In light of global change, there is an increasing urgency to successfully harness restoration to safeguard biodiversity and yield resilient and functioning ecosystems. In measuring biodiversity, approaches that incorporate species’ functional traits (i.e. measures of functional diversity) are crucial in linking biodiversity with ecosystem functioning in ways richness-based measures alone cannot. However, there lacks a comprehensive global assessment of the effectiveness of restoration in the recovery of functional diversity.

I conducted a meta-analysis of 30 restoration projects (freshwater and terrestrial) by extracting species lists from published studies and matching these to publicly available trait data. I compared actively and passively restored sites with degraded and pristine control sites with respect to three key measures of functional diversity (functional richness, evenness and dispersion) and two measures of species diversity (species richness and evenness). I conducted separate analyses for longitudinal studies (which monitored control and restoration 15 sites through time) and space-for-time substitutions, which compared control sites with restoration sites of different ages at one point in time.

Overall, restoration appeared to be effective in space-for-time studies, with restored sites improving across multiple diversity measures over time. However, the studies that were best able to detect a difference (i.e. replicated longitudinal data) did not find sustained benefits of restoration for any measure of functional diversity, suggesting that the positive results found in space-for-time data may have been an artefact of the inability of the study design to control for regional changes across all sites. Further, active measures (i.e. guided recovery) were no more effective than passive measures (i.e. unassisted regeneration) at restoring species diversity or functional diversity. My findings on differences across study designs explain the variable results found by recent studies that directly measured the response of functional diversity to restoration, as many did not have these controls for temporal changes, whereas the study that did found no long-term effect of restoration. Further to this, functional richness and functional dispersion increased logarithmically with species richness, though this observed relationship was no different than could be expected if assemblages of species had been generated at random. Patterns were consistent across the six taxonomic groups, six ecoregions and two realms (freshwater and terrestrial) included in this work.

Based on these findings, I stress the indispensability of including negative degraded controls in ongoing monitoring to distinguish the consequences of restoration efforts from unassisted temporal changes. Additionally, the failure of active restoration to outperform passive restoration suggests that allocating resources towards less intensive measures over larger areas may be a successful strategy to optimise gains for functional biodiversity.