You are here:
PREDICTING CLIMATE-INDUCED RANGE SHIFTS FOR MAMMALS: HOW GOOD ARE THE MODELS?
Citation:
LAWLER, J. PREDICTING CLIMATE-INDUCED RANGE SHIFTS FOR MAMMALS: HOW GOOD ARE THE MODELS? Presented at Bringing Climate Into Natural Resource Management, Portland, OR, June 28 - 30, 2005.
Description:
In order to manage wildlife and conserve biodiversity, it is critical that we understand the potential impacts of climate change on species distributions. Several different approaches to predicting climate-induced geographic range shifts have been proposed to address this problem. We investigated the potential implications of using these different approaches for conclusions drawn about future range shifts and extinctions. Using the current ranges of 100 mammal species found in the western hemisphere, we compared six methods for modeling their predicted future ranges. All approaches modeled current ranges as a function of current climate and current land cover. Future ranges were predicted using predicted future climatic conditions from a global circulation model and predicted future land cover from a process-based equilibrium vegetation model. Predicted future distributions differed markedly across the alternative modeling approaches, resulting in estimates of extinction rates that ranged between 0 and 7%, depending on which modeling approach was used. Random Forest Predictors, a model-averaging approach, consistently outperformed the other techniques (correctly predicting > 99% of current absences and 86% of current presences). We conclude that the types of models used in a study can have dramatic effects on predicted range shifts and extinction rates; and that model-averaging approaches appear to have the greatest potential for accurately predicting range shifts in the face of climate change.