https://cpw.cvlcollections.org/files/original/eb7532e985f2c051f1d72f6a5f423c40.pdf 94feda63fc87590df5f8340aded09d7c PDF Text Text Census of Swift Fox (Vulpes velox) in Canada and Northern Montana: 2000-2001 Alberta Species at Risk Report No. 24 �Census of Swift Fox (Vulpes velox) in Canada and Northern Montana: 2000-2001 A. Moehrenschlager and C. Moehrenschlager Alberta Species at Risk Report No. 24 August 2001 �Publication No.: I/034 ISBN: 0-7785-1828-0 (Printed Edition) ISBN: 0-7785-1829-9 (On-line Edition) ISSN: 1496-7219 (Printed Edition) ISSN: 1496-7146 (On-line Edition) Illustration: Brian Huffman For copies of this report, contact: Information Centre – Publications Alberta Environment / Alberta Sustainable Resource Development Main Floor, Great West Life Building 9920 108 Street Edmonton, Alberta Canada T5K 2M4 Telephone: (780) 422-2079 OR Information Service Alberta Environment / Alberta Sustainable Resource Development #100, 3115 12 Street NE Calgary, Alberta Canada T2E 7J2 Telephone: (403) 297-3362 This publication may be cited as: Moehrenschlager, A. and C. Moehrenschlager. 2001. Census of Swift Fox (Vulpes velox) in Canada and Northern Montana: 2000-2001. Alberta Sustainable Resource Development, Fish and Wildlife Division, Alberta Species at Risk Report No. 24. Edmonton, AB. 21 pp. ii �Project Partners and Funding Agencies: Wildlife Preservation Trust Canada iii �DISCLAIMER The views and opinions expressed are those of the authors and do not necessarily represent the policies or positions of the Department or the Alberta Government. iv �TABLE OF CONTENTS ACKNOWLEDGEMENTS............................................................................................. viii EXECUTIVE SUMMARY ............................................................................................... ix 1.0 INTRODUCTION ........................................................................................................ 1 1.1 Historical Distribution and Abundance .................................................................... 1 1.2 Canadian Swift Fox Reintroduction.......................................................................... 1 1.3 Population Status (1994-1999) ................................................................................. 2 1.4 Census Objectives (2000-2001)................................................................................ 3 2.0 METHODS ................................................................................................................... 4 2.1 Study Area ................................................................................................................ 4 2.2 Training..................................................................................................................... 4 2.3 Catch-and-Release Trapping..................................................................................... 5 2.4 Supplementary Indicators of Swift Fox Presence: Snow-tracking, Spot-lighting and Incidental Sightings.................................................................................................. 7 2.5 Data Analyses ........................................................................................................... 8 3.0 RESULTS ..................................................................................................................... 9 3.1 Catch-and-Release Trapping Effort .......................................................................... 9 3.2 Population Composition.......................................................................................... 11 3.3 Catch-and-Release Trapping Success and Population Changes ............................. 12 3.4 Population Distribution and Connectivity .............................................................. 13 4.0 DISCUSSION ............................................................................................................. 15 4.1 Recommendations................................................................................................... 17 5.0 LITERATURE CITED ............................................................................................... 19 v �LIST OF TABLES Table 1. Weather criteria protocols for live-trapping swift foxes...................................... 7 Table 2. Age, sex, and origin of swift foxes trapped in 1996-1997 and 2000-2001........ 11 Table 3. Comparative survey effort, capture success, and proportion of foxes that were recaptured at least once in the Alberta/Saskatchewan border subpopulation, the Grasslands National Park subpopulation, and Montana in 1996-1997 and 2000-2001 ................................................................................................................. 13 Table 4. Survey effort, captures, estimated densities, and estimated population sizes for the Alberta/Saskatchewan border, Grasslands, Montana, and the total survey area............................................................................................................................ 13 vi �LIST OF MAPS Map 1. Suspected range/sampled range for Canada and Montana in 2000-2001.............. 4 Map 2. Trap locations for Canada and Montana in 2000-2001 ......................................... 6 Map 3. Alberta and Alberta/Saskatchewan border area locations in 2000-2001............... 9 Map 4. Grasslands National Park locations of swift fox captures .................................. .10 Map 5. Montana locations of swift fox captures in 2000-2001 ....................................... 10 Map 6. Sign-tracking sample areas and results/locations of swift foxes in 2000-2001... 14 vii �ACKNOWLEDGEMENTS The commitment of many individuals and organizations has allowed for the successful completion of this Canada and northern Montana 2000-2001 swift fox population census. Overall census coordination is attributable to Dr. A. Moehrenschlager (Conservation Research, Calgary Zoo). A great deal of preparation and organization took place prior to this endeavour including the instructional and technical training at the Calgary Zoo to a dedicated group of field crew members: Jason Benko, Kelly Comishin, Jason Flaten, Dave Fuller, Marc Kloker, Brian Meagher, Dave Quinn, Corey Rasmussen, Ron Stoneberg and Tanya Wolowski. Funding was provided by Alberta Conservation Association; Alberta Sustainable Resource Development; Alberta Sports, Recreation, Parks and Wildlife Foundation; Calgary Zoo; Parks Canada; Saskatchewan Environment and Resource Management; Wildlife Preservation Trust Canada; and Wildlife Trust. Further Montana funding was provided by Montana Fish, Wildlife and Parks; National Fish and Wildlife Foundation; and the United States Bureau of Land Management. Logistical planning and endless support was received from Steve Brechtel, Sue Cotterill and Joel Nicholson (Alberta Sustainable Resource Development, Fish and Wildlife Division); Tian Dalgleish and Nicole Peters (Calgary Zoo Conservation Research – Calgary Zoological Society); Dr. Sandie Black, Dr. Maud Lafortune, Lori Rogers, Dr. Todd Shury, and Heather Spratt (Calgary Zoo Veterinary Services – City of Calgary); Brian Giddings and Ron Stoneberg (Montana Fish, Wildlife, and Parks); Pat Fargey (Parks Canada in Grasslands National Park); Wayne Harris (Saskatchewan Environment and Resource Management); and Elaine Williams (Wildlife Preservation Trust Canada). Also, thank you to Dr. Alonso Aguirre (Wildlife Trust) and Dr. Alex Aguirre for their hard work regarding the Conservation Medicine aspects of this project. Traps were provided by Alberta Sustainable Resource Development, Fish and Wildlife Division; Calgary Zoo; Canadian Wildlife Service; Montana Fish, Wildlife and Parks; and Parks Canada. Our special thanks to all landowners for permission to trap on their land, continual support, knowledge of swift foxes, and most of all for their hospitality. Moreover, thanks to the Saville/Moebis family for a temporary home base field station and their continual involvement and interest in the future of the swift fox. Funding for this report was provided by Alberta Sustainable Resource Development. Many thanks to Ron Stoneberg and Nicole Peters for their contribution towards data entry and verification. Jane Horb (Ducks Unlimited Canada) produced the maps for this report. Editorial reviews were submitted by Steve Brechtel, Sue Cotterill, Pat Fargey, and Elaine Williams. viii �EXECUTIVE SUMMARY The swift fox (Vulpes velox) is a rare, house-cat-sized carnivore that can race across native prairie at speeds of up to 60 km/hr. Although swift foxes were once so abundant in Canada that 117 025 were trapped between 1853 and 1877, this species was extirpated from Canada and northern Montana by the late 1930s. Since 1983, a reintroduction program has been underway to restore this species to Canada and the most recent releases were made in Grasslands National Park, Saskatchewan in 1997. A Canadian swift fox census during the winter of 1996-1997 revealed that the reintroduced population was located within two regions: 1) approximately 192 foxes were estimated to span the Alberta/Saskatchewan border south of the Cypress Hills; and 2) approximately 89 foxes were thought to exist along the United States border in and around Grasslands National Park, Saskatchewan. Concurrent with the Canadian swift fox reintroduction program, mounting evidence suggested that Canadian fox releases had also established a small swift fox population in north-central Montana. However, a coordinated international effort has not been previously conducted to assess the extent and composition of the shared swift fox population in Canada and Montana. The focus of the 2000-2001 census was: 1) to estimate changes in the distribution and abundance of swift foxes within Canada since the 1996-1997 census; 2) to estimate the distribution and abundance of swift foxes in adjacent areas of Montana. The 1996-1997 census area of 108 Canadian townships was resurveyed and supplemented with 80 Montana townships to form a total study area of 17 326.1 km2. Following training at the Calgary Zoo, six field teams conducted catch-and-release surveys in 80.3% of the study area townships from November 4, 2000 until February 15, 2001. Significant results were as follows: 1. In total, 149 swift foxes were live-trapped: 97 in the Alberta/Saskatchewan border area population, 14 in the Grasslands National Park region, and 38 in adjacent Montana areas. By comparison, 32 swift foxes were caught during catch-and-release efforts during the Canadian 1996-1997 swift fox census. In 2000-2001, 98.6% of captured foxes were unmarked, which means that they were wild-born in the Canadian/Montana population. This is a greater proportion of wild-born foxes than that recorded in 1996-1997, when 81.3% were unmarked. 2. The known distribution of swift foxes in Canada and Montana has substantially increased through the results of this census. In the Alberta/Saskatchewan border area swift foxes were found in 18 townships in 1996-1997 and during the 20002001 census they were found in 38 townships. In the Grasslands National Park region, swift foxes were found in 7 townships during the previous census, whereas they have now been located in 13 townships. In Montana, where a previous census of this kind had not been conducted, swift foxes were found in 25 townships. Hence, the 1996-1997 census found swift foxes in 25 townships ix �3. 4. 5. 6. 7. 8. 9. whereas the current survey yielded evidence of swift foxes in 76 townships; this represents a three-fold increase in the known swift fox distribution. The number of swift fox captures in Canada has tripled since 1996-1997, in areas that were previously surveyed at the same time of year. Similar recapture rates between the 2000-2001 census and the 1996-1997 census indicate that the threefold increase of the Canadian swift fox population in replicated areas is not due to higher trapability of foxes during this census but, in fact, that this represents a statistically significant, three-fold increase in the fox population in these areas. Fox body condition and age ratios are similar to those of the 1996-1997 census, but there has been a significant shift from a male-biased sex ratio previously to a female-biased population during this census. The increase in swift fox population size since 1996-1997 differed significantly between the Canadian swift fox subpopulations. Captures in the replicated regions of the Alberta/Saskatchewan border subpopulation have significantly increased by a factor of 3.5. By comparison, the 1.6-fold increase in the Grasslands National Park region is not statistically significant. In newly surveyed areas, 50% of 16 townships in the Alberta/Saskatchewan border area had successful captures for a total of 22 foxes. By comparison, only 9% of 11 newly surveyed townships in the Grasslands National Park area had successful captures totalling one fox. In Montana, which had not been previously surveyed, 31.8% of the 66 townships had swift foxes totalling captures of 38 individuals. Capture success for replicated and new areas combined was highest in the Alberta/Saskatchewan border area with 1 new capture every 10.0 trapnights, intermediate in Montana with 1 new capture/33.2 trapnights, and lowest for the Grasslands National Park region with 1 new capture/38.6 trapnights. Application of the same population estimation technique utilized in 1996-1997 suggests that the Alberta/Saskatchewan border population consists of 560 individuals (compared to 192 previously), the Grasslands National Park area contains 96 individuals (compared to 87 previously), and the sampled Montana area contains 221 foxes. This suggests a total population size of 877 foxes. Previously the Canadian swift fox population has been thought to consist of two isolated subpopulations. The comparative results in Canada suggest that this population has experienced a significant increase since 1996-1997. Furthermore, the present census suggests that the foxes in Canada and Montana now form one loosely-connected population. x �1.0 INTRODUCTION 1.1 Historical Distribution and Abundance Swift foxes historically ranged from Canada, to eastern Wyoming, the Dakotas, Nebraska, Kansas and Colorado through to north-western Texas, the Oklahoma panhandle, and eastern New Mexico (Scott-Brown et al. 1987). Naturalists and explorers considered swift foxes abundant within their historical range, and 10 614 swift fox pelts were traded by the American Fur Company at the Missouri and Sioux outfits between 1835 and 1838 (Hillman and Sharps 1978) while only 1989 red fox and 108 gray fox furs were traded during the same period (FaunaWest Wildlife Consultants 1991). Swift fox pelts had little value compared to those of other furbearers, selling for $0.25 in Kansas by 1861 (FaunaWest Wildlife Consultants 1991) and for between $0.30 and $1.32 in London in 1906 (Seton 1925); consequently, swift foxes were probably captured incidentally as trappers targeted more lucrative furbearers. Before European settlers arrived, swift foxes were found in Canada from the Pembina Hills in Manitoba across southern Saskatchewan to the southern foothills of the Rocky Mountains in Alberta (Carlington 1980). An average of 4876 pelts was sold in Canada by the Hudson Bay Company annually between 1853 and 1877 for a total of 117 025 specimens (Rand 1948) and the species was rare in the northern portions of its range by 1900 (Hillman and Sharps 1978). Between 1922 and 1925, an average of only 508 swift fox pelts were taken in Canada and record-keeping discontinued in 1925 because of these low numbers (Carlington 1980). The last Canadian museum specimen was collected in 1928 near Govenlock, Saskatchewan and the last sighting was made near Manyberries, Alberta in 1938 (Soper 1964). The species was officially designated as extirpated from Canada in 1978 (COSEWIC 2001). Today, swift foxes only exist in approximately 40% of their historical North American range (Kahn et al. 1997). 1.2 Canadian Swift Fox Reintroduction In 1973 two swift fox pairs from Colorado were sent to the home of the Smeeton family near Cochrane, Alberta (Herrero et al. 1991) where both pairs bred in 1974 (Herrero et al. 1986). In 1976, Dr. Steve Herrero of the University of Calgary and the Smeetons agreed to assess the feasibility of reintroducing swift foxes to Canada. The political and biological aspects of this potential reintroduction were consequently investigated with the assistance of three University of Calgary graduate students (Carlington 1980, Reynolds 1983, Schroeder 1985). In 1983, the first releases of captive-bred foxes were attempted in Alberta (Reynolds 1983) followed by releases in Saskatchewan in 1984, and agreements subsequently signed between federal and provincial governments to delineate responsibilities for further swift fox reintroductions (Pruss 1994). Swift foxes were released annually from 1983 until 1997. A feasibility study was completed in 1993 which concluded that, based on previous successes, a self-sustaining population of swift foxes could be established and that the most effective method of 1 �achieving this goal would be to conduct and monitor 3 to 5 more years of swift fox releases (Brechtel et al. 1993, Carbyn et al. 1994). Swift fox reintroductions occurred from 1983-1996 in the Alberta/Saskatchewan border area and from 1984-1997 in the Grasslands National Park region of south-central Saskatchewan. An attempted reintroduction into the Milk River ridge region in southern Alberta was discontinued because of a rabies outbreak in the skunk population. In total 942 foxes were released through captive-breeding and, more recently, translocation from the United States. As a result of Canadian reintroductions, the number of swift fox reports in north-central Montana have steadily increased, and Zimmerman (1998) established that swift foxes were establishing territories and breeding successfully in northern counties of the State. Reintroductions of swift foxes have also occurred on the Blackfoot Indian Reserve in Montana, and have been ongoing since 1998 through the collaboration of the Blackfoot Nation, Defenders of Wildlife, the Cochrane Ecological Institute, and Montana Fish, Wildlife, and Parks. Currently, foxes from these reintroductions are thought to be isolated from the swift fox population in Canada and north-central Montana. 1.3 Population Status (1994-1999) The ecology, status, and threats of Canadian swift foxes were intensively studied from 1994 until 1998. During this time, 76 swift foxes and 11 coyotes were radio-tracked for up to 3.5 years. Methods of reducing swift fox injuries were developed, factors influencing swift fox survival and reproduction were identified, the efficacy of swift fox translocations was evaluated, the effects of pipeline construction on swift foxes were determined, and factors reducing coyote predation on swift or kit foxes throughout North America were identified (Moehrenschlager 2000). These investigations also showed that, although survival and reproductive rates were variable between years, the swift fox population can sustain short-term fluctuations in predator pressure and climatic extremes. As one part of the ongoing recovery program, a national swift fox census was conducted in Alberta and Saskatchewan during the winter of 1996-1997 to determine the composition, distribution, and abundance of Canada’s swift fox population. The population is split into two subpopulations; the first, which spans the Alberta / Saskatchewan border, was estimated to have 192 foxes while the second, which is centred around Grasslands National Park in south-central Saskatchewan was estimated to contain 89 foxes (Cotterill 1997). Moreover, census results combined with ongoing live-captures for the 3.5 year-long research study (Moehrenschlager 2000), showed that 88% of individuals trapped from 1995 – 1998 were born in the wild in Canada. Hence, the Canadian swift fox population no longer relied on the reintroduced founder stock. In 1999, the Alberta / Saskatchewan border population assessment (Moehrenschlager and Moehrenschlager 1999) showed that this subpopulation was stable or increasing in size. 2 �1.4 Census Objectives (2000-2001) The primary goal of the National Recovery Plan for the Swift Fox is to restore populations to self-sustaining levels in the Canadian prairies by the year 2000 (Brechtel et al. 1996), and to remove swift fox from the endangered species list. Since Canadian swift fox releases ended in the Alberta/Saskatchewan border area before the 1996-1997 swift fox census, one can now ascertain whether the foxes can persist for a four-year time span without population supplementation. The Grasslands subpopulation did have releases in 1997 but the lack of releases in subsequent years still allows an examination of the state of this subpopulation without annual releases. Consequently, the Canadian Swift Fox Recovery Team and Montana Fish, Wildlife and Parks identified the following objectives for this international census: Key Objectives 1. To estimate changes in the distribution and abundance of swift foxes within the Canadian area surveyed during the 1996-1997 census. 2. To estimate the distribution and abundance of swift foxes in Montana, south of the Canadian Grasslands National Park and Alberta/Saskatchewan subpopulations. The census results will also be utilized to investigate secondary Canadian National Recovery Team objectives, which focus on environmental and demographic factors that determine the potential growth or extinction of the reintroduced population: 1. To determine exposure of swift foxes to disease in Canada. This is conducted through the collaboration of the Calgary Zoo and the Wildlife Trust. 2. To conduct preliminary swift fox habitat comparisons between sites with captures vs. those sites that did not have captures. The Calgary Zoo is partnering with Parks Canada and Alberta Sustainable Resource Development - Fish and Wildlife Division in this endeavour. 3. To determine the genetic relatedness and connectivity between separate regions of the Canadian and Montana swift fox population. The Calgary Zoo is partnering with the University of Alberta on this question. 4. To develop a population viability analysis which will model the likelihood of swift fox population sustainability for the Canadian/Montana swift fox population that was surveyed during the census. The Calgary Zoo will partner with Oxford University and additional collaborators to address this issue. Research is ongoing regarding these four secondary objectives, but the focus of this report is to address the two key objectives. This analysis is intended to be a timely follow-up to field work conducted from November 2000–February 2001. As such, population size analyses are preliminary and will be subject to further review over the course of the next year. 3 �2.0 METHODS 2.1 Study Area The study area spanned the borders of Alberta, Saskatchewan and Montana (48o 25’N; 49o 29’N; 106o 35’W; 110o 48’W), an area with scarce human habitation that is primarily used for cattle ranching. The Alberta/Saskatchewan border area spanned from Manyberries, Alberta to Claydon, Saskatchewan (Map 1, Map 3). The Grasslands National Park region spanned from Climax in the west to Killdeer in the east (Map 1, Map 4). The Montana study area spanned from Wild Horse in the west to Opheim in the east (Map 1, Map 5). This area is within the Brown Soil Zone characterized by shallow profiles, low amounts of organic matter, nitrogen and phosphorus, and a compact calcium carbonate layer averaging about 30 cm below the surface. While a small proportion of crop-land (<5%) is present on the periphery of the study area, the main vegetation types are representative of the Mixed Prairie Association which is characterized by an abundance of mid- and shortgrasses, numerous forbs, and few scrubs (Smoliak 1985). 2.2 Training Training of field staff was conducted at two sites. Five crews were at the Calgary Zoo from October 30–November 3 inclusively where they were instructed in fox handling 4 �procedures, landowner relations, safety, GPS use, mapping, data collection, snowtracking evaluations, and reporting requirements. In addition, a swift fox was handled by the field crews at the Calgary Zoo. From November 4 until November 9, all trapping crews moved to Consul, Saskatchewan. During this time period, traps were lined with hard-board to reduce injuries, equipment was assembled, and field training was conducted. During training, 18 foxes were trapped in 3.5 nights on three townships and all recaptured foxes were handled to maximize training opportunities; thereafter recaptured foxes were not handled. Subsequently, teams split into respective jurisdictions. Swift fox trapping was conducted by ten field staff, the team coordinator (CM), the census coordinator (AM), and two veterinarians from the Wildlife Trust. Field staff formed five teams, and of these, two surveyed the Alberta/Saskatchewan border subpopulation, one surveyed the Grasslands National Park region, and two sampled the Montana areas. The sixth team, which consisted of AM and the Wildlife Trust veterinarians, sampled swift foxes in the core of the Alberta/Saskatchewan border population in January 2001. To enhance Conservation Medicine components, serological health assessments were conducted of captured swift foxes at this time but a discussion of these findings is beyond the scope of this report. 