an observed set of xt's are random. where X~ - log N~ (t-1,2, ... ,n). �221 To test the null hypothesis of random (density-independent) ~'s first use the observed values Xl. Xz •...• Xn to compute the value of the test statistic T. an appropriate likelihood ratio test-statistic (see Pollard et al. 1987). Either calculate the dt - (~+1 - ~) values and enumerate the (n-l)1 possible sets of ~ values. or randomly permute the dt values. and corresponding to N such permutations. obtain a sample of N simulated sets of ~ values. For each simulated set of ~ values the test statistic T is computed. If less than 5% of the T values calculated under the simulated ~ values are smaller than or equal to the computed T value under the observed ~ values. then reject the null hypothesis of density independence at 5% level of significance. Bartmann et al. (1992) experimentally manipulated fawn mule deer (Odocoileus hemionus hemionus) population densities and found a strong density-dependent mortality process existed. This experimental approach with temporal and spatial controls provides a cause and effect relationship that does not exist in the previously described tests. Population parameters as covariates: Any of the tests developed to detect density dependence in natural populations can be modified to include population parameters as covariates. If data are collected to include population parameters such as age and sex that may respond to density differentially then any test can be conducted for that particular segment of the population that may be exhibiting density dependence. For example. Lebreton et al. (1992) tested for a possible density effect on survival (.) of roe deer (Capreolus capreolus) using SURGE. an iterative model fitting computer software package for the estimation of survival probabilities from mark and recapture (or resighting) data. Data were collected such that sex and age could be tested as covariates effecting survival. as well as density. The general model included age. sex. and density as covariates effecting survival. This general model was tested against a simpler model using only age and sex effects to estimate survival. Because the more general model was selected there was evidence that density, age, and sex all had significant effect on survival rate. Had the simpler model been selected there would have been no support for density effecting survival rate. Clutton-Brock (1987) used logistic models (Cox 1970) to investigate how changes in population density interacted with age. reproductive status, dominance rank, and matriline size to affect fecundity or calf survival in a resource limited population of red deer (Cervus elaphus). This technique converts binary data into probability values by fitting a logistic curve through the available points (Clutton-Brock 1987). Only variables that significantly improved the fit were considered to interact with population density. The parameters of the logistic model were estimated by maximum likelihood. A goodness-of-fit of a model was determined by comparing the difference between deviance values of 2 models which were distributed approximately as X2 with degrees of freedom equivalent to the differences in number of parameters fitted in each model (Clutton-Brock 1987). �222 Environmental factor. as covariate.: The statistical technique for detecting key factors and density-dependent effects is k-factor analysis· (Varley and Gradwell 1960). Factors which playa key role in population change are those sources of mortality which are correlated with total mortality (Morris 1959). Thus, key factors are sources of mortality that contribute most to perturbation of densities away from population equilibria. Density-dependent factors are those that tend to occur aore heavily at higher densities tO,bring the population back toward an equilibrium . Density dependence is detected from plots of individual k-factors against the logs of the densities of the life stages on which they act. Significant correlations indicate positive density dependence (Varley et al. 1973). Therefore, the data include population size estimates for t years and annual mortality that has been partitioned into m factors. Sources of mortality are expressed in terms of killing power, k-values, numerically equivalent to the difference between log(t+l) and log(t), the population densities before and after a designated mortality factor has operated. The overall generational mortality is referred to as K, and individual sources of mortality are designated ko, k1, ••• , k_. The key factor is that whose variation makes the greatest contribution to total mortality variation. Thus, Xl+1J=X,J- • J-1 ".j AI.1 E 1U- E 1;+1,1 where Xi,J represents the logarithm of the population density on which mortality factor j is about to act in generation i; and k.w - Xi,J - Xi,J+l. Heterogeneity models: In my review of the literature, I was unable to find any tests for detecting density dependence that addressed individual heterogeneity. However, spatial heterogeneity, as exhibited in metapopulations, and its effect on masking density dependence has been addressed (De Jong 1979, Hassell 1985, Mountford 1988, Stewart-Oaten and Murdoch 1990). These models then express heterogeneity on a larger scale. Instead of individuals within a patch exhibiting heterogeneity in the population growth parameters, the metapopulation is a group of subpopulations within the population and heterogeneity exists among subpopulations. A more realistic model would include both levels of heterogeneity - the subpopulation and individuals within each subpopulation expressing heterogeneity in growth parameters. Because the heterogeneity among subpopulations may be due to different densities within the subpopulations, simpler models of the dynamics of animal numbers are no longer applicable, as they assume that the density experienced by each individual is the same (De Jong 1979). The focus however, has been on modeling various forms of heterogeneity rather than developing tests sensitive to this masking effect. For example, Mountford (1987) developed a model of a univoltine insect population incorporating non-random dispersal of adults and larval density-dependence to test if heterogeneity obscured the detection of density dependence. To test �223 for density dependence he used the permutation test developed by Pollard et al. (1987). De Jong (1979) explored the consequences of the subdivision of a population of juveniles confined to patchy food sources and dispersing adults on the dynamics of the entire population. He developed models with varying probability distributions of dispersers and juvenile density-dependence. Hassell (1985) and Hassell et al. (1987) suggest that inter-generation density dependence can be obscured by limited amounts of heterogeneity. When heterogeneity was introduced into the models, the populations fluctuated wildly and no direct density dependence could be detected with k-factor analysis. However, Dempster and Pollard (1986) repeated Hassell's (1985) computation with different heterogeneity values and found the k-value showed anticlockwise spiralling, characteristic of delayed density dependence. Stewart-Oaten and Murdoch (1990) reevaluated the data from Hassell et al. (1987) and concluded that temporal as well as spatial density dependence could be detected. Models with variation in population parameters: Bulmer's (1975) autoregression test as described above was expanded to account for measurement error of the parameters (N) used in the in the models. Bulmer (1975) however, found that including sampling variance reduced the power of the test. Thus, this model was only suitable for long runs of data. Models with delayed density dependence: A time-delayed logistic model was presented by Berryman (1978) to demonstrate density dependence in a population when the delay is longer than the natural periodicity of the system. The model to describe time-delayed density dependence is: where Nt is the population size at time t, 1 is the maximum per capita recruitment rate per generation, and f(ut) is a density-dependent feedback function in which ut is the density of the population at time t. Any number of functions could be used to describe f(ut). For example, if an exponential decay function is substituted for the function f N 1=N1-1At! -""-1 where a is a rate parameter. To define a, set Nt - Nt-1 - N*, where N* is the steady state population density. At equilibrium a=ln'J./N* �224 N =N t l'-(N,_,IN") t-1 If there is a time delay of 1-year in the density-dependent then N =N t t-1 feedback process l'-(N,.JN") When 1 < 2.72 (1n1 < 1) the model has a stable point, N*, but when 1 > 2.72, it exhibits stable limit cycles with frequency determined by 1 (Berryman 1978). Turchin (1990) developed 2 methods for detecting delayed density dependence. The first method used the diagnostic techniques of time-series analysis and: x~ - 8Q + 81Xkl + 8~t-2 + ... + ~b + E~ where Xt - log Nt is the log-transformed population density at time t, Xt-l log Nt-l, al...Bp are weights that quantify the influence of past densities on the population change, and Et is a random variable with mean zero and variance a2. This approach allows for multiple lag-effects but requires a linear relationship among the log-transformed population densities. Therefore, the second approach specifies a non-linear model for Nt but considers only 2 time lags: The parameters of the model roo al. Q2. and Et are estimated by regressing the rate of population change r - 10g(Nt/Nt-l) on Nt-l and Nt-2• Each series of population censuses are analyzed using a stepwise regression: first regressing r on Nt-l, and then testing whether adding the term Nt-2 significantly reduced the unexplained variance (Turchin 1990). The ability of these proposed methods to detect density dependence has however, proven to be controversial (Strong 1986, Gaston and Lawton 1987, Lomnicki 1987, Pollard et a1. 1987, Reddingius 1990, Solow 1990, Vickery and Nudds 1991). For example, most techniques are unsuitable for detecting delayed density dependence, and may underestimate the overall level of density dependence (Hanski 1990). The chances of detecting density dependence increases with increasing number of years of data, thus some data sets may range over too brief a time span (Hassell 1986, Solow and Steele 1990). Power to detect density dependence is also decreased if the parameters are estimated (Bulmer 1975). Scale at which such density dependence can operate also affects the performance of density-dependence tests (Hassell 1986). Further difficulties arise in detecting density dependence from the models themselves. As Vickery and Nudds (1984) suggest the current available tests for density dependence address the question "Are density-dependent effects strong enough in the population to outweigh the counteracting effects of inverse density dependence and the stochastic effects of density �225 independence?' The tests do not address the question, "Do density-dependent effects operate in this population?" To better answer the first question we could improve tests for detecting density dependence by addressing the 4 population growth parameters (birth, death, immigration, and emigration rates) instead of simply population size, including population and/or environmental covariates, account for heterogeneity of population growth parameters, improve estimates of population parameters used in the tests or adjust for the sampling variance, and testing for delayed density responses. To truly answer the latter question requires experimental manipulation of the population with detailed information of the effect of the manipulation on birth, death, immigration and/or emigration rates. Few studies have quantitatively assessed the numerous difficulties inherent in attempting to detect density dependence in natural populations. Yet, management and research agencies continue to collect information such as population size estimates, reproductive effort, and harvest and attempt to relate these factors within a density-dependent framework. Rigorous testing is needed to determine the types of information necessary to detect density dependence and the effect of innate and sampling variance associated with these parameters on the robustness of these models. Thus, the primary objective of this study is to quantitatively assess the feasibility of detecting density dependence in natural populations. Monte Carlo simulation are the mos't appropriate approach to evaluate tests for the detection of density dependence because the underlying model of the data is known. Tests can then be evaluated on their robustness to detect the true model. The current controversy surrounding the effectiveness of tests for detecting density dependence stems primarily from using field data in the evaluations where the true underlying models are unknown. Without- this information there is no way to truly evaluate the given test. Thus, simulated data from a known model will be used to evaluate the robustness of tests developed in this study as well as tests currently in the literature to detect density dependence in natural populations. Within each simulation, population size and number of years of data will vary. B. OBJECTIVES 1. Evaluate the potential of detecting the presence or absence of density dependence in a natural population by testing if popUlation size (Nt+1), birth (bt), death (dt), immigration (it), and/or emigration (et) rates are functions of population density (Nt). �226 2. 3. 4. Evaluate the efficiency of detecting the presence of density dependence when population parameters are confounding factors in response to density. Hz.: Tests perform acceptably in the detection of density dependence when population parameters are confounding factors of birth rate in response to density. H~: Tests perform acceptably in the detection of density dependence when population parameters are confounding factors of death rate in response to density. Hzc: Tests perform acceptably in the detection of density dependence when population parameters are confounding factors of immigration rate in response to density. HZd: Tests perform acceptably in the detection of density dependence when population parameters are confounding factors of emigration rate in response to density. Evaluate the efficiency of detecting the presence of density dependence when environmental covariates are confounding factors in response to density. H3.: Tests perform acceptably in the detection of density dependence when environmental covariates are confounding factors of birth rate in response to density. H3b: Tests perform acceptably in the detection of density dependence when environmental covariates are confounding factors of death rate in response to density. H3c: Tests perform acceptably in the detection of density dependence when environmental covariates are confounding factors of immigration rate in response to density. H3d: Tests perform acceptably in the detection of density dependence when environmental covariates are confounding factors of emigration rate in response to density. Evaluate the efficiency of detecting the presence of density dependence when heterogeneity confounds the response to density. H4.: Tests perform acceptably in the detection of density dependence when heterogeneity confounds birth rate in response to density. H4b: Tests perform acceptably in the detection of density dependence when heterogeneity confounds death rate in response to density. H4c: Tests perform acceptably in the detection of density dependence when heterogeneity confounds immigration rate in response to density. �227 H4d: 5. 6. Tests perform acceptably in the detection of density dependence when heterogeneity confounds emigration rate in response to density. Simulate a range of variances in the estimates of the parameters (population size and birth, death, immigration, and emigration rates) used to test for detecting the presence of density dependence to evaluate the robustness of the test. Hs.: Tests perform acceptably despite variance in population size estimates used to conduct the test. H~: Tests perform acceptably despite variance in birth rate estimates used to conduct the test. HSc: Tests perform acceptably despite variance in death rate estimates used to conduct the test. HSd: Tests perform acceptably despite variance in immigration rate estimates used to conduct the test. Hs.: Tests perform acceptably despite variance in emigration rate estimates used to conduct the test. Evaluate the efficiency of detecting the presence of density dependence when time delays occur between the density stress and the density response in population size, and birth, death, immigration, and emigration rates. H6.: Tests perform acceptably in the detection of density dependence when a time delay occurs in the population size response to density. H6b: Tests perform acceptably in the detection of density dependence when a time delay occurs in the birth rate response to density. H6c: Tests perform acceptably in the detection of density dependence when a time delay occurs in the death rate response to density. H6d: Tests perform acceptably in the detection of density dependence when a time delay occurs in the immigration rate response to density. H6.: Tests perform acceptably in the detection of density dependence when a time delay occurs in the emigration rate response to density. 7. Repeat objectives 1-5 with models developed during the study. 8. Develop guidelines on the design and methods of studies to detect the presence of density dependence in a population. �228 C. EXPECTED RESULTS OR BENEFITS Detecting density dependence within a population could provide detailed information to be used as a management tool in setting harvest limits, altering population age structure, identifying limiting habitat .features, pest control, and conservation issues. Evaluation of the efficiency of various tests of density dependence will provide guidelines for the feasibility of such tests to detect density dependence in a natural population. Evaluation of these tests will also determine if the data collected by the eDOW on large ungulate populations is sufficient to detect density dependence. D. APPROACH Simulations The Statistical Analysis System (SAS) will be used with all Monte-Carlo simulations. Four popUlation sizes (50, 100, 500, 1000) and 4 data set lengths (5, 10, 20, and 40 years) will be used in all simulation sets. Objective 1 - Evaluate the potential of detecting the presence or absence of density dependence in a natural population by testing if population size (Nt+1), birth (bt), death (dt), immigration (it), and emigration (et) rates are functions of population size (Nt). The first objective is to evaluate the overall performance of tests for the presence of density dependence given that all assumptions of the various tests are met. The null hypothesis to be tested is that density dependence will be detected in a population if it exists and the absence of density dependence will likewise be detected. The detection can occur within any of the 5 population responses: population size, birth, death, immigration, and/or emigration rates. Data will be generated under conditions of density dependence and density independence. These data will be evaluated by each test and results tabulated as either correct or incorrect. A correct result is one where density dependence is confirmed by the test when it does in fact exist in the data or when density dependence is not detected by the test and there in fact is no density dependence in the data. Simulations will be run with 1000 sets of generated data for both density dependence and density independence. Numbers of correct and incorrect results will provide an estimate of efficiency for each test. Objective 2 - Evaluate the efficiency of detecting the presence of density dependence when population parameters are confounding factors in response to density. The second objective consists of evaluating the performance of the tests for the detection of density dependence when population parameters are confounding factors in response to density. Data will be simulated with differential density responses first by sex, then by age (juvenile, subadult, and adult). Each simulation will be conducted for 1000 data sets for each test of detection of density dependence. Numbers of correct and incorrect results will provide an estimate of precision for each test. Objective 3 - Evaluate the efficiency of detecting the presence of density dependence when environmental covariates are confounding factors in response to density. �229 The third objective addresses the effect of environmental covariates in response to density and the performance of the tests for the detection of density dependence. Data will be simulated with a density dependent covariate (i.e. weather). Each simulation will be conducted for 1000 data sets for each test of detection of density dependence. Numbers of correct and incorrect results will provide an estimate of robustness for each test. Objective 4 - Evaluate the efficiency of detecting the presence of density dependence when heterogeneity confounds the response to density. The fourth objective is to introduce heterogeneity into birth, death, immigration, and emigration rates and evaluate the performance of the tests for the detection of density dependence. Data will be simulated with heterogeneity in the various population growth responses. Each simulation will be conducted for 1000 data sets for each test of detecting density dependence. Numbers of correct and incorrect results will provide an estimate of robustness for each test. Objective 5 - Simulate a range of variances in the estimates of the parameters (population size and birth, death, immigration, and emigration rates) used to test for detecting the presence of density dependence to evaluate the robustness of the test. For objective 5, adding variance to the parameters used in the various tests, the density dependence tests will be evaluated on their ability to detect density dependence when it exists in the data. The coefficient of variation of 5 parameters (population size, birth, death, immigration, and emigration rates) will be set at 5, 10, 20, and 40%. Each simulation will be conducted for 1000 data sets for each test of detection of density dependence. Numbers of correct and incorrect results will provide an estimate for each test. Objective 6 - Evaluate the efficiency of detecting the presence of density dependence when time delays occur between the density stress and the density response in population size, and birth, death, immigration, and emigration rates. The sixth objective addresses the robustness of the tests to detecting density dependence when there is a time delay between the density stress and the population response to density. Data will be simulated with time delays of I, 2, and 3 generations. Each simulation will be conducted for 1000 data sets for each test of detection of density dependence. Numbers of correct and incorrect results will provide an estimate of robustness of each test. Objective 7 - Repeat the approaches in objectives 1-5 for models developed during the study. Objective 8 - Develop guidelines on the design and methods of studies to detect the presence of density dependence in a population. �230 Schedule Period Activity Design study Complete objective Complete objective Complete objective Complete objective Complete objective Complete objective Complete objective Complete objective [Field work] Completion date 1 5 6 2 3 4 7 8 April - June 1992 November 1992 January 1993 April 1993 July 1993 November 1993 March 1994 November 1994 May 1995 ? May 1995 Personnel Tanya M. Shenk R. Bruce Gill Gary C. White Graduate Research Assistant Co-Principal Investigator Co-Principal Investigator Estimated Annual Cost (01) Personal Services $16,500 (21) Operating Supplies and Services o (28) Travel Expenses o (31) Capital Expenditures o TOTAL EST. ANNUAL COSTS $16,500 FIE Requirements PFTE o TFTE o TOTAL FTE (All work by contract) o 1992-95 Person Days Graduate Student 820 Dr. Gary C. White 200 R. Bruce Gill 100 TOTAL 1120 E. LOCATION The simulations will take place in the Department of Fishery and Wildlife Biology Department at Colorado State University, Fort Collins, Colorado. Literature Cited Anderson, D. R., and K. P. Burnham. 1976. Population ecology of the mallard: VI. The effect of exploitation on survival. U. S. Fish Wi1dl. Serv., Resour. Pub1. 128. 66 pp. Andrewartha, H. G., and L. C. Birch. 1954. The distribution and abundance of animals. University of Chicago Press, Chicago, Illinois. �231 Bartmann, R. M., G. C. White, and L. H. Carpenter. 1992. Compensatory mortality in a Colorado mule deer population. Wildl. Monog. 121. pp. 39 Berryman, A. A. 1978. Population cycles of the Douglas-fir tussock moth (Lepidoptera: Lymantriidae); the time-delay hypothesis. Canadian Entomologist 110:513-518. Bulmer, M. G. 1975. The statistical analysis of density dependence. Biometrics 31:901-911. Burnham, K. P. and D. R. Anderson. 1984. Tests of compensatory vs. additive hypotheses of mortality in mallards. Ecology 65:105-112. Burnham, K. P., G. C. White, and D. A. Anderson. 1984. Estimating the effect of hunting on annual survival rates of adult mallards. J. Wildl. Manage. 48:350-361. Clutton-Brock, T. H., S. D. Albon, and F. E. Guinness. 1987. Interactions between population density and maternal characteristics affecting fecundity and juvenile survival in red deer. J. of animal Ecology 56:857-871. Cox, D. R. 1970. l42pp. The Analysis of Binary Data. Methuen & Co., London, U.K. De Jong, G. 1979. The influence of the distribution of juveniles over patches of food on the dynamics of a population. Netherlands J of Zoology 29:33-51. Eberhardt, L. L. 1970. Correlation, regression, and density dependence. Ecology 51:306-310. Errington, P.L. 1945. Some contributions of a fifteen-year study of the northern bobwhite to a knowledge of population phenomena. Ecol. Monogr. 15:1-34. Fowler, C. W. 1987. A review of density dependence in populations of large mammals. Pages 401-441 in H. H. Genoways, ed. Current mammalogy. Vol. 1. Plenum Press, New York, N. Y. Fretwell, S. D. and H. L. Lucas, Jr. 1970. On territorial behavior and other factors influencing habitat distribution in birds. I. Theoretical development. Acta Biotheoretica 19:16-36. Gaston, K. J., and J. H. Lawton. 1987. A test of statistical techniques for detecting density dependence in sequential censuses of animal populations. Oecologia 74:404-410. Ginzburg, L. R., S. Ferson, and H. R. Akcakaya. 1990. Reconstructibilityof density dependence and the conservative assessment of extinction risks. Conservation Biology 4:63-70. �232 Hanski, I. 1990. Density dependence, regulation and variability populations. Phil. Trans. of R. Soc. Lond. 330:141-150. in animal Hassell, M. P. 1985. Insect natural enemies as regulating factors. Animal Ecology 54:323-334. J of Hassell, M. P. and R. M. May. 1985. From individual behavior to population dynamics. Behavioural Ecology eds. R. M. Sibley and R. H. Smith. pp 322. Blackwell Scientific Publications, Oxford. Hassell, M. P. 1986. Detecting density dependence. TREE 1:90-93. Hassell, M. P., T. R. E. Southwood, and P. M. Reader. 1987. The dynamics of the viburnum whitefly (Aleurotrachelus jelinekii): a case study of population regulation. J of Animal Ecology 56:283-300. Lebreton, J. D., K. P. Burnham, J. Clobert, and D. A. Anderson. 1992. Modeling survival and testing biological hypotheses using marked animals: a unified approach with case studies. Ecolog. Monog. 62:1. Lomnicki, A. 1987. The density dependence debate. TREE 3:76. Maelzer, D. A. 1970. The regression of log Nu+l on log Nn as a test of density dependence: an exercise with computer-constructed densityindependent populations. Ecology 51:810-822. Morris, R. F. 1959. 40:580-588. Single-factor analysis in population dynamics. Ecology Mountford, M. D. 1988. Population regulation, density dependence, and heterogeneity. J. of Anim. Ecology 57:845-858. Nichols, J. D. and K. H. Pollock. 1990. Estimation of recruitment from immigration versus in situ reproduction using Pollock's robust design. Ecology 71:21-26. Nicholson, A. J. 1954. An outline of the dynamics of animal populations. Australian J. Zool. 2:9-65. Pie10u, E. C. London. 1974. 424p. Population and community ecology. Pollard, E., K. H. Lakhani, and P. Rothery. 1987. dependence from a series of annual censuses. Gordon and Breach, The detection of density Ecology 68:2046-2055. Pollock, K. H., W. L. Kendall, and J. D. Nichols. 1992. The "robust" capture-recapture design allows components of recruitment to be estimated. in press??? Reddingius, J. 1971. Gambling for existence. A discussion of some theoretical problems in animal population ecology. Acta Biotheor. 20:1208. �233 Reddingius, J. and den Boer, P. J. 1989. On the stabilization of animal numbers. Problems of testing 2. Confrontation with data from the field. Oeco1ogia 79:143-149. Reddingius, J. 52. 1990. Models for testing: a secondary note. Oeco1ogia 83:50- Ro1stad, J. 1991. Consequences of forest fragmentation for the dynamics of bird populations: conceptual issues and the evidence. in M. Gilpin and I. Hanski eds. Metapopulation dynamics: empirical and theoretical investigations. Academic Press London. Biological J. of the Linnean Society (1991) 42:149-163. Royama, T. 1977. Population persistence and density-dependence. Monographs 47:1-35. Ecological Skogland, T. 1990. Density dependence in a fluctuating wild reindeer herd; maternal vs. offspring effects. Oeco1ogia 84:442-450. Slade, N. A. 1977. Statistical detection of density dependence from a series of sequential censuses. Ecology 58:1094-1102. Sokal, R. R., and F. J. Rohlf. 776p. 1969. Biometry. W. H. Freeman, San Francisco. Solow, A. R. 1990. Testing for density dependence: a cautionary note. Oecologia 83:47-49. Solow, A. R., and J. H. Steele. 1990. On sample size, statistical power, and the detection of density dependence. Journal of Animal Ecology 59:10731076. St. Amant, J. L. S. 1970. Ecology 51: 823-828. The detection of regulation in animal populations. Stewart-Oaten, A. and W. W. Murdoch. 1990. Temporal consequences of spatial density dependence. J. of Animal Ecology 59:1027-1045. Strong, D. R. Turchin, P. lags? 1986. Density-vague population change. TREE 1:39-42. 1990. Rarity of density dependence or population regulation with Nature 344:660-663. Varley, G. C. and G. R. Gradwe11. 1960. J of Animal Ecology 29:399-401. Key factors in population studies. Varley, G. C., G. R. Gradwe11, and M. P. Hassell. 1973. Insect population ecology, an analytical approach. Blackwell, Oxford London. Vickery, W. L., and T. D. Nudds. 1984. Detection of density-dependent effects in annual duck censuses. Ecology 65:96-104. Vickery, W. L., and T. D. Nudds. 1991. Testing for density-dependent effects in sequential censuses. Oeco1ogia 85:419-423. �234 Yi1ox, B. A. and D. D. Murphy. 1985. Conservation strategy: the effects of fragmentation on extinction. Amer. Nat. 125:879-887. \:). ~- ~ Prepared by _.;~''~\) __ ':;_~-=---='-A--==;...._,I...;;.ctk_i(Y Tanya M. Shenk Graduate Student m. ~~ k__ �235 Colorado Division of Wildlife Wildlife Research Report July 1992 JOB PROGRESS REPORT State of ~C~o~l~o~r~a~d~o~_ W-IS3-R-S Project No. Mammals Research Work Plan No. 8A Small Carnivorous Mammals Investi~ations Job No. 1 Development of River Otter Reintroduction Procedures Period Covered: Author: July 1, 1991 - June 30, 1992 T. D. I. Beck Personnel: J. White, G. White Abstract Seven adult river otters (Lutra canadensis) were received in good health from Oregon in 1991 for release into the Dolores River. Three males died prior to release, one of unknown sex escaped, and 2 males and 1 female were released. Dispersion of river otters is widespread throughout the Dolores and San Miguel Rivers. The agreement with Oregon Dept. of Fish & Wildlife for obtaining river otters has been completed. Survival of reintroduced juvenile river otters seems to be less than for adults. A third population estimation procedure was tested for estimating crayfish (Orconectes virilis) density. Mark-recapture estimates indicate a minimum density of 17.9 adults/m2.for a biomass estimate of 2,830 kg/ha. Sightings of river otter at previous release sites suggest the reintroductions have been successful in the short-term and dispersal is occurring. ��237 DEVELOPMENT OF RIVER OTTER REINTRODUCTION PROCEDURES Thomas D. I. Beck P. N. OBJECTIVE Develop procedures for river otter reintroductions in Colorado and establish a self-sustaining population of river otters from which to collect river otters for future translocations. SEGMENT OBJECTIVES 1. Introduce up to 10 river otters into the Dolores River drainage. 2. Develop techniques to monitor survival, reproduction, dispersion, and dispersal of river otters after introduction. 3. Monitor other river otter release sites in Colorado to evaluate success of past reintroductions. METHODS AND MATERIALS Dolores River Release River otters were captured throughout Oregon by personnel of the Oregon Dept. of Fish & Wildlife. The cage-type trap developed by Oregon staff was used. In addition, 2 river otters were donated from a wildlife rehabilitation facility near Grants Pass, OR. The river otters were held in captivity for varying periods ranging from 2 weeks to 2 months for 6 wild animals, and 17 months for the rehabilitated animals. River otters were shipped in transport cages (41 X 41 X 76 cm) via chartered airline to Cortez, CO, on 18 November 1991. Cost of shipment by this means was donated by Evergreen Federal Savings of Grants Pass, OR. The river otters were transported to Dolores and released in holding pens (2.6 X 5.2 m) with running water and food (crayfish and suckers). Prototype radio transmitters from AVM Equipment Co. (26 gms, 70 mm long, 17 mm diameter, 185 mm whip antenna) were surgically implanted subcutaneously along the back by Dr. Mike Miller, DVM, on 19 Nov. 1991. The river otters were transferred to a squeeze box where immobilization was obtained from intramuscular injection of ketamine (100 mg/ml) and xylazine (20 mg/ml) at a dose rate of 15 mg ketamine per kg of animal. A paste made from betadine and petroleum jelly was thoroughly massaged into the incision areas. This allowed the hair to be combed away from the incision line, omitting the need to shave the area. Two incisions were made along the dorsal mid-line. The larger (2-3 cm) was made at a point between the shoulders while the smaller (1 cm) was made 210 mm posterior to the first. A pocket large enough to hold the transmitter was made under the skin at the anterior incision. Blunt, longnecked forceps were used to push under the skin from the posterior to anterior incision. The transmitter antenna was pulled under the skin from the anterior �238 incision. Both incisions were closed with subcuticular sutures and VetSet (a type of super glue). Each river otter was tagged with a Monel #3 metal tag in the web between front toes and was measured for total length; tail length; neck, chest, and head girth; and weight. All injuries were recorded. River otters were placed in transport boxes in a cool room until recovered from anesthesia. They were transported to the Dolores River for release early on 20 Nov. Radio tracking was conducted on foot, by canoe, and from the air. All locations were recorded to the nearest 0.1 mi. One aerial search of the San Miguel, Dolores, and Colorado Rivers downstream to Moab was conducted after coordination with Utah Div. of Wildlife Resources. Crayfish Studies Data from the crayfish mark-recapture study were analyzed by Dr. Gary White, Colorado State University, with a variety of estimation programs available through program CAPTURE. Estimates were computed for both all 4 consecutive trapping occasions and only the first 3 trapping occasions in an effort to examine what was gained by more trapping occasions. Details of trapping and marking were reported in the 1991 Job Progress Report. A relative abundance survey of crayfish along a 99-mi reach of the Dolores River was repeated in August-September 1991. The methods were identical to those used in 1988 and reported in the July 1989 Job Progress Report. Basically it required setting a baited wire crayfish trap at 1.0 mi intervals for a 24-hr period at each site and recording the catch by sex and length. The reason for repeating the survey was to evaluate the hypothesis that the distribution of crayfish observed in 1988 was a result of crayfish dispersal downstream and not a habitat related phenomenon. Mann-Whitney U-tests were used to compare catch rates in various reaches of the river. Release Site Monitoring A standardized river otter sighting form was distributed to District Wildlife Managers in areas of river otter release. All reported sightings are filed by river drainage. Periodic contact with Utah Division of Wildlife Resources personnel was continued to monitor movement of radio-transmittered river otters released by Utah DWR in the Green River along the state borders. Contacts were made with commercial boating companies and BLM river rangers to obtain river otter sightings in the Colorado River in the Utah-Colorado border area. RESULTS Dolores River AND DISCUSSION Release One adult male river otter died while in captivity in Oregon just 6 days prior to shipment. He had been held for over 4 weeks. Seven river otters were safely transported to Colorado. One wild-caught otter escaped the pens by climbing 2 m up a corner pole and breaking a hole into the metal roofing. This animal was tracked to a beaver den 2 km south. The den was monitored and �239 traps set for 8 days but no sign of the river otter was seen. However, a reliable observer did see a river otter 5 km SE of the site in April so it is likely this animal survived. Three male river otters died during surgery from respiratory arrest. Two died within 5 minutes of anesthesia while the third was alive when placed in the transit box but never recovered. Necropsy indicated massive hemorrhaging in the lungs and no other noticeable impairments. Three river otters (2 male, 1 female) were successfully telemetered and released in the Dolores River (Table 1). Both of the animals from the rehabilitation center (male & female) were successfully released. Table 1. River otters released in the Dolores River, CO, 1991. Yt(kg} Length (cm) Circumference (cm) Total Tail Head Neck Chest Tag # Origin ID Sex RO-38 M 12.1 117 47 25.0 28.5 49.0 38 Klamath River CA RO-39 F 8.9 114 45 24.5 30.5 42.5 39 Klamath River CA RO-40 M 8.1 109 45 27.0 28.0 41.0 40 Oregon The surgical procedure for radio transmitter implant appears to be improved in that all 3 transmitters stayed in the otters. The 3 released river otters were tracked throughout December and into mid-January 1992. All were alive at that time. The 2 males were within 6 mi of the release site while the female had moved downstream 27 mi. The female was observed catching and feeding on catfish (Ictalurus punctatus) for 4 consecutive days in January. Poor ice conditions restricted tracking during the winter and no signal was received from any of the 3 during 36 days of tracking during Mar-May 1992. An untagged subadult female river otter was reportedly found dead alongside the Dolores River at RM 113 on 18 January 1992. The carcass was delivered to eDOY on 29 Jan. 1992. Necropsy determined the animal had been bitten across the chest by a large canid sometime after death. Death was caused by puncture wounds to the ventral part of the neck reSUlting in massive hemorrhaging. There was also a large hematoma on the dorsal portion of the skull. There. were no obvious trap marks on the feet nor the skin. There was no evidence of feeding upon the carcass. The female weighed 6.3 kg, was III cm in total length, and did not have any fetuses in utero. Since she was untagged it is presumed she was born on the Dolores River. Twenty-seven river otters have been released into the Dolores River during 1988-91; 15 males and 12 females. Seven are believed to have died within 2 weeks of release (26%); 5 along the river and 2 which had climbed out of the canyon and gone cross-country. One adult female died of natural causes 9 months post-release. One male became stranded on a cliff and would have died had not research personnel radio-tracked him and caught and returned him to the river. At least 2 moved into other river systems, traveling 124 to 160 �240 miles. It seems reasonable to assume at least 10 animals were lost to the reintroduction effort before they had the opportunity to contribute offspring, a loss of 37%. While this may at first seem high, data on mortality postrelease for mustelid reintroductions is scarce. A fisher (Martes pennanti) reintroduction into Montana suffered a comparable loss of about 50% (Kim Heinemeyer, pers. comm.). Clearly juvenile river otter suffered the greatest loss whether it be in transit, during surgery, or immediate post-release. It seems prudent to assume a 40-50% loss in all future reintroductions of river otters. Therefore, the number initially released should be increased appropriately. Inadvertent catch in traps set for beavers (Castor canadensis) will continue to plague reintroduction and future management programs. Based on discussions with beaver trappers and measurements of river otters, the technique of using snares with a stop-lock deserves attention for trapping beaver. If the stoplock were set 45 cm (17.75 in) from the snare lock, 28 of 30 river otters measured would not have been captured. If the lock were set at 43 cm (17.0 in) then 26 of 30 would have been protected. A reasonable approach to protecting river otters in a reintroduction area would be to set up regulations which would protect nearly all of the river otters but not significantly hinder the bona fide beaver trapper. Historically we have set regulations which may have been overly restrictive but which were designed to protect all river otters from trapping. I suggest a combination of trapper information and education combined with less restrictive regulations, i.e. snares with stop-locks, warrants an evaluation. Crayfish Studies The mark-recapture study resulted in a total of 3,100 crayfish greater than 50 mm in total length being marked in 4 days of capture (1714 males, 1386 females). Six hundred fifteen individuals were recaptured a total of 741 times (500 twice, 104 three times, 11 four times). Only 2 crayfish were recaptured with unreadable marks; both of these were initially marked on day 1 and recaptured on day 4. The jackknife estimator [M(h)] from Program Capture and the M(h) model of Chao (1989) are the 2 most appropriate estimators (G. White, pers. comm.). The capture characteristics observed in this study (4 capture occasions and capture probability of about 0.1) are intermediate of the characteristics for which either of the 2 above estimators perform best. White's suggestion (based on simulations) was that the probable positive bias of the Chao estimator is much preferred to the larger negative bias of the jackknife estimator. The estimate for the 0.1 acre study reach was 9,275 crayfish greater than 50 mm total length with a standard error of 379. The 95% Confidence Interval was 8,575-10,064; or roughly within 8% of the estimate. Estimated density was 17.9 crayfish/m2 for a biomass estimate of 2,830 kg/ha (2,526 lb/ac). It is important to stress that total crayfish biomass may be much greater since the traps used rarely catch crayfish smaller than 50 mm in total length. The lower estimate of the M(h) model still represents a lot of crayfish. The. estimate was 6,539 with a standard error of 96 and a 95% Confidence Interval of 6,356-6,732. By shortening the number of trap occasions to 3 the estimate �241 was reduced by 9% and the Confidence Interval increased by 11%. More importantly, reducing the number of trap occasions gets one farther from the characteristics where the Chao M(h) estimator works best; i.e. over 5 trap occasions. The relative abundance survey of 1991 was truly a surprise, for the results were almost a mirror image of the 1988 survey (Fig 1). Three reaches of the Dolores River are readily identifiable from the crayfish survey. The upper reach, from Bradfield Bridge to Disappointment Creek (RM 173-129), clearly supports the greatest numbers of crayfish. The number of crayfish caught above Disappointment Creek, median catch of 40 in '88 and 61 in '91, was significantly greater (P<O.OOl) than for either of the lower reaches in both years. The most obvious habitat differences include low siltation and high proportion of cobble and boulder substrate above the confluence with Disappointment Creek. The middle reach, from Disappointment Creek to near Leach Creek (RM 129-98), still has significantly more crayfish than the lower reach. in both years (P<O.OOl) , with median captures of 17 in '88 and 10 in '91. The near disappearance of crayfish below mile 95-100 was consistent in both years; with a median catch of 2 in both years. While the sharp reduction in catch per unit effort seen at Disappointment .Creek can be readily explained by habitat changes, the reduction between Bull Canyon and Leach Creek cannot be. A comparison of substrate, water temperature, silt load, bank configuration, and riparian vegetation suggest no significant differences between the middle and lower reaches. The middle reach is characterized by many abandoned uranium mines and milling facilities. While the rapid reduction in crayfish numbers does not occur consistently at the mouth of a tributary canyon, it may be that cumulative loads of toxic materials may be changing water chemistry sufficiently to impact the crayfish. Whether such materials come from mine wastes or leach from the bedrock is conjecture. Certainly the sharp decline, rather than a scatter downstream, suggests a major change in habitat rather than a dispersal pattern. This work would suggest that water chemistry would be a more lucrative area of investigation rather than the physical structure of the habitat. The consistent decline in crayfish occurs dramatically between Bull Canyon (RM 105) and Leach Creek (RM 98). While differences in CPUE between years within a reach were statistically significant (P<O.OS), the changes are less striking than the consistent truncations. The intensity of sampling with CPUE data does not ring strong confidence in me despite the moderate statistical difference. I believe the shape of the curve, rather than the absolute levels, is the biological message in this data set. Release Site Monitoring Observations by outdoorsmen strongly indicate that river otters are present in Navajo Reservoir and all the major tributaries of this large reservoir. While the majority of sightings occur in the reservoir and the Piedra River, reputable sightings also come from the San Juan, Navajo, and Pine Rivers. River otters are consistently sighted throughout the Piedra River drainage and the small tributaries. At least 3 river otters released in the Green River in Utah have ventured into Colorado, traveling in both the Green and Yampa Rivers. Since only a small proportion of the animals released in Utah were radio transmitter equipped, �N +:-- N .Relative abundance of crayfish Dolores River No. of Crayflsh/24 10 0 t- 80 r _..-.._ _.._.._.._ __ ._._._ .. hrs ._.._ _ _.-.-.- .._.._.._.-.._.._.._- .._ ~ _.- -.._-_ _.-_.._._ ____ _.__.._.._._ _ _..,.._ , _.__.. "'P ••••••••••••• _ _- .,.._ ••••••• _._ .._.._.._ _ •••••••••••••••••. l ..• , ••••• _ _._. •••••••••••••••••••••••••. • 60 40 20 o l-watt~ 76 I 86 96 106 I IV. 116 126 I I 136 146 River Mile -1991 -t-1988 I 166 I I 166 176 I �243 there could be others. It appears likely that continued releases in Utah will result in some river otters taking permanent residence in Colorado's portion of the Green River. Sightings of river otter are consistently made on the Colorado River between Westwater Ranger Station and Dewey Bridge in Utah. The regularity of sightings is probably a response to the high recreational use of this river segment. While at least I river otter from the Dolores release has joined this population, the appearance of this population predates the initial Dolores release. It is speculated that this population is the result of dispersal from either or both the North Fork Colorado River or the Gunnison River releases. Sightings of river otter in the Granby-Grand Lake-N. Fork Colorado River remain common. Increasing numbers of complaints are received each year from boathouse owners in the lakes who do not like sharing their boathouses with river otters. Also, 2 river otters were inadvertently killed by beaver trappers near Grand Lake in 1991. LITERATURE CITED Chao, A. 1989. Estimating population size for sparse experiments. Biometrics 45:427-438. Thomas D. I. Beck Wildlife Researcher data in capture-recapture ��L4~ Colorado Division of Wildlife Wildlife Research Report July 1992 JOB PROGRESS REPORT State of ~C~o~l~o~r~a~d~o~_ Project No. Mammals Research W-153-R-4 Work Plan No. Elk Investigations 9A Job No. Impact of Elk Winter Grazing on Livestock Production and Work Plan No. 3~ Job No. 5 Period Covered: Author: _ July 1, 1991 - June 30, 1992 N. T. Hobbs, D. L. Baker, and G. Bear Personnel: D. Bowden Abstract Populations of elk using sagebrush steppe rangeland during winter and early spring are believed to compete with cattle for limited supplies of spring forage. We examined effects of variation in elk population densi ty (0, 8, 15, 31 animalsfkm2) on elk competition with cattle in a randomized complete block experiment with 3 replications. We repeated the experiment over four years. We observed effects of elk populations on the supply of forage available to cattle, and on cattle growth and reproduction. We observed linear declines in forage supplies available to cattle as elk population density increased. Utilization of standing dead perennial grass by elk increased at a rate of about 2% per unit increase in elk density (Fl,6 80.1, P < 0.0001); utilization of live grass increased at about 0.5% per unit increase in elk density (Fl,6 - 6.4, P - 0.04). Consequently, standing crops of dead perennial grass declined in direct proportion to elk density (Fl,6 9.9, P - 0.02) from an average of about 10 g/m2 in the controls to less than 5 g/m2 in the high density (31 elk/km2) treatment. We also observed linear trends in effects of elk on the early spring standing crop of live perennial grass, but these trends only approached significance (Fl,6 - 3.2, P - 0.12). We failed to detect effects of elk grazing on primary production of perennial grasses and other live herbs. �246 Rates of removal of live perennial grasses by cattle declined in direct proportion to elk density (F16 - 6.3, P - 0.04); removal of dead perennial grass showed similar downward trends (F1•6 - 13.4, P - 0.01). Averaged across treatment levels, elk grazing depressed removal of live perennial grass by cattle by as much as 2.6 kg/cow/day, and reduced removal of dead perennial grass by as much as 2.1 kg/cow/day. Rates of gain of cows and calves during the spring grazing season were directly related to rates of forage removal by cows. Over the 4 study years, forage removal rates (an index to dry matter intake) accounted for more than a third of the variation in rates of gain of cows during the spring (r2 - 0.36, F1•47 - 25.9, P - .0001) but only accounted fo 8% of the variation in calf rates of gain (r2 - 0.08, F1•47 - 4.0, P 0.05). Calf spring weights declined in direct proportion to elk density (F1•6 - 6.0, P - 0.05). However, by fall, we failed to detect linear trends in calf weights (F1•6 - 0.0, P - 0.95) while quadratic effects approached significance (F1•6 - P - 0.14). We found the largest effect of elk grazing on fall calf weights at the 8 elk(km2 level, where calves were 4.8 - 23.2 kg lighter than those in the control. We failed to observed repeatable effects of treatment on rates of gain by cows. However, we saw a negative linear trend in the cow rates of gain relative to treatment during the spring grazing season and a positive trend during summer and early fall. Consequently, the was a highly significant negative relationship between rate gain during the spring grazing season, and rate of gain thereafter (r2 - 0.49, F147 - 44.5, P < 0.0001). This suggests that growth rates of cows during s~er compensated for treatment effects during the spring. Natality rates of cows ranged from a high of 96% in the controls to a low of 85% in the low elk density (8fkm2) treatment. We were unable to detect repeatable effects of elk grazing on cattle natality rates, although effects approached significance for the 8 e1kjkm2 level (F1•6 - 2.8, P - 0.14). Total secQndary production by cattle was reduced by about 10% as a result of competition'with elk, but the effects of competition on cattle production were not proportionate to elk density (F1 6 0.04 , P - 0.84). • We conclude that competition between cattle and elk on sagebrush steppe rangelands represents the composite outcome of several weak forces. Taken collectively, these composite forces drive strong effects on cattle growth and reproduction. The ability of cattle to compensate for competitive effects by compensatory growth produces non-linear relationship in the effects of elk population density on the intensity of interspecific competition with cattle. �247 IMPACTS OF WINTER GRAZING BY ELK ON CATTLE PRODUCTION P. N. OBJECTIVES 1. To test the hypothesis that elk grazing during winter influences the productivity and botanical composition of herbage on sagebrush grassland ranges during spring. 2. To test the hypothesis that elk grazing during winter influences the body weights and rates of gain of cows and calves using sagebrush grassland ranges during spring. METHODS AND MATERIALS Study Area We conducted experiments at the Colorado Division of Wildlife Little Snake Wildlife Management Area in northwestern Colorado (lat. - 40, long. - 108). Topography and climate are typical of the high, cold deserts of the Intermountain Sagebrush Steppe. The area includes slopes, gullies, and flats ranging in elevation from 1800 to 2000 m. Aspects are predominantly southern and southwesterly with an average slope of 15 degrees. Soils are mostly sand and sandy loam with a rooting depth of about 100-120 cm and moderate to high permeability. Brown's Park sandstone forms the underlying substrate. Climate of the area is dry and cold; annual mean temperature is 6.06 C and mean annual precipitation is 27.5 cm, about two thirds of which usually falls as snow. The growing season averages only 81 days. Winter snow depths are highly variable, but accumulations of 10-30 cm are common on level ground during December through March. Wind scours snow from ridge tops, depositing it on lee slopes. Growing season precipitation during our study tended to be below average, while snow accumulation exceeded the average (Table 1). Vegetation is representative of the Wyoming Basin Province. Big sagebrush (Artemisia tridentata ) dominates the overstory; other important shrubs include rabbit brush (Chrysothamnus vicidiflorus, Chrysothamnus nauseous), horse brush (Tetradymia spinosa), and snake weed (Gutierzia glomerella). Predominant grasses are needle and thread (Stipa comata), western wheatgrass (Agropyron smithii), Indian Junegrass (Koleria cristata), bluegrass (Poa spp.), Indian ricegrass (Oryzopsis hymenoides) and cheatgrass (Bromus tectorum). Important forbs include wallflower (Erysimum asperum), peppergrass (Lepidium perfoliatum), silver lupine (Lupinus argenteus), and scarlet globe mallow (Sphaeralcea coccinea). Our study area was protected from domestic grazing during the five years previous to the start of our work. Condition of the range at the beginning of our studies was good-excellent. Experimental Design We examined responses of forage and cattle to 4 levels of elk population density (0, 8, 15, 31 animalsfkm2) in a randomized complete block experiment �248 with 3 replications. We repeated the experiment annually during four consecutive years (1988 through 1991). Experimental units consisted of 12 fenced pastures constructed specifically to implement our research design. Each pasture was 32 ha in area and triangular in shape. Although competitive interactions between cattle and elk in nature are not bounded in fenced pastures as they were in our studies, we were willing to sacrifice the realism offered by free-ranging animals to justify the experimental control provided by constraining their movements. Treatments were allocated to experimental units as follows. Four pastures were assigned to each of 3 blocks on the basis of pretreatment biomass of perennial grasses estimated by harvesting and weighing thirty 0.25 m2 plots in each pasture on June 1 of the year previous to initiating the experiment. Thus, there was one block with low biomass of perennial grass, one with medium biomass, and one with high. Treatment levels (0, 8, 15, 31 elkfkm2) were randomly assigned within each of the three blocks. Stocking Rates of Elk and Cattle We varied the intensity of elk grazing by stocking pastures with different densities of elk during winter and early spring. Controls consisted of pastures that received no elk grazing. Treatment levels were chosen to reflect local population densities of elk. Density clearly depends on the spatial scale over which it is measured. On a broad scale, elk populations in northwest Colorado typically average about 5-10 animalsfkm2 of winter range. We chose one treatment level (8 elkfkm2) to represent this large scale average. However, elk densities can be much higher at finer scales. Thus, we choose higher treatment levels (IS, 31 elkfkm2) to mimic local concentrations of animals and to examine forage and cattle responses to extraordinary levels of elk grazing. Treatment pastures were stocked with 3 elk to achieve a density equivalent to 8 animalsfkm2 (5.1 ha/animal unit month), 5 elk to achieve 15 animalsfkm2 (3.04 ha/animal-unit month) and 10 elk to achieve the 31 animalfkm2 level (1.5 ha/animal unit month). All elk introduced into pastures were females> 2 years old. Each year, animals were trapped from the surrounding area in portable corral-traps baited with alfalfa hay. Average date of release of elk into the experimental pastures was December 27. Elk were held in pastures until approximately April 15 when they were expelled to the surrounding rangeland. After removing elk, we stocked experimental pastures with hereford cattle during May and June of each year. We chose season of use by cattle to reflect prevalent patterns in Colorado. The same number of cattle (7 cow-calf pairs and 1 heifer) were stocked in all pastures, including the control, during all four study years. Thus, the 0 elkfkm2 pastures controlled for the effects elk grazing but were not intended to control for influences of cattle grazing. We randomly allocated 84 cows and calves and 12 heifers among the replicated treatments. Randomization was accomplished by first stratifying cows herd by age, and then randomly choosing animals using to accomplish similar age distributions among pastures. Stocking rates of cattle were chosen to achieve about 50% utilization of the net annual aboveground production of perennial grasses (about 25% of the net aboveground primary production of all vegetation) in the control pastures. �249 This stocking rate is typical for sagebrush steppe on public and private land in the region, but probably exceeds stocking rates needed to achieve optimum economic returns. Duration of cattle grazing was adjusted each year to achieve annually consistent levels of forage utilization by cattle in the control pastures despite annual fluctuations in forage production. Thus, pastures were stocked with cattle for 6 weeks during years 1 and 2 (May 8-June 18 1987 and May 12-June 21 1988, 2.8 ha/animal-unit month), for 3 weeks during year 3 (May 10-May 31 1989, 6.9 ha/animal unit month), and for 5 weeks during year 4 (May 9-June 13 1990, 3.3 ha/animal unit month). Hereafter, we will refer to the time interval that cattle were stocked in the pastures as the "spring grazing season." Cattle were obtained via contract with a local rancher (Mr. Bruce Seely of Craig, Colorado). At the end of the spring grazing season, cattle were returned to the Mr Seely's Ranch where they were maintained in the same group with the same general management regime for the remainder of the year. They were maintained on native rangeland in excellent condition during the summer and fall, and were fed hay through out the winter. All cows were exposed to fertile bulls immediately after their removal from pastures. Each year's calf crop any cows that failed to breed were sold after weaning, usually in midNovember. After cattle were removed from pastures, pastures remained ungrazed by cattle or elk until they were restocked with elk during the following December. Measurements of Forage Production and Utilization We estimated primary production and rates of forage removal and utilization of herbaceous forage using movable exclosures. We defined production as the amount of live plant tissue produced annual before the end of the spring grazing season. Thus, our estimate of production approximates aboveground net primary production, presuming that production after early July was minimal. (Given that virtually all the grasses on our study site were cool season species, this is a reasonable presumption.) We defined removal as the amount of forage consumed and trampled by cattle and elk. We used removal as a general indication of intake rate (kg/animal/day). Utilization of standing dead forage was defined as the percentage of the annual, ungrazed standing crop (i.e. the residual of the previous year's grazed production) that was removed by herbivores. Utilization of live forage was the percentage of the annual production of live herbaceous tissue that was removed. To estimate production, removal, and utilization, we harvested above ground standing crop contained in 40 paired, 0.75 m2 grazed and ungrazed plots in each of the 12 pastures on May 1, June 1 and July 1 of each study year. Grazing was excluded from one member of each paired plot by enclosing it with a movable, cone-shape cage (diameter - 1.2 m. height - 1.4 m) formed of heavy gauge concrete reinforcing wire (mesh size - 10 cm). Cages were staked to the ground with 64 cm lengths of steel rebar. Although cages such as the ones we used are known to influence primary production, we assumed that such influences were constant across levels. Thus, by using the same methods to estimate production in the control as well as the treatment pastures, we obtained reliable estimates of the relative effects on·production, even if our absolute estimates of production were biased by spurious effects of cages. �250 Plot pairs were initially established during September before the first study year. Locations of plots were chosen by girding a map of each pasture into 10xlO m cells, numbering all grid intersections, and randomly choosing (without replacement) among numbered locations. Plots were located in the field by using compass bearings from known locations and pacing required distances. After locating one plot randomly, a second plot was subjectively chosen within a 30 m radius to mimic the standing crop of the initial plot. Both plots were marked with a 25 cm surveyor's stake and one was randomly chosen by coin flip to receive the cage. This process was repeated for each of the 40 paris in each pasture. On each sample date plots were clipped and moved. New plot pairs were subjectively chosen to mimic the grazed condition in the previous open (uncaged) plot. After the last sample date, plot locations for the next study year were chosen randomly within a 30 m radius of the last established plot. Plots were harvested as follows. All herbaceous vegetation contained within a .7 m diameter steel hoop centered on the plot stake was clipped to stubble height (about 1 cm) and sorted into 3 categories (perennial grass, annual grass, forbs) in the field. After returning clipped samples to our facilities in Fort Collins, they were dried at 55 C for 48 hours and were subsequently separated by hand (many hands) into live and dead components. We then weighed each of the 6 categories (the 3 above x live and dead) to the nearest 0.1 g. Composites of material in each category harvested from each pasture was subsampled to form a 20 g sample for nutritional analysis. We ground these samples to pass a .5 mm mesh screen, and analyzed them for total dry matter, ash, nitrogen content, in vitro organic matter digestibility. Dry matter and ash were determined following A.O.A.C. Total nitrogen was determined using kehldahl procedures. Organic matter digestibility was estimated using a two stage, in vitro procedure. We obtain rumen inocula for these procedures from a fistulated holstein cow fed native grass hay. Primary production was estimated as the sum of increments in live standing crop biomass within caged plots observed on the 3 sample dates. Forage removal was estimated as the sum of the difference between grazed and ungrazed plots on each sample date. Utilization was calculated as the ratio of removal to production for live forage categories, and as the ratio of removal to the ungrazed standing crop for standing dead. (For details on these calculations, see Appendix A). As a result of exceedingly high variability in forbs and annuals, particularly for the standing dead component, we focussed on three vegetation categories: live perennial grass, dead perennial grass and other live herbs. The "other live herbs" category included pooled data for annual and perennial forbs and annual grasses. Pretreatment measurements on shrub biomass indicated that sampling intensity required to obtain estimates the biomass of palatable shrubs was prohibitive. Although these plants were relatively rare, we believed they were potential important to cattle. Consequently, we chose to estimate availability of shrubs to cattle based on shrub canopy coverage. We estimated percent canopy cover of 2 groups, Artemisia tridentata (which is not palatable to cattle) and other shrubs which included a variety of palatable species. Our estimates �251 were based on intercepts of shrub canopies along 25, 12-m line transects in each pasture during approximately July 5-10 of each study year. Locations of transects were established during year 1 and were maintained through out the study. Transect locations were chosen by girding a map of each pasture into 10x10 m cells, numbering all grid intersections, and randomly choosing without replacement among numbered locations. Transect endpoints were marked with steel pins (1 x 30 cm). In addition, a 1 m wooden surveyor's stake (painted high visibility orange) was placed 10 m south of a transect endpoint to facilitate locating pins. Readings were taken as follows. After locating pins, we stretched a surveyor's tape between transect endpoints. We then read intersections of shrub canopies with the tape to the nearest 1 cm. Measurements of Cattle Production It is important to emphasize that although all pastures were stocked with cattle and elk, stocking rates of cattle were constant across pastures, while elk stocking rates varied. Thus, performance of cattle offered a controlled response to different levels of elk grazing. Cattle performance was measured in several ways. During each study year, we weighed calves born to cows in the experiment with 2 days of their birth. (Throughout this paper, we will use the term "calf" to refer to young of the year born to cattle. At no point will we discuss elk calves.) Calves and cows were also weighed when they were introduced to pastures, when they were removed, and at weaning. We also recorded birth dates of all calves and reproductive status (pregnant vs open) of all cows during each year. Natality rates were calculated for each pasture each year as the number of cows producing calves divided by the number of total number of cows in the pasture (- 8). We estimated total secondary production by cattle in each treatment over the four study years as the product of the average natality rate times the average weaning weight of calves plus the average total weight gain of adults (Appendix A). To the extent possible, we maintained the same individuals in each pasture from year to year and followed their weight gains and reproductive output throughout the study. This allowed some expression of carry-over effects from year to year. On average, cattle were kept in the study for 2.7 years each; 34 of the original 96 cows remained in the study for all 4 years. Twelve older animals were replaced each year to accommodate addition of one heifer in each pasture. Sixteen additional animals were replaced because of failure to breed, death losses, and other miscellaneous sources of attrition. Statistical Analysis We analyzed weight responses and conception rates of cattle with a split plot, factorial analysis of variance. Factors were elk density, block, and year all of which were assumed to be fixed. Treatment formed the whole plot and repeated measures (i.e. years) formed the split plot. We used responses of individual animals as observations and calculated all F ratios using the mean square error of the density x block interaction in denominators. This allowed us to preserve individual responses in the analysis (which was particularly �254 year for both shrub classes (P < 0.03) but the magnitude of year effects did not interact with treatment (year x level P > 0.51). We were unable to detect effects of elk grazing on primary production of perennial grasses and other live herbs during the spring grazing season. However, we cannot exclude the possibility that elk grazing (averaged across treatment levels and years) reduced primary production by as much as 4.9 g/m2, nor can we rule out the possibility of enhancing effect on production that could be as large as 1.9 g/m2 at the intermediate grazing levels (15 elkjkm2) levels. Effects of elk grazing on other live herbs were substantially more variable than effects on perennial grass, but we can be reasonably certain that elk grazing, averaged across treatment levels, did not change primary production of live herbs by more than 3.3 g/m2. Summed differences between open and caged plots during the spring grazing season provided an estimate of forage removal by cattle (kg/cow/day, Appendix A). Forage removal is strongly influenced by cattle forage intake (removal also includes trampling losses) and thus provides an index of the effect of treatment on total daily forage ingestion (kg/cow/day) by cattle. Rates of removal of live perennial grasses by cattle declined in direct proportion to elk density (linear contrast F1.6 - 6.3, P - 0.04); removal of dead perennial grass showed similar downward trends in relation to elk density (Fig.6B, linear contrast F16 - 13.4, P - 0.01). Averaged across study years, removal of standing dead grass by cattle in the controls was 4 times greater than in the high density (31 elkjkm2) treatment, while removal of live was, on average, about 25% greater in the control compared with the high elk density treatment. Averaged across treatment levels, elk grazing depressed removal of live perennial grass by cattle by as much as 2.6 kg/cow/day, and reduced removal of dead perennial grass by as much as 2.1 kg/cow/day. We failed to detect effects of elk grazing on removal of other live herbs. Total forage removal (the sum of removal of dead perennial grass, live perennial grass, and other live herbs) declined with treatment, but linear effects were weak (linear contrast F1.6 - 3.76, P - 0.10) Forage removal rates by cattle declined significantly with year for all forage classes (year effect P < 0.03). However effects of year on removal of standing dead were particularly large-- removal of standing dead grass by cattle during year 1 was four times greater than removal during the fourth study year. The effects of year on forage removal by cattle interacted with the effects of treatment; the inhibiting effect of elk density on grass removal by cattle tended to intensify as the study proceeded (year x treatment interaction P < 0.07). Cattle utilization of standing dead perennial grass (removal/standing crop, Appendix A) declined in direct proportion to elk density (linear contrast F1.6 - 24.1, P - 0.003) and, thus, opposed the trends in utilization of standing dead grass by elk. Similarly, utilization of live perennial grass by cattle was virtually a mirror image of utilization by elk. As a result of these opposing trends, the proportion of the annual production of perennial grass collectively removed by cattle and elk remained relatively constant across all levels of elk density. �255 Effects of Elk Grazing on Cattle Weight Performance and Reproduction Calves born to cows exposed to competition with elk during the previous spring tended to be born slightly later if their mothers were in moderately grazed treatments (8 and 15 elkfkm2) than if they were in the control or the high elk density treatment. Quadratic effects on birth date approached significance (F1•6 - 3.2 , P - 0.12). On average, calves in the 15 elkfkm2 group were born 5.5 days later than calves in the control. We can be reasonably sure that the true effect of moderate elk grazing delayed calf births dates in the range of o to 10 days. Effects of elk on performance of calves varied with season of the year. We failed to detect effects of elk population density on calf birth weights. However calves born to cows who were in control pastures the previous spring tended to be slightly heavier (on average, about 1 kg) than calves born to cows who were in the elk-grazed pastures. These tendencies were more apparent at the beginning of the spring grazing season when weights of calves whose mothers were in the 15 elkfkm2 treatment during the previous year were lower than weights of calves whose mothers were in the control (F1•6 - 6.2 P0.05). Quadratic effects on calf weights at the beginning of the spring grazing season were significant (F1•6, , P - 0.06), but linear effects were not (F1•6 - 1.2, P - 0.31). At the end of the spring grazing season, calf weights declined in direct proportion to elk density (linear contrast F1•6 6.0, P - 0.05). By fall, we once again failed to detect linear trends attributable to treatment (linear contrast F1•6 - 0.0, P - 0.95) while quadratic effects once again approached significance (F1•6 - P - 0.14). We found the largest effect of elk grazing on fall calf weights at the 8 elkjkm2 level, where calves were 4.8 - 23.2 kg lighter than those in the control. The difference in fall weights of calves in the control relative to the average of the 3 elk-grazed treatments was weakly significant (control vs others contrast, F1•6 - 3.9, P - 0.097). We can be reasonably confident that, on average, elk grazing reduces calf weights in the fall by no more than 15 kg, which is about 8% of the fall weight they would be expected to attain in the absence of competition with elk. Calves showed positive rates of gain thought the spring and summer. Rates of gain of calves during the spring grazing season decreased in direct proportion to treatment (linear effect contrast F1•6 - 8.0; P - 0.03) and the control differed significantly from the average of other levels (F1•6 - 0.08). We were unable to detect significant effects of treatment on calf rates of gain after they were removed from experimental pastures, but rates of gain of animals in the moderate and high density treatments were slightly higher than those in the control. Year effects were significant for all calf responses (year effect P < 0.10); calf weights and rates of gain tended to be the lowest during year 3 when spring precipitation (Table 1), production of perennial grass, and live grass intake by cows were also at a low point. However, the magnitude of treatment effects did not depend on year for any calf response (year x treatment interaction P > 0.45). �256 At the beginning of the spring season, average weights of cows who were in control pastures during the previous spring exceed weights of cows who were in the elk-grazed pastures (control vs others contrast Fl•6, , P -). However, we can be sure the average effect of elk on early spring weights of cows was less than 21 kg (or about 5% of the weight of animals who did not compete with elk). We were unable to detect any significant effects of elk grazing on cow weights at the end of the spring grazing season, despite some linear trends in the data. It would appear logical to compare average weights at the beginning of spring with average weights at the end of spring to draw conclusions about weights gains. However we caution that such comparisons of means are misleading because the average of early spring weights is based on 3 years data, while the average of late spring weights is based on 4 years data. This was the case because cows during the early spring of the first study year had not ever been exposed to treatment and were used as covariates rather than as responses in the analysis, while cows at the end of spring had been exposed to the elk density treatment during all 4 study years. We failed to observed repeatable effects of treatment on rates of gain by cows. However the there was a negative linear trend in the cow rates of gain relative to treatment during the spring grazing season and a positive trend during summer and early fall. Consequently, the was a highly significant negative relationship betWeen rate gain during the spring grazing season, and rate of gain thereafter (r2 - 0.49, Fl•47 - 44.5, P < 0.0001). This suggests that growth rates of cows during summer compensated for treatment effects during the spring. We did not observed compensatory effects in growth rates of calves. Calf rates of gain duig the summer and early fall were positively related to rates of gain during spriong. However, this relationship failed to account for much variation in calf growth rates (r2 - 0.08, Fl•47 - 2.9, P 0.09). Rates of gain of cows and calves during the spring grazing season were directly related to rates of forage removal by cows. Over the 4 study years, forage removal rates accounted for more than a third of the variation in rates of gain of cows during the spring (r2 - 0.36, Fl•47 - 25.9, P - .0001). Although calf rates of gain were statistically related to forage removal rates by cows, this relationship failed explain much variation in calf rates of gain (r2 - 0.08, Fl•47 - 4.0, P - 0.05). We measured natality rates of cattle as the number calves born per 100 cows. Natality rates ranged from a high of 96% in the controls to a low of 85% in the low elk density (8fkm2) treatment. We were unable to detect repeatable effects of elk grazing on cattle natality rates, although effects approached significance for the 8 elkfkm2 level (Fl•6 - 2.8, P - 0.14). However, we were also unable to rule out treatment effects that could be quite large (about 20% points). Year effects were not significant (year effect F2•S - 0.08, P0.95) and the effect of elk density on natality rate did not depend on year (year x treatment interaction Fl•6 - 1.08; P - 0.42). Total secondary production by cattle was reduced as a result of competition with elk, but the effects of competition on cattle secondary production were not proportionate to elk density (linear contrast Fl•6 - 0.04 , P - 0.84). We observed the greatest effects of competition at the 8 elkfkm2 level, and the smallest effects at the 31 elkfkm2 level (quadratic contrast Fl•6 - 5.6, P 0.06). Four year means for cattle production at the moderate and heavy elk �257 grazing levels were not significantly different from the control, but the control did differ significantly from the mean of the three elk density levels (control vs others contrast F1•6 - 4.4, P - 0.08). Year effects on total cattle production approached significance (year effect F1•6 - 3.4, P - 0.13) but the magnitude of treatment effects did not depend on year (year x treatment interaction F1•6 - 0.60, P - 0.77). ! ..~ Prepared by __·~ ~ l_~ N. Thompson Hobbs Wildlife Researcher C Dan L. Baker Wildlife Researcher C _ ��259 Colorado Division of Wildlife Wildlife Research Report July 1992 JOB PROGRESS REPORT State of ~C~o~l~o~r~a~d~o~_ Project No. Work Plan No. Job No. Mammals Research lOA Kit Fox Studies Kit fox (Vulpes velox macrotis) Status in Colorado 1 Period Covered: Author: W-153-R-5 July 1, 1991 - June 30, 1992 T. D. I. Beck and M. A. Link Personnel: R. B. Gill, J. P. Fitzgerald Abstract. A detailed study plan was approved after internal review. Three kit fox (Vu1pes ve10x macrotis) were captured during 1,584 trap nights between 9 March and 30 June 1992. The 2 males and 1 female were captured NE of Montrose. Ten non-target animals were also captured. Only 1 fox was observed in 51 hrs of spotlight surveys. Extensive trapping and searching in apparently suitable habitat with little return suggests either technique problems or especially low densities of kit fox. ��261 KIT FOX (YULPES !ELQX KACROTIS) STATUS IN COLORADO MICHELLE LINK AND THOMAS BECK P. N. OBJECTIVE Document the geographic distribution and relative abundance of kit fox in western Colorado. SEGMENT OBJECTIVES 1. Develop a detailed study plan with the contractor and graduate student, J. P. Fitzgerald and M. Link, respectively. 2. Identify the geographic extent of kit fox distribution in western Colorado. 3. Evaluate kit fox survey techniques. METHODS AND MATERIALS The detailed study plan was prepared by the contractors and submitted to the Division of Wildlife Research Group peer review committee and to the contract statistical consultant. Historic records of mammalogists and trapping records were reviewed in order to designate priority search areas. Initial work in an area focused on interviews with landowners, trappers, and Division of Wildlife personnel. Field surveys began on 9 March 1992. Potential survey areas were driven through and foot reconnaissance was conducted in search of fox sign. Depending on availability of observers, spotlight surveys were conducted at night in search of foxes and potential prey items. Baited cage traps were set in what appeared to be suitable habitat for 3-5 consecutive nights. Dead 2day-old turkeys were used most often as bait. Records of trap nights, target and non-target catches, weather conditions, and habitat type were recorded. Non-target captures were photographed and released. Captured kit fox were weighed, sexed, ear-tagged, and photographed. More detailed descriptions of the study methods are available in the detailed study plan. RESULTS AND DISCUSSION A detailed study plan was reviewed by the peer group and approved by the statistician. Copies are available from the contractor, James P. Fitzgerald, Univ. Northern Colorado, or the Mammals Research Section Leader, Colorado Division of Wildlife. Only a portion of the possible kit fox habitat in western Colorado was surveyed during this fiscal year. Primary areas searched include the Dolores �262 River valley near Gateway, the Grand Valley and Rabbit Valley NNW of Grand Junction, Sinbad Valley, the Gunnison River desert areas between Delta and Whitewater, and areas east of Delta and Montrose. Slightly over 51 hrs of spotlighting was conducted over 24 occasions. While numerous wildlife species were observed, only 1 fox was sighted and positive species of fox was not determined. Gray fox (Urocyon cinereoargenteus) were known to inhabit the area of the sighting. Thirteen animals were trapped in 1,584 trap nights. These included 3 kit fox, 4 striped skunks (Mephitis mephitis), 3 white-tailed prairie dogs (Cynomys leucurus), 1 raccoon (Procyon ~), 1 coyote pup (Canis latrans), and 1 domestic cat (Felis sp.). The kit fox had been located in den areas based on conversations with landowners prior to trapping. All kit fox were trapped at 2 den sites in the Peach Valley area NE of Montrose. More foxes were at the dens but it was not an objective to capture all foxes so traps were pulled after 2 nights per site. The 2 dens were approximately 5 km apart. Possible causes of the low catch in live traps are low density of animals, poor bait, and trap wariness. Different baits and trap techniques will be used during further field work. The fact that no kit fox was trapped without prior knowledge of the den site is a point of concern. The poor return on spotlight surveys suggest further efforts are probably not warranted. Interviews with local landowners, trappers, road workers, and DOW personnel have proved productive in several ways. Two dens were found as a result of interviews. Few people have ever seen a kit fox, only slightly more have seen gray foxes, but many have seen red foxes (Vulpes). It is believed this pattern is more a result of sightability during daylight than population numbers. While some local landowners were very much aware of the kit fox dens, some of their neighbors were unaware that such animals ranged in the region. Trapping records are being collected and evaluated. By ,j-Al5nVLJ) 1fJ. --0 I3JZt2/:_. i Yr'Jt_~~_)~;?;l U1 c:)__j_, Thomas D. I. Beck and Michelle Link Wildlife Researcher Graduate Research Assistant � https://cpw.cvlcollections.org/files/original/e2a6daf297c0716c8a9ccd4387cff713.pdf 4e2f261bedf791233a9fa2490734d0fb PDF Text Text 5 Colorado Division of Wildlife Wildlife Research Report October. 1992 JOB FINAL REPORT State of __~C~o~l~o~r~a~d~o~ Project Work Plan Avian Research - Migratory Game Birds W-166-R-1, _1_ : Job _ 19 Job Title: Evaluation of nesting habitat management for ducks Period Covered: 1 April 1991 through 31 March 1992 Authors: David W. Gilbert. James K. Ringelman. Michael R. Szymczak. and David R. Anderson Personnel: D. Anderson, Colorado Cooperative Fish and Wildlife Research Unit; J. Ringelman, M. Szymczak, Colorado Division of Wildlife; D. Gilbert, Colorado State University; S. Brock, S. Berlinger, A. Morkill, R. Schnaderbeck, U.S . .Fish and Wildlife Service. ABSTRACT The numbers of mallard (Anas platyrhnchos) and northern pintail (Anas acuta) were equally distributed between the May and June survey periods. Clutch sizes of mallards did not vary among management units and varied little by year (~<0.001). June mallard clutch sizes (7.64) were smaller than May clutches (7.94; ~<0.001), but the difference was not substantial. Nest transect data from the Monte Vista National Wildlife Refuge were further evaluated to determine the effects of grazing, burning, haying, and predator control on duck nest density, nest success, and species composition. Dormant season grazing (15 September to early winter) in a 3-year °restrotation cycle was conducted on 19 units during 1976-90. Nest density declined (~ - 0.037) with increasing grazing intensity, with large (38%) declines apparent the first season after only light grazing (0.5 AUM/acre). Three years after grazing, nest density was still depressed at least 17% below pre-grazing levels. In addition, nest success declined as grazing intensity increased (~- 0.070). Records of prescribed bUrns with >50%. ground coverage revealed that nest density was depressed (l < 0.05) the ne.!;;ting season after burning, but recovered to pre-burn densities the second year. Haying also had a detrimental effect in duck nesting densities (~- 0.035), although inferences were weakened because of confounding with other management treatments. Predator management was conducted at 3 intensities: an early (1964-70) period of intensive predator control using poisons and traps, a middle (1971-80) period of less intensive predator control relying primarily on aerial gunning, and a late (1981-90) period of renewed intensity during which mammalian and avian predators were shot and trapped by refuge personnel. Duck nest success varied among predator management periods (l- 0.004), with greater success evident during the first period than during the second; success during the third period did not differ from the other two. Several meetings have been held with personnel in the Division of Refuges, U.S. Fish and Wildlife Service Region 6, to discuss the implications of study �6 findings to management practices on the Monte Vista National Yildlife Refuge. A manuscript for Yildlife Monographs has been prepared and is presently under review. Any additional analyses and publication preparation will be conducted under York Plan 22, Job 2, Migratory Bird Publications. �7 EVALUATION OF NESTING HABITAT MANAGEMENT FOR DUCKS David W. Gilbert James K. Ringelman Michael R. Szymczak David R. Anderson P. N. OBJECTIVES 1. Relate nesting duck species composition, nest success rate, and nest density to habitat management practices. 2. Assess changes in wetland and upland vegetation between 1962 and 1985 by contrasting digitized habitat information derived from aerial photographs. 3. Determine nesting habitat preferences of duck species by comparing usage of nesting habitat with relative habitat availability. SEGMENT OBJECTIVES 1. Complete categorical analyses of nest transect data, employing log-linear models, logistic regression and contingency table analyses to test hypotheses concerning the effects of grazing, burning, and water application on duck species composition, nest success and nest density. 2. Convey findings to agency personnel involved with management of state and federal wildlife refuges in the San Luis Valley and North Park. 3. Complete annual reports, and prepare a monograph on study results. METHODS Note: A complete description of study area and methods and results of nest success was presented in last year's (October 1991) Wildlife Research Report. Methodology, results and discussion reported here reflect only updated analyses completed during this segment. Clutch Size Average completed clutch size was calculated using only the number of eggs in nests recorded as "under incubation". May clutches for all species were classified as early season clutches because in the SLV little nesting occurs before the first week in May (Pospahala 1969, Schroeder et al. 1976). June nests were considered late nests for mallards and pintails, intermediate in the nesting cycle for teal, and early nests for gadwall (Anas strepera). Teal begin nesting the last week in May and gadwall the first week in June (Schroeder et al. 1976). Samples to measure management unit and annual variation in clutch sizes within species were too small for all species except mallards. Variation in mallard clutch sizes between units and between years was examined by regressing the number of mallard eggs (E) on the number of mallard nests (N) �8 for each unit (i) or year (j) as: We compared mean clutch sizes for all species between May and June survey period and between MVNWR and other geographic areas (2-tailed t-test, Simpson et al. 1960) for mallards, northern pintail, and gadwall. Clutch size standard deviations were obtained from published literature or calculated from published information if sufficient data were available. Blue-winged (Anas discors)/cinnamon teal (Anas cyanoptera) nests were eliminated from speciesspecific analyses. Because survey periods were short, the date on which nests were found was a function of search scheduling and effort rather than a reflection of species-specific nesting chronology. Therefore, an average median date for the number of .incubated nests found was determined for each species for each survey period (May and June), then the number of days between the 2 median dates was used to calculate an average daily change in clutch sizes between the 2 periods. . Wetland Development MVNWR is an intensively managed refuge with an extensive system of contour dikes, water conveyance systems, pumps, and water control structures. Because of differences in topography, soils, proximity to water sources and electricity, and ease of access, the extent of wetland development and intensity of wetland management varies among units. Past and ongoing management practices have been instrumental in changing vegetative composition and landscape patterns. Although the suite of wetland development and management practices collectively result in varying management intensity, such differences are ·difficult to quantify. Nonetheless, the degree to which management actions provide waterfowl benefits by promoting high nest densities and species richness is of great interest to refuge managers. Accordingly, the senior author (DWG) ranked the level of managed development (Qr) in 4 categories (0 no development to 3 - intense development) based on his knowledge of MVNWR management units. Our objective was to test the null hypothesis that rank of wetland development was not related to either duck nest densities or waterfowl species richness on MVNWR. Analysis of variance was used to evaluate these relationships (PROC GLM; SAS Institute Inc., 1987). Grazing Cattle have grazed on portions of the MVNWR annually since data collection began in 1964. However, a 3-year rest-rotation grazing system used during 1977-1987 (through 1990 in some units) presented a unique opportunity to assess the effects of grazing on nesting ducks. During this period, individual wetland management units were moderately grazed by cattle to remove decadent, residual vegetation (usually baltic rush). Grazing usually occurred between 15 September and 31 December. Except for a few units that were burned after grazing was completed, units usually were left undisturbed for 3 years, then grazed again. Grazing treatments therefore resembled a Latin square statistical design, in which treatment is replicated spatially and temporally among units (Fig. 1). Grazing intensity was measured in animal unit months (AUM). Our evaluation relied only on the years and units under �9 Year Unit 1 78 79 80 81 82 83 84 85 86 87 88 89 90 G G G G 2 3 G G G G G G G G G G G 4 6 1 1 G 7 1 G G 8 1 G G 11 1 14 G G G G G 10 G G G G 9 * G G G G G G G G G G * G G G * 15 G G G * 16 G G G * 17 1 G G G 18 19 20 1 G 1 1 G G 1 G G G G G G G G G G G G 21 22 1 G G G G * * * * * * * * Fig. 1. Dormant season, moderate intensity grazing regime evaluated for the Monte Vista National Wildlife Refuge. "G" denotes grazing, "_" denotes rest period, "1" denotes unknown activity, "*" denotes Savory System of grazing. �10 the rest-rotation grazing system and did not include those units switched to short-term, high intensity, growing season grazing. Duck nest density and nest success were used as response variables to evaluate the effects of grazing on 19 of 24 units in 12-14 years. Unsurveyed units were excluded from analyses. For analyses of the effects of grazing under a 3-year rest-rotation system, u~its and years were treated as class variables and estimated nest density (Q) was regressed on grazing intensity (I), a continuous variable measured in AUM/acre, and sub-period (~). Sub-period denoted either' the first, second, or third nesting season following grazing in the 3-year restrotation schedule. This approach yielded the regression model where i-year - 1977-1990 j - unit 1,2,3,4,6,7,8,9,10,11,14,15,16,17,18,19,20,21,22. Qo - intercept b1 vector of partial regression coefficient for the 19 units b2 vector of partial regression coefficient for the 14 years. ~ and Q4 - partial regression coefficients for I and S2, respectively. Eij - error terms, assumed to have mean of zero, independent and constant variance. If nest density increased in a year following increased grazing, we expected ~ to be positive; if nest density decreased following increased grazing intensity, then ~ would be negative. The null hypothesis of no effect of grazing is ~ - O. Our s priori belief was nest density decreased as grazing intensity increased, therefore we treated the test of the null hypothesis as one-sided. Sub-period was treated as an ordered independent variable. The null hypothesis of interest was that Q4 - 0; the one-sided alternative hypothesis was that nest density increased each sub-period (year) following grazing (thus Q4 > 0). The response of nest density to years of non-grazing was slightly nonlinear, therefore S2 was used in the model. We explored a logtransformation of Q, but found no improvement in model fit; because this made interpretation more difficult, those results are not reported here. Recognizing that large amounts of variation in nest density are attributable to the effects of weather and quality of management uni.ts, we treated year (i) and unit (j) as class variables to remove these sources of variation, allowing us to focus on the main hypothesis of interest. The effects of the rest-rotation grazing system on nest success were studied using the same multiple regression model, except S was used as an independent variable (instead of S2). Specifically, where p is the proportion successful and logit(Pij) is the empirical logit transformation. �11 Burning Burning was first used as a habitat management technique in 1981. Prescribed burns were conducted in late winter or early spring, often after units were grazed, to remove dead, residual vegetation. The approximate size of burns was estimated visually by refuge staff. Cases in which coverage of the burn was <50% of a unit were excluded from analyses. The effect of burning was evaluated by comparing nest densities the year prior to burning to nest densities each of the following 2 years using paired T-tests. Haying Hay was cut on the MVNWR during 1964-1970, but records were incomplete with respect to the amount of each unit that was hayed. We considered "hayed" units those in which some hay cutting took place during a given year. We included in our analyses nest densities in hayed versus unhayed units for the 1964-70 period. To account for unit differences independent of the haying treatment, we also examined nest densities in these same units during a second period (1984-1990). In doing so, we assumed that similar, inherent unit differences existed during both periods. The latter period was selected because nest densities and water conditions varied little during the interval, unlike the highly variable conditions that prevailed during the 1970's. We analyzed the effect of haying on nest density by treating the data as a 2-period quasi-experiment (Cook and Campbell 1979). An analysis of variance model was used with treatment (haying, only during the first period), period (1964-70 or 1984-90), and unit as main effects. All units were used except unit 5, which often had nest densities of 0, and unit 24, which was not surveyed until 1981. Our g priori belief was that by removing vegetative cover, haying would reduce duck nest densities. Therefore, we performed a 1tailed test on the null hypothesis of haying effect, evaluating significance using Type III sums of squares (SAS 1987). A logarithmic transformation of nest density, the dependent variable, was used to normalize the distribution and stabilize variances because of the small sample size (Hosmer and Lemshow 1989). Nest densities of 0 were assigned densities of 1 to enable log transformations. Predator Control Records of quantitative measures of predator control on MVNWR, such as number of trap-days or predators killed, were often unavailable. Thus, we reviewed refuge records and recorded mostly qualitative indicators of effort such as the use or non-use of poison baits, aerial gunning activity, the relative intensity of trapping activity, and whether trapping was directed at mammalian predators, avian predators, or both. Periods of years with apparently similar levels of effort and methodology were grouped for comparative purposes. We used the number of duck nests that failed from predation divided by the total number of failed nests as the response measure to evaluate the effectiveness of predator control. RESULTS Clutch Size Substantial numbers of mallard, blue-winged/cinnamon teal, northern �12 pintail and gadwall incubated nests were found during surveys (Table 1). The comparative distribution of the number of nests found by species between survey periods generally reflects the species chronology of nesting reported on the MVNWR by Schroeder et al. (1976). Unexpectedly, however, the numbers of mallard and northern pintail nests found was similar in May and June surveys. The median date for incubated nests found during the 26 year period varied only slightly between species (16-20 May and 16-18 June). Table 1. Average clutch sizes and number of incubated duck nests found during May and June surveys on the Monte Vista National 'Wildlife Refuge, 1964-90. Species Mean Mallard Tealb Pintail Gadwall Shoveler Redhead G-'W teal aClutch size bBlue-winged CClutch size dClutch size 7.94 8.71 7.09 7.92 8.59 8.20 6.00 Combined June May SD N 1.80 1.65 1.47 2.36 1.70 1.92 Mean 581 98 119 13 17 5 1 SD 7.64a 1.84 8.73 1.48 2.00 7.22 8.91C 1.73 8.71 1.96 9.52d 1.85 8.50 1.02 N Mean 572 183 90 150 51 44 6 7.79 8.72 7.15 8.83 8.68 9.39 8.14 SD N 1.83 1.54 1.72 1.80 1.89 1.88 1.02 1,153 281 209 163 68 49 7 different (f < 0.001) than in May. teal and cinnamon teal combined. different (f < 0.02) than in May. different (f < 0.005) than in May. (16-20 May and 16-18 June). Clutch sizes of mallard nests on the MVNWR did not vary among management units (F-25,039; 1,20 df; P<O.OOl, R2 - 0.999) and varied little by year (F-l,438; 1,24 df; P<O.OOl, R2-0.983). The intercept for management units was near zero (a - -8.67, se - 4.01) and the slope was 7.93 (se - 0.050). The intercept for years was not significantly different from zero (a - 19.4, se 16.8). Only in 1976 did clutch size vary from normal (Standardized residual -2.82). 'With the results for mallard clutches as an indicator, clutches for all species were pooled across units and years for subsequent analyses. For most species, May clutches were similar to those found in June (Table 1). Smaller May than June clutches were recorded for gadwall (~- 2.57, 161 df, f < 0.02) and redhead (Aythyaamericana~ - 2.05,47 df, f < 0.05), which are character-istically late nesting species. Mallard clutches were smaller in June than May (~- 3.78, 1151 df, f < 0.001), but the difference was not substantial. Decline in average clutch size for mallards between the 2 survey periods was only 0.3 eggs, or a reduction of 0.0096 eggs/day (0.30 eggs + 31 days) . 'Wetland Development Simple regressions of wetland development with duck nest.density and duck species richness were ~tatistically significant and biologically reasonable. Average nest density (Q) was correlated (R2 -AO.22; r - 6.75; 1,20 df; P 0.017) with the rank of wetland developmenk (Q - 224.6 + l56.l(4r); Fig. 2). Similarly, average duck species richness (Bs) and the rank of wetland �13 -- ~ R 2= 0.21 o en -- en ten P = 0.01 o 1,500 o W Z ~ en o 1,000 ······················-0··· - -_ ... z w o o t- en w Z Z o ................................ 0 500 B -c W o B o o ~ 0 0 0.5 1 1.5 2 2.5 3 DEVELOPMENT RANK Fig. Z. Mean duck nest density in relation to intensity of wetland development by management unit. �14 development (2:5 _ 1.07 + 0.769(4r)) were also positively correlated (R2_ 0.46; r - 19.2; 1,20 df; P < 0.001; Fig. 3). Based on the success of the simple wetland development model, we developed a multiple regression model based on the hypothesis that duck nest densities should be (1) positively related to wetland development, and (2) the proportion of a unit vegetated with preferred nesting vegetation, and (3) negatively correlated with the proportion of the unit consisting of avoided nesting vegetation. This resulted in the model A 28.4 + 14.6 (RQ) + 1.1 (GH) - 0.66 (PH) where A E(Qj) average (over years) nest density for unit j, Ro - ranked measure of the amount of development on each unit j. Ro - (0,1,2,3) where 0 represents an undeveloped unit and 3 represents a unit with dikes, borrow pits, and several water control structures, GH - preferred nesting vegetation, expressed as the proportion of each unit j consisting of rush and cattail, and PH - avoided nesting vegetation, expressed as the proportion of each unit j consisting of sa1tgrass and greasewood. This model was significant (F - 6.95; df - 3,14; f - 0.004; R2 - 0.51), as were all estimated coefficients (f < 0.05). Grazing There was a reduction in total animal unit months (AUM) during the rest rotation period (Fig. 4), but average annual grazing intensity increased. Grazing intensity varied from 0.1 to 1.8 AUM/acre. Nest density varied across years (f - 0.0002) and units (f - 0.0001; Table 2). Effects of the grazing treatment were evaluated by a 32 parameter regression equation. The equation was expressed for an average year and unit by taking the means of the partial regression coefficients for years and units (b1 and~, respectively), then expressing estimated nest density (Q) in nests/square mile as Q - 590 - 263.7(1) + 20.25(s2). A �15 6~------------------------------------------------~ o (J) 5 (J) W o = 1 observation Z :::c o o o 4 c: a: =2 observations (J) W o 3 o W a, (J) 2 ......................... § Z -c W ~ - -~ _ ......•.. _ .•.... _ 1 __ -_ __ o. o _ _ _ _ - . o LJ.... ..... _-'. o~----~------~------~----~------~----~~----~ o 0.5 1 1.5 2 2.5 3 3.5 DEVELOPMENT RANK Fig. 3. Number of duck species nesting in management units as a function of the intensity of wetland development. �16 6000 , . . 5,000 :aE 4,000 :J < 3,000 2,000 1,000 o 1964 1968 1972 1976 1980 ., 1984 1988 YEAR Fig. 4. Annual grazing rates in animal units months on the Monte Vista National Wildlife Refuge. �17 Table 2. Analysis of variance and regression results for the 3 year, restrotation grazing program at MVNWR. Model: E(nij) - 12.0 + b1(i) + b2(j) Analysis of Variance .., .., + 12.J(I) + ]4(S2) + Eij Sum of Squares (Type III) Mean Square F value Pr > F 32 44,206,979.84 1,381,468.12 8.36 0.0001 year (j) 12 6,663,819.74 555,318.31 3.36 0.0002 unit (i) 18 36,249,385.53 2,013,854.75 12.19 0.0001 intensity (I) 1 529,481.83 529,481.83 ·3.20 0.0752 sub-period (S2) 1 854,069.90 854,069.90 5.17 0.0242 Error 175 28,920,182.12 165,258 Corrected total 207 73,127,161.96 Source Model R2 - 0.605 df RMSE - 406.5 The partial regression coefficient for grazing intensity (-263.7, SE147.3) indicates that nest density declined (t - 1.79, one-tailed l - 0.037) as grazing intensity (I) increased. The partial regression coefficient for sub-period; the first, second, or third nesting season following grazing in the 3-year rest-rotation schedule, (20.25, SE - 8.9) reveals that nest density increased (t - 2.27, one-tailed l - 0.01) during the 3 years following the grazing treatment (Fig. 3). Examination of residuals from the model (Draper and Smith 1981) revealed that the 8 most extreme points (all positive) were associated with the 3 best units (units 3, 9, 18) in good years. Thus, the model failed to predict the density of nests in the best units in generally good years, but otherwise fit the data well (R2 - 0.60, l < 0.001). We used the regression model to predict the decline in nest density following moderate grazing, and to simulate increases in nest densities each year following the termination of grazing. At 1.0 AUM/acre, predicted densities in years 1, 2, and 3 are 347, 407, and 509 nest/square mile, respectively. At 0.5 AUM/acre, comparative densities are 478, 539, and 640 nests/square mile. No control (ungrazed) units were available, therefore we predicted the average nest density in an average year and unit by setting I - 0 (no grazing) and S - 3 years (a predicted control). This allowed an estimate of nest density in the absence of grazing without extrapolating beyond the data, but provided a minimum estimate of average nest density under the long-term absence of grazing. The resulting estimate of 772 nests/mi2 indicates that nest density decreased 38% the first season after grazing at 0.5 AUM/acre, and would decrease 55% the first season after grazing at 1 AUM/acre (Fig. 5). We believe the estimate of average nest density on ungrazed areas is very conservative, because total recovery in nest density would take >3 years. The effect of the 3-year, rest-rotation grazing system on nest success can �18 1,000 -- C\I ~ 800 ~ en w z 600 ~ en z w 400 '0 I- en W Z 200 o 0.5 1.0 1.5 2.0 1 SUBPERIOD GRAZING INTENSITY (AUM) Fig. 5. Estimated nest density as a function of cattle grazing intensity' (AUM/acre) and number of nesting seasons since the cessation of grazing (subperiod) on the Monte Vista National Wildlife Refuge. �19 be summarized by the simple model logit(p) - 0.591 - 0.407(1) or, equivalently, 1 P = ----~~~~~~ 1+e - (0.591-0 .407 (I» where p is the proportion of successful nests and I is the grazing intensity (ADM/Acre). This indicates that nest success decreases as grazing intensity increases (t - -1.48, f - 0.070; Fig. 4). Subperiod and unit were not significant (f> 0.05), but nest success varied by year (f- 0.003). Using the equation for ~, the estimate of nest success in the absence of grazing is 0.644, or approximately 64% (Fig. 6). Burning Ten prescribed burns occurred on MVNWR during 1964-90. Some burns covered parts of 2 units but ~50% aerial coverage occurred in only 8 units (Table 3). Nest densities in those 8 units before burn treatments (~- 971 nestimi2, SD 689) were greater (f - 0.05, paired t-test) than nest densities in the nesting season following burning (~- 534 nests/mi2, SD ~ 520). However, by the second nesting season after burning, duck nests increased (~ - 904 nests/mi2, SD - 762) to densities similar (f - 0.82, paired t-test) to those observed before burning. Table 3. Unit 4 9 19 14 15 17 3 Prescribed burning treatments on the MVNWR, 1964-1990.~ % burned 39 ioo28 42 48 75 33 Year Unit 1982 1983 1983 1984 1984 1984 1984 4 18 6 8 17 9 10 % burned 47 60 100 100 73 50 72 Year 1984 1984 1989 1989 1989 1989 1989 BAcreage burned was generally estimated by refuge staff. blncomp1ete unit burns usually resulted from greasewood vegetation not burning. CReported as: "burned unit 9", assumed to be a complete burn. Haying In 1964, hay was cut on 14.4% (1,665 acres) of the refuge in units 4, 7, 14, 20, 22, and 23 (Bill McDermith, Monte Vista National Wildlife Refuge, �20 0.7 ..----------------------'-------. _J ::J 11. en en w o o ::J en 1 p= 1 + Ef (0.591 - 0.407 ( I » ~ en w z 11. o Z o 0.5 . t= a: o a.. o a: n, --L.. 0.4 L....- o 0.5 _,..,_ 1 I...1.5 ......,.j_, 2 GRAZING INTENSITY (AUM / ACRE) Fig. l:,. Duck nest success as a function of cattle grazing intensity on the Monte Vista National Wildlife Refuge. ~ 2.5 �21 pers. commun.) The program was reduced, and eventually eliminated, by 1977 (Fig. 7). During period 1 (1964-70), the untransformed nest density on hayed 2 areas (~ - 257 nests/mi , SD - 176, N - 42) was higher (~ - 122.8, f - 0.0001) than that on unhayed units (~ - 814 nests/mi2, SD - 676, N - 98; units 5 and 24 excluded). However, during the second period (1984-1990) when no units were hayed, nest densities also differed (~ - 4.27, f - 0.0001) between the unit groups (~- 356, SD - 323, N - 42 and ~ - 721, SD - 687, N - 98, respectively). The ANOVA model (g2 - 0.35) revealed significant unit (E 5.47, f - 0.0001), period (I- 8.68, f - 0.0035), and treatment (haying) effects (I- 3.32, I-tailed f - 0.035) on log transformed nest density. Unfortunately, the rigor of our analysis is reduced by the absence of a randomized design and confounding with other management treatments. Nevertheless, even after accounting for variation in nest density attributable to unit and period differences, haying appeared to have detrimental effects on duck nesting densities. Predator Control Predator management 1964-1990 (Table 4). records revealed 3 levels of predator control From 1964-1970 (period 1), nest predators Table 4. Summary of predator control efforts Specific records are on file at the MVNWR. Control Poison Magpie Destroy Mammal Aerial Overall Period 1 bait trapping magpie nests trapping coyote gunning control effort 4Probably not as effective yes no no yes yes High as period 1964-1990. on MVNWR, Period during 2 Period no yes yes limited yes no yes some yes yes Low High4 3 1 when poison was used. (mainly ravens, magpies, skunks, raccoons, and coyotes) were poisoned using strychnine bait and trapped, usually with large, conibear bait traps. Because poisoned animals may wander from the site after ingesting toxin, carcasses were difficult to locate and the effectiveness of the control effort was largely unquantified. However, based on partial records of animals killed and the known lethality of poison bait control methods, we categorized this first period as one of very intensive predator control. From 1971-1980 (period 2), incomplete records were found on the numbers of trapped mammals. Extensive aerial gunning for coyote (Canis latrans) occurred during this period. Poisoning ceased, but magpies (Pica pica) were controlled by deploying bait traps for adults and destroying nests with young. We considered 1971-1980 a less aggressive predator control period than period 1. In period 3 (1981- �22 - 1,500 en w a: U :$, 0 1,000 W ~ :c « w a: « 500 o 1964 1966 1968 1970 1972 1974 1976 YEAR Fig. 7. Area of hayground on the Monte Vista National Wildlife Refuge. �23 1990), predator management intensified from 1971-80. Predator removal was quantified arid data were retained on file. Mammals and avian predators were trapped and shot during this period. We considered predator control efforts high for this period, but at a level between periods 1 and 3 .. The percent of nests destroyed by predators varied considerably by year (Fig. 8). Predated nests accounted for 16.6%, 33.7% and 25.1% of all nest losses during periods 1, 2, and 3, respectively. The percentage of nests that failed from predation varied among predator management periods (I - 6.89, 2 df, f - 0.005). Nest failure through predation was less during period 1 (f < 0.05) than during period 2, but period 3 did not differ from either earlier period. DISCUSSION The success or failure of habitat management programs can only be understood within the context of local habitat characteristics and the biological requirements of waterfowl populations. San Luis Valley waterfowl encounter a breeding habitat fundamentally different than that in the northern Great Plains, the major continental breeding area, largely because of unique climatic conditions and topography. Extremely cold winter temperatures result from the high elevation of the SLV and from downslope drainage of frigid air off surrounding mountains. The deeply frozen soils and short growing season during the warmer months promote a simplified plant community. A precipitation-evaporation balance <1 allows development of pedoca1 soils and alkali buildup (Marr 1967), which further restricts plant distributions. Although precipitation is low due to the rain shadow effect of surrounding mountains, spring wetlands are often abundant because of snowpack runoff and seasonally high water tables. During the brief summer period, wetland primary productivity is high and aquatic invertebrate populations develop quickly, similar to the bloom in productivity more typical of arctic environments (Ringelman 1992). Banding information (Hopper et al. 1975, Szymczak 1986) and radiotelemetry data (Jeske 1991) indicate that about half of the mallards that breed in the SLV are resident year-round. Non-resident mallards and other duck species have a high homing rate to the SLV and the MVNWR, consistent with their general life history strategies (Johnson and Grier 1988). Resident mallards incur high energetic costs during winter, particularly for thermoregulation and flight (Jeske 1991). Winter food resources are limited, primarily consisting of barley and field peas remaining as residue after harvest (Jeske 1991). Consequently, SLV resident waterfowl incur a negative energy balance, resulting in endogenous lipid reserves during winter and early spring that are substantially lower than other mallard populations (Jeske 1991). Non-resident breeding mallards, however, likely originate from southern wintering habitats, such as the Playa Lakes of Texas, where mallards retain high levels of endogenous reserves during winter (Whyte et a1. 1986). Breeding behavior, resource use, and response to habitat management are consistent with the notion that 2 distinct mallard sub-populations contribute to recruitment on the MVNWR. Spring Nesting Chronology and Resource Use Jeske (1991) documented nesting chronology of mallards on the MVNWR by determining the age of mallard broods according to plumage development (Gol10p �24 END OF POISON DECREASED TRAPPING / ACTIVE TRAPPING AVIAN AND MAMMAL \/ \ 801-·············································· POISON AND ACTIVE TRAPPING \ / •.•.60 ~ . Z W o a: w CL 40 ~ . 20 o L- L... 1964 L.. 1968 1972 1976 1980 1984 1988 YEAR Fig. e. Percentage of duck nests that were destroyed by predators during three periods of varying predator control effort. �25 and Marshall 1954), then backdating to the date of hatch. Although nesting chronology varied annually, early clutches began hatching in mid-May. Assuming a 28-day incubation period (Girard 1941), an 8-day laying interval (based in 1 egg/day; Be11rose 1976), and 10 days for nest site selection, these early-nesting mallards must have begun establishing territories in midMarch. Typically, ducks during this pre-nesting period are aggressive towards conspecifics (Sowls 1955, Dzubin 1969). However, mallards on the MVNWR exhibit little evidence of territoriality such as 3-bird-flights (McKinney 1965). Even while water is pumped into wetland units beginning in Late March, field-feeding flights to barley fields continue up until the time they are plowed in late April (J. K. Ringe1man, Colorado Division of Wildlife, unpublished data). Notably, aggressive interactions do not increase markedly even after field-feeding subsides (C. W. Jeske, u.s. Fish and Wildlife Service, Lafayette, LA, pers. commun.). We believe these late field-feeding flocks are composed largely of resident mallards that are attempting to reacquire endogenous fat reserves depleted during winter. As such, they are heavily dependent upon high carbohydrate foods such as waste grain and seeds in recently flooded wetland units. Because these birds must first acquire a threshold of endogenous reserves prior to initiating egg laying (Krapu 1981), we suspect that SLV resident mallards may nest later than non-resident birds, thus accounting for the similar number of mallards nests found during May and June surveys, as well as the relatively protracted distribution of hatching dates (Jeske 1991). Application of early water by pumping and artesian flow is critical to attracting and retaining breeding ducks on the MVNWR (Schroeder et al. 1976). Shallow flooding not only makes seeds and plant materials available, but also promotes the development of aquatic invertebrates (Schroeder 1973), which are requisite foods for breeding ducks (Swanson and Meyer 1973). Abundant potential nest sites are created as water surrounds small, elevated hummocks of baltic rush. Given the sudden abundance of both food and nest sites, as well as the short breeding season imposed by climatic conditions, SLV mallards appear to forego establishment of traditional breeding territories typical of prairie-nesting mallards. Compression of territory size has been observed for mallards breeding in high (>40 pairs/mi2) densities, a phenomena attributable to incomplete dominance of drakes (Dzubin 1969). Certainly, dramatic compression or even elimination of breeding territories must occur among mallards as well as other duck species to account for the extremely high pair densities (>300 pairs/mi2 average, assuming at least 1 pair/nest) found on the MVNWR. Excluding island-nesting situations, several units on the MVNWR have higher duck nest densities than any other region in North America (Table 5). Nest Success High nest densities, particularly when nests are dispersed on hummocks within flooded wetlands, have the potential to temporarily overwhelm local predator communities and result in high nest success rates. Nest success appears to be further enhanced by predator control activities on the MVNWR. As a context for evaluating MVNWR nest success rates, we reviewed reported duck nest success for other regions of North America (Table 6). Contemporary studies, which reflect recent drought and habitat degradation in the northern Great Plains, indicate that nest success averages <20%. Only 2 locations that were not under intensive predator management had nest success rates greater than those on the MVNWR, but these data were derived only within a short period of study (Table 6). Because duckling production is the product of nest �Table 5. Reported duck nest densities in North America, exclusive of island-nesting situations. Cover type and location ~et meadow, Unit 6, MVN~R, Colo. Idle agricultural land, South Dakota" Unmowed hwy. right-of-way, North Dakota ~et sedge meadow, Manitoba Cat ta it, Alberta Baltic rush, Alberta ~et meadow, MVN~R, Refuge-wide Idle agricultural lands, South Dakota Retired croplands, ~isconsin Seeded grasslands, North Dakota Idle agricultural lands, South Dakota Native prairie, North Dakota Mixed prairie, Alberta Untilled upland, North Dakota "Total nlJ11berof nests estimated using nne "Area surveyed each year. <Predator control area. Table 6. Density (nests/mi') No. nests 1,164 899 454 427 400 376 329 312 311 214 198 104 74 52 _" 1,062 447 238 64 480 _" 61 691 1,455 620 2,292 155 93 Size of area (mi") - Idle agricul tural lands, South Dakota" ~et sedge meadow, Manitoba· Idle agricultural lands, South Dakota· Ungrazed grassland, North Dakota· Untilled upland, North Dakota· ~et meadow, MVN~R, refuge-wide, Colo. Seeded grassland, North Dakota Native prairie, North Dakota Grassland, North Dakota Planted cover, North Dakota Rainwater basin, Nebraska· Wetlands, North Dakota Retired croplands, ~isconsin Right-of-way, North Dakota nests are reported. Present Study Duebbert and Lokemoen 1980 Oetting and Cassel 1971 Oetting and Dixon 1975 Keith 1961 Kei th 1961 Present Study Duebbert 1969 Livezey 1981 Higgins et al. 1992 Duebbert and Lokemoen 1976 Higgins et al. 1992 Duncan 1987a Higgins 1977 6 3 1 5 5 26 1 3 16 3 16 2 22.40 0.64" 1.78 >50 Source 26 1.05 6 transect theory. Reported duck nest success in North America, exclusive of island-nesting Cover type and location No. years 0.97" 0.20 0.98 0.56 0.03" 0.25" 18.32" 0.20 0.91 6.90 . Only studies with Success (X hatched) No. nests 85 73 56 28 25 52 18 16 14 12 11 9 9 8 1,062 238 620 221 93 4,156 1,366 2,081 1,744 6,031 206 271 691 542 situations. Size of study area (mi2) 0.20 0.56 1.05 0.24 1.78 18.32 ·Nest success not Mayfield corrected ·Predator control area ·Size of study area was not given dSize of study area was not given, included 61 sites in east and central North Dakota. - -• --- dd c -9.00 -0.91 -- Only studies with >50 nests are reported. No. years. 6 1 3 4 6 26 16 16 8 8 5 8 3 8 Source Duebbert and Lokemoen 1980 Oetting and Dixon 1975 Duebbert and Lokemoen 1976 Kirsch 1969 Higgins 1977 Present Study Higgins et al. 1992 Klett et al. 1988 Evans and Wolfe 1967 Klett et al. 1988 Livezey 1981b Klett et al. 1988 N 0'1 �27 density and nest success, our reviews suggest that the MVNWR is one of the most, if not the most, productive duck breeding area in North America. Notably, it has sustained this level of production over a 26-year period. Clutch Size The low endogenous reserves of SLV resident mallards and the brief breeding season typical of intermountain basin habitats prompted us to examine the hypothesis that the clutch size of mallards in the SLV is smaller than the clutch size of mallards elsewhere in North America. If optimum clutch size represents a balance between endogenous reserves available for egg production and time available for completion of the breeding cycle, we reasoned that SLV mallards may lay fewer eggs because of sub-maximal reserves and/or the necessity to proceed with nesting in a timely manner. Despite variation in unit habitat quality, mallard clutch size did not vary across management units. Similarity in clutch size across the MVNWR was not unexpected, because refuge water supply and water management was generally consistent from year to year. Moreover, the home range of nesting mallards would have overlapped unit boundaries, thereby confounding unit effects. Below normal mallard clutch sizes were found in dry years in North Dakota (Krapu et a1. 1983). Mallard clutch sizes in 1976 on the MVNWR were below normal, however 1976 was not especially dry in the SLV (Fig. 9); water use on the MVNWR was the fifth lowest recorded during the study period (Fig. 10). Generally, clutch sizes of all duck species on the MVNWR were smaller than range-wide averages reported by Bellrose (1980), particularly for mallards (avg. -1.18 eggs, Table 6) and gadwall (avg. -1.31, Table 7). Mallard clutch sizes recorded SLV-wide by Ryder (1951), although larger than those recorded on the MVNWR (Table 6), were generally more comparable to MVNWR mallards than to those in other geographic areas. Only northern pintail clutches on the MVNWR did not show consistent deviation in size from clutches in other areas (Table 8). Sufficient range-wide data were unavailable to adequately compare early and late clutch sizes on the MVNWR with similar data from other geographic areas. However, early clutches were smaller, late clutches were larger, and the differences in clutch size between the 2 periods was smaller, than observed in other areas (Table 9). Reductions in clutch sizes as the nesting season progresses have been demonstrated for most prairie-nesting duck species (Blohm 1979, Cowardin et al. 1985, Duncan 1987b, Lokemoen et a1. 1990 and Higgins et al. 1992). But on the MVNWR, only mallard clutch sizes declined and the reduction, although statistically significant, was not large. Linear declines in within-season clutch sizes in populations of wild mallards have ranged from 0.027 eggs/day (Lokemoen et a1. 1990) to 0.06 eggs/day (Cowardin et a1. 1985). Batt and Prince (1979) found that clutch sizes for first nesting attempts of captive mallards were reduced 0.1 egg per day as the nesting season progressed. The latter authors also found that the clutch size/initiation date relationship was slightly curvilinear when all nests initiated were considered, indicating a possible minimum clutch size. Although data were not available to measure the shape of the curve describing the within season decline in mallard clutch sizes on the MVNWR, the reduction in clutch size between early and late nests was only 0.01 egg/day, far less than reported in wild populations elsewhere. �28 Table 6. Comparative clutch size of mallards from selected North American studies. Differences in mean clutch sizes were compared between this study and other areas when standard deviations and sample sizes were available. Clutch size Early Area Mean SLY (MVNWR) SLY (valley-wide) Roseneath, Alb. Kindersley, Sask. S.E Alberta North Dakota New England Cal ifornia 7.94 (1.80) "clutch size different bClutch size different ·Clutch size different Season-long Late (SO) Mean 581 8.55 (1.69)' 111 9.12 (1.61 )' 529 9.6 (SO) Mean 7.64 (1.84) 572 7.21 (1.48)b 7.70 (1.98) 8.1 55 98 (SO) Source 7.79 (1.83) 1,153 This study 8.21 (1.62)' 165 Ryder 1951 8.11 (1.62)· 166 Ozubin and Gollop 1972 8.90 (1.67)' 627 Ozubin and Gollop 1972 8.8 (1.51)' 97 Keith 1961 9.5 (1.49)' 150 Krapu et al. 1983 9.6 (1.85)' 131 Coulter and Miller 1968 178 Miller and Collins 1954 8.9 (f < 0.001, 2-tailed) than comparative clutches in this study. (f < 0.02, 2-tailed) than comparative clutches in this study. (f < 0.01, 2-tailed) than comparative clutches in this study. Table 7. Comparative clutch size of gadwall from selected North American studies. Differences in mean clutch sizes were compared between this study and other areas when standard deviations and sample sizes were available. Clutch size Early Late .!! Area Mean SLY (MVNWR) S. Saskatchewan S. Alberta S. Manitoba Utah North Dakota 7.92 (2.36) 13 10.37 (1.37)' 295 9.8 'Clutch size different bClutch size different (SO) Mean Season-long (SO) 8.91 (1.73) 8.00 (0.99) 8.8 Mean 150 27 8.83 10.17 9.4 10.4 10.05 10.7 (SO) (1.80) (1.80)' (1.47)b (1.51)' (1.25)' .!! Source 163 322 55 77 141 161 This study Hines and Mitchell 1983 Keith 1961 Blohm 1979 Gates 1962 Ouebbert et al. 1983 (f < 0.0005, 2-tailed) than comparative clutches in this study. (f < 0.025, 2-tailed) than comparative clutches in this study. �29 TabLe 8. Comparative cLutch size of northern pintaiL from seLected North American studies. Differences in mean cLutch sizes were compared between this study and other areas when standard deviations and sampLe sizes were avaiLabLe. CLutch size EarLy (SO) Area Mean SLV (MVNWR) 7.09 (1.47) 7.4 7.2 9.0 s. ALberta S. ALberta s. Manitoba Utah s. Saskatchewan -CLutch size different bCLutch size different Season-Long Late (SO) !! Mean 119 7.22 (2.00) 6.0 6.2 7.1 45 (SO) !! Mean 90 7.15 (1.n) 209 6.9 (1.70)- 290 6.7 (1.33)b 79 7.9 105 8.26 (1.83)- 58 8.0 112 14 !! Source This· study Duncan 1987b Keith 1961 SowLs 1955 Fuller 1953 Stoudt 1971 (f < 0.05. 2-taiLed) than comparative cLutches in this study. (f < 0.01. 2-taiLed) than comparative cLutches in this study. TabLe 9. Average cLutch sizes and number of incubated duck nests found during May and June surveys on the Monte Vista NationaL ~iLdLife Refuge. 1964-90. May June Species Mean SO N MaL Lard TeaLb PintaiL GadwaL L ShoveLer Redhead G-~ teaL 7.94 8.71 7.09 7.92 8.59 8.20 6.00 1.80 1.65 1.47 2.36 1.70 1.92 581 98 119 13 17 5 1 -CLutch size bBLue-winged "CLutch size dCLutch size Mean Combined SO 7.64- 1.84 8.73 1.48 7.22 2.00 8.91" 1.73 8.71 1.96 9.52d 1.85 8.50 1.02 N Mean SO 5n 7.79 183 90 150 51 44 6 8.n 1.83 1.54 7.15 8.83 8.68 9.39 8.14 N 1.153 281 1.n 209 1.80 163 1.89 68 1.88 49 1.02 7 different (f < 0.001) than in May. teaL and cinnamon teaL combined. different (f < 0.02) than in May. different (f < 0.005) than in May. Proposed determinants of clutch size in temperate breeding waterfowl are numerous (Duncan 1987b) and have been subject of much study (Krapu 1981, Ankney and Afton 1988, Ankney and A1isauskas 1991), experimentation (Rohwer 1985), and discussion (see Drobney 1991, Ankney et a1. 1991, Arnold and Rohwer 1991). Hypotheses for seasonal clutch size decline in waterfowl are numerous (Duncan 1987b), but focus on the interrelated factors of body condition and available nutrient resources (Krapu 1974, 1981), later nesting attempts by first year females (Krapu and Doty 1979, Cowardin et al. 1985), the increased presence of renests (which usually have smaller clutches [Keith 1961, Batt and Prince 1979, Duncan 1987b]) later in the cycle, and a possible genetic component (Batt and Prince 1979). The small average clutch size of mallards �30 600,000,-----------------------., 28-YEAR MEAN 500,000 / a: w 400,000 ~ ~ u, 0 ~ ____ / '______ ___ --------- 300,000 W W u, I W a: 200,000 o « 100,000 o I 1965 I 1970 I I 1975 1980 I 1985 I 1990 YEAR Fig. 9. Acre-feet of water at the Del Norte gauging station of the Rio Grande River, an index to water availability in the San Luis Valley. �31 35,000 r----------------------------. 30,000 a: w ~ ~ u, 25,000 20,000 ...........................................................••.•.•............................................................................. o tuW 15,000 U. I W a: 10,000 ~ 5,000 ° I 1965 1970 1980 1975 1985 YEAR Fig. 10. Acre-feet of water used by the Monte Vista National Wildlife Refuge. 1990 �32 and lack of substantial reduction in measured clutch size between early and late nests supports our ~ priori hypothesis that optimum clutch size for SLV mallards may strike a balance between time available to acquire endogenous reserves and environmental constraint on the time available for nesting. Atypically small and seasonally consistent mallard clutch sizes make SLV mallards unique among the North American mallard population. MANAGEMENT IMPLICATIONS As evidenced by our qualitative evaluations of wetland development in relation to nest density, habitat management results in enhanced waterfowl production on the MVNWR. However, the strength of this inference is weakened by the non-randomized, ~ posteriori nature of our analyses. Units selected for intensive management may have also been those with the highest production before management intervention. Moreover, multiple treatments were applied to most management units, confounding cause and effect relationships. Some of these treatments, adapted from traditional agricultural practices, were not beneficial to breeding waterfowl. Other researchers have also concluded that many agricultural practices were detrimental to breeding ducks, either directly through destruction of the clutch and/or incubating hen, or indirectly by habitat destruction or modification. Earliest works (Bennett 1937, Bue et al. 1952, Murdy 1953) condemned all but light grazing in waterfowl nesting habitat. After an extensive literature review, Kirsch (1969) found no conclusive evidence that supported the use of grazing as a beneficial waterfowl management tool, although Kirby et al. (1992) point out that poor studies designs often preclude valid scientific inferences. Evaluations of rest-rotation grazing systems suggested that ducks responded to pastures ungrazed the previous year by nesting at higher densities and with greater success rates than on similar grazed pastures (Gjersing 1975, Mundinger 1976). However, recent work (Ruyle et al. 1980) indicates that delayed grazing may improve waterfowl habitats. Enright (1971) speculated that grazing on the Monte Vista National Wildlife Refuge benefitted nesting waterfowl by setting back vegetative succession and breaking up monotypic stands of nesting cover. Only because of the extensive spatial and temporal replication of grazing treatments were we able to unequivocally determine the effects of this treatment on duck nest density and success. Clearly, grazing on the MVNWR had severe and long-lasting effects on both nest density and success. Units grazed at 2 ADM/acre had nest success rates that were 69.2% of those predicted for ungrazed areas (45% versus 65%), and nest densities that were 16.7% of those predicted for ungrazed areas (100 nests/mi2 versus 600 nests/mi2). The product of these differences, 0.116 (0.692 times 0.167), indicates that duck production on grazed areas was only 11.6% of that expected for ungrazed units. Mowing grasslands for hay is thought to produce detrimental effects similar to those described for grazing (Labisky 1957, Martz 1967, Kirsch et al. 1978, Voorhees and Cassel 1980). Burning nesting cover, often as a follow-up treatment to grazing, has also been used to prevent encroachment of woody plants and rejuvenate rank vegetation (Kantrud 1986, Kirby et al. 1988). Although burning releases nutrients, sets back natural succession, and breaks up monotypic stands of vegetation, it also destroys ground litter that is an important component of attractive nesting cover (Enright 1971), and may negatively effect nesting densities in subsequent years (Bouffard et a1. 1987). Our evaluations of the effects of burning and haying are somewhat �33 equivocal, but tend to mirror those observed for grazing. All 3 treatments result in the removal of dead residual vegetation, primarily baltic rush, the most abundant vegetative type as well as a preferred nesting vegetation type. Cover that remains after these treatments is short and spare, not unlike the structure of saltgrass, a vegetation type that was avoided by nesting ducks. Such low-growing or sparse cover also had lower nest success rates than found in denser vegetation. The importance of residual nesting vegetation of sufficient height and density, a well established relationship for prairienesting areas (Dwernychuk and Boag 1972, Schranck 1972), appears equally valid for intermountain basin habitats like the San Luis Valley. Characteristic of Preferred Units Microhabitat evaluations of nest site concealment, canopy cover and similar physical attributes have fostered conflicting views on the importance of morphometric features to duck nesting density and success. Some studies have found no relationship between nest success and nest site vegetative characteristics (Dwernychuk and Boag 1972, Byers 1974), whereas others (Odin 1957, Choate 1967, Chesness at al. 1968, Jones and Hungerford 1972, Enright 1971, Krasowski and Nudds 1986, Crabtree et al. 1989) found a positive relationship between these factors. The discrepancy may be caused by sitespecific differences and the role of vegetation in discouraging movements of nest predators or inhibiting the visual or olfactory detection of prey. This interaction between predators and nesting vegetation is an important consideration in managing nesting habitat, and has also received much attention by researchers. Despite high average nest densities and success, some units on the MVNWR were clearly preferred by nesting ducks. Typically, such units contained large amounts of baltic rush, a preferred nesting vegetation, surrounding the nest site. The two best units, 9 and 18, also had a large amount of cattail interspersed within rush. The amount of cattail throughout the MVNWR appears to be increasing (D. R. Anderson, W. O. McDermith, pers. commun.). None of the highly productive units had extensive open water or a large amount of greasewood. The best units were gently sloped with numerous small earthen dikes, which keep much of the vegetation flooded by shallow water. Enright (1971) reported a 2 foot drop west to northeast in unit 18, the second best nesting unit on the MVNWR. Where steeper gradients exist, such as in unit 19 and some western units, dikes tended to make large ponds but not areas of shallow flooded vegetation. Low nesting density units were typified by large expanses of greasewood with a limited herbaceous understory or a high percentage of saltgrass. Often, these units were dominated by large expanses of upland habitat complimented by relatively small areas of developed wetlands. Our efforts to account for variation in nest density in relation to vegetation patterns at the landscape scale were largely unsuccessful. Part of our inability to discern patterns may be a function of the 50 m2 cell size we selected as our sampling unit. Further progress on this issue will necessitate information on individual nest locations in relation to vegetation patches, followed by a variable scale approach to correlating nest site location and density with landscape features. Intensive predator control appeared to reduce the percentage of duck nets destroyed by predators on the MVNWR. Direct control of predators through trapping or shooting has also been observed to increase duck nesting success in other regions of the country (Balser et al. 1968; Duebbert and Lokemoen �34 1980). However, these activities require a continuing effort or animals quickly recolonize the area (Greenwood 1986). An alternative approach is to manipulate land use practices and vegetation to minimize attractiveness to predators (Capel 1965, Ladd 1969, Miller 1971, Duebbert and Kantrud 1974, Higgins 1977, Holm 1984). Because avian predators hunt by visual cues (Dwernychuk and Boag 1972, Picozzi 1975, Strang 1980, Sugden and Beyersbergen 1987) but mammals often use olfactory detection (Schrank 1972, Duebbert and Kantrud 1974, Duebbert and Lokemoen 1976, Livezey 1981a, Hines and Mitchell 1983), strategies for minimizing predation rates may differ according to the nature of the habitat, composition of the predator community, and abundance of alternate prey (Crabtree and Wolfe 1988). Thus, the complex interactions among predators, vegetation and nesting ducks suggest that effective habitat management prescriptions must be matched to the characteristics of individual nesting regions and backed up by evaluations. MANAGEMENT RECOMMENDATIONS Management Treatments Ducks prefer dense, residual baltic rush vegetation for nesting cover on the MVNWR. Consequently, if refuge objectives include maximizing duck production, activities such as grazing, haying, and burning should be curtailed. The short growing season in the SLV results in a long recovery time for vegetation removed by these activities. Because decomposition rates are slow in the SLV, residual vegetation may eventually become so dense as to insulate the soil and block light penetration, thereby choking out new growth. In such a case, residual vegetation removal may be needed to rejuvenate the nesting cover. However, based on our results, the frequency of such treatments should be longer than the 3 years practiced in the past; perhaps on the order of every 6 or 7 years. We recommend periodic height-density measurements of vegetation at nest sites and random locations as a means to assess changes in height-density preferences by ducks. Together with continual tracking of nesting densities using line transects, these heightdensity measures can be used to detect declines in vegetation quality which signal a need to initiate vegetation renewal. When removal is deemed necessary, consideration should be given to using endemic herbivores such as bison, rather than domestic livestock. Native herbivores may not only provide suitable grazing treatments, but are consistent with new initiatives to maintain biotic integrity and enhance biodiversity on National Wildlife Refuges (Kirby et al. 1992). Application of water in late March is a valuable management technique, both for attracting migrant ducks and for providing food resources to resident mallards who are striving to rebuild fat reserves depleted during the winter. Water run across wetland units flushes alkali salts from the soil, warms quickly and thereby promotes rapid development of aquatic invertebrates necessary for laying hens and ducklings, and saturates soil and vegetation so hens can readily distinguish dry nesting site locations. As in other studies of breeding ducks (Doty and Rondeau 1987, Duebbert and Kantrud 1974, Balser, et al. 1968) nest success on MVNWR appeared to increase in response to intensive predator control. Often, however, high nest losses caused by predation are a symptom of imbalances in the predator-prey community that might be addressed through habitat enhancements. For example, enhancement of the height and density of vegetation through cessation of grazing may mitigate nest losses caused by predation. If predator control is �35 deemed desirable, we suggest a multi-year experiment using treatment and control areas in a sophisticated design to evaluate the effectiveness of the control program. Much of this could be done within the framework of the existing transect monitoring program. Clearly, predator control also bears on ethical issues of wildlife management on National Wildlife Refuges, a system that is newly committed to maintaining a modicum of biotic integrity and balance in its faunal community. Intensive wetland development on the MVNWR results in high nest densities and duck species richness. Given the large variation in nest success and density among management units, additional wetland development could boost duck production on several areas of the MVNWR. Extremely high nest densities and nest success typical of this refuge would result in substantial increases in production per unit of ground developed, particularly when compared to more costly developments in rich agricultural lands of the northern Great Plains where nest success is much lower. Such developments would also restore wetlands in a valley that has experienced extensive wetland losses in the past, thus increasing habitat for other waterbirds and wetland wildlife and enhancing the regional biotic diversity of the San Luis Valley. Biomonitoring Wetland succession and attendant changes in plant communities occur relatively slowly, particularly in montane wetland systems where growing seasons are short. Thus, long-term data are needed for monitoring gradual changes over decades. The value of such datasets becomes even greater when attempting to evaluate treatment effects such as grazing, burning, or predator control, because large annual and spatial variation in water and available wetland resources create large changes in response measures such as nest density and success. The importance of biomonitoring has been the focus of much recent attention (Likens 1989, Risser 1991, Kirby et al. 1992). Appropriately, 6 of the 9 management alternatives being considered in Refuges 2003 Plan for the Future of the National Wildlife Refuge System call for increased emphasis on monitoring resources. The 27 years of management and biological data reported from the MVNWR result from >10,000 miles of transect and >300 weeks of time. By virtue of its design as well as level of effort, the MVNWR undoubtedly has one of the best biomonitoring programs in the National Wildlife Refuge System. Nonetheless, minor modifications could greatly improve the quality and quantity of data obtained for relatively slight costs. Line transects should replace "strip transects" as the operational survey method. This change would allow attainment of the original survey objectives (±10-l5% and the 95% level of confidence) even though every other transect line has been excluded from the survey. The change would have no effect on the distances surveyed, but would necessitate more careful measurement of nest distances from the centerline. To further increase the precision of density estimators, standard protocol should be to measure nest distances for nests up to 20 feet from the centerline of the transect. Species other than waterfowl should be enumerated during surveys, although the additional time required for this inclusion must be balanced against the gain in information. Transect surveys should n2! be suspended in dry or very wet years with low nesting effort, because such extreme conditions often provide the greatest insights into resource use and waterfowl biology. We recommend reinstatement of 2 biomonitoring efforts that were conceived in the past but subsequently abandoned. Low-level aerial photographs covered a �36 with mylar film were used on past surveys to pinpoint nest site location within each management unit. Although these overlays were apparently discarded prior to our analyses, the method has proven workable under field conditions and would provide insightful data on microhabitat selection of nest sites in relation to vegetative features and local landscapes. A second technique, first initiated in 1963, used elevated photo plots to record vegetation type and distribution from 80 photo stations throughout the MVNWR. The resultant color photos were also later discarded, but would have provided an excellent means to quantify long-term vegetation change in response to habitat treatments and natural succession. We recommend that the MVNWR retake these elevated photos as detailed in initial guidelines, then establish a regular vegetative monitoring program based on repeated elevated-plot photography taken at 5-year intervals. Effective biomonitoring requires that well-trained field personnel consistently follow standardized procedures for data collection and evaluation. Our experiences attempting to recover lost or discarded data underscore the need to develop systems to archive and safeguard data for longterm storage. The cost of obtaining biomonitoring data is relatively low «5% of the annual budget at MVNWR), yet the value of historic data is priceless for managers working in an adaptive management environment (Walters 1986). In the adaptive management approach, data are routinely collected on the effects of various management actions (e.g., grazing, predator control, burning), then interim evaluations are used to assist in deciding the course of future management. We highly recommend this iterative process, which allows decision making, monitoring and review, followed by further (refined or corrective) decisions and then additional monitoring on a continuing basis (Holling 1978, Macnab 1983, Bailey 1992). Lastly, any biomonitoring program is complicated by the application of multiple treatments to individual management units. When possible, managers should impose only a single treatment, followed by an evaluation of the efficacy of that treatment. If the treatment is deemed ineffective, a different, single treatment should then be imposed. Throughout the iterative evaluation process, a control (untreated) area should be maintained whenever possible. LITERATURE Ankney, C. D., and A. D. Afton. 1988. Shovelers: diet, nutrient reserves, Condor 90:459-472 . . and R. T. Alisauskas. 1991. -----limiting waterfowl production. CITED Bioenergetics of breeding Northern clutch size, and incubation. The role of nutrient Condor 93:1029-1032. reserves in Arnold, T. W., and F. C. Rohwer. 1991. Do egg formation costs limit clutch size in waterfowl? A skeptical view. Condor 93:1032-1038. Bailey, J. A. 1982. Implications of "muddling management. Wildl. Soc. Bull. 10:363-369. through" for wildlife 1968. Effect of predator Balser, D. S., Dill, H. H. and H. K. Nelson. J. Wildl. Manage. 32:669-682. reduction on waterfowl nesting success. Batt, B. D. J., and H. H. Prince. 1979. Laying dates, clutch size and �37 egg weight of captive mallards. Condor 81:35-41. Bellrose, F. C. 1980. Ducks, geese, and swans of North America. Stackpole Books, Harrisburg, PA. 540pp. Bennett, L. J. 1937. Grazing in relation to the nesting of the bluewing teal. Trans. N. Am. Wildl. Nat. Resour. Conf. 1:494-498. Blohm, R. J. 1979. The breeding ecology of gadwall in southern Manitoba. Ph.D. Thesis,. Univ. of Wisconsin, Madison. l52pp. Bue, I. G., L. Blankenship, and W. H. Marshall. 1952. The relationship of grazing practices to waterfowl breeding populations and production on stock ponds in western South Dakota. Trans N. Am. Wildl. Nat. Resour. Conf. 17:396-414. Byers, S. M. 1974. Predator-prey relationships on an Iowa waterfowl nesting area. Trans. N. Am. Wildl. Nat. Resour. Conf. 39:223-229. Capel, S. W. 1965. The relationships between grazing and predator activity in four types of waterfowl nesting cover. M. A. Thesis, Univ. Missouri, Columbia. 78pp. Chesness, R. A., M. M. Nelson, and W. H. Longley. 1968. the effect of predator removal on pheasant reproductive success. J. Wildl. Manage. 32:683-697. Choate, J. S. 1967. Factors influencing nesting success of eiders in Penobscot Bay, Maine. J. Wildl. Manage. 31:769-777. Cook, T. D., and D. T. Campbell. 1979. Mifflin Co., Boston, Mass. 405pp. Quasi-experimentation. Houghton Coulter, M. W. and W. R. Miller. 1968 Nesting biology of black ducks and mallards in northern New England. Vermont Fish and Game Bulletin No. 68-2. 74pp. Crabtree, R. L., and M. L. Wolfe. 1988. Effects of alternate prey on mammalian predation of upland waterfowl nests. Wildl. Soc. Bull. 16:163-169. _____ ., L. S. Broome, and M. 1. Wolfe. 1989. Effects of habitat characteristics on gadwall nest predation and nest-site selection. Wildl. Manage. 53:129-137. J. Cowardin, L. M., D. S. Gilmer, and C. W. Shaiffer. 1985. Mallard recruitment in the agricultural environment of North Dakota. Wildl. Monogr. 92. 37pp. Doty H. A. and A.J. Rondeau. 1987. Predator management to increase duck nest success. Paper presented at 8th Great Plains wildlife damage control workshop. Rapid City, S.D. April 27-30., 1987. �38 Draper, N. R. and H. Smith. 1981. Applied Regression Analysis. 2nd. ed. John Wiley & Sons. New York. 709pp. Drobney, R. D. 1991. Nutrient limitations of clutch size in waterfowl: is there a universal hypothesis? Duebbert, H. F. 1969. High nest density and hatching success of ducks on South Dakota CAP land. Trans. North Am. Wildl. Nat. Resour. Conf. 34:218-228. . and H. A. Kantrud. 1974. Upland duck nesting related to land use and predator reduction. J. Wildl. Manage. 38:257-265 . . and J. T. Lokemoen. 1976. Duck nesting in fields of undisturbed grass-legume cover. J. Wildl. Manage. 40:39-49. _________ , and _ 1980. High duck nesting success in a predator-reduced environment. J. Wildl. Manage. 44:428-437. _____ and D.E. Sharp. 1983. Concentrated nesting of mallards and gadwal1s on Miller Lake Island, North Dakota. J. Wi1dl. Manage. 47:729-740. Duncan, D. C. 1987a. Nest-site distribution and overland brood movements of northern pintails in Alberta. J. Wildl. Manage. 51:716-723. 1987b. Nesting of northern pintails in Alberta: laying date, clutch size, and renesting. Can. J. Zool. 65:234-246. Dwernychuk, L. W. and D. A. Boag. 1972. How vegetative cover protects duck nests from egg-eating birds. J. Wildl. Manage. 36:955-958. Dzubin, A. 1969. Comments on carrying capacity of small ponds for ducks and possible effects of density on mallard production. Pages 138-160 in Saskatoon wetlands seminar. Can. Wi1dl. Servo Rep. Ser. 6. _____ , and J. B. Gollop. 1972. Aspects of mallard breeding ecology in Canadian parkland and grassland. Pages 113-152 in Population ecology of migratory birds--a symposium. U.S. Fish and Wild1. Servo Wi1d1. Res. Rep. 2. Enright, C. A. 1971. An analysis of Mallard nesting habitat on the Monte Vista National Wildlife Refuge. M. S. Thesis, Colorado State Univ. , Ft. Collins. 113pp. Evans, R. D. and C. W. Wolfe. 1967. Waterfowl production in the Rainwater Basin area of Nebraska. J. Wildl. Manage. 31:788-794. Fuller, R. W. 1953. Studies in the life history and ecology of the American pintail in Utah. M.S. Thesis, Utah State Agric. ColI. l81pp. Gates, J. M. 1962. Bull. 74:43-67. Breeding biology of Gadwall in northern Utah. Wilson �39 Girard, G. L. 1941. The mallard: its management in western Montana. Wildl. Manage. 5:223-259. J. Gjersing, F. M. 1975. Waterfowl production in relation to rest-rotation grazing. J. Range Manage. 28:37-42. Gollop, J. B., and W. H. Marshall. 1954. A guide for aging duck broods in the field. Miss. Flyway Tech. Sect. Rep. l4pp. Greenwood, R. A. 1986. Influence of striped skunk removal on upland duck nest success in North Dakota. Wildl. Soc Bull. 14:6-11. Higgins K. F. 1977. Duck nesting in intensively farmed areas of North Dakota. J. Wildl. Manage. 41:232-242. _____ ., L. M. Kirsch, A. T. Klett, and H. W. Miller. 1992. Waterfowl production on the Woodworth Station in south-central North Dakota. U.S. Fish and Wildl. Servo Resour. Publ. 180. 79pp. Hines, J. E. and G. J. Mitchell. 1983. Breeding ecology of the gadwall at Watehen Marsh, Saskatchewan. Can J. Zool. 61:1532-1539. Holling, C. S. (ed.) 1978. Adaptive environmental assessment and management. Wiley Intern. Ser. on Applied Systems Analysis. Vol. 3. Holm, J. W. 1984. Nest success and cover relationships of upland-nesting ducks in northcentral Montana. M. S. Thesis, Univ. Montana, Missoula Hopper, R. M., A. D. Geis, J. R. Grieb, and L. Nelson, Jr. 1975. Experimental duck hunting seasons, San Luis Valley, Colorado. Wildl. Monogr. 46. 68pp. Hosmer, D. W., and S. Lemshow. 1989. Applied logistic regression. John Wiley & Sons, Inc., N.Y., USA. 328pp. Jeske, C. W. 1991. Condition, survival, and management of mallards wintering in the San Luis Valley, Colorado. Ph.D. Thesis, Colorado State University, Fort Collins. l50pp. Johnson, D. H. and J. W. Grier. 1988. Determinants of breeding distributions of ducks. Wildl Monogr. 100:1-37. Jones, R. E. and K. E. Hungerford. 1972. Evaluation of nesting cover as protection from magpie predation. J. Wildl. Manage. 36:727-732. Kantrud,. 1986. Effects of vegetation manipulation of breeding waterfowl in prairie wetlands-a literature review. U.S. Fish and Wildl. Serv., Fish Wildl. Tech. Rep. 3. l5pp. Keith, L. B. 1961. A study of waterfowl ecology on small impoundments in southeastern Alberta. Wildl. Monogr. 6. 88pp. �40 Kirby, R. E., S. J. Lewis, and T. N. Sexson. 1988. Fire in North American wetland ecosystems and fire-wildlife relationships: An annotated bibliography. u.S. Fish and Wildl. Serv., BioI. Rep. 88. 146pp . ., J. K. Ringelman, D. R. Anderson, and R. S. Sodja. 1992. Grazing ----o-nNational Wildlife Refuges: Do the needs outweigh the problems? Trans. N. Am. Wildl. and Nat. Resour. Conf. 57: (in press). Kirsch, L. M. 1969. Waterfowl production in relation to grazing. J. Wildl. Manage. 33:821-828. , H. F. Duebbert, and G. L. Heismeyer. 1978. Use of seeded native grasses as nesting cover by ducks. Trans. N. Am. Wildl. Nat Resour. Conf. 43:486-497. Klett, A. T., T. L. Shaffer, and D. H. Johnson. 1988. Duck nest success in the prairie pothole region. J. Wildl.Manage. 52:431-440. Krapu, G. L. 1974. Auk 91:278-290. 1981. 98:29-38. Feeding ecology of Pintail hens during reproduction. The role of nutrient reserves in mallard reproduction. , and H. A. Doty. Wildfowl 30:35-39. 1979. Auk Age-related aspects of Mallard reproduction. _____ ., G. L., A. T. Klett, and D. G. Jorde. 1983. The effect of variable spring water conditions on mallard reproduction. Auk 100:689-698. Krasowski, T. P., and T. D. Nudds. 1986. Microhabitat structure of nest sites and nesting success of diving ducks. J. Wildl. Manage. 50:203208. Labiskey, R. F. 1957. Relation of hay harvesting to duck nesting under a refuge-permittee system. J. Wildl. Manage. 21:194-200. Ladd, W. N., Jr. 1969. relationship of predation and land use practices to duck nesting activities on Valentine National Wildlife refuge, Nebraska. M. S. Thesis, Colorado State Univ., Fort Collins. 110pp. Likens, G. E. (ed.) 1989. Long-term studies in ecology: Approaches and alternatives. Springer-Verlag, London. 214pp. Livezey, B. C. 1981a. Locations and success of duck nest evaluation through discriminant analysis. Wildfowl 32:23-27. 1981b. Duck nesting in retired croplands at Horicon National Wildlife Refuge, Wisconsin. J. Wildl. Manage. 45:27-37. �41 Lokemoen, J. T., H. F. Duebbert, and D. E. Sharp. 1990. Homing and reproductive habits of mallards, gadwalls, and blue-winged teal. Wildl. Monogr. 106. 28pp. Marr, J. 1967. Ecosystems of the east slope of the Front Range of Colorado. Univ. of Colo. Studies, Ser. BioI. 8. Univ. Colo. Press, Boulder. 134pp. Martz, G. F. 1967. Effects of nesting cover removal on breeding puddle ducks. J. Wildl. Manage. 31:236-247. McKinney, F. 1965. Spacing and chasing in breeding ducks. Trust Ann. Rep. 16:92-106. Wildfowl Macnab, J. 1983. Wildlife management as scientific experimentation. Wildl. Soc. Bull. 11:397-401. Miller, A. W., and B. D. Collins. 1954. A nesting study of ducks and coots on Tule Lake and Lowere Klamath National Wildlife Refuges. Calif. Fish and Game 40:17-37. Miller, H. W. 1971. Relationships of duck nesting success to land use in North and South Dakota. Proc. Int. Congr. Game BioI. 10:133-141. Mundinger, J. G. 1976. Waterfowl response to rest-rotation grazing. Wi1dl. Manage. 40:60-68. J. Murdy, H. W. 1953. Some of the relationships between livestock grazing and duck nesting in the saltgrass vegetation type in Utah. M. S. Thesis, Utah State Agric. ColI., Logan. 58pp. Odin, C. R. 202. 1957. California gull predation on waterfowl. Auk 74:185- Oetting, R. B. and J. F. Cassel. 1971. Waterfowl nesting on interstate highway right-of-way in North Dakota. J. Wildl. Manage. 35:774-781 . C. C. Dixon. 1975. Waterfowl nest densities and success -----Oak. and Hammock Marsh, Manitoba. Wildl. Soc. Bull. 3:166-171. Picozzi, N. 1975. 39:151-155. Crow predation on marked nests. at J. Wildl. Manage. Pospaha1a, R. S. 1969. Waterfowl production and vegetative response to early water. M. S. Thesis, Colorado State Univ., Ft. Collins. 96pp. SAS Institute Inc. 1987. SAS/STAT user's guide release 6.0. SAS Inst., Inc., Cary, NC. 1,028pp. Ringe1man, J. K. 1992. Ecology of montane wetlands. Wi1d1. Serv., Fish and Wi1d1. Leaf. 13.3.6. 7pp. U.S. Fish and Risser, P. G. 1991. Long-term ecological research: An international perspective. John Wiley & Sons. N.Y. 299pp. �42 Rohwer, F. C. 1985. The adaptive significance of clutch size in prairie ducks. Auk 102:354-360. Ruyle, G. B., J. W. Menke, and D. L. Lancaster. 1980. Delayed grazing may improve upland waterfowl habitat. Calif. Agric. 34:29-31. Ryder, R. A. 1951. Waterfowl production in the San Luis Valley, Colorado. M.S. Thesis, Colorado A. and M. Col1., Ft. Collins. 166pp . Schranck, B. W. 1972. Waterfowl nest cover and some predation relationships. J. Wi1dl. Manage. 36:182-186. Schroeder, L. J. 1973. Effects of invertebrate utilization on waterfowl production. M.S. Thesis, Colorado State Univ., Ft. Collins. 44pp. _____ ., D. R. Anderson, R. S. Pospahala, G. W. Robinson, and F. A. Glover. 1976. Effects of early water application on waterfowl production. J. Wi1d1. Manage. 40:227-232. Simpson, G. G., A. Rowe, and R. C. Lewontin. 1960. Quantitative Zoology. Harcourt, Brace and World, Inc., New York, NY. 440pp. Sow1s, L. K. 1955. and management. Prairie ducks: a study of their behavior, ecology, Stackpole Books, Inc., Harrisburg, Pa. 193pp. Stoudt, J. H. 1971. Ecological factors affecting waterfowl production in the Saskatchewan parklands. U. S. Fish and Wildl. Servo Resour. Publ. 99. 58pp. Strang, C. A. 1980. Incidence of avian predators near people searching for waterfowl nests. J. Wildl. Manage. 44:220-222. Sugden, L. G. and G. W. Beyersbergen. 1987. Effect of nesting cover density on American crow predation of simulated duck nests. J. Wildl. Manage. 51:481-485. Swanson, G. A., and M. I. Meyer. 1973. The role of invertebrates in the feeding ecology of Anatinae during the breeding season. Pages 143-185 in The waterfowl habitat management symposium. U.S. Fish and Wildlife Service, Jamestown, ND. Szymczak, M. R. 1986. Characteristics of duck populations in the intermountain parks of Colorado. Colorado Div. Wi1dl., Div. Rep. 6. l3pp. Voorhees, L. D., and J. F. Cassel. 1980. Highway right-of-way: mowing versus succession as related to duck nesting. J. Wild1. Manage. 44:155-163. Walters, C. J. 1986. Adaptive management of renewable resources. MacMillan Pub1. Co., New York, N.Y. 374pp. Whyte, R. J., G. A. Baldassarre, and E. G. Bolen. 1986. Winter condition of mallards on the southern high plains of Texas. J. Wi1d1. Manage. 50:52-57. �43 Prepared by: Wildlife Researcher C Michael R. Sz Wildlife Researcher C ��45 Colorado Division of Wildlife Wildlife Research Report October, 1992 JOB PROGRESS REPORT State of Colorado Project W-166-R-l Work Plan __l__: Job Avian Research - Mi&ratory Game Birds 20 Job Title: Development and evaluation of moist-soil mana&ement techniques in Colorado Period Covered: Author: 1 April 1991 through 31 March 1992 James K. Rin&elman Personnel: R. Brown, D. Ferrin, J. Goettl, L. Swift, T. Ostertag, J. Ringelman, M. Szymczak, Colorado Division of Wildlife. ABSTRACT The Moist Soil Advisor computer package was obtain from D. B. Hamilton of the National Ecology Research Center, U.S. Fish and Wildlife Service. Preliminary evaluations of the software indicate that this expert system is applicable to moist-soil management in Colorado. Meanwhile, the search for suitable field sites to verify model expectations and plant responses continued during this segment. The site at Wellington Wildlife Management Area was deemed unsuitable unless costly earthmoving work was performed (additional detail on impoundments and attendant complications on this area are detailed in the October 1991 Wildlife Research Report). However, a wetland on private property just south-east of Fossil Creek Reservoir has been identified as a potential moist-soil unit. The landowner has indicated a willingness to cooperate with water manipulations and follow-up evaluations at this site. This wetland will ,be flooded in late April into May 1992, and a ..drawdown performed about 1 June 1992. . Prepared by: James K. Ringelman Wildlife Researcher C ��47 Colorado Division of Wildlife Wildlife Research Report October, 1992 JOB FINAL REPORT State of Colorado Project W-166-R-l Work Pl an ---,1:!:....-_ : Job Job Title: Migratory Game Bird Research 21 Analysis of Mallard Winter Banding Date Collected in Eastern and Western Colorado Period Covered: 1 April 1991 through 31 March 1992 Personnel: M. Wotawa, D. Anderson, K. Burnham, Colorado Cooperative Fish and Wildlife Research Unit; B. Wunder, Colorado State University; H. Funk, D. Hopper, M. Szymczak, J. Ringelman, Colorado Division of Wildlife. ABSTRACT All data from this project have been analyzed. The results will be prepared as a publication next segment by Mark Wotawa, Graduate Student, as the final report for this project. This work will be accomplished under Work Plan 22, Job 2, Migratory Bird Publications. Prepared by: ��49 Colorado Division of Wildlife Wildlife Research Report October, 1992 JOB PROGRESS REPORT State of _-"C=o"""l=or,_,a=d=o'--Project _1_: Job Title: Job 22 Harvest distribution of mallards and pintails banded preseason in western Colorado Period Covered: Author: Migratory Game Bird Research W-166-R-1 Work Plan _ 01 April 1991 through 31 March 1992 Michael R. Szymczak Personnel: J. Gamble and staff, Browns' Park National Wildlife Refuge; T. Fratt, U. S, Forest Service; D. Coven, P. Creeden, J. Ellenberger, V. Graham, J. Gray, J. Gumber, D. Masden, J. Miller, B. Motz, J. Olterman, R. alterman, J. Ringelman, N. Smith, P. Will, and M. Szymczak, Colorado Division of Wildlife ABSTRACT Ducks were trapped in modified Salt Plains bait traps and banded on 18 different wetlands located in 5 areas across western Colorado in August and September,1991. Nearly 1,100 mallards (Anas platyrhnchos) were banded, with the number captured generally well distributed between trapping areas. Mallard trapping efficiency was greatest in the Colorado River area as 8.7 mallards were banded per trap/day. All traps averaged 5.6 mallards banded per trap/day. Only 36 northern pintail (Anas acuta) were banded. wood ducks (Aix sponsa) trapped at the S. Walker site in the Colorado River area and ring-necked ducks (Aythya co77aris) trapped at Gardner Park Res. were also banded. ��51 HARVEST DISTRIBUTION OF MALLARDS AND PINTAILS BANDED PRESEASON IN WESTERN COLORADO P. N, OBJECTIVE 1. Document the distribution of band recoveries of mallards and pintails captured preseason in western Co lorado. . ' 2. Determine if the geographic location of recovery of mallards and pintails is dependent on the area of banding in Colorado. 3. Determine the relationship between the recovery distribution of western Colorado banded mallards and pintails and the distribution of recovery of those species banded in other areas of the Pacific Flyway. 4. Cooperate in analysis of Pacific Flyway - wide band recovery data and preparation of reports. SEGMENT OBJECTIVES . 1. Trap and band mallards and pintail in 4-5 areas of western Colorado in late August-early September using salt plains bait traps (Szymczak and Corey 1976). Recommended areas are: (l)Browns' Park National Wildlife Refuge, (2)Yampa River Valley below Craig, (3)White River Valley below Meeker, (4}Colorado River Valley below Glenwood Springs, (5) Uncompahgre-lower Gunnison River Valley, and (6)the DurangoCortez area. 2. Submit banding schedules and recapture reports to the U. S. Fish and Wildlife Services' Bird Banding Laboratory. Summarize and file band return reports. INTRODUCTION Recent drought in the northern Great Plains of the United States and Canada has reduced many waterfowl species to historically low population levels in surveyed habitats. Mallards and northern pinta i ls have been especially effected. Mallards reached near record low levels in Spring 1990, 27% below the 1955-89 average, while northern pintail populations reached a record low of 52% below the 1955-89 average (Reynolds et al. 1990). Despite low population levels and restrictive regulations, both species continue to be popular among hunters. In 1989, mallards composed 38% of the U.S. harvest and 40% in the Pacific Flyway harvest. Pintails, which made up nearly 25% of the Pacific Flyway harvest in the late 1960's and 1970's (U. S. Dept. Interior 1988:44) when all of the bag of 5 to 7 birds could be that species, still composed 9% of the harvest in 1989 (Martin et al. 1990) when only 1 of 4 birds could be a pintail. In western Colorado, pintail are not important in the hunters' bag (0.6%), but mallards made up 67% of the take in 1989 (Martin et al. �52 1990) It has been assumed that many of the mallards and pintails wintering in the Pacific Flyway originate in unsurveyed breeding areas, west of the continental divide. Water regimes, and thus waterfowl production, in these areas are usually more stable and not tied to the wet-dry cycles characteristic of the midcontinent Great Plains. Childress (1986) examined relationships among duck breeding populations and fall flight estimates in surveyed areas, and duck harvests in the Pacific Flyway during the stabi 1ized regulat ion period (1975-83) He found that recruitment in standard survey areas differed from that in unsurveyed areas. Pospahala et al.(1974: compiled from Table 8- 20) estimated that over 600,000 mallards occupied western breeding habitats (excluding Alaska) outside the standardized surveyed breeding areas. Deficiencies in pre-hunting-season (preseason) mallard banding in western breeding areas was first mentioned by Anderson and Henny (1972:86). Munro and Kimball (1982:19) noted that continental mallard banding data suggested that a larger portion of the mallard harvest occurred in the Central Flyway than in the Pacific Flyway; however harvest surveys indicated the opposite. Trost (1985:4-5) reported that Pacific Flyway mallard harvests, derived from analysis of banding data, were underestimated because the appropriate breeding populations were not properly represented by banded samples. Smith and Powell (1989:4) concluded that a complete picture of the Pacific Flyway mallard harvest cannot be completed until additional samples of mallards are banded in the states and provinces of the western U. S. and Canada. Colorado has breeding mallards scattered throughout the Pacific Flyway portion of the state. Annual surveys along the Yampa and Green rivers indicate nearly 1,500 mallards breed along those riparian zones (CDOW unpubl.). Frary (1954) and Rutherford and Hayes (1976) found respective rnallard breeding densities of 0.5 and 0.8 pairs/km2 on the White River Plateau and among a variety of hab ttats in the mountainous areas west of the San Luis Valley. More recently, Ringelman et al. (1990) found mallards nesting at densities of 5.7 pairs\kffi2in montane habitats containing beaver flowages and kettle lakes just west of North Park. Although the latter 2 studies were conducted immediately east of the continental divide (Central Flyway), there is no reason to believe that habitats west of the divide (Pacific Flyway) don't support similar mallard nesting densities. With pintail populations at record low levels, obtaining information on derivation of harvest and harvest rates would be of value in adjusting harvest regulations and measuring population response.. A special preseason banding project for pintails, including banding in western breeding reference areas was proposed in the continental duck banding program (U. S. Fish and Wildl. Servo 1989). Ident ifying and understanding distribut iona 1 characterist ics of Co lorado's major waterfowl populations are important in des igning strategies to reach population objectives (Colo Div. Wildl. 1989). Preseason mallard populations in western Colorado have never been banded, and therefore population and harvest distribution characteristics are unknown. With continued drought in the Great Plains and dependence upon prairie duck breeding recruitment to develop fall flight duck forecasts, there is a definite . need to identify those breeding populations that winter in the Pacific Flyway and determine the specific areas of harvest. A companion step, not part of this study, is to develop standardized surveys that will result in annual estimates of population size and recruitment in those breeding populations. Then annual assessments of ducks will include those specifically identified as Pacific Flyway birds and regulatory adjustments can be based on the status of those populations. �53 The Pacific Flyway Study Committee has formulated a 5-year cooperative mallard and pintail preseason banding program that has been endorsed by the Pacific Flyway Council. This program is designed to address banding needs throughout the western U. S., including Alaska, and in the provinces of British Columbia and Alberta (northwestern). Because Colorado supports Pacific Flyway breeding mallards and harvests Pacific Flyway mallards (x- = 18,567, 1961-88), the Colorado Division of Wildlife has committed to cooperate in this program. Few pinta ils breed or are harvested in western Co lorado. Therefore, banding in Colorado will be directed toward mallards. METHODS Trap Area Selection Most breeding mallards in western Colorado are associated with small wet lands that are widely distributed throughout high e levat ion, mountainous areas. Only in Browns' Park, located in the extreme NW corner of Colorado, are there extens ive wet land comp lexes capab le of support ing breeding ducks. Trapping on widely distributed wetlands with low densities of breeding ducks was assumed to be an inefficient method for banding western Colorado mallards. Therefore, strategica lly located areas, to which post-breeding and fledged Co lorado ma llards would move, were selected as primary trapping areas. In addition to the Browns' Park Nat iona 1 Wildl ife Refuge (BPNWR), trapping occurred a long the Co lorado River Valley from Debeque to near Fruita, in the Uncomphagre River Valley from Montrose to near Delta, in the Cortez-Mancos area, and at Gardner Park Reservoir, about 5 miles W. of Yampa. Trapping Period Because the degree of success expected was unknown, ducks were trapped and banded in each area for a consecutive period of about 10 days beginning near the end of August and ending near 20 September. The actual trapping periods were: BPNWR - 17 thru 27 September; Gardner Park - 19 thru 25 September; Colorado R. Valley - 5 thru 16 September; Uncomphagre R. Valley - 29 August thru 7 September; Cortez- Mancos 7 thru 17 September. Trapping and Recording Technique All birds were trapped in modified Salt Plains bait traps (Szymczak and Corey 1976) using whole shelled corn for bait. A workshop was conducted for all CDOW personnel involved in trapping to review use of the trap, age and sex characteristics of mallards and pintails, and recording and maintaining records. Traps were visited once a day and only mallards, pintails and wood ducks (Aix sponsa) were banded. Banded birds were recorded by wetland site. Band numbers of all birds captured that were banded in previous years or outside the specific area of trapping were recorded. Records were also maintained on the number of traps operated by wetland in order to evaluate capture/unit of effort at each trap site. �54 RESULTS Trap locations Trapping was distributed over a total of 18 different wetlands in the 5 areas (Table 1). Pre-baiting was well distributed geographically throughout each area, but ultimately trapping site locations were dependent on bird distribution. Nearly all trap sites were located in Palustrine Emergent Persistent Wetlands, but some sites in the Colorado River Valley were in Riverine Upper Perenial Rock Bottom wetlands (Cowardin et ale 1979). Banding and trapping efficiency Nearly 1,100 mallards were banded during trapping in western Colorado in 1991 (Table 2). Mallards banded were well distributed, but more birds were banded in the southwest one-quarter of the state than in the northwest onequarter. Nearly 70% of the banded sample was composed of immature birds. Only 'in Browns' Park were equal numbers of adults and young banded. Nearly 6 mallards were banded per trap day (Table 2). Traps were most efficient at 3 locations in the Colorado River area, although Markley's Pond in the Uncompahgre River Valley produced more banded ducks than any other trap site. Only 36 northern pintail, the secondary target species, were banded (Table 3). Wood ducks exclusively captured at 1 trap site in the Colorado River area and ring-necked ducks captured at Gardner Park Reservoir were also banded, because of local inter~st. Band Reporting and Record Keeping All new banded birds and recaptures were submitted to the U. S. fish and Wildlife Service's Bird Banding laboratory on standard forms. Computor files . containing the number of birds banded by area, site, day, age and sex were constructed at the Colorado Division of Wildlife's Research Center. �55 Table 1. Trapping locations during preseason banding in western Colorado, August- September, 1991. Wetland Area Browns' Park Natl Wildl. Ref. Name Location Butch Cassidy TI0N, RI04W, Sec 12, SW\ Flynn Marsh TI0N, RI03W, Sec 16, SE\ Hog Lake TI0N, RI03W, Sec 9, SW~ Spitzie Slough TI0N, RI03W, Sec 15, S~ Gardner Park Res. Gardner Park Res. TIN, R86W~ Sec 22, NE\ Colorado R. Valley Latham's Slough T8S, R97W, Sec 27, SW\ Morse's Pond TIS, RIE, Sec 34, NE% Ute Meridian Walker Wildl Area South TI1S, RI01W, Sec 14, NW\ Walker Wildl Area North Uncompahgre R. Valley CortezjMancos TIN, R2W, Sec 36, SW\ Ute Meridian Skippers' Island TIN, R3W, Sec 14, NE% Ute Meridian Porter's Feedlot T51N, RI0W, Sec 27, SW% Grett's Dairy T50N, RI0W, Sec 2, NW% Markley's Pond T50N, R9W, Sec 30, NW% Kintz's Pond T48N, R9W, Sec 3, SE% Weir's Pond T36N, RI6W, Sec 13, SE\ Totten Res. T36N, RI5W, Sec 20, NW\ Mancos Wetland T36N, R13W, Sec 27, SW\ Weber Res. T36N, R13W, Sec 12, NE\ �56 Table 2. Number of Mallards banded by area, site, age and sex and trapping efficiency during pre-season trapping in western Colorado, 1991. Number of locals included in parentheses. Area Brown's Park Site AM B. Cassidy Flynn Marsh Hog lake Spitzie Slough Sub-total 0 13 2 Gardner Park Colo. River Uncomp. River CortezMancos S. Walker N. Walker Skippers Is. Morse's Pd. latham Slough SUb-total Markley's Grett's Porter's Kintz Pd. SUb-total Weir's Totten Weber Res. Mancos Wetland Sub-total GRAND TOTALS AgeLsex AF 1M IF Total No. trap days No. banded per trapLday 26 1 19 4 II 37 0 9 10 J. 26 1 17 6 II 36 2 58 22 43 125 '1 9 18 22 50 2.0 6.4 1.2 2.0 2.5 6 4 17. 12 39 16 2.4 2 30 5 5 1 12 1 3 7 50 37(2) 26 33 _5. 50 22 153(2) 13 119 53 68 87 340 8 8 5 6 II 39 1.6 14.9 10.6 11.3 7.3 8.7 54 0 15 21 0 5 75 3 13 53 4 14 203 7 47 20 6 9 _Q _Q _5. _l4 _§. 69 26 100 76 271 41 10.2 1.2 5.2 2.3 6.6 30 29 2 _! 65 13 20 1 20 34 45 _.1 37 107 ~ 28(3) 91(3) ·32(2) 115(2) 32 80 _! .is 96(5) 305(5) 11 14 14 7 46 8.3 8.2 5.7 2.7 6.6 126 403 335(7) 1080(7) 192 5.6 11 .a 216 j 3 40(2) 21 27 24 115(2) �57 Table 3. Number of northern pintail, wood ducks and, ring-necked ducks banded by area. site. age. and sex. during preseason trapping jn western Colorado. 1991. Species Northern pintail Area Colo. R. Site AM AgeLSex AF 1M IF 2 Total N. Walker Latham Slough 0 1 0 0 1 1 1 1 Uncomp. R. Markley's Pd. 2 0 2 1 5 18 2 6 0 26 21 2 10 3 36 8 9 4 3 24 0 0 4 1 5 CortezMancos Totten Res. SUb-total Wood duck Colo. R. Ring-necked duck Gardner Pk. S. Walker 3 LITERATURE cnED Anderson, D. R., and C. J. Henny. 1972. Population ecology of the mallard. I I. A review of previous studies and the distribution and migration from breeding areas. U.S. Fish and Wildl. Servo Resour. Publ. 105. 166pp. Carney, S. M. 1964. Preliminary keys to waterfowl age and sex identification by means of wing plumage. U. S. Dep. Inter., Fish and Wildl. Servo Spec. Sci. Rep. - Wi1dl. 82. 47pp. Childress, D. (Subcomrn. Chmn). 1986. Evaluation of stabilized regulations in the Pacific Flyway 1975-1983. Pacific Flyway Study Comm. Unpub1. rep. 40pp. Colorado Division of Wildlife. 1989. Colorado statewide waterfowl management plan 1989 - 2003. Colo. Div. Wildl., Terrestrial Wild. Sect., Migratory Game Bird Program Unit. 97pp. Frary, l. G. 1954. Waterfowl production on the White River Plateau, Colorado. M.S. Thesis, Colorado State University. 93pp. Martin, E. M., P. H. Geissler, and A. N. Novara. 1990. Preliminary estimates of waterfowl harvest and hunter activity in the United States during the 1989 hunting season. U. S. Fish Wildl. Serv., Admin. Rep. -- July. 34pp. �58 Munro, R. E.~ and C. F. Kimball. 1982. Population ecology of the mallard. VII. Distribution and derivation of the harvest. U. ·S. Fish and Wi1d1. Servo Resour. Publ. 147. 127pp. Pospahala, R. S., D. R. Anderson, and C. J. Henny. 1974. Population ecology of the mallard. II. Breeding habitat conditions, size of the breeding populations, and production indices. U. S. Fish and Wildl. Servo Resour. Publ. 115. 73pp. Ringe1man, J. K., M. A. Wotawa, and R. S. Langley. 1990. Waterfowl abundance and production on the Routt National Forest, Colorado, 1990. Unpub1. rep. Colo. Div. Wild1., Fort Collins. Reynolds, R. E., R. J. Blohm, F. A. Johnson, and J. B. Bortner. 1990. 1990 status of waterfowl and fall flight forecast. U. S. Fish and Wi1dl. Servo July. 43pp. Rutherford, W. H. and C. R. Hayes. 1976. Stratification as a means for improving waterfowl surveys. Wildl. Soc. Bull. 4:74-78. Smith, R. I., and B. H. Powell. 1989. Distribution of band recoveries from huntershot adult female mallards in western North America. U. S. Fish and Wild1. Serv., Laurel, MD. Unpub1. rep. 20pp. Szymczak, M.R., and J. F. Corey. 1976. Construction and use of the Salt Plains duck trap in Colorado. Colo. Div. Wi1d1., Div. Rep. 6. 13pp. Trost, R. E. 1985. A preliminary assessment of the recent distribution and derivation of the mallard harvest in the United States based on recoveries from breeding ground bandings, 1975-1984. U. S. Fish and Wild1. Serv., Laurel, MD. Unpubl. rep. 66pp. U. S. Dept. Interior. 1988. Issuance of annual regulations permitting the sport hunting of migratory birds. U. S. Fish and Wildl. Servo final supplemental· environmental impact statement. Wash. D. C. 340pp •. Unpub1. Rep. 40pp. U. S. Fish and Wild. Servo 1989. The North American duck banding program - a revised approach. Canadian Wildl. Servo and U. S. Fish and Wildl. Serv., laurel MD. Unpubl. rep. 23pp. Weller, M. W. 1976. Molts and plumages of waterfowl. Pages 34-38 in F. C. Bellrose. Ducks, geese and swans of North America. Stackpole Books, Harrisburg, Pa. 543pp. Prepared by: -----m .."1..4 --C ~-cr'£ Michael R. Szymczak Wildlife Reasercher C �59 Colorado Division of Wildlife Wildlife Research Report October, 1992 JOB PROGRESS REPORT State of _--"C"-,,o:...:.l~orwa~d~o:....__ _ W-152-R-l Project 10 Work Plan Job Title: _1_ Cooperative Management Programs Period Covered: Author: : Job Avian Research - Migratory Game Birds 01 April 1991 through 31 March 1992 Michael R. Szymczak Personnel: James K. Ringelman and Michael R. Szymczak, Wildl ife Colorado Division of ABSTRACT Methods of implementation of the San Luis Valley Waterbird Plan were discussed with Colorado Division of Wildlife {CDOW} regional and U. S. Fish and Wildlife personnal. Recommendations for wetland habitat improvements on present holdings or proposed acquisitions were provided for Russell Lakes and South Republican SWA's, Walden Reservoir and Hebron Sloughs in North Park, Cooper Slough in the City of Ft. Collins, and to Soil Conservation Service and CDOW personnal in western Colorado. Presentations on various aspects of waterfowl ecology were given at training and educational schools, workshops, and short courses. Responsibilities as Colorado's representative on Pacific Flyway Study Committees and Council were fulfilled; Colorado chaired the Pacific Flyway committees through a portion of this reporting: .. segment. Participation in Central 'Flyway matters include appointment and participation as a Central Flyway representative to the Stabilized Duck Regulations Committee. Waterfowl counts designed to help identify population limiting factors were conducted on ducks and/or geese in portions of North Park. Methods for surveying breeding ducks and assistance with surveys were provided to U. S. Forest Service biologists in the Routt, White River and Roosevelt National Forests. Technical assistance was provided through appointment to the Continental Evaluation Team for the North American Waterfowl Management Plan, toxicity testing of a new alternative to lead shot, and other waterfowl and wetland issues. ��61 Cooperative Migratory Bird Management Programs Michael R. Szymczak James K. Ringelman In 1988, the Colorado Division of Wildlife (CDOW) created the Migratory Game Bird Program Unit (MBPU) within the Terrestrial Wildlife Section. This administrative change combined all individuals having statewide responsibilities for research and management of migratory game birds~ Members of the MBPU work in concert to improve migratory bird management in Colorado. This job was created to allow team members to participate in these management programs. P. N. OBJECTIVES 1. Participate in developing and implementing habitat-based waterfowl management plans on a statewide, habitat region, and project basis. 2. Advise state and federal land managers on beneficial habitat acquisitions and/or developments and provide expertise in preparation of development and/or management plans. Advise private land managers in developing habitat management plans and assessing impacts on waterbird populations. 3. Present information on the principles of waterfowl management to workshop attendees, educational classes, and conservation organizations. 4. Participate in migratory bird management meetings at the state and flyway levels. 5. Cooperate in developing surveys and techniques that will assess the impact of migratory bird management programs. SEGMENT OBJECTIVES 1. In conjunction with the statewide waterfowl management plan, begin work on waterfowl habitat region plans. 2. Provide biological expertise for wetland development programs on the Brush, Russell Lakes, South Republican State Wildlife Areas, Hebron Ponds and Walden Reservoir (BLM-CDOW), various wetland complexes managed by the U. S. Forest Service, and other areas where requested. 3. Prepare and present informational programs on migratory bird management when requested. 4. Compile appropriate population status information and represent Colorado at Pacific Flyway Technical Committee and Council meetings. Attend migratory bird management program meetings in Colorado when requested. 5. Provide methodology for wetland habitat and migratory bird population surveys when requested. 6. provide technical information an offer court testimony when necessary. �62 RESULTS Waterfowl Management Plans Discussions were undertaken with CDOW, U. S. Fish and Wildlife Service and Bureau of Land Management personnel concerning meeting the San Luis Valley Waterbird Plan objectives for breeding and wintering ducks. Emphasis was on evaluation of current methods for measuring breeding ducks and development of new methods measuring breeding ducks in various habitats. The results of habitat based winter duck dispersion efforts and future plans were also reviewed. An updated version 0 the recently completed San Luis Valley Waterbird Plan was forwarded to CDOW personnel responsible for developing a Waterfowl Management Plan for North Park. Wetland Developments and Acquisitions Potential sites for wetland developments and existing wetlands were visited on the Russell Lakes SWA in the San Luis Valley, the South Republican SWA at Bonny Reservoir and Walden Reservoir and Hebron Sloughs in North Park. Recommendations for wetland construction, water management, and management of existing wetlands were given to CDOW management personnel. The City of Ft. Collins was advised on the value and potential for additional wetland development of Cooper Slough. CDOW Southwest and Northwest Regional personnel were advised of available wetland acquisition funds, and some potential projects were identified in their respective areas. We consulted with U. S. Soil Conservation Service biologists on their mitigation program for wetlands lost because of de-salinization projects in the Uncomphagre River Valley and toured some wetlands recently developed under that program. Project personnel, as members of the multi-agency Waterfowl Habitat Project Review Committee, reviewed and rated wetland enhancement and acquisition proposals from land management agencies for funding with Colorado State Duck Stamp monies. Waterfowl Hunter Attitude Survey This survey will be conducted during 1992-93 through the Human Dimensions Research Unit at Colorado St~te University. Informational Programs Informational Programs Formal presentations on waterfowl recruitment and managing wetland habitat were presented to U.s. Forest Service, Rocky Mountain Region Waterfowl and Wetlands Training Workshop and the BLM Wetlands Training Workshop. Presentations on waterfowl ecology and management were made to CDOW District Wildlife Management Trainees, the Wildlife Management Shortcourse and the Avian Management course and Non-domestic Animal Management and Disease course at Colorado State University. Presentations on duck identification were given at the CDOW District Wildlife Management Trainees. Waterfowl Technical Committee and Council Meetings Colorado was represented at. the July 1991 Pacific Flyway Study Committee �63 meetings by project personnel (Szymczak). Waterfowl population status was reviewed and hunting season recommendations forwarded to the U. S. Fish and Wildlife Service's Regulation Committee. Populations of specific interest to Colorado whose status was reviewed in July were (1) breeding and wintering mallards inhabiting the Pacific Flyway portion of Colorado and (2) the Rocky Mountain Canada goose population. In October 1991, the Chairmanship of the Pacific Flyway Council's Waterfowl and Western Migratory Upland Bird Study Committees rotated to Colorado. Colorado chaired a special meeting of the Waterfowl Study'Committee in January to discuss specific Pacific Flyway issues and formulate recommendations for approval at the regular March 1992 meeting. In March, in addition to chairing the Pacific Flyway Committee, project personnel (Ringelman) attended the Central FlYway Technical Committee meeting. At both flyway meetings general information on migratory game bird populations was exchanged by committee members. Regulatory recommendations were made to both Flyway Councils. Ringelman was appointed as one of two Central Flyway representatives to the Stabilized Duck Regulations Committee, which consists of 2 technical representatives from each flyway and personnel from the U. S. Fish and Wildlife Service's Migratory Bird Management Office. The chairmanship of the Pacific and Western Committees required the Colorado representative to attend the Pacific Flyway Council meeting at the North American Wildlife and Natural Resources Conference in late March. Population Survey Methodology Surveys were conducted or methodology breeding duck surveys were developed with or for U. S. Forest Service biologists for the Kawaneeche Bench (Routt N. F), Shorty/Cataract Creek (Routt N. F.), Ammons and Thomas Creek drainages (Routt N. F.), Swamp Creek (Roosevelt N. F), and Aldrich Lakes (White River N. F.). Surveys were continued on 5 study units in the Big Creek Lakes region (Routt N. F.). The results of all surveys on Forest Service lands were used to refine and further develop a computer model for estimating duck breeding pair use of forested montane habitats (see Appendix A). ' Surveys of nesting and brood rearing Canada geese on Walden and MacFarlane reservoirs in North Park during spring 1991 continued to documented poor productivity at both sites. Goose nests visited, marked, and mapped in early May just prior to hatch, and revisited after all had hatched showed good hatchability. Periodic brood counts made post-hatch indicated gosling mortality probably occurred shortly after hatch. A detailed report was submitted to Northeast Regional and Bureau of Land Management personnel. Technical Assistance Ringelman was appointed to and began participation on the Continental Evaluation Team for the North American Waterfowl Management Plan. Project personnel provided input to The Nature Conservancy on designation of the Big Creek Lakes Research Natural Area and the Fossil Creek Reservoir goose hunting closure. The toxic effects of spent lead shot in the environment caused the USFW to mandate the exclusive use of steel shot for waterfowl hunting. However, some hunters are still reluctant to accept steel shot as a viable alternative, large because of perceived ballistic shortcomings. Consequently, when research personnel were approached by a manufacturer of a proposed substitute nontoxic shot with better potential ballistic properties than steel, we agreed �64·· to undertake toxicity testing of the new shot material. Results of the testing are reported in the form of a draft manuscript (Appendix B). DISCUSSION Project personnel proved to be a useful resource in planning and evaluating waterfowl management and habitat enhancement programs in Colorado and educating land management agency personnel about the habitat requirements of waterfowl. We experienced increased involvement in long-range planning for waterfowl on specific areas. We anticipate that with increased emphasis on habitat enhancement in Colorado as outlined in the statewide Waterfowl Management Plan that our services will be more in demand. Continued participation on Flyway committees ensures that Colorado will remain informed on migratory bird matters and have input in migratory bird hunting regulations. Prepared by: 2UJ~. Michael R. Szymczak Wildlife Researcher C �APPENDIX Title: 65 A Verification of a Model to Estimate Vaterfowl Pair-use on Montane Vetlands Investigator: Richard S. Langley Department of Wildlife Biology Colorado State University Fort Collins, Colorado 80521 INTRODUCTION The Colorado Division of Wildlife conducted waterfowl pair censuses and brood counts on 4 adjacent study areas in the Routt National Forest, Colorado, during 1989. The study was designed to determine waterfowl production in montane wetland communities consisting of kettle ponds and beaver ponds. Each study area was censused weekly for 6 weeks beginning the third week in May to document pair use on wetlands. Brood counts were done for the following 6 weeks beginning the second week in July. All wetlands were mapped and measured for basin size and surface water area. Other variables recorded were percent aquatic vegetation, vegetation type, wetland type, grazing impacts, human disturbance, water permanency, and pond-bottom type. Wetland variables were tested for correlation with pair and brood use. Surface water area and wetland type (beaver or kettle pond) correlated highly with pair use. Brood use was mostly related to pair use. J. K. Ringelman (Unpublished data, Colorado Division of Wildlife) developed a computer model to estimate waterfowl breeding pair use on montane wetlands using the calculated regression equations for surface water area and wetland type. The model also estimates the composition of waterfowl species based on the frequency of occurrence of different species on wetlands of different sizes. Models have been found to be applicable in predicting population size, determining causes for low populations, predicting population changes with habitat alterations, identifying areas for mitigation of habitat loss, and determining functional relationships between populations and the habitat. �66 Desirable characteristics of a model include realism - it includes the functional attributes of the population, precision - it accounts for a large percentage of variation, and generality - it is based on many years of data from a broad geographic area. Potential problems with models include low sample size, errors in measuring independent variables, biological assumptions, minimum limiting factors, compensation factors, lack of verification, and lack of testing. The validity of the montane, breeding pair model has been tested on several areas close to the original study site. The pair-use estimates produced by the model were very close to what was actually found when the areas were censused. Since then, the U. S. Forest Service has censused several more montane wetland areas for waterfowl use. to six new areas to test its predictive ca~bilities Here I apply the model when applied to areas widely separated from the original study site. The objectives of this study were to help validate a model that is potentially useful to land-use managers and state and federal agencies in planning, research, management, or development. This study also helps to validate the model by testing it for accuracy, since waterfowl pair numbers have already been verified on the study areas. METHODS During the spring and summer of 1991, Fo:;-estService biologists collected waterfowl pair and brood-use data on five montane forest areas in Colorado. These study areas include Swamp Creek in Roosevelt National Forest, Shorty/Cataract Creeks and the Kawaneeche Bench in Routt National Forest, Aldrich Lakes in White River National Forest, and ThomaS/Ammons Creeks in Arapaho National Forest. In addition, J. K. Ringelman and M. R. Szymczak of �67 the Colorado National Division of Wildlife and recensused Waterfowl indicated planimeter pair numbers were estimated 1991). Wetland Wetland with a rangefinder, by counting on each study area during the census size (in square meters) and rough maps were drawn. was measured and provided Wetland information collected during the censuses. using a that have a on-site size for the Aldrich by the Forest Service. (beaver or kettle pond) was determined The type of from aerial photos and These are the same methods used in tests of the model. Wetland size and type were entered run for each area. Pair estimates and predicted into input files, then the model was were summarized of pairs using a paired T-test observed the number of size for the Swamp Creek complex was measured Lakes complex was measured previous Pond area in Routt on maps of the surface water made from aerial photos known scale. number the South Kettle four other study areas. pairs, pairs and lone males, (see Ringelman wetland censused and compared to the actual to test the null hypothesis values did not differ (P that > 0.05) from zero. RESULTS The waterfowl areas 2). censuses found a variety (Table 1) and varying numbers Some variation second censuses in waterfowl of the original of waterfowl pair use between study areas and Schafer Creek) and the North Kettle areas, Aldrich Lakes had the highest predicted pair-use. predicted pair-use The Kawaneeche (Table 2). of waterfowl study areas (Table years was found during the (N~ Fork N. Platte, Ponds observed between using the new study (Table 2). pair-use Goose Creek, Of the new study and- the highest Bench had the lowest observed The model produced reasonable three of the six new study areas and, again, performed and estimates on well on the other study �68 areas based on previous predicted and observed all study areas years censuses (Table 2). The difference pair-use were not significantly (T - 1.45, P - 0.1733) different and the null hypothesis between the from 0 for was not rej ected . .Species on the species perform composition predictions composition were not analyzed in the original since they were based study areas and do not appear to well on other areas (Table 1). DISCUSSION For all of the study areas combined, estimates that reflect actual waterfowl the pair-use pair-use. model produced Three of the six estimates for the 1991 study areas were close to the observed numbers. model was accurate of waterfowl study areas. Kettle in predicting relative The second censuses abundance of the original for some of the variation There may be several Shorty/Cataract pairs on Shorty/Cataract estimate Lakes, and Swamp Creek. in the area. and drakes had begun to move to molting counts. This may also be the case with Aldrich intensively early beginning (May 15), the wetland attributed The overestimation later (June areas, resulting Lakes, because in low even though complex was cerrsus ed most was underestimated to field measurement of to the lack of an accurate June 13. Swamp Creek pair-use on that most hens were already nesting began use and this Pair counts were started 6) than for other areas, which may indicate .censuses in waterfowl causes to the poor predictions Creeks may be attributed of actual pair-use on the new in the model estimates. contributing Creeks, Aldrich the study areas and the North Ponds show that there may be yearly variation may account Overall, by the model. This may be errors since an aerial photo was not �69 available, thereby affecting planimeter measurements. subsequent errors in map development There may be factors that are more limiting wetland that affect waterfowl Bench complex aquatic storing nutrients foods. that low invertebrate Peterson to climatic oxygen, In montane or edaphic of Utah, that prevent of wetlands forests, use (Peterson Average and aquatic plants such as temperature, reservoirs during nest initiation at higher pH, Lakes complex, that undergo or on small ponds by pond aspect, pond depth and size, in mean temperature Waterfowl to the and s~ow depth. use due to lack of hiding were found to occupy wetlands> than those < 0.4 ha in the Uinta Mountains, Utah wetlands and wetland complexes 0.4 ha (Peterson Low 1977). Individual large can also affect wetland elevations pond size may also affect waterfowl more frequently a the growth of aquatic plants. especially cover and food availability. to plants, which provide In the case of the Aldrich Ice-out may also be influenced and yearly variations and Low use was related with dense tree cover, ice-out may occur too late to attract birds wetland. for rapid it was also found of invertebrates are large, deep containment of to hens and also to broods aquatic factors to the paucity are essential in low waterfowl or absence and nutrients. fluctuations Availability use. and incubation, of seed-producing The presence the largest wetlands yearly use may be related invertebrates resulted of the Kawaneeche and Low (1977) also found that waterfowl source of food. dissolved numbers than size and type of investigations In the Uinta Mountains or abundance may be related Aquatic for egg-laying growth and development. the presence Initial suggest that low waterfowl invertebrate 1977). use. and may also have a lack of and �70 nesting cover, either from absence of low growing vegetation, cover, or degradation and Shorty/Cataract waterfowl by cattle grazing. Creeks complexes production Observations indicate presence in the Aldrich of snow Lakes that grazing may be reducing potential. RECOMMENDATIONS The results waterfowl using indicate that the model can produce use in montane wetlands. It may be beneficial the recent data to recalculate size and.type. to a greater variety on specific states with similar Mountains the correlations to modify for pair-use This would increase sample size and generality, model more applicable accuracy good estimates areas. Satellite wetland independent variables. of areas but probably imagery or GIS may also be useful Lack of verification encountered in the development becaus~they reduce The species difficult species reducing data may for size in an attempt to reduce error in the measuring of the model. of models. composition are very selective estimator accurate Systematic of habitat may occur in only a few areas. to the surveys' are essential, estimations. does h~ve potential, estimates is essential of is one of the problems the chance of errors in pair-use to get to produce the One such area would be the Uinta The need for accurate pair counts on the study areas validation to wetland making study area from Peterson and Low's study, where wetland calculating the model The model should also be tested on areas in other characteristics. already be available. of but would be over a wide area since some type (e.g. bufflehead, goldeneye) Much of the error in the estimates attributable to predicting the number of pairs for species in an area. Species which are not even predicted and was that do not occur by the model, but found on �/1 the areas, caused errors in the other direction. A first step would be to incorporate these "new" species into a new generation model. Secondly, when a user enters their wetland data, they would also include a list of AOU numbers of species known or thought to breed in the region. Before a final e. ~termination, the model would zero out any estimated pairs for species not thought to occur in an area, then only report (and sum) the number of pairs for those that do. That wouljd eliminate all cases where observed was zero but predicted was> O. Another problem in the format of the model occurs when a wetland complex is very small or the wetlands themselves are very small: the model produces a negative total number of pairs. Kawaneeche Bench complex. This was actually the case with the A change in the program could be made so that it produces 0 total pairs when the calculations come up negative. LITERATURE CITED Peterson, S. R., and J. B. Low. wetlands in Utah. Ringelman, J. K. 1991. J. vriai. 1977. Waterfowl use of Uinta Mountain Manage. 41:112-117. Evaluating and managing waterfowl habitat. Division of Wildlife rep. no. 16. Fort Collins, co. 46pp. Colorado �Table 1. 1991 species composition of waterfowl on the six study areas observed (0) and predicted (P) by the model. Aldrich L. Kawaneeche Shorty/Cat. S. Kettle P. Swamp Cr. Thomas/Ammons Species (0) (P) (0) (P) (0) (P) (0) (P) (0) (P) (0) (P) Mallard 17 18 2 0 15 10 7 5 7 2 6 4 Green-wing teal 0 4 0 0 0 5 1 2 7 0 1 1 Cinnamon teal 0 6 0 0 0 3 0 2 0 0 0 1 Wigeon 0 2 0 0 0 1 0 1 0 0 0 0 Wood duck 0 0 0 0 0 0 0 0 0 0 0 0 Ring-necked duck 1 20 0 0 2 9 1 5 1 1 0 1 Bufflehead 0 8 0 0 0 4 2 2 0 0 0 0 Cornmonmerganser 4 2 0 0 0 1 0 0 1 0 0 0 Canada goose· 0 0 0 - 2 - 0 - 0 Lesser scaup· 10 0 : - 0 0 - 0 - 0 0 - 0 0 - 0 - 0 0 - 0 0 - 0 - 0 2 0 17 32 13 16 15 3 7 __7 Blue-wing teal· 2 Redhead· 2 - Totals 36 60 • These species were not modeled since they were not observed in 1989. 'Table 2 . Predicted versus observed numbers of breeding duck pa:f.rson National . J -..J N �Forest study units, 1989-1991. No. of resident pairs Observed Unit No. ponds 1989 1990 1991 Predicted I~~;z.. [06.1.) N. Fork N. Platte 19 Forester Creek 22 7 Goose Creek 41 23 35 18 ~~ Schafer Creek 22 10 9 9 8 25 37 31 North Kettle Ponds 23 17 IS 160 6 135 32 Newcomb Creek 57 57 61 Livingston Park 37 17 17 Colorado Creek 47 27 35 Lone Pine/Bear Crks. 59 8 19 Shorty/Cataract Crks. 75 17 32 Swamp Creek 17 15 3 Kawaneeche Bench 60 2 0 Thomas/Ammons Crks. 52 7 7 Aldrich Lakes 62 36 60 South Kettle Ponds 44 13 16 -- No pair counts conducted that year. I!: �74 APPENDIX B EFFECTS OF INGESTED TUNGSTEN-BISMUTH-TIN SHOT ON MALLARDS (Draft paper) JAMES K. RINGELMAN, Wildlife Research Center, Colorado Division of Wildlife, 317 W. Prospect Road, Fort Collins, CO 80526 MICHAEL W. MILLER, Wildlife Research Center, Colorado Division of Wildlife, 317 W. Prospect Road, Fort Collins, CO 80526 WILLIAM F. ANDELT, Department of Fishery and Wildlife Biology, Colorado State University, Fort Collins, CO 80523 Abstract: Mallards (Anas platyrhynchos) were orally dosed with 1 g of tungsten-bismuth-tin shot pellets and monitored for 32 days for evidence of intoxication. Minimum percentages of retained shot revealed in radiographs 1 day (87%) and 11 days (56%) post-dosing indicated acute exposure to the constituent metals, but no treatment or control birds died during the trial. Behavior of treatment and control birds remained normal throughout the trial, and food consumption and body mass changes did not differ between groups. Except for blood calcium, none of the 24 hematology or blood chemistry parameters measured 3, 11, 21, and 32 days post-dosing differed significantly between treatment and control birds. Differences in blood calcium were also judged to be biologically insignificant. No significant gross lesions were observed during postmortem examinations, nor did histopathologic examinations reveal any evidence of toxicity tissue damage. Tin and tungsten were not detectable «1 mgfkg) in kidney or liver samples. Bismuth concentrations in kidney and liver were low «114 mgfkg) and did not differ between control and treatment birds. We conclude that tungsten-bismuth-tin shot presents virtually no potential for acute intoxication in mallards under the condition of our study. In 1991, lead shot was banned for all waterfowl hunting in the United States because it is toxic when ingested by birds and other wildlife (Bel1rose 1959, U.S. Fish and Wildlife Service 1976). Sanctions against lead shot are also being considered elsewhere. Searches for nontoxic alternatives to lead, or for coatings, alloys, or composites to minimize the toxic effects of lead, have spurred the development and testing of >30 alternative shot formulations since 1950 (cited in U.S. Fish and Wildlife Service 1976, Haseltine a~d Sileo 1983, Boddington 1992). At present, steel shot is the only nontoxic shot approved for waterfowl hunting in the United States; Recently, a new shot pellet was formulated using tungsten (39.05% by weight), bismuth (44.49%), and tin (16.46%). Finely-powdered tungsten, a high-density (19.25 g/cc) metal with a very high melting temperature, is mechanically suspended in molten tin and bismuth to yield TBT shot with similar density and hardnessas lead (V. C. Oltrogge, Denver, Colo., a). Previous studies suggest that none of the metals in TBT shot are likely to be toxic in their inorganic state. (Grayson 1983, V&L 1978, Grandy et a1 1968) apatents are pending on a range of proportions of these and other metals as well as a range of densities, and on other manufacturing processes. �75 Ringelman et al. Biomethylated tin can be toxic (Zuckerman et al. 1978, Eisler 1989) but low solubility, poor absorption, low accumulation, and rapid excretion . diminish toxicity of inorganic tin compounds (Venugopal and Luckey 1978, Eisler 1989). Before a candidate nontoxic shot can.be approved for use in the U.S., it must undergo corrosion tests and acute, chronic, and reproductive toxicity testing (U.S. Fish and Wildlife Service 1989). We conducted an acute toxicity trial that generally followed the protocol set forth in federal guidelines to determine whether mallards that ingest TBT shot would experience (1) increased mortality, (2) sublethal physiological effects as indicated by some combination of behavioral changes, decreased food consumption, loss of body mass, changes in blood chemistry, or gross and/or histopathological abnormalities, or (3) increased concentrations of either tin, bismuth, or tungsten in liver or kidney. We express special appreciation to V. C. Oltrogge for supplying TBT shot and technical expertise, to M. R. Szymczak for his assistance in handling live birds, necropsy, and manuscript review, and to E. S. Williams for histopathological evaluation of tissue samples. L. S. Butler, K. Cook, and M. R. Stratman cared for captive ducks throughout the trial, and together with A. L. Case, C. W. McCarty, Dr. M. A. Wild and several other persons provided valuable assistance during necropsies. The eager cooperation of J. W. Van Cleave, L. M. Vap, J. R. Self, and other personnel at Colorado State University's Veterinary Teaching Hospital and Soil Testing Laboratory was greatly appreciated. D. C. Bowden consulted on statistical methodology, and W. J. Adrian, N. T. Hobbs, R. M. Hopper, V. C. Oltrogge, and M. R. Szymczak provided helpful comments on early drafts of the manuscript. This study was cooperatively funded by V. C. Oltrogge and the Colorado Division of Wildlife through Federal Aid in Wildlife Restoration Project W-166-R. METHODS We used 20 male and 20 female, 5-month-old game farm mallards with plumage and body conformation resembling wild mallards. Birds were marked with uniquely numbered bands, and 1 male and 1 female randomly assigned to each of 20 3 x 4-m outdoor pens at the Colorado Division of Wildlife's Foothills Wildlife Research Facility (Fort Collins, Colo.) on 8 July 1991. Water for swimming and drinking was provided from a single source, flow-through well system. Grit from natural sources was available in each pen. Ducks were fed a nutritionally balanced, pelletized commercial ration (crude protein ~20.0%, crude fat ~2.0%, crude fiber ~10.5%, calcium 3.0 to 3.5%, phosphorus ~1.0%) ad libitum during both the acclimation (8 Jul-29 Sep 1991) and trial (30 Sep-1 Nov 1991) periods. We randomly assigned 10 pens to both treatment and control groups. From 1.00 to 1.04 g of TBT shot pellets (~- 1.03 g; equivalent in mass to 5 number 4 lead shot) suspended in about 0.4 m1 carboxymethy1ce11u10se was introduced into the esophagus of each treatment bird with a modified tuberculin syringe following the methods of Trust et a1. (1990). Fifteen birds received 12 shot each, but because experimental shot size varied somewhat and its supply was limited the remaining 5 birds each received 13-17 pellets. Control birds were sham-dosed by introducing about 0.4 m1 carboxymethy1ce11u1ose into the esophagus, and were otherwise handled identically to treatment birds �76 Ringelman et al. throughout the experiment. After dosing, all birds were observed daily for changes or abnormalities in coordination, locomotor abilities, posture, or avoidance behavior that might indicate intoxication. Food consumption by treated and control birds was measured volumetrically by pen at 24-hour intervals, then averaged for 5 successive 6-day intervals for statistical analyses. Ducks were weighed (±l g) on days 0 (pre-dosing), 11, 21, and 32 (end of trial). An average body mass/bird for the 3 latter weighing dates, as well as analyses of body mass within each date, were used to assess the effects of ingested shot on mass. On trial days 1 and 11, all birds were captured and transported to the nearby Colorado State University Veterinary Teaching Hospital (CSUVTH) for radiography to determine shot retention. All birds were immobilized on a plexiglass restraining board (J. K. Ringelman, Colorado Division of Wildlife, unpublished manuscript), but only treatment ducks were radiographed. We collected blood from 12 randomly chosen (without replacement) mallards on day 0 (pre-dosing), 8 birds on day 3, 12 birds on day 11, and 8 birds on day 21; the 12 birds sampled on day 0 were resampled on day 32. Numbers and sex ratios of birds were balanced between treatment and control groups during each sampling period. We collected about 3 ml of blood from each bird via jugular venipuncture with heparinized syringes. Blood was transferred to heparinized glass tubes and transported to the Clinical Pathology Laboratory at CSUVTH. Federal guidelines for the approval of a new nontoxic shot require examination of 6 blood parameters, some of which (e.g., zinc protoporphyrin) are most appropriate if toxicity from a specific metal (lead in the latter example) is suspected. However, because we were evaluating shot composed of metals with unknown potential effects, we believed complete diagnostic blood screening would be more thorough and informative (Campbell and Coles 1986, Lewandowski et al. 1986, Fairbrother et al. 1990). Packed cell volumes, hemoglobin concentrations, and complete blood counts were determined for whole blood samples (Campbell and Coles 1986, Campbell 1988). In addition, plasma was subjected to a panel of tests (Table.1; see Boehringer Mannbeim Diagnostics Division [1988] for details of assay methods) performed using an automated analyzer (Hitachi 704, Boehringer Mannbeim Corporation, Indianapolis, Ind.). We did not perform analyses specifically intended to reveal toxic effects of metals not in TBT shot. At the end of the trial (day 32, 1 Nov 1991), all treatment and control birds were euthanized by CO2 asphyxiation and necropsied. In particular, the liver, kidneys, spleen, heart, lungs, air sacs, skeletal muscle, integument, gastrointestinal tract, gonads, thyroid glands, and brain from each bird were examined visually for gross lesions. Representative samples of these tissues were preserved in 10% buffered formalin. Microscopic histopathological examinations were performed on hematoxylin- and eosin-stained tissue sections. Gizzard contents were removed, rinsed with water, then radiographed to determine presence of eroded shot. Kidney and liver samples (1.25 ± 0.02 g) from each of the 20 treatment birds were pooled at random (birds from each pen remained together) into 5 groups of 4 samples. Samples of kidney and liver from the 20 control birds were pooled in an identical manner. Kidney and liver from individual birds were retained in the event individual re-analyses were warranted based on pooled sample results. All samples were stored at -20 C until analyzed. Dry �77 Ringelman et al. Table 1. Plasma chemistry and hematology parameters used in evaluating the physiological effects of ingested tungsten-bismuth-tin shot. Parameter (Abbreviation) Physiological indications4 Units .Metabolic state, stress Nutrition, renaljhepatic disease Nutrition, renaljhepatic disease Nutrition, hepatic disease Trauma, myopathy, lead toxicity Parathyroid function, lead toxicity Severe hemolytic or hepatic disease Severe hepatic disease Hepatocellular disease, myopathy Renal function Renal damage Severe renal damage Parathyroid and renal function Nutrition, renal disease Electrolyte balance, renal disease Electrolyte balance, renal disease mg/dl g/dl g/dl mg/dl lUll lUll mg/dl lUll lUll mg/dl lUll mg/dl mg/dl mg/dl meq/l meq/l Plasma Chemistries Glucose (GLU) Total Protein (TP) Albumin (ALB) 'Cholesterol (CHOL) Creatinine Phosphokinase (CPK) Alkaline Phosphatase (ALP) Total Bilirubin (TBILI) Alanine Amino Transferase (ALT) Aspartate Amino Transferase (AST) Uric Acid (UA) Gamma Glutamyl Transferase (GGT) Blood Urea Nitrogen (BUN) Calcium Phosphorous (P) Sodium Potassium (K) Hematology Packed Cell Volume (PCV) Hemoglobin (HGB) Total White Blood Cells (WBC) Heterophils (HET) Lymphocytes (LYM) Monocytes (MONO) Eosinophils (EOS) Basophils (BASOS) Anemia, Anemia, Stress, Stress, Stress, Stress, Stress, Stress, dehydration toxicity infectious disease infectious disease infectious disease infectious disease parasitic disease infectious disease percent mg/dl cells X 103 cells X 103 cells X 103 cells X 103 cells X 103 cells X 103 4See Campbell and Coles 1986, Lewandowski et al. 1986, Campbell 1988, Fairbrother et al. 1990. mass concentrations of tin, bismuth, and tungsten in the pooled samples were determined by atomic absorption spectrophotometry at CSU's Soil Testing Laboratory. Accuracy of the spectrophotometer was ascertained by testing with National Bureau of Standards Reference Materials; detection limits were about 1 mg/kg for each element assayed. Blood chemistry parameters, body masses, and food consumption were compared using ANOVA (PROC GLM, SAS lnst., Inc. 1988). Treatment (dosed versus undosed), day, and sex were used as main effects and combined as �78 Ringelman et al. interaction terms in blood chemistry analyses. Body treatment and sex as main effects, with days used as the food consumption analyses, we used treatment and and days as repeated measures. Metal concentrations compared between treatment and control g~oups with a Inst., Inc. 1988). mass analyses considered repeated measures. For pen as the main effects in kidney and liver were ~-test (PROC TTEST, SAS RESULTS TBT pellets in the gizzard and intestines were evident in plain-film radiographs taken on days 1 and 11. Because radiographs lack a third dimension (so some pellets may have been concealed by others) radiographs indicate the minimum number of shot retained 'by a bird. However, we observed radiographic evidence of shot passage in 1 bird on day 1. Of 258 shot pellets given to mallards, 225 (87%) were identifiable on day 1; an average of 11.3 shotfbird was retained (range - 8-16). On day 11, 114 shot pellets were visible, a 49% decrease from day 1 retention and a 56% decrease from the original dose. Birds retained a mean of 5.7 shot (SD - 3.2, range - 1-13) on day 11, but severe gizzard erosion was evident. A radiograph of gizzard contents 32 days after dosing revealed that 2 birds retained identifiable pellets and at least 3 other birds had residual shot material adhering to the surface of grit particles. This adhered metal was difficult to distinguish by visual examination, but was evident on radiographs. All treated and control ducks survived. Daily food consumption ranged from 145 to 657 cc per pen (2 birds per pen), and varied markedly in response to disturbance from handling the previous day (Fig. 1), and to ambient temperature. However, dosed and undosed birds did not differ (f > 0.39) in food consumption over time (Fig. 1). Body masses of males and females were greater than those reported for wild birds (Owen and Cook 1977, Whyte and Bolen 1984, Delnicki and Reinecke 1986), and were slightly higher at the end of the trial than at the beginning (Table 2). Differences in mean body mass for days 11, 21, and 32 were attributable only to bird sex (E - 19.92, f - 0.0001); body mass did not vary in relation to treatment (E - 0.14, f - 0.71) or as an interaction between sex and Table 2. Body mass (xiSE, g) of mallards dosed with tungsten-bismuth-tin shot. Sample size is 10 birds per treatment-sex group. Treatment Control No. days after dosing Males 0 11 21 32 1,379 ± 31.9 1,343 ± 34.4 1,371 ± 34.0 1,496 ± 37.1 Females 1,160 1,132 1,159 1,260 ± ± ± ± 19.4 20.4 20.3 22.2 Males 1,314 ± 49.0 1,286 ± 56.4 1,313 ± 60.7 1,417 ± 65.6 Females 1,164 ± 20.0 1,154 ± 21.5 1,200 ± 27.3 1,307 ± 21.0 �I') • 700 CONTROL 600 - o o z !: ;.~ . ,: " ............................................................................................................................... DOSED 500 (N = 10) " , (N = 10) .................................................................................................................................. ~ ,.... ·0 b: ~ :J 400 C/) Z o o c o o LL. , , " ..:" 300 . , 200 100 5 10 15 20 25 DAY Fig. 1. Mean daily food consumption of mallard pairs dosed with tungstenbismuth-tin shot and undosed (control) mallards. 30 �80 Ringelman et al. treatment (I - 1.78, f - 0.19). Analyses of body mass differences for each weighing date confirmed these same trends (all treatment f ~0.35). One female in the control group began laying eggs on day 18 of the trial, and continued laying intermittently through the end of the experiment. This female's food consumption and body mass dynamics were similar to those of the other control birds, but her blood chemistry values differed and were eliminated from analyses. In addition, measures of AST, ALT, ALP, and GGT were unavailable for 1 treatment bird. None of the 24 hematology or blood chemistry parameters differed between treatment and control groups prior to dosing (~ values < 1.86, 10 df, f values > 0.09). Except for blood calcium, none of these 24 parameters obtained 3, 11, 21, and 32 days post-dosing differed (f > 0.05) between treatment and control birds (Table 3). Blood calcium of control birds (~- 10.27) was greater (f - 0.02) than treatment birds (~- 9.71; Table 3). Day effects (f ~ 0.05) were noted for 13 parameters, whereas sex was a significant (f < 0.05) effect for 4 parameters (Table 3) .. Except for blood calcium, all interaction effects of day and sex with treatment were insignificant (f> 0.05). Ingestion of TBT shot had no apparent pathological effects on mallards. No significant gross lesions were observed during postmortem examinations of treatment or control birds. Similarly, histopathologic examinations revealed no evidence of toxicity or other tissue changes attributable to shot ingestion. Ingestion of TBT shot did not affect levels of tin, tungsten, or bismuth residues in either liver or kidney tissue. Neither tin nor tungsten were detectable (concentrations < 0.1 mgfkg) in any of the pooled liver or kidney samples from either treatment or control mallards. Bismuth concentrations in pooled liver samples did not differ (~ - 1.38, 8 df, f - 0.21) between treatment (~- 95 mgfkg) and control (~- 114 mgfkg) groups; similarly, bismuth concentrations in pooled kidney samples from treatment (~- 83 mgfkg) and control (~- 88 mgfkg) birds were indistinguishable (~ - 0.54, 8 df, f 0.60). DISCUSSION Ingestion of TBT shot caused no apparent acute toxicity in mallards under these experimental conditions. Although some (~13%) TBT shot were quickly voided by dosed mallards, nearly half were retained in the gizzard through day 11. In addition to individual variability in digestion rate and random voiding of shot, the quantity and quality of grit also affect shot erosion and retention rates (Longcore et al. 1974). We offered all birds the same type of grit ad libitum, but the variability in grit size «1 to 5 mm diameter) apparent at necropsy could have affected acute exposure rates. Nevertheless, the large number of small shot pellets ingested, which resulted in a large surface area of exposure for a relatively small volume, along with the flattened appearance of shot on day 11, suggest that our experiment induced considerable erosion resulting in some acute exposure to tungsten, bismuth, and tin early in the trial. This exposure did not cause mortality, behavioral or physiological abnormalities, changes in food consumption, or loss of body mass. Timing of this trial coincided with the onset of courtship and pairing in mallards and early pre-basic molt in females. This may account for variation �81 Ringelman et al. Table 3. Clinical chemistry and hematology parameters and significance levels of ANOVA main effects for control and treatment mallards dosed with tungstenbismuth-tin shot. Sample sizes are 19 control birds and 20 treatment birds for each parameter, except AST, ALT, ALP, and GGT where treatment sample size equals 19 birds. f va l ue s" Par ame t e r" Treatment Control means means Day" Treat. Sex Plasma Chemistries GLU TP ALB CHOL CPK ALP TBILI ALT AST UA GGT BUN C P N K 190.95 4.19 1.93 236.00 281.15 297.26 0.53 23.42 17.31 6.16 9.95 1.45 9.71 2.74 149.60 3.73 ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± 3.93 0.15 0.20 7.37 42.33 33.59 0.31 1.40 4.82 0.95 1.81 0.20 0.21 0.40 0.96 0.41 194.79 4.09 1. 73 234.37 268.59 259.16 0.15 21. 95 8.74 5.18 12.74 1.26 10.27 2.40 151. 21 3.29 ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± 3.57 0.06 0.03 6.77 30.50 21.11 0.02 1. 69 0.81 0.64 2.33 0.13 0.14 0.22 0.83 0.20 0.51 0.58 0.37 0.76 0.80 0.34 0.26 0.46 0.09 0.14 0.14 0.25 0.02 0.43 0.18 0.36 0.05 0.16 0.28 0.41 0.12 0.05 0.35 0.02 0.27 0.01 0.01 0.01 0.80 0.03 0.29 0.92 0.35 0.63 0.37 0.01 0.60 0.03 0.32 0.84 0.42 0.99 0.03 0.49 0.03 0.66 0.80 0.34 46.40 13.96 7.00 3.06 3.34 0.20 0.22 0.25 ± ± ± ± ± ± ± ± 0.51 0.22 0.79 0.23 0.68 0.05 0.07 0.04 46.42 13.75 6.32 2.78 2.98 0.15 0.14 0.25 ± ± ± ± ± ± ± ± 0.68 0.31 0.71 0.25 0.57 0.04 0.05 0.04 0.98 0.47 0.42 0.42 0.57 0.34 0.34 0.87 0.16 0.01 0.01 0.01 0.01 0.01 0.17 0.01 0.49 0.94 0.11 0.29 0.13 0.92 0.92 0.67 Hematology PCV HGB WBC HET LYM MONO EOS BASOS ·See Table 1 for parameter descriptions and units. ~ype III sum of squares; f values ~0.01 expressed as f ~ata from birds sampled on days 3, 11, 21, and 32. = 0.01. �82 Ringelman et al. in several blood chemistry parameters with sex and/or time (Driver 1981, Fairbrother et al. 1990). Only blood calcium appeared affected by treatment; however, the difference between treatment and control bird means was only 5.8%, and means for both groups were within the range of values reported elsewhere (e.g., Fairbrother et al. 1990). If these differences are real, we doubt that they are biologically significant given the normalcy of calcium values of treatment birds. However, we believe the apparent difference in calcium levels between control and treatment groups is more likely indicative of a Type I error exacerbated by the large number of statistical comparisons we performed. Gross and histolopathogical examinations of mallard tissues revealed no evidence of toxicity, further supporting interpretation of clinical chemistry data. In animals given oral doses of tin and bismuth, the highest concentrations have been found in the kidneys and liver (Venugopal and Luckey 1978, Hassett et al. 1984). Highest concentrations of tungsten occur in the kidneys, liver, and spleen of animals, but this element is poorly retained by these organs. Tin and tungsten were undetectable in liver or kidney samples from any of the mallards examined here, and ingestion of TBT shot had no effect on bismuth concentrations in liver and kidney tissues from treated mallards. These findings, combined with lack of any demonstrable behavioral or physiological compromise, suggest TBT shot ingestion presents virtually no potential for acute intoxication in mallards. MANAGEMENT IMPLICATIONS Ingested tungsten-bismuth-tin under the conditions of our acute approved for use as nontoxic shot outlined by federal regulations. the acute toxicity test. shot appeared harmless to mallard ducks exposure trial. However, before TBT can be it must pass more lengthy testing procedures Our study suggests that TBT shot will pass LITERATURE CITED Bartels, K. E., E. L. Stair, and R. E. Cohen. 1991. Corrosion potential of steel bird shot in dogs. J. Am. Vet. Med. Assoc. 199:856-863. Bel1rose, F. C. 1959. Lead poisoning as a mortality factor in waterfowl populations. Illinois Nat. Hist. Surv. Bull. 27:235-288. Boddington, C. 1992. Breakthrough in nontoxic shot. Magazine, Jan. 1992. Petersen's Hunting Boehringer Mannheim Diagnostics Division. 1988. Hitachi 704 operator's manual, Section 3, Chemistry application sheets. Boehringer Mannheim Corp., Indianapolis, Ind. mimeo. Campbell, T. W. 1988. Avian hematology and cytology. Press, Ames. 10lpp. Iowa State Univ. �83 Ringe1man et a1. Campbell, T. W., and E. H. Coles. 1986. Avian clinical pathology. Pages 279-301 in E. H. Coles, ed. Veterinary clinical pathology. W. B. Saunders Co., Philadelphia, Pa. De1nicki, D., and K. J. Reinecke. 1986 .. Mid-winter food use and body weights of mallards and wood ducks in Mississippi. J. Wi1d1. Manage. 50:43-51. Driver, E. A. 1981. Hematological and blood chemical values of mallard, Anas p1atyrhynchos p1atyrhynchos, drakes before, during, and after remige moult. J. Wi1d1. Dis. 17:413-421. Eisler, R. 1989. Tin hazards to fish, wildlife, and invertebrates: a synoptic review. U.S. Fish vn.ai , Serv. Bio1. Rep. 85 (1.15). 83pp. Enck, J. W., and D. J. Decker. participation in New York. 1990. Influences on waterfowl hunting Human Dimensions Res. Unit Rep. 90-2. Fairbrother, A., M. A. Craig, K. Walker, and D. 0'Loughlin. 1990. Changes in mallard (Anas p1atyrhynchos) serum chemistry due to age, sex, and reproductive condition. J. Wi1d1. Dis. 26:67-77. Grandy, J. W. IV, L. N. Locke, and G. E. Bagley. 1968. Relative toxicity of lead and five proposed substitute shot ·types to pen-reared mallards. J. Wi1d1. Manage. 32:483-488. Grayson, M., editor. 1983. Kirk-Othmer encyclopedia of chemical technology: tin and tin alloys. Vol. 23, third ed. John Wiley and Sons, New York, N.Y. 979pp. Haseltine, S. D., and L. Sileo. 1983. Response of American black ducks to diet~ry uranium: a proposed substitute for lead shot. J. Wi1d1. Manage. 47:1124-1129. Hassett, J. M., D. L. Johnson, J. A. Myers, A. A1-Mudamgha, M. E. Me1cer, C. L. Kutcher, and M. M. Sembrat. 1984. The exposure of rats to inorganic tin: behavioral and systemic effects of different levels and modes of exposure. Pages 487-496 in D. D. Hemphill, ed. Trace substances in environmental health-XVIII. Univ. Missouri, Columbia. Lewandowski, A. H., T. W. Campbell, and G. J. Harrison. 1986. Clinical chemistries. Pages 192-200 in G. J. Harrison and L. R. Harrison, eds. Clinical avian medicine and surgery. W. B. Saunders Co., Philadelphia, Pa. Longcore, J. R., R. Andrews, L. N. Locke, G. E. Bagley, and L. T. Young. 1974. Toxicity of lead and proposed substitute shot to mallards. U.S. Fish and Wi1d1. Serv., Spec. Sci. Rep. Wi1d1. 183. 23pp. Martin, E. M., P. H. Geissler, and A. N. Novara 1991. Preliminary estimates of waterfowl harvest and hunter activity in the United �84 Ringelman et al. States during the 1990 hunting season. Admin. Rep., June 1991. 33pp. Matunas, E. A. 1987. U.S. Fish and Wildl. Servo Coping with steel shot. Am. Rifleman 135:44-49. Owen, M., and W. A. Cook. 1977. Variations in body weight, wing length, and condition of mallard (Anas platyrhynchos) and their relationship to environmental changes. J. Zool. (London) 183:377-95. Roster, T. 1990. Tungsten/polymer shot. Cooperative lead poisoning control information program, third quarter newsletter. 7pp. (mimeo.). SAS Institute, Inc. 1988. SAS/STAT user's guide, release 6.03 ed. Inst., Inc., Cary, N.C. 1028pp. Sanderson, G. C., S. G. Wood, J. D. Brawn, and G. L. Foley. preliminary study of bismuth shot in gamefarm mallards. Wildl. Nat. Resour. Conf. 57 (in press). SAS 1992. A Proc. N. Am. Trust, K. A., M. W. Miller, J. K. Ringelman, and I. M. Orme. 1990. Effects of ingested lead on antibody production in mallards (Anas platyrhynchos). J. Wildl. Dis. 26:316-322. U.S. Fish and Wildlife Service. 1976. Final environmental statement: proposed use of steel shot for hunting waterfowl in the United States. U.S. Fish and Wildl. Serv., Washington, D.C. 276pp. U.S. Fish and Wildlife Service. 1989. Nontoxic shot. 50 CFR § 20.134, 10-1-89 edition, 289-293. Federal Register Venugopal, B. and T. D. Luckey. 1978. Metal toxicity in mammals. Vol. 2: Chemical toxicity of metals and metalloids. Plenum Press, New York, N.Y. 409pp. Whyte, R. J., and E. G. Bolen. 1984. Variation in winter fat depots and condition indices of mallards. J. Wildl. Manage. 48:1370-1373. Zuckerman, J. J., R. P. Reisdorf, H. V. Ellis, III, and R. R. Wilkinson. 1978. Organotins in biology and the environment. Pages 388-424 in F. E. Brinckman and J. M. Be11ama, eds. Organometa1s and organometa1loids, occurrence and fate in the environment. Am. Chem. Soc. Symp. Ser. 82. Washington, D.C. 447pp. �85 Colorado Division of Wildlife Wildlife Research Report October, 1992 JOB PROGRESS REPORT State of _-->:C=o....:..;lo=-:.r-=a=do><--_ Project 22 Work Plan Job Title: : Job _2_ Migratory Game Bird Publications Period Covered: Author: Avian Research - Migratory Game Birds W-166-R-l 01 April 1991 through 31 March 1992 Michael R. Szymczak Personnel: James K. Ringelman and Michael R. Szymczak, Wi ldl ife Colorado Division of ABSTRACT The following list contains those articles that were prepared and/or submitted for publication or published during this segment: Gilbert, D. W., D. R. Anderson, J. K. Ringelman, and M. R. Szymczak. Response of nesting ducks to habitat management on the Monte Vista National Wildlife Refuge. Wildlife Monograph (In review) Jeske, C. W., D. W. Gilbert, D. R. Anderson, J. K. Ringelman, and M. R. Szymczak. Use of a restraining board and wing bands to immobilize and mark mallard ducks. J. Field Ornith. (In press). Jeske, C. W., M. R. Szymczak, D. R. Anderson, J. K. Ringelman, and J. A. Armstrong. Mortality factors and body condition/survival attributes of wintering mallards in the San Luis Valley, Colorado. J. Wildl. Manage. (In review) . Ringelman, J. K. 1991. Managing beaver to benefit waterfowl. and Wildl. Serv., Fish and Wildl. Leaf. 13. Ringelman, J. K. 1991. Ecology of montane wetlands. Serv., Fish and Wildl. Leaf. 13. U.S. Fish U.S. Fish and Wildl. Ringelman, J. K. 1991. Identifying factors that limit duck recruitment. U.S. Fish and Wildl. Serv., Fish and Wildl. Leaf. 13. Ringelman, J. K., M. R. Szymczak, C. W. Jeske, and K. E. Ragotzkie. Ulnar lipid as an indicator of depleted fat reserves in mallards. J. Wildl. Manage. 56:315-319. �86 Ringelman, J. K., M. W. Miller, and W. F. Andelt. 19_. Effects of ingested tungsten-bismuth-t in shot on rnallards. J. Wildl. Marnage. (in review) . Shenk, T. M., and J. K. Ringelrnan. 1992. Habitat use by cross-fostered whooping cranes in Colorado. J. Wildl. Manage. 56(4) (in press). Szymczak, M. R., and E. A. Rexstad. 1991. Harvest distribution survival of a gadwall population. J. Wildl. Manage. 55:592-600. Prepared by: ~R1t *n--" Michael R. Szymczak Wildlife Reasercher C and � https://cpw.cvlcollections.org/files/original/38e90aabbfc6387e7a572df302ce9dfa.pdf cb837d3c68208eac5763ecd39dcfd909 PDF Text Text Colorado Division Wildlife Research January 1993 1 of Wildlife Report JOB PROGRESS REPORT State of __ ~Co=lo~rad=o,-__ Project: __ ..I..(W.:.:..;;.-1~5::.:.:0~R~-....::6:.1-) __ : Peregrine FaIcon Restoration Program Period Covered: 1 July, 1992 - 30 June, 1993 Personnel: G.R. Craig, Colorado Division of Wildlife and J.R. Enderson, The Colorado College. ABSTRAcr In the 1993 peregrine breeding season, 61 territories were occupied by 53 breeding pairs that fledged 92 young. Productivity averaged 1.51 young fledged per total pair. Eggshell thicknesses averaged 7.2% thin. Contents of 11 nonviable eggs have been preserved for future analysis. This Job Progress Report represents a preliminary analysis and is subject to change. For this reason, information presented herein MAY NOT BE PUBLISHED OR QUOTED without permission of the author. ��3 PEREGRINE FALCON RESTORATION PROGRAM Gerald R. Craig SEGMENT OBJECTIVES 1. Annually monitor the number of breeding pairs of peregrines and their reproductive success in Colorado. 2. Annually monitor organochlorine pesticide levels in wild breeding peregrines in Colorado. 3. Monitor breeding population turnover through band recoveries, presence of color markers, and telephotographic identification of individual breeding adults. 4. Augment poor wild production by placement of captive hatched wiid young and captive produced young into occupied wild nests. 5. Release captive hatched and captive produced young at potential and vacant wild territories. 6. Monitor recruitment of reintroduced peregrines into the wild breeding population of Colorado. ME1HODS AND MATERIALS 1. VISitall known peregrine breeding territories throughout Colorado and observe them from a distance to establish the presence of breeding adults. Breeding pairs will be kept under surveillance to determine initiation of egg laying. Depending upon the individual female's reproductive history and eggshell condition (obtained through measurement of previous year's eggshell thicknesses) and availability of captive hatched young for release, breeding pairs either will monitored or manipulated as outlined in approach 4. Those pairs not designated to be manipulated will be revisited periodicaI1y throughout the nesting season to document reproductive success. When a pair's behavior indicates that egg laying has occurred and incubation is underway, the eyrie will be visited to document the number of eggs produced. The eggs will be candied to ascertain viability and approximate age. All nonviable eggs will be collected for chemical analysis. A second visit will be made to determine productivity, band nestlings, and collect eggshell fragments and unhatched eggs for thickness measurement and analysis under 2a and 2b. 2a. Eggshell fragments encountered during eyrie visits described in approaches 1 and 4a will be measured for index to thickness following standardized procedures. 2b. Whole, nonviable eggs which are encountered during eyrie visits will be collected, preserved and subtnitted to the appropriate Fish and Wildlife Service approved laboratory for pesticide analysis. Eggs collected from the wild in the course of Approaches 4a, 4b and 4c that are artificially at the Peregrine Fund's Boise, Idaho facility also will be submitted for shell thickness measurement and chemical analysis. 3. Peregrines present at breeding territories will be examined to determine the presence of bands or color markers. Band confirmation will be accomplished through observation from a distance with telescopes and concealed remote controlled cameras. When banded falcons are encountered, every effort will be made to read band numbers without trapping or handling the birds. It is possible this can be accomplished in most situations with a Questar field model telescope (80-130x). When band numbers cannot be discerned, attempts will be made to trap and examine the falcon at a time when capture will have least impact upon breeding activities. �4 4a. In accordance with an annual release plan developed and approved by the State, U.S. Fish and Wildlife Service, Bureau of Land Management, National Park Service, and the Forest Service, a predetermined number of wild breeding pairs will be manipulated to augment natural productivity. Pairs with a history of reduced clutch size, cracked eggs, or infertile or dead eggs will be candidates for fostering efforts. 4b. On occasion, it may be necessary to recycle several early breeding pairs in order to delay them until captive hatched young of the proper age are available for placement into wild sites. No later than 10 days after the last egg has been deposited, the eyrie will be visited and the entire clutch removed without replacement. Approximately 14 days after removal of the clutch, the pair will recycle, select another nest ledge, and deposit a second clutch of eggs. If the eggs are thin shelled, they may be replaced with plastic replicas and treated as outlined in approach 4a. This technique also works well to augment captive production with wild produced eggs. 4c. At times, pairs will select inferior eyrie ledges that may compromise nest success such as ledges that are too narrow to support a brood of large nestlings, the site may be vulnerable to predators, or it may be exposed to the elements. If the ledge cannot be mechanical1y improved, pairs can be relocated to other ledges through the recycling method described in approach 4b since they invariably relocate and select a new ledge when recycled. 5. In accordance with an annual release plan developed and approved by the State, U.S. Fish and Wildlife Service, Bureau of Land Management, National Park Service, and the Forest Service, a predetermined number of captive produced falcons wiII be released at unoccupied or potential sites through the technique of hacking. This technique is employed at locations that do not have the benefit of protection or care from adults. Young falcons of about 35 days of age will be placed in a hack box on a suitable cliff ledge at the reintroduction site. They will be fed and cared for by attendants until they are flying and capable of fending for themselves. This technique assures that the young become familiar with their surroundings and hopefully will return to the site as adults and take up residency. Hacking requires constant attendance and observation to protect the vulnerable young and assure they have sufficient food while they are dependent upon the hack site. While the hack sites will be operated by the State, actual costs to operate the sites will be borne by the appropriate land administering agency (Forest Service, Bureau of Land Management, and National Park Service). 6. Confirmed breeding territories and selected potential breeding sites will be surveyed annual1y to document the presence of released falcons and ultimately determine the success of recovery efforts. RESULTS AND DISCUSSION Survev Effort In 1993, 5 teams comprised of 2 observers each were assigned to monitor breeding activities and survey potential cliffs as time permitted. Three teams were assigned regions west of the Continental Divide and 2 teams were located east of the Divide. A total of 3,154 hours were expended monitoring 80 documented breeding territories for an average of 39 hours of observation per site. Remoteness and lack of time prevented teams from visiting 2 documented territories. An average of 5 visits were made to each site during courtship, incubation and young rearing phases. Approximately 4.4 hours of observation were expended per team in the course of each visit. As schedules permitted, the teams were instructed to survey important potential nests sites within their region to locate additional pairs. In all, 564 additional hours were devoted to survey 55 potential breeding cliffs. �5 Approximately 10.25 hours were expended at each site by the teams. This effort resulted in documentation of 4 additional territories. Territorv Occupancy Breeding territory occupancy at 61 from 1992 to 1993 (Table 1) in spite of occupancy of 4 previously undocumented sites. The addition of sites 82, 83, 84 and 85 and reoccupancy of 6 were offset by the vacancies at sites 65, 76, 78, 79, and 80. Those 1992 site that were vacated in 1993 were occupied by a combination of lone adults (site 79), pairs comprised of 1 member that was subadult (site 76), or were first time nesters (site 80). Failure and subsequent relocation can be expected when inexperienced birds first establish territories. The stabilization in the rate of occupancy should be expected since augmentation was curtailed after 1990. Sites on the West Slope and East Slope exhibited similar occupancy rates (Table 2). Reproduction Peregrine productivity in 1993 averaged 2.56 young fledged per successful pair (36 pairs fledged 92 young) and 1.51 young fledged per total pair (61 pairs fledged 92 young)(Table 3). Although remedial actions were not taken, productivity continued to remain well above the threshold of 1.25 considered necessary for population stability. Eggshell Condition Twelve whole, nonviable eggs were collected from 10 sites. Shell thicknesses from these eggs averaged 72% thin (.333mm) and ranged from 6.1% greater than pre-DDT era thickness (381mm) to 192% thin (290mm). Although overall shell thicknesses continue to improve, wide variability among sites is still cause for caution. Eggshell fragments were collected from 27 territories. All of the fragments were coJlected in the course of visits to band young and by that time, fragments were intermixed making it impossible to assign fragments to particular eggs, Because shell thickness may vary as much as 5% from the egg's pole to equator and possibility of shells intermixing within the scrape, thickness measurements of fragments are judged to be only an approximation of individual egg thickness. Where multiple eggs were represented from particular sites, the measurements were combined for a clutch average. The shell thickness measurements (with membrane) from the sample represented by fragments averaged 72% thin (0.333mm). Within clutch averages varied from 17.3% thin to 1.7% thinner than pre DDT era measurements. Organochlorine Residue in Eggs The 12 nonviable eggs collected during the 1993 season have been preserved along with 10 eggs encountered in 1992 and 9 from 1991. They will be submitted for pesticide analysis at a future date when funding is available. Release and Augmentation Efforts Due to the population increase, no remedial efforts such as fostering or hacking have been undertaken since 1990. Prepared BY' (j ,/( .~ Gerald R. Craig Wildlife Researcher C �6 Tabla 1. OCCUPANCY OF PEREGRINE BREEDING TERRITORIES IN COLORADO SITE PRE NO. 1 2 3 4 5 6 7 1964 1964 1965 1966 1967 1968 1969 1970 1971 191'2 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 P + + + P P P P P A A A M P P P P P P P P P M P P P P A P P P P 8 9 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 37 38 39 40 41 42 43 44 45 V + V P A M V F P V V V V V V + + + + + + + + + V + V P A P P P V P P P P P V V A F' V P P P P P P P P P P V A V V V V V A M V P P P V P P V V P P p' V P p' V V V P P V V V V V V A V V M V V V V V V V V V V P V V V V V V V V V V isIoricaI p V V V V p' P P V M V V V p' V V V V V M P P V P V V V V V V V V p' P V P V V V P V P V V V V P p' V V V V p' M p V P V V V P' P' V P V V V P V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V sites above this line P P V V P P P P P P V P' P M P P p' M V V V V V V M F F V P P V V V V V V V V P P F V V M V V V V V V V V V V V V V V A P M M P V V V V V V V V V V V V P V V V V V V V V V V V V V V V V V V V V V V M P V V V V P V V V V V V V V V V V V V V V V V V V V V V P P P P P V P V P V V V V V V V V V F V V V V V V V V V V V V V V V P P P V V V V P P P V V V V P P P P P V V V V V V V P P P P P P P P P V P V V V V V V V V V V V V V V V V V V V V V P P P P P P P P P P P P P P P V V V V V V V V P P P V M P P P P P P P P P P P P P P P P P P P P P p' P P P P V P P V V P P P P P P P P P P p' p' P P p' P V p' P pO P P P p' P P P P P V P P P V V P P p' P P P P P P P P P P P P P P P p' P P P P P P' v P V P P pO V P V V V V V P P P P V V P P p P pO P P P P P V V A p P p' P V P P P F V P P P P P P P P P P P P P P P P P p V V F F V V V V V V V V V P V P P P P p' P P P P V V V V P P P M V V P P P P P P P A 69 70 71 1'2 73 74 75 7fS 77 78 79 V V V V V V V M P P P V V P p' V V V V V V V P V M V V F P V V V V V V V V V V V P P P P A •• Lane adult. 3:Ad. female F •• Lane adult tamale. replaced mate & female. by imm. femaIe~. 8:1mm. mate. M ,. Lane adult •.•••. P •• Adult pair. 4:Oead ad. hImaIe found in lricinity. 9:Ad. male paifed with female pIairie falcon. V os Vacant site. l:Ad. 5:Ad. male reptaced 10: Ages undetermined. male & imm. female. by imm. maIe~. 2:1mm. male & ad. female. 6:Ad. •.••• dead in Iricinity. p' P P P P P V P P M V V V P' p' 81 82 83 84 85 P P' P 80 P P P P P P V P P P P P P V P V P P V P P P P P P P P V V P P P P P V 68 P P P P P P P P P P P P P P V V 67 P P P P P V A V 63 64 65 66 P P P V P V P P p P P P p P V V P P p P pO P P M P P pO p 62 7:lmm. p P V V V P V V V 60 V P V P V 81 V P V P V V V P V V V V P P P P P P P V P P M" V V P P P P P V P V P P P V V V 48 49 V V V V V F V 46 P V V V P P P P P V P V P P P V V P V P V P P P P P V P V P P p' V F P 51 52 53 54 55 56 57 58 59 P P P P P V P V P P V p' 50 P P p p' P V P V P P P V V V V V P V V V P V p P P V V P V P V P + V P V p P V V V P V V V V V V P V p P P V P V P P V V p' V V V V P V P V V V P M P P V V V V V V M V V p' V p' V V P P' V V V P V p'. P p' p' P �7 Table 2. COMPARISON OF EAST AND WEST SLOPE S:I'l'E OCCUPANCY Statewide Occupied Sites Adult Pairs Breeding Pairs Young Produced Young Hacked 73 74 75 76 77 78 79 11 8 8 8 12 10 12 10 7 7 7 11 9 9 6 6 5 4 9 7 4 2 13 5 7 11 20 22 0 0 0 0 0 4 10 86 21 21 18 46 14 87 29 27 23 64 19 88 33 29 24 75 26 89 39 34 26 91 25 90 44 42 38 74 12 91 58 53 49 91 0 92 61 53 50 86 0 93 CUM 61 56 53 92 945 0 221 West Slo:ge Occupied Sites Total Pairs Breeding Pairs Young Produced Young Hacked 73 74 75 76 77 78 79 80 81 82 83 84 85 86 7 6 7 7 10 8 9 11 10 10 11 12 13 19 7 5 6 6 9 8 7 9 8 7 11 12 13 19 3 4 4 3 8 7 4 5 6 7 7 11 12 17 0 10 5 4 11 16 17 23 25 28 36 44 36 34 0 a 0 0 0 0 5 7 10 8 15 18 12 4 87 24 24 20 38 0 88 26 25 22 44 89 33 28 27 60 90 35 34 33 51 91 41 37 40 75 92 44 37 37 61 a 93 CUM 45 42 39 68 694 a 79 East Slo:ge Occupied Sites Total Pairs Breeding Pairs Young Produced Young Hacked 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 4 2 1 1 2 2 3 2 1 1 2 1 1 2 5 7 6 9 3 2 1 1 2 1 2 1 a 1 1 1 1 2 3 6 6 8 3 2 1 1 1 a a 0 a a a 1 1 1 3 2 4 5 2 3 0 3 a 4 5 4 3 5 8 12 14 12 26 31 31 23 0 0 0 0 0 4 5 4 3 5 8 9 12 10 19 26 25 12 91 17 15 13 16 0 92 17 16 12 25 0 93 CUM 16 14 14 24 251 o 142 80 13 10 5 27 11 81 11 8 6 28 13 82 11 8 7 33 13 83 13 12 7 44 23 84 13 13 12 56 27 85 14 14 13 50 24 a a a a �8 Table 3. Site 1 2 3 4 5 6 7 9 11 12 16 18 23 25 27 29 30 31 32 33 34 35 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 5S 57 58 59 60 61 62 63 64 66 68 69 71 72 73 74 75 81 82 83 84 85 Summary of 1993 Peregrine Production Age Male Female A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A Young Hatched Eggs Ist 2nd Clutch Clutch 3+ A A A A A A A A A A A A A A A A A A A A A A 2 3 2 4 2+ 0 2+ 3+ 4 0 3 2+ 3+ 2+ 2+ 3+ 1+ 1+ 3+ 2+ 0 2+ 1+ 4 2+ 4 Youn~ at 3w s. 2 3 2 4 2+ 0 2+ 3+ 4 Young Fledged 2 3 2 4 2 0 2 1+ 3+ 3+ 3+ 3+ 1+ 3+ 1+ 3+ 3+ 4+ 4 2+ 2+ 2+ 3+ 0 4 3+ 1+ 2+ 3+ 0 0 A 2+ 1+ 2+ 3+ 4 1+ 4 2+ 3+ 2+ 2+ 3+ 2+ 1+ 3+ 2+ 3+ 3+ 1+ 0 1+ ? 2+ 1+ 3+ + 1+ 1+ 3+ 3+ 3+ 3+ 1+ 4 1+ 3+ 3+ 4 4 2+ 2+ 2+ 3+ 1+ 4 3+ 1+ 0 4 3+ 2 3 2 3 A A A ? ? ? A 2+ 2+ 2 0 A A A A A 0 0 3+ 3+ 3 A A A A A A A A A A A A A A A A A A A A A A A ? A I A A A A A A A A A A A A A A A A A A A A A A A A A A A I A ? I I A 0 ? 2+ 1+ 3+ ? 6 0 3+ Total Sites Occupied: 61 Total Breeding Pairs: 53 Total Young Hatched: 113+ Average Fledged Brood Size:2.36* Total Fledglings Divided by Total Total Total Total Young Successful 0 3 2 3 2 2+ 3 1 1+ 3 2+ 0 2 1+ 0 2+ 3 4 0 2 2 3 2 2 0 1 1 3 2 0 2 1 0 ? ? ? 2 0 0 :3 ? 0 1+ 3 3+ 3 3 1+ 3 1 3 3+ 4 4 2 2+ 2 3 0 4 3 ? 6 1 2 3 1 3 1 3 ? 3 3 4 4 1 2 ? 2 0 4 3 Adult Pairs: 56 Successful Pairs: 39 Young Fled¥ed: 92 Fledged / otal Pair:l.51 Pairs �Colorado Division wildlife Research January 1993 9 of Wildlife Report JOB PROGRESS REPORT State of __ ...::Co=lo:.:..ra=d=o~ __ Project: __ (WI,...!.:._-1~5:..:;1-,-R:..:....::-5:..t.) __ : Bald Eagle Nest Site Protection and Enhancement Program Period Covered.: 1 July, 1992 - 30 June, 1993 Personnel: G.R. Craig, Colorado Division of Wildlife and R.L Knight, Colorado State University. ABSTRACT Bald eagles occupied 18 Colorado nesting territories in 1993. Four new territories were added of which 3 were successfu.l. Twelve pairs produced eggs and nine pairs were confirmed to have successfully fledged 18 young. Productivity averaged 1.00 young per occupied territory. This Job Progress Report represents a preliminary analysis and is subject to change. For this reason, information presented herein MAY NOT BE PUBLISHED OR QUOTED without permission of the author. ��11 BALD EAGLE NEST SITE PROTECTION AND ENHANCEMENT PROGRAM Gerald R. Craig SEGMENT OBJECTIVES 1. Monitor nest site occupancy and reproductive success. 2. Document survival rates and mortality factors. 3. Determine migration and wintering areas. 4. Determine if philopatry occurs in breeding eagles. 5. Determine nest site tenacity by individual breeding eagles. 6. Quantify nesting habitats and associated foraging areas in an effort to document nest site parameters conducive to improved reproduction. 7. Document pesticide contamination through eggshell measurement and chemical analysis of nonviable eggs. 8. Where necessary, implement actions to stabilize nests and maintain occupancy. METHODS AND MAlERIALS This work will be a cooperative endeavor between the Division and Dr. Richard Knight of Colorado State University. 1. Annually visit all documented breeding sites to determine the presence of bald eagles. Pairs at territories will be documented by DWMs and other field personnel. Previously unrecorded pairs will probably be revealed in the course of aerial eagle and waterfowl flights. DWMs will confirm actual incubation from ground visits. 2. Occupied territories will be visited by DWMs periodically throughout the breeding season to determine hatch of young, nesting failures, etc. 3. In May arid June, a Utility Worker will observe breeding eagles from a distance and endeavor to follow their movements to locate important foraging areas. Responses of eagles to various human activities and land uses will be recorded. 4. In June, when the young are determined to be old enough to band, sites will be visited by Craig and Knight to place a federal band on one leg and a colored, alpha numeric marker on the other. The color markers will permit identification if the young return in subsequent years. During the same nest visit the following will be recorded: Physical parameters such as tree species, height, DBH, condition, and dominance. Nest condition, size, and location. Vegetative community and land use practices. In addition, collect prey remains, nonviable eggs and eggshell fragments. 5. Approximately Sec's of blood will be collected from each nestling. The blood will be analyzed at the Savannah River Ecology Lab in Aiken, South Carolina. Electrophoretic examination will permit �12 genetic comparison with samples collected from other populations in Saskatchewan, the Lake States and Arizona, as well as determine the heterogeneity of the Colorado birds. 6. When necessary, remedial actions will be taken to stabilize nests that are threatened by wind throw. Should the tree be decadent and in danger of falling, an artificial nest base may be placed in a suitable, adjacent tree. Action will be taken only after it has been deemed desirable to encourage the eagles to nest at the same location. RESULTS AND DISCUSSION Territory Occupancy Bald eagle nesting activities for Colorado are summarized in Table 1. In 1993, 14 territories (Adams, Archuleta, Fremont, Gunnison. 1..aPlata #1, Mesa, Mineral, Moffat #1, #2 and #3, Montezuma #2 and #3, Rio Blanco #1, #3, and #4, Morgan #1 and Weld #3) were occupied. Four new territories (Fremont, Gunnison. Routt, and Weld #3) were added in 1993. Reports from local landowners suggest that the Fremont site was used in 1991 and 1992, and the Weld #3 site produced 2 young in 1992. Fma11y,a new site was occupied in Montezuma County by a pair that constructed a nest and incubated eggs but failed when the nest blew down. It is possible that this site is an alternate to the Montezuma #2 site since that pair did not occupy their nest. It is suspected that the pair relocated due to construction of a housing development adjacent to their historical nest. It is definite that the Gunnison site is new in 1993. The pair at the Mineral site were observed frequenting the area in 1992,but did not nest. In 1993,the female was paired with a subaduh male that did not contribute significantly to courtship. The pair at the Routt site may have relocated upstream from an unconfirmed nest. Land Status All the new territories are on private land and the landowners are aware of and protective of their eagles. Land use at the Fremont, Gunnison and Routt sites is livestock ranching. The Mineral site is associated with a private lake. Reproduction Reproductive efforts are summarized in Table 2. Nineteen young were hatched by 11 pairs and 18 were fledged by the pairs (1.64 young per successful pair) which yielded an overall productivity of 1.00 young per territory occupying pair. The reduced production from the previous 4 years was partially due to the presence of new pairs that typically fail in their early nesting attempts, as well as the failure of Moffat #2, and the lack of broods with 3 young. The Mesa County site failed during incubation for the third consecutive year, probably because of the pair's inexperience., Eight young were banded and color marked at 4 locations (Adams, Fremont, Morgan, and Rio Blanco #1). At the landowner request, the Gunnison and Rio Blanco #4 were not climbed. Fish and Wildlife bands were affixed to the nestlings' right legs and red alpha-numeric bands with yellow vinyl flags were affixed to their left legs. Culmen length and foot pad length measurements were obtained from the eaglets that were banded. Genetic Variation Tissue samples, including pin feathers, blood cells and serum, that were collected 1988-91were subjected to electrophoretic analysis for genetic loci. The results were compared to similar work done in Canada. There was little genetic variation between the Canadian and Colorado populations. A manuscript presenting the results is underway. �13 genetic comparison with samples collected from other populations in Saskatchewan, the Lake States and Arizona, as well as determine the heterogeneity of the Colorado birds. 6. When necessary, remedial actions will be taken to stabilize nests that are threatened by wind throw. Should the tree be decadent and in danger of falling, an artificial nest base may be placed in a suitable, adjacent tree. Action will be taken only after it has been deemed desirable to encourage the eagles to nest at the same location. RESULTS AND DISCUSSION Territorv Occupancy Bald eagle nesting activities for Colorado are summarized in Table 1. In 1993, 14 territories (Adams, Archuleta, Fremont, Gunnison, La Plata #1, Mesa, Minerai, Moffat #1, #2 and #3, Montezuma #2 and #3, Rio Blanco #1, #3, and #4, Morgan #1 and Weld #3) were occupied. Four new territories (Fremont, Gunnison, Routt, and Weld #3) were added in 1993. Reports from local landowners suggest that the Fremont site was used in 1991 and 1992, and the Weld #3 site produced 2 young in 1992. Fmally, a new site was occupied in Montezuma County by a pair that constructed a nest and incubated eggs but failed when the nest blew down. It is possible that this site is an alternate to the Montezuma #2 site since that pair did not oa:upy their nest. It is suspected that the pair relocated due to construction of a housing development adjacent to their historical nest. It is definite that the Gunnison site is new in 1993. The pair at the Mineral site were observed frequenting the area in 1992, but did not nest. In 1993, the female was paired with a subaduIt male that did not contribute significantly to courtship. The pair at the Routt site may have relocated upstream from an unconfirmed nest. Land Status All the new territories are on private land and the landowners are aware of and protective of their eagles. Land use at the Fremont, Gunnison and Routt sites is livestock ranching. The Mineral site is associated with a private lake. Reproduction Reproductive efforts are summarized in Table 2. Nineteen young were hatched by 11 pairs and 18 were fledged by the pairs (1.64 young per successful pair) which yielded an overall productivity of 1.00 young per territory occupying pair. The reduced production from the previous 4 years was partially due to the presence of new pairs that typically fail in their early nesting attempts, as well as the failure of Moffat #2, and the lack of broods with 3 young. The Mesa County site failed during incubation for the third consecutive year, probably because of the pair's inexperience .. Eight young were banded and color marked at 4 locations (Adams, Fremont, Morgan, and Rio Blanco #1). At the landowner request, the Gunnison and Rio Blanco #4 were not climbed. Fish and Wildlife bands were affixed to the nestlings' right legs and red alpha-numeric bands with yellow vinyl flags were affixed to their left legs. Culmen length and foot pad length measurements were obtained from the eaglets that were banded. Genetic Variation Tissue samples, including pin feathers, blood cells and serum, that were collected 1988-91 were subjected to electrophoretic analysis for genetic loci. The results were compared to similar work done in Canada. There was little genetic variation between the Canadian and Colorado populations. A manuscript presenting the results is underway, �16 Table 2. Colorado Bald Eagle Nesting Efforts - 1993 Site Age of Birds Male Female Adams Co. Archuleta Co. Fremont Co. Gunnison Co. La Plata Co.#1 Mesa Co. Mineral Co. Moffat Co. #1 Moffat Co. #2 5/25. Moffat Co. #3 Montezuma Co.#2 Montezuma Co. #3 Morgan Co. Rio Blanco Co. #1 Rio Blanco Co. #3 Rio Blanco Co. #4 Routt Co. Weld Co. #3 Adult Adult Adult Adul Adult Adult Imm. Adult Adult Adult Adult Adult Adult Adult Adult Adult Adult Adult Total 18 Adult Adult Adult Adult Adult Adult Adult Adult Adult Adult Adult Adult Adult Adult Adult Adult Adult Adult 18 Young Young Produced Fledged Comments 2 0 2 1 Pair renested. Pair incubated, failed. Nest used for past 2-3 years. ? ? 0 0 1 1 2 0 2 2 2 2 0 0 1 0 2 0 2 2 2 2 Pair incubated, outcome undetermined. Egg(s) failed to hatch. Female added sticks to pinnacle in lake. Young observed from aicraft, assumed fledged. Nestling observed from aircraft 4(2.7,nest empty Large young, assumed fledged. Pair present, failed to nest, may have relocated. 2 0 2 1 ? ? 2 0 2 0 19 18 Nest produced 2 young in 1992, disturbed at laying. �17 Colorado Division Wildlife Research January 1993 of Wildlife Report JOB PROGRESS REPORT State of Project: _:::CO=lo~rad=o::..__ (W-156-R-3) :Effects of Human Disturbance on Grassland Raptors Period Covered: Personnel: University 1 July, 1992 - 30 June, 1993 G.R. Craig, Colorado Division of Wildlife and R.L. Knight and T. Holmes, Colorado State ABSTRACT Flushing distances were recorded for nesting Swainson's hawks, red-tailed hawks, and ferruginous hawks that were subjected to experimental disturbances by humans approaching their nests on foot. For all species, flushing distances and call rate increased during the brooding phase. Nest height and distance to dwellings and roads were also correlated. Predictive models were developed and tested. This Job Progress Report represents a preliminary analysis and is subject to change. For this reason, information presented herein MAY NOT BE PUBUSHED OR QU01ED without permission of the author. ��19 RESULTS AND DISCUSSION Wintering Raptors Flush distance data colJected the winter of 1991-92 was reanalyzed and edited into a manuscript entitled "Responses of wintering grassland raptors to human disturbance" that was accepted for publication in The Wildlife Society Bulletin. Nest Studv In the spring of 1993, response to disturbance was observed for 16 pairs of nesting red-tailed hawks (8 incubating pairs and 8 brooding pairs), 16 pairs of Swainson's Hawks (8 incubating pairs and 8 brooding pairs), and 16 pairs of ferruginous hawks (8 incubating pairs and 8 brooding pairs). Each nest was disturbed 4 times. Nests were randomly assigned to 1 of 4 treatment combinations created by varying the duration of disturbances (5 or 15 minutes) and the interval between disturbances (10 or 30 minutes). Two replications of treatment were performed during each reproductive phase of each species. Disturbance consisted of approaching raptor nests on foot, always within the raptor's sight, and measuring the distance from disturber and raptor nest when: (1) the parent first called and (2) the parent flushed. Other measured variables were: whether the mate was present during the disturbance; the direction taken by the flushing raptor; the number of parental calls and dives within a 3 minute period; the nearest source of permanent disturbance (e.g. occupied dwelling, road, etc.) and its distance from the nest; nest height; the time required for the parent to return to the nest; temperature; wind speed; and cloud cover. Preliminary data analysis was performed. Since no year effects were found using ANOV, data from 1992 and 1993 were pooled. Dependent variables include: flush distance, call distance, call rate, and return time. Data from each species/reproductive phase combination was analyzed separately in repeated measures ANOVA to examine the effect of disturbance frequency (# visits», duration, interval, and any interactions between these 3 factors. Nest height and/or distance from the nest to nearest permanent disturbance were used as covariates if significant (P < .05). Effects of disturbance are often species-specific and specific to the reproductive phase. For example, in all species, brooding individuals flushed at greater distances and called more than incubating individuals; conversely, incubating individuals took longer to return to the nest than brooding individuals. InCllbating Swainson's and ferruginous hawks exhibited a quadratic trend in flush distances whereas brooding red-tailed hawks showed a linear increase in flush distance with successive disturbances. Nest height was often a significant covariate. For example, in incubating Swainson's hawks, the closer the nest to the ground, the greater the flush distance and return time. In tests for correlations between call rate and mate presence and between call rate and flush distance, only incubating red-tailed hawks showed a correlation. Both mate presence and flush distance increased with increased call rate. Regression analysis was used to build predictive models of Swainson's, red-tailed. and ferruginous hawk flush distances, call distances, call rates, and return times when exposed to human disturbance. Significant models (P< .05) were obtained for each species and dependent variable. As distance to nearest source of permanent disturbance increased, return times decreased and flush distance increased for all species, suggesting that species which nest further from human activity are less tolerant of disturbance. �20 EFFECTS OF HUMAN DIS1URBANCE ON GRASSLAND RAPTORS Gerald R. Craig and Tamara L Holmes SEGMENT OBJECTIVES 1. Review the literature concerning the concepts of: flushing response, flight distance, buffer zones and produce a summary report. In addition, review and incorporate information regarding habitat requirements of sensitive grassland raptors such as ferruginous hawks and burrowing owls. 2. Determine flushing responses and flight distances of a community of diurnal grassland raptors during both the breeding season and winter period. During the breeding season, determine flushing responses and flight distances at nests (for species that will not be adversely impacted) and at perch sites. 3. Describe the learning and decay rates of flushing responses and flight distances of selected raptor species and understand the mechanisms that influence flushing responses and flight distances. METHODS AND MA1ERIALS This work is a cooperative effort between the Division and Dr. Richard Knight of Colorado State University. 1. Perched raptors will be located during the winter and breeding seasons. During the nesting period, sensitive raptors such as Ferruginous hawks, will be approached only while perched away from their nest. More tolerant species will be approached while on the nest. In the course of the investigation, all raptors will be approached in a clear line-of-sight and the distance of the observer to the perch will be recorded once they flush. Where possible, the distance that the raptor flies will also be measured. Environmental variables including wind speed, temperature, and cloud cover will be noted. 2. For selected species, learning experiments will be conducted for perched raptors at and away from their nest. Again. sensitive species will not be flushed while present at their nests. The raptors will be approached at varying frequencies with the number of visits, duration of visits, and the interval between visits varied in a £3 factorial experiment (Box et aI. 1978). � https://cpw.cvlcollections.org/files/original/edfc9858ec234efa1f12978157b3c7a6.pdf e88c1479449ef8e848cc2e10386729f5 PDF Text Text JOB PROGRESS REPORT State of: Colorado Project: ~W_-~1~6~7_-~R~ _ Upland Bird Research Work Plan: 1__ Job Title: Evaluation of Habitat Development for Ring-necked Pheasants in Eastern Colorado Period Covered: Author: Job ~ 01 January through 31 December 1992 Thomas E. Remington and Warren D. Snyder Personnel: L. L. Bixler, C. E. Braun, T. J. Davis, G. E. Mekelburg, J. W. Moore, T. E. Remington, W. D. Snyder, M. Trujillo, M. M. Warmoth, L. L. Whitmore, and J. D. Wieland, Colorado Division of Wildlife ABSTRACT The Pheasant Habitat Improvement Program (PHIP) was implemented by contracting with local Pheasants Forever Chapters in northeastern Colorado during 1992. About 55 sorghum plantings ($11,278), 13 no-till wheat stubble fields ($2,068), 13 annual weed plots ($3,300) and 38 shrub thickets (some with windbreaks; $38,306) were contracted during spring and summer 1992 to enhance pheasant survival. Nearly all of the $55,000 provided to three Pheasants Forever Chapters was used in habitat work. Nine, 9-section blocks for evaluation of the Pheasant Habitat Improvement Program were selected, mapped, and landowner contacts for consigning habitat supplements were begun in portions of Phillips, Logan, Yuma, and Washington counties. Comparable control blocks (9) were selected anq mapped. Contracted annual plantings (sorghums and annual forbs) generally provided good winter survival cover for ring-necked pheasants (Phasianus colchicus). Woody plantings had good initial survival. Hunting pressure was assessed during the opening day of the 1992 pheasant season and was relatively low in both treatment and control blocks. Major efforts were directed toward preparations for implementing the PHIP within treatment blocks during 1993. A technical guide series for habitat development options under the Cooperative Habitat Improvement Program (CHIP). was prepared for use by Division personnel. ��3 EVALUATION OF HABITAT DEVELOPMENT FOR RING-NECKED IN EASTERN COLORADO Thomas E. Remington PHEASANTS and Warren D. Snyder INTRODUCTION Pheasants are pursued by more small game hunters than any other small game species in Colorado (83-88% of small game license buyers). In a recent survey, 74% of pheasant hunters rated their hunting trips in Colorado as poor (45%) or fair (29%), while only 10% rated their trips as very good or excellent. Lack of birds and places to hunt were identified as the most significant reasons why some hunters did not hunt pheasants in Colorado. Small game license sales have declined by about 70,000 (40%) in the last 10 years. It is apparent that if the Division is going to turn this decline around pheasants will be a key species. Presumably, recruitment and retention of hunters will increase if the quality of pheasant hunting is improved, i.e., increases in pheasant numbers and places to hunt. Previous research has indicated that over-winter survival of pheasants is the most critical factor limiting pheasant populations. The Pheasant Habitat Improvement Program was created to establish over-winter survival cover within historically good pheasant range in eastern Colorado. The program was conceptually designed to overcome significant obstacles to developing habitat, mainly a lack of manpower and a burdensome contractual system (costs of administering contracts exceeded costs of developments). Under PHIP, the Division of Wildlife contracts with individual Pheasants Forever chapters in eastern Colorado to contact landowners and develop habitat on private lands following specific guidelines. Each chapter develops contracts with individual landowners and pays them when the habitat work is completed and verified. Division of Wildlife personnel inspect a subsamp1e of habitat developments and verify completion and compliance with guidelines. P. N. OBJECTIVES To determine if habitat developments offered through the Pheasant Improvement Program increase pheasant survival, breeding density, harvest within selected northeast Colorado study areas. Habitat and pheasant SEGMENT OBJECTIVES 1. Work with the CHIP program coordinator, Pheasants Forever Chapters, management personnel, and landowners to develop habitat within treatment sites. 2. Evaluate Pheasants Forever and landowner acceptance of program guidelines and implementation, and consider modifications as suggested or needed. 3. Conduct evaluations of the quality of annual plantings as survival cover. �4 4. Monitor hunting pressure and pheasant harvest within treatment and control sites. 5. Map habitat within treatment and control sites. 6. Prepare and publish a CHIP Technical Guide series. 7. Prepare progress report. METHODS Cooperative Agreements were signed with Pheasants Forever Chapters in February 1992. A PHIP sub-contract form was prepared and included as Exhibit A of the contract (Appendix A) with each chapter. Division habitat guidelines were included as contract specifications (Exhibit B) and are attached (Appendix B). Members of Pheasants Forever chapters and Division personnel contacted farmers and prepared PHIP sub-contract forms. Individual landowners were paid by the contracting Pheasants Forever Chapters as soon as the habitat work was completed. Annual plantings that would provide satisfactory survival cover during the first winter were needed for evaluation purposes within the treatment blocks. Tall sorghums and tall annual forbs were about the only options. Switchgrass (Panicum virgatum) can also provide survival cover, but generally takes 1-2 years to establish. Shrubs thickets, with or without a windward windbreak, were offered to Pheasants Forever Chapters and were a priority item for most Chapters. However, they did not provide immediate cover for evaluation. Division personnel, involved with PHIP, inspected all plantings to verify completion. Subsequent inspections were also conducted to determine the relative quality of annual plantings as winter survival cover and to determine survival of seedlings within shrub thickets and associated windbreaks. To evaluate PHIP, 18, 9-section blocks (4.8 km/side) were selected within the primary study area (Fig. 1). Blocks were paired and randomly selected as either a treatment or control. Nine potential pairs were established based on the assumption that not all would obtain enough cooperating landowners to be included in the evaluation. Temporary personnel were assigned to mapping cover types, determining land ownership, and contacting farmers to present the PHIP program. Equipment needs were determined and equipment was ordered for use during the 1993 work year. Pre-treatment hunting pressure surveys were conducted within most treatment and control blocks during the first weekend of the 1992 hunting season. A route was established and driven within each block. Beginning and ending odometer readings and times were recorded. All hunting parties, including number of hunters when possible, were recorded as (1) on the road, (2) in the field, or (3) at farm residences. Contacts concerning hunter success were made when possible. �~ tm wm Treatment Fig. 1. Location and layout of PHIP treatment and control evaluation tm Control blocks. Ln �6 RESULTS The Terrestrial Section of the Division of Wildlife established contracts with four Pheasants Forever Chapters in early February 1992. Several months were required to process these contracts through the Division of Wildlife and other State offices involved with contractual services. Funds were received by the Phillips County, Yuma County, and Northeast Colorado (Logan County) chapters by early to mid-May. The check sent to the East-Central Colorado Chapter (Burlington in Kit Carson County) was apparently lost in the mail. That Chapter, lacking funds, adequate personnel, and a strong commitment toward the effort did not complete habitat work in 1992. The three Chapters within the proposed evaluation area in northeastern Colorado requested and were allocated $55,000 (Table 1). Nearly all of the Table 1. Total 1992 Pheasant Habitat Improvement Program funds allocated, used, and remaining by Pheasants Forever Chapter, northeastern Colorado. Chapter Allocated Used Remaining Phillips County $25,000 $24,984.09 $ 15.91 Yuma County $15,000 $14,938.19 $ 61.81 Northeast Colorado $15,000 $15,030.53· $ $55,000 $54,952.81· $ 77.72 Grand Total 0.00 $54,922.28 paid out of PHIP and $30.53 paid by Northeast Colorado Chapter. allocated funds were used in habitat work during spring and summer 1992 (Table 2). Members of the Phillips County Chapter donated both manpower and equipment and planted 26 shrub thickets (16 of these were accompanied by small windbreaks). Approximately 18,300 m of polypropoline weed-barrier fabric were laid. Other Pheasants Forever chapters contracted woody plantings to the Yuma County Soil Conservation District. Since the Phillips County Chapter of Pheasants Forever used nearly all of their $25,000 for woody plantings, most of the other habitat work completed in Phillips and Sedgwick counties was paid for by the Northeast Colorado Chapter. It als? paid for some woody plantings in Yuma County. �7 Table 2. Pheasant habitat planted and/or contracted by Pheasants Forever (PF) chapters during 1992, northeastern Colorado. Number HabitatLChaEter Contracts Plantings Acres 26 38 201 $7,388 2 2 6 360 11 __ll 94 3,530 39 55 301 $11,278 2 5 90 $1,000 Pament Sorghum Plantings Northeast Colorado· Phillips County Yuma County Subtotal No-Till Wheat Stubble Northeast Colorado Yuma County ---l± Subtotal 6 _8 1,068 ~ 13 166 $2,068 Disturbance Tillage (Annual Forbs) Northeast Colorado 1 2 22 $520 Phillips County 5 5 24 1,170 _.2.§. 1,610 102 $3,300 Yuma County _l Subtotal 7 13 5 5 $ 6,122 13 26 23,454 _9 Shrub Th i cket.s" Northeast Colorado Phillips County Yuma County Subtotal _7 25 _7 38 Grand Total 8,730 $38,306 $54,953 • Northeast Colorado Chapter of Pheasants Forever payments were primarily in Phillips, Yuma, and Sedgwick counties. b Includes several windbreaks placed windward of plum thickets. Weather Factors Influencing 1992 Efforts Conditions during April and May 1992 were extremely dry and were not conducive to planting woody plantings or sorghums. Watering of shrub thickets was required in early May in attempts to minimize mortality. A hard freeze in May �8 damaged winter wheat, some so severely that it was not harvested in portions of the area. It was severely stunted by dry weather in all areas resulting in short growth and subsequent short stubble when harvested. Some rains were received in late May and average or better moisture was received from June through August which promoted growth of planted sorghums. Evaluations of Habitat Types Most sorghum plantings were completed by farmers during the first two weeks of June; late summer evaluations indicated that most obtained tall growth (2 m) and provided fair to excellent survival 'cover into winter. Plantings that were completed late (in late Jun and early Jul) and those that were planted in more arid western parts of the region were the poorest. Annual weeds, when present, enhanced winter cover quality within the sorghum plots. Only a couple of plots failed to provide adequate survival cover. Several small areas containing tall annual forbs were purchased (to prevent them from being disced and destroyed) in mid- to late-summer. Most also contained standing wheat (that had frozen in May) that stood well because the heads had not filled. No-till wheat stubble was purchased if it possessed adequate height and was left standing through the spring nesting season. Only a few fields, primarily in Yuma County, were enrolled. Survival of both plums (Prunus americana) and junipers (Juniperus scopulorum) was excellent (approximating 90%) in spite of severe dry planting conditions. Those that survived dry spring weather received sufficient rainfall to sustain them through the summer and fall. A major snow in mid-November also provided winter protection and moisture to the seedlings. Hunting pressure within proposed treatment and control sites was surveyed during opening day of pheasant season on 14 November 1992 (Table 3). General observations indicated hunting pressure was light and averaged only a couple of hunting parties per 9-section block. The approximately IS-mile long census routes allowed near total observation of hunters during the interval of inventory. However, a few parties that were road hunting or traveling through the blocks may have been missed. Observations on the second day of hunting season revealed almost no hunters afield. Most contacted parties reported seeing pheasants but both pheasant numbers and harvest were considered low. Because of hunter mobility it would be difficult to determine and compare actual harvest among treatment and control areas. Hunters do not always know or remember the location where they observed or harvested pheasants. General observations indicated pheasant numbers were slightly higher in most areas than in previous years but a significant population increase had not occurred. Hunting pressure, which has reflected low densities in past years, tended to remain low. Pre-hunting season observations indicated considerable use of the sorghum and weed plots by pheasants. �9 Table 3. Pheasant hunting pressure recorded within the 9-section treatment and control blocks during opening weekend of pheasant season, 14-15 November 1992, northeastern Colorado. Block name Route miles min. Observed parties hunters Treatment Blocks Holyoke SEa Mai1ander Mailander" Kurtzer St. Petersburg 32.8 15.2 31.7 15.1 15.4 75 30 75 45 60 6 3 3 2 4 15 12 Control Blocks Paoli Paoli" Haxtun Haxtun" Phillips Cty. S SE Logan Cty. 14.6 26.1 14.2 30.2 15.0 19.0 30 60 45 70 30 50 0 4 3 5 2 6 0 11+ 13 6+ 13 18 " Location road parked field 5 2 1 1 3 ? 4+ 34 1 2 1 3 1 5 Birds 1 1 1 1 4 3 2 1 2 1 1 1 3 Completed during the morning of 11-15-92 (2nd day of the hunting season) CHIP Technical Guide Series Specifications for the Division's Cooperative Habitat Improvement Program (CHIP) were completed, submitted, and provided to field management personnel. Guidelines were completed for pheasants, quail, and prairie grouse, however, additional guidelines to cover other wildlife will be added in the future. A form for preparing pre-approval plans for CHIP contracts was also developed and distributed to field personnel. Preparations for 1993 Evaluations Because it took several months for internal processing of cooperative agreements, funding efforts for 1993 were initiated in October 1992. Pheasants Forever Chapters in northeastern Colorado showed considerable enthusiasm toward the program as indicated by the amounts of funds requested. The Phillips County Chapter requested $100,000, the Yuma County Chapter requested $60,000, the Northeastern Colorado Chapter requested $40,000, and a newly formed Washington County Chapter initially requested $30,000, and then requested an additional $20,000 in early 1993. The East-Central Colorado Chapter (Burlington) requested $20,000 when a cooperative agreement with the Division of Wildlife was signed in £arly December 1992. The PHIP was offered to other Pheasants Forever Chapters in southeastern Colorado, however, they elected to not participate in the program. Prepared bY~[' Thomas E. Remington Wildlife Researcher" Prepared by ~ru:D, ~~u (~ ') Warren D. Snyder Wildlife Researcher C ��11 I 10/92 Pheas. 1192 No. 3908-93 $100~000 COOPERATIVE AGREEMENT TInS COOPERATIVE AGREEMENT, made this ') +h day of &- f,.,J.,p y , 199.:;L._, by and between the State of Colorado acting by and through the Department of Natural Resources for the use and benefit of the WilDLIFE CO?v.IMISSIONAND TIIE DMSION OF WILDUFE, hereinafter referred to as the "State", and the A,;/{, 125 C'ot(r1~ Charier (')£ mrtfC.nis FlueJey ,whose address is _ 10. D C .{)e.11'Ue y= ;Sf, fit' LgQke , Co Ie)m dog' d 73 4f hereinafter referred to as the Contractor. 0 WHEREAS, authority exists in the Laws and Funds have been budgeted, appropriated and otherwise made available and a sufficient unencumbered balance thereof remains available for payment in Fund Number 461 , Agency Number PBA, Organization Unit 4300 , Appr. Code 376 , Program HUNT ,Function 0600 , Object Code 5281 ' Contract Encumbrance Number C ~ I '[).) C; q .2._ , and P9270NE-T WHEREAS, required approval, clearance, and coordination has been accomplished from and with appropriate agencies; and WHEREAS, the parties desire to increase pheasant numbers in eastern Colorado by providing for habitat improvements on private lands; and WHEREAS, the Contractor has sufficient manpower and resources available for contacting private landowners and negotiating agreements for habitat improvements on private lands, and WHEREAS, the State has funds available which could be used by private landowners for habitat improvement on private lands. WHEREAS, this agreement is awarded in full compliance with procurement code, 24103-205, C.R.S. NOW TIIEREFORE, it is hereby agreed that: I. The Contractor: A. Agrees to contact private landowners within the primary pheasant range in eastern Colorado (east of 1-25) for the purpose of obtaining landowner agreements for establishing winter survival plantings for pheasants in accordance with the cost share rates and practices set forth in Exhibit B, attached and incorporated by reference. 