Using environmental DNA and occupancy modelling to identify drivers of eastern hellbender (Cryptobranchus alleganiensis alleganiensis) extirpation

Population declines and local extirpation trends are widespread among freshwater species, but the responsible drivers of these trends are poorly understood. Identifying the potential drivers of population declines is essential to effective conservation planning. However, conventional detection methods used to monitor cryptic and elusive freshwater species are inefficient. Integrating new surveying and modelling techniques may allow for a more comprehensive assessment of population declines. We used environmental DNA (eDNA) sampling methods and detailed historical records to identify drivers of local extirpation in a declining, long‐lived giant salamander, the eastern hellbender (Cryptobranchus alleganiensis alleganiensis) in West Virginia, U.S.A. We used a site occupancy and detection modelling framework (SODM) to test the influence of current land use, historical mining, hydrogeomorphic and water quality variables on model‐based predictions of hellbender extirpation and detection. We failed to detect hellbender eDNA at 51% (naïve 1 – Ψ) of historical sites, suggesting local extirpation at a broad spatial scale in West Virginia. Our best‐supported SODM model suggested catchment‐scale road density was the best predictor of hellbender extirpation, and that 38% (predicted 1 – Ψ) of historical sites may be locally extirpated. Estimates of hellbender occupancy probability were extremely low in highly developed catchments. Water turbidity and conductivity were the best predictors of eDNA detection, both negatively influencing detection probability. Roads can increase sedimentation rates and alter water chemistry of freshwater ecosystems, identifying landscape alteration/human development and water quality declines as possible drivers of hellbender extirpation trends in West Virginia. Our findings also suggest that water conductivity and turbidity may act as polymerase chain reaction inhibitors and decrease eDNA detection in lotic systems. This study emphasises the negative impacts of urban development on freshwater ecosystems and the sensitivity of long‐lived amphibian species to rapid environmental change. Our findings may aid in conservation planning by providing a sampling framework that integrates eDNA data within a SODM framework to rapidly and accurately assesses relational changes in aquatic species' occupancy at historical sites.