2.3 Catch-and-Release Trapping The Canadian swift fox census area was determined by the National Swift Fox Recovery Team in 1996 based on habitat criteria and background information (Cotterill 1997). The suspected swift fox range consisted of 65 townships in the Alberta/Saskatchewan border region and 43 townships in and around Grasslands National Park in south-central Saskatchewan. Of these 108 townships, 81 (75%) were randomly selected for subsequent sampling, of which 2 were excluded at the request of the resident landowners. Similarly a census area of 80 townships was selected by Montana Fish, Wildlife and Parks staff in 2000, and 75% of these townships were selected for subsequent sampling (Map 1). Montana wildlife officials and the Canadian Swift Fox Recovery Team agreed not to include recent reintroduction areas on the Blackfoot Indian Reserve in north-western Montana, which may be disconnected from the current continuous Canada/Montana swift fox population. During the 1996-1997 census, live-trapping in 58 of the 81 randomly selected, Canadian townships was successfully completed. Catch-and-release trapping priorities for the 2000-2001 census, with diminishing priority, were as follows: 1) replicate townships trapped in 1996-1997; 2) sample the remaining 75% of randomly selected townships in Canada and Montana; and 3) survey the final 25% of townships. Although trap effort significantly increased from the previous census, not all townships could be surveyed due to access, time, or equipment restrictions. Swift fox trapping was conducted from November 4, 2000 until February 15, 2001 in Alberta, Saskatchewan, and Montana. Individual live-traps were placed at one-kilometre intervals along a five kilometre continuous section of the trail closest to the center of respective townships. The inter-trap distance was adjusted by up to 100 m to allow trap 5 �placement along fences or on top of hills (Map 2). Each township was surveyed with six traps for three nights, for a total of 18 trap nights per township. Catch-and-release trapping was conducted on consecutive nights when possible, but this was dependent on weather conditions. Two sizes of fox live-traps were used for catch-and-release procedures; they were 109 cm x 39 cm x 39 cm Tomahawk (Tomahawk Live Trap Co., Tomahawk, WI) double-door or 83 cm x 31cm x 31 cm Tomahawk single door box traps. Trap bottoms and corners were lined with 3 mm hard-board to reduce the likelihood of jaw, canine, or paw injuries (Moehrenschlager 2000). Catch-and-release trapping was conducted at night to reflect fox activity periods, to avoid heat-stress, and to prevent fox disturbance by people. Traps were generally set between 1800 and 2000, checked between 2400 and 0200, and closed following a second check between 0600 and 0800. Trapping was not conducted at temperatures colder than –20o C or when snow, rain, and wind conditions were potentially hazardous to captured foxes (Table 1). Foxes were handled by two field workers in each team. The first positioned the animal on his/her lap to shelter it from the wind, one hand restrained the head and covered the eyes, and the second hand restrained the body. The second field worker sexed and aged the fox, conducted parasite counts, scored body condition through palpation on an index ranging from 1 (poor) to 5 (excellent), collected a hair sample for genetic analyses, checked for injuries, and tattooed the ear for identification. Foxes were uniquely marked with tattoo dye so that recaptured individuals could be easily identified. 6 �Table 1. Weather criteria protocols for live-trapping swift foxes. Weather Conditions No Wind or Snow Low Wind Moderate Wind Strong Wind Low Wind and Light Snow Moderate or High Wind and Moderate, Heavy or Blowing Snow Rain Trap Setting Temperature (oC) > - 17 > - 17 > - 15 > - 13 > - 12 Traps not set Trap Closing Temperature (oC) < - 20 < - 20 < - 18 < - 16 < - 15 --- Traps not set --- Age classifications were based on the size, colour, and wear of teeth (see Ralls et al. 1990). Swift foxes emerge at approximately three weeks of age and emergence dates in the Canadian swift fox population range from May 25–June 9 (Pruss 1994, Moehrenschlager 2000). Given the variation in parturition dates and the extent of the census trapping period, captured juveniles were likely 5.5–10 months old. Accordingly, adults were 17.5 months or older. 2.4 Supplementary Indicators of Swift Fox Presence: Snow-tracking, Spot-lighting and Incidental Sightings Snow-tracking has been used as an indicator of swift fox presence in Canada (Mamo 1994, Moehrenschlager and Moehrenschlager 1999), and in this census, snow-tracking was conducted to supplement trapping information on the distribution of Canadian swift foxes. Within surveyed townships, each accessible trail was searched for the tracks of swift foxes, other carnivores and lagomorphs. No time restrictions were imposed for searches and trackers travelled by truck, ATV or on foot. The choice of townships was influenced by snow conditions but the emphasis with decreasing priority was on townships that had not been trapped, townships that had been trapped but where no captures had been made, and townships where swift foxes had been captured. Spot-lighting has been used to locate swift foxes (Mamo 1994, Moehrenschlager 1994, Moehrenschlager and Moehrenschlager 1999) in Canada. While it may not accurately reflect changes in population abundance, it is useful for presence/absence surveys. During this census, spot-lighting was conducted to find swift foxes in areas where they were not know to exist or to visually confirm their presence in areas where snow-tracking or scat sign had been found. The priorities for spot-lighting were: 1) to survey townships that had swift fox sign during snow-tracking but where no captures had been made; 2) to spotlight townships where neither snow-tracking nor trapping were conducted; 3) to survey townships where foxes had been trapped. For priorities 1 and 2, all accessible trails were sampled up to 7 �three times with at least two hours between spot-lighting passes. For priority 3 areas, trails were only sampled once. To conduct spot-lighting, two field workers drove no faster than 40 km/hr while scanning both sides of the road with a one-million candlelight spotlight. Carnivores were identified by eye or using binoculars, while jackrabbits or cottontails were counted on all surveyed trails. Incidental sightings were made opportunistically at times that spotlighting was not conducted. Locations of swift foxes, red foxes, and coyotes obtained through the various survey techniques were recorded with a GPS. 2.5 Data Analyses The number of sampled townships was compared relative to the survey area size of the Alberta/Saskatchewan border, the Grasslands National Park area, and Montana survey areas. The proportion of wild-born individuals in the population was compared between the 2000-2001 and the 1996-1997 censuses. Body weights were compared between the censuses with unpaired t-tests for juvenile and adult foxes respectively. Moreover, body weights were compared between the Alberta/Saskatchewan border, Grasslands National Park, and the Montana regions using a General Linear Model that also incorporated fox age. These tests, in combination with a Mann-Whitney U test that compared body assessment scores between the censuses, were conducted to determine if the condition of trapped foxes had changed over time. Age ratio differences between the present census and former census, and sex ratio comparisons, over time and between areas, were tested using chi-square. The number of captures on replicated Canadian townships was compared between 20002001 and 1996-1997 using Wilcoxon paired rank tests individually for the Canadian subpopulations and the combined population. The change in capture numbers on replicated townships was compared between the subpopulations with a Mann-Whitney U test. Two townships in the Alberta/Saskatchewan border population were excluded from paired analyses because they were only trapped for 0.5 nights during 2000-2001 presently compared to 3.0 nights during the previous census. Two foxes were trapped in these townships during the half-night in 2000-2001. Trapping success was compared between the Alberta/Saskatchewan border, Grasslands National Park, and the Montana regions by comparing the number of trapnights/new capture. For the Canadian areas, this measure was also compared between the 2000-2001 and the 1996-1997 censuses. Swift fox trapability was also compared between the two censuses by examining the proportion of captured foxes that was subsequently recaptured (for background, see Moehrenschlager and Moehrenschlager 1999). To allow for standardized comparisons over time, subpopulation and total population fox densities as well as respective abundance estimates were derived for the Canadian and Montana areas using the methods outlined in Cotterill (1997). Data collected in Alberta during the census were entered into the Biodiversity/Species Observation Database (BSOD). Initial capture locations of swift foxes were mapped for the Alberta/Saskatchewan border area, the Grasslands National Park region, and Montana; recaptures were not mapped. 8 �Moreover, swift fox locations obtained through spot-lighting, sign-tracking, or incidental sightings were mapped for all regions. 3.0 RESULTS 3.1 Catch-and-Release Trapping Effort The census area consisted of three connected regions: 1) Alberta/Saskatchewan border: 65 townships – 5990.4 km2 ; 2) Grasslands National Park region: 43 townships – 3962.9 km2 ; 3) Montana: 80 townships - 7372.8 km2 . The total census area spanned 17 326.1 km2 (Map 1). Swift fox trapping was conducted in 151 (80.3%) of the 188 study area townships. In the Alberta/Saskatchewan border area, 84.6% (55/65) of townships were surveyed during 969 trapnights (Map 1, Map 3) and, in the Grasslands National Park area, 69.8% (30/43) of townships were surveyed in 540 trapnights (Map 1, Map 4). In 1996-1997, 75% of the townships within the Canadian sample area were randomly selected; 95.9% (47/49) of the randomly selected Alberta/Saskatchewan border area townships were trapped and 93.3% (28/30) of randomly selected Grasslands National Park region townships were surveyed during the current census (Map 1). With the greater trapping effort during the 2000-2001 census, 23% more townships were sampled in Canada than during the 1996-1997 census. All 58 townships that were sampled in Canada previously were re-trapped; ie. 39 in the Alberta/Saskatchewan border area and 19 in the Grasslands National Park region (Map 3, Map 4). In Montana 82.5% (66/80) of study area townships were sampled in 1188 trap nights (Map 1, Map 5). 9 �10 �3.2 Population Composition In total, 149 swift foxes were captured; 97 in the Alberta/Saskatchewan border area, 14 in Grasslands, and 38 in Montana (Table 2). In 1996-1997, 81.3% (26/32) of captured foxes were wild-born while 12.5% were captive-bred and 6.3% were translocated foxes from Wyoming. Compared to the 1996-1997 census, the proportion of wild-born foxes has increased. Of 142 foxes with known origin, 98.6% were wild-born while one capture each of a captive-bred and a translocated fox, constituted the remaining 1.4% (Table 2). Table 2. Age, sex, and origin of swift foxes trapped in 1996-1997 and 2000-2001. Age* Area Sampled AB/Sask. 1996-1997 Grasslands 1996-1997 Total 1996-1997 AB/Sask. 2000-2001 Grasslands 2001-2002 Montana 2001-2002 Total 2000-2001 # of Foxes Sex* Female Wild-born 11 20 Origin* Captivebred 2 Translocated 2 24 Adult 13 Juvenile 11 Male 13 8 3 5 7 1 6 2 0 32 16 16 20 12 26 4 2 97 55 39 39 56 91 1 1 14 9 4 7 6 13 0 0 38 17 20 16 21 36 0 0 149 81 63 62 83 140 1 1 * In 2000/2001, the age, sex, and origin of one escaped fox in each of the Border, Grasslands, and Montana areas was unknown. In addition, the sex and age of one fox, the age of a second fox, and the origin of two foxes in the border area is unknown. Moreover, one fox in the border area and one in Montana were either wild-born in Canada or translocated from Wyoming. Body weights were similar between this and the previous census for adults (1996-1997: 2.3 ± 0.2 kg; 2000-2001: 2.3 ± 0.2 kg; t = 0.60, df = 95, p = 0.55) and juveniles (19961997: 2.1 ± 0.2; 2000-2001: 2.2 ± 0.2; t = 0.71, df = 77, p = 0.48). Similarly, body condition scores were similar between the censuses (z = 1.8, n1 = 32, n2 = 143, p = 0.07). Moreover, once the significant age effect was accounted for in a general linear model (F1, 144 = 13.0; p < 0.0001), there was no significant difference in fox body weights between the Alberta/Saskatchewan border population, the Grasslands area, and Montana (F2, 144 = 1.7; p = 0.19). The age ratio in Canada is similar now to that recorded during the previous census (19961997: 50% adults, 2000-2001: 56% adults; x2 = 1.0, df = 1, p = 0.32). However, the sex ratio has changed significantly from a male bias (63% males) in the 1996-1991 census to a female bias (57% females) during the 2000-2001 survey (x2 = 3.9, df = 1, p < 0.05; Table 2). There was no difference in the sex ratio between the Alberta/Saskatchewan, Grasslands, and Montana sampling regions during the present census (x2 = 0.8, df = 2, p = 0.68). 11 �3.3 Catch-and-Release Trapping Success and Population Changes In 2000-2001 significantly more foxes were captured in Canada on the townships that were originally sampled in 1996-1997 (Wilcoxon paired rank: z = 4.0; n = 56; p < 0.0001). While 29 were captured in these areas previously, 86 were caught during the present census; this represents a three-fold increase. The increase was primarily due to higher capture rates in the Alberta/Saskatchewan border population. Here the number of captures increased significantly (Wilcoxon paired rank: z = 3.7; n = 37; p < 0.0001) from 21 to 73, a 3.5-fold increase. Although the number of captures also increased in the Grasslands area from 8 to 13, this 1.6-fold increase was not statistically significant (Wilcoxon paired rank: z = 1.3; n = 19; p = 0.19). In fact, the change in the number of captures on replicated townships was significantly greater in the Alberta/Saskatchewan border area than in the Grasslands area (z = 2.1; n1 = 37, n2 = 19; p < 0.05). While these numbers reflect changes in previously sampled Canadian townships, trapping success also differs between regions in townships that were surveyed for the first time during the 2000-2001 census. In the Alberta/Saskatchewan border region, 16 new townships were surveyed of which 8 (50%) had captures for a total of 22 foxes. Comparatively, of 11 new townships surveyed in Grasslands, only one (9%) had a swift fox capture. In Montana, the 66 newly surveyed townships had 21 (31.8%) with swift foxes, totalling 38 individuals. Area differences over time are also apparent when examining trapping success/trapnights of effort. In 1996-1997, approximately three times as many trapnights were required for each new capture in the border area compared to the current census (Table 3). The success per effort ratio in the Grasslands National Park subpopulation on the other hand has only slightly improved. Currently, capture success is approximately 3 times higher in the border area than in Montana, but capture success in Montana is slightly higher than in the Grasslands population (Table 3). The difference in captures between years and between areas was not due to a difference in swift fox trapability. Recapture rates, which are an indicator of trapability (Moehrenschlager and Moehrenschlager 1999), were similar for each of the study areas between years (Table 3). The 1996-1997 trapability correction factor based on the home ranges and catchability of radio-tracked swift foxes (Moehrenschlager 2000), was applied when estimating the swift fox population (Cotterill 1997). This is reasonable due to the similarity in recapture likelihoods for the Canadian areas over time (Table 3). Estimated fox densities ranged from a low of 2.4 foxes/100 km2 in the Grasslands area to a high of 9.2 foxes/100 km2 in the Canadian border population (Table 4). Respective subpopulation estimates were 560 for the Alberta/Saskatchewan border area population, 96 for the Grasslands area, and 221 for Montana. The total, preliminary population estimate using this technique is estimated at 877 individuals. 12 �Table 3. Comparative survey effort, capture success, and proportion of foxes that were recaptured at least once in the Alberta/Saskatchewan border subpopulation, the Grasslands National Park subpopulation, and Montana in 1996-1997 and 2000-2001. Area AB/Sask. 1996-1997 Grasslands 1996-1997 Total 1996-1997 AB/Sask. 2000-2001 Grasslands 2001-2002 Montana 2001-2002 Total 2000-2001 Number of Townships Surveyed 39 # of Foxes Caught % of Foxes Recaptured # of Trapnights / New Capture 24 33 29.5 19 8 25 41.3 58 32 31 32.4 55 97 33 10.0 30 14 29 38.6 66 38 32 31.3 151 149 31 18.1 Table 4. Survey effort, captures, estimated densities, and estimated population sizes for the Alberta/Saskatchewan border, Grasslands, Montana, and the total survey area. Region Border Grasslands Montana Total Region Townships 65 43 80 188 Region Area (km2) 5990.4 3962.9 7372.8 17326.1 Townships Surveyed 53* 30 66 149* Total Area Sampled (km2)1 3690.5 2013.0 4428.6 10132.1 Foxes Caught in Sample 95* 14 38 147* Estimated Fox Density2 (foxes/100km2) 9.3 2.4 3.0 5.1 Estimated Population Size 560.1 96.4 221.4 877.9 1 Area sampled in region = (Area sampled per trapline * # of surveyed townships); (Area sampled per trapline = 67.1 km2) 2 Estimated Fox Density=[(# Foxes Caught in Sample * Correction Factor)/Area Sampled in Region] * 100 (Correction Factor = 3.5) 3 Estimated Population Size = (Total Area of Swift Fox Range * Estimated Density) * Two townships with a total of 2 fox captures were excluded in the border area because these areas were only trapped for 0.5 nights each. 3.4 Population Distribution and Connectivity The known swift fox distribution in Canada and Montana has significantly increased, in part because of a greater survey area and also because of a greater occurrence of swift foxes in re-surveyed regions. In the 1996-1997 census, swift foxes were trapped in 14 townships and additional sign was found in four townships to yield a known distribution of 18 townships in the Alberta / Saskatchewan border area. Comparatively, swift foxes were trapped in 32 and otherwise located in six border area townships for a total known distribution of 38 townships in 2000-2001. In the Grasslands National Park region, swift 13 �foxes were trapped in six and otherwise found in one township, yielding a 1996-1997 known census distribution of seven townships; in 2000-2001, foxes were trapped in nine and additionally found in four townships for a total known range of 13 townships. During the current census, swift foxes were also trapped in 21 and otherwise found in four Montana townships. In the Alberta/Saskatchewan border subpopulation, all trapped and sighted swift foxes were within 2 townships of another swift fox location (Map 3, Map 6). In the Grasslands National Park region, there is an apparent split between the eastern and western areas of the population as two clusters of swift foxes are separated by three empty townships (Map 4). This split might be exaggerated since no trapping, sign-tracking, or spotlighting surveys were conducted in township 1, range 11 and township 1, range 12 (Map 4, Map 6). Nevertheless, this apparent gap continues into Montana, where no foxes are found in ranges 33, 34, or 35 (Map 5, Map 6). However, the incidental sighting of a swift fox in township 37, range 32 just south of the Canadian border (Map 6), reduces the apparent connectivity gap of the Canadian side to two townships. The surveyed Montana population is artificially split in two because the pre-defined survey area had a gap in the center, and the random selection of townships within this zone excluded the southern, connecting region of this area (Map 1). In the western portion of the Montana survey area, all swift fox sightings or captures were within one township of each other and this block of swift fox presence is well connected to the Alberta/Saskatchewan border population. 14 �4.0 DISCUSSION The primary focus of this international swift fox census was to estimate changes in the distribution and abundance of swift foxes within Canada since 1996-1997 and to estimate these parameters for adjacent areas in northern Montana. Over a study area that spanned 17 326 km2, 149 swift foxes were captured from November 4, 2000 until February 15, 2001 and additional swift fox sightings or snow-tracking sign were documented. The fact that 98.6% of captured foxes were unmarked suggests that swift foxes are reproducing successfully in the wild and that the population primarily consists of wildborn individuals. This proportion of wild-born foxes is higher than that previously recorded in other Canadian catch-and-release studies (Brechtel et al. 1993, Mamo 1994, Cotterill 1997, Moehrenschlager and Moehrenschlager 1999, Moehrenschlager 2000). This likely results from two factors: 1) the fact that the last swift fox releases in the Alberta/Saskatchewan border were conducted in 1996 and in Grasslands National Park in 1997 and; 2) the population is persisting and expanding through the recruitment of wild young instead of released foxes. In resurveying Canadian townships that were originally sampled in 1996-1997, there was a statistically significant increase in capture rates while fox trapability was similar. Overall the population appears to have tripled in these areas over four years. This increase is primarily attributable to a surge in swift fox numbers in the Alberta/Saskatchewan border area. Perhaps the increase in the border population would not be surprising if 1996 had been a poor year, but daily radio-tracking from January 1995 until October 1997 and survival monitoring until February 1998, revealed that swift fox survival was higher in 1996 than 1995 and 1997 respectively (Moehrenschlager 2000). Hence, the increase in swift fox numbers documented during this census is already relative to a period of swift fox abundance. The Grasslands National Park population showed smaller population increases in replicated areas and lower trapping success in newly surveyed areas than the Alberta/Saskatchewan border population. This explains why the marginal increase in swift fox captures on replicated Grasslands townships did not result in a significantly larger population estimate than that of the previous census. In contrast, capture increases in replicated Alberta/Saskatchewan border townships were complemented by high capture successes in newly surveyed areas. The fact that the Grasslands National Park subpopulation has had smaller increases than the border subpopulation is especially unexpected since over 50 captive-bred swift foxes were released into the Grasslands area in 1997, whereas no releases were made in the border region. The lower numbers of swift foxes in the Grasslands National Park area seems to suggest one of or a combination of three occurrences since the last census: 1) higher mortality than in the border area; 2) lower reproduction than in the border area; or 3) higher net dispersal from the region than in the border area. 15 �The shift towards a female-biased sex ratio in 2000-2001 from a male-biased ratio in 1996-1997, suggests a higher ratio of effective population size/total population size now than before. Although skewed, the current sex ratio is better balanced and, as such, more effective than that recorded previously for this primarily monogamous species. However, swift foxes appear to occasionally be polygynous (Covell 1992; Olson et al. 1997; Moehrenschlager 2000) which may favour slightly female-biased populations. Hence, the increase in population size is also complemented by a per capita increase in reproductive potential. Swift foxes in the Alberta/Saskatchewan border area, Grasslands National Park region, and Montana are closely connected. The largest gaps in the distribution span three townships, which is within the maximum dispersal distance of 34.3 km recorded for naturally dispersing Canadian swift foxes (Moehrenschlager 2000). Even so, the gaps in the distribution are likely exaggerated because of limited surveys. Future presence/absence assessments should concentrate on apparent gaps between the eastern and western regions of the Grasslands National Park area and adjacent regions in Montana, between the Canadian border and Grasslands National Park populations, and between the eastern and western Montana survey area regions where no trapping was conducted. The implications of a loosely connected but continuous population are numerous. The population should be less prone to genetic drift in isolated fragments, but the susceptibility to extinction at the hands of disease will now be greater for the population as a whole. The primary factor driving small populations of canids to extinction is disease, as recent outbreaks of canine distemper or rabies have shown in African wild dogs, Ethiopian wolves, Blanford’s foxes, and Channel island grey foxes. During this 2000-2001 census, the study of wildlife disease in Canadian swift foxes was initiated and laboratory testing of collected blood samples will commence shortly. Overall the Canadian swift fox population has increased in abundance and in its known distribution. The previous perceptions that the Canadian population consists of two subpopulations and that swift foxes may only be found in scattered pockets in Montana, no longer hold. Zimmerman’s (1998) results combined with the findings from this census, show that the Montana population in the survey area is well established, consists almost exclusively of wild-born individuals, and is widespread. Moreover, the extent of the Canadian swift fox population reaches beyond the census area since occasional sightings have been documented in areas such as Suffield Alberta, north of Swift Current in Saskatchewan, and a single case in Manitoba. Moreover, the present Montana swift fox population, which originally drew its founders from Canadian releases, appears to be expanding as well. Recent surveys in Montana have shown a significant expansion as far west as the Sweetgrass Hills (Giddings, pers. comm.) and a southward reach into the Fort Belknap Indian Reserve (Stoneberg, pers. comm.). Indeed, the Montana population appears regionally well established, and it had higher capture success rates during this census than the Canadian Grasslands National Park population where releases were conducted for 14 years. One crucial finding is that the Canadian swift fox population has substantially increased over a four-year span without supplementation from swift fox releases. Consequently, 16 �swift fox releases are not necessary for the Canada/Montana swift fox population to increase in abundance or distribution, although such population supplementation may help to reduce the inbreeding effective population size (Ryman et al. 1995) and aid population subsistence over time. While the population has clearly increased in abundance and distribution, the factors driving the increase are not understood. It is safe to say, however, that an absence of such favourable conditions and/or the occurrence of stochastic events could cause future population crashes. For example, if high prey abundance, favourable winter conditions, or low predator numbers allowed for strong population growth, then low prey densities, harsh winter conditions and high predator numbers would cause a decrease in population size. The magnitude of future increases or decreases will depend on the additive effects of such parameters and the impact of stochastic factors such as disease or drought. Because of high capture densities in January 1995 and subsequently high mortality rates in the spring and summer of that year (Moehrenschlager 2000), it is the opinion of the authors, that the high densities in the border population are beyond the carrying capacity of the area and we expect increased mortalities and decreased per capita reproduction in these core zones in 2001. Equally, however, these increased densities will likely drive increased dispersal and colonization along the edges of the swift fox distribution. 