Species diversity patterns are shaped by multiple ecological processes across the range of the longleaf pine (Pinus palustris Mill.) ecosystem

Background/Question/Methods

A key challenge for ecologists has been to identify which ecological processes structure communities at local scales.  Historically, studies have focused on the influence of local, deterministic processes (e.g., biotic interactions), but recently, ecologists have recognized that regional and historical processes also shape species diversity and composition in local communities.  Here, I explore the relative contribution of local, regional, and historical processes structuring longleaf pine plant communities.  Specifically, I focus on local environmental filtering (soil properties), regional environmental filtering (mean annual temperature, MAT), and biogeographic history (approximated as the size of the regional species pool). In addition, I explore how the relative importance of these processes changes across ecoregions and at different grain sizes (1 m² and 1000 m²).  Species richness and beta-diversity was calculated for 858 longleaf pine plots from NC to FL. Principal components analysis was used to extract the major axes of variation in soil properties. Species pools were built for the four ecoregions in the southeastern US from Weakley 2012 and from species occurrences in the data set. Multiple linear regression and variance partitioning were used to quantify the unique variation in richness and beta-diversity explained by soil properties, MAT, and species pool size.

Results/Conclusions

The size of the regional species pool varied across ecoregions from 1107 species in the Florida Peninsula to 1373 species in the East Gulf Coastal Plain. Species richness at 1 and 1000m² increased slightly as latitude decreased (R²=.02, R²=.09; p<.001), suggesting longleaf pine plant communities in southern latitudes are more species-rich. In contrast, beta-diversity did not increase with decreasing latitude (R²<.001, p=.45).  For all plots, soil properties were the most important predictor of species richness and beta-diversity, regardless of spatial scale (R²=.28 1000m², R²=.28 1m², R²=.21 beta-diversity), with higher richness and beta-diversity on sites with greater nutrient content, silt %, and soil moisture. Additional variation was explained by MAT, although significantly less than soil properties. Interestingly, the size of the regional species pool only explained additional variation in richness at 1000m². Across ecoregions, soil properties were the most important predictor, but explained less variation in richness and beta-diversity in southern ecoregions (R²=.48 for species richness at 1000m² in Mid-Atlantic Coastal Plain vs. R²=.16 in East Gulf Coastal Plain).  This work suggests longleaf pine plant communities are structured simultaneously by local, regional, and historical processes, which shift in their relative importance across space and grain size.