Quantitative insights into the spatio‐temporal variation of Atlantic salmon ( Salmo salar ) biomass in a river catchment using eDNA metabarcoding

Abstract

Effective species conservation and management requires comprehensive biomonitoring, enhanced by combining traditional and newer methodologies, such as environmental DNA (eDNA) analyses. A seasonal pulse of spawning adult Atlantic salmon ( Salmo salar ) was detected by normalised eDNA 12S reads from metabarcoding, which facilitated estimation of spatial patterns in salmon biomass. A strong relationship was found between normalised reads in the lower section of the River Conwy (Wales, UK) and whole‐river adult biomass (estimated from rod catch data and a fish counter), explaining 61% of the variation in a linear regression. Moreover, the positive linear relationship between adult biomass and partial effect on normalised reads occurred after the biomass estimate exceeded 1500 kg, indicating a threshold where normalised reads become representative of biomass. The relationship observed between normalised reads and biomass, as well as the unique profiles of normalised reads at each of the sites, supports the hypothesis of limited eDNA transport among sampling sites that were 2–4 km apart. River pH showed a significant non‐linear relationship with normalised reads, with a peak in partial effect on normalised reads at pH 6.5. Partial effect on normalised reads also showed a positive linear relationship with flow (discharge), while also peaking at the highest average monthly air temperatures (14°C). These trends are contrary to what would be expected from eDNA decay, dilution or transport, demonstrating that metabarcoding is robust to such influences, reinforcing the interpretation of trends driven by Atlantic salmon ecology and physiology. For example, pH effects reflect beneficial conditions for eggs and perhaps habitat preference for spawners, flow effects reflect the annual return of salmon during higher flows which aid upstream migration, tributary entry and spawning, and, finally, temperature effects reflect higher metabolic rates and greater shedding of eDNA.

Bibliographic Information