4.1 Recommendations This international swift fox census has shown that the distribution and abundance of swift foxes has dramatically increased in Canada since 1996-1997. Moreover, swift foxes are clearly present in adjacent areas of northern Montana. This suggests that a genetically connected Canadian/Montana swift fox population has been established which now consists almost exclusively of foxes that have been born in these regions in the wild. While these results are an encouraging sign that swift foxes may one day be forever returned to Alberta, Saskatchewan and Montana after decades of extirpation, one cannot assume that a minimum viable population size has been established at this point. A minimum population size of 500 is thought to maintain sufficient genetic variability in quantitative characters (Franklin 1980, Reed and Bryant 2000). However, this number has also been debated extensively. While Franklin and Frankham (1998) believe an effective population size of 500–1000 is generally appropriate, Lynch and Lande (1998) maintain that 1000–5000 individuals should be considered a minimum. To ensure the protection and growth of the swift fox population, the authors believe that the following actions should be taken: I. II. Continue to monitor the population to assess future growth or declines in abundance and distribution. Develop rigorous education programs with the aid of local school teachers, farmers, and ranchers to increase swift fox awareness among children and adults in the very communities where swift foxes reside. The majority of 17 �III. IV. V. VI. VII. the swift fox distribution that was determined during this census lies in unprotected habitat. Consequently, the goodwill and continued support of local people is absolutely crucial towards the long-term survival of this species. This means that conservation planning for swift foxes must integrate the opinions and needs of local human communities. Devise a habitat model that will define crucial habitat parameters for swift foxes in this population and subsequently allow for the protection of critical swift fox areas. Outside parks, the primary form of environmental protection should be landowner stewardship incentives. A rigorous habitat model will also delineate potential areas for future swift fox establishment. Reduce human-caused mortalities of swift foxes. Road-kills, accidental trapping, and poisoning of swift foxes are factors that can be more easily controlled than environmental conditions that naturally impact the population. Solutions to these problems should be devised with the involvement of local people. Determine gene flow throughout the population to determine if isolated fragments exist that might be prone to inbreeding depression over time. In addition, conduct presence/absence surveys in areas where gaps were apparent in the swift fox distribution during this census. Determine age-specific exposure of swift foxes to canine diseases and the likelihood of disease contraction from sympatric canids. Integrate existing demographic data into a population viability model to determine time frame-specific likelihoods of population perseverance, identify primary threats that could drive the population to extinction, and establish swift fox-specific estimates of minimum viable population size. 18 �5.0 LITERATURE CITED Brechtel, S., L. Carbyn, G. Erickson, D. Hjertaas, C. Mamo, and P. McDougall. 1996. National Recovery Plan for the Swift Fox. Report No. 15. Recovery of Nationally Endangered Wildlife Committee, Ottawa, ON. 29 pp. Brechtel, S. H., L. N. Carbyn, D. Hjertaas, and C. Mamo. 1993. Canadian swift fox reintroduction feasibility study: 1989-1992; report and recommendations of the National Recovery Team. Unpublished report, Alberta Fish and Wildlife Services. 95 pp. Carbyn, L. N., H. Armbruster, and C. Mamo. 1994. The swift fox reintroduction program in Canada from 1983 to 1992. Pages 247-271 in M. Bowles and C. J. Whelan, eds. Symposium Proceedings on Restoration of Endangered Plants and Animals. University of Cambridge Press. Carlington, B. G. 1980. Re-introduction of the swift fox (Vulpes velox) to the Canadian prairies. Master's Thesis, Faculty of Environmental Design, University of Calgary, AB. COSEWIC. 2001. Canadian species at risk, May 2001. Committee on the Status of Endangered Wildlife in Canada, Ottawa, ON. 31 pp. Cotterill, S. E. 1997. Population census of swift fox (Vulpes velox) in Canada: Winter 1996/1997. Prepared for Swift Fox National Recovery Team. Alberta Environmental Protection. Natural Resources Service, Wildlife Management Division, Edmonton AB. 50 pp. Covell, D. F. 1992. Ecology of the swift fox (Vulpes velox) in southeastern Colorado. Thesis, University of Wisconsin-Madison, Madison, WI. FaunaWest Wildlife Consultants. 1991. An ecological and taxonomic review of the swift fox (Vulpes velox) with special reference to Montana. Prepared for Montana Department of Fish, Wildlife, and Parks; Montana State University Campus, Bozeman, MT. Franklin, I.R. 1980. Evolutionary changes in small populations. In Conservation Biology. An Evolutionary-Ecological Approach, eds. M.E. Soulé and B.A. Wilcox, pp. 135149. Sinauer Associates, Sunderland, MA. Franklin, I.R. and R. Frankham. 1998. How large must populations be to retain evolutionary potential? Animal Conservation 1: 69-70. Herrero, S., C. Mamo, L. Carbyn, and M. Scott-Brown. 1991. Swift fox reintroduction into Canada. Pages 246-252 in: Proc. of the Second Endangered Species and Prairie 19 �Conservation Workshop (Eds. G.L. Holroyd, G. Burns, and H.C. Smith) Provincial Museum of Alberta Natural History Section, Occasional Paper No. 15, Edmonton, AB. Herrero, S., C. Schroeder, and M. Scott-Brown. 1986. Are Canadian foxes swift enough? Biological Conservation 36: 159-167. Hillman, C. N., and J. C. Sharps. 1978. Return of swift fox to northern great plains. Proceedings of the South Dakota Academy of Science 57: 154-162. Kahn, R., L. Fox, P. Horner, B. Giddings, and C. Roy. 1997. Conservation assessment and conservation strategy for swift fox in the United States. Swift Fox Conservation Team. 54 pp. Lynch, M. and R. Lande. 1997. The critical effective size for a genetically secure population. Animal Conservation 1: 70-72. Mamo, C. C. 1994. Swift fox (Vulpes velox) population survey assessment. Alberta Fish and Wildlife Services and the Swift Fox Conservation Society. 28 pp. Moehrenschlager, A. 2000. Effects of ecological and human factors on the behaviour and population dynamics of reintroduced Canadian swift foxes (Vulpes velox). D. Phil thesis, University of Oxford, England. Moehrenschlager, A. 1994. Population monitoring of swift foxes (Vulpes velox) by means of scent-posting and spotlighting in Wood Mountain, Saskatchewan. Department of Saskatchewan Environment and Resource Management. Moehrenschlager, C. A. J., and A. Moehrenschlager. 1999. Canadian swift fox (Vulpes velox) population assessment: Winter 1999. Report to Alberta Environmental Protection. Edmonton, AB. 37 pp. Olson, T. L., J. S. Dieni, and F. G. Lindzey. 1997. Swift fox survey evaluation, productivity, and survivorship in southeast Wyoming. Wyoming Cooperative Fish and Wildlife Research Unit. 30 pp. Pruss, S. D. 1994. An observational natal den study of wild swift fox (Vulpes velox) on the Canadian Prairie. Thesis, Department of Environmental Design, University of Calgary, Calgary, AB. Ralls, K., White, P.J., Cochran, J., and D. B. Siniff. 1990. Kit-fox coyote relationships in the Carrizo Plain Natural Area. Annual Report to the U.S. Fish and Wildlife Service. U.S.A. Rand, A. L. 1948. Mammals of the eastern Rockies and western Plains of Canada. Natl. Mus. Can. Bull. 108. 20 �Reed, D.H.B., and H. Edwin. 2000. Experimental tests of minimum viable population size. Animal Conservation 3: 7-14. Reynolds, J. 1983. A plan for the reintroduction of swift fox to the Canadian prairie. Master's Thesis, Faculty of Environmental Design, University of Calgary, Calgary, AB. Ryman, N., Jorde, P.E., and L. Laikre. 1995. Supportive breeding and variance effective population size. Conservation Biology 9(6): 1619-1628. Scott-Brown, J. M., T. P. O'Farrell, and K. L. Hammer. 1987. Swift fox. In Novak, M., Baker, J.A., Obbard, M.E., and B. Malloch, eds. Wild Furbearer Management and Conservation in North America. Ontario Ministry of Natural Resources, ON. Schroeder, C. 1985. A preliminary management plan for securing swift fox reintroductions into Canada. Master's Thesis, Faculty of Environmental Design, University of Calgary, Calgary, AB. Seton, E. T. 1925. Lives of Game Animals. Vol. I. Cats, Wolves, and Foxes. Doubleday, Doran and Co., New York. Smoliak, S. 1985. Flora of the Manyberries Research Substation. Lethbridge Research Station Contribution No. 6. Lethbridge, AB. 31 pp. Soper, J. D. 1964. The mammals of Alberta. Hamly Press, Edmonton, AB. Zimmerman, A. L. 1998. Reestablishment of swift fox in northcentral Montana. M.S. Thesis, Montana State University, Bozeman, MT. 21 �List of Titles in This Series (as of August 2001) No. 1 Alberta species at risk program and projects 2000-2001, by Alberta Sustainable Resource Development, Fish and Wildlife Division. (2001) No. 2 Survey of the peregrine falcon (Falco peregrinus anatum) in Alberta, by R. Corrigan. (2001) No. 3 Distribution and relative abundance of the shortjaw cisco (Coregonus zenithicus) in Alberta, by M. Steinhilber and L. Rhude. (2001) No. 4 Survey of the bats of central and northwestern Alberta, by M.J. Vonhof and D. Hobson. (2001) No. 5 2000 survey of the Trumpeter Swan (Cygnus buccinator) in Alberta, by M.L. James and A. James. (2001) No. 6 2000/2001 Brassy Minnow inventory at Musreau Lake and outlet, by T. Ripley. (2001) No. 7 Colonial nesting waterbird survey in the Northwest Boreal Region – 2000, by M. Hanneman and M. Heckbert. (2001) No. 8 Burrowing owl trend block survey and monitoring - Brooks and Hanna areas, by D. Scobie and R. Russell. (2000) No. 9 Survey of the Lake Sturgeon (Acipenser fulvescens) fishery on the South Saskatchewan River, Alberta (June-September, 2000), by L.A. Winkel. (2000) No. 10 An evaluation of grizzly bear-human conflict in the Northwest Boreal Region of Alberta (19912000) and potential mitigation, by T. Augustyn. (2001) No. 11 Harlequin duck monitoring in the Northern East Slopes of Alberta: 1998-2000 preliminary results, by J. Kneteman and A. Hubbs. (2000) No. 12 Distribution of selected small mammals in Alberta, by L. Engley and M. Norton. (2001) No. 13 Northern leopard frog reintroduction. Raven River - Year 2 (2000), by K. Kendell. (2001) No. 14 Cumulative effects of watershed disturbances on fish communities in the Kakwa and Simonette watersheds. The Northern Watershed Project. Study 3 Progress report, by T. Thera and A. Wildeman. (2001) No. 15 Harlequin duck research in Kananaskis Country in 2000, by C.M. Smith. (2001) No. 16 Proposed monitoring plan for harlequin ducks in the Bow Region of Alberta, by C.M. Smith. (2001) No. 17 Distribution and relative abundance of small mammals of the western plains of Alberta as determined from great horned owl pellets, by D. Schowalter. (2001) No. 18 Western blue flag (Iris missouriensis) in Alberta: a census of naturally occurring populations for 2000, by R. Ernst. (2000) No. 19 Assessing chick survival of sage grouse in Canada, by C.L. Aldridge. (2000) No. 20 Harlequin duck surveys of the Oldman River Basin in 2000, by D. Paton. (2000) �No. 21 Proposed protocols for inventories of rare plants of the Grassland Natural Region, by C. Wallis. (2001) No. 22 Utilization of airphoto interpretation to locate prairie rattlesnake (Crotalus viridis viridis) hibernacula in the South Saskatchewan River valley, by J. Nicholson and S. Rose. (2001) No. 23 2000/2001 Progress report on caribou research in west central Alberta, by T. Szkorupa. (2001) No. 24 Census of Swift Fox (Vulpes velox) in Canada and Northern Montana: 2000-2001, by A. Moehrenschlager and C. Moehrenschlager. (2001) � https://cpw.cvlcollections.org/files/original/d9447e904b35deb0c6d483431e629460.pdf 5e864e8d48ac79f3119980920b0ff7f0 PDF Text Text Population Census of Reintroduced Swift Foxes in Canada and Northern Montana 2005/2006 By Axel Moehrenschlager and Cynthia Moehrenschlager Centre for Conservation Research �Population Census of Reintroduced Swift Foxes (Vulpes velox) in Canada and Northern Montana 2005 / 2006 By 1 2 Axel M oehrenschlager and Cynthia M oehrenschlager 1 Centre for Conservation Research, Calgary Zoo, 1300 Zoo Road N.E., T2E 7V6, Calgary, Alberta, Canada; Email: axelm@calgaryzoo.ab.ca 2 Wildlife Preservation Canada, RR 5, 5420 Highway 6 North, Guelph, Ontario, N1H 6J2, Canada; Email: cyns3@telus.net Funded by: �2 Published by: Centre for Conservation Research, Calgary Zoo Copyright: © 2006 Centre for Conservation Research, Calgary Zoo and Wildlife Preservation Canada This publication may be cited as: M oehrenschlager, A. and C. M oehrenschlager. 2006. Population census of reintroduced swift foxes (Vulpes velox) in Canada and northern M ontana 2005/2006. Centre for Conservation Research Report No. 1, Calgary Zoo. Calgary, Alberta, Canada. For copies of this report, contact: Centre for Conservation Research Calgary Zoo 1300 Zoo Road N.E. Calgary, Alberta T2E 7V6 Email: axelm@calgaryzoo.ab.ca Phone: (403) 232-7771 Fax: (403) 232-9370 www.calgaryzoo.com Printed on 100% recycled paper. �3 This report is dedicated to the late Wayne Harris, whose love for wildlife and commitment to conservation served as an example for all of us to follow. Among many impressive leadership roles, Wayne was a passionate member of the National Swift Fox Recovery Team. During our swift fox census work, Wayne was sorely missed, but never forgotten. �4 Ta ble of Contents Acknowledgments ...................................................................................................................... 6 Executive Summary................................................................................................................... 7 1.0 Introduction .......................................................................................................................... 9 1.1 Species Description............................................................................................................ 9 1.2 Historical Distribution And Abundance ............................................................................10 1.3 Canadian Swift Fox Reintroduction...................................................................................10 1.4 Population Status (1996 – 2001)........................................................................................11 1.5 Census Objective (2005 – 2006)........................................................................................12 2.0 Methods.................................................................................................................................13 2.1 Study Area.........................................................................................................................13 2.2 Training.............................................................................................................................14 2.3 Catch And Release Trapping............................................................................................14 2.4 Data Analyses ...................................................................................................................15 3.0 Results ...................................................................................................................................18 3.1 Catch And Release Sampling Intensity.............................................................................18 3.2 Geographic Distribution And Connectivity......................................................................18 3.3 Population Composition....................................................................................................20 3.4 Population Abundance......................................................................................................23 4.0 Discussion..............................................................................................................................27 5.0 Literature Cited....................................................................................................................30 �5 Lis t of Table s Table 1: Comparative capture rates of swift foxes on 146 townships that were surveyed in 2000/2001 and re-surveyed in 2005/2006 ..................................................................23 Table 2: Area-specific success of trapping at least one fox on replicated townships in 2005/2006 relative to fox presence in 2000/2001 .......................................................24 Table 3: Survey effort, captures, estimated densities, and estimated population sizes for the Alberta/Saskatchewan Border, Grasslands, M ontana, and combined survey areas in 2005/2006.......................................................................................................25 Lis t of Maps Map 1: Study area depicting the regions designated as the Alberta/Saskatchewan Border Region (A), the Grasslands National Park Region (B) in Canada and the M ontana Region (C) in the United States...................................................................................13 Map 2: Trap locations, numbers of individual foxes caught, and incidental sightings of foxes in the ‘Border’ region ........................................................................................19 Map 3: Trap locations, numbers of individual foxes caught, and incidental sightings of foxes on townships in the ‘Grasslands’ region............................................................20 Map 4: Trap locations, numbers of individual foxes caught, and incidental sightings of foxes on townships in the M ontana region..................................................................21 Map 5: Trap and capture locations for the entire 2005/2006 census area. Dotted line illustrates extent of occurrence....................................................................................22 Map 6: Capture success in 2005/2006, representing townships with captures with an ‘X’, townships without captures with an ‘O’, and townships with an ‘O’ and a single solid dot that had no captures but did have a swift fox sighting. Background shading represents relative fox presence in 2000/2001 ..........................26 �6 Ack nowle dgme nts The dedication and commitment of many individuals and organizations has allowed for the successful completion of the 2005/2006 international swift fox census in Canada and M ontana. Census coordination is attributable to Pat Fargey (Parks Canada), Ryan Rauscher (M ontana Fish, Wildlife, and Parks), Cynthia M oehrenschlager (Wildlife Preservation Canada), and Axel M oehrenschlager (Calgary Zoo). A remarkably dedicated crew of field research staff spent countless hours in challenging winter conditions to survey swift foxes carefully and effectively. Thank you to Ken Cheek, Corey Corbett, Leah Darling, Vanessa Fogal, M arcie Gareau, Christy Hansen, Janis Hooge, Dru Osterhout, and Chris Reed of the Calgary Zoo Centre for Conservation Research and Cody Dix, Becky Filipowicz, Scott Lind, Nick Toth, and Kirk Perseyk of M ontana Fish, Wildlife, and Parks. At the Calgary Zoo we would like to thank Wanda Angermeyer, Dr. Sandie Black, Kati Hrynewich, Lynn Klassen, Glenda M isurelli, Bob Peel, Lori Rogers, Kathy Twamley, and Dr. Doug Whiteside for excellent veterinary and animal care as well as proficient instruction of animal care procedures using the zoo’s swift foxes. Thank you as well to Jon Harrison and Ted Trewella for instruction regarding adequate safety protocols, to Diane Casimir and to Dr. Sherry Rainsforth regarding human resource issues. Logistical planning and endless support was received from Bill Bristol and Don Sweet (Prairie Farm Rehabilitation Association), Ursula Banasch, Dave Duncan and Paul Gregoire (Canadian Wildlife Service), Dr. Bob Fischer (Valley Pet Hospital), Brian Giddings (M ontana Fish, Wildlife and Parks), M aple Creek Veterinary Services, Sue M cAdam (Saskatchewan Environment and Resource M anagement), and Joel Nicholson (Alberta Fish & Wildlife). Traps were provided by the Calgary Zoo, Canadian Wildlife Service, Grasslands National Park, and M ontana Fish, Wildlife and Parks. Traps were lined by Ken Cheek, Corey Corbett, and M arcie Gareau (Calgary Zoo Centre for Conservation Research) and preliminary trap designs were tested by David Ausband (University of M ontana). M aps in the report were prepared with the assistance of Typhenn Bricheri-Colombi and Dr. Shelley Alexander (University of Calgary). We are extremely grateful to all of the landowners for permission to trap on their land, continual support to assist the swift fox, knowledge of swift foxes, and most of all for their continual generosity and hospitality. Additionally, we are thankful to all of those who housed the trapping crews for short or long periods throughout the four months. Our special thanks to Chris Reed for his artistic creation of an international census 2005/2006 swift fox ceiling tile which now dons the ceiling of the Consul, Saskatchewan bar. Thank you Consul for displaying a part of your history and heritage to locals and visitors. �7 Exe cutive Summary The swift fox (Vulpes velox) is a rare, house cat-sized carnivore that can race across native prairie at speeds of up to 60 km/hr. Although swift foxes were once so abundant in Canada that 117,025 were trapped between 1853 and 1877, this species was extirpated from Canada and northern M ontana by the late 1930s. Since 1983, reintroduction efforts have been underway to restore this imperilled carnivore to sustainable levels. Comparative surveys revealed the following demographic dynamics between 1997 and 2001: 1) the abundance and distribution of the Canadian swift fox population had increased substantially; 2) Canadian reintroductions had helped reestablish swift foxes in many regions of M ontana; 3) the previously fragmented population appeared to be connected; 4) the population had changed towards a female-biased sex ratio, thereby increasing effective population size; and 5) the core of the population had a high rate of exposure to canine distemper and parvovirus. By 2005, it was unclear whether the reintroduced swift fox population would be able to sustain its 2001 status, or if changes in population abundance, distribution, or composition would occur. This report outlines the results of the third comprehensive catch-and-release census of reintroduced swift foxes since 1996. A study area of 237 townships was sampled through catchand-release trapping from October 18, 2005 to February 17, 2006 in southeastern Alberta, southwestern Saskatchewan and north-central M ontana. In total, 196 foxes were caught and released: 84 in the Canadian Border Region, which spanned from M anyberries, Alberta to Frontier, Saskatchewan; 20 in the Canadian Grasslands Region, which spanned from Frontier, Saskatchewan to the eastern periphery of Grasslands National Park, Saskatchewan; and 92 in the M ontana Region, which was contiguous with the Canadian survey area. The sex ratio of foxes was balanced and foxes had higher body weights than those captured in 2000/2001. For the first time, zero released foxes from the original reintroductions were caught and 100% of captured foxes were wild-born. Of 146 replicated townships spanning an area of 2 21,954 km , the proportion of townships with fox captures increased from 39.7% in 2000/2001 to 52.1% during this census. This increase was primarily driven by an expansion of foxes in M ontana, where foxes were found on over half of the townships that did not have captures previously. There was a statistically significant increase in fox abundance, as the number of foxes caught on replicated townships increased by 30.5% from 139 in 2000/2001 to 181 during this census. When townships that had not been surveyed previously were included, the total number of townships with swift fox captures increased by 38.7% from 62 in 2000/2001 to 86 during this census. Population connectivity improved as all townships with fox captures were separated by no more than one township, whereas 2000/2001 captures were separated by up to three townships. Canadian fox numbers and densities were similar to the high levels documented in 2001 with a �8 2006 population abundance estimate of 647.3 foxes that indicates population stability, while M ontana numbers increased significantly to an estimated 515.2 foxes, for a combined 2006 population estimate of 1162.5 foxes. Using IUCN (World Conservation Union) indicators of population trend, abundance, and geographic distribution, a preliminary assessment of IUCN threat categories suggests that the Canadian and M ontana populations would respectively qualify for a ‘Vulnerable’ status. However, it is the discretion of federal, provincial, and state agencies to decide what the equivalent protection status and management regime would be for swift foxes in each jurisdiction. Population status is entirely dependent upon the impact of threats. Potential degradation of primary habitat areas, potential mortality increases due to poisoning or trapping, and potential increases in deaths due to increased vehicle traffic in core population zones could still have negative impacts on the population. While further conservation actions are still needed, the results of the 2005/2006 swift fox census indicate that this is the most successful reintroduction of a nationally extinct carnivore to date. This success story is one of cooperation between Canadians and Americans, government and non-government agencies, biologists and landowners. These results show that the recovery of endangered species across geographic and social borders is possible if science, management, and stewardship are utilized effectively. �9 1. 