AbstractThe influence of bottom‐up food resources and top‐down mortality risk underlies the demographic trajectory of wildlife populations. For species of conservation concern, understanding the factors driving population dynamics is crucial to effective management and, ultimately, conservation. In southeastern British Columbia, Canada, populations of the mostly omnivorous grizzly bear (Ursus arctos) are fragmented into a mosaic of small isolated or larger partially connected sub‐populations. They obtain most of their energy from vegetative resources that are also influenced by human activities. Roads and associated motorized human access shape availability of food resources but also displace bears and facilitate human‐caused mortality. Effective grizzly bear management requires an understanding of the relationship between habitat quality and mortality risk. We integrated analyses of bottom‐up and top‐down demographic parameters to understand and inform a comprehensive and efficient management paradigm across the region. Black huckleberry (Vaccinium membranaceum) is the key high‐energy food for grizzly bears in much of southeastern British Columbia. Little is known about where and why huckleberries grow into patches that are useful for grizzly bears (i.e., densely clustered fruiting shrubs that provide efficient access to high energy food) and how forage supply and mortality risk influence population vital rates. By following 43 grizzly bears tracked with global positioning system (GPS) collars (57 bear years) in a 14,236‐km2 focal area spanning the Selkirk and Purcell mountain ranges, we developed a model to identify huckleberry patches from grizzly bear use data. Over 2 years we visited 512 sites used by bears, identifying more than 300 huckleberry patches. We used boosted regression tree modeling associating geophysical, ecological, soil, climate, and topographical variables with huckleberry patches. We integrated this modeled food layer depicting an important pre‐hibernation resource, into broader bottom‐up and top‐down analyses. In addition to berries, we examined bottom‐up variables indexing vegetative productivity that were previously found to be predictive of bear use (e.g., alpine, canopy cover, greenness, riparian). We also examined top‐down variables including road presence, road density, distance‐to‐road, secure habitat (defined as 500 m away from a road open to vehicular access), highways, human development, and terrain ruggedness. We evaluated the relationship of these variables to female habitat selection, fitness, and population density, testing the predictability and interrelatedness of covariates relative to bottom‐up and top‐down influences. We estimated resource selection functions with 20,293 GPS telemetry locations collected over 10 years from 20 female grizzly bears. We modeled fitness using logistic regression of spatially explicit reproductive data derived from genetically identified family pedigrees consisting of a mother, father, and offspring. Data included 33 mothers and 72 offspring (1–8 offspring per female). We estimated density through spatial capture‐recapture analysis of 126 grizzly bears detected with hair‐sampled DNA 287 times between 1998 and 2005. In all 3 analyses (habitat selection, fitness, and density), huckleberry patches were the most influential bottom‐up factor and secure habitat was the most consistent top‐down variable (road density was similarly predictive). All of the best supported models contained bottom‐up and top‐down variables except the male density model, which only contained a top‐down variable (secure habitat). These results suggest that both bottom‐up and top‐down forces drive several population processes of grizzly bears in the region, especially for females. We found that 38% of huckleberry patches (235 km2) predicted by the top model were in non‐secure habitat and that these patches were associated with lower fitness and density relative to those in secure habitat. Grizzly bear density was 2.6 times higher in habitat with road densities <0.6 km/km2, supporting the use of this road density target for management. The models predict that applying motorized access controls to backcountry areas with huckleberry patches would increase grizzly bear abundance by 23% on average across the region and 125% in the lowest density portion of the study area (Yahk). Managing both bottom‐up and top‐down influences is necessary to best mitigate the expanding human footprint, which is affecting many carnivore species worldwide. We provide evidence that bottom‐up forces were more influential for female habitat selection, fitness, and density than top‐down effects. We also uncovered a critical pattern in the magnitude of top‐down and bottom‐up influences on behavioral (habitat selection) and demographic (population density and fitness) responses. We show that the relative influence of top‐down influences on habitat selection and fitness are relatively weak compared to bottom‐up influences, whereas top‐down pressures exert much stronger limiting forces on population density. Forming conservation decisions around behavioral responses alone may misdirect actions and have limited benefits to populations. This insight can facilitate more effective decision‐making for grizzly bear conservation. Our findings highlight the importance of considering both bottom‐up and top‐down influences, suggesting cautious interpretation of habitat selection models for any species. A comprehensive examination with population‐level metrics such as density, vital rates, and fitness may be needed for effective management.
Search from 250 M+ research papersSearch from 250 M+ research papers
Open Access
Sep 30, 2023
Gaurav Kumar 
