November 28, 2006
Movements, Connectivity, and Resource Selection of Rocky Mountain Bighorn Sheep
By DeCesare, Nicholas J; Pletscher, Daniel H
Species that exist in naturally fragmented subpopulations can maintain long-term viability through interpopulation connectivity and recolonization of suitable habitat. We used radiotelemetry to study movements of 3 herds of bighorn sheep (Ovis canadensis) that recently colonized previously unoccupied parts of western Montana. These herds also provided a unique opportunity to compare resource- selection patterns in newly colonized habitats, and we used logistic regression in a global information system framework to generate predictive models for females in each herd. We detected relatively long (19- to 33-km) extra-home range movements by males in all 3 herds, and connectivity with nearby bighorn and domestic sheep herds. An information-theoretic approach to model selection revealed greater differences in resource selection among herds than anticipated. Initial evaluation of resource-selection models by resubstituting data showed excellent predictive accuracy (P ≤ 0.002), but testing models across sites gave mixed results, and in many cases, poor fit (0.001 ≤ P ≤ 0.960). High vagility of males and variability in resource selection by females suggests increased potential for future recolonization and connectivity.
Species with specialized habitat requirements often are segregated into disjunct subpopulations because of both natural and anthropogenic fragmentation of suitable habitat. In the spectrum of large herbivores, bighorn sheep (Ovis canadensis) provide a good example of such a species (Bleich et al. 1990; DeForge et al. 1979). Unlike other North American ungulate species, bighorn sheep have not fully recovered from dramatic population declines during the 19th century. After the extinction of 1 subspecies (O. c. auduboni) and local extirpations in several states, restoration efforts brought population estimates in the contiguous United States to 49,000 in 1991 (Valdez and Krausman 1999). Although overall numbers are improving, local populations remain fragmented and small (Douglas and Leslie 1999), with the majority at numbers far below minimum viable population estimates (Berger 1990; Krausman et al. 1992; Lande 1995).
Habitat specialists are typically poor colonizers (MacArthur and Wilson 1967) and sheep have been perceived as such, with strong site fidelity (Geist 1970, 1971; Shackleton et al. 1999). Thus, costly reintroductions and supplemental transplants have been a primary tool in augmenting populations during recent decades (Hansen 1980). Shackleton et al. (1999) suggested that anthropogenic changes to the landscape have forced this species into more specialized habitats than previously occupied. In addition, bighorn sheep may have a greater capacity for extra-home range or dispersal movements than previously envisioned (Bleich et al. 1996; Festa-Bianchet 1986b; Schwartz et al. 1986). Schwartz et al. (1986) reviewed a host of studies noting intermountain movements of desert bighorn (O. c. nelsoni) males and females across highways and rivers, and Bleich et al. (1996) used mitochondrial DNA to reveal metapopulation dynamics of desert bighorn populations within 15 km of one another. Interpopulation movements of Rocky Mountain bighorn sheep (O. c. canadensis) in northern ecosystems are less well documented. Festa- Bianchet (1986b) documented males up to 48 km from the site of their capture in southern Alberta, and Singer et al. (2000a) monitored 24 cases of dispersal of translocated Rocky Mountain bighorn sheep from release patches to neighboring unoccupied patches. Although movements of this type can enable gene flow and rescue effects, they also may serve as a disease vector, to which bighorn populations have proven very vulnerable (Bunch et al. 1999; DeForge et al. 1979; Fitzsimmons et al. 1995; Sausman 1984). As a resuit, issues of habitat quality (Johnson and Swift 2000; Turner et al. 2004), connectivity (Krausman 1997; Singer et al. 2000b), and disease (Gross et al. 2000) all play a role in the viability of each local population.
Existing bighorn populations in western Montana include native and reintroduced lineages (McCarthy 1996). During the 1980s bighorn sheep colonized 3 previously unoccupied areas of western Montana. We radiotracked animals from these young herds to study movements and connectivity in a northern landscape occupied by both native bighorn and domestic sheep. These populations also provided a unique opportunity for studying the degree of similarity in resource selection at newly pioneered habitats. We developed and evaluated resource-selection functions for each herd, and compared and rigorously tested all models with independent data from the other 2 herds, as recommended by Verbyla and Litvaitis (1989), Fielding and Bell (1997), and Boyce et al. (2002). We quantitatively asked whether recolonized populations selected resources similarly, and assessed the effectiveness of applying site-specific models across areas.
MATERIALS AND METHODS
Study area.-We studied 3 newly colonized bighorn sheep populations (Bearmouth, Garrison, and Skalkaho) in the Rocky Mountains of western Montana at elevations of roughly 1,100-2.000 m. The Beurmoulh (4643'N, 11327'W) herd occupied parts of the southern Garnet Range, 50 km east of Missoula, Montana, and numbered approximately 12-17 individuals during the study period (DeCesare 2002). The Garrison (4631'N, 11250'W) herd was located in the northeastern foothills of the Flint Creek Range, 100 km east of Missoula, and contained an estimated 55-74 individuals (DeCesare 2002). The Skalkaho (4610'N, 11359'W) herd was in the western Sapphire Mountains, 75 km south of Missoula, and consisted of roughly 58-62 individuals (DeCesare 2002). Each site was roughly 70 km from the other 2 sites, and other bighorn herds existed in between.