0 I ntroduction 1.1 Species Description Swift foxes are primarily known for their small size, averaging 2.4 kg (M oehrenschlager and Sovada 2004), and their fast speed, as they can run at speeds exceeding 60 km/hr. Swift foxes are among the most den-dependent canids and, unlike most others, depend on dens every day of the year (Kilgore 1969, Egoscue 1979, Hines 1980, Pruss 1999, Tannerfeld et al. 2003). Swift foxes will excavate their own dens and modify the burrows of other species. Dens serve several functions, such as providing escape cover from predators, protection from extreme climate conditions in both summer and winter, and shelter for raising young. Swift foxes are primarily monogamous (Kilgore 1969) although additional females that act as helpers in raising pups are occasionally observed at den sites (Kilgore 1969, Sovada et al. 2003, Tannerfeld et al. 2003). Swift foxes are monoestrus and the timing of breeding is dependent upon latitude. Breeding occurs from December to January in Oklahoma (Kilgore 1969) and in M arch among wild and captive Canadian foxes (Pruss 1994, M oehrenschlager 2000). The mean gestation period is 51 days. Average litter sizes of 2.4−5.7 have been reported based on counts of pups at natal dens. In Colorado, litter sizes were greater for mated pairs with helpers than for those without (Covell 1992). Pups open their eyes at 10−15 days, emerge from the natal den after approximately one month, and are weaned at 6−7 weeks of age (Kilgore 1969, Hines 1980). Both parents provide for the young who remain with the adults for 4−6 months (Rongstad et al. 1989, Covell 1992), which is longer than other North American canids. Swift fox dispersal generally commences in August in Canada (Pruss 1994, M oehrenschlager 2000), September/October in southern Colorado (Covell 1992), and October/November in Kansas (Sovada et al. 2003). As many as 67% of juveniles in Canada still remained in natal home ranges during the 1.5 months that preceded the subsequent breeding season. By comparison, all foxes that were 18 months or older had dispersed from natal territories (M oehrenschlager 2000). Average dispersal distances for swift foxes in Colorado, Kansas, and Canada are generally less than 15 km (M oehrenschlager et al. 2004). Swift foxes are opportunistic foragers, feeding primarily on a variety of small mammals, but also birds, insects, plants, and carrion (Kitchen et al. 1999, Sovada et al. 2001, M oehrenschlager et al. 2 2004). Swift fox home range sizes, calculated with adaptive kernels, vary from 7.6 km on the Piñon Canyon M aneuver site in Colorado to 40.8 km2 in Alberta and Saskatchewan (Kitchen et al. 1999, M oehrenschlager et al., in press). Andersen et al. (2003) found that neighbouring foxes of the same sex, excluded each other from core home range areas. Pechacek et al. (2000) and Sovada et al. (2003) found that adjacent family groups had minimal home range overlap and exclusive core areas. Annual mortality rates of swift foxes range from 0.44−0.63 for adults and 0.31−0.87 for juveniles throughout North America (M oehrenschlager et al 2004). The annual proportion of coyote kills ranged from 0.17−0.56 in Canada, while the proportion of eagle kills ranged from �10 0.22−0.75 (M oehrenschlager et al., in press). Of 23 Canadian foxes that were radio-collared in Alberta and Saskatchewan in January 1995, one could not be relocated after September 1997 and the remaining 22 were known to be dead by February 1998. The combination of relatively high mortality rates, large home range sizes, and low dispersal distances raises questions whether the reintroduced swift fox population in Alberta/Saskatchewan/M ontana can maintain high densities, what the potential for population growth might be, and how quickly the population distribution can change. 1.2 Historical Distribution and Abundance The swift fox is native to the short-grass and mixed-grass prairies of the Great Plains in North America (Egoscue 1979). On the northern limit of its range, swift foxes were present in the Canadian provinces of Alberta, Saskatchewan, and M anitoba. The southern species boundary was New M exico and Texas in the United States. Historical records also exist for areas in M ontana, Wyoming, North Dakota, South Dakota, Nebraska, Kansas, Colorado, and Oklahoma. Naturalists and explorers considered swift foxes abundant within their historical range, and 10,614 swift fox pelts were traded by the American Fur Company at the M issouri and Sioux outfits between 1835 and 1838 (Hillman and Sharps 1978) while only 1989 red fox and 108 gray fox furs were traded during the same period. Recent estimates for the United States suggest that swift foxes are located in 39−42% of their historic range depending on conservative versus liberal estimates of historic range and the time span of records that are considered. As such, the conservative estimate, based on the relative presence or absence of swift foxes in counties throughout individual states, is that swift foxes are 2 2 distributed across 505,149 km while the liberal estimate is 607,767 km (Sovada and Scheick 1999). Before European settlers arrived, swift foxes were found in Canada from the Pembina Hills in M anitoba across southern Saskatchewan to the southern foothills of the Rocky Mountains in Alberta (Carlington 1980). A total of 117,025 swift foxes were traded by the Hudson’s Bay Company between 1853 and 1877 (Rand 1948), a yearly average of only 508 swift fox pelts were taken between 1922 and 1925 (Carlington 1980), the last Canadian museum specimen was collected in 1928 near Govenlock, Saskatchewan, and the last sighting was made near M anyberries, Alberta in 1938 (Soper 1964). The species was officially designated as extirpated from Canada in 1978 (COSEWIC 1978). 1.3 Canadian Swift Fox Reintroduction In 1973, two swift fox pairs from Colorado were sent to the home of the Smeeton family near Cochrane, Alberta (Herrero et al. 1991) where both pairs bred in 1974 (Herrero et al. 1986). In 1976, Dr. Steve Herrero of the University of Calgary and the Smeetons agreed to assess the feasibility of reintroducing swift foxes to Canada. The political and biological aspects of this potential reintroduction were consequently investigated with the assistance of three University of Calgary graduate students (Carlington 1980, Reynolds 1983, Schroeder 1985). In 1983, the first �11 releases of captive-bred foxes were attempted in Alberta (Reynolds 1983) followed by releases in Saskatchewan in 1984, and agreements were subsequently signed between federal and provincial governments to delineate responsibilities for further swift fox reintroductions (Pruss 1994). Swift foxes were released annually from 1983 until 1997. A feasibility study was completed in 1993 which concluded that, based on previous successes, a self-sustaining population of swift foxes could be established and that the most effective method of achieving this goal would be to conduct and monitor 3−5 more years of swift fox releases (Brechtel et al. 1993; Carbyn et al. 1994). Swift fox reintroductions occurred from 1983–1996 in the Alberta/Saskatchewan border area and from 1984–1997 in the Grasslands National Park region of south-central Saskatchewan. An attempted reintroduction into the M ilk River ridge region in southern Alberta was discontinued because of a rabies outbreak in the skunk population. In total, 942 foxes were released through captive-breeding and translocation from the United States. M oreover, 13−15 captive-bred swift foxes were released on the Blood Indian Reserve in southwestern Alberta in 2004. Concurrent to the Canadian reintroduction phases, the number of swift fox reports in northern Montana began increasing and Zimmerman (1998) found that swift foxes were establishing territories and breeding successfully in northern counties of the State. Reintroductions of swift foxes on the Blackfoot Indian Reserve in M ontana were conducted from 1998−2002 through the collaboration of the Blackfoot Nation, Defenders of Wildlife, the Cochrane Ecological Institute, and M ontana Fish, Wildlife, and Parks. These releases have established a small population of swift foxes (Ausband 2005) on the reserve that has the potential of eventually joining the contiguous Alberta/Saskatchewan/M ontana swift fox population. 1.4 Population Status (1996 – 2001) As one part of the ongoing recovery program, a national swift fox census was conducted in Alberta and Saskatchewan during the winter of 1996/1997 to determine the composition, distribution, and abundance of Canada’s swift fox population. At the time, the population was located in two distinct subpopulations; the first, which spanned the Alberta/Saskatchewan border, was estimated to have 192 foxes while the second, which was centred around Grasslands National Park in south-central Saskatchewan was estimated to contain 89 foxes (Cotterill, 1997). Of trapped foxes, 81.3% were born in the wild while the remainder were either captive-bred or translocated foxes that had been reintroduced. A second census that replicated the methodology of the 1996/1997 survey was conducted in 2000/2001. During the 2000/2001 census 149 swift foxes were live-trapped of which 98.6% were wild born. In M ontana, which had not been previously surveyed, 31.8% of townships had swift foxes totalling captures of 38 individuals. The known total distribution of townships containing swift foxes increased from 25 in 1997 to 76 in 2001, showed a male-biased sex ratio, the number of foxes caught on the same townships at the same time of year increased threefold compared to 1996/1997, and the previously fragmented population showed signs of initial connectivity (M oehrenschlager and M oehrenschlager 2001). �12 1.5 Census Objective (2005 – 2006) The long-term recovery goal of the Canadian Swift Fox Recovery Strategy (Pruss et al., in review) is “By 2026, restore a self-sustaining swift fox population of 1000 or more mature, reproducing foxes that do not experience a greater than 30% population reduction in any 10-year period”. The associated short term recovery goal over the next five years is to “Ensure a mature reproducing swift fox population size of at least 250 foxes by 2011”. Key draft recovery objectives over the next five years necessitate assessments of swift fox abundance and spatial distribution, a quantitative assessment of long term population viability, identification of habitat quantity and spatial configuration required to achieve short and long term population goals, and an initiation of swift fox habitat securement (Pruss et al., in review). A connecting requirement for the achievement of these objectives is an improved understanding of the distribution, population trend, and habitat use of the swift fox population. Such evaluations are also crucial for recovery efforts in M ontana where population status information is needed to guide additional research, potential translocation efforts, or potential fur harvest considerations. To address these needs, the objective of this 2005/2006 census was to determine changes in the distribution and abundance of swift foxes in Canada and M ontana, relative to the 1996/1997 and 2000/2001 censuses. Since the 2001 report (M oehrenschlager and M oehrenschlager 2001) evaluated changes between the first and second census, the primary focus of the current report is to document demographic changes from 2000/2001 to the present. �13 2. 0 Me thods 2.1 Study Area The study area spanned the borders of Alberta, Saskatchewan and M ontana (48o24’N−49o31’N; 106o33’W−110o50’W), roughly from M anyberries, Alberta in the west to Glentworth, Saskatchewan in the east, Eastend, Saskatchewan in the north to Dodson, Montana in the south (M aps 1, 5). In the early 1990s, the Canadian swift fox population was located in two distinct subpopulations (Cotterill 1997). To track demographic changes relative to previous censuses, the Canadian portion of the study area was divided into a ‘Border Region’, which ranged from M anyberries, Alberta to Frontier, Saskatchewan and a ‘Grasslands National Park Region’, which ranged from Frontier, Saskatchewan to Glentworth, Saskatchewan; the ‘M ontana Region’ was maintained as a single region which was previously designated (M oehrenschlager and M oehrenschlager 2001) and which had not had swift fox releases (M ap 5). Map 1: Study area depicting the regions designated as the Alberta/ Saskatchewan Border Region (A), the Grasslands National Park Region (B) in Canada and the M ontana Region (C) in the United States. Trap lines depict 2005/2006 trap locations, white townships were randomly excluded during subsampling, solid shaded townships were trapped in 2000/2001 with successful swift fox captures, and striped shaded townships did not have captures in 2000/2001. �14 2.2 Training Four Canadian and two American teams consisting of two people each were trained at the Calgary Zoo and in the field. Through the collaboration of Canadian and American field coordinators, teams were instructed in GPS and map use, field safety and survival techniques, data management, and landowner relations at the Calgary Zoo from October 11–October 15, 2005. All individuals had an opportunity to handle two Calgary Zoo swift foxes while receiving animal care instructions from a Zoo veterinarian regarding manual restraint, body condition assessments, pain management, and emergency aid procedures. Study permits were obtained from Alberta Fish & Wildlife, Canadian Wildlife Service, Saskatchewan Environment and Resource M anagement, Prairie Farm Rehabilitation Administration, Calgary Zoo Biological Research Review Committee, and M ontana field work was sanctioned by M ontana Fish, Wildlife and Parks. From October 18–October 23 all trapping crews moved to Consul, Saskatchewan. During this time, 14 foxes were caught on high-density PFRA townships and all teams were given handling experience until they felt comfortable to work on their own. After this time, trapping teams split into respective jurisdictions. 2.3 Catch and Release Trapping The original Canadian swift fox census area was determined by the National Swift Fox Recovery Team in 1996, based on habitat criteria and background information (Cotterill 1997). The suspected swift fox range consisted of 65 townships in the Alberta/Saskatchewan border region and 43 townships in and around Grasslands National Park in south-central Saskatchewan. Of these 108 townships, 81 (75%) were randomly selected for subsequent sampling, of which two were excluded at the request of the resident landowners. Similarly, a census area of 80 townships was selected by M ontana Fish, Wildlife and Parks staff in 2000, and 75% of these townships were selected for subsequent sampling. During the current census, the targeted census area was further expanded by 49 townships on the periphery of the 2000/2001 area of swift fox occupancy. These townships were adjacent to townships with 2000/2001 captures to assess potential population expansion (M ap 1). Trapping priorities for the current census with diminishing importance were: 1) Resample townships that had been trapped in both 1996/1996 and 2000/2001; 2) Resample townships that had only been surveyed in 2000/2001; and 3) Sample additional peripheral townships that had not been trapped before. Swift fox trapping was conducted October 18, 2005−February 13, 2006 in Alberta, Saskatchewan, and M ontana using replicated methodology from Cotterill (1997) and M oehrenschlager and M oehrenschlager (2001). Individual live-traps were placed at onekilometre intervals along a five kilometre continuous section of the trail closest to the centre of respective townships. The inter-trap distance was adjusted by up to 100 meters to allow trap placement along fences or on top of hills. Each township was surveyed with six traps for three nights, for a total of 18 trap nights per township. Catch-and-release trapping was conducted on consecutive nights when possible, but this was dependent on weather conditions. �15 Two sizes of fox live-traps were used for catch-and-release: large traps were 109 cm x 39 cm x 39 cm Tomahawk (Tomahawk Live Trap Co., Tomahawk, WI) double-door; or 109 cm x 39 cm x 47 cm Safe Guard (Safe Guard Products Inc., Lancaster County, PA) traps; small traps were 83 cm x 31 cm x 31 cm Tomahawk single door box traps; or 78 cm x 29 cm x 31 cm Safe Guard traps. All traps were lined with 3 mm hardboard, as this has been shown to reduce injuries (M oehrenschlager et al. 2003). Unlike before, Canadian traps were lined completely with hardboard that was custom made with small holes; the floor, ceiling, and sides were lined with hardboard while the door and back were lined with plexiglass so that captured foxes could be seen easily. In contrast, traps in M ontana were lined only on the floor, lower corners, and lower portions of the door, which is the way that traps were lined previously in Canada and M ontana (Cotterill et al. 1997, M oehrenschlager and M oehrenschlager 2001). Catch-and-release trapping was conducted at night to reflect fox activity periods, to avoid heat-stress, and to prevent fox disturbance by people. Traps were generally set between 1800 hrs and 2000 hrs, checked between 2400 hrs and 0200 hrs, and closed following a second check between 0600 hrs and 0800 hrs. Trapping was not conducted at temperatures colder than –20oC or when snow, rain, and wind conditions were potentially hazardous to captured foxes (as in M oehrenschlager and M oehrenschlager 2001). Foxes were coaxed into a denim handling bag which was placed over the end of the box trap. Every team alternated between four handling bags, which were washed frequently to reduce the transmission of fleas between foxes. Foxes were handled by both field workers of respective trapping teams. The first team member positioned the animal on his/her lap to shelter it from the wind, one hand restrained the head and covered the eyes, and the second hand restrained the body. The second field worker sexed and aged the fox, conducted parasite counts, scored body condition through palpation on a standardized index ranging from 1 (poor) to 5 (excellent), collected a hair sample for genetic analyses, checked for injuries, and tattooed the ear for identification. Foxes were uniquely marked with tattoo dye so that recaptured individuals could be easily identified. Swift foxes emerge at approximately three weeks of age and emergence dates in the Canadian swift fox population range from M ay 25–June 9 (Pruss 1994, M oehrenschlager 2000). Given the variation in parturition dates and the extent of the census trapping period, captured juveniles were likely 5–10 months old. Accordingly, adults were 17.5 months or older. Incidental sightings of swift foxes were recorded, but no systematic snow-tracking or spotlighting sampling regimes were pursued so that the primary emphasis for trapping teams could be the catch-and-release efforts. 2.4 Data Analyses IUCN criteria, which are used for species assessments globally and can be applied regionally (IUCN 2001), involve estimations of geographic distribution, population trend, and population abundance. The first relevant geographic criterion the IUCN utilizes is the ‘Extent of Occurrence’ which is ‘defined as the area contained within the shortest continuous imaginary boundary which can be drawn to encompass all the known, inferred or projected sites of present �16 occurrence of a taxon, excluding cases of vagrancy’ (IUCN 2001). Since a fox caught on the periphery of the population will range further than the site of its capture, extent of occurrence in the context of this census was calculated as the outside minimum convex polygon that envelops all townships containing swift fox captures. Since extent of occurrence calculations consist of connections among the most peripheral occurrence points, the sum of the smaller region extents of occurrence do not total to the overall value of the entire population (see IUCN 2001). The second relevant estimate of geographic distribution is the ‘Area of Occupancy’, which is ‘defined as the area within its ‘Extent of Occurrence’ which is occupied by a taxon, excluding cases of vagrancy. The measure reflects the fact that a taxon will not usually occur throughout the area of its extent of occurrence, which may contain unsuitable or unoccupied habitats.’ Areas of occupancy are estimated using grid systems (IUCN 2001), and assessments are influenced by grid scale. Decisions of grid scale for estimations of area of occupancy are species-specific and dependent on individual sampling regimes (IUCN 2001). Depending on utilized kernel estimators, swift fox 2 2 home ranges in Alberta and Saskatchewan envelop 31.8 km −40.8 km , mates share 74.3% of their home ranges, and neighbours share 28.8% of home ranges (M oehrenschlager et al. in press). The presence of one fox in a 92.16 km2 township consequently does not necessarily mean the entire township is occupied, but, depending on the distribution of captures and individual home range size variation, two or three unmated adult foxes can occupy a township. The number of foxes in a township that contains at least one capture, will be underestimated by the census methodology because 32.9% of foxes will not be trapped by three nights of trapping, even if the traps fall inside the foxes’ home range (Cotterill 1997). For the same reason, some townships with low densities of foxes will not have any captures although foxes are present. Given these respective trade-offs, which can alternatively balance overestimates or underestimates of areas of occupancy, and the fact that the township scale was chosen to provide sample block independence for the census sampling (Cotterill 1997), the township scale was utilized as the grid scale for estimating areas of occupancy. Within the respective Border, Grassland, and M ontana Regions, the area of occupancy is restricted to respective extents of occurrence and, within these areas approximately 75% of the areas were sampled due to the original subsampling regime. Sampled townships that had captures contributed their areas to the area of occupancy, sampled townships that did not have captures excluded their entire area from the area of occupancy or, if the extent of occurrence polygon subdivided a peripheral township, a portion of the township area was excluded from the area of occupancy. The proportion of occupied vs. unoccupied areas within sampled regions of the extent of occurrence was used to estimate occupancy on unsampled townships. �17 Consequently, area of occupancy was calculated using the following equation: AOcc = APresent + [(APresent / ANone )* ANoSample ] Where: AOcc APresent ANone ANoSample = Area of occupancy = Total area of townships where at least one fox was captured = Total area of townships without captures within extent of occurrence = Area within extent of occurrence of townships where no trapping was conducted In terms of evaluating population trend, we concentrated primarily on comparative capture rates on townships in Alberta, Saskatchewan and M ontana that were sampled in the same season and at the same sites as the 2000/2001 census. The number of trap nights per fox capture are presented for each region relative to the previous census. The number of foxes caught per township were compared using paired Wilcoxon signed ranked tests with the township as the sample unit for replicated comparisons. Differences in numbers of foxes caught were compared for the census population overall, and then for each of the subregions. The sex ratio of captured foxes was compared between regions and relative to that of the previous census using Pearson chi-square. Fox body weights were compared to the 2000/2001 census using independent sample t-tests for adult and juvenile foxes respectively. For consistency, fox abundance was estimated for the three sample regions and the overall population using the same techniques employed in Cotterill (1997) and M oehrenschlager and M oehrenschlager (2001). Since recapture rates of individual foxes are high within the three night trapping period, maps and statistics of fox numbers in the individual sample regions and the overall population only involve the first capture of individual foxes within a township. �18 3. 