Print PageOne of the great challenges of community ecology is understanding the relative importance of neutral and niche-driven community assembly mechanisms. Large forest plot data paired with mapped environmental variables can enhance our understanding of the contribution of niche processes to local community assembly. Here we present an analysis of eight long term tropical forest dynamics plots 24-50 ha in size, for which topographic and soil resource variables have been mapped. We used a method of variation partitioning via canonical redundancy analysis to decompose variation in community composition into fractions explained by spatial, soil resource, and topographic variables. We also used a technique from landscape mapping to visualize tree community structure within our plots. These maps provide a novel glimpse into the structure of tropical tree communities and help to gauge the ecological significance of the variation partitioning results.
Results/Conclusions:
All variables taken together explained between 32-74% of compositional variation within a study site, all environmental variables (soil resource and topographic) explained 14-39%, soil variables explained 9-34%, and topographic variables explained 5-29%. The great majority of explained variation consistently belonged to the spatially structured environmental and purely spatial components. There was a large amount of variability among study sites in our ability to explain compositional variation, though this may be partially explained by the degree of habitat heterogeneity in a study site. In general, soil resource variables explained a greater fraction of variation than topographic variables, though this is partly due to the larger number of soil resource variables included in the analysis. We used the beta diversity maps to inform our interpretation of the variation partitioning results, and argue from them that local habitat heterogeneity may be more important than previously thought.
