Stressor interactions at the community level: explaining qualitative mismatches between observations

Abstract

Perturbations such as climate change, invasive species and pollution, impact the functioning and diversity of ecosystems. But because there is no unique way to measure functioning and diversity, this leads to a ubiquitous and overwhelming variability in community-level responses, that is often seen as a barrier to prediction in ecology. Here, we show that this variability can instead provide insights into hidden features of a communitys functions and responses to perturbations. By first analysing a dataset of global change experiments in microbial soil systems we show that variability of functional and diversity responses to a given perturbation is not random: aggregate properties that are thought to be mechanistically similar tend to respond similarly. We then formalise this intuitive observation to demonstrate that the variability of community-level responses to perturbations is not only predictable, but that it can also be used to access hidden and useful information about population-level responses to perturbations (i.e., response diversity and scaling by species biomass). Our theory offers a baseline expectation for the variability of community-level responses to perturbations and helps to explain the complexity of ecological responses to global change. Significance StatementMeasures of biodiversity and ecosystem functioning show highly variable responses to a given perturbation. This variability is traditionally thought of as reflecting our inability to predict ecological responses to global change. Our work, however, finds that variability of community-level responses is itself predictable and can even be used to gain insights about how species respond to perturbations and collectively contribute to ecosystem functions.