COS 193-3 - Proactive disease management outperforms reactive action, regardless of action chosen

Abstract: The Anthropocene can be characterized by the sentiment expressed in Joni Mitchell’s Big Yellow Taxi (1970): “We don’t know what we’ve got, til it’s gone.” Conservation is often characterized by reactive responses to losses, yet mitigating threats may be more efficient, and lead to improved outcomes. Despite this, proactive conservation actions are rarely implemented. This is particularly true for emerging infectious diseases, like the Batrachochytrium pathogens that have devastated and continue to threaten amphibians. To date, Batrachochytrium salamandrivorans (Bsal) has not been detected in North America but salamander diversity is threatened by its arrival, providing a unique and rare opportunity to be proactive with management. Initiating proactive management is often challenged by limited knowledge about how a threat will manifest in a new system, limited empirical data on proposed interventions, and a lack of validated predictive models. Despite this, managers must decide if, when and how to act; delaying action is a decision and should be based on sound science. We developed a dynamic multi-state occupancy model it to predict salamander persistence and pathogen prevalence under different management intervention tactics (20 unique actions) and implementation timings (proactive vs reactive; early vs late detection) in the face of high uncertainty. To parameterize the models, we used an alternative approach - expert judgement. By harnessing expert knowledge in a formal, structured process we gathered estimates for unknown parameters using methods that reduce bias and allow characterization of a full error distribution. We found that proactive management outperformed reactive management across nearly all evaluated actions by maintaining higher levels of host occupancy through Bsal emergence. Host outcomes varied across specific proactive and reactive actions, and reactive actions exhibited greater uncertainty. No actions had the capacity to eradicate Bsal, but reactive management strategies did overall exhibit low levels of Bsal persistence at high prevalence. Furthermore, early pathogen detection which can be accomplished via robust surveillance reduced Bsal occupancy, but had negligible effects on host occupancy. Here we have built a quantitative framework to predict host-pathogen outcomes across 20 management actions and translate our ecological knowledge into information useful for management decision-making. We provided strong quantitative evidence that proactive disease management is by far a better strategy for maintaining salamander populations threatened by Bsal emergence. This framework can serve as a tool across wildlife disease systems offering a way for sound science to advance decision making and minimize delays of optimal management intervention.