Search

There is limited information available on wolverine (Gulo gulo) population density and trends in the boreal forest of North America. We estimated wolverine density using spatial capture‐recapture methods across 2 boreal forest study areas in Red Lake, Ontario (26,568 km2) and Rainbow Lake, Alberta (19,084 km2), Canada. We also used radio‐telemetry data to estimate annual survival of adult and sub‐adult wolverines and evaluated population trends with a stage‐based matrix model. We used an array of run poles and live traps to detect wolverines. In Red Lake over 3 winter field seasons (2019–2022), we detected 56 individual wolverines (17 females, 32 males, and 7 unknown sex), and in Rainbow Lake over 2 field seasons (2014–2016), we detected 48 individuals (19 females, 18 males, and 11 of unknown sex). Average densities in Red Lake and Rainbow Lake were 3.64 and 6.74 wolverines/1,000 km2, respectively. Adults and sub‐adults occurred at equal abundance. Spring snow cover, roads, and industrial developments were not associated with spatial patterns of wolverine density. Most deaths occurred near roads; wolverines were killed in fur traps set along roads, by wolves using roads to travel, and by vehicles. The largest source of death was from incidental (n= 6 in Red Lake) or licensed fur trapping (n= 8 in Rainbow Lake) and we report 8 injuries from fur trapping sets. Red Lake survival estimates for adults (0.87) and sub‐adults (0.86) contributed to a stable population trend. Rainbow Lake survival estimates for adults (0.66) and sub‐adults (0.50) contributed to a declining population trend based on a relatively low sample of radio‐days. Red Lake and Rainbow Lake combined survival estimates for adults (0.77) and sub‐adults (0.73) also contributed to a declining population trend. Our survival and population modeling suggests that human‐caused mortality is a significant risk to these populations. Our results can be applied to wolverine status assessments and used as benchmarks for future monitoring. Wolverine population stability or growth might be achieved by reducing incidental trapping deaths or injury and hindering human access to wolverine habitats through decommissioning or limiting development of industrial roads or other anthropogenic linear features. We estimated wolverine density, survival, and population trend at boreal forest field sites in northwestern Alberta and northwestern Ontario, Canada. Wolverine density was 6.74 individuals/1,000 km2in Alberta and 3.64 individuals/1,000 km2in Ontario. Our wolverine survival estimates contributed to evidence of population decline in Alberta and stability in Ontario. Wolverine population growth could be achieved by limiting human access to wolverine habitats.

For many ecosystems, feral horses are increasingly becoming an important if not dominant component of ungulate biomass and hence influence on community dynamics. Yet we still know little of how horses contribute to key ecological interactions including predator‐prey and indirect competitive relationships at a community level. Notably, feral species like horses can exhibit life‐history traits that differ from that of native (mainly artiodactyl) herbivore competitors. Artificial selection for traits like increased, early, or extended reproduction that have yet to be reversed by natural selection, coupled with naturally selected differences in anatomy and behavior, in addition to unique management objectives for horses compared to other species, means that the dynamics of feral horse populations are not likely to align with what might be expected of other large herbivores. Unexpected population dynamics and inherent biological asymmetries between native ungulates and feral horses may therefore influence the former via direct competition for shared resources and through enemy‐mediated interactions like apparent competition. In several localities feral horses now co‐exist with multiple native prey species, some of which are in decline or are species at risk. Compounding risks to native species from direct or indirect competitive exclusion by horses is the unique nature and socio‐political context of feral horse management, which tends towards allowing horse populations to be limited largely by natural, density‐dependent factors. We summarize the inherent asymmetries between feral horse biology and that of other ungulate prey species with consequences for conservation, focusing on predator‐prey and emerging indirect interactions in multi‐prey systems, and highlight future directions to address key knowledge gaps in our understanding of how feral horses may now be contributing to the (re)structuring of food webs. Observations of patterns of rapid growth and decline, and associated skews in sex ratios of feral horse populations, indicate a heightened potential for indirect interactions among large ungulate prey species, where there is a prevalence of feral horses as preferred prey, particularly where native prey are declining. In places like western North America, we expect predator‐prey interactions involving feral horses to become an increasingly important factor in the conservation of wildlife. This applies not only to economically or culturally important game species but also at‐risk species, both predators (e.g., wolves [Canis lupus], grizzly bears [Ursus arctos]) and prey (e.g., woodland caribou [Rangifer tarandus caribou]), necessitating an ecological understanding of the role of horses in natural environments that goes beyond that of population control. © 2021 The Wildlife Society. Feral horses now constitute a significant component of prey biomass in many ecosystems, yet little is known about predator‐prey interactions involving horses despite their potential to negatively affect native species. Understanding such interactions and in particular indirect interactions like apparent competition will be critical for managing horses and their role in contemporary food webs.

