# Responses of African Savanna Trees to Large Herbivore Extinction and Rewilding

## TL;DR

A long-term experiment in an African savanna shows that removing large herbivores increases tree cover, but reintroducing them only partially reverses these changes.

## Contribution

The study experimentally demonstrates that rewilding can partially reverse ecosystem changes caused by herbivore extirpation, but some effects may be persistent.

## Key findings

- Excluding herbivores for 18 years increased tree cover, canopy area, and growth rates.
- Rewilding reduced tree cover and reversed some phenotypic changes but did not fully restore original conditions.
- Tree density remained elevated despite increased mortality after herbivore reintroduction.

## Abstract

The global decline or extinction of large mammals over the last 50,000 years has caused sweeping changes in the ecosystems they once inhabited. Trophic rewilding holds promise for returning lost ecological function and restoring processes that support ecosystem resilience, but there remains considerable uncertainty surrounding the efficacy of rewilding. To address this uncertainty, we experimentally excluded a diverse African savanna mammal community from replicated plots for 18 years to simulate extinction. Herbivore exclusion caused a rapid increase in tree cover, which was underlain by shifts in community composition and increases in canopy area, growth rate and density. We then removed the exclosure fences, simulating rewilding. Reintroducing herbivores rapidly reduced tree cover and largely reversed individual phenotypic shifts, but tree density remained elevated despite increased mortality rates after reintroduction. Our results suggest that even short‐term extirpation can cause complex shifts in vegetation communities, some of which may be resistant to rewilding.

Trophic rewilding may be a restoration ‘win‐win’ if the return of extirpated wildlife also restores lost ecosystem function, but few studies have addressed whether wildlife reintroduction is capable of reversing changes that occurred during extirpation. We experimentally excluded a diverse African savanna large herbivore community for 18 years, which caused changes at individual, population and community levels that ultimately increased tree cover. After reintroducing wildlife into formerly fenced plots, many, but not all, of these changes were reversed, suggesting that some impacts of extirpation may be resistant to rewilding.