The influence of spatial and temporal scale on the relative importance of biotic vs. abiotic factors for species distributions

Aim: The scales of space and time over which biotic interactions influence distribution patterns remain an area of debate. Biotic interactions may be particularly influential in the ecology of mammalian carnivores, which engage in strong predator–prey and competitive interactions. Regional, multi-scale data on distribution patterns of interacting carnivore species are key to informing our understanding of this issue. Location: Washington State, USA. Methods: Using a spatially extensive camera-trapping array, we examined the factors influencing distribution patterns of seven carnivore species at multiple spatial and temporal scales. We used single-species occupancy models to assess the relative influence of abiotic and biotic covariates on distribution of individual carnivore species, and two-species occupancy models to assess how dominant carnivores influence subordinate carnivore occupancy and detection. Results: Carnivore occupancy patterns responded more strongly to abiotic than biotic covariates at both spatial grains and extents of analysis. The influence of biotic variables decreased as the grain of analysis increased, while constraining our study area extent substantially decreased the explanatory power of abiotic variables. Interspecific interactions among carnivores influenced occupancy and detection across spatial scales. However, there was little evidence that interactions were more pronounced at finer temporal scales. Main Conclusions: Results demonstrate that at broad spatial extents, species distributions are largely dictated by abiotic factors, particularly climate. Although biotic factors related to habitat and prey were important factors for some species, they generally declined in importance as grain size of analysis increased, suggesting these interactions play out at finer resolutions. However, competitive and mutualistic interactions exerted an independent influence on distribution over broad extents and coarse grains of analysis, suggesting that failure to account for interactions may limit our ability to accurately model distributions of species and their responses to future large-scale disturbances.