Translocation is an important management tool that has been used for >50 years in Arizona, USA, to increase bighorn sheep (Ovis canadensis) population densities and to restore herds to suitable habitat throughout their historical range. Yet, translocation can also alter the underlying genetic diversity and spatial structure of managed wildlife species in beneficial or detrimental ways. To evaluate the long‐term effect of translocation actions on bighorn sheep, we characterized statewide genetic structure and diversity using microsatellite and mitochondrial DNA data in 16 indigenous and translocated (supplemented or reintroduced) Arizona populations sampled between 2005 and 2012. Populations that were recipients of translocated animals showed no reduction in genetic diversity with allelic richness and heterozygosity estimates equivalent to, and in some cases greater than, indigenous source populations. The indigenous population occupying the Silver Bell Mountains population displayed the lowest indices of genetic diversity but shared mitochondrial DNA haplotypes with the Mohawk Mountains, Sierra Pinta, and Cabeza Prieta populations, indicating past connectivity and potential opportunities for genetic management if warranted. Bayesian clustering on genetic similarity and genetic divergence estimates corroborated previous work differentiating Rocky Mountain bighorn sheep (O. c. canadensis) and 2 desert lineages corresponding with Nelson's (O. c. nelsoni) and Mexican desert bighorn sheep (O. c. mexicana). In northern Arizona, assignment tests confirmed the presence of 2 indigenous metapopulations of Nelson's desert bighorn sheep in the Black Mountains and Grand Canyon and indicated that gene flow from the Grand Canyon population has likely played a role in maintaining genetic diversity and mitigating founder effects among multiple translocated populations in the area. In southern Arizona, we detected genetic structure consistent with 2 metapopulations of Mexican desert bighorn sheep representing a departure from current management practices that consider this lineage to be a single genetic unit. Several lines of genetic evidence presented in this study suggest that the Bill Williams River area is the contemporary contact zone for the 2 desert lineages; however, the degree to which translocation has enhanced introgression is unknown. Despite relative isolation from other herds, the translocated Rocky Mountain bighorn sheep population in eastern Arizona had high levels of allelic richness and heterozygosity and a negative inbreeding coefficient, conceivably as a result of multiple translocation events from sources in Colorado and New Mexico, USA. Although translocation management has successfully contributed to the reestablishment of bighorn sheep populations in Arizona without diminishing genetic diversity, future translocation should proceed with caution to preserve the genetic integrity and potential local adaptation within the Nelson's and Mexican desert bighorn sheep metapopulations identified in this study. © 2019 The Wildlife Society. We evaluated microsatellite and mitochondrial DNA diversity in indigenous and translocated populations of bighorn sheep in Arizona to assess the genetic outcomes of over half a century of translocation management. Our study revealed that there was no reduction in genetic diversity of translocated populations compared to sources and that previously undetected hierarchical genetic structure was present within 2 desert bighorn sheep ancestral lineages.

Pronghorn (Antilocapra americana) in North Dakota have experienced habitat fragmentation due to agricultural practices, roads, and oil development. We analyzed patterns of female pronghorn habitat selection in 2006 and 2014, years with contrasting pronghorn density and oil production in western North Dakota. We quantified resource selection and fawn:female ratios relative to proximity to active wells, road density, land cover, development, normalized difference vegetation index (NDVI), and agricultural practices. We also assessed patterns of well placement relative to the same environmental variables. Pronghorn selected sagebrush and areas with low NDVI but avoided developed areas, roads, forests, and wetlands. Pronghorn selected areas close to oil and gas wells because wells were located in high‐value habitats (e.g., native sagebrush‐steppe ecosystems selected by pronghorn). For the majority of variables tested, selection was stronger when pronghorn density was low, consistent with current resource selection theory. Although females selected relatively open habitats, fawn:female ratios within areas selected by females were positively correlated with NDVI. Our results demonstrate that pronghorn avoid human development and roads but not oil and gas wells. Although wells are not actively avoided by pronghorn, their placement in high‐value habitat for this species leads to significant habitat fragmentation. In light of these results, we recommend efforts to conserve pronghorn habitat such as constructing wells away from sagebrush, using existing roads to service newly constructed wells, and re‐vegetating well pads with sagebrush plantings once they are no longer in use. © 2016 The Wildlife Society. Pronghorn selected sagebrush and avoided developed areas, roads, forests, and wetlands. Pronghorn also selected areas close to well pads, which were associated with sagebrush landscapes.

We respond to Festa‐Bianchet (2019) and caution against using interpretations from the unique Ram Mountain history to justify management of mountain sheep throughout their range. Because harvest management at Ram Mountain is atypical, it is not useful in informing the management of most mountain sheep herds.

Wolverines (Gulo gulo) in the contiguous United States have been considered for protection under the Endangered Species Act, most recently based on the value of deep snow for the duration of the wolverine's denning season. We examined evidence for an obligate relationship between wolverines and spring snow cover using camera traps and long‐term fur harvests in Alberta. The proportion of traplines that harvested ≥1 wolverine was highest in the northwest Boreal Forest (0.3), where mean wolverine harvest density increased by 75% from the 1990s to 2000s. There was no difference in percent spring snow cover on traplines with a female (n= 81) or no female (n= 416) wolverine harvest in the Boreal Forest. Further, all female wolverines (n= 8) positively identified from camera traps in the Boreal Forest, including 5 lactating females, were located within townships predicted to have no spring snow cover. Long‐term harvests and evidence of reproduction in areas with low amounts of spring snow cover in the Boreal Forest of northern Alberta suggest that wolverines may be more flexible in their distribution than previously assumed. © 2016 The Wildlife Society.

