Decomposing functional trait associations in a Chinese subtropical forest

Functional traits, properties of organisms correlated with ecological performance, play a central role in plant community assembly and functioning. To some extents, functional traits vary in concert, reflecting fundamental ecological strategies. While “trait syndromes” characteristic of e.g. fast-growing, early-successional vs. competitive, late-successional species are recognized in principle, less is known about the environmental and genetic factors at the source of trait variation and covariation within plant communities. We studied the three leaf traits leaf half-life (LHL), leaf mass per area (LMA) and nitrogen concentration in green leaves (N(green)) and the wood trait wood density (WD) in 294 individuals belonging to 45 tree or shrub species in a Chinese subtropical forest from September 2006 to January 2009. Using multilevel ANOVA and decomposition of sums of products, we estimated the amount of trait variation and covariation among species (mainly genetic causes), i.e. plant functional type (deciduous vs. evergreen species), growth form (tree vs. shrub species), family/genus/species differences, and within species (mainly environmental causes), i.e. individual and season. For single traits, the variation between functional types and among species within functional types was large, but only LMA and N(green) varied significantly among families and thus showed phylogenetic signal. Trait variation among individuals within species was small, but large temporal variation due to seasonal effects was found within individuals. We did not find any trait variation related to soil conditions underneath the measured individuals. For pairs of traits, variation between functional types and among species within functional types was large, reflecting a strong evolutionary coordination of the traits, with LMA, LHL and WD being positively correlated among each other and negatively with N(green). This integration of traits was consistent with a putative stem-leaf economics spectrum ranging from deciduous species with thin, high-nitrogen leaves and low-density wood to evergreen species with thick, low-nitrogen leaves and dense wood and was not influenced by phylogenetic history. Trait coordination within species was weak, allowing individual trees to deviate from the interspecific trait coordination and thus respond flexibly to environmental heterogeneity. Our findings suggest that within a single woody plant community variation and covariation in functional traits allows a large number of species to co-exist and cover a broad spectrum of multivariate niche space, which in turn may increase total resource extraction by the community and community functioning.