Landscape context and behavioral clustering contribute to flexible habitat selection strategies in a large mammal

Abstract

An understanding of individual variation in animal habitat selection is important for effective conservation and management as well as predicting species responses to a rapidly changing world. Functional responses to habitat availability can explain some of this variation, but not accounting for behavioral grouping may oversimplify inference and mask the diversity of habitat use strategies present in a population. We investigated within-home range habitat selection variation at the group level in a reintroduced population of elk (Cervus canadensis) in Kentucky, USA, during 2020–2022, analyzing satellite tracking data from 103 individuals to quantify variance in responses to seven landscape variables. We used group-level selection coefficients estimated with mixed-effects resource selection functions to model population-level functional responses and classify groups into within-season behavioral clusters. We then used cross validation to assess if these methods of generalizing group-level variation improved predictions of space use. We found that elk had highly variable responses to several covariates, and that some of this variation could be attributed to functional responses to either cover type availability or configuration. When we generalized behavioral tactics via clustering, we found significant increases in group-level predictive ability over the global model and, in some cases, the functional models. This suggests that clustered behaviors are also driving individual heterogeneity in this population. Our results highlight the importance of considering individual differences when studying wildlife-habitat relationships and underscore the need for a more complete understanding of the mechanisms behind this variation to inform habitat management and conservation efforts.