Multivariate identification of plant functional response and effect traits in an agricultural landscape.

Plant functional traits have been proposed as a linkage between the environmental control of vegetation and ecosystem function. Identification of traits that mediate the response of plant species to the environment is well established, but the identification of effect traits and the linkage between the two sets is less developed. This was attempted for a study of eight contrasting land uses in a marginal agricultural landscape where data on vegetation, management controls of the disturbance regime, and soil characteristics, including nitrogen release, were measured simultaneously with measures of ecosystem function such as litter decomposition rates and primary productivity on 30 sites. Trait data were assembled from databases, and an iterative multivariate approach using the three table (species, trait, environment) method RLQ was employed to identify a parsimonious set of traits that predict plant species responses to the environment and a parsimonious set of traits that link vegetation to ecosystem function. The lists of response and effect traits were similar, and where differences were observed, traits were usually highly correlated with at least one trait in the other list. This approach identified a small number of traits (canopy height, leaf dry matter content, leaf size, and specific leaf area) that provide a means of linking vegetation responses to environmental change with changes in ecosystem function. Other response traits included vegetative spread strategy, start of flowering, and seed terminal velocity, but within the system studied these traits were all significantly correlated to the traits shared between the response and effect lists.