Habitat availability and gene flow influence diverging local population trajectories under scenarios of climate change: a place‐based approach

Ecological niche theory holds that species distributions are shaped by a large and complex suite of interacting factors. Species distribution models (SDMs) are increasingly used to describe species' niches and predict the effects of future environmental change, including climate change. Currently, SDMs often fail to capture the complexity of species' niches, resulting in predictions that are generally limited to climate-occupancy interactions. Here, we explore the potential impact of climate change on the American pika using a replicated place-based approach that incorporates climate, gene flow, habitat configuration, and microhabitat complexity into SDMs. Using contemporary presence-absence data from occupancy surveys, genetic data to infer connectivity between habitat patches, and 21 environmental niche variables, we built separate SDMs for pika populations inhabiting eight US National Park Service units representing the habitat and climatic breadth of the species across the western United States. We then predicted occurrence probability under current (1981-2010) and three future time periods (out to 2100). Occurrence probabilities and the relative importance of predictor variables varied widely among study areas, revealing important local-scale differences in the realized niche of the American pika. This variation resulted in diverse and - in some cases - highly divergent future potential occupancy patterns for pikas, ranging from complete extirpation in some study areas to stable occupancy patterns in others. Habitat composition and connectivity, which are rarely incorporated in SDM projections, were influential in predicting pika occupancy in all study areas and frequently outranked climate variables. Our findings illustrate the importance of a place-based approach to species distribution modeling that includes fine-scale factors when assessing current and future climate impacts on species' distributions, especially when predictions are intended to manage and conserve species of concern within individual protected areas.