4..,_ j... I' �12 B. 2. C. Shall use the State's monies provided herein only to compensate individual landowners who have agreed to establish winter survival plantings for pheasants in accordance with cost share rates and practices set forth in Exhibit B. D. Agrees that all landowner agreements shall require the private landowner to comply with the practices set forth in Exhibit B. E. Agrees that all news releases or public statements relating to information acquired as a result of this agreement must have prior approval by the CDOW or its authorized representative prior to submission for publication. F. Agrees to provide the State with a monthly report showing number of agreements entered into with amount of money spent for the individual month and running totals. G. Agrees to permit the State to audit and/or inspect its records during the term of the contract. and for a period of three years following the termination of the contract, to assure compliance with the terms hereof, or to evaluate contractor's performance hereunder. The State: A. 3. Shall enter into agreements with individual landowners to develop winter survival plantings. A copy of sample agreement between the Contractor and an individual landowner is hereto attached as Exhibit A and incorporated by reference. The Contractor shall provide copies of signed agreements to the State within fifteen (15) days of signature. Agrees to provide the Contractor with funds not to exceed -5::::d?tu~...::.loo~' _ ~>J~~ ~ ?f;t - - dollars ($ 1t7~ @,aq) fqr this project. Payment shall be made upon execution of this contract. . It is Mutually Agreed: A. The State and the Contractor shall conduct verification and evaluation surveys confirming plantings were completed and assessing their values as pheasant habitat. Verifications shall be completed by August 31 of each year. page 2 of 5 pages .: �13 B. This agreement shall commence upon the signature of the State Controller or his designee and be in full force and effect through /Je (~mlee c3I; 119.3 . C. Either party shall have the right to terminate this Agreement by giving the other party thirty (30) days written notice by certified mail, return receipt requested. If notice is so given, this Agreement shall terminate on the expiration of the thirty days, and the liability of the parties hereunder for the further performance of the terms of this Contract shall thereupon cease, but the parties must continue to perform their obligations up to the .date of termination. All funds not paid to landowners by the Contractor for the purpose of this agreement shall be refunded to the State within thirty (30) day of termination of said agreement. Upon termination, the 'Contractor shall assign all Contractor's rights, title and interest' to any and all agreements entered into between the Contractor and individual landowners pursuant to this contract to the State. Contractor shall prepare and execute all documents necessary to accomplish the assignment. page 3 of 5 pages �14 Form 6-Ae.o:B SPECIAL PROVISIONS CONTROLLER'S APPROV.AL I. This contract shall not be deemed valid until it shall have been approved by the Controller of the State of Colorado or such assistant as he may designate. provision is applicable to any contract involving the payment of money by the State. This FUND AVAILABILITY 2. Financial obligations available. of the State payable after the fiscal year are contingent upon funds for that purpose being appropriated, budgeted and otherwise made BOND REQUIREMENT 3. [fthis contract involved the payment of more than fifty thousand dollars for the construction. erection. repair. maintenance. or improvement of any building. roac. bridge. viaduct. tunnel. excavation or other public works for this Slate. the contractor shall. before entering the performance of any such work included in this contract. duly execute and deliver to and file with the official whose signature appears below for the State. a good and sufficient bond or other acceptable surety to be approved by said official in a penal sum not less than one-half of the total amount payable by the terms of this contract. Such bond shall be duly execute by a qualified corporate surety, conditioned forthe due and faithful performance of the contract. and in addition. shall provide that if the contractor or his subcontractors fail to duly pay for any labor. materials. team hire. sustenance. provisions, provendor or other supplies used or consumed by such contractor or his subcontractor in performance of the work contracted to be done. the surety will pay the same in an amount not exceeding the sum specified in the bond. together with interest at the rate of eight per cent per annum. Unless such bond. when so required. is executed. delivered and filed. no claim in favor of the contractor arising under this contract shall be audited. allowed or paid. A certified or cashier's check or a bank money order payable to the Treasurer of the State of Colorado may be accepted in lieu of a bond. This provision is in compliance with 38·26·106 CRS. as amended. INDEMNIFICATION 4. Tu the extent authorized by law. the contractor shal! indemnify. save ant! hold harmless the State. its employees and agents. against any and all claims. damages. liability and court awards including costs. expenses. and attorney fees incurred as 3 result of any act or omission by the contractor. or its employees. agents. subcontractors. or assignees pursuant to the terms of :his contract. DISCRIMINATION AND AFFIRMA nVE ACTION S. The contractor agrees to comply with the letter and spirit of the Colorado Antidiscrimination Act of 1957. as amended. and other applicable law respecting discrimination and unfair employment practices (24·34-402. CRS 1988 Replacement Vcl.), and as required by Executive Order. Equal Opportunity and Affirmauve Action. dated April 16. 1975. Pursuant thereto, the following provisions shall be contained in 311State contracts or sub-contracts. During the performance of this contract. the contractor agrees as follows: (I) The contractor will not discriminate against any employee or applicaru for employment because of race. creed. color. national origin. sex. marital status. religion. ancestry. mental or physical handicap, or age. The contractor will take affirmative action to insure that applicants are employed. and th3t employees are trearec during employment. without regard to the above mentioned characteristics. Such action shall include, but not be limited to the following: employment. upgradin g. demotion. or transfer, recruitment or recruitment advertising; lay-offs or terminations: rates of payor other forms of compensation; and selection for trairung. including apprenuceship. the contractor agrees to post in conspicuous places. available to employees and applicants for employment. 2) The contractor will. in all solicitations or advertisements for employees placed by or on behalf of the contractor, state that all qualified applicants will receive consideration for employment without regard 10 race. creed. color. national origin. sex, marita! status, religion. ancestry, mental or physical handicap. or age. (3) The contractor will send to each labor union or representative of workers with which he has collective bargaining agreements or other contract or understanding. notice to be provided by the contracting officer. advertising the labor union or workers' representative of the contractor's commianent under the Executive Order. Equal Opportunity and Affirmative Action. dated April 16. 1975. and of the rules. regulations. and relevant Orders of the Governor. (4) The contractor and labor unions will furnish all information and reports required by Executive Order. Equal Opportunity and Affirmative Action of April 16. 1975. and by the rules. regulations and Orders of the Governor. or pursuant thereto. and will permit access to his books. records. and accounts by the contracnng agency and the office of the Governor or his designee for purposes of investiganon to ascertain compliance with such rules. regulations and orders. (5) A labor organizauon will nor exclude any individual otherwise qualified from full membership rights in such labor organizations, from membership In such labor organization or discriminate against any of its members in the full enjoyment of work opportunity. color. sex. age. national origin. or ancestry. I 24·34-402 (I) (c) ) or expel any such indrvidua: because of handicap, race. creec. (6) A labor crganizauon. or the employees or members thereof will nOI aid. abet. incite. compel or coerce the doing of any act defined in this contract to be discriminatory or obstruct any person from complying with the provisions of this contract or any order issued thereunder: or attempt either directly or indirectly. to commit any act defined in this contract to be discriminatory. ( 24·34---402 (I) (e) ) Page 4 Revised 51'11 J'J5·53·0 1·1 022 of _5_Pages �15 form 6-AC-02C (7) In the event of the contractor's non-compliance with the non-discrimination clauses of this contractor or with any of such rules. regulations. or orders. this conrrac: may be cancelled. terminated or suspended in whole or in part and the comractor may be declared ineligible: for further State COntl':lC1Sin accordance with procedures. authorized in Executive Order. Equal Opportunity and Affirmative Action of April 16.1975 and the rules. regulations. or orders promulgated in accordance there- .•.uh. and such others sanctions as may be imposed and remedies as may be invoked as provided in Executive Order. EqU:lI Opportunity and Affirmative Action of April 16. 1975 or by rules. regulations. or orders promulgared in accordance therewith. or as otherwise provided by law. . (8) The contractor will include the provisions of paragraph (I) through (8) in every sub-contract. subcomracrcr and purchase order. pursuant to Executive Order. Equal Opportunity and Affirmative Action of April 16. 1975. so that such provisions will be binding upon each subcontractor or vendor. The contractor will we such action with respect to any sub-contracting or purchase order as the contracting agency m:ly direct; as a means of enforcing such provisions. including sanctions for ncn-comptiance; provided. however. that in the event the contractor becomes involved in. or is threatened with. litigation with the subcontractor or vendor as a result of such direction by the contracting agency. the contractor may request the State of Colorado to enter into such litigation to protect the interest of the State of Colorado. COLORADO LABOR PREFERE:'ICE 6a. Provisions of 8-17-10 I & 102. CRS for preference . and are financed in whole or in part by State funds. of Colorado labor are applicable . to this contract if public works within the Stare are undertaken hereunder b. When construction contract for a public project is to be awarded to a bidder. a resident bidder shall be allowed a preference against a non-resident bidder from a Sl3le or foreign country equal to the preference given or required by the state or foreign country in which the non-resident bidder is a resident. If it is determined by the officer responsible for awarding the bid that compliance with this subsection.06 may cause denial of federal funds which would otherwise be available or would otherwise be inconsistent with requirements of federal law • this subsection shall be suspended. but only to the extent necessary to prevent denial of the moneys or to eliminate the . inconsistency with federal requirements (section 8-19-102. CRS). GENERAL 7. The laws of the State of Colorado and rules and regulations issued pursuant thereto shall be applied in the interpretation. execution and enforcement of this contract. Any provision of this contract whether or not inccrporared herein by reference which provides for arbitration by any extra-judicial body or person or which is otherwise in conflict with said laws. rules and regulations shall be considered null and void. Nothing contained in any provision incorporated herein by reference which purports to negate this or any other special provision in whole or in part shall be valid or enforceable or aV:lil:lble in any action at law whether by way of complaint. defense or otherwise. Any provision rendered null and void by Ihe operation of this provision will nOI invalidate the remainder of this ccntract to the extent that the contract is capable of execution. 8. At all times during the performance been or may hereafter be cs!ablished. sh:lll strictly adhere to :III applicable of this Contract. the Contractor 9. The signatories hereto aver th:llthey are familiar with 18-8-301. et seq .• (Bribery and Corrupt Influences) 1986 Replacement VoL. and th:lt no violation of such previsions is present. 10. The signatories IN WITNESS Contractor: (Full Leg:!l Name) aver Ih:lt to their knowledge, WH/i !~ no state employee has :I personal federal and stare laws. rules and regulations that have and 18-8-401. et, seq .• (Abuse of Public Office). CRS or beneficial interest whatsoever in the service or properly described herein: F. the parties hereto have executed this Contract on the day first above written. ~""'"~rpt.---+-~:;":'=:""';"':_~'-=-'24':"""'---By----~~~~~~~~~;L--------------'5' Position (Tille ,__ f)L-!.r...lie,--,i'-'/w~~t'~"z..-Z-;:;._ __ ""C:o.i~rttl!.L..___'~~7).L1 _ 7'/-lf&' '[/'19' (lfCorporatiO~ • • OF Natural ReSOyrC9S l J- ( ~!lest (Seal) E.XEctmVE DIRECTOR Deputy Director, Divis' n of W,ldl,fe \F8~lT~~eth Salazar, Executive Director \ ru J-.--£ By APPROVALS ATIORNEY GENERAL By CONTROLLER By _ \:tRIFlED lNFORMATION _ cor~ ·!-r.e oligina! and two copies of Hus eornract rava been signed by all state oiucials required tiy law to approve rn±: conracts, .1<l5·53-UI·I030 (Revised 5~ I) Pile 5 "See _hlch f1 In,,,,,Chont. the last of on rever-se 5 \Ide'. plCCi �neasant J. :_:~~" ~~ EXHIBIT Habitat I1Jlprovement A PHEASANTS D~_~ H_A_B_IT_A_T_P_R_OJ_E_C_T_A_G_RE_E_M_EN_T V AGREEMENT tI ADDRESS: Rte. _ LANDOWNER'-=" Town CHAPTER NAME R._ 8_ =_::__ . PHONE Zip _ COUNTY _ HABITAT DEVELOPMENTS Item Number Size lac Cost ----;.. __ ___.,:I~ Combined Cost +_-- I -:--__ Roadsides Tall annual weeds Tall Wheat Stubble ~18n No-till Wheat Stubble Custom Operations/Other Show Location of Habitat'OevelQP~ents .,..- =-:---:-- __ State ADDRESS Item Sorghums Shrub tRkket (.sO.2 ac.) Shrub thicket + windbreak Switchgrass ~I_-_ x,~ • "_ : s_ . 1 I C I <:.C\~.....------...;-1--S\,v I 1 I on Section Maps . n._ S._ I have read and understand the above statement and the speCifications (Exhibit B) and fully agree to the term~ of this agreement. T._ ·: · . "_ S._ for development Landowner's Signature Date: Pheasants Forever Authorized Signature Date: #I Inspected by: (Signature) _ I The landowner agrees to repay Pheasants Forever the establishment cost for contracted items listed above that are destroyed prior to the term of projects set forth in this agreement. RECORD OF PAYMENT: Check Years of Contract +-I O\''\ T._ ·· · •••••••• :•.••••••.. _ Program _ Amount Date Inspected: Date: Date Planted: . �17 APPENDIX B SHRUB THICKETS AND SUPPLEMENTAL WINDBREAKS Shrub (plum) thickets are the priority item. Small windbreaks, if planted, must be associated with a thicket and will not be funded if planted alone. Plantings will be eligible for funding only in farmed areas and must be within 0.1 mile of cultivated cropland. Plantings must remain undisturbed for at least 10 years. Maximum Funded: No more than 1 thicket (with or without wind barrier) can be planted per 80 acres. Each thicket must be at least 1/4 mile from another thicket. Size: Shrub thickets must be at least l/lOth acre (4,300 ft2) and no larger than 2/l0ths acre (8,800 ft2) in size, and must include at least 8 rows (excluding windbreak rows). Twelve hundred (1,200) feet is the maximum linear feet of fabric funded per thicket. Supplemental windbreaks, if planted, must be placed on the north and west side of the thicket and must include no more than 900 linear feet total if straight and 1,200 linear feet total if L-shaped. They must include at least 3 rows, one of which must be juniper or cedar. Spacing between the thicket and the windbreak should approximate 100 feet (range 60 ft. minimum; 120 ft. maximum). Payment Rate: Payment will be at $0.58 per linear foot of fabric (6 ft. wide) to the maximums listed above. Changes in rates charged by the Colorado State Forest Service will be used to adjust payment rates. The maximum payment rate for approved private contractors will be $0.64 per linear foot of fabric. Supplemental payment rates/linear foot will be: $0.03 for application of polymer, $0.05 if 8 ft. wide fabric is used, and $0.01 for band application of an approved herbicide along the exterior edge of the fabric to reduce weed competition. Use of fertilizer will not be funded. Planting Dates: Between March 20 and May 15. Pre-Plant Treatment: Sites must be tilled, preferably the fall prior to planting. Tillage must be to bare soil with little residue remaining and must be deep enough to kill existing vegetation. Approved Species: American Plum (priority) and Russian olive (bare root), Rocky Mt. Juniper, E. Red Cedar (potted only). Between-row Spacing: A maximum of 10 feet will be permitted (6 to 8 feet spacing is recommended) for shrub thickets. A maximum of 12 feet will be permitted for wind barriers. In-row Spacing: Mulching: A maximum of 8 feet will be permitted for shrub thickets (6 feet is recommended for plums) and 10 feet will be permitted for evergreens within wind barriers. Woven polypropylene fabric is required for all plantings. Minimum fabric �18 width is 6 ft. (3 ft. on each side of the row). Eight preferred. Cost share is not available for drip systems. PERENNIAL GRASS AND GRASS-LEGUME ft. width is PLANTINGS Switchgrass provides tall cover that stands well over winter and has high value for pheasants. Small unfarmed tracts, currently in short, sodded grasses, are recommended for revegetation to swi tchgrass . Other shorter, cool- season grass -legume mixtures may be used in roadsides where snowdrift is a.problem. This practice is funded only in farmland (not rangeland) settings. Payment Rate: Pre-plant $40.00 per acre as a one-time payment for sites up to 10 acres. For each additional acre (in sites larger than 10 acres [40 acres maximum]) the rate is $35.00. An additional $15.00 per acre will be paid for breaking out sod in heavily sodded sites and supplemental discing prior to planting switchgrass (this does not apply to roadsides) . Preparation: Adequate tillage to completely destroy existing perennial vegetation and to establish a moist, weed-free, firm seed bed is required. Interseeding is not approved. A preemergent herbicide (e.g. atrazine at up to 1 lb/acre) is recommended when planting switchgrass. Planting a tall-sorghum mix (for which payment is available) is recommended the first year. Switchgrass can be seeded into the residual sorghum without tillage during the subsequent spring. Plantin& Procedures: Planting procedures outlined in the Division's Game Information Leaflet #113 should be considered when planting switchgrass. In general, about 20 pure live seeds/ft2 (2 - 3 lbs/acre) should be planted using a drill with double-disk furrow openers, 1-inch depth bands, and packer wheels. If a herbicide is not used, up to 1 1b/acre of an adapted dry 1and alfalfa and up to 1/2 lb of sweet clover should be added. Approved Species: In plots, switchgrass should comprise at least 75% of the live seed (alfalfa and sweet clover are approved additions). Within roadsides, switchgrass is the priority species where snowdrift is not a problem. Other approved warm-season grasses include bluestems and Indian grass. Where these can not be used the tallest wheatgrasses (tall, intermediate, or standard crested) the roadside site will allow, should be used in combination with alfalfa (1 to 2 1bs/acre). P1antin& Dates: Plot Duration: Warm-season Cool-Season grasses including switchgrass: March 15 to May Grass-legume Mixtures: March 15 - July 15 15: Grass and grass-legume plantings must remain ungrazed and undisturbed for at least 7 years. Roadsides should remain unmowed unless essential to reduce snowdrift. If essential, mowing should be delayed until after 1 August and restricted to the road shoulder. �19 Prescribed burning, thinning tillage, or other renovation treatments to rejuvenate grass stands may be applied after 7 years. Grass stands that are relatively thin provide taller, better cover for pheasants. Legumes provide nitrogen and increase growth and quality when added to mixtures. DISTURBANCE TILLAGE AND TALL WILD ANNUALS This practice is primarily designated for 'small unfarmed odd areas within cropland. Wild sunflowers and other tall annuals which attain 3 - 6 ft. height in moderately open stands, stand better through winter than other herbaceous vegetation, and provide excellent cover for broods, protection from blizzards and predators, and supplemental food. This is the most effective and least expensive approach for increasing pheasants and other upland game birds. Maximum Funded: 10 acres/quarter section. Funding Rate: $lS.OO/acre for breaking out sod. $30.00/year in patches 0.1 to O.S acre for subsequent annual disturbance (1/2 acre or larger will be considered 1 acre). $SO.OO/acre/year for sites up to 2 acres. Seeding wild sunflower or other approved wild annuals at 2 to 4 lbs. per acre will be funded at direct seed costs (see seed sources below). $30.00/acre/year will be paid for up to S-acre patches of uncut wheat that contains moderate to dense stands of tall (4 to 6 ft.) annual weeds. Plot Dimensions: Short, relatively wide patches, which will not be easily inundated by drifting snow, are preferred. Placement: Adjacent to woody cover when possible. Draw bottoms that already contain weeds and above average moisture are ideal. Sites containing noxious perennials should be avoided. Specifications: Initial tillage with a disk plow or mold-board plow is needed in sites containing perennial grass to destroy all perennial cover, preferably, immediately after the ground has thawed in early March. Large clod size is preferred to retain thin stands of annual forbs. Initial tillage in subsequent years should be conducted prior to May 1. A second thinning tillage may be used prior to the 1st of June. Spring tillage is needed each year to retain tall annuals. Annual grasses usually dominate if tillage is not used each spring. Wild sunflowers and annual ragweeds can be drilled or broadcast and harrowed at low rates to help establish tall annuals, if they are not already present. Known sources in Colorado include the Arkansas Valley Seed Company - Denver & Longmont, and Sharpe Bros. Seed Company - Greeley. Retention: Tall annuals must remain undisturbed through March of the following year. Sites can then be prepared for the next year's growth. �20 ANNUAL SURVIVAL PLANTINGS - SORGHUMS APPLICATION: On CRP, Annual Set Aside, and other cropland or tilled wasteland. When applied within CRP fields, SCS specifications for CP-12 must be used (see supplement). MAXIMUM FUNDED: 1 plot/80-acre field, 2 plots/160 acres. Plots must be at least 1/4 mile apart. PAYMENT RATE: $40.00/acre/year for 1-5 acres (7 acres within center pivot corners) and $25.00/acre/year for additional acreages in tracts larger than 5 acres (12 acre maximum). Payment will increase by $5.00/acre for application of 30 lbs of nitrogen/acre. An additional one-time payment of $15.00 per acre will be paid for breaking out sod in CRP or heavily sodded sites and supplemental discing prior to planting. PLACEMENT: Plots should be placed within or near cropland and placed crosswise to prevailing winds. SPECIFICATIONS: Initial treatment: Adequate tillage to destroy existing perennial vegetation in early spring prior to annual growth. Subsequent years: Preferably minimum tillage shredding of old materials as needed prior to April 25. Annual application of nitrogen at 40 lbs./ac. is recommended. Plot Dimensions: Minimum total plot width shall be 150 feet. Wider strips are preferred to reduce impacts of drifting snow. (See special dimensions on CRP). Row Spacing: 18 to 36 inches. Seed Specifications: At least 50% (75% preferred) of an adapted tall forage sorghum that will stand well with minimal lodging and will mature before frost. Up to 40% can be adapted varieties of grain sorghum. These can be mixed or planted in separate rows (i.e., 2 rows of grain sorghum to 6 rows of forage sorghum. These sorghums should equal a minimum of 75% of the total weight. Maximum amounts for other grains include: Dryland corn (25%), sunflowers (10%) and proso millet (10%). Addition of 1 to 2 lbs./ac. of wild sunflower seed is recommended (Source: Arkansas Valley Seed Company - Denver). Planting: Between April 25 and June 05; Late April to mid-May recommended. Sorghums should be planted at 4-8 lbs./acre. PLOT DURATION: One year. Sorghum plantings must remain undisturbed through March of the following year. �21 SUPPLEMENT FOR SORGHUM PLANTINGS WITHIN CRP SCS Notification: The CRP contract must be amended at the local SCS office prior to implementing CP-12 and breaking out food plots within CRP. This requires filling out a one-page form at your SCS office. The ASCS must be advised of the change for their records. Once a winter cover-food plot is broken out within CRP it must remain as such until the end of the CRP contract. Payments will be made annually based on seeded acres. If the farmer wishes to discontinue this practice he must reestablish grass (required by the ASCS). Reimbursement will be at $40.00/acre to cover reseeding grass. Maximum Size: The maximum size is 3 acres per site. CRP fields must contain at least 40 acres to be eligible for a CP-12 food plot. Plot Dimensions: Plantings must be between 66 and 99 feet wide, therefore, plots should be laid out as 2 or 3 (66 to 99 ft. wide) adjacent strips with a 30 ft. buffer of untilled grass between strips to attain the 150 ft. minimum width. For example, a plot 99 ft wide x 440 ft. long equals 1 acre. Three adjacent plots will equal the maximum of 3 acres/site. Three 66 x 660 ft. plots or two 99 x 660 ft strips will also yield 3 acres. Placement: Usually within the southeast corner when in CRP, preferably within 50-100 yds of edge, but location can vary depending on soil, wind, and moisture, and location of other winter covers if they occur. Sorghum plantings are not permitted in soils containing free lime (shows effervescence), or soils that are deep sands or choppy sands. POST-HARVEST RETENTION OF TALL WHEAT STUBBLE APPLICATION: Primarily in northeastern Colorado where stubble is usually left standing over winter. The objective is to provide taller, more secure cover for night roosting, feeding, loafing, and escape by pheasants through summer, fall, and winter. This practice must be applied during wheat harvest by raising the combine header in selected locations within the wheat field. PAYMENT RATE: $30.00 per acre with 2 acres maximum per site. MAXIMUM FUNDED: One site per 80 acres and 2 per 160 acres of small grain stubble. SPECIFICATIONS: Retention of at least 18 inches of wheat, rye, triticale, or barley stubble during harvest. Harvest of heads is essential to prevent lodging of stubble under snow. �22 PLACEMENT: Patches of tall stubble should be near corn, sorghum, or other cropland preferably within the southeast part of the stubble fields. RETENTION: This treatment should not be applied unless the entire wheat stubble field is to be left undisturbed through the subsequent fall and winter. Tillage, if used, should not be initiated until after April 5. SPRING NO-TILL RETENTION OF WHEAT STUBBLE This practice will not be funded in 1993 because of the extremely poor quality wheat stubble remaining after the 1992 wheat harvest. SUPPLEMENTAL PAYMENTS FOR CUSTOM SITE PREPARATION PURPOSE: To prepare planting sites when the landowner does not have the proper equipment or does not have time to prepare the site. TREATMENT: Breaking out small tracts within CRP or sodded waste areas with a moldboard plow or heavy discing to completely destroy existing vegetation for reseeding to switchgrass or planting sorghum patches. Tillage must be to a depth of at least 6 inches. PAYMENT RATE: Option 1: Option 2: Mold-board Plowing: Payment will be $16.00/acre Primary heavy tillage with a tandem or offset disc. - $lO.OO/acre. Replicate discing - $8.00/acre/treatment. Equipment transportation to and between small tracts. $15.00/hour. Supplemental payment will be SPRING NO-TILL RETENTION OF WHEAT STUBBLE Purpose: To provide secure stubble for nesting pheasants through spring using either reduced tillage or no-tillage fallow methods. PAYMENT RATE: $lO.OO/acre for up to 20 acres/quarter section. MAXIMUM FUNDED: A maximum of 20 landowner/year. SPECIFICATIONS: Stubble must have remained undisturbed since harvest, must remain untilled until after July I, and must average at least 12 inches in height to qualify. If the remainder of the stubble field is to be conventionally fallowed, it must be tilled to remove >90% of the standing stubble prior to May 1. The retained tract must remain undisturbed, other than application of herbicides. Only herbicides acres/quarter section, 4 tracts/section and �23 of low toxicity that are approved for chemical fallow in wheat stubble may be used to control weeds and volunteer wheat (Roundup, Landmaster, Cyclone, etc.). Restricted use herbicides such as Paraquat may not be used. Dense annual weeds which exceed 12 inches within stubble fields may be substituted for wheat stubble. PlACEMENT: Adjacent to green wheat surrounded by fallow. PLOTDURATION: Until after July 1. or other cropland. It should not be �37 JOB PROGRESS REPORT State of: Project: Colorado W-167-R Work Plan: Job Title: 12 Job _!2__ Effects of Mycoplasma Infection on Reproductive Performance and Survival of Merriam's Wild Turkeys Period Covered: Author: Upland Bird Research 01 January through 31 December 1992 Richard W. Hoffman Personnel: Clait E. Braun, Amanda Clements, Renzo Del Piccolo, John H. Ellenberger, Van K. Graham, John P. Gray, Anthony W. Hoag, Richard W. Hoffman, Robert T. Magill, and Walter J. Miller, Colorado Division Wildlife; William R. Davidson and Page M. Luttrell, Southeastern Cooperative Disease Study. ABSTRACT Fifty-three Merriam's wild turkeys (Meleagris gallopavo merriami), including 20 adult and 33 subadult females, were surveyed by serologic and cultural methods for evidence of Mycoplasma gallisepticum (MG), M. synoviae (MS) , M. meleagridis (MM), and M. gallopavonis (MGp). Their reproductive performance and survival were subsequently monitored using radio-telemetry techniques. No clinical signs of Mycoplasma infection were evident in any of the birds examined. The majority of birds showed no rapid plate agglutination (RPA) reactions to laboratory (MG - 92%, MM - 89%, MS - 58%) or commercially prepared (MG - 72%, MS - 40%) antigens. All reactions to MM and most reactions to MG and MS were classified as weak responses. Hemagglutination inhibition{HJ:) tests were uniformly negative. MGp, but not MG, MS, or MM, was isolated from 8 of 9 tracheal cultures. RPA tests can only be ·interpreted as suspicious of MS infection because of the lack of confirmation by HI tests and cultures. In contrast, 16 chickens living in close proximity to the wild turkeys showed strong RPA reactions to MS.that were supported by positive (titers ~ 40) HI results. The lack of clinical disease in the chickens was considered characteristic of a carrier state. Clutch size for first and second nest attempts, nesting success, hatching success, and fertility did not differ between seropositive and seronegative wild turkeys. If the RPA positive reactions in the wild turkeys were due to MS infection, that there was no overt disease, widespread infection, or suppressed reproductive output suggests the infection was old and.no longer active or the organisms involved were of low virulence and infectivity. ��39 EFFECTS OF MYCOPLASMA INFECTION ON REPRODUCTIVE PERFORMANCE AND SURVIVAL OF MERRIAM'S WILD TURKEYS Richard W. Hoffman INTRODUCTION Restoration of the wild turkey has been one of the most noteworthy successes of the wildlife management profession (Kennamer and Kennamer 1990). However, only recently have wildlife managers realized the potential risks of disease introduction and dissemination associated with this practice (Nettles and Thorne 1982, Nettles 1984, Amundson 1985). Although disease problems have been linked to wild turkey restoration efforts using wild-trapped stock (Davidson 1987), without data to prove otherwise, such efforts could be erroneously blamed for disease outbreaks in other wildlife populations or more importantly, for unrelated disease problems in the domestic poultry industry (Amundson 1985). no Mycoplasmosis is the disease that has prompted these concerns. Three species of Mycoplasma are known to be pathogenic for domestic poultry including M. gallisepticum (MG), M. meleagridis (MM), and M. synoviae (MS). MG is the most detrimental, producing sinusitis, reduced egg production, decreased hatching success, and poor juvenile growth; however, it seldom causes direct mortality (Yoder 1984). Mycoplasma infection can produce chronic carriers and may be transmitted through the respiratory tract, venereally via semen, or transovarially through the oviduct to the egg. Prior to 1980 there were few reports of mycoplasmosis in either wild or semiwild, free ranging turkeys. Trainer (1973) first reported the isolation of Mycoplasma from wild turkeys captured in Texas and Wisconsin. The isolates were not identified nor associated with any disease. Hensley and Cain (1979) conducted a serologic survey of wild turkeys in Texas and detected antibodies to MG in counties supporting commercial poultry operations; isolation was not attempted. In 1980, antibodies to MG and MM were detected in wild turkeys trapped in Missouri for release in Wisconsin; further serologic testing from 1980 to 1984 disclosed the presence of Mycoplasma seropositive birds in Wisconsin, Minnesota, and Missouri (Amundson 1985). Clinical MG infection was also found in small flocks of free ranging, wild-type turkeys in California (Jessup et al. 1983) and Georgia (Davidson et al. 1982) that were associating with domestic fowl. More recently, antibodies to MG, MM, and MS have been detected in flocks from Texas, New Mexico, Arizona, Colorado, Oklahoma, and North Dakota (Rocke and Yuill 1987, Fritz et al. 1992). Isolation attempts for serotypes which included these species were mostly unsuccessful, but numerous isolates were serotyped as M. gallopavonis (MGp). Antibody to MS was also detected in 85 of 94 wild turkeys transported from Arizona to Idaho in 1984; all were condemned by the Idaho Department of Agriculture (Anon. 1985g). Birds from the same population in Arizona were released in Utah without any prior testing (J. A. Roberson, Utah Div. Wildl. Resour., pers. commun.). Extensive disease �40 monitoring in the southeastern United States suggests wild turkeys in that region are not important in the epizootiology of MG, MS, or MM (Davidson et al. 1982, 1985, 1988). In Colorado, evidence of MG, MM, and MS infections was found in declining populations of wild turkeys on the Uncompahgre Plateau and Devil's Creek State Wildlife Area; conversely, serological tests on wild turkeys trapped from stable populations near Trinidad and Pueblo were negative (Adrian 1984). MG was isolated from the trachea of 1 bird from Devil's Creek and MM was isolated from the oviduct of another bird captured on the Uncompahgre Plateau. Mycoplasma was cultured from several other birds, but the serotype was unknown. Whereas mycoplasmosis was implicated in the decline of wild turkeys on the Uncompahgre Plateau (Adrian 1984), there was only indirect evidence to support this contention. Rocke and Yuill (1987) found no evidence to link mycoplasmosis with the decline of wild turkeys on the Welder Wildlife Refuge in Texas. Although free-ranging wild turkeys are susceptible to Mycoplasma, the consequences of infection (i.e., transmission between wild and domestic birds, pathologic effects, persistence, and long-term population impacts) are unclear. Experimental MG infections in captive-reared wild turkeys reduced production, fertility, and hatching success of eggs compared to noninfected controls (Rocke et al. 1988). Egg production was unaffected by MG infection 2 years postinoculation, but fertility remained low both years. No mortality occurred in either the infected or control groups. Currently, there is only circumstantial ev,idence to support the contention that wild turkey populations harbor or perpetuate MG, MM, or MS, or that such infections result from contact with domestic fowl and suppress wild populations through subtle changes in reproductive performance. Despite the uncertainly surrounding the Mycoplasma issue, the benefits of testing not only for Mycoplasma spp., but also for other avian pathogens, outweigh the negative aspects (Davidson 1987). The Wildlife Disease Association (YDA) has prepared an advisory statement on disease monitoring of wild turkeys including suggested guidelines for conducting disease testing (YDA 1985). These guidelines have been endorsed by the International Association of Fish and Wildlife Agencies (Nettles 1984) and The United States Animal Health Association (Nettles and Thorne 1982). P. N. OBJECTIVES Compare survival and reproductive performance (i.e., egg production, fertility, nesting success, and hatching success) of free-ranging, female wild turkeys that are seropositive for Mycoplasma·with those that are negative. SEGMENT OBJECTIVES 1. Trap and collect blood samples from 50 wild turkeys known to be associating with domestic fowl. 2. Collect blood samples with wild turkeys. from 15 domestic chickens known to be in contact �41 3. Test blood samples for antibodies to 3 pathogenic mycoplasmas: Mycoplasma gallisepticum, M. synoviae, and M. meleagridis. 4. Obtain tracheal swaps from 10 wild turkeys and culture for Mycoplasma gallopavonis. 5. Attach radio transmitters to a sample of seropositive and seronegative wild turkeys and monitor their reproductive performance. 6. Compile and analyze data, and prepare progress report. STUDY AREA Trapping was confined to the Hittle Ranch approximately 6 km northeast of Collbran, Colorado in Mesa County. From here, radio-marked birds ranged over 120 km2 of surrounding areas during the breeding and brood rearing periods. The primary areas where turkeys were found included Hawxhurst Creek, Buzzard Creek, Salt Creek, Kimball Creek, Smalley Gulch, and Big Creek. These north/south oriented drainages were dominated by narrowleaf cottonwood (Populus ansustifolia) along the bottoms giving way to Gambel oak (Quercus gambelii) interspersed with pinyon pine-juniper (Pinus edulis-Juniperus spp.) woodlands on the drier slopes. Most of the mesas were cleared and planted to hay meadows. Livestock were pastured on the meadows during winter. About 85% of the area was privately owned. Turkeys first appeared at the Hittle Ranch about 25 years ago. Over the past 5 years, approximately 120-140 wild turkeys have wintered at the ranch. Although not intentionally fed, the turkeys consume large quantities of artificial foods such as oat hay, alfalfa hay, beef feed, and poultry feed provided for the domestic livestock. At least 30 chickens roam freely about the ranch; others (50+) are confined to a holding pen. Wild turkeys were frequently observed feeding with the free-roaming chickens and commonly scratched around the holding pen for spilled poultry feed. METHODS Trapping, Marking, and Radio-tracking Turkeys were baited with oat hay and corn, and live-trapped with cannon nets or Clover traps during February and March 1992. Captured birds were weighed on an electronic scale, classified to age and sex, and banded with serially numbered aluminum leg bands. Allflex livestock eartags were attached to the patagium. Ages were recorded as subadult (8-10 months) or adult (>18 months). Fifty-three females were equipped with lithium battery-powered transmitters (model HLPB-2l50-LD, Wildlife Materials, Carbondale, IL) attached with a poncho collar (Amstrup 1980). The radio package weighed < 40 g. Tracking was conducted from the ground using a 3-element Yagi antenna and Telonics TR-2 receiver with a TS-l scanner attachment. All locations were verified by visual observation and recorded to the nearest 50 m as Universal Transverse Mercator coordinates. One aerial search was conducted in late April to locate birds not found during ground searches. Birds found during the aerial search were subsequently located from the ground. �42 Disease Monitoring All birds were examined for clinical signs of mycoplasma infection and their general physical condition was noted. Blood (8-10 cc) was collected by jugular venipuncture from each radio-marked turkey and from the wing vein of 8 free-roaming and 8 confined chickens. The plasma was separated by centrifugation, pipetted into a separate container, and mailed overnight express to the Southeastern Cooperative Wildlife Disease Study (SCDS), College of Veterinary Medicine, University of Georgia, Athens. Tracheal swaps were obtained from 9 wild turkeys, placed into Frey's media with swine serum, and mailed overnight express to SCDS. Rapid plate agglutination (RPA) tests were performed on plasma from each bird. Two plate antigens were used for testing for MG and MS, a commercial antigen prepared by Salsbury Laboratories, Inc. (Charles City, IA) and a laboratoryprepared antigen made by the Poultry Disease Research Center (PDRC), College of Veterinary Medicine, University of Georgia, Athens. One plate antigen (PDRC) was used for MM testing. Agglutination was scored on a 0 to 4 scale, with 0 being a negative reaction, 1 and 2 a weak reaction, and 3 and 4 a strong reaction. When the 2 antigens produced different results, those from the antigen yielding the stronger reaction were used. Appropriate control sera were used at intervals throughout the sampling period. Plasma samples also were tested for antibodies to MG, MS, and MM with the hemagglutination inhibition (HI) test using a laboratory-prepared antigen from PDRC. The test was performed as described in the National Poultry Improvement Plan (Anon, 1985Q). HI titers ~ 40 were considered positive. Sixteen sera samples showing some reactivity to MG or MS on RPA tests were sent to North Carolina State University, Raleigh for immunoplot testing for antibodies to MG and MS (Avakian et al. 1992). A subsample (n - 6) of these sera was tested for MGp. A second set of 8 sera samples that showed no reactivity to MG, MS, and MM by RPA and HI tests was also sent to North Carolina State University for additional immunoblotting tests. Reproductive Parameters During late April and May, hens were located once every 2-3 days to ascertain if they were nesting. Suspected nest sites were circled and flagged from> 20 m away. Some nests were visually observable from this distance. Others were monitored but not approached for 30 days unless the radio signal indicated the hen was gone. Nest sites were visited daily as the anticipated hatch date approached. Most hens were located often enough just before and during the early stages of incubation to approximate within 2 days of when they started incubating. For successful nests (~ 1 egg hatched), onset of incubation was estimated by backdating 28 days (incubation period) from the date of hatch. Clutch size, fertility (unhatched eggs were broken open and examined for developing embryos), nest success (X hens that hatched ~ 1 egg), and hatching success (X eggs in successful nests that hatched) were determined from egg shell characteristics after the eggs hatched or after the nest was abandoned or depredated. Clutch size was also determined by visiting nest sites when the hen was away from the nest. �43 RESULTS Capture and Marking f Sixty-eight turkeys, including 5 adult males, 6 subadu1t males, 24 adult females, and 33 subadult females were captured with Clover traps (8) or cannon nets (60); 20 adult and 33 subadu1t females were weighed (Table 1), banded, bled, equipped with radio transmitters, and released at the trap site. One adult male died as a result of trapping. Two adult females and two subadult males were banded and held in captivity for 8 weeks until they recovered from severe feather loss due to trapping. One of these males subsequently died in captivity. The other 3 birds were released near the trap site. Seven birds (1 adult male, 4 subadult males, and 2 adult females) were transplanted to another site. Three adult males were banded and released at the trap site. Recoveries and Mortalities Seven radios were recovered between 18 February (earliest trap date) and 31 July (termination of field work), 5 from subadults and 2 from adults. Radiocontact was lost with one adult and one subadult hen prior to nesting and one adult hen after she lost her first nest. One radio was recovered in early April from an adult hen that slipped the poncho collar. Another radio was recovered in early July from a road-killed subadult hen. Five recoveries (2 adults, 3 subadults) were classified as natural mortalities, including 2 hens killed prior to nesting, one killed on the nest, and 2 killed after having successfully nested. Table 1. Weights (kg) of female wild turkeys captured for serological monitoring of Mycoplasma spp., Collbran, Colorado, 1992. Descriptive statistic n ~ SD Range Adult 20 4.94 0.33 4.35-5.92 Subadult 33 4.19 0.33 3.42-4.88 Disease Monitoring All birds captured appeared to be in good physical condition. None had clinical signs of Mycoplasma infection. The majority of birds showed no RPA reaction to MG (PDRe - 92%, Salsbury - 72%), MS (PDRe ~ 58%, Salsbury - 40%), or MM (PDRe - 89%) (Table 2). There were no differences of biological significance between reactions of adults and subadults. All reactions to MM, and most reactions to MG (PDRe = 100%, Salsbury = 80%) and MS (PDRe = 73%, Salsbury - 71%) were weak. Only 2 of 8 birds with strong reactions to MS also had strong reactions to MG. Results for HI testing were uniformly negative for MG and MS. There were two MM titers of 20, but these were not considered positive reactions. �44 Thirteen of 16 sera samples showing some reactivity to MG and/or MS on RPA tests had bands considered positive (MG - 4, MS - 0) or suspicious (MG ~ 9, MS - 3) based on the immunoblot test. Three samples had bands suspicious of MG and MS. Four samples had bands corresponding with the proteins p64 and p56, which have been identified as being species-specific for MG (Avakian et al. 1992). There was no evidence of antibodies to MM. The 6 samples tested for MGp were positive. Three of 8 sera samples showing no reactivity on RPA tests reacted positively on the immunoblot tests. Table 2. Seropreva1ence of antibody to-Mycoplasma ga11isepticum (MG) , M. synoviae (MS) , and M. me1eagridis (MM) among female wild turkeys as determined by rapid plate agglutination tests (RPA) , Collbran, Colorado, 1992. MS MG Score· RPA-P' RPA-2 RPA-l RPA-2 MM RPA-l 6 13 1 18 2 0 16 10 7 29 4 0 22 20 8 47 6 0 ADULTS 0 1-2 3-4 19 1 0 15 5 0 13 5 2 SUBADULTS 0 1-2 3-4 30 3 0 23 7 3 18 11 4 ADULTS AND SUBADULTS 0 1-2 3-4 49 4 0 38 12 3 31 16 6 ·0 - no reaction, 1-2 - weak reaction, 3-4 - strong reaction. ~A-l - Poultry Disease Research Center prepared antigen, RPA-2 commercially prepared Salsbury antigen. Mycoplasma spp. organisms were isolated from 8 of 9 tracheal cultures. All were negative for pathogenic mycoplasmas by fluorescent antibody tests, whereas, the nonpathogenic MGp was identified from each of the 8 samples. Seven eggs from 2 hens with strong MS reactions (one hen also had a strong MG reaction) were cultured for mycoplasmas, but no isolations were made. Sixteen chickens tested by RPA showed strong reactions to MS (PORC - 87%, Salsbury - 19%) and to a lesser extent, MG (PORC = 19%, Salsbury - 31%) (Table 3). HI results supported the RPA tests for MS but not MG (Table 3). HI titers ~ 40 were found for 13 of 16 chickens tested for MS. �Table 3. Seroprevalence of antibody to Mycoplasma gallisepticum and M. synoviae among domestic chickens as determined by rapid plate agglutination (RPA) and hemagglutination inhibition (HI) tests, Collbran, Colorado, 1992. RPA-lo oc 0 0 1 4 0 0 3 1 0 0 3 0 0 0 0 MG RPA-2 3 2 3 1 4 2 1 3 2 1 0 3 2 0 2 0 HI" 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 RPA-l 4 4 4 4 4 4 4 4 4 4 4 4 4 0 3 1 MS RPA-2 2 0 0 0 4 4 0 2 0 0 1 3 0 0 0 0 HI 40 80 20 80 80 40 40 80 20 80 80 160 40 0 40 40 °RPA-l Poultry Disease Research Center prepared antigen, RPA-2 commercially prepared Salsbury antigen. ~I titers ~ 40 are considered positive. cO - no reaction, 1-2 - weak reaction, 3-4 - strong reaction. Reproductive Performance Nineteen adult hens were monitored into the nesting period of which 12 (63%) nested successfully and 7 attempted to nest but failed. Six of the 7 unsuccessful hens renested and 3 were successful. Nesting success for first and second nest attempts was 79%; 146 of 164 eggs (89%) in successful nests hatched. In comparison, 10 of 27 (37%) subadu1t hens followed into the nesting period nested successfully and 17 attempted to nest but failed, including 4 hens that abandoned their nests after being disturbed. Six of 17 unsuccessful subadults renested and 2 were successful. Nesting and hatching success of subadults for first and second nest attempts was 44 and 90% (122/136 eggs hatched), respectively. Only 4.3% (n - 13) of the 300 eggs laid in successful nests were infertile. Clutch size for first (f - 0.34) and second (f - 0.69) attempts did not differ between age classes (Table 4). Both adult and subadult hens laid more eggs (f - 0.17 for adults, f - 0.003 for subadu1ts) on first than second nest attempts. Adults initiated incubation approximately one week earlier (f < 0.001) than subadults (Table 5). An average of 25.6 ± 5.5 days (n = 10) elapsed between when a hen lost its first nest and when it started incubating a second clutch. Subadults (24.2 ± 4.7 days, n -5) recycled slightly faster (l - 0.46) than adults (27.0 ± 6.3 days, n -5). Onset of incubation for �46 renest attempts ranged from 23 May to 15 June (median - 10 June) for adults and 25 May to 11 June (median - 6 June) for subadults. Clutch size for first (l- 0.50) and second (l- 0.25) nest attempts, nesting success (l- 0.79), hatching success (f - 0.80), and fertility (f ~ 0.84) did not differ between serologically positive and serologically negative birds (Table 6). Five of the 10 birds that renested and 1 of 2 birds that presumably did not attempt to nest were serologically positive. Table 4. Clutch size for first and second nest attempts of adult and subadult wild turkeys, Collbran, Colorado, 1992. Descriptive statistic First attem~t Adult Subadult 11 12.3 1.2 10-14 n Z. SD Range 19 11.7 1.4 10-15 Second attem~t Subadult Adult 4 9.7 2.2 7-12 6 9.2 2.2 5-11 Table 5. Chronologie distribution for onset of incubation for first nest attempts by adult and subadult wild turkeys, Collbran, Colorado. Time Period 26-30 April 1-5 May 6-10 May 11-15 May 16-20 May Median Range Adults % N 10 5 1 2 56 28 5 11 30 April 26 April-15 May Subadults % N 1 8 8 7 2 4 31 31 27 7 9 May 29 April-20 May �47 'Table 6. Comparative reproductive parameters of serologically positive score ~ 2) and negative wild turkeys, Collbran, Colorado, 1992. Parameter Positive Clutch size (X ± SD) First nest Second nest Nesting Success, % Hatching Success, % Fertility, % 11.6 ± 1.6 8.6 ± 2.1 71 91 94 (MG/MS Negative 12.0 ± 1.2 10.2 ± 1.3 56 88 97 Movements Winter (trap site) to breeding area (nest site) movements averaged 3.3 ± 1.8 km and did not differ (f - 0.43) between adult and subadult hens. The median distance between first and second nest sites was 971 m (range, 255-4793 m). Nests occurred in all areas where turkeys were observed during winter in addition to areas at higher elevations where turkeys did not occur during winter. DISCUSSION Much of the information for interpreting serologic tests performed on wild turkeys is still based on experimental testing of domestic birds. Therefore, caution must be applied in interpreting the results. In this study, RPA test results can only be interpreted as suspicious of MS infection because of the lack of confirmation by HI tests and cultures. The HI results are not surprising since HI activity in wild turkeys has been shown to decline to low or negative titers within a few months postexposure (Rocke et a1. 1985, Rocke and Yuill 1988). Weak RPA reactions in conjunction with negative HI titers are not indicative of exposure based on accepted test interpretations. If the RPA positive reactions were due to MS infection, that there was no evidence of overt disease, widespread infection, or suppressed reproductive output, suggests the infection was old and no longer active or the organisms involved were of low virulence and infectivity. Further, the optimal conditions (i.e., easy access to unlimited food supplies) under which these birds lived may have prevented any clinical manifestation of Mycoplasma infection. Experimental MG infections in captive-reared wild turkeys reduced egg production, fertility, and hatching success (Rocke et al. 1988). These effects were not apparent in the Collbran population. Reactions Salsbury (Avakian the PDRC reactions handling to MG and MS were greater for the Salsbury than the PDRC antigen. antigens are acknowledged to be highly sensitive with low specificity et al. 1988) and, therefore, may have given more false positives than antigen. There is also the possibility that some of the weak were due to non-specific agglutination resulting from improper of the serum (Rocke et al. 1985). �48 Immunoblots are generally regarded as more sensitive and usually more specific than RPA or HI tests. This may explain why several sera samples that were MG negative by RPA and HI tests reacted to MG specific proteins on the immunoblot tests. However, this does not explain why some of the sera samples that were RPA positive, especially for MS, showed negative reactions on the immunoblot tests. The immunoblot tests, like the RPA tests, are only suggestive of exposure. Their interpretation for free-ranging wild turkeys requires further research. The MG and MS responses on the RPA tests may have been the result of crossreactions with MGp, which has been identified as a common, nonpathogenic mycoplasma in free-ranging wild turkeys (Cobb et al. 1992, Fritz et al. 1992, Luttrell et al. 1992, this study). However, the immunoblot tests did not support this hypothesis. The responses to MG and MGp were clearly distinct (D. H. Ley, College of Veterinary Medicine, North Carolina State Univ., pers. commun.). There remains the possibility the reactions,were due to some unknown or undetected Mycoplasma spp. In contrast to the test results for wild turkeys, RPA and HI tests for chickens were highly suggestive of MS infection. Since MG and MS share some antigens, RPA tests are often positive for both organisms even when the birds are only infected with one of them. This may explain the positive RPA reactions to MG that were not substantiated with positive HI results. The strong seroprevalence of MS infection in chickens but not in wild turkeys was unexpected. However, a similar finding involving MG infection was reported for wild turkeys and chickens living in close association on Cumberland Island, Georgia (Luttrell et al. 1991). MG infection had been confirmed by culture and serology in this population in 1980 (Davidson et al. 1982). Despite the continued association between wild turkeys and domestic chickens on the island, surveys in 1988 indicated the MG infection had not persisted in the turkeys even though the chickens still tested strongly positive. The lack of clinical disease in the chickens, low rate of spread, low virulence, and high RPA reactivity were considered characteristic of a carrier state. The MG organism was presumed localized in the lung and reproductive tissues of the chickens from where it could not be readily transmitted to other birds. LITERATURE CITED Adrian, W. J. 1984. Investigation of disease as a limiting factor in wild turkey populations. Ph.D. Diss., Colorado State Univ., Fort Collins. 63pp. Amstrup, S. C. 1980. 44:214-217. A radio-collar for game birds. J. Wildl. Manage. Amundson, T. E. 1985. Health management in wild turkey restoration programs. Proc. Natl. Wild Turkey Symp. 5:285-294. Anonymous. 1985g. Idaho turkey relocations halted. Anonymous. 1985h. National poultry improvement plan. Publ. 91-40 (147.7), Washington, D. C. 82pp. Turkitat 3(1):4. U. S. Dep. Agric. �49 Avakian, A. P., D. H. Kleven, and J. R. Glisson. 1988. Evaluation of the specificity and sensitivity of two commercial enzyme-linked immunosorbent assay kits, the serum plate agglutination test, and the hemagglutination-inhibition test for antibodies formed in response to Mycoplasma gallisepticum. Avian Dis. 32:262-272. ____________ , D. H. Ley, and M. A. T. McBride. 1992. Humoral immune response of turkeys to strain S6 and a variant Mycoplasma gallisepticurn studied by immunoblotting. Avian Dis. 36:69-77. Cobb, D. T., D. H. Ley, and P. D. Doerr;- 1992. Isolation of Mycoplasma gallopavonis from free-ranging wild turkeys in coastal North Carolina seropositive and culture-negative for Mycoplasma gallisepticurn. J. Wildl. Dis. 28:105-109. Davidson, W. R. programs. 1987. Disease monitoring in wild turkey restoration Proc. West. Assoc. Fish and Wildl. Agencies 67:113-118. ____________ , V. F. Nettles, C. E. Couvillion, and H. W. Yoder, Jr. 1982. Infectious sinusitis in turkeys. Avian Dis. 26:402-405. ____________ , and E. W. Howerth. 1985. Diseases diagnosed in wild turkeys (Meleagris gallopavo) of the southeastern United States. J. Wildl. Dis. 21:386-390. ____________ , H. W. Yoder, M. Brugh, and V. F. Nettles. 1988. Serological monitoring of eastern wild turkeys for antibodies to Mycoplasma spp. and avian influenza viruses. J. Wildl. Dis. 24:348-351. Fritz, B. A., C. B. Thomas, and T. M. Yuill. 1992. Serological and microbial survey of Mycoplasma gallisepticum in wild turkeys (Meleagris gallopavo) in six western states. J. Wildl. Dis. 28:10-20. Hensley, T. S., and J. R. Cain. 1979. Prevalence of certain antibodies to selected disease causing agents in wild turkeys in Texas. Avian Dis. 23:62-69. Jessup, D. A., A. J. Damassa, R. Lewis, and K. R. Jones. 1983. Mycoplasma gallisepticurn infection in wild-type turkeys living in close contact with domestic fowl. J. Am. Vet. Med. Assoc. 183:1245-1247. Kennamer, J. E., and M. C. Kennamer. 1990. Current status and distribution of the wild turkey, 1989. Proc. Natl. Wild Turkey Sypm. 6:1-12. Luttrell, M. P., S. H. Kleven, and W. R. Davidson. 1991. An investigation of the persistence of Mycoplasma gallisepticum in an eastern population of wild turkeys. J. Wildl. Dis. 27:74-80. ____________ , T. H. Eleazer, and S. H. Kleven. 1992. Mycoplasma gallopavonis in eastern wild turkeys. J. Wild1. Dis. 28:288-291. Nettles, V. F. 1984. Report of the fish and wildlife health committee. Proc. IntI. Assoc. Fish and Wi1dl. Agencies 74:89-101. �50 ______ , and E. T. Thorne. 1982. Annual report of the wildlife disease committee. Proc. United States Animal Health Assoc. 86:64-65. Rocke, T. E., and T. M. Yuill. 1987. Microbial infections in a declining wild turkey population in Texas. J. Wi1d1. Manage. 51:778-782. ______ , and 1988. Serologic responses of Rio Grande Wild Turkeys to experimental infections of Mycoplasma gallisepticum. J. Wild1. Dis. 24:668-671. ______ , and T. E. Amundson. 1985. Evaluation of serologic tests for Mycoplasma gal1isepticum in wild turkeys. J. Wild1. Dis. 21:58-61. ______ , and 1988. Experimental Mycoplasma ga11isepticum infections in captive-reared wild turkeys. J. Wild1. Dis. 24:528-532. Trainer, D. O. 1973. Some diseases of wild turkeys from Texas and Wisconsin. Pages 160-173 in G. C. Sanderson and H. C. Schultz, eds. Wild turkey management: current problems and programs. Univ. Missouri Press, Columbia. Wildlife Disease Association. 1985. Advisory statement on disease monitoring in wild turkeys. Wi1dl. Dis. Newsletter 21:1-3 Yoder, H. W., Jr. 1984. Mycoplasma ga11isepticum infection. Pages 190-202 in M. S. Hofstad, H. J. Barnes, B. W. Calnek, W. M. Reid, and H. W. Yoder, eds. Diseases of poultry, 8th ed. Iowa State Univ. Press, Ames. �51 JOB PROGRESS REPORT State of: Colorado Project: W-167-R Work Plan: Job Title: 13 Job: Movements. Plains sharp-tailed Period Covered: Author: Upland Bird Research 10 Reproductive Success. and Habitat Use by Introduced Grouse 01 January through 31 December 1992 Kenneth M. Giesen Personnel: Clait E. Braun, Kenneth M. Giesen, Division of Wildlife and David B. laBelle, Colorado ABSTRACT Surveys in southeastern Wyoming resulted in 18 dancing grounds being located as potential sites for obtaining plains sharp-tailed grouse (Tympanuchus phasianellus jamesi) for transplant into Colorado. Most leks observed had 1020 males present. The transplant was postponed until 1993 because permission to release grouse at the primary and secondary sites in Colorado was not obtained. Surveys of a pioneering population of plains sharp-tailed grouse near the Tamarack State Wildlife Area resulted in 24 male plains sharp-tailed grouse or sharp-tailed grouse X greater prairie-chicken (I. cupido) hybrids being observed at 7 distinct leks. At least 13 of 24 sharp-tailed grouse observed and 6 of 12 sharp-tailed grouse captured showed plumage characteristics of hybridization. Home ranges estimates for 3 hybrids and 4 sharp-tailed grouse ranged from 1.02 to 4.73 km2• ��53 MOVEMENTS, REPRODUCTIVE SUCCESS, AND HABITAT USE BY INTRODUCED PLAINS SHARP-TAILED GROUSE Kenneth M. Giesen INTRODUCTION Plains sharp-tailed grouse (Tympanuchus phasianellus jamesi) historically occurred along the Front Range of Colorado. Sharp-tailed grouse populations declined with human settlement and were "extirpated from most of their range in eastern Colorado by the late 1800's. In recent years breeding populations were documented only in Douglas County, although winter migrants or transients have been reported from Yuma, Logan, and Weld counties (Hoag and Braun 1990). Plans to increase distribution and populations of plains sharp-tailed grouse in Colorado will rely primarily on transplants (Braun et al. 1992). While numerous transplants of prairie grouse have occurred, few have been successful (Toepfer et al. 1990, Rodgers 1992, Hoffman et a1. 1992). Thus, it is desirable to document responses of sharp-tailed grouse to experimental transplants and evaluate parameters potentially affecting success including movements, habitat use, mortality, and reproduction. P. N. OBJECTIVES The objectives of this project are to assist with trapping and transplanting of plains sharp-tailed grouse into selected sites along the Front Range of Colorado and evaluate transplant success. Population characteristics of the transplanted population including movements and home range size, mortality, and production will be compared to a naturally pioneering sharp-tailed grouse population in Logan County near the Tamarack Ranch State Wildlife Area. SEGMENT OBJECTIVES 1. Review literature habitat use. on prairie 2. Coordinate efforts with Wyoming Game and Fish personnel and affected landowners in southeastern Wyoming to locate potential trapping sites for plains sharp-tailed grouse. 3. Transplant up to 40 plains sharp-tailed grouse from southeastern into suitable habitats along the Front Range of Colorado. 4. Radiomark up to 12 sharp-tailed grouse in the transplanted population and monitor movemen~s, habitat use, reproduction, and mortality. 5. Radiomark up to 12 sharp-tailed grouse from the pioneering population near the Tamarack State Wildlife Area and monitor movements, habitat use, reproduction, and mortality. 6. Prepare annual progress grouse introductions, report. movements, and Wyoming �54 METHODS Sharp-tailed grouse dancing grounds were located in Wyoming by driving along primary and secondary roads north of Burns in late March and stopping every 12 km to listen for courtship display. Landowners in the area were interviewed to ascertain seasonal occurrence of grouse and possible lek locations. Sharptailed grouse on and near the Tamarack State Wildlife Area were inventoried by surveying all historic and occupied greater prairie-chicken leks (Benson 1987, Schroeder 1990, Hoffman et al. 1992, L. Crooks, pers. commun.) and identifying greater prairie-chickens, sharp-tailed grouse, and hybrids. Surveys for additional active leks were made by driving secondary roads within 5-10 km of known leks and stopping at 1 to 2 km intervals and listening 3-5 minutes for displaying grouse and scanning with binoculars to search for grouse on leks. Sharp-tailed grouse and hybrids were captured on leks on the Tamarack area using walk-in funnel traps (Toepfer et al. 1988, Schroeder and Braun 1991). A sample of captured birds was fitted with miniature transmitters (11-13 g) on a necklace. Radio-marked birds were relocated on the ground by approaching them until they flushed. Attempts were made to locate radio-marked birds weekly through August and bi-monthly thereafter. Aerial searches were conducted for dispersing birds and those whose signals were lost. RESULTS AND DISCUSSION Transplant of sharp-tailed grouse Three days were spent in southeastern Wyoming in late March to meet with Wyoming Fish and Game personnel and explore habitats occupied by sharp-tailed grouse east of Cheyenne. We located 18 active sharp-tailed grouse dancing grounds in Laramie County between Burns and Albin. Most leks were in areas having a combination of winter wheat, native rangelands, and CRP plantings. The number of males on most leks ranged from 10 to 20. Permission to release sharp-tailed grouse was not obtained from the two primary transplant sites, Rocky Flats and Fort Carson (Braun et al. 1992) Thus, plans to trap and transplant plains sharp-tailed grouse from Wyoming into Colorado WEre postponed until 1993. Sharp-tailed grouse at the Tamarack Wildlife Area Sharp-tailed grouse were initially observed on the South Platte Management Area (Tamarack) during November and December, 1980 (Miller 1981). In 1989 sharp-tailed grouse or sharp-tailed grouse X greater prairie-chicken hybrids were observed on leks with greater prairie-chickens and breeding populations have been increasing annually'(Table 1). In 1992 we recorded 60 male greater prairie-chickens on 13 active booming grounds and 24 male plains sharp-tailed grouse on 7 leks (Table 1). Two leks had sharp-tailed grouse or hybrids only (leks 17 & 25), and 5 leks (leks 2, 11, 15, 19, & 22) had both greater prairie-chickens and sharp-tailed grouse or hybrids. The extent of hybridization is unknown but at least 13 of the 24 male sharp-tailed grouse observed on leks and 6 of 12 sharp-tailed grouse captured had plumage characteristics of hybridization. Only 3 female sharp-tailed grouse were �Table l. Annual maximum counts of male greater prairie-chickens and plains sharp-tailed grouse X greater prairie-chicken hybrids on leks in Logan and Sedgwick counties, Colorado, 1984-1992. 1984 Lek 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 GPC STG" 1 5 Totals 6 3.0 Avg. 1985 1986 1987 1988 1989 1990 1991 1992 GPC STG GPC STG GPC srG GPC STG GPC STG GPC srG GPC srG GPC STG 7 9 1 1 2 11 6 12 7 7 7 9 6 7 6 1 7 7 - 2 8 2 1 7 4 2 6 2 2 7 8 3 6 2 5 9 3 5 5 2 5 1 1 10 6 1 1 1 10 3 4 5 3 7 1 2 7 1 5 5 4 5 5 4 7 5 6 3 2 3 3 6 4 20 4.0 19 6.3 35 5.0 32 6.4 40 5.7 2 41 2.0 5.1 7 1.8 13 47 5.2 4.3 60 24 4.6 3.4 • Includes plains sharp-tailed grouse and hybrids. i..Jl i..Jl �56 identified on leks although it is likely that we failed to distinguish them from female greater prairie-chickens. Home ranges of radio-marked sharp-tailed grouse and hybrids Home ranges were calculated for 3 plains sharp-tailed grouse males and 4 sharp-tailed grouse X greater prairie-chicken hybrid males. The sharp-tailed grouse had larger home ranges than did the hybrids (Table 2). There were no obvious differences in habitats used by the grouse (sandsage grasslands were the primary habitat used by all grouse)·'and the reasons for the difference in home range size are unknown. Home ranges observed in this study were within the ranges of those reported in other studies of prairie grouse (Robel et a1. 1970, Toepfer and Eng 1988, Schroeder and Braun 1992). Table 2. Home range estimates of sharp-tailed grouse· and sharp-tailed grouse X greater prairie-chicken hybrids banded at Tamarack SWA, northeastern Colorado, 1992. Home Range Estimate (km2) Band No. Age Sex Species 304 306 310 311 312 330 1095 2+ 2+ 2+ 2+ 2+ 2+ 4+ M M M M M M M STG Hybrid STG Hybrid STG STG Hybrid Dates 4/14 4/14 4/16 4/16 4/16 4/22 4/23 - 11/20 7/9 11/20 11/20 11/20 11/6 12/18 N Obs. 23 14 15 23 21 20 23 MCP" HMb 3.40 l.02 2.84 l.08 4.73 3.67 l.12 5.32 2.00 4.40 l.10 6.31 6.31 l.16 95XE11ipse 7.50 3.80 8.71 2.51 13.44 13.52 3.10 • Minimum convex polygon (Mohr 1947) b Harmonic mean (Dixon and Chapman 1980) LITERATURE CITED Benson, L. A. 1987. Greater prairie-chicken survey, northeast Colorado, 1987. Unpubl. Rep. Colorado Div. Wi1dl., Fort Collins. 10pp. Braun, C. E., R. B. Davies, J. R. Dennis, K. A. Green, and J. L. Sheppard. 1992. Plains sharp-tailed grouse recovery plan. Colorado Div. Wild1., Denver. 33 pp. Dixon, K. R., and J. A. Chapman. 1980. Harmonf,c mean measure of animal activity areas. Ecol. 61:1040-1044. Hoag, A. W., and C. E. Braun. 1990. Status and distribution of plains sharptailed grouse in Colorado. Prairie Nat. 22:97-102. Hoffman, R. W., W. D. Snyder, G. C. Miller, and C. E. Braun. 1992. Reintroduction of greater prairie-chickens in northeastern Colorado. Prairie Nat. 24:197-204. �57 Miller, G. M. 1981. Development of a preservation program for three species of prairie grouse. Colorado Div. Wildl., Wildl. Res. Rep., Fed. Aid Proj. SE-3-3. Jan. 38-50. Mohr, C. O. 1947. Table of equivalent populations of North American small mammals. Am. Midl. Nat. 37:223-249. Robel, R. J., J. J. Cebula, N. J. Silvy, C. E. Viers, and P. G. Watt. 1970. Greater prairie chicken ranges, movements, and habitat usage in Kansas. J. Wildl. Manage. 34:286-306. Rodgers, R. D. 1992. A technique for establishing sharp-tailed grouse in unoccupied range. Wildl. Soc. Bull. 20:101-106. Schroeder, M. A. 1990. Greater Prairie-chicken survey, Tamarack State Wildlife Area, 1990. Unpubl. Rep. Colorado Div. Wildl., Fort Collins. 8pp. _______ , and C. E. Braun. 1991. Walk-in traps for capturing greater prairiechickens on leks. J. Field Ornithol. 62:378-385. _______ , and 1992. Seasonal movement and habitat use by greater prairie-chickens in northeastern Colorado. Colorado Div. Wildl. Spec. Rep. 68. 44pp. Toepfer, J. E., and R. L. Eng. 1988. Winter ecology of the greater prairie chicken on the Sheyenne National Grasslands, North Dakota. Pages 32-48 in A. J. Bjugstad, Tech. Coord. Prairie chickens on the Sheyenne National Grasslands. U. S. Dep. Agric., For. Servo Gen. Tech. Rep. RM159. _______ , J. A. Newell, and J. Monarch. 1988. A method for trapping prairie grouse hens on display grounds. Pages 21-23 in A. J. Bjugstad, Tech. Coord. Prairie chickens on the Sheyenne National Grasslands. U. S. Dep. Agric., For. Servo Gen. Tech. Rep. RM-159. _______ , R. L. Eng, and R. K. Anderson. 1990. Transplanting pra1r1e grouse: what have we learned? Trans. No. Am. Wildl. and Nat. Resourc. Conf. 55:569-579. Prepared by \)ffiI~ if\. ctA~ Kenheth M. Giesen Wildlife Researcher C (~ ') ��59 INTERIM JOB FINAL REPORT State of: Colorado Project: W-167-R Work Plan: Job Title: 14 Job: 01 January Bird Research 4 Movements. Reproductive Success. Greater Prairie-chickens Period Covered: Author: Upland and Habitat Use by Introduced 1991 through 31 December 1992 Grant M. Beauprez Personnel: C1ait E. Braun, Shane Briggs, Larry Budde, Courtney Crawford, Tim Davis, Francie Pusateri, Mike Schroeder, Mike Trujillo, Colorado Division of Wildlife; Grant M. Beauprez, Jennifer Clarke, University of Northern Colorado ABSTRACT The goals of this study were to: (1) successfully transplant greater prairiechickens (Tympanuchus cupido) to two sites in northeastern Colorado, (2) evaluate the success of greater prairie-chickens in establishing breeding populations at the two sites and in surrounding areas, and (3) improve guidelines for introduction of greater prairie-chickens into new or previously occupied habitats. In 1991, 43 birds (23 females, 20 males) were released at Pinneo, Washington County, while 50 (23 females, 27 males) were released near Wells Ranch, Weld County. In 1992, 41 birds (22 females, 19 males) were released at Pinneo, while 50 (27 females, 23 males) were released at Wells Ranch. Six males and six females were radiomarked at each site for each successive year and were relocated using radiotelemetry 1-2 times per week for approximately 12 months. Birds at Pinneo established 12 different leks with a mean distance from the release site of 7.12 km, while birds at Wells Ranch established 9 different leks. The mean distance of leks from the release site at Wells Ranch decreased from 1991 to 1992. Seven of 13 (53.8%) nests located at Pinneo were successful while 5 of 6 (83.3%) nests at Wells Ranch were successful. Mean clutch size for first nests was 12.6 eggs and 8.7 for second nests. The mean distance of nests from release. sites was 6.8 km at Pinneo and 11.5 km at Wells Ranch. Recruitment of juveniles was docUmented at 'Pinneo in 1992. Mortality was 38% at Pinneo and 44% at Wells Ranch. Of the . mortalities, 69% were caused by mammalian predators, 21% by avian predators, and 10% by miscellaneous causes. Mean dispe'rsal -distance of birds was 6.6 km at Pinneo and 15.0 km at Wells Ranch. These data indicate the Pirtneo releases were more successful in establishing populations of greater prairie-chickens than releases at Wells Ranch. The higher predation rate and lo~ger dispersal distance at Wells Ranch may be an indication the habitat at Wells Ranch was not as suitable as the habitat at Pinneo for sustaining a population of greater prairie-chickens. Differences between the birds' native habitat and habitats at the transplant site, and the length of time that birds are held in captivity prior to release may be the two primary factors influencing success of reintroductions. Prepared '-in· J>~l_~ by ~ ('OinJ:) Grant M. Beauprez Graduate Research l__o_LD) Assistant Approved by (fJy-~ ~--~~--~----_ Clait E. Braun Wildlife Research Leader ��61 JOB PROGRESS REPORT Colorado State of: Project: Upland Bird Research Y-167-R Work Plan: 17 Job Title: Population Dynamics of White-tailed Ptarmigan Period Covered: Author: Job ~7~ 01 January through 31 December 1992 C1ait E Braun and Kenneth M. Giesen Personnel: Kathy Martin, University of Toronto; C1ait E. Braun and Kenneth M. Giesen, Colorado Division of Wildlife ABSTRACT Long-term studies of populations of white-tailed ptarmigan (Lagopus 1eucurus) were continued at hunted (Mt. Evans) and unhunted (Rocky Mountain National Park) areas in Colorado through 1992. Breeding densities of ptarmigan at Rocky Mountain National Park increased slightly in 1992 while those at Mt. Evans decreased. Nest success at Rocky Mountain National Park increased in 1992 while that at Mt. Evans increased only slightly (33% vs. 20%) over that in 1991. Harvest increased dramatically at Mt. Evans and was at least 47-62% of the fall population. Breeding densities at Mt. Evans are expected to decrease dramatically in 1993. ��63 POPULATION DYNAMICS OF WHITE-TAILED PTARMIGAN C1ait E. Braun and Kenneth M. Giesen Long-term studies of trends in population size and investigation of reasons for fluctuations in size of tetraonid populations are lacking. Studies on the population dynamics of unhunted and hunted populations of white-tailed ptarmigan were initiated in Colorado in 1966 and have continued essentially uninterrupted at 2 sites. Studies of the unhunted population (Rocky Mountain National Park) identified possible short-term cycles of 7-8 years with an amplitude of 25-30% between high and low breeding densities. Conversely, studies of the manipulated population (hunted) at Mt. Evans have not indicated any cyclic pattern and it would appear that controlled hunting may mask any long-term trend that may occur. This study is designed to examine the question whether white-tailed ptarmigan are truly cyclic and whether hunting affects the apparent oscillations. P. N. OBJECTIVES The goals of this investigation are to be able to predict the length and amplitude of cycles in white-tailed ptarmigan in Colorado, to examine the impact of hunting on cycles, and to clarify underlying causes of the apparent cycles. SEGMENT OBJECTIVES 1. Conduct breeding (May-Jun) and brood (Aug-Sep) censuses of white-tailed ptarmigan using tape-recorded calls of males (breeding) and chicks (brooas). 2. Censuses will be conducted on previously established, defined study areas at Mt. Evans (hunted) and at Rocky Mountain National Park (unhunted). 3. Capture (noose poles) and band (aluminum and plastic color-coded bands) all unmarked white-tailed ptarmigan encountered on study areas at Mt. Evans and at Rocky Mountain National Park. 4. Individually identify all ptarmigan observed on study areas at Mt. Evans and Rocky Mountain National Park through use of binoculars. 5. Make hunting season and bag limit recommendations for Mt. Evans and collect hunting data through use of volunteer wing barrels and hunter field checks. 6. Compile data, analyze results, and prepare progress reports. STUDY AREA AND METHODS Areas investigated were Mt. Go1iath-Mt. Evans in Clear Creek County and at Tombstone Ridge-Sundance Mountain to Fall River Pass in Rocky Mountain National Park in Larimer County. The physiography, geology, location, and �64 vegetation of these study areas have been previously described (Braun 1969, 1971; Braun and Rogers 1971; Giesen 1977). Ptarmigan were located through use of tape-recorded calls (Braun et al. 1973), captured through use of telescoping noose poles (Zwickel and Bendell 1967) as described by Braun and Rogers (1971), and classified to age and gender and banded following Braun and Rogers (1971). Age of chicks was estimated following Giesen and Braun (1979). Numbered plastic bandettes were not used as in earlier years (Braun and Rogers 197.1) as a color-code system using up to 4 different colored plastic bandettes was instituted in 1977-78. A check station was operated on the Mt. Evans highway during the opening weekend of the ptarmigan season in that area. A volunteer wing collection station was available to hunters in the area when the check station was not in operation until the season closed. RESULTS AND DISCUSSION Breeding Densities Mt. Evans. -- Timing of breeding events in the Mt. Evans area in 1992 was about the same as in 1991. During the May-early June interval, 12 pairs and 2 single males were identified. Of the 14 males identified 6 were yearlings (42.8%) while 6 (50%) of 12 hens identified were yearlings. These data differ only slightly from that collected in 1991 (38% yearling males, 40% yearling females). However, adult survival from 1991 to 1992 decreased and the overall population decreased (Table 1). Rocky Mountain National Park. -- Timing of breeding events on the Trail Ridge study area was similar to the long term average and about one week earlier than in 1991. Surveys of ptarmigan on breeding territories along Trail Ridge Road in May and June indicated a m1n1mum population of 42 birds, similar to the levels documented in 1990. Overall, this population has remained low since 1982 (Table 1). The breeding density reflected low survival of banded adult males (33 of 59, 55.9%) and females (12 of 36, 33.3%) from 1991. Two yearlings banded as chicks in 1991 (both males) recruited to the study area in 1992 although yearlings comprised 35.4% of all adult ptarmigan identified. Nesting Success and Brood Size Mt. Evans. -- Twenty-four hens were located during August-early September 1992 on or immediately adjacent to the study area. Sixteen (67%) were without broods while only 8 had broods. Average brood size to 1 September was average (3.3 chicksfhen). Hatch dates for 22 chicks varied from 6 July to 5 August with 13 of 22 (59%) hatching after 21 July. These chicks probably were progeny of renests. Rocky Mountain National Park. -- Nest success was estimated from the proportion of hens with broods and without broods observed in July and August. Eleven of 16 hens observed during summer surveys were with broods for an estimated nest success rate of 69%, up from 50% in 1991. The median hatch date calculated from wing molt of 12 juveniles was 7 July (range 4-16 July) and was a week earlier than the 1966-1991 average. Brood size in August averaged 3.3 chicksfhen (range 1-7). �65 Table 1. 1966-92 White-tailed ptarmigan breeding densities (birdsjkm2), Colorado Study area Year Rocky Mountain National Park (5.5 km2) Mt. Evans (4.0 km2) 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 11.3 9.8 11.5 12.0 9.6 9.1 8.7 7.8 8.0 11.1 13.5 12.9 10.7 8.7 8.4 8.2 7.8 6.7 5.8 6.0 4.5 6.0 5.4 6.2 7.6 6.7 7.6 3.0 2.7 2.7 2.2 2.0 4.2 7.5 6.2 6.2 6.2 6.7 > 6.0 7.5 10.3 9.5 9.0 6.5 6.5 8.0 8.0 6.5 5.0 7.5 8.0 8.0 8.2 6.5 Harvest Mt. Evans. -- The hunting season at Mt. Evans in 1992 opened on 12 September and closed on 4 October (23 days) with a bag and possession limit of 3 and 9. This was the earliest opening date since 1972 when the season opened on 11 September. The early opening resulted from an administrative decision to open the statewide ptarmigan season on 1 September instead of the traditional opening date of the second Saturday. Weather in fall 1992 was mild and the Mt. Evans road remained open to vehicles through 4 October. This provided excellent access for hunters. A check station was operated on 12-13 and 19-20 September to measure hunting pressure and harvest. Twenty hunters with 38 ptarmigan (30 banded) were checked on 12-13 September while 7 hunters with 9 ptarmigan (7 banded) were checked on 19-20 September. Field checks resulted in 6 birds being examined on 14 September (2 hunters with 4 banded ptarmigan). Three additional banded �66 birds were reported harvested for a total of at least 45 banded birds. Wings of an additional 16 ptarmigan were recovered from the wing collection station. No bands were reported for these birds although it is probable that many were banded. Thus, at least 75 ptarmigan were harvested along the Mt. Evans highway in 1992. Known band recoveries represented birds banded from 1987 through 1992. Minimum harvest based on reported banded birds (45 of possibly 95 alive) was 47.4%. However, using a minimal estimate of 75 birds harvested and the known banded sample from the check station (37 of 47 checked were banded - 78.7%), it is probable that at least 62% of the fall population was harvested. This is the highest harvest 'at Mt. Evans since 1968 and was directly related to the early opening date (the birds had not dispersed), mild weather, and the road remaining open throughout the season. It is clear that not all bands recovered are reported. This is illustrated by the recovery of 5 bands (possibly 6) in 1992 from ptarmigan harvested at Mt. Evans in 1991 through a law enforcement operation. Thus, harvest estimates at Mt. Evans are minimal. The harvest of 47-62% (plus) of the fall population at Mt. Evans in 1992 was the highest since hunting seasons were manipulated starting in 1970 to reduce harvest levels to sustainable levels. While much of the harvest was of transient birds from south (Mt. Epaulet south) and west of the study area, it is expected that breeding populations from Mt. Warren west and south will be greatly reduced in spring 1993. LITERATURE CITED Braun, C. E. 1969. Population dynamics, habitat, and movements of whitetailed ptarmigan in Colorado. Ph.D. Thesis, Colorado State Univ., Fort Collins. 189pp. 1971. Habitat requirements of Colorado white-tailed ptarmigan. West. Assoc. State Game and Fish Comm. 51:284-292. Proc. _____ , and G. E. Rogers. 1971. The white-tailed ptarmigan in Colorado. Colorado Div. Game, Fish and Parks Tech. Pub1. 27. 80pp. _____ , R. K. Schmidt, Jr., and G. E. Rogers. 1973. Census of Colorado whitetailed ptarmigan with tape recorded calls. J. Wi1d1. Manage. 37:90-93. Giesen, K. M. 1977. Mortality and dispersal of juvenile white-tailed ptarmigan. M.S. Thesis, Colorado State Univ., Fort Collins. 55pp. _____ , and C. E. Braun. 1979. A technique for age determination of juvenile white-tailed ptarmigan. J. Wi1d1. Manage. 43:508-511. Zwicke1, F. C., and J. F. Bende11. J. Wi1dl. Manage. 31:202-204. Prepared by _.=:;a;..;;~=~_·.__ L.._ ~.......,=....;;.~ C1ait E. Braun Wildlife Research Leader 1967. _ A snare for capturing blue grouse. ~;nQth ~. cM\'wQ,v (2M) Ke neth M. iesen Wildlife Researcher C �67 JOB FINAL REPORT State of: Colorado Project: W-167-R Work Plan: 21 Job Title: Evaluation of Habitat Quality on Conservation Reserve Lands in Eastern Colorado Period Covered: Author: Upland Bird Research Job _5_ 01 January through 31 December 1992 Warren D. Snyder Personnel: L. L. Bixler, J. W. Moore, T. E. Remington, and W. D. Snyder, Colorado Division of Wildlife ABSTRACT Vegetation sampling to monitor quality within 105 Conservation Reserve Program (CRP) fields within eastern Colorado was initiated in early spring 1988 and continued through 1992 as part of a Division of Wildlife evaluation and as part of a national evaluation coordinated by the National Ecology Research Center (NERC). Seven of 105 NERC fields were contained within another sample of 42 fields to evaluate the impact of the Division of Wildlife's cost-share effort to enhance cover quality within CRP fields. Only a partial sample containing both data sets (41 fields) was monitored during 1991. This federal farm program, initiated in 1986, converted 14% of eastern Colorado farmland to perennial grass vegetation by 1987 and 18.2% by 1990. Density of CRP increased to the maximum (slightly> 25%/county) in southeastern Colorado and density was inversely related to the quality of the land for dry1and farming. Visual obstruction readings (VOR) conducted during pre-greenup (early spring) and nesting (late Jun-early Jul) were important components of both evaluations which were primarily directed toward ring-necked pheasants (Phasianus colchicus). Findings show that extensive use of herbicides and mowing suppressed annual weeds and enhanced establishment of grasses within CRP. Precipi~ation was.genera1ly favorable for gra~s establishment and growth during the interval.· Highest VQR indices ·were obtained in the northeastern corner of Colorado because of more favorable climate and soils. Nesting cover quality for pheasants within the NERC sample during pre-greenup was very poor to poor through 1991 and was rated fair in 1992. During the nesting season, ratings increased steadily among years from very poor to good for pheasants. Canopy cover of grasses increased, that of annuals decreased, and vegetation height increased with pr.ogression of time. Switchgrass (Panicum virgatum) ·produced the highest VOR indices during early spring and nesting intervals. Alfalfa-grass mixtures ranked second during nesting season and warm-season mixes were third. Smooth brome (Bromus inermis), tame wheatgrasses (Agropyron spp.), and native mid and short grasses generally yielded marginal cover for nesting pheasants. Division cost-share increased VOR indices during early spring (f < 0.05) and nesting intervals (f < 0.001). Approximately 32% of the CRP fields increased diversity by establishing grass within cropland, 38% �68 reduced diversity by reducing or eliminating cropland within rangeland situations, and 30% did not markedly change diversity. Wheat stubble and CRP possessed similar early spring VOR indices but the latter was more secure. Green wheat fields were more abundant and their VOR indices generally exceeded those of CRP by late April. Security was near equal until wheat harvest. About 60% of the CRP lands were within historic pheasant range in eastern Colorado although much of this was marginal range. Survival cover for wintering wildlife was rated as poor and was lacking both within and adjacent to CRP fields in most instances. The CRP.possibly reduced habitat quality for scaled quail (Callipepla squamata) by reducing food sources, may have had similar impacts on prairie grouse, and primarily was outside the range of these species and northern bobwhite (Colinus virginianus) in eastern Colorado. Analyses of data from breeding bird survey routes, mourning dove (Zenaida macroura) call-count routes, and pheasant crowing routes prior to and during the interval of CRP failed to discern detectable increases in avifauna (possible declines were indicated). Lack of survival cover was considered the primary reason pheasants did not respond to the CRP in eastern Colorado. �69 RECOMMENDATIONS Much of the land enrolled in the Conservation Reserve Program has already been under contract for 5-7 years of the 10-year contract. Therefore, there is limited opportunity to alter existing vegetation for wildlife. Based on the possibility the program will be extended and expanded, or that a similar program may evolve in the future, the following recommendations for wildlife enhancement are presented. 1 The Colorado Division of Wildlife should concentrate efforts specifically for pheasants only within Colorado's historic better pheasant range. It is doubtful the program, in its present form, can markedly benefit other game species in eastern Colorado. 2. The CP-12 food-cover provision should be the primary habitat modification funded by the Division of Wildlife. This practice should be used to establish. tall sorghums and annual forbs as survival cover for pheasants within small « 2 ha) plots. Any CRP tract larger than 6 ha should be considered for a food-cover planting. Such plantings should be at least 50 m wide to reduce the impact of drifting snow and should be near tilled cropland. At least one plot per 32 ha should be used. The Division of Wildlife should be prepared to conduct much of the site preparation and planting, or to contract it, since many landowners will have neither the equipment, time, nor inclination to do the work. 3. Assuming the CRP is extended, or a similar program is implemented in the future, the Division of Wildlife should work closely with the U.S.D.A. Soil Conservation Service in attempts to obtain seed mixtures which are compatible with the needs of both agencies. Switchgrass, if it can not be planted singularly, should be planted only in combination with other tall, warm-season species and should comprise at least 50% of any mixture. If such mixtures can not be used because of SCS specifications in some soil types and locations, alfalfa should be a primary component in cool-season grass/legum~ mixtures. Monocultures of smooth brome and tame wheatgrass species should not be permitted on CRP. The Division of Wildlife should not accept or approve grass seed mixtures that will not achieve its cover quality objectives. 4. The U. S. Department of Agriculture seems committed to massive crop reduction programs for wheat, corn, and other over-produced crops. In the past, these have usually been annual reductions (annual set aside) that promote erosion, promote production of non-target crops, and in eastern Colorado, are often detrimental to pheasants and other wildlife. The Conservation Reserve Program, in contrast, is long-term (10 years), reduces erosion, but still has limited value for wildlife in eastern Colorado. Implementation of a shorter multi-year program (3-5 years) is recommended wherein forage sorghums, sweet clovers, alfalfa, other legumes in combination with cool-season grasses could be used. Efforts should be designed to provide taller, better quality cover more rapidly which was of greater value to wildlife. 5. Although livestock grazing to remove excess cover may be warranted in higher rainfall areas of the Midwest, it is not recommended as a management tool within CRP in the High Plains Region. ��71 EVALUATION OF HABITAT QUALITY ON CONSERVATION IN EASTERN COLORADO RESERVE LANDS Warren D. Snyder INTRODUCTION AND BACKGROUND Data collection continued during the fifth and final year to evaluate vegetation quality within Conservation Reserve Program (CRP) fields in eastern Colorado. Two data sets were under assessment. The first contained a random sample of 105 fields for which data were supplied to the U.S. Fish and Wildlife Service, National Ecology Research Center (NERC) in Fort Collins as part of a nationwide e'valuation effort. Seven of those fields were included within a second data set of 42 CRP fields which the Division of Wildlife cost shared to obtain better cover quality. The Conservation Reserve Program (CRP) was created under the 1985 Federal farm bill with the primary objectives of reducing crop surpluses by taking cropland out of production for extended (10 year) intervals while reducing erosion on these lands. Lands that possessed a high risk for erosion were eligible and most nonirrigated farmland in eastern Colorado qualified. Farmers bid their lands to the program at rates up to ceilings established by the Agricultural Stabilization and Conservation Service (ASCS). These limits, set subsequent to initial bids, were $40.00jacre and $45.00jacre respectively in southeastern and northeastern Colorado. Nearly all CRP lands in eastern Colorado were enrolled during the first several years of the program. Over 14% (1,412,659 acres) of the eligible cropland in eastern Colorado was consigned by March 1987, a little over a year after the program was implemented (Table 1). More marginal farmlands in southeastern Colorado were consigned to the maximum (25%jcounty) within 1 to 2 years (some counties were allowed to exceed the limit by a small percentage). Acceptance of the program in northeastern Colorado was more gradual and at a lower level. As of January 1990, 18.2% (1,805,477 acres) of cropland were accepted into the program within 23 eastern Colorado counties (Table 1). The CRP was continued as part of the 1990 federal farm program, but with modified criteria for eligibility and bidding. As a consequence only a few thousand acres have been consigned to the program since then in Colorado. The Soil Conservation Service (SCS) within the U.S. Department of Agriculture was responsible for developing technical specifications for the CRP. They required planting a sorghum cover crop during the first year and perennial grasses were then seeded into this cover during the subsequent late fall, winter or early spring. In a few instances, tame cool-season grasses were planted directly into wheat stubble. Herbicides, primarily Chlorsulfuron (Glean), were permitted for weed suppression and used extensively. Mowing was an alternate option to suppress annuals. The Division of Wildlife requested that crested wheatgrass not be permitted as a perennial cover in CRP; a request that was enacted. Within extremely marginal farmlands, primarily in southeastern Colorado, the SCS required that native species be planted. In less marginal areas, primarily in northeastern Colorado, tame cool-season species (smooth brome and wheatgrasses) could be used. �72 Table l. Acres and percent of cropland by county accepted into the Conservation Reserve Program, eastern Colorado, 1987 and 1990. Total cropland {acres} 621,800 1987 21,499 1990 34,669 1987 3.5 1990 5.6 201,400 31,947 35,231 15.9 17.5 Baca' 1,032,000 278,757 276,661 27.0 26.8 Bent 116,000 26,808 28,524 23.1 24.6 Cheyenne 642,000 119,724 151,016 18.7 23.5 Crowley' 84,700 23,672 23,962 28.0 28.3 208,000 52,031 51,645 25.0 26.3 96,500 12,963 15,548 13.4 16.1 6,800 1,950 1,773 28.7 26.1 Kiowa' 685,000 192,108 175,549 28.0 25.6 Kit Carson 885,000 103,465 149,631 11.7 16.9 Las Animas 90,000 19,359 20,468 2l. 5 22.7 Lincoln' 512,000 88,359 142,207 17.2 27.8 Logan 654,700 22,636 59,520 3.5 9.1 Morgan 242,300 21,926 50,344 9.1 20.8 84,600 2,597 2,694 3.1 3.2 Phillips 373,000 8,800 16,515 2.4 4.4 Prowers' 530,000 144,847 155,721 27.3 29.4 Pueblo' 114,700 32,239 35,264 28.1 30.7 Sedgwick 211,000 4,828 4,016 2.3 l.9 Washington 838,000 76,580 169,824 9.1 20.3 Weld 1,037,000 93,295 151,097 9.0 14.6 Yuma 644,300 32,448 53,548 5.0 8.3 9,910,800 1,412,659 1,805,477 14.3 18.2 County Adams Arapahoe Elbert' E1Paso Huerfano' Otero Total/Mean Acres , Designates counties where the maximum Conservation Reserve was attained. % allowable percent of �73 P. N. OBJECTIVES Determine distribution and quantity of Conservation Reserve Program land in eastern Colorado in relation to distribution of selected wildlife species, evaluate the quality of vegetation on these lands for selected wildlife species, measure response of selected wildlife species to the Conservation Reserve Program using existing annual surveys, and evaluate the impact of the Colorado Division of Wildlife's cost-share program on cover quality. SEGMENT OBJECTIVES 1. Conduct evaluations of randomly selected CRP fields within eastern Colorado as part of a regional and national assessment of the Conservation Reserve Program coordinated by the National Ecology Research Center (NERC) of the U.S. Fish and Wildlife Service. 2. Conduct intensive visual obstruction readings sample of CRP fields and proximal controls. (VOR) in a stratified random 3. Conduct intensive visual obstruction readings (VOR) in a sample of fields cost shared by the CDOW (for enhancement of cover quality) for comparison with CRP fields not cost-shared. 4. Estimate response of selected wildlife population surveys. 5. Monitor effects of precipitation 6. Compile and analyze data and prepare species to the CRP based on annual on CRP VOR. the final report. METHODS The contract data base provided by NERC was used during contacts of county offices of the Agricultural Stabilization and Conservation Service to obtain photocopies of CRP contracts, aerial photos of fields, legal descriptions, and cropping history for a random sample of fields. An authorization letter form the State ASCS Director was obtained to facilitate release of data from county ASCS offices. Landowners and/or operators within 18 eastern Colorado counties representing 78 contracts and 104 CRP fields were contacted to gain permission for access to fields and to explain the purpose of the study. The visual obstruction reading (VOR) Robel et. al (1970) was used to measure nesting cover quality obtained from CRP fields during pre-greenup (Mar) and late spring (mid to late Jun) nesting intervals. NERC requested 8 VOR samples per field, however, 20 samples per field were obtained in all Colorado fields to reduce within-field variability. Sampling was conducted on an accessible side of each field and extended into the field based on random selection of target objects (if available). The 1st sample was obtained 25 paces from the field edge and subsequent samples were obtained at 10 pace intervals thereafter. Procedures for vegetation sampling during the nesting interval were similar to those used during pre-greenup and sampling was conducted from south to north across eastern Colorado. �74 The VOR or height-density index (HDI) sampling technique has been used extensively in eastern Colorado (Snyder 1984Q, 1985, 1991, 1992) for herbaceous vegetation. The VOR index is based on procedures developed by L. M. Kirsch (unpubl. rep., U.S. Fish and Wildl. Serv., Jamestown, N.D., 1977). NERC required a different (higher) interpretive reading of VOR for use in regional evaluations than the Kirsch method previously used in Colorado. To maintain continuity in Colorado sampling, the Kirsch method was retained as the basic method and an adjustment was made to correct it to the NERC method. All VOR presented within this text are based on the Kirsch method which recorded the highest obstructed 0.5 dm using a square pole. Percent canopy cover of herbaceous vegetation within a 0.5 x 1.0 m (0.5 m2) Daubenmire frame was estimated twice at each of the 4 locations used in sighting toward the Robel pole (Daubenmire 1959). The 0.5-m: plot immediately to the front and right of the sighting stick was sampled and the sample frame was flipped forward to obtain a 2nd sample. The 1st 8 samples were obtained at the 25-pace location and a 2nd group was recorded at the last location along the VOR sample transect. Perennial and annual grasses were distinguished from annual and perennial forbs. Additional information concerning dominant seeded species, rating of stand establishment, associated cover types, amount of edge, distance to feeding and winter sites, and other data were collected during the 5-year study. Vegetation VOR data were segmented into 6 strata for analysis based on potential differences due to soils, precipitation, and species planted (Fig. 1). Analysis of variance was used to detect differences among vegetation types and strata. Unpaired ~ tests were used to test for differences between NERC and CDOW samples. Sampling was conducted from early spring through early summer 1988 to provide da.ta on quantity, quality, and security (for nesting wildlife) of land use types, including CRP in eastern Colorado. Sampling through spring was nonrandom and restricted by time constraints, primarily to the eastern tier of counties. Pastures and rangelands were not included within spring samples but were included in the June samples. During mid- to late June sampling, roadside listings of cover types were obtained to 20 kID from each sampled CRP field during travel from one field to the next. Since CRP fields were randomly selected, this provided a random distribution of associated cover types. The Colorado Division of Wildlife (CDOW) cost shared directly with farmers on numerous CRP fields in eastern Colorado attempting to increase use of grass species possessing greater herbaceous cover quality for ring-necked pheasants and other ground-nesting wildlife. Seven of these fields were within the initial NERC sample of 104 fields monitored beginning in 1988. Thirty-six additional fields were randomly selected in February 1989 and landowners were contacted for access permission to monitor vegetation conditions. It was learned later that one of these fields had not been cost shared by the CDOW. Therefore, that field was included within the NERC sample bringing the total fields not cost shared to 98 among 105 total NERC fields. Forty-two fields, cost shared by the CDOW, were monitored bringing the overall sample to 140. The number of CRP fields sampled in 1991 was reduced to 41 (contrasted to 140 in preceding years). These were concentrated within strata 1 through 4 (within the eastern part of Colorado). Efforts were coordinated with T. E. �75 wet. 0 ) ". Fig. 1. eastern Stratification Colorado. for random.sampling conservation reserve fields in �76 Remington, who conducted a companion study assessing use of Conservation Reserve Program fields by avian wildlife. He conducted some of the analysis of variance presented within this report. Monthly and annual precipitation data were obtained from 9 U.S. Weather Bureau stations distributed across eastern Colorado. Locations were at or near Springfield, Lamar, Rocky Ford, Limon, Burlington, Akron, Holyoke, Sterling, and Greeley (Fig. 1). Conservation Reserve Program distribution was contrasted to pheasant range based on 1982-83 pheasant distribution-density estimates (Snyder 1985). Comparison of the CRP distribution with that of northern bobwhite was based on bobwhite distribution presented in Snyder (1984~). Scaled quail distribution and density were based on Hoffman (1965) and modified to exclude range he classed as very poor, since most of this was farmland or shrubless shortgrass unsuited for quail habitation. Hoffman rated a large proportion of the scaled quail range as good based on habitat characteristics rather than population densities existing there. RESULTS Environmental Conditions Distribution of CRP.--Landowner consignment of cropland to the Conservation Reserve Program occurred rapidly in southeastern Colorado. By the end of the 4th sign-up interval in 1987 the maximum acreage per county had been committed in several southeastern Colorado counties (Table 1). A lower percentage of the farmland in northeastern Colorado was contracted to the CRP at a slower rate. Establishment of Vegetation.-- The requirement that a cover crop be planted prior to seeding grass delayed grass seeding. Only a few CRP tracts were seeded to grass during 1986, primarily in southeastern Colorado. Fall 1987 inspections revealed the majority of CRP fields had been planted to sorghum in summer 1987 and would be seeded to perennial grasses from November 1987 through April 1988. Up to a third of the tracts had been seeded to perennial grasses, primarily sideoats grama (Bouteloua curtipendula), in southeast Colorado by summer 1987. Wild annuals dominated in those fields and varied widely in species composition, height, and density. Among 104 sampled NERC fields, 1 was seeded in 1986, 42 in 1987, 57 in 1988, and 4 were seeded in 1989. Establishment of perennial grasses, which is usually slow, was enhanced by weed suppression (herbicides), mowing of weeds, and above average precipitation. The percentage of perennial grass canopy cover increased steadily from 1988 through 1991 whereas the percentage of annual forbs declined (Table 2). Sandbur (Cenchrus spp.), which herbicides did not control, formed dense stands within many CRP fields, especially within sandy soils. Sandburs gradually diminished and were not a major problem after 1989. Field bindweed (Convolvulus arvensis) was a major problem in southern Colorado, primarily in Prowers, Baca, and Bent counties where whole fields were occupied. Satisfactory grass stands were not established on several fields within the NERC sample because of field bindweed. In two instances it was assumed the fields were subsequently dropped from the CRP. �77 Table 2. Canopy cover (%) of total vegetation, perennial grass, and annual vegetation, and mean vegetation height (em) within Conservation Reserve fields during pre-greenup and nesting intervals, 1988 - 1992, eastern Colorado. Year Perennial ~rass Annual forbs Total canopv cover Hei~ht Pre-greenup 1989 8.7 28.6 37.3 17.7 1990 20.8 3l. 5 52.3 16.4 1988 5.3 18.7 24.0 9.7 1989 14.5 28.1 42.6 16.3 1990 27.8 26.4 54.2 19.7 1991 39.3" 12.4" 5l. 7" 24. sa 1992 39.3 10.9 50.2 28.1b Nesting Projected from a subsamp1e of 41 fields (18 NERC and 23 CDOW) based on 1990 to 1991 changes. b Vegetation was sampled later in 1992 than in preceeding years potentially increasing height. Preciuitation.-- Monthly precipitation averaged &~ong 4 southeastern weather stations and 5 northeastern weather stations for 1987-92 was plotted in comparison to the long-term mean for eastern Colorado (Fig. 2). Considerable fluctuation during spring and summer, in comparison to the longterm mean, was evident. Annual precipitation in 1987 was above average at 7 of 9 locations in eastern Colorado (Table 3). In contrast, only 1 of these (Springfield) recorded above average precipitation in 1988. However, precipitation was not markedly deficient during 1988 in eastern Colorado. Early 1989 remained extremely dry throughout eastern Colorado. As a result most of the winter wheat crop was either severely stunted or lost. However, considerable rainfall was received, starting in May, in southeastern Colorado and continued through most summer months (Fig. 2). Limited precipitation was received during mid-May in northeastern Colorado, but it remained dryer than southeastern Colorado until June. July was moderately dry at most locations. Precipitation averaged through the first 4 months of 1990 was greater than that received during the same interval in 1989. May precipitation was near average, however, June was extremely hot and dry throughout eastern Colorado. Above average rainfall was received from July through fall 1990. �78 14 14 12 12 '[10 - ~10 sa ,Sc,_ e 'Q ct 4 2 , JFMAMJ I JA I I ! I I saND J I FMA 1987 14 12 12 10 '[10 -6 -a -a " , , § «5 'Q ct 4 2 ",SE 4 2 ......... .. I JFMAMJJA I I JFMAMJ saND 14 14 12 12 -s '[10 JA . I saND - '[10 NE § §a ",SE i:t: e c.. 'Q I 199o 19e9 ct I i:t: e c, 'Q ct I saND , § c.. I JA tssa 14 i:t: I MJ ~ 6 c, 'g ct 4 2 I I I I JFMAMJ I JA 1991 I I I saND I 4 2 I I J I I F M A M J J I A SON i D 19~ Fig. 2. Monthly precipitation (cm) averaged among weather stations in southeastern and northeastern Colorado in relation to the long-term mean, 1987-92. �79 Table 3. Annual prec~p~tation (em) at 9 U.S. Weather Bureau locations in eastern Colorado, 1987-92. Location Lon~-term 1987 1988 1989 1990 1991 1992 ~ Springfield 42.3 55.5 44.3 46.0 47.6 38.7 54.0 Lamar 39.1 38.4 29.4 34.2 39.9 40.1 47.5 Rocky Ford 31.8 29.2 23.9 24.4 45.4 25.0 31.3 Limon 38.3 50.4 36.9 29.7 49.8 44.1 42.1 Burlington 44.2 46.3 33.3 41.2 40.9 44.9 41.2 Akron 44.0 50.4 4l.2 34.4 53.0 33.0 42.8 Greeley 31.6 33.1 31.3 39.9 34.8 30.7 40.1 Holyoke 44.7 5l.0 41.8 42.6 41.7 46.6 39.S Sterling 38.0 37.9 37.3 34.5 46.4 31.3 49.7 Combined ~ 39.3 44.1 35.S 36.3 44.4 37.2 43.1 Precipitation fluctuated extensively among months and locations in 1991. Large areas within strata 1 and 2 (Fig. 1) received considerable rainfall from early May t~rough mid June 1991. Most stations recorded average or better precipitation in July, however, it was below that received the previous year. Overall, spring-summer rainfall averaged near or above normal in 1991. Springfield (Baca County) did not lead in total annual rainfall for the first time in several years based on data through the first 10 months of 1991; however, it like most stations in the eastern part of Colorado, received above average precipitation (Table 3). Markedly deficient rainfall was recorded during April and May 1992 followed by above average moisture in subsequent months in a pattern similar to that in 1989 (Fig. 2). Evaluation of Vegetation Quality within CRP Visual Obstruction Readings (VOR) among Years and Strata.-- During pregreenup (Mar) sampling, VOR remained relatively constant throughout the study (Table 4, Fig. 3). The primary increase occurred in 1992 within both the NERC and CDOW samples. Data for 1991 were projected based on a subsample of 41 fields within the eastern tier of strata. Nesting season (late Jun-early Jul) VOR's within the NERC sample increased steadily during the study (E < 0.05), and a similar trend was evident within the CDOW sample (Table 4). Slightly later sampling during the latter years of �80 1.2 , , , 'E 1.O "C .•....... CJ z 0 L5 a:: z 0.6 TI 0 0.4 , , , , , , , 0 ~ 0 :::> a: r(J) CD _J « :::> (J) > Ixl j,,/ 0.8 ,, , .; .,. , .,. .,. *' , , , , , , , *' *' I 1 0.2 o ~I------------~------------~I------------~ 1992 1988 1989 1990 YEAR Fig. 3. Trends in visual obstruction readings (VOR-dm) with 95% CI for NERC pre-greenup and nesting season samples, eastern Colorado, 1988-92. �81 Table 4. Average visual obstruction readings (dm) within Conservation Reserve fields sampled during pre-greenup and nesting season intervals, eastern Colorado, 1988-90. Pre-greenup Nesting Fields Within the NERC Sample (n - 105) 1988 0.39 0.20 1989 0.34 0.56 1990 0.39 0.67 1991 0.38" 0.86" 1992 0.57 1.05 Fields Within the CDOW Cost Shareb (n - 42) 1939 0.51 1.13 1990 0.66 1.09 1991 0.65' 1.40' 1992 1.00 2.08 Combined NERC and CDOW Cos~ Shared Fields (n - 140) 1989 0.36 0.77 1990 0.47 0.83 1991 0.46' 1.07" 1992 0.68 1.31 • Projected from a subsamp1e of 41 fields (18 NERC and 23 CDOW) based on 1990 to 1991 changes. b Seven fields were also contained within the NERC sample. �82 study was partially responsible. season grasses was also noted. However, progressive development of warm- During 1988 (CDOW fields were not yet sampled), fields within the southern strata (3 and 4), which had generally been planted a year earlier, contained higher VOR indices during pre-greenup sampling than those to the north (Table 5). Strata 1 and 2, in the northeastern part of the State contained the highest pre-greenup indices in subsequent years. Differences in nesting season VOR indices among strata were less pronounced from 1988 through 1990 but strata 1 and 2 contained higher indices in 1992 (f < 0.05, Table 5). There was a transition toward better quality cover with study progression in both pre-greenup and nesting intervals (Fig. 4). Most cover was rated as very poor for nesting pheasants in 1988. However, by 1992, most was poor to fair during pre-greenup and fair to moderate during nesting intervals. Table 5. Average visual obstruction reading (VOR-dm) per stratum within CRP fields during pre-greenup and nesting intervals, 1988-92, eastern Colorado. Year 1 Stratum 3 2 4 5 6 Pre-greenup 1988 0.11 BC' 0.04 C 0.74 A 0.52 ABC 0.20 BC 0.06 BC 1989 0.51 AB 0.72 A 0.19 B 0.21 B 0.31 B 0.17 B 1990 0.52 AB 0.69 A 0.48 AB 0.50 AB 0.42 AB 0.25 B 1992 0.93 0.87 AB 0.56 BC 0.48 C 0.57 BC 0.36 C A Nesting 1988 0.22 A 1989 1990 A 1.01 AB 1. 28 0.97 0.89 A 1992 0.07 1. 88 A A A 1.72 A 0.24 0.06 A A 0.12 A 0.17 A 0.59 B 0.81 AB 0.46 B 0.59 B 0.79 0.73 0.73 0.57 A 1.11 B Means with the same letter do not differ A 0.77 B (1: > 0.05). A l.08 B A 1. 20 B �83 80 PRE-GREEN UP III 1988 b3 1989 LB 1990 o 1992 60 I- Z ill 040 0: ill 0.. 20 <.26 V.POOR .2t3-.M .51-1.0 POOR FAIR VISUAL OBSTRUCTION 1.1-2.0 >2.0 MODERATE . GOOD READING (dm) 100 NESTING SEASON eo II 1988 &1 1989 [§1 o 1990 1992 I- 60 Z ill o 0: ill 0.. 40 20 <.26 V.POOR .26-.50 .51-1.0 POOR FAIR VISUAL OBSTRUCTION 1.1-2.0 >2.0 MODERATE READING (dm) GOOD Fig. 4. Rating of cover qual ity (VOR-dm) for nesting pheasants during pre-greenup and nesting season intervals, eastern Colorado, 1988-92. �84 An inverse relationship was evident between the percentage of available dry farmland consigned to the CRP and the VOR index (~- 0.75, Table 6). Only 8.8% of eligible cropland was in CRP in extreme northeastern Colorado where the summer 1992 VOR index averaged 1.88 dm. The VOR index was 1.0 dm lower in the southeastern corner of Colorado where 26.9% of the farmland was in CRP. Table 6. The 1992 nesting season VOR index (dm) in relation to the average amount of CRP (%)' within grouped county areas, eastern Colorado. Location Counties VOR 40 1.88 8.8 % CRP Northeast Phillips, East-central Kit Carson, Cheyenne, Kiowa 27 1.30 22.0 Northwest Morgan, Weld, Arapahoe 14 1.18 11.3 Southwest Lincoln, Elbert, El Paso, Crowley, 17 1.17 25.8 Southeast Baca, Bent, Prowers 42 0.86 26.9 Based on January, Service. Sedgwick, Logan, Yuma, Washington N Pueblo 1990 acreages of CRP provided by the Soil Conservation Cover Dualitv in Relation to Species.-- The VOR indices of grass mixtures and grass-alfalfa mixtures were compared during pre-greenup and nesting intervals from 1988 through 1992 (Table 7). Differences were relatively insignificant during pre-greenup until 1992 when switchgrass provided markedly better standing residual. With progression of grass establishment, nesting season samples again showed that switchgrass was the best cover. Alfalfawheatgrass and alfalfa-brome were primarily associated with CDaW cost share within strata 1 and 2 and most fields provided good nesting cover (CDOW fields were not sampled in 1988). Warm-season grass mixtures developed slowly but ranked 2nd and 3rd in quality during 1992 respectively in pre-greenup and nesting season samples (Table 7). Warm-season grass mixtures dominated in 13 fields during summer 1989 (VOR = 0.77 dm) but increased to 38 by 1992. Most of these fields were dominated by annual forbs in 1989 indicating the transition to grasses during the 3-year interval. Introduced wheatgrasses and smooth brome VOR indices during the nesting season from 1989-92 were significantly lower than when combined with alfalfa. Native grasses, dominated by sideoats grama, blue grama, and western wheatgrass, usually provided marginal cover, whereas annual forbs yielded diverse cover. Forbs dominated in 20 of 140 fields during 1992. Location influenced VOR indices. Smooth brome was primarily planted within loam soils in strata 1 and 2. Soil Conservation Service specifications prevented its use in marginal soils to the west and south where native grass mixtures were mandated. The VOR index for smooth brome would have been lower �85 Table 7. Average visual obstruction readings (VOR-dm) in relation to dominant vegetation within CRP fields during pre-greenup and nesting intervals, 198892, eastern Colorado. Year Switchgrass Alf/cool season' W. season mix Smooth brome Wheatgrassesb Native grasses Annual forbs Pre-g:reenup 1988 0.11 A 0.00 A 0.16 A 0.17 A 0.14 A 0.50 A 0.61 A 1989 0.64 A 0.49 A 0.44 A 0.60 A 0.29 A 0.16 A 0.38 A 1990 0.87 A 0.50 AB 0.43 AB 0.56 AB 0.31 B 0.43 AB 0.64 AB 1992d 1. 80 A 0.79 B 0.45 B C 0.72 B 0.45 B 0.44 B 0.41 B Nesting: 1988 0.02 A 0.02 A 0.17 A 0.10 A 0.11 A 0.10 A 0.23 A 1989 1.72 A 1.41 A 0.58 C 1. 32 AB 0.57 C 0.45 C 0.80 BC 1990 1.52 A 1.75 0.82 B 0.87 B 0.54 B 0.53 B 0.61 B 3.04 A 2.34 B 1.49 C 1.12 CD 1.02 CD 0.78 D 0.91 D 1992 b d A Cool season grasses = smooth brome and tame wheatgrasses. Wheatgyasses = intermediate, pubescent, and tall species. Means with the same letter do not differ (f >0.05). Only a partial sample was obtained in 1991 so data are omitted. if it had been planted in more marginal soils. Its primary use was within Colorado's historically better pheasant range. Intermediate, pubescent, and other introduced wheatgrasses were widely used in strata 1 and 6. Plantings dominated by switchgrass were primarily within sandy soils in the eastern tier of strata, whereas other warm-season mixes were used there and within sandy soils in all strata. Thus, site, precipitation, and species must all be considered in interpreting VOR indices among species and species groups. Division of Wildlife Cost Share.--The Colorado Division of Wildlife costshared with farmers to increase the quality of vegetation within CRP fields in eastern Colorado. Based on combined pre-greenup data for 1990 and 1992, �86 comparing cost-shared fields with nearby NERC fields, the cost-share program increased herbaceous vegetation quality (NERC ~ - 0.60 vs CDOW ~ - 0.82, f < 0.05). Using the same groups of fields, the 1990-92 combined nesting season data yielded mean indices of 1.04 (NERC) and 1.69 (CDOW) (f < 0.001). Cost of this program is varied by county (Table 8). Average cost per acre for herbaceous cover enhancement in southeastern Colorado was $6.25 for a total expenditure of $188,387. That for northeastern Colorado was $141,029 for an average of $9.20/acre. About $17,289 was paid for 1,665 acres in the Central Region by fall 1991. Approximately 2.4% of the CRP in Colorado was included in the Division of Wildlife cost share program. Cost-shared field size was Table 8. Amount and cost for perennial grass and woody plantings within the Conservation Reserve Program cost-shared by the Colorado Division of Wildlife with private landowners, eastern Colorado. County Grass Cost Acres Woodv planting:s n Cost Adams 2,378 $11,306 o Baca 2,471 28,852 7 $1,102 10,402 60,931 2 1,000 205 1,735 Elbert 1,513 7,819 E1Paso 326 4,418 Kiowa 2,686 19,060 o o o o Kit Carson 3,420 23,479 2 Las Animas 303 889 Lincoln 308 2,694 o o 1,394 14,253 5 2,823 Morgan 584 4,604 5 3,330 Otero 133 2,375 o Phillips 527 6,870 5 8,378 36,624 o Pueblo 170 1,953 2 414 Sedgwick 210 3,020 2 2,406 Washington 2,920 18,961 13 13,845 Weld 2,438 23,919 2 587 Yuma 2,940 40,337 3 6,000 43,706 $314,099 48 $40,155 Cheyenne Crowley Logan Prowers Totals 1,000 7,648 �87 much smaller (usually 40 acres) in northeastern Colorado pheasant range (larger fields were cost shared within prairie grouse range). In addition to herbaceous plantings, the Division of Wildlife spent $36,640 for 35 woody plantings in the Northeast Region and $3,500 for 13 plantings in southeastern Colorado. Canopv Cover and Height.-- Perennial grasses (a few samples included alfalfa) increased from 5% (1988) to 39% (1992) whereas the percentage of annual forbs declined during nesting season (Table 2). Total canopy cover of combined vegetation increased to about 50% by 1990 but did not show evidence of continued increase in 1992. Vegetation height (excluding seed head) increased steadily (Table 2). Increases in 1991 and 1992 were partially because sampling during the nesting season was 1-2 weeks later than in previous years. However, greater growth of slow developing warm-season grasses was also evident. Edge and Interspersion About 33.4% of the CRP field borders were adjacent to rangeland or pasture and ano:her 28.8% were adjacent to other CRP fields. This grouping represented 62.2% of the total edge based on sampling conducted during 1988. Residual stubble, fallow, and small grains bordered 32.0% of the sampled CRP tract edges; sorghum and corn bordered an additional 3.6%. Developed land (building sites, etc.) and other (mostly minor crops) represented 2.2%. Based on general observations and assessments, approximat~ly 32% of the 104 fields increased diversity by extending grass into cropland, whereas 38% decreased diversity by converting former cropland within rangeland into grass. About 30% of the CRP tracts did not markedly change diversity. Approximately 22% of the sampled CRP tracts extended rangeland into cropland, 24% reduced cropland in range-dominated sites, 9% eliminated cropland in range-dominated sites, and an additional 45% provided no change. As the study progressed, a few more CRP fields were consigned adjacent to the sample fields further reducing diversity of the sampled group. Quantity, Quality, and Security of Land Use Types Percentages of vegetation types sampled by strata in early spring 1989 varied by strata (Table 9). These data reflect land use in areas near CRP fields, which, in turn, were formerly cropland. Percent rangeland would be still higher, especially in strata 4 and 5 , if sampling was completely random. If small grains, standing stubble, and mulched stubble were combined, they would equal about 42% of the total. Comparisons of standing and mulched stubble show that most wheat stubble was left standing over winter in northeastern Colorado (strata 1 and 6), whereas little remained standing in southeastern Colorado. Some land classed as fallow was formerly wheat stubble tilled after harvest to provide volunteer wheat pasture for livestock through fall and winter. Corn was a minor crop except in strata 1 and 2, whereas sorghums and millets were cornmon in all areas. The amount of land in annual set-aside varied among years and influenced the amount of summer fallow, millet, tame sunflowers and other non-target crops planted on acres diverted from wheat, corn, and sorghum production. �88 Table 9. 1989. Major vegetation types (%) among strata, eastern Colorado, Vegetation early spring Stratum 1 2 3 4 5 6 13 .6 26.0 21.1 58.6 51. 7 22.7 27.8 3.8 4.8 6.1 0.3 1.6 7.3 4.5 Small grains 32.1 26.9 25.2 8.6 10.4 28.2 24.2 Stubble Standing 23.1 13.7 3.8 1.5 5.9 17.3 12.4 2.5 6.1 11.2 2.9 2.0 4.4 5.3 Conservation Reserve % of total 7.7 9.1 19.2 23.3 20.4 13.3 14.0 9.0 12.3 24.4 56.2 42.2 17.3 19.5 3.7 4.9 6.8 4.7 7.9 3.9 5.3 11.4 7.2 2.5 1.7 4.8 Beans/beets 0.8 0.5 Sunflowers 0.2 0.6 Alfalfa 0.7 0.2 Unfarmed 0.4 Rangeland Fallow Mulched % of cropland Sorghum/millet Corn n samples Wheat Stubble 1,390 0.1 3.8 0.3 0.3 0.4 0.3 0.9 1.1 0.4 1,498 1,292 and Green Wheat Comparisons 408 1,000 1,012 6,600 with CRP The VOR of wheat stubble in early spring varied widely among years (Table 10). This was primarily because of the influence of varying precipitation patterns on the growth of green wheat during the preceding growing season. Sampling among strata during 1989 and 1990 showed that highest VOR indices were within strata 1 and 2 in the northeastern corner of Colorado. The quality of wheat stubble exceeded that of CRP only during spring 1989 (Table 10). In 1989, wheat stubble was much more abundant than CRP within stratum 1 whereas CRP was equal or more abundant than stubble in most other strata. Mulched (undercut) wheat stubble consistently yielded extremely low VOR indices. �89 Table 10. Average visual obstruction readings (VOR-dm) for standing mulched stubble, and CRP (pre-greenup) in eastern Colorado, 1988-90, stubble, 1992. Cover type 1983 1989 1990 1992 Standing stubble 0.40 0.62 0.30 0.40 Mulched stubble 0.17 0.14 0.13 0.17 CRP (pre-greenup) 0.39 0.36 0.47 0.68 Progression of preliminary stubble tillage was monitored along an established route in northeastern Colorado during 1988-90 and 1992 (Fig. 5). Data show tillage commenced in April and was usually complete by late May. The VOR of green wheat consistently exceeded that of CRP by late April during the study and was far greater by late May (Fig. 6). Wheat fields occupied considerably more acreage than CRP in all strata except 4 and 5 during 1989 (Table 9). Wheat harvest usually started in late June in southeastern Colorado and in late June to early July in the northeastern Colorado, at which time the nesting security of this cover was reduced. Conservation Reserve Association with Wildlife Distribution Conservation Reserve Program acreage distribution within eastern Colorado, based on data through the 1st 4 CRP signups, was compared with the distribution and general density classifications for ring-necked pheasants, northern bobwhite, and scaled quail. Ring-necked Pheasant.-- Analysis indicated 43.8% of the CRP in eastern Colorado was outside pheasant range, whereas 32.6% occurred within range classed as remnant, or it contained only scattered occurrences of the species. About 16.9% of the CRP was within range classed as low, 2.5% was within fair range, and 4.2% was within moderate to good pheasant range. Since remnant range in eastern Colorado provided almost no hunting opportunity, only 23.6% of the CRP acreage occurred within huntable range. Most pheasant range rated moderate to high in 1982-83 had deteriorated to low to fair status by 1988-92. A higher proportion of land, committed to the CRP within latter signups, was within low or better pheasant range. However, the amount of land committed to the CRP during more recent signups was much less than that during the 1st 4 signups; thus, overall percentages would not markedly change. VOR cover quality in most NERC fields was very poor to poor for pheasant nesting at study initiation (Fig. 4). However, cover conditions gradually improved during both pre-greenup and nesting sample intervals. Winter cover on CRP fields, rated for its over-winter 1989-90 value for pheasants, was predominately poor quality. Among 140 fields, over 74% were rated poor, 13% were fair, and <9% were good to excellent. Anticipated quality during winter 1990-91 was estimated during summer sampling and showed similar results. However, some fields containing tall annuals, were mowed �90 100 l 1992, / ,.,. ."" ,", •• • ,./ .. • " •• / I" • • 080 ' W ~ • /1 9 0.. /, J- 60 Z W / o 0: W / 0..40 / " I , ~' " I " , , , 20 •• • •• •• • •• • •• .. •• • • •• ,1990 ..• 1989 ,.;' o I 10 15 20 25 30 APRIL Fig. 5. Progression of wheat stubble northeastern Colorado, 1988-92. 5 10 15 20 25 30 MAY tillage along an establ ished route, �91 5 . . .. . . . . . .. . . .. ,: 1990: .: / ,,/ /' / I :/ ,, I .Y /.: 1992 ~/ , I , I /, ..: , I .. / .. : / .. / / / / / .. .. / / / ..... o~--~~~--~--~--~--~--~--~--~--_._ 10 15 20 25 30 5 APRIL Fig. 6. 1988-92 . Progression 10 15 20 25 30 MAY of wheat growth (VOR-dm) in eastern Colorado, �92 subsequent to sampling. Total NERC and additional cost-shared fields were compared with those within pheasant range (Table 11). Winter cover conditions during 1989-90 and 1990-91 were rated slightly better within pheasant range than within total samples. A higher proportion of the CRP fields within pheasant range contained food within 0.25 miles and were associated with good wintering sites «2 miles) (Table 11). A higher percentage of the fields within pheasant range also contained or were associated with pheasant brood habitat. Table 11. Winter and brood conditions on, and associated rated for pheasants, eastern Colorado, 1990. with, CRP fields Percent of CRP fields within Variable Total NERC Sample size, fields NERC within Additional CDOW pheasant range cost share CDOW within pheasant range 105 76 35 1989-90 cover on CRP Poor Fair Good Excellent 80.0 15.2 2.9 1.9 77 .6 15.8 3.9 2.6 57.1 22.9 17.1 2.9 54.5 24.2 18.2 3.0 1990-91 cover on CRP Poor Fair Good Excellent 76.2 20.0 3.8 0.0 72.4 25.0 2.6 0.0 54.3 28.6 14.3 2.9 54.5 27.3 15.2 3.0 Food <0.25 mi. 58.1 72.4 80.0 84.8 46.7 63.2 54.3 54.4 Brood cover on CRP 46.7 53.9 91.4 87.9 Brood cover <0.5 mi. 78.1 88.2 97.1 97.0 33 Winter Cover Conditions Good wintering within 2 mi. area Summer Brood Conditions Pheasant crowing indices along established census routes in eastern Colorado were contrasted to the percentage of CRP adjacent to the route (Table 12). Meaningful pre-CRP data were not available and only limited census was conducted during the CRP interval.. However, no evidence of marked increases in breeding pheasants in association with routes adjacent to CRP was evident. �93 The Two Buttes Route in northern Baca County was associated with the highest conversion to CRP but pheasant populations did not increase subsequent to implementation of the program (Table 12). (i Table 12. Ring-necked pheasant crowing count indices calls/station) and the percentage of Conservation Reserve along established census routes', eastern Colorado 1986-92. Census route (County) Stratum Amherst-Paoli (Phillips) Julesburg-Amherst (Sedgewick) Julesb.-Crook (Sedg-Log.) Holyoke-Fleming (Phil.-Log.) ;";ages -Haxcun (Phil.-Yuma) Fleming-Leroy (Logan) Eckley-Yuma (Yuma) Lonestar-Akron (Washington) Sterling Proctor (Logan) Ft. Morgan-Narrows (Morgan) Platner-Elba (Washington) Ida1ia-Joes (Yuma) Bonny Reservoir (Yuma) Burlington North (Kit Carson) Smokey Hill (Kit Carson) Lamar-Bristol (Prowers) Las Animas 4 (Bent) Two Buttes (Baca) % CRP 1986 1987 Crowing count index 1988 1989 1990 1991 1992 1 2.4 16.8 23.1 12.5 18.3 1 o 18.3 13.4 15.4 1 <l. 0 14.1 19.1 18.3 1 3.3 13.4 7.0 1 5.3 10.4 16.7 9.2 8.1 8.4 18.2 1 7.7 8.1 9.0 7.8 5.4 5.0 21.8 1 3.3 16.4 23.8 15.7 7.0 11.1 1 16.3 5.8 10.4 14.5 1 <l.0 12.7 14.7 17.7 15.5 19.8 12.6 6.9 11.0 15.2 6 2.7 9.8 8.7 8.7 6.0 2 3.3 7.1 2.5 7.2 2 2.5 8.8 10.2 10.5 6.1 2 0.9 4.1 3.0 6.3 2.7 7.6 2 3.2 6.2 5.6 12.1 6.3 7.6 2 0.5 l.3 5.5 2.9 5.4 3 5.4 11.5 11.9 10.6 &5 l.7 16.0 3.0 4.0 2.9 3 33.3 7.6 6.8 3l.3 2l.0 3 12.0 9.3 2.6 39.0 19.2 5.8 14.7 11.3 13.8 12.6 1.4 25.2 5.8 5.7 7.0 4.2 11.8 • The percentage CRP was determined using a 1 or 2-mi. wide strip bisected by the census route. �No evidence of increased pheasant numbers was evident along several breeding bird survey routes (Table 13) in eastern Colorado. Average pheasant indices along the Briggsdale and Two Buttes routes, which both possessed much higher percentages of CRP than other routes, declined from pre-CRP to CRP intervals. Table 13. Average birds/route observed during pre-CRP (1968-86) and CRP (1987-92) intervals along Breeding Bird Survey routes in relation to the amount of CRP adjacent to the route, eastern Colorado. Species PreCRP CRP Percent CRP 005 ? 007 012 Route" 013 014 020 027 028 2 1 3.6 2.2 1.9 3.4 18.3 Pheasant PRE CRP 15 8 125 54 <1 5 5 19 26 <1 <1 <1 85 19 ~1ourning dove PRE CRP 72 72 94 ll8 60 25 82 95 95 130 109 68 33 49 264 123 W. meCldowlark PRE CRP ll4 103 174 152 165 115 136 128 162 168 166 115 87 199 577 226 Horned PRE CRP 236 253 254 254 29 15 197 142 156 179 223 178 192 237 432 329 Lark bunting PRE CRP 268 227 127 116 42 24 224 145 89 90 232 131 54 12 357 160 Grasshopper sparrow PRE CRP 2 16 54 46 15 17 33 63 8 7 2 7 52 45 Cassin's sparrow PRE CRP 2 8 42 87 123 108 45 67 82 Brewer's sparrow PRE CRP 3 7 <1 2 Lark sparrow PRE CRP 1 5 Red-winged blackbird PRE CRP 2 5 Vesper PRE CRP lark sparrow Dickcissel Swainson's hawk <1 2 <1 1 <1 <1 <1 10 9 8 10 4 19 1 21 65 75 20 50 26 72 16 22 6 <1 1 1 77 43 T T 4 4 12 <1 2 4 3 II <1 1 3 PRE CRP PRE CRP 6 2 <1 <1 <1 <1 3 3 2 2 4 6 3 3 4 4 6 4 2 3 �95 Table 13 Species PreCRP CRP Percent CRP 005 ? Ferruginous hawk PRE CRP <1 <1 N. Harrier PRE CRP 1 2 PRE CRP 3 4 Burrowing owl Scaled quail PRE CRP ~. bobwhite PRE CRP Upland sandpiper PRE eRP Longbilled curlew PRE CRP Totals PRE CRP % Change Route" 013 012 2 1 3.6 2.2 <1 1 <1 <1 <1 <1 2 <1 <1 <1 0 1 <1 <1 0 0 <1 2 1 3 12 17 4 5 2 1 2 8 1 0 1 7 1 014 2 3 <1 <1 <1 T 020 028 007 1 1 1 1 T 2 <1 <1 1 027 1.9 3.4 18.3 5 2 <1 6 2 <1 16 1 724 710 907 825 327 245 703 618 640 781 884 630 442 623 1,963 997 -1. 9 -9.0 -25.1 -12.1 +22.0 -28.7 +40.9 -50.7 012 Briggsdale (Weld Co.); 007 = Fleming (Logan/Phillips); • Route 005 013 Last Chance (Washington/Adams); 014 = Abarr (Yuma); 020 = Boyero 027 = Gilpin (Bent), and 028 = Two Buttes (Baca). Kiowa (Elbert) (Lincoln); = Northern Bobwhite.-- Cnly 6% of the CRP acreage was within northern bobwhite range, primarily remnant or marginal habitats. The best bobwhite range was along riparian habitats in Colorado. Therefore, it was separated from the CRP, especially along the South Platte River. Some overlap of marginal bobwhite range and CRP occurred within sandhill rangelands in extreme eastern Colorado. Scaled Quai1.-- About 2.2% of the CRP acreage was within poor scaled quail range, 7.4% was within fair range, 5.5% was within good range, and 1.0% was within range rated as very good or excellent. Thus, 16.2% of the CRP was within or closely associated with scaled quail range. This analysis used townships as the basic unit of measure and scaled quail range was often isolated along rangeland-dominated drainages distributed as fingers of habitat �96 within sections and townships. If a more refined analysis was conducted, the percentage of the CRP acreage within actual scaled quail range would be smaller. If southeastern Colorado scaled quail range was reclassified based on long-term densities, a vast majority would be rated poor or low rather than good range. Mourning Doves.--Considerable nesting activity by mourning doves was obserJed while sampling vegetation during nesting seasons within CRP fields. A comparison of the average number of mourning doves heard and observed along 8 eastern Colorado doves routes during pre-CRP (1976-87) and CRP (1988-92) intervals in relation to the percentage CRP was made (Table 14). These data do not show evidence that CRP impacted indices. The Fort Morgan and MathesonSimla routes, with respective 5% and 3.8% adjacent CRP, increased in both doves heard and seen. In contrast, declines were indicated along both the Las Animas and Cheyenne Wells routes which contained higher percentages of CRP (Table 14). Because of declining hunting pressure on mourning doves in recent years, overall harvest data are not believed usable as an index concerning impacts of CRP. Table 14. Mourning doves heard and seen (i) along 8 census routes during pre-CRP (1876-87) and CRP (1988-92) intervals in relation to che amount of CRP adjacent to the census route, eastern Colorado. Route % CRP Doves heard Pre-CRP CRP Doves seen Pre-CRP CRP Greeley -0 24.3 36.4 38.2 24.0 Trinidad -0 2l. 7 16.6 9.2 4.4 Ft. Morgan 5.0 38.8 56.8 40.8 64.8 Las Animas 5.9 15.6 11.6 22.5 3.4 Matheson-Simla 3.7 47.3 55.0 29.8 38.0 Crowley 4.2 42.5 60.6 49.1 28.0 Yuma l.7 35.8 49.3 6l.0 67.7 15.8 49.3 33.8 35.1 34.3 Cheyenne Wells Mourning dove information from breeding bird survey routes (Table 13) show similar results. The Two Buttes route was associated with a markedly higher percentage of CRP but mourning doves declined about 50% from pre-CRP to CRP intervals. Other routes, associated with minor amounts of CRP showed little overall change. Breeding Bird Survevs.-- Indices of birds observed and heard along 8 breeding bird survey routes in eastern Colorado were averaged for pre-CRP �97 (1968-86) and CRP (1987-92) intervals. CRP adjacent to the routes (Table 13). Data were compared with the percentage However, samle sizes were small. Data for several passerines, raptors and other species were compared. Although the data are weak, there is no evidence that any species was enhanced by conversion to CRP along the route. DISCUSSION This evaluation of the Conservation Reserve Program in eastern Colorado focused primarily on its values for ring-necked pheasants. Pheasants, the primary upland game species within eastern Colorado farmlands, were anticipated as being most benefitted by the program in Colorado and other states in the Great Plains Region. The Colorado Division of Wildlife cost shared directly with farmers in eastern Colorado in attempts to enhance the quality of herbaceous cover for nesting pheasants and prairie grouse and to increase the suantity of woody cover associated with CRP. Findings illustrate that much of eastern Colorado was marginal for both farming and pheasants limiting the value of the CRP for the species. Since pheasants were primarily found within the better farmlands and the CRP was more extensively committed to marginal farmland, an inverse relationship between pheasant abundance and the CRP was noted. One exception was in portions of Baca County where extensive conversion to the CRP occurred within what was Colorado's best pheasant range at the time. However, pheasant populations crashed there at the same time that up to a third of the land was being converted to grass. Sorghums, required as a cover crop for grass establishment when the CRP was initiated, provided brood cover for pheasants during summer and early fall. However, winter to spring seeding (Nov-Apr) flattened much of this cover limiting its value for winter protection. Some sorghum fields remained standing over-winter and provided survival cover to pheasants during the initiation of CRP, but as a whole, winter survival cover was poor. Extensive use of herbicides, primary Glean and 2,4-D, and mowing severely restricted growth and quality of annual forbs and markedly reduced the VOR of herbaceous cover within CRP fields during the first and second years after grass was seeded. As a consequence of these treatments in conjunction with favorable rainfall, perennial grasses established rapidly on most fields. Growth of tall weeds, anticipated as high value for pheasant survival, existed only within a few of the sampled fields and did not persist. Most CRP fields provided poor survival cover for pheasants through winter and survival cover adjacent to CRP was also lacking. This was the most limiting factor to pheasants in eastern Colorado and the primary reason that pheasant populations did not respond favorably to the CRP. Pre-greenup VOR indices within the NERC sample showed that early spring nesting quality on CRP remained very poor to poor throughout most of the evaluation. However, cover quality by late June and July continued to improve on average among NERC fields. Nesting cover quality was determined by several factors including soils, precipitation-evaporation ratios, and species that were seeded. Sideoats grama, western wheatgrass, and blue grama dominated �98 within many fields within strata 3, 4, and 5 in southeastern Colorado. Much of that area lay outside historic pheasant range or was classed as remnant range. Greater evaporation loss of precipitation also impacted vegetation growth. Wheatgrasses were dominant within marginal lands in strata 6 and yielded relatively low VOR indices. Smooth brome and wheatgrasses were used extensively within the loam soils in strata 1 and 2 in the northeastern corner of Colorado. The better soils and moisture there were more influential in increasing VOR indices, whereas the brome and wheatgrasses that were used tended to have a negative impact. Smooth brome established rapidly to become dense, short, root bound, and nitrogen deficient. The Division of Wildlife provided free alfalfa seed but it was not used extensively because of herbicide use and because personnel within the U.S.D.A. Soil Conservation Service discouraged its use. However, among the small sample of fields where it occurred with brome and wheatgrasses, it markedly increased the vegetation VOR during the early summer sample (Table 7). In more sandy soils where the Soil Conservation Service permitted planting of switchgrass and other tall, warm-season species, VOR indices were much greater and observations indicated pheasant use was much higher. However, these were primarily in rangelands sites more suited for prairie grouse than for pheasants. The Colorado Division of Wildlife's cost-share program apparently resulted in increased nesting cover quality as indicated by higher VOR indices. However, only 2.4% of the CRP was impacted and in many instances, similar grass species would have been planted without cost share. Compromises with SCS specifications severely hampered the program. Pheasants, and to a much lesser extent, prairie grouse, would have been the primary benefactors of increased herbaceous cover quality. However, some of the cost share occurred outside of historic pheasant or grouse range and much of it occurred in marginal pheasant range. Efforts sho~ld have centered within the historic better range in stratum 1, to replace smooth brome, but this did not occur. Lack of survival cover in association with this better nesting cover was also a factor resulting in no apparent or detectable pheasant population increase. Hindsight indicates that tall sorghums, planted within small food-cover patches (practice CP-12) should have been incorporated into, and combined with the grass enhancement program to achieve pheasant population increases. The Division of Wildlife did not have manpower, equipment, or a committed longterm funding source to initiate this effort. There is no evidence that the cost-share program affected pra~r~e grouse eastern Colorado. Many of the same warm-season grass species would have seeded in these sandy rangelands whether the Division cost-shared or not payment for large fields dramatically increased program costs. Plantings rangelands enhanced nesting cover quality but reduced interspersion of croplands within rangelands. We lack information as to which was more limiting, the need for additional nesting cover or the need for proximal Some data indicate food may have been more limiting. in been and in food. The CRP reduced interspersion of shrub-covered rangelands and croplands within southeastern Colorado scaled quail range. There, lack of disturbance and annual forbs or grains was considered a primary limitation to scaled quail. Rapid grass establishment in most CRP fields probably was more detrimental than beneficial to this species. However, the CRP probably had little overall impact to the species primarily because there was not major overlap. Northern bobwhite, which were most abundant in riverbottoms, were seldom effected by CRP. �99 Observations indicate mourning doves, among eastern Colorado game species, were most benefitted by the CRP but census route information does not show that this occurred. Additional insight concerning this species is provided by Remington (1992). The CRP provided vast areas of nesting cover for native wildlife as about 18% of the cropland in eastern Colorado was converted to perennial cover. For seme species, such as pheasants, cover quality in most fields was marginal. Possibly, it increased rodent populations increasing the prey base and therefore increased populations of coyotes, kit fox, great-horned owls, and other predators. Breeding bird survey data do not indicate that raptors increased (Table 13). Data and discussion concerning the CRP influence on passerines, based on field research, are presented by Remington (1992). CRP fields were more secure, except for predation, than wheat stubble fields and possessed similar early spring nesti~g quality. Green wheat fields provided much more nesting cover for pheasants and possessed greater VOR indices by the end of April. Wheat harvest reduced security of this cover type by late June or early July but often allowed time for nesting and reproduction without disturbance. LITERATURE CITED Daubenmire, R.F. 1959. A canopy-coverage Northwest Sci. 33:43-64. method of vegetational analysis. Hoffman, D. M. 1965. The scaled quail in Colorado - Range, population, status, and harvest. Colorado Dep. Game, Fish and Parks. Tech. Bull. 18. 47pp. Remington, T. E. 1992. Evaluation of wildlife responses ot the Conservati0n Reserve Program. Colorado Div. Wildl., Final Rep., Fed. Aid Proj. W167-R Apr: In Press Robel, R. J., J. N. Briggs, A. D. Dayton, and L. C. Hulbert. 1970. Relationship between visual obstruction meansurements and weight grassland vegetation. J. Range Manage. 23:295-297. Snyder, W. D. 1984~. Management procedures for northern bobwhites COLorado. Colorado Div. Wildl. Spec. Rep. 56. 22pp. of in eastern ______ . 1984£. Ring-necked pheasant nesting ecology and wheat farming on the High Plains. J. Wildl. Manage. 48:878-888. ______ . 1985. Management procedures for ring-necked Colorado Div. Wildl. Spec. Rep. 59. 53pp. pheasants in Colorado. ______ . 1991. Wheat stubble as nesting cover for ring-necked northeastern Colorado. Wildl. Soc. Bull. 19:469-474. pheasants ______ . 1992. Sandsage-bluestem Prairie Renovation. Colorado Final Rep., Fed. Aid Proj. W-167-R Apr. In Press Div. Wildl., . , ~= - ----..._ \ Prepared by: ~j_j.l..\r\Q, -1), 'fJVlli(t Warren D. Snyder -/ Wildlife Researcher C - (J LD \ ~~ in ��101 JOB PROGRESS REPORT State of: Colorado Project: W-167-R Work Plan: 21 Job Title: Period Author: Job Warren 01 January Bird Research 8 Increasing foods for wildlife eastern Colorado Covered: Personnel: of Wildlife Upland within through 31 December the rangelands of 1992 D. Snyder J. L. Mekelburg, J. W. Moore, and W. D. Snyder, Colorado Division ABSTRACT Evaluations were conducted to determine establishment, survival, growth, and seed production of selected herbaceous species as food sources for avifauna within the Tamarack Prairie in eastern Logan County. Eight rangeland sites each <1 ha, were tilled in early spring in preparation for planting. Each site contained a randomly selected block of seeded annuals, a block of seeded perennials (primarily legumes), and disturbance-tillage blocks within 5 of the 8 sites. An undisturbed control accompanied all sites. Planting was initiated in late April but most was delayed until late May because of extremely dry April-May weather. Monitoring was conducted through summer, fall, and winter 1992. Plentiful moisture after planting resulted in good to excellent establishment, survival, and growth of most species. Annuals had fair to excellent seed production with considerable variation among species and sites. Nearly all perennials on all sites survived and retained green vegetation into late fall. Time of planting prevented seed production of one species and seed quality from wild sources was suspected in the marginal establishment of 2 or more species. Annual forbs within disturbance-tillage blocks produced limited food and good to excellent cover. Consistent use of seeded annuals and disturbance. tillage by numerous ~vifauna was observed from summer into late wint~r ...Passerines ~ete most abundant. _Mournj_-rig. doves (Zenaida macroura) were frequently recorded in late summer and early fall.. Fall-winter use by greater prairie-chickens· (Tympanuchus cupido) at one site was noted and ring-necked pheasants (Phasianus colchicus) were consistently present in fall and winter at 2 locations. Considerable deer (Odocoileus spp.) use of most sites was observed with moderate browsing of alfalfa and other legumes. Kangaroo rats (Dipodomys spp.) were the most abundant rodent and were major consumers of seed-producing annuals in fall. ��103 INCREASING FOODS FOR WILDLIFE WITHIN OF EASTERN COLORADO Warren THE RANGELANDS D. Snyder INTRODUCTION Greater prairie-chickens, lesser prairie-chickens (I. pallidicinctus), scaled quail (Callipepla squamata), northern bobwhite (Colinus virginianus), sharptailed grouse (I. phasianellus), mourning doves, other avifauna, and a few ring-necked pheasants reside in eastern Colorado rangelands, especially in sandhill rangelands. In most instances, their densities are relatively low and all species of prairie grouse are listed as endangered or threatened in Colorado. It is hypothesized that lack of food may be a primary limitation to avifauna in eastern Colorado rangelands. It is possible that a long history of grazing has reduced quantities of rangeland forbs. All the above-listed species are primarily seed eaters as adults and consume the energy-rich seeds of annuals in priority. Prairie grouse also consume green vegetation including legumes if available. Except for mourning doves, the young of all species begin life on a diet of insects. Perennial and annual forbs comprise <10% of the total vegetation species composition within the ungrazed Tamarack Prairie and seed-producing forbs are even less available on most grazed sites. Prairie grouse have not used the Tamarack Prairie extensively even though livestock grazing has not been permitted and the height-density of residual vegetation has increased. Greater prairie-chickens have increased in Yuma county following extensive development of corn and other irrigated crops using center-pivot irrigation systems. These factors provide evidence that food may be limiting in some rangeland situations. There is need to determine what food species are best suited for use in dry rangeland situations in eastern Colorado. Both seed-producing species and those that will yield green vegetation, especially into winter, need to be evaluated. P. N. OBJECTIVE Evaluate the establishment, survival, growth, and seed production of selected herbaceous species as food sources for avifauna within the Tamarack Prairie in eastern Logan County. Ascertain the presence/abundance of selected wildlife during brood rearing (summer) and fall-winter intervals. Assuming objectives land 2 are attained, a potential 3rd objective will be to test their adaption to other rangeland sites in eastern Colorado for scaled quail, lesser prairiechickens, mourning doves, and other species. SEGMENT OBJECTIVES 1. Review literature potential species and contact knowledgeable persons and varieties for testing. 2. Select and prepare 3. Monitor the relative establishment, qualities of tested species. to develop a list of sites for planting. survival, growth, and food producing �104 4. Monitor the presence/abundance of selected wildlife using the test sites. 5. Monitor precipitation and other environmental factors. 6. Prepare job progress report. METHODS Literature concerning avifauna food habits was reviewed and horticulturists and range specialists were contacted to develop a list of plant species and varieties potentially of high food value for wildlife and suited for rangelands in eastern Colorado. Seeds of several species and varieties were obtained using both acquisition and hand collection (Table 1). Initially,S sites, each <1 ha in size were placed linear and perpendicular to prevailing winds among the Dailey and Valent loamy sand soils (Amen et al. 1977) within the eastern part of the Tamarack Prairie. These (3 south of 1-76 and 2 north of 1-76) (Fig. 1) were mold-board plowed in March by J. L. Mekelburg and subsequently tilled with a cultipacker prior to planting. A management directive precluded planting certain exotic species south of 1-76. Therefore, 3 additional plots were prepared by disc tillage north of 1-76 in April. Each of the 5 initial sites contained 3 randomly selected plots: (1) seeded, (2) disturbance tillage, and (3) undisturbed control. The last 3 sites did not contain a disturbance tillage site due to prevailing dry conditions. Within seeded plots, species were separated for planting and evaluation into 2 groups: (1) those of value for seeds (primarily annuals), and (2) those of value for green leafy vegetation during summer brood rearing and fall-winter intervals (primarily perennials). Selected species were planted within each of the 8 sites using randomized patterns. Each seed-producing species was planted in a block containing 10 - 12 rows, spaced 5 - 6 dm apart, and 10 - 12 m long within each site. A push-type garden seeder, equipped with seeder plates adapted for different seed sizes, was used for planting within the 5 initial sites. Efforts were made, based on seed size, plant size, and past experience, to seed at proper rates and depths. Planting by hand started on 28 April (site 2), continued in early May (sites 3 -5), and was completed on 14 May (site 1). Species blocks were marked with metal pins and labeled with metal tags for identification. Species for production of green leafy vegetation were planted within linear belts approximately 6-m wide and as long as the tilled site was wide (approximately 40 m) on each site. A drill equipped with double-disk furrow openers (1.8-dm spacing between openers), 2.5-cm band attachments, packer wheels, and an alfalfa seed box was used. Planting was conducted on 26 and 28 May. Species blocks were marked with metal pins and labeled with metal tags for identification. Within sites 6 - 8 (Fig. 1) all species were seeded by this method. A shop vacuum, powered by a portable generator, was used to remove surplus seed from the drill box before planting the next species. Disturbance tillage was established in plots at least 10 x 50 m [0.05 hal) strips at each site perpendicular to prevailing winds within sites 1 - 5. Because of dry soil and weather conditions prevailing at the time and concern for subsequent wind erosion, wild sunflower, (He1ianthus spp.) was drilled into approximately one-half of each disturbance site (100% of the disturbance �:, :, :, :, LEG ND 0 N 2ICm 1 tiCAL t 7 12 11 18 15 SITE 3D ,~ 11: 18 I CK PRAIRIE I .' I "'; --.".------: r I 24 I 111 I I 20 21 R4.-W- - - - - - - - ,. SITE 1 SI'lj2 0 - - - - - I -----il - R411W I •.... o Fig. 1. Locat.Lonand sequence of rangeland food sites County,Colorado, 1992. within the TamarackPrairie, Logan VI �106 Table 1. Vegetation species and varieties planted within 8 test sites, Tamarack Prairie 1992 Site Triticale 'W.'W.millet Grain sorghum 'Wild sunflower Euphorbia spp. I. bundleflower Sweet pea Common ragweed Proso millet Bee plant Ladak alfalfa Common vetch Sweet clover Hairy vetch 5 Grain sorghum F. penny cress Common ragweed Lewis blue flax Triticale Proso millet Bee plant 'Wild sunflower 'W.'W.millet Euphorbia spp. Croton Safflower Small burnet 'Wild flower mix Calif. poppy Per. sweet pea Alf., Spreador II Sainfoin Hairy vetch Common vetch Ladino clover Cicer milkvetch 4 3 2 1 Triticale Proso millet Grain sorghum Common ragweed 'Wild sunflower Sweet pea Bee plant Euphorbia spp. Field pennycress 'W.'W.millet Common vetch Sweet clover Alf., spreador II Hairy vetch 6 Safflower Grain sorghum Triticale Proso millet Common ragweed Annual buckwheat 'W.'W.millet 'Wild sunflower Calif. poppy 'Wild flower mix Lewis blue flax Hairy vetch Sainfoin Small burnet P. prairie clover Ladino clover Cicer milkvetch Ladak alfalfa Triticale Proso millet 'W.'W.millet Euphorbia spp. Sweet pea Grain sorghum Common ragweed 'Wild sunflower Bee plant Leadplant Common vetch Sweet clover Ladak alfalfa Hairy vetch 7 A. buckwheat P. Pro clover L. blue flax 'W. flower mix Calif. poppy 'W. sunflower 'W.'W.millet Common ragweed Proso millet Safflower Grain sorghum Triticale Hairy vetch Sainfoin Small burnet Ladino clover Cicer milkvetch F. Pennycress 'W. Sunflower Safflower C. ragweed 'W.'W.millet Proso millet Grain sorghum Triticale Calif. poppy L. blue flax 'Wild flower Small burnet Bundle flower Alf., spread. Common vetch Sweet clover Ladak alfalfa Sainfoin Ladino clover Hairy vetch P. Pro clover A. buckwheat 8 A. buckwheat Triticale Grain sorghum Safflower Proso millet C. ragweed 'W.'W.millet 'W. Sunflower L. blue flax Hairy vetch Sainfoin Small burnet Ladino clover C. milkvetch �107 tillage area within site 3). Disturbance tillage plots were not established on sites 6-8 due to concern for wind erosion. Transects southwest to compare vegetation within controls were placed approximately of each of the test sites and parallel to the site. 30 m The relative establishment, survival, growth, and food producing qualities of species within seeded and disturbed sites were monitored periodically from June through October. Within plots planted for seed production, post-seeding inspections were conducted to determine status as (1) no establishment, (2) established but failed to survive, and (3) established and survived. Canopy cover was used to measure establishment, survival, and growth. A point-frame sampling procedure (Floyd and Anderson 1983) that was initially proposed within the Program Narrative for sampling was omitted due to limitations in time and manpower. A 0.5 x 1.0 m Daubenmire frame was substituted which permitted more rapid ocular estimation of vegetation (Daubenmire 1959). Sampling was replicated 12 times per plot at 1-2 pace intervals (depending on plot size) in diagonal direction across each plot. Sampling was conducted between 28 August and 3 September. Vegetation was classified as the percentage of seeded species, competing annual forbs, annual grasses, perennial forbs, perennial grasses, bare ground, and dead vegetation. This procedure was used within seeded annuals, seeded perennials, disturbance tillage, and control plots on all sites. a Height (dm) of seeded vegetation and total vegetation were sampled at the same time within the Daubenmire plots. Seed production within all plots was rated as: 0) - no production; 1) - low production; 2) - moderate production; and 3) - high production, based on the expected potential of the species to yield seed under favorable growing conditions. Wildlife use of seeded annuals, seeded perennials, disturbance tillage, and control plots was sampled during periodic visits to the 8 sites from mid-June 1992 through March 1993. Wildlife presence or absence was determined, categorized by wildlife group: gallinaceous, passerines, mourning doves, and other avifauna (data were listed by species when possible). Abundance, when present, was delineated as ~ 5, and> 5/species or group. Narrow «2 m), disturbed soil buffers were maintained around seeded and disturbed tracts using a tractor-mounted rototiller in August. Observations of tracks within the buffer strips were recorded during periodic visits from mid-June through March to aid in ascertaining use by deer, rabbits (Lagomorpha), rodents (Rodentia), and gallinaceous birds. Establishment, survival, and food producing problems by species and variety, occurring because of depredation by deer, rodents, rabbits, or other wildlife, were documented. Automatic precipitation recorders were used to monitor precipitation. Gauges were placed at sites 2 and 5 in early spring. Precipitation during winter months was obtained from U.S. Weather Bureau records from nearby recording stations. �108 RESULTS Acquisition of Seed Seed for most species (or varieties) was purchased from several seed companies. Seed from snow-on-the-mountain (Euphorbia marginata), texas croton (Croton texensis) , field pennycress (Thlaspi arvense), and Rocky Mountain bee plant (Celome serrulata) were collected by hand-stripping in fall 1991. Hand collection usually did not yield enough seed to adequately seed more than one or two plots. Establishment of satisfactory stands with hand collected seeds was more difficult than with purchased seeds. Site Preparation, Planting, and Weather Sites 1 - 5 (Fig. 1) that were plowed and subsequently tilled using a cultipacker in March 1992 received little precipitation to aid in preparing a moist, firm seed bed prior to planting. Dry weather persisted in northeastern Colorado during April and May 1992 and delayed initiation of seeding (Table 2). Wind erosion occurred within some plots even though they were in narrow strips perpendicular to prevailing winds. Sites 6 - 8 (Fig. 1), that were disked and harrowed in mid-April, retained enough residual vegetation to prevent significant erosion. However, disking, as opposed to plowing, failed to kill sandsage (Artemisia filifolia), prickly pear cactus (Opuntia spp.), and some deep-rooted grasses. Table 2. Precipitation (in.) received in the vicinity of the Tamarack Prairie through 1992 in relation to the long-term average. Month 1.10b 0.49b 2.lP 0.22 0.96 5.74 1.32 2.57 trace 0.78 0.37 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Totals • b c South Tamarack 15.66 North Tamarack 1.10 0.49 2.11 0.18 0.67 5.74c 2.75 4.74 trace 1.03 0.37 19.18 Long-term i· 0.32 0.35 0.78 2.02 2.55 2.52 2.09 2.04 1.37 0.99 0.44 0.49 15.96 Based on U.S. Weather Bureau records during a 44-year interval . Data averaged between Sterling and Sedgwick stations. Rain gauge destroyed by hail. Precipitation believed < 5.74 inches. Rainfall averaged 8.43 inches for Sterling and Sedgwick weather stations. �109 Soil conditions were dry when planting of annuals and a few perennials, using the hand seeder, was initiated at site 2 on 28 April. Hand seeding was completed on sites I, 3, 4, and 5 between 4 and 14 May, again, into dry soil. Limited emergence of millet, triticale, and sorghum was noted within Site 2 by mid May but no significant establishment occurred on the remaining plots until after rainfall was received during late May and early June. Planting of perennials (and some annuals) on 26-28 May was conducted using a tractor and drill after light rains had moistened the top soil. Rains in early June resulted in excellent establishment of all perennials and most annuals. Subsequent rains through the summer enhanced survival, growth, and seed production. A hail storm in June, containing large hail, lightly to moderately damaged test vegetation in sites 4 and 5 and destroyed the collecting portion of the rain gauge on site 5. As a consequence accurate monitoring of precipitation for June was not possible. Less rainfall was received on sites 6 - 8 than on sites to the east (Fig. 1) but a precipitation gauge was not present to monitor differences. Weeding by hand (hoe) was conducted several times in June on all sites to reduce the number of competing large annuals such as Russian thistle (Salsoli kali) , lambsquarter (Chenopodium album), and pigweed (Amaranthus spp.) which tended to dominate the plots. Smaller annuals such as Texas croton were not removed. Establishment, Growth, Survival, Seed Production, and Use by Species Nearly all planted species established satisfactory to dense stands after precipitation was received. Exceptions were Rocky Mountain bee plant, snowon-the-mountain, and Texas croton which were hand-collected from wild sources. Buckwheat, triticale, and sweet pea (Lathyrus polymorphus) did not establish well from purchased seed. The number of test plots planted to each species, percent canopy cover for planted species, competing annual and perennial forbs, competing annual and perennial grasses, and bare ground per test plot varied (Table 3). Average height of planted species, overall plot height (including all species) and the seed production rating for annual (and some biennial species) also varied. These data provide an overview of the relative food producing qualities of the planted species. White Wonder Millet. -- This species established dense growth, grew to nearly 1 m in height and produced abundant seed (Table 3). Rodents, primarily kangaroo rats and passerines were the primary consumers from early fall (when it matured) to winter, but considerable seed remained on standing stalks into winter. Rodents pulled about one-half of it down by mid November. Ringnecked pheasants fed on this species extensively during winter. It was not significantly consumed by deer. This millet matured later than proso millet and contained much smaller seed. Proso Millet. -- This early maturing (late summer) variety established dense growth and yielded abundant seed at 4 - 6 dm (Table 3). It received extensive use, primarily by rodents and passerines, and was nearly all consumed by I September. Mourning doves were frequently observed feeding within these plots. Rodents pulled the stalks over to obtain the seed. �110 Table 3. Canopy cover (%) for tested species and varieties, disturbance tillage, and vegetation within controls, Tamarack Prairie, Colorado, 1992. i Vegetation/ No. Cover Variable Plots W.W. Millet Occur. 8 ~ SE Proso Occur. 7 ~ SE Sorghum Occur. 8 ~ SE Triticale Occur. 7 ~ SE Safflower Occur. 5 ~ SE Sunflower Occur. 8 ~ SE C. Ragweed Occur. 7 ~ SE Buckwheat Occur. 4 ~ SE Calif. Poppy Occur. 4 ~ SE Wildfowers Occur. ~ SE 4 Bare Planted grnd. species Canopy Cover Height(dm) Annual Peren. Peren. Annual forbs forbs grass grass 8 21.7 3.8 8 61.5 4.5 8 15.8 3.7 1 0.2 1 0.3 0 7 29.1 3.7 7 58.6 6.5 6 9.1 3.2 4 0.9 0.5 3 0.4 1.2 1 0.2 8 20.6 3.3 8 56.1 4.3 8 20.2 3.6 3 0.8 0.7 6 2.2 0.9 0 7 38.9 4.5 7 26.9 4.8 7 26.8 6.4 5 1.8 0.8 6 5.6 1.7 5 30.3 5.1 5 39.8 7.2 5 18.1 4.5 4 2.8 1.3 5 7.2 1.8 8 17.8 2.7 8 69.3 4.5 8 8.8 2.2 4 0.8 0.4 5 3.3 2.5 6 8.9 4.4 8 57.2 11.8 6 29.4 6.7 3 0.8 0.3 3 3.2 4.0 4 35.2 7.2 4 31.9 8.3 4 22.8 2.8 2 2 1.1 4 7.9 4.0 3 18.2 12.9 4 61.4 19.6 2 9.9 8.1 2 1.7 2.9 2 6.4 11.1 2 1.1 1.5 4 20.0 6.1 4 56.8 16.2 4 13.7 5.2 3 2.5 1.5 4 6.8 3.7 1 0.2 1 0.1 3 0.5 0.3 Planted species Seed Total prod. veget. rate· 9.5 10.1 2.9 5.8 0.9 4.3 1.0 2.9 0.1 5.8 0.7 6.2 0.6 2.3 0.2 6.1 0.6 6.5 0.7 2.5 0.2 4.9 0.8 5.4 0.5 2.8 0.2 14.8 1.4 14.5 1.4 2.9 0.1 9.8 1.4 10.0 1.1 1.0 3.0 0.4 3.7 0.1 2.0 0.4 2.7 0.5 3.2 0.6 2.3 0.5 3.9 1.3 4.5 1.0 2.3 0.2 �111 Table 3 cont. i Vegetation/ No. Cover Variable Plots Alfalfa Occur. (spreador II) ~ SE 4 Alfalfa (ladak) 4 Occur. ~ SE Hairy vetch Occur. 7 ~ SE Common vetch Occur. 3 ~ SE Cicer milkvetch Occur. 3 ~ SE Sainfoin Occur. 5 ~ SE Sweet clover Occur. 4 ~ SE Purple pro clover Occur. 3 ~ SE Ladino clover Occur 3 ~ SE Small burnet Occur. 5 ~ SE Lewis blue flax Occur. 5 ~ SE Bare Planted &rnd. species Canopy Cover Height(dm) Annual Peren. Peren. Annual forbs forbs &rass &rass 4 14.7 4.4 4 62.1 9.1 4 17.6 4.0 3 2.4 1.1 3 2.9 0.5 4 32.2 5.5 4 44.5 14.0 4 14.6 8.5 2 1.6 3 10.2 5.0 6 15.9 6.8 7 54.8 10.9 7 21.5 5.7 4 2.7 2.6 6 5.0 2.5 3 12.1 10.4 3 70.8 4.1 3 15.8 6.7 1 0.8 1 0.4 3 30.7 10.6 3 29.3 14.2 3 22.8 11.2 3 3.6 1.5 3 13.2 5.7 1 0.4 5 26.7 5.5 5 27.1 7.8 5 29.5 4.2 4 5.3 2.4 5 11.0 3.2 2 0.4 4 16.3 7.5 4 65.7 15.5 4 14.6 7.1 1 0.1 3 3.1 1.6 1 0.3 3 37.0 7.0 3 25.7 11.7 3 17.2 9.0 2 3.9 1.7 3 12.7 4.7 3 3.6 1.9 3 12.5 4.0 3 45.6 9.9 3 33.3 2.7 2 2.1 2.5 3 5.0 4.4 1 0.3 5 26.2 6.0 5 42.4 10.5 5 17.0 8.5 5 4.4 0.4 5 9.6 1.9 2 0.4 0.3 5 22.7 7.8 5 41.4 12.3 5 20.3 8.2 3 1.3 1.2 5 10.5 3.7 1 2.2 2 0.3 1 0.3 Planted species Seed Total Prod. ve&et. rate' 2.0 0.5 1.7 0.6 1.3 0.3 2.2 0.5 2.2 0.6 3.4 0.6 3.8 0.8 4.0 0.9 0.8 0.2 2.1 0.6 1.2 0.2 3.5 0.6 1.2 0.4 2.0 0.4 1.9 0.7 3.0 0.7 1.2 0.6 3.3 0.2 1.1 0.3 2.9 1.2 2.0 0.7 4.0 0.9 Sweet pea ~ 1 10.8 22.5 9.6 44.2 5.8 7.1 2.4 2.7 Bundelflr. ~ 1 10.8 70.8 10.4 2.1 5.0 0.8 2.8 2.8 0 1.9 0.3 2.2 0.8 �112 Table 3 cont. ~ Canopy Cover No. Vegetation/ Cover Variable Plots Dist. till. Occur. 5 ~ SE Rangeland (control) Occur. ~ SE 8 Height(dm) Bare Planted Annual Peren. Peren. Annual grnd, sl!ecies forbs forbs grass grass 5 0 5 4 2 1 12.4 80.9 4.4 1.8 0.1 2.3 2.5 0.6 1.6 8 24.7b 1.5 o 8 8 6.4 1.6 4.0 2.6 8 64. 7~ 4.0 Seed Planted Total prod. sl!ecies veget, rate· 11.0 2.0 2.5 0.3 3.3 0.1 0.3 1 0.2 Seed production was rated on an ascending scale from 1 to 3 based on the estimated capability of the species to produce seed. Most legumes did not produce seed in 1992. b Includes bare ground and dead vegetation. ~ Includes 11.3% sandsage and 1.3% cactus canopy cover. WGF Grain Sorghum. -- This was a short (6 - 8 dm tall) "wild game food" variety of grain sorghum. It matured in mid fall yielding moderate amounts of seed and remained standing into winter. Some seed heads did not fill completely. Pheasants, present in sites 4 and 5, used it extensively in late fall and winter. Rodent and deer use was low. Triticale. -- This wheat-rye cross established poor stands in all plots. Under favorable moisture conditions it should be planted in early April so time of planting influenced stand establishment. It produced considerable grain which matured in late summer. Most was pulled down and consumed by kangaroo rats. Volunteer establishment from scattered seed was noted in fall providing green vegetation into winter. Safflower. -- This spiny-tipped species established fair (open) stands, grew to about 5 dm, matured in early fall, and retained seed into winter. Pheasants learned to extract the seeds from the standing seed heads and consumed all that was present. Minor feeding by rodents occurred. It remained standing through winter but provided marginal protective cover when planted alone. Wild Sunflower. -- This species established satisfactory stands, attained excellent height (1.5 m), and yielded good seed production. Small passerines, primarily sparrows (Spizella spp. and others) were consistently present from summer through winter within sunflower and disturbance tillage plots. Sunflower ranked first in standing over-winter and cover and retained much of its seed into winter. Common Ragweed. -- This species, while related to western ragweed (perennial) and giant annual ragweed, which grow in eastern Colorado, was not resident to the area. It established and grew slowly. Because of late maturity the seed probably was not fertile. Seed production was low and no �113 use was noted into early winter. It grew to about 1 m and provided good cover. Earlier planting might yield more satisfactory results, however, Colorado's growing season may be too short. Seeds from resident giant annual ragweed are preferable if a source could be found. Annual Buckwheat. -- This species was planted in mid June in plots where bee plant and others did not emerge. It grew and matured rapidly and yielded seeds with high oil content, and like other Polygonums, was highly preferred by wildlife. Deer, rodents, and insects browsed this species before it reached maturity. It provided little food for avifauna and would have to be planted in much larger areas to overcome the high use by other wildlife. It did not stand well. California Poppy. -- This floral species established in dense stands, grew to about 3 dm, and yielded abundant small seed from mid summer to fall. It bloomed profusely yielding a carpet of brilliant golden orange from mid to late summer. It did not stand well into fall and seed use by wildlife was not recorded. If used, it was primarily by mice and passerines. Wildflower Mix. -- This was a dryland mix of annual and perennial wild flowers however, few if any were native to northeastern Colorado. Several species, primarily annuals, bloomed extensively from mid summer to fall, attained heights averaging about 4 dm, but seed consumption by wildlife was not extensive. Much of the seed was light and chaffy. Rocky Mountain Bee Plant. -- Failure to establish a stand of this native annual forb was the most disappointing aspect of the study. Reasons for this failure remain uncertain. Potentially, the wild seed had not matured enough to yield viable seed for planting. Time, depth of planting, or hard seed may have been factors. Wild stands, to 1 m height, grew well within disturbed rangeland sites in northeastern Colorado during 1992 and produced considerable seed. Insects were an observed problem on individual plants established within the test sites. Maturity was in early fall. Use of its seed by wildlife was uncertain. Snow-on-the-Mountain. -- Only a few plants emerged within test plots planted to this native species. Those that established grew to about 8 dm and produced considerable seed which matured in early fall. Seed was retained into early winter. Additional tests should be conducted if commercial sources can be found. Texas Croton. -- This native annual, commonly called "dove weed", did not establish well within test plots, but grew extensively throughout all test sites and produced abundant seed. Most mourning doves had left northeastern Colorado by the time the majority of croton seed was mature, however, some use was noted. It was probably a factor in holding doves on the test sites into fall. Two doves were flushed from 1 site in mid November. Seed consumption by greater prairie-chickens was noted on site 3 in late November. Consumption by deer, rodents, and passerines did not appear extensive. If commercial seed sources were available, this would be a priority species for use in eastern Colorado rangelands. Field Pennycress. -- This mustard, collected from wild sources, established dense stands. However, typical of mustards, it is an extremely early species and should be planted in March or early April. Because of the �114 late establishment, it did not produce seed. The species produced abundant seed in wild stands during 1992 and matured in mid to late June. Alfalfa. -- Two varieties, spreador II and Iadak , were each planted in 4 test plots. Seedlings emerged in early June in dense stands and grew throughout the relatively wet summer. Spreador II canopy cover averaged greater than that for Iadak , but site differences may have been a factor. Deer grazed most plots from summer to winter but consistently left about 1 dm of growth standing. Use was most pronounced on sites 4 and 5 where it attained the most growth and was close to the riverbottom. Sign indicated prairie grouse fed on alfalfa, other legUmes, small burnet, and sorghum within site 3 during early winter and 3 greater prairie-chickens were flushed from the location in early December. Hairy Vetch. -- This winter annual or biennial established well on all plots. It spread in its decumbant growth form into dense mats within sites 2, 3, 4 and 5. It was not grazed markedly by deer although the blooms were consumed. It retained some green vegetation into early winter. It is not certain that this species will over-winter when planted in spring; however, it is a winter annual. Common Vetch. -- This species appeared to be an annual which produced considerable seed on several test plots. It had a decumbant growth form but broader leaves than hairy vetch. Use by wildlife was relatively minor, primarily by deer. It did not retain green vegetation into winter. Cicer Milkvetch. -- This perennial species established slowly into marginal stands because of its "hard seed" characteristic. Growth in 1992 was limited with <30% canopy cover. Previous observations of the species revealed that it established more slowly than alfalfa, grew more slower and later into summer, and produced larger plants. Its growth form was partially decumbant and it retained green vegetation into early winter. Observed wildlife use in 1992 was relatively minor. Sainfoin. -- This perennial established satisfactory stands, made fair first-year growth, and retained green vegetation into winter. It produced seed heads which, along with some leafy stems, were consumed by deer. Its growth form was decumbant. Although it looked promising, additional years of evaluation are needed. Sweet Clover. -- This fine-seeded biennial established dense stands which survived well and retained green vegetation into winter. Stands may be too dense to yield the tall growth, characteristic of the species during its second year. It sustained limited grazing by deer. Purple Prairie Clover. -- This perennial established well, made limited growth in 1992, produced some seed heads, and retained green vegetation into winter. Plants remained small and received little use by deer or other wildlife. Additional years of evaluation are needed. Ladino Clover. -- This small perennial established good stands of small plants which yielded considerable seed, and retained limited green vegetation into winter. It received relative minor use by deer and other wildlife. �115 Preliminary observations indicate eastern Colorado rangelands. this species probably is not well adapted to Small Burnet. -- Small burnet has not sustained itself for more than a few years in previous plantings in northeastern Colorado. However, first-year results within these test plots were favorable. Plants remained small, relatively decumbant, and green through winter. Seed heads and some green vegetation were consumed by deer and other wildlife from summer into winter. Lewis Blue Flax. -- This perennial established good stands, attained limited growth, produced considerable seed, and retained green vegetation winter. However, no significant use by wildlife was observed. into Perennial Sweet Pea. -- This perennial (Lathyrus polymorphus) was believed to be the native species when planted within 3 test plots, however, it was the domesticated flowering garden variety. It established a satisfactory stand within site 5 where it attained about 22% canopy cover by late summer in competition with Physalis subglabrata, Agropyron smithii, and other vegetation. Blooms and use by wildlife were not observed. Illinois Bundleflower. -- This species was planted in one test plot on site 4 and established and grew well attaining about 70% canopy cover. Its growth form, where observed along the nearby riverbottom, was ascending. Within the test plot it was browsed extensively by deer and by late summer it averaged 2.7 dm in height with no seed production. Eastern Colorado rangelands probably do not receive enough moisture to sustain this species, unless it is cultivated. Disturbance Tillage. -- Modified disturbance tillage tracts were established on sites 1 through 5. Space did not permit establishment on sites 6 through 8. Wild sunflower was drilled into about one half of sites 1, 2, 4, and 5 and all of site 3 during late May because of prevailing dry weather and concern for erosion. Sparse to moderate stands of sunflowers established, grew to about 1.5 m, and produced considerable seed. Cover and height was relatively poor within site 1 where sunflower was the primary species. Texas croton established naturally within site 1 and produced considerable seed. Sunflower (seeded), in relatively tall, dense stands, dominated within site 3. Lambsquarter, pigweed, and Russian thistle dominated within sites 2, 4, and 5 although sunflower, Texas croton, and other species were present. Annual forbs averaged about 80% canopy cover and grew to 1.1 m. Extensive use by passerines and other wildlife was noted. Rangeland Controls. -- Summer rains stimulated growth of sunflower, pigweed, Russian thistle, horseweed (Conyza canadensis), and other annuals resulting in about 6% canopy cover by annual forbs. Growth of most, in competition with perennials, was small, however, sunflowers averaged over 1 m in height. Perennial forbs (Mentzelia nuda, Sphaeralcea coccinea), and a few others yielded 4% canopy cover. Perennial grasses, sandsage, and prickly pear cactus in combination totaled 64.7% canopy cover. Overall vegetation height averaged about 3.3 dm. Growth and food production by forbs was above average due to favorable summer rains. �116 Wildlife Occurrence Time constraints prevented frequent monitoring of wildlife use, however, limited observations were recorded (Table 4). Use by wildlife was noted at all 8 sites but was least common within sites 1 and 6 through 8. Site use was related to food-cover quality. Disturbance tillage was lacking on sites 6-8 and impacted wildlife use. Passerines, primarily Spizella spp., were consistently present in disturbance tillage, sunflowers, and other annuals within sites 2 through 5 from summer into winter in numbers averaging between 25 and 50. They were present during varied weather, including snow cover, at all times of day. Both food and cover values of the sites were obvious and sunflowers seemed especially important as food. Mourning doves, until they migrated, were common within all disturbance tillage plots and within plots containing other annuals. Two were flushed from site 6 on 11 November 1992. Bedded deer and pheasants were flushed from sites 4 and 5 on several occasions from summer to winter and observation of tracks indicated all sites received some use by deer. Kangaroo rats were common within all sites. Rabbits were not observed but tracks indicated sites 2 and 5 received minor use in winter. Table 4. Wildlife observed within the 8 rangeland food test sites, Tamarack Prairie, late summer 1992 - January 1993. Date 26 Aug 2 Sep 1 l5p·, ld 2p, ld 2 12p, ld l2p, 5d 3 10d Site 4 5 6 lOp, 8d 7 6p, 9d 8 lOp, 13d 11 Nov 5p 0 >50p <20p, 7d 8p, ld lOp 5d 18 Sep <30p, PGtr ld >20p, 7d <40p,4PH 3 Dec 10 Jan 20p 0 <20p >50p <30p <lOOp 22 Nov <30p, PGtr 25p,<5PH 2PH 25p,<5PH 2d 3 deer <20p, 1 dr <30p <SOp <SOp <SOp, 3 PG <lOp, lPH,3dr 3PH 0 5p,3PH <50p,5PH 0 l5p <lOp, 3 dr <lSp p - passerines, d- mourning doves, PG - prairie grouse, PH - Pheasants, tr - tracks. >30p <75p dr - deer �117 DISCUSSION Findings during this first year of study show that numerous annual and perennial species can be established within sandy rangelands in eastern Colorado. Annuals, established for seed production, generally yielded good results. Native species, including wild sunflower and Texas croton, were among the best and probably could be sustained using only disturbance tillage. More work is needed with Euphorbia, Rocky Mountain bee plant, and other wild annuals. A primary problem is seed availability and cost. Mechanical harvesting of wild seed should be considered as an option if small harvest equipment can be obtained. Grain sorghums, millets, and safflower all have potential but probably require annual planting. Safflower is grown commercially in Colorado for oil and bird seed, and millets are grown for bird seed and other uses. They probably have potential in range sites and elsewhere in eastern Colorado for food plots if planted in combination with other taller species. Kangaroo rats are a major problem with many grains planted in rangelands, especially in small plots. Grain sorghum and safflower were not consumed extensively, apparently because of growth form. Little use of millets or grain sorghum by deer was noted. Pheasants consumed grain sorghum and white wonder millet in priority over other foods. Planting several species in mixtures appears the most practical method. Furrow openings within grain drills can be separated to reduce within-row competition among species. One problem is that some species should be planted earlier (or later) than others. This must be considered in selections for mixtures. Several species of perennial and biennial legumes and other species were successfully established. Deer seemed to prefer alfalfa but used sainfoin and some others as well. It may be essential to exclude deer using electric fences in some situations. Rodent use was not considered important. Legumes were believed a factor in attracting and holding a small flock of greater prairie-chickens at site 3 through fall and early winter. Their use of legumes, Texas croton, and grain sorghum was noted based on tracks. Preliminary findings indicate disturbance tillage and planting wild and tame annuals, biennials, and perennials can be successfully used to attract and increase survival of wildlife in rangelands. These techniques, supplemented with brush shelters and thickets could be used to attract and sustain scaled quail and northern bobwhite. Efforts were not directed toward deer, however, they were attracted to the sites for both food and cover. Insects were not a significant problem in 1992 but they may be during some years and more testing and evaluation is needed. The sites were not large enough to provide an abundance of food throughout winter. Pheasants and other wildlife depleted most available food by January within sites 4 and 5. Larger plantings would be needed where pheasants and prairie grouse were to be sustained through winter. �118 LITERATURE CITED Amen, A. E., D. L. Anderson, T. J. Hughes, and T. J. Weber. 1977. Soil Survey of Logan County, Colorado. U. S. Dep. Agric., Soil Conserv. Serv., Washington, D.C. 252pp. Daubenmire, R.F. 1959. A canopy-coverage method of vegetational analysis. Northwest Sci. 33:43-64. Floyd, D. A., and J. E..Anderson. 1983. A new point intercept frame for estimating cover of vegetation. Pages 107-113 in Idaho Nat1. Eng. Lab. Radioecology and Ecology Programs. U.S. Dep. Energy DOE/ID 12098. Prepared by: \.(Q'l\N.,}__; ~,:1rru_don) Warren D. Snyder ,~ ~9ln) �119 JOB PROGRESS REPORT State of: Project: Colorado W-167-R Upland Bird Research Work Plan: 22 Job Title: Upland Bird Research Publications Period Covered: Author: Job _1_ 01 January through 31 December 1992 Clait E. Braun Personnel: Clait E. Braun, K. M. Giesen, R. W. Hoffman, T. E. Remington, and W. D. Snyder, Colorado Division of Wildlife ABSTRACT The following articles were published in 1992: Beauprez, G. M., J. A. Clarke, and C. E. Braun. 1992. Movements, reproductive success, and habitat use by introduced g!eater prairiechickens. J. Colo.-Wyo. Acad. Sci. 24:12. Giesen, K. M. 1992. Body mass of Columbian sharp-tailed grouse in Colorado. Prairie Nat. 24:191-196. _____ , and C. E. Braun. Winter home range and habitat characteristics of white-tailed ptarmigan in Colorado. Wilson Bull. 104:263-272. Hoffman, R. W., W. D. Snyder, G. C. Miller, and C. E. Braun. 1992. Reintroduction of greater prairie-chickens in northeastern Colorado. Prairie Nat. 24:197-204., Schroeder, M. A., and C. E. Braun. 1992. Greater prairie-chicken attendance ," a:t leks"and stability of leki in 'Colorado. Wilson Bull. 10'4;273..,284. _____ , and . 1992. Seasonal movement and habitat use by greater prairie-chickens in northeastern Colorado. Colorado Div. Wi1dl. Spec. Rep. 68. 44 pp. _____ , K. M. Giesen, and C. E. Braun. 1992. Use of helicopters for estimating numbers of greater and lesser prairie-chicken leks in eastern Colorado. Wildl. Soc. Bull. 20:106-113. Snyder, W. D., and G. C. Miller. 1992. Changes in riparian vegetation along the Colorado River and Rio Grande, Colorado. Great Basin Nat. 52:357363. Prepared by C1ait E. Braun Wildlife Research Leader ��121 JOB PROGRESS REPORT Colorado State of: Project: W-167-R Upland Bird Research Work Plan: 26 Job Title: Analysis of Upland Bird Population Trends Period Covered: Author: Job _1_ 01 January through 31 December 1992 Clait E. Braun Personnel: C1ait E. Braun, Kenneth M. Giesen, Richard W. Hoffman, Thomas E. Remington, and Warren D. Snyder, Colorado Division of Wildlife ABSTRACT The following were prepared in 1992. Braun, C. E., and T. A. Artiss. data, 1976-92. _____ , and _ 1992. 1992. Blue Mountain sage grouse harvest Cold Spring Mountain harvest data, 1976-92. _____ , and 1992. Sage grouse harvest data, Eastern Moffat and Western Routt counties, Colorado, 1976-92. _____ , and 1992. Eagle County sage grouse harvest data, 1992. -----, and 1992. Gunnison Basin sage grouse harvest data, 1992. _____ , and 1992. Middle Park sage grouse harvest data, 1992. 1992. Northcentral Moffat County sage grouse harvest data, _____ , and 1992. North Park sage grouse harvest data, 1973-92. and 1992. Piceance Basin sage grouse harvest data, 1977-92. _____ , and 1992. Yampa Area sage grouse harvest data, 1992. and 1976-92. _____ , R. B. Davies, J. R. Dennis, K. A. Green, and J. L. Sheppard. 1992. Plains sharp-tailed grouse recovery plan. Colorado Div. Wi1dl. Denver. 61 pp. �122 Giesen, K. M. 1992. Columbian sharp-tailed grouse harvest data, northwest Colorado, 1976-92. --- 1992. Hoffman, R. W. --- Prairie grouse lek surveys, Logan and Sedgwick counties, 1992. 1992. Blue grouse wing analyses, Durango Area. 1992. Blue grouse wing analyses, Gunnison Basin. 1992. Blue grouse wing analyses, Northeast Region. 1992. Blue grouse wing analyses, Northwest Region. Prepared by r.~ ~ Clait E. Braun Wildlife Research Leader � https://cpw.cvlcollections.org/files/original/a3bd1dafb6edfc965ecfc52a73c3b2d5.pdf 1eb4f61aeb0fff43e3425ba7349d431b PDF Text Text 1 JOB PROGRESS state of REPORT Colorado Project No. Mammals Research W-153-R-6 Work Plan No. 1 Multispecies Job No. 7 Mammals Publication, Editing, Library services Period Covered: Author: and July 1, 1992-June 30, 1993 Jacqueline Personnel: Investigations A. Boss Jacqueline A. Boss and Nancy McEwen ABSTRACT During the segment the following were accomplished: ~ 9 publications were purchased at the request of Mammals Researcher personnel and placed into the Colorado Division of Wildlife Research Center Library collection. ~ 28 free reports and short publications from state or federal agencies or from private sources were located, ordered, and obtained for use by Mammals Research personnel. ~ 19 theses or books were obtained on Interlibrary for use by Mammals Research personnel. ~ 931 individual articles were located and delivered use by Mammals Research personnel. ~ 12 manuscripts by Mammals Research personnel accepted for publication. ~ 5 manuscripts were prepared personnel for peer review. and submitted Loan or as gifts on request were published or by Mammals Research for ��3 MAMMALS PUBLICATION, EDITING Jacqueline AND LIBRARY SERVICES A. Boss P. N. OBJECTIVE To provide a centralized support program for manuscript editing services to facilitate publishing results of research conducted Federal Aid Project W-1S3-R. SEGMENT and library by starf of OBJECTIVE To provide a centralized support program for Mammals Research editing, library, and publishing services so that Mammals Research personnel can be most efficient in publishing results of their research. SUMMARY OF SERVICES Publications Purchased with Mammals Research Funds and Placed in the Research Center Library Allaby, M., ed. 1992. Oxford University Brown, R. D., ed. 596pp. The concise Oxford Press. S08pp. 1992. Buckley, G. P. 1989. Press. 363pp. The biology Biological dictionary of deer. habitat of zoology. New York reconstruction. New York Springer-Verlag. New York Belhaven Dawley, R. M. and J. P. Bogart. 1989. Evolution and ecology of unisexual vertebrates. New York State Museum bulletin; 466. Albany, NY : New York State Museum. 302pp. Dunlap, J. R. 1988. Saving America's University press. 222p. wildlife. Mighetto, L. 1991. Wild animals and American University of Arizona Press. 177pp. Princeton Princeton enviro.nmental ethics. Tucson Rolston, Holmes, III. 1989. Philosophy gone wild: Buffalo, NY : Prometheus Books. 269pp. environmental Schmidt, J. o. 1989. Special Albuquerque : University in the arid Southwest. 152pp. biotic relationships of New Mexico Press. Skalski, J. R., and D. S. Robson. 1992. investigations: design and analysis Academic Press, Inc. 237pp. Techniques of capture ethics. for wildlife data. New York �4 In addition to books purchased with Federal Aid Funds, 28 free reports and short publications from state or federal agencies or from private sources were located, ordered, and obtained for use by Mammals Research personnel. Theses and Books Loan or as Gifts Derr, Obtained on Interlibrary for Use by Researchers J. M. 1990. Genetic interactions between two species deer, Odocoileus virqinianus and Odocoileus hemionus. Dissertation, Texas A & M University, College Station, of North American Ph.D. TX. 118pp. Goodrich, John M. 1990. Ecology, conservation and management of two western Great Basin black bear populations. M.S. Thesis, University of Nevada, Reno, NV. 84pp. Gresswell, R. E., B. A. Barton, and J. L. Kershner. [1989]. Practical approaches to riparian resource management: an educational workshop, 8-11, 1989, Billings, Montana. Billings, MT : U.s. Bureau of Land Management. 193pp. May Guerouali, A. 1990. Digestion and metabolism during pregnancy and lactation in prolific sheep: dman ewes. Ph. D. Dissertation, Colorado state University, Fort Collins, co. 232pp. Harris, H. T. 1991. Habitat use by dispersing western Montana. M.S. Thesis, University 40pp. Hasbrouck, J. J. 1991. Demographic responses exploitation rates. Ph.D. Dissertation, IA. 144pp. and transplanted beavers in of Montana, Missoula, MT. of raccoons to varying Iowa State University, Ames, Harmoning, A. K. 1976. White-tailed deer dispersion and habitat utilization in central North Dakota. M.A. Thesis, North Dakota state University, Fargo, ND. 56pp. Hygnstrom, S. E. 1988. Animal damage control in Wisconsin. University of Wisconsin -.Madison" WI. 84pp. Ph.D. Thesis, Kelly, B. T. 1991. Carnivore scat analysis: ari evatuation of existing techniques and the development of predictive models of prey consumed. M.S. Thesis, University of Idaho, MOSCOW, ID. 200pp. Laake, J. L. 1992. Catch-effort models and their application to elk in Colorado. Ph.D. Dissertation, Colorado state University, Fort Collins, co. 104pp. J. C. 1977. Distribution and habitat (Vulpes macrotis) in Utah. M.S. Thesis, UTe 92pp. McGrew, characteristics of the kit fox Utah state University, Logan, �Plumb, G. E., ed. 1992. National Park Service Research Center. 15th Arinual Report 1991. [Laramie, WY): University of Wyoming, National Park Service Research Center. 309pp. Sahanatien, V. A. M. 1990. Mediating human interactions with nature: of ecological concepts in the management to furbearing mammals. Thesis, York University, Downsview, Onto 129pp. the use M.S. Samuel, M. D. 1984. An evaluation of elk sightability in north central Idaho with application to aerial census and herd composition counts. Ph.D. Dissertation, University of Idaho, Moscow, ID. 118pp. Schooley, R. L. 1990. Habitat use, fall movements, and denning ecology of female black bears in Maine. M.S. Thesis, University of Maine, Orono, ME. 115pp. Taylor, B. L. 1991. Perspectives on dynamics of small popUlations: diverse ways of ,apprehending pattern and details from variable and incomplete data. Ph.D. Dissertation, University of California-San Diego. 151pp. Van Deelen, T. R. 1991. Dispersal patterns of juvenile beavers in western Montana. M.S. Thesis, University of Montana, Missoula, MT. 91pp. Vogel, W. O. 1983. The effects of housing developments and agriculture on the ecology of white-tailed deer and mule deer in the Gallatin Valley, Montana. M.S. Thesis, Montana State University, Bozeman, MT. 86pp. Wylie, S. R. 1981. Effects Hyla arenicolor Cope. Tempe, AZ. 123pp. Reference Document The Research Center individual articles segment. Location of basking on the biology of the canyon treefrog, Ph.D. Dissertation, Arizona State University, and Delivery Library staff also located and delivered approximately on request for Mammals Research personnel during this Manuscripts Published FY 1992-93 (Includes those published and those 'in press.) Job Progress All studies. Reports; Federal Aid. 931 Cassirer, E. F., D. J. Freddy, and E. D. Ables. 1992. Elk responses to disturbance by cross-country skiers in Yellowstone National Park. Wildl. Soc. Bull. 20:375-381 Cassirer, E. F., D. J. Freddy, and E. D. Ables. 1993. ranging elk. J. Wildl. Manage. (in press) Heart rates of free- �6 Freddy, D. J., D. L. Baker, R. M. Bartmann, and R. C. Kufeld. 1993. Deer and elk management analysis guide, 1992-94. Colo. Div. of Wildl. Division Report No. 17. 77pp. Gross, J. E., N. T. Hobbs, and B. A. Wunder. 1993. Independent variables for predicting intake rate of mammalian herbivores: biomass density, plant density, or bite size? Oikos. (in press) Hobbs, N. T., and D. E. Spalinger. 1992. Herbivore functional response: a mechanistic model of food intake rate in patchy environments. Bull. Ecol. Soc. Am. (Suppl.) 73:209. (Abstract) Miller, M. W., and E. T. Thorne. 1993. captive cervids as potential sources of disease for North America's wild cervid popUlations: avenues, implications, and preventive management. Trans. N. Am. Wildl. Nat. Res. Conf. 58. (in press) Neal, A. k., G. C. white, R. B. Gill, D. F. Reed, and J. H. Olterman. 1993. Evaluation of mark-resight popUlation estimates using simulations and field data~ J. Wildl. Manage. 57: (in press) Ringelman, J. K., M. W~ Miller, and W. F. Andelt. 1993. Effects of ingested tungsten-bismuth-tin shot on mallards. J. wi1dl. Manage. 57: (in press) Saltz, D., G. C. White, and R. m. Bartmann. 1992. Urinary cortisol, urea nitrogen excretion, and winter survival in mule deer fawns. J. Wildl. Manage. 56:640-644. Shipley, L. A., J. E. Gross, D. E. Spalinger, N. T. Hobbs, and B. A. Wunder. 1993. Mechanisms of foraging: the scaling of maximum intake rate in mammalian herbivores. Am. Nat. (in press) Shipley, L. A., J. E. Gross, D. E. Spalinger, N. T. Hobbs, and B. A. Wunder. 1992. The scaling of intake rate in vertebrate herbivores. Bull. Ecol. Soc. Am. (Suppl.) 73:343. (Abstract) Thorne, E. T., M. W. Miller, D. A. Jessup, and D. L. Hunter. 1992. Disease as a consideration,i~ translocating and reintroducing wild animals: western state wildlife management ageJ~cy.perspective. 'pp•.18:-25 in R. E. Junge, ed., Proc. of the Joint Me~ting of the Am."Assoc. Zoo Vet. and Am. Assoc. Wild1- Vet., 404pp. Manuscripts in Review FY 1991-92 (Includes those in review at publishing review. ) agency. Not 'peer' Kraabel, B. J., M. W. Miller, D. M. Getzy, and J. K. Ringe1man. 1993. Corrosion of and inflammatory responses to embedded tungsten-bismuth-tin shot and steel shot in mallards. J. Wildl. Dis. (in review) �Pojar, T. M., D. C. Bowden, and R. B. Gill. 1993. Comparison of methods estimate pronghorn numbers. J. Wildl. Manage. (in review) to Reed, D. F., J. Vayhinger, E. B. Brekke, and T. P. Huber. 1993. Mountain sheep habitat use in the Arkansas River Canyon, Colorado - west of Canyon City and north and east of Salida. Colo Div. Wildl. Rep. (in review) White, G. C., and R. M. Bartmann. 1993. Drop nets versus helicopter for capturing mule deer fawns. J. Wildl. Manage. (in review) net guns Wild, M. A., and M. W. Miller. 1993. Effects of modified Cary and Blair medium on recovery of nonhemolytic Pasteurella haemolytica from bighorn sheep (Ovis canadensis) pharyngeal swabs. J. Wildl. Dis. (in review) Wild, M. A., M. W. Miller, D. L. Baker, R. B. Gill, N. T. Hobbs, and B. J. Maynard. 1993. Comparison of growth rate and milk intake of bottleraised and dam-raised bighorn sheep, pronghorn antelope and elk neonates. J. Wildl. Manage. (in review) Prepared by b1~ q~§t» Librarian W£ ��9 JOB PROGRESS Colorado State of Project Work No. Job No. Covered: Author: Personnel: Mammals W-153-R-6 Plan No. Period REPORT July Research 1 Multispecies 9 Mammals 1, 1992-June 1 Research Administration 30, 1993 R. B. Gill R. B. Gill and D. K. Hall ABSTRACT Only 94% of the fiscal resources allocated to the Mammals 1 Research section for FY 92-93 were expended. The primary reason for the difference between resource allocation and expenditures was due to the time lag between manuscript preparation, acceptance for publication, and billing by professional journals. During FY 92-93, 1 new project was started, 3 ongoing projects were and 1 project proposal was prepared and submitted for decision. concluded, ��11 MAMMALS 1 RESEARCH ADMINISTRATION R. Bruce Gill P.N. OBJECTIVE Administer research studies within the Mammals productivity at the lowest cost. Agreement 1. Supervise and administer Research Section. 1 Research Unit for the highest Objectives research on deer, elk, and moose in the Mammals RESULTS ~ The total budget request for FY 1992-93 for Mammals 1 Research was $560,798. Of that amount, $525,401.66 was expended to accomplish research programmed for FY 1992-93. The majority of the unexpended balance (> 90%) resulted from savings in 3 projects; Work Plan 1 Job 7 (45%), Work Plan 1 Job 9 (21%), and Work Plan 4 Job 1 (25%). Work Plan 1 Job 7 was under expended because several expected manuscripts and publications were delayed in the journal review and publication process. Work Plan 1 Job 9 was underexpended because the majority of the administrative costs for both Mammals Research Projects were charged to the Mammals 2 Research administration project since both Research Projects temporarily are being managed by a single Wildlife Research Leader. Work Plan 4 Job 1 was underexpended because several of the project expenses were cost-shared with the Northeast Region. ~ One new research start was proposed and planned in FY 92-93 and funded for FY 1993-94: Work P~an 3 Job 9 - Es~ima~ing survival ra~es of elk and deve~oping ~echniques ~o es~ima~e popu~a~ion size. ~ Three research projects, Work Plan 3 Jobs 5, 6, and 7 were concluded in FY 92-93. One, Work Plan 3 Job 5 has been summarized in manuscript format and is in peer review process for the journal Eco~ogical App1.ica~ions. The other 2, Work Plan 3 Jobs 6 and 7 will be summarized in manuscript form during FY 93-94 and submitted to jqurnals yet to be selected·~ .. ~ The acting Mammals 1 Research Leader. actively participated in the Colorado Division of Wildlife's Long Range Plan update process, the Mount Evans management planning process, the Terrestrial Wildlife Reorganization planning process, and the Human Dimensions Advisory Committee. �12 ~ A research proposal was prepared outlining research protocol and anticipated costs for a study to evaluate potential resource competition between mule deer and wapiti populations. If the proposal receives a "green light", the more detailed Program Narrative planning process will be initiated with a tentative project start-up date planned for FY 9596. :7 /) Prepared by r> _\..l-(_,_IJ_Ait__;:·t.(:...;:f.'--...;::(():...;c1{=_ .•..•~"'97 R. Bruce Gill Wildlife Research Leader _ �_U Colorado Division Wildlife Research July, 1993 of Wildlife Report JOB PROGRESS State of Project Work Colorado No. ~W_-~1~5~3~-~R~-~5~ Plan No. Job No. Period REPORT ~2~ 7 Covered: Author: Personnel: July _ Mammals _ Deer Research Investigations Development of Census in Plains Riverbottom Methods for Deer habitats 1, 1992 - June 30, 1993 R. C. Kufeld D. Bowden, D. Younkin ABSTRACT A Colorado Division of Wildlife Technical Publication, entitled "Mule and White-tailed Deer in Plains Riverbottom Habitats of Eastern Colorado" is being prepared, which describes results of all aspects of this study. When completed it will constitute the final report. ��15 DEVELOPMEHT OF CENSUS METHODS FOR DEER IN PLAINS RIVERBOXTOM HABITATS Roland C. Kufeld P. N. OBJECTIVES 1. To determine seasonal movements mule deer in plains riverbottom and home range size of white-tailed habitats. 2. To develop and test methods for estimating plains riverbottom habitats. size of deer populations and in SEGMENT OBJECTIVE To determine seasonal movements and home range size of white-tailed deer in plains riverbottom habitats. and mule STUDY AREA The South Platte River study area was described Arkansas River Study area by Kufeld (1991). by Kufeld (1989), and the METHODS AND MATERIALS A Colorado Division of Wildlife Technical Publication, entitled ·Mule and White-tailed Deer in Plains Riverbottom Habitats of Eastern Colorado· is being prepared, which describes results of all aspects of this study. When completed it will constitute the final report. LITERATURE CITED Kufeld, R. C. 1989. Development of census methods for deer in plains riverbottom habitats. Colo. Div. Wildl., Wildl. Res. Rep. July (1):1117. Kufeld, R. C. 1991. Development of census methods for deer in plains riverbottom habitats. colo. Div. Wildl., Wildl. Res. Rep. July:9-18. Prepared bY~)C~ Roland C. Kufeld Wildlife Researcher C. ��II Colorado Division Wildlife Research July 1993 of Wildlife Report JOB PROGRESS state of Project Colorado No. Mammals W-153-R-6 Work Plan No. Job No. Period REPORT 2 Deer Author: Personnel: July Investigations Compensatory Effects Mule Deer Population 15 Covered: Research of Harvest in a 1, 1992 - June 30, 1993. R. M. Bartmann, G. C. White. G. D. Bear, L. H. Carpenter, W. Devergie, B. L. Dupire, S. E. Fairbairn, D. J. Freddy, J. Frothingham, R. B. Gill, V. K. Graham, B. D. Gray, J. P. Gray, R. Harthan, T. Lytle, J. A. LaFleur, J. D. Madison, M. M. Mellaci, J. E. Morris, B. K. Rush, D. G. Saltz, C. L. Vardaman, E. B. White, and numerous others from the Colorado Division of Wildlife and Colorado State University. Abstract Estimates of deer density from aerial line transects were 26.0 and 17.0 deer/krn2 on the control and treatment units, respectively. Density on the control was 55% lower (~ = 0.008) than in 1991, but this is not consistent with survival data. Helicopter net guns were used to supplement drop nets to capture deer for radio collaring and enabled achieving desired sample sizes for the first time in 3 years. Seventy-nine fawns and 9 does were captured with helicopter net guns and 86 fawns and 38 does were captured with drop nets. Estimated fawn survival rates on the control (0.112, SE 0.036) and treatment units (0.148, SE 0.043) were not'different (~ = 0.102). Estimated adult doe survival from 1 December 1992 to 1 July 1993 was 0.761 (SE 0.063) on the control unit and 0.807 (SE 0.058) on the treatment unit. Three does and 6 fawns moved off the Ridge study area. All the fawns and 2 does died and the other doe moved back through the study area during spring migration. The late hunting season in December was shortened to 9 days and only 100 licenses were issued. The estimated kill was 105 deer and consisted of 79% does and 21% fawns. No differences were found in body condition indices between the control and treatment units for fawns (~~ 0.824), yearling does (~ ~ 0.195), or adult does (~~ 0.329). ��19 COMPENSATORY EFFECTS OF HARVEST IN A MULE DEER POPULATION Richard M. Bartmann and Gary C. White P. N. OBJECTIVES 1. Increase the winter survival rate of mule deer fawns by lowering deer density to reduce competition for forage during winter. 2. Increase the harvest rate of deer through increased productivity of adult does and decreased natural mortality of fawns resulting from closer alignment of popUlation size with carrying capacity. SEGMENT OBJECTIVES 1. Maintain the winter population a density <40jkm2 for 5 years. of mule 2. Estimate winter survival rates of fawns on control 3. Estimate units. annual survival rates 4. Estimate units. harvest 5. Estimate condition 6. Estimate age structure of adult females on the treatment unit condition of adult females on control and treatment units. of adult rates of bucks, total does, of fawns on control deer on the Ridge females treatment and treatment on control and fawns on control and treatment unit at units. and treatment and treatment units. and METHODS Except for deer trapping, methods remained the same as previously reported by Bartmann (1990) and Bartmann and White (1991) with modifications by Bartmann and White (1992). In addition to drop nets, fawns and does were captured with helicopter net guns by Helicopter Wildlife Management, Inc. RESULTS Maintain AND DISCUSSION Population Aerial line transects were flown on the Ridge study area 9-10 January and again 9-11 February 1993. During the January flight, there was 100% snow cover due to recent heavy snowfalls. Deer were noticeably less responsive to helicopter disturbance than in previous years, a presumed consequence of the deep snow. Also, wind conditions prevented flying at desired altitudes, �20 particularly in steeper canyons, and forced canceling the final set of transects on each unit. Resultant density estimates were much lower than expected; 33.0 and 18.2 deerjkm2 on control and treatment units, respectively. Weather for the second flight in February was more favorable. However, deer were still not very responsive to the helicopter and density estimates changed little from the first survey; 26.0 and 17.0 deerjkm2 on control and treatment units, respectively (Fig. 1). Estimated density on the treatment unit was lower than on the control unit and on the treatment unit in 1991, but differences were non-significant (~~ 0.055). Estimated density on the control unit was significantly lower than in 1991 (~ = 0.008). --t- COIITROL ................... ,i 1 2 0 -....... ~ -.. ;, 100 ._-+--- __ ..•................................... T REA - __ -- - ~·--·-·-··---··i--···················--·-- r~~·~·.~ T K It It T ---_ --- ..----._-..-----.----------..--_._---- ... i·············!··········· l··············I·············j························ • 0 '" c .............. _.... : - , 0 •... • 0 '" z 2 0 . ~~~:J~~~~~~~~ 2~:~;·~~ ..~~~~~~~~<~L::i:~··;:···~:::::::~~~ . :::::L::::::::.:~::::::::::j::::::::::::::::::::::::: .. ..... _--. __ _---_.: ~::::::::~~t~~~~~t~~ _- -_._------_ _----_ _--_._--_ _---_! . c 1'85 1'86 1987 1988 19.9 1990 1991 1"2 YEAR Fig. 1. Deer density estimates (wj95% confidence intervals) from aerial line transects on control and treatment units during early to mid-winter. A slight decline in deer density could be expected on the treatment unit as a result of the late season in December. However, a 55% reduction on the control unit seems unreasonable, especially since the estimated 46% fawn survival for the 1991-92 winter was above the previous 9-year average of 35% and there was no indication of large-scale movement between the 2 units. Fawn Survival Dropnetting of deer occurred from 11 November to 4 December 1992. We radio collared 31 fawns on the control unit and 55 fawns on the treatment unit. Another 64 fawns, 51 and 28 on the control and· treatment units, respectively, were captured with helicopter net guns from 1-5 December 1992. Netgunning was done on an experimental basis to determine if it was a viable alternative to dropnetting that had become less effective in recent years. A manuscript comparing helicopter netgunning with dropnetting is described in the report for Work Plan lA, Job 5. The 1992-93 winter was considerably harsher than other recent depths on north slopes averaged close to 90 cm and snow cover persistent on south slopes. These conditions resulted in the survival since the 1983-84 winter on the control unit (0.112, since 1987-88 on the treatment unit (0.148, SE 0.043) (Table rates were not different from each other (~= 0.102). winters. Snow was more lowest fawn SE 0.036) and 1), but these Predation was the major mortality cause on the control unit (64%) and took a greater percentage of fawns than during any of the past 10 years (Table 2). Except for the 31% and 40% during the 1986-87 and 1990-91 winters, respectively, predation has been less than 20% and was 0% during 2 winters. Predation has generally been lower on the treatment unit than on the control, but even here it was higher (39%) than during any previous year. Predation percentages are minimal estimates because they represent.confirmed losses. When evidence is inconclusive, mortalities are placed in the "other" category. �21 Table 1. Kaplan-Meier estimates of survival rates (§.) for radio-collared mule deer fawns on control and treatment units of the Ridge study area in Piceance Basin, Colorado, from time of collaring in November and December until the following 15 June 1982-83 through 1992-93. Hunting mortalities are censored. Winter 1982-83 1983-84 1984-85 1985-86 1986-87 1987-88 1988-89 1989-90 1990-91 1991-92 1992-93 n 28 28 34 59 60 32 34 38 34 28 80 Control §. 0.321 0.071 0.196 0.537 0.431 0.241 0.270 0.779 0.320 0.456 0.112 unit SE(§.) 0.088 0.049 0.078 0.070 0.064 0.077 0.083 0.078 0.090 0.107 0.036 n 31 32 26 58 58 28 28 44 36 42 74 Treatment §. 0.387 0.033 0.431 0.439 0.471 0.107 0.445 0.745 0.339 0.548 0.148 unit SE(§.) £. of equal §.ct.) 0.087 0.033 0.105 0.070 0.067 0.058 0.096 0.070 0.106 0.098 0.043 0.578 0.774 0.075 0.157 0.565 0.006 0.509 0.659 0.909 0.481 0.102 Six fawns from the treatment unit and 1 from the control unit were shot during the late season in December. The fawn from the control unit was a net-gun capture. We do not have precise locations for these captures, but presume it was caught close to the boundary and may have been on the treatment unit when killed. For this study, hunting mortalities are treated as censored animals and survival estimates have been recalculated to reflect this status. Six fawns and 3 does left the Ridge study area. Three fawns and 2 does moved off the treatment unit by early January and wintered about 7 km farther down the White River near the mouth of Yellow Creek. All the fawns and 1 doe eventually died there. The other doe was again heard on the Ridge study area in April 1993, presumably as she began spring migration. Two fawns and 1 doe from the treatment unit were found dead north of the White River and 1 fawn from the control unit was found dead immediately south of the west end of the treatment unit. These deer were all excluded from survival analyses. Movements off the Ridge study area have been rare and have not previously occurred so early in the winter and pose a possible bias in survival analyses. Because these deer left the Ridge study area, they could no longer be considered part of the experiment. However, only with the 5 deer that moved farther down the White River was it known, before any deaths occurred, they had left the area. They were eventually located by aerial tracking. The other 4 deer were in locations where they could still be heard during monitoring on the study area, and it was not until they died that their true locations were determined. Therefore, it is possible that other deer that did not die also could have been off the study area and their radio signals still heard. These movements would not always be detected during location efforts during January, February, and March because specific locations were not identified. Rather, it was only determined if radio signals originated east or west of the boundary between the control and treatment units. �22 Table 2. Cause of mortality for radio-collared mule deer fawns on control and treatment units of the Ridge study area in Piceance Basin, Colorado, from time of collaring in November and December until the following 15 June 1982-83 through 1992-93. Percentages are of total uncensored· fawns. No. of fawns Winter 1982-83 1983-84 1984-85 1985-86 1986-87 1987-88 1988-89 1989-90 1990-91 1991-92 1992-93 29 28 34 59 60 32 34 38 34 28 80 1982-83 1983-84 1984-85 1985-86 1986-87 1987-88 1988-89 1989-90 1990-91 1991-92 1992-93 31 32 26 58 58 28 28 44 36 42 74 Censored Hunting other 1 7 11 6 2 5 14 9 7 1 2 1 1 1 4 9 5 5 9 9 6 3 9 7 9 1 starvation No. % Control unit 54 15 22 79 16 59 10 21 14 26 22 73 10 34 3 14 6 24 7 33 12 15 Treatment unit 15 50 27 87 8 36 17 35 16 30 19 68 36 9 6 20 8 40 7 29 23 34 Mortality cause Predation No. % 4 4 5 7 17 14 14 19 15 31 5 17 10 3 50 40 14 64 2 3 2 11 13 1 1 7 10 9 22 25 4 4 3 2 26 15 8 39 other No. % 1 7 3 2 7 3 3 3 7 4 15 6 7 24 14 12 14 9 2 7 3 1 1 5 5 4 4 4 7 14 2 2 18 20 13 20 17 10 • Uncensored fawns are those that were not killed by hunters, that had nonfailing radios, or that had collars that did not drop off prematurely. Adult Doe Survival Fifty-four does radio collared in previous years, including 9 female fawns radio collared in 1991 that survived to 15 June 1992, were alive as of 1 December 1992; 29 and 25 on the control and treatment units, respectively. During November and December 1992, 38 does were captured with drop nets and 9 with helicopter net guns bringing the total number radio collared on the control and treatment units to 46 and 55, respectively. Another 24 does, 15 yearling bucks, and 7 mature bucks were captured with drop nets and released without radio collars. The number of bucks caught is a substantial increase over the 8 yearling and 1 mature bucks captured in 1991; a probable result of the restrictive 3-day buck seasons set for 1992. Six does were killed on the treatment unit during the December late season. As in the past, these were treated as censored animals in survival estimates. �23 During the 7-month period ending unit and 9 died on the treatment (SE 0.063) and 0.807 (SE 0.058), 30 June 1993, 11 does died on the control unit for preliminary survival rates of 0.761 respectively (Table 3). Table 3. Kaplan-Meier estimates of annual (1 Dec-30 Nov) survival rates (~) for radio-collared adult female mule deer on control and treatment units of the Ridge study area in Piceance Basin, Colorado, 1982-83 through 1992-93. Hunting mortalities are censored. Winter 1982-83 1983-84 1984-85 1985-86 1986-87 1987-88 1988-89 1989-90 1990-91 1991-92 1992-93- !! 10 15 9 25 27 14 7 23 39 41 46 - Survival Control Censored 1 4 2 1 unit ~ 0.800 0.779 1.000 0.917 0.756 0.818 0.857 1.000 0.969 0.758 0.761 rate estimates .!! 0.126 0.113 0.056 0.087 0.116 0.132 0.031 0.071 0.063 for 1992-93 11 15 10 21 18 10 Treatment Censored ~ 2 1 5 28 41 43 52 9 12 6 6 0.909 0.929 1.000 0.900 0.878 1.000 0.800 1.000 0.906 0.809 0.807 are only through R of equal ~ unit SE(~) 0.448 0.271 1.000 0.821 0.432 0.329 0.854 1.000 0.474 0.608 0.580 0.087 0.069 0.067 0.081 0.179 0.065 0.071 0.058 30 June 1993. All radio-collared deer were located during mid-month in January, February, and March to check for movements between the control and treatment units. A doe and fawn that moved from the treatment to the control unit in 1991 remained on the control unit during most of the 1992-93 winter but had returned to the treatment unit by mid-March. The only other movement between units was by 2 fawns that moved from the control to the treatment unit. Both were captured with net guns, presumably close to the boundary between units. Harvest The 1992 late season on the treatment unit was shortened because the targeted population reduction was believed accomplished. A 9-day season occurred 12-20 December with only 100 licenses, each valid for 2 antlerless deer (Table 4). The estimated harvest based on a random survey was 105 deer and consisted of 83 does and 22 fawns. Percentages of success for people taking 1 and 2 deer were within the range of values for past seasons. The estimated 21% fawns in the harvest was the lowest of the 4 seasons, although only marginally lower than the 25% estimated the first year. �24 Table 4. Hunter participation and success and deer harvest estimates for the December late season on the treatment unit of the Ridge study area, 1989-92. 1989 No. Licenses· Survey respondents Did not hunt Unsuccessful Harvested 1 deer Harvested 2 deer Harvest-Does Fawns Total • Hunters Condition 1990 % 375 60 (SE) 1992 1991 No. No. % 400 92 % % 100 37 350 69 81 38 56 200 21.7 10.0 15.0 53.3 48 109 104 139 12.0 27.2 26.0 34.8 41 76 122 112 11.6 21.7 34.8 31.9 342 114 456 75.0 25.0 (47.5) 263 120 383 68.6 31.4 (40.0) 214 131 345 62.1 37.9 (39.4) were allowed to take 2 antlerless No. 32 30 38 83 22 105 32.4 29.7 37.8 78.9 21.1 (15.8) deer on a license. of Fawns Body condition of fawns was found not to differ (~ ~ 0.824) between the control and treatment units with regard to weight, total body length, left hind foot length, and a weight/length ratio (Table 5). Age Structure and Condition of Does Despite efforts to eartag all fawn and yearling deer beginning in 1988, only 2 known-age jaws >2-years old have been collected. One was from a 3-2/3-yearold doe and the other from a 9-2/3-year-old doe captured as a fawn in 1982. This lack of known-age jaws for reference prompted combining the 4-7-year-old age classes although we still retained the >7-year-old category. Except possibly for 1989, adult female age structures for regular and late seasons in Fig. 2 are not truly comparable. Most does killed during regular seasons were on the. control unit. This area seemed more popular with hunters and the deer population wa:s not subjected to the extreme reductions imposed on the treatment unit. In addition, relatively small sample sizes probably contributed to some of the inconsistencies in age structures across years. During late seasons, proportions of deer in the 4-7 and >7-year-old age classes declined over the years. The other prominent aspect of these data is the carry-over to the 3-year-old age class of the high fawn survival during the 1989-90 winter. As in past years, no significant differences were found in body condition indices for yearling (~~ 0.195) or mature does (~~ 0.329) between control and treatment units (Table 6). �Table 5. Weights (kg) and body measurements (cm) of mule deer fawns trapped on control and treatment units of the Ridge study area in Piceance Basin, Colorado, 1982-92. n Weight so .!. Total body length so R Left hind foot lengl;h so R Unit Year Control 1982 1983 1984 1985 1986. 1987 1988 1989 1990 1991 1992 28 28 34 60 58 33 34 40 35 28 82 34.6 31.7 32.3 32.6 31.9 29.9 29.5 32.7 30.8 30.7 30.4 3.10 4.40 4.65 4.02 3.89 3.60 3.10 3.31 4.29 3.58 3.80 124.0 124.2 123.9 124.4 128.1· 127.3 123.8 131.0 126.5 128.2 126.8 4.64 5.65 7.25 6.26 6.53 6.12 7.83 5.81 9.02 5.47 6.67 41.1 40.6 40.8 41.1 41.0· 40.8 41.0 41.8 40.8 40.6b 40.6 1.08 1. 73 1.53 1.48 1.95 1.72 1.37 2.16 1.75 1.32 1.58 Treatment 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 30 32 26 60 61 28 30 47 36 43 83 32.8 32.3 32.3 32.3 31.7 30.2 28.8 30.6 30.7 32.6 30.3 4.18 3.12 5.07 4.62 4.13 5.34 4.13 3.33 4.41 3.54 4.68 121.7c 123.6 124.7 124.4 126.0· 127.5 124.9 126.6 127.6d 129.8 126.7 5.45 5.53 7.25 6.28 6.62 8.86 7.07 5.33 6.57 6.11 7.94 41.1c 40.6 40.8 40.8 41.0· 41.2 40.6 40.7 41.1d 40.8 40.5 1.65 1.34 1.89 1. 77 2.11 1.66 1.88 1.41 1.69 1.29 1.77 • !! = 60 b !! = 27 c !! = 31 d n = 32 �26 II REGULAR • SEASONS LATE SEASON 50 1 989 • 4 0 !!. • 33 • !!. • 223 .............................................................................................................................. 3 0 p. 0.003 . ,. 20 10 o 5 0 1990 4 0 p • 0.'07 1····················································· , • !!. • 201 3 0 2 0 10 Eo-< Z JO.:I t...> p::; JO.:I p., 0 5 0 1991 4 0 1.. · .. · 3 0 1 · · · · ··· · · · .. · .. ·· .. · .. · · .. · P · ···· .. ·· • ·· .. · .. ····· 0.543 ····· ·· .. · ···· .. ·· .. ···· .. ··· .. · .. ······· .. · ··· · .. · .. ·· . 2 0 10 0 5 0 1 992 • ~ • P 15 • 0.117 4 0 3 0 2 0 10 0 1 4 - 7 2 AGE > 7 (YEARS) Fig. 2. Estimated age structure of adult female regular and late hunting seasons, 1989-1992. mule deer killed during · .. 1 �1..1 Table 6. Weights (kg) and body measurements (cm) of yearling mule deer trapped on control and treatment units of the Ridge Piceance Basin, Colorado, 1988-92. Left hind foot length SD R Unit Year n Control 1988 1989 1990 1991 1992 2 2 10 8 4 46.2 46.0 49.450.9 50.6 Yearlings 1.91 6.43 2.38 2.50 4.85 146.1 148.4 147.9 149.2 145.6 10.75 2.26 2.49 3.38 4.19 45.8 47.8 45.8 45.7 45.5 0.99 3.04 1.08 1.02 1.59 Treatment 1988 1989 1990 1991 1992 2 7 8 10 10 50.2 49.3 50.8 49.4 50.3 0.35 3.58 3.38 3.45 5.38 151.2 154.3b 150.1 149.8 149.8 3.18 6.20 4.94 5.31 5.34 46.0 46.6 45.6 45.3 47.2 0.00 0.69 1.26 1.51 6.37 Control 1989 1990 1991 1992 19 21 30 31 67.7 66.9 62.9 63.0 Adults 5.22 5.51 5.18 5.67 168.6 166.7 162.3 164.3 5.20 6.98 6.40 7.89 48.1 47.3 47.3 47.5 1.10 1.09 2.19 1.29 Treatment 1989 1990 1991 1992 39 25 32 36 65.8 67.6 62.8 62.4 5.04 5.14 6.24 6.91 166.6 166.8 162.2 162.4 6.57 5.81 7.07 7.32 47.8 48.1 47.0 47.2 1.32 1.44 1.06 1.47 b n n ~ 9. 6. LITERATURE CITED Bartmann, R. M. 1990. Compensatory effects population. Colo. Div. Wildl., Wildl. of harvest Res. Rep. _____ , and G. C. White. population. Colo. 1991. Compensatory Div. Wildl., Wildl. effects of harvest in a mule Res. Rep. July:27-40. deer _____ , and G. C. White. 1992. Compensatory population. Colo .•Div. Wildl., Wildl. effects of harvest in a mule Res. Rep. July: (in press). deer Prepared by i ?~""S---Richard '-,--- Total body length SD R Weight SD and adult female study area in M. Bartmann (;l(t~: Dr. Gary C. White Professor - in a mule deer July:187-196. ��29 Colorado Division Wildlife Research July 1993 of Wildlife Report JOB FINAL REPORT State of Project C=o.l..:::o""r""'a .•• •.• d::,:o::...,_ _ No. W-153-R-6 Mammals Research Work Plan No. 3 Elk Investigations Job No. 6 Effect of Elk Harvest Breeding Period Author: Covered: Systems on Elk Biology July 1, 1992 - June 30, 1993 D. J. Freddy Personnel: M. Miller, CDOW, D. Bowden and G. White, Forbes Trinchera Ranch. Abstract CSU, and E. Ryland of We monitored harvests of male and female deer and elk and reproduction in elk on Forbes Trinchera Ranch from 1986 through 1992 by estimating ages of all animals harvested, determining Boone and Crockett antler scores for bucks and bulls, and collecting reproductive tracts from female elk. During 1986-1992, harvests increased for buck deer from 84 to 118, for bull elk from 68 to 117, for female deer from 69 to 186, and for female elk from 33 to 109 animals. There were no definitive trends in ages or antler scores of bucks harvested across years although there were yearly differences in ages (P -0.001) and antler scores (P ~ 0.06). Average age declined across years when age was based on replacement-wear criteria (P -0.024) but increased when based on dental cementum (P ~0.06). Antler scores of bucks remained stable across years (P -0.11) but trends in scores were dependent on which technique was used to determine age of deer. Ages and antler scores of bulls harvested remained stable or increased slightly across years (P ~ 0.002). Average age of female deer harvested increased across years (P -0.001) and for female elk, average age remained stable (P ~ 0.11). Increasing age of female deer may reflect poor recruitment of young adults. Pregnancy rates for elk averaged 80%, ranged from 62-92% among years and were 0% for calves, 19% for yearlings, 84-96% for prime-aged cows, and 77% for cows ~ II-years old. Fetal sex ratios deviated from 50 M:50 F in 4 of 7 years and inconsistently favored males or females (P ~0.05) but the sex ratio of 54 M:46 F pooled across years was not different from 50 M:50 F (P >0.05). Mean and median conception dates ranged from 23 September to 4 October with earliest and latest conceptions occurring 6 September and 17 November, respectively. Earlier conception dates were associated with lower snow depths the previous March (P -0.03). Since 1987, 100% of 248 serum samples collected from elk tested negative for brucellosis. We recommend continued monitoring of ages of females harvested and ages and antler scores of bucks and bulls harvested to interpret future trends in harvests. ��31 JOB FINAL REPORT EFFECT OF ELK HARVEST SYSTEMS ON ELK BREEDING David J. Freddy BIOLOGY P. N. OBJECTIVE Evaluate effects of harvest systems on breeding biology of elk. SEGMENT OBJECTIVES 1. Determine reproductive status of elk on the Forbes Trinchera fetal collections from hunter-killed elk. Ranch using 2. Determine physical condition of elk and deer on the Forbes Trinchera Ranch by collecting body weights, antler weights and measurements, and ages of animals harvested. 3. Complete a publication entitled "Pregnancy testing elk using progesterone assays" and present a summary of reproductive data at an appropriate workshop, symposium, or professional meeting if locally available. INTRODUCTION During the 1980s there was concern throughout Colorado that low numbers of bull elk (Cervus elaphus nelsoni) during the breeding season were reducing conception rates in adult cows and survival of calves. This concern was based primarily on correlations between declining postseason bull:cow ratios and calf:cow ratios in some herds in Colorado and other states. An auxiliary concern was that rising hunting pressure on bull elk during the rut in September could interfere with breeding and subsequently reduce conception rates or delay breeding of cows until their second or third estrus cycle later in the Fall. Part of this concern stemmed from insufficient knowledge about changes over time, if any, in the in vivo reproductive status of elk. Our knowledge of pregnancy rates and conceptions dates for elk in Colorado is limited primarily to one survey conducted in 1965-1966 in southwestern Colorado (Boyd and Ryland 1971) when total hunting pressure and hunting of elk during the ruc"were comparatively low. . As a participant in Colorado's Wildlife Ranching program, the Forbes Trinchera Ranch in south-central Colorado instituted controlled fee-hunting for bull elk and buck deer (Odocoileus hemionus) from September through early December and initiated controlled public hunting for antlerless elk and deer in December beginning in 1986 (Freddy et al. 1991). This hunting system afforded the opportunity to: 1) intensely monitor the harvesting of male and female elk and deer in a "trophy" hunting system, and 2) monitor reproduction in an elk population subjected to chronic hunting of mature bulls during the rut in September by obtaining reproductive organs from female elk harvested by hunters in December when in vivo pregnancy status of elk could be easily determined. -:-_-_- __ . �32 Our objectives were to monitor the number and age composition of male and female harvests of deer and elk and measure in vivo pregnancy rates, fetal rates, and fetal sex and weights of elk and to seek correlative relationships between these reproductive variables and numbers and ages of bulls harvested, age and body condition of cows harvested, and trends in precipitation by month and year. Ancillary objectives were to use blood samples obtained by hunters from harvested female elk to monitor disease status of elk and to evaluate accuracy of serum progesterone hormonal assays to predict pregnancy in elk. METHODS Ages of Animals Harvested We estimated age of deer and elk harvested from 1986 to 1992 using replacement and wear (RW) (Robinette et al. 1957, Quimby and Gaab 1957) and dental cementum (DC) (Stevens 1987, Keiss 1969). We used both techniques because each estimates true age with an unknown degree of bias. Two observers (E. Ryland, primary; D. Freddy, secondary) independently aged lower jaws removed from harvested buck deer and bull elk; heads and antlers of these animals were not present when jaws were examined and observers examined jaws at different times. These same observers examined intact jaws from antlerless deer and elk harvested and pooled their estimates of age. A median incisor was removed and age determined via dental cementum by the same technician each year at the CDOW research laboratory in Ft. Collins. For most analyses, ages were pooled into classes often representing young, mature, and old individuals. Antler Measurements We measured antlers from harvested bucks and bulls according to Boone and Crockett criteria (Nesbitt and Reneau 1986). Three individuals measured antlers from 1987-1992 with one person common to all years. Antler weights, including the frontal bone, were measured to the nearest 0.1 kg. Body Measurements and Fetal Collections Eviscerated body weights (nearest 0.5 kg) were obtained on deer and elk harvested from 1987 to 1992 using scales located at mandatory check stations. Total body length and hind foot length were also obtained for antlerless animals. Reproductive tracts from female elk, including calves, harvested primarily in December were collected by hunters who were provided step-by-step illustrated instructions prior to their hunt. Tracts were deposited at check stations and kept cool until processing. Pregnancy status was determined from the presence of fetuses, embryos, and developed uterine tissue. Questionable uteri were preserved for later examination. Fetal.measurements were made on fresh specimens (subsequently preserved) and followed definitions of Armstrong (1950). Fetal age was estimated from growth curves of Morrison et al. (1959). Blood Assays �33 We monitored progesterone levels of antlerless elk harvested by hunters in December 1987-1991. We could associate serum progesterone levels with known pregnancy status because hunters also collected reproductive organs from their elk allowing us to develop predictive relationships for determining pregnancy from blood assays. Serum progesterone levels were determined by radioimmunoassay (RIA) at the Physiology Laboratory, Go lorado State University, Ft. Gollins. Serum from elk harvested by hunters was tested for the presence of brucellosis by the USDA Laboratory, Denver, GO. Serum was stored at -18G from centrifuging until tested for brucellosis. Hunting Seasons Separate hunting seasons were established yearly for private fee-only hunters and public hunters. Private seasons for bucks and bulls were: 13 Sep-9 Nov 1986, 5 Sep-9 Oct and 14 Nov-II Dec 1987, 10 Sep-9 Dec 1988, 9 Sep-8 Dec 1989, 9 Sep-5 Dec 1990, 7 Sep-6 Dec 1991, and 12 Sep-ll Dec 1992. Public seasons primarily for female deer and elk were: 29 Nov-5 Dec 1986, 12-14 and 19-21 Dec 1987, 10-12 and 17-19 Dec 1988, 9-11 and 16-18 Dec 1989, 8-17 Dec 1990, 7-15 Dec 1991, and 12 Dec-21 Dec 1992. Statistical Analysis We used SAS (SAS Institute, Inc. 1988) to evaluate trends in age composition of harvests (PROG FREQ - chi-square, PROG GLM - anova and linear regression), antler measurements (PROG GLM), body measurements (PROG GLM), elk conception dates (PROG UNIVARIATE), elk fetal sex ratios (PROG FREQ) , and relationships between fetal sex ratios and ages and body sizes of pregnant cows (PROG LOGISTIG). We generally used P ~0.05 as the critical level for statistical significance. RESULTS AND DISCUSSION Harvests Yearly harvests of male dee~ and elk ranged from 84 to 118 and 68 to 117 animals, respectively during 1986-1992 (Table 1). Harvests were controlled yearly by adjusting numbers of permits available for fee-hunting. Harvests male deer and elk were intentionally reduced about 10% in 1992 by reducing available permits. For all years, about 54% of the bucks were harvested in November-December and 85% of the bulls were harvested in September-October. Public harvest of antlerless deer in 1992 was reduced about 45% compared to average harvest from 1987":1991 because of reduced numbers of permits while public harvest of antlerless elk in 1992 was the lowest since 1986 also because of reduced numbers of permits (Table 1). For all years, proportions of the buck, bull, and female deer and elk harvests occurring within the "Blanca" portion of the Ranch were: 68%, 56%, and 78-95%, respectively. Bucks-younger Age (RW) frequencies of bucks differed among years primarily due to a age structure in 1991 (P =0.001). Mean age differed among years, 0 �34 especially between 1987 and 1991 (ANOVA P -0.0001) and exhibited a weak decline across years (R2 -0.007, P ~0.024; Table 2). This weak decline in yearly RW age was corroborated by the second field observer (R2 -0.041, P -0.0001). Changes in DC ages among years were somewhat contrary to trends in RWages. Age frequencies (DC) of bucks differed among years primarily due to erratic proportions of 1-3 year old deer harvested in 1989 and 1990 and reduced harvest of ~ 7 year-old deer in 1986 (P -0.026). Yearly mean DC ages were always younger than RW means, differed among years because of younger bucks harvested in 1986 and 1989 (ANOVA P - 0.005), and exhibited a weak increase across years (R2 -0.005 P -0.06, Table 3). Neither aging method indicated definitive changes in age sturcture, and thus we conclude that age structure of buck harvest was relatively stable from 1986 to 1992. Age frequencies of bucks differed between RW and DC ages each year with DC indicating fewer bucks ~ 7 years old were harvested (P <0.04, Table 4). Differences in techniques cannot be solely attributed to the field observer. The 2 independent estimates of RW age frequencies differed only in 1986 (Table 6). However, a paired t-test indicated that observers differed in RW ages in 4 of 7 years (P <0.007, Table 6) but average differences (+0.34 - -0.41 years) were smaller than differences in average RW and DC ages (0.55 - 1.59 years, Tables 2, 3). Bulls-- Age (RW) frequencies of bulls differed among years due to an inconsistent harvest of bulls ~ 7 years old in 1988, 1990, and 1992 (P -0.001). Mean age was lower in 1986-88 (ANOVA P -0.001) and weakly increased across years (R2 =0.025 P =0.001; Table 8). This increase in RW age across years was not corroborated by the second field observer (R2 -0.0004 P -0.630) whose data indicated a stable age structure. Changes in DC ages among years also indicated an increasingly older aged harvest. Age frequencies (DC) of bulls differed among years due to increasing proportion ~ 7 years old especially in 1990 and 1992 (P ~0.001, Table 9). Yearly mean DC ages were always younger than RW means but increased among years, especially since 1990, and across years (ANOVA P - 0.001; R2 -0.114 P -0.001). We suggest that ages of bulls harvested increased, or at least remained stable, from 1986 to 1992. Age frequencies of bulls differed between RW and DC ages each year except in 1991 and 1992 with DC indicating fewer bulls ~ 7 years old were harvested (P <0.056, Table 5). Again, differerices in techniques cannot be solely attributed to the field observer. The 2 independent estimates of RW age frequencies differed only in 1990 (Table 7). However, a paired t-test again indicated that observers differed each year in RW ages except 1986 and 1989 (P <0.010, Table 7) but average differences (+0.33 - -0.72 years) were generally smaller than differences in average RW and DC ages (0.29 - 1.21 years, Tables 8, 9). Females-Deer-- Age frequencies of female deer differed among years due to a high harvest of 2-3 year-olds in 1988 and ~ 7 year-olds in 1990 based either on adult age classes (P -0.001) or adults plus fawns (P -0.001). Mean age of adults increased across years (ANOVA P =0.001; R2 -0.030 P -0.001; Table 10). We conclude that ages of female deer harvested increased from 1986 to 1992 based on changes in ages of adult animals. We believe basing our conclusions on changes in adult ages is more appropriate than analyses including fawns �35 because hunters likely select against fawns. Age frequencies of female deer were similar between RW and DC ages in most years (P >0.172) but differed when all years were pooled (P -0.009, Table 11). This difference was primarily caused by a lower proportion of yearlings and ~ 7 year-olds in the DC sample. We assumed that our RW criteria for yearlings were correct and that yearlings could be considered "known" age. The "error" with DC involving yearlings was caused by the presence of a cementum line resulting in some yearlings being classed as 2-years old. We therefore recommend that female deer be aged by RW for fawn and yearling classes and by DC or RW for older ages. Females-Elk-- Age frequencies of female elk were stable among years based either on adult age classes (P =0.115) or adults plus calves (P -0.130). Mean age of adults was stable across years (ANOVA P =0.648; R2 =0.008 P ~0.08l; Table 12). We conclude that ages of female elk harvested remained stable from 1986 to 1992 based on ages adult animals. Age frequencies of female elk were similar between RW and DC ages each year (P >0.091) and in all years pooled (P <0.169, Table 13). When all years were pooled there were no differences in frequencies for age classes but discrepancies were more evident for classes ~ 8 years (Table 13). We conclude that DC or RW provide acceptable estimates of age for female elk of all ages. We therefore found that the relative performance of RW and DC aging techniques was dependent on sex of animal. For adult bucks and bulls, DC consistently provided younger ages than RW but the techniques provided similar ages for adult female deer and elk. We believe our subjective criteria for wear on teeth was the same for both sexes. If so, younger ages provided by DC would suggest that wear patterns on male teeth were accelerated compared to females. Another explanation is that male teeth often had crowded cementum lines which were harder to read than the more orderly cementum lines found in females thus raising the possibility of missing lines in teeth of males. Antler Measurements Deer-- Antler scores and weights increased until plateauing at age 6. Trends were similar based on either RW or DC ages except that scores, and weights were slightly higher at younger ages for DC. Gross antler scores near 200 occurred primarily at ages ~ 5 but net scores near 200 occurred at age 6. Trends in scores and weights suggested ages 5, 6, and 7 were the most efficient for producing high scoring antlers. Trends in average antler scores and weights were inconclusive. Gross and net scores were different in 1988 and 1990 (ANOVA P ~0.062) but did not decline across years (R2 =0.004 P ~O.ll, Table 14). Antler weights declined (ANOVA P =0.001) but this was possibly due to a different scale used in 1987 which was the year contributing to the decline (Table 14). Within age classes ~ 7 and 4-6 years old, yearly differences in gross and net scores along with regressions to detect trends across years were erratically significant (ANOVA P ~0.06) depending upon whether RW or DC ages were used. The principle consistency was the negative slopes of marginally significant regressions -,-- �36 (Table 16). Ages of bucks and year of harvest were 2 potential variables affecting antler measurements and age, whether based on RW or DC, was the primary contributor to yearly differences in scores and weights (P -0.001, Table 18). Elk-- Antler scores and weights increased through age 8. Trends were similar based on either RW or DC ages except that scores and weights were again slightly higher at younger ages for DC. Gross antler scores near 375 occurred primarily at ages ~ 6 but highest net scores near 350 occurred at ages ~ 5. Trends in scores and weights suggested ages 6, 7, and 8 were equally likely to produce the highest scoring antlers. Yearly average antler scores were higher in 1992 than 1987 (ANOVA P 50.072) and showed a weak increase across years (R2 ~0.012 P ~0.01) (Table 15). Antler weights were stable among years (ANOVA P =0.297, R2 =0.002, P =0.29, Table 15). A similar pattern of stability in antler scores and weights occurred for age classes ~ 7 and 4-6 years old (Table 17). As with bucks, age of bulls harvested, whether based on RW or DC, was the primary contributor to yearly fluctuations in antler measurements (ANOVA P ~O.OOI for age and >0.38 for year, Table 18). Body Measurements Bucks-- Eviscerated weights of bucks differed in 1988, 1990, and 1992 and tended to decrease from 1987-1992 (ANOVA P ~0.001; R2 -0.019 P ~0.01; Table 19). Yearly fluctuations resulted primarily from effects of age (RW or DC), date of harvest, and the interaction of date and harvest (P <0.03) on body weight. We would expect bucks to lose body weight between September and the rut in December because of declines in body fat. Bulls-- Eviscerated weights of bulls differed in 1987 and 1991 and decreased from 1987-1992 (ANOVA P =0.033; R2 =0.019 P =0.01; Table 19). As with bucks, yearly fluctuations resulted primarily from effects of age (RW or DC), date of harvest, and the interaction of date and harvest (P <0.001) on body weight. We would expect bulls to lose body weight from before the rut in early September to after the rut in mid-October because of rapid declines in body fat. Females-Deer-- Eviscerated weight, total body length, and hind foot length were stable among years for male (ANOVA P >0.14) and female (ANOVA P >0.38) fawns. Male fawns were heavier than females (P -0.004) but hind foot lengths were similar between sexes (P >0.22, t-tests, Table 20). There was no evidence of increasing or decreasing fawn weights across years (R2 -0.004, P =0.47). Yearling females were also stable in body weight and lengths among years (ANOVA P >0.07, R2 =0.029, P =0.09). Adult weights stabilized after age 4 but hind-foot lengths stabilized at age 2-3 (Table 20). Females-Elk-- Eviscerated weight, total body length, and hind foot length were stable among years for male (ANOVA P >0.10) and female (ANOVA P >0.11) calves. Body weight and lengths of calves for sexes pooled was also stable across years (R2 -0.04, P ~0.08) Male and female calves were similar in weight and hind-foot lengths (P >0.43, t-tests, Table 20). Yearling females were stable �37 in body weight and size among years (ANOVA P >0.31). to increase through age 10 (Table 20). Adult weights continued Elk Reproduction Pregnancy Rates-- Pregnancy rates for adult cows (~ 1 yr old) averaged 80% and ranged from 62-92% during 1986 to 1992. Frequency of pregnancy differed among years for all adults ~ 1 year-old (P -0.004) and for adults ~ 2-years old (P ~0.020) with differences attributable primarily to high rates in 1986 and 1992 and low rates in 1990. The declining trend in pregnancy rates from 1986 through 1990 may reflect the drought conditions of those years. Pregnancies were not detected in calves (6 mos old) and rates for yearlings were variable and averaged 19%. For mature adults, rates were lowest (77%) in cows ~ 11years old (Table 21). For all years, litter size was 1 except for 4 sets of twins (1.4% of 282 litters). Twins consisted of 3 sets of 2 females (1986, 1989, 1990) and 1 set of 1 male and 1 female (1986). Cows with twins were 4, 6, 9, and 12 years old. Infected uteri not capable of supporting pregnancy occurred in 5 (1.4%) of 353 adults examined; these occurred in 1 yearling and 4 adults ~ 13 years old with 3 in 1986 and 1 each in 1988 and 1989. Fetal Sex Ratios and Body Size-- Fetal sex ratios deviated from 50 M:50 F in 1986, 1987, 1989, and 1991 inconsistently favored males or females (P ~0.05). However, the pooled ratio among years of 54 M:46 F was not different from 50:50 (P >0.05). Fetal sex did not deviate from unity within age classes of cows (P -0.05) although there was a tendency for cows ~ 11 years-old to produce male calves (Table 22). Changes in fetal sex among years could not be explained by the variables of mother body weight or mother age (P >0.23, logistic regression, n =199). We tentatively conclude that yearly variation in fetal sex ratios reflects sampling error. From 1987 to 1992, fetal weights and crown-rump lengths differed by year but not sex although there was a significant fetal sex by year interaction (ANOVA P <0.06, Table 23). Males were heavier in 1987 and 1990 (ANOVA P -0.007) and their weights were more erratic than weights of females whose weights remained stable among years (ANOVA P -0.488). Male weights declined from 1987 to 1992 (Rz ~0.042, P -0.022) unlike females who showed a tendency to increase in weight (Rz =0.023, P =0.078) We are cautious about interpreting . differences in fetal weights until fetal body measurements are adjusted for slight yearly differences in fetal ages (Table 24). Conception Dates-Mean and median conception dates ranged from 23 September to 4 October with the earliest breeding dates occurring in 1990 and 1991 (Table 24, Fig. 1, top). For all years, earliest and latest breeding dates were 6 September and 17 November, respectively. Yearly variation in median conception dates, pregnancy rates, and'fetal sex ratios suggested that yearly changes in amounts of precipitation and subsequent forage production or, numbers of mature bulls harvested was affecting reproduction. In the arid climate typical of the Ranch, precipitation is highly variable and cyclic as evidenced by patterns of yearly snow depth (Fig. 1, bottom). We could not correlate changes in pregnancy rates and fetal sex ratios with precipitation �38 in July and August when most rain is received (P >0.56), with changes in March snow depths which reflect the severity of winter (P >0.16), nor with numbers of bulls harvested (P >0.15). Estimated median conception dates were, however, correlated with snow depths in March (P >0.03, R2 - 0.582) but not precipitation in July and August or bulls harvested (P >0.19). Earlier conception dates were associated with lower snow depths the immediate previous March and were earliest when snow depths were below average in 1990 and 1991. This correlation suggested that severity or length of winter affects timing of conception the following Fall. Disease Surveys-- All serum samples collected during December 1992 from elk on Forbes Trinchera Ranch tested negative for brucellosis. The 48 samples were from 3 male and 7 female calves and 38 adult females. Since 1987, 100% of 248 serum samples have tested negative for brucellosis. Proges~erone Assays-- Final assay validation trials for elk progesterone RIA were finally completed in April 1992 thus delaying final preparation of the manuscript on using progesterone assays to determine pregnancy status in elk. CONCLUSIONS Yearly harvests of 100-120 mature bucks and 90-110 mature bulls were sustained from 1987 through 1992 without causing declines in ages of animals harvested or antler quality. Harvest objectives within these ranges appear sustainable provided the hunting system does not change so as to increase the vulnerability of older bucks and bulls or narrow the focus of animals selected. Yearly harvests have included some young 2-4 year-old bucks and bulls that were not of high trophy value but were satisfactory to some hunters and sustaining current harvests is predicated on taking some younger males each year. Yearly harvests of 55-110 antlerless elk were sustained without causing declines in ages of animals harvested and apparently withou~ affecting availability of suitable bulls for harvest. Evidence supports continuing these levels of antlerless harvest. High harvests of 160-186 antlerless deer from 1989 through 1991 were associated with increases in age of deer harvested from 1990 through 1992. The increase in average age of female harvested did not decline in 1992 when harvest was reduced about 50% suggesting the possibility that hunters may be selecting for older females or recruitment into young adult age classes has declined. If recruitment has declined, then reduced availability of mature bucks for harvest may become evident in 1993 or 1994. Caution would argue for maintaining reduced female harvests until there is evidence of improved recruitment. High harvests of females were instituted in part because densities were perceived to be high and reducing density could have favorable impacts on recruitment of fawns because of potentially improved nutrition. To date, weights of harvested fawns and yearling females have remained unchanged since 1986 suggesting that reduced densities of adult females has not been associated with improved body condition of young deer. �39 Pregnancy rates in elk were acceptably high in all mature adult age classes but differences in rates among years suggested there could be considerable variation in numbers of calves recruited annually. Conception dates appear to be influenced by snow depths the previous winter, and at this time, not influenced by hunting bulls during the rut. We recommend that ages of bucks and bulls harvested be estimated by replacement-wear criteria and if possible, dental cementum, and that measurements of antler quality continue so as to provide an objective basis for interpreting trends in male harvests. We also recommend monitoring ages of females harvested: for elk, dental cementum would be adequate for all age classes while for deer, replacement-wear should be used to judge fawns and particularly yearlings, and dental cementum could be used for older age classes. Continuing to measure weights on harvested fawns, calves, and yearlings of both species may provide insight into major changes in habitat conditions. LITERATURE CITED Armstrong, R. A. 1950. Fetal development of northern white-tailed deer. Amer. Mid. Nat. 43:650-666. Boyd, R. J., and E. E. Ryland. 1971. Breeding dates of Colorado elk as estimated by fetal growth curves. Colo. Game, Fish and Parks Div. Game Info. Leaflet 88. 2pp. Freddy, D. J., E. E. Ryland, andR. M. Hopper. 1991. Colorado's wildlife ranching program: the Forbes Trinchera experience. In: Wildlife Production: conservation and sustainable development. eds. L. A. Renecker and R. J. Hudson, pp. 336-343. AFES misc. pub. 91-6. Univ. Alaska Fairbanks, Fairbanks. Keiss, R. E. 1969. Comparison of eruption-wear patterns and cementum annuli as age criteria in elk. J. Wi1d1. Manage. 33:175-180. Morrison, J. A., C. E. Trainer, and P. L. Wright. 1959. Breeding seasons in elk as determined from known-age embryos. J. Wildl. Manage. 23:27-34. Nesbitt, W. H., and J. Reneau (eds.). 1986. Boone and Crockett Club's 19th big game awards. Boone and Crockett Club, Dumfries, Vermont. Quimby, D. C., and J. E. Gaab. 1957. Mandibular dentition as an age indicator in Rocky Mountain elk. J. Wild1. Manage. 21:134-153. Robinette, W. L., D. L. Jones, G': Rogers, and J. S. Ga shwe Ll.e r . 1957. Notes on tooth development and wear for Rocky Mountain mule deer. J. Wild1. Manage. 21:134-153. SAS Institute, Inc. 1988. SAS user's guide. SAS Inst. Inc., Cary, N. C. 1028pp. Stevens, M. L. 1987. Apparent accuracy of cementum annuli for estimating ages of mule deer. M. S. Thesis, Colorado State University, Fort Collins. Prepared by David J.' eddy Wildlife/Researcher �I o "'" Table 1. Harvests of male and female deer and elk on Forbes Trinchera Ranch during private fee-only and public hunting seasons. 1986-1992. Seas./Spec. 1986 Fb M8 Private/Deer Private/Elk 84 68 5 3 Public/Deer Public/Elk M 1987 F M 1988 F M 1989 F M 1990 F M 1991 F 0 0 107 93 0 0 109 87 0 0 118 103 0 0 108 117 0 0 116 104 0 0 69 33 6 5 137 63 5 3 133 79 5 3 186 61 17 9 160 56 16 9 168 109 M 1992 F 108 92 0 0 24 9 86 55 8Inc1udes only males ~ 1 year old. blnc1udes adult females and male and female fawns or calves. Table 2. Ages of buck deer harvested on Forbes Trinchera Ranch during private fee-only hunting seasons, 1986-1992. Age based on replacement and wear as estimated by E. Ryland. Age Category 1-3 4-6 ~ 7 Avg. Age (SE) n 1986 1987 Year (Yearly Percent) 1988 1989 1990 11 (13.3) 40 (48.2) 32 (38.6) 5 ( 4.8) 42 (40.0) 58 (55.2) 3 ( 2.8) 49 (45.4) 5~ (51.8) 16 (13.6) 50 (42.4) 52 (44.1) 5.84bc (0.19) 83 6.62a (0.18) 105 6.56ab (0.17) 108 e.oz= (0.18) 118 1991 1992 17 (15.7) 43 (39.8) 48 (44.4) 18 (15.5) 64 (55.2) 34 (29.3) 7 (6.5) 49 (45.4) 52 (48.2) 5.98abc (0.21) 108 5.52c (0.18) 116 6.Babe (0.18) 108 Total 77 337 332 746 a,b,eDifferent letters in rows denote yearly means that were significantly different based on Tukey's HSD test, P -0.05. �Table 3. Ages of buck deer harvested on Forbes Trinchera Ranch during private fee-only hunting seasons, 1986-1992 . Age based on dental cementum, Age Year (Year1~ Percent} Category 1986 1987 1988 1989 1990 Total 1991 1992 ( 1-3 4-6 ~ 7 Avg. Age (SE) n 17 (23.6) 51 (70.8) 4 ( 5.6) 19 (18.1) 69 (65.7) 17 (16.2) 19 (18.6) 66 (64.7) 17 (16.7) 34 (30.4) 62 (55.4) 16 (14.3) 10 ( 9.8) 71 (69.6) 21 (20.6) 19 (17.9) 72 (67.9) 15 (14.2) 17 (16.0) 69 (65.1) 20 (18.9) 4.49b (0.17) 72 5.038b (0.16) 105 5.058b (0.16) 102 4.63b (0.17) 112 5.408 (0.17) 102 4.978b (0.17) 106 5.088b (0.15) 106 135 460 110 705 Different letters in rows denote yearly means that were significantly different based on Tukey's HSD test, P =0.05. a,b +:-f-' I �42 Table 4. Yearly comparisons between replacement and wear (RW) and dental cementum (DC) ages of buck deer harvested on Forbes Trinchera Ranch during private fee-only hunting seasons, 1986-1992. Data represent only those deer aged by both techniques. RW age by E. Ryland. Year 1986 1987 1988 1989 1990 1991 1992 Aging Technique RW DC RW DC RW DC RW DC R'W DC RW DC RW DC Age Categor;:!(;:!rsl 4-6 1-3 ~ 7 11 17 5 19 3 19 16 33 15 10 16 17 6 17 33 50 42 66 48 66 44 60 41 71 57 71 48 69 27 4 54 16 51 17 48 15 46 21 30 15 52 20 n 71 71 101 101 102 102 108 108 102 102 103 103 106 106 Chi-square Yearly (P) 0.001 0.001 0.001 0.001 0.001 0.038 0.001 Table 5. Yearly comparisons between replacement and wear (RW) and dental cementum (DC) ages of bull elk harvested on Forbes Trinchera Ranch during private fee-only hunting seasons, 1986-1992. Data represent only those elk aged by both techniques. RW age by E. Ryland. Year 1986 1987 1988 1989 1990 1991 1992 \-- Aging Technique RW DC RW DC R'W DC RW DC RW DC RW DC RW DC Age Categor;:!(;:!rsl 1-3 4-6 ~ 7 10 25 16 46 15 27 15 42 19 21 18 18 8 13 32 26 59 38 63 56 61 53 48 64 52 60 47 55 10 1 15 6 9 4 24 5 45 27 22 14 36 23 n 52 52 90 90 87 87 100 100 112 112 92 92 91 91 Chi-square Yearly (P) 0.001 0.001 0.056 0.001 0.032 0.309 0.096 �43 Table 6. observers Trinchera represent Year 1986 1987 1988 1989 1990 1991 1992 Table 7. observers Trinchera represent Year 1986 1987 1988 1989 1990 1991 1992 Ages of deer based on replacement and wear (RY) compared between E. Ryland (ER) and D. Freddy (DF) for bucks harvested on Forbes Ranch during private fee-only hunting seasons, 1986-1992. Data only those deer aged by both observers. Aging Technique RY-ER RY-DF RY-ER RY-DF RY-ER RY-DF RW-ER RW-DF RW-ER RW-DF RW-ER RW-DF RW-ER RW-DR Age Category (yrs) 4-6 1-3 ~ 7 11 4 5 4 3 4 16 14 17 17 18 23 7 8 38 53 41 32 49 42 49 61 43 52 64 65 49 63 29 21 58 68 56 62 52 42 48 39 34 28 52 37 n 78 78 104 104 108 108 117 117 108 108 116 116 108 108 Chi-square Yearly (P) Avg. Age Yrs. (SE) 0.030 5.78(0.20) 5.88(0.21) 6.63(0.18) 6.96(0.18) 6.56(0.17) 6.64(0.17) 6.02(0.18) 5.80(0.18) 5.98(0.21) 5.71(0.19) 5.52(0.18) 5.11(0.17) 6.13(0.18) 5.74(0.17) 0.365 0.611 0.286 0.410 0.549 0.114 Ages of elk based on replacement and wear (RW) compared between E. Ryland (ER) and D. Freddy (DF) for bulls harvested on Forbes Ranch during private fee-only hunting seasons, 1986-1992. Data only those elk aged by both observers. Aging Technique RW-ER RW-DF RW-ER RW-DF RW-ER RW-DF RW-ER RW-DF RW-ER RW-DF RW-ER RW-DF RW-ER RW-DF Age Category (yrs) 1-3 4-6 ~ 7 12 12 13 10 15 9 16 16 19 24 22 25 7 9 42 44 54 52 63 63 61 64 49 66 54 57 48 60 11 91 14 19 96 15 25 22 48 26 27 21 36 22 n 65 65 81 81 87 87 102 102 116 116 103 103 91 91 Chi-square Yearly (P) 0.884 0.552 0.223 0.877 0.008 0.600 0.084 Avg. Age Yrs. (SE) 4.88(0.21) 4.89(0.23) 4.99(0.18) 5.27(0.20) 4.80(0.16) 5.14(0.17) 5.18(0.18) 5.24(0.18) 5.75(0.19) 5.03(0.17) 5.16(0.20) 4.92(0.18) 5.84(0.19) 5.40(0.17) ~ \ �1:- -'" -'" Table 8. Ages of bull elk harvested on Forbes Trinchera Ranch during private fee-only hunting seasons, 1986 -1992. A.lte_b_ased on replacemenLand_ wear as estimated bv E. RvLand . Age (yrs) Year (Yearly Percent) Category 1986 1987 1988 1989 1990 1991 1992 Total 1-3 4-6 ~ 7 Avg. Age (SE) n 13 (19.4) 43 (64.2) 11 (16.4) 16 (17.4) 61 (66.3) 15 (16.3) 15 (17.2) 63 (72.4) 9 (10.35 16 (15.5) 61 (59.2) 26 (25.2) 19 (16.4) 49 (42.2) 48 (41.4) 22 (21.2) 55 (52.9) 27 (26.0) 8 (8.7) 48 (52.2) 36 (39.1) 4.85b (0.20) 67 4.92b (0.17) 92 4.80b (0.16) 87 5.19ab (0.18) 103 5.75a (0.19) 116 5.15ab (0.20) 104 5.80a (0.19) 92 109 380 172 661 a,D Different letters in rows denote yearly means that were significantly different based on Tukey's HSD test, P -0.05. Table 9. ,Ages of bull elk harvested on Forbes Trinchera Ranch during private fee-only hunting seasons, 1986-1992. Age based on dental cementum, Year (Yearly Percent} Age 1986 1987 1988 Category 1989 1990 1991 1992 Total 1-3 4-6 ~ 7 Avg. Age (SE) n 28 (48.3) 29 (50.0) 1 (1.7) 46 (50.6) 39 (42.9) 6 ( 6.6) 27 (31.0) 56 (64.4) 4 ( 4.6) 42 (42.0) 53 (53.0) 5 ( 5.0) 21 (18.6) 65 (57.5) 27 (23.9) 19 (19.4) 64 (65.3) 15 (15.3) 13 (14.3) 55 (60.4) 23 (25.3) 3.64c (0.17) 58 3.79c (0.15) 91 4.17bc (0.15) 87 4.07c (0.15) 100 5.05a (0.18) 113 4.86ab (0.18) 98 5.45a (0.18) 91 196 362 80 638 a,h,cDifferent letters in rows denote yearly means that were significantly different based on Tukey's HSD test, P -0.05. �Table 10. Ages of antlerless deer harvested on Forbes Trinchera Ranch during public hunting seasons, 19861992. Age based on replacement and wear for fawns and yearlings and dental cementum for ages> 2 years. Age (yrs) Category Year (Yearlv Percent) 1988 1989 1990 1986 1987 13 (20.6) 22 (16.5) 14 (11.8) 32 (17.9) Yearling 4 ( 6.4) 19 (14.3) 10 ( 8.4) 2-3 23 (36.5) 49 (36.8) 4-6 18 (28.6) 2:: 7 1991 1992 Total 22 (14.3) 32 (20.1) 8 143 24 (13.3) 16 (10.4) 15 ( 9.4) 11 (12.9) 99 58 (48.7) 67 (37.2) 29 (18.8) 42 (26.4) 26 (30.6) 294 34 (25.6) 28 (23.5) 40 (22.2) 53 (34.4) 42 (26.4) 25 (29.4) 240 5 9 9 ( 7.6) 34 (22.1) 28 (17.6) 117 ( 6.8) 17 ( 9.4) 15 (7.9) Avg. Age" a.ao= (SE) (0.31) 50 3.35c (0.19) 105 3.59bc (0.20) 148 4.82a (0.26) 132 4.42ab (0.23) 4.48ab (0.33) n 3.33c (0.19) 111 127 77 Total Aged 63 133 119 180 154 159 85 Fawnd (9.4) (17.7) 893 Qlncludes male and female fawns. BAverage age excluding fawns. a,b,c Different letters in rows denote yearly means that were significantly different based on Tukey's HSD test P -0.05. ..,.. lJ1 1 �46 Table 11. Yearly comparisons between replacement and wear (RW) and dental cementum (DC) ages of female deer harvested on Forbes Trinchera Ranch during public hunting seasons, 1987-1992. Data represent only those deer aged by both techniques. Year Aging Technique Age Categor~ (~rs) Fawn" 1 2-3 4-6 ~ 7 n Chi-square Yearly (P) 1987 RW DC 12 12 18 9 31 38 19 24 10 7 90 90 0.306 1988 RW DC 10 10 ~ 57 57 35 26 16 7 105 105 0.051 5 1989 RW DC 9 9 19 16 49 64 37 36 28 17 142 142 0.292 1990 RW DC 9 9 15 5 36 34 41 51 29 31 130 130 0.184 1991 RW DC 20 20 13 12 43 41 33 40 32 28 141 141 0.906 1992 RW DC 2 2 9 8 22 25 20 25 22 15 75 75 0.172 ALL RW DC 62 62 81 55 218 259 185 202 137 105 683 683 0.009 8Includes male and female fawns. �Table 12. Ages of antlerless elk harvested on Forbes Trinchera Ranch during public hunting seasons, 19861992. Age based on replacement and wear for calves, yearlings, and 2-yr olds and dental cementum for ages> 3 years. Year (Yearlv Percent) Age (yrs) Category 1986 1987 1988 1989 1990 1991 1992 Total CalfC Yearling 2 2 ( 5.7) 14 (22.2) 14 (18.9) 9 10 (18.2) 22 (21.0) 12 (22.2) 83 (15.3) 6 7 7 6 9 7 2 44 (17.1) (11.1) ( 9.5) (10.2) (16.4) ( 6.7) (3.7) 7 10 (15.9) 8 4 7 ( 6.8) (12.7) 11 (10.5) 5 (10.8) 6 8 12 (20.3) 22 (21.0) 8 (12.7) 18 (24.3) 8 (17.1) 5 15 (23.8) 16 (21. 6) 18 _ (30.5) 8 14 (13.3) 11 (20.4) 87 (14.3) 2 4 5 4 8 ( 6.4) ( 6.8) ( 6.9) (14.6) 14 (13.3) 12 (22.2) 49 ( 5.7) 7 5 6 6 5 48 ( 7.9) ( 8.1) (10.2) ( 9.1) 15 (14.3) 4 (20.0) 5.398 (0.79) 33 35 5.148 (0.59) 49 63 5.138 (0.49) 60 74 5.608 (0.57) 50 59 5.368 (0.65) 45 6.038 (0.47) 83 105 6.408 (0.56) 42 54 .(20.0) 3-4 5-7 8-10 ~11 Avg. Aged (SE) n Total Aged (14.6) (14.6) 55 52 (9.3) 82 (14.8) (7.4) 358 445 Clncludes male and female calves. dAverage age excluding calves. 8 Different letters in rows denote yearly means that were significantly different based on Tukey's HSD test, p -0.05. I .I>'-.J �48 Table 13. Yearly comparisons between replacement and wear (RW) and dental cementum (DC) ages of female elk harvested on Forbes Trinchera Ranch during public hunting seasons, 1986-1992. Data represent only those elk aged by both techniques. Year Aging Technique Age Categor:l (:lrs) 1 2 3-4 5-7 CalfA 8-10 ~11 n Chi-square Yearly (P) 1986 RW DC 2 2 4 2 1 5 5 3 5 4 2 1 1 3 20 20 0.509 1987 RW DC 12 12 7 6 6 9 14 10 8 15 13 4 1 5 61 61 0.091 1988 RW DC 11 6 5 8 8 12 17 16 16 9 5 7 6 69 69 0.899 12 RW DC 5 5 4 4 4 3 9 9 4 5 6 49 49 0.849 11 13 16 1990 RW DC 8 8 7 8 6 4 8 10 8 7 8 7 4 5 49 49 0.988 1991 RW DC 7 7 3 3 9 10 29 22 11 18 13 7 15 84 84 0.532 14 RW DC 10 10 2 2 4 4 6 9 15 11 10 49 49 0.974 11 2 2 RW DC 55 56 33 30 38 43 83 82 76 83 69 45 27 42 381 381 0.169 1989 1992 ALL Alnc1udes male and female calves. �Table 14. Gross and net Boone and Crockett antler scores and antler weights for buck deer harvested during private fee-only hunting seasons on Forbes Trinchera Ranch, 1987-1992. Results of l-way ANOVA (Type III SS) and linear reareaaton __ of antler variables among years are given. Antler Year F-Test Linear Regression Variable 1987 1988 1989 1990 1991 1992 P P R2 Slope Gross Score Net Score n Wt. (kg) n l64ab 1698 162ab l60b l638b 1638b 0.035 o . 11 153ab 106 1588 108 l52ab 113 150b 108 153ab 116 152ab 107 0.062 0.12 0.004 Neg. 2.668 103 2.07b 102 1.93b 107 1.95b 106 2.09b 115 1.99b 103 0.001 0.01 0.072 Neg. 0.004 Neg. 8bDifferent letters in rows denote yearly means that were significantly different based on Tukey's HSD test P - 0.05. Table 15. Gross and net Boone and Crockett antler scores and antler weights for bull elk harvested during private fee-only hunting seasons on Forbes Trinchera Ranch, 1987-1992. Results of 1-way ANOVA (Type 111'SS) and linear regression of antler variables among years are given. Antler Year F-Test Linear Regression Variable 1987 1988 1989 1990 1991 1992 P P R2 Slope Gross Score Net Score n Wt. (kg) n 266b 2768b 278sb 2838b 280ab 2908 0.057 o . 01 256b 89 266ab 87 267ab 103 2738b 115 268ab 101 2788 91 0.072 0.01 0.012 Pos. 6.348 88 6.228 85 5.958 94 6.578 110 6.228 96 6.658 82 0.297 0.29 0.002 Pos. 0.014 Pos. 8DDifferent letters in rows denote yearly means that were significantly different based on Tukey's HSD test P - 0.05. II ~ \0 �50 Table 16. Gross and net Boone and Crockett antler scores and antler weights for buck deer aged ~ 7 years and 4-6 years old by replacement and wear (RW) and dental cementum (DC). Animals harvested during private fee-only hunting seasons on Forbes Trinchera Ranch, 1987-1992. Results of l-way ANOVA (Type III SS) and linear regression of antler variables among years are given. RW age based on E. Ryland. Antler Year F-Test Linear Regression p P R2 Slope Variable 1987 1988 1989 1992 1990 1991 RW AGE >7 Years Gross Score 168ab 173' Net Score n 157' 57 162' 56 Wt. (kg) n 0.008 0.15 0.007 Neg. 153' 49 152' 48 156' 34 154' 52 0.056 0.09 0.010 Neg. 2.83' a .os- 2.20b 34 2.lsb 49 0.01 0.112 Neg. 46 2.06b 48 0.001 58 DC AGE >7 Years Gross Score 167' 175' 163' 160' 168' 163' 0.087 0.13 0.022 Neg. Net Score n 162' 17 146' 15 150' 21 157' 15 149' 20 0.193 0.20 0.016 Neg. 2.38ab 17 2.07b 14 2.23b 15 2.20b 18 0.001 0.01 0.098 Neg. 165' 164' 162' 163' 161" 0.593 0.77 0.000 Pos. 154' 48 155' 48 151' 43 154' 64 152' 48 0.309 0.49 0.002 Pos. 1.88b 47 0.001 0.02 0.020 Neg. 0.023 0.04 0.011 Neg. ls6ab 68 0.081 0.14 0.006 Neg. 2.03b 66 0.001 0.01 0.094 Neg. 156' 15 wt. (kg) 2.81' n 16 'W AGE 4-6 Years ross dcore 157' Net Score n 146' 42 wt. (kg) 2.40' n 39 DC AGE 4-6 Years Gross Score 171' Net Score n 160'. 65 Wt. (kg) n 2.76' 65 2.06b 63 1.97b 69 2.12b 70 '~ifferent letters in rows denote yearly means that were significantly different based on Tukey's HSD test p = 0.05. �51 Table 17. Gross and net Boone and Crockett antler scores and antler weights for bull e~K aged ~ 7 years and 4-6 years old by replacement and wear (RW) and dental cementum (DC). Animals harvested during private fee-only hunting seasons on Forbes Trinchera Ranch, 1987-1992. Results of 1-way ANOVA (Type III SS) and linear regression of antler variables among years are given. RW age based on E. Ryland. Antler Year Linear Regression F-Test p P R2 Slope Variable 1987 1988 1989 1992 1990 1991 RW AGE >7 Years Gross Score 303" 325" 321" 321" 326" 320" 0.220 0.13 0.014 Pos. Net Score 312" 309" 26 310" 48 313" 26 307" 36 0.191 0.13 0.015 Pos. n Wt. 287" 15 (kg) 9 8.51" 14 DC AGE >7 Years Gross Score 315" 8.04' 22 8.40" 46 8.62" 26 8.24" 33 0.851 0.86 0.001 Neg. 9 319" 342' 324" 326" 320" 0.795 0.93 0.000 Neg. Net Score n 299' 17 326" 15 312' 21 313" 15 304' 23 0.695 0.74 0.001 Pos. 8.70' 17 8.96" 14 8.26' 21 8.66' 15 8.46" 22 0.800 0.44 O. 008 Neg. 286' 280' 277" 281" 283' 0.174 0.41 0.002 Pos. -, 275" 63 268' 61 268' 49 269' 55 273' 48 0.272 0.48 0.002 Pos. n 291" 15 9 • 20" 16 RW AGE 4-6 Years Gross Score 271' Wt • (kg ) n Net Score n 261" 58 wt • ( kg ) n 8.65' ( 6 • 31" 58 DC AGE 4-6 Years Gross Score 286" 6.55" 61 5.88" 57 5.93" 47 5.94" 51 6.02' 42 0.203 0.08 0.010 Neg. 293' 303' 290' 288' 296' 0.178 o . 91 O. 000 Pas. Net Score n 277' 38 283" 56 291" 53 282" 64 277" 60 287" 55 0.224 0.90 0.000 Pos. 7.22" 37 6.85" 55 6.95" 47 6.72' 62 6.31" 55 6.60' 48 0.221 0.03 0.016 Neg. significantly different based Wt. (kg) n '~ifferent on Tukey's letters in rows denote HSD test p = 0.05. yearly means that were �52 Table 18. Results of analysis of variance (Type III SS) for effects of animal age and year on gross and net Boone and Crockett antler scores and antler weights (kg) for buck deer and bull elk aged by replacement and wear (RW) and dental cementum (DC). Animals harvested during private fee-only hunting seasons on Forbes Trinchera Ranch, 1987-1992. RW age based on E. Ryland. F-Test P Net Score Weight Variation Source Gross Score RW Age Year Year*Age Model 0.001 0.965 0.634 0.001 0.001 0.911 0.737 0.001 RW Age Year Year*Age Model 0.001 0.495 0.185 0.001 0.001 0.661 0.312 0.001 F-Test P Net Score Weight Variation Source Gross Score 0.001 0.113 0.799 0.001 DC Age Year Year*Age Model 0.001 0.643 0.345 0.001 0.001 0.567 0.358 0.009 0.001 0.162 0.883 0.001 0.001 0.482 0.375 0.001 DC Age Year Year*Age Model 0.001 0.383 0.155 0.001 0.001 0.243 0.116 0.001 0.001 0.353 0.684 0.001 �Table 19. Eviscerated body weights (kg) for buck deer and bull elk harvested during private fee-only hunting seasons on Forbes Trinchera Ranch, 1987-1992. Results of 1-way ANOVA (Type III SS) and linear regression of body weight among years are given. Antler Variable Year 1989 F-Test Linear Regression P R2 Slope 1990 1991 1992 p 82.3b 108 84.1ab 116 8l.3b 108 0.001 0.01 0.019 Neg. 2l7.9ab 0.033 89 0.01 0.019 Neg. 1987 1988 84.8ab 107 87.7a 108 225.9a 87 223.0ab 224.0ab 220.2ab 211.4b 85 102 112 100 Deer Body Weight n 82.4b 117 Elk Body Weight n aD Different letters in rows denote yearly means that were significantly different based on Tukey's HSD test P - 0.05. Vl W T �54 Table 20. Eviscerated body weights (kg) and hind foot lengths (cm) for female deer and elk harvested during public seasons on Forbes Trinchera Ranch, 19871991. Age-Years Fawns Deer Weight All Avg. (SE) n Hind Foot Avg. (SE) n F M 1 2-3 4-6 z 7 22.3 0.28 122 23.38 21.7b 0.47 0.33 44 77 36.6 0.42 84 42.8 0.28 245 44.5 0.33 197 44.1 0.49 97 41.1 0.23 122 41.6c 40.8c 0.53 0.20 44 78 46.2 0.39 84 47.6 0.43 245 47.9 0.24 197 47.6 0.18 97 Age-Years Calves Elk Weight All Avg. (SE) n Hind Foot Avg. (SE) n a,b,c,d,e M F 1 2 3-4 5-7 8-10 ~ll 65.5 1.56 69 66.7d 64.5d 108.1 134.0 2.44 2.05 2.75 2.00 34 36 40 27 140.8 1.78 68 148.3 1.47 71 155.0 1.79 35 151.1 2.49 37 56.0 0.30 69 56.1 e 56.0e 60.1 2.44 2.05 0.46 28 38 36 62.4 0.33 68 62.4 0.28 71 63.0 0.36 35 63.0 0.44 37 62.2 0.43 40 Different letters within rows denote means that were different between male and female fawns or calves. i _.. ,-.-- �Table 21. Age" Pregnancy rates (PG) for female elk on the Forbes Trinchera Ranch in December. 1986-1992. 1986 PG 5 2 7 7 6 6 5 5 2 2 5 5 1 1 (yrs) 1 2 3-4 5-7 8-10 ~11 Uk-Adb n Totals XPG 31 28 90 1987 PG 6 0 10 10 8 8 14 12 4 4 4 2 2 2 n 48 38 79 1988 PG 7 2 6 4 17 15 16 14 4 4 6 4 4 3 n 60 46 77 1989 PG 5 1 4 2 12 10 16 15 4 4 6 3 1 1 n 48 36 75 1990 PG 8 0 7 3 8 6 8 7 8 8 5 4 1 0 n 45 28 62 1991 PG 4 1 10 9 22 21 12 12 14 14 14 13 3 3 1992 n PG 2 1 5 5 7 5 11 8 12 10 4 3 1 1 79 42 n 73 92 33 79 1986-1992 PG XPG 37 7 19 49 40 84 80 71 89 82 73 89 48 46 96 44 34 77 13 11 85 n 353 282 80 80 SAge for ~ 2 from replacement and wear; for ~ 3 from dental cementum. bAdults of unknown age. Table 22. AgeS (yrs) 1 2 3-4 5-7 8-10 ~11 Fetal sex ratios observed in female elk on the Forbes Trinchera Ranch in December. 1986-1992. 1986 1987 1988 F U M F U M F U 0 1 0 0 0 1 0 1 2 4 4 6 0 2 2 0 1 1 2 5 1 8 5 2 0 0 2 9 1 11 3 0 0 1 3 1 0 1 3 0 1 o 1 1 0 1 2 1 2 0 0 2 0 2 0 1 Uk-Ad'' M 1 1 4 5 1 5 0 Totals XMale 17 6 74 7 .12 24 2 33 26 15 5 63 M 1 1 3 4 0 3 0 1989 1990 1991 1992 F U M F U M F U M F U 0 0 0 0 0 0 1 0 0 0 1 2 0 1 2 0 5 3 1 4 1 0 7 0 3 4 0 13 7 1 2 3 0 11 1 2 5 0 9 3 0 3 5 0 3 1 6 1 1 10 4 0 3 6 1 0 0 1 3 0 9 3 1 3 0 0 1 0 0 0 0 2 1 0 1 0 0 12 24 2 33 13 15 1 46 48 22 3 69 16 15 2 52 1986-92 M F U 3 1 3 18 18 5 35 32 5 36 36 2 24 18 4 23 10 2 5 6 1 144 121 22 54 x Male 75 50 52 50 57 70 46 54 SAge for ~ 2 from replacement and wear; for ~ 3 from dental cementum. bAdults of unknown age. Vl Vl r �56 Table 23. Measurements of elk fetuses collected from Forbes Trinchera Ranch in December 1986-1991 Date/ Statistic Bod:£ Weight (g} Male Female Crown-rum:g (mm} Male Female Hind Foot (mm} Female Male 29 Nov-5 Dec 1986 Mean SE n 27.5 3.32 17 20.8 2.82 6 89.9 4.06 17 85.8 4.22 5 20.4 1.14 17 19.5 1.20 6 12-14 & 19-21 Dec 1987 Mean 118.7 SE 15.78 n 12 59.7 7.04 23 141.4 6.10 12 115.4 4.76 24 42.3 3.15 12 30.3 1.81 24 10-12 & 17-19 Dec 1988 Mean 74.5 SE 6.91 n 26 66.7 7.68 15 120.0 3.90 26 121. 7 4.63 15 32.6 1.51 26 33.1 1.80 15 9-11 & 16-18 Dec 1989 Mean 57.6 SE 17.31 n 12 60.4 9.27 24 102.4 10.82 12 111. 5 6.81 24 27.9 3.70 11 29.9 2.26 23 8-17 Dec 1990 Mean SE n 92.2 14.55 13 78.4 11.50 14 128.0 6.96 13 127.6 5.85 15 35.1 2.69 13 35.7 2.72 15 7-15 Dec 1991 Mean SE n 65.3 5.95 47 77.7 11.93 22 115.6 3.92 48 123.0 5.87 22 30.4 1.41 48 34.1 2.35 22 12-22 Dec 1992 Mean SE N 61.8 14.44 16 73.7 9.17 15 103.4 9.74 16 127.1 6.46 15 27.3 3.33 16 35.1 2.40 15 �57 Table 24. Estimated conception dates and fetal ages for elk on Forbes Trinchera Ranch 1986-1992. Year Mean Conce~tion Date Median Mode Min. Max. n Average Fetal Age (Da~s} Male Female Unk. 1986 1 Oct 30 Sep 2 Oct 19 Sep 23 Oct 26 67 65 46 1987 3 Oct 2 Oct 29 Sep 18 Sep 17 Nov 37 82 74 34 1988 30 Sep 29 Sep 27 Sep 15 Sep 25 Oct 44 75 76 51 1989 3 Oct 4 Oct 4 Oct 18 Sep 24 Oct 35 70 73 0 1990 24 Sep 23 Sep 21 Sep 14 Sep 7 Oct 27 78 78 0 1991 26 Sep 25 Sep 26 Sep 6 Sep 2 Nov 73 74 76 44 1992 2 Oct 29 Sep 28 Sep 17 Sep 5 Nov 32 71 78 38 �58 30 ......::::::_ MEDIAN DATES 25 Z 1988 t 1989 0 20 1991 0 1992 Z 15 _._._---- _--------- .......... 1986-1992 I- Z W _ ................. 1990 W o o a: _ .._ .._ .._ .. 1987 C/) o ----- 1986 n =274 10 W a. 5 0 9/18 10/8 11/17 10/28 MONTH & DAY (S-DAY INTERVALS) 140 120 ~ - 100 o t 80 W 0 ~ 0 60 Z C/) 40 LAVETA PASS SNOW DEPTH MARCH 1 20 AVERAGE 1965-1992 ............... 0 1980 82 84 86 88 90 1992 YEAR Fig. 1. Conception dates for elk on Forbes Trinchera Ranch, 1986-1992 (TOP) and depth of snow on March 1 at the LaVeta Pass snow course monitored by the Soil Conservation Service, 1980-1992 (BOTTOM). �59 Colorado Division Wildlife Research July 1993 of Wildlife Report JOB FINAL REPORT State of ~C~o~l~o~r~a~d~o~_ Project No. W-1S3-R-6 Mammals Research Work Plan No. 3 Elk Investigations Job No. 7 Elk Census Methodology Period Covered: Author: Personnel: July 1, 1992 - June 30, 1993 D. J. Freddy R. Bartmann, CDOW; G. White, CSU Abstract We terminated efforts to use line transects to estimate densities of elk and completed evaluating line transects and quadrats to estimate densities of mule deer. The following abstract is from a draft manuscript (enclosed) to be submitted to the Wildlife Society Bulletin or Journal of Wildlife Management. \ Densities of mule deer (Odocoileus hemionus) on sagebrush (Artemisia tridentata wyomingensis) winter range in Middle Park, Colorado were estimated with helicopter surveys using systematically spaced line transects and randomly selected 2.59 km2 quadrats from 1990 through 1993. We found the negative exponential polynomial, exponential polynomial, and Fourier series sighting distance models adequately fit (£ ;;::0.20) sighting data although no sighting model provided the best fit during all years. Population estimates for line transects ranged from 2,000 to 7,500 deer and varied up to 39% among models within years. Satisfactory precision of ± 22-30% of mean estimates was achieved only when;;::177 groups of deer were detected during surveys. Population estimates for quadrats ranged from 1,647 to 2,361 deer with precision of ± 30-54% and were significantly lower (£ ~ 0.07) than line transect estimates during 3 years. Although both sampling systems suggested similar trends in population status, the numeric discrepancy between population estimators was greater than expected. Our densities of deer may represent the lower bound for efficiently using line transects which cost 57% more in helicopter time to implement and improved precision 4-7% only during 1 year. We suggest that implementation of either sampling system be based on likelihood of adequately meeting the fundamental assumptions of each system. We recommend experiments to test observer ability to accurately estimate distances to groups of deer in rough terrain. ��61 JOB FINAL REPORT ELK CENSUS METHODOLOGY David J. Freddy P. N. OBJECTIVE Evaluate methods to estimate numbers of elk during winter. SEGMENT OBJECTIVES 1. Determine efficacy of line transect methodology to estimate numbers elk and mule deer sharing sagebrush-aspen-conifer winter ranges. 2. Compare estimates on line transects Wildlife of mule deer density in sagebrush winter and 1 mi2 quadrats. ddy Researcher of range based ��63 Estimating Deer Densities. Freddy et al. 7/21/93 DRAFT David J. Freddy Colorado Division of Wildlife 317 W. Prospect Ft. Collins, CO 80526 303-484-2836 x346 RH: Estimating ESTIMATING deer densities. DENSITIES Freddy et al. OF MULE DEER USING LINE TRANSECTS AND QUADRATS David J. Freddy, Colorado Division of Wildlife, Research Section, 317 W. Prospect Rd., Fort Collins, CO 80526 Richard M. Bartmann, Colorado Division of Wildlife, Research Section, 317 W. Prospect Rd., Fort Collins, CO 80526 Gary C. White, Department of Fishery and Wildlife Biology, Colorado State University, Fort Collins, CO 80523 A. Eugene Byrne, Colorado Division of Wildlife, 50633 Highway 6 & 24, Glenwood Springs, CO 81601 Abstract: Densities of mule deer (Odocoileus hemionus) on sagebrush (Artemisia tridentata wyomingensis) winter_range in Middle Park, Colorado were estimated with helicopter surveys using systematically spaced line transects and randomly selected 2.59 kmz quadrats from 1990 through 1993. We found the negative exponential polynomial, exponential polynomial, and Fourier series sighting distance models adequately fit (f ~ 0.20) sighting data although no sighting model provided the best fit during all years. Population estimates for line transects ranged from 2,000 to 7,500 deer and varied up to 39% among models within years. Satisfactory precision of ± 22-30% of mean estimates was achieved only when ~ 177 groups of deer were detected during surveys. Population estimates for quadrats ranged from 1,647 to 2,361 deer with precision of ± 30-54% and were significantly lower (f $ 0.007) than line transect estimates during 3 years. Although both sampling systems suggested similar trends in population status, the numeric discrepancy between population estimators was greater than expected. Our densities of deer may represent the lower bound for efficiently using line transects which cost 57% more in helicopter time to implement and improved precision 4-7% only during 1 year. We suggest that implementation of either sampling system be based on likelihood of adequately meeting the fundamental assumptions of each system. We recommend experiments to test observer ability to accurately estimate distances to groups of deer in rough terrain. Key words: census, quadrats density, ~ Wildl. Manage. 00(0):000-000 line transects, mule deer, Odocoileus hemionus, Accurate estimates of population density or size are fundamental to managing harvests of many ungulate populations (Caughley 1977). Several studies have shown that most survey systems underestimate true densities of ungulates (Caughley 1974, Floyd et al. 1979, Bartmann et al. 1986, Samuel et �64 al. 1987). Approaches to account for this negative bias include developing correction factors for numbers of animals missed (Ackerman 1988, Samuel et al. 1987, Bartmann et al. 1986) or using line transects that inherently account for animals missed during surveys (Burnham et al. 1980). White et al. (1989) and Bartmann et al. (1986) demonstrated that line transects were less negatively biased than area quadrats for estimating densities of mule deer in pinyon-juniper habitats (Pinus edulis-Juniperous osteosperma) of western Colorado. They found that estimates based on line transects and quadrats represented 90 and 66% of the true densities of deer, respectively. These results suggested that if both line transects and quadrats were used to estimate densities of deer over the same area of pinyonjuniper habitat, that average densities based on quadrats would represent about 70% of the average densities derived from line transects. Sampling systems using quadrats are routinely used in Colorado to estimate densities of mule deer in sagebrush and pinyon-juniper habitats during winter (Gill 1969, Kufeld 1980). In open sagebrush habitats, we expect the degree of negative bias inherent in counts of deer to be less than similar counts in pinyon-juniper habitats because of the lack of vegetative overstory that interferes with detecting deer. We hypothesized that in sagebrush habitats, quadrats would provide estimates of density that would be ~ 70% of estimates derived from line transects. We therefore used both aerial line transect and quadrat sampling systems to estimate densities of mule deer in sagebrush habitat from 1990 through 1993. We compared estimated densities and their precision and costs associated with these sampling systems. Each sampling system has assumptions fundamental to proper application. For quadrats those assumptions are: 1) all animals on quadrats are detected and counted only once, 2) animals off quadrats are not counted, i.e., there are no boundary errors. For line transects, primary assumptions are: 1) animals directly on transect lines will never be missed, 2) animals do not move from their initial position before being detected and are counted only once, 3) sighting distances to animals are measured accurately, and 4) sightings are independent events (Burnham et al. 1980). We thank pilots from Ptarmigan, Thunderbird, and High Country helicopters for competent flying and J. Gerrans, C. Craig, and R. Firth of the Colorado Division of Wildlife for field support. -Constructive comments on this paper were- provided by several reviewe-rs. This study is a contribution from Colorado Federal Aid in Wildlife Restoration Project W-153-R STUDY AREA Aerial counts occurred in the Troublesome subunit of deer winter range in the Middle Park basin of northcentral Colorado. During winter, mule deer usually migrate in December in response to increasing snow depths (Gilbert et al. 1970) and concentrate into this 166 km2 area of rolling hills and steep canyons dominated by sagebrush-steppe grassland having isolated pockets of aspen (Populus tremuloides) and conifer (Pinus contorta, Pseudotsu&a menziesii). Within this segment of winter habitat, elevations range from 2,240 to 2,590 m. Since 1968, densities of deer within this area in January have been 8-17 deerfkm2 based on counts on randomly selected 2.59 km2 quadrats. �65 Estimating Deer Densities . Freddy et al. Snow cover and depth were variable among years and may have affected our ability to detect deer during aerial counts. In 1990, 1992, and 1993 snow cover was nearly 100% and counting conditions excellent when we flew in midJanuary but in 1991, many south-facing slopes frequented by deer were devoid of snow even though we delayed flying until early March so that snow might accumulate. Minimum snow depths were 0, 15, 25, and 30 cm in 1991, 1990, 1992, and 1993, respectively. Deer were most concentrated in distribution in 1991 and 1993 and most reluctant to move in response to the helicopter in 1992 and 1993. Temperatures (C) while flying were 2 - 7 in 1990 and 1991, -23 - -7 in 1992, and -7 - 0 in 1993. METHODS We attempted to standardize procedures for aerial counts during all years. Line transects and quadrats were flown with a Bell-Soloy helicopter at 65-80 kmph and 35-50 m above the ground or tree canopy. A pilot, navigator, and primary observer were used on all flights. Three pilots experienced in flying transects and quadrats were used during the 4 years. One primary observer was used in 1990 and a second primary observer was used during 19911993 while the navigator was the same person in all years. These 3 persons each had ~ 5 years of previous experience in flying quadrats but each had ~2 years experience in flying line transects. Transects and quadrats were flown during 9-12 January 1990, 1-8 March 1991, 14-17 January 1992, and 15-23 January 1993. Line Transects Two sets of parallel line transects, each having 26 transects systematically spaced at 1,000 m intervals, were flown. These 2 sets of transects were offset from each other by 500 m which resulted in transects occurring every 500 m across the sampled winter range. Individual transects were 1-14 km long and total length of all transects combined was 602 km. We flew each set of transects twice yearly (4 replicate flights) and generally flew one set in the morning and the second set on the same day but in the afternoon. We attempted to fly each set once in the morning and once in the afternoon but in some years weather affected this schedule. Transects were oriented on true north-south bearings perpendicular to changes in elevation and expected gradients in deer densities (White et al. 1989). Transects were delineated on 1:24,000 scale topographic maps but were not marked with flight markers on the ground. The navigator and observer were responsible for detecting deer. The observer, seated on the right, estimated perpendicular distance to the geometric center of each group of deer and counted animals for each group located from the transect center line to the right (White et al. 1989). The navigator, seated in the middle, maintained course bearing using topographic maps, decided whether groups of deer were to the right or left of the center line, and aided in locating groups on or near the transect center line. The pilot was not routinely used as an observer but occasionally spotted groups that may have been missed by observers. The helicopter did not leave the course centerline to aid in counting deer. �66 Estimating Deer Densities . Freddy et a1. The centerline interval for transects included all groups of deer from 0-15 m to the right of course bearing. Beyond 15 m, estimated distances to groups of deer were placed into 10 m intervals out to 155 m. A deer group was defined as ~ 1 animal in proximity at the same distance from the centerline (White et al. 1989). Distances to groups and deer per group were recorded on a portable tape recorder by the primary observer. Observers practiced estimating distances each day that flights were conducted by flying the helicopter along practice transect lines and observing markers placed at known locations 0 to 150 m from the line on level and steep terrain. Quadrats We counted deer on 30 randomly selected 2.59 km2 quadrats which represented a 47% sample of the 166 km2 area. At least 1 corner of each quadrat was marked on the ground to aid in locating quadrats from the helicopter. The navigator delineated each quadrat by first flying the perimeter in a clockwise manner and along with the observer, counting those groups of deer close to the boundary and on the quadrat. After completing the perimeter, the navigator and observer both searched for and counted deer while systematically flying the interior of each quadrat (Kufeld et al. 1980). Both persons attempted to completely count each group with the highest count recorded by the primary observer. Groups were summed to estimate total deer per quadrat. All quadrats were flown once after completing replicates 1 and 2 but prior to completing replicates 3 and 4 of the line transects. Data Analysis Estimates of population size and variance based on line transects followed methods outlined by White et al. (1989) using program TRANSECT. We used the exponential polynomial (EXPL) , exponential power series (EXPS) , negative exponential polynomial (NEXP), and Fourier series (FSER) sighting probability models with strip-widths of 95, 115, and 155 m. Combinations of these models and strip-widths provided 12 sighting functions for each year. Beyond the centerline interval, distance intervals were 10 m for strip-width of 95 m, 20 m for strip-width of 115 m, and 10 m out to 95 m and then pooled beyond 95 m for strip-width of 155 m. Models were considered to adequately fit observed data when f ~ 0.20 for chi-square goodness-of-fit values (White et al. 1989). We considered replicate sets of-transects to be independent and thus pooled results from the 4 replicate flights to derive population estimates. We had no indication that counting deer caused them to move from 1 transect to another while flying one set of transects and thus we feel duplicated counts within a set of transects did not occur. Estimates of population size derived from quadrats followed methods of Gill (1969) and Mendenhall et al. (1971). We used finite population correction factors for calculating variances because we sampled from a known universe of km2. We expressed precision of population estimates for line transects and quadrats as the ± 95% confidence interval expressed as a percentage of the mean estimate. I_-- We compared deer densities (individualsfkm2) among years using a series of ~-tests for line transects and ANOVA for quadrats (PROC GUM, SAS 1988). I - �01 Estimating Deer Densities. Deer densities were using ~-tests. When we opted to use the though this sighting data. Freddy et al. compared between line transects and quadrats each year comparing yearly estimates derived from line transects, EXPL model as recommended by White et al. (1989) even model did not always provide the best fit of our transect Group size observed among years was compared using ANOVA (PROC GUM). Frequencies of group size observed on line transects and quadrats were compared with chi-square tests (PROC FREQ). Relationships between group size and sighting distances were assessed using linear regression (PROC REG). RESULTS Line Transects No single sighting model consistently fit observed data during all years. Nearly all models failed to fit (f S 0.20) data in 1990 whereas all models fit adequately in 1993 (Table 1). Of the 27 models that adequately fit, 33% were NEXP, 26% were FSER, 22% were EXPS, and 19% were EXPL. Acceptable models were more uniformly distributed among strip-widths with 37% at 155 m, 33% at 95 m, and 30% at 115 m. The greater tendency for the NEXP model to fit was related to sighting data being highly spiked at the centerline interval (Fig. 1). Estimates of population size during all years ranged from about 2,000 to 7,500 deer depending on the sighting model selected. Estimates within years varied up to 39% among models that adequately fit data. Satisfactory precision (± 22-30%) was achieved with EXPL, NEXP, and FSER models only when number of groups was ~ 177 as in 1992 and 1993. Precision varied considerably among models such as in 1993 when all models fit data and precision ranged from ± 22 to 47% (Table 1). Based on EXPL models, there were no differences in population estimates from 1991 to 1993 (f> 0.10, Tables 1, 2). The high population estimate for 1990 was different from most EXPL estimates for 1991 through 1993 (f < 0.07) but caution must be exercised in this interpretation as no EXPL model adequately fit 1990 data. Apparent trends in population size were little affected by which strip-width was used in the EXPL model (Table 2). Average group size ranged from 3.9 to 7.3 deer/group and was larger in 1990 and 1991 (f < 0.001, Table 3). Large groups observed in 1991 corresponded to flying in early March when deer were most concentrated and snow depths lowest. ;--- Group size increased with distance from the centerline when using a strip-width of 155 m (f - 0.002) but correlations were weak (~2 < 0.011, n 774). Effects of group size diminished at strip widths of 115 m (f - 0.09, n - 663) and 95 m (f - 0.97, n - 603). Significant effects of group size were dependent on groups of ~30 animals which accounted for 1% of the total groups detected. Quadrats �68 Estimating Deer Densities . Freddy et al. Estimates of deer density were stable among years (f > 0.20) with population sizes ranging from 1,647 to 2,361 (Table 4). Precision of population estimates ranged from ± 30 to 54%. Best precision was achieved in 1992 when snow cover was complete and deer were dispersed in distribution. Average group size ranged from 6.4 to 13.1 deer/group and, similar to results on line transects, was largest in 1991 when deer were concentrated in distribution (f <0.001, Table 3). Quadrats vs. Line Transects Both sampling systems indicated a stable population; quadrats for all y~ars and line transects for 1991-1993, excluding the possible spurious estimate in 1990 (Fig. 2). Population estimates based on quadrats were $ 60% of estimates derived from line transects and significantly lower (f ~ 0.07) in all years except 1992 when quadrats estimated 79-86% of the transect estimate (Table 4). Comparing preC1S1on of population estimates derived from quadrats and transects is complicated by the multiple models for line transects. For EXPL models compared to quadrats, precision was similar in 1990, 1991, and 1993 (± 32-55%). In 1992, precision of EXPL models was 4-7% better than quadrats and in 1992 and 1993 FSER models provided the most precise estimates (22-27%) of any estimator and were 7-17% better than quadrats. The least precise estimates for both sampling systems occurred in 1991 and 1993 when deer were most concentrated and group sizes generally larger. Conversely, the most precise estimates occurred for both systems in 1992 when deer were most dispersed and in smaller groups (Tables I, 3, 4). Group sizes were larger on quadrats primarily due to observing fewer groups of 1-2 deer and more groups of> 15 deer (f < 0.001). Larger groups on quadrats may reflect the tendency to herd deer when flying quadrats as opposed to passing by groups when flying transects. Flying time for line transects was 57% greater than for quadrats to obtain estimates of population density over the same geographic area (Table 5). Extra costs of transects resulted in more precise estimates (4-17%) in some years and for some sighting models. DISCUSSION We hypothesized that quadrats would provide density estimates ~ 0.70 of the estimates derived from line transects but this occurred in only 1 of 4 years. Because true density of deer was unknown, we cannot explain the reasons for greater than anticipated discrepancies between quadrats and line transects. One cause may be a higher level of negative bias on quadrats than anticipated. If EXPL models provide estimates representing 90% of true density (White et al. 1989) then counts on quadrats represented only 29-49% of true density in 1990, 1991, and 1993 and 74% in 1993 (Table 4). If comparisons were based on the NEXP model, then a lower percentage of deer were detected on quadrats because NEXP estimates represented 100.2 % of true density (White et al. 1989). This potential degree of negative bias on quadrats during the 3 most discrepant years appears extreme. In habitats with tree overstory, �69 Estimating Deer Densities . Freddy et al. Bartmann et al. (1986) and Floyd et al. (1979) found 66 and 50-56% of the deer, respectively, while in sagebrush habitats, Ackerman (1988) detected 68% of the deer. Alternatively, our sighting curve was highly spiked at the centerline interval, similar to data of White et a1. (1989), suggesting the possibility that we inaccurately placed groups of deer in this interval which would inflate estimates of density (White et a1. 