Determinants of tree species turnover in a southern Amazonian rain forest

Questions What is the relative importance of environmental variables and geographical distances to explain tree species turnover? Are these patterns consistent for different tree categories, i.e. all trees (DBH = 2.5 cm), large trees (DBH = 10 cm), small trees sensu lato (DBH < 10 cm) and small trees sensu stricto (strictly understorey species, DBH < 10 cm)? Location Department of Beni, Bolivia, southwestern Amazon. Methods A total of 55 0.1-ha plots were inventoried in old-growth terra firme forest in seven sites. Composite soil samples from each plot were analysed for physical and chemical properties. Environmental and geographical influences on tree species turnover were quantified with Mantel correlations and variation partitioning based on multiple regressions on distance matrices. Results Floristic differences between sites yielded significant correlations with both geographical distances and environmental variables (pH, Ca, Mg, exchangeable acidity, C:N ratio, sand content) for all tree categories. Phosphorus was correlated with floristic patterns only for small trees sensu stricto. Together, geographical distances and environmental variables explained 62% of the floristic variation for all trees. Environmental variables explained more variation for large trees than for small trees sensu stricto. Conclusions The results support the hypothesis that species distributions are driven by both geographical distances (as quantifiers of dispersal processes) and environmental variables (niche factors) in similar proportions. Spatial floristic patterns of large trees and small trees sensu stricto were in general terms congruent.