eDNA metabarcoding-based source attribution of fecal indicator bacteria exceedances in urban freshwater beaches, sand and rivers

Abstract

Freshwater beach quality is routinely tested by measuring fecal indicator bacteria, which can assess water quality but cannot identify sources of fecal contamination. We compared eDNA metabarcoding and microbial source tracking (MST) digital PCR methods to identify fecal contamination sources in water and sand at four urban Lake Ontario beaches and two nearby river mouth locations. eDNA sequences matched mammal, bird, and fish taxa known in the study area. Human eDNA sequences were prominent in all water and sand samples such that they had less value for discriminating between sewage occurrence at sites. Mallard duck, muskrat, beaver, raccoon, gull, robin, chicken, red fox, and cow eDNA sequences were common across all locations. Dog, Canada goose, and swan eDNA sequences were more common in Toronto beach waters, suggesting localized sources. MST results were generally consistent with eDNA, such as finding the Gull4 DNA marker and the human mitochondrial DNA marker in most water and sand samples. Chicken, cow, and dog eDNA sequences and the human bacterial MST DNA marker often showed a higher frequency of occurrence on Beach Action Value (BAV) exceedance days. The surprisingly widespread detection of chicken and cow eDNA sequences was likely from incompletely digested human food, raising caution for interpreting eDNA results related to food animals in sewage-contaminated urban settings. Combined use of MST and eDNA methods provided a more comprehensive characterization of potential fecal contamination sources, including diverse wildlife species at the human–animal One Health interface, that can guide targeted beach-specific water monitoring and risk management strategies.