Documenting biodiversity, species occurrence, and species status require reliable monitoring techniques, but the complex life history and cryptic behavior of many anurans create challenges for conventional monitoring approaches. Environmental DNA (eDNA) surveys are a promising alternative (or complement) to conventional anuran monitoring, but their relative success has not been fully tested. We assessed the comparative efficacy of targeted eDNA detection via quantitative PCR (qPCR) and three conventional amphibian survey methods (visual encounter, breeding call, and larval dipnet surveys) for detecting nine anuran species in natural wetlands in southern Ontario, Canada. Our analyses revealed that all assessment methods yielded imperfect detection, with visual encounter and eDNA surveys detecting the greatest species richness and eDNA surveys requiring the fewest sampling events. Amphibian community composition results differed among survey methods and sampling events, and detection efficacy was markedly variable, with some species requiring two to three methods to maximize detection success. Notably, two relatively terrestrial species (Anaxyrus americanus and Hyla versicolor) had relatively low and seasonally variable eDNA detection rates, suggesting that species-specific ecology likely affects eDNA presence or detection. These findings suggest that optimized monitoring for complex anuran communities may require application of multiple monitoring methods, which may need to be tailored to individual target species or communities.

Environmental DNA (eDNA) monitoring is rapidly becoming an established approach for detecting the presence of aquatic organisms and may also be useful for indexing or estimating species abundance. However, the link between eDNA concentration and abundance of individuals (i.e., density or biomass) remains tenuous and may vary widely across species and environmental conditions. We investigated the relationship between eDNA concentration and abundance in two common and closely related amphibians in eastern North America, the wood frog (Rana sylvatica), and northern leopard frog (R. pipiens). We manipulated tadpole density in 80-L mesocosms and documented the relationship between tadpole density and biomass and eDNA concentration through time. The two species differed in the amount of detectible genetic material produced, despite having comparable biomass. Concentration of eDNA increased with tadpole numbers and was primarily correlated with tadpole density in wood frogs and biomass in leopard frogs. eDNA degradation rates were rapid and comparable between species, with tadpoles becoming indetectable within 5 days post-removal from the mesocosm, irrespective of tadpole density. Overall, our findings support that eDNA concentration has potential for tracking amphibian abundance in wetlands, but that indices of abundance are likely to be coarse and speciesspecific calibration will be required. Future research should address how biotic and abiotic factors influence eDNA production, degradation, and recovery across species and through time before relying on eDNA for monitoring amphibian abundance in nature.