Removing invasive species followed by passive recovery of ecological systems is a common management practice. This practice assumes that ecological systems are resilient to invasions, and that systems should positively recover to some pre-invaded state. However, recent studies find that removal and passive recovery results in various outcomes, including no recovery or unintended negative consequences, such as “surprise or secondary effects.” Given the high variation in resilience, we expect that when different types of management are employed or when systems vary in ecological context they are more or less likely to recover. For example, certain elements of management might facilitate recovery, including when multiple invaders are removed, or time for recovery is long. Additionally, characteristics of the invader or invaded ecosystem might influence the likelihood of recovery, including the trophic level of the invader, the ecosystem type, or if there are other anthropogenic disturbances. We calculated the effect of invader removal as the log response ratio and measured changes in various types of ecological responses, including abundance, species richness, or nutrient concentrations. We collected data on predictor variables and compared models with different combinations using AIC (Akaike information criterion) to uncover in what contexts ecological recovery is most likely occur.
Results/Conclusions
We calculated effect sizes for 453 comparisons from 163 studies, with a mean effect size of 0.525 (+/- 0.158). We examined in variation in effect sizes were explained by various predictor variables. We found that conservation unit significantly explained variation in effect sizes with strong positive effects of invader removal on population-level conservation units (e.g., native species cover) with 75% of response types significantly increasing post-removal. In contrast, more complex conservation units, such as community-level (e.g., species richness), and ecosystem-level (e.g., nitrogen concentrations) responses were fewer and weaker, with only 18% of response types significantly changing. We also found that level of anthropogenic disturbance influenced recovery, including with areas possessing multiple invaders less likely to recover. Given the importance of conservation unit, we are currently running models with predictor variables within different conservation units and will present the results of these analyses. Overall, we are finding that ecological systems in certain contexts are more resilient than others and our synthesis provides important guidance for managing invaded ecosystems. We show that in some cases, invader removal alone will not be sufficient for alleviating impacts of invaders and provide guidance for when managers might need to anticipate employment of active management.