Adaptive management of species recovery programs: A real-world application for an endangered amphibian

Abstract Adaptive management (AM) is often cited as an innovative approach which could improve recovery plans for threatened species, yet real-world applications remain rare. Here, we describe an adaptive program for the reintroduction of a threatened amphibian in Italy. We used an adaptive approach to decide whether to use captive breeding, headstarting of tadpoles or direct translocation of eggs, with the aim of establishing a new population while minimizing the impact of harvesting from wild populations. We built a quantitative model of the system, articulated uncertainty around model parameters using probability distributions, and evaluated the benefit of learning using simulations. Following the results of simulations, we began by implementing all actions in parallel, monitoring survival of released individuals, and updating our priors using monitoring data in a Bayesian framework. Based on the updated knowledge, captive breeding and egg translocation were discontinued after the first and second year, respectively. We found little impact of harvest on the source population, so headstarting and releases are ongoing after three years. Although survival has been lower than expected, it remains within the predicted range of possible outcomes. Our experience reinforces the potential of AM for amphibian conservation. Although AM could not guarantee success, it easily scaled to our small project, it increased our efficiency in monitoring and changing actions, and ensured a safe approach against possible negative impacts. We encourage managers of amphibian recovery programs worldwide to adopt AM by following the steps we illustrate. Disseminating the results of real-world applications is the best way to fill the current implementation gap.