Experimental assessment of optimal lotic eDNA sampling and assay multiplexing for a critically endangered fish

Designing eDNA tools to detect and quantify rare species includes inherent assumptions about the spatial distribution of the organism, spatial nature of eDNA dynamics, and the real‐world performance of alternate assays under field conditions. Here, we use cage experiments with small numbers of Atlantic salmon (Salmo salar), to reveal that eDNA detection rates and eDNA quantities follow a predictable, but nonlinear relationship with distance from a point source. In contrast to the common assumption of consistent eDNA degradation moving away from a source, eDNA detections and concentrations increased up to roughly 70 m downstream before declining steadily. We apply our eDNA distance functions to selection of stream sampling intervals for detecting fish without known locations and find that even a single juvenile salmon can be reliably detected with intervals up to 400 m spacing. Finally, we show that two different qPCR eDNA assays provide very different detection probabilities in nature despite similar efficiency in laboratory testing, demonstrating the importance of experimentally assessing assay efficiencies in the wild as well as the capacity for multiplexing as a strategy to ensure high detection efficiency when monitoring rare species. Here, we investigate the relationship between downstream sampling distance and eDNA detection rate and quantity for lotic organisms. Using an in‐stream cage experiment, we demonstrate that eDNA detection rate and quantity increase, then decrease downstream from focal organisms. We provide recommendations for eDNA sampling designs to detect organisms in specific stream reaches and in whole streams.