Local perceptions of grizzly bear (Ursus arctos) numbers in southwestern Alberta, Canada are incongruent with their threatened status. We used non‐invasive genetic sampling to estimate grizzly bear density and abundance in southwestern Alberta. We established 899 bear rub objects (e.g., tree, power pole, fence post) for hair sample collection across the study area by surveying trail networks, using geographic information system layers, and working with >70 landowners to identify priority sampling areas. The study area included 2 management zones: the Recovery Zone where the objective was to recover the grizzly bear population, and a Support Zone intended to maintain those bears not exclusively within the Recovery Zone. We visited rub objects every 3 weeks from late May through early November for 8 visits (7 sampling occasions) per field season. We also allowed for opportunistically collected hair samples (e.g., trapped bears, hair at agricultural bear‐conflict sites). We identified species, individual identity, and sex based on nuclear DNA extracted from hair follicles. From 2013 through 2014, we identified 164 individual grizzly bears. Using spatially explicit capture–recapture models (SECR), we estimated density in 2 ways. First, we estimated density for each sex and year separately (2013: M = 9.2/1,000 km2in the Recovery Zone and 8.1/1,000 km2in the Support Zone, F = 14.9/1,000 km2in the Recovery Zone and 13.6/1,000 km2in the Support Zone; 2014: M = 7.2/1,000 km2in the Recovery Zone and 5.7/1,000 km2in the Support Zone, F = 9.0/1,000 km2in the Recovery Zone and 8.5/1,000 km2in the Support Zone). Second, we did not allow density to vary across years and instead estimated a single density for the study area (M = 8.0/1,000 km2in the Recovery Zone and 7.1/1,000 km2in the Support Zone, F = 12.4/1,000 km2in the Recovery Zone and 10.0/1,000 km2in the Support Zone). Though yearly variation occurred, we derived from our density estimates an expected abundance of approximately 67.4 resident grizzly bears, indicating a 4% per year increase since a 2007 estimate of 51 bears. These SECR density estimates pertain only to bears with home ranges that were centered within the study area. Using traditional capture‐mark‐recapture (CMR) models with the same data yielded a higher estimate of bears because it included all bears that were using the study area (2013: F = 68.9, M = 102.6; 2014: F = 63.0, M = 108.6), including >50% of bears previously genotyped in Montana or British Columbia. In contrast with the SECR estimates, the CMR estimates represent the number of bears that southwestern Alberta residents could have encountered (i.e., the population of bears that had potential to have been involved in conflict). Shifts in grizzly bear distribution resulted in large changes in our SECR density estimates between years, whereas our estimate of the number of bears using the area remained constant. We recommend increased inter‐jurisdictional monitoring and management of this international grizzly bear population. © 2016 The Wildlife Society.

In southern California, wildfire and urbanization influence landscape condition and features. We explored the synergistic effects of these disturbances on puma (Puma concolor) movement and habitat preference in southern California, USA using a telemetry dataset of >40 individuals from 2001 to 2011. We examined post‐fire movement of tracked pumas to consider the relationship between habitat use and availability as animals used the landscape. We also explored puma response to burned areas to examine the influence of vegetation cover, time since fire, fire frequency, and the synergistic effects of urbanization on puma landscape use. We compare our results to those of others by using a complementary approach to evaluate how habitat relationships change with additional puma data and with consideration of the fire history of the landscape. Our analyses suggest that after fires, pumas use burned habitats opportunistically, responding to post‐fire changes in prey density. Additionally, time since fire and fire frequency were useful predictors of puma presence at the individual‐ and population‐level. Our puma habitat models showed a strong relationship between puma habitat use, vegetation, and human developments. Our analysis suggests that fire is an important landscape element to consider when evaluating puma habitat; pumas opportunistically used recently burned areas in the first few years after fire. We identified that pumas avoided grassland vegetation types and areas of sparse cover, suggesting that increasing fire frequency could affect puma habitat suitability and connectivity in the future. Fire frequencies in parts of southern California have already increased beyond the threshold to which the ecosystem is resilient, and if this trend continues across the region, suitable puma habitat may be lost through vegetation‐type conversion to homogenous non‐native annual grasslands, a habitat avoided by pumas during foraging and resting. We suggest monitoring responses to increasing fire frequency to assess how pumas and other carnivores will be affected by large‐scale changes that may pose a threat to landscape integrity and persistence of puma populations in southern California. © 2015 The Wildlife Society.

We analyzed a unique 51-year time series for a population of mule deer in the North Dakota badlands, USA to examine the effects of seasonal weather on autumn recruitment. Winter weather recorded prior to birth of fawns and weather conditions recorded during spring the previous year (lagged effect), but not during spring or summer after birth, were related to observed patterns in autumn recruitment. When deer density was low (approx. 1?deer/km2) during the 1960s, fawn/female ratios were high ranging from 1.1 to 1.4 when minimum temperatures during the prior winter averaged -16°?C and -8°?C, respectively. Likewise, during the 2000s, when deer density was high (approx. 3?deer/km2), fawn/female ratios ranged from 0.6 to 0.9 when minimum daily temperatures during the previous winter were -16°?C and -8°?C, respectively. Large-scale Pacific-based climatic indices (>2,000?km to the Pacific coast) were correlated with local weather and helped explain variability in autumn recruitment. Higher values of the multivariate el niño southern oscillation and Pacific decadal oscillation were correlated with warmer and drier winters in the North Dakota badlands, whereas the North Pacific Index was correlated with colder and snowy winters. The ability to predict recruitment from local weather or from broad-scale climate indices (oscillations) provides greater opportunities for conservation and management, such as adjusting harvest quotas prior to autumn harvest. © 2015 The Wildlife Society.

Fragmentation is a growing threat to wildlife worldwide and managers need solutions to reverse its impacts on species' populations. Populations of grizzly bears (Ursus arctos), often considered an umbrella and focal species for large mammal conservation, are fragmented by human settlement and major highways in the trans-border region of southern British Columbia, northern Montana, Idaho, and northeastern Washington. To improve prospects for bear movement among 5 small fragmented grizzly bear subpopulations, we asked 2 inter-related questions: Are there preferred linkage habitats for grizzly bears across settled valleys with major highways in the fragmented trans-border region, and if so, could we predict them using a combination of resource selection functions and human settlement patterns? We estimated a resource selection function (RSF) to identify high quality backcountry core habitat and to predict front-country linkage areas using global positioning system (GPS) telemetry locations representing an average of 12 relocations per day from 27 grizzly bears (13F, 14M). We used RSF models and data on human presence (building density) to inform cost surfaces for connectivity network analyses identifying linkage areas based on least-cost path, corridor, and circuit theory methods. We identified 60 trans-border (Canada–USA) linkage areas across all major highways and settlement zones in the Purcell, Selkirk, and Cabinet Mountains encompassing 24% of total highway length. We tested the correspondence of the core and linkage areas predicted from models with grizzly bear use based on bear GPS telemetry locations and movement data. Highway crossings were relatively rare; however, 88% of 122 crossings from 13 of our bears were within predicted linkage areas (mean?=?8.3 crossings/bear, SE?=?2.8, range 1–31, 3 bears with 1 crossing) indicating bears use linkage habitat that could be predicted with an RSF. Long-term persistence of small fragmented grizzly bear populations will require management of connectivity with larger populations. Linkage areas identified here could inform such efforts. © 2015 The Wildlife Society.

