We modeled and mapped, using the predictive data mining tool Random Forests, 134 tree species from the eastern United States for potential response to several scenarios of climate change. Each species was modeled individually to show current and potential future habitats according to two emission scenarios (high emissions on current trajectory and reasonable conservation of energy implemented) and three climate models: the Parallel Climate Model, the Hadley CM3 model, and the Geophysical Fluid Dynamics Laboratory model. Since we model potential suitable habitats of species, our results should not be interpreted as actual changes in ranges of the species. We also evaluated both emission scenarios under an "average" future climate from all three models. Climate change could have large impacts on suitable habitat for tree species in the eastern United States, especially under a high emissions trajectory. Of the 134 species, approximately 66 species would gain and 54 species would lose at least 10% of their suitable habitat under climate change. A lower emission pathway would result in lower numbers of both losers and gainers. When the mean centers, i.e. center of gravity, of current and potential future habitat are evaluated, most of the species habitat moves generally northeast, up to 800 km in the hottest scenario and highest emissions trajectory. The models suggest a retreat of the spruce-fir zone and an advance of the southern oaks and pines. In any case, our results show that species will have a lot less pressure to move their suitable habitats if we follow the path of lower emissions of greenhouse gases. The information contained in this; paper, and much more, is detailed on our website: http://www.nrs.fs.fed.us/atlas. Published by Elsevier B.V.
