Modeling Contemporary Climate Profiles and Predicting their Response to Global Warming for Biotic Communities in Western United States
Authored By: G. E. Rehfeldt, N. L. Crookston, M. V. Warwell, J. S. Evans
Gerald E. Rehfeldt, Nicolas L. Crookston, Marcus V.Warwell, and Jeffrey S. Evans
USDA Forest Service Rocky Mountain Research Station
The Random Forests multiple regression tree was used to develop bioclimatic models of 25 biotic communities of western USA. Independent variables included 35 simple expressions of temperature and precipitation and their interactions. Analyses of the communities were based on a gridded sample of about 140,000 points, while those of species distributions used presence-absence data from about 120,000 permanent sample plots. Errors of classification for the distribution of plant communities averaged about 2 % when adjusted for ecotones between communities and misalignment between geographic data sets. Climate variables of most importance for segregating the communities were those that generally differentiate maritime from continental climates. Unmitigated global warming should increase the abundance of the climate profiles typifying the montane forests and grasslands largely at the expense of those of the alpine-tundra arid woodlands. However, nearly 47 % of the future landscape may be outside the climate profiles of the contemporary communities.
corresponding author:
Marcus V. Warwell
USDA Forest Service
Rocky Mountain Research Station
1221 S. Main
Moscow, ID 83843
208-883-2322
mwarwell@fs.fed.us
