Extreme weather reveals divergence in syntopic mammals mediated by moisture

Background/Question/Methods Although increased frequency of extreme-weather events is one of the most-secure predictions associated with contemporary climate change, effects of such events on distribution and abundance of climate-sensitive species remain poorly understood. Montane ecosystems may be especially sensitive to weather anomalies because of complex abiotic and biotic interactions that propagate from climate-driven reductions in snowpack. Snowpack not only provides insulation from extreme cold, but also influences forage availability through timing of melt-off and water availability. We related relative abundances of two species of alpine mammals, the American pika (Ochotona princeps) and hoary marmot (Marmota caligata), to measures of weather and snowpack dynamics over an 11-year period that included a year of record-low snowpack in Washington, USA. We sought to 1) quantify any change in mammal abundance associated with the snowpack anomaly, and 2) identify aspects of weather and snowpack that influenced inter-annual abundance of pikas and marmots.

Results/Conclusions Whereas relative abundance of pikas declined markedly at elevations below 1400 m, it increased slightly at higher elevations that retained snow during the low-snow winter. Pikas showed a 1-year lag response to the snowpack anomaly suggestive of reproductive failure at low elevations following the weather anomaly, rather than direct mortality from exposure to extreme cold, in the absence of snowpack. Relative abundances of marmots declined by 69% between 2007 and 2016; the most-severe losses occurred at warm, dry sites. Moisture was surprisingly important, evidenced by strong support for top-ranked models that had interactions of vapor pressure deficit with snowpack duration (pikas) or cold exposure (marmots). Notably, our novel application of vapor pressure deficit from gridded climate data for analyses of animal abundances shows strong potential for improving species distribution models. A post-hoc assessment of vegetative phenology and productivity indicated that mammals were affected more strongly by physiological stress rather than biotic interactions associated with weather and snowpack dynamics; however, responses were context-dependent and species specific. Our results indicate that weather extremes, warming temperatures, and increasing aridity can lead to rapid loss of species in montane ecosystems.