Patterns of taxonomic and phylogenetic beta diversity and the process structuring tree species communities in Amazonian forests

Abstract

Beta diversity is based on species identity differences between communities/localities, and is important for community ecology theory, biogeographic studies, and for defining sampling voids and systematic conservation planning. This proposal will model the taxonomic and phylogenetic beta diversity of tree species on Amazonian forests and mapping they spatial distribution to infer the relative influence of environmental and geographic gradients on tree community assembly. We will implement three approaches to build site vs. species matrices that allow beta diversity computation: 1) occurrence records; 2) species rage polygons; 3) inventory plots. To measure turnover on species composition between communities, we will compute the Simpson's dissimilarity index - ²sim - which controls for species richness gradients. To measure phylogenetic beta diversity, we will apply a synthesis dated phylogeny for vascular plants and compute the "true" phylogenetic turnover index - PhyloSorTURN -which removes the influence of phylogenetic diversity gradients. The compositional and phylogenetic beta diversities will be related to ecological and geographic distances between locality pairs using Generalized Dissimilarity Modeling (GDM). The ecological gradient measurements will rely on climatic and edaphic descriptors. Model deviance explanation will be partitioned between the climatic, edaphic and geographic distances to infer the relative role of processes structuring compositional and phylogenetic beta diversity, and to assess the congruence between those predictors to explain the compositional and phylogenetic beta diversity. The fitted GDM's will be useful to predict the beta diversity distribution throughout the Amazonian forests, building baseline maps for expected species compositional and phylogenetic turnover.