An Empirical Typology of Tree Species Assemblages across the U.S. for Assessing Climate Change Threats to Forest Communities

An understanding of how the composition of forest tree communities will change in the future is critical for assessing the effects of climate change on forests. Recent research suggests that simply knowing the responses of individual species to changing climate does not adequately capture likely community-level response. In part, this is because some tree species in a community play a larger role than others, and thus, any climate change effects on those species will have greater effects on the associated forest tree community. As the basis of an assessment of climate change impacts to forest communities, we used hierarchical cluster analysis of over 120,000 recent forest inventory plots to empirically define forest tree communities across the U.S. We then used indicator species analysis to define the important tree species in each community. The result was a U.S.-wide typology of 29 clusters representing empirical forest communities. Clusters contained between six and 30,000 inventory plots each, and were associated with between one and twelve indicator species. We show how these clusters can be used to assess forest community-level climate change impacts. For a subset of clusters in the eastern U.S., we overlaid projected habitat suitability maps from USDA Forest Service Climate Change Tree Atlas for the indicator species under two climate change scenarios for the end of the century. Climate change threat results ranged widely for both scenarios, showing an average loss of habitat across all indicator species in some communities, and an average gain for others. For example, the balsam fir-quaking aspen community in the northeastern U.S. was projected to see the most loss of habitat for indicator species, with a change of 47.0% and 73.4% in importance value under the low and high climate change scenarios, respectively. Conversely, the indicator species in the green ash-American elm community were projected to see gains of 32.7% and 117.4% in importance value in the two scenarios. This approach puts existing tree species range projections in the context of tree communities. The results of our cluster analysis could similarly be used to assess forest community threat from other global change drivers such as invasive species or urbanization.