Nearshore fish diversity changes with sampling method and human disturbance: Comparing eDNA metabarcoding and seine netting along the Upper St. Lawrence River

Aquatic ecosystems are deteriorating, with the most impacted species and populations having insufficient data to inform conservation and management. Monitoring aquatic biodiversity and evaluating anthropogenic impacts typically rely on time-consuming, logistically challenging, and invasive methods (e.g., seining, trawling or electrofishing). Environmental DNA (eDNA) methods have been touted as an important advancement, especially in fish biodiversity assessment. We compare eDNA metabarcoding and seining methods along a 350 km section of the Upper St. Lawrence River (Canada) for: i) mapping the distribution of invasive and threatened fish species, ii) describing species richness and fish community structure, and iii) assessing the effect of habitat type and anthropogenic degradation of riparian zones on fish community composition and richness. eDNA detected more fish species (n = 67) than seining (n = 38) and revealed higher fish diversity in samples adjacent to intact, natural riparian zones. Fish assemblages were influenced by habitat type according to analyses using eDNA, while no effect of any environmental predictor on fish community composition was found using seining. Altogether, our results support eDNA metabarcoding as a powerful, complementary tool in fish monitoring and testing for the impacts of anthropogenic disturbances.