Predicting stability: how resistance and recovery contribute to the temporal stability and resilience of ecosystems and species

Understanding what allows natural and managed systems to maintain function in the face of disturbances is a central challenge in ecology and beyond. Ecosystems, agricultural systems, climate processes, and even financial markets all depend on some degree of temporal stability—the tendency of a system to fluctuate relatively little through time. Temporal stability has underlying components: the capacity to resist disruption when a disturbance occurs, and the capacity to recover afterward. These same components underpin resilience, defined here as the extent to which a system has returned toward unperturbed levels following a perturbation. Although these concepts are widely invoked, their quantitative relationships and predictive power have remained unclear.

In this presentation, I introduce and test a new framework that clarifies how resistance and recovery combine to influence both temporal stability and resilience. A key insight emerging from this work is that temporal stability can often be estimated from resistance alone, even when information on recovery rates is sparse or unavailable. In contrast, and consistent with previously theoretical work, resilience depends most strongly on recovery.

To empirically test these new theoretical predictions, we analyzed more than 25 years of plant productivity data from the world’s longest-running biodiversity experiment, which allowed us to test relationships at both the ecosystem and species levels. Consistent with the theoretical predictions, resistance alone provided moderately accurate predictions of long-term stability, and additionally incorporating recovery improved predictions only slightly. Resilience, however, was predicted by the combined contributions of resistance and recovery at the ecosystem level. We also found that ecosystems exhibiting greater temporal stability before a drought tended to be more resistant during the drought, suggesting that routine monitoring of variability may provide an early warning and forecasting of system responses during future perturbations.

Our findings advance the field from simply acknowledging that stability encompasses multiple dimensions to showing how four key stability metrics fit within a coherent hierarchy. In this structure, two integrated types of stability, temporal stability and resilience, each arise from two underlying components, resistance and recovery, whose relative contributions are predictable.