All 3 herds occupied low- to midelevation habitats where common native grassland species included bluehunch wheatgrass (Agropyron spicatum). Idaho fescue (Festucu idahoensis), and rough fescue (Festuca scabrella). Shrub species included sagebrush (Artemesia tridentata) and mountain mahogany (Cercocarpus ledifolius), and forests were dominated by open stands of ponderosa pine (Pinus ponderosa) and denser stands of Douglas-fir (Pseudotsuga menziesii). Exotic species such as cheatgrass (Bromus tectorum) and spotted knapweed (Centaurea maculosa) also were common.
Capture and monitoring.-We captured 16 adult female and 5 adult male bighorn sheep at 3 sites in March 2001, using a net-gun tired from a helicopter (Krausman et al. 1985) and fitted all animals with Telonics MOD-500 VHF radiocollars (Telonics, Mesa, Arizona). We attempted to capture animals from different subgroups within each herd and radiocollared 2, 7, and 7 females in Bearmouth, Garrison, and Skalkaho herds, respectively. Although males were valuable for movement studies, small sample sizes of radiocollared males (n = 2, 2, and I, respectively) prevented inclusion of male data in resourceselection analysis. We used primarily ground radiotelemetry to collect 1.034 locations for collared females and 219 for males between March 2001 and August 2002. All handling and monitoring of animals followed guidelines of the American Society of Mammalogists (Animal Care and Use Committee 1998) and were approved by an institutional animal care and use committee.
Movements.-We quantified movements in reference to home ranges of resident females. We calculated 5% fixed kernel estimates using the Animal Movement 2.0 Extension (Hooge and Eichenlaub 1997) for ArcView 3.2a software (ESRI, Redlands, California) and all female locations for each herd. We used the centroids of these core female home ranges as reference points for measuring movement. We measured the furthest distance traveled by each male and female from core female home ranges and documented all associations with nearby bighorn and domestic sheep herds.
TABLE 1.-Descriptions of variables derived from a global information system to quantify resource selection of bighorn sheep, western Montana, 2001-2002.
Resource selection.-We maintained a minimum of 3 days between bighorn relocations to ensure temporal independence of data (McNay et al. 1994; Otis and White 1999: Swihart and Slade 1985; Swihart et al. 1988). A modified nearest-neighbor analysis of spatial independence revealed a lack of independence in relocations among individuals (DeCesare and Pletscher 2005), which can bias habitat- selection results (Dasgupta and Alldredge 2000; Millspaugh et al. 1998). Therefore, we pooled data within each herd and the sampling unit became the location of a group of sheep, which included 1 or more collared animals. Sample size was reduced from 1,034 individual female locations to 523 group locations.
Bighorn sheep generally select resources differently by season (Festa-Bianchet 1986a, 1986b; Geist 1971; Shannon et al. 1975). We divided data for each herd into 3 biologically meaningful seasons (winter, lambing, and autumn) using shifts in movements or behavior to delineate each season. Generally, the lambing season lasted from early May through late July, the autumn season from early August throughlate November, and the winter season from early December through late April. This division resulted in 3 sets of group locations each for the Bearmouth (n^sub winter^ = 47, n^sub lambing^ = 49, and n^sub autumn^ = 20), Garrison (n^sub winter^ = 69, n^sub lambing^ = 72, and n^sub autumn^ = 51), and Skalkaho (n^sub winter^ = 95, n^sub lambing^ = 76, and n^sub autmn^ = 44) study sites.
We selected explanatory variables for resource-selection modeling from the literature (Table 1), and compiled these data in a global information system (GIS) using ArcView 3.2a and the Spatial Analyst extension. We derived 30 x 30-m grids of elevation, slope, and aspect from the United States Geological Survey National Elevation Dataset (http://ned.usgs.gov/). We also derived a solar radiation index combining latitude, slope, and aspect that has proven a significant predictor of bighorn sheep habitat use (Dicus 2002; K. Keating, pers. comm.). We defined areas with slopes >27 as escape terrain (Andrew et al. 1999; Dunn 1996; McCarty 1993; Smith et al. 1991) and quantified distance (m) from escape terrain patches. We created a minimum patch size for escape terrain (0.7 ha) based on the smallest patch we observed sheep using for escape. We used TIGER 2000 Census data (http://www.census.gov/geo/www/tiger/) to measure distance to water and distance to roads.
TABLE 2.-Maximum distances moved from core female home ranges and estimated ages at the time of movement for individual males and all collared females at 3 study sites in western Montana, 2001-2002.
