Spatial heterogeneity of eDNA concentration as a predictor of small biomass of fish in a mountain stream

Environmental DNA (eDNA) is a powerful tool for detecting target species in aquatic systems. While many studies have demonstrated the effectiveness of eDNA in estimating the abundance and/or biomass of organisms in lotic systems, the spatial correspondence between detected eDNA and actual fish biomass remains unclear. In this study, we verified the effectiveness of eDNA as a tool for assessing the biomass of juvenile masu salmon (Oncorhynchus masou masou) using cage experiments in a natural stream that does not host salmon. We attempted to clarify the spatial correspondence between the target fish biomass and the eDNA concentration on a fine scale. By developing species-specific primers and a probe for O. masou, we successfully detected masu salmon eDNA even from 500 m downstream of the cages containing eight salmon juveniles. However, the pattern of longitudinal variation in eDNA concentrations differed significantly depending on the fish biomass and the distance from the cages. The results of tracer experiments, followed by an occupancy modeling, suggested that the eDNA can be laterally mixed more than 20 m downstream of the fish cage, and that the spatial heterogeneity of eDNA concentration in the vicinity of the target fish can be mainly explained by turbulent streamflow. These results indicate that the distribution of eDNA can be highly heterogeneous in streams when they host a small number of target fish. Therefore, when using eDNA as a predictor of fish biomass under such conditions, a well-designed sampling method that considers the spatiotemporal heterogeneity of eDNA should be adopted.