1989). In open habitats like sagebrush where forward visibility from the helicopter down the transect line is not hindered by overs tory vegetation, decisions regarding whether deer are to the right or left of the transect line must often be made when deer are near the course bearing and ~ 100 m ahead of the helicopter because deer begin to move in response to the helicopter. If there is even a slight and unperceived tendency for the helicopter to alter course towards deer to be counted, estimated perpendicular distances to deer would be underestimated. Furthermore, the rapid decline in numbers of groups observed at distances >15 m (Fig. 1) cannot be explained by dense vegetation hampering observation of deer as in pinyon-juniper habitats (White et a1. 1989). An important source of centerline error could be observer experience. The degree of differences between quadrat and line transect estimates declined from the first to the fourth year (Fig. 2). Transect estimates likely declined because proportions of groups assigned to the centerline interval steadily decreased from 39 to 22% from 1990. to 1993 suggesting that observers' abilities to conduct line transects may have changed with experience because observers were initially less experienced with line transects than quadrats. Our results support the concerns of White et al. (1989) that selecting an appropriate sighting curve model for estimating population size is somewhat arbitrary. We observed up to 39% variation in population estimates within years among models that adequately fit data according to chi-square criteria. Because we found no consistent model, we support using the EXPL model (White et a1. 1989) although the NEXP model may have been more appropriate for our data. During 1992 and 1993, FSER models having 1 parameter also fit (Table 1) suggesting that a desirable shape criterion (White et. al 1989) was occurring during years when observers were more experienced at flying line transects. Unfortunately, neither system consistently provided satisfactorily precise estimates of population size, even with reasonably high sampling intensity. We flew a 47% sample of potential quadrats and obtained precision of 30-55%. Four replicate flights of line transects yielded the desirable 200 groups of deer in only 1992 and 1993 but in these years precision did approach ± 20% as predicted by White et a1. (1989) (Table 1). Precision was best for both sampling systems when deer were dispersed in small groups and worst when deer were concentrated into larger groups. '_, __ . Line transects cost 57% more in helicopter time to complete than did quadrats but in 1993 provided 4-7% better precision than quadrats when using EXPL sighting models. White et al. (1989) predicted that quadrats would be less costly than transects at low densities of deer because of anticipated difficulty in detecting adequate numbers of groups along transects to generate acceptable precision. We suggest that our difficulty in obtaining 200 groups of deer with minimum deer densities of 9.9-14.2 deerfkm2 (Table 4) indicates these densities were at the lower bound for efficiently using line transects. / �70 Estimating Deer Densities . Freddy et al. When using line transects in sagebrush habitats, we suggest avoiding a long truncation distance of 155 m even though more groups of deer will be detected. Bias associated with detecting larger groups of deer (Drummer and McDonald 1987) was minimal at truncation distances of 95 or 115 m and observers can focus their scanning time on detecting groups of deer near the centerline. Precision and mean estimates were not appreciably affected by shorter truncation distances in those years when models adequately fit data (Table 1). MANAGEMENT AND RESEARCH IMPLICATIONS The use of line transects or quadrats to estimate deer densities in sagebrush habitats is not a choice based upon distinct criteria. Which system is used, is in part, dependent on the likelihood of meeting or violating the assumptions of each sampling system. We undoubtedly missed deer on quadrats, thus violating the primary assumption of area based sampling systems. Similarly, we violated the primary assumptions of line transects, i.e., we probably missed deer on the centerline of transects and made errors in estimating distances to groups of deer because of the tendency for groups to move in response to the helicopter and because observers constantly coped with changes in terrain that affected their perceptions of distances to groups. Because our data suggests that observer training may play a larger role in successfully executing line transects than quadrats, we recommend experimental tests of observers abilities to accurately estimate distances to groups of deer in rough terrain. LITERATURE CITED Ackerman, B. B. 1988. Visibility bias of mule deer aerial census procedures in southeast Idaho. Phd. Thesis. University of Idaho, Moscow. 106pp. Bartmann, R. M., L. H. Carpenter, R. A. Garrott, and D. C. Bowden. 1986. Accuracy of helicopter counts of mule deer in pinyon-juniper woodland. Wildl. Soc. Bull. 14:356-363. Burnham, K. P., D. R. Anderson, and J. L. Laake. 1980. Estimation of density from line transect sampling of biological populations. Wildl. Mono. 72. 202pp. Caughley, G. 1974. Bias in aerial survey. J. vriai. Manage. 38:921-933. Caughley, G. 1977. Analysis of vertebrate populations. John Wiley & Sons, New York. 232pp. Drummer, T. D., and L. L. McDonald. 1987. Size bias in line transect sampling. Biometrics 43:13-21. Floyd, T. J;, L. D. Mech, and M. E. Nelson. 1979. An improved method of censusing deer in deciduous-coniferous forests. J. Wildl. Manage. 43: 258-261. Gilbert, P. F., O. C. Wallmo, and R. B. Gill. 1970. Effect of snow depth on mule deer in Middle Park, Colorado. J. Wildl. Manage. 34:15-23. Gill, R. B. 1969. A quadrat count system for estimating game population. Colo. Game, Fish, & Parks Game Info. Leaflet 76. 2pp. Kufeld, R. C., J. H. Olterman, and D. C. Bowden. 1980. A helicopter quadrat census for mule deer on Uncompahgre Plateau, Colorado. J. Wildl. Manage. 44:632-639. �71 Estimating Deer Densities. Freddy et al. Mendenhall, W., L. Ott, and R. L. Scheaffer. 1971. Elementary survey sampling. Duxbury Press, Belmont, Calif. 247pp. Pollock, K. H., and W. L. Kendall. 1987. Visibility bias in aerial surveys: a review of estimation procedures. J. Wildl. Manage. 51:502-510. Samuel, M. D., E. O. Garton, M. W. Schlegel, and R. G. Carson. 1987. Visibility bias during aerial surveys of elk in northcentral Idaho. J. Wildl. Manage. 51:622-630. SAS Institute, Inc. 1988. SAS/STAT user's guide for personal computers. SAS Inst. Inc., Cary, N. C. 1028pp. White, G. C., R. M. Bartmann, L. H. Carpenter, and R. A. Garrott. 1989. Evaluation of aerial line transects for estimating mule deer densities. J. Wildl. Manage. 53:625-635. �72 Estimating Deer Densities Freddy et al. Table 1. Population estimates for IIIJledeer based on sighting probabil ity IIOdels for line transects in the Troublesome subunit, Middle Park, Colorado, 1990-1993. Deer Model Estimated Distance/ Groups Fit No. POl:!!:!lation Size Density b Cf)c Year Model· Intervals Parameters Mean Precision~ Deer/km2 n 1990 EXPL EXPL EXPL 155/10 95/9 115/6 185 139 155 0.001 0.091 0.001 2 2 2 4833 7012 5854 32 37 35 29.1 42.2 35.3 EXPS EXPS EXPS 155/10 95/9 115/6 185 139 155 0.001 0.202 0.001 2 2 2 6102 8925 7462 68 84 75 36.8 53.8 44.9 NEXP NEXP NEXP 155/10 95/9 115/6 185 139 155 0.001 0.142 0.001 4833 7012 5854 27 45 39 29.1 42.2 35.3 FSER FSER FSER 155/10 95/9 115/6 185 139 155 0.001 0.744 0.821 2 3 4 4507 6618 7250 32 38 37 27.1 39.9 43.7 EXPL EXPL EXPL 155/10 95/9 115/6 123 101 109 0.180 0.188 0.144 2 2 2 4358 3912 4209 39 55 67 26.3 23.6 25.4 EXPS EXPS EXPS 155/10 95/9 115/6 123 101 109 0.257 0.196 0.178 2 2 2 5363 4376 5096 79 87 89 32.3 26.4 30.7 NEXP NEXP NEXP 155/10 95/9 115/6 123 101 109 0.254 0.271 0.247 4357 3910 4208 38 42 40 26.2 23.6 25.4 FSER FSER FSER 155/10 95/9 115/6 123 101 109 0.153 0.177 0.151 2 1 2 3684 2842 3600 34 35 36 22.2 17.1 21.7 EXPL EXPL EXPL 155/10 95/9 115/6 254 186 212 0.489 0.377 0.185 2 2 2 3006 2759 2937 23 26 24 18.1 16.6 17.9 EXPS EXPS EXPS 155/10 95/9 115/6 254 186 212 0.567 0.021 0.249 2 2 2 3347 64 1999 24 3231 73 20.2 12.0 19.5 NEXP NEXP NEXP 155/10 95/9 115/6 254 186 212 0.597 0.490 0.305 3006 2758 2941 27 31 29 18.1 16.6 17.7 FSER FSER FSER 155/10 95/9 115/6 254 186 212 0.378 0.356 0.165 2587 2490 2614 23 27 25 15.6 15.0 15.8 EXPL EXPL EXPL 155/10 95/9 115/6 212 177 187 0.957 0.915 0.917 2 2 2 3405 3366 3016 40 46 44 20.5 20.3 18.2 EXPS EXPS EXPS 155/10 95/9 115/6 212 177 187 0.961 0.921 0.908 2 2 2 3303 3261 2943 40 47 34 19.9 19.6 17.7 NEXP NEXP NEXP 155/10 95/9 115/6 212 177 187 0.971 0.955 0.708 3589 3466 3615 26 30 28 21.6 20.9 21.8 FSER FSER FSER 155/10 95/9 115/6 212 177 187 0.769 0.978 0.905 2610 2953 2862 22 25 23 15.7 17.8 17.2 1991 1992 1993 "Models are: EXPL =exponential polynomial, EXPS = exponential power series, NEXP = negative exponential, FSER = Fourier Series. ~istance/Intervals = strip width (m) and number of sighting distance intervals. CChi-square goodness-of-fit tests; models adequately fit data when f ~ 0.20. dprecision is ! 95% CI expressed as percentage of mean estimate. -;--" �IJ Estimating Deer Densities . Freddy et a1. Table 2. Results of ~-tests comparing estimates of deer population size derived from line transects using the EXPL model with strip-widths of 95, 115, and 155 m in the Troublesome subunit. Middle Park. Colorado. 1990-1993. EXPL Model Years Compared @95 @115 @155 1990 vs. 1991 f < 0.07 NSS& NSS 1990 vs. 1992 f < 0.01 f < 0.01 f < 0.05 1990 vs. 1993 f < 0.05 f < 0.05 NS 1991 vs. 1992 NS NSS NS 1991 vs. 1993 NSS NSS NSS 1992 vs. 1993 NSS NSS NSS & NSS - f > 0.20; NS - 0.19 > f > 0.10. Table 3. Average group sizes of mule deer observed on line transects having strip-width of 155 m and on 2.59 km2 quadrats, Troublesome subunit, Middle Park Colorado 1990-1993. Sampling Group Size Year System Mean Max. n 95% UCI 95% LCI Min. 1990 1991 1992 1993 Transects Transects Transects Transects 1990 1991 1992 1993 Quadrats Quadrats Quadrats Quadrats 5.9& 7.3& 4.1b 3.9b 185 6.5 123 9.2 254 4.5 212 4.6 Not Available l3.1c 61 17.1 6.4d 172 7.3 7.0d 117 8.2 5.2 5.5 3.6 3.2 1 1 1 1 30 68 32 53 9.0 5.5 5.7 1 1 97 34 35 1 &,b Different letters denote significantly different sizes of groups observed on transects (f < 0.001). c,d Different letters denote significantly different sizes of groups observed on quadrats (f < 0.001). �Estimating Deer Densities. -...J Freddy et a1. .J::- Table 4. Estimates of deer population size based on 30-2.59 km2 quadrats for the Troublesome subunit, Middle Park, Colorado, 1990-1993 with results of ~-tests comparing yearly deer densities estimated with quadrats and line transects using the EXPL model and strip-widths of 95, 115, and 155 m and quadrat density estimates expressed as percentages of corresponding transect estimates. Estimated Population Size Density Quadrats vs, EXPL Models Quadrats Percent of Transectsb a Year Mean Precision Deer/km2 AVG. @115 @155 @95 @115 @155 @95 1990 1991 1992 1993 a b c d 1848c 1647c 2361c 1739c Precision See Table Means not NSS - f > 32 54 30 39 11.2 9.9 14'.2 10.5 £ < 0.001 £ - 0.05 £ < 0.001 £ < 0.001 £ - 0.06 f < 0.01 NSSd f - 0.06 NSS f - 0.07 NSS f < 0.05 is ± 95% CI expressed as percentage of mean estimate. 1 for population estimates based on transects. different among years (ANOVA, f > 0.20) 0.20 Table 5. Time devoted to flying line transects and quadrats to estimate deer densities in the Troublesome subunit of Middle Park, Colorado, 1990-1993, Time (minutes} Mean (CV) Flight Method 1990 1991 1992 1993 Transects Set 1 Rep 1 Transects Set 2 Rep 1 Transects Set 1 Rep 2 Transects Set 2 Rep 2 Transects Yearly Total Quadrats Yearly Total r 159 172 159 167 657 475 182 186 185 171 724 385 193 179 185 170 727 472 151 176 151 156 634 417 171 178 170 166 685 437 (11) ( 3) (10) ( 4) ( 7) (11) 26 42 86 52 32 39 80 58 38 38 79 51 32 40 82 54 �75 200 1990 -1993 CIJ 0..150 => o a: ~ L1.. 0100 a: W a::l :::! =>50 Z o DISTANCE FROM CENTER LINE (m) Fig. 1. Groups of deer observed within distance intervals on line transects In the Troublesome subunit, Middle Park, Colorado, 1990-1993 �76 8,000 0 EXPL@155 m EXPL@115m EXPL@95m 0 0 QUADRATS MEAN +/- 95% CI w 6,000 - * • 0 N en Z * 0 ~ 4,000 ~ ~ - u ::> c.. 0 c.. €!l 0 !~ ....•. 2,000 o z-= -.-- .-.-. I I 1990 1991 ~ ~ I I 1992 1993 YEAR Fig. 2. Mean population estimates of total deer based on line transects for EXPL models at 95 m, 115 m, and 155 m and quadrats, Troublesome subunit, Middle Park, Colorado, 1990-1993. Mean estimate +/- 95% CI shown for quadrats. �������83 Colorado Division of Wildlife Wildlife Research Report July 1993' JOB PROGRESS REPORT Project ,No. W-lS3-R-6 MammalsResearch Elk Investigations Work Plan No. Estimating Survival Rates of Elk and Developing Techniques to Estimate Population Covered:; .Author: D.. J. July Size 1, 1992 - June 30, 1993 Freddy Persbru'l.e1:< .R. Bar tmann, and C. McCarty, CDOW; C. Vardeman, D. Bowden and G. Wh-i t.e , CSU;< Abstract A detailed study plan for this project was completed and approved through the required.peer-;review system. The study area selected was that portion of Gaine Management Unit 42 south of NewCastle and Rifle, Colorado. Two elk corral traps were refurbished for use in December 1993 and arrangements made to contract tI:'app~ng :of 100 elk using a helicopter. Radio-collars for adult cows and male and female calves were designed and purchased. ��JOB PROGRESS REPORT ESTIMATING SURVIVAL RATES OF ELK AND DEVELOPING TECHNIQUES TO ESTI~1ATE POPULATION SIZE David J. Freddy P. N. OBJECTIVE Estimate survival rates of adult female and calf elk and develop techniques to estimate population size. SEGMENT OBJECTIVES 1. Complete detailed and approved study plan and select study area. 2. Refurbish elk traps and identify potential trap-site locations within the selected study area. 3. Design expandable radio-collars for calves and purchase and assemble collar materials transmitter units. A complete copy of the approved study plan presented. Prepared by �86 STUDY PLAN FOR RESEARCH FOR FY 1992-93 - 1996-97 State of Colorado Project No. 4400 Work Plan No. Job No. Hunt 0715 Mammals I Research Elk Investigations 3 Estimating Survival Rates of Elk and Developing Techniques to Estimate Population Size 9 ESTIMATING SURVIVAL RATES OF ELK AND DEVELOPING TECHNIQUES TO ESTIMATE POPULATION SIZE Principal Investigators David J. Freddy, Wildlife Researcher, Mammals Research R. Bruce Gill, Wildlife Research Leader, Mammals Research Cooperators John E. Ellenberger, Wildlife Biologist, Northwest Region James H. Olterman, Wildlife Biologist, Southwest Region David C. Bowden, Professor Statistics, Colo. St. Univ. Gary C. White, Professor Wildlife Biology, Colo. St. Univ. STUDY PLAN APPROVAL Prepared by: Date: _ Submitted by: Date: _ Reviewed by: Date: ~ ___________________________ Date: _ ___________________________ Date: _ ___________________________ Date: _ Date: _ Approved by: Biometrician ___________________________ Date: _ { - -1-- --. �87 'FREDDY STUDY PLAN PROGRAM NARRATIVE State of Project No. Colorado 4400 3 Work Plan No. Job No. I. Hunt 9 0715 Mammals I Research Elk Investigations Estimating Survival Rates of Elk and Developing Techniques to Estimate Population Size NEED Elk (Cervus elapbus nelsoni) are a high-profile wildlife resource throughout much of Colorado because they provide recreation for persons who hunt, watch, and photograph wildlife (Freddy et al. 1993). The popularity of elk with the public is undoubtedly related to the expanding geographic distribution and increasing size of the State's elk population. This burgeoning elk resource has many benefits but frequent social, political, and economic conflicts suggest elk, to some degree, have reached a "social" carrying capacity. Balancing these uses and conflicts by maintaining acceptable numbers of elk is a challenge for future management of this species in Colorado. The positive and negative economic impacts of elk are often the most influential forces directing management of elk. In 1990 and 1991, 193-194,000 hunters harvested 46-51,000 elk in Colorado (CDOW 1990, 1991). In doing so, hunters generated at least $20 million in yearly license revenue to the Colorado Division of Wildlife (CDOW) and contributed about $250 million in direct and secondary expenditures to the yearly economy of Colorado (Freddy et al. 1993). Revenues derived from elk hunting provide the financial resources for many wildlife programs of the CDOW and for many private businesses in the State. However, rising numbers of elk and attendant harvests have also brought about conflicts with agricultural interests. Demands to reduce elk populations and expand payments for damage by elk to cultivated and native forage on private lands have resulted in the CDOW altering statewide objectives for elk from increasing to decreasing the total population and creating a Habitat Partnership Program with private land owners. This partnership to cooperatively manage elk that impact private lands may reach expenditures of $1 million annually. ,--- The core of the conflict in elk management, in part, stems from inadequate estimates of population size resulting in an inability to establish management objectives for specific populations that are agreeable to competing interests. Even when objectives are established, we can not readily quantify impacts of management decisions because current methodology for monitoring populations is neither precise nor accurate. �88 FREDDY STUDY PLAN The cnow has relied on computer modeling, rather than direct measurements, to generate estimates of population size. This approach has been inadequate because model outputs have not been validated with independent estimates of population size. Estimates of population size and subsequent proposed harvest levels have, therefore, been subjected to endless debate. Computer models based on POP-II (Bartholow 1992) or POPMOD (White 1992) function best when natural rates of survival or mortality, hunter harvests, and population size are reliably measured. In the current modeling process, survival of calves from June 1 until September 1 is implicitly estimated by measured post-hunting season calf:cow ratios, which along with bull:cow ratios, are used to estimate composition of the pre-hunting season population. However, the size of the preseason population is not measured. From this preseason population, hunter harvests, which are adequately measured, are subtracted resulting in a prewinter population. The prewinter population is multiplied by rates of survival, which are not measured, to estimate a postwinter population and subsequently the next preseason population. Obvious missing links in this modeling process are measured estimates of survival for calves and adult females during winter (Nelson and Peek 1982) and population size at some time during the annual cycle. Importantly, estimates of population size independent of the modeling process could be used to align models on measured values. The current modeling process is prone to many subjective estimates regarding population size and survival rates which has resulted in diminished confidence in using models to guide management of elk populations (Freddy 1987). The CDOW recognized the need to improve monitoring status of ungulate populations during 3 planning efforts. First, the statewide Long Range Plan (LRP), approved in 1988 and updated in 1991 (CnOW 1991b), highlighted the need for improved population data to accommodate increasing recreational demands and reduce conflicts among competing interests. The LRP targeted 25% of the revenues generated from increasing hunting license fees towards improving and expanding efforts to monitor ungulate populations. In 1991, increased fees associated with elk licenses generated approximately $4.5 millon in new revenue potentially res~lting in $1.1 millon towards expanded population monitoring. Second, CDOW terrestrial biologists at,a statewide meeting held in July 1991 stated.that obtaining quantified estimates of surviv:al rates and population size ~erE~'their priority needs for developing more r~liable models of elk populations. Third, the. GDOW, Deer and Elk Management Analysis Guide, approved in November 1991, identified as priority issues the .need to obtain reliable estimates of population size and survival, rates of calves and adult female elk (Freddy et a l., 1993). Estimating both survival rates and popu La t Lon size are fundamental to developing systems for monitoring and predicting changes in elk populations. We believe estimating calf and adult female survival rates during winter and annual survival rates of adult females are a higher priority than estimating adult male survival rates primarily because most males are harvested when they reach legal age and contribute little to problems of either popUlation growth or decline. Models having valid estimates of survival rates for calves and adult females along with currently obtained estimates of harvests and population composition will provide more defensible estimates of population �FREDDY STUDY PLAN size and trend than similar estimates derived from current models. However, an independent estimate of population size generated from a sampling system designed to enumerate elk would provide the opportunity to validate population estimates derived from models. Ideally, estimates of both survival and size would be obtained simultaneously. Changes in either calf or adult female survival have pronounced effects on population growth and greater effects than changes in fecundity (Nelson and Peek 1982). Sensitivity analyses indicated population growth is more affected by changes in survival rates of adult females than to equivalent changes in calf survival. A 15% increase in adult female survival resulted in a population increase 7x that of the increase from a 15% change in calf survival (Table 1). Although small changes in adult female survival can have major effects on population growth if compounded for 10-20 years, we expect calf survival to be more variable among years and our ability to detect changes in calf survival should be greater than detecting smaller, but important changes in adult female survival (White et al. 1987, Bear 1989, Bartmann et al. 1992). Because there are few estimates of either calf or adult female survival rates, we need to measure both simultaneously to document the relative differences in survival rates and patterns of mortality in order to develop more reliable population models. An example of how measurements of survival can alter our perception and modeling of populations involves the Piceance deer population in northwest Colorado. Measured overwinter survival of fawns (0.22) was much lower than perceived, and survival of adult females (0.82) was more stable among years than anticipated (White et al. 1987). Using measured rates of survival in existing models of this deer population resulted in radical changes in estimated population size, and served to focus wildlife managers on probable causes of low fawn recruitment and solutions to improve recruitment (Bartmann and White 1991, Bartmann et al. 1992). �90 FREDDY STUDY PLAN Table 1. Sensitivity analysis for 3, 10, and 15% increases in survival rates of calves and adult females for a beginning population of 1000 elk simulated for 20 years. Initial survival rates for calves (0.705) and adult cows (0.800) essentially stabilized population growth for 20 years. Growth rate increase equals the percent increase in yearly average growth rate compared to stabilized average growth rate. Change in Calf Survival Rate Calf Survival Ad. Female Survivala Ad. Male Survival Avg. Yrly. Growth Rate Growth Rate Increase End 20 Yr. Pop. Size 0% +3% +10% +15% 0.705 0.800 0.650 0.82% 0.00% 990 0.726 0.800 0.650 l.49% 82.5% 1109 0.775 0.800 0.650 3.21% 293.0% 1441 0.811 0.800 0.650 4.61% 465.0% 1742 Change in Adult Cow Survival Rate +3% +15% 0% +10% Calf Survival Ad. Female Survival Ad. Male Survival Avg. Yrly. Growth Rate Growth Rate Increase End 20 Yr. Pop. Size a 0.705 0.800 0.650 0.82% 0.00% 990 0.705 0.824 0.650 3.35% 310% 1531 0.705 0.880 0.650 12.1% 1378% 4093 0.705 0.920 0.650 22.1% 2609% 8038 Ad.-adu1t ~ 1 year old. Our lack of adequate population data on elk is not without just cause as known methodologies are limited and expensive. As resource ,managers, we have likely reached a threshold where we must expand our investment in monitoring the processes that affect the growth or decline of the elk resource to accommodate the ,rising and often conflicting demands for managing elk. II . OBJECTIVES Our objectives are to provide reliable estimates of survival rates for calves and adult females overwinter and adult females throughout the year, and develop and test a system for estimating population size. Our efforts will focus on elk inhabiting winter ranges south of NewCastle and Rifle, Colorado. Our specific objectives are: 1) Estimate survival rates of calves from 1 December - 31 May within ± 15% of the true survival rate at the 95% confidence interval and identify sources of mortality using adequate numbers of radio-collared calves. 2) Estimate winter (1 December - 31 May) and yearly (1 December - 30 November) survival rates of adult females within ± 10% of the true �91 FREDDY STUDY PLAN survival rate at the 95% confidence interval and identify sources mortality using adequate numbers of radio-collared adults. of 3) Monitor the fate of yearling bull elk radio-collared as calves in relation to antler-point restrictions designed to protect yearling bulls from harvest. 4) Develop and test a system to estimate population size during winter within ± 20% of the true population size ,at the 90% confidence interval using radio-collared elk to evaluate sighting bias and to provide population estimates based on random plot searches and mark-resight theory. 5) Provide data on general movements, distribution, and dispersal of radio-collared elk within the study area. Establish cooperative studies with universities and land management agencies to investigate specific aspects of habitats used by elk and to relate genetic characteristics of elk to their survival. III. EXPECTED RESULTS OR BENEFITS This investigation will provide estimates of survival rates for calf and adult female elk in a selected elk population during 4 consecutive years. These estimates will immediately assist CDOW biologists not only in refining population models for this specific population but also in providing estimates of survival that may be applicable to modeling other elk populations inhabiting similar habitats. Quantified estimates of survival will allow objective assessment of whether major restructuring of current population models is necessary. In the process of estimating survival rates, we will document sources of mortality. Our efforts to estimate population size are inherently more difficult than estimating survival rates. If we obtain relatively precise and unbiased estimates of true popUlation size, the CDOW will have a method to obtain valid independent estimates of population size to compare with estimates derived from modeling, and therefore, provide another objective basis for assessing the current modeling process. If our efforts provide less than satisfactory results, we will have an objective basis for either continuing with refinements or discontinuing efforts to estimate population size and invest in measuring other population parameters. IV. APPROACH A. Survival ,-"-- General Approach We will estimate survival rates for calves (6 mos old) and adult females (~ 1 yr old) during winter, 1 December - 31 May, and for adult females annually, 1 December - 30 November, by using radio-telemetry collars that emit a mortality pulse code when animals remain motionless for 4-6 hours (White 1983, Bear 1989, White et al. 1987). Radios provide the ability to know the fate of individual animals (alive or dead) over discrete periods of time unlike marking animals with other types of tags that are seldom recovered upon the animal's death (White and Garrott 1990). Improved longevity of batteries �92 FREDDY STUDY PLAN for radio collars will allow us to monitor status of individual animals to 4+ years. Survival will be monitored from 1993-94 through 1996-97. for up We assume that radio-collaring calf and adult female elk does not bias estimates of survival rates by jeopardizing or enhancing the welfare of individuals. This assumption can be violated in studies involving birds when transmitter packages exceed critical weightfbody size ratios (Burger et al. 1991, Foster et al. 1992) but does not appear to be a detectable bias with ungulates when radio collars weigh about 0.8% of an animal's body weight (Garrott et al. 1985, White et al. 1987). Although some level of bias may occur, radio collars will likely provide much less positively biased estimates of survival compared to other indirect methods (Bartmann 1984, Zager and Leptich 1991). Radio collars used in this project will weigh about 1.1 kg and represent about 0.8% of calf and 0.5% of adult female body weights. Collars for adult females will be of fixed circumference and fitted for each individual. Collars for male and female calves will allow for expansion to adult size (Bear 1986, White et al. 1987, Appendix I). We assume that survival of those animals captured unbiasedly represents survival of individuals in the population. Individual behavior, social behavior, trapping method, and distribution of trapping effort all potentially bias those individuals trapped and marked (White el al. 1982, Garrott and White 1982). Recognizing these problems, we will capture elk with the objectives of marking animals throughout the distribution of the population and reducing influences of social hierarchies. The study area will be divided into several zones each having multiple capture sites to assure that marked elk represent the population. Capture quotas for calves and adult females will be established for each zone. A Hughes 500 helicopter and a net-gun will be used to capture individuals from groups of elk in more remote areas where portable corral traps would be inefficient to use. In those areas inhabited by people or having agricultural activities, we will use portable corral traps to capture groups of elk. Elk caught in corral traps are likely less affected by social dominance hierarchies than elk captured with single animal Clover traps (Garrott and White 1982, Bear et al. 1989). For groups of elk captured in traps, only a fraction of each group will be randomly selected and radio collared. We will attempt to collar equal numbers of adult females and calves within.each capture method. Trapping will occur from .approximately 29 November-20 December each year. While handling animals we will mi"nimize capture stress and follow guidelines for animal welfare (Appendix II). Each collared animal will be individually recogniz~ble based on radio collar frequency, an external symbol/number on the top surface of the collar, and a numbered metal ear-tag placed in each ear. Radio collars will be white with a black symbol/number (Appendix III). Additionally, male and female calves or yearlings captured but not collared will be ear-tagged to provide additional data on dispersal when and if recovered as harvested animals during hunting seasons. Collared calves, and when possible, yearling females, will be weighed to nearest 0.5 kg. Additionally, hind foot lengths and total body lengths will be measured for calves. Blood samples will be obtained via venepuncture to assess pregnancy status of adult females using progesterone assays (Freddy _'--' �93 F~DDYsroDYPUW 1991) and from calves for potential genetic analyses related to measured survival rates (Pemberton et al. 1988). All collared elk will be aged as calves, yearlings, or young (2-3 yis) , prime (4-10 yrs) , and old (~ 10 yrs) adults based on visual inspection of tooth replacement and wear (Quimby and Gaab 1957). We will use periodic and systematic aerial and ground searches throughout areas inhabited by collared elk to determine'their life or death status based on pulse rates of collars (Gilmer et al. 1981). Animals will be monitored daily from the ground during winter with aerial searches done twice per month depending upon frequency of mortalities and relative ability to monitor all animals from the ground. During spring, summer, and fall we will conduct monthly or bi-monthly aerial searches. During fall hunting seasons, aerial searches may be conducted weekly in conjunction with intensified ground searches. Mortalities coarsely located during aerial surveys will be found from ground searches using hand-held antennas. Criteria for assigning probable cause of death (primarily during winter) will include body position, presence of bite or claw marks and subdermal hemorrhaging, tracks, drag marks, and when necessary rumen or tissue samples (Wade and Browns 1982). Potential causes of death include starvation, accidental trauma, plant poisoning, predation by black bears, mountain lions, coyotes, and domestic dogs, and legal and illegal hunter harvest (Bear 1989, Schlegel 1977). Analyses Number of animals collared depends on expected survival rates, desired precision, and degree of statistical power desired for comparative tests. The few studies that have estimated survival rates for elk used either age structure analyses or radio collars. Annual survival rates estimated from age structure analyses were: 64% for all calves (overwinter only) in Yellowstone National Park (Houston 1982); 61% for female and 26% for male calves in Jackson Hole, Wyoming (Boyce 1989); 90% for all calves in Colorado's White River population (Laake 1992); and for adult females, 97% (Houston 1982), 68% (Boyce 1989), and 79% (Laake 1992). Annual survival rates based on radio telemetry were: 93% for male and female calves pooled with adult females (nonhunted segmen~s of the population) in northern New Mexico (White 1983); 75-94% for calves through their first 9 months in Rocky Mountain National Park, Colorado (Bear 1989); 100% for calves from shortly after birth through their first winter in central Colorado (de Vergie 1989); 88% for adult females in northern Idaho where all mortality was attributed to hunting (Leptich and Zager 1991); and nearly 100% for adult females in northern Colorado (overwinter only, G. D. Bear pers. comm.). These results suggest that survival may be > 70% for calves and> 85% for adult females overwinter with annual survival rates for adult females varying due to intensity of hunting. We will radio-collar 75 calves (38 female, 37 male) yearly and initially collar 75 adult females and maintain at least 75 adults (~ 1 year old) collared in the population yearly to estimate rates of survival for these 2 age classes of animals. Assuming 75 collars remain functioning and survival is ~ 0.7 or ~ 0.9, 95% confidence intervals will be ~ ± 15% or ~ ± 8% of the �94 FREDDY STUDY PLAN mean survival rate, respectively. If survival for either calves or adults approaches 0.5 then considerably more collars must be deployed to maintain high precision (Appendix IV). These confidence intervals are calculated as 1.96 ± VVAR(S) where survival (S) is a binomial with a variance, VAR(S) - S(lS)/n, and n is number of collars deployed (White et al. 1982). Power (l-beta) to detect differences in survival rates ~ 0.20 among years or age/sex classes will be ~ 0.70 but power will be S 0.25 for differences in survival S 0.10 when survival rate is 0.7 and alpha - 0.05. If survival is 0.9, power will be ~ 0.80 for differences in survival ~ 0.20 and 0.40 for differences S 0.10 (Appendix IV, White and Garrott 1990:32). Reliably detecting differences of 0.05 in survival rates would require deployment of about 1,000 collars yearly for each age class. Analyses of survival rates assumes that individual animals are known to be either alive or dead during specific time intervals of interest and that their individual life status is independent among animals. This binomial approach is most correct if all animals of interest enter and leave the population simultaneously and there are no animals of unknown status (censored) due to emigration or failure of radio collars (White and Garrott 1990). Typically, animals are trapped and marked over varying time periods and some animals emigrate or radio collars fail. To account for these problems, we will use a staggered entry Kaplan-Meier analysis, which is a modification of the binomial estimator, to estimate survival rates (Pollock et al. 1989, White and Garrott 1990, Bartmann et al. 1992). If no censoring occurs, the Kaplan-Meier approach will provide an estimate identical to the binomial estimator. We believe the binomial approach and attendant assumptions provides a more realistic basis for estimating survival rates than alternative approaches that assume constant survival rates for specified time intervals (Heisy and Fuller 1985). Staggered entry Kaplan-Meier analyses are readily available in programs written for SAS (SAS 1988, White and Garrott 1990, G. C. White, pers. comm.). The PC database program RADIOS (G. C. White pers. comm) will be used to store data on each collared animal. We will test the following generalized hypotheses: Ho: Survival Rate - S Scalves- Sadultfemalesduring winter Ho: Smalecalves- Sf_aleca~~es'during'winter Ho: Sadultfemales is equal among years Ho: Scalves is equal among years We anticipate the following comparative analyses of survival rates using data from animals of known fate. We will conduct pair-wise comparisons using log-rank tests (Pollock et al. 1989, Bartmann et al. 1992) to compare survival functions for: calves and adult females during winter and for male and female calves during winter. Program SURVIV (White and Garrott 1990) will be used to test for differences in survival rates among and between age classes and years (Table 2). For example, to compare survival rates of adult females among �95 . i FREDDY STUDY PLAN years, constraints for program SURVIV would be Sl - S2, Sl - S3, Sl - S4, and to compare survival of adult females and calves within years, constraints would be Sl - S5, S2 - S6, S3 - S7, S4 - Sa. Similar constraint pairings would occur for comparisons between male and female calves. These reduced models of survival would be tested against a generalized model that estimates all survival parameters individually, using log-likelihood ratio and goodness-offit tests in program SURVIV. Tests will be significant at P ~ 0.05. Table 2. SURVIV. Age/Sex Class Generalized structure for analyzing survival rates with program Year Radios Deployed Proportions Lived Died Adult Females 93-94 94-95 95-96 96-97 75 75+ 75+ 75+ Sl S2 S3 S4 (l-Sl) (1-S2) (1-S3) (1-S4) Calves 93-94 94-95 95-96 94-95 75 75 75 75 S5 S6 S7 Sa (1-S5) (1-S6) (1-S7) (I-Sa) Male Calves 93-94 94-95 95-96 94-95 37 37 37 37 S9 S10 S11 S12 (1-S9) (1-S10) (l-S11) (1-S12) Female Calves 93-94 94-95 95-96 94-95 38 38 38 38 S13 S14 S15 S16 (1-S13) (1-S14) (1-S15) (1-S16) We will assess whether calf survival can be predicted from body weight, hind foot length, or total body length at time of trapping (continuous variable) and sex, year, and trapping zone(s) (categorical variables) using logistic regression. We will also assess whether calf and/or ~dult survival can be predicted from weather variables such as mean monthly snow depth and temperature using logistic regression (PROC LOGISTIC, PROC CATMOD, SAS 1988; Bartmann et al. 1992). T.ests will be significant at P ~ 0.05. We will use NOAA weather data from Rifle, Collbran, and Bonham Reservoir, Colorado, USDASCS snow depth coarses at Overland and Park Reservoirs on the Grand Mesa, and establish our own snow measurement coarse and weather station on the Garfield Creek State Wildlife Area deer/elk winter range. We will monitor fate of yearling bulls collared as calves immediately before and during hunting seasons to ascertain whether yearling bulls are illegally taken by hunters in areas where these bulls are protected from �96 FREDDY STUDY PLAN harvest by antler-point restrictions. Our selected study area is subject to antler-point restrictions during all hunting seasons but is adjacent to management areas where antler restrictions are not in effect during all seasons. Some y~arling bulls may disperse into areas where they will be legal as yearlings, some will undoubtedly reside along the border where combinations of restrictions apply, and some will remain within the antler-point restriction area. We must assume hunters would select against yearling bulls having white radio collars compared to yearling bulls without collars resulting in negatively biased estimates of illegal harvest. However, if several collared bulls are taken illegally, we have evidence that a problem exists and the basis for designing a study specifically to test hypotheses regarding harvests of yearling bulls. If no bulls are illegally taken we cannot conclude that illegal harvest does not exist because the collars may effectively cause hunters to avoid taking marked bulls. B. Population Size General Approach Counting all animals in a specific area is the intuitive approach to determining population size. Unfortunately, total counts over large areas are prohibitively costly in time and money and generally highly inaccurate because many animals present are not seen and counted (Caughley 1977). Efforts to improve population counts have followed 2 basic paths: designing sampling systems with statistical rigor to obtain precise estimates that often underestimate true population size (Siniff and Skoog 1964, Kufe1d et. a1. 1980), and developing procedures to correct for missed animals to improve accuracy of estimates (Bartmann et al. 1986, Samuel et al. 1987). Efforts to minimize numbers of animals missed during aerial counts often focused on variables that affect actual counting conditions such as experience of observers, height and speed of aircraft, type of aircraft, search time, weather conditions, and extent of snow background (Erickson and Siniff 1963, Caughley 1974, Gasaway et al. 1986). Although standardizing procedures may reduce variability between sampled counts, animals are still missed, often in percentages exceeding 30% (Bartmann et al. 1986). Other efforts at improving accuracy attempted to model the probability detecting animals. This probability function depends on group size, animal behavior, vegetation type and density, and age and sex class of individual (Floyd et a1. 1979, Burnham et a1. 1980, Samuel and Pollock 1981, Pollock and Kendall 1987, Samuel et a1. 1987, Unsworth et al. 1990, Samuel et al. 1992). One solution to correcting for bias therefore lies in calculating probabilities of detection and correcting observed counts. of We propose a multiple-step approach to "developing a system to estimate elk densities: 1) estimate degree of sighting or counting bias on sample quadrats and develop an equation for predicting sightability using radiocollared elk as a known subpopulation, 2) apply predictive sighting bias correction factors to adjust counts of elk on sampled quadrats used to estimate density of elk over a large area and use mark-resight estimators based on radio-collared elk to provide an additional estimate of elk density from counts of marked and unmarked elk on the same quadrats, and 3) conduct multiple counts of elk on quadrats to test the repeatability of estimates based on the predictive sighting equation and mark-resight estimators. Elk radio-collared to estimate rates of survival will be used in developing methods to estimate population size. �97 FREDDY STUDY PLAN Fundamentally, we believe a sampling system using quadrats instead of transects will be most appropriate for estimating densities of elk. Although line transect theory accounts for missing animals (negative sighting bias) ('White et al. 1989), our experience with line transects to estimate densities of elk in portions of Middle Park, Colorado indicates too few groups of elk will likely be detected to provide reasonable precision, even with intensive repeated flights of transects (Freddy 1991). Additionally, quadrats are easier to fly than transects in rough terrain inhabited by elk. Initially, we will use a 2.59 km2 quadrat. Developing a Sightability Model Given the premise of using a 2.59 km2 quadrat as a sampling unit, we will estimate the probability, or degree of bias, of detecting elk on this size of quadrat using radio-collared elk (Samuel et a1. 1987, Ackerman 1988). These authors used logistic regression to develop sighting models to predict the probability of detecting elk and mule deer (Odocoileus hemionus). Samuel et al. (1987) found group size and percent vegetation canopy coverage were the most useful variables to predict the probability of detecting elk in forested terrain of northern Idaho (n - 111 trials) while Ackerman (1988) found group size, animal activity, and vegetation type most useful for predicting detection of mule deer in brush-dominated habitats in southeastern Idaho (n 277 trials). Probability of sighting elk has been 0.35-0.95 in dense or-cpen vegetation, respectively (Samuel et al. 1987) and 0.6-0.7 in mixed conifersagebrush (Artemisia tridentata) habitats (Bear et al. 1989). The logistic regression for predicting sightability is p ~ exp(u) 1 + exp(u) where p - the sighting probability, (exp) - 2.718, and u - bo + b1xl +b2x2 + ...bnXn, the mUltiple linear regression equation of (n) variables affecting sightability. Correction factors for each group of elk detected are obtained by inverting the estimated sighting probability (Samuel et al. 1987). The primary advantage of this sighting bias model approach is .that marked animals are not needed in the elk population once the model has been developed unlike traditional mark-resight models that require marking animals before every survey. ,--- The primary assumptions of this approach are: 1) Behavior of elk is not altered by the radio collar, 2) Visual detection of elk is not enhanced by the presence of the radio collar as concluded by Bartmann et al. 1987, Samuel et al. 1987, Ackerman 1988, and Bear et al. 1989, 3) Selected radio-collared elk can be located to within a quadrat, 4) Collared elk can be individually identified either by radio frequency or collar symbol/numbers when detected, 5) Observers attempting to visually detect collared elk do not know the locations of marked elk, 6) The dependent variable, sighting probability, is asymptotic between values of 0 and 1, �98 FREDDY STUDY PLAN 7) The independent variables, whether numeric or categorical, are measured accurately, 8) Selected elk represent heterogenous habitats typically frequented elk. by Major hypotheses to be tested are: 1) Logistic regression improves precision of sighting probability compared to a simple binomial sighting probability model, and 2) Predictive sighting models are repeatable among periods of time (years) . The entire winter range within the study area will be divided into 2.59 km2 units having boundaries based on topographic features instead of cadastral surveys. These irregularly shaped units will be approximately equal in size. Sightability tests will be conducted during 2 5-day periods in Januaryearly February 1994 and during 3 5-day periods in December, January, and February 1994-95. Our objective is to complete 80 test trials (16/day) on marked animals (seen or missed) during each 5-day period. We anticipate searching 16 quadrats/day at a rate of 15 min/quadrat resulting in 4-5 hours of helicopter use per day. During the 5 days, we will systematically cover the study area. During the first year, up to 150 different collared elk will be available which should reduce our need to repeat trials on individual elk during any 5-day period. We will conduct tests during the morning and afternoon and avoid mid-day to approximate traditional time periods when elk counts are conducted. Because sighting bias is largely associated with small groups of elk, we will attempt to fly when elk are widely distributed in small groups (Appendix VI). A primary observer in a Cessna 185 will locate selected radio-collared elk (Gilmer et al. 1981) and assign each elk to a quadrat or quadrats if animals are near a quadrat boundary. Within 2 hours of locating an elk, 2 observers along with the pilot in a Bell-Soloy helicopter will search assigned quadrates) and enumerate all groups of elk and identify all radio-collared elk on the quadrat. Radio-collared elk will be identified either by the symbol/number on the collar and/or radio frequency using a receiver-scanner in the helicopter (Appendix I). When observers have completed their search of each quadrat they.will check via radio wit:h the prim~ry.observer to determine if they found the assigned elk. If the elk was found they will proceed to the next quadrat, but if the elk was missed the helicopter crew will immediately find the marked elk using telemetry and document its location as on or off the assigned quadrat. If more than 1 assigned elk occurs in a group only one animal will be used because observations may not be independent. If an assigned elk is missed and then found off the quadrat; the trial will be voided. To maintain vigilant observers, the helicopter crew will be assigned quadrats where there are no known radio-collared elk to serve as controls. For each group detected on a quadrat, observers will assign classification variables to describe conditions potentially affecting detection of each group whether or not the group contained a marked elk. These independent variables will be group size, activity of animal first detected in the group, vegetation type, percent vegetation canopy density, _,_-----and �99 FREDDY STUDY PLAN percent snow cover. These variables must also be noted for those groups containing assigned marked elk that are missed during the initial search of the quadrat. Additional independent variables assigned to all groups observed during a flight will be year, flight, pilot, observer, search time per quadrat, and age/sex of marked animals detected (Appendix V). During the first year we will attempt to use only one experienced observer, navigator, and pilot in the helicopter with the intent of controlling these variables. The helicopter crew will be trained prior collecting data. to We anticipate group size, activity, vegetation type, and percent vegetation canopy cover may significantly influence sighting probability (Samuel et al. 1987 and Ackerman 1988). We also anticipate that percent vegetation canopy cover may be difficult to judge. Samuel et al. (1987) worked primarily in habitats having tall conifers. They considered vertical canopy cover to be any vegetation obstructing the observer's view of an elk group for up to 30 m around the group (Unsworth et al. 1991). Ackerman (1988) correctly noted that animals are seldom viewed from vertical but rather from oblique angles. Confounding this perception of visual obstruction are ..deciduous habitats, such as oakbrush (Quercus gambelii) and aspen (Populus tremuloides), that can be seen through during winter and only partially obscure one's view of elk even when canopy coverage may approach 100%. A similar problem involves sagebrush which can effectively hide a bedded animal with little distinguishable vertical canopy. We will follow suggestions by Unsworth et al. (1991) but anticipate this to be a difficult variable to estimate. Using stepwise logistic regression (SAS 1988), independent variables will be evaluated as to their significance in predicting the probability of sighting elk (groups seen or missed) to develop a predictive sightability model. Model fit will be assessed using AIC and log-likelihood values. We anticipate transforming some independent numeric variables .to improve conformity of data to model assumptions. We will compare simple binomial and complex sightability models using computer simulations to assess whether independent variables meaningfully improve precision of the sighting probability estimator and also compare the mean and variance of estimates of elk density derived from corrected and uncorrected counts of elk on quadrats. Sightability models among flight periods will be compared to assess the repeatability of the sighting model when pilot and observers remain constant. If we achieve a sample of 80 trials/flight period, power (I-beta) at alpha - 0.05 to detect a difference between sighting models will be about 0.57 (SE=0.016) (sighting models from Samuel et al. 1987 for sighting functions at 20 and 40% vegetation cover). Power improves to about 0.87 (SE=0.006) when n > 160 trials. These power calculations assume surveys will encounter groups of elk at frequencies of group sizes similar to those observed during preliminary sampling flights (Appendix VI, Table 2). If observed sizes of groups shift towards large groups then power of tests would be less because differences in sighting probabilities between models is most pronounced for small groups (Appendix VI) . �100 FREDDY STUDY PLAN Comparing Population Estimators We will estimate total population size within the study area during winter using 3 approaches based on a random quadrat sampling system. This effort will begin in 1994-95 with one flight of quadrats to evaluate logistics, adequacy of sampling, and initial estimates of population size and then continue in 1995-96 and 1996-97 with 3 repeated flights each year if quality of results justifies the effort. We will randomly select about 15% of the potential quadrats and search these quadrats each flight following procedures suggested by Kufeld et al.(1980). Sampling may be stratified according to expected animal density based on distribution of radio-collared elk in 1993-94. For quadrats, we will use 3 estimators to calculate density of elk and all 3 estimators will use the same counts of elk generated from each flight: 1) uncorrected counts of elk on quadrats, 2) counts of elk on quadrats corrected for sighting bias based on the appropriate sighting bias model, and 3) counts of elk on quadrats corrected for sighting bias based on Lincoln-Petersen mark-resight estimator combined for repeated flights using the joint hypergeometric maximum likelihood estimator (JHE) (White and Garrott 1990). The Lincoln-Petersen estimator for each individual flight is: Ni - _illl+ 1) (ni + 1) (m, + 1) -1 with an estimated variance of " " Var(Ni) - _illl+ 1) (ni + 1) (nl_=_.!!hli!h_:_!!hl (mi + 1)2 (mi + 2) " " and a nominal 95% confidence interval of Ni ± 1.96 VVar(Ni) where nl is number. of radio-collared e~k, ni is number of elk observed on each aerial survey, and mi is number of marked elk observed on·each survey. We will compare means and variances of estimates derived from each method. . Assumptions for each estimator are: Uncorrected Counts: a) All elk on the quadrats are counted and b) only elk on quadrats are counted (no boundary errors). Assumption a) is likely violated and is the reason for developing a sighting bias model to correct counts. Corrected Counts: c) assumptions are listed in the previous section. We will apply sighting bias corrections to each observed group of elk according to the sighting bias model. Lincoln-Petersen Mark-Resight: d) probability of sighting marked and unmarked elk is the same, e) marked and unmarked elk are correctly classified (marks are discernible), f) marked elk are randomly selected and distributed throughout the population or at least resighting effort is randomly distributed throughout the population, g) each animal has an �101 FREDDY STUDY PLAN identical but independent probability of being resighted, h) number of marked elk (radio collars) in the population is known at the time of the survey which implies that marks are not lost or unaccounted for, and i) the population is geographically and demographically closed; i.e., there is no immigration or emigration, recruitment or death during the survey in a geographically defined area (Otis et al. 1978, Bear et al. 1989). Assumptions for Lincoln-Petersen mark-resight estimates may be the most difficult to achieve and evaluate. To achieve adequate numbers of marked animals in the population, elk marked each year must accumulate and be used during the years of population surveys. We implicitly assume that detection of white radio collars from yeara is equal to detection of white collars from yearb when surveys are flown in yearc, otherwise there will be different sighting probabilities for each cohort of marked elk. We also must assume there will be some (1-5%) nonoperative radio collars on live elk which would create an error in knowing how many elk are marked within the area surveyed (fixed-winged flights will be conducted prior to population surveys to determine number of functioning radios in the sampled area using and observer independent of counters). This type of error would cause positive bias in population estimates. Bear et al. (1989) identified marker visibility, random sampling of the population, and heterogeneous sighting probabilities due to group size or vegetation type as problems associated with applying mark-resight estimators to elk populations. We have attempted to address these concerns by using high-contrast white collars not prone to color bias, helicopters to capture elk throughout the distribution of the population, use of random sampling to survey the population, and use of multiple independent aerial resight (recapture) surveys to reduce recapture bias associated with other recapture techniques and to improve precision of estimates (Otis et al. 1978, Bartmann et al. 1987, Bear et al. 1989, Neal et al. 1993). Effects of group size and vegetation type are sources of sighting heterogeneity which can bias the variance of the estimate and may negatively bias the estimate (Bear et al. 1989, Neal et al. 1993). Comparing variances of estimates derived from markresight and the sighting probability model may provide some insight into the severity of heterogenous sighting probabilities. Population estimates based on the JHE mark-resight approach will be calculated using the program NO REMARK (G. White, pers. comm.). We also used NOREMARK to simulate expected quality of population estimates for the JHE estimator given the following constraints. If the real population consists of 2,000 elk, 200 elk are marked with radio collars, about 10% of all elk are observed on each flight ([sighting probability - 0.60] x [proportion of area sampled - 0.15]), and 3 replicate flights are completed, we should achieve a 90% confidence interval about the true population of ± 20% with a 1% positively biased estimate of population size. Confidence interval coverage would be 89% (n = 500 Monte Carlo simulations). During the second year of the project when mark-resight surveys will be initiated on quadrats, there could be about 200 marked elk in the population. Precision of JHE estimates would improve to ± 15% if 300 marked elk were available or 5 replicate flights were completed. These would be potential design options if marked elk continued to accumulate in the third or fourth year of the project or helicopter hours were �102 FREDDY STUDY PLAN reallocated among uses or increased. The JHE estimator does not require that marked elk are individually identified, only that marked and unmarked elk are discerned. If we can consistently identify individual elk by collar numbers or radio frequency, the Minta-Mangel mark-resight model can be used to estimate population size and associated variance (Minta and Mangel 1989) using program NOREMARK. Variance of this estimator can be calculated as described by Minta-Mangel (1989) or in a new approach developed by Bowden (1993, unpubl. ms.). Complete capture histories on each marked elk would also allow an assessment of bias associated with the estimate of N due to unequal individual capture probabilities. Unequal capture probabilities can be caused by individual animal behavior, time intervals between initial capture and resight surveys, and heterogeneity of capture probabilities due to factors of age, sex, social dominance, or habitat. Program CAPTURE will be usedAto evaluate the most appropriate capture probability model to estimate N (Rexstad and Burnham 1991). C. Distribution and Movements of Elk General Approach All visual sightings of live and dead collared elk will be assigned UTM coordinates as determined from USGS topographic maps. Visual sightings of collared elk during winter will occur during ground surveys to determine life/death status of elk and during aerial surveys to estimate population density. These locations will provide a general perspective of the distribution of elk on their winter range. Seasonal patterns of movement will be obtained primarily from aerial surveys. Accurate seasonal locations on all collared elk would be cost prohibitive and not commensurate with the main objectives of the project. Therefore we plan to randomly select 20% of the calves (8 M, 8 F) and 20% of the adult females (15) and monitor their locations periodically to reveal patterns of seasonal movement. Yearling males marked as calves will be located prior to hunting seasons to monitor their fates in relation to ant_ler-point restrictions. Distribution and movement data will be plotted following suggestions of White and Garrott (1990). V. SCHEDULE Activity Approximate Date Phase I--Complete detailed study plan, July 1992-June 1993 select study location, refurbish elk traps, purchase radio collars for calves and design expandable collars for calves. Phase II--Complete initial purchase of July 1993-June 1994 radio collars; trap and radio-collar Trapping 12/93 elk to estimate survival; begin surveys Sighting bias to design sighting bias estimator. 1 & 2/94 '---' �103 FREDDY STUDY PLAN Phase III--Continue radio-collaring July 1994-June 1995 Trapping 12/94 elk for estimating survival, Sighting bias 12/94, continue testing sighting bias estimator, estimate population 1/95, 2/95; Pop. Est. 1 & 2/95 size using quadrats and mark-resight. Phase IV-- Estimate pop. size July 1995-June 1997 size using sighting bias correction, Trapping 12/95 & quadrats, and mark-resight. Continue 12/96; Pop. Est. trapping to estimate survival. 1/96 &1/97 VI. PERSONNEL PROGRAM RESPONSIBILITIES David J. Freddy: Principal Investigator responsible for final project design, organ~z~ng field personnel, obtaining and organizing data, financial control, coordinating publications. R. Bruce Gill: Provide administrative support, input for study design, and liaison with other administrative sections within the Division of Wildlife or other natural resource agencies. John H. Ellenberger and James H. Olterman: Provide input for study design and location, coordinate study with Regional activities, and logistical support. David C. Bowden and Gary C. White: Provide input for study design and statistical protocol, conduct data analyses, and provide software support. �104 FREDDY STUDY PLAN VII. BUDGET-SUMMARY ACTIVITY FY92-93 Personal Operating Travel Capital Total 55,773 26,800 1,200 4,700 88,473 'WI 4% inflation FY93-94 FY94-95 FY95-96 FY96-97 72,236 72,270 4,680 5,700 72,236 93,045 5,590 0 70,706 66,020 4,940 0 70,706 66,370 4,940 0 154,886 170,871 141. 666 142,016 161,081 177,706 147,333 147,697 BUDGET-ITEMIZED PERSONAL SERVICES A. D.J. FREDDY 12 MOS B. MAMMALS SECR. 1 MOS C. UTILITY I 6 MOS OCT-MAR D. UTILITY I, TRAPPING E. WORK STUDY STUDENT, 1 YEAR F. VOLUNTEER FOR 20 DAYS SUB-TOTAL OPERATING G. VEHICLES MILEAGE, 4X4, 3 H. VEHICLES, SNOWMOBILES, ATV I. NEW RADIO COLLARS J. REFURBISHED RADIO COLLARS J. TRAPPING, BAIT, SUPPLIES K. MISC. SUPPLIES, SERVICES L. FIXEDWINGED, MORT. SURV. M. FIXEDWINGED, SIGHT. SURV. N. HELICOPTER, SIGHT. SURVEYS o. HELICOPTER, QUAD. SURVEYS P. HELICOPTER, NONRANDOM SURV. Q. HELICOPTER TRAPPING SUB-TOTAL TRAVEL R. DJF 10 DAYS, STANDARD S. DJF DAYS TRAPPING T. VOLUNTEER, TRAPPING U. REGION PERSONS, TRAPPING V. REGION PERSONS, SURVEYS SUB-TOTAL CAPITAL EXPENSES W. RADIO TELEMETRY RECEIVER X. PACK-SETS/LAPTOP COMPUTER SUB-TOTAL PROJECT YEARLY TOTAL WITH 4% INFLATION FY92-93 FY93-94 FY94-95 FY95-96 FY96-97 $49,841 $2,509 $0 $2,723 $700 $0 $55,773 $56,796 $2,500 $9,180 $3,060 $700 $0 $72,236 $56,796 $2,500 $9,180 $3,060 $700 $0 $72,236 $56,796 $2,500 $9,180 $1,530 $700 $0 $70,706 $56,796 $2,500 $9,180 $1,530 $700 $0 $70,706 $2,200 $0 $19,500 $0 $3,000 $1,500 $600 $0 $0 $0 $0 $0 $26,800 $5,720 $350 $8,600 $0 $2,500 $2,500 $3,150 $2,700 $21,750 $0 $0 $25,000 $72,270 $5,720 $350 $15,000 $4,050 $2,500 $2,500 $3,150 $4,050 $32,625 $4,350 $0 $18,750 $93,045 $5,720 $350 $6,000 $5,400 $2,500 $2,500 $3,150 $1,800 $0 $21,750 $4,350 $12,500 $66,020 $5,720 $350 $5,000 $6,750 $2,500 $2,500 $3,150 $1,800 $0 $21,750 $4,350 $12,500 $66,370 $650 $550 $0 $0 $0 $1,200 $650 $1,365 $0 $1,365 $1,300 $4,680 $650 $1,365 $0 $1,365 $2,210 $5,590 $650 $1,365 $0 $1,365 $1,560 $4,940 $650 $1,365 $0 $1,365 $1,560 $4,940 $3,200 $1,500 $4,700 $3,700 $2,000 $5,700 $0 $0 $0 $0 $0 $0 $0 $0 $0 $88,473 $154,886 $161,081 $170,871 $177,706 $141,666 $147,333 $142,016 $147,697 ******************************************************************************************* REGION REGION REGION REGION REGION REGION TRAPPING MDAYS PILOT, HRS BIOL.,FLYING MDAYS PICKUP(S)+TRAP TRAILER(S) SKIDOO ALPINE AND ATV PROP. TECH. MDAYS 0 65 20 TRAP A/N 3 21 65 20 TRAP A/N 3 21 80 34 TRAP A/N 3 21 55 24 TRAP A/N 3 . 21 55 24 TRAP A/N 3 . -_ _. �.lU.) FREDDY STUDY PLAN VIII. LOCATION We selected the eastern portion of GMU 42 within the Grand Mesa DAU (E14) for conducting this project (Appendix VII). This area encompasses about 1,028 km2 (397 mi2) in the Divide Creek drainages located south of the towns of NewCastle and Rifle in west-central Colorado. We estimate that 1500-3000 elk inhabit about 642 km2 (248 mi2) of winter range within this area. This area meets the following necessary criteria: 1) the elk population is relatively "closed" during the winter, that is, subject to minimal emigration or immigration, and assumes a somewhat static distribution during winter; 2) the winter range generally has reliable snow cover to facilitate detecting elk during population surveys and mixed vegetation types consisting of oakbrush, aspen, sagebrush, juniper-pinyon woodland (Juniperus osteosperma-Pinus edulis), mixed conifers (Picea sp., Abies sp.) and agricultural fields typical of many elk winter ranges in western Colorado; 3) there is dependable access for research purposes to both private and public lands and dependable public access to public lands for hunting; 4) there is minimal large-scale land-use conflict problems such as agricultural damage; 5) there is local, area, and regional CDOW support for conducting the project in this area; and 6) the airport at Rifle will provide readily accessible support services. IX. LITERATURE CITED Ackerman, B. B. 1988. Visability bias of mule deer aerial census procedures in southeast Idaho. Phd. Thesis. University of Idaho, Moscow. 106pp. Bartholow, J. 1992. Pop-II system documentation. Fossil Creek Software, Fort Collins, CO. 50pp. Bartmann, R. M. 1984. Estimating mule deer winter mortality in Colorado. J. Wildl. Manage. 48:262-267. _____ , Carpenter, L. H., R. A. Garrott, and D. C. Bowden. 1986. Accuracy of helicopter counts of mule deer in pinyon-juniper woodland. Wildl. Soc. Bull. 14:356-363. _____ , and G. C. White. 1991. Compensatory effects of harvest in a mule deer population. Colo. Div. Wildl. Game 'Res. Rep. July:28-40. _____ , , L. H. Carpenter. 1992. Compensatory mortality in a Colorado mule deer population. Wildl. Monogr. 121. 39pp. _____ , and R. A. Garrott. 1987. Aerial mark-recapture -----, estimates of confined mule deer in pinyon-juniper woodland. J. Wildl. Manage. 51:41-46. Bear, G. D. 1986. Expanding telemetry collar for elk calves. Colo. Div. Wildl. Game Info. Leaflet 112. 2pp. 1989. Seasonal distribution and population characteristics of elk in Estes Valley, Colorado. Colo. Div. Wildl. Spec. Rep. 65. l4pp. _____ , G. C. White, L. H. Carpenter, R. B. Gill, and D. J. Essex. 1989. Evaluation of aerial mark-resighting estimates of elk populations. J. Wildl. Manage. 53:908-915. Bowden, D. C. A simple technique for estimating population size. Unpubl. ms. l7pp. Boyce, M. s. 1989. The Jackson elk herd, intensive wildlife management in North America. Cambridge Univ. Press, Cambridge. 306pp. �106 FREDDY STUDY PLAN Burger L. W., Jr., M. R. Ryan, D. P. Jones, and A. P Wywialowski. 1991. Radio transmitters bias estimation of movements and survival. J. Wildl. Manage. 55:693-697. Burnham, K. P., D. R. Anderson, and J. L. Laake. 1980. Estimation of density from line transect sampling of biological populations. Wildl. Mono. 72. 202pp. Caughley, G. 1974. Bias in aerial survey. J. Wildl. Manage. 38:921-933. 1977. Analysis of vertebrate populations. John Wiley & Sons, New York. 232pp. Colorado Division of Wildlife. 1990. Colorado big game harvest. Colo. Div. Wildl., Denver. l73pp. 1991. Colorado big game harvest. Colo. Div. Wild1., Denver. l72pp. 1991b. Long Range Plan (revised). Colo. Div. Wildl., Denver. l6pp. de Vergie, W. J. 1989. Elk movements, dispersal, and winter range carrying capacity in the upper Eagle River valley, Colorado. MSc. Thesis, Colorado St. Univ., Fort Collins. l24pp. Erickson, A. W., and D. B. Siniff. 1963. A statistical evaluation of factors influencing aerial survey results on brown bears. Trans. N. Am. Wildl. Conf. 28:391-409. Foster, C. C., E. D. Forsman, E. C. Meslow, G. S. Miller, J. A. Reid, F. F. Wanger, A. B. Carey, and J. B. Lint. 1992. Survival and reproduction of radio-marked adult spotted owls. J. Wildl. Manage. 56:91-95. Floyd, T. J., L. D. Mech, and M. E. Nelson. 1979. An improved method of censusing deer in deciduous-coniferous forests. J. Wildl. Manage. 43:258-261. Freddy, D. J. 1987. The White River elk herd: a perspective, 1960-85. Colo. Div. Wildl. Tech. Publ. 37. 64pp. 1991. Elk census methodology. Colo. Div. Wildl. Game Res. Rep. July: 59-72. _____ , D. L. Baker, R. M. Bartmann, and R. C. Kufeld 1993. Deer and elk management analysis guide, 1992-1994. Colo. Div. Wildl., Div. Rep. 17. 77pp. (In press). Garrott, R. A., and G. C. White. 1982. Age and sex selectivity in trapping mule deer. J. Wildl. Manage. 46:1083-1086. Garrott, R. A., R. M. Bartmann, and G. C. White. 1985. Comparison of radiotransmitter packages relative to deer fgwn mortality. J. Wildl. Manage. 49:758-759. Gasaway, W. C., S. D. DuBois, D. J. Reed, and S. J. Harbo. 1986. Estimating moose population parameters from aerial surveys. Institute of Arctic BioI., BioI. Papers Univ. of Alaska No. 22. 108pp. . Gilmer, D. S., L. M. Cowardin, R. L. Duval, L. M. Mechlin, C. W. Shaiffer, and V. B. Kuechle. 1981. Procedures for the use of aircraft in wildlife biotelemetry studies. USDI Fish & Wildl. Servo Resource Publ. 140. 19pp. Heisey, D. M., and T. K. Fuller. 1985. Evaluation of survival and causespecific mortality rates using telemetry data. J. Wildl. Manage. 49:668-674. Houston, D. B. 1982. The northern Yellowstone elk, ecology and management. Macmillan Publ. Co., Inc. New York. 474pp. Kufeld, R. C., J. H. Olterman, and D. C. Bowden. 1980. A helicopter quadrat census for mule deer on Uncompahgre Plateau, Colorado. J. Wildl. Manage. 44:632-639. 1-. . �.lUI FREDDY STUDY PLAN -,-- Laake, J. L. 1992. Catch-per-unit-effort models: an application to an elk population in Colorado. Pages 44-55 in D. R. McCullough and R. H. Barrett, eds., Wildlife 2001: Populations. Elsevier Publ., London. Leptich, D. J., and P. Zager. 1991. Road access management effects on elk mortality and population dynamics. Pages 126-131 in A. G. Christensen, L. J. Lyon, and T. N. Lonner, comps., Proc. Elk Vulnerability Symp. Montana St. Univ., Bozeman. Minta, S. and M. Mangel. 1989. A simple population estimate based on simulation for capture-recapture and capture resight data. Ecology 70:1738-1751. Neal, A. K., G. C. White, R. B. Gill, and D. F. Reed. 1993. Evaluation of mark-resight methods to estimate mountain sheep numbers. J. Wildl. Manage. (In press). Nelson, L. J., and J. M. Peek. 1982. Effect of survival and fecundity on rate of increase of elk. J. Wi1d1. Manage. 46:535-540. Pemberton, J. M., S. D. Albon, F. E. Guinness, T. H. Clutton-Brock, and R. J. Berry. 1988. Genetic variation and juvenile survival in red deer. Evolution 42:921-934. Pollock, K. H., and W. L. Kendall. 1987. Visibility bias in aerial surveys: A review of estimation procedures. J. Wildl. Manage. 51:502-510. _____ , S. R. Winterstein, C. M. Bunck, and P. D. Curtis. 1989. Survival analysis in telemetry studies: the staggered entry design. J. Wildl. Manage. 53:7-15. Quimby, D. C., and J. E. Gaab. 1957. Mandibular dentition as an age indicator in Rocky Mountain elk. J. Wildl. Manage. 21:134-153. Rexstad, E., and K. Burnham. 1991. User's guide for interactive program CAPTURE. Coop. Fish & Wildl. Res. Unit, Colorado St. Univ., Fort Collins. 29pp. Samuel, M. D. and K. H. Pollock. 1981. Correction of visibility bias in aerial surveys where animals occur in groups. J. Wildl. Manage. 45:993997. _____ , E. O. Garton, M. W. Schlegel, and R. G. Carson. 1987. Visibility bias during aerial surveys of elk in northcentral Idaho. J. Wildl. Manage. 51:622-630. _____ , R. K. Steinhorst, E. O. Garton, and J. W. Unsworth. 1992. Estimation of wildlife population ratios incorporating survey design and visibility bias. J. Wildl. Manage. 56:718-725. SAS Institute, Inc. 1988. SAS Language guide for personal computers. SAS Institute Inc., Cary, NC. 1028pp. Sauer, J. R., and M. S. Boyce. 1983. Density dependence and survival of elk in northwestern Wyoming. J. Wildl. Manage. 47:31-37. Schlegel, M. W. 1977. Factors affecting calf elk survival on Coolwater Ridge in north central Idaho. Pages 35-37 in Proc. Western States Elk Workshop, Estes Park, Colo. Siniff, D. B., and R. O. Skoog. 1964. Aerial censusing of caribou using stratified random sampling. J. Wildl. Manage. 28:391-401. Unsworth, J. W., L. Kuck, and E. O. Garton. 1990. Elk sightability model validation at the National Bison Range, Montana. Wildl. Soc. Bull. 18: 113 -115. _____ , F. A. Leban, G. A. Sargeant, E. O. Garton, M. A. Hurley, J. R. Pope, and D. J. Leptich. 1991. Aerial survey: user's manual with practical tips for designing and conducting aerial big game surveys. Idaho Dep. �108 FREDDY STUDY PLAN Wade, Game & Fish, Boise. 54pp. D. A., and J. E. Browns. 1982. Procedures for evaluating predation on livestock and wildlife. Texas Agric. Expt. Sta. Publ. B-1429. 42pp. White, G. C. 1983. Numerical estimation of survival rates from band recovery and biotelemetry data. J. Wildl. Manage. 47:716-728. 1992. DEAMAN database manager and population modeling procedures; Colorado Division of Wildlife User's manual and reference. Colorado St. Univ., Fort Collins. 109pp. _____ , D. R. Anderson, K. P. Burnham, and D. L. Otis. 1982. Capturerecapture and removal methods for sampling closed populations. Los Alamos Natl. Lab. LA-8787-NERP. 235pp. _____ , R. M. Bartmann, L. H. Carpenter, and R. A. Garrott. 1989. Evaluation of aerial line transects for estimting mule deer densities. J. Wildl. Manage. 53:625-635. _____ , and R. A. Garrott. 1990. Analysis of wildlife radio-tracking data. Academic Press, Inc., San Diego. 383pp. _____ , , R. M. Bartmann, L. H. Carpenter, and A. W. Alldrege. 1987. Survival of mule deer in northwest Colorado. J. Wildl. Manage. 51:852859. Zager, P., and D. J. Leptich. 1991. Comparing two methods of calculating elk survival rates. Pages 106-109 in A. G. Christensen, L. J. Lyon, and T. N. Lonner, comps., Proc. Elk Vulnerability Symp. Montana St. Univ., Bozeman. APPENDIX I Specifications for Radio Collars Pulse Rate Normal: 60-65 ppm Pulse Rate Mortality: 120-130 ppm Motion Sensor Delay: 4-6 hrs Batteries: 4+ year life, 2 lithium D-cells adults, 1 lithium D-cell calves Antenna: External whip, pvc coated Collar Material: 7.6 cm (3") wide white colored smooth surfaced conveyor belting, 2 layers sewn together, 0.64 cm (1/4") total thickness: Additional Material: Bright white with black core Ritchie All-Flex rubber material for identifcation symbol/number placed as a sleeve over top' portion of collar Collar Size: 61-81 cm (24-32") adult females, individually fitted 61-81 cm (24-32") expandable for female calves 61-97 cm (24-38") expandable for male calves Collar Weight: About 1.1 kg �109 FREDDY STUDY PLAN APPENDIX Animal Welfare There are 3 potential issues regarding radio collars, and capture techniques. II Concerns animal welfare: use of ear-tags, use of Ear-tags: To permanently mark radio-collared animals and some non-collared animals, an aluminum, numbered and colored ear-tag will be placed in each ear. Tags are self-piercing using special pliers and have been used for many years on elk. To reduce potential problems of tag loss, tags will be placed on front edge of the ear about 1/3 to 1/2-way above the base of the ear and the edge of the tag will be flush with the edge of the ear. Radio collars: Fitted size of radio collars is based on measurements from hunter-killed elk and captive research elk and testing of actual collars on captive elk. The basic expandable design for calves has been extensively field tested on deer and moose. Capture Techniques: We will capture elk in traditional corral traps capable of holding several elk at one time and by net-gunning individual elk from a helicopter. We expect a 1-2% serious injury/mortality rate with either technique. Animals having a broken leg, neck, or pelvis will be euthanized with a gunshot to the head following euthanasia procedures in our Animal Welfare guide. All persons involved in capture will be trained to properly euthanize appropriate animals .. Capture techniques will be constantly monitored and changed if necessary to insure that minor injuries to animals do not chronically occur. Animals captured with a helicopter will be hobbled and blindfolded and slung under the helicopter to a nearby field processing point where they will be collared, weighed, etc. and then released at that site. We aniticapte 1-5 minutes to capture the elk, 1-2 minutes to ferry the elk to .processing point, and 3-4 minutes to process the elk before release. This technique has been successfully used to capture and radio-collar elk in Arizona, Colorado, Idaho, Montana, Oregon, and Washington. &ecause of sho~t handling times, mortalities due to capture have been rare. Animals captured in corral traps will be blindfolded as they are processed in the chutes. Calves may be hobbled or walked into a holding box for weighing. We prefer to trigger trap~ manually using a solenoid system so we can be selective about animals captured and reduce the time animals are in traps. However, traps may be triggered by animals entering and in such cases animals may be in the trap overnight. Food will be provided via alfalfa used as bait. We will use the minimum number of persons to process elk to reduce capture stress. �llO FREDDY STUDY PLAN APPENDIX III Visual Identification System for Radio Collars Numbers, symbols, and letters will be used in ordered combinations to identify individual elk primarily during mark-resight aerial surveys. No more than 3 characters will be used to identify an individual and there will always be a 2-digit number. Numbers Used (8): 0, I, 2, 3, 4, 5, 6, and 7 Symbols Used (5): .,.,., ~, + Letters Used (11): A, C, F, H, K, N, P, T, V, X, Y 10, ll, 12, 13, 14, 15, 16, 17, 20, 21, 22, 23, 24, 25, 26, 27, Number Matrix 30, 40, 50, 60, 31, 41, 51, 61, 32, 42, 52, 62, 33, 43, 53, 63, 34, 44, 54, 64, 35, 45, 55, 65, 36, 46, 56, 66, 37, 47, 57, 67, 70 71 72 73 74 75 76 77 Each Symbol with each number represents 56 identification codes; all symbols with each number equals 280 codes or animals. Each letter with each number represents 56 codes; all letters with each number equals 616 codes or animals. Therefore, 896 different animals can be individually marked using this system. �FREDDY STUDY PLAN APPENDIX IV Precision and Power of Survival Estimates and Tests 30 ~ S - 0.50 S-D.w ~ -_~~ = : .. S -0.70 S - 0.80 • ••• S -0.90 I ••• .- 5 50 100 75 RADIO COUARS DEPLOYED 2 ci • DIFF-O.25 ?!~~~~'.~.. ~~=':~ .~~~:'.~~ 0.8 __________ .i->': I ! -----+--- DETECT DIFFERENCES IN SURVIVAL OF: .•... f . -- 0.8 ~ ® m 0.04 .... ~ .•.... ..' 15 i 7 ..••..•.•.....•....••. ...... ..... .................................. + __ . 0.2 a.. a I 20 040 1.2 DIFF-O.2 DIFF-O.1 -------------------------- ...__ .- •...:-_ .. I 0.8 t; ~ 15 ! I DEPLOYED DIFF-O.OS 0.8 @ 80 1 ci • 75 DETECT DIFFERENCES IN SURVIVAL OF: DIFF-O.25 i 'r 100 I 60 RADIOCOUARS 0.04 0.2 . ~ ~ . .' ............................. .. .>" _ ' ........................................................................•................ :.::.= - - - - - - - - - - - - - - - - - - - - - - - - - - - -1----------•••••••••••••• i a 20 _ . ' 60 RADIO COLlARS 75 DEPLOYED _ I T 80 100 �112 FREDDY STUDY PLAN APPENDIX V Classification variables for elk groups and flight conditions for sightability trials. Group Size 1 2 3 Activity Level Vegatation Type Oakbrush-dense/continuous Oakbrush-scattered/c1umped Aspen Sagebrush Pinyon-Juniper Tall Conifer Agri. Fields/Clearings Riparian Bedded Standing Moving 4 5 6 7 8 9 10-14 15-19 20-24 ~25 % Vegetation Canopy Vertical Cover %Snow Cover o 0-15 16-25 26-35 36-45 46-55 56-65 66-75 76-85 86-100 Snow Type Fresh New Snow Old snow with old tracks 1-25 26-50 51-75 76-99 100 Lighting Conditions Age/Sex Marked Elk Bright Sun/High Contrast Male Calf Hazy Sun Female Calf Overcast, dull light Adult Female Yearling Male Adult Male Year/Period Pilot Observer 1993-94-1 1993-94-II 1994-95-1 1994-95-II 1994-95-III Search Rate 1993 1993 1994 1994 km2 / min /_ ~ . �113 FREDDY STUDY PLAN APPENDIX VI Power Considerations for Sightability Dr. David C. Bowden Models Consider the logistic model (1) below which describes the probability of observing elk groups for a given set of observational conditions. Assume that the same model also describes the probability of observing elk groups for a different set of conditions but with a different value for the parameter Po. Suppose data is obtained under both sets of conditions. This section describes the results of a simulation study which investigates the power of a test of the hypothesis that the parameter value for Po is the same for both sets of conditions. In the simulation study, observations were generated as if two sets of conditions applied. Then a test of equality of the 2 Po values was performed. This process was repeated 1000 times. The proportion of times that the null hypothesis was rejected estimates the power of the given test. Following Samuel et a1. (1987), it is assumed that the probability of sighting elk groups in uniform habitat relative to percent vegetation cover can be adequately modeled as p = 1/(1 + exp(-«(3o +Pi1n(g» ) (1) where p is the probability of sighting a group of size g in the specified habitat, Po and Pi are parameters and In(g) is the natural logarithm of g. Those authors estimated the parameter values to be .22 and 1.55, respectively for habitat locations with 20 % vegetation cover and -0.78 and 1.55 with 40% vegetation cover. These sets of estimates were used to specify the parameter values for the simulation study. Table 1 gives values of p for the 2 sets of parameter values and a few choices of group sizes. Thus the problem of interest is, "What is the chance that a change in sighting probabilities as given in Table 1 will be detected if the conditions of the simulation study apply?" �114 FREDDY STUDY PLAN Table 1. Probability of sighting given 2 sets of parameter values at different group sizes (g). Sample 1 uses ~o = .22 and Pi = 1.55 in function (1) and Sample 2 uses ~o = -.78 and ~i = 1.55. Group size(g) PROBABILITY SAMPLE 1 SAMPLE 2 1 .555 .314 2 .785 .573 3 .872 .716 4 .914 .797 5 .938 .847 6 .952 .881 10 .978 .942 15 .988 .968 20 .992 .979 30 .996 .989 Generation of observations for a single iteration in the simulation study is described next. A set of n elk groups were used to generate 2 samples of elk group observational data. Sample 1 was generated by taking an observation from a Bernoulli random variable for each group where the probability of success is given by probability function (1) with ~o = .22 and ~i = 1.55 , that is, a random uniform number on the interval [0,1), say u, was selected, if u~p then the group was said to be sighted otherwise it was not sighted. The second sample of n observations was independently generated with the same group sizes using parameter values ~o = -.78 and ~i = 1. 55 and function (1) to obtain the needed probability of success. ·The 2 samples of sighting data were pooled and then analyzed using the SAS logistic regression procedure. The following model was fi·t to the pooled data set: p = 1 / ( 1 + exp (- ( Po + Ii 1 In (g) + ~ 2 Z ) ) ) (2 ) where z is an indicator variable which is equal to 0 if the observation came from sample 1 and is equal to 1 if the observation came from sample 2. The question of interest is, "What is the power of a 4-.05 test of the hypothesis that li2 = 0 versus an alternative hypothesis li2 ". 0 when actually ~2 = -1?" This null hypothesis is equivalent to stating that the 2 samples were generated from the same logistic function. Of course the power of the test depends on the distribution of group sizes and the number of groups in each sample. It should also be clear from Table 1 that groups of sizes 1-10 are needed to obtain reasonable power unless very large numbers of groups are observed in both samples. Group sizes used in the study are those obtained from a sample survey " �115 FREDDY STUDY PLAN of a winter elk population. This sample survey was a stratified random quadrat count in game management unit 23, 24 on January 18-19, 1989 by C. Reichert, J.Ellenberger, and D. Freddy. During the sample survey, 84 elk groups were observed and group size recorded (Table 2). Table 2. Frequency distribution of the 84 elk group sizes. Group size 1 2 3 4 5 6 7 8 9 Frequency 9 12 10 9 10 5 4 4 6 Group size 10 11 12 13 16 18 19 20 39 Frequency 1 3 2 3 1 1 1 2 1 Two independent simulation studies were performed using the frequency distribution'of group sizes from Table 2 in the data generation step. Each study consisted in 1000 replications of pairs of samples generated as described above. In the first study, for each of the 84 listed group sizes, a determination of sighting or not was made, independently, using the 2 probability sighting functions. In the second study the frequency of groups of a given size-in Table 2 was doubled to give 2 samples of size 168, replicated 1000 times. For each replication the 2 samples were pooled and model (2) fitted. Then the null hypothesis that P2 = 0 was tested. The number of -rejections of this null hypothesis is shown for the 1000 replications in Table 3 as 10 sets of 100 tests each. The standard error of the estimated power was calculated using variation among the 10 sets. Thus given 2 independent samples of observations each based on 84 groups with frequency distribution listed in Table 2, a 5% significance level test has an estimated power of . 57 ( se-.016) of rejecting Ho: P2 = 0 when P2= -1. The estimated power when the number of groups is doubled (n-168 each ) is estimated to be .873 (se-.006). �116 FREDDY STUDY PLAN Table 3. Proportion of rejections of the null hypothesis Ho: P2 = 0 VB HA: [32 '" 0 when ~2 = -1 , for 10 sets of 100 replications each, CIt = .05. Sample I Sample II 1 .58 .89 2 .59 .89 3 .62 .88 4 .66 .86 5 .57 .91 6 .58 .88 7 .52 .85 8 .51 .88 9 .57 .86 10 .50 .90 Mean .57 .88 .016 .006 Set SE Reference Samuel, M., E. Garton, M. Schlegel, and R. Carson. 1987. Visibility bias during aerial surveys of elk in northcentral Idaho. J. Wildl. Manage. 51: 622-630. \-.-- �FREDDY STUDY PLAN APPENDIX VII Location of the study area in the eastern portion of Game Management Unit 42 within the Grand Mesa Data Analysis Unit E-14. 50 MILES NEWCASTLE GRAND MESA DAU GRAND JUNCTION ��119 Colorado Division Wildlife Research July, 1993 of Wildlife Report JOB PROGRESS State of Project Work Colorado No. Plan No Job No. Period Author: REPORT Covered: ~W~-~1~5~3~-~R~-~5~ _ Mammals Research 4 Moose 1 Development of census methods and determination of movements, habitat selection, and degree of calf mortality of moose in North Central Colorado. July Investigations 1, 1992 - June 30, 1993 R. C. Kufeld Personnel: D. Bowden, J. Bredehoft, Schoonveld, K. Snyder, S. Steinert, S. Kerr, M. Miller, S. Porter, D. Younkin, C. Wood. G. ABSTRACT Fourteen moose were captured and marked in North Park, Colorado, during October, 1992, and December, 1992. Twelve received radio-collars and eartags and 2 received only eartags. Instrumented moose captured during 1991 and 1992 were located at approximately 2-week intervals from January, 1992, through June, 1993. ��121 DEVELOPMENT OF CENSUS METHODS AND DETERMINATION OF MOVEMENTS, HABITAT SELECTION, AND DEGREE OF CALF MORTALITY OF MOOSE IN NORTH CENTRAL COLORADO Roland C. Kufeld P. N. OBJECTIVES 1. To determine the proportion of moose counting moose in North Park. 2. To determine 3. To determine the degree of dispersal of young animals, and seasonal movements, home range size, and habitat selection of North Park moose. the extent of moose SEGMENT the extent actually observed calf mortality when aerially in late winter. OBJECTIVES 1. To determine of moose calf mortality in late winter. 2. To determine the degree of dispersal of young animals, and seasonal movements, home range size, and habitat selection of North Park moose. STUDY AREA The study area was described by Kufeld METHODS (1992). AND MATERIALS Moose were captured throughout the eastern and southern part of North Park during October, 1992, and December, 1992, by tranquilizing-them with dart guns from the ground. The immobilizing drug for adult animals was 2.7 mg (0.9 cc) of Carfentanil and 40 mg (0.1 cc) of Xylazine, and for calf moose 1.35 mg (0.45 cc) of Carfentanil and 20 mg (0.05 cc) of Xylazine. This was loaded into alec dart. Tranquilized adult moose were reversed with 500 mg of Naloxone, of which 150 mg is administered interveniously (IV) and 350 mg administered subcutaneously (SQ). For calves, the Naloxone dosage was 300 mg of which 100 mg was administered IV and 200 mg SQ. Penecillin at a dosage of 30 cc for adults and 20 cc for calves was given to tranquilized moose to counteract any adverse affects of 'immobilization. Sex and age composition of the darted animals was proportional to sex and age categories occurring in the population. captured moose were fitted with 2 numbered eartags and a numbered radiocollar. Yellow eartags were used for moose captured on the east side of North Park and north of Highway 14, whereas orange eartags were used for moose captured on the east side of North Park but south of Highway 14. Each radiocollar was equipped with a mortality switch which will acitvate if the animal has not moved for 5 hours, suggesting that the animal has died. 1--- Instrumented moose captured during 1991 and 1992 (Kufeld 1992) were located at approximately 2-week intervals from January, 1992, through June, 1993, and plans call for such monitoring to continue fpr at least 3 years. Most locations were made by aerial telemetry using a Cessna 185 aircraft with a 2 element, "H" configuration receiving antenna mounted on each strut. A switchbox permitted the telemetry operator, a passenger in the aircraft, to operate antennas jointly or separately. Some locations were made by tracking on the ground until the animal was observed when the airplane was not �122 available. Moose locations were plotted on USGS 1:50,000 scale maps and recorded by UTM coordinates. Vegetation type was also recorded for each moose location. RESULTS oose capturing and monitoring Fourteen moose were marked in North Park during October, 1992, and December, 1992 (Table 1). Twelve received numbered eartags and a radio-collar. One adult bull and one bull calf, received only eartags. The radio-collared animals included 1 adult bull, 5 adult cows, 3 bull calves, and 3 cow calves. Analysis of data for movements, home range size, and habitat use will be presented in a future report when periodic monitoring of moose is completed. �Table 1. Moose radio-collared and eartagged in North Park during October and December, 1992. Collar No. No collar No collar 13' 232 393 46 47 50 51 Unnumbered Unnumbered Unnumbered Unnumbered Unnumbered , No. 38 71 20 44 47 46 72 89 76 49 77 48 39 37 Eartag Color Org. Yel. Yel. Org. Org. Org. Yel. Yel. Org. Org. Org. Org. Drg. Org. Sex Age Date Captured M M F F M F F 3.5+ Calf 2.5+ Calf Calf Calf Calf Calf Calf 3.5+ 3.5+ 1.5 3.5+ 3.5+ 12-07-92 12-07-92 09-24-92 10-08-92 12-07-92 10-09-92 12-07-92 12-07-92 12-07-92 12-07-92 12-08-92 12-07-92 12-07-92 12-07-92 M M F F F F M Capture Location S. Fork Michigan R. ).5 mi. N. Pines C.G. State For. Moose Observation Platform. State For. 0.25 mi. E. of KOA Campground. Illinois R. at Big Bottoms. State For. Custer Draw. Illinois R. at Big Bottoms. State For. at KOA Campground. Jack Cr. at Bridge North of Old Homestead. 0.2 mi. SW of Aspen Campground. State For. Moose Observation Platform. Jack Cr. at Bridge North of Old Homestead. State For. Custer Draw. 0.2 mi. SW of Aspen Campground. 0.3 mi. NW of Aspen Campground. This is the 2nd moose to wear this collar. The first one was eartag.No. 9, which lost the collar during the spring of 1992. 2 This is the 2nd moose to wear this collar. The first one was eartag .No. 30, which died during the spring of 1992. 3 This is the 2nd moose to wear this collar. The first one was eartag No. 34, which lost its collar during the spring of 1992. A • Location UTM Coords E-W N-S 413.55 414.78 412.35 411.81 413.30 411.61 412.35 406.68 412.65 414.78 406.68 413.30 412.65 412.24 483.60 494.82 • 489.52 A 473.64 493.10 474.19 489.45 A 473.28 c 485.20 b 494.82 • 473.28 c 493.10 485.20 485.20 , b , c ,.• A'pair of these symbols indicates a mother - offspring relationship. I-' tv W �124 LITERATURE CITED Kufeld, R. C. 1992. Development of census methods and determination of movements, habitat selection, and degree of calf mortality of moose in North Central Colorado. Colo. Div. Wildl. Wildl. Res. Rep. July:(in press). Prepared by Roland C. Kufeld Wildlife Researcher C � https://cpw.cvlcollections.org/files/original/3a55388eb7c2439ca26b33ae350832a2.pdf 5961ef3abc16bfc2b572238e694de1ab PDF Text Text 125 Colorado Division Wildlife Research July 1993 of Wildlife Report JOB PROGRESS State of Project Colorado No. ~W~-~1~5~3~-~R~-~3~ Work Plan No. Job No. Period REPORT Covered: Author: Personnel: _ Mammals Research lA Multispecies Investigations 1 Animal and Pen Support Facilities·for Mammals Research July 1, 1992 - June 30, 1993. M. A. Wild. T. R. Anderson, D. L. Baker, P. E. Bleicher, A. L. Case, R. B~ Gill, W. S. Graffam, B. J. Kraabel, M~ J. McArtor, C. McCarty, M. A. Miller, M. W. Miller, A. N. Torres, and J. A. Yost. ABSTRACT The Colorado Division of Wildlife's Foothills Wildlife Research Facility (FWRF) maintained captive animals (up to 139 wild and domestic ungulates of 7 species and 80 migratory and upland game birds of 6 species) and facilities supporting 9 major mammalian and avian research projects. Financial support was provided by various federal agencies for maintenance of a portion of the mule deer herd, elk calves, and whitetailed deer. In fall 1992, we recruited 28 individuals into our captive herd. During FY 1993, mortalities occurred in 8 adults (6 individuals ~10 years old) and 8 neonates (4 individuals <48 hr old). Routine animal care and facility maintenance programs were conducted as previously described with an emphasis on quality and conservation. FWRF passed 2 inspections by USDA APHIS for compliance with federal animal welfare standards, Numerous improvements were made to .fiicilities at FWRF to: increase usefulness, efficiency, quality, and/or security.· several . management experiments provided useful information on efficacy and co~t effectiveness of nutritional maintenance of captive hoofstock. Feeding grass hay rather than alfalfa hay to adult bighorn sheep reduced ov.erall costs of feed (15%) and trash removal (50%). We showed that adult mule deer and pronghorns at FWRF select an alfalfa hay diet with a relative feed va.Lue : (REV) >200 regardless of overall quality of the hay (when· offered good or excellent hay). Mule deer and pronghorns wasted more hay as hay quality decreased, indicating that obtaining the highest quality hay is most cost effective. No apparent differences occurred in plasma vitamin E levels or clinical health of bighorn sheep ewes, or their lambs, supplemented at low or high levels of vitamin E orally in feed. Addition of browse to the diets of mule deer fawns aged 52-76 days at study initiation did not effect (P > 0.05) growth rate or milk intake as compared to a nonsupplemented group. Longterm supplementation of browse may be beneficial, but is not practical under management conditions at FWRF. Copper levels provided to animals at FWRF were shown to be �126 adequate using guidelines established for cattle (10 ppm copper in feed) assuming that other minerals, such as molybdenum, do not interfere with copper utilization. Serum copper levels of captive pronghorn at FWRF were similar to those in a free-ranging herd in Wyoming (1 ppm and 0.94 ppm, respectively). We used real time, B-mode ultrasonography to accurately diagnose pregnancy status in 7 pronghorns and 6 mule deer examined 104-124 days prior to parturition. We were unable to consistently identify multiple pregnancies. �127 ANIMAL AND SUPPORT FACILITIES MAMMALS RESEARCH Margaret FOR A. Wild P. N. Objectives 1. To provide and maintain captive wildlife populations and facilities supporting CDOW's Terrestrial Wildlife Research Program, as well as programs of CDOW cooperators. 2. To develop improved animal and facility management practices will provide maximum research opportunities at minimum cost. 3. To enhance facilities research needs. to serve a growing SEGMENT improve, and expand diversity that of anticipated OBJECTIVES 1. Maintain, animal 2. Coordinate all rearing, activities. 3. Maintain up to 30 elk, 30 mountain sheep, 30 pronghorns, 45 mule deer, 12 white-tailed deer, 80 ducks and 5 upland birds in suitable health to perform required research experiments, and in accordance with federal and institutional animal welfare regulations. 4. Conduct management experiments to increase feeding and maintenance activities related operations. 5. Establish a conservation-oriented and services to operate research 6. Establish a standard program for evaluating and documenting health status of captive wildlife that will increase quality of care and reduce cost per individual. training, METHODS research maintenance, and holding facilities. and research efficiency and efficacy to research facility approach for providing facilities. of utilities AND MATERIALS Routine animal care and facility maintenance programs supporting new and ongoing Terrestrial Wildlife Research Program projects were conducted as previously described. We emphasized a quality and conservation-oriented approach in this work by striving to increase efficiency and longterm benefit from the programs and projects undertaken. Specifically, we performed the following tasks: ANIMAL MAINTENANCE General: Again this year, routine feeding and caretaking of research animals, including health observations, training, weighing, and clean-up, was performed primarily by well trained work-study and temporary employees, as well as volunteers. Beginning 1 January, volunteers maintained "timesheets" to monitor their contributions. FWRF was inspected by USDA APHIS for compliance with federal animal welfare regulations twice during FY 1993. A set of facility protocols covering cleaning and maintenance of animal enclosures and of feed storage areas, rodent control, facility security, and daily animal �128 feeding and observation inspection. were implemented as a response to the first During FY 1993, white-tailed deer were added to the species held at FWRF, and the mountain goat and domestic cow were eliminated. Animal crowding was alleviated somewhat with completion of 4 new animal paddocks. The US Fish and Wildlife Service provided financial support to maintain 17 mule deer at FWRF. The USDA Animal Damage Control provided financial support for 12 white-tailed deer fawns at FWRF. NUTRITIONAL MAINTENANCE Feeding protocols: pronghorns, mule deer, and some bighorn sheep continued to receive longstem alfalfa hay as their primary diet, with pelleted feed (Baker and Hobbs 1985) supplemented to meet energy requirements and maintain ideal body condition (Wild et al. 1992). Adult female bighorn sheep were fed grass hay in place of alfalfa hay. Elk received cubed alfalfa as their primary diet with grass hay and pelleted feed supplemented to meet scratch and vitamin/mineral requirements, respectively. Core vs. bite sample analvsis: We evaluated shipments/cuttings of alfalfa hay using two techniques: by comparing crude protein, TDN, and relative feed value (RFV) of core samples from each shipment, and by comparing RFV of core samples with "bite samples" (bite samples represented that portion of the hay actually consumed by mule deer or pronghorn and excluded waste such as large diameter hay stems). Bottle-raising neonates: In 1992, we hand-raised 19 mule deer fawns for use on the fertility control project (See WP2J16). Fawns were obtained as orphans or collected from wild does held for several months in captivity at the Little Hills Facility and transferred to FWRF when about 2 days to 3 weeks of age. Additionally, 2 orphaned pronghorn fawns were hand-raised. In 1993, 12 white-tailed deer fawns were collected from Rocky Mountain Arsenal and hand-raised. Eight pronghorn fawns and 8 mule deer fawns born to captive does at FWRF or obtained as orphans were also hand-raised. All bottle-raised neonates were fed according to Wild and Miller (1991), except that in 1992 we limit fed milk to fawns >2 months of age. In 1992, weaning was performed on a more arbitrary basis than in previous years. On 1 September (fawns aged 65-89 days) fawns were limited to 900 ml evaporated milk/day. Volume offered was reduced by 100 ml every forth day, until fawns were weaned on 24 September (fawns aged 88-112 days). Response of captive Rocky Mountain bighorn sheep to oral vitamin E supplementation: This pilot study was performed to monitor plasma vitamin E levels, as well as some other related vitamins and trace minerals, in pregnant bighorn sheep receiving 2 levels of oral vitamin E supplementation. In December, 12 potentially bred ewes were paired and then randomly assigned either to a group receiving low level (160 mg/kg pelleted feed) or high level (500 mg/kg pelleted feed) vitamin E supplementation. Although the total quantity of vitamin E provided varied over time with changes in supplemental feeding rate (from 46-218 and 143-682 IU/animal/day for low and high level, respectively), the high level group always received 3 times more vitamin E than the low level group. Blood samples collected in heparinized tubes were obtained from bighorn sheep ewes monthly. We also collected blood from ewes and lambs about 24 hr post-partum, and biweekly from lambs thereafter. Plasma alpha tocopherol levels were determined using the technique of Miller and Chung (1984). Clinical health of all bighorn sheep was monitored daily. Addition of browse to fawn diets: We performed a management experiment to assess the utility of feeding browse in addition to standard feed stuffs (evaporated milk, alfalfa hay, calf manna) to mule deer fawns. We �129 hypothesized that availability of browse may increase dry feed intake by neonates and thus decrease milk intake and/or increase growth rates. Sixteen bottle-raised fawns, age 52-76 days at study initiation, were randomly assigned to control/standard management or treatment/browse added groups. Ad libitum quantities of mountain mahogany were provided to the treatment group daily for 28 days. Daily milk intake and weekly body weights were recorded for all individuals, and daily group intake of mountain mahogany (excluding calculated evaporative loss) was determined. Investigations into possible copper deficiency: We investigated possible trace mineral and vitamin deficiencies in our captive ruminants. Copper deficiency has been suspected of causing, or exacerbating, subtle health problems in captive pronghorns and possibly elk held at FWRF. Clinical signs observed that could be attributed to copper deficiency include: slightly dull and brittle hair coats in elk; and ataxia in young pronghorn, and immunocompromise and suboptimal reproductive success in adult pronghorns. We collected data to identify levels of copper offered in feed, and serum copper levels of the pronghorn herd. HEALTH MAINTENANCE General: We continued to use and revise the Daily Animal Observation Form to facilitate communication of health problems from caretakers to veterinarians, and to provide a written health history for each animal. Animal care protocols were written to further educate caretakers. Professional veterinary care was provided by CDOW employees. An annual preventive medicine protocol was implemented to assure timely vaccination and anthelmintic treatment of research animals. Neonatal illnesses: All mule deer fawns collected at the Little Hills facility and relocated to FWRF arrived diarrhetic, or developed diarrhea shortly after arrival. At FWRF, we implemented strict sanitation and isolation procedures to prevent disease transmission. III fawns received intensive therapy (Wild et ale 1992). Two other health problems recognized in the fawns were chronic eye infections in most fawns and recurrent seizures in 3 fawns. Fawns arrived at FWRF from Little Hills with mild to copious mucopurulent ocular discharge. Triple antibiotic ophthalmic ointment was used initially, then later oxytetracycline ophthalmic ointment (Terramycin®) .. Seizures occurred sporadically over a 2 week period in 3 fawns (C92, D92, and 192) aged 3-5 weeks. Seizures were characterized by circling, then collapse, loss of consciousness, muscle rigidity, and periodic apnea lasting ~l min. Pregnancy diagnosis: We investigated the utility of ultrasonography to diagnose and stage pregnancy in captive pronghorns and mule deer. Pronghorns were sedated with Telazol (about 130 mg/animal, IV) and mule deer with xylazine (about 50-80 mg/animal, 1M) prior to ultrasonographic examinations. We performed uterine examinations using a real-time, Bmode, diagnostic ultrasound scanner with a linear-array transducer designed for intrarectal placement. The transducer was manipulated intrarectally using a probe extension fabricated for use in llamas (L. Johnson, Colorado State University, Fort Collins, CO). We measured head width and crown-nose length of fetuses that were in appropriate body position using electronic calipers. FACILITY MAINTENANCE/REPAIRS/IMPROVEMENTS A variety of scheduled and unscheduled maintenance and repair activities were necessary to support facility operation and ongoing research programs. We worked toward a conservation-oriented approach for facility care by undertaking preventive maintenance projects, and performing highquality new construction and repairs to existing facilities. �130 RESEARCH PROJECTS Facility operations offered support for pilot studies, student special studies, and CDOW and cooperative research experiments that were initiated, conducted, or continued using FWRF animals and facilities throughout the year. EDUCATIONAL CONTRIBUTIONS Facility tours and educational lectures were provided to high school, university, and professional groups visiting FWRF. We emphasized the importance of maintaining captive wildlife for performing controlled experiments and the contributions made by research projects performed at FWRF. FWRF animals and facilities were also used occasionally for handson training for professional groups. RESULTS ANIMAL AND DISCUSSION MAINTENANCE General: Detailed feeding instructions, feed consumption forms and health records facilitated communication between caretakers and managers/veterinarians, thus optimizing overall quality of animal care. When volunteers were carefully selected and trained similarly to paid employees, their contribution was remarkable. Four volunteers contributed a total of 194 hours of work at FWRF between 1 January-3D June. Volunteers performed primarily caretaker tasks and also assisted in weighing and collecting samples from animals. Their contribution represented a savings of about $1800 to FWRF over a 6 month period (vs. cost of temporary employees). The 29 July 1992 inspection by USDA APHIS revealed minor deficiencies in 3 areas. All breaches were corrected within 48 hours. Implementation of facility protocols apparently aided in avoiding repeat violations. No facility violations were recorded on the next inspection on 6 May 1993. New animal paddocks added needed space for research animals at FWRF. Mule deer and elk were moved into 4 new animal paddocks. Moving these animals created more space for pronghorns in the east side pastures, and decreased stocking density of elk. In fall 1992, we recruited 6 bighorn sheep, 18 mule deer, 2 pronghorns, and 2 elk into our captive herd. We complied with RFAC imposed population limits of number of animals financially supported by FWRF's budget. At the close of FY 1993, FWRF housed 26 bighorn sheep, 30 elk, 29 pronghorns, 43 mule deer, and 12 white-tailed deer. However, alternate sources of funding financially supported 10 of the elk, 4 of the pronghorns, 31 of the mule deer, and all white-tailed deer. NUTRITIONAL MAINTENANCE Feeding protocols: All species maintained reasonable body condition on available diets. Bighorn sheep readily consumed high quality grass hay as an alternative to longstem alfalfa hay. Although grass hay was more expensive on a per ton basis, waste was reduced by about 50%. This resulted in overall savings on costs of hay (15%), labor to clean pens (40%), and trash removal (50%). When high quality grass hay can be obtained, we recommend that it be used as the primary feed source for adult bighorn sheep. Core vs bite sample analysis: Comparison of alfalfa hay core and bite sample analysis provided useful information on the true quality of hay �131 consumed by mule deer and pronghorns. Core samples provided necessary information that allowed us to compare hay with published values and assure quality standards. Analysis of bite samples gave a more accurate description of actual nutrient intake and also served as an indicator of waste associated with the hay. Mule deer and pronghorns appeared to select diets with RFV greater than 200 regardless of overall quality of the hay (at least when offered good and excellent quality hay; Table ). Discrepancy in RFV of core and bite samples was greater for hay that was observed to be stemmy than for leafy hay with small stems (Table 1). Mule deer and pronghorns consumed diets with >200 RFV by wasting more hay as RFV of core samples fell. This observation aids in the justification of purchasing the highest quality alfalfa hay available for feeding mule deer and pronghorns. Bottle-raising neonates: Evaporated milk fed ad libitum again appeared to be an adequate diet for hand-raised neonates. We attributed occurrence of enteric disease primarily to the husbandry practice of obtaining fawns from wild does held in captivity (Trindle et al. 1978) rather than to inadequacy of the evaporated milk diet (Wild and Miller 1991). We used the ad libitum evaporated milk diet in a new species, whitetailed deer, in 1993. All hand-raised neonates in 1993, 12 white-tailed deer, 8 mule deer, and 8 pronghorns, remained healthy. Although clinical health of mule deer raised in 1992 and 1993 differed markedly, growth rate during the first month were similar between the 2 groups (Fig. 1). Response of captive Rocky Mountain bighorn sheep to oral vitamin E supplementation: Mean monthly plasma vitamin E (alpha tocopherol) levels did not differ markedly between groups supplemented with low and high levels of vitamin E (Fig. 2). Although supplementation was at its peak and green forage was available to bighorn sheep in May and June, both groups experienced an apparent decline in plasma vitamin E during this time. This decline could be attributed to sequestering of vitamin E in the milk of dams, but declines occurred in barren ewes as well. Vitamin E requirements of bighorn sheep and/or bioavailability of vitamin E to bighorn sheep may change throughout the year. Alternately, plasma alpha tocopherol may not be an accurate indicator of vitamin E status in bighorn sheep. Future studies of vitamin E supplementation in adult bighorn sheep should focus on alternate techniques for measuring alpha tocopherol, the importance of gamma tocopherol, and the interaction of vitamin E with other trace minerals (e.g. copper, selenium). Small sample sizes precluded thorough investigation into the effects of vitamin E supplementation in bighorn lambs (via supplementation to the dam). Only 4 (2 low and 2 high level supplemented) of 12 potentially bred ewes gave birth to lambs, and one of these lambs (from a high level supplemented ewe) died at about 48 hr of age. Vitamin E levels and immune competence appeared similar in the 3 lambs during the first month of life. Addition of browse to fawn diets: All fawns readily consumed browse, but daily intakes were quite low. Mean daily intakes ranged from 0-146 g/fawn, and averaged 81 g/fawn (SE = 5) or about 0.5% of their body weight. No difference (P > 0.05) in growth rate between control and treatment groups was observed. Milk intake of treatment and control fawns appeared similar. Browse intake did not increase markedly when quantity of milk offered was reduced. More subtle effects of supplementation, such as stimulation of rumen development or forage selection patterns, are difficult to quantify, but may be beneficial in the longterm. Because browse was supplemented for only a limited period in this study, questions concerning the benefit of browse to neonates' remain. Further investigation should include evaluation of longterm browse supplementation to neonates' diet; however, acquisition of fresh browse, at least at FWRF, is a major obstacle to this study. �132 Investigations into possible copper deficiency: No recommended daily allowances for copper are available for wild ungulates, but requirements may be similar to the 10 ppm recommended for cattle. Our pelleted supplemental feed contained 66 ppm copper. Copper levels in alfalfa hay varied widely with source, from 61 ppm in our first cutting hay to 11 ppm in fourth cutting hay. Hay cubes contained 15 ppm copper. Therefore, copper levels presented should be adequate. Additionally, in this geographic location (and locations of hay sources) we would not suspect high levels of other trace minerals that would cause secondary copper deficiency. A mean serum copper level of 1.0 (SE = 0.06) in the herd also indicated that copper supplementation was probably adequate (Table 1). Normal values for serum copper in all species of animals are generally reported as about ~1 ppm. Haigh (1989) reported normal values in elk as >0.76 ppm, and the mean value in a healthy free-ranging pronghorn herd in Wyoming was 0.94 ppm (Gonzales, unpub. data). These data suggest that copper deficiency is probably not a significant problem in our captive pronghorn herd; however, we will continue to monitor the situation by investigating more sophisticated sample analysis techniques (i.e. serum copper:protein and liver copper levels), monitoring feed copper and molybdenum levels, and identifying normal levels in other free-ranging pronghorn herds. HEALTH MAINTENANCE General: Our system for recording daily health status for research animals continued to improve communications, detection and treatment of animal health problems at FWRF. Overall, captive wildlife maintained at FWRF remained healthy throughout the year. During FY 1993, 8 mortalities occurred in adult hoofstock and 8 in neonates at FWRF (Table 2). Six of the 8 adults that died were ~10 years old, and 4 of the 8 neonates that died were <48 hr old. Neonatal illnesses: Strict sanitation and isolation procedures coupled with intensive therapy resulted in survival of 16 of 19 mule deer fawns in 1992. Twelve of the 16 survivors required antibiotic therapy, and Salmonella sp. (serogroup B, serotype S. heidelberg), rotavirus and/or coronavirus were detected in a subsample of these fawns. At weaning all fawns appeared healthy; however, one fawn died shortly thereafter from apparent viral enteritis. Two fawns obtained at weaning from rehabilitators remained healthy when placed with recovered fawns in our herd. Triple antibiotic ophthalmic ointment was ineffective in treating eye infections in mule deer fawns. Scrapings for chlamydia detection were unsuccessful, culture revealed no mycoplasma, but heavy growth of Branhamella ovis occurred. Oxytetracycline ophthalmic ointment (Terramycin®) treated signs, but discharge frequently recurred after therapy was discontinued, requiring treatment to continue longterm. Neurologic exam and blood diagnostic panels from seizuring mule deer fawns revealed no abnormalities. Occurrence of seizures ceased without therapy and cause was not determined. Clinical illness in the 1992 mule deer fawns, and associated postmortem findings on 3 mortalities, suggested some degree of immune compromise. Immune compromise likely resulted from maternal stress and/or inadequate passive transfer of immunoglobulins to fawns. Similar observations were made by Parkinson et ale (1982) with fawns obtained from wild does held in captivity under conditions analogous to those at Little Hills. In the future, collecting fawns from wild does held in captivity should be discouraged, or at the least, visual barriers should be provided for shelter and to reduce stress on does (Trindle et ale 1978). �133 Pregnancy diagnosis: Pregnancy status was accurately diagnosed in 7 pronghorn and 6 mule deer does. Six pronghorn does were diagnosed as pregnant 113-124 days prior to parturition; the other doe was not pregnant. Two fetuses were observed on ultrasound in only 2 of 5 does that gave birth to twins. In another doe (BL91), two fetuses that were very disparate in size (thorax length of one about one third that of the sibling) were observed on ultrasound, but only one fawn was born. Six mule deer does were confirmed pregnant at 104-122 days prior to parturition. Multiple pregnancies were not confirmed in these does, but 4 of 6 gave birth to twins. We measured head width and crown-nose length of fetuses that were in appropriate body position using electronic calipers (Table 3). These measurements from known-age fetuses may be useful in the future for estimating age of fetuses in utero. FACILITY MAINTENANCE/REPAIRS/IMPROVEMENTS In addition to numerous daily repairs performed several major improvements. repair/improvement projects completed - - - and maintenance projects, we Significant maintenance/ at FWRF this year included: Stripping and repainting metabolic cages to prevent contamination of samples and deterioration of cages; constructing caps to cover inground collection wells outside the cages. Designing, planning, and contracting for installation of automatic waterers for new west side paddocks. Renovation of FWRF office, including: electrical rewiring and plumbing as necessary, constructing an office partition and closet, placing adequate insulation and weather stripping, sealing the roof, installing a suspended ceiling with light panels, and finishing interior walls using sheetrock. Completion of the replacement of the south section of the west alleyway. Construction of animal shelter and feed area in new paddock F. Construction of animal shelters in new paddocks B, C, and E. Replacement of all ramp-type shelters in bighorn sheep pens. Landscaping of pastures WI and W2 to reclaim land upset by previous construction. Construction of a new drive-thru gate on the north perimeter fence to provide easier access and increased security. Construction of a new entrance and gate to the west side of the facility for increased security and clearance for large vehicles. Replacement of entrance gate to east side after wind damage. Construction/replacement of several gates in animal pastures. Installation of a new heater in the west side scaleroom. Installation of lighting for east and west scales. Modifications to east scale to improve drainage. RESEARCH PROJECTS In addition to ongoing facility management experiments and improvements described above, the following (listed in no particular order of importance) pilot studies, special studies, and research experiments were initiated, conducted, or continued using FWRF animals and facilities this year: - Regulation of mule deer population growth by fertility control: laboratory, field, and simulation experiments (initial GnRH studies on mule deer, pronghorns, and elk) -- Baker, Nett, Miller, Hobbs, and Gill. - Serum and fecal progesterone levels in elk, mule deer, through gestation -- Baker, Wild, and Case. and pronghorns �134 Detection detection of breeding activity in elk and mule deer using devices -- Baker and McArtor. - Response of captive Rocky Mountain bighorn supplementation -- Wild, Graffam, Irlbeck, remote sheep to oral vitamin Miller, and Nockels. E - Effects of refrigeration on recovery of Pasteurella haemolytica from bighorn sheep pharyngeal swabs transported in modified Cary and Blair medium -- Wild and Miller. - Comparison of corrosion potential and the inflammatory response tungsten-bismuth-tin shot and steel shot in mallards -- Kraabel, Miller, Getzy, and Ringelman. - Addition of browse and Wild. to the diets of hand-raised fawns -- M. A. Miller - Hemorheological experiments on pronghorn blood with applications human biomedical research -- Popel, Kameneva, and Wild. - of to New models of the functional response in vertebrate herbivores: the role of plant availability and animal morphology (locomotion studies) -- Spalinger, Hobbs, Wunder, and Gross. EDUCATIONAL CONTRIBUTIONS FWRF provided formal educational instruction for 4 high school vocational education classes, 4 university classes, 2 Project Wild workshops, and 3 professional groups. Animals and facilities were used for hands-on training with 2 professional groups. Numerous other informal tours were provided individually to visiting professionals. LITERATURE CITED BAKER, D. L., AND N. T. HOBBS. 1985. Emergency feeding of mule deer during winter: tests of a supplemental ration. J. Wildl. Manage. 49: 934-942. HAIGH, J. C. 1989. Copper - deficiency and toxicity, in Game Practice. University of Saskatchewan. pp. 7-01 - 7-04. Farming MILLER, K. W., AND C. S. CHUNG. 1984. An isocratic high performance liquid chromatography method for the simultaneous analysis of plasma retinol, alpha tocopherol and various carotenoids. Analytical Biochemistry 138:340. PARKINSON, D. E., R. P. ELLIS, AND L. D. LEWIS. 1982. Colostrum deficiency in mule deer fawns: Identification, treatment and influence on neonatal mortality. J. Wildl. Dis. 18:17-28. TRINDLE, B. D., L. D. LEWIS, AND L. H. LAUERMAN. 1978. Evaluation stress on the immune system of hand-reared mule deer fawns (Odocoileus hemionus). J. Wildl. Dis. 14:523-537. WILD, M. A., and M. W. MILLER. 1991. Bottle-raising wild ruminants captivity. Outdoor Facts #114, Colo. Div. Wildl., Denver, 6pp. of in WILD, M. A., MILLER, M. W., MAYNARD, B. J., AND MAGNUSON, D. R. 1992. Animal and pen support facilities for terrestrial wildlife research. Colo. Div. Wildl. Res. Rep. Fed. Aid Proj. W-153-R3, WP1A J1, Job Progr. Rep., July 1991-June 1992, Fort Collins. �135 Table 1. Comparison of dry matter RFV of core and bite samples from shipments/cuttings of alfalfa hay fed to mule deer and pronghorns at FWRF. RFV Species Hay description Mule deer First cutting-large stems; loose, crushed leaves Forth cutting-small many leaves Bite 175 242 176 206 187 236 212 217 stems; Third cutting-medium leafy Pronghorn Core stems; Forth cutting-few small stems; very leafy Table 2. Summary Species Bighorn of mortalities Age (yrs) Cause 083 G77 G78 H83 M82 Q93 10 15 14 10 11 0 Bacteremia, Trauma Age related Trauma Trauma Septicemia P78 14 Meningitis, Animal sheep Mountain goat in hoof stock at FWRF during ID FY 1993. of Death lumpy jaw lesions changes, neuritis Pronghorns B088 BLa93 SEa93 SEb93 4 0 0 0 Trauma Hypoglycemia Hypoxia at birth Septicemia Mule deer B91 J92 1 2 3 2 0 0 0 0 Trauma Enteritis Enteritis Enteritis Enteritis weight loss �136 Table 3. In utero fetal measurements mule deer does. (cm) from captive Ultrasound Date Days prepartum Crownnose length Pronghorn BL91 DD91 SE91 BN91 NK91 Y091 2/17/93 2/17/93 2/17/93 2/24/93 2/24/93 2/24/93 -113 -124 -118 -120 -117 -114 6 6 5.3 5.5 Mule deer A91 B91 E91 R91 R86 S90 2/19/93 2/19/93 2/19/93 2/19/93 2/19/93 2/19/93 -120 -117 -122 -113 -104 -113 Animal ID "Not determined due to improper pronghorn Head width 5.3 3.2 2.9 2.8 3.6 nd 3.2 6 6.9 nd nd 6.5 nd 3.7 2.6 nd nd 3.5 nd nd" fetal position. and �.J...J I 30 --... 20 Cl ~ lI o w 3: 10 o ~ __~ __~ __~ __ o 7 14 21 -L__~ 28 __~ 35 __~ 42 __~ 49 ~ 56 63 __~ 70 __~ __~ rt 83 __~ __~ 90 AGE (days) Fig. 1. Mean and 1993. body weight (±1 SE) of mule deer fawns hand-raised in 1992 __~ 97 104 �138 200 w z ~ « I> 100 ---- -c ---- ---I -- ~ en « _J n, o Nov Dec Jan Feb Mar Apr May MONTH Fig 2. Mean plasma vitamin E levels (±1 SE) of bighorn sheep ewes supplemented with oral vitamin E at a low level (-) and a high level (--). Jun �139 JOB PROGRESS REPORT state of Project Colorado No. W-153-R-6 Mammals Research Work Plan No. 1A Multispecies Job No. 3 Mammals Period Covered: Author: Personnel: July 1, 1992-June 2 Research Administration 30, 1993 R. B. Gill R. B. Gill and D. K. Hall ABSTRACT Fiscal resources allocated to the Mammals 2 Research Section for FY 92-93 overexpended by $5,725.12. The primary reason for the cost overrun was unanticipated equipment needs for 2 research projects. During FY 92-93, 1 new project was started, 2 ongoing projects were and 1 project proposal was prepared and submitted for decision. were concluded, ��141 MAMMALS 1 RESEARCH ADMINISTRATION R. Bruce Gill P.N. OBJECTIVE Administer research studies within productivity at the lowest cost. the Mammals Agreement 1. 2 Research Unit for the highest Objectives Supervise and administer research on species moose in the Mammals Research Section. other than deer, elk, and RESULTS The total budget request for FY 1992-93 for Mammals 2 Research was $883,297. A total of $889,022.14 was expended to accomplish research programmed for FY 1992-93. The majority of the cost overrun (> 90%) resulted from overexpenditures in 2 projects; Work Plan lOA Job 1 (34%) and Work Plan 11A Job 1 (60%). Overexpenditures in both jobs resulted from unanticipated equipment needs. One new research start was proposed and planned in FY 92-93 and inaugurated FY 1993-94: Work P~an SA Job 2 - Deve~opmen~ of B~ack Bear Inven~ory Techniques. Planning on another new research start - Work P~an 3A Job 6 - Pronghorn Win~er Whea~ Damage S~udy was started in FY 92-93, will be concluded in FY 93-94 and field work will commence in winter-spring of FY 93-94. ~ Two research projects, Work P~an SA Job 1 and Work P~an 9A Job 1, were concluded in FY 92-93 and results of Work P~an S A Job 1 will be summarized in a Division of Wildlife Standard Operating Procedure. Results of Work P~an 9A Job 1 is in the peer review process for the journal Eco~ogic~ App~ica~ions. ~ The acting Mammals 1 Research Leader actively participated in the Colorado Division of Wildlife's Long Range Plan update process, the Mount Evans management planning process, the Terrestrial Wildlife Reorganization planning process, and the Human Dimensions Advisory Committee, the Furbearer Management Analysis process, the Post-Election Evaluation of Amendment 10 - the Black Bear Management Initiative, the North Park Wildlife Viewing Pilot Project, and the Rocky Mountain Arsenal fertility control project. �142 ~ A research proposal was prepared outlining research protocol and anticipated costs for a study to evaluate post-dispersal mortality pronghorn populations. The proposal is tentatively scheduled for implementation during FY 94-95. Prepared by R. Bruce Gill Wildlife Research Leader in �143 Colorado Division Wildlife Research July 1993 of Wildlife Report JOB PROGRESS State of Project Colorado No. ~W_-~1~5~3~-~R~-~2~ Work Plan No. __~l~A~ Job No. Period REPORT 5 Covered: Author: _ Mammals _ Multispecies Investigations Consulting Services for Mark-Recapture Analysis Research July 1, 1992 - June 30, 1993 G. C. White Personnel: R. M. Bartmann, R. B. Gill, T. D. I. Beck, D. J. Freddy ABSTRACT Progress towards the objectives of this job include: 1. A user-friendly computer program (Program NOREMARK) was developed operate on personal computers for the computation of population estimates based on resightings of marked animals. 2. Alternative estimators and sampling designs for estimating numbers at the Bear Management Unit level were explored. 3. A study of compensatory effects of harvest on the Piceance Basin mule deer population was continued as part of Federal Aid Project W-153-R Work Plan 2 Job 15, entitled Compensatory Effects of Harvest in a Mule Deer Population. Experimental harvests have been conducted in December 1989, 1990, and 1991. Radio collars to monitor over-winter survival of fawns were placed on the animals during November 1989, 1990, 1991, and 1992. 4. Estimates of deer and elk harvest were compared between at check station interviews and phone interviews. 5. Consultation has been provided in the design and analysis of a statistical protocol for estimation of the statewide black bec;tr population, and of procedu~es to estimate elk Burvival and eiightability from the air. . 1 black hunter to bear surveys ��145 CONSULTING SERVICES FOR MARK-RECAPTURE ANALYSES G. C. White P. N. OBJECTIVES Evaluate compensatory population. effects of harvest SEGMENT 1. on the Piceance Basin mule deer OBJECTIVES Evaluate the impact of capture by drop nets and helicopter net gunning on the survival of mule deer fawns 2 and 4 weeks post capture. RESULTS AND DISCUSSION Introduction. The assumption that a sample of marked animals from a population is a representative (or random) sample of the population is required for any inference from the sample to the population. However, this necessary assumption is seldom tested because biologists have no opportunity to compare a sample of animals to the remainder of the population. Usually, a sample of animals is compared to another sample. For example, Garrott and White (1982) found that door height and bait placement with Clover traps (Clover 1956) affected the age and sex of mule deer captured. Adult females were more likely to be captured when the door was fully open and most of the bait was at the rear of the trap. When the door was set lower and most of the bait was placed in front of the trap, more female fawns were captured. A variety of trapping techniques have been developed for capturing larger ungulates. The advantages and disadvantages of each for a particular situation must be carefully considered when arriving at a choice. Thus, additional information on biases and animal safety can only help in these decisions. In this study, we evaluate the use of drop nets (Ramsey 1968, Schmidt et al. 1978) and helicopter net guns (Barrett et al. 1982, van Reenen 1982) to capture mule deer fawns during winter. Despite the longer history of use for drop nets, there has been little formal evaluation of their effectiveness and incidence of capture-related mortality. Net gunning from a helicopter, on the other hand, has received more attention. Some initial evaluations of helicopter net gunning to capture various ungulate species were based on sample sizes ranging only from 2-17 (Barrett et al. 1982, Andryk et al. 1983, Krausman et al. 1985) and the few deaths reported seemed largely a result of the learning experience with a new technique. No deaths attributable to injuries or capture myopathy were confirmed in these studies after animals were released. Kock et al. (1987) evaluated trapping data from the western U.S. between 1980-86 for bighorn sheep (Ovis canadensis) captured with drop nets, net guns, drive nets, and chemical immobilization. They found a significant (~ < 0.05) relationship between mortality and capture method with net guns having the lowest total mortality (including accidents and capture myopathy) followed closely by drop nets. They also found that higher proportions of younger animals were captured with drop nets and drive nets than with net gunning and chemical immobilization, but no differences in sex ratios were apparent among methods. They acknowledged the selectivity potential with net gunning and chemical immobilization, but some selectivity is also possible with the other 2 methods. For instance, in our study, we were primarily interested in capturing fawn mule deer and we would drop a net as soon as fawns were present. Because of our desire to capture fawns, a concern with helicopter net gunning was a bias in the size of fawns captured. We thought the helicopter crew, to avoid confusing larger fawns with smaller adults, might tend to select the smallest deer in a group. Therefore, we included body size in addition to sex ratio and post-capture survival in our evaluation of drop nets and helicopter net gunning. 2 �146 Methods. Mule deer fawns were captured with 21.5 X 21.5-m drop nets baited with alfalfa hay and fermented apple pomace. Up to 10 nets were operated each afternoon until dark for 19 days from 11 November to 4 December 1992 to capture 86 fawns. Animals were blind-folded and hobbled under the net, placed on a stretcher, and weighed to the nearest 0.1 kg. They were taken off the stretcher and total body length (cm) and left hind foot length (cm) measured. Fawns were then radio collared, ear-tagged, and released. Radios contained a motion sensor that triggered after 3-4 hours of inactivity to indicate mortality. Net gunning from a Hughes 500C helicopter occurred from 2-5 December 1992. seventy-nine fawns and 9 does were captured in 19.5 hours of air time. The does were captured intentionally and were not a result of misidentifying fawns. Net-gunned fawns were hobbled and blindfolded and then transported in the rear compartment of the helicopter to a processing location <2.5 km from the capture site. Measurements and marking procedures were the same as for drop-netted fawns. The fawns were then either transported back to the capture site and released or released at the processing location if <0.5 km from the capture site. After capture, radio-collared fawns were monitored for mortality 3-5 days/week. When a mortality signal was heard, the deer was located to verify mortality and cause of death. statistical analyses were performed with SAS (SAS Institute Inc. 1987) procedures FREQ and GLM. Sex ratio was tested with a 2 X 2 Chi-square table categorized by capture method and sex. Differences in means of body size measurements were tested with an ANOVA model consisting of capture method, sex, and their interaction term. Variances of body measurements were tested with the Levene median test suggested by Conover et ale (1981). Survival was tested with a 2 X 2 Chi-square table categorized by method of capture and survival status of each fawn 2 and 4 weeks after capture. A longer period was not used because most mortality related to capture myopathy tends to occur within 2-4 weeks post-capture (Spraker 1982). Too, as the post-capture period increases, capture myopathy becomes more confounded with other mortality causes. Results. Sex ratios of net-gunned (52.3% males) and drop-netted fawns (54.4% males) were not different (~ = 0.787) (Table 1). Neither were there differences in mean weight (~ = 0.708) or left hind foot length (~ = 0.613) between capture methods, but total body length of drop-netted fawns averaged 3.0 cm longer than net-gunned fawns (~ = 0.010). Sex was significant for all 3 body measurements (~~ 0.015), but the interaction of sex and capture method was not (~ > 0.838). Variances also did not differ across the 4 sex and capture method classes for the 3 variables (~~ 0.223). Nine of 86 drop-netted fawns were dead ~2 weeks after capture compared to 4 of 79 net-gunned fawns (~ = 0.192). At 4 weeks after capture, 15 of 86 and 8 of 79 were dead (~ = 0.172). Capture-related deaths of 5 drop-netted fawns occurred within 3-9 days post-capture compared to no such mortality for net-gunned fawns. Cause of death for 1 fawn in each group could not be determined although predation was suspected in both cases. All other deaths in both groups were from predation or hunting, although capture stress may have predisposed some fawns to these causes. Drop netting took 365 person-days versus 6 (exclusive of the helicopter crew) for net gunning (Table 2). These totals do not include travel time to and from the study area. Assigning monetary costs to the days for drop netting is difficult because of the number of different people that participated for varying periods. There were 3 permanent project personnel, 7 technicians hired only for trapping, numerous CDOW non-project personnel, and student volunteers. Our best estimate of cost/fawn captured is $448 with drop nets and $270 with net guns. If non-project personnel were replaced with additional temporary technicians, the cost/fawn for drop netting would drop to $367. Not included in these calculations are 81 does and 22 bucks also captured with drop nets and 9 does captured with net guns. Including the these extra deer would reduce the cost/animal for drop nets but would have little effect for net gunning because the helicopter contract, the major expense, was on a per-animal basis. 3 �147 Discussion. We attribute the difference in total body length between dropnetted and net-gunned fawns as an artifact of the method of hobbling. Dropnetted fawns were hobbled by securing each leg with a strap attached to a central ring. Legs of net-gunned fawns were hobbled in pairs. The front and rear legs on 1 side were held parallel to each other, hooves pointing in opposite directions, and secured with 2-3 wraps of a 5-cm-wide leather belt. This greater overlap of the front and rear legs created more spinal curvature that yielded a slightly longer body length. Other than this measurement, body size of fawns captured by the 2 methods did not differ with regard to the mean and variance of the 3 measurements. Based on capture-related deaths, net gunning was a safer capture method for mule deer fawns than drop nets. However, based on total mortal~ty, dropnetted and net-gunned fawns were equally likely to survive the first 4 weeks after capture. With the sample sizes in this study and a survival rate of 0.899 for the 4-week interval for net-gunned fawns, this study had power equal to 0.80 to detect a 0.17 difference in survival for the 2 groups of fawns. Therefore, we are 80% confident the actual difference in survival the first 4 weeks after capture was <0.17 and conclude survival was not different between the 2 methods. Several other factors to consider to evaluate the 2 capture methods cannot be assigned a monetary value. With drop netting, there is no guarantee the desired number of fawns (160 for this project) could be caught in the allotted 5-week period due primarily to weather conditions that affect trapping success. In contrast, with net gunning, the sample size should easily be attained within a 1-2-week period based on a proven capture rate of 20-25 deer/day. Another factor is the better distribution over the study area of fawns captured from a helicopter. Although vegetation and terrain obstacles exclude some areas as capture sites, most deer can be pushed to locations where capture is possible. On the other hand, drop nets must be placed near roads for vehicular access and sites must be reasonably level and easily cleared of shrubs and trees. The deer also must encounter the bait sites in their normal movements. Thus, some deer are never susceptible to capture when trap sites are too widely spaced. Finally, there is the effect of stress on survival. One way to quantify this is to compare the time animals are restrained with both methods. With drop nets, we tried to process and release all fawns first. The range in times varied from about 5 to 20 minutes, depending on the number and kinds of deer captured and the number of handlers available. We did not record times for all net drops so we cannot calculate an average. With net gunning, the range in restraint times was much less with most variation associated with pursuit and capture. Once captured, fawns were delivered to the processing location in ~2 minutes, processed in about 3 minutes, and returned to the capture site either immediately or ~3 minutes. Assuming 3-4 minutes for pursuit and capture, total restraint times for most fawns varied from 7-12 minutes. Total times were slightly greater on the few occasions when >1 fawn was captured. However, the increase amounted to <5 minutes because multiple captures were made only with fawns in close proximity to each other. These restraint times are reasonable based on an average capture rate of 4 fawns/hour with some allowance for search time, and compare favorably with times reported by Andryk et ale (1983) and Krausman et ale (1985). The longer restraint times with drop nets probably contributed to the 5 capture-related deaths. Fawns tended to struggle longer under drop nets because it usually took longer to untangle them, especially when too few handlers were available. Also, more than 1 fawn was usually captured so each had to wait its turn for processing. These delays seldom occurred with net gunning because most fawns were brought in separately. Under the conditions of this study, helicopter net gunning was a cheaper and more efficient method of capturing mule deer fawns than drop netting. Although we found no significant difference in survival to 4 weeks postcapture, we also believe net gunning to be safer because we could not verify any capture-related mortality. 4 �148 Summary. Mule deer fawns (Odocoileus hemionus) were captured with drop nets (n = 86) and helicopter net guns (n = 79). Sex ratio, weight, and left hind foot length did not differ between the 2 capture methods (~~ 0.613). The longer body length of drop-netted fawns (~ = 0.010) was presumed due to different methods of hobbling the animals while they were measured. Variances were not significantly different (~~ 0.223) for any body measurements of the 4 sex and method classes. Survival at 2-weeks post-capture for drop-netted (89.5%) and net-gunned fawns (94.9%) was not different (~ = 0.192), nor was survival different at 4 weeks (82.6% versus 89.9%, ~ = 0.172). Literature Cited. Andryk, T. A., L. R. Irby, D. L. Hook, J. J. McCarthy, and G. Olson. 1983. Comparison of mountain sheep capture techniqUes: helicopter darting versus net-gunning. Wildl. Soc. Bull. 11:184-187. Barrett, M. W., J. W. Nolan, and L. D. Roy. 1982. Evaluation of a hand-held net-gun to capture large mammals. Wildl. Soc. Bull. 10:108-114. Clover, M. R., 1956. Single-gate deer trap. Calif. Fish and Game 42:199-201. Conover, W. J., M. E. Johnson, and M. M. Johnson. 1981. A comparative study of tests of homogeneity of variances, with applications to the outer continental shelf bidding data. Technometrics 23:351-361. Garrott, R. A., and G. C. White. 1982. Age and sex selectivity in trapping mule deer. J. Wildl. Manage. 46:1083-1086. Kock, M. D., D. A. Jessup, R. K. Clark, C. E. Franti, and R. A. Weaver. 1987. Capture methods in five subspecies of free-ranging bighorn sheep: an evaluation of drop-net, drive-net, chemical immobilization and the netgun. J. Wildl. Dis. 23:634-640. Krausman, P. R., J. J. Hervert, and L. L. Ordway. 1985. Capturing deer and mountain sheep with a net-gun. Wildl. Soc. Bull. 13:71-73. Ramsey, C. W. 1968. A drop-net deer trap. J. Wildl. Manage. 32:187-190. SAS Institute, Inc. 1987. SAS/STAT™ guide for personal computers, version 6 ed. SAS Inst., Inc., Cary, N.C. 1028pp. Schmidt, R. L., W. H. Rutherford, and F. M. Bodenham. 1978. Colorado bighorn sheep-trapping techniques. Wildl. Soc. Bull. 6:159-163. Spraker, T. A. 1982. An overview of the pathophysiology of capture myopathy and related conditions that occur at the time of capture of wild animals. Pages 83-118 in L. Nielsen, J.C. Haigh, and M. E. Fowler, eds. Chemical immobilization of North American wildlife. Wisconsin Humane Soc., Milwaukee. van Reenen, G. 1982. Field experience in the capture of red deer by helicopter in New Zealand with reference to post-capture sequela and management. Pages 408-421 in L. Nielsen, J.C. Haigh, and M. E. Fowler, eds. Chemical immobilization of North American wildlife. Wisconsin Humane Soc., Milwaukee. 5 �149 Table 1. Mean weight (kg) , total body length (TBL) (cm) , and left hind foot length (LHFL) (cm) of mule deer fawns captured with drop nets and helicopter net guns on pinyon-juniper winter range in northwest Colorado, 11 November-5 December 1992. Method Sex n Variable Drop net M 45 Weight 45 TBL 45 LHFL 40.99 1.66 34.9 44.5 41 Weight 29.18 4.09 20.0 35.3 41 TBL 41 LHFL 40.00 1.65 35.1 42.2 86 Weight 30.45 4.59 18.7 42.9 86 TBL 86 LHFL 40.51 1. 72 34.9 44.5 43 Weight 31.34 4.12 22.6 39.2 43 TBL 42 LHFL 41.25 1.47 38.3 44.4 36 Weight 28.97 3.16 18.6 33.4 36 TBL 36 LHFL 39.98 1.12 38.1 42.9 79 Weight 30.26 3.88 18.6 39.2 79 TBL 78 LHFL F Total Helicopter M net gun F Total R 31.61 126.6 124.0 125.3 129.6 126.7 128.3 40.66 6 SD 4.76 7.42 8.04 7.78 7.04 5.39 6.47 1.46 Minimum 18.7 108 102 102 117 113 113 38.1 Maximum 42.9 143 138 143 144 135 144 44.4 �150 Table 2. comparative helicopter costs for capturing mule deer fawns with drop nets and net guns. capture Source Drop net method Helicopter net gun Non-monetary Person-days Vehicle Monetary miles Subsistence (@ $0.123/mile) Helicopter Other (bait, etc.) 50 33,600 1,500 3,200 100 615 6 19,750 1,100 38,535 of fawns captured Cost/fawn 5,000 contract Total Number 6 ($) Personnel Vehicle 325 ($) 21,356 86 79 448 270 7 �151 Colorado Division Wildlife Research July 1993 of Wildlife Report JOB PROGRESS state of Project REPORT Colorado No. W-153-R-4 Mammals Research Work Plan No. lA Multispecies Job No. 6 Monitoring and Managing in Colorado Period Covered: Investigations Wildlife Health July 1, 1992 - June 30, 1993 Authors: M. W. Miller, M. L. Stevens, W. J. Adrian, Williams, C. W. McCarty, and D. M. Getzy T. R. Spraker, E. S. Personnel: J. Bredehoft, G. Byrne, A. Case, D. Clarkson, M. Cousins, B. Davies, H. Dietrick, L. Evans, R. Forde, D. Freddy, T. Fulk, K. Green, J. Jackson, K. Kinney, M. Lamb, C. Leonard, K. Madriaga, C. W. McCarty, C. A. Mehaffy, B. Olmstead, J. Ritchie, K. Scott, H. Spear, and R. Spowart Abstract Wildlife populations throughout Colorado were monitored for occurrence of disease using a combination of extensive and intensive approaches. We continued to develop and modify a statewide surveillance program for acquiring, examining, reporting on, and summarizing sporadic wildlife disease cases occurring throughout Colorado. At least 40 carcasses and/or tissue samples representing 30 wildlife cases were submitted for diagnostic examination during July 1992-June 1993. Malnutrition/starvation (9/22), locoism (5/22), or chronic wasting disease (3/22) were the most common diagnoses in wild cervids submitted; all 4 bighorn sheep submitted showed gross and/or histologic lesions of bronchopneumonia. Among carnivore cases, canine distemper (3/4 raccoons submitted) and feline panleukopenia (1/2 bobcats and 1 mountain lion) were diagnosed. Hepatitis was diagnosed in a great blue heron. Other mammalian' .and avian cases appeared to represent isolated incidents of trauma. For a cases, cause of death could not be determined. We continued the annual statewide survey of deer and elk hunters to collect sera for brucellosis screening, and also continued modifying and evaluating our survey program. Of 11,205 deer and 4,865 elk hunters surveyed, 2,375 (15%) returned blood samples for brucellosis screening from animals harvested throughout Colorado during October 1992-January 1993. Of samples returned, only 1,228 (48%) were usable; marked hemolysis and/or contamination precluded evaluation of the remaining samples. All 795 deer and 352 elk sera tested were negative for antibodies to Brucella spp. on the standard card test. Overall, about 7% of the survey kits distributed to deer or elk hunters in 1992-1993 provided usable samples, as compared to 5% in 1991-1992. Increases in both sample return rates (15% in 1992-1993, up from 12% in 1991-1992) and proportions of returned samples that were usable (48% in 1992-1993, up from 42% in 1991-1992) appeared to contribute to improvement in the overall effectiveness of the 1992-1993 survey. �152 Three cases of chronic wasting disease (CWO), a spongiform encephalopathy, were confirmed in free-ranging deer and elk in Larimer county during FY 19921993, and 6 additional suspect cases are still pending final diagnosis; 15 free-ranging cases have been confirmed since 1985. To date, all of these cases are. from GMUs 9, 191, 19, or 20. To obtain reliable estimates for distribution and prevalence of CWO in wild cervids, we continued to survey for cwo in select deer and elk populations throughout Colorado. Brains from about 25 mule deer and 29 elk harvested in GMUs 19 and 20 (DAUs D4, DI0/E4, E9), from about 67 mule deer and 32 elk harvested on the Forbes Trinchera Ranch near Ft. Garland (DAU D31/E33), and from about 80 mule deer and elk harvested in GMUs 66 and 67 (GMU D25/E25) were collected for examination for CWO. Histologic evaluation of samples has not been completed, but all brains from hunter-killed deer and elk examined to date have been negative for spongiform encephalopathy. Retropharyngeal and other cranial lymph nodes and tonsils from about 80 mule deer and elk harvested in GMUs 66 and 67 (GMU D25/E25), from about 67 mule deer and 32 elk harvested on the Forbes Trinchera Ranch (DAU D3l/E33), and from about 25 mule deer and 29 elk harvested in GMUs 19 and 20 (DAUs D4 ,DI0/E4, E9) were examined for gross lesions of bovine tuberculosis and collected for histologic evaluation and culture. We occasionally observed tonsillar and/or retropharyngeal cysts, abscesses, and/or granulomas in samples from all 3 areas (17 cases total). Histologic evaluations have not been completed, but no microscopic lesions compatible with bovine tuberculosis have been observed in samples examined to date. We began developing a generalized, stochastic, individual-based simulation model of infectious disease in wild ungulate populations. Initially, we incorporated parameters to simulate introduction of bovine tuberculosis into a wild elk population and examined preliminary results of SO-year simulations (n = SOD) where 2 infected elk were introduced into a population of SOD wild elk. Transmission coefficient (tc) assumptions markedly influenced outcomes: assuming tc = 0.3 new infections/infected individual/year, the probability that tuberculosis became established in simulated populations was about 0.2, and prevalence in infected populations averaged about 0.03; assuming a slightly higher tc (O.S new infections/infected individual/year), the probability that tuberculosis became established increased to about 0.6, and mean prevalence in infected populations reached about 0.7. Our preliminary results suggest introduction of bovine tuberculosis into wild elk populations could represent a significant obstacle to national eradication goals. �153 MONITORING AND MANAGING WILDLIFE HEALTH IN COLORADO M. W. Miller M. L. stevens W. J. Adrian T. R. Spraker E. S. Williams C. W. McCarty and D. M. Getzy P. N. OBJECTIVES Develop and implement a program for enhancing statewide efforts manage health of Colorado's terrestrial wildlife populations. AGREEMENT to monitor OBJECTIVES 1. Modify and improve systems for submitting, diagnosing and reporting sporadic disease cases in wild animals throughout Colorado. 2. Develop and use databases for assimilating and analyzing problems identified through surveillance and surveys. and managing wildlife on data on disease 3. Design, conduct, and report results of surveys for brucellosis, tuberculosis, and chronic wasting disease in specific deer and/or populations. 4. Provide assistance in investigating outbreaks in Colorado. and elk disease 5. Design experiments to develop and/or improve techniques for investigating wildlife diseases; begin conducting approved and funded research. Maintaining healthy wildlife populations is a fundamental component of sound wildlife management practices. Habitat degradation, high animal density, extreme weather, and disease can act singly or in combination to compromise the overall health of a wildlife population. As Colorado's wildlife managers, we have developed a variety of tools for monitoring and assessing the effects of habitat loss, animal numbers, and weather on wildlife populations. We have also invested considerably in developing tools to manage these factors to optimize performance of the wildlife populations in our stewardship. In contrast, monitoring and managing the effects of disease on wildlife population performance have received relatively little attention (with a few notable exceptions). This lack of attention may be rooted to some extent in a widely-held belief that wildlife diseases are symptoms of larger underlying population problems that will be resolved if those larger problems are managed properly. Despite this belief, disease can be a significant obstacle to effective and efficient wildlife management in Colorado. Disease outbreaks account for substantial mortality in some wildlife populations. Introduced pathogens have potential to decimate local wildlife populations. Some diseases depress wildlife population performance to levels below resource-based carrying capacity. Many wildlife diseases are shared with domestic animals and/or humans, and in some cases wildlife populations serve as reservoirs for these agents. Disease also detracts from the aesthetic value of wild animals, and may convey a perception of mismanagement to uninformed publics. For these �154 reasons, diseases should be regarded as an integral population dynamics and wildlife management. part of wildlife Select wildlife health problems have been monitored in Colorado for more than 30 years. These longstanding efforts have provided useful information on the diseases studied. However, because these efforts have not always been coordinated on a statewide basis, and because some findings have not been widely available to managers and policy makers, applications to overall management programs have been limited. In order to improve our collective ability to manage wildlife health in Colorado, we need a more coordinated and systematic approach for monitoring, investigating, and reporting on health problems in free-ranging wildlife. A more complete understanding of wildlife diseases and their effects on population performance is fundamental to comprehensive wildlife management. Enhanced surveillance efforts will provide a mechanism for detecting health problems throughout the state before serious impacts to wildlife populations occur. Assimilating diagnostic data will aid in assessing trends suggestive of population-level disease problems. Programs for conducting extensive and intensive surveys for potential and realized wildlife diseases will provide reliable prevalence and distribution data for managers and administrators to use in decision making. Expertise in investigating and managing epizootics and epornitics will ameliorate efficacy and efficiency of efforts to control outbreaks. Improved techniques for diagnosing and studying wildlife diseases will provide a firm foundation for health management programs designed to enhance the quality of Colorado's wildlife populations. MATERIALS Disease AND METHODS Surveillance We monitored wildlife populations throughout Colorado for occurrence of disease using a combination of extensive and intensive approaches. These were organized and conducted as follows: Statewide Surveillance We continued to develop and modify a program for acquiring, examining, reporting on, and summarizing sporadic wildlife disease cases occurring throughout Colorado. All carcass submissions were subjected to necropsy. Ancillary diagnostics, including histopathology, bacteriology, virology, serology, parasitology, and toxicology were performed at the discretion of CDOW personnel and/or the attending pathologist. Preliminary examination and/or test results were telephoned to CDOW's Wildlife Research Center Laboratory, usually within 3-5 days of completion, and a final report were usually provided within 15 business days of submission. Pertinent data from preliminary and final reports were entered into a permanent database (described below), and copies of reports were filed as well as sent to appropriate field personnel. Pertinent data, including species, age, sex, location, number affected, diagnosis, and other information (as available) were entered into a computerized database. This database was used to generate quarterly and annual wildlife morbidity and mortality reports. In addition, data are available for analysis of long-term trends in select wildlife disease problems. Surveys Brucellosis Survev: We continued the statewide survey of deer and elk hunters to collect sera for brucellosis screening. Over the next several years, however, we plan to continue developing and implementing strategies for expanding utility and improving efficacy and efficiency of this survey. In particular, we will focus on improving return rates on sampling kits, �155 quality of samples returned, and ability populations for surveillance. to target specific areas or We continued examining performance of the existing survey to determine average return rates and sample usability by species and season. In addition to the modifying the statewide survey, we also continued developing a process for intensively sampling specific geographic areas or populations using modified hunter surveys focused on sampling in select DAUs and/or GMUs. These surveys were constructed such that the probability of failure to detect at least 1 case of brucellosis in the selected population was ~0.1 even if herd prevalence is 1%. We will compare return rates and sample usability among seasons and collection methods, and use these comparisons to guide future survey efforts. Data from this year's survey will be analyzed in combination with those from FY 1991/1992 and FY 1993/1994 to compare sampling strategies. Results of survey modifications will be reported in future annual Job Progress Reports. We mailed about 11,205 blood sampling kits to deer hunters and 4,865 kits to elk hunters in selected GMUs to gather samples for CDOW's annual brucellosis surveillance program conducted in cooperation with the Colorado Department of Agriculture's State/Federal Brucellosis Laboratory in Denver. Kits went to sportsmen with antlerless or either-sex deer permits for early, second, regular, and late seasons in mountain and plains GMUs statewide. In addition, second and third regular and late season hunters with antlerless elk permits in DAU E2, E25, Ell, E27, E28, and E33 received kits as part of an intensive sampling effort in the vicinity of a recent bovine brucellosis and tuberculosis outbreaks. Returned samples were identified by GMU of harvest. Usable samples were centrifuged, and sera were tested for antibodies to Brucella spp. using a standard card test. Unused sera were banked and stored at -20 C for future use. Chronic Wasting Disease Survey: To obtain reliable estimates for distribution and prevalence of CWO in wild cervids, we continued to survey for CWO in select deer and elk populations throughout Colorado. Brains from mule deer and elk harvested in various seasons during October 1992-January 1993 in GMUs 19 and 20 (DAUs D4 ,D10/E4, E9), in GMUs 66 and 67 (GMU D25/E25), and on the Forbes Trinchera Ranch near Ft. Garland (DAU D31/E33) were collected for examination for CWO. Brains from hunter harvested mule deer and elk were collected, usually within 12 hours of death, and fixed in 10% buffered formalin contained in 4 L plastic bags for at least 3 months. Sections of medulla at the obex and frontal portion of the brain including basal ganglia, olfactory cortex and tract, and some frontal cortex were processed routinely for paraffin embedment. Histologic sections were cut at 5-6 ~m, stained with hematoxylin and eosin, and examined under a light microscope. In addition to formal surveys, we continued to encourage increased surveillance efforts by field personnel statewide and submission of carcasses from deer or elk showing clinical signs resembling CWO. Bovine Tuberculosis Survey: Bovine tuberculosis was diagnosed in captive elk held on a game ranch near Powderhorn, CO in June 1991. Since that time, we have continued to investigate the possibility that tuberculosis might have spread to free-ranging wildlife outside the infected premises. Retropharyngeal and other cranial lymph nodes and tonsils from mule deer and elk harvested in GMUs 66 and 67 (GMU D25/E25), on the Forbes Trinchera Ranch (DAU D31/E33), and in GMUs 19 and 20 (DAUs D4, D10/E4, E9) were examined for gross lesions of bovine tuberculosis and collected for histologic evaluation and culture. Subsamples of parotid, mandibular, and retropharyngeal lymph nodes and tonsils, as available, were preserved in 10% buffered formalin and frozen and submitted to the Wyoming State Veterinary Laboratory in Laramie for histologic examination (and culturing, when warranted). When possible, �156 eviscerated carcasses tuberculosis. Disease were also examined for gross lesions suggestive of Investigations No significant 1993. Experimental disease outbreaks were investigated during July 1992-June Approaches We began developing a generalized, stochastic, individual-based simulation model of infectious disease in wild ungulate populations (Fig. 1). We plan to use this model in predicting consequences of disease introductions, improving understanding of the epizootiology of select disease problems, and evaluating potential disease management strategies. In this model, populations display density-dependent sigmoid growth in the absence of disease or other limiting processes. We employed a novel mathematical approach for estimating pathogen transmission within simulated populations, and assumed transmission probabilities are a function of prevalence. Initially, we incorporated parameters to simulate introduction of bovine tuberculosis into a wild elk population and examined probable consequences of such introductions. As a preliminary step, we examined results of replicated SO-year simulations (n = SaO/parameter set) where 2 infected elk were introduced into a population of 500 wild elk. Our model incorporated population parameters estimated from a lightly hunted elk population (Forbes Trinchera Ranch). We assumed a constant cow-calf transmission rate (0.95) and a 2-year incubation period before newly infected animals became infectious. We then made replicated simulations, varying transmission coefficient (tc = 0.3 or 0.5 new infections/infected individual/year) to assess the influence of transmission on potential outcome of tuberculosis introductions. RESULTS Disease AND DISCUSSION Surveillance Statewide Surveillance At least 40 carcasses and/or tissue samples representing 30 wildlife cases were submitted for diagnostic examination during July 1992-June 1993 (Table 1). Malnutrition/starvation (9/22), locoism (5/22), or chronic wasting disease (3/22) were the most common diagnoses in wild cervids submitted; all 4 bighorn sheep submitted showed gross and/or histologic lesions of bronchopneumonia. Among carnivore cases, canine distemper (3/4 raccoons submitted) and feline panleukopenia (1/2 bobcats and 1 mountain lion) were diagnosed. Hepatitis was diagnosed in a great blue heron. Other mammalian and avian cases appeared to represent isolated incidents of trauma. For 8 cases, cause of death could not be determined. We will continue adding new accessions throughout the coming fiscal year to our computerized database for diagnostic case information, as well as data from archived reports as they become available. Surveys Brucellosis Survev: Of 11,205 deer and 4,865 elk hunters surveyed, 2,375 (15%) returned blood samples for brucellosis screening. Of samples returned, only 1,228 (48%) were usable; marked hemolysis and/or contamination precluded evaluation of the remaining samples. All 795 deer and 352 elk sera tested were negative for antibodies to Brucella spp. on the standard card test. �157 Overall, about 7% of the survey kits distributed to deer or elk hunters provided samples usable in this year's brucellosis survey, as compared to 5% in 1991-1992. Increases in both sample return rates (15% in 1992-1993, up from 12% in 1991-1992) and proportions of returned samples that were usable (48% in 1992-1993, up from 42% in 1991-1992) appeared to contribute to improvement in the overall effectiveness of the 1992-1993 survey. These data, combined with those collected during FY 1991/1992 and in FY 1993/1994, will be used in assessing strategies for improving the efficiency of statewide serologic surveys that depend on blood samples submitted from harvested animals. Chronic Wasting Disease Survev: Three cases of chronic wasting disease (CWO), a spongiform encephalopathy, were confirmed in free-ranging deer and elk in Larimer county during FY 1992-1993, and 6 additional suspect cases are still pending final diagnosis; 15 free-ranging cases have been confirmed since 1985. To date, all of these cases are from GMUs 9, 191, 19, or 20. To obtain reliable estimates for distribution and prevalence of CWO in wild cervids, we continued to survey for CWO in select deer and elk populations throughout Colorado. Brains from about 25 mule deer and 29 elk harvested in GMUs 19 and 20 (DAUs D4 ,D10/E4, E9), from about 67 mule deer and 32 elk harvested on the Forbes Trinchera Ranch near Ft. Garland (DAU D31/E33), and from about 80 mule deer and elk harvested in GMUs 66 and 67 (GMU D25/E25) were collected for examination for CWO. Histologic evaluation of samples has not been completed, but all brains from hunter-killed deer and elk examined to date have been negative for spongiform encephalopathy. Bovine Tuberculosis Survey: Retropharyngeal and other cranial lymph nodes and tonsils from about 80 mule deer and elk harvested in GMUs 66 and 67 (GMU D25/E25), from about 67 mule deer and 32 elk harvested on the Forbes Trinchera Ranch (DAU D31/E33), and from about 25 mule deer and 29 elk harvested in GMUs 19 and 20 (DAUs D4, D10/E4, E9) were examined for gross lesions of bovine tuberculosis and collected for histologic evaluation. We occasionally observed tonsillar and/or retropharyngeal cysts, abscesses, and/or granulomas in samples from all 3 areas (17 cases total). Histologic evaluations have not been completed, but no microscopic lesions compatible with bovine tuberculosis have been observed in samples examined to date. We plan to further develop and refine ongoing surveillance programs for both tuberculosis and CWO during FY 1993-1994. Disease Investigations No significant 1993. Experimental disease outbreaks were investigated during July 1992-June Approaches In examining preliminary results of 500 50-year simulations where 2 infected elk were introduced into a population of 500 wild elk, transmission coefficient (tc) assumptions markedly influenced outcomes. Under conservative assumptions (tc = 0.3 new infections/infected individual/year), the probability that tuberculosis became established (i.e., infection still present 50 years after initial introduction) in simulated populations was about 0.2 (Fig. 2), and prevalence in infected populations averaged about 0.03 (Fig. 3). Using a slightly higher tc (0.5 new infections/infected individual/year), the probability that tuberculosis became established increased to about 0.6 (Fig. 2), and mean prevalence in infected populations reached about 0.7 (Fig. 3). Our preliminary results suggest introduction of bovine tuberculosis into wild elk populations could represent a significant obstacle to national eradication goals. We plan to further refine parameter estimates for elk-tuberculQsis simulations, and to explore application of this modeling approach to other real and potential disease problems affecting wild ungulate populations. �158 Acknowledgments The statewide wildlife health monitoring and surveillance program described above relies heavily on efforts of dedicated field personnel throughout the Colorado Division of Wildlife, and truly represents a division-wide effort to improve our understanding and management of wildlife disease problems. In addition to those specifically listed, we collectively thank all of those regional and area biologists, district and area wildlife managers, and others who assisted by submitting diagnostic cases throughout the year. In particular, we thank personnel from areas 2, 4, 10, and 16, and from the Forbes Trinchera Ranch for assistance and logistical support in tuberculosis and CWO surveys and surveillance activities, and personnel from the stateFederal Cooperative Brucellos' Laboratory for their continued cooperation, assistance and logistical ort in conducting annual brucellosis surveys. Prepared by Researcher �159 Table 1. Summary of wildlife July 1992 -June 1993. REGION SPECIES HE diagnostic cases submitted for diagnosis CAUSE during OF DEATH SEX AGE Mule Deer Mule Deer Mule Deer Mule Deer Mule Deer Mule Deer Elk Bobcat Bobcat Raccoon M F M M M A A I M I I A Raccoon Red Fox Heron Pelican (2) Blackbirds F A I F A RM BHS RM BHS Raccoon Raccoon M M A RM BHS Blackbirds M Y Pneumonia Undetermined Elk RM BHS M A Locoweed Toxicity Pneumonia Undetermined cwo Undetermined CWO CWO Malnutrition Malnutrition Trauma/Starvation Feline Panleukopenia Gunshot-induced septicemia SE CE Y SW Elk (7 ) Hawk M/F I/A Unknown Elk Elk Elk Elk (4) Mt. Lion F M M M A A I A I Canine Distemper Undetermined Hepatitis Gunshot-induced trauma Undetermined Pneumonia Pneumonia Canine Distemper Canine Distemper Starvation Undetermined septicemia Undetermined Undetermined Locoweed Toxicity Feline Panleukopenia ��161 Colorado Division Wildlife Research July 1993 of Wildlife Report JOB PROGRESS state of Project REPORT Colorado W-153-R-4 Mammals Research Work Plan No. 2A Mountain Sheep Job No. 4 Strategies for Managing Infectious Disease in Mountain Sheep Populations Period No. Covered: July Investigations 1, 1992 - June 30, 1993 Authors: M. W. Miller, M. A. Wild, B. J. Kraabel, Personnel: K. P. Snipes, E. S. Williams, R. Silflow, and T. R. Spraker and T. R. Anderson A. Boeger-Fields, K. W. Mills, ABSTRACT Examination of new and archived Pasteurella haemolytica isolates from 8 indigenous wild bighorn herds continued. Using ribosomal RNA gene restriction patterns, at least 11 distinct genotypes (A, B, I, J, K, L, M, N, 0, P, and T) of P. haemolytica have been identified among isolates (n= 17) from 3 herds (Avalanche Creek, Tarryall, Taylor River) analyzed to date. Ribotypes varied both within and among the 5 distinct phenotypes (T3; 4; 3,4; 3,4,10; and untypable) of P. haemolytica identified among these 3 herds using rapid plate agglutination. In contrast to phenotypes, ribotypes of P. haemolytica appeared unique for each of these geographically isolated herds. Preliminary in vitro evaluation of potency of cytotoxins derived from genotypicallydistinct P. haemolytica isolates (A, B, I, K, 0, T) revealed further differences: cytotoxin from one Taylor River isolate (0) collected 2 months before a recent pasteurellosis epizootic showed greater potency (>50% neutrophil death at 50pg) than cytotoxin from other genotypes examined to date (all S30% neutrophil death at 50pg). Sera from 5 indigenous herds (Almont/Taylor River, Chalk Creek, Cottonwood Creek, Tarryall/Cotton Gordon's, and Tarryall/Sugarloaf Mountain) have also been screened for antibodies to a P. haemolytica T4 strain (WSVL Antigen 14) using an enzyme-linked immunosorbent assay (ELISA). Mean optical density (OD) readings (indicating relative levels of antibody to that strain) were comparable among 4 of 5 herds, but appeared somewhat lower for the Almont/Taylor River herd •. The latter herd suffered a: pasteurellosis outbreak and minor die-off during MarchApril 1991, about 2 months after capture. Whether relatively low antibody levels observed are indicative of herd susceptibility to pasteurellosis or are coincidental warrants further investigation. For the fifth consecutive summer, we made observations on a pasteurellosis epizootic in captive bighorn lambs. Six healthy lambs were born to captive ewes between mid-May and mid-June 1992; 2 additional lambs died within 24 hrs of birth, 1 from systemic pasteurellosis. We first observed nasal discharge in 1 of 4 7-week-old lambs on 7 July. By 10 July, 4 of 5 lambs pastured together had developed pneumonia. Consequently, pneumonic lambs were housed in isolation pens with their dams and treated with amoxicillin and gentamicin for 5 days followed by long-acting oxytetracycline every other day. Respiratory disease in the fifth lamb never progressed beyond nasal discharge. A sixth lamb, housed in an isolation pen with her dam, remained clinically normal until 26 July, about 2 weeks after she was pastured with other lambs. She showed occasional nasal discharge thereafter. Pneumonic lambs initially �162 appeared to respond to antibiotic therapy, but they did not recover completely despite over 2 weeks of daily or alternate-day treatment. Because their pneumonia seemed somewhat refractory to the antibiotic regime prescribed, we conducted a small clinical trial to evaluate a new long-acting antibiotic. Beginning on 30 July, the 2 most severely affected lambs were treated with tilomicosin (Micotil®; 10mg/kg injected subcutaneously (SC) every third day); the other 2 received long-acting oxytetracycline (Liquamycin®, LA200®; 10mg/kg SC every third day). Within 4 days (2 treatments), 2 blinded observers independently noted that lambs treated with tilomicosin showed greatest clinical improvement. We subsequently began treating all 4 lambs with tilomicosin and vitamin E. After about 10 days, lambs appeared healthy enough to return to pasture, although periodic nasal discharge and coughing persisted in all 4 for 2-6 months. We conducted a second experiment evaluating utility of modified Cary and Blair medium (MCB; Port-A-Cul® transport tubes) for improving recovery of P. haemolytica from bighorn pharyngeal swabs. We collected 4 pharyngeal swabs from each of 10 healthy captive bighorns and stored them in MCB transport tubes. One tube from each sheep was cultured within a few hours of collection; others were stored for 24, 48, or 72 hrs at 5 C. Nonhemolytic P. haemolytica was recovered from all 40 swabs in relatively heavy growth, regardless of the duration of refrigeration. Our findings suggest recovery of P. haemolytica from field studies may be improved dramatically by refrigerating bighorn pharyngeal swabs stored in MCB transport tubes. Since incorporating Port-A-Cul® tubes into recommended procedures for sampling bighorns, isolation rates for P. haemolytica among Colorado's wild sheep herds have exceeded rates in previous years when pharyngeal (1989-90) or nasal (1988-89) swabs were transported in modified Amies with charcoal, suggesting previous sampling efforts probably underestimated prevalence of P. haemolytica infections in free-ranging herds in Colorado and elsewhere. Refrigerating Port-A-Cul® tubes between sample collection and processing for intervals ~72 hr appears to be a practical way to optimize recovery of P. haemolytica from bighorn pharyngeal swabs. Following such procedures should provide more reliable prevalence estimates for studying epizootiology and management of pasteurellosis in bighorns. Preliminary evaluation of additional data from ongoing monthly collections of feces from captive bighorns supported previous observations that reproductive status may influence fecal cortisol excretion in bighorn ewes; by May, fecal cortisol concentrations from pregnant ewes (mean±SE=39.S±6.9 ng/g dm) were about 77% higher than those from open ewes (22.2±3.0 ng/g dm)(P<O.OS). Evaluation of seasonal influences on fecal cortisol measurements will continue through September 1992 to provide 2 complete years of data. �163 EXPERIMENTS TO IDENTIFY AND MANAGE STRESS IN MOUNTAIN SHEEP POPULATIONS M. W. Miller, M. A. Wild, B. J. Kraabel, and T. R. Anderson P. N. OBJECTIVE To develop population strategies for managing performance. infectious SEGMENT diseases affecting bighorn sheep OBJECTIVES 1. Compare rates for tonsillar carriage and nasal shedding of Pasteurella spp. among free-ranging bighorn populations; compare serum antibody titers to Pasteurella spp. among populations; use phenotypic and genotypic characterizations to compare Pasteurella spp. isolates from different bighorn populations. 2. Use computer simulation managing pasteurellosis 3. Monitor sheep. MANAGEMENT seasonal OF BACTERIAL changes modeling to examine alternative in bighorn sheep populations. in fecal cortisol AND VIRAL DISEASES excretion IN MOUNTAIN strategies in captive SHEEP for bighorn POPULATIONS Inability to control infectious disease outbreaks and subsequent mortality in mountain sheep populations represents a significant obstacle to long-term success in their management. Although the "bighorn pneumonia complex" has been studied intensively for over 3 decades, little is known about many aspects of its etiology and epizootiology. Moreover, management interventions recommended for preventing or controlling this problem remain untested. Most previous efforts to improve understanding and management of the epizootiology of pneumonia in bighorns involved post hoc investigations of dieoffs occurring in free-ranging sheep herds. These studies identified various etiological agents associated with known mortalities and attempted to determine predisposing causes and population consequences of individual outbreaks. From these investigations, comparisons of real or perceived patterns became the basis for hypotheses on the epizootiology of pneumonia in bighorns. Recognition of similar patterns in other outbreaks served as evidence supporting these as unifying hypotheses. Unfortunately, several of these hypotheses have failed to withstand rigorous experimental testing. And, despite our best management efforts, bighorns continue to die. Our strategy for developing a better understanding of the epizootiology and management of bacterial and viral diseases in bighorn populations differs -generally, we propose to take an adaptive environmental assessment approach for studying the bighorn pneumonia complex. As a foundation for our research strategy, we have assimilated existing knowledge on bighorn population dynamics (including the epizootiology and consequences of infectious disease) into a computer simulation model (Hobbs and Miller 1991). Because pasteurellosis appears to underlie virtually all respiratory disease problems reported for bighorns, our modeling efforts have focused on the epizootiology of pasteurellosis in sheep populations. We have constructed a model that reflects dynamics of bighorn populations seen in nature using the simplest assumptions necessary to reproduce those behaviors. We plan to conduct �164 simulation experiments to identify variables that might be particularly sensitive to management perturbations in altering the dynamics of disease bighorn populations. Those results will serve as the basis for designing management level experiments in the future. in In parallel with our modeling efforts, we are conducting a series of experiments to develop, improve and standardize methods for collecting and interpreting diagnostic data to provide better estimates of key parameters driving our models. In particular, we have been developing tools for identifying strains of Pasteurella haemolytica and quantifying immunological responses of bighorns to infection by these pathogens. These tools will be key components of laboratory and field experiments designed to evaluate potential tactics (including vaccination and/or treatment) for managing pasteurellosis in wild sheep, and appear prerequisite to initiating management level experiments. To this end, our recent efforts have focused on both simulation modeling and on improving tools available for use in future management experiments that will be designed to study etiology, epizootiology, and prevention or control of disease outbreaks in bighorn popUlations: METHODS Management of Bacterial and Viral AND MATERIALS Diseases in Mountain Sheep Populations In conjunction with numerous cooperators, we continued developing and improving tools available for use in studying etiology, epizootiology, prevention or control of disease outbreaks in bighorn popUlations: and Epizootiology of pasteurellosis in indigenous bighorn populations (Miller, Spraker, Mills, Snipes, and Kraabel): Examination of new and archived Pasteurella haemolytica isolates from 8 indigenous wild bighorn herds (Almont/Taylor River, Avalanche Creek, Chalk Creek, Cottonwood Creek, Grant, Tarryall, Texas Creek, Waterton Canyon) continued. Genomic fingerprinting of remaining untyped isolates (n ~ 50) is in progress. We also began evaluating potency of cytotoxins derived from genotypically-distinct P. haemolytica isolates in vitro using methods described by Silflow et ale (1993). In addition, sera from 5 indigenous herds (Almont/Taylor River, Chalk Creek, Cottonwood Creek, Tarryall/Cotton Gordon's, and Tarryall/Sugarloaf Mountain) were screened for antibodies to a P. haemolytica T4 strain (WSVL Antigen 14) using an enzyme-linked immunosorbent assay (ELISA). Epizootiology of pasteurellosis in bighorn lambs (Miller, Wild, and Kraabel): For the fifth consecutive summer, we made observations on a pasteurellosis epizootic in captive bighorn lambs. Six healthy lambs were born to captive ewes between mid-May and mid-June; 2 additional lambs died within 24 hrs of birth, 1 from systemic pasteurellosis. We observed lambs daily for premonitory signs of respiratory disease. Pneumonic lambs were housed in isolation pens with their dams and treated with amoxicillin and gentamicin for 5 days followed by long-acting oxytetracycline every other day as described previously. Because lambs did not recover completely under this treatment regime, we conducted a small clinical trial to evaluate a new long-acting antibiotic developed for treating pasteurellosis in cattle. Beginning on 30 July, the 2 most severely affected lambs (L92, Q92) were treated with tilomicosin (Micotil®; 10mg/kg injected subcutaneously (SC) every third day); the other 2 received long-acting oxytetracycline (Liquamycin®, LA200®; 10mg/kg SC every third day). Two blinded observers independently observed all lambs twice each day and noted their clinical condition. Improving Recovery of P. haemolytica From Bighorn Pharyngeal Swabs Usina Modified Cary and Blair Medium (Wild and Miller): We conducted a second experiment evaluating the utility of modified Cary and Blair medium (MCB; Port-A-Cul® transport tubes) for improving recovery of P. haemolytica from bighorn pharyngeal swabs. In an earlier study, we compared recovery rates from MCB transport tubes with rates from direct streaking onto blood agar. We �165 collected 3 pharyngeal swabs from each of 25 healthy captive bighorns and recovered nonhemolytic P. haemolytica from 23 of 25 (92%) swabs streaked onto blood agar plates and incubated immediately, from 16 of 25 (64%) swabs held in MCB transport tubes for 24 hr, and from 1 of 25 (4%) swabs held in MCB tubes for 48 hr. Although the recovery rate from swabs held in MCB tubes for 24 hr was only about 70% of that from direct swabs, rates were markedly higher than those previously observed for other transport media (Wild and Miller 1991). Based on our findings, Port-A-Cul® transport tubes were used in sampling wild bighorns during several field studies in 1991-1992. Of field samples received, we observed that a high proportion of swabs from one set inadvertently chilled during transport yielded P. haemolytica isolates 36-48 hrs after collection. Refrigerating other sets of field samples collected subsequently produced similar results. These observations, being contrary to manufacturer's recommendations and conventional wisdom, inspired our most recent experiment. We collected 4 pharyngeal swabs from each of 10 healthy captive bighorns and stored them in Port-A-Cul® transport tubes. One tube from each sheep was cultured within a few hours of collection, the others were stored for 24, 48, or 72 hrs at 5 C. Use computer simulation model ina to examine alternative strategies for managing pasteurellosis in bighorn sheep populations (Miller and Hobbs): Inability to access and manipulate the bighorn sheep/pasteurellosis model described previously precluded progress on using computer simulation modeling to examine alternative strategies for managing pasteurellosis in bighorn sheep populations. Experiments on Measuring and Managing Stress in Bighorns (Miller and Anderson): In planning field studies of stress responses in wild bighorns, it has become apparent that further development and understanding of fecal cortisol measurements is necessary in order to credibly apply these techniques to free-ranging populations. In order to anticipate potential confounding influences on field applications of fecal cortisol measurements, we have collected feces from captive bighorns at monthly intervals to examine seasonal influences on cortisol excretion. Beginning in October 1990, samples were collected over a 3-5 day period at the beginning of each month from individual captive bighorns of varied age/sex classes. Feces were stored at -20 C until processed for cortisol determination. Methods for processing feces and measuring cortisol were as described by Miller et al. (1991). We continued collecting feces from captive bighorns at monthly intervals to examine seasonal influences on cortisol excretion. Our 2 year study ended with October's collections. Processed samples of feces from captive bighorns collected at monthly intervals from March-September 1992 were submitted for cortisol assay in late October. Samples from adult ewes were subsequently analyzed for progesterone using a validated radioimmunoassay (RIA) to evaluate utility and reliability of this approach for detecting pregnancy in bighorns. RESULTS Management of Bacterial and Viral AND DISCUSSION Diseases in Mountain Sheep Populations Epizootiology of pasteurellosis in indiaenous bighorn populations: Using ribosomal RNA gene restriction patterns, at least 11 distinct genotypes (A, B, I, J, K, L, M, N, 0, P, and T) of P. haemolytica have been identified among isolates (n= 17) from 3 herds (Avalanche Creek, Tarryall, Taylor River) analyzed to date. Ribotypes varied both within and among the 5 distinct phenotypes (T3; 4; 3,4; 3,4,10; and untypable) of P. haemolytica identified among these 3 herds using rapid plate agglutination. In contrast to phenotypes, ribotypes of P. haemolytica appeared unique for each of these geographically isolated herds. Preliminary in vitro evaluation of potency of cytotoxins derived from genotypically-distinct P. haemolytica isolates (A, B, I, K, 0, T) revealed further differences: cytotoxin from one Taylor River �166 isolate (0) collected 2 months before a recent pasteurellosis epizootic showed greater potency (>50% neutrophil death at 50~g) than cytotoxin from other genotypes examined to date (all ~30% neutrophil death at 50~g). Analyzing sera from 5 indigenous herds for antibodies to a P. haemolytica T4 strain (WSVL Antigen 14) revealed that mean optical density (00) readings (indicating relative levels of antibody to that strain) were comparable among 4 of 5 herds, but appeared somewhat lower for the Almont/Taylor River herd. The latter herd suffered a pasteurellosis outbreak and minor die-off during MarchApril 1991, about 2 months after capture. Whether the relatively low antibody levels we observed are indicative of herd susceptibility to pasteurellosis or are coincidental warrants further investigation. Epizootiology of pasteurellosis in bighorn lambs: We first observed nasal discharge in 1 of 4 7-week-old lambs (L92) on 7 July. By 10 July, 4 of 5 lambs pastured together had developed pneumonia. Respiratory disease in the fifth lamb (M92) never progressed beyond nasal discharge. A sixth lamb, housed in an isolation pen with her dam, remained clinically normal until 26 July, about 2 weeks after she was pastured with other lambs. She showed occasional nasal discharge thereafter. Neither of the latter 2 lambs developed clinical pneumonia. The 4 pneumonic lambs initially appeared to respond to antibiotic therapy, but they did not recover completely despite over 2 weeks of daily or alternate-day treatment. Within 4 days (2 treatments) of initiating tilomicosin therapy, both blinded observers independently noted that lambs treated with tilomicosin showed greatest clinical improvement. We subsequently began treating all 4 lambs with tilomicosin and vitamin E. After about 10 days, lambs appeared healthy enough to return to pasture, although periodic nasal discharge and coughing have persisted in all. Imorovina Recoverv of P. haemolytica From Bighorn Pharyngeal Swabs Using Modified Carv and Blair Medium: Nonhemolytic P. haemolytica was recovered from all 40 swabs in relatively heavy growth, regardless of the duration of refrigeration in MCB medium (Fig. 1). These rather surprising findings suggest recovery of P. haemolytica from field studies may be improved dramatically by refrigerating bighorn pharyngeal swabs stored in Port-A-Cul® transport tubes. Since incorporating Port-A-Cul® tubes into recommended procedures for sampling bighorns, isolation rates for P. haemolytica among Colorado's wild sheep herds have exceeded rates in previous years when pharyngeal (1989-90) or nasal (1988-89) swabs were transported in modified Amies with charcoal. This suggests previous sampling efforts probably underestimated prevalence of P. haemolytica infections in free-ranging herds in Colorado and elsewhere. Refrigerating Port-A-Cul® tubes between sample collection and processing for intervals ~72 hr appears to be a practical way to optimize recovery of P. haemolytica from bighorn pharyngeal swabs. Following such procedures should provide more reliable prevalence estimates for studying epizootiology and management of pasteurellosis in bighorns. A manuscript summarizing accepted for publication findings from both of our Port-A-Cul® in the Journal of Wildlife Diseases. experiments Use computer simulation modeling to examine alternative strategies for managing pasteurellosis in bighorn sheep popUlations: Inability to access manipulate the bighorn sheep/pasteurellosis model described previously precluded progress on using computer simulation modeling to examine alternative strategies for managing pasteurellosis in bighorn sheep populations. Experiments on Measuring and Managing Stress was and in Bighorns Preliminary evaluation of data from monthly collections of feces from captive bighorns suggested reproductive status may influence fecal cortisol excretion �167 in bighorn ewes (Fig. 2); by May, fecal cortisol concentrations from pregnant ewes (mean±SE=39.5±6.9 ng/g dm) were about 77% higher than those from open ewes (22.2±3.0 ng/g dm)(P<O.05). Based on these findings, further evaluation of seasonal and other influences on fecal cortisol measurements is warranted before these techniques are applied to studies of stress responses in freeranging bighorns. LITERATURE CITED Snipes, K. P., R. W. Kasten, M. A. Wild, M. W. Miller, D. A. Jessup, R. L. Silflow, W. J. Foreyt, and T. E. Carpenter. 1991. Using ribosomal RNA gene restriction patterns in distinguishing isolates of Pasteurella haemolytica from bighorn sheep (Ovis canadensis). J. Wildl. Dis. 28: 347354. Prepared by Wildlife Researcher ��169 Colorado Division Wildlife Research July 1993 of Wildlife Report JOB PROGRESS Colorado State of Project Work No. Job No. Covered: Authors: Personnel: Mammals Research 2A Mountain Sheep 7 Experimental Evaluation of Mountain Transplanting and Disease Treatment W-153-R-4 Plan No. Period REPORT Investigations Sheep July 1, 1992 - June 30, 1993 M. W. Miller and J. Vayhinger K. Alderman, J. Backstrand, R. Dobson, R. Green, V. Jurgens, R. Hancock, D. Lovell, K. McLaughlin, S. Ogi+vie, G. Roberts, S. Roush, A. Stencel, A. Torres, T. Verry, C. Yoder, R. Zaccagnini. ABSTRACT We continued monitoring lamb survival among radiocollared ewes from 4 freeranging bighorn herds as part of a 4-year management experiment to examine effects of alternative lungworm treatment strategies on lamb survival and population performance. Since December 1991, 2 bighorn herds in the Tarryall Mountains [Twin Eagles (TE) and Sugarloaf Mountain (SL)] and 2 in the Collegiate Peaks [Chalk Creek (CH) and Cottonwood Creek (CW)] have been managed under 1 of 4 alternative lungworm treatment regimes: baiting with alfalfa hay and apple pulp treated with fenbendazole, baiting with alfalfa hay and apple pulp without fenbendazole, placing fenbendazole-treated salt blocks on bait stations, and withholding bait and fenbendazole. Treatments have been rotated annually under a predetermined, randomly-selected schedule. We monitored lamb production and survival among radiocollared ewes in each herd from May through October 1992 to complete the first field season. Year 2 started in midDecember 1992 with baiting and treated salt block distribution at scheduled sites. We began monitoring lamb production and survival among radiocollared ewes in each herd to assess year 2 treatment responses in May 1993. Because this study will not be completed until October 1995, all data presented are preliminary and largely descriptive. Relatively consistent and predictable range use and movement patterns for each of the 4 study herds have.emerged since monitoring began in 1991. Plots of May-October 1992 location data (UTM coordinates) for·radiocollared ewes revealed apparent differences in distribution and movement patterns among herds. Ewes from the CW herd showed widest distribution and greatest movements; CH ewes were the most limited in their range use and movement. Disturbances by hikers (CW) and hunters (CH, SL) influenced movements of ewe/lamb groups on occasion, and probably accounted for erratic movements sometimes observed in these herds. As many as 72 sheep fed at the SL bait site and as many as 34 sheep fed at the CH bait site during December 1992-February 1993. All marked ewes fed at least 14 days. Marked ewes at SL (n = 14) averaged 38 days on bait (se = 0.5; missing data for 6 days) and marked ewes at CH (n = 15) averaged 51 days on bait (se = 2.8; no missing data). Visitors harassment of sheep at the SL site apparently disrupted bait site attendance on occasion. Attendance at CH was somewhat higher and more consistent than in 1992 [mean(se) = 42(±3.8) days). In addition to bait, 14 of 15 marked ewes at CH also received fenbendazole at least once and 13 received 2 fenbendazole treatments, for a total of 13.5 ewe treatments. �170 Eight fenbendazole-treated blocks disappeared almost entirely at the CW site between December 1992 and May 1993; we attributed the observed 57% weight loss in a control block from the same site to dissolving effects of snowfalls in February-April. Adjusting for estimated environmental losses, about 95.7 kg of treated blocks (about 157.9 g fenbendazole) were consumed at CWo Block consumption equated to about 26.5 ewe treatments at CW in 1993, compared to 6 ewe treatments at SL in 1992; under a more moderate dosing regime, block consumption equated to about 70 ewe treatments at CW in 1993. We never observed use of blocks at CW directly, but did observe radiocollared sheep within 25 m of blocks and sheep tracks around blocks; observations and tracks also suggested elk and mule deer were probably using these blocks, thereby confounding estimates of drug delivery to bighorns. Both reduced production and mortality apparently affected recruitment (lambs/marked ewes) to 6 months among lamb cohorts monitored during May-october 1992. Lamb production, as estimated by observations of marked ewes with new (~ 2 wk old) lambs at heel, appeared to vary somewhat among the 4 herds studied. All 13 CH ewes were seen with lambs at least once last summer; in contrast, only 8 of 11 (0.73) SL ewes were ever observed with lambs at heel. Most, but not all, marked ewes in the other 2 herds also produced lambs (TE = 0.82, CW = 0.92). Although most lambs were born in May (TE, SL) and June (CH, CW), at least 1 new lamb was observed after 1 July in each herd. In addition to reproductive/perinatal losses, some apparently viable lambs disappeared in each of the experimental herds during the summer. Lamb survival to 6 months ranged from 0.91 in CW to 0.62 in CHi lamb survival rates in SL (0.88) and TE (0.79) were both relatively high and comparable to CWo Field observations suggest factors affecting lamb survival at CH may differ somewhat from those affecting the other 3 study herds: one sick and 3 coughing lambs were observed in CH during late August through early October. Although the sick lamb at CH subsequently disappeared, 3 others observed coughing survived through October. Sick or coughing lambs have not been observed in the 3 other herds studied, even though lambs have disappeared from all of those herds. Although the proportion of ewes observed with lambs and the proportion of those lambs surviving to 6 months of age varied somewhat among herds, overall recruitment of lambs through October 1992 appeared to be relatively high across all 4 herds being monitored, regardless of management treatment. Overall recruitment rates to 6 months ranged from 0.62-0.83 lambs/marked ewes. For 3 of the study herds (TE, CH, SL), the proportion of marked ewes with lambs at the end of October was indistinguishable. Other than 2 mid-summer mortalities in CW, no health problems were detected among marked ewes during May-October 1992. However, 3 Tarryall ewes were found dead in early December .1992-mid-June 1993 (SL-9). Between February 1991 and June 1993, a total of 13 of-76 radiocollared ewes from our 4 study herds have died. Four losses were probably either directly or indirectly related to capture, altho\lgh 3- of these 4 ewes survived at least 2 months after being captured and radiocollared •. The other 9 losses (4 at TE, 3 at SL, and 2 at CW) were likely not influenced by previous capture. Of these, predation may have been involved in 4 losses in the Tarryalls (3 at TE and 1 at SL), but causes of death or disappearance for 4 other ewes (1 at TE, 2 at SL, and 2 at CW) have not been determined. Overall, noncapture mortality rates of adult ewes in the 4 study herds have averaged about 0.05 annually over the past 28 months, although causes and rates of ewe mortality seem to vary somewhat among herds. Winter range counts indicate all 4 bighorn herds under study have remained stable or grown since our experiment began in 1991. Intensive monitoring will continue through October 1993 with emphasis on documenting survival of known lambs, as well as movement patterns. Additional monitoring is planned in conjunction with experimental treatments in year 3, and intensive monitoring will resume in May 1994. �171 EXPERIMENTAL EVALUATION OF MOUNTAIN SHEEP TRANSPLANTING AND DISEASE TREATMENT M. W. Miller and J. Vayhinger P. N. OBJECTIVE Design, conduct, and report on management experiments to evaluate efficacy transplanting and disease treatment practices for managing mountain sheep populations. AGREEMENT continue parasite a management level experiment control program. of OBJECTIVE evaluating Colorado's mountain sheep We continued monitoring lamb survival among radiocollared ewes from 4 freeranging bighorn herds as part of a management experiment to examine effects of alternative lungworm treatment strategies on bighorn lamb survival and population performance. Year 1 of this 4-year study ended with completion of the summer field season in October 1992; year 2 began in mid-December 1992 with baiting and treated salt block distribution at scheduled sites and will continue through October 1993. Our study will be completed in October 1995 •. MATERIALS AND METHODS Beginning in December 1991, we began managing each of 4 study herds [Tarryall Mountains: Twin Eagles (TE) and Sugarloaf (SL); Collegiate Mountains: Chalk Creek (CH) and Cottonwood Creek (CW») under 1 of 4 alternative lungworm treatment regimes: Control - no treatment -- bait and fenbendazole withheld; Treat Onlyfenbendazole-treated salt blocks placed on bait stations; Bait Only - baited with alfalfa hay and apple pulp but not treated with fenbendazole; Bait/Treatbaited with alfalfa hay and apple pulp and treated with fenbendazole. Treatments were assigned to study herds as prescribed rotating schedule (Table 1.; Year 1 = 1992). by a randomly selected, In year 1, we baited 2 herds (CH, TE) daily with alfalfa hay and apple pulp from mid-December 1991 through 21 February 1992; we baited for 64 days (ending at TE and 62 days at CH. We also treated sheep at the TE site with fenbendazole (about 3 gjadult ewe; Schmidt et ale 1979) added to apple pulp on 31 January and 6 February. Attendance of radiocollared ewes at bait sites was recorded daily. Fenbendazole-treated salt blocks (1.65 g fenbendazole/kg, 15 kg/block; n = 4 total) were available from mid-January through early May at the SL bait site; 1 block held in a wire cage was used as an environmental control. That bait site was observed daily from mid-January through mid-February and periodically thereafter to determine whether sheep were using treated salt blocks. We weighed blocks into and out of the field to estimate consumption by sheep and losses to the environment. Radiocollared ewes at CW were also observed periodically during December-April. From mid-May through October 1992, radiocollared ewes from all 4 herds were observed about once every 2 weeks to determine whether they produced lambs, and whether their lambs were still alive. In addition to lamb survival data, we �172 recorded approximate UTM coordinates, habitat type, and group size and composition for each radiocollared ewe observed. All field data were transcribed into a computerized database to aid in mapping seasonal range movements and determining annual lamb production and survival rates. We initiated treatments for year 2 (1993) in mid-December 1992 with baiting and treated salt block distribution at scheduled sites (Table 1). We baited for 64 days (ending 18 February 1993) at SL and 66 days (ending 20 February) at CHi sheep at the CH site were also treated with fenbendazole (about 3 g/adult ewe) added to apple pulp on 17 January and 17 February. Treated salt blocks (1.65 g fenbendazole/kg, 15 kg/block; 8 blocks total) were available to sheep at cw during December-May -- our initial offering of 3 blocks in December was supplemented with 3 blocks in February and 2 blocks in April. One block held in a wire cage during December-May was used as an environmental control. We radiocollared 12 additional ewes between January and March (7 at SL, 2 each at CW and CH, and 1 at TE); all but 3 were darted over bait. Radiocollared ewes were also monitored every 2-4 weeks to detect mortality and movements during February-May in conjunction with a USFS/CDOW cooperative project to identify critical winter and transitional ranges of these 4 herds. Biweekly observations of radiocollared ewes and their lambs began again on 3 May, and will continue through October. Year 3 (1994) will begin in December with baiting and treated salt block distribution at scheduled sites (Table 1). RESULTS AND DISCUSSION Range Use and Movement Patterns Relatively consistent and predictable range use and movement patterns for each of the 4 study herds have emerged since monitoring began in 1991. Plots of May-October 1992 location data (UTM coordinates) for radiocollared ewes revealed apparent differences in distribution and movement patterns among herds (Fig. 1). During May-October, ewes from the CW herd showed widest distribution and greatest movements; CH ewes were the most limited in their range use and movement. Disturbances by hikers (CW) and hunters (CH, SL) influenced movements of ewe/lamb groups on occasion, and probably accounted for erratic movements sometimes observed in these herds. Despite these disturbances, marked ewes in all 4 study herds have shown remarkable fidelity to summer, winter, and transitional ranges, even in instances when they've been in close proximity to other ewe groups -- this observation is particularly noteworthy in the Tarryall Mountains, where the TE and SL ewe/lamb bands were observed within about 150 m of one another one day in september 1992 but did not intermingle. As in 1991-1992, marked ewes in the CW herd showed greater mobility than ewes in other herds during winter 1992-1993. Based on periodic observations, most marked CW ewes spent at least some time on alpine winter ranges despite relatively heavy snowfall in the Collegiate Mountains. Many of these ewes moved from alpine to lower elevation winter ranges and back several times during the winter. Wider distribution and greater mobility may be at least partially attributable to absence of bait stations, although these sheep tend to be more mobile than those in other herds included in our study. However, unbaited TE ewes were also frequently observed in a large group that ranged 0.5-5 km south and east of the traditional bait site regarded as their preferred winter range. Movement and distribution patterns for all 4 herds were recorded during February-April and May-June 1993, and those data will be added to update range use maps at the end of the summer field season. Additional field assistance during 1993-1994 will help provide a more complete data set for use in developing winter distribution maps for all 4 study herds. Treatment Rates As many as 72 sheep fed at the SL bait site and as many as 34 sheep fed at the CH bait site during December-February. All marked ewes fed at least 14 days. Marked ewes at SL (n = 14) averaged 38 days on bait (se = 0.5; missing data for 6 days) and marked ewes at CH (n = 15) averaged 51 days on bait (se = 2.8; no �173 missing data). Visitors harassment of sheep at the SL site apparently contributed to their somewhat erratic bait site attendance, although mean attendance might have approximated that of TE sheep in 1992 [mean(se) = 47(±1.1) days] if all 1993 data were available. Attendance at CH was somewhat higher and more consistent this year than in 1992 [mean(se) = 42(±3.8) days]. In addition to bait, 14 of 15 marked ewes at CH also received fenbendazole at least once and 13 received 2 fenbendazole treatments, for a total of 13.5 ewe treatments (Schmidt et al. 1979). All 8 fenbendazole-treated blocks disappeared almost entirely at the CW site between December and May (we recovered a total of 339 g of remnants in May). However, a control block from the same site weighed only 6.44 kg in May; we attributed the observed 57% loss largely to dissolving effects of heavy, wet snowfalls in February-April. Adjusting for estimated environmental losses (about 20% for blocks in December-February and February-April and about 17% for April-May), about 95.7 kg of treated blocks (about 157.9 g fenbendazole) were consumed at CW, more than a 4-fold increase over total consumption recorded at SL in 1992 (about 36.3 g fenbendazole). Using Schmidt et al.'s (1979) dose recommendation (3 g fenbendazole/ewe/day, twice), block consumption equated to about 26.5 ewe treatments at CW in 1993, compared to 6 ewe treatments at SL in 1992; under a more moderate dosing regime (about 0.75 g fenbendazole/ewe/day for 3 consecutive days; Foreyt and Coggins 1990), block consumption equated to about 70 ewe treatments at CW in 1993 and 16 ewe treatments at SL in 1992. We never observed use of blocks at CW directly, but did observe radiocollared sheep within 25 m of blocks and sheep tracks around blocks. Unfortunately, observations and tracks also suggested elk and mule deer were probably using these blocks, thereby confounding our estimates of drug delivery to radiocollared bighorn ewes. It follows that actual delivery of fenbendazole to CW bighorns in 1993 was probably only a fraction of the estimated total. In contrast, we believe estimated fenbendazole delivery at SL in 1992 was considerably more reliable because only bighorns appeared to use treated blocks at that site. Based on these experiences, assuring delivery of effective fenbendazole doses to target animals may represent a significant obstacle to field use of treated salt blocks in many parts of Colorado where bighorn, mule deer, and elk populations are sympatric. Lamb Production and Survival Our field crews have done an outstanding job of monitoring marked sheep to collect reliable data with minimal impacts on the study herds since monitoring began in 1991. All but 3 marked ewes were observed at least monthly from May to October 1992. Two of these ewes eluded field crews periodically, a third has not been seen (or a transmitter signal received) since late June 1992 and is presumed dead. Both reduced production and mortality apparently affected recruitment (lambs/marked ewes) to 6 months among lamb cohorts monitored during May-October 1992. Lamb production, as estimated by observations of marked ewes with new (~ 2 wk old) lambs at heel, appeared to vary somewhat among the 4 herds studied (Fig. 2). All 13 CH ewes were seen with lambs at least once last summer; in contrast, only 8 of 11 (0.73) SL ewes were ever observed with lambs at heel. Most, but not all, marked ewes in the other 2 herds also produced lambs (TE = 0.82, CW = 0.92). Although most lambs were born in May (TE, SL) and June (CH, CW), at least 1 new lamb was observed after 1 July in each herd; lambs may have been born during August in SL, TE, and CWo We recognize that our approach for estimating lamb production cannot discern between failures to bear live lambs and perinatal mortality among viable lambs. However, because lungworm treatment is directed specifically at reducing mortality in otherwise viable 26 month old lambs, we believe it important to partition the contributions of reproductive failure and perinatal mortality from mortality in older, viable lambs as sources of reduced recruitment among lamb cohorts. In addition to reproductive/perinatal losses, some apparently viable lambs disappeared in each of the experimental herds during the course of the summer. Lamb survival to 6 months ranged from 0.91 in CW (10 of 11 known lambs �174 excluding the dead and missing ewe/lamb pairs) to 0.62 in CH (8 of 13) (Fig. 2); lamb survival rates in SL (0.88) and TE (0.79) were both relatively high and comparable to CWo Field observations suggest factors affecting lamb survival at CH may differ somewhat from those affecting the other 3 study herds: one sick and 3 coughing lambs were observed in CH during late August through early October. (Coughing lambs were also observed in CH during pretreatment summer field work in 1991). Although the sick lamb at CH subsequently disappeared, the 3 others observed coughing survived through October. Sick or coughing lambs have not been observed in the 3 other herds studied, even though lambs have disappeared from all of those herds. Although the proportion of ewes observed with lambs and the proportion of those lambs surviving to 6 months of age varied somewhat among herds (Fig. 2), overall recruitment of lambs through October 1992 appeared to be relatively high across all 4 herds being monitored, regardless of management treatment (Fig. 3). Overall recruitment rates to 6 months ranged from 0.62-0.83 lambs/marked ewes (Fig. 3). For 3 of the study herds (TE, CH, SL), the proportion of marked ewes with lambs at the end of October was indistinguishable. Only preliminary data are available on lamb production and survival for the 1993 summer field season (Table 2). Lambing patterns appeared more consistent among the 4 herds than in 1992: for all 4 herds, new lambs were first observed in mid-May, and lambing peaked in late May-early June. As of late June, the overall proportion of radiocollared ewes that had produced lambs was relatively high, ranging from 0.71 at TE to 0.88 at SL (CH = 0.87, CW = 0.79). Population Parameters and Performance Other than 2 mid-summer mortalities in CW reported previously, no health problems were detected among marked ewes during May-October 1992. However, 3 Tarryall ewes were found dead in early December 1992 (SL-.), early May 1993 (TE-E), and mid-June 1993 (SL-9) -- no cause could be confirmed for any of these cases, but TE-E most likely fell victim to predation. Between February 1991 and June 1993, a total of 13 of 76 radiocollared ewes from our 4 study herds have died. Four losses (1 each at TE, CH, CW, and SL) were probably either directly or indirectly related to capture, although 3 of these 4 ewes survived at least 2 months after being captured and radiocollared. The other 9 losses (4 at TE, 3 at SL, and 2 at CW) were likely not influenced by previous capture. Of these, predation may have been involved in 4 losses in the Tarryalls (3 at TE and 1 at SL), but causes of death or disappearance for 4 other ewes (1 at TE, 2 at SL, and 2 at CW) have not been determined. Overall, noncapture mortality rates of adult ewes in the 4 study herds have averaged about 0.05 annually over the past 28 months, although causes and rates of ewe mortality seem to vary somewhat among herds. Winter range counts indicate all 4 bighorn herds under study have remained stable or grown since our experiment began in 1991. In the Tarryall Mountains, as many as 72 sheep (L:E:R = 20:27:25) (including all collared ewes) fed at the SL bait site in early 1993; about 55-60 unclassified sheep were observed in the Sugarloaf Mountain vicinity in early 1992. At least 61 TE sheep (L:E:R = ) (including all collared ewes) were observed in the Twin Eagles-Spruce Grove vicinity during December 1992-April 1993; as many as 60 sheep (L:E:R = 12:30:18) (including all collared ewes) had fed at the TE bait site in 1992. Despite good recruitment (>0.7) and modest adult mortality (about 0.08) rates, the actual number of bighorn sheep in the Tarryall Mountains (BHS GMU 23S) appears to be somewhat below recent estimates of about 250 head. Based on our data from and observations of the SL and TE herds gathered since 1991, it seems unlikely that the collective bighorn population in the Tarryall Mountains exceeds ~70 head in total unless other distinct (and as yet unidentified) herds occur in other parts of this mountain range. In the Collegiate Peaks, as many as 34 sheep (L:E:R = 8:18:8) (including all collared ewes) fed at the CH bait site in early 1993; 31 sheep (L:E:R = 8:16:7) �17) (including all collared ewes) had fed at the CH bait site in 1992. Reliably estimating bighorn numbers for the CW herd has been complicated by tendencies of these sheep to be highly mobile and remain on alpine ranges during much of the winter. However, our limited field observations indicate high recruitment and low adult mortality are characteristic of this herd, and it is likely growing. Intensive monitoring will continue through October 1993 with emphasis on documenting survival of known lambs, as well as movement patterns. Additional monitoring is planned for November 1993-April 1994 in conjunction with the experimental treatments in year 3, and intensive monitoring will resume in May 1994. Prepared by Table 1. Treatment assignments (A-D) for bighorn herds included management experiment to examine effects of alternative lungworm strategies on bighorn lamb survival and population performance. in a 4-year treatment HERD COLLEGIATE MOUNTAINS YEAR CHALK CREEK COTTONWOOD CREEK TARRYALL MOUNTAINS SUGARLOAF MOUNTAIN TWIN EAGLES 1992 B' C T BfT 1993 BfT T B C 1994 C B BfT T 1995 T BfT C B , Treatment assignments: BfT bait with alfalfa hay and apple pulp treated with fenbendazole; B = bait with alfalfa hay and apple pulp without fenbendazole; T = fenbendazole-treated salt blocks on bait stations; and C = withhold all bait and fenbendazole (control). �176 Table 2. May-June 1993 lamb production and survival for radiocollared bighorn ewes. COLLEGIATE - CHALK CREEK' 1993 10 (FRO) MAY • (.040) ? + + ? • (.080) - + + + • (.100) ? + + + 1 (.120) + + + 2 (.148) - + + + 4 (.180) ? ? + ? 5 (.202) - - - ? 6 (.220) ? + + + 7 (.240) ? - + 8 (.270) - + + 9 (.300) ? - - - *E (.318) - + + + H (.340) + + + M (.380) - + + + T (.442) ? - + - LIE 0 9/14 (0.64) 12/15 (0.80) 10/12 (0.83) JUN , Blue/white collars; freq. 149.040 - 149.492; black/yellow * Collar not functioning but ewe observed frequently. eartags. �177 Table 2. (continued) CREEK2 COLLEGIATE - COTTONWOOD 1993 ID (FRQ) MAY • (.520) 7 - + + (.540) 7 + + + ••• (.560) 7 - - - 1 (.582) **7 7 7 7 2 (.620) 7 + + + 4 (.642) 7 - - + 5 (.660) 7 + + + 6 (.697) 7 - - 7 (.722) 7 - 7 + 8 (.740) 7 + + + *9 (.782) 7 + + + E (.800) 7 + + H (.820) 7 - + + K (.858) 7 + 7 7 M (.880) 7 - 7 7 LIE 0 • 2 JUN 6/14 8/11 10/12 (0.43) (0.73) (0.83) Black/yellow collars; freq. 148.520 - 148.980; white/red eartags. * Collar not functioning but ewe observed frequently. * * Disappeared 7/1/92; presumed dead. �178 Table 2. (continued) TARRY ALL - SUGARLOAF MOUNTAIN3 1993 10 (FRQ) MAY • (.530) ? - + + (.562) ? + + + *1 (.600) ? ? + + 2 (.620) ? ? + + 4 (.642) ? + + - 5 (.660) ? + + + 6 (.682) ? + + + 7 (.700) ? + + + 8 (.738) ? + + E (.780) ? + - H (.800) ? + + + K (.818) ? - + M (.840) ? + + R (.880) ? - - - *T (.940) ? ? ? ? X (.980) ? ? + + LIE 0 9/11 (0.82) 12/14 (0.86) 11/15 (0.73) • 3 * JUN + Black/blue collars; freq. 149.500-149.999; black/orange eartags. Collar not functioning but ewe observed frequently. �179 Table 2. (continued) TARRY ALL - TWIN EAGLES' 1993 10 • • 4 (FRO) MAY (.020) ? + + + (.040) - - - + + JUN ••• (.060) - 1 (.080) ? + + + 2 (.100) - - ? - 4 (.120) ? + + + 5 (.140) - - - - 6 (.160) ? - + + 7 (.180) + + + 9 (.220) + + + H (.262) - + + K (.302) ? - - - M (.340) - + + P (.360) - + + R (.380) - - - - T (.418) ? - + + X (.460) - + + + LIE 0 7/17 (0.41) 12/16 (0.75) 11/17 (0.65) Black/white collars; freq. 148.020-148.460; black/yellow eartags. �f-' 00 o BIGHORN SHEEP TREATMENT STUDY CHALK CREEK HERO (MAY 1992 - OCTOBER 19921 ++ A MCM<T PAfNCfTON BIGHORN SHEEP TREATMENT STUDY COTTONWOOD CREEK HEAD (HAY 1992 - OCTOBER 19921 ••_ .• 'U, + X x< + X it_ ANT£AO A x EL~EY X SILVE'!SBfi GUlCHX ~ '1.ouHf~U ~ + ••con_DOlI ~NNY CREEK 'IAR~tlLAKE X X X X~~X PAIl X X MIDDLE COTTONWOOD + CREEK" A~BO~ ~ L~KJC ~ + + ~ Figure 1. Plots of May-October 1992 UTM coordinates (+ or X) for individual radiocollared ewes revealed apparent differences in distribution and movement patterns among the 4 study herds. During May-October, ewes from the CW herd showed widest distribution and greatest movements; CH ewes were the most limited in their range use and movement. Despite range overlap, we have not observed commingling of SL and TE ewes. ~ �BIGHORN SHEEP TREATMENT STUDY COTTON GORDON HERD (MAY 1992 - OCTOBER 19921 BIGHORN SHEEP TREATMENT STUDY SUGARLOAF MOUNTAIN HERD (MAY 1992 - OCTOBER 19921 cMURDY CREEK o\cCUll)Y HOUHTlIN + AwINOY _ + + + * x ++:fSAND CREEK IISOH ~' ++ + ~ '\!'WlLEY _ i+ "';!r"AlL- ~• ~ TAMYAU. PIAl! PILOT ~ + ~ I-' co I-' Figure 1. (continued) �182 100 r;::::'l en PROCUCI'" LMI!l5 IIIIIIIIIII WITH LAI/IIS AT8111O'Tt·S 80 LLI 3: LLI 60 LL 0 .... Z 40 LLI 0 IX: LLI a.. 20 COTTONWOOD (Control) SUGARLOAF (Treat Only) CHALK CREEK (Bait Only) TWIN EAGLES (BaltfTreat) HERD (Treatment) Figure 2. The proportion of ewes producing lambs and the proportion of those lambs surviving to 6 months of age appeared to vary somewhat among 4 bighorn herds managed under alternative lungworm treatment regimes in 1992. ................... OOTTOIlWOOD o SUGARLOAF ~'ClfALJ('ClIEEK 100 -* D ~ TWI! 80 VJ UJ iI:: 60 - 40 ::I: < ...I 20 UJ VJ III 0 MAY JUN JUL AUG SEP OCT NOV Figure 3. Overall, the proportion of ewes last observed with lambs in late October 1992 appeared relatively consistent across herds managed under 4 alternative lungworm treatment regimes, despite some variability in lamb production and survival among herds. EAGLES . �183 Colorado Division Wildlife Research July 1993 of Wildlife Report JOB PROGRESS State of Project Colorado No. W-153-R-5 Work Plan No. __~2~A~ Job No. Period Author: REPORT 9 Covered: _ Mammals Research Mountain Sheep Investigations Quantity and Quality of Mountain Sheep Habitat with Regard to Minimum Viable Populations and Response of Mountain Sheep to Human Activity July 1, 1992 - June 30, 1993 D. F. Reed ABSTRACT Contract work was completed with the University of Colorado, Colorado Springs, Geography and Environmental Studies Department, in mapping and evaluating mountain sheep habitats. This was done by using MIPS (Map and Image Processing System) for GIS (Geographic Information Systems) integration of mountain sheep distribution and movements, vegetation, viewshed, and water sources. Data collected on mountain sheep included group size, location, movements, marks (collars and/or eartags), and sex and age classification. vegetation classification was done using an algorithm which was "trained" to recognize the statistical characteristics of each class in each band of satellite imagery. Field work was completed for Area B, Nathrop through Brown's Canyon and the county line area east of Salida. Based on continued observations of the sheep north of the river in Area A, where it had been hypothesized that an estimated 50-60 ewes had failed to breed in 2 successive mating seasons, one of the three 2-year old rams that were transplanted into the area north of the river, returned from south of the river during the mating season. He died of pneumonia by early December, 1992. Another ram (yearling) from the Ramparts was released north of the river in November and he also died of pnemonia. Two "native" rams that crossed the river to the north during the fall of 1992 were fitted with radio collars in February and have since crossed the river to the south. ��185 QUANTITY AND QUALITY OF MOUNTAIN SHEEP HABITAT WITH REGARD TO MINIMUM VIABLEPOPULATIONS AND RESPONSE OF MOUNTAIN SHEEP TO HUMAN ACTIVITY Dale F. Reed P. N. OBJECTIVE Evaluate the quantity and quality of mountain sheep habitat with regard to minimum viable populations and test the response of mountain sheep to human activity. SEGMENT OBJECTIVES 1. Complete contracts between University of Colorado, Colorado Springs, and the Colorado Divsion of Wildlife for establishing a Geographical Information sytem (GIS) data-base for Areas A and B • 2. Prepare a draft report for Area B. ACKNOWLEDGMENTS I thank co-principal investigators M. W. Miller. R. B. Gill, J. Vayhinger, and S. ogilvie for their ideas and support. Division personnel D. Finch, W. Travnicek, J. Backstrand, and L. Spicer were helpful in collecting field data. Division personnel D. L. Schrupp and D. C. Lovell were helpful for their ideas and in coordinating GIS. BLM E. Brekke made field observations and provided important coordination. University of Colorado, Colorado springs, T. P. Huber developed the GIS procedures and products. DESCRIPTION The study areas have been described METHODS Methods and materials are described RESULTS OF AREAS by Reed (1991, 1992). AND MATERIALS in Reed (1992) and in the appended report. AND DISCUSSION Results and discussion for Area A and B are described the appended report, respectively. in Reed (1992) and in �186 LITERATURE Reed, CITED D. F. 1991. Quantity and quality of mountain sheep habitat with regard to minimum viable populations and response of mountain sheep to human activity. Colo. Div. of Wildl. Game Res. Rep. July:157-160. 1992. Quantity and quality of mountain sheep habitat with regard to minimum viable populations and response of mountain sheep to human activity. Colo. Div. of Wildl. Game Res. Rep. July: 173-193. preparedS), 4 A_~ ale F. Reed Wildlife Researcher �187 Draft 1 Sep 93 (w/o figs) MOUNTAIN SHEEP HABITAT USE IN THE ARKANSAS RIVER CANYON THE LOWER END OF BROWN'S CANYON AND THE COUNTY-LINE - PART B: NATHROP TO AREA SE OF SALIDA Dale F. Reed, Colorado Division of Wildlife, Prospect, Fort Collins, CO 80526 Center, Jack Vayhinger, Park, CO 80863 Colorado Erik B. Brekke, 81212 Bureau Division Researcher of Wildlife, of Land Management, 498 Old Wagon Canon Trail, City District, Thomas P. Huber, Department of Geography and Environmental of Colorado, Colorado Springs, CO 80933-7150 317 W. Woodland Canon Studies, City, CO University The Arkansas River from crosses the river is part of the recently established Arkansas Headwaters Recreation Area (AHRA). A Cooperative Management Agreement was signed between the Bureau of Land Management (BLM) and the Colorado Division of Parks and Outdoor Recreation (DPOR) on 27 October 1989 to establish a partnership for management of recreation resources in the AHRA. The AHRA has experienced increased recreational use, and further increases in recreation and facility development are projected. One of the more publicly visible, and perhaps sensitive, resources in the canyon is mountain sheep (Ovis canadensis canadensis). As recreational use (white water boating in particular) increases and recreational sites are developed, their impact on mountain sheep remains unknown. Specifically, impacts of human disturbance on mountain sheep habitat use and behavior had become BLM and Colorado Division of Wildlife (CDOW) management concerns. This generated needs for information on 1) mountain sheep habitat use, 2) quantity and quality of mountain sheep habitats available, 3) possible habitat enhancement for mountain sheep, and 4) mountain sheep response to human disturbance. To address these needs BLM, Canon City District, committed to a mountain sheep habitat study and subsequently engaged the CDOW in a cooperative effort. Further, the CDOW engaged the Department of Geography and Environmental studies, University of Colorado, Colorado Springs (UCCS) in implementing Geographical Information System (GIS) technologies to expedite and enhance data capture and display. The purpose of this report is to delineate mountain sheep habitat use, habitats available, and possible habitat enhancements in Area B located 1) east of the Arkansas River from Sugarloaf Mountain to the lower end of Brown's Canyon and 2) north of the Arkansas River where the Chaffee and Fremont county line crosses the river southeast of Salida. STUDY AREA Area B includes about 19 km (12 mi) of the Arkansas River Canyon from Sugarloaf Mountain to the lower end of Brown's Canyon and about 3 km (2 mi) of the river in the county-line area southeast of Salida. The study involved the areas specifically east and north of the Arkansas River because no recent distribution of mountain sheep were known to occur west or south of the river in these areas. Geological features of the canyon in these areas include granite and some cenozoic sediments (Chronic 1980). Rock outcrops are mainly of igneous rock, intermingled with shallow soils, terraces, and gravelly alluvia. Side drainages are numerous, often with steep sides, and occasionally have alluvial fans at their terminus. The Arkansas River and the Denver and Rio Grande Western railroad generally parallel one another creating a relatively narrow corridor through the areas. Common vegetation in the lower part of the canyon includes pinyon pine (Pinus edulis)-juniper (Juniperus spp.) (P-J); shrubs littleleaf mockorange (Philadelphus microphyllus), wafer-ash (ptelea trifoliate), currants (Ribes spp.), and �188 skunkbush (Rhus trilobata); grasses sand dropseed (Sporobolus cryptandrus), several Stipa and Boutelua spp.; and cholla and pricklypear cactus (Opuntia spp.). Mountain mahogany (Cercocarpus montanus) was common about a third of the way up the slope east and north of the river. Gambels oak (Quercus gambelii) was common at higher elevations east and north of the river. Climate was semi-arid and during winter snow usually melts off the southern exposed slopes quickly. METHODS Mountain Sheep Distribution and Movements Ground counts were made from selected locations 1-2 times per week during the summer of 1990. These were incorporated into the data even though the study had not begun. Monthly to bi-monthly counts were then made during 1991 except during the summer when again roughly 1-2 counts were made per week. Counts were made roughly 1-2 times per month January-April 1992. Searches were made from vehicle access areas at the lower end of Brown's Canyon, at Hecla Junction Recreation Site, at Sugarloaf and Ruby Mountain, and from u.S. Highway 50 with binoculars. All individuals were classified to sex and age. Group size, location, movements, marks (collars and/or eartags), interactions with kayaks and other major stimuli, date, and time were recorded. Locations were plotted on 7.5 minute quad maps and universal transverse mercator (UTM) coordinates determined. Most of these data were entered into dBase IV and then into MIPS (Map and Image Processing System; MicroImages Corp., Lincoln, NE) for GIS integration. MIPS has raster, vector, or CAD file capabilities. Sightings, movements, escape terrain, lambing areas, and home range were done using vector capabilities. Home range was calculated using the minimum polygon method. The 3-dimension (3-D) was done using 3-D raster display function and digital elevation model (OEM; U.S.G.S. survey 3-arcsec 1:250,000) data. Calibration of images was done in UTM - Zone 13 which allowed superimposing of images. MIPS printouts were run on a HP Paint jet XL printer and a Lasergraphics personal film recorder. Vegetation Vegetation classification was done using a maximum likelihood classification algorithm. This means that training sites were chosen to represent each vegetation class. The algorithm was "trained" to recognize the statistical characteristics of each vegetation class in each band of the satellite imagery (Landsat TM scene YS120117021XO, 15 Jun 87). Training sites were done by conducting line-pace transects (vegetation sampled every other step at the tip of one's boot) through 100 m diameter circle plots (0-360 degrees, direction chosen randomly, n = 50 data points) to verifiy the predominant plant species for given vegetation classes. These were transferred to the computer system and the algorithm was programed to classify all other pixels in the study area according to the "trained" classes. The actual classification was done using the Gaussian curves and standard deviations of each class to assign each pixel to that class where it was statistically most likely to fit. Five classes were used, namely, P-J, mountain shrub, grassland, riparian, and bare rock. Visibility Sheep habitat visibility (converse of percent vegetation and topographic obstruction) was sampled at most of the selected training sites for PJ. Five directions (0-360 degrees) and 5 distances (10-50 m) were chosen randomly for each of the 100 m diameter circle plots. The observer estimated by eye the percent of obstruction of aIm square white target at each of the selected directions and distances from the center of the circle plot (n = 5 data points). These percentages were then averaged for an overall percent visual obstruction for the given site. �189 Water and Mineral Ground searches for water were made up the numerous side drainages that fed into the Arkansas River during the summer when dry conditions were prevalent. RESULTS Mountain Sheep Distribution and Movements A total of 78 sheep sightings were made from May 1990 through January 1993. Most of these sightings were made east of the river (Table 1) because of the larger area and greater number of animals. However, access and visibility were limited. Based on the highest number of sightings on given days, an estimated 40-50 sheep were distributed in 2-5 subgroups east of the river from Sugarloaf Mountain to the lower end of Browns's Canyon and 20-30 in 2-5 subgroups north of the river in the county-line area. During the summers of 1990-92, incremental increases occurred with new lambs east of the river (lamb:ewe ratios: '90 = 32:100, '91 = 72:100, '92 = 32:100) and north of the river (lamb:ewe ratios: '90 = 83:100, '91 = 44:100, '92 = 75:100). The ram:ewe ratios east of the river were consistent ('90 = 16:100, '91 = 13:100, '92 = 13:100), but north of the river the ratios were variable ('90 = 63:100, '91 = 30:100, '92 = 22:100). Although sampling error may be the cause of some, if not much, of this variability, it is possible that the sheep north of the river were more productive (i.e. > no. of both rams and lambs per no. of females). Table 1. Number of sightings through 8 February 1993. of mountain Number Year 1990 1991 1992 1993 Total East of Arkansas River (Nathrop to lower end of Brown's Canyon) sheep in Area B from May 1990 of sightings North of Arkansas River (County-line area SE of Salida) Total 14 15 15 3 9 10 4 8 23 25 19 11 47 31 78 Distribution of sheep was likely influenced by escape terrain throughout the year and lambing areas during May and June. Escape terrain of 30 degree slopes were adjacent to the Arkansas River in nearly every km along the river of both areas, but escape terrain of 45 degrees was essentially negligible. Lambing was consistently observed at the lower end of Brown's Canyon, and occasionally at Sugarloaf Mountain and near Green Gulch. Most of the sightings occurred at the lower end of Brown's Canyon because of the ease of access and better visibility. Although this biases the distribution of data points (observers spending more time at this location and better able to see sheep), it was apparent that larger groups of sheep occurred at this location (largest group = 37; largest group elsewhere = 18). Limited sample sizes of groups sighted preclude making any reasonable inferences in regard to seasonal differences in distribution. �190 Movements of sheep in the two areas of Area B were also difficult to ascertain because of limited access and visibility. However, observations of marked animals and the finding of one collar provided an indication of where more extensive movements may have occurred (Table 2). Blue collared sheep 94 and 98 (Table 2) were from a release at Sugarloaf Mountain 11 January 1985. Table 2. Marked mountain sheep and dates when they were sighted or collar found east of the river from Sugarloaf Mountain to the lower end of Brown's Canyon, brief description of location, and the UTM coordinates. collarjEartag Location Dates Blue collar 94 18 23 9 23 Blue collar 98 21 Sep 92 Red eartag Red eartag 43 44 Radiocollar , Shown Jul Jul Aug Aug 90 90 90 90 SW Green Gulch NE Waterfall Sugarloaf Sugarloaf 40900 40892 40662 40667 428104 428046 429100 429064 Above 41425 428340 40789 40798 40785 40821 40799 40830 40798 40792 40791 427405 427407 427404 427445 427425. 427490 427485 427411 427420 40821 40782 427445 427420 40821 40916 40862 427445 428109 427696 Spg Gul 90 90 90 90 91 91 91 91 92 End of Brown's 8 Aug 90 17 Jul 92 End of Brown's 8 Aug 90 11 Sep 90 27 Sep 90 End of Brown's Below Green Gul S Railroad Gulch 5 11 26 8 24 25 4 12 17 Jul Jul Jul Aug Feb Feb Mar Aug Jul only to nearest UTM' " " " Comment Collar found 100 meters. The size of sheep groups in Area B North and South were not significantly different (range 1-37, mean = 9.6, ±SE 1.4, and range 1-18, mean = 6.4, ±SE 1.0, respectively; t = 1.54, df = 65, P > 0.10). No detectable pattern is evident from the distribution by group size. Assuming a sheep population of 70 (40-50 east and 20-30 north of the river; using medians of 45 and 25, i.e. 45 + 25 = 70) and an area of 25 sq km delineated by the minimum polygon method, the density of sheep per sq km = 2.8. These estimates may be biased because of the lack of information on movements of sheep east of and along the river and possibly between Area B North and Area B South. Habitat between these areas was primarily PJ and was not investigated. Interactions between sheep and kayaks and other selected stimuli suggest that sheep response was variable and that they probably habituated to people and trains (Table 3). Although the number of interactions noted between sheep and various stimuli were few (n = 14, Table 3), most exhibited no response or walked or trotted away from the stimulus. However, sheep responded dramatically to a train when half a group of 20 was in a fenced pasture between the river and the railroad tracks and when the other half was just west of and on the railroad tracks, by crossing fences and running about 150 and 110 m to near escape terrain at the lower end of Brown's Canyon (Table 3). Conversely, when sheep were on the escape terrain side of the train, they �191 showed no reaction or only slight similar to the behavior typically previously. Vegetation, visibility, water, flight reaction (Table 3). This was very exhibited by sheep to trains as reported and mineral Five classes of vegetation and bare rock were classified. These were PJ, mountain shrub, grassland, riparian, and rock in order to simplify and be consistent with classes typically used by BLM. Variability in the density of dominant vegetation occasionally complicated judging the classification. For example, some low density PJ with grass understory could be classified as grassland. Percent of dominant vegetation or rock (percent of PJ hits in the PJ class, etc.) ranged from 36-48% for PJ, 38-42% for mountain shrub, and 4672% for grassland, and 40-62% for bare rock (Table 3). Sheep east of the river were sighted in grassland (47%) more often than in either PJ (33%) or mountain shrub (20%) and sheep north of the river were most often sighted in mountain shrub (83%). Few sightings occurred on the railroad, in the railroad right-of-way, or between the railroad and the river. Some grasses appeared to green-up in the pastures (grassland) at the lower end of Brown's Canyon early as February and March and sheep use was observed. These pastures alone accounted for the greater percentage of sightings in the grassland type. Visibility was estimated at 8 PJ sites (Table 4). The lack of visibility or the percent of vegetation and topographic obstruction ranged from 28-92% (mean = 61.8, ±SE 7.9) for PJ (Table 4). As sampled, there was no significant difference in visual obstruction between PJ in Area B as compared to PJ in Area A (t = -1.370, df = 11, P > 0.10). More importantly, however, this vegetation class covered about 60% of the approximate 25 sq km in Area Band may have been one of the limiting factors in providing optimum visibility either within or between habitat patches. water occurred in most named drainages east of the river including Middle Cottonwood, Cottonwood, Spring Gulch, and Green Gulch about 0.5 to 2 km east of their confluences with the Arkansas River, and in an unnamed drainage I km northeast of the lower end of Brown,s Canyon. Water also occurred in Area B North at two locations up an unnamed side drainage of Wells Gulch. However, water availabilty at most of these sources was estimated to be temporal, often drying up during late summer or early fall. Conversely, an irregation ditch at the lower end of Brown's Canyon and the Arkansas River were almost always available and regularly used spring through fall. The only times these sources of water were not available was during periods in winter when water in the ditch was turned off or when the river froze over or formed ice shelfs along the shore. At those times there was usually some snow available for the sheep to eat. Periods when some open water sources dried up or ice shelves formed during cold periods without snow cover may have indeed affected the disturbution of sheep. However, our observations were not sufficiently numerous or refined to detect this. No natural mineral licks were found. Salt was probably provided for horses pastured at the lower end of Brown's Canyon and was probably used by sheep. �192 Table 3. Interactions between mountain sheep, people, other animals, kayaks, and vehicles June 1990 through June 1993. Date Distance (m) To Group Reaction' Moved2 stimulus3 size Year stimulus 1990 people 1991 people 12Aug none 50 26 horse 21May 12Aug none none 20 20 3 26 people deer 1993 people/animal activity walking 11 kayak 1992 Marks 17Jul 17Nov sfr mfr 25 125 50 120 20 22 pickup 17Jul mfr 5 50 20 train 17Jul ifr 150 450 10 17Jul ifr 110 200 10 17Jul sfr 30 190 20 dog 12Jun ifr train 19Jan 12Jun sfr none >30 50 15 5 100 17 85 15 o Red ear tag 43 none Red ear tag 43 none Red ear tags 43 & 44 Red ear tag ? Red ear tags 43 & 44 Red ear tag 44 Red ear tag 43 Red ear tags 43 & 44 Red ear tag 43 none Red ear tag 43 man working in field grazing look @ none none none none none none none none , Reaction includes: sfr = slight flight reaction where animals walk away from stimulus; mfr = moderate flight reaction where animals trot away from stimulus; ifr = intense flight reaction where animals ~ away from stimulus. 2 Distance moved away from and in apparent response to stimulus. 3 Distance between animals and stimulus when initial response detected. �193 Table 4. Vegetation type, number of transects or sites, percent dominant vegetation (Pinus edulis and Juniperus spp. for PJ, shrub spp. for mountain shrub, grass spp. for grassland, and cottonwood spp. for riparian), percent obscured (how much of sq m target was obscured by vegetation and/or topography), and general location of site. Type PJ No. Sites % Dominant Vegetation or rock % Obscured 8 36 49 E of river 48 60 Top lower Brown's 85 48 Lower Brown's Upper Green Gulch 50 Upper 28 92 82 Above Brown's Creek Top lower Brown's Top lower Brown's General Location at Hecla 1 8 Mtn Shrub Grassland 2 6 Riparian 1 Bare rock 2 Total Green Gulch 38 42 N of Railroad Gulch Upper Grn Gulch 62 66 72 48 2 46 80 Bottom of W of Ruby N end Mid S of upper Brown's Mountain Cottonwood Grn Gulch Comment Moderate density Moderate density High density Moderate density Moderate density Low density High density High density Wide range spp Grazed of pasture Snakeweed 24% E river SE Hecla Above Brown's Creek 2 94 In RR Gulch 40 62 N side RR Gulch Bottom of Brown's drainage Cottonwood Mostly talus 19 Measurement not taken. Transect parallel to drainages directions. I 2 rather than at randomly chosen DISCUSSION The information on sheep numbers and sex/age composition indicates that seemingly healthy and "normal" subpopulations of sheep are represented in the two areas. Area B North was largely vegetated with PJ and has widely dispersed, limited escape terrain. Consequently, for this area it is not surprising that sheep distribution is characterized by a high degree of patchiness. Area B South was vegetated with grassland or mountain shrub at lower elevations and PJ at higher elevations and appears to have adequate escape terrain adjacent to where most sheep were sighted. What the carrying �194 capacity is and how the range, including movement corridors, might be improved or expanded for both areas were not determined in this study. Although there could have been movement between these 2 areas, the lack of sighting any marks from Area B North in Area B South, suggest that there may not have been much interchange between them. Had there been data suggesting a movement corridor, recommendations might have included expanding such corridors by clearing PJ vegetation. A more extensive study involving telemetry (see Clark et al. 1993) would be needed to resolve this uncertainty. However, the use of corridors for some species has been debated (Beier and Loe 1992, Simberloff et al. 1992). The information on interactions suggest that sheep in this study generally responded to familiar and predictable stimuli including kayaking and peopleon-foot activities by exhibiting only slight to no apparent overt reactions. When they were on or west of the railroad tracks and trains came, the response was heightened. If surprised by humans, or if human activity was less predictable, it is estimated that their response would have also been heightened. During summer, boating became daily and relatively predictable phenomena (typical hours, etc.) and habituation would have been expected. Despite the only one case of sheep being observed responding to kayaks, there is a question of whether the interactions with boaters deterred sheep from getting water and how important that was to sheep. The timely use of water during early to mid summer would be especially important to lactating females and new lambs. Futhermore, there may have been physiological impacts and energy costs that occurred in the absence of overt behavioral reactions (i.e. elevated heart rate and excitement levels) (MacArthur et al. 1982). For example, MacArthur et al. (1979) reported that the continued presence of a human within 50 m of sheep resulted in a 20 percent rise in mean heart rate. Additionally, King and Workman (1986) suggested that responses of desert bighorn (~ ~ nelsoni) to encounters with humans were more severe and energy costly for animals that had been previously exposed to relatively high levels of human disturbance. Hence, for animals in this study, the effects of increased boating acitivity may not diminish, but rather increase over time. The habitats that were used by sheep in this study can generally be described as either PJ, mountain shrub, grassland, or bare rock located in a relatively narrow river canyon with adjacent steep terrain. Some of the components of these habitats that were likely important to sheep were not unlike those often reported by other workers (Boyd et al. 1986): - water, - steep, rocky, - visibility - mineral terrain, for predator detection and visual communication, licks, - suitable - forage escape thermal environments, of adequate quantity and and quality. Favorable combinations of these components existed at best in patchy distributions. For habitats east and north of the river, the river problably served as the primary source of water and possibly as some kind of western and southern boundary. Similarly, the distances that sheep venture away from water « 3.2 km [Smith et al. 1991:212.]), in this case the river especially during dry periods, and the distances that sheep venture away from escape terrain (0.5 km [Van Dyke et al. 1983]), may have strongly influenced the location of "high use" areas, movement corridors, and the home range boundaries of these sub-populations. Additionally, the steep terrain used for lambing, the openness or visibility used for predator detection and visual communication, the southern exposed slopes used for suitable thermal �195 environments, and the forage quantity (density and diversity of plant species) and quality (palatability and nutritional levels), probably contributed to a mosaic of habitats within the delineated home range. How sheep subpopulations have responded to this patchy environment (Weins 1976) and whether critical thresholds in spatial patterns (e.g. long-term closure of the PJ vegetation type) have. altered ecological processes (Turner 1990) is uncertain. A missing component for habitats in both of these areas may be regularly used mineral licks. Only one area, just north of the east pasture at the lower end of Brown's Canyon where salt was supplied (w. Travnecek, per. comm.), was kwown to have minerals for sheep. Escape terrain and lambing areas for both east and north of the river were judged to be adequate, although specific measurements were not made. Escape terrain has been variously described as cliffs with at least 8 x 200 m (ht x length) dimensions (McCollough et al. 1980) and areas of at least 0.16 ha (Van Dyke et al. 1983). Van Dyke et al. (1983) suggested a relationship between distance from escape terrain and extent of habitat use. Wakelyn (1987) indicated that ranges supporting greater numbers of sheep had more habitat on or within 0.25 km of escape terrain. A number of smaller cliffs could probably be substituted for a large one and terraced cliffs that allow sheep a variety of escape routes are likely superior. Lambing areas have been described as cliffs with 8 x 260 m (ht x length) dimensions and areas of at least 2 ha (Van Dyke et al. 1983). However, it may take a computer model to adequately describe the array of variables associated with escape terrain and lambing areas. Good visibility or the lack of visual obstructions is generally espoused as an important component for predator detection, visual communication, and efficient foraging in sheep (Boyd et al. 1986, Wakelyn 1987), but how it has been measured has varied between workers. Risenhoover and Bailey (1985) estimated the percent of each quarter of the compass over which an object 90 cm in height could be seen at 40 m from the observer. Thomson (1975) used a more exact measure of the degree of obstruction by using a screen behind the vegetation, photographing it, and later projecting the image on a quadrat matrix for analysis. The technique used in this study was considered sufficiently accurate for estimating the visibility of approaching predators (of which would likely be < 1 m in ht) and other sheep, but the measurements were not taken in habitats where sheep were necessarily observed. Hence, considering that these training sites were not representative of habitats used by sheep (possibly having> vegetation and topographic relief), it may not be suprising that measurements of their visibility were relatively low (i.e. for PJ and mountain shrub). Habitat evaluation procedures have been developed for desert bighorn sheep (Q. Q. mexicana) (Grunigen 1980, Hansen 1980, Holl 1982, Armentrout and Brigham 1988, Cunningham 1989) and more recently for Rocky Mountain bighorn sheep (Smith et al. 1991). However, only the procedure by smith et al. (1991) critially examines the minimum viable population (MVP; defined as the smallest isolated population having at least a 95% probability of surviving at least 100 years [Shaffer 1983) in sheep. Using information from Berger (1990) and suggestions of others (e.g. Van Dyke et al. 1983), Smith et al. (1991) suggested that 125 individuals represent a "best estimate" MVP. In the present study, sheep south of the river likely interact with other sub-groups, thereby relaxing the MVP requirement, but for those north of the river of which appear to be isolated, it is another matter. Important considerations in any habitat evaluation are the relationships between habitat use/availability data, habitat quality, and population density. Van Horne (1983) suggested that our understanding of individual species-habitat relationships is still rudimentary. She indicated that without sufficient data it would be difficult to distinguish "source" and "sink" habitats and that habitat quality is inversely related to MVP size. She gave an extreme example where one could imagine a habitat in which all the animals were immigrants and none emigrated or reproduced, and thus, where the quality of the habitat would be zero. Does some of this example apply to the �196 habitat north of the river in our study? Additionally, Hobbs and Hanley (1990) suggested that we do not yet have a scientifically sound and complete method of evaluating habitat quality. They indicated that habitat evaluation systems capable of predicting, for example the effects of human impacts, require understanding the mechanisms by which the environment influences the distribution and abundance of animals. They further suggested that an understanding of the cause-and-effect relations linking the performance of populations to resources in their habitats is fundamentally important. Predictability could potentially be greatly increased by the use of habitat models. However, one of the problems in attempting to validate habitat models is the difficulty of obtaining an independent measure that can serve as a suitable comparison (Schamberger and O'Neil 1986). MANAGEMENT IMPLICATIONS AND RECOMMENDATIONS Habitat use by bighorn sheep in both areas of Area B were likely related the patchiness of several habitat components, namely, escape terrain, preferred forage species (shrub and grassland spp.), proximity to water, lambing areas, and visibility. Of these components, only forage species, water availability, and visibility could be reasonably manipulated (improved for sheep). Application of any such manipulation presumes knowledge about the degree to which it would be effective. Three methods might be used to increase or improve forage species, i.e., irrigation, fertilization, and clearing of PJ. Irrigation is likely impractical considering costs, topography, water rights, etc. and fertilization on sheep ranges has had mixed results (Bear 1978). Clearing of PJ could be used both for increasing shrub and grassland species and for improving visibility and hence improving preferred habitats and movement corridors. Although this study did not identify specific habitats or corridors that could be improved, the potential may exist. Development of water and mineral sources may be the most practical means of improving and possibly extending some of the sheep range. Specific recommentations for water and mineral are: - develop water sources in most named drainages about 500 to 1000 m east of the river in order to provide water during dry periods - insure that water sources occur between the lower end of Brown's Canyon and the county-line area east of Salida provide saltlicks at: 1) north of the river east of Salida 2) Sugarloaf - maintain and railroad tracks in the county line area and Ruby Mountains a saltlick at the lower end of Brown's Canyon. The question of the effects of disturbance by rafting, kayaking, and other recreation on the river was probably not a major concern in the areas of Area B. Interactions between river recreationists and sheep were limited. It is probably impractical to validate, refine, or develop habitat models for a dispersed, patchily distributed sheep population such as the one for Area B. SUMMARY Mountain sheep habitat use was evaluated in the Arkansas River Canyon from Sugarloaf Mountain north of Nathrop to the lower end of Brown's Canyon and the area north of the Chaffee and Fremont county line east of Salida. GIS �197 technologies (MIPS) were used to summarize and produce color and 3-dimensional plots. An estimated 60-80 sheep in 4-8 subgroups were distributed along the east and north sides of the river. New and innovative approaches may be needed if the sheep ranges are to be improved and/or expanded. Management recommendations are provided for improving availability of saltlicks and water. Acknowledgments. We thank T. Grette and L. M. Berta of the BLM for technical advice and management and budget support, and S. Ogilvie, W. Travnicek, D. C. Finch, D. C. Lovell, D. L. Schrupp, M. W. Miller, and R. B. Gill of the CDOW for field support, advice, and management and budget support. J. Backstrand collected bighorn sheep field data during the summers of 1990 and 1991. C. Huber at UCCS Image Lab assisted with GIS technologies. LITERATURE CITED Armentrout, D. J. and W. R. Brigham. 1988. Habitat suitability rating system for desert bighorn sheep in the Basin and Range Province. USDA Bureau of Land Management Technical Note 384. 18pp. Bear, G. D. 1978. Evaluation of fertilizer and herbicide applications on two Colorado bighorn sheep winter ranges. Colo. Div. Wildl. Div. Rep. 10. 75pp. Beier, P. and S. Loe. 1992. A checklist for evaluating impacts to wildlife movement corridors. Wildl. Soc. Bull. 20:434-440. Berger, J. 1990. Persistence of different-sized populations: an empirical assessment of rapid extinctions in bighorn sheep. Conserv. Biol. 4:9198. Boyd, R. J., A. Y. Cooperrider, P. C. Lent, and J. A. Bailey. 1986. Ungulates. Pages 519-564 in Inventory and Monitoring of Wildlife Habitat. A. Y. Cooperrider, R. J. Boyd, and H. R. Stuart, eds. U.S. Dep. Int., Bur. Land Manage., Denver, CO. 858pp. Chronic, H. 1980. Roadside geology of Colorado. Mountain Press, Missoula, MT. 322pp. Clark, J. D., J. E. Dunn, and K. G. Smith. 1993. A multivariate model of female black bear habitat use for a geographic information system. J Wildl Mange 57:519-526. Cunningham, S. 1989. Evaluation of bighorn sheep habitat. Pages 135-160 in The desert biighorn sheep in Arizona, R. M. Lee, ed., Arizona Game and Fish Dep. 265pp. Grunigen, R. E. 1980. A system for evaluating potential bighorn sheep transplant sites in northern New Mexico. Bienn. Symp. North. Wild Sheep and Goat Counc. 2:211-228. Hansen, C. G. 1980. Habitat evaluation. Pages 320-335 in Monson, G. and L. Sumner. eds. The desert bighorn -- its life history, ecology, and management. Univ. Arizona Press, Tucson. 370pp. Hobbs, N. T. and T. A. Hanley. 1990. Habitat evaluation: do use/availability data reflect carrying capacity? J. Wildl. Manage. 54:515-522. Holl, S. A. 1982. Evaluation of bighorn sheep habitat. Desert Bighorn Sheep Counc. Trans. 26:47-49. King, M. M. and G. W. Workman. 1986. Response of desert bighorn sheep to human harassment: management implications. Trans. N. Am. Wildl. Nat. Resour. Conf. 51:74-85. MacArthur, R. A., R. H. Johnston, and V. Geist. 1979. Factors influencing heart rate in free-ranging bighorn sheep: a physiological approach to the study of wildlife harassment. Can. J. Zool. 57:2010-2021. MacArthur, R. A., V. Geist, and R. H. Johnston. 1982. Cardiac and behavioral responses of mountain sheep to human disturbance. J. Wildl. Manage. 46:351-358. McCollough, S. A., A. Y. Cooperrider, and J. A. Bailey. 1980. Impact of cattle grazing on bighorn sheep at Trickle Mountain, Colorado. Bienn. Symp. Northern Wildl. Sheep and Goat Counc. 2:42-58. �198 Risenhoover, K. L. and J. A. Bailey. 1985. Foraging ecology of mountain sheep: implications for habitat management. J. Wildl. Manage. 49:797804. Schamberger, M. L. and L. J. O'Neil. 1986. concepts and constraints of habitat-model testing. Pages 5-10 in J. Verner, M. L. Morrison, and C. J. Ralph, eds. Wildlife 2000: modeling wildlife-habitat relationships of terrestrial vertebrates. Univ. Wisconsin Press, Madison. Shaffer, M. L. 1983. Determining minimum viable population sizes for the grizzly bear. International Conf. Bear Research and Manage. 5:133-139. Simberloff, D., J. A. Farr, J. Cox, and D. W. Mehlman. 1992. Movement corridors: conservation bargains or poor investments? Conserve BioI. 6:493-504. Smith, T. S., J. T. Flinders, and D. S. Winne 1991. A habitat evaluation procedure for Rocky Mountain bighorn sheep in the intermountain west. Great Basin Nat. 51:205-225. Thomson, W. R. 1975. A photographic technique to quantify lateral cover density. J. South Afr. Wildl. Manage. Assoc. 5:75-78. Turner, M. G. 1990. spatial and temporal analysis of landscape patterns. Landscape Ecology 4:21-30. Van Dyke, W. A., A. Sands, J. Yoakum, A. Polentz, and J. Blaisdell. 1983. Wildlife habitat in managed rangelands--the Great Basin of southeastern Oregon: bighorn sheep. USDA Forest Service General Technical Report PNW159. Pacific Northwest Forest and Range Experiment Station, Portland, OR. 37pp. Van Horn, B. 1983. Density as a misleading indicator of habitat quality. J. Wildl. Manage. 47:893-901. Wakelyn, L. A. 1987. Changing habitat conditions on bighorn sheep ranges in Colorado. J. Wildl. Manage. 51:904-912. Wiens, J. A. 1976. Population responses to patchy environments. Ann. Rev. Ecol. Syst. 7:81-120. �199 Colorado Division Wildlife Research July 1993 of Wildlife Report JOB PROGRESS Colorado State of Project No. W-153-R-4 Work Plan No. 3A 2 Job No. Period Author: REPORT Covered: Mammals Research Pronghorn Investigations Habitat Selection ahd Population Performance of a pioneering Pronghorn Population July 1, 1992 - June 30, 1993 T.M. Pojar Abstract The distribution of the Middle Park pronghorn (Antilocapra americana) herd has been monitored since January 1, 1987. During this time, the population has increased from 80 to approximately 425 animals (summer 1993). The populations rate of increase (ROI) has ranged from 0.52 in 1986-87 to 0.10 in 1990-91 and averaged 0.27 over 7 years. There is a significant (P=0.0432) negative correlation between ROI and population size; ROI reaches zero when population size is approximately 423 animals. This can be construed to roughly estimate the K-value for the population. The seasonal distribution of this herd has not changed substantially since monitoring began in 1987 even with a 5 fold increase in the population. The summer range includes sagebrush habitat that extends from north of Granby (Coffee Divide) west to Kremmling, then north to Muddy Pass mostly north of the Colorado River (Fig. 1). The winter range encompasses approximately 65 km2 (25 mi2) north and east of Kremmling on ridgetops west of the Troublesome Creek drainage (Fig 1). There are 33 radios functioning as of this writing; 4 on adult males and 29 on adult females. Of the five solar powered ear tag radios put on male fawns in December 1991, one of the fawns died the following spring and 2 of the remaining 4 radios cannot be located. This means that either the animal has "disappeared" or that the radio has ceased to function. If the animal was predated and the solar panel was covered or buried by the predator, no signal would be emitted and the radios would be impossible to loqate. The third consecutive hunting season was authorized in 1992 with 10 buck permits and 5 doe/fawn permits issued •. -There was 100 percent hunter success with no documented evidence of wounding loss. . ��201 HABITAT SELECTION AND POPULATION PERFORMANCE PRONGHORN POPULATION Thomas OF A PIONEERING M. Pojar P.N. OBJECTIVE Describe population dynamics pronghorn population. and habitat SEGMENT and annual use of a pioneering, expanding OBJECTIVES 1. Describe seasonal population. distribution of the Middle 2. Determine necessary 3. Monitor population dynamics of Middle Park pronghorn with: a. Ground counts to describe changes in population size. b. Ground counts to quantify population sex and age composition. sample sizes of radio-collared animals to describe habitat preferences. Park pronghorn and observations STUDY AREA The study area is described in Pojar (1988:183-184). For orientation of Middle Park in relation to the state of Colorado see Figure 1. The approximate area of sagebrush steppe habitat in Middle Park is outlined in Figure 1. This is considered to be the potential area of distribution for pronghorn in Middle Park although, with few exceptions, the population inhabits only the area north of the Colorado River. METHODS SEASONAL AND ANNUAL AND MATERIALS DISTRIBUTION Tracking was done mostly from the ground to increase the probability of observing and identifying animals with numbered plastic collars. Fixed wing aircraft was used if an animal could not be located after a reasonable effort from the ground. Legal descriptions of animal locations were recorded to the nearest quarter mile then converted to UTM (U.S. Army 1973) coordinates for computer processing. All radioed animals have been located biweekly (with very few exceptions) since January I, 1987. POPULATION SIZE AND STRUCTURE Herd structure estimates were obtained in late summer or early autumn by locating all radioed animals and classifying all animals that accompanied them. The herd structure estimate that is used in population projections the one with the largest sample size obtained in August or September. Classification after October 1 is not used because the probability of mistaking early fawns for does increases. is Pronghorn congregate into large groups during winter and by locating all radioed animals it is possible to obtain a nearly complete census. A total count of mature bucks (age 1.5 years and older) is also possible which permits verification or correction of the buck:doe ratio from the late summer herd structure estimates. For population projections, the total population count and number of mature bucks counted during winter are used and recruitment is based on the fawn to doe ratios from late summer. Population 1. projections Winter mature are based on the following assumptions: counts represent the total population bucks in Middle Park. and the total number of �202 2. Late summer age ratio estimates represent "recruitment" into the population. Annual survival of mature bucks and does and female fawns is 92.5%. Annual survival of males in their first year (after weaning) is 50%. (This severe mortality on male fawns is arbitrary, however, it allows the number of mature males in subsequent years to match fairly well with winter counts.) 3. 4. RESULTS SEASONAL AND ANNUAL DISTRIBUTION The distribution of the population, as determined by re-location of radioed and marked animals, has not changed substantially from previous years. In general, summer area of habitation includes sagebrush habitat from Granby west to Kremmling and north to Muddy Pass. All summer observations in the first years of tracking this population have been north of the Colorado River (Figure 1). However, in the summer of 1992, one radioed doe accompanied by a yearling buck and 2 other does stayed in the area of the Taussig Ranch (west of Williams Fork Reservoir) south of the Colorado River. This same doe, accompanied by 2 to 3 other pronghorn (including 1 buck) summered in the same area in 1993. Another radioed doe that summered north of the river in the Parshall Divide area during 1992 moved south of the river during the summer of 1993. She was in a group of 6 does, 1 buck, and 3 fawns that inhabited an area east of Williams Fork Reservoir about 2 miles south of the Colorado River. Totally, 13 or 14 pronghorn summered south of the Colorado River during 1993 which is the largest number thus far documented summering south of the river. The same 6 radioed animals that summered in the Granby area in 1992 returned in 1993. By mid-summer there were slightly over 50 pronghorn (including fawns) in the Granby area. Two radioed does, both marked as fawns in Middle Park in December 1991, spent the summer of 1992 and the winter of 1992-93 in North Park. They wintered with a large band (400+) of pronghorn on the Araphoe National Wildlife Refuge 3-4 miles southwest of the town of Walden. One of the does remained in North Park during the summer of 1993 and the other one returned to Middle Park and summered near Whitley Peak. POPULATION SIZE AND STRUCTURE Total population size estimates are obtained during winter and herd structure estimates are obtained in late summer (Table 1). The annual changes in population size are used to calculate the rate of increase which is regressed on population size (a significant negative correlation, P=0.0432) to project the K-value for the population. The ROI is calculated as ROI = P2-P1 P1 where PI is the population size at time 1 and P2 is the population at time 2 (Table 2). The rate of increase for 1992-93 is 0.14 (Table 2). With a projected late summer population of 425 animals for 1993 (Table 3) there are 7 points available for the regression to project the populations K-value. Because of a relatively high fawn to doe ratio of 66:100 mid-summer 1993, the ROI increased in 1993 over 1992 (Table 2) which raises the projected K-value to approximately 423 animals (Figure 2). The population projection for late summer 1993 is presented in Table 3. The projection is based on the age ratio obtained during mid-summer 1993 and the total count of the herd and the number of bucks obtained during the winter of 1992-1993. In this projection it is assumed that the age ratio for the 1993 �203 production is 66 fawns:100 does. This is a critical assumption and subsequent ratios based on larger sample sizes may have a significant impact on this projection. The winter of 1992-93 was moderately severe with deep snows, 38 cm (15 in.) or greater that persisted throughout the winter. During most of the winter the herd was confined to a small portion of their customary winter range. Mortality on the winter range was not noted and the general condition of the animals was rated as fair to good throughout the winter. Although mortality was not observed on the winter range, latent mortality from the stress of winter may have taken place during spring dispersal which would not have been detected. The animals most likely to succumb to latent winter stress are the young of the year and old age classes. It is possible that the relatively high fawn to doe ratio observed in late summer 1993 is the result of mortality of 1992 fawns reducing the number of non producing yearling does in the "doe" segment of the herd structure sample of late summer 1993. If this is the case, the actural population will be lower than the projected and, consequently, the estimated ROI for 1993 would be less than the observed value of 0.23. The total population count for the winter of 1993-94 will reveal the true ROI for 1993. Table 1. Herd structure of Middle Park pronghorn based on a sample obtained by locating radioed animals in late summer. The population size is from the subsequent winter counts, i.e. herd structure obtained in late summer 1992 is related to the population size for the winter of 1992-93. YEAR POP. SIZE NO. RADIO RADIO B: IOOD RATIO F:IOOD RATIO SAMPLE % OF POP. 1986' 1987 1988 1989 1990 1991 1992 80 122 160 223 261 308 347 7 24 22 17 13 39 5.7 15.0 10.2 6.5 4.2 11.2 36 54 40 56 22 23 26 77 77 32 50 47 65 48 47 63 108 161 148 148 286 59 52 68 72 66 48 82 , This year's 1986. data based % on the sample of the population trapped 16 December Table 2. Population size of the Middle Park pronghorn herd during winter and the calculated rate of increase. Population size reflects the removal of 15 animals per year by harvest beginning in 1990, i.e. the 1990-91 winter population was 261 before harvest and 246 after harvest. YEAR 1986-87 1987-88 1988-89 1989-90 1990-91 1991-92 1992-93 1993-94 POP. SIZE (Projected) 80 122 160 223 246 292 332 410 RATE OF INCREASE .52 .31 .39 .10 .19 .14 .23 �204 Table 3. Population projection text for the assumptions. I POPULATION I BUCKS for the Middle I DOES Park pronghorn population. I I FAWNS 73 179 80 332 WINTER MORTALITY 73 X .075 = 5 MORT 179 X.075 = 13 MORT 40X.5=20B 40X.075=3D 41 PREFAWNING 1993 73 - 5 = 68 MATURES + 20 YRLS TOTAL = 88 179 - 13= 166 MATURES + 37 YRLS TOTAL = 203 LATE SUMMER 1993 MATURE 68 YRLS 20 TOTAL 88 MATURE 166 YRLS 37 TOTAL 203 Pojar, 291 @ 66F:100D 203 X .66 = 134 FAWNS 425 CITED T.M. 1988. Habitat selection and population performance of a pioneering pronghorn population. Colo. Div. Wildl. Res. Rep. July, 181-192. U.S. Army. 1973. Technical Manual: Headquarters, Dep. of the Army, Prepared by Thomas M. Pojar Wildlife Researcher I TOTAL WINTER '92-93 REFERENCES See pp Universal Transverse Mercator Grid. Washington D.C. TM No. 5-241-8, 64 pp. �205 ,--, ,,, I ' '- , I I 1 N . ~ ...-DENVER MIDDLE PARK Colorado Figure 1. Location of Middle Park in relation to the state of Colorado. The dashed line circumscribes the general area of sagebrush steppe habitat. The wintering area north and east of Kremmling is outlined with a solid line and the summer area is all sagebrush habitat north of the Colorado River. �1 N o 0\ 0.9 0.8 0.7 Q) m ...0 Q) c • '87 0.5 '0 '89 Q) 1ii a: y = 0.5281 - 0.OO1248x R2 = 0.5916 0.6 0.4 • 0.3 • • 0.2 0.1 Projected K-Value = 423 '93 '88 '90 • 0 0 100 200 300 400 500 600 700 Population Size Figure 2. Correlation of rate of increase and population size of the Middle Park pronghorn population. Where the regression line intersects the x axis is the estimated K-value for the population. �207 APPENDIX I Middle Park pronghorn with radios that are functioning 1993. Animal ages without a "+" are known aged. SEX AGE F F F F F F F F F F F F F F F F F F F F F F F F F M M M F F F M F 6 4+ 5+ 6 9+ 5+ 4+ 9+ 4+ 5+ 5+ 6 3 3 3 6 3 5+ 5+ 2 2 2 2 5+ 5+ 2 2 4+ 5 7 6 8+ 8+ EAR TAG OLD NEW COLLARS NEW OLD B12 B29R W5R W22R B23R Y4R W11R W8R Y1R W17R W19R W27R B11R W13R W2R W37R B42R W36R W38R W7R W6R W12R W18R SUEF (DAM W35R W26R W28L W33L W4R B1R B34R B5R B3R Y12R as July COMMENTS 148.110 Yrl in Dec '88 148.130 148.200 148.630 148.210 Yrl in Dec '88 148.400 148.220 148.230 148.240 148.530 148.250 Prior radio 148.420 148.260 148.270 148.290 148.580 148.310 Yrl IN Dec '88 148.330 Yrl in Dec '91 148.340 148.350 B23 148.360 YRL IN Dec '88 148.380 148.420 Telonics refurbished radio 148.660 148.670 Fawn in Dec '91 148.680 Fawn in Dec '91 148.690 Fawn in Dec '91 148.710 B14L & GlORi FAWN IN Dec '91 = YC17) 148.720 148.738 ATS factory refurbished 148.790 Fawn in Dec '91, solar ear tag 148.840 Fawn in Dec '91. solar ear tag 148.950 149.250 Telonics refurbished radio 149.330 Telonics refurbished radio B17 148.570 Deployed Dec '88 148.590 Deployed Dec '88 148.450 Deployed Dec '88 ��209 Colorado Division Wildlife Research July 1993 of Wildlife Report ) JOB PROGRESS State of Project Work Colorado No. Plan No. Job No. Period Author: REPORT Covered: W-153-R-4 Manunals Research 3A Pronghorn 5 The Feasibility of Detecting Density Dependence in Natural Populations Investigations July 1, 1992 - June 30, 1993 T. M. Shenk Abstract Numerous tests have been developed over the last twenty years to detect density dependence in populations from analysis of temporal trends in density. Performance of these tests for data from natural populations has not, however, been rigorously evaluated. From an extensive literature review, three tests were selected for evaluation based on: (1) accepted use in the literature, and (2) recent development. These tests include an autoregression method, a randomization test, and a parametric bootstrap likelihood ratio test. Each selected test was coded in a statistical software package in preparation for Monte Carlo simulations. Each test code was verified through extensive simulation to reproduce results reported in the literature. Equations to be used to generate biologically realistic data for the Monte Carlo simulations are presented. ��211 THE FEASIBILITY OF DETECTING DENSITY Tanya DEPENDENCE IN NA~~ POPULATIONS M. Shenk P. N. OBJECTIVE Evaluate the feasibility of detecting density populations through Monte Carlo simulations. SEGMENT dependence in natural OBJECTIVES 1. Through an extensive literature search select existing tests for detecting density dependence to be evaluated through Monte Carlo simulations. 2. Program SAS. 3. Verify 4. Develop equations to be used to generate the Monte Carlo simulations. selected, existing tests into the statistical the SAS code for each selected software package test. biologically realistic data for STUDY AREA The study will take place in the Department of Fishery Colorado State University, Fort Collins, Colorado. METHODS SELECTION OF EXISTING and Wildlife Biology at AND MATERIALS DENSITY-DEPENDENCE TESTS FOR EVALUATION It has been generally accepted for several decades that density dependence can be detected by the analysis of long-term life-table data that provide estimates of temporal trends in the density of different life-history stages (e.g. Varley and Gradwell 1960, Eberhardt 1970, Bulmer 1975, Royama 1977, Slade 1977, Berryman 1978, Vickery and Nudds 1984, Pollard et al. 1987, Reddinguis and den Boer 1989, Turchin 1990, Dennis and Taper 1993, Holyoak and Crowley 1993). However, existing tests developed to analyse temporal trends in population density may not be sensitive enough to allow for detection of density-dependence (see Strong 1986, Gaston and Lawton 1987, Hassell et al. 1987, Lomnicki 1987, Mountford 1988, Solow 1990, Bartmann et al. 1992). The influence of confounding popUlation parameters such as age and sex, environmental factors such as weather, natural heterogeneity of the population parameters influencing density such as birth, death, immigration, and emigration rates, and sampling variance all serve to mask the effects of density within a popUlation. The influence of these masking effects on test performance for detecting density dependence has not been thoroughly investigated. Therefore I selected 3 existing tests for evaluation through Monte Carlo simulation. Bulmer's (1975) autoregression method was selected because it is the classical method and is still being used today, Pollard et al.'s (1987) randomization test is currently the most popular method, and a parametric bootstrap likelihood ratio test recently developed by Dennis and Taper (1993). Each selected test was coded in SAS in preparation for the Monte Carlo simulations. The following is a description of the 3 tests and the results of the verification of the SAS code. Lastly, I will present the equations to be used to generate biologically realistic data for the Monte Carlo simulations. �212 DESCRIPTION AND CODING VERIFICATION OF BULMER'S (1975) TEST Bulmer (1975) proposed two autoregression methods for the detection dependence describing density independence as: of density (1) and density dependence as: (2) where Xt and Xt 1 are the natural logarithms of population size at time t and t+l respectively, €t'S are independent normal random variables with mean zero and variance a2• The parameter ~ is an equilibrium value and la I < 1. Thus, for a density-independent population there will be no deterministic trend in population size, nor will there be any te~dency for it to return to an equilibrium value. Alternatively, in the density- dependent model, population size will constantly return to the equilibrium value (Bulmer 1975) . + The models are then used to test the null hypothesis of density independence against the alternative model of density dependence based on observations N1, ••• Nn• The test statistic is: (3) R=V/U where: (4 ) n-a U = L (Xi+1 - Xi) 2 i=l (5) n V = L (Xi - X)2 i=l and (6) n X = L Xi / n i =1 The null hypothesis is rejected for small values of R. I coded Bulmer's test in SAS for known data sets (Appendix 1) and for generated data (Appendix 2). To verify my SAS code I ran 42 of ~he 59 data sets cited in Vickery and Nudds (1991). A paired t-test was conducted to test the hypothesis of no difference in results from my SAS code from the results reported by Vickery and Nudds (1991). �213 To complete the verification from a random walk model: of my code for Bulmer's test I generated data (7) where the ee's are independent random variables with mean zero and variance a2 = 0.333. Data were generated with initial values (NO) of 50, 100, SOD, and 1000 for 10, 25, 50, and 100 years. Each set of initial values was simulated 2500 times to compare with simulations run by Bulmer (1975) to verify a Type I error rate of 5% for R. A paired t-test was used to test the null hypothesis that Bulmer's (1975) results and results from my simulations were comparable. DESCRIPTION AND CODING VERIFICATION OF POLLARD ET AL.'S (1987) TEST Pollard et al. (1987) developed a distribution-free test for detecting density dependence from a series of annual population censuses. The rationale for their randomization test is to consider whether, using an appropriate test statistic, T, the observed set of Xi (Xi = log(Ni)) values should be judged as a significantly extreme arrangement when compared with a set of Xi values simulated from a sample of random arrangements of the di values (di = Xi•l - xJ. The procedure (1987) : for the randomization a) Use the observed values test-statistic, T. b) Either calculate the 1)! possible sets of and corresponding to simulated sets of Xi di test is as follows Xl' X2, X= from Pollard to compute et al. the value of the x) values and enumerate the (n or randomly permute the d; values, N such permutations, obtain a sample of N values (a Monte Carlo test) . = (Xi•l - Xi values, c) For each simulated set of Xi values, compute the test-statistic d) If less than 5% of the T values calculated under (c) are smaller than or equal to the computed T value under (a), then provisionally reject the null hypothesis of density-independence, at 5% level of significance. T. The randomization test was programmed in SAS for data from a series of annual censuses (Appendix 3) and for data generated from a density-independent or density-dependent model (Appendix 4) . TO verify the SAS code for Pollard et al.'s compared my results with the results of den longterm (11-29 years) population counts in compare the p-values from my simulations to (1989) a z-statistic was calculated to test proportions (Ott 1988) : (1987) randomization test I first Boer and Reddingius (1989) for 16 univoltine insect species. To those from Den Boer and Reddingius for differences in 2 binomial (8 ) z To further verify the SAS code I mimicked the simulations performed et al. (1987) for 3 types of data. Pollard et al. (1987) described by Pollard the �214 following family density-dependent of time data: series processes to define density-independent and (9) (10) (11) (P "* In these year i. variance 1987) . 