We used the SILC3 Land Cover Classification Data (Wildlife Spatial Analysis Lab, in litt.) to quantify vegetation types at 30 30-m resolution. We simplified cover types into 3 basic categories (xeric shrub and grassland, open canopy forest, and closed canopy forest). Two study sites experienced wildfires during summer 2000, after SILC3 data were generated. We used fire severity data (Wildlife Spatial Analysis Lab, in litt.) to edit SILC3 vegetation data and create 3 additional vegetation classes (burned grass, low- severity burned forest, and high-severity burned forest). We included these cover types in habitat-selection models as a set of 6 continuous variables by quantifying distance from locations to each cover type (Miller et al. 2000) and removing categories not biologically relevant to each study site.
Numerous bighorn habitat suitability models have highlighted horizontal visibility as a key component of bighorn habitat (Johnson and Swift 2000: Risenhoover and Bailey 1980, 1985; Smith et al. 1991; Zeigenfuss et al. 2000). We found no strong preference for areas with high horizontal visibility in separate analyses using the same study animals (DeCesare and Pletscher 2004) and did not include horizontal visibility as an explanatory variable in CIS-based modeling.
We modeled Johnson's (1980) 3rd-order selection (habitat selection within the home range) comparing habitat used by animals to that available. Female group locations provided a sample of habitats used. Definitions of available habitat are somewhat arbitrary (Aebischer et al. 1993: Wilson et al. 1998); we used a 100% minimum convex polygon with an additional 100-m buffer to approximate the area available to each herd during the study period (McCorquodale 1999). Seasonal home ranges overlapped greatly, so we used a single, year-round home range to define available habitat at each site. Three locations gathered during forays to outlying areas not considered part of the home range (Burt 1943) were removed before calculating availability minimum convex polygons. We drew a systematic sample of points (n^sub Bearmouth^ = 2,428, n^sub Garrison^ = 1,204, and n^sub Skalkaho^ = 1,953) from a 100 100-m grid within minimum convex polygons to estimate habitat availability (Erickson et al. 1998). To compensate for large and autocorrelated samples, we randomly selected 200 locations from the systematic samples to quantify availability for initial model comparisons. We then reran best models with the full samples to obtain more precise and unbiased coefficient estimates (Burnham and Anderson 1998; T. McDonald, West Inc., pers. comm.).
We used logistic regression to estimate 9 resource-selection functions (Manly et al. 2002). or 1 for each site in each season. Using logistic regression with use-availability data presents some problems; predicted values are not scaled between 0 and 1, and are not necessarily valid resource-selection probabilities (Keating and Cherry 2004; Manly et al. 2002). Instead, resource-selection functions provided an index for probability of use and a method to rank habitats selected by bighorn sheep (Keating and Cherry 2004).
We used the information-theoretic methods of model selection (Burnham and Anderson 1998) to avoid overfitting models to site- specific data. We developed IO a priori candidate models for each season based on literature review and field observation. We fit global models with all meaningful explanatory variables and used the Hosmer and Lemeshow goodness-of-fit test to evaluate general model fit before selecting among candidate models (Burnham and Anderson 1998; Hosmer and Lemeshow 2000). We also tested for correlation among independent variables to ensure that correlations among included variables did not affect results.
We used the same set of seasonal candidate models for each site (with some variation in applicable cover type categories), and calculated the Akaike information criterion, corrected for small sample sizes (AIC^sub c^), to discriminate fit and parsimony of candidate models (Burnham and Anderson 1998). We calculated differences in AIC^sub c^ values (Δ^sub i^) between the lowest- scoring model and each candidate model. When multiple models scored Δ^sub i^
For initial model evaluation, we generated predictions from the same data used to create the model. This resubstitution can lead to optimistic estimates of model accuracy (Chatfield 1995; Fielding and Bell 1997; Verbyla and Litvaitis 1989), but remains a common approach to model validation. Second, we applied models across independent study sites for a more robust and unbiased validation of each model. Testing models across sites also was a means to assess how similarly females of recently colonized populations selected habitat. To evaluate the prediction success of models, we calculated the degree of correlation between predicted probabilities in a GIS landscape and the measured use of that landscape. We used a Spearman rank correlation to test for correlation between the density of use across the landscape and the predicted probability of use from each model (Boyce et al. 2002). We repeated this multiple times but did not use a Bonferroni correction (Cabin and Mitchell 2000; Garcia 2004; Moran 2003); the value of these tests was in comparing the internal and external validation of models, not generating specific probability values.
Movements.-Bearmouth, Garrison, and Skalkaho females remained localized within home ranges, making maximum movements of 3.7, 7.4, and 9.8 km from core female home ranges (Table 2), respectively. None of these movements overlapped with home ranges of nearby herds of bighorn or domestic sheep.
