Ecosystems vary widely in their responses to biodiversity change, with some losing function dramatically while others are highly resilient. However, generalizations about how species- and community-level properties determine these divergent ecosystem responses have been elusive because potential sources of variation (e.g., trophic structure, compensation, functional trait diversity) are rarely evaluated in conjunction. At a species level, ecosystem vulnerability, or the likely change in ecosystem function following biodiversity change, is influenced by trophic position and two types of traits: response traits that determine species’ individual sensitivities to environmental change, and effect traits that determine a species’ contribution to ecosystem function. At a community level, ecosystem vulnerability is influenced by extinction order and compensatory dynamics. Here we extend the response-effect trait framework to quantify ecosystem vulnerability by developing an “ecosystem stress test,” in silico trait-based simulations for evaluating the outcomes of different extinction scenarios on ecosystem function. Using this approach, we evaluate how trophic structure, within-trait variance, and among-trait covariance affect ecosystem vulnerability by linking extinction order and functional compensation.
Results/Conclusions
We report four key findings. First, in conjunction with compensation, the community level covariance of traits that influence an individual species’ probability of extinction and its contribution to ecosystem functioning predict the extent to which an ecosystem loses or gains function if biodiversity declines. Second, trophic structure restricts the degree to which remaining species can compensate for lost species; consequently, trophic structure increases ecosystem vulnerability when response and effect traits covary positively, and decreases resilience when response and effect traits covary negatively. Third, while some species within communities may compensate for the loss of function associated with community disassembly, the strength of functional compensation is determined by the frequency distribution of traits in the community. Finally, while the positive relationship between functional diversity and ecosystem function is well established, we show that functional diversity also increases ecosystem vulnerability when species’ extinction probabilities and contributions to function covary positively. While biodiversity may increase the resistance and resilience of ecosystem function, our results suggest that this positive effect is conditional on the relationship between trophic structure, and trait covariance and distribution. Taken together, these three factors can predict when ecosystems are poised to lose or gain function with ongoing biodiversity change.