1) equations, Xi denotes the natural logarithm of the population size in The terms ei are independent random variables with mean zero and equal to a2, while r and ~ are the model parameters (Pollard et al. Equation 9 describes a density-independent population displaying a random walk. Equation 10 also describes a density-independent population however the growth of the population has a trend (r). Density dependence is defined in equation 11 with the addition of the ~ term when ~ ¢ 1. Z-tests as described above were used to test for differences in P-values from the different simulations and a paired t-test was performed for each table. The 2 densi~y-independent models should not be sensitive to initial values (Xl). Therefore, to further test my code I used 2 different initial values (Xl = 1, Xl = 2). The number of the permutations done within the randomization test were first performed 100 times to match the simulations done by Pollard et al. (1987) and were repeated for Xl = 2 for 1000 permutations. The density-dependent model is sensitive to initial values. Whatever the initial value Xi' the population will eventually fluctuate around the expected value, therefore the results are dependent upon the choice of Xi (Pollard et al. 1987). Therefore, 3 different initial values (Xi = 1.0, 2.0, and 3.0) were used for the density-dependent simulations to verify the randomization test perform as expected. DESCRIPTION AND CODING VERIFICATION OF DENNIS AND TAPER'S (1993) TEST Dennis and Taper (1993) present a likelihood ratio test based on a discrete time stochastic logistic model to detect density dependence in univariate time series observations of population abundances. The null hypothesis is that the population is undergoing stochastic exponential growth, stochastic exponential decline, or random walk. The distribution of the test statistic under both the null and alternate hypothesis is obtained through parametric bootstrapping (Dennis and Taper 1993) . �215 Dennis and Taper (1993) use the following model to relate Nt+1 to Nt: (12) where Nt and Nt+1 are population size at time t and t+l respectively, a is a constant drift parameter, b is the slope of the linear function, a is a positive constant, and Z: is a random shock to the population growth rate (Zt - N(O, 1)) . This stochastic logistic model becomes a first-order nonlinear autoregression model when transformed to a logarithmic scale. By letting Xt = In N: the following model results: (13) Both Bulmer (1975) and Pollard et al. (1987) used the Gompertz model as the alternate hypothesis in their density dependence tests but have used different test statistics. The Gompertz model is as follows: (14) On a logarithmic scale this becomes: (IS) The difference between the 2 sets of equations is the use of Nt in the stochastic logistic model and the use of In Nt in the Gompertz model to define the rate of growth when b ¢ o. Thus, in the Gompertz model growth rate depends only logarithmically on population density. By contrast, in the stochastic logistic model the growth rate depends on the population size itself. This feature of the model structurally allows for stronger density dependence (Dennis and Taper 1993) . The parametric bootstrap likelihood ratio test (PBLR) distinguishes the stochastic logistic equation (Eg. 13). The cases form a series nested hypotheses as described by Dennis and Taper (1993): The simplest is Model He: a = 0, 3 cases of 3 0: b o. thus, (16) This is a random walk model with zero drift (a popu~ation density to the growth rate (b = 0). Modell is: He: a ¢ 0, b o. 0) and no feedback of of �216 thus, (17) Modell defines a stochastic exponential growth or decay, (b = 0) model with a positive or negative drift parameter Model 2 is the density-dependent Ho: a ~ 0, b ~ density independent (a ~ 0). model: o. thus, (18) with a positive or negative drift parameter (a ~ 0) and a density-dependent growth rate (b ~ 0). A one-sided test is often more appropriate where b < o. To test for Modell vs. Model 2 the likelihood ratio statistic (T12) is identical to the t-statistic used for testing whether the slope parameter is nonzero in a linear regression. However, this T12 does not have a Student's t distribution due to the time-dependence of the observations. Instead the distribution of T12 can be estimated from the data through parametric bootstrapping. Thus, the PBLR test can be conducted in SAS following steps (Appendix 5, Appendix 6) : 1. A time series of population (SAS Inst. Inc. 1987) using count data are input the as the variable 'N' . 2. The variable y is calculated 3. The T12 statistic as y=log(nt1/nt). is calculated using the following equation: 1 "2 q T12 = B2 (q-2) L (n:-l ~t=~1 __ ~ - fl) 2 _ qo~) which is equivalent to the t-statistic parameter when you regress y on n. used for the slope 4. To extract the T12 value from the PRoe REG output code developed by Dennis and Taper (1993). I repeated the 5. Calculate the mean of the y's (a) and the standard deviation of the y's (s) from the original data, q = the number of one-step transitions (number of observations - 1), and nO is the value of the first observation. 6. To perform and simnum perform. 7. Calculate the T12 values from each of the bootstrapped data sets and compare it to the T12 value from the true data. If the T12 value from the bootstrapped data is smaller than the T12 from the true data then record this as detecting density dependence in the population. the PBLR test retain the values in step (5), set b=O to the number of desired bootstraps you want to �217 8. Estimates The percent of bootstrapped T12 values lower than the true T12 is the P-value of the test. If the P-value is s 0.05 then the null hypothesis of density independence is rejected. of at by definition, b, and a2 for Model 0 are as follows: a 0 b 0 and (19) This ML estimate for a". Estimates 0 ln Lo(a2)/oa2 is found by setting of a, b, and a2 for Modell equal to zero and solving are as follows: b = 0 (20) (J~ 1 where (21) 1 q Y = - q :E Yc .=1 The ML estimates of a and a2 are found by setting 0 ln L1(a,a2)/oaand L1(a,a2)/oa2 simultaneously equal to zero and solving for a and a2• Estimates n represent Let n1• of a, b, and a2 for Model •• , the sample 2 are as follows: mean of the initial values nq_1: (22) 1 q n = 0 ln -:E n q t-1 .=1 of the transitions, no' �218 (23) q L (Y C -.y) (nC-1 - ill C=l q L (nC-1 - ill 2 C=l (24) (25) Dennis and Taper (1993) however, found a slight but detectable using the unbiased estimate of a2 defined as follows: advantage to (26) To verify my SAS code was performing the PBLR test correctly I first compared my results with the results of Dennis and Taper (1993) for the Yellowstone grizzly bear population, the northern Yellowstone elk population, the elk population in the central valley of Grand Teton National Park, and the 16 insect data sets listed by den Boer and Reddingius (1989). Because of problems to be discussed below, simulations were run using the uncorrected value of a2, the corrected value of a2, the SAS function NORMAL to generate a normally distributed pseudo-random variate and the SAS function RANNOR to generate a normal deviate. To compare the P-values from these simulations to those from Dennis and Taper (1993) paired t-tests were conducted. To further ver~Iy the SAS code I mimicked the simulations performed by Dennis and Taper (1993) to estimate sizes and powers of various one-sided densitydependence tests. Monte Carlo simulations were conducted to determine the size or power of the PBLR test of Modell against Model 2. Using a kno~~ set of parameters (a, s, q, b, and no) a time series of population densities were generated. This simulated time series was then subjected to the PBLR test exactly as if it were data from field observations (200 bootstrapped data sets were generated for each simulated data set as follows Dennis and Taper (1993)). For each set of parameters 1000 trials were conducted. Therefore, each set of 200 bootstrapped data sets provided a Pvalue. If the P-value was s 0.05 then the trail rejected the null hypothesis of density independence. The proportion of the 1000 trials that rejected the null hypothesis was then recorded. For data generated with b = 0 this proportion ~s the size of the test and should be close to the set alpha level (a = 0.05 for these simulations). For data generated with b ¢ 0 this proportion is the power of the test. The size and power values from my simulations were tested for differences from those in Dennis and Taper (1993) by the z-statistic. �219 EQUATIONS FOR GENERATING GROWTH DATA DENSITY-INDEPENDENT AND DENSITY-DEPENDENT POPULATION Data will be generated for evaluating Bulmer's (1975) test, Pollard et al.'s (1987) randomization test, and Dennis and Taper's (1993) test for detecting density dependence in natural populations. Standardized notation for all parameters used in the equations is summarized in Appendix 7 and where applicable follows that used in the DEAMAN Database Manager and Population Modelling Procedures (White 1992). The equations are separated into 3 primary groups, those generating deterministic population growth data, those generating stochastic population growth, and age and sex structured population growth. In general, the equations within each group increase in biological reality and thus complexity. Initial equations for each of these primary groups describes growth in a single independent parameter, r, with no sampling variance. The growth parameter r is then defined as a combination of birth rate (b), death rate (d), emigration rate (e), and immigration rate (i). Covariates such as density, sex, age, and temporal variation are added to each of the growth parameters. Stochasticity is introduced as demographic variation and heterogeneity to more readily mimic natural populations. Because all 3 tests use only population size counts (Nt'S) sampling variance will only be added to the estimates of population size. Further complexity will be introduced as the 4 growth parameters are defined as functions of each other. Age- and sex-structured population growth for 4 age classes (juveniles, yearlings, 2-yr. olds, and all individuals greater than 2 yrs. old) is described. All of the above confounding factors are added to these basic population equations for both density-independent and density-dependent growth. The equations will be used to generate time series of population estimates for 5, 10, 40, and 100 years. Initial population sizes will be 50, 100, 500, and 1000. Parameter values for b, i, d, and e will be determined from the literature for realistic values of big game species. Extreme values and values of parameters that may cause the tests to have low power will also be used to fully evaluate the tests for detecting density dependence. A time series of population estimates from density-independent growth will be generated from the equations described in the Results section below. The equations are presented in the following 3 categories: deterministic growth, stochastic growth, and age- and sex~structured population growth. Within each of these 3 categories density-independent and density-dependent growth will be defined and population densities will be determined with and without sampling variance. Population growth will first be individual growth parameters b, then be defined as functions of growth will be defined with age defined as r, then as the combination of the d, i, and e. The 4 growth parameters will possible covariates. Finally, population and sex structure. For all equations it is implicitly understood that the parameters Nt and Nt•, refer to numbe~s of individuals within a specified area. With this assumption we can then assume we are defining densities. RESULTS CODING VERIFICATION OF BULMER'S (1975) TEST The results of the comparisons were confusing to 3 decimal places and yet other comparisons as there were exact matches were off by a full integer out �220 (Table 1). I rechecked my data entry, verified the SAS code was doing what requested with answers from a spreadsheet and still had the same discrepancies. However, a paired t-test did not reject the null hypothesis that the results were comparable (£ = 0.3593). I Holyoak (1993) also reported using Bulmer's test, therefore I contacted him for his data sets and R-values for confirmation of my code. He sent me 3 data sets and his calculated R-value for each set. Comparison with my R-values indicated verification of my SAS code (Table 2, £ = 0.4226). Note that one of these data sets (Aleurotrachelus jelinekii) was also reported in Vickery and Nudds (1991) under Whitefly eggs with an R-value of 7.143 - Holyoak (1993) and I report R-values of 7.12065 / 7.12066 respectively. R-values from den Boer and Reddinguis (1989) were also compared (Table 3, £ = 0.8556). A paired ttest was used to test the null hypothesis support the null hypothesis that Bulmer's (1975) results and results from my simulations were comparable (Table 4, £ = 0.6376). CODING VERIFICATION OF POLLARD ET AL.'S (1987) TEST The results were comparable (Table 5). A paired t-test also confirmed comparable results (£ = 0.2734). The results from the simulations done with my SAS program were comparable for the random walk model (Table 6, £ = 1.000), the random walk model with trend (Table 7, £ = 1.000) and the densitydependent model (Table 8, £ = 1.000). Three different initial values (Xi = 1.0, 2.0, and 3.0) were used for the density-dependent simulations to verify the randomization test performed as expected (Table 8) . CODING VERIFICATION OF DENNIS AND TAPER'S (1993) TEST All paired t-tests support the null hypothesis of comparable results (Table 9, Dennis and Taper vs. SAS code uncorrected a2NORMAL £ = 0.9450, Dennis and Taper vs. SAS code corrected a2 RANNOR £ = 0.4282, Dennis and Taper vs. SAS code corrected a' NORMAL P = 0.9304, SAS corrected a2 NORMAL vs. SAS corrected a2 RANNOR £ = 0.2417). Although the P-values generated from my SAS code were not significantly different from those reported by Dennis and Taper (1993), all the SAS code P-values were higher than those of Dennis and Taper (1993, Table 10). The paired t-test conducted on the 16 sets of P-values in Table 10 was significant (Dennis and Taper vs. SAS uncorrected a2 NORMAL £ = 0.0107). Using the corrected a2 value increased the p-values which is the opposite effect it should have on the results. Thus, the SAS code is not comparable to what Dennis and Taper (1993) have done. Because the SAS code is working correctly when actual data sets are used in the program the bug would appear to be in the data generation step of my program. I will be pursuing this to complete verification of my SAS code for Dennis and Taper's test. EQUATIONS FOR GENERATING GROWTH RATE DENSITY-INDEPENDENT Deterministic growth population Density-independent ages, and years: growth with constant AND DENSITY-DEPENDENT growth rate (r) across POPULATION all sexes, (27) where NTt+1 is the total population at time t+l. size across all sex, age classes and years �221 Table code. 1. Comparison Species Winter eggs nhl from Vickery R-value Vickery and Nudds (1991) R-value mine 1.160 l.127 1.16039 Reason for missing R-value l.142 l.145 1.300 0.937 0.896 budworm 3.704 data not found the reference 7.143 7.194 10.526 10.417 10.204 7.407 7.12066 7.18501 10.5475 10.4168 10.1558 7.39093 Cinnabar moth eggs larvae I II larvae III-IV larvae V pupae adults 0.438 0.444 0.396 0.423 0.474 0.427 0.40747 0.40934 0.37364 0.39589 0.45287 0.42058 Pine looper eggs larvae I larvae II-III 0.589 0.571 0.624 0.57062 0.65367 larvae III-IV larvae IV-VI pupae (Dec) pupae (Apr) adults 1.195 0.707 0.705 0.847 0.681 2.710 2.717 2.755 2.053 2.183 l. 89 in data not reported in these age classes 0.61725 0.58911 Eye-spotted bud moth on MacIntosh E (1991) and from my SAS data not reported in these age classes Whitefly eggs larvae I larvae II larvae III larvae IV adults NL LPW LEH SL SLspa and Nudds moth lsC lsop lsdp psap adults Spruce of R-values 2.71285 2.75710 2.75621 2.05341 2.18385 1.89022 �222 Table Nudds 1 (cont.). Comparison of R-values (1991) and from my SAS code. Species from Vickery R-value R-value Vickery and Nudds (1991) mine 2.703 2.538 2.801 2.096 2.049 1. 791 3.04228 3.13702 2.49036 1. 90629 2.36671 2.13552 2.227 2.169 2.353 2.336 1.728 1.013 2.22808 2.16999 2.35466 2.33813 1.45687 1. 01362 1.406 1.178 0.963 1. 060 2.141 1.227 1. 40638 1.17790 0.96325 1. 06030 2.14110 1.22659 and Reason for missing R-value Eye-spotted bud moth on Delicious E NL LPW LEH SL SLspa Codling moth MacIntosh on E NL LEF LEFF LRCS LP On Delicious E NL LEF LEFF LRCS LP data from personal communication Ditropis 1. 931 1. 052 Spring Autumn data from personal communication Plodia Bran and glycerine Bran, glycerine, yeast and flour 1. 075 0.935 Great 1.016 data not recorded in the reference 1.309 data not recorded in the cited reference Blue Tawny tit tit owl 7.463 7.45821 �223 Table 2. Comparison of R-values R-values from my SAS code. from M. Holyoak R-value Holyoak R-value mine 1.42962 1.42962 Aleurotrachelus jelinekii 7.12065 7.12066 Andraca 0.67833 0.67833 Species Acleris variana bipunctata Table 3. SAS code. Comparison of R-values Species (personal from den Boer and Reddingius R-value den Boer and (1989) Reddinguis R-value mine Winter moth; larvae 1.160 1.15845 Winter moth; adults 0.952 0.95214 Pine looper; August larvae 0.620 0.61997 Pine looper; September larvae 0.626 0.62649 Pine looper; April pupae 0.809 0.80889 Pine looper; adults 0.581 0.58137 Pine looper; UK, 0.979 0.97934 pupae communication) and (1989) and my �224 Table 4. Comparison of simulation results from this study with simulation results from Bulmer (1975) to determine the relative frequency with which R was less than the lower critical value (R1) and greater than the upper critical value (R) Number of years (Nyrs) Initial value (NO) Lower Bulmer (1975) 10 ? 0.044 25 50 100 Lower this study Upper Bulmer (1975) Upper this study 0.049 50 0.0636 0.0608 100 0.0568 0.0476 500 0.0472 0.0588 1000 0.0540 0.0512 ? 0.056 0.053 50 0.0568 0.0524 100 0.0596 0.0436 500 0.0572 0.0516 1000 0.0428 0.0508 ? 0.056 0.056 50 0.0536 0.0552 100 0.0460 0.0532 500 0.0548 0.0544 1000 0.0440 0.0560 ? 0.048 0.051 50 0.0420 0.0624 100 0.0504 0.0528 500 0.0476 0.0488 1000 0.0568 0.0604 �225 Table 5. Comparison of P-values for Pollard et al.'s (1987) randomization test from den Boer and Reddinguis (1989) and the SAS code. Data were permuted 500 times in den Boer and Reddinguis (1989) and 5000 times in the SAS code. r is the number of times the permutated value of the test statistic fell below the true value of the test statistic. Species P-value r/501 (den Boer and Reddingius 1989) P-value r/5001§ SAS code (Appendix 1) Winter moth; larvae 0.415 0.37872 Winter moth; adults 0.274 0.2604 Pine looper; larvae August 0.100 0.09878 Pine looper; larvae 0.126 0.11118 Pine looper; pupae 0.246 0.23115 Pine looper; adults 0.060 0.04539 Pine looper; UK, pupae 0.557 0.55709 September April Garden chafer; Rydal Farm 0.611 0.63667 Garden chafer; Hawes End 0.158 0.18476 0.519 0.49130 Grey larch bud moth Spruce budworm; Plot G4 0.521 0.51050 Spruce budworm; Plot G5 0.289 0.28994 Viburnum whitefly; pop. 1 0.828 0.52430 Viburnum whitefly; pop. 2 0.896 0.90742 Viburnum whitefly; pop. 3 0.186 0.18096 Nebria brevi collis; adults 0.359 0.38472 § P-values significantly and Reddingius (1989). different (a = 0.05) from those reported in Den Boer �226 Table 6. The frequency distribution of the probability that the observed data set is from a density-independent population, based upon 200 data sets from the random walk model with 10 census points. Results are classified by the total displacement values (observed range) IXn - xII. Comparison of simulations from Pollard et al. (1987, Table 1) and 3 sets of simulations from my SAS code. T = Total Mean slope displacement b Pollard et al. (1987) x1=1, pe rme t.uo x2=2, pe rme t.o o x2=2, perm=1000 Pollard et al. (1987) x1=1, perm=100 x2=2, pe rme t.o o x2=2, perm=1000 Pollard et al. (1987) x1=1, pe rme f o o x2=2, pe rme t oo x2=2, perm=1000 Pollard et al. (1987) x1=1, perm=100 x2=2, perm=100 x2=2, perm=1000 Pollard et al. (1987) x1=1, perm=100 x2=2, perm=100 x2=2, perm=1000 Pollard et al. (1987) x1=1, perm=100 x1=2, perm=100 x1=2, perm=1000 < 0.1 ± SE+ T = b rdx Probability intervals+ 1 2 3 4 1 2 3 4 0.30 ± 0.0412(2)# 10 5 12 11 (2) 10 8 10 o .25 16 8 13 12 17 22 12 15 9 11 (4) 13(2) 14(2) 8(4) 20 13 18 17 7 7 15 16 8 13 19 9 13 17 10 11 8 17 10 10(0) 17(0) 11(5) 6(2) 15 12 0.1-0.2 ± 0.04 12 (3) 0.33 ± 0.0412(2) 0.35 ± 0.0313(2) 0.44 ± 0.049(2) ± ± ± ± 0.044(2) 0.0419(2) 0.0410(4) 0.0412(3) 11 10 11 11 18 0.2-0.3 0.51 0.35 0.45 0.62 ± ± ± ± 0.058(0) 0.049(1) 0.077(2) 0.0511(2) 17 12 8 8 14 13 5 12 0.3-0.4 0.58 0.58 0.57 0.69 7 8 >0.4 0.69 0.65 0.75 0.83 ± ± ± ± 0.058(3) 0.057(3) 0.045(1) 0.0412(2) 9 10 13 8 8 11 15 15 14 11 11 8(2) 14(2) 17 10 12 11 8 6 9 (3) 5§ 9 7 9 12 (2) 6 8 9 9 7 6 5 6 (0) 9 5 10 10 9 4 8 8 4 5 7 9 10 10 6 8 11 (7) 4§(3) 12 (2) 6 8 9 13 12 (3) 0.86 ± 0.038(3) 4§(1) 7 7 10 6 0.86 ± 0.057(1) 6 (2) 11 10 7 6 0.82 ± 0.047(2) TOTAL 56 (11) 48 48 48 49(12) 58 8 12 11 8 5 5 8 44 49 40§(11) 59 49 52 47(9) 59 45 49 54 (9) 52 48 46 59(12) 52 49 40 42 (11) 53 55 50 44(15) 55 52 49 Mean slope b is the average of the slopes of Xi+lon Xi in the given displacement class. + The probability interval values are 1:0-<0.25; 2:0.25-<0.50; 3: 0.50-<0.75; 4:0.75-1.0. # The numbers in parentheses show the number of cases when the null hypothesis of density independence is rejected at 5% level of testing. § p-values significantly different (a = 0.05) from those reported in Pollard et al. (1987) + �227 Table 7. The frequency distribution of the probability that the observed data set is from a density-independent population, based upon 200 data sets from the density-independent model 2 (Eg. 2) with r = 0.4, i.e. Xi+l = 0.4 + Xi + ei. Comparison of simulations from Pollard et al. (1987, Table 2) and 3 sets of simulations from my SAS code. Probability intervals' Simulation Procedur e used 1 2 3 4 Pollard (1987) rax 51(12)+ 47 59 43 b 53 (11) 48 55 44 rax 55(9) 52 45 48 b 49 (10) 50 43 58 rax 55(14) 48 49 48 b 53(16) 49 56 42 rax 55(13) 55 43§ 47 b 60(14) 44 48 48 xl= et al. 1, perm=100 xl=2, xl=2, perm=100 perm=1000 • Probability intervals as in Table 2. 0.05) = 10i observed number + Expected number of significant results (a given in parentheses. (a = 0.05) from those reported in Pollard § P-values significantly different et al. (1987) �228 Table 8. The frequency distribution of the probability that the observed data set is from a density-independent population, based upon 200 data sets from the density-dependent model (Eq. 3), with 10 census points. Comparison of results from Pollard et al. (1987, Table 3) and 3 sets of simulations from my SAS code. T=rdx Simulation Pollard (1987) et al. xl=l, perm=100 x1=1, perm=1000 Pollard et al. (1987) x1=2, perm=100 x1=2, perm=1000 Pollard et al. (1987) x1=3, perm=100 x1=3, perm=1000 probability Starting x + value of 1 1 2 3 4 164 (76)# 27 1.0 6 3 positive trend 146§ (71) 35 12 7 149§ (66) 35 14 2 54(12) 42 56 48 2.0 no trend 54 (20) 45 41 60 47(10) 60 53 40 159(55) 26 12 3.0 3 negative trend 149(68) 31 14 6 151(63) 29 10 10 T=b intervals' 1 2 3 4 162(52) 30 5 3 143§ (58) 43 156(53) 35 59(15) SO 11 7 46 3 2 45 53 (21) 44 51(12) 64 139 (38) 45 43 42 13 60 43 3 148(58) 152(48) 10 12 0 1 42 35 The expected value of x is 2 . • Probability intervals as in Table 2. # The numbers in parentheses show the number of cases when the null hypothesis of density-independence is rejected at the 5% level of testing. (ex = O. OS) from those reported in Pollard § P-values significantly different et al. (1987) + �229 Table 9. Comparison of values from Dennis and Taper (1993) and the SAS code for the likelihood ratio test statistic (T12), the number of 1-step transitions (q), and the P-values estimated by parametric bootstrapping. The first 3 data sets are from Dennis and Taper (1993), the 16 insect data sets were from den Boer and Reddingius (1989) Table 1. Dennis and Taper (1993) Grizzly bear -0.60 Elk; Northern-5.37 Yellowstone Elk; Grand -2.92 Tetons Winter moth; -1.84 larvae Winter moth; -2.54 adults Pine looper; -3.14 larvae August Pine looper; -3.07 larvae Sept. Pine looper; -1.51 pupae April Pine looper; -2.15 adults Pine looper; -2.52 UK, pupae Garden -0.914 chafer; Rydal Farm Garden -1.45 chafer; Hawes End -1. 29 Grey larch bud moth Spruce -1.61 budworm; Plot G4 Spruce -2.07 budworm; Plot G5 Viburnum +2.17 whitefly; pop. 1 Viburnum -0.622 whitefly; pop. 2 Viburnum -1.86 whitefly; pop. 3 Nebria -1.66 brevicollis; adults Species p q T12 SAS code Dennis and Taper (1993) 16 -0.604 -5.372 11 SAS code 16 11 Dennis and Taper (1993) 0.72 0.0025 SAS code SAS code SAS uncorr s corr s code normal normal corr s rannor 0.709 0.718 0.726 0.002 0.001 0.002 -2.925 22 22 0.044 0.042 0.048 0.046 -1.838 18 18 0.24 0.244 0.235 0.243 -2.544 18 18 0.089 0.094 0.087 0.089 -3.142 14 14 0.033 0.034 0.035 0.033 -3.071 14 14 0.033 0.036 0.033 0.038 -1.511 13 13 0.44 0.434 0.411 0.417 -2.146 13 13 0.15 0.172* 0.162 0.157 -2.516 12 12 0.095 0.100 0.095 0.093 -0.914 28 28 0.78 0.777 0.775 0.778 -1. 452 18 18 0.35 0.334* 0.343 0.350 -1.285 19 19 0.46 0.464 0.454 0.462 -1. 608 14 14 0.38 0.380 0.371 0.382 -2.074 13 13 0.20 0.206 0.196 0.197 +2.167 11 11 0.99 0.994 * 0.995 0.996 -0.622 11 11 0.76 0.759 0.766 0.768 -1.863 11 11 0.23 0.219 0.236 0.228 -1.663 10 10 0.40 0.394 0.396 0.398 P-values significantly different (a = 0.05) , assuming 8000 bootstrap samples for Dennis and Taper (1993), from those reported in Dennis and Taper (1993) . �230 Table 10. Comparison of estimated sizes and powers of various density dependence tests from Dennis and Taper (1993) and from Data are generated from the stochastic logistic models. Tests at a nominal significance level of 0.05 and used 200 bootstrap estimate was obtained from 1,000 trials. For all tests, q =9. b C1 a PBLR (logistic) SAS code' uncorr s normal SAS code corr s normal o o o o o o o o o o o o 100 0.25 0.5 0.039 0.056 0.056 100 0.25 1.0 0.051 0.057 0.063 100 0.25 1.5 0.045 0.060 0.063 100 0.50 0.5 0.052 0.064 0.051 100 0.50 1.0 0.050 0.052 0.063 100 0.50 1.5 0.046 0.052 0.046 100 0.75 0.5 0.041 0.065' 0.077 100 0.75 1.0 0.049 0.058 0.074 100 0.75 1.5 0.056 0.053 0.050 100 1. 00 0.5 0.046 0.064 0.065 100 1. 00 1.0 0.065 0.047 0.054 100 1. 00 1.5 0.054 0.060 0.053 -0.01 -alb 0.25 0.5 0.104 -0.01 -alb 0.25 1.0 0.362 -0.01 -alb 0.25 1.5 0.759 -0.01 -alb 0.50 0.5 0.152 -0.01 -alb 0.50 1.0 0.424 -0.01 -alb 0.50 1.5 0.807 -0.01 -alb 0.75 0.5 0.191 -0.01 -alb 0.75 1.0 0.494 -0.01 -alb 0.75 1.5 0.775 -0.01 -alb 1.00 0.5 0.237 -0.01 -alb 1. 00 1.0 0.562 -0.01 -alb 1. 00 1.5 0.788 The stochastic logistic P-values significantly and Taper (1993). § one-sided my SAS code. were conducted samples. Each model is different 0.115 0.387 0.383 0.468' 0.521 0.564 Nt+1 Nt exp (a + bNt + C1Zt) . (a = 0.05) from those reported in Dennis �231 Density-independent growth (r) may be a function of sex, age, and/or year: (28) r Density-dependent (sex, age, year) NTt growth: (29) r where r is a function (NTC' sex, of density (NTC)' age, year)) NTC sex, age, and year. growth where r = b + i - d - e. The parameter b is the across all years, i is the constant immigration rate is the constant death rate across all years, and e is the rate across all years: Density-independent constant birth rate across all years, d constant emigration (30) NTC+1 = (b + i - d - e) NTC Density-independent growth occurs when each of the 4 demographic parameters may be functions of sex and/or age. For example, birth rate may vary as a function of age and year: (31) bac = b ( age, year) where b is the mean per capita birth rate and bat is birth of age and year (yearling females at time t). Thus: rate as a function (32) NTC+1 = [b(age, d(sex, age, year) year) + i (sex, - e(sex, age, age, year) year)] Density-dependent growth where each of the 4 demographic functions of density, sex, age, and year: NTC parameters are (33) NTC + 1 = [b{NTC' d(NTC' where NTC is population sex, i age, time) + (NTc' sex, age, time) - e(NTt, sex, age, time) age, time)] NTC size at time t. The above equations define the true population size at time t+l. However, natural populations are rarely censused and only estimates of population size (N) are observed. Sampling variance occurs in any population parameter that is estimated from a sample of that population. To mimic sampling variance in any of the population densities above: �232 (34) where the Et'S are normal, independent variance cNTt+1 and 0 < C < 1. Stochastic population random variables with mean zero and growth The deterministic equations above are simplistic descriptions of growth in natural populations. Growth in natural populations is a stochastic process rather than deterministic. Stochasticity arises from process variation including spatial variation, demographic variation, and individual heterogeneity. Any combination of the process variances may occur in the 4 growth parameters: birth rate, death rate, immigration rate, and emigration rate. For example, let each female in the population have a multinomial probability of not giving birth, having 1 young, or having 2 young. The deterministic equations above allow the proportion p of females in the population to give birth to one offspring. The following stochastic equation introduces demographic variation in the birth rate by giving each female a multinomial probability of not giving birth, having 1 young, or having 2 young. This multinomial will be expressed as demographic stochasticity in the birth rate (DSB) . Thus: (35) (b (DSB) + i - d - e) Nrc Combining the demographic stochasticity of birth rate with the fact that birth rates are a function of age and year as described above: (36) (b (DSB, age, year) + i - d - e) NTt Demographic stochasticity in immigration rate (DSI) , death rate (DSD) , and emigration rate (DSE) are all binomial (e.g. either an animal dies or doesn't die). Therefore, density-dependent, stochastic growth would define N'!'t+l as: (37) N-:C + 1 = (b (NTt, DSB, age, d(N=-=, DSD, sex, To add sampling age, year) year) + i (NTt, DSI, ) - e(NTt, sex, DSE, sex, age, age, year) year) ) )) N-:t variance: (38) where the Et'S are normal, independent variance cNT"+land 0 < c < 1. random variables with mean zero and Heterogeneity of individuals within an age class will also be addressed. Heterogeneity varies growth parameters by individual animals. To model heterogeneity each animal in the simulated population would have its own distribution of values for the growth parameters. �233 Time lags can also be introduced in the density-dependent model: (39) Nn+1 = (b(Nn-L, d(Nn-L, DSB, age, DSD, sex, where L is the number growth. age, year) i + year) (Nn-L, ) - e(Nn-L, of years before DSI, sex, DSE, sex, the density age, age, affects year) year) ) )) N-:: the population As in the equations described for deterministic growth all of the above models for stochastic growth assume independence of the 4 growth parameters (b, d, i, e). This is not biologically realistic and models will be developed to better mimic natural populations. For example, if birth rates and death rates were dependent of each other then: (40) NTt+1 d(NTt, where Age- = (b (NTt, DSB, age, DSD, sex, age, year, dt)) year, bt)) dt and b: are death and birth and sex-structured population + - i (NTt, e(NTt, DSI, sex, DSE, sex, rates respectively age, age, year)) year))) at time N-:: t. growth Age- and sex-structured population growth has been described above in general by defining each of the demographic parameters as functions of sex and age (d(sex,age)). Below I outline specifically the age- and sex-structure of a population with 4 age classes. Deterministic, density-independent functions of age and sex: Number of juvenile females growth in time with the 4 demographic parameters t+l: (41) N JF t+1 Number of juvenile = [NYF :-1 males * * b2yo] * SRF * b2yO] * SRM by + N2YOFt+1 + N>2YOFt+1 b>2YO in time * t+l: (42) = [NYF :+1 N JM t+1 * by + N2YOFt+1 + N>2YOFe-r b>2YO * where by is the birth rate for yearlings, b2yO is the birth rate for 2 yr. olds b,2YO is the birth rate for individuals > 2 yrs. old, SRF is the female sex ratio, and SRM (SRM = 1 - SRF) is the male sex ratio. I Number (43) of yearling females in time t+l: as �234 Number of yearling males in time t+l: (44) Number of 2 yr. old females in time t+l: (45 ) Number of 2 yr. old males in time t+l: (46) Number of females> 2 yrs. old in time t+l: (47) = N>ZYOF t+l NZYOF t N>ZYOF t+l Number of males> * * (1 - (1 - * dZYOF) d>ZYOF) * (1 - (1 - 2 yrs. old in time + iZYOF eZYOF) e>ZYOF) + + i>2YOF t+l: (48) = N>2YOM t+l N2yOM t N>2YOM t+l Total population * * * (1 - d2yOM) (1 - d>2YOM) size at time * (1 - (1 - + i2YOM e2yOM) e>2YOM) + + i>2YOM t+l: (49) + Adding sampling variance N2yOM t+l + N>2YOF to the population + N>2YOM t+l t+l density: (50) where the Et'S variance cNTt+1 are normal, independent and 0 < c < 1. random variables with mean zero and Adding density dependence, covariates, environmental stochasticity (Eb~)' and demographic stochasticity the age- and sex-structured model is as follows: �235 Number of juvenile females in time t+l: (51) * NYF t+l + NJF c-a of juvenile * N2YOFt+l * + N)2YOF t+l Number by(NTt, males DSB, b2YO(NTt' year, €bt) DSB year, b)2YO (NTt, DSB, in time t+l: * DSB, €b:) year, * SRF * SRM €b:) (52) NYF t+l. + NJM t+l by(NTt, €bt) * b2yO (NTt, DSB, year, €b:) * b)2YO (NTt, DSB, year, €b:) N2YOFe-a + N)2YOF t+l year, where by is the birth rate for yearlings as a function of density at time t and year with demographic variation, b2yO is the birth rate for 2 yr. olds as a function of density at time t and year with demographic variation, b,2YO is the birth rate for individuals > 2 yrs. old as a function of density at time t and year with demographic variation, SRF is the female sex ratio, and SRM is the male sex ratio. Number of yearling females in time t+l: (53) NYF t+l = NJF t: (1 - eJF(NTt, Number of yearling * (1 - dJF(NTt, DSE, males year, DSD, in time €et)) year, * €=:)) + iJF(yea:, DSI, €it) t+l: (54) NYMt+l = NJM t: (1 - eJM(NTt, Number * (1 - dJM(NTt , DSD, DSE, of 2 yr. old females year, €et)) year, * €=:)) + iJM(yea:, DSI, €it) in time t+l: (55) N2YOFt+l = NYFt (1 - eyF(NTt, Number * (1- DSE, of 2 yr. old males dyF(NTt' year, DSD, in time €et)) year, + iyF(yea:, E=:)) DSI, t+l: (56) N2yOMt+l = NYMt * (1 - eYM(NTt, (1 - dYM(NTt , DSD, DSE, year)) yea:)) + iYM(DSI, * year) * €it) �236 Number of females> 2 yrs. old in time t+l: (57) N>ZYOF t+l = NZYOF t * * (1 - dzyoF(Nn, (1 - eZYOF(NTt' N>ZYOF t+l * * (1 - d>ZYOF(NTti (1 - e>ZYOF(Nn, of males> 2 yrs. old in time + year, year, year, Edt)) Eet)) Eit)) DSD, DSE, + i>ZYOF(DSI, Number year, DSE, year, + iZYOF(DSI, year, DSD, Edt))) Eet))) Eit) t+l: (58) N>ZYOMt+l = NZYOMt * * (1 - dZyOM(Nn, (1 - eZYOM(NTt' year, + iZYOM(DSI, N>ZYOM t+l * * (1 - d>ZYOM(NTt, (1 - e>ZYOM(Nn, DSE, + i>ZYOM(DSI, Total population size at time year, DSE, year, DSD, Eit)) + year, year, Ee:)) year, DSD, Edt)) Edt))) Eet))) Eit) t+l: (59) + NZYOM t+l + N>ZYOF t+l + N>ZYOMt'l Adding sampling variance to the population density: (60) where the Et'S are normal, independent and variance cNTt+1 and 0 < c < 1. Modelling the growth random variables with mean zero parameters Modelling of the 4 growth parameters (b, i, d, and e) will be an ongoing process throughout the simulations. In evaluating the existing tests for detecting density dependence it is most appropriate to begin with the simplest models of population growth. As biological reality (complexity) is added to these simpler models I predict they will become more and more unreliable. Therefore, I may only use a subset of the above described equations to evaluate these tests. However, the more realistic population models will be used in the testing or developing of new tests. �237 For the initial Sampling as: simulations: variance in the population density estimates has been defined (61) where the Et'S are normal, independent random variables with mean zero and variance cNTt+1 and 0 < c < 1. Therefore, the equation could be rewritten as: (62 ) taking the natural log: (63) where E is N - (O,a") and a = 0.OSNTt+1, 0.10NTt+1, and 0.20NTt+1• Birth rates and immigration rates will be bounded at (O,x) where x is greater than I, thus requiring continuous distributions of domain (0 s x s 00 ). Gamma distributions may be the most appropriate with mean a~ and variance a~" (~> 1). Death rates and emigration rates will be continuous variables o s x s 1. Therefore, beta distributions are appropriate. bounded by �238 REFERENCES CITED Bartmann, R. M., G. C. White, and L. H. Carpenter. 1992. Compensatory mortality in a Colorado mule deer population. Wildl. Monog. 121. 39pp. Berryman, A. A. 1978. Population cycles of the Douglas-fir tussock moth (Lepidoptera: Lymantriidae); the time-delay hypothesis. Canadian Entomologist 110:513-518. Bulmer, M. G. 1975. The statistical analysis of density dependence. Biometrics 31:901-911. Den Boer, P. J. and J. Reddingius. 1989. On the stabilization of animal numbers. Problems of testing: 2. Confrontation with data from the field. Oecologia 79:143-149. Dennis, B. and M. Taper. 1993. Density dependence in time series observations of natural populations: estimation and testing. Ecology (in prep). Eberhardt, L. L. 1970. Correlation, regression, and density dependence. Ecology 51:306-310. Eberhardt, L. L., R. R. Knight, and B. M. Blanchard. 1986. Monitoring grizzly bear population trends. Journal Wildl. Mngmt. 50:613-618. Gaston, K. J. and J. H. Lawton. 1987. A test of statistical techniques for detecting density dependence in sequential censuses of animal populations. Oecologia 74:404-410. Hassell, M. P., T. R. E. Southwood, and P. M. Reader. 1987. The dynamics of the viburnum whitefly (Aleurotrachelus jelinekii): a case study of population regulation. J. of Animal Ecology 56:283-300. Holyoak, M. 1993. New insights into testing for density dependence. Oecologia 93:435-444. Lomnicki, A. 1987. The density dependence debate. TREE 3:76. Mountford, M. D. 1988. Population regulation, density dependence, and heterogeneity. J. of Anim. Ecology 57:845-858. ott, L., 1988. An introduction to statistical methods and data analysis. Third ed. PWS-Kent Publ. Co., Boston, M.A. 835pp. Pollard, E., K. H. Lakhani, and P. Rothery. 1987. The detection of densitydependence from a series of annual censuses. Ecology 68(6):2046-2055. Reddinguis, J. and P. J. den Boer. 1989. On the stabilization of animal numbers. Problems of testing 2. Confrontation with data from the field. Oecologia 79:143-149. Royama, T. 1977. Population persistence and density-dependence. Ecological Monographs 47:1-35. SAS Institute Inc. 1987. SAS/Language guide for personal computers. Version 6 ed. SAS Inst. Inc., Cary, N. C. 429pp. Slade, N. A. 1977. Statistical detection of density dependence from a series of sequential censuses. Ecology 58:1094-1102. Solow, A. R. 1990. Testing for density dependence: a cautionary note. Oecologia 83:47-49. Strong, D. R. 1986. Density-vague population change. TREE 1:39-42. Turchin, P. 1990. Rarity of density dependence or population regulation with lags? Nature 344:660-663. Varley, G. C. and G. R. Gradwell. 1960. Key factors in population studies. J. of Animal Ecology 29:399-401. Vickery, W. L. and T. D. Nudds. 1991. Testing for density-dependent effects in sequential censuses. Oecologia 85:419-423. white, G. P. 1992. DEAMAN Database manager and population modeling procedures, Colorado Division of Wildlife user's manual and reference. 109pp. �239 Appendix~. SAS program to perform Bulmer's (~975) test for detecting density dependence with known population counts. libname library '.'; %macro bulmer; proc means noprint data=di; var Utemp xt; output out=distats sum=U data bulmer; set distats; Nyrs=62; R = css=udummy v/u; Ru = 0.25 + (Nyrs - 1)*0.461; Rl = 0.25 + (Nyrs - 2)*0.0366; BDD = 'no'; if R <= Rl then BDD = 'yes'; if R > Rl and R <=Ru then BDD keep Nyrs U V R Ru Rl BDD; run; %mend bulmer; data di; input bear raccoon; xt=log(bear); Utemp = (xt - lag(xt»**2; cards; 6003 2091 6342 1289 6261 1338 6262 1847 7205 1255 7484 1695 6331 1193 9266 1676 9346 1798 8182 1895 8130 2295 8922 1273 8144 2434 7474 3397 8214 3640 7571 3883 7878 1794 7337 3335 8931 4710 7603 11678 6920 21321 8661 4894 8420 1696 8589 3341 8569 4011 8172 3636 7431 3152 7120 7241 7804 2149 7543 1042 7415 514 7796 613 5951 15 8531 830 8021 538 11188 841 5515 354 10765 139 8386 124 8279 325 10080 250 9606 217 11719 153 10454 172 13419 171 11163 194 8968 218 9891 743 9685 575 10425 1642 10279 6466 10033 2916 10152 13544 8665 9177 1* 1* V; P=0.05 P=0.05 'inc' ; 8045 7267 6780 5131 5685 4750 4548 4532 1973 1024 718 404 281 602 243 141 proc print; run; %bulmer; proc print; run; from Table from Table 1 - Bulmer 1 - Bulmer 1975 1975 *1 *1 �240 Appendix 2. SAS program to perform dependence with simulated data. libname library'.'; 'omacro bulmer; proc means noprint data=di; var Utemp xt; output out=distats sum=U css=udummy V; data bulmer; set distats; Nyrs=&Nyrs; NO=&NO; R = V/U; /* P=0.05 Ru = 0.25 + {Nyrs - l)*0.46l; Rl = 0.25 + (Nyrs - 2)*0.0366; /* P=0.05 BDD = 'no'; if R <= Rl then BDD = 'yes'; if R > Rl and R <=Ru then BDD = 'inc'; keep Nyrs NO U V R Ru Rl BDD; proc append base=library.bulran data=bulmer; run; 'omend bulmer; 'omacro popsim; 'odo isim = l 'oto 2500 'oby l; 'odo N=l 'oto 4 'oby l; 50; 'oif &N l 'othen 'olet NO lOO; 'oif &N 2 'othen 'olet NO 'oif &N 3 'othen 'olet NO 500; lOOO; 'oif &N 4 'othen 'olet NO data ditot; NO = &NO; Nt = NO; lambda .; newpop = ., oldpop = .; keep NO Nt newpop oldpop lambda xt xtpl; do yrs = l to lOl; Ntpl = Nt + rannor{O)*0.577; oldpop = Nt; Nt = Ntpl; newpop = Ntpl; lambda = newpop/oldpop; xt = log (oldpop) ; xtpl = log{newpop); if yrsA=l then output; end; 'odo yrobs = l 'oto 4 'oby l; 'oif &yrobs l 'othen 'olet Nyrs lO; 'oif &yrobs 2 'othen 'olet Nyrs 25; 'oif &yrobs 3 'othen 'olet Nyrs 50; 'oif &yrobs 4 'othen 'olet Nyrs lOO; data di; set ditot (obs = &Nyrs); if &Nyrs= N then xtpl=.; Utemp = {xtpl - xt)**2; 'obulmer 'oend; 'oend; 'oend; proc freq data=library.bulran; tables BDD BDD*Nyrs BDD*Nyrs*NO; 'omend popsim; %popsim; run; Bulmer's from Table from Table (1975) test for detecting l - Bulmer l - Bulmer 1975 1975 */ */ density �241 Appendix 3. SAS program to perform Pollard et al. (1987) randomization test for detecting density dependence with population counts of viburnum whitefly, fourth ins tar larvae from den Boer and Reddingius (1989). /* Pollard et al.'s (~987) randomization test with /* data from annual censuses */ */ libname library'.'; \-macropollard; data pollard; \-letnsim = 5000; Uet Nyrs = ~2; array t{&Nyrs} t~-t&Nyrs array d{&Nyrs} d~-d&Nyrs retain nobs LERt23 LEb 0; retain t~-t&Nyrs d~-d&Nyrs; set di; by NO; Nyrs = &Nyrs; seed=O; if first.NO then do; nobs=O; end; nobs=nobs+~; t{nobs} = Nt; if Afirst.NO then d{nobs-~}=dx; if last.NO then do; link calc; Rt~2 = tM~2; Rt~3 = tM~3; Rt23 = tM23; do isim=~ to &nsim; do n = &Nyrs-~ to 2 by -~; j = int«n-~)*ranuni(seed»+~; ~{~~ : ~!jl~ d{j ~ = temp; end; do n = ~ to &Nyrs-~; t{n+~}=t{n}+d{n}; end; link calc; if tM23<= Rt23 then LERt23=LERt23+~; if b<= Rb then LEb=LEb+~; keep LEb Rb LERt23 Rt23; end; output; end; return; calc: m~=sum(of t~-t&Nyrs); m2=(m~-t~)/(&Nyrs-~); m~=(m~-t&Nyrs)/(&Nyrs-~); btop=O; bbot=O; ttop=O; tbot=O; do i=~ to &Nyrs-~; btop=btop+«t{i}-m~)*(t{i+~}-m2»; bbot=bbot+«t{i}-m~)**2); ttop=ttop+(t{i+~}-m2)**2; tbot=tbot+(t{i+~}-t{i})**2; end; b=btop/bbot; tM~2=(t(&Nyrs)-t(~»/sqrt(tbot); tM~3=(ttop-(b*btop»/tbot; tM23=(ttop-(b*btop»/(tbot-«t(&Nyrs)-t(~»**2)/(&Nyrsreturn; run; 'tmendpollard; data di; input NO N; xt = log(N); dx = xt - lag(xt); cards; ~ ~3~90 ~ ~937 ~ 3904 1 13919 ~ ~ ~ ~ ~ ~ ~ ~ 27562 34~36 3~1l0 50658 78943 ~64027 300583 330390 run; proc print; 'tpol1ard; proc print; �242 Appendix 4. SAS program to perform Pollard et al.'s (1987) randomization test to detect density dependence using generated data. This example generates data from a random walk model, see text for other models used. /* Pollard et al.'s (1987) randomization test with /* simulated data from a random walk model (see text) */ */ libname library I.'; \-macro pollard; data pollard; \-let nsim = 100; array t{&Nyrs} t1-t&Nyrs; array d{&Nyrs} d1-d&Nyrs; retain nobs LERt23 LEb 0; retain t1-t&Nyrs d1-d&Nyrs; set di; by xO; Nyrs = &Nyrs; seed=O; if first.xO then do; nobs=O; end; nobs=nobs+1; t{nobs} = xt; if Afirst.xO then d{nobs-1}=dx; if last.xO then do; link calc; Rt12 = tM12; Rt13 = tM13; Rt23 = tM23; Rb = b; Rdspl = dspl; do isim=1 to &nsim; do n = &Nyrs-1 to 2 by -1; j = int«n-1)*ranuni(seed»+1; ~{~ : ~ijJ~ d{j'j = temp; end; do n = 1 to &Nyrs-1; t{n+1}=t{n}+d{n}; end; link calc; if tM23 <= Rt23 then LERt23=LERt23+1; if b<=Rb then LEb=LEb+1; keep LEb Rb LERt23 Rt23 Rdspl; end; output; end; return; calc: m1=sum(of t1-t&Nyrs); m2=(m1-t1)/(&Nyrs-1); m1=(m1-t&Nyrs)/(&Nyrs-1); btop=O; bbot=O; ttop=O; tbot=O; do i=1 to &Nyrs-1; btop=btop+«t{i}-m1)*(t{i+1}-m2»; bbot=bbot+«t{i}-m1)**2); ttop=ttop+(t{i+1}-m2) **2; tbot=tbot+(t{i+1}-t{i})**2; end; b=btop/bbot; tM12=(t{&Nyrs}-t{1})/sqrt(tbot); tM13=(ttop-(b*btop»/tbot; tM23=(ttop-(b*btop»/(tbot-«t{&Nyrs}-t{1})**2)/(&NyrS-1»; dspl = t{&Nyrs} - t{1}; return; proc append base=library.pollard data=pollard; run; %mend pollard; %macro popsim; seed., 0; %do isim = 1 %to 200 %by 1; keep xO xt; \-let Nyrs = 10; do yrs ., 1 to &Nyrs; data ditot xtp1 ., xt + rannor(seed)*O.1; xO 2 output; xt = xO xt = xtp1; end; data di; set ditot; dx '"'xt-Iag(xt); %pollard \-end; %mend popsim; %popsim; run; �243 Appendix 5. SAS program to perform Dennis and Taper's (1993) parametric bootstrap likelihood ratio test for detecting density dependence from population counts of the Yellowstone grizzly bear population (Eberhardt et al. 1986) . /* /* Dennis and Taper's (1993) PBLR test with data from */ the Yellowstone grizzly bear population. */ libname library '.'; %-let NO=33; %-let q=16; data di; input oldn; y=log(oldn/lag(oldn»; n=lag(oldn); q=&q; cards; 33 36 34 39 35 34 38 36 37 41 39 51 47 57 48 59 64 run; %-macro dennis; proc reg data=di outest=Rtsest noprint; model y=n; run; proc means data=di vardef=N noprint; var n y; id q; output OUT=Rtsvarn var=varn me an =dummy1 a std=dummy2 s; data Rtdist; merge Rtsest(keep=n RMSE rename=(n=bhat» Rtsvarn(keep=varn RT=bhat*sqrt(q*varn)7(_RMSE_); run; data tseries; set Rtdist (keep=RT a s q); retain nt1 0 n 0 nO &NO b 0 simnum 8000; do group=1 to si~lum; do obser = 1 to q; if obser = 1 then n=nO; nt1 = n*exp(a+normal(O)*s+b*n); y = log (nU/n) ; output; n=nt1; end; end; run; proc reg data=tseries outest=tsest noprint; by group; model y=n; proc means data=tseries vardef=N noprint; var n; by group; id q RT; output OUT=tsvarn var=varn; data library.denbear; merge tsest(keep=n _RMSE_ rename=(n=bhat» tsvarn (keep=varn group q RT); T=bhat*sqrt(q*varn)/(_RMSE_) ; DDD = 'no I; if T >= RT then DDD='yes'; keep group q RT T DDD; proc univariate data=library.denbear var T; pctldef=1; q as); proc freq data=library.denbear; tables DDD; run; %-mend dennis; %-dennis �244 Appendix 6. SAS program to perform Dennis and Taper's (1993) PBLR test to detect density dependence using data generated from a stochastic logistic model. /* /* Dennis and Taper's (1993) PBLR test with from a stochastic logistic model. data generated */ */ libname library' .'; %macro trial; %do isim=1 %to 1000; %macro dennis; proc reg data=di outest=Rtsest noprint; model y=n; run; proc means data=di vardef=N noprint; var n y; id q; output OUT=Rtsvarn var=varn mean=dummy1 a std=dummy2 s; data Rtdist; merge Rtsest(keep=n RMSE_ rename=(n=bhat» Rtsvarn(keep=varn q as); RT=bhat*sqrt(q*varn)7(_RMSE_); run; data tseries; set Rtdist (keep=RT a s q); retain nt1 0 n 0 nO 100 b 0 simnum 200; do group=1 to simnum; do obser = 1 to q; if obser = 1 then n=nO; nt1 = n*exp(a+normal(O)*s+b*n); y = log (nU/n) ; output; n=nt1; end; end; run; proc reg data=tseries outest=tsest noprint; by group; model y=n; run; proc means data=tseries vardef=N noprint; var n; by group; id q RT; output OUT=tsvarn var=varn; run; data tdist; merge tsest(keep=n RMSE rename=(n=bhat» tsvarn(Keep=varn group q RT); T=bhat*sqrt(q*varn)7( RMSE ); DDD=O; __ if T <= RT then DDD=1; keep group q RT T DDD; run; proc means sum noprint; var DDD; output out=sum sum=sumDDD; data end; set sum; if sumDDD <= 10 then DD=1; else DD=O; proc append base=library.temp data=end; run; %mend dennis; data di; a=0.5; s=0.25; seed=O; b=-0.01; nO=-a/b; q=9; do obser = 1 to q; if obser=1 then n=nO; nt1 = n*exp(a+normal(seed)*s+b*n); y = log(nu/n); output; n=ntl; end; run; %dennis %end; %mend trial; %trial proc means sum data=library.temp; var DD; run; �245 Appendix 7. Notation and definitions for all variables used to generate density-independent and density-dependent data. Notation Definition r population growth b population birth rate, rate, constant constant birth rate of female yearlings, birth rate of yearlings birth rate of 2 year old females birth rate of females> across across all years all years constant across all years in year t 2 yrs. old d population definition death rate (see birth rates above of specific subscripts) e population definition emigration rate (see birth of specific subscripts) i population definition immigration of specific SRF female SRM male rates rate (see birth subscripts) rates for above for above for sex ratio sex ratio number of juveniles females in year t number of yearlings females in year t number of 2 year old females number of females NJMt number of juveniles males in year t NYMt number of yearlings males in year t number of males > 2 years old in year t number of males > 2 years old in year t N,2YOMt in the equations total population > 2 years in year t old in year t size for all age and sex classes in year size for all age and sex classes in year t total population t+l total estimated population classes in year t+l size for all age and sex ��247 Colorado Division Wildlife Research July 1993 of Wildlife Report JOB PROGRESS REPORT State of __~C~o~l~o~r~a~d~o~ _ Project No. __~W_-~1~5~3_-~R~ _ Mammals Research Work Plan No. _ Pronghorn Job No. 6 Period Covered: Author: ~3~A~ Pronghorn Winter Wheat Damage Study July 1, 1992 - June 30, 1993 D. C. Strohmeyer, Personnel: Research G. C. White, and R. B. Gill G. C. White, and R. B. Gill Abstract We had 2 segment objectives. The first was to mark wild pronghorn so that collection of detailed movement data could began in the winter 1993. Fifteen adult, female pronghorn were marked in the Pawnee National Grassland on 17 February 1993. Two of these animals moved to Wyoming while the rest have stayed in the Pawnee National Grassland. Two mortalities have occurred from poaching. The second objective was to design an experiment concerning pronghorn movement from winter wheat to native range. Several experimental designs have been submitted and are being reviewed. Approval of the telemetry section of the study plan is expected by November 1993. Choice of the design for the experimental manipulation is expected by September 1993. Approval of the design for the experiment is expected by July 1994. ��249 Colorado Division wildlife Research July 1993 of Wildlife Report JOB PROGRESS state of Project Colorado No. W-153-R-6 Work Plan No. Job No. Period Author: REPORT 5A 2 Covered: Mammals Research Black Bear Research Development Techniques of Black Bear Inventory July 1, 1992 - June 30, 1993 T. D. I. Beck Personnel: G. White, D. Bowden, L. Willmarth, C. Parmeter, S. Lechman, M. Caughlan, M. McLain, R. Hays, J. Beach Abstract A detailed study plan on the development of black bear (Ursus americanus) census techniques was approved. A research prospectus for testing hypotheses on protein metabolism during hibernation of black bears was developed. A crew was fielded in June 1993 to trap and tag black bear in a 180 mi2 study area on the Uncompahgre Plateau. Thirty-two bears w~re trapped between 21 June and 30 June. ��251 DEVELOPMENT OF BLACK BEAR INVENTORY Thomas TECHNIQUES D. I. Beck P.N. OBJECTIVE 1. Evaluate a capture-sight stations for estimating program utilizing cameras black bear density. 2. Document age and gender bias in vulnerability during autumn hunting seasons. 3. Obtain density estimates of black bears in 3 heavily areas of markedly different vegetation communities. SEGMENT set on bait of black bears hunted OBJECTIVES 1. Develop a detailed study plan in program narrative form on the development of techniques to estimate black bear density and population composition. 2. Acquire 1993. the necessary 3. Capture and tag black bears equipment METHODS to begin field studies in the selected in June study area. AND MATERIALS Study Plan Development Discussions were held with researchers in Montana who are currently evaluating a similar approach to estimating grizzly bear (Ursus arctos horribilis) density. An array of computer simulations were evaluated to examine the impact of proportion initially captured and resighting probabilities on population estimates for population sizes of 100-150. These simulations were prepared by Dr. Gary White, Colorado state University. A detailed study plan was prepared. The draft was submitted to the peer review committee, 3 regional biologists, Dr. Gary White, and Dr. David Bowden, contract statistician, for critical comment. Comments were incorporated into the program narrative. The final study plan was submitted to the Animal Welfare Committee for the Colorado Division of wildlife Terrestrial Research section for approval. Acquisition of Equipment All necessary field equipment was purchased or borrowed. The most significant procurement was the development of a new type of bear trap. A prototype was constructed by Jack Beach, Northwest Region �252 technician, based on initial designs by the author. The prototype was studied by a mechanical engineer, who provided numerous suggestions and detailed line drawings. Further modifications were suggested by the fabricator, Bill Welfelt, and were incorporated. The trap uses a spring-loaded side-swing door released by a treadle pedal. Walls are of wire mesh fabric attached to angle iron frame, which provides an open appearance. The fabricator was issued a bid for 44 traps. capture and Tagging General procedures are outlined in the Program Narrative. More detailed explanations will be provided in next years report when the entire 1993 summer field season will be summarized. RESULTS AND DISCUSSIONS Study Plan Development Final study plan was approved by all committees reviewers. copies are available upon request. and outside Acquisition of Equipment Delivery of traps was 2 weeks late, only allowing 10 trapping days in June. Dur ing deployment of the traps, 2 minor weaknesses in the door release system were discovered. These were corrected. A final detailed blueprint was provided by the fabricator. Two trailers were constructed for transporting the traps behind ATV's. Two ATV's were generously loaned to the project by Glenn Davis, Davis Tire & Service Center, of Montrose, CO. capture and Tagging During the first 10 days of trapping 32 black bears were captured in 248 trap nights. The new traps worked exceptionally well. There was one incident of human interference where someone vandalized a trap, causing us to miss a bear. No injuries to bears were observed. Handling of bears can be easily done by a single technician. Although the traps weigh 182 kg, they can be transported and set up by a single technician when using the ATV and trailer. Two workers can transport them in a conventional truck. More complete summaries of the trapping effort will be detailed in the next annual report when the entire summer season can be presented. Prepared by Thomas D. I. Beck wildlife Researcher �253 Colorado Division Wildlife Research July 1993 of Wildlife Report JOB PROGRESS REPORT State of ~C~o~l~o~r~a~d~o~ _ Project No. W-lS3-R-6 Mammals Research Work Plan No. 8A Small Carnivorous Job No. 1 Development of River Otter Reintroduction Procedures Period Covered: Author: Mammals Investigations July 1, 1992 - June 30, 1993 T. D. I. Beck Abstract All data on this project has been summarized but analyses and writing of guidelines and job final report were not completed during this segment. ��255 DEVELOPMENT OF RIVER OTTER REINTRODUCTION PROCEDURES Thomas D. I. Beck P. N. OBJECTIVE Develop procedures for river otter reintroductions in Colorado and establish a self-sustaining population of river otters from which to collect river otters for future translocations. SEGMENT OBJECTIVES 1. Develop guidelines monitoring. for river otter population 2. Prepare job final report. establishment and METHODS AND MATERIALS All data pertinent to river otter establishment in the Dolores River has been summarized. Analyses of this release, other Colorado reintroductions, and other North American introductions were incorporated in developing river otter reintroduction guidelines. RESULTS AND DISCUSSION Draft guidelines for population establishment and monitoring were discussed with regional biologists. Information on monitoring current release sites was prepared for regional biologists. The final guidelines were not completed during this segment because of conflicting commitments of time. While all pertinent data has been summarized, analyses have not been completed. Therefore, a job final report was not completed. Again, too many assignments, so little time; and perhaps a temporary decline in efficiency. Completion of these 2 items remains my highest priority for the next segment. Prepared by Thomas D. I. Beck Wildlife Researcher ��257 Colorado Division Wildlife Research July 1993 of Wildlife Report JOB PROGRESS State of Project REPORT Colorado No. W-153-R-4 Work Plan No. 9A 1 Job No. Mammals Research Elk Investigations Impact of Elk Winter Livestock Period Author: Covered: Grazing on Production July 1, 1992 - June 30, 1993 N. T. Hobbs, D. L. Baker, G. Bear, and D. Bowden Abstract In many areas of western North America, populations of Rocky Mountain elk (Cervus elaphus canadensis) avoid snow at high elevations during winter by migrating to sagebrush steppe communities in mountain valleys, communities that are used by cattle in the spring and early summer. As a result of these patterns of habitat use, the impact of elk on cattle has emerged as an important issue in policy and management throughout the West. We examined effects of variation in the population density of elk on the availability and use of forage resources by cattle an on cattle production in a randomized complete block experiment conducted in sagebrush steppe during 4 years. We manipulated elk numbers to achieve 4 levels of population density (0, 9, 15, and 31 elk/km2), replicated each level 3 times, and observed responses of vegetation and cattle vegetation to these manipulations. High densities of elk (31 animals/km2) annually removed 57% of the standing crop of dead perennial grass and 12% of the total annual production of live perennial grass. The proportion of the standing dead biomass of perennial grass removed by elk increased at a rate of 1.7% per unit increase in elk density (.E1.6 = 80.1, £ < .0001); utilization of the production of live grass (kg removed/kg net primary production) increased at about .3% per unit increase in elk density (.E'.6 = 6.4, £ = .04). As a result of these increased levels of use, standing crops of dead perennial grass declined in direct proportion to increasing e Lk density (.E,.6 = 10.0, ~ = .02) from a mean of 8.7 g/m2 in th$ controls to 3.3 g/m2 in the high density (31 elk/km2.) treatment. Early spring standing, crops of live perennial grass also declined as elk. population density increased, but these trends' only approached significance (.E'.6 = 3.2, £ = 0.12). We did not detect effects of elk grazing on primary production of perennial grasses or other live herbs. Canopy cover of shrubs was least and canopy cover of grass was greatest at intermediate levels of elk grazing (quadratic effect .E,.6 = 9.4, £ = .0;3). However, despite these shifts in canopy cover, the total supply of herbaceous dry matter available to cattle' declined in linear relation to elk density (linear effect .E,.6 = 5.0, £ = .07). We found weak enhancing effects of elk populations on nutritional quality of spring forage. Elk grazing caused linear increases in the digestibility (.E'.6 = 5.0, £ = .07) and nitrogen content (.E'.6 = 8.9, £ = .002) of the standing crop of perennial grass available to cattle. Improvements in forage quality resulting from elk grazing caused minor improvements in nutritional quality of cattle diets. Nitrogen content of cattle increased in the moderately grazed treatments (.E'.6 = 4.3, £ = .08), but dietary digestibility and fiber content did not change significantly. ��259 IMPACTS OF WINTER GRAZING BY ELK ON CATTLE PRODUCTION P. N. OBJECTIVES 1. To test the hypothesis that elk grazing during productivity and botanical composition of herbage ranges during spring. winter influences the on sagebrush grassland 2. To test the hypothesis that elk grazing during winter influences the body weights and rates of gain of cows and calves using sagebrush grassland ranges during spring. RESULTS Two mauscripts reporting results of tests of hypotheses were prepared for submisssion to Ecological Applications (see copies, attached). These manuscripts are in the final stages of internal review and will be submitted for publication in September, 1993. ��261 UNGULATE GRAZING IN SAGEBRUSH STEPPE I: MECHANISMS OF RESOURCE COMPETITION N. Thompson IMammals Research 2Statistics Corresponding Hobbs I, Dan L. Bakerl, George and David C. Bowden2 Section, Colorado Division Fort Collins, Colorado, Department, Colorado D. Bearl of Wildlife, 80526 USA State University, 80523 USA 317 W. Prospect, Fort Collins Colorado, Author: N. T. Hobbs Mammals Research Section Colorado Division of Wildlife 317 W. Prospect Fort Collins Colorado, 80526 USA Telephone: 303-484-2836 ext. 360 Fax: 303-490-2621 Internet: nthobbs@ariel.dowfc.colostate.edu Abstract. In many areas of western North America, populations of Rocky Mountain elk (Cervus elaphus canadensis) avoid snow at high elevations during winter by migrating to sagebrush steppe communities in mountain valleys, communities that are used by cattle in the spring and early summer. As a result of these patterns of habitat use, the impact of elk on cattle has emerged as an important issue in policy and management throughout the West. We examined effects of variation in the population density of elk (Cervus elaphus canadensis) on the availability and use of forage resources by cattle in a randomized complete block experiment conducted in sagebrush steppe during 4 years. We manipulated elk numbers to achieve 4 levels of population density (0, 9, 15, and 31 elk/km2), replicated each level 3 times, and observed responses of vegetation and cattle vegetation to these manipulations. High densities of elk (31 animals/km2) annually removed 57% of the standing crop of dead perennial grass and 12% of the total annual production of live perennial grass. The proportion of the standing dead biomass of perennial grass removed by elk increased at a rate of 1.7% per unit increase in elk density (ll,6 = 80.1, ~ < .0001); utilization of the production of live grass (kg removed/kg net primary production) increased at about .3% per unit increase in elk density (ll,6 = 6.4, ~ = .04). As a result of these increased levels of use, standing crops of dead perennial grass declined in direct proportion to increasing elk density (ll,6 = 10.0, ~ = .02) from a mean of 8.7 g/m2 in the controls to 3.3 g/m2 in the high density (31 elk/km2) treatment. Early spring standing crops of live perennial grass also declined as elk population density increased, but these trends only approached significance (ll~ = 3.2, ~ = 0.12). Effects of elk grazing on primary production of perennial grasses or other live herbs were not significant. Canopy cover of shrubs was least and canopy cover of grass was greatest at intermediate levels of elk grazing (quadratic effect ll,6 = 9.4, ~ = .03). However, despite these shifts in canopy cover, the total supply of herbaceous dry matter available to cattle declined in linear relation to elk density (linear effect ll,6 = 5.0, ~ = .07). We found weak enhancing effects of elk populations on nutritional quality of spring forage. Elk grazing caused linear increases in the �262 digestibility (,[.,6 = 5.0, ~ = .07) and nitrogen content (,[.,6 = 8.9, ~ = .002) of the standing crop of perennial grass available to cattle. Improvements in forage quality resulting from elk grazing caused minor improvements in nutritional quality of cattle diets. Nitrogen content of cattle diets increased in the moderately grazed treatments (,[.,6 = 4.3, ~ = .08), but dietary digestibility and fiber content did not change significantly with treatment. Daily forage intake by cattle (kg dry matter cow" d-·) declined in direct relation to elk density (,[.,6 = 6.3, ~ = .04), primarily as a result of reductions in intake of standing dead grass. Consequently, cattle daily intake of digestible energy (,[1,6 = 4. 8, ~ = .03) and nitrogen (g cow -1 d-·, ,[.,6 = 15.1, ~ =.008) declined as elk population density increased. The mechanism responsible for this decline was a Type II functional response of cattle to forage biomass. We conclude that effects of elk on cattle represent a composite of facilitative and competitive effects. When forage production is low and cattle density is high, competition is a much stronger force than facilitation is. In systems similar to those we studied, competition is likely to be operative when elk reduce total forage supply available to cattle below 40 g/m2 of live and dead herbaceous biomass. INTRODUCTION Patterns in the abundance and distribution of vertebrates have often been explained as the outcome of competition for resources shared among species (for review see Diamond 1978, Law 1989, Keddy 1989). However, ambiguities remain on the importance of competition as ecological force and on its mode of action (Grant 1972, Wiens 1977, Underwood 1986, Tilman 1987, Tilman 1989). These' uncertainties result in part from an insufficient opportunity to observe conditions where competition can be reasonably expected to occur. It should not be surprising that such conditions are rare--seminal theoretical work (MacArthur 1972:21) acknowledged that the circumstances necessary to produce competition may arise infrequently. This is the case in many ecosystems because environmental disturbances reduce populations of potential competitors to densities where demand for resources is small relative to resource supply (Wiens 1977, Huston 1979, Denslow 1985). In such systems, competition is likely to be weak and difficult to detect except during periods of unusual resource scarcity, that is, during "ecological bottlenecks" (sensu Wiens 1977). Managed grazing systems offer an opportunity to examine workings of competition between species of large herbivores under conditions where resources can be reasonably presumed to limit the growth of their populations. Contemporary grazing management in North American attempts to achieve equilibria between large herbivores and the plants they consume by adjusting stocking rates of herbivore populations (Holechek et ale 1989, Valentine 1990). In theory, such adjustments can maximize secondary production of large herbivores by preventing excessive inter- and intraspecific competition for forage over the short term, and by preventing harmful effects on plant production over the long term. Thus, secondary production in these systems should respond to changes in population densities of large herbivores if those herbivores compete for limiting resources. In addition, managed grazing systems are amenable to experimental manipulation and permit detailed observation of herbivore responses to those manipulations. Competition between populations of elk (Cervus elaphus) and cattle in the Western United States is widely believed to harm the production of cattle (Smith 1961, Wagner, 1978, Powell et ale 1986, Hogan 1990, Cool 1992, Williamson 1992). A particularly important interaction between elk and cattle occurs in sagebrush steppe communities of the Rocky Mountains during winter and spring. During winter, elk populations avoid snow at high elevations by migrating to mountain valleys (Sweeney and Sweeney 1984), areas where sagebrush is the predominant cover-type. During early spring, elk usually return to montane and alpine communities at higher elevation as snow recedes and green forage becomes available (Frank and McNaughton 1992). In contrast, �263 cattle are usually absent from sagebrush rangelands during winter, but depend on them for pasture during spring because rangeland able to support cattle grazing is in short supply at that time (Blaisdell et al. 1982, Powell et al. 1986). Pasture is scarce for cattle during spring because rangelands at high elevation fail to provide forage in a phenological condition appropriate for grazing cattle (Powell et al. 1986), while valley bottoms are largely committed to hay production. As a result, sagebrush steppe at mid-elevations offers spring forage for cattle that is substantially limited elsewhere. This limited availability of forage creates an ecological and economic "bottleneck" for livestock production. Use of sagebrush steppe by elk populations during winter and early spring is believed to reduce the quantity of forage available to cattle during spring (Smith 1961, Powell et al. 1986, Hogan 1990). Such reductions could occur as a result of the direct effects of consumption of senescent forage by elk during winter and from their consumption of live forage in early spring. In addition, excessive defoliation of young plants could reduce forage availability indirectly by retarding rates of plant growth during the remainder of the growing season. If standing crops of forage and forage production available to cattle decline as a result of winter grazing by elk, it is plausible that elk populations could compete with cattle, even though cattle and elk use sagebrush steppe at different times. Opposing these potential, competitive effects is the potential for a facilitative relationship between elk and cattle. Consumption of standing dead grass by elk during winter could enhance the nutritional quality of spring forage for cattle by shifting live/dead ratio in bunch grasses to favor live tissue (Willms et al. 1980a, Willms et al. 1981) and could increase grass production by removing litter (Allaye-Chan 1984, Willms 1986, but also see Sauer 1978). Such shifts could facilitate selection of high quality diets by cattle (Willms et al. 1980b). Moreover, shrubs can contribute a substantial portion of the winter diets of elk (reviewed by Kufeld 1972), and defoliation of shrubs during winter can increase production of grasses during spring by reducing competition for water and nutrients (Robertson 1947, Rittenhouse and Sneva 1976, but also see Wright 1970). Thus, browsing during winter might enhance conditions for grazing during summer (Laycock 1967, 1970, 1979). Resource competition can be partitioned into three, general processes: effects of consumption on resource supply, effects of changes in resource supply on resource acquisition, and effects of changes in resource acquisition on growth and reproduction. We conducted a manipulative experiment to examine how changes in elk population density affect these processes in sagebrush steppe. Here, we examine the influence of elk grazing during winter on forage available to cattle in the spring and on cattle forage intake. In a companion paper (Hobbs et al. submitted), we focus on the effects of elk on cattle growth and reproduction. METHODS AND MATERIALS Study Area We conducted experiments at the Colorado Division of Wildlife Little Snake Wildlife Management Area in northwestern Colorado (lat. = 40, long. = 108). Topography and climate are typical of the high, cold deserts of the Intermountain Sagebrush Steppe (West 1983). The area includes slopes, gullies, and flats ranging in elevation from 1800 to 2000 m. Aspects are predominantly southern and southwesterly with an average slope of 15 degrees. Soils are mostly sand and sandy loam with a rooting depth of about 100-120 cm and moderate to high permeability. Brown's Park sandstone forms the underlying substrate. Climate of the area is cold and dry; annual mean temperature is 6.06 C and mean annual precipitation is 27.5 cm, about two thirds of which usually falls as snow. The growing season averages only 81 days. Winter snow depths are highly variable, but accumulations of 10-30 cm are common on level ground during December through March. Wind scours snow from ridge tops, depositing it on lee slopes. Growing season (May - July) �264 precipitation during our study was slightly below the annual average of 5.9 cm (Table 1). Vegetation is representative of the Wyoming Basin Province (McKell and Garcia-Moya 1989). Big sagebrush (Artemisia tridentata) dominates the overstory; other important shrubs include rabbit brush (Chrysothamnus vicidiflorus, Chrysothamnus nauseous), horse brush (Tetradymia spinosa), and snake weed (Gutierzia glomerella). Predominant grasses are needle and thread (Stipa comata), western wheatgrass (Agropyron smithii), Indian Junegrass (Koleria cristata), bluegrass (Poa spp.), Indian ricegrass (Oryzopsis hymenoides) and cheatgrass (Bromus tectorum). Important forbs include wallflower (Erysimum asperum), peppergrass (Lepidium perfoliatum), silver lupine (Lupinus argenteus), and scarlet globe mallow (Sphaeralcea coccinea). Our study area was protected from livestock grazing during the five years previous to the start of our work. Condition of the range at the beginning of our studies was good-excellent. Experimental Design We examined responses of forage and cattle to 4 levels of elk population density (0, 9, 15, 31 animals/km2) in a randomized complete block experiment with 3 replications. We repeated the experiment annually during four consecutive years (1987 through 1990). Experimental units consisted of 12 fenced pastures constructed specifically to implement our research design. Each pasture was 32 ha in area and triangular in shape (Fig. 1). We acknowledge that interactions between cattle and elk in nature are not controlled by fences as they were in our studies. However, we were willing to sacrifice the realism offered by free-ranging animals to achieve the experimental control provided by constraining their movements. Our experiment was specifically designed to examine the case where elk and cattle use the same habitats and, hence, our findings will not apply to cases were habitat segregation is strong. Treatments were allocated to experimental units as follows. Four pastures were assigned to each of 3 blocks on the basis of pretreatment biomass of perennial grasses estimated by harvesting and weighing thirty .25 m2 plots in each pasture on June 1 of the year previous to initiating the experiment. Thus, there was one block with low biomass of perennial grass, one with medium biomass, and one with high. Treatment levels (0, 9, 15, 31 elk/km2) were randomly assigned within each of the three blocks. stocking Rates of Elk and Cattle We varied elk population density by stocking pastures with different numbers of elk during winter and early spring. Controls were pastures that contained no elk (Fig. 1). Treatment pastures were stocked with 3 elk to achieve a density equivalent to 9.3 animals/km2 (5.1 ha/animal-unit-month), 5 elk to achieve 15.4 animals/km2 (3.04 ha/animal-unit-month) and 10 elk to achieve the 31.4 animal/km2 (1.5 ha/animal-unit-month) (Fig. 1). Treatment levels were chosen to reflect a plausible range of elk population densities. Elk populations in northwest Colorado typically average about 5-12 animals/km2 of winter range. We chose one level (9 elk/km2) to represent this large scale average. However, elk densities can be much hi~her at finer scales. Thus, we choose the highest treatment level (31 elk/km) to mimic local concentrations of animals and to examine forage and cattle responses to extraordinary levels of elk grazing. We acknowledge that the 31 elk/km2 treatment is an unusually high concentration of animals; however, densities of some non-hunted elk populations on sagebrush steppe winter ranges are known to attain such densities (e.g, Houston 1982: Fig. 4.1, Table 4.3). All elk introduced into pastures were females> 2 years old. Each year, animals were trapped from the surrounding area in portable corral-traps baited with alfalfa hay. Average date of release of elk into the experimental pastures was December 27. Elk were held in pastures until approximately April 15 when they were released to the surrounding rangeland. �265 Because our objectives were to examine effects of elk on vegetation in sagebrush steppe that was also grazed by cattle, pastures were stocked with cattle immediately after elk were removed. Stocking rates of cattle were chosen to achieve about 50% removal of the net annual aboveground production of perennial grasses in the control pastures. Although this stocking rate is typical for sagebrush steppe on public and private land in the region, it should probably be considered heavy rather than moderate grazing by cattle (Laycock and Conrad 1981). Cattle- stocking rates were adjusted by changing the duration of cattle grazing each year to achieve annually similar levels of forage utilization in the control pastures despite annual fluctuations in forage production. Pastures were stocked with cattle for 6 weeks during years 1 and 2 (May 8-June 18 1987 and May 12-June 21 1988, 2.8 ha/animal-unit-month), for 3 weeks during year 3 (May 10-May 31 1989, 6.9 ha/animal-unit-month), and for 5 weeks during year 4 (May 9-June 13 1990, 3.3 ha/animal-unit-month). Hereafter, we will refer to the time interval that cattle were stocked in the pastures as the "spring grazing season." Details on management of cattle are reported in Hobbs et al. (submitted). Primary Production, Forage Intake, and Utilization Field Measurements--. We estimated primary production, rate of forage intake by cattle, and utilization of herbaceous forage by elk and cattle using 480 pairs of movable 0.55 m2 plots (Klingman et al. 1943). Forty plots were randomly placed in each of the 12 pastures. Grazing was excluded from one member of each paired plot by enclosing it with a movable, cone-shaped cage (diameter = .83 m height = 1.2 m) constructed of heavy-gauge, concrete reinforcing wire (mesh size = 10 cm). Cages were staked to the ground with 64 cm lengths of steel rebar. Although cages like the ones we used are known to influence primary production (Daubenmire 1940, cowlishaw 1951, Williams 1951, Owensby 1969, Sharrows and Motazedian 1983, Parsons et al. 1984), we assumed that such influences were constant across all treatment levels. Thus, by using the same methods to estimate production in the control as well as the treatment pastures, we obtained reliable estimates of the relative effects of elk grazing on production, even if our absolute estimates of production were biased by spurious effects of cages. Plot pairs were initially established during september before the first study year. Locations of plots were chosen by gridding a map of each pasture into 10 x 10 m cells, numbering all grid intersections, and randomly choosing (without replacement) among numbered locations. Plots were located in the field by using compass bearings from known locations and pacing required distances. After locating one plot randomly, a second plot was subjectively chosen within a 30 m radius to mimic the standing crop and vegetative composition of the initial plot. Both plots were marked with a 25 cm surveyor's stake and one was randomly chosen by coin flip to receive the cage. This process was repeated for each of the 40 pairs in each pasture. Aboveground, herbaceous biomass was harvested from each plot on May 1, June 1, and July 1 of each study year. On each sample date, plots were clipped and moved. New plot pairs were subjectively chosen to mimic the grazed condition in the previous open (uncaged) plot. After the last sample date, plot locations for the next study year were chosen randomly within a 30 m radius of the last established plot. Plots were harvested as follows. All herbaceous vegetation contained within a steel hoop (.75 m diameter) centered on the plot stake was clipped to stubble height (about 1 cm) and sorted into 3 categories (perennial grass, annual grass, forbs) in the field. After returning clipped samples to our facilities in Fort Collins, they were dried at 55 C for 48 hours and were subsequently separated by hand (many hands) into live and dead components. We then weighed each of the 6 categories (the 3 above x live and dead) to the nearest .01 g. To increase precision, we lumped data into three classes, live perennial grass, dead perennial grass, and other live herbs. The "other live herbs" category included pooled data for annual and perennial forbs and annual �266 grasses. Standing dead annual reasonable level of precision. grass and forbs were too rare to measure at a Estimating forage production, supply, and offtake--. Aboveground net primary production of each forage class was estimated as the sum of positive increments in live standing crop biomass within caged plots observed on the 3 sample dates. Total biomass of herbaceous forage available to cattle during the spring grazing season was estimated as the sum of the annual production of live tissue and the residual biomass crop of standing dead that remained after elk grazing. Energy supply in herbaceous forage· was the total digestible energy contained in the residual dead forage and the live produced after elk were removed from pastures; nitrogen supply was estimated similarly. Forge intake rate was estimated as the sum of the differences between grazed and ungrazed plots on each sample date weighted by the number of cows in a pasture (Van Der Kley 1955). Using differences in biomass between caged and uncaged plots tends to overestimate forage consumption by large herbivores because such differences include effects of trampling and wastage (Van Der Kley 1955). Moreover, we must assume that such effects are constant across treatments, which may not be the case (Allison et al. 1983). Nonetheless, all techniques for estimating intake rate of herbivores involve difficult assumptions (e.g., see review of Van Dyne et al. 1980), and we believe (as do others, e.g., Meijs 1986, Mitchell et al. 1986, Frank and McNaughton 1992), that paired plots provide reasonable estimates of average daily intake given large sample sizes like those in our study. We estimated digestible energy intake rate as the product of daily dry matter intake and digestible dry matter coefficients, using a constant of 18.4 kjoules/g to convert dry matter to energy. Nitrogen intake was calculated similarly. Utilization was calculated as the ratio of dry matter removal to production for live forage categories, and as the ratio of removal to the ungrazed standing crop for standing dead. Forage Nutritional Quality Subsamples of material in each forage class were composited to form a 20 g sample for each pasture. We ground samples to pass a .5 mm screen and analyzed them for total dry matter, ash, nitrogen content, and in vitro dry matter digestibility. Dry matter and ash were determined gravimetrically. Total nitrogen was determined using Kjeldahl procedures (A. O. A. C. 1980). Dry matter digestibility was estimated using two-stage in vitro procedures of Tilley and Terry (1963) as modified by Pearson (1970). We obtained rumen inocula for these procedures from a fistulated Holstein cow fed native grass hay. Canopy Cover Pretreatment measurements indicated that sampling effort required to estimate biomass of shrubs was prohibitive. However, we were interested in the effects of elk on the balance between and shrubs and grasses in sagebrush communities and were also interested in the effects of elk on the availability of palatable shrubs to cattle. To address these questions, we estimated percent cover of herbs and shrubs based on intercepts of canopies (Canfield 1941) along 25, 12-m line transects in each pasture during July 6-11 of each study year. Locations of transects were established during year 1 and were maintained throughout the study. Transect locations were chosen by gridding a map of each pasture into 10 x 10 m cells, numbering all grid intersections, and randomly choosing without replacement among numbered locations. Transect endpoints were marked with steel pins (1 x 30 cm). In addition, a 1 m wooden surveyor's stake (painted high visibility orange) was placed 10 m south of a transect endpoint to facilitate locating pins. After locating pins, we stretched a surveyor's tape between transect endpoints and read intersections of plant canopies with the tape to the nearest 1 cm. �267 Cattle Diet Composition and Quality During the last two years of the study (1989, 1990), we examined effects of elk grazing on nutritional quality and botanical composition of diets of nine cows fistulated at the esophagus. Grazing trials were separated into two periods; early and late spring. Early spring trials were conducted 5 days before introducing resident cows into experimental pastures (May 2 in 1989; May 5 in 1990) and late spring trials were conducted 5 days after resident cows were removed from pastures (June 23 in 1989; June 19 in 1990). Fistulated cows were randomly assigned to treatments within each block. One block of treatments was sampled each day so that by the end of the grazing trial all groups of cows experienced all treatments. Between grazing periods, fistulated cows were held in an adjacent pasture resembling experimental pastures. Supplemental feed was provided between sampling periods in order to maintain condition of fistulated animals. Pasture forage was sampled one day after cows were introduced to treatment pastures and every day thereafter for three days. Between 0600 and 1100, each group of cows was gathered and fitted with screen-bottomed collection bags attached to fistulas. They were then lead to a random starting location and allowed to graze freely for 30-40 minutes. Following these collections, an aliquot of each extrusa sample was frozen in a plastic bag and stored for later analysis. At the end of each morning of grazing, cows were gathered and moved to the block of pastures to be sampled the next day. This protocol was repeated each day until all pastures were sampled. Cows were not fasted before extrusa collections. Esophageal samples were freeze-dried and ground in a Wiley mill through a 1-mm screen. Samples were subdivided and one portion was used to determine botanical composition and the other portion was subjected to chemical analyses. Botanical composition of diets was estimated by microhistological examination of esophageal masticate samples (Composition Analysis Laboratory, Department of Range Science, Colorado State University, Fort Collins, Colorado). Ten systematically located fields per slide and five slides per sample (individual cow samples) were examined at 100X magnification. Extrusa samples were analyzed for dry matter, ash, and nitrogen by standard procedures (AOAC 1980). Neutral detergent fiber, acid detergent fiber, and acid detergent lignin were determined using sequential analysis according to procedures of Goering and Van Soest (1970). Digestibility of organic matter was estimated following Van Soest et al. (1987). Rumen inoculum was taken from a rumen fistulated Holstein cow fed high quality grass hay. Statistical Analysis Responses to elk grazing were examined using a randomized complete block analysis of variance with a repeated measures structure. Repetitions over years were treated as within subject effects using a multivariate approach (Cole and Grizzle 1966, Gill and Hafs 1971). Values for individual plots and transects were treated as subsamples and were pooled within pastures before analysis, thereby providing one data value, after averaging, for each pasture/vegetation response category for each year. Covariance using pretreatment observations of forage biomass failed to significantly reduce experimental error in post-treatment responses, largely because the range in pretreatment values was substantially less than the range in values post-treatment and because our blocking presumably reduced the effect of pretreatment variation. Consequently we did not use pretreatment biomass as a covariate • .Annual rainfall could be used as a covariate to reduce variation among years, but it would not be useful in reducing experimental error among blocks because rainfall was presumably quite similar among pastures within years. We wished to portray the magnitude of the year effect relative to the magnitude of the effect of treatment, so we did not adjust data using rainfall as a covariate. However we did separate the effects of rainfall from the effect of year using regression, and we report that analysis separately from the analysis of variance. �268 We tested a priori hypotheses using planned, single degree of freedom contrasts (Maize and Schultz 1985, Toothaker 1991). In our subsequent discussion of contrasts, the following terminology will be used. The control mean (~o) will refer to the true mean of the response at level 0 (elk/km). A treatment mean (~I) will refer, generically, to the true mean of the response of any of the 3 non-zero elk density levels or to the mean of the 3 treatment levels. We examined each of 4 contrasts identified as a control mean minus a treatment mean. We also used contrasts to test for linear and quadratic changes in response means with increasing elk density. To achieve a reasonable compromise between the probability of a Type I error and the power of our tests (Bransby 1989), we chose critical values of F at Q = .10 for all contrasts. However, we report the calculated significance of individual test statistics to allow the reader to use an alternative significance level if desired. Failure to reject a null hypotheses of no treatment effect does not infer that no effect exists. Further, finding a statistically significant result does not imply that the result is biologically significant. This creates ambiguity in interpretation of non-significant results when analyses are reported as simple hypothesis tests. Interpreting non-significant results has caused problems in the study of competitive interactions (Rotenberry and Wiens 1985, Connor and Simberloff 1986). To deal with these problems, we report 90% confidence intervals on the magnitude of the effect of treatment for all responses (Toothaker 1991). Confidence intervals on effect size should be interpreted as follows. The effect size is the value of the contrast of interest, that is the true treatment mean subtracted from the true control mean (~o - ~I). The estimated effect size is the value of the contrast when the estimated means (Xo' XI) are used to replace the true means (i.e., Xo - ~). Thus, an effect size whose value is positive (Xo - ~ > 0) indicates that that the control mean exceeds the treatment mean; similarly, a negative effect size (Xo - ~ < 0) indicates that the treatment mean exceeds the control mean. It follows that the sign of the upper and lower bound of a confidence interval on the estimate of effect size reveals whether treatment effects, as examined by the contrast, were significant and in what direction. When the lower bound is negative and the upper bound is positive (i.e., when the interval overlaps 0), effects of treatment, as examined by the contrast, are not significant at ~ < .10. However, in this case we stress that although repeatable differences between treatment and control means were not detected, differences as large as those indicated by upper and lower bound of the confidence interval are as plausible as a zero difference. Alternatively, when effect confidence intervals fail to include zero, treatment effects, as examined by the contrast, were significant at the Q = .10 level. If both bounds are negative, then the treatment significantly exceeded the control. If both bounds are positive then the control exceeded the treatment. We performed all statistical analyses using the SAS System for General Linear Models (Freund et ale 1986) and the SAS Interactive Matrix Language. RESULTS Effects of Elk Grazing on Forage Supplies Elk Grazing Intensity--. On average, elk stocked at 31 elk/km2 annually removed 57% of the standing crop of dead perennial grass and removed 12% of the annual production of live perennial grass. Utilization of standing dead perennial grass increased linearly at a rate of 1.7% per unit (elk/km2) increase in elk density (Fig. 2A, linear contrast £:.,6 = 82, .f < .0001); utilization of live grass increased at .3% per unit increase in elk density (Fig. 2B, linear contrast £:.,6 = 6.5, .f = .04). Although average utilization of other live herbs by elk differed from the control (where there was no elk grazing, Fig. 2C, control vs others £:.,6 = 4.5, .f = .08), linear effects were weak (linear contrast £:.,6 = 3.2, .f = .13). �269 Elk utilization of live and dead perennial grass declined significantly with year (minimum ~ = .006) but the magnitude of the effect of elk density on utilization rates did not depend on year (maximum ~ = .39). Herbaceous Standing Crop Biomass--. Grazing by elk reduced standing crops of forage available to cattle at the beginning of the spring grazing season. Averaged across years, standing crops of dead perennial grass declined in direct proportion to elk density (Fig. 3A, linear effect £:1,6 = 10.0, ~ = .02) from a mean of 8.7 g/m2 in the controls to 3.3 g/m2 in the high density (31 elk/km2) treatment. We also observed linear trends in effects of elk on the early spring standing crop of live perennial grass, but these trends only approached significance (Fig. 3B, linear effect £:1,6 = 3.2, ~ = .12). Effects of elk population density on the early spring standing crop of other live herbs (annual grasses + forbs) were not significant (main effect £:3,6 = 0.91, ~ = .49, Fig. 3C). Year effects on standing crops were strong for all forage categories (maximum ~ < .05 Fig. 3 A-C). The strength of the effect of elk on standing crops of dead perennial grass diminished as the study progressed (year x linear effect L,4 = 7. 9 ~ = .04). Year x linear effect interactions were not significant for live forage (minimum ~ = .14). The strong year x linear effect interaction for dead grass resulted in large part because the spring standing crop of dead grass aprroached 0 in all pastures during the final study year. Herbaceous Primary Production--. Effects of elk grazing on above ground net primary production (ANPP) of perennial grasses were not significant (main effect £:3,6 = 1.9, ~ = .38, Fig. 4A). Confidence intervals on the effect size of the 31 elk/km2 treatment ranged from -.6 to 4.9 g/m2. Thus, plausible values ranged from an enhancing effect on primary production of .6 g/m2 to a reduction of 4.9 g/m2 when comparing mean primary production at the 31 elk/km2 level to that of the control. Similarly, we can be 90% confident that the effect size at the 15 elk/km2 level was in the range -2.4 to 3.2 g/m2. Effects of elk grazing on other ANPP of other live herbs were substantially more variable than effects on perennial grass, but we can be 90% certain that elk grazing, averaged across treatment levels, did not reduce ANPP of live herbs by more than 3.3 g/m2 (Fig. 4B). Total herbaceous ANPP did not change with treatment (main effect L,6 F .23, ~ = 0.90). Effects of year on total herbaceous ANPP were strong (L,4 = 196, ~ < .0001) but year effects did not interact with the effects of treatment (£:9,18 = 1.1, ~ = .42). Annual variation in rainfall accounted for 57% of the year main effect (Fig. 5) and the effect of rainfall on ANPP was not modified by the elk grazing treatment (difference among slopes, minimum ~ = .28, Fig. 5). To examine the year effect after the effects of rainfall were removed, we regressed the residuals of the ANPP vs. rainfall regression on year (Fig. 6). This regression showed a consistent downward trend and residuals were predominantly negative during the fourth study year, indicating that rainfall progressively underestimated ANPP as the study proceeded. Treatment did not affect this trend (difference among slopes, minimum ~ = .58, Fig. 6). This suggests that the harmful effects of grazing on ANPP accumulated over time in all pastures, including the controls. Canopy Cover of Herbs and Shrubs--. Canopy coverage of the dominant shrub, Artemisia tridentata, was greatest in the control and in the highest density elk treatment (Fig. 7A). As a result, quadratic of effects approached significance (£:1,6 = 3.2, P = .13, Fig. 7A). We failed to observe significant effects of elk population density on canopy coverage of palatable shrubs (Fig. 7B). However, differences among levels were obscured by high levels of variation among blocks. Canopy cover declined with year for both shrub classes (~ < .03) and the magnitude of year effects did not interact with treatment (year x level ~ > .51). Canopy cover of perennial grasses tended to be greatest in the moderately grazed (9, 15 elk/km2) treatments, but this tendency was not significant (£:1,6 = 2.3 ~ = .18, Fig. 7C). Because the canopy coverage of �270 grass was greatest at the moderate grazing levels and the cover of shrubs was least, we found that grass contributed the greatest proportion of the total plant canopy at intermediate levels of elk grazing (£:'1,6 = 4.5, P = 0.07, Fig. 70) • Herbaceous canopy cover declined steeply with year (£:'3,4 = 63, ~ = .0008) as did the total cover of shrubs (L,4 = 9.4, ~ = .03). The magnitude of the effect of treatment on herbaceous cover and shrub cover did not change as the study proceeded (minimum ~ = .25). Forage Nutritional Quality--. We did not detect effects of elk grazing on digestibility of live perennial grass (main effect £:'1,6 = .95, ~ = .51, Fig. 8A). Digestibility of standing dead perennial grass (averaged across years and sample dates) increased in direct relation to elk density (linear effect £:'1,6 = 4.6, ~ .08, Fig. 8B), but these effects were small. We can be 90% confident that, on average, digestibility of dead grass in high density (31 elk/km2) grazed pastures did not exceed control values by more than.2 percentage points. However, effects elk grazing on the digestibility of the total grass biomass (live + dead) were more substantial. Linear effects were significant (£:'1,6 = 7.7, ~ = .03, Fig. 8C) and digestibility values of grass biomass in the 31 elk/km2 treatments were as much as 5.5% points higher than control values. Digestibility of other live herbs also increased linearly with elk grazing level (£:'1,6 = 4.2, ~ = .08, Fig. 80). Year effects on dry matter digestibility were strong for all forage categories (maximum ~ < .001, Fig. 8). However, year x level interactions only approached significance (minimum ~ = .11). The effect of year on forage digestibility can be attributed to annual differences in phenology and in dead/live ratios. Effects of elk grazing on nitrogen content of live perennial grass were not significant (main effect L,6 = .25, ~ = 0.86, Fig. 9A). Nitrogen content of standing dead perennial grass (averaged across years and sample dates) increased in direct proportion to treatment (linear contrast £:'1,6 = 8.8, ~ = .02, Fig. 9B) as did the nitrogen concentration of the total grass biomass (linear effect £:'1,6 = 15.6, ~ = .008, Fig. 9C). We did not detect effects of treatment on nitrogen content of other live herbs (main effect £:. = .89, ~ = .54). Year effects on forage nitrogen content were strong for all forage categories (maximum ~ < .001, Fig. 8) but effects of year and treatment did not interact (minimum ~ .50). = = Total Forage and Nutrient Supply--. Total supply of forage dry matter available to cattle declined in direct proportion to elk population density (linear effect F1,6 = 5.0, ~ = .07, Fig. lOA), as did the supply of digestible energy (linear effect £:1,6 = 4.2, ~ = .09, Fig. lOB) and nitrogen (linear effect £:'1,6 = 4.5, ~ = .08, Fig. 10C). Year effects were significant for all measures of forage supply (maximum P < 0.0003) and the size of the effect of elk grazing on forage supply were consistent across all years (year x level interaction, minimum P > .37). Foraae Quantity/Quality Interactions-The average concentration of digestible energy in the total biomass available to cattle declined 1.7% per 10 kg increase in that biomass (r2 = .59, ~ < .0001, Fig. 11). This decline occurred primarily as a result of changes in the dead to live ratio of the total available biomass, which increased asymptotically as biomass increased (Fig. 12). Thus, when biomass available to cattle was high, a large component of the biomass was contributed by residual standing dead and, consequently, the average nutritional quality of the available biomass was low. Effects on Cattle Foraging Cattle Diet Composition and Quality--. Herbaceous plants dominated cattle diets (Table 2); shrubs uniformly contributed less than 12% of dry matter consumed. Although the contribution of shrubs to cattle diets more than doubled in response to elk grazing (Table 2), these effects were not significant (main effect F48 = 1.1, P = .37). We found no year effect on the �271 botanical composition of cattle diets (minimum R > .12), and the effect of treatment did not depend on year (minimum R > .18). Nitrogen content of cattle diets increased significantly in response to elk grazing in the moderately grazed treatment (£:1,6 = 4.3, P = 0.06, Table 2) and approached significance over all treatments (control vs others £:1,6 = 3.1 R = 0.1). We did not detect effects of elk on the digestibility or fiber constituents of cattle diets (Table 2). Nutritional quality of cattle diets did not change with year (minimum R > 0.48), and we found no interactions between treatment and year (minimum R > 0.33). Daily forage intake bv cattle: Daily dry matter intake of live forage by cattle tended to decline as elk density increased, but this tendency was not significant (level main effect lJ,6 = .62, R = .62, Fig. 13). In contrast, cattle intake of standing dead declined in direct proportion to elk density (linear effect £:1,6 = 5.2, R = .06 Fig. 13A). Consequently, total (live + dead) dry matter intake of cattle also declined in response to increasing elk density (linear contrast £:1,6 = 5.0, R = .07, Fig. 13A). Reductions in dry matter intake rates caused declines in daily intake of energy (linear contrast £:1,6 = 4.7, R = .07, Fig. 13B) and nitrogen (control vs others £:1,6 = 5.4, R = .06, Fig. 13C), despite enhancements in concentrations of energy and nitrogen in forage and in cattle diets that were attributable to treatment (Fig. 8,9, Table 2). Year effects on digestible energy and nitrogen intake of cattle approached significance (energy: lJ,4) 1.8, R = .11, nitrogen: lJ,4 = 3.5, P = 0.13). The magnitude of treatment effects on digestible energy intake changed significantly with year (year x quadratic effect lJ,4 = 5.5, R .07; year x control vs others lJ,4 = 5.7, R = .06; year x control vs 15 lJ,4 = 6.9, R = .05). Utilization of the total forage supply by elk accounted for 18% of the variation in daily digestible energy intake by cattle (Fig. 14). On average, each % point increase in utilization of the herbaceous forage supply by elk was associated with a 2.9±.9 Mjoule reduction in cattle daily energy intake. However, scatter about the regression was large. For example, utilization rates in the range of 10 to 15% were associated with a 50 fold range in digestible energy intake by cattle. Daily intake of forage dry matter and digestible energy by cattle were asymptotically related to the total herbaceous forage supply available to cattle during the spring grazing season (Fig. 15). Increases in total forage supply above about 40 gm/2 exerted negligible effects on cattle intake rates of dry matter or energy (Fig. 15). However, when forage supplies dropped below this level, daily intake rates of dry matter and energy declined steeply as forage supplies declined. Cattle Grazing Intensity: Reductions in cattle dry matter intake were reflected in their utilization of forage. Utilization of standing dead perennial grass (removal/standing crop) by cattle declined in direct proportion to elk density (Fig. 16A, linear contrast £:1,6 = 24.1, R = .003) and, thus, opposed the trends in utilization of standing dead grass by elk (i.e., Fig. 2A). Similarly, utilization of live perennial grass by cattle was virtually a mirror image of utilization by elk (i.e., compare Fig. 14B vs. 2B). As a result of these opposing trends, the proportion of the annual production of perennial grass removed by cattle and elk remained relatively constant across all levels of elk density (Fig. 17) at about 60%. DISCUSSION Effects of Elk on Forage Resources for Cattle A necessary condition for the expression of interspecific, resource competition is a reduction in availability of resources for one species resulting from use of those resources by another species (Tilman 1982). Elk grazing during winter and spring reduced the total supply of forage dry matter, digestible energy, and nitrogen available for consumption by cattle in �272 spring. These reductions occurred as a direct result of removal of live and dead biomass by elk; we found no indirect effects of elk grazing on primary production. The failure of elk grazing to influence primary production can be explained in part by interactions between elk and cattle. Consumption of live perennial grass by cattle declined with increasing elk density. Thus, the increasing intensity of forage utilization by elk grazing was opposed by reductions in utilization by cattle. As a result, total utilization of live tissue by cattle and elk remained relatively constant at slightly less than 60% of the annual production, despite increases in elk population density and despite increased rates of forage use by elk. It follows that we should not expect large changes in net primary production as a result of grazing simply because the overall intensity of grazing was relatively constant (ca 60%) across all treatment levels. The magnitude of the effect of year on net primary production was far greater than the effect of elk population density. A large portion of the year effect was attributable to annual variation in rainfall, but a significant effect of year on primary production remained even after the influence of rainfall was removed. We emphasize that this remaining annual effect was evident in the control as well as the treatment pastures. We suggest these effects can be explained as follows. Because our study area was not grazed by livestock for 5 years before the study began, the introduction of grazing represented a significant perturbation of the system we studied. We suggest that part of the effect of year on ANPP was a response to this perturbation, and that downward trends in ANPP were, in part, a response to the introduction of grazing after a history of rest. Because effects of year (independent of rainfall effects) were evident in the control pastures, it appears that cattle grazing may have contributed to a downward trend in ANPP on this site. Effects on Resource Acquisition A sufficient condition for the expression of interspecific competition for resources is a reduction in acquisition of limiting resources by one species that results from use of those resources by another species (Tilman 1982). Elk used resources that otherwise have been available to cattle, and as a result, the amount of dry matter and energy consumed by cattle declined as elk population density increased. Thus, the mechanism linking elk population density to cattle resource use was a Type II functional response (Holling 1965) of cattle to the total supply of forage available during spring (i.e., Fig. 15). On the surface, it would appear paradoxical that a nonlinear mechanism like a Type II functional response of cattle to forage availability would produce a linear response of cattle forage intake rate to elk population density (i.e., Fig. 13 but also see Hobbs et ale submitted: Fig. 9). However, we emphasize that given only 4 levels in the population density treatment, it is simply infeasible to separate a linear response from more complex, curvelinear ones, particularly at low population densities. Similar responses of large herbivores to changes in forage availability have been observed in other grazing systems (Van Der Kley 1956, Allden 1962, Trudell and White 1981, Allison 1985, Forbes and Hodgson 1985, Short 1985, Birrell 1991). A Type II response of daily intake to total forage biomass occurs for large herbivores because they are able to compensate for the effects of reduced availability on instantaneous intake by increasing daily grazing time (Arnold 1970, Allden and Whittaker 1970, Chacon and Stobbs 1976, Allison 1985, Bunnell and Gillignham 1985, Short 1986, Penning et ale 1991). This compensation allows daily intake to remain relatively constant over a broad range of forage availabilities. However, animals cannot increase grazing time without limit--the need to ruminate and rest constrains maximum grazing time of cattle to no more than about 12 hours per day (Stobbs 1975). When this constraint is reached, daily intake must decline as availability declines because animals can no longer offset reduced instantaneous intake by expanding grazing time. We emphasize that such compensatory mechanism are not without cost. Prolonged grazing elevates energy expenditures (Young and Corbett 1972; Havstad and Ma1echeck 1982), which may reduce animal growth and �273 reproduction by harming energy balance, even if energy intake remains constant. Based on the relationship between forage supply and cattle intake rate, we conclude that under conditions similar to those we studied, a threshold in forage supply in the vicinity of 40 gm/m2 of live and dead herbaceous biomass appears to determine the nature of the interaction between elk and cattle in sagebrush steppe. When available biomass exceeds this threshold, further increases in biomass are not likely to impact cattle energy and dry matter intake, and competition will be relatively weak. When elk numbers and spatial distribution cause forage available to cattle to decline below this threshold, then resource competition between elk and cattle will be strong and will increase in intensity as forage availability declines. Thus, we suggest that under conditions similar to those we studied, elk populations in sagebrush steppe can be managed to reduce competition with cattle by controlling their spatial distribution and population density to assure forage available to cattle during spring exceeds about 40 g/m2 of herbaceous live and dead herbaceous biomass. This result has important implications for rangeland monitoring. Forage utilization is often used to assess impacts of wild herbivore populations on rangelands; excessive utilization is viewed as an indicator of excessive population densities. However, forage utilization by elk was only weakly related to energy intake by cattle and was not related to net primary production. Thus, in systems similar to those we studied, it is unlikely that utilization data will prove useful in predicting when elk grazing will harm the nutritional status of cattle. Such data will not predict impacts of elk on forage production. This is the case because forage utilization rates are only one of the factors that influence variation in forage availability. Other factors, including site characteristics, precipitation, and previous grazing history, contribute to that variation as well. Thus, measurements of the residual forage remaining after grazing, measurements that reflect all of these sources of variation, are likely to prove more useful than measures of utilization alone in evaluating impacts of one herbivore species on another. The Balance Between Competition and Facilitation Much effort has been invested in understanding facilitative relationships among herbivores who share forage resources (Bennett et al. 1970, Bell 1971, McNaughton 1976, Jarman and Sinclair 1979, Sinclair and Norton-Griffiths 1982, Coppock et al. 1983, McNaughton 1984, Gordon 1988, Gordon and Lindsay 1990). There is ample evidence that grazing by one herbivore species can enhance the nutritional quality of forage supplies for other species by maintaining vegetation in an immature, rapidly growing state, and by removing standing dead (for review, see Gordon and Lindsay 1990). In so doing, grazera retard the accumulation of structural tissue in plant communities, tissue that would otherwise dilute nutrients and reduce digestibility of forage plants (McNaughton 1984). However, Hobbs and Swift (1988) urged caution in interpreting facilitative effects. They showed that nutritionally beneficial effects of grazing on forage quality must be weighed against nutritionally detrimental effects of grazing on the amount of forage available. As we demonstrated above (i.e, Fig. 15), sufficient reductions in forage amount can reduce energy intake rate of herbivores even when the nutritional quality of forage is enhanced (i.e., Fig. 8,9). This is apparently what occurs when winter grazing by elk grazing precedes spring grazing by cattle in sagebrush steppe. We showed that grazing by elk could enhance the average digestibility and nitrogen content of forage available to cattle, and could increase the nitrogen content of cattle diets. These enhancements resulted primarily from shifts in the dead to live ratio of the forage available to cattle. We showed further that elk can increase the proportion of the total plant canopy that is contributed by grasses. Presumably, such shifts favor conditions for foraging by grazers like cattle. However, we also found that elk grazing caused substantial reductions in cattle daily intake of dry matter, digestible energy, and nitrogen, reductions that appeared to be caused by reductions in forage biomass. Reductions in �274 intake, in turn, appeared to cause significant declines in growth by cattle during spring (Hobbs et al. submitted: Fig. 10). Thus, although facilitation and competition operated simultaneously in the same grazing system, the effects of competition prevailed. competitive effects of elk on cattle in sagebrush steppe resulted in large part because elk grazing reduced the residual standing crop of dead perennial grass that was carried forward from year to year. Although dead plant tissue can be viewed as diluent of nutrients in the standing crop, it also represents an important buffer against annual variation in forage supply. Standing dead plant biomass mitigates the effect of low rainfall on the availability of total herbaceous forage. This is to say that large residuals of standing dead resulting from high levels of primary production during high rainfall years can buffer the effect of reduced production during low rainfall years. Elk grazing weakened this buffering effect. However, residual standing dead declined precipitously with year in the con~rol as well as in the treatment pastures. We surmise from this result that stocking rates of cattle were probably excessive--that grazing sagebrush grassland ranges to achieve 50% removal of the production of live perennial grass will not provide a forage residual adequate to sustain grazing during low rainfall years. CONCLUSIONS We conclude that interactions between elk and cattle in sagebrush grassland can be viewed as a balance between competitive effects on forage quantity and facilitative effects on forage quality. Excessive population density of elk can tip this balance in the direction of competition, particularly when annual primary production is low and cattle stocking rates are high. Table 1. Data on precipitation for Maybell, Colorado years. Total precipitation averages 27.5 cm. Precipitation during growing season (MayJuly, cm) Year 1 during Total annual precipitation 1 1987 6.9 37.0 1988 4.8 28.0 1989 3.8 13.0 1990 5.7 15.9 Data from Craig Colorado, ca 30 km east. the four study (cm) �Table 2. Effects of elk winter and spring grazing on composition cattle diets during the spring and summer. NItrogen o vs 31 16 16 31 ~ ~ ~ SE1 1: Effcct Cll 1: 70.2 70.4 69.4 0.7 O.BB -2.7 .2.1 0.45 0.06 O.OB -0.25.0.03 0.40 1.47 1.40 Effcct CI -1.3.2.3 -0.31 • -0.01 o vs others 1: Effcct CI 0.72 -1.4. 2.1 0.14 -.23. -0.09 NDr 55.1 52.7 ·54.9 2.3 0.4B -4.1. B.7 0.94 -6.2.6.6 0.66 -4.3.6.B ADF" 24.9 24.7 25.5 1.2 O.lB -9.2.5.4 0.3B -1.7.4.6 0.20 -0.90 .• 0.46 0.28 O.BO -0.67.0.61 0.61 -O.9B.0.57 O.BB -0.71 .0.61 0.67 -14.7.22.9 0.43 0.B1 -17.0.21.9 0.95 -17.4. O.OB -14.2. -0.49 LIgnIn 4.92 4.B1 5.13 Qrass 65.4 56.1 61.3 6.7 0.35 Forbs 29.9 33.4 27.4 7.1 0.72 4.B 11.7 11.B 3.1 0.13 Shrubs I o vs 0 1.32 of Contrast Elk densIty (antmats/km') IVDDM" (' of dry matter) Pooled estimate of standard -23.0.15.9 -15.3. 1.7 0.12 -15.5 • 1.6 ~9.6. 23.0 16.3 error of mean. 90% confldence Interval on the difference ) In vitro dry matter digestibility 2 -9.5 - 2B.1 between the control and treatment means. • Neutral detergent "her S Acid detergent "ber N " V1 �276 FIGURE CAPTIONS Figure 1. Layout of 32 ha pastures used in studies of effects of winter and spring grazing by elk on cattle performance during the spring. Numerals within each pasture show the number of adult elk females stocked each year during December 27 - April 15. The 0 elk pastures served as controls; the pastures with 3 animals achieved a density of 9.3 animals/km2 (5.1 ha/animal unit month); the pastures containing 5 elk achieved a density equivalent to density of 15.4 anima1s/km2 (3.04 ha/animal unit month) and the 10 elk pastures achieved a density of 30.9 animals/km2 (1.5 ha/animal unit month). All pastures were stocked with cattle (7 cow/calf pairs and 1 heifer) during May-June of each study year. See text for explanation of cattle stocking rates. Pastures were constructed of high tensile electric fence. Water was provided near the center hub in each pasture. Figure 2. Rates of forage utilization by elk on sagebrush steppe rangeland during winter and early spring plotted against elk population density. (A) Utilization of standing dead perennial grass (8) Utilization of live perennial grass (C) Utilization of other live herbs (forbs + annual grasses). Open symbols are level means for each year. Solid squares show the 4 year average with 90% confidence intervals (vertical bars) based on variation among blocks (n = 3). Effect confidence intervals (CI) enclose the true difference between level means (control ~ - treatment ~) with 90% confidence. Significant differences between treatment means and the control are indicated by * (~< .10), ** (~ < .05), and *** (~< .01). Figure 3. Spring (ca May 1) standing crops of (A) standing dead perennial grass, (8) live perennial grass, and (C) other live herbs (forbs + annual grasses) on sagebrush steppe rangeland plotted against population density of elk using that rangeland during winter and early spring. Open symbols are level means for each year. Solid squares show the 4 year average with 90% confidence intervals (vertical bars) based on variation among blocks (n = 3). Effect confidence intervals (CI) enclose the true difference between level means (control ~ - treatment ~) with 90% confidence. Significant differences between treatment means and the control are indicated by * (~< .10), ** (~< .05), and *** (~< .01). Figure 4. Aboveground net primary production of (A) perennial grass and (8) other live herbs (forbs + annual grasses) on sagebrush steppe rangeland plotted against population density of elk using that rangeland during winter and early spring. Open symbols are level means for each year. Solid squares show the 4 year average with 90% confidence intervals (vertical bars) based on variation among blocks (n = 3). Effect confidence intervals (CI) enclose the true difference between level means (control ~ - treatment ~) with 90% confidence. Significant differences between treatment means and the control are indicated by * (~ < .10), ** (~< .05), and *** (~< .01). Figure 5. Relationship between aboveground net primary production (ANPP) and rainfall in sagebrush steppe during 4 years. Equation for the best fit line including all treatment levels is y = -22 + a.ax. Figure 6. Relationship between year and the residuals from ANPP vs. rainfall regression. A substantial year effect on ANPP remained after the effects of rainfall were removed. Equation for the best fit line including all treatment levels is y = 2.63 - 3.5x. Figure 7. Canopy coverage of (A) Artemisia tridentata, (8) other shrubs, and (C) perennial grass at the end of the growing season in sagebrush steppe plotted against population density of elk using that rangeland during winter and early spring. The canopy cover of perennial grass as proportion of the total cover was greatest at the intermediate elk densities (D). Open symbols are level means for each year. Solid squares show the 4 year average with 90% confidence intervals (vertical bars) based on variation among blocks (n = 3). �277 Effect confidence intervals (CI) enclose the true difference between level means (control ~ - treatment ~) with 90% confidence. Significant differences between treatment means and the control are indicated by * (~< .10), ** (~ < .05), and *** (~< .01). Figure 8. Average in vitro dry matter digestibility of forages collected during May-July on sagebrush steppe rangeland plotted against population density of elk using that rangeland during winter and early spring. Forage categories include live perennial grass (A), dead perennial grass (B), total perennial grass (C), and other live herbs (D). Open symbols are level means for each year. Solid squares show the 4 year average with 90% confidence intervals (vertical bars) based on variation among blocks (n = 3). Effect confidence intervals (CI) enclose the true difference between level means (control ~ - treatment ~) with 90% confidence. Significant differences between treatment means and the control are indicated by * (~ < .10), ** (~ < .05), and *** (~< .01). Figure 9. Average nitrogen content (g/g dry matter) of forages collected during May-July on sagebrush steppe rangeland plotted against population density of elk using that rangeland during winter and early spring.. Forage categories include live perennial grass (A), dead perennial grass (B), total perennial grass (C), and other live herbs (D). Open symbols are level means for each year. Solid squares show the 4 year average with 90% confidence intervals (vertical bars) based on variation among blocks (n = 3). Effect confidence intervals (CI) enclose the true difference between level means (control ~ - treatment ~) with 90% confidence. Significant differences between treatment means and the control are indicated by * (~ < .10), ** (~< .05), and *** (~ < .01). Figure 10. Amounts of dry matter (A), digestible energy (B), and nitrogen (C) available to cattle during May-July on sagebrush steppe rangeland plotted against population density of elk using that rangeland during winter and early spring. Open symbols are level means for each year. Solid squares show the 4 year average with 90% confidence intervals (vertical bars) based on variation among blocks (n = 3). Effect confidence intervals (CI) enclose the true difference between level means (control ~ - treatment ~) with 90% confidence. Significant differences between treatment means and the control are indicated by * (~ < .10), ** (~< .05), and *** (~< .01). Figure 11. Relationship between average forage digestibility and total herbaceous biomass available to cattle during the spring and early summer in sagebrush steppe. Each data point shows a pasture mean during one year. The solid line is y = 64 - 0.17x. Dashed lines are 90% confidence intervals on the prediction of the average y at a given x. Figure 12. The live/dead ratio in total herbaceous biomass available to cattle increased asymptotically with increasing total herbaceous biomass. Each data point is a value for 1 pasture for 1 year. Figure 13. Rates of intake of dry matter (A), digestible energy (B), and nitrogen (C) of cattle using sagebrush steppe rangeland during the spring grazing season (ca May 10 - June 20) plotted against population density of elk using that rangeland during winter and early spring. Forage categories, include (A) standing dead perennial grass, (B) live perennial grass, and (C) other live herbs. Open symbols are level means for each year. Solid squares show the 4 year average with 90% confidence intervals (vertical bars) based on variation among blocks (n = 3). Effect confidence intervals (CI) enclose the true difference between level means (control ~ - treatment ~) with 90% confidence. Significant differences between treatment means and the control are indicated by * (~< .10), ** (~< .05), and *** (~< .01). Panel A shows effect confidence intervals on total dry matter intake. �278 Figure 14. Relationship between digestible energy intake of cattle and forage utilization rates by elk in sagebrush steppe. Solid line is y = 167 - 291x. Dashed lines are 90% confidence intervals on the mean of y for a x. Figure 15. Functional response of cattle to changes in forage supply during spring and summer in sagebrush grassland. Forage supply is the sum of the live and dead biomass available to cattle during the spring grazing season. Data points are pasture means for cattle daily intake of dry matter (A) and digestible energy (B) paired with pasture means for forage supply for each of the 4 study years. Figure 16. Rates of forage utilization by cattle on sagebrush steppe rangeland during the spring plotted against population density of elk using that rangeland during winter and early spring. Forage categories are standing dead perennial grass (A), live perennial grass (B), and other live herbs (C). Open symbols are level means for each year. Solid squares show the 4 year average with 90% confidence intervals (vertical bars) based on variation among blocks (n = 3). Effect confidence intervals (CI) enclose the true difference between level means (control ~ - treatment ~) with 90% confidence. Significant differences between treatment means and the control are indicated by * (~< .10), ** (~< .05), and *** (~< .01). Figure 17. Total forage utilization by cattle + elk on sagebrush steppe rangeland during the spring plotted against population density of elk using that rangeland during winter and early spring. Forage categories are standing dead perennial grass (A), live perennial grass (B), and other live herbs (C). Open symbols are level means for each year. Solid squares show the 4 year average with 90% confidence intervals (vertical bars) based on variation among blocks (n = 3). Effect confidence intervals (CI) enclose the true difference between level means (control ~ - treatment ~) with 90% confidence. Significant differences between treatment means and the control are indicated by * (~< .10), ** (~< .05), and *** (~ < .01). �279 LITERATURE CITED A.O.A.C. 1980. Official methods of analysis (13th Ed.). Association of Analytical Chemists. Washington, D.C. Allden, W. G. 1962. Rate of herbage intake and grazing time in relation to herbage availability. Proceedings Australian Society for Animal Production. 4:163-166. Allden, W. G. and I. A. McD. Whittaker. 1970. The determinants of herbage intake by grazing sheep: the interrelationship of factors influencing herbage intake and availability. Australian Journal of Agricultural Research (Cambridge) 21:755-766. Allaye-Chan, A. C. 1984. A computer simulation of bluebunch wheatgrass (Agropyron spicatum) growth dynamics and implications to integrated management of livestock and wildlife. M. S. Thesis, The University of British Columbia, 144pp. Vancouver, British Columbia, Canada. Allison, C. D, M. M. Kothmann, and L. R. Rittenhouse. 1983. Forage intake by cattle as affected by grazing pressure. Proceedings of the International Grasslands Congress 14:670-672. Allison, C. D. 1985. Factors affecting forage intake by ruminants:a review. Journal of Range Management 38:305-311. Bell, R. H. V. 1971. A grazing ecosystem in the Serengeti. Scientific American 225:86-93. Bennett, D., F. H. W. Morley, K. W. Clark, and M. L. Dudeinski. 1970. The effect of grazing cattle and sheep together. Australian Journal of Experimental Agriculture and Animal Husbandry 10:694-709. Birrell, H. A. 1991. The effect of stocking rate on the grazing behavior of corriedale sheep. Applied Animal Behavior Science 28:321-331. Blaisdell, J. P., R. B. Murray and E. D. McArthur. 1982. Managing Intermountain Rangelands: Sagebrush-grass Ranges U.S.D.A. Forest Service General Technical Report Int-134. Bransby, D. I. 1989. Compromises in the design and conduct of grazing experiments. Crop Science Society of America, Special Publication 16:53-67. Bunnell, F. L. and M. P. Gillingham. 1985. Foraging behavior: dynamics of dining out. Pages 53-80 in R. J. Hudson and R. G. White eds. Bioenergetics of Wild Herbivores. CRC Press. Boca Raton, Florida. Canfield, R. H. 1941. Application of the line-interception method in sampling range vegetation. Journal of Forestry 39:25-29. Chacon, E., and T. H. Stobbs. 1976. Influence of progressive defoliation of a grass sward on the eating behavior of cattle. Australian Journal of Agricultural Research 27:709-727. Cole, I. W. L., and I. E. Grizzle. 1966. Applications of multivariate analysis of variance to repeated measures experiments. Biometrics 31:22:810-828. Cool, K. L. 1992. Seeking common ground on western rangelands. Rangelands 14:90-92. Connor, E. F. and D. Simberloff. 1986. Competition, scientific method, and null models in ecology. American Scientist 74:155-162. Coppock, D. L., J. E. Ellis, J. K. Detling, and M. I. Dyer. 1983. Plant-herbivore interactions in a North American mixed-grass prairie. II. Responses of bison to modification of vegetation by prairie dogs. Oecologia 56:10-15. Cowlishaw, S. J. 1951. The effects of sampling cages on the yields of herbage. Journal of the British Grassland Society 6:179-184. Coppock, D. L., D. M. Swift, J. E. Ellis, and K. Galvin. 1986. Seasonal patterns of energy allocation to basal metabolism activity and production for livestock in a nomadic pastoral ecosystem. Journal of Agricultural Science (Cambridge) 107:357-365. Daubenmire, R. F. 1940. Exclosure technique in ecology. Ecology 21:514-515. Denslow, J. S. 1985. Disturbance-mediated coexistence of species. Pages 307-323 in S. T. A. Pickett and P. S. White. The Ecology of Natural Disturbances and Patch Dynamics. Academic Press, San Diego California, U.S.A. �280 Diamond, J. R. 1978 Niche shifts and the rediscovery of interspecific competition. American Scientist 66:322-331. Forbes, T. D. A., and J. Hodgson. 1985. Comparative studies on the influence of sward conditions on the ingestive behavior of cows and sheep. Grass and Forage Science 40:69-77. Frank, D. A. and S. A. McNaughton. 1992. The ecology of plants, large mammalian herbivores, and drought in Yellowstone National Park. Ecology 78:2043-2058. Freund, R. J., R. C. Littell, and P. C. Spector. 1986. SAS System for Linear Models. SAS Institute. Cary N. C. 211pp. Gill, J. L., and H. D. Hafs. 1971. Analysis of repeated measurements of animals. Journal of Animal Science 33:331. Gordon, I. J. 1988. Facilitation of red deer grazing by cattle and its impact on red deer performance. Journal of Applied Ecology 25:1-10. Gordon, I. J., and w. K. Lindsay. 1990. Could mammalian herbivores manage their resources? Oikos 59:270-280. Goering, H. K., and P. J. Van Soest. 1970. Forage fiber analyses: apparatus, procedures, and some applications. USDA-ARS Agr. Handbk. No. 379. Grant, P. R. 1972. Interspecific competition among rodents. Annual Review of Ecology and Systematics. 4:79-106. Havstad, K. M. and J. C. Malechek. 1982. Energy expenditure by heifers grazing crested wheatgrass of diminishing availability. Journal of Range Management 35:447-450. Hobbs, N. T. and D. M. Swift. 1988. Grazing in herds: when are nutritional benefits realized? American Naturalist 131:760-764. Hobbs, N. T., D. L. Baker, G. D. Bear, and D. C. Bowden. submitted. Ungulate grazing in sagebrush steppe: effects of resource competition on secondary production. Ecological Applications: under review. Hogan, Z. 1990. Elk and cattle: a conflict in land use? Rangelands 12:268-270. Holechek, J. L., R. D. Pieper, and C. H. Herbel. 1989. Range Management: Principles and Practices. Prentice Hall. Englewood Cliffs, New Jersy. Holling, C. S. 1965. The functional response of predators to prey density and its role in mimicry and population regulation. Memoirs of the Entomological Society of Canada 48:1-46. Huston, M. 1979. A general hypothesis of species diversity. American Naturalist 113:81-101. Keddy, P. A. 1989. Competition. Chapman and Hall London, U. K. Englewood Cliffs, New Jersy, U.S.A. 501 pp. Klingman, D. L., S. R. Miles, and G. o. Mott. 1943. The cage method for determining consumption and yield of pasture herbage. Journal of American Society of Agronomy 35:739-746. Kufeld, R. C. 1972. Foods eaten by the Rocky Mountain elk. Journal of Range Management 26:106-113. Law, R. 1989 Competition. Pages 243-284 in J. M. Cherrett editor. Ecological Concepts: The Contribution of Ecology to an Understanding of the Natural World. Blackwell Scientific, London U. K. Laycock, W. A. 1967. How heavy grazing and protection affect sagebrush-grass ranges. Journal of Range Management 20:206-213. Laycock, W. A. 1970. The effects of spring and fall grazing on sagebrush grass ranges in eastern Idaho. International Grassland Congress Proceedings 11:52-54. Laycock, W. A. 1979. Management of sagebrush. Rangelands 1:207-210. Laycock, W. A., and P. W. Conrad. 1981. Responses of vegetation and cattle to various systems of grazing on seeded and native mountain rangelands in eastern Utah. Journal of Range Management 34:52-58. MacArthur R. H. 1972. Geographical Ecology. Harper and Row, New York, New York, U.S.A. McNaughton, S. J. 1976. Serengeti migratory wildebeest: facilitation of energy flow by grazing. Science 191:92-94. McNaughton, S. J. 1984. Grazing lawns: animals in herds, plant form, and coevolution. American Naturalist 124:863-886. Meijs, J. A. C. 1986. Effects of concentrates on herbage intake and milk production. Grass and Forage Science 41:229-235. �281 Mitchell, R. L., J. B. McLaren, and H. A. Fribourg. 1986. Forage growth, consumption, and performance of steers grazing bermudagrass and fescu mixtures. Agronomy Journal 78:675-680. Mize, C. W. and R. C. Schultz. 1985. Comparing treatment means correctly and appropriately. Canadian Journal of Forest Research 15:1142-1148. Owensby, C. E. 1969. Effect of cages on herbage yield in true prairie vegetation. Journal of Range Management. 22:131-132. Parsons, A. J., B. Collett, and J. Lewis. 1984. Changes in the structure and physiology of a perennial ryegrass sward when released from a continuous stocking management: implications for the use of exclusion cages in continuously stocked swards. Grass and Forage Science. 39:1-9. Pearson, H. A. 1970. Digestibility trials: in vitro techniques. Pages 120-127 in Range and Wildlife Habitat Evaluation. United States Department of Agriculture Forest Service Miscellaneous Publication Number 1147. Penning, P. D., A. J. Parsons, R. J. Orr, and T. T. Treacher. 1991. Intake and behavior responses by sheep to changes in sward characteristics under continuous stocking. Grass and Forage Science 46:15-28. Powell, J., G. Godbolt, and W. G. Hepworth. 1986. Questions about livestock-big game relations. Rangelands 8:281-283. Rittenhouse, L. R. and F. A. Sneva. 1976. Expressing the competitive relationship between Wyoming big sagebrush and crested wheatgrass. Journal of Range Management. 29:326-327. Robertson, J. H. 1947. Responses of range grasses to different intensities of competition with sagebrush. Ecology 28:1-16. Rotenberry, J. T., and J. A. Wiens. 1985. Statistical power analysis and community-wide patterns. American Naturalist 125:164-168. Sauer, R. H. 1978. Effect of removal of standing dead material on growth of Agropyron spicatum. Journal of Range Management. 31:121-122. Sharrows, S. H. and I. Motazedian. 1983. A comparison of three methods for estimating forage disappearance. Journal of Range Management. 36:459-471. Short, J. 1985. The functional response of kangaroos, sheep, and rabbits in an arid grazing system. Journal of Applied Ecology 22:435-447. Short, J. 1986. The effects of pasture availability on food intake, species selection and grazing behavior of kangaroos. Journal of Applied Ecology 23:559-572. Sinclair, A. R. E. 1985. Does intraspecific competition or predation shape the African ungulate community? Journal of Animal Ecology 54:899-918. Sinclair, A. R. E., and M. Norton-Griffiths. 1982. Does competition or facilitation regulate migrant ungulate populations in the Serengeti? A test of hypotheses. Oecologia 53:364-369. Smith, D. R. 1961. Competition between cattle and game on elk winter range. Agricultural Experiment Station, University of Wyoming, Laramie, Wyoming. 16 pp. Sweeney, J. M., and J. R. Sweeney. 1984. Snow depths influencing winter movements of elk. Journal of Mammalogy 65:524-526. Stobbs, T. H. 1975. Factors limiting the nutritive value of grazed tropical pastures for beef and milk production. Tropical Grasslands 4:237-245. Tilley, J. M. A., and R. A. Terry. 1963. A two-stage technique for iq vitro digestion of forage crops. Journal of the British Grassland Society 18:401-411. Tilman, D. 1982. Resource Competition and Community Structure. Princeton University Press, Princeton, New Jersy, U.S.A. Tilman, D. 1987. The importance of the mechanisms of interspecific competition. American Naturalist 129:769-774. Tilman, D. 1989. Discussion: population dynamics and species interactions. Pages 89-100 in J. Roughgarden, R. M. May, and S. A. Levin eds, Perspectives in Ecological Theory. Princeton University Press, Princeton, New Jersy, USA. Toothaker, L. E. 1991. Multiple Comparisons for Researchers. Sage Publications, London. 168 pp. �282 Trudell, J., and R. G. White. 1981. The effect of forage structure and availability on food intake, biting rate, bite size, and daily eating time of reindeer. Journal of Applied Ecology 18:63-82. Underwood, T. 1986. The analysis of competition by field experiments. Pages 240-268 in J. Kikkawa and D. J. Anderson eds. Community Ecology: Pattern and Process. Blackwell, Boston MA. Valentine, J. F. 1990. Grazing Management. Academic Press, San Diego, California, 533 pp. Van Der Kley, F. K. 1955. A simple method for the accurate estimation of daily variations in the quality and quantity of herbage consumed by rotationally grazed cattle and sheep. Netherlands Journal of Agricultural Science 4:197-204 Van Dyne, G. M., N. R. Brockington, Z. Szocs, J. Duek, and C. A. Ribic. 1980. Large herbivore subsystem. Pages 269-537 in A. J. Breymeyer and G. M. Van Dyne, editors. Grasslands Systems Analysis and Man. Cambridge University Press. Cambridge, U.K. Van Soest, P. J., R. H. Wine, and L. A. Moore. 1987. Estimation of the true digestibility of forage by the in vitro digestion of cell walls. Proceedings of the 10th International Grassland Congress 10:438-441. Wagner, F. H. 1978. Livestock grazing and the livestock industry. Pages 121145 in H. P. Brokaw, editors. Wildlife in America: Contributions to an Understanding of American Wildlife and its Conservation. Council on Environmental Quality. United States Government Printing Office Document Number 041-001-00043-2. West, N. E. 1983. Western intermountain sagebrush steppe. Pages 351-374 in N. E. West, editor. Temperate Deserts and Semideserts. Ecosystems of the World Series. 5. Elsevier, New York, New York. Wiens, J. A. 1977. On competition and variable environments. American Scientist 65:590-597. Williams, S. s. 1951. Microenvironment in relation to experimental techniques. Journal of British Grassland Society 6:207-217. Williamson, L. 1992. Identifying the Shinola. Rangelands 14:81-87. Willms, W. D. 1986. Herbage production following litter removal on Alberta native grassland. Journal of Range Management. 39:536-540. Willms, W., A. W. Bailey, and A. Mclean. 1980a. Effects of clipping or burning some morphological characteristics of Agropyron spicatum. Canadian Journal of Botany 58:2309-2312. Willms, W., A. W. Bailey, and A. Mclean. 1980b. Effect of burning or clipping Agropyron spicatum in the autumn on the spring foraging behavior of mule deer and cattle. Journal of Applied Ecology 17:69-84. Willms, W., A. W. Bailey, A. McLean, and C. Kalnin. 1981. Effects of fall clipping or burning on the distribution of chemical constituents in bluebu~ch wheatgrass in spring. Journal of Range Management 34:267-269. Wright, H. A. 1970. Response of big sagebrush and three-tip sagebrush to season of clipping. Journal of Range Management 23:20-22. Young, B. A., and J. L. Corbett. 1972. Maintenance energy requirement of grazing sheep in relation to herbage availability. I. Calorimetric estimates. Australian Journal of Agricultural Research 23:57-76. �~ . C'? o o . (D o o �r :j.,l,oi,_, 284 80 A: DEAD PERENNIAL EFFECT GRASS t 60 o vs o o -32.5 0 -9.4·· -40~6 0 -17.6··· vs 15: o vs 40 8: 31: -65.0 0 -42.0·"" vs OTHERS: 20 CI LINEAR -43.9 0 -25.1""" P - 0.0001 QUADRATIC P - 0.46 0 0 -;;,te _. >al W C!) < a: 0 IL. IL. 31 15 40 EFFECT B: LIVE PERENNIAL :x::: W 9 GRASS o vs 30 o 0 20 0 1 <> 10 0 Z 0 •... 8: vs 15: o vs o vs 31 : OTHERS: CI -12.2 0 3.26 -16.40-0.9-17.6 0 -2.2-14.0 0 -1.4· 1 LINEAR 0 QUADRATIC P - 0.04 P - 0.36 0 0 < N 9 15 31 :::J ~ ::> 40 EFFECT C: OTHER LIVE HERBS CI o vs 8: -9.1 0 -0.4" 30 o vs 15: -9.70 o vs o vs 31: -10.80-0.1. 20 OTHERS: 1.0 -9.1 0 -0.4" 0 10 f 0 0 9 -.Y! 15 ! LINEAR P - 0.13 QUADRATIC P - 0.32 31 ELK DENSITY (animals j km") YEAR 1 {;:. 20 30 4\7 X ALL. ./. �2-85 20 A: DEAD PERENNIAL GRASS EFFECT A A 15 A 5 T 1---+--T~l 0 1 V V 0 9 0 -- '"E 10 0 0 o vs 8: -1.7,5.1 o vs 15: -0.9 , 5.9 o vs o vs 0 10 A CI 2.0,8.8·· 31: 0.4,6.0·· OTHERS: LINEAR P - 0.02 o 1 15 31 QUADRATIC P - 0.92 B: LIVE PERENNIAL GRASS EFFECT CI CI o vs 8: -0.7 , 2.4 o vs 15: -0.9,2.2 'P'" > -e ::::l: a, 0 A 5 a: o o T t---T A T %-:____ V V V Z o vs o vs 0.0 , 3.1 • A T OTHERS: -0.3 , 2.3 , V LINEAR P - 0.12 0 QUADRATIC P - 0.89 Z -c •.... 31: 0 en 0 9 15 31 w o -c a: 0 10 C: OTHER LIVE HERBS EFFECT LL o vs A 5 0 A t-~ o 8: CI -0.5,1.9 o vs 15: -1.2,1.8 o vs 31: -0.3,2.7 o vs OTHERS: -0.9 , 1.9 LINEAR P - 0.18 0 ~ QUADRATIC P - 0.81 0 0 9 15 31 ELK DENSITY (anirnalsjkm") YEAR 1 t:.. 20 30 4<:] X ALL. �286 A: PERENNIAL GRASS 30 EFFECT A -- ••• E CI i.•. 0 o 10 ~ v <> 0 0 a: a.. -.--... 9 ••• .•. OTHERS: -1.4 • 3.2 " LINEAR P - 0.21 T v <> <> -1.9.3.7 -3.2 • 2.4 '------0• v 8: o vs 15: o vs 31: o vs A 20 :z .... o vs A CI -0.6.4.9 QUADRATIC P - 0.47 0 > a: 0 -c 9 31 15 ::E a: a.. .... 30 B: OTHER HERBS W :z A 0 :z ~ 20 A a: o W > 10 0 ~--II--T 1 t ~ ! <> <> m A A 0 CI EFFECT 1 o vs 8: -5.9.4.1 o vs 15: o vs 31: o vs -4.8.3.9 -3.8. 5.0 OTHERS: -3.8 • 3.3 ~ <> <> LINEAR P - 0.72 c( QUADRATIC P - 0.60 0 0 9 31 15 ELK DENSITY (animals j krn") YEAR 1 6 20 30 4\7 X ALL. �(.0 0 . ~ 1\ • 0 1\ • :\. • ~ \. ,.- 0 0 0 V •... e, C\J II 0 I"-- .- .\~.;. \. \. \'"\". \ • It) 0 ( c;UJ / ,. ~ \ '\ 0 C\I '\.' ~ '~ _ .. "\) y\\ ('I) 0 6 ) ddNV '\ 0 ,.... 0 co r- (.0 It) ~ ('I) -E 0 ...J ...J ~ Z ~ Z 0 en US en 0)lY5 I I o ,.... It) 0 O C\I E c E ~ (!J .._ z C/) 3: as a: 0 (!J -as ~ en Z 0 w ~ ...J W F '.'lre.. 287..J s- �288 0 C\I L() ..q- II C\I •... • • ,.... 0 0 0 V CL 0 T"" , •• • 1' I,' I,' •• •• ' f,', /', +i •• •• me;; I I T"" OlD , , " {, a: L5 -'w ~ 0 w z (/) ~ -as ~ E (U C/) ~ (II // T"" >- 0 I // /,.' .. - " ' • 0 C\I E -. _. ~': :. , / I, , ( ,,: I':, l', : J',__'' .. •• 0 I T"" NOISS38838 ddNV NI SlvnOlS38 �-~ 0 25 20 a: W > 0 15 a. >- 10 z < U 5 U A EFFECT V t--+ Y 0 i CI o vs 8: -3.5.6.7 o vs 15: -0.4 , 9.8 o vs o vs 31: -4.6,5.6 - -a:~ 2.0 > 0 1.5 >- 1.0 0 -1.9 ,6.4 OTHERS: a. Z P • 0.85 -c 9 15 0 EFFECT tt- t- 5 Z T __ ._~ ~ b 8 e -c U 1 • o v ~ o W CI > -2.3,1.2 U o vs -2.3,1.3 >n, 15: 31: -0.7,2.9 OTHERS: -1.4,1.5 LINEAR 20 O 10 0 5 u, P • 0.18 -0.15.0.71 o vs 15: -0.38,0.47 31: -0.15 , 0.69 OTHERS: -0.15 , 0.55 LINEAR P • 0.38 QUADRATIC P ·0.87 15 YEAR 1 'V 26 (animals 30 I .I o e v v "'i EFFECT vs 8: -0.10,0.03 o vs 15: -0.12,0.10 o vs o vs o CI o 31: -0.05 , 0.07 OTHERS: -0.08, 0.03 LINEAR P • 0.62 QUADRATIC 0 DENSITY 1 a VI 1 31 ELK 31 P ·0.07 0 '$. 9 I 15 I 0 0 0 1 < o _J -c II- 9 D t- 15 Z P • 0.20 QUADRATIC 25 0 O vs 8: o vs o vs t- U 0 0 ~ V!----~ 31 C ~ >a. vs 8: 0.0 0 > 0 t CI o o vs o vs o ~<> 0.5 U P • 0.13 QUADRATIC 10 o 0 a: a: w EFFECT B U LINEAR 0 -- r W 0 15 2.5 9 15 31 krn") 40 X ALL • I LJ;. f ") ~I" I.D -- �N \.0 65 r -"#. A g ~ 60 ~ m -•.... C/J W o o a: w •.... •.... « a: a: •.... > v •<>.. T 6 6 B: -2.2 , 0.3 o vs 15: -1.B, 0.5 <> V 10 45 60 ~ 55 ~ v EFFECT o , o <> <> Ii<>___io 6 0 6 o vs o 15: vs 31: 50 t 6 -1.5,0.7 o vs 31: o vs -2.0,0.2 EFFECT o r-~! -3.0, -5.5" 60 vs B: CI -4.2 , 0.3 Ovs1S: -1.4,3.1 o -5.1, -0.6 vs 31: o vs OTHERS: -3.2 , 0.5 6 6 vs 15: 31 OTHERS: -1.3, -1.0" 6 o OTHERS: -1.2 , 0.2 <> o vs -0.5,1.7 D 65 -2.2,0.26 OvsB: QUADRATIC P • 0.37 -2.0 , 0.5 vs B: CI LINEAR P ·O.OB 70 CI EFFECT 6 o ~~--------------~-o 9 15 C v T 10 31 v V 0 6 QUADRATIC P ·0.14 v V 6 LINEAR P • 0.91 65 r <> ~-----~,---------~ 6 50 -1.2,1.4 <> <> OTHERS: -1.5 , -0.3" 6 o ~~--~--~------~-o 9 15 B 55 CI o vs o vs 31: o vs 6 55 . o o v r--~i ~ >- EFFECT 0 T >•.... -I 0 o 55 LINEAR P • 0.03 LINEAR P • 0.09 Z 10 QUADRATIC P ·0.79 10 QUADRATIC P • 0.24 o ~~--~--~------~-- o ~~--~--~------~-o 9 15 o 31 ELK YEAR 1 \l DENSITY 26 3 (animals 0 4 <> 9 15 31 I krn") X ALL • II c: ~ to �2.5 r A , A 2.3 ~ .--+--~ g 'i - 2.0 ~ I- 0.1 Z W 0.0 ~~--~--~------~ ---'eft. I- o A v 1 EFFECT o 0 vs 8: o -0.13, Ovs31: 0 1.0 0.11 -0.16,0.08 0.8 vs OTHERS: o .-.__ . . B v -0.16,0.82 vs 15: o 1.2 CI v ~ A o EFFECT CI o vs 8: -0.09 , 0.06 ~ ~ V o vs 15: -0.12, 0.02 A ~ L 0 0 o vs 31: -0.18, -0.03·· o vs e -0.13,0.07 OTHERS: -0.12, 0.0 0.6 1.8 LINEAR QUADRATIC o Z o P • 0.64 9 LINEAR 0.1 P - 0.97 P • 0.03 QUADRATIC P - 0.B2 0.0 15 o 31 9 31 15 o z w c 2.5 EFFECT (!) o a: I- 1.8 v v v e-a-·- Z ~ ~ o v _*a vs 8: o vs o 1.2 15: 0.1 v -0.2, 0.04 vs -0.2, 9 15 2.0 o 1 \l 1 A o vs 31: OTHERS: LINEAR 3 0 9 -0.3,0.1 o vs o QUADRATIC 26 -0.2,0.3 vs 15: 1.5 (animals -0.2 , 0.2 o 0.5 0.0 L-_._ __ ---'....___.._ DENSITY vs 8: ; LINEAR P • 0.008 P • 0.50 CI o A o o ELK ~ ,-~-~! 31 YEAR ? 0.02 0.0 ~~--~--~------~ o EFFECT 2.5 -0.3,-0.1*** OTHERS: 0.5 D -0.2,0.1 Ovs31: ~ 3.0 CI 15 -0.2, 0.2 P • 0.49 QUADRATIC P • 0.44 __'o. 31 I krn") 40 X ALL • ~ ~ i-' T f.. ;l ..- ~ �-_ I-,~\.>.K.IO 292 J 75 '"c; E 50 (/) (/) ct ~ 25 m EFFECT A +--.-k V o l V o l V y ~ CI o vs 8: -5.7,11.3 o -5.2,12.0 vs 15: o vs 31: o vs 1.2,18.3· OTHERS: -1.7,12.3 l v o LINEAR P - 0.08 o QUADRATIC P - 0.75 o o - '" 9 31 15 .E. 5 700 600 o vs 8: -70 , 111 >- 500 o -54,125 Q) ~ EFFECT B CJ Ovs31: a: 400 w z w 300 ~ m 200 vs 15: ~ -25, 123 OTHERS: o <> LINEAR P - 0.09 QUADRATIC P - 0.75 0 o 5 9 15 1.0 - ... -- '"E 31 EFFECT C CI Ovs8: -0.09,0.13 o vs -0.07,0.13 0.8 Cl 15: 0.6 o vs 31: o vs o 0.4 OTHERS: z 0.2 w CJ 0.86,183· o vs i= 100 (/) z CI a: •.... o o 0.004, 0.23· -0.03,0.15 v o e LINEAR P - 0.08 QUADRATIC P - 0.75 0.0 o 9 31 15 ELK DENSITY (animals/km2) YEAR 1 ~ 20 30 4\7 X ALL. �I"- 0 • • / / • / LO (0 / •• • I I / I I • • • I I / / / , • •• • • • • •• 0 (0 I • I I I I I I I I I • LO LO I I I • I I I I / I / I / ,• • • • ,_ V . § 0) LO II •... e, C\I (%) A.l.11181J.S3810 ~3llVII\J A~O 38~3J\V / / • / / / 0 LO 0 CO I"- 0 0 (0 LO 0 C\I - 0) .€ 0 L5 0 + W > -:J « en en ~ 0 0 v m en :J 0 l- 0 ~ _J I W m c: ~ 0 W ('I) 0 C\I 0 T"" 0 r .j 'J.-rr"_ 293 I' �294 o eo. o • o •••••. o • • • • • • • • • • • • •• • •• • • • • • • • • C\I • (Y) ci • -.::t ci • LO ci • • • <0 ci 00000 ci (6/6) OI.L't11:J 3A1l1aV'3a • • ••• • o o. o • • ••• • • • • o T"': o 0 eo ('II - 0 ::J w en ffi 0 ~ en en -c -:J > + w 0 Lli 0 - C)) 0 ~ r- <0 0 0 LO -.::t 0 0 (Y) m ~ a: w ~ ~ _j 0 I C\I 0 T"'" o �t- .lh_~ J 295 ~ til "C - 25 Cl .:.:. 20 W ~ e:( 15 c:: 10 ~ ~ 5 W e:( TOTAL -0- LIVE - ?---??~~ . 0 0 DEAD ~T i_ >c:: 0 - . . 9 15 o vs 15: -0.6,6.0 o vs 15: -1.8,4.8 o vs o vs 31: 1.0 , 7.S" 0.8 ,5.5" OTHERS: -..__.---_. .•.. CI EFFECT A T I- z -- LINEAR P - 0.06 QUADRATIC P - 0.87 31 >til- - "C 225 s 0 CI EFFECT 8 t:. w o vs 8: 4.40 - 71.S* o vs 1S: -17.0, SO.O o vs o vs 31: 11.9 , 79.4* Q) ::J 150 0 ~~ <p ~ W ~ W •,b. <> <> 75 I- Z <> LINEAR >- " z c:: w -- QUADRATIC 0 9 15 300 >til Cl t:. 200 EFFECT 0 ~ e:( l- Z Z w ~ 100 o T • ! r e CI o vs 15: o vs 15: -14.0, 120.S o vs o vs 31: 11.5,120.5* -0.3,108.7 8.6, 97.4* OTHERS: ~ <> LINEAR 0 P - 0.21 <> c:: QUADRATIC I- Z P - 0.74 31 C "C ~ P - 0.07 0 w W S.S , SO.7* OTHERS: e:( P - 0.25 0 0 9 31 15 ELK DENSITY (antmalsjkrn") YEAR 1 £:. 20 30 4\7 X ALL. I~ �296 0 0 eo • • • • I I • • I • I I I I • I • I • I I I • 0 0 C\I I •• I I I •• I I I I I I I I • •• I I I I • I I I I I • • I • • • • • "• I ,, ,, I , •., I • I I 0 0 T"'" I I I (Aep/Moo/alnofll\l) 3>iV.lN1 30 I I • I • I ,_ Q) II I I I • • ,_ § V •... e, C\I I I I • • 0 10 · (t) 0 0 co · 0 -. '-*' « en en 10 0 W 0 :J en ffi ~ · C\I 0 · 0 C\I 0 ·W ~ _J :J J= ~ 0 Z 0 ~ LL ~ _J J: W a: m 10 ~ T"': 0 0 0 T"': 10 · 0 0 · 0 0 0 10 0 I 0 0 T"': I 0 �;'::;297 30 A 'S: CIJ I 20 * + * ~ * * *.•. ~ +**:tt' ,....* .•.... "!Ir * a: * * * + 10 ~ >a: .•.+ + + **,. .•.* 0 0 0 10 20 30 40 50 60 TOTAL HERBACEOUS BIOMASS (UVE + DEAD, ~ ELK DENSITY *** 0 *** 9 70 80 2) .•..•..•. 15 +++ 31 300 ... B ... 'S: CIJ 1 200 + 0 * ~ w 100 l ** +++ * * *"..•.•. •••• + ~ ... + .•. + + III '3 ... ** * * * .;, ....... w 0 *A .•.* 0 0 10 20 30 40 50 TOTAL HERBACEOUS BIOMASS (UVE ELK DENSITY *** 0 *** + 60 70 DEAD, ~ 2) 9 .•..•..•. 15 +++31 80 �298 A: DEAD PERENNIAL GRASS 80 EFFECT o vs 60 o T ~ 0 ~ 5.4 , 21.9** 11.6,28.2** OTHERS: 6.3, 19.8*** 'il T LINEAR P - 0.003 J. o ~ o -- -2.7,13.9 o vs If-.--..~T & ~ 20 vs 15: Ovs31: ~ 40 8: CI o 9 15 QUADRATIC P - 0.53 31 :::fie a W -l B: LIVE PERENNIAL GRASS 80 ee o > m 60 w (!) -c 40 o o o t-+--.. _ ~ o v '1,-----, LL o Z o 'il o a: o LL EFFECT o tt- o o 20 CI o vs 8: o vs 15: 5.2 , 13.7**· o vs 31: 2.4 , 10.8** o -3.1, 5.3 vs OTHERS: 2.3, 9.2** LINEAR P - 0.01 QUADRATIC P - 0.07 o o t- ee 9 15 31 N ::i t- :::> 80 C: OTHER LIVE HERBS EFFECT 60 1 40 J 20 CI Ovs8: -15.0,28.3 o -28.9,14.4 vs 15: o vs 31: -21.1, 22.2 o vs OTHERS: -17.7,17.6 LINEAR P - 0.86 o o QUADRATIC P - 0.81 o o 9 15 31 ELK DENSITY (animals/km2) YEAR 1 b. 20 3<> 4\7 X ALL. �!=-; ~ 1Are.. 17. ...J 299 A: DEAD PERENNIAL GRASS 90 , v 'V V 60 ~-~ 30 ~ vs 8: -27.4 , -3.4·· o vs 15: -27.4 , -3.4*· 31: -45.0 , -20.1··· OTHERS: -31.3, -11.7··· ~ LINEAR P - 0.002 ~ -- CI o o vs o vs 0 ~ EFFECT QUADRATIC P - 0.76 0 0 0 9 31 15 ~ ...J W c < Z 80 B: LIVE PERENNIAL GRASS 0 W 0 ...J tt- 60 > m 40 ~ < 0 _" i--+- 'f__ EFFECT 0 ¢ V ~ 0 t- 20 o vs 15: -6.8 , 8.3 o vs o vs 31: -10.9 , 4.3 -10.9,4.2 QUADRATIC P - 0.78 :;j t- 8: LINEAR P - 0.57 < N :::l o vs OTHERS: -8.1 ,4.2 Z 0 CI O 0 w 9 31 15 C!l < a:: 0 LL. 80 C: OTHER LIVE HERBS 60 V EFFECT o vs 40 20 0 8: CI -20.6,24.8 v o 1 o vs 31: o vs t OTHERS: -23.0,13.9 0 LINEAR P - 0.55 vs 15: -34.4 , 11.1 -27.0, 17.9 0 QUADRATIC P - 0.57 0 0 9 15 31 ELK DENSITY (animalsjkm") YEAR 1 c: 20 30 4\7 X ALL. ��301 UNGULATE EFFECTS N. Thompson OF RESOURCE Hobbsl, IMammals Research GRAZING IN SAGEBRUSH COMPETITION Dan L. Bakerl, II: ON SECONDARY George D. Bearl PRODUCTION and David Section, Colorado Division of Wildlife, Fort Collins, Colorado, 80526 USA 2Statistics Department, Colorado Fort Collins Colorado, Corresponding STEPPE C. Bowderr' 317 W. Prospect, State University, 80523 USA Author: N. T. Hobbs Mammals Research Section Colorado Division of Wildlife 317 W. Prospect Fort Collins Colorado, 80526 USA Telephone: 303-484-2836 ext. 360 Fax: 303-490-2621 Internet: nthobbs@ariel.dowfc.colostate.edu Abstract. The potential for resource competition between wild and domestic ungulates on arid rangelands in Western North America has emerged as an important issue influencing land management and policy. In particular, populations of elk are believed to harm production of cattle by competing with them for limited supplies of native forage. We examined effects of variation in the population density of elk (Cervus elaphus canadensis) during winter on growth and reproduction of cattle during spring in a randomized complete block experiment conducted in sagebrush steppe during 4 years. We manipulated elk numbers to achieve 4 levels of population density (0, 9, 15, and 31 elk/km2), replicated each level 3 times, and observed responses of cattle to these manipulations. Average birth dates of calves born to cows in the intermediate (9, 15 elk/km2) treatment levels were delayed by 5 days relative to birth dates of calves born to cows in controls, but these trends were not statistically signif icant (£:1,6 = 3.2, F. = .12). Calf body mass at birth was not significantly influenced by treatment, but calf body mass at the end of spring declined linearly (£:1,6 = 6.0, F. = • OS) with increasing elk population density from a mean of 80.2 kg in the control to 73.0 kg in the 31 elk/km2 treatment. Calf body mass at weaning was weakly depressed by treatment, with the largest treatment effects occurring at the 9 elk/km2 level (FI,6 = 8.8, F. =.02). Body mass of cows at the end of the spring grazing season tended to decline linearly with treatment, but these tendencies were not significant (linear contrast £:1,6 = 2.3, F. = .18). We did not find significant effects of treatment on cow body mass at weaning (control vs. others contrast £:1,6 = 3.3, F. = .12) or on natality rates (control vs other contrast £:1,6 = 2.8, F. = .14), although in both cases, values for the control tended to exceed the treatments. Body mass of cows (~I = -3.9, F. = .0003) and calves (~I = -3.9, F. < .0001) at the end of the spring grazing season were quadratically related to the biomass of available herbaceous forage during spring. We observed a threshold in the effects of forage supply on cattle production at about 40 g/m2. Cattle production declined with declining forage biomass only when forage supply fell below this threshold. Total cattle production, which incorporated all of the above effects on growth and reproduction, was quadratically related to elk popula~ion density (£:1,6 = 5.8, F. = .05). Average cattle production in the control (x = 248 kg �302 = cow" year:' exceeded the mean of the other treatment levels (i 224 kg cow" year-), .[),6 = 6.7, g = .04). Quadratic responses in cattle performance were apparently caused by compensatory growth after the spring grazing season. Growth rates of cows during spring were inversely related to their subsequent growth rates during summer and fall (.[),46 = 33.5, g < .0001, r2 = .50). We conclude that the elk grazing exerts a relatively strong effect on cattle production. This strong effect resulted from a composite of influences that were individually difficult to demonstrate statistically. Cattle production declined as a result of competition with elk, but these declines were not proportionate to elk population density. Under rangeland conditions and stocking rates similar to those we studied, we expect that elk grazing will not harm cattle ~rowth when live and dead herbaceous forage available to cattle exceeds 40 g/mINTRODUCTION Shared use of resources by vertebrates has often been interpreted as defacto evidence of competition (e.g., Grant 1972, Hudson 1977, Dunbar 1978, Jarman and Sinclair 1979, Singer 1979, Seegmiller and Ohmart 1981, Hanley and Hanley 1982, Belovsky 1984, Sinclair 1985, Spowart and Hobbs 1985, Coppock et ale 1986, Ambrose and Deniro 1986, Llewellyn and Jenkins 1987, Jenkins and Wright 1988, Bodmer 1991). However, many uncertainties result from inferring competition from patterns in resource use (Colwell and Futuyma 1971, Colwell and Fuentes 1975, May 1975, Connell 1975, Wiens 1977, Thompson 1980, Connor and Simberloff 1986). Overcoming these ambiguities requires manipulative experiments that examine the direct effects of one species on the growth and reproduction of another when resources are limiting (Grant 1972, Connell 1975, Pianka 1976, Hastings 1987, Glasser and Price 1988, Tilman 1989). Such experiments are rare for vertebrates (Sinclair and Norton-Griffiths 1982, Brown and Batzli 1985, Brown et ale 1986, Dickman 1986). This is particularly true for large herbivores because shared use of resources by these animals may actually facilitate, rather than harm their acquisition of resources (Bennett et ale 1970, Bell 1971, McNaughton 1976, Coppock et ale 1983, McNaughton 1984, Danell and Huss-Danell 1985, Gordon 1988). In many cases, such facilitative relationships occur because defoliation of plants can enhance their nutritional quality (reviewed by Gordon and Lindsay 1990). Consequently, it is particularly uncertain whether shared use of forage resources is harmful or beneficial to herbivore populations that share them. Elk populations in western North America are believed to compete with cattle for limited supplies of forage on sagebrush steppe rangelands, and these competitive relationships are believed to harm cattle production (Smith 1961, Powell et ale 1986, Hogan 1990). We have shown that increasing densities of elk populations using sagebrush steppe during winter and spring caused reduced daily intake of dry matter, energy, and nitrogen by cattle during the spring and summer (Hobbs et ale submitted). These reductions in energy and nitrogen intake occurred despite enhancing effects of elk on the nutritional quality of forage available to cattle, enhancements that facilitated selection of high nitrogen diets by cattle (Hobbs et ale submitted). However, it remains uncertain whether changes in resource use by cattle cause changes in cattle growth or reproduction. Here, we report a manipulative experiment designed to examine the effects of increasing population density of elk on production of cattle in sagebrush steppe. METHODS Study Area and Experimental Design Details of our study area and experimental design have been presented elsewhere (Hobbs et ale submitted). In brief, we examined response of vegetation and cattle to increasing population densities of elk (0, 9, 15, 31 animals/km2) in a randomized complete block experiment conducted during 4 consecutive years at the Colorado Division of Wildlife Little Snake Wildlife �303 Management Area, near Maybell, Colorado. Three replications of each treatment level were established annually by stocking 12 fenced pastures (each 32 ha) with different numbers of elk (Hobbs et al. submitted: Fig. 1). Treatments were randomly assigned within low, medium, and high biomass blocks. Elk were held in pastures during December through April and were replaced with cattle in early May. Management of Cattle During early May we stocked pastures with cattle. Cattle stocking rates were chosen to achieve about 50% utilization of the above ground net primary production of perennial grasses in the control (0 elk/km2) pastures (see Hobbs et al. submitted, for details). The same number of cattle (7 cow-calf pairs and 1 heifer) were stocked in all pastures, including the control, during all four study years. Thus, cattle densities were held constant across pastures, while elk densities varied. This allowed performance of cattle in the 0 elk/km2 pastures to control for performance of cattle who shared rangeland with elk. Cattle remained in pastures during May - June, an interval we will refer to as the "spring grazing season." Cattle were obtained via contract with a local rancher, Mr. Bruce Seely of Craig, Colorado. We randomly allocated 84 cows and calves and 12 heifers among the replicated treatments. Randomization was accomplished by first stratifying the cow herd by age, and then randomly choosing animals to accomplish similar age distributions among pastures. Cows were herefords. Calves were predominantly here fords and hereford x angus crosses. To the extent possible, we maintained the same adults in each pasture from year to year and followed their growth and reproductive output throughout the study. This allowed effects of competition with elk to carry-over from year to year. On average, cattle were kept in the study for 2.7 years each; 34 of the original 96 cows remained in the study for all 4 years. Twelve older animals were replaced each year to accommodate addition of one heifer in each pasture. Sixteen additional animals were replaced because of failure to breed, death losses, and other miscellaneous sources of attrition. Rates of replacement were similar across all treatments. At the end of the spring grazing season, cattle were returned to Mr. Seely's ranch where they were maintained in the same group with the same management regime for the remainder of the year. They were maintained on native rangeland in excellent condition during the summer and fall, and were fed grass hay throughout the winter. All cows were exposed to fertile bulls immediately after their removal from pastures. Each year's calf crop was sold each fall, after weaning, except for some females that were retained as replacements •. Any cows that failed to breed were sold after weaning, usually in mid-November. After cattle were removed from pastures, pastures remained ungrazed by cattle or elk until they were restocked with elk during the following December. Measurements of Cattle Production Cattle growth and reproductive performance was measured as follows. During each study year, we weighed calves born to cows in the experiment within 2 days of their birth. Calves and cows were also weighed when they were introduced to pastures (ca May 15), when they were removed (ca July 1) and at weaning (ca Nov. 10). Data on weaning weights of calves are limited to 3 study years (1987, 1989, 1990) because calves were sold prematurely during 1988. To examine reproductive performance, we recorded birth dates of all calves and calculated natality rates for each pasture each year as the number of cows producing calves divided by the number of total number of cows in the pasture (n = 8). We estimated total annual production by cattle in each pasture as the product of the average natality rate times the average body mass of calves at weaning plus the average annual gain in mass of adults. �304 Statistical Analysis We analyzed effects of treatment on cattle body mass, birth dates and natality rates using a split plot, factorial analysis of variance. Factors were elk density, block, and year, all of which were assumed to be fixed. Pastures formed the whole plot and repeated measures (i.e., years) formed the split plot. We used responses of individual animals (sampling units) as observations but we calculated all F ratios using the mean square error of the treatment x block interaction in denominators. This allowed us to preserve individual responses in the analysis (which was particularly important when covariates were used), but simultaneously provided an estimate of experimental error appropriate to a randomized complete block design where pastures, rather than individuals, were the popper experimental units. We considered natality rate to be a pasture measurement where the popper experimental error is based on the block x treatment interaction. Thus, independence of cows within pastures is not assumed in our analysis. We used covariates to reduce experimental error. Pre-treatment body mass of cows and calves (measured when they were first introduced to pastures) were used as concomitant observations in analysis of covariance of body mass at the beginning of spring, at the end of spring, and at weaning. Thus, we emphasize that concomitant observations were taken during an animal's initial study year, before animals had any opportunity to be influenced by treatment. When covariates were used, we report least-squared means adjusted to the average value of the pretreatment covariate. We did not use covariates to adjust birth dates or body mass of calves at birth because they did not increase precision in the analysis. Average rates of growth were calculated as the difference between least squared estimates of cow and calf body masses at the beginning and end of a time interval divided by the duration of the interval. We tested a priori hypotheses using planned, orthogonal contrasts (Mize and Schultz 1985, Toothaker 1991). contrasts included a comparison of each treatment level versus the control, the control versus the average of all treatment levels, as well as tests for linear and quadratic effects (see Hobbs et al submitted for details). To achieve a reasonable compromise between the probability of a Type I error and the power of our tests (Bransby 1989), we chose critical values of F at a = .10 for all contrasts. We also report 90% confidence intervals on the size of treatment effects. The effect size is the value of the contrast of interest, that is, the true treatment mean subtracted from the true control mean. Thus, positive effect sizes indicate that the control mean exceeded the treatment; negative effect sizes indicate that the treatment exceeded the control. The estimated effect size is the value of the contrast when the estimated means are used to replace the true means. Effect confidence intervals place bounds on the estimate of the effect size. See Hobbs et al. (submitted) for a detailed discussion of effect confidence intervals. We performed all statistical analyses using the SAS System for General Linear Models (Freund et al. 1986) and the SAS Interactive Matrix Language. RESULTS Effects of Elk on Cattle Growth Calf bodv mass.--Effects of elk grazing on growth of calves varied with season of the year. We did not observe significant effects of elk population density on calf body mass at birth (Fig. 1A). However, calves born to cows who were in control pastures (during the previous year) tended to be slightly heavier (on average, about 1 kg) than calves born to cows who were in the elkgrazed pastures. These trends were also seen at the beginning of the spring grazing season (approx. May 10) when calves whose mothers were in the controls during the previous year were 1-9 kg (90% CI) heavier than calves whose mothers were in the elk-grazed pastures (£:1,6 = 6.2, ~ = .05, Fig. lS). Quadratic effects on calf body mass at the beginning of the spring grazing �305 season were significant ('£:1,6 = 5.2, R = .06); linear effects were not ('£:1,6 = 1.2, R = .31). At the end of the spring grazing season (approx. July 1), calf body mass declined in an approximately linear fashion with increasing elk density (Fig. 1C, linear contrast '£:1,6 = 6.0, R = .05). On average, calves in the control were 2.9-11.5 kg (90% CI) heavier than calves in the 31 elk/km2 treatment. However, at weaning (approx. Nov. 10), we did not detect linear effects of elk density on calf growth (linear contrast F1,6 = .0, R = .95, Fig. 1D) while quadratic effects approached significance (F1,6 = 3.6, R = .11). The largest estimated effect of elk grazing on fall body mass of calves occurred at the 9 elk/km2 level where the 90% confidence interval on the effect size showed mean body masses of calves were 4.9 to 23.3 kg lighter than those in the control (Fig. 1D). 80dy mass at weaning tended to be higher in the control pastures than in the treatments (Fig. 1D) and the difference in fall body mass of calves in the control relative to the average of the 3 elk-grazed treatments was marginally significant (control vs. others contrast, '£:1,6 = 3.9, R = .097). We can be 90% confident that, averaged across treatment levels, elk grazing reduced mean calf body mass in the fall by .1 - 15 kg. This is about 0 - 8% of the growth they would be expected to attain in the absence of competition with elk. Year effects were significant for all calf responses (maximum year effect R < .08). Calf body mass tended to be the lowest during year 3 when spring precipitation (Table 1), production of perennial grass, and energy intake by cows were also at a low point (Hobbs et al. submitted). However, although year effects were significant for all growth responses, the magnitude of the effect of elk density did not depend on year for any response (minimum year x treatment interaction R > .45). Calf Growth Rates.--Calves showed positive rates of growth throughout the spring and summer (Fig. 2). Growth rates of calves during the spring grazing season decreased in direct proportion to elk density (linear effect contrast '£:1,6 = 8.0, R = .03, Fig. 2A) and the control differed significantly from the average of other levels ('£:1,6 = 4.4, R = .08). However, these results should be interpreted with caution. When the general linear model for calf body mass at the end of spring was analyzed using body mass at the beginning of spring as a covariate, we found that early spring body mass accounted for 55% of the variance in mass at the end of spring, while treatment accounted for only 4% of that variance. Thus, although the level effect on growth rates was linearly related to treatment, much of the effect of treatment on growth rate was attributable to its effect on body mass at the beginning of spring. We did not detect statistically significant effects of treatment on calf rates of growth after calves were removed from experimental pastures (i.e. during July-November), but rates of growth of animals in the moderate and high density treatments were slightly higher than those in the control (Fig. 28). Year effects were significant for all measures of calf growth rate (maximum year effect R < .10), but the magnitude of the effect of elk density did not depend on year for any growth rate response (minimum year x treatment interaction R > .37). Cow body mass.--At the beginning of the spring grazing season, cows who were in control pastures during the previous spring weighed more than cows who were in the elk-grazed pastures (control vs. others contrast '£:1,6 = 5.9, R = .05, Fig. 3A). However, we can be 90% confident that the average effect of elk was at most 21 kg (or about 5% of the body mass of animals who did not compete with elk), and could be as small as 2.3 kg. We did not observe significant effects of elk grazing on cow body mass at the end of the spring grazing season, despite linear trends in the data (Fig. 38). It would appear logical to compare average body mass at the beginning of spring (Fig. 3A) with average mass at the end of spring (Fig. 38) to draw conclusions about growth. However we caution that such comparisons are misleading because the early spring average (Fig. 3A) is based on 3 years of data, while the late spring average (Fig. 38) is based on 4 years of data. This was the case because cows during the early spring of the first study year had never been exposed to treatment and, hence, were used as covariates rather �306 than as responses in the analysis. In contrast, cows at the end of spring had been exposed to the elk density treatment during all 4 study years and, hence, could be used as responses to treatment. We did not detect effects of elk grazing on cow body mass in the fall (Fig. 3C). Year effects were significant for cow body mass for all sample dates (minimum ~ ~ .002) but the effect of treatment did not depend on year (maximum ~ > .61). Cow rates of growth.--We did not observe statistically significant effects of the elk grazing on growth rates of cows during the spring grazing season (Fig. 4A). However, there was a negative linear trend in cow rates of gain relative to elk density during the spring grazing season (Fig. 4A) and the magnitude of the linear effect depended on year (year x linear effect L,4 = 7.7, ~ = .03). During all years, we also observed a positive trend in cow growth rate with respect to elk density during summer and early fall (Fig. 4B). Consequently, there was a highly significant negative relationship between rate of gain during the spring grazing season, and rate of gain thereafter (,[1,46 = 46.0, ~ < .0001, r2 = 0.50, Fig. 5). Effects on Reproduction Natalitv Rates.--We measured natality rates of cattle as the number of calves born per 100 cows. Natality rates ranged from a high of 96% in the controls to a low of 85% in the low elk density (9/km2) treatment (Fig. 6A). We did not detect repeatable effects of elk grazing on cattle natality rates, although effects approached significance for the 9 elk/km2 level (,[1,6 = 2.8, ~ = .14, Fig. 6A). However, we were also unable to rule out treatment effects that could be quite large (about 20% points, Fig. 6A). Year effects were not significant (year effect ,[45 = .08, ~ = .95) and the effect of elk density on natality rate did not depend on year (year x treatment interaction '[1,6 = 1.08; ~ = .42). Timing of Reproduction.--Birth dates of calves born to cows in the moderate elk density treatments (9 and 15 elk/km2) averaged about 5 days later than birth dates of calves whose mothers were in the control or the high elk density treatments (Fig. 6B). Quadratic effects on birth date approached significance (,[1,6 = 3.2 , ~ = .12, Fig. 6B). The 90% confidence interval places the true effect of moderate (15 animals/km2) elk grazing as a delayed mean birth date of 0-12 days (Fig. 6B). We emphasize that these analyses include only calves whose mothers were in the experiment during the previous year, and thus had an opportunity to be influenced by treatment. Effects on Secondary Production Total production by cattle was reduced as a result of competition with elk, but the effects of competition on cattle production were not linearly related to elk density (linear contrast '[1,6 = .02, ~ = .55, Fig. 7). We observed the greatest effects of competition at the 9 elk/km2 level, and the smallest effects at the 31 elk/km2 level (quadratic contrast '[1,6 = 5.8, ~ = .05, Fig. 7). Cattle production in the control differed from the average of the other 3 levels (control vs others contrast '[1,6 = 6.8, ~ = .04) and control values exceeded values for all treatments. We can be 90% certain that, on average, elk grazing reduced cattle production by 6-40 kg/cow/year. However, we caution that these effects depended in large part on the large response of secondary production to the 9 elk/km2 level. Cattle production at the moderate and heavy elk grazing levels were not significantly different from the control (control vs 15, 31 F1,6 2.96, ~ = .14). Year effects on total cattle production approached significance (year effect '[1,6 = 3.4, ~ = .13), but the magnitude of treatment effects did not depend on year (year x treatment interaction '[1,6 = .60, ~ = .77). �307 Mechanisms Causing Changes in Cattle Growth and Reproduction Body mass of calves and cows at the end of the spring grazing season were curvelinearly related to the biomass of total herbaceous forage available to cattle during the spring grazing season (Fig. 8). Quadratic effects were highly significant for both calves (~I = -4.7, ~ < .0001, r2 = .52) and cows (~I = -3.9, ~ = .0003, r2 = .64). When herbaceous biomass exceeded 40-50 g/m2, further increases in biomass were associated with reductions in cattle body mass at the end of spring. The relationship between cattle body mass at the end of spring and forage biomass during spring appeared to be mediated by the effects of energy intake on growth rate. Rates of gain of cows during the spring grazing season were directly related to their daily intake of digestible energy (EI,46 = 22.4, P < .0001, r2 = .33, Fig. 9). These affects, in turn, appeared to influence the performance of calves--calf rates of gain were linearly related to the rates of gain of their mothers (EI,46 = 33.5, .f < .0001, r2 = .42, Fig. 10). DISCUSSION Ecological Separation in Time? The use of resources during different seasons has been proposed as an important mechanism allowing species of ungulates to avoid the deleterious effects of interspecific competition (Lamprey 1963, Leuthold 1978, Dinerstein 1979). Although use of resources by elk and cattle were completely distinct in time, such temporal separation failed to prevent reductions in resource acquisition by cattle in response to elk grazing (Hobbs et ale submitted), and did not prevent reductions in cattle production. The extent of shared used of habitat by elk and cattle was brief--on average, cattle grazed for only 5 weeks a year in pastures that had been used previously by elk. Nonetheless, this shared use was sufficient to cause reductions in annual production by cattle averaging 10% of control values. Thus, we found that use of forage resources by elk during winter and early spring caused measurable effects on cattle production, even when cattle used those resources only briefly during the late spring and early summer. Competition as a Composite of Forces Effects of elk grazing on calf birth date, on cow conception rate, and on end of spring and fall body of cows were not statistically significant. However, our failure to detect statistically significant effects for these variables should not be interpreted as an absence of biologically meaningful effects on cattle production. This is the case because secondary production is a relatively high order response composed of several lower order ones, and even though many of those low-order responses were not statistically detectable, their collective effects were relatively large and were quite significant. This suggests that competition can act as a composite of forces, forces that may be difficult to detect when observed one at a time. Although competition with elk reduced production by cattle, we suggest that it is unlikely that the presence of elk on sagebrush grassland caused a net decline in secondary production in the system we studied. To make this point, we estimated reasonable values for production by elk in a manner analogous to our calculations of cattle production. We first assumed that the proportion of elk cows producing calves that survive to weaning declines linearly with increasing elk density. Based on data in Houston 1982 (Table 4.3, Fig. 5.2) we estimated this proportion as .68 + .02 * elk density (in animals/km2). Presuming growth of elk cows of 30 kg/cow during summer and fall and estimating calf body mass in fall as 125 kg, (Flook 1982), we calculated total production per elk cow as (.68 + .02(elk/km2)(125 kg»+ 30 kg, and estimated total production per km2 as the sum of per capita elk and cattle production multiplied by the stocking density (Fig 11). We freely admit these calculations provide crude approximations, but they nonetheless suggest that production by elk populations could more'than compensate for �308 competitive effects on cattle production. However, such compensation can be realized on land that is privately owned only if the economic harm caused by elk grazing can be offset by economic benefits accruing from elk to landowners. Population Density and Resource Competition Resource competition can be dissected into two sets of effects. The first effect is the direct influence of changing population density of one species on the current supply and rate of renewal of resources. A secondary effect is the influence of changes in resources on the performance of another, competing species. The expression of this second, higher order effect clearly depends on the operation of the first. Virtually all measurements of the amount of forage available to cattle declined approximately linearly with increasing elk population density. These effects on forage quantity appeared to cause reduced rates of intake of forage dry matter and energy by cattle as a result of a typical Type II response of cattle to changes in forage biomass (Hobbs et ale submitted). Reductions in forage intake, in turn, appeared to retard cattle growth (i.e., Fig. lOA). However, although effects of elk on forage supplies were consistently linear, the effects of elk on the growth and reproductive performance of cattle were not. Although, we found that competition with elk caused linear reductions in cattle body mass at the end of the spring grazing season, many responses measured after cows were released from experimental pastures (i.e., cow rates of gain, cow and calf fall body mass, conception rates, calf birth dates and early spring body mass of cows and calves the subsequent year) were not linearly related to elk population density. We suggest that the absence of linear trends in these responses resulted, at least in part, from the ability of cattle to compensate for effects of competition when it is subsequently relaxed, that is, following an "ecological bottleneck" in resource supply. We observed an inverse relationship between rates of gain of cows during the spring grazing season and their subsequent rates of gain during summer and early fall (Fig. 5). We interpret this relationship as an illustration of compensatory growth (Allden 1970, Butler-Hogg and Tulloh 1982, Suttie et ale 1983, Wright et ale 1986). Reduction in energy intake by vertebrates can cause increased rates of growth when energy intake is increased following periods of energy depravation (reviewed by Wilson and Osbourn 1959). Such increased growth allows animals to compensate for effects of surprisingly long periods of reduced energy intake. Treatment-induced increases in rates of gain during summer could explain the observed reductions in the effect of treatment that occurred after the spring grazing season. Compensatory growth is usually associated with increased rates of food intake (e.g, Wright et ale 1989). Thus, the mechanisms of compensatory growth may depend on relatively high levels of energy availability following energy shortage (Milne et ale 1987, Wright et ale 1989). It follows that the effects we observed may have depended on the relatively abundant forage conditions that were available to cattle after they were removed from spring pastures. In the absence of such conditions, we might fail to observe compensatory effects. Implications for Population Management The absence of a linear relationship between elk population density and production by cattle (Fig. 7) makes it difficult to predict the effects of reducing elk populations on cattle production. This difficulty is compounded by questions of scale. Population objectives for elk populations in the Western United States are usually set at relatively large spatial scales, often at 103 km2• In contrast, conflicts between livestock and cattle are frequently more localized, occurring at scales closer to 10's of km2• It is possible that large scale population reductions will have no impact on local conflicts. For example, our data (Fig. 7) suggest that harvest regimes causing local densities of elk to decline by as much as 3 fold (i.e., from 31 to 9 elk km2) would fail to produce a measurable change in the effects of elk �309 grazing on cattle production. This implies that management of the spatial distribution of elk populations to move animals away from areas of conflict may be more effective than reducing large scale population densities in ameliorating competitive effects of elk on livestock. Our results do not directly address the question of impact of elk grazing at densities < 9 animals/km2, but we can nonetheless infer that some densities of elk are not likely to harm cattle production. We observed a quadratic relationship between cattle body mass at the end of the spring and total forage supply available to cattle during the spring (Fig. 8). When total herbaceous biomass available to cattle (residual dead + live produced) was below about 40 g/m2, cattle body mass at the end of spring declined as forage biomass declined, but when forage biomass available to cattle exceed 40 g/m2, cattle body mass was largely insensitive to changes in biomass, and may have increased as biomass declined (Fig 8). These results are consistent with our findings (Hobbs et al. submitted: Fig. 1S) on the relationship between cattle energy intake and forage supply. Thus, we surmise that when cattle stocking rates and rangeland conditions resemble those we studied, rates of elk grazing that allow for about 40 g/m2 of herbaceous biomass for cattle are not likely to cause measurable competitive effects. The quadratic relationship between cattle body mass at the end of spring and total biomass available can be explained by relationships between forage quantity and quality. High levels of biomass were negatively correlated with the digestible energy concentration of that biomass as a result of the diluting influence of standing dead (Hobbs et al. submitted: Fig. 11). This occurred because when herbaceous biomass was large, the proportion of standing dead in that biomass was disproportionately great (Hobbs et al. submitted: Fig 12). However, this effect may depend on the fact that our study area was not grazed for several years before the beginning of our experiment. CONCLUSIONS We conclude that elk can compete with cattle for forage resources in sagebrush steppe communities, despite temporal separation in their use of those communities. This competitive relationship causes measurable reductions in total cattle production. However, we caution that the magnitude of the competitive effects of elk on cattle production are not proportionate to elk population density. FIGURE CAPTIONS Figure 1. Body mass of calves using sagebrush steppe rangeland during the spring plotted against population density of elk using that rangeland during the winter and early spring. Calves were weighed at birth (A), during early spring (B), at the end of spring (C), and in the fall (D). Open symbols are level means for each year. Means for all responses except body mass at birth adjusted by covariance using the early spring body mass of calves, pretreatment, as a covariate. Solid squares show the 4 year, least squares mean with 90% confidence intervals (vertical bars) based on variation among blocks (n = 3). Effect confidence intervals (CI) enclose the true difference between level means (control ~ - treatment ~) with 90% confidence. Significant differences between treatment means and the control are indicated by * (~ < .10), ** (~< .OS), and *** (~ < .01). Figure 2. Growth rates of calves using sagebrush steppe rangeland during the spring (A during the spring grazing season and B during the summer and early fall) plotted against population density of elk using that rangeland during the winter and early spring. Open symbols are least squared, level means for each year. Solid squares show the 4 year average with 90% confidence intervals (vertical bars) based on variation among blocks (n = 3). Effect confidence intervals (CI) enclose the true difference between level means (control ~ - treatment ~) with 90% confidence. Significant differences between treatment means and the control are indicated by * (~< .10), ** (~ < .OS), and *** (~< .01). �310 Figure 3. Body mass of cows using sagebrush steppe rangeland during the spring plotted against population density of elk using that rangeland during the winter and early spring. Cows were weighed during early spring (A), at the end of spring (B), and in the fall (C). Open symbols are least squares, level means for each year. Solid squares show the 4 year average with 90% confidence intervals (vertical bars) based on variation among blocks (n = 3). Effect confidence intervals (CI) enclose the true difference between level means (control ~ - treatment ~) with 90% confidence. Significant differences between treatment means and the control are indicated by * (~< .10), ** (~ < .05), and *** (~ < .01). Figure 4. Growth rates of cows (A) during the spring grazing season and (B) during the summer and early fall plotted against population density of elk sharing rangeland with cows. Open symbols are least squared, level means for each year. Solid squares show the 4 year average with 90% confidence intervals (vertical bars) based on variation among blocks (n = 3). Effect confidence intervals (CI) enclose the true difference between level means (control ~ - treatment ~) with 90% confidence. Significant differences between treatment means and the control are indicated by * (~ < .10), ** (~< .05), and *** (~< .01). Figure 5. Cow rate of gain during the summer and fall regressed on rate of gain during the spring and early summer. Data points are annual pasture means. Dashed lines show a 95% confidence interval on the prediction of the rate of gain during summer and fall for a given rate of gain during spring. Figure 6. Natality rates of cows (A) and birth dates of calves (B) plotted against population density of elk sharing rangeland with cows. Open symbols are level means for each year. Solid squares show the 4 year average with 90% confidence intervals (vertical bars) based on variation among blocks (n = 3). Effect confidence intervals (CI) enclose the true difference between level means (control ~ - treatment ~) with 90% confidence. Significant differences between treatment means and the control are indicated by * (~< .10), ** (~< .05), and *** (~< .01). Figure 7. Annual production of cattle using sagebrush during spring plotted against population density of elk using that rangeland during winter and early spring. Open symbols are level means for each year. solid squares show the 4 year average with 90% confidence intervals (vertical bars) based on variation among blocks (n = 3). Effect confidence intervals (CI) enclose the true difference between level means (control ~ - treatment ~) with 90% confidence. Significant differences between treatment means and the control are indicated by * (~ < .10), ** (~< .05), and *** (~ < .01). Figure spring during 8. Relationship between body mass of cows and calves at the end of the grazing season and total supply of live and dead herbaceous biomass the spring grazing season. Figure 9. Growth rates of cows during the spring grazing season declined in direct proportion to reductions in their rates of forage intake. Data points are annual pasture means. Dashed lines show a 95% confidence interval on the prediction of the mean rate of gain. Figure 10. Rates of gain of calves during the spring grazing season declined in direct proportion to rates of gain of their mothers. Data points are annual pasture means. Dashed lines show a 95% confidence interval on the prediction of the mean rate of gain. Figure 11. Estimated production by elk and measured production by cattle relation to elk population density on sagebrush steppe rangelands. in �311 LITERATURE CITED Allden, W. G. 1970. The effects of nutritional deprivation on the subsequent productivity of sheep and cattle. Nutrition Abstracts and Reviews 40: 1167-1184. Ambrose, S.H. and M. J. Deniro. 1986. The isotopic ecology of East African mammals. Oecologia 69:395-406. Bell, R. H. V. 1971. A grazing ecosystem in the Serengeti. Scientific American 225:86-93. Belovsky, G. E. 1984. Moose and snowshoe hare competition and a mechanistic explanation from foraging theory. Oecologia 61:150-159. Bennett, D., F. H. W. Morley, K. W. Clark, and M. L. Dudeinski. 1970. The effect of grazing cattle and sheep together. Australian Journal of Experimental Agriculture and Animal Husbandry 10:694-709. Bodmer, R. E. 1991. Influence of digestive morphology on resource partitioning in Amazonian ungulates. Oecologia 85:361-365. Bransby, D. I. 1989. compromises in the design and conduct of grazing experiments. Crop Science Society of America, Special Publication 16:53-67 Brown B. W. and G. o. Batzli. 1985. Field manipulation of fox and gray squirrel populations: how important is interspecific competition? Canadian Journal of Zoology 63:2134-2140. Brown, J. H., D. W. Davidson, J. C. Munger, and R. S. Inouye. 1986. Experimental community ecology: the desert granivore system. Pages 4461 J. Diamond and T. J. Case editors. Community Ecology, Harper and Row, New York. Butler-Hogg, B. W., and N. M. Tulloh. 1982. Growth patterns in sheep: the effects of weight losses on compensatory growth and feed intake in Corriedale sheep. Journal of Agricultural Science of Cambridge 99:641-649. Colwell, R. K. and E. R. Fuentes. 1975. Experimental studies of the niche. Annual Review of Ecology and Systematics 6:281-310. Colwell, R. K. and D. J. Futuyma. 1971 On the measurement of niche breadth and overlap. Ecology 52:567-576. Connell, J. H. 1975. Some mechanisms producing structure in natural communities: a model and evidence from field experiments. Pages 460-490 in M. L. Cody and J. M. Diamond editors. Ecology and the Evolution of Communities. Harvard University Press, Cambridge, Mass. Connor, E. F. and D. Simberloff. 1986. Competition, scientific method, and null models in ecology. American Scientist 74:155-162. Coppock, D. L., J., E. Ellis, J. K. Detling, and M. 1. Dyer. 1983. Plant-herbivore interactions in a North American mixed- grass prairie. II. Responses of bison to modification of vegetation by prairie dogs. Oecologia 56:10-15. Coppock, D. L., D. M. Swift, J. E. Ellis, and K. Galvin. 1986. Seasonal patterns of energy allocation to basal metabolism activity and production for livestock in a nomadic pastoral ecosystem. Journal of Agricultural Science (Cambridge) 107:357-365. Danell, K. and K. Huss-Danell. 1985. Feeding by insects and hares on birches earlier affected by moose browsing. Oikos 44:75-81 Dickman, C. R. 1986. An experimental manipulation of the intensity of interspecific competition - Effects on a small marsupial. Oecologia 70:536-543. Dinerstein, E. 1979. An ecological survey of the royal Karnali-Bardia wildlife reserve, Nepal.Part II: Habitat/animal interactions. Biological Conservation 25:265-300. Dunbar, R. I. M. 1978. Competition and niche separation in a high altitude herbivore community in Ethiopia. East African. Wildlife Journal 16:183-199. Flook, D. R. 1982. A study of sex differential in the survival of wapitii. Canadian Wildlife Service Report, Series No. 11, 71pp. Glasser, J. W. and H. J. Price. 1988. Evaluating expectations deduced from explicit hypotheses about mechanisms of competition. Oikos 57:57-70. �312 Gordon, I. J. 1988. Facilitation of red deer grazing by cattle and its impact on red deer performance. Journal of Applied Ecology 25:1-10. Gordon, I. J., and W. K. Lindsay. 1990. Could mammalian herbivores manage their resources? Oikos 59:270-280. Grant, P. R. 1972. Interspecific competition among rodents. Annual Review of Ecology and Systematics. 4:79-106. Hanley, T. A., and K. A. Hanley. 1982. Food resource partitioning by sympatric ungulates on Great Basin rangeland. Journal of Range Management. 35:152-158. Hastings, A. 1987. Can competition be detected using species co-occurrence data? Ecology 68:117-123. Hobbs, N. T., D. L. Baker, G. D. Bear, and D. C. Bowden. submitted. Ungulate grazing in sagebrush steppe I: mechanisms of resource competition. Ecological Applications: under review. Hogan, Z. 1990. Elk and cattle: a conflict in land use? Rangelands 12:268-270. Houston, D. B. 1982. The Northern Yellowstone Elk: Ecology and Management. Macmillian Publishing, New York. 474 pp. Hudson, R. J. 1977. Habitat utilization and resource partitioning by wild ruminants: multi-variate analysis of nominally scaled attribute data. Northwest Science 51:101-110. Jarman, P. J., and A. R. E. Sinclair. 1979. Feeding strategy and the pattern of resource partitioning in ungulates. Pages 130-163 in A. R. E. Sinclair and M. Norton-Griffiths, editors. Seregeti: Dynamics of an Ecosystem. University of Chicago Press, Chicago, Illinois, U.S.A. Jenkins, K. J. and R. G. Wright. 1988. Resource partitioning and competition among cervids in the northern Rocky Mountains. Journal of Applied Ecology 25:11-24. Lamprey, H. F. 1963. Ecological separation of the large mammal species in the Tarangire Game Reserve, Tanganyika. East Africa Wildlife Journal 1:63-92 Leuthold, W. 1978. Ecological separation among browsing ungulates in Tsavo East National Park, Kenya. Oecologia 35:241-252. Llewellyn, J. B. and S. H. Jenkins. 1987. Patterns of niche shift in mice: seasonal changes in microhabitat breadth and overlap. American Naturalist 129:364-381. McNaughton, S. J. 1976. Serengeti migratory wildebeest: facilitation of energy flow by grazing. Science 191:92-94. McNaughton, S. J. 1983. Serengeti grassland ecology: The role of composite environmental factors and contingency in community organization. Ecological Monographs 53:291-320. McNaughton, S. J. 1984. Grazing lawns: animals in herds, plant form, and coevolution. American Naturalist 124:863-886. Milne, J. A., A. M. Sibbald, H. A. McCormack, and A. S. I. Loudon. 1987. The influences of nutrition and management on the growth of red deer calves from weaning to 16 months of age. Animal Production 45:511-522. Mize, C. W. and R. C. Schultz. 1985. Comparing treatment means correctly and appropriately. Canadian Journal of Forest Research 15:1142-1148. Pianka, E. R. 1976. Competition and niche theory. Pages 114-143 in R. M. May ed. Theoretical Ecology: Principles and Application. Saunders, Philadelphia, Pennsylvania, U.S.A. Powell, J., G. Godbolt, and Wm. G. Hepworth. 1986. Questions about livestock-big game relations. Rangelands 8:281-283. Seegmiller, R. F., and R. D. Ohmart. 1981. Ecological relationships of feral burros and desert bighorn sheep. Wildlife Monograph 78. Sinclair, A. R. E. 1985. Does intraspecific competition or predation shape the African ungulate community? Journal of Animal Ecology 54:899-918. Sinclair, A. R. E., and M. Norton-Griffiths. 1982. Does competition or facilitation regulate migrant ungulate populations in the Serengeti? A test of hypotheses. oecologia 53:364-369. Singer, F. J. 1979. Habitat partitioning and wildfire relationships of cervids in Glacier National Park, Montana. Journal of Wildlife Management 43:437-444. �313 Smith, D. R. 1961. competition between cattle and game on elk winter range. Agricultural Experiment Station, University of Wyoming, Laramie, Wyoming. 16 pp. Spowart, R. A. and N. T. Hobbs. 1985. Effects of fire on diet overlap between mule deer and mountain sheep. J. Wildlife Management 49:942-946. Suttie, M., E. D. Goodall, K. Pennie and R. N. B. Kay. 1983. Winter food restriction and summer compensation in red deer stags. British Journal of Nutrition 50:737-747. Thomson, J. D. 1980. Implications of different sorts of evidence for competition. American Naturalist 116:719-726. Tilman, D. 1989. Discussion: population dynamics and species interactions. Pages 89-100 in J. Roughgarden, R. M. May, and S. A. Levin eds, Perspectives in Ecological Theory. Princeton Univ. Press, Princeton, N. J., U. S. A. Toothaker, L. E. 1991. Multiple comparisons for Researchers. Sage Publications, London. 168 pp. Wiens, J. A. 1977. On competition and variable environments. American Scientist 65:590-597. Wilson, P. N. and D. F. Osbourn. 1959. Compensatory growth after undernutrition in mammals and birds. Biological Review 35:324-363. Wright, I. A., A. J. F. Russel, and E. A. Hunter. 1986. The effects of winter food level on compensatory growth of weaned, suckled calves grazed at two sward heights. Animal Production. 43:211-224. �W t-' 40 35 A: BIRTH f--,-t i o vs 8: -1.2 , 3.8 o vs 15: -1.4,3.5 Ovs31: o 30 C) ~ .......• en en « ::E > C 65 r B: BEGINNING SPRING EFFECT CI o vs ~t-t--f 55 -1.6,3.5 vs OTHERS: -0.9, 3.2 <> 45 LINEAR 10 QUADRATIC o m 15 0 !'-~---t <> vs 8: o vs 15: vs 15: 1.4, o vs 31: -1.4,8.6 11.0·· o vs OTHERS: 0.9, 8.7· P - 0.29 QUADRATIC 10 0 D: FALL 200 -2.9,5.6 .I. A A T 2.9,11.5** 175 ~ 1.3 , 8.2** OTHERS: P - 0.06 I 10 r 0 31 , A <> <> P - 0.97 o vs o 9: vs 15: Ovs31: o - 4.8 , 23.2** -4.8,14.2 -4.9,13.7 vs LINEAR -0.08, 9 15 15.4" P - 0.83 QUADRATIC . o 31 EFFECT CI OTHERS: 150 P • 0.03 QUADRATIC 15 15 9 1.5,9.8** o vs 31: o vs LINEAR 9 225 <> <> o 10 o o T 60 ~ .42 EFFECT CI 0 0 80 ~ P - 0.0 , 9.5* o LINEAR 31 C: END SPRING 0 60 8: o ~~----~~~----~ o ~~----~--~----~ o 9 100 r p •. ~ EFFECT CI P - 0.11 31 ELK DENSITY (anlmats rkrn") YEAR 1 f::::. 20 3<> 4\7 X ALL. ::-n L.G ~ ;t �1.2 EFFECT CI A: SPRING 1.0 0 0.8 o vs 8: 0.01 ,0.16" o vs o vs o vs 15: -0.07,0.10 31 : 0.1 , 0.23"" OTHERS: - 0.6 LINEAR P - 0.03 >- -lU 'C 0 ~ 0.01 ,0.15" 0.1 0.0 QUADRATIC 0 Z 9 P - 0.51 31 15 -c C!) u, 0 1.2 EFFECT CI 8: SUMMER •... W -c a:: o vs , 1.0 0.8 r-~ T o vs 15: o vs 31: o vs , V V T -0.08,0.04 -0.08,0.04 OTHERS: 0.6 -0.05 , 0.05 LINEAR P - 0.30 0.1 0.0 QUADRATIC 0 9 1 b. 20 P - 0.73 31 15 ELK DENSITY YEAR -0.03,0.09 8: (animalsjkm") 30 4\7 X ALL. �f=.:q ..•.re. 3, 316 -.J 450 A: BEGINNING SPRING EFFECTCI +--_Li 400 ~ • ~ 1 0 o vs 8: 0.0 - 23.1" o vs 15: 0.5 ,23.6" o vs o vs 31: -0.4 , 21.5 OTHERS: LINEAR P - 0.16 350 10 0 QUADRATIC P - 0.14 0 500 -CI ~ 0 CD 31 EFFECTCI t:. t 425 T t:. t--I-._ + ! v <C 0 15 9 B: END SPRING C/J C/J ::: > C 2.3,20.4" v t:. -. f ! 0 o vs 8: -14.5 - 39.3 o vs 15: -19.3 - 34.7 o vs o vs 31: -6.3 - 46.2 OTHERS: -8.3 - 35.0 LINEAR P - 0.22 350 10 0 QUADRATIC P - 0.91 9 0 15 31 500 EFFECTCI ~ ~! 450 0 o vs 8: r o vs 15: -4.7 , 24.8 o vs o vs 31: -5.1 - 23.6 • ! -0.6 , 29.1 OTHERS: -12.93, 10.59 C: FALL LINEAR P - 0.63 400 10 0 QUADRATIC P - 0.20 0 9 15 31 ELK DENSITY (anlmalsjkm-) YEAR 1 6- 20 30 4\7 X ALL. �2.0 A: SPRING EFFECT CI o vs 8: 1.5 o vs 1.0 -0.34 , 0.85 15: -0.45,0.78 Ovs31: -0.15,1.08 o vs 0.5 >. as 0.0 -- -0.5 - 1.0 "'C o 9 a: 0.6 o o o _.,. .---. ~ o 1 ~ P - 0.96 EFFECT CI TtY 0.3 P - 0.22 QUADRATIC B: SUMMER o -0.20 , 0.80 31 15 G) as LINEAR " " o Cl oX - OTHERS: ! 1 ~ o -0.1 o vs 8: -0.19,0.13 Ovs15: -0.17,0.15 o vs 31: o vs -0.08, 0.25 OTHERS: -0.18,0.09 LINEAR P - 0.35 QUADRATIC P - 0.77 -0.5 o 9 ELK DENSITY YEAR 1 b. 20 31 15 (animalsjkm-) 3<> 4\7 X ALL. �318 q T"'" / / / / / • / / . CO 0 / / • / / / • / / • I I • / I I / I I I I / I • I / • / • • • • • I / I / • I / I . <0 0 I. • I •• I ~ 0 • I • I I • I / I / ./ I / I / / / / / / / ..,. . • •• • • • ,..; V § , C to II •... e, C'II C'! 0 (Aep/6>t i\ON - A.nr) NIVE>.::103.lV1::l MOO / . 0 0 0 q C\I as 5: 32. 0) s: 0 U. (!) « Z -s ~ I J 0 >: 0 :.J · ::J T"'" · 0 ~ ~ 0 W 0 0 · 0 I T"'" �:0::R '-'" EFFECT CI w 100 « CC 85 ~---,--i I- >- I_J -c I-c Z 70 A 10 o vs 8: -2.1,17.3 o vs 15: -1.6 , 21.6 o vs o vs 31: OTHERS: 0 LINEAR 0 31 15 9 -4.7 , 20.0 -2.1 , 17.3 P = 0.75 QUADRATIC ELK DENSITY YEAR l::. 0 1 2 (animals 03 '\]4 P - 0.17 I km2) ···X c: ~ 120 o ~ Q) as f/J EFFECT CI B 110 ~ '-'" 100 w ~ 10 :r: o lCC CO o vs 15: o -10.7, 1.7 -11.5, 0.6 31: -8.6, 3.4 vs OTHERS: -9.1, 1.0 LINEAR P - 0.62 o 0 31 9 15 ELK DENSITY YEAR vs 8: o vs "'C c o 2 0 (animals 3 0 I km2) 4 ••• 'x QUADRATIC P - 0.13 �••......... w T"" I N 0 'm ~ 300 I EFFECT CI T"" I ~ 0 0 I::. 250 0) 6 ~ 0 200 Io :::> 0 0 a: a.. 10 1 T/! •1 .••......•. Z • ~ t 1 vs 8: 16.0 , 58.0** o vs 15: -3.7 , 39.2 o vs o vs <> <> 0 o 31 : -6.3 , 36.6 OTHERS: I ! ! 0 9 15 I 6.0 , 40.0** LINEAR P = 0.55 31 QUADRATIC P = 0.05 ELK DENSITY (animals I krn") YEAR I::. 1 0 3 <> 4 -.- X -r=7 c;. ~ i\ --4 �500 A ~ en ~ • •• ++ 450 + • * +.to * *+ * 400 t:i0 * m 350 * *. • • ** (2 -.64 P < .0001 3OO~----~----r---~~---'----~----~-----r----~ o ro ~ ~ ~ ~ ~ TOTAL HERBACEOUS BIOMASS (LIVE 8...K DENSITY *** 0 *** 9 + ro 80 DEAD, g/m :.I) '&'&'&15 +++ 31 100 B 90 ~ en en ~ 8 •* *+ *+ ••• •* 80 * • 70 m * ** 60 * ** (2 - .62 P < .0001 50 0 10 ao so ~ so TOTAL HERBACEOUS BIOMASS (LIVE 8...K DENSITY *** 0 *** eo + 70 OEAC, g/m :.I) 9 .&.&.& 15 +++ 31 80 �322 \ \ \ ,, , , , , , , \ \ \ \ \ •• • • •• \ \ \ \ \ \ • •• \ \ \ .. \ • \ ~ \ \ • \ \ \ \ \ \ \ \ \ \ \ • \ 'e \ \ • . c e, V . 8 CW) CW) II C\I ~ • • , , , ', , ,• , , , , ., , •• • • \. \ , \ \ a o • • AVII\J)NIV8 :10 3.lV1:l MOO \ ,, .',, a. 0r- • .\ , , , • , , • ••• • , '. ,, a. C\I (Aepj6:>t 'Alnr- a. I or- a a (W) - 0 ~ ~ :J a ); a :.J C\I I J ~ -s W ~ ~ 0 0 $ W Z a ~ a a: W or- a �q T"" \ \ \ \ . •• • • • (j) \ \ \ C\I , C\1 II ~ . 0 \ \ \ \ \ • \ • • • ,.... § V Q.. \ \ \ \ • \ \ \ \. • . \ \. \ .\ •••• \ \. \ . \ \ • • \ \ \ \ •• \ •• \ , \ \ \ \ .\. . \ \ \ \ \ \. \ \ \ \ \ \ \ • \ \ <q \ 0 \ CO 0 I': · 0 \ \ \ • . \ \ \ \ \ \ \ • \ \ \ • • \ \ 10. 0 • \ (Aep/f»t 'l\ON - Ainr) NIVD .::1031.\fCJ .::IiVO \ • \ 0 ~ ); ~ C) ~ 5; as r'j\)..~JLJI 0 \1. C\I · 0 0 0 0 ~ ~ W 0 LL G « Z - ~ ~ I J 0 :.J · ::> T"" 0 q 0 · 0 I T"" �324 75 50 z o - I- 25 o ............. ... . . . . . . . . . . . .. .. . ........ ......... ... .... . . . . . . . . . . .... ... .. ............. ........... ::::> o o (C o ........... .......................... o o, 9 31 15 ELK DENSITY (animals I krn") Cattle Elk ·. ·. · (measured) (estimated) �325 Colorado Division Wildlife Research July 1993 of Wildlife Report JOB state of Project PROGRESS REPORT Colorado No. W-lS3-R-4 Work Plan No. 9A 2 Job No. Mammals Elk Investigations Spatial Wildlife Period Author: Covered: July N. T. Hobbs, Research Analysis of Conflicts Between and Livestock 1, 1992 - June 3D, 1993 D. Schrupp Abstract A spatial database was created in the GRASS geographic information system analysis of elk livestock conflicts in North Park Colorado. The database contained map layers for roads and streams, major vegetation types, distribution of elk populations, and areas of conflict between elk and livestock. for ��327 SPATIAL ANALYSIS OF CONFLICTS BETWEEN WILDLIFE AND LIVESTOCK P. H. OBJECTIVES 1. Prepare a study plan for the spatial analysis of elk-livestock conflicts that will identify spatial features were conflicts will occur and will support decisions on allocating habitat treatments to resolve conflict. RESULTS A spatial database was created in the GRASS geographic information system for analysis of elk livestock conflicts in North Park Colorado. The database contained map layers for roads and streams, major vegetation types, distribution of elk populations, and areas of conflict between elk and livestock. This database will be used in the Owl Ridge Ecosystem Management Project. All future efforts in spatial analysis will be conducted as a part of the Owl Ridge Project. This project ��329 Colorado Division of Wildlife Wildlife Research Report July 1993 JOB PROGRESS REPORT State of __ C::.::o~l~ora~do:.:....._ __ Project No. W-152-R Work Plan No. Job No. 9A 3 Mammals Research : Elk Investigations Spatial Analysis of Elk Survival Period Covered: July 1, 1992 - June 30, 1993 Author: K. R. Wilson and N. T. Hobbs ABSTRACf Procedures for analyzing the survival of elk were outlined as a function of the animals spatial history and movement through time. The intent of the approach is to detect differences in population processes as a function of factors such as vegetation, elevation, aspect, age, etc. A proportional hazard rate model and a logistic regression model were examined as potential methods in detecting how the variance in spatial history might influence survival. Preliminary Monte Carlo simulation results indicate that a proportional hazard rate model may be useful for detecting differences in survival based upon age, only. ' ��331 SPATIAL ANALYSIS OF ELK SURVIVAL K. R. Wilson and N. T. Hobbs P.N.OBJECTIVES List in contract 1. Develop procedures for simulating landscape heterogeneity. Sources of heterogeneity will include slope, aspect, and vegetation cover types. Use procedures to create simulated maps containing different levels of heterogeneity. 2. Simulate movement of elk over the maps and their survival over time. 3. Simulate movement of elk over the maps and their survival over time. 4. Using 1-3 above, develop the "best" possible sampling approach for determining animal locations. 5. Using 1-3 above, develop a statistical approach for evaluating the effect of landscape heterogeneity on elk survival. The statistical approach will be evaluated by computing bias, precision, power, and goodness of fit. RESULTS In wildlife studies, new approaches to spatial analysis can offer powerful new insight into relationships between animals and the habitats that support them. Traditionally, systems for habitat evaluation have relied on relationships between patterns of habitat use and availability to make inferences on habitat quality. However, it has become clear that such inferences may be misleading (Van Home 1983, Hobbs and Hanley 1991). The fundamental problem with use/availability data is that patterns of habitat use alone cannot reveal the role of spatial variation in processes that regulate the abundance of animal populations. Our focus is on analyzing how variation in the use of space by animals contributes to variation in population processes such as survival rates. We are basing our approach on the perfection of a system that allows for almost perfect locations of animals (within several meters). Currently, this is not feasible for radio-telemetry, but the possibility exists that in the near future global positioning systems (GPS) can be used to acquire very accurate locations. Using GPS, very accurate location information could be stored using a geographic information system along with data layers such as vegetation type, slope, aspect, etc. The data recorded for each animal would consist of locations in space and time until mortality occurs or the study is completed. Two approaches seem feasible for modeling survival as a function of covariates (such as vegetation type, slope, aspect, age, weight, etc.). Cox's (1984) proportional hazard model is a well-developed approach, and the use of logistic regression is another approach. The �332 logistic regression approach is useful if relocations will occur over large discrete time blocks (e.g. if animals are relocated weekly), but this approach is not feasible if the number of time intervals is large relative to the number of animals in the study. An advantage of both approaches is that maximum likelihood theory can be employed. This is useful for parameter estimation, goodness of fit testing, and model selection. The proportional hazard rate model is as follows. Assume we have n animals and that animals will be followed "somewhat" continuously over a specified period. The model is where 1; is the time of death of animal i, Aa(t) is the underlying hazard function for all animals, B is the vector of regression parameters, and X is the vector of covariates (age, sex, weight, and habitat). The final regression form becomes The framework for the logistic approach is as follows. Assume we have n animals and that the animals will be followed over K time periods. When animals are first captured, we record age, sex, and weight. At each relocation, we record habitat type (as mentioned earlier other covariates could also be measured such as elevation, slope, weather, etc., but for simplicity assume only habitat type is measured). A model can be built with the following parameters: where m, is the probability of death in the kut interval given that the animal was alive in the k-lst interval with covariates x a function of age, sex, wt, and habitat. A logistic regression model for fit is then The logit of m, is a linear function in the covariates: this would be the linear function for one animal during interval k. So, in general animal i dies with probability mk(xJ and lives with probability I-mk(xi). If the study had k= 10 intervals, then animals still alive at the end of the study would have 10 records of data (Y =0 for each interval, i.e. the animal survived each interval). An �333 animal that died during the 4th interval would have 4 records of data (Y =0 for 3 intervals, i.e. the animal survived 3 intervals and Y= 1 for the 4th interval, i.e. the animal died during the 4th interval). In addition, the covariate information would be included with each interval record. Monte Carlo simulations using the proportional hazard model outlined above were performed. A simple model based on survival of young and adults was constructed to evaluate the simulation process and the analysis potential using SAS's PROC LIFEREG. A simple model with )...o{t) = 1 and only one covariate, age was simulated. Age was divided into 2 age classes (young and adult), and survival rates (S) for the young and adults along with the percent coefficient of variation (%CV{S» for the simulations are shown below. Each simulation was run with a total of 50 and 100 animals, half were young and half were adults. The study was assumed to run for 120 days (4 months). Using Monte Carlo simulation techniques, times of death {TJ for each animal were generated. Simulation case 0 resulted in a nonsignificant effect (a = 0.05 level) as expected, but only case 4 with S = 0.6 for young and S = 0.9 for adults and %CV{S) = 15 resulted in a significant age effect (p=O.OOOI). Cases 1-3 all would have resulted in Type II errors, indicating that the power of the tests is poor under those conditions. Power of the tests can be increased by: 1) increasing the number of animals in the study, 2) possibly the ratio of young to adults, 3) length of the study, 4) increasing the difference in the survival rates, and 5) decreasing the %CV{S). Still, preliminary results indicate that the proportional hazard rate model can be used to detect differences in survival rate. SIMULAnON # AGE 0 YOUNG 0.8 25 ADULT 0.8 25 YOUNG 0.6 25 ADULT 0.8 25 YOUNG 0.6 15 ADULT 0.8 15 YOUNG 0.6 25 ADULT 0.9 25 YOUNG 0.6 15 ADULT 0.9 15 1 2 3 4 avg{S) %CV{S) �334 REFERENCES Cox, D.R. 1984. Analysis of survival data. Chapman and Hall, London, England. Hobbs, N.T., and T.A. Hanley. 1990. Habitat evaluation: do use/availability data reflect carrying capacity? Journal of Wildlife Management 54:515-522. Van Home, B. 1983. Density as a misleading indicator of habitat quality. Journal of Wildlife Management 47:893-901. �335 JOB PROGRESS state of Project Colorado No. ~W_-~1~5~3~-~R~-~5~ Work Plan No. __~l~O~A~ Job No. Period Author: REPORT _ Mammals _ Kit Fox studies 1 Covered: Research Kit Fox Status in Colorado July 1, 1992 to June 30, 1993 J. P. Fitzgerald and M. Link, Univ. Personnel: T. Beck, C. Parmeter, of Northern Colorado). (volunteers Northern Colorado, M. Reddy, T. Hugo Greeley from University ABSTRACT Since 9 March 1992, 9 kit fox have been trapped in 2,725 trap nights of effort including 1,141 trap nights in the current reporting period. All individuals (5 males and 4 females) have been captured NE of Montrose in Delta and Montrose counties in desert-shrub/greasewood habitat. Three males and 4 females have been radio-collared and are periodically being monitored. One additional kit fox has been observed while spotlighting in Browns Park in extreme northwestern Moffat county. Den sites used by foxes are in steeper, rougher terrain than reported in the literature. Few kit fox have been reported harvested by trappers responding to the annual survey questionnaire distributed by the Colorado Division of Wildlife. Contacts with ranchers, land management agency personnel, CDOW employees, and local trappers have resulted in only three reports of kit fox in western Colorado. Live trapping efforts in areas from which kit fox have been historically reported have not resulted in any captures or visual observations. ��337 KIT FOX (VULPES MACROTIS) DISTRIBUTION James P. Fitzgerald AND ABUNDANCE and Michelle IN WESTERN COLORADO Link P.N. Objective Document the geographic western Colorado. distribution and relative Segment 1. 2. 3. 4. 5. abundance of kit fox in Objectives Identify geographic extent of kit fox distribution in western Colorado. Determine habitat use by the species and general prey species abundance in areas used by kit fox. Compile general harvest information. Test different survey techniques in areas with kit fox populations. Identify suitable kit fox populations for more intensive ecological research. Methods Methods are generally the same as previously reported (Link and Beck 1992). However, captured kit fox are now being collared with transmitters (Model No. HLPM 2180 LD - Wildlife Materials, Inc.) for monitoring purposes. Results Determination and Discussion of Distribution Since the project began 9 March 1992 a total of 2,794 trap nights of effort have been expended (Table 1) resulting in capture of 9 kit fox. Seven of the captured fox, 3 males and 4 females have been radio-collared. Two of the females have been recaptured and their collars have been replaced. One adult female captured on 31 May 1992, rehabilitated in captivity with a broken jaw, and radioed and released to the wild in late summer 1992, was found dead from injuries (probably inflicted by a coyote) in mid December 1992. All animals have been captured in a 2590 ha area of Peach Valley in Montrose and Delta counties. Live trapping (1,210 trap nights) and spotlighting efforts (54 hours) during the fiscal year were concentrated in the Big Gypsum, Disappointment, Paradox and Peach valleys in southwestern Colorado and the Rangely-Dinosaur and Brown's Park areas of northwestern Colorado. Because of the capture of kit fox in the Peach Valley area in May and June of 1992 the area continued to receive periodic, concentrated search effort during the present reporting period. One animal identified by Link as a kit fox was observed by spotlight in the Browns Park area in late June 1993 but no animals were captured during 144 trap nights of effort in that location. Concerted effort will be expended in that area in 1993-94 to attempt to capture and verify presence of fox in Moffat County. Of particular interest is whether animals in Moffat County are "Swift or Kit" foxes as it is possible that "Swift" fox could occupy the "Red Desert" area of Wyoming and filter south into Moffat County rather than represent extensions of Kit fox populations from the Colorado River drainage of Utah. Sampling has included trapping in a number of different habitat types including: sagebrush-grasslands; margins of pinyon-juniper woodland, and desert-shrubjgreasewood areas. To date, kit fox have only been captured in the desert-shrubjgreasewood type common to the Peach Valley area. Much of the area occupied by kit fox in Peach Valley is dominated by mat saltbush (Atriplex corrugata) on the uplands and ridges with greasewood (Sarcobatus �338 Table 1. trapping, County Areas searched, dates of search, kit fox study, western Colorado, and Area Montezuma Co. McElmo Canyon San Miguel Co. McIntyre Canyon Disappointment Valley Big Gypsum Valley Montrose Paradox Co. Valley Dates Searched and trap nights of effort 1992 to June 30, 93. Total Nights Trap Effort 60 o 3/5-8/93 77 o 8/3-9/92 141 o 7/5-8/92; 7/29-8/1/92 174 o 8/10-20/92 141 o Montrose Co., Delta & Mesa Co. Sinbad Valley 3/15-22/92 Delta to Grand Junction 4/28-5/16/92 Peach Valley Various - May to Present 168 o 192 836 o Mesa Co. Dolores River Valley-Gateway to Utah border 114 o 78 o 528 o Colo. Natl. Mon. 4/20-23/92 N.W. Grand Junct. & Rabbit Valley Various - 1992-93 in Fox Captures 100 Trap Nights 9/9-12/92 3/6-12/92 spent 1 Rio Blanco Co. Rangely/Dinosaur 6/16-19/93 72 o Moffat Co. Brown's Park 6/22-28/93 144 o vermiculatus), Nuttall saltbush (~ nuttallii), shadscale (~ confertifolia, and sagebrush (Artemisia ~) occurring in more localized sites. A considerable amount of the Peach Valley area occupied by foxes is on lands used heavily by off road vehicles (ORV's). Habitat Use and Relationship of Prey Abundance to Kit Fox Abundance Determining the diets of kit foxes in the Peach Valley area will be part of the 93-94 work effort. Link has collected numerous kit fox scat samples from Peach Valley. Remains of white-tailed prairie dogs (Cynomys leucurus), cottontail (Sylvilagus audubonii), and cricetid mice (Peromyscus) have also been found at kit fox dens. No usable relationship appears to exist between numbers of lagomorphs spotlighted at night and distribution of kit fox. An average of .2 lagomorphs (mostly cottontails) have been recorded per road km in over 916 km surveyed (range 1.6 to 0 animals/km). In Peach Valley the average has also been .2/km in a total of 68 km censused. Lagomorph populations have been very low across �339 western Brown's Colorado with highest Park areas. numbers observed in the Sinbad Valley, Mack and Den sites located in the Peach Valley area are not similar to those reported by other workers (Egoscue 1962, O'Neal et a1. 1986, McGrew 1977) who have found most dens on relatively flat ground with multiple burrow entrances. Out of 13 dens being used in Peach Valley, 4 are located in deep, dry washes, 4 are on the sides of hills with good drainage, two are half way up hills in washes, two are on top of ridges at 5,600 feet, and one is on the side of a large U-shaped bowl. Most dens are within 300 m of a well-traveled road, jeep trail, or pathway (ORV, horse). Use of these "atypical" den sites has led us to revise search efforts to focus on more broken terrain than suggested by the literature. General Harvest Information/ Other Reports of Kit Fox Table 2 indicates reported harvest data for kit fox from western Colorado. Take is estimated from trapper survey data and numbers may include other fox species. The numbers of animals trapped in Montezuma county in 1981, and reports of kit fox from La Plata county are doubtful. Table 2. Estimated harvest of kit fox from western Colorado based on trapper questionnaire returns to the Colorado Division of Wildlife, 1975-1992. Year County Delta Garfield Gunnison La Plata Mesa Moffat Montezuma Montrose Rio Blanco Totals 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 2 5 3 2 6 2 3 1 13 33 5 2 3 26 4 10 5 192 2 7 5 6 10 1 2 33 41 198 4 41 13 2 Link has made a concerted effort to question ranchers, agency personnel, or local trappers regarding their knowledge of kit fox. Few reports have been made. The district wildlife manager (DWM) in the Mack/Grand Junction West area reported seeing young kit fox in an arroyo north of Mack in 1988. Extensive trapping has not resulted in any captures in that area. The Fish and Wildlife Service at Browns Park National Wildlife Refuge had no records of kit foxes, but one employee (Brad Pech) reported that trappers had taken kit fox in that area several years ago. Link did spotlight an animal believed to be a kit fox in that area but none were trapped. The DWM (Divergie) and the Chevron Oil Company's on-site environmentalist (Sellars) at Rangely have never observed or heard of kit foxes in that area even after extensive spotlighting and surveying for Black-footed Ferrets. Ranchers (Tozer Bros.) from McElmo Canyon, Montezuma County, have not observed any kit fox and trapping efforts did not produce captures in 1992 in that area, although there are past records of kit fox in that area (Egoscue 1964). Likewise, trapping efforts in other areas reported to have kit fox by Miller and McCoy (1965) including the Colorado National Monument and near the Utah state line in the Grand Valley) have not resulted in any captures. The BLM ranger (Sering) assigned to the Delta and Montrose area reported to Link that he observed a kit fox on Flattop Mountain south of the Peach Valley area in June of this year. The Flattop area will be trapped and surveyed in 1993-94. �340 Test sampling techniques in areas with Kit Fox The Peach Valley area is the only site with a known population of kit fox. Those animals have been located by daytime sightings at dens close to roads or by live-trapping. Trapping success in Peach Valley is low with one capture per 100 trap nights effort. Two of the radio-collared animals (both females) have been recaptured 5 months after their first capture allowing for replacement of their radios. Males appear to be more difficult to recapture using present techniques. Scent stations baited with standard lure (scented predator survey discs with FAS, Pocatello Supply Depot) were placed at 10 sites in Peach Valley in June 1992. No kit fox tracks were identified at any stations. The scent station surveys were abandoned because it was more efficient for one person to work a trap line than to maintain a station line. Spotlighting has resulted in Browns Park. Identify suitable in observations kit fox populations of 2 kit fox in Peach Valley for more intensive and 1 research The Peach Valley area has from 4-8 adult fox based on monitoring of radioed animals. Since these animals represent the only kit fox presently known to occur in western Colorado the decision was made to radio-collar all animals captured on the site and to increase efforts to better understand habits of the animals in that area with the expectation that it may enhance efforts to locate other kit fox populations. A temporary worker (Parmeter) was hired from mid-January to mid-April to monitor fox in Peach Valley and to try winter trapping in the Grand Junction and McIntyre Canyon areas. Attempts to radiotrack animals at night in the Peach Valley area by both Link and Parmeter have not been effective for determining distances ranged from dens or areas being hunted. The foxes are aware of the researcher and their behaviors seem to be more intent in evading the tracker rather than on night hunting. In the 93-94 field season attempts will be made to track some radioed animals using teams of trackers stationed at high points of ground rather than a single individual. By: ames P. Fitzgerald Contractor, UNC and Michelle Link Graduate Research Assistant �341 Colorado Division Wildlife Research July 1991 of Wildlife Report JOB PROGRESS state of Project Colorado No. W-153-R-4 Work Plan No. llA 1 Job No. Period Author: REPORT Covered: Mammals Research Multi-Species Investigations Predicting the Impacts of Environmental Change: Simulations of Genetic and Species Diversity at Landscape and Regional Scales July 1, 1992 - June 30, 1993 N. T. Hobbs, J. E. Gross, and J. Miller Abstract Our long term goal is to understand and predict the effects of changes in spatial arrangements of plant communities on the diversity of vertebrate species within ecosystems (Fig. 1). In addition, we seek to develop ways to identify individual species that are placed at particular risk of extinction by a given change in land cover. Here, we outline our progress toward these goals. ��343 Predicting the Impacts of Environmental change: Simulations of Genetic and Species Diversity at Landscape and Regional Scales P. N. OBJECTIVES 1. 2. Develop an object oriented, individual-based simulation model of vertebrate population dynamics. Apply the model to determine effects habitat fragmentation on genetic diversity and species extinction probabilities using bighorn sheep as an example. Prepare pattern a study plan for simulating effects on vertebrate species diversity. of changes of in landscape RESULTS This report is divided into six sections. We begin by clarifying our objectives. Subsequently, we relate those objectives to the goals of other initiatives underway at the Division of Wildlife, particularly GAP analysis. We then outline a theoretical approach to forecasting risks of species extinctions based on metapopulation dynamics, and we describe a prototype model illustrating that approach. After describing our model, we discuss how it could be made more realistic by deriving input data from a geographical information system. We then briefly examine an approach to model testing. Finally, we outline several unresolved issues in model development and sketch our current efforts to resolve them. I. Objectives Understanding our objectives requires some background. A changing spatial context is fundamental to the function of virtually all ecological systems. By spatial context, we mean the geographic arrangement topographic, hydrological, and biotic features that influence ecological processes. These features change at different rates. The very notion of "geological time" reflects the slow rate of change in physical features of landscapes (e.g., topography), features that change only at multi-thousand year time scales. An exception to these slow changes in physical features can occur when man alters the physical environment by building roads, cites, dams, etc. Biotic features (e.g., plant cover) may be also shift relatively rapidly, over time scales of 1-100 years. Such shifts may occur, for example, as a result of plant succession or from natural disturbances like fire and flood. Here, we focus on the effects of these relatively rapid changes in biotic and physical features on the diversity of vertebrate fauna. In the past, a changing spatial context has been referred to rather generically as "habitat fragmentation", a concept that unifies many aspects of the emerging discipline of Conservation Biology. However, the concept of habitat fragmentation has limitations--it reflects a an emphasis on singlespecies approaches to management, and on anthropogenic effects on their habitats. The idea of fragmenting space requires that we think of land cover in terms of islands of suitable habitat surrounded by an unsuitable matrix. Distinctions on "suitability" must be made in reference to some standard, usually the habitat requirements of individual species or guilds. The emphasis on man-caused shifts in the spatial context has emerged because the activities of man are, and will continue to be, one the most important sources of variation in land cover. However, although the idea of fragmentation is clearly a necessary concept in understanding a changing spatial context, it is probably not a sufficient one. Global changes in climate are likely to cause fundamental shifts in composition of plant communities as well as in their juxtaposition--in essence we are confronted with understanding the consequences of a changing matrix as well as changing islands. Moreover, to understand and predict changes in wildlife diversity, we must expand our outlook to include communities of �344 species. This complicates the idea of fragmentation because cover types that are hospitable to one member of a community are hostile to another. Thus, understanding the effects of changes in land cover on the diversity of wildlife species challenges us to develop a broader concept of the spatial context, a context driven not only by the activities of man, but also by the forces of a changing climate. We are challenged to think of the effects of fragmentation on multiple species rather than single ones. Wildlife diversity represents the quantitative balance between rates of speciation (the creation of new species) and rates of extinction (the loss of existing species). For simplicity, we will assume here that the rate of speciation is effectively constant. It follows from this assumption that understanding variation in diversity requires that we understand the ecological processes controlling extinctions at the level of the individual species and at the level of multi-species communities (Fig. 2). We propose to synthesize current understanding of the processes controlling rates of extinction of terrestrial vertebrates. The product of our synthesis will be a simulation model of vertebrate population and community dynamics, the objectives of which are as follows: 1) To understand and predict the effect of variation in vegetative cover types on the probability of extinction of vertebrate species with a given life history (Fig. 3) Thus, we want to understand how differences in spatial context influence the persistence of a single species with a fixed life history. 2) To understand and predict the effects of variation species' life histories on their probability of extinction within a given spatial context (Fig. 4). Thus, we want to understand how differences in life history within communities influence extinction in a fixed spatial context. 3) To understand and predict the consequences of changes in the spatial context for the richness and composition of vertebrate communities (Fig. S) • II. Relationship to Ongoing Projects The Division of Wildlife has invested heavily in analyzing the availability of habitats for the state's vertebrate fauna. A large portion of this investment is manifested in GAP analysis, an approach which provides a comprehensive inventory of the geographical distributions of native fauna. This inventory is based on the relationship vegetative cover types and the likelihood that a species will be found at given geographic location. GAP analysis provides a much needed survey of the current status of habitats for wildlife in Colorado. However, the view of species diversity provided by GAP is static; GAP does not attempt to infer the status of populations or communities. It emphasizes the current presence or absence of species, and it makes no qualitative or quantitative distinctions between species for which habitat conditions are adequate to assure the long term survival, and species for which current conditions will almost assuredly lead to extinction. What is needed, then, is an extension of GAP, an approach that extends GAP's spatial inventory to make predictions on the future status of species and communities (Fig. 6). We propose to develop such an extension. III. A Theoretical Model and Allometry. Based on Metapopulation Dynamics, Risk Assessment, A.. Background We are interested in developing a spatially explicit model predicting effects of changes in land cover on wildlife diversity, that is, on the number of species present in some spatially bounded area at some time in the future. �345 What are the consequences of landscape change? LANDSCAPE A LANDSCAPE B eoosa.ra Figure 1. Changes in landscapes are likely to occur as a result of impacts of man and as an outcome of plant succession. We are working to predict the consequences of those changes for wildlife diversity. �346 Modeling Biodiversity Requires Modeling Extinctions til til 4) C s: 0 n: ..._ ._- , ... __ •.... til 4) ._ 0 4) c. (J) Years i -.2 c - c x • Sum across species .----. .. ~ .• • .- •• -- . .• 0 x 10 4) Q. Years x 10 Figure 2. Changes in wildlife diversity result collectively from changes in the rate of extinction of individual species. �347 What are the consequences of the spatial context for a given species? • • -2 c Species --->~ .-g A ••x ..., ~\ ~ ...... ..~_.- _.- . Q.L.....- _ Years x 10 • • • • •4 • • ~ • • •••••• c Species A • - ,.' • 0 •• )g .••>< Q) I I .. .. • • - , Q. Years x 10 spaspc1.tc Figure 3. Our first objective is to predict how changes in spatial pattern of plant communities influences the probability of extinction of a species with at given set of life history characteristics. �348 What are the consequences of species life history characteristics in a given spatial context? c: -o ._oe xQ) - ~ Years Species I • 10 B • • x . •4 • •- -2 •• ... c:: Species _ ••• I .,• ---->~.~•...,.. A # •• x Q) a. Years x 10 spaspc2.tc Figure 4. Our second objective is to understand how a species' life history influences its probability of extinction in a given spatial arrangement of plant communities. �349 (f.) Q) c:: ..c:: o ....... _---- l_, __ ------ 0: .• _--- • .. " (f.) .. Q) o Q) c. en Years x 10 1 o g 1 (f.) Q) c:: s: o .0: (f.) Q) I •• , .... I 1 - I - . •...... .... _ o 1_- _ .• .. Q) c. en I Years x 10 ftCIIWttn.tC Figure 5. Our third objective is to understand and predict effects of changes in spatial pattern on the richness of communities of vertebrates. We are particularly interested how spatial scale and body size affect this prediction. in �350 GAP Models Use vegetation data to predict presence/absence of species. Use vegetation data and species life history characteristics to predict population trajectories and extinctions. Metapopulation Models Use vegetation data and food web theory to predict community interactions. Community Models Figure 6. We plan to extend GAP models in two steps. First, we will use patch maps derived from land cover data to model metapopulation dynamics. Later, we will model interspecific interactions, including predation and competition. �351 Before we describe are as follows: our modeling approach we need to define some terms. These A landscape is a geographic area of sufficient size to include one or more populations of several species. It could be a watershed, a county, or tract of public land. We view a landscape as defining the spatial limits, that is, the domain of model predictions. A cover type is a vegetation canopy that defines the spatial extent of a plant community. Thus, we view a landscape as mosaic of cover types. We presume that cover types can be delineated using remotely sensed data obtained from aircraft or satellites. A patch is an area of a landscape that is capable of supporting at least one breeding pair of a given species. Thus, a patch is defined by the requirements of a species for cover types the availability of those types within a landscape. A landscape is composed of patches (the aggregations of suitable cover types) and matrix (the aggregation of unsuitable cover types). Initially, we will assume that a patch is spatially continuous (i.e., is uninterrupted by matrix) and that individuals of a population move freely within a patch. Each patch can support a finite number of individuals of a species. Dispersal is the process of individuals moving among spatially distinct patches. Whether animals move freely among patches (i.e., across matrix) depends on characteristics of the mosaic of patches and on characteristics of the species. Extinction landscape. is the loss of all individuals The probability will go extinct of a species from the of extinction is the likelihood that a given during a specific time interval. species We begin by assuming that we can define a patch structure for any given landscape by making the traditional distinction between patch and matrix, between areas that offer suitable habitat and areas that do not. Defining patch structure is most easily understood by example (Fig. 7). A patch consists of 1 or more spatially contiguous vegetation cover types that offer habitat suitable for a species. It follows that for any given landscape, specialist species (those requiring a narrow range of cover types) will have more fragmented patch mosaics than generalist species (whose requirements can be met by a broad range of types). Thus, the habitat requirements of a species are used to define the spatial distribution of a set of populations connected by immigration and emigration. Developing such definitions is the essence of GAP analysis. GAP analysis can be extended as follows. Given a patch structure, we can model the dynamics of populations within each patch as a function of speciesspecific life history characteristics that determine survival and reproduction, the carrying capacity of the patch, and dispersal among patches. Differences in population dynamics are driven by differences in patch carrying capacities and distances among patches. Thus, carrying capacities and interpatch distances mediate the effect of space on the probability of extinction by linking patch structures to population trajectories over time. Once this link is achieved, we can examine the effects of a changing spatial context on a single species by changing the patch structure (Objective 1, Fig. 3) and can examine the effect of species life histories in a fixed spatial context by changing rates of survival and reproduction as well as dispersal abilities (Objective 2, Fig. 4). Assuming that species extinction probabilities are independent (an admittedly infirm conjecture--see section A in Unresolved Issues, below), we can estimate the expected change in the number of species within a community, and their relative abundance (Objective 3, Fig. 5). Our approach to these estimates is illustrated in the next section. �352 B. Model Formulation We are now developing a prototype model representing the linkage between spatial arrangements of patches and the dynamics of populations in a stochastic, density dependent fashion: 2 N .. ~Jt'l =N .. ~Jt +(r.+Z.)N ~ .. ~Jt ~t +I )_E .. _r.(N1Jt ~Jt ~ K1J (1) ... ~Jt where t = index of time steps, = index of species, j = index of patches, Nij= number of animals i ri = rate of increase of species of species i in patch j, i, Z = a normally distributed random variable with mean 0 and variance S2. (S2 is a measure of the variation in r resulting from temporal variation in the environment), Iij number of animals ~j carrying Eij number capacity of animals of species of patch i immigrating j for species of species emigrating into patc~, i, from patch j. This equation is not as complicated as it might appear. It simply says that the number of animals of species i in patch j at time t+l is given by the population at time t (term 1 on right hand side of =), plus a stochasitc rate of increase (term 2), plus the rate of emigration into the patch (term 3), minus a density dependent mortality rate (term 4) minus the dispersal from the patch (term 5). In this simple model, the spatial context is represented in the patch structure of the landscape. The patch structure, in turn, controls population dynamics and the risk of extinction by influencing K, E, and I. C. Model Implementation We are coding the model in c++ using object oriented techniques. Initial versions are targeted for the UNIX operating system (Sun OS 4.1.2), but we have our eye on portability, and are writing code that should compile on DOS or Mac systems with minimal modification. We are constructing both molar and individual based versions, both incorporating the dynamics represented in equation 1. Molar models will use the population as the fundamental unit of simulation; individual based models will represent many, discrete animals. Both approaches will incorporate environmental and demographic variation over time following eq. 1. The molar version will apply best at large spatial scales, while the individual-based version will offer greater flexibility for adding detail at finer scales. In our initial model exercises, we will examine how the effect of spatial arrangement of habitat types depends on animal body size. We will implement the model using Monte Carlo simulation as follows. We will construct a broad array of spatial arrangements of patches using techniques of fractal geometry. Each arrangement will be defined by a mean and a variance in patch size and a mean and a variance in distances among patches. For simplicity in our initial prototype, patches will be treated as circles (Fig. 8). Carrying capacities will be proportional to the area of each circle, and patch mosaics will be specified by the circle diameters and the distances among their centroids. �353 Patch Mosaic Cover Mosaic D Type 1 II Type 2 MMml Type 3 m Type 4 species adapted to type 2,3,4 species adapted to type 3 • - ~ • - • •4 • - • Figure 7. Knowledge of a species' habitat requirements allows us to covert a map of vegetation types into a map of patches suitable for the species. �354 = Habitat Patch Figure 8. Our model will be driven by spatial data on patch size and distance among patches. During model development, we will assume all patches are circular in shape. �355 (Later, we will show how this obviously simplified approach can be made more realistic using geographic information systems). Emigration and immigration rate will be calculated as a function of minimum distances among patch boundaries (see section B in Unresolved Issues, below). To incorporate demographic stochasticity, birth rate will be represented as a poison random variable and survival rate will be represented as a binomial random variable. Drawing values randomly from these distributions, we will conduct 100 model runs simulating population trajectories over 150 year time intervals for each set of patch conditions and for each species. For each simulation, we will record the maximum and minimum predicted population size. We will use these model runs to accumulate a frequency distribution (e.g., Fig. 9) of predicted times to extinction for each species. We will also plot the probability of the population falling below a range of threshold population sizes (Fig. 10). These plots will be used to illustrate the effects of changing spatial context for a species with a given combination of life history traits (determining r, K, E, an I) and will be used to illustrate the effects of variation in life histories within a given spatial context. We can estimate the effects of changes in spatial relationships at the community level by deriving the expected value of species richness (R) based on the potential number of species in a community (N) and their individual probabilities of extinction [P(e»), N R = N' (1-11 P(e) j) j-1 This statistic departs from traditional, species specific approaches to population viability by accumulating small effects across the entire community. It embodies the idea that species diversity (a community level phenomena) is the cumulative result of extinctions (a population level phenomena) (Fig. 2). Thus, minor effects on the probability of extinction of individual species could be manifested collectively in major impacts on the expected value of species richness at the community level. D. Estimation of Model Parameters We confront formidable problems in estimating values for parameters in eg. 1. A community of vertebrates may contain a hundred or more species, and the life history information needed to choose reasonable values for the parameters in eg. 1 may be available for only a small portion of them. In the absence of such information how can we project the effect of a changing spatial context on a species with poorly described life history? How can we infer which species might be at particular risk of extinction in a given spatial context? How can we infer the impacts of a changing spatial context at the community level? A potential solution to the problem of parameterizing the model for many species is to use allometric relationships rather than detailed life histories to estimate parameter values. For example, we can reasonably assume that the intrinsic rate of increase (rmu) of mammals scales to the body mass raised to the .27 power, a relationship that is supported by both theoretical arguments and empirical data. Similarly, maximum population density (K) of mammalian herbivores has been shown to scale to body mass to the power of .77. If we can derive scaling relationships for dispersal abilities and for the environmental variance in r (see sections Band C in Unresolved Issues, below), then we can estimate values for equation 1 entirely on the basis of species body mass. In essence, we are using information on life histories of well studied species to estimate the values of parameters for poorly studied ones via relationships between body mass and demographic processes. Our scaling-based approach might appear tenuous--obviously, animals of the same size mayor may not have similar demographics. This is simply to say that there is substantial scatter around any scaling relationship. However, this not defeating, it simply means that we need to be careful in how we use �356 >. o c: CD ::J 0CD 10 aL: o @ - Habita. 0 0 - N M ~ ~ ~ ~ 000 = e=~~ 0 0 0 0 0 0 Q Time to Extinction Patch 1 0 0 How does patch structure affect population viability? 1 >. o c: CD ::J 0CD 10 u: 000 _ 0 0 0 000 N M ~ ~ ~ ~ = 000 m Time to Extinction 000 _ N M ---0 Figure 9. Output from many simulations will be accumulated estimate a frequency distribution for time until extinction each species in a given landscape. We will examine effects changes in simulated land pattern on these distributions. to for of �357 .?: 0.8 :c ~ 0.5 o a: 0.3 ®. Minimum Population Size During 150 Years Habil.t Patch 1 1 How do changes in patch structure affect population viability? ~0.8 :c ~ 0.5 ~ a.. 0.3 o _ ®- Hllbi'.t Patch 0 N coo M • ~ 0 ~ 0 ~ = ~ ~ ~ : ~ 0 0 0 0 0 0 Minimum Population Size During 150 Years Figure 10. Output from many simulations will be accumulated to estimate the probability that a population will fall below a minimum size. We will examine effects of changes in simulated land pattern on these probabilities. �358 the inferences of the model. We freely acknowledge that there is no substitute for detailed life history data in building detailed, species specific models and for comparing the status of similar species. However, if our purpose is to make inferences about many species in a variety of trophic groups, then such detail is likely to obscure rather than illuminate the broad patterns we seek. One way to look at our scaling approach is that it provides a "coarse filter"--a model based on scaling allows us to identify body sizes that are put at particular risk by a given change in the spatial context. Once those have been identified, more detailed models faithful to the biology of the species in that range may be required to support decisions on specific management actions. IV. Linking Imaginary Landscapes to Real Ones: Patch structures and GIS Thus far, we have offered, schematically, a theoretical approach for linking variation in the spatial distribution of populations to their probability of extinction. However, animals don't live in circles of habitat (except, perhaps, those inhabiting fields irrigated by center pivots), and we need a way to connect our theoretical model with information from the real world. We envision that this can be accomplished as follows (Fig 11). Two sets of information define a patch structure needed to 'drive the model represented in equation 1. First, each patch inhabited by each species has a carrying capacity (K) determined by patch size and by the body size of the species. Thus, for each species, we would require a list of the sizes of patches within a landscape. This assumes that each polygon defines the spatial boundaries on a population, which, of course, is not likely to be true. The ranges of some species could easily embrace several polygons. How to deal with these situations has not yet been resolved (see section D in Unresolved Issues, below). We will also need information in the juxtaposition of the patches because we will model dispersal (and hence, E and I, eq. 1) as some function of distances among patches (see section B in Unresolved Issues, below). Thus, we will also require a matrix that specifies the distance from each patch to all other patches in the landscape. Geographic information systems routinely provide algorithms for determining the areas of polygons, and for calculating minimum distances between polygon boundaries. Thus, using these algorithms, we can provide the data on patch structure from any given map of habitat types. In essence then, we plan to develop a general, abstract model, and apply it to specific, real-world locations using spatial data (Fig. 11). V. Model Testing Model testing is the process of creating opportunities for predictions to fail relative to objectives. The most revealing test of a model compares utility of its predictions against those of other, competing abstractions. One of our primary objectives is to develop tools that improve over GAP analysis in their ability to predict the consequences of landscape change for wildlife diversity. We envision a series of tests aimed at determining whether that objective is met. GAP analysis and our model both predict species diversity in a given spatial context. Our approach to model testing is to observe the frequency distributions of body sizes predicted by our model and by GAP for a range of spatially divergent landscapes. We predict that our model will predict that specific ranges of body sizes will be poorly represented in some spatial arrangements of vegetation as a consequence of the effect of patch size on K and as a consequence of the effect of distances among patches on I and E (eq. 1.) GAP analysis will not be sensitive to these differences. Thus we will generate two predictions, species richness and frequency distributions of body sizes, from each of the two models. We will compare these predictions with observations in a variety of landscapes. Initially, we may be able to rely on previous studies to glean information on species abundances in areas where �359 Step 1: Create patch map based on GIS data and speceis habitat requirements. • • •.. •- • • • • • Step 2: Derive model inputs (patch areas, distances among patches) from patch map. ~r--------------. Step 3. Simulate population d in patch network. Predict probabilities n,.. of extinction. tD Ex'i.clion 0-"........•.. gixoltc Figure 11. steps in modeling extinctions based on GIS data. We will create patch maps and use them to derive patch areas and distances among patches. These areas and distances will be used to drive our model (eq. 1). �360 areal photographs or other types we would like to design explicit VI. A. Unresolved Community of spatial data are available. Eventually, field tests of model predictions. Issues Level Interactions: Predation and Competition We assumed above that the population dynamics of all species that we model are independent of the dynamics of all other species. Made plain, this means that predation and competition do not change the way each population behaves. We quickly admit that this assumption is, without doubt, wrong. However, we believe that the problem of modeling species extinctions is dauntingly complex even when we make this simplifying conjecture. In the face of such complexity, we argue that it will be most productive to build a model we are satisfied with in the absence of interactions among species before .we attempt to include them. The alternative is bewilderment. Thus, we propose a stepwise approach. We will first extend GAP models to allow current spatial data to predict the relative likelihood of species extinctions based on metapopulation dynamics. We will then attempt to link the dynamics of populations at the community level by explicitly representing predation and competition (Fig 6). Current food-web models potentially offer tactics for achieving this linkage. B. Modeling Emigration and Immigration: Details of Dispersal Representing dispersal among patches is clearly a critical challenge in modeling the effect of spatial pattern on metapopulation dynamics. We confront a variety of unresolved issues in doing that. An attractive approach to predicting the likelihood that an animal will successfully disperse from patch a to patch b is to represent the probability of its success as a negative exponential function of distance between the 2 patches (Fig. 12). There are a couple of problems in implementing this approach. First, it is necessary to scale the relationship to body size to make it consistent with the approach we outlined above. We are now working on the theory of such scaling relationship and are gathering empirical results to test our ideas. In brief, we suggest that the animal's dispersal ability should scale with the ratio of its travel velocity to its energy reserves, where energy reserves are expressed in units of survival time at 0 energy balance. Both of velocity and survival time have known allometries. Second, it is unclear how differences in patch shape and size might affect the relationship between the probability of successful dispersal and distance between patches. We plan to clarify the effects of patch shape and size independent of distance using simulated animal movements in a variety of patch mosaics. C. Scaling of Environment Variance in Rate of Increase The behavior of equation 1 is affected by both the mean and the variance in species' rates of increase. Although intrinsic rate of increase can clearly be scaled to body mass, it is not clear how the effects of environmental variation on r depend on body mass. In other words, we need to develop a scaling relationship for the effect of environmental variation on r. Theoretically, we are exploring the possibility that this variation also scales inversely with endogenous reserves of energy. The reasoning behind this scaling is that high reserves of energy provide a buffer against energy shortages, thereby mitigating the effect of environmental variation on r. At some point, it will also be necessary to deal with spatial autocorrelation among patches in their effects on r. High levels of such correlation will lead to higher rates of extinction than low rates will. It follows that �361 (I) :::s Q) (I) Q) - en "fi 1 .0 tU c a.. o c: :::s en Q) Q) o _ ;: >.Q) _(0 ..c tU ..c o tU (I) •... Q) •... a. (I) a.. .- o Distance Between Patches 4iaperat.tc Figure 12. Several empirical studies show that the probability of successful dispersal between patches declines exponentially with increasing distance between patches. �362 maximum probabilities of extinction will be obtained by assuming that all patches are identical in the effects of stochastic environmental variation population dynamics. In the short term, we assume that such perfectly correlated environments prevail, but eventually we will need to understand proximity of patches in space affects how similarly they behave over time. D. What is dispersal? When are groups on how populations? If polygons representing the areas of habitats are sufficiently large, and if the polygons are sufficiently far apart, then we can reasonably presume that each polygon represent an area occupied by a spatially distinct population. However, precisely defining what we mean by "sufficiently" is a prerequisite to implementing our model. Our line of attack on this problem is as follows. We will define a threshold distance based on animal velocities and foraging times that defines the maximum distance between habitat patches that allows them to be used as part of the animal's daily feeding regime (Fig. 12). If we plot probability of successful travel as a function of distance between patches, this threshold distance creates a "shoulder" in the exponential decline of the likelihood of success as a function of distance. We presume that when the distance between patches exceeds this threshold, then there is some declining probability that the animal will move successfully between them. We are currently trying to define relationship both theoretically and empirically. If we can do that successfully, then we may escape from the distinction between "dispersal" and "daily movements" and thus, may be able to model metapopulation dynamics without specify the exact spatial boundaries on each population. � https://cpw.cvlcollections.org/files/original/5c2c9c4b1ae5831d38a3b174fdef886b.pdf 576d9638fcb53b79b6de00e0aeace51d PDF Text Text 1 Colorado Division Wildlife Research October 1993 of Wildlife Report JOB PROGRESS State of __~C~o~l~o~r~a~d~o~ Project Author: Migratory 1__ : Job Job Title: Development Colorado Period _ W-166-R-2 Work Plan Covered: REPORT Game Bird Investigations 20 and evaluation I April of moist-soil 1992 through 31 March management techniques in 1993 James K. Ringelman Personnel: J. Goettl, T. Ostertag, Colorado Division of Wildlife. J. Ringelman, M. Szymczak, L. Swift, ABSTRACT A 4.0 ha (200 m by 200 m) shallow impoundment was selected for study in 1992. Located just southeast of Fort Collins, this wetland is on private property owned by Mr. Louis Swift, a Wildlife Commissioner for the State of Colorado. Physical characteristics and a water depth profile for the impoundment were obtained on 4 June 1992. Water depth, determined from measurements taken along transects at 10 m intervals, ranged from 15-70 cm, with a mean of 40 cm. On 7 June, stoplogs on the single, half-round riser were removed to achieve a drawdown of 45 cm. However, the Fort Collins area began experiencing abnormally high rainfalls shortly thereafter, a trend that continued through the month of June. Ultimately, June rainfall totaled 230% of normal (4.45 cm normal; 14.68 June 1992). Because the small control structure could not accommodate the increased flows and direct precipitation, a partial drawdown was not achieved until mid-July. Instead of prqmoting germination of annual, se~d-producing plants,. the shallow marsh environment resulted in a vigorous' stand of rhr ee squaxe bul.rusb' (Scirpus americana) ,·.with minimal germination of·smartweeds (Polygonum spp.)" curlydock (Rumex crispus), wild millet (Echinochloa crusgalli), and other moist-soil plants. Later, . filamentous green algae became common, due in part to water stagnation. Because bulrush is not considered a moist-soil plant, seed production and fall waterfowl use were not measured in 1992. This impoundment will be manipulated in spring 1993 in a followup attempt to control bulrush and achieve moist-soil plant germination. Prepared by: James K. Ringelman Wildlife Researcher C ��3 Colorado Division of Wildlife Wildlife Research Report October 1993 JOB PROGRESS REPORT State of Colorado Project W-166-R-2 Work Plan Job Title: _1_ 22 Harvest distribution of mallards and pintails banded preseason in western Colorado Period Covered: Author: : Job Mi~ratory Game Bird Investi~ations 01 April 1992 through 31 March 1993 Michael R. Szymczak Personnel: J. Gamble and staff, Browns' Park National Wildlife Refuge; K. Nelson, U. S, Forest Service; D. Coven, P. Creeden, J. Ellenberger, V. Graham, J. Gray, J. Gumber, D. Masden, J. Miller, B. Motz, J. alterman, R. alterman, N. Smith, M. Szymczak, and K. Wagner, Colorado Division of Wildlife ABSTRACT Ducks were trapped in modified Salt Plains bait traps and banded on 17 different wetlands located in 5 areas across western Colorado in August and September, 1992. About 1,700 mallards (Anas platyrhnchos) were banded, with the number captured generally well distributed between trapping areas. Mallard trapping efficiency increased in 1992 over 1991 levels. Increases in total number and trapping efficiency were greatest on the Browns' Park National Wildlife Refuge. Excluding Gardner Park, an average of 8 mallards were banded per trap/day. Only 42 northern pintail (Anas acuta) were banded. Most birds.reported recovered were taken in western Colorado. ��5 HARVEST DISTRIBUTION OF MALLARDS AND PINTAILS COLORADO BANDED PRESEASON IN WESTERN P. N. OBJECTIVE 1. Document captured the distribution qf band recoveries preseason in western Colorado. of mallards and pintails 2. Determine if the geographic location of recovery of mallards pintails is dependent on the area of banding in Colorado. 3. Determine the relationship between the recovery distribution of western Colorado banded mallards and pintails and the distribution of recovery of those species banded in other areas of the Pacific Flyway. 4. Cooperate in analysis of Pacific Flyway - wide band recovery and preparation of reports. and data SEGMENT OBJECTIVES 1. Trap and band mallards and pintail in 4-5 areas of western Colorado in late August-early September using salt plains bait traps (Szymczak and Corey 1976). Recommended areas are: (1) Browns' Park National Wildlife Refuge, (2) Yampa River Valley below Craig, (3) Colorado River Valley below Glenwood Springs, (5) Uncompahgre-lower Gunnison River Valley, and (6) the Cortez - Mancos area. 2. Submit banding schedules and recapture reports to the U. S. Fish and Wildlife Services' Bird Banding Laboratory. Summarize and file band return reports. INTRODUCTION Recent drought in the northern Great Plains of the United States and Canada has reduced many waterfowl species to historically low population levels in surveyed habitats. Mallards and northern pintails have been especially effected. Mallards reached near record low levels in Spring 1990, 27% below the 1955-89 average, while northern pintail populations reached a record low of 52% below the 1955-89 average (Reynolds et al. 1990). Despite low population levels and restrictive regulations, both species continue to be popular among hunters. In 1989, mallards composed 38% of the U.S. harvest and 40% in the Pacific Flyway harvest. Pintails, which made up nearly 25% of the Pacific Flyway harvest in the late 1960's and 1970's (U. S. Dept. Interior 1988:44) when all of the bag of 5 to 7 birds could be that species, still composed 9% of the harvest in 1989 (Martin et al. 1990) when only 1 of 4 birds could be a pintail. In western Colorado, pintail are not important in the hunters' bag (0.6%), but mallards made up 67% of the take in 1989 (Martin et al. 1990) �6 It has been assumed that many of the mallards and pintails wintering in the Pacific Flyway originate in unsurveyed breeding areas, west of the continental divide. Water regimes, and thus waterfowl production, in these areas are usually more stable and not tied to the wet-dry cycles characteristic of the mid-continent Great Plains. Childress (1986) examined relationships among duck breeding populations and fall flight estimates in surveyed areas, and duck harvests in the Pacific Flyway during the stabilized regulation period (1975-83). He found that recruitment in standard survey areas differed from that in unsurveyed areas. Pospahala et al. (1974: compiled from Table B- 20) estimated that over 600,000 mallards occupied western breeding habitats (excluding Alaska) outside the standardized surveyed breeding areas. Deficiencies in pre-hunting-season (preseason) mallard banding in western breeding areas was first mentioned by Anderson and Henny (1972:86). Munro and Kimball (1982:19) noted that continental mallard banding data suggested that a larger portion of the mallard harvest occurred in the Central Flyway than in the Pacific Flyway; however harvest surveys indicated the opposite. Trost (1985:4-5) reported that Pacific Flyway mallard harvests, derived from analysis of banding data, were underestimated because the appropriate breeding populations were not properly represented by banded samples. Smith and Powell (1989:4) concluded that a complete picture of the Pacific Flyway mallard harvest cannot be completed until additional samples of mallards are banded in the states and provinces of the western D. S. and Canada. Colorado has breeding mallards scattered throughout the Pacific Flyway portion of the state. Annual surveys along the Yampa and Green rivers indicate nearly 1,500 mallards breed along those riparian zones (CDOW unpubl.). Frary (1954) and Rutherford and Hayes (1976) found respective mallard breeding densities of 0.5 and 0.8 pairs/krn2 on the White River Plateau and among a variety of habitats in the mountainous areas west of the San Luis Valley. More recently, Ringelman et al. (1990) found mallards nesting at densities of 5.7 pairs\krn2 in montane habitats containing beaver flowages and kettle lakes just west of North Park. Although the latter 2 studies were conducted immediately east of the continental divide (Central Flyway), there is no reason to believe that habitats west of the divide (Pacific Flyway) don't support similar mallard nesting densities. With pintail populations at record low levels, obtaining information on derivation of harvest and harvest rates would be of value in adjusting harvest regulations and measuring population response. A special preseason banding project for pintails, including banding in western breeding reference areas was proposed in the continental duck banding program (D. S. Fish and Wildl. Servo 1989). Identifying and understanding distributional characteristics of Colorado's major waterfowl populations are important in designing strategies to reach population objectives (Colo. Div. Wildl. 1989). Preseason mallard populations in western Colorado have never been banded, and therefore population and harvest distribution characteristics are unknown. With continued drought in the Great Plains and dependence upon prairie duck breeding recruitment to develop fall flight duck forecasts, there is a definite need to identify those breeding populations that winter in the Pacific Flyway and determine the specific areas of harvest. A companion step, not part of this study, is to develop standardized surveys that will result in annual estimates of population size and recruitment in those breeding populations. Then annual assessments of ducks will include those specifically �7 identified as Pacific Flyway birds and regulatory adjustments can be based on the status of those populations. The Pacific Flyway Study Committee has formulated a 5-year cooperative mallard and pintail preseason banding program that has been endorsed by the Pacific Flyway Council. This program is designed to address banding needs throughout the western U. S., including Alaska, and in the provinces of British Columbia and Alberta (northwestern). Because Colorado supports Pacific Flyway breeding mallards and harvests Pacific Flyway mallards (x = 18,567, 1961-88), the Colorado Division of Wildlife has committed to cooperate in this program. Few pintails breed or are harvested in western Colorado. Therefore, banding in Colorado will be directed toward mallards. This report covers the second year of the banding effort. METHODS Trap Area Selection Most breeding mallards in western Colorado are associated with small wetlands that are widely distributed throughout high elevation, mountainous areas. Only in Browns' Park, located in the extreme NW corner of Colorado, are there extensive wetland complexes capable of supporting breeding ducks. Trapping on widely distributed wetlands with low densities of breeding ducks was assumed to be an inefficient method for banding western Colorado mallards. Therefore, strategically located areas, to which post-breeding and fledged Colorado mallards would move, were selected as primary trapping areas. In addition to the Browns' Park National Wildlife Refuge (BPNWR) , trapping occurred along the Colorado River Valley from Debeque to near Fruita (CRV) , in the Uncomphagre River Valley from Montrose to west of Delta (URV) , in the Cortez-Mancos area (CM) , and at Gardner Park Reservoir, about 5 miles west of Yampa. Trapping Period Ducks were trapped and banded in each area for a consecutive period of about 10 days beginning near the end of August and ending near 20 September. The actual banding periods were: BPNWR - 23 August thru 4 September; Gardner Park - 9 thru 23 September; CRV - 2 thru 12 September; URV - 31 August thru 10 September; CM 2 thru 11 September. Trapping and Recording Technique All birds were trapped in modified Salt Plains bait traps (Szymczak and Corey 1976) using whole shelled corn for bait. Traps were visited once a day and only mallards and pintails were banded except in Browns' Park. Banded birds were recorded by wetland site. Band numbers of all birds captured that were banded in previous years or outside the specific area of trapping were recorded. Records were also maintained on the number of traps operated by wetland in order to evaluate capture/unit of effort at each trap site. Band Recovery Locations A preliminary quarterly printouts and Rates listing of reported band recoveries was compiled from received from the U. S. Fish and Wildlife Service's Bird �8 Banding Laboratory. Recoveries were compiled according to area of banding age and sex and the results interpeted in a general sense. by RESULTS Trap Locations Trapping was distributed over a total of 17 different wetlands in the 5 areas (Table 1). Four non-productive 1991 sites (Szymczak 1992), Hog Lake (BPNWR), Walker Wildlife Area South (CRV), Grett's Dairy (URV), and Kintz's Pond (URV), were not trapped in 1992. Three new sites, Swietzer Lake and Sander's Pond in the URV area and Summit Lake in the CM area were added in 1992. Nearly all trap sites were located in Palustrine Emergent Persistent Wetlands, but some sites in the Colorado River Valley were in Riverine Upper Perenial Rock Bottom wetlands (Cowardin et al. 1979). Banding and trapping efficiency Over 1,700 mallards were banded during trapping in western Colorado in 1992 (Table 2). The number of birds banded per trap/day increased substantially in most areas. Total trapping success for mallards increased about 350% in BPNWR, 29% in the CRV, and 86% in the URV over 1991 levels. Only 33% of the mallards banded in BPNWR were immature birds. However, immatures mallards comprised 67 %, 78% and 72% of the of the CRV, URV, and CM samples, respectively. Throughout the trapping area 62.6% of the mallards banded were immatures compared to 68.3% in 1991. Excluding Gardner Park, 8 mallards were banded per trap day in 1992 (Table 2) compared with 5.6 in 1991. Trapping efficiency improved in BPNWR and the CRV. As in 1991, more mallards were banded at Markley's Pond in the URV than at any other trap site. Only 42 northern pintail, the secondary target species, were banded (Table 3). Addition species banded were wood ducks (Aix sponsa), redheads (Aythya americana), blue-winged teal (Anas discors) and cinnamon teal (Anas cyanoptera) (Table 3). Band Reporting and Record Keeping All new banded birds and recaptures were submitted to the U. S. Fish and Wildlife Service's Bird Banding Laboratory on standard forms. Computor files containing the number of birds banded by area, site, day, age and sex were constructed at the Colorado Division of Wildlife's Research Center. Band Recovery Locations and Rates First year band recovery locations following the first 2 years of banding have been generally predominately in western Colorado, with most other recoveries restricted to adjacent Wyoming, Utah and New Mexico. Most 2nd year recoveries occurred in the same area with limited movement, particularly by birds banded as immature males, into Central Flyway areas. First year recovery rates, which were 2.6% for 1991 bandings, increased to 3.6% for 1992 bandings. As expected recovery rates were highest for immature males and lowest for adult females. �9 Table 1. Trapping locations during preseason banding in western Colorado. August- September. 1992. 'Wetland Area Name Browns' Park Natl 'Wildl. Ref. Location Butch Cassidy TlON, Rl04'W, Sec 12, S'W~ Flynn Marsh TlON, Rl03'W, Sec 16, SE~ Spitzie Slough T10N, Rl03'W, Sec 15, S~ Gardner Park Res. Gardner Park Res. TIN, R86'W, Sec 22, NE~ Colorado R. Valley Latham's Slough T8S, R97'W, Sec 27, S'W~ Morse's Pond TIS, RlE, Sec 34, NE~ Ute Meridian 'Walker 'Wildl Area South Uncompahgre R. Valley Cortez/Mancos TllS, RlOl'W, Sec 14, N'W~ Skippers' Island T1N, R3'W, Sec 14, NE~ Ute Meridian Porter's Feedlot TslN, RlO'W, Sec 27, S'W~ Markley's Pond TsON, R9'W, Sec 30, N'W~ Sander's Pond T14S, R94'W, Sec 34, N'W~ Sweitzer Lake TlsS, R9s'W, Sec 28, S'W~ 'Weir's Pond T36N, R16'W, Sec 13, SE~ Totten Res. T36N, Rls'W, Sec 20, N'W~ Mancos 'Wetland T36N, R13'W, Sec 27, S'W~ 'Weber Res. T36N, R13'W, Sec 12, NE~ Summit Lake T37N, R14'W, Sec 33, S'W~ �10 Table 2. Number of Mallards banded by area, site, age and sex and trapping efficiency during pre-season trapping in western Colorado, 1991. Number of locals included in I!arentheses. Site AM Agebex AF 1M B. Cassidy Flynn Marsh Hog Lake Spitzie Slough Sub-total 0 13 2 11 26 1 19 4 13 37 6 4 2 30 5 5 1 12 1 3 ~ .2 50 22 54 0 15 21 0 5 _Q _Q 69 26 75 3 13 _9 100 30 29 2 13 20 1 _!± 65 Area Brown's Park Gardner Park Colo. River Uncomp. River CortezMancos S. Walker N. Walker Skippers Is. Morse's Pd. Latham Slough Sub-total Markley's Grett's Porter's Kintz Pd. Sub-total Weir's Totten Weber Res. Mancos Wetland Sub-total GRAND TOTALS 216 No. trap days No. banded per trap/day IF Total 26 1 17 6 12 36 2 58 22 43 125 1 9 18 22 50 2.0 6.4 1.2 2.0 2.5 17 12 39 16 2.4 13 119 53 68 8 8 5 6 12 39 1.6 14.9 10.6 11.3 7.3 8.7 20 6 9 41 10.2 1.2 5.2 2.3 6.6 8.3 8.2 5.7 2.7 6.6 5.6 0 9 10 .: 3 7 37(2) 40(2) 26 21 33 27 50 24 153(2) 115(2) .sz 340 76 203 7 47 14 271 20 34 45 28(3) 32(2) 32 91(3) 115(2) 80 _l ~ ...l± ~ 37 107 96(5) 305(5) 11 14 14 7 46 335 (7) 1080(7) 192 126 403 53 4 14 .2 _.§. �11 Table 3. Number of northern pintail and other species banded by area, site, age, and sex in western Colorado, I!reseason 1992. AgeLSex Area Site IF Total SI!ecies AM AF 1M Northern pintail Brown's Pk. Flynn Marsh 0 0 1 0 1 Colo. R. Skipper's Is. Latham Slough 1 1 0 1 1 1 0 0 2 3 Uncomp. R. Markley's Pd. Sander's Pd. 5 7 6 1 2 0 1 4 14 12 Totten Res. Weir's Weber Res. 2 3 2 2 0 0 0 1 0 0 0 0 4 4 2 21 10 5 5 42 CortezMancos Sub-total Wood duck Brown's Pk. Flynn Marsh 0 1 0 0 1 Redhead Brown's Pk. Spitzie Marsh 0 0 1 0 1 Blue-winged Teal Brown's Pk. Spitzie Marsh 0 0 1 0 1 Cinnamon Teal Brown's Pk. Flynn Marsh 0 0 1 0 1 LITERATURE CITED Anderson, D. R., and C. J. Henny. 1972. Population ecology of the mallard. I I. A review of previous studies and the distribution and migration from breeding areas. U.S. Fish and Wildl. Servo Resour. Publ. 105. l66pp. Carney, S. M. 1964. Preliminary keys to waterfowl age and sex identification by means of wing plumage. U. S. Dep. Inter., Fish and Wildl. Servo Spec. Sci. Rep. - Wildl. 82. 47pp. Childress, D. (Subcomm. Chmn). 1986. Evaluation of stabilized regulations in the Pacific Flyway 1975-1983. Pacific Flyway Study Comm. Unpubl. rep. 40pp. Colorado Division of Wildlife. 1989. Colorado statewide waterfowl management plan 1989 - 2003. Colo. Div. Wildl., Terrestrial Wild. Sect., Migratory Game Bird Program Unit. 97pp. Frary, L. G. 1954. Waterfowl production on the White River Plateau, M.S. Thesis, Colorado State University. 93pp. Colorado. �12 Martin, E. M., P. H. Geissler, and A. N. Novara. 1990. Preliminary estimates of waterfowl harvest and hunter activity in the United States during the 1989 hunting season. U. S. Fish Wildl. Serv., Admin. Rep. -- July. 34pp. Munro, R. E., and C. F. Kimball. 1982. Population ecology of the mallard. VII. Distribution and derivation of the harvest. U. S. Fish and Wildl. Servo Resour. Publ. 147. l27pp. Pospahala, R. S., D. R. Anderson, and C. J. Henny. 1974. Population ecology of the mallard. II. Breeding habitat conditions, size of the breeding populations, and production indices. U. S. Fish and Wildl. Servo Resour. Publ. 115. 73pp. Ringelman, J. K., M. A. Wotawa, and R. S. Langley. 1990. Waterfowl abundance and production on the Routt National Forest, Colorado, 1990. Unpubl. rep. Colo. Div. Wildl., Fort Collins. Reynolds, R. E., R. J. Blohm, F. A. Johnson, and J. B. Bortner. 1990. 1990 status of waterfowl and fall flight forecast. U. S. Fish and Wildl. Servo July. 43pp. Rutherford, W. H. and C. R. Hayes. 1976. Stratification as a means for improving waterfowl surveys. Wildl. Soc. Bull. 4:74-78. Smith, R. I., and B. H. Powell. 1989. Distribution of band recoveries from huntershot adult female mallards in western North America. U. S. Fish and Wildl. Serv., Laurel, MD. Unpubl. rep. 20pp. Szymczak, M.R., and J. F. Corey. 1976. Construction and use of the Salt Plains duck trap in Colorado. Colo. Div. Wildl., Div. Rep. 6. l3pp. Trost, R. E. 1985. A preliminary assessment of the recent distribution and derivation of the mallard harvest in the United States based on recoveries from breeding ground bandings, 1975-1984. U. S. Fish and Wildl. Serv., Laurel, MD. Unpubl. rep. 66pp. U. S. Dept. Interior. 1988. Issuance of annual regulations permitting the sport hunting of migratory birds. U. S. Fish and Wildl. Servo final supplemental environmental impact statement. Wash. D. C. 340pp. Unpubl. Rep. 40pp. U. S. Fish and Wild. Servo 1989. The North American duck banding program - a revised approach. Canadian Wildl. Servo and U. S. Fish and Wildl. Serv., Laurel MD. Unpubl. rep. 23pp. Weller, M. W. 1976. Molts and plumages of waterfowl. Pages 34-38 in F. C. Bellrose. Ducks, geese and swans of North America. Stackpole Books, Harrisburg, Pa. 543pp. Prepared by: Michael R. Szymczak Wildlife Reasercher C �13 Colorado Division of Wildlife Wildlife Research Report October 1993 JOB PROGRESS REPORT State of __~C~o~l~o~r~a~d~o~ Project Work Plan W-166-R-2 ___1___ Job _ Migratory Game Bird Investigations 23 Job Title: Ecology of waterfowl in montane wetland communities Period Covered: Author: 1 April 1992 through 31 March 1993 James K. Ringelman Personnel: J. Ringe1man, M. Szymczak, Colorado Division of Wildlife; L. Fredrickson, R. Sanders, University of Missouri, Columbia. ABSTRACT Breeding pairs were surveyed on 3 wetland units during 27-29 May 1992 to provide supplemental data for sampling stratification (used versus unused wetlands). Breeding chronology was at least 3 weeks ahead of previous years, so observed numbers of pairs provided minimum estimates. Twenty-five duck pairs and 1 Canada goose were observed. Literature reviews on montane wetlands and waterfowl were conducted, and an initial sampling design was developed based on previous waterfowl use information. A detailed study plan was written, and preparations were made to initiate field work in May, 1993. ��15 ECOLOGY OF WATERFOWL IN MONTANE WETLAND COMMUNITIES P. N. OBJECTIVES 1. Obtain baseline information on wetland hydrology in relation to precipitation, inflow, outflow, and wetland origin (glacier or beaver). 2. Determine the effects of hydrology, water chemistry, conditions on wetland plant communities. 3. Relate the seasonal composition and abundance of aquatic and fish to information from objectives 1 and 2. 4. Relate waterfowl migrant, breeding pair, brood, and post-breeding use to the findings from objectives 1, 2, and 3. and edaphic invertebrates bird SEGMENT OBJECTIVES 1. Review literature on waterfowl biology and wetlands ecology in montane habitats. Draft a detailed study plan within the context of objectives and approaches contained in the program narrative. Determine sample sizes and sampling protocol, and develop testable statistical null hypotheses. 2. Based on data from pilot research in the study area (J. K. Ringelman, Colorado Division of Wildlife, unpublished data), a listing of high waterfowl use wetlands and low waterfowl use wetlands will be developed. High and low use ponds will be paired on the basis of size and origin. 3. Within both high and low use categories, sample wetlands with permanent water and those with non-permanent (seasonal) water will be selected in proportion to their abundance in the study area. 4. Within the permanent water classification, an equal number of kettle ponds and beaver ponds will be selected as sample units. In the nonpermanent classification, only representative kettle ponds will be selected for study (beaver ponds are usually seasonally permanent wetlands). 5. Three representative plant associations (floating-leaved, submersed aquatic, and emergent herbaceous) will be selected for sub-sampling within the permanent wetland classes identified in 4. Only the emergent association will be sampled in the non-permanent class. 6. Plant associations identified in 5 will be sampled for aquatic invertebrates in 1993 shortly after ice-out and in mid-summer (permanent wetlands only). 7. Waterfowl pairs and broods will be surveyed in a systematic quiet observation techniques beginning in mid-May 1993. 8. Install weather sample wetlands manner using stations and permanent water level markers in or near to quantify hydrology and water budgets on site. �16 9. Indicator species will be used to designate plant associations for aerial coverage, seed sampling, and invertebrate sampling. F1oatingleaved plant associations will be defined as that type dominated by cow1i1y (Nuphar 1uteum ssp. po1ysepalum). Submersed plant associations will contain pondweeds (Potamogeton natans) and watermi1foi1 (Myriophyllum spp.). The herbaceous emergent association will be dominated by sedges (Carex spp.), mannagrass (G1yceria borealis), and blue-joint grass (Ca1amagrostis canadensis). 10. Benthic aquatic invertebrates will be sampled with core samplers. Standardized sweep net procedures will be used to collect invertebrates in the water column and at the surface. Statistical software will be used to compute power and sample sizes necessary for invertebrate sampling. Invertebrates will be preserved in a 90% ethanol solution prior to picking and sorting. 11. Variable mesh size gill nets will be deployed in sample wetlands to determine the presence or absence of fish. Numbers and species of fish captured will be recorded for each sample wetland. 