https://cpw.cvlcollections.org/files/original/51fe0817135c12325d486b515051e3d2.pdf e00f8d0767c1240024ab0ad9941d56cd PDF Text Text COLORADO PARKS AND WILDLIFE Brown-capped Rosy-Finch Population Assessment Pilot Study 2018 Amy Seglund, Jon Runge, Michelle Flenner, and Kathryn Bernier 12/1/2018 �2 Introduction The Brown-capped Rosy-Finch (BCRF; Leucosticte australis) is a small, hardy passerine that resides in high elevation alpine environments throughout the year. It has the smallest range of the three closely related species of Rosy-Finch (L. atrata, L. tephrocotis) with it being almost completely endemic to Colorado, though populations extend into northern New Mexico and southern Wyoming (Johnson et al. 2000). Their breeding habitat occurs predominantly on U.S. Forest Service lands with many acres in designated wilderness based on Colorado Parks and Wildlife predicted range model developed for the species (Figure 1). Suitable nest sites are normally found near steep cliff faces or rocky, talus slopes away from excessive human disturbance (Johnson et al. 2000). The species is monogamous with males defending floating territories around their mates. All Rosy-Finch species have been found to have skewed sex ratios with females being the limiting resource (Shreeve 1980a, Johnson et al. 2000). BCRF produce only one clutch per season with clutch size small varying from 3-6 eggs (mean 4.22; Johnson et al. 2000). Like other alpine avian species this reduced reproductive output, in response to high elevation environmental conditions, has lead to increased plasticity and alteration of life history traits to favor adult survival over high reproductive output to maintain population numbers (Bears et al. 2009, Martin 2014). During the early breeding season, BCRFs forage at the lower edges of snowfields where insects and seeds are deposited and in fell fields, cliffs, and alpine tundra in close proximity to nesting sites (Kingery 1998; Johnson et al. 2000, Stanek 2009). The BCRF remains at high elevations throughout the year unless severe winter storms events push them down to lower elevations to feed on exposed vegetation or at well stocked bird feeders (Johnson et al. 2000). The BCRF along with other alpine endemic species (e.g., white-tailed ptarmigan - Lagopus leucura, Seglund et al. 2018) are thought to be susceptible to climate change. The Climate Change in Colorado Report (Lukas et al. 2014) found that the state has warmed substantially in all four seasons over the last 30 years (>2.0 °F), and this same report predicts that future warming could be as high as 6.5°F by 2050. Alpine species are thought to be more susceptible to climate change because increases in temperatures are accentuated at higher elevations (Diaz and Eischeid 2007). Specific concerns for the BCRF are depletion of late lying snowfields as temperatures increase and precipitation patterns are altered, changes in insect abundance and phenology, and increases in fragmentation of the alpine ecosystem due to predicted treeline expansion upslope (Rosenberg et al. 2016). Declining BCRF population trends have been reported using Christmas Bird Count survey information (Johnson et al. 2000), however, these data may not provide an accurate assessment of �3 the species’ population trends as winter numbers are eruptive and nomadic. Changes in breeding information from the first Breeding Bird Atlas in Colorado to the Atlas II showed that the general distribution of BCRFs had not changed, but there was a 3% decline in the number of breeding blocks reported to be occupied (Wickersham 2016). The BCRF is a Tier 1 species in Colorado’s State Wildlife Action Plan (SWAP; CPW 2015) and is one of 19 species on the Partners in Flight’s Red List (Rosenberg et al. 2016). Little has been published regarding BCRF basic biology, and lack of information was identified as a primary concern for this species in the Colorado SWAP. Estimating population parameters was listed as one of the highest priority conservation actions in the plan. Potential threats impacting the species at both the summer breeding range and at wintering sites also needs to be investigated to understand potential negative impacts to BCRF populations. The current pilot project was undertaken to evaluate populations at both local and statewide levels to develop a baseline population status assessment. The baseline data collected from this survey effort will help inform