Beta-diversity components change after a severe natural disturbance in a subtropical broad-leaved forest of China
In early 2008, an ice storm caused massive damage to the trees in a 24-ha stem-mapped species-rich plot in the subtropical evergreen broad-leaved forest of the Gutianshan National Nature Reserve in Eastern China. This event occurred between two planned surveys, in 2005 and 2010.What was the effect of the ice storm on the distribution of the species (biomass data) across the Gutianshan plot?
The variation in species composition among sites, or beta diversity (BD), can be decomposed into replacement (Repl) and biomass difference (BiomDiff). The Repl and BiomDiff indices can be interpreted and related to ecosystem processes. Here, we focus on the percentage-difference index (D), which is the quantitative equivalent of the Sørensen dissimilarity index; it describes the magnitude of changes in species biomass. We will illustrate how this index, and its replacement and abundance difference components, changed between two surveys after a severe natural disturbance at four different scales in Gutianshan plot.
In each survey, the tree community was dominated by replacement processes. D, Repl and BiomDiff showed significant differences between the two surveys through a paried t-test for dissimilarity matrices. Because of the ice storm, the pairwise dissimilarities in D significantly increased between the first and the second survey for four quadrat sizes (scales), while the contribution of Repl decreased and the contribution of BiomDiff increased. The variation of D, Repl and BiomDiff matrices were significantly explained by four topographic factors through db-RDA. Explanation (R2adj) of topographic variables on D, Repl and BiomDiff did not signficantly change between the two surveys. Paired t-tests showed that there were significant differences in the Local Contributions to Beta Diversity (LCBD) of D, Repl and BiomDiff at all scales between the two surveys. For D and BiomDiff, more quadrats increased LCBD in all scales; in contrast, for Repl, more quadrats decreased LCBD. There were three identifiable groups of quadrats in each survey for all four scales. There were also significant differences in group memberships and group indicators species between two surveys.