12. An annual report covering the period 1 July 1992 through 31 March 1993 will be prepared and submitted by 1 August 1993. INTRODUCTION Most waterfowl managers env~s~on typical waterfowl habitat as the undulating or flat terrain characteristic of the prairie pothole region of the central U.S. or the aspen park1ands of Canada. However, several other habitats in North America provide valuable resources for migrating, breeding and post-breeding waterfowl. Among these is the Rocky Mountain region of the western U.S., parts of which contain waterfowl breeding densities that equal or exceed those of prairie breeding habitat (Ringe1man 1992). The U.S. government, through the Forest Service, Bureau of Land Management, and Fish and Wildlife Service, owns much of this montane habitat. As a partner in the North American Waterfowl Management Plan, the United States Forest Service has recently initiated an ambitious, nationwide, waterfowl habitat management program on its lands (Hohmann 1989). Forested montane habitat covers about 56,000 km2 in Colorado and 244,500 km2 in the central Rocky Mountain region (Bailey 1980, United States Forest Service 1988). Although comparable in size to the United States portion of the prairie pothole region (280,000 km2), Rocky Mountain montane waterfowl habitats have attributes that set them apart from their grassland counterparts. First, montane wetlands have been less impacted by human activities than wetlands of the northern Great Plains (CanadafUnited States Steering Committee 1986), although severe localized damage has occurred within narrow floodplain valleys (Windell et a1. 1986). Second, except in these valleys, upland plant communities still possess a compliment of native plants despite some grazing and timber harvest. Third, reliable snowfalls provide a dependable source of spring water, thereby creating waterfowl breeding habitat that is stable compared to prairie breeding grounds (Ringe1man 1992). Such stability may be particularly important during times of widespread drought on the prairies, such has occurred in recent years (Anonymous 1991). �17 As in other habitats, waterfowl use montane aquatic plant communities for nesting, food and shelter, and aquatic invertebrates for food during breeding and post-breeding periods (Fredrickson and Dugger 1992). However, the geology and topography of montane regions creates a higher diversity of wetlands types than found in the prairies. Rocks weather slowly and annual primary production decreases with elevation, therefore wetland succession proceeds much more slowly in montane wetlands than in low elevation ponds. Elevational gradients interacting with precipitation patterns and growing season determine soil types and wetland hydrology, which in turn affect nutrient availability, aquatic plant flora, and aquatic invertebrate communities (Ringelman 1992). Fish may compete for invertebrates consumed by adult and young ducks (Eriksson 1983, Beattie and Nudds 1989), and therefore must also be considered in evaluations of wetland habitats. Two forested, montane wetland types -- beaver ponds and glacial ponds -receive the greatest use by waterfowl (Ringelman 1992). Beaver ponds most commonly occur in mid-elevation, montane valleys where slope is <15%. Because beaver ponds are often clustered in "flowages" along suitable lengths of streams and rivers, they provide a valuable wetland community well-suited to the needs of breeding waterfowl (Ringelman 1991). Wetlands created by beaver possess relatively stable water levels maintained by runoff and direct precipitation. In addition, beaver ponds act as nutrient sinks, trapping sediments and organic matter that would otherwise be carried downstream. This enhances wetland fertility and the aquatic invertebrate communities exploited by waterfowl. However, beaver flowages are dynamic wetland systems because they are abandoned after beaver deplete their food resources in 10-30 years. Glacial ponds include wetlands formed behind lateral and terminal moraines and kettle ponds created by the same glacial process that made the prairie potholes -- large chucks of ice embedded in glacial outwash melt after a glacier retreats, forming depressions that later fill with water (Flint 1971). Often, kettle wetlands have a water budget that is dependent solely on spring runoff and summer precipitation, and therefore experience receding water levels during the summer with a concomitant increase in the density and abundance of herbaceous, emergent vegetation. Despite dynamic water levels, natural succession is slow; peat accumulations indicate that some kettle ponds have persisted as wetlands for >7,000 years (Windell et al. 1986). Few studies have been conducted on waterfowl in forested montane habitats. Previous surveys conducted in the San Juan mountains (Rutherford and Hayes 1976) and White River Plateau (Frary 1954) of Colorado and the Uinta mountains of Utah (Peterson and Low 1977) primarily focused on seasonal waterfowl populations and distribution, although some quantitative data on the relationships between the wetland community and breeding waterfowl were described in the latter study. In beaver ponds and kettle wetlands, mallards (Anas platyrhynchos) and green-winged teal (Anas crecca) are usually the most common nesting species, but ring-necked ducks (Aythya collaris), Barrow's goldeneye (Bucephala islandica), buffleheads (Bucephala albeola) and cinnamon teal (Anas cyanoptera) are also common. The peak of nest initiation for early-nesting ducks (mallards and green-winged teal) varies from early May to early June, depending on snow conditions and wetland availability. Latenesting species such as ring-necked ducks begin nesting nearly a month later than early-nesters. Pilot studies (J. K. Ringelman, Colorado Division of Wildlife, unpublished data) during 1989 and 1990 in the northern Park Range of Colorado (Fig. 1) provided information on waterfowl abundance, species composition, and wetland communities. Wetland complexes in this 30 km2 study area contained �40' G . of 8 wetland Fig. 1. Loc~t10n Routt National study units 1n the Forest, Colorado. , I,.. -, ,:'1 _ y "7,-<\ - 0.., ..-:, _ ~. __ __ ~ __ H �19 519 beaver and kettle ponds and an estimated 149 resident waterfowl pairs (Table 1). Waterfowl pair density, however, varied markedly among units (Table 1), probably as a function of physical and hydrologic features as well as floristic and invertebrate communities. This preliminary study, along with information from floristic surveys conducted by The Nature Conservancy (Neeley 1990), provides valuable baseline information useful for enhancing the design of more detailed investigations on the interrelationships among wetland hydrology and wetland origin, aquatic plant flora, and aquatic invertebrate communities. Knowledge of these ecological relationships is fundamental to understanding patterns in waterfowl distribution and abundance. Only by understanding natural processes in unaltered wetland communities can one develop a basis for identifying deficiencies in altered wetland systems and prescribing appropriate management actions. METHODS Breeding AND RESULTS Pair Surveys Breeding pairs were surveyed on 3 wetland units during 27-29 May 1992 to provide supplemental data for sampling stratification (used versus unused wetlands). Breeding chronology was at least 3 weeks ahead of previous years, so observed numbers of pairs are minimum estimates for each unit. Five mallard pairs were observed on the North Fork of the North Platte unit; 4 mallards, 1 ring-necked duck, 1 cinnamon teal, and 1 Canada goose were observed on the Goose Creek unit; 6 mallards, 2 ring-necked ducks, 2 greenwinged teal, and 4 bufflehead were enumerated on the North Kettle Ponds unit. The 1992 total of 25 ducks and 1 goose are far short of the 3-year mean of 79 duck pairs. Two mallard nests and 1 mallard brood were also observed during the survey. Table 1. Number of wetlands and mean number of breeding pairs/year in 8 drainages on the Routt National Forest, 1989-91 (J. K. Ringelman, Colorado Division of Wildlife, unpublished data). Drainage or unit Lower Beaver Creek Upper Beaver Creek North Kettle Ponds North Fork of the North Platte Forester Creek Goose Creek Shafer Creek South Kettle Ponds Totals No. wetlands No. pairs 55 40 135 160 22 41 22 44 23 17 29 21 7 29 10 13 519 149 Pairs/wetland 0.42 0.42 0.21 0.13 0.32 0.71 0.45 0.29 �20 Study Proposal Segment objectives 1-5 were accomplished during the first-year reporting period 1 July 1992 to 31 March 1993. M.S. graduate student Robert Sanders prepared a detailed study proposal (Appendix A) in cooperation with Colorado Division of Wildlife personnel and Dr. Leigh Fredrickson of the Gaylord Memorial Laboratory, University of Missouri. Results of the initial field season, which will commence in early May, will be presented in the next annual report. LITERATURE CITED Anonymous. 1991. 1991 status of waterfowl and fall flight forecast. U.S. Fish and Wildl. Serv., Wash. D.C. and Can. Wildl. Serv., Ottawa. 39 pp. Bailey, R. G. States. 1980. Description of the ecoregions of the United U.S. Forest Serv., Miscell. Publ. 1391. 77pp. Beattie, L. A., and T. D. Nudds. 1989. Differential habitat occupancy by goldeneye ducklings and fish: predator avoidance or competition? Can. J. Zool. 67:475-482. CanadafUnited States Steering Committee. 1986. North American waterfowl management plan, May 1986. Can. Wildl. Serv., Ottawa, and U.S. Fish and Wildl. Serv., Wash. D.C. 19pp. Eriksson, M. O. G. 1983. The role of fish in the selection of lakes by nonpiscivorous ducks: mallard, teal and goldeneye. Wildfowl 34:27-32. Flint, R. F. 1971. Glacial and quaternary geology. Inc. New Yory, NY. J. Wiley and Sons, Frary, L. G. 1954. Waterfowl production on the White River Plateau, Colorado. M.S. Thesis, Colorado State University, Fort Collins. 93pp. Fredrickson, L. H., and B. D. Dugger. 1992. Management of montane wetlands in Arizona. U.S. Forest Servo Rep. (in press). Hohmann, K. 1989. Innovation and teamwork in the North American waterfowl management plan. U.S. Fish and Wildl. Serv., Fish and Wildl. Leaf. 13.2.2. 5pp. Neeley, B. E. 1990. A biological reconnaissance of Big Creek Lakes and Arapaho Lakes areas. The Nature Conservancy, Boulder, CO. 22pp. Peterson, S. R., and J. B. Low. 1977. Waterfowl use of Uinta Mountain wetlands in Utah. J. Wildl. Manage. 41:112-117. Ringelman, J. K. 1991. Managing beaver to benefit waterfowl. U.S. Fish and Wildl. Serv., Fish and Wi1dl. Leaf. 13.4.7. 7pp. �21 Ringelman, J. K. 1992. Ecology of montane wetlands. u.s. Wildl. Serv., Fish and Wildl. Leaf. 13.3.6. 6pp. Fish and Rutherford, W. H., and C. R. Hayes. 1976. Stratification as a means for improving waterfowl surveys. Wildl. Soc. Bull. 4:74-78. United States Forest Service. 1988. Land areas of the National Forest system. U.S. Forest Servo Publ. FS-383. 87pp. Windell, J. T., B. E. Willard, D. J. Cooper, S. Q. Foster, C. F. KnudHansen, L. P. Rink, and G. N. Kiladis. 1986. An ecological characterization of Rocky Mountain montane and subalpine wetlands. U.S. Fish and Wildl. Servo Biol. Rep. 86(11). 298pp. Prepared by: c!= y. ~ James K. Ringelman Wildlife Researcher C ��23 APPENDIX STUDY PLAN: ECOLOGY OF WATERFOWL A IN MONTANE WETLAND COMMUNITIES Goal: To describe physical and chemical characteristics of montane wetlands relate those characteristics to vegetative composition, aquatic macro invertebrate community composition and waterfowl use. and Objectives: 1) Describe montane wetland habitats based on physical, hydrological and vegetative characteristics. chemical, 2) Describe aquatic macro invertebrate communities as they relate various wetland habitat types as determined in Objective 1. 3) Determine waterfowl and 2. to habitat use of migrating, breeding, and brood rearing in relation to habitat data collected in Objectives 1 JUSTIFICATION Traditional sites associated with waterfowl breeding include the prairie pothole and parkland regions of the north-central United States and southern Canada. Because this vast area produces over one-half of North America's ducks, (Bellrose 1976) it has been a primary focus of waterfowl researchers and managers over the last half century. However, dramatic increases in agricultural intensity combined with long-term drought have reduced the potential for this region to continue to contribute to the continental waterfowl population or to provide sites for investigations relating to life history strategies. Because of this loss of breeding potential in prairie regions, interest has increased in non-prairie nesting habitat. One such area of interest is the forested montane region of the central Rocky Mountains. This area encompasses approximately 244,500 km2 (Bailey 1980) and is comparable in size to the Prairie Pothole Region (280,000 km2) (Belrose 1976). Although this area has a wide dispersion of wetlands when compared to the Prairie Pothole region, high densities of breeding ducks are common in these montane wetlands (Peterson and Low 1977, Ringelman et al. 1989, 1990). These wetlands are of interest on a regional scale because they provide habitat for migratory birds and recruitment contributes to regional duck populations as well as providing opportunities to understand waterfowl response to habitats and life history strategies. Several attributes of montane wetlands make them valuable in understanding waterfowl habitat relationships. Human impacts are minimal in high elevation wetland systems compared to the prairies. Prairie habitats continue to experience the effects of habitat degradation through wetland draining, chemical pollution and intensification of agriculture (Tiner 1984). In contrast, agriculture has little influence on montane wetland systems. Modifications within wetland basins are limited and usually less severe than in the prairies. Impacts in many montane wetlands are generally minimal, highly localized, and primarily limited to timber harvest, mining, recreational, and livestock grazing activities. Water supply in montane wetlands is more dependable than in the prairie region. Snowfall accounts for as much as 80% of the annual precipitation at higher elevations in the Rocky Mountains (Windell et al. 1986). Although snow accumulations vary among years, spring runoff from snow melt provides a reliable source of water, and periods of prolonged drought are less common than in the northern Great Plains. �24 Mammalian predators such as raccoon (Procyon lotor) and red fox (Vulpes vulpes) are abundant in prairie habitats and adversely impact female duck survival and nest success (Sargeant et ale 1984). Predator populations in montane habitats appear to be more balanced, with less severe impacts on waterfowl populations than in prairie habitats. Despite their apparent value as breeding and migration habitat, montane areas in the central Rocky Mountains have been the focus of few waterfowl studies. One of the earliest montane waterfowl studies focused on the White River Plateau area of central Colorado (Frary 1954). Later, waterfowl production in the mountains west of the San Luis Valley of southern Colorado was identified (Rutherford and Hayes 1976), as was waterfowl use of high elevation wetlands in the Uinta Mountains of Utah (Peterson and Low 1977). These studies have brought attention to the potential value of high elevation montane wetlands to breeding waterfowl. Recent studies in the Big Creek Lakes area have corroborated these findings (Ringelman et ale 1989, 1990). Breeding pair densities in this area were high (6.1 pairs/km2), indicating that favorable spring habitat conditions exist for several species of ducks. Nest success was also high (50-70%). Duckling survival, however, was relatively low (4.1 young/brood) and is believed to be the factor most limiting to duck recruitment in this area. This low survival rate may be attributable to shortages of food and/or cover. During periods of low water, emergent cover in beaver (Castor canadensis) ponds and in kettle ponds results in a reduced availability of escape and thermal cover. This reduction in the structural features of the habitat likely reduces invertebrate production as well. The relationship between limnology, hydrology, plant composition, invertebrate community composition and distribution, and waterfowl use in montane wetlands is complex and poorly understood. Past efforts have identified these areas as having value to waterfowl but have not attempted to quantify the habitat variables that affect waterfowl use. Similarly, research efforts to quantify variables in prairie habitats are often hampered by prolonged drought and human related impacts. Under these conditions the interrelationships of physical, chemical, floral, and faunal components of wetland systems may be obscured. The high density of relatively unimpacted wetlands in the Big Creek Lakes area, combined with a high degree of differential waterfowl use presents a unique opportunity to quantify and compare a variety of wetland variables and their impact on the reproductive cycle of breeding waterfowl. I propose to study wetland characteristics and waterfowl habitat use in north-central Colorado to examine several hypotheses regarding habitat variables that influence duck use of high elevation montane wetlands. The following objectives will be addressed in this study: 1) 2) 3) EXPECTED Characterize montane wetlands based on physical, chemical, hydrological and vegetative characteristics. Describe aquatic macro invertebrate communities as they relate to habitat variables determined in Objective 1. Determine habitat use of migrating, breeding and brood rearing waterfowl in relation to habitat data collected in Objectives 1 and 2. BENEFITS Knowledge of high elevation montane wetland systems and associated waterfowl populations in the central Rocky Mountains is incomplete. Information from this study will provide insight into the composition of these communities and the interrelationship between habitat variables and waterfowl use. This information will allow managers to make more informed decisions regarding the regional values of these wetlands and will provide a basis for effective wetland protection and mitigation of wetland degradation and losses. In addition, information on waterfowl use in these relatively pristine, unimpacted systems can be used to provide a comparison to waterfowl life history strategies in highly altered systems. This study will also provide baseline data for observing long-term effects of acid rain, environmental pollutants, and other perturbations on high elevation montane wetlands. �25 STUDY AREA Located in the Routt National Forest approximately 33 km northwest of Walden, Colorado, the Big Creek Lakes area of north-central Colorado contains one of the highest wetland densities in the state. The 20.7 km2 study area encompasses the upper reaches of the North Fork of the North Platte River, Forester, Goose and Shafer creeks (Fig.1), and contains approximately 160 beaver and glacial (kettle) ponds. Annual precipitation ranges from 64-76 cm (25-30 in.). Elevations range from 2573-2774 m (8520-9100 ft.). Upland areas are dominated by lodgepole pine (Pinus contorta) with Engelmann spruce (Picea engelmannii) and subalpine fir (Abies lasiocarpa) more prevalent at higher elevations and on cooler north-facing slopes. Some of the more common plants in wetland areas include willow (Salix spp.), yellow pond lily (Nuphar luteum ssp. polysepalum), pondweed (Potamogeton gramineus and P. filiformis), mannagrass (Glyceria borealis and G. elata) and sedges (Carex utriculata, C. vescaria and C. aquatilis). For a more detailed description of the area's vegetation see Neely (1990). Eight species of ducks frequent the area, of these mallard (Anas plat yrhynchos) , green-winged teal (A. crecca carolinensis), cinnamon teal (A. cyanoptera), American wigeon (A. americana), ring-necked duck (Aythya collaris), and bufflehead (Bucephala albeola) are common nesters. Nonbreeding species include wood duck (Aix sponsa) and common merganser (Mergus merganser) • This area was selected for several reasons. First, a large number of kettle and beaver ponds of various sizes are present in a relatively small area. This facilitates field operations while allowing comparison of waterfowl use among wetlands without the problems associated with bird use when wetlands are distributed over a wide area. Second, information on waterfowl use (Ringelman et al. 1989) and vegetational composition (Neely 1990) are available. Finally, this portion of the Routt National Forest is under consideration by the U.S. Forest Service for designation as a Research Natural Area (RNA), which will protect this area from activities that directly or indirectly modify ecological processes (USFS 1990). This protection will minimize perturbations from activities such as timber harvests, intensive grazing and other disturbances during the duration of this project as well as providing opportunities to examine long term variations in montane wetlands and duck populations. OBJECTIVE 1) 1 Characterize hydrological montane wetlands based on physical, and vegetative characteristics. chemical, Questions: Do significant differences in physical, chemical and vegetational composition exist among high and low waterfowl use wetlands? Does wetland origin (glacial vs. beaver) significantly influence physical and chemical composition? To what extent do these differences (if any) influence vegetative composition and density? What effect does hydrology have on the composition of montane wetland plant communities? What effect does pond age have on the composition and productivity of beaver ponds? Background Physical characteristics of a wetland influence water chemistry and hydrology which in turn influence vegetative composition (Windell et al. 1986). To effectively isolate variables having the greatest effect on wetland productivity, each variable must be quantified and both the individual and cumulative effects assessed. Limnological and hydrological data for wetlands in the Big Creek Lakes area are lacking. Nutrient availability plays an important role in regulating primary productivity in wetlands (Kadlec 1979). Phosphorous appears to be the nutrient in least supply relative to demand in most cold climate freshwater wetlands (Windell et al. 1986) and may be limiting in this area. Wetland plant species have specific tolerance ranges for pH (Jeglum 1971), soil temperature (Chapin 1981), and hydrologic regime (Fredrickson and Taylor 1982). Differences in one or more of these variables may influence �26 vegetative composition, which in turn influence food and cover availability and waterfowl use. Baseline data on physical and chemical composition is a prerequisite to understanding plant, invertebrate, and waterfowl distributions within and among wetlands. Although the entire waterfowl community of the Big Creek Lakes area will be observed and evaluated, the focus of this study will be on mal.lards and ring-necked ducks. These species were chosen for several reasons. First, due to time and manpower limitations, intensive sampling techniques such as time budget and food habit analysis must focus on a few representative species of which mallards and ringneck ducks represent both dabbling duck (Anatini) and diving duck (Aythyini) respectively, and a wide range of foraging strategies, dietary requirements, and habitat needs can be observed and assessed. Second, mallards and ring-necked ducks are the most abundant species on the area (40.6% and 23.7% of the breeding pairs respectively) (Ringelman et al. 1989) and ample opportunities exist to observe and document habitat use and activities of these species. Physical Ho: and Chemical Characteristics No significant differences exist in the physical or chemical composition of wetlands in the Big Creek Lakes area. Wetlands receiving high waterfowl use have one or more physical or chemical characteristics that differ significantly from low waterfowl use wetlands. Wetlands of glacial origin have one or more physical or chemical characteristics that differ significantly from wetlands of beaver origin. Vegetation Ho: No significant differences exist in vegetational composition of wetlands in the Big Creek Lakes area. Wetlands receiving high waterfowl use have significantly different vegetational composition than wetlands receiving low waterfowl use. Wetlands of glacial origin have significantly different vegetational composition than wetlands of beaver origin. Methods Ponds were divided into high and low use categories based on waterfowl pair and brood counts (Ringelman, unpublished data). To equally weight wetlands of various sizes, the number of waterfowl pairs observed on each wetland was divided by the area of the wetland basin covered by water to obtain a ratio of pairs observed: hectare of surface water. Ponds were then placed into high use ( > 7 pairs/ha) and low use ( < 4 pairs/ha) categories and further subdivided by pond origin. From these categories 12 high use and 12 low use wetlands were randomly selected for intensive sampling (see Appendix A for a list of ponds to be sampled). Physical, chemical, vegetational, and invertebrate sampling will be conducted during three periods during each field season. Due to annual fluctuations in snow depth and timing of ice-out, exact dates cannot be assigned to nest initiation and subsequent life history events. A combination of field observations and data from previous studies (Ringelman et al. 1989) will be used to divide sampling into three periods corresponding with major life history events in the waterfowl breeding cycle. Period 1 (Prenesting) will correspond with the stabilization of mallard pairs numbers on the area (departure of migrants) and before lone male numbers (indicated pairs, female on nest) peak. Samples from this period will provide information on habitat conditions and invertebrates when female requirements for protein are high (Krapu 1979). Period 2 (Early Mallard Brood Rearing) will correspond with the 10 day period following the peak of mallard brood hatch. Data from field observations of pair numbers and Class I mallard broods will be used to determine sampling dates. Data from this sampling period will provide information regarding habitat conditions and invertebrate food resources during early brood development. �27 Period 3 (Late Mallard/Early Ring-necked Duck Brood Rearing) sampling will be conducted within 10 days after the peak of ring-necked duck brood hatch. This will provide information on habitat conditions and invertebrate food availability during early brood rearing for ring-necked ducks and for mallard ducklings during mid to late brood rearing (Class II and III). Physical and Chemical Characteristics Physical and chemical data will be collected for the 24 sample wetlands beginning in May 1993. Sampling times will be between 0900-1600 hours to minimize disturbance to waterfowl during periods of peak activity. Parameters to be recorded on site for each wetland include: air and water temperature, average water depth, pH, water color, and conductivity. Temperature will be recorded with a pocket thermometer. Water depth will be measured using a graduated rod at all invertebrate sampling plots (see Objective 2). Water color will be estimated using a Hach color test kit (Model CO-1). Conductivity and pH will be recorded using a water testing meter (model and Hanna Instruments pHep+). Shoreline length will be estimated using aerial photographs. Data for each wetland will be recorded on data sheets (Fig. 2). Water samples will be collected from each wetland and submitted to Colorado Division of Wildlife (CDOW) laboratories for analysis according to Division protocol. Total nitrogen (TN), total phosphate (TP), cation concentrations (Mg, Ca, Na, and K), and algal chlorophyll (AC) levels will be determined by CDOW personnel. Pond age will be estimated for all beaver ponds sampled using historic aerial photographs. If complete aerial photograph records are unavailable, tree growth ring counts (Lawrence 1952) will be used to estimate pond age. Hydrology Two weather stations and four ground water monitoring stations will be installed in 1993 near selected study wetlands and will be monitored on a weekly basis. Weather stations will include max-min thermometers, rain gauges, hygrothermographs and evaporation pans. Evaporation pans will be placed on covered platforms in or near the wetland basin, filled with a known depth of water, and monitored to determine evaporation rates. Water level markers will be placed in all wetlands and will be monitored bi-weekly to determine water fluctuations within the wetland basins. Plant transpiration rates will be estimated using standard tables (Boyd 1987). Due to the nature of isolated precipitation events in the area, weather station rain gauges will be supplemented with the placement of an additional 10 gauges installed at various locations on the study area. Ground water levels in two wetland basins of glacial origin will be monitored using wells as described by Cooper (1990). Well casings will be constructed of polyvinylchloride (PVC) pipe with an inside diameter of 5.1 cm. Vertical slots 10 cm long and 0.1 cm wide will be cut into each casing to allow water movement. Wells will be dug using a 3 inch diameter bucket auger. Wells will be installed in homogenous stands of vegetation, through the entire peat column down to the underlying alluvium, glacial outwash or till. Vegetation Maps delineating vegetational zones and dominant plant type(s) have been obtained for all wetlands on the study area from the CDOW and will be field checked for accuracy during the first field season. Low level aerial photographs of study wetlands will be taken in 1993 using a modification of methods described by Cowardin and Meyers (1974). These photographs will be digitized and entered into a Geographical Information System (GIS) format which will allow calculation of shoreline length and percent composition of wetland habitat types. Seed production estimates will be made using methods described by Laubhan (1992). Two high use and two low use wetlands will be selected for sampling in 1993. Sampling will be conducted at six sites in homogenous stands of emergent vegetation in each wetland. A 25 x 25 cm frame will be placed at each point and the species present, average height for each species, number of seed heads, dimensions of seed heads (length and width) within the �28 frame will be recorded. All seed heads placed in paper bags and labeled. Materials 1622111144210101241181100- Needed - Objective in the plot will then be clipped, 1. Laptop computer and Database software Thermometer, min.-max. Thermometers, pocket type, °c Rulers, folding, 2 m. pH meter, Hanna Instruments model pHep+ Conductivity meter, model Hach Kit, color test model CO-l Cubitanors, 1 1. Weather stations w/rain gauges, min-max thermometers and evap. pans pvc pipe, 2 in.ID, 10 ft., w/vertical slots PVC caps, 2 in. bucket auger, 3 in. diameter Water level markers, 1/2 in. PVC Aerial photograph, 36x36 of study area Camera, 3Smm w/50mm lens Film, Kodachrome ASA 64, 36 expo Seed sampling frame, 2SX25 cm. Bags, paper, small Objective 2) 2 Describe aquatic macro invertebrate various wetland habitats. populations as they relate Questions: What factors (physical, chemical, vegetative) affect composition and density? What is the relationship between plant and invertebrate community composition? to invertebrate associations Background The abundance of aquatic macro invertebrates and composition of invertebrate communities are influenced by several habitat variables. Pond age, size, and water chemistry (Friday 1987) and aquatic macrophyte composition (Krull 1970, Voights 1976, Chilton 1990) all playa major role in aquatic invertebrate abundance and distribution which in turn influences waterfowl use of a particular wetland or habitat type (Joyner 1980, Murkin et al. 1982, Murkin and Kadlec 1986). Invertebrates are an excellent source of protein (Driver et al. 1974), an essential component of the diets of prelaying and laying hens (Krapu 1979, Swanson et al. 1979) and ducklings during periods of rapid development (Chura 1961, Street 1978, Swanson 1984). A detailed invertebrate study is very time and labor intensive and is far beyond the resources of this project. Rather the objective of this study is to gather baseline data on the relative abundance and distribution of invertebrates within and among wetlands and wetland habitat types and to compare that data with physical, chemical and vegetative composition of montane wetlands and to waterfowl habitat use. Ho: Hal: Ha2: Ha3: No significant differences exist in aquatic macroinvertebrate community composition or relative abundance among wetlands or within wetland habitat types in the Big Creek Lakes area. Wetlands receiving high waterfowl use have a significantly greater relative abundance of aquatic macro invertebrates than wetlands receiving low waterfowl use. Wetlands of glacial origin have a significantly different aquatic macro invertebrate composition and relative abundance than wetlands of beaver origin. Significant differences exist in the aquatic macro invertebrate composition and relative abundance among wetland vegetation types. �29 Methods Invertebrate sampling will be conducted during each of three periods corresponding with waterfowl life cycle chronology as determined by field observations (see Objective 1). Plot locations will be determined using stratified random sampling techniques as described by Elliott (1971). Each wetland will be divided into strata according to vegetative community type. The following six vegetation types will be sampled; Carex spp., Potamogeton spp., Myriophyllum spp., Nuphar spp., Glyceria spp., and open (non-vegetated) water. Vegetation types will be located using wetland maps and field observations and plots within habitat types will be selected by walking a random number of paces into the habitat type as determined by a random number table. Samples will be collected in shallow water zones «0.5 m) and in deep water zones (>0.5 m) where vegetation growth is within 0.5 m of the water surface. Sweep sample data (Ringelman unpublished data) were analyzed to determine required sample sizes. Using the formula from Elliott (1971): where n is the required sample size, S2 is the sample variance, D is the index of precision (power), and x is the sample mean, the required number of samples (D = 0.80) was calculated to be 25 per vegetation type in each waterfowl use category during each sampling period for a total of 900 sweep samples per field season. Because no preliminary sampling data is available for benthic sampling, an initial sample size of 25 cores per vegetation type in each waterfowl use category during each sampling period will be used for a total of 900 cores per field season. Data from both water column and benthic samples will be analyzed prior to the second field season to determine if an adequate number of samples have been taken and sample size will be adjusted accordingly in 1994. To further maximize sampling efforts, waterfowl habitat use data from the 1993 field season will be analyzed to determine which habitats are being utilized by feeding waterfowl and sampling efforts will focus on these types during 1994. Water Column Plots will be sampled using an aquatic dip net and a standard 1 m. sweep. A metal stake with a 1 m. cord fastened to the net will be used to standardize the length of the sweep. Upon reaching the sampling location, the stake will be pushed into the bottom substrate, the sweep net lowered into the water such that the top 25 cm. of the water column is included in the sample, and a 1 m. sweep taken. Samples will be cleaned using a 1.0 mm self-cleaning screen (Euliss and Swanson 1989) with all coarse, non-animal material being removed, the sample placed into a plastic bag, and labeled. Immediately following each sweep, a vegetation sample will be taken along the sweep transect line. A garden shears will be used to first remove all vegetation above the water line, followed by three standardized cuts at a depth of 25 cm. below the surface. All vegetation collected by these cuts will be blotted dry and weighed immediately and recorded on data sheets (Fig.). Benthos Benthic sampling will be conducted concurrent with water column sampling. A 5.1 cm diameter core sampler will be used to sample benthic macro invertebrates according to the methods described by Swanson (1984). The sampler will be inserted into the substrate adjacent to the sweep sample site to a depth of 5 cm, the sample brought to the surface, cleaned with a 1.0 mm self-cleaning screen (Euliss and Swanson 1989), contents placed in a plastic bag, and labeled. �30 Sorting and Analysis Samples will be taken to field camp where each will be sorted to taxonomic order and recorded on data sheets (Fig. 4). Samples will then be preserved in 80% ethanol until they can be returned to the laboratory, dried at 100°C for 1 hour and weighed. Sampling data will be analyzed to determine if significant differences in invertebrate composition and/or relative abundance exist among habitat types within and among wetlands. Vegetation sample weights will be entered as a covariate to reduce variance due to differences in plant substrate available to invertebrates. Materials 222100010030112222526211- Needed - Objective 2 Sweep nets, aquatic invertebrate Core samplers, aquatic, 5.1 cm Screens, 1.0 rom. Bags, plastic, ziplock, 16oz. Bags, plastic, ziplock, 32 oz. Vials, glass, 25 ml. Graduated cylinder, 250 ml. Balance, electronic Dissecting scopes, 0.7-3x, w/light Pans, white enamel, 10x14 in. Dissecting kits Ethanol, 5 gal. Marking tape, 150 ft. rolls Clipboards Markers, waterproof Waders, chest, neoprene, size 10 and 9 Float tube Shears, 10 in. Objective 3) 3 Determine habitat waterfowl. use of migrating, breeding and brood rearing Questions: Does waterfowl use show a significant relationship to the parameters measured under objectives 1 and 2? Which parameters have the greatest effect on use? Does use vary by species? time period? Background Waterfowl use of wetlands is related to various physical characteristics (Gilmer et ale 1975, Godin and Joyner 1981, Rotella and Ratti 1992), water chemistry (Patterson 1976, Murphy et ale 1984, Swanson et ale 1988), vegetative composition (Patterson 1976, Swanson 1988), and invertebrate abundance (Joyner 1980, Murkin et ale 1982, Murkin and Kadlec 1986). These parameters along with factors such as behavioral spacing (Patterson 1976) often form complex interrelationships with no single variable responsible for influencing duck use of a particular wetland. Under the first two objectives of this study several habitat types in the wetlands of the Big Creek Lakes area will be quantified. In order to determine which of these types are used by waterfowl to fulfill their life history requirements, waterfowl use of these wetlands must also be quantified. Wetland use by waterfowl varies by species, sex, age, and stage in the annual cycle. Based on high pair use, species diversity, general lack of repairing, and high proportion of pairs observed with broods, such factors as pond numbers, upland nesting cover availability, nest success, and early season invertebrate numbers appear not to be limiting duck production in the Big Creek Lakes area (Ringelman et ale 1989). Mean brood size at fledging was lower than expected (4.1 young/brood) suggesting that duckling survival is the factor most limiting duck recruitment on the area. The greatest decline in duckling survival occurred during the first 2 to 3 weeks following hatch, then stabilized at about 4.0 ducklings per brood. The high attrition rate of young ducklings during this period corresponds with the stage of development when �31 ducklings are highly dependent on high protein invertebrate food resources. Duckling mortality is often highest during the first 12 days of life and may be attributable to a lack of invertebrates on oligotrophic ponds (Street 1977). A similar study by Street (1978) demonstrated that a diet deficient in animal protein resulted in reduced growth rates and high duckling mortality, even when plant foods were provided ad lib. Relatively low primary and secondary productivity in high elevation wetlands (Windell 1986) points towards invertebrate distribution and abundance during brood rearing as a major factor in influencing waterfowl use of wetlands in this area. Ho: Hal: Ha2: Ha3: Ha4: Has: All waterfowl select wetlands and wetland habitats at random and demonstrate no patterns of use within or among wetlands. Waterfowl selection of wetlands is positively correlated with one or physical or chemical wetland variables. Waterfowl selection of wetlands is positively correlated with wetland vegetative composition. Waterfowl selection of wetlands is positively correlated with the relative abundance of aquatic macroinvertebrates within the wetland regardless of invertebrate community composition. Waterfowl selection of wetlands is positively correlated with invertebrate community composition of a wetland regardless of relative abundance of invertebrates. Waterfowl use of wetlands and wetland habitat type is species specific and varies among duck species. Methods Pair Counts Pair observations will be conducted on the 24 sample wetlands beginning in May 1993. Each wetland will be observed a minimum of two times per week. Pair counts will be conducted between 0800 and 1800 hours to avoid confusion between non-nesting paired hens and hens accompanied by males on early morning or late afternoon incubation breaks. Wetlands in adjacent drainages will be surveyed by a single observer, one drainage in the morning and the other in the afternoon. The sequence will be reversed on subsequent visits to avoid surveying the same wetlands at the same time of day. Observations will be conducted so as not to conflict with limnological, vegetative and invertebrate sampling. Wetlands will be approached using vegetation and terrain to conceal the observer and avoid flushing ducks to adjacent wetlands and possibly duplicating counts. Sufficient time should be spent at several vantage points around each wetland to ensure that all waterfowl on that wetland are accounted for. If no waterfowl are visible, the observer will carefully approach the wetland and attempt to flush any concealed birds. Species, sex, location, habitat type, time, breeding status (paired vs. unpaired), and activity, will be recorded for each observation. In case of inclement weather, observations will be postponed at the discretion of the primary researcher. Regular nest searches will not be conducted. In the event of nest being located, the observer will record the species (if known), number of eggs, stage of incubation (Weller 1956), distance to nearest water and vegetation within five meters of the nest. Nest location will be noted on an aerial photo or topographic map. A piece of marking tape will be secured to vegetation five meters north of the nest to facilitate relocation. Nests will be revisited shortly after hatching to determine fate. Broods Brood observations will begin on or about 1 June depending on hatching chronology as determined by broods observed during pair counts. Methods will be similar to those described for pair observations except that survey times will be limited to periods of greatest brood activity (before 0800 hours and after 1800 hours). In addition to data collected for pair observations, brood �32 size and age (Gollup and Marshall 5) for all duck broods observed. 1954) will be recorded on data sheets (Fig. Time Budgets One water body of each use category (high and low) within each origin class (beaver and glacial) will be randomly selected for weekly time activity observations on both mallard and ring-necked ducks. A time budget computer program on field computers (Tandy model) will be used to record diurnal activity using continuous sampling methods (Tacha et ale 1985). Observations will be made during 15 minute sampling sessions during three time periods (0600-0800 hours, 1000-1500 and 1800-2000 hours). Activities will be divided into nine categories: 1) surface feeding, 2) sub-surface feeding (tiping up), 3) diving 4) locomotion (swimming, walking, and flying not associated with other activities), 5) resting/sleeping, 6) comfort movements, 7) alert, 8) social interactions, and 9) out of sight. Bird locations will be recorded using a rangefinder (model 160) and compass and recorded on wetland maps. Analysis Materials 22222228TIME of Waterfowl Needed Data - Objective 3 Spotting scopes, 20x, w/stock mount Binoculars, lOx Field Computers, Tandy Model w/time budget software Rangefinders, model 160 Compasses, Silva Ranger Egg candling devices (2 in. ID automotive radiator hose) Microcassette recorders Microcassette tapes SCHEDULE Project Duration: October 1992 - May 1995 October 1992 - February 1993: Conduct a detailed literature review, write initial draft of research proposal and conduct first committee meeting. April 1993: Submit final draft of research proposal and conduct committee meeting. Prepare for first field season. May - August 1993: First field season. September 1993 - April 1994: Conduct analysis of first season data. Identify and correct problems encountered. Submit progress report. Prepare for second field season. April - September 1994: Second field season. October 1994 - May 1995: Analyze collected data. Prepare and submit draft of thesis. Final draft of thesis will be submitted and final committee meeting (thesis defense) will be conducted. MONTHLY SCHEDULE 1) 2) 3) 4) Total Conduct pair/brood counts 2x/week Collect water samples lx/month Monitor weather stations and wells lx/week Collect invertebrate samples lx/month 24 wetlands x 30 samples/wetland 5) Wash and sort invertebrate samples 6) Time activity budget 12 hr/week 7) Misc. duties, weather delays, etc. Hours/month 492 hr/mo - 4 weeks/mo = 123 hr/week Primary Researcher Field Technician (CDOW) CDOW Researcher TOTAL 70 40 13 123 hr/week hr/week hr/week hr/week Hours/month 64 24 16 120 180 48 ~ 492 (approx.) �33 BUDGET Personnel 1) Graduate Research Assistant 2) Field Assistant (2 4-mo field seasons) Total Salaries FY 1993 $ $ 5832.00 FY 1994 FY 1995 5832.00 Travel Round trip from Missouri to Colorado 120.00 120.00 Supplies Materials MATERIALS 1 1 2 2 1 1 2 2 2 2 4 500 500 50 10 1 1 2 4 5 2 2 6 2 10 from CDOW or UM-C NEEDED - LITERATURE Afton, not available Measuring tape, fiberglass, 50 m. Rangefinder, model 400 Hach kits, model Hach kits, model Hach kit, model CO-1, Color Test Kit Dissecting scope Thermometers, Sweep nets, D-frame Core Samplers, 5 cm Sieves, #30 Stream depth gauges Sample jars, Nalgene, 125 ml Plastic bags, zip-lock, 1 qt. Vials, glass 25ml Ethanol, gal. Dissecting tray, white enamel Float tube Waders, neoprene, size 10 Rain gauges, Marking tape, 150 ft. rolls Spotting scopes, 20X w/stock mount Clipboards w/ waterproof cover Permanent markers Microcassette recorders Tapes, microcassette $ 39.95 117.75 22.50 31.90 31.90 6.75 9.00 80.00 CITED A. D. 1979. Time budget of breeding northern shovelers. Wilson Bull. 91(1):42-49. Altmann, J. 1974. Observational study of behavior: sampling methods. Behaviour 49:227-267. Bailey, R. G. 1980. Description of Ecoregions of the United States. USDA Forest Service. Intermtn. Reg. Ogden, Utah. 77pp. Bellrose, F. C. 1976. Ducks, geese, and swans of North America. Stackpole Books, Harrisburg, PA. 544 pp. Boyd, C. E. 1987. Evapotranspiration/evaporation (E/Eo) ratios for aquatic plants. J. Aquat. Plant Manage. 25:1-3. Chapin, F. S. 1981. Field measurement of growth and phosphate absorption in Carex aquatilus along a latitudinal gradient. Arct. Alp. Res. 13:83-94. Chilton, E. W. 1990. Macroinvertebrate communities associated with three aquatic macrophytes (Ceratophyllum demersum, Myriophyllum spicatum, and Vallisneria americana) in Lake Onalaska, Wisconsin. J. Freshwat. Ecol. 5(4):455-466. �34 N. J. 1961. Food availability and preferences of juvenile mallards. Trans. N. Am. Wildl. and Nat. Resour. Conf. 26:121-34. Cooper, D. J. 1990. Ecology of wetlands in Big Meadows, Rocky Mountain National Park, Colorado. u.s. Fish Wildl. Serv., Biol. Rep. 90 (15). 45pp. Cowardin, L. M. and V. I. Meyers. 1974. Remote sensing for identification and classification of wetland vegetation. J. Wildl. Manage. 38(2):308314. Driver, E. A., L. G. Sugden, and R. J. Kovach. 1974. Calorific, chemical and physical values of potential duck foods. Freshwat. Biol. 4:281-292. Elliott, J. M. 1971. Some methods for the statistical analysis of samples of benthic invertebrates. Freshwater Biol. Assoc. Sci. Publ. 25. 144pp. Euliss, N. H. and G. A. Swanson. 1989. Improved self-cleaning screen for processing benthic samples. Calif. Fish and Game 75(2):124-128. Frary, L. G. 1954. Waterfowl production on the White River Plateau, Colorado. M.S. Thesis, Colorado State University, Ft. Collins. 93 pp. Fredrickson, L. H., and T. S. Taylor. 1982. Management of seasonally flooded impoundments for wildlife. u.S. Fish Wildl. Serv., Resour. Publ. 148. 29pp. Friday, L. E. 1987. The diversity of macro invertebrate and macrophyte communities in ponds. Freshwater Biol. 18:87-104. Gilmer, D. S., I. J. Ball, L. M. Coward in , J. H. Riechmann, and J. R. Tester. 1975. Habitat use and home range of mallards breeding in Minnesota. J. Wildl. Manage. 39(4):781-789. Godin, P. R. and D. E. Joyner. 1981. Pond ecology and its influence on mallard use in Ontario, Canada. Wildfowl 32:28-34. Gollup, J. B. and W. H. Marshall. 1954. A guide for aging duck broods in the field. Mississippi Flyway Tech. Council, Sect. Rep. Mimeo. Jeglum, J. K. 1971. Plant indicators of pH and water level in peatlands in Candle Lake, Saskatchewan. Can. J. Bot. 49:1661-1676. Joyner, D. E. 1980. Influence of invertebrates on pond selection by ducks in Ontario. J. Wildl. Manage. 44(3):700-705. Krapu, G. L. 1979. Nutrition of female dabbling ducks during reproduction. Pages 59 - 70 in T. A. Bookhout, ed. Waterfowl and wetlands - an integrated review. Proc. 1977 Symp., Madison, WI, N. Cent. Sect., The Wildlife Society. Krull, J. N. 1970. Aquatic plant-macro invertebrate associations and waterfowl. J. Wildl. Manage. 31(4):707-718. Laubhan, M. 1992. A technique for estimating seed production of common moist-soil plants. u.S. Fish Wildl. Leafl. 13.4.5. 8pp. Murkin, H. R. and J. A. Kadlec. 1986. Relationships between waterfowl and macro invertebrate densities in a northern prairie marsh. J. Wildl. Manage. 50(2):212-217. Murkin, H. R., R. M. Kaminski, and R. D. Titman. 1982. Responses by dabbling ducks and aquatic invertebrates to an experimentally manipulated cattail marsh. Can. J. Zool. 60:2324-2332. Murphy, S. M., B. Kessel, and L. J. Vining. 1984. Waterfowl populations and limnologic characteristics of taiga ponds. J. Wildl. Manage. 48(4):1156-1163. Neely, B. E. 1990. A biological reconnaissance of Big Creek Lakes and Arapaho Lakes areas. Unpublished report to The Nature Conservancy, Boulder, CO. Patterson, J. H. 1976. The role of heterogeneity in the regulation of duck populations. J. Wildl. Manage. 40:22-32. Peterson, S. R. and J. B. Low. 1977. Waterfowl use of Uinta Mountain wetlands in Utah. J. Wildl. Manage. 41:112-117. Ringelman, J. K., M. A. Willms, and R. S. Langley. 1989. Waterfowl abundance, production, and habitat use on the Routt National Forest, Colorado. Unpublished field report, Colorado Division of Wildlife, Fort Collins, CO. Ringelman, J. K., M. A. Wotawa, and R. S. Langley. 1990. Waterfowl abundance and production on the Routt National Forest, Colorado, 1990. Unpublished field report, Colorado Division of Wildlife, Fort Collins, CO. Chura, �35 Rotella, J. J. and J. T. Ratti. 1992. Mallard brood survival and wetland habitat conditions in southwestern Manitoba. J. Wildl. Manage. 56(3):499-507. Rutherford, W. H., and C. R. Hayes. 1976. Stratification as a means for improving waterfowl surveys. Wildl. Soc. Bull. 4:74-78. Sargeant, A. B., S. H. Allen, and R. T. Eberhardt. 1984. Red fox predation on breeding ducks in midcontinent North America. Wildl. Monogr. 89. 41 pp. Street, M. 1977. The food of mallard ducklings in a wet gravel quarry, and its relation to duckling survival. Wildfowl 28:113-125. Street, M. 1978. The role of insects in the diet of mallard ducklings - an experimental approach. Wildfowl 29:93-100. Swanson, G. A. 1983. Benthic sampling for waterfowl foods in emergent vegetation. J. Wildl. Manage. 47:821-823. Swanson, G. A. 1984. Invertebrates consumed by dabbling ducks on the breeding grounds. J. Minn. Acad. Sci. 50:37-40. Swanson, G. A. 1988. Aquatic habitats of breeding waterfowl. Pages 195-202 in D. D. Hook et al. eds. Ecology and management of wetlands, Vol. 1: Ecology of wetlands. Timber Press, Portland. 592pp. Swanson, G. A., G. L. Krapu, and J. R. Serie. 1979. Foods of laying female dabbling ducks on the breeding grounds. Pages 47-57 in T. A. Bookhout, ed. Waterfowl and wetlands - an integrated review. Proc. 1977 Symp., Madison, WI, N. Cent. Sect., The Wildlife Society. Swanson, G. A., M. I. Meyer, and J. R. Serie. 1974. Feeding ecology of breeding blue-winged teals. J. Wildl. Manage. 38:396-407. Swanson, G. A., T. C. Winter, V. A. Adomaitis, and J. W. LaBaugh. 1988. Chemical characteristics of prairie lakes in south-central North Dakotatheir potential for influencing use by fish and wildlife. U.S. Fish and Wildl. Serv., Fish Wildl. Tech. Rep. 18. 44pp. Tacha, T. C., P. A. Vohs, and G. C. Iverson. 1985. A comparison of interval and continuous sampling methods for behavioral observations. J. Field Ornith. 56(3):258-264. Tiner, R. W., Jr. 1984. Wetlands of the united States: current status and recent trends. National Wetlands Inventory. U.S. Fish and Wildl. Serv., Washington, D.C. 58pp. U.S. Forest Service. 1990. Forest Service manual 4063.3. Voights, D. K. 1976. Aquatic invertebrate abundance in relation to changing marsh vegetation. Am. MidI. Nat. 95(2):313-322. Weller, M. W. 1956. A simple field candler for waterfowl eggs. J. Wildl. Manage. 20:111-113. Windell, J. T., B. E. Willard, D. J. Cooper, S. Q. Foster, C. F. Knud-Hansen, L. P. Rink, and G. N. Kiladis. 1986. An ecological characterization of Rocky Mountain montane and subalpine wetlands. U.S. Fish and Wildl. Servo BioI. Rep. 86(11). 298 pp. �36 Wetlands selected for intensive sampling The following wetlands have been selected for intensive sampling. Ponds have been placed in high and low pair use categories based on 1989 data and have been subdivided by pond origin. PAIRS Low Use High Use brood Kettle Beaver Kettle NF 79 NF 84 S 14 S 18 NF 6 NF 9 NF 13 NF 53 G 18 G 19 G 21 G 50 NF NF NF F For comparison, the same ponds observation data. are placed Beaver 80 83 86 43 NF NF NF NF F 12 19 37 65 50 G 24 G 53 S 28 into categories based on 1989 BROODS Low Use High Use Kettle Beaver Kettle NF 79 NF 83 F 43 S 18 NF NF NF NF F NF 80 NF 86 S 14 6 12 13 37 50 G 18 G 21 G 24 Beaver NF 9 NF 19 NF 53 NF 65 G 19 G 50 G 53 S 28 �37 Colorado Division of Wildlife Wildlife Research Report October 1993 JOB PROGRESS REPORT State of __~C~o~l~o~r~a~d~o~ Project Job Title: 10 : Job _1_ Cooperative Management Programs Period Covered: Author: Migratory Game Birds Investigations W-166-R-2 Work Plan _ 01 April 1992 through 31 March 1993 Michael R. Szymczak Personnel: James K. Ringelman and Michael R. Szymczak, Wildlife Colorado Division of ABSTRACT National Wetland Inventory maps of the San Luis Valley (SLV) were used to delineate the wetland portion of the SLV. Recommendations for wetland habitat improvements and/or management were provided for Russell Lakes, San Luis Lakes, Yampa and Horsethief State Wildlife Area's; the Partners in Wildlife Program; Hebron Sloughs in North Park; the Soil Conservation Service's wetlands program in southwest Colorado; and the Rio Grande National Forest. A draft Colorado Waterfowl Hunter Attitude Survey was written. Presentations on various aspects of waterfowl ecology were given at training and educational schools, workshops, and shor~ courses. Responsibilities as .Colorado's . representative on Pacific Flyway Study Committees and Central Flyway Techniccll Committee were fulfilled. Waterfowl surveys were conduc t ed Ln 2 areas. of North Park and the Flat·Tops Wilderness Area in 1;he white River. NatioriaT ; Forest. A 3-5 year cooperative post breeding goose banding program was initiated in the Co.rtez-Mancos area. Technical assistance was provided as a member of the Continental Evaluation Team for the North American Waterfowl Management Plan and the Adaptive Harvest Management Working Group. Participation in Central Flyway matters include appointment and participation as a Central Flyway representative to the Stabilized Duck Regulations Committee. ��39 Cooperative Migratory Bird Management Programs Michael R. Szymczak James K. Ringelman In 1988, the Colorado Division of Wildlife (CDOW) created the Migratory Game Bird Program Unit (MBPU) within the Terrestrial Wildlife Section. This administrative change combined all individuals having statewide responsibilities for research and management of migratory game birds. Members of the MBPU work in concert to improve migratory bird management in Colorado. This job was created to allow team members to participate in these management programs. P. N. OBJECTIVES 1. Participate in developing and implementing habitat-based waterfowl management plans on a statewide, habitat region, and project basis. 2. Advise state and federal land managers on beneficial habitat acquisitions and/or developments and provide expertise in preparation of development and/or management plans. Advise private land managers in developing habitat management plans and assessing impacts on waterbird populations. 3. Present information on the principles of waterfowl management to workshop attendees, educational classes, and conservation organizations. 4. Participate levels. 5. Cooperate in developing surveys and techniques of migratory bird management programs. in migratory bird management meetings at the state and flyway that will assess the impact SEGMENT OBJECTIVES 1. In conjunction with the Statewide Waterfowl on habitat region plans. 2. Provide biological expertise on waterfowl biology and wetland development programs on the Brush, Yampa, Russell Lakes, and Bonny State Wildlife Areas, Hebron Ponds, Walden Reservoir (BLM-CDOW), various wetland complexes managed by the u.S. Forest Service, and other areas where requested. 3. Conduct a survey of waterfowl hunters that purchased a Colorado waterfowl stamp for the 1990-91 hunting season but did not buy a stamp for the 199192 season, to determine their reasons for not hunting waterfowl in Colorado in 1991-92. 4. Prepare and present requested. lectures on migratory 5. Compile population appropriate Management Plan, continue game bird management status information and represent work when Colorado �40 at Pacific and Central Flyway Technical Committee and Council meetings. Attend migratory game bird program and biologist meetings in Colorado when requested. 6. Provide methodology for wetland habitat and migratory game bird population surveys when requested. 7. Cooperate in Canada goose trapping and banding operations in the DurangoCortez area, Middle Park, and the Gunnison area. 8. Serve on the U.S. Fish and Wildlife Service's Stabilized Regulations Committee and North American Waterfowl Management Plan Evaluation Team. RESULTS Waterfowl Management Plans Recently completed National Wetland Inventory maps of the San Luis Valley (SLV) were obtained from the U. S. Fish and Wildlife Service (USFWS). Using those maps, the wetlands of the SLV were traced onto transparencies, along with identifying physical features. These maps are needed to evaluate current wetland status in the SLV and design a new sampling system for measuring breeding duck. The SLV Waterbird Plan includes both wetland and duck breeding pair objectives. Wetland Developments and Acquisitions Potential sites for wetland developments and existing wetlands were visited on the Russell Lakes and San Luis Lakes State Wildlife Area's (SWA) in the SLV, and the Yampa and Horsethief SWA's in northwest Colorado. Recommendations for wetland construction, water management, and management of existing wetlands were given to CDOW management personnel. Recommendations on future development were given for the Blanca Wildlife Area in the SLV and the Hebron Sloughs in North Park, both Bureau of Land Management (BLM) areas. The relation of upland nesting area and wetland vegetation to water level was evaluated at Eighteen Island Reservoir at Hebron Sloughs which resulted in a decision on maintenance of water levels in the short-term. Wetland projects developed under Partners in Wildlife, a USFWS program that utilizes some CDOW duck stamp funds, were visited in the SLV and the Grand Valley of western Colorado. Existing projects were reviewed and potential for new projects was evaluated. We consulted with U. S. Soil Conservation Service (SCS) biologists on their mitigation program for wetlands lost because of de-salinization projects in the Cortez area, and toured some wetlands recently developed under that program. We recommended that developing a cooperative program between SCS and Partners in Wildlife (USFWS) would be an advantage to both programs and would result in better wetland design for the mitigation projects. Project personnel, as members of the multi-agency Waterfowl Habitat Project Review Committee, reviewed and rated wetland enhancement and acquisition proposals from land management agencies for funding with Colorado State Duck Stamp monies. Existing wetlands and wetland development sites were visited in the Rio Grande National Forest. Advise was given on future developments. �41 Yaterfowl Hunter Attitude Survey In cooperation with the Human Dimensions Research Unit at Colorado State University (CSU) , a draft survey instrument was developed. A time table was established for exposing the general intent of the survey to focus groups, pretesting the survey, developing the final survey, conducting the survey and preparing the final report. The final report should be completed by early fall 1993. Informational Programs A workshop on waterfowl ecology and management for the U. S. Forest Service was sponsored and organized by project personnel. Formal presentations were given: on montane wetlands to the above workshop; on waterfowl ecology to the CSU Wildlife Management Short Course; on cattail management and control to the CDOW Southeast Region; on evaluating factors that limit waterfowl production to a meeting of waterfowl managers at the Gaylord Laboratory, University of Missouri; on integrated wetlands management to the Mississippi Flyway Technical Committee; and on the future of waterfowl management to the CSU Fisheries and Wildlife graduate students and faculty. Presentations on duck identification were given to the CDOW District Wildlife Management Trainees. Yaterfowl Technical Committee and Council Meetings The July 1992 Pacific Flyway Study Committee meetings were chaired by Colorado. Waterfowl population status was reviewed and hunting season recommendations forwarded through the Pacific Flyway Council to the USFYS Regulation Committee. Populations of specific interest to Colorado whose status was reviewed in July were (1) breeding and wintering mallards inhabiting the Pacific Flyway portion of Colorado and (2) the Rocky Mountain Canada goose population. Project personnel participated in the special winter meetings and regular March meetings of the Pacific Flyway Study Committee and Central Flyway Technical Committee. Special winter meetings are scheduled to discuss specific Flyway issues and formulate recommendations for approval at the regular March meetings. In March, flyway committee members exchange general information on migratory game bird populations and formulate regulatory recommendations for both Flyway Councils, for species hunted before Oct. 1. The Pacific Flyway representative was appointed to an ad hoc committee of the Western Management Unit to look into the need for a Four Corners bandtailed pigeon population management plan. A meeting involving representatives of the four corners states and the USFYS was held to exchange population information, outline a possible management plan, and make writing assignments for various portions of the plan. Population Survey Methodology Surveys were conducted in cooperation with the White River National Forest to delineate the breeding range of Barrow's goldeneye (Bucephala islandica) in the Flat Tops Wilderness Area. Methods for evaluating wetland developments at Hebron Sloughs in North Park and on Partners for Wildlife areas in the SLV were discussed and recommended to BLM and USFYS biologist's during on the ground visits to the respective areas. Surveys were continued on 5 study �42 units in the Big Creek Lakes region in the Routt National Forest. These were the last surveys prior to initiation of an in depth study of wetland characteristics and waterfowl use in the area (see Ringelman, J. K. 1993. Ecology of Waterfowl in Montane Communities. Colo. Div. Wildl., Fed. Aid Rep. Oct.). Surveys of nesting and brood rearing Canada geese on Walden Reservoir in North Park during spring 1992 found recruitment to the population to be better than in 1991, but still at atypically low levels in relation to the number of nesting pairs. Goose nests visited, marked, and mapped in early May just prior to hatch, and revisited after all had hatched showed good hatchability. Periodic brood counts indicated better gosling survival than in 1991. A detailed report was submitted to CDOW Northeast Regional and BLM personnel. Cooperative Canada Goose Banding Canada geese were banded in the Mancos-Cortez area in southwest Colorado in cooperation with personnel of the CDOW Southwest Region. A total of 118 goslings and 50 adults were banded at 7 locations in late June 1992. Technical Assistance Ringelman attended 2 meetings each as a member of the Continental Evaluation Team for the North American Waterfowl Management Plan, and as a Central Flyway representative on the Adaptive Harvest Management Working Group. Each assignment requires extensive analysis of information both during and between meetings. As Central Flyway representative, Ringelman must prepare reports and/or presentations to the Central Flyway Technical Committee and Council. Project personnel conducted Hunter Performance Surveys during the 1992 September teal season in an effort to evaluate the effect of allowing teal hunting one-half hour before sunrise on non-target species. DISCUSSION Project personnel provide useful information in planning and evaluating waterfowl management and habitat enhancement programs in Colorado and educating land management agency personnel about the habitat requirements of waterfowl. We expect that with increased emphasis on habitat enhancement in Colorado as outlined in the statewide Waterfowl Management Plan, our services will be more in demand. Conducting and/or formulating surveys and banding efforts and informing management agency personnel about various aspects of waterfowl and wetland ecology provides a valuable service to management agencies, the waterfowl resource and in some cases the hunting public. Continued participation on Flyway and National committees ensures that Colorado will remain informed on migratory bird matters, have input in migratory bird hunting regulations, and have influence on habitat programs affecting migratory game birds. Prepared by: lil.,:. /Z...Lx *,(j...f.. Michael R. Szymczak Wildlife Researcher C �43 Colorado Division of Wildlife Wildlife Research Report October 1993 JOB PROGRESS REPORT State of __~C~o~l~o~r=a=d~o Project Job Title: 22 : Job _2_ Migratory Game Bird Publications Period Covered: Author: Migratory Game Birds Investigations W-l66-R-2 Work Plan _ 01 April 1992 through 31 March 1993 Michael R. Szymczak Personnel: James K. Ringelman and Michael R. Szymczak, Colorado Division of Wildlife ABSTRACT The following list contains those articles that were prepared and/or submitted for publication or published during this segment: Gilbert, D. W., D. R. Anderson, J. K. Ringelman, and M. R. Szymczak. of nesting ducks to habitat management on the Monte Vista Wildlife Refuge. Wildlife Monograph (In revision) Response National Kirby, R. E., J. K. Ringe1man, D. R. Anderson, and R. J. Sojda. 1992. Grazing on National Wildlife Refuges: Do the needs outweigh the problems. Trans. N. Am. Wildl. Nat. Resour. Conf. 57:611-626. Jeske, C. W., D. W. Gilbert, D. R. Anderson, J. K. Ringe1man, and M. R. Szymczak. 1993. Use of a restraining board and wing bands .toimmobilize and mark mallard ducks . J. Field Ornith. 64 :·84,.. 89.· . Jeske, C. W., M. R. Szymczak, D. R. Anderson, J. K. Ringe1man, and J. A. Armstrong. Mortality factors and body condition/survival attributes of wintering mallards in the San Luis Valley, Colorado. J. Wild1. Manage . .(In revision). Ringe1man, J. K., M. R. Szymczak, C. W. Jeske, and K. E. Ragotzkie. 1992. Ulnar lipid as an indicator of depleted fat reserves in mallards. J. Wi1dl. Manage. 56:317-321. Ringelman, J. K., M. W. Miller, and W. F. Andelt. 1993. Effects of ingested tungsten-bismuth-tin shot on mallards. J. Wildl. Manage. 57:725-732. �44 Shenk, T. M., and J. K. Ringe1man. 1992. Habitat use by cross-fostered whooping cranes in Colorado. J. Wildl. Manage. 56:769-776. Prepared by: -ia~~ -m.~I4,,-IT Michael R. Szymczak Wildlife Reasercher C �45 Colorado Division of Wildlife Wildlife Research Report October, 1992 JOB PROGRESS REPORT State of Project: Work Plan ~C~o~l~o~r~a~d~o_ W-164-R-2 1 Laboratory Investigations Job__ l Job Title: MONOCLONAL ANTIBODIES TO DISTINGUISH AMONG WHITE-TAILED DEER. MULE DEER AND ELK Period Covered: 1 July, 1992 30 June, 1993. Personnel: W.J. Adrian, R.P. Ellis, and R.J. Todd. ABSTRACT Research on the production and use of monoclonal antibodies (MAbs) directed against albumin for direct and positive identification of the species of origin of blood, blood stains and meat was initiated last fiscal year. Substantial progress was made in these first two years. We have obtained purified elk. mule deer and white-tailed deer albumins for use as antigen in our MAb studies. We have completed numerous fusions with mule deer and elk albumin. ��47 MONOCLONAL ANTIBODIES TO DISTINGUISH AMONG WHITE-TAILED MULE DEER AND ELK DEER, William J. Adrian 1. Produce species specific monoclonal 2. Determine 3. Produce species specific monoclonal 4. Determine specificity deer and mule deer. 5. Initiate specificity antibodies against elk and mule deer. of the monoclonal for elk and mule deer. antibodies against white-tailed of the monoclonal use of this technology antibodies antibodies between deer. white-tailed for ongoing forensic cases. METHODS AND MATERIALS This work is a cooperative endeavor between the Division and Dr.Robert Ellis of Colorado State University. The technique for producing MAbs was developed by Kohler and Milstein (1975). Schulman el at. (1978) developed modifications which simplified the technique. It is a technique in use in many immunology laboratories throughout the world. The advantages of MAbs over polyconal antibodies (those produced by animals and harvested from serum of that animal) are (1) each individual hybridoma clone produces antibody molecules of a single specificity, (2) many hybridoma clones can be generated, each producing a particular antibody specificity directed against a particular antigenic epitope, and (3) hybridoma clones can be frozen in liquid nitrogen (LN2) and stored for future use, thus once the hybridoma clones are developed, they are available indefinitely. The method of visualization of MAb binding to albumin antigen epitopes is the Western Blot technique. This technique employs electrophoretic separation of the sample protein, followed by transfer of the electrophoresed sample onto a nitrocellulose membrane. The membrane is reacted with a MAb, rinsed, then reacted with an enzyme (usually horseradish peroxide) conjugated antibody. The presence of the bound peroxidase is detected by adding peroxide plus a chromophore. The development of a brown spot indicates MAb bound to its specific antigenic epitope. Failure of development of a brown spot indicates that the MAb did not bind to its specific epitope. Positive controls are run on the same electrophoresis gels as the sample to ensure accuracy. �48 RESULTS AND DISCUSSION Research on the production and use of monoclonal antibodies (Mabs) directed against albumin for direct and positive identification of the species of origin of blood, blood stains and meat was initiated last fiscal year. Sixty-two white-tailed deer serum samples were obtained from the southeastern United States. Dr. Peter Dratch, from the U.S. Fish and Wildlife Forensics Laboratory at Ashland, Oregon assisted us in obtaining the serum samples. The white-tailed deer samples were obtained form the southeastern U.S. area to ensure that we had pure white-tailed deer serum and no serum from mule deer - white-tailed deer hybrids. The serum samples were pooled and albumin was purified from the pool. We are now ready to immunize mice for the production of white-tailed deer monoclonal antibodies. We have produced several hybridoma clones which produce antibody against elk and mule deer albumin. As reported, there are cross-reactions with other albumins. We have done further Antigen-Antibody reactions utilizing a modification of the western blot technique referred to as checkerboard blotting. This technique is less sensitive than ELISA, and the decrease in sensitivity is an asset to our obj ectives. Some of the slight cross -reactions which were evident in the ELISA were not seen in the checkerboard blot. Also, many sample antigenic extracts can be compared with standard albumins in reactions with several MAbs allan the same blot. Thus, this is an advance in our ability to detect and identify deer and elk albumins. In addition to the above, we have obtained purified moose and pronghorn albumin. Thus we currently have purified albumin from elk, white-tailed deer, mule deer, moose, pronghorn, and cattle. We are currently utilizing these albumins as standards in our assays. Prepared by:~~~~~~~z=~~~~' Xl.an Wildlife Researcher C � https://cpw.cvlcollections.org/files/original/8893686234555e740d50e32dcc9060e4.pdf 66d6dea37bb889d0475f19c5f67da300 PDF Text Text 1 JOB PROGRESS State of: Project: Work Plan: Job Title: Period Colorado W-167-R 1 Personnel: Upland : Job Thomas Bird Research _1!__ Evaluation of Habitat in Eastern Colorado Covered: Authors: REPORT 01 January through E. Remington L. L. Bixler, C. M. Giesen, L. K. G. E. Mekelburg, Remington, D. E. Trujillo, M. M. Younkin. Development for Ring-necked 31 December 1993 and Warren Pheasants D. Snyder E. Braun, J. J. Brim, T. J~' Davis, M. A. Etl, K. Haynes, R. W. Hoffman, M. W. Hooker, T. J. Legg, J. L. Mekelberg, J. T. Pelletier, T. E. Ripp, B. J. Rosenbach, W. D. Snyder, M. L. Warmoth, L. L. Whitmore, J. D. Wieland, and D. J. ABSTRACT The Pheasant Habitat Improvement Program (PHIP) was expanded during 1993 and a intensive evaluation of the program was initiated. Contracts were signed with 5 Pheasants Forever Chapters in northeastern and eastcentral Colorado. Through their joint efforts over 500 habitat developments were planted for ring-necked pheasants (Phasianus colchicus) during 1993. These included: 138 plum thickets (96 with a juniper windb~eak), 300 sorghum plots, 42 disturbance tillage plots, 24 switchgrass plots, and 4 tall wheat stubble plots. Within· (or immediately adjacent to) the 9 treatment blocks established as part of our evaluation effort, 157 sorghum plots, 8 disturbance tillage ~lots, and 3 switchgrass plots were established to provide survival cover. A severe hail storm impacted most plots within one block and one block lacked adequate plots of sufficient quality because of lack of time, manpower, and equipment needed to effectively contract, prepare, and plant all sites. Much of the seed used in initial plantings was found to be too old and of low vigor necessitating extensive replanting. A cool summer and early fall freeze retarded maturity which resulted in lodging of much of the sorghum stalks. Spring crowing counts ranged from 7 to 28 crows/station, indicating densities were generally low to fair. Crowing counts did not differ (R = 0.51) between treatment and control blocks, indicating pre-treatment pheasant breeding densities were similar. Hunting pressure and harvest were monitored opening weekend in all treatment and control blocks. Hunting pressure was relatively light. Hunters averaged from a to 1.2 birds/hour (overall average 0.10). Pheasants used, and . _ were harvested from, many of the annual plantings through fall and early winter but no severe winter storms occurred to markedly impact survival. Heni> pheasants (154) were trapped within treatment and control blocks from mid:~ Is [J October through December and radiomarked to allow estimation of survival t m rates. I :r==<tJ ~=-D i~ CJ ~=::::::a 1~=3 0 '9 i~-A '~_m <g 0-- u ��3 EVALUATION OF HABITAT DEVELOPMENT FOR RING-NECKED IN EASTERN COLORADO Thomas E. Remington PHEASANTS & Warren D. Snyder INTRODUCTION Pheasants are pursued by more small game hunters than any other small game species in Colorado (83-88% of small game license buyers). In a recent survey, 74% of pheasant hunters rated their hunting trips in Colorado as poor (45%) or fair (29%), while only 10% rated their trips as very good or excellent. Lack of birds and places to hunt were identified as the most significant reasons why some hunters did not hunt pheasants in Colorado. Small game license sales have declined by about 70,000 (40%) in the last 10 years. It is apparent that if the Division of Wildlife is going to turn this decline around pheasants will be a key species. Presumably, recruitment and retention of hunters will increase if the quality of pheasant hunting is improved, i.e., increases in pheasant numbers and places to hunt. Previous research has indicated that over-winter survival of pheasants is the most critical factor limiting pheasant populations. The Pheasant Habitat Improvement Program was created to establish over-winter survival cover within historically good pheasant range in eastern Colorado. The program was conceptually designed to overcome significant obstacles to developing habitat, mainly a lack of manpower and a burdensome contractual system (costs of administering contracts exceeded costs of developments). Under PHIP, the Division of Wildlife contracts with individual Pheasants Forever chapters in eastern Colorado to contact landowners and develop habitat on private lands following specific guidelines. Each chapter develops contracts with individual landowners and pays them when the habitat.work is completed and verified. Division of Wildlife personnel inspect a subsample of habitat developments and verify completion and compliance with guidelines. 0 P. N. OBJECTIVES To determine if habitat developments offered through the Pheasant Improvement Program increase pheasant survival, breeding density, harvest within selected northeast Colorado study areas. SEGMENT 1~ 2. 3. 4. 5. Habitat and pheasant OBJECTIVES Work with Pheasants Forever Chapters, management personnel, and landowners to develop habitat within treatment sites and elsewhere in the primary pheasant range in northeast Colorado. Evaluate Pheasants Forever and landowner acceptance of program guidelines and implementation and consider modifications as suggested or needed. Establish census routes and conduct pre-treatment crowlng counts within all treatment and control blocks during April-May 1993. Conduct evaluations of the qUality of annual plantings as survival cover. Monitor hunting pressure and pheasant harvest within treatment and control sites. �4 6. 7. 8. Map habitat, land ownership, and other site conditions within all treatment and control blocks. Trap 100 hen pheasants within treatment blocks and 100 within control blocks during fall, radiomark with mortality sensing transmitters, and relocate periodically to estimate survival rates. Prepare an annual progress report. METHODS Cooperative Agreements were signed with Pheasants Forever Chapters in OctoberDecember 1992. A PHIP Habitat Project Agreement form was prepared and included as Exhibit A of the contract (Appendix A) with each chapter. Division habitat guidelines were included as contract specifications (Exhibit B) and are attached (Appendix B). Members of Pheasants Forever chapters and Division personnel contacted farmers and completed PHIP Habitat Project Agreement forms. Individual landowners were paid by the contracting Pheasants Forever Chapters as soon as the habitat work was completed. Reference is made to Remington and Snyder (1993) for a review of methods previously used in this study. Several temporary personnel were hired by the Division of Wildlife beginning in fall 1992, to map cover and land use types, roads, and land ownership within the treatment and control blocks (Fig. 1) that had been selected for evaluation. Many habitat development sites were within Conservation Re~erve Program (CRP) fields requiring that landowners modify their CRP contracts at u.s. Department of Agriculture Soil Conservation Service offices. Division personnel assisted in this effort and in marking and measuring plots. Michael L. Trujillo, District Wildlife Manager in Yuma provided major assistance in this effort. Division of Wildlife personnel began disking sod within CRP fields during March 1993. Because this was the initial treatment year and well over 150 sites needed to be prepared for planting, much of the initial tillage was contracted to local farmers with larger equipment. Extermination of coolseason tame grasses, primarily smooth brome (Bromus enermis), was difficult (partly because of the cool, wet spring) and required repeated discing. Most replicate discing was completed by Division of Wildlife personnel. Granular nitrogen fertilizer was applied (30 lbs. a.i./ac.) to most plots within CRP in treatment blocks in Phillips and Logan counties during early spring, 1993. Sorghum planting was initiated in mid-May and continued through the first week of July. Most plots were planted by mid-June, however, sparse stands due to poor seed, crusting of top soil because of heavy rains, and late preparation of a few plots in Washington and Yuma counties necessitated late and replicate planting. Most landowners, who planted their own plots, completed their efforts during early to mid-June. Census routes were established within all treatment and control blocks during early 1993. Crowing counts were conducted from late April through May using standard census procedures (Snyder 1985). Hunter pressure and success were evaluated in treatment and. control blocks through personal contacts and interviews. Upland Bird Program personnel (6) were each assigned 3 adjacent treatment and/or control blocks. Interviewers criss-crossed each block along improved roads searching for hunters and then moved on to the next block, repeating this pattern from about 8:00 a.m. until �CONTROL AREAS 1. Haxtun NE 2. Pao II HE 3. Kelly 4. 5. 6. 7. 8. 9. St. Petersburg Paoli S Washington Lonestar Platner Yuma 6 ~ 8 ~ 9 18 tm TREATMENT AREAS 10. Fleming I I. Haxtun S 12. MaII ander I 3. Ho 1 yo ke 14. 15. 16. 17. 18. Fig. SE Paull Otis Curve Clarkvllie Y-W Co. LI ne Kuntz 1. Location, .Treatment layout, and name of PHIP treatment and control ~ evaluation Control blocks. U1 �6 sunset. A fixed wing plane with a pilot and observer flew on Saturday (from 8:30.a.m.-12:00 and 1:30 p.m.-4:00) over each area (once) and communicated locations of hunting parties to interviewers on the ground. The observer also kept a log of hunters observed in the field as well as a count of vehicles parked and driving in each block. When a hunting party was contacted, interviewers explained the purpose of the interview. Hunters were then asked a series of questions to ascertain how many were hunters were in their party, how long they had hunted inside and. outside the block, how many birds they had flushed, bagged, and crippled, type of cover they were hunting in, and whether they had harvested any banded birds. If they indicated they would be hunting in that area on Sunday they were given a pheasant hunting diary (containing the same questions) to complete and a map of the block and asked to return them in the stamped, self-addressed envelope provided. If a vehicle(s) was located which appeared to belong to hunting parties that we had not contacted we left a letter explaining our interest and instructions along with a diary on the windshield. . . Hen pheasants were located night roosting in CRP or wheat stubble fields with the aid of vehicle-mounted lights and captured with long handled nets (Labisky 1968). Trapping began on 11 October 1993, and continued through the end of the year. We attempted to trap 12 to 15 hens within each treatment and control block. Landowners and rural residents proximal to trapping operations w~re contacted to obtain permission for access or to inform them of the operation. Two to 3 vehicles, each containing 2 persons, were used in trapping. captured hens were weighed, a proximal primary was collected for age information, and a serially numbered aluminum leg band was attached. Prior to release, each hen was fitted with a battery-,Powered transmitter (Lotek, Inc., New Market, Ontario, Canada) which was attached to the base of the bird's neck with either dacron cord or a nylon cable tie. The transmitter and attachment device weighed approximately 12 grams (1.2-1'.9% of body weight) and had an expected operational life of 20 months. Captured males were weighed, f~tted with a serially numbered aluminum leg band, and released. RESULTS The Pheasant Habitat Improvement Program (PHIP), which was initiated in 1992, was well received by rural farmers and sportsmen in northeastern Colorado. This resulted in a major expansion in 1993, and the creation of a new Pheasants Forever chapter in washirigton County. This Chapter immediately became involved in the program. Habitat funds were received by chapters in January to May 1993. Amounts requested and spent by chapter varied, but over $220,000 of the requested $270,000 was utilized during 1993 (Table 1). Money not spent during 1993 remained in Chapter accounts and will be used during 1994. Types and amounts of habitat developed during 1993 varied among Pheasants Forever Chapters (Table 2). Most funds were used by chapters to establish 138 shrub thickets (most with small windbreaks established on the north and west sides). Polypropoline weed barrier was required on all woody plantings, and over SO miles was laid as a result. Above average precipitation resulted in excellent first-year survival and growth of seedlings. Some loss of plums was noted, although this was attributed to receiving dead bare root stock from the state Forest Service, and was not extensive enough to be of concern. �7 Table 1. Expendi~ures and balance remaining Improvement Program during 1993 by Pheasants and eastcentral Colorado. Chapter Phillips 1992 Balance County Northeast Yuma $15.91 Colorado County Washington Eastcentral County Colorado Totals a Over-expenditure funds. for the Pheasant Forever chapters 1993 Contract Expenditure Habitat in northeastern Unused balance $100,000 $85,352 $14,663.91 - 0 - 40,000 25,800 14,200.00 $61.81 60,000 60,243a -°- - 0 - 50,000 44,077 5,923.00 - 0 - 20,000 5,556 14,444.00 $270,000 $221,028 $49,230.91 $77.72 of $181.19 by the Yuma County chapter was paid from chapter Sorghum plots accounted for 300 of the 517 total plantings (Table 2) and were the primary cover used within the 9 intensive treatment blocks. Remaining plots were either disturbance tillage (annual weeds) or switchgrass plots. Division of Wildlife personnel conducted most site preparation (including tillage to destroy perennial vegetation within CRP plots) and planted most sorghum plots within treatment blocks. They also completed most disturbance tillage to create annual weeds. Number and composition of habitats developed ovaried among treatment blocks (Table 3). Quality of survival cover created by sorghum plantings varied within and among blocks for several reasons. Free sorghum-sudangrass seed, provided by the national Pheasants Forever organization, was used in a majority of the plantings. Most sites planted with this seed had to be replanted because of poor germination, sparse stands, and slow initial growth. Replanting was necessitated by heavy rains in some areas, .and as a result planting was extended beyond mid-June when we would have preferred to have all the sorghum planted. Late and inadequate site preparation, followed by late planting and lack of adequate nitrogen in the soil resulted in poor stands in some plots, especially in blocks in northwestern Yuma and northeaster~ Washington counties. Precipitation was above average through the planting and growing season but was accompanied by one of the coolest summers on record and an early (mid-Sep.) frost in northeastern Colorado. As a consequence, many of the sorghum plots, that had made considerable growth, did not mature sufficiently and form strong, lodgeresistent stalks: High winds in early November caused extensive lodging of sorghum primarily in central and eastern Phillips County. However, lodging continued in most blocks as fall and early winter progressed. Thus, most of the sorghum plots were rated poor to marginal as survival cover for pheasants �8 Table 2. Allocation and cost of 517 pheasant habitat plantings impacting 1,827.6 acres planted and/or contracted by Pheasants Forever chapters during 1993 in northeastern and eastcentral Colorado through the Pheasant Habitat Improvement Program. Habitat/Chapter Plantings Sorghum Plantings Northeast Colorado Phillips County Yuma County Washington County Subtotal (Logan Cty) 52 113 131 ___i 300 Switchgrass Plantings Northeast Colorado Phillips County Yuma County Subtotal 6 11 __ 7 24 Disturbance Tillage (Annual Forbs) Northeast Colorado Phillips County Yuma County Washington County Subtotal 7 19 14 _.1. 42 Tall Wheat Stubble Retention Phillips County Washington County Subtotal No-till Wheat Stubble Yuma County a Designates b Contracted Payment 246.6 465 532 17.5 1,261.1 $9,472 17,385 22,102 700 $49,659 19 70 132.1 221.1 $754 2,800 1,676 $5,130 25 83.5 105 _8_ 221.5 $1,200 3,615 5,250 300 $10,365 5 10 $300 _1 __b2 _§Q 4 12.5 $360 2 40 $400 (4)a 5 (40) (17) (29) 28 12 18 _3 $11,992 59,661 27,?25 41,760 5,556 $146,494 (Nesting Season) Plum Thickets and (Windbreaks) Northeast Colorado. Phillips County Yuma County Washington County Eastcentral Colorado (Kit Carson Subtotal Custom Tillage b Northeast Colorado Phillips County Yuma County Washington County Subtotal Number Acres 20 51 30 31 Cty) __§ .L§.l 138 65 39 25 _.ll 150 66 (96) 236 124 105.5 34 499.5 number of 3-row windbreaks placed windward labor to prepare sites for planting. $2,382 1,591 3,290 1.257 $8,520 of some thickets. �9 Table 3. Sorghum, disturbance tillage, and switchgrass plots established within and immediately adjacent to the 9 treatment blocks, northeastern Colorado, 1993. Block Cover type Plots Holyoke SE Mailander Kurtzer Fleming Pauli Quality Number Acres Good ratinga Fair Poor Sorghum 19 75 8 9 2 Sorghum Switchgrass Subtotal 24 2 26 71 12 83 21 2 21 2 1 2 3 Sorghum Disturbance 10 till. ~4 Subtotal 29 ~1~5~ 3 3 ~3~ 4 _=1 14 44 3 6 5 28 -=1 127 ~2~ 15 11 2 --:1 Subtotal 29 129 15 11 Sorghum Disturbance 12 till. -=3 5 ~2~ sorghum switchgrass 17 Subtotal 54.5 ~2~.~S~ 20 57 12 7 3 _=1 1 8 Clarkville Sorghum 22 100 14 YW Treatment Sorghum 15 47 15 Kuntz Sorghum Disturbance 37 ~2~ 10 2 ~1~ 1 _ 39 10 3 1 13 till. ~1~ Subtotal Otis Curve Sorghum Total Sorghum Total Disturbance Total Switchgrass Grand Total a Subjective rating tillage of survival 14 9 33.5 5 2 2 157 8 3 168 574 19.5 14 607.5 103 34 6 103 40 20 2 3 25 cover quality in early fall 1993. �10 by mid- to late winter 1993-94 (Table 4). A severe hail storm occurred in early August impacting about two-thirds of the Holyoke Southeast Treatment Block. 'Twelve of the sorghum plots were severely damaged, and although their was some recovery, they provided poor cover by early fall and did not stand over winter. Most of the irrigated corn in the block was also destroyed and provided marginal brood and fall habitat. The storm may have caused pheasant mortality but this remains unknown. Pheasant Crowing Census Counts of crowing males did not differ (Table 5; ~ = 0.98 and 0.51 using average of replicate counts and high count of replicate counts, respectively) between treatment and control blocks. Counts in the 15-20 crows/station range are fairly low when compared to long term crowing count data collected in northeastern Colorado. While this suggests pre-treatment pheasant densities were low to moderate, the trend in crowing count censuses conducted in this area in recent years is slowly increasing. Table 4. Survival quality rating of sorghum plots within the 9 treatment blocks, northeastern Colorado, January - February 1994. Sorghum height Block Tall Short Percent standing Weed assoc. Overall rating 2.0 1.2 14.7 1.6 1.6 Mailander 2.8 1.5 24.6 1.7 1.7 Kurtzer 2.4 1.4 32.8 1.6 1.4 Pauli 3.2 1.8 30.0 1.9 1.8 Fleming 2.2 1.1 27.9 2.4 1.3 Clarkville 2.8 1.3 15.5 1.5 1.2 Y-W 3.3 1.5 32.0 2.0 1.7 Kuntz 3.8 1.8 40.5 2.1 2.0 otis Curve 2.5 1.0 18.6 1.9 1.3 Average 2.8 1.4 26.3 1.9 1.6 Holyoke SE �Table 5. Pheasant crowing census data among treatment northeastern Colorado, spring 1993. and control blocks, Date Block 2 1 Average of 2-3 countsa 3 Highest count High count/ stationb Treatments Holyoke SE 13.8 16.5 15.1 16.5 18.2 Mailander 23.0 14.4 18.7 23.0 23.1 Kurtzer 16.1 16.0 16.0 16.1 18.1 5.1 6.1 12.6 15.1 15.0 12.4 17.4 17.4 16.2 19.0 22.8 25.5 27.7 22.0 27.7 27.7 6.2 8.0 8.0 C1arkville 6.1 Pauli 4.2 11.2/11.5/15.1 Y-W Co. Line Fleming Kuntz 17.4 7.4 22.8/19.0 9.0 16.4 otis Curve 4.5 8.0 ,.-14.1 Average ± 5.7 17.0 7.1 . ± 6.S 17.S Controls Paoli NE Haxtun NE 12.7 Paoli South St. Pete Kelly 16.5 16.5 16.5 16.S 14.7 16.S lS.9 13.6 13.3 13.6 15.1 11.9 13.7 15.6 17.0 20.4 21.9 17.2 17.2 17.2 lS.2 17.3 19.0 22.1 6.6 12.1 12.1 12.2 6.6 6.6 6.7 3.8/16.5 13.0 9.7{12.3 13.0/20.4 Yuma Co. 17.2 Lonestar 19.0 Platner 12.1 Wash-West Average 6.6 13.7/7.0 11.0 6.9/14.6 17.6 r.e 14.1 ± 3.5 15.1 ± 4.1 a The lowest count was excluded when wide variance among counts occurred. b Obtained using the highest count per station among counts before averaging. 16.2 �12 Hunter pressure and success Hunting pressure in treatment and control blocks was generally light (Table 6). ,While this probably reflects declining trends in small game license sales and in pheasant hunting in Colorado (Braun et al. 1994), heavy rains on the day before the opener (which made roads difficult to travel on) followed by a windy and cold snowstorm on Sunday probably exacerbated low participation. Sixty-five hunters were observed Saturday from the airplane. These hunters were found in an area of approximately 173 mi2, for an apparent hunter density of 0.38 hunters/mi2 or 2.7 mi2 per hunter. Eighty hunters were contacted on the ground, either directly or through diaries, who hunted in control blocks, while 120 were contacted who hunted in treatment areas. No statistical analyses were attempted because hours hunted within blocks were low, and flush and harvest rates were too highly variable to detect treatment and control block differences at these sample sizes. Birds/hunter/hour is the only reasonable standard of comparison because typically hunters hunted over a wide area, spending relatively little time within blocks. Hunters were contacted in 6 of 9 control blocks and 8 of 9 treatment blocks, and averaged 0.04 ± 0.04 and 0.22 ± 0.38 birds/hunter/hour hunting within, these blocks, respectively. While success in treatment areas appears, on average, to have been much better, most of this difference was attributed to the 1.2 birds/hunter/hour in the Mailander block where 5 hunters shot 3 pheasants in half an hour. Without this extraordinarily successful estimate, treatment block harvest rates declined to 0.08 ± 0.09. Combining information from hunts nearby blocks with hunts conducted within blocks, birds/hunter/hour averaged 0.032 ± 0.03 in control blocks, and 0.18 ± 0.37 in treatment blocks. The combined treatment block estimate is inflated because of the highly successful half hour hunt in the Mailander tract. Hunting success was obviously quite poor, since the average hunter would have to hunt 10 to 25 hours to bag a pheasant. Hunters flushed an average of 1.4 roosters each, but bagged only 0.3 roosters each. Pheasant Trapping and Survival Pheasant trapping began on 11 October and was concluded on 29 December. Trapping was not continuous throughout this period as heavy rains followed by wet snow made fields inaccessible and forced suspension of trapping efforts from 11 to 29 November. Transmitters were attached tC3 154 hen phaaaarrts , In addition, 67 males were banded although no transmitters were at.t.ached , We determined the age of 131 hens for which completeiy grown feathers were available. Of these, 80 (61%) were j~veniles and 51 were adults, a ratio of 1.6 juveniles per adult hen which is indicative of reasonably good reproductive success. The dacron cord mounted on the Lotek transmitters proved to be extremely difficult to attach to the birds, and unreliable as several became untied or were pulled over the birds head. We switched to sewing nylon cable ties on the transmitters with dental floss which was then coated with super glue for strengthening. This proved to be a quick and ,convenient attachment harness. Unfortunately, we discovered that the nylon deteriorated and became brittle after the super glue had completely cured (~ 5 days), thus the transmitters were at risk of falling off. We recaptured 21 bi~?s and replaced their susceptible transmitters with transmitters attached with cable ties coated �Table 6. Pheasant hunters 14 November 1993. Block Hunters Hours/ Hunter contacted, Flush/ Hunter Inside hunter effort, Bag/ Birds/Hour/ Hunter Hunter and hunter Sample Size success Hours/ Hunter Block in Treatment Flush/ Hunter Outside and Control Bag/ Birds/Hour/ Hunter Hunter Block blocks, 13- Birds/Hour/ Hunter Combined Treatments Holyoke SE Mailander 12 5 0.16 1.3 0.58 0.29 0 0.0 0.0 0.0 0.0 0.29 0.1 1.2 0.6 1.2 0 0.0 0.0 0.0 0.0 1.2 Kurtzer Clarkville 12 0.17 0.33 0.17 0.08 12 2 1.25 0.42 0.02 0.01 Pauli, 15 0.73 3.0 0.0 0.0 12 1~58 1.92 0.17 0.008 0.002 0.0 0.0 0.0 0.0 7 0.9 0.0 0.0 0.0 0.0 0.47 1.14 0.21 1.01 5 0.2 0.0 0.0 0.0 0.009 1.3 1.17 0.5 0.06 7 2.21 0.85 0.14 0.009 0.01 18 0.27 2.05 0.27 0.05 15 0.36 0.8 0.26 0.050 0.02 8 0.69 0.88 0.25 0.05 4 0.93 5 1.5 0.4 0.07 Y-W Co. 0 Fleming 42 6 Kuntz Otis Curve Controls Paoli NE Haxtun NE 7 0.5 1. 71 0.43 0.12 4 0.25 0.0 0.0 0.0 0.05 Paoli South 4 0.5 0.75 0.0 0.0 2 1 2.5 0.0 0.0 0.0 st. Pete 10 1.7 1.9 0.03 0.02 2 1 1 0.5 0.25 0.02 Kelly 15 0.7 2.8 0.93 0.08 7 0.43 2.71 1. 43 0.48 0.08 0.0 0.0 0.0 0.5 0.5 0.17 0.0 0.03 32 1.03 0.88 0.19 0.006 6 0.58 1.83 1 0.28 0.008· Platner 4 0.75 0.25 0.0 0.0 4 2.25 1.25 0.0 0.0 0.0 Wash-West 0 0.0 0.0 0.0 0.0 8 0.5 0.25 0.125 0.0 0.03 Yuma Co. Lonestar 0 ' 0.0 12 f-' LV �1:+ with an alternate lost because glue, glue which did not make them brittle. of mounting leaving difficulties with the dacron a total of 138 functioning C. E., K. M. Giesen, 1994. Upland Wildl., Labisky, Division R. F. chickens, Notes Remington, 62. R. W. Hoffman, Bird Management 1968. Rep. 19. Guide, and W. D. Snyder. 1994-1998. Colorado Div. its use in capturing and cottontails. Illinois pheasants, Nat. Hist. prairie Surv., BioI. Urbana .. 12pp. for ring-necked pheasants 1993. in eastern Prog. Rep., Fed. Aid Proj. W-167-R. Colorado. T. E. Remington, 48pp. T. E., and W. D. Snyder. Snyder, W. D. or the cable ties/super CITED Analysis Nightlighting: bobwhites, radios were transmitters. LITERATURE Braun, Sixteen 1985. Survival J. Wildl. / Colorado. Apr. of radio-marked Manage. /-:. Evaluation of habitat Colorado development Div. Wildl., 1-24. hen ring-necked pheasants in 49:1044-1050. '/ .:) : ; L:': .; -.r-;!"·~- Prepared by Warren D. Snyder Wildlife Researcher C �-v r" r C.I'~l) :" 1""'\ Pheasant 15 ~ PHEASANTS ~_~ HA_B_I_TA_T __PR_O_J_EC_T __AG_R_E_EM_E_N __ T AGREEMENT , ADDRESS: Rte. _ lANDOWNER,-=Town _ ----:::-:---:--:7"" PHONE.---'-~--State Zip ·COUNTY CHAPTER NAME, ADDRESS _ HABITAT DEVELOPMENTS Item Number Size/ac Cost Item Combined Cost I Sorghums Shrub thicket (~O.2 ac.) Shrub thicket + windbreak. I I ___! +- ..!.I ,;-.---- Switchgrass I Roadsides Tall annual weeds Tall Wheat Stubble ~18n No-~ill Wheat Stubble Custom Operations/Other Show Location of Habitat Developments on Section Maps ,..R._ 8._ T._ •••••••• ! ·· ··· . : R __ Years of Contract •••••••• : •••••••• s__ R._ s._ I, I I ·. · .............•... T._ R._ s._ · The landowner agrees to repay this Pheasants Forever Chapter for the establishment cost for contracted items listed above that are destroyed, or that are included within a commercial hunting or fee-for-hunting access area, prior to the term of projects set forth in this agreement. I have-read and understand the above statement and the specifications (Exhibit 8) and fully agree to the terms of this agreement. for development Landowner's Signature Date: Pheasants Forever Authorized Signature Date: RECORD OF PAYMENT: Check , Amount Inspected by:(Signature) Date Inspected: ~hite copy to ~arren Snyder. Yellow copy to Pheasants Forever Chapter. Pin~ cony to Landowner. Date: Date Planted: See attached worksheet (blUe). �16 APPENDIX B. PHIP Specifications - 1993 SHRUB THI CKETS AND SUPPLEMENTAL WINDBREAKS Shrub (plum) thickets are the priority item. Small windbreaks, if planted, must be associated with a thicket and will not be f.unded if planted alone. Plantings will be eligible for funding only in farmed areas and must be within 0.1 mile of cultivated cropland. Plantings must remain undisturbed for at least 10 years. Maximum planted Funded: No more than 1 thicket (with or without wind barrier) can be per 80 acres. Each thicket must be at least 1/4 mile from another thicket. Size: Shrub thickets must be at least 1/10th acre (4,300 ft2) and no larger than 2/10ths acre (8,800 ft2) in size, and must include at least 8 rows (excluding windbreak rows). Twelve hundred (1,200) feet is the maximum linear feet of fabric funded per thicket. Supplemental windbreaks, if planted, must be placed on the north and west side of the thicket and must include no more than 900 linear feet total if straight and 1,200 linear feet total if L-shaped. They must include at least 3 rows, one of which must be juniper or cedar. Spacing between the thicket and the windbreak should approximate 100 feet (range 60 ft. minimum; 120 ft. maximum). Payment Rate: Payment will be at $0.58 per linear foot of fabric (6 ft. wide) to the maximums listed above. Changes in rates charged by the Colorado State Forest Service will be used to adjust payment rates. The maximum payment rate for approved private contractors will be $0.64 per linear foot of fabric. Supplemental payment rates/linear foot will be: $0.03 for application of polymer, $0.05 if 8 ft. wide fabric is used, and $0.01 for band application of an approved herbicide along the exterior edge of the fabric to reduce weed competition. Use of fertilizer will not be funded. Planting Dates: Between March 20 and May 15. Pre-Plant Treatment: Sites must be tilled, preferably the fall prior to planting. Tillage must be to bare soil with little residue remaining and must be deep enough to kill existing vegetation. Approved Juniper, Species: American Plum (priority) E. Red Cedar (potted only). and Russian olive (bare root), Rocky Mt. Between-row Spacing: A maximum of 10 feet will be permitted (6 to 8 feet spacing is recommended) for shrub thickets. A maximum of 12 feet will be permitted for wind barriers. In-row Spacing: A maximum of 8 feet will be permitted for shrub thickets (6 feet is recommended for plums) and 10 feet will be permitted for evergreens within wind barriers. Mulching: Woven polypropylene fabric is required for all plantings. Minimum fabric width is 6 ft. (3 ft. on each side of the row). Eight ft. width is preferred. Cost share is not available for drip systems. �17 PERENNIAL GRASS AND GRASS-LEGUME PLANTINGS Switchgrass provides tall cover that stands well over winter and has high value pheasants. Small unfarmed tracts, currently in short, sodded grasses, are recommended for revegetation to switchgrass. Other shorter, cool-season grasslegume mixtures may be used in roadsides where snowdrift is a problem. This practice is funded only in farmland (not rangeland) settings. for Pavment Rate: $40.00 per acre as a one-time payment for sites up to 10 acres. For each additional acre (in sites larger than 10 acres [40 acres maximum]) the rate is $35.00. An additional $15.00 per acre will be paid for breaking out sod. in heavily sodded sites and supplemental discing prior to planting switchgrass (this does not apply to roadsides). Preplant Soil Preparation: Adequate tillage to completely destroy existing perennial vegetation and to establish a moist, weed-free, firm seed bed is required. Interseeding is not approved. A preemergent herbicide (e.g. atrazine at up to 1 lb/acre) is recommended when planting switchgrass. Planting a tall-sorghum mix (for which payment is available) is recommended the first year. Switchgrass can be seeded into the residual sorghum without tillage during the subsequent spring. Planting Procedures: Planting procedures outlined in the Division's Game Informa~ion Leaflet #113 should be considered when planting switchgrass. In general, about 20 pure live seedsjft2 (2 - 3 lbs/acre) should be planted using a If a drill with double-disk furrow openers, 1-inch depth bands, and packer wheels. herbicide is not used, up to 1 lb/acre of an adapted dryland alfalfa and up to 1/2 lb of sweet clover should be added. Approved Species: In plots, switchgrass should comprise at least 75% of the live seed (alfalfa and sweet clover are approved additions). Within roadsides, switchgrass is the priority species where snowdrift is not a problem. Other approved warm-season grasses include bluestems and Indian grass. Where these-can not be used the tallest wheatgrasses (tall, intermediate, or standard crested) the roadside site will allow, should be used in combination with alfalfa (1 to 2 lbs/acre) • Planting Dates: Warm-season grasses Season Grass-legume Mixtures: March including switchgrass: 15-July 15. March 1S-May 15: Cool- Plot Duration: Grass and grass-legume plantings must remain ungrazed and undisturbed for at least 7 years. Roadsides should remain unmowed unless essential to reduce snowdrift. If essential, mowing should be delayed until after 1 August and restricted to the road shoulder. Prescribed burning, thinning tillage, or other renovation treatments to rejuvenate grass stands may be applied after 7 years. Grass stands that are relatively thin provide taller, better cover for pheasants. Legumes provide nitrogen and increase growth and quality when added to mixtures. DISTURBANCE TILLAGE AND TALL WILD ANNUALS This practice is primarily designated for small unfarrned odd areas within cropland. Wild sunflowers and other tall annuals which attain 3 - 6 ft. height in moderately open stands, stand better through winter than other herbaceous vegetation, and �provide excellent cover for broods, protection from blizzards and predators, and supplemental food. This is the most effective and least expensive approach for increasing pheasants and other upland game birds. Maximum Funded: 10 acres/quarter section. Fundincr Rate: $15.00/acre for breaking out sod. $30.00/year in patches 0.1 to 0.5 acre for subsequent annual disturbance (1/2 acre or larger will be considered 1 acre). $50.00/acre/year for sites up to 2 acres. Seeding wild sunflower or other approved wild annuals at 2 to 4 lbs. per acre will be funded at direct seed costs (see seed sources below). $30.00/acre/year will be paid for up to 5-acre patches of uncut wheat that contains moderate to dense stands of tall (4 to 6 ft.) annual weeds. Plot Dimensions: Short, relatively by drifting snow, are preferred. wide patches, which will not be easily inundated Placement: Adjacent to woody cover when possible. Draw bottoms that already contain weeds and above average moisture. are ideal. Sites containing noxious perennials should be avoided. Specifications: Initial tillage with a disk plow or mold-board plow is needed in sites containing perennial grass to destroy all perennial cover, preferably, immediately after the ground has thawed in early March. Large clod size is preferred to retain thin stands of annual forbs. Initial tillage in subsequent years should be conducted prior to May 1 A second thinning tillage may be used prior to the 1st of June. Spring tillage is needed each year to retain tall annuals. Annual grasses usually dominate if tillage is not used each spring. Wild sunflowers and annual ragweeds can be drilled or broadcast and harrowed at low rates to help establish tall annuals, if they are not already present. Known sources in Colorado include the Arkansas Valley Seed Company - Denver & Longmont, ana Sharpe Bros. Seed Company - Greeley. Retention: Tall annuals must remain undisturbed through March year. Sites can then be prepared for the next year's growth. ANNUAL SURVIVAL PLANTINGS of the following - SORGHUMS APPLICATION: On CRP, Annual Set Aside, and other cropland·or tilled wasteland. When applied within CRP fields, SCS specifications for CP-12 must be used (see supplement) • MAXIMUM FUNDED: 1/4 mile apart. 1 plot/SO-acre field, 2 plots/160 acres. Plots must be at least PAYMENT RATE: $40.00/acre/year for 1 ~ 5 acres (7 acres within center pivot corners) and $25.00/acre/year for additional acreages in tracts larger than 5 acres (12 acre maximum). Payment will increase by $5.00/acre for application of 30 lbs of nitrogen/acre. An additional one-time payment of $15.00 per acre will be paid for breaking out sod in CRP or heavily sodded sites and· supplemental discing prior to planting. PLACEMENT: Plots should be placed within or near cropland and �19 placed crosswise to prevailing winds. SPECIFICATIONS: Preplant Soil Preparation: Initial treatment: Adequate tillage existing perennial vegetation in early spring prior to annual growth. to destroy Subsequent vears: Preferably minimum tillage shredding of old materials needed prior to April 25. Annual application of nitrogen at 40 lbs./ac. is recommended. Plot Dimensions: are preferred to reduce Row Spacing: Minimum impacts as total plot width shall be 150 feet. Wider strips of drifting snow. (See special dimensions on CRP). 18 to 36 inches. Seed Specifications: At least 50% (75% preferred) of an adapted tall forage sorghum that will stand well with minimal lodging and will mature before frost. Up to 40% can be adapted varieties of grain sorghum. These can be mixed or planted in separate rows (i.e., 2 rows of grain sorghum to 6 rows of forage sorghum. These sorghums should equal a minimum of 75% of the total weight. Maximum amounts for other grains include: Dryland corn (25%), sunflowers (10%) and proso millet (10%). Addition of 1 to 2 lbs./ac. of wild sunflower seed is recommended (Source: Arkansas Valley Seed Company - Denver). Planting Dates & Rates: recommended. Sorghums should PLOT DURATION: 1 year. the following year. Between April 25 and June 05; Late April be planted at 4 - 8 lbs./acre. Sorghum SUPPLEMENT plantings FOR SORGHUM must remain PLANTINGS undisturbed WITHIN to mid-May through March of CRP SCS Notification: The CRP contract must be amended at the local SCS office prior to implementing CP-12 and breaking out food plots within CRP. This requires filling out a one-page form at your SCS office. The ASCS must be advised of the change for their records • .Once a winter cover-food plot is broken out within CRP it must remain as such until the end of the CRP contract. Payments will be made annually based on seeded acres. If the farmer wishes to discontinue this practice he must reestablish grass (required by the ASCS). Reimbursement will be at $40.00/acre to cover reseeding grass. Maximum Size: The maximum size is 3 acres per site. least 40 acres to be eligible for a. CP-12 food plot. CRP fields must contain at Plot Dimensions: Plantings must be between 66 and 99 feet wide, therefore, plots should be laid out as 2 or 3 (66 t9 99 ft. wide) adjacent strips with a 30 ft. buffer of untilled grass between strips to attain the 150 ft. minimum width. For example, a plot 99 ft wide x 440 ft. long equals 1 acre. Three adjacent plots will equal the maximum of 3 acres/site. Three 66 x 660 ft. plots or two 99 x 660 ft strips will also yield 3 acres. �20 Placement: Usually within the southeast corner when in CRP, preferably within 50-10' yds of edge, but location can vary depending on soil, wind, and moisture, and location of other winter covers if they occur. Sorghum plantings are not permitted in soils containing free lime effervescence), or soils that are deep sands or choppy sands. POST-HARVEST RETENTION OF TALL WHEAT (shows STUBBLE APPLICATION: Primarily in northeastern Colorado where stubble is usually left standing over winter. The objective is to provide taller, more secure cover for night roosting, feeding, loafing, and escape by pheasants through summer, fall, and winter. This practice must be applied during wheat harvest by raising the combine header in selected locations within the wheat field. PAYMENT RATE: $30.00 MAXIMUM FUNDED: per acre with 2 acres maximum per site. One site per 80 acres and 2 per 160 acres of small grain stubble. SPECIFICATIONS: Retention of at least 18 inches of wheat, rye, triticale, or barley stubble .during harvest. Harvest of heads is essential to prevent lodging of stubble under snow. PLACEMENT: preferably Patches of tall stubble should be near corn, sorghum, within the southeast part of the stubble fields. or other cropland RETENTION: This treatment should not be applied unless the entire wheat stubble field is to be left undisturbed through the subsequent fall and winter. Tillage, used, should not be initiated until after AprilS. SPRING NO-TILL RETENTION OF WHEAT Purpose: To provide secure stubble for nesting reduced tillage or no-tillage fallow methods. PAYMENT RATE: MAXIMUM FUNDED: landowner/year. $10.00/acre A maximum of 20 acres/quarter STUBBLE pheasants for up to 20 acres/quarter if through spring using either section. section, 4 tracts/section and SPECIFICATIONS: Stubble must have remained undisturbed since harvest, must remain untilled until after July 1, and must average at least 12 inches in height to qualify. If the remainder of the stubble field is to be conventionally fallowed, it must be tilled to remove >90% of the standing stubble prior to May 1. The retained ,tract must remain undisturbed, other than application o.f herbicides. Only herbicides of low toxicity that are approved for chemi~al fallow in wheat stubble may be used to control weeds and volunteer wheat (Roundup, Landmaster, Cyclone, etc.). Restricted use herbicides such as Paraquat may not be used. Dense annual weeds which exceed 12 inches within stubble fields may be SUbstituted for wheat stubble. PLACEMENT: Adjacent to green wheat or other cropland. It should not be surrounded �21 by fallow. PLOT DURATION: Until after July 1. SUPPLEMENTAL PURPOSE: equipment PAYMENT FOR CUSTOM WORK To prepare and plant sites when the landowner or does not have time to do the work. does not have the proper TREATMENT: Breaking out small tracts within CRP or sodded waste areas with a moldboard plow or heavy discing to completely destroy existing vegetation for reseeding to switchgrass or planting sorghum patches. Tillage must be to a depth of at least 6 inches. PAYMENT RATE: Option 1: Mold-board Option 2: Primary Replicate discing Planting: Payment = Plowing: Payment will be S16.00/acre heavy tillage with a tandem S8.00/acre/treatment. will Equipment transportation S1S.00/hour. or offset disc. = SlO.OO/acre. be at S6.00/acre to and between DEVELOPMENT WITHIN small tracts. INTENSIVE Supplemental payment will be STUDY AREAS PURPOSE: The Terrestrial Section of the Colorado Division of Wildlife (CDOW) has been assigned to evaluate the Pheasant Habitat Improvement Program. To conduct this evaluation at least 6 nine-section blocks of land have been selected for intensive development. Efforts will be made to attain 1 to 2 developments (sorghum survival plantings) per section within these blocks. Since annual set-aside of winter wheat will not be in effect during 1993 it is anticipated that the CDOW will need to rent small tracts of farmground and pay at rates competitive with crop production to accomplish its goals for evaluation. PAYMENT RATE: Payment for rental of small tracts, up to 5 acres in size, will be made at rates competitive with normal crop production only within sites designated as treatment blocks and where CRP, wasteland, or set aside for corn or feed grains is not available. Payment rates will not exceed S80.00/acre/year. Payment for land rental can be made using Pheasant Forever Chapter funds provided by this Cooperative Agreement but must be enacted by CDOW personnel and must be approved by the CDOW PHIP Coordinator. ��JOB PROGRESS REPORT State of: Colorado Project: ~W~-~1~6~7_-~R~ Work Plan: _1_ Job Title: Upland Bird Research ___f2_ Farming for Ring-necked Evaluation Period Covered: Author: Job _ 01 January Pheasants through 31 December, - Program Development and 1993 Thomas E. Remington Personnel: C. E. Braun, S. M. DeMasso, T. E. Remington, and W. D. Snyder, Colorado Division of Wildlife; M. J. Manfredo, and J. J. Vaske, Colorado State University. ABSTRACT A telephone survey was conducted of Colorado small game license buyers to investigate their experience, success, and satisfaction with ring-necked pheasant (Phasianus colchicus) hunting in Colorado. Respondents were also asked several questions designed to ascertain their experience with, and willingness to participate in, fee hunting for pheasants. While 92.4% of respondents had hunted pheasants in the past, only 63% hunted pheasants in Colorado in 1991. Colorado pheasant hunters generally hunted 3 days or less, bagged 5 or fewer pheasants (31% bagged none), and rated hunting as poor or fair. Many hunters (23%) had hunted pheasants on shooting preserves, clubs, on land they had leased, or had paid landowners for access. Twenty-two percent of respondents reported hunting pheasants out-of-state in 1991. Barriers to participation in pheasant hunting identified by hunters included lack of access to hunting areas, lack of birds, and distance to hunting areas. Only 9% of respondents had no interest in pheasant hunting the following year. Hunter willingness to participate in a hypothetical fee hunt of wild pheasants ranged from 75 to 20% as daily fee increased from $15 to $70. Flush rate, a purported measure of quality, did not influence participation. These data demonstrate substantial interest in pheasant hunting, and willingness to pay for it, by Colorado small game hunters. ��25 FARMING FOR RING-NECKED PHEASANTS Thomas - PROGRAM DEVELOPMENT AND EVALUATION E. Remington INTRODUCTION Small game license sales have declined markedly in Colorado over the past 10 years, from a high of about 200,000 in 1982 to a low of about 127,000 in 1993. Declines in participation in small game hunting are even more striking when considered as a percentage of Colorado's population. A recurring theme in surveys conducted to identify barriers to participation in hunting has been access to places to hunt, particularly places that have reasonably good hunting (Peterson and Manfredo 1993). Pheasants (Phasianus colchicus) are the most commonly hunted small game species in Colorado (Braun et al. 1992). Declines in small game hunter numbers may be attributable, in part, to difficulties in acquiring access to places to hunt (pheasants) (Enck et al. 1993, Rounds 1975) and/or hunter dissatisfaction because of declines in pheasant populations (Farris and Cole 1981). Thus, pheasants were identified as a pivotal species in any strategy to reverse declines in small game hunter participation, and access to quality pheasant hunting was identified as a key management goal (Braun et al. 1993). Pheasant populations have declined in eastern Colorado because of a lack of survival cover and secure nesting cover (Snyder 1984, 1985, 1991) caused by intensive farming. Landowners are unlikely to alter agricultural practices to benefit pheasants or other wildlife, or in some cases allow hunting access, without significant financial incentives (Matulich and Bagwell 1979, Bishop 1981, Rasker 1989). The Pheasant Cooperative Program was developed, so far only as a conceptual model, as a .means to develop habitat, increase local pheasant populations, and provide hunting access to interested hunters for a fee (Remington 1993). The objective of this study was to ascertain small game hunter experience, satisfaction, and future interest in pheasant hunting in Colorado, determine their willingness to pay to hunt wild pheasants, and to predict the impact that fee rate and quality of hunting would have on rates of participation in a fee hunting program for pheasants. P. N. OBJECTIVES .Develop a program to link hunters willing to pay for pheasant hunting opportunity with landowners willing to: 1) provide access for a fee;, 2) ,develop habitat for pheasants .within the program area, and 3) amend farming practices to make them more compatible with production and survival of -pheasants. SEGMENT 1. 2. 3. 4. 5. OBJECTIVES Investigate feasibility of CDOW participation in forming a trial Community Cooperative fee hunting program in Burlington, Yuma, and/or other interested communities. Consider developing a detailed study plan for implementation and evaluation of this program depending upon feasibility analysis. Evaluate landowner and hunter int~rest with the program. Recommend changes to improve the program development. Prepare annual report. �26 METHODS A telephone survey of 1991 small game license buyers was conducted between August and November 1992. A contingent valuation model was used to determine willingness to pay. Respondents expressing interest in pheasant hunting in the "next" year were asked how many days they would be likely to participate in a hypothetical pheasant hunt requiring them to pay a daily access fee of $15, $30, $50, or $70/day, and at variable levels of hunting quality. The hunting quality variable used was the number of pheasants likely to be flushed by the hunting party in a day of hunting; 1-5, 6-10, 11-15, or 16-20. Each respondent was asked how many days they were likely to participate under only 1 of the 16 possible permutations of daily fee and flush rates. The combination of fee and flush rate used was determined randomly prior to each call. Each combination of flush rates and fees was exposed to 30 potential hunters, for a total sample of 480. Sample sizes of responses to other questions varied. The survey sample was randomly .se1ected from 1991 resident, combination (small game and fishing) license buyers and resident small game license buyers, in proportion to the percentage of each in total license sales (75 and 25%, respectively). The sample frame was restricted to residents of northeastern Colorado counties (Adams, Arapahoe, Boulder, Cheyenne, Douglas, Elbert, El Paso, Jefferson, Kit Carson, Larimer, Lincoln, Logan, Morgan, Phillips, Sedgwick, Washington, Weld, and Yuma) who were 25 years of age or older at the time the survey was conducted. We presumed this group would have the greatest interest in a fee hunting program in northeastern Colorado because of proximity to the area described in the hypothetical hunt question «250 kilometers), and ability to pay. Phone numbers were taken directly from license stubs. If the number was incorrect or the phone had been disconnected, we attempted to find it in the phone directory. License holders were deleted from the survey if we were unable to acquire a correct listing, or after five unsuccessful contact attempts. The relationship between fee and flush rate and willingness to pay was evaluated in 2 ways; by analyzing the response measure (number of days likely to participate in hypothetical fee hunt) as a dichotomous response (0 days vs. ~ 1 day) using logistic regression, and as a continuous response variable (using subset of responses ~ 1 day) using analysis of variance. This allowed us to model the influence of fee and flush rate on decisions whether to participate or not, and also to model the influence of these factors on extent of participation (number of days) for those who said they were likely to hunt . .RESULTS Most (93.8%) license buyers contacted were willing to participate in the survey. The majority of the respondents were male (98%). Age distribution of respondents was skewed towards older hunters, with three~fourths over 45 years of age. Twenty-eight percent of the respondents earned an annual income of $35,000-$50,000, 25% earned $20,000-$35,000, 11% earned $60,000-$85,000, and 12% did not wish to answer this question. Most respondents (92.4%) had hunted pheasants in the past. Pheasant hunting experience ranged from 1 to 50 years, 69% had hunted 5 years or more (Fig. 1). Most (41%) rated themselves as intermediate hunters, 25% as advanced, 25% as novice, and 9% as experts. Eighty-eight percent of the respondents reported having hunted pheasants in Colorado in the past, but only 63% of these had hunted pheasants in Colorado in 1991. Hunters commonly hunted 1 (16%), 2 (22%), or 3 (16%) days �27 for pheasants in Colorado, but some hunters reported hunting as many as 30 days (Fig. 2). Most hunters (75%) reported harvesting 5 or fewer pheasants; 31% harvested no birds (Fig. 3). Pheasant hunting trips in Colorado were rated primarily as either poor (45%) or fair (29%), but 16% rated hunting as good, 6% very good, and 4% excellent. Twenty-three percent of those who hunted pheasants in Colorado in 1991 also hunted pheasants in another state. About 22% of respondents who hunted pheasants in Colorado in 1991 had participated in some form of fee hunting for pheasants. Fee hunters hunted on shooting preserves (60%), on land leased by a club (5%), on private land where they paid landowners directly for access (23%), or on land they had personally leased (16%). One in 4 respondents reported hunting pheasants in another state in 1991 (Fig. 4). Most (75%) had hunted in Kansas and/or Nebraska, although many had hunted in South Dakota or Iowa, and 20% had hunted in 2 or.3 other states. Hunters most frequently hunted 3 (16%), 4 (13%), or 5 (11%) days for pheasants out-of-state during the 1991 season (Fig. 4). Seventy-six percent of respondents said that acquiring access was difficult. Difficulties included landowners refusing permission (26%), lack of cover to hunt (14%), inability to determine ownership nr contact owner, and leased land. Other reasons for access difficulty were volunteered, including other hunters had ruined the opportunity by poor behavior in the past, too few birds, lack of time to search for areas/permission, lack of public land, and distance to hunting areas. Individuals who had never'hunted pheasants were asked why. Reasons included; no place to go' (32.5%); lack of interest (22.5%); not enough time (12.5%); too few birds (10%); or they preferred other types of hunting (15%). Respondents who had hunted pheasants, but not in Colorado, were asked why they did not hunt pheasants in Colorado. Access to places to hunt was the sole reason for 20%, and part of the reason for 32%. Thirty-one percent said that lack of pheasants was their primary reason, and lack of time was the reason for 25%. Other reasons included too far to travel to hunt, recently'arrived in Colorado, and availability of access in other states. Only 9% of respondents were not interested in pheasant hunting "next" year, while 40% were highly interested, 27% moderately interested, and 23% slightly interested. Twenty-four of 40 who had never hunted pheasants before, and 53 of 58 who had not hunted pheasants in Colorado, were at least slightly interested in hunting pheasants "next" year. Fee was a significant (£ ~ 0.01) factor influencing likelihood to participate in the hypothetical fee hunt while flush rat:e and income were not. Seventy-five percent of the respondents who were interested in'hunting pheasants "next" year indicated they were likely to participate in the hypothetical fee hunt at $15/day, 48% at $30/day, 41% at $50/day, and 20% at $70/day (Fig Sa-d). Likely participants included 8 of 24 who were at least slightly interested in pheasant hunting but had never hunted pheasants before, and 21 of 53 hunters who had not hunted pheasants in Colorado before but had an interest. Flush rate (£ = 0.10), but not fee (£ - 0.53), influenced how many days hunters were likely to participate. Extent of like.ly participation increased from an average of 2.9 days/willing hunter to 2.8, 3.4, and 3.9 days as flush rate increased from 1-5, 6-10, 11-15, and 16-20 birds/day, respectively. Expenditures per hunter did not vary with fee (£ - 0.36), but increased (£ = 0.08) from $36 to $39, $58, and $61 as flush rate increased from 1-5 to 6-10, 11-15, and 16:20 birds/day, respectively. �12 10 w .(!J ~ 8 Z W (.) 6 a: W a.. 4 2 o o 3 6 9 12 15 18 21 24 27 30 33 36 39 42 45 48 YEARS HUNTED Fig. 1. Pheasant hunting experience of Colorado pheasant hunters living in northeastern Colorado. 15 w C!J ~ Z 10 W o a: w a. 5 o o 1 2 3 4 5 6 7 8 9 10 11 12 15 16 18 20 21 24 30 DAYS HUNTED Fig. 2. Number of days spent hunting pheasants by hunters in Colorado during . 1991. �)(l DID NOT HUNT PHEASANTS OUT-OF-STATE LA.STYEAR 75.7% 151----W C!j ~ Z W 101----- o a: W Q..5 o 7 8 9 10 11 12 15 16 18 20 21 24 30 DAYS HUNTED Fig. 3. Harvest (birds/hunter) rates of pheasants by hunters during 1991. �30 35 r--------------------------------------------------- 30 25 --------- ------.----------------1 20 ..... --_.. -.. -... -- w (!J ~ -_.. - .--.--.----.-------1 Z W o c:: 15 - --------.-- .... -.----------~ W n, 10 5 o - --. o -- •.------ 1 2 3 4 5 6 7 8 9 1011 12 1415 16 172021 22 23 24 25 30 PHEASANTS HARVESTED Fig. 4. Percentage of survey respondents who hunted pheasants out-of-state 1991 and the number of days spent hunting pheasants out-of-state. in �31 DISCUSSION This survey was not completely representative of attitudes or pheasant hunting experiences of Colorado small game hunters as a whole because residents of southeastern Colorado and the west slope were excluded. It is reasonable to assume the counties that were included represented most of the resident pheasant hunters in 1992, so trends should be indicative. The positively skewed distributions of hunter ages and years of pheasant hunting experience suggest poor recruitment of new, young pheasant hunters. There was a fair degree of dissatisfaction with Colorado pheasant hunting, evidenced by low participation in 1991 by those who had hunted pheasants in the past, extent of hunting out of state, extent of preserve hunting, and the high percentage of hunters who rated pheasant hunting trips in Colorado as poor or fair. Hunting pheasants out of state is an indirect form of fee hunting because hunters incur additional or substantially higher costs for non-resident licenses, lodging, transportation, and meals. Many (34%) hunters who had participated in fee hunting for pheasants also hunted out of state. Thirty-nine percent of respondents had either paid a fee to hunt pheasants or hunted out of state. Thus, many Colorado pheasant hunters have had experience with paying for quality hunting. Hunter drop out, participation in fee hunting, and out of state hunting are probably all due in large part to reported difficulties in acquiring access to good pheasant hunting in Colorado. Although the amount of the fee charged-in the hypothetical hunt question strongly influenced participation rates, it is interesting to note that 75% of respondents would participate at $15/day, suggesting little philosophical opposition to paying a fee to hunt pheasants. Average expenditure per hunter did not differ among fee rates, because hunters dropping out as fee increased offset revenue increases due to higher fees per hunter. Expenditures per hunter were maximized at the highest flush rates because hunters said they were li~ely to hunt an average of 1 additional day at the highest flush rate. Flush rate was the only factor influencing number of days hunters were likely to participate, probably because hunters uncomfortable with the fee they were offered chose not to participate at all. Hunters seemed to make the decision to participate or not based on the fee charged, and decisions on how many days to participate based on the perceived quality of the trip. The flush rate variable was designed as a way to vary the quality of the hunt, although the description of the hunt itself may have predisposed individuals to expect a moderately high quality even if we superimposed a low flush rate. MANAGEMENT IMPLICATIONS This study demonstrated that Colorado small game hunters will support a fee hunting program for pheasants, both philosophically and by participating. This is true to the extent that hunters who said they'would participate actually do participate. These data also indicate that participation in pheasant hunting will increase with some type of fee hunting program, because some small game hunters who have not hunted pheasants, or have not hunted them in Colorado, will participate. Because this study was restricted to small game license buyers, how former hunters who have dropped out of small game hunting completely will react to a fee hunting program for pheasants is unknown_ 'Participation by this group should increase to the extent that barriers to their participation are removed. A fee hunting program that matches hunters willing to pay to hunt pheasants with landowners willing to provide access for a fee should be linked to habitat development because of �32 the quality aspect demanded by hunters. Hunters' willingness to pay as much as $70 per day for quality pheasant hunting should provide the financial incentive for landowners to develop habitat and perhaps alter agricultural practices most detrimental to pheasant populations. LITERATURE CITED Bishop, R. C. 1981. Economic considerations affecting landowner behavior. Pages 73-87 in R. T. Dumke, G. V. Burger, and J. R. March, eds. Wildlife management on private lands. Wisconsin Chapt. The Wildl. Soc., Madison. Braun, C. E., K. M. Giesen, R. W. Hoffman, T. E. Remington, W. D. Snyder. 1994. Upland Bird Management Analys.is Guide, 1994-1998. Colorado Div. Wildl., Denver 48 pp. Farris, A. L., and S. H. Co Le, 1981. Strategies and goals for wildlife habitat restoration on private agri'cultural lands. Trans. North Am. Wildl. and Nat. Resour. Conf. 46:130-135. Matulich, G. C., and G. Bagwell. 1979. On-farm pheasants enhancement potentials in irrigated agriculture. West J. Agric. Econ. 4:99-109. Peterson, M. R., and M. J. Manfredo. 1993. are recreationists' perceived problems Dimensions Persp. 15. Public access and preferred Rasker, R. 1989. Agriculture and wildlife: habitat management on farms in western State Univ., Corvalis. 24lpp. an economic analysis of waterfowl Oregon. Ph.D. Diss., Oregon in Colorado: what solutions? Human Remington, T. E. 1993. Farming for ring-necked pheasants - program development and evaluation. Colorado Div. Wildl., Prog. Rep., Fed. Aid Proj. W-167-R. Apr.: 25-36. Rounds, R. C. 1975. Public access to private Div. Wildl. Spec. Rep. 2. l79pp. lands for hunting. Snyder, W. D. 1984. Ring-necked pheasant nesting ecology on the high plains. J. Wildl. Manage. 48:878-888. 1985. Survival J. Wildl. Manage. of radio~marked 49:1044-1050. hen ring-necked and wheat farming pheasants 1991. Wheat stubble as nesting cover for ring-necked northeastern Colorado. Wi1d1. Soc. Bull. 19:469-474. U. S. Bureau of Census. 1993. Statistical Ed. Washington D. C. abstract Colorado' in Colorado. pheasants of the United States: in 113th �~r_------------------------~ ~ r_----------------------------------------------~ 5a. $15/DAY 5c. $50/DAY • 1-5 BIRDS II1II6-10BIRDS [J 11-15 BIRDS 016-20 BIRDS ------- 30 :~~:17~~~····· ~ . D11-15BIRDS o 16-20 BIRDS ------ 60 .- FLUSH RATE FLUSH RATE 40 20 10 ~ .• ~ . m........•....... I ..I......... u ......•. !HI':l..I ..• ill .........• !!II':'I 1 ..I 20 !:'! I I 0 W C!} ~ Z W 0 a: w a, o I .'111:'1 I 0 110 .url I -filI"I I _ r! I 2 3 1 n .HI I I 4 .1I1,:j I 5 .mn 6 .!UFII 7 5b. $30/DAY I 8 9 0 10 ~[li 1 3 4 5 6 7 8 9 10 11 5d. $70/DAY 12 13 14 • 1-5 BIRDS II1II 6-10 BIRDS 011-15 BIRDS 016'20 BIRDS oo~l~ FLUSH RATE FLUSH RATE 60 III ................... 40 ~-.".'" I 0 2 100 • 1-5 BIRDS 1liI6-10BIRDS [J 11-15 BIRDS, o 18-20 BIRDS 50 I- fll:ll 30 Ilirl 20 .- 0 0 1 2 3 4 5 II 7 a II 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 II 10 11 12 13 14 DAYS UKELYTO PARTICIPATE Fig. 5a-d. Number of days hunters interested in hunting pheasants were likely to participate in a hypothetical fee hunt at variable flush rates and daily fees of 15 (a), 30 (b), 50 (c) and $70 (d). w w ��35 JOB FINAL REPORT State of: Project: Colorado Upland Bird Research W-37-R/W-152-R/W-167-R : Jobs 8. 9. II. 12. 13 (In Part) Work Plan: 3 Job Title: Survival Estimates of Sage Grouse in North Park. Colorado Period Covered: 01 January 1973 through 31 December 1993 Author: Mari1et A. Zablan Personnel: Clait E. Braun, District Wildlife Managers, temporary assistants, and graduate students, Colorado Division of Wildlife; Gary C. White and Marilet A. Zab1an, Colorado State University ABSTRACT Leg-band recovery data from sage grouse (Centrocercus urophasianus) in North Park, Colorado were analyzed to obtain estimates of annual survival rates for the period 1973-87. Survival was significantly different between sexes. I found no differences in female surviya1 between subadu1ts and adults, ,nor any year-to-year differences in survival. Estimated survival rate for both age classes of females, over years, was 54.7%, and the estimated recovery rate was 7.8%. I found significant differences in male annual survival across years and between subadult and adult age classes. Mean estimated survival rate for males banded as adults was 38.4%, and was 51.7% for males banded as subadults. Mean estimated recovery rate of males banded as adults was 9.6%, and 10.2% for males banded as subadu1ts. Sage grouse band recovery data were used to assess the impact of weather variables on survival rates. None of the.variables tested (spring precipitation, winter precipitation, spring temperatures, winter temperatures) significantly predicted annual survival in males. However, statistical power of hypothesis tests was low due to the relatively small number of males banded (average number banded per year per age class 121). I found approximately 600 banded birds per year are required in each age- and sex-c Lass to···obtaincoefficients of variation of 20% for annual survival rates, allowing detection of a difference of 0.4 in annual survival rates. An assumption inherent in modern band-recovery models is that fate of each banded individual is independent of fates of other banded individuals. This assumption is probably violated in the practical application of sage grouse banding studies, due to interaction of trapping and banding techniques, behavior of birds during banding and hunting seasons, and harvest of birds by hunters in groups. I examined effects of aggregated �36 recoveries on survival and recovery estimator bias and confidence interval coverage. I simulated 36 recovery scenarios, consisting of varying annual banding sample size, average recovery group size, and years of banding and recovery. Accuracy of survival and recovery estimates was low. Coverage of 95% confidence intervals of survival and recovery estimates was well below 95%. Precision of estimates was greatly affected, as a standard deviation of empirical estimates was greater than the mean of standard errors calculated analytically. Bias was significantly different than zero; however, this was not as great a concern as effect on estimate precision. Extent of recovery aggregation was not determined in this study. However, the results indicate that even if aggregation is occurring at levels lower than those simulated, effects on survival and recovery estimate precision may be severe. �) I THESIS EVALUATION OF SAGE GROUSE BANDING PROGRAM IN NORTH PARK, COLORADO Submitted by Marilet A. Zablan Fishery and Wildlife Biology In partial fulfillment of requirements for the Degree of Master of Science Colorado State University Fort Collins, Colorado Fall 1993 COLORADO STATE UNIVERSITY JULy 7,1993 WE HEREBY RECOMMEND THAT THE THESIS PREPARED UNDER OUR SUPERVISION BY MARILEf A. ZABLAN ENTITLED EVALUATION OF SAGE GROUSE BANDING PROGRAM IN NORTH PARK, COLORADO BE ACCEPTED._AS FULFILLING IN PART REQumEMENTS FOR THE DEGREE OF MASTER OF SCIENCE. Cominittee on Graduate Work Department Chair �38 ABSTRACT OF THESIS EVALUATION OF SAGE GROUSE BANDING PROGRAM IN NORTH PARK,· COLORADO Leg-band recovery data from sage grouse tCentrocercus urophasianus) in North Park, Colorado were analyzed to obtain estimates of annual survival rates for the period 1973-87. Survival was found to be different between sexes (P=0.013). I found no differences in female survival between subadults and adults, nor any year-to-year differences in survival. I found significant differences in male annual survival across years and between subadult and adult age classes. Sage grouse band recovery data were used to assess the impact of weather variables on survival rates. I found survival differed between sexes (P=0.013). I found no differences in survival between female subadults and adults, nor any differences in survival across the 17-year period. Small sample sizes (average number banded per year per age class=57) probably caused Type II errors. I found significant differences in annual survival of males across the period 1973-1988, and between subadult and adult male age classes. None of the variables tested (spring precipitation, winter precipitation, spring temperatures, winter temperatures) significantly predicted annual survival in males. However, power of the tests was low due to the relatively small number of males banded (average number banded per year per age class = 121). I found approximately 600 banded birds per year are required in each age- and sex-class to obtain coefficients of variation of 20% for annual survival rates, allowing detection of a difference of 0.4 in annual survival rates. An assumption inherent in modem band-recovery models is that fate of each banded individual is independent of fates of other banded individuals. This assumption is probably violated in the practical application of sage grouse banding studies, due to interaction of trapping and banding techniques, behavior of birds during banding and hunting seasons, and harvest of birds by hunters in groups. Research on effects of assumption violation on sage grouse survival analysis is important for definition of extent of inference from banding studies and for comparison with alternative methods for study and monitoring of sage grouse populations. I examined effects of aggregated recoveries on survival and recovery estimator bias and confidence interval (Cl) coverage. Precision of survival and recovery estimates was greatly reduced. As aggregation of recoveries effectively reduces sample size of birds for analysis, these results indicate a need for quantification of possible aggregation effects and consideration of any aggregation effects into sage grouse banding study design and analysis. Department Marilet A Zablan of Fishery and Wildlife Biology, Colorado State University, Fort Collins, Colorado 80523 Fall 1993 ACKNOWLEDGMENTS This project would not have been possible without the continued guidance and support of my academic adviser, Dr. Gary C. White. I thank him also for his professional direction and forbearance throughout my work.. Dr. Clait E. Braun of the Colorado Divisi;n of Wildlife (CDOW) initiated the banding program and was instrumental in collecting the data used in this analysis. I am grateful for his enthusiasm, and dedication to long-term wildlife studies. This study also would not have been possible without the assistance of many field assistants, graduate students, and district wildlife managers over the years. I am also indebted to my graduate committee members, Dr. Kenneth P. Burnham (Colorado Cooperative Fish and Wildlife Research Unit) and Dr. Beatrice van Home (Department of Biology), for their assistance throughout this project. Finally, I thank my family and friends for their unconditional support, encouragement. Their support has been, and continues to be, essential in all my life endeavors. This project was funded through CDOW Federal Aid to Wildlife Restoration Project W-152-R, Upland Bird Research, Work Plan 3, Job 13a. . ii �39 TABLE OF CONTENTS Chapter 1. Estimation of Sage Grouse Survival in North Park, Colorado 1.1 Introduction 1.2 Study Area ; 1.3 Methods 1.4 Results 1.5 Discussion 1.6 Literature Cited Chapter 2. 2.1 2.2 2.3 2.4 2.5 2.6 2.7 Assessment of Weather Effects on Survival Estimates of Sage Grouse in North Park, Colorado _ Introduction Approach Objectives · Methods Results Discussion 2.6.1 Banding Analysis Sample Sizes 2.6.2 Sample Sizes for Detection of Weather Effects Literature Cited Page . . . . ·. . . Chapter 3. Evaluation of Effect of Aggregated Recoveries on Sage Grouse Survival Estimation 3.1 Introduction 3.2 Methods ........................................•......................... 3.2.1 Simulations 3.2.2 Analysis of Simulations 3.3 Results •................................................................. 3.4 Discussion 3.5 Literature Cited . . . . . . . . . . . . . . . . LIST OF TABLES Page 1-1. Number of sage grouse banded in North Park, Colorado, 1973-88 . 1-2. Recoveries by hunting season of male sage grouse banded as adults in North Park, Colorado, 1973-87 . 1-3. Recoveries by hunting season of male sage grouse banded as subadults in North Park, Colorado, 1973-87 . ~ ~ . 1-4. Recoveries by hunting season of female sage grouse banded as adults in North Park, Colorado, 1973-87 . 1-5. Recoveries by hunting season of female sage grouse banded as subadults in North Park, Colorado, 1973-87 . 1-6. Model tests from Program BROWNIE for male sage grouse in North Park, Colorado, 1973-87 . . 1-7. Model tests from Programs BROWNIE and ESTIMATE for female sage grouse in North Park, Colorado, 1973-87 ' . 1-8. Survival estimates from Program BROWNIE for male sage grouse banded as adults in North Park, Colorado, 1973-87 . 1-9. Survival estimates from Program BROWNIE for male sage grouse banded as subadults in North Park, Colorado, 1973-87 . 2-1. Contingency table chi-square test for annual survival and recovery rate differences due to sex of sage grouse, North Park, Colorado . 2-2. Key to names of models used to test weather effects on annual survival . 2-3. Goodness-of-fit and likelihood ratio tests for models of weather effects on male sage grouse survival in North Park, Colorado, 1973-87 . iii �40 LIST OF TABLES (continued) 2-4. Input to simulations of models testing weather effects on annual survival of sage grouse 2-5. Results of simulations of models "k!sting.,}Veathereffects on annual survival of sage grouse .... 3-1. Survival and recovery estimates (S and f) confidence interval coverage, based on estimates (%) generated from 1,000 simulated_recov_gry matrices with five years of recovery . 3-2. Survival and recovery estimates (S and f) confidence interval coverage, based on estimates (%) generated from 1,000 simulated recovery matrices with ten years of recovery . 3-3. Maximum likelihood analysis of variance for survival rate estimate 95% confidence interval coverage . 3-4. Maximum likelihood analysis of variance for recovery rate estimate 95% confidence interval coverage ' . 3-5. Survival and recovery estimator bias, based on estimates (%) generated from 1,000 simulated recovery matrices with five years of recovery . 3-6. Survival and recovery estimator bias, based on estimates (%) generated from 1,000 simulated recovery matrices with ten years of recovery ....: . 3-7. Analysis of variance for percent relative bias of survival estimates (PRB ~ . 3-8. Analysis of variance for percent relative bias of recovery estimates (PRB f) . LIST OF FIGURES Page 1-1. Model parameters and relationship between models for birds banded as adults or as subadults and adults (based on Brownie et al. 1985) . 2-1. Average required banding sample size of male and female sage grouse in North Park, Colorado, based on desired coefficient of variation (CV) of annual survival rate estimates . .. ' 2-2. Simulated survival estimates for subadults and adults as a function of winter precipitation .', iv �CHAPTER 1: ESTIMATION OF SAGE GROUSE SURVIVAL IN NORTH PARK, COLORADO 1.1 INTRODUCTION The sage grouse tCentrocercus urophasianus) population in North Park, Colorado is managed by the Colorado Division of Wildlife (CDOW) for autumn harvest by hunters and for year-round public viewing. Hunting season objectives include maximum hunter recreation with no detriment to breeding population levels. Meeting this objective requires knowledge of population status in the face of changing environmental conditions, due to both anthropogenic and stochastic factors. Understanding dynamics of a wild population requires estimates of such critical population parameters as recruitment and survival. Precise estimates of survival and recovery rates are required to detect differences in a population among years. In 1973 CDOW began a long-term banding study of the North Park sage grouse population. Through 1988, CDOW personnel banded 5,627 sage grouse as part of this banding study. My objective was to estimate and compare annual recovery and survival rates for male and female age classes for the period 1973-1987 from band recovery data. Sample size of sage grouse banded was also reviewed with respect to precision of estimates of annual recovery and survival rates. 1.2 STUDY AREA The area referred to as North Park is located in Jackson County, in northwestern Colorado. North Park is a large intermontane basin of rolling topography at an elevation of about 2,500 m surrounded by mountains ranging upwards to 3,800 m (Braun 1984). Drainage is to the north by a series of small streams that flow into the North Platte River. Climate is typified by long, cold winters, and a short growing season (about 46 days). Mean annual precipitation is 23 em and mean annual temperature is 2.5 C (U.S. Dep. Commerce 1973-88). Prevailing winds are from the southwest with frequently high velocities, particularly in winter and spring (Emmons and Braun 1984). Big sagebrush, primarily Artemisia tridentata uryomingensis and A t. uaseyana, is the dominant shrub in North Park and occurs in mixed stands with native bunchgrasses and perennial forbs. Other locally common shrubs include sagebrush (A argilosa, A cana, and A longiloba), rabbitbrush (Crysothamnus spp.), broom snakeweed tGutierrezia sarothrae), black greasewood tSarcobatus vermiculatus), willows (Salix spp.), and antelope bitterbrush tPurshia tridentata) (Emmons and Braun 1984). Herbaceous vegetation consists primarily of low-growing perennial forbs and perennial bunchgrass with few annual forbs (Beck 1977). Potential sage grouse range encompasses 1,870 km2 including 1,250 km2 of uplands and 620 km2 of native hay meadows (Beck 1975). No annual cultivated crops are grown in North Park. Dominant land uses are livestock (primarily cattle) grazing and hay production. 1.3 METHODS During breeding season, which usually peaks from mid-March to late-April, birds were captured at night near active leks using spot-lighting techniques described by Giesen et al. (1982). Birds were banded during a 2-week period at the same locations each spring; starting dates for banding varied, as did dates of peak activity on leks. There was no programmatic banding effort at any other time of the year. With very few exceptions throughout the length of the study, dates of banding were from the :first week in March through the third week in May. Each aluminum leg-band had a unique number inscribed on it, permitting unique identification of banded individuals. Almost all band recoveries came from birds hunters shot and reported to CDOW; band recoveries also came from other reports of banded birds found dead. CDOW operated sage grouse hunter check stations at Willow Creek Pass on Colorado Highway 125 and near Cameron Pass, east of Gould, on Colorado Highway 14. These check stations were operated on both days of opening weekend of each hunting season. Hunter check stations were also operated on the second Sunday in earlier years (1975-83 except Gould check station was not . operated on the second Sunday in 1976). Stateline check station was operated on opening weekend 1 �42 and the second Sunday from 1974 through 1979 (except the second Sunday in 1979). Muddy Pass check. station was operated on opening weekend in 1974 and only on the first Sunday in 1980. placed volunteer wing collection barrels and signs along Colorado Highway 14 northeast of Muddy Pass and near the top of Cameron Pass, northeast of Three-Way on Colorado Highway 127, and at Willow Creek Pass. Volunteer wing collection barrels were available to hunters during the entire season, including all days check. stations were not operated east of Gould and at Willow Creek Pass. Barrels were placed at Arapahoe Creek in 1982 and 1983, and in 1982 at Seymour Reservoir. Birds were classified as subadult (less than 1 year old) or adult (more than 1 year old) at time of banding. Captured birds were classified to sex and age by wing molt and primary length (Beck. et al. 1975). An age-Classification scheme for assigning ages of birds when recaptured or recovered was devised. Birds banded as sub adults were classified as sub adults if recaptured or recovered before or during the first autumn after banding (01 August through 31 October). Otherwise, birds banded as subadults were classified in the database as adults if they were recaptured or recovered during the subsequent winter or spring (01 November through 31 July) or beyond. Age classification was an important consideration, as survival models tested included an age component. The database was also designed to facilitate entry of any future recapture and recovery data. While other data were 'available from the banding study, only recoveries of dead birds were used for this survival analysis. The interval to which the term "annual survival" applied was from midpoint to midpoint of each banding season, with the midpoint defined as 23 April of each year. CDOW personnel banded a total of 5,627 birds during the period 1973-88 (Table 1-1). I constructed band recovery matrices for each age and sex class. These triangular matrices (Tables 1-2 to 1-5) summarized, for each age and sex class banded, band recoveries by hunting season. These recovery matrices constituted the input for programs BROWNIE and ESTIMATE (Brownie et al. 1985). I used programs BROWNIE and ESTIMATE to test for differences in survival and recovery rates due to age and sex, and to calculate estimates of annual recovery and survival rates by sex and age class. cnow 1.4 RESULTS The null hypothesis that adult males and females had equal recovery and survival rates was rejected (P = 0.013). Survival rates for males differed by year (P = 0.007) and goodness-of-fit and likelihood ratio tests indicated that Model HI fit the _data for males (Table 1-6). Assumptions of oModel HI are that annual survival and recovery rates are year-specific, and that subadults have different survival and recovery rates than adults (Table 1-7). Model He was selected for females; assumptions of Model He are that survival, hunting, reporting, and hence recovery rates, are year-specific but independent of age. Model Ho is equivalent to assuming Model 1 applies to pooled adult and young recovery data (Figure 1-1). Program ESTIMATE (Brownie et al. 1985) further demonstrated that the simplest model, Model 3 (constant survival and recovery rates across years, Figure 1-1), fit the combined female recovery data (P = 0.336). I calculated survival and recovery rate estimates using programs BROWNIE (males) and ESTIMATE (females). Survival rate estimates for males banded and released as adults (Table 1-8) ranged from 9% (95% confidence interval (CD: 0 - 19.9%) in 1985 to 84.6% (95% CI: 33.8 - 100%) in 1979. Mean estimated survival rate of males banded as adults was 38.4% (95% CI: 32.9 - 43.8%). Estimated survival rates for males banded and released as sub adults (Table 1-9) ranged from 15.0% (95% CI: 0 - 37.5%) in 1986 to 100% (95% CI: 27.3 - 100%) in 1974 and averaged 51.7% (95% CI: 40.063.3%). Mean estimated recovery rate for males banded as adults was 9.6% (95% CI: 8.2 - 10.9%) and estimates ranged from 6.7% (95% CI: 3.5 - 9.9%) in 1980 to 14.7% (95% CI: 5.5 - 24.0%) in 1986. Mean estimated recovery rate for males banded as subadults was 10.2% (95% CI: 8.8 - 11.7%) and estimates ranged from 4.8% (95% CI: 0.7 - 8.9%) in 1986 to 13.2% (95% CI: 6.8 - 19.7%) in 1978. Estimated survival rate for both age classes of females; over years, was 54.7% (95% CI: 50.1 - 59.2%), and estimated recovery rate for females was 7.8% (95% CI: 6.7 - 8.9%). To clarify, each value is one estimate for all years, rather than an average of several different annual estimates. 2 �43 1.5 DISCUSSION While significant differences were found between survival and recovery rates of males of both age classes, and between years, survival estimates had unacceptably wide confidence intervals (Tables 1-8 and 1-9). Several annual survival rate estimate confidence intervals exceeded 1.0 on the upper end and 0 on the lower end. Even after truncating confidence interval limits in all such cases, average confidence interval width for males was 0.509. Any conclusions based on these point estimates of survival would be extremely tenuous, especially population-level inferences about effects of events such as changes in harvest regulation or severe weather, as examples. While Model 3 of program ESTIMATE (Brownie et al. 1985), with constant survival and recovery rates across years, fit the female recovery data, I do not conclude that age and time differences were nonexistent. Through identification of one survival rate estimate for both female age classes over all 17 years of the study, the model selected would imply subadult and adult female sage grouse undergo identical responses to mortality factors. Subadult and adult female sage grouse have significantly different nesting success and chicks per hen ratios (C. E. Braun, pers. commun.); it is not likely that annual survival rate is identical for both age classes. While Model 3 was adequate for observed data, it also implied that any differences during the 17-year period were not statistically significant. It is far more likely that biologically significant differences during 17 years were undetected due to low statistical power. Were sample sizes of banded birds larger, more credence could be given to results of statistical tests from programs BROWNIE and ESTIMATE. WIlson et al. (1989) have suggested that a minimum of 200 birds in each age and sex category be banded in any banding study of this type. Contrary to that guideline, an average .of 120 males and 57 females of each age class were banded each spring. Thus, sample sizes were inadequate to determine whether or not significant differences in annual survival existed. 1.6 LITERATURE CITED Beck, T. D. 1. 1975. Attributes of a wintering population of sage grouse, North Park, Colorado. M.S. Thesis, Colo. State Univ., Fort Collins. 49pp. Beck, T. D. 1., R. B. Gill, and C. E. Braun. 1975. Sex and age determination of sage grouse from wing characteristics. Colo. Div. WIldl. Game Info. Leafl. 49 (revised). Beck, T. D. 1. 1977. Sage grouse flock characteristics and habitat selection in winter. J. Wildl. Manage. 41:18~26. Braun, C. E. 1984. Attributes of a hunted sage grouse population in Colorado, U.S.A. Pp. 148-162 in P. J. Hudson and T. W. 1. Lovel, eds. 3rd Int. Grouse Symp., World Pheasant Assoc., York Univ., U.K. Braun, C. E., and T. D. 1. Beck. 1985. Effects of changes in hunting regulations on sage grouse harvest and populations. Pp, 335-343 in S. L. Beasom and S. F. Roberson, eds., Game Harvest Management. Proc. 3rd Int. Symp., Caesar Kleberg Res. Inst., Kingsville, Tex. Brownie, C., D. R. Anderson, K. P: Burnham, and D. S. Robson. 198'5. Statistical inference from band recovery data - a handbook. Second ed. U.S. Dep. Inter., Fish Wildl. Servo Resour. Publ, 156. 305pp. Eberhardt, L. L. 1985. Assessing the dynamics of wild populations. J..WIldl. Manage. 49:997-1012. Emmons, S. R,· and C. E. Braun. 1984. Lek attendance of male sage grouse. J. Wildl. Manage. 48:1023-1028. Giesen, K. M., T. J. Schoenberg, and C. E. Braun. 1982. Methods for trapping sage grouse in Colorado. Wildl. Soc. Bull. 10:224-231. U.S. Department of Commerce. 1973-88. Climatological data: annual summary, Colorado. Nat. Ocean. Atmosph. Admin. 78-93. Wilson, K. R., J. D. Nichols, and J. D. Hines. 1989. A computer program for sample size computations for banding studies. U.S. Dep. Inter., Fish. WIldl. Tech. Rep. 23. 19pp. 3 �44 Table 1-1. Number of sage grouse banded in North Park, Colorado, 1973-88. Eema1~ Ma1~ Year Subadults Adults All Subadults Adults All Total 109 288 49 191 1973 80 99 179 41 1974 54 88 142 22 68 27 . 1975 138 153 291 62 68 130 421 1976 120 114 234 71 74 145 379 1977 183 123 306 101 133 234 540 1978 106 98 204 32 22 54 258 1979 111 146 257 48 52 100 357 1980 127 173 300 72 94 166 466 1981 110 190 300 31 38 69 369 1982 110 190 300 42 69 111 411 1983 152 157 309 33 32 65 374 1984 102 92 194 123 111 234 428 1985 163 88 251 46 28 74 325 1986 104 51 155 33 30 63 218 1987 117 85 202 71 29 100 302 1988 115 88 203 48 59 107 310 Average 118 121 239 55 58 113 352 1,892 1,935 3,827 876 ·934 1,800 5,627 Total 4 �Table 1-2. Recoveries of male Sage grouse banded as adults, by hunting season, North Park, Colorado, 1973-87. Recoveriesby huntingseason Year Banded 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 1973 7 4 1 0 1 0 0 0 0 0 0 0 0 0 0 8 5 1 0 0 0 0 0 0 0 0 0 0 0 10 4 2 0 1 0 0 0 0 0 0 0 16 3 2 0 0 0 0 0 0 0 0 0 12 3 2 3 0 0 0 0 0 0 0 10 9 3 0 0 0 0 0 0 0 14 9 3 3 0 0 0 0 0 9 5 2 1 0 1 0 0 16 5 2 0 1 0 0 19 6 2 1 0 1 15 3 0 0 0 8 5 1 0 10 1 0 8 1 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 10 Table 1-3. Recoveries of male sage grouse banded as subadults, by hunting season, North Park, Colorado, 1973-87. Recoveriesby huntingseason Year Banded 73 74 75 1973 6 4 6 6 5 2 18 6 17 1974 1975 1976 1977 1978 76 77 79 80 81 82 83 84 85 86 87 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 6 2 0 0 0 0 0 0 0 0 5 6 2 1 0 0 0 0 0 0 20 9 6 2 1 0 0 0 0 0 14 4 3 1 0 0 0 0 0 0 13 4 0 1- 0 0 0 0 0 13 5 3 1- 0 0 0 0 13 5 4 0 0 0 0 1 0 78 0 1979 1980 1981 1982 7 1983 3 15 1984 1985 1986 10 2 -0 12 4 0 16 4 5 0 0 2 8 1987 5 �Table 1-4. Recoveries of female sage grouse banded as adults, by hunting season, North Park, Colorado, 1973-87. Recoveries by hunting season Year Banded 1973 73 5 1974 74 75 76 77 78 79 80 81 82 83 84 85 86 87 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 1 0, 2 0 0 6 2 4 1 0 2 2 3 0 1975 1976 1977 15 3 0 5 1"978 0 0 0 0 0 1979 3 3 0 0, 0 7 5 2 2 0 0 0 0' 0 0 1 2 0 2 0 1 0 0 4 2 0 1 2 0 1 1 1980 4 1981 1982 5 1983 1 1984 6 1985 1986 , 1987 6 Table 1-3. Recoveries of male sage grouse banded as subadults, by hunting season, North Park, Colorado, 1973-87. Recoveries by hunting season Year Banded 77 78 79 80 81 82 83 84 85 86 87 4 2 0 0 0 0 0 0 0 0 0 0 2 0 0 0 0 0 0 0 0 0 0 0 2 0 0 0 0 0 0 0 0 0 4 2 1 0 0 0 0 0 0 0 7 6 6 0 1 0 0 0 5 0 0 0 0 0 0 0 4 0 0 0 0 0 0 8 4 1 2 0 0 0 0 4 3 0 0 0 0 0 2 0 2 0 2 0 0 0, 0 1983 1984 6 7 2 2 6 1 4 1973 1974 1975 1976 1977 1978 73 74 75 2 2 6 76 2 2 1979 7 1980 1981 1982 1985 1986 0 0 3 1987 8 6 �4/ Table 1-6. Model tests from program BROWNIE for male sage grouse in North Park, Colorado, 1973-87. Models are described in Figure 1-1. Goodness-of-Fit i E'(;(- > x: calculated) calculated d.f. s, 75.49 59 0.073 Hoz 75.85 50 0.011 HI 34.34 30 0.268 H2 19.46 17 0.303 Model Tests between Models Table 1-7. x! NX > x: calculated) calculated d.f. Ho vs. HI 41.16 29 0.067 . HI vs. H2 14.88 13 0.315 H2 vs. H3 15.45 13 0.280 Hoi vs. Hoz 27.57 . 28 ·0.487 a, vs. HI 46.78 26 0.007 Case Model tests from program BROWNIE for female sage grouse in North Park, Colorado, 1973-87. Models are described in Figure 1-1. Goodness-of-Fit x: calculated) Model i calculated d.f. a, 53.63 57 0.602 Hoz 35.61 38 0.581 HI 23.91 28 0.686 H2 17.35 15 0.289 E'(;(- > Tests between Models x: calculated) Case i calculated d.f. a, vs. HI 29.72 29 0.428 HI vs. H2 6.56 13 0.923 H2 vs. H3 16.89 13 0.204 Hoi vs. Hoz 30.94 28 0.320 Hoz vs. HI 25.93 26 0.467 7 E'(;(- > �· ~8 Table 1-8. Survival estimates from program BROWNIE for male sage grouse banded as adults in North Park, Colorado, 1973-87. Year Estimate SE 95% CI 1973 0.360 0.167 0.032 - 0.688 1974 0.645 0.242 0.171 - 1.119 1975 0.220 0.079 0.064 - 0.375 1976 0.448 0.148 0.158 - 0.737 1977 0.325 0.102 0.126 - 0.524 1978 0.558 0.155 0.254 - 0.863 1979 0.846 0.259 0.338 - 1.353 1980 0.284 0.098 0.093 - 0.475 1981 0.319 0.098 0.127 - 0.511 1982 0.430 0.144 0.149 - 0.712 1983 0.226 0.090 0.050 - 0.403 1984 0.395 0.171 0.061 - 0.729 1985 0.090 0.056 -0.020 - 0.199 1986 0.230 0.156 -0.075 - 0.535 Average 0.384 0.028 0.329 - 0.438 Note: Annual survival is defined by average midpoint of banding period, or, from 23 April to 22 April of the following year. Table 1-9._ .Survival estimates from program BROWNIE for male sage grouse banded as subadults in North Park, Colorado, 1973-87. . Year Estimate SE 1973 .0.890 0.322 0.259 - 1.521 1974 1.201 0.474 0.273 - 2.129 1975 0.568 0.178 0.219 - 0.917 1976 0.738 0.234 0.280 - 1.196 1977 0.447 0.128 0.195 - 0.699 1978 0.370 0.140 0.096 - 0.644 1979 0.413 0.202 0.016 - 0.809 1980 0.541 0.202 0.145 - 0.938 1981 0.521 0.189 0.151 - 0.891 1982 0.454 0.206 0.049 - 0.858 1983 0.490 0.181 0.134 - 0.845 1984 0.291 0.165 -0.032 - 0.613 1985 0.160 0.085· -0.008 - 0.327 1986 0.150 0.114 -0.074 - 0.375 Average 0.517 0.059 0.400 - 0.633 8 95% CI �Parnmeters S;, ~, ff S;, ~ S-, fI S, f Birds banded as adults Birds banded as two age classes Parameters Model H3 s, ~,st, fi, Si', t S~" I , f," I fi', I I Model 0 ---- Model Hz t t I I I I Model 1 Model HI t t I I I I Model 2 Model t t I I I I Model 3 U .L..1.Q2 Model HoI S, ~, S', fi S,f,S',f' Kev to model parameters S ~= Constant annual survival rate for adults. f = Constant band recovery rate for adults. S' = Constant annual survival rate for young. f' = Constant band recovery rate for young. S, = Survival rate for adults in year i. t; = Band recovery rate for adults in year L ff = First year band recovery rate for adults. Si = Survival rate for young in year i. fi = Band recovery rate for young in year L Si' = Survival rate in yr:M i for subadults (i.e., survivors of young). fi' = Band recovery rate in year i for subadults. S]" = Survival rate in yr:M i for adults banded in year L fi" = Band recovery rate in yr:M i for adults banded in year i. Figure 1-1. Model parameters and relationship between models for birds banded as adults or as two age classes (based on Brownie et ale 1985). Arrows indicate simplest to most general models. 9 �so CHAPTER 2: stravrv ASSESSMENT OF WEATHER EFFECTS ON AL ESTIMATES OF SAGE GROUSE IN NORTH PARK, COLORADO 2.1 INTRODUCTION The sage grouse (Centrocercus urophasianus) population in North Park, Colorado is managed by the Colorado Division of WIldlife (CDOW) for autumn harvest by hunters and for year-round public viewing. Hunting season objectives include maximum hunter recreation with no detriment to breeding population levels. Establishment of hunting season regulations results from several desired objectives including providing maximum hunter recreational opportunity without detrimentally affecting breeding population levels. Meeting this objective requires knowledge of population status in the face of changing environmental conditions, due to both anthropogenic. and stochastic factors. Thus, development of an optimal management plan for sage grouse requires predictive population models. Autenrieth et al. (1982) proposed that important components of a simple predictive model are: 1. sex ratio of breeding population, 2. estimate of breeding population size, 3. nesting success by age class of female, 4. .young survival per brood to a certain time, 5. percentage of females that reproduce by age class, 6. average clutch size by age class of female, 7. egg fertility, . 8. composition of annual harvest, 9. estimate of annual harvest size, 10. climatic variables during important events (nesting, hatching, early brooding, winter), 11. habitat variables (alteration patterns, quality of food, availability, ete.), and 12. land ownership patterns and attitudes of landowners. While all these components may be important, sage grouse populations may be limited to a level that existing habitat can Support during the critical period of a year (Autenrieth et al. 1982). This "critical period" may vary within and among areas. While summer forb availability may be limiting in one area, adequate sagebrush (Artemisia spp.) cover above snow may be limiting elsewhere. Snow depth during some years may affect sage grouse survival in North Park, for example, by limiting food accessibility and/or availability. Addressing a similar question, Bartmann and Bowden (1984) conducted a study to predict mule deer (Odocoileus hemionus) mortality from weather data in Colorado and found that snow depths during early winter were most associated with deer winter mortality whereas no particular period of mean temperatures could be similarly identified. Weather has been shown to greatly affect reproductive success and survival of gallinaceous game bird species. Larsen and Lahey (1958) determined a correlation between spring temperatures and population size the following spring for rUfred grouse tBonasa-umbellusr and ring-necked pheasants tPhasianus colchicus). Martinson and Grondahl (1966) found that high reproductive success of pheasants was associated with relatively cool weather during May and June and concluded (pg. 78) that "productivity and survival among pheasants in southwestern North Dakota were related to May and June weather. Specifically, high productivity and high survival were closely associated with, and perhaps dependent on, relatively high precipitation during May and June. Consequently, the amount of rainfall occurring during May and June could determine the trend of the pheasant population." Similarly, in North Park, precipitation during April and May may be a good predictor of lek. counts the following year (C. E. Braun, pers. commun.). Survival of British grey herons (Ardea cinerea) was modeled by North and Morgan (1979) using weather data. Survival rates for first-year birds were weather-dependent (and therefore timespecific) and the constant survival rate for second-year birds differed from the constant annual survival rate assumed for all older birds. Based on their knowledge of life history and behavior of grey herons, North and Morgan (1979) formulated separate models of survival which incorporated 3 measures of weather severity: 1) averages of January, February, and March mean temperatures, 2) lengths of 10 �51 periods above and below freezing during winter (Oct - Mar), and 3) lengths of periods during which each daily minimum temperature was no greater than some threshold value. I investigated effects of weather on annual survival estimates of sage grouse in North Park, Colorado. I evaluated sample size of birds banded, and determined sample sizes required to detect certain changes in annual survival (e.g., due to weather effects). 2.2 APPROACH From 1973-1988, personnel captured and banded 5,627 sage grouse in North Park, Colorado as part of that agency's long-term monitoring study. Trapping occurred in spring, primarily by spotlighting at night (Giesen et al. 1982). Numbers of grouse banded by age- and sex-class varied (Table 1-1). Almost all band recoveries came from birds hunters shot and reported to band recoveries also came from other reports of banded birds found dead (for details, see Chapter 1). The first step in analysis of fluctuating survival in response to weather conditions was determining whether survival rates actually varied among sexes, age classes, and years. I used program B:ROWNIE (Brownie et al. 1985) to test for differences in survival between sexes. The null hypothesis that adult males and females have the same annual survival and recovery rates was rejected (P = 0.013, Table 2-1), so data were not pooled over sexes. I then used Program BROWNIE to fit a series of models to subadult and adult band recovery data for males and females separately to explore differences in survival and recovery rates between subadults and adults and across years (for details, see Chapter 1). Model HI fit band recovery data best for males while Model Ho fit the data best for females (Tables 1-6 and 1-7). Assumptions of Model HI are that annual survival and recovery rates are year-specific, and that subadult birds have different survival and recovery rates from those of adults. Assumptions of Model Ho are that survival, hunting, reporting, and hence recovery rates, are year-specific but independent of age. Because I concluded that sub adult and adult male birds had different survival and harvest rates that were year-specific (Model HI)' the source of variation in survival rates was investigated as described in the null hypotheses listed below. Further, program ESTIMATE (Brownie et al. 1985) was used to show that Model 3 (constant survival and recovery rates across years, Figure 1-1) fit the combined female recovery data (P = 0.336). Because this model that assumes constant survival and recovery across years and female age-classes, fitted the observed data did not indicate biological differences were nonexistent and that subadult and adult female sage grouse underwent identical responses to mortality factors. Further, fit of the model ili3), that includes constant survival and recovery rates over years, to the observed data did not imply that differences across the 17 years were not important. In fact, quite the opposite is likely true for both hypotheses. The total number of birds banded were much lower for females (N = 1,800 versus N = 3,827 for males); hence, number of band recoveries was also lower. Program BAND2 (Wilson et al. 1989) suggests banding studies not be conducted with less than 200 individuals of each category banded each year. Particularly in the case of females, banding and recovery data were too sparse to detect age-specific and year-specific variation in survival (Table 1-1). Incorporation of covariates into survival models is an approach which seeks to explain sources of variation which have been demonstrated to exist. It is possible, however, that tests of non-specific time effects, such as those tested for with programs BROWNIE and ESTIMATE, may not be sufficient to ascertain differences in survival over time. Tests of specific time effects (e.g., weather) are more powerful than non-specific tests (K. P. Burnham, pers. commun.). I elected to not incorporate covariates in survival models for females, as significant variation in survival rate was not demonstrated previously and female recovery data were so sparse. Therefore, the following null hypotheses were tested only for male sage grouse. cnow cnow; 2.3 OBJECTIVES The following hypotheses were. tested for male sage grouse banded in North Park, Colorado: Hasp: Spring precipitation has no effect on annual survival, ~: Winter precipitation has no effect on annual survival, Host: Spring temperature has no effect on annual survival, and Howt: Winter temperature has no effect on annual survival, 11 �52 where "spring precipitation" was total accumulated precipitation during April and May, "winter precipitation" was total accumulated precipitation during January, February, and March, "spring temperature" was the average of daily low temperatures during April and May, and "winter temperature" was the average of daily low temperatures during January, February, and March. 2.4 METHODS To test the hypotheses that weather variables had no effect on annual survival of male sage grouse, I used the numerical optimization program S~VIV (White 1983). Annual weather differences were incorporated into user-specified cell probability functions for survival estimation. Spring and winter precipitation and temperature were incorporated as covariates in a logistic model: where Wi is the appropriate weather variable for a particular year, i. ~o and ~ 1 are unknown parameters estimated by numerical maximum likelihood methods to predict annual survival Si. Models tested with program SURVIV ranged from the simplest 2-parameter model, f S ,which includes 1 recovery rate and 1 survival rate for both age classes over all years (with ~ = 0), to the most general, ftaSwa, which allows 34 parameters, including year-specific recovery rates for subadults and adults, and weather-related (year-specific, i.e., ~ 1 ::f 0) survival rates for sub adults and adults (Table 2-2). Four input files were created to test weather effects (spring precipitation, winter precipitation, spring temperature,· and winter temperature) on survival of males. Weather covariates were not combined; each null hypothesis was tested separately. Each of the resulting 4 input files included 16 different models. Program SURVIV generated the log-likelihood function value, degrees of freedom (d.f.), Akaike's Information Criterion (AlC) (Sakamoto et al. 1986), and Chi-square goodness-of-fit significance probability for each model tested. The smaller the AlC value (equal to -2 • (log-likelihood) + 2 • (number of estimated parameters», the better the fit of that model to the observed data. In addition, program SURVIV provided likelihood ratio test results of all pairs of models that do not have equal degrees of freedom. Because many models were not nested, likelihood ratio tests were inappropriate and the AlC was used to select the most parsimonious model for the data. 2.5 RESULTS The model consistently selected in all tests incorporating weather covariates for the male recovery data was f as a. This model assumes year-constant but age-specific recovery rates and yearconstant but age-specific survival rates (Table 2-3). . Likelihood ratio tests did not reject the reduced sub model f as a when compared with four models incorporating weather covariates: spring precipitation (P = 0.695), winter precipitation (P = 0.787), spring temperature (P = 0.439), and winter temperature (P = 0.194). Therefore, hypotheses HoS1?'How." Host>and Howt were not rejected and incorporation of weather covariates did not sigriificaIitly improve fit of the previously selected model (f_as_a) to the recovery data. . 2.6 DISCUSSION 2.6.1 Banding Analysis Sample Sizes Age-specific differences were not detected between subadult and adult females; because this is not likely true in wild populations, further investigation is warranted. Likely, data were too sparse (i.e., number of female birds banded) to detect age-specific survival; in that case, a Type II error was made in the conclusion (i.e., a false null hypothesis was not rejected) and age-specific survival rates actually exist. Especially with sparse data sets, danger lies in accepting conditions of the null hypothesis simply because it was not rejected with the data on hand (White 1983). In the following: section, I quantify sample sizes necessary to detect differences in survival. I used Program BAND2 Wilson et al. 1989) to determine required banded sample sizes of males and females, based on specified levels of precision. A measure of dispersion of the data (i.e., 12 �53 precision) is provided by the coefficient of variation (CV), which is the standard deviation expressed as a proportion or percentage of the mean of the sample. The CV is a dimensionless measure and provides a useful descriptive index of the variability relative to the mean. Precision of estimates may be specified with desired CVs of annual and mean annual survival rate estimates. Required banding sample size was calculated based on survival and recovery rates from previous years of the study. That is, program BAND2 produced banding sample sizes necessary to achieve a certain level of precision (CV) while maintaining apparent variation in survival seen previously. In years of severe weather conditions, an annual difference of survival of 0.2 may be expected; certainly survival estimates from program BROWNIE vary much more than that, although these estimates include substantial sampling error. In order for an annual survival rate difference of 0.2 to be considered significant, a 10% CV of annual survival would be adequate. Program BAND2 calculations showed an average of 2,362 male and 2,609 female adult birds would have to be banded each year, for a 15-year study with specified CV of 10% of annual survival. Naturally, higher specified coefficients of variation yielded lower required numbers of banded birds. For example, with a 20% CV for annual survival, a significant difference of 0.4 in survival could be detected; for example, between 0.3 (95% CI = 0.18 0.42) and 0.7 (95% CI = 0.42 - 0.98). In that scenario, 590 and 652 males and females, respectively, of each age class would have to be banded each year (Figure 2-1). Further analysis demonstrated that specified Cvs of 45% for males and 60% for females roughly coincided with actual sample sizes banded. That is, a specified 45% CV for males would require 116 bands per year, and actual average number banded was 121. Likewise, a specified 60% CV for females would require 73 bands per year, and the actual average number banded was 57. In this manner, statistical power was determined indirectly, and it is clear that ability to detect assumed significant differences in annual survival with given sample sizes is very low. These sample sizes are inadequate to detect even 0.9 differences in survival rate with 90% confidence. 2.6.2 Sample Sizes for Detection of Weather Effects Failure of any weather variables to explain observed variation in survival rates over years suggests that weather variables tested were not particularly important in explaining the variation present. However, statistical power of my procedures to detect significant relationships may be low, and I cannot conclude that weather is not an important factor without evaluating power of statistical tests. To evaluate power of the procedures used, I simulated a reduced case with program SURVIV. Instead of 17 years of data collection, I simulated 5 years with parameters in Table 2-4 to make the problem more tractable. Sample sizes are representative of what could potentially be accomplished in North Park, but recovery rates were about 3 times greater than recovery rates realized in North Park. Adult and subadult survival functions were created to be strongly related to winter precipitation (Figure 2-2), and are in the same range as observed survival rates. The range of winter precipitation was about the observed range of winter precipitation over the 17-year period of this study. Statistical power was > 90% for IX = 0.05 (Table 2-5) to reject the model of constant survival across years with weather as a covariate explaining all variation in survival. Had sample sizes used in this study been at least 300 birds per year per age and sex class (to account for recovery rates 31imes .those observed), procedures used in this analysis would have detected the effect. However, these analyses assume that weather explains 100% of annual variation in survival, and that adequate data are taken to reject the null hypothesis of constant survival across years. Thus, adequate sample sizes required are at least as large as estimates presented above, i.e., - 600 birds banded per year for each age- and sex-class. I cannot surmise what percentage of annual variation in survival is explained by weather effects, but suspect it is not high based on the limited results of this study. Note that a second approach to improving power of the present study is to increase annual band recovery rates. Some increase might be expected by more intensive and extensive hunter check .station activity. However, the most likely approach to achieve large increases would be to use bands offering significant monetary rewards. Although notable effort and expense were incurred to obtain the data analyzed in this thesis, number of birds banded was inadequate to estimate annual survival rates with CVs of even 20%. Realized Cvs were >40%. To detect reasonable variation in annual survival rates of 0.4, a CV of at least 20% would be required, with sample sizes in the neighborhood of 600 birds per year in each age13 �54 and sex-class. The amount of information contained in present recovery matrices is inadequate to allow detection of even large differences in survival (> 0.8) predicted by weather variables. 2.7 LITERATURE CITED Autenrieth, R, W. Molini, and C. E. Braun. 1982. Sage grouse management practices. West. States Sage Grouse Comm. Tech. Bull. 1. 42pp .. Bartmann, R M., and D. C. Bowden. 1984. Predicting mule deer mortality from weather data in Colorado. Wildl. Soc. Bull. 12:246-248. Beck, T. D. 1. 1977. Sage grouse flock characteristics and habitat selection in winter. J. Wildl. Manage. 41:18-26. Brownie, C., D. R. Anderson, K P. Burnham, and D. S. Robson. 1985. Statistical inference from band recovery data - a handbook. Second ed. U.S. Dep. Inter., Fish WIldl. Servo Resour. Publ. 156. 305pp. Emmons, S. R., and C. E. Braun. 1984. Lek attendance of male sage grouse. J. Wildl. Manage. 48:1023-1028. Giesen, K M., T. J. Schoenberg, and C. E. Braun. 1982. Methods for trapping sage grouse in Colorado. WIldl. Soc. Bull. 10:224-231. Larsen, J. A, and J. F. Lahey. 1958. Influence of weather upon a ruffed grouse population. J. Wildl. Manage. 22:63-70. Martinson, R K, and C. R Grondahl. 1966. Weather and pheasant populations in southwestern North Dakota. J. Wildl. Manage. 30:74-81. North, P. M., and B. J. T. Morgan. 1979. Modelling heron survival using weather data. Biometrics 35:667-681. Sakamoto, Y., M. Ishiguro, and G. Kitagawa. 1986. Akaike information criterion statistics. KTK Scientific Publ., Dordrecht, Boston. 290pp. U. S. Dep. Commerce. 1973-88. Climatological data: annual summary, Colorado. Nat. Ocean. Atmosph. Admin. White, G. C. 1983. Numerical estimation of survival rates from band recovery and biotelemetry data. J. Wildl. Manage. 47:716-728. WIlson, K R, J. D. Nichols, and J. E. Hines. 1989. A computer program for sample size computations for banding studies. U.S. Dep. Inter., Fish WIldl. Tech. Rep. 23. 19pp. 14 �Table 2-1. Contingency table chi-square test for annual survival and/or recovery rate differences due to sex of adult sage grouse, North Park, Colorado. Year 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Em Matrix 1 Table (I d.f.) 13 86 60 8 0.068 x: Table 7 5 Matrix :; (1 d.f.) 6 0.151 3 x: 14 3 74 24 0.378 12 3 8 3 0.189 18 14 135 54 2.960 16 6 10 11 2.833 21 9 93 65 1.311 21 5 10 15 8.887 20 27 103 106 0.696 18 23 12 19 0.196 22 1 76 21 3.717 15 6 19 14 1.056 29 7 117 45 1.056. 26 12 22 9 0.052 18 16 155 78 2.400 25 H 15 14 2.131 24 7 166 31 0.903 24 11 15 10 0.471 29 10 161 59 0.023 29 8 15 12 3.784 18 3 139 29 0.118 24 6 9 9 4.713 14 12 78 99 0.875 13 9 10 '12 0.820 11 3 77 25 0.064 18 8 3 7 4.573 9 2 42 28 1.941 10 6 2 3 0.788 10 75 Total chi-square = 48.58 (29 d,.f.) 23 6 1.428 p(:x- > 48.58) = 0.013 !QWSQf mt!1~~ mal~ !2and dam; ~ecQnd !QWSm femal~ data, 15 �56 Table 2-2. Key to names of models used in tests of weather effects on survival. Parameters are recovery (f) and survival rates (S), where subscripts denote sources of variation and underscores denote no effect. Recovery component f_ Model assumptions is year-constant and age-constant Parameter! s) f f_a is year-constant but age-specific fy, fA ft_ is year-specific but age-constant f; fta is year-specific and age-specific f;y, fiA Survival component S Model assumptions is not weather-related and is age-constant {3o, {31' = 0 S_a is not weather-related but is age-specific {30Y, {30A, {31 = 0 Sw_ is weather-related but is age-constant {3o, {31 Swa is both weather-related and age-specific . {30Y, {31Y if: 0, {30A, {31A :1= 0 - Parameter(s) if: 0 Note: Model names include 2 components, recovery and survival. As an example, consider model oft 8 a." The recovery component is"ft ," while the survival component is "8_8.'- Model "ft_8_a" includes assumptions that recovery rate is year-specific (per "t" of recovery component) but is constant across age classes (per"" of recovery component), Model"ft 8 a" also includes the assumptionS that survival rate is constant across and hence, is not weather-related (per "_" of survival component) but is agespecific (per "a" of survival component), YM 16 �) Table 2-3. i Goodness-of-fit and likelihood ratio tests for models of weather effects on male sage grouse survival in North Park, Colorado, 1973-87. Weather variables are spring (5), winter (w), precipitation (p), and temperature (t) (see text). Best-fitting model by AlC is in bold. Wc:ather LogName Var. likelihood f S -250.87395 -247.79054 f S a f_Sw_ sp -250.72855 wp -250.87389 st -250.31725 wt -249.1-8433 f_Swa sp -247.43260 wp -247.56206 st -246.94083 wt -245.02603 -250.02764 faS -246.66738 raSa f_aSw_ sp -249.87013 wp -250.02752 st -249.46941 wt -248.28447 f_aSwa sp -246.30321 wp -246.42719 st -245.84324 wt -245.02603 ft S -240.20251 ft_S_a -237.02162 ft_Sw_ sp -240.16916 wp -240.12256 st -239.37464 wt -239.21819 ft_Swa sp -236.75421 wp -236.77069 st -235.96515 wt -236.04497 ftaS -236.27840 ftaS_a -232.88220 ftaSw sp -236.07733 wp -236.26849 st * wt * ••• ftaSwa sp wp -232.60252 ••• st ••• wt Pr(Lgr. Total Pooled Pearson d.f. d.f. ATC Chi-Sg. Chi-Sg.) 0.0097 505.74791 84 238 116.468 0.0210 501.58107 78 104.875 237 0.0090 507.45711 83 115.683 237 0.0078 507.74778 116.454 237 83 0.0087 506.63450 114.692 82 237 0.0154 504.36865 81 110.237 237 0.0163 106.406 504.86520 78 235 0.0162 104.070 505.12413 76 235 0.0244 105.121 503.88166 79 235 234 502.05205 rt 100.651 0.0345 0.0117 114.198 237 506.05527 83 0.0277 102.024 77 236 501.33475 0.0110 113.378 236 507.74027 82 0.0095 114.179 236 508.05504 82 0.0127 506.93882 112.531 236 82 0.0190 504.56894 107.817 236 80 0.0204 504.60643 102.376 234 76 0.0216 234 504.85437 75 101.221 100.599 0.0348 234 503.68647 77 0.0345 100.651 234 502.05205 77 0.0304 224 512.40501 66 88.712 0.0727 508.04323 82.096 223 65 88.873 0.0360 67 223 514.33832 0.0306 88.688 514.24512 66 223 0.0409 88.080 223 512.74929 67 0.0279 64 86.887 512.43638 223 80.848 0.0623 511.50842 63 221 0.0580 81.311 221 ·511.54137 63 80.956 0.0859 509.93031 65 221 . 0.0612 80.969 221 510.08993 63 0.0053 209 534.55681 83.018 53· 0.0106 208 529.76439 ·50 75.8420.0035 208 536.15465 51 82.373 0.0032 208 536.53696 82.857 51 (* indicates a failure to optimize, usually the result of trying to fit too many parameters to the given data) 0.0077 206 533.20505 48 74.968 (failure to optimize) (failure to optimize) 17 �58 Table 2-4. Input to simulations of models testing weather effects on annual survival of sage grouse. Annual survival was predicted as log(S/(l-Si)] = 1.5-0.25(winter precipitation) for adults and log(S/(1-Si)] =2.0-0.50(winter precipitation) for subadults. Number Winter Preci12imtiQn Ygr B!Ylded Adult R~v!m! Yearling RecQveo: 1983 100 0.15 0.17 2.00 1984 100 0.17 0.19 4.00 1985 100 0.19 0.21 6.00 1986 100 0.21 0.23 8.00 1987 100 0.23 0.25 - Table 2-5. Results of simulations of models testing weather effects on annual survival of sage grouse. Values shown are percentage rejection (N = 1,000) of reduced model by likelihood ratio test. Simulations were conducted with program SURVIV. A key to model names is shown in Table 2-2 and input parameters for this simulation are shown in Table 2-4, Reduced General rnodel : model i:s.: i:»: . f_S_ s:»: f_aS_a f_aS_a ft_Sw_ ftaS_a f_aS_ ft_Sw_ ftaS_a ftaSwa ftaS_a ftaSwa ftaSwa ftaSwa J.&vel .<0.01 or si~ifignce <0.05 <0.10 50 93 54 97 32 94 56 96 62 97 69 99 43 97 66 98 23 81 30 89 13 81 31 90 18 <0.50 91 100 96 100 87 100 . 94 100 < 1.00 100 100 100 100 100 100 100 100 �3,000 ~------------------, 2,500 "0 uuu. .uuuuuJF~ . 2,000 .----------.--~ = \ tU ~ ~ Lu u u u u uu uu .\. __ __ . - ~ S 1,500 ................ ~ . IZI "0 •.. is 1,000 500 10 20 30 40 CV( annual survival) 50 60 Figure 2-1. Average required- banding sample size per year of male and female sage grouse in North Park, Colorado, based on desired coefficient of variation (CV) of annual survival rate estimates. ~~------------------------------------~~-, " 0..6 -.-- .. - ".•..•. -.--- ..•.. -~, ........................................................•........... .•..•. '-.•..•..•..•. .•..•.... .•..•..•. •••••••••••••••••.••.•• __ .•..•. .•. ••••••••••••••••••••.•• :'..., ••••••.••••••••••••••• " " "" .r •• _ ....• \\ . 0.2 ................ -_ _ ~, ~~"',~ _ . "''''--"•..•.•.•.. o~------~------~------~~------~------~ 10 o 2 4 6 8 Wmter Precipitation Figure 2-2. Simulated survival estimates fOr sub adults and adults as a function of winter precipitation. All variation in annual survival rates is explained by winter precipitation. 19 �60 CHAPTER 3: EVALUATION OF EFFECT OF AGGREGATED RECOVER.IES ON SAGE GROUSE SURVIVAL ESTIMATION 3.1 INTRODUCTION There are important assumptions involved in making inferences from banding data.. Assumptions relating to study planning, field procedures, and type of species are: 1) sample is representative of target population, 2) age and sex of individuals are correctly determined, 3) there is no band loss, 4) survival rates are not affected by banding or tagging itself, and 5) year (hunting season) of band recoveries is correctly tabulated (Brownie et al. 1985). Additionally, when specific hypotheses about survival and recovery rates are made, it is assumed that fate of each banded animal is independent of (not correlated with) fates of other banded individuals (Brownie et al. 1985). Several published papers examine robustness of estimators to partial failure of particular assumptions (Anderson and Burnham 1980, Krementz et al. 1989), Nelson et al. 1980, Nichols et al. 1982, Pollock. and Raveling 1982, Smith and Anderson 1987). The assumption of independent fates is probably violated in many practical applications of bird-banding models (Pollock and Raveling 1982). Birds are not independent entities in terms of survival or other characteristics (Sulzbach and Cooke 1978). Pollock and Raveling (1982) stated that failure of this assumption will not bias any estimators, but will mean that true sampling variances are larger than those given by the statistical models. Thus, any calculated confidence intervals will be narrower than they should be. A population composed of independent pairs of birds that behave as though they are a single individual would mean that a sample of N individuals from this population is effectively only one-half of N. Lack of fate independence is even more apparent in upland birds such as sage grouse tCentrocercus urophasianus) than in migratory waterfowl, as sage grouse exhibit fidelity to natallek sites. Probably over half of all sub adult grouse attend their natal area lek (Dunn and Braun 1985), suggesting that sage grouse may be philopatric. Following natal dispersal, birds are generally faithful to their initial breeding area in successive years (Dunn and Braun 1985). Almost all matings are performed by "adult males that return during successive years to the same lek and territory within the lek (Wlley 1978). Beck (1977) concluded that distribution of sage grouse in North Park during winter "was primarily a reflection of availability of sagebrush above snow, slope, and aspect Likelihood of aggregation of sage grouse recoveries is increased when hunters travel in groups, and recover multiple bands from a single flock. The objective of this study was to quantify effects of non-independent fates of banded sage grouse on survival and recovery estimate bias and confidence interval (CD coverage for different sample sizes of sage grouse banded, average group size recovered, and number of years of banding. This study examined robustness of a survival model for sage grouse in North Park, Colorado. 3.2 METHODS 3.2.1 Simulations I assessed accuracy, precision, and bias of recovery and survival rate estimates through comparison of known (observed) and simulated recovery and survival rates. I generated several simulated recovery matrices from a range of sample sizes banded, group sizes, and number of years of banding and recovery. The Statistical Analysis System (SAS) (SAS Institute Inc. 1988) was used for Monte Carlo simulations. Banded sample sizes of 200, 300, 400, 500, 600, and 700 per year were used with average recovered group sizes of 2, 3, and 4, for 5 and 10 years of both banding and recovery. Recovery and survival rates were assumed to be equal and constant for the whole population, following Model 3 of Brownie et al. (1985). Model 3 assumes constant survival and recovery rates across years and recovery and survival rates were assigned values of 0.0955 and 0.3839, respectively. These assumed time-constant values were average adult male survival and recovery rate estimates for North Park sage grouse generated by Program BROWNIE (Brownie et al. 1985) and were thus used as reasonable parameter values for the simplest model. 20 �61 Each simulated scenario consisted of one of 6 average annual banding sample sizes, one of 3 average recovery group sizes, and either 5 or 10 years of banding and recovery, resulting in 36 simulated recovery scenarios. Each simulated recovery of a banded bird was generated as a random result from a multinomial distribution, then assigned a random group size from a Poisson distribution with lambda equal to mean group size (2, 3, or 4). To keep possible variation due to sampling small, each of the 36 scenarios was simulated 1,000 times. 3.2.2 Analysis of Simulations The mean of estimates of survival or recovery rates for a given scenario is an approximation of the expected value of the estimator of survival or recovery rate. I used mean estimates of survival and recovery rates for each scenario to evaluate the effect of aggregated recoveries. Coverage was defined as percentage of the estimates' 95% confidence intervals covering the true parameter (survival or recovery rate). That coverage percentage was calculated for each simulated scenario, with each survival and recovery rate estimate classified as occurring within or outside the 95% confidence interval. Coverage was used as an indicator of accuracy of survival and recovery rate estimates. Percent relative bias (PRB) was used to evaluate bias of the estimates of recovery and survival rates: (mean of estimates - true parameter value) PRB = -------------------x 100. true parameter value Student's t value was used to test the hypothesis that bias equals zero. 3.3 RESULTS In this study, accuracy is defined as the closeness of a computed value to its true value (Sokal and Rohlf 1981). In all 36 simulated scenarios, coverage of 95% confidence intervals was well below 95%. With 5 years of banding and recovery, coverage of survival estimates' confidence intervals ranged from 59.0% for group size of 4 and 300 birds banded per year, to 82.1% for group size of 2 and 700 birds banded per year (Table 3-1). As indicated by CI coverage, accuracy of survival and recovery estimates declined with increasing group size for both scenarios of 5 and 10 years of recovery data (Tables 3-1 and 3-2). Maximum likelihood analysis of variance of survival rate confidence interval coverage revealed that group size was a significant factor with respect to coverage (P < 0.001), and number of birds banded per year was significant (P = 0.003) (Table 3-3). Number of years of banding and recovery data was not significant. Source term interactions also were not significant; however, the interaction of number of birds banded and average group size recovered (Band" Group) might be generally useful (P = 0.064). Maximum likelihood analysis of variance of recovery rate confidence interval coverage uncovered four significant sources of variance: number of birds banded (P < 0.001), the interaction of number of years of data and number of birds banded (Year·Band; P = 0.008), average group size recovered (P < 0.0001), and the interaction of number of birds banded and average group size recovered (Band" Group; P = 0.011) (Table 3-4). Intercept terms aside, average group size recovered was the primary source of variance for both survival and recovery rate estimate confidence interval coverage (Tables 3-3 and 3-4). Precision is the closeness of repeated measurements of the same quantity (Sokal and Rohlf 1981). In this study, standard deviation of the empirical estimates (Tables 3-5 and 3-6) was greater than the mean of the standard errors (Tables 3-1 and 3-2), calculated analytically. As an example, averaged over number of birds banded and number of years of data, survival estimates generated with group size of 4 had a mean standard error of estimates of 2.57 (from Tables 3-1 and 3-2), while standard deviation of the estimates was 14.70 (from Tables 3-5 and 3-6). This difference is usually the result of violating a theoretical assumption regarding independence. Percent relative bias (PRB) was overall significantly different than zero for survival rate estimates, for all levels of average grouped recoveries examined, and for both 5 years (P < 0.001, Table 3-5), and 10 years (P < 0.001, Table 3-6) of banding and recovery data. PRB also was significantly different than zero for recovery rate estimates, for all levels of average grouped recoveries examined with 5 years of data (P < 0.001, Table 3-5), but not significantly different than zero with 10 years of data (P = 0.156, Table 3-6). 21 �62 Analysis of variance for PRB of survival rate estimates revealed three significant sources of variance: number of years of banding and recovery (P = 0.004), number of birds banded annually (P = 0.003), and the interaction of those parameters (Year'tBand; P = 0.006) (Table 3-7). While average group size recovered was not statistically significant, it might have affected bias (P = 0.065). Analysis of variance for PRB of recovery rate estimates uncovered only one significant source of variance, number of years of banding and recovery (P = 0.004; Table 3-8). 3.4 DISCUSSION Accuracy of survival and recovery estimates, as measured by coverage of respective 95% confidence intervals, did not approach the 95% level. This was true for all 36 simulated scenarios, and while recovery aggregation was responsible for a significant portion of the variance, lack of coverage was also affected by other variables. Interestingly, number of years of banding and recovery was not significant in coverage of either survival or recovery estimates. It is possible that analysis of additional scenarios of even longer banding periods may have yielded different results; however, it is not likely, as doubling the number of years of the study (5 versus 10) did not result in appreciably different levels of coverage. Another factor affecting accuracy may have been the relatively low constant survival rate (38.39%) obtained from the simple Model 3 of Brownie et al. (1985) and used in the simulations (see Chapter 1 for background of model selection). The effect of violating the assumption regarding independence is seen clearly in the large differences between the estimates' empirical standard deviations and average of standard errors, as computed by Program ESTIMATE. Precision of survival and recovery estimates was greatly affected, as empirical standard deviations and mean theoretical standard errors should have been equivalent. Instead, empirical standard deviations were far greater than mean standard errors. In other words, repeated measurements- of the same quantity varied widely. This effect on precision was not alleviated with the additional number of years of data (5 versus 10). The problem of excess variance has been studied by many statisticians and quantitative ecologists. Burnham et al. (1987) and Lebreton et al. (1992) recommended use of some type of empirical variance if there is substantial excess variation, with the simplest approach being estimation of a variance inflation factor (Burnham et al. 1987:256246 and Lebreton et 31. 1992:84-85). With the sage grouse data set simulated in this study, however, the question remaining is whether the model structure used was correct. Values used for male sage grouse survival and recovery rates were from one of the simplest models detailed in Brownie et al. (1985); certainly, it is questionable that male sage grouse survive at a rate of only 38% per year. More comprehensive approaches exist for testing effects of assumption violation. The approach utilized in this study, however, was a simpler one of selecting a single model structure and setting parameter values based on observed data and Program ESTIMATE calculations, rather than protracted testing of an array of hypothesized survival and recovery parameter values. Although bias 'was affected, this was not as great a concern as was effect on precision. _Average group size recovered was not a significant factor in either PRB of survival or recovery estimates. However, an important question follows naturally: how can bias in real studies be evaluated? While the source terms were defined in this study, source terms, and of course, actual parameter values, in a real data set remain unknown. In this study, hypotheses were made about what might be occurring under natural conditions, and those hypotheses were the basis for testing effects on somewhat realistic survival and recovery estimates. This approach does not limit ability to evaluate effects of assumption violation, but necessitates caution in any conclusions about real-world processes. The relatively small size of the sage grouse population in North Park, and nature of sage grouse to gather on leks in spring and flush in flocks when hunted in autumn, makes it difficult to meet the assumption of independent fates of individuals. This difficulty is not limited to sage grouse studies; with manywildlife species, strict conformity with the banding analysis requirement of independent fates of banded individuals is very difficult to achieve in the field. Likely this assumption is not often met, for several possible reasons in addition to grouping of birds and grouping of hunters (the major source of band recovery reports). If, however, a significant cause of assumption violation can be uncovered, that effect can be quantified and considered when survival and recovery estimates are calculated and interpreted. 22 �63 Finally, the extent of recovery aggregation in sage grouse populations was not determined in this study. However, the implication of the results is that, if aggregation is occurring even at levels lower than those simulated here, effects on survival and recovery estimate precision may be severe. 3.5 IlTERATURE CITED Anderson, D. R, and K. P. Burnham. 1980. Effect of delayed reporting of band recoveries on survival estimates. J. Field Ornithol. 51:244-247. Beck, T. D. I. 1977. Sage grouse flock characteristics and habitat selection in winter. J. Wildl. Manage. 41:18-26. Brownie, C., D. R Anderson, K. P. Burnham, and D. S. Robson. 1985. Statistical inference from band recovery data - a handbook. U.S. Dep. Int, Fish Wildl. Serv., Resour. Publ. 156. 305pp. Burnham, K. P., D. R Anderson, G. C. White, C. Brownie, and K. H. Pollock. 1987. Design and analysis methods for:fish survival experiments based on release-recapture. Am. Fish. Soc. Monogr. 5. 437pp. Dunn, P.O., and C. E. Braun. 1985. Natal dispersal and lek fidelity of sage grouse. Auk 102:621-627. Krementz, D. G., J. D. Nichols, and J. E. Hines. 1989. Post-fledging survival of European starlings. Ecology 70:646-655. Lebreton, J. D., K. P. Burnham, J. Clobert, and D. R. Anderson. 1992. Modeling survival and testing biological hypotheses using marked animals: a unified approach with case studies. Ecol. Monogr.62:67-118. Nelson, L. J., D. R Anderson, and K. P. Burnham. 1980. The effect of band loss on estimates of annual survivals. J. Field Ornithol. 51:30-38. Nichols, J. D., S. L. Stokes, J. E. Hines, and M. J. Conroy. 1982. Additional comments on the assumptions of homogeneous survival rates in modern bird banding estimation models. J. Manage. 46:953-962. Pollock, K. H., and D. G. Raveling. 1982. Assumptions of modern band-recovery models, with emphasis on heterogeneous survival rates. J. Wildl. Manage. 46:88-98. SAS Institute Inc. 1988. SAS procedures guide, release 6.03 edition. SAS Institute, Inc., Cary, N.C. 441pp. SAS Institute Inc. 1988. SAS/STAT user's guide; release 6.03 edition. SAS Institute Inc., Cary, N.C. 1029pp. Smith, D. R, and D. R Anderson. 1987. Effects of lengthy ringing periods on estimators of annual survival. Acta Ornithol. 23:69-77. Sokal, R. R, and F. J. Rohlf. 1981. Biometry: the principles and practice of statistics in biological research. Second ed. W. H. Freeman and Co., New York. 859pp. Sulzbach, D., and F. Cooke. 1978. Elements of nonrandomness in mass-captured samples of snow geese. J. Wildl. Manage. 42:437-441. U.S. Dep. Commerce. 1973-88. Climatological data: annual summary, Colorado. Nat. Ocean. Atmosph. Admin. 78-93. Wiley, R H. 1978. The lek mating system of the sage grouse. Sci. Amer. 238:114-125. wun. 23 �64 Table 3-1. - - Survival and recovery estimates (S and f), mean standard errors, and coverage of 95% confidence interval, based on estimates (%) generated from 1,000 simulated recovery matrices with 5 years of recovery data. Mean Mean SE S f Coverage S f Group Size Number Banded S f 2 200 37.81 9.62 4.63 0.89 76.9 76.2 2 300 38.11 9.56 3.79 0.73 77.3 78.8 2 400 38.37 9.59 3.29 0.63 80.5 77.4 2 500 38.49 9.54 2.95 0.56 80.8 78.7 2 600 38.47 9.54 2.69 0.51 78.4 77.8 2 700 38.20 9.60 2.48 0.48 82.1 78.5 38.24 9.57 3.31 0.63 79.3 77.9 66.9 68.2 Group Mean 3 200 37.51 9.66 4.62 3 300 38.12 9.60 3.79 0.73 69.1· 68.0 . 3 400 37.96 9.59 3.28 0.63 70.7 71.7 3 500 38.10 9.55 2.94 0.56 68.9 69.1 3 600 38.18 9.62 2.68 0.52 68.8 69.4 3 700 38.33 9.54 2.49 0.48 71.1 66.3 38.03 9.59 3.30 0.63 69.3 68.8 .0.89 Group Mean 4 200 37.80 9.59 4.67 0.89 62.7 55.2 4 300 37.58 9.67 3.77 0.73 59.0 63.0 4 400 38.28 9.59 3.29 0.63 62.5 62.7 4 500 38.42 9.58 2.94 0.56 65.7 62.0 4 600 38.27 9.61 2.68 0.52 63.1 61.5 4 700 37.95 9.63 2.48 0.48 64.2 61.4 38.05 9.61 3.30 0.63 62.9 61.0 Group Mean 24 �65 Table 3-2. - - Survival and recovery estimates (S and 0, mean standard errors, and coverage of 95% confidence interval, based on estimates (%) generated from 1,000 simulated recovery matriees with 10 years of recovery data. Group Size Number Banded S Me:m 2 200 38.07 2 300 2 Mean SE Coverage f S f 9.58 2.58 0.62 76.8 76.4 38.31 9.57 2.11 0.50 78.7 77.1 400 38.40 9.55 1.82 0.43 78.9 76.1 2 500 38.41 9.53 1.63 0.39 78.7 79.2 2 600 38.38 9.54 1.49 0.35 77.1 77.6 2 700 38.32 9.55 1.38 0.33 80.2 80.6 38.32 9.55 1.84 0.44 78.4 77.8 Group Mean f S 3 200 38.34 9.51 2.59 0.61 67.8 64.7 3 300 38.16 9.61 2.10 0.50 69.3 64.1 3 400 38.15 1.83 0.43 69.9 66.4 3 500 38.16 9.56 1.63 0.39 67.3 67.0 3 600 38.36 9.55 1.49 0.35 70.3 69.9 3 700 38.27 9.55 1.38 0.33 69.3 69.2 38.24 9.55 1.84 0.44 69.0 ·66.9 Group Mean ·9.54 4 200 38.56 9.53 2.58 0.61 61.4 54.2 4 300 37.98 9.57 2.11 0.50 60.9 60.8 4 400 38.22 9.59 1.82 0.43 62.3 61.5 4 500 38.19 9.58 1.63 0.39 63.9 62.4 4 600 38.30 9.61 1.49 0.36 65.0 62.6 4 700 38.19 9.56 1.38 0.33 61.3 65.5 38.24 9.57 1.83 0.44 62..5 61.2 Group Mean 25 �66 Table 3-3. Maximum likelihood analysis of variance for survival rate estimate 95% confidence interval coverage. Source terms are number of years of banding and recovery (Year), annual banded sample size (Band), average recovered group size (Group), and interactions. Source of Variance d.f. Intercept. Year Table 3-4. Chi-square P-value 5594.65 <0.0001 1.59 0.2077 Band 5 18.23 0.0027 Year*Band 5 6.66 0.2473 Group 2 750.24 <0.0001 Year=Group 2 0.69 0.7070 Band=Group 10 17.50 0.0640 year*Band'"Group 10 2.19 0.9935 Maximum likelihood analysis of variance for recovery rate estimate 95% confidence interval coverage. Source terms are number of years of banding and recovery (Year), annual banded sample size (Band), average recovered group size (Group), and . interactions. Source of Variance Intercept d.f. Chi-square P-value 1 4955.93 <0.0001 Year 1.29 0.2553· Band 5 33.79 <0.0001 year*Band 5 15.77 0.0075 Group 2 782.86 <0.0001 Year*Group 2 3.58 0.1674 Band*Group 10 22.95 0.0109 Year*Band*Group 10 2.22 0.9944 26 �6;: Survival and recovery estimator bias, based on estimates simulated recovery matrices with five years of recovery. Table 3-5. (%) generated from 1,000 N PRB f r» It I f s S S r» It I 37.81 9.62 -1.51 19.70 0.0157 0.70 15.24 0.1443 300 38.11 9.56 -0.73 15.87 0.1473 0.15 11.97 0.6865 2 400 38.37 9.59 -0.04 13.08 0.9144 0.45 10.50 0.1735 2 500 38.49 9.54 0.25 11.38 0.4864 -0.15 9.25 0.5962 2 600 38.47 9.54 0.20 10.92 0.5682 -0.12 8.67 0.6554 2 700 38.20 9.60 -0.50 9.59 0.1026 0.51 7.97 0.0435 38.24 9.57 -0.39 Group Banded Size per yr 2 200 2 PRB S SD PRB PRB SD f f Group Mean 0.26 3 200 37.51 9.66 -2.30 23.97 0.0025 1.17 17.94 0.0388 3 300 38.12 9.60 -0.70 . 18.68 0.2442 0.49 14.81 0.2993 3 400 37.96 9.59 -1.13 15.73 0.0235 0.40 12.79 0.3177 3 500 38.10 9.55 -0.74 15.II 0.1199 -0.01 11.64 0.9864 3 600 38.18 9.62 -0.55 13.38 0.1905 0.70 10.29 0.0327 3 700 -38.33 9.54 -0.16 12.19 0.6811 -0.09 9.73 0.7656 38.03 9.59 -0.93 Group Mean 0.44 4 200 37.80 9.59 -1.54 26.24 0.0638 0.42 22.82 0.5618 4 300 37.58 9.67 -2.11 22.22 0.0027 1.28 16.50 0.0142 4 400 38.28 9.59 -0.29' 18.96 0.6333 0.38 14.76 0.4149 4 500 38.42 9.58 0.07 16.52 0.9003 0.34 13.27 0.4119 4 600 38.27 9.61 -0.32 15.16 0.5079 0.58 11.77 0.1201 4 ·700 37.95 9.63 -1.14 13.97 0.0097 0.86 11.09 0.0147 38.05 9.61 -0.89 0.64 -0.74 0.45 Group Mean Mean 0.0001 t-test 27 0.0001 �68 Table 3-6. Survival and recovery estimator bias, based on estimates (%) generated from 1,000 simulated recovery matrices with ten years of recovery. Ii PRB SO PRB S PRB SO f s s f PRB f r» It I f r» It I 38.07 9.58 -0.84 11.03 0.0163 0.33 10.89 0.3362 300 38.31 9.57 -0.21 8.48 0.4307 0.22 8.61 0.4248 2 400 38.40 9.55 0.03 7.47 0.8840 -0.05 7.37 0.8325 2 500 38.41 9.53 0.05 6.67 0.8205 -0.21 6.45 0.3010 2 600 38.38 9.54 -0.02 6.29 0.9119 -0.09 6.11 0.6441 2 700 38.32 9.55 -0.18 5.57 0.2965 0.03 5.36 0.8539 38.32 9.55 -0.20 Group Banded Size per yr 2 200 2 S Group Mean 0.03 3 200 38.34 9.51 -0.12 12.81 0.7606 -0.46 13.16 0.2707 3 300 38.16 9.61 -0.61 10.46 0.066Z- 0.61 10.91 0.0779 3 400 38.15 9.54 -0.63 9.15 0.0302 -0.08 9.13 0.7749 3 500 38.16 9.56 -0.60 8.10 0.0204 0.10 8.18 0.7037 3 600 38.36 9.55 -0.07 7.39 0.7553 0.03 7.38 0.8897 3 700 38.27 9.55 -0.31 . 7.01 0.1647 0.04 6.59 0.8392 38.24 9.55 -0.39 Group Mean 0.04 4 200 38.56 9.53 0.45 15.09 0.3421 -0.17 16.37 0.7460 4 300 37.98 9.57 -1.06 12.31 0.0068 0.17 11.92 0.6533 4 400 38.22 9.59 -0.44 10.56 0.1904 0.38 10.45 0.2534 4 500 38.19 9.58 -0.52 8.99 0.0683 0.27 8.91 0.3322 4 600 38.30 9.61 -0.23 8.28 0.3795 0.64 8.20 0.0140 4 700 38.19 9.56 -0.52 8.06 0.0410 0.10 7.29 0.6517 38.24 9.57 -0.39 0.23 -0.32 0.10 Group Mean Mean o.oooi r-test 28 0.1557 �Table 3-7. Analysis of variance for percent relative bias of survival estimates (PRE S). Source terms are number of years of banding and recovery (Year), annual banded sample size (Band), average recovered group size (Group), and interactions. d.f. Sum of Squares Mean Square F-value Year 1 1528.49 1528.50 8.19 0.0042 Band 5 3391.92 678.38 3.63 0.0028 Year*Band 5 3048.57 609.71 3.27 0.0060 Group 2 1019.02 509.51 2.73 0.0653 Year*Group 2 219.26 109.63 0.59 0.5559 Band*Group 10 3023.98 302.40 1.62 0.0943 Year*Band*Group 10 1209.62 120.96 0.65 0.7736 Model 35 13440.87 384.02 2.06 0.0002 Source of Variance Table 3-8. - Analysis of variance for percent relative bias of recovery estimates (PRE f). Source terms are number of years of banding and recovery (Year), annual banded sample size (Band), average recovered group size (Group), and interactions. d.f. Sum of Squares Mean Square F-value Year 1 1088.80 1088.80 8.14 0.0043 Band 5 589.01 117.80 0.88 0.4926 Year*Band 5 615.40 123.08 0.92 0.4663 Group 2 543.05 271.52 2.03 0.1312 Year*Group 2 63.72 31.86 0.24 0.7880 Band*Group 10 995.04 99.50 0.74 0.6831 Year*Band*Group 10 1348.19 134.82 1.01 0.4332 Model 35 5243.22 149.81 1.12 0.2865 Source of Variance Prepared by F Pr> . "'IQ lI]O-tLu_r . 2.a.i.X0i)'v (, Co) Approved by Marilet A. Zablan Graduate Research Assistant ~~ Pr> F ~ Clait E. Braun ~ildlife Research Leader 29 ��71 JOB PROGRESS State of: Project: Work Plan: Job Title: Colorado Personnel: Upland W-167-R 3 Job: Grazing and Residual 01 January through 31 December Kenneth Bird Research 18 Evaluation of Livestock Sage Grouse Nest Success. Period Covered: Author: REPORT Herbaceous Cover on 1993. M. Giesen C. E. Braun, K. M. Giesen, Division of Wildlife. D. B. LaBelle, P. D. Rosales, Colorado ABSTRACT Six sage grouse (Centrocercus urophasianus) strutting grounds in North Park, Colorado (Boettcher Junction, Coalmont, Delaney Butte, Lost Creek, Raven, and Spring Creek #1) were selected for documentation of nesting sites and movements of hens to nests in 1993. Seventy-one hens were trapped and radiomarked on 4 strutting grounds (Boettcher - 14, Coalmont - 22, Delaney Butte - 20, Spring Creek #1 - 15); no hens were trapped on Raven and Lost Creek strutting grounds because access was delayed by persistent snow. Fortythree hens were followed to nests; 4 hens were depredated prior to nesting and 12 hens could not be located during the nesting season. No nests were located for another 12 hens. Movements from lek-of-capture to nest site ranged from 0.45 to 29.0 km with a mean movement of 3.16 ± 4.39 km (median 1.73 km). More than one-fourth of the hen~ moved farther than 3.0 km to nest. Hens captured at Spring Creek #1 strutting ground moved farther to nests (6.54 ± 9.25 km) than hens from Boettcher Junction (2.39 ± 1.95 km) and Delaney Butte (1.57 ± 0.89 km) strutting grounds. Hens marked at Coalmont strutting ground moved 3.64 ± 3.40 km; similar to hens marked at other locations. Nest success was 21.6% (including 8 renests) and hen success was 25.6%. Most nest loss was due to depredation, primarily by ground squirrels (Spermophilus richardsonii), and 6 nests were abandoned. Shrub height, primarily sagebrush (Artemisia spp.), at nests avexaged 54;6 ± 14.0 cm and visual obstruction was 65.0·± 15; 8%. Mean sagebruSh canopy cover and density at nests Were 35.3 ± 14.9% and 14,700 ± 6,300 plantsfha, respectively. There were no differences in habitats selected for nesting by hens from any of the leks. ��73 EVALUATION OF LIVESTOCK GRAZING AND RESIDUAL ON SAGE GROUSE NEST SUCCESS HERBACEOUS COVER Kenneth M. Giesen INTRODUCTION Populations of sage grouse in Colorado have declined since the early 1980's. This decline is reflected in the North Park, Colorado sage grouse population by low numbers of male sage grouse on strutting grounds in spring and reduced harvests in fall (C. E. Braun, pers. commun). During this population decline, nest success of hens and production and survival of chicks has been below the long-term average. Considerable evidence suggest nesting success of sage grouse is related to amount of residual herbaceous cover in nesting areas which affects depredation of nests (Gregg et al. 1994). The amount of residual herbaceous cover available to sage grouse in North Park is thought to be related to winter-spring precipitation and patterns of forage use by livestock (C. E. Braun, pers. commun.). The Bureau of Land Management reports that under optimum conditions, the best range sites in North Park produce only 894-1345 kilograms of forage per hectare, according to Soil Conservation Service range site descriptions. When late-winter and spring precipitation levels are below average, herbaceous forage production is less. The current pattern of livestock grazing in North Park may exacerbate the problem of lack of residual herbaceous cover needed for secure nesting cover by sage grouse in spring. Experiments to evaluate sage grouse nest success in relation to changes in residual herbaceous cover will provide information useful for evaluating and enhancing sage grouse nesting habitats in Colorado and elsewhere. These experiments could also result in recommendations to develop new grazing systems and range enhancements. P. N. OBJECTIVES The primary objective of this study is to experimentally evaluate the relationships between residual herbaceous nesting cover and nest fa~lure for sage grouse in North Park, Colorado. Distribution of sage grouse nesting habitats in relation to strutting grounds will be identified and nest success rates will be ascertained. SEGMENT OBJECTIVES 1. Conduct a literature review related to nesting -cover and prepare a detailed study plan. success and residual 2. Contact Bureau of Land Management (BLM) personnel, other affected land management agencies and ranchers in North Park, Colorado, and obtain approval to conduct the study with experimental grazing exclosures adjacent to 6 selected sage grouse strutting grounds. 3. Trap and radiomark 20 hens at each of 6 selected strutting grounds for identification of nesting habitat and documentation of nest success. �4. Monitor radio-marked hens to document 5. Measure residual cover at nest sites and within nesting habitats analysis of cover quality and effects on nest fate. 6. Prepare annual progress timing and causes of nest failure. for and other reports. STUDY AREA Six strutting grounds in North Park (Jackson County) were selected as sites for trapping and radiomarking hens based on criteria of access, consent of landowners or managers, lek size (number of males counted in 1992), and geographic distribution within North Park. These strutting grounds were Boettcher Junction and Delaney Butte in the northwest, Coalmont and Lost Creek in the southwest, Spring Creek 111 in the southeast, and Raven in the northeast. Topography, soils, vegetation, and climate of the North Park area have been summarized by Beck (1975), Emmons (1980), Petersen (1980), Schoenberg (1982), and Remington (1983). METHODS Trapping and radiomarking Spotlighting (Giesen et al. 1982) and walk-in funnel traps (Toepfer et al. 1988) were used to capture female sage grouse on or adjacent to selected strutting grounds in April and early May. Captured birds were placed in burlap bags and processed within 30 minutes. Each captured hen was weighed on an electronic balance and banded with a numbered aluminum band and red plastic bandettes. A Telemetry Systems solar or Holohil battery-powered transmitter was attached to hens using a poncho (Amstrup 1980) or necklace attachment. Transmitter packages weighed 12-18 gms with the Holohil transmitters being lighter than the Telemetry Systems transmitt~rs. Birds were released near the site of capture. Portable hand-held receivers and a 3-element Yagi antenna were used to monitor radio signals and locate hens once or twice weekly. Aerial tracking was used 4 times in May-June in an attempt to locate hens whose radio signals could not be detected from the ground. Hens were approached to within 20 m but not intentionally flushed when radio tracked on the ground. Locations were plotted on 1:24,000 U. S. Geological Survey topographic maps and dis.tances. and directions. from .. capture site were measured. Comparisons of movements to nests by hens from different leks' ·were made using Kruskal-Yallis one-way analysis of variance by ranks and pair-wise multiple comparisons (test alpha - 0.20) to test for differences in median movements (Dunn 1964). Microhabitat Sampling at Nest Sites Vegetation at nests was measured within a week of nest depredation, abandonment, or hatch. Heights of the nearest shrub, forb, and grass were measured at nest bowls to the nearest centimeter. A 10-m north-south transect was centered on the nest bowl and used to measure line-intercept of canopy cover for sagebrush and other shrubs (Canfield 1941). Heights of all shrubs intersected by the transect were also measured. Sagebrush density at the nest bowl and at 25 m in each cardinal compass direction was measured using a O.OOl-ha circular plot. Visual obstruction of vegetation at the nest was �75 measured using a 3-sided cover board (Jones 1968) placed on the nest bowl and measured from 45° above horizontal from 0°, 120°, and 240° aspects. No comparative measures of vegetation were obtained from random plots within nesting habitats. RESULTS Trapping and Radiomarking Seventy-one hens were captured and radiomarked on or within 1.0 km of strutting grounds. Most (n - 50; 70.4%) were captured by spotlighting and nearly all (n ~ 56; 78.9%) were adults. The number of radio-marked hens was 22, 20, 14, and 15 at Coalmont, Delaney Butte, Boettcher Junction, and Spring Creek #1 strutting grounds, respectively. Telemetry Systems solar-powered transmitters were placed on 25 hens and Holohil battery-powered transmitters were placed on the remaining 46 hens. No hens were captured at Raven and Lost Creek leks because vehicular access to these sites was delayed until after the peak of hen attendance due to snow accumulation and persistent snowdrifts. Access into Spring Creek III and Boettcher Junction leks was also delayed and resulted in less trapping effort and reduced success in trapping hens. The average body mass of 47 adult hens captured was 1634 ± 94 gms and the average mass of 13 yearlings was 1466 ± 92 gms. Radio transmitters were 0.8-1.3% of the body mass of hens and were not thought to markedly affect their movements or choice of nesting habitat. Movements of Radio-marked Hens to Nest Sites Forty-three of 71 radio-marked hens were followed to nests (Table 1). Eight hens (all adults) whose first nests were depredated renested; these nests are included in the data on lek-to-nest-movements. All 8 renesting hens selected a site within 1.0 km of their initial nest for their second nest. Four hens were killed by predators prior to nesting and 10 hens could not be radio located from the ground 2 weeks after being marked. Because of intensive search efforts, it is likely that most of the missing hens moved farther than 5.0 km to nests or their radios failed after being released. Transmitter signals were received from 3 of these hens during an aerial search on 27 May (4-5 weeks after capture) and indicated these hens had apparently moved 6-9 km. However, they could not be located from the ground. One transmitter was recovered (no sign of mortality) in August nearly 8 km from the capture site. Two hens moved onto private lands· (approximately 12 and 20 km from capture site) where access was restricted. . The remaining 12 hens (8 adults, 4 yearlings) were located at least twice weekly but no nests were detected. It is likely that most of these hens attempted to nest but the nests were depredated or abandoned prior to incubation and no nests were located. Movements of these hens during the post-capture period were similar to movements of nesting hens from the same strutting grounds. Little information could be obtained from 10 hens for which radio signals were not heard. Radio signals from these hens were monitored for 2 to 34 days before being lost. During this period the birds were located from 400 to 3,700 m from capture sites. Transmitter type did not appear to be a factor in the loss of these hens as 4 transmitters were solar powered (16%) and 6 were battery powered (17%). �,~ - Table 1. 1993. Fates of radio-marked sage grouse hens in North Park, Colorado, Fates of radio-marked Strutting Ground Boettcher Jct. Coalmont Delaney Butte Spring Creek #1 Totals a Includes Includes b hens n Successful Depredated Abandoned Lost signal No nest found 14 22 20 15 71 2 3 5 1 11 5 lOa 9a 6 30 1 3 2 0 6 4 2 1 5b 12 2 4 3 3 12 2 hens killed by predators prior to nesting. 2 hens moving onto private land where access was denied. Distances and directions of movements of hens from leks to nest sites were unique for each strutting ground (Figs. 1-4) and reflected habitat types adjacent to each area. Movements from 1ek-of-capture to nest site ranged from 0.45 to 29.0 km with a mean movement of 3.16 ± 4.39 km (median = 1.73 km). There was no relationship between capture date and distance moved to initial nests (Table 2). Overall, 14 of 51 hens (27.5%) moved ~3.0 km to nest (Fig. 5). Hens from Delaney Butte strutting ground had the shortest average movements (1.57 ± 0.89 km) followed by hens from Boettcher Junction (2.39 ± 1.95 km), Coalmont (3.64 ± 3.40 km), and Spring Creek #1 (6.54 ± 9.25 km). Median movements to nests were 1.31, 1.81, 2.34, and 4.35 km, respectively. The hens radiomarked at Spring Creek 1ek moved farther (Q < 0.20) to nests than hens from Delaney Butte and Boettcher Junction leks. No other ~ignificant differences in 1ek-to?nest movements were documented. Table 2. Relationship between date of capture and movement of hens from capture location to initial nest site, North Park, Colorado, 1993. Capture date 01-10 11-20 21-30 01-11 Apr Apr Apr May N hens 8 33 25 5 N nests 4 23 13 3 ~ Distance (km2 moved to nests SD Range 2.84 2.83 5.12 1.29 1.44· 2.91 7.57 0.74 1. 31 - 3.10 0.45 - 11.24 0.79 - 29.00 0.45 - 1.83 Nest Fates Only 11 of 43 hens (9 adult, 2 yearling) whose nests were locate~ were successful in hatching one or more eggs in their clutch for an estimated hen success of 25.6% (11/43 nests). Renesting was documented for 8 hens (all adult) that lost their initial clutches, all renests were unsuccessful due to depredation. Overall nest success was 21.6% (11/51 nesting attempts). Most nest loss (34 of 40 nests, 85%) was due to depredation, primarily from �I Boettcher Jct. Fig. 1. Sage grouse nest sites in relation to Boettcher Junction strutting ground, Jackson County, Colorado, 1993. Coalmont 4 km Fig. 2. Sage grouse nest sites in relation to Coalmont strutting ground, Jackson County, Colorado, 1993. I �-" ~ Delaney Butte 4 km N: I Fig. 3. Sage grouse nest sites in relation to Delaney Butte strutting ground, Jackson County, Colorado, 1993. Spring Creek 4 km , NI Fig. 4. Sage grouse nest sites in relation to Spring Creek #1 strutting ground, Jackson County, �30 Ul I- Ul 15 w Z ZI 10 < 1.0 • ~ SPRING CREEK 1 ~ BOrnCHER II COALMONT DELANEY SUITE 1.1-2.0 2.1-3.0 3.1-4.0 4.1-5.0 DISTANCE JCT. >5.0 (km) Fig. 5. Lek-to-nest movements of 51 sage grouse captured at 4 strutting grounds in Jackson County,. Colorado, 1993. Richardson's ground squirrels, although 6 nests were abandoned, possibly due to investigator disturbance. If we assume the 12 hens which were not found on nests were also unsuccessful in nesting (they were not found on nests or with broods), then estimated hen success declines to 20% (11/55 hens). There was no relationship between date of capture and nest success. Seven of 27 hens (25.9%) captured prior to 21 April were successful and 31.2% of hens captured on or after 21 April were successful. Adults appeare~ to be less succeas fu'l, (9 of 37'nests, 24.3%) than yearlings (2 of·6 nests, 33.3%). but too few yearlings were included for meaningful analysis. Vegetative Characteristics at Nest Sites Vegetative characteristics were documented for 50 nest sites (1 nest at Delaney Butte could not be relocated after depredation). Vegetative heights at nest bowls averaged 55.7 ± 11.7 cm for shrubs, primarily big sagebrush (~. tridentata), 6.4 ± 6.6 cm for forbs, and 20.8 ± 15.2 cm for grasses and was similar among hens from different strutting grounds (Table 2). Average VOR was 65.0 ± 15.8%. There were no apparent differences in vegetative characteristics at nest sites between adult and yearling hens. Although not quantified, nests wer~ typically located in sites (e.g., roadsides, mima mounds, irrigation ditches, other disturbed areas) having taller vegetative cover than adjacent areas. �80 Live sagebrush canopy cover along 10-m nest transects averaged 35.3 ± 14~9%; dead sagebrush averaged 8.1 ± 7.4%. Average sagebrush height was 37.5 ± 16.0 cm and density averaged 14,700 ± 6,300 p1ants(ha (Table 3). There were no differences in sagebrush height, canopy cover, or density among nest sites from hens banded at different strutting grounds. Table 3. Vegetative height and VOR at 50 sage grouse nest sites in North Park, Colorado, 1993. Values are presented as means ± 1 S.D. Height n Lek tf. Boettcher Jct. Coalmont- . Delaney Butte Spring Creek #1 Average 10 16 16 8 Shrub SD 48.8 56.9 60.4 52-.~654.6 (cm) Grass Forb 11.3 6.8 13.8 11.9 14.0 tf. SD tf. 3.4 8.9 6.4 4.9 6.4 2.6 10.2 3.8 3.4 6.6 13.9 19.9 27.2 18.6 20.8 of sagebrush at 10-m transects Table 4. Characteristics grouse nests in North Park, Colorado, 1993. n Lek (plants(ha) Boettcher Jct. Coalmont Delaney Butte -Spring Creek #1 Average 10 16 16 8 % Cano:QY Cover SD x Heig;ht (cm} SD tf. 34.3 35.4 35.7 35.4 35.3 29.2 38.2 44.0 38.0 37.5 14.6 17.0 13.0 17.3 14.9 11.1 11.6 17.9 16.3 16.0 VOR SD 4.3 7.6 18.9 11.1 15.2 65.2 63.2 65.5 66.4 65.0 centered 5.9 7.0 14.8 15.7 15.8 on sage Density tf. 18,900 15,200 11,800 14,600 14,700 SD 7,700 6,300 5,300 3,100 6,300 DISCUSSION Movements to Nests Previous studies of nesting sage grouse have documented a close relationship between strutting grounds and nest sites, with most nests located within 3.2 km of leks (Patterson 1952, Schlatterer 1960, Gill 1965, Martin 1970, Braun et al. 1977). Prior to the use of radiotelemetry, however, the spatial relationship between lek-of-mating.and nest site was poorly known. Studies ·of female sage grouse using radiotelemetry indicate the range of movements of hens from lek-of-capture to nest site recorded in this study was similar to that reported elsewhere (Montana, Yallestad and Pyrah 1974; Jackson County, Colorado, Petersen 1980; Idaho, Wakkinen et al. 1992). While the average movement by hens to nest sites is typically < 3.0 km, some hens move much farther to nest sites. The reasons for these longer movements are unknown because there is a lack of knowledge concerning seasonal home ranges of hens in relation to strutting grounds and nests. �81 If strutting grounds are a focal point within female home ranges (Bradbury 1981), then nests should be clustered around strutting grounds where hens breed. However, a recent study (Wakkinen et al. 1992) indicates the distribution of sage grouse nests may be randomly located with respect to strutting grounds. Consequently, spring home ranges of-hens and nests should be randomly distributed within suitable habitats adjacent to leks. Therefore, if the distribution of suitable nesting habitat is generally uniform around. strutting grounds, then hens should select nest sites uniformly or randomly around leks. Petersen (1980) and Schoenberg (1982) reported that adult hens nested farther from strutting grounds than did yearlings. Since adult hens attend strutting grounds earlier than do yearlings (Petersen 1980), they might be expected to have first selection of nesting habitats and chose the better quality habitats. Longer movements by adult hens was not documented in this study, possibly due to small samples of yearlings marked, but the longest movement between a capture site and nest site was of a yearling hen. Differences in hen movements among strutting grounds may be related to density of hens. If the number of hens attracted to a strutting ground is positively related to the number of displaying males, then the number of hens associated with Delaney Butte strutting ground was the least, followed by Boettcher Junction, Coalmont, and Spring Creek #1 (C. E. Braun, unpubl. data). If hens have exclusive nesting territories within their home range, and their home ranges include the strutting ground where they breed, then we expect the shortest movements by hens to nests to occur at the smallest leks, and the farthest movements to occur at the largest leks. This relationship was apparent in this study. However, the shortest movements to nests was observed at Delaney Butte strutting ground, where hens were able to use nearby ungrazed rangeland for nesting. Nest Depredation The high rate.of nest depredation or failure observed in this study is not unusual, and may have been affected by the study. Evidence from examination of hunter harvested hens in North Park in 1993 suggests nest success was 63% overall (adults - 69%, yearlings ~ 40%) (C. E. Braun, unpubl. data). This difference suggests radio-marked hens may have been less successful than un-marked hens, and the effects of radio tracking and nest' monitoring may have decreased their nest success. Excluding the 6 hens which abandoned their nests, the estimated hen success (30%) is half that as reflected in the hunter harvest. The estimate~of nest success from wings. collected in the hunter harvest may be biased by renesting hens. Since hens which were unsuccessful with initial and renesting attempts would not begin molting their primary feathers until the loss of their second nest, their primary molt may have been delayed and have been similar to or later than in hens successfully hatching their initial clutch. The number of young as determined from the hunter harvest (1.7. chicks/successful hen) was the lowest recorded in 20 years in North Park (C. E. Braun, unpubl. data) which may indicate that actual nest success as determined from wings collected in the hunter harvest may have been overestimated. The major cause of nest depredation was the Richardson's or Wyoming ground squirrel. Other s~udies of nesting sage grouse in North Park have also documented high rates of nest depredation by this rodent (Gill 1965, Petersen 1980). Densities of ground squirrels in North Park are unknown, but ground squirrels and their burrows were commonly observed throughout the area. �52 Fagerstone (1982) documented a positive relationship between ground squirrel densities and low herbaceous cover, which may occur with excessive livestock grazing. Thus, it is likely that ground squirrel densities have increased with high grazing levels in North Park and the recent drought which resulted in less forage for cattle. Increased ground squirrel densities and reduced vegetative cover for sage grouse nests could have increased the levels of ground squirrel depredation observed in 1993. LITERATURE Amstrup, S. C. 1981. 44:214-217. A radio-collar CITED for game birds. J. Wildl. Manage. Beck, T. D. I. 1975. Attributes of a wintering population of sage grouse, North Park, Colorado. M. S. Thesis. Colorado State Univ., Fort Collins . .49pp. Bradbury, J. W. 1981. The evolution of leks. Pp. 138-169 in R. D. Alexander and D. W. Tinkle, eds. Natural selection and social behavior: recent research and new theory. Chiron Press, New York. Braun, C. E., T. Britt, and R. O. Wallestad. 1977. Guidelines of sage grouse habitats. Wildl. Soc. Bull. 5:99-106. for maintenance Canfield, R. 1941. Application of the line interception of range vegetation. J. For. 39:386-394. method Dunn, O. J. 1964. 252. Technometrics Multiple comparisons using rank sums. in sampling 6:242- Emmons, S. R. 1980. Lek attendance of male sage grouse, North Park, Colorado. M. S. Thesis, Colorado State Univ., Fort Collins. 69pp. Fagerstone, K. A. 1982. Ethology and taxonomy of Richardson's ground squirrel (Spermophilus richardsonii). Ph.D. Diss., Univ. Colorado, Boulder. 298pp. Giesen, K. M., T. J. Schoenberg, and C. E. Braun. 1982. Methods sage grouse i~ Colorado. Wildl. Soc. Bull. 10:224-231. Gill, R. B. 1965. Distribution in North Park, Colorado. Collins. 185pp. and abundance M. S. Thesis, of a population of sage grouse Colorado State Univ., Fort Gregg, M. A., J. A. Crawford, M. S. Drut, and A. K. DeLong. Vegetational cover and predation of sage grouse nests Wildl. Manage. 58:162-166. Jones, R. E. 1968. A board to measure Wildl. Manage. 32:28-31. for trapping cover used by prairie Martin, N. S. 1970. Sagebrush control related to habitat occurrence. J. Wildl. Manage. 34:313-320. 1994. in Oregon. grouse. J. J. and sage grouse �Patterson, R. L. 1952. Colo. 34lpp. The sage grouse in Wyoming. Sage Books, Inc., Denver, Petersen, B. E. 1980. Breeding and nesting ecology of female sage grouse in North Park, Colorado. M. S. Thesis, Colorado State Univ., Fort Collins. 86pp. Remington, T. E. 1983. Food selection, nutrition, and energy reserves of sage grouse during winter, North Park, Colorado. M. S. Thesis, Colorado State Univ., Fort Collins. 89pp. Schlatterer, E. F. 1960. Productivity and movements of a population of sage grouse in southeastern Idaho. M. S. Thesis, Univ. Idaho, Moscow. 87pp. Schoenberg, T. J. 1982. Sage grouse movements and habitat selection in North Park, Colorado. M. S. Thesis, Colorado State Univ., Fort Collins. 86pp. Toepfer, J. E., J. A. Newell, and J. prairie grouse hens on display Tech. Coord. Prairie chickens S. Dep. Agric. For. Servo Gen. Monarch. 1988. A method for trapping grounds. Pp. 21-23 in A. J. Bjugstad, on the Sheyenne National grasslands. U. Tech. Rep. RM-159. Wallestad, R. 0., and D. Pyrah. 1974. Movement and nesting of sage grouse hens in central Montana. J. Wildl. Manage. 38:630-633. Wakkinen, W. L., K. P. Reese, and J. W. Connelly. 1992. Sage grouse nest locations in relation to nests. J. Wildl. Manage. 56:381-383. Prepared by: Kenneth M. Giesen Wildlife Researcher ��(j) CD Iv rc--Jo vJ; ld~feJOB PROGRESS REPORT State of: Colorado Project: W-167-R Work Plan: Job Title: 12 Job Bird Research _lI__ Effects of Mycoplasma Survival of Merriam's Period Covered: Author: Upland 01 January Infection on Reproductive Wild Turkeys through 31 December Performance and 1993 Richard W. Hoffman Personnel: Thomas A. Artiss, Clait E. Braun, Renzo Del Piccolo, Richard W. Hoffman, and Robert T. Magill, Colorado Division Wildlife; William R. Davidson and Page M. Luttrell, Southeastern Cooperative Disease Study. ABSTRACT Fifty-eight Merriam's wild turkeys (Melea~ris ~allopavo merriami), including 30 adult and 28 subadult females, were surveyed by serologic and cultural methods for evidence of Mycoplasma ~allisepticum (MG) , M. synoviae (MS) , M. melea~ridis (MM), and M. ~allopavonis (MGp). Reproductive performance was subsequently monitored using radiotelemetry techniques. No clinical signs of Mycoplasma infection were evident in any of the birds examined. Most birds showed no rapid plate agglutination (RPA) reactions to laboratory (MG.~ 66%, MS - 74%) or commercially prepared (MG - 59%, MS = 34%, MM = 52%) antigens. When the strongest reaction was used regardless of the antigen yielding that reaction, 29 (50%) birds were classified as positive reactors to MG or MS, including 17 (61%) subadults and 12 (40%) adults. Hemagglutination inhibition (HI) tests were uniformly negative for MG and MS. Mycoplasma spp. organisms were isolated from 41 of 47 tracheal cultures. All were negative for MG by fluorescent antibody tests. A subsample (n - 19) of the 41 isolates was tested for MS, MM, and MGp, plus 11 other species of Mycoplasma. Four isolates were identified as MGp and 6 were identified as M. ~allinaceum. Twelve of 15 chickens showed positive RPA reactions to MS that were supported by positive (titers ~ 40) HI results; however, none of the Mycoplasma organisms isolated from the chickens was identified as MS. Isolates obtained from the chickens were identified as M. pul1orum, M. gal1inaceum, or M. gallinarum. Clutch size, nesting success, and fertility did not differ between seropositive and seronegative wild turkeys, whereas, hatching success was lower in the positive group. Evidence for MG or MS infection in this population was inconclusive. If the RPA positive reactions to MS in the wild turkeys were due to contact with carrier chickens, the fact that there was no overt disease, widespread infection, or suppressed reproductive output suggests the infection was old and no longer active or the organisms involved were of low virulence and infectivity. ��87 EFFECTS OF MYCOPLASMA INFECTION ON REPRODUCTIVE PERFORMANCE AND SURVIVAL OF MERRIAM'S WILD TURKEYS Richard W. Hoffman INTRODUCTION Restoration of the wild turkey has been one of the most noteworthy successes of the wildlife management profession (Kennamer and Kennamer 1990). However, only recently have wildlife managers realized the potential risks of disease introduction and dissemination associated with this practice (Nettles and Thorne 1982, Nettles 1984, Amundson 1985). Although no disease problems have been linked to wild turkey restoration efforts using wild-trapped stock (Davidson 1987), without data to prove otherwise, such efforts could be erroneously blamed for disease outbreaks in other wildlife populations or more importantly, for unrelated disease problems in the domestic pOUltry industry (Amundson 1985). Mycop1asmosis is the disease that has prompted these concerns. Three species of Mycoplasma are known to be pathogenic for domestic poultry including M. gallisepticum (MG), M. meleagridis (MM), and M. synoviae (MS). MG is the most detrimental, producing sinusitis, reduced egg production, decreased hatching success, and poor juvenile growth; however, it seldom causes direct mortality (Yoder 1984). Mycoplasma infections can produce chronic carriers and may be transmitted through the respiratory tract, venerea11y via semen, or transovarially through the oviduct to the egg. Prior to 1980 there were few reports of mycoplasmosis in either wild or semiwild, free ranging turkeys. Trainer (1973) first reported the isolation of Mycoplasma from wild turkeys captured in Texas and Wisconsin. The isolates were not identified nor associated with any disease. Hensley and Cain (1979) conducted a serologic survey of wild turkeys in Texas and detected antibodies to MG in counties supporting commercial poultry operations; isolation was not attempted. In 1980, antibodies to MG and MM were detected in wild turkeys trapped in Missouri for release in Wisconsin; further serologic testing from 1980 to 1984 disclosed the presence of Mycoplasma seropositive birds in Wisconsin, Minne~ota, and Missouri (Amundson 1985). Clinical MG infec~ion was also found in small flocks of free ranging, wild-type turkeys in California (Jessup et al. 1983) and Georgia (Davidson et al. 1982) that were associating with domestic fowl. More recently, antibodies to MG, MM, and MS have been detected in flocks from Texas, New Mexico, Arizona, Colorado, Oklahoma, and North Dakota (Rocke and Yuill 1987, Fritz et al. 1992). Isolation attempts for serotypes which included these species were mostly unsuccessful, but numerous isolates were serotyped as M. ga1lopavonis (MGp). Antibody to MS was also detected in 85 of 94 wild turkeys transported from Arizona to Idaho in 1984; all were condemned by the Idaho Department of Agriculture (Anon. 1985~). Birds from the same population in Arizona were released in Utah without any prior testing (J. A. Roberson, Utah Div. Wi1d1. Resour., pers. commun.). Extensive disease monitoring in the southeastern United States suggests wild turkeys in that �88 region are not important in the epizootiology of MG, MS, or MM (Davidson et a1. 1982, 1985, 1988). In Colorado, evidence of MG, MM, and MS infections was found in declining populations of wild turkeys on the Uncompahgre Plateau and Devil's Creek State Wildlife Area; conversely, serological tests on wild turkeys trapped from stable populations near Trinidad and Pueblo were negative (Adrian 1984). MG was isolated from the trachea of 1 bird from Devil's Creek and MM was isolated from the oviduct of another bird captured on the Uncompahgre Plateau. Mycoplasma was cultured from several other birds, but the serotype was unknown. Whereas mycoplasmosis was implicated in the decline of wild turkeys on the Uncompahgre Plateau (Adrian 1984), there was only indirect evidence to support this contention. Rocke and Yuill (1987) found no evidence to link mycop1asmosis with the decline of wild turkeys on the Welder Wildlife Refuge in Texas. Although free-ranging wild turkeys are susceptible to Mycoplasma, the consequences of infection (i.e., transmission between wild and domestic birds, pathologic effects, persistence, and long-term population impacts) are unclear. Experimental MG infections in captive-reared wild turkeys reduced production, fertility, and hatching success of eggs compared to noninfected controls (Rocke et al. 1988). Egg production was unaffected by MG infection 2 years postinoculation, but fertility remained low both years.. No mortality occurred in either the infected or control groups. Currently, there is only circumstantial evidence to support the contention that wild turkey populations harbor or perpetuate MG, MM, or MS, or that such infections result from contact with domestic fowl and suppress wild populations through subtle changes in reproductive performance. Despite the uncertainly surrounding the Mycoplasma issue, the benefits of testing not only for Mycoplasma spp., but also for other avian pathogens, outweigh the negative aspects (Davidson 1987). The Wildlife Disease Association (YDA) has prepared an advisory statement on disease monitoring of wild turkeys including suggested guidelines for conducting disease testing (YDA 1985). These guidelines have been endorsed by the International Association of Fish and Wildlife Agencies (Nettles 1984) and The United States Animal Health Association (Nettles and Thorne 1982). P. N. OBJECTIVES Compare survival and reproductive performance (i.e., egg production, fertility, nesting success, and hatching success) of free-ranging, female wild turkeys that are seropositive for Mycoplasma with those that are negative~ SEGMENT OBJECTIVES 1. Trap and collect blood samples from 50 wild turkeys known to be associating with domestic fowl. 2. Collect blood samples from 15 domestic with wild turkeys. chickens known to be in contact __ . �89 3. Test blood samples for antibodies to 3 pathogenic gallisepticum, M. synoviae, and M. meleagridis. 4. Obtain tracheal swaps from at least 10 wild turkeys and culture Mycoplasma gallisepticum, M. synoviae, M. meleagridis, and M. gallopavonis. S. Attach radio transmitters to a sample of seropositive and seronegative wild turkeys and monitor their reproductive performance. 6. Compile and analyze data, and prepare progress mycoplasmas: Mycoplasma for report. STUDY AREA Trapping was confined to the Hittle Ranch approximately 6 km northeast of Collbran, Colorado in Mesa County. From here, radio-marked birds ranged over 120 km2 of surrounding areas during the breeding and brood rearing periods. The primary areas where turkeys were found included Hawxhurst Creek, Buzzard Creek, Salt Creek, Kimball Creek, Smalley Gulch, and Big Creek. These north/south oriented drainages were dominated by narrow1eaf cottonwood (Populus angustifolia) along the bottoms giving way to Gambe1 oak (Quercus gambe1ii) interspersed with pinyon pine-juniper (Pinus edu1is-Juniperus spp.) woodlands on the drier slopes. Most of the mesas had been cleared and planted to hay meadows. Livestock were pastured on the meadows during winter. About 85% of the area was privately owned. Turkeys first appeared at the Hittle Ranch about 25 years ago. Over the past 5 years, approximately 120-140 wild turkeys have wintered at the ranch. Although not intentionally fed, the turkeys consume large quantities of artificial foods such as oat hay, alfalfa hay, corn, beef feed, and poultry feed provided for domestic livestock. Domestic chickens roam freely about the ranch, while others are confined to holding pens. Domestic turkeys, guinea fowl, and-peacocks also have been raised on the Hittle Ranch, but were not present during this study. Wild turkeys were frequently observed feeding with the free-roaming chickens and commonly scratched around the holding pens for spilled poultry feed. METHODS Trapping, Marking, and Radiotracking Turkeys were baited with oat hay and corn, and live-trapped with cannon nets during February and March 1993. Captured birds were weighed on an electronic scale, classified to age and sex, and banded with serially numbered aluminum leg bands. Allf1ex livestock eartags were attached to the patagium. Ages were recorded as subadu1t (8-10 months) or adult (>18 months). Fifty-three females were equipped with lithium battery-powered transmitters (model HLPB2150-LD, Wildlife Materials, Carbondale, IL) attached with a poncho collar (Amstrup 1980). The radio package weighed < 40 g. Tracking was conducted from the ground using a 3-element Yagi antenna and Te10nics TR-2 receiver with a TS-I scanner attachment. All locations were verified by visual observation and recorded to the nearest 50 m as Universal Transverse Mercator coordinates. One aerial search was conducted in late April to locate birds not found during �90 ground searches. Birds found during the aerial search were subsequently located from the ground. Disease Monitoring All birds were examined for clinical signs of Mycoplasma infection and their general physical condition was noted. Blood (8-10 cc) was collected by jugular venipuncture from each radio-marked turkey and from 7 free-ranging and 8 confined chickens. The plasma was separated by centrifugation, pipetted into a separate container, and mailed overnight express to the Southeastern Cooperative Yildlife Disease Study (SCDS), College of Veterinary Medicine, University of Georgia, Athens. Tracheal swaps were obtained from 47 wild turkeys and 15 chickens, placed into Frey's media with swine serum, and mailed overnight express to SCDS. Since MGp was identified from wild turkeys captured at this site in 1992, a subsample (n - 28) of media tubes received antisera to MGp to inhibit its growth. Rapid plate agglutination (RPA) tests were performed on plasma from each bird. Two plate antigens were used for testing for MG and MS, a commercial antigen prepared by Salsbury Laboratories, Inc. (Charles City, IA) and a laboratoryprepared antigen made by the Poultry Disease Research Center (PDRC), College of Veterinary Medicine, University of Georgia, Athens. One plate antigen (Salsbury) was used for MM testing. Agglutination was scored on a 0 to 4 ' scale, with 0 being a negative reaction, 1 and 2 a weak reaction, and 3 and 4 a strong reaction. Any reaction> 1 was considered-positive. Yhen the 2 antigens produced different results, those from the antigen yielding the stronger reaction were used. Appropriate control sera were used at intervals throughout the sampling period. Plasma samples also were tested for antibodies to MG, MS, and MM with the hemagglutination inhibition (HI) test using a laboratory-prepared antigen from PDRC. The test was performed as described in the National Poultry Improvement Plan (Anon. 1985Q). HI titers ~ 40 were considered positive. Twelve sera samples representing a range of reactivity to MG or MS on RPA tests were sent to North Carolina State University, Raleigh for immunoblot testing for antibodies to MG and MS (Avakian et al. 1992). Reproductive Parameters During late April and May, hens were located once every 2-3 days to ascertain if they were nesting. Suspected nest sites were circled and flagged from> 20 m away. Some nests were visually observable from this distance. Others were monitored but not approached for 30 days unless the radio signal indicated the hen was gone. Nest sites were visited daily as the anticipated hatch date approached. Most hens were located often enough just before and during the early stages of incubation to approximate within 2 days of when they started incubating. For successful nests (~ 1 egg hatched), onset of incubation was estimated by backdating 28 days (incubation period) from the date of hatch. Clutch size, fertility (unhatched eggs were broken open and examined for developing embryos), nest success (X hens that hatched ~ 1 egg), and hatching success (X eggs in successful nests that hatched) were determined from egg shell characteristics after the eggs hatched or after the nest was abandoned or depredated. Clutch size was also determined by visiting nest sites when the hen was away from the nest. �91 RESULTS Capture and Marking Eighty-five turkeys, including 6 adult males, 11 subadult males, 32 adult females, and 36 subadult females were captured with cannon nets; 30 adult and 28_subadult females were weighed, banded, bled, equipped with radio transmitters, and released at the trap site. Two adult females and 1 subadult female died as a result of trapping. The remaining birds were banded and released at the trap site. Recoveries and Mortalities Seven radios were recovered between 1 February (earliest trap date) and 31 July (termination of field work), 3 from subadults and 4 from adults. Radiocontact was lost with 1 adult and 4 subadult hens prior to nesting. All 7 recoveries were classified as natural mortalities, including 1 hen killed prior to nesting, 4 killed on the nest, and 2 killed after having unsuccessfully nested on their first attempt. Disease Monitoring Although the Salsbury antigen produced more (E = 0.0001) MS positive reactions than the PDRC antigen, the majority of birds still showed no RPA reaction-to MG (PDRC - 66%, Salsbury - 59%), MS (PDRC - 74%, Salsbury - 34%), or MM (PDRC - 52%) (Table 1). However, when the strongest reaction was used regardless of the antigen yielding that reaction, 29 birds (50%) had a RPA score > 1 for MG or MS. Proportionally more subadults (17/28, 61%) than adults (12/30, 40%) were classified as positive reactors (RPA score> 1), but the difference was not significant (E - 0.18). Thirteen of 17 birds with strong reactions (RPA score> 2) to MS also had strong reactions to MG. All birds clinical positive positive captured appeared to be in good physical condition. None had signs of Mycoplasma infection. 'Weights did not differ between and negative reactors for adults (E - 0.25), whereas for subadults, reactors weighed more (E - 0.08) than negative reactors (Table 2). Results for HI testing were uniformly negative for MG and MS. There was one MM titer of 20 and another of 40; the latter reaction was considered positive. Seven of 12 sera samples subjected to immunoblot tests were classified as suspicious of infection; 3 had bands corresponding to MG-specific proteins and 4 had bands indicative of MG and MS (Avakian et al. 1992). Immunoblot tests were not performed for MM. The immunoblot profiles were only considered suspicious of infection because the responses were weak in comparison to the control sera and to responses of domestic birds with known MG and MS infections. Furthermore, the immunoblot and RPA tests were not in agreement. The 5 sera samples showing no reactivity on the immunoblot tests were positive for MG and MS on the RPA tests. �92 Table 1. Seroprevalence of antibody to Mycoplasma gallisepticum (MG) , M. synoviae (MS) , and M. meleagridis (MM) among female wild turkeys as determined by rapid plate agglutination tests (RPA) , Collbran, Colorado, 1993. MS MG Score- RPA-lb RPA-2 RPA-l ...l1!:L_ RPA-2 RPA-2 ADULTS o 1-2 3-4 19 19 22 7 5 6 8 4 o 12 10 15 8 11 4 SUBADULTS 19 15 21 8 16 4 9 6 12 5 4 1 8 7 5 20 22 16 30 12 16 ADULTS AND SUBADULTS o 38 1-2 3-4 11 9 43 14 34 14 10 1 ~ - no reaction, 1-2 - weak reaction, 3-4 - strong reaction. ~A-l - Poultry Disease Research Center prepared antigen, RPA-2 commercially prepared Salsbury antigen. Table 2. Comparative weights (kg) of serologically positive (RPA score> and negative female wild turkeys, Collbran, Colorado, 1993. Descriptive statistic n ~ SD Range Adult Subadul t + 12 4.86 0.31 4.27-5.25 1) + 18 5.02 0.40 4.28-5.91 17 4.35 0.24 4.02-4.85 11 4.19 0.21 3.79-4.53 Mycoplasma spp. organisms were isolated from 41 of 47 tracheal cultures .. All were negative for MG by fluorescent antibody tests. A subsample of 19 cultures was tested for MS, MM, MGp, M. gallinarum, M. pullorum, M. gallinaceum, M. iners, M. iowae, M. cloacale, M. lipofaciens, M. glycophilum, M. columbinum, M. columborale, and M: columbinasale. Four isolates were identified as MGp and 6 were identified'as M. gallinaceum. Based on these �(I·") :; .J results, it is likely the majority of birds were infected with one or both of these organisms. Twelve of 15 chickens showed positive reactions to MS by the RPA tests compared to 5 of 15 chickens with MG positive reactions (Table 3). HI results supported the RPA tests for MS but not MG (Table 3). HI titers ~ 40 were found for 1.3of the chickens tested for MS. Isolates were obtained from allIS tracheal cultures from the chickens and were identified as M. pullorum, M. gallinaceum, or M. gallinarum. None was identified as MG or MS. Table 3. Seroprevalence of antibody to Mycoplasma gallisepticum and M. synoviae among domestic chickens as determined by rapid plate agglutination (RPA) and hemagglutination inhibition (HI) tests, Collbran, Colorado, 1993. RPA-l" 3c 0 1 0 0 0 0 0 2 0 0 0 0 0 4 MG RPA-2 4 0 4 3 0 0 0 0 3 0 0 0 0 0 2 Hlb 0 0 20 0 0 0 0 0 0 0 0 0 0 0 0 RPA-l 4 0 3 3 1 4 4 4 4 0 0 4 4 4 4 MS RPA-2 4 0 0- 0 0 1 0 0 1 0 .0 4 4 4 4 HI 160 40 80 40 160 160 160 160 160 0 0 160 80 80 80 "RPA-1 Poultry Disease Research Center prepared antigen, RPA-2 commercially prepared Salsbury an~igen. "HI titers ~ 40 are considered positive. 'CO - no reaction, 1-2 - weak reaction, 3-4 strong reaction. Reproductive Performance Twenty-eight adult hens were monitored into the nesting period of which 9 (32%) nested successfully, 16 (including 3 hens killed during incubation) attempted to nest but failed, 2 did not attempt to nest, and 1 abandoned its nest after being disturbed by us. Nine of 16 adult hens that had the opportunity to renest did so, but only one was successful. Nesting success for first and second nest attempts was 36%; 94 of 116 eggs (81%) in successful nests hatched. In comparison, 7 of 24 (29%) subadult hens followed into the nesting period nested successfully, 17 (including 1 hen kil1ea during incubation) attempted to nest but failed, and 5 did not attempt to nest. Only �94 1 subadult hen attempted to renest and it was successful. Nesting and hatching success of·subadu1ts for first and second nest attempts was 33 and 86% (74/86 eggs hatched), respectively. Nineteen (9.4%) of the 202 eggs laid in successful nests were infertile. Clutch size did not differ between first and second nest attempts of adults (E - 0.33) or between age classes (E - 0.28) for first nest attempts only (Table 4). Adults initiated incubation approximately one week earlier than subadults (Table 5). An average of 22.2 ± 3.0 days (n - 9) elapsed between when a hen lost its first nest and when it started incubating a second clutch. Onset of incubation for renest attempts ranged from 21 May to 16 June (median = 5 Jun). Clutch size (E - 0.87), nesting success (E - 0.13), and fertility (E - 0.71) did not differ between serologically positive and negative birds (Table 6). However, because more fertile eggs did not hatch from nests of positive hens, overall hatching success was lower (E - 0.05) for positive than negative birds. Seven of the 10 birds that renested and 5 of 7 birds that presumably did not attempt to nest were serologically positive. Table 4. Clutch size for first and second nest attempts wild turkeys, Collbran, Colorado, 1993. of adult and subadult Descriptive statistic Second attem12t Adult Subadult First attem12t Adult Subadu1t 17 11.8 1.7 9-15 n ~ SD Range 14 11.0 2.2 8-16 8 11.0 2.0 9-15 1 9.0 Table 5. Chronologie distribution for onset of incubation for first nest attempts by adult and subadult wild turkeys, Collbran, Colorado, 1993 Adults Time period N % 26-30 Apr 1-5 May 6-10 May 11-15 May 16-20 May 21-25 May 20-30 May 31 May-4 Jun 2 5 7 8 3 0 0 0 8 20 28 32 12 0 0 0 Subadults % N l Median Range 8 May 27 Apr-19 May 0 1 3 7 1 2 3 2 0 5 16 37 5 10 16 10 14 May 2 May-3 Jun �95 Table 6. Comparative reproductive parameters of serologically positive score ~ 2) and negative wild turkeys, Collbran, Colorado, 1993. Parameter Clutch size (z ± SD) First nest Nesting success, % Hatching success, % Fertility, % Positive 11.5 ± 2.2 26 77 90 (MG/MS Negative 11.4 ± 1.7 46 87 91 DISCUSSION Collectively, the results are inconclusive concerning the status of MG and MS in this population. Much of the information for interpreting serologic tests performed on wild turkeys is still based on experimental testing of domestic birds. Therefore, caution must be applied in interpreting the results. In this study, RPA test results can only be interpreted as suspicious of MS infection because of the lack of confirmation by HI tests and cultures. The HI results are not surprising since HI activity in wild turkeys has been shown to decline to low or negative titers within a few months postexposure (Rocke et al. 1985, Rocke and Yuill 1988). Weak RPA reactions in conjunction with negative HI titers are not indicative of exposure based on accepted test interpretations. Reactions to MS were greater for the Salsbury than the PDRC antigen. Salsbury antigens are acknowledged to be highly sensitive with low specificity (Avakian et al. 1988) and, ·therefore, may have given more false positives than the PDRC antigen. There is also the possibility that some of the weak reactions were due to non-specific agglutination reSUlting from improper handling of the serum (Rocke et al. 1985). If the RPA positive reactions were due to MS infection, that there was no evidence of overt disease, widespread in~ection, or suppresse~ reproductive output, suggests the infection was old and no longer active or the organisms involved were less pathogenic, variant strairis.of MS or MG. Further, the optimal conditions (i.e., easy access to unlimited food supplies) under which these birds lived may have prevented any clinical manifestation of Mycoplasma infection. Variant strains of pathogenic mycoplasmas have been difficult to diagnose in commercial poultry. Birds infected with variant strains often show no clinical signs of disease and the organisms involved are difficult to isolate because they are slow-~owing and may be localized in the reproductive or lower respiratory tracts where they would not be detected using tracheal cultures (Mallinson 1983, Yoder 1986, Dingfelder et al. 1991). Another explanation for the inconclusive results is that the serological responses of the wild turkeys may have been false-positives caused by crossreactions of the MG or MS antigen with other mycoplasmas such as MGp or M. gallinaceum, which were both isolated from the wild turkeys examined in this �96 study. MGp has been identified as a common, nonpathogenic mycoplasma in freeranging wild turkeys (Cobb et al. 1992, Fritz et al. 1992, Luttrell et al. 1992), whereas, M. gallinaceum has not been previously reported to occur in wild turkeys. M. gallinaceum has been most commonly found in chickens and indeed was isolated from the chickens cultured in this study. These findings suggest the wild turkeys became infected with M. gallinaceum by interacting with the chickens and demonstrates the possibility of contact transmission between domestic and wild birds. Immunoblot tests are generally regarded as more sensitive and usually more specific than RPA or HI tests. This may explain why several sera samples that were MG or MS negative by RPA and HI tests reacted to MG or MS specific proteins on the immtinoblot tests. However, this does not explain why some of the sera samples that were RPA positive, showed negative reactions on the immunoblot tests. The immunoblot tests, like the RPA tests, are only suggestive of exposure. Their interpretation for free-ranging wild turkeys requires further research. In contrast to the test results for wild turkeys, RPA and HI tests on blood samples from the chickens were highly suggestive of MS infection. Since MG and MS share some antigens, RPA tests are often positive for both organisms even when birds are only infectedwith one of them. This may explain the positive RPA reactions to MG that were not substantiated with positive HI results. The strong seroprevalence of MS infection in chickens but not in wild turkeys was unexpected. However, a similar finding involving MG infection was reported for wild turkeys and chickens living in close association on Cumberland Island, Georgia (Luttrell et al. 1991). MG infection had been confirmed by culture and serology in this population in 1980 (Davidson et al. 1982). Despite the continued association between wild turkeys and domestic chickens on the island, surveys in 1988 indicated the MG infection had not persisted in the turkeys even though the chickens still tested strongly positive. The lack of clinical disease in the chickens, low rate of spread, low virulence, and high RPA reactivity were considered characteristic of a carrier state. The MG organism was presumed localized in the lung and reproductive tissues of the chickens from where it could not be readily transmitted to other birds. LITERATURE CITED Adrian, W. J. 1984. Investigation of disease as a limiting factor in wild turkey populations. Ph.D. Diss., Colorado State Univ., Fort Collins. 63pp. Amstrup, S. C. 1980. 44:214-217. A radio-collar for game birds. J .. Wildl. Manage. Amundson, T. E. 1985. Health management in wild turkey restoration Proc. Natl. Wild Turkey Symp. 5:285-294. Anonymous. 1985~. Idaho turkey relocations halted. Turkitat programs. 3(1):4. �97 Anonymous. 1985Q. National poultry improvement plan. Publ. 91-40 (147.7), Washington, D. C. 82pp. U. S. Dep. Agric. Avakian, A. P., D. H. Kleven, and J. R. Glisson. 1988. Evaluation of the specificity and sensitivity of two commercial enzyme-linked immunosorbent assay kits, the serum plate agglutination test, and the hemagglutination-inhibition test for antibodies formed in response to Mycoplasma gallisepticum. Avian Dis. 32:262-272. ______ , D. H. Ley, and M. A. T. McBride. 1992. Humoral immune response of turkeys to strain S6 and a variant Mycoplasma gallisepticum studied by immunoblotting. Avian Dis. 36:69-77. -. Cobb, D. T., D. H. Ley, and P. D.-Doerr. 1992. Isolation of Mycoplasma gal1opavonis from free-ranging wild turkeys in coastal North Carolina seropositive and culture-negative for Mycoplasma ga11isepticum. J. Wi1d1. Dis. 28:105-109. Davidson, W. R. programs. 1987. Disease monitoring in wild turkey restoration Proc. West. Assoc. Fish and Wi1d1. Agencies 67:113-118. ______ , V. F. Nettles, C. E. Couvillion, and H. W. Yoder, Jr. 1982. Infectious sinusitis in turkeys. Avian Dis. 26:402-405. ______ , and E. W. Howerth. 1985. Diseases diagnosed in wild turkeys (Meleagris gal1opavo) of the southeastern United States. J. Wi1dl. Dis. 21:386-390. ______ , H. W. Yoder, M. Brugh, and V. F. Nettles. 1988. Serological monitoring of eastern wild turkeys for antibodies to Mycoplasma spp. and avian influenza viruses. J. Wi1dl. Dis. 24:348-351. Dingfelder, R. S., D. H. Ley, J. M. Mclaren, and C. Brownie. 1991. Experimental infection of turkeys with Mycoplasma ga11isepticum of low virulence, transmissibility, and immunogenicity. Avian Dis. 35:910-919. Fritz, B. A., C. B. Thomas, and T. M. Yuill. 1992. Serological and microbial survey of Mycoplasma ga11isepticum in wild turkeys (Me1eagris gal1opavo) in six western states. J. Wi1d1. Dis. 28:10-20. Hensley, T. S., and J. R. Cain. 1979. Prevalence of certain antibodies to selected disease causing agents in wild turkeys in Texas. Avian Dis. 23:62-69. Jessup, D. A., A. J. Damassa, R. Lewis, and K. R. Jones. 1983. Mycoplasma gallisepticum infection in wild-type turkeys living in close contact with domestic fowl. J. Am; Vet. Med. Assoc. 183:1245-1247. Kennamer, J. E., and M. C. Kennamer. 1990. Current status and distribution of the wild turkey, 1989. Proc. Natl. Wild Turkey Sypm. 6:1-12. Luttrell, M. P., S. H. Kleven, and W. R. Davidson. 1991. An investigation of the persistence of Mycoplasma gal1isepticum in an eastern popUlation of wild turkeys. J. Wildl. Dis. 27:74-80. �98 ______ , T. H. Eleazer, and S. H. Kleven. 1992. MYcoplasma eastern wild turkeys. J. ~ildl. Dis. 28:288-291. Mallinson, E. T. 1983. Atypical serologic reactions breeding flocks. Avian Dis. 27:330-331. gallopavonis for mycoplasma in in Nettles, V. F. 1984. Report of the .fish and wildlife health committee. Proc. IntI. Assoc. Fish and ~ildI. Agencies 74: 89 -101. ____ , and E. T. Thorne. 1982. Annual report of the wildlife disease committee. Proc. United States Animal Health Assoc. 86:64-65. Rocke, T. E., and T. M. Yuill. 1987. Microbial infections in a declining wild turkey population in Texas. J. ~ildl. Manage. 51:778-782. _____ , and experimental 24: 668-671. 1988. Serologic responses of Rio Grande wild turkeys to infections of Mycoplasma gallisepticum. J. Wildl. Dis. ____ , and T. E. Amundson. 1985. Evaluation of serologic tests for Mycoplasma ga11isepticum in wild turkeys. J. Wi1d1. Dis. 21:58-61. _____ , and 1988. Experimental Mycoplasma ga1lisepticum infections in captive-reared wild. turkeys. J. Wi1d1. Dis. 24:528-532. Trainer, D. O. 1973. Some diseases of wild turkeys from Texas and Wisconsin. Pages 160-173 in G. C. Sanderson and H. C. Schultz, eds. Wild turkey management: current problems and programs. Univ. Missouri Press, Columbia. Wildlife Disease Association. 1985. Advisory in wild turkeys. Wi1dl. Dis. Newsletter statement 21:1-3 on disease monitoring Yoder, H. W., Jr. ·1984. Mycoplasma ga11isepticum infection. Pages 190-202in M. S. Hofstad, H. J. Barnes, B. W. Ca1nek, W. M. Reid, and H. W. Yoder, eds. Diseases of poultry, 8th ed. Iowa State Univ. Press, Ames. 1986. A historical account of the diagnosis and characterization strains of Mycoplasma ga11isepticum of low virulence. Avian Dis. 30:510-518. Richard W. of �99 JOB PROGRESS REPORT State of: Colorado Project: W-167-R Work Plan: Job Title: 13 Upland Bird Research Job: 10 Movements, Reproductive Success, and Habitat Use by Introduced Plains sharp-tailed Grouse Period Covered: 01 January through 31 December 1993 Author: Kenneth M. Giesen Personnel: Clait E. Braun, Kenneth M. Giesen, and Justin T. Pelletier, Colorado Division of Wildlife ABSTRACT Plans to transplant plains sharp-tailed grouse (Tympanuchus phasianellus jamesi) into eastern Colorado were temporarily postponed in 1993 because a suitable release site was not available. Monitoring of the pioneering sharptailed grouse at the Tamarack Ranch area indicated a decline of 58% from 1992. Ten male sharp-tailed grouse or sharp-tailed grouse X greater prairie-chicken (I. cupido) were observed on 4 different leks in April, 1993. ��101 MOVEMENTS, REPRODUCTIVE SUCCESS, AND HABITAT USE BY INTRODUCED PLAINS SHARP-TAILED GROUSE Kenneth M. Giesen INTRODUCTION Plains sharp-tailed grouse historically occurred along the Front Range of Colorado. Sharp-tailed grouse populations declined with human settlement and were extirpated from most of their range in eastern Colorado by the late 1800's. In recent years breeding populations were documented only in Douglas County, although winter migrants or transients have been reported from Yuma, Logan, and Yeld counties (Hoag and Braun 1990). Plans to increase distribution and populations of plains sharp-tailed grouse in Colorado will rely primarily on transplants (Braun et al. 1992). While. numerous transplants of prairie grouse have occurred, few have been successful (Toepfer et al. 1990, Rodgers 1992, Hoffman et al. 1992). Thus, it is desirable to document responses of sharp-tailed grouse to experimental transplants and evaluate parameters potentially affecting success including movements, habitat use, mortality, and reproduction. P. N. OBJECTIVES The objectives of ·this project are to assist with trapping and transplanting of plains sharp-tailed grouse into selected sites along the Front Range of Colorado and evaluate transplant success. Population characteristics of the transplanted population including movements and home range size, mortality, and production will be compared to a naturally pioneering sharp-tailed grouse population in Logan County near the Tamarack Ranch State Yildlif~ Area. SEGMENT OBJECTIVES 1. Review literature on prairie grouse introductions, movements, and habitat use. 2. . Coordinate efforts with Yyoming Game and Fish personnel and affected landowners in southeastern Yyoming to locate potential trapping sites for plains sharp-tailed grouse .. 3. Transplant up to 40 plains sharp-tailed grouse from southeastern Yyoming into suitable habitats along the ~ront Range of Colorado. 4. Radiomark up to 12 sharp-tailed grouse in the transplanted population and monitor movements, habitat use, reproduction, and mortality. 5. Radiomark up to 12 sharp-tailed grouse from the pioneering population near the Tamarack State Yildlife Area and monitor movements, habitat use, reproduction, and mortality. 6. Prepare annual progress report. �102 METHODS Sharp-tailed grouse on and near the Tamarack State Yildlife Area were inventoried by surveying all historic and occupied greater prairie-chicken leks (Benson 1987, Schroeder 1990, Hoffman et al. 1992, L. Crooks, pers. commun.) and identifying greater prairie-chickens, sharp-tailed grouse, and hybrids. Surveys for additional active leks were made by driving secondary roads within 5-10 km of known leks and stopping at 1 to 2 km intervals and listening 3-5 minutes for displaying grouse and scanning with binoculars to search for grouse on leks. ,RESULTS AND DISCUSSION Transplant of sharp-tailed grouse Permission to release sharp-tailed grouse was not obtained from the two primary transplant sites, Rocky Flats and Fort Carson (Braun et al. 1992). Additional release sites in Larimer County near Livermore and in Las Animas County near Raton Mesa will need to be evaluated., The transplant of plains sharp-tailed grouse into suitable habitats in eastern Colorado was postponed until suitable habitats can be identified and permission to release the grouse obtained. Sharp-tailed grouse at the Tamarack Wildlife Area Sharp-tailed grouse were initially observed on the South Platte Management Area (Tamarack) during November and December, 1980 (Miller 1981). In 1989 sharp-tailed grouse or sharp-tailed grouse X greater prairie-chicken hybrids were observed on leks with greater prairie-chickens and breeding populations have been increasing annually. The highest counts of prairie grouse were obtained in 1992 when we observed 60 male greater prairie-chickens on 13 active booming grounds and 24 male plains sharp-tailed grouse on 7 leks. Counts of both greater prairie-chickens and sharp-tailed grouse declined in 1993 (Table 1). Four leks (6, 16, 21, and 27) had both greater prairiechickens and sharp-tailed grouse or hybrids. LITERATURE CITED Benson, L. A. 1987. Greater prairie-chicken survey, northeast Colorado, 1987. Unpub1. Rep. Colorado Div. Yi1dl., Fort Collins. 10pp. Braun, C. E., R. B. Davies, J. R. Dennis, K. A. Green, and J. L. Sheppard. 1992. Plains sharp-tailed grouse recovery plan. Colorado Div. Yildl., Denver. 33 pp. Hoag, A. Y., and C. E. Braun. 1990. Status and distribution of plains sharptailed grouse in Colorado. Prairie Nat. 22:97-102. Hoffman. R. Y., Y. D. Snyder, G. C. Miller, and C. E. Braun. 1992. Reintroduction of greater prairie-chickens in northeastern Colorado. �103 Miller, G. M. 1981. Development of a preservation program for three species of prairie grouse. Colorado Div. Wildl., Wildl. Res. Rep., Fed. Aid. Proj. SE-3-3. Jan. 38-50. Rodgers, R. D. 1992. A technique for establishing sharp-tailed grouse in unoccupied range. Wi1d1. Soc. Bull. 20:101-106. Schroeder, M. A. 1990. Greater prairie-chicken survey, Tamarack State Wildlife Area, 1990. Unpubl. Rep. Colorado Div. Wildl., Fort Collins. 8pp. Toepfer, J. E., R. L. Eng, and R. K. Anderson. 1990. Transplanting prairie grouse: what have we learned? Trans. N. Am. Wild1. and Nat. Resourc. Conf. 55:569-579. Prepared by _,-~...o::::::;.L:==::.......;:....:....:;..:...-=~:::..:::.~=.=-:....._ Kenneth M. Giesen Wildlife Researcher C �104 Table l. Annual maximum councs of male greacer prairie-chickens and plains sharp-cailed grouse X greacer prairiechicken hybrids on leks in Logan and Sedgwick councies, Colorado, 1984-1993. 1984 Lek 1 2 3 4 '5 6 7 ---' GPC STC" 1 5 1985 --- GPC STG 7 9 1 1 2 1986 GPC STC GPC src 1988 1989 GPC STC GPC STG II 12 7 6 7 7 ---. 9 6 1990 1991 1992 1993 GPC STG GPC STC GPC STG Gl'C STC 7 6 1 2 8 7 7 2 1 7 8 6 2 5 5 5 2 5 1 1 10 6 1 1 1 2 3 7 4 2 6 2 2 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 Totals 6 3.0 Avg. 1987 5 3 5 7 1 2 1 1 1 1 ·3 9 3 10 3 7 4 1 5 5 4 5 5 4 7 5 6 3 2 3 3 8 2 6 3 4 4 1 4 20 4.0 19 6.3 3S 5.0 "Includes plains sharp-cai1ed grouse and hybrids. 32 6.4 40 5.7 41 2 2.0 5.1 7 1.8 47 13 5.2 4.3 60 24 4.6 3.4 4 6' 37 3.1 1 10 2.5 �105 JOB FINAL REPORT Colorado State of: Project: W-167-R Work Plan: Job Title: Period Author: 14 Upland Job _4_ Movements, Reproductive Success, Greater Prairie-chickens Covered: Bird Research 01 January and Habitat Use by Introduced 1991 through 30 June 1994 Grant M, Beauprez Personnel: C1ait E. Braun, Shane Briggs, Larry Budde, Courtney Crawford, Tim Davis, Jim Dennis, Frances Pusateri, Larry Rogstad, Gene Schoonveld, Mike Schroeder, Mike Trujillo, Colorado Division of Wildlife; Grant M. Beauprez, Jennifer Clarke, University of Northern Colorado. ABSTRACT Greater Prairie-chickens (Tympanuchus cupido) were declared endangered in Colorado in 1973. As part of their recovery plan, birds were released to two different sites in northeastern Colorado during April 1991 and 1992. Prairiechickens released near Pinneo established nine leks and six hens had successful nests while birds released at Wells Ranch established five leks and two hens had successful nests. Recruitment of juveniles was documented at Pinneo while no recruitment was documented at Wells Ranch. Mortality was 38% at Pinneo and 48% at Wells Ranch, and mean maximum dispersal distance of birds was 8.5 km at Pinneo and 13.1 km at Wells Ranch. These data indicate the Pinneo releases were more successful in establishing a population of Greater Prairie-chickens than releases at Wells Ranch. If the transplants are successful in establishing self-sustaining populations, Greater Prairiechickens may be delisted from endangered status in Colorado. ��107 UNIVERSITY OF NORTHERN COLORADO Greeley, Colorado The Graduate School MOVEMENTS, REPRODUCTIVE SUCCESS, AND HABITAT USE BY INTRODUCED GREATER PRAIRIE-CHICKENS IN NORTHEASTERN COLORADO - A Thesis Submitted in Partial Fulfillment of the Requirements for the Degree of Master of Arts Grant Matthew Beauprez College of Arts and Sciences Department of Biological Sciences May 1994 �ABSTRACT Beauprez, Grant Matthew. Movements, Reproductive Success, and Habitat Use by Introduced Greater Prairie-chickens in Northeastern Colorado. Published Master of Arts Thesis, University of Northern Colorado, 1994. Greater Prairie-chickens (Tympanuchus cupido) were declared endangered in Colorado in 1973. As part of their recovery plan, birds were released to two different sites in northeastern Colorado during April 1991 and 1992. Prairie-chickens released near Pinneo established nine leks and six hens had successful nests while birds released at Wells Ranch established five leks and two hens had successful nests. Recruitment of juveniles was documented at Pinneo while no recruitment was documented at Wells Ranch. Mortality was 38% at Pinneo and 48% at Wells Ranch, and mean maximum dispersal distance of birds was 8.5 Jan at Pinneo and 13.1 Jan at Wells Ranch. These data indicate the Pinneo releases were more successful in establishing a population of Greater Prairie-chickens than releases at Wells Ranch. If the transplants are successful in establishing self-sustaining populations, Greater Prairie-chickens may be delisted from endangered status in Colorado. AUTHORED BY: THESIS COM1vfITIEE: THESIS SPONSOR COMlvITITEE MEMBER EX OFFICIO MEl\1BER Clait E. Braun, Ph.D. ii �109 ACKNOWLEDGMENTS I begin by dedicating this thesis to the memory of Cullen "G-Money" Glosson who was killed in a car accident in August 1992. Cullen was invaluable to me in the field. Though he was only 15, he was very bright, a quick learner, and a hard worker. He also became my very good friend. The summer of 1992 would have been much less enjoyable without him. "Dude, you were huge!" A special thanks goes to my wife, Christy. Her patience, support, and help in the field were greatly appreciated. I would not have gotten this far without her love and encouragement. I also sincerely thank my entire family for support and having confidence that I could complete this degree. I thank Dr. Jennifer A. Clarke for seeing me through to the end of this project.. Her patience and help was astronomical, but her friendship and kindness were even greater. Throughout my "tenure" as a graduate student, she was always there if I needed any help, or just a friend to talk to. Many thanks! I thank Dr. Clait E. Braun of the Colorado Division of Wildlife for guidance, encouragement, and patience throughout the length of this project. His knowledge of grouse and expertise in the field were invaluable. His accomplishments and professionalism will always be an inspiration to me. I also thank him for giving me the opportunity to complete this research, and for providing funding for the project. I thank the entire biology faculty at UNC for making my stay educational and enjoyable. Special thanks go to Dr. James P. Fitzgerald and Dr. Ivo E. Lindauer for serving on my committee and reviewing this thesis. I also thank Dr. Michael A. Schroeder for helping me "learn the ropes" during the initial phases of this project and for teaching me the intricacies of radiotelemetry • It would not have been possible to conduct this research without the permission of the many landowners allowing me access to their property. I especially thank Jac~ Wells and Tom Kneedler for allowing us to release birds on their property, and for pulling me out of the sand on a number of occasions! My life as a graduate student would not have been nearly as much fun without the camraderie of all the grad students I have known along the way. could always count on going in the "grad lab" and finding friendship, laughter, and stimulating cqnversation! Special thanks to my Canadian the field was greatly appreciated. chili, and birding ventures! I thank Keri Fox, Jody Nelson, friend, Pthomas Artiss. His assistance I also thank him for all the movies, and Eric Powers I in for field assistance. I thank the Colorado Division of Wildlife and all the personnel that helped me during various phases of this project. Funding for this project was provided by the Colorado Division of Wildlife. iii �110 TABLE OF CONTENTS Chapter I. INTRODUCTION II. LITERATURE REVIEW Species Description Distribution Reproductive Behavior Nesting Mortality Movements Habitat Transplants III. METHODS Study Site Descriptions Trapping and Release Methodology Radiotelemetry Lek Surveys Nesting Studies Habitat Use Statistics IV. RESULTS Trapping and Release Methodology Movements Lek Establishment Behavior on Leks Nest Establishment and Success Mortality Habit.at Use V. DISCUSSION Trapping and Release Methodology Movements Lek Establishment Nest Establishment and Success Mortality Habitat Use Summary and Conclusions Recommendations APPENDIX APPENDIX APPENDIX APPENDIX LITERATURE A B C D CITED LIST OF TABLES 1. 