future monitoring efforts by supplying derived estimates of detection probabilities, abundance estimates, and percent occupancy with sufficient precision to evaluate the current species status and help predict conservation concerns as climate change impacts become more prominent and human use of the alpine increases. This project was also designed to provide information on nest site selection, breeding behavior, and habitat use during the breeding season to gain a better understanding of the species in Colorado. Sampling Frame We developed a predicted range model for the BCRF (Figure 1). This model incorporated areas > 3292 m in elevation and Colorado Gap Analysis Project (CO-GAP; Schrupp et al. 2000) vegetation types that included Mixed Tundra, Meadow Tundra, Prostrate Shrub Tundra, Bare Ground Tundra, Exposed Rock, Shrub Dominated Wetland/Riparian, and Graminoid/Forb Dominated Wetland. This model was intersected with a terrain roughness layer, pulling out the extreme change in elevation areas (i.e., cliffs). All areas that occurred within 500 m of cliffs were included in the sampling frame as breeding BCRF have been found to preferentially forage within 300 to 400 m of cliff areas during the breeding season (Stanek 2009). The final BCRF predicted range model encompassed 7255 km2, with the dominant land manager being the United States Forest Service. We used the predicted range model to overlay a 4x4 km grid and calculated the percentage of habitat within each grid cell. We selected only grids that had a minimum of 60% suitable habitat to randomly select 80 grids to sample across the state. Within each grid cell, we identified an alpine basin (i.e., restricted area within which the rock strata dip toward the center forming a hollow �4 depression that holds and drains water) to conduct line transect surveys (Figure 2; example photo 1). Basins sampled varied in size, shape and directional position which impacted persistence of snow cover, vegetation, and microsite weather patterns. �5 Figure 1. Predicted range model used to define the Brown-capped Rosy-Finch sampling frame. Grid cells selected for sampling in 2018 in Colorado are identified. �6 Figure 2. Example of survey transect sampled at the Lost Lake alpine basin for Brown-capped RosyFinch during 2018 in Colorado. �7 Photo 1. Example of the Lost Lake alpine basin in Colorado surveyed in 2018 for Brown-capped RosyFinch. �8 Field Methods Two observers conducted separate line transect surveys within a basin each independently recording BCRFs. One observer walked high under cliffs and in talus while the second observer walked a lower section along and into the alpine tundra. Depending on topography and access, at times observers had to alter their line of travel and switched positions for survey efforts, though we attempted to maintain our set designated transects as much as possible. Each observer completed an individual transect and recorded their tracks using assigned Garmin GPS units. Transects followed the contour of the topography and varied in length depending on the size of the basin surveyed. We avoided double counting birds by having observers radio one another when a bird observation was made to ensure the second observer had not also detected a BCRF. Since BCRFs maintain floating territories around a nest site (Johnson et al. 2000) and we followed the topography of a basin, it became possible to map out breeding pairs and avoid recounting birds encountered. While surveying along transects, observers recorded auditory and visual detections of BCRFs for occupancy and measured distance perpendicular to the transect (using laser range finders) to each individual or group of birds seen. For observations we record behavior of birds (e.g., foraging, courtship behavior, nest building etc.), location of nesting sites, and insects and plants that we saw the birds foraging on to increase our knowledge of the basic biology of the species. Finally, observers recorded sex of each bird encountered if it could be determined and habitat it was using (e.g., snowfield, talus, tundra etc.). Sampling of transects occurred from the beginning of June (6/8/2018) when access was feasible and continued until early August (8/12/2018) when we began to detect fledglings. Surveys were conducted starting around 7:30 am when the sun was up and the alpine began to warm. Birds rarely became active until insects began to move around thus we waited until later in the morning to initiate surveys. We did not restrict a time when surveys had to be completed as we found that birds were active all day, thus surveys continued until observers finished running a transect. Analysis We used Distance analysis (Buckland et al. 1993) with program DISTANCE (v 6.2 rel. 1, Thomas et al. 2010) to estimate density of Brown-capped Rosy-Finch (birds per km2). Distance sampling and analysis included four assumptions (Buckland et al. 1993). 