Spatial variation in bioclimatic relationships for a snow‐adapted species along a discontinuous southern range boundary
Sean M. Sultaire, John Humphreys, Benjamin Zuckerberg, Jonathan N. Pauli, Gary J. Roloff
Abstract
Abstract Aim Variation in the relationship between variables across space, known as spatial non‐stationarity, is a common phenomenon in ecology. Species distribution models (SDMs), however, typically assume stationarity in species–environment relationships across space. Spatial non‐stationarity may be particularly apparent in populations along range boundaries where species are actively adapting to changes in climate, and geographically isolated areas where isolation can facilitate adaptation. We used data collected from the southern range boundary of a winter‐adapted mammal to evaluate spatially varying environmental relationships in a region with a geographically isolated population. Taxon Lepus americanus . Location Laurentian Great Lakes region, USA. Methods We surveyed 320 sites for snowshoe hares near their regional southern range boundary. We related snowshoe hare occurrence to short‐term climate variables (snow cover duration, maximum temperature and their interaction) with a spatially explicit model that allowed the effect of climate variables on snowshoe hare occurrence to vary spatially. Results We found highest support for an SDM that incorporated geographic barriers when estimating spatial effects, in addition to spatially varying effects of maximum temperate, snow cover duration and their interaction. Hares were more likely to occupy sites with longer snow cover duration but the influence of temperature was more variable. Increasing temperature in northern portions of our study area had a positive influence on hare occurrence, but exhibited a negative relationship along the southern range boundary. Main Conclusions We provide a robust evaluation of stationarity assumptions in bioclimatic relationships and highlight spatially variable relationships between a cold‐adapted species and short‐term climate variables across a range boundary. However, spatial non‐stationarity was not most apparent across dispersal barriers but coincided with differences in the relationship between climate variables across space. Our study provides evidence that the distribution of a winter‐adapted species is strongly constrained by snow cover, but prevailing temperature modulates the strength of this relationship across space.