Functional diversity (FD) metrics are widely used to assess invasion ecosystem impacts, but we have limited theory to predict how FD should respond to invasion. A key challenge to effectively using FD metrics is the complexity of conceptualizing alterations to multidimensional trait space, making it difficult to select a priori the most appropriate metric for specific ecological questions. Here, we provide expectations on how invasion should change four commonly used FD metrics—functional richness (FRic), evenness (FEve), divergence (FDiv) and dispersion (FDis)—and then test these expectations in a laboratory decomposition experiment. We simulate invasion of a forest by understorey plants by adding leaf litter from 18 natives and non-natives to a representative canopy tree litter mixture to test changes in FD and decomposition.
Results/Conclusions
All four metrics changed predictably with invasion. Species that were more functionally unique or when added at greater proportions had larger impacts on FD. Overall, FRic, FEve and FDiv were poor choices for understanding impacts of non-native species. FDis was the only metric that both changed predictably with addition of understorey litter and correlated intuitively with changes in carbon mineralization. Furthermore, ranking species based upon how much they changed FDis of the litter mixture provided a fair assessment of which species had the largest impact on decomposition. As such, functional dispersion may be a key tool for predicting a priori which non-natives will have the greatest impact on ecosystem processes. We highlight the need to assess the suitability of each functional diversity metric for the specific ecological question at hand. Our work reveals the pitfalls of considering multiple metrics or randomly choosing a single metric without suitability assessments. At the same time, it suggests a framework for metric assess- ment that should help lead to selection of a metric or metrics that provide robust a priori insights into how invasion by non-native species can impact ecosystem processes.