1. All animals directly on the transect line are detected. �9 2. All animals are noted and distances are measured at their original detection point. 3. Measurements from the transect line to the animals are exact. 4. Double counting of animals on a transect line does not occur. To satisfy assumption #3, we only included observations taken that had exact distance to birds. All other assumptions were met. Many observations included more than one bird, so we used the ‘cluster of objects’ option in DISTANCE for analysis. We initially analyzed data with the actual distance measured perpendicular to transect lines, but the Kolmogorov-Smirnoff test showed that goodness of fit assumptions were not satisfied (p < 0.0001 for all models fit). We therefore binned data into 15 m intervals; 0 to 225 m from the transect line with the furthest bin distance being 225280 m (the maximum distance measured for an observation was 278 m). Binning the data in this manner enabled some models to satisfy the goodness of fit test (p > 0.05). We additionally provided an effective transect width to quantify the distance at which the number of unseen BCRFs within the effective strip equals the number of detected birds outside that distance. This strip width applied to only one side of the transect line, thus it would need to be doubled if applying the measure to both sides of a transect line (Buckland et al. 1993). For model selection, we used all four key functions available in program DISTANCE (halfnormal, hazard rate, negative exponential, and uniform) in combination with the three series expansions available (cosine, simple polynomial, and Hermite polynomial). This resulted in 12 models in the largest possible model set. We eliminated from contention all models that failed a chisquare GOF test used for binned data, and if two models produced identical results, one was eliminated. We used Akaike’s Information Criterion (Akaike 1973), hereafter AIC, to rank the models and to develop Akaike weights (AICw), which quantify the relative support of each model on a scale of 0-1 (Burnham and Anderson 2002). We used the AICw to average density estimates across all models remaining to produce a statewide density estimate for BCRF that accounted for uncertainty in model selection. We applied the predicted range model area of 7255 km2 to estimate total birds in Colorado with 95% confidence intervals. Results We were able to sample 52 basin areas within grid cells during the summer of 2018 with transect length varying from 1.69 km to 7.12 km. All of the basins sampled were occupied (100% occupancy). We recorded 434 observations of male BCRFs, 170 observations of females, 60 �10 juveniles, and 389 birds of unknown sex. It was apparent that some basins held very high densities of birds (>70 individual detections) and some with substantially lower numbers where we recorded only one or two observations of a BCRF. We had very high detection probability for occupancy with only one observer not detecting a BCRF during one survey. We located eight nest sites; one in a talus slope (photo 2) and the remaining seven in cliffs or rock crevices. Most nests were located while observing a female building the nest cup and the male guarding the female as she worked. Eggs were found in two of the nests; one with two eggs and the other with three eggs. Basins sampled had little human disturbance occurring where transects were completed. Disturbances were relegated to abandoned mines and hiking trails. Predators detected in the area included common raven (Corvus corax), golden eagle (Aquila chrysaetos), prairie falcon (Falco mexicanus), peregrine falcon (Falco peregrinus), red-tailed hawk (Buteo jamaicensis), coyote (Canis latrans), and long- tailed weasel (Mustela frenata). We retained 346 observations for density estimates for which exact distance was measured; there were an additional 115 observations that could not be included because