We detected relatively long extra-home range movements for males in all herds and 4 of 5 collared animals (Table 2). These movements occurred primarily during summer months (May-September) but 1 male continued moving back and forth between core female home ranges and neighboring herd home ranges during October and December as well. The longest movement was 32.9 km from the core female home range, and 2 individuals crossed interstate I-90 and the Clark Fork River a total of 7 times. We found direct association with bighorn sheep in neighboring populations in 2 of 3 herds (3 of 5 animals) and found direct association with domestic sheep in the same 2 herds (2 of 5 animals).
TABLE 3.-Candidate models and selection criteria for information- theoretic model for bighorn sheep selection at 3 study sites and seasons in western Montana, 2001-2002. Best models, selected as those with the fewest variables when Δ^sub i^
Resource selection.-We found both similarities and differences in best models across seasons and study sites. The same candidate models (with differences in included cover type categories) were used for each site, but in only 1 of 9 cases were the same "best" models chosen between 2 sites (Table 3). Increasing slopes and decreasing distances to escape terrain were important explanatory variables in most cases (Table 4). Although distance to water was a significant explanatory variable in 2 of the 3 lambing models, the sign of the coefficient revealed that females during the lambing period were actually further from water than expected. The solar radiation index proved a valuable explanatory variable in several cases, while aspect was not in a single "best" model.
Models performed well when applied to the site where they were fit (P ≤ 0.002; Table 5), which indicated that "best" models were reliable estimates of habitat selection for given sites and seasons. Applying models to independent data from different study sites produced mixed results (P = 0.001-0.960; Table 5); we excluded Bearmouth autumn data from model comparisons because of low sample sizes. Although some models did quite well at predicting use at other sites, this was not consistent.
Movements.-We found that females exhibited strong fidelity to home ranges. This was c\onsistent with past findings (Geist 1971) and suggested that these populations represented colonizations of new habitat rather than range expansions from existing populations.
TABLE 4.-Variables and coefficients included in season and site- specific best models for bighorn sheep resource selection at 3 sites in western Montana, 2001-2002. Variables used in models are defined in Table 1.
Bleich et al. (1996) defined a metapopulation of desert bighorn sheep as a set of subpopulations occupying mountain ranges less than 15 km apart, based on genetic evidence that metapopulations defined as such had unique haplotypes from one another. We observed movements of 18.9-32.9 km from core female home ranges in 4 of 5 collared males. Although the majority of these movements were not during the breeding season, these results suggest that Rocky Mountain bighorn sheep may be able to maintain genetic connectivity among subpopulations at distances greater than observed by Bleich et al. (1996) for desert subspecies. Genetic techniques similar to those used by Bleich et al. (1996) could be used to derive an alternate distance, but considerable translocations of bighorns in and out of most of western Montana's populations might blur genetic distinctions that once existed across the region (McCarthy 1996).
Our small sample of marked males (n = 5) also revealed a high degree of connectivity between bighorn and domestic sheep populations. Thus, we cautiously suggest that distinct populations of Rocky Mountain bighorn sheep may be more connected than previously known. If so, connectivity can create an efficient path for disease spread. We encourage further research on metapopulation dynamics of this species in northern ecosystems, because it remains unclear whether connectivity among populations serves as a genetic boost or a disease-laden liability.
TABLE 5.-Spearman rank correlations (R values) and associated P values testing correlation between season- and site-specific resource-selection model predictions and relocation data for bighorn sheep at 3 sites, 2001-2002. Values in bold represent unbiased tests of models with independent data from other sites. We did not include the Bearmouth autumn model because of small sample size of locations.
Resource selection.-Females of these 3 recently colonized herds were different in their selection patterns as evidenced by our model selection and validation results. Although the presence of steep, escape terrain was consistent across sites, large differences among sites may be indicative of a lower degree of habitat specialization than commonly expected. When combined with increasing knowledge about dispersal and colonization abilities (Singer et al. 2000a; this study), these results suggest that the potential for natural recolonization of the West may be higher than previously expected.
We used resource-selection modeling to compare 3rd-order selection (Johnson 1980) within herd home ranges, and we encourage future research on 2nd-order selection, or selection of herd home ranges within larger landscapes. Singer et al. (2000a) modeled the probability of colonization for different habitat patches after sheep transplants and found that population growth rates in source herds and vegetation characteristics in corridors between source and colonized patches were predictive parameters. A resource-selection function analysis of the habitat characteristics of colonized and uncolonized patches themselves (instead of the corridor habitat) would be valuable for future habitat modeling and conservation.
All habitat-selection analyses fall under the broad assumption that selection equates with better habitat quality and population fitness (Garshellis 2000; Van Home 1983). Comparing successful and unsuccessful reintroductions addresses this problem and could be further explored (Johnson and Swift 2000). Incorporating fitness into measures of habitat quality should be a goal for all studies of wildlife and habitat, but funding and logistics make data collection problematic.
In agreement with past bighorn sheep habitat studies (Geist 1971; McCarty and Bailey 1994; Tilton and Willard 1982) and modeling (Dicus 2002; Hughes 1997; Johnson and Swift 2000; Sweanor et al. 1996), we found slope and distance to escape terrain were important habitat variables across seasons and sites. These parameters appear universally important to bighorn sheep habitat in our study areas. However, these were not the only parameters that explained variation in resource selection, and more site-specific variation occurred in other parameters.
