Temporal vs. spatial variation in stress-associated metabolites within a population of climate-sensitive small mammals.

Temporal variation in stress might signify changes in an animal's internal or external environment, while spatial variation in stress might signify variation in the quality of the habitats that individual animals experience. Habitat-induced variations in stress might be easiest to detect in highly territorial animals, and especially in species that do not take advantage of common strategies for modulating habitat-induced stress, such as migration (escape in space) or hibernation (escape in time). Spatial and temporal variation in response to potential stressors has received little study in wild animals, especially at scales appropriate for relating stress to specific habitat characteristics. Here, we use the American pika (), a territorial small mammal, to investigate stress response within and among territories. For individually territorial animals such as pikas, differences in habitat quality should lead to differences in stress exhibited by territory owners. We indexed stress using stress-associated hormone metabolites in feces collected non-invasively from pika territories every 2 weeks from June to September 2018. We hypothesized that differences in territory quality would lead to spatial differences in mean stress and that seasonal variation in physiology or the physical environment would lead to synchronous variation across territories through time. We used linear mixed-effects models to explore spatiotemporal variation in stress using fixed effects of day-of-year and broad habitat characteristics (elevation, aspect, site), as well as local variation in habitat characteristics hypothesized to affect territory quality for this saxicolous species (talus depth, clast size, available forage types). We found that temporal variation within territories was greater than spatial variation among territories, suggesting that shared seasonal stressors are more influential than differences in individual habitat quality. This approach could be used in other wildlife studies to refine our understanding of habitat quality and its effect on individual stress levels as a driver of population decline.