0 Re s ults 3.1 Catch and Release Sampling Intensity The total area that encompassed the Border, Grasslands, and M ontana Regions spanned 237 townships (21,742 km2) and trapping was conducted in 191 (80.6%) of these townships containing 17,534 km2 (M aps 1, 6). The total study area increased by 26.1% from 188 townships in 2000/2001 and the sampling area increased by 26.5% from 151 townships in 2000/2001. Of 151 townships that were trapped for three nights in 2000/2001, 146 were replicated during the present census. Specifically, the sampling was conducted as follows: 2 1) Alberta/Saskatchewan Border: Spanning 78 townships (7,089 km ) that contained nine townships which were excluded randomly during the previous census and consequently excluded during this census; of the remaining 69 townships, 62 were trapped containing 5,645 km2. 2) Grasslands National Park Region: 53 townships (4,884 km2) that contained 12 previously excluded townships; of the remaining 41 townships, two that were not trapped in 2000/2001 were not trapped again and one that was surveyed in 2000/2001 was not surveyed; the remaining 38 townships contained 3,502 km2. 3) M ontana Region: Spanning 106 townships (9,769 km2) that contained 15 previously excluded townships and two townships that had been sampled outside the original 2000/2001 survey area; 91 townships were trapped containing a total area of 8,387 km2. In the Saskatchewan portion of the Border region, census trapping was closely coordinated with a University of M anitoba/Parks Canada graduate research study which is developing DNA faecal censusing techniques relative to success in live-capture techniques. DNA has successfully been isolated from summer and winter samples (Curteanu, pers. comm.) on Prairie Farm Rehabilitation Agency (PFRA) townships that had had swift fox captures during the current census. Hair samples were collected from captured swift foxes to complement these genetic assessments. Serological investigations have not been completed at this point, but plans are in place to do so. Trap level and trap-line level GIS layers of swift fox captures have been constructed for the 2005/2006 data set by University of Calgary and Calgary Zoo researchers (Alexander and Brichieri Colombi, pers. comm). 3.2 Geographic Distribution and Connectivity The known swift fox distribution in Canada and M ontana has significantly increased, in part because of a greater survey area and also because of a greater occurrence of swift foxes in resurveyed regions. Similar to 2000/2001, swift foxes were trapped in 32 townships in the Border Region in 2005/2006. In the Grasslands Region, swift foxes were trapped in eight townships in 2005/2006 compared to only six townships in the previous census. In M ontana, 46 townships had fox captures in 2005/2006 compared to only 21 townships in 2000/2001. Consequently the total �19 known distribution, confirmed through catch-and-release, has increased by 38.7% from 62 to 86 townships. The largest gap between townships with captures was one township in the Border Region (M ap 2), two townships in the Grasslands Region (M ap 4), and two townships in M ontana (M ap 4). However, through the combination of the Canadian and M ontana regions (M ap 5), the largest gap between capture townships decreases to just one township for the entire census population. Since the 2000/2001 census, a connective corridor adjacent to the international border has been established in M ontana with a minimum span of ranges 20–29 on the M ontana grid (M aps 4, 6). Map 2: Trap locations, numbers of individual foxes caught, and incidental sightings of foxes in the ‘Border’ region. Extent of occurrence and area of occupancy are presented separately for each of the three study regions in Table 3. While M ontana, the Canadian Border population, and the Canadian Grasslands region differ in extent of occurrence, these regional classifications, which represent former population fragments (Cotterill 1997, M oehrenschlager and M oehrenschlager 2001), are now increasingly arbitrary given the great connectivity between these zones. Nevertheless, since one key cut-off point for determining levels of threat under the IUCN Red List approach (IUCN 2 2001) is a minimum extent of occurrence of 20,000 km , it is worth noting that the Canadian extent of occurrence based on captures alone is 14,038 km2, the M ontana extent of occurrence of the entire population is 11,495 km2 and the combined extent of occurrence is 21,954 km2. In �20 contrast to extent of occurrence, area of occupancy is additive so the combined area of occupancy in Canada is 6,343 km2, while that in M ontana is 6,606 km2; in both jurisdictions a critical IUCN threshold of 2,000 km2 has been crossed for area of occupancy. Map 3: Trap locations, numbers of individual foxes caught, and incidental sightings of foxes on townships in the ‘Grasslands’ region. 3.3 Population Composition In total, 196 swift fox individuals were captured and released. Foxes were trapped with the following frequencies: • 84 in the Border Region, 20 in the Greater Grasslands Region, and 92 in M ontana; • Two foxes were caught on more than one township; • In the Border Area, one fox was recaught on neighbouring township (original capture: Township 4, range 1; recapture: Township 4, range 30; M ap 2) and a second fox was caught in the Border Area and then recaptured in M ontana (original capture: Township 2, range 27; recapture Township 35, Range 17; M aps 2, 4). �Map 4: Trap locations, numbers of individual foxes caught, and incidental sightings of foxes on townships in the Montana region. 21 �Map 5: Trap and capture locations for the entire 2005/2006 census area. Dotted line illustrates extent of occurrence. 22 �23 Whereas 1.4% of trapped foxes in 2000/2001 had originally been captive-bred or translocated, 100% of foxes caught in the present census were unmarked, meaning that they were wild born. One stunning finding was the capture of a fox that was born on the Blackfoot Reserve, which is located about 150 km west of the census population in M ontana. A female that was captured on February 13, 2003 and last relocated on February 21, 2003 on the Blackfoot Reserve (Ausband, pers. comm.) was recaptured on November 6, 2005, 18 km west of the Saskatchewan border in township 1, range 2, where two males were also captured. She had travelled a straight-line distance of 191 km. Body weights were greater for this census than 2000/2001 for both adults (2000/2001: 2.30±0.03 kg; 2005/2006: 2.41±0.03 kg; t = 2.8, df = 140, p < 0.01) and juveniles (2000/2001: 2.17±0.02; 2005/2006: 2.25±0.03; t = 2.2, df = 165, p = 0.04). During this census 52.3% of 195 foxes of known sex were male compared to 42.9% males in 2000/2001; this difference was not 2 statistically significant (Pearson χ = 3.0, df = 1, p = 0.08). The sex ratio was also similar between Border, Grasslands, and M ontana sample regions in 2005/2006 (Pearson χ 2 = 1.2, df = 2, p = 0.55). 3.4 Population Abundance Of 146 townships that were replicated from the 2000/2001 census, the percentage of townships that had at least one capture increased from 39.7% in 2000/2001 to 52.1% in 2005/2006; this increase was primarily driven by an increase in fox occupancy from 30.9% in 2000/2001 to 57.4% in 2005/2006 in M ontana (Table 1). Of 88 townships that had no captures in 2000/2001, 36.4% now had captures and, of 58 townships that had had previous captures, 75.9% had captures in 2005/2006 (Table 2). In M ontana, over half of townships that had not had captures previously did have fox captures in 2005/2006, compared to only 19.5% of Canadian townships (Table 2). Table 1: Comparative capture rates of swift foxes on 146 townships that were surveyed in 2000/2001 and re-surveyed in 2005/2006. Region Border (48 townships) Grasslands (30 townships) Montana (68 townships) Total (146 townships) % of Townships with Captures 2000/ 2005/ 2001 2006 Total Number of Foxes Caught 2000/ 2005/ 2001 2006 Trap nights/ New Capture 2000/ 2005/ 2001 2006 58.3 60.4 87 78 10.0 11.1 30.0 26.7 14 20 38.6 27.0 30.9 57.4 38 83 32.2 14.7 39.7 52.1 139 181 19.0 14.5 �24 Table 2: Area-specific success of trapping at least one fox on replicated townships in 2005/2006 relative to fox presence in 2000/2001. 2000/2001 Foxes Present on Township No Yes Overall 2005/2006 % Of Townships With Foxes Present Border Townships 25.0% Grassland Townships 14.3% Montana Townships 51.1% 85.7% 60.4% 55.6% 26.7% 71.4% 57.4% Fox numbers on replicated townships were significantly greater in 2005/2006 than in 2000/2001 (Wilcoxon paired signed rank test; Z = 2.45, p = 0.01), as 2005/2006 fox numbers were higher on 53 townships, equal on 64 townships, and lower on 29 townships. The number of foxes on replicated townships increased by 30.5% from a total of 139 (mean/township = 0.95+0.14) foxes in 2000/2001 to 181 foxes (mean/township = 1.24+0.13) foxes in 2005/2006 (Table 1). This change was primarily due to M ontana regions where the population increased significantly (Wilcoxon paired signed rank test; Z = 3.13, p < 0.01) and the number of trap nights per capture decreased (Table 1), while the Border region (Wilcoxon paired signed rank test; Z = 0.28, p = 0.78) and Grasslands region (Wilcoxon paired signed rank test; Z = 1.00, p = 0.32) did not change significantly in swift fox abundance or the number of trapnights per capture (Table 1). Trapping success of 18.6% was relatively low on 43 primarily peripheral townships (12 Border, 8 Grasslands, and 23 M ontana) that were trapped in 2005/2006, but not in the previous census (M ap 6). The frequency of recapturing foxes within the same trapping interval was higher in Canada in 2005/2006 (64%; Table 3) compared to 2000/2001 (31%; Pearson χ 2 = 22.0, df = 1, p < 0.01). In contrast, in M ontana, the recapture rate was 27.2% in 2005/2006 compared to 31.5% in 2000/2001 (Pearson χ 2 = 0.26, df = 1, p = 0.61; Table 3). The only discernible difference between Canada and M ontana that was not also accounted for in previous censuses, was the difference in trap lining. Canadian foxes were often harder to coax out of the fully lined traps, as the foxes were apparently sheltered and comfortable. Given the relatively positive experience during the first capture, foxes were apparently more likely to re-enter fully lined traps than traps that were not fully lined in M ontana during this census, and in Canada during previous censuses. The total number of townships with fox captures increased from 62 in 2000/2001 (41.0% of 151 surveyed townships) to 86 in 2005/2006 (45.0% of 191 surveyed townships). After excluding one township that was trapped for only 0.5 trap nights in 2000/2001, the number of foxes on townships that had at least one capture did not differ between 2000/2001 (mean = 2.45+0.25; median = 2) and 2005/2006 (mean = 2.30+0.15; median = 2; n2000/2001 = 60; n2005 /2006 = 86; Z = 0.11; p = 0.91), indicating that fox densities have not changed on townships where foxes are present. The maximum number of foxes caught per township increased from three to five in the Grasslands National Park Region and from four to six in M ontana since the last census. While the Border Region still had one township with a higher density (M ap 2), this indicates that highdensity townships are now well-distributed throughout the census region. �25 Table 3: Survey effort, captures, estimated densities, and estimated population sizes for the Alberta/Saskatchewan Border, Grasslands, M ontana, and combined survey areas in 2005/2006. Region Region Region Townships Area Townships Extent of Area of Distinct % of Townships Area with Occurrence Occupancy Foxes Distinct Surveyed sampled 2 2 1 2 (km ) (km ) Captures (km ) (km2) Caught Foxes Recaught Estimated Fox Density (foxes/ 100km2)2 Estimated Population Size3 B order 78 7,089 62 4,061 32 6,536 4,726 84 65.4 7.2 513.2 G rasslands 53 4,884 38 2,550 8 2,654 1,617 20 60.0 2.7 134.1 Montana 106 9,769 91 6,106 46 11,495 6,606 92 27.2 5.3 515.2 Total 237 21,742 191 12,717 86 21,954 15,704 196 47.9 5.4 1,162.5 1 Area sampled = (Area sampled per trapline * #of surveyed townships); (Area sampled per trapline = 67.1 km 2) 2 Estimated Fox Density=[(# Foxes Caught in Sample * Correction Factor)/Area Sampled in Region] * 100 (Correction Factor = 3.5) 3 Estimated Population Size = (T otal Area of Swift Fox Range * Estimated Density) For consistency with past censuses, the 1996/1997 trapability correction factor (Coterill 1997) based on the home ranges and catchability of radio-tracked swift foxes (M oehrenschlager 2000), was applied. For 2005/2006, the Border Region estimate of 513.2 foxes, combined with an estimated population of 134.1 foxes in the Grasslands Region combines to an estimate of 647.3 foxes in Canada (Table 3). This number combined with an estimated 515.2 foxes in M ontana adds to an overall population estimate of 1162.5 foxes for 2005/2006. Since critical habitat protection under the Canadian Species at Risk Act, is primarily pertinent to federal lands, it is worth noting that 38 of 84 (40.4%) fox captures in the Canadian border region were on townships that contain Prairie Farm Rehabilitation Agency (PFRA) land within them. No captures were made on associated Agriculture Canada land at the Onefour Research Station in Alberta, although swift fox observations have been made in the area this summer (Reed, pers. comm.). Since PFRA only operates in Saskatchewan it is worth noting that 38 of 52 (73.1%) foxes caught on Saskatchewan townships in the Border region contain PFRA land. Not all captures were directly in PFRA pastures, but all were within a potential home range diameter of captured home ranges. In the Grasslands Region, 9 of 20 (45%) foxes were from townships that contained PFRA pastures so overall 47 of 104 (45.2 %) of Canadian foxes were caught on townships that contained PFRA pastures. In the Grasslands Region, nine further foxes were trapped on townships containing Parks Canada land so that a total of 90% of the 20 foxes trapped in the Grasslands Region were found on townships containing federal land. Overall then 56/104 (53.8%) of foxes trapped in Canada were located on townships containing federal land. �Map 6: Capture success in 2005/2006, representing townships with captures with an ‘X’, townships without captures with an ‘O’, and townships with an ‘O’ and a single solid dot that had no captures but did have a swift fox sighting. Background shading represents relative fox presence in 2000/2001. 26 �27 4. 0 Dis cuss ion The 2005/2006 census has been the third standardized assessment of the Canadian swift fox reintroduction program in the last ten years. Since swift foxes were sampled using similar methods at a similar time of year during each of the three time censuses, population changes can be evaluated rigorously over time. Like the interval between 1996/1997 and 2000/2001, the population has once again crossed numerous significant thresholds as it grows towards sustainability. Since 2000/2001, the contiguous Alberta, Saskatchewan, M ontana swift fox population has increased in its area of occupancy, extent of occurrence, number of foxes, and the proportion of wild-born individuals in the population. This continued growth indicates that this is the most successful reintroduction of a nationally extirpated carnivore in the world to date. The Canadian swift fox population had undergone large increases in abundance, density, and distribution between 1996/1997 and 2000/2001 (M oehrenschlager and M oehrenschlager 2001), and it was questionable whether this population surge was sustainable over time. The population has indeed maintained its 2001 levels, the primary regional predictor of swift fox presence in 2006 was presence in 2001, and 100% of trapped individuals are now born in the wild. Clearly, the population has been sustaining itself without a contribution of reintroduced individuals. In M ontana, there has been an extensive settlement of areas where foxes were not captured in 2001, and these increases in distribution and abundance have driven significant population growth for the entire Canadian/M ontana population. Increases in fox weight since 2001, were likely caused in part by mild fall and winter weather conditions. Since the body weight of females before the breeding season is correlated to subsequent litter size (M oehrenschlager 2000), the summer of 2006 may have been especially productive for swift fox pup production. Increases in population spread have further improved the connectivity of the swift fox population. In the 2000/2001 census, the eastern border of the study area was apparently split from the rest of the population as there was a gap of three townships between the eastern Grasslands Region/eastern M ontana study areas and the rest of the captures. Due to the original sampling region chosen for M ontana, which excludes a central area that is largely used for cropland farming, there also appeared to be a potential split between eastern and western trapping blocks in M ontana in 2000/2001. While the sampling gap still exists, captures in townships that previously did not have foxes on the periphery of this gap have now increased connectivity in this region as well. No exchange of foxes between the contiguous Canadian/M ontana population and the Blackfoot Reserve in M ontana had been documented previously despite extensive marking of individuals in both populations. The discovery in this census that one fox, which dispersed further than any previously documented swift fox in North America (M oehrenschlager et al. 2004), indicates that continuous genetic exchange may be possible with these populations or, potentially, Wyoming populations in the future. In Canada, the level of oil and gas industrial developments was at unprecedented levels within the census area compared to swift fox field work that has been ongoing in the area since 1993. While the previous development of an oil pipeline in Alberta did not cause discernible mortality �28 or space-use differences, disturbances apparently caused a decline in reproductive success (M oehrenschlager 2000). Logistic-regression GIS habitat modelling of 1996/1997 and 2000/2001 census data clearly show that fox abundance declines as habitat fragmentation increases (M oehrenschlager and Alexander, unpublished data). During the present census, oil and gas industrial development activity was intense in many areas and particularly on some Prairie Farm Rehabilitation Agency (PFRA) townships. On township 2 range 28 which lies in the core of the Alberta/Saskatchewan border region, a 150 man camp was in place to install 200 wells/township and an extensive grid work of trails had been established across an area that was previously relatively pristine. Clearly, such activities are of concern to swift foxes either by degrading habitat quality, disturbing residences, or by increasing the probability of direct mortalities due to vehicles. Since 45% of Canadian foxes were trapped in townships that contained PFRA land, a large challenge and opportunity exists to ensure that developments on Canadian federal land do not impair the potential for Canadian swift fox recovery. As swift foxes are sympatric with numerous other Canadian species-at-risk, the resolution of this opportunity will also be valuable for species such as burrowing owls or sage grouse for example. Assessments of species sustainability by the IUCN (World Conservation Union) utilize population indicators of geographic distribution and animal abundance to determine the Red List status of imperilled species worldwide (IUCN 2001) and these have been applied to the IUCN listing of swift foxes (M oehrenschlager and Sovada 2004). IUCN criteria can also be applied on smaller scales. In Canada, IUCN criteria have also been adapted by COSEWIC (Committee on the Status of Endangered Wildlife in Canada) to determine national levels of threat and the Alberta Endangered Species Conservation Committee applies IUCN guidelines to arrive at province-specific designations. IUCN threat designations are derived through a combination of demographic parameters: population trend, geographic range, population size, and quantitative analysis of extinction probability. These lead to a potential designation within the ‘Threatened’ categories, which are: ‘Critically Endangered’, ‘Endangered’, and ‘Vulnerable’. This census, combined with the previous two censuses yield 10 years of swift fox population trend information in Canada, and it is clear that this population has not experienced an observed, estimated, inferred or suspected population size reduction of > 30% over the last 10 years, which would otherwise qualify the population for at least a ‘vulnerable’ IUCN status. In terms of geographic range, a population will fall into at least a ‘vulnerable’ category if its area of occupancy is less than 2,000 km2, but the Canadian captures clearly exceeded this requirement. A population will classify for at least ‘vulnerable’ status if the extent of occurrence is < 20,000 km2 and estimates for at least two of the following three population parameters are true: 1) severely fragmented or known to exist at fewer than 10 locations; 2) continuing decline in area, extent, or quality of habitat; 3) continuing decline or extreme fluctuations in the extent of occurrence, area of occupancy, number of mature individuals, or number of locations or subpopulations. 2 The Canadian extent of occurrence is below 20,000 km , but the population has improved connectivity through the M ontana population, and the number of individuals in the current census is similar to the high densities and numbers recorded in the 2001 census; however, there �29 may be decreasing habitat quality due to increased resource extraction in some prairie regions. Although the potential decline in habitat quality is of concern, this is only one of the three population parameters considered here and, consequently, the extent of occurrence evaluation would not trigger a ‘vulnerable’ status for Canada at this point. However, the number of mature individuals in Canada is clearly under 1000 mature breeding individuals, which classifies for an IUCN-equivalent listing of ‘vulnerable’. The higher listing category of ‘Endangered’ would only be adequate if the population had fewer than 250 mature breeding individuals. The current assessment meets the Canadian recovery strategy objective of maintaining a population of at least 250 mature individuals for a minimum of five years. While population estimates will also be verified with alternative population size estimates, similar sex ratios, the ability of swift foxes to breed in their first year of life, and the fact that over 600 individuals were estimated in both the 2001 and the current census indicates that the 250 individual mark has clearly been met and maintained. Comparative assessments for M ontana also indicate that the population trend documented through this census and graduate student studies (Zimmerman 1998, Ausband 2005) has been 2 positive over the last ten years. The area of occupancy exceeds 2,000 km , the extent of occurrence including the Blackfoot Reserve population (Ausband 2005), which is severely fragmented from the contiguous Alberta/Saskatchewan/M ontana population, exceeds 20,000 km2 but the total M ontana population is clearly still smaller than 1000 mature breeding individuals, which would also qualify it for an IUCN ‘vulnerable’ status. Fortunately, animals tend to ignore human provincial, state, and country boundaries. Given the current assessments, the combined Canadian/M ontana census population which had a capture 2 area exceeding 20,000 km in extent of occurrence, a positive population trend, and more than 1000 estimated individuals, the population would arguably not fall into any IUCN threatened criteria any longer. However, this finding needs to be interpreted with caution due to uncertainty variance around the abundance estimation, the fact that some individuals will have died before the breeding season which commenced in February/M arch, and the future vulnerability of the population to increases in mortality due to accidental trapping, designated fur harvest, poisoning, increased vehicle traffic, or potential declines in habitat quality or quantity. M ore importantly, species listing and delisting is the responsibility of distinct jurisdictions on provincial, state, or country scales and, as such, the population would need to meet delisting criteria separately in each jurisdiction. Biologically, the results of this survey would not support a status change in Alberta and Saskatchewan, and hence Canada at this point. In M ontana, the increase in swift fox numbers and distribution suggests an excellent recovery potential and the possibility that this expanding population might eventually connect with the rest of the United States contiguous swift fox range. �30 5. 