Distance to cover type variables revealed differences among sites. The presence of recently burned grasslands and forest only in Bearmouth and Skalkaho created obvious site-specific differences. Best models often included these cover type variables and sheep were likely responding to effects of the fires. Areas in or near high- visibility habitats such as grasslands and high-severity burned forests were generally preferred and low-visibility denser forests avoided. However, we did not detect selection based on horizontal visibility alone (DeCesare and Pletscher 2004) and are hesitant to associate selection of cover types to horizontal visibility. We hypothesize that horizontal visibility may be more important at the 2nd-order scale of selection (Johnson 1980) for these herds, and that within the home range, other parameters are more important. We were unable to include all potential variables in our analysis. Other factors such as predator densities, food quality, and forage abundance also may be important.
Many recent efforts identifying and modeling suitable bighorn habitat (Hughes 1997; Johnson and Swift 2000; Zeigenfuss et al. 2000) are based on the habitat-evaluation procedure of Smith et al. (1991). The habitat-evaluation procedure of Smith et al. (1991) was a stepwise procedure that identified core habitat based on escape terrain patches and subsequently removed areas from this core habitat that were unsuitable for other reasons (e.g., dense vegetation, close proximity to human development, close proximity to domestic sheep, high elk or cattle concentrations, etc.). Zeigenfuss et al. (2000) identified the 5 key habitat criteria upon which this habitat-evaluation procedure was based: close proximity to large patches of escape terrain, close proximity to water, lack of highway or water barriers, high horizontal visibility, and minimum of 150 m from human development. The importance of slope and distance to escape terrain in our models reaffirms the consideration of these variables when identifying core bighorn sheep habitat. Surface sources of water have been shown to be seasonally important to desert bighorn sheep (Bleich et al. 1997; Leslie and Douglas 1979; Turner and Weaver 1980), but this relationship is less clear for the Rocky Mountain subspecies (Shackleton et al. 1999). We found no consistent relationships between bighorn use and proximity to water, although in 2 of 3 lambing models females were actually further from water than random expectation. This result may be an artifact of high water availability in the home range relative to lambing areas, or of avoidance of dense, riparian habitats commonly associated with water in our study areas. Comparison to the habitat-evaluation procedures of Zeigenfuss et al. (2000) and other such procedures is difficult because of differences in selection scale. Most bighorn sheep habitat-evaluation procedures are developed to identify suitable bighorn home ranges within larger landscapes (Johnson's  2nd-order selection), whereas our models reflect 3rd-order selection within home ranges. We found that horizontal visibility (DeCesare and Pletscher 2004) and proximity to water and roads were not consistent factors in habitat selection within home ranges, but we cannot assess how these variables affected selection of home ranges within the larger landscape.
Dicus (2002) used similar logistic regression procedures to assess 3rd-order winter habitat selection by bighorn sheep in northwestern Montana. Parameters in his best winter range model were slope, distance to escape terrain, solar radiation index, snow cover, and values of cover types. These parameters were similar to those we found important in our study areas and were measured at similar scales, but signs and magnitudes of coefficients were not consistent with our models. Model validation procedures such as those in this paper would provide a clearer comparison of resource selection across study sites.
Validating resource-selection functions with resubstituted data resulted in high prediction success. Had this study been limited to any 1 of these 3 sites, as is often the case, we might have been overly confident in our models' validity. The danger of applying such models to new landscapes is evident by our between-site comparisons. Although some models predicted bighorn sheep use in new areas well, site-specific variation can strongly affect their external validity. These modeling procedures elucidated differences in selection patterns across recently colonized habitats, and we encourage future use of such techniques both to compare selection across sites and to externally validate local models.
We thank J. Firebuugh, K. Foresman, L. Hicks. J. Hogg, and J. W. Thomas for their contributions in time and thought to this study. Funding came from partnership between The University of Montana, Montana Department of Fish, Wildlife and Parks, The Bitterroot National Forest, The Bureau of Land Management, The Boone and Crockett Wildlife Conservation Program, Plum Creek Timber Company, The Welder Wildlife Foundation, The Five Valleys Chapter and National Headquarters of Safari Club International, and The Montana Chapter of the Foundation for North American Wild Sheep. We also thank the many private landowners whose cooperation made this resea\rch possible. A. J. Kroll, D. G. Whittaker, and anonymous reviewers provided valuable comments on earlier drafts of this manuscript.
AEBISCHER, N. J., P. A. ROBERTSON, AND R. E. KENWARD. 1993. Compositional analysis of habitat use from animal radio-tracking data. Ecology 74:1313-1325.
ANDREW, N. G., V. C. BLEICH, AND P. V. AUGUST. 1999. Habitat selection by mountain sheep in the Sonoran Desert: implications for conservation in the United States and Mexico. California Wildlife Conservation Bulletin 12. California Department of Fish and Game, Sacramento, California.