exact distance was not measured due to observations being a flyover or auditory detection, or detections were greater than 300 m from the transect line. Of the 12 models fit, five demonstrated sufficient goodness-of-fit and also had unique results. Six models failed the chi-square goodness of fit test, and two models showed identical results, thus only one of those was retained for the model set. All five models were within 2.21 AIC units of the best ranked model. AIC weight ranged from 0.13-0.35 (Table 1). The top-ranked model used a negative exponential key function with a simple polynomial expansion (ΔAIC = 0.00, AICw = 0.35, density estimate 22.3 birds per km2, density SE = 2.8). The model-averaged estimate for density was 20.3, with a SE of 3.0 and a 95% confidence interval of 15.2-27.2 (Table 1). Extrapolated to the 7255 km2 suitable habitat in Colorado, the statewide population estimate was 150,058 individuals (95% CI: 112,738-199,730). The detection functions (Figures 3-7) showed the expected 1.0 (i.e., 100%) detection probability on the transect line, with decreasing detection probability for birds further from the line. For all detection functions, detection probability had decreased to 0.24-0.28 near 40 m from the transect line. Effective strip width was 32-39 m depending upon model (Table 1), thus if applied to both sides of the transect line, the effective strip width would be 65-78 m, assuming both sides of the transect have equal sighting probability (but see Discussion below). �11 Photo 2. Brown-capped Rosy-Finch nest in Colorado located in talus slope. �12 Table 1. Results from DISTANCE analysis where k is the number of parameters in the model, ESW is effective strip width, and CV is percent coefficient of variation. Model Negative exponential, simple polynomial Hazardrate, Hermite polynomial Negative exponential, cosine Hazardrate, simple polynomial Hazardrate, cosine k ΔAIC AICw ESW Density/km2 Density 95% LCL Density 95% UCL Density SE Density CV 3 0 0.35 31.98 22.32 17.45 28.55 2.81 13% 3 0.48 0.27 36.85 19.7 14.99 25.89 2.75 14% 3 1.87 0.14 33.56 21.37 16.79 27.19 2.63 12% 3 2.05 0.12 37.87 19.26 14.87 24.93 2.54 13% 2 2.21 0.12 39 18.81 14.7 24.07 2.37 13% Model Averaged 20.68 15.54 27.53 3.03 15% Statewide Total 150,058 112,739 199,730 �13 Figure 3. Detection function for model with negative exponential key function, simple polynomial series (ΔAIC=0.00). �14 Figure 4. Detection function for model with Hazard-rate key function, Hermite polynomial series (ΔAIC=0.48). �15 Figure 5. Detection function for model with negative exponential key function, cosine series (ΔAIC=1.87). �16 Figure 6. Detection function for model with hazard-rate key function, simple polynomial series (ΔAIC=2.05). �17 Figure 7. Detection function for model with hazard-rate key function, cosine series (ΔAIC=2.21). �18 Next Steps The 2018 survey effort was the first step in an attempt to assess the population status of the BCRF in Colorado as well as gain a better understanding of breeding behavior and resource needs. Contemporary surveys indicated that the BCRF occupied all mountain ranges in Colorado, though there appears to be areas with higher densities than at other local sites. Quantifying these differences will be our next step. Funding is available for an additional two years of survey efforts and with the additional surveys, we will complete sampling of the randomly selected 80 grids cells (28 remain) and revisit sites across the spectrum of bird densities detected to evaluate if 2018 was an exceptional year for occupancy and density estimates. We will make an effort to locate additional nests sites to continue to characterize nesting substrates. We will work to determine predictor variables of occurrence for the species and develop a model of resource use to help decipher “hot spots” of occurrence. Our survey methodology of placing transects within a basin following the topography along a cliff face potentially biased our density estimates because cliffs obscure available sampling area resulting in a non-random effect of possible observation distances (Buckland et al. 2004). Our 2018 estimates may have also been inflated because transects positioned near cliffs sampled preferred areas used by BCRF during breeding; abundance of the focal species is likely higher here than areas further from cliff walls. Stanek (2009) found that breeding BCRF foraged predominantly