Background/Question/Methods The use of molecular methods detecting environmental DNA (eDNA) is gaining importance for the assessment of local biodiversity. However, the capability to draw quantitative inferences beyond presence/absence data is essential for any approach to ever be considered a legitimate alternative to conventional monitoring. Quantitative inferences based solely on eDNA measurements remain challenging without a priori knowledge as numerous factors influence both the shedding and distribution of eDNA in the environment. In lotic systems, biological, ecological, and environmental variables influence eDNA signal strength and distribution. Here, we examine I) how fish physiology and behavior affect eDNA shedding and II) how the incorporation of hydrological variables is key for the correct interpretation of eDNA-based data on different spatial scales. eDNA shedding rates were correlated with species-specific activity in an aquarium experiment. The importance of discharge for the quantitative interpretation of eDNA signals was investigated during spawning migrations of two potamodromous fish species, and the influence of hydraulics on eDNA distribution was examined using a caged-fish experiment. Ultimately, these findings were used for the design and execution of a field sampling campaign to semi-quantitatively survey the fish diversity of a large river system.
Results/Conclusions Our results from the aquarium experiment show species-specific eDNA shedding rates and a positive correlation between eDNA concentration and fish activity after controlling for the effect of fish mass. During the investigated spawning migrations, the number of adults present at the spawning site on a given day was described via an eDNA-based time series model after accounting for daily discharge fluctuations. In the caged fish experiment, the direct effect of local hydraulics on lateral and longitudinal eDNA distribution was shown by the distinct position and shape of the eDNA plumes downstream of the source. By incorporating both biological as well as hydrological factors in the analysis of eDNA-based data, it was possible to portrait the fish community in a heavily fragmented river system beyond site-specific presence/absence data. This exemplifies how combined knowledge on eDNA shedding and distribution in the environment can enable quantitative inferences and promote the use of molecular methods for management and conservation.