USDA Climate Change Tree and Bird Atlas

The USDA Northern Research Station recently released a new website predicting the effects of climate change on 147 bird species.  It is based on their previous similar atlas for tree species.  Methods to be published in an upcoming issue of Ecography.

Website citation:
Matthews, S.N., L. R. Iverson, A.M. Prasad, A. M., and M.P. Peters. 2007-ongoing. A Climate Change Atlas for 147 Bird Species of the Eastern United States [database]., Northern Research Station, USDA Forest Service, Delaware, Ohio.

Ecography Citation:
Matthews, S. N., Iverson, L. R., Prasad, A. M. and Peters, M. P. 2011. Changes in potential habitat of 147 North American breeding bird species in response to redistribution of trees and climate following predicted climate change. Ecography, 34: no. doi: 10.1111/j.1600-0587.2010.06803.x

Mounting evidence shows that organisms have already begun to respond to global climate change. Advances in our knowledge of how climate shapes species distributional patterns has helped us better understand the response of birds to climate change. However, the distribution of birds across the landscape is also driven by biotic and abiotic components, including habitat characteristics. We therefore developed statistical models of 147 bird species distributions in the eastern United States, using climate, elevation, and the distributions of 39 tree species to predict contemporary bird distributions. We used randomForest, a robust regression-based decision tree ensemble method to predict contemporary bird distributions. These models were then projected onto three models of climate change under high and low emission scenarios for both climate and the projected change in suitable habitat for the 39 tree species. The resulting bird species models indicated that breeding habitat will decrease by at least 10% for 61–79 species (depending on model and emissions scenario) and increase by at least 10% for 38–52 species in the eastern United States. Alternatively, running the species models using only climate/elevation (omitting tree species), we found that the predictive power of these models was significantly reduced (p