2. 3. 4. s. Movements of Greater Prairie-chickens released April-30 September 1991 Movements of Greater Prairie-chickens released April-30 September 1992 Movements of Greater Prairie-chickens released 1 April-3D September 1991 Movements of Greater Prairie-chickens released 1 April-30 September 1992 Mean home range size of Greater Prairie-chickens Ranch, Colorado, 1 April - 30 September 1991 - iv at Pinneo, Colorado, 1 at Pinneo, Colorado, 1 at Wells Ranch, Colorado, at Wells Ranch, Colorado, at Pinneo 1992 and Wells �111 LIST OF TABLES 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. (continued) Maximum attendance of male Greater Prairie-chickens at display sites at Wells Ranch and Pinneo, Colorado, 1991-92 Number of active leks, mean number of males per lek, and mean distance of leks from the release site based on Greater Prairie-chicken lek surveys at Pinneo and Wells Ranch, Colorado, 1991-92 Nesting activity of Greater Prairie-chickens released at Pinneo, Colorado 1991 Nesting activity of Greater Prairie-chickens released at Wells Ranch, Colorado, 1991 Nest-lek and nest-release site distances of Greater Prairie-chicken nests at Pinneo and Wells Ranch, Colorado, 1991-92 Nesting activity of Greater Prairie-chickens released at Pinneo, and Wells Ranch, Colorado, 1992 Fates of radio-marked Greater Prairie-chickens released at Pinneo and Wells Ranch, Colorado, 1991-92 Fates of radio-marked Greater Prairie-chickens released at Pinneo and Wells Ranch, Colorado, 1 April - 1 August, 1991-92 Causes of mortality of Greater Prairie-chickens released at Pinneo and Wells Ranch, Colorado, 1991-92 Habitat at observed and random locations of male Greater Prairie-chickens. at Pinneo, Colorado, 1992 Habitat at observed and random locations of female Greater Prairie-chickens at Pinneo, Colorado, ·1992 Habitat at observed and random locations of Greater Prairie chicken nest sites at Pinneo, Colorado, 1991-92. Habitat at observed locations for successful and unsuccessful nests of female Greater Prairie-chickens at Pinneo, Colorado, 1991-92 Habitat at observed locations for female Greater Prairie-chickens with and without broods at Pinneo, Colorado, 1992 Habitat at observed and random locations of Greater Prairie chicken leks at Pinneo, Colorado, 1991-92 LIST OF FIGURES 1. 5. Historical and. present distribution of Greater Prairie-chickens in North America Historical and present distribution of Greater Prairie-chickens in Colorado Means, ranges, and 95% confidence intervals for weights of female and male Greater Prairie-chickens released at Pinneo and Wells Ranch, Colorado, 1991-92 Mean maximum dispersal distance of Greater Prairie-chickens released at Pinneo and Wells Ranch, Colorado, 1991-92 Locations of leks in relation to release site at Pinneo, Colorado, 6. ~ocations 2. 3. 4. 1991-92 of leks in relation to release site at Wells Ranch, Colorado, 1991-92 7. 8. 9. 10. 11. 12. Locations of leks and nests in relation to release site at Pinneo, Colorado, 1991 Locations of leks and nests in relation to release site at Wells Ranch, Colorado, 1991 Locations of leks and nests in relation to release site at Pinneo, Colorado, 1992 . Survival of radio-marked Greater Prairie-chickens 120 days post-release at Pinneo and Wells Ranch, Colorado, 1991-92 Distribution of 116 male and 154 female Greater Prairie-chicken locations by habitat type at Pinneo, Colorado, 1 April 1 October 1991-92 Distribution of 66 male and 42 female Greater Prairie-chicken locations by habitat type at Wells Ranch, Colorado, 1 April - 1 October 1991-92 v ��113 CHAPTER I INTRODUCTION The Greater Prairie-chicken (Tympanuchus cupido) originally ranged over much of the central and eastern Great Plains of North America (Aldrich 1963); its distribution has been greatly reduced due to destruction of native grassland habitats (Aldrich 1963, Jones 1963, Christisen 1969, Johnsgard 1983). The first record of Greater Prairie-chickens in Colorado was in 1897 in the extreme northeast part of the state (Cooke 1898). Evidence suggests that Greater Prairie-chickens expanded their range west and north with human settlement and advent of grain farming during the late 1800's and early 1900's (Sclater 1912, Schorger 1944, Beck 1957, Stempel and Rodgers 1961, Christisen 1969, Horak 1985). However, with intensified farming, overgrazing, and a series of drought years in the 1930's, their distribution and numbers in Colorado markedly decreased (Evans and Gilbert 1963). By 1963, the reported number of Greater Prairie-chickens in Colorado had decreased to 700-800 birds (Evans 1964). In 1973, population estimates were as low as 600 birds (Graul 1975) and the Greater Prairie-chicken was declared endangered under Colorado's Nongame, Endangered, and Threatened Species Conservation Act (Title 33, Article 8, Colorado Revised Statutes) (Pusateri 1990). The most current published population estimates indicate that at least 3000 and possibly 6000 birds were present in Yuma, Washington, and Phillips counties in 198183 (VanSant and Braun 1990). In 1990, the Colorado Division of Wildlife prepared a Greater Prairiechicken recovery plan (Pusateri 1990). The main objective of the plan was to remove the Greater Prairie-chicken from the state's endangered and threatened list by 1995 and eventually reclassify it as a game species. Attainment of these goals involves increasing the Greater Prairie-chicken's distribution in the state by transplanting birds into previously occupied and unoccupied range. Since 1950, at least 26 attempts have been made to re-establish Greater Prairie-chickens within their former range (Toepfer et ale 1990). The majority of these attempts have met with minimal success. Many researchers have noted that dispersal of reintroduced grouse from release sites, possibly because of inadequate habitat, decreased the success of trans locations (Hamerstrom and Hamerstrom 1951; Patterson 1952; Jacobs 1959; Toepfer 1976, 1988; Wooley 1985; Hoffman 1986; Musil 1989). A number of researchers have recognized the need for additional information concerning habitat requirements for Greater Prairie-chickens Jones 1963, Robel et ale 1970b, Toepfer et ale 1990). Schwartz (1945), Hamerstrom et al. .(1957), and Kirsch (1974) note that, of the many seasonal habitat needs of Greater Prairie-chickens, nesting and broodrearing habitat seem to be important limiting factors. Griffith et ale (1989) and Toepfer et ale (1990) suggest the amount of quality habitat is the primary factor in determlning whether a translocation effort will be successful. Translocations which have had the greatest success are those in which grassland habitat has been restored and intensively managed for several years before prairie-chickens were released (Hoffman et ale 1992). Toepfer et ale (1990) also noted that,. even though an abundance of information is available concerning the basic biology and ecology of all species of prairie grouse, this information has been inadequately used when attempting to re-establish prairie grouse populations. Another factor contributing to the ineffectiveness of prairie grouse reintroductions is the lack of thorough documentation of results (Toepfer et ale 1990). Although a great deal is known about established populations of prairie grouse, little has been documented on the biology, ecology, and behaviors of translocated populations. �114 The goals of this study were to (A) successfully transplant Greater Prairie-chickens to two sites in northeastern Colorado, (B) evaluate the success or failure of Greater Prairie-chickens to establish breeding populations at the two sites and in surrounding areas, and (C) improve guidelines for introduction of Greater Prairie-chickens into new or previously occupied habitats. Specific objectives were to: I 1. 2. 3. 4. document establishment and attendance at leks, determine nest establishment and estimate nest success, document juvenile recruitment, document and measure movements of translocated prairie-chickens during late spring (1 Apr to 15 May), early summer (16 May to 30 Jun) , late summer (1 Jul to 15 Aug), and early fall (16 Aug to 30 Sep), 5. identify and compare the floristic and structural features of sites used by grouse and random sites within occupied ranges, 6. identify habitat characteristics that could be managed to enhance suitability of habitat for prairie-chickens, and 7. compare movements, nesting success, juvenile recruitment, and habitat use of transplanted Greater Prairie-chickens with those from native populations. This study will provide essential information for wildlife managers and researchers on movements, habitat use, reproductive success, and survival of introduced Greater Prairie-chickens. Evaluation of release methodology used, plus information on movements, habitat use, reproductive success, and survival will lead to increased success of future prairie grouse translocations. CHAPTER LITERATURE Species II REVIEW Description The Greater Prairie-chicken is a member of the Phasianidae, which includes pheasants and grouse. The Greater Prairie-chicken's closest relatives include Attwater's Prairie-chicken (T. c. a~~wa~eri), Lesser Prairiechicken (T. Pallidicinc~us), and Sharp-tailed Grouse (T. Phasianellus) (Howard and Moore 1991). The extinct Heath Hen (T. c. cupido) was a race of the Greater Prairiechicken (Schroeder and Robb 1993). The Greater Prairie-chicken varies in length from 40 to 48 cm. Males and females have nearly identical plumage. Both sexes are heavily barred above with buffy, brown, black, and white, while the underparts are more buffy on the abdomen and white under the tail. Males have long, ornamental neck feathers or "pinnae", corispicuouf?·.yellowcombs above the eye, orange air sacs on the fleck, and short, black, rounded, tails with the central tail..,.feathers being barred. During sexual display the pinnae are erected and the eye combs and air sacs are inflated. Females have relatively shorter pinnae and extensively barred outer tail feathers Johnsgard 1983). Distribution The Greater Prairie-chicken's current distribution in North America is drastically different from its historic range (Figure 1). They once occurred in 17 states in the Great Plains region of the United States and three prairie provinces in Canada. Greater Prairie-chickens have been extirpated from Indiana, Ohio, Kentucky, Iowa, Texas, Michigan, Massachusetts, Arkansas, Manitoba, Saskatchewan, and Alberta. Only remnant populations remain in Minnesota, Wisconsin, Illinois, Missouri, and North Dakota. Currently, the largest populations of Greater Prairie-chickens are in Oklahoma, Kansas, Nebraska, and South Dakota Johnsgard 1983). 2 �Figure 1. Historical (diagonal tion of Greater Prairie-chickens 1993). lines) and present in North America. (black shading) distribu(from Schroeder and Robb In Colorado, Greater Prairie-chickens had a population of 6,000 birds in 1983 (VanSant and Braun 1990). They currently have a breeding population of as many as 10,000 birds (C.E. Braun, pers. commun.) centered in Yuma County and following the sandhills into Phillips and Washington counties. A small population is also present in Logan and sedgwick counties due to a transplant conducted in 1984-85 (Hoffman et al. 1992) (Figure 2). Reproductive Behavior The Greater Prairie-chicken has a polygynous mating system in which males establish territories on booming grounds or "leks". Hamerstrom and Hamerstrom (1973) defined a lek as a display ground used by two or more male prairie-chickens. Hamerstrom and Hamerstrom (1960) listed the following functions of booming grounds: (1) advertisement to facilitate mating; (2) aggression, including territorial defense; (3) release pent up energy; (4) courtship display; and (5) as a means of sexual recognition. Robel (1967) suggested the role of the Greater Prairie-chicken booming ground was to: (1) select for the fittest males in the population; and (2) attract and stimulate females for breeding. Schwartz (1945) reported that males typically begin visiting the 1ek in January and that peak hen attendance and booming 9 ground activity was during the first week of April declining steadily until the end of the booming season in June. Watt (1973) reported that activity on the lek early in the breeding season was irregular, of low intensity, and dependent on the· weather. Males gather on leks in the mornings and evenings, but attendance at the morning display period is more constant than during the evening period, except during the height of the booming season. Morning displays usually begin 50 mfnutes before sunrise and continue for two to four hours. Evening displays begin one hour before sunset and end at dusk (Schwartz 1945). . 3 �Figure 2. Historical (diagonal lines) and present (black shading) distribution of Greater Prairie-chickens in Colorado (after VanSant and Braun 1990). 1 = Wells Ranch study area. 2 = Pinneo study area. Fall establishment of territories also occurs (Schwartz 1945, Ammann 1957). Johnsgard (1983) reported that territorial boundaries are re-established by mature and experienced males during fall attendance, and that young males learn the locations of the leks at that time. However, Hamerstrom and Hamerstrom (1949) did not regar~ lek attendance of males in the fall to be typical in Wisconsin. - Schwartz (1945) and Robel (1967) found that number of males occupying a booming ground varied. Schwartz (1945) found an average of 11.7 males on 61 booming grounds in the 1942 spring booming season in Missouri, and an average of 12.1 males on 187 booming grounds in 1943. The number of males attendingany individual booming ground ranged from two to 42._ Arthaud (1968) found that in 1967 in Missouri, number of cocks using a single booming ground ranged from two to 62 and between 1958 to 1967 the average number of males per booming ground ranged from 7.6 to 15.0. Schroeder and Braun (1992a) found the numbers of males visiting leks -ranged between two and 28 il1 northeastern co Loxado, and average attendance of males was 10.8 for leks active at least six consecutive years (N = 20). During territorial displays, males charge forward a short distance, stop, and stamp their feet often rotating in a half circle. They stoop forward with tail feathers erect, orange air sacs and eye combs inflated, arid pinnae erected over the head. The booming calls are made as the air sacs are inflated, with the tail feathers often clicked during booming (Schwartz 1945). Territorial disputes can often be intense. Males will face one another and -make a series of vocalizations and display postures. Actual fighting will often include short jumps into the air and striking with the feet, beak, and wings. Display behavior of males on leks intensifies when females begin attending (Robel 1967). The frequency of booming increases, and the eye combs are enlarged and pinnae fully erected. Males may also exhibit extreme 4 �117 posturing, and may "flutter-jump" to attract hens. Territorial boundaries often break down when hens are on leks. Males will rush toward them, ignoring territories, and position themselves around the female (Ammann 1957). Arthaud (1968) reported the earliest observed visits by females on booming grounds in Missouri was 14 March with peak hen attendance between 6 and 12 April. Svedarsky (1979) reported peak female attendance for 1975, 1976, and 1977 in northwestern Minnesota to be about 12 April. Most mating takes place during this time, although several peaks of female attendance often occur. For' example, Hamerstrom and Hamerstrom (1973) reported peak hen attendance between 18 and 26 April in Wisconsin, and a lesser peak between 7 and 13 May. Additionally, Watt (1969) reported peak female attendance in northeastern Kansas to be during the first two weeks of April; however he also reported a second, lesser peak of female attendance during the second and third weeks of May. The later peaks are thought to represent renesting attempts for females that lost their initial nest (Watt 1973, Svedarsky 1979). Mating of Greater Prairie-chickens is promiscuous with ,dominant males typically holding the largest territories and accounting for the majority of copulations (Robel 1966). Copulation generally occurs after the male has boomed. Copulation is often preceded by a display in which the male stops booming after circling a female, spreads his wings, and lowers his bill to the ground while keeping his pinnae erect (Sharpe 1968). This display is called the "nuptial bow" (Hamerstrom and Hamerstrom 1960). If the female is receptive, she will squat with wings slightly spread, head raised, and neck outstretched. The male will mount the female by grasping the nape of her neck, and lowering his wings on both sides. Copulation is brief, and females usually run forward a few steps, and stop to ruffle their feathers. Males do not have a specific postcopulatory display_except they begin booming again within a few seconds (Sharpe 1968). Hamerstrom and Hamerstrom (1973), and Svedarsky (1979) note the majority of copulations generally occur a few days after peak hen attendance. Nesting Females begin nesting activities during the peak of the mating display period. Svedarsky (1979) reported an average date of nest initiation of 24 April for a 3-year period in Minnesota. Watt's (1969) estimated dates of laying of the first egg for 21 observed nests in Kansas ranged from 11 April to 7 June with seven of. the clutches being started between 11 and 27 April. Svedarsky (1979) observed the laying rate for five hens to be one egg per day. Hamerstrom (1939) reported Greater Prairie-chicken clutches in Wisconsin averaged 12.0 eggs over six years with a range of 5-17 eggs. Silvy (1968) reported an average clutch size of 12.0 eggs in Kansas with a range of 7-15 eggs. Schwartz (1945) reported a range of 8-25 eggs for 55 nests in Missouri. Greater Prairie-chickens will renest up to two times (a total of three nesting attempts). ·.Renesting hens typically have smaller clutch 'sizes (Silvy 1968). Watt (1969) found,that clutches laid before 1 May (probable first nesting attempts) averaged 13.8 eggs, and clutches laid after 1 May (probable second nesting attempt) averaged 10.3 eggs. Svedarsky (1979) found an average of 14.6 eggs for first nests and 12.8 eggs for probable second nests in Minneso- ta. Females often nest closer to a booming ground other than where copulation occurred (Robel et ale 1970b, Svedarsky 1979). In colorado, Schroeder (1990) found 66 of 89 hens (74%) nested closer to a lek other than the lek on which they were captured and 67 of 79 hens (85%) visited> 1 lek during the breeding season. Incubation is usually 21-23 days (Watt 1973). Svedarsky (1979) found the average length of incubation to be 25.5 days. He also noted that of 246 eggs, 226 were fertile (91.9%). Silvy (1968) reported 100% fertility of 56 eggs in four nests. Hamerstrom (1939) reported nesting success of approximately 50%, while Svedarsky (1979) reported nesting success of 62.4%, and Watt (1969) had 4 of 20 nests (20%) successfully hatch. Schroeder and Braun 5 �118 (1992b) reported nesting northeastern Colorado. success to be 40.5% for Greater Prairie-chickens in Young Greater Prairie-chickens are precocial (Schwartz 1945). The nest is generally abandoned by the hen and chicks within 24 hours after the last chick has hatched aohnsgard 1983). Chicks feed on small insects rather than on herbaceous material (Schwartz 1945). When chicks are 3-4 weeks of age they are capable of flight for short distances (Schwartz 1945). Chicks typically stay with the brood 6-8 weeks, after which they gradually separate Johnsgard 1983). In contrast, Bowman and Robel (1977) reported that six of eight radiomarked chcks dispersed from the brood behveen 10 and 14 weeks of age. Mortalitv Hamerstrom and Hamerstrom (1973) found an average annual mortality rate of 54% in Wisconsin when combining all ages and sexes of Greater Prairiechickens. Svedarsky (1979) found that brood mortality of radio-marked hens in Minnesota was high with broods surviving an· average of 23.9 days. He attributed this high mortality to unfavorable precipitation at time of hatching, long movements, disturbance caused by the researcher, lack of quality brood habitat, and predation. Bowman and Robel (1977) also found high mortality among juveniles in Kansas with 58% mortality in the first 10 weeks after hatching. They found that juveniles were vulnerable to predation during fall dispersal of broods due to their large movements. Berger et ale (1963) found that predation of Greater Prairie-chickens on booming grounds was low with only three cocks being killed in 1,379 encounters w.ith raptors. Movements Hamerstrom and Hamerstrom (1949), working without the aid of radiotelemetry, found that movements of Greater Prairie-chickens in Wisconsin varied seasonally with the least amount of movement occurring during summer and the greatest movement during fall. Robel et ale (1970b) conducted an extensive study on movements and rangs of Greater Prairiechickens in Kansas from 1964 to 1968 in which they obtained 2,229 successive days of location data. Their findings supported the Hamerstrom's study. Typically, the least amount of movement occurs during the summer when birds are molting and hens are rearing broods. Males are typically solitary in summer or in small flocks. However, during fall, larger, sexually segregrated flocks are formed and longer movements are made, espedally by juveniles. Robel et ale (1970b) also found that females averaged larger daily movements than males (738 vs. 603 m) during fall. In winter, large flocks of both sexes form, with some flocks containing as many as 350 birds (Schwartz 1945). Movements of adult males were greatest in February when they were moving to booming grounds (maximum range of 513 ha in Mar). However, home range size of males in April and May decreased to 108 ha and 37 ha, respectively. Females had mean ranges in March, April, and May of 182, 192, and 234 ha, respectively. In Colorado, the average date of spring migration was 27 March for females and 20 February for males, while the average date of autumn migration was 4 July for females and 28 July for males. Females had an average migration distance of 10.58 km and males a distance of 2.86 km (Schroeder 1990). Habitat Grassland habitat is essential for populations of Greater Prairie chickens in North America (Schwartz 1945, Hamerstrom et ale 1957, Christisen 1969, Hamerstrom and Hamerstrom 1973, Kirsch 1974). Historically, Greater Prairie-chickens inhabited mid- and tall-grass prairie intermixed with oak (Quer~us spp.) woodland in the central and eastern Great Plains of North America (Aldrich 1963). Much of the decline of the Greater Prairie-chicken has been attributed to the conversion of native grassland to cropland (Hamerstrom and Hamerstrom 1961, Christisen 1969). 6 �119 In Colorado, habitat for Greater Prairie-chickens was described as "unique" (Schroeder and Braun 1992b) when compared with prairie-chicken habitat in other states due to the prevalence of sand sagebrush (Artemisia filifolia) and small soapweed (Yucca glauca). Evans (1964) noted that in the absence of suitable tall grass (Andropogon hallii, A. scoparious, Calamovilfa longifolia, Eragrostis trichodes, Panicum virgatum, Sporobolus cryptandrus) cover for protection, sand sagebrush may serve as a substitute. Schroeder and Braun (1992b) found that Greater Prairie-chickens selected sagebrush habitats that had a strong component of grass. They also found the presence of cereal grains, specifically corn, was important for Greater Prairie-chickens in Colorado, especially in times of severe winter weather, because it provided a consistent supply of food. Hamerstrom et al. (1957) and Kirsch (1974) suggested that'nesting and broodrearing habitat is the main factor limiting populations of Greater Prairiechickens. Jones (1963) described the nesting habitat of Greater Prairie chickens to be taller and denser than normal for the mid to tal1grass community with mean height of vegetation at 45.7 cm. Svedarsky (1979) found that in northwestern Minnesota, brome (Bromus spp.) and redtop (Agrostis spp.) were preferred for nesting (mean 100% visual obstruction readings of 27 and 18 cm, respectively) (Robel et al. 1970a) while aspen (Populus spp.) and cropland were avoided. Schroeder and Braun (1992b) found that female Greater Prairie-chickens in northeastern Colorado selected nest sites with strong components of sand sagebrush and mid-tall grass. Hens selected against short/mid-grass and cropland. Nest sites had a mean height-density index of 5.9 dm and average heights of sand sagebrush, grass, and forbs were 83.79, 111.32, and 82.64 cm, respectively. Svedarsky (1979) found that 24 week old broods preferred alfalfa and avoided cropland. Schroeder and Braun (1992b) found that in Colorado, females with broods used areas with greater heightdensity indices, species richness, and taller sand sagebrush, grass, and forbs than females without broods. Booming grounds of Greater Prairie-chickens are generally on areas of high ground (i.e., open ridges, grassy knolls) where vegetation is low and sparse and visibility is good (Schwartz 1945, watt 1973). Jones (1963) found the mean height of vegetation on Greater Prairie-chicken booming grounds was 15.1 cm. Schwartz (1945) noted that prairie-chickens can be persistent in their use of a 1ek .site with two grounds in Missouri being used for 40 and 20 years, respectively, despite being mowed, cultivated, and grazed. There is some evidence that Greater Prairie-chicken males respond positively to mowing or burning of leks in that numbers of males on the leks may increase after alteration (Anderson 1969). In Colorado, habitat at leks consisted of short, sparse cover with low species richness. Leks were located on relatively low, overgrazed ridgetops with reduced slopes (Schroeder and Braun 1992b). The minimum amount of grassland habitat needed to support Greater Prairiechickens is unknown (Kirsch 1974). He believed the goal Of prairiechicken management should·be to maintain populations of at least 100 booming males per 1.6 km2 of managed habitat. Kirsch (1974) recommended 518 ha as a minimum for prairie-chicken management units. These units need not be contiguous although he suggested a minimum block of 65 ha and that all blocks be within 12.8 km2. Hamerstrom et al. (1957) suggested 1,036 ha as the minimum area for management in Wisconsin. Arthaud (1968) concluded that in Missouri, the minimum amount of land needed to support prairie-chickens was 256 ha, with a minimum of 64 ha of suitable nesting/brood-rearing habitat. There have been a number of studies which indicate a relationship betweeen populations of Greater Prairie-chickens and the proportion of grassland to cropland (Stempel and Rodgers 1961, Yeatter 1963, Arthaud 1968, Hamerstrom and Hamerstrom 1973, Kirsch et al. 1973, Schroeder and Braun 1992b). Mitchell (1984) suggested that when cropland exceeded 40% of the total area, prairiechickens began to decline. 7 �120 Transplants Since 1950, at least 26 attempts have been made to re-establish Greater Prairie-chickens within their former range (Toepfer et al. 1990). The majority of these attempts have met with minimal success (Kruse 1973, Toepfer et ale 1990). Many researchers have noted that dispersal of birds from release sites contributed to the lack of success of trans locations (Hamerstrom and Hamerstrom 1951; Jacobs 1959; Toepfer 1976, 1988; Wooley 1985; Hoffman 1986; Lawrence and Silvy 1987). Release methodology may be a significant factor contributing to ,success or failure of a translocation. Several attempts have been made to release pen-reared gallinaceous birds into the wild (Hessler et al 1970, Roseberry et ale 1987, Toepfer 1988). Although this method allows for release of large numbers of birds, pen-reared birds have high mortality (Toepfer et ale 1990). Another method is to trap wild birds and hold them temporarily until release (Rodgers 1992). However, Toepfer (1988) noted that penning of wild birds is expensive and leads to weight loss and muscle atrophy due to lack of flight exercise. Several attempts have been made to transplant birds during summer (Toepfer et at , 1990). The logic in transplanting during summer is that it Iriinimizes dispersal from release sites because birds ,are molting and sexually inactive (Toepfer et ale 1990). The easiest and most widely used method for trans locating prairie grouse is to capture birds on display grounds during the breeding season. Schroeder and Braun (1991) found they could capture substantial numbers of both male and female Greater Prairie-chickens in a short period of time with minimal disturbance to leks. ' Griffith et ale (1989) and Toepfer et ale (1990) suggest the amount of quality habitat is the ultimate factor in determining whether a translocation effort will be successful. Translocations which have had the greatest success are those in which grassland habitat has been restored and intensively managed for several years before prairie-chickens were released. In Colorado, 76 Greater Prairie-chickens were trapped on leks during 1984 (36) and 1985(40) and released on the 1,520-ha Tamarack State Wildlife Area which was managed intensively between 1978 and the release date by burning, tilling, grass reseeding, discontinuing grazing, and mowing artificial leks (Hoffman et ale 1992). Leks were established (15 between 1984 and 1991) and reproduction and recruitment were documented. 8 �1' ~ CHAPTER III METHODS Study Site Descriptions Releases of Greater Prairie-chickens were conducted at two separate sites. The Wells Ranch site was 26 kIn east of Greeley, in Weld County, Colorado (553300 m east, 4475280 m north). The Pinneo site was 19 kIn southeast of Brush, in Washington County, Colorado (634240 m east, 4447900 m north). The landscape at both sites was level to gently rolling sandhills Slopes were less than 10% with the majority 2 to 3% sloping in any direction. The elevation ranged from 1,200 to 1,500 m. The habitat at both areas consisted of a mixture of grassland and sand sagebrush with the Pinneo area having a much stronger component of sand sagebrush. Warm-season and cool-season grasses were present at both study sites with the most abundant species being blue grama (Bouteloua gracilis), needle-and-thread (Stipa comata), prairie-sandreed (Calamovilfa longifolia), and sand dropseed (Sporobolus cryptandrus). Other species present were wheatgrass (Agropyron spp.), bluestems (Andropogon spp.), sandhill muhly (Muhlenbergia pungens), Indian ricegrass (Oryzopsis hymenoides), red threeawn (Aristida longiseta), switchgrass (Panicum virgatum), and hairy grama (Bouteloua hirsuta). Agriculture was intermixed in both areas with corn, wheat, and alfalfa being the most abundant crops. The South Platte River lies along the southern edge of the Wells Ranch area and, farther east, borders the northern edge of the Pinneo study site. Cornfields occur along the river at both sites. Ranching and livestock grazing occur at both Pinneo and Wells Ranch. Grazing levels range from light to overgrazed. A number of potential predators occur in the study areas including: Coyote (Canis latrans), Golden Eagle (Aquila chrysaetos), Ferruginous Hawk (Buteo regalis), Red-tailed Hawk (Buteo jamaicensis), Rough-·legged Hawk (Buteo lagopus), and Great Horned Owls (Bubo virginianus). _Potential competitors include Ring-necked Pheasants (Phasianus colchicus). Trapping and Release Methodology Transplant stock for the Wells Ranch site was obtained in Cowley County in southcentral Kansas near Winfield, and stock for Pinneo was obtained in Yuma County, Colorado near Wray. Greater prairie-chickens were captured on leks using walk-in traps (Schroeder and Braun 1991). Captured birds were classified to age by examining patterns of feather wear (Ammann 1944) and to sex and marked with serially numbered aluminum bands on _the right leg and red (Pinneo, 1991), .green (Wells Ranch, ·19.91), yellow (Pinneo, 1992), or white (Wells Ranch, .1992) plastic bandettes on both legs. Individuals caught near Wray were held in burlap sacks and placed in ventilated cardboard boxes, transported.by truck, and released at Pinneo within 15 hours of capture. Individuals caught in Kansas were tested for Mycoplasma spp. to ensure release of healthy birds. Birds were held in ventilated wooden boxes, transported by truck, and released at Wells Ranch within 7 days of capture. The releases were completed between 2 and 9 April 1991, and between 2 and 8 April 1992. In 1991, 43 birds (23 hens, 20 males) were released at Pinneo, while 50 (23 hens, 27 males) were released at Wells Ranch. In 1992, 41 birds (22 hens, 19 males) were released at Pinneo, while 50 (27 hens, 23 males) were released at Wells Ranch. Radiotelemetry In 1991, ~4 birds (six hens and six males at each site) were equipped with battery-powered, necklace attached radio transmitters (Holohil Inc., Woodlawn, Ont.), while 25 birds (seven hens and six males at Wells Ranch, six 9 �122 hens and six males at Pinneo) were radiomarked in 1992. Radiomarked prairie-chickens were relocated--using a Telonics TR-2 portable receiver and a hand-held, 3-element yagi antenna. Locations of radio-marked birds, leks, nests, and non-radio-marked birds were recorded to the nearest 20-m interval using Universal Transverse Mercator (UTM) grid coordinates (Grubb and Eakle 1988), and plotted on u.s. Geo~ogical Survey 7.5-minute topographic maps. An attempt was made to locate radio-marked birds via ,triangulation at least two times every seven days (one each from 0000-1200 and 1201-2400 hrs), and by visual observation at least once every 21~28 days. Three or more azimuths were obtained'<1.5 km of target transmitters and at angles-of-incidence >35 degrees and <145 degrees (Schroeder.1991); ·Periodic flights using a Cessna 185 fixed-wing aircraft with strut-mounted antennae and a Telonics receiver/scanner were made to locate missing radio-marked birds. Seasonal and total ranges of all radio-marked Greater Prairie-chickens were calculated using the minimum convex polygon method (Mohr 1947). Lek Surveys Leks were located in March, April, and May 1991-92 by systematic search of suitable habitats using a spotting scope and a parabolic microphone listening device during morning display periods. Counts of male and female Greater Prairie-chickens on all known leks were made one to two times weekly during the morning display periods from March through May 1991 and 1992. Recruitment of juveniles into the population was estimated in March, April, and May 1992 from temporary blinds near leks by counting numbers of birds with and without leg bands. Nesting Studies Greater prairie-chicken nests were located by following radio-marked hens. Clutch size, incubation period, number of eggs hatched, distance to nearest 1ek, nest fate, and renesting attempts were recorded. A nest was considered successful if at least one egg hatched (Svedarsky 1979). Hens with broods were located twice weekly via triangulation to minimize the chance of brood abandonment by the hen. Brood size was estimated at 45 days posthatching by flushing the broods and counting the number of chicks that flew. An attempt was made to calculate juvenile mortality by comparing brood sizes at bi-weekly ~ntervals after 45-days. Habitat Use As birds were flushed, the habitat they were in was classified to one of five categories: sagebrush, grass, grass/forb, agriculture, or trees/residential. Vegetative structure and species composition were measured on leks and at nests using a..Robel pole (Robel et al. 1970a), and 0.1 m2 Daubenmi.jre plots (Daubenmi.r'e-1959)., Two 18-m transects oriented along north-south and east,-west·axes were placed on each lek ~ith the center of each transect coinciding with the approximate center of the lek. Ten point-Intercept locations, 2 m apart, were located along each transect (total of 20 points). A height-density-index (HOI) was recorded from a height of 1 m and at a distance of 4 m ,to one side of each transect for all 20 points. Height-density indices were recorded as the height of vegetation obstructed on a Robel pole (to the nearest 5 cm). A single 25 rn2 circle was also centered on each site (2.82 m diameter). All plant species within the circle were identified and recorded (n of plants listed as 'species richness'). Additionally, heights for sand sagebrush, grass, and forbs were recorded to the nearest 5 cm. The Daubenmire plot frame was placed in the vegetation at each of the 20 point-intercept locations along the.transect and canopy cover was estimated for each plot. Parameters measured included percent canopy cover of shrubs (sand sagebrush), grasses, forbs, and bare ground~ Vegetative structure and species composition at random Prairie-chicken use sites were measured in 1992 as described nests. At Greater Prairie-chicken use sites, transects were 10 and Greater for leks and centered on the �,', 1_..) point at which the bird was observed. Random sites were chosen relative to observed sites within a 0.5 kIn circle (0.5 kIn is representative of a typical prairie-chicken flight distance) centered over the observed site. Both sites were sampled on the same day to eliminate possible problems associated with measuring habitat variables in a rapidly changing environment (i.e~, p Larrc growth, grazing pressure). An attempt was made to sample two random ·and two use sites per month for each radio-marked bird on each study area. statistics Descriptive statistics and the WI.lcoxon matched-paiiif, -signed-:rank test (SAS program) was used to analyze weight, movement, and habitat data. A significance level of < 0.05 was used. CHAPTER IV ., RES~TS Trapping and Release Methodology A total of 184 Greater Prairie-chickens was captured using walk-in traps and released during early April 1991 and 1992 (Appendix A-D). It generally took three to five days to capture 40 to 50 birds for each study site. Birds trapped in Kansas and released at Wells Ranch were held an average of 4.4 days (range 2-7) while birds trapped in Colorado and released at Pinneo were held less than one day «15 hours). Five birds died as a result of trapping and handling. In 1991, one male trapped in Colorado died of asphyxiation due to a radio-collar that was attached too tightly. No birds .. trapped in Kansas in 1991 died due to trapping procedures. In 1992, two males trapped in Colo~ado died due to heat stress within two hours of capture. Two males trapped in Kansas in 1992 were unable to fly upon release at Wells Ranch. These males were held in captivity and died within two days due to dehydration and. weight loss. Means, ranges, and 95% confidence intervals for weights of Greater Prairie-chickens trapped and released at Pinneo and Wells Ranch in 1991 and 1992 varied (Figure 3). There were differences (P < 0.0001) between weights of males and femal@s released at Pinneo and Wells Ranch in 1991 and 1992 • Weights of males released at Pinneo were heavier than those of males released at Wells Ranch in 1991 (P = 0.0001) and 1992 (P = 0.052). However, there was no difference in weights of females released at Pinneo and 'Wells Ranch in 1991 (P = 0.5667) or 1992 (P = 0.3605). Weights of males released at Wells Ranch in 1991 = 958 g) were less than males released at Wells Ranch in 1992 = 1025 g) = 0.0001). However, there were no differences between weights of males released at Pinneo in 1991 = 1040 g) and 1992 = 1056 9i P = 0.4107). There were no di'fferences in weights of females between sites oz between years. (x (x (x (x (x Movements A total of 441 locations of radio-marked Greater Prairie-chickens were. recorded at Pinneo and Wells Ranch between 1 April and 30 September, 1991 and 1992 (Tables 1-4). In 1991, seven birds (three males, four females) accounted for 91.7% of the locations at Pinneo (Table 1), and five birds (three males, two females) accounted for 87.2% of the locations at Wells Ranch (Table 3). In 1992, seven birds (four males, three females) accounted for 90.9% of the . locations at Pinneo (Table 2), while at Wells Ranch one male accounted for 40% of the locations and twelve birds (five males, seven females) had three locations or fewer (Table 4). 1991 The mean maximum dispersal distance (MMDD) for birds at Wells Ranch in (17.19 kIn) and 1992 (9.78 kIn) was greater than at Pinneo (1991-8.42 kIn, 11 �-0 81 1 1991 1992 WELLS RANCH T U o1 1991 1992 WELLS RANCH Figure 3. Means, ranges, and 95% confidence intervals for weights of female (top) and male (bottom) Greater Prairie-chickens released at Pinneo and Wells Ranch, Colorado, 1991-92. 1992-8.64 km), but not significant for either y.ear (1991 P = 0.3595, 1992 P = 0.7290) (Figure 4). Females generally dispersed farther than males, but this was only significant for Pinneo in 1991 (P = 0.0277). There was no difference in dispersal of males between sites in 1991 (3.93 km at Pinneo vs. 11.22 km at Wells Ranch, P = 0.0988), or 1992 (11.15 vs. 5.43, P = 0.3358). There was no difference in dispersal of females between sites in 1991 (Pinneo - 13.80 km, Wells Ranch - 23.16 km, P 0.4633), or 1992 (6.13 km vs. 12.45 km, P 0.1495). = = 12 �125 Table 1. Movements of Greater April-3D September 1991. released Prairie-chickens Freq. Sex Age 0018 0048 0079 0109 0280 0301 0554 0896 0906 0968 0976 1096 M M M M M M F F F F F F 12+ 2+ 12+ 12+ 12+ 2+ 11- Table 2. Movements of Greater 1 April-3D September 1992. N radio locations 21 4 20 3 1 12 21 14 16 0 3 17 Prairie-chickens :;,'- ..••._ Freq. Sex Age 0190 0268 0309 0349 0369 0640 0208 0228 0246 0286 0330 0520 M M M M M M F F F F F F 2+ 2+ 112+ 1112+ 12+ 2+ Table 3. Colorado, Movements of Greater 1 April-3D September Freq. 0089 0319 0339 0349 0410 0419 0288 0310 0369 0378 0440 0449 Sex Age M M M M M M F F F F F F 2+ 2+ 12+ 112+ 12+ 2+ 11- _ Colorado, 1 Maximum distance from release site (kro) Distance from release site where bird localized (kro) 6.8 2.8 5.9 1.5 1.5 5.1 14.8 17.1 4.5 4.9 2.5 14.1 15.4 2.1 17.1 15.5 14.0 released .. at Pinneo, 4.9 at Pinneo, Colorado, N radio locations Maximum distance from release site (kro) Distance from release site where bird localized (kro) 19 16 25 2 17 2 7 3 27 21 34 2 5.5 5.5 5.0 30.3 17.7 2.9 12.5 1.4 6.0 5.8 3.4 7.7 3.8 2.5 1.9 Prairie-chickens 1991. N radio locations 1 0 22 21 3 12 5 2 3 24 16 0 13 released 4.1 3.9 2.3 5.2 1.3 at Wells Maximum distance from release site (kro) Ranch, Distance from release site where bird localized (kro) 4.4 16.2 15.6 4.3 15.6 78.1 7.7 2.3 15.4 12.3 13.7 15.0 15.0 77 .9 1.7 14.2 11.9 �Table 4. Colorado, Movements of Greater 1 April-3D September Freq. Sex 0070 0408 0479 0560 0600 0619 0049 0088 0280 0440 0460 0500 M M M M M 0539 - N radio locations Age 2+ I2+ l1- I- I F F F F F F 2+ 12+ 2 2 1 I- I 2+ 2+ 1 0 f I- I 20 released • MALES ~ FEMALES at Wells Ranch, Maximum distance from release site (km) Distance from release site where bird localized (Jan) 2.2 16.7 0.8 16.4 6.1 0.4 9.5 22.0 23.1 10.7 6.8 1 I 1 ID 3 M 25 ~ ~ ~ Prairie-chickens 1992. 12.7 . 2.6 U 15 Z ~ 10 C/) 5 E-< 0 0 PINNEO 1991 PINNEO 1992 - WELLS WELLS RANCH RANCH 1991 1992 Figure 4. Mean ~aximum dispersal distance of Greater Prairie-chickens released at Pinneo and Wells Ranch, Colorado, 1991-92. Dispersal from the release site at Pinneo in 1991 was oriented west and north. Four hens (1096, 0976, 0896, 0554) moved approximately 14 km north. Two males (0079, 0018) localized approximately 5 km west of the release site where they were observed displaying on lek #2. One female (0906) localized within 1.5 km of the release site while the final hen (0968) was not located by aerial or ground search after release. One male (0301) localized within 1 km of the release site and was observed displaying on lek #1 but later moved 5 km west of the release site near males 0079 and 0018. Two males (0280, 0109) were found dead within 1.5 km of the release site within 30 days post-release. The final male (0048) lost his radio 0.8 Jan from the release site. The radio was found on 23 April 1991. At Wells Ranch in 1991, dispersal was oriented east, south, and southeast. Four birds (3 males, 1 female) localized 16 km east of the release site in an agricultural area of wheat and summer fallow. All 3 males (0339, 0349, 0419) were observed displaying on leks. Bird 0089, an adult male, was found dead on 18 April 1991, 3.7 Jan west of the release site. On 20 April, 14 �1-')I 0410, a yearling male, was found 3.8 km west of the release site. The bird did not fly when approached, but tried to run away. He was subsequently pursued and caught by hand. Upon examination, the bird was found to have an injured right wing. The bird was released where it was captured on the same day. On 25 April the radio collar of this bird was found 40 m west of its previous capture site. No evidence of predation could be found. Bird 0319, an adult male, was located by aerial search on 23 April approximately 5 km southeast of the release site, however, ground contact was not made with this bird. The second aerial search on 29 May also failed to locate this bird. Bird 0449, a yearling female, was not located after release by air or ground search. Bird 0310, a yearling female, was found dead on 18 April, 1 km west of the release site. Bird 0369, an adult female, was located by aerial search on 23 April, 26 km east of the release site. On 30 April, this same bird was located 0.7 km south of the release site. Contact was lost with this bird after 3 May. Bird 0440, a yearling female, localized 11.7 km southeast of the release site. The maximum distance any bird was located from the release site was 78 km. This bird was an adult female (0288) released at Wells Ranch. She was located by aerial search on 29 May at the Pinneo study area. She was found dead on 20 June. It is unknown if she was nesting. In 1992 at Pinneo, dispersal was greatly reduced due to birds from the 1991 release "holding" the newly released birds in the area. Eight birds (4 males, 4 females) localized within 6 km of the release site (Table 2). Bird 0369, an adult male, was found 17.7 km east-southeast of the release site on 30 May 1992, however, this bird moved back near the release site and was 1.1 km west of the release site on 16 June. Bird 0349, a yearling male, was 4.3 km north of the release site on 4 May 1992, but was found dead on 13 May, 30.3 km north northeast of the release site. Bird 0640, a yearling male, was found dead on 26 April 1992, 1.3 km from the release site. Bird 0208, a yearling female, nested 3.9 km west of the release site, but the nest was depredated on 5 May. This hen was found dead on 30 May, 12.5 km southeast of the release site. The radio-collar of bird 0228, a yearling female, was found on 12 April, 1.4 km from the release site. Only three locations were obtained for this bird. Bird 0520, an adult female, was found dead on 20 June, 7.7 km west of the release site. Only two locations were obtained for this bird. At Wells Ranch in 1992, dispersal was high. Contact was not made with 0500 (adult female) by ground or aerial search. contact was lost with 0440 (yearling female), 0460 (adult female), and 0600 (yearling male) within 30 days post-release. However, 0440 was found on 15 August, 10.7 km east southeast of the release site. Three birds (1 male, 2 females) slipped their radio-collars within 45 days post-release. Bird 0479's (adult male) collar was found 0.8 km from the release site, while 0088's (yearling female), and 0049's (adult female) radio-collars were found 22.0 km southeast and 9.5 km northwest of the release site, respectively. Five birds (3 males, 2 females) were depredated within 2 months post-release. Bird 0070 (adult male) was found dead on 10 April, 2.2 km from the release site. Birc;i0619 (yearling male) was found dead 0.4 km from the release site on 13 April. Bird 0408 (yearling male) was found dead on 31 May, 16.7 km east of the release site. Bird 0280 (adult female.) was found dead on 8 May, 23.1 km southeast of the release site. Bird 0539 (yearling female) was found dead on 31 May, 2.6 km east of the release site. Movements between summer range and winter range varied. In 1991 at Pinneo, female 1096 moved 11.2 km from her summer range to her winter range, 25.6 km from the release site, while in 1992 at Pinneo three males (0190, 0268, 0309) moved 6.0, 5.6, and 4.0 km respectively, to their winter range, 2.0 km from the release site. These three males wintered together in the same flock of 40+ birds. At Wells Ranch in 1991, one female (0378) moved 12.9 km to her winter range, 26.0 km from the release site, while in 1992 one male (0560) moved 10.5 km to his winter range, 23.1 km from the release site. The winter ranges of all birds at both sites centered around grain fields, specifically corn. 15 �128 Birds at Pinneo had slightly larger (9.9 krn2) home ranges than birds at Wells Ranch (9.2 krn2). Males had larger home ranges than females at both sites (Table 5). Females without broods at Pinneo had larger home ranges (8.4 km2) than females with broods (3.4 km2). The calculated home ranges of the birds do not include the dispersal distances from the release sites. Table Wells 5. Mean home range size (km2) of Greater Prairie-chickens Ranch Colorado, 1 April - 30 September 1991 - 1992. Pinneo Category N Males Females With broods Without broods Total 7 8 4 4 15 Wells X 14.4 5.9 3.4 8.4 9.9 at Pinneo and Ranch SD N X 13.2 5.5 2.8 6.8 10.4 4 2 2 0 6 13.2 1.3 1.3 20.4 0.40 0.40 9.2 ,-".,,,,;~i''''I'6 • 9 SD Lek Establishment seventy-seven mornings were spent searching for and observing leks during spring 1991 and 1992. Leks were numbered in order in which they were found at each site. Fourteen leks were documented (nine at Pinneo, five at Wells Ranch) during 1991-92 (Table 6, Figures 5 and 6). Two of nine (22.2%) leks at Pinneo were active for both years while one of five (20%) were active both years at Wells Ranch. There was some instability in lek locations at Pinneo in 1991 as lek #1 only formed for several weeks in April before it broke up, while lek #3, although it had four males in attendance, was found to be active on only one occasion. Following the breakup of lek #1, male 0301, which had been in attendance at this lek, moved to the vicinity of lek #2, but was not seen displaying on this lek. Male 0339 at Wells Ranch, was seen displaying on two different leks (#2 and #3) within two days in 1991. In Table 6. Maximum attendance of male Greater Prairie-chickens at Wells Ranch and Pinneo, Colorado, 1991-92. Universal Transverse Mercators (Zone 13) Lek 1 2 3 4 5 6 7 8 9 1 2 3 4 5 East (m) 633700 629960 632960 630120 630680 633560 633420 636380 630620 553340 565260 568260 568260 557040 North (m) Legal Township Range sites Male attendance description Section at display Quarter 91 92 4447300 4448020 4446800 4461700 4461280 4448540 4446680 4451600 4450840 2N 2N 2N .4N 4N 3N 2N 3N 3N Pinneo 54W 55W 54W 54W 54W 54W 54W 54W 54W 4 1 5 19 19 33 4 26 30 NW NE SE SW SW SW SW SE NW 4 4 4 7 3 0 0 0 0 0 0 0 9 5 7 3 3 3 4473080 4476940 4477200 4476800 4476520 Wells Ranch 5N 63W 6N 61W 6N 61W 6N 61W 6N 62W 7 32 34 34 33 SW SE SE SE SE 6 6 4 0 0 5 0 0 6 3 16 �4.!c::0CO _.---------------.-_._------------.--_ __ , 4460000 . N , ---------------'----------------'----.-.----.----...•. .-.... - Eoc:::: o z ::2: - 4455000 _ .•••.•.•••.•••••••••.•••••.• J_ , ••.•••••••.• _ •••••••••••••• _,_ .•••.•.•••••.•.•.•.•••••••.••• ' .: , ~ 4450000 . . , --------------,----------------r---------------, 6 • ~ 3•• 0 RELEASE SITE •• 7 4445000 625000 Figure S. 1991-92. Locations of leks in relation 1992, this same male was active the entire display season. Pinneo 635000 630000 to release on a new lek 640000 site at Pinneo, Colorado, (#4), 400 m south of lek #3, for 1991 A total of five leks formed, averaging 4.4 males/lek (range 3-7), 7.01 km from the release site at Pinneo (range 0.81-14.40) (Table 7). and Lek #1 was found on 17 April approximately 12 days post-release in a disturbed area 0.81 km from the release site next to a 2-track road and Conservation Reserve Program (CRP) section. This lek had a maximum of four 17 �4480000 ~ - -. .. - - .- .- - , - . •3 :2 •• ~ ~ 4 • 4475000 o z g g 4470000 _________ • . A. --'----0- . .• .• __ __ .- _ ....•.•. 4465000 -~~~--~~--1--~--~~~---~~--~ 550000 560000 555000 565000 570000 UTM (KM EAST) Figure 6. Locations Colorado, 1991-92. of leks in relation to release Table 7. Number of active leks, mean number distance of leks from the release site based surveys conducted at Pinneo and Wells Ranch, Year Active leks 1991 1992 5 6 1991 1992 3 Males/lek Pinneo 4.4 5.0 Wells 3 Ranch 5.3 4.7 site at Wells Ranch, of males per lek, and mean on greater prairie-chicken Colorado, 1991-92. lek Distance 7.0 6 .•6 10.0 7.1 males in attendance, including radio 0301, through 28 April, but had little or no attendance during May. The early breakup of this lek may have been due to heavy traffic on the road next to the lek and to the quick growth of vegetation on the lek. Lek #2 was found 4.28 km west of the release site with a maximum of four males in attendance including radios 0018 and 0079. The lek was in a low, flat area with short-grass surrounded by sand sagebrush and grass. Lek #3 was found i4 May 1991 in an alfalfa field, 1.69 km southwest the release site with four displaying males in attendance. However, subsequent visits to this site showed no males in attendance. 18 of �131 Although leks #4 and #5 were not found until 3 June 1991, they had seven and three males in attendance, respectively. Both leks were on exposed ridgetops with sparse vegetation, surrounded by sand sagebrush and grass and were 14.40 and 13.85 km north of the release site, respectively. No females were observed on leks at Pinneo in 1991, however, radio 0906 was near lek #1 on several occasions during the morning display period. Unlike other Greater Prairie-chicken populations, males in this translocated population were not observed to display during evening hours. The display season was generally finished by 1 - 10 June, however one non-displaying male was observed sitting on lek #4 as late as 20 June. Wells Ranch 1991 A total of three leks formed averaging 10.02 km (range 2.20-15.08) from the release 5.3 males/lek site at Wells (range 4-6) and Ranch (Table 7). Lek #1 was found within 21 days post-release 2.20 km south of the release site, with six males in attendance. This lek was active through May and was on a grassy knoll with sparse vegetation. Leks #2 and #3 were not found respectively, from the release site 0419, and 0339 were seen displaying 0339 was observed displaying on lek 30 until 4 June, 12.07 km and 15.08 km, with four males on each lek. Males 0349, on lek #3 on 6 June. Two days previously, #2 with three other males. Lek #2 had six males in attendance on 10 June, but by 19 June there were no males in attendance. Lek #3 had four males in attendance on 4 June. Lek #2 was in a fallow wheat field on ~ exposed hilltop surrounded by strips of wheat. Lek #3 was'in a flat, grassy pasture surrounded by wheat fields. No females were observed on leks. However, female 0369 was flushed 2.8 km west of 1ek #1 with a non-radio-marked male on one occasion during the display season. Pinneo 1992 Three leks active in 1991 (#1, #2, and #3) were inactive in 1992. Two leks (#4, #5) active in 1991 were again active in 1992. Additionally, four new leks were formed. Thus, in 1992 there were six active leks averaging 5.0 males/lek (range 3-9) and 6.6 km (0.93 km-14.4 km) from the release site (Table 7). Lek #4 increased from seven males in 1991 to nine males in 1992, while lek #5 increased from three males in 1991 to five males in 1992. At least three males, on lek #4 were unbanded indicating recruitment. Also, five unhanded males formed a new lek (#6) prior to release of birds in 1992. Following release of birds, two yellow banded (1992 release) males succeeded in establishing territories on the periphery of the lek. One copulation was observed with an unbanded hen on this lek, and three copulations with yellow banded hens (including radio-marked female 0330) were observed. This lek was 0.93 km north of the release site in a flat, grassy area surrounded by sand sagebrush. Lek #7 was formed 1.47 km south of the release site on a hilltop in an irrigated alfalfa field with three males in regular attendance (including male 0309) . Lek #8 was located 4.27 km northeast of the release site in a flat, grassy, open area with three males in regular attendance. Males 0349 and 0369 were in the vicinity of this lek on several'occasions, but were not seen displaying on the lek. Lek #9 was also in an alfalfa field on a hilltop 4.66 km northwest of the release site with three males in regular attendance (including male 0190). 19 �132 Wells Ranch 1992 Two leks active at Wells Ranch in 1991 (#2, #3) were inactive in 1992. However, a new lek (#4) was established in the vicinity of leks #2 and #3, with six males in attendance. Male 0339 (1991 release) was in regular attendance on this lek throughout the display season. This lek was in a wheat field with relatively flat topography, 15.04 km east of the release site. Lek #5 was also newly formed in 1992 with three males in attendance. This lek was on an exposed ridgetop with sparse grass surrounded by sand sagebrush, 3.94 km northeast of the release site. Lek #1, which was active in 1991 with six males, was also active prior to the 1992 release with three green-banded (1991 release) males. Following the 1992 release, two white-banded males succeeded in establishing territories on the periphery of this lek. No recruitment was documented for any of the three active leks obser~ed in 1992. Hens were observed visiting leks #1 and #4, but no copulations were observed. Behavior on Leks Behavior on leks was typical, with males establishing and defending territories. Females were not seen on any leks in 1991, however, when hens were observed to visit leks in 1992, four copulations were documented. The nuptual bow was observed on one occasion on lek #6 at Pinneo in 1992. Numerous flutter jumps were also seen while observing leks. When females did visit leks, territories of males broke down. The males would group around the female, and booming intensity would increase. Intensity of display decreased in May and June and attendance was irregular later in the season. Males, released in 1992, attempted to establish territories on leks that were active prior to 1992. Initially, these males were chased off the leks. However, these males were successful in establishing territories on the peripheries of the leks within one week, but they were not observed to copulate with any hens. Two newly released males were successful in establishing territories on the yearling lek (#6) at Pinneo in 1992, but it is unknown if these males were adults or. yearlings. Nest Establishment and Success In 1991, nests of seven radio-marked hens were located (five at Pinneo, two at Wells Ranch) (Figures 7 and 8, Tables 8 and 9). Nest success of radiomarked hens was 80% (4/6) at Pinneo in 1991. At Pinneo, four hens successfully nested and a fifth hen (1096) was on a clutch of 11 eggs until 21 May. She was no longer on the nest after this date, but was still in the area until the end of October. It is unknown if she was successful as her nest was not relocated after she left it, and ~he was not seen with chicks when flushed from 31 May to 31 August. Therefore, it is assumed that female 1096 was unsuccessful. The two radio-marked hens that nested at Wells Ranch were also successful. Average dutch size for seven nests was 11.1 eggs (range 9-16). Clutch size for two hens (0976, 0554) was estimated from shell remains because they were not observed away from their nests prior to hatch. Three eggs failed to hatch (1 of 11 from 0896, 2 of 11 from 0378). Egg fertility was not determined for those eggs which failed to hatch. The mean distance of nests from leks was 1.8 km at Pinneo and 5.6 km at Wells Ranch (Table 10). Five of seven nests were in areas with heavy sand sagebrush and were well hidden within the vegetation. Another nest was in dense, tall grass while the final nest was in a wheat field. Brood counts conducted on 21-22 July revealed that three hens at Pinneo (0554, 0896, 0906) had broods of at 20 �-+-;7'2000 -- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - .. - - . - .. - - - - - - - I i = Nests I ........... -Wi5000 . _ ~ __ . ,,-I _._R_eJ_ea_se_si_r2_ ·N •••• j - -WiOOOD Eoc:::: C Z ~ --::::: 4450000 • 4445000 625000 630000 635000 UTM (KM EAST) Figure 7. Locations Colorado, 1991. of leks and nests in relation to release site at Pinneo, least one, three, and ten chicks, respectively (Table 8). Females 1096, 0440, and 0378 were not flushed with a brood. Brood hens stayed in areas with dense sand sagebrush and grass cover, and high concentrations of insects (i.e. , grasshoppers) • In 1992, five radio-marked hens at Pinneo nested, while no radiomarked hens nested at Wells Ranch (Figure 9, Table 11). Nest success for five radio-marked hens was 37.5\ (3/8) (Table 11). Four nests were depredated by weasels (Must:ela spp.) while the fifth nest was abandoned due to human disturbance (i.e., farming activity). Average clutch size for first nests (n 5) was 11.0 eggs (range 6-13), and 8.7 eggs (range 7-12) for renests (n = 3). Mean distance of nests from leks was 1.1 km. = 21 �4480000 ... °. po • ••• . _. , • Leks o Nests • _ •••••• • Release site -::c 4475000 - _ •• __ •••••• - - _ ••• - - - ,_ - - - - - - ~ - - - - _ • - - - l...- _.l - - - - _ •••• - - - •• • _ _ •• - - - - - t_ • • • _ • _ • _ •••• _ • _, _ •• _ - • - • _' , ~ ~ 0 z ~ ~ , 4470000 •• •••••••• Po. •••• , •• • __ •••• ••• __ • • '. __ • ••• _ •• - • - -, • I -~ ~ ;:J 4465000 - .......... - - '_ , - - - .. - .. - - ..•.................. I - - '- 0 I - '- - - - • ··,__ "---·i_.L....._.__L-- __ J __ 4460000 550000 _ 5600011 555000 ' • 565(KKI j l- __ 57(XKKI UTM (KM EAST) Figure Ranch, 8. Locations of leks and nests Colorado, 1991. in relation to release site at Wells Female 0208 was on a clutch of 12 eggs on 4 May, but the nest was depredated by weasels on 13 May. Following the failure of her nest, she could not be located for several weeks. After an aerial search, she was found dead 16 km southeast of her nest site. Female 0246 was unsuccessful on her first nest attempt due to weasel predation. She renested within 20 days, 840 m from her first nest, but her second ne~t was also depredated by weasels. She did not attempt to nest a third tLme. Female 0330 was also unsuccessful on her first nest due to weasel predation on 23 May. She renested on 7 June, 620 m from her first nest, but this second nest was abandoned on 12 June due to human disturbance. She attempted a third nest on 16 June, 610 m from her second nest and 80 m from her first nest. This third nest was successful as four of seven eggs hatched on 10 July. She was flushed with all four chicks on 12 August. Female 0286 was also successful with four of six eggs hatching on 28 June. Female 0554 (1991 release) was successful for the second consecutive year. She nested 360 m south of her 1991 nest site, with eleven of twelve eggs hatching. The one egg that did not hatch was fertile; the chick died during pipping of the shell. Fertility of eggs that did not hatch in other nests was not determined. 22 �135 Table 8. Colorado, Female Nesting 1991. Age activity Date of nestingll of Greater Date of hatching Prairie-chickens Clutch size released N eggs hatched 10 0976 1- 31 May 11 Jun 0906 2+ 10 May 8 Jun 16 16 0896 1- 31 May 27 Jun 11 10 0554 2+ 31 May 27 Jun 9b 9 1096 1- 13 May Unknownc 11 at Pinneo, Comments Lost contact with hen after 11 Jun. Successful nest. Flushed hen with 10 chicks on 21 Jul Successful nest. Flushed hen with 3 chicks on 21 Jul. Successful nest. Observed hen with 1 chick on 21 Jul. Was not on nest after 31 May. Flushed without any chicks. Unknownc aDate when first observed on a nest. bClutch size was estimated from shell remains because hen was not observed away from nest. cFailed to relocate nest after hen left. Assumed unsuccessful because hen was not seen with chicks when flushed from 31 May to 31 August. Table 9. Nesting activity Ranch, Colorado, 1991. Female Prairie-chickens Age Date of nestinif Date of hatching Clutch size 12+ 17 May 4 Jun 19 Jun 6 Jul 10 11 0440 0378 aDate when Table nests of Greater first observed N eggs hatched 10 9 Freq. at Wells Comments Flushed Flushed without without any chicks. any chicks. on a nest. site distances 10. Nest-lek and nest-release at Pinneo and Wells Ranch, Colorado, 1991-92. Year released Nearest lek (#) of Greater Distance (km) Prairie-chicken Distance from release site (km) Pinneo 1991 1991 1991 1991 1991 1992 1992 1992 1992 1992 1992 1992 1992 0976 0906 0896 0554 1096 0208 0246 0246 0286 0330 0330 0330 0554 1991 1991 0378 0440 5 1 4 5 5 7 7 7 9 6 7 6 5 Wells 5.5 0.6 1.0 0.7 1.2 3.1 0.4 1.1 1.6 0.5 0.8 0.6 0.6 17.1 1.4 14.8 14.0 12.8 3.9 1.3 0.4 5.8 1.1 1.0 1.0 13.6 1.2 10.0 13.3 12.0 Ranch 2 1 23 �- - ') ); 4465C-00 -"" - - - - - - - - - - - - - - -- - - - -""" - --------- ----------------- .A ~ : ________ " 4460000 " " __; .- i 5 :; Nests i I " ~ • Release site 1 ~ -..•. Ec:::: C Z ~ 4455000 ---------------~----------------------------------. . 4450000 o . . ---------------~-----------------~----------------~ . e gW:J. 0 0 :0 IlL 7 4445000 625000 635000 630000 640000 UTM (KM EAST) Figure 9. Locations Colorado, 1992. of leks and nests in relation to release site at Pinneo, Mortality Of the 49 radio-marked Greater Prairie-chickens, 21 (42.9%) were depredated, 12 (24.5%) slipped their radio-collars, and contact was lost with 13 (26.5%) (Table 12). Of the 21 mortalities of radio-marked birds, 15 (71.4%) occurred within 120 days post-release (Table 13). In 1991, mortality at Pinneo was 16.7% within 120 days post-release, and 25% at Wells Ranch 120 days post release (Table 13, Figure 10). In 1992, mortality was 33.3% and 46.2% at Pinneo and Wells Ranch, respectively, 120 days post-release (Table 13, Figure 10). Mortality at Pinneo from 1991-1992 was 37.5% compared with 48% at Wells Ranch. Of the 23_documented deaths (21 radio-marked birds, 2 non-radiomarked birds), mammalian predators (Coyotes) were the suspected cause of death for sixteen (69.6%), raptors caused four (17.4%), and three (13.0%) were lost to miscellaneous causes (Table 14). 24 �137 Table Wells Radio Freq. 11. Nesting activity of Greater Ranch, Colorado, 1992. Age Date of nestinga Clutch size Date of hatch Prairie-chickens released N eggs hatched at Pinneo and Comments Pinneo 4 May 0208 0228 0246 0246 0286 0330 0330 0330 112+ 2+ 2+ 2+ 2+ 4 2 S 4 7 16 0520 0554b 2+ 2+ 13 May I- May Jun Jun May Jun Jun 12 0 10 Jul 13 12 6 12 7 7 0 0 4 0 0 4 27 May 12 11 28 Jun Depredated 13 May. Found radio-collar 12 Apr. Depredated 13 May. Depredated 12 Jun. Not flushed with any chicks. Depredated 23 May. Abandoned 12 Jun. Flushed with 4 chicks on 12 Aug. Found dead on 2 Jun. Not flushed with any chicks. Wells Ranch No nests were documented in 1992. aDate when first observed on a nest. hrhis hen was released in April 1991. and 1992. She was successful in both 1991 A yearling female (0440) released at the Wells Ranch in 1991 was shot by hunters on 27 November 1991, approximately 15 km southeast of the release site. This hen had been in the area since May 1991, and also had a successful nest during 1991. A yearling male (0349) released at Pinneo in April 1992, was found dead in Prewitt Reservoir (apparently drowned) 30.5 km north of the release site during the second week of May. One week previously, this male was flushed 27 km south of the reservoir in association with adult male 0268. On 5 July 1992, a non-radio-marked hen (yellow band #61; aluminum band #1111) released at Pinneo, was found dead on u.S. 34 approximately 800 m east of the Morgan-Washington County line, 9.5 km northwest of the release site. The bird had been struck by a car. Table 12. Fates of radio-marked Greater and Wells Ranch, Colorado, 1991-92. Prairie-chickens Pinneo Wells 1991 Fate Known depredated Slipped collar Signal lost N 3/12 4/12 5/12 25 33.3 41.6 % N 6/12 1/12. 3/12 25 at Pinneo Ranch 1992 1991 1992 % released 50 8.3 25 N 6/12 2/12 4/12 % 50 16.7 33.3 N % 6/13 5/13 1/13 46.2 38.5 7.7 �Table 13. Fates of radio-marked Greater Prairie-chickens and Wells Ranch, Colorado, 1 April - 1 August, 1991-92. Depredated Slipped collar .•..•.. Signal at Pinneo lost Totals .. - Pinneo 1991 1992 1991-92 Wells released 2/12 4/12 6/24 2/12 1/12 3/24 2/12 0/12 2/24 6/12 5/12 11/24 3/12 6/13 9/25 2/12 3/13 5/25 2/12 3/13 5/25 7/12 12/13 19/25 Ranch 1991 1992 1991-92 12 1991 10 en -< 0 8 0 c::: ~ :::.> - fU < 6 0 30 60 90 120 14 1992 21 12 10 8 6~--'--~------~------~~--~ o 30 60 90 120 NDAYS SURVIVED Figure 10. Survival of radio-marked Greater Prairie-chickens post-release at Pinneo and Wells Ranch, Colorado, 1991-92. 26 120 days �Table 14. Causes of mortality of Greater and Wells Ranch, Colorado, 1991-92. Mammal Prai=~e-chicker.s Raptor released Other at Pinneo Totals Pinneo a a 8 1 1 2 2 4 4 8 1 2 3 a 3 5 1991 1992 1991-92 3 8 11 Wells Ranch 1991 1992 1991-92 Habitat 1 1 6 6 12 Use Habitat use varied between sites and between sexes. Birds at Pinneo tended to use more sagebrush habitat throughout the spring and summer, while at Wells Ranch birds switched from sagebrush habitat during the display and nesting period. to grass habitat (CRP) during the rest of the summer (Figures 11 and 12). At Pinneo, males used a combination of sagebrush, grass/forb, and IJ SAGEBRUSH IIGRASSiFORB OGKASS IAGRlCULTIJRE 100 en Z - 80 o 60 ~ 40 U o ,..J 20 o 1 APR lIUN lAUG 1 APR lJUN ixuc 1ocr 1ocr Figure 11.· Distribution of 116 male (top) and 154 female (bottom) Greater Prairie-chicken locations by habitat type at Pinneo, Colorado, 1 April-l October, 1991-92. 27 �o SAGEBRUSH i OCRASS GRASS/FORB I.!,GRICULRRE 100 o: Z .:;. - a t< U 80 60 40 0 ,...J 20 ~ a a 1 APR lJUN 1 APR lJUN lAue Iocr t- Z IJ,J U 100 ~ 80 ~ 60 IJ,J 40 20 0 i xuc 1 OCT Figure 12. Distribution of 66 male (top) and 42 female (bottom) ,Greater Prairie-chicken locations by habitat type at Wells Ranch, Colorado, 1 April October, 1991-92. -1 agricultural are~s in spring, but at the end of the display season and as summer progressed males switched their use to more sagebrush habitats along with some agriculture. Females were in sagebrush habitat 60-80% o·f the locations at Pinneo from April to October. Females were found in agricultural habitats less often than males from April to October, however, females were found more often than males in grass areas throughout the spring and summer. At Wells Ranch" males were found more often than females in sagebrush habitats, however, after the display season males switched to more grassy areas (CRP) for the remainder of the summer. Female use of sagebrush habitat peaked during the nesting period, but they switched to grass habitat for the brood-rearing period. Data for winter (1 Oct-3~ Mar.) habitat use were, pooled for both sites and both years. Of 59 locations obtained during winter, 30 (50.8%) were in sagebrush habitat, 18 (30.5%) were in agriculture (corn), five (8.5%) were in grass areas, three (5.1%) were in disturbed (forbs) areas, and on three (5.1%) occasions birds were found roosting in trees. Habitat was compared for observed and random sites for Greater Prairie-chickens at Pinneo during spring and 'summer of 1992. Habitat was not compared at Wells Ranch in 1992 due to high mortality and dispersal of birds (92% of the birds were gone by 1 Aug.). Of the thirteen variable examined at each site, none'was significant for males or females (P > 0.05) (Tables 15 and 16). Greater Prairie-chicken nest sites had more sagebrush cover (P 0.031) and less bare ground (P = 0.045) than random sites (Table 17). Successful = 28 �nests had taller sagebrush (P = 0.018), with higher species richness (P = 0.006) than unsuccessful nests (Table 18). Height-density indices at observed sites for brood females were less than for observed sites of nonbrood females (P = 0.033) (Table 19). Habitat at leks did not differ significantly from random sites for any of the variables examined (Table 20). Table 15. Habitat at observed (N = 18) and random (N Greater Prairie-chickens at Pinneo, Colorado, 1992. Observed Habitat X variable 18) locations of male Random SD Height-density index, dm 1.41 Height, cm Sand Sagebrusha 55.22 Grasses 70.28 ""...<~~ Forbs 37.06 Cover, % 15.36 Bare Sand sagebrush 10.62 Grasses 49.79 Blue Gramab 16.85 .12.49 Need1e-and-threadc Sand Dropseedd 6.46 Prairie Sandreede 3.94 15.38 Forbs 12.83 Species Richness = X SD P 0.98 1.21 0.79 0.751 29.28 24.05 19.54 53.50 71.39 37.89 31.69 32.07 23.36 0.899 0.950 0.962 9.46 9.18 23.63 13.20 9.20 12.11 4.76 28.96 5.40 19.83 6.89 51.99 17.44 15.71 2.63 6.73 12.01 11.78 20.95 8.62 29.76 14.63 17.09 4.95 7.48 24.29 4.83 0.975 0.101 0.457 0.824 0.962 0.174 0.506 0.516 0.445 aArtemisia filifolia bBouteloua gracilis CStipa comata dsporobolus cryptandrus eCalamovilfa longifolia Table 16. Habitat at observed female Greater Prairie-chickens (N = 14) and random (N = 14) locations at P~nneo, Colorado, 1992. Observed Habitat X variable Height-density index, Height, cm Sand Sagebrusha Grasses Forbs Cover, % Bare Sand Sagebrush Grasses Blue Gramab Needle-and-threadc Sand Dropseedd Prairie Sandreede Forbs Species Richness of Random SD X SD P 1.54 0.68 1.54 0.93 0.713 58.56 76.93 50.86 27.97 21.63 29.73 36.29 63.71 51.71 31.86 42.96 24.87 0.055 0.312 0.713 10.34 11.95 52.96 14.14 15.97 6.76 5.58 19.21 12.07 6.19 11.29 24.25 14.64 13.41 9.78 7.18 27.41 3.83 14.34 7.91 46.84 16.73 9.06 1.52 6.94 20.55 10.71 7.79 10.78 29.18 16.27 11.80 2.67 10.21 32.73 5.81 Q.232 0.355 0.566 0.854 0.090 0.158 0.925 0.999 0.447 dm aArtemisia fillfolia bBouteloua gracilis CStipa comata dsporobolus cryptandrus eCalamovilfa longifolia 29 �Table 17. Habitat at observed (N = 11) and random (N = 11) locations Greater Prairie-chicken nest sites at Pinneo, Colorado, 1991-92. Observed Habitat X variable Random so P 0.78 0.192 X SO 1.50 Height-density index, dm Height, cm Sand Sagebrusha 56.64 79.00 Grasses -Forbs 30.00 Cover, % Bare 6.76 Sand Sagebrush 12.97 Grasses 53.00 6.45 Blue Gramab Need1e-and-threadc 15.39 Sand Oropseedd 11.90 7.82 Prairie Sandreede 5.96 Forbs Species Richness 11.55 of 0.40 1.28 26.39 13.82 23.49 41.20 64.70 28.60 36_.21 25.40 14.62 0.645 0.191 0.672 6.36 9.44 17.72 7.19 13.53 12.89 10.77 5.12 5.85 18.31 4.54 36.33 6.48 13.13 6.67 2.19 20.98 10.50 13.53 6.17 25.09 7.66 14.39 7.09 4.61 27.97 5.91 Q.045 0.031 0.159 0.972 0.341 0.417 0.314 0.181 0.832 sArtemisia filifolia bBouteloua gracilis CStipa comata dsporobolus cryptandrus eCalamovilfa ~ongifolia Table 18. Habitat at observed locations for successful (N = 6) and unsuccessful (N = 5) nests of female Greater Prairie-chickens at Pinneo, Colorado, 1991-92. Nests Successful Habitat X variable Height-density index, Height, cm Sand Sagebrusha -_. Grasses Forbs Cover, % Bare Sand Sagebrush Grasses Blue Gramab Needle-and-threadc Sand Dropseedd Prairie Sandreede Forbs Species Richness dm Unsuccessful SO X SO P -0.714 1.45 0.37 1.56 0.48 71.83 80.17 33.00 19.10 13.24 26.34 38.40 77.60 26.40 22.86 15.93 21.96 0.018 0.714 0.999 7.68 17.97 47.60 5.61 14.75 12.37 5.18 6.42 15.50 8.03 7.65 16.98 4.76 12.35 14.56 10.05 5.92 4.93 5.66 6.97 59.48 7.47 16.15 11.35 10.98 5.41 6.80 4.22 8.21 18.13 9.93 16.30 12.24 11.86 4.59 1.92 0.784 0.068 0.273 0.714 0.999 0.855 0.320 0.999 0.006 sArtemisia filifolia bBouteloua gracilis CStipa comata dSporobolus cryptandrus eCalamovilfa longifolia 30 �143 Table 19. Habitat at observed locations for female Greater Prairie-chickens with (N = 8) and without (N = 6) broods at Pinneo, Colorado, 1992. Hens Brood Habitat variable so X Height-density index, Height, cm Sand Sagebrusha Grasses Forbs Cover, % Bare Sand Sagebrush Grasses Blue Gramab Needle-and-threadC Sand Oropseedd Prairie Sandreede Forbs Species Richness Non-brood X so P 0.46 1.98 0.72 0.033 56.25 78.13 56.25 25.83 15.56 37 •. 58 62.33 75.33 43.67 32.77 29.53 14.56 0.437 0.605 0.747 12.10 8.99 53.60 15.90 20.86 7.73 4.70 20.66 11.63 7.44 10.51 24.12 17.64 13.38 11.45 8.03 21.34 3.54 7.98 15.90 52.11 11.80 9.46 5.48 6.77 17.28 12.67 3.21 12.01 26.71 10.48 11.31 7.85 6.40 36.17 4.46 0.220 0.291 0.796 0.949 0.155 0.792 0.509 0.121 0.605 dm 1.22 aAr~emisia filifolia bBou~eloua gracilis cS~ipa Coma~a dsporobolus cryp~andrus eCalamovilfa longifolia Table 20. Habitat at observed (N = 7) and random Prairie-chicken leks at Pinneo, Colorado, 1991-92 (N = Observed Habitat variable x - Height-density index, Height, cm Sand Sagebrusha Grasses Forbs Cover, % Bare Sand Sagebrush Grasses Blue Gramab Needle-and-threadc Sand oropseedd Prairie Sandreede Forbs Species Richness dm 7) locations of Greater Random so X SO P 0.45 0.27 0.77 0.37 0.141 7.14 47.43 20.86 18.90 31.38 11.95 35.29 65.14 27.14 34.32 33.03 14.36 0.087 0.306 0.370 23.16 0.02 38.44 16.37 10.37 2.14 1.04 23.90 9.71 11.12 0.05 28.53 16.48 11.30 3.00 2.76 32.51 5.44 19.41 4.60 52.76 10.29 22.80 6.37 3.41 12.10 12.00 12.02 7.26 29.70 8.02 17.12 5.97 5.91 18.30 5.13 Q.565 0.074 0.306 0.607 0.155 0.163 0.424 0.482 0.304 aAr~emisia filifolia bBou~eloua gracilis cS~ipa comet:« dsporobolus cryp~andrus eCalamovilfa longifolia 31 �CHAPTER V DISCUSSION Trapping and Release Methodology The methods used for trapping and releasing Greater Prairie-chickens in this study followed Hoffman et al. (1992). They recommended trapping and releasing birds during the peak of breeding activities so that birds could immediately breed and nest. Toepfer et al. (1990) recommended releasing birds during summer when they were molting and sexually inactive to minimize dispersal from release sites. However, as Hoffman et al. (1992) explained, there is no reliable method for trapping large numbers of birds during summer, and the instinct to breed and nes~ should limit movements more than molting or sexual inactivity. Furthermore, birds released during summer must survive at least eight months before they can reproduce. Birds in this study reproduced within two months post-release. Rodgers (1992) recommended releasing> 100 birds (Sharp-tailed Grouse) within one season to compensate for high dispersal and mortality, and to increase breeding densities. However, this may have a negative effect on the source population. Additionally, release of >100 grouse in one location may result in increased movement because of artificially high densities. We were able to capture <50 birds each year over two successive years for each site with little apparent impact on the source populations. Supplemental releases near leks established in the initial year may help reduce dispersal of newly transplanted birds and, thus, increase survival. Walk-in traps were effective in capturing Greater Prairie-chickens on leks during the breeding season. We captured 89 males and 95 females with minimal disturbance to leks. Of the 184 birds trapped and released during the study period, only five (2.7%) died due to trapping and handling procedures. Schroeder and Braun (1991) reported 3.5% mortality when using walk-in traps to capture Greater Prairie-chickens. Rodgers (1992) reported a mean mortality of 10% when holding Sharp-tailed Grouse in pens an average of 40 days prior to release. Toepfer (1988) stated penning of wild birds is expensive and leads to weight loss and muscle atrophy due to a lack of flight exercise. Holding birds more than three days prior to release may have negative effects on immediate survival. Specifically, birds may suffer from dehydration, weight loss, and injuries, from being held for an extended length of time. However, Kurzjeski and Root's (1988) data on reintroduced Ruffed Grouse (Bonasa umbellus) indicated that a 10% weight loss during holding may not be critical to survival. Little and Sheets (1982) found that weight loss in wild-trapped Ruffed Grouse continues at least through the fifth day of captivity. Therefore, Kurzejeski and Root (1988) suggested the procedure of holding grouse <2-4 days before transport and release is acceptable. In this study, two males trapped in Kansas could not fly upon release in 1992. Upon examination, they were extremely underweight and suffering from dehydration. These birds died within two days. It is unknown if these birds were in this condition prior to capture or if they lost weight and became dehydrated during holding. It is advisable to weigh the birds at the time of release in addition to weighing them at time of capture to calculate weight loss. Unfortunately, I only had weights from time of capture. In addition, numerous birds trapped in Kansas were suffering from 'scalping' injuries upon release due to banging their heads on the ceilings of the wooden holding boxes. Kurzejeski and Root (1988.) observed no difference in survival estimates between Ruffed Grouse with or without injuries (exposed skull and/or muscle tissue) attained during holding. However, it is advisable to transport birds in cardboard holding boxes to prevent these injuries. Weights of males released at Pinneo were significantly higher than those of males released at Wells Ranch in 1991 (1040 vs. 958g) and 1992 (1058 vs. 1025 g (Figure 3). Males released at Wells Ranch in 1991 were significantly 32 �145 lower in weight than those released in 1992. This may be the result of being held longer (first birds caught) which could have caused more dehydration, or, it may be possible that Greater Prairie-chickens from Colorado have larger body mass than prairie-chickens from Kansas. Schroeder and Robb (1993) reported an average of 1024 g for males and 885 g for females in Colorado, while Horak (1985) reported an average of 933 g for males and 795 g for females in Kansas. These lower average weights for Kansas birds, coupled with weight loss during holding time, may have made the Kansas birds more susceptible to predation, which could have contributed to the minimal success of the Wells Ranch transplant. Dumke and Pils (1973) found similar mortality rates for juvenile Ring-necked Pheasant hens under and over 850 g. However, Stokes (1954) found that survival of "light" pheasants (pheasants weighing less than the median weight for all pheasants captured and weighed) was less than for "heavy" pheasants. Movements Dispersal from the release site has been proposed as one of the primary reasons for lack of success of translocations (Toepfer et ale 1990). Musil et ale (1993) found Sage Grouse (Centrocercus urophasianus) translocated to central Idaho moved widely following release. Lawrence and Silvy (1987) found Attwater's Prairie-chickens transplanted in Texas dispersed in all directions after release and only two of 25 birds established ranges in the vicinity of the release site. In this study, dispersal was high following release in 1991 at both Pinneo and Wells Ranch. However, in 1992 dispersal generally decreased at both sites. Hoffman et ale (1992) also found that birds dispersed extensively (up to 29 km) from the release site within the first three weeks following release, but with a supplemental release, birds did not disperse greatly, and were instead attracted to leks within 5 km of the release site. Birds in this study dispersed up to 78 km from the release site in the first year, but with the second year's release, dispersal distances decreased. The presence of birds released in 1991 may have minimized the dispersal of birds released in 1992. Toepfer (1976) noted resident Greater Prairie-chickens did not prevent spring transplanted males from wandering~or moving away from their release sites. However, the presence of a resident population of displaying males appeared to be critical to the rapid establishment and successful br~eding of transplanted hens. Toepfer et al. (1990) suggested that amount of quality habitat is the ultimate factor in determining whether a translocation will be successful. The relatively higher dispersal distances at Wells Ranch, especially in 1991, may be an indication that habitat at Wells Ranch is not as suitable as the habitat at Pinneo for sustaining a population of Greater Prairie-chickens. However, with the transplant stock for each study site coming from two completely different sources, this hypothesis may not be valid. An alternative hypothesis would be that birds from two totally dissimilar areas would' select for different habitats. Toepfer et al~ (1990) recommended the habitat the birds were obtained from should closely match that of the area in which they are transplanted. Birds released at Pinneo were trapped in northeastern Colorado which had. similar components of sagebrush, grass, and agriculture to that of the Pinneo area, while the birds released at Wells Ranch were trapped in southcentral Kansas in a mid-tall grass prairie habitat. The habitat at Wells Ranch contained more sagebrush and had shorter grasses than the habitat in Kansas. The Pinneo birds had lower dispersal distances and the Wells Ranch birds had higher dispersal distances possibly because the Wells Ranch birds dispersed farther to find habitat that more closely matched that of Kansas. Once birds dispersed and finally localized, home ranges at both sites were similar. Males tended to have larger home ranges than females and females without broods had larger home ranges- than females with broods. This is in contrast to Toepfer's (1988) findings in which brood females had larger ranges than both hens without broods and males. Schroeder and Braun (1992b) also reported that females with broods tended to have larger home ranges. It 33 �146 may be possible that hens without broods moved more looking for suitable nesting habitat after having lost a previous nest. Although females generally dispersed farther than males, once localiz~d, males had larger ranges than females. This is in contrast to Schroeder and Braun (l992b), in which males generally exhibited smaller home ranges than females. Musil et ale (1993) found movements of female Sage Grouse following release to be greater than those of males. They found one hen moved 37 km from the release site in 27 days and then moved back to the release site within two weeks. However, they found that home ranges of male and female translocated Sage Grouse during their first summer following release did not differ. The small sample size in this study, and the fact these transplanted birds may have been exhibiting atypical movement patterns when compared to established populations, may explain the differences between home ranges in this study when compared to other studies. Lek Establishment The number of active leks increased from five to six at Pinneo between 1991 and 1992 while there were three active leks each year at Wells Ranch. Hoffman et al. (1992), in an earlier study in Colorado, reported three active leks following the first year of release (36 birds released) and five leks in the second year (40 birds released). Musil et ale (1993) found five leks of translocated Sage Grouse (196 birds released) within two years post-release. These leks averaged 3.5 km from the release site. Lawrence and Silvy (1987) did not find any lekking activity by transplanted Attwater's Prairie-chickens (29 birds released) following release. At Pinneo, males per lek increased (from 4.4 to 5.0) between 1991 and 1992, while at Wells Ranch, males per lek decreased (from 5.3 to 4.7) between 1991 and 1992 (Table 7). Hoffman et al. (1992) reported three males per lek in 1984 and four males per lek in 1985. The average distance of leks from the release sites decreased at both Pinneo (1991-7.0 km, 1992-6.6 km) and Wells Ranch (1991-10.0 km, 1992-7.1 km) between 1991 and 1992 (Table 1). The decrease in the mean distance of leks from release sites in 1992 was probably due to birds from the 1991 release displaying in the vicinity of the release site and attracting birds released in 1992. Lek #6 at Pinneo and lek #1 at Wells Ranch were both" active prior to the 1992 release and were 0.93 and 2.20 km from the respective release site. All leks located (N = 14) were in areas with good visibility~ sparse vegetation, and nearby escape cover. Eight of 14 (57.1%) leks were on exposed hills or ridgetops while six of 14 (42.9%) were in low, flat areas. Nine (64.3%) leks were in areas of natural grass and other vegetation, three (21.4%) were in alfalfa fields, while two (14.3%) were in wheat fields. Other researchers have also reported selection of short cover types for leks. Toepfer (1988) described general booming ground cover as short and disturbed in Wisconsin. Schroeder and-Braun (1992b) found that Greater Prairie-chicken display sites in Colorado were generally in open habitats with short and sparse vegetation, often on ,ridgetops. Jones (1963') found that Greater Prairie-chickens preferred display sites with short grass vegetation on level prairie or on elevated areas. Ther.e was some instability in the formation of leks at Pinneo in 1991 as two leks formed, but then broke up well before the end of the display season. Schroeder and Braun (1992a) reported instability of leks between years, but not within a single display season. Without the advantage of already having established permanent leks at which males are displaying, a pioneering population of Greater Prairie-chickens may form and disband leks over a short period of time until they find more permanent lek sites. I documented two yearling males (0301 at Pinneo, 0339 at Wells Ranch) visiting two different leks within the same display season. This supports Schroeder and Braun's (1992a) findings that yearling Greater Prairie-chickens visited more leks than adults. Other researchers have also observed multiple lek visitations by males of unknown age in other populations (Hamerstrom and H~erstrom 1949, Arthaud 1968, Silvy 1968). 34 �1-+,-/ Cannon and Knopf (1981) noted that formation of new leks in response to increasing populations can involve mostly yearling males. I documented the formation of a new lek (#6) in 1992,at Pinneo by five presumed yearling (unmarked) males. Four copulations were observed on this lek in 1992. Hamerstrom and Hamerstrom (1973) found that yearling males performed only 18% of the copulations during their 22-year study. In a newly established population, it may be possible that yearling males from the first year's hatch have a better opportunity to establish leks and to account for more copulations than in an established population. Nest Establishment and Success Nest success for 13 nests in 1991-92 at Pinneo was 53.8%. This is comparable to the findings of Hoffman et ale (1992) in which Greater Prairiechicken nest success was 44.4% for the first two years following release. Also, Toepfer (1988) reported nesting success of 52% for his study of Greater Prairie-chickens in central Wisconsin. Lawrence and Silvy (1987) did not observe any nesting activity by transplanted Attwater's Prairie-chickens in Texas following release. Musil et ale (1993) reported 43% nest success (three of seven nests produced chicks) for translocated Sage Grouse in Idaho. Nest success at Wells Ranch was 100%, but only two nests were found in 1991, and none were found in 1992. The low number of nests found at Wells Ranch may be attributed to the higher dispersal distance of females. The long movements of females at Wells Ranch may be an indication there was poorer nesting habitat there than at Pinneo, or that nesting habitat at Wells Ranch is different from that in Kansas. The mean distance of nests from the release site decreased dramatically at Pinneo between 1991 (12.02 km) and 1992 (3.5 km). The presence of birds released in 1991 may have minimized dispersal of hens in 1992. Specifically, lek #6, which formed prior to release in 1992, was only 0.93 km north of the release site., Following release, three copulations with yellow-banded (1992 release) hens were observed on this lek. Toepfer (1976) stated the presence of a resident population of displaying males appeared to be instrumental in the rapid establishment and successful breeding of transplant hens. Hens showed a tenden~y to nest close to leks (Table 10). The mean distance of nests from the nearest lek at Pinneo was ,1.4 km for 1991-92 and 5.6 km at Wells Ranch for 1991. Hoffman et ale (1992) reported a mean distance of nests from leks of 1.9 km for 1984-85. Toepfer (1988) found Greater Prairie-chicken nests in Wisconsin to average 0.9 km from the nearest booming ground. Musil et ale (1993) reported a mean distance of nests from the nearest active lek of 0.9 km for translocated Sage Grouse. The fact hens at Pinneo nested closer to leks than hens at Wells Ranch indicates there was less suitable nesting habitat surrounding leks at Wells Ranch and, consequently, hens had to disperse farther to find suitable nesting habitat. Renests (N = 3) averaged 690 m from previous nests. This is similar to Svedarsky's (1979) results of renests (N = 6) averaging 760 m from previous nests. Hoffman et ale (1992) documented only one renest attempt. One hen at Pinneo (0554) in 1992 nested 360 m south of her 1991 nest site. Svedarsky (1979) reported two hens nesting 4.6 m and 29.8 m from their respective nest sites from the previous year. Mortality Mammalian 'predators were the main cause of death (69.6% of deaths) for released birds while avian predators caused 17.4% of the deaths. Durnke and Pils (1973) found that mammalian predators accounted for 48.9% of annual mortality for radio-marked hen pheasants in Wisconsin while avian predators accounted for 20.2% of annual mortality. In this study, it is suspected the majority of deaths caused by mammalian predators were by Coyotes as they were 35 �148 seen on a regular basis at both study sites. However, there was no distinction made between Coyote, Fox, dog, or other mammals, so this is purely speculation. Likewise, there was no distinction made between types of raptor kills. Mortality differed between years and sites. At Pinneo, mortality was· 25% within 120 days post-release for 1991-92, while mortality was 36% at Wells Ranch 120 days post-release for 1991-92. Hoffman et al. (1992) reported 47% mortality within 110 days post-release for transplanted birds in northeastern Colorado. , Wilson et al. (1992) reported 45% mortality for 89 radio-marked hen pheasants in Missouri while Dumke and Pils (1973) reported 74% annual mortality for radio-marked hen pheasants in Wisconsin. The higher mortality at Wells Ranch may be attributed to the higher dispersal distances of birds at Wells Ranch. It is possible that dispersing individuals exhibit higher mortality (Murray 1967). Kurzejeski and Root (1988) found that Ruffed Grouse with greater rates of movements had higher mortality than sedentary birds. However, Beaudette and Keppie (1992) found Spruce Grouse (Dendragapus canadensis) that dis'persed did not exhibit lower survival than birds that did not disperse. The difference in mortality between Pinneo and Wells Ranch may also be attributed to differences in habitat at the two sites. Thus, if the habitat provides poor cover from predators, predation may be higher. More data are needed to test this hypothesis. Lawrence and Silvy (1987) reported 56% mortality for a five-month period for transplanted Attwater's Prairie-chickens in Texas. One of the reasons they gave for this high mortality was unfavorable (overgrazed) habitat at the release site. Janes (1985) has stated that ground-level vegetation can affect the ability of a raptor to detect prey. Twenty-eight percent of the deaths at Wells Ranch were caused by raptors, whereas only 9% of the deaths at Pinneo were caused by raptors. It is possible that raptors, being highly visual predators, had higher success at Wells Ranch because of poor cover. Also, ~he Wells Ranch area was relatively closer to the South Platte River, and the density of raptors may have been higher in this riparian area. Again, more data are needed to test these hypotheses. Habitat Use Habitat used by prairie-chickens was classified to one of five types: sagebrush, grass, grass/forb, agriculture, trees/residential. Selection for habitats is not known because the availability of each amount of available habitat type was not measured. However, some trends were apparent. Sagebrush habitat was used throughout the spring ,and summer at Pinneo, while at Wells Ranch use of sagebrush habitat decreased'after the breeding and nesting periods. The use of more 'grassy habitats (CRP) at Wells Ranch after 1 June indicates these areas were more suitable than sagebrush areas for brood-rearing and cover during the summer months. These differences may indicate the sagebrush habitat at Pinneo was more suitable for Greater Prairie-chickens than the sagebrush habitat at Wells Ranch, and that there was more suitable sagebrush habitat available at Pinneo. Schroeder and Braun (1992b) showed that Greater Prairie-chickens selected sagebrush habitats in northeastern Colorado, but did not heavily use these habitats unless there was also a strong compC?nent of grass. Additional study is needed at Pinneo and Wells Ranch to determine ~election for habitat types. Schroeder and Braun (1992b) indicated that corn is important for Greater Prairie-chickens, especially during severe winter weather. Of 59 locations obtained during winter at Pinneo and Wells Ranch, 18 (30.5%) were in cornfields. On numerous additional occasions when the birds were actually flushed from sagebrush habitat in winter, they were within 100-200 m of a cornfield. Evidently, corn provides winter food for Greater prairie-chickens 36 �149 at the transplant areas. Future transplants should winter food available for Greater Prairie-chickens. consider the amount of Habitat was compared for observed and random sites for Greater Prairie-chickens at Pinneo during spring and summer 1992. Habitat was not compared at Wells Ranch in 1992 due to high mortality and dispersal of birds (92% of the birds were gone by 1 Aug). Birds at Pinneo used more sagebrush habitat throughout the spring and summer, while at Wells Ranch birds used sagebrush during spring and more grassy areas during summer. Jones (1963) indicated Greater Prairie-chickens in Oklahoma used tallgrass (46%), midgrass (29%), and shortgrass (17%) for loafing and roosting during spring while during summer they used midforbs (51%), midgrass (20%), shortgrass (12%), and tallgrass (10%) for loafing and roosting. The lack of significant differences among males and females for the variables examined indicates there was little selection for habitat during summer. This may also indicate the habitat birds used at Pinneo during summer was homogenous. Nesting habitat in this study was similar to nesting habitat found throughout the range of Greater Prairie-chickens. Nests were generally in dense vegetation which had an average height of 55 em in this study. Schroeder and Braun (1992b) reported an average height of vegetation of 59 cm at nests in northeastern Colorado while Kirsch (1974) reported average vegetation height at nests to be 51 cm throughout their range. Observed nest sites had more sagebrush cover and less bare ground than random sites. Of 15 nests found, 10 (67%) sites were dominated by sand sagebrush. Schroeder and Braun (1992b) found 74 of 83 (89%) nests in heavy sand sagebrush in northeastern Colorado. They also found more sagebrush cover and less bare ground at observed nest sites. Successful nests in this study had taller and denser sagebrush, and higher species diversity than unsuccessful nests. This is in contrast to Schroeder and Braun (1992b) who found no differences in height or cover of sand sagebrush between successful and unsuccessful nests. They also did not find higher species richness at successful nest sites. It is possible that sand sagebrush at Pinneo and Wells Ranch was more important for nesting than for other sites in northeastern Colorado. The higher species diversity at successful nest sites was indicative of sagebrush habitat with a strong component of grass and forb cover. Schroeder and Braun (1992b) found taller sagebrush and forbs, less bare ground, more grass, and'higher species richness for observed sites of hens with broods than for hens without broods. I found a greater height-density index at observed sites for non-brood females than for females with broods. Toepfer (1988) found that brood hens moved from dense cover to relatively less dense cover through which the hen and chicks could move easily to find food. This may also be the case for brood hens at Pinneo. Habitat at leks did not differ from random sites for any of the variables examined. However, there tended to be taller height-density indices, sagebrush, and grass at random sites,' and less sagebrush cover at observed lek sites. Schroeder and Braun (1992b) found habitat at leks in northeastern Colorado differed dramatically from available habitat in that lek sites had short, sparse cover, low species richness, and reduced slopes. The lack of significant differences between observed lek sites and random sites in this study can be attributed to the time of year in which they were sampled. None of the leks were sampled until June following completion of the display season. The vegetation was taller and denser on the leks at this time of year compared to earlier in the season. 37 �150 Summary and Conclusions MOl?t attempts to re-establish Greater Prairie-chickens within their former range have been unsuccessful and/or poorly documented (Toepfer et ale 1990). The methods used in this study developed by the Colorado Division of Wildlife were relatively successful when compared with many other translocation efforts. Birds at Pinneo established nine leks, and six hens had successful nests while birds at Wells Ranch established five leks and two hens had successful nests. Recruitment of juveniles was documented at Pinneo while no recruitment was documented at Wells Ranch. Mortality was 38% at Pinneo and 48% at Wells Ranch, "and mean maximum dispersal distance of birds was 8.5 km at Pinneo and 13.1 km at Wells Ranch. These data indicate the Pinneo releases were more successful in establishing a population of Greater Prairie-chickens than releases at Wells Ranch. The low number of hens that nested, lack of recruitment, higher mortality, and higher dispersal distances of birds at Wells Ranch may indicate the habitat at this site is not as suitable as the habitat at Pinneo for sustaining a population of Greater Prairie-chickens. However, it is premature to consider either transplant a success or failure. Additional data need to be collected regarding lek establishment, lek attendance, and juvenile recruitment Recommendations 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. Releases should be discontinued at the Pinneo and Wells Ranch study areas. Lek surveys should be continued through 1995 to obtain data on recruitment and population expansion. Potential habitat should be identified and managed prior to release to insure that birds will not disperse from the area. Future releases should consider the amount of winter food available for Greater Prairie-chickens. Transplant stock should be obtained from habitat that closely matches that of the release area to decrease dispersal, lower predation, and increase nest success. Future releases of Greater Prairie-chickens should be conducted during early April to promote immediate display by males and nesting by hens. Between 40-50 birds released in two successive years should be sufficient to establish a breeding population of Greater Prairie-chickens if the habitat is adequate. However, the genetics of founder populations should be considered as·relatively few males are responsible for most matings. Supplemental releases should be conducted in the vicinity of an active lek. The time birds are held in captivity should be kept to a minimum «1 day) to decrease stress and weight loss. Birds: 'ahouLd be' weighed at time of, capture and at time of release to calculate weight loss. _ Birds should be transported in cardboard holding boxes to reduce injuries. Additional study should be conducted to determine habitat use and selection throughout the year at these transplant sites. A study should be conducted to determine the relationship between predation and cover for Greater Prairie-chickens. A study should be conducted to determine the effect of weight loss and injuries attained during captivity on survival of Greater Prairie-chickens. 38 �151 APPENDIX Greater Prairie-chickens captured bands on both legs), 1991. Band # 1054 1055 1056 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 2772 2773 2774 2775 2776 2177 2778 2779 2780 2782 2783 2784 2785 2787 2788 2789 2790 2791 2792 2793 Sex F F F F F F F F F F F F F F F F F F F F F F F M M M M M M M M M M M M M M M M M M M M Age 112+ 2+ 2+ 2+ 2+ 2+ 112+ 112+ 2+ 2+ 2+ 2+ 12+ 1- 2+ 2+ 12+ 2+ 2+ 2+ 12+ 112+ 12+ 2+ 2+ 2+ 2+ 112+ 2+ A in Colorado-and Radio Frequency 0976 1096 0554 0968 released a 0896 a a a a 0 a a 0 a 0 0 0 0 0 0301 0 0280 a a 0018 a a 0109 0 0 a a 0048 a 0079 a a a 0 39 (red Date Captured Released 2 3 3 3 3 3 4 4 4 4 4 4 4 4 4 2 3 3 3 3 3 4 4 4 4 4 4 4 4 Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr 4 Apr 4 Apr 4 Apr 4 Apr 4 Apr 4 Apr 4 Apr 4 Apr 2 Apr 2 Apr 2 Apr 2.Apr 2 Apr 2 Apr 2 Apr 2 Apr 2 Apr 3 Apr 3 Apr 3 Apr 3 Apr 3 Apr 3 Apr 3 Apr 3 Apr 4 Apr 4 Apr 4 Apr 0 0906 0 at Pinneo 4 4 4 4 4 4 4 4 4 2 2 2 2 2 2 3 3 3 3 3 3 3 3 3 4 4 4 4 4 Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Mass 835 755 835 805 910 840 905 880 765 830 860 845 805 800 930 910 910 895 890 875 765 830 990 1030 1025 1050 1015 1085 965 1105 1100 1050 1010 1000 1005 1060 1120 1035 1115 1080 980 915 1060 (g) �15~ Appendix Greater Prairie-chickens captured bands on both legs), 1991- Band # 2721 2722 2723 2724 2725 2726 2727 2728 2729 2730 2731 2732 2733 2734 2735 2736 2737 2738 2739 2740 2741 2742 2743 2745 2746 2747 2748 2749 2750 2751 2752 2753 2754 2755 2756 2757 2758 2759 2760 2761 2762 2763 2764 2765 2766 2767 2768 2769 2770 2781 Sex Age M 2+ 112+ 2+ 2+ 12+ 2+ 2+ 1112+ 2+ 12+ 2+ M. U F F F F F F F F F F 12+ 12+ 112+ 12+ 112+ 12+ 112+ 12+ 2+ 2+ 2+ 12+ 2+ 112+ 2+ 12+ M M M M M M M M F M M M M F M M M M M M F F F M M F M M F F M F F F F F M M B in Kansas and released Radio Frequency a Captured 0349 0449 0089 0369 0440 0310 0378 0288 0419 0319 7 7 7 7 7 7 7 7 7 a a a a 0410 a a a a a a a a a a a a a a .0 a a a 0 a a a 0 0 0 a 0 a a a a a a 5 40 Ranch (green Date 2 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 4 4 4 4 4 4 6 6 6 6 5 5 5 5 5 5 6 6 5 6 6 0339 at Wells Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Released 9 Apr 9 Apr 9 Apr 9 Apr 9 Apr 9 Apr 9 Apr 9 Apr 9 Apr 9 Apr 9 Apr 9 Apr 9 Apr 9 Apr 9 Apr 9 Apr 9 Apr 9 Apr 9 Apr 9 Apr 9 Apr 9 Apr 9 Apr 9 Apr 9 Apr 9 Apr '9 Apr 9 Apr 9 Apr 9 Apr 9 Apr 9 Apr 9 Apr 9 Apr 9 Apr 9 Apr 9 Apr 9 Apr 9 Apr 9 Apr 9 Apr 9 Apr 9 Apr 9 Apr 9 Apr 9 Apr 9 Apr 9 Apr 9 Apr 9 Apr Mass (g) 965 1015 915 1040 915 915 965 1015 915 965 865 965 1015 865 940 940 965 1015 915 765 840 865 840 815 890 840 865 915 865 965 915 965 765 815 765 990 890 865 1015 940 715 865 965 915 740 815 840 890 915 905 �153 APPENDIX C -" Greater Prairie-chickens captured bands on both legs), 1992. Band # 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 Sex F F F M F M F M M F F M M F M M F F F F M F F M M M M F M F M F F F F M F M M F M Age 112+ 2+ 2+ 112+ 1112+ 112+ 12+ 2+ 112+ 12+ 2+ 12+ 2+ 12+ 2+ 2+ 12+ 1111112+ 1- in Colorado and released at Pinneo Date Radio Frequency Captured Released 0228 0208 0246 0268 . 0330 0349 0286 0369 0309 0 0 0 0 0 0 0640 0520 2 2 2 2 2 2 3 3 3 3 3 3 3 3 3 3 3 2 2 2 2 2 2 3 3 3 3 3 3 3 3 a a 4 4 0 0190 0 0 0 0 0 0 0 4 4 4 4 4 4 4 4 4 a 4 4 0 0 0 0 4 4 4 4 0 5 5 a a a a 5 5 5 5 5 0 a 0 41 (yellow Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr 3 3 3 4 4 4 4 4 4 4 4 4. 4 4 4 4 4 4 4 4 5 5 5 5 5 5 5 Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Mass (g) 800 875 875 1100 850 1000 ·.900 1100 950 850 800 1050 950 750 1000 1100 900 1000 900 900 1050 875 850 1000 1050 1100 1050 950 1100 950 1100 800 900 800 900 1100 850 1050 1150 950 1100 �1.)-+ APPENDIX Greater Prairie-chickens captured bands on both legs), 1992. Band # 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 . 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 Sex M F M M M F F M F M M M M M F F F F F F F M F F F F M F F F F M M M M M F F M F M M M M M F F F F F Age 12+ 12+ 2+ 1112+ 12+ 2+ 112+ 2+ 2+ 2+ 1111.2+ 2+ 2+ 2+ 2+ 2+ 2+ 2+ 2+ 2+ 2+ 1112+ 2+ 2+ 2+ 111112+ 112+ 2+ in Kansas Rad i.o Frequency a a 0408 a 0479 a a a a a a a 0560 0600 .0049 0460 0500 a 0088 0440 0539 0619 a a a a a a ·0 a a a a a a a a D and released Captured Released 2 2 1 1 1 2 2 2 2 2 2 2 2 3 3 3 3 3 3 3 3 3 4 4 4 4 4 4 4 4 3 3 3 3 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 5 5 5 5 5 4 a a a a a a a a a a a 5 5 6 6 6 6 6 6 6 42 (white Date 5 0280 0070 at Wells Ranch Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Apr Mass (g) 1100 850 1000 1000 1100 900 800 975 800 1100 1000 1125 1100 1050 850 925 875 900 850 850 850 1075 925 900 925 850 1000 875 825 850 800 950 1000 1000 1050 950 875 900 1025 850 1025 950 950 1000 1000 825 850 850 875 850 �155 LITERATURE CITED Aldrich, J.W. Geographic orientation of North American Tetraonidae. Journal of Wildlife Management. 27:529-545; 1963. Ammann, G.A. Determining age of pinnated and sharp-tailed grouse. Journal of Wildlife Management 8:170-171; 1944. Ammann, G.A. The prairie grouse of Michigan. Michigan Department of Conservation Technical Bulletin. 200 pp; 1957. Anderson, R.K. Prairie chicken responses to changing booming-ground cover type and height. Journal of Wildlife Management 33:636-643; 1969 Arthaud, F.L. Populations of the prairie chickens related to land use in southwestern Missouri. Columbia, MO; University of Missouri; 1968. M.S. Thesis. 134 pp. Beaudette, P.D.; Keppie, D.M. Survival of dispersing spruce grouse. Canadian Journal of Zoology 70:693-697. 1992. Beck, J.V. The greater prairie chicken in history. Nebraska Bird Review 25:8-12; 1957. Berger, D.D.; Hamerstrom F.; Hamerstrom, F.N., Jr. The effect of raptors on prairie chickens on booming grounds. Journal of Wildlife Management 27:778-791. 1963. Bowman, T.J.; Robel, R.J. Brood break-up, dispersal, mobility, and mortality of juvenile prairie chickens. Journal of Wildlife Management 41:27-34; 1977 • Cannon, R.W.; Knopf, F.L. Lek numbers as a trend index to prairie grouse populations. Journal of Wildlife Management 45:776-778; 1981. Christisen, D.M. National status.and management of the greater prairie chicken. Transactions of the North American Wildlife and Natural Resources Conference 34:207-217; 1969. Cooke, W.W. Further notes on the birds of Colorado. An appendix to Bulletin 37. Colorado Agriculture Collection Bulletin 44, Technical Series 4:147-176; 1898. Daubenmire, R. A canopy-coverage method of vegetational analysis. Northwest Scientist 33:43-64; 1959. Dumke, R.T.; Pils, C.M. Mortality of radio-tagged pheasants on the waterloo wildlife area. Technical Bulletin Number 72, Department of Natural· Resources, Madison, Wisconsin. 1973: 52 pp. Evans, K.E. Habitat evaluation of the greater prairie chicken in Colorado. Fort Collins, CO; Colorado State University; 1964. M.S. Thesis. 98 pp. Evans, K.E.; Gilbert D.L. Grouse of the grasslands: the greater prairie chicken in Colorado. Colorado Outdoors 12(6):15-18; 1963. Graul, W. Grassland boomers. Colorado Outdoors 24:24-28; 1975. Griffith, B.; Scott J.M.; Carpenter J.W.; Reed C. Translocation as a species conservation tool: status and strategy. Science 245:477-480; 1989. Grubb, T.G.; Eakle W.L. Recording wildlife locations with the universal transverse mercator (UTM) grid system. United State,s Department of Agriculture, Forest Service, Research Note RM-483; 3· p •. 198"8. Hamerstrom, F.N., Jr. A study of Wisconsin prairie-·chicken and sharp-tailed grouse. Wilson Bulletin 51:105-120; 1939. Hamerstrom, F.N., Jr.; Hamerstrom F. Daily and seasonal movements of Wisconsin prairie chickens. Auk 66:313-337; 1949. Hamerstrom, F.N., Jr.; Hamerstrom F. Mobility of sharp-tailed grouse in relation to ecology and distribution,. American Midland Naturalist 47:174226; 1951. Hamerstrom, F.N., Jr.; Hamerstrom F. Comparability of some social displays of grouse. Transactions of the International Ornithological Congress 12:274293; 1960. Hamerstrom, F.N., J~.; Hamerstrom F. Status and problems of North American grouse. Wilson Bulletin 73:284-294; 1961. Hamerstrom, F.N., Jr.; Hamerstrom F. The prairie chicken in Wisconsin highlights of a 22-year study of counts, behavior, movements, turnover, and habitat. Wisconsin Department of Natural Resources Technical Bulletin 64. 52pp. 1973. - 43 �156 Hamerstrom, F.N., Jr.; Mattson O.E.; Hamerstrom F. A guide to prairie chicken management. Wisconsin Conservation Department Technical Bulletin 15. 128 pp. 1957. Hessler, E.; Tester J.R.; Siniff D.B.; Nelson M.M. A biotelemetry study of survival' of pen-reared pheasants in selected habitat. Journal of Wildlife Management 34:267-274; 1970. Hoffman, R.W. Greater prairie-chicken transplant. Colorado Division of Wildlife Federal Aid Report. N-1-R. January. pp 27-41. 1986. Hoffman, R.W.; Snyder W.O.; Miller G.C.; Braun C.E. Reintroduction of greater prairie-chickens in northeastern Colorado. Prairie Naturalist 24:197-204; 1992. Horak, G.J. Kansas prairie chickens. Kansas Fish and Game Commission Wildlife Bulletin 3. 65 pp. 1985. Howard, R.; Moore, A. A complete checklist of the birds of the world. San Diego, CA: Academic Press; 622 pp. 1991. Jacobs, K.F. Restoration of the greater prairie chicken. Oklahoma Department of Wildlife and Conservation: Federal Aid Project W-65-R. 42 pp. 1959. Janes, S.W. Raptor habitat selection. In Cody, M.L., ed. Habitat selection in birds. Orlando: Academic Press,Inc.; 1985: pp. 159-188. Johnsgard, P.A. The grouse of the world. Lincoln, NE: University of Nebraska Press; 413 pp. 1983. Jones, R.E. Identification and analysis of lesser and greater prairie chicken habitat. Journal of Wildlife Management 27:757-758; 1963. Kirsch, L.M. Habitat management considerations for prairie chickens. Wildlife Society Bulletin 2:124-129; 1974. Kirsch, L.M.; Klett A.T.; Miller H.W. Land use and prairie grouse population relationships in North Dakota. Journal of Wildlife Management 37:449-453; 1973. Kruse, A.D. Prairie chicken restoration projects. Pages 40-46 in Svedarsky, D.W. ed. The prairie chicken in Minnesota. .Minneapolis: University of Minnesota Press; 1973. Kurzejeski, E.W.; Root, B.G. Survival of reintroduced ruffed grouse in north Missouri. Journal of Wildlife Management 52:248-252; 1988. Lawrence, J.S.; Silvy N.J. Movement and mortality of transplanted Attwater s prairie chickens. Journal of the World Pheasant Association 12:57-65; 1987. Little, T.W.; Sheets, R. Transplanting Iowa ruffed grouse. Proceedings of .the Iowa Academy of Science 89:172-175; 1982. Mitchell, J. Managing grouse - looking to the future. In Prairie grouse in Nebraska. Nebraskaland 62:10-13; 1984. Mohr, c.o. Table of equivalent populations of North American small mammals. American Midland Naturalist 37:233-249; 1947. Murray, B.G., Jr. Dispersal invertebrates. Ecology 48:975-978; 1967. Musil, D.O. Movements, survival and habitat use of sage grouse translocated into the Sawtooth Valley, Idaho. Moscow, ID; University of Idaho; 1989. M.S. Thesis. 72 pp. ° Musil, D.O.; Connelly, J.W .•;·.Reese, K.P. Movements, survival, and reproduction of' sage grouse translocated into central Idaho. Journal of Wildlife Management 57:85-91; 1993. Patterson, R.L. The sage grouse in Wyoming. Denver, CO: Sage Books; 341 pp. 1952. Pusateri, F.M. Greater prairie-chicken recovery plan. Colorado Division of Wildlife. Fort Collins. 25 pp. 1990. Robel, R.J. Booming territory size and mating success of the greater prairie chicken (Tympanuchus cupido pinnatus). Animal Behaviour 14:328-331; 1966. Robel, R.J. Significance of booming grounds of greater prairie chickens. Proceedings of the American Philosophical Society. 111:109-114; 1967.' Robel, R.J.; Briggs J.N.; Dayton A.D.; Hulbert L.C. Relationships between visual obstruction measurements and weight of grassland vegetation. Journal of Range Management 23:295-297; 1970a. Robel, R.J.; Briggs, J.N.; CebulaoJ.J.; Silvy N.J.; Viers C.E.; Watt P.G. Greater prairie chicken ranges, movements, and habitat usage in Kansas. Journal of Wildlife Management 34:286-306; 1970b. 44 �157 Rodgers, R.D. A technique for establishing sharp-tailed grouse in unoccupied range. Wildlife Society Bulletin 20:101-106; 1992. Roseberry, J.T.; Ellsworth D.L.; Klimstra W.D. Comparative post-release behavior and survival of wild, semi-wild and game farm bobwhites. Wildlife Society Bulletin 15:449-455; 1987. schorger, A.W. The prairie chicken and sharp-tailed grouse in early Wisconsin. Transactions of the Wisconsin Academy of Sciences, Arts, and Letters 35:1-59; 1944. Schroeder, M.A. Movement and dispersion of greater prairie-chickens in northeastern Colorado. Fort Collins, CO; Colorado State University; 1990. Ph.D. Dissertation. 135 pp. Schroeder, M.A. Movement and lek visitation by female greater prairie chickens in relation to predictions of Bradbury's female preference hypothesis of lek evolution. Auk 108:896-903; 1991. Schroeder, M.A.; Braun, C.E. Walk-in traps for capturing greater prairie chickens on leks. Journal of Field Ornithology 62:378-385; 1991. Schroeder, M.A; Braun, C.E. Greater prairie-chicken attendance at leks and stability of· leks in Colorado. Wilson Bulletin 104:273-284; 1992a. Schroeder, M.A.; Braun, C.E. Seasonal movement and habitat use by greater prairie-chickens in northeastern Colorado. Special Report Number 68. Colorado Division of Wildlife. Fort Collins. 44 pp. 1992. Schroeder, M.A.; Robb, L.A. Greater Prairie-chicken. In The Birds of North America, No. 36. Poole, A., Stettenheim, P., Gill, F., eds. Philadelphia, PA: The Academy of Natural Sciences; Washington, DC: American Ornithologists' Union. 24 pp. 1993. Schwartz, C.W. The ecology of the prairie chicken in Missouri. University of Missouri Studies 20:1-99; 1945. Sclater, W.L. A history of the birds of Colorado. London: Witherby and Company; 536 pp. 1912. Sharpe, R.S. The evolutionary relationships and comparative behavior of prairie chickens. Lincoln, NE; University of Nebraska; 1968. Ph.D. Dissertation. 187 pp. Silvy, N.J. Movements, monthly ranges, reproductive behavior, and mortality of radio-tagged greater prairie chickens (Tympanuchus cupido pinna~us). Manhattan, KS; Kansas State University; 1968. M.S. Thesis. 135 pp. Stempel, M.E.; Rodgers, S., Jr.; History of prairie chickens in Iowa. Proceedings of the Iowa Academy of Sciences 68:314-322; 1961. Stokes, A.W. Population studies of the ring-necked pheasants on Pelee Island, Ontario. Ontario Department of Lands and Forests, Technical Bulletin Wildlife Series Number 4. 1954: 154 pp. Svedarsky, W.D. Spring and summer ecology of female greater prairie chickens in northwestern Minnesota. Minneapolis, MN; University of Minnesota; 1979. Ph.D. Dissertation. 166 pp. Toepfer, J.E. Movements and behavior of transplanted radioed prairie chickens in central Wisconsin. Stevens Point, WI; University of Wisconsin; 1976. M.S. Thesis. 42 pp. Toepfer, J.E. Ecology of the prairie-chicken as related to reintroductions. Bozeman, MT; Montana State University; 1988. Ph.D. Dissertation. 536 pp. Toepfer, J.E.; Eng R.L.; Anderson R.K. Translocating prairie grouse: what have we learned? Transactions of the North American Wildlife and Natural Resources Conference 55:569-579; 1990. Van Sant, B.F.; Braun C.E. Distribution and status of greater prairie-chickens in Colorado. Prairie Naturalist 22:225-230; 1990. Watt, P.G. Mobility patterns, habitat relationships, and reproductive success of greater prairie chickens (Tympanuchus cupido pinna~us) in northeastern Kansas. Manhattan, KS; Kansas State University; 1969. M.S. Thesis. 146 pp. Watt, P.G. Life history and behavior of greater prairie chickens. In Svedarsky D.W. ed. The prairie-chicken in Minnesota. Minneapolis: University of Minnesota Press; 1973: pp. 5-14. Wilson, R.J.; Drobney, R.D.; Hallett, D.L. Survival, dispersal, and site fidelity of wild female ring-necked pheasants following translocation. Journal of Wildlife Management 56:79-85; 1992. 45 �Wooley, J. Prairie chicken update. Iowa State Conservation Commission. Des Moines. 8 pp. 1985. Yeatter, R.E. Population responses of prairie chickens to land-use changes in Illinois. Journal of Wildlife Management 27:739-757; 1963. Prepared by ..L.Ift:.~;t.._-""'-......LI.oi~~~'--_ Grant M. Beaupre Graduate Research Approved by I _-,-~....;;..V_·_:Z_'_'''''''~~;;'''_ __ Clait E. Braun Wildlife Research Leader 46 �159 JOB PROGRESS REPORT Colorado State of: Project: W-167-R Upland Work Plan: 17 Job Title: Population Period Covered: Author: Bird Research : Job _7_ Dynamics 01 January of White-tailed through 31 December Ptarmigan 1993 Clait E. Braun and Kenneth M. Giesen Personnel: Kathy Martin, University of British Columbia; Clait E. Braun and Kenneth M. Giesen, Colorado Division of Wildlife ABSTRACT Long-term studies of populations of white-tailed ptarmigan (Lagopus leucurus) were continued at hunted (Mt. Evans) and unhunted (Rocky Mountain National Park) areas in Colorado through 1993. Breeding densities of ptarmigan decreased at both sites in 1993, most markedly at Mt. Evans because of excessive harvest in fall 1992. Density of breeding ptarmigan was the lowest since 1970 (following excessive harvests in 1966-69) at Mt. Evans. The breeding density at Rocky Mountain National Park, while low, appeared poised for rebound because of the high proportion of unmated, territorial males in spring 1993. Nest success at both sites was lower in 1993 than in 1992, probably the result of extensive late storms in June and early July. Because of excessive harvest in fall 1992 at Mt. Evans, harVest regulations were changed. There was no known harvest along the Mt. Evans road in fall 1993 as a result of early snow storms that resulted in closure of the road and a permit system with a 1 bird daily and 2 per season bag/possession limit. Populations at both areas are expected to increase in spring 1994. ��161 POPULATION DYNAMICS OF WHITE-TAILED PTARMIGAN Clait E. Braun and Kenneth M. Giesen Long-term studies of trends in population size and investigation of reasons for fluctuations in size of tetraonid populations are lacking. Studies on the population dynamics of unhunted and hunted populations of white-tailed ptarmigan were initiated in Colorado in 1966 and have continued essentially uninterrupted at 2 sites. Studies of the unhunted population (Rocky Mountain National Park) identified possible short-term cycles of 7-8 years with an amplitude of 25-30% between high and low breeding densities. Conversely, studies of the manipulated population (hunted) at Mt. Evans have not indicated any cyclic pattern and it would appear that controlled hunting may mask any long-term trend that may occur. This study is designed to examine the question whether white-tailed ptarmigan are truly cyclic and whether hunting affects the apparent oscillations. P. N. OBJECTIVES The goals of this investigation are to be able to predict the length and amplitude of cycles in white-tailed ptarmigan in Colorado, to examine the impact of hunting on cycles, and· to clarify underlying causes of the apparent cycles. SEGMENT OBJECTIVES 1. Conduct breeding (May-Jun) and brood (Aug-Sep) ~ensuses ptarmigan using tape-recorded calls of males (breeding) (broods) . 2. Censuses will be conducted on previously established, defined study areas at Mt. Evans (hunted) and at Rocky Mountain National Park (unhunted). 3. Capture (noose poles) and band (aluminum and plastic color-coded bands) all unmarked white-tailed ptarmigan encountered on study areas at Mt. Evans and at Rocky Mountain National Park. 4. Individually identify all ptarmigan observed on study areas at Mt. Evans and Rocky Mountain National Park through use of binoculars. 5. Make hunting season and bag limit recommendations for Mt. Evans and collect hunting data through use of volunteer wing barrels and hunter field checks. 6. Compile data, analyze results, and prepare progress of white-tailed and chicks reports. �162 STUDY AREA AND METHODS Areas investigated were Mt. Goliath-Mt. Evans in Clear Creek County and at Tombstone Ridge-Sundance Mountain to Fall River Pass in Rocky Mountain National Park in Larimer County. The physiography, geology, location, and vegetation of these study areas have been previously described (Braun 1969, 1971; Braun and Rogers 1971; Giesen 1977). Ptarmigan were located through use of tape-recorded calls (Braun et al. 1973), captured through use of telescoping noose poles (Zwickel and Bendell 1967) as described by Braun and Rogers (1971), and classified to age and gender and banded following Braun and Rogers (1971). Age of chicks was estimated following Giesen and Braun (1979). Numbered plastic bandettes were not used as in earlier years (Braun and Rogers 1971) as a color-code system using up to 4 different colored plastic bandettes was instituted in 1977-78. A check station was not operated on the Mt. Evans highway during the opening weekend of the ptarmigan season in that area as the road was closed. A volunteer wing collection station was available to hunters in the area until the season closed. RESULTS Breeding AND DISCUSSION Densities Mt. Evans. -- Seven pairs and 2 single birds (1 male, 1 female) were recorded on the original study area in 1993. Thus, the breeding density decreased markedly from 1992 (Table 1), especially in the expanded study area (Mt. Warren - Summit Lake Flats). This was the result of the heavy harvest at Mt. Evans in fall 1992. Of the 8 males identified on breeding territories in 1993, 4 were yearlings as were 4 of 8 hens. Rocky Mountain National Park. -- Timing of breeding events on the Trail Ridge study area was one week later than in 1992 and equal to the 1966-92 average. Surveys of ptarmigan on breeding territories along Trail Ridge Road in May and June indicated a minimum population of 66 birds and included 21 pairs and 22 unpaired males. Densities on the original study area remained low (Table 1) but the increased number of unmated males suggests the population is about to expand. There was high survival of banded adult males. (4~ of 61. 68.9%) and females (15 of 26, 57.7%)'from 1992 .. One yearling banded as a chick in 1992 (male) recruited to the study area and yearlings comprised 20.2% (22 of 109) of all adult ptarmigan identified. Nesting Success and Brood Size Mt. Evans. -- Timing of breeding in 1993 was "normal" until mid June when a series of late storms caused nest loss and renesting. These storms continued until 4-5 July. Thus, nest success was poor and hatching was delayed. Only 5 of 15 (33.3%) hens observed on the study area and adjacent sites in August September were successful. These hens had an average of 1.8 chicks into September. Timing of hatch was late with 5 of 7 chicks hatching on 7 August;_ the other 2 hatched on 18 July. One hen was known to incubate a clutch with dead embryos into August (56 days) before the cl~tch was taken (KM). �163 Rocky Mountain National Park. -- Nest success was estimated from the proportion of hens with broods and without broods observed during July and August. Four of 9 hens observed during summer surveys were with broods for an estimated nest success rate of 44.4%, down from 69% in 1992. The median hatch date calculated from wing molt of 15 juveniles was 12 July (range 6-17 Jul) and was equal to the 1966-1992 average. Brood size in August averaged 4.2 chicksjhen (range 3-6). Table 1. 1966-93. White-tailed ptarmigan breeding densities (birds/km!), Colorado Study area Year 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 Rocky Mountain National Park (5.5 km") 11.3 9.8 11.5 12.0 9.6 9.1 . 8.7 7.8 8.0 11.1 13.5 12.9 10.7 8.7 8.4 8.2 7.8 6.7 5.8 6.0 4.5 6.0 5.4 6.2 7.6 6.7 7.6 6.0 Mt. Evans (4.0 km2) 3.0 2.7 2.7 2.2 2.0 4.2 7.5 6.2 6.2 6.2 6.7 > 6.0 7.5 10.3 9 .. 5 9.0 6.5 6.5 8.0 8.0 6.5 5.0 7.5 8.0 8.0 8.2 6.5 4.0 �16~ Harvest Mt. Evans. -- Due to the excessive harvest in 1992, regulations for the Mt. Evans area were modified for 1993. The season was delayed until 20 September (Monday) and closed on 1 October (Friday). In addition, all hunters were required to obtain a free permit which allowed them to harvest one ptarmigan each day with a season limit of two birds. All birds taken were to be tagged. By the time the season opened in the Mt. Evans area in 1993, a major snow storm on 12-13 September had closed the entire road above Echo Lake. Even though much of this snow melted, the road was not reopened to the public. Thus, there was no public vehicle access throughout the l2-day season. A wing collection station was maintained on the Mt. Evans road throughout the season. No wings were received at the station and no hunters were contacted in the area on 20, 24, 25, 26, 28 September or 1 October. Thus, there was no known harvest along the Mt. Evans road in 1993. LITERATURE CITED Braun, C. E. 1969. Population dynamics, habitat, and movements of whitetailed ptarmigan in Colorado. Ph.D. Thesis, Colorado State Univ., Fort Collins. l89pp. 1971. Habitat requirements of Colorado white-tailed West. Assoc. State Game and Fish Comm. 51:284-292. ptarmigan. Proc. _____ , and G. E. Rogers. 1971. The white-tailed ptarmigan in Colorado. Colorado Div. Game, Fish and Parks Tech. Publ. 27. 80pp. _____ , R. K. Schmidt, Jr., and G. E. Rogers. 1973~ Census of Colorado whitetailed ptarmigan with tape recorded calls. J. Wildl. Manage. 37:90-93. Giesen, K. M. 1977. Mortality and dispersal of juvenile white-tailed ptarmigan. M.S. Thesis, Colorado State Univ., Fort Collins. 55pp. _____ , and C. E. Braun. 1979. A technique for age determination white-tailed ptarmigan. J. Wildl. Manage. 43:508-511. Zwickel, F. C., and J. F. Bendell. J. Wildl. Manage. 31:202-204. tla_~_-_~ __ Prepared by __ -I-~~;;";;'_...:......:.... Clait E. Braun Wildlife Research."Leader 1967. of juvenile A snare for capturing blue grouse. _ Kenneth M. "Giesen Wildlife Researcher C �165 JOB PROGRESS REPORT Colorado State of: Project: W-167-R Work Plan: 21 Job Title: Warren Personnel: Job 01 January Bird Research 8 Increasing foods for wildlife eastern Colorado Period Covered: Author: Upland within through 31 December the rangelands of 1993 D. Snyder L. L. Bixler, and W. D. Snyder, Colorado Division of Wildlife ABSTRACT Annual plantings completed in May 1993 failed to establish. Time constraints prevented replanting, therefore, little progress was made on this study in 1993. ��167 INCREASING FOODS FOR WILDLIFE WITHIN OF EASTERN COLORADO Warren THE RANGELANDS D. Snyder P. N. OBJECTIVE Evaluate the establishment, survival, growth, and seed production of selected herbaceous species as food sources for avifauna within the Tamarack Prairie of eastern Logan County. Ascertain the presence/abundance of selected wildlife during brood rearing (summer) and fall-winter intervals. Assuming objectives 1 and 2 are attained, a potential 3rd objective will be to test their adaption to other rangeland sites in eastern Colorado for scaled quail, lesser prairiechickens, mourning doves, and other species. SEGMENT 1. 2. 3. 4. 5. OBJECTIVES Prepare sites for planting. Monitor the relative establishment, survival, growth, and food producing qualities of tested species. Monitor the presence/abundance of selected wildlife using test sites. Monitor precip~tation and other environmental factors. Prepare a job progress report. METHODS Snyder (1993) described the methods used which were the same as in 1992. The annual food plots were rototilled using a tractor-mounted rototil1er in April and planted in May 1993. One-half of each disturbance tillage plot was rototilled. Annual food plots were planted with a hand planter. RESULTS Precipitation remained below average through most of the growing season within the Tamarack Prairie. However, both precipitation recorders on the Tamarack malfunctioned through part of the growing season and accurate data are not available. Data from nearby weather stations were not available for the entire growing season at the time of thi's report. Dry soil conditions at and after planting caused a failure in stand establishment among all plots and sites. Although a few seedlings emerged among some species, satisfactory Stands were ,not attained. Lack of manpower and time constraints prevented replanting. Some volunteer of white wonder millet occurred within plots planted the previous year. 'No other annuals reestablished including sunflowers (Helianthus ~). Russian thistle (Salsola kali) was the dominant cover on nearly all plots including disturbance tillage. It formed relatively dense stands that attained a height averaging 0.3 to 0.5 dm. Absence of tall cover including wild sunflower resulted in minimal use by passerines and other wildlife in contrast to findings during 1992. Observations indicated that one tillage in April, using a rototiller, �168 was too early to yield excessively pulverized dense. sparse stands of annuals, and the rototiller probably the ground resulting in forbs stands that were too Lewis blue flax produced the most vigorous growth among the perennials that had been planted the previous year. It formed dense, relatively weed-free stands and considerable flowering and seed production. However, this species while attractive in appearance, apparently had little value to wildlife. It was not grazed, and produced only limited cover. Dense stands of alfalfa continued to survive through 1993 but were intensively grazed by deer and other wildlife and little growth was attained under the dry conditions. Small burnet, and other perennials, established in 1992, showed low vigor, and attained little growth in 1993; Lack of precipitation and competition with annual forbs were primary factors with sainfoin and cicer milk vetch. Yellow sweet clover, a biennial, had formed dense stands in 1992. It attained about 0.3 to 0.5 dm of height in 1993 before being defoliated by grasshoppers. Hairy vetch re-established in only one plot during 1993 and it and common vetch were not observed where planted in 1992 at other sites. LITERATURE CITED Snyder, W.D. 1993. Increasing foods for wildlife within the rangelands of eastern Colorado. Colorado Div. Wildl., Progress Rep., Fed. Aid Proj. W-167-R. Apr. pp. 101-118. �169 JOB PROGRESS State of: REPORT Colorado Project: W-167-R Upland Bird Research Work Plan: 22 Job Title: Upland Bird Research Publications Period Covered: Author: Job _1_ 01 January through 31 December 1993 Clait E. Braun Personnel: Clait E. Braun, K. M. Giesen, R. W. Hoffman, T. E. Remington, and W. D. Snyder, Colorado Division of Wildlife ABSTRACT The following articles were published in 1993: Beauprez, G. M., J. A. Clarke, and C. E. Braun. 1993. Reintroduction of greater prairie-chickens in northeastern Colorado. Proc. Joint Meeting, Prairie Grouse Tech. Council and West. States Sage/Columbian Sharp-tailed Grouse Workshop. Abstract. Braun, C. E. 1993. The status of sage grouse: are they endangered, threatened, or? Proc. Joint Meeting, Prairie Grouse Tech. Council and West. States Sage/Columbian Sharp-tailed Grouse Workshop. Abstract. 1993. White-tailed ptarmigan habitat investigations in Oregon, Oregon Birds 19(3):72-73. ___ Northeast , K. Martin, and L. A. Robb. 1993. White-tailed ptarmigan (La~opus leucurus). in A. Poole and F. Gill, eds. The Birds of North America. Philadelphia Acad. Nat. Sci. and Am. Ornithol. Union. No. 68. 24 pp. Giesen, K. M., and C. E. Braun. 1993. Considerations for reintroduction of sharp-tailed grouse in Colorado. Proc. Joint Meeting, Prairie Grouse Tech. Council and West. States Sage/Columbian Sharp-tailed Grouse Workshop. (Abstract) _____ , and 1993. Natal dispersal and recruitment of juvenile whitetailed ptarmigan in Colorado. J. Wild1. Manage. 57:72-77. ____ , and 1993 Status and distribution of Columbian sharp-tailed grouse in Colorado. Prairie Nat. 25:237-242. ___ , and J. W. Connelly. 1993. Guidelines for management of Columbian sharp-tailed grouse habitats. Wi1dl. Soc. Bull. 21:325-333. �170 Hoffman, R. W., H. G. Shaw, M. A. Rumble, B. F. Wake ling , C. M. Mollohan, S. D. Schemnitz, R. Engel-Wilson, and D. A. Hengel. 1993. Management guidelines for Merriam's wild turkeys. Colorado Div. Wildl., Div. Rep. 18. 24 pp. Schroeder, M. A., and C. E. Braun. 1993. tailed pigeons captured in Colorado. Movement and philopatry of bandJ. Wildl. Manage. 57:103-112. _____ , and . 1993. Partial migration in a population of greater prairie-chickens in northeastern Colorado. Auk 110:21-28. _____ , and L. A. Robb. 1993. Greater prairie-chicken (Tympanuchus cupido). in A. Poole, P. Stettenheim, and F. Gill, eds. The Birds of North America. Philadelphia Acad. Nat. Sci. and Am. Ornithol. Union. No. 36. 24 pp. _____ , and G. C. White. 1993. Dispersion relation to lek location: evaluation evolution. Behav. Ecol. 4:266-270. Prepared by _.....;c&J'-""'.......;;.-__ Z=-rI~...c...;;;;;:;;..__....:;.;;;.. Clait E. Braun Wildlife Research Leader _ of greater prairie-chicken of the hot-spot hypothesis nests in of lek �171 JOB PROGRESS REPORT State of: Colorado Project: W-167 -R Upland Work Plan: 26 Job Title: Analysis .Period Covered: Author: Bird Research Job _1_ of Upland Bird Population 01 January through 31 December Trends 1993 Clait E. Braun Personnel: Clait E. Braun, Kenneth M. Giesen, Richard W. Hoffman, Thomas E. Remington, and Warren D. Snyder, Colorado Division of Wildlife ABSTRACT The following were prepared Braun, 1993. C. E. Blue Mountain sage grouse harvest 1993. Cold Spring Mountain 1993. Eagle County sage grouse harvest 1993. counties, Sage grouse harvest Colorado, 1976-93: harvest 1993. Middle 1993. Northcentral 1993. North Park sage grouse harvest 1993. Piceance 1993. Yampa Area sage grouse harvest R. W. 1993. Basin sage grouse harvest Park sage grouse harvest and Western data, 1977-93. County sage grouse ?arvest sharp-tailed 19.93. Blue grouse wing analyses, data, 1976-93. data, 1973-93. data, 1977-93. data, 1993. grouse harvest Blue grouse wing analyses, Blue grouse wing analyses, Routt data, 1975-93. Basin sage grouse harvest 1993. Columbian 1976-93. 1993. data, 1977-93. data, Eastern Moffat Gunnison Moffat data, 1976-93. data, 1976-93. 1993. Giesen, K. M. Colorado, Hoffman, in 1993. Gunnison Northeast Durango data, northwest Area. Basin. Region. �-"') 11- 1993. Blue grouse wing analyses, Northwest Remington, T. E. in Colorado, 1993. Analysis 1970-93. 1993. Final report. Committee. Snyder, W. D. completed Region. of trends in small game hunter participation Colorado Division of Wildlife 1993. Summary of pheasant habitat through Pheasant Forever Chapters, Prepared by Clait E. Braun Wildlife Research Leader Laboratory improvement 1993. projects Review � https://cpw.cvlcollections.org/files/original/3326c9c6ec7f97ab0203b6a4de69d734.pdf 8d8d9eea841954171dd680788ea4d13b PDF Text Text 1 Colorado Division Wildlife Research July 1994 of Wildlife Report JOB PROGRESS State Work Colorado of Project W-153-R-7 No. Plan No. Job No. Period REPORT Covered: Author: July Jacqueline Personnel: Mammals Research 1 Multispecies 7 Mammals Library Investigations Publication, Services Editing, and 1, 1993 - June 30, 1994 A. Boss Jacqueline A. Boss and Nancy W. McEwen Abstract During the Segment the following were accomplished: * 1 publication was purchased at the request of Mammals Researcher personnel and placed into the Colorado Division of Wildlife Research Center Library collection. * 30 free reports and short publications from state or federal agencies or from private sources were located, ordered, and obtained for use by mammals Research personnel. * 6l theses or books were obtained on Interlibrary use by Mammals Research personnel. * 890 individual articles were located by Mammals Research personnel. * 17 manuscripts by Mammals accepted for publication. * 6 manuscripts were prepared personnel for peer review. Research Loan and delivered personnel and submitted or as gifts on request were published by Mammals or Research for for use ��3 MAMMALS PUBLICATION, EDITING Jacqueline AND LIBRARY SERVICES A. Boss P. N. OBJECTIVE To provide a centralized support program for manuscript editing services to facilitate publishing results of research conducted Federal Aid Project W-153-R. SEGMENT and library by staff of OBJECTIVE To provide a centralized support program for Mammals Research editing, library, and publishing services so that Mammals Research personnel can be most efficient in publishing results of their research. SUMMARY OF SERVICES Publications Purchased with Mammals Research Funds and Placed in the Research Center Library Kerasote, T. 1993. Random House. Bloodties: 277pp. nature, culture, New York: and the hunt. Theses and Books Obtained on Interlibrary Loan or as Gifts for Use by Researchers Ackerman, E. and L. C. Gatewood. sciences: a computer-aided Minnesota Press. 357pp. 1979. Mathematical models in the health approach. Minneapolis: University of Ackerman, E., L. R. Elveback, and J. P. Fox. 1984. Simulations disease epidemics. Springfield, IL : Charles Thomas. Allen, of infectious R. B. 1985. Research and management implications of the pursuit of black bears with trained bear dogs. M.S. Thesis, University of Montana, Missoula, MT. 51pp. Anderson, E. M. 1987. Bobcat behavioral ecology in relation in southeastern Colorado. Ph.D. Dissertation, Colorado University, Fort Collins, CO. 107pp. to resource State use Anderson, J. R. 1992. Monitoring the effects of clearcutting on wildlife communities in Rocky Mountain aspen stands. Professional paper, Colorado State University, Fort Collins, CO. 96pp. Anderson, R. M. 1982. and applications. The population dynamics of infectious New York : Chapman & Hall. 368pp. diseases: theory Anderson, R. M. and R. M. May, eds. 1982. Population biology of infectious diseases: report of the Dahlem Workshop on population biology of infectious disease agents Berlin 1982, March 14-19. New York: Springer-Verlag. 314pp. Anderson, R. M. and R. M. May. 1991. Infectious diseases of human: and control. New York : Oxford University Press. 757pp. dynamics Atkinson, A. C. and S. E. Fienberg, eds. 1985. A celebration of statistics: the lSI centenary volume. New York Springer-Verlag. 606pp. Bailey, N. T. J. 1982. The biomathematics Griffin & Company Ltd. 210pp. of malaria. London: Charles �4 Bailey, N. T. J. 1975. its applications. The mathematical London : Charles theory of infectious diseases and Griffin & Company Ltd. 413pp. Barigozzi, C., ed. 1980. Vito volterra symposium on mathematical models in biology. Lecture notes in biomathematics; 39. New York: springerVerlag. 417pp. Berger, J. 1978. Social development and reproductive strategies sheep. Ph.D Thesis, Univ. Colo., Boulder. 143pp. in bighorn Berger, J., W. Buhler, R. Repges, and P. Tautu, eds. 1976. Mathematical models in medicine: workshop, Mainz, March 1976. Lecture notes in biomathematics; 11. Berlin: Springer-Verlag. 281pp. Bithell, J. F. and R. Coppi, eds. 1981. Perspectives proceedings of the European symposium on medical New York: Academic Press. 330pp. in medical statistics: statistics, Rome, 1980. Cardus, D. and C. Vallbona. 1981. Computers and mathematical models in medicine: medical sessions of the first conference on mathematics at the Service of Man, Barcelona, July 11-16, 1977. Lecture notes in medical informatics; 9. New York: Springer-Verlag. 312pp. Clark, J. D. 1991. Ecology of two black bear (Ursus americanus) populations in the interior highlands of Arkansas. Ph.D Dissertation, University of Georgia, Athens, GA. 228pp. Cliff, A. D., P. Haggett, J. K. Ord, K. A. Bassett, Elements of spatial structure: a quantitative Cambridge University Press. 258pp. Cottam, D. F. 1985. Lamb production Mountain, Oregon. M.S. Thesis, 60pp. and R. B. Davies. 1975. approach. Cambridge differences of bighorn sheep on Hart Oregon State University, Corvallis, OR. Cvjetanovic, B., B. Grab, and K. Uemura. 1978. Dynamics of acute bacterial diseases; epidemiological models and their application in public health. Bulletin of the World Health Organization; Vol.56 (Supp. 1). 143pp. River otter Dronkert-Egnew, A. E. 1991. northwestern Montana. M.S. Thesis, 112pp. population University Duellman, W. E. and L. Trueb. 1986. Biology McGraw-Hill Book Company. 670pp. Dyke, status and habitat use in of Montana, Missoula, MT. of amphibians. B., and J. W. MaqCluer. 1973. Computer simulation studies. New York: Academic Press, Inc. 518pp. New York in human population Eisenfeld, J. and C. DeLisi, eds. 1985. Mathematics and computers in biomedical applications: a publication of the IMACS Technical Committee TC-5 on 'modelling of biomedical systems'. New York: North-Holland. 389pp. Enderle, J. D. 1980. A stochastic communicable disease model with age specific states and applications to measles. Ph.D. Thesis, Rensselaer Polytechnic Institute, Troy, NY. 121pp. Fager, C. w. 1991. Harvest dynamics and winter marten in southwest Montana. M.S. Thesis, Bozeman, MT. 87pp. Gans, C., and R. B. Huey, Alan R. Liss, Inc. eds. 1988. 659pp. Biology habitat Montana use of the pine State University, of the reptilia. New York �5 Gremillion-Smith, C. A. 1986. Skunk rabies: aspects of epizootiology and simulation modeling. Ph.D. Dissertation, Southern Illinois University at Carbondale. 189pp. Hallam, T. G. and S. A. Levin. .Biomathematics; Vol. 17. 1986. Mathematical ecology; an introduction • Berlin: Springer-Verlag. 457pp. Hein, Eric Walter. 1992. Evaluations of coyote attractants and a density estimate on the Rocky Mountain Arsenal. M.S. Thesis, Colorado State University, Fort Collins, CO. 58pp. Hogg, J. T. 1984. Mating behavior in Rocky Mountain bighorn sheep; male and female strategies in reproduction. Ph.D Dissertation, University of Montana, Missoula, MT. 181pp. Jager, W., H. Rost, and P. Tautu. 1980. Biological growth and spread: mathematical theories and applications. Lecture notes in biomathematics; 38. New York : Springer-Verlag. 511pp. Kane, D. M. 1989. Factors influencing the vulnerability of black bears to hunters in northern New Hampshire. M.S. Thesis, University of New Hampshire, Durham, NH. 48pp. King, M. M. 1985. Behavioral response of desert bighorn sheep to human harassment: a comparison of disturbed and undisturbed populations. Ph.D. Dissertation, Utah State University, Logan, UTe . 137pp. Kranz, J. 1974. Epidemics of plant diseases: mathematical analysis modeling. Ecological studies 13. New York: Springer-Verlag. Lakshmikantham, V., ed. 1977. international conference. Nonlinear New York and 170pp. systems and applications: an : Academic Press, Inc. 700pp. Lornnicki, A. 1988. Population ecology of individuals. Princeton University Press. 223pp. Princeton, Longini, I. M., Jr. 1977. Optimal control of influenza Thesis, University of Minnesota, Minneapolis, MN. A epidemics. 242pp. N.J.' Ph.D. Ludwig, D. and K. L. Cooke, eds. 1975. Proceedings of SIMS Conference Epidemiology. Philadelphia: Society for Industrial and Applied Mathematics. 158pp. MacDonald, G. 1957. The epidemiology Oxford University Press. 201pp. and control of malaria. on New York MacIntyre, A. A. 1982. The politics of non incremental domestic change: reform in federal pesticide and predator contro1 policy. Ph.D. Dissertation, University of California, Davis, CA. 876pp. Mack, J. 1988. Ecology of black bears on the Beartooth Face, south Montana. M.S. Thesis, Montana State University, Bozeman, MT. McGlashan, N. D., ed. . London : Methuen 1972. Medical geography: 336pp. techniques and field major central studies • & Co. Ltd. Madsen, A. and P. Willeberg, eds. 1972. Proceedings of the international summer school on computers and research in animal nutrition and veterinary medicine. Elsinore, Denmark, August 13 to 26, 1972. Copenhagen : Frederiksberg Bogtrykkeri. 563pp. Matthews, J. W. 1973. Ecology of bighorn sheep on Wildhorse Island Flathead Lake, Montana. M.S. Thesis, University of Montana, Missoula, MT. 88pp. �6 Mazaika, R. 1989. Desert bighorn sheep and nutritional carrying capacity Pusch Ridge Wilderness. M.S. Thesis, University of Arizona, Tucson, 25pp. in AZ. Miller, M. c. 1992. Reintroduction of river otters into Great Smoky Mountains National Park. M.S. Thesis, University of Tennessee, Knoxville, TN. 58pp. Molineaux, L. and G. Gramiccia. 1980. The Garki Project: research on the epidemiology and control of malaria in the Sudan Savanna of West Africa. Geneva : World Health Organization. 311pp. Muench, H. 1959. Catalytic models University Press. 110pp. Ogren, in epidemiology. Cambridge, MA Harvard H. A. 1954. A population study of the Rocky Mountain bighorn sheep (Ovis canadensis SHAW) on Wildhorse Island. M.S. Thesis, Montana State University, Missoula, MT. 77pp. Pitzman, M. S. 1970. Birth behavior and lamb survival in mountain sheep Alaska. M.S. Thesis, University of Alaska, college, AK. 116pp. Riley, R. L., and F. O'Grady. 1961. Airborne. infection: control. New York: Macmillan Co. 180pp. Ross, Ronald. edition. 1911. The prevention 711pp. of malaria. London transmission in and John Murray. 2nd Rossell, C. R., Jr. 1990. The influence of human demographic factors on black bear harvests in New Hampshire. M.S. Thesis, University of New Hampshire, Durham, NH. 51pp. Scott, M. G. 1991. Body fat prediction, nutrition bears in the interior highlands of Arkansas. Arkansas, Fayetteville, AR. 58pp. Selby, P. 1982. Lancaster Influenza models: prospects MTP Press Ltd. 259pp. and reproduction of black M.S. Thesis, University of for development and use. Simmons, N. M. 1969. The social organization, behavior, and environment the desert bighorn sheep on the Cabeza Prieta Game Range, Arizona. Dissertation, University of Arizona, Tucson, AZ. 145pp. Waller, A. J. Montana. 1992. Seasonal habitat use of river otters in northwestern M.S. Thesis, University of Montana, Missoula, MT. 75pp. Warren, E. R. 1942. 2nd rev. ed. The mammals of Colorado, distribution. Norman, OK :.Univ. of Oklahoma Press. Wells, W. F. Harvard 1955. Airborne contagion University Press. 423pp. and air hygiene. their habits 330pp. Cambridge, Wolstenholme, G. E. W. and M. O'aonnor, eds. 1962. CIBA foundation on Bilharziasis. Boston: Little, Brown and Co. 433pp. Reference of Ph.D Document Location and MA Symposium and Delivery The Research Center Library staff also located and delivered individual articles or free documents on request for Mammals personnel during this segment. approximately Research 920 �7 Manuscripts Job Progress Published Reports; FY 1993-94 Federal Aid. All studies. Andelt, W. F., K. P. Burnham, and D. L. Baker. 1994. Effectiveness of capsaicin and bitrex repellents for determining browsing by captive deer. J. Wildl. Manage. 58:330-334. mule Barone, M. A., M. E. Roelke, J. Howard, J. L. Brown, A. E. Anderson, and D. E. Wildt. 1994. Reproductive characteristics of male Florida panthers: comparative studies from Florida, Texas, Colorado, Latin America, and North American zoos. J. Mammal. 75:150-162. Gross, J. E., L. A. Shipley, N. T. Hobbs, D. E. Spalinger, and B. A. Wunder. 1993. Functional response of herbivores in food-concentrated patches: tests of a mechanistic model. Ecology. 74:778-791. Kraabel, B. J., M. W. Miller, D. M. Getzy, and J. K. Ringelman. 1995. Corrosion of and inflammatory responses to embedded tungsten-bismuth-tin shot and steel shot in mallards. J. Wildl. Dis. 31: (in press) Kufeld, R. C. and D. C. Bowden. 1994. 'Mule and white-tailed Colorado plains river bottoms. Colo. Div. Wildl. Tech. (in press) deer of eastern Publ. No. 41. McCarty, C. W. and J. A. Bailey. 1994. Habitat requirements of desert bighorn sheep. Colo. Div. of Wildl. Special Report No. 69. 27pp. Miller, J. R. and N. T. Hobbs. during invitro digestion 110. 1994. Effect of forage hydration on lag time of meadow hay. Grass & Forage Sci. 49:107- Miller, M. W., M. A. Wild, and W. R. Lance. 1995. Efficacy and safety of naltrexone hydrochlorine in antagonizing carfetanil citrate immobilization in captive Rocky Mountain elk (Cervus elaphus nelsoni). J. Wildl. Dis. 31:(in press). alterman, J. H., D. W. Kenvin, and R. C. Kufeld. southwestern Colorado. Alces. (in press) Pojar, Pojar, 1994. Mopse transplant T. M., D. C. Bowden, and R. B. Gill. 1994. Aerial counting experiments to estimate pronghorn density and herd structure. Manage. 58: (in press) to J. Wildl. T. M., K. D. Pojar, C. H. Wagner, and R. Firth. 1994. Observations pronghorn response to an electric fence. Pronghorn Worksho~, Proc. 16: (in press) of Shipley, L. A., J. E. Gross, D. E. Spalinger, N. T. Hobbs, and B. A. Wunder. 1994. The scaling of intake rate in mammalian herbivores. Am. Nat. 143:1055-1082. Thorne, E. T., M. W. Miller, D.' L. Hunter, and E. S. Williams. 1993. Wildlife management agency concerns about bovine tuberculosis in captive cervidae. pp. 47-51 in M. A. Essey, ed., Bovine tuberculosis in Cervidae: proceedings of a symposium. July 16-17, 1991, Denver, Colorado. USDA/APHIS/VS, Misc. Publ. 1506, Washington, D. C., 71 pp. Torbit, S. C., R. B. Gill, of pronghorn grazing 57: 173-181. White, A. W. Alldredge, and J. C. Liewer. 1993. Impacts on winter wheat in Colorado. J. Wildl. Manage. G. C. and R. M. Bartmann. 1994. Drop nets versus helicopter for capturing mule deer fawns. Wildl. Soc. Bull. 22:248-252. net guns �8 Wild, M. A. and M. W. Miller. 1994. Effects of modified Cary and Blair medium on recovery of nonhemolytic Pas~eurella haemoly~ica from Rocky Mountain bighorn sheep (Ovis canadensis) pharyngeal swabs. J. Wildl. Dis. 30:16-19. Wild, M. A., M. W. Miller, D. L. Baker, N. T. Hobbs, R. B. Gill, and B. J. Maynard. 1994. Comparing growth rates of dam- and hand-raised bighorn sheep, pronghorn, and elk neonates. J. Wildl. Manage. 58:340-347. Manuscripts Baker, in Review Py 1993-94 D. L., M. W. Miller, and T. M. Nett. 1995. GnRH analog-induced patterns of LH secretion in female wapiti (Cervus elaphus) during the breeding season, anestrus, and pregnancy. BioI. Reprod. (in review). Bowden, D. C. and R. C. Kufeld. 1994. Generalized mark-sight population size estimation applied to Colorado moose. J. Wildl. Manage. (in review) Hobbs, N. T., D. L. Baker, G. D. Bear, and D. C. Bowden. 1994. Ungulate grazing in sagebrush steppe I: mechanisms of resource competition. Ecolog. Appl. (in review). Hobbs, N. T., D. L. Baker, G. D. Bear, and D. C. Bowden. grazing in sagebrush steppe II: effects of resource secondary production. Ecolog. Appl. (in review). Kufeld, R.- C. review) 1994. status and management of moose 1994. Ungulate competition on in Colorado. Alces. Popel, A. S., P. C. Johnson, M. V. Kameneva, and M. A. Wild. 1994. for red blood cell aggregation is higher in athletic mammalian than in sedentarY species. J. Appl Physiol. (in review). -L~~ lJs~ Jac~nelt. Prepared by . Boss LibrtJanu . . (in capacity species �9 Colorado Division Wildlife Research July 1994 of Wildlife Report JOB PROGRESS REPORT Colorado state of Project Work No. W-153-R-7 Plan No. Job No. Period Covered: Author: Personnel: R. Bruce July Mammals Research 1 Multispecies 9 Mammals Investigations 1 Research Administration 1, 1993 - June 30, 1994 Gill R. Bruce Gill and Diane K. Hall ABSTRACT Personnel time and budgets were allocated among 6 Mammals '1 Research Projects. Highlights of preliminary results and management implications are summarized. All research objectives were successfully met within the assigned budgets. Four technical manuscripts were accepted publications and/or journals. for publication in scientific ��11 MAMMALS 1 RESEARCH ADMINISTRATION R. Bruce Gill P .N. OBJECTIVE Administer research studies within productivity at the lowest cost. the Mammals segment 1. Assign, supervise, and moose projects and administer in the Mammals 1 Research Unit for the highest Objectives research projects concerning 1 Research Section. deer, elk, RESULTS Six projects were active during the segment. Segment completed for all 6 projects. Highlights include: short publications, objectives • Acquisition of 31 reports, Research Library holdings. • Publication or acceptance publications. • Preparation and submission peer review. • Preliminary evidence for" increased survival of mule deer fawns in response to reductions in deer population density, in turn suggesting that managing density of mature females may result in increased survival and recruitment of male fawns to adulthood. • Strong evidence was obtained linking early season elk hunting to elk distributional responses. Elk apparently respond to the opening of archery and muzzle loading rifle seasons by retreating to functional refugia where hunting pressure is greatly reduced. These refugia consist of both private and public lands where hunter access is denied by landowners or where hunter access is limited by scarcity of access roads. • Survival rates of calf and adult cow elk were virtually identical in a west central elk population with mean annual survival rates for both groups exceeding 90%. These survival rates include only natural sources of mortality. Hunting induced mortality will be estimate during the 1994 hunting season. Experiments in elk counting revealed that visual sightings of elk on counting surveys missed approximately 15% of the elk on sample units. • Eighteen additional moose, 1 adult cow and 17 calves, and each was fitted with a radio transmitter equipped mortality switch. • Initial experiments to test the feasibility of hormonal toxins to control fertility in mule deer indicated that hormonal toxins would deliver effective doses of toxins to deer pituitaries during winter periods when live-capture is most effective. Both pregnant and nonpregnant does .responded to doses of synthetic GnRH in ranges which would be practical to deliver remotely with syringe projectiles or biobullets. Simulation modelling exercises suggest that fertility control can be efficaciously used to control populations provided for publication of 3 technical and books were for the of 4 technical publication manuscripts were with to captured a �12 that initial treatments sexually mature females females can be rendered • Prepared result in infertility of 2/3rds of the and that 20-30% of the remaining fertile infertile each year. Segment objectives of all Mammals 1 Research successfully accomplished within the budgets projects. by: R. Bruce Gill Wildlife Research Leader Projects were allocated to those �13 colorado Division Wildlife Research July 1994 of Wildlife Report JOB PROGRESS Colorado state of Project Work Mammals W-153-R-7 No. Plan No. 2 Deer Covered:. Author: Personnel: Research Investigations Compensatory Effects Mule Deer Population 15 Job No. Period REPORT of Harvest in a July 1, 1993 - June 30, 1994. R. M. Bartmann, G. C. White. G. D. Bear, L. H. Carpenter, J. Frothingham, R. B. Gill, Harthan, R. H. Kahn, T. Lytle, C. L. Vardaman. R. Abstract Line transects were not flown on the Ridge study area during the 1993-94 winter. However, deer densities on the control and treatment units were not expected to change much from the previous year because of low fawn survival on both units during the 1992-93 winter. Helicopter net gunning was used to capture 85 fawns and 14 does on the control unit and 81 fawns and 10 does on the treatment unit between 17 November and 2 December, 1993. The estimated fawn survival rate on the treatment unit (0.767, SE 0.048) was significantly higher (~= 0.006) than on the control unit (0.550, SE 0.056). As of 30 June, only 3 does died on each unit yielding similar survival rate estim.ates: control - 0.939, SE 0.034 and treatment - 0.935, SE 0.036. One fawn from the control unit and 2 from the treatment unit died off the Ridge study area. Essentially all remaining fawns were on the units where captured during most of the winter. Treatment unit fawns were significantly larger (~~ 0.027) than control unit fawns with respect to weight, total body length, and left hind foot length. No differences in these same measurements were found for does from the control and treatment units (~~ 0.099). ��15 COMPENSATORY EFFECTS OF HARVEST IN A MULE DEER POPULATION Richard M. Bartmann and Gary C. White P. N. OBJECTIVES 1. Increase the winter survival rate of mule deer fawns by lowering deer density to reduce competition for forage during winter. 2. Increase the harvest rate of deer through increased productivity of adult does and decreased natural mortality of fawns resulting from closer alignment of population size with carrying capacity. SEGMENT total OBJECTIVES 1. Maintain the winter population a density <40/krnf for 5 years. of mule deer on the Ridge 2. Estimate winter survival rates of fawns on control 3. Estimate units. annual survival rates of adult 5. Estimate condition of fawns on control 6. Estimate condition of adult females females and treatment on control and treatment on control treatment unit at units. and treatment units. and treatment units. METHODS Except for deer trapping, methods remained the same as previously reported by Bartmann (1990) and Bartmann and White (1991) with modifications by Bartmann and White (1992). All deer were captured with helicopter net guns by Helicopter Wildlife Management, Inc. RESULTS Maintain AND DISCUSSION Population Line transects were not flown on the Ridge study area. Estimated fawn survival was less than 15% on both ends of the Ridge. during the 1992-93 winter, so little or .no increase in deer density 'was expec;:ed. Fawn Survival In 1993, all deer were captured by helicopter netgunning. Netgunning was done in 2 sessic:ms; the first occurred 17-21 November and the second 29 November-2 December. During each session, capture effort was alternated daily between control and treatment units. A hand-held Global Positioning System unit was used in the helicopter to get capture locations for each deer. We captured and radiocollared 85 fawns on the control unit and 81 on the treatment unit. Five fawn deaths on the control unit and 1 on the treatment unit were considered capture-related. The 1993-94 winter was relatively mild and fawn survival on both units was high compared to previous years. Fawn survival on the control unit (0.555, SE 0.055) was only 'exceeded during the mild 1989-90 winter and was comparable to that for the 1985-86 winter (Table 1) •. On the treatment unit, fawn survival �16 (0.780, SE 0.043) was about the same as the previous high during 1989-90. It was also significantly higher (~ = 0.002) than on the control unit suggesting a possible effect of the reduced deer density. Table 1. Kaplan-Meier estimates of survival rates (~) for radio-collared mule deer fawns on control and treatment units of the Ridge study area in Piceance Basin, Colorado, from time of collaring in November and December until the following 15 June 1982-83 through 1993-94. Hunting mortalities are censored. Control ~ Winter 1982-83 1983-84 1984-85 1985-86 1986-87 1987-88 1988-89 1989-90 1990-91 1991-92 1992-93 1993-94 28 28 34 59 60 32 34 38 34 28 80 80 0.321 0.071 0.196 0.537 0.431 0.241 0.270 0.779 0.320 0.456 0.112 0.550 unit SE (~) 0.088 0.049 0.078 0.070 0.064 0.077 0.083 0.078 0.090 0.107 0.036 0.056 Treatment ~ 31 32 26 58 58 28 28 44 36 42 74 80 0.387 0.033 0.431 0.439 0.471 0.107 0.445 0.745 0.339 0.548 0.148 0.767 unit SE(~) P of equal ~(:t.) 0.087 0.033 0.105 0.070 0.067 0.058 0.096 0.070 0.106 0.098 0.043 0.048 0.578 0.774 0.075 0.157 0.565 0.006 0.509 0.659 0.909 0.481 0.102 0.006 As during the 1992-93 winter, fawn mortality attributed to predation was a major factor on both units accounting for 60% of the deaths on the control and 44% on the treatment unit (Table 2). These are minimal e·stimates as. they only represent confirmed losses. When evidence was inconclusive, mortalities were placed in the "other" category. No fawns were shot as there was no late hunting season because the desired population reduction on the treatment unit was deemed to have been achieved. Three fawns, 1 from the control unit and 2 from the treatment unit, died off the Ridge study area and were deleted from survival analyses. Location checks during mid-month in January, February, and March indicated most deer remained on the units where they were captured. The few that were off their units, were captured near the boundary and were usually close to it during subsequent locations. Adult Doe Survival Adult doe survival on both units from 1 December 1993 to 30 June 1994 was quite high. Only 3 does died on each unit for preliminary survival rate estimates of 0.934 (SE 0.034) on the control and 0.935 (SE 0.036) on the treatment unit (Table 3). As with fawns, there were no hunting mortalities because there was no late season. Condition of Fawns For the first time during this study, all 4 condition indices for fawns on the treatment unit were larger than for fawns on the control unit (~S 0.027). On average, treatment unit fawns were 2.2 kg heavier, 2.5 cm longer, and had a left hind foot length 0.8 cm longer than control unit fawns (Table 4). Their weight/length ratio of 0-.250 (SO 0.023) was also greater than the 0.237 (SO 0.022) for control unit fawns. �17 Table 2. Cause of mortality for radio-collared mule deer fawns on control and treatment units of the Ridge study area in Piceance Basin, Colorado, from time of collaring in November and December until the following 15 June 1982-83 through 1993-94. Percentages are of total uncensoreda fawns. Winter a 1982-83 1983-84 1984-85 1985-86 1986-87 1987-88 1988-89 1989-90 1990-91 1991-92 1992-93 1993-94 29 28 34 59 60 32 34 38 34 28 80 80 1982-83 1983-84 1984-85 1985-86 1986-87 1987-88 1988-89 1989-90 1990-91 1991-92 1992-93 1993-94 31 32 26 58 58 28 28 44 36 42 74 80 Censored Hunting Other 1 7 11 6 2 5 14 9 7 1 4 2 1 1 1 4 9 5 3 9 7 9 1 10 5 9 9 6 Starvation No. % Control unit 54 15 22 79 16 59 10 21 14 26 22 73 34 10 3 14 6 24 7 33 12 15 4 3 Treatment unit 15 50 87 27 8 36 17 35 16 30 19 68 9 36 20 6 8 40 7 29 23 34 3 4 Mortality cause Predation No. % 4 4 5 7 17 14 14 19 15 31 5 17 10 3 50 22 40 14 64 29 2 3 2 11 13 1 1 7 10 9 22 25 4 4 3 2 26 8 15 8 39 11 Other No. % .1 7 3 2 7 3 3 3 7 11 4 15 6 7 24 14 12 14 9 14 2 7 3 1 1 5 5 4 4 4 7 8 14 2 2 18 20 13 20 17 10 11 a Uncensored fawns are those that were not killed by hunters, that had nonfailing radios, or that had collars that did not drop off prematurely. A Table 3. Kaplan-Meier estimates of annual (1 Dec-30 Nov) survival rates (~) for radio-collared adult female mule deer on control and treatment units of the Ridge study area in Piceance Basin, Colorado, 1982-83 through 1993-94. Hunting mortalities are censored. Winter 1982-83 1983-84 1984-85 1985-86 1986-87 1987-88 1988-89 1989-90 1990-91 1991-92 1992-93 1993-94a a n 10 15 9 25 27 14 7 23 39 41. 46 49 Survival Control Hunting 1 3 2 1 unit ~ SE(~) n 0.800 0.779 1.000 0.917 0.756 0.818 0.857 1.000 0.969 0.758 0.716 0.939 0.126 0.113 11 15 10 21 18 10 5 28 41 42 52 46 rate estimates 0.056 0.087 0.116 0.132 0.031 0.071 0.067 0.034 for 1993-94 Treatment Hunting 2 1 9 12 6 7 0.909 0.929 1.000 0.900 0.878 1.000 0.800 1.000 0.906 0.806 0.719 0.935 are only through unit .SE(~) A 0.448 0.271 1.000 0.821 0.432 0.329 0.854 1.000 0.474 0.641 0.910 0.933 0.087 0.069 0.067 0.081 0.179 0.065 0.072 0.066 0.036 30 June ~ of equal 2 1994. �18 Condition of Ooes Only 14 adult does were captured on the control unit and 9 on the treatment unit. These small sample sizes together with high variability resulted in no significant differences for any of .the 4 condition indices (~~ 0.099) (Table 5). Only 1 yearling doe was captured on the treatment unit which precluded similar comparisons between units. Table 4. Weights (kg) and body measurements (cm) of mule deer fawns trapped on control and treatment units of the Ridge study area in Piceance Basin, Colorado, 1982-93. Weight .2. so 28 28 34 60 60 33 34 40 35 28 82 85 34.6 31. 7 32.2 32.6 31.9a 29.9 29.5 32.7 30.8 30.7 30.4 30.7 3.10 4.40 4.65 4.02 3.89 3.60 3.10 '3.31 4.29 3.58 3.80 3.91 31 32 26 60 61 28 30 47 36 43 83 80 32.8c 32.3 32.3 32.3 31. 7 30.2 28.8 30.6 30.7 32.6 30.3 32.9 Unit Year n Control 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 Treatment 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 a n _ 58 = 27 n = 30 bii c 4.18 3.12 5.07 4.62 4.13 5.34 4.13 3.33 4.42 ·3.54 4.68 3.96 TOtal bod:r:lenoth SO .2. Left hind foot leng!;h SO .2. 124.0 124.2 123.9 124.4 128.1 127.3 123.8 131.0 126.5 128.2 126.8 129.1 4.64 5.65 7.25 6.26 6.53 6.12 7.83 5.81 9.02 5.47 6.67 6.17 41.1 40.6 40.8 41.1 41.0 40.8 41.0 41.8 40.8 40.6b 40.6 40.5 1.08 1.73 1.53 1.48 1.95 1.72 1.37 2.16 1. 75 1.32 1.58 1.68 121.7 123.6 124.7 124.4 125.9 127.5 124.9 126.6 127.8 129.8 126.7 131.6 5.45 5.53 7.25 6.28 6.58 8.86 7.07 5.33 6.54 6.11 7.94 5.76 41.1 40.6 40.8 40.8 41.0 41.2 40.6 40.7 41.1 40.8 40.5 41.3 1.65 1.34 1.89 1.77 2.11 1.66 1.88 1.41 1.69 1.29 1. 77 1.43 �19 Table 5. Weights (kg) and body measurements (cm) of yearling mule deer trapped on control and treatment units of the Ridge Piceance Basin, Colorado, 1988-93. unit Year n Control 1988 1989 1990 1991 1992 1993 2 2 10 8 4 46.2 46.0 49.4a 50.9 50.6 Treatment 1988 1989 1990 1991 1992 1993 2 7 8 10 10 1 50.2 49.3 50.8 49.4 50.3 48.7 Control 1989 1990 1991 1992 1993 19 21 30 31 14 67.7 66.9 62.9 63.0 65.3 Treatment 1989 1990 1991 1992 1993 39 25 32 36 9 65.8 67.6 62.8 62.4 61.9 ~ Total body: length SD ~ Weight SD Yearlings 1.91 6.43 2.38 2.50 4.85 and adult female study area in Left hind foot length SD ~ 146.1 148.4 147.9 149.2 145.6 10.75 2.26 2.49 3.38 4.19 45.8 47.8 45.8 45.7 45.5 0.99 3.04 1.08 1.02 1.59 151.2 154.3b 150.1 149.8 149.8 150.0 3.18 6.20 4.94 5.31 5.35 46.0 46.6 45.6 45.3 47.2 45.1 0.00 ·0.69 1.26 1.51 6.37 Adults 5.22 5.51 5.18 5.67 8.32 168.6 166.7 162.3 164.3 169.3 5.20 6.98 6.40 7.89 7.70 48.1 47.3 47.3 .47.5 47.3 1.10 1.09 2.19 1.29 1.32 5.04 5.14 6.24 6.91 2.58 166.6 166.8 162.2 162.4 163.8 6.57 5.81 7.07 7.32 7.08 47.8 48.1 47.0 47.2 47.4 1.32 1.44 1.06 1.47 1.04 0.35 3.58 3.38 3.45 5.38 a !! = 9. 6. !! b LITERATURE CITED Bartmann, R. M. 1990. Compensatory effects population. Colo. Div. Wildl., Wildl. of harvest Res. Rep. in a mule deer July: 187-196. _____ , and G. C. White. 1991. Compensatory effects of harvest in a mule deer population. Colo. Div. Wildl., Wi.ldl. Res. Rep. Ju1y:27-40. _____ , and G. C. White. 1992. Compensatory population. Colo. Div. Wildl., Wildl. effects of harvest in a mule deer Res. Rep. July: (in press). ��������'<'1 Colorado Division Wildlife Research July 1994 of Wildlife Report .•. - JOB PROGRESS State of Project Work Colorado No. Job No. Covered: Author: Personnel: Mammals W-153-R-7 Plan No. Period REPORT Research 3 Elk Investigations 9 Estimating Survival Rates and Developing Techniques Estimate Population Size July 1, 1993 - June of Elk to 30, 1994 D. J. Freddy F. Barnes, J. Broderick, G. Byrne, D. Crane, J. Ellenberger, J. Frothingham, V. Graham, J. Gray, R. Hays, C. Kolus, G. Loucks, J. Pelletier, S. Rivera, P. Will, CDOW; D. Bowden, C. Vardeman, G. White, CSU; M. Miller, C. McCarty, R. Witt, Volunteers; USFS Rifle, BLM Glenwood Springs, Cooperating. Abstract We radio-collared 73 calf elk (Cervus elaphus nelsoni) (6 mos old) and 68 adult female elk (~ 1 year old) in December 1993 to estimate survival rates during winter and used these same elk in 110 aerial sighting bias trials to develop models for estimating degree of negative sighting bias when counting elk with a helicopter on sample quadrats. Elk were captured using portable corral traps and helicopter net-gunning. Survival rates (± 95% CI) from December 1993 to June 1994 were 0.918 + 0.063 for calves and 0.956 + 0.049 for adult females. Suspected causes of death were primarily malnutrition and predation for calves and shooting and predation for adult females. Calves suspected of dying from malnutrition had marrow fat values < 13% at death and body weights < 102 kg at capture. Blood tests for brucellosis were negative in all 43 adult females sampled while progesterone assays indicated 74-80% of the adult females were pregnant. Observers did not count 15.4% of the elk during sighting trials resulting in a simple binomial correction factor of 0.846 ± 0.071 (± 95% CI). Univariate tests, however, indicated elk age, log(ln)initial group size, log(ln) total group size, elk behavior, vegetation type, percent occlusion cover, percent snow cover, and wind conditions affected sightability of elk (P ~ 0.05). Step-down regression analyses produced a complex 12 parameter sighting model inclusive of elk age, elk behavior, vegetation type, log total group size, and percent snow cover. Simpler models inclusive of log total group size or log total group size and percent occlusion cover may be more utilitarian and provide acceptably precise correction factors. The simple binomial correction model provided unacceptably inflated and imprecise correction factors which supported incorporating the variable total group size into sighting bias models. ��29 ESTIMATING SURVIVAL JOB PROGRESS REPORT RATES OF ELK AND DEVELOPING ESTIMATE POPULATION SIZE TECHNIQUES TO David J. Freddy P. N. OBJECTIVE Estimate estimate survival rates of adult population size. female and calf elk and develop SEGMENT techniques to OBJECTIVES 1. Radio-collar Management 75 adult female and 75 calf elk during Unit 42 south of Rifle, Colorado. 2. Estimate winter and annual survival from known fates of radioed elk. 3. Estimate probability of sighting and counting elk during surveys using radioed elk as a known population. 4. Analyze survival and sightability Aid Job Progress reports rates December of calf and adult data and summarize 1993 in Game female elk helicopter annually in Federal INTRODUCTION Elk (Cervus elaphus nelsoni) are a high-profile and burgeoning wildlife resource throughout much of Colorado. This burgeoning resource has many benefits but frequent social, political, and economic conflicts suggest elk, to some degree, have reached a "social" carrying capacity. Many conflicts arise, in part, from an inadequate ability to predict the dynamics of elk populations. Key population parameters such as survival rates of calves and adult females and total population size are seldom measured resulting in computer models of elk population dynamics that are based on minimal data. Such models can be of little predictive value and, as such, diminish confidence in using models to guide management of elk populations. Our objectives are to provide reliable estimates of survival rates for calves and adult females during winter and for adult females throughout the year for the period 1993-94 through 1997-98. Additionally, we will develop and test a system for estimating population size that will incorporate estimates of sighting bias in conjunction with a random sampling system using search quadrats as sample units. Our winter study area encompasses about 839 kmt (324 mi2) in the eastern half of Game Management Unit 42 south and east of Rifle, Colorado~ Elk winter range vegetation types include juniper-pinyon woodland (Juniperus osteosperma-Pinus edulis), oakbrush-mountain shrub (Quercus gambelii-Amelanchier alnifolia), aspen (Populus tremuloides), sagebrush ~_rtemisia tridentata), and agricultural fields (Freddy 1993). �30 METHODS Marking We placed radio collars (172-176MHz) having mortality sensors on 73 calves (6 mos old), of which 36 were males and 37 were females, and 68 adult females (~ 1 year old). Of these, 54 calves and 46 adults were trapped from 2-6 December 1993 using helicopter net-gunning and 19 calves and 22 adults were trapped from 10-20 December using portable corral traps. Helicopter capture occurred at 17 remote sites located primarily on public lands while corral-trapping occurred at 3 sites primarily on private lands. Trapping effort was allocated among 8 geographic trap zones to assure that radioed elk were representative of most if not all segments of the population (Table 1). Number of elk to be captured in each zone was established prior to trapping based on distribution and relative numbers of elk observed during sex and age ratio classification flights completed annually in January from 1989 to 1993. Collars for adults were fixed in circumference while those for both male and female calves were expandable to accommodate growth when animals became adults. Additionally, all radio collars had a 2 character visual identification code on the dorsal surface of the collar consisting of letters, numbers, or symbols. Collars were white with black symbols. Calves captured by net-gunning were ferried by helicopter to processing points usually within 1.6 km of capture sites. At processing points, body weight, total body length, and hind foot length were measured and calves were then radio-collared and released. Similar measurements were also made on calves that were corral-trapped and then released at the trap site. Body measurements for calves were compared between sexes using Proc FREQ and Proc GLM (SAS 1988). Physiological Assays Serum samples were obtained from adult females captured by net-gunning for brucellosis tests (USDA Lab, Denver) and progesterone assays (Physiology Lab, Colo. St. Univ.). Fat content (percent dry matter) of femur marrow or lower jawbone marrow obtained from dead radioed elk was determined (Colo. Div. Wildl. Lab, Ft. Collins). Survival We monitored life or death .status of radioed elk during daily ground surveys and aerial surveys conducted at 2-4 week intervals from December 1993 through April 1994 and via monthly aerial surveys in May and June 1994. Survival rates (S) of radioed elk.were calculated using the binomial estimator with ~ variance, VAR(-S) = S(l"';'S)/n' (White and Garrot 1990) (Proc FREQ, SAS 1988). We chose not to use the staggered entry approach and did not use a Kaplan-Meier approach because no animals were censored (White and Garrot 1990). Life or death status of all radioed calves was known for the period 2 December through 15 June. On 15 June, calves become yearlings for purposes of calculating rates of calf survival. Life or death status was known for 65 of 68 adult females through 15 June with all females of known status through 28 April. Sighting Bias We conducted 110 aerial sighting bias trials that targeted individual radioed elk located on search quadrats (Samuel et ale 1987). Trials were conducted on 25, 27, 28, and 31 January, 1, and 14-17 February, and on 9 March. Search quadrats were approximately 2.6 km (1 mi2) in size with boundaries based on topographic features We used a Bell B-1-Soloy helicopter, the same pilot, and observation teams composed of 2 persons selected from 3 observers and 2 navigators for all �31 trials. The Bell-Soloy offers good forward and lateral visibility. Navigators were seated in the middle with the pilot to their left and the primary observer to their right. Navigators were highly experienced at flying sample quadrats and observers had high, moderate, low experience in flying sample quadrats. However, all observers were well experienced in conducting sex and age ratio helicopter survey flights for elk. The helicopter was flown at speeds of 65-95 krnph at 25-40 m above the vegetation canopy when flying the perimeter of search quadrats and weather permitting, at slower speeds and lower heights when searching the interior of quadrats. searching time per quadrat ranged from 10-50 minutes. Sighting bias trails were usually conducted in 2 sessions each day. Up to 8 different radioed elk were located via telemetry with a Cessna 185 during each early morning and mid-day session. Locations of elk were assigned to a previously determined search quadrat. Assignment to a quadrat was confirmed by a technician on the ground who checked Loran-C locations obtained by plane with latitude-longitude locations of quadrats plotted on USGS 7.5 minute topographic maps. The navigator was given a list of 5-8 target elk and quadrats per session but no information as to where target elk were located within quadrats. Usually within 2 hours of locating target elk with the plane, the helicopter team searched the quadrat and counted all groups of elk until either'the target radioed elk was seen or until the entire quadrat had been searched and the target elk not seen. If a target elk was missed, the helicopter team located the elk via telemetry immediately after completing their search of the quadrat to determine whether the elk was on or off the quadrat at the time they flew the quadrat. The helicopter team recorded initial group size, total group size, activity behavior, vegetation type, percent occlusion cover, and percent snow cover for the portion of each group initially detected whether or not a group contained a radioed elk (Appendix I). Snow type, light conditions, wind conditions were assessed for each quadrat. When radioed elk were observed, they were usually individually identified by the number/symbol code, and if a target elk, the search was then terminated. Several radioed elk were often seen per quadrat. After completing the search of a quadrat, observers located target elk that were missed, and, if found on the quadrat, observers recorded all sighting variables for the missed group. Usually 10-12 trials were conducted each day. Successful trials occurred when the target elk was on the assigned quadrat whether or not that elk was seen. Unsuccessful trials 'occurred when target elk were found off the assigned quadrat immediately subsequent to the time the quadrat was flown. Elk were off quadrats primarily due to their movements that occurred between the time elk were located with the plane and the time the helicopter team flew the quadrat. We wanted to avoid ~ny positive effects on sightability that might be gained if an observer located a target elk during 1 trial and then was assigned that same target elk during another trial. Thus, elk were targets for an individual observer only once. However, the same navigator was involved in 2 trials with each of 2 elk, but the 2 trials were generally 2-4 weeks apart and we suspect-the navigator had little influence on the ability of the .. ·.. observation team to detect -the target elk.' We subsequently targeted 90 different radioed ~lk: 70 were targeted 1 time and 20 were targeted 2 times during the 110 trials and for 106 trials,' there was a unique combination of observer and navigator attempting to detect the targeted elk. Sighting bias models were developed using logistic regression (Proc GENMOD, SAS). The dichotomous classifioation of groups seen or missed was the dependent variable. Initial group size, total group size, activity behavior, vegetation type, percent occlusion cover, and percent snow cover, etc. were independent variables. Significance level for independent variables was P ~ 0.05 for univariate tests and P ~ 0.10 for stepwise regression in multivariate models. Relative efficiency of models was assessed using AIC values (Akaike Information Criterion = -2[-log likelihood value] + 2[No. model parameters]). Observer, navigator, trap method, elk sex, elk age, activity, vegetation type, �32 snow type, and wind were treated as class variables and group size, percent occlusion cover, and percent snow cover were treated as continuous variables. Variances for estimates of total elk that were derived from sighting models followed a modified version of that presented by Steinhorst and Samuel (1989)(pers. comm. D. Bowden and G. White). Movements We located 34 radioed elk (17 .aduLt; females, 17 calves [9 males, 8 females] at least once per month since capture via telemetry using a Cessna 185. These elk were selected at random from within trap zones and equalized by age class. These elk will be monitored seasonally to document general movements and areas of use with results reported next year. RESULTS AND DISCUSSION Survival From 2 December 1993 through 15 June 1994, 6 calves died (4 male, 2 female) and 3 adult females died (Table 2). Overwinter survial rate (± 95% Cl) for calves was 0.918 + 0.063 and for adult females was 0.956 + 0.049. These survival rates were associated with a winter considered to be mild in both temperature and snow depth. Suspected' causes of death were primarily malnutrition and predation for calves and shooting and predation for adult females (Table 2). Those calves dying from malnutrition had marrow fat values < 13% at death and body weights < 102 kg at capture. The 2 smallest male calves weighed (Table 3) died of suspected malnutrition (Table 2). Male calves had larger body weights (P = 0.03), longer hind leg lengths (P = 0.07), and higher condition indexes (P = 0.03) than female calves. Total body lengths were not different between sexes (P = 0.36) (Table 4). Blood Assays Brucellosis tests were negative for all 43 adult females sampled. Progesterone assays indicated 25 (74%) of 34 adult females were pregnant (1.37-4.69 ng/ml), 2 (6%) were possibly pregnant (1.02-1.04 ng/ml), and 7 (20%) were not pregnant (~ 0.37 ng/ml) (Freddy 1989). Of the females judged not pregnant, 5 were judged to be 2-4 years old and 2, 5-9 years old. Sighting Bias We completed 99 successful sighting trials involving 45 adult females and 54 calves of which 26 were males and 28 were.females. Elk involved in 11 unsuccessful trials were 6 adult females.and 5 calves. Overall, observers missed 15.4% of the targeted elk (Table 5). The degree of negative sighting bias was not different among observers (P = 0.53) rang~ng from 12.2 to 21.9%. Therefore, the simple binomial correction factor (± 95% CI) for detecting elk was 0.846 ± 0.071. .. Of the 16 target elk missed during successful trials (representing 16 groups), 14 (88%) were calves (5M, 9F) and 2 (12%) were adult females. Frequency of missing male and female calves was not different between = 0.28). Calves missed were in groups of ~ 4 elk, except for 1 group elk. Capturing calves with a helicopter or corral trap had no effect sightability (P = 0.66). target sexes (P of 30 on their Univariate tests indicated elk age (calf or adult), initial group size, log(ln) initial group size, total group size, log(ln) total group size, activity behavior, vegetation type, percent occlusion cover, percent snow cover, and wind conditions affected the probability of sighting elk (P ~ 0.05, Tables 6, 7). The In total group size was the most efficient of these single variable models (AIC 74.678; intercept 0.1215, group size coefficient = = = �33 1.3073). provided Log transformations of either initial group size or total group more efficient models than untransformed values (Table 7). size The initial multivariate model for stepwise regression incorporated elk age, In total group size, activity behavior, vegetation type, percent occlusion cover, percent snow cover, and wind conditions. Vegetation type, percent occlusion cover, and wind were not significant (Model A, Table 8). Proceeding in a step-down process, wind was dropped as the least significant variable from model A, then percent occlusion cover from model B, to derive model C in which elk age, activity behavior, vegetation type, In total group size, and percent snow cover were all significant. However, model C necessitated estimates of coefficients for 12 parameters which makes this model unwieldy and likely subject to highly imprecise estimates of corrected numbers of elk although this model had a relatively low AIC value (Table 8). Other models having fewer parameters were assessed. In model D having 3 primary variables, elk age was not significant, but In total group size, percent occlusion cover, and the interaction of elk age and In total group size were significant (Table 8). This significant interaction term reflected that although calves were the predominate class of elk missed, they were almost always in small groups when found after completely searching quadrats. We are not sure whether calves were missed because they were in separate small groups (assumed, true sighting bias) or because they represented a fraction of a larger group that was detected but the calves were not seen (counting bias). Missed groups involving targeted calves were seldom the only group of elk on the quadrat. Model E includes ,In total group size and percent occlusion cover which mimics variables used in sighting models developed in Idaho for elk (Samuel et ale 1987). These models were surprisingly similar considering that the Idaho model was developed in tall conifer habitats and based only on sightability of adult elk (Samuel et ale 1987) (Table 9). Model F includes only ln total group size and had the highest AIC value but probability of detecting a group increased rapidly to ~ 0.90 for groups ~ 6 elk (Table 8, Fig. 1). The degree to which selected models corrected number of elk counted and the variances of corrected counts were assessed. The simple binomial correction, model G, inflated counts the most (18%) and had nearly the highest variance (Table 10). Inflated counts resulted from applying the 0.846 correction factor to large groups of elk which had a sighting probability of nearly 1.0 (Fig. 1) and needed no adjustment in numbers. Models E and F having In total group size or In total group size and percent occlusion cover had similarly low variances and low corrections for counts of 3-4% (Table 10). These 2 models highlight the need to include group size as an adjustment variable for correcting counts. Adding percent snow cover as an additional variable, model H, markedly increased the variance and provided minimal additional correction to counts (Table 10). We caution, however, that small variances and potentially precise estimates about corrected counts per quadrat based on adjustments for group size may not result in highly precise estimates of population size because of variance associated with total counts of elk among sample quadrats when a sampling system is employed (Otten et ale 1993). CONCLUSIONS We obtained acceptably precise estimates of calf and adult survival rates during winter and recommend continuing our current sampling effort of monitoring 75 radioed calves and 75 radioed adult females in 1994-95. Efforts to measure and develop criteria to adjust for negative bias in counts of elk were promising and we recommend conducting an additional 100 sighting bias trials in 1994-95 to refine a sighting bias model. We are particularly interested in evaluating effects of elk age (calves) on sighting bias and assessing potential heterogeneity of sighting probabilities for individual elk among years by resighting elk in 1994-95 that were involved in trials in 199394. �34 LITERATURE CITED Freddy, D. J. 1989. Effect of elk harvest systems on elk breeding Colo. Div. Wildl. Game Res. Rep. July(1): 35-60. biology. Freddy, D. J. 1993. Estimating survival rates of elk and developing techniques to estimate population size. Colo. Div. Wildl. Game Res. Rep. July: 83-117. Samuel, M. D., E. o. Garton, M. W. Schlegel, and R. G. Carson. 1987. Visibility bias during aerial surveys of elk in northcentral Idaho. Wildl. Manage. 51:622-630. Otten, M. R., J. B. Haufler, S. R. Winterstein, and L. C. Bender. 1993. aerial censusing procedure for elk in Michigan. Wildl. Soc. Bull. 73-80. SAS Institute Inc. 1988. SAS/STAT .Cary, NC. 1028pp. User's Steinhorst, R. K., and M. D. Samuel. for aerial surveys of wildlife White, Guide, 6.03. SAS Institute, by An 21: Inc., 1989. Sightability adjustment methods populations. Biometrics 45:415-425. G. C., and R. A. Garrot. 1990. Analysis of wildlife data. Academic Press, Inc., San Diego. 383pp. Prepared J. radio-tracking �35 Table 1Management Trap Zone A B c D E F G H Capture objectives and numbers Unit 42, December, 1993. Capture Objective Name Garfield Gibson Uncle Bob West Divide Hightower Middle Mamm West Mamm Dry Hollow 8 10 10 8 44 141 150 100 in 8 trapzones, Calves Males Females 3 1 8 5 10 7 6 5 4 3 0 0 3 2 9 7 f;lk CollaredTotal Helio Corral 8 0 8 19 6 25 29 0 29 0 21 21 0 17 17 0 0 0 6 0 6 35 0 35 20 24 26 All of elk radioed 41 36 Game Adult Females 4 12 12 10 10 0 1 19 68 37 • Helio - Helicopter net-gunning, Corral - corral-trap. Table 2. Causes of mortality December-1S June 1993-94. Radiocollar Fregyenc:l1: 172.090/93 172.542/93 172.690/93 172.899/93 173.000/93 173.262/93 173.289/93 173.461/93 173.469/93 for radioed elk in Game Sex Age· Date Dead Estimated Cause of Death F F 11 yrs 6 yrs 8 yrs 10 mos 11 mos 10 mos 10 mos 9'mos· 11 mos 1/16/94 2/01/94 1/24/94 3/18/94 4/25/94 3/18/94 3/22/94 2/07/94 4/25/94 Wounding Loss Cougar kill Gunshot, poached Cougar kill Malnutrition Possible predation Undetermined Malnutrition Malnutrition F F F M M M M Management Body Wt. (kg]b Unit Marrow Fat Percent D~ 42, 1 Matter not available l8.0c not available 28.5e1 12.6" 28.9e1 not available 2.1c 2.7" 86 102 108 III 82 77 Approximate age at death. b Whole body weight at capture in December 1993. " Fat content of either the right or left femur bone marrow. eI Fat content of lower jaw bone marrow. Table 3. Frequency distribution of whole body weights for male and female elk calves trapped in Game Management Unit 42, December, 1993. Percentage per weight class shown in parentheses. 90-99 Body Weight Class <kgl 100-109 110-119 Sex 70-79 80-89 Hale 1 (2.9) 1 (2.9) 3 (8.6) 6 (17.1) Female 1 (2.9) 4 (11.4) 8 (22.9) 12 (34.3) 12 (34.3) 4 (11.4) 120-129 10 (28.6) 130-139 1 (2.9) 6 (17.1) 0 (0.0) 140-149 Total 1 (2.9) 35 (100) o 3S (100) (0.0) �36 Table 4. December, Body measurements 1993. for elk calves trapped Male Calves SD Max Min Measurement Mean Body Weight (kg) Body Length (cm) Hindfoot Length (cm) Condition Index (Wt./Body Length) 112.7 191.2 56.3 0.59 13.7 10.2 2.2 0.05 77.0 164.0 52.0 0.43 141.0 210.0 61.0 0.67 in Game Management n -Mean 35 36 36 35 103.8 188.9 54.8 0;55 Unit 42, Female Calves Min Max SD 12.2 9.8 2.1 0.05 76.0 167.0 51.0 0.46 n 123.0 207.0 59.0 0.64 35 35 34 34 Table 5. Summary of elk sighting bias trials conducted in Game Management Unit 42, January-March 1994. Percentages shown in parentheses. Observer Trials· Trials Not Atte!!!l2tedUseable Trials Useable Trial Elk Detected Trial Elk Not Detected OM 29 36 45 3 4 4 26(90) 32(89) 41(91) 22(84.6) 25(78.1) 36(87.8) 4(15.4) 7(21.9) 5(12.2) Totals 110 11 99(90) 83(84.6) 16(15.4) JB JE Table 6. Elk sightability survey. results by major Game Management Unit 42, January-March, 1994. Ho, c[s!UI!!! !:u.ssed See!! VaJ::j,.abl!! Group va Size 6 7 1 1 0 0 0 1 0 1 2 3 4 5 6 7-15 16-30 31+ Vaqa. Type CleariDq/Aq Ripar.ian Sagabru.h O&Icbru.h PinyoOll-Juni.p&r A8pa1l Tall COnifer Occl. COVer(~) 0-19 20-39 40-59 60-79 80-100 • v-vI.Lbillty 8 10 11 6 5 5 17 14 7 0 0 0 5 6 3 2 9 1 3 40 26 3 1 1 3 6 3 3 (group. 22 25 23 12 1 seen / 0.57 0.59 0.92 0.86 1.00 1.00 1.00 0.93 1.00 1.00 1.00 1.00 0.89 0.81 0.50 0.33 0.96 0.89 0.79 0.80 0.25 groups VaJ::iable Behavior Bedded Standing Hoving Navigator CB VG Observer JB JE OM No, C£QUJ2S tUssed See!! Trap (All elk) COrral Trap Helicopter Trap (calves) COrral Trap Helicopter seen plu. groups va VaJ::i§ble Missed from See!! va (') 3 3 1 9 2 4 29 48 0.40 0.57 0.97 0.84 0.80 0.88 Snow Type Fresh Old 4 12 21 62 0.84 0.84 22 25 36 0.85 0.78 0.88 Wind Light Moderate Strong 16 0 0 67 14 2 0.81 1.00 1.00 43 40 0-.96 0.74 Light Bright Dull Hazy 12 0 4 52 20 0.81 1.00 0.83 2 53 28 0.29 0.86 0.93 - 11 5 45 38 4 7 5 5 9 21 19 0.81 0.68 4 12 27 56 0.87 0.82 3 11 11 29 0.79 0.73 m~ssed). variables Snow COver 0-19 20-49 50-99 100 5 9 2 Elk.·Age -Adult l1'ema1e 2 calf (M + F) 14 calf Sex MIlle l1'emale independent 11 �37 Table 7. regression 1994. Sununary of variables tested in univariate tests using logistic for elk sightability trials, Game Management Unit 42, January-March, Likelihood Ratio Probe >Chi-Square Variable 0.7152 0.5936 0.2763 0.5399 0.5650 0.6257 0.0022 0.5453 0.0166 0.0044 0.0027 0.0001 0.0014 0.0405 0.0010 0.0275 0.3127 0.0449 0.1233 0.9797 Date of Sight Trial Sight Trial Navigator Observer Individual Elk Sex of Elk Age of Elk Capture Type Initial Group Size Log(ln) Initial Group Size Total Group Size Log(ln) Total Group Size Elk Activity Vegetation Type Occlusion Cover % Snow Cover % Temperature Wind Conditions Light Conditions Snow Type a AIC - Akaike's Information Sign./ Nonsign. AIea . 91.450 91.299 90.398 90.351 12.318 91.345 82.170 93.218 85.844 83.468 82.589 74.678 82.435 90.417 80.703 86.725 91.023 89.376 91.397 93.583 NSign NSign NSign NSignNSign -NSign SIGN NSign SIGN SIGN SIGN SIGN SIGN SIGN SIGN SIGN NSign SIGN NSign NSign criterion Table 8. Multivariate analyses of factors affecting sighting probability based on stepwise logistic regression for elk in Game Management Unit 42, JanuaryMarch, 1994. No. Parameters Model A + Intercept B 13 C D E F 12 6 3 14 2 Parameters Useda Age, Act, Vege, Wind, LnGroup, Cover', Snow%· Age·, Act·, vege·, LnGroup·, Cover%, Snow%· Age·, Act·, vege·, LnGroup·, Snow% Age, LnGroup·, Cover%· LnGrOup·, Cover%· LnGroup • • • AIC Value 58.32 • 58.96 • 59.16 62.61 68.89 74.68 i Age-elk age class, Act-elk activity class, vege=vegetation type class, wind=wind class, LnGroup=log total group size, Cover%=occlusion cover percent, Snow\=snow cover percent, * Denotes variable as significant in model, P 0.10). bLower AIC value denotes better fitting model. =~ �38 Table 9. Comparison of logistic regression models incorporating log(ln) group size and percent occlusion cover or percent vegetation cover for sightability trials involving elk in Colorado and Idaho. variable Coeffiecent constant-colorado 1.98 constant-Idaho 1.22 In Group Size-Colorado 1.23 In Group Size-Idaho 1.55 % accl. Cover-Coloraod -0.04 % Vege. Cover-Idaho -0.05 a Calculated SE 0.86 0.67· 0.40 0.37· 0.02 o.oi- from Table 2 in Samuel et ale 1987. Table 10. Estimated variance components and estimated corrected numbers of elk for selected sightability models for elk in Game Management Unit 42, 1994. Variance Corrected Percent Model Estimate5 Covariancec Grou:e TotalQ Elkc Correction G-Binomial1 46,121 610 2,534 1,666 49,264 118 F-LnGroup 284 246 1,454 12 543 103 E-LnGroup+ 264 111 498 873 1,462 104 Covert H-LnGroup+ -805,323 806,283 283,588 284,549 1,489 106 Cover%+ Snow!!! • Variance from binomial sighting probability of groups. b Variance from predicted sighting probability~ Variance from covariance of predicted sighting probability; d Variance of estimate of total elk. f Binomial model uses simple sighting bias correction of 0.846 applied to all observed groups of elk. e Total elk observed in groups u~ed to develop models was 1,405 elk in 99 groups. A . C �39 1.0 Z 0·0.9 ...................... J(J 0.8 w tu 0.7 .: / C 0.6 ••' .1" •••• I I , MEAN LL. 00.5 >J- 0.4 ...J OJ 0.3 -c 0.2 o 0.1 OJ a: ................................................ 95 % CI a. o 1 1 3 2 3 LOG (In) TOTAL 7 20 7 456 GROUP SIZE 60 150 TOTAL GROUP SIZE FIG. 1. Probability of detecting elk by group size, + t- 95% CI, Game Management Unit 42, January - March, 1994. Group sizes observed during sighting bias trials ranged from 1 - 400 . ", ' 400 1100 �40 15 Appendix assi~ed A. I. Procedures for flying quadrats and definitions to each group of elk observed on quadrats. of variables PROCEDURES FOR FLYING QUADRATS: 1. OBTAIN PROPER MAPS, ARRANGE IN ORDER OF NEED, DECIDE ON GENERAL ROUTE TO QUADRAT, HIGHLIGHT QUADRAT BORDERS WITH FELT MARKER, AND DETERMINE LIKELY STARTING POINT. MAKE SURE NAVIGATOR HAS LISTING OF FREQUENCIES AND NECKBANDS TO DETECT (MINIMIZE ANY TALK ABOUT TARGET ANIMALS) AND THAT TELONICS TR-2 RECEIVER IS LOADED AND READY TO GO. CHECK TELEMETRY SYSTEM ON HELICOPTER WITH . DUMMYCOLLAR. 2. FLY PERIMETER OF QUADRAT FIRST IN A CLOCKWISE MANNER SO THE INSIDE OF THE QUADRAT IS TO THE RIGHT OF THE PRIMARY OBSERVER. 3. DETERMINE STATUS OF GROUPS OF ELK ON PERIMETER. A. ELK MOVING OFF THE QUADRAT WHEN DETECTED ARE CONSIDERED ON THE QUADRAT. B. ELK MOVING ONTO THE QUADRAT WHEN DETECTED ARE CONSIDERED OFF THE QUADRAT. C. IF A GROUP IS STANDING ON THE PERIMETER, COUNT THOSE ELK INSIDE THE QUADRAT. D. RESIST THE TEMPTATION TO LEAVE THE PERIMETER TO COUNT A GROUP ON THE INSIDE OF THE QUADRAT. USE YOUR BEST JUDGEMENT AT THE TIME OF DETECTION. E. IF A MARKED ELK IS DETECTED NEAR THE BORDER BE SURE TO NOTE ALL IMPORTANT DATA VARIABLES AS THIS ELK COULD BE THE TARGET ELK. 4. FLY INTERIOR OF QUADRAT. A. FLY THE INTERIOR OF THE QUADRAT SYSTEMATICALLY IN STRIPS OR STRIPCONTOURS. USE PROMINENT TERRAIN FEATURES TO DIVIDE THE QUADRAT INTO SMALLER COUNTING BLOCKS. IF DECIDE TO FLY STRIP-CONTOURS, WORKING FROM THE HIGHEST TO LOWEST ELEVATION USUALLY·VORKS BEST. B. BE PATIENT AND STRIVE FOR 100% COVERAGE. S. WHEN MULTIPLE GROUPS OF ELK ARE DETECTED SIMULTANEOUSLY, FOLLOW THESES GUIDELINES. A. PRIMARY OBSERVER SHOULD BEGIN OBTAINING DATA ON GROUP NEAREST THE HELICOPTER AND USE HELICOPTER TO KEEP GROUPS SEPARATED. B. NAVIGATOR SHOULD FOCUS ~OMENTARILY ON SECOND GROUP OBSERVED AND NOTE LOCATION, ·ACTIVITY AND NtiMB·ER ·OF ANIMALS FIR·ST SEEN. ., C. AFTER COMPLETING DATA COLLECTION ON FIRST GROUP, PROCEED TO LOCATION OF SECOND GROUP, DETERMINE VEGETATION TYPE, OCCLUSION%, AND SNOW COVER% AT SITE OF DETECTION, THEN FIND AND COUNT. SECOND GROUP. 6. RULES FOR MARKED ANIMALS A. FOR EVERY MARKED ELK SEEN (UNTIL THE TARGET ELK IS DETECTED) OBTAIN A POSITIVE IDENTIFICATION VIA THE NECKBAND NUMBER OR WITH TELEMETRY TO DETERMINE THAT THE MARKED ELK IS OR IS NOT THE TARGET ANIMAL. B. IF MARKED ELK ARE SEEN AFTER DETECTING THE TARGET ELK, OBTAIN AN IDENTIFICATION FROM THE .NECKBAND WITHOUT UNDULY HARASSING THE ELK OR USING TOO MUCH TIME. 7. RULES FOR DETECTING THE TARGET MARKED ELK A. YOU WILL ESTABLISH A FLIGHT PATTERN TO COVER THE ENTIRE QUADRAT. WHEN YOU HAVE COMPLETED THIS PATTERN, YOU ARE DONE FLYING THE QUADRAT. IF YOU HAVE NOT DETECTED THE TARGET ANIMAL DURING THIS PATTERN, DO NOT �41 16 RE-SEARCH THE QUADRAT AS THIS WOULD BE DISHONEST; IE, DO NOT SEARCH FOREVER TO FIND THE TARGET ELK, IT MAY NOT BE THERE OR THERE MAY NOT BE ONE TO FIND! B. IF YOU DO NOT FIND THE TARGET ELK AFTER COMPLETING YOUR FLIGHT PATTERN, TURN ON THE RECEIVER AND START SEARCHING THE QUADRAT FOR A SIGNAL. C. IF THE TARGET ELK IS ON THE QUADRAT, COUNT THE GROUP AS THOUGH IT WAS ANY OTHER GROUP OF ELK, AND BE SURE TO OBTAIN ALL VARIABLES ASSOCIATED WITH THAT GROUP. THIS IS THE MOST IMPORTANT DATA WE CAN COLLECT IN REGARDS TO SIGHTING BIAS ADJUSTMENT FACTORS AS IT REPRESENTS AN ELK YOU DID NOT SEE. D. IF THE TARGET ELK IS DEFINITELY OFF THE QUADRAT SPEND SOME TIME LOCATING IT BECAUSE IT COULD BE NEAR THE BOUNDARY AND MAY HAVE RUN OFF THE QUADRAT (BACKTRACK TO DETERMINE IF LEFT QUADRAT). SPEND ENOUGH TIME SEARCHING TO DETERMINE TO YOUR SATISFACTION THAT THE ELK WAS LIKELY OFF THE OUADRAT DURING YOUR TIME ON THE OUADRAT. E. NOTE NUMBER OF MARKED ANIMALS SEEN IN EACH GROUP, AND ID' S OF THOSE ELK IF OBTAINED. LOCATIONS OF ELK SEEN ON A QUADRAT WILL BE ASSUMED TO BE THE MIDDLE OF THE QUADRAT FOR PURPOSES OF DISTRIBUTION AND MOVEMENT INFORMATION UNLESS YOU OBTAIN A LORAN-C LOCATION FROM THE HELICOPTER. B. DEFINITIONS THAT APPLY TO EACH GROUP OF ELK DETECTED ON A QUADRAT. OUR DEFINITIONS ARE BASED ON THE HYPOTHESIS THAT DETECTING INITIALLY DETECTING SOME FRACTION OF EACH GROUP. ELK IS DEPENDENT ON 1. DETECTION GROUP SIZE: NUMBER OF ELK INITIALLY DETECTED WHEN AN OBSERVER SEES A GROUP. THIS NUMBER WILL REPRESENT SOME PORTION OF THE TOTAL GROUP SIZE. 2. ANIMAL ACTIVITY: THE ACTIVITY OF THE ELK INITIALLY DETECTED OR THE ACTIVITY OF THE MAJORITY OF THE ELK INITIALLY DETECTED. ACTIVITY CLASSES WILL-BE BEDDED-B, STANDING-S, MOVING-M. 3. VEGETATION TYPE: THE PREDOMINATE TYPE OF VEGETATION FOR 30 METERS AROUND THE INITIAL ELK SEEN. VEGETATION TYPES ARE: OK--OAKBRUSH-TALL MOUNTAIN SHRUB SG--SAGEBRUSH-(INCLUDING GREASEWOOD & RABBITBRUSH) AS--ASPEN PJ--PINYON-JUNIPER TC - - TALL CONIFER (DOUGLAS FIR, ENGELMANN SPRUCE, LODGEPOLE PINE) RP - - RIPARIAN (COTTONWOOD/WILLOW BOTTOMS) CA--CLEARINGS/AGRlCULTURAL FIELDS (CLEARINGS CAN BE NATURAL MEADOWS, CLEAR-Cu~S COVERED WITH SNOW, DRILL-PADS, OR ANY SIZEABLE OPENING THAT APPEARS TO BE A CLEARING DUE TO SHORT VEGETATION AND/OR SNOW COVER) 4. % OCCLUSION COVER: AN ESTIMATE OF THE PERCENTAGE OF THE OBSERVER'S VIEW BLOCKED BY VEGETATION INTERFERING WITH THE OBSERVERS ABILITY TO SEE THE INITIAL ELK DETECTED IN A SPACE OF 30 METERS AROUND THE ELK. OBSERVERS WILL ESTIMATE IN INCREMENTS OF 10%, IE; 0, la, 20, 30, ••• , 100. DATA WILL BE POOLED LATER. VEGETATION INCLUDES STEMS AND TRUNKS OF DECIDUOUS SHRUBS AND TREES, TRUNKS AND LEAVES OF EVERGREEN CONIFERS, LEAVES AND STEMS OF ANY OTHER EVERGREEN SPECIES. GRASS AND SHORT SAGEBRUSH SHOULD HAVE 0% OCCLUSION COVER �42 17 EXCEPT POSSIBLY WHEN ELK ARE BEDDED'~ 5. %SNOW COVER: THE PERCENTAGE OF BARE GROUND COVERED BY SNOW FOR 30 M AROUND ,THE INITIAL ELK DETECTED. OBSERVERS WILL ESTIMATE IN INCREMENTS OF 10%, IE;' 0, 10, 20, 30, ....100. DATA WILL BE POOLED LATER. 6 . TOTAL ELK IN GROUP: THE TOTAL NUMBER OF ELK COUNTED IN A GROUP. A GROUP IS DEFINED AS A CLUSTER OF ELK ACTING INDEPENDENTLY EITHER IN BEHAVIOR OR SPACE-FROM ANOTHER CLUSTER OF ELK. GUIDELINES FOR DEFINING A GROUP ARE: CLUSTERS OF ELK SEPARATED BY >30 METERS OR CLUSTERS SHOWING DIFFERENT PATTERNS OF BEHAVIOR. DEFINING A GROUP CAN BE SUBJECTIVE IF ELK ARE AT HIGH DENSITIES ON THE QUADRAT. THESE DEFINITIONS APPLY TO AMBIENT CONDITIONS EXISTING ON EACH QUADRAT AT THE TIME OF FLYING 1. WIND: WIND WAS: LIGHT-L, MODERATE=M, OR STRONG-S IN RELATION TO THE RELATIVE EFFECT OF THE WIND ON' THE ABILITY TO FLY THE QUADRAT IN A SLOW, LOW AND DELIBERATE MANNER. 2. LIGHT CONDITIONS: LIGHT CONDITIONS WERE EITHER: BRIGHT SUNSHINE WITH HIGH CONTRAST-B, HAZY SUNSHINE WITH HIGH THIN CLOUDS-H, OR DULL SUNSHINE DUE TO OVERCAST OF CLOUDS-D. 3. SNOW TYPE: SNOW TYPE IS EITHER FRESH SNOW THAT HAS FALLEN WITHIN THE PAST 48 HRS=F, OR OLD SNOW THAT IS OLDER THAN 48 HRS-O. SNOW CONDITIONS WILL LIKELY BE THE SAME FOR ALL GROUPS OF ELK DETECTED ON A QUADRAT. : 3. TEMPERATURE: TEMPERATURE (F) DURING THE MORNING AND AFTERNOON FLIGHTS AS MEASURED AT THE BASE OF OPERATIONS FOR EACH FLIGHT �43 Colorado Division Wildlife Research July 1994 of Wildlife Report JOB PROGRESS State of Project REPORT Colorado No. ~W~-~1~5~3~-~R~-_7~ _ Mammals Research Work Plan No 4 Moose Job No. 1 Development of census methods and determination of movements, habitat selection, and degree of calf mortality of moose in North Central Colorado. Period Author: Covered: July Investigations 1, 1993 - June 30, 1994 R. C. Kufeld Personnel: D. Bowden, D. Younkin. ABSTRACT Eighteen moose were captured and marked with radio-collars and eartags in North Park, Colorado, during January, 1994. Instrumented moose captured during 1991,. 1992, .1993, .. and-1994 .were located at .approximately 2-week intervals from January, 1992, through June, 1994. ��45 DEVELOPMENT OF CENSUS METHODS AND DETERMINATION OF MOVEMENTS, HABITAT SELECTION, AND DEGREE OF CALF MORTALITY OF MOOSE IN NORTH CENTRAL COLORADO Roland C. Kufeld P. N. OBJECTIVES 1. To determine the proportion of moose counting moose in North Park. 2. To determine 3. To determine the degree of dispersal of young animals, and seasonal movements, home range size, and habitat selection of North Park moose. the extent of moose SEGMENT the extent actually observed calf mortality when aerially in late winter. OBJECTIVES 1. To determine of moose calf mortality in late winter. 2. To determine the degree of dispersal of young animals, and seasonal movements, home range size, and habitat selection of North Park moose. STUDY AREA The study area was described by Kufeld METHODS (1992). AND MATERIALS Moose were captured throughout the eastern and southern part of North Park during January, 1994 by net gunning from a helicopter. Captured ~oose were fitted with 2 numbered eartags and a numbered radiocollar. Yellow eartags were used for moose captured on the east side of North Park and north of Highway 14, whereas orange eartags were used for moose captured on the east side of North Park but south of Highway 14 •. Each radio-collar was equipped with a mortality switch which will acitvate if the animal has not moved for 5 hours. Instrumented moose captured during 1991, 1992, 1993, and 1994 (Kufeld 1992, 93) were located at approximately 2-week intervals from January, 1992, through June, 1994, and plans call for such monitoring to continue for at least 1.5 more years. Most locations were made by aerial telemetry using a Cessna 185 aircraft with a 2 element, "H" configuration receiving antenna mounted on each strut. A switchbox permitted the telemetry operator, a passenger in the aircraft, to operate antennas jointly or separately. Some locations were made by tracking on the ground until the animal was observed when the airplane was not available. Moose locations were plotted on USGS 1:50,000 scale maps and recorded by UTM coordinates. Vegetation type was also recorded for each moose location. �46 RESULTS Moose capturing and monitoring Eighteen moose were radio-collared and eartagged in North Park during January, 1994 (Table 1). All, except for one adult cow, were calves born during the spring of 1993. Analysis of data for movements, home range size, and habitat use for all tagged moose will be presented in a future report when periodic monitoring of moose is completed. LITERATURE CITED Kufeld, R. C. 1992. Development of census methods and determination of movements, habitat selection, and degree of calf mortality of moose in North Central Colorado. Colo. Div. Wildl. Wildl. R~s. Rep. July:95108. . Kufeld, R. C. 1993. Development of census methods and determination of movements, habitat selection, and degree of calf mortality of moose in North Central Colorado. Colo. Div. Wildl. Wildl. Res. Rep. July:(in press). Prepared by Roland C. Kufeld Wildlife Researcher C �Table 1. Moose radio·collared and eartagged in North Park during January, 1994. collaf No. Unm.lnbered Unm.mbered Unnumbered Unnumbered Unnumbered Unnumbered Unnumbered Unnumbered Unnumbered 48 Unnumbered Unnumbered Unnumbered Unmrnbered Unnumbered Unnumbered Unnumbered Unnumbered 1 No. 73 93 79 86 87 94 78 90 74 88 85 84 80 98 99 81 82 83 Eartag co~or Yel. Yel. Org. Org. Org. Yel. Org. Yel. Org. Yel Org. Org. Org. Yel. Yel. Org. Org. Org. Sex F F F F M H H M H F F F F F F F H H Age Date Captured 2.5+ Calf Calf Calf Calf Calf Calf Calf Calf Calf Calf Calf Calf Calf Calf Calf Calf Calf 01-03-94 01-03-94 01-03-94 01-03-94 01-03-94 01-03-94 01-03-94 01-03-94 01-03-94 01-03-94 01-03-94 01-03-94 01-03-94 01-03-94 01-03-94 01-03-94 01-03-94 01-03-94 Capture Location Michigan R. 0.5 mi. NW Custer Mtn. Michigan R. 1 mi. W. Jackson Rd. 30. S.·Fork Michigan R. at Aspen Campground Michigan R. 0,25 mi. S. KOA Campground Michigan R. 1 mi. N. Custer Mtn. 1.5 mi. N. Custer Mtn. N. Fork Michigan R. 0.5 mi. E. KOA Campgr. Michigan R. 0.5 mi. NilCuster Mtn. Michigan R. 0.1 mi. SW Custer Mtn. Illinois R~ at Big Bottoms Indian Cr. at USFS Boundary Upper Illinois R. at Big Bottoms Middle Badgero Creek Spring Cr. 0.3 mi. N. Willow Cr. Rd. Spring Cr. 0.5 mi. N. Willow Cr. Rd. Illinois R. at Haney Oraw Jack Cr. 0.4 mi. E. of USFS Boundary Willow Cr. 0.2 mi. N. of Willow Cr. Road Location UTM Coords E-W N-S 409.43 406.93 412.31 411.82 409.35 411.06 412.81 409.78 410.25 410.14 405.44 411.84 404.13 401.28 401.53 406.22 408.79 399.99 493.05 497.27 485.51 488.93 493.59 495.78 489.80 492.60 492.07 473.41 481. 78 473.43 481.52 472.11 472.28 474.92 474.69 472.49 Prior to 1994 most radio-collars had a number visible from a distance. ~ ��49 Colorado Division Wildlife Research July 1994 of Wildlife Report JOB PROGRESS Colorado State of Project REPORT W-1s3-R-7 Mammals Work Plan No. SP1 Deer Job No. 1 Regulation of- Mule De·er Population Growth by Fertility Control: Laboratory, Field, and Simulation Experiments Period No. Covered: Authors: Personnel: July Research Investigations 1, 1993 - June 30, 1994 Dan L. Baker, A. C. Case, N. T. Hobbs, T. M. Nett, J. R. Ritchie, M. W. Miller M. A. Wild Abstract We conducted research to develop a practical and acceptable method of controlling deer populations at the Rocky Mountain Arsenal using GnRH-toxin conjugates as contraceptives. In this study, we integrate laboratory and field experiments to answer questions of efficacy, side effects and delivery ·and develop an interactive analytical model to allow wildlife managers to choose alternative contraceptive regimes on simulated populations before applying them to real animals. Here we report on two phases of this study; laboratory experiments with captive mule deer and-preliminary model development and analysis. We conducted two controlled laboratory experiments with captive, tame mule deer to determine the most effective dose of GnRH-toxin conjugate during the breeding season, anestrus, and pregnancy. In the first experiment, we measured LH secretion in cycling and anestrus mule deer challenged with five doses of GnRH analog (.3, 1, 3, 10, and 30 pg/sOkgBW). In the second experiment, LH response was determined in pregnant deer treated with six doses of GnRH (.5, 1, 2, 4, 8, and 16pg/sOkgBW). Animals were fitted non-surgically with indwelling catheters and blood samples were collected at 0, 30, 60, 90, 120, 180, 240, 300, 360, 420, and 480 min post-injection. Pituitary sensitivity to treatments of exogenous GnRH was greatest during the breeding season compared to anestrus. Peak LH concentrations were lowest for pregnant mule deer and the magnitude of the response decreased during gestation. Initial model results suggest that given the proper circumstances, fertility control can be a feasible and efficacious means of regulating growth when compared to culling. --. ��51 REGULATION OF MULE DEER POPULATION GROWTH BY FERTILITY LABORATORY, FIELD, AND SIMULATION EXPERIMENTS CONTROL: Dan L. Baker and N. T. Hobbs P. N. OBJECTIVES 1. To develop a practical and acceptable method populations using GnRH-toxin conjugates. 2. To demonstrate the feasibility the Rocky Mountain Arsenal. 3. To predict simulation population modeling. impacts for controlling of such control of alternative SEGMENT in a field contraceptive mule deer application regimes at using OBJECTIVES 1. Begin conducting controlled experiments with captive mule determine the most effective dose of GnRH-toxin conjugate preventing normal production of reproductive hormones. deer to in 2. Begin developing an analytical model that can be used by wildlife managers at the Rocky Mountain Arsenal to evaluate probable consequences of decisions on implementing fertility control technology to regulate mule deer populations. INTRODUCTION Controlling the growth of wildlife populations is fundamental to maintaining proper balance between animals and the habitats they use. Hunting has traditionally been used to maintain this balance; but there are an increasing number of circumstances where hunting wild ungulates to regulate their numbers is not feasible or where hunting is inconsistent with. goals for land management especially when populations are managed primarily for nonconsumptive uses. The Rocky Mountain Arsenal offers rich opportunity for an urban population to view and enjoy Colorado's wildlife. However, the Arsenal presents unique problems as well as unique opportunities. For reasons of security, the perimeter of the area was fenced in 1990. While this fence does not impede the usual movements of birds and small mammals, it does create an unnatural barrier to the movements by ungulates, particularly mule deer (Odocoileus hemionus). As a result, preventing those movements will almost certainly cause the deer population within the Arsenal boundaries to rise exponentially during the next decade. Experience with enclosed deer populations elsewhere has revealed that such increases will lead relentlessly to degradation of habitat, to widespread starvation, and eventually to catastrophic declines in animal numbers. The only way to prevent this trajectory is to control the abundance of enclosed deer populations. This is usually done by public hunting or by professional culling of animals. However, public hunting cannot be allowed on the Arsenal because of concerns for security. Moreover, although professional culling would eliminate excess animals, it would also reduce the value of the deer population as a watchable resource. In this situation, alternatives to hunting as a means of regulating ungulate numbers is needed. Fertility control offers a viable alternative to hunting as a means of population control when hunting is infeasible. However, current fertility ) �52 control technology does not provide a means of controlling ungulate numbers practically and economically. Here, we propose to develop a practical and economical method of fertility control in mammalian wildlife that overcomes many of the shortcomings of current technology, particularly problems of treatment duration and environmental safety. We propose to use conjugates of gonadotropin-releasing hormone (GnRH) and cellular toxins to selectively destroy gonadotropin-producing cells in the anterior pituitary gland, thereby preventing gamete production by the ovaries and testes. This research project consists of laboratory, field, and modeling phases, of which is designed to address questions that must be answered before widespread application of hormonal-toxin conjugates is possible. Laboratory each Phase The sexual cycle in mammals is completely dependent on the gonadotropin hormones secreted by cells in the anterior pituitary gland. Two important gonadotropic hormones include follicle stimulating hormone (FSH) and luteinizing hormone (LH). These two hormones control proper functioning of ovaries in females and testes in males. In all mammals studied, control of gonadotropin secretion is mediated by production and release of GnRH in the hypothalamus. By coupling a superactive analog of GnRH to a cytotoxin, it should be possible to specifically target that toxin to FSH- and LH- secreting cells in the anterior pituitary, thereby preventing successful reproduction. At present, manufacturing GnRH-toxin conjugates is done at laboratory rather than commercial scales. Because hormonal-toxin production is costly, it' is imperative to determine the minimum dose that will provide consistent effects. To accomplish this, analogs of GnRH that have a biological potency sixty times greater than natural GnRH will be linked to a toxin to form a potential chemical sterilant. Using an analog with greater biological potency will increase the efficiency of delivering toxin conjugates to gonadotrophs in the anterior pituitary, thus· reducing the quantity of toxin needed for treatment of an individual animal. During this phase of research, we conducted controlled experiments with mule deer to determine the most effective dose of GnRH-toxin conjugate. All studies during this phase were conducted at the Dept. of Physiology, Colorado state University, and the Colorado Division of Wildlife's Foothills Wildlife Research Facility, Ft. Collins, Colorado. Modeling Phase Applying GnRH-~oxin conjugates to control the growth of deer populations will require that wildlife- managerSl_choose specific-tactics for treating animals. Choices must be made on the number and age to treat, the frequency of treatment, and so on. Decisions on the best.tactics will depend on comparing the effects of alternative actions on population behavior. We will provide support for these decisions by developing an interactive model of deer population dynamics. This model will combine knowledge of deer biology with an understanding of the constraints intrinsic in the GnRH-toxin conjugate technique. We used an individually based approach to simulate effects of different contraceptive regimes on population dynamics. Parameters for the model were derived from laboratory experiments, from the mule deer literature, and from consultation with biologist and managers from the Rocky Mountain Arsenal. �53 METHODS AND MATERIALS Laboratory Phase Experiment 1: Breeding Season The amount of GnRH analog needed to induce half-maximal release of gonadotropic hormones is not known for wild ungulates. In order to estimate the dose of GnRH-toxih conjugate required for sterilization, it is essential to determine the potency of GnRH analog for each species of concern. This experiment was conducted during the peak of the breeding season to insure that the pituitary gland was at its most active state when stimulated by GnRH analog. We conducted experiments to measure pituitary response to GnRH analogs in mule deer during the breeding season (Nov 9-Dec 11). We estimated peak of conception to be about December 1. We measured LH responses of mule deer to five levels of GnRH analog in a randomized complete block design with a repeated measures structure. Levels of GnRH analog administered were 0, 0.3, 1.0, 3.0, and 10 micrograms/50kgBW. Based on information from studies on other mammals, we chose a range of doses of this analog that should include the biologically effective range for mule deer. All treatments were administered on the same day to all animals. Trials were conducted every 8 days. Five tame, sexually mature mule deer (3 non-pregnant females; 2 male castrates) were sedated with xylazine hydrochloride and fitted nonsurgically with indwelling jugular catheters. Concentrations of LH in serum were measured immediately prior to treatment and at 30, 60, 90, 120, 180, 240, 300, and 360 minutes post-treatment. Blood was allowed to clot at room temperature, then refrigerated for 24 hours prior to centrifugation. Serum was harvested the next day and frozen at -40 C. All serum samples were analyzed for LH concentrations using radioimmunoassay (Niswender et al. 1969). The LH response to analogs of GnRH was assessed by 1) estimating the maximum concentration released (peak of LH dose response curve), 2) time interval from injection of GnRH bolus to peak LH concentration, and 3) the total amount of LH released (estimated by calculati?g the area under the LH curve). ~ Experiment 2: Anestrus Cells in the pituitary are most susceptible to GnRH-toxin treatment when they are most active. This likely occurs during the peak of the breeding season, in late autUmn. For a variety of reasons, however, it may be necessary to treat animals in the field at other times of the year when animals are less reproductively active. This experiment will evaluate effects of season on performance of GnRH analog. If response of .the pituitary is similar to that observed during the breeding season, then GnRH-toxin conjugate could be administered without regard to season of year. We conducted experiments with mule deer to evaluate the responsiveness of the pituitary gland to stimulation by GnRH analogs during the anestrous phase of the reproductive cycle (April 4 - May 4). Treatment with GnRH analogs, blood sampling, and LH analysis were similar to the methods described in Experiment 1 with the exception that two additional sampling times (150 and 420 min) were added to better describe the LH secretion curve. Experiment 3: Pregnancy For free-ranging wild ungulates, effective treatment with contraceptives will require treating animals when they are most vulnerable to trapping. For mule deer, this will most likely occur during winter (Jan-Mar). At this time, most females are pregnant. In domestic animals, pituitary response to challenge with GnRH analog is lower during pregnancy than during the breeding season. �54 There have been no comparable studies with wild ungulates, thus the objective of this experiment was to assess the responsiveness of pituitary gonadotrophs in female mule deer during pregnancy and compare this response to that measured during, the breeding season and anestrous. This experiment was conducted during the stage of gestation when mule deer would likely be trapped most effectively (Jan 3 - Feb 7). Based on previous studies, we estimated peak of conception of captive mule deer to be approximately December 1. Approximately one estrus cycle (25 days) prior to this date (Nov 6), four sexually mature male deer were released into a pasture containing six female mule deer. Males were removed approximately one cycle after the estimated peak of conception. Pregnancy of female mule deer was confirmed using pregnancy specific-protein B (Wood et ale 1986) and transrectal ultrasonic scanning (Mulley et ale 1987). We determined the pattern of luteinizing hormone (LH) secretion following intravenous administration of GnRH analog to pregnant mule deer during December 15 to January 19, 1993. Animal handling, blood sampling and laboratory assay followed procedures described in Experiment 1. RESULTS Laboratory AND DISCUSSION Phase Cycling and anestrus mule deer challenged with analogs of GnRH responded with a measurable increase in serum LH by the first post-treatment blood collection (30 min), reached a peak about 2 h later, and then declined to pretreatment levels after 6 h post-injection. Time from injection of GnRH to peak LH response occurred between 180 and 237 min post-injection. During the breeding season, peak serum concentration of LH (23.09 ng/ml) was induced with a dose of l~g GnRH analog/50kgBW (Fig. '1). Peak LH response (13.64 ng/ml) during anestrus was achieved with a dose of 3 ~g GnRH analog/50kgBW (Fig. 2). No further increases in peak LH were observed with increasing doses of GnRH analog for mule deer during the breeding season or anestrus. Treatment of pregnant mule deer with GnRH analogs resulted in well-defined LH respqnse curves. Serum concentrations of LH in pregnant mule deer increased in response to GnRH analogs up to a dose of 4 ~g/50kgBW; then remained relatively unchanged with increasing doses. Pituitary response to GnRH showed a marked decline with advancing gestation (Fig. 4). Serum LH concentrations had declined by 86% over the course of the six week experiment. Pituitary sensitivity of mule deer was greater during the breeding season than during anestrus or pregnancy however, the magnitude of the GnRH-induced LH release was also more variable compared to the other reproductive states. Estrous cycles were not synchronized~ and as a conSequence treatments were administered randomly throughout the cycle. Therefo~e, variation in peak LH response may be largely attributable to the phase of the cycle when deer were challenged with GnRH and the influence of fluctuating concentrations of estradiol and progesterone (Goodman et ale 1980). The diminished responsiveness to GnRH observed during anestrus may be related to the inhibitory effects of longer photoperiod on the reproductive axis. Increasing photoperiod inhibits melatonin secretion from the pineal gland which in turn decreases secretion of GnRH from the hypothalamus. This leads to reduced LH and FSH secretion resulting in lack of ovarian activity. Low levels of estradiol and reduced secretion of GnRH both lead to a reduction in the number of receptors for GnRH and ultimately to decreased pituitary responsiveness (Gregg and Nett 1989). Collectively, this suggests that a larger dosage of GnRH-toxin conjugate will be needed to induce sterility in mule deer during anestrus compared to the breeding season. �55 To our knowledge, pituitary responsiveness to GnRH during pregnancy has not been previously reported for any wild ungulate. In the domestic ewe, however, several studies have documented a similar depression in pituitary sensitivity during pregnancy and progressive decline in pituitary response during gestation (Jenkin et al. 1977, Crowder et al. 1982). Furthermore, it has been reported in sheep that there is a constant percentage of LH released in response to a maximally stimulatory dose of GnRH. That is, more LH is released early in gestation because more LH is contained in the pituitary, not because of a change in pituitary sensitivity to GnRH. Our observations that peak LH release occurred at the same dose of GnRH throughout gestation suggest that a similar mechanism may occur in mule deer. Modeling ,Phase Preliminary modeling efforts were aimed at developing a simple compartmental model representing the dynamics of the mule deer population at the Rocky Mountain Arsenal. We then compared the effects of different contraceptive regimes on population dynamics relative to the effects of culling to regulate population growth. Model development and analysis are presented in Appendix A. LITERATURE CITED Crowder, M. E., P. A. Giles, C. Tamanini, G. E. Moss, and T. M. Nett. 1982. pituitary content of gonadotropins and GnRH receptors in pregnant, postpartum, and steroid-treated ovx ewes. J. Anim. Sci. 54:1235-1242. Goodman, R. L., and F. J. Karsch. 1980. Pulsatile secretion of luteinizing hormone: differential suppression by ovarian steroids. Endocrinology 112: 1286-1290. Gregg, D. W., and T. M. Nett. 1984. Direct effects of estradiol-17 on the number of gonadotropin-releasing hormone receptors in the ovine pituitary. Bio. Reprod. 40:288-293. Jenkin, G., R. B. Heap, D. B •.A. Symons. 1977. Pituitary responsiveness synthetic LH-RH and pituitary LH content at various reproductive in sheep. J. Reprod. Fertil. 49:207-214. to stages Mulley, R. C., A. W. English, R. J. Rawlinson, and R. S. Chapple. 1987. Pregnancy diagnosis of fallow deer by ultrasonography. Aust. Vet. J. 64:257-258. Niswender, G. D., L. E. Reichert, Jr., A. R. Midgley, Jr., and A. V. Nalbandov. 1969. Radioimmunoassay for bovine and ovine luteinizing hormone. Endocrinology 84:1166-1173 •. Wood, A. K., R. E. Short, A. E. Darling, G. L. Dusek, R. G. Sasser, and C. A. Rudder. 1986. Serum assays for detecting pregnancy in mule deer and white-tailed deer. J. Wildl. Manage. 50:684-687. Prepared by Dan L. Baker Wildlife Researcher C �56 32 - -E Cl c: 24 :c ...J ~ ca a.l 16 o, 8 o 10 1 0.1 LN GnRH Dose (micrograms/50kgBW) Figure 1. Peak serum concentrations of LH for female mule deer challenged with GnRH analogs during the breeding season. 20 16 c E Cl c: 12 :r: ...J .::.:: ca 8 III c, 4 o 0.1 1 10 40 LN GnRH Dose (micrograms/50kgBW) Figure 2. Peak serum concentrations of LH for female mule deer challenged with GnRH analogs during anestrus. �57 3 :::::E -- 2 Cl c: :I: -l .:.:: as· Q) c, 1 o L- ~ __ ~~ __ ~~~~L_ ~ __ L_L_~~L_ ~ 10 1 0.1 Figure ~_J LN GnRH Dose (micrograms/50kgBW) 3. Peak serum concentrations of LH for female mule deer challenged GnRH analogs during pregnancy. 20 with 3 :::::·E -CI c: :I: 2 ...J .:.:: as CD a, 1 60 68 76 84 Gestation (days) Figure 4. Average serum concentrations of LH in pregnant with GnRH analogs during gestation. 92 mule deer treated 100 ��59 APPENDIX A Regulation of Mule Deer Population Growth by Fertility Control: Field, Laboratory and Simulation Experiments Progress Report-1993 We developed analytical models to describe dynamics of the Rocky Mountain Arsenal mule population. Our objectives were: 1) To estimate carrying capacity of the Arsenal using existing data on population composition and performance. 2) To estimate the proportion of the population that must be infertile to maintain the population at approximately half of carrying capacity. 3) To estimate the delivery rate and the:number of animals that must be treated annually to maintain the population at half of carrying capacity. 4) To compare the effort required to stablize the population using liftetime duration contraceptives with the effort required to stabilize the population using culling. Here, we summarize our progress toward these objectives. Model development and analysis are described in detail in Appendix A. Models are based on a very approximate relationship between fawn survival and the total number of females in the population (Figure 1). This approximation was derived from studies of Don Whittaker on the Arsenal population. Using on this relationship, we estimated the carrying capacity of the Arsenal to be about 525 females or about 1050 in Fawn Survival vs Total Female Numbers 1 51 100 200 300 400 500 soo N Figure1. Hypoth_ized relationship between the per capita rate of survival of fawns (S1) and the total number of females (N) in the mule deer population at the Rocky Mountain Arsenal. in total. It should be understood that this is a very crude estimate. �60 Models predict that approximately 66% of the population must be infertile to stabilize the population at half of carrying capacity (Figure 2). However. the proportion of the fertile population Number Females at Steady State vs Proportion Infertile 600 400 Neq 300 200 100 0.2 0.4 0.8 0.6 PP.a Figure 2. Steady state population size (Neq) in relation to the proportion of the population that is infertile (Peq). that must be treated annually to maintain a given population size is much smaller than the proportion that must be infertile toachieve a given reduction in population size (Figure 3). This is the case because mule deer does are long lived. When the effects of contraceptives are permanent. then the proportion of animals that are infertile will always exceed the delivery rate required to maintain that proportion in the population (Figure 4). Population Density vs Delivery Rate 600 500 400 Neq 300 200 100 o._------------------------------~---o 0.1 0.7 r. o.s 0.4 O.!i Figure 3. Steady state population size (Neq) in relation to proportion of the fertile females that are treated annually (c). �61 Proportion Infertile vs Delivery Rate r. Figure 4. Proportion of the population infertile at steady state (Peq) in relation to proportion that are treated annually (c, femaleJfemalelyear). of the fertile females Based on modeling results, we estimate that fewer animals will have to be to be treated with contraceptives than will have to be culled to maintain a given target density for the population. For example, maintaining the population at 250 females (i.e., about 112of carrying capacity) would require shooting 71 females/yr but would require that only 24 animals be treated with lifelong contraceptives (Figure 5). �62 Number Treated or Culled Annually vs Target Population Size 80 100 200 Neq 300 400 500 Figure 5. Number of animals that must be treated or culled annually to maintain a target, steady state population size (Neq). The upper curve gives the number of animals that must be culled. The lower curve gives the number of animals that must be treated. . > �63 %2 ( 1-S2 - %2 S2 -mSI---- m Sl %2 %1 J %1 ~tH:=--~--~--------------~--~ A % 1 ( m S1 fJ %1 :=NeqmSI mSI ~ %2 + ----"---- J %1 ~-S2 -mSI %2 := -Neq m Sl ~ - 1 + S2 + m Sl o Derivation and Analysis of Contraceptive Model: To derive a model of the population regulated by fertility sometime during the 'interval t--->t+ I; I define the number of animals treated (trtL) as trtl. :» c (F[tl (1 - ~ N) Sl + S2 A[t]) It follows from this equation that the delivery rate, c, is the number of females treated per surviving female. I presume that some animals are treated before breeding occurs, thereby preventing them from breeding and from contributing to the number of fawns at t[+ 1]. I define alpha as the proportion of animals that are treated before breeding occurs. It follows from these ideas that the number of adults that will not produce fawns at t+1 is the number that are treated by contraceptives before breeding. Define this number as trtB: trtE := aA[t] S2 c Thus, the number of fawns at time t+ 1 is a function of the number of adults at time t, the rate of delivery of contraceptives, c, and the timing of delivery, alpha: F[t+ II = mA[t]- F[t+ II =A[tl aA[tl S2 c (m - a S2 c) The number of fertile adults at time t[t+ 1] is a function of adult and fawn survival during t--t+ 1 and the. rate of delivery of contraceptives to adults and fawns: A[t+ II = F[tl R Sl A[t+ II = F[tl R Sl + S2 A[tl - trtL + S2 A[tl- c (F[t] (1 - ~ N) Sl + S2 A[tl) A[t+ 11 = (R Sl - c (1 - ~ N) Sl) F[tl + (S2 - c S2) A[tl The number of infertile adults at time t+ 1 (i.e., I[t+ 1]) is simply the number of infertiles . that survive during t-->t+ 1 plus the number that are treated: I[t+ 11 = I[tl S2 + trtL �64 I[ t + 1] = I[ t] S2 + c S1 F[ t} - cS 1 F[ t] !3 N + c S2 A[ t] As before, I place these in matrix form: S2 - a. S2 c Om C=, SlR(1-c) [ S2(1-c) c R Sl c S2 v = [F[t]' A[t]' I[t]] To simplify things a bit, I assume that all animals are treated after breeding has occurred (i.e., alpha = 0). As above, I solve for the steady state, solL, by finding the characteristic polynomial for C, setting Iarnbdae l, and solving for N, thereby obtaining: a.:= 0 soll: 1 - S2 + c S2 - m Sl + m SIc :=---------mSl!3{-l+c) The rate of treatment CcL) needed to stabilize a population at a given target density (=Neq) is solved by setting the rhs of solL=Neq and solving for c Neq m Sl !3+ 1 - S2 - m Sl cL=--~--~----------Neq m Sl !3- S2 - m Sl Remember that this particular solution for cL depends ori alpha = 0.. I now find the composition of the steady state population. I substitute solL for N in C, and find the dominant eigenvector (ev): I -S2 + c S2 - m S1 + m SIc N:=-------------------mSl!3(-I+c) m o S2 (1 - c) 0 cS2 S2 o S1 ( 1 - 1- S2 + c S2 - m Sl + m Sl c ) c ( 1sol:= [ 1,1, { [ m Sl (-1 + c) (1 - c) 1 - S2 + c S2 - m Sl + m Sl c ) m Sl (-1 + c) m p -S2 - c + c S2) . 1 - S2 - c + c S2 ccI, 1-S2+cS2 -1 + S2 - c S2, 1, 1 - ---{[ [ . m Sl ] }] [S2, 1, {[O 0 I]}], 2 ev := [ m (l - S2 - c + c S2 ) 1 - S2 - c + c S2 ccI 2 2cZ)]}] C(1-2S2+2CS2+S2 -2CS2 +S2 ____;,------------___;_ m(-I+c-cS2+S2c2) ] , �65 Colorado Division Wildlife Research July 1994 of Wildlife Report JOB PROGRESS State of Project Work Colorado No. Plan W-1S3-R-7 No. Job No. Period REPORT Covered: Authors: Personnel: Mammals Research 1A Multispecies 1 Animal and Pen Support Facilities for Mammals July Investigations Research 1, 1993 - June 30, 1994. M. A. Wild and W. S. Graffam. T. R. Anderson, J. A. Yost. D. L. Baker, P. E. Bleicher, A. L. Case, and ABSTRACT The Colorado Division of Wildlife's Foothills Wildlife Research Facility (FWRF) maintained captive animals (up to 128 wild ungulates of 5 species and 70 migratory game birds of 4 species) and facilities supporting 4 major mammalian research projects. Financial support was provided by various federal agencies for maintenance of a portion of the mule deer herd, elk calves, and white-tailed deer. In fall 1993, we recruited 37 individuals into our captive herd. Ad libitum evaporated milk was fed to white-tailed deer for the first time and appears to be a safe and effective diet. During FY 1994, mortalities occurred in 9 adults and 7 neonates. Additionally, 24 individuals were transferred (18 permanently and 8 temporarily) to other research facilities. FWRF was inspected by USDA APHIS and found to be in compliance with federal animal welfare standards. A revised Chronic Wasting Disease protocol was implemented and appears functional and effective. Routine animal care and facility maintenance programs were conducted as previously described with an emphasis on quality and conservation. Numerous improvements were made to facilities at FWRF to increase usefulness, efficiency, quality, and/or safety. A computer database was constructed to assure that important information on research animals is recorded and that the information is easily accessible for reference. currently, vital information, health history, reproductive history, and animal use on experiments is contained in the database. Phase I of the experiment to provide oral vitamin E supplementation to bighorn sheep was completed. We did not observe differences in vitamin E levels between treatment groups (P=0.2S); vitamin E levels in all bighorn sheep remained at levels below those observed in normal domestic sheep. However, level of vitamin E supplementation was not strictly controlled in this experiment and the sample size of lambs was low. High copper levels were observed in all bighorn_sheep and were attributed to high level of copper in pelleted supplement. Copper level in pelleted feed for bighorn sheep was reduced from 66 ppm to 11 ppm based on these data. Phase II of the experiment was initiated. Strict control of vitamin E intake and increased sample size will provide further information on response of bighorn sheep to vitamin E supplementation. ��67 ANIMAL AND SUPPORT FACILITIES MAMMALS RESEARCH FOR Margaret A. Wild and Wendy S. Graffam P. N. Objectives 1. To provide and maintain captive wildlife populations and facilities supporting CDOW's· Terrestrial Wildlife Research Program, as well as programs of CDOW cooperators. 2. To develop improved animal and facility management practices provide maximum research opportunities at minimum cost. 3. To enhance needs. facilities to serve a growing SEGMENT improve, and expand diversity that will of anticipated research OBJECTIVES 1. Maintain, 2. Coordinate 3. Maintain up to 30 elk, 35 mountain sheep, 20 pronghorn, 50 mule deer, 12 white-tailed deer, and 70 ducks in suitable health to perform required research experiments, and in accordance with federal and institutional animal welfare regulations. 4. Conduct management experiments to increase feeding and maintenance activities related operations. 5. Follow a conservation-oriented approach for providing services to operate research facilities. 6. Establish a computer data base so that information complete and easily accessible. all rearing, animal training, METHODS research maintenance, and holding and research facilities. activities. efficiency and efficacy to research facility utilities on research of and animals is AND MATERIALS Routine animal care and facility maintenance programs supporting new and ongoing Terrestrial Wildlife Research Program projects were conducted as previously described. We. emphasized a .quality and conservation-oriented approach in this work by striving to increase efficiency and longterm benefit from the programs and projects undertaken. Specifically, we performed the following tasks: ANIMAL MAINTENANCE General: Again this year, routine feeding and caretaking of research animals, including health observations, training, weighing, and clean-up, was performed primarily by well trai~ed work-study and temporary employees, as well as volunteers. A database was constructed using Paradox for Windows (Version 1.0) to manage information on research animals at FWRF. The database consists of files ~n vital information, clinical health, reproductive history, and animal use on experiments. �68 FWRF was inspected by USDA APHIS regulations on 5 May 1994. for compliance with federal animal welfare The US Fish and Wildlife Service provided financial support to maintain 37 mule deer at FWRF. The USDA Animal Damage Control provided financial support for 12 white-tailed deer at FWRF. ADC also financed construction of 1 new animal paddock for housing white-tailed deer. NUTRITIONAL MAINTENANCE Feedina protocols: Pronghorn and mule and white-tailed deer continued to receive longstem alfalfa hay as their primary diet. Bighorn sheep received grass-alfalfa mix hay. Elk received cubed alfalfa as their primary diet with grass hay supplemented to meet scratch requirements. All ruminants were supplemented with pelleted feed (Baker and Hobbs 1985) to meet energy requirements and maintain ideal body condition (Wild et ale 1992). Bottle-raising neonates: Twelve white-tailed deer, 8 mule deer, and 4 pronghorn fawns were hand-raised at FWRF in 1993. Fawns received ad libitum evaporated milk until 50-60 days of age. At that time, milk offered to fawns was limited to 900 ml/day. Thereafter, the amount of milk offered was reduced by 100 ml/day each week until fawns received 300 ml/day. Feeding was maintained at 300 ml/day for several weeks to continue training practices. Five additional mule deer fawns that were hand-raised by rehabilitators were received by FWRF at weaning. Additionally, 6 male mule deer fawns were damraised by tame does at FWRF. Two orphaned pronghorn fawns are being handraised in 1994. Response of captive Rocky Mountain bighorn sheep to oral vitamin E supplementation: Methods and design of the first phase (conducted October 1992-0ctober 1993) have been reported (Wild 1993). Statistical analyses were performed on SAS using GLM procedures to determine significance levels. To further study effects. of vitamin E supplementation in bighorn sheep, Phase II of the experiment was initiated in December 1993. Levels of vitamin E and copper available to bighorn sheep were more rigorously controlled in this phase of the experiment, and sample sizes were increased. Ten ewes were paired and placed in breeding groups with 2 rams per group. After ewes were removed from the breeding groups, they were placed into low and high vitamin E level groups. Low vitamin E group diet contained approximately 63.40 IU/day and the high vitamin E group received 102.54 IU/day. Copper levels were reduced to 11 ppm in pelleted supplement. Bighorn sheep were fed a controlled amount of vitamin E containing pelleted feed (500 g/head/day) over the length of the study. Energy needs were met through an increase in pelleted feed which contained no vitamin E. To ensure that vitamin E and copper levels were similar to those calculated, hay and pelleted supplement were analyzed monthly for these elements and vegetation in the pastures was also monitored and analyzed periodically. Sample analyses were similar to that of the initial phase of the study. HEALTH MAINTENANCE General: A revised protocol for managing Chronic Wasting Disease (CWO) at FWRF was finalized (See Addendum A). The revised protocol incorporated basic ideas and practices from the original document, but acknowledged changes in management since the diagnosis of CWO in 3 captive elk. Neonatal illnesses: All 4 pronghorn hand-raised at FWRF in 1993 died in October. Clinical signs included acute onset of mild diarrhea, muscular weakness, and depression. Three fawns showed clinical signs for 8 to 48 hr prior to death; 1 fawn survived 10 days while receiving supportive care. �69 FACILITY MAINTENANCE/REPAIRS/IMPROVEMENTS A variety of scheduled and unscheduled maintenance and repair activities were necessary to support facility operation and ongoing research programs. We worked toward a conservation-oriented approach for facility care by undertaking preventive maintenance projects, and performing high-quality new construction and repairs to existing facilities. RESEARCH PROJECTS Facility operations offered support for pilot studies, student special studies, and CDOW and cooperative research experiments that were initiated, conducted, or continued using FWRF animals and facilities throughout the year. EDUCA~IONAL CONTRIBUTIONS Facility tours and educational lectures were provided to high school, university, and professional groups visiting FWRF. We emphasized the importance of maintaining captive wildlife for performing controlled experiments and the contributions made by research projects performed at FWRF. FWRF animals and facilities were also used occasionally for hands-on training for professional groups. RESULTS ANIMAL AND DISCUSSION MAINTENANCE General: When volunteers were carefully selected and trained similarly to paid employees, their contribution was remarkable. Nine volunteers contributed a total of 536 hours of work (range 2.5-160 hr) at FWRF during FY 1994. Volunteers performed primarily caretaker tasks and also assisted in weighing and collecting samples from animals. Their contribution represented a savings of about $5114 toFWRF (vs •.. cost .ot: -temporary employees). . The Paradox database was constructed and contains information on all animals that have been maintained at FWRF since 1991, the time when complete written information is available. Additionally, the database contains incomplete information on events prior to 1991, but only for animals living at FWRF after 1992. Maintenance of this database will assure that important information on research animals is recorded and that the information is easily accessible for reference. The animal welfare inspection by USDA APHIS revealed no deficiencies in research animal care or facilities. No violations have been identified in the last 2 inspections; this reflects the high level of dedication and training of FWRF employees and volunteers. In fall 1993, we recruited 3 bighorn sheep, 19 mule deer, 11 white-tailed deer, and 4 pronghorn into our captive herd. Nine elk calves were transferred to the USDA National Animal Disease Center, Ames, Iowa. Four fawn and 5 adult male pronghorn were transferred to Wyoming Game and Fish Department's Sybille Wildlife Research Unit. An additional 6 adult female pronghorn were temporarily relocated to Sybille. At the close of FY 1994, FWRF housed 32 bighorn sheep, 20 elk, 9 pronghorn, 43 mule deer, and 11 white-tailed deer. We complied with RFAC imposed population limits on number of animals financially supported by FWRF's budget by obtaining alternate funding for 3 bighorn .sheep, 32 mule deer, and all white-tailed deer. NUTRITIONAL MAINTENANCE Feedina orotocols: All species maintained reasonable body condition on available diets. OVerconditioning in elk has been controlled by limiting all feedstuffs, especially during winter, and in bighorn sheep and pronghorn by �70 limiting pelleted supplement. Mule deer continue to maintain generally fair body condition; however, alternate feedstuffs that more closely resemble native diets should be investigated in an attempt to improve nutritional condition of mule deer. Bottle-raising neonates: Growth of white-tailed deer (Fig. 1) hand-raised on ad libitum quantities of evaporated milk appeared similar to that reported for dam-raised fawns (Silver 1961, Robbins and Moen 1975). Although white-tailed deer performed well, and pronghorn and bighorn sheep neonates hand-raised previously with ad libitum evaporated milk grew at rates comparable to damraised neonates (Wild et al. 1994), mule deer raised using this protocol are consistently smaller than expected (Fig. 2). In 1993, all fawns were limit fed starting at about 60 days of age. The change in feeding protocol resulted in a more gradual weaning of fawns as compared. to previous years (Wild et al. 1992, Wild et al. 1994). Weaning weights of fawns raised on the standard method and the gradual weaning method did not appear to differ (Fig. 3). The gradual weaning method is more similar to natural weaning and should stimulate fawns to consume more dry feeds prior to weaning. Response of captive Rocky Mountain bighorn sheep to oral vitamin E supplementation: In Phase I, no differences in vitamin E values were observed between groups of bighorn sheep supplemented with low and high levels of vitamin E (P=0~2487, n=10) (Fig. 4). Values of all bighorn sheep were below the recommended normal of 500 ug/dl for domestic sheep (Hamar, pers. comm.). Although no statistical differences were seen between the high and low groups, individual variation was significant (P=O.OOOl, n=10) and may have masked treatment effects. Seasonal effects in vitamin E levels were observed. These may have been associated with increases in the amount vitamin E containing pelleted supplement being fed to meet physiological needs of pregnancy and lactation. Also, fresh vegetation was available in the enclosures as spring arrived. This vegetation was not analyzed for vitamin E or copper content. Cholesterol to vitamin E ratios are suggested to be a more accurate measure of the actual status of animals, especially non-domestics (Dierenfeld and Jessup 1990; Dierenfeld and Traber 1992). These ratios are suggested to be approximately 1.0-3.0 ug/mg when vitamin E is adequate. Bighorn sheep on this study had vitamin E:cholesterol ratios of about 0.3-2.2 ug/mg. This suggests that when blood lipids are considered, the vitamin E values fall more closely to the normal range. .Copper values between the high and low vitamin E groups were not statistically different (P=0.5803, n=10). Individual variation was significant (P=O.OOOl, n=10) and may, again, be masking some of the group effects. Plasma analyses revealed that the sheep appear to be chronically copper toxic as compared with normal values for domestic sheep (Hamar, pers. comm.). Plasma copper levels were consistently >1 ppm, and means over time were about 1.2 ppm. Liver copper levels are most indicative of· true body copper, and q liver.copper level of 621.20 ppm (normal for domestic sheep is 100-400) was found in one bighorn sheep that died of unrelated causes. Domestic sheep have been shown to be highly sensitive to the amount of copper present in their diets. NRC (1985) finds sheep to be vulnerable to even low levels of copper, such as 2535 ppm per day. Bighorn sheep in this study received 38.12 ppm/day when maintenance levels of the pelleted supplement were being fed. This high level of copper may increase the number of free radicals that are produced in the body and increase the requirement for antioxidant nutrients, such as vitamin E. The ewes may have increased vitamin E requirement due to the elevated copper in the diet, consequently, they may not have received the elevated level of vitamin E which was required for their increased needs even with supplementation. Plasma copper levels between the 2 groups showed no significant differences therefore affirming that both groups were receiving the same level of mineral and did not have different plasma levels due to vitamin E supplementation. The chronic copper toxicity in this particular herd is believed to result from the amount of copper supplied by the pelleted supplement. Based on these data, the amount of copper supplied in the pelleted diet was reduced from 66 ppm to 11 ppm for all. bighorn sheep at FWRF. �71 Data were collected on only 3 lambs (2 in low vitamin E group and 1 in high vitamin E group). Due to the small sample size, lamb data was not statistically analyzed. None of these lambs showed clinical pneumonia. optical density analyses for Pasteurella spp. are pending. Hay analyses were performed after the study was completed. These analyses showed an unexpectedly high level of vitamin E (29.27 IU/kg hay). Further, peileted supplement was calculated to contain 160 IU/kg and 500 IU/kg at the onset of the study. Upon analysis at the end of the study, it only contained 40.27 IU/kg and 98.54 IU/kg, respectively. Therefore, sheep were not receiving the desired of amount vitamin E from feed, and vitamin E levels were more similar in the 2 groups than anticipated. The results from this study showed no statistical differences between the high and low vitamin E groups. ·Possible causes for this include, inordinately high levels of copper in the diet and poorly controlled vitamin E intake. These factors serve to decrease the ability to find statistical differences between the groups. These factors have been controlled in Phase II of the experiment. In May 1994, 8 lambs were born to treatment ewes, 5 in the low and 3 in the high vitamin E group. We are currently monitoring and collecting samples from these lambs and ewes. HEALTH MAINTENANCE General: The revised CWO protocol was implemented. functional and effective. No CWO cases or suspects 1994. The plan appears were identified in FY Overall, captive wildlife maintained at FWRF remained healthy throughout the year. During FY 1993, 9 mortalities occurred in adult hoofstock and 7 in neonates at FWRF (T.able 1). An unident.ified illness resembling hemorrhagic .disease accounted for 4 of the neonatal mortalities. Chronic diseases were responsible for euthanasia of 7 of the adult animals. A minor but widespread anomaly occurred in most white-tailed and adult female mule deer.· The index case occurred in late December 1993 in an 8 mo old white-tailed deer male (C93). He showed bilaterally symmetrical truncal alopecia. Hair loss began in patches and progressed to large areas of thinning and loss between the shoulder and rump. There was no skin involvement apparent and no pruritus. Fungal culture and skin scrapings for parasites were negative. The buck remained in good health otherwise. In late January in a separate section of the facility, 2 female 8 mo old white-tailed deer and 3 female 1.5 yr old mule deer showed similar clinical signs. By March, 10 of 11 white-tailed deer (male, castrated male, and female) in 2 pens and 13 of 15 female mule deer in 1 pen were affected. Thirty mule deer in 4 other pens remained unaffected. Outbreaks did not occur in adjacent pens, and unaffected pens occurred next to pens with affected animals. In general, hair loss remained patchy in mule deer, but was more extensive in white-tailed deer. Partial regrowth of hair was usually apparent about 1 mo following initial hair loss. Repeated fungal cultures and skin scrapings for bacterial culture and parasite examination were negative. Skin biopsy results were nonspecific and did not identify etiologic agents. A metabolic/endocrine derangement was suggested, but not identified. By late April 1994, regrowth of hair had occurred and no further sign of disease was noted. Neonatal illnesses: Postmortem findings on the 3 acute cases of mortality in pronghorn fawns were nonspecific and etiologic agents were not identified. Due to the vascular nat.uxe of the lesions (hemorrhage and serositis), an acute vasculopathy was suspected; however, virus isolation and serology were . negative for bluetongue virus/epizootic hemorrhagic disease, as was FA for IBR, BVO, BRSV, and PI-3. Inclusion bodies suggestive of Adenovirus were absent. Death of the fourth fawn was likely due to metabolic sequelae of severe emaciation and dehydration from the illness. Postmortem signs were not �- .~ 1_ consistent with those in acute cases. No etiologic agents were identified. Serologic responses to bluetongue virus (as measured by AGID) were not identified in adult pronghorn held in pastures adjacent to fawns. Tentative diagnosis remains hemorrhagic disease. FACILITY MAINTENANCE/REPAIRS/IMPROVEMENTS In addition to numerous daily repairs and maintenance projects, we performed several major improvements. Significant maintenance/ repair/improvement projects completed at FWRF this year included: - Construction of a new hay storage area with capacity for an additional 35 tons of hay, and repair of existing hay.shed. - Construction of 2 new feed storage buildings to service deer pastures and elk pastures on the west side of the facility. - Construction of a new animal paddock (pen W) on the west side of the facility for white-tailed deer. - Construction of a divider fence in mule deer pasture B (to create pen Bl and B2). Construction of animal shelters in pens Wand B2. - Construction of covered feed areas for pastures W, A, Bl, B2, and C. - Construction of cut-off fences to provide catch pens in pastures B1, B2, and W. - Renovation of east side scale room. Installation of electric service to east side pens, handling areas, and feed storage areas. - Revegetatiori of pens E1, E3, E6, and E7. Replacement of pen fencing and visual barrier in pens E1, E6, and E7. - Addition of gravel to portions of east and west side alleyways. - Replacement of roof on west side scale room. - Removal of deteriorated elk calf rearing pens ahd replacement with a small elk holding area. - Installation of new carpet and baseboards in facility office. Replacement of motor and repairs to pump of the booster system that supplies water to office and west side of the facility. - Prairie dog depopulation and control measures for pens W3, W4, W7, and pasture F. _ - Tree planting along north perimeter fence for landscaping and visual barrier. . - Addition of road base to roughest/muddiest portions of facility roads. - Repairs to old roofs and fencing after wind damage. RESEARCH PROJECTS In addition to ongoing facility management experiments and improvements described above, the following pilot studies, special studies, and research experiments were initiated, conducted, or continued using FWRF animals and facilities this year: - Regulation of mule deer population Nett, Miller, Hobbs, and. Gill. growth by fertility Effects of yeast culture supplementation Baker, Murphy, and Irlbeck on digestion control -- Baker, in captive - Response of captive Rocky Mountain bighorn sheep to oral vitamin supplementation -- Graffam, Wild, and Irlbeck. Characterization of the estrous cycle progesterone levels -- Wild and Drew. Additionally, animals research projects: or samples in bighorn sheep using from FWRF were provided elk -- E fecal to the following �73 - USDA National Animal Health - University of Wyoming - University of Wyoming Cornell - Syracuse EDUCATIONAL University University Center -- Cervid tuberculosis Selenium toxicity Pronghorn -- Bighorn in pronghorn maternal care and neonatal nutrition sheep genetics -- Nitrogen ratios in elk CONTRIBUTIONS FWRF provided formal educational instruction for 4 high school classes/special interest groups, 3 university classes, 1 Project Wild workshop, and 2 professional groups. Animals and facilities were used for hands-on training with 1 professional group. Numerous other informal tours were provided individually to visiting professionals. LITERATURE CITED Baker, D. L., AND N. T. Hobbs. 1985. Emergency feeding of mule deer during winter: tests of a supplemental ration. J. Wildl. Manage. 49: 934-942. Dierenfeld, E. S. and D. A. Jessup. 1990. Variation in serum alphatocopherol, retinol, cholesterol, and selenium of free-ranging mule (Odocoileus hemionus). J. Zoo Wildl. Med. 21(4):425. deer Dierenfeld, E. S. and M. G. Traber. 1992. Vitamin E status of exotic animals compared with livestock and domestics. In L. Packer and J. Fuchs, eds. Vitamin E in Health and Disease. Marcel Dekker, New York. National Research Council. 1985. Nutrient Academy Press. Washington, D.C. Requirements Robbins, C. T., and A. N. Moen. 1975. Milk consumption white-tailed deer. J. Wildl. Manage. 39:355-360 •. Silver, H. 1961. Wildl. Manage. Deer milk 25:66-70. compared with substitute of Sheep. National and weight gain of milk for fawns. J. Wild, M. A., M. W. Miller, B. J. Maynard, and D. R. Magnuson. 1992. Animal and pen support facilities for terrestrial wildlife research. Colo. Div. Wildl. Res. Rep. Fed. Aid Proj. W-l53-R5, WPlA Jl, Job Progr. Rep., July 1991-June 1992, Fort Collins. Wild, M. A. 1993. Animal and pen support facilities for terrestrial wildlife research. Colo. Div. Wildl. Res. Rep. Fed. Aid Proj. W-l53-R6, WPIA Jl, Job Progr. Rep., July 1992-June 1993, Fort Collins. Wild, M. A., M. W. Miller, D. L. Baker, N. T. Hobbs, R. B. Gill, and B. J. Maynard. 1994. Comparing growth rates of dam- and hand-raised bighorn sheep, pronghorn, and elk neonates. J. Wi1dl. Manage. 58:340-347. �/ .. Table 1. Summary Species Bighorn of mortalities Elk Cause of Death L72 T82 21 11 Old age changes, Old age changes, E193 o 203 Pronghorn MI88 JE91 BN93 DD93 NK93 Y093 Mule deer White-tailed L92 Q93 Qa91 T92 U92 deer "Euthanatized ~nconfirmed diagnosis at FWRF during FY 1994. Age (yrs) Animal,ID sheep in hoofstock J93 13 5 2 o o o o 1 Peracute, chronic chronic sinusitis· sinusitis' Undetermined Chronic dental disease, weight loss' Intraspecific trauma Chronic neurologic signs; excess' Hemorrhagic diseaseb Hemorrhagic diseasea.b Hemorrhagic diseaseb Hemorrhagic diseaseb 3 1 1 Acute cardiovascular collapse Trauma Chronic poor doer, pleural adhesions" Chronic enteritis, failure to thrive" Failure to thrive" o Intestinal o volvulus �75 30 T: ..-.. 20 C) ..:.:: I- :r: T ~ W ~ 10 oL-~~--~~--~_L--~~~--~~--~~--~~ o 7 14 21 28 35 42 49 56 63 70 77 83 90 97 104 AGE (days) Fig. J:. Growth rates of white-tailed deer hand-raised 1993. Error bars indicate +1 SE of mean observations. at FWRF in 40 30 t- :r: o 20 W ~ 10 o~~~--~~--~-L--~~~--J_~--~~--~~ o 7 14 21 28 35 42 49 56 63 70 T7 83 90 97 104 AG E (days) Fig. 2. Growth rates of mule deer females hand-raised in 1992 (-) and 1993 (--), and males dam-raised in 1993 (-.-) at FWRF. Error bars indicate +1 SE of mean observations. �76 A 30 20 ~ 10 ..... ~--- ~ .... !: .•.... 7-- _ •••••.•.... ...c::: .s> I ~ -- 20 Q ..::.:: l- ::I: C!J W s: 10 o~~--~~--~~--~~~~~~--~~--J_~--~ o 7 14 21 ~8 35 42 49 56 63 70 77 83 90 97· 104 AGE (days) Fig. 3. Growth rates of hand-raised (A) mule deer and (B) pronghorn weaned abruptly (--) and more gradually (-) at FWRF in 1993. Error bars indicate +1 SE of mean observations. �77 300 -1:1 Cl ~ w 200 J, z - I I I I I :::::iE -c l- s -c :::::iE I I I I I I I I I I I 100 CIJ -c -l c, o N o J F M A M J J A \ \ \ \1 s MONTH. Fig. 4. Plasma vitamin E levels of bighorn sheep ewes supplemented with oral vitamin E at a low (-) and a high level (--) at FWRF in 19921993. �78 ADDENDUM A REVISED PROTOCOL FOR MANAGING CHRONIC AT FOOTHILLS WILDLIFE RESEARCH WASTING DISEASE FACILITY Despite a comprehensive program initiated in 1985 to eradicate Chronic Wasting Disease (CWO) from cervids at Foothills Wildlife Research Facility (FWRF), CWO apparently remains endemic in elk at the facility. CWO was diagnosed in a 3-year-old cow elk from the facility in 1989. Two more cases of CWO in elk have been diagnosed since that time, one in 1991 and one in 1992. CWO has not been diagnosed in deer at FWRF since 1985. Based on these observations, guidelines established in 1985 for maintaining a CWO-free facility are largely obsolete. Here, we provide revisions to those guidelines that are directed at managing the disease reduce risk to personnel and research animals at FWRF. to OBJECTIVES 1. Prevent transmission of CWO to deer. 2. Minimize potential for further exposing elk to cwo. 3. Minimize potential for exposing personnel to CWO or other potential pathogens. . 4. Maintain healthy animals for experimental wildlife research. 5. Prevent transmission of CWO to animals outside FWRF. ASSUMPTIONS 1. CWO is an infectious disease of cervids (especially deer and elk); CWO is not widespread in free ranging cervids. 2. Mode of transmission for CWO is not known, and may be direct, via animal/animal contact, or indirect, through contact with excreta (saliva, urine, feces); animate and inanimate objects may serve as fomites (vehicles) in transmitting CWO. 3. Noncervid wildlife and domestic species are potenti.ally inapparent carriers of CWO. Elk and some noncervid research animals (bighorn sheep and pronghorn) have been exposed to· CWO. APPROACH Overview: 1. 2. 3. 4. 5. 6. Animals: 1. 2. 3. Continue to follow general guidelines that minimize contact of freeranging cervids and domestic livestock with captive research animals. Continue .to minimize common use of facilities and materials between cervids and noncervids. Continue to minimize potential for transmission of CWO by contaminated personnel or equipment. House elk separate from deer. Minimize common use of facilities and materials between elk and deer. Continue to educate animal caretakers about CWO (hazards, protocols, and clinical signs exhibited by affected animals). Perform daily animal observations and maintain detailed records of animal health as a portion of the CWO surveillance program. Continue to exclude wild or captive cervids from endemic CWO areas from entering the captive herd. Endemic areas will now include: Larimer County, Colorado, Park and Albany Counties, Wyoming, and Denver Zoo. Raise and maintain deer separately from elk. Orphans, and neonates raised outside FWRF, will occasionally be accepted from areas that are not CWO endemic. �79 Animal Maintenance: Continue all aspects of guidelines concerning use and maintenance of research animals with emphasis on separation of elk and deer. These include: 1. House and maintain cervids separate from noncervids. 2. House and. maintain deer and elk separately and with no sharing of bedding, watering, or feeding areas. 3. Maintain accurate records for all animals. This information includes (but is not limited to): birth date, origin, body weights (monthly), vaccinations, health problems, research projects, travel. Additionally, i. Tag all animals for easy individual identification. ii. Train FWRF personnel to recognize clinical signs of CWO. FWRF personnel will maintain daily animal observation records describing animal status and will report abnormal observations to the facility manager. 4. Weigh and briefly examine every animal at least once monthly. 5. Follow a preventative medicine program that includes routine vaccination, anthelmintic treatment, hoof trimming, nutritional evaluation, and other measures to optimize overall health of research animals. 6. To prevent transmission of CWO from FWRF to facilities where CWO is not endemic, animals from FWRF will be transferred or donated to other facilities only if the following criteria are met: 1) animals are scheduled for a specific research project, 2) the destination is a closed facility (no egress of live animals) or FWRF-source animals are kept in isolation and sanitation measures followed. Recipients will be notified of CWO risks associated with accepting animals from FWRF. Domestic Livestock Maintenance: 1. Cattle will be regarded as noncervids and will be maintained separate from cervids. 2. Domestic sheep or goats will not be maintained in CDOW research facilities. Use of Research Animals Outside FWRF: 1. The loaning of animals to facilities outside FWRF is discouraged. 2. Procedures for isolating cervids at other CDOW facilities will be the same as those at FWRF. 3. RFAC will approve movement of FWRF cervids out of, and back in to, FWRF. 4. Animals maintained at FWRF will not be permanently released into the wild. 5. During field studies, precautions should be taken to minimize chances of exposure to free-ranging cervids and livestock. 6. Researchers are responsible for maintaining accurate records of animals taken outside FWRF. CWO SURVEILLANCE PROGRAM 1. Continue to euthanitize any animal showing clinical signs of CWO and examine tissues grossly and histologically. 2. Continue to perform complete postmortem examinations and histologically examine brain tissue of any animal that dies at FWRF. 3. If CWO is confirmed in a deer at FWRF, isolation procedures as previously described for elk will be implemented for the group of deer exposed to the affected CWO animal. RFAC will determine if further actions will be taken. Facilities: 1. Continue to exclude free-ranging wildlife or livestock from the facility or from contact with captive animals using interior and perimeter fencing. 2. Maintain deer separately from elk with no fenceline contact. 3. Prevent contact between cervids and noncervids. �80 4. S. 6. 7. 8. Minimize runoff from noncervid to all cervid pens, and from elk to deer pens through appropriate pen assignment and drainage control. Feed and handle animals or clean pens using the following traffic pattern: deer, then elk, then noncervids. Clean animal pens (especially feed areas and waterers) weekly. Isolation pens, and other areas where animals are held for extended periods, will be cleaned of organic matter and disinfected with a chlorine solution after use. The researcher last using the area will be responsible for cleanup. Cooperative compliance will be made a condition of all study plans which use FWRF ungulates. Different species may be held concurrently in isolation pens if a buffer zone (empty pen) is used. Equipment: 1. Minimize common use of equipment, especially between deer and all other areas. 2. Disinfect common use equipment prior to use in deer or elk areas. Feed: 1. Hay will not be accepted on cultivated pastures. from areas where domestic sheep have grazed Personnel: 1. Wash hands before and after handling each species of anLmal. 2. All researchers and collaborators and their subordinates will comply with this protocol. 3. Unsupervised access to FWRF will be lLmited to authorized personnel. 4. RFAC will evaluate and amend this program as necessary. �81 Colorado Division Wildlife Research July 1994 of Wildlife Report JOB state of Project Work REPORT Colorado No. W-153-R-7 Plan No. Mammals lA Covered: Author: Personnel: Research Multispecies 3 Job No. Period PROGRESS Mammals Investigations 2 Research Administration July 1, 1993 - June 30, 1994 R. Bruce Gill R. Bruce Gill and Diane K. Hall ABSTRACT Personnel time and budgets were allocated among 6 Mammals 1 Research Projects. Highlights of preliminary results and management implications are summarized. All research objectives were successfully met within the assigned budgets. Four technical manuscripts were accepted publications and/or journals. . . for publication in scientific ��83 MAMMALS 2 RESEARCH ADMINISTRATION R. Bruce Gill P. N. OBJECTIVE Administer research studies within productivity at the lowest cost. the Mammals segment 1. 2 Research Unit for the highest Objectives Assign, supervise; and administer research projects concerning bighorn sheep, pronghorn antelope, black bears, kit foxes, ungulate survival modeling, and biodiversity management modeling. RESULTS Fourteen projects were active during the segment. Segment completed for all 14 projects. Highli9hts include: objectives were • Maintenance and expansion of the Foothills Wildlife Research Facility to accommodate research on 5 species of wild ungulates and 4 species of migratory game birds. Facilities were modified to accommodate cooperative work with the USDA APHIS to test the feasibility of immunocontraceptives to control fertility of white-tailed deer. • Publication or acceptance publications. • Preparation and submission peer review. • Development of a user-friendly software program (NOREMARK) to assist wildlife biologists in estimating the number of marked animals and resightings of marked'animals necessary to reliably estimate populations of free-ranging big game animals. • Wildlife disease outbreaks were monitored by analyzing specimens from samples submitted to the Wildlife Research Laboratory (WRL). A computerized database was developed to record each diagnostic case history submitted to the WRL and hard copies of historic case histories were added to the database. Chronic wasting disease has been diagnosed from 24 mule deer, 1 white-tailed deer, and 4 Rocky Mountain elk submitted for examination since 1981. So far all cases have been confined to Larimer County. No microscopic lesions compatible with bovine tuberculosis have so far been detected in free-ranging populations of mule deer and Rocky Mountain elk. • At least 26 strains of Pasturella haemolytica have been isolated from 8 free-ranging populations of bighorn sheep. These preliminary testing results suggest the combination of genomic RNA fingerprinting and cytotoxicity evaluations may offer a useful approach to detect and predict the severity of pneumonia. outbreaks among free-ranging bighorn populations. • Preliminary results of experiments to increase bighorn sheep lamb production and/or survival with drug treated baits reveal no consistent increases in either production or survival which can be attributed to drug treatment. for publication of 1 technical of 13 technical publication manuscript to �84 • A final report evaluating the use of bighorn sheep habitat suitability models and GIS habitat classification systems to describe and categorize habitats suggests that GIS systems and habitat suitability models in combination are very effective and efficient tools to classify and evaluate bighorn habitats. • A continuing study of the population dynamics of a pioneering pronghorn antelope population in Middle Park Colorado indicate the population continues to exhibit characteristics of density dependent regulation of population growth. Rate of growth continues to decline and density increases. In addition, the population seems to be expanding its area of distribution in response to increasing density. • Evaluat-ion of several tests too-detect density dependence in big game populations suggest the most promising variables to measure are reproduction and survival. Direct measures of density are not sufficiently sensitive to reliably detect density effects. • Comparisons of pronghorn density estimates from quadrat count sampling vs. line transect sampling indicate that line transect sampling significantly underestimates true populations size (ca 30%). • A new black bear live trap was developed and tested. Only 1 injury resulted in 134 live-captures. This compares to an injury rate of 1 injury for every 3 live-captures in a previous study where leg-hold snares were used. • Four new groups of kit foxes have been located in addition to the 1 group located in Peach Valley. Distribution of all groups suggests that kit fox are widely distributed throughout potential habitats in western Colorado, but in widely dispersed, small family groups. Study results were used to modify trapping regulations to safeguard kit foxes from inadvertent accidental capture. • A research proposal was developed and successfully competed for funding with GoColorado funds. The proposal is a computerized decision model to aid land use planners in evaluating the consequences of land use decisions and to__compare trade offs among land use alternatives. • Objectives for all 14 Mammals 2 Research Section projects were completed successfully. Funding requirements were underestimated the black bear census project, but internal reallocations were sufficient to cover the shortfall. Prepared by: R. Bruce Gill Wildlife Research Leader for �85 Colorado Division Wildlife Research July 1994 of Wildlife Report JOB PROGRESS State of Project Work Colorado No. ~W~-~1~5~3~-~R~-~2~ Plan No. __~l=A~ Job No. Period Covered: _ Mammals _ Multispecies Research Investigations Consulting Services for Mark-Recapture Analysis 5 Author: REPORT July 1, 1993 - June 30, 1994 G. C. White Personnel: R. M. Bartmann, R. B. Gill, T. D. I. Beck, D. J. Freddy ABSTRACT Progress towards the objectives of this job include: 1. A user-friendly computer program (Program NOREMARK) was developed to operate on personal computers for the computation of population estimates based on resightings of marked animals. A User's Manual has been written and the package submitted to the Wildlife Society Bulletin. 2. A study of compensatory effects of harvest on the Piceance Basin mule deer population was continued as part of Federal Aid Project W-153-R Work Plan 2 Job 15, entitled Compensatory Effects of Harvest in a Mule Deer Population. Experimental harvests have been conducted in December, 1989, 1990 and 1991. Radio collars to monitor over-winter survival of fawns were placed on the animals during November, 19891993. 3. Consultation has been provided in the design and analysis of a statistical protocol for estimation of the statewide black bear population. 4. Consultation has been provided in the design and analysis of a statistical protocol for procedures to estimate elk survival and sightability from the' air. . 5. Consultation has been provided in the design and analysis of a study of the timing of migration and distribution of elk in response to hunting. ��87 CONSULTING SERVICES FOR MARK-RECAPTURE ANALYSES G. C. White P. N. OBJECTIVES Evaluate compensatory population. effects of harvest SEGMENT 1. on the Piceance Basin mule deer OBJECTIVES Develop a user-friendly program for the estimation of mark-resight population estimators, produce a User's Manual, and submit for publication in the Wildlife Society Bulletin. RESULTS AND DISCUSSION Introduction. Estimation of population size of a geographically and demographically closed but free-ranging population is a common problem encountered by wildlife biologists. The earliest approaches to this problem were developed by Petersen (1896) and later Lincoln (1930), where capturerecapture techniques were applied. Extensions to the simple 2 occasion Lincoln-Petersen estimator were developed for multiple occasions (Schnabel 1938, Darroch 1958), for removal experiments (Zippin 1956, 1958), and for heterogeneity of individual animals (Burnham and overton 1978, 1979, Chao 1988). For the capture-recapture technique, Otis et ale (1978) and White et ale (1982) provided a summary of the available methods, and others (White et ale 1978, Rexstad and Burnham 1991) describe Program CAPTURE for computing these estimators of population size. More technologically advanced approaches to the problem of abundance estimation have incorporated animals marked with radio transmitters. The initial sample of animals is captured and marked with radios,. but recaptures of these animals are obtained by only observing them, not actually recapturing them. The limitation·of this procedure is that unmarked animals are not marked on subsequent occasions. The advantage of this procedure is that resighting occasions are generally much cheaper to acquire than when the animals must be physically captured and handled. The mark-resight procedure has been tested with known populations of mule deer (Odocoileus hemionus) (Bartmann et ale 1987), and used with white-tailed deer (Q. virginianus) (Rice and Harder 1977), mountain sheep (Ovis canadensis canadensis) (Furlowet ale 1981, Neal et ale 1993), black (Ursus americanus) and grizzly (g. arctos) bears (Miller et ale 1987), and coyotes (Canus latrans) (Hein 1992). Arnason et ale (1991) describe a method where the number .of marked animals is not known, whereas the mark-resight estimators described here assume the number of marked animals is known. I herein describe Program NOREMARK, a program to compute mark-resight estimators of population abundance. Four estimators of abundance can be computed, providing models that allow heterogeneity of capture probabilities, and immigration and emigration from a fixed study area (Eberhardt 1990). The first estimator in NOREMARK is the joint hypergeometric maximum likelihood estimator (JHE) (Bartmann et ale 1987, White and Garrott 1990, Neal 1990, Neal et ale 1993). JHE is the value of H which maximizes the joint hypergeometric likelihood for k occasions. The estimate N can be found by iterative numerical methods. Confidence intervals are determined with the profile likelihood method (Hudson 1971, Venzon and Moolgavkar 1988). This estimator assumes that all the marked animals are on the area surveyed for each survey, i.e., that the population is geographically closed. Hence, the number of marked animals (M) is constant for each survey, although the sighting probability is not assumed to be constant for each survey. An extension allowed in Program NOREMARK is to allow additional marked animals to �88 be added to the population between sighting occasions, animals to be marked between sighting occasions. i.e., to allow unmarked The JHE estimator has been extended to accommodate immigration and emigration (Neal et al. 1993) through a binomial process. This estimator is labeled IEJHE, and does not assume that the population is geographically closed. Assume that the total population with any chance of being observed on the study area is H*, and that at the time of the iilisighting survey, ~ animals occur on the study area. I am interested in estimating the mean number of animals on the study area, N, and possibly H·. At the time of the iilisighting occasion, a known number of the marked animals (~) are on the study area of the possible ~ animals with transmitters. The probability that an individual is on the study area on the iilioccasion can be estimated as Hi/~, or in terms of the parameters of interest as ~/H*. The likelihood function for this model that includes, temporary immigration and emigration from the study area is a product of the binomial distribution for the probability that the animal is on the study area times the joint hypergeometric likelihood. The parameters H* and ~ for i=l to k can be estimated by numerical iteration to maximize this likelihood, with the constraints that Hi > (~ + !!j) and H* > Hi for i=l to k. Profile confidence intervals can be obtained for the k+1 parameters. I was not interested in the k population estimates for each sighting occasion, but rather desired the mean of the ~ est,imates. Therefore, I re-parameterized the likelihood to estimate the total population and mean population size on the study area directly, and their profile likelihood confidence intervals. Minta and Mangel (1989) suggested a bootstrap estimator (MM) of population size based on the sighting frequencies of the marked animals, 1i• For unmarked animals, sighting frequencies are drawn at random from the observed sighting frequencies of the marked animals until the total number of sightings equals the number of unmarked animal sightings. The number o~ animals sampled is then an estimate of the number of unmarked animals in the population, so that M plus the number sampled is an estimator for H. Only bootstrap samples where the number of sightings was exactly equal to the number of unmarked animal sights were used, i.e., cases where the cumulative sightings exceeded y. were excluded. Minta and Mangel (1989) accepted the first value where the cumulative sightings equalled or exceeded the number of unmarked animal sightings. The stopping rule I used results in less bias than the rule -used, by Minta and Mangel (1989). Minta and Mangel (1989) suggested the mode of the bootstrap replicates as the population estimate. Confidence intervals were computed as probability intervals with the 2.Siliand 97.Silipercentiles from the bootstrapped sample of estimates. White (1993) demonstrated that the MM estimator is basically unbiased, but that the confidence interval coverage was not the expected 9S% for a=O.OS. A modified procedure was suggested, but coverage still was not satisfactory. Bowden (1993) suggested an estimator for the Minta-Mangel model where the confidence intervals on the estimate were computed based on the variance of resighting frequencies of marked animals. He approached the problem from a sampling framework, where each animal in the population has the attribute ~ of the number of times it was resighted. The values of ~ are known for the marked animals, and the sum of the ~'s are known for the unmarked animals. Bowden presented an unbiased estimator and its variance, and suggested that confidence intervals should be computed using a log transformation. NOREMARK contains a design option to assist the user with determining the number of resighting occasions, proportion of the population to mark, and proportion of the population to resight on each occasion. This design routine is implemented with simulations from the JHE estimator. �89 Simulations of the 4 estimators can also be performed with NOREMARK. Output from the simulations includes expected bias, confidence interval length, and coverage. Program NOREMARK is written to be used interactively, but with options store data once they have been entered and to save results to a file or printer. The interface is user-friendly, providing context-sensitive help any time, and the ability to back up and re-enter or re-check previous entries. to at Program Availability and SYstem Requirements. Copies of the program and related documentation are available on Internet on the anonymous ftp address picea.cnr.colostate.edu. NOREMARK is written in CA-Clipper (user interface) and Microsoft Fortran (numerical optimization procedures), and runs on MSDOS PC computers. The program (executable files and source code) are accompanied by an ~lectronically stored manual and by auxiliary files, including data files containing the mountain sheep observations described by Neal et ale (1993) • The system requirements are minimal: an IBM-pc-compatible with 640k of base memory and approximately 1M of disk space will suffice. Simulation of estimators can be very time consuming, particularly for the immigrationemigration estimator. Users will want a math coprocessor to perform a reasonable number of simulations. Those contemplating many simulations will want to have a high-speed 80486 or Pentium machine. Literature Cited Arnason, A. N., Schwarz, C. J., and Gerrard, J. M. 1991. Estimating closed population size and number of marked animals from sighting data. J. Wildl. Manage. 