Degradation and dispersion limit environmental DNA detection of rare amphibians in wetlands: Increasing efficacy of sampling designs

The detection of rare macroorganisms using environmental DNA (eDNA) is a powerful new method for conservation and management; the efficacy of this method is affected by physiological, ecological, and hydrological processes. Understanding the processes limiting eDNA detection and accounting for those factors with optimized sampling designs is critical for realizing the potential of this tool. Amphibians are a focus of conservation programs globally and are often difficult to detect, presenting a challenge for effective action. To increase the ability of eDNA techniques to inform conservation and management programs, we investigated the eDNA detection of amphibians compared with field surveys for six species across a gradient of environmental factors expected to affect eDNA detection in three different systems: perennial wetlands, intermittent wetlands, and acidic intermittent wetlands. We applied a baseline sampling design in each wetland and used an occupancy modeling approach to evaluate evidence for processes limiting detection for each species given the presence of the target species. Evidence weights indicated that limiting processes varied across systems and included those associated with increased degradation (pH < 5, temperature > 25 degrees C) and limited dispersion (wetland area > 1200 m(2), sample volume < 200 mL). Optimized sampling protocols based on model results included an increased number of sampling locations in large and highly degradative (acidic) wetlands and increased filter pore size in high-particulate systems. These improved designs compensated for the previously limiting factors and yielded average detection rates of 0.62-0.86 per water sample. Degradation and dispersion processes appear to strongly influence the detection of amphibians in wetlands. Optimized, adaptive sampling designs can greatly increase the efficacy of eDNA monitoring approaches. (C) 2018 Elsevier B.V. All rights reserved.