Cougar (Puma concolor) sightings have increased markedly throughout much of Midwestern North America and breeding populations have re-established in areas where there has not been a viable population of cougars for much of the past century. Using satellite telemetry data, we estimated resource selection functions (RSF) to examine shifts in seasonal habitat selection of a recently re-established population of cougars relative to human activity in Cypress Hills Interprovincial Park (CHIP) in southwest Saskatchewan and southeast Alberta, Canada. We modeled human activity predictors, used in cougar RSFs, from data collected at a network of motion-triggered trail cameras on the road and trail system within CHIP. Using the same network of remote cameras, we quantified cougar use of trails to examine the potential for human–cougar encounters. Cougar habitat selection models that included human activity outperformed models including only environmental variables. Cougars avoided areas near roads and trails during seasonal peaks in human activity but selected those areas during seasons of low human activity. Within each season, we found cougars avoided areas of increased motorized and non-motorized human activity, particularly in spring (motorized and non-motorized) and summer (non-motorized). Although resource selection models did not include covariates for prey, selection results were consistent with the expected distribution of prey on the landscape. Cougar use of trails was prevalent throughout CHIP and use was concentrated during dusk and nighttime periods. Thus, the potential for human–cougar interactions are present throughout the year, although they are likely highest during periods of low human-use and in the evenings. Further, individual variation in habitat selection among cougars makes human–cougar interactions difficult to predict. Our results highlight the adaptable nature of cougars to varying levels of human disturbance, which will facilitate their continued eastward range expansion. © 2014 The Wildlife Society.

ABSTRACTGreater sage-grouse (Centrocercus urophasianus) are dependent on sagebrush (Artemisiaspp.) for food and shelter during winter, yet few studies have assessed winter habitat selection, particularly at scales applicable to conservation planning. Small changes to availability of winter habitats have caused drastic reductions in some sage-grouse populations. We modeled winter habitat selection by sage-grouse in Alberta, Canada, by using a resource selection function. Our purpose was to 1) generate a robust winter habitat-selection model for Alberta sage-grouse; 2) spatially depict habitat suitability in a Geographic Information System to identify areas with a high probability of selection and thus, conservation importance; and 3) assess the relative influence of human development, including oil and gas wells, in landscape models of winter habitat selection. Terrain and vegetation characteristics, sagebrush cover, anthropogenic landscape features, and energy development were important in top Akaike's Information Criterion-selected models. During winter, sage-grouse selected dense sagebrush cover and homogenous less rugged areas, and avoided energy development and 2-track truck trails. Sage-grouse avoidance of energy development highlights the need for comprehensive management strategies that maintain suitable habitats across all seasons.

Abstract:Assessing the impact of large carnivores on ungulate prey has been challenging in part because even basic components of predation are difficult to measure. For cougars (Puma concolor), limited field data are available concerning fundamental aspects of predation, such as kill rate, or the influence of season, cougar demography, or prey vulnerability on predation, leading to uncertainty over how best to predict or interpret cougar-ungulate dynamics. Global Positioning System (GPS) telemetry used to locate predation events in the field is an efficient way to monitor large numbers of cougars over long periods in all seasons. We applied GPS telemetry techniques combined with occasional snow-tracking to locate 1,509 predation events for 53 marked and an unknown number of unmarked cougars and amassed 9,543 days of continuous predation monitoring for a subset of 42 GPS-collared cougars in west-central Alberta, Canada. Cougars killed ungulates at rates near the upper end of the previously recorded range, and demography substantially influenced annual kill rate in terms of both number of ungulates (subad F [SAF] = 24, subad M [SAM] = 31, ad M = 35, ad F = 42, ad F with kittens <6 months = 47, ad F with kittens <6 months = 67) and kg of prey (SAF = 1,441, SAM = 2,051, ad M = 4,708, ad F = 2,423, ad F with kittens <6 months = 2,794, ad F with kittens >6 months = 4,280). Demography also influenced prey composition; adult females subsisted primarily on deer (Odocoileusspp.), whereas adult males killed more large ungulates (e.g., moose [Alces alces]), and subadults incorporated the highest proportion of nonungulate prey. Predation patterns varied by season and cougars killed ungulates 1.5 times more frequently in summer when juveniles dominated the diet. Higher kill rate in summer appeared to be driven primarily by greater vulnerability of juvenile prey and secondarily by reduced handling time for smaller prey. Moreover, in accordance with predictions of the reproductive vulnerability hypothesis, female ungulates made up a higher proportion of cougar diet in spring just prior to and during the birthing period, whereas the proportion of males increased dramatically in autumn during the rut, supporting the notion that prey vulnerability influences cougar predation. Our results have implications for the impact cougars have on ungulate populations and have application for using cougar harvest to manage ungulates.