55:716-730. Bartmann, R. M., G. C. White, L. H. Carpenter, and R. A. Garrott. 1987. Aerial mark-recapture estimates of confined mule deer in pinyon-juniper woodland. J. Wildl. Manage. 51:41-46. Bowden, D. C. 1993. A simple technique for estimating population size. Dept. of Statistics, Colorado State Univ.,· Fort Collins, Colo. 17pp. Burnham, K. P., and W. S. Overton. 1978. Estimation of the size of a closed population when capture probabilities vary among animals. Biometrika 65:625-633. Burnham, K. P., and W. S. Overton. 1979. Robust estimation of population size when capture probabilities vary among animals. Ecology 60:927-936. Chao, A. 1988. Estimating animal abundance with capture frequency data. J. Wildl. Manage. 52: 295-300. Chapman, D. G. 1951. Some properties of the hypergeometric distribution with applications to zoological sample censuses. University of California Publication in Statistics 1:131-160. Eberhardt, L. L. 1990. Using radio-telemetry for mark-recaptur,e studies with edge effects. J. Applied Ecol. 27:259-271. _ Furlow, R. C., M. Haderlie, and R. Van den Berge. 1981. Estimating a bighorn sheep population by mark-recapture. Desert Bighorn Council Transactions 1981:31-33. Hein, E. W. 1992. Evaluations of coyote attractants and a density estimate on the Rocky Mountain Arsenal. M. S. Thesis, Colorado State Univ., Fort Collins. 58pp. Hudson, D. J. 1971. Interval estimation from the likelihood function. J. Royal Stat. Soc. Series B 33:256-262. Leslie, D. 'M., Jr. and C. L. Douglas. 1979. Desert bighorn sheep of the River Mountains, Nevada. Wildl. Monogr. 66:1-56. Leslie, D. M., Jr. and C. L. Douglas. 1986. Modeling demographics of bighorn sheep: current abilities and missing links. North American Wildlife and Natural Resources Conference Transactions 51:62-73. Miller, S. D., E. F. Becker, and W. H. Ballard. 1987. Black and brown bear .density estimates using modified capture-recapture techniques in Alaska. International Conference on Bear Research and Management 7:23-35. �90 Minta, S. and M. Mangel. 1989. A simple population estimate based on simulation for capture-recapture and capture-resight data. Ecology 70: 1738-175l. Neal, A. K. 1990. Evaluation of mark-resight population estimates using simulations and field data from mountain sheep. M. S. Thesis, Colorado State Univ., Fort Collins. 198pp. Neal, A. K., G. C. White, R. B. Gill, D. F. Reed, and J. H. Olterman. 1993. Evaluation of mark-resight model assumptions for estimating mountain sheep numbers. J. Wildl. Manage. 57:436-450. otis, D. L., K. P. Burnham, G. C. White, and D. R. Anderson. 1978. Statistical inference from capture data on closed animal populations. Wild1. Monogr. 62:1-135. Schnabel, Z. E. 1938. estimation of the size of animal populations by marking experiments. U. S~ Fish and Wildlife service Fisheries Bull. 69:191-203. Rexstad, E., and K. Burnham. 1991. Users' guide for interactive program CAPTURE. Colo. Coop. Fish and Wildl. Res. Unit, Colo. State Univ., Fort Collins. 29pp. Rice,. W. R. and J. D. Harder. 1977. Application of multiple aerial sampling to a mark-recapture census of white-tailed deer. J. Wildl. Manage. 41:197-206. Venzon, D. J. and Moolgavkar, S. H. 1988. A method for computing profilelikelihood based confidence intervals. Appl. stat. 37:87-94. White, G. C. 1993. Evaluation of radio tagging marking and sighting estimators of population size using Monte Carlo simulations. Pages 91-103 in J. D. Lebreton and P. M. North, eds. Marked Individuals in the study of Bird Population, Birkhauser Verlag, Basel, Switzerland. User's White, G. C., K. P. Burnham; D. L. Otis, and D. R. Anderson. 1978. manual for program CAPTURE. Utah state Univ. Press, Logan, UT. 40pp. White, G. C., D. R. Anderson, K. P. Burnham, and D. L. Otis. 1982. CaptureLos Alamos recapture and removal methods for sampling closed populations. National Laboratory. LA-8787-NERP. Los Alamos, N.M. 235pp. White, G. C. and R. A. Garrott. 1990. Analysis of wildlife radio-tracking data. Academic Press, New York. 383pp. Zippin, C. 1956. An evaluation of the removal method of estimating animal populations. Biometrics 12 :163-169. Zippin, C. 1958. The removal method of population estimation. J. Wildl. Manage. 22:82-90. �Colorado Division Wildlife Research July 1994 of Wildlife Report JOB PROGRESS state of Project Work Colorado No. W-153-R-7 Plan No. Job No. Period REPORT Covered: Mammals Research 1A Multispecies Investigations 6 Monitoring and Managing in Colorado Wildlife Health July 1, 1993 - June 30, 1994 Authors: M.W. Miller, C.W. McCarty, D.M. Getzy, C.A. Mehaffy, E.S. Williams, W. J. Adrian, and T. R. Spraker M.L. Stevens, Personnel: J. Bredehoft, G. Byrne, A. Case, D. Clarkson, M. Cousins, B. Davies, H. Dietrick, L. Evans, R. Forde, D. Freddy, T. Fulk, K. Green, J. Jackson, K. Kinney, M. Lamb, C. Leonard, K. Madriaga, C. W. McCarty, B. Olmstead, J. Ritchie, K. Scott, H. Spear, and R. Spowart Abstract Wildlife populations throughout Colorado were monitored for occurrence of disease using a combination of extensive and intensive approaches. We continued to develop and modify a statewide surveillance program for acquiring, examining, reporting on, and summarizing sporadic wildlife disease cases occurring throughout Colorado. At least 65 carcasses and/or tissue samples representing 53 wildlife cases were submitted for diagnostic examination during July 1992-June 1993. Malnutrition/starvation (9/22), locoism (5/22), or chronic wasting disease (3/22) were the most common diagnoses in wild cervids submitted; all 7 bighorn sheep submitted showed gross and/or histologic lesions of bronchopneumonia. Among carnivore cases, canine distemper (3/4 raccoons submitted) and feline panleukopenia (1/2 bobcats and 1 mountain lion) were diagnosed. Salmonellosis was diagnosed in passerine birds from 7- locations throughout Colorado. Other mammalian and avian cases appeared to represent isolated incidents of trauma. For 8 cases, cause of death could not be determined. We continued the annual statewide survey of deer and elk hunters to collect sera for brucellosis screening, and also continued modifying and evaluating our survey program. Of 10,000 elkr and 500 deer hunters surveyed, 1,375 (14%) returned blood samples for brucellosis screening from animals harvested throughout Colorado during October 1993-January 1994. Of samples returned, 723 (53%) were usable; marked hemolysis and/or contamination precluded evaluation of the remaining samples. All elk and deer sera tested were negative for antibodies to Brucella spp. on the standard card test. Overall, about 13% of the survey kits distributed to deer or elk hunters in 1993-1994 provided usable samples, as compared to 7 % in 1992-1993 and 5% in 1991-1992. Nine cases of chronic wasting disease (CWO), a spongiform encephalopathy, were confirmed in free-ranging deer and elk in Larimer county during FY 1993-1994. Twenty-nine free-ranging CWO cases have been confirmed in Colorado since 1981; to date, all of these cases have been from Larimer County (GMUs 9, 191, 19, or 20). We prepared a manuscript summarizing these cases. To obtain reliable estimates for distribution and prevalence of CWO in wild cervids, we continued to survey for CWO in select deer and elk populations throughout Colorado. L__ �92 Brains from about 100 mule deer and 49 elk harvested in GMUs 19 and 20 (DAUs D4, D10/E4, E9), and from about 57 mule deer and 29 elk harvested in GMUs 66 and 67 (GMU D25/E25) were collected for examination for cwo. Histologic evaluation of samples has not been completed, but all brains from hunterkilled deer and elk examined to date have been negative for spongiform encephalopathy. Retropharyngeal and other cranial lymph nodes and tonsils from about 57 mule deer and 29 elk harvested in GMUs 66 and 67 (GMU D25/E25), and from about 100 mule deer and 49 elk harvested in GMUs 19 and 20 (DAUa D4 ,D10/E4, E9) were examined for gross lesions of bovine tuberculosis and collected for histologic evaluation and culture. We occasionally observed tonsillar and/or retropharyngeal cysts, abscesses, and/or granulomas in samples from both areas (36 cases total). Histologic evaluations have not been completed, but no microscopic lesions compatible with bovine tuberculosis have been observed in samples examined to date. We continued developing a generalized, stochastic, individual-based simulation model of infectious disease in wild ungulate populations. We incorporated parameters to simulate introduction of bovine tuberculosis into a wild elk population and conducted sensitivity analyses of 50-year simulations (n = 500) where an infected elk was introduced into a population of 200 wild elk under assumptions of varying conservancy. As with previous analyses, transmission coefficient (tc) assumptions markedly influenced outcomes: assuming tc = 0.3 new infections/infected individual/year, the probability that tuberculosis became established in simulated populations ranged from 0 to 0.4, depending on survivorship of infectious indivisuals and likelihood of cow-calf transmission. Our preliminary results suggest introduction of bovine tuberculosis into wild elk populations could represent a significant obstacle to national eradication goals. �93 MONITORING AND MANAGING WILDLIFE HEALTH IN COLORADO M.W. Miller, C.W. McCarty, D.M. Getzy, ,C.A. Mehaffy, M.L. stevens, E.S. Williams, W.J. Adrian, and, T.R. Spraker P. N. OBJECTIVES Develop and implement a program for enhancing s'tatewide efforts manage health of Colorado's terrestrial wildlife populations. AGREEMENT to monitor OBJECTIVES 1. Modify and improve systems for submitting, diagnosing and reporting sporadic disease cases in wild animals throughout Colorado. 2. Develop and use databases for assimilating and analyzing problems identified through surveillance and surveys. and managing wildlife on data on disease 3. Design, conduct, and report results of surveys for brucellosis, tuberculosis, and chronic wasting disease in specific deer and/or populations. 4. Provide assistance in investigating outbreaks in Colorado. and elk disease 5. Design experiments to develop and/or improve techniques for investigating wildlife diseases; begin conducting approved and funded research. Maintaining healthy wildlife populations is a fundamental component of sound wildlife management practices., Habitat degradation, high animal density, extreme weather, and disease can act singly or in combination to compromise the overall health of a wildlife population. As Colorado's wildlife managers, we have developed a variety of tools for monitoring and assessing the effects of habitat loss, animal numbers, and weather on wildlife populations. We have also invested considerably in developing tools to manage these factors to optimize performance of the wildlife populations in our stewardship. In contrast, monitoring and managing the effects of disease on wildlife population performance have received relatively little attention (with a few notable exceptions). This lack of attention may, be rooted to some extent in a widely-held belief that wildlife diseases are symptoms of larger underlying population problems that will be resolved if those larger problems are managed properly. ' Despite this belief, disease can be a significant obstacle to effective and efficient wildlife management in Colorado. Disease outbreaks account for substantial mortality in some wildlife populations. Introduced pathogens have potential to decimate local wildlife populations. Some diseases depress wildlife population performance to levels below resource-based carrying capacity •. Many wildlife diseases are shared with domestic animals and/or humans, and in some cases wildlife populations serve as reservoirs for these agents. Disease also detracts from the aesthetic value of wild animals, and may convey a perception of mismanagement to uninformed publics. For these reasons, diseases should be regarded as an integral part of wildlife population dynamics and wildlife management. Select wildlife health problems have been monitored in Colorado for more than 30 years. These longstanding efforts have provided useful information on the diseases studied. However, because these efforts have not always been �94 coordinated on a statewide basis, and because some findings have not been widely available to managers and policy makers, applications to overall management programs have been limited. In order to improve our collective ability to manage wildlife health in Colorado, we need a mor~ coordinated and systematic approach for monitoring, investigating, and reporting on health problems in free-ranging wildlife. A more complete understanding of wildlife diseases and their effects on population performance is fundamental to comprehensive wildlife management. Enhanced surveillance efforts will provide a mechanism for detecting health problems throughout the state before serious impacts to wildlife populations occur. Assimilating diagnostic data will aid in assessing trends suggestive of population-level disease problems. Programs for conducting extensive and intensive surveys for potential and realized wildlife diseases will provide reliable prevalence and distribution data for managers and administrators to use in decision making. Expertise in investigating and managing epizootics and epornitics will ameliorate efficacy and efficiency of efforts to control outbreaks. Improved techniques for diagnosing and studying wildlife diseases will provide a firm foundation for health management programs designed to enhance the quality.of Colorado's wildlife populations. MATERIALS Disease AND METHODS Surveillance We monitored wildlife populations throughout Colorado for occurrence of disease using a combination of extensive and intensive approaches. These were organized and conducted as follows: Statewide Surveillance We continued to develop and modify a program for acquiring, examining, reporting on, and summarizing sporadic wildlife disease cases occurring throughout Colorado. All carcass submissions were subjected to necropsy. Ancillary diagnostics, including histopathology, bacteriology, virology, serology, parasitology, and toxicology were performed at the discretion of CDOW personnel and/or the attending pathologist. Preliminary examination and/or test results were telephoned to CDOW's Wildlife Research Center Laboratory, usually within 3-5 days of completion, and a final report were usually provided within 15 business days of submission. Pertinent data from preliminary and final reports were entered into a permanent database (described below), and copies of reports were filed as well as sent to appropriate field personnel. Pertinent data, including species, age, sex, location, number affected, diagnosis, and other information (as available) were entered ~nto a computerized database. This database was used to generate quarterly and·~nual wildlife morbidity and mortality reports. In addition, data are avaLf.ab Le for anaLyed.s of long-term trends in select wil,dlife disease problems. . Surveys Brucellosis Survev: We continued the statewide survey of deer and elk hunters to collect sera for brucellosis screening. Over the next several years, however, we plan to continue developing and implementing strategies for expanding utility and improving efficacy and efficiency of this survey. In particular, we will focus on improving return rates on sampling kits, quality of samples returned, and ability to target specific areas or populations for surveillance. We continued examining performance of the existing survey to determine average return rates and sample usability by species and season. In addition to the modifying the statewide survey, we also continued developing a process for intensively sampling specific geographic areas or populations using modified �95 hunter surveys focused on sampling in select DAUs and/or GMUs. These surveys were constructed such that the probability of failure to detect at least 1 case of brucellosis in the selected population was ~0.1 even if herd prevalence is 1%. We will compare return rates and sample usability among seasons and collection methods, and use these comparisons to guide future survey efforts. Data from this year's survey will be analyzed in combination with those from FY 1991/1992 and FY 1993/1994 to compare sampling strategies. Results of survey modifications will be reported in future annual Job Progress Reports. We mailed about 11,205 blood sampling kits to deer hunters and 4,865 kits to elk hunters in selected GMUs to gather samples for CDOW's annual brucellosis surveillance program conducted in cooperation with the Colorado Department of Agriculture's State/Federal Brucellosis Laboratory in Denver. Kits went to sportsmen with antlerless or either-sex deer permits for early, second, regular, and late seasons in mountain and plains GMUs statewide. In addition, second and third regular and late season hunters with antlerless elk permits in DAU E2, E25, Ell, E27, E28, and E33 received kits as part of an intensive sampling effort in the vicinity of a recent bovine brucellosis and tuberculosis outbreaks. Returned samples were identified by GMU of harvest. Usable samples were centrifuged, and sera were tested for antibodies to BruceLLa spp. using a standard card test. Unused sera were banked and stored at -20 C for future use. Chronic Wasting Disease Survey: ·To obtain reliable estimates for distribution and prevalence of cwo in wild cervids, we continued to survey for CWO in select deer and elk populations throughout Colorado. Brains from mule deer and elk harvested in various seasons during October 1992-January 1993 in GMUs 19 and 20 (DAUs D4 ,D10/E4, E9), in GMUs 66 and 67 (GMU D25/E25), and on the Forbes Trinchera Ranch near Ft. Garland (DAU D31/E33) were collected for examination for CWO. Brains from hunter harvested mule deer and elk were collected, usually within 12 hours of death, and fixed in 10% buffered formalin contained in 4 L plastic bags for at least 3 months. Sections of medulla at the obex and frontal portion of the brain including basal ganglia, olfactory cortex and tract, and some frontal cortex were processed routinely for paraffin embedment. Histologic ,sections were cut at 5-6 JJm, stained with hematoxylin and eosin, and examined under a light microscope. In addition to formal surveys, we continued to encourage increased surveillance efforts by field personnel statewide and submission of carcasses from deer or elk showing clinical signs resembling cwo. Bovine Tuberculosis Survey: Bovine tuberculosis was diagnosed in captive elk held on a game ranch near Powderhorn, CO in June 1991. Since that time, we have continued to investigate the possibility that tuberculosis might have spread to free-ranging wildlife outside the infected premises. Retropharyngeal and other cranial lymph nodes and tonsils from mule deer and elk harvested in GMUs 66 and 67 (GMU D25/E25), on the Forbes Trinchera Ranch (DAU D31/E33), and in GMUs 19 and 20 (DAUs D4, D10/E4, E9) were examined for gross lesions of bovine tuberculosis and collected for histologic evaluation and culture. Subsamples of parotid, mandibular, and retropharyngeal lymph nodes and tonsils, as available, were preserved in 10% buffered formalin and frozen and submitted to the Wyoming State VeterinarY Laboratory in Laramie for histologic examination (and culturing, when warranted). When possible, eviscerated carcasses were also examined for gross lesions suggestive of tuberculosis • .Disease Investigations No significant disease outbreaks were investigated during July 1992-June 1993. �96 Experimental Approaches We began developing a generalized, stochastic, individual-based simulation model of infectious disease in wild ungulate populations (Fig. 1). We plan to use this model in predicting consequences of disease introductions, improving understanding of the epizootiology of select disease problems, and evaluating potential disease management strategies. In this model, populations display density-dependent sigmoid growth in the absence of disease or other limiting processes. We employed a novel mathematical approach for estimating pathogen transmission within simulated populations, and assumed transmission probabilities are a function of prevalence. Initially, we incorporated parameters to simulate introduction of bovine tuberculosis into a wild elk population and examined probable consequences of such introductions. As a preliminary step, we examined results of replicated 50-year simulations (n = SOO/parameter set) where 2 infected elk were introduced into a population of 500 wild elk. Our model incorporated population parameters estimated from a lightly hunted elk population (Forbes Trinchera Ranch). We assumed a constant cow-calf transmission rate (0.95) and a 2-year inCUbation period before newly infected animals became infectious. We then made replicated simulations, varying transmission coefficient (tc = 0.3 or 0.5 new infections/infected individual/year) to assess the influence of transmission on potential outcome of tuberculosis introductions. RESULTS Disease AND DISCUSSION Surveillance statewide Surveillance At least 40 carcasses and/or tissue samples representing 30 wildlife cases were submitted for diagnostic examination during July 1992-June 1993 (Table 1). Malnutrition/starvation (9/22), locoism (5/22), or chronic wasting disease (3/22) were the most common diagnoses in wild cervids submitted; all 4 bighorn sheep submitted showed gross and/or histologic lesions of bronchopneumonia. Among carnivore cases, canine distemper (3/4 raccoons submitted) and feline panleukopenia (1/2 bobcats and 1 mountain lion) were diagnosed. Hepatitis was diagnosed in a great blue heron. other mammalian and avian cases appeared to represent isolated incidents of trauma. For 8 cases, cause of death could not be determined. We will continue adding new accessions throughout the coming fiscal year to our computerized database for diagnostic case information, as well as data from archived reports as they become available. Surveys Brucellosis Survey: Of 11,205 deer and 4,865 elk hunters surveyed, 2,375 (15%) returned blood samples for brucellosis screening. Of samples returned, only 1,228 (48%) were usable; marked hemolysis and/or contamination precluded evaluation of the remaining samples. All 795 deer and 352 elk sera tested were negative for antibodies to Brucella spp. on the standard card test. overall, about 7% of the survey kits distributed to deer or elk hunters provided samples usable in this year's brucellosis survey, as compared to 5% in 1991-1992. Increases in both sample return rates (15% in 1992-1993, up from 12% in 1991-1992) and proportions of returned samples that were usable (48% in 1992-1993, up from .42% in 1991-1992) appeared to contribute to improvement in the overall effectiveness of the 1992-1993 survey. These data, combined with those collected during FY 1991/1992 and in FY 1993/1994, will be used in assessing strategies for improving the efficiency of statewide serologic surveys that depend on blood samples submitted from harvested animals. �97 Chronic Wasting Disease Survey: Nine cases of chronic wasting disease (CWO), a spongiform encephalopathy, were confirmed in free-ranging deer and elk in Larimer County during FY 1993-1994. Twenty-nine free-ranging Cwo cases have been confirmed in Colorado since 1981 (Table 2); to date, all of these cases have been from Larimer County (GMUs 9, 191, 19, or 20) (Fig. 1). We prepared a manuscript summarizing these'cases; the abstract of that manuscript follows: CHRONIC WASTING DISEASE IN FREE-RANGING DEER AND ELK IN LARIMER COUNTY, COLORADO, 1981-1994: CLINICAL, PATHOLOGICAL, AND EPIZOOTIOLOGICAL OBSERVATIONS. T.R. sprakexi, M.W. Mille?, E.S. Williams3, D.M. Getzy, W.J. Adrian2, 2 Gene G. Schoonveld , and R.A. spowart2• 2 3 Colorado State University Diagnostic Laboratory, Fort Collins, Colorado 80523, .USA. Colorado Division of Wildlife, 317 West Prospect Road, Fort Collins, Colorado 80526, USA. Department of Veterinary Sciences, University of Wyoming, 1174 Snowy Range Road, Laramie, Wyoming 82070, USA. Abstract: Between March 1981 and June 1994, we diagnosed 29 cases of chronic wasting disease in 24 mule deer (Odocoileus hemionus), four Rocky Mountain elk (Cervus elaphus nelsoni), and a white-tailed deer (0. virginianus) submitted from throughout Larimer County, Colorado. Emaciation accompanied by excessive salivation, behavioral changes, ataxia, or weakness were reported in 12 affected animals observed prior to death. Severe emaciation was the only consistent gross finding; bronchopneumonia and/or other intercurrent diseases were also observed in 14 (48%) of 29 cases. Spongiform encephalopathy, characterized by microcavitation of grey and/or white matter, intraneuronal vacuolation, and neuronal degener~tion, was diagnosed microscopically in all cases. Although histologic lesions and their distributions were indistinguishable from those in captive cervids, clinical and gross findings suggested the course of disease was more acute in free-ranging deer and elk than in captive conspecifics. Most affected animals were young adults; estimated ages ranged from 2.5 to 7.5 yrs (median = 3.5 yrs) for mule deer and from 1.5 to 10.5 yrs for elk. Mule deer appeared to be the primary species affected, and accounted for 83% of all cases. Moreover, males were disproportionately overrepresented among mule deer cases: bucks comprised only about 19% of adult mule deer populations in Larimer County in recent years, but 15 (63%) of 29 affected mule deer were male (P = 0.00015). Some seasonality of case submissions was also apparent: 26 of 29 cases were submitted during October-April. Although cases originated from throughout the county, mule deer submissions were clustered near two population centers (Fort Collins, Estes Park). Twenty-six of these 29 cases were submitted since 1990; this pattern may be a product of intensified detection efforts and/or increasing prevalence. Prevalence estimates, host range, distribution, origins, and management implications of chronic wasting disease in wild deer and elk remain undetermined, and warrant further investigation. To obtain reliable estimates for distribution and prevalence of CWO in wild ce=vids, we continued to survey for CWO in select deer and elk populations throughout Colorado. Brains from about 25 mule deer and 29 elk harvested in GMUs 19 and 20 (DAUs D4 ,D10/E4, E9), from about 67 mule deer and 32 elk harvested on the Forbes Trinchera Ranch near Ft. Garland (DAU D31/E33), and from about 80 mule deer and elk harvested in GMUs 66 and 67 (GMU D25/E25) were collected for examination for CWO. Histologic evaluation of samples has not been completed, but all brains from hunter-killed deer and elk examined to date have been negative for spongiform encephalopathy. Bovine Tuberculosis Survey: Retropharyngeal and other cranial lymph nodes and tonsils from about 80 mule deer and elk harvested in GMUs 66 and 67 (GMU D25/E25), from about 67 mule deer and 32 elk harvested on the Forbes Trinchera �98 Ranch (DAU D31/E33), and from about 25 mule deer and 29 elk harvested in GMUs 19 and 20 (DAUs D4, D10/E4, E9) were examined for gross lesions of bovine tuberculosis and collected for histologic evaluation. We occasionally observed tonsillar and/or retropharyngeal cysts, abscesses, and/or granulomas in samples from all 3 areas (17 cases total). Histologic evaluations have not been completed, but no microscopic lesions compatible with bovine tuberculosis have been observed in samples examined to date. We plan to further develop and refine ongoing surveillance programs for both tuberculosis and CWO during FY 1993-1994. . Disease Investigations No significant Experimental disease outbreaks were investigated during July 1992-June 1993. Approaches In .examining preliminary results of 500 50-year simulations where 2 infected elk were introduced into a population of 500 wild elk, transmission coefficient (tc) assumptions markedly influenced outcomes. Under conservative assumptions (tc = 0.3 new infections/infected individual/year), the probability that tuberculosis became established (i.e., infection still present 50 years after initial introduction) in simulated populations was about 0.2 (Fig. 2), and prevalence in infected populations averaged about 0.03 (Fig. 3). Using a slightly higher tc (0.5 new infections/infected individual/year), the probability that tuberculosis became established increased to about 0.6 (Fig. 2), and mean prevalence in infected populations reached about 0.7 (Fig. 3). Our preliminary results suggest introduction of bovine tuberculosis into wild elk populations could represent a significant obstacle to national eradication goals. We plan to further refine parameter estimates for elk-tuberculosis simulations, and to explore application of this modeling approach to other real and potential disease problems affecting wild ungulate populations. Acknowledgments The statewide wildlife health monitoring and surveillance program described above relies heavily on efforts of dedicated field personnel throughout the Colorado Division of Wildlife, and truly represents a division-wide effort to improve our understanding and management of wildlife disease problems. In addition to those specifically listed, we collectively thank all of those regional and area biologists, district and area wildlife managers, and others who assisted by submitting diagnostic cases throughout the year. In particular, we thank personnel from areas 2, 4, 10, and 16, and from the Forbes Trinchera Ranch for assistance and logistical support in tuberculosis and CWO surveys and surveillance activities, and personnel from the StateFederal Cooperative Brucellosis Laboratory for their continued cooperation, assistance and logistical support in conducting annual brucellosis surveys. �99 Table 1. Summary of wildlife diagnostic during July 1992 - June 1993. REGION SPECIES HE Mule Deer M Mule Deer F Mule Deer M Mule Deer M Mule Deer M Mule Deer Elk M Bobcat Bobcat Raccoon Raccoon F Red Fox Heron Pelican (2) F Blackbirds SE SEX AGE A A I I I A A I A Y cases submitted for diagnosis CAUSE OF DEATH Undetermined CWO Undetermined CWO CWO Malnutrition Malnutrition Trauma/Starvation Feline Panleukopenia septicemia Gunshot-induced Canine Distemper Undetermined Hepatitis Gunshot-induced trauma Undetermined RM BHS RM BHS Raccoon Raccoon M M A Pneumonia Pneumonia Canine Distemper Canine Distemper CE RM BHS Blackbirds M y Pneumonia Undetermined HW Elk RM BHS M A Locoweed Toxicity Pneumonia SW Elk (7 ) Hawk M/F I/A Unknown Elk Elk Elk Elk (4 ) Mt. Lion F M M M A A I A I Starvation Undetermined Septicemia Undetermined Undetermined Locoweed Toxicity Feline Panleukopenia �Table 2. Twenty-nine cases of chronic wasting disease have been documented in free-ranging deer and elk in Larimer County, CO between March 1981 and June 1994. 6/31/94 REVISED [§JI I C' II CASE I III SPP DATE LOCATION " MDF020-062694 WSVL MD 6/26/94 MDM020060694 CSUDL II MDM020042394 6/6/94 II MDF009032494 II MDM020020894 WTDM020- CSUDL MD 4/23/94 934-23602 WSVL WSVL MD ESTES PARK 3/24/94 " EFOI9- WSVl "II 112893 W_ V EFOI9- CSUDl RAWHIDE MD 2/8/94 MD II MDM020- elK 11/28/93 S II ELK 11/27/93 MD 11/16/93 II (SITEI) ESW (RS) WSVL 032893 W2331 MD MD· 4/8/93 4/4/93 II MDFOI9021993 CSUDl MD 3/28/93 II F 3_6 E E,ICT, SE PLAC 4_6+ D E, AP, SE T(LK) 6_6+ O(OBS) E, P SE F 10_6 D E SE CK F 7_6 E E, P SE M 3_6 E E SE M 4 0 E SE M 4 E E, P SE F 7.6 0 (P) (P) M 3.6+ 0 E, Abo SE M 3.6+ D nd SE BO) ESTES PARK ESTES PARK lOVELAND GLADE ROAD (DC, BO) MD 2/19/93 923-16323 HORSETOOTH SO & INLET BAYS (LE) DMG P SE STANLEY CIRCLE (RS) ESW a CK (VAN ORSDEL; WSVL MDF020- E M MASONVILLE W9660 WSVl 0 SE on REDSTON E CK ESW R BUCKHORN 040893 W2837 4+ STOVE PRAIRIE JX (BO) MDM020- 040493 M NW of STOVE PRAIRIE JX (BO) 934-11462 MDM020- SE 01 34 @ BT RIV BUCKHORN NWof DMG II E (GS, BO) 934-12113 CSUDL 0 M LOVELAND W, S ESW 111693 T ESTES PARK 12/8/93 934-12926 DMG U 6+ (WEST @ ELKHORN) (RS) TRS 112893 M PRPA FACILITY (ML) 94Wl037 CSUDL E (STANLEY CIRCLE) (RS) 94W2606 120793 W SE 4 (RS) ESW X E, REN F PINE KNOLL DR ESW Y HISTO HEIGHTSI ESTES PARK 934-24772 TRS Z GROSS (RS) MD TRS A' DlEDI EUTH ESTES PARK (STANLEY II AGE PAlli ESW B' SEX (SUBMITTER) 10/30/92 MDM020- WSVl 103092 W2838 (SITEI) ESW (RS) MD ESTES PARK II f-' 0 0 �I 0 II JMOI9101992 CSUDL MD 10/19/92 923-8118 TRS N II MOFI91- CSUOL 101392 923-7743 II MDF020- CSIJOL 031392 MD 10/13/92 II MOMOI9022692 MD 3/13/92 II MOM020121791 2/28/92 II MOM009110891 MD 12/17/82 CSUOL 901-18888 MD 11/8/91 II MOMI91- CSUDL 011891 901-11667 MD 4/8/91 II EM020011191 CSUDL II MOFOI912 E II -90 1/18/91 MOMOI9080290 ELK 1/11/91 II MOM020020390 MD 12/_/90 II MD 8/2/90 Wl028 CSUDL 110988 889-16373 MD 2/3/90 A II II FT. COLLINS/LOVELAND N of RAWHIDE ESTES PARK LIVERMORE US36@ 8ELLVUE BELLVUE ESTES PARK MD l1m88 HORSETOOTH @ DIXON M 3.6 0 E SE M 3.6+ E (h) E SE F 3-4 0 E, Abo SE M 2,6 D E, Abo SE M 4-6 E E, P SE F 3.6 D E SE M 4 0 E, P SE F 3.6 0 E SE M 3.6+ E, SE E (h) SI (fb) (HS) RO 23 S of 60E TRS (HS) TRS SE LAKE 18411 6344 E, P (RS) CSUOL CSUDL 0 MTN 021486 EF020- 3.6+ 17 MDF020- 032081 M & VINE PROSPECT TRS B FT, COLLINS CO OW HATCHERY (LE) 890-11723 MOMOI9- SE (HS) TRS C E, P COOW HATCHERY TRS CSUOL 0 (RS) ESW 0 3,6 BIG ELK MEADOWS 901-9922 WSVL F @JACKSON'S (JJ) 901-11111 CSUOL SE (RS, GS) MD TRS F E (SITE7) TRS G 0 (SOAPSTONE RANCH?) (GS) TRS H 7.6 (GS, BO) 912-8930 040891 ESTES PARK Wof DMG MOF020- F (GSI 912-11469 CSUDL SE LEI OVERLAND TRS J E, Abo (injuryl (RS, GSI MD 912-16979 CSUDL E (SITE7) OMG K PRAIRIE DIVIDE (KENNEY; 912-17171 CSUOL 3.6+ N of 82E, NE of MINE TRS L M PARK (HSI DMG M .sr CANYON STRATTON MD ELK 2/14/86 3/20/81 BELLVUE RMNP (HORSESHOE F 3.6 E E SE F 1.6 E E, P (SE) PARK 7) (G8; NPS7) t-' 0 t-' �102 LarimerCounty Colorado Figure 1. Twenty-three cases of chronic wasting disease (CWO) have been confirmed in free-ranging deer and elk in Larimer County, CO, from March 1981June 1994. Although cases have occurred throughout the county, apparent foci of infection have emerged near Estes Park, Bellvue, and along Buckhorn and stove Prairie Creeks. �lUj Colorado Division Wildlife Research July 1994 of Wildlife Report JOB PROGRESS State of Project REPORT Colorado No. W-153-R-7 Mammals Work Plan No. 2A Mountain Job No. 4 Strategies for Managing Infectious Disease in Mountain Sheep Populations Period Covered: Research Sheep Investigations July 1, 1993 - June 30, 1994 Authors: M. W. Miller, B. J. Kraabel, and K. P. Snipes Personnel: N. R. Canon, R. W. Kasten, A. Boeger-Fields, Silflow, T. R. Anderson, and T. R. Spraker K. W. Mills, R. ABSTRACT We used ribosomal RNA fingerprinting and in vitro measures of cytotoxin production to compare Pasteurella haemolytica isolates from eight indigenous Rocky Mountain bighorn sheep herds (Almont/Taylor River, Avalanche Creek, Chalk Creek, Cottonwood Creek, Grant, Tarryall Mountains, Texas Creek, Waterton Canyon) in Colorado. Using ribosomal RNA gene restriction patterns, at least 26 distinct strains of P. haemolytica were identified among isolates (n = 59) from these herds; we identified one to seven distinguishable ribotypes within individual herds. Of the 26 ribotypes identified, 21 appeared unique to individual herds, four others (E, N, T, BB) were shared by only two herds, and one (A) was common to 3 herds. In vitro evaluation of cytotoxin production by genotypically-distinct P. haemolytica isolates revealed further differences among strains: 4 ribotypes (AA, B, E, 0) showed marked cytotoxin production -- bighorn neutrophil death rate @ 150 ~g culture supernatant was 4-9 times that of an Enterobacter sp. control (7.1±0.4% neutrophil death @ 150 ~g supernatant); cytotoxicity of the other 22 strains examined approximated control levels. All three indigenous bighorn herds that yielded markedly cytotoxic P. haemolytica strains have recent histories of pneumonia epizootics: pasteurellosis outbreaks occurred in the Taylor River herd in 1979 and again in 1991, in the Waterton Canyon herd in 1980, and in the Chalk Creek herd in 1981. One of these strains (E) was also recovered from dead bighorns during recent epizootics in the Alamosa Canyon (1989) and Rock Creek (1990) herds -- these latter herds can be linked to Taylor River by bighorn translocation activities during the last decade. These findings support hypotheses that strains of P. haemolytica carried by healthy bighorn sheep may vary within and among wild populations. Our preliminary results also suggest the combination of genomic fingerprinting and cytotoxicity determination may offer a useful approach for studying the epizootiology of pasteurellosis within and among bighorn herds and may provide insights into strategies for effectively preventing or managing pneumonia epizootics. ��105 EXPERIMENTS TO IDENTIFY AND MANAGE STRESS IN MOUNTAIN SHEEP POPULATIONS M. W. Miller, B. J. Kraabel, and K. P. Snipes P. N. OBJECTIVE To develop population strategies for managing performance. infectious SEGMENT 1. diseases affecting bighorn sheep OBJECTIVES Compare rates for tonsillar carriage and nasal shedding of Pasteurella spp. among free-ranging bighorn populations; compare serum antibody titers to Pasteurella spp. among populations; use phenotypic and genotypic characterizations to compare Pasteurella spp. isolates from different bighorn populations. MANAGEMENT OF BACTERIAL AND VIRAL DISEASES IN MOUNTAIN SHEEP POPULATIONS Inability to control infectious disease outbreaks and subsequent mortality in mountain sheep populations represents a significant obstacle to long-term success in their management. Although the "bighorn pneumonia complex" has been studied intensively for over 3 decades, little is known about many aspects of its etiology and epizootiology. Moreover, management interventions recommended for preventing or controlling this problem remain untested. Most previous efforts to improve understanding and management of the epizootiology of pneumonia in bighorns involved post hoc investigations of dieoffs occurring in free-ranging sheep herds. These studies identified various etiological agents associated with known mortalities and attempted to determine predisposing causes and population consequences of individual outbreaks. From. these investigations, comparisons of real or perceived patterns became the basis for hypotheses ori the epizootiology of pneumonia in bighorns. Recognition of similar patterns in other outbreaks served as evidence supporting these as unifying hypotheses. Unfortunately, several of these. hypotheses have failed to withstand rigorous experimental testing. And, despite our best management efforts, bighorns continue to die. Our strategy for developing a better understanding of the epizootiology and management of bacterial and viral diseases in bighorn populations differs -generally, we propose to take an adaptive environmental assessment approach for studying the bighorn pneumonia complex. As a foundation for our research strategy, we have assimilated existing knowledge on bighorn population dynamics (including the epizootiology and consequences of infectious disease) into a computer simulation model (Hobbs and Miller 1991). Because pasteurellosis appears to underlie virtually all respiratory disease problems reported for bighorns, our modeling efforts have focused on the epizootiology of pasteurellosis in sheep populations. We have constructed a model that reflects dynamics of bighorn populations seen in nature using the simplest assumptions necessary to reproduce those behaviors. We plan to conduct simulation experiments to identify variables that might be particularly sensitive to. management perturbations in altering the dynamics of disease in bighorn populations. Those results will serve as the basis for designing management level experiment~ in the future. In parallel experiments interpreting driving our identifying with our modeling efforts, we are conducting a series of to develop, improve and standardize methods for collecting and diagnostic data to provide better estimates of key parameters models. In particular, we have been developing tools for strains of Pasteurella haemolytica and quantifying immunological �106 responses of bighorns to infection by these pathogens. These tools will be key components of laboratory and field experiments designed to evaluate potential tactics (including vaccination and/or treatment) for managing pasteurellosis in wild sheep, and appear prerequisite to initiating management level experiments. To this end, our recent efforts have focused on both simulation modeling and on improving tools available for use in future management experiments that will be designed to study etiology, epizootiology, and prevention or control of disease outbreaks in bighorn populations: METHODS Management of Bacterial AND MATERIALS and Viral Diseases in Mountain Sheep Populations In conjunction with numerous cooperators, we continued developing and improving tools available for use in studying etiology, epizootiology, prevention or control of disease outbreaks in bighorn populations: and Epizootiology of pasteurellosis in indigenous bighorn populations (Miller, Spraker, Mills, snipes, and Kraabel): We used ribosomal RNA fingerprinting and in vitro measures of cytotoxin production to compare Pasteurella haemolytica isolates from eight indigenous Rocky Mountain bighorn sheep herds (Almont/Taylor River, Avalanche Creek, Chalk Creek, Cottonwood Creek, Grant, Tarryall Mountains, Texas Creek, waterton Canyon) in Colorado. Genomic fingerprinting (Snipes et ale 1992) of remaining untyped isolates (n ~ SO) was completed. We also continued evaluating potency of cytotoxins derived from genotypically-distinct P. haemolytica isolates in vitro using methods described by Silflow et ale (1993). RESULTS Management of Bacterial AND DISCUSSION and Viral Diseases in Mountain Sheep Populations Epizootiology of pasteurellosis in indigenous biahorn populations: Using ribosomal RNA gene restriction patterns, at least 26 distinct strains of P. haemolytica were identified among isolates (n = 59) from these herds; we identified one to seven distinguishable ribotypes within individual herds. Of the 26 ribotypes identified, 21 appeared unique to individual herds, four others (E, N, T, BB) were shared by only two herds, and one (A) was common to 3 herds. . In vitro evaluation of cytotoxin production by genotypically-distinct P. haemolytica isolates revealed further diff.erences among strains: Four ribotypes (AA, B, E, 0) showed marked cytotoxin production -- bighorn neutrophil death rate @ 150 ~g culture supernatant was 4-9 times that of an Enterobacter sp. control (7.1±0.4% neutrophil death @ 150 ~g supernatant) (Fig. 1); cytotoxicity of the other 22 strains' examined approximated control levels. All three indigenous bighorn herds that yielded markedly cytotoxic P. haemolytica strains have recent histories of pneumonia epizootics: pasteurellosis outbreaks occurred in the Taylor River herd in 1979 and again in 1991, in the Waterton Canyon herd in 1980, and in the Chalk Creek herd in 1981. One of these strains (E) was also recovered from dead bighorns during recent epizootics in the Alamosa Canyon (1989) and Rock Creek (1990) herds these latter herds can be linked to Taylor River by bighorn translocation activities during the last decade. Our findings support hypotheses that strains of P. haemolytica carried by healthy bighorn sheep may vary within and among wild popUlations. Our preliminary results also suggest the combination of genomic fingerprinting and cytotoxicity determination may offer a useful approach for studying the epizootiology of pasteurellosis within and among bighorn herds and may provide insights into strategies for effectively preventing or managing pneumonia �107 epizootics. We are currently reported here. preparing a manuscript LITERATURE summarizing the findings CITED Snipes, K. P., R. W. Kasten, M. A. Wild, M. W. Miller, D. A. Jessup, R. L. Silflow, W. J. Foreyt, and T. E. Carpenter. 1992. Using ribosomal RNA gene restriction patterns in distinguishing isolates of Pasteurella haemolytica from bighorn sheep (Ovis canadensis). J. Wild1. Dis. 28: 347-354. Silflow, R. M., W. J. Foreyt, and R. W. Leid. 1993. Pasteurella haemolytica cytotoxin-dependent killing of neutrophils from bighorn and domestic sheep. J. Wild1. Dis. 29: 30-35. Prepared by veterinarian �108 10 - ------------------------------------- >o B -x o o 6 >oQ) > 4 ~ Q) c::r 2 o OTHERS AA EcoRI B E o ribotypes Figure 1. In vitro evaluation of cytotoxin production by genotypicallydistinct P. haemolytica isolates revealed differences among strains: 4 ribotypes (AA, B, E, 0) showed cytotoxin production 4-9 times that of an Enterobac~er sp. control; cytotoxicity of the other 22 strains examined approximated control levels. �109 Colorado Division Wildlife Research July 1994 of Wildlife Report JOB PROGRESS State of Project Work Colorado No. W-153-R-7 Plan No. Job No. Period REPORT Covered: Authors: Personnel: Mammals Research 2A Mountain 7 Experimental Evaluation of Mountain Transplanting and Disease Treatment Sheep Investigations Sheep July 1, 1993 - June 30, 1994 M. W. Miller, J. Vayhinger, S. Roush, and D. Lovell A. Archuleta, J. Backstrand, F. Barnes, R. Dobson, J. Duran, B. Elkins, W. Fey, D. Getzy, R. Green, V. Jurgens, R. Hancock, M. Lamb, K. McLaughlin, S. Ogilvie, G. Roberts, G. Rudd, A. Stencel, A. Torres, T. Verry, C. Yoder, R. Zaccagnini. ABSTRACT We continued monitoring lamb survival among radiocollared ewes from 4 freeranging bighorn herds as part of a 4-year management experiment to examine effects of alternative lungworm treatment strategies on lamb survival and population performance. Since December 1991, 2 bighorn herds in the Tarryall Mountains [Sugarloaf Mountain (SL) and Twin Eagles (TE)] and 2 herds in the Collegiate Peaks [Chalk Creek (CH) and Cottonwood Creek (CW)] have been managed under 1 of 4 alternative lungworm treatment regimes: baiting with alfalfa hay and apple pulp treated with fenbendazole, baiting with alfalfa hay and apple pulp without fenbendazole, placing fenbendazole-treated salt blocks on bait stations, and withholding bait and fenbendazole. Treatments have been rotated annually under a predetermined, randomly-selected schedule. We monitored lamb production and survival among radiocollared ewes in each herd from May through October 1993 to complete the second field season. Year 3 started in midDecember 1993 with baiting and treated salt block distribution at scheduled sites. We began monitoring lamb production and survival among radiocollared ewes in each herd to assess year 3 treatment responses in May 1994. Because this experiment will not be completed until October 1995, data presented here relative to treatment effects are preliminary and we have made no attempt to interpret them. Both production and mortality affected recruitment (lambs/marked ewes) through October among lamb cohorts monitored during May-October 1993. Lamb production, as estimated by observations of marked ewes with new (~ 2 wk old) lambs at heel, ranged from 0.76 at TE to 0.94 at SL. In addition to reproductive/perinatal losses, some lambs disappeared in each of the experimental herds during the summer. Lamb survival (lambs in October/lambs born to marked ewes) through October ranged from 0.73 at SL to 1.0 at TE. Although the proportion of ewes observed with lambs and the proportion of those lambs surviving varied somewhat among herds, overall recruitment of lambs through October appeared to be relatively high across all 4 herds in 1993, ranging from 0.67-0.79 lambs/marked ewes. As many as 73 sheep fed at the SL bait site and as many as 90 sheep fed at the CW bait site during December 1993-February 1994. All marked ewes fed on at least 7 days. Marked ewes at SL (n = 17) averaged 45 days on bait (sd = 1.1) and marked ewes at CW (n = 15) averaged 26 days on bait (sd = 15.1). �110 Radiocollared ewes visited the SL site almost daily; in contrast, most marked ewes at CW visited the bait site infrequently during December-January. In addition to bait, all 17 marked ewes at SL also received 2 fenbendazole treatments. Four 15 kg fenbendazole-treated salt blocks were available to sheep at 3 sites within the TE winter range between January and May 1994. Block consumption ranged from 3.2-15 kg; by comparison, a control block lost about 2.3 kg via environmental effects during January-May. Adjusting for estimated environmental losses, about 23.7 kg of treated blocks (about 39.1 g fenbendazole) were consumed at TE. Block consumption equated to about 6.5 ewe treatments at TE in 1994, compared to 26.5 at CW in 1993 and 6 at SL in 1992; under a more moderate dosing regime, block consumption equated to about 17 ewe treatments at TE in 1994. We observed 33 marked and unmarked TE sheep using a single block in mid-February; mule deer may also have used treated blocks at TE occasionally. Field data for the third lambing season span only May-June 1994, and consequently are quite preliminary. ,Observed lamb production through June 1994 ranged from 0.59 at SL to 1.0 at TE; lamb survival through June ranged from 0.91 at CH to 1.0 at both CW and SL. Necropsy of a sick lamb collected at CH in June revealed bronchopneumonia that probably arose secondary to a congenital heart defect. Pasteurella haemolytica, biotype A, was isolated from pneumonic lung tissue, but no lungworms were observed grossly or histologically. Relatively consistent and predictable range use and movement patterns for each of the 4 study herds have emerged since monitoring began in 1991. Plots of May 1991-March 1994 location data (UTM coordinates) for radiocollared ewes revealed apparent differences in distribution and movement patterns among herds. Ewes from the CW herd showed wide~t distribution and greatest movements; CH ewes were the most limited in their range use and movement. Although ranges of the SL and TE herds appeared to overlap considerably, to date we have observed no exchange of radiocollared ewes between these 2 herds. Disturbances by hikers (CW) and hunters (CH, SL) appeared to influence movements of ewe/lamb groups on occasion. No health problems have been detected among marked ewes. However, 5 radiocollared ewes (3 at CW, 1 each at SL and TE) died or disappeared during January-June 1994, and a total of 14 radiocollared ewes (5 at TE, 4 at SL, and 5 at CW) have died of noncapture causes since 1991. Of these, lion predation may have been involved in 5 losses in the Tarryalls (3 at TE and 2 at SL) and injuries from a fall may have claimed 1 loss (at CW); causes of death or disappearance for 8 other ewes (2 at TE, 2 at SL, and 4 at CW) have not been determined, although we speculate lightning strikes may have been involved in 2 deaths and 2 disappearances at CW during late June-early July. Losses appeared to occur relatively uniformly across seasons. Overall, noncapture mortality rates of adult ewes in the 4 study herds have averaged about 0.07 (se = 0.002) annually over the past 36 months, but causes and rates (ranging from 0 at CH to 0.17 at CW in 1994) of ewe mortality appeared to vary somewhat among herds. Despite observed variation in recruitment and adult mortality rates, winter range counts during 1991-1994 suggest all 4 bighorn herds under study have remained stable or grown since our experiment began in 1991. �111 EXPERIMENTAL EVALUATION OF MOUNTAIN SHEEP TRANSPLANTING AND DISEASE TREATMENT M. W. Miller, J. Vayhinger, S. Roush, and D. Lovell P. N. OBJECTIVE Design, conduct, and report on management experiments to evaluate efficacy transplanting and disease treatment practices for managing mountain sheep populations. AGREEMENT Continue parasite a management level experiment control program. of OBJECTIVE evaluating Colorado's mountain sheep We continued monitoring lamb survival among radiocollared ewes from 4 freeranging bighorn herds as part of a management experiment to examine effects of alternative lungworm treatment strategies on bighorn lamb survival and population performance. Year 2 of this 4-year study ended with completion of the summer field season in October 1993; year 3 began in mid-December 1993 with baiting and treated salt block distribution at scheduled sites and will continue through October 1994. Our study will be completed in October 1995. MATERIALS AND METHODS Beginning in December 1991, we began managing each of 4 study herds [Tarryall Mountains: Twin Eagles (TE) and sugarloaf (SL); Collegiate Mountains: Chalk Creek (CH) and Cottonwood Creek (CW)] under 1 of 4 alternative lungworm treatment regimes: Control Treat OnlyBait Only Bait/Treat-:- no treatment -- bait and fenbendazole withheld; fenbendazole~treated salt blocks placed on bait stations; baited with alfalfa hay and apple pulp but not treated with fenbendazole; baited with alfalfa hay and apple pulp and treated with fenbendazole. Treatments were assigned to study herds as prescribed rotating schedule (Table 1.; Year 1 = 1992). by a randomly selected, Experimental treatments during the winters of years 2 (1993) and 3 (1994) were applied at scheduled sites (Table 1). In year 2, we baited for 64 days (ending 18 February) at SL and 66 days (ending 20 February) at CHi sheep at the CH site were also treated with fenbendazole (about 3 g/adult ewe) added to apple pulp on 17 January and 17 February. In addition, treated salt blocks (1.65 g fenbendazole/kg, 15 kg/block; 8 blocks total) were available to sheep at CW during December 1992-May 1993-- we initially offered 3 blocks in December, and added 3 more blocks in February and 2 more blocks in April. One block held in a wire cage during December-May was used as an environmental control. During year 3, we baited for 55 days (ending 7 February) at SL and 76 days (ending 28 February) at CW; sheep at the SL site were also treated with fenbendazole (about 3 g/adult ewe) added'to apple pulp on 31 January and 7 February. Treated salt blocks (1.65 g fenbendazole/kg, 15 kg/block; 4 blocks total) were available to sheep at TE during January-May; 1 block held in a wire cage during that same period was used as an environmental control. We assessed effects of winter treatments on lamb production and survival by observing radiocollared ewes from all 4 herds about once every 2 weeks from May �112 through October to determine whether they produced lambs, and whether their lambs were still alive. In addition to lamb survival data, we recorded approximate UTM coordinates, habitat type, and group size and composition for each radiocollared ewe observed. All field data were transcribed into a computerized database to aid in mapping seasonal range movements and determining annual lamb production and survival rates. Radiocollared ewes were also monitored every 2-4 weeks to detect mortality and movements during November through April in conjunction with a USFS/CDOW cooperative project to identify critical winter and transitional ranges of these 4 herds. Sixty-two radiocollared ewes were located and observed biweekly during JulyOctober 1993 (15 at CH, 14 at CW, 16 at SL, and 17 at TE) to assess treatment effects for year 2. Two marked ewes died during December 1993-January 1994. We radiocollared 6 additional ewes during January and February 1994 (2 at SL, 4 at CW) to increase sample sizes and/or replace losses; all were immobilized with a combination of carfentanil HCl (1.5 mg), ketamine HCl (100 mg), and xylazine HCl (20 mg) delivered via syringe darts over bait and reversed with na1trexone HCl (50 mg IV + 100 mg SC). Consequently, 66 radiocollared ewes (15 at CH, 17 at CW, 17 at SL, and 17 at TE) were available for biweekly observation when assessment of year 3 treatment effects began in May 1994. RESULTS Because this experiment presented here relative no attempt to interpret AND DISCUSSION will not be completed until October 1995, data to treatment effects are preliminary and we have made them. Treatment Rates Year 2: As many as 72 sheep fed at the SL bait site and as many as 34 sheep fed at the CH bait site during December 1992-February 1993. All marked ewes fed on at least 14 days. Marked ewes at SL (n = 14) averaged 38 days on bait (sd = 1.9) and marked ewes at CH (n = 15) averaged 51 days on bait (sd = 10.8) • .Visitors harassment of sheep at the SL site apparently contributed to their .somewhat erratic bait site attendance. Attendance at CH was somewhat higher and more consistent in year 2 than in year 1 [mean(sd) = 42(±13.7) days]. In addition to bait, 14 of 15 marked ewes at CH also received fenbendazole at least once and 13 received 2 fenbendazole treatments in year 2. Eight fenbendazole-treated blocks disappeared almost entirely at the CW site between December 1992 and May 1993 -- we recovered a total of 339 g of remnants in May. A control block from the same site weighed only 6.44 kg in May; we attributed the observed 57% loss largely to dissolving effects of heavy, wet snowfalls in February-April. Adjusting for estimated environmental losses (about 20% for blocks in December-February and February-April and about 17% for April-May), about 95.7 kg of treated blocks (about 157.9 g fenbendazole) were consumed at CW, more than a 4-fold increase over total consumption recorded at SL in 1992 (about 36.3 g fenbendazole). Using Schmidt et al.'s (1979) dose recommendation (3 g fenbendazole/ewe/day, twice), block consumption equated to about 26.5 ewe treatments at CW in 1993, compared to 6 ewe treatments at SL in 1992; under a more moderate dosing regime (about 0.75 g fenbendazole/ewe/day for 3 consecutive days; Foreyt and Coggins 1990), block consumption equated to about 70 ewe treatments at CW in 1993 and 16 ewe treatments at SL in 1992 ••.We never observed use of blocks at CW directly, but did observe radiocollared· sheep within 25 m of blocks and sheep tracks around blocks. Observations and tracks also suggested elk and mule deer were probably using these blocks, thereby confounding estimates of drug delivery to radiocollared bighorn ewes. It follows that actual delivery of fenbendazole to CW bighorns in 1993 was probably only a fraction of the estimated total. In contrast, we believe estimated fenbendazole delivery at SL in 1992 was considerably more reliable because only bighorns appeared to use treated blocks at that site. It follows that assuring delivery of effective fenbendazole doses to target animals may represent a significant obstacle to field use of treated salt blocks in many parts of Colorado where bighorn, mule deer, and elk populations are sympatric. �113 Year 3: As many as 73 sheep fed at the SL bait site and as many as 90 sheep fed at the CW bait site during December 1993-February 1994. All marked ewes fed on at least 7 days. Marked ewes at SL (n = 17) averaged 45 days on bait (sd 1.1) and marked ewes at CW (n 15) averaged 26 days on bait (sd 15.1). Radiocollared ewes visited the SL site almost daily. In contrast, most marked ewes at CW visited the bait site infrequently during December-January. Many ewes in this herd stayed on alpine winter ranges until heavy snows apparently forced them to lower elevations in February, when CW bait site attendance increased markedly. In addition to bait, all 17 marked ewes at SL also received 2 fenbendazole treatments (31 January and 7 February). = = = Four 15 kg fenbendazo1e-treated salt blocks were available to sheep at 3 sites within the TE winter range between January and May 1994. Block consumption ranged from 3.2-15 kg; by comparison, a control block lost about 2.3 kg via environmental effects during January-May. Adjusting for estimated environmental losses, about 23.7 kg of treated blocks (about 39.1 g fenbendazole) were consumed at TE. Block consumption equated to about 6.5 ewe treatments at TE in 1994, compared to 26.5 at CW in 1993 and 6 at SL in 1992; under a more moderate dosing regime, block consumption equated to about 17 ewe treatments at TE in 1994. We observed 33 marked and unmarked TE sheep using a single block in mid-February; mule deer may also have used treated blocks at TE occasionally. Lamb Production and Survival Year 2: Both production and mortality affected recruitment (lambs/marked ewes) through October among lamb cohorts monitored during May-October 1993. Lamb production, as estimated by observations of marked ewes with new « 2 wk old) lambs at heel, ranged from 0.76 at TE to 0.94 at SL (Fig. 1) •. We recognize that our approach for estimating lamb production cannot discern between failures to bear live lambs and perinatal mortality among viable lambs. However, because lungworm treatment is directed specifically at reducing mortality in otherwise healthy 2-6 month old lambs, we believe it important to partition the contributions of reproductive failure and perinatal mortality from mortality in older, viable lambs as sources of reduced recruitment among lamb cohorts. In addition to reproductive/perinatal losses, some lambs disappeared in each of the experimental herds during the summer of year 2. Lamb survival (lambs in October/lambs born to marked ewes) through October 1993 ranged from 0.73 at SL to 1.0 at TE (Fig. 1). Field observations support previous indications that factors affecting lamb survival may vary among herds. At CH, 5 coughing lambs were observed in CH during mid-August through October 1993. (Coughing lambs were also observed in CH during summer field work in 1991 and 1992). Although 1 sick lamb at CH subsequently disappeared, 4 others observed coughing survived through October and lamb survival at CH in 1993 was 0.77. Sick or coughing .lambs have not been observed in the 3 other herds studied, even though lamb survival rates from 2 of those herds (TE = 0.79 in 1992 and SL = 0.73 in 1993) have approximated those observed at CH (Fig. 1). Although the proportion of ewes observed with lambs and the proportion of those lambs surviving varied somewhat among herds, overall recruitment of lambs through October appeared to be relatively high across all 4 herds in 1993, ranging from 0.67-0.79 lambs/marked ewes (Fig. 2). Year 3: Field data for the third lambing season span only May-June 1994, and consequently are quite preliminary. Observed lamb production through June 1994 ranged from 0.59 at SL to 1.0 at TE (CH = 0.73, CW = 0.94); lamb survival through June~ranged from 0.91 at CH and 0.94 at TE to 1.0 at both CW and SL (Fig. 1). Necropsy of a sick lamb collected at CH in June revealed bronchopneumonia that probably arose secondary to a congenital heart defect. Pasteurella haemolytica, biotype A, was isolated from pneumonic lung tissue, but no lungworms were observed grossly or histologically. No other sick or coughing lambs were observed at CH or elsewhere through June. �114 Range Use and Movement Patterns Relatively consistent and predictable range use and movement patterns for each of the 4 study herds have emerged since monitoring began in 1991. Plots of May 1991-March 1994 location data (UTM coordinates) for radiocollared ewes revealed apparent differences in distribution and movement patterns among herds (Fig. 3). Ewes from the CW herd have consistently shown widest distribution and greatest movements; CH ewes were the most limited in their range use and movement. Although ranges of the SL and TE herds appeared to overlap considerably, to date we have observed no exchange of radiocollared ewes between these 2 herds. Disturbances by hikers (CW) and hunters (CH, SL) appeared to influence movements of ewe/lamb groups on occasion. Location data gathered since March 1994 will be added to further define key ranges and migration corridors for these 4 herds. population Parameters and Performance No obvious health problems were detected among marked ewes during summer or winter observation. However, 5 radiocollared ewes (3 at CW, 1 each at SL and TE) died or disappeared during January-June 1994, and a total of 14 radiocollared ewes (5 at TE, 4 at SL, and 5 at CW) have died of noncapture causes since 1991. Of these, lion predation may have been involved in 5 losses in the Tarryalls (3 at TE and 2 at SL) and injuries from a fall may have claiIDed 1 loss (at CW); causes of death or disappearance for B other ewes (2 at TE, 2 at SL, and 4 at CW) have not been determined, although we speculate lightning strikes may have been involved in 2 deaths and 2 disappearances at CW during late June-early July. Losses appeared to occur relatively uniformly across seasons. Overall, noncapture mortality rates of adult ewes in the 4 study herds have averaged about 0.07 (se = 0.002) annually over the past 36 months, but causes and rates (ranging from 0 at CH to 0.17 at CW in 1994) of ewe mortality appeared to vary somewhat among herds. Despite observed variation in recruitment and adult mortality rates, winter range counts during 1991-1994 suggest all 4 bighorn herds under study have remained stable or grown since our experiment began in 1991. Intensive monitoring will continue through October 1994 with emphasis on documenting survival of known lambs and movement patterns. Additional monitoring is planned in conjunction with experimental treatments in year 4, and intensive monitoring will resume in May 1995. prepared by �115 Table 1. Treatment assignments for 4 bighorn herds included in a 4-year management experiment to examine effects of alternative lungworm treatment strategies on bighorn lamb survival and population performance. HERD COLLEGIA TE MOUNTAINS YEAR CHALK CREEK 1992 B' 1993 TARRY ALL MOUNTAINS SUGARLOAF MOUNTAIN TWIN EAGLES C T BfT BfT T B C 1994 C B BfT T 1995 T BfT C B COTTONWOOD CREEK , Treatment assignments: BfT = bait with alfalfa hay and apple pulp treated with fenbendazole; B = bait with alfalfa hay and apple pulp without fenbendazole; T = fenbendazole-treated salt blocks on bait stations; and C = withhold all bait and fenbendazole (control). �116 100 ..•..........•.......... Year 1 (May-October 1992) BO 60 n - 12) ..-.. ..... 0 0........ _.... 40 20 co COTTONWOOD (Control) SUGARLOAF (Treat Only) CHALK CREEK (Bait Only) TWIN EAGLES (Bait/Treat) > > '::J en 100 ......•................................ Year 2 (May-October 1993) "0 BO C CO c o •••• 60 40 20 o ::J "0 COTTONWOOD (Treat OnIy) SUGARLOAF (Bart Only) CHALK CREEK (Bait/ Treat) TWIN EAGLES (Control) COTTONWOOD (Bait Only) SUGARLOAF (Bail/Trsat) CHALK CREEK (Control) TWIN EAGLES (Treat Only) o '0. .0 E CO ...J 100 BO 60 40 20 Herd (treatment) Figure 1. Both production and mortality affected lamb recruitment during years 1 and 2 of our 4year study. Lamb production (open bars) was estimated by observations of marked ewes seen with lambs at heel; lamb survival {shaded bars} was estimated by following survivorship of lambs born to marked ewes through October. Values in parentheses are the number of marked ewes in each herd during the May-October observation period. Because data for year 3 {1994} cover only May-June, they probably underestimate lamb production and overestimate lamb survival. �117 CD 100 ........................................................... o 80 ..................... ..c o - o O'y~~~'1" ~.Year.2. ..c Ol ::s 60 o •.... ..c - 40 C CD E ::s •.... o CD 20 o .__-'-- __ COTTONWOOD SUGARLOAF CHALK CREEK TWIN EAGLES II Herd Figure 2. Although the proportion of ewes observed with lambs and the proportion of those lambs surviving varied somewhat among herds, overall recruitment of lambs through October appeared to be relatively high across all 4 herds in years 1 and 2, ranging from 0.62-0.84 lambs/marked ewes. �118 BIGHORN SHEEP TREATMENT STUCY COTTONWOOD CREEl< HERD (JUL Y 1991 - MARCH 1994) BIGHORN SHEEP TREATMENT STUOY CHALK CREEl< HERD (MAY 1991 - MARCH 1994) + + + 10000 .. I10UNT PfU He!.! CH + ++ .... "it~~E + + + + + .•"""'" AHTERO + ++ + LITILE BROWN' S CREEl< + +++ + + + c, + + SICAL£ 1: 110000 BIGHORN SHEEP TREATMENT STUCY SUGARLOAF MOUNTAIN HERO (MAY 1991 - MARCH 1994) BIGHORN SHEEP TREATMENT STUDY TWIN E.lGLE HERD (MAY 1991 - MARCH 1994) LOST CREEK McCURDY CREEK x ~ SISON REFRIGERATOR ULCH X%}" ..X PEAK X xX ·X X ~X McCURDY CREE.r X .. )(;:tlRly X X ,." •••• "X .. X •. PILOT Figure 3. Plots of May 1991-March 1994 location data (UTM coordinates) for radiocollaredewes revealed apparent differences in distribution and movement patterns among herds. Ewes from the CW herd showed greatest distribution and movements; CH ewes' range use and movements were the most limited. Although ranges of the SL and TE herds overlap, no exchange of radiocollared ewes between these 2 herds has been observed. x 3 1C11..urw �119 Colorado Division Wildlife Research July 1994 of Wildlife Report JOB FINAL State of Project Work Colorado No. ~W~-..:::1~5~3~-~R._-....:.7 _ Plan No. __~2~A~ Job No. Period Author: REPORT 9 Covered: _ Mammals Mountain Research Sheep Investigations Quantity and Quality of Mountain Sheep Habitat with Regard to Minimum Viable Populations and Response of Mountain Sheep to Human Activity July 1, 1993 - June 30, 1994 D. F. Reed ABSTRACT Cooperative Agreement work was completed with the Bureau of Land Management, Canon City District, in mapping and evaluating mountain sheep habitats. This was done by using MIPS (Map and Image Processing System) for GIS (Geographic Information Systems) integration of mountain sheep distribution and movements, vegetation, viewshed, and water sources. Data collected on mountain sheep included group size, location, movements, marks (collars and/or eartags), sex and age classification, ram translocations, and disease paradigms. Recommendations were made for two water developments. Planned reports for Area A (west of Parkdale) and Area B (Brown's Canyon north of Salida and county-line area east of Salida) were combined into one report titled "Mountain sheep habitat use in the Arkansas River Canyon, Colorado" (Reed et al. 1994). ��121 QUANTITY AND QUALITY OF MOUNTAIN POPULATIONS AND RESPONSE SHEEP HABITAT WITH REGARD TO MINIMUM OF MOUNTAIN SHEEP TO HUMAN ACTIVITY VIABLE Dale F. Reed P. N. OBJECTIVE Evaluate the quantity and minimum viable populations activity. quality of mountain sheep habitat and test the response of mountain SEGMENT with regard to sheep to human OBJECTIVES 1. Complete cooperative agreement between Bureau of Land Management, District, and the Colorado Division of Wildlife for mapping and mountain sheep habitats in Areas A and B • 2. Prepare combined report Canon City evaluating for Area A and Area B. ACKNOWLEDGMENTS I thank co-principal investigators M. W. Miller. R. B. Gill, J. Vayhinger,· and S. Ogilvie for their ideas and support. Division personnel D. Finch, W. Travnicek, J. Backstrarid, and L. Spicer were helpful in collecting field data. Division personnel D. L. Schrupp and D. C. Lovell were helpful for their ideas and in coordinating GIS. BLM E. Brekke made field observations and provided important coordination. University of Colorado, Colorado Springs, T. P. Huber developed the GIS procedures and products. DESCRIPTION The study (1994). areas have been described METHODS Methods and materials are described RESULTS Results and discussion Reed et al.(1994). OF AREAS by Reed (1991, 1992, 1993) and Reed et al. AND MATERIALS in Reed (1992,1993) and Reed et al. (1994). AND DISCUSSION for Area A and B are described in Reed (1992, 1993) and �122 'LITERATURE Reed, CITED D. F. 1991. Quantity and quality of mountain sheep habitat with regard to minimum viable populations and response of mountain sheep to human .activity. Colo. Div. of Wildl., Wi1dl. Res. Rep. July:157-160. 1992. Quantity and quality of mountain sheep habitat with regard to minimum viable populations and response of mountain sheep to human activity. colo. Div. of Wildl., Wildl.. Res. Rep. -July:173-193. 1993. Quantity and quality of mountain sheep habitat with regard to minimum viable populations and response of mountain sheep to human activity. colo. Div. of Wildl., Wildl. Res •.Rep. July: 183-198. ______~~~, J. Vayhinger, E. B. Brekke, and T. P. Huber. 1994. habitat use in the Arkansas River Canyon, Colorado. Colo. pp. Prepared by 9Ja~ Dale F. Ree Wildlife Researcher Mountain sheep Div. of Wildl. ���125 Colorado Division Wildlife Research July 1994 of Wildlife Report JOB PROGRESS State of Project REPORT Colorado No. W-153-R-7 Mammals Research Work .Plan No. 3A Pronghorn Job No. 2 Habitat Selection and population Performance of a Pioneering Pronghorn Population Period Author: Covered: July Investigations 1, 1993 - June 30, 1994 T. M. Pojar Abstract The rate of increase (ROI) for the Middle Park pronghorn population is the lowest (0.08) since observations began in 1987. The projected late summer 1994 population size is 458 (Table 3) which is the 8th consecutive year of positive growth. Using the linear regression technique (ROI on population size) to estimate carrying capacity, the projected K-value for this population is 525 animals (Figure 1). There were distinct differences in winter .di.stribution .of the .. population. from past years, which may be due to density approaching carrying capacity. Twenty 1993 fawns (10 males and 10 females) were equipped with radio collars in December 1993 initiating a new objective for estimating differential natural mortality between males and females. At this writing, all radios are functional and all 20 yearlings are alive. ��127 HABITAT SELECTION AND POPULATION PERFORMANCE PRONGHORN POPULATION OF A PIONEERING Thomas M. Pojar P.N. OBJECTIVE Describe population dynamics pronghorn population. and habitat SEGMENT use of a pioneering, expanding OBJECTIVES 1. Describe seasonal population. and annual distribution 2. Monitor natural mortality males and females. 3. Map areas of habitation 4. Monitor population dynamics of Middle Park pronghorn with: a. Ground counts to describe changes in population size. b. Ground counts to quantify population sex and age composition. and movement of the Middle patterns Park pronghorn of radioed yearling using the GIS format. STUDY AREA The study area is described (1993) in Pojar METHODS SEASONAL AND ANNUAL (1988) and a map of the area is in Pojar AND MATERIALS DISTRIBUTION Tracking was done mostly from the ground; fixedwing aircraft was used if an animal could not be located after a reasonable effort from the g~ound. Legal descriptions of animal locations were recorded to the nearest quarter mile then converted to UTM (U.S. Army 1973) coordinates for computer processing. All radioed animals have been locat~d biweekly (with very few excepti~ns) since January 1, 1987. POPULATION SIZE AND STRUCTURE Herd structure estimates were obtained by classifying all animals that accompanied the animals that are radioed. The herd structure estimate used in population projections is the one with the largest sample size obtained in August or September. Total counts are made during winter by counting all animals associated with radioed animals. With the increased population size, it is not always possible to get an accurate count of total mature bucks (1.5 yrs and older) in the population. However, it is still possible to get very accurate counts of bucks in 60-80% of the population. The proportion of bucks in this portion of the population is then extrapolated to the total population to estimate total mature bucks. Total population count during winter, estimated number of mature bucks from the winter count, and recruitment based on fawn to doe ratios from late summer are used for the population projection. Population 1. 2. 3. projections are based on the following assumptions: Winter counts represent the total population and the estimated number of mature bucks in Middle Park. Late summer age ratio estimates represent "recruitment" into the population. Annual survival of mature bucks and does and female fawns is 92.5%. �128 4. NATURAL Annual survival of males in their first year (after weaning) is 50%. (This severe mortality on male fawns is arbitrary, however, it allows the number of mature males in subsequent years to match fairly well with winter counts.) MORTALITY OF MALES AND FEMALES The methods for estimating differential natural female pronghorn are outlined in Appendix I. mortality between male and RESULTS SEASONAL AND ANNUAL DISTRIBUTION The winter distribution during 1993-94 exhibited some distinct differences from past years. A group of 115 spent the entire winter in the Sulphur Gulch area (T1N,R79W,S6). This group included all radioed animals that migrate east to Corral Creek and Granby in summer. Another somewhat distinct group of 70 ± wintered north of Antelope Pass (T2N,RSOW,S6). The remainder of the population wintered in the customary area north and east of the town of Kremmling. POPULATION SIZE AND STRUCTURE Total population size estimates are obtained during winter and herd structure estimates are obtained in late summer (Table 1). The annual changes in population. size are used to calculate the rate of increase which is regressed on population size to project the K-value for the population. The ROI is calculated as where Pl is the population size at time land P2 is the population at time 2 (Table 2). The rate of increase for 1993-94 is O.OS (Table 2), which is the lowest yet observed for this population. Based on the relationship of population size and ROI, the projected K-value is 525 (Figure 1). The population projection for late summer 1994 is presented in Table 3. In this projection it is assumed that the 1994 fawn production is 44 fawns:100 does based on a mid-summer 1994 age structure estimate. This is a critical assumption and subsequent ratio estimates based on larger sample sizes may have a significant impact on this projection. The herd structure estimates obtained in this study should not be subject to the shortcomings of herd structure estimates discussed by MCCullough (1994) because of accurate population size data. The winter of 1993-94 was relatively mild with· low snow fall and accumulation, and no extended periods of sub-zero (~) temperatures. These conditions may have contributed to the expanded distribution of the wintering population. Lack of snowpack and an extremely dry spring and early summer may have reduced the growth of forbs and subsequently adversely affected fawn survival (Ellis 1970) resulting in the relatively low observed fawn to doe ratio. The above mentioned phenomena may also be the manifestation of population density approaching carrying capacity (Figure 1). NATURAL MORTALITY OF YEARLING MALES AND FEMALES Since this is a new objective to this project, a study plan was prepared and is included in Appendix I. A trapping operation, using the conventional drive trap, was conducted on December 14, 1993. The objective was to radio 10 male and 10 female fawns. Totally 125 animals were trapped to obtain the number of target animals to radio (Table 4). �129 Table 1. Herd structure of Middle Park pronghorn based on a sample obtained by locating radioed animals in late summer. The population size is from the subsequent winter counts with harvest added back into the population to get the pre-hunt population size, e.g. 1993 pre-hunt population was 425, 410 winter count plus 15 harvest. YEAR POP. SIZE NO. RADIO RADIO B:100D RATIO F:100D RATIO SAMPLE % OF POP. 19861 1987 1988 1989 1990 1991 1992 1993 80 122 160 223 261' 308 347 425 7 24 22 17 13 39 31 5.7 15.0 10.2 6.5 4.2 11.2 7.3 36 54 40 56 22 23 26 10 77 77 32 50 47 65 48 66 47 63 108 161 148 148 286 266 S9 S2 68 72 66 48 82 63 This year's 1986. 1 % data based on the sample of the population trapped 16 December Table 2. Population size of the Middle Park pronghorn herd during winter and the calculated rate of increase. Population size reflects the removal of 15 animals per year by harvest beginning in 1990, i.e. the 1990-91 winter population was 261 before harvest and 246 after harvest. YEAR POP. SIZE 1986-87 1987-88 1988-89 1989-90 1990-91 1991-92 1992-93 1993-94 1994-95 80 122 160 223 246 292 332 410 443 (Projected) Table 3. Population projection text for the assumptions. I POPULATION I RATE OF INCREASE BUCKS .52 .31 .39 .10 .19 .14 .23 .08 for the Middle I DOES Park pronghorn population. I I FAWNS TOTAL WINTER '93-94 78 198 134 410 WINTER MORTALITY 78 X .075 = 6 MORT 198 X.075 = 15 MORT 67X.5=33B 67X.075=5D 60 PREFAWNING 1994 78 - 6 = 72 MATURES + 33 YRLS TOTAL =105 198 - 15= 183 MATURES + 62 YRLS TOTAL = 245 LATE SUMMER 1994 MATURE 72 YRLS 33 TOTAL 105 MATURE 183 YRLS 62 TOTAL 245 350 @ 44F:IOOD 245 X .44 = 108 FAWNS 458 See I �130 The radios were on expandable collars developed by Dick Bartmann (CDOW). The collars put on female fawns had an original circumference of 15.5 inches (39.4 cm) and were designed to expand to 17.5 inches (44.5 cm). Male fawn collars were 15.75 inches (40 cm) originally and will expand to 19.5 inches (49.5 cm). The radio packages used on the fawns in this study weighed 250 g. This compares to packages used in the past that weighted 450-500 g. The lighter weight was accomplished by using lighter belt material for the collar, smaller batteries (2 LTC30 batteries rather than 2 C or 1 D battery), and circuitry that turns the radio off for 10 hours during nighttime (2030 to 0630) to preserve battery life. As of this writing, all 20 radios are functional and all 20 yearlings are alive. Movement of the yearlings essentially reflects that of the population with summer distribution following that or older radioed animals. The possible exception to this generalization is that 3 of the yearlings (2 females and 1 male) are summering in North Park whereas only 1 of the older animals (female) spent the early part of the summer in North Park but has since moved back into Middle Park. Table 4. Record of ear tag numbers and radio frequency of fawns radio collared on December 14, 1993 in Middle Park, Colorado (T2N,R80W,S36). tag designation Y=yellow and B=blue. REFERENCES Ellis, Sex Ear Tag F F F F F F F F F F M M M M M M M M M M Y3 Y5 Y8 Y9 Y10 Y11 Y12 B53 B54 B55 Y1 Y2 Y4 Y6 Y7 B56 B57 B58 B59 B60 Radio Ear Frequency 148.500 149.230 149.150 149.272 149.502 149.512 149.490 148.760 148.730 149.430 149.650 149.172 149.450 149.410 149.470 149.390 149.190 149.550 149.530 149.132 CITED J. E. 1970. A computer analysis of fawn survival in the pronghorn antelope. Ph.D. Thesis, Univ. of Calif., Davis. 70 pp. McCullough, D. R. 1994. What do herd composition counts tell us? Wildl. Soc. Bull. 22:295-300. Pojar, T.M. 1988. Habitat selection and population performance of a pioneering pronghorn population. Colo. Div. Wildl. Res. Rep. July, pp 181-192. . 1993. Habitat selection and population performance of a pioneering pronghorn population. Colo. Div. Wildl. Res. Rep. July, pp 199-207. u.s. A;my. 1973. Technical Manual: Universal Transverse Mercator Grid. Headquarters, Dep. of the Army, Washington D.C. TM. No. 5-241-8, 64 pp. �0.9 0.8 y = 0.5219 - 0.000994 x R 2= 0.5449 0.7 Q) 0') Q) Q) 0.6 •... CJ .5 0.5 0 '89 Q) ~al 0.4 a: • 0.3 '93 • 0.2 0.1 '90 .•.• • • ~'94 <, 0 0 100 200 300 Projected K-Value = 525 400 500 I 600 Population Size Figure 1. population Projected carrying capacity (K-Value) of Middle based on regression of ROI on population size. Park Pronghorn 700 ��133 APPENDIX I PROGRAM NARRATIVE PRONGHORN INVESTIGATIONS State: Colorado T. M. Pojar Project Title: Project Pronghorn sex difference characteristics. Number: in natural Work Plan 3Ar Job 5 June 16, 1993 mortality and movement Sample based survey methods to estimate pronghorn (Antilocapra americana) n~mbers have consistently resulted in higher population sizes than managers' "intu~tive" estimates and higher than conventional "total" count estimates (Johnson et al. 1991, Pojar et al. 1993). Attempts to build population simulations based on the larger population sizes have met with a common problem. The reported harvest of males is not sufficient to align the observed buck to doe ratio with the simulated ratio (T.M. Pojar, unpubl. data, 1986, J. Emmerich, Wyoming Game and Fish Dep., pers. comm., 1990). The simulated ratio gravitates toward unity if the natural mortality rates are kept equal for males and females and if unity is assumed in the sex ratio at birth. To align simulated buck to doe ratios with observed ratios, it is necessary to use substantially higher natural mortality rates on males compared to females. Higher dispersal rates of sub-adult males have been documented in white-tailed deer (Odocoileus virginianus) (Dusek et al. 1989, Hawkins et al. 1971, Kammermeyer and Marchington 1976, and Nelson and Mech 1992), and mule and black-tailed deer (~hemionus) (Robinette 1966, Bunnell and Harestad 1983). However, the impact of movement characteristics has not been related to survival. The magnitude of difference in sex ratio at birth is not sufficient to reconcile the difference between the observed and simulated buck to doe ratios. Therefore, differential mortality, possibly because of the risks of movement characteristics, between males and females may be the best explanation for the disparity. Sub-adult and young adult males (marginal breeding capability because of social status) may have different movement patterns than females. The survival cost of movement characteristics is the _ question that has not been addressed in reported studies. Knowledge of movements and mortality of young males would assist managers in reconciling population estimates and sex ratio estimates through the simulation process. Fetal or neonatal sex ratio does not differ substantially from unity according to published reports. From does that were somewhat randomly collected by investigators, the sex ratio of 52 fetuses from Colorado was 27 males to 25 females (Hoover et al. 1959); in Montana, of 65 fetuses, 34 were males and 31 were females (O'Gara 1968). The widest disparity was in a sample of 57 fetuses from road killed does from Alberta, Canada where 35 were males and 22 were females (Mitchell 1980). Summarizing records of 11 studies that included fetal sex ratios, Bodie (1979) reports 377 male and 343 female fetuses. A sample of 228 fawns caught a few days after birth from Colorado yielded a ratio of 111 males to 117 females (Hoover et al. 1959) and in a captive Utah population, there were 48 males and 51 females born (Smith and Beale ca. 1979). Like o'ther ungulates, pronghorn fawns encounter relatively high mortality rates during the first few weeks after birth. Although data are limited, it appears that males and females aucv Lve the first few weeks and months of life at about the same rate. From a sample of 465 fawns classified (ground survey using binoculars and spotting scope) in August, 229 were males and 236 were females (V. Graham, Colorado Div. Wildl., Pers. Comm., 1989) and based on a sample of 55 fawns Fairbanks (1993) concluded that the mortality rate of male and female fawns is not different (~~ 0.74) through weaning. Two studies on Moose (Alces alces) found no difference in survival rates between male and female calves during their first 6 months of life (Ballard et al. 1991, Osborne et al. 1991). �134 The weaning process (termination of nursing) is usually completed by the time of the rut in late September or early October (Autenrieth and Fichter 1975). After this, fawns become part of a wintering group which offers protection at small social cost because of lack of social interactions during this time of year (Kitchen 1974). However, by the time of spring dispersal, juvenile males are emancipated from their mothers and tend to avoid large males because of the onset of territorial defense. Juveniles (yearlings) are then relegated to lone existence or to joining bachelor groups that wander widely during May, June, and July (Kitchen 1974). Social interactions in these groups predominantly revolve around relative dominance leading to expenditure of energy for threats and sparing. Establishing a dominance hierarchy and wandering through unfamiliar territory, many times without being accompanied by an older ~ow+edgeable individual, may be a source of risk to yearl4:!lg..!ll_ales .that is not imposed on.'yearl),~gfelI!~les.. This investigatiC?n. would focus on the differential movement characteristics and mortality of young .and sub-adult male and female pronghorn. Objectives: 1. Test for differential natural mortality between male and female pronghorn through 2 years of age. 2. Test for different home range size and movement characteristics between male and female pronghorn through 2 years of age. Expected Results and Benefits: A better understanding of the timing and extent of natural mortality of pronghorn will allow managers to more precisely manage this species. With the advent of better (in terms of less bias) survey methods for estimating population size, knowledge of timing and extent of natural mortality by sex will foster more realistic population simulations. Population simulations are the hub of population management and demands on the resource are at an intensity that requires improved population data and more realism in population simulations. Pronghorn hunting in Colorado is based on the Division issuing permits that are specified by sex. Since the hunting permits are issued by sex, it is possible for managers to control and affect the sex ratio in the population through the harvest. Thus far, unreliable buck to doe ratio data from aerial surveys and lack of information on natural mortality by sex has stymied advances in optimum management and use of the pronghorn resource in Colorado. Approach: The hypotheses to be tested are: 1. Ho: There is no difference in the survival between male and female pronghorn fawns (older than 5-6 months), yearlings, and 2-year-olds. 2. Ho: There is no difference in home range size between male and female pronghorn fawns (older than 5-6 months), yearlings, and 2year-olds. 3. Ho: There is no difference in movement characteristics as measured by distance between locations at standardized time intervals between male and female pronghorn fawns (older than 5-6 months), yearlings, and 2-year-olds. The power. of the test, at a=.1 significance level, chosen for the first hypothesis is the ability to detect a difference in survival at the 90% confidence level. The sample needed to meet the prescribed power for a delta (difference between male and female mortality) of 0.50 would be 10 animals of each sex; for a delta of 0.40, 20 animals of each sex; and for a delta of 0.30, 30 animals of each sex (White and Garrott 1990). These sample sizes may or may not be adequate for the second and third hypotheses depending on home . range size consistency and movement characteristics between males. and females. To account for possible mortality differences between years and to augment the overall efficiency of the study (i.e. handle minimum number of animals), target animals will be radioed at a rate of 20 per year (10 females and 10 males). The maximum number to be radioed is 60 which will provide the power, at a=O.1 significance level, to detect a differ~nce of 30% in mortality �135 between males and females at the 90% confidence level. To reach the ability to detect a difference of 20% would require an overall sample of approximately 160 radioed fawns (G.C. White, Colorado State Univ., Pers. Comm.) which would be out of the realm of possibility for the Middle Park herd unless the study spanned more than 4 years. Several approaches for testing the above hypotheses have been considered, all of which involve capturing animals and equipping them with radios (either neck collars or solar powered ear tags). Factors considered in selecting animal capture method were efficiency of capture, maximizing the probability of obtaining the desired sample size, and minimizing the risk to the subject animals. The conventional drive trap will be used in December-February and care will be taken to drive groups slowly so that smaller animals (i.e. fawns) will not drop out of the group before reaching the trap. Once in hand, age can be determined by incisor examination which would minimize bias.associatEi!d with differences in fawn sLze, using this approach for capturing target animals will avoid the initial high post-parturition mortality period of newborn fawns greatly improving the efficiency of marking target animals. The negative aspect of the drive trap capture method is that many non-target animals may have to be processed through the trapping operation putting them at risk for injury. . As a supplement to the drive trap capture, the use of a helicopter and This net gun to capture animals during December or January may be necessary. method would be highly selective for the target animals, would avoid the high mortality post fawning period, and would minimize handling of non-target animals. However, this method is expensive and may have selection bias for fawn size so it would be used only if necessary to obtain the desired sample. Data Collection and Analysis: Pertinent data for determining the difference in male and female survival is whether they live or die. The subject animals will be radio tracked bi-monthly (or more often, see below) and their survival status recorded. The test for equal survival will be accomplished using Fisher's exact test (Zar 1984:390-395, Gustafson 1991:8-8) on a 2x2 contingency table consisting of sex by status (live or dead). Location data will be collected on subject animals every second or third day beginning April 1 and continuing through November 15 of each year. Each location will be determined using GPS (Global Positioning System) technology so nearly exact coordinates. can be obtained. A.minimum target sample of 30 locations per animal is desirable to characterize movement patterns. Home range size will be calculated using either program HOMER or MCPAAL. Home range size of males and females by age class will be compared for like time periods using Student's t-test. An index to movement characteristics will be derived by calculating distances between consecutive locations; these distances will be described statistically. The median of a sample of n points for each animal will be calculated as that point that has minimum average distance from the n points. The 0.5 and 0.9 quantiles of the median location will be calculated. Sex and age class comparisons of distances from median locations that contain 50% of the locations will be made testing the null hypothesis that 2 or more utilization distributions are similar using the MRPP procedure (Anderson et ale 1992). A difference in movement characteristics will be construed as a factor in survival. Location: Middle Park provides a relatively discrete population for this study. Accurate information on population size and structure is available for this herd for past years and can be monitored in the future. If population density impacts mortality or movement characteristics it will be possible to account for it. Schedule: 1992-93 1993-94 1994-95 June - complete Program Narrative December capture and radio first tier of fawns December - radio second tier of fawns �136 1995-96 1996-97 Data collection and preliminary report December - radio third tier of fawns Data collection and interim report Final report and publication preparation Personnel: Principal Investigators: Tom Pojar and Bruce Gill Consultants: Gary White and Dave Bowden Cooperators: Chuck Wagner, Jim Liewer, Rob Firth, Cathy Craig, Chuck Cesar Literature Bob Thompson, Cited . Anderson, A. E., D. C. Bowden,·and D.·Mo·,Kattner. 1992. The puma on····· Uncompahgre Plateau, Colorado. Tech. Bull. 40. Colorado Div. wildl. 116pp. Autenrieth, R. E., and E. Fichter. 1975. On the behavior and socialization of pronghorn fawns. Wildl. Monogr. 42. 111pp. Ballard, W. B., J. S. Whitman, and D. J. Reed. 1991. Population dynamics of moose in south-central Alaska. Wildl. Monogr. 114. 49pp. Bodie, W. L. 1979. Factors affecting pronghorn fawn mortality in central Idaho. M.S. Thesis. Univ. of Montana, Missoula. 98pp. Bunnell, F. L., and A. S. Harestad. 1983. Dispersal and dispersion of blacktailed deer: Models and observations. J. Mamm. 64:201-209. Dusek, G. L., R. J. Mackie, J. D. Herriges, Jr., and B. B. Compton. 1989. Population ecology of white-tailed deer along the lower Yellowstone River. Wildl. Monogr. 104. 68pp. Fairbanks, W. S. 1993. Birthdate, birthweight, and survival in pronghorn fawns. J. Mamm. 74:129-135. Gustafson, T. L. 1991. True Epistat Manual. Epistat Services, Richardson, Texas, USA. Hawkins, R. E., W. D. Klimstra, and D. C. Autry. 1971. Dispersal of deer from Crab Orchard National Wildlife Refuge. J. Wildl. Manage. 35:216-220. Hoover, R. L., C. E. Till, and S. Ogilvie. 1959. The antelope of co Lozado , Colorado Dep. of Fish and Game. Tech. Bull. No.4. 110pp. Johnson, B. K., J. G. Lindzey, and R. J. Guenzel. 1991. Use of aerial line transect surveys to estimate pronghorn populations in Wyoming. Wildl. Soc. Bull. ~9:315-321 Kammermeyer, K. E. and R. L. Marchinton. 1976. Notes on dispersal of male white-tailed deer. J. Mamm. 57:776-778. Kitchen, D. W. 1974. Social behavior and ecology of the pronghorn. Wildl. Monogr. 38. 96pp. Mitchell, G. J. 1980. The pronghorn antelope in Alberta. The Alberta Dep. of Energy and Nat. Res. Fish and Wildl. Div. 165pp. Nelson, M. E., and L. D. Mech. 1992. Dispersal in female white-tailed deer. J. Mamm. 73:891-894. O'Gara, B. W. 1968. A study of the reproductive cycle of the female pronghorn (Antilocapra americana) Ord. Ph.D. Thesis, Univ. of Montana, Missoula. 161pp. Osborne, T. 0., T. F. Paragi, J. L. Bodkin, A~ J. Loranger, and W. N. Johnson. 1991. Extent, cause, and timing of moose calf mortality in western interior Alaska. Alces 27:24-30. Pojar, T. M., D. C. Bowden, and R. B. Gill. Submitted 1993. Aerial counting experiments to estimate pronghorn density and herd structure. J. Wildl. Manage. Robinette, W. L. 1966. Mule deer home range and dispersal in Utah. J. Wildl. Manage. 30:335-349. Smith, A. D., and D. M. Beale. ca. 1979. Pronghorn antelope in Utah: Some research and observations. Utah Di v. of Wildl. Res. -Publ. No. 80-13. 88pp. White, G. C., and R. A. Garrott. 1990. Analysis of wildlife radio-tracking data. Academic Press, Inc., NY. 383pp. Zar, J. H. 1984. Biostatistical Analysis. Second Edition. Prentice Hall, NJ. 718pp. �13/ Colorado Division Wildlife Research July 1994 of Wildlife Report JOB PROGRESS state of Project Work Colorado No. Plan No. Job No. Period Author: REPORT Covered: W-153-R-7 Mammals Research 3A Pronghorn 5 Detecting Density Dependence Natural Populations Investigations in July 1, 1993 - June 30, 1994 T. M. Shenk Abstract An historical review of methods developed to detect density dependence from temporal trends in natural populations was drafted to highlight similarities, strengths, weaknesses and applications of the various techniques. Past evaluations and criticisms of the methods were included in the review. Attempts to detect density dependence from regression techniques relating either population abundance or segments of population abundance (key-factors) at time t+l to population abundance at time t were criticized for spurious correlations resulting from the non-independence of observations. Techniques developed to address this time-series dependence in population. abundances (e.g., autoregession, and tests for randomness, limitation and attraction) lack power. Tests for direction and significance of curvature in population growth curves show promise as a management tool to determine if harvested populations are above or below a maximum net productivity level. Techniques developed to relate survival or reproduction to population abundances through correlation or logistic regression are valid if conducted on independent parameter estimates to avoid spurious results. Detecting density effects on survival may also be done within the extensive theory developed for parameter estimation from mark-recapture data. Testing for compensatory mortality as a specific form of density dependence has served as a management tool for harvesting of waterfowl. For all tests, power increases with increasing sample size, strength of density-dependent response, and intrinsic growth rate. ��139 DETECTING DENSITY DEPENDENCE IN NATURAL POPULATIONS Tanya M. Shenk P. N. OBJECTIVE Evaluate the feasibility of detecting density populations through Monte Carlo simulation. SEGMENT dependence in natural OBJECTIVES 1. Prepare first draft of paper evaluating Bulmer's, Pollard et al.'s, Dennis and Taper's tests for detecting density dependence. 2. Complete first draft of review paper dependence in natural populations. 3. Begin to prepare meetings. 4. Complete annual on tests the 2 oral presentations report for detecting for the INTECOL to the CDOW, due July and density and ICBE 31, 1994. STUDY AREA The study will take place in the Department of Fishery Colorado State University, Fort Collins, Colorado. and Wildlife Biology at METHODS DRAFT OF PAPER EVALUATING BULMER'S, POLLARD TESTS FOR DETECTING DENSITY DEPENDENCE ET AL.'S, AND DENNIS AND TAPER'S Simulations were conducted to evaluate the robustness of Bulmer's, Pollard et al.'s, and Dennis and Taper's tests for detecting density dependence from series of population abundances to the addition of sampling variance. Population abundances were generated from random walk, density-independent, and density-dependent growth models. Sampling variances were distributed as lognormal to mimc the skewed, non-negative nature common to true sampling variance when estimating abundance. Sampling variance was added to population abundances generated from random walk models (the null hypothesis for all 3 tests) with coefficients of variation ranging from 0.00 to 0.01. Series of population abundances were then generated from 6, deterministic densityindependent, constant growth models where growth ranged from R=O.OO to R=0.7S. Analogous simulations were also conducted for populations generated from a density-dependent growth model to estimate power of the 3 tests. Results of the simulations were summarized. DRAFT OF A REVIEW ..POPULATIONS PAPER ON TESTS FOR DETECTING DENSITY DEPENDENCE IN NATURAL An exten~ive literature review was conducted to locate tests to detect density dependence from temporal trends in natural populations. An historical review was drafted to highlight similarities, strengths, weaknesses, and applications of the various techniques. Past evaluations and criticisms of the methods were included in the review. Future research directions-were suggested based on weaknesses of current tests. RESULTS DRAFT OF PAPER EVALUATING BULMER'S, POLLARD TESTS FOR DETECTING DENSITY DEPENDENCE ET AL.'S, AND DENNIS AND TAPER'S �140 Simulations were completed to quantitatively assess the Bulmer's, Pollard et al.'s, and Dennis and Tapers tests variance. Preliminary results were reported in my June Report. The draft of the manuscript, however, is still will not be included in this annual report. ORAL PRESENTATIONS FOR THE INTECOL AND ICSE MEETINGS Oral presentations are in preparation. ANNUAL REPORT Annual report to the CDOW was completed DRAFT OF A REVIEW POPULATIONS on July 31, 1994. PAPER ON TESTS FOR DETECTING The following is a draft of the a review papwer density dependnece in natural populations. DETECTING DENSITY DEPENDENCE robustness of to effects of sampling 1994 CDOW Quarterly in preparation and FROM TEMPORAL DENSITY DEPENDENCE on methods TRENDS IN NATURAL for detecting IN NATURAL POPULATIONS INTRODUCTION The concept of density-dependent population growth is central to population dynamics, especially processes such as compensatory mortality, competition, predation, natural selection, management of harvested populations, and conservation biology. The general acceptance of densitydependent regulation of population abundance has not, however, been paralleled by detection of such regulation from temporal trends in natural populations. An historical overview of the development of the theory of density-dependent population growth, and consequent attempts to test for its existence in nature, may shed light on this discrepancy as well as suggest more fruitful directions for future attempts at detection. The initial controversy on the regulation of natural populations was a primary force that resulted in the development of ecology as a formal discipline (Taylor 1994). Some early attempts to predict population growth stemmed from specific concerns about human population growth and economically important species (see Cole 1957). However, before the 18th century no formulation of any general concepts other than the balance of nature was inferred from studies on population growth (McIntosh 1985). Malthus (1798) in his 'Essay on the Principle of Population' recognized that population abundances were constrained within bounds set by available space and resources required by individuals. Implications of this essay stimulated both Charles Darwin and Alfred Russell Wallace to formulate their concepts of natural selection and the idea that environment limits populations. Thus, recognition of the discrepancy between the capacity and actual increase in population abundances led to the concept of population regulation. Although the concept of population regulation existed, Howard and Fiske (1911) were the first to propose the role of density to explain changes in .... population abundance. In.their work on the. control of pest insect populations '. by parasites they distinguished between 'facultative' .and 'catastrophic '.. mortality factors. Catastrophic factors, such as storms, were defined as those that destroyed a fixed percentage of a population without regard to population density. Facultative factors, such as predators or parasites, destroy a greater proportion of individuals within a population as the population increases in abundance. Formal recognition of these 'facultative' factors later become known as the theory of density-dependent population regulation. Replacement of the terms catastrophic and facultative with 'densityindependent' and 'density-dependent', respectively, followed the work of Smith (1935). Parallel to the development of the theory of density-dependent popUlation regulation, mathematical descriptions of population growth with respect to some equilibrium density were suggested. The most successful �141 attempt, the logistic equation, was proposed by Verhulst in 1838 (Kingsland 1985). However, the equation was not widely acknowledged at the time. Later, Pearl (1927) generated and successfully promoted the same population growth equation. Several laboratory experiments supported the logistic equation as an accurate predictor of population growth (Pearl and Parker 1922, Gause 1934). Field data, however, were equivocal in their support of density-dependent population growth. Lack (1954) provided evidence of logistic population growth in many bird populations. Andrewartha and Birch (1954) however, concluded population growth of Australian thrips was more related to weather than to density. They argued that most populations, insects and small invertebrates in particular, are influenced primarily by density-independent factors. strong opposition followed the claims made by Andrewartha and Birch. Nicholson (1954), in particular argued that within a balance .. of _nature framework, population densities continually move toward a stable level in relation to fluctuating environmental conditions. He suggested a regulating mechanism limits population abundance, operating with greater severity as population levels more closely approach the stable level. Nicholson (1954) proposed that inter-specific competition was the density-dependent controlling factor. The Cold Spring Harbour Symposium on Quantitative Biology held in 1957 addressed the controversy over population regulation. A general acceptance of the existence of density-dependent regulating factors followed and interest turned toward the detection of density-dependence in natural populations. An early, and continuing, approach to the detection of density dependence is the analysis of long-term rife-table data that provide estimates of temporal trends in the density of different life-history stages (e.g., Varley and Gradwell 1960, Bulmer 1975, Royama 1977, Slade 1977, Berryman 1978, Vickery and Nudds 1984, Pollard et al. 1987, Reddinguis and den Boer 1989, Turchin 1990, Holyoak and Crowley 1993, Dennis and Taper 1994). However, closer examination of tests developed to analyze temporal trends in population density suggest they may not be sensitive enough to allow for detection of density dependence (see Strong 1986, Gaston and Lawton 1987, Hassell et al. 1987, Lomnicki 1987, Mountford 1988, Solow 1990, Bartmann et al. 1992, Holyoak 1993). The following review describes methods used to detect density dependence from temporal trends in either population abundances or specific life history parameters such as survival or reproduction. The focus of this paper is to highlight the similarities, strengths, weaknesses and applications of the various techniques. Summarizing the curr~ntly availabl~ techniques we suggest the more fruitful avenues for future research. We address only those tests applicable to observational data. Therefore, whatever the apparent strengths of the techniques, inference cannot be made concerning any cause and effect relationships resulting from population density. To truly address the question of density-dependent effects operating within a population requires experimental manipulation of the population with detailed information on demographic parameters (e.g., see Bartmann et al. 1992). ANALYSIS OF POPULATION DENSITIES Several techniques have been proposed for detecting the presence of density-dependent processes from a series of sequential population estimates •.. :'-. The. logic underlying these· tests relies -on the· following population ,growth s.;; .. models. Density-independent population growth can be generated from the following 4 growth models. stationary population growth can be defined as: .. . c __ •• Xt.l • x; (1) where Xt is the natural log of the population density (Nt) at time t. If populations follow random walk growth through time then: (2) ~ .--- �142 where Xt is defined as above and €t is a random normal variable with mean zero and variance if. Random walk growth with a trend can be described as: Xe•1 • i: where Xt and €t are defined results when: where + Xe + ee (3) as above and r is a constant. Constant growth and. rare defined as above Density-dependent growth can be defined from any number of models (e.g., logistic, Gompertz, Ricker equation). Most tests developed to detect density dependence agree on the above models for describing density-independent growth. Primary differences in the tests stemmed from the form of the density-dependent growth model, and also the estimation and evaluation of the test criterion used to distinguish between the various models of densityindependent growth and their specific model of density-dependent growth. As initial tests were presented and evaluated, later tests were developed to address biases discovered in these early attempts. Following is an overview of tests used to detect density dependence from series of population abundances and their specific modifications to evaluate population growth. Xt e, Regression techniques .Several early tests developed simple direct density dependence: where as Xt, r, and €t are defined for detecting density dependence as above and ~ is a constant. (13 "* 1) relied Rewriting on Eq.5 (6) shows more clearly how Xt+l is dependent on Xt• Initial efforts to test the random walk model (Eq. 2) against the alternative of direct density dependence (Eq. 5) were based on the assumption that the regression coefficient, b, of ~+1 on Xt was a reasonable estimate of ~. Test of the hypothesis ~ = 1 against the alternative ~ ~ 1 then distinguished between density-independent and density-dependent growth. If ~ < 1 the population is negatively density dependent, thus the relative growth rate decreases as the population increases. If ~ > 1 the relative growth rate increases as the population size increases, the population is positively density-dependent. The latter. case is typically of .less concern and the ..., alternative hypothesis if often· stated· as 13 <.1. ..~ . . Morris (1959) used such a regression approach but further refined the response variable (Xt+1) by categorizing mortali~y factors within a given population as either those that cause a relatively constant mortality from year to year and contribute little to population variation or those that cause a variable mortality .and appear largely responsible for the observed changes in population. Mortality factors falling in the latter category are called 'key factors' because they contribute most to perturbation of densities away from population equilibria. To determine key-factors and density-dependent effects, data must include population size estimates for n years and annual mortality partitioned into m factors. Sources of mortality are expressed in terms of killing power, k-values, numerically equivalent to the difference between the log of the •• "0: ••• �143 population density at time t (Xt) and the log of the population time t after a designated mortality factor has operated (Xt+d: k-value • Xt•1 - Xt density at (7) The total generational mortality is referred to as K, and individual sources of mortality are designated ko, k1, ••• , km. The key factor is that whose variation makes the greatest contribution (highest correlation coefficient)to total mortality variation.. A slope of 1 in the regression of a key factor and the log of population abundance is to be interpreted as complete density independence. Deviations from 1 in either direction are a measure of density response in a population. Varley and Gradwell (1960) addressed the difficulty of identifying keyfactors in a population (see Royama 1981) by using a least-squares regression of all the k-values (Xt+l- Xt) on either Xt or Nt. F-tests are used to determine significance for a given a-level. The slope is then tested against a slope of zero using a Student's t-test (errors in the ordinate are assumed to be normally distributed). A significant slope indicates density dependence. Varley and Gradwell (1963) also addressed the bias of the simple regression coefficient, b, as an estimate of ~ in the presence of random variation. Recognizing that population abundances are estimated and thus subject to sampling error, density dependence was modeled as: Xt•1 • r + 13 [Xt - Yt-1] + Yt (8) where ~, r, and ~ are defined as above and Yt is a random normal variable with mean zero and variance 02• Equation 8 simulates population growth with only measurement errors. TheYfollowing model simulates both environmental and measurement error: (9) where .., all parameters are defined as above. Assuming both sources of error, Varley and Gradwell (1963) recommended regressing Xt+l against Xt and vice versa, then rejecting the hypothesis of 13 = 1 (density independence) only when both calculated regression slopes are significantly different from 1 in the same direction. The combined confidence intervals around bX(t+l)'X(t) and 1/bx(t)'x(t+1) provide the most conservative (widest) confidence intervals for the true slopes of ~+1 regressed against Xt when error occurs in either or both variables. Power of the test, however, has been shown to be low (Slade 1977, Vickery and Nudds 1984). Like Varley and Gradwell, Slade (1977) modified Morris's (1959) keyfactor regression analysis to address the bias in the value of the regression _coefficient ..(b) when. random .variation occurs in the data. Slade. (1977) c proposed the ,use of ,the slope. of the principal·· axis· (l3p) of -CL.bivariate. scatter plot of Xt and ~+1 instead of the regression coefficient (b) as an estimate of 13 in Eq. 5. The slope of the principal axis should be less sensitive to random variation in the data when the Xt's have equal variance. As in regression analysis, if I3p = 1 then there is no density dependence, if I3p < 1 then there is evidence of density dependence in the population. Solomon (1964) simplified Morris's (1959) key-factor analysis to detect density dependence in natural populations by regressing the log of the total population size at time t+l (~+1) on the log population size at time t (~). Thus, diffirent sources of mortality (key-factors) were not isolated. Again, density independence exists if the plot yields a straight line with a slope (13) of 1, indicating per capita population growth rate (r) is constant. If negative density dependence is occurring in the population the plot should '_ .•.... , ,'. -- .... �144 result in a line with 0 < 1. Autoregression techniques All tests developed to detect density dependence based on regressing Xt+1 on Xt were criticized for their statistical properties. Maelzer (1970) pointed out the bias and subsequent spurious correlations resulting from the non-independence of observations of k-factors and population abundance. Eberhardt (1970) also illustrated the dangers of inferring causation, the existence of density dependence, from spurious correlation. st. Amant (1970) determined Type II error rates increased when random variation (defined above as et) occurred in population densities of unregulated populations and when sample sizes (number of years of population densities available) were small. Kuno (1971), .Ito (1972).,.Benson (1973), and Bulmer (1975) suggested this bias was a result of choosing the simple linear regression coefficient to estimate 0,- the coefficient of ~ in equation 5. When 0 = 1, b has a strong negative bias and an intractable sampling distribution because b is nearly equivalent to ru the first serial correlation coefficient that cannot exceed 1 (Bulmer 1975). Bulmer (1975) addressed this negative bias by approaching the problem as a time series analysis and proposed two autoregression methods for the detection of density dependence. Density independence is defined as: (10) identical to Eg. 2. However, density-dependent growth was defined as a time series and modeled as discrete time, stochastic Gompertz first-order population growth: (11) where Xt, et, and r are defined as above, and b is a constant. Densitydependent growth is then defined when b ~ O. Dennis and Taper (1994) point out a statistical motivation for use of this model is that it can be written in the form of a linear, first-order autoregressive model: x t.l - Jl • tl •••(X t - p) + e t: (12) where Xt and et are defined as above, and 0 = 1+b and p = -rib where rand b are the constants defined in Eg.11, p is an equilibrium value. Thus, for a density-independent population there will be no deterministic trend in population size, nor will there be any tendency for it to return to an equilibrium value. The population will follow a random walk. Alternatively, in the density-dependent model ·population size will constantly return to the equilibrium value p. The models are then used to test the null _~"hypothesis .of.density.independence against the alternate model of density .,_0- dependence ·based .on observations Nlt· .•.•• Nn• ;·The.. test.··statistic0 is: R· V/U (13) where: (14) �145 (15) x··tx/n tol (16) The ratio R is almost equivalent to using the first serial correlation c.oefficient, rl (R = [1.- N-1]/2[1 - rt), the difference being only in the weights given to the first and last observations (Bulmer 1975). The null hypothesis is rejected for small values of R. Distribution of R was determined from computer simulation and approximations of the upper and lower boundaries for assessing the size of R are given by 1 4 RL• + (N - 2) XL (17) where values of XL and Xu are given in Bulmer (1975). Bulmer (1975) also proposed a second test when measurement error of population densities is considered. The test statistic, R', is as follows: (18) R* • W/V where V is defined in Eq. 15 and W • ~ (Xt.2 e-i - Xe•l) (Xt - X·) (19) where Xt and x' are def ined above and N is the number of population abundances. Small values of R' (critical values given in Bulmer [1975]) support the hypothesis of density dependence. Randomization techniques An alternative to viewing density dependence as population abundances achieving and then fluctuating around a single equilibrium point (Eq. 12) is _-.. to consider .. the properties .. of. limitation .. (i.e.•.. boundedness) and attraction in --_.a .series of abundances •..,Limitation .La defined as--the,.propensity for a population to remain within narrow limits for long periods (Reddinguis and Den Boer 1989). Attraction is defined as a restricted interval of population densities (the attractor) toward which density tends to shift from generation to generation (Crowley 1992). Testing for either limitation or attraction eliminates the need to depend on any equilibrium-centered theory (see Wiens 1977, Schoener 1982). Two predictions of the limitation concept of population abundances are: (1) densities nearest the limits (boundaries) should be most strongly repelled, and (2) the observed density range over the whole sequence should be conspicuously lower than could otherwise be achieved (Crowley 1992). Within this framework of limitation, Reddinguis (1971) developed a permutation test that may be used as a nonparametric alternative to Bulmer's test. The method �146 tests the hypothesis "the bounds between which the population fluctuates are, or are not, significantly narrower than one would expect given the circumstances 'X'" (Wolda 1989) and proceeds as follows. The permutation test uses the rank of the log-range [LR = log (highest density) - log(lowest density)] of the field data within the collection of log-ranges obtained from all possible permutations of population abundance sequences derived from the field data (Reddinguis and den Boer 1989). Thus, for density dependence the actual sequence of abundances can be expected to have a lower log-range than in a significant majority of the permuted series. To perform this test for limitation let N1, Nz,.•• ,Nn denote population densities at equidistant times t = 1, 2, •.. ,n. Let Xt = 1nNt, t=1, 2, ... ,n, and let d, = Xt+l - Xt for t = 1,2, •••n-1. The null hypothesis of the permutation test states that the order in which the d's occur is random, the alternative hypothesis is.that the d-values occur in such an order that the resulting fluctuations of the X-sequences are restricted (Reddinguis and Den Boer 1989). To test the null hypothesis a random sample of all possible permutations ([n-1]1, where n is the number of observed abundances) of the observed d's is taken, and the X's and LR's computed (see Table 1). If the observed log-range (LRoBS) is a random drawing from the same population as the sample of LR1,... k'S (where k = number of permutations) obtained by permutations then the null hypothesis is not rejected. Let the observed LR be denoted as LOBH and the first, second, ••• , k+1~t order statistic of the combined set of LR' s be L1, Lz, .•. ,Lk+1• Let r be the rank of LOBS in the sequence of Li's (i=1,2, •••k+1); then if r/(k+1) ~ a, where a is the Type I error rate, the null hypothesis is rejected (Reddinguis and den Boer 1989). Note that r is the number of LR' s in the sample permutations that are not greater than LOBS' Table 1. Comparison of observed and 2 examples of permuted log population abundances (Xt's) and resulting observed differences in sequential abundances (dt's). Log-range (LR) values are computed as LR=[log(highest density )log(lowest density)] for each set of sequential abundances. LR-values are used to conduct the limitation test for detecting density dependence presented by Reddinguis and Den Boer (1989). Permutation Observed dt's (Xt+1- Xt) Randomized ~'s 1 Permutation Resulting Xt's (Xt+1- (Xt+l-Xt) 6.0 Randomized dt's 2 Resulting Xt's Xt) 6.0 6.0 6.2 0.2 -0.2 5.8 -0.2 5.8 6.4 0.2 0.1 5.9 -0.1 5.7 6.3 -0.1 0.0 5.9 0.0 5.7 6.4 0.1 -0.1 5.8 0.1 5.8 6.3 -0.1 -0.1 5.7 0.2 6.0 6.4 0.1 0.2 5.9 0.1 6.1 6.4 0.0 0.2 6.1 -0.1 6.0 6.2 -0.2 0.1 6.2 0.2 6.2 LRoBS = (6 •4- 6.0 )=0 •4 LRl =(6.2-5.7)=0.5 LRz =(6.2-5.7)=0.5 �147 Both Bulmer's test and that of Reddinguis and Den Boer (1989) are heavily influenced by temporal trends in population abundances (Holyoak 1993). Unless a satisfactory method of detrending the data is available these tests have limited usefulness and are restricted to detecting density dependence when abundances remain relatively constant. Thus, Pollard et ale (1987) developed a test robust to temporal trends in population abundances. Pollard et ale (1987) proposed a distribution-free randomization method to test if an observed set of Xt'S (Xt = log Nt (t=1,2, ••• ,n)) are random. The test is similar to that of Reddinguis (1971) and Reddinguis and Den Boer (1989), the only difference being the test statistic is not based on the logarithmic range (LR = x",ax - Xmn). To test the null hypothesis of a random sequence of Xt's (density independence) the observed values Xl' X2, ••• ,Xn are first used to compute the value of the test statistic T23, an appropriate likelihood ratio test-statistic. The likelihood ratio test-statistic is based on the 3 nested time series models used by Pollard et ale (1987) to define the density independence and density dependence hypotheses. The models are as follows: Model 1 (20 ) Model 2 (21) and Model 3 (22) where all parameters have been defined above. Modell, a random walk, is identical (see Eq. 10) to the density independence null hypothesis proposed by Bulmer (1975). Model 2, also density independent, is a random walk with trend. The last, Model 3, defines density-dependent growth. Model 3 can be rewritten as (23) where b=~-l and all other parameters are as defined above. This is the same discrete time, stochastic Gompertz first-order population growth model (see Eq. 11) used by Bulmer (1975). Under the proposed randomization method, the likelihood ratio teststatistic to select the best fit between Model 1 and Model 3 (T13) and Model 2 and Model 3 (T23) give identical results. Thus, to test for density dependence (Modell or 2 against Model 3), the likelihood ratio test-statistic is (24) ~ ~ t.l (X ~l - X t ) 2 - (Xn - X I) 2/ (n - 1) where (25) �1.+8 (26) b • (27 ) Xt is defined as above and n is the number of·observations. The sampling distribution of T23 is not tractable mathmatically, but Pollard et ale (1987) show that T23 is equivalent to using (1-r~) where rdxis the correlation coefficient between the dt and Xt values, with d • (Xe•l - X ) • Then to compute the distribution of T23 under the nuli hypothesis, calculate the dt = (Xt+1 - Xt) values and either enumerate the (n-1) 1 possible sets of Xt values, or randomly permute the dt values, and corresponding to N such permutations, obtain a sample of n simulated sets of Xt values (as in Table 1 for the limitation test of Reddinguis and Den Boer (1989». For each simulated set of Xt values the test statistic T23 is computed. If less than 5% of the T23 values calculated under the simulated Xt sequences are smaller than or equal to the computed T23 value under the observed Xt values, the null hypothesis of density independence is rejected at 0=5% level of significance. Crowley (1992) approached the problem of detecting density dependence by testing for attraction, defined when population abundances are pulled towards a hypothetical attractor band. The attraction approach predicts density' shifts are toward the attractor, whether they overshoot, undershoot, or just 'reach the attractor or remain within the attractor band. The following procedure for attempting to detect attraction in a temporal sequence of population densities was developed by Crowley (1992). The test statistic is the 'minimum violation number.' To obtain the test statistic, the n observed abundances are first sorted into ascending order, independent of the temporal sequence. These abundances are then used as n+1 density intervals. The n-l density increments between the sorted density observations were each calculated as In(Ni+l+l)-ln(Ni+1) for changes in density between abundances i and i+1. Each of the n+l density intervals represent trial attractors. The violation number is obtained by counting the number of observed abundances in the time-series that were outside the range of abundances encompassed by the attractor for each of the n+1 density intervals. The trial attractor that gave the smallest violation number is then the minimum violation number. Two methods were used to interpret the violation-number results. The first, the random-walk attraction test is based on computer-simulated density sequences generated from the following equation: (28) " where Xt is defined as above, et is the environmental log-density perturbation immediately preceding time t, Xo is the log-density of the attractor at time t, Yt is the sampling error in estimating log-density at time t, and k is the dimensionless resilience coefficient, a measure of linear density-dependence. To determine the critical violation number for the test, 25,000 randomwalk density sequences of the same length as the data sequence were generated and violation numbers determined for each. From these violation numbers, the critical violation number was designated as the value equivalent to the 0smallest (i.e. the 251st smallest if 0 = 0.01). The second method to evaluate the violation-number results was a ;',. �149 randomized attraction test. Randomization was used to generate the frequency distribution of violation numbers, as in Pollard et al. (1987) and Reddinguis and den Boer (1989) described above. The criterion for detecting density dependence is to tabulate shifts in sequential densities inconsistent with the location of the apparent attractor. In the same spirit of randomization techniques used by Reddinguis (1971), Reddinguis and Den Boer (1989), and Pollard et al. (1987), Dennis and Taper (1994) present a likelihood ratio test based on a discrete time stochastic logistic model to detect density dependence in time series observations of population abundances. The difference in tests being primarily concerned with the model of density-dependent growth against which to test the null hypothesis of density-independent growth. Density-dependent growth can be projected by numerous models. Reddinguis (1971), Bulmer (1975), and Pollard et al. (1987) used the Gompertz first-order population growth model (see Eq. 11). Dennis and Taper (1994), however, use the following model to relate Nt+1 to Nt: (29 ) where Nt is population abundance at time t and all other parameters are as defined above. However, this stochastic logistic model becomes a first-order nonlinear autoregression model when transformed to a logarithmic scale. By letting Xt = ln Nt, (30) where all parameters are as defined above. The null hypothesis of Dennis and Taper's (1994) test is the population is undergoing random walk (Eq. 2), stochastic exponential growth or stochastic exponential decline (Eq. 3). Thus, the parametric bootstrap likelihood ratio test (PBLR) distinguishes 3 cases of the stochastic logistic equation: Modell as a random walk (r=b=O), Model 2 as a stochastic exponential or decay (r=O, b~O), and Model 3 (Eq. 30, r~O, b~O) as density-dependence. These cases form a series of 3 nested hypotheses. Therefore, to test for Modell vs. ~odel 2 the likelihood ratio statistic (T12) is identical to the t-statistic used for testing whether the slope parameter (b) is nonzero in a linear regression of X 1 - Xt • r + b exp (X) + €t. However, this T12 does t not have a student's t d~stribution due .to the time-dependence of the observations. Instead the distribution of T12 is estimated from the data through parametric bootstrapping (Dennis and Taper 1994). If Model 3 is the model that best fits the data, the alternative hypothesis of density dependence is accepted. Testing for delayed density-dependent response Density-dependent responses that affect population abundance may occur following a time lag. Most tests described above can be modified to test for a delayed response. For example, Turchin (1990) extended Varley and . _...._"Gradwell' S" single regression test;·to .. ·testing, for delayed density dependence by .- .. ' ~,".2...different .methods. ··The.first method is. identical. to Varley' and Gradwell's test except that Turchin (1990) regresses the k-value (In(Nt+l/Nd) against Xt-1 while partialling out Xt• Thus, (31) where Xt and Bt are defined as above and ao•••ap are weights that quantify the influence of past densities on the population change. This approach allows for multiple lag effects but requires a linear relationship among the logtransformed population densities. The second approach by Turchin (1990) specifies an extension (considers 2 time lags) of the same non-linear model for population density as developed �150 by Ricker (1954) and used by Dennis Nt.l - Nt exp[rO and Taper + Ci.1Ne + (1994). Ci.2Ne_1 + ee] The model is: (32) where the parameters of the model ro, aI' az, and €t are estimated by regressing the rate of population change r = log(Nt+dNt) on Nt and Nt-I. Each series of population censuses are analyzed using a stepwise regression: first regressing r on Nt, and then testing whether adding the term Nt-1 significantly reduced the unexplained variance (Turchin 1990). However, the first regression is an induced correlation because of the common estimate of Nt (see Au~oregression ~echniques above). Both regressions suffer from increased Type II error rates due to environmental variation (€t) in population abundances (St. Amant 1970). Dynamic response analysis Dynamic response analysis was developed as a management tool for harvested populations at the Southwest Fisheries Center, National Marine Fisheries Service, USA (see Goodman 1988). The method is used to determine whether a population is above or below its maximum net productivity level (MNPL). A population below MNPL exhibits exponential (i.e. densityindependent, Eq. 3) growth and if above MNPL the population has a decelerating growth rate (i.e. density-dependent, Eq. 29). The dynamic response method depends on using series of population sizes (or indices of population size) to trace out a segment of the population growth curve. If the population is assumed to follow logistic growth (Eq. 29), in a plot of population densities by time, early stages of population growth are displayed with the curve concave upwards, beyond the inflection point the curve becomes concave downward. DeMaster et al.(1982) proposed fitting a second-order polynomial to .the trend data and assumed the population to be below maximum net productivity level (MNPL) if the second-order term is positive (curve concave upwards) and above MNPL if the second-order term is negative (curve concave downwards). Gerrodette (1988) formalized the test procedure by fitting a simple linear regression to the trend data. If the slope is negative or not significantly different from zero, assume the population is above MNPL. If the regression slope is significant and positive, assume the population is below MNPL; if it is not significant, assume the population is "near" MNPL. Boveng et al. (1988) modified dynamic response analysis to use a series of fits, using first 4 of the available observations, then 5, and so on up to a possible maximum or the number of points available. Each such sample size was used in a moving-interval fashion. That is, the first four observations were fitted to a second-order polynomial, then the first point was dropped and the next point in the observed series included, a new fit calculated, and so on until all of the observed series of points had been used. The same procedure was used for 5 points, and so on, using an increasing number of data points in each fitting. The "correct" number of points to use is determined by the criterion that the second-order coefficient changes sign only once as the interval is moved along the data set, or does not change sign at all so that .the entire set is presumably above or below MSY. After the correct sample size is determined a· linear regression is computed and the direction· and significance of curvature is evaluated as in Gerrodette's test. Eberhardt (1992) modified Gerrodette's (1988) method of implementing dynamic response by using the natural logarithms of the observed population size values. A simple linear regression and a second-order polynomial are then fitted to these transformed data. An F-test is used to test if the second-order coefficient is different from zero. Following an evaluation of all 3 methods of implementing the dynamic response method Eberhardt (1992) combined the Gerrodette (1988) method with his own to provide a test with better performance. The combination was a series of F-tests: use the F-test for significance on data on the arithmetic scale. If this was non-significant, then the natural logarithms of the population trend data were used in the same test. An F-test was used to test �151 the linear regression stronger curvature. ANALYSIS OF BIRTH, against DEATH, a third-order IMMIGRATION, polynomial AND EMIGRATION to better test for RATES Use of total population abundance, or even abundances of specific age classes, as the response variable to detect density effects in a population addresses the problem on a fairly course level. Fluctuations in population abundance result from density-dependent effects on birth, death, immigration, and/or emigration rates. Therefore, attempting to detect density dependence from these life history parameters may prove more fruitful. The following techniques are examples of such attempts. Correlation A significant correlation between a given demographic parameter (e.g., fecundity, survival) and population density supports the hypothesis of density-dependent growth. Correlations of this sort must however, be conducted on independent parameter estimates. Eberhardt (1970) and Anderson and Burnham (1976, 1981) caution against 'forced' correlations, a result of using 2 demographic parameters estimated from the same data using a common sampled variable. Be.cause the variables are estimated, results of the correlation between the demographic parameters force an apparent relationship based on the additional sampling correlation. The resulting correlation is not a valid estimation of the relationship between the 2 parameters. Testing for compensatory mortality Anderson and Burnham (1976) and Burnham and Anderson (1984) investigated expressed as compensatory mortality in a specific form of density-dependence They relate survival and kill rates as hunted mallard (Anas pla~yrhynchos). follows: (33) where Si is the annual survival rate in year t, So is annual survival rate in the absence of hunting, ~ is the annual kill rate in year t, and 0 ~ b ~ 1 (Burnham and Anderson 1984). If b = 0 complete compensation is represented, if b = 1 complete additivity is represented, if 0 < b < 1 then an intermediate relationship exists between survival and kill rate. Estimation of survival and kill rates are based on fitting band recovery models representing competing hypotheses and testing between these models (see Anderson 1975, Brownie et ale 1985). The slope parameter, b, is defined as: b.~bs,K/SO (34) where 0 ~ b ~ 1 and bs,K is the slope of the linear relationship between annual survival. rate and annual kill rate (see Burnham and Anderson 1984 and Anderson and Burnham 1976: Appendix A). Estimation of b from simple regression is not _.' -. ,.. feasible. because survival. and kill rate estimates .are obtained from the same ._ .. ".data, .thus _introducing:.sampling correlation· (see Correla~ion .. above). . . Nichols (1991) reviewed the various techniques to estimate and test hypotheses about b. These techniques include: (1) a variance components approach which . used estimated sampling variances and covariances of survival and kill rates, permitting the estimation of b (Anderson and Burnham 1976); (2) an ultrastructural band recovery model in which survival was modelled as in Eq. 33 and b was estuimated directly (Anderson .et ale 1982); and (3) obtain independent estimates of survival and kill rate for use with standard linear regression techniques. Mark-recapture modeling The technique of mark-capture survival estimation is based on following the fates of individually marked animals through time (see Lebreton et ale �152 1992). Numerous models are structured to estimate survival based on biological and statistical considerations specific to the data. Global models may include potentially explanatory covariates such as age, sex, and/or environmental variables such as rainfall in estimating survival. Density effects on survival can also be included in these models. Parameter estimation is based on maximum likelihood theory, models are subjected to goodness-of-fit tests and model selection is based on a function optimization framework using Akaike's Information Criterion, AIC (see Lebreton et al. 1992). AIC is defined as AIC - -21nL + 2. (number of parameters) (35) where lnL is the log-likelihood of the parameter. The term (.2*number of parameters) is a penalty for addition parameters, thus engaging the 'principle of parsimony', selecting the simplest model that fit the data. An example of how this technique tests for a possible density effect on survival was given by Lebreton et al. (1992) for roe deer (Capreolus capreolus). Data were collected such that sex and age could be tested as covariates effecting survival, as well as density. The general model included age, sex, and density as covariates effecting survival. This general model was tested against a simpler model using only age and sex effects to estimate survival. Because the more general model was selected there was evidence that density, age, and sex all had significant effect on survival rate. Had the simpler model (age and sex only) been selected there would have been no support for density effecting survival rate. Logistic regression Logistic regression converts binary data into probability values by fitting a logistic curve through the available points. Only variables that significantly improve fit are considered to interact with the response variable. Parameters of the logistic model are estimated by maximum likelihood. A goodness-of-fit of a model is determined by comparing the difference between deviance values of 2 models which were distributed approximately as x2 with degrees of freedom equivalent to the differences in number of parameters fitted in each model (Clutton-Brock 1987). Logistic regression may be used with density as a potential variable to significantly improve model fit for estimation of various life history parameters. Fo'r example, Clutton-Brock (1987) used logistic models (see Cox 1970) to investigate how changes in population density interacted with age, reproductive status, dominance rank, and matriline size to affect fecundity or calf survival in a resource limited population of red deer (Cervus elaphus). Selection of a model inclusive of density over a simpler model without density is support for a density-dependent response variable. Modeling techniques Modeling can· be used as a tool to support the existence of densitydependent demographic parameters. Comparison of a specific population parameter generated from a model with built-in density-dependent population dynamics to that same population parameter observed in nature supports the hypothesis of, density dependence operating within that population. For example, DeAngelis (1991) developed an individual-based model with density-dependent feedback processes acting within the population. The model begins with an initial cohort of individuals of a given size distribution. The simulation follows each individual through the series of feedback processes for a given time, varying recruitment with population density. The resulting size distribution of individuals from the model is then compared to the size distributions found in nature of the system being modeled. Comparability of the 2 distributions was used as supporting evidence for the existence of density-dependent compensatory mortality. �153 SUJlMARY To truly answer the question, 'Is density dependence operating in this population?' requires experimental manipulation of the population, coupled with precise estimates of demographic parameters. This review, however, is limited to our current attempts to detect density dependence from temporal trends in natural populations. Thus, the most significant weakness of these techniques is inherent in any analysis of observational data; cause and effect relationships cannot be established. Manipulative experiments however, are not always feasible. Therefore, accepting the limitation of weaker inference, future research is needed to resolve the current controversy surrounding the effectiveness of existing tests and to develop new tests robust to the biases already acknowledged in current methods. Several general conclusions from existing evaluations of these tests can serve as immediate guidelines for future research. These conclusions are as follows: (1) the chance of detecting density dependence increases with increasing number of years of data; (2) power to detect density dependence is decreased if the parameters ·are estimated; and (3) scale at which density dependence operates affects the performance of the tests (i.e. the rate of detection of density dependence is correlated with the value of the autocorrelation coefficient which is heavily influenced by the value of the intrinsic growth rate). Not surprisingly, these results reflect the inherent difficulties of analyzing observational data which are subject to environmental and innate variation as well as sampling error. Therefore, we suggest future research to develop and evaluate new techniques to detect density dependence from temporal trends in natural populations should focus on analyzing the demographic parameters most sensitive to density dependence (e.g. survival, reproduction). Ideally, tests should be designed t9 be robust to sampling error and effectively handle environmental variation as well as innate heterogeneity of demographic parameters. Assumptions of each test should be stated explicitly. Finally, before tests are implemented, evaluation of their robustness to assumption violations should be documented through techniques such as Monte Carlo simulation where the true model is known. LITERATURE CIDD Anderson, D. R. 1975. Population ecology of the mallard. V. Temporal and geographic estimates of survival, recovery and harvest rates. u.s. Fish and Wildlife Service Resource publication, 125, 1-110. Anderson, D. R. and K. P. Burnham. 1976. Population ecology of the mallard. VI. The effect of exploitation on survival. U.S. Fish and Wildlife Service Resource Publication, 128, 1-66. Anderson, D. R. and K. P. Burnham. 1981. Bobwhite population responses to exploitation: two problems. J. Wildl. Manage. 45(4):1052-1054. Anderson, D. R., K. P.Burnham, and G. C. White. 1982. The effect of exploitation on annual survival of mallard ducks: An ultrastructural model. Proceedings International Biometrics Conference 11:33-39. Andrewartha, H. G. and L. C. Birch. 1954. The distribution and abundance of animals. University Chicago Press, Chicago, IL. '... .;"" "Bartmann, R. M., Go' C. White, and L. H. Carpenter •. 1992. Compensatory mortality in a Colorado mule deer .population. Wildlife Monog. 56(1) 39pp. . Benson, J. F. 1973. Some problems of testing for density-dependence in animal populations. Oecologia 13:183-190. Berryman, A. A. 1978. Population cycles of the Douglas-fir tussock moth (Lepidoptera: Lymantriidae); the time-delay hypothesis. Canadian Entomologist 110: 513-518. Boveng,P., D. P. DeMaster,. and B. S. Stewart. 1988. Dynamic response analysis. III. A consistency filter and application to four northern elephant seal colonies. Marine Mammal Science 4:210-222. Brownie, C,. D. R. Anderson, K. P. Burnham, and D. S. Robson. 1985. statistical inference from band recovery data: A handbook, 2nd ed. U.S. Fish and Wildlife Service Resource Publication, 156, 1-305. �154 Bulmer, M. G. 1975. The statistical analysis of density dependence. Biometrics 31:901-911. Burnham, K. P. and D. R. Anderson. 1984. Tests of compensatory vs. additive hypotheses of mortality in mallards. Ecology 65:105-112. Clutton-Brock, T. H., S. D. Albon, and F. E. Guinness. 1987. Interactions between population density and maternal characteristics affecting fecundity and juvenile survival in red deer. Journal of Animal Ecology 56: 857-871. cole, L. C. 1957. Sketches of general and comparative demography. Cold Spring Harbor Symposia on Quantitative Biology 22:1-15. Cox, D. R. 1970. The analysis of binary Data. Methuen, London. Crowley, P. H. 1992. Density dependence, boundedness and attraction: detecting stability in stochastic systems. Oecologia 90:246-254. DeAngelis, D. L., L. Godbout, .and B. J. Shuter. 1991. An individual-based approach to predicting density-dependent dynamics in smallmouth bass populations. Ecological Modelling 57: 91-115. DeMaster, D. P., D. J. Miller, D. Goodman, R. L. DeLong and B. S. Stewart. 1982. Assessment of California sea lion fishery interactions. Pages 253-264 in Transactions of the 47th North American Wildlife and Natural Resources Conference. Wildlife Management Institute, Washington, D.C. Dennis, B. and M. Taper. 1994. Density dependence in time series observations of natural popUlations: estimation and testing. Ecology (in prep). Eberhardt, L. L. 1970. Correlation, regression, and density dependence. Ecology 51(2):306-310. Eberhardt, L. L. 1992. An analysis of procedures for implementing the dynamic response method. Marine Mammal Science 8(3):201-212. Gaston, K. J. and J. H. Lawton. 1987. A test of statistical techniques for detecting density dependence in sequential censuses of animal populations. Oecologia 74:404-410. Gause, G. F. 1934. The struggle for Existence. Williams and Wilkens, Baltimore. Gerrodette, T. 1988. Dynamic response analysis. II. Evaluation of dynamic response analysis in a simulated no-harvest case. Marine Mammal Science 4:196-209. Hassell, M. P., T. R. E. southwood, and P. M. Reader. 1987 •. The dynamics of the viburnum whitefly (Aleuro~rachelus jelinekii): a case study of population regulation. J. of Animal Ecology 56:283-300. Holyoak, M. 1993: New insights into testing for density dependence. Oecologia 93:435-444. Holyoak, M. and P. H. Crowley. 1993. Avoiding erroneously high levels of detection in combinations of semi-independent tests: an application to testing for density dependence. Oecologia 95(1):103. Howard, L. 0., and W. F. Fiske. 1911. The importation into the United States of the parasites of the gipsy moth and the brown-tail moth. U.S. Dept. of Agriculture, Bureau of Entomology, Bulletin No. 91. Reprinted New York: Arno Press (1977). Ita, Y. 1972. On the methods for determining density-dependence by means of regression. Oecologia 10:347-372. Kingsland, S. E. 1985. Modeling Nature. Univ. of Chicago Press, Chicago. 267pp. Kuno, E. 1971. Sampling error as a misleading artifact in "key, factor analysis." Res. Populo Ecol. 13:28-45. Lack, D. 1954. The Natural Regulation of Animal Numbers. Oxford Univ. Press, London. Lebreton, J. D., K. P. Burnham, J. Clobert, and D. R. Anderson. 1992. Modelling survival and testing biological hypotheses using marked animals: a unified approach with case studies. Ecological Monog. 62(1): 67-118. Lomnicki, A. 1987. ,The density dependence debate. TREE 3:76. McIntosh, R. P. 1985. The Background of Ecology: Concept and theory. Cambridge University Press, Cambridge. 383pp. Maelzer, D. A. 1970. The regression of log Nn+1 on log Nn as a test of density dependence: an exercise with computer-constructed densityindependent populations. Ecology 51: 810-822. �155 Malthus, T. R. 1798. An essay on the principle of population. London: Johnson. Morris, R. F. 1959. Single-factor analysis in population dynamics. Ecology 40:580-588. Mountford, M. D. 1988. Population regulation, density dependence, and heterogeneity. J. of Anim. Ecology 57:845-858. Nichols, J. D. 1991. Responses of North American duck populations to exploitation. Pages 498-525 in C. M. Perrins, J. D. Lebreton, and G. J. M. Hirons, eds. Bird Population Studies, Oxford, New York, N.Y. Nicholson, A. J. 1954. An outline of the dynamics of animal populations. Austr. J. Zool. 2:9-65. Pearl. R. 1927. The growth of populations. Quart. Rev. Biol. 2:532-548. Pearl, R. and S. L. Parker. 1922. on the influence of density of population upon the rate of reproduction in Drosophila. Proc. of the Nat. Academy of Sciences. 8:212-218. Pollard, E., K. L. Lakhani, P. Rothery. 1987. The detection of density dependence from a series of annual censuses. Ecology 68: 2046-2055. Reddingius, J. 1971. Gambling for existence. A discussion of some theoretical problems in animal population ecology. Acta Biotheor. 20:1208. Reddingius, J. an P.J. den Boer. 1989. On the stabilization of animal numbers. Problems of testing 2. Confrontation with data from the field. Oecologia 79:143-149. Journal of the Fisheries Research Ricker, W. E. 1954. Stock and recruitment. Board of Canada 11: 559-623. Roya.ma, T. 1977. Population persistence and density-dependence. Ecological Monographs 47:1-35. Royama, T. 1981. Evaluation of mortality factors in insect life table analysis. Ecological Monographs 5(14):495-505. St. Amant, J. L. S. 1970. The detection of regulation in animal numbers. Ecology 51:823-828. Salt, G. W. 1966. An examination of logarithmic regression as a measure of population density response. Ecology 47(6):1035-1039. Schoener, T. W. 1982. The contr~versy over interspecific competition. American Scientist 70:586-595. Slade, N. A. 1977. Statistical detection of density dependence from a series of sequential censuses. Ecology 58:1094-1102. Smith, H. S. 1935. The role of biotic factors in the determination of population densities. J. Econ. Entomol. 28:873-898. Solomon, M. E. 1964. Analysis of processes involved in the natural control of insects, pI-58 in J. B. Cragg ed. Advances in ecological research. vol. 2. Academic Press, London. Solow, A. R. 1990. Testing for density dependence: a cautionary note. Oecologia 83:47-49. Strong, D. R. 1986. Density-vague population change. TREE 1:39-42. Taylor, M. 1994. Density-dependent population regulation in black, brown, and polar bears. Int. Conf. Bear Res. and Manage. Monogr. Series No. 3 (in press). Turchin, P. 1990. Rarity of density dependence or population regulation with lags? Nature 344:660-663. Varley, G. C. and G. R. Gradwell. 1960. Key factors in population studies. Journal of Animal Ecology 29: 399-401. Varley, G. C. and G. R. Gradwell. 1963. Predatory insects as density dependent mortality factors. Proc. XVI Int. Congo Zool., Washington, D.C. 1:240. Vickery, W. L. and T. D. Nudds. 1984. Detection of density-dependent effects in annual duck censuses. Ecology 65:96-104. Wiens, J. A. 1977. On competition and variable en~ironments. Amer. Sci. 65:590-597. Wolda, H. 1989. The equilibrium concept and density dependence tests: what does it all mean? Oecologia 81:430-432. Prepared by -~____:_{ 9-r-(111....;...........c.=--.-·;..__;~=-k_ Tanya Shenk Graduate Research Assistant 7 ~r¥-V'""""1~r-~ ��157 Colorado Division Wildlife Research July 1994 of Wildlife Report JOB PROGRESS Colorado State of Project Work No. Plan No. Job No. Period Covered: Authors: REPORT W-153-R-7 Mammals Research 3A Pronghorn Research 6 Pronghorn Winter Wheat Damage Study July 1, 1994 - June 30, 1994 D. C. Strohmeyer, G. C. White, and R. B. Gill Abstract Wildlife managers have responded to winter wheat damage complaints by reducing pronghorn (Antilocapra americana) numbers via hunting and trapping removals. Recent research (Torbit et ale 1993) has suggested pronghorn may not damage winter wheat, implying reducing pronghorn populations may not be necessary. For this suggestion to be true, free-ranging pronghorn must stop foraging on wheat as wheat enters the jointing stage. Our research focuses on evaluating this suggestion. There were several important points learned this year. One, the pronghorn shift in vegetation use from wheat to native range was strong enough to be detectable in small data sets. Two, the last marked-animal relocations on wheat occurred before wheat entered the jointing phenological stage. Three, males appeared to less sensitive to the nutritional differences between wheat and native plants. Four, the timing of the pronghorn shift in vegetation use seemed to correspond to the period when cell contents of wheat exceeds that of native plants. ��159 PRONGHORN WINTER D. C. Strohmeyer, WHEAT DAMAGE G. C. White, STUDY and R. B. Gill P.N. OBJECTIVES Elucidate pronghorn movement the telemetry patterns concerning SEGMENT OBJECTIVES data collected wheat use. 1. Analyze 2. Revise the telemetry study plan concerning winter wheat use of radio-collared, wild pronghorn, then submit it for peer review. 3. Submit 4. Collect a draft of the feeding-trial data for the telemetry METHODS spring winter 1993. study plan for peer review. study plan. AND MATERIALS PRONGHORN MOVEMENTS, SPRING 1993 The study area encompasses the Pawnee National Grassland in Weld County, Colorado. It is at the northern edge of the short-grass steppe region. Vegetation use of marked animals was monitored during spring 1993. Marked animals were visually relocated using aerial and ground telemetry. Vegetation type and marked animal activity were recorded at each relocation. Pronghorn use of vegetation types (native range or winter wheat) was the dependent variable. This discrete, nominally scaled variable was recorded as use versus non-use and was binomially distributed. Time (t), a continuous variable recorded as Julian dates, was limited to February through May. Time was the predictor variable for the logistic regression model: logit(1tN) = 11::) = ~o + ~lt 1 where ~w = probability that pronghorn use wheat, and ~N probability that pronghorn use native range. The null hypothesis was that there were no changes in the vegetation type used by pronghorn over time, Ho: PI = O. The alternative hypothesis was that the logit of the ~w increased linearly over time, Ha: PI > O. Model fit was assessed by examining deviance. = The number of true replications in this analysis was the number of animals. To remove pseudoreplication across time, the above model was used to estimate coefficients for slopes for each animal and the resulting slopes were used in a sign test. The null hypothesis for the sign test was that the slopes are zero, ~: PI = 0, so 1/2 of the coefficients would be expected to be positive, and 1/2 negative. The alternative hypothesis was that pronghorn use of native range would increase with time, Ha: PI < O. �160 RESULTS AND DISCUSSION PRONGHORN MOVEMENTS, 1993 Pronghorn on winter wheat were moved to native range using CDOW aircraft several times during the 1992-93 winter. These hazing operations conditioned pronghorn to run from aircraft and when marked pronghorn were relocated aerially they were often running when seen. It was often unclear whether or not the marked animals were on wheat before they responded to the aircraft. In these instances no vegetation type was recorded, otherwise the data would have erroneously indicated higher use levels of native range then actually occurred. Consequently, only 4 of 13 marked animals had relocations on both winter wheat and native range, with the rest only relocated on native range. All animals that were relocated on both vegetation types used wheat then shifted to native range. The last marked animal relocation on wheat was April 10th. Unmarked pronghorn were observed on wheat on May 4th. (Marked animals were again relocated on wheat on October 20th.) Goodness-of-fit tests for the models did not appear violated; although, the values of deviance/df were 1.6-1.7, suggesting a tendency for overdispersion. A sign test of the slopes was marginally significant (~= 0.0625). In summary, the pronghorn shift in vegetation use from wheat to native range was detectable in this sparse data set. This shift for the marked animals occurred between April lOth and May 5th. Mean date of the period of no difference in pronghorn use of vegetation types (~N = ~w) in the converged regression models was April 4th. Future data collection efforts will need to emphasize getting wheat relocations early in the field season. PRONGHORN ~TRY STUDY For the statement, "pronghorn do not damage winter wheat," to be true, pronghorn must be shown to stop foraging on wheat as wheat enters the jointing stage (Dunphyet al. 1982). Observational studies involving pronghorn and wheat have noted seasonal use patterns (Cole and Wilkins 1958, Hoover et al. 1959, Torbit et al. 1993). Pronghorn use of wheat is heavy in the fall and winter, while light in spring and summer. Monthly aerial surveys in northeastern Colorado found unmarked, free-ranging pronghorn used wheat fields from November through April, then abandoned them by early May. Data from 5 radiocollared pronghorn also support this pattern. Marked pronghorn abandoned wheat fieldS by 19 April 1985 and by 12 May 1986, but the exact timing of the shift in vegetation type was not well documented. Explicit ties between the shift in the vegetation used by pronghorn and winter wheat phenology are also lacking because winter wheat phenology was not reported for these studies. We are also interested in examining mechanisms that stimulate pronghorn to stop foraging on wheat, because we would like to infer that the results of our study may extend beyond northeastern Colorado and include all areas where pronghorn and winter wheat overlap. Combining forage quality differences in wheat versus native plant species with pronghorn preference for high crude protein and high cell content implies pronghorn will use winter wheat until the forage quality of native plants exceeds that of the winter wheat. This hypothesis requires that forage quality of native plants must exceed wheat forage quality as wheat begins to joint. A pronghorn telemetry study plan was There are 2 objectives in this study 1. demonstrate that pronghorn jointing phenological stage; and 2. quantitatively compare the shortgrass prairie forage. submitted for review in January 1994. plan: stop using wheat before'wheat enters the nutritional dynamics of wheat and �161 PRONGHORN MOVEMENTS, 1994 Problems with the quality of aerial relocations were not bad because pronghorn were not intensively hazed from wheat fields during the 1993-94 winter. Relocations from 10 marked animals were collectable during the 1994 field season. The last marked-animal relocation on wheat was April 16th (Table 1). Wheat entered the jointing phenological stage during the week of April 17th. As of May 25th, a few unmarked animals were still seen in wheat fields. Most of these observations occurred in the same wheat field. Of the unmarked groups seen in wheat fields from April 16th through May 27th, males dominated the observations. Only 4 groups of females (15 females total) were observed in wheat fields. Alldredge et al. (1987) indicated cell contents of native plants began to exceed that of wheat during April, while crude protein content of native plants became higher than that of wheat in May. Combining this with the telemetry data suggests that the marked animals may be responding to differences in cell contents. (Chemical analyses for this field season will not be completed until July 1994.) Whatever the mechanism is, males appear less sensitive to it than females which are usually in the last trimester of pregnancy at this time. Table 1. Date of last relocation County, Colorado, 1994. Collar Date of Most Recent Relocation 108 150 March Never 171 270 March Wheat wheat of 10 marked # Wheat Relocations pronghorn, 17 0 4 12 3 16 3 1 3 on wheat 292 March 27 2 18 340 March 13 1 16 550 April 16 3 14 570 April 14 5 19 598 February 2 13 1 10 660 FEEDING-TRIAL March 15 3 Weld Total # of Relocations 2 27 relocated April on winter STUDY As previously stated, we are interested in exam~n~ng potential mechanisms that stimulate pronghorn to stop foraging on wheat, because we want a broad range of inference. We suggest that the mechanism of interest is forage quality which implies pronghorn can respond to differences in forage quality. Purpose of the feeding-trial study is to provide experimental support for the correlational data collected in the Telemetry Study Plan. A study plan was submitted in June 1994 for peer review. The objective in this study plan is to test the hypothesis that pronghorn respond to differences in nutrition by manipulating winter wheat phenology. NEXT QUARTER'S OBJECTIVES There are 2 objectives for next quarter: (1) analyze the 1994 data, and begin growing wheat in the Rangeland and Ecosystem sciences Department greenhouse to verify length of the phenological stages of interest. (2) �162 LITERATURE CITED Alldredge, A. W., S. C. Torbit, J. A. Liewer, and R. B. Gill. 19B7. Pronghorn foraging on winter wheat: final report. Agric. Exp. Stat., Colo. State. Univ., Fort Collins. 75pp. Co+e, G., and B. Wilkins. 1958. The pronghorn antelope - its range use and food habits in central Montana with special reference to wheat. Mont. Fish and Game Dep. Tech. Bull. 2. 39pp. Dunphy, D. J., M. E. McDaniel, and E. C. Holt. 1982. Effect of forage utilization on wheat grain yield. Crop Sci. 22:106-108. Hoover, R. L., C. E. Till, and S. G. Ogilvie. 1959. The antelope Colorado. Colo. Dep. Game and Fish Tech. Bull. 4. 100pp. of Schwartz, C. C. 1977. Pronghorn grazing strategies on the shortgrass prairie, Colorado. Ph.D. Thesis. Colo. State Univ., Fort Collins. 113pp. Torbit, S. C., R. B. Gill, A. W. Alldredge., and J. F. Liewer. 1993. Impacts of pronghorn grazing on winter wheat in Colorado. J. Wildl. Manage. 57:173-181. �163 colorado Division Wildlife Research July 1994 of Wildlife Report JOB PROGRESS REPORT state Project work Colorado of No. W-153-R-7 Plan No. 3A Job No. Period Author: 7 Covered: T.M. July Mammals Research Pronghorn Investigations Experimental Pronghorn Surveys Using Fixedwing Line Transects and Helicopter Quadrats 1, 1993 - June 30, 1994 Pojar Abstract The first of 3 years data were collected to compare pronghorn population density estimates obtained by fixedwing line transect and helicopter quadrat surveys. The line transect estimate, using mean group size of 3.1667, was 6,085 animals with the 90% confidence interval ± 30%. The helicopter quadrat survey estimated the population to be 8,465 ± 31%. Given tests that detected no overcount bias in quadrat surveys (Pojar et al •.in press), it is likely the line transect survey is not meeting the key assumption that all subjects in the first distance interval are being detected. The cost of the fixedwing line transect survey was approximately $277.50 and the helicopter survey cost was $4,162.50. The coordinates of 9 brass caps of the United states cadastral survey were estimated to the nearest hundredth minute from 7.5 minute maps. Navigation with a GPS receiver to these points during a separate helicopter flight differed on average by 22.6 m, SO = 16.5 m, from the map point estimate. ��165 EXPERIMENTAL PRONGHORN SURVEYS USING FIXEDWING HELICOPTER QUADRATS LINE TRANSECTS AND Thomas M. Pojar P .N. OBJECTIVE Compare fixedwing line transect pronghorn density. and helicopter quadrat surveys in estimating SEGMENT OBJECTIVES 1. 2. 3. Compare pronghorn density estimate and precision of fixed-wing transect survey with helicopter quadrat survey. Evaluate consistency of line transect data analysis. Test accuracy of a Global Positioning System (GPS) in locating geographic points. line known STUDY AREA The study area is 452 square miles of sagebrush steppe pronghorn habitat north and west of Craig. It is described in Pojar et ale 1994 (in press). METHODS AND MATERIALS The methods are outlined in the Program Narrative; see Appendix I. For the line transect survey, a Maule N91AR fixedwing aircraft was used. Survey techniques include: 1. 2. 3. 4. 5. 6. Two observers, each counting opposite sides of the aircraft. Each line is offset 65 m from aircraft at 300 ft altitude. The latitude and longitude is recorded at the beginning and end of each transect and at the location of each group of pronghorn observed. Altitude of the aircraft above ground level (AGL) is recorded at each group location with a radar altimeter. Coordinates are recorded in both degrees, minutes to the hundredth of a minute, and northing and easting in meters (UTM coordinates). Distance intervals were as follows: A. The line (see item #2 above) to 25 m B. 25 to 50 m C. 50 to 100 m D. 100 to 200 m All perpendicular distances for pronghorn groups were corrected for altitude with the altimeter reading for that location. If the correction resulted in a group being outside the 200 m limit, the group was not included in the data set. The quadrats were searched using a Bell Jet Ranger helicopter equipped with ARNAV Global Positioning System receiver. The navigation was done by the pilot and one observer who recorded observed groups on a tape recorder. an RESULTS The line transect survey was done on May 18 and 19, 1994 by Western Air Research Incorporated (Driggs, Idaho) with Jeff Madison (CDOW) as primary observer. Methodology followed that described by Johnson et ale 1991. The line transect data set was submitted to 2 experts in line transect analysis. Each person followed their interpretation of the criteria of Buckland et ale (1993) and made independent estimates of pronghorn density variance using options available in the program DISTANCE (Laake 1993). and �166 Pojar et al. 1994 (in press) surveyed this same area by helicopter quadrat and helicopter line transect sampling in 1987. Helicopter line transect surveys may be subject to different biases than fixedwing surveys as discussed by Pojar et al. 1994 (in press). For comparison of possible changes in population density, data from the 1987 surveys are included in Table 1. Table 1. Pronghorn population estimates (N), standard errors (SE), and 90% confidence intervals (CI-)using helicopter quadrats, and helicopter and fixedwing line transects, on 452 square miles of sagebrush steppe habitat north and west of Craig, Colorado. The model used in program DISTANCE (Laake et al. 1993) for line transect estimates was a half-normal with 1 cosine term. CI Method Year 1987 1994 N SE df Lower Upper Helicopter quadrat 6,764 1,061 18 4,924 8,604 Helicopter line transect 9,663 1,898 9 6,185 13,141 Helicopter Quadrat 8,465 1,513 20 5,855 11,075 Fixedwing line transect8 6,085 1,061 23 4,266 7,904 Fixedwing line transectb 6,838 1,197 23 4,786 8,890 Fixedwing line transectC 7,740 1,827 41 4,664 10,816 8 Follows analysis procedure of Pojar et al. 1994 (in press) with the estimate based on mean cluster size (3.1667) out to 200 m. The regression of group size on distance was not significant (t=1.445) b Estimate is based on expected cluster size (3.5590) from regression estimate out to 200 m width, although the regression was not significant (t=1. 445) • C Estimate is based on expected cluster size (4.0280) from regression estimate out to SO m, although the regression was not significant (t=0.110). Two line transect experts analyzed this data set independently using~rogram DISTANCE. Although they selected different models (uniform with 1 polynomial adjustment term and half-normal with no adjustment term) the major difference in their estimates was the result of the group size parameter each used. One chose to use the mean group size of all groups in the 200 m width of the transect (3.1667) because the regression of group size and distance from the line was not significant. This approach yielded an estimate of 6,048 animals. The other expert choose to use the results of the regression group size estimate (3.5578) (although the regression was not significant) and estimated the population at 6,916 animals. Program TRANSAN (Routledge and Fyfe 1992) implements the shape-restricted. estimator of Johnson and Routledge (198.5). The shape-restricted estimator is purported to be less prone to bias and instability than other estimators when data do not fit the function form closely. The observational units were groups of pronghorn in this survey and the program is capable of analysis only by equal sized intervals, therefore, the program analyzed frequency counts by groups in 2 intervals, 0-100 m and 100-200 m. I converted estimated group density from program results to total animals by multiplying by average group size (3.1667). The population estimate using this approach was 5,716. During a helicopter flight separate from the survey flight, 9 brass caps from the United States cadastral survey were located by Global Positioning System (GPS) navigation. The coordinates for these points, to the nearest hundredth minute, were obtained from USGS 7.5 minute maps. Mean distance from the helicopter location and the brass cap was 22.6 m (SD = 16.5 m). �167 Flight time to do the fixedwing line transect survey was approximately 1.5 hrs at S185.00 per hour (S277.50). Time to complete the quadrat survey by helicopter was 9.25 hrs at S450 per hour (S4,162.50). REFERENCES CITED 1992. TRANSAN: Line transect Wi1dl. Soc. Bull. 20:455-456. estimates ��169 APPENDIX I PROGRAM NARRATIVE PRONGHORN INVESTIGATIONS state: Colorado T. M. Pojar Project Title: Project Number: Experimental pronghorn surveys using transects and helicopter quadrats. Work Plan 3Ar Job 7 March 16, 1994 fixedwing line Transect surveys are popular because they make efficient use of air time in terms of surveying geographic area. Fixedwing aircraft are most commonly used for transect surveys because they are usually available to biologists and they are more economical to operate than helicopters. Conventional strip transect surveys attempt to count all pronghorn on contiguous strip transects that cover 100% of the area (Springer 1950, Hailey 1979, Gill et ale 1983, and Allen and Samuelson 1987). The total number observed is then taken as the total for that population with no adjustments for detection errors, despite abundant evidence that subjects of aerial surveys are nearly always under counted (Bergerud 1963, Graham and Bell 1969, Pennycuick and Western 1972i .caughley 1974, Parker 1975; Norton-Griffiths 1976, Melton 1978, Bayliss and Giles 1985, Packard et al. 1985, Pollock and Kendall 1987, Bayliss and Yeomans 1989, Firchow et ale 1990, Johnson et ale 1991, Lefebvre and Kochman 1991). Line transect surveys are attractive because, to be effective, they do not require that all subjects within the bounds of the transect are seen (Burnham et ale 1980). Other attractive features of line transects is that they require minimal effort to establish (White et ale 1989), they can be done from fixedwing, and they make efficient use of air time. The most stringent requirement, however, is that all subjects in the first interval are accurately enumerated. Accurate delineation of all other distance interval boundaries is important to ensure that the detection function is fit to the real distribution of observations. . Line transect methodology was used to estimate mule deer (Odocoileus hemionus·) density using a helicopter as the observation platform (White et ale 1989). Pojar et ale (In Press) used helicopter line transects to estimate pronghorn (Antilocapra americana) density in 2 areas of Colorado. A spec~ally equipped fixedwing aircraft was used by Johnson et ale (1991) to survey pronghorn herds in Wyoming and they obtained adequate line transect samples in ~ 50% of the flight time required for conventional trend counts. Results from fixedwing line transect surveys of large ungulates have not been compared with known densities or other methods where bias has been estimated. Bias was estimated for helicopter quadrat surveys by pojar et ale (In Press). Overcount bias was not detected and undercount bias would have been proportional to the number of animals that did not flush and were not detected during a search of the quadrat by 2 helicopter search crews. These authors believe that helicopter surveys of quadrats is the least biased of any method they tested - helicopter line transect, wide.strip transect, and narrow strip transect. Relative to helicopter quadrat surveys, line transects are perceived to be easier to execute and operation costs of fixedwing aircraft are about onethird that of helicopter costs. Therefore, if fixedwing line transect surveys produce estimates of comparable density and precision to helicopter quadrat . surveys at less cost, then they can be considered for more extensive application as a management tool in Colorado. Line transect data analysis is burdened with subjective decisions on grouping, truncation, and model selection any of which can dramatically affect density and variance estimates. In contrast, quadrats feature a finite population sampling approach and the analysis procedure is direct with no subjectivity in the calculation of density and variance. From an analysis perspective, quadrats offer a standardized procedure which does not add to the variability of the density and variance estimates. �170 However, it is expensive to locate and mark quadrat corners which is one of the principle obstacles to more widespread use of quadrat surveys. Recent technology known as the Global Positioning System (GPS) may provide sufficiently accurate navigation to eliminate the need to physically mark quadrats. Although the typical error for a GPS receiver is claimed to be 2035 m (Hurn 1989), preliminary field tests indicate errors may be somewhat larger (See Appendices I and II). Most field surveys are based on United States Geological Survey maps that have as a basic point of reference permanent field markers (brass caps) which represent relatively accurate survey points (and elevation). The repeatability of locating these markers with a GPS receiver would be useful information for anyone applying this technology for mapping or surveys. Objectives: 1. Compare the pronghorn density estimate and precision of fixedwing line transect surveys with helicopter quadrat surveys. 2. Evaluate consistency of line transect data analysis. 3. Test the accuracy of a Global Positioning System (GPS) in locating known geographic points. Expected Results and Benefits: Good management of any species is anchored to reliable estimates of density. Fixedwing aircraft cost about one-third as much to operate as a helicopter. Therefore, if a survey method using fixedwing produces comparable estimates of density and variance to a helicopter quadrat survey using equal flight hours, then the management agency will realize a significant savings. Quadrat surveys are used extensively in Colorado to estimate mule deer density and experimentally to estimate pronghorn density. Accurately locating quadrat corners is critical to the execution of a quadrat survey. New navigational technology using GPS receivers holds promise to eliminate the cost of establishing and maintaining quadrat corner markers (See Appendices I and II). However, further evaluation of GPS accuracy is desirable before widespread application is acceptable. This experiment provides an opportunity to test GPS accuracy. Approach: The following hypotheses will be tested: , 1. Ho: There is no difference in the den~ity estimate between fixedwing line transect and helicopter quadrat surveys. 2. Ho: Density and variance estimates do not differ by analysis conducted on the same data set by 3 scientists experienced in line transect analysis. The following criteria will be used to subjectively evaluate the 2 survey methods: 1. Precision as measured by the 90% confidence interval. 2., Cost, in terms of time and money, of the surveys. The GPS receiver accuracy wil~ be evaluated as follows: 1. Ho: Direction of GPS error from known points is random. 2. Ho: Distance of GPS error from known points is ~ 35 meters. The fixedwing line transect will be conducted as outlined by Johnson et al. 1991. Navigation will be with LORAN or GPS equipment. Quadrats will be surveyed as described by Pojar et al. (In Press) with the exception that quadrats will be located with a GPS and adjustments in quadrat size will be made if necessary. Latitude and longitude coordinates will be estimated for each quadrat corner from USGS topographic maps. A GPS receiver mounted in the helicopter, using an external antenna, will be used to locate the quadrat corners. If existing markers or brass caps are located, they will be used instead of GPS locations. A 3 person crew (pilot, navigator, and primary observer) will conduct the quadrat survey. ' The t-test will be used to compare estimateq pronghorn densities from line transect and quadrat surveys. For purposes of this experiment, quadrat results will be the standard. �171 Three persons experienced in the intricacies of line transect data analysis will be asked to analyze the same line transect data sets. Each person will be familiar with the analysis criteria of Buckland et al. (1993) and will make independent estimates of pronghorn density and variance using options available in the program DISTANCE (Laake 1993). These persons will not have any familiarity with the actual survey, pronghorn population, or the study area. The 3 density estimates will be compared with the t-test. Coordinates of a sample of 25 USGS brass caps will be estimated from topographic maps of the study area and located on the ground with a handheld GPS. This sample of locations will be selected from quadrat sample units so they will be points that will be encountered during the quadrat survey~ When these points are encountered during the survey, the point indicated by the GPS will be marked on the ground. These points will then be re-located immediately after the survey to measure the distance and angle from the brass cap. The angles will be tested for randomness and the distance will be compared with the t-test against a 35 m standard, which is the amount of expected GPS error. Location: A 452 square mile area north and west of craig will be used as the study area. Forty randomly drawn quadrats (a 10% sample) from a previous quadrat survey will be used in this experiment for the quadrat survey. Some of the corner markers from the previous survey are still in place which will assist in navigation. The line transect survey will be done on the same area within 1 week of the quadrat survey. Schedule: 1993-94 1993-94 1993-94 1994-95 1994-95 1995-96 1995-96 April - estimate quadrat corner coordinates from USGS maps. May - conduct both fixedwing line and helicopter quadrat surveys. June-July - data analysis and preliminary report. May - conduct both surveys. June-July - data analysis and interim report. May - conduct both surveys. June-December - data analysis and final report. Personnel: Principal Investigator: Tom Pojar Consultants: Dave Bowden, Gary White, Dave Anderson, -and Ken Burnham \.. Cooperators: Jeff Madison, Mike Bauman, and John Ellenberger Literature Cited: Allen, S. H. and J. M. Samuelson. 1987. Precision and bias of a summer aerial transect census of pronghorn antelope. Prairie Nat. 19:19-24. Bayliss, P., and J. Giles. 1985. Factors affecting the visibility of kangaroos counted during aerial surveys. J. Wildl. Manage. 49:686-692. Bayliss, P., and K. M. Yeomans. 1989~ Correcting bias in aerial survey population estimates of feral livestock in northern Australia using the double-count technique. J. Appl. Ecol. 26:925-933. Bergerud, A. T. 1963. Aerial winter census of caribou. J. Wildl. Manage. 27:438-449. Buckland, S. T., D. R. Anderson, K. P. Burnham, and J. L. Laake. 1993. Distance Sampling: Estimating Abundance of Biological Populations. Chapman and Hall, NY. 471pp. Burnham, K. P., D. R. Anderson, and J. L. Laake. 1980. Estimation Of Density From Line Transect Sampling Of Biological populations. Wildl. Monogr. 72. 202pp. Caughley, G. 1974. Bias in aerial survey. J. Wildl~ Manage. 38:921-933. Firchow, K. M., M. R. Vaughan, W. R. Mytton. 1990. Comparison of aerial survey techniques for pronghorn. Wildl. Soc. Bull. 18:18-23. Gill, R. B., L. H. Carpenter, and D. C. Bowden. 1983. Monitoring large animal populations: The Colorado experience. N. Am. Wildl. Conf. 48: 330-341. �172 Graham, A., and R. Bell. 1969(68). Factors influencing the countability of animals. E. Afr. Agric. For. J. Special Issue 34:38-43. Hailey, T. L. 1979. A Handbook For Pronghorn Antelope Management In Texas. Texas Parks and Wildl. Dep. No. 20. 66pp. Hurn, J. 1989. GPS A Guide to the Next Utility. Trimble Navigation, P.O. Box 3642, Sunnyvale, California. 76pp. Johnson, B. K., F. G. Lindzey, and R. J. Guenze1. 1991. Use of aerial line transect surveys to estimate pronghorn populations in Wyoming. Wildl. Soc. Bull. 19:315-321. Laake, J. L., S. T. Buckland, D. R. Anderson, and K. P. Burnham. 1993. DISTANCE User's Guide V2.0. Colorado Coop. Fish and Wildl. Res. Unit, Colorado State Univ., Fort Collins, CO. 72pp. Lefebvre, L. W., and H. I. Kochman. 1991. An evaluation of aerial survey replicate counting methodology to determine trends in manatee abundance. Wildl. Soc. Bull. 19:298-309. Melton, D. A. 1978. Undercounting bias of helicopter censuses in Umfolozi Game Reserve. Lammgergeyer 26:1-6. Norton-Griffiths. M. 1976. Further aspects of bias in aerial census of large mammals. J. Wildl. Manage. 40:368-371. Packard, J. M., R. C. Summers, and L. B. Barnes. 1985. Variation of visibility bias during aerial surveys of manatees. J. Wildl. Manage. 49:347-351. Parker, G. R. 1975. A review of aerial surveys used for estimating the numbers of barren-ground caribou in northern Canada. Polar Record 17(111):627-638. Pennycuick, C. J., and D. Western. 1972. An investigation of some sources of bias in aerial transect sampling of large mammal populations. E. Afr. Wildl. J. 10:175-191. Pojar, T. M., D. C. Bowden, and R. B. Gill. Aerial counting experiments to estimate pronghorn density and herd structure. J. Wildl. Manage. (in press). Pollock, K. H., and W. L. Kendall. 1987. Visibility bias in aerial surveys: A review of estimation procedures. J. Wi1dl. Manage. 51:502-509. Springer, L. M. 1950. Aerial census of interstate antelope herds of California, Idaho, Nevada, and Oregon. J. Wildl. Manage. 14:295-298 • .White, G. C., R. M. Bartmann, L. H. Carpenter and R. A. Garrott. 1989. Evaluation of aerial line transects for estimating mule deer densities. J. Wildl. Manage. 53:625-635. �173 Colorado Division Wildlife Research July 1994 of Wildlife Report JOB PROGRESS state of Project Work REPORT Colorado No. Plan No. W-1S3-R-7 SA 2 Job No. Period Covered: Author: Thomas Mammals Research Black Bear Research Development of Black Inventory Techniques Bear July 1, 1993 - June 30, 1994 D. I. Beck Personnel: A. Anderson, T. Beck, J. Beach, J. Broderick, M. Caughlan, R. Hays, S. Lechman, Mark McLain, Mike McLain, J. Olterman, C. Parmeter, L. Willmarth, CDOW; R. Stevens, Colo. State Patrol; T. Holland, USFS. Abstract A total of 134 captures of black bear (Ursus americanus) were made over an 88day period with a total of 2,161 trap days. Eighty-nine individual bears over the age of l-yr were tagged; 23 subadult females, 18 adult females, 30 subadult males, and 18 adult males. A total of 17 cubs were captured and released untagged. Only 1 injury to a bear was recorded; a broken canine tooth. Of the 89 tagged bears, 73 were radio-collared. A high proportion of the collared bears leave the formal study area in mid-August., returning in early October. This will affect conduct of the sighting phase of the study. ��175 DEVELOPMENT OF BLACK BEAR INVENTORY Thomas TECHNIQUES D. I. Beck P •N. OBJECTIVE 1. 2. 3. Evaluate a capture-sight program utilizing cameras set on bait stations for estimating black bear density. Document age and gender bias in vulnerability of black bears during autumn hunting seasons. Obtain density estimates of black bears in 3 heavily hunted areas of markedly different vegetation communities. SEGMENT 1. 2. 3. 4. OBJECTIVES Capture, "tag and radio-collar as many black bears as possible in a 465 km2 study area on the Uncompahgre Plateau. Evaluate predator calling as a method for observing/hunting black bears in Colorado. Test the hypothesis that protein tissues, in concert with fat, are vital reserves for bears in winter, especially while lactating. "Test the hypothesis that skeletal muscle is defended during hibernation by retaining protein content and integrity of muscle fiber ratios. METHODS AND MATERIALS Study Area Description The Uncompahgre Plateau study area was located on the northern end of the plateau, in parts of Game Management Units 61 and 62. The east and west boundaries of the study area roughly coincide with the change from oak shrublands (Quercus gambelii) to pinyon-juniper woodlands (Pinus edulis, Juniperus monosperma). The northern boundary is the rim on the south margin of Unaweep Canyon while the south boundary is non-topographical and occurs where we reached a total of 45 quadrants (near Uncompahgre Butte). The area is roughly 22 km north-south and 25 km east-west. Elevations vary from t195 to 2895 m. Vegetation communities vary markedly throughout the area. Oak shrub lands and aspen stands (Populus tremu16ides) are most common although extensive stands of Ponderosa pine (Pinus ponderosa), sagebrush-grassland (Artemisia tridentata), pinyon-juniper, and Douglas fir (Pseudotsuga menziesii) are present. Timber type maps (U. S. Forest Service) for the area have been obtained ~nd vegetation communities for each quadrant will be determined. The study area was subdivided into 45 quadrants, each 10.4 km2 in area. The size was based on the average annual range of subadult female black bears on the Black Mesa Study Area, an area of similar habitat (Beck 1991). The use of this size quadrant was to insure that each bear on the study area would be exposed to at least one trap, since a trap would remain in each quadrant during the entire trapping period. Capture and Marking Black Bears A new cage-type trap for live capture of bears was developed and 45 were manufactured. A trap measures 1.8 m long and 1.0 m high and wide. The frame is constructed of angle iron and all side and top panels are wire mesh, with a 1.9 em mesh size. The floor is 16 gauge steel. A spring-powered, solid aluminum door is mounted on a full-length side hinge at one end. A fulllength latching mechanism holds the door closed. The door is triggered via a �176 treadle pedal mounted on the floor 1.0 m from the door. A standard garage door coil spring provides the closing power. Along one side of the trap is a hinged panel measuring 1.8 m X 0.3 m. Behind this panel are vertical bars placed on 0.3 m centers. Swinging the window up allows access through the barred area for administering immobilizing drugs by jabpole. Each tr~p weighs approximately 236 kg. Two trailers were custom built to transport the traps behind ATV's. The trailers have an angle iron frame mounted on a 682-kg load axle with 30 cm high ATV-type low-pressure tires. Two Suzuki ATV's were loaned to the project by Davis Tire & Service center, Montrose, CO. One person can load, transport, and set up the traps with this arrangement. Two people could set a trap in a pickup truck and a few traps were set out this way. Traps were set on the ground and tied to nearby trees or steel stakes to prevent tipping. Traps were distributed, one per quadrant, during 3-20 June 1993. It was initially planned to operate all traps simultaneously with each crew member checking 15 traps per day. This proved inoperable with the number of captures we had. Therefore, we set 30 traps initially. After 1 week, we closed 15 and set the remaining 15. We then maintained a schedule for each trap of 2 weeks set, 1 week closed. On any day, we usually had 30 traps operable. While each quadrant did not receive equal trapping pressure, differences were less than 10% of the average trap days effort. Most set traps were checked daily. Traps at the more remote sites were equipped with trap-door transmitters manufactured by Advanced Telemetry Systems, Isanti, MN. The transmitter emitted a slow-pulse signal with the door in the open position and a fast-pulse signal when closed. These traps were checked at 2-day intervals when the transmitter indicated open doors. Traps were baited with rotten fish, butcher shop scraps, or spoiled fruit contained in burlap bags. The bags of bait were wired to the end mesh panel opposite the door. The initial trapping period began on 21 June and ended on 27 July. The fall trapping period ran from 12 August to 21 September. At initial capture, all black bears older than cubs were immobilized with a mixture of ketamine hydrochloride and xylazine hydrochloride. The K/Xy ratio was 5:1 and the concentration of ketamine was 200 mg/ml. The dosage rate for black bears was 8.8 mg ketamine/kg bear. The drug was administered by a _ spring-powered jabpole. Because underdosing with ketamine/xylazine causes no response in the bear, crew members were instructed to dose for 20 kg over the estimated weight of the bear. There is a wide safety margin with this drug mixture and this procedure usually prevents high doses as a result of multiple administrations of drug. Cubs and recaptured bears with intact ~ags/collar were released without immobilizing. The spring was removed from the door by disconnecting a clevis pin. This allowed the door to be easily opened. Crew members usually opened the door from the opposite end of the trap with the aid of a pole. Upon immobilization, a bear was removed from the trap and placed in a shady area in a dorsally recumbent position. An opthalmic solution was placed in each eye to inhibit drying and the eyes were covered with a cloth. A unique combination ear tag/streamer was placed in a specific ear. Radio collars (Advanced Telemetry System, Isanti, MN) were placed on most bears. A few yearlings were too small for collars. During July we ran out of collars so quit collaring subadult males for a short period until a supplemental order arrived. Each collar had a canvas section bolted into the belting. This section is designed to rot through, thus allowing the collar to falloff. The canvas sections were obtained from Blue Star Canvas, Missoula, MT. Ear tags were AIIFlex cattle tags (5.5 X 4 cm), unique~y numbered in 5 colors: red, white, 'orange, yellow, blue. A pair of streamers was attached to each ear tag post. The streamers were made from 2 layers of 22-oz PVC glued together for stiffness. The PVC was provided free (manufacturing scraps) by �177 Jack's Plastic Welding, Aztec, NM. The 2 streamers were glued together at one end to form a chevron with about a 30 degree angle. Each streamer was 2.5 X 13 cm. A number of colors were evaluated for clarity in photographs taken in varying light conditions.' The streamer combinations used were: red/yellow, red/red, red/white, red/teal, yellow/yellow, yellow/white, yellow/teal, blue/yellow, white/white, orange/white, teal/white. Each bear was uniquely identifiable by which ear was tagged, tag color, and streamer combination. The tag number is not necessary for accurate identification. The following characteristics were recorded for each bear: sex, age group (subadult, adult, cub), color, weight, total length, chest girth, number of incisors, broken teeth, nipple size' and color on females, pelage condition, breathing rate. Date, quadrant, and UTM location were recorded. In cooperation with a study of cementum annuli for accurate age estimation black bears, a number of bears were injected with tetracycline compounds (Biomycin, Liquimycin) at a dosage of 10 mg tetracycline/kg body weight. These injections were subcutaneous in the neck region. of Crew members left immobilized bears in shady areas for recovery. Crew usually vacated the area to allow uninterrupted recovery. The exceptions were when people and/or domestic cattle were nearby; then crew members stayed within .sight of the bear until the bear was fully recovered. All crew members received detailed written and oral instructions on the safe handling of trapped bears, emphasizing safety to the bears foremost. Additionally, all crew members worked at least 7 bears with an experienced handler (Beck or Broderick) before handling a trapped bear by themselves. Monitoring seasonal movements Aerial radio-tracking was conducted at biweekly intervals beginning in early July 1993 and continuing through early December. Aerial locations were recorded by longitude and latitude through onboard Loran navigational units. Locations were converted to UTM's later. Evaluation of predator calls for attracting black bears No work was conducted toward this objective because of the trapping commitment. This objective will be addressed in the next segment. Winter black bear physiology study Because of delays in procuring equipment effort was postponed until FY 94-95. RESULTS capture and finalizing contract, the primary AND DISCUSSION and marking 'black bears OVer an 88-day trapping period with 2,161 trap days, we made 134 captures of black bears (Tables 1,2). Eighty-nine individual black bears older than cubs were ear-tagged. Of these, 73 were radio-collared. All adults of both sexes, 18 of 23 subadult females, and 19 of 30 subadult males were radio-collared. Since cubs were not marked, we do not know how many individuals were represented in the 17 cub captures. �178 Table 1. Composition INITIAL AGE/SEX ADULT of captured FEMALES black bears, CAPTURES Uncompahgre INDIVIDS Plateau, RECAPT. 1993. TOTAL RECAPTURES 18 5 7 23 3 4 ADULT MALES 18 4 7 SUBADULT 30 6 10 SUBADULT FEMALES MALES CUBS· 17 TOTAL 106 Table 2. Trapping Plateau, PERIOD success 1993. DID NOT TAG 18 by time on black NEW CAPTURES TRAP DAYS bears> 28 1-yr old, Uncompahgre TO/NEW CATCH CUM. CATCH 6/21--6/29 27 237 8.8 27 6/30--7/08 17 278 16.4 44 7/09--7/18 17 302 17.8 61 7/19--7/27 14 288 20.6 75 8/12--8/22 3 241 80.3 78 8/23--9/02 3 251 83.7 81 9/03--9/12 5 305 61.0 86 9/13--9/21 3 259 86.3 89 89 -2161 24.3 TOTAL There was only 1 injury to a bear in 134 captures. An adult male broke a canine tooth while in the trap. ThiEf is in welcome contrast to the 37% injury rate to bears caught in snares during the Black Mesa study (Beck 1991). The primary reason to develop the new trap was to reduce injuries to black bears. In this regard, the new trap is a wonderful success. Except for apprehension by crew, there were no problems encountered releasing cubs and older bears from the trap directly, even when the mother of cubs was visibly present. When exiting the trap, the near-universal pattern was tor the bear to approach the open door tentatively, slowly setting one foot out at a time. Then once all 4 feet were on solid ground, all bears exhibited enthusiasm in leaving the area. It did not appear that captured bears were expending much effort trying to break out of the traps. The only damage to any traps was inflicted by a freeranging adult female in attempting to free her captured cubs; she stretched a spring on the treadle peddle. The treadle peddle was designed to flop backward to lay flat on the floor. This is the position it was always found in. Most bears appeared outwardly calm when in the trap. Often small subadult bears paced from end to end. We captured a subadult male at our camp in June 1994 where we heard the door slam shut and were able to immediately begin observing his behavior without being seen by the bear. He flipped the treadle peddle, paced for <10 minutes, then laid down. The capture rates for the June-July period were unexpectedly high, based on prior experience with snaring. I speculate that the open appearance of the wire mesh panels contributes greatly to the efficiency of capture and the apparent calm of captured bears. In an effort to further evaluate the �179 efficiency of the trap, 20 traps were loaned to the New Mexico Game & Fish Dept. and Hornocker Wildlife Research Institute. The traps will be used on an area with extensive snaring (>5 yrs) and an area with only 1 years prior snaring. Of interest is not only capture rate, but age and gender composition of captured bears. capture rates in August-september were quite low. Two plausible explanations developed were 1) most of the bears had migrated out of the 45-quadrant study area and 2) a very high proportion of the bears had already been captured and were displaying trap avoidance. Radio-tracking data supported that at least 75% of the collared bears were out of the trapping grid for most of the trapping period. The collared bears within the trapping area were in close proximity to traps daily. The camera-based sighting data in 1994 should clarify the more probable explanation. Initial captures were not distributed uniformly throughout the 45 quadrants, ranging from 0 to 8. In general, the capture frequencies were close to what was expected based on habitat considerations. The exceptions were situations where fewer bears were caught than expected based on the amount of aspen-oak woodlands. A more detailed examination will be made once the USFS timber type maps have been converted to our quadrant grid. The age and gender composition of the captured bears was another surprise. Nearly all studies demonstrate a strong bias for capture of subadult males, especially if one only traps for a single season. The sample captured appears to be appropriately representative for a moderately exploited black bear population. The only anomaly is the greater number of subadult females caught than adult females. Among most bear researchers, the capture of subadult females is viewed as a valued measure of trapping effectiveness. Since they move less than other groups of bears, they are harder to catch. Thus if you are catching subadult females, you are probably catching the other groups as well. The large numbers of subadult females caught suggest that the combination of trapping by quadrants and the new trap was very effective. The relatively low numbers of adult females is presently unexplainable. Examination of hunter kill records for the previous 13 years does not offer an explanation. Pe~haps data from the 1994 camera sightings will help clarify. Typical for western Colorado, 80 of the 89 tagged bears were broWn in color (40 females, 40 males). Of the 9 black colored bears, 8 were male. Weights of Uncompahgre Plateau black bears were similar to those recorded in the Black Mesa study (Table 3). Twenty-nine bears were inoculated with tetracycline. The sample was composed of 8 subadult females and 7 each of subadult males, adult males, and adult females. Unfortunately, sufficient supplies of light-proof bottles for transporting'the compound were not received until mid-way throughout the first trapping session. Table 3. Comparison Uncompahgre Plateau of black bear weights between and Black Mesa, in kg. 2 western N AVG WT areas, BLK MESA ONC AGE/SEX Colorado RANGE AD FEMALE 15 74.1 52-111 SUB FEMALE 20 46.4 23-84 AD MALE 15 118.2 SUB MALE 26 56.4 91-152 28-90 AVG WT RANGE 9 76.1 61-98 13 40.8 27-57 9 127.3 91-159 15 61.0 41-86 N �180 Monitoring seasonal movements Aerial radio tracking indicated a large proportion of the collared bears left the study area in mid-August. Such a migration was anticipated based on the findings from the Black Mesa study (Beck 1991). The greatest number of .collared bears moved into Unaweep Canyon where much of their foraging was on a steep north-facing escarpment. All access to the lower sites was limited by private land ownership. While these bears were outside the 45 quadrants, they were all within 1 Jan of the main study area, but about 850 m lower in elevation. A significant number of collared bears moved south of the study area into the Escalante Breaks and Kelso Mesa areas. These areas are 1-8 Jan from the formal study area. Fall seasonal use of areas continues to be difficult to predict just based on vegetation communities. While it was expected that many bears would go into Unaweep Canyon, there were ample soft and hard mast supplies in oakbrushdominated areas outside the canyon; yet many of these areas received limited use by black bears. I wonder how much prior experience, especially when young, dictates later behavior? Regardless, we were able to identify 2 seasonally critical areas for black bear on the north end of the Plateau. While this fall movement probably negatively impacted our trapping success, it was quite clear that photographic sighting of bears in the 45 quadrants during September will be most unproductive since a high proportion of tagged bears will be absent from the study area. Thus the study design for the sightings will be modified to omit September. This will also allow more time for evaluation of predator calling as a viewing/hunting technique. Of great concern was the inaccuracy of many of our aerial locations. There did not appear to be a pattern between experienced trackers (>15 yrs) .and novices, nor between 2 different pilots. We encountered great difficulty in taking aerial locations of dens (December) and even finding a radio signal when on the ground. Two slipped collars had been tracked throughout the summer and fall, unknown to the trackers that they were not on a bear. One collar was located 13 times with an average error between known location and estimated location of 4.6 km, and a range of 0.5 to 13.4 km. The other collar was located 3 times after slipping, with errors of 1.1, 4.0, and 7.3 km. Errors between reported location and actual location of 3.dens was 0.7, 1.0, and 0.7 Jan. The radio telemetry equipment used is simirar and/or identical to that which was used on earlier studies of black bears in topographically more difficult terrain, yet these types of inaccuracies were not encountered earlier. Considerable effort will need to be expended in 1994 to improve the accuracy of locations, especially during denning season. In the interim, all locations will be marked on topographic maps while in the air rather than rely on the Loran system coordinates. LITERATURE CITED Beck, T.D.I. 1991. Black bears of west-central Tech. publ. 39. 86 p. Prepared Colorado. by Thomas D. I. Beck, Wildlife Researcher Colo. Div. Wildlife �181 Colorado Division Wildlife Research July 1994 of Wildlife Report JOB PROGRESS State Work Colorado of Project No. W-153-R-7 9A Plan No. Author: Covered: Mammals Research Elk Investigations 3 Job No. Period REPORT spatial Analysis of Elk Survival July 1, 1992 - June 30, 1993 K. Wilson, N. T. Hobbs Abstract We prepared a detailed study plan to develop methods for analysis relationship between animal survival and habitat use. of the ��183 SPATIAL ANALYSIS OF ELK SURVIVAL P. H. OBJECTIVES The objective of this project is to evaluate methods for the statistical analysis of effects of habitat use on survival rates of mammals. We will determine the feasibility of detecting differences in population processes attributable to variations in landscape patterns. This will be accomplished by evaluating various survival analysis models using spatial simulation modeling. SEGMENT OBJECTIVES 1) Develop a geographic information system for a landscape used by elk. Build the system to include data layers for elevation, slope, aspect, vegetation types, roads, and streams. Based on this system, simulate artificial landscapes showing different levels of heterogeneity in landscape pattern. 2) Simulate ~ovement of elk over the landscape and their survival over time. 3) Simulate detection of animals by radio telemetry. Exact locations animals will be based upon error polygons, using different confidence shapes and sizes. of ellipse 4) Develop a statistical approach for evaluating the effect of landscape heterogeneity on elk survival. Alternative analytical approaches will be evaluated by comparing bias, precision, power, and goodness of fit. 5) Using 1-3 above, develop alternative sampling designs for obtaining data needed for statistical approach identified in objective 4. Evaluate economic feasibility of implementing sampling designs. RESULTS Traditionally, habitat use studies have relied on 3 general designs: (1) Population level designs where used, unused, and available resource units are sampled or censused for an entire study area and for a collection of animals (but individual animals are not identified), (2) An individual level design where individuals are measured, but resource availability is measured at the population level, and (3) individual" level design where individuals are measured and resources are measured for each individual animal (Thomas and Taylor 1990). Studies using these designs usually try to infer some preference or avoidance for a particular habitat, but selection for a particular habitat does not necessarily mean the habitat is critical for survival and/or reproduction (White and Garrott 1990). In fact, several authors have been critical of the use o( use/availability data to infer fitness (increased population size or survival) (see Van Horne 1983, Hobbs and Hanley 1991). Research has continued on methoQs of analyzing the survival of elk as a function of spatial movement in space and time; the first step in determing the mechanistic link between an animal and its habitat. Our assumption is that variation in the use of space (habitat type, elevation, distance to water, etc.) contribute to variations in population processes (e~g. survival). Our objectives a~e to evaluate statistical approaches for evaluating the effect of landscape heterogeneity on elk survival. Procedures for simulating .landscape heterogeneity will be developed with sources of heterogeneity such as slope, aspect, and vegetation cover types, and movement of animals will be simulated over the maps with survival over time varying as a function of landscape heterogeneity. We have completed the investigation into potentially useful survival models, and these models are outlined below. �184 SURVIVAL ANALYSIS MODELS Our assumption is that the survival of animals will vary as a function of landscape heterogeneity. As with any estimator, an ideal survival estimator will have little bias and good precision (small variances). We have primarily restricted survival estimators to estimators that utilize maximum likelihood techniques. Maximum likelihood estimation is a powerful tool for deriving point estimators and estimators of sampling variances and covariances (Mood et ale 1974). In addition, these models allow testing of model fit and a variety of hypotheses. The robustness of an estimator to failure of model assumptions is also critical, and it is important that methods exist for evaluating failures of these assumptions. There are two components of variance to be considered in this study. The first is spatial or temporal variation. Survival rates of animals will vary from area to area and individual to individual. This variation may be due to landscape heterogeneity, age, sex, etc. The other variance component is sampling variation. This variation is a result of the stochastic nature of the sampling process used to estimate survival. It is a measure of the repeatability or precision of the result, and can be controlled by increasing sample sizes, stratification of the population, or use of a "better" estimation method. The following are assumptions that apply to all of the survival estimation techniques that will be discussed: (1) the distributions of the measured variables for the resources selected do not change during the study period and (2) the population of resource units available to the animals are correctly identified. The first assumption is often difficult to obtain, especially if the study is over several seasons. When this assumption is not met then the results are averages over the period (Manly et ale 1993). Assumption (2) is a very critical and difficult aspect of any study. This assumption assumes that the "important" variables are being sampled, but the only true method of ascertaining whether or not a variable is important to an animals survival or reproduction is through a manipulative experiment (White and Garrott 1990). In many respect,s, this research is designed to give researchers tools that will help in defining potentially "important" variables. Assuming that we have an ideal method of gathering animals such as a global positioning system collar, each spatial history comprised of the habitats traversed over might have habitat type, elevation, slope, etc. When an of death is recorded. Survivorship can be modeled using function based on the times of death for each animal, T, S( t) = # animals surviving longer than total # of animals spatial locations of animal will have a time. Each location animal dies, the time a survivorship where t ' The hazard rate, h(t), or probability of dying during a very small interval, given that the animal survived to the beginning of the interval, is another function commonly used in survival analysis. The hazard rate is sometimes referred to as the instantaneous death rate. Most of the methods developed for survival analysis come from medical studies (Lee 1992). For most medical studies, the time component is not important, but in wildlife studies environmental factors related to time can be very important (White and Garrott 1992). For example, seasonal weather patterns ~nd even year to year variations in weather can significantly impact survival •. Most survival models are based on smooth functions such as the exponential, Weibell, lognormal, and gamma distributions. The stochastic nature of wildlife data, due to the time variations mentioned above, can often violate these assumptions, and thus it is important to evaluate the robustness of these models in w~ldlife studies (White and Garrott 1992). The survival models that have been chosen for study use multiple regression �185 techniques to determine the relationship, if any, between survival time and possible covariates such as habitat type, slope,'aspect, age, etc. Standard regression methods would require the assumption of a no+mal distribution which is usually not appropriate for survival data, thus our focus is on the use of alternative distributions. There are 2 overall approaches for analyzing survival data: (1) continuous time approach (survival times are recorded) and (2) discrete or dichotomous approach (an animals is considered either alive or dead at the end of the study). CONTINUOUS hazard form TIME APPROACHES Cox Proportional Hazard Model (see Cox 1984): This model assumes the functions for all animals are proportional to one another and takes the p =ho (t) exp (E ~jXj) , j-1 where xl' x2' ••• ' xp are the covariates of interest, h(t:x) is the hazard function, and hart) is considered the baseline hazard function when the X variables are ignored. There is no .underlying distribution assumed, so the proportional hazard model is a nonparametric approach. By dividing both sides of the equation by hart) and taking the logarithm, the following regression equation is obtained for individual i. Log-Linear Model (see Lee 1992) : The log':"linearapproach is based on" the' assumption that the underlying distribution is exponential with a survivorship function of where P '-1 =exp ( a E b x i+ 1 11) • 1-1 The exp(aj) .ignored. term represents the underlying DICHOTOMOUS hazard when the covariates OR DISCRETE are APPROACHES Discriminant Function Model (see Johnson and Wichern 1988): In discriminant analysis the goal is to separate (discriminate) the population into 2 classes .(alive and dead) based on the covariates, Xj' of interest. The covariates are assumed to be multivariate normal. Discrete covariates such as habitat type or sex can be included in discriminant analysis, but the results can be highly variable depending on the correlations of the covariates (Johnson and Wichern 1988). White and Garrott (1990) recommend logistic regression over discriminant analysis, because of this fact. Linear Loaistic Regression (see Hosmer and Lemeshow 1989): In logistic regression, only the fate of the animal, alive or dead, is used. The probability of survival for each individual is defined as Sj and the probability of death is l-st • Standard regression assumes that the dependent �186 variable is normally distributed, violated. Instead Sj is modeled Yi and since Sj is binomial this assumption using the logistic transform as follows: = is In{_5_} l-Si where Yj is often referred to as the logit transform or log odds. The importance of the transform is that Yi is linear and can range from -00 to +00 depending of the Xj; this puts the regression equation in a linear regression framework. . For the survival models outlined above standard regression techniques are employed such as testing for significance of the ~ terms. In addition, regression fitting methods such as stepwise, backward, and forward selection procedures are possible. Programs for analysis of survival data are included in most statistical packages and program SAS (SAS Institute 1990) we will be used for analysis of our data. SIMULATIONS The computer simulation model is currently being developed. The model will include algorithms for generating a spatially heterogeneous landscape and animal movement. The program will be general in nature which will allow its utility for a variety of animals besides elk. The simulation model will be used to evaluate which survival models perform better under various sampling designs, and to evaluate the effect of landscape heterogeneity on survival estimation. The landscape will be created as a grid or raster map. Map cells will have certain attributes such as vegetation and slope assigned to them based on desired patchiness of the distribution. Distributions will be allowed to range from random to fixed to contagious. In addition the scale (cell size) will be allowed to vary. Animal movements will be simulated on the landscape according to various models. Models proposed for incorporation are: (1) simple random walk where movement is independent of the landscape and (2) rule-based movements where movement is a function of the animals direction, speed, memory (previous history), etc. The rule-based approach is more realistic and should provide the most interesting results. Survival of animals in the simulations is then based on their movement through the landscape over time. For example, all animals might have an initial survival rate of 1, but as they move through the landscape there probability of survival will change as a function of their "spatial history" (Le. movements through different cells within the landscape). Each cell, depending on the cell's attributes, that an animal enters can potentially decrease, increase, or maintain the animal's current survival probability. At each time step an animal has the chance of living or dying. At the end of the simulation, an array of animal movement locations and the times of death are recorded. This data will then be used to evaluate the survival models outlined above. Each simulation setup will be replicated numerous times (e.g. 100) in order to obtain an estimate of the true overall survival for a particular landscape. In the "real" world the entire spatial history of an animal is not known. Instead, a sample of this spatial history is obtained (e.g. an animal.s may only be located once a day or once a week), because it would be quite costly and time consuming to record all.animal movements. Since the entire "spatial history" is known for each animal, we will be able to evaluate the survival models under various sampling regimes. This should be very usefUl in �187 determining sampling intervals and sample sizes survival estimates based on resource usage. for obtaining reliable As landscapes become more heterogeneous, it should be more.difficult to reliably estimate survival. By changing the distribution of the cell attributes (e.g. random versus regular or clumped), the survival of animals will also change. The reliability of the survival models will also be investigated as a function of landscape heterogeneity. This approach.will also be evaluated as a function of samplin~ intervals. A detailed (Appendix A). Program Narrative was prepared to investigate these questions REFERENCES Cox, D.R. Hobbs, 1984. Analysis of survival data. Chapman and Hall, England. N.T., and T.A. Hanley. 1990. Habitat evaluation: do use/availability data reflect carrying capacity? Journal of Wildlife Management 54:515522. Hosmer, D.W., and S. Lemeshow. and Sons, New York, NY. 1989. Applied logistic regression. Johnson, R.A., and D.W. Wichern. 1988. Applied multivariate analysis. Prentice Hall, Englewood Cliffs, NJ. Lee, .E.T. 1992. Statistical Sons, New York, NY. Manly, London, B.F.J., animals. methods for survival statistical data analysis. L.L. McDonald, and D.L. Thomas. 1993. Resource Chapman and Hall, London, England. Mood, A.M., F.A. Graybill, and D.C. Boes. 1974. Introduction statistics, 3rd ed. McGraw-Hill. New York, NY. SAS Institute. 1990·. SAS/STAT Gary, NC. User's Guide, Version 64th John Wiley John Wiley selection and by to the theory of ed. SAS Institute, Thomas, D.L., and E.J. Taylor. 1990. Study designs and tests for comparing resource use and availability. J. Wildl. Manage. 54: 322.-330. Van Horne, B. 1983. Density as a misleading indicator Journal of Wildlife Management 47:893-901. White, G.C., and R.A. Garrott. 1990. Analysis Academic Press, San Diego, CA. ·of habitat of wildlife quality. radio-tracking data. �188 APPENDIX A PROGRAM NARRATIVE State of _--'=C=oc:..;lo=r=a=do=--_ Project No. W-1S3-R-8 Mammals Research Work Plan No. _---..<:9~A~ __ ElK Investigations Job No. Spatial Analysis of Elk Survival 3 A. NEED Understanding the link between an animal and its habitat (basic resources such as water, food, cover, etc.) has been a principle area of interest in ecology for a long time (for an overview see Morrison et al. 1992). The present distribution of plants and animals is function of past and present adaptations to biotic and abiotic factors. Abiotic factors include temperature, precipitation, soil, etc. whereas biotic factors include factors such as competition and predation. Our understanding of evolutionary factors that link a species with its habitat are limited though. Therefore, most wildlife studies have primarily focused on a correlative approach comparing an animals habitat use to the availability of habitat. The premise being that animals use habitats in proportion to their area.' and that "better" populations can be maintained by protecting or creating habitat that is suitable for an animals survival. Consequently, many habitat studies have focused on the idea of habitat suitability (quality) in order to predict "potential" areas of habitat use (Flood et al. 1977, Verner et al. 1986). These approaches have been criticized because there is rarely, if ever, a link between the "fitness" of the animals and the habitat in question. For example, Van Horne (1983) showed that there is not always a direct linkage between density of a population and habitat quality. Further, Hobbs and Hanley (1990) showed that use/availability data can be misleading if both habitat quality and quantity are not taken into account. A recent study comparing ground counts of beavers and habitat suitability index values confirmed the criticisms outlined by Van Horne (Robel et al. 1993). Both Van Horne (1983) and Hobbs and Hanley (1990) have cautioned against using habitat use/availability data without establishing mechanistic links between patterns of habitat use and population processes such as birth and death. In addition, ecologists have become increasingly aware of the importance that the landscape plays in our understanding of the relationships between animals and habitat. The ecology of the landscape such as the amount of heterogeneity, fragmentation, landscape diversity, connectivity, etc. are all important in defining a population (Forman and Godron 1986). �189 Recent progress in remote sensing and geographical information systems (GIS) has greatly increased the amount of spatial data available to ecologists. This and the increase in the use global positioning systems (GPS) has increased our ability to detect habitat usage by animals. The combination of precise animal movement data and GIS layers such as vegetation type, elevation, slope, precipitation, etc. offer a unique opportunity of coupling population processes (e.g. survival) with habitat usage data. B. OBJECTIVE The objective of this project is to evaluate methods for the statistical analysis of effects of habitat use on survival rates of mammals. We will determine the feasibility of detecting differences in population processes attributable to variations in landscape patterns. This will be accomplished by evaluating various survival analysis models using spatial simulation modeling. C. EXPECTED RESULTS AND BENEFITS Understanding the significance of data on habitat use has been hampered by poor accuracy of radio telemetry equipment (see Garrot and White 1992). This is particularly true for large mobile ungulates like elk. However, global positioning systems (GPS) will in the near future provide habitat usage data with an accuracy of 5 m or less; in fact, a GPS collar for large ungulates has recently been introduced by a Canadian company at approximately $5000 per collar. The current collar has the ability to store locations of animals as' frequently as every hour, but the costs of collar maintenance and data retrieval increase. with increasing sampling frequency. This study will help assess sampling frequency and will help researchers and managers assess potential study cost. In addition, LANDSAT data has provided users with a relatively easy method of acquiring and updating landscape data over large areas. This project would give resource managers the ability to directly assess the effect of landscape heterogeneity on wildlife survival. This is especially important for evaluating the impacts of management practices such as clear cutting, road building, and habitat improvement. Finally, as our knowledge of animal habitat usage is improved, survival models can be developed using the GIS and GPS data that can ultimately be used to predict the effect of natural and human disturbances on animal survival. D. APPROACH A spatial simulation model will be developed to evaluate methods for quantifying differences in survival based upon landscape characteristics. A square raster (cell by cell) map will be created. The landscape will have the flexibility to be varied from as small as a 2 x 2 to 1000 x 1000 array of cells. Cell size can vary, but the initial size will be 1 ha. Each cell can be assigned any number of attributes such as vegetation size, elevation, slope, etc. Spatial distribution of the attributes can be random, regular, or clumped (Whittaker 1975). A random pattern would be characteristic when the location of one individual is completely �190 independent. A regular pattern might arise because of territoriality or competition; this type of pattern might be argued for some types of vegetation, but not for an attribute such as soil type. A clumped distribution might arise from the aggregation of similar attribute types; aspen trees are a good example. In addition, clumps may be distributed in a random or regular pattern. Any number of animals can be placed within the landscape in random, regular, or clumped distributions. Animals can move through the landscape using a random walk, random walk limited by a home range or a rule-based approach. In a random walk an animal is allowed to move in any direction and each movement is independent of previous movements. A rulebased approach uses foraging behavior to model movements, and is more realistic. For example, specific rules about distance, direction, speed, affinity to vegetation types, etc. will be incorporated into animals movement. In this case a general rule based model will be developed that incorporates features of mammal movement. An example of how a simulation run might proceed is outlined below. Initially, a landscape with one attribute such as vegetation is created. The size of the landscape is chosen, e.g. 50 x 50 cells at 1 ha each for a total area of 2500 ha. Two vegetation types such as aspen and sage are randomly distributed in equal proportion on the landscape. A population of animals is then distributed according to a desired spatial distribution, e.g. random clumps with an . average of 5 animals per clump and 100 total animals in the population. An animal is allowed to move/forage throughout the landscape dependent upon the movement model and the time step. The time step might be 1 hour or 1 day. At each time step an animal has a probability of survival based upon the attributes of a cell. For example, an animal might have an 0.95 survival rate if the cell at' any time step contains aspen and an 0.85 survival rate if the cell at any time step contains sage. A uniform random number between 0 and 1 is generated for each animal at each time step, and if the random number is less than or equal to the survival probability for that animal then the animal survives. An animal continues to forage/move through the landscape until the end of a predetermined time period or the animal dies. At the end of the simulation each animal the exact proportion of animals surviving is known and each animal has a unique history of cells/attributes traveled through. The landscape attribute history will be used to evaluate the survival models outlined below. The survival rate of each animal is modeled as where S, is the survival rate for an animal at time t, St-l is the survival rate at the previous time step, and S, is the effect on survival due to the k attributes (in the example given above k= 1 'because the only attribute was vegetation type). All animals can begin with the same survival rate, So (e.g. 0,9) or survival rates can be heterogeneous (i.e. each animals has a different survival rate based on a known distribution). In the above equation any S, < 1 would decrease the overall survival of an animal for that cell. �191 E. ANALYSIS Of interest in our work is the estimation of survival, but more importantly the factors (covariates) that affect survival. Several regression approaches for relating covariates to survival will be examined. Common procedures for investigating the relationship between covariates and survival are proportional hazard rate models (Cox 1984, Lee 1992) and logistic regression (Hosmer and Lemeshow 1989). The proportional hazard rate model can be expressed as where w(tJ is a function of the survival time of the ith individual and f(xJ is a function of the covariates (landscape attributes) of interest (see Lee 1992:250). In logistic regression, the outcome is categorical (i.e. 0 if an animal dies and 1 if an animal dies) and survival of an individual can be modeled as where g(x) is the log it transform based upon an animal dying or surviving and covariates or landscape attributes of interest (Hosmer and Lemeshow 1989:5-7). Xi are the The following factors will be investigated for the two survival models (the number of levels at each factor are in parentheses): number of landscape attributes (2), spatial distribution of attributes (3), sampling interval (3), and type of movement (2). Again, landscape attribute refers to landscape features such as vegetation type, slope, aspect, etc. The· spatial distributions examined will be random, regular, and clumped. Sampling interval refers to the fact that in nature it is virtually impossible to continuously relocate an animal, and in fact animals are only located daily or weekly. The survival models will be examined at 3 sampling intervals: continuously (i.e. at each time step), daily, and weekly. Finally, the type of movement (foraging) will be either a random walk or a rule based movement as outlined above. The combinations of the above factors will provide a wide range of potential outcomes for evaluating which survival models "best" represent the true survival of the animals as represented by the simulations. The true survival rate for any particular simulation setup is obtained by completing a large number of simulation runs, e.g. 100 repetitions of the example outlined in the previous paragraph. After 100 repetitions the true survival rate is the average proportion of animals alive for the simulations (also known as the empirical survival rate). The percent relative bias (pRB), percent confidence interval width coverage (COV), and percent coefficient of variation (CV) will be measured in order to evaluate the performance of the survival models. The goal is to determine which survival models best represent the landscape characteristics simulated with the least amount of bias and precision. �192 F. SCHEDULE August 1994: Complete survival model development September January 1994: Begin development of simulation model 1995: Initial version of spatial simulation model complete May 1995: Final version of simulation model complete with linkages to survival models June 1995: Simulation runs begin September 1995: Prepare manuscript for publication . G. LITERATURE CITED Cox, D.R. 1984. Analysis of survival data. Chapman and Hall, London, England. Flood, B.S., M.E. Sangster, R.D. Sparrowe, and T.S. Baskett. 1977. A handbook for habitat evaluation procedures. U.S. Fish Wildl. Servo Resour. Publ. 132. 77 pp. Forman, R.T.T., and M. Godron. 1986. Landscape ecology. John Wiley & Sons. New York. 619 pp. Hobbs, N.T., and T.A Hanley. 1990. Habitat evaluation: do use/availability data reflect carrying capacity? Journal of Wildlife Management 54:515-522. Hosmer, D.W., and S. Lemeshow. 1989. Applied logistic regression. John Wiley and Sons, New York. Lee, E.T. 1992. Statistical methods for survival data analysis. John Wiley and Sons, New York. Morrison, M.L., B.G. Marcot, and R.W. Mannan, 1992. Wildlife-habitat relationships: concepts and applications. The University of Wisconsin Press, Madison. 343 pp. Robel, R.J., L.B. Fox, and KE. Kemp. 1993. Relationship between habitat suitability index values and ground counts of beaver colonies in Kansas. Wildl. Soc. Bull. 21:415-421. Van Horne, B. 1983. Density as a misleading indicator of habitat quality. Journal of Wildlife Management 47:893-901. Verner, J., M.L. Morrison, and C.J. Ralph, eds. 1986. Wildlife 2000: modeling wildlifehabitat relationships of terrestrial vertebrates. Univ. Wisconsin Press, Madison. 470 pp. White, G.C., and R.A Garrott. 1992. Analysis of wildlife radio-tracking data. Academic Press, San Diego. Whittaker, R.H. 1975. Communities and ecosystems, 2nd ed. MacMillan Publishing, New York. 385 pp. . �l':!3 Colorado Division Wildlife Research June 30 1994 of Wildlife Report JOB PROGRESS State of Project Work Colorado No. W-153-R-7 Plan No. Author: Mammals lOA Covered: July Research Kit Fox Studies 1 Job No. Period REPORT Kit Fox (Vulpes macrotia) Status in Colorado 1, 1993 - June 30, 1994 J. P. Fitzgerald Personnel: Eussen, J. P. Fitzgerald, M. Link, T. Verbeck, A. Anderson, J. Prather, M. Reddy, D. Watson, T. Beck, B. Gill L. Dent, J. Abstract A total of 1754 trap nights of effort were expended in the project year to capture 18 kit fox. Trapping efforts included parts of Moffat, Mesa, Garfield, Delta, and Montrose counties. All captures were made in Montrose and Delta counties in Peach Valley (8), East of Montrose (9), or North of Delta (1). A reliable report of a "kit or swift" fox was received from u.S. Fish and Wildlife Service and BLM personnal in the Vermillion Creek area in northern Moffat County. Trapping to date in northern Moffat County has not resulted in any fox captures. Monitoring of radio-collared kit foxes in the Peach Valley study area in Montrose and Delta county continued. Four (F1a, F5, M13, Ma) of 6 radio-collared adults are still alive in that population. On 26 March and 13 April carcasses of two kit fox, M12 and an untagged animal of unknown sex, were recovered in T50N, R9W, NE corner Section a within 100 m of each other. The cause of death could not be determined for either animal. East of Montrose, nine kit fox (4 adult females, 1 adult male, 1 male pup, 3 female pups) were captured and either radio-collared or ear tagged. Gill and field crew personnel have taken photographs of animals in that population. In March, the Wildlife Commission closed the season on kit fox and imposed trapping regulations for the Peach Valley area. The Director of the Division of Wildlife placed the species on the state special concern list. ��195 KIT FOX (VULPES MACROTIS) STATUS IN COLORADO James P. Fitzgerald P. N. Objective Document the geographic Colorado. distribution and relative Segment abundance of kit fox in Western Objectives 1. Continue to monitor radio-collared foxes in the Peach Valley Area. 2. Write ~p project results (M.A. thesis - Michelle Link in progress). 3. Begin extensive summer trapping effort. Methods Methods continue to be similar to those reported previously (Fitzgerald and Link 1993, Fitzgerald and Verbeck 1993) .-Field Personnel: At the start of the project year Michelle Link was replaced by Tom Verbeck to take over the kit fox field research. Verbeck worked from mid-August to January. Allan Anderson monitored fox populations in Peach Valley from March through late May. Verbeck resigned from the project in April. Three field crews: Matt Reddy-David Watson; Lonnie Dent-Jim Eussen; and Jake Prather (alone) began work on the project in late May. Results and Discussion Trapping Efforts: Between July 1, 1993 and June 30, 1994 nights of effort have resulted in a total of 20 captures fox. All captures were made in Montrose or Delta County. locations are summarized in Table 1. a total of 1754 trap of 18 individual kit Trapping efforts and In areas with known kit fox populations (Peach Valley, North of Delta, and Montrose East) a total of 1012 trap nights of effort-were expended in trapping the 18 animals captured (56 trap nights/fox). The capture rate of foxes in traps placed near « 200 m) known fox dens is much higher. The 18 individual animals were captured in a total of 45 trap nights, a capture rate of 2.5 trap nights per fox. It appears that the methods we are employing catch kit foxes efficiently if foxes are present. It appears kit fox populations across the areas surveyed are very low resulting in many trap nights of effort with no captures. In comparing this years trapping effort to efforts by Link, she captured 9 kit fox during 2,725 trap nights of effort, with all 9 animals captured during 836 trap nights (93 trap nights/fox) of effort in Peach Valley (Fitzgerald and Link 1993). The SW Regional Office of CDOW has hired Mr. Doug McCauley, a taxidermist from Delta, to work with the project crew to find additional kit fox populations. Mr. McCauley believes that many more kit foxes exist .than our trapping results would indicate. The understanding with Mr. McCaughley is that he will try to locate kit fox populations using predator calling and night lighting. He has to show either the project crew or someone from the SW Regional Office the exact location of occupied dens and verify the presence of kit fox in order to be paid his fee. To date he has not taken the crew out but has indicated to Dent and Eussen that he knows of 2 or 3 active dens on the Colorado-Utah border. We are still trying to schedule a time when he can go out with us. �196 Table 1. Areas study, western searched, Colorado, dates of searches, and trap nights July 1, 1993 - June 30, 1994. Area Searched and County Dates Searched Number of Trap Nights Peach Valley Montrose.Co. 8-29 to 12-6-93 4-5 to 4-24-94 232 139 North of Delta Delta Co. 11-19-93 to 1-6-94 5-27 to 5-30-94 245 39 Montrose Montrose 5-27 to 6-18-94 East Co. captures of effort, kit fox Total Trap Nights and Success by Area 8 o 371/8 foxes o 284/1 fox 357 9 357/9 foxes 60 o 60/0 1 West Gunnison Gorge. Delta Co. 12-14 to 12-16-93 SEW of Wells Gulch. Delta 6-21 to 6-27-94 230 o 230/0 10-5 to 10-12-93 151 o 151/0 10-20 to 10-28-93 129 o 129/0 172 o 172/0 1754 18 1754/18 Brown's Moffat Co. Park Co. Grand Junction NW, Mesa Co. DeBeque/parachute 5-30 to 6-7-93 Mesa/Garfield Co's Totals Trapping and Monitoring Effort - Peach Valley Population: Verbeck from 29 August to 6 December conducted 232 trap nights of effort in Peach Valley resulting in 8 captures of foxes: M13 (captured twice), F14, MIS, M16, M8, F18, and an animal that escaped before it could be sexed or marked (Table 2). Animals which were recaptures (F14, MIS, M8, F18) were given new radio-collars and ear tags. Invariably ear tags were torn out on most of the foxes recaptured. Anderson conducted 139 trap nights of effort in Peach Valley from 5 April to 24 April with no animals captured and had difficulty picking up radio signals from collared animals •.Anderson picked up radio-signals from 5 (F18, M12, M8, F5, M13) of the 6 adults radioed in the area. However, he was unable to locate their precise den sites because of field conditions and problems with radio-receivers. On 26 March, Anderson recovered the carcass of M12 within a hundred meters of its den but could not verify cause of death. He recovered the carcass of a kit fox that had no ear tags or collar about the same distance from that same den on 13 April. Again it was not possible to estimate cause of death. F14 has been missing since January unless she is represented by the carcass (no radio-collar found) located near M12 - they were both trying to pair in December and January. Field crews verified presence of M13, F18, M8 on 19'June in T51N, R9W, 529,30. This is 4-6 Jan north of locations Verbeck found them using in late fall and early' winter (Fitzgerald and Verbeck 1993). They also reported picking up good signals in T50N, R9W, 54 from 150.037 radio - a radio frequency that Anderson attributed to M12 found dead on 26 March. This error has to be investigated and resolved, and attempts made to recapture and recollar an~als in the valley. Field crews report that M13 is with an apparently unmarked female in a den with 4 pups close to the water tank in lower (north) Peach Valley. They have also seen several �197 foxes crossing the road at night in that same part of Peach Valley. be made later this summer to capture and radio those animals. Montrose Attempts will East Population: Anderson on 23 May received word of a family of foxes along the Landfill Road (Bostwick Park Road) in T49N, R8W, NE7 and began to observe animals in early morning and late afternoons. He estimated as many as 12 animals in the area including pups of the year. He showed the summer field crews the site on 27 May. Since 28 Maya total of 9 kit fox have been trapped and marked in that population (Table 3). The population includes several adult females which may represent yearling animals that did not breed. The whelping den included pups from apparently two different litters based on size differences and estimated numbers of pups (7-8). The Montrose East population is located southeast of Flattop Mesa. Habitat is similar to Peach Valley and there is nothing to preclude mixing of the two populations which are about 12 km apart. However, trapping efforts from Flattop Mesa north to upper Peach Valley have not resulted in any other fox captures. The large numbers of females present in the population are encouraging provided sufficient males are available for mating in 1994-9S. Several enthusiastic local residents are aware of this population and watch it carefully attempting to protect animals from being shot since they are located close to a heavily used county road. Table 2. Kit fox trapped in Peach Valley, Montrose 1, 1994 and dates of last sightings/radio-signals. Ear tags replaced. Capture or Date Located Sex S/14/92 , M S/31/92 11/23/92 F 6/2S/92 M 4 2.2 9/28/92 M 2 9/28/92 3/1/93 1/3/94 4/6/94 F S 9/28/92 3/2/93 9/22/93 1/6/94 F 2/23/93 ,10/26/93 12/6/93 1/6/94 3/9/94 S/22/94 6/19/94 F 2/23/93 10/26/93 12/6/93 1/6/94 3/26/94 Ear Tag Number 1 6 ETR14 Weight Kg 2.3 not collared 2.0 lS0.238 ETR18 M 8 Location Fate TS1N/R9W/S29 Unknown TS1N/R9W/S29 T1SS/R94W/S27 .coyoce not collared TSON/R9W/S16 Unknown 2.7 lS0.309 TSON/R9W/S9 Unknown 2.S lS0.338 CR1S0.9S6 2.4 2.6 7 Radio Collar Frequency County, Colorado, 1992-July CR=Collar replaced; ETR = lS0.379 CR1S0.8S0 CR1S0.889 2.S lS0.638 2.6 CR1S0.S94 3.0 lS0.468 3.1 CRlSO.la9 TSON/R9W/S9 TSON/R9W/S17 TSON/R9W/S17 TSON/R9W/S9 Killed Signal Weak TSON/R9W/S9 TSON/R9W/S9 TSON/R9W/S16 TSON/R9W/S17 Unknown TS1N/R9W/S29 TSON/R9W/SS TSON/R9W/SS TSON/R9W/SS TSON/R9W/S16 TSON/R9W/S18 TS1N/R9W/S29 Alive TS1N/R9W/S29 TSON/R9W/SS TSON/R9W/SS TSON/R9W/SS TSON/R9W/SS by �198 Table 2 cont. TSON/R9W/SS TS1N/R9W/S30 S/21/94 6/19/94 4/21/93 9/24/93 10/23/93 1/6/94 3/26/94 9/20/93 11/9/93 12/14/93 3/31/94 5/21/94 6/19/94 9/30/93 11/22/93 M 12 ETR15 2.S 2.8 lS0.708 CR1S0.037 M 13 2.7 150.813 M 16 2.3 150.499 Alive TSON/R9W/S22 T50N/R9W/S8 T50N/R9W/S17 T50N/R9W/S16 T50N/R9W/S8 T50N/R9W/S16 T50N/R9W/S16 T50N/R9W/S16 T50N/R9W/S5 T50N/R9W/SS T51N/R9W/S29 Alive T50N/R9W/S17 T50N/R9W/S17 Unknown Dead Table 3. Kit foxes trapped in the Montrose East population, Montrose County, females not pups of the year; L = lactating. 1994. NL = non-lactating capture or Date Located 5/29/94 5/30/94 Sex Ear Tag Number 21 F(NL) recaptured Weight Kg 2.5 Radio Collar Frequency 150.947 Location Fate T49N/R9W/S7 Alive 5/29/94 F(NL) 22 2.5 150.403 T49N/R9W/S7 Alive 5/29/94 F(NL) 23 2.25 150.338 T49N/R9W/S7 Alive 5/29/94 M pup 24 1.7S not-collared T49N/R9W/S7 Alive 5/30/94 F pup L176/R177 1.5 not-collared T49N/R9W/S7 Alive 5/30/94 F pup L178/R179 1.45 not-collared T49N/R9W/S7 Alive 5/31/94 6/7/94 F (L) L180/R181 recaptured 2.3 T49N/R9W/S7 T49N/R9W/S7 Alive 6/1/94 F pup L183/R184 1.8 not-collared T49N/R9W/S7 Alive 6/3/94 M Ad 3.0 not-collared T49N/R9W/S12 Alive L185/R186 150.940 Use of Baits: Table 4 summarizes the baits used in traps which captured kit fox. Verbeck made a detailed analysis of what baits were used during his 817 trap nights of effort and we are following that procedure this field season. Anderson conducted 139 trap nights of effort in Peach Valley with no successful captures during the month of April. During that period he tried a variety of baits including commercial fox scent, sardines, commercial raccoon-fox scent, and dead chicks. He also put out carcasses of dead white leghorn chickens near several dens in the Valley with foxes paying little attention to such carrion. Since late May crews have conducted an additional 798 trap nights using a variety of baits. During �199 trapping efforts at East Montrose, emphasis was given to taking as many animals as possible in as few a trap nights as necessary. As a result of that effort many traps were baited with multiple baits consisting of road killed cottontails and/or prairie dogs and turkeys with the idea being that the foxes might respond best to a variety of fresh killed meats, a technique that worked well. Although baits from local road-killed mammals and birds might work better than turkeys or turkey-lure combinations the numbers of traps being run precludes exclusive use of such local carrion. . Table 4. Summary of kit fox captures/recaptures 1992 to June 30, 1994. Investigator/Year and types of baits used. May 14, Verbeck I Link/Parmetet,r 92 to July 93 Aug 93 to Jan 94 Present Crews April - Present Baits Cottontail Cottontail/Turkey Dog. Cottontail/Turkey/po Prairie Dog/Turkey Prairie Dog Turkey Turkey/Commercial Lure Pheasant Turkey/Big Game Commercial Lure or w Sardines Chicken/Commercial Lure Total Trap Nights Effort 1/<50 5/77 3/41 5/70 0/39 4/10 0/6 6/>2,600 0/408 2/679 2/<200 1/32 0/66 0/100 0/39 2,725 817 937 Other Activity: Based on recommendations made by Fitzgerald and presented to CDOW personnel 11/15/93, the Colorado Wildlife Commission on 3/10/94 closed the season on kit fox in Colorado and imposed trap and trapping restrictions for the Peach Valley, area. The Director of the CDOW on 22 March, placed the kit fox on the list of species of "special concern." The agency is presently contemplating future management, including whether it should be listed as "threatened." A meeting of CDOW personnel and the contractor will take place in late summer 1994 to discuss these issues. Link is finishing her thesis incorporating some of Verbeck's findings into the paper. Gill has photographed animals in the Montrose East group. ��201 Colorado Division Wildlife Research July 1994 of Wildlife Report JOB PROGRESS State of Project Work Colorado No. 11A Author: Research Multi-Species 1 Job No. Covered: Mammals W-153-R-7 Plan No. Period REPORT Investigations Predicting the Impacts of Environmental Change: Simulations of Genetic and Species Diversity at Landscape and Regional Scales July 1, 1993 - June 30, 1994 N. T. Hobbs, J. Miller, and J. A. Wiens Abstract Loss of intact, natural habitat is the foremost threat to wildlife diversity in Colorado and the West. Historically, the prevailing source of habitat loss in the Rocky Mountain region has been harvest of natural resources (e.g., logging, mining, and agriculture). However, changing demographic and economic trends will drive a fundamental shift in the source of environmental change affecting wildlife habitat. As a result of these trends, residential development is likely to become the predominant human influence on the diversity of Colorado's wildlife during the coming decade and beyond. It follows that protecting wildlife habitat will depend in a pivotal way on developing wise policy on land use in the places where people live. To foster such policy, we propose to develop a System for Conservation Planning (hereafter, SeoP). The initiating idea of SCoP is that the success of habitat protection will depend on offering wise alternatives for land use, alternatives that meet human needs for economic vitality as well as the needs of wildlife for natural landscapes. Providing information needed to develop these alternatives is the primary goal of the SCoP project: The goal of SCoP is ~o ob~ain, assemb~e, and dis~ribu~e s~a~e of ~e ~ infor.ma~ionon effec~s of ~and use on w~d~ife diversi~y,p~icu~ar~y ~and use associa~edwi~ residen~ial expansion in Co~orado and ~e Wes~. Meeting this goal requires that we enhance access to current knowledge needed to support decisions on land use while we simultaneously strive to improve that knowledge. To this end, SCoP will initiate three efforts. First, we will make information more readily available to decision makers by developing an expert system that forecasts effects of changes in land use on wildlife diversity. The expert system will be composed of a geographic information system, a generalized population dynamics model, a database of vertebrate life history characteristics, and a user interface. Second, we will initiate a Citizen Inventory Program to enhance data on the distribution and abundance of Colorado's wildlife species. In this program, we will enlist, train, and organize citizen volunteers to conduct rigorous, scientific monitoring of wildlife populations and their habitats. Third, we will initiate a program of extramurally funded research that strives to understand processes regulating wildlife distribution and abundance in human dominated ecosystems. SCoP will coordinate these three initiatives in a cogent, tightly linked program designed to meet the challenges of the next century in preserving protecting the vital wildlife resources of Colorado and the West. and ��203 Predicting the Impacts of Environmental Change: Simulations of Genetic and Species Diversity at Landscape and Regional Scales P. N. Objective 1. Develop analytical tools to support decisions on management of wildlife diversity in Colorado. These tools will include simulation models and research results that predict consequences of changes in land cover types and land use for maintaining wildlife diversity. Segment Objectives 1. Complete preparation of a study plan for simulating effects of changes landscape pattern on vertebrate species diversity in managed ecosystems. 2. Develop spatially vertebrate wildlife. explicit simulation 3. Begin model experiments to examine diversity of vertebrate wildlife. model of meta-population role of spatial pattern dynamics in controlling Results We prepared 2 study plans to implement appended as reference. the above objectives. in copies are of ��205 Appendix A SCoP: A System for Conservation Planning Project Proposal Mammals Research. Colorado Division of Wildlife Habitat Resources. Colorado Division of Wildlife Natural Resource Ecology Laboratory. Colorado State University N. T. Hobbs J. E. Gross J. R. Miller D. Malkinson R. B. Gill D. L. Schrupp March 1, 1994 �206 PROJECT SUMMARY Loss of intact, natural habitat is the foremost threat to wildlife diversity in Colorado and the West. Historically, the prevailing source of habitat loss in the Rocky Mountain region has been harvest of natural resources (e.g., logging, mining, and agriculture). However, changing demographic and economic trends will drive a fundamental shift in the source of environmental change affecting wildlife habitat. As a result of these trends, residential development is likely to become the predominant human influence on the diversity of Colorado's wildlife during the coming decade and beyond. It follows that protecting wildlife habitat will depend in a pivotal way on developing wise policy on land use in the places where people live. To foster such policy, we propose to develop a System for Conservation Planning (hereafter, SCoP). The initiating idea of SCoP is that the success of habitat protection will depend on offering wise alternatives for land use, alternatives that meet human needs for economic vitality as well as the needs of wildlife for natural landscapes. Providing information needed to develop these alternatives is the primary goal of the SCOP project: The goal of SCOP is to obtain., assemble. and distribute state of the art information on effects of land use on wiJdlife diversity, particularly land use associated with residential expansion. in Colorado and the west. Meeting this goal requires that we enhance access to current knowledge needed to support decisions on land use while we simultaneously strive to improve that knowledge. .To this end, SCoP will initiate three efforts. First, we will make information more readily available to decision makers by developing an expert system that . forecasts effects of changes in land use on wildlife diversity. The expert system will be composed of a geographic information system, a generalized population dynamics model, a database of vertebrate life history characteristics, and a user interface. Second, we will initiate a Citizen Inventory Program to enhance data on the distribution and abundance of Colorado's wildlife species. In this program, we will enlist, train, and organize citizen volunteers to conduct rigorous, scientific monitoring of wildlife populations and their habitats. Third, we. will initiate a program of extramurally funded research that strives to understand processes regulating wildlife distribution and abundance in human dominated ecosystems. SCoP will coordinate these three initiatives in a cogent, tightly linked program designed to meet the challenges of the next century in preserving and protecting the vital wildlife resources of Colorado and the West. TABLE OF CONTENTS NEED................. ..•......................................................... GOAL ....•....................................................................... ORGANIZATION OF PROPOSAL .... ,.................................................. TASKS ..........•.................................................................. Expert System ........................................•..............•.••......... Geographic Information System . . . • . . . . • • • . . . . . . . . . . . . . . . . . . . . . . . . . . • • . . . Generalized Vertebrate Population Dynamics Model. . . . .. . . . . . . . . . . . . . . . . . . . . . Species Database • . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . • . • . . . . . . . . . . . User Interface . . . . . . . . . • . . . . . . . . • • . . . . ... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A Citizen's Wildlife Inventory Program .; .••..........................•...... , Monitoring scope and scale ......•...............•..............................•... Implementation . • • . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Research on Human Dominated Ecosystems .•.......................................... PROJECT IN'I'EGRATION . . . . . . . . . . . . . . • . . . . . . . . . . . . • . . . . . . . . . . . . . . . . . . . . . . . . RELEVANCE TO THE LONG RANGE PLAN OF THE COLORADO DIVISION OF WILDLIFE RELATIONSHlP TO CURRENT PROGRAMS ...............................•.............. DEMONSTRATION PROJECTS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A NATIONAL FUNDING INITIATIVE ..... ; . . . . . . . . . . . . . . . . . . . • . . . . . . . . .. . . . . . . . . . PROJECT ORGANIZATION AND SCHEDULE .................•......•.................... BUDGET •••...•............•..........................•........•••..•.............. LITERATURE CITED •...•..••...........................•...........•...•..........•. . . . . . . . . . . . . . . . . . . . . . . . . . ... .•. ... ... ••. . . . ... . . .. . . . . . . • .. . . . . . . . .. . . . . . • . .. .. .. .. .. .. 1 4 6 6 6 7 7 8 8 9 9 10 11 13 14 16 17 17 19 21 23 �207 seop: A System for Conservation Planning NEED The mission of the Colorado Division of Wildlife is to perpetuate the wildlife resources of the state and to provide people the opportunity to enjoy them. Fulfilling this mission depends in a fundamental way on protecting and enhancing wildlife habitat. The importance of protecting habitat is revealed by noting that 30~60 square miles of land in Colorado are annually developed for use by people (Soil Conservation Service 1989, Rennicke 1990). Such use includes economic activity (mining, agriculture, logging) as well as expansion of housing, roads, and services. These changing land uses have emerged as the prevailing influence on wildlife diversity in Colorado (Armstrong et al., in press) because development of land to serve people often conflicts with the habitat needs of wildlife. Resolving this conflict requires that we achieve workable compromises between human needs for economic well-being and the needs of wildlife for natural landscapes (Baharav et al. 1991). Expansion of urban and suburban areas in Colorado will accelerate as a result of demographic and technological changes during the coming decade. By the year 2000, retirements of the baby boom generation will begin in ernest (Colorado Department of Health 1993). These retirements will allow a major component of the nation's population to migrate from one region to another. In addition, the growth of the service sector in the nation's economy and the development of telecommunications that allow rapid exchange of information across long distances will permit many employees to work where they choose (Riley 1993). Freed from the need to be close to markets, factories, and organization headquarters, these employees will select their work place based on quality of life (Riley 1993). As an illustration of these trends, consider the economy of Montana Sixty-four percent of Montana's work force is employed in the service sector while only 7% work in industry and only 4% in agriculture (Snow 1994). Employment provides 60% of the state's personal income, while an astonishing 40% accrues from "unearned" sources such as retirement benefits, investments, and pensions (Snow 1994). These trends are significant because they illustrate the increased opportunity for regional mobility in the nation's population and the concomitant rise in the importance of the service sector of the economy. In such an economy, environmental values offered by states like Montana and Colorado become a tangible economic asset (Randall 1986), an asset capable of attracting people and businesses seeking alternatives to the congestion and environmental decay of the population centers on the east and west coasts. However, it is also clear that the quality of life that attracts people to the West is put at risk by a rapidly growing population. Changes in economies to favor service industries and changes in demographics to favor people who are supported by investment and pension income will produce a qualitative shift in the causes of environmental change in Colorado and the West (Fig. 1,2). Historically, the most. important source of loss of wildlife habitat has been harvest of natural resources (Fig. 1)(Harris 1984, OTA 1987, U.S. Department of Commerce 1989, Cairns and Lackey 1992). Agriculture, logging, :mjning, and water development cause habitat loss as a result of the removal of commodities (timber, crops, minerals, water) from natural landscapes. The effects of these activities are intensive because they force type conversions; areas of land surface are changed from one cover type to another. Such large scale conversion has given rise to the prevailing metaphor of conservation biology--habitat patches are like islands in a sea of human disturbance. (e.g., Ambuel and Temple 1983, Harris 1984, WJlcove 1985, Saunders et aL 1991, Murphy and Noon 1992). However, a different source of disturbance, a source that is more extensive and less intensive, looms large in the future of Colorado. Expansion of residential communities to accommodate regional shifts in population will play an increasingly important role in determining environmental quality in Colorado and the West (Fig. 1,2) (Fitzgerald et al., in press). Such expansion does not necessarily 1 �208 50 Colorado 40· 30 20 10 til 0) -c:c: ••• ~ w -- o 1942 1947 1952 1957 1982 1987 1972 1977 1982 1987 100 80 Montana CD ~ eo ••• c, 40 > - c 20 c: CD o o ••• CD 1942 1947 1952 1957 1982 1987 1972 ;977 1982 1987 Il. 80 50 Rocky Mountain Region 3D 20 10 o 1~1~1~1~1~1~1~1~1~1~ Year • !§ill Services Natural Resources Figure 1. The economic contribution of industries based on harvest of natural resources is declining rapidly in the West, while the contribution of service-based industries is rising. This means that urban and suburban expansion are likely to replace commodity-driven industries as the primary source of habitat loss in the West. 2 �209 40 e E 0 CJ c 30 ~ e c as e c 0- - 20 ~ c m CJ 10 0- m c, o 1942 1947 1952 1957 1962 1967 1972 1977 1982 1987 o Year Rocky Mtn. Region !=tIIIII Colorado _ Montana Figure 2. The rise in unearned income from pensions and investments reflects increasing regional mobility. People supported by such income are able to choose their residence solely on the basis of quality of life, independent of economic constraints. The environmental values of the West are likely to attract many people seeking relief from the dense populations on the east and west coast. Thus, those values are put at risk by a rapidly growing population. force type conversions. This is the case because houses can be built in forests, shrublands, or meadows without converting those cover types to pavement. Thus, while historical sources of disturbance resulted in patches of habitat embedded in a matrix of disturbed land, future disturbance is more likely to produce patches of disturbance (i.e., residential development) embedded in a matrix of habitat (Fig. 3) (Batty 1991, Hobbs and Miller 1994). Changes in the forces driving habitat loss simultaneously offer a problem and an opportunity. Residential development is a problem because its effects can be much more eaensiue than the effects of commodity harvest, particularly in arid areas of the West where agriculture is infeasible. However, a shift from commodity driven disturbance to disturbance driven by settlement offers an opportunity because the impacts of residential development, if properly managed, can be less intensive than impacts of commodity harvest (Stenberg and Shaw 1986, Hobbs and Miller 1994). More importantly, there are many alternatives for how settlement can proceed, and these alternatives provide opportunitieS for managing growth (Adams and.Dove 1989, Gordon 1990, O'Connell and Noss 1992, Majeske 1993). As a society that values environmental quality, we are obliged to seize these opportunities. 3 �210 ~ ~ 8 A Ii- fi. ~ Ii- ? f:;. ~ B Figure 3. Harvest of natural resources (e.g., logging) often produces fragments of habitat in a matrix of disturbance because harvest causes conversion of one cover type to another (A). Residential development can produce fragments of disturbance in a matrix of habitat because no type conversion occurs (B). GOAL Here, we propose to develop a System for Conservation Planning (hereafter, SCoP-pronounced ."scope"). Our goal can be described as follows: The goal of the SCoP project is to obtain. assemble. and distribute state of the art information on effects of land use on wildlife diversity, particularly land use associated with residential expansion in Colorado and the West . We envision that SCoP will be used by citizens, politicians, planners, and policy makers to evaluate impacts of alternative land uses on wildlife. SCOP will be used to enhance the wisdom of a variety of policies, including zoning regulations and designs for residential developments. SCOP will be used to plan private and public investment in nature reserves, open space, parks, and corridors. (e.g., McHarg 1971, OTA 1987, Soule 1991, Beier 1993). 4 �211 SCoP is needed for a several reasons. Habitat protection has traditionally been synonymous with opposition to development, Such opposition leads inexorably to conflict between people who value environmental quality and people who promote economic growth. We believe that such conflict is mutually defeating. SCoP offers an alternative to this adversarlal relationship-we accept a priori that humans are an integral part of ecosystems and that human welfare will always be a significant dimension of land use policy. We believe that progressive policy on land use will emerge if and only if we are able to offer wise alternatives for development, alternatives that meet human needs as well as the needs of wildlife (Table 1). We believe that development of sound environmental policy begins by fostering cooperation between those who promote conservation of environmental values and those who promote economic vitality . .It follows logically from these ideas that SCoP will focus on the ecosystems where people live. This focus is important because the loss of wildlife habitat at large scales is often the result of many small scale decisions on land use. This is to say that the 60 square miles of development that occurs each year in Colorado (Rennicke 1990) is the outcome of thousands of decisions on the uses of 10's of acres of natural landscapes. These decisions, made one at time, appear eminently reasonable. Unfortunately, we are learning that the collective result of these many small scale decisions is clearly unreasonable-collectively they produce a slow but certain impoverishment of the biota of natural ecosystems <Lovejoy 1986).. Coloradans urgently need tools that can place these small scale decisions on land use in a larger, more ecologically complete perspective. Because environmental values can be viewed as an important economic asset, such understanding is needed not only by environmentalists, but also by businesses, by investors, and by industry. The focus of SCoP on human dominated ecosystem has an additional motivation. It is our view that the success of habitat protection depends on the extent to which people value wildlife. Fostering such values, in turn, depends on offering abundant, diverse, and meaningful opportunities for citizens to enjoy and understand wildlife communities. As we shall show, SCoP will offer such opportunities close to home. Table 1. Principles that will guide the SCoP project. • We will be successful in protecting habitat only to the extent that we are able to offer wise alternatives for development, alternatives that meet human needs as well as the needs of wildlife. • One of the most important barriers to wise decisions on land use is inadequate access to current information on the processes that control wildlife diversity in human dominated ecosystems. • WIldlife diversity must be made real and tangible by allowing people the chance to experience a diverse fauna. If this doesn't happen, then diversity will likely remain an abstract, academic concept, a concept that plays an insignificant role in influencing land use decisions. • In a democratic society, the only political force stronger than the stakeholder is the informed stakeholder, • The future wisdom of land use decisions depends on current investment in inventory and monitoring of wildlife populations. • Ecological research will be most meaningful to land use policy if we strive to improve our understanding of the processes that control wildlife diversity in the places where people live. • The Colorado Division of Wildlife, operating in vital connection with Colorado Universities, can attract meaningful funding from national foundations if we are willing to aggressively attack environmental problems of national importance. 5 �212 ORGANIZATION OF PROPOSAL This proposal will address three barriers that impede development of wise policy on land use. The first barrier is poor access to current information on the ecological processes that control wildlife diversity. To overcome this obstacle, we will develop a system for decision support. The second impediment is insufficient data on the distribution and abundance of wildlife. To meet this challenge, we will initiate a Citizen's Wildlife Inventory Program. The third barrier is an insufficient understanding of the processes that control wildlife distribution and abundance in the places where people live. To enhance such understanding, we will conduct research on wildlife diversity in human dominated ecosystems. In the next section, we describe each of these initiatives individually and we then discuss how the individual parts connect to form a comprehensive, cogent program. We subsequently describe how SCoP will contribute to achieving the goals of the Long Range Plan of the Colorado Division of Wildlife. We then sketch how SCoP will relate to other, ongoing efforts in habitat and wildlife protection. Next, we propose a pilot, demonstration project and we outline a national fund raising initiative to support expansion of SCoP during and beyond the demonstration phase. Finally, we describe project organization, scheduling, and budget. Throughout the document, we will offer principles that we believe are fundamental to conserving wildlife diversity in ecosystems where human welfare is an important dimension of policy (Table 1). TASKS SCoP will consist of 3 components: an expert system for decision support, a citizen based inventory and monitoring regime, and a program of research (Fig. 4). Expert System We will build an expert system (Coulson et al. 1987, Coulson et al. 1991, Steyaert 1993, Fedra 1993) to support decisions on land use by urban planners, developers, and local governments. To explain this system, we begin by defining the term "analysis area". Hereafter, we will use this term to mean a locality where human habitation and associated activities influence wildlife populations in a significant way. Thus, an analysis area might be a city and its surroundings (e.g., Durango), it might be a county (e.g., Summit) or it might be geographic area transcending several political Human boundaries (e.g., the Roaring Fork Valley). Dominated The analysis area sets boundaries on the use of the expert system, described below. The expert system will consist of 4 parts: a raster-based geographic information system (GIS), a generalized population dynamics model, a relational database of vertebrate life history characteristics and inventory data, and an interactive userFigure 4. The SCOP project will consist of 3 interface (Fig. 5). components: an expert system, an inventory regime, and a research program. 6 �213 Generalized Species Database Population Dynamics Model IIIIIII~&~~ CS~dJ 1 User ~D l-fuEifriJ /<) I I I I i I ! iii \ ","" Figure 5. seoP's expert system will be composed of a geographic information system, a population simulator, and a database of species life history characteristics and inventory results. 1. Geographic Information System: Any system for supporting decisions on land use must be founded. on spatial data <Dueker and Delacy 1990). Spatial data provides an accurate description of the current landscape and thus offers a necessary first step in forecasting effects of land-use change. For each analysis area, we will build a GIS containing data layers on land ownership, land use, vegetation types, slope and elevation, transportation networks, and hydrology. For a given analysis area, we will obtain information needed to assemble these layers by reviewing public records and through the use of remotely-sensed data. We will use existing satellite imagery (Landsat Thematic Mapper) for coarse-grained analysis and to put an analysis area in a geographic context, and 1:40,000 color infra-red aerial photographs (NAPP) for analysis at finer scales. This imagery allows discrimination among landscape elements that is not possible with conventional photographs (1:40,000 photos can be enlarged to permit identification of individual buildings). Spatial data will be ground-truthed, using a global positioning system, and subsequently digitized using widely accessible GIS software (e.g., Idrisi, GRASS, or ArcjInfo). 2. Generalized Vertebrate Population Dynamics Model. Spatial data will be used to drive a population simulator. During the last decade, ecologists have invested much effort in building models that represent effects of changes in the spatial arrangement of habitats on the long term viability of wildlife populations (e.g., Fahrig et aL 1982, Gilpin 1987, Fahrig 1988, Urban and Smith 1989, Hansen et aL 1991, Pulliam et aL 1992, Doak et al. 1992, McKelvey et aL 1993, Hastings in press). These models have been widely used in formulating and justifying land management policy (Thomas et aL 1990, Burgman et aL 1992, McKelvey et aL 1993, Beier 1993). However, the insights of these models are limited because each model focuses on a single species, or, at best, on a small group of species. 7 �214 SCoP will borrow from these approaches while expanding their predictive power. To do that, we will build a generalized population simulator (Hobbs et al. 1993) linking spatial data on habitat types (see above, "Geographic Information System") to data on the life history characteristics and habitat requirements of a variety of vertebrate species (see below, "Species Database"). The simulator will predict how changes in land use will affect the population trajectories of vertebrate species endemic to an analysis area. We will use Mote-Carlo techniques (Burgman et al. 1992, Burgman et al. 1993) to assess how predicted changes in population trajectories will affect species viability under a range of land use alternatives. In essence, we will use the model (Hobbs et al. 1993) to answer the question. "How will a given change in land use affect the ability of a landscape to sustain wildlife diversity?" 3. Species Database: We will build a database assembling life history information on vertebrate species endemic to an analysis area. Important fields in the database will include taxonomic classification, body mass, life span, age of first reproduction, average litter/clutch size, juvenile survival rate, dispersal capability, relationships to vegetative cover types, confirmed locations in the analysis area and other inventory data, and statewide population status. There will be species for which such data are not currently available. In these cases, we will estimate missing data values using allometric relationships derived from previous, multi-taxa studies (Fenchel 1974, Blueweiss et al. 1978, Lindstedt and Calder 1981, CluttonBrock and Harvey 1983, Peters 1983, Hennemann 1983, Schmidt-Nielsen 1984, Lindstedt and Boyce 1985, Lindstedt et al. 1986, Millar and Hickling 1991, Charnov 1992, Cotgreave and Harvey 1992, Currie 1993). 4. User Interface: Information, no matter how reliable and revealing, cannot support decisions unless decision makers have access to that information. Thus, the second principle of SCoP is that we can enhance the wisdom of land use decisions by enhancing access to state of the art information (Table 1) while we simultaneously strive to improve it. We have been highly successful in the past in making information accessible to decision makers by creating interactive computer programs that represent the consequences of ecologicalprocesses relevant to management decisions (Hobbs 1989,·Hobbs et al. 1990, Hobbs and Miller 1990). These programs shield the user from the myriad of scientific details that are necessary to portray an ecological process, while simultaneously allowing that user to manipulate aspects of the process that are necessary for understanding its consequences. Thus, a well designed user interface portrays the details relevant to a decision and hides those that are not. We will develop an interface that allows decision makers to examine the consequences of a given shift in land cover or land use. Users will be able to input scenarios for changes in density of housing, construction of roads, creation of open space and natural areas, water impoundments, industrial development, shopping centers, etc. We will strive to create the simplest feasible means of input. For example, we will make it possible for users to interact directly with the GIS by drawing polygons on a map displayed on a video terminal (using a mouse) and by choosing land uses from a menu. Thus, the interface will allow user-specified polygons to represent a given land use change. In response to user input on alternative land uses, the expert system will produce easily understood statistics and graphs summarizing the best, currently available projections (sensu Caswell 1989) of the effects of a given land change on the viability of all species of vertebrates that could reasonably be expected to inhabit an analysis area. These statistics will be summarized by body mass, guild, and trophic group to reduce the volume of output, but individual species projections will always be available. In addition, the expert system will 8 �215 identify changes that might be expected to impact threatened or endangered species. We will assure that all output produced by the expert system is simple, attractive, and revealing. The system will respond to user-questions on the logic and data responsible for its projections. We will assure that the mechanisms that produce a given prediction can be described in response to queries by users. Thus, the expert system will not only predict what will happen under a given set of input values, it will also produce output explaining the basis for its prediction. By so doing, SCoP's expert system becomes an educational instrument as well as an instrument of policy analysis. It provides a way for citizens to understand their impacts on the ecosystems where they live (see section below "Relationship to Long Range Plan"). It _. _ illuminates fundamental principles of conservation biology. These four components (Fig. 5) will act in concert to increase the accessibility of relevant information to citizens, planners, politicians, developers, and policy makers. It should be firmly understood, however, that the objective of the expert system is to augment professional judgement, not to replace it. Our system is simply a tool for bringing the best available ecological data to bear on a local decision. It is simply a tool for exploring the consequences of different scenarios in land use. It is not a surrogate for human decisions. A Citizen's Wildlife Inventory Program: Systems for decision support like the one described above lean heavily on ecological theory, but the credibility of such systems ultimately depends on empirical data. Consequently, the fifth principle of SCoP is that the future wisdom of land use decisions depends on current investment in inventory and monitoring of wildlife diversity in an analysis area (Table 1). Monitoring scope and scale Biological monitoring and inventory provides the foundation for conservation of wildlife diversity (Spellerberg 1991). Effects of habitat fragmentation or loss are frequently reflected by declines in abundance and distribution of species. When these trends occur slowly, as they often do, they will only be detected by data collected over long periods, data that can distinguish population trends from year-to-year fluctuations caused by weather or other non-human influences (Burgman et aL 1993). Such data is essential to the development of informed management plans to assure the persistence of a diverse fauna. To be worthwhile, such monitoring must occur at relatively large scales and must be sustained over extended time periods (Verner 1983, Clark et al. 1988, Noss 1990, Spellerberg 1991). In essence, we must design monitoring systems that" illuminate the "big picture." Monitoring at large scales and over long time intervals is difficult because it is expensive. A related problem is that monitoring efforts, if they are insufficiently sensitive, may fail to detect wildlife responses to environmental change until it is too late to do anything about those changes. Consequently,. we need monitoring regimes that are sufficiently intensive to detect relatively short term changes, but that are sufficiently extensive to reveal trends over wide areas on a sustained basis. It follows that the monitoring program must be sensitive over different temporal and spatial scales. For some species, it will be sufficient to simply document their occurrence and relative abundance. For other species, it will be important to collect information useful for determining animal density, reproduction, or absolute population numbers. Data from the" citizen monitoring program will permit much more extensive and up-to-date analysis of trends in populations and communities than is currently possible. g" �216 Implementation To implement the above goals, SCoP will organize a Citizen's Wildlife Inventory Program (Fig. 6). The program will invest in rigorous, scientifically credible monitoring of wildlife populations by citizens. River Watch Adopt a Hig,way. Citizen's Wildlife We will recruit citizens to the monitoring Inventory process for three reasons. The most obvious Program motivation, but the least important one, is to reduce costs=monitoring is expensive largely as a result of the H.nter EdJcation costs of labor. A far more significant rationale is that citizen involvement offers an opportunity for the people to experience, enjoy, and understand their Figure 6. The Colorado Citizens Wildlife wildlife. Thus, the third principle of SCoP is that Inventory Program will borrow elements from wildlife diversity must be made real and tangible to existing successful initiatives: River Watch, people by allowing them the chance to experience a Hunter Education, and Adopt a Highway. diverse fauna (Table 1). If this doesn't happen, then diversity will likely remain an abstract, academic concept, a concept that plays no significant part in influencing land use decisions. WIldlife diversity becomes tangible to people if they can experience and understand it. Monitoring offers a way to do that. The final motivation is that involving citizens in data collection gives them a stake in how that information is used. If citizens are responsible for obtaining information on wildlife, if they understand why that information is important and what it means for preserving wildlife diversity, then they will have a vested interest in its use. SCoP will give citizens tools to assure that data are translated into policy and action. The Citizen Wildlife Inventory Program will borrow features of River Watch, the Hunter Education Program, and the Adopt A Highway Initiative (Fig. 6). Here is how it will work. In each analysis area, we will use SCoP's geographic information system to identify a variety of habitat patches varying in size from 10's to 1000's of ha. As in the Adopt a Highway Initiative, we will seek out local groups concerned about the environment (e.g., high school and university classes, Audubon and Sierra Club chapters, sportsman's organizations, senior citizens groups) and offer them the responsibility for monitoring specific areas. The idea is that each group will "own" the enterprise of monitoring wildlife populations on a specific habitat patch. Initially, the DOW will train individuals in the relevant data collection techniques, will provide the necessary field equipment (binoculars, traps, nets, data collection forms, species guides, etc.), and will give strong technical support during monitoring efforts, Eventually, we will train citizen instructors (as we now train Hunter Education Instructors). These citizen instructors, in tum, will educate people on inventory techniques and on how to interpret findings. Throughout the process, the DOW will assure that efforts are properly timed, coordinated, and standardized. As in River Watch, we will provide computer software and hardware for the storage and communication of results to and from a central server. The species database will be updated regularly via phone network to that server. Computers given to inventory teams will house the SCoP expert system and species data base. The DOW will train citizen teams in the use of this software. Such training will allow citizens to interact with the expert system (described above) in the same way that decision makers do. Thus, the people will have the same opportunity as their governments to explore how changes in land use might affect wildlife populations and habitats. Moreover, SCoP's expert system will allow each inventory team to view the findings of other groups in the analysis area. Thus, each inventory team will be able 10 �217 to project the fate of individual wildlife populations and communities locally and throughout the analysis area. It follows that citizens will see the fine scale picture of wildlife diversity on the habitat patch they monitor and they will see the large scale picture from SCoP's expert system. This leads to the fourth principle of SCoP--in a democratic society, the only political force stronger than the stakeholder is the informed stakeholder (Table 1). It can be said, of course, that volunteers will not be forthcoming. In response, we argue that if citizens are willing to pick up garbage along a highway, then there is reason to wager that they will be willing to invest in a sustained way in gathering data on wildlife. It can be said, of course, that the participation of "amateurs" will not provide reliable scientific data and that such enterprises should, be left to the "professionals" (Robbins et al. 1986, Terborgh 1989). In response, we point to a wealth of irreplaceable data (e.g., Christmas Bird Count, Breeding Bird Census, Breeding Bird Survey, Terborgh 1989) that were obtained by volunteer citizens. Despite their limitations, these data reveal population trends over large areas and long time spans (Temple and Wiens 1989) and have proven useful to many researchers in conservation biology (e.g., Whitcomb et al. 1981, Hansen and Urban 1992). SCoP will strive to improve such efforts by choosing inventory techniques appropriate for use by citizens and by providing sustained investment in citizen training (Temple and Wiens 1989, Terborgh 1989, Miller et al. 1994). We also point out that wildlife management outside of the U.S. often relies on citizen participation in a fundamental way. For example, the entire enterprise of wildlife management in Sweden is conducted by non-professional users. In Sweden, citizens do the monitoring and data analysis; they formulate harvest recommendations. 