Resilience, reactivity and variability : A mathematical comparison of ecological stability measures

TL;DR

This paper introduces new variability-based stability measures for ecosystems, linking theoretical resilience and reactivity with empirical data, and demonstrating their predictive power for ecosystem responses to disturbances.

Contribution

It defines invariability measures based on stochastic and deterministic perturbations, connecting theoretical stability concepts with empirical variability measures.

Findings

Invariability measures predict worst responses to broad disturbances.

They are intermediate between resilience and reactivity.

Invariability relates to the entire transient response regime.

Abstract

In theoretical studies, the most commonly used measure of ecological stability is resilience: ecosystems asymptotic rate of return to equilibrium after a pulse-perturbation $-$or shock. A complementary notion of growing popularity is reactivity: the strongest initial response to shocks. On the other hand, empirical stability is often quantified as the inverse of temporal variability, directly estimated on data, and reflecting ecosystems response to persistent and erratic environmental disturbances. It is unclear whether and how this empirical measure is related to resilience and reactivity. Here, we establish a connection by introducing two variability-based stability measures belonging to the theoretical realm of resilience and reactivity. We call them intrinsic, stochastic and deterministic invariability; respectively defined as the inverse of the strongest stationary response to white-noise and to single-frequency perturbations. We prove that they predict ecosystems worst response to broad classes of disturbances, including realistic models of environmental fluctuations. We show that they are intermediate measures between resilience and reactivity and that, although defined with respect to persistent perturbations, they can be related to the whole transient regime following a shock, making them more integrative notions than reactivity and resilience. We argue that invariability measures constitute a stepping stone, and discuss the challenges ahead to further unify theoretical and empirical approaches to stability.