ABSTRACT Although habitat attributes of black-tailed prairie dog (Cynomys ludovicianus) colonies have been described for central and northern portions of the species' geographic range, little is known about these associations at the southern edge of this species' distribution. Because high-quality habitats are expected to be scarcer at the edge of the species' geographic range, different patterns of habitat selection might emerge in these populations. We analyzed habitat selection by black-tailed prairie dogs in a human-disturbed mosaic of desert grasslands and shrublands in northwestern Chihuahua, Mexico. We contrasted 151 used and 133 unused habitat units producing 11 case-control logistic regression models to explain site occupancy by prairie dogs with different combinations of environmental variables. Prairie dogs from Chihuahua occupy sites similar in most respects to sites in more northern regions, although these prairie dogs appear to be more tolerant of increased shrub density and reduced herbage cover. We found that site occupancy was best modeled by positive effects of soil moisture level, cover of forbs, cover of unpalatable vegetation, cover of bare ground, and amount of prairie-dog colony area within 1 km and by the inverse of altitude, shrub density, herbage height, and amount of hostile habitat within 1 km. The 2 most significant variables were herbage height and shrub density, which might reflect the prominent role that visibility plays in habitat selection by prairie dogs. In contrast, we found weak evidence that human features have significant impacts on site occupancy by prairie dogs. Our results support the prediction that environmental conditions of sites used by prairie dogs in edge regions partially differ from those observed in more northern latitudes. We suggest that reserve managers focus conservation efforts on areas with short vegetation, low density of shrubs, and high herbage cover, conditions that could be promoted by controlled burns, herbage mowing, and mechanical removal of shrubs.

Abstract:Gray wolf (Canis lupus) populations are recovering in many parts of the world and managers from various jurisdictions will be faced with difficult decisions as wolf populations continue to increase. Wolf management in Alberta, Canada, is achieved mostly through trapping on a registered trapline system. Wolf harvest increased over the last 22 years relative to population increases. Most wolf harvests occurred in the Rocky Mountains and surrounding foothills area and this pattern was consistent over time. On average, trappers only harvested an estimated 9.8% of the provincial population annually despite the lack of bag limits or quotas. Harvests are spatially autocorrelated with peak autocorrelation coinciding with average home-range size for wolves in Alberta. When wolf control actions are deemed necessary, trappers are unlikely to remove a sufficient number of wolves over a large enough area to limit subpopulations under the registered trapline system.

Abstract:In western Canada it is illegal to trap or snare cougars (Puma concolor), but cougars are sometimes caught accidentally in snares placed near carrion baits, a technique commonly used by trappers to harvest wolves (Canis lupus). We studied cougar foraging ecology and survival in west-central Alberta to estimate the propensity for cougars to scavenge, their susceptibility to snaring at trapper bait stations, and the implications these have for managing cougar populations. During 2005–2008, we used data from visits to 3,407 Global Positioning System (GPS) location clusters and >400 km of snow tracking of 44 cougars to locate foraging events and calculate scavenging rates. We identified 83 instances of scavenging, and 64% of monitored cougars scavenged at least once. Scavenging rates were higher in winter (0.12 events/week) than in summer (0.04 events/week), reflecting seasonal variation in carrion availability. Individual cougars scavenged at different rates, and winter feeding on carrion occupied up to 50% of total carcass handling time for some cougars. Based on these results we conclude that cougars are facultative scavengers. A propensity to scavenge made cougars susceptible to snaring causing high annual mortality in radiocollared cougars (0.11, 95% CI = 0.03–0.21). Provincial cougar mortality data demonstrate that snaring has increased dramatically as a mortality source in Alberta over the last 2 decades. Mortalities of radiocollared cougars during our study were 100% human caused and the addition of snaring mortality to already high hunting mortality resulted in low annual survival (0.67, 95% CI = 0.53–0.81). Our study is one of the first to identify population-level consequences for nontarget animals killed unintentionally by indiscriminate harvest techniques in a terrestrial ecosystem. Maintaining sustainable cougar harvest where snaring at carrion baits is permitted may require flexible hunting quotas capable of accommodating high cougar snaring mortalities in some years.

ABSTRACT We developed predictive habitat models for a bighorn sheep (Ovis Canadensis) population in the Peninsular Ranges of southern California, USA, using 2 Geographic Information System modeling techniques, Ecological Niche Factor Analysis (ENFA) and Genetic Algorithm for Rule-set Production (GARP). We used >16,000 Global Positioning System locations from 34 animals in 5 subpopulations to develop and test ENFA and GARP models, and we then compared these models to each other and to the expert-based model presented in the United States Fish and Wildlife Service's Recovery Plan for this population. Based on a suite of evaluation methods, we found both ENFA and GARP to provide useful predictions of habitat; however, models developed with GARP appeared to have higher predictive power. Habitat delineations resulting from GARP models were similar to the expert-based model, affirming that the expert-based model provided a useful delineation of bighorn sheep habitat in the Peninsular Ranges. In addition, all 3 models identified continuous bighorn sheep habitat from the northern to southern extent of our study area, indicating that the Recovery Plan's recommendation of maintaining habitat connectivity throughout the range is an appropriate goal.

ABSTRACT Using clusters of locations obtained from Global Positioning System (GPS) telemetry collars to identify predation events may allow more efficient estimation of behavioral predation parameters for the study and management of large carnivore predator-prey systems. Applications of field- and model-based GPS telemetry cluster techniques, however, have met with mixed success. To further evaluate and refine these techniques for cougars (Puma concolor), we used data from visits to 1,735 GPS telemetry clusters, 637 of which were locations where cougars killed prey >8 kg in a multi-prey system in west-central Alberta. We tested 1) whether clusters were reliably created at kill locations, 2) the ability of logistic regression models to identify kill occurrence (prey >8 kg) and multinomial regression models to identify the prey species at a kill cluster, and 3) the duration of monitoring required to accurately estimate kill rate and prey composition. We found that GPS collars programmed to attempt location fixes every 3 hours consistently identified locations where prey >8 kg were handled, and cluster creation was robust to GPS location acquisition failures (poor collar fix success). The logistic regression model was capable of estimating cougar kill rate with a mean 5-fold cross validation error of <10%, provided the appropriate probability cutoff distinguishing kill clusters from non-kill clusters was selected. Logistic models also can be used to direct visits to clusters, reducing field efforts by as much as 25%, while still locating >95% of all kills. The multinomial model overpredicted occurrence of primary prey (deer) in the diet and underpredicted consumption of alternate prey (e.g., elk and moose) by as much as 100%. We conclude that a purely model-based approach should be used cautiously and that field visitation is required to obtain reliable information on species, sex, age, or condition of prey. Ultimately, we recommend a combined approach that involves using models to direct field visitation when estimating behavioral predation parameters. Regardless of the monitoring approach, long continuous monitoring periods (i.e., >100 days of a 180-day period) were necessary to reduce bias and imprecision in kill rate and prey composition estimates.

