Environmental DNA provides quantitative estimates of a threatened salmon species

Species of conservation interest are often rare or elusive, and often require labor-intensive population surveys for management. Sampling genetic traces of such species from environmental media such as water, air, or soil (environmental DNA; eDNA) can provide noninvasive and cost-effective means of monitoring. However, eDNA results may not align with traditional survey methods (e.g., visual, net) making it difficult to interpret eDNA results. We present the results of parallel beach seine and quantitative-PCR (qPCR) surveys of a threatened Chinook salmon (Oncorhynchus tshawytscha) from Skagit