Predicting and mitigating the impacts of global change on species' distributions

Abstract

Global change is expected to accelerate extinction rates substantially. Accurately predicting species responses to future climate and land use changes and the conservation effectiveness of protected areas are critical. Here, I test whether species distribution models can predict how species' ranges shift through time and if protected areas are more robust to recent global change impacts than areas lacking formal protection. Purely spatial species distribution models are able to predict how species' distributions have changed over the 20th century for many species. However, because this predictive ability was not strongly related to biological or sampling characteristics considered here, there is no a priori way to determine which species' models will accurately predict range shifts through time. Protected areas rarely performed differently than randomly selected, unprotected areas in terms of species richness change and species composition change over the past century. Conservation strategies should focus on improving landscape connectivity to facilitate species' geographical responses to future global changes and should account for uncertainty in predictions of those responses.