ABSTRACT The sex-age-kill (SAK) model is widely used to estimate abundance of harvested large mammals, including white-tailed deer (Odocoileus virginianus). Despite a long history of use, few formal evaluations of SAK performance exist. We investigated how violations of the stable age distribution and stationary population assumption, changes to male or female harvest, stochastic effects (i.e., random fluctuations in recruitment and survival), and sampling efforts influenced SAK estimation. When the simulated population had a stable age distribution and ? > 1, the SAK model underestimated abundance. Conversely, when ? < 1, the SAK overestimated abundance. When changes to male harvest were introduced, SAK estimates were opposite the true population trend. In contrast, SAK estimates were robust to changes in female harvest rates. Stochastic effects caused SAK estimates to fluctuate about their equilibrium abundance, but the effect dampened as the size of the surveyed population increased. When we considered both stochastic effects and sampling error at a deer management unit scale the resultant abundance estimates were within ±121.9% of the true population level 95% of the time. These combined results demonstrate extreme sensitivity to model violations and scale of analysis. Without changes to model formulation, the SAK model will be biased when ? ? 1. Furthermore, any factor that alters the male harvest rate, such as changes to regulations or changes in hunter attitudes, will bias population estimates. Sex-age-kill estimates may be precise at large spatial scales, such as the state level, but less so at the individual management unit level. Alternative models, such as statistical age-at-harvest models, which require similar data types, might allow for more robust, broad-scale demographic assessments.

Abstract:Stratification is commonly used to improve sampling efficiency of aerial surveys of ungulate populations with strata typically based on a priori information, such as preflight animal observations or vegetation attributes as surrogates for animal densities. We evaluated the usefulness of stratifying survey units for elk (Cervus elaphus) in the Rocky Mountain foothills of Alberta, Canada, using a resource selection function (RSF). We compared precision and design efficiency (DEFF) of population estimates from stratification approaches based on an RSF model to the past approach using amount of forest cover. We used a sample of telemetry relocations taken over a 3-year period from 165 elk, rarified to times of the day and months of the year when aerial surveys are conducted, to develop the RSF. We then used the top RSF model, based on Akaike's Information Criterion, to derive the average RSF value for an 8-km2survey unit. Using survey data from the first year, we evaluated binning schemes to define RSF-oriented strata based on poststratification and showed that Jenks natural breaks in the RSF values provided the greatest improvement in DEFF and increased precision, compared to 2 other stratification schemes. We then used this approach with data from 2 additional surveys to find that stratification by RSF consistently improves relative precision and design efficiency of elk population estimates, whether we employ pre- or poststratification. Where a RSF is available it could be used as a surrogate for animal densities when conducting stratified sampling for population surveys.

Abstract:Because of increasing concerns about puma (Puma concolor) attacks on people and the desire to minimize dangerous puma-human encounters while conserving puma populations, we examined spatial and temporal relationships between pumas and people that used Cuyamaca Rancho State Park (CRSP), California, USA. From 2001 to 2003, we studied 10 adult pumas outfitted with Global Positioning System collars. Although number of visitors to CRSP was increasing, no dangerous puma-human encounters were reported during our study. Male and female pumas typically moved short distances during the day (mean of means of individual hourly movements = 168 m and 131 m each hr, respectively) and moved the most at night (mean of means = 690 m and 390 m each hr, respectively). Of 10 pumas, 9 were least active during the day and most active during the evening or at night. In contrast, most visitor use of trails (x¯ = 85%) occurred during the day. Based on puma and human activity patterns, risk of a puma-human encounter was greatest during the evening. Puma prey caches were randomly distributed in relation to trails and park facilities; however, 8 of 33 caches were still within 100 m of a trail and 2 were within 300 m of a facility. Individual puma behavior relative to human activity areas was variable. Some pumas appeared to temporally avoid human activity areas; others used the park randomly in relation to human activity areas; none appeared to be attracted to human activity areas. Pumas that did not show detectable responses to human activity may have been exhibiting some level of habituation; if so, this level of habituation did not result in puma-human conflicts. When human activity peaked during the day, adult male and female pumas were within 100 m of a trail an average of 9% and 19% of the time they were located in the park, respectively. Thus, there were opportunities for puma-human encounters. Management personnel can take a proactive approach to deal with puma-human interactions through education and protocols that help to minimize probability of conflicts; this may provide the best chance for a continued puma presence in habitat used by pumas and people.

