Functional diversity is more sensitive to biotic filters than phylogenetic diversity during community assembly

Abstract

Ecologically important functional traits and phylogenetic relatedness may provide mechanistic insight into biotic filters influencing community assembly. To assess this, we examined the relationship between functional diversity (FD, functional trait relatedness) and phylogenetic diversity (PD, evolutionary relatedness) during grassland restoration. Temporal changes in FD and PD were used to examine the effect of two biotic filters during community assembly: intraspecific variation in dominant species (differ-ent population sources) and interspecific variation among subordinate species (phylogenetically distinct species pools). The experimental grassland restoration contained whole plots sown with either cultivars or local ecotypes of three dominant grasses (Sorghastrum nutans, Andropogon gerardii, and Schizachyrium sco-parium). Each whole plot contained subplots sown with three phylogenetically distinct pools of subordinate species. Cover of all species was measured in permanent sampling areas for ten years, and 10 functional traits were measured for 88 species, allowing calculation of PD and FD, respectively. Overall, the communities establishing with local ecotypes of the dominant grasses were functionally more dissimilar than when established with cultivars, particularly among the Asteraceae, suggesting competitive exclusion of functionally dissimilar species by cultivars. This result was opposite to our prediction that cultivars through limiting similarity would exclude more similar species more so than would local ecotypes. The effect of the dominant grass population source was contingent upon interactions with species pools. A supplemental propagule addition of functionally dissimilar species four years since initial sowing increased FD, but this effect varied among species pools. A lack of relationship between FD and PD in this system indicated that measuring PD alone without inclusion of functional traits may miss additional factors affect-ing species coexistence. In addition, the variation in FD between population sources and among species pools indicated that the measured traits were more sensitive to these factors than to their phylogenetic rela-tionships. This analysis of long-term data from a field experiment showed new consequences of using different seed sources and species pools (as biotic filters), and supplemental seeding on PD and FD of restored grassland. Quantifying relevant functional traits in a phylogenetic framework could help identify plant population sources that enhance coexistence of desirable species.