within 300 to 400 m of cliffs. Adult BCRFs have been found to feed several kilometers from a nesting area, but are less numerous as distance from cliffs increases (Shreeve 1980b). To improve our density estimates and reduce bias in future surveys, observers will record the side of transect the bird was observed (i.e., cliff or non-cliff side). Though our density estimate for the state may have incurred some biases, the estimate appears to represent a realistic approximation based on winter banding efforts currently being undertaken by Colorado Parks and Wildlife and the University of Santa Cruz in which high capture but low recapture rates appear to indicate a fairly sizable population of BCRF in Colorado. Our estimates are much lower than the one study that has attempted to quantify densities of breeding BCRFs in the San Juan Mountains (Stanek 2009). This study estimated an average density of 1.91 birds/ha (95% C.I. of 1.77 to 2.07) whereas we estimated 0.21 birds/ha (95% CI of 0.16 to 0.28) statewide. A proposal and protocol will be developed based on this pilot project to outline future survey efforts. A graduate student will be hired to help develop the protocol for future surveys and will be responsible for analysis and publication of findings. The final project proposal will be available April 2019. The overall objective of the next steps in our research is to ascertain trends in �19 populations, develop a resource selection model for breeding birds using identified predictor variable of occurrence, and better evaluate nest site selection and success. The specific objectives are: 1. Continue occupancy surveys and distance sampling along transects at identified grid cells to assess populations statewide. 2. Develop hypotheses using resource selection models to identify “hot spots” of breeding habitat use and use this model to predict shifts in available habitat due to climate change and increased recreation in the alpine areas in Colorado. 3. Provide data on nest site selection and reproductive output to increase our understanding of important nest site characteristics and nest success rates. There is currently very limited data on nesting of the BCRF. �20 Literature Cited Akaike, H. 1973. Information theory as an extension of the maximum likelihood principle, Pages 267 – 281 in B.N. Petrov and F. Csaki, (eds.) Second International Symposium on Information Theory. Akademiai Kiado, Budapest. Bears, H. K. Martin and G.C. White. 2009. Breeding in High-Elevation Habitat Results in Shift to Slower Life-History within a Single Species. J. of Animal Ecology. Vol. 8 No. 3. Buckland, S.T., D.R. Anderson, K.P. Burnham, and J.L. Laake. 1993. Distance Sampling: Estimating Abundance of Biological Populations. Chapman and Hall, London, reprinted in 1999 by RUWPA, University of St. Andrews, Scotland. 446pp. Burnham, K.P., and D.R. Anderson. 2002. Model Selection and Multimodel Inference: A Practical Information-Theoretic Approach. 2nd Edition. Springer, New York. 488pp. Colorado Parks and Wildlife. 2015. State Wildlife Action Plan: Prepared for the citizens of Colorado and Its Visitors by Colorado Parks and Wildlife. Denver, CO. cpw.state.co.us Diaz, H.F. and Eischeid, S.H. 2007. Disappearing “alpine tundra” Köppen climactic type in the western United States. Geophysical Research Letters. VOL. 34, L18707, doi:10.1029/2007GL031253 2007 Johnson, R. E., P. Hendricks, D. L. Pattie, and K. B. Hunter (2000). Brown-capped Rosy-Finch (Leucosticte australis), version 2.0. In The Birds of North America (A. F. Poole and F. B. Gill, Editors). Cornell Lab of Ornithology, Ithaca, NY, USA. https://doi.org/10.2173/bna.536 Lukas, J., J. Barsugli, D. Doesken,. I. Rangwala, and K. Wolter. 2014. Climate Change in Colorado: a synthesis to support water resources management and adaptation. A report for the Colorado Water Conservation Board. Western Water Assessment, Cooperative Institute for Research in Environmental Sciences, University of Colorado Boulder. Martin, K. 2014. Avian strategies for living at high elevations: life history variations and coping mechanisms in mountain habitats. BOU Proceeding–Ecology and conservation of birds in upland and alpine Habitats http://www.bou.org.uk/bouproc--‐net/uplands/martin.pdf Rosenberg,K. V., J. A. Kennedy, R. Dettmers, R. P. Ford, D. Reynolds, J.D. Alexander, C. J. Beardmore, P. J. Blancher, R. E. Bogart, G. S. Butcher, A. F. Camfield, A. Couturier, D. W. Demarest, W. E. Easton, J.J. Giocomo, R.H. Keller, A. E. Mini, A. O. Panjabi, D. N. Pashley, T. D. Rich, J. M. Ruth, H. Stabins, J. Stanton, and T. Will. 2016. Partners in Flight Landbird Conservation Plan: (2016) Revision for Canada and Continental United States. Partners in Flight Science Committee. �21 Schrupp, D.L., W.A. Reiners, T. Thompson, F. D’Erschia, T. Owens, K. Driese, C. Buoy, A. Cade, J. Kindler, J. Lowsky, L. O’Brien, L. Satcowitz, M. Wunder, J. Stark and S. Russo. 