Abstract:We responded to the claim by Greenwald et al. (2005) that the management recommendations for the northern goshawk in the Southwestern United States (MRNG; Reynolds et al. 1992), a food web-based conservation plan that incorporated both northern goshawk (Accipiter gentilis) and multiple prey habitats, may be inadequate to protect goshawks. Greenwald et al. (2005) based this claim on their review of 12 telemetry studies of goshawk habitat selection and 5 nontelemetry studies of the effects of vegetation structure at the home range scale on goshawk nest occupancy and reproduction that appeared after the 1992 publication of the MRNG. Greenwald et al. (2005) summarized their review as showing that 1) goshawks were habitat specialists limited to forests with mature and old-growth structures including large trees, high canopy cover, multiple canopy layering, and abundant woody debris; 2) habitats were not selected on the basis of prey abundance and, therefore, managing for prey habitats diluted goshawk habitats; and 3) selection for openings, edges, and habitat diversity was inconclusive. Our review found that when the studies' respective authors pooled their radiotagged goshawks there were weak to strong selections for old forest structures. However, the studies also documented extensive variation in use of vegetation types and structures by individual goshawks; some avoided openings, edges, young forests, and old forests, whereas others selected for these characteristics. Additionally, by virtue of their wide geographic distribution, the studies showed that the focal populations themselves occurred in a variety of forest types, some with large structural differences. We found no evidence in Greenwald's et al. (2005) review that the MRNG are inadequate to protect goshawks. Rather, the studies reviewed by Greenwald et al. (2005), as well as many studies they missed, supported the MRNG. The suggestion of inadequacy by Greenwald et al. (2005) appeared rooted in misunderstandings of goshawk habitats described in the MRNG, a discounting of the extent of variation in vegetation structural and seral stages used by goshawks, a limited understanding of the extent to which prey limits goshawks, a failure to recognize the dynamic nature of forests, and an incomplete review of the literature. We believe the MRNG are adequate because they maximize the sustainable amount of mature and old forests in goshawk home ranges and specify the kinds and intermixtures of prey habitats within home ranges. Implementation of MRNG should reduce the likelihood that the availability of vegetation structures suited to goshawk nesting and foraging, as well as abundance and availability of prey, will limit goshawk nest occupancy and reproduction.

Abstract:We used resource selection functions (RSF) to estimate the relative probability of use for grizzly bears (Ursus arctos) adjacent to the Parsnip River, British Columbia, Canada, 1998-2003. We collected data from 30 radiocollared bears on a rolling plateau where a large portion of the landscape had been modified by human activities, primarily forestry. We also monitored 24 radiocollared bears in mountain areas largely inaccessible to humans. Bears that lived on the plateau existed at less than one-quarter the density of bears in the mountains. Plateau bears ate more high-quality food items, such as meat and berries, leading us to conclude that food limitation was not responsible for the differences in densities. We hypothesized that plateau bears were limited by human-caused mortality associated with roads constructed for forestry activities. Independent estimates of bear population size from DNA-based mark-recapture techniques allowed us to link populations to habitats using RSF models to scale habitat use patterns to population density. To evaluate whether differences in land-cover type, roads, or mortality risk could account for the disparity in density we used the mountain RSF model to predict habitat use and number of bears on the plateau and vice versa. We predicted increases ranging from 34 bears to 96 bears on the plateau when switching model coefficients, excluding land-cover types; when exchanging land-cover coefficients, the model predicted that the plateau population would be 9 bears lower than was observed. Large reductions in the numbers of mountain bears were predicted by habitat-selection models of bears using the plateau landscape. Although RSF models estimated in mountain and plateau landscapes could not predict bear use and abundance in the other areas, contrasts in models between areas provided a useful tool for examining the effects of human activities on grizzly bears.

ABSTRACTTo maximize success, reintroduction programs generally select predator-free release areas having high habitat quality. Past studies provide little insight into recovery efforts where multiple, potentially novel, mortality hazards occur. The ability of translocated animals to cope with novel environments can be affected by both pre- and postrelease experiences with habitat and mortality risks. We experimentally released elk (Cervus elaphus) having different background experiences into an area where predators and hunters were prevalent and habitat quality varied. Using a competing risks approach, we predicted the postrelease survival of individuals and their fidelity to release areas as a function of animal source and postrelease encounters with forage resources and areas used by wolves (Canis lupus) or humans. Mortality patterns were consistent with prerelease exposure to mortality risks but not habitat differences among source areas. Wolf predation, poaching, and legal Native hunting were equivalent in magnitude and accounted for the majority of elk mortalities. Familiarity with either wolves or hunters prior to release yielded first-year survival rates 1.9-2.2 times greater than observed for animals naive to both risks. These 2 primary sources of mortality traded off temporally as well as spatially given the proximity of roads, which wolves avoided. The prevalence of forage resources in release areas increased fidelity to release sites but coincided with higher mortality risk during the critical first year, potentially setting an ecological trap for animals naïve to local risks. Translocated individuals largely mediated their respective vulnerabilities over time, showing second-year survival rates equivalent to resident elk. In addition to using source populations that are able to adjust to mortality risks in release areas, spatial and temporal variation in mortality risks might be exploited when planning releases to increase the success of translocations into risky landscapes. (JOURNAL OF WILDLIFE MANAGEMENT 71(2):541–554; 2007)

Applications of logistic regression in a used-unused design in wildlife habitat studies often suffer from asymmetry of errors: used resource units (landscape locations) are known with certainty, whereas unused resource units might be observed to be used with greater sampling intensity. More appropriate might be to use logistic regression to estimate a resource selection function (RSF) tied to a use-availability design based on independent samples drawn from used and available resource units. We review the theoretical motivation for RSFs and show that sample “contamination” and the exponential form commonly assumed for the RSF are not concerns, contrary to recent statements by Keating and Cherry (2004; Use and interpretation of logistic regression in habitat-selection studies. Journal of Wildlife Management 68:774–789). To do this, we re-derive the use-availability likelihood and show that it can be maximized by logistic regression software. We then consider 2 case studies that illustrate our findings. For our first case study, we fit both RSFs and resource selection probability functions (RSPF) to point count data for 4 bird species with varying levels of occurrence among sample blocks. Drawing on our new derivation of the likelihood, we sample available resource units with replacement and assume overlapping distributions of used and available resource units. Irrespective of overlap, we observed approximate proportionality between predictions of a RSF and RSPF. For our second case study, we evaluate the classic use-availability design suggested by Manly et al. (2002), where availability is sampled without replacement, and we systematically introduce contamination to a sample of available units applied to RSFs for woodland caribou (Rangifer tarandus caribou). Although contamination appeared to reduce the magnitude of one RSF beta coefficient, change in magnitude exceeded sampling variation only when >20% of the available units were confirmed caribou use locations (i.e., contaminated). These empirically based simulations suggest that previously recommended sampling designs are robust to contamination. We conclude with a new validation method for evaluating predictive performance of a RSF and for assessing if the model deviates from being proportional to the probability of use of a resource unit.