1 Research on Human Dominated Ecosystems The discipline of conservation biology strives to understand the ecological, genetic, and demographic processes that determine the persistence of wildlife populations and, hence, sustain the diversity of wildlife communities. It can be reasonably argued that despite tremendous investment in research in conservation biology, we understand relatively little about how to practice conservation where it matters most, in the places where people live (but see Leedy et al. 1978, Stenberg and Shaw 1986, Adams and Dove 1989, Goode 1989, Gordon 1990, Baharav et al. 1991, Soule 1991, Beier 1993, Majeske 1993). We believe this is because ecologists have traditionally avoided human dominated ecosystems as locations for study. To support this belief, we recently reviewed papers published in seven, major ecological journals2 during the last three years. We examined how many studies were conducted in areas where human habitation was a major environmental influence vs. those that were conducted in areas where human habitation was largely absent. Over 97% of the hundreds of field studies we reviewed were conducted in areas where human settlement and associated activities played an insignificant role in the ecological process under study. In most of these cases, the study area was specifically chosen so as to be remote {rom human impacts. These findings amplify the conclusions of Cronon (1993) that "Ecologists continued to view the human modification of ecological systems mainly as a source of disturbance, a factor to be controlled and eliminated from scientific analysis if they had to acknowledge it at all." 1 These Swedish "amateurs" sustain annual harvests of 50,000-70,000 moose. 2 Ecology, Ecological Applications, Journal of Wildlife Management, Animal Ecology, Conseniation Biology, Biological Conservation 11 Journal of Applied Ecology, Journal of �218 The knowledge of ecologists has been disconnected from many aspects of environmental policy simply because researchers have failed to view habitation by man as an integral part of ecological systems (Pickett et al. 1992; but also see papers in McDonnell and Pickett 1993). The SCoP project is founded on an opposing view: our sixth principle is that ecological research will be most meaningful to land use decisions affecting wildlife if we strive to understand the factors that control wildlife diversity where people live (Table 1). We will use SCoP as a context in which to attack a variety of questions on ecological, sociological, and economic processes in such systems. Vitally important resources in these endeavors will be contributed by the GIS of SCoP as well as by inventory data that will accumulate over time. We are confident that we can use these resources to attract extramural funding to support research (see below, "A National Funding Initiative"). Human dominated landscapes offer particularly useful experimental opportunities to study ecological processes at large scales. Landscape ecologists have often excused their lack of experimental rigor on the basis of their inability to manipulate the systems they study (Borcard et al. 1992, Wester 1992, Hargrove and Pickering 1992, Underwood 1993). However, landscape manipulation is perfectly feasible in areas that are developed for human uses. The very name of the discipline ''landscape architecture" suggests opportunities for manipulation of horizontal as well as vertical structure of landscapes. For example, roads, trails, wind breaks, irrigation ditches, and settlement itself provide ways to alter the permeability of landscapes to animal movement, and such permeability is widely believed to mediate the effects of landscape structure on many ecological processes, particularly the effects of habitat fragmentation on population viability (Wiens et al. 1985, Holland et al. 1991, Hansen and di Castri 1992, Martin 1992). Often changes in landscapes can be anticipated, thereby facilitating before and after experimental designs (Stewart-Oaten et al. 1986, Underwood 1993). In the near term, we will seek partnerships with universities, foundations, conservation organizations and local governments to fund and execute individual studies that attack particularly important, unresolved questions on ecological processes in human dominated ecosystems. Thus, an important function of SCoP will be as a focal point for research. SCoP will stimulate inquiry and integrate it. In the long term, some of the questions that might be addressed under the general umbrella of SCoP include. the following: Interactions between human density and wildlife diversity: What is the effect of variation in density and spatial arrangement of residential development on wildlife diversity in human dominated landscapes? Are there thresholds where effects of habitation are rapidly amplified? (e.g., Holbrook and Vaughan 1985, Thiel 1985, Brocke et al. 1989). Does the effect of human population density have predictable effects on population densities of selected vertebrates? In what way do these effects depend on animal body size and trophic position? Shifts in anthropogenic sources of landscape disturbance: To what extent does the traditional ecological view of habitat as a series of isolated fragments fit current and historic data on land pattern in the Rocky Mountain Region? Does this view depend on the presence of typeconverting economic activity (e.g., logging, agriculture, etc)? To what extent does the appropriateness of the traditional fragmentation model depend on levels of primary production in ecosystems? That is, are we likely to find an appropriate fit of the fragmentation model only in areas that are sufficiently productive" to support agriculture and intensive forestry? Effects of landscape permeability on vertebrate population dynamics: How is landscape permeability influenced by linear disturbance (roads, trails, etc). How does such disturbance affect dispersal of mammals? Does this effect depend on trophic position? Does permeability influence population viability as would be predicted by fragmentation models? Can the effects of disturbances be altered by changes in vertical habitat structure (i.e., the presence of . 12 . �219 "cover")? What are the structural and spatial characteristics of wildlife? of corridors that promote dispersal Effects of non-consumptive use on wildlife distribution and abundance: Which human activities in parks and reserves exert the greatest influence on the viability of wildlife populations? How can such effects be mitigated? How can use of urban corridors (i.e. trails, open space) by man be reconciled with dispersal needs of wildlife? Human related effects on wildlife community structure: What is the role of introduced predators on the structure and function of natural food webs in human dominated landscapes? Conversely, how does control of natural predators (coyotes, mountain lions) at the urban interface affect wildlife diversity and community dynamics? Does supplemental feeding influence wildlife diversity by altering competitive relationships and tropic structure of food webs? Human dimensions research: What are the features of the environment that motivate political action by citizens? At what point do changes in those values initiate feed-back to policy and law? To what extent does environmental education foster political action? What demographic groups are best served by wildlife recreation in human dominated landscapes? How can environmental values that are fundamentally a collective asset of society be translated to individual economic assets in free market economies? To what extent are citizens willing to forgo short term economic benefit to foster long term environmental quality in urban and suburban areas? PROJECT INTEGRATION Thus far, the components of SCoP have been described as distinct parts. However, in practice these components will be tightly linked into a cogent, interactive program (Fig. 7). For example, the expert system that we propose above will initially contain a blend of ecological theory, empirical . results, and well educated guesses on the impacts of changes in land use on wildlife diversity. Our system will offer no more and no less than the best predictions possible with currently available data. However, investment in monitoring and research will allow the expert system to evolve, in essence, to learn. In so doing, our expert system will offer better and better predictions as research provides enhanced understanding and as inventory provides more comprehensive data. The credibility of the system will be built by repeatedly testing its predictions at different scales of time and space and by modifying its logic in response to the outcome of those tests (we use "test" here in the sense of Oreskes et al. 1994). Moreover, even models as elementary as the ones we propose above can illuminate and communicate needs for inventory and research (Doak et al. 1992, Hanski and Gyllenberg 1993). Thus, we expect that the expert system will prove useful as a repository for information and as an analytical tool for making sense of it (Fig. 7). It will help identify monitoring and research priorities and will help us communicate those priorities to funding agencies. and foundations. The participation of citizens remains central to the connections among inventory, research, and modeling (Fig. 7). This is the case because all of these activities, if properly executed and . .communicated, offer an opportunity to educate people about the wildlife they value. More importantly, .these activities provide the chance for meaningful participation by citizens in decisions on land use--the expert system of SCoP is a tool for citizen education and for policy analysis. 13 �220 Citizen Inventory Research on Human Dominated Ecosystems Figure 7. The expert system of SCOP will serve as a repository of information generated by research and inventory. Research and inventory will be used to improve the logic of the expert system. RELEVANCE TO THE LONG RANGE PLAN OF THE COLORADO DIVISION OF WILDLIFE In the section above, we outlined how the components of SCoP fit together. In this section and in the subsequent one, we describe how SCoP meets the strategic goals of the DOW, and in so doing, compliments other ongoing efforts. The Long Range Plan of the Colorado Division of Wlldlife describes the strategic thrust of the Division for the future. It describes the emphasis and direction of our activities. SCoP will playa pivotal role in meeting strategic goals in four major areas: Habitat and Wildlife Protection, Wlldlife Related Recreation, Wlldlife Information and Education, and Responsive Management (Table 2). The DOW's overriding goal for the next decade will be to protect and enhance the viability of all of Colorado's wildlife species. Clearly, SCoP offers an aggressive, unambiguous response to these strategic directions (Table 2). Because SCoP will focus on places where people live, it will "be highly visible. SCoP will demonstrate a fundamental commitment to meeting the Habitat and Wlldlife protection goals of the Long Range Plan. SCoP also offers a new opportunity for citizens to view and enjoy wildlife (Table 2). Wildlife inventory is defined as the systematic recording of observations of the abundance and distribution of animals. It follows easily from this definition that the distinction between bird watching and avian inventory is not large--participation in monitoring efforts is viewed by many citizens as an opportunity for recreation (Boxall and Mcfarlane 1994). Consequently, the Citizen's Inventory component of SCoP will not only contribute to our efforts to preserve and enhance wildlife populations and their habitats, but will also offer a new, vital thrust in providing Coloradans a chance to enjoy, understand, and value their wildlife. 14 �221 Table 2. Relationship of the SCoP project to goals of the Long Range Plan of the Colorado Division of Wildlife. Contribution of SCOP Goal of Long Range Plan Expert System CItizen Inventory Program Research on Human Dominated Ecosystems • • • • • • • • Habitat and Wildlife Protection Lead statewide efforts involving state, county, and local gOllemments, private landowners, and other organizations to define and identify high priority habitats in the state at various geographic, administrative, and ecological lelle/s. Coordinate sharing, and use of data on wildlife habitats. the collection, Identify high priority habitats and strategies to protect, enhance, and restore them. Provide information to land use decision makers regrading potential impacts of their decisions on wildlife and wildlife habitat, and ensure that they use this information appropriately in established review processes. Lead statewide efforts involving federal, state, county, and local gOllemment agencies, private landowners, and other organization understand and monitor the status of Colorado's wildlife species. to • • • • • • • Wildlife Related Recreation • By providing a dillfJrsity of quality wildlife viewing opportunities, increase the percentage of the state's population that participates in wildlife viewing. DBllelop and implement user groups. strategies to increase participation • by dillfJrse Wildlife Education and Information • Increase the number of Colorado Residents who are knowledgeable about wildlife issues, how human actions affect wildlife and wildlife habitat and the role of humans in the environment. • By 1999, increase volunteer participation by 30% by creating opportunities and projects which will encourage Coloradans to become more involved with and better educated about wildlife and wildlife habitat. DBllelop volunteer educators and wildlife habitat. • to help educate the public about wildlife DeIIf1lop wildlife-related education and information activities and material which capitalize on the use of public resources - parks, natural areas, wildlife areas - as outdoor classrooms that provide a variety of learning experiences to citizens and visitors of all ages. • • • .- Responsive Management Worlc cooperatively with interested groups and the public to dellelop additional funding sources for protecting and enhancing habitat, managing native wildlife species, providing wildlife-related education, and providing wildlife viewing opportunities. 15 • �222 Finally, the DOW has committed itself to an aggressive program in Wildlife Information and Education. SCoP fulfills these committments by providing a meaningful context for citizen education as well as educational tools (Table 2). We believe that people learn best by doing--SCoP offers a chance for people to do meaningful ecological measurements and interpret them. We are genuinely excited about the idea that these measurements can provide a focal point for leaning vital principles in conservation biology and community ecology. For example, presume an inventory group studies a particular habitat patch and finds that a particular species of songbird consistently totals less than 10. If that population is isolated from others, then the expert system will predict that it will disappear in short order. Thus, the experience of people in the field coupled with the model's predictions provide a natural opportunity to explore why small isolated populations are not likely to persist. It offers the opportunity to understand the values of large blocks of habitat and the role of corridors and the importance of genetics and demographic trends. It raises questions on effects of predators and competitors and environmental variability. It provides an entry point for understanding the importance of life history traits. It can stimulate discussion of why it is that large animals at the top of the food chain are not likely to be found in small fragments of habitat. Thus, SCoP offers a way to link people's personal experience with the fundamental processes that underlie the behavior of complex, ecological systems. In the absence of personal experience, these processes are likely to remain distant and abstract. . In' summary, SCoP will meet many needs in a single enterprise. This is feasible, of course, only because those needs are tightly coupled and as a result, SCoP can simultaneously serve several purposes well (Table 2). It is a tool for habitat protection. It is a recreational opportunity. It is a vehicle for education. It is a, research program. RELATIONSHIP TO CURRENT PROGRAMS SCoP shares purposes and approaches with several existing programs, most notably, our efforts in GAP analysis (Scott et aI. 1993), county land use planning, and the Colorado Natural Heritage Program. Here, we make it clear how SCoP compliments these efforts. GAP analysis provides a comprehensive inventory of the geographical distributions of native fauna. This inventory is based on the relationship between vegetative cover types and the likelihood that a species will be found at a given geographic location. GAP analysis provides a much needed survey of the current status of habitats for wildlife in Colorado. We will use some of the products of GAP for coarse scale data layers in SCoP's expert system (see above, "Geographic Information System"). 'However, the view of wildlife diversity provided by GAP is static; GAP does not attempt to infer the status of populations or communities. It emphasizes the current presence or absence of species, and it makes no qualitative or quantitative distinctions between species for which habitat conditions are adequate to assure long term survival, and species for which current conditions will almost assuredly lead to extinction. What is needed, then, is an extension of GAP, an approach that extends GP..P's spstial inventory to make predictions on the future status of species and communities. SCoP offers such an extension. Although the emphasis of SCoP on expanding residential development requires that our efforts focus on a smaller scale than GAP analysis does, the analytical techniques we develop can be used at a variety of scales. Thus, SCoP's expert system will be developed at the scale of municipalities and counties, but it can be applied at a much larger scale, for example, at the scale of a watershed. Such application would, of course, require modification of the system, but the fundamental logic and approach would be the same. With respect to the county land use planning process, it should be clear that SCoP will provide tools to enhance the effectiveness of that process. We seek to facilitate that process. We do not seek to compete with it. 16 �223 Similarly, SCoP will compliment the activities of the Colorado Natural Heritage Program (CNHP). One of the key functions of the CNHP is to provide an objective, fact-based review of impacts on biodiversity that result from projects in the public and private sectors. We will work closely with the CNHP in the design and implementation of SCoP's expert system. As SCoP matures, we will link our databases to theirs. SCoP compliments CHNP by providing analytical tools and a research thrust that are currently missing from policy analysis on human impacts on wildlife diversity. DEMONSTRATION PROJECTS We will initiate projects in three analysis areas to demonstrate that SCoP offers a viable, costeffective approach to wildlife and habitat protection In areas of rapid residential development. The goal of these demonstration projects is to prove the concept of SCoP. Analysis areas will be chosen to represent prevailing patterns of development. We identify two characteristic patterns of residential expansion in Colorado--the growth seen along the east slope of the Front Range, and the development occurring in western slope communities. In the former case, growth proceeds along the interface between mountains and plains, in the latter case growth tends to follow river valleys. Our first analysis area will focus on a rapidly growing Front Range city, including the surrounding plains and foothills (probably Ft Collins). A second and third analysis area will be chosen to represent conditions on the western slope. The greater Durango area, the Eagle River Valley, and the Upper Colorado River Basin have been suggested as excellent candidates. We are now in the process of visiting these areas and discussing SCoP with relevant DOW personnel and local governments. Projects will be initiated in FY 1994-95. We expect completion of the demonstration phase in FY 1996-97 (Fig. 8). A NATIONAL FUNDING INITIATIVE One of the goals of the DOW Long Range Plan (Table 2) is to ''Work cooperatively with interested groups and the public to develop additional funding sources for protecting and enhancing habitat, managing native wildlife species, providing wildliferelated education, and providing wildlife viewing opportunities. " This goal is particularly significant. It s significance stems from the widely held view that any fiscal resources we bring to bear on new problems must necessarily subtract resources from ongoing, traditional efforts. In essence, this view sees wildlife management in Colorado as a zero-sum game. Our view differs. We believe the Colorado Division of Wildlife, in vital connection with Colorado's Universities, can attract meaningful funding from national foundations (e.g., Table 3) if we are willing to aggressively attack problems of national importance. Funding obtained this way is not restricted by the spending limitations of Amendment 1. We are confident of ability of the SCoP project to attract such support. We believe that the problems we have identified rank among the foremost environmental issues confronting society, particularly in the WeStern U.S. We believe that we offer an imaginative and novel approach to addressing these issues, an approach that currently has few competitors. We are aware of a variety of foundations and agencies that seek proposals addressing the issues and questions we described above (Table 3). We are confident that we can prepare marketable grant proposals'', Thus, one the primary 3 This document is an example of the ability of our project team to assemble salable, targeted proposals of the highest quality. Moreover, since 1984, Hobbs has authored or coauthored similar grant proposals generating $1,700,000 in cash awards to the Division of Wildlife and to the Natural Resource Ecology Laboratory. Funding sources have included the National Science Foundation, the U.S. Fish and Wildlife Service, and the National Park Service. 17 �224 Table 3. Organizations that support programs like SCoP. Organization Program Program Objective ARCO Foundation Environmental Grants Support development of rational land-use and natural resource policy; conservation of wildlife; increase participation of youth in community ............................................................................................................................... j ~!:~~!: ..~.?.~P.E?..~.':.~~!: ..:~~ ..~!:~~: __ Compton Foundation .................................................................. Environmental Protection ...~:~5Y. i Land; river, and watershed protection for i purposes of long-term preservation, through J ~ ~~~~.!~ ..~~~~::..~~.p.~~~: ..:~.~~~~?~: i Exploratory Research ~ ~~~ _ .1 Environmental Exxon Corporation _ Research ~ i The Environment ~ ~.«!.~~~~~ ..~~!:.~~':~.~: 1 Contributes to environmental research, i especially that i Support ~ Environment _ . relating to public policy and _.. _._.. research and education to conserve and ~ Research in the west that integrates rural and environmental protection, 1 and demonstration projects of regional ~ l development 1 Jones Foundation Sustainable Society Program Further efforts to maintain biological diversity, especially programs that anticipate impacts of human activities and find sustainable alternatives. National Geographic Society Grants for Field-Based Research Supports research that involves time, space, and scale, with a particular emphasis on multidisciplinary projects of an environmental nature. National Science Foundation Computational Biology Facilitate development and use of computational tools and algorithms that enhance biological research; develop novel computational or mathematical approaches to .................................................................. 1. _.- _ 1 ~.~~!?..!P.~.p.!:.~E.~.~.~..:.. _ 1 Conservation and National Science Foundation l Restoration . on environmental policy and its social ~ P.E~~~~.~?~ ..?~.~~~~~.. .................................................................. ~ Hewlett Foundation i Research ~~~::?: ~ p..:.?~..~~:..:~~?~:~~: .................................................................. ~ Frost Foundation Limited _ . i Environment i _ . 1 Supports research to elucidate principles that l underlie Biology the conservation and restoration of ..-_ --National Science Foundation ~.._-_ ..-.1 Decision, Risk, and Management Science ~ !?~~!.ogi~..~!~~~~ _ _ __. 1 Supports research on decision making under certain and uncertain conditions, management, _ ~ i ~~!~P..':~~!~.~ ..~:.~ t~.~:e~!!..P..~i~ __ . 1 Supports research on natural and managed i ecological systems. Includes experimental, l l emphasis i _ i ! Ecological Studies National Science Foundation 1 ....... _ _ _ _ ....... _ :. i _ _ -_._-- Elementary, Secondary, and Informal Education ~ i _ _ Experience for Undergraduates i_ _ •...•.••...__ theoretical and modeling studies, with a current on the role of biological diversity in ~~.!.~~-.~~:~-_ i i _ _-.__.__ ._ _-_ .. Improve educational experience by providing opportunities to explore science and technology ~~.~~~2..~.~:..~~~! ..~~~~: i Research National Science Foundation _ ~ _ i National Science Foundation l _ __ . _.. i Provide opportunity for undergraduate students to experience hands-on participation in research i or related scientific.. activity. .__ .........•...... . .. .._-_.. i _--_ _ __ __ _ _ __ �225 Organization 1 Program 1 National Science Foundation ! Systematic and ~ Population Biology ! Supports research on the patterns and causes of ~ diversity among and within organisms and 1 populations. Program Objective 1 --------_. ...._._-.-----------.----------------- ! Ecosystem ! Program Nature Conservancy 1 1 1 Research ~ Investigate ecosystem function; foster working relationships between research partners; communicate knowledge about ecosystems; 1 provide for permanent ecosystem research ~ funding. l Social Concerns i Support 1 National Research 1 Support research on the structure, function, ~ 1 -------.--.- ....-~------.--.-----~-----.--.----.---....,----Skaggs Foundation United States Department of Agriculture World Wildlife Fund ~ 1 1 Initiative Competitive 1 Grants P~gram __ ._..... ~ WWF Conservation 1 Programme projects relating to ecology and education. __ . i. and sustainability 11 of ecosystems. ~---.----.-------_---.-1 Strives to preserve target species as well as 1 entire communities. aims of SCoP is to assure that the initial investment of the DOW in the project will eventually be multiplied by extramural funds. In seeking the initial support of the DOW, we are committed to increasing extramural funding for SCoP by no less than double the initial investment by no later than FY 1997-98~ PROJECT ORGANIZATION AND SCHEDULE SCoP will be developed by team. of personnel from the DOW and Colorado's universities during 1994-1997 (Fig. 8). The core development team will include N. T. Hobbs, R. B. Gill, and D. L. Schrupp. Hobbs will lead the project, and will oversee development of the expert system, and will. direct fund raising efforts (Fig. 8). Gill will design and coordinate the Citizen Inventory Program and will develop a process for using SCoP to influence policy. Schrupp will work with Hobbs and a PostDoctoral Fellow to build the expert system. By letting a contract to the Natural Resource Ecology Laboratory", we will recruit a postdoctoral fellow with expertise in GIS and knowledge engineering to support expert system development. In addition, we will hire a computer programmer to write code for the user interface and to implement the SCoP computer network. We will let 4 additional contracts to Colorado universities. One of these contracts in now in place; we are working with John Wiens of the Biology Department of C.S.U. and his doctoral student James Miller. Wiens is widely admired as one of the foremost community ecologists working tcida.y. .Wiens and Miller will assist in the development of the population simulator and will conduct research to enhance and test it. We will also contract with an avian ecologist, a mammalogist, and herpetologist to help us choose and apply the most appropriate inventory techniques and to develop training modules for educating citizens in their use. 4 Hobbs in a member of the Senior Staff at the Natural Resource Ecology Laboratory and can oversee funds and supervise employees. . 19 �226 Figure 8. Schedule of tasks of SCoP Project during demonstration will determine whether project is continued. phase. Project review by in 1997 General Receive enabling Select analysis funding areas .' Flrepare grant proposals for national Project foundations Review by Director's • • Slall Expert System Development Recruit GIS specialist. knowledge ~ engineer Obtain Images for anafysis ~ areas Classily > images :> Build GIS' Develop general >: population simulator' Develop spcecies Develop user interface :> database :>: Citizen's Wildlife Inventory PubliciZIt ----? SCoP and Citizen Inventory Teams Distribute requests methods and training Develop inventory ~ for proposals for inventory modules :> methOds and tra,ining modules ,Recruit Citizen Train CItizen Condud Inventory Inventory Demonstrate invenotry Inventory ~ Teams ~ --7 Teams :--7 >: and Monitoring expert system to citizen teams. >: >: local governments 07/01/94. 20 07/01/95 07/01/96 07/01/97 �227 Figure 8. Continued. Research on Human Dominated Dominated Ecosystems Conduct proof of concept research to attract extramural > funding Develop joint ventures universities with > and NGO's Prepare grant proposals Conduct process studies Publish books. papers to . to publisize motivate Seo? further and funding Final Review of Demonstration Project by Director's Wildlife Commission • Staff and 07/01/94 07/01/95 07/01/96 07/01/97 We will initiate demonstration projecm in 1994 and will deliver a completed product at the end of FY 1997 (Fig. 8). This product will consist of the following parts: • • • • • • • .• A folly operational expert system, A cadre of citizen inventoey teams, Initial monitoring data and research findings, A series of inventoey training modules, Funding commitments from national foundations, International recognition of the SCOP approach, Results from pilot research projects, Ongoing, extramurally funded research. . BUDGET In considering the budget (Table 4), bear in mind that these are start-up costs of the program. As a result, the cost per analysis area is quite high. However, marginal costs will decline steeply after development of the expert system, and after inventoey training modules have been produced. Mo~er, we are confident that we will be able to leverage initial investments to expand the program beyond FY 1996 (see above, -A National Fund Raising Initiative"), 21 �Table 4. Estimated costs of implementing SCoP in 3 analysis areas during FY 1994-1997. N N ------- ------------ ------- -------- - co --- 1994-95 mo. I 1995-96 FTE cost 12 1.0 52,027. 31,072. 6 .5 32,159. 12 15,000. 12 15,000. Contract with Mammalogist 1.5 7,500 1.5 7,500 Contract with Avian Eoologist 1.5 7,500 1.5 7,500 Contraot with Harpatologist 1.5 7,500 1.5 7,500 12 30,000 12 30,000. 12 32,500. 12 24,000. 6 13,000. Personnel Servioes N. T. Hobbs R. B. Gill Contraot with J. A. Wiens, C.S.U. Contraot with Natural Resouroe Eoology Laboratory Post Dootoral Fellow (GIS speoialist-knowledge engineer) FTE cost 12 1.0 50,268. 6 .5 Computer Programmer Overhead on contracts (@ 10%) mo. 1996-97 cost 12 1.0 52,027. 6 .5 32,159 . 4,500. 10,000. Sampling Equipment 3,000. Computer for post-doc 6,000 35,000 Computers for citizen inventory teams, computer programmer Project Cost Over 3 Years: 4.5 FTE's $523,152.00 FTE 9,675. 6,750. Eguie!I!ent Imagery and olassifioation Total mo. 46.50 1.5 $174,590.00 58.5 1.5 $205,376 36 1.5 $143,186 �229 . In addition, it is important to understand that these are the costs of support of the core team. We anticipate close cooperation with many other people in the DOW as the project matures. For example, we expect that District Wildlife Mangers and Biologists will be able to contribute some time to support citizen inventory efforts during start-up phases. LITERATURE CITED Adams, L. W., and Dove, L. E. (eds.) 1989. Wildlife Reserves and Corridors in the Urban Environment: A Guide to Ecological Landscape Planning and Resource Conservation. National Institute for Urban Wildlife, Columbia, Maryland. 91 pp. Ambuel, B. and S.A. Temple. 1983. Area-dependent changes in the bird communities and vegetation of southern Wisconsin forests. Ecology 64:1057-1068. Baharav, D., J. A Kent, and E. A- Baharav. 1991. Eco-mapping: planing and management of bio. social ecosystems. Pages 207-10 in RD. Comer et al. eds., Issues in the Management of Impacted Wildlife. Thorne Institute, Boulder Colorado. Batty, M. 1991. Cities as fractals: simulating growth and form. Pages 43-69 in A. J. Crilly, R A. Hernshaw and H. Jones, eds. Fractals and Chaos. Springer-Verlag, New York. Beier, P. 1993. Determining minimum habitat areas and habitat corridors for cougars. Conservation Biology 7:94-107. Blueweiss, L., H. Fox, V. Kudzma, D. Nakashima, R Peters, and S. Sams, 1978. Relationships between body size and some life history parameters. Oecologia 37:257-272. Boxall, P. C., and B. L. McFarlane. 1994. Human dimensions of Christmas Bird Counts: implications for nonconsumptive wildlife recreation programs. The Widlife Society Bulletin 21:390-397. Borcard, D., P. Legendre, and P. Drapeau. 1992. Partialling out the spatial component of ecological variation. Ecology 73:1045-1055. Brocke, R, J. O'Pezio, and K Gustafson. 1989. A forest management scheme mitigating impact of road networks on sensitive wildlife species. Pages 13-17 in R DeGraaf and W. Healy, compilers. Is forest fragmentation a management issue in the Northeast? General Technical Report NE-140. U.S. Department of Agriculture, Forest Service, Northeast Forest Experiment Station, Radnor, Pennsylvania. Burgman, M., D. Cantoni, and P. Vogel. 1992. Shrews in suburbia: an application of Goodman's extinction model. Biological Conservation 117-123. Burgman, M. A., S. Ferson, and H. R Akcakaya. 1993. Risk Assessment in Conservation Biology. Chapman and Hall, New York. 314 pp. Cairns, M. A, and R. T. Lackey. 1992. Biodiversity and management of natural resources: the issues. Fisheries 17:6-10. Caswell, H. 1989. Matrix Population Models. Sinauer. Sunderland, Massachusetts. 328 pp. Charnov, E. L. 1992. Allometric aspects of population dynamics: a symmetry approach. Evolutionary Ecology 6:307-311. Clark, T., K Clark, S. Paterson, D. Mackay, and R J. Norstrom. 1988. Wildlife monitoring, modeling, and fugacity. Environmental Science and Technology 22:120-127. Clutton-Brock, T. H. and P. H. Harvey. 1983. The functional significance of variation in body size among mammals. Special Publications of the American Society of Mammalogists 7:632-663. Colorado Department of Health. 1993. Colorado Vital Statistics 1991. Health Statistics and Vital Records Division, Department of Health, Denver, Colorado. Cotgreave, P., and P. H. Harvey. 1992. Relationships between body size, abundance and phylogeny in bird communities. Functional Ecology 6:248-256. Coulson, R N., L. J. Folse, and D. K Loh. 1987. Artificial intelligence and natural resource management. Science 237:262-267. Coulson, R N., C. N. Lovelady, R O. Flamm, S. L. Spradling, and M. C. Saunders. 1991. Intellegent geographicinformation systems for natural resource management. Pages 153-173 in M. G. Turner and R H. Gardner, eds. Quantitative Methods in Landscape Ecology: The Analysis and Interpretation of Landscape Heterogeneity. Springer-Verlag. New York, N.Y., U.S.A. C 23 �230 Cronon, W.J. 1993. Foreword: the turn toward history. Pages vii-x in M.J. McDonnell and S.T.A. Pickett, editors. Humans as components of ecosystems. Springer-Verlag, New York. Currie, D. J. 1993. What shape is the relationship between body size and population density? Oikos 66:353-358. Doak, D. F., P. C. Marino, and P. M. Kareiva. 1992. Spatial scale mediates the influence of habitat fragmentation on dispersal success: implications for conservation. Theoretical Population Biology 41:315-336. Dueker, K. J., and P. B. deLacy. 1990. GIS in the land development planning process. Balancing the needs of land use planners and real estate developers. American Planning Association Journal 56:483-491. Fahrig, L. 1988. A general model of populations in patchy habitats. Applied Mathematics and Computation 27:53-66. Fahrig, L., L. P. Letkovitch, and H. G. Merriam. 1982. Population stability in a patchy environment. Developments in Environmental Modelling, 5. Pages 61-67 in W. K. Lauenroth, G. V. Skogerboe, and M. Flug, editors. Analysis of Ecological Systems: State-of-the-Art in Ecological Modelling. Elsevier Scientific Publishing Company, New York. Fedra, K. 1993. GIS and environmental modeling. Pages 35-50 in M. F. Goodchild, B. O. Parks, and L. T. Steyart, Environmental Modeling with GIS. Oxford University Press. Oxford, U. K. Fenchel, T. 1974. Intrinsic rate of natural increase: the relationship with body size. Oecologia 14:317-326. Fitzgerald, J.P., C.A Meaney, and D.M. Armstrong. In Press. Mammals of Colorado. Colorado Associated University Press. Boulder, CO. Gilpin, M. E. 1987. Spatial structure and population vulnerability. Pages 125-139 in M. E. Soule, editor. Viable populations for conservation. Cambridge University Press, Cambridge. Goode, D. A 1989. Urban nature conservation in Britan. Journal of Applied Ecology 26:859-873. Gordon, D. 1990. Green Cities: Ecologically Sound Approaches to Urban Spaces. Black Rose Books, Montreal. 299 pp. . Hansen, A. J. and F. di Castri. 1992. Landscape Boundaries: Consequences for Biotic Diversity and Ecological Flows. Springer Verlag, New York. 452 pp. Hansen, A J., and D. L. Urban. 1992. Avian response to landscape pattern: the role of species' life histories. Landscape Ecology 7:163-180. . Hansen, A J., T. A Speis, F. J. Swanson, and J. L. Ohmann. 1991. Conserving biodiversity in managed forests. BioScience 41:382-392. Hanski, 1., and M. Gyllenberg. 1993. Two general metapopulation models and the core-satellite species hypothesis. American Naturalist 142:17-41. Hargrove, W. W., and J. Pickering. 1992. Pseudoreplication: a sine qua nan for regional ecology. Landscape Ecology 6:251-258. Harris, L. D. 1984. The Fragmented Forest: Island Biogeography Thoeory and the Preservation of Biotic Diversity. Univeristy of Chicago Press, Chicago, Illinois, U.S.A 211 pp. Hastings, A In Press. Conservation and spatial structure: theoretical approaches. Lecture Notes in Biomathematics 100: in press. . Hennemann, W. W. m. 1983. Relationship among body mass, metabolic rate and the intrinsic rate of natural increase in mammals" Oecologia 56:104-108. Hobbs, N. T. 1989. Linking energy balance to survival in mule deer: development and test of a simulation model. WIldlife Monographs 101:1-39. Hobbs, N. T. amd M. W. Miller. 1990. Users Guide to CERVGENE: Software for analysis of the impacts of introduced red deer on Colorado's elk populations. Report to a Joint Workshop of the Colorado Agriculture and WIldife Commissions, August 1990. Hobbs, N. T., and J. R Miller. 1994. Quarterly Report, January 1994--Predicting the Impacts of Environmental Change: Simulations of Genetic and Species Diversity at Landscape and Regional Scales. Federal Aid Proj llA-1, Job Progress Report. Hobbs, N. T., M. W. Miller, J. A. Bailey, D. F. Reed, and R B. Gill. 1990. Biological criteria for introductions of large mammals using simulation models to predict impacts of competition. Transactions of the North American Natural Resources Conferance 55:620-632. 24 �231 Hobbs, N. T., J. E. Gross and J. R. Miller. 1993. Quarterly Report, March 1993--Predicting the Impacts of Environmental Change: Simulations of Genetic and Species Diversity at Landscape and Regional Scales. Federal Aid Proj 11A-1, Job Progress Report. Holbrook, H., and M. Vaughan .. 1985. Influence of roads on turkey mortality. Journal of Wildlife Management 9:611-614. Holland, M. M., P. G. Risser, and R. J. Naiman. 1991. Ecotones: The Role of Landscape Boundaries in the Management and Restoration of Changing Environments. Chapman and Hall, New York. 142 pp. Leedy, D. L., R. M. Maestro, and T. M. Franklin. 1978. Planning for Wildlife in Cities and Suburbs. U. S. Fish and Wildlife Service Rep. OBS - 77/66. . Lindstedt, S. L., and M. S. Boyce. 1985. Seasonality, fasting endurance, and body size in mammals. The American Naturalist 125:873-878. Lindstedt, S. L. and W. A. Calder. 1981. Body size, physiological time, and longevity of homeothermic animals. Quarterly Review of Biology 56:1-16. Lindstedt, S. L., B. J. Miller, and S. W. Buskirk. 1986. Home range, time, and body size in mammals. Ecology 67:413-418. Lovejoy, T. E. 1986. Species leave the ark one by one. Pages 13-27 in B. G. Norton, ed. The Preservation of Species: The Value of Biological Diversity. Princeton University Press. Princeton, New Jersy. U.S.A. Majeske, C. 1993. 118 pages. Wilderness Planning and Management at the Urban Interface. M.S. Thesis, Ft. Collins, Colorado. Martin, T. E; 1992. Landscape considerations for viable populations and biological diversity. Transactions of the 57th North American Wildlife and Natural Resources Conference 1992:283-291. McDonnell, M.J., and S.T.A Pickett (eds.). 1993. Humans as components of ecosystems. SpringerVerlag, New York. McHarg,1. 1971. Design with Nature. DoubledayjNatural History Press. Garden City, New Jersy, U.S.A McKelvey, K, B. R. Noon, and R. H. Lamberson. 1993. Conservation planning for species occupying fragmented landscapes: the case of the Northern spotted owl. Pages 424-450 in P. M. Kareiva, J. G. Kingsolver, and R. B. Huey, eds. Biotic Interactions and Global Change. Sinauer, Sunderland, Mass. Millar, J. S., and G. J. Hickling. 1991. Body size and the evolution of mammalian life histories. . Functional Ecology 5:588-593. . Miller, J.R., D.L. Urban, R.L. Knight, AJ. Hansen, and M.F. Carter. 1994. Avian life-history traits and landscape change: a comparison among three biomes. Conservation Biology (in review). Murphy, D. and B. Noon. 1992. Integrating scientific methods with habitat conservation planning: reserve design for northern spotted owls. Ecological Applications 2:3-17. Noss, R. F. 1990. Indicators for monitoring biodiversity: a hierachial approach Conservation Biology 4:355-364. O'Connell, M. A, and R. F. Noss. 1992. Private land management for biodiversity conservation. Environmental Management 16:435-450. Oreskes, N., K Shrader-Frechette, and K Belitz. 1994. Verification, validation, and confirmation of numerical models in the earth sciences. Science 263:641-646. . OTA 1987. Technologies to maintain biological.diversity. OTA-F-330. Government Printing Office, Washington, D.C. 334 pp. Peters, H.R. 1983. The Ecological Implications of Body Size. Cambridge University Press, New York. 329 pp. Pulliam, H. R., J. B. Dunning, and J. G. Liu. 1992. Population dynamics in complex landscapes - a case study. Ecological Applications 2:165-177. . Randall, A 1986. Human preferences, economics, and the preservation of species. Pages 79-109 in B. G. Norton, ed. The Preservation of Species: The Value of Biological Diversity. Princeton University Press. Princeton, New Jersey. U.S.A. Rennicke, J. 1990. Colorado wildlife. Colorado Geographic Series No.6. Falcon Press Publishing Company, Helena and Billings, Montana. 25 �232 Riley, M. 1993. Byte by byte and fax by fax, the West is being transformed. High Country News 25:12. Robbins, C., D. Bystrak, and P. Geissler. 1986. The breeding bird survey: its first 15 years. 19651979. Resource Publ. 157, U.S. Dept of Interior, Fish and Wildlife Service, Washington, D.C. Saunders, D. A., R J. Hobbs, and C. R Margules. 1991. Biological consequences of ecosystem fragmentation: a review. Conservation Biology 5:18-32. Schmidt-Nielsen, K. 1984. Scaling: why is animal size so important? Cambridge Univiversity Press, Cambridge. 241 pp. Scott, J. M., F. Davis, B. Csuti, R. Noss, B. Butterfield, C. Groves, H. Anderson, S. Caicco, F. Derchia, T.C. Edwards, J. Ulliman, and RG. Wright. 1993. GAP analysis: a geographical approach to protection of biological diversity. Wildlife Monographs 123:1-41. Snow, D. 1994. Wise Use. Northern Lights 9:10-13. Soil Conservation Service. 1989. Summary Report: 1987 National Resources Inventory. USDA, Soil Conservation Service, Statistical Bulletin No. 790. United States Government Printing Office, Washington, D.C. Soule, M. E. 1991. Land use planning and wildlife mainenance: guidlines for conserving Wildlife in urban landscapes. Journal of the American Planning Association 57:313-323. Spellerberg, I. F. 1991 Monitoring Ecological Change Cambridge University .Press, Cambridge U.K., 334 pp. Stenberg, K., and W. W. Shaw. 1986 Wildlife Conservation and New Residential Development. Proceedings of the National Symposium on Urban Wildlife. ARCO Foundation, Tucson, Arizona. 203 pp. Stewart-Oaten, A., W. W. Murdoch, and K. R Parker. 1986. Environmental impact assessment: "Pseudoreplication" in time? Ecology 67:929-940. Steyart, L. T. 1993. A perspective on the state of enviornmental simulation. Pages 16-30 in M. F. Goodchild, B; O. Parks, and L. T. Steyart eds, Environmental Modeling with GIS. Oxford University Press, Oxford. Temple, S. A., and J. A. Wiens. 1989. Bird populations and environmental changes: can birds be bioindicators? American Birds 43:260-270. Terborgh, J. 1989. Where have all the birds gone? Princeton University Press, Princeton, New Jersey. Thiel, R 1985. Relationship' between road densities and wolf habitat Suitability in Wisconsin. American Midland Naturalist 113:404-407. Thomas, J. W., E. D. Forsman, J. B. Lint, E. C. Meslow, B. R Noon, and J. Verner. 1990. A conservation strategy for the Norther Spotted Owl. Interagency committee to address the conservation of the Northern Spotted Owl (USDA: Forest Service; USDI: Bureau of Land Management, Fish and Wildlife Service, and National Park Service). 1990-791-171/20026. U.S. Government Printing Office, Washington, D.C. Underwood, A. J. 1993. The mechanics of spatially replicated sampling programmes to detect environmental impacts in a variable world. Australian Journal of Ecology 18:99-116. Urban, D. L., and T. M. Smith. 1989. Microhabitat pattern and the structure of forest bird communities. American Naturalist 133:811-829. U.S. Department of Commerce. 1989. State Personal Income, 1929-1987. United States Government Printing Office, Washington, D.C. Verner, J. 1983 An integrated system for monitoring wildlife on the Sierra National Forest. Transactions of the North American Wildlife Conference 48:355-366. Wester, D. B. 1992. Replication, randomization, and statistics in range research. Journal of Range Management 45:285-290. Whitcomb, R F., J. F. Lynch, M. K. Klimkiewicz, C. S. Robbins, B. L. Whitcomb, and B. Bystrak. 1981. Effects of forest fragmentation on avifauna of the eastern deciduous forest. Pages 125206 in: R. L. Burgess and D. M. Sharpe, eds. Forest Island Dynamics in Man-dominated Landscapes. Springer-Verlag, N.Y. Wiens, J. A., C. S. Crawford, and J. R. Gosz. 1985. Boundary dynamics: a conceptual framework for studying landscape ecosystems. Oikos 45:421-427. WIlcove, D~ S. 1985. Nest predation in forest tracts and the decline of migratory songbirds. Ecology 66:1211-1214. 26 �233 Appendix B Draft 1 8-1-94 Dissertation Proposal Avian Diversity Across a Gradient of Human Density J. R Miller . Graduate Degree Program in Ecology Department of Biology _ Colorado State University Fort Collins, Colorado 80523 Introduction: Habitat fragmentation has been called "...the most serious threat to biological diversity" and "... the primary cause of the present extinction crisis" (Wilcox and Murphy 1985). Many workers have described fragmentation in terms of two components, the loss of habitat and increased insularization of the remnants (\VIlcove et al. 1986, Saunders et al. 1991, Primack 1993). Studies of fragmentation in North America focused initially on birds in eastern deciduous forests, largely to the exclusion of other taxa in other habitats (Rosenberg and Raphael 1986). More generally, such studies have been undertaken in areas where human land use has resulted in large-scale land clearing for agriculture, urban development, or timber harvest. While much discussion of habitat fragmentation has centered on habitat "islands" embedded in a highly modified or degraded matrix, such landscapes represent only one endpoint of the fragmentation continuum. Landscape trajectories that lead to this endpoint of habitat "isolation" begin necessarily with the first agricultural field, clearcut, or residential development. This initial phase or starting point may be referred to as habitat "perforation" (Fig. 1). Movement along this continuum may be in either direction and is tightly coupled to land-use trends and economic factors. For instance, Foster (1992, 1993) has shown that a New England landscape underwent a period of deforestation for agricultural purposes after European settlement, followed by a phase of population migration and reforestation paralIelling the rise of agricultural opportunities and urban jobs in the midwest Since the 1930's, there has been a reversal in the downward population trend with an increase in residential growth (Foster 1992, 1993). Similarly, large-scale conversion of short-grass prairie to agricultural land and clearing of forested lands, particularly at low elevations, followed settlement in Colorado (Veblen and Lorenz 1991). These traditional sources of habitat loss, however, are currently being supplanted by residential growth (Veblen and Lorenz 1991, Fitzgerald et al: in press), as thousands of hectares of agricultural land are being converted annually to ur~an use (U.S. Bureau of the Census 1993) and below-cost timber sales are phased out on Colorado's National Forests (Obmascik 1993). At the "isolation" end of the fragmentation continuum, habitats may be characterized as islands in a sea of development and edge effects, such as those involving predation, parasitism, and changes in microclimate (Temple 1986, Wilcove et al. 1986, Saunders et al. 1991) emanate from the matrix into the habitat patches. With "perforation", edge effects or disturbance originates in the patches and moves into the habitat matrix. Corridors, such as roads, that connect these patches may also serve as sources of disturbance in addition to acting as barriers to animal movement (Bennett 1991, Schonewald-Cox and Buechner 1992). �234 Miller While the negative impacts of "perforation" and disturbance corridors on the native fauna may be substantial (e.g., Small and Hunter 1988, Askins 1994, Engels and Sexton 1994), most empirical work has been concentrated at the "isolation" end of the spectrum. Indeed, I recently surveyed studies published during the last 20 years that focused on fragmentation of terrestrial habitats using the BIOSYS database. Over 83% of these investigations were conducted in areas where habitat islands existed in a hostile matrix, while less than 8% examined "perforated" landscapes or the effects of connecting rights-of-way. This survey serves to underscore the need for further study of this important and widespread phase of habitat fragmentation. Objective: I propose to examine avian species composition along a gradient of human density (McDonnell and Pickett 1990). One would expect a higher proportion of human-associated species (such as starlings, blue jays, grackles, crows, and magpies) as human density increases. In addition, I will examine changes in avian predator assemblages along this gradient. Nest predation is the primacy cause of nesting mortality for many species (Martin 1987) and has been implicated in some population declines <BohningGaese et al. 1993). There is evidence to suggest increased predation rates near residential areas (Wilcove 1985), probably attributable to the presence of species that thrive in human-dominated environments, such as dogs, cats, raccoons, foxes, skunks, and corvids. By examining species presence along a gradient of human density, it may be possible to identify thresholds relating to the presence of human-adapted species. This, in turn, may allow the determination of potential thresholds above which "perforations" connected by disturbance corridors (i.e., roads) percolate across the landscape (Gardner et al. 1987, Turner et al. 1988), resulting in effective habitat "isolation." Such information may prove useful for land managers in planning for residential development or the acquisition of open space. Study Area: Fieldwork will be conducted in Douglas County, Colorado. Douglas County has the second highest growth rate of any county in the U.S. - with the highest growth rate being experienced in adjacent Jefferson County (U.S. Bureau of the Census 1993). Most of this growth is relatively recent, however, and the county provides a wide gradient of human density. The western third of the county is characterized by the interface of the foothills of the Rocky Mountains, primarily the Rampart Range, and the eastern plains. The vegetation consists primarily of coniferous forest, dominated by ponderosa pine (pinus ponderosa) and Douglas-fir (pseudotsuga menziesii) in the overstory and Gambel's oak (quercus gambelii) in the understory. The eastern portion of the county consists of rolling hills and mesas with vegetation characteristic of montane shrublands, primarily Gambel's oak with ponderosa pine interspersed. The combination of Gambel's oak and ponderosa pine provides for high avian diversity (Andrews and Righter 1992, Mutel and Emerick 1984) and an especially rich mammalian fauna, especially where it occurs on rocky substrates (Mute! and Emerick 1992, Fitzgerald et aL in press). Methods: This study will be conducted using a stratified design whereby 50 census points will be randomly selected in each of four strata and divided between two study areas per strata (i.e., each study area will have 25 points). For each study area, the 25 census points will be evenly divided between five randomly selected sites (i.e. five points/site). The four strata will represent a gradient of human density from high density (e.g., areas near the towns of Castle Rock, Parker, or Larkspur) to the lowest density. Strata, study areas, sites, and census points will be selected using TIGER data files �235 Miller . ™ (U.s. Bureau of the Census 1991), Landsatsatellite imagery, 1:24,000 color infrared photographs, and a Geographic Information System (GIS). Census points will be located on private and public land in mixed ponderosa pine/Gambel's oak stands, and will be visited three times per field season from mid-May to mid-July. Avian species will be censused over a 20 minute period at each point, beginning with a 10 minute cool-down period, with observers randomly assigned to and rotated among study sites in order to avoid observer bias. Predator assemblages will be sampled at one census point per study site using dummy nests, scent posts, or smoked aluminum track beds, or a combination thereof. Vegetation will be sampled at one plot per study site from mid-July to mid-August. In addition, other variables such as actual distance from residential developments, distance from roads, surrounding land-use, and human density of surrounding areas (in addition to the density in which the site is located) using a GIS and the data layers mentioned above. A hierarchical regression model will be constructed for selected species in order to examine the predictive power of both study site and landscape variables. Budget: I am requesting $15,000 annually for two field seasons. This money will cover costs associated with hiring three field technicians, housing, transportation, and materials. . Literature Cited: Andrews, R, and R Righter. Colorado. 1992. Colorado Birds. Denver Museum of Natural History, Denver, ..•... Askins, R 1994. Open corridors in a heavily forested landscape: impact on shrubland and forest-interior birds. Wildlife Society Bulletin 22:339-347. Bennett, A. }991. Roads, roadsides and wildlife conservation: a review. Pages 99-117 in D. A. Saunders and R J. Hobbs, editors. Nature conservation 2: the role of corridors. Surrey Beatty & Sons, Chipping Norton. Bohning-Gaese, K., M. Taper, and J. Brown. 1993. Are declines in North American insectivorous songbirds due to causes on the breeding range? Conservation Biology 7:76-86. Engels, T., and C. Sexton. 1994. Negative correlation of blue jays and golden-cheeked warblers near . an urbanizing area. Conservation Biology 8:286-290. . Fitzgerald, J. P., C. A. Meaney, and D. M. Armstrong. Associated University Press, Niwot, Colorado. In press. Mammals of Colorado. Colorado Foster, D. 1992. Land-use history (1730-1990) and vegetation dynamics in central New England, USA. Ecology 80:753-772. Foster, D. 1993. Land-use history and forest transformations in central New England. Pages 91-110 in M. J. McDonnell and S. T. A. Pickett, editors. Humans as components of ecosystems. Springer-Verlag, New York. . Gardner, R, B. Milne, M. Turner, and R. O'Neill. 1987. Neutral models for the analysis of broad-scale landscape pattern. Landscape Ecology 1:19-28. �236 Miller Martin, T. 1987. Artificial nest experiments: effects of nests appearance and type of predator. 89:925-928. McDonnell, M., and S. Pickett. 1990. Ecosystem structure and function along urban-rural an unexploited opportunity for ecology. Ecology 71:1232-1237. Mute!, C., and J. Emerick. Colorado. Condor gradients: 1984. From grassland to glacier. Johnson Publishing Company, Boulder, Obmascik, M. 1993. Feds may fell timber industry. Denver Post, June 20, pp. 1A, lOA Primack, R. 1993. Essentials of conservation biology. Sinauer Associates Inc., Sunderland, Massachusetts. Rosenberg, K., and M. Raphael. 1986. Effects of forest fragmentation on vertebrates in Douglas-fir forests. Pages 263-272 in J. Verner, M. L. Morrison, and C. J. Ralph, editors. Wildlife 2000. University of Wisconsin Press, Madison, WlSconsin. Saunders, D., R. Hobbs, and C. Margules. 1991. Biological consequences of ecosystem fragmentation: a review. Conservation Biology 5:18-32. Schonewald-Cox, C., and M. Buechner. 1992. Park protection and public roads. Pages 373-395 in P. Fielder and S. Jain, editors .. Conservation biology. Chapman and Hall, New York, New York. Small, M., and M. Hunter. 1988. Forest fragmentation landscapes. Oecologia 76:62-64. and avian nest predation in forested Temple, S. A 1986. Predicting impacts of habitat fragmentation on forest birds: a comparison of two models. Pages 301-304 in J. Verner, M. Morrison, and C. Ralph, editors. Wildlife 2000. University of Wisconsin Press, Madison, Wisconsin. Turner, M., R. Gardner, V. Dale, and R. O'Neill. 1988. predicting the spread 'of disturbance across heterogeneous landscapes. Oikos 55:121-129. U.S. Bureau of the Census. 1993. Statistical abstract of the United States. Office, Washington, D.C. Veblen, T., and D. Lorenz. City, Utah. c U.S. Government Printing 1991. The Colorado Front Range. University of Utah Press, Salt Lake Wilcove, D., C. McLellan, and A Dobson. 1986. Habitat fragmentation in the temperate zone. Pages 237-256 in M. Soule, editor. Conservation biology: the science of scarcity and diversity. Sinauer Associates, Inc., Sunderland, Massachusetts. Wilcox, B., and D. Murphy. 1985. Conservation strategy: the effects of fragmentation American Naturalist 125:879-887. on extinction. � https://cpw.cvlcollections.org/files/original/622b1b1cfb4b9ff5577aca7004faa0ca.pdf 35c5b71da5d0c11ef26ed9b33b49b7c4 PDF Text Text 1 Colorado Division of Wildlife Wildlife Research Report October 1994 JOB PROGRESS REPORT State of _ _;C=o"-l,-"o,-,-r~a=do,,,-_ Project Miaratory Game Bird Investiaations W-166-R-3 Work Plan _1_: Job 20 Job Title: Development and evaluation of moist-soil manaaement techniques in Colorado Period Covered: Author: 1 April 1993 through 31 March 1994 James K. Rinaelman Personnel: J. Goettl, J. Ringelman, M. Szymczak. L. Swift, Colorado Division of Wildlife. ABSTRACT Uncontrollable high water during the 1992 growing season in the 4 ha study pond resulted in a vigorous stand of threesqua~e bulrush (Scirpus americana), with minimal germination of smartweeds (Polygonum spp.), curlydock (Rumex crispus), wild millet (Echinochloa crusaalli), and other moist-soil plants. In late summer, filamentous green algae became common, due in part to water stagnation. In 1993, an attempt to control bulrush by maintaining water at a comparatively high level resulted in limited success. Some smartweed germinated during the high water period. Plans are to begin drawdown early spring in 1994. Prepared by: James K. Ringelman Researcher/Scientist V ��3 Colorado Division of Wildlife Wildlife Research Report October 1994 JOB PROGRESS REPORT State of _.....:C~o!.-'l'-"'o~r~ad~o:!._ _ Project Work Plan 1 Job Title: : Job 22 Harvest distribution of mallards and pintails banded preseason in western Colorado Period Covered: Author: Miaratory Game Bird Investigations W-166-R-3 01 April 1993 through 31 March 1994 Michael R. Szym~zak Personnel: J. Gamble and staff, Browns' Park National Wildlife Refuge; K. Nelson, U. S, Forest Service; J. Broderick, J. Corey, D. Coven, P. Creeden, K. Dillinger, J. Ellenberger, V. Graham, J. Gray, J. Gumber, D. Masden, J. Miller, B. Motz, J. Olterman, R. Olterman, J. Slater, N. Smith, M. Szymczak, J. Todd and K. Wagner, Colorado Division of Wildlife ABSTRACT Ducks were trapped in modified Salt Plains bait traps and banded on 19 different wetlands located in 5 areas across western Colorado in August and September, 1993. About 2,200 mallards (Anas p1atyrhnchos) were banded, with the number captured generally well distributed between trapping areas. Where trapping effort was measured, mallard trapping efficiency was similar to 1992 levels, with about 7.5 mallards banded per trap/day. Only 14 northern pintail (Anas acuta) were banded in 1993. ��5 HARVEST DISTRIBUTION OF MALLARDS AND PINTAILS BANDED PRESEASON IN WESTERN COLORADO P. N. OBJECTIVE 1. Document the distribution of band recoveries of mallards and pintails captured preseason in western Colorado. 2. Determine if the geographic location of recovery of mallards and pintails is dependent on the area of banding in Colorado. 3. Determine the relationship between the recovery distribution of western Colorado banded mallards and pintails and the distribution of recovery of those species banded in other areas of the Pacific Flyway. 4. Cooperate in analysis of Pacific Flyway - wide band recovery data and preparation of reports. SEGMENT OBJECTIVES 1. Trap and band mallards and pintail in 4-5 areas of western Colorado in late August-early September using salt plains bait traps (Szymczak and Corey 1976). Recommended areas are: (1) Browns' Park National Wildlife Refuge, (2) Yampa River Valley below Craig, (3) Colorado River Valley below Glenwood Springs, (5) Uncompahgre-lower Gunnison River Valley, and (6) the Cortez - Mancos area. 2. Submit banding schedules and recapture reports to the U. S. Fish and Wildlife Services' Bird Banding Laboratory. Summarize and file band return reports. 3. Contribute manpower and equipment. to cooperative duck banding crews in Alberta. INTRODUCTION In 1990, the Pacific Flyway Study Committee formulated a 5-year cooperative mallard and pintail preseason banding program that was endorsed by the Pacific. Flyway Council. This program was designed to address banding needs throughout the western U. S., including Alaska, and in the provinces of British Columbia and Alberta. Through the first 2 years, nearly 172,000 ducks were banded under this program. This report covers the third year of banding during the preseason period in western Colorado. Background information can be found in Szymczak (1992). �6 METHODS Trap Area Selection Most breeding mallards in western Colorado are associated with small wetlands that are widely distributed throughout high elevation, mountainous areas. Only in Browns' Park, located in the extreme NW corner of Colorado, are there extensive wetland complexes capable of supporting breeding ducks. Trapping on widely distributed wetlands with low densities of breeding ducks was assumed to be an inefficient method for banding western Colorado mallards. Therefore, strategically located areas, to which post-breeding and fledged Colorado mallards would move, were selected as primary trapping areas. In 1993, trapping occurred on the Browns' Park National Wildlife Refuge (BPNWR), along the Colorado River Valley from Debeque to near Fruita (CRV), in the Uncomphagre River Valley from.Montrose to Delta including some locations east of Delta in the Gunnison River Valley (URV), in the Cortez-Mancos area (CM), and at Gardner Park and Allen Basin reservoirs, about 5 miles west and 7 miles Northwest, of Yampa, respectively. Trapping Period Ducks were trapped and banded in each area for a consecutive period of about 10 days beginning near the end of August and ending near 20 September. The actual banding periods in 1993 were: BPNWR ~ 17 August thru 2 September; Yampa Area - 9 thru 20 September; CRV ~ 30 August thru 12 September; URV - 29 August thru 10 September; CM 3 thru 12 September. Trapping and Recording Technique All birds were trapped in modified Salt Plains bait traps (Szymczak and Corey 1976) using whole shelled corn for bait. Traps were visited daily. Mallards and pintails were the target species, but green-winged teal (Anas caro7inensis) were also handed in all areas, ring-necked ducks (Aythya co77aris) were banded near Yampa and blue-winged teal (Anas discors) and/or cinnamon teal (Anas cyanoptera) were banded in the CRV, and blue-winged teal, gadwall (A. strepera) and redheads (Aythya americana) were banded in Browns' Park. Banded birds were recorded by wetland site. Band numbers of all birds captured that were banded in previous years or outside the specific area of trapping were retorded. Records were also maintained on the number of traps operated by wetland in order to evaluate capture/unit of effort at each trap site. Alberta Banding Two volunteers were recruited from the Colorado Division of Wildlife's permanent staff to work on cooperative duck banding crews in southern Alberta. The crews were active in Alberta throughout the month of August. All banding records were submitted to the U. S. Fish and Wildlife Service's Bird Banding Laboratory by the Canadian Wildlife Service. Therefore the.results of the Alberta trapping effort will not be reported here. �, J RESULTS Trap Locations Trapping was distributed over a total of 19 different wetlands in the 5 areas (Table 1). Hog Lake (BPNWR), Allen Basin (Yampa), Hall Pond (URV), Nolan's Pond (CM), Hooten's Pond (CM), and Williamson's Pond (CM) were added as trap sites in 1993, while Butch Cassiday (BPNWR), Porter's Feedlot (URV), Sander's Pond (URV), and Weir's Pond were dropped. Nearly all trap sites were located in Palustrine Emergent Persistent Wetlands, but some sites in the Colorado River Valley were in Riverine Upper Perennial Rock Bottom wetlands (Cowardin et ale 1979). Allen Basin is a high mountain water storage reservoir that contained mostly dead storage water at the time of trapping. Banding and trapping efficiency Over 2,200 mallards were banded during trapping in western Colorado in 1993 (Table 2) bringing the 3-year total to slightly over 5,000 mallards. Compared to 1992 the number of birds banded per trap/day in 1993 increased in the CRY (9.7 to 15.6) and the CM (5.2 to 6.8) areas, decreased on the BPNWR (13.3 to 4.8) and remained similar in UVR (6.8 to 6.4). Excluding the Yampa area, overall trapping success was similar in 1993 (7.5 mallards/trap-day) to 1992 (8.0 mallards/ trap-day). Only 28% of the mallards banded in BPNWR were immature birds. However, immatures mallards comprised 64 %, 59%, 79% and 79% of the CRY, URV, CM, and Yampa samples, respectively. Throughout the trapping area 61.9% of the mallards banded were immatures compared to 62.6% in 1992, and 68.3% in 1991. Only 14 northern pintail, the secondary target species, were banded (Table 3)~ Addition species banded were green-winged teal, redheads, bluewinged teal and/or cinnamon teal, gadwall and ring-necked ducks (Table 3). Band Reporting and Record Keeping All new banded birds and recaptures were submitted to the U. S. Fish and Wildlife Service's Bird Banding Laboratory on standard forms. Computer files. containing the number of birds banded by area, site, day, age and sex were constructed at the Colorado Division of Wildlife's Research Center. �8 LITERATURE CITED Szymczak, M. R. 1992. Harvest distribution of mallards and pintails banded preseason in western Colorado. Job Prog. Rep., Colo. Div. Wildl., Fed. Aid Wildl. Rest. Oct. Pp. 49-53. Szymczak, M.R., and J. F. Corey. 1976. Construction and use of the Salt Plains duck trap in Colorado. Colo. Div. Wildl., Div. Rep. 6. 13pp. Prepared by: ~J 1! 1;'(j~ Michael R. Szymczak Researcher/Scientist III �9 Table 1. Trapping locations during preseason banding in western Colorado. August- September, 1993. Wetland Area Browns' Park Natl Wildl. Ref. Gardner Park Res. Colorado R. Valley Name Hog Lake TION, RI03W, Sec 9, SW~ Flynn Marsh TION, RI03W, Sec 16, SE~ Spitzie Slough TION, RI03W, Sec 15, S~ Gardner Park Res. TIN, R86W, Sec 22, NE~ Allen Basin Res. T2N, R86W, Sec 9. N~ Latham's Slough T8S, R97W, Sec 27, SW~ Morse's Pond TIS, RIE, Sec 34. NE\ Ute Meridian Walker Wildl Area North Skippers' Island Uncompahgre R. Valley Markley's Pond Sweitzer Lake .Ha 11 Pond Cortez/Mancos Location TIN, R2W, Sec 36, 5W~ Ute Meridian TIN, R3W, Sec 14. NE% Ute Meridian T50N, R9W~ Sec 30, NW% TI5S, R95W, Sec 28, SW~ TI45, R94W, Sec 21. SE~ Totten Res. T36N, R1SW, Sec 20, NW~ Mancos Wetland T36N, R13W, .Sec 27, SW~ Weber Res. T36N, R13W, Sec 12, NE% Summit Lake T37N, R14W, Sec 33, SW~ Hooten's Pond T36N, R16W, Sec 14, N~ Williamson's Pond T36N, R1SW, Sec 31 SE~ Nolan's Pond T~7N, R16W, Sec 8, N~ �10 Table 2. Number of Mallards banded by area, site, age and sex and trapping efficiency during pre-Season trapping in western Colorado, 1993. Number of locals included in parentheses. No. Area Brown's Park Yampa Area Colorado River Uncomp. River CortezMancos Site AM Age[sex AF 1M Hog Lake Flynn Marsh Spitzie Slough Sub-total 87 75 54 83 Gardner Park All en Basin Sub-total N. Walker Skippers Is1. Morse's Pond Latham Slough Sub-total Markley's Sweitzer Lake Hall Pond Sub-total Nolan's Pond Totten Res. Weber Res. Mancos Wetld. Summit Lake Hooten's Pond Williamson's Pd. Sub-total GRAND TOTAL ~ Exclude Yampa -area sites. ~ 177 5 .s 14 76 0 9 _§_ 91 .iz 154 4 ~ 7 20 37 lQ 67 .sa 458 394 4 58 105 561 109(1) 91(2) 39 71 44 40 224 (1) 160(2) 281 215 151 647 .sz 12 32 11 0 2 2 12 12 10 1 1 2 _l .i _li 319 60 124(1) 1 24 44 193(1) 24 18 20 . 62 533 11 175 230 102 32 162 1 2 11 14 li 57 220 39 14 35 Total 23 li(3) 36(3) 19(2) 26(2) 45(4) 57 87 47 191 60 IF 24 70(2) 44 5 16 26 200(2) 23 68 32 1 12 29 _5 170 trap days 32 48 12 95 No. banded per traQ[day 5.5 4.8 3.5 4.8 51 u 71 182 97 7 31 59 .zz 469 756(7) 629(6) 2237 20 1 3 11 36 63 20 11 95 10 20 10 2 10 10 .i 69 29~/ 19.7 4.0 19.3 8.8 15.6 3.9 10.8 12.6 6.4 7.1 9.1 9.7 3.5 3 ..1 5.9 3.1 6.8 7.~ �11 Table 3. Number of northern pintail and other species banded by area, site, age, and sex in western Colorad0 greseason 1993. AgeLSex Total Sgecies Area Site AF 1M IF AM 2 Northern pintail Brown's Pk. Flynn Marsh Hog Lake 1 0 0 0 0 0 0 1. 1 1 Colo. R. Latham Slough 1 0 0 0 1 Uncomp. R. Markley's Pd. 3 2 2 1 8 0 CortezMancos Totten Res. Hooten's Pd. Total Green-winged Teal Yampa 0 0 Q Q 2 2 1 3 2 .i 14 Gardner Park All en Basi n 0 0 1 0 1 1 0 0 0 1 Walker Wildl. Latham Slough 4 1 1 1 1 0 3 1 3 Uncomp. R. Hall Pd. Sweitzer L Markley's Pd. 2 0 1 0 0 1 0 2 6 0 0 0 0 0 2 4 2 1 _! 1 Colo. R. CortezMancos Totten Res. Nolan's Pd. Williamson's Total 12 Blue-winged/ Cin. Teal Brown's Pk. Spitzi e Marsh Flynn Marsh Hog L~ke Colo. R. Morse's Pd. Total Ring-necked Duck 2 Q 7 Yampa Gardner Park Allen Basin Total 4 0 5 . ...J. _l 20 13 9 2 2 13 4 5 .s 49 0 3 1 0 0 0 1 4 0 0 3 0 1 10 1 Q Q 0 Z Z _! 4 0 2 Q Q 0 2 7 5 16 3{2} 2(2} 7 1 Q 4{2} 2(2) 1 8 Gadwall Brown's Pk. Flynn Marsh 0 0 1 0 1 Redhead Brown's Pk. Flynn Marsh 0 0 4 0 4 ��13 Colorado Division of Wildlife Wildlife Research Report May 1994 JOB PROGRESS REPORT Colorado State of Project Migratory Game Bird Investigations W-166-R-3 Work Plan _1_ Job 23 Job Title: Ecoloay of Waterfowl in Montane Wetland Communities Period Covered: 1 April 1993 through 31 March 1994 Authors: Robert L. Sanders, James K. Ringelman, Leigh H. Fredrickson Personnel: Robert L. Sanders, University of Missouri - Columbia; James K. Ringelman, Colorado Division of Wildlife; Leigh H. Fredrickson, University of Mi ssouri - Gaylord Memorial Laboratory; Joseph Todd, Co lorado Divi sion of Wi ldlife ABSTRACT Data on waterfowl use and habitat characteristics of 24 high elevation, montane wetlands were collected during the period 20 May -9 September 1993. Chemical and physical features of wetlands, vegetative composition, and invertebrate community composition and biomasses were recorded during three sampling periods corresponding with distinct phases in the reproductive cycles of mallard and ring-necked ducks. Semi-weekly waterfowl surveys were conducted throughout the entire field season .. Several differences in chemical and physical characteristics among beaver and glacia 1 (kett le) wet lands were apparent. Beaver wet lands genera lly had higher total, calcium and magnesium hardness, dissolved oxygen, and conductivity levels than did kettle wetlands. Beaver wetlands were circumneutral to slightly basic (pH=6.8-8.0) while kettle wetlands were slightly acidic (pH=5.8-6.9). Invertebrate numbers and biomass were also higher for beaver wetlands than for glacial wetlands. Beaver wetlands typically had an early season maximum in invertebrate biomass followed by a decline throughout the second and third sampling periods. Kettle wetlands had low initial invertebrate biomass, reached a mid-season peak and declined during the third sampling period. Beaver wetlands had higher invertebrate densities and.biomass than kettle wetlands. Eleven species of waterfowl were observed on the study wetlands. . Breeding species observed included ring-necked ducks (42.2%), mallards (28.4%), buffleheads (12.4%), green-winged teal (9.6%), wigeon (5.0%), cinnamon teal (1.4%), and Canada goose (1.0%). Duckling composition closely paralleled that of adults with the exception of bufflehead broods which were not observed on the study wetlands. Habitat use by waterfowl demonstrated a preference of Nuphar sp. habitats by ring-necked ducks, mallards were observed most frequently in emergent Carex sp. habitats, buffleheads preferred open water habitat types, while green-winged teal used these three habitats in equal proportions. Surveys and sampling will be conducted during 1994 using modifications of the methods employed in 1993. Analysis of all data will be continued through 1994 and a final report will be submitted in mid-1995. ��15 ECOLOGY OF WATERFOWL IN MONTANE WETLAND COMMUNITIES Robert L. Sanders James K. Ringelman Leigh H. Fredrickson P. N. OBJECTIVES 1. Describe montane wetland habitats based on physical, chemical, hydrological and vegetative characteristics. . 2. Describe aquatic macroinvertebrate communities as they relate to wetland habitat characteristics as determined in Objective 1. 3. Determine habitat use of waterfowl in relation to habitat data collected under Objectives 1 and 2. SEGMENT OBJECTIVES 1. Review literature on waterfowl biology and wetlands ecology in montane habitats. Draft a detailed study plan within the context of objectives and approathes contained in the program narrative. Determine sample sizes and sampling protocol, and develop testable statistical null hypotheses. 2. Based on data from pilot research in the study area (J. K. Ringelman, Colorado Division of Wildlife, unpublished data), a listing of high waterfowl use wetlands and low waterfowl use wetlands will be developed. High and low use ponds will be paired on the basis of size and origin. 3. Within both high and low use categories, sample wetlands with permanent water and those with non-permanent (seasonal) water will be selected in proportion to their abundance in the study area. 4. Within the permanent water classification, an equal number of kettle ponds and beaver ponds will'be selected as sample units. In the nonpermanent classification, only representative kettle ponds will be selected for study (beaver ponds are usually seasonally permanent wetlands). 5. Three representative plant associations (floating-leaved~ submersed aquatic, and emergent herbaceous) will be selected for sub-samplin~ within the permanent wetland classes identified in 4. Only the eme~gent association will be sampled in the non-permanent class. 6. Plant associations identified in 5 will be sampled for aquatic invertebrates in 1993 shortly after ice-out and in mid-summer (permanent wetlands only). 7. Waterfowl pairs and broods will be surveyed in a systematic manner using quiet observation techniques beginning in mid-May 1993. 8. Install weather stations and permanent water level markers in or near �16 sample wetlands to quantify hydrology and water budgets on site. 9. Indicator species will be used to designate plant associations for aerial coverage, seed sampling, and invertebrate sampling. Floatingleaved plant associations will be defined as that type dominated by cowlily (Nuohar luteum ssp. polysepalum). Submersed plant associations will contain pondweeds (Potamogeton natans) and watermilfoil (Myriophyllum spp.). The herbaceous emergent association will be dominated by sedges (Carex spp.), mannagrass (Glyceria borealis), and blue-joint grass (Calamaqrostis canadensis). 10. Benthic aquatic invertebrates will be sampled with core samplers. Standardized sweep net procedures will be used to collect invertebrates in the water column and at the surface. Statistical software will be used to compute power and sample sizes necessary for invertebrate sampling. Invertebrates will be preserved in a 90% ethanol solution prior to picking and sorting. 11. Variable mesh size gill nets will be deployed in sample wetlands to determine the presence or absence of fish. Numbers and species of fish captured will be recorded for each sample wetland. 12. An annual report covering the period 1 July 1993 through 31 March 1994 will be prepared and submitted by 1 August 1994. STUDY AREA Data were collected in the vicinity of Big Creek Lakes on the Routt National Forest approximately 33 km northwest of Walden, Colorado (Fig. 1). The 20.7 kmZ study area encompasses the upper reaches of the North Fork of the North Platte River, Forester, Goose and Shafer Creeks and contains approximately 160 wetlands of both beaver and glacial origin. Study area elevations range from 2573-2774 m (8520-9100 ft.). Annual precipitation, primarily in the form of snow, ranges from 64-76 cm (25-30 in.), making it one of the wettest areas in Colorado. Upland areas are dominated by lodgepole pine (Pinus contorta) with Engelmann spruce (Picea engelmannii) and subalpine fir (Abies lasiocaroa) more prevalent at higher elevations and on cooler north-facing slopes. Some of the more common plants in wetland areas include willow (Salix spp.), yellow pond lily (Nuphar luteum ssp. polysepalum), pondweed (Potamogeton gramineus and ~ filiformis), mannagrass (Glyceria borealis and ~ elata) and sedges (Carex· utriculata, ~ vescaria and ~ aguatilis). . Eleven species of ducks have been observed on the area, of these mallard (Anas plat yrhynchos) , green-winged teal (A. crecca carolinensis), cinnamon teal (A. cyanoptera), American wigeon (A. americana), ring-necked duck (Aythya collaris), and bufflehead (Bucephala albeola) are common nesters. Nonbreeding species include northern pintail (Anas acuta), blue-winged teal (A~ discors), wood duck (Aix sponsa), Barrow's goldeneye (Bucephala islandica), and common merganser (Mergus merganser). �17 METHODS A subsample of 24 wetlands (14 beaver and 10 glacial) were selected for intensive physical, chemical, vegetative, and aquatic invertebrate sampling and waterfowl observations. These wetlands were divided into high use and low use categories based on waterfowl pair and brood counts (Ringelman, unpublished data). Timing of sampling corresponded with three periods in the life cycle of mallard and ring-necked ducks as determined by field observations of birds on the area; P~riod 1 (Mallard Prenesting), Period 2 (Early Mallard Brood Rearing/Ring-necked Duck Prenesting), and Period 3 (Late Mallard Brood Rearing/Early Ring-necked Duck Brood Rearing). The following sections describe each portion of the project. Chemical and Physical Characteristics Water chemistry data were collected for the 24 study wetlands during three periods; 28 May - 2 June, 28 June - 1 July, and 26-30 July. All water samples were collected at the surface from the open water zone of each wetland. The following ten parameters were measured; nitrate, orthophosphate, hardness (total, calcium and magnesium), dissolved oxygen, hydrogen sulfide, color, pH, and conductivity. Nitrate (NO)) levels were measured using a Hach low-range nitrate test kit, model NI-14 (Hach Co., Loveland,CO). ·Orthophosphate (as PO.) was measured using a Hach orthophosphate test kit model PO-19. Total, calcium and magnesium hardness (as CaCO), Mg by. difference) was measured using a Hach total hardness and calcium test kit. model HA-4P. Dissolved oxygen (D.O.) was measured using a Hach dissolved oxygen test kit model OX-2P. Hydrogen sulfide (H S) was measured using·a Hach hydrogen sulfide test kit model HS-C. Water color was measured using a Hach color test kit model CO-I. pH was measured using a Hanna Instruments pHep+ probe (Hanna Instruments Inc.)'. Conductivity was measured using a Corning PS17 conductivity meter (Corning Inc.). 2 Physical parameters measured include water and air temperature, total basin area and shoreline length. Water temperature was measured concurrent with water chemistry sampling. Temperature readings were taken with a pocket case thermometer in the shallow water zone of each wetland. Air temperature was recorded using a hygrothermograph located at base camp and a maximumminimum thermometer located at the highest elevation wetland on the study area. Total basin area and shoreline length were determined using low level aerial photographs. Slides of each wetland were taken on 18 August 1993 from a fixed-wing aircraft using a hand-held 35 mm camera with a 50 mm lens and Kodachrome 64 ASA film. Wetland maps were drawn to scale using known reference points. Maps were digitized using Earth Resources Data Analysis Systems (ERDAS) and both basin area and shoreline length were determined. Hydrology Water budget data were collected on a weekly basis beginning in early June. Water level markers constructed of 1/2 in. PVC tubing were marked with 10 cm graduations and placed in each wetland. Water levels were measured concurrent with semiweekly waterfowl observations and recorded to the. nearest 1 cm. PreCipitation levels were measured within 24 hours of precipitation events using plastic tube-type rain gauges. Evaporation rates were recorded using three metal evaporation pans located at lower,.middle and upper �18 elevations on the study site. Pans were filled to a known depth with water and levels were monitored semi-weekly. Vegetation Wetland vegetational zones were mapped using low level aerial photography. Slides of each wetland were projected on graph paper and dominant vegetational zones were delineated. Wetland maps were then digitized using ERDAS and habitat acreages were recorded for each wetland. Maps will be field checked for accuracy during the J994 field season. Aquatic Invertebrates Aquatic invertebrate sampling was conducted monthly on all 24 wetlands concurrent with water chemistry sampling. Wetlands were stratified by vegetation type present the day of sampling. Ocular estimation of percent of the basin occupied by a vegetation type was made and four sweep samples and four core samples were taken in each vegetation type that comprised a minimum of 10% of the basin inundated by water. Sweep samples were made using a 0.5 mm mesh net with a opening size of 288 cm A metal rod was attached to the net by aim cord to standardize the length of the sweep. A random number of paces were taken in each vegetational strata to determine plot location. The metal rod was pushed into the substrate at the beginning of the sweep, the net was passed through the water. column between the surface and 15 cm depth, and the net containing invertebrates and associated organic material was raised at the end of the 1 m sweep. Excess vegetation collected in the net was cleaned of attached invertebrates, the sample was rinsed into a ziplock bag, labeled and returned to camp. . Core samples were taken adjacent to randomly selected sweep sampling plots. A 5.12 cm diameter PVC tube was inserted into the substrate to a depth of 10 cm, the sample was brought to the surface, rinsed using a 0.5 mm mesh screen and inspected for invertebrates. Sweep samples containing invertebrates were returned to camp, all non-animal matter was removed and samples were preserved using an 80% ethanol solution~ Samples were transported to Gaylord Memorial Laboratory and sorted during January and February 1994. Aquatic invertebrates were placed into one of following 20 taxa: amphipoda, anostraca, arachnida, chironomidae, coleoptera, copapoda, daphnia, diptera,' egg masses (amphibian and unknown), ephemeroptera, gastropoda, hemiptera, hirudinea, lepidoptera, odonata, oligochaeta, orthoptera, ostracoda, pelecypoda, and trichoptera. Individual organisms in each sample were counted, the sample was dried at 95°C to a constant mass and weighed to the nearest 0.001 g. 2 _ Waterfowl Semiweekly waterfowl surveys were conducted on the 24 study ponds. Breeding pair surveys were conducted between 0800-1800 hrs. beginning on 28 May. Ponds were approached 4uietly using vegetation and terrain to conceal the observer. Information recorded included pond number, time, number of birds, species, status (lone bird, pair, hen with brood, lone brood, multiple birds in a group), activity (surface feed, sub-surface feed, loaf, swim, sleep, courtship, alert), habitat type, and weather conditions. Brood surveys were conducted using similar quiet observation techniques between 0500-0830 hrs. and 1800-2030 hrs. during periods of peak brood activity. In addition to �19 the information recorded for pair surveys, brood size and age were also recorded for all broods observed. Time budget data were recorded for selected waterfowl whenever conditions were favorable. Fifteen minute budgets were conducted on mallard and ring-necked ducks using continuous sampling techniques. Activities were divided into nine categories; surface feeding, subsurface feeding (tipping up), diving, locomotion, resting/sleeping, comfort movements, alert, social interactions/courtship, and out of sight. Cassette recorders were used to record activities in the field and data were entered into a time budget software program upon return to camp. Food habit collections were made for mallard, ring-necked ducks, and green-winged teal. Birds were observed feeding for a minimum of 15 minutes, collected by shotgun, the esophageal and proventricular contents removed and placed in an 80% ethanol solution. Bird car~asses were frozen and digestive tract contents were stored pending analysis. RESULTS Chemical and Physical Characteristics Water chemistry parameters collected for the 24 study wetlands demonstrate a marked difference in chemical composition based on pond orlgln (see Appendix A). Nitrate and orthophosphate levels were too low to be detected by the test kits. Total hardness ranged(from 17-85 mg/l with ponds of beaver origin generally higher in hardness than those of kettle origin (Fig . 2). Dissolved oxygen levels also showed marked differences between beaver and kettle ponds (Fig. 3) and ranged from 1-10 mg/l. Hydrogen sulfide levels were too low to be detected by the test kit .. Color values for wetlands ranged from 20-115 units, but no distinct differences are apparent between beaver and kettle ponds (Fig. 4). The pH values of study wetlands were close to circumneutral for all ponds, with beaver ponds ranging from 6.7-8.1 and kettle ponds ranging from 5.8-6.9 (Fig. 5). Conductivity values showed marked differences between beaver and kettle ponds and ranged from 20-170 ms (Fig. 6) . Water temperatures for sampling period 1 averaged 16°C for beaver ponds and 14.8°e for kettle ponds. Water temperature data for periods 2 and 3 were discarded due to equipment problems. Air temperatures ranged from a low of 4°C (24°F) recorded on 6 July to a high of 28°C (82°F) recorded on 29 July. Mean monthly temperatures were as follows; 18-30 June 13°C (55.7°F), 1-31 July 9.5°e (49.2°F), 23-29 August 8.2°e (46.7°F). Due to equipment damage, temperature data for 1-22 August were not collected. Physical features were calculated from aerial photographs. Total basin areas for the 24 study wetlands ranged from 629 - 80,402 or (Table 1). Wetland perimeter (shore length) ranged from 156 - 1724 m. Hydrology Water levels in the study wetlands were recorded on a weekly basis beginning in early June. Typically, beaver ponds had relatively stable water levels throughout the spring and summer (Fig. 7), while kettle ponds reached maximum water level shortly after snowmelt with water levels declining gradually over the summer period (Fig. 8). Precipitation over the period 1 June through· 31 August averaged 7.07 em �20 (2.78 in.) among the three weather stations. Evaporation over the same period averaged 38.69 cm (15.23 in.) for an average net evaporative loss of 31.62 cm (12.45 in.)(Table 2). Veaetation Vegetative composition of each study wetland was determined using aerial photographs taken on 18 August 1993. Five habitat types were distinguishable on aerial photographs; emergent (primarily Carex spp.), nuphar, floating leaf glyceria, broad-leaf potamogeton, and open water. Several other vegetation types were present on the study wetlands including myriophyllum, narrow-leaf pondweed, chara, ranunculus, and menyanthes, but resolution of aerial photographs did not allow accurate delination. The number of habitat types in each wetland ranged from one to seven and vegetative composition varied considerably among wetlands (Table 1). Aquatic Invertebrates A total of 34,195 organisms (65.725 g dry weight) were collected in 868 sweeps during the three sampling periods. The composition of aquatic invertebrate communities varied among wetland types (Fig. 9). Carex habitat type was the most productive invertebrate habitat during all three sampling periods for beaver wetlands and during periods 1 and 2 for kettle wetlands (Fig. 10). Gastropods, amphipods, ephemeropterans, odonates, and hirudineans comprised 94% of the total invertebrate biomass found in beaver ponds (Fig. 11) while gastropods, coleopterans, trichopterans, hirudineans, odonates, and ostracods comprised 83% of the total biomass in kettle ponds (Fig. 12). Gastropods were highest in dry biomass during all three periods for both beaver and kettle wetlands, but this difference is due to the weight of the shell which was included when calculating total dry biomass.Seasonal population trends indicated a ~pring maximum biomass followed by a decline over the second and third sampling periods in beaver ponds (Fig. 13), while invertebrate biomass in kettle ponds generally demonstrated an early season low, a mid-season peak and a decline in biomass through the third sampling period (Fig. 14). Core samples taken concurrent with sweep sampling indicated an almost complete absence of invertebrates in the bottom substrate of both beaver and kettle ponds .. Because of this lack of invertebrates, samples taken in each strata were inspected for invertebrates, invertebrates present were noted and the core was discarded. Waterfowl Eleven species of waterfowl were observed on the study area .. Based on the number of birds observed during all periods combined, ring-necked ducks were the most commonly observed breeding species (42.2%), followed by mallard (28.4%), bufflehead (12.4%), green-winged teal (9.6%), wigeon (5.0%), cinnamon teal (1.4%), and Canada goose (1.0%}(Fig. 15). Non-breeding species observed included Northern pintail, blue-winged teal, wood duck and Barrow's goldeneye. Waterfowl use of individual study wetlands ranged from 0-77 birds observed on each wetland throughout the study period. Bird use ranged from DIDO birds/ha. (Fig. 16). Waterfowl use of wetland habitat types varied by species (Fig. 17). Mallards demonstrated a preference for emergent carex spp. types, ring-necked �21 ducks preferred nuphar sp. habitats, while buffleheads were most commonly observed in open water. A lack of.use by waterfowl of submergent habitats may be due to difficulties in distinguishing these habitats from open water during waterfowl observations. . Nesting chronologies were estimated from known age broods using literature values for average clutch size, incubation periods, and fledging dates (Bellrose 1980). Mallards generally initiated nests earliest followed by green-winged teal, wigeon, and ring-necked ducks (Table 3). One Canada goose brood was observed on a nest just after hatching but no further observations of goose.broods were made. Other common breeding species on the study area were not observed with broods on any of the study wetlands. Duckling species composition on the area closely paralleled that of adult birds with the exception of bufflehead broods which were not observed on the study wetlands. Ring-necked duck broods were most ~ommonly observed (58.8%), followed by mallard (19.8%), green-winged teal (12.3%), wigeon (7.5%), and Canada goose (1.6%)(Fig. 18). FUTURE DATA COLLECTIONS AND ANALYSIS ChemicaL physical, vegetative and invertebrate sampling and waterfowl observations will be continued through the 1994 field season. Several modifications will be made to procedures used in 1993. Water chemistry analysis will be intensified and will include the use of a Hach DR2000 spectrophotometer to increase the accuracy of nitrate, phosphate, dissolved oxygen, color, and hardness measurements. A mid-season water sample will be . collected and submitted to the University of Missouri limnology laboratory for nitrate, phosphorous, and algal chlorophyll analysis. Depth contour maps will be constructed for all study wetlands. Vegetation maps constructed from 1993 . aerial photographs will be field checked for accuracy. Monthly aerial· photographs of study wetlands will be taken to determine seasonal changes in surface water acreages and vegetational composition. Maximum-minimum thermometers will be placed in representative wetlands and water temperature fluctuations will be recorded. Insect emergence traps will be placed in a subsample of study wetlands to record emergence patterns of aquatic insects. Waterfowl food habits and time budget data collections will also be intensified during the 1994 field season. Analysis of all data will be continued through April 1995. Invertebrate samples collected in 1994 will be sorted, counted and weighed in Fall 1994. All data collected during the 1994 field season will be compiled using methods similar to those used on 1993 field data. Statistical analyses will be performed on all data. A final report (thesis) will be submitted in mid-1995. Prepared by: Robert L. Sanders Research Assistant . ~.7JJ.(1~ �22 Table 1. Wetland basin size, perimeter and aerial coverage of major vegetation tvpes, 18 August 1994·. Vegetation Type Pond# Basin Perim. Emera. Nuphar FI.Gly. FI.Pot. Ooen Beaver NF-06 09 12 13 19 37 53 65 G-16 18 19 21 24 50 1796 491 922 3959 1908 915 739 646 1826 31766 9357 2949 1651 1948 692 205 156 748 512 287 396 258 593 925 882 435 362 525 173 51 39 187 128 72 99 64 148 5193 221 108 90 131 2034 12814 15583 1284 7231 4111 20961 629 80402 18072 221 643 698 201 464 332 1724 193 1700 686 1592 6038 5341 715 4437 3659 12429 248 17151 11418 536 53 586 381 111 733 17219 4866 70 68 52 143 68 200 1242 387 883 3133 1399 843 529 582 893 9354 4270 2560 1493 1617 Kettle NF-79 80 83 84 86 F-43 50 S-14 18 28 327 566 746 4671 4528 1148 4240 342 35557 3479 66 881 765 167 2402 30 933 9 861 17771 1398 • All values expressed as (m") except perimeter (m). Submergent vegetation types are included in the open water habitat type. �,., ., _.) Table 2. Precipitation and evaporation Lakes Studv area, June-August 1993. data Weather 1993 station Summary - Big Creek Lakes Ave. Evaporation/Day Station (rom) June July August 3.2 5.0 5.5 Ave. Precipitation/Day 0.67 0.58 1.09 Net Monthly Water Balance (Estimated) - 2 station 4.5 4.6 5.6 3.4 4.1 1.9 0.14 0.54 0.75 1.04 0.76 1.33 (mm) Precipitation Evaporation Net July - Precipitation Evaporation Net August station (mm) June July August June 1 for the Big - Precipitation Evaporation Net Total Net Water Balance Cmm> June - August 1993 4.2 -102.0 97.8 20.1 96.0 75.9 31. 2 -135.0 -103.8 - 18.0 -155.0 -137.0 23.6 -142.6 -119.0 16.7 -127.1 -110.4 33.8 -170.5 -136.7 41.2 -173.6 -132.4 - - -349.6 -355.2 23.3 58.9 35.6 -243.8 Creek 3 �24 Fiq. L Location of Biq Creek Lakes study area, Jackson County, Colorado. �25 Fiaure '2. Hardness - Total, Ca, and Mg mg/! CaC03 'oo~----------------------------------------~ i--------- -- ...... ---. 06 09 12 13 19 37 53 65 16 18 19 21 24 50 79 80 83 84 86 4~ 50 14 18 28 Beaver _ Kettle Calcium &1\\\\\\1 Magnesium Measured wi Hach Total Harcness and Calcium test kit HA-.P Figure 3. Dissolved Oxygen Big Creek Lakes Study Area 1993 mg/l 12r-------------------------------------------~ 10~----~~----------~----------------~---4 c- -!il-------------i ,. t-,-,::~--------------l;:f- ,;~ .jH..~---1"r;~:Hiif+-*FH\'l 4 r-.': .-. 2 ,. '. .: . . - . .: 1 06091213193753651618 19212450 ~\\\\'"6/20/93 " fa -l~! 79808384864350141828 Seaver Measurements taken wi Hac!! D.O. test kit OX-2.P : '. Kettle D 7/26/93 �26 Figure 4. Color Big Creek Lakes Study Area 1993 units 120 •• -~J~~ .._J _~ 1 100:-; -------- I 80 :;f,,----I··- 60 40, 20 I I I - -_ ... _-- r- '. - -{I----, -- fil' ~I--I-- :.~. ~ :·t--~rtil:l~-~ .1--- 'I- r 1 O~~~~wm~~~~~~um~mu~UD~~~MU~~ 0609121319375365161819212450 79808384864350141828 Seaver Measured w/Hach Kettle 5/28/931 ~\\\\,\j 6/20/932 I 7/26/933 I Calor test kit CO-l Figure 5. pH Big Creek Lakes Study Area 1993 pH 9 .~-~ .. a 7 : a 5 I ;-- r' j J , , : , 06091213193753651618 19 212450 Seaver Measurements take wi Hanna Instruments pHep· 5/28/93 probe II 79808384864350141828 Kettle _ 6/20/93 CJ 7/26/93 '. �27 Figure 6. Conductivity Big Creek Lakes Study Area 1993 mS 200 j 150;,_--,...---H----n-----i~--·-·- -... _... - .--.-._.'-' I ~~ sr O~~~~~MaUM~~KUUD~~~~~~~~~ 0609121319375365161819212450 Beaver Meaaurements taken wi C,jrning ?S-17 Conductivity 5/28/93 meter 79808384864350141828 Kettle _ 6/20/93 0 7/26/93 . �28 Figure 7.. Mean Water Levels Beaver vs. Kettle - 1993 em 10~, ----------------------------------------~ 0~1~~~~~==~-=~===========i I -1 0 ,,-------__:~. ... _._---------j , -20, -_.. _._- -30 -- - 1 -40 ,------------- -:;0 I-- -- .. ..--.-- 2 -- ._- -_. .. . .--._-'._.. .-----.---.... .__. _;ol~·--~~--~~~~ 1 _--- 3 __ ~~_~ 4 5 6 7 - .. -- .. - _.-.- __ ~~ __ ~ 8 9 10 11 12 ~ 13 14 Week - Beaver -.- Kettle Week 1 corresponds with.30 May - S June Mean weekly water level ot NF-09.12.13.19 vs. NF-79.83.84.aS Figure 8. Mean Water Levels . Kettle ponds - 1993 -ao~-----------------100~~--~~--~--~_.--~~~~--~~----~ 12345 6 7 a 9 ro n ~ Week W•• 1e 1 corresponds NF-79 with -:- NF-83 30 May - 5 Jun. - NF-84 --- NF-66 ~ M �29 Figure 9. Invert Composition Big Creek Lakes Study Area Period 1 ~ 50~----------~------------~--------------~ 40~---------------30 _ .. ------. --- 20~--------------- O~~~~~~-=~~~~~~L-~~~~~~~~ Amp Ara Chi Col Cap Oap Dip Eon Gas Hem Hir Odo Ost Pel Tri Taxa Period 2 ~ 60-. -------------------------------------------I I so .-------------------~ Amp Chi Col Oap Eon --------------------~ Gas Hem. Hir Odo Ost Pel Tri Taxa Period 3 ~ 60~------------------------------------------~ 50 1--------- 40~-------------30~----------------20~------------10 Ou.----~~~~--~~~~~~~~~~~~~--~ Amp Ara Chi Col Cop Cap Cip E~h Gas Hem Hir Odo Ost Taxa _ or. Composition, all haaitat types comained. by dry wt. Beaver ~~~~\\'I Kettle Pel Tri �30 Fiaure 10. Invert Habitat Use ._.. Period 1 o Carex Chara Nuphar Glyceria ocen Water Habitat type Period 2 O.3--------------------------------~----------~ __ . ---------------- o Colrex Cllara Eleac ._--_.- __ ----- Glyc Myrta . ... ---- .- ---_., -_- Nuphar N. Pot. Oaen Water Habitat Type Period 3 0.3.---------------------------------------------. ---- ----_. - - -_._- 0.2 Carex Chara Eleoc Glyc Myrto Nuphar N. Pot. Habitat Type - Beaver Pond _ Kettle Pond B. PotOgen Water �.." .)1. Figure 11. Invert Composition Beaver Ponds Period 1 (g) 0.16-~------------------------------------------------------------------~ 0.14 0.12JI-----j;'I~--------------------------~---------~ 0.1 0.08.11----;·;·; --------.--!: 0.06 ":I---j 0.04 ~!---= 0.02 o Carex Chara Nuchar . Open Water Habitat type Period 2 (g) Carex C:tara Myriophyllum N. PotamoQ. Nuphar 0::18" Water Habitat type Period 3 (g) 0.14 0.12 0.1 ~I----,;;;;,.---------------... U-t,It---------------- ..- ----..---..... - _.--.- . 0.08 0.06 0.04 0.02 0 C.arex Chara MyrloQhyllum N. PotamoQ. Nuphar OQ." Water Habitat type _ Amphlpoda ·a Odonata Averaqe biomass/sweep _ El)hem a Hirudin •• CJ C Gastropoda Other �32 Figure 12. Invert Composition Kettle Ponds Period 1 (g) ~:~: L- -----.--.-. --.--. .-.. 0.04 ~- -.-0.03 - -. ----- --- - -_ -- -... ... J--- .. -_ ..---.------- 0.02 0.01 ~ .----. ... _.-.--- '" .----------1 ---- o Carex Glyceria Open Water Nucnar Habitat type Period 2 ._J --- ._- ------. Glyceria Carex --- ._ _j _ Open Water Nuphar Habitat type Period 3 (g) 0.06 -.------------------------" ----------f.=:J 0.05 0.04 .----- .-- --i:li: J:, ----I . n Ji·,I-·--------- 0.03 ---I.:!:~If__-------- Carex Glyceria Open Water Nuphar Habitat type - Coleopt.ra _ Dlpt.ra 0 Gaatropoda _ 5 Trichopt ~ Ostracoda IE Hirudinea IlliiiiiI Other Average biomass/swe.p Odonata . �33 Figure 13. Invertebrate Trends Beaver Ponds - 1993 r----------------------~ 0.03 0.0 15-;----=----.:....;:___;~..- I 0.0 1I~----:-----=:::"..--.--------==-- 0.005''-----,--- o~·I ~===============X=_ 05/30/93 ~ 07/30/9 06/30/93 - Amphipoda -+- Oiptera -:- Hirudinea ~ Odonata ~ E;Jhmeropt ave. biomass/sweep Figure 14. Invertebrate Trends Kettle Ponds - 1993 (g) 0.014r-----------------------------~ 0.012~------~---~~---------~ 0.01~---------~~----~~------~ 0.008 ~-------~':-------------:::,....,.--------0.004 0'006~~~ 0.002 o 05/30/93 06/30/93 07/30/9 Date Sampled - Coleoptera -- .Dlptera -- Hirudinea -- Odonata - Ostracoda -- Trlchopt mean Invertecrate biomass/sweep .11 ".cltats combined �3c. Figure 15. Waterfowl Observed - Bia Creek Lakes 1993 Ring-neck· 211 Gin. tea! Total birds observed. periods combined Euffleheac all Figure 16. Waterfowl Use Big Creek Lakes 1993 8irds/ha 120~----------------------------------------~ 100 ~------------- -- ao~----------------. 60~---------------40 20 o 0609121319375365161819212450 Beaver 79808384864350141828 Kettle _ Waterfowl use/ha. all species combined. all periods Waterfowl/ha �35 Figure 17. Habitat Use Waterfowl # Obs. 100r------------------------------------------, Carex _ Nuphar Mallard Open Water &\W~\!:·: Ring-neck ' Submergent I r;:.WTeal _ Terrestrial 8ufflehead All periods combined Figure 18. Ducklings Observed Big Creek Lakes 1993 Ring-neck 149 C. goose 4 Wigeon. 19 Totai numbers ob.served, all periods �Ap;:endix A. Water c:-:emisuy data fer Sig Creek L3l<es srudy wetlands. 00 1 1 1 1 1 1 06 09 1 1 . 5.3 :: :: :: 3 3 1 1 1 13 19 6: 15 13 19 ::. .... "'~ a a a s a B B B 3 3 3 sa is eo .. ...-~ 13 83 _::I • 86 ..; . ~ 4 1• 4 06 09 1 1 13 1 19 37 53 65 3 3 3 3· a 13 B B B B B 3 19 B 3 3 .3 1 1 2~ 24 50 B B 1 1 3 3 1 1 1 16 79 80 83 84 86 2 2 43 4 14 18 4 4 1 1 • !O 2a 06 09 12 13 B It It It It It It It It It It B B B B OS/23/93 OS/28/93 OS/23/93 OS/23/93 OS/28/93 OS/28/93 0.5/28/93 OS/28/93 OS/28/93 OS/28/93 OS/23/93 OS/28/93 OS/28/93 OS/28/93 OS/28/93 OS/29/93 OS/28/93 OS/28/93 OS/28/93 OS/25/93 OS/29/93 05;:Z8/93 OS/23/93 OS/28/93 06/20/93 06/20/93 06/20/93 06/20/93 06/20/93 06/20/93 06/20/93 06/20/93 06/20/93 06/20/93 06/20/93· 06/20/93 06/20/93 06/'2.0/93 06/'2.0/93 06/'2.0/93. 06/'2.0/93 06/'2.0/93 06/'2.0/93 06/'2.0/93 06/'2.0/93 06/'2.0/93 06/'2.0/93 06/'2.0/93 07/'2.6/93 07/'2.6/93 07/'2.6/93 07/'2.6/93 7.1 aa -1._ 1 100 7.0 o 7.1 6.S 6.S 100 6.9 6.i 140 -1._., 100 60 60 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 C.O 0.0 0.0 C.O C.O C.O C ., 0.0 100 0.0 90 90 160 0.0 0.0 0.:1 0.0 0.0 0.0 0.0 :.S 30 30 30 0.0 6.C 30 6.0 20 -, 1._ 7.1 6.9 6.1 6.5 0.0 6.3 30 O.C O.C O.C O.C O.C 6.0 30 0.0 6.: 40 6.4 0.0 0.0 7.1 60 90 110 iO O.C O.C 5.C 5.0 O.Q 7.3 1.:0 7.2 7.2 i.3 SO 60 90 a.o a.o 7.4 140 6.: 7.:' 7.3 6.9 so 7.1 90 100 100 7.2 7.1 7.1 7.4 6.3 6.1 6.8 6.5 6.5 100 150 60 40 0.0 8.0 9.0 7.0 7.0 5.0 '2..5 4.0 4.5 SO 4.0 4.0 0.0 0.0 1.0 6.4 50 6.6 6.2 40 40 6.7 30 6.4 40 100 140 7.3 7.2 7.3 7.2 5.0 6.5 4.0 30 40 150 170 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 60 90 -I. _ 1 MG 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 i.1 c.-; 5.0 2.0 9.0 1.0 5.0 1.0 c.o 0.0 c.o 0.0 C.O C.O C.O c.O 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 ccrca 53 as 95 105 53 sa ao as 11.0 ue 92 C.O C.O 0.0 c.: c.O O.C C.C C.C O.C C.: C.O C.O C.O O.C C.O C.O l.O iC :.0 45 50 45 60 63 53 c.: C.: C.: c.... -,., f" ,., 3.0 -•.....,., 4.0 3.0 3.0 l.O 3.0 1.5 2.0 2.0 2.: 4.0 '2..0 3.0 3.0 3.0 3.0 5.0 2.0 1.0 1.0 '2..0 1.0 2.0 2.0 2.0 1.0 1.0 4.0 4.0 4.0 4.0 1.0 1.': 1.5 3.0 1.0 2.0 2.0 1.5 1.5 3.5 1.0 0.5 0.5 1.0 0.5 Z.O 0.5 0.5 0.5 0.5 iJ a.: sa 83 SO 35 1.0 1.0 1.0 1.0 1.5 1.5 1.5 1.0 80 90 1~0 US 0.5 0.5 50 1.0 0.5 60 5: 95 65 as 55 us ·1.0 1.0 1.0 1.0 0.5 0.5 0.5 0.5 55 2.5 45 2.5 3.5 1.5 1.5 0.5 2.5 1.5 30 40 1.0 1.0 65 95 85 so �..,.J D~ I'OND CRIt::::3 1 1 1 1 19 37 c:_oJ 65 15 S S S S S 13 a 07/26/93 07/26/93 07/26/93 07/26/93 .07/25/93 07/25/93 19 2~ 24 3 07/25/93 B B B A 07/25/93 07/25/93 07/25/93 07/26/93 07/26/93 07/26/93 07/25/93 07/26/93 07/26/93 07/26/93 07/26/93 07/26/93 07/26/93 .3 - :3 3 .3 .3 1 1 1 1 1 '2 :2 4 4 4 so 79 ao a3 a4 86 43 50 14 13 28 x x: x X A x: x: z x: DA...-:':: !'!:: ccxn 7.2 7C 7.2 7.0 7.2 7.2 40 -- ,., I. sa Isa 7C _ 90 I. _ EO 8.0 a.l 7.4 6.3 6.7 6.j 6.3 6.4 6.3 6.5 6.4 6.9 6.5 90 140 120 40 70 20 DO 6.0 8.0 6.0 5.0 6.0 8.0 9.9 9.9 9.0 6.0 4.0 5.0 s.o sa 5.0 40 50 40 30 30 40 4.0 4.0 4.0 7.0 5.0 5.0 'l:!:: 4.0 2.0 2.0 4.0 3.0 3.0 4.0 4.0 4.0 5.0 1.0 1.0 1.0 2.0 1.0 2.0 2.0 2.0 1.0 2.0 a 2.': 1.0 1.5 2.5 2.0 2.0 2.0 2.0 KG 1.5 1.0 0.5 1.5 1.0 La 2.0 2.0 2.5 1.S 3.S 1.5. o.s o.s o.s 0.5 La o.s 1.0 1.0 1.0 0.5 1.0 o.s 0 •.5 1.0 0.5 1.0 1.0 1.0 0.5 1.0 c=I.OR s: 75 35 20 sa 65 '60 liO lie 65 lie 8.5 75 1~C! 70 llO 95 95 ---- a.5 I ��39 Colorado Division of Wildlife Wildlife Research Report October 1994 JOB PROGRESS REPORT State of _--'C:<..::oe...;l'-"o'-'-r=a_,._do::<._ Avian Research - Miaratorv Game Bird Project W-166-R-3 Investigations Work Plan _1_: Job 24 Job Title: Integrated Waterbird Management Studies Period Covered: Author: 1 April 1993 through 31 March 1994 James K. Rinaelman Personnel: J. Rinaelman. M. Szymczak. Colorado Division of Wildlife: D. Cooper, Colorado State University; M. Laubhan, National BiologicaJ Survey. ABSTRACT Russell Lakes State Wildlife Area (RLSWA), a 1,550 hectare wetland complex located in the northern San Luis Valley, was selected for a study of wetland resource dynamics and use by waterbirds. In November, personnel from the Division of Wildlife and Colorado State University (CSU) met at RLSWA to discuss study objectives and logistics. A site was selected for construction of experimental ponds that could be flooded with surface water or well water. Weather conditions throughout the winter and early spring prevented use of a .levee plow and pond construction was delayed. In January of 1994, personnel from the Mid-continent Ecological Science Center, National Biological Survey· (NBS), joined the project in a cooperative effort. NBS personnel will assist in fieldwork and production of maps of the study area in a Geographic Information System (GIS). Previous studies and censuses conducted on the study area were reviewed to determine likely waterbird species composition and chronology of use. Based on this information, 10 species, representing 9 foraging guilds, were tentatively selected for study of seasonal food habits in relation to reproductive status and aquatic invertebrate community characteristics. These and 3 additional species were selected for study of diurnal activities in various habitats, including shallow emergent, tall emergent, permanent open water, seasonally flooded open water, saltgrass, upland grass and upland shrub. Sites in wetland habitats will be selected for intensive aquatic invertebrate sampling by personnel from CSU. Fieldwork was set to begin in May of 1994. Prepared by: C!f-.,:z:~ James K. Ringelman Researcher/Scientist V ��Colorado Division of Wildlife Wildlife Research Report October 1994 JOB PROGRESS REPORT State of _--'C=o"-l,__"o,_,_r_",a""'d=o _ Project W-166-R-3 Work Plan 10 Job Title: : Job _1_ Cooperative Management Programs Period Covered: Author: Miaratory Game Birds Investigations 01 April 1993 through 31 March 1994 Michael R. Szymczak Personnel: James K. Ringelman and Michael R. Szymczak. Wildlife Colorado Division of ABSTRACT Recommendations for wetland habitat improvements and/or management were provided for public and private land managers across the state. Proposals for funding projects with Duck Stamp monies were evaluated and rated. A survey of Colorado waterfowl hunters was conducted and the results tabulated. Presentations on various aspects of waterfowl ecology were given at training and educational schools, workshops, and short courses. Responsibilities as Colorado's representative on Pacific Flyway Study Committees and Central Flyway Technical Committee, including chairmanship of the Central Flyway committee, were fulfilled. Waterfowl surveys were conducted in North Park and on selected private lands in the San Luis Valley. A 3-5 year cooperative post breeding goose banding program was continued in the Cortez-Mancos area. Technical assistance was provided as a member of the Continental Evaluation Team for the North American Waterfowl Management Plan and the Adaptive Harvest Management Working Group. Surveys to determine the current status of Colorados' band-tailed pigeon population were initiated and analyzed. ��43 COOPERATIVE MIGRATORY BIRO MANAGEMENT PROGRAMS Michael R. Szymczak James K. Ringelman In 1988, the Colorado Division of Wildlife (CDOW) created the Migratory Game Bird Program Unit (MBPU) within the Terrestrial Wildlife Section. This administrative change combined all individuals having statewide responsibilities for research and management of migratory game birds. Members of the MBPU work in concert to improve migratory bird management in Colorado. This job was created to allow team members to participate in these management programs. P. N. OBJECTIVES 1. Participate in developing and implementing habitat-based waterfowl management plans on a statewide, habitat region, and project basis. 2. Advise state and federal land managers on beneficial habitat acquisitions and/or developments and provide expertise in preparation of development and/or management plans. Advise private land managers in developing habitat management plans and assessing impacts on waterbird populations. 3. Present information on the principles of waterfowl management to workshop attendees, educational classes, and conservation organizations. 4. Participate in migratory bird management meetings at the state and flyway levels. 5. Cooperate in developing surveys and techniques that will assess the impact of migratory bird management programs. SEGMENT OBJECTIVES 1.. In conjunction with the Statewide Waterfowl Management Plan, continue work on habitat region plans. 2. Provide biological expertise on waterfowl biology and wetland development programs on the Brush, Yampa, Russell Lakes, Horsetheif Canyon· and Red Lion State Wildlife Areas, Hebron Ponds, Walden Reservoir (BLM-COOW), various wetland complexes managed by the U.S. Forest Service, and other areas where requested. 3. Prepare and present lectures on migratory game bird management when requested. 4. Compile appropriate population status and hunter behavior information and represent Colorado at Western and Central Migratory Upland Game Bird meetings and Pacific and Central Flyway Technical Committee and Council meetings. Colorado chairs the Central Management and Council this year and will act as host state. Attend migratory game bird program and �44 biologist meetings in Colorado when requested. 5. Provide methodology for wetland habitat and migratory game bird population surveys when requested. 6. Cooperate in Canada goose trapping and banding operations in th~ DurangoCortez area, Middle Park, and the Gunnison area. 7. Serve on the U.S. Fish and Wildlife Service's Adaptive Harvest Management Committee and North American Waterfowl Management Plan Evaluation Team. 8. Conduct preliminary mail and observational surveys on the location and general abundance of band-tailed pigeons (Columba fasciata) using late summer/early fall feeding areas described by Braun (1976). 9. Conduct experimental surveys of waterfowl breeding pairs on wetland development units under management by the U. S. Fish and Wildlife through the Partners in Wildlife program. RESULTS Waterfowl Management Plans Very little was accomplished on habitat region plans. The San Luis Valley (SLV) Waterbird Plan continues to be used to make mahagement decisions in the SLV. I Wetland Developments and Acquisitions Potential sites for wetland developments and existing wetlands were visited on the Oak Ridge, Yampa and Horsethief SWA's in northwest Colorado and the Wellington SWA in northeast Colorado. Recommendations for wetland construction, water management, and management of existing wetlands were given to CDOW management personnel. . . The potential of Cooper Slough near Ft. Collins, and an area just north of Windsor as wetland open space was communicated to the cities of Ft. Collins and Windsor, respectively. Advise on improving wetlands in Beebe Draw was conveyed to private landowners near Greeley. We toured and consulted with U. S. Bureau of Reclamation (BOR) on additional wetland developments on the Colorado River Wildlife Area near Grand Junction and toured and commented on proposed developments at Dawson Draw {Soil Conservation Service - BOR mitigation for de-salinization project} near Cortez. A potential wetland development proposed by the U. S. Forest Service on·Sawls Creek northeast of Bayfield was visited. . In November 1993, project personnel assumed additional responsibility for leading and administering the Duck Stamp wetland development program. Ringelman became the CDOW'S representative on the Colorado Waterfowl Stamp Committee (CWSC), which deals with developing contracts with publishers and/or artists for production and sale of art work, and decides how money obtained from duck stamp print sales will be spent. Ringelman retained his membership, �while Szymczak became the chairman of the multi-agency Waterfowl Habitat Project Review Committee (WHPRC), which reviews an~ rates wetland enhancement and acquisition proposals from land management agencies for funding with Colorado State Duck Stamp monies. As chairman, Szymczak obtained the status of projects funded in previous years by the WHPRC, obtained additional information needed to complete project proposals submitted for 1994-95 funding, distributed proposal packets to committee members, and scheduled the meeting to review proposals. Waterfowl Hunter Attitude Survey In cooperation with the Human Dimensions Research Unit at Colorado State University (CSU), a waterfowl hunter attitude survey was conducted using a random sample of 1991 Colorado waterfowl stamp purchasers that indicated they were waterfowl hunters. The sample was stratified by Central and Pacific flyway hunters. A follow-up telephone survey was conducted of non-respondents to the initial survey, to examine non-response bias. Most data analysis has been completed, but the final report has not been written. On a similar assignment Ringelman attended 4 meetings of the committee that is examining the reduction in the number of small game hunters in Colorado. Informational Programs Formal presentations were given: on wetland management and waterfowl ecology to the CSU Wildlife Management Short Course; on wetland management to Project Wild teachers; on Adaptive Management and Waterfowl Ecology at the Ducks Unlimited Symposium in Winnipeg, Canada; on duck identification and general waterfowl management to the CDOW District Wildlife Management Trainees; and on designing wetlands for waterfowl and applying for Colorado Duck Stamp and Ducks Unlimited MARSH monies at the CDOW Wildlife Technicians Workshop. Waterfowl Technical Committee and Council Meetings Project personnei attended the July 1993 meetings of the Central Flyway Technical Committee (Ringelman), the Pacific Flyway Study Committee (Szymczak), and the respective Councils. Waterfowl population status was reviewed along with characteristics of the 1992-93 waterfowl hunting season harvest, and proposed 1993-94 hunting season recommendations were formulated and forwarded through the 2 Councils to the USFWS Regulation Committee. Populations of specific interest to Colorado whose status was reviewed in July were (1) breeding and wintering mallards inhabiting both eastern and western Colorado and (2) the Rocky Mountain, Hi-Line, and shortgrass prairie Canada goose populations and (3) the western segment of the Mid-Continent snow goose population. . The chairmanship of the Central Flyway Council and Technical Committee rotated to Colorado in October 1993 for a 1 year term. As chairman of the Technical Committee, Ringelman organized and co-chaired the 1st ever joint meeting between the Central Flyway Technical Committee and the Pacific Flyway Study Committee. The joint meeting, held in January 1993, featured topics of �I ' 40 mutual interest to both flyways with emphasis on the use of adaptive waterfowl harvest management, and other topics on which the 2 committees could develop a working relationship. In addition, the 2 committees met independently to discuss issues specific to the respective flyways. In March, Colorado hosted and chaired (Ringelman) the Central Flyway Technical Committee and Central Management Unit meeting while Szymczak attended the Pacific Flyway Study Committee meeting. March meetings allow committee members to exchange general information on migratory game bird populations and formulate regulatory recommendations for both Flyway Councils, for species hunted before Oct. 1, including doves, band-tailed pigeons, snipe, rails, and cranes. Early season special Canada goose seasons are also considered as well as early teal seasons. Of special interest at both meetings was a workshop and evaluation of a new program to train waterfowl in techniques to reduce crippling losses of waterfowl. For all meetings reports containing topics pertinent to Colorado written, compiled and distributed to CDOW personnel. In addition, project personnel were active in the newly formed Intermountain West Joint Venture of the North American Waterfowl Management Plan. Pooulation Survey Methodoloav Protocol for surveys for observing neck-collared Canada geese and maintaining and monitoring bufflehead boxes were presented to volunteers. Surveys of nesting and brood rearing Canada geese on Walden Reservoir in North Park during spring 1993 found the number of nests decreased, hatching success increased, but overall recruitment below the 1992 level. Detailed tables are submitted to CDOW Northeast Regional and BLM personnel annually. Project personnel conducted Hunter Performance Surveys durin9 the 1993 September teal season in an effort to evaluate the effect of allowing teal hunting one-half hour before sunrise on non-target species. Cooperative Canada Goose Banding For the second consecutive year, Canada geese were banded in the MancosCortez area in southwest Colorado in cooperation with personnel of the CDOW Southwest Region. A total of 176·gos1ings and 45 adults were banded at 5 locations in late June 1993 bring the total number banded in 2 years to 389 (Appendix A). Adaptive Harvest Management and North American Waterfowl Plan Evaluation Team Ringelman attended 1 meeting of the Continental Evaluation Team for the North American Waterfowl Management Plan, and 2 meetings as a Central Flyway representative on the Adaptive Harvest Management Working Group. Each . assignment requires extensive analysis of information both during and between meetings. As Central Flyway representative, Ringelman must prepare reports and/or presentations to the Central Flyway Technical Committee and Council. �47 Band-tailed Piqeons As a member of the Pacific Flyway ad hoc committee of the Western Management Unit, investigations into determining the current status of the Four-Corners band-tailed pigeon population were initiated by Szymczak in cooperation with Howard Funk (CDOW). In Colorado, surveys requesting bandtailed pigeon sighting information were distributed in spring 1993 to field personnel within the band-tail population range; required hunting permits and wing envelopes were sent to region and area CDOW offices for distribution to 1993 band-tailed pigeon hunters; and a harvest survey was designed and sent to pigeon permit holders. All data collected was compiled and analyzed and presented in the report in Appendix B. Beginning in March 1993, Szymczak assumed the chairmanship of the Four-Corners band-tailed pigeon sub-committee of the Western Management Unit of the Western Migratory Upland Game Bird Committee. Duck Breeding Pairs - Partners for Wildlife Areas In May 1993 duck breeding pair counts were conducted from the ground and the air on 13 Partners for Wildlife wetland development areas in the San Luis Valley. These counts were designed to evaluate methods of enumerating breeding pairs on these small, intensively managed private land areas that are leased.by the U. S. Fish and Wildlife Service. It is assumed that special surveys of these areas will be needed to evaluate the effectiveness of the program and provide reliable duck breeding pair numbers from these areas for the valley-wide breeding pair survey. Two or 3 counts were conducted on each wetland complex from the ground and 2 counts were made from the air. Only 1 ground count was made per day on each wetland. Aerial counts were conducted on consecutive days. Initial analysis indicates there was little variation between ground counts in the number of breeding pairs observed. Aerial count totals were similar to ground counts in areas on which there were few breeding birds. However, aerial counts recorded a smaller percent of the birds observed from the ground, as the total number of birds on the area increased. These comparative counts are scheduled to be conducted for 2 mare years. DISCUSSION Project personnel provide useful information in planning and evaluating waterfowl management and habitat enhancement programs in Colorado and educating land management agency personnel about the habitat requirements of waterfowl. We expect that with increased emphasis on habitat enhancement in Colorado as outlined in the statewide Waterfowl Management Plan, our services will be more in demand. Additional activity on the CWSC and the WHPRC will help insure that the money raised through the Colorado Duck Stamp program is spent in accordance with the objectives of the program. Committee and the. Conducting and/or formulating surveys and banding efforts and informing management agency personnel about various aspects of waterfowl and wetland ecology provides a valuable service to management agencies, the waterfowl resource and in some cases the hunting public. Continued participation on Flyway and National committees ensures that Colorado will remain informed on migratory bird matters, have input in �48 migratory bird hunting regulations, and have influence on habitat programs affecting migratory game birds. LITERATURE CITED Braun, C. E. 1976. Methods for locating, trapping and banding band-tailed pigeons in Colorado. Colo. Div. Wildl., Spec. Rep. 39. 20pp. Prepared by: ~J?£ ~L Michael R. Szymczak Researcher/Scientist III �-., /. Q APPENDIX A Table l. Numbers of Canada geese trapped and banded, by location in the Cortez - Mancos area 1992-93. Lac. M Lac. F Location 92 93 92 Thomas's Pd. 12 8 17 Verde' Pds. 1 Dolores' Hat. 6 Baikie Pds. Ad. F 93 92 92 93 92 93 14 8 5 4 42 28 14 11 93 2 0 7 6 Total Ad. M 13 14 2 9 1 1 3 3 4 2 22 23 34 Colbert's Pd. 7 35 7 30 4 8 5 9 23 82 Duddleson Pds. 9 34 9 26 4 3 6 8 71 28 McPhee Res. 1 Browning Pd. Total 3 5 50 88 5 68 88 2 23 10 3 3 16 27 5 17 29 168 221 �50 APPENDIX B BAND-TAILED PIGEON REPORT COLORADO 1993 Mike Szymczak and Howard Funk Colorado Division of Wildlife INTRODUCTION From 1967 through 1972, research on the four corners band-tailed pigeon population, coordinated through the Four Corners Cooperative Band-tailed Pigeon Technical Committee (FCPC), was conducted in the states of Colorado, Utah, Arizona and New Mexico. Information on distribution, habitats, migration patterns and chronology, survival, hunting pressure, harvest, hunter success, crippling loss and age composition of the harvest was collected (Braun et ale 1975). Since experimental work was terminated in 1973, hunting seasons have continued, but little information on population status has been collected. Recently, concern for Four Corners bandtails has been voiced by biologists and other field personnel in the respective conservation agencies and the U. S. Fish and Wildlife Service (USFWS). This concern has prompted the 5 management agencies to form an adhoc committee, through the Central and Pacific Flyway councils, to examine the current status of the Four Corners pigeons and possibly construct a management plan for that population. The following are the results obtained from field observations, harvest questionnaires, and wing collection surveys, in Colorado in 1993. METHODS Field Observations Band-tailed pigeon (Columba fasciata) observation packets were distributed to District Wildlife Managers, whose districts included pigeon feeding areas and trap sites in Colorado, described in Braun (1976, Appendix). They were asked to visit the sites when 1n the vicinity, and note pigeon activity. They were also asked to record pigeon numbers observed at any other locations. In addition, some of these sites were visited by Szymczak, and some unsolicited observations were received. As part of the harvest questionnaire, hunters were asked to record the number of band-tailed pigeons they observed while hunting. If hunters recorded a range in numbers seen, the highest figure was recorded. Small grain production Many of the band-tailed pigeon feeding and trapping sites described by Braun (1976) were associated with grain fields. Responses on many of the questionnaires sent to field personnel indicated that the described feeding �51 sites were no longer present and grain production was reduced in the general area. Therefore, planting statistics obtained from the annual report, Colorado Agricultural Statistics, were compiled for 1970-74 and 1987-91, for Fremont and Huerfano counties in south-central Colorado, Archuleta and La Plata counties in southwest Colorado, Pitkin and Eagle in central Colorado, and Garfield and Rio Blanco in northwest Colorado to obtain a quantitative value of grain reduction. Revised estimates which were published 2 years following the actual planting year were used. Crops include winter and spring wheat, barley, oats, corn for grain and sorghum for grain. Harvest Survey All band-tailed pigeon hunters were required to obtain a free permit from the Colorado Division of Wildlife (CDOW). Permits were issued through personal contact or by mail from the CDOW offices in Grand Junction, Montrose, Glenwood Springs, Durango, Monte Vista, Salida, Fort Collins, Denver and Pueblo. Each permit contained a hunting diary on which the hunter could record hunting and harvest information. On October 12, 1993, following the Sept. 1 - Sept. 30 hunting season, a questionnaire (Appendix) was sent to permit holders. Nine were returned as undeliverable. A second questionnaire was sent on October 27, to those permit holders that had not responded. The final response used in analysis was received in early January, 1994. Responses were compiled and analyzed initially based on whether the hunter responded to the first or second questionnaire. Since permits were required to hunt band-tails on a statewide basis, we anticipated that some hunters of rock doves (Columba livia).would obtain a permit and report their hunting activity. If a hunter reported hunting pigeons in areas where band-tailed pigeons did not exist or report~d as seeing their quarry in non-bandtail - habitat, such as under bridges br on billboards, their responses were not used. Non-respondents were assumed to include rock dove hunters also. The number of permittees projected as hunting band-tails were adjusted according to the number of rock dove hunters in the sample. Because the questionnaire was not structured to compile data according to county, if the respondent reported hunting in more than one county, the first county mentioned was the county of record. We solicited comments from pigeon hunters. Those comments relating to comparisons of birds seen in 1993 compared to 1992 or earlier seasons, weather related factors, movements of birds, hun~ing season dates, etc. were tallied by category and are presented in narrative form. Wing Survey Each permit holder was issued a wing packet containing 10 pre-addressed envelopes. The hunter was instructed to remove one wing from each band-tailed pigeon harvested, enclose all wings from pigeons harvested in anyone day in the envelopes provided, and mail the envelope. Wings were received at the Coleman National Fish Hatchery, Red Bluff, California. On February 22, 1994, wings were examined by Roy Tomlinson, Western Dove/Pigeon Coordinator, U. S. Fish and Wildlife Service, and Clait Braun, Wildlife Research Leader, Colorado �Division of Wildlife to document the bird's age at harvest through examination of wing feather molt (White and Braun 1978). RESULTS Field Observations Reports of presence or absence of pigeons was reported for 42 of 86 (48.8%) sites/areas listed by Braun (1976) (Table 1). Some of the other sites may have been checked, and because there were no pigeons, no report was filed. Pigeons were found at or near only 17 of 42 (40.5%) of the sites that were checked. Pigeons were observed at 14 new sites in 1993 and 1 new site was reported from 1992 observations. Observations were distributed widely throughout the pigeon range (Fig. 1). Sixty-one of the 67 questionnaire respondents that reported they went hunting indicated the number of band-tailed pigeons they observed. Forty-seven (77.0%) of the 61 hunters saw pigeons ranging in numbers from 1 to 600 (Table 2). Pigeons were not seen in 7 of the 19 counties in which there was reported hunting. Only 5 of 52 respondents (9.6%) hunting in counties where pigeons were observed, reported not seeing pigeons. Substantial sized flocks were seen in Mesa (500), Montrose (150), and La Plata (600) counties. Small Grain Production Substantial acreages of small grain were planted in all sample counties in 1970-74 except Eagle-Pitkin. Winter wheat, barley and oats were the most common grains planted, but there were very few fields planted to corn or sorghum for grain. Most Grain acreage was declining during the 1970-74 period (Fig 2) and by 1987 had declined to levels that were about one-half of 1970 levels. Declines were recorded in all counties, but were most dramatic in south central Colorado. In Huerfano County, where 2,204 pigeons were banded in or near grain fields near La Veta from 1969 - 75 tBraun 1976), no grain was reported planted in 1989, 1990 or 1991. Harvest Survey A total of 185 permits were issued through the following offices: Grand Junction (13), Montrose (43), Glenwood Springs (25), Durango (15), Monte Vista (11), Salida (9), Fort Collins (15), Denver (32), Colorado Springs (13) and Pueblo (9). Of the 185 questionnaires sent, 9 were undeliverable, 92 were returned following the initial mailing and 46 were returned following the second mailing for an overall response rate of 78.4 percent. Thirteen respondents (9.4%) were considered rock dove hunters. There was no measurable difference between respondents to the first or second mailing in terms of the ratio that hunted (V2 = 1.87, P > 0.10), the ratio of hunters who were successful (v = 0.445, P > 0.25), the mean number of days hunted (t = 0.303, P > 0.70) nor the number of pigeons harvested/successful hunter (t = 0.091, P > 0.90) (Table 3). Therefore, all usable questionnaires were pooled. The projected number of permittees that were considered band-tailed pigeon hunters was 168. 2 Only a projected 82 hunters, about 50% of the permit holders, reported hunting band-tailed pigeons in 1993 (Table 3). They hunted an average of 2.4 days during the season, harvesting 80 birds and crippling an additional 10. �53 Only 13 hunters reported harvesting band~tailed pigeons (Table 4). Pigeon harvest was confined to 7 counties but was distributed throughout the season (Fig 3). Fifty-three permittees that hunted band-tailed pigeons offered thoughts or impressions of the 1993 season. Some mentioned only one item whereas others covered several subjects. Comments came from Montrose County hunters (19), followed by LaPlata (10), Mesa (5), and Delta (3). Other counties had comments from only one or two hunters. Fourteen hunters mentioned seeing fewer pigeons in 1993 compared to previous years and 3 thought the pigeon population was decreasing. Nine hunters specifically mentioned cold, rainy, and windy weather in late August to very early September, 1993 as responsible for decreasing the number of birds seen. Nine hunters mentioned birds leaving prior to September 1 or just after, with four hunters stating they had seen birds in their hunting areas prior to September 1. Most weather-related comments came from Montrose, LaPlata, and Mesa County hunters, but hunters from high elevation counties also mentioned weather. Eight of 10 hunters suggesting that the season begin prior to September 1 hunted in Montrose, LaPlata, and Mesa counties. Yet, 8 out of 37 hunters from Montrose, LaPlata, Mesa, and Delta counties mentioned seeing enough pigeons for good hunting. Some said they saw more birds than ever in the hunting season. Two other hunters from two different counties mentioned seeing good numbers to hunt, while twelve hunters from eleven different counti~s simply stated they didn't find any birds. Extension of the season beyond September 30 was mentioned by three hunters. Wing Survey Only 31 w.ings were co llected from Colorado hunters through the Parts Collection Survey. Nearly one-third (29%) of the wings collected were classified as ~oming from immature bi~ds while 71% were adult wings. DISCUSSION Field Observations The response. from field personnel was not adequate to measure present bandtailed pigeon distribution. Many respondents indicated that they had not seen band-tails for quite a few years, but some said they had not been looking specifically for pigeons either. Many of the trap sites and/or areas where pigeons were observed in the 1970's were associated with small grain fields or corrals where livestock were fed (Braun 1976). Comments on returned observation forms indicated that many of the grainfields had been converted to pasture or alfalfa fields. A preliminary examination of agricultural statistics confirms the reduction in grainfields. Pigeons feeding on native foods are difficult to locate (Braun 1976). The decline in small grain production generally throughout the pigeon range has most likely resulted in greater dependence of pigeons on native foods. If pigeons were concentrated on the remaining grain fields, their occurrence at these sites would probably be noted by field personnel. Hunters reported seeing more band-tailed pigeons than did CDOW field �personnel. A substantial number of pigeons were observed by hunters in some counties, such as Mesa and La Plata, in which very few pigeons were reported by field personnel. Many hunters mentioned that pigeons, responding to cold weather, had left hunting areas. However, large numbers of pigeons were reported seen by hunters in Mesa County, September 18 - 19, 1993 and in La Plata County, September 27 - 30, 1993. Hunters and Harvest According to the number of permits issued, the number of band-tailed pigeon hunters were below historical numbers. During 1970-74 and 1975-80, when special permits were required for hunting, the number of active pigeon hunters averaged 265 (Table 5) and 226 (H. Funk, unpubl. data), respectively. Using comparative results from the Colorado Small Game Harvest Survey for 1975-80, Funk (unpubl.) found that the Small Game Survey over estimated the number of pigeon hunters by an average of 43.8%. Using the calculated correction, Funk estimated an average of 337 active hunters during 1981-90, using results of the Small Game Survey. Confidence intervals on Small Game Survey estimates of the number of pigeon hunters state-wide are normally in the ± 70% range. However, pigeon hunter numbers did not show a downward trend during the 198190 period. Therefore, either the hunter decline has been most dramatic in recent years, or a number of pigeon hunters did not obtain a permit in 1993. The questionnaire response rate was lower in 1993 than during the 1970-74 period, but the percent of permit holders that hunted was similar (Table 5). The percent of successful hunters differed considerably between the comparative seasons (V2 = 9.48, P > 0.01), but the number of days· hunted/hunter was similar. The estimated reported bag of 80 pigeons in 1993 was about 10% of mean estimated harvest during 1970-74 and about 5% of 1981-90 corrected estimate using Small Game Survey results. The number of pigeons harvested/successful hunter was similar in 1993 to the earlier period, but the range in pigeons shot in 1993 was variable (Table 4), as 1 hunter accounted for nearly 29% of the pigeons reported harvested. c The age composition of the pigeons harvested in 1993 was similar to what was found during the late 1960's early 1970's (V2 = 0.059, P < 0.75) The percent of immature pigeons in the harvest in Colorado (29.0%) was slightly less than was found during 1970-72 period (32.8%; Braun et al. 1975). In the Four Corners States combined, during the 1968-72 period, 24.8% of the pigeons harvested were classified as immatures compared to 26.1% in 1993. Sample sizes in 1993 were small compared to the earlier period (Colorado = 674 wings/yr.; population = 2,225 wings/yr.). x x CONCLUSION The distribution of feeding band-tailed pigeons has changed considerably since initial studies in the late 1960's, early 1970's. Some grain fields that were present during the initial period have now been converted to other uses. If birds are still present in the vicinity of some previous feeding sites they are using alternative, most likely natural food sources. Birds feeding on natural food supplies are probably wide-spread, difficult to find and count, �and may change foraging areas annually depending on the presence of natural food supplies. The impact of changing foraging strategies on breeding distribution, nest success and population size is not known. Band-tailed pigeon hunter numbers have declined considerably since the initial study period. A portion of this decline is probably the result of the reduced concentrations of pigeons during the hunting season. Additionally, hunter numbers for all small game species have been declining in Colorado. Based on hunter activity and pigeon harvest during the 1993 hunting season, harvest currently is having a minimal impact on populations. RECOMMENDATIONS 1. Continue the required permit for band-tail pigeon hunting, the wing envelopes and the hunter questionnaire. Since pigeon hunters are a source for population distribution information, restructure the questionnaire to obtain more precise information on number of pigeons observed. 2. Obtain the current status of all feeding/trap sites listed in Braun (1976) through personnel reconnaissance or contacting those field personnel that failed to respond to the pigeon observation questionnaire. 3. Continue cooperative work with conservation agencies of the 4-corners states, Utah, New Mexico and Arizona, and the U. S. Fish and Wildlife Service to develop strategies for monitoring 4-corners band-tailed pigeon population status. LITERATURE CITED Braun, C. E. 1976. Methods for locating, trapping and banding band-tailed pigeons in Colorado. Colo. Div. Wildl. Spec. Rpt. 39. 20pp. ____ , D. E. Brown, J. C. Pederson, and T. P. Zapatka. 1975. Results ofthe Four Corners cooperative band-tailed pigeon investigation. U.S. Dep. Inter., Fish and Wildl. Servo Resour. Publ. 126. 20pp. White, J. A. and C. E. Braun. 1978. Age and sex determination of juvenile band-tailed pigeons. J. Wildl. Manage. 42:564-569.2 �56 Table 1. Results of questionnaire distributed to Colorado Division of Wildlife field personnel requesting bandtailed pigeon observations Spring-FaU1993 at feeding or trap sites described by Braun (1976). Some notes are observations by the authors. Observations County Location Report Archuleta Arboles yes No Pigeons Observed (NPO) No pigeons observed for 6-7 years Pagosa Sprgs. Piedra yes no NPO No grainfields (9/6) No grainfields (9/6) Boulder Niwot no Chaffee Missouri Park Salida yes yes Clear Creek Upper Bear Cr. Pine Valley no no Conejos Antonito Area yes Romeo Area yes Costilla Ft. Garland Area no Custer Westcliffe Area Wetmore Area yes yes Austin Area Delta Area Hotchkiss Area Paonia Area no yes no no General lone Cone SWA yes new site Douglas larkspur Area Louviers Area Perry Park Area Sedalia Area no no no no Eagle Avon Area Beaver Cr. Ski Area yes new site Eagle Area El Jebel Vail Deep Creek yes no new site new site Delta Dolores Date Time Aug 20 Sep 26 Sep 29 Jun 30 Jul03 Jul27 Aug 17 Sep 26 Sep 28. 1800 1630 900 1100 1430 800 1200 930 No. Notes 20 6 At bird feeders, 5m W Poncho Springs Traditional pigeon hunters stopped hunting 10 8 7 6 15 8 6 12 Roosting Feeding Feeding Feeding Feeding Feeding Feeding Feeding Barley fields decline, conv. to alfalfa (9/6) Forty sighted E of Fort Garland (4/25) NPO Sep 12 1300 15 NPO Sep 30 1200 20 Jun 20 600 No grainfields (5/27) Babcock Hole No grainfields observed along CO 96 Greenwood to Wetmore (5/27) NPO Initial feed sites being developed At bird feeder, 2Q..3()sighted in previous years on seeded ski area; previous yeats observations along 1-70 near Minturn turnoff. NPO Jut Sep 15 4 8 Aying over town Water at Colo. R. and Deep Cr. �) Table 1. Continued. Observations County Location EI Paso Air Force Acad. Chipita Park Area Manitou Spg.Area Woodmen Valley Ar. Monument Fremont Garfield . Canon City Florence Cotopaxi Area Hillside-Texas Cr. Lower Hardscrabble Upper Bear Cr. Howard Area Report no yes no no new site Jun 11Sep 28 NPO yes yes yes yes new site NPO NPO NPO NPO Aug 10 Aug 13-17 no yes no Grand Williams Fork Res. no Huertano Farsita-Red Wing La Veta Area yes yes Jefferson Aspen Park Bergen Park Conifer Area Deer Creek Area Evergreen Area Homewood Pk. Area Indian Hills Area Lookout Mtn. Area Marshdale Area no no no no no no no no no La Plata Bayfield Durango Area Bayfield South Hesperus yes no yes no Larimer Estes Park Area Glen Haven Las Animas Aguilar Area StoneMrcU~Weston Area Trinidad Area San Miguel Cr. Whiskey Cr. Collbran Area Mesa Area Molina Area Unaweep Canyon Lands End ·no no no no new site TIme No. Notes NPO yes Carbondale Area Glenwood Spgs Rifle Mesa Date All day 1-70 1-70 Bird feeder; Mean numbers June (10.3); July (10.1); Aug (24.3); Sept (16.3) Pigeons decreased in last 5 years 12-15 25-30 NPO Fewer grainfields NPO NPO Only one grainfiled-near Gardner (5/27) No grainfields NPO No grainfields N of Bayfield Aug 20 1000 40 Feeding in oak brush No grainfields (6/16) yes yes Jul 12 NPO 1500 10 Near Prospect Mountain yes yes NPO Aug 25 1600 10 No grainfields (9/5) Number of pigeons down yes new site new site NPO 900 16 15 Sep04 1 I �~~ .J'-- Table 1. Continued. Observations County Location Montezuma Dolores Area Lewis Area Mancos Area McElmo Canyon Stoner Area yes yes no no new site NPO NPO Montrose Area yes Jul08 Aug 15 Sanborn Park Montrose Gun Range no new site Robideau Tr. new site Jun Sep Sep Jun Ouray Colona Area yes NPO Park Estabrook no Pitkin Aspen Area Capitol Cr. Crystal R. Dinkle L Sopris Cr. Woody Cr. yes yes no no yes Apache City Beulah Area yes yes Colorado City Rye Area yes yes _Buford Area yes Uttte Beaver Area Coal Creek yes Del Norte Area yes Montrose Pueblo Rio Blanco Rio Grande Report new site South Fork new site Monte Vista Area no Saguache La Garita Cr. new site San Miguel Norwood Area no Teller Woodland no Park Date Time Notes No. No grainfields (Jun 16) Olterman's bird feeder 07 03 09 28 Sep OS NPO 700 830 200 400 1545 900 1445 1030 4 36 4 3 Flying toward Montrose Roosting Flying over oak brush 1030 4 Flying 800 900 1000 1000 930 720 50 75 50 8 15 20 In oak brush 3R Road North Creek Road Twelve mile cut off Greenhorn Cr., roosting in cottonwoods At bird feeder Old San Isabel Rd-5t.Charies Ridge Rd. NPO NPO Aug 21 Sep 04 Sep 11 Sep OS May 25 Sep 04 Jun 05 Jun 21 Jun 28 Jul07 Aug 13 Sep 03 Apr 28 Jun 18 May 27 Jun 26 Jun 31 900_ 1200 1000 1300 1100 1000 630. 530 1300 1800 33 Oct 10 Oct 12 SepOS Sep 10 5 12 6 8 4 2 22 12 12 -150 6 2S 1000 900 78 82 Marvin Cr. - Adams Lodge Marvin Cr. - Fritzlers South Fork Cabin Marvin Cr. - Adams Lodge South Fork White R. - White Ranch South Fork White R. - White Ranch At grouse lek site Grain in area decreasing On dove route Dunham Ranch Dunham Ranch Six 1/2 miles W of Del Norte Davies site on Co 1ro, feeding both sides of hwy; observations at same site in 1992 Dakota Pk. development. Nov 10-27, 34 sighted in 1992. Areas S. of Monte VIStavisited on 5/24-26 with no obser. Farmers 5a!f no pigeons for quite a few years. MVNWR pars. say no pigeons on 8 Mile Road in recerityrs. Feeding in alfalfa LaGarita Ranch �59 Table 2. Number of band-talled pigeons observed, by county, by hunters reporting observations during the 1993 hunting season. Area County No. of hunters Reporting Obs. Pigeons Number Pigeons obs. x Range STD 102.3 10-500 182.6 28.5 22.7 7-50 1-150 30.4 35.7 99.6 12-600 189.4 5.5 3.5 3-8 2-5 3.5 2.1 48.0 1-600 112.4 Northwest Moffat Garfield Mesa Pitkin 3 2 8 North-central Jackson Jefferson Douglas Clear Creek 0 7 0 0 30 716 0 1 0 0 24 10 0 0 2 19 0 9 57 432 0 896 1 2 2 Southwest Delta Montrose Archuleta LaPlata South-central Chaffee Teller EI Paso Fremont Las Animas Costilla Rio Grande 3 20 9 1 2 2 2 1 1 1 0 20 11 7 3 50 0 0 0 61 47 2256 �60 Table 3. Hunting pressure and harvest statistics, 1993 band-tailed pigeon season. Percent of total in parenth_. standard deviations in brackets. Response Projected for all First Second Category Mailing Mailing Total Number of permittees 92(52.3) 46(26.1) 138(78.4) Number of usable returns 85(68.0) 40(32.0) 125 Number of respondents hunting 42(49.4) 25(62.5) 67(53.6) 82(48.6) Number of respondents not hunting 43(50.6) 15(37.5) ~8(46.4) 71 Number of succesaful 10(23.6) 3(12.0) 13(19.4) 16 Number of hunter days 102 58 160 195 Days hunted per hunter 2.41 [1.6) 2.3[1.1J 2.4[1.5) 2.4 Number of pigeons bagged 47 19 66 80 Pigeons p••. suec_ 4.7[5.4) 6.3[4.2] 5.1[5.0] 5.1 Pigeons per hunter 1.1[3.3) 0.8[2.4] 1.0[3.0] 1.0 Number of pigeons 6 2 8 10 0.14(0.60J 0.08(0.41J 0.12[0.53] 0.12 53 21 74 90 11.3 9.5 10.8 11.1 permittees responding hunters ful hunter crippled and lost Pigeons crippled and lost per hunt••. Total harv_t (bagged crippled and lost) Percent crippling lou �61 Table 4. Band-talled pigeon harvest by county, 1993 hunting season. Number of Pigeons Bagl Number of Area County Northwest Garfield Mesa Crippled Successful Hunter Successful Hunters Bagged 2 0 2.00 3 12 0 4.00 7 0 7.00 0 2.50 North-central Jackson Southwest Delta 2 5 Montrose 3 15 2 5.00 19 5 19.00 LaPlata South-cantral * Las Animas State wide average 2 6 13 66 3.00 6 5.04- �0' Table 5. Band-t.llect I" pigeon harvea. aurvey reaulta. 1870-74.1893. Valuea In parentheaea are percentagea. Vears Number of permltteea 1970 1971 384 524 1973 1974 x 1993 ,·562 455 518 485 185 1972 Number of permittees r•• pondlng 338 477 520 419 465 443(91.3) 138(74.6) Number of permltteea hunting 182 344 298 212 288 265(54.6) 82(53.8) Number of permittees not hunting 182 180 264 243 230 220 71 77 201 124 80 132 119 18 374 851 762 841 776 681 195 2.1 2.5 2.6 3 2.7 2.6 2.4 458 1537 729 329 734 757 80 Number of succ •••• ul hunters Number o' hunter days Day. per hunter Numb.r 0' pigeon. bagged Pigeons per .ucce ••• ul hunter 5.9 7.8 5.9 5.5 5.6 8.4 5.1 Plgeona par hunter 2.5 4.5 2.4 1.8 2.5 2.9 1 Number 0' crippled and loat 83 188 93 37 58 91.4 10 Pigeon. crippled and 10lt per hunter 0.5 0.5 0.3 0.2 0.2 0.35 0.12 541 1723 822 366 792 849 90 11.3 10.1 10.5 11.1 pigeon. Total harve.t (bagged Percent crippling 10•• + crippled and l08t) 15.3 108 7.6 �,' 0,,) !~~ POSITIVE L ..... :::.:-:.:.:.\ .-. NEGATIVE NO REPORT NO SITES o < 1 0 PIG. • 1 0 - SOP CJ > SOP I G. I G. FIG 1. SIGHTINGS OF BAND-TAILEDPIGEONS BY CDOW FIELD PERSONNEL, BY COUNTY, 1993. SITES EXAMINED WERE PRIMARILY FROM BRAUN 1976. �64 30,000 ,....---------------------, ,, ARCHULETA - LA PLATA ------- GARFIELD - RIO BLANCO - - - - FREMONT - HUERFANO , , , o w t- , 20,000 Z ::sc. sa EAGLE - PITKIN . ,. , ...._, ._ - ,_ , _, ..",. 15,000 _.. \ ~""" /, ' / '/ " /,' , , ,, ,",, , '" ..... ... ..... _-_ ...•. '" -_ .... \ W c: ~ o «: '\ \ \\ 10,000 '.\ "" '- ...••. ..... -_ '---------~~ --.. -,' - -- " 5,000 ............................... ........... o ~-~-~-~~---~-~-~-~~~ 1970 1972 1971 1974 1987 1973 . 1989 1991 1988 1990 YEARS FIG 2. ACRES PLANTED TO SMALL GRAIN IN SELECTED COLORADO COUNTIES. �65 30 ~ I 25 L ~ Cf) ill > a:: 20 -c J: LL a 15 ~ Z W o a:: w a, 10 5 o SEPTEMBER DATE FIG. 3. BAND-TAILED PIGEON HARVEST BY TIME PERIOD, 1993 �STATE OF COLORADO Roy Romer, Governor DEPARTME..1\fTOF NATURAL RESOlJRCES DIVISION AH EQU"," OPPORTUNITY OF WILDLIFE I3\oIPlOYER Perry D. Olson, Director 6060 Broadway Denver, Colorado 80216 Telephone: (303) 297-1192 l\UGRATORY GAME BIRD UNIT 317 West Prospect Fort Collins, Colorado 80526· (303) 484-2836 October 12, 1993 For Wi1dJif~ For pf!Ople X362 Dear Band-tailed Pigeon Hunter: In order to evaluate Colorado's band-tailed pigeon season, it is necessary for us to obtain specific information regarding hunter effort and success. Please take time now to supply the information requested below as accurately as possible and return this questionnaire in the enclosed self-addressed, postage-paid envelope. If you did not hunt, please note this by answering the first question and returning the questionnaire because this information is .important as well. If you hunted and recorded results on the diary on your permit, you can utilize this information for the questionnaire. Otherwise, please try to remember results of your hunting effort as accurately as possible. Please record only v6ur activities and results; not those of others in your hunting party. Your cooperation will be appreciated. BAND-TAILED PIGEON HUNTER QUESTIONNAIRE # --- Yes 1. Did you hunt band-tailed pigeons in the 1993 hunting season? 2. What county did you hunt in most? _ Second? No -------- 3. How many days did you hunt pigeons this season? _ 4. How many pigeons did you bag during the entire season? _ 5. How many pigeons did you knock down within sight but could not find and/or retrieve? ------------- 6. Please indicate the number of pigeons you bagged during each September period below that you hunted. Place a "0" in the box for periods you hunted but did not harvest birds. Leave blank those periods which you did not hunt. 1-3 4-5 l 7. Approximately 8. 6-10 11-12 13-17 18-19 CJ CJ LJ 20-24 25-26 27-30 CJ .__( __, how many pigeons did you observe during your hunt(s)? In what type(s) of habitat or area did you observe most pigeons? _ _ 9. Please give your thoughts on and/or impressions of the band-tailed pigeon hunting season. -------------------------- ICSIOI DEPARTMENT OF NATURAL RESOURCES. Kczmctb Salazar. Executive Dim:tor WILDLIFE COMMISSION, Thomu M. Eve:, Claimwt. Louis F. Swilt, Vice Chainnan. Arnold Salazar. Secretary L Boyd. Ir •• Member· Edon W. Cooper. Member· Rebecu M. Frank. Member. William R. Hegbcrr. Member. Marie LeValley, Member �67 Colorado Division of Wildlife Wildlife Research Report ·October 1994 JOB PROGRESS REPORT State of _-'C"""o:;._:l_::::o_,_r=ad=o"-_ Project Work Plan 22 Job Title: : Job _2_ Migratory Game Bird Publications Period Covered: Author: Migratory Game Birds Investigations W-166-R-3 01 April 1993 through 31 March 1994 Michael R. Szymczak Personnel: J~mes K. Ringelman and Michael R. Szymczak, Colorado Division of Wndl ife ABSTRACT The following list contains those articles that were prepared and/or submitted for publication or published during this segment: Ball, I. J., T. E. Martin, and J. K. Ringelman. 1994. Conservation of nongame birds and waterfowl: conflict or compliment? Trans. N. Am. Wildl. Nat. Resour. Conf. 59 (In press). Gilbert, D. W., D. R. Anderson, J. K. Ringelman, and M. R ..Szymczak. Response of nesting ducks to habitat management on the Monte Vista National Wildlife Refuge. Wildlife Monograph (In review) Jeske, C. W., M. R. Szymczak, D. R. Anderson, J. K. Ringelman, and J. A. Armstrong. Mortality factors, body condition and survival rates of Wintering mallards in the San Luis Valley, Colorado. J. Wildl. Manage. 58 (In press). Prepared by: 2J1,;,j_,.. j Jf~<:r- M.ichae·lR. Szymczak Reasercher/Scientist III i .I. ~.-~ ��69 Colorado Division of Wildlife Wildlife Research Report september 9, 1994 JOB PROGRESS State of Project: Work Colorado W-164-R-1 Plan __ -=l__ ~ Job Title: Period Laboratory Job 1 July, W.J. Adrian, Investigations l MONOCLONAL ANTIBODIES MULE DEER AND ELK Covered: Personnel: REPORT 1993 TO DISTINGUISH 30 June, R.P. Ellis, AMONG WHITE-TAILED DEER, 1994. and R.J. Todd. ABSTRACT Research on the production and use of monoclonal antibodies (MAbs) directed against albumin for direct and positive identification of the species of origin of blood, blood stains and meat was continued this fiscal year. We have obtained purified elk, mule deer and antelope albumins for use as antigen in our MAb studies. We have completed three fusions with mule deer albumin, one fusion with elk albumin and two fusions with antelope albumin. ; ------ ��71 MONOCLONAL ANTIBODIES TO DISTINGUISH AMONG MULE DEER AND ELK William 1. Produce species specific 2. Determine 3. Produce 4. Determine specificity deer and mule deer. 5. Initiate specificity species monoclonal monoclonal antibodies METHODS against antibodies antibodies of the monoclonal use of this technology DEER, J. Adrian of the monoclonal specific WHITE-TAILED for elk and mule deer. against antibodies for ongoing elk and m~le deer. forensic white-tailed between deer. white-tailed cases. AND MATERIALS This work is a cooperative endeavor between the Division and Dr. Robert Ellis of Colorado State University. The technique for producing MAbs was developed by Kohler and Milstein (1975). Schulman el at. (1978) developed modifications which simplified the technique. It is a technique in use in many immunology laboratories throughout the world. The advantages of MAbs over polyconal antibodies (those produced by animals and harvested from serum of that animal) are (1) each individual hybridoma clone produces antibody molecules of a single specificity, (2) many hybridoma clones can be generated, each producing a particular antibody specificity directed against a particular antigenic epitope, and (3) hybridoma clones can be frozen in liquid nitrogen (LN2) and stored for future use, thus once the hybridoma clones are developed, they are available indefinitely. The method of visualization of MAb binding to albumin antigen epitopes is the Western Blot technique. This technique employs electrophoretic separation of the sample protein, followed by transfer of the electrophoresed sample onto a nitrocellulose membrane. The membrane is reacted with a MAb, rinsed, then reacted with an enzyme (usually horseradish peroxide) conjugated antibody. The presence of the bound peroxidase is detected by adding peroxide plus a chromophore. The development of a brown spot indicates MAb bound to its specific antigenic epitope. Failure of development of a brown spot indicates that the MAb did not bind to its specific epitope. Positive controls are run on the same electrophoresis gels as the sample to ensure accuracy. RESULTS AND DISCUSSION . Research on the production and use of monoclonal antibodies (Mabs) directed against albumin for direct and positive identification of the species of origin of blood, blood stains and meat was continued this fiscal year. We have obtained purified elk, mule deer and antelope albumins for use as antigen in our MAb studies. We have completed three fusions with mule deer albumin, one fusion with elk albumin and two fusions with antelope. The MAbs produced to date show a wide range of reactivity. Several clones have been generated which produce anti-mule deer albumin (MDA), anti-elk albumin (AE), anti-whitetailed deer serum (WTS), anti-moose serum (MS) and anti-antelope albumin (AA), which do not cross react with bovine serum albumin (BSA). A similar clone shows the same pattern of reaction except it is not producing anti-AA or antiBSA. �72 A clone from the third MDA fusion, which has been dilution cloned once produces antibody which reacts strongly with WTS, weakly with MDA< EA, AA and moose, but not with BSA. Another clone's antibody reacts with MDA, MS, WTS, and BSA, but not with AA or EA. One reacts strongly with MDA and WTS and weakly with AA, moose and BSA, but not EA. Several clones react with MS and WTS, but not AA, MDA, EA or BSA. These are currently being dilution-cloned and will undergo further screening. White tail deer albumin (WTA) was purified from serum as described previously. Purity was determined by SDS-PAGE on Phastgels (Pharmacia) and on mini-gels (Bio-Rad). The pure protein was mixed with RAS (Ribi Adjuvant System). Female 6 week old Balb/C mice received the. initial dose of 0.1 ml in 2 sites each subcutaneously. One mouse received 0.·2 ml of the adjuvant-antigen mixture that had been at room temperature for 3 days. The other mouse received 0.2 ml of RAS/WTA mixture made fresh 6124194. The SP2/0 myeloma cells were thawed from liquid nitrogen storage and checked daily for viability. On 6/29 and 6/30 the mice were tail bled and the serum tested by ELISA for antibody production. Both were producing approximately equal amounts of antibody. The final injection was l.V. to the tail vein one week later. The spleen of one mouse was fused to the SP2/0 myeloma cells using polyethylene glycol on 7/11/94. On 7/22 the fusion products were screened by ELISA against WTA, mule deer albumin (MDA), and elk albumin (EA). Those wells containing cells producing antibody reactive with WTA but little or no reactivity with the other albumins were transferred from the 96-well plates to 24-well, 2 ml plates. A fusion was performed on the second mouse spleen as described for the first mouse above 3 days after one additional boost of WTA (no adjuvant) intraperitoneally. These 2nd fusion products and the dilution clones of the 1 st fusion were discarded due to a fungal contamination. The original 1st fusion plates had been saved and these were fed, grown and dilution cloned. Thirteen of 201 single-clone wells were kept as a result of another ELISA screening. The 13 clones showed various reactivities with WTA, MDA, EA, moose albumin (MA) and antelope albumin (AA). The computer program (BioRad Microplate Manager) and printer gave the same symbol on the printed data for ~ the top cut-off value as for under the lower cutoff value. Thus some clones were .not as specific as originally tho·ught. However, these clones were all grown, frozen and retested for specificity against the albumins listed above, in. addition to bovine serum albumin (BSA). . These ELISA results indicate: 3 clones are no longer producing any antibody. 3 clones are producing anti-WTA weakly, anti-MDA (weaker) and anti-AA (very weak) but no anti-EA,MA,or BSA. 2 clones are producing anti-WTA weakly, anti-MDA (weaker) and no other antibody. 3 clones that are producing the most anti-WTA are producing what appears to be an equal amount of anti-EA and about 1/2 as much anti-MDA. No clones react with BSA. It appears that there are some shared antigenic epitopes among the albumins from different species, as shown by some cross-reactivity. The differences reactivity will be examined by additional dilution cloning and specificity testing. Prepared bY'~£ " William Wildlife C in �73 JOB PROGRESS REPORT State of Project: __ C",-,o=I'-"o;;:_r""ad;:;,;o"'-__ ...J(_:W_:_-_:1",,5..:::_O_:-R:.,.:_-_,,6o,L) __ : Peregrine Falcon Restoration Program Period C-overed: 1 July, 1993 - 30 June, 1994 Personnel: G.R. Craig, C-olorado Division of Wildlife and J.H. Enderson, The C-olorado College. ABSTRACT In the 1994 peregrine breeding season, 71 territories were occupied by 61 breeding pairs that fledged 105 young. Productivity averaged 1.58 young fledged for those pairs that were monitored. Contents of 5 nonviable eggs were collected and have been preserved for future analysis. Shell thicknesses of the whole eggs averaged 1% greater than the pre-DDT era average. Shell fragments collected at 9 sites averaged 10% thin. This Job Progress Report represents a preliminary analysis and is subject to change. For this reason, information presented herein MAY NOT BE PUBLISHED OR QUOTED without permission of the author. ��75 PEREGRINE FALCON RESTORATION PROGRAM Gerald R. Craig SEGMENT OBJECTIVES 1. Annually monitor the number of breeding pairs of peregrines and their reproductive success in Colorado. 2. Annually monitor organochlorine pesticide levels in wild breeding peregrines in Colorado. 3. Monitor breeding population turnover through band. recoveries, presence of color markers, and. telephotographic identification of individual breeding adults. 4. Augment poor wild production by placement of captive hatched wild young and captive produced young into occupied wild nests. 5. Release captive hatched and captive produced young at potential and vacant wild territories. 6. Monitor recruitment of reintroduced peregrines into the wild breeding population of Colorado. METHODS AND MATERTALS 1. Visit all known peregrine breeding territories throughout Colorado and observe them from a distance to establish the presence of breeding adults. Breeding pairs will be kept under surveillance to determine initiation of egg laying. Depending upon the individual female's reproductive history and eggshell condition (obtained through measurement of previous year's eggshell thicknesses) and availability of captive hatched young for release, breeding pairs either will monitored or manipulated as outlined in approach 4. Those pairs not designated to be manipulated willbe revisited periodically throughout the nesting season to document reproductive success. When a pair's behavior indicates that egg laying has occurred and incubation is underway, the eyrie will be visited to document the number of eggs produced. The eggs will be candled to ascertain viability and approximate age. All nonviable eggs will be collected for chemical analysis. A second visit will be made to determine productivity, band nestlings, and collect eggshell fragments and unhatched eggs for thickness measurement and analysis under 2a and 2b. 2a. Eggshell fragments encountered during eyrie visits described in approaches 1 and 4a willbe measured for index to thickness following standardized procedures. 2b. Whole, nonviable eggs which are encountered during eyrie visits will be collected, preserved and submitted to the appropriate Fish and Wildlife Service approved laboratory for pesticide analysis. Eggs collected from the wild in the course of Approaches 4a, 4b and 4c that are artificially at the Peregrine Fund's Boise, Idaho facility also will be submitted for shell thickness measurement and chemical analysis. 3. Peregrines present at breeding territories will be examined to determine the presence of bands or color markers. Band confirmation will be accomplished through observation from a distance with . telescopes and concealed remote controlled cameras. When banded falcons are encountered, every effort will be made to read band numbers without trapping or handling the birds. It is possible this can be accomplished in most situations with a Questar field model telescope (80-BOx). When band numbers cannot be discerned, attempts will be made to trap and examine the falcon at a time when c.apture will have least impact upon breeding activities. �76 4a. In accordance with an annual release plan developed and approved by the State, U.S. Fish and Wildlife Service, Bureau of Land Management, National Park Service, and the Forest Service, a predetermined number of wild breeding pairs will be manipulated to augment natural productivity. Pairs with a history of reduced clutch size, cracked eggs, or infertile or dead eggs will be candidates for fostering efforts. 4b. On occasion, it may be necessary to recycle several early breeding pairs in order to delay them until captive hatched young of the proper age are available for placement into wild sites. No later than 10 days after the last egg has been deposited, the eyrie will be visited and the entire clutch removed without replacement. Approximately 14 days after removal of the clutch, the pair will recycle, select another nest ledge, and deposit a second clutch of eggs. If the eggs are thin shelled, they may be replaced with plastic replicas and treated as outlined in approach 4a. This technique also works well to augment captive production with wild produced eggs. 4c. At times, pairs will select inferior eyrie ledges that may compromise nest success such as ledges that are too narrow to support a brood of large nestlings, the site may be vulnerable to predators, or it may be exposed to the elements. If the ledge cannot be mechanically improved, pairs can be relocated to other ledges through the recycling method described in approach 4b since they invariably relocate and select a new ledge when recycled. . 5. In accordance with an annual release plan developed and approved by the State, U.S. Fish and Wildlife Service, Bureau of Land Management, National Park Service, and the Forest Service, a predetermined number of captive produced falcons will be released at unoccupied or potential sites through the technique of hacking. This technique is employed at locations that do not have the benefit of protection or care from adults. Young falcons of about 35 days of age will be placed in a hack box on a suitable cliff ledge at the reintroduction site. They will be fed and cared for by attendants until they are flying and capable of fending for themselves. This technique assures that the young become familiar with their surroundings and hopefully will return to the site as adults and take up residency. Hacking requires constant attendance and observation to protect the vulnerable young and assure they have sufficient food while they are dependent upon the hack site. While the hack sites will be operated by the State, actual costs to operate the sites will be borne by the appropriate land administering agency (Forest Service, Bureau of Land Management, and National Park Service). 6. Confirmed breeding territories and selected potential breeding sites will be surveyed annually to document the presence of released falcons and ultimately determine the success of recovery efforts. RESULTS AND DISCUSSION Survey Effort In 1994, 4 teams comprised of 2 observers each were assigned particular regions of the state to monitor breeding activities and survey potential cliffs as time permitted. Three teams were assigned regions west of the Continental Divide and 1 team was located east of the Divide. Fifty-two occupied territories were monitored by the teams; 19 other occupied sites were observed by cooperators. A total of 1,308 team hours were expended. An average of 4.8 visits were made to each site during courtship, incubation and young rearing phases. Approximately :?.26hours of observation were expended per team in.the course of each visit. Given the capabilities and dedication of these teams, this can be considered the optimum level of performance necessary to monitor 66 occupied territories. Although 5 additional occupied sites were intermittently visited, their remoteness and lack of time prevented teams from confirming actual reproductive success. �77 Ten previously unoccupied breeding territories were also observed by the teams. Visits to these sites averaged 1.6 hours with an average expenditure of 4.4 hours of observation per visit. This level of effort was sufficient to determine presence of peregrines and their reproductive stage, but was not sufficient beyond that level. Two of these previously unoccupied sites were reoccupied and added to the schedule of regular surveillance described above. As schedules permitted, the teams surveyed potential nest cliffs within their region of responsibility to locate additional pairs. An additional 314 hours were devoted to survey of 65 potential breeding cliffs and approximately 4.4 hours were expended at each site by the teams. This effort resulted in documentation of 4 additional territories. Territory Occupancy Breeding territory occupancy increased from 61 in 1993 to 71 in 1994 (Table 1). Five sites ( 8, 14,23, 76 and 79) were reoccupied and 4 previously undocumented sites( 87, 88, 89 and 90) were confirmed. When site 23 was reported occupied in 1989, the pair actually occupied an adjacent canyon that was thought to be an alternate to territory 23. However, both sites were occupied in 1994 causing them to be separated into sites 23 and 86. Only 1 vacancy occurred (site 77). The rate of occupancy was 84% . .Reproduction Peregrine productivity in 1994 averaged 2.31 young fledged per successful pair and 1.58young fledged per total pair (Table 2). Since 1991, no fostering or hacking has taken place in Colorado, and yet productivity continues to remain well above the threshold of 1.25 considered necessary for population stability (Table 3). Eggshell Condition Five whole, nonviable eggs were encountered in the course of visits to 5 sites. Shell thicknesses from these eggs averaged .363mm which is 1.1% thicker than pre DDT era egg shell thicknesses, and ranged from 10.3% (.396mm) thicker to 5.0% thinner (.341mm) than pre-DDT era thicknesses. Although this sample is too small to be representative, the overall shell thickness improvement is cause for optimism. Eggshell fragments were also collected from 9 additional nesting sites in the course of visits to band young. Because shell thickness may vary as much as 5% from the egg's pole to equator and possibility of shells intermixing within the scrape, thickness measurements of fragments are judged to be only an approximation of individual egg thickness. Where multiple eggs were represented from particular sites, the measurements were combined for a clutch average. The shell thickness measurements (with membrane) from the 1994 sample represented by fragments averaged 10.0% thin (0.323mm) with extremes of 0.8% (0.375mm) above normal to 16.7% (0.299mm) below pre DDT era averages. As with 1994 whole egg measurements, the sample - size is too small to draw any conclusions. Organochlorine Residue in Eggs The 5 nonviable eggs collected during the 1994 season have been preserved along with eggs encountered in 1993, 1992 and 1991. All are awaiting scheduling for pesticide analysis by the Fish and Wildlife Service will when funding is available. �78 Release and Augmentation Efforts or Remedial management efforts such as fostering hacking have not been undertaken since 1990. The Colorado peregrine population has recovered sufficiently and eggshell thinning has improved sufficiently for the species to be down listed to threatened within the state. Prepared By:__ C=--..::~+-/?.l.-.:..-' ~~=-::":~'1-n Gerald R. Craig Wildlife Researcfier C U . _ �Table SITE NO. PRII: 196. 1. Ocupancy 196. 1965 - 1966 of Peregrine 1967 - 1968 1969 1970 Breeding 1971 1972 - -- - - --- - ---------- - --- - - --- - - - ------------ ------ ------------------------ ---- - ------------------ ---- - -- -- -- - -- ---- -- - -- -- --- ---- ---- -- --- ---- ---- ---- ---- --- - --- ---- --- --- -- -- 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1987 19S8 1989 1990 1992 1993 P P P P P V V p p V p p V p p P P P P P 1/ P P P P p p' p p p p p p p P P P P P V V V V V P P V V II V 1/ V V V p P p p V V V V p' V V V V P P P V V V V V V V V V p' V p' p p' V V V V M Ii' Ii' P P V V V V p' p V V V V V p' 1/ P V V V V M p V V V V p V V V V p' p P M V V V M V P P P P P P P P P P P 1/ 1/1 P P 1/ V V V V V V V V V V V V V P P P P P V V V V V V P P P 1/ 1/ V V V V V V V V p' V V P P P P P P V V V V V P P P V V V V V V V v v v v v V adult femalo. - - -- ----------------------- M D ~one W - Lona A - Lone adult. 2 :l:Dm. male , ad. female. l:M. male' imn. female. 6:Ad. male 5:M. male replaced by imn. male midway. 9:Ad. male paired with female prairie falcon. in Colorado 1973 1 P P P P P 14 P P p p p 2 P II P + 3 A P P P II P + A 4 P A A P + P 5 M P P P P p p p 6 p 7 p 8 p 9 v + 11 p V 12 P A V V V V 13 A M v v + V 14 Ii' P v A + 15 v v + Ii" 16 + 17 + v 18 + 19 v + 20 V + 21 V + v 22 A A 23 V + ---------------------------------------------------------(Historical 24 25 26 27 28 211 30 31 32 ' 33 34 35 37 + 38 39 .0 U .2 .3 U .5 t6 t8 t9 50 ,51 52 53 54 55 56 57 58 511 ----------------------------- Territories - p p p V A V P v p p V V P P M P p' p' 1/ v v v v v v P 14 P P P 1/ v. v v v V V p' V V 'V V V P P V V V V p' V V V I! V V M p' V V P P P P v P - v v v v v v v v v v v v ---- V v v V V V V V V V V V V V v v v v v v v v v v v v v v v V V -- - - -------, --- - v v ------- --- -- -- -- --- ---- -- v --- --- ---- --- --- - v v v v v v v v v v v v v v v v v v v V V V V V sites p - ---- -- this v v V V V M P II p' p M M V 10' V V - ---- -1/ - P P 1/ -V -- - ----- V --P - - 1/ P - P P A 1/ p' M M V V V - 14 p -- ---- ---- -- --- --- - -- - -- V V V V V V V V V v v v v v v v V v v v v v V v v P P P P v v v V Ii' 1/ p, v v v v v v v v v v v v v v v v V V v v v v v V V v V v P 1991 v P P P v v v v P l' P P v v v v v v V V v v v v v v v v v v v V V V V V p line)-----------------------------------------------------------------"-------V p v v v v V - v v v v v v v p -- --- -- -- - above v p -- ---- -- -- v v v v v 1986 V - V p V p. 'V V M p V It V V V p V p V V 11 P P 1/ P P P V p p' p V V V V P P P P p p --- -- V - --- 1/ p - - -- --- - -- - --- --- - - - -- V - V V V - - - - - V - P V -- V - P V P V V V V V P 1/ P 1/ 1/ P P P P P V V V V V V V V V V V V 1/ P 1/ P P 1/ P V V V V V V V M p' 1/ P P P P 1/ P P V P P P P P V P P P 1/ 1/ P P P P 1/ 1/ P P 1/ P 1/ P P P P P p p p' p p' 1/' P P p' P 1/ p 1/' P P p p p p' P p P P P p p p V p p p p p' p p p p p V p 1/ 1/ 1/ P P P 1/ 1/ 1/ 1/ 1/ p' V M p p' p p p p p' p p' p --- P V V V p V - - -- P P P P P P P 1/ P P P -P -1/ V p V p' -- 1994 ------ - -- V - --- V V V V p V V p V p 1/ V V -V V - V V V p V P 1/ V p p' V V - V V V -Ii' V V V - V V p V V V V P P P P P P P V P P P V V V p p M p p P P P P P P P P P P P P p' p' P II V p P p p P l' 11 P 11 P P P P P P P P P P P P 1/ 1/ 1/ 1/ P P 1/ 1/ 1/ P P 1/ P P 1/ P 1/ 1/ 1/ P 1/ P P P P 1/ 1/ 1/ P 1/ P p p 1/' P 1/ V I! V 1/ 1/ 1/ P P P 1/ P 1/, 1/ 1/ 1/ P P V V V V V 1/ P 1/ ). 1/ 1/ 1/ P 1/ 1/ P P P P p V - P 1/ 1/ 1/ adult 1/ • Adult paIr. V m Vacant alto. ~io. 4:Dcad ad. femalo 3:Ad.'famalo replaoed by imn. female midway. 7:l:Dm. male' female. 8:l:Dm. m.le. doad in vicinity. found in vicinity. '-l <.D �Table 1 (cont.) . of Peregrine Occupancy SI!l'Ii: PRII: NO. 1964 1964 1965 1966 1967 1968 1969 1970 19711972 60 61 62 63 64 65 66 67 68 69 70 71 72 73 H 75 76 77 78 79 80 81 82 83 84 85 86 81 88 89 90 A 1 5 9 -------------- -------- - ----- ----- ------ - - ------ ------ -- --- -- --- -- ----- ------ ---- - --- ---- --- --- -- --- --- - --- ---- -- -.- - -- -- - - - ----- ------ --- - - ----- -------- Breeding 1973 19741975 ----- - ----- --- -- -- ------ ----- - -- ------ ---- - Territories co in Colorado 0 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987.1988 ------------ - -- V - ---- - ---------- ---------- V V ---- V V - ----. - - - -- -- -- --- -- - V --- - -- V - ------- .V --- --- --- --- - - V - V V - V ---- - -- - -- - --- -V - -------- V . V ---V V - V - -- V ---- V V - V V ---- -- - - - -- -- - - -- -- V -- --- -- -- - - - - -- -- -- --- -- V V V V -- V - --- - V ------ V P ••Adult pair. V I: Vacant site. II • Lone adult female. M a Lone adult male. • Lone adult. 4:Dead ad. female found in vicinity. 3:Ad. female replaced by imn. female midway. Ad. 1M1e , imn. female. 2:~. male, ad. female. 6:Ad. 1M1e dead· in vicinity. 1:Imm. male' female. 8:~. 1M1". Ad. ...de replaced by imn. male midway. Ad. 1M1e paired with femalo prairie falcon. 10 Ad. male died, replaced with imn. male midway. 1989 1990 19911992 --- - --- V V V - --II - - P P P P P --- V " V - - V ---- --- M P P P A p p P P p P P M If P V P P P P V p 1~ P p 1993 1994 P P P P P V p V P p P P P V P V P p'. -P V P p' p' p' P - P p P p J? P P P p' V p' -- V p' P V V -- - p' p' p lOt -- P -- P P V V -P - M p p' J? V P P P V V p P P -P P P P J? P �81 _-_ Table 2. Summ~ Site Male A A A A A A A A A ..•.• 1 2 3 4 5 6 7 8 9 A 11 12 14 16 18 23 25 27 29 30 31 Eemale A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A 32 .--._ of 1994 Peregrine.Production Age Eggs Young A A A 33 A 34 A 35 A 37 A 38 A A 39 A I 40 A A 41 A A 42 A A 43 A A? 44 A A 45 A A 46 A A A 48 A 49 A A A 50 A 51 A A A A 52 53 A A 54 A A A 55 A A A 57 58 A A 59 A A 60 A A A 61 A 62 A A 63 A A 64 A A 66 A A 68 A A 69 A,I A 71 A A 72 A A 73 A A 74 A A 75 A A 76 T A 79 A A 80 A A 82 A A 83 A T 84 A A 85 A A 86 A A 87 A A 88 A A 89 A A 90 A A IotaI SItes Occu~Ied: 71 Total Breedin:§: airs: 61 Total atched: 125+ Pairs with own Outcome: 66 Young Fledgedrrotal Pair:1.58** YOU¥& '* I au] Hatched Young Young Eledged at 3~'los 2+ 4 + 2+ 2+ 2+ 2 2 3 3 3 0 2+ 2+ 0 0 2+ 4 4 4 2+ 2+ 2+ ? ? ? ? ? 2+ + 3+ 4 2+ 3+ 4 0 2+ 0 3+ 3 2 2+ 3+ 4 0 2 0 3 3 2 2 4 4 2+ 2+ + 4 + + 3+ + 0 2+ 2+ 1+ 2+ 2+ + 2 + 0 3+ 1+· 0 2+ 2+ 1+ 0 3+ 4 2+ 3+ 3 3+ 2+ 3+ 3+ 2+ 3+ 2+ 3+ 2+ 2+ 3 + 1+ 2+ 3+ 2+ 0 + 0 4 0 2+ 0 3+ 3 2 2 3 4 4 2+ 2 ? 2 4 ? 0 3+ ? ? ? 2+ 0 2 2+ 1+ 0 3+ 3+ 2+ 2+ 0 + 2+ 1+ 0 3+ 3+ 2+ 0 2 2+ 1+ 0 3+ 4 2+ 3 3 3 2+ 3+ 3 2+ 3+ 2+ 3 2 2 2 + 1+ 2+ 3 2 0 0 0 4 ? 2 0 0 2+ 1+ 0 3 3+ 0 ? ? ? ? ? ? ? 3+ 4 2+ 3+ 3 3+ 2+ 3+ 3+ 2+ 3+ 2+ 3+ 2+ 2+ 4 + 1+ 2+ 3+ 2+ + + + 4 3+ 3+ 3+ 3+ 3 3+ Total Adult Pam;: 65 Total Successful Pairs: 45 Total Young Fled§ed: 105 . Average Fledged rood Size:231 * Total Fledglings Divided by Total Successful Pairs with Known Outcomes. 0 0 2 4 2 3 4 1 2 0 0 2 0 0 3 ? 0 ? ? 1 2 1 0 3 4 2 2 3 3 2 3 3 2 2 2 3 2 0 1 ? 1+ 2 3 2 0 0 0 1 ? 2 0 0 2 1 0 0 3 0 3 3 ** Pairs with known outcomes. �82 Table 3. Reproductive summary of Colorado peregrines. SUitable SItes on Record Occupied Sites Adult Pairs Breeding Pairs! Successful Pairs'' Young Hatched Young Fledged Occupancy Rate" Percent Pairs Breeding Productivity" Nest Success' I~~i II)Y~ lYY~ I~~4 69 58 51 49 40 100 91 84% 92% 1.72 82% 76 60 51 47 36 114 86 79% 82% 151 77% 80 60 56 52 39 113 92 75% 87% 1.61 75% Pairs that prOduced eggs Pairs that fledged young 3 Occupied sites/Total sites 4 Young fledged/Sites occupied by pairs with known outcomes i 2 85 71 68 61 45 125 105 84% 86% 1.58 72% �83 JOB PROGRESS REPORT State of Project: __ C"".()=lo:'-!r.::>ad""'o""-__ .•.. (W.:..,:_-~15::...:1:....-=-R:....-6.::..)'___ : Bald Eagle Nest Site Protection and Enhancement Program Period Covered: 1 July, 1993 - 30 June, 1994 Personnel: G.R. Craig, Colorado Division of Wildlife ABSTRACT Bald eagles occupied 17 Colorado nesting territories in 1994. Two new territories were discovered and two were unoccupied. Sixteen nesting territories produced eggs and nine pairs were confirmed to have successfully fledged 16 young. Productivity averaged 0.94 young per occupied territory. This Job Progress Report represents a preliminary analysis and is subject to change. For this reason,· information presented herein MAY NOT BE PUBLISHED OR QUOTED without permission of the author. ��85 BALD EAGLE NEST SITE PROTECTION AND ENHANCEMENT PROGRAM Gerald R. Craig SEGMENT OBJECTIVES 1. Monitor nest site occupancy and reproductive success. 2. Document survival rates and mortality factors. 3. Determine migration and wintering areas. 4. Determine if philopatry occurs in breeding eagles. 5. Determine nest site tenacity by individual breeding eagles. 6. Quantify nesting habitats and associated foraging areas in an effort to document nest site parameters conducive to improved reproduction. 7. Document pesticide contamination through eggshell measurement and chemical analysis of nonviable eggs. 8. Where necessary, implement actions to stabilize nests and maintain occupancy. METHODS AND MATERIALS This work will be a cooperative endeavor between the Division and Dr. Richard Knight of Colorado State University. 1. Annually visit all documented breeding sites to determine the presence of bald eagles. Pairs at territories will be documented by DWMs and other field personnel. Previously urirecorded pairs will probably be revealed in the course of aerial eagle and waterfowl flights. DWMs will confirm actual incubation from ground visits. 2. Occupied territories will be visited by DWMs periodically throughout determine hatch of young, nesting failures, etc. 3. In May and June, a Utility Worker will observe breeding eagles from a distance and endeavor to follow their movements to locate important foraging areas. Responses of eagles to various human activities and land uses will be recorded. 4. In June, when the young are determined to be old enough to band, sites will be visited by Craig and . Knight to place a federal band on one leg and a colored, alpha numeric marker on the other. The color markers will permit identification if the young return in subsequent years. During the same nest visit the following will be recorded: . Physical parameters such as tree species, height, DBH, condition, and dominance. Nest condition, size, and location. Vegetative community and land use practices. In addition, collect prey remains, nonviable eggs and eggshell fragments. 5. Approximately Sec's of blood will be collected from each nestling. The blood will be analyzed at the Savannah River Ecology Lab in Aiken, South Carolina. Electrophoretic examination will permit the breeding season to �86 genetic comparison with samples collected from other populations in Saskatchewan, the Lake States and Arizona, as well as determine the heterogeneity of the Colorado birds. 6. When necessary, remedial actions will be taken to stabilize nests that are threatened by wind throw. Should the tree be decadent and in danger of falling, an artificial nest base may be placed in a suitable, adjacent tree. Action will be taken only after it has been deemed desirable to encourage the eagles to nest at the same location. RESULTS AND DTSCUSSTON Territory Occupancy Bald eagle nesting activities for Colorado are summarized in Table 1. In 1994,17 territories were occupied (Adams, Jefferson, Gunnison, La Plata #1, Mesa #1 and #2, Moffat #1, #2, #3 and #4, Montezuma #3, Morgan, Rio Blanco #1, #3, and #4, Routt, and Weld #3). Two new territories (Mesa #2 and Moffat #4) were added in 1994. The Jefferson site was occupied by an adult pair in 1993, but since they did not produce eggs that year, were not included in the 1993 summary. In 1994 the pair reoccupied the site and by their behavior on numerous occasions, were judged to have incubated eggs. No prior history of use was discovered for either the Mesa #2 or Moffat #4 sites. The Mineral site reported in 1993 was again occupied, but reproduction could not be confirmed. Pairs could not be located at the Fremont or Montezuma #4 sites. Although the Routt site was frequented by an adult pair and a subadult, it was classified as inactive since the nest was usurped by Canada geese and the pair did not nest elsewhere. Land Status The new territory at Mesa Co. #2 is administered by BLM and the Moffat Co. #4 site is on state land bordering private property. Livestock grazing is the primary land use at both sites although the Mesa Co. #2 is adjacent to a river and receives rafting pressure. Reproduction Reproductive efforts are summarized in Table 2. Nineteen young were hatched by 10 pairs and 16 were fledged by 9 pairs (1.78 young per successful pair) which yielded an overall productivity of 0.94 young per territory occupying pair. Reduced productivity resulted when 5 pairs failed during incubation, 2 due to wind throw and 3 for unknown causes. The pair at the Routt County site apparently were displaced by Canada geese and did not build an alternate nest. Sixyoung were banded and color marked at 3 locations (La Plata Co. #1, Moffat Co. #4 and Rio Blanco #3). Fish and Wildlife Service bands were affixed to the nestlings' right legs and red alpha-numeric bands with yellow vinyl flags were affixed to their left legs. Culmen length and foot pad length measurements were obtained from the eaglets that were banded. Genetic Variation A manuscript presenting the results of the electrophoretic analysis of blood and tissue samples for genetic variation has been. accepted for publication. Banded Adults No observations of color marked individuals or band recoveries were reported during this segment. �87 Nest Stabilization Efforts No nest stabilization efforts were undertaken in 1994. The nest at Rio Blanco Co. #4 slid off the supporting limbs during incubation. This was unanticipated given that the nest was placed on substantial support boughs near the trunk of a conifer. The pair at Moffat Co. #2 relocated across the river to a heron colony and built upon an old heron nest. As had occurred in 1987, nest placement was not substantial and it blew down midway through incubation. Pairs at Rio Blanco Co. #1 and #3 remained secure in spite of the decadence of the cottonwood trees in which they were situated. As the young at Rio Blanco #1 neared fledging a . lightning fire destroyed almost all the cottonwoods in the vicinity. The local fire department went beyond the call of duty to wrestle pumps overland to the site, climbed and adjacent tree and put out the nest tree at dusk. One young was lost but the other survived. The tree has ben badly damaged and will probably succumb. The nest tree at Rio Blanco #3 is almost completely dead. Arrangements are underway with the Rio Blanco #1 and #3 landowners to develop alternate nest platforms to encourage the eagles to remain on their properties. Prepared by: a.£.C~..;. Gerald R. Craig J Wildlife Researcher C �co co Table 1. Bald Eagle Nesting Site La Plata Co. #1 Moffat Co. #1 La Plata Co. #2 Grand Co. Montezuma Co. #1 Rio Blanco Co. #1 Rio Blanco Co. #3 Weld Co. #1 Montezuma Co. #2 Moffat Co. #2 Moffat Co. #3 Adams Co. Archuleta Co. Montezuma Co. #3' Weld Co. #2 La Plata Co. #3 Rio Blanco Co. #4 Morgan Co. #1 Mesa Co. #1 Fremont Co. Routt Co. Gunnison Co. Mineral Co. Weld Co. #3 ·Montezuma Co. #4 Jefferson Co. Mesa Co~ #2 Moffat Co. #4 Total Young Total Pairs Young/Occ. Terr. IA = Inactive A 1974 1975 legg fA -- 1yng -- --- --- --- --- -~ -- --- --- --- --- --- --- --- --- --- --- --- --- --- --- --- --- --- --- --- --- -0 1 1 1 0.00 1.00 = Active 1976 fA 2yn9 2yng -------------------------- Efforts 1977 IA 2yng 2yng --------' ------_-----------4 4 2 2 2.00 2.00 1978 IA 2yng 2yng Oyng ------------------------- in Colorado 1979 IA 1yng Oyng Oyng ------------------------4 1 3 3 1.33 0.33 1980 ? -IA IA A ------------------------ 1981·1982 1983 ? ? ? 2yn9 2yng -IA IA IA IA IA IA A A A 1yng 1yng? -- 3yng 2yng -- -- --- -- --- -- --- -- --- -- --- -- --- -- --- -- --- -- --- -- --- -- --- -- --- -- --- -- --- -- --- -- --- -- --- -- --- -- --- -- --- -- -0 3 6 2 1 3 4 2 0.00 1.00 1.50 1.00 1984 ? 2yn9 IA IA IA eggs 2yng 2yng --------------------6 4 1.50 1985 ? Oyn9 IA IA IA Oyn9 2yng 2yng 2yng --~ -----------------6 5 1.20 1986 eggs 1yn9 IA IA IA 2yng Oyng eggs 1yng 1yng legg eggs eggs ---------------5 10 0.50 1987 2yng 2yng A IA IA 2yng lyn9 IA 1yng Oyng IA 1egg 2yng ---------------10 9 1.11 1988 1989 1yng 2yng 1yng 3yng IA IA IA IA IA IA 2yng 2yng A 2yng IA CIA lyn9 1yng 2yng 3yng 7 eggs eggs 2yng IA IA 1yng lyng -- eggs -- --- --- --- --- --- --- --- --- --- --- --- --- -8 16 8 10 1.00 1.60 1990 Oyng 2yng IA IA IA 2yng lyng Oyng lyng 2yng IA 2yng IA 1yng IA 2yng 2yng ---------' ---13 10 1.30 1991 1992 1yn9 1yng 2yng 2yng IA IA IA IA IA IA 2yng 1yng 3yng 3yng IA IA Oyng 1yng 2yng 3yng Oyng Oyng 3yng 3yng Oyng Oyng 2yng 1yng IA IA 2yng? lyng 2yng 2yng 3yng Oyng eggs -- --- --- --- --- --- --- --- --- -19 .. 20 13 14 1.46 1.43 1993 1994 2yng 3yng 2yn9 2yn9 IA IA IA IA lA "IA 2yng 2yng 2yng 1yng IA IA IA IA Oyng Oyng 2yng 2yng 2yng 3yng eggs lA 2yng 1yng lA IA ? IA 3yng Oyng 2yng Oyng. eggs Oyng 2yng IA 2yng lA 1yn9 2yng A ? A Oyng 0 IA 0 Oyng -- 2yng -- 1yng 18 16 18 16 1.00 1.00 �89 Table 2. Colorado Bald Eagle Nesting Age at Bi rds Male Female Site Adams Co. #1 Jefferson Co. Gunnison Co. La Plata Co. #1 Mesa Co. n Mesa Co. #2 Moffat Co. #1 Moffat Co. #2 Moffat Co. #3 Moffat Co. #4 Montezuma Co. #3 Morgan Co. Rio Blanco Co. n Rio Blanco Co. #3 Rio Blanco Co. #4 Routt Co. vicinity. Weld Co. #3 Total Adult Adult Adult Adult Adult Adult Adult Adult Adult Adult Adult Adult Adult Adult Adult Adult Adult Adult Adult Adult Adult Adult Adult Adult Adult Adult Adult Adult Adult Adult Adult Adult Efforts - 1994 Corrment s : Young Young Produced Fledged 3 3 0 0 1 3 1 3 0 0 2 2 2 Pair incubated, failed. Female incubating, fail ed. 2 0 0 2 1 1 ? New nest, blew out wi th small young. 1 1 Pair incubated, failed. Tree nearly destroyed by fire, 1 young died. 0 0 2 2 1 2 0 0 0 0 Nest sl ipped out of tree early in incubation. Goose took over nest, pair and yearling in Pair disturbed Adult Adult 0 0 17 17 19 16 at egg laying. ��91 JOB PROGRESS REPORT State of Project: Period Covered: ---"=C:,.><o=lo"-'-r=ad=o"--__ ----'.,-(W~.•..• 15"-"6'---R~-4!..,l.)--_ :Effects of Human Disturbance on Grassland Raptors 1 July, 1993 - 30 June, 1994 Personnel: G.R. Craig, Colorado Division of Wildlife and R.L. Knight and T. Holmes, Colorado State University ABSTRACT The data was compiled and analyzed and thesis prepared by Tamara Holmes entitled Behavioral Responses of Grassland Raptors to Human Disturbance in partial fulfillment of a Master of Science degree at Colorado State University. ;~~ ~::==LI1 , f=~r- i(,)_ cO .J1 ,~_=c :5_-0 ~__ :3 :=i~ This Job Progress Report represents a preliminary analysis and is subject to change. For this reason, ".~ information presented herein MAY NOT BE PUBLISHED OR QUOTED without permission of the author. .-.-~--- ��93 THESIS BEHAVIORAL RESPONSES OF GRASSLAND RAPTORS Submitted Tamara Department In partial fulfillment Colorado by Lynn Holmes of Fishery and Wildlife of the requirements state University, Spring COLORADO TO HUMAN DISTURBANCE Biology for the Master Fort Collins, of Science Degree Colorado 1994 STATE UNIVERSITY 21 January 1994 WE HEREBY RECOMMEND THAT. THE THESIS PREPARED UNDER OUR SUPERVISION BY TAMARA LYNN HOLMES ENTITLED BEHAVIORAL RESPONSES OF GRASSLAND RAPTORS TO HUMAN DISTURBANCE BE ACCEPTED AS FULFILLING IN PART REQUIREMENTS FOR THE MASTER OF SCIENCE DEGREE. Committee on Graduate Adviser .Department Head Work �94 ABSTRACT BEHAVIORAL RESPONSES OF THESIS OF GRASSLAND RAPTORS TO HUMAN DISTURBANCES As human activities increase on public lands, managers must ensure that wildlife can forage, rest and reproduce undisturbed. Often, spatial and temporal buffer zones are used to minimize disturbance of wildlife populations. Spatial buffer zones are radii of restricted human activity around foraging or nesting areas, whereas temporal buffer zones are limitations on the times (e.g., hours, days, months) that human activities are allowed in sensitive areas. Understanding how raptors respond to various types of disturbance and how responses change over repeated disturbances enables managers to calculate adequate species-specific buffer zones. In the winters of 1990-1991 and 1991-1992, I measured the flushing responses and flush distances of 6 species of diurnal raptors [American kestrels (Fa.lco sparverius), merlins (F. co.lumbarius), prairie falcons (F. mexicanus),. rough-legged hawks (But:eo .lagopus), ferruginous hawks (B. rega.lis), and golden eagles (Aqui.la chrysaet:os)] exposed to walking and vehicle disturbances in northern Colorado. Walking disturbances resulted in more flushes than vehicle disturbances for all species except prairie falcons. Although flush distance did not vary with disturbance type for the 3 falcon species, rough-legged hawks and golden eagles flushed at greater distances for walking disturbances and ferruginous hawks flushed at greater distances for vehicle disturbances. Merlins and prairie falcons perched along paved roads had shorter flush distances to walking disturbances than individuals perched along gravel roads; individuals along paved roads may have habituated to 'human activity or, alternatively, were more tolerant of disturbance. Rough-legged hawks perched nearer to the road flushed at greater distances than those further away. Similarly, American kestrels, prairie falcons, and ferruginous hawks perched closer to the ground had greater flush distances than those perched higher. Dark-morph ferruginous and rough-legged hawks flushed at greater distances than light morphs. For walking disturbances, a linear relationship existed between flight distance and body weight, with lighter species flushing at shorter distances; however, this trend did not hold for vehicle disturbances. In the summers of 1992 and 1993, I measured the flush distances, firstcall distances, call rates, and return-to-nest times of 3 species of nesting buteos exposed to controlled human disturbances that differed in nest-visit . frequency, duration of nest visit, and interval between visits. With repeated visits, red-tailed hawk (B. jamaic.ensis) call rates decreased and ferruginous hawk flush distances increased. Repeated nest visits with the longest durations resulted in decreasing red-tailed hawk return times and visits with the shortest intervals resulted in increasing return-to-nest times for redtailed hawks and increasing flush distances for ferruginous hawks. Red-tailed and Swainson's hawk (B. swainsoni) first-call distances were shortest for disturbances of long duration, and both species were more likely to flush away when the mate was absent. Regardless of species, individual call rates increased when the mate was pre~ent. Reaction intensities for all species increased at nests which were closer to the ground and at nests closer to sources of permanent human activities. Overall, nest defense was greatest during the nestling phase; however, ferruginous hawks exhibited the lowest levels of nest defense regardless of reproductive phase. During incubation, all species were more likely to flush away from the researcher, but, during the nestling phase, red-tailed and Swainson's hawks were more likely to flush toward the researcher. Red-tailed hawks called the most, but red-tailed and ferruginous hawk call rates were greater during the nestling period. Regardless .. of reproductive phase, Swainson's hawks required the smallest buffer zones. Because of �95 interspecific and intraspecific variation in reaction to disturbance, managers should examine population, habitat, and disturbance characteristics of species communities before implementing buffer zones. . Although anecdotal evidence suggests that incubating ferruginous hawks desert the nest when disturbed, nest failures due to experimental disturbances did not occur in this study. However, ferruginous hawks often disappeared from the nest area for several hours. The lack of an incubating or brooding parent during this time may have lowered productivity, something which this study does not address. Furthermore, the long-term impacts of disturbance were not examined, but may include failure to return to the nesting area in future reproductive attempts and long-term habituation or sensitization to disturbance. . Tamara L. Holmes Department of Fishery and Wildlife Colorado state University Fort Collins, CO 80523 Spring 1994 Biology ACKNOWLEDGMENTS Support for this research came from the Colorado Division of Wildlife. (CDOW), through the Federal Aid in Wildlife Restoration Project W-150-R, Colorado State University (CSU), and the U. S. Forest Service (USFS). The advice and encouragement of my adviser, Dr~ Richard L. Knight, refreshed me on several occasions, and the statistical aid of Dr. Kenneth P. Burnham eased numerous uncertainties. Wildlife biologists Gerald R. Craig (CDOW) and Dennis P. Lowry (USFS) served on my graduate committee and were always available for assistance, reassurance, and conversation. I thank CDOW District Wildlife Manager, John Wagner, for introducing me to the Pawnee National Grasslands and trusting me to practice ethical science. Libby Stegall initiated the winter research and helped me to make a smooth transition into graduate research. I appreciate the field assistance of Laurie Masten and Erik Sparling. Fellow graduate students and friends, especially Richard J. Camp, Claire L. DeLeo, Jon Judson, Kay Neumann, Russ Pickering, and Daniel E. Varland, were always willing to share their experiences and opinions • .. ";.- �96 TABLE OF CONTENTS TITLE PAGE SIGNATURE PAGE ••• ABSTRACT OF THESIS TABLE OF CONTENTS ACKNOWLEDGMENTS CHAPTER 1 RESPONSES OF WINTERING .' GRASSLAND RAPTORS TO HUMAN DISTURBANCE INTRODUCTION METHODS Study area Data collection Data Analyses RESULTS Walk Disturbance Vehicle Disturbance Body Weight DISCUSSION LITERATURE CITED CHAPTER 2 RESPONSES OF 3 SYMPATRIC BUTEOS ,. '\ TO DISTURBANCE DURING THE BREEDING INTRODUCTION METHODS Study Area Study Design Data Collection Data Analysis RESULTS • Between-Species Comparisons Flushing Response Distance Call Rates Return-to-nest Times Buffer Zones Between-Pair Comparisons Flushing Response. Distance Call Rates Return-to-nest Times Within-Pair Comparisons The Role of Nest Visit Frequency DISCUSSION Disturbance Duration, Interval, and Frequency Nest Characteristics and Mate Location Reproductive Phase Species Differences and Buffer Zones Conclusion LITERATURE CITED SEASON ,. "" _,' h" _ ~'~_'.. ~ _.' :•.. ,. �97 CHAPTER RESPONSES OF WINTERING GRASSLAND 1 RAPTORS TO HUMAN DISTURBANCE INTRODUCTION Human activities can affect the geographic distribution, dispersion patterns, habitat use, fecundity, survival, and activity and energy budgets of wildlife populations (Knight and Cole 1991). In most cases, the magnitude and frequency of disturbances determine whether or not the fitness of affected individuals will be altered. Additive and synergistic effects from multiple sources of disturbance can result in reductions of fitness, even when individual types of disturbance alone have no impact. Experiments are necessary to understand which aspects of human activity affect wildlife, and to develop management schemes that minimize conflicts between animals and human activities (Knight and Skagen 1988, Gutzwiller 1991). One such management practice is the creation of restricted areas that spatially or temporally separate wildlife from potential disturbance. For these buffer zones to be effective, they need to be based on empirical evidence of wildlife responses to disturbance (Knight and Skagen 1988). Most studies on wildlife disturbance have focused on the reproductive period and have addressed single species. As a result, we lack understanding of how concepts relating to human disturbance apply during the nonbreeding season. In addition, comparative studies that examine species assemblages can provide insights not evident from single-species studies (Skagen et al. 1991). I recorded flushing responses (whether animals flee because of a disturbance) and flush distances (the distance between the disturbance and the animal when flushed) of wintering grassland raptors disturbed by pedestrians and vehicles. My objectives were to compare species-specific responses to the 2 disturbance types, to examine the hypothesis that flush distances positively correlate with species' body size (Cooke 1980), and to calculate minimum distances £or species-specific buffer zones. METHODS Study area I conducted the study on a 5,870 km2 portion of northern Weld County, Colorado. Ranches and farms comprise 4,760 km2 of this area; farms consist primarily of alfalfa and winter wheat. The remaining 1,110 km2 consists of state-owned land (265 km2), the Central Plains Experimental Range (65 km2), and the Pawnee National Grassland (PNG) (780 km2). Walking and vehicle .disturbances do not represent every form of disturbance, but are the most common forms of disturbance where I conducted my studies. Ranchers, maintenance crews, and birdwatchers frequently walk and drive within sight of pe.:ochedraptcr--s. The PNG is managed. by the U. S. Forest Service and .is shortgrass prairie consisting primarily of blue grama (Bouteloua gracilis) and buffalo grass (Buchloe dactyloides). This area is described by Jameson and Bement (1969) and Olendorff (1973). Data collection I collected data during 50 visits to the study area between 16 December 1990 and 23 February.1991, and between 4 December 1991 and 9 January 1992. To reduce the probability of sensitizing or habituating raptors (Knight and Cole 1991), I minimized repeated encounters of individuals by dividing my study area into 10 units, each approximately 585 km2 and systematically rotatin.g:my. visits among these sites. During each visit, I systematically surveyed every j" 1 �98 road within a unit, and I was careful to identify landed to avoid resampling individuals. where the flushed raptors Previous trials indicated that perched raptors flushed before I began my approach if I initiated wal~ disturbances closer than 450 m (1000 m for roughlegged hawks [Bu~eo lagopus]), so I used a Leitz range finder to determine starting points. I used the same range finder to measure flush distances and calibrated it each morning. I also wore the same clothing, a brown coat and blue jeans, for all disturbances. A walk disturbance consisted of me driving to within 450 m (or 1000 m) of a perched raptor, leaving the parked vehicle, and walking along the center of the road at a constant 1.5 mJsecond rate towards the perched bird until it flushed or, if it did not flush, until parallel with the perch (i.e., birds were always perched 1 m to 400 m away from the road's edge). I.only approached raptors where there was an unobstructed line of sight to them. If the raptor flushed, I ceased walking and used the range finder to measure the distance from me to the base of t~e vacated perch. I also used the range finder to measure the perpendicular distance from the perch to the road center. A Suunto clinometer was used to estimate perching height. I sampled raptors perched <400 m from the road because, at distances >400 m, I had difficulty determining whether a raptor flushed as a result of my disturbance. Similarly, I did not sample raptors perched on the ground or at heights >20 m.. After each trial I visually estimated cloud cover as <50%, 50%-90%, or >90%, measured wind speed to the nearest 1.6 kph with an anemometer, and measured temperature to the nearest 1 C with an outdoor thermometer. Two similar station wagons (1 per year) were used for all vehicle disturbances. I proceeded as for walk disturbances, except I remained in the vehicle and maintained a constant speed of 70 kmJh while driving on the road center, approaching a point parallel with the perched raptor if it did not flush. If the raptor flushed, I dropped a marker onto the road, stopped the vehicle, and measured the distance from the marker to the base of the vacated perch with the range finder. All other measurements were made as for the walk .trials. I recorded species, sex (for American kestrels [Falco sparverius] and merlins [F. columbarius]), age (adult vs. immature for golden eagles [Aquila chrysae~os]), color morph (for rough-legged and ferruginous hawks [B. regalis]), perching height (m), distance from center of road to perch (m), perch type (pole, fencepost, tree, or windmill), and road type (gravel, dirt, or paved) for each trial. Disturbance types were alternated as raptor species were encountered to ensure balanced sample sizes. Data Analyses I tested for a year effect for flush distances using a univariate ~test, and for flushing responses using a chi-square test. There were no detectable between-year differences for flush distances (~ = 0.334, 324 df, P = 0.7383) or flushing responses (X2 = 0.024, 1 df, P = 0.8778), so I pooled the data over years. I verified that flush distance data from each species fit normality by plotting histograms and examining residuals. After separating data by species and disturbance type, I performed multiple linear regressions (with flush distance [m] as the dependent variable) and logistic regressions (with flushing response [0 = no flush, 1 = flush] as the dependent variable). Each regression initially included time of day, cloud cover, wind speed, location, road type, perch type, age (for golden eagles), sex (for kestrels and merlins), color morph (for rough-legged and ferruginous hawks), perching height (mj , and perch distance from road center (m) as independent v.ariables. Backward elimination was used to remove insignificant (P > 0.05) variables. No strong .collinearities were. present .among .. the ... significant ... remaining independent variables. I used li2 values from linear regressions and Goodman- 2 �Kruskal Gamma (GKG) values power of each model. from logistic regressions to assess the predicti.ve To detect interactions, I ran separate analyses of covariance for flush distance and flushing response from walking and vehicle disturbances. I included independent variables from the regression analyses (Table 1-1) and all possible interactions between these variables, except that perching height (PERHT) and perch distance from road center (PERDIS) were included as covariates and not included in interaction terms. For each of the 4 disturbance type/dependent variable combinations, 2 analyses of covariance were performed, 1 using PERHT and the other using PERDIS as the covariate. If neither variable was a covariate (P > 0.05), I performed an analysis of variance without PERHT and PERDIS; all other independent variables and Table 1-1. Independent variables selected in multiple linear regression and logistic regression analyses for flush distances and flush~ng responses of raptors exposed to walking disturbances in northern Colorado during the winters of 1990-1991 and 1991-1992. . variable symbol Variable description and coding ROAD Road type: 1 2 3 = = gravel dirt paved PERCH ·Perch type: PERHT PERDIS 1 pole 2 = fencepost 3 tree 4 = windmill Distance from ground to location of perched Perpendicular distance from perch to center raptor (m) of road (m) interaction terms remained.. I considered independent variables and interactions significant if the Type III Sums of squares P < 0.05. For each disturbance type, I used Tukey's HSD tests (experiment-wise error rate = 0.05) (Montgomery 1991:78) to test whether mean flush distances differed, and chi-square tests to test whether the frequency of the flushing responses (0 = no flush, 1 = flush) differed, between pairs of species, perch types, or road types. SAS procedures FREQ, GLM, LOGISTIC, and REG were used to analyze the data (SAS .Inst. Inc. 1988:544, 623, 195, 864). RESULTS Three-hundred twenty-six trials were run for 6 species (Table 1-2). Northern harriers (Circus cyaneus) and.red-tailed hawks (B. jamaicensis) were sampled but-were not included in the analyses because sample sizes were ~ 11. Kestrels, merlins, rough-legged hawks, ferruginous hawks, and golden eagles were more likely to flush when approached by a human on foot than an automobile (X2 = 33.57, 23.84, 41.05, 25.07, and 23.68, respectively, 1 df each and all P ~ 0.0001), but prairie falcons were equally sensitive to both disturbance types (X2 = 2.01, 1 df, P = 0.1564). Although flush distance did not vary with disturbance type for kestrels, merlins, and prairie falcons (t = 0.45, 35 df, P = 0.6558; t = 0.49, 15 df, P = 0.6282; and t = 0.60, 49 df, P = 0.5524, respectively), it was significantly greater during walk disturbances for rough-legged hawks and golden eagles (t = 4.18, 68 df, P = 0.0001 and t = 2.78, 19 df, P 0.0119, r~spectively) and greater: during vehicle disturbances for ferruginous hawks (t' == 3~67, 22 df; P':'; 0.0013). Overall, ·97% of raptors I approached on foot flushed with a mean flush distance of 118 m, whereas, 38% of raptors approached by car flushed with a mean flush distance of 75 m (Table = 3 �100 1-2). Date, time, and location of disturbance (i.e., road coordinates), as well as temperature, wind speed, cloud cover, age of golden eagles and sex of kestrels and merlins were eliminated from the flushing response and flush distance models for all species and both disturbance types (all P > 0.05). Walk Disturbance Because flush distance varied among species (13.~= 20.72, P = 0.0001), I developed individual regression models for each species (Table 1-3). Distance of perch to road center was a significant covariate (P = 0.0001), and scatterplots indicated a general trend of decreasing raptor flush distances with increasing perch-to-road distances. Analysis of covariance for flushing response failed to detect any significant variables, presumably because 97% of the raptors flushed at my approach. Regression models for flushing response were only meaningful (contained explanatory variables) for the ferruginous hawk (probability of flushing = 1 /(1 + e r-3.665+0.029PERDIS1), GKG = 0.818), Flushing responses and flush distances Table 1-2. walk and vehicle disturbances in northern Colorado 1990-1991 and 1991-1992. (m) of raptors exposed during the winters of Flush distances to (m) n % flushed X SE Range 28 14 100 100 44 76 5 13 10-100 17-180 prairie falcon rough-legged hawk ferruginous hawk golden eagle Total 33 45 24 18 162 91 100 92 100 97 92 177 63 225 118 8 19 9 19 8 24-185 55-900 13-165 105-390 10-900 Vehicle disturbance American kestrel merlin prairie falcon rough-legged hawk ferruginous hawk golden eagle Total 33 10 27 62 16 16 164 27 30 78 40 13 19 38 40 62 .85 71 195 82 75 11 12 11 8 85 54 7 12-115 44-85 18-200 9-170 110-280 14-190 9-280 Species Walk disturbance American kestrel merlin Regression equations for flush distances Table 1-3. exposed to walking disturbances in northern Colorado 1990-1991 and 1991-1992. Species American merlin prairie kestrel falcon rough-legged ferruginous golden eagle hawk hawk !l B2 28 14 of 6 raptor species during the winters of P Equation 0.125 0.556 0.0408 0.0116 30 0.253 0.0125 45 0.291 0.0007 22 0.241 0.0456 0.296 0.0196 68.152-2.879(PERHT) 221.093-47.001(PERCH) -34.961(ROAD) 185.739-43.219(ROAD) -4.949 (PERHT) 113.422+87.766(PERCH) -2.672(PERDIS) 96.916-4.455(PERHT) . +0.432(PERDIS) .173.654+0.390 (PERDIS). 18 .. "-." 4 �101 and I found no significant interactions with flush distance or flushing response (flush distance: £:3,99 = 0.29, P = 0.8325; flush response: £:3,99 = 0.07, P = 0.9758). Golden eagles flushed at greater distances than kestrels, merlins, prairie falcons, rough-legged hawks, and ferruginous hawks (t = 18.47, 43 df; t = 12.95, 30 df; t = 14.40, 49 df; t = 5.27, 61 df; and t = 15.95, 40 df, respectively, all P < 0.0001). In turn, rough-legged hawks flushed at greater distances than kestrels, merlins, prairie falcons, and ferruginous hawks (t = 17.05, 70 df; t = 10.28, 57 df; t = 12.00, 76 df; and t = 13.86, 67 df, respectively, all P < 0.0001), and prairie falcons flushed at greater distances than kestrels (t 5.40, 58 df, P < 0.0001). Dark-morph roughl~gged and ferruginous hawks flushed at greater distances than light morphs (t = 2.60, 43 df, P = 0.0127; t = 2.38, 22 df, P = 0.0264, respectively). = Vehicle Disturbance I developed individual regression models for each species because raptor flushing response and flush distance varied among species (£:5,123 = 7.13, P = 0.0001 and £:5,110 = 5.33, P = 0.0002, respectively). I found no significant covariates or variable interactions for flushing response or flush distance from vehicle disturbance. Prairie falcons were more likely to flush than kestrels, merlins, roughlegged hawks, ferruginous hawks, and golden eagles (X2 = 14.45, 1 df, P = 0.0001; x2 = 7.31, 1 df, P = 0.0070; x2 = 10.57, 1 df, P = 0.0010; x2 = 17.21, 1 df, P = 0.0001; x2 = 15.20, 1 df, P = 0.0001, respectively) and flushed at greater distances from the approaching vehicle than kestrels, merlins, roughlegged hawks, and golden eagles (t = 7.00, 57 df; t = 7.10, 35 df; t = 4.99, 87 df; t = 4.50, 41 df, respectively, all P < 0.0001). Flushing responses and flush distances of dark-morph rough-legged and ferruginous hawks did not differ from those of light morphs (X2 = 0.70 and 0.04, 1 df each, P = 0.4040 and 0.8489; t = 0.77 and 0.75, 59 df and 14 df, P = 0.4444 and 0.4657, respectively, for flushing response and flush distance). Body Weight I examined the relationship between raptor body weight and flush distance for each disturbance type. Mean body weights for each species were obtained by averaging the weights given for males and females in Johnsgard (1990). For walking disturbance data, a positive linear relationship existed (P = 0.0188), with lighter species flushing at shorter distances than heavier species; however, no such trend was evident in the vehicle disturbance data (P = 0.7451) (Fig. 1-1). Pooled data were used to examine species tolerance to disturbances by comparing the distances from disturbance with the cumulative flushing percentages for each species. Thresholds varied among species; small species rarely flushed at distances >125 m, whereas large species often flew when disturbances were >200 m away (Fig. 1-2). DISCUSSION Human disturbance activities may elicit various responses from wildlife and differences in persecution histories result in diverse wildlife responses to different activity types (Knight et ale 1989, Knight and Cole 1991). In my study, most species were more likely to flush when approached by a human on foot than when approached by an automobile. Skagen (1980) reported that bald eagles (Haliaeetus leucocephalus) were less likely to flush from humans approaching in vehicles than from pedestrians. Similar behavior has been reported for waterbirds (Klein 1993). That a person approaching in plain view of a raptor elicits a stronger response than a person within an approaching vehicle suggests thatnunians approaching slowly are viewed as a greater disturbance than vehicles, which are moving rapidly and screen humans. 5 �102 250 y= 63.05 + 0.04x R2 200 6 = 0.713 (WALKING) f R ill ~ 150 ~ (f) o :r: 100 ::J (f) 9 r -l F u, a o y= 83.08 + 0.01x R2 = 0.019 (VEHICLE) o 1,000 2,000 3,000 4,000 5,000 BODY WEIGHT (G) Figure 1-1. Relationship between mean flush distances and 'mean body masses for 6 raptor species in northern Colorado during the winters of 1990-1991 and 1991-1992. Mean masses of males and females (Johnsgard 1990) were averaged to obtain mean body masses. Each species is represented by the first letter of its common name (A = American kestrel, M = merlin, P = prairie falcon, R = rough-legged hawk, F = ferruginous hawk, G = golden eagle); uppercase letters symbolize walking disturbance data and lower case letters symbolize vehicle disturbance data. 120~---------------------------------------~---------------------~ Figure 1-2. Relationship between distances from disturbance· (combined walking and vehicle data) and cumulative percentages of individuals flushing for populations of 6 raptor species in northern .Colorado during the winters of 1990-1991 and 1991-1992. Each species is represented by the first letter of its ~ommon name (A = American kestrel, M = merlin, P = prairie falcon, R = rough-legged hawk, F = ferruginous hawk, G = golden eagle). 6 �103 The spatial context in which disturbance occurs can influence the response shown by wildlife (Knight and Cole. 1991). I found that rough-legged hawks perched nearer to the road flushed at greater distances during walking disturbances than individuals perched further away~ Similarly, American kestrels, prairie falcons, and ferruginous hawks perched closer to the ground flushed at greater distances than those perched higher. This suggests that tolerance to disturbance decreases when the stimulus is closer and, thus, more focused. Other studies of raptors support this pattern (Russell 1980, Skagen 1980, Knight and Knight 1984). For merlins and prairie falcons, walking disturbances on paved roads resulted in shorter flight distances than a similar disturbance on gravel roads. Presumably, raptors perched along paved roads have habituated to the greater traffic volume associated with paved roads. Alternatively, individuals with greater tolerance limits to disturbance may be using areas with greater disturbance levels (Fraser 1983, Fraser et ale 1985, Buehler et ale 1991, MCGariga1 et ale 1991). Wintering bald eagles showed lower flushing responses along rivers and estuaries with high levels of recreational boating activity than along adjacent areas with little boating activity (Knight and Knight 1984, Buehler et ale 1991). Passerines in rural areas are less approachable than those in suburban areas where human activity is ubiquitious (Cooke 1980, Knight 1984, Knight et ale 1987, Kenney and Knight 1992). In the absence of persecution, it is probably adaptive for wildlife to habituate to frequent potentially-disruptive human activities, as this allows individuals adequate time to perform necessary biological functions (Knight et ale 1987). Flushing response and flush distance of golden eagles did not differ between adults and immatures. Stalmaster and Newman (1978) reported that adult bald eagles were more intolerant of disturbance, but other studies have not reported differences attributed to age (Russell 1980, Knight and Knight 1984, Buehler et ale 1991). Dark-morph rough-legged and ferruginous hawks exposed to walk disturbances flushed more often and at greater distances than light morphs. Because dark-morph raptors are uncommon in Colorado, they may be more conspicuous; consequently, they may experience more intentional (i.e., more focused) human disturbances. Dark-morph buteos can be difficult to identify except at close range; therefore, birdwatchers may flush these raptors while trying to get a better view, or, as raptor identification manuals suggest, they may intentionally flush the· birds in order to view underwing diagnostic characteristics (Clark and Wheeler 1987:75, Dunne et ale 1988:16, Johnsgard 1990:255). If this is the case, dark-morphs may be more likely to become sensitized to disturbance. For walking disturbance, flush distances increased with rapt or body size. Cooke (1980) reported a similar relationship in passerines, and suggested that larger birds were more wary of humans because of greater human persecution of larger, more visible species. Alternatively, the flush distance/body weight relationship may be based on the differing energetics of large versus small raptors. Smaller raptors have greater surface area to body weight ratios, therefore they expend relatively more energy than larger raptors (Hayes and Gessaman 1980, Koplin.et al. 1980, Wasser 1986). Disturbance results in increased energy expenditure caused by avoidance flights and a decreased energy intake caused by shortened foraging and feeding times (Stalmaster 1983). Small raptors, therefore, may be more energetically stressed if they are repeatedly forced to expend energy in avoidance flights. These birds might be expected to show a greater tolerance to disturbance in order to minimize energy expenditures. The lack of a flush distance/body weight relationship for vehicle disturbances may indicate that individual variation in flush distance is greater for less focused forms of disturbance. For example, the ferruginous hawk was the species. least likely to flush due to vehicle disturbance, but individuals that flushed did so at great distances from the approaching vehicle. Perhaps there were 2 behavioral groups within the species as Klein (1993) proposed for great egrets (Casmerodius albus), green-backed night 7 �104 (Butorides striatus), and yellow-crowned night herons (Nycticorax violaceus). Individual variation may explain the lack of meaningful herons regression models for flushing response and flush distance of raptors exposed to vehicle disturbances. Alternatively, other, unmeasured variables may be important, which also may explain the low B2 values of flush distance models and the lack of meaningful flushing response models for walking disturbances. For example, I measured no variables which explained why the 3 falcon species' flush distances were similar regardless of disturbance type. whereas flush distances for the hawks and the eagle correlated with disturbance type. Human behavioral restrictions can be used to allow coexistance of wildlife and human activities (Knight and Cole 1991). However, because raptor response to disturbance varies among species and between populations, management plans should be tailored to each species, habitat, and season. Furthermore, because humans in vehicles are less disruptive to raptors than pedestrians, management plans can offer different restrictions based on disturbance type. Spatial buffer zones are commonly used to protect nesting sites from disturbance (Knight and Skagen 1988); however, buffer zones for wintering raptors also could be effective if placed around sensitive foraging areas. From this study, the minimum radii of circular buffer zones that would prevent flushing by approximately 90% of the wintering individuals of a species are: American kestrel, 75 m; merlin, 125 m; prairie falcon, 160 m; rough-legged hawk, 210 m; ferruginous hawk, 140 m; and golden eagle, 300 m (Fig. 1-2). In addition, this study suggests that, for some species, horizontal spatial restrictions can be shortened if perching sites with greater vertical height are made available. Buffer zone recommendations could serve as guidelines that landmanagement agencies could present to birdwatchers, recreationists, and others who visit areas with sensitive species. Although my buffer zone recommendations are for use by wildlife managers working with the same raptor species in similar habitats (e.g., prairies, rangelands, or agricultural areas), my protocol could be used to collect data for most other species and habitats. LITERATURE CITED Buehler, D. A., T. J. Mersmann, J. D. Fraser, and J. K. D. Seegar. 1991. Effects of human activity on bald eagle distribution on the northern Chesapeake Bay. Journal of Wildlife Management 55:282-290. Clark, W. S., and B. K. Wheeler. 1987. A field guide to hawks of North America. Houghton Mifflin Co., Boston, Mass. 198pp. Cooke, .A. S. 1980. Observations on how close certain passerine species will tolerate an approaching human in rural and suburban areas. Biological Conservation 18:85-88. Dunne, P., D. Sibley, and C. Sutton. 1988. Hawks in flight. Houghton Mifflin Co., Boston, Mass. 254pp. Fraser, J. D. 1983. The impact of human activities on bald eagle populations--a review. Pages 68-84 in J. M. Gerrard and T. M. Ingram, eds. The bald eagle in Canada. proceedings of Bald Eagle Days, Winnepeg, Manitoba. Fraser, J. D., L. D. Frenzel, and J. E. Mathisen. 1985. The impact of human activities on breeding bald eagles in northcentral Minnesota. Journal of Wildlife Management 49:585-592. Gutzwiller, K. J. 1991. Assessing recreational impacts on wildlife: the value and design of experiments. Transactions of the North American Wildlife and Natural Resource Conference 56:248-255. Hayes, S. R., and J. A. Gessaman. 1980. The combined effects of air· temperature, wind and radiation on the resting metabolism of avian raptors. Journal of Thermal ~.ic)loCJY 5:119~125. 8 �Jameson, D. A., and R. E. Bement. 1969. General description of the Pawnee Site. u.s. IBP Grassland Biome Technical Report 1. Natural Resource Ecology Laboratory, Colorado State University, Fort Collins. 24pp. Johnsgard, P. A. 1990. Hawks, eagles, and falcons of North America. smithsonian Institution Press, Washington, D.C. 404pp. Kenney, S. P., and R. L. Knight. 1992. Flight distances of black-billed magpies in different regimes of human density and persecution. Condor 94:545-547. Klein, M. L. 1993. Waterbird behavioral responses to human disturbances. Wildlife Society Bulletin 21:31-39.· . Knight, R. L. 1984. Responses of nesting ravens to people in areas of different human densities. Condor 86:345-346. Knight, R. L., and D. N. Cole. 1991. Effects of recreational activity on wildlife in wildlands. Transactions of the North American Wildlife and Natural Resources Conference 56:238-247. Knight, R. L., and S. K~ Knight. 1984. Responses of wintering bald eagles to boating activity. Journal of Wildlife Management 48:999-1,004. Knight, R. L., and S. K. Skagen. 1988. Effects of recreational disturbance on birds of prey: a review. Pages 355-359 in R. L. Glinski, B. G. Pendleton, M. B. Moss, M. N. LeFranc, Jr., B. A. Millsap, and S. W. Hoffman, eds. Proceedings of the Southwest Raptor Management Symposium and Workshop. National Wildlife Federation, Washington, D.C. Knight, R. L., D •.J. Grout, and S. A. Temple. 198i. Nest-defense behavior of American crows in urban and rural areas. Condor 89:175-177. Knight, R. L., D. E. Andersen, M. J. Bechard, and N. V. Marr. 1989. Geographic variation in nest-defence behaviour of the red-tailed hawk Bu~eo jamaicensis. Ibis 131:22-26. Koplin, J. R., M. W. Collopy, A. R. Bammann, and H. Levenson. 1980. Energetics' of two wintering raptors. Auk 97:795-806. McGargial, K., R. G. Anthony, and F. B. Isaacs. 1991. Interactions of humans and. bald eagles on the Columbia River estuary. Wildlife Monographs 115. 47pp. Montgomery, D. C. 1991. Design and analysis of experiments. Third ed. John Wiley and Sons, Inc., New York, N.Y. 649pp. Olendorff, R. R. 1973. The ecology of the nesting birds of prey of northeastern Colorado. U.S. IBP Grassland Biome Technical Report 211. Natural Resource Ecology Laboratory, Colorado State University, Fort Collins. 223pp. Russell, D. 1980. Occurrence and human disturbance sensitivity of wintering bald eagles on the Sauk and Suiattle Rivers, Washington. Pages 165-174 in R. L. Knight, G. T. Allen, M. V. Stalmaster, and C. W. Servheen, eds. Proceedings of the Washington bald eagle symposium. The Nature Conservancy, Seattle, Washington. SAS Institute Inc. 1988. SAS/STAT user's guide, release 6.03 ed. SAS Institute Inc., Cary, N.C. 1,028pp. Skagen, S. K. 1980. Behavioral responses of wintering bald eagles to human activity on the Skagit River, Washington. Pages 231-241 in R. L. Knight, G. T. Allen, M. V. Stalmaster, and C. W. Servheen, eds. Proceedings of the Washington bald eagle symposium. The Nature Conservancy, Seattle, Washington. Skagen, S. K., R. L. Knight, and G. H. Orians. 1991. Human disturbance of an avian scavenging guild. Ecological Applications 1:215-225. Stalmaster, M. V. 1983. An energetics simulation model for managing wintering bald eagles. Journal of Wildlife Management 47:349-359. Stalmaster, M. V., and J. R. Newman. 1978. Behavioral responses of wiptering bald eagles to human activity. Journal of Wildlife Management 42:506513. Wasser, J. S. 1986. The relationship of energetics of falconiform birds to body mass and climate. Condor 88:57-62. ;.:;' . '~'.~ , ".' '.:.. 9 �106 CHAPTER 2 RESPONSES OF 3 SYMPATRIC BUTEOS TO DISTURBANCE DURING THE BREEDING SEASON INTRODUCTION With multiple-use of public lands, and increasing levels of outdoor recreation, areas must exist where wildlife can rest, reproduce, and forage undisturbed. For managers, this may mean limiting the times or areas in which human act~vities can occur by imposing temporal or spatial buffer zones (Knight and Temple 1994). To ensure persistence of sensitive populations, we need to understand potentially harmful characteristics of different disturbance types, and the reactions of species assemblages to these disturbance types (Klein 1993). No disturbance-related studies on nesting birds have simultaneously manipulated disturbance frequency, interval, and duration. Often, researchers repeatedly visited nesting birds to determine whether habituation occurs (Grier 1969, Fraser et ale 1985, White and Thurow 1985, Andersen et ale 1989, Andersen 1990, Ferrer et ale 1990). In some studies, repeated disturbances varied in duration and interval, but these factors were not manipulated because disturbances were observed and not experimental (Keller 1989, Fernandez and Azkona 1993, Klein 1993); however, such information is useful for building models and developing management scenarios (Grubb and King 1991). I conducted an experiment to examine the behavior of nesting redtailed hawks (Bu~eo jamaicensis), Swainson's hawks (B. swansoni), and ferruginous hawks (B. regal is) exposed to disturbances of varying frequency, duration, and interval. I compared the flushing behavior between species and between individuals in incubation and nestling phases. From this information, I calculated spatial buffer zones and determined how aspects of disturbance and habitat influence behavior. METHODS Study Area I conducted my study on a 5,870 km2 portion of northern Weld County, Colorado. Ranches and farms comprise 4,760 km2 of this area; farms con~ist primarily of alfalfa and winter wheat. The remaining 1,110 km2 consists of state-owned land (265 ~), the Central Plains Experimental Range (65 km2), and the Pawnee National Grassland (PNG) (780 km2). The PNG is managed by the U. S. Forest Service and is shortgrass prairie consisting primarily of blue grama (Bou~eloua gracilis) and buffalo grass (Buchloe dac~yloides). I selected equal numbers of nests from public and private lands for all species except red-tailed hawks, as these nests were associated with riparian areas found mostly on private lands. Ranching, oil exploration, and wildlifewatching are common activities throughout the region. This area is described by Jameson and Bement (1969) and Olendorff (1973). Study Design I collected data between 1 June - 18 July 1992 and between 24 April - 6 July 1993. Before collecting data, I searched the study area to locate active nest sites. I located approximately 66% of all active Swainson's hawk nests (n = 100) and 90% of all active red-tailed (n = 23) and ferruginous (n = 31) hawk nests. Experimental nests were randomly selected from this pool. I used different nests each year to avoid resampling of adults. A nest was an experimental unit. I manipulated 3 experimental variables: 1) duration of nest visit, 2) interval between nest visits, and 3) frequency of nest visits. I defined duration as the amount of time the researcher spent at the base of the nest tree, and interval as the amount of time allowed to pass between the parent bird' s return to the nest tree ... after a nest visit and ... the. initiation of the next visit. During the interval, the researcher was hidden from the parents' view. The frequency of nest visits was defined as the number of 10 �107 successive per nest. visits made to each nest tree, and was held constant at 4 visits The treatment structure was a 22 factorial, with nests being randomly assigned to 1 of 4 disturbance treatments: 5 minute duration, 10 minute interval; 5 minute duration, 30 minute interval; 15 minute duration, 10 minute interval; or 15 minute duration, 30 minute interval. I blocked data collection by reproductive phase (i.e., incubation or nestling) and, in 1992, performed 2 replications of each treatment on incubating and nestling phase Swainson's hawks and nestling phase red-tailed hawks. In 1993, I performed 2 replications of each treatment per phase on Swainson' s, red.-tailed, and ferruginous hawks. I completed the experiment for each species/reproductive phase combination within one week to increase the likelihood that the nests were at the same stage in the reproductive phase; this lessened the chance that treatment effects were confounded with variation in nest defense resulting from variation in age of nest contents. I wore the same clothing for all disturbances to ensure treatment effects were not confounded with reactions due to clothing color. Furthermore, to reduce the probability of desertion, each nest was disturbed within one day. Data Collection Within 300+ m of the nest, I used binoculars to determine mate presence or absence and the best approach to the nest for maximum visibility of me to the:raptor. Then, I started a stopwatch and walked quietly toward the nest at constant speed, keeping eye contact with the nest. When a parent first called, I recorded the time, placed a numbered flag at my position, and continued walking. As the parent flushed from the nest, I recorded the time, noted whether the bird flew toward or away from me, placed a numbered flag at my position, and continued walking to the base of the nest tree. Upon reaching the tree, I recorded the number of calls and dives performed by each parent in the initial 3 minutes. Dives were defined as breaks in horizontal flight directed toward the researcher; the closest approach of the diving bird to the researcher was estimated to the nearest 1 m. Temperature, wind speed, and cloud cover data were recorded while waiting for the 5 or 15 minute duration to elapse. Following a nest visit, I withdrew to the vehicle and recorded the time for a parent to return to the nest tree. Following this, I reset the stopwatch and waited for the 10 or 30 minute interval to elapse before repeating the disturbance protocol. After performing 4 nest visits, I used a Leitz rangefinder to measure the distances from the nest to the numbered flags representing the first-call and flush distances from each disturbance. The height of the nest above the ground was estimated with a Suunto clinometer. In addition, I noted the nearest permanent source of human activity (e.g., occupied house, oil well, road, etc.) and measured its distance from the nest. I collected data during mid-incubation .and midnestling periods to avoid uncertainties as to whether nest contents were eggs or chicks; however, if I was unsure, I checked the nest contents postdisturbance. Lastly, I checked nests for activity 24 to 48 hours postdisturbance. Data Analysis To screen the data for normality, I averaged each response variable (flush distance, first-call distance, call rate, return-to-nest time) over the 4 visits and examined residuals and normal probability plots for each variable from all species/reproductive phase combinations. Then, I tested for a year effect in the flush distances, first-call distances, call rates (calls/3 minutes), and return~to-nest times of incubation and nestling phase Swainson's hawks and nestling phase red-tailed hawks using a univariate t-test. I tested the response variables from each of the 4 visits in separate t-tests. In Table.2:-1,;I report only.:the fourth visit t-test. for brevity; however, regardless of visit, no variables were significantly different between years, so I pooled the data. Furthermore, since fewer than 3% of the pairs within 11 �108 Results of tests for year effects between 1992 and 1993 response Table 2-1. variables from the fourth disturbance of nesting raptors in northern Colorado. Species Phase swainson's incubation nestling red-tailed nestling t df P flush distance call distance call rate return time flush distance call distance call rate return time 1.25 0.18 0.98 1.01 .1.02 1.05 1.81. 0.55 14 14 14 14 14 14 14 14 0.2311 0.8634 0.3463 0.3307 0.3286 0.3114 0.0920 0.5896 flush distance call distance call· rate return time 0.69 0.45 0.70 1.29 14 14 14 14 0.5029 0.6620 0.4940 0.2190 Response Variable any species dove at me, both dive rate (dives/3 minutes) distance (m) were excluded from all analyses. and nearest dive I pooled data by species and performed 4 multiple linear regressions per species (i.e., one regression for each of the dependent variables). I also performed 1 logistic regression per species using flush response (1 = toward researcher, 2 = away) as the dependent variable. All regressions initially included date, wind speed, temperature, cloud cover, mate presence/absence, reproductive phase, nest height, type of nearest permanent human activity, distance to nearest permanent human activity, nest visit duration, and interval between visits as independent variables. Backward elimination was. used to remove insignificant (P > 0.05) variables. Remaining variables were tested for collinearities. I used R2 values from linear regressions and Gooctman-Kruskal Gamma (GKG) values from logistic regressions to assess the predictive power of each model. To ascertain the within-pair effects of nest-visit frequency and factor interactions, I ran repeated measures analyses of covariance (RMS-ANCOVA) for flush distances, first-call distances, call rates, return times, and flush responses. I used reproductive phase as a blocking factor, pooled data by species, included all significant variables except MATE from the regression analyses (Table 2-2), and also included frequency of visits as the repeated Table 2-2. Independent variables selected in multiple linear regression and logistic regression analyses for flush distances, first call distances, call rates,· return times, and flushing responses of nesting raptors subjected to repeated disturbances in northern Colorado during the summers of 1992 and 1993. Variable symbol Variable description and coding INT DUR PHASE Time between successive nest visits 1 = 10 minutes 2 = 30 minutes Length of time of nest visit 1 = 5 minutes 2 = 15 minutes Stage in reproductive cycle 1 = incubation 2 nestling Mate location 1 = present 2 = absent Height of nest above ground (m) Distance from nest to nearest permanent. human. activity (m) = MATE NESTHT DHUM 12 �109 measure. For each species, I performed 2 repeated measures analyses of covariance, 1 using nest height (NESTHT) and the other using distance to nearest permanent human activity (DHUM) as the covariate. If neither variable was a covariate (P > 0.05), I performed a repeated measures analysis of variance without NESTHT or DHUM; all other independent variables and interactions remained. I considered independent variables and interactions, except those containing frequency, significant if tests of between subjects effects P < 0.05; frequency and its interactions were significant if tests of within subjects effects P < 0.05. By pooling all data, I was able to make species comparisons of mean flush distances, first-call distances, call rates, and return-to-nest times using Tukey's HSD tests and species comparisons of flushing responses using chi-square tests. SAS procedures FREQ, GLM, LOGISTIC, and REG were used to analyze the data (SAS Inst. Inc. 1988: 544, 623, 195,864). RESULTS I visited 16 Swainson's, 8 red-tailed, and 8 ferruginous hawk nests during incubation and 16 Swainson's, 16 red-tailed, and 8 ferruginous hawk nests during the nestling phase; thus, the between-pair analyses were based on 32, 24, and 16 pairs of Swainson's, red-tailed, and ferruginous hawks, respectively. Since I approached each nest 4 times, the within-pair analyses were based on 128, 96, and 64 observations for Swainson's, red-tailed, and ferruginous hawks, respectively. For each species, regression and RMS-ANCQVA analyses indicated that date, wind speed, cloud cover, temperature, and type of nearest permanent disturbance were insignificant (P > 0.05) explanatory variables for flushing responses, flush distances, first-call distances, call rates, and return-to-nest times. Between-Species Comparisons Flushing Response. -- During incubation, the mean percentages (averaged over 4 visits) of individuals in each. species which did not flush, flushed away from, or flushed toward the researcher were: 0%, 97%, and 3% for red-tailed hawks; 31%, 50%, and 19% for Swainson's hawks; and 0%, 72%, and 28% for ferruginous hawks. During the nestling period, the mean percentages of individuals in each species which did not flush, flushed away from, or flushed toward the researcher were: 0%, 33%, and 67% for red-tailed hawks; 11%, 42%, and 47% for Swainson's hawks; and 9%, 63%, and 28% for ferruginous hawks. For both reproductive phases, flushing responses significantly differed between species (X2 33.12 and 20.33, 4 df each, and both P < 0.0001, for incubation and nestling periods, respectively). = Distance. -- Regardless of reproductive phase, red-tailed and ferruginous hawks flushed at greater distances than Swainson's hawks (incubation: t = 3.81 and 2.68, 29 df each, P = 0.0007 and 0.0119; nestling: t = 2.45 and 3.09, 37 df each, P = 0.0192 and 0.0038, respectively, for redtailed and ferruginous hawks) (Figure 2-1). Red-tailed and ferruginous hawk flush distances did not significantly differ from one another in either the incubation or nestling phase (t = 1.40 and 1.64, 29 df and 37 df, P = 0.1709 and 0.1105, respectively). Red-tailed hawk first-call distances were greater than those of Swainson's and ferruginous hawks during incubation (t = 8.85 and 8.25, 29 df each, both P < 0.0001, respectively), and greater than those of Swainson's hawks during the nestling phase (t = 3.67, 37 df, P = 0.0008) (Figure 2-2). call Rates. -- Call rates of incubating red-tailed hawks were higher than incubating Swainson's and ferruginous hawks (t = 3.75 and 3.94, 29 df each, P = 0.0008 and 0.0005, respectively) (Figure 2-3). During the nestling phase, red-tailed hawks called at greater rates than ferruginous hawks (t = 3.56,·37 df, P = 0.0011), and both red-tailed and ferruginous hawks called at higher rates than Swainson's hawks (t = 3.42 and 2.29, 37 df each, P = 0.0016 and 0.0276, respectively). to-nest Return-to-nest Times. -- Regardless of reproductive phase, returntimes did not differ between Swainson's and ferruginous hawks, 13· �no 400 ~ INCUBATING IINESTLING ~ 300 z ~ (f) 0 200 I (f) ::> _J LL Z « w 100 ~ o SWA FER RTH SPECIES Figure 2-1. Mean flush distances (m) and standard errors incubating and nestling phase buteos in northern Colorado of 1992 and 1993. . ..--.. 400 .-----------------~ species for the summers ~ INCUBATING IINESTLING --~ ill o z 300 versus during 1------------ ~ (f) o _J ~ 200 o ~ (f) 0::: LL 100 1-----1--- Z « W ~ o SWA· FER RTH SPECIES Figure 2-2. Mean first-call distances (m) and standard errors versus species for incubating and nestling phase buteos in northern Colorado during the summers of 1992 and 1993. 14 �111 ----:--25 [Ii INCUBATING Z ~ - • NESTLING 20 1---------,----- 15 1-------- C'0 (f) _j _j <3 w ~ 10 1------- _j _j <{ o z « w ~ FER RTH SPECIES Figure 2-3. Mean call rates (calls/3 min.) and standard errors versus species for incubating and nestling phase buteos in northern Colorado during the summers of 1992 and 1993. Swainson's and red-tailed hawks, or ferruginous and red-tailed hawks (incubation: t = 1.40, 1.72, and 1.37, 29 df each, P = 0.1709, 0.0957, and 0.1721; nestling: t = 1.59, 1.01, and 0.78, 37 df each, P = 0.1194, 0.3181, and 0.4386) (Figure 2-4). 40 ~----------------------------------~ r---------------, • INCUBATING .• NESTLING z ~30 w ~ I- ~ 20 :::J rW 0:: ~ 10 w ~ o SWA FER RTH SPECIES Figure 2-4. Mean return times (min.) and standard errors incubating and nestling phase buteos in northern Colorado of 1992 and 1993. 15 versus during species for the summers �112 Buffer Zones. -~ Spatial buffer zones were calculated as the radius of restricted activity around a nest which would prevent flushing by approximately 90% of the individuals within a species. I used flush distance data to calculate the following buffer zone radii for incubation and nestling periods, respectively: Swainson's hawk, 171 m and 316 m; red-tailed hawk, 458 m and 510 m; and ferruginous hawk, 397 m and 640 m. Between-Pair Comparisons Flushing Response. -- Logistic regression models for flushing response contained explanator~ variables for Swainson's hawks (probability of . flushing away = 1 / [1 + e [3.+ 1.77PHASE+1. 85INT-O.65MATE-O.03DHUMI l . GKG = 0.489) and red-tailed hawks (probability of flushing away = 1 / [1 + e [-4.75+4.98PHASE-1.90MATEO.02DHUM]],GKG = 0.887). Individuals of both species had increased probabilities of flushing to my approach during mate absence (P = 0.0156 and 0.0001 for Swainson's and red-tailed hawks, respectively), during incubation (P = 0.0011 and 0.0001, respectively), and if nests were further from sources of human activity (P = 0.0151 and 0.0442, respectively). Distance. -- Although behavioral responses to visit duration and between-visit interval varied across species and by reproductive phase (Table 2-3), pooling data by reproductive phase and running multiple linear regressions (Table 2-4) indicated that duration and interval were significant explanatory variables. With long between-visit intervals (30 minutes), redtailed hawk flush distances increased (P = 0.0025), whereas ferruginous hawk flush distances decreased (P = 0.0001) (Figure 2-5). First~call distances of Swainson's and red-tailed hawks decreased for long (15 minutes) nest visit durations (P = 0.0003 and 0.0312, for Swainson's and red-tailed hawks, respectively) (Figure 2-6). Flush distances were greatest for Swainson's and Table 2-3. Mean flush distances (m) , first-call distances (m) , call rates (calls/3 min.), and return-to-nest times (min. ) of incubating and nestling phase swainson's, red-tailed, and ferruginous hawks subjected to 1 of 4 duration (min.)/interval (min.) treatments. The study was conducted in northern Colorado during the summers of 1992 and 1993. Treatment (duration/interval) (5/10) (5/30) (15/10) (15/30) Species n Phase Swainson's 16 inc_ 16 red-tailed 8 16 ferruginous 8 8 Variable flush distance first-call distance call rate return time nest. flush distance first-call distance call rate return time inc. flush distance first-call distance call rate return time nest. flush distance first-call distance call rate return'time inc. flush distance first-call distance call rate return time nest. flush distance first-call distance call rate return time 16 X SE X SE X 45 0 0 12 208 149 4 17 135 173 4 23 268 417 21 7 154 6 1 38 502 313 22 29 14 0 0 4 35 47 2 3 27· 47 1 4 41 36 3 1 63 6 1 15 56 120 9 6 92 15 4 22 169 137 7 4 247 242 15 31 313 346 21 8 75 0 0 8 270 259 6 18 16 6 2 5 39 39 2 1 65 62 2 5 45 31 10 0 0 12 97 49 4 12 158 154 13 21 178 303 19 6 206 0 0 41 493 183 26 7 3 2 23 0 0 2 49 62 2 1 SE 4 0 0 5 26 21 2 3 28 28 1 2 32 49 2 1 65 0 0 15 58 70 10 1 X 44 16 6 11 90 14 2 5 315 311 16 25 285 312 17 10 126 0 0 52 230 236 9 3 SE 16 10 3 5 28 6 1 1 42 41 3 3 50 47 2 3 34 0 0 15 87 89 4 1 �jJ 3 Table 2-4. Regression equations for 4 behavioral response variables of 3 nesting raptor species subjected to repeated disturbances in northern Colorado during the summers of 1992 and 1993. Species Dependent Swainson's hawk red-tailed hawk Variable n Equation flush distance 32 0.438 0.0001 153.60 - 78.03DUR +.81.04PHASE call distance 32 0.323 0.0001 92.18 - 60.49DUR + 72.93PHASE call rate 32 0.201 0.0001 8.08 + 3.11INT return time 32 0.120 0.0001 29.16 - 2.44NESTHT flush distance 24 0.126 0.0012 -30.29 + 97.59INT + 36.78MATE call distance 24 0.242 0·.0001 60.50 - 7l.43DUR + 144.80PHASE call rate 24 0.192 0.0002 21.33 + 6.68PHASE return time 24 0.581 0.0001 47.50 + 3.96INT 16 0.533 0.0001 302.99 + 265.81PHASE ferruginous hawk flush distance - 2.24~~TE - 15.43NESTHT - 13.31NESTHT + 58.06MATE - 2.76MATE + 0.99NESTHT - 18.76PHASE - 2.57MATE - O.OlDHUM - 191.29INT 16 0.677 0.0001 -163.31 + 230.12PHASE 16 0.460 0.0001 -0.45 + 14.48PHASE - 4.10MATE - 1.51NESTHT return time 16 0.356 0.0001 89.10 - 9.84MATE 4.45NESTHT - 0.02DHUM c - 82.22MATE - 74.27MATE call distance ..-~ 300 + O.llDHUM - 0.62NESTHT call rate 350 + 0.17DHOM + 0.22DHUM + O.OlDHOM III RTH -- liFER ill ~ 250 ~ (j) 200 V5 .150 LL 100 o ::J _J z « ill 2 50 o 30 10 INTERVAL (MIN) Figure 2-5. Mean flush distances (m) and standard errors versus between-visit intervals for red-tailed and ferruginous hawks in northern Colorado during the summers of 1992 and 1993. 17 �114 ..--.. 160 III SWA ~ ~140 liFER 0 z 120 ;:; (f) 0 _J _J « 0 100 80 I r 60 (f) n:: 40 LL z « w 20 ~ 0 5 15 DURATION (MIN) Figure 2-6. Mean first-call distances (m) and standard errors versus visit durations for Swainson's and ferruginous hawks in northern Colorado during the summers of 1992 and 1993. Incubation and nestling phase data are pooled. ferruginous hawks (both P = 0.0001) and first-call distances were greater for all species (all P = 0.0001) during the nestling phase. During mate absence, red-tailed hawks flushed and called at greater distances (P = 0.0433 and 0.0030, for flush and first-call distances, respectively), but ferruginous hawks flushed and called at shorter distances, than when a mate was present (P = 0.0015 and 0.0001, respectively). Swainson's hawks with nests closer to the ground began calling at greater distances than those at nests further above the ground (P = 0.0001). First-call distances were greatest for Swainson's and ferruginous hawks nesting further from sources of permanent human activity (P = 0.0009 and 0.0001, for Swainson's and ferruginous hawks, respectively). Call Rates. -- Call rates for all species were less when mates were absent (P = 0.0007, 0.0075, and 0.0268 for Swainson's, red-tailed, and ferruginous hawks, respectively) (Figure 2-7). At nests further above the ground, call rates of Swainson's and ferruginous hawks were greater (P = 0.0057 and 0.0029, respectively), but red-tailed hawk call rates were less (P = 0.0168). Red-tailed and ferruginous hawks called at greater rates during the nestling phase than during incubation (P = 0.0008 and 0.0001, respectively), whereas call rates of Swainson's hawks did not differ between periods (P = 0.3549) (Figure 2-8). Return-to-nest Times. -~ Red-tailed and ferruginous hawks exhibited similarities in return-to-nest times. Individuals of both species returned more quickly during mate absence (both P = 0.0081) and if nests were further from sources of permanent human activity (P ·0.0004 and 0.0080, respectively). Swainson's and ferruginous hawks returned more quickly if nests were further above the ground (both P = 0.0001), whereas red-tailed hawk return times were consistent over the range of nest heights (P = 0.1314). = 18 �us .-.20 z 18SWA IIRTH FmFER [2J ~ ('() (f) __J __J 15 « o ....._.. w 10 ~ n:: __J __J « o 5 z « w ~ o PRESENT ABSENT MATE LOCATION Figure 2-7. Mean call rate (calls/3 min.) and standard errors versus mate location for Swainson's, red-tailed; and ferruginous hawks in northern Colorado during the summers of 1992 and 1993. Incubation and nestling phase data are pooled. _25 z IIlSWA .RTH ETI l:illj FER ~ ~ 20 (f) __J __J c3 15 ~ . ~ 10 __J __J c;( o z 5 « w ~ o ·INC NEST REPRODUCTIVE PHASE Figure 2-8. Mean call rates (calls/3 min.) and standard errors versus reproductive phase (incubation or nestling) for Swainson's, red-tailed, and ferruginous hawks in northern Colorado during the summers of 1992 and 1993. Incubation and nestling phase data are pooled. 19 �ll6 Within-Pair Comparisons The Role of Nest Visit Frequency. -- Nest visit frequency and interactions between frequency and other independent variables were significant explanatory variables for flush distances, first-call distances, call rates, and return-to-nest times, but no consistent patterns existed across all species (Table 2-5). As the number of nest visits increased, ferruginous hawk flush distances increased (~310 = 3.90, P = 0.0182) and redtailed hawk call rates decreased (~3,18= 3.13, 'P = 0.0332) (Figure 2-9). Although each species showed a frequency and reproductive phase interaction, the species differed in the affected behavior. As number of visits increased, Swainson's hawks maintained constant first-call distances during incubation, but increased first-call distances during the nestling phase (~3,26= 3.05, P = 0.0319) (Figure 2-10). With repeated visits, red-tailed hawk call rates remained constant during incubation, but decreased during the nestling phase (~3,18= 3.75, P = 0.0160) and ferruginous hawk flush distances increased during incubation but remained constant during the nestling phase (~3,10= 3.11, P = 0.0412). Ferruginous hawk flush distances increased with repeated visits if intervals between visits were short, but remained constant if intervals were long (~3,10= 5.27, P = 0.0048) (Figure 2-11). Similarly, redtailed hawk return-to-nest times increased if repeated visits were separated by short intervals, but remained constant for long intervals (~3,18= 3.24, P = 0.0291). When repeated visits were of long duration, red-tailed hawk returnto-nest times decreased, but' return-to-nest times increased if visit durations were short (~3,18= 4.80, P = 0.0049) (Figure 2-12). Table 2-5. Mean flush distances (m) , first-call distances (m) , call rates (callsj3 min.), and return-to-nest times (min. ) of repeatedly visited incubating and nestling phase Swainson's, red-tailed " and ferruginous hawks. This study was conducted in northern Colorado during the summers of 1992 and 1993. visit 1 Species n Phase Swainson's 16 inc. 16 red-tailed 8 16 ferruginous 8 8 Variable X flush distance first-call distance call rate return time nest. flush distance first-call distance call rate return time inc. flush distance first-call distance call rate return time nest. flush distance first-call distance call rate return time inc. flush distance first-call distance call rate return time nest. flush distance first-call distance call rate return time 62 14 2 15 114 39 3 7 173 188 14 25 260 352 28 9 83 0 0 46 339 211 12 14 20 visit 2 SE SE X 19 9 1 4 34 21 2 1 22 26 3 4 46 47 3 3 21 0 50 4 3 15 148 94 4 9 232 242 12 24 263 314 19 7 169 0 1 28 348 243 16 16 19 70 82 6 5 0 16 4 2 5 31 33 2 2 58 59 2 4 38 42 3 2 59 0 1 10 74 85 7 7 visit 3 X SE 41 13 5 4 2 2 15 4 156 36 106 40 2 5 11 3 233 57 234 57 11 2 26 3 258 42 358 38 17 2 7 2 176 56 6 6 0 0 3;1.12 383 80 261 92 16 8· 4 13 visit 4 SE X 39 9 2 11 147 110 4 11 217 216 10 25 263 354 14 8 133 0 0 33 426 275 19 14 14 6 1 4 37 42 2 3 55 56 2 4 50 44 2 2 59 0 0 14 88 97 9 ·4 �117 ___26 z - .-'~. ~ 24 - - (Y) (f) 22 - _j _J __. u ~ 20 w 18 - ~ IX: _J _J 16 - d 14 II t - z J f 3 4 ~ 12 w ~ 10 _l 0 I 2 1 I 5 NEST VISIT Figure 2-9. Mean call rates (calls/3 min.) and standard errors versus nest visit for red-tailed hawks in northern Colorado during the summers of 1992 and 1993. Incubation and nestling phase data are pooled. 120 INC __. ~ 100 • .....••...... - NEST w u Z • 80- ~ (f) o 60 _J _J ~ U I t- 40 - .- (f) n:: 20 LL o I. + I 1 2 .3 4 ·NESTVISIT Figure 2-10. Mean first-call distances visit for incubating and nestling phase during the summers of 1992 and 1993. 21 (m) and standard errors versus nest Swainson's hawks in northern Colorado �118 450 ..-.. ~ ..._ 400 - w l- ~ 350 ~ 300 - ~ 250 - 200 - ~ 150 - (f) o z t t 10 MINUTE T 30 MINUTE • + ~ ::J _j LL t + I <{ + I I 2 3 NEST VISIT 4 I I 100 1 t.. Figure 2-11. Mean flush distances (m) and standard errors versus nest visit for ferruginous hawks subjected to between-visit intervals of 10 and 30 minutes. Data were collected in northern Colorado during the summers of 1992 and 1993. Incubation and nestling phase data are pooled. 17 "-"16 z . ~ ..._ w 15 - t + -r 5 MINUTE • 15 MINUTE + ~ r 14 I- Z 0:: ::J 13 r w 0::: z 12 - <{ , w + + t I I I 2 3 NEST VISIT 4 ~ 11 10 I 1 t Figure 2-12. Mean return times (min.) and standard errors versus nest visit for red-tailed hawks subjected to visits of 5 and 15 minute durations. Data were collected in northern Colorado during the summers of 1992 and 1993. Incubation and nestling phase data are pooled. 22 �119 DISCUSSION Disturbance Duration, Interval, and Frequency I found that long durations of disturbance seemed to increase tolerance: Swainson's and red-tailed hawk call distances were shortest for long durations and, over repeated visits, red-tailed hawk return-to-nest times remained constant for short durations but decreased for long durations. Since I performed each disturbance, I caution that the birds' behaviors may have been shaped by repeated exposure to me. Male red-winged blackbirds approached closer and responded more aggressively to familiar humans than to unfamiliar humans (Knight and Temple 1986a). Furthermore, even though I did not find a significant interaction between duration and interval, tolerance may decrease if disturbances are infrequent. Grubb and King (1991) found that duration. was the second-most important variable in determining whether pedestrian, aquatic, vehicle, and aircraft activities disturbed bald eagles (Haliaeetus leucocephalus). Red-tailed and ferruginous hawk flush distances were shortest when the interval between disturbances was greatest and, over repeated visits, ferruginous hawk flush distances increased for short intervals but remained constant for long intervals. I suggest that 30-minute intervals allowed agitation levels to subside, whereas 10-minute intervals did not; ferruginous hawk avoidance behavior may have increased because birds were still agitated from the previous visit. Over repeated visits, red-tailed hawk return-to-nest times increased for short intervals but remained constant for· long intervals, further indication that agitation levels may vary with interval length. Frequency of disturbing events was relatively unimportant; however, interactions between frequency and other disturbance characteristics had greater explanatory power than frequency alone. My results may have been different, however, had visits occurred over days or weeks instead of within 1 day. As number of visits increased, red-tailed hawk call rates decreased and ferruginous hawk flush distances increased, suggesting that tolerance to disturbance may have increased for .red-tailed hawks and decreased for ferruginous hawks. Alternatively, red-tailed hawk call rates may have decreased because calls were ineffective in driving off the predator (Patterson et al. 1980). White and Thurow (1985) found that ferruginous hawk flush distances increased over repeated visits during incubation. Flush distances for nesting bald eagles also increased with successive disturbances (Fraser et al. 1985). Nest Characteristics and Mate Location Wildlife responses to disturbance may vary depending upon habitat characteristics and the spatial context of the disturbance (Knight and Skagen 1988, Knight and Cole 1991). Where green-backed herons (Butorides striatus) could hear but not see boaters, they did not alter their foraging behavior (Kaiser and Fritzell 1984). Perched bald eagles flushed at shorter distances when vegetation concealed human activity (Stalmaster and Newman 1978), but .preferred to feed further from vegetatio_D. in high humarr-use .,areas (Skagen et al. 1991). Bird species are_ less likely to flush if approached by humans in vehicles rather than humans on foot (Skagen 1980, Klein 1993, Holmes et al. 1993). Raptors perched nearer to an approaching disturbance flushed at greater distances than birds perched further away, suggesting that spatial context can influence the perceived intensity of a disturbance (Holmes et al. 1993). This study indicates that, regardless of species, most reactions to disturbance increased as distance to nest contents (i.e.; nest height) and as likelihood of previous disturbance (i.e., distance to nearest human activity) decreased. Swainson's and ferruginous hawks which nested closer to the ground called at greater distances, had increased call rates, and had greater return times than individuals nesting higher above the ground, indicating that tolerance decreases as disturbance becomes more focused. Red-tailed hawks 23 �120 were the exception, because first-call distances were greater at nests further above the ground. This may have resulted from increased visibility of the disturbance. Unlike Swainson's and ferruginous hawks, which nested in open portions of lone trees and could see a human approaching from any direction, red-tailed hawks nested in the canopies of trees located within large groves and, thus, may have visually benefited from greater nest height. At nests closer to permanent sources of human activity, red-tailed hawks called at greater distances and red-tailed and Swainson's hawks were more likely to flush toward the intruder. These results suggest that these individuals are more sensitive than those nesting further from human activity, and support findings from other studies which show that birds with higher rates of neutral encounters with humans defend more vigorously than persecuted birds (Knight 1984, Knight et al. 1987, Kenney and Knight 1992). However, red-tailed and ferruginous hawks returned less quickly if nesting closer to human activity, indicating that human encounter rate and past persecution may not be the only important considerations. For example, birds may defend vigorously until they realize their actions are ineffective, and then may choose in favor of their own safety. Those with conspicuous· and easily accessible nests (Le., Swainson's and ferruginous hawks) may not call immediately, so as not to. attract attention, but may wait until the predator has moved beyond a threshold of safety. Alternatively, birds may vary behavior to prevent predator habituation (Winkler 1992). Regardless of species, call rates increased when the mate was present, perhaps serving to warn the mate (Patterson et al. 1980) or to rally the mate to help in defending the nest (Knight and Temple 198Gb). In this study, when the mate was absent, ferruginous hawk flush and call distances were shorter, those of red-tailed hawks were greater, and red-tailed and Swainson's hawks were more likely to flush away from the researcher. Such reversals in behavior may depend on which parent is present at the nest and whether one sex defends more aggressively than the other; however, I did not examine this factor. Defense intensity may vary within a sex depending upon habitat and predator characteristics (Sproat and Ritchison 1993) or may vary between and within the sexes as a result of past persecution. Conceivably, persecution could select a.gainst female aggressiveness since females are most often at the nest and, therefore, most likely to be killed (Newton 1979, Knight et al. 1989). Contrary to my findings, Andersen (1990) found that the overall intensity of nest defense was not influenced by mate presence. Reproductive Phase Because behavior is dynamic, individual and population-level responses to disturbance may change over the life cycle, within a year, or during one reproductive period. This study found that red-tailed and Swainson's hawks were 22 and 2 times, respectively, more likely to flush toward the researcher during nestling phase than during incubation. These species also initiated defense sooner during nestling phase, as indicated by increased flush and call distances. Eastern screech owls (Otus asio) (Sproat and Ritchison 1993) and Spanish imperial eagles (Aquila adalberti) (Ferrer et al. 1990) defended more vigorously in the nestling period, whereas roughlegged hawks (S. lagopus) and merlins (Falco columbarius) did not (Andersson arid Wiklund ~987~ Wiklund 1990). Species variability was also evident in my study. -Unlike-the other 2 buteos, ferruginous hawks did not defend their nests; most individuals called and flushed away from me at increasing distances over the reproductive period. White and Thurow (1985) found avoidance behavior in ferruginous hawks subjected to repeated disturbances but defensive behavior in pairs visited only once. These results suggest that sensitization occurred as individuals gained more experience with the disturbance, supporting the hypothesis that birds may respond to familiar stimuli differently than to novel stimuli (Knight and Temple 198Ga). For ferruginous and red-tailed hawks, call rates were greater during the nestling phase, suggesting that calls function as warnings to the offspring. Call rates of American goldfinches (Carduelis tristis) were greater for the nestling phase, and calls silenced the young and caused them 24 �121 to crouch, thus making them less conspicuous (Knight and Temple 1986b). Patterson et ale (1980) documented similar results for the white-crowned sparrow (Zonotrichia leucophrys) and Andersen (1990) reported that red-tailed hawk call rates were positively correlated with nestling age. Ferruginous and red-tailed hawk return times were shorter during the nestling period. Holthuijzen et ale (1990) noted that, regardless of reproductive phase, most prairie falcons (F. mexicanus) returned to the nest within minutes after disturbance. Fernandez and Azkona (1993) reported similar findings for marsh harriers (Circus aeruginosus). White and Thurow (1985) documented a 33% desertion rate for incubating ferruginous hawks, and anecdotal evidence suggests that this species frequently deserts the nest due to human disturbance. Regardless of species, none of the pairs disturbed in my study deserted. However, 3 pairs of ferruginous hawks remained away from the nest for over 4 hours, and were not seen incubating until the nest was checked 3 days later. Palmer (1988) states that ferruginous hawks may be more likely to abandon the nest in years of food scarcity. Species Differences and Buffer Zones Although these 3 species are congeneric, differences in reaction to disturbance exist. Regardless of reproductive phase, ferruginous hawks exhibited low levels of nest defense. Ferruginous hawks have, on average, larger clutches and broods than any other North American buteo (Palmer 1988), and are also long-lived. This suggests that the species has higher future reproductive potential than its congeners, and models predict that, as residual reproductive value increases, nest·defense decreases (Redondo 1989). However, if current reproductive output is considered, one would predict that species with larger broods should defend more vigorously (Ricklefs 1977), which my findings contradict. Because of high adult survivorship and high probability of future reproduction, ferruginous hawks may show a reduced nest defense response, favoring enhanced survival and future reproductive opportunities over current reproductive attempts (Grier 1969, Buitron 1983, Curio et ale 1983, Wallin 1987, Andersen 1990). Studies on passerines (Cooke 1980) and wintering raptors (Skagen et a1. 1991, Holmes et ale 1993) have shown a positive linear relationship between body size and flush distance. This study, however, indicates that during incubation and nestling phases, the ferruginous hawk and the smaller red-tailed hawk flushed at similar distances, with both species flushing at greater distances than the smaller Swainson's hawk. Such differences may result from differing degrees of habituation. On average, Swainson's hawks nested closer to human activity (x = 227 m) than did ferruginous or red-tailed hawks (x = 538 m and 516 m, respectively), suggesting that Swainson's hawks are more tolerant of human activities. Grier (1969) suggested that nesting bald eagles accustomed to human activity may not be as disturbed by researchers climbing to the nests. Wintering bald eagles perched near popular boating areas were more tolerant of disturbance than those perched along rivers with less activity (Knight and Knight 1984) and urban magpies flushed bess often and at shorter distances than rural magpies (Kenney and Knight 1992). Habituation, in the absence of persecution, may be adaptive by decreasing energy expenditures and increasing parental attentiveness to eggs and young (Knight et ale 1987, Keller 1989). Habituation, however, is maladaptive if individuals are actively persecuted since it increases the likelihood that nest contents and adults will be destroyed (Knight 1984). Red-tailed hawks always called more than Swainson's or ferruginous hawks. Red-tailed hawks may call more because their nests are less conspicuous and calling may distract a potential predator (Knight and Temple 1988). Finding the active nest was made difficult by the calling pair, not only because they soared above the researcher instead of the nest, but also because calling silenced the chicks and caused them to crouch below the nest rim (pers. obs.). In contrast, ferruginous hawk chicks never crouched in 25 �122 response to a calling parent (pers. obs.). For species with highly visibl~ nests (Le., Swainson's and ferruginous hawks), calling may be detrimental since it alerts humans that the nest is active and may result in the destruction of the nest contents and/or the parents (Knight 1984). My buffer zones are consistently larger than .others reported for nesting buteos. For nesting ferruginous hawks, White and Thurow (1985) reported a 250 m radius, as compared to my radii of 397 m for incubating pairs and 640 m for nestling phase pairs. The buffer zone radius for wintering ferruginous hawks is even smaller, at 140 m (Holmes et ale 1993). Seasonal differences in requirements are also evident between wintering bald eagles, which need between 350 m and 450 m (Knight and Knight 1984) and breeding bald eagles, which need 500 m (Fraser et ale 1985). Wintering prairie falcons required a buffer zone radius of 160 m (Holmes et ale 1993), whereas breeding birds required 500 m for low level disturbances and 1000 m for long-term or intense disturbances (Holthuijzen et ale 1990). We.could not find buffer zone information for Swainson's or red-tailed hawks in the literature. Conclusion Public lands are subject to a myriad of commodity and amenity uses. The result is a growing concern among conservationists that wildlife sensitive to disturbance will experience significant population declines, either because of decreased fecundity and survival, or increased habitat alterations and dispersion (White and Thurow 1985, Andersen et ale 1990, Knight and Cole 1991). Insights into the responses of wildlife to experimental disturbances can assist in determining which components of disturbance most alter behavior, the potential consequences of such alterations, and viable mitigation techniques. For example, temporal buffer zones may be developed to control the time interval between successive disturbances or spatial buffer zones may be increased or decreased depending on the species and its current reproductive phase, thereby allowing human use. of an area while managing for viable wildlife populations. Such information is also vital in educating the public as to why access to an area is restricted. Due to between-population variation in response to disturbance, my results are most applicable to the Swainson's, red-tailed, and ferruginous hawk populations of northeastern Colorado. My protocol, however, is relatively simple, inexpensive, and could be tailored to most locations and habitat types. It seems prudent for those who manage sensitive populations on multiple-use lands to examine not only the characteristics of the population and habitat, but also those of the disturbances, before deciding what kind of information to gather and what type of mitigation technique to use. LITERATURE Andersen, CITED D. E. 1990. Nest-defence behavior of Red-tailed Hawks. Condor 92: 991-997. Andersen, D. E., O. J. Rongstad, and W. R. Mytton. 1989. Response of nesting Red-tailed Hawks to helicopter overflights. Condor 91: 296-299. Andersen, D. E., O. J. ·Rongstad, and W. R. Mytton. 1990~ Home range changes in raptors exposed to increased human activity levels in southeastern Colorado. Wildlife Society Bulletin 18: 134-142. Andersson, S., and C. Wiklund. 1987. Sex role partitioning during offspring protection in the Rough-legged Buzzard (Buteo lagopus). Ibis 129: 103-107. Cooke, A. S. 1980. Observations on how close certain passerine species will tolerate an approaching human in rural and suburban areas. Biological Conservation 18: 85-88. Curio, E., G. Klump, and K. Regelmann. 1983. An anti-predator response in the great tit (Parus major): Is it tuned to predator risk? Oecologia 60: 83-88. 26 �123 Fernandez, C., and P. Azkona. 1993. Human disturbance affects parental care of marsh harriers and nutritional status of nestlings. Journal of Wildlife Management 57: 602-608. Ferrer, M., L. Garcia, and R. Cadenas. 1990. Long-term changes in nest defense intensity of the Spanish Imperial Eagle, Aquila adalberti. Ardea 78: 395-398. Fraser, J. D., L. D. Frenzel, and J. E. Mathisen. 1985. The impact of human activities on breeding bald eagles in northcentral Minnesota. Journal of Wildlife Management 49: 585-592. Grier, J. W. 1969. Bald eagle behavior and productivity responses to climbing to nests. Journal of Wildlife Management 49: 585-592. Grubb, T. G., and R. M. King. 1991. Assessing human disturbance of breeding bald eagles with classification tree models. Journal of wildlife Management 55: 500-511. Holmes, T. L., R. L. Knight, L. Stegall, G. R. Craig. 1993. Responses of wintering grassland raptors to human disturbance. Wildlife Society Bulletin 21: 461-468. Holthuijzen, A. M. A., W. G. Eastland, A. R. Ansell, M. N. Kochert, R. D. Williams, and L. S. Young. 1990. Effects of blasting on behavior and productivity of nesting prairie falcons. Wildlife Society Bulletin 18: 270-281. Jameson, D. A., and R. E. Bement. 1969. G~neral description of the Pawnee Site. U. S. IBP Grassland Biome Technical Report 1, Natural Resource Ecology Laboratory, Colorado State University, Fort Collins. 24pp. Kaiser, M. S., and E. K. Fritzell. 1984. Effects of river recreationists on green-backed heron behavior. Journal of Wildlife Management 48: 561-566. Keller, V. 1989. ,Variations in the response of Great Crested Grebes Podiceps cristatus to human disturbance -- a sign of adaptation? Biological Conservation 49: 31-45. Kenney, S. P., and R. L. Knight. 1992. Flight distances of black-billed magpies in different regimes of human density and persecution. Condor 94: 545-547. Klein, M. L. 1993.' Waterbird behavioral responses to human disturbances. Wildlife Society Bulletin 21: 31-39., Knight, R. L. 1984. Responses of nesting ravens to people in areas of human different densities. Condor 86: 345-346. Knight, R. L., D. E. Andersen, M. J. Bechard, and N. V. Marr. 1989. Geographic variation in nest-defense behavior of the Red-tailed Hawk Buteo jamaicensis. Ibis 131: 22-26. Knight, R. L., and D. N. Cole. 1991. Effects of recreational activity on wildlife in wildlands. Transactions of the North American Wildlife and Natural Resource Conference 56: 238-247. Knight, R. L., D. J. Grout, and S. A. Temple. 1987. Nest-defense behavior of American crows in urban and rural areas. Condor 89: 175-177. Knight, R. L., and S. K. Knight. 1984. Responses of wintering bald eagles to boating activity. Journal of Wildlife Management 48: 999-1,004. Knight, .R. L., and S. K. Skagen. 1988. Effects of recreational disturbance on birds of prey: a review. Pages 355-35.9 in R. L. Glinski, B. G. Pendleton, M. B. Moss, M. N. LeFranc, Jr., B. A. Millsap, and S. W. Hoffman, eds. --"Proceedingsof the Southwest raptor management symposium and workshop. National Wildlife Federation, Washington, D. C. Knight, R. L., and S. A. Temple. 1986a. Methodological problems in the study of avian nest defense. Animal Behaviour 34: 561-566. Knight, R. L., and S. A. Temple. 1986b. Nest defense in the American goldfinch. Animal Behaviour 34: 887-897. Knight, R. L., and S. A. Temple. 1988. Nest-defense behavior in the redwinged blackbird. Condor 90: 193-200. Knight, R. L., and S. A. Temple. 1994. Wildlife and recreationists can exist. Pages in R. L. Knight and K. Gutzwiller, eds., Wildlife and recreatiOnIsts: coexistence through research and management. Island Press, Covelo, California. pp. 27 �124 Newton, I. 1979. Population ecology of raptors. Buteo Books, Vermillion, S. D. 399pp. . Olendorff, R. R. 1973. The ecology of the nesting birds of prey of northeastern Colorado. U. S. IBP Grassland Biome Technical Report 211. Natural Resource Ecology Laboratory, Colorado state University, Fort Collins. 223pp. Palmer, R. S. 1988. Handbook of North American birds. Vol. 5. Yale University Press, New Haven, Conn. 465pp. Patterson, T. L., L. Petrinovich, and D. K. James. 1980. Reproductive value and appropriateness of response to predators by White-crowned Sparrows. Behavioural Ecology and Sociobiology 7: 227-231. Redondo, T. 1989. Avian nest defense: theoretical models and evidence. Behaviour 3: 161-195. Ricklefs, R. E. 1977. Reactions of some Panamanian birds to human intrusion at the nest. Condor 9: 376-379. SAS Institute Inc. 1988. SAS/STAT user's guide, release 6.03 ed. SAS Institute Inc., Cary, N. C. 1,028pp. Skagen, S. K. 1980. Behavioral responses of wintering bald eagles to human activity on the skagit River, Washingto Pages 231-241 in R. L. Knight, G. T. Allen, M. V. Stalmaster, and C. W. Servheen, eds. Proceedings of the Washington bald eagle symposium. The Nature Conservancy, Seattle, Wash. Skagen, S. K., R. L. Knight, and G. H. Orians. 1991. Human disturbance of an avian scavenging guild. Ecological Applications 1: 215-225. sproat, T. M., and G. Ritchison. 1993. The nest defense behavior of eastern screech owls: effects of nest stage, sex, nest type, and predator location. Condor 95: 288-296. Stalmaster, M. V., and J. R. Newman. 1978. Behavioral responses of wintering bald eagles to human activity. Journal of Wildlif~ Management 42: 506-513. White, C. M., and T. L. Thurow. 1985. Reproduction of ferruginous hawks exposed to controlled disturbance. Condor 87: 14-22. Wiklund, C. G. 1990. Offspring protection by merlin Falco columbarius females; the importance of brood size and expected offspring survival for defense of young. Behavioural Ecology and Sociobiology 26: 217-223. Winkler, D. W. 1992. Causes and consequences of variation in parental defense behavior by tree swallows. Condor 94: 502-520. 28 � Dublin Core The Dublin Core metadata element set is common to all Omeka records, including items, files, and collections. For more information see, http://dublincore.org/documents/dces/. Title A name given to the resource Federal Aid Wildlife Quarterly/Annual Reports Description An account of the resource In 1937 the U.S. Congress enacted, and President Franklin D. Roosevelt signed, the Federal Aid to Wildlife Restoration Act (50S-917). This milestone in wildlife financing is popularly known as the Pittman-Robertson act. The Colorado Federal Aid to Wildlife Restoration program began in 1937, shortly after enabling legislation by the Congress in 1936. Reports of progress, as required by the Federal Aid Act, first appeared in 1938. Federal Aid reports have been prepared continuously since. Beginning in 1947 and continuing through 2000, Colorado Federal Aid Reports were issued quarterly, in January, April, July and October. <strong>Starting in 2001</strong>, reports were issued annually.&nbsp;<br /><br />Current reports cover both federal aid and non-federal aid research and summarize (≤6 pages each with tables and figures) preliminary results of wildlife research projects and support services updates conducted by the Mammals<br />Research Team.&nbsp;&nbsp;<br /><br />The title of the reports has changed over the years. Reports from 1938-1945 were issued under specific job title.&nbsp; For a listing of individual projects see <a href="https://cpw.cvlcollections.org/exhibits/show/mammals-research/progress-reports">Mammals Research: Progress Reports (1939-current)</a>.<br /><br />Quarterly Progress Report (1946-1956)<br /><a href="https://cpw.cvlcollections.org/items/show/674">1946-1956</a><br /><br />Quarterly Report (1957-1962)<br /><a href="https://cpw.cvlcollections.org/items/show/673">1957-1962</a><br /><br />Game Research Report (1963-1979)<br /><a href="https://cpw.cvlcollections.org/items/show/454">1963-1969</a><br /><a href="https://cpw.cvlcollections.org/items/show/455">1970-1979</a> <p>Wildlife Research Report (1980-2000)<br /><a href="https://cpw.cvlcollections.org/items/show/451">1980-1987</a><br /><a href="https://cpw.cvlcollections.org/items/show/452">1988-1994</a><br /><a href="https://cpw.cvlcollections.org/items/show/453">1995-2000</a></p> <p>Wildlife Research Report. Mammals (2001-2023)<br /><a href="https://cpw.cvlcollections.org/items/show/611" class="permalink">2001-2023</a><br /><br />Mammals Research Summary Report (2024-&nbsp; )<br /><a href="https://cpw.cvlcollections.org/items/show/612">2024-present</a></p> Text A resource consisting primarily of words for reading. Examples include books, letters, dissertations, poems, newspapers, articles, archives of mailing lists. Note that facsimiles or images of texts are still of the genre Text. Dublin Core The Dublin Core metadata element set is common to all Omeka records, including items, files, and collections. For more information see, http://dublincore.org/documents/dces/. Title A name given to the resource Wildlife Research Report (1988-1994) Description An account of the resource <p>Published under title: <em>Wildlife Research Report</em>&nbsp;from 1980-2000. This set contains <strong>1988-1994</strong>.<br /><a href="https://cpw.cvlcollections.org/items/show/451">Wildlife Research Report (1980-1987)</a><br /><a href="https://cpw.cvlcollections.org/items/show/453">Wildlife Research Report (1995-2000)</a><br /><br />Research topics issued in specific months. January 1983-1986: Non-game investigations; January 1987-1993: Raptor investigations. April 1980-1982: Game bird survey; April 1983: Small game investigations; April 1984: Game bird survey; April 1986-2000: Avian research. July 1980-1984, July 1985- Mammals. September 1983-1984: Law Enforcement investigations; September 1986: Wildlife law enforcement research. October 1980-1984: Migratory bird investigations; October 1999: Avian research, migratory birds; October 1989-1993: Migratory game bird research.</p> <p>Progress reports for mammals and avian Federal Aid research. See&nbsp;<a href="https://cpw.cvlcollections.org/exhibits/show/mammals-research/progress-reports">Mammals Research: Progress Reports (1939-current)</a>&nbsp;for a listing of collated reports for each mammals project (avian to be added in the future).</p> <p>Some quarters contain multiple volumes.<br /><br />Continues: Game Research Report (1963-1969)<br />&nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp;&nbsp;<a href="https://cpw.cvlcollections.org/items/show/454">1963-1969</a>&nbsp; |&nbsp;&nbsp;<a href="https://cpw.cvlcollections.org/items/show/455">1971-1979</a><br /><br />Continued by:&nbsp;<a href="https://cpw.cvlcollections.org/items/show/611">Wildlife Research Report. Mammals (2001-current)</a>. Avian research did not publish quarterly/annual report from 2001-2009.</p> <p>Print copy: Federal Aid binders</p> Rights Information about rights held in and over the resource <a href="http://rightsstatements.org/vocab/InC-NC/1.0/">IN COPYRIGHT - NON-COMMERCIAL USE PERMITTED</a> Type The nature or genre of the resource Text Format The file format, physical medium, or dimensions of the resource application/pdf Language A language of the resource English