Building evolutionary resilience as a framework for managing safe haven populations

Safe havens are widely used to mitigate the impacts of invasive predators on threatened species, yet their isolation often disrupts evolutionary processes and can leave these species vulnerable to environmental change and stochastic events. Effective approaches to overcome these limitations and promote adaptive capacity in safe haven populations are lacking in conservation biology. Thus, we developed a pragmatic, heuristic management framework for integrating evolutionary principles into safe haven management with a focus on establishing a metapopulation of individuals distributed across geographically separated safe havens spanning environmental gradients. We used genetic simulations to demonstrate how this approach can enhance adaptive potential and persistence and to explore the optimal configuration of safe havens and number of individuals in each haven. Using demographic simulations for two threatened Australian mammals, we illustrated how this approach can be practically implemented. Based on spatially explicit population models, we predicted that many existing safe havens for these species are at risk of collapse within decades. However, our simulations showed that selection based on standing genetic variation can drive fitness gains within tens of generations, improving population resilience to novel conditions. Integrating these processes revealed that safe haven networks can promote adaptation and resilience, substantially improving long-term persistence and providing robust sources for reintroductions.