Abstract: Prey species are thought to select habitats to obtain necessary resources while also avoiding predation. We examined whether habitat selection by elk (Cervus elaphus) changed following the reintroduction of wolves (Canis lupus) into Yellowstone National Park in 1995. Using conditional fixed-effects logistic regression to build habitat-selection models, we compared seasonal habitat selection by elk based on weekly elk radiolocations taken in 1985–1990 (without wolves) and 2000–2002 (with wolves). Fire-related habitat changes and climate likely interacted with wolf avoidance in shaping habitat selection by elk. In summer, when wolf activity was centered around dens and rendezvous sites, elk apparently avoided wolves by selecting higher elevations, less open habitat, more burned forest, and, in areas of high wolf density, steeper slopes than they had before wolf reintroduction. In winter, elk did not spatially separate themselves from wolves. Compared to the pre-wolf period, elk selected more open habitats in winter after wolf reintroduction, but did not change their selection of snow water equivalents (SWE) or slope. Elk appear to select habitats that allow them to avoid wolves during summer, but they may rely on other behavioral antipredator strategies, such as grouping, in winter. This study provides evidence that wolves can alter seasonal elk distribution and habitat selection, and demonstrates how the return of wolves to Yellowstone restores important ecosystem processes.

Abstract: Identifying how habitat use is influenced by environmental heterogeneity at different scales is central to understanding ungulate population dynamics on complex landscapes. We used resource selection functions (RSF) to study summer habitat use in a reintroduced and expanding elk (Cervus elaphus nelsoni) population in the Chequamegon National Forest, Wisconsin, USA. Factors were examined that influenced where elk established home ranges and that influenced habitat use within established home ranges. We also determined grain sizes over which elk responded to environmental heterogeneity and the number of categories of habitat selection from low to high that the elk distinguished. At a large spatial extent, elk home-range establishment was largely explained by the spatial distribution of wolf (Canis lupus) territories. Forage abundance was also influential but was relatively more important at a small spatial extent when elk moved within established home ranges. Areas near roads were avoided when establishing a home-range, but areas near roads were selected for use within the established home range. Elk distinguished among 4 different categories of habitat selection when establishing and moving within home ranges. Spatial and temporal cross validation demonstrated that to improve the predictive strength of habitat models in areas of low inter-annual variability in the environment, it is better to follow more individuals across diverse environmental conditions than to follow the same individuals over a longer time period. Last, our results show that the effects of environmental variables on habitat use were scale-dependent and reemphasize the necessity of analyzing habitat use at multiple scales that are fit to address specific research questions.

Abstract: Wildlife ecologists often use the Kaplan-Meier procedure or Cox proportional hazards model to estimate survival rates, distributions, and magnitude of risk factors. The Andersen-Gill formulation (A-G) of the Cox proportional hazards model has seen limited application to mark-resight data but has a number of advantages, including the ability to accommodate left-censored data, time-varying covariates, multiple events, and discontinuous intervals of risks. We introduce the A-G model including structure of data, interpretation of results, and assessment of assumptions. We then apply the model to 22 years of radiotelemetry data for grizzly bears (Ursus arctos) of the Greater Yellowstone Grizzly Bear Recovery Zone in Montana, Idaho, and Wyoming, USA. We used Akaike's Information Criterion (AICc) and multi-model inference to assess a number of potentially useful predictive models relative to explanatory covariates for demography, human disturbance, and habitat. Using the most parsimonious models, we generated risk ratios, hypothetical survival curves, and a map of the spatial distribution of high-risk areas across the recovery zone. Our results were in agreement with past studies of mortality factors for Yellowstone grizzly bears. Holding other covariates constant, mortality was highest for bears that were subjected to repeated management actions and inhabited areas with high road densities outside Yellowstone National Park. Hazard models developed with covariates descriptive of foraging habitats were not the most parsimonious, but they suggested that high-elevation areas offered lower risks of mortality when compared to agricultural areas.

Abstract: Of the approximately 2,500 California bighorn sheep (Ovis canadensis californiana) in Oregon, USA, the majority descend from a single transplant of 20 animals from British Columbia, Canada, in 1954. Recently, several populations have experienced poor recruitment, raising concerns that populations may be experiencing inbreeding depression resulting from a genetic bottleneck. We sampled 117 animals from 5 populations in Oregon and 1 population in Nevada to determine genetic variability within and among populations. We found that Oregon populations had fewer mean alleles per locus (2.2–2.4), lower heterozygosity (0.28–0.36), and higher inbreeding potential than animals from Nevada (3.8 alleles/locus, H = 0.53). These results now provide the baseline for rigorous ongoing evaluation of changes to allelic variability, inbreeding potential, variation among populations, and their effects on population demographics for Oregon's California bighorn sheep program. We suggest that evaluation of genetic variability in other source and recipient populations should be used to further understand how and when genetic management can be used for bighorn sheep conservation and management.