0 Lite rature Cite d Andersen, D.E., Laurion, T.R., Cary, J.R., Sikes, R.S., M cLeod, M .A., & Gese, E.M . (2003). Aspects of swift fox ecology in southeastern Colorado. In: Swift fox conservation in a changing world (eds. M . Sovada & L. Carbyn). Canadian Plains Research Center, University of Regina, Regina, Canada. Ausband, D.E. (2005). Assessing the success of swift fox reintroductions on the Blackfeet Indian Reservation, M ontana. M .S. thesis, University of M ontana, M issoula, M ontana, USA. Brechtel, S.H., Carbyn, L.N., Hjertaas, D., & M amo, C. (1993). Canadian swift fox reintroduction feasibility study: 1989−1992; report and recommendations of the National Recovery Team. Unpublished report, Alberta Fish and Wildlife Services. Carbyn, L. N., Armbruster, H. & M amo, C. (1994). The swift fox reintroduction program in Canada from 1983−1992. In: Symposium Proceedings on Restoration of Endangered Plants and Animals (eds. M . Bowles & C.J. Whelan). University of Cambridge Press. Cambridge, UK. Carlington, B.G. (1980). Re-introduction of the swift fox (Vulpes velox) to the Canadian prairies. M aster's Thesis, Faculty of Environmental Design, University of Calgary, Alberta, Canada. Cotterill, S.E. (1997). Population census of swift fox (Vulpes velox) in Canada: Winter 1996−1997. Prepared for Swift Fox National Recovery Team. Alberta Environmental Protection. Natural Resources Service, Wildlife M anagement Division. Covell, D. F. (1992). Ecology of the swift fox (Vulpes velox) in southeastern Colorado. M .Sc. Thesis. University of Wisconsin-M adison, M adison, Wisconsin, USA. Egoscue, H. J. (1979). Vulpes velox. Mammalian Species 122:1−5. Herrero, S., M amo, C., Carbyn, L.N., & Scott-Brown, M . (1991). Swift fox reintroduction into Canada. In: Proceedings of the Second Endangered Species and Prairie Conservation Workshop (eds. G.L. Holroyd, G. Burns, and H.C. Smith). Provincial M useum of Alberta Natural History Section, Occasional Paper No. 15, Edmonton, Alberta, Canada. Herrero, S., Schroeder, C. & Scott-Brown, M . (1986). Are Canadian foxes swift enough? Biological Conservation 36: 159−167. Hillman, C.N., & Sharps, J.C. (1978). Return of swift fox to northern Great Plains. Proceedings of South Dakota Academy of Science 57:154−162. Hines, T.D. (1980). An ecological study of Vulpes velox in Nebraska. M .Sc.Thesis, University of Nebraska, Lincoln, Nebraska, USA. IUCN. (2001). IUCN Red List Categories and Criteria: Version 3.1. IUCN Species Survival Commission. IUCN, Gland, Switzerland and Cambridge, UK. �31 Kilgore, D.L. (1969). An ecological study of the swift fox (Vulpes velox) in the Oklahoma panhandle. American Midland Naturalist 81: 512−534. Kitchen, A. M ., Gese, E.M ., & Schauster, E.R. 1999. Resource partitioning between coyotes and swift foxes: space, time, and diet. Canadian Journal of Zoology 77: 1645−1656. M oehrenschlager, A. (2000). Effects of ecological and human factors on the behavior and population dynamics of reintroduced Canadian swift foxes (Vulpes velox), D Phil Thesis. University of Oxford, Oxford, UK. M oehrenschlager, A., List, R., & M acdonald, D.W. In press. Escaping interspecific killing: M exican kit foxes survive while coyotes and golden eagles kill Canadian swift foxes. Journal of Mammalogy. M oehrenschlager, A., M acdonald, D.W., & M oehrenschlager, C. (2003). Reducing capturerelated injuries and radio-collaring effects on swift foxes. In: Swift fox conservation in a changing world (eds. M . Sovada & L. Carbyn). Canadian Plains Research Center, University of Regina, Regina, Canada. M oehrenschlager, A., Cypher, B., Ralls, K., Sovada, M .A., & List, R. (2004). Comparative ecology and conservation priorities of swift and kit foxes. In: Biology and Conservation of Wild Canids (eds. D.W. M acdonald and C. Sillero-Zubiri). O xford University Press, Oxford, UK. M oehrenschlager, A. & Sovada, M .A. (2004). Swift fox (Vulpes velox), pp. 109-116. In: Canids: foxes, wolves, jackals, and dogs. Status survey and conservation action plan. (C. Sillero-Zubiri, M . Hoffmann, & D. W. M acdonald, eds.). IUCN/SSC Canid Specialist Group, Gland, Switzerland and Cambridge, UK. Pechacek, P., Lindzey, F.G., & Anderson, S.H. (2000). Home range size and spatial organization of swift fox Vulpes velox (Say, 1823) in southeastern Wyoming. Zeitschrift für Saügetierkunde 65: 209−215. Pruss, S.D. (1994). An observational natal den study of wild swift fox (Vulpes velox) on the Canadian prairie. M .S. Thesis. University of Calgary, Alberta, Canada. Pruss, S.D. (1999). Selection of natal dens by the swift fox (Vulpes velox) on the Canadian prairies. Canadian Journal of Zoology 77 (4): 646−652. Pruss, S.D., Fargey, P. & M oehrenschlager, A. Submitted. Recovery Strategy for the swift fox (Vulpes velox) in Canada. Prepared in consultation with the Canadian Swift Fox Recovery Team. Species at Risk Act Recovery Strategy Series. Parks Canada Agency. Reynolds, J. (1983). A plan for the reintroduction of swift fox to the Canadian prairie. M aster's Thesis, Faculty of Environmental Design, University of Calgary, Calgary, Alberta, Canada. Rongstad, O.J., Laurion, T.R., & Andersen, D.E. (1989). Ecology of swift fox on the Pinon Canyon M aneuver Site, Colorado. University of Wisconsin, U.S.A. �32 Schroeder, C. (1985). A preliminary management plan for securing swift fox reintroductions into Canada. M aster's Thesis, Faculty of Environmental Design, University of Calgary, Calgary, Alberta. Soper, J. D. (1964). The mammals of Alberta. Hamly Press, Edmonton, Alberta, Canada. Sovada, M .A., Roy, C.C., & Telesco, D.J. (2001). Seasonal food habits of swift foxes in cropland and rangeland habitats in western Kansas. American Midland Naturalist 145: 101−111. Sovada, M .A., & Scheick, B.K. (1999). Preliminary report to the swift fox conservation team: historic and recent distribution of swift foxes in North America. In: Swift fox conservation team 1999 annual report (ed. C.G. Schmitt). New M exico Department of Game and Fish, Albuquerque, New M exico, USA. Sovada, M . A., Slivinski, C.C. & Woodward, R.O. (2003). Home range, habitat use, litter sizes, and pup dispersal of swift foxes in two distinct landscapes of western Kansas. In: Swift fox conservation in a changing world (eds. M . Sovada & L.N. Carbyn). Canadian Plains Research Center, University of Regina, Regina, Canada. Tannerfeldt, M ., Moehrenschlager, A., and Angerbjörn, A. (2003). Den ecology of swift, kit and arctic foxes. In: Swift fox conservation in a changing world (eds. M . Sovada & L.N. Carbyn). Canadian Plains Research Center, University of Regina, Regina, Canada. �� https://cpw.cvlcollections.org/files/original/aee463241c7b0c94bb06d3e96514389d.pdf 5afa9e1329cd9bd12eb784a3ebe72fd1 PDF Text Text Population Survey of Reintroduced Swift Foxes (Vulpes velox) in Canada and Northern Montana 2014/2015 By A. Moehrenschlager and C. Moehrenschlager �This publication may be cited as: Moehrenschlager, A. and C. Moehrenschlager. 2018. Population survey of reintroduced swift foxes (Vulpes velox) in Canada and northern Montana 2014/2015. Centre for Conservation Research, Calgary Zoological Society, Calgary, Alberta, Canada For copies of this report, contact: Centre for Conservation Research Calgary Zoological Society 1300 Zoo Road Calgary, Alberta, Canada T2E 7V6 e-mail: axelm@calgaryzoo.com Telephone: 403-232-7771 Fax: 403-232-9370 Internet: www.calgaryzoo.org �Population Survey of Reintroduced Swift Foxes (Vulpes velox) in Canada and Northern Montana 2014 / 2015 By A. Moehrenschlager and C. Moehrenschlager Funded by: �Acknowledgments Primary acknowledgement goes to Heather Harris of Montana Fish, Wildlife, and Parks who coordinated and supported surveys in Montana with exceptional diligence; Bob Inman also provided valuable overarching support for the agency. In Canada, Paul Gregoire of the Canadian Wildlife Service facilitated time-sensitive permits. Joel Nicholson of Alberta Fish & Wildlife and Kristy Bly from World Wildlife Fund (United States) were kind enough to lend motion-sensor camera equipment to field crews. Joel Nicholson and Pat Fargey of Alberta Fish & Wildlife provided insights and advice regarding data analyses. Elaine Williams and Jessica Steiner of Wildlife Preservation Canada provided administrative support. Calgary Zoological Society and Montana Fish, Wildlife, and Parks field staff braved harsh winter conditions during day or night to collect the valuable data that grace the following pages. Thank you to the following for their tenacity, commitment, community engagement, and dedication to animal welfare: Matthew Bell, Jarek Bernt, Melanie Finch, Genevieve Fuller, Lacey Hébert, Tegan May, Michael Miskolczi, Leslie Norris, Ethan Page, Carolyn Prentice, Christopher Reed, Lauren Riches, and Sian Wilson. Special thanks to Calgary Zoological Society staff who assisted with data analyses: Typhenn BrichieriColombi for continuing her decade of swift fox GIS and demographic analysis, Lea Randall for occupancy analyses, and Laura Keating for SECR sampling. Thanks also go to Tom Moorhouse of Oxford University’s Wildlife Conservation Research Unit for mark-recapture comparisons, and to New Zealanders Darryl McKenzie of Proteus Consulting and Murray Efford of Otago University for advice regarding statistical analyses. Thank you to all landowners for granting access to their land and for frequently assisting swift fox field crews in both countries. Finally, a huge thank you to ATB Financial Services, specifically to Sandra Huculak and Holly Regel. Swift fox surveys could not be conducted until ATB Financial Services stepped forward to help. Their tireless support for the program allowed us to collect swift fox information that is now making a real difference for this precious species. �Executive Summary Swift foxes (Vulpes velox) originating from reintroductions were surveyed across an area of 32,609 km2 on the prairies of Canada and Montana from October 31, 2014 to February 22, 2015. Catch-and-release and camera-trapping determined changes in demography and geographic distribution from population surveys that had been done up to 2006. Surveys were conducted in 102 Canadian and 87 Montana townships, and 38 Canadian townships were sampled with both survey techniques. Site occupancy analyses were used for camera data, and catch-and-release results were analyzed using mark-recapture or spatially explicit capture-recapture estimation (SECR). Similar likelihood of detection between methods suggested that camera-trapping could be used as a suitable alternative to catch-and-release for occupancy objectives. Catch-and-release at 94 townships where foxes had been previously located in 2005/2006 indicated reduced abundance since that time. Foxes were also detected using cameras in townships without previous swift fox documentation, but 45% fewer replicated townships had evidence of foxes from catch-and-release or camera-trapping than in the previous survey. Incidental sightings were frequent even in areas where foxes had not been detected, and these occurrences contributed to an extent of occurrence of 19,779 km2 which is similar to that of the population in 2005/2006. Swift fox abundance across the population was estimated at 870.1 + 191.7. Foxes were well distributed within an inferred area of occupancy of 6,410 km2. Causes for recent declines are difficult to discern, but reductions likely occurred between 2009 and 2013 and winter severity may have been a contributing factor. Although swift fox reintroductions have shown evidence of success, recent population changes suggest that the frequency of population monitoring should be increased again. A comprehensive population survey will be conducted in 2018 in Canada and Montana. �Table of Contents 1 2 Introduction .................................................................................................................... 1 Methods .......................................................................................................................... 3 2.1 2.2 Sampling approach ............................................................................................................... 3 Survey techniques ................................................................................................................. 5 2.2.1 Release-trapping .................................................................................................................. 5 2.2.2 Camera-trapping .................................................................................................................. 5 2.3 Analyses ................................................................................................................................ 6 3 4 2.3.1 Detection Efficacy of Release vs. Camera-Traps ................................................................... 6 2.3.2 Demographic Changes since 2005/2006 .............................................................................. 7 2.3.3 2014/2015 Abundance and Geographic Estimates .............................................................. 7 Results ............................................................................................................................ 9 3.1 3.2 3.3 3.4 Sampling Effort...................................................................................................................... 9 Detection Efficacy of Release vs. Camera-Traps ..................................................................... 9 Demographic Changes since 2005/2006............................................................................... 10 2014/2015 Abundance and Geographic Estimates ............................................................... 13 Discussion ......................................................................................................................16 Literature Cited .....................................................................................................................19 �1 Introduction Conservation translocations such as reintroductions have returned species to sites of former extirpation, including ones that were extinct in the wild such as the scimitar-horned oryx (Oryx dammah), Socorro dove (Zenaida graysoni), and black-footed ferret (Mustela nigripes). Carnivores can be difficult to restore because prey populations must be sufficient to support released animals, competitors may have filled a niche left empty by the extirpated species, captive-bred individuals must exhibit hunting behaviours after release, mesocarnivores need to retain predator avoidance behaviours, and many carnivores are the target of human persecution when the safety of people, their pets, or their livestock is in peril. Nevertheless, success is possible as conservation translocations have restored carnivores such as bobcats (Lynx rufus) to regions of the United States (Diefenbach et al. 2015), brown bears (Ursus arctos) to areas of Italy (Peters et al. 2015), and African wild dogs (Lycaon pictus) to southern Africa (Gusset et al. 2007). Conservation translocations involve different stages of population growth and regulation. Restored populations are created through phases of establishment, growth and regulation (Brichieri-Colombi & Moehrenschlager 2016) which may nevertheless be characterized by periods of dramatic growth or decline. Historical species presence does not imply current habitat suitability (Osborne & Seddon 2011), and lack of historical presence does not connote current habitat unsuitability. Consequently carrying capacity of reintroduction sites can be difficult to estimate, particularly in severely altered ecosystems or following long periods of species extirpation. The swift fox (Vulpes velox) is a mesocarnivore of the North American plains historically ranging from southern Canada to the southern United States. Like its close relative the kit fox (Vulpes macrotis) which ranges from northwestern areas of the United States to northern areas of Mexico, it has undergone large-scale declines in distribution and abundance as a result of direct persecution, habitat loss, or habitat degradation (Moehrenschlager & Sovada 2016). Current site suitability is affected by the presence of relatively undisturbed native grasslands, available prey, the presence of predators such as golden eagles (Aquila chrysaetos), and the effect of generalist competitors such as the red fox (Vulpes vulpes) or coyote (Canis latrans) which have benefitted from human expansion and the associated extirpation of wolves (Canis lupus) in prairie environments (Moehrenschlager et al. 2004). The swift fox was extirpated from Canada by 1938, and foxes in Montana suffered a similar fate. Reintroduction programs were initiated in Canada in 1983 and ranged until 1997 with releases of 942 animals, initially involving only captive-bred foxes, and gradually increasing the role of translocated animals which exhibited relatively high survival rates (Moehrenschlager & Macdonald 2003). Populations were sufficiently well established by 1996 (Cotterill 1997) that further releases were discontinued in the hope that populations would persist or grow without further augmentation. Indeed, increasing evidence suggested that swift foxes were beginning to establish in areas of northern Montana that were adjunct to Canadian populations. By 2001, Canadian populations had grown in distribution and abundance and some contiguous Montana regions had clearly been occupied by swift foxes (Moehrenschlager & Moehrenschlager 2001). A follow-up population survey suggested further dramatic increases in range and abundance by 2006, and all release-trapped foxes were wild born suggesting waned dependence on original reintroduction releases (Moehrenschlager & Moehrenschlager 2006). Genetic analyses showed that the reintroduced population was sufficiently diverse and that an eastern cluster of Canadian/Montana swift foxes had differential genetic structure from a western cluster of Canadian/Montana foxes (Cullingham & Moehrenschlager 2013). Habitat modelling not only revealed predictive characteristics of swift fox presence, but also suggested that habitat surrounding 1 �the reintroduced population may be limited (Moehrenschlager et al. 2006). Although one swift fox was found to disperse nearly 200 km from a separate swift fox reintroduction site in northwestern Montana to the contiguous Canada/Montana population, such potential movements among these distinct populations are thought to be rare (Ausband & Moehrenschlager 2009). Swift foxes were downlisted in Canada from ‘Extirpated’ to ‘Endangered’ and subsequently from ‘Endangered’ to ‘Threatened’ in 2012 based on available demographic evidence entailing population trends and geographic parameters. Given the period of sustained population growth, the question remained whether the swift fox population would continue to grow, stabilize, or decline. The Canadian federal swift fox recovery strategy articulates short and long-term goals towards population recovery (Pruss et al. 2008). The short-term recovery goal is to ensure a mature, reproducing swift fox population size of at least 250 foxes by 2012. The long-term recovery goal by 2027 is to restore a self-sustaining swift fox population of 1,000 or more mature, reproducing foxes that does not experience greater than a 30% population reduction in any 10-year period. The recovery of swift foxes in Canada and in Montana is interdependent, particularly as movements occur across the international border and foxes in respective eastern and western clusters of the population are genetically more similar across the border than they are within respective countries (Ausband & Moehrenschlager 2009; Cullingham & Moehrenschlager 2013). Consequently, achieving swift fox recovery goals in different jurisdictions necessarily depends upon an understanding of the population across its range. In conjunction with the responsible government agencies we aimed to determine population trends and geographic distributions of swift foxes for the first time since comprehensive surveys had last been conducted in 2005/2006. 2 �2 Methods 2.1 Sampling Approach (Adapted from Moehrenschlager & Moehrenschlager 2006) The original Canadian swift fox population survey area was determined by the National Swift Fox Recovery Team in 1996, based on habitat criteria and background information (Cotterill 1997). The suspected swift fox range consisted of 65 townships in the Alberta / Saskatchewan border region and 43 townships in and around Grasslands National Park in south-central Saskatchewan. Of these 108 townships, 81 (75%) were randomly selected for subsequent sampling, of which two were excluded at the request of resident landowners. Similarly a population survey area of 80 townships was selected by Montana Fish, Wildlife and Parks staff in 2000, and 75% of these townships were selected for subsequent sampling. In 2005/2006, the population survey area was further expanded by 49 townships on the periphery of the 2000/2001 area of swift fox occupancy. These townships were adjacent to townships with 2000/2001 captures to assess potential population expansion. In 2014/2015, a mix of catch-and-release trapping (hereafter termed “release trapping”) and camera-trapping were used to determine the relative efficacy of the methods in detecting foxes, and to optimize logistical considerations for surveys across a swift fox population that could be expanding further. In Canada, both methods were tested within 2 weeks of one another at 38 townships. Other regions with previous swift fox evidence were surveyed using release trapping, while areas without previous detections were re-surveyed using camera-traps in Canada and Montana. The survey area was expanded slightly beyond 2005/2006 areas by sampling additional townships on the western edge of the population in Alberta and the northern edge of population clusters which connected Alberta and Saskatchewan populations. Some targeted townships could not be sampled because of landowner access restrictions or winter weather conditions. Swift fox surveys were conducted from October 31, 2014 until February 22, 2015 in Alberta, Saskatchewan, and Montana using site selection based on replicated methodology from previous comprehensive surveys of the region (Moehrenschlager & Moehrenschlager 2006). Sample sites for release traps or camera-traps were placed at one-kilometre intervals along five kilometre continuous sections of trail that were closest to the centre of respective townships. Inter-trap distances were adjusted by up to 100 meters to allow trap placement along fences or on top of hills. Each township was surveyed with six camera or release traps for three nights. Incidental sightings of swift foxes were recorded. 3 �4 Figure 1: Sampling sites using release-traps, camera-traps, or both methods in 2014/2015 �2.2 Survey Techniques 2.2.1 Release-Trapping (Adapted from Moehrenschlager & Moehrenschlager 2006) Catch-and-release trapping was conducted on consecutive nights when possible, although interruptions in the sequence happened occasionally due to weather. Two sizes of fox release-traps were used for catch-and-release: large traps were 109 cm x 39 cm x 39 cm Tomahawk (Tomahawk Live Trap Co., Tomahawk, WI) double-door or 109 cm x 39 cm x 47 cm Safe Guard (Safe Guard Products Inc., Lancaster County, PA) traps; small traps were 83 cm x 31cm x 31 cm Tomahawk single door box traps or 78 cm x 29 cm x 31 cm Safe Guard traps. Lining of traps with wood has been shown to reduce injuries (Moehrenschlager et al. 2003), and lining further evolved during this survey by using puck board to completely encase the interior of traps aside from the door which was lined with Plexiglas. Release trapping was conducted at night to reflect fox activity periods, to avoid heat-stress, and to prevent fox disturbance by people. Traps were generally set between 1800 and 2000, checked between 2400 and 0200, and closed following a second check around 0600. Trapping was not conducted at temperatures colder than –20 Co or when snow, rain, and wind conditions were potentially hazardous to captured foxes (Moehrenschlager & Moehrenschlager 2001). Foxes were coaxed into a denim handling bag which was placed over the end of the box trap. Every team alternated between four handling bags which were washed frequently to reduce flea transference. Foxes were handled by both field workers of respective trapping teams. The first team member positioned the animal on his/her lap to shelter it from the wind, one hand restrained the head and covered the eyes, and the second hand restrained the body. The second field worker sexed and aged the fox, conducted parasite counts, scored body condition, collected a hair sample for genetic analyses, checked for injuries, and tattooed the ear for identification. Age classifications were based on the size, colour, and wear of teeth. Foxes were uniquely marked with tattoo dye so that recaptured individuals could be easily identified. 2.2.2 Camera-Trapping Camera-trapping sites used mackerel oil or fatty acid attractants and scent-posts. Scent-posts consisted of lumber stakes measuring 5 x 5 x 40 cm and pointed on one end to facilitate mounting in the ground. Approximately 5 cm from the top of the stake, a plaster of Paris disk was mounted on the wood using a nail. Plaster disks were soaked in mackerel oil at the time of deployment. Reconyx motion-sensor cameras were mounted on custom-made iron stakes and positioned approximately 40 cm from the ground facing the scent-post. Cameras faced the scent-posts at a distance of 2 m (Camaclang et al. 2010). 