ANIMAL CAKE AND USE COMMITTEE. 1998. Guidelines for the capture, handling, and care of mammals as approved by the American Society of Mammalogists. Journal of Mammalogy 79:1416-1431.
BERGER, J. 1990. Persistence of different-sized populations: an empirical assessment of rapid extinctions in bighorn sheep. Conservation Biology 4:91-98.
BLEICH, V. C., R. T. BOWYER, AND J. D. WEHAUSEN. 1997. Sexual segregation in mountain sheep: resources or predation? Wildlife Monograph 134:1-50.
BLEICH, V. C., J. D. WEHALLSEN, AND S. A. HOLL. 1990. Desertdwelling mountain sheep: conservation implications of a naturally fragmented distribution. Conservation Biology 4:383-390.
BLEICH, V. C., J. D. WEHAUSEN, R. R. RAMEY II, AND J. L. RECHEL. 1996. Metapopulation theory and mountain sheep: implications for conservation. Pp. 353-373 in Metapopulations and wildlife aconservation (D. R. McCullough, ed.). Island Press, Covelo, California.
BOYCE, M. S., P. R. VERNIER, S. E. NIELSEN, AND F. K. A. SCHMIEGELOW. 2002. Evaluating resource selection functions. Ecological Modelling 157:281-300.
BUNCH, T. D., W. M BOYCE, C. P. HIBLER, W. R. LANCE T. R. SPRAKER, AND E. S. WILLIAMS. 1999. Diseases of North American wild sheep. Pp. 209-237 in Mountain sheep of North America (R. Valdez and P. R. Krausman, eds.). University of Arizona Press, Tucson.
BURNHAM, K. P., AND D. R. ANDERSON. 1998. Model selection and inference: a practical information-theoretic approach. Springer- Verlag, New York.
BURT, W. H. 1943. Territoriality and home range concepts as applied to mammals. Journal of Mammalogy 24:346-352.
CABIN, R. J., AND R. J. MITCHELL. 2000. To Bonferroni or not to Bonferroni: when and how are the questions. Bulletin of the Ecological Society of America 81:246-248.
CHATFIELD, C. 1995. Model uncertainty, data mining, and statistical inference. Journal of the Royal Statistical Society, Series A (Statistics in Society) 158:419-466.
DASGUPTA, N., AND J. R. ALLDREDGE. 2000. A chi-square goodness- of-fit analysis of dependent resource selection data. Biometrics 56: 402-408.
DECESARE, N. J. 2002. Movements and resource selection of recolonizing bighorn sheep in western Montana. M.S. thesis, University of Montana, Missoula.
DECESARE, N. J., AND D. H. PLETSCHER. 2004. Bighorn sheep, horizontal visibility, and GIS. Proceedings of the Biennial Symposium of the Northern Wild Sheep and Goat Council 14:181-190.
DECESARE, N. J., AND D. H. PLETSCHER. 2005. A dynamic test of spatial independence among bighorn sheep. Intermountain Journal of Sciences 11:25-30.
DEFORGE, J. R., C. W. JENNER, A. J. PLECHNER, AND G. W. SUDMEIER. 1979. Decline of bighorn sheep (Ovis canadensis), the genetic implications. Desert Bighorn Council Transactions 23:63-66.
DICUS, G. H. 2002. An evaluation of GIS-based habitat models for bighorn sheep winter range in Glacier National Park, Montana, USA. M.S. thesis, University of Montana, Missoula.
DOUGLAS, C. L., AND D. M. LESLIE, JR. 1999. Management of bighorn sheep. Pp. 238-262 in Mountain sheep of North America (R. Valdez and P. R. Krausman, eds.). University of Arizona Press, Tucson.
DUNN, W. C. 1996. Evaluating bighorn habitat: a landscape approach. Technical Note 395. Bureau of Land Management, Washington, D.C.
ERICKSON, W. P., T. L. MCDONALD, AND R. SKINNER. 1998. Habitat selection using GIS data: a case study. Journal of Agricultural, Biological, and Environmental Statistics 3:296-310.
FESTA BIANCHET, M. 1986a. Seasonal dispersion of overlapping mountain sheep ewe groups. Journal of Wildlife Management 50:325- 330.
FESTA-BIANCHET, M. 1986b. Site fidelity and seasonal range use by bighorn rams. Canadian Journal of Zoology 64:2126-2132.
FIELDING, A. H., AND J. F. BELL. 1997. A review of methods for the assessment of prediction errors in conservation presence/ absence models. Environmental Conservation 24:38-49.
FITZSIMMONS, N. N., S. W. BUSKIRK, AND M. H. SMITH. 1995. Population history, genetic variability, and horn growth in bighorn sheep. Conservation Biology 9:314-323.
GARCIA, L. V. 2004. Escaping the Bonferroni iron claw in ecological studies. Oikos 105:657-663.
