Caged fish experiment and hydrodynamic bidimensional modeling highlight the importance to consider 2D dispersion in fluvial environmental DNA studies

Abstract The analysis of environmental DNA (eDNA) is a powerful tool to increase the efficiency of species detection and monitoring in aquatic ecosystems. Yet, several points remain to be clarified in order to estimate with better precision the distribution and abundance of targeted species, such as the dispersion and dilution of eDNA in large lotic systems. This study aimed to document the dispersion patterns of eDNA in the St. Lawrence River, the largest fluvial system in eastern North America. Caged Brown trout (Salmo trutta) were placed in two different water masses present in this part of the river, the Ottawa River, and water from the outlet of the Laurentian Great Lakes. eDNA detection of the caged fish was performed for two days following cage removal at 53 sampling stations located at 500 upstream, and 10, 100, 500, 1,000, and 5,000 m downstream from the cages. Quantitative PCR analysis using a Brown trout specific assay revealed a positive detection only at downstream stations and up to 5,000 m. To further investigate patterns of dispersion, the relative concentrations of eDNA were predicted using a bidimensional hydrodynamic model, calibrated for downstream advection and lateral mixing of particles (i.e., quantification of 2D dispersion). The detection and the quantities of eDNA obtained by qPCR analyses were compared with the model predictions. Our model which predicts a low lateral mixing and a downstream flow in direct line from the eDNA source best fits the results. We discuss how such studies can improve our capacity to produce more precise estimates of species abundance and distribution in order to better interpret eDNA signals in large lotic systems.