5 �2.3 Analyses 2.3.1 Detection Efficacy of Release vs. Camera-Traps Survey intervals fell into two sessions per night that spanned from approximately 6 pm to midnight, and midnight to 6 am respectively, for release and camera sampling. Only one fox could be caught per trap in each release trapping session. In contrast, camera data could have multiple or repetitive images in a 6 hour period. Unlike species with distinctive markings, swift fox individuals cannot be reliably differentiated from camera images; even sexes cannot be distinguished. To effectively compare the release trapping and camera methods, camera data were coded as binary based on detection in the 6 hour periods. If one or several swift fox images were recorded at a camera, that sample session was categorized as having swift fox evidence for the individual sample site. Since release or camera-traps spaced at 1 km intervals are spatially nonindependent, respective release and camera data across the 6 sample sites were lumped for each sample session. This means that even if multiple release-traps or camera-traps had evidence of foxes within the same township in the same session, each township would only have one positive coding indicating swift fox evidence per session. Comparative method sampling was conducted at 38 townships in Canada for three consecutive nights per method. Both methods were sampled within 2 weeks of one another to ensure reasonable similarity in seasonal shifts. However, at least 3 days buffer were used between methods to allow swift fox behaviour to normalize since previous sampling. The order of utilized methods was randomized across townships to further reduce bias. Respective method-specific probabilities of detection, and subsequent occupancy estimates for the 38 townships were estimated using program Presence 12.6. To determine relative likelihoods of detection, a multi-method analysis (Guzy et al. 2014) was run (psi, theta(.),p(method). As all sampling was conducted across 3 consecutive nights consisting of two 6 hour sampling sessions respectively, probabilities of detection are presented as cumulative values that represent the entire three-night sampling interval. To compare if detection probabilities differed substantially between methods, a model was run that allowed the probability of detection to differ by method, and this was compared to an alternative model where probability of detection was kept constant. These candidate models were compared using ranked AIC, model weights, and an evidence ratio. As a further exploratory method-comparison test, occupancy was estimated with comparative models, and again compared with ranked AIC, model weights, and an evidence ratio. Given that foxes are either present or not at a site during a trapping interval, occupancy estimates should be similar across sites regardless of the trapping method. To explore if camera-trapping vs. releasetrapping would result in similar occupancy estimates, we used a multi-season analysis where occupancy was coded as a separate season for each method. The first model was run keeping occupancy constant (psi(.),gamma(),p(method)), while the alternative model allowed occupancy to differ by method (psi(method),gamma(),p(method)). 6 �2.3.2 Demographic Changes since 2005/2006 Release trapping results were compared for townships that were surveyed in both 2005/2006 and 2014/2015. Unlike camera data, release trapping allowed for the identification of trapped individuals. Recaptures could be identified in traps and released without additional re-handling. Recaptures always occurred in the same townships as initial captures, which further corroborates that neighbouring townships were spatially independent. Individual foxes were frequently caught at more than one trap within townships, corroborating lack of spatial independence within townships. The capture history of individual foxes was categorized across the six sampling sessions that comprised the 3 days of release capture sampling that was particular to each township. Similarly, all 2005/2006 release capture data were categorized into individual capture histories. The surveyed swift fox population was not assumed to be closed across the interval of the entire population survey, but it was assumed to be closed across the 3 day sampling intervals for release traps, camera-traps, or two-week intervals between camera and release-traps for comparative method analyses; these assumptions seem reasonable based on annual survival rates of swift foxes in Canada (Moehrenschlager et al. 2007). Closed capture mark-recapture models were run in program Mark without co-variates to estimate and compare abundance among 2005/2006 and 2014/2015. For each time frame, capture histories were tested to determine whether the likelihood of first capture (p) and recapture (c) varied substantially (ie. constant p: p(.)=c(.); time varying p: p(t) = c(t); behavioural response: p(.), c(.)). Estimates of p and c from best-fit models were examined to determine potential differences in capture or recapture between 2005/2006 and 2014/2015. Finally abundance estimates (fo) and associated 95% confidence intervals were derived with best-fit models for each time frame. 2.3.3 2014/2015 Abundance and Geographic Estimates While effective at showing trend over time in a subset of sample transects, one disadvantage of the mark-recapture estimates was that the abundance estimates are not spatially explicit, and hence cannot be extrapolated to accurately estimate abundance across the population. While neighbouring townships were spatially independent, it was uncertain what proportion of the total area within respective townships was effectively sampled. For example, it was unclear from how far foxes would be effectively attracted to sample sites. Package Secr (Spatially Explicit Capture Recapture) of Program R (Efford 2017) is the R implementation of Program Density and uses a maximum likelihood approach. Coordinates of all sample sites, capture locations, and recapture locations were integrated. The habitat mask integrated a grid size of 400 metres, which appeared optimal in terms of resolution and adherence to analyses assumptions (Efford, pers. comm.). The habitat mask was also based upon a sampling buffer of 3.5 km that was fit to individual trap sites; this distance coarsely approximated the radius of Canadian swift fox home ranges (Moehrenschlager et al. 2007) and is the approximate distance beyond which habitat selection models loose predictive capacity (Moehrenschlager et al. 2006). While 34 townships in 2014/15 had swift fox evidence from release trapping, 12 had evidence from townships only surveyed with cameras. Three of these were from sites that had been sampled with both methods, where cameras detected foxes but release trapping did not. For abundance estimation, evidence of foxes from these three sites was not included to prevent overestimation based on double sampling. The other nine townships that had evidence from cameras but had not had release trapping were relevant to abundance estimation. The distribution of these nine transects was dispersed within the study area, and the habitat associated with these transects was broadly similar to that of areas where release trapping yielded swift fox captures (Figure 3). Consequently 7 �spatially-explicit capture dynamics were simulated for these nine (21%) of 43 townships that were used for abundance estimation across the entire study area. The resulting swift fox mark recapture data set used for SECR calculations used 63 known individuals from actual release capture sites, and 15 inferred individuals from the nine camera-trapping transects that had swift fox evidence. With a similar balance of actual and inferred data, 116 total detections were estimated for the total of 78 individuals. These data were used to parameterize a half-normal detection function (intercept go = 0.17, spatial sigma = 618.8 m) using the grid spacing of 400 m with the 3.5 km sampling site buffer. Locations for simulated captures were generated using the distribution of captures in the release trapping data set. Finally, fox density and associated errors were calculated by dividing estimated abundance by the SECR-generated estimated sampling area. Abundance for Canada, Montana, and the combined population were then determined by extrapolating across the entire study area. To inform a pending COSEWIC review of swift fox status in Canada, and to yield updated information for IUCN reassessments, extent of occurrence for the sampled population was calculated for Canada, Montana, and the combined population. While IUCN calculates Area of Occupancy (AO), COSEWIC terms its calculation as the Inferred Area of Occupancy (IAO) which must be based upon a projection of 2 x 2 km grid squares that are overlapped entirely or in part by the inferred species range. IAO was calculated excluding incidental observations using the following process: 1. Initial captures from release-trapping and simulated camera data sets were plotted; 2. All trap sites on transects that had captures were buffered by 3.19 km (approximate radius of 32 km2 swift fox home range); the number of 2 x 2 km grid cells overlapped by occupied buffer zones were determined; 3. The number of inferred occupied grid cells was adjusted using a probability of detection of 0.865, which is the average of estimated release-trapping and camera-trapping samples; 4. The proportion of occupied grid squares was determined by dividing the number of squares inferred in step 3 by those, determined in step 2; 5. The total # of grid squares overlapping the sampled population were counted; 6. The proportion of occupied grid squares from step 4 was multiplied by the total number of grid squares comprising the swift fox study area; 7. The # of grid squares inferred to have swift fox presence were multiplied by 4 km2 to determine the COSEWIC required Inferred Area of Occupancy. IAO calculations are presented for Canada, Montana, and the combined swift fox population. As COSEWIC requires a graphical representation of IAO, a 4 x 4 km grid is depicted in Figure 5; however, this grid is a gross underestimate of true IAO as inferred areas can be calculated but not mapped. 8 �3 Results 3.1 Sampling Effort During the winter of 2014/2015 189 townships were sampled in Canada and Montana; 102 were sampled in Canada and 87 in Montana. Of these, replicated sites were surveyed in 170 of 191 townships that had been sampled in 2005/2006; 19 townships were sampled that had not been surveyed in 2005/2006. Release-trapping was conducted in 94 townships (42 in Canada, 52 in Montana), and camera-trapping in 133 townships (98 in Canada, 35 in Montana). Of the 102 townships surveyed in Canada, 38 were sampled using both methods (Figure 1). The total sampling area, calculated as a minimum convex polygon bordered by sample transects of either method, enveloped 32,608.5 km2. 3.2 Detection Efficacy of Release vs. Camera-Traps Release and camera-traps had high detection probabilities and similar efficacies in detecting swift foxes. Among 38 townships sampled with both camera and release-traps in Canada (Figure 1), 21 did not document foxes with either technique, 10 had evidence with both methods, 4 detected foxes using only release trapping, and 3 had detections only using camera-traps. Among the 14 townships where foxes were documented with release traps, 7 had the first detection in the first night (ie. first 2 trap checks), 5 had the first detection in the second night (3rd and 4th trap check), and the first detection occurred in the third night (5th and 6th check) on 2 occasions. Among 13 townships where foxes were documented with cameras, first detection was in the first night on 6 townships, in the second night in 6 townships, and in the third night on one occasions. Among the 10 townships where foxes were detected using both methods, the first detection occurred earlier with cameras than traps on 4 occasions, earlier with traps than cameras on 4 occasions, and at the same time on 2 occasions. The fact that most first detections occurred in the first 2 nights compared to a lower proportion on the third night for both methods seems to suggest that foxes were detected relatively early within the sampling periods. Multi-method occupancy modeling determined that the cumulative detection probability across 3 nights of sampling were 0.81 for camera-trapping and 0.93 for release traps. Comparative models psi,theta(.),p(method) compared to psi,theta(.),p(same) revealed no significant support (delta AIC = 1.33) to suggest that such differences were statistically important. The combined occupancy estimate was 0.39 +/- 0.08 (95% Confidence Interval: 0.24 – 0.54). Just as a point of comparison, if occupancy had been calculated separately for each method at the 38 townships where both methods were used, the occupancy estimates would nevertheless have been nearly identical (Release-trapping: 0.38 +/- 0.08, 95% Confidence Interval: 0.22 – 0.54; Camera-trapping: 0.40 +/- 0.10, 95% Confidence Interval: 0.21 – 0.60). This again suggests that release-trapping and camera-trapping methods yielded similar information about the presence of swift foxes. In addition to the 17 townships with swift fox evidence from one or both sample methods, swift foxes were also detected on 29 townships which were either release-trapped or camera-trapped, but not sampled with both methods. Of these 29, release traps detected foxes in 17 Montana townships and 3 Canadian townships, and cameras detected foxes on 4 Montana and 5 Canadian townships. In total foxes were detected using sampling techniques on 46 townships, but incidental sightings occurred in additional areas. 9 �3.3 Demographic Changes Since 2005/2006 Release trapping was conducted at 94 townships in 2014/2015 where foxes had been detected in 1995/1996, 2000/2001 or 2005/2006 surveys (Figure 2). In 2005/2006, 76 of these townships had release trap captures for a total of 172 foxes. In 2014/2015, only 35 of these townships had captures for a total of 51 foxes. In 2014/15, males made up 56.9% of captures compared to 52.6% in 2005/2006; juveniles made up 60.2% of foxes in the earlier survey compared to 45.1% in the more recent assessment. Likelihood of fox capture and recapture in mark-recapture models was similar between 2005/2006 and 2014/2015. The highest ranking model for 2005/2006 suggested that capture and recapture probabilities varied over the 6 successive trap opportunities. The probability of first capture averaged 0.184 (range : 0.131 and 0.243), and the likelihood of recapture averaged 0.194 (range: 0.152 – 0.243). The highest ranking model for 2014/2015 suggested that probabilities of first capture and recapture were constant (0.198; SE: 0.147 – 0.261). Models testing for a behavioural response to trapping within respective sampling periods ranked low, indicating that foxes did not become more likely or less likely to be recaptured over the three-day sample periods. Accounting for respective capture and recapture probabilities, mark-recapture abundance estimates at the 94 sampling sites where release trapping was replicated were 243.4 +/- SE 15.4 (95% Confidence Interval: 218.8 – 280.1) for 2005/2006 compared to 68.9 +/- SE 15.4 (95% Confidence Interval: 59.3 – 90.0) for 2014/2015. The decrease in replicated townships with captures and/or fox abundance was slightly offset by the fact that foxes were detected on townships where they had not been documented in 2014/2015. Twelve replicated townships had evidence of foxes in 2014/2015 which did not have evidence in 2005/2006. Of these 5 were release-trapped and these were already included in the markrecapture abundance comparisons described above. However, in addition to the 94 townships with fox captures in 2005/2006, further 76 townships were sampled with cameras in 2014/2015 where no foxes had been captured in 2005/2006 (Figure 3). Camera-trapping revealed evidence of swift foxes on 7 townships where foxes had not been documented in 2005/2006. On the total 170 replicated townships, 76 (44.7%) had evidence of foxes in 2005/2006 compared to 42 (24.7%) in 2014/2015, representing an approximate drop of 45% in townships with evidence of swift foxes from systematic sampling. Nevertheless, swift foxes were noted in additional areas. Nineteen townships were sampled in 2014/2015 that had not been sampled in 2005/2006. Swift foxes were detected on one of these townships using release-trapping and another using camera sampling, increasing the total number of townships with evidence of foxes from such sampling to 44. In addition, incidental sightings were made on, or within 3 km, of 10 replicated sample transects in 2014/2015 where release capture or camera surveys showed no evidence of foxes (Figure 3). One incidental sighting without sampling evidence in 2014/2015 was on a township that had not had evidence in 2005/2006 either. 10 �Figure 2: Comparative number of unique swift foxes caught at replicated release-trap sites in 2005/2006 vs. 2014/2015 11 �12 Figure 3: Evidence of swift foxes on townships that were replicated in 2014/2015 (from release-trapping or camera-trapping) from previous sampling in 2005/2006 �3.4 2014/2015 Abundance and Geographic Estimates Spatially explicit capture-recapture estimation (SECR) yielded fox abundance estimates for 2014/2015. A total of 63 individual foxes were release-captured, and an additional 15 foxes were inferred to occur at similar densities for sample transects where swift foxes were photographed but could not be individually identified. Program SECR used distances between initial captures and recaptures of individual release-trapped foxes to estimate a total effective sampling area (esa) of 2,148.96 km2. Fox density across the Canadian/Montana study area range was estimated at 0.0036 + 0.008 foxes / km2. In Canada, 522.9 + 115.2 foxes were estimated to occur across 14,402.1 km2. In Montana, 346.9 + 79.5 foxes were estimated to occur across 9,561.6 km2. Across 23,963.5 km2 of sample range for the study area, fox abundance was estimated at 870.1 + 191.7. Abundance estimates using SECR methodology are not available for 2005/2006 at this time, but abundance is thought to have decreased since that time frame. Extent of Occurrence from sampled and incidental observations during the 2014/2015 winter survey period spanned 19,779 km2 for the swift fox population. Extent of Occurrence of the Canadian population was 10,998 km2 in Canada and 6,916 km2 in Montana (Figure 4). Using refined methodology from the current survey, Inferred Area of Occupancy was recalculated for the 2005/2006 period. In Canada alone, IAO was 4,655 km2 for 2005/2006 which is similar to the 4,411 km2 estimate for 2014/2015. IAO for the entire Canadian/Montana range was 6,410 km2 in 2014/2015 compared to 8,969 km2 previously due especially to fewer occupied areas in Montana in the later time-frame. An under-representation of IAO is depicted in Figure 5 with an IUCNcompliant 2 x 2 km grid; this is an under-representation of calculated IAO because it cannot illustrate areas that are inferred to exist when accounting for subsampling and imperfect detection. 13 �14 Figure 4: Extent of Occurrence on population, Canadian, and Montana levels from sampling evidence and incidental observation during the winter of 2014/2015 �15 Figure 5: Under-represented graphical interpretation of Area of Occupancy, indicating sampling buffers surrounding transects. Graphic represents 2x2 km gridded occupancy from sampling or incidental sightings; this does not fully depict Inferred Area of Occupancy (IAO) �4 Discussion Swift foxes were extirpated from Canada for over 40 years until reintroductions were initiated in 1983. By 1997, intensive releases had founded a population that was growing in abundance and distribution in Canada and adjoining areas of Montana. After releases in Canada were stopped in 1997, the population continued to grow over the following decade in Canada and adjunct regions of the United States (Moehrenschlager & Moehrenschlager 2001, 2006). Current results based on data from 2014/2015 suggest that the population has a similar geographic extent as in 2005/2006 and remains composed of foxes born in the wild, but demographic declines appear evident. Given the nine year gap between comprehensive surveys and distribution over nearly 20,000 km2, causal mechanisms are difficult to discern. Differences in sampling methodology do not appear to explain observed demographic changes. Release trap designs have been improved continuously and effectively to minimize injury risks by progressively lining more of the steel mesh with wood or plastic (Moehrenschlager et al. 2003; Moehrenschlager & Moehrenschlager 2006). In the current survey, the entire trap interior was lined with puck board or Plexiglas. A potential confound was that this lining would limit scent dispersion or that foxes would be less likely to enter traps. The fact that calculated probabilities of initial capture and recapture were similar between 2014/2015 and 2005/2006 in mark-recapture models based on sampling at identical locations, suggests otherwise. Moreover, results from comparative method tests in 2014/2015 yielded swift fox captures in 10 townships where cameras had also detected foxes and seven townships with differential results between methods included four townships where release traps caught foxes but cameras did not detect them. A second potential confound was that previous comprehensive population surveys only used release traps, whereas the current survey used a combination of this technique with camera-traps. Occupancy models based on comparisons in 38 Canadian townships, showed no evidence that the likelihood of detection differed significantly between techniques. Indeed, initial fox detections primarily occurred within the first two nights for both methods, suggesting similar low latencies of detection. These results suggest that camera-trapping is an effective alternative to release trapping to estimate changes in swift fox occupancy. Camera-trapping probability of detection was similar to Colorado findings where nightly rates ranged from 0.62 to 0.8 and most initial detections had occurred after two nights (Stratman & Apker 2014). Camera-trapping is non-invasive, cheaper, and easier for surveying multiple townships than release trapping. However, due to a lack of distinctive markings and dimorphism, demographic information such as sex ratios or abundance cannot be effectively determined with camera-traps for swift foxes. Aside from the 38 townships where both methods were used, camera-trapping was utilized at 95 townships whereas release-trapping was only used in 56 townships. Nevertheless, almost 80% of detected foxes were located on townships with release trapping because this method was utilized for sites where foxes had been detected previously. Previous population surveys of swift foxes in Canada and Montana attempted to integrate imperfect detection based on the capture success of a small sample of radio-tracked foxes in 1996/97 for which home range locations and sizes were known (Cotterill 1997; Moehrenschlager 2002). These derived correction factors were not only utilized for the 1996/1997 population survey estimates; they were also applied to subsequent population surveys to enable replicable comparisons (Moehrenschlager & Moehrenschlager 2001, 2006). However, these methods did not use formal mark-recapture estimation or spatially explicit estimators, unlike the methods applied for the current survey. The estimated sampling area derived by SECR analyses for the 2014/2015 data seems to suggest that previous correction factors may have been too conservative in their population abundance projection. Maximum distances between initial captures and recaptures generally did not 16 �exceed 2 km in the current survey, indicating that release-trapping transects were only attracting foxes from relatively proximate sites. This finding implies that a higher proportion of townships were not sampled with the central trapping transects. Future analyses will aim to recode data from the previous population surveys to apply SECR techniques, and future release trapping surveys will apply similar analyses. Surveys utilizing camera traps should continue to use occupancy analyses. Release-trapping mark-recapture comparisons for 94 previously sampled townships suggested dramatic decreases in swift fox abundance since 2005/2006. This result was not entirely representative of the entire population since foxes were also detected in some townships using camera-trapping where they had not been detected previously. Nevertheless, foxes were detected in 45% fewer replicated townships than during the last comprehensive survey which is a matter of concern. Numerous sites had incidental sightings without detections during release-trapping or camera surveys, particularly in eastern areas of Montana and Canadian sample regions. The swift fox population remains well distributed, which bears promise for extirpated sites to be colonized from extant sites under favourable conditions. Given perpetual growth in the distribution and abundance of the swift fox population from 1996 to 2006, it is difficult to determine if current levels constitute a dramatic decline from normal levels or whether 2006 densities exceeded the yet unknown carrying capacity of the reintroduced population. Moreover, mechanisms driving population reduction to 2014/2015 are difficult to infer. Reductions in abundance or localized distributions appeared uniform across the population, spanning jurisdictional boundaries and those of established western and eastern population clusters. Consequently, localized small-scale habitat loss, habitat degradation, poison use, or swift foxes harvested for fur in Montana cannot sufficiently explain population changes across the population. Disease could theoretically cause wide-spread drops in population abundance, but such effects in canids tend to be patchy (van de Bildt et al. 2002; Knobel et al. 2007; Timm et al. 2009; King et al. 2014; Marino et al. 2017). More importantly, eastern and western regions of the swift fox population have been genetically clustered as two distinctive subpopulations with limited gene flow (Cullingham & Moehrenschlager 2013), which makes the ubiquitous spread of disease unlikely. The period between comprehensive surveys spanned nine years, which makes it difficult to discern causal drivers for observed population reductions. Nevertheless, localized swift fox surveys may be useful in discerning the time period where population changes were greatest. Swift fox camera sampling related to an assessment of natural gas exploration sites in southwestern Saskatchewan had high numbers of swift fox detections (Camaclang et al. 2009) that were in line with 2005/2006 population densities. In the spring of 2009, camera-trapping was also conducted on 32 townships that had been surveyed in 2005/2006 as a component of habitat model validation (Moehrenschlager et al. 2006). Swift foxes were detected in 23 of these townships in 2009 compared to detections at only 14 of these townships in 2005/2006. As such dramatic decreases did not seem to have occurred by that time. However, camera surveys that preceded 2014/2015 during the winter of 2013/2014 in adjunct populations of Alberta and Saskatchewan indicated substantially fewer swift fox occurrences than during previous surveys. These results seem to suggest that decreases in the swift population may have occurred particularly between 2009 and 2013. Swift foxes are at the northern edge of their range in Canada with home range sizes that are the largest recorded for the species. Indeed Canadian swift fox home ranges are approximately 4 times larger than those of Colorado conspecifics and these home range expansions occur primarily in winter (Moehrenschlager et al. 2007). Energetic constraints and/ or increased vulnerability to killing by predators or intraguild competitors can limit populations, and starvation during winter has been documented previously (Moehrenschlager 2002). The winter of 2010/2011 was severe, with 17 �numerous anectodal reports of ubiquitous livestock losses. One possible explanation for the population decline observed in 2014/2015 is that winter severity during 2010/2011 may have reduced the population and insufficient time or suboptimal environmental conditions since then limited population recovery to 2014. If winter severity was indeed responsible for population reduction, the potential for a population rebound would remain under more favourable conditions. Swift foxes have the reproductive potential to bounce back under optimal conditions as they can breed at one year of age, and produce litters of up to seven young in Canada (Moehrenschlager 2002). Broadly distributed areas of occupancy within the population’s extent could serve as source sites to recolonize townships where foxes were not detected in the recent survey. Montana Fish & Wildlife reacted proactively to initial survey results by limiting the legal harvest of swift foxes in the state, and by precluding the use of the swift fox population as a source for conservation translocations in other regions (Harris, pers. comm.). Additional population monitoring is necessary to determine demographic changes since 2015, and a range-wide Canadian/Montana swift fox survey will be conducted in 2018. 