2000. The Colorado GAP analysis project: a geographic approach to planning for biological diversity. CDOW and USGS final report. Seglund, A.E., P.A. Street, K. Aagaard, J. Runge, and M. Flenner. 2018. Southern White-tailed Ptarmigan (Lagopus leucura altipetens) Population Assessment and Conservation Considerations in Colorado. Colorado Parks and Wildlife Final Report. Shreeve, D. F. (1980a). Differential mortality in the sexes of the Aleutian Gray-crowned Rosy-Finch. Am. Midl. Nat. no. 104:193-197. Shreeve, D. F. (1980b). Behaviour of the Aleutian Grey-crowned and Brown-capped Rosy Finches Leucosticte tephrocotis. Ibis no. 122:145-165. Stanek, J.R. 2009. Breeding habitat selection by rosy-finches in the San Juan Mountains, Colorado, M.S., Department of Zoology and Physiology University of Wyoming. Thomas, L., S.T. Buckland, E.A. Rexstad, J. L. Laake, S. Strindberg, S. L. Hedley, J. R.B. Bishop, T. A. Marques, and K. P. Burnham. 2010. Distance software: design and analysis of distance sampling surveys for estimating population size. Journal of Applied Ecology 47: 5-14. Wickersham, L.E. 2016. The second Colorado breeding bird atlas. Colorado Bird Atlas Partnership and ColoradoParks and Wildlife, Denver, CO. �22 � 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 Reports Text A resource consisting primarily of words for reading. Examples include books, letters, dissertations, poems, newspapers, articles, archives of mailing lists. Note that facsimiles or images of texts are still of the genre Text. 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 Brown-capped rosy-finch population assessment pilot study Description An account of the resource The Brown-capped Rosy-Finch (BCRF; <em>Leucosticte australis</em>) is a small, hardy passerine that resides in high elevation alpine environments throughout the year. It has the smallest range of the three closely related species of Rosy-Finch (<em>L. atrata</em>, <em>L. tephrocotis</em>) with it being almost completely endemic to Colorado, though populations extend into northern New Mexico and southern Wyoming (Johnson et al. 2000). Their breeding habitat occurs predominantly on U.S. Forest Service lands with many acres in designated wilderness based on Colorado Parks and Wildlife predicted range model developed for the species (Figure 1). Suitable nest sites are normally found near steep cliff faces or rocky, talus slopes away from excessive human disturbance (Johnson et al. 2000). The species is monogamous with males defending floating territories around their mates. All Rosy-Finch species have been found to have skewed sex ratios with females being the limiting resource (Shreeve 1980a, Johnson et al. 2000). BCRF produce only one clutch per season with clutch size small varying from 3-6 eggs (mean 4.22; Johnson et al. 2000). Like other alpine avian species reduced reproductive output, in response to high elevation environmental conditions, has lead to increased plasticity and alteration of life history traits to favor adult survival over high reproductive output to maintain population numbers (Bears et al. 2009, Martin 2014). Creator An entity primarily responsible for making the resource Seglund, Amy Runge, Jon Flenner, Michelle Bernier, Kathryn Subject The topic of the resource Brown-capped Rosy-Finch (BCRF) <em>Leucosticte australis</em> Colorado New Mexico Southern Wyoming Breeding habitat Population Extent The size or duration of the resource. 22 pages Date Created Date of creation of the resource. 2018-12-01 Rights Information about rights held in and over the resource <a href="http://rightsstatements.org/vocab/NoC-NC/1.0/">No Copyright - Non-Commercial Use Only</a> Type The nature or genre of the resource Text Format The file format, physical medium, or dimensions of the resource application/pdf Language A language of the resource English Publisher An entity responsible for making the resource available Colorado Parks and Wildlife