COS 63-9
Small mammal metacommunity responses to a century of climate change revealed by multispecies occupancy models

Wednesday, August 12, 2015: 10:50 AM
321, Baltimore Convention Center
Sean P. Maher, Department of Biology, Missouri State University, Springfield, MO
Steven R. Beissinger, Environmental Science, Policy & Management, University of California at Berkeley, Berkeley, CA

Recent work examined responses of bird and mammal species to 20th century climate change in the Sierra Nevada by comparing elevational profiles of occupancy between historic and modern surveys to demonstrate both upwards and downslope  shifts. We extend theses analyses to quantify effects of climate change on site-level persistence and colonization in small mammal communities using a Bayesian multispecies, multiseason occupancy model using data collected as part of the Grinnell Resurvey Project. We identified paired sites and modeled historical occupancy as function of elevation. In the modern era, turnover (persistence and colonization) was initially modeled as a function of elevation and then compared using AUC to models in which climate change variables were included. Turnover with elevation was tested for a linear relationship and more complex patterns. Finally, we evaluated species-specific trajectories under future scenarios of change to identify the most vulnerable taxa.


Models incorporating only elevation terms describing turnover were highly accurate (mean AUC > 0.85), and the quadratic model fit best. Mean modern occupancy was lower at middle elevations as a result of lower persistence at these sites. Sørenson distances showed a decreasing trend with elevation, indicating higher elevations experienced less turnover than lower elevations. Models including climate change terms were more accurate than models with only elevation-related terms (mean AUC > 0.90 in all cases). Change in mean annual temperature, temperature of the warmest quarter, annual precipitation, and precipitation of the driest quarter yielded highly accurate models of turnover. Combining changes in mean annual temperature and annual precipitation into a single model yielded the most accurate model overall. Continued warming is expected to decrease mean occupancy for most species, whereas changes in annual precipitation provided less clear patterns of future occupancy. Our results show that small mammal metapopulations are sensitive to changes in climate and the pattern of turnover suggests high elevation occupancy has been stable. By using a multi-species occupancy model, we delineated patterns more effectively than using a species-by-species approach.