Effective conservation planning relies on accurate species detection. However, conventional sampling methods used for detecting rare and cryptic aquatic species suffer from low probabilities of detection. Environmental DNA (eDNA) has emerged as an innovative and powerful sampling tool for detecting aquatic species, with previous studies suggesting a detection advantage over conventional sampling. However, comparative studies often fail to consider the appropriate sampling frameworks to adequately compare sampling methodologies and account for the influence of environmental variables on eDNA detection probabilities. In this study, we paired two detection methods (eDNA and physical sampling) at 22 sites in West Virginia, USA, to compare the probability of detecting a cryptic, elusive, and imperiled species of giant salamander, the Eastern Hellbender (Cryptobranchus alleganiensis alleganiensis). We used a multimethod occupancy modeling framework to compare method-specific detection probabilities using a suite of predictor variables based on environmental conditions thought to influence hellbender detection. We detected hellbenders at 19/22 sites using eDNA and at 13/22 sites using physical sampling methods. The best supported model indicated that detection probability for eDNA (0.84 ± 0.06) was three times higher than conventional methods (0.28 ± 0.07). Water turbidity was the best predictor of hellbender detection and negatively impacted our ability to detect eDNA. We failed to detect an association between eDNA concentration and hellbender catch per unit effort. Our study supports previous findings that suggest eDNA sampling methods greatly increase the probability of detecting aquatic species. However, with little known about the influence of environmental variables on eDNA detection, our results highlight the negative influence turbidity and other physiochemical factors have on eDNA detection and suggest that further research on eDNA detection in turbid environments is needed. | 87 WINELAND Et AL.