Quantifying the impacts of rewilding on ecosystem resilience to disturbances: A global meta-analysis

Rewilding is one approach to restoration that aims at restoring natural self-sustaining ecosystems, allowing natural processes to resume by targeting an increase in trophic complexity, disturbance stochasticity, and dispersal, while minimizing human interventions. These components have also been argued to enhance ecosystem resilience, yet this claim has barely been specifically addressed. We conducted a meta-analysis to explore whether rewilding interventions aimed at increasing biodiversity (i.e., trophic complexity), disturbance stochasticity or connectivity increase ecosystem resilience to future abiotic and biotic disturbances. We integrated two recently developed operational frameworks to address rewilding and resilience and scrutinized the outcomes of 42 case studies (305 observations). We found that, overall, the three abovementioned rewilding components increased resilience of variables related to demography, biodiversity, biophysical characteristics and the disturbance regime characteristics (70% of observations). Yet, this result was influenced by the nature of the disturbance and the resilience approach, with lower success reported for abiotic disturbances (drought and fire) and social-ecological resilience. While interventions targeting only disturbance stochasticity or biodiversity and disturbance stochasticity together showed positive effects, interventions targeting the trophic complexity alone contributed less to system variables related to biodiversity. The most common rewilding interventions, such as domestic and wild herbivore introductions and invasive plant removals, enhanced resilience towards biotic disturbances (i.e., invasions). We also found that some particular resilience contexts (social-ecological systems) lack sufficient observations to allow clear conclusions. Overall, our results empirically demonstrate the predominantly positive effects of rewilding on ecosystem resilience, underpinning the potential of this approach for preparing ecosystems for the uncertain effects of increasing climate change and associated disturbances yet acknowledging some limitations depending on the nature of the disturbance.