18 �Literature Cited Ausband, D., and A. Moehrenschlager. 2009. Long-range juvenile dispersal and its implication for conservation of reintroduced swift fox Vulpes velox populations in the USA and Canada. Oryx 43:73-7. Brichieri-Colombi, T. A., and A. Moehrenschlager. 2016. Alignment of threat, effort, and perceived success in North American conservation translocations. Conservervation Biology 30:11591172. Camaclang, A. E., A. Moehrenschlager, and P. Fargey. 2010. Swift fox use of gas structure sites at a small Community Pasture in southwest Saskatchewan. Centre for Conservation Research Report No. 3., Calgary Zoological Society, Calgary, Alberta, Canada. Cotterill, S. E. 1997. Population census of swift fox (Vulpes velox) in Canada: Winter 19961997.Prepared for Swift Fox National Recovery Team. Alberta Environmental Protection. Natural Resources Service, Wildlife Management Division. Pages 50. Cullingham, C. I., and A. Moehrenschlager. 2013. Temporal analysis of genetic structure to assess population dynamics of reintroduced swift foxes. Conservervation Biology 27:1389-1398. Diefenbach, D., L. Hansen, J. Bohling, and C. Miller-Butterworth. 2015. Population and genetic outcomes 20 years after reintroducing bobcats (Lynx rufus) to Cumberland Island, Georgia USA. Ecology and Evolution 5:4885-4895. Efford, M. G. 2017. secr: Spatially explicit capture-recapture models. R package version 3.1.0. https://CRAN.R-project.org/package=secr. Gusset, M., et al. 2007. Efforts going to the dogs? Evaluating attempts to re-introduce endangered wild dogs in South Africa. Journal of Applied Ecology 45:100-108. Guzy, J. C., S. J. Price, and M. E. Dorcas. 2014. Using multiple methods to assess detection probabilities of riparian-zone anurans: implications for monitoring. Wildlife Research 41:243257 King, J. L., C. L. Duncan, and D. K. Garcelon. 2014. Status of the Santa Catalina Island fox thirteen years after its decline. Monographs of the Western North American Naturalist 7:382-396. Knobel, D. L., A. R. Fooks, S. M. Brookes, D .A. Randall, S. D. Williams, K. Argaw, F. Shiferaw, L. A. Tallents, and M. K. Laurenson. 2007. Trapping and vaccination of endangered Ethiopian wolves to control an outbreak of rabies. Journal of Applied Ecology 45:109-116. Marino, J., C. Sillero-Zubiri, A. Deressa, E. Bedin, A. Bitewa, F. Lema, G. Rskay, G. Banyard, and A. Fooks. 2017. Rabies and distemper outbreaks in the smallest Ethiopian Wolf population. Emerging Infectious Diseases 23: 2102-2104. Moehrenschlager, A. and C. Moehrenschlager. 2001. Census of swift fox (Vulpes velox) in Canada and northern Montana:2000-2001. Alberta Sustainable Resource Development, Fish and Wildlife Division. Alberta Species at Risk Report No. 24. Edmonton, AB. 21.pp. Moehrenschlager, A. 2002. Effects of ecological and human factors on the behaviour and population dynamics of reintroduced Canadian swift foxes (Vulpes velox), D Phil Thesis. University of Oxford, Oxford, UK. Moehrenschlager, A., D. W. Macdonald, and C. Moehrenschlager. 2003. Reducing capture-related injuies and radio collaring effects on swift foxes. In: Swift fox conservation in a changing world (eds. M. Sovada & L. Carbyn). Canadian Plains Research Centre, University of Regina, Regina, SK. Moehrenschlager, A., and D. W., Macdonald. 2003. Movement and survival parameters of translocated and resident swift foxes Vulpes velox. Animal Conservation 6:199-206 Moehrenschlager, A., B. L. Cypher, K. Ralls, R. List, and M. A. Sovada. 2004. Comparative ecology and conservation priorities of swift and kit foxes. In: D. W. Macdonald and C. Sillero-Zubiri 19 �(eds), The biology and conservation of wild canids, pp. 185-198. Oxford University Press, Oxford, UK Moehrenschlager, A., S. Alexander, and T. Brichieri-Colombi. 2006. Habitat suitability and population viability analysis for reintroduced swift foxes in Canada and northern Montana. Centre for Conservation Research, Calgary Zoolgical Society. Calgary, Alberta, Canada. Moehrenschlager, A. and C. Moehrenschlager. 2006. Population census of reintroduced swift foxes (Vulpes velox) in Canada and northern Montana 2006/2006. Centre for Conservation Research Report No. 1. Calgary Zoo, Calgary, Alberta, Canada. Moehrenschlager, A., R. List, and D. W. Macdonald. 2007. Escaping intraguild predation: Mexican kit foxes survive while coyotes and golden eagles kill Canadian swift foxes. American Society of Mammalogists 88:1029-1039. Moehrenschlager, A., and M. Sovada. 2016. Vulpes velox. The IUCN Red List of Threatened Species 2016: e.T23059A57629306 http://dx.doi.org/10.2305/IUCN.UK.20163.RLTS.T23059A57629306.en Osbourne, P. E. and P. J. Seddon. 2011. Selecting suitable habitats for reintroductions: variation, change and the role of species distribution modelling. In Reintroductions Biology: integrating science and management, eds J.G. Ewen, D.P. Armstrong, K.A. Parker & P.J. Seddon, Chapter 3. Wiley-Blackwell, Oxford, UK. Peters, W., M. Hebblewhite, M. Cavedon, L. Pedrotti, A. Mustoni, F. Zibordi, C. Groff, M. Zanin, and F. Cagnacci. 2015. Resource selection and connectivity reveal conservation challenges for reintroduced brown bears in the Italian Alps. Biological Conservation 186:123-133. Pruss, S., P. Fargey, and A. Moehrenschlager. 2008. National swift fox recovery strategy. Prepared in Consultation with Canadian Swift Fox Recovery Team. Species at Risk Act Recovery Strategy Series. Parks Canada Agency, 23 pp. Stratman, M. R. and J. A. Apker. 2014. Using infrared cameras and skunk lure to monitor swift fox (Vulpes velox). The Southwestern Naturalist 59:502-510. Timm, S.F., L. Munson, B. A. Summers, K. A. Terio, E. J. Dubovi, C. E. Rupprecht, S. Kapil, and D. K. Garcelon. 2009. A suspected canine distemper epidemic as the cause of a catastrophic decline in Santa Catalina Island foxes (Urocyon littoralis catalinae). Journal of Wildlife Disease 45:333-343. van de Bildt, M.W., T. Kuiken, A. M. Visee, S. Lema, T. R. Fitzjohn, and A. D. Osterhaus AD. 2002. Distemper outbreak and its effect on African wild dog conservation. Emergent and Infectious Diseases 8:211-213. 20 ��� https://cpw.cvlcollections.org/files/original/0159c1e81e1a37d31511338431e49f3b.pdf 96f8eafdd90bed14c32f38cb998e8c1b PDF Text Text Population Survey of Reintroduced Swift Foxes (Vulpes velox) in Canada and Northern Montana 2018 A. Moehrenschlager and H. Harris �Population Survey of Reintroduced Swift Foxes (Vulpes velox) in Canada and Northern Montana 2018 By A. Moehrenschlager and H. Harris Funded by: 1 �This publication may be cited as: Moehrenschlager, A. and H. Harris. 2019. Population survey of reintroduced swift foxes (Vulpes velox) in Canada and northern Montana 2018. Centre for Conservation Research, Calgary Zoological Society, Calgary, Alberta, Canada For copies of this report, contact: Centre for Conservation Research Calgary Zoological Society 1300 Zoo Road Calgary, Alberta, Canada T2E 7V6 e-mail: axelm@calgaryzoo.com Telephone: 403-232-7771 Fax: 403-232-9370 Internet: www.calgaryzoo.org 2 �Acknowledgments Primary acknowledgement goes to dedicated field crews who yielded precious information regarding the demography of swift foxes in Canada and Montana during the summers of 2015 and 2018. Canadian swift fox camera sampling in 2015 was conducted by Lauren Riches and Michael Miskolczi. Camera trapping in 2018 was led by Lacey Hebert in Canada and John Kuntz in Montana. Special thanks to Calgary Zoological Society staff who assisted with data analyses: Typhenn Brichieri-Colombi for GIS and demographic analysis and Lea Randall for occupancy analyses. Thank you to all government regulatory agencies and to all landowners for granting access to their land and for frequently assisting swift fox field crews in both countries. Finally, a huge thank you to ATB Financial Services, specifically to Sandra Huculak and Holly Regel. Swift fox surveys once again could only be done due to the leadership support of ATB Financial Services. Tireless support for the program allowed us to collect swift fox information that is now making a real difference for this precious species. 3 �Executive Summary Swift foxes (Vulpes velox) originating from reintroductions were surveyed in Canada during the summer of 2015 and in Canada and Montana during the summer of 2018. Demographic analyses, aimed at determining distribution and occupancy changes since 2015, were conducted to determine whether declines observed between 2005 and 2015 had continued. Occupancy analyses of successive data-sets from the winter of 2014/2015, the summer of 2015, and the summer of 2018 indicated that swift fox distribution and occupancy was similar between 2015 and 2018. While the swift fox population has not recovered to the high abundance levels of 2005/2006, stability despite a severe winter in 2017/2018 suggests that the population continues to harbour further recovery potential. To enable such recovery, actions to protect the current contiguous population should be reflected in conservation planning of all responsible government jurisdictions. 4 �1 Introduction Following previous years of growth, the contiguous reintroduced swift fox population in Canada and Montana decreased in abundance and occupancy between 2005/2006 and 2014/2015, while maintaining a similar extent of occurrence (Moehrenschlager and Moehrenschlager, 2018). Genetically the population seemed healthy (Cullingham and Moehrenschlager 2013), but demographic changes highlighted the need for additional monitoring to determine population trends since 2015. The objective of this monitoring was to assess whether there had been further changes to the distribution or occupancy of swift foxes on both sides of the international border. To address this objective, coordinated camera trapping was conducted in Canada during the summer of 2015 and in both Canada and Montana from June to September in 2018. Paired tests between cameras and live traps showed that the two methods are similar in their likelihood of detection (Moehrenschlager and Moehrenschlager 2018). Key questions to address overarching population trends included: 1. Was the probability of detection and occupancy similar between camera trapping in the winter of 2014/2015, and the summer of 2015? and 2. Has occupancy changed between the summers of 2015 and 2018 in Canada? 2 Methods 2.1 Sampling Approach During the summer of 2015, camera trapping was conducted with the aim of replicating sampling at all sites that were live or camera trapped during the winter of 2014/2015. During the summer of 2018, camera trapping was conducted at all sites in Canada and Montana where live and/or camera trapping had been conducted during the winter of 2014/15 (Moehrenschlager and Moehrenschlager 2018; Figure 1). 2.2 Survey Techniques Camera trapping methods replicated those of the winter 2014/2015 survey in Canada and Montana (Moehrenschlager and Moehrenschlager 2018). As in past surveys, the contiguous reintroduced swift fox population was subsampled in previously surveyed townships at previously sampled sites. Sampling involved camera traps, spaced at 1 km intervals along 5 km transects, placed as centrally as possible within respective townships, and spaced at least 6 km apart (Cullingham and Moehrenschlager, 2003; Cullingham and Moehrenschlager, in press). 5 �Camera trapping occurred for 3 consecutive 24-hour periods at sample sites. Attractants were mackerel oil or fatty acid attractants and scent-posts. Scent-posts consisted of lumber stakes measuring 5 x 5 x 40 cm and pointed on one end to facilitate mounting in the ground. Approximately 5 cm from the top of the stake, a plaster of Paris disk was mounted on the wood using a nail. Plaster disks were soaked in mackerel oil at the time of deployment. Reconyx motion-sensor cameras were mounted on custom-made iron stakes and positioned approximately 40 cm from the ground facing the scent-post. Cameras faced the scent-posts at a distance of 2 m (Camaclang et al. 2010; Moehrenschlager and Moehrenschlager 2018). 6 �Figure 1: Sampling sites using camera-traps in Canada and Montana during the summer of 2018. 7 �Figure 2: Comparative swift fox evidence between 2015 and 2018 (Canadian data-sets compare summer camera data over time; the Montana data-set compares 2018 summer camera data with 2014/2015 camera/winter data, as supported by preliminary analysis). 8 �2.3 Analyses Analyses of population distribution and demography among seasons or years focused on camera trapping results, but also included catch-and-release data from Montana during the winter of 2014/2015. As over 95% of swift fox detections in previous camera trapping surveys occurred between 6 pm and 6 am, and as catch-and release only occurs through trap checks focused at midnight and at 6 am, all camera occurrences were coded into respective 6 pm to midnight and midnight to 6 am sample intervals. If one or several swift fox images were recorded at a camera, that sample session was categorized as having swift fox evidence for the individual sample site. Since release or camera-traps spaced at 1 km intervals are spatially nonindependent (Moehrenschlager et al. 2007), camera data were lumped across the 6 sample sites within each township. Comparisons among time frames included the following input data sets: 1) Camera data in Canada during the winter of 2014/2015; 2) A combination of catch-and-release and camera data in Montana during the winter of 2014/2015; 3) Camera data in Canada during the summer of 2015; 4) Camera data in Canada and Montana during the summer of 2018. Since camera and catch-and-release sampling yielded similar probabilities of detection in a paired experiment of 38 sample sites during 2014/2015 (Moehrenschlager and Moehrenschlager 2018), analyses including 2014/2015 Montana catch-and-release data were integrated in a similar manner to camera-trapping data. Swift fox occupancy was estimated for comparative time frames using single-season models in Program Presence 12.6. Such analyses incorporate relative estimates of the probability of detection to approximate occupancy. To better understand occupancy estimates, we present associated cumulative 3 day-sample interval probabilities of detection for respective time frames. 3 Results 3.1 Sampling Effort During the winter of 2014/2015 189 townships were sampled in Canada and Montana through catch-and-release or camera trapping; 102 were sampled in Canada and 87 in Montana (Moehrenschlager and Moehrenschlager 2018). During the summer of 2015, 75 townships in Canada that had been sampled in the winter of 2014/2015 were sampled using cameras. During the summer of 2018, 147 townships in Canada and Montana that had been sampled using cameras or catch-and-release in the winter of 2014/2015 were sampled using cameras. 9 �3.2 Capture data and Occupancy Estimates Camera trap data from 67 townships in Canada found swift foxes on 17 transects in the summer of 2015 and on 14 transects in the summer of 2018 (Figure 2). In 75 replicated townships, the cumulative probability of detection and occupancy between the summer of 2015 (p. det. = 0.84; occupancy = 0.26 +/- 0.06), and the preceding winter of 2014/2015 (p. det. = 0.81; occupancy = 0.24 +/- 0.06) were similar. This suggests that the use of winter 2014/2015 Montana data was a sound proxy for range-wide comparisons between 2015 and 2018 summer surveys. Of 147 townships replicated between 2018 and 2014/2015 across the Canadian/Montana contiguous swift fox population, 36 townships had foxes in 2014/2015 compared to 40 with foxes in 2018 (Figure 2). The extent of occurrence including townships that were not replicates was 14,743 km2. Results suggest similar occupancy across this three year time frame (20142015 Montana winter with 2015 Canada summer data: p.det. = 0.80; occupancy = 0.26 +/- 0.04; 2018: p.det. = 0.91; occupancy = 0.27 +/- 0.04).0 4 Discussion The Canadian and contiguous Montana swift fox population is interdependent in its viability. Demographic growth since reintroductions began in 1983 has been accompanied by sufficient genetic diversity. Eastern and western portions of the reintroduced swift fox range show discreet genetic structure, suggesting that the dispersal between these population clusters is limited (Cullingham and Moehrenschlager 2013). Indeed, recent analyses illustrate that swift fox juveniles disperse in both directions across the international border. As dispersal rates for males have been documented to exceed 100 km., management decisions in one country can directly affect recovery potential within the other country (Cullingham and Moehrenschlager in press). The contiguous swift fox population declined between 2005/2006 and 2014/2015 by approximately 45% in township occupancy (Moehrenschlager and Moehrenschlager 2018). The winter survey of 2014/2015 included a comparison of catch-and-release and camera trapping which illustrated that camera-trapping is a sound correlate to catch-and-release data. This information has been advantageous for subsequent comparisons. Similarities in the probability of detection and occupancy between the 2014/2015 winter and 2015 summer in Canada suggest that the 2014/2015 combination of catch-and-release and camera data in Montana is likely a sound proxy for 2015 Montana summer demographics. Comparisons of the 2018 summer survey with previous time frames indicate that the population has neither decreased nor increased since 2015 in terms of distribution and occupancy. One primary hypothesis for the decline between 2005/2006 and 2014/2015 was that the severe winter of 2010/2011 had a population limiting effect. The winter of 2017/2018 10 �was similar in severity in Canada and Montana. It is possible that the swift fox population recovered from 2015 to 2017, but was affected negatively by the subsequent winter. Swift fox populations in Canada and Montana are not yet sustainable without close monitoring and management. Future recovery plans on state, provincial, and federal levels of both countries should emphasize research, protection, and conservation actions that continue to increase the density, distribution, and connectivity of the reintroduced population. Management decisions since 2015, such as reduction of the fur harvest quote in Montana by two thirds, and precluding captures of swift foxes in this population as a source for new reintroductions have likely supported population stability. These precautionary actions should continue until demographics improve further, so that recovery of this iconic prairie carnivore can continue effectively in Canada and in the United States. 11 �Literature Cited Cullingham, C. I., and A. Moehrenschlager. 2013. Temporal analysis of genetic structure to assess population dynamics of reintroduced swift foxes. Conservervation Biology 27:1389-1398. Cullingham, C. I. and A. Moehrenschlager. In press. Genetics of reintroductions across borders: Swift fox movements within distinct subpopulations require co-management by Canada and the United States. Animal Conservation Moehrenschlager, A. and C. Moehrenschlager. 2018. Population survey of reintroducted swift foxes (Vulpes velox) in Canada and northern Montana 2014/2015. Centre for Conservation Research, Calgary Zoological Society, Calgary, Alberta, Canada. 12 �� https://cpw.cvlcollections.org/files/original/c0bfd8e157793bb43bf23cf9292dd3a4.pdf 06c50c5d972d2455f800d65cdfd3314f PDF Text Text ������ https://cpw.cvlcollections.org/files/original/95d23fa3f0dfa32f87402ef9a19adb63.pdf bf227aeab2b7299aa3f65072cc645ce9 PDF Text Text March 2016 == s...=::= =--T & ~• • _,......:::;-_: parks wildfif~ Wild Thing: The Other Foxes Though less prevalent than reds and grays, swift and kit foxes are found in some areas of Texas. By Jonah Evans You may have seen two familiar species of foxes in Texas, but the state actually has four types. Gray and red foxes, the most common, are found across most of the state. The swift and kit fox are much more specialized — the swift fox occurs only in the Panhandle, and the kit fox is found in the deserts of West Texas. Swift and kit foxes are closely related; scientists have long argued whether they are different species. For now, most seem to agree that although the two sometimes hybridize where their ranges overlap, they are separate species. The kit fox of West Texas is less than half the size of a gray fox and at only 3 to 6 pounds is the smallest of Texas’ foxes. Highly adapted for living in open desert with large ears for keeping cool, the kit fox can also get all the water it needs from its diet, which consists of kangaroo rats, mice, lizards, birds, insects, berries, seeds, and other desert flora and fauna. Swift foxes are slightly (but not much) larger than kit foxes and are found in the short-grass prairies of the Texas Panhandle. They eat mostly rabbits, small rodents, birds, lizards and insects. �Both swift and kit foxes use burrows, and it isn’t uncommon to see a family group of five or more foxes sitting outside the opening during dawn and dusk hours. Burrows serve not only as shelter from the elements, but also as safe havens to escape coyotes and other predators. While the swift fox historically occurred in 77 counties in North Texas, today it is found only in the two most northwestern Panhandle counties: Dallam and Sherman. More research is needed to determine if it still occurs in other parts of the state. This apparent decline is not limited to just Texas — it is estimated that swift foxes currently inhabit only 40 percent of their historic range. This decline is primarily attributed to shrub encroachment, changing land uses, historical hunting practices and predation from coyotes. Very little research has been done on kit foxes in Texas. Anecdotal accounts from oldtimers in West Texas seem to agree that they used to be much more common than they are today. They are known to be declining in many parts of their range, though not to the extent of the swift fox. TPWD is currently working to learn more about swift and kit foxes in the state, but in a large state with primarily private land, finding a small fox isn’t easy. Reports from public sightings are a valuable source of information, so please share your observations on iNaturalist.com or by contacting TPWD at jonah.evans@tpwd.texas.gov. Common Names: Swift fox, kit fox �Scientific Names: Vulpes velox, Vulpes macrotis Habitat: Desert scrub, chaparral and grasslands Diet: Rabbits, prairie dogs, ground squirrels, mice, birds, reptiles, amphibians, berries and seeds Did You Know? Biologists had thought swift and kit foxes might be the same species, but now it's generally agreed that they are separate species. � Dublin Core The Dublin Core metadata element set is common to all Omeka records, including items, files, and collections. For more information see, http://dublincore.org/documents/dces/. Title A name given to the resource Swift Fox Conservation Team Description An account of the resource Swift Fox Conservation Team reports and newsletters Dublin Core The Dublin Core metadata element set is common to all Omeka records, including items, files, and collections. For more information see, http://dublincore.org/documents/dces/. Title A name given to the resource Swift Fox Conservation Team: Agency / Partner Documents Type The nature or genre of the resource Text