Inference of ecological processes structuring tree communities in eastern North America using functional and phylogenetic diversity

Background/Question/Methods

Recent research suggests that examination of functional and phylogenetic diversity will yield insight into processes shaping ecological communities. However, the type of data used to assess these patterns may influence the resulting conclusions. We investigated geographic variation in functional and phylogenetic diversity in two contrasting data sets of eastern North American tree communities; the Carolina Vegetation Survey (CVS), which focuses on intensively sampling native plant communities in the southeastern United States, and the Forest Inventory and Analysis national program (FIA), whose goal is to assess the current state of U.S. forests through broad scale stratified random sampling. Using seed mass, wood density, leaf nitrogen content, and maximum height to characterize functional diversity, we evaluated whether the differing goals and methodologies of the two surveys resulted in similar answers to three questions: Do gradients in climate predict patterns of phylogenetic and functional diversity across broad regional extents? Are communities with high functional diversity also characterized by high phylogenetic diversity? Do particular traits exhibit convergence indicative of the environment imposing filters on particular aspects of tree physiology and life-history? 

 Results/Conclusions

Temperature and precipitation were generally weak predictors of phylogenetic and functional diversity across large spatial extents. In the CVS data (n=2177), phylogenetic diversity of tree communities tended to increase with temperature (R²=0.16, P<0.001), but functional diversity had little relationship with temperature or precipitation (R²=0.001 and 0.02, respectively). Diversity of single traits did respond to climate variables, though responses differed between traits. For example, variation in wood density increased with precipitation (R²=0.08, P<0.001), but was not correlated with temperature (R²=0.001), while variation in maximum height increased with temperature (R²=0.11, P<0.001), but was uncorrelated with precipitation (R²=0.02). In contrast to the CVS data, tree communities in the FIA data (n=4926) displayed no relationship between phylogenetic or functional diversity and any temperature or precipitation variable, even when analyses were restricted to the same spatial extent as the CVS data.  In both data sets, high phylogenetic diversity did not correspond to high functional diversity; 91% (FIA) and 68% (CVS) of phylogenetically overdispersed communities were functionally clustered. The CVS data had more phylogenetically clustered tree communities than the FIA data, even though functional diversity was generally higher in CVS communities. These initial analyses suggest that sampling goals and methodologies impact observed diversity patterns and can lead to different inferences about processes structuring communities.