GARSHELLIS, D. L. 2000. Delusions in habitat evaluation: measuring use, selection, and importance. Pp. 111-164 in Research techniques in animal ecology: controversies and consequences (L. Boitani and T. K. Fuller, eds.). Columbia University Press, New York.
GEIST, V. 1970. On the home range fidelity of bighorn rams. Desert Bighorn Council Transactions 14:51-53.
GEIST, V. 1971. Mountain sheep: a study in behavior and evolution. University of Chicago Press. Chicago. Illinois.
GROSS, J. E., F. J. SINGER, AND M. E. MOSES. 2000. Effects of disease, dispersal, and area on bighorn sheep restoration. Restoration Ecology 8(4S):25-37.
HANSEN, C. G. 1980. Habitat evaluation. Pp. 320-335 in The desert bighorn (G. Monson and L. Sumner, eds.). University of Arizona Press, Tucson.
HOOGIE, P. N., AND B. EICHENLAUB. 1997. Animal movement extension to Arcview ver. 1.1. Alaska Science Center-Biological Science Office. United States Geological Survey. Anchorage.
HOSMER, D. W., AND S. LEMESHOW. 2000. Applied logistic regression. John Wiley and Sons Inc., New York.
HUGHES, L. G. 1997. A GIS-based evaluation of the Big Horn Mountains for reintroduction of Rocky Mountain bighorn sheep. M.S. thesis. University of Wyoming, Laramie.
JOHNSON, D. H. 1980. The comparison of usage and availability measurements for evaluating resource preference. Ecology 61: 65-71.
JOHNSON, T. L., AND D. M. SWIFT. 2000. A test of a habitat evaluation procedure for Rocky Mountain bighorn sheep. Restoration Ecology 8(4S):47-56.
KEATING, K. A., AND S. CHERRY. 2004. Use and interpretation of logistic regression in habitat-selection studies. Journal of Wildlife Management 68:774-789.
KRAUSMAN, P. R. 1997. The influence of landscape scale on the management of desert bighorn sheep. Pp. 349-367 in Wildlife and landscape ecology: effects of pattern and scale (J. A. Bissonette, ed.). Springer-Verlag, New York.
KRAUSMAN, P. R., W. M. BOYCE, M. C. WALLANCE, AND R. M. LEE. 1992. What is minimum viable population size? Desert Bighorn Council Transactions 36:68-74.
KRAUSMAN, P. R., J. J. HERVERT, AND L. L. ORDWAY. 1985. Capturing deer and mountain sheep with a net-gun. Wildlife Society Bulletin 13:71-73.
LANDE, R. 1995. Mutation and conservation. Conservation Biology 9:782-791.
LESLIE, D. M., JR., AND C. L. DOUGLAS. 1979. Desert bighorn sheep of the River Mountains. Nevada. Wildlife Monograph 66:1-66.
MACARTHUR, R. H., AND E. O. WILSON. 1967. The theory of island biogeography. Princeton University Press, Princeton, New Jersey.
MANLY, B. F. J., L. L. MCDONALD, D. L. THOMAS, T. L. MCDONALD, AND W. P. ERICKSON. 2002. Resource selection by animals. 2nd edition. Kluwer Academic Publishers, Dordrecht. The Netherlands.
MCCARTHY, J. J. 1996. History of transplanting mountain goats and mountain sheep-Montana. Proceedings of the Biennial Symposium of the Northern Wild Sheep and Goat Council 10: 176-181.
MCCARTY, C. W. 1993. Evaluation of a desert bighorn sheep habitat suitability model. M.S. thesis, Colorado State University, Fort Collins.
MCCARTY, C. W., AND J. A. BAILEY. 1994. Habitat requirements of desert bighorn sheep. Special Report 69. Colorado Division of Wildlife, Denver.
MCCORQUODALE, S. 1999. Landscape and patch scale habitat use by migratory black-tailed deer in the Klickitat Basin of Washington. Northwest Science 73:1-11.
MCNAY, R. S., J. A. MORGAN, AND F. L. BUNNELL. 1994. Characterizing independence of observations in movements of Columbian black-tailed deer. Journal of Wildlife Management 58: 422- 429.
MILLER, D. A., B. D. LEOPOLD, G. A. HURST, AND P. D. GERARD. 2000. Habitat selection models for eastern wild turkeys in central Mississippi. Journal of Wildlife Management 64:765-766.
MILLSPAUGH, J. J., J. R. SKALSKI, B. J. KERNOHAN, K. J. RAEDEKE, G. C. BRUNDIGE, AND A. B. COOPER. 1998. Some comments on spatial independence in studies of resource selection. Wildlife Society Bulletin 26:232-236.
MORAN, M. D. 2003. Arguments for rejecting the sequential Bonferroni in ecological studies. Oikos 100:403-405.
OTIS, D. L., AND G. C. WHITE. 1999. Autocorrelation of location estimates and the analysis of radiotracking data. Journal of Wildlife Management 63:1039-1044.
RISENHOOVER, K. L., AND J. A. BAILEY. 1980. Visibility: an important factor for an indigenous, low elevation bighorn herd in Colorado. Proceedings of the Biennial Symposium of the Northern Wild Sheep and Goat Council 2:18-28.
RISENHOOVER, K. L., AND J. A. BAILEY. 1985. Foraging ecology of mountain sheep: implications for habitat management. Journal of Wildlife Management 49:797-804.
SAUSMAN, K. 1984. Survival of captive born Ovis canadensis in North American zoos. Zoo Biology 3:111-121.
SCHWARTz, O. A., V. C. BLEICH, AND S. A. HOLL. 1986. Genetics and the conservation of mountain sheep: Ovis canadensis nelsoni. Biological Conservation 37:179-190.
SHACKLETON, D. M., C. C. SHANK, AND B. M. WIKEEM. 1999. Natural history of Rocky Mountain and California bighorn sheep. Pp. 78-138 in Mountain sheep of North America (R. Valdez and P. R. Krausman, eds.). University of Arizona Press, Tucson.
SHANNON, N. H., R. J. HUDSON, V. C. BRINK, AND W. D. KITTS. 1975. Determinants of spatial distribution of Rocky Mountain bighorn sheep. Journal of Wildlife Management 39:387-401.
SINGER,F. J., M. E. MOSES, S. BELLEW, AND W. SLOAN. 2000a. Correlates to colonizations by translocated populations of bighorn sheep. Restoration Ecology 8(4S):66-74.
SINGER, F. J., C. M. PAPOUCHIS, AND K. K. SYMONDS. 2000b. Translocations as a tool for restoring populations of bighorn sheep. Restoration Ecology 8(4S):6-13.
SMITH, T. S., J. T. FLINDERS, AND D. S. WINN. 1991. A habitat evaluation procedure for Rocky Mountain bighorn sheep in the intermountain West. Great Basin Naturalist 51:205-225.
SWEANOR, P. Y., M. GUDORF, AND F. J. SINGER. 1996. Application of a GIS-based bighorn sheep habitat model in Rocky Mountain region of national parks. Proceedings of the Biennial Symposium of the Northern Wild Sheep and Goat Council 10:118-125.
SWIHART, R. K., AND N. A. SLADE. 1985. Testing for independence of observations in animal movements. Ecology 66:1176-1184.
SWIHART, R. K., N. A. SLADE, AND B. J. BERGSTROM. 1988. Relating body size to the rate of home range use in mammals. Ecology 69:393- 399.
TILTON, M. E., AND E. E. WILLARD. 1982. Winter habitat selection by mountain sheep. Journal of Wildlife Management 46:359-366.
TURNER, J. C., C. L. DOUGLAS, C. R. HALLUM, P. R. KRAUSMAN, AND R. R. RAMEY. 2004. Determination of critical habitat for the endangered Nelson's bighorn sheep in southern California. Wildlife Society Bulletin 32:427-488.
TURNER, J. C., AND R. A. WEAVER. 1980. Water. Pp. 100-112 in The desert bighorn: its life history, ecology, and management (G. Monson and L. Sumner, eds.). University of Arizona Press, Tucson.
VALDEZ, R., AND P. R. KRAUSMAN. 1999. Description, distribution, and abundance of mountain sheep in North America. Pp. 3-22 in Mountain sheep of North America (R. Valdez and P. R. Krausman, eds.). University of Arizona Press, Tucson.
VAN HORNE, B. 1983. Density as a misleading indicator of habitat quality. Journal of Wildlife Management 47:893-901.
VERBYLA, D. L., AND J. A. LITVAITIS. 1989. Resampling methods for evaluating classification accuracy of wildlife habitat models. Environmental Management 13:783-787.
WILSON, S. F., D. M SHACKLETON, AND K. L. CAMPBELL. 1998. Making habitat-availability estimates spatially explicit. Wildlife Society Bulletin 26:626-631.
ZEIGENFUSS, L. C., F. J. SINGER, AND M. A. GUDORF. 2000. Test of a modified habitat suitability model for bighorn sheep. Restoration Ecology 8(4S):38-46.
Submitted 29 July 2005. Accepted 10 November 2005.
Associate Editor was Floyd W. Weckerly.
NICHOLAS J. DECESARE* AND DANIEL H. PLETSCHER
Wildlife Biology Program, Department of Ecosystem and Conservation Sciences, University of Montana, Missoula, MT 59812, USA
Present address of NJD: United States Department of Agriculture/ United States Forestry Service, Rocky Mountain Research Station, 800 East Beckwith, Missoula, MT 59801, USA
* Correspondent: [email protected]
2006 American Society of Mammalogists
Copyright Alliance Communications Group, A Division of Allen Press, Inc. Jun 2006
(c) 2006 Journal of Mammalogy. Provided by ProQuest Information and Learning. All rights Reserved.