Rising temperatures globally are pressuring species to track cooler environments, with low-elevation species generally shifting poleward and mountain-dwelling species moving upwards in elevation. Examining changes in both abundance and occupancy across bioclimatic gradients can inform forecasts of such range shifts. Although occupancy and abundance reflect similar aspects of species-environment relationships, they are governed by different underlying processes. Abundance is often more reflective of shorter-term conditions that affect vital rates, whereas site occupancy often reflects habitat suitability and is typically governed by prevailing climate. By directly comparing results of both response types, we offer a robust method for assessing complex species-climate relationships. In this study, we test how populations of the American pika (Ochotona princeps), a small montane lagomorph, respond to climate change by quantifying and comparing the drivers of site occupancy, abundance, and amount of upslope retraction across 760 sites, nested within 64 watersheds, in the Northern Rocky Mountains. We also test the importance of seasonal means, extremes, thresholds, and inter-annual variability in climate, as well as habitat characteristics. Using mixed effects modelling, paired with an Information-Theoretic approach, we tested several a priori model suites to assess the best predictors of our three response classes.
Results/Conclusions
We found that the top predictors differed greatly among site occupancy, abundance, and amount of upslope retraction. For site occupancy, the top model included summer acute heat stress, actual evapotranspiration, and habitat availability metrics. For abundance, the preceding winter’s mean temperature (i.e. chronic cold stress) was the best predictor, suggesting rapid responses of populations to recent winter conditions. Furthermore, we found that both chronic heat and chronic cold stress best predicted the total amount of vertical retraction across the surveyed watersheds. Collectively, these results emphasize how species often respond to climate change at divergent temporal and spatial scales, and that results from occupancy analyses should only be used with caution when interpolating to predicting species abundances on the landscape, and vice versa. Moreover, our novel method assessing the drivers of elevational retraction across many replicated units (i.e., watersheds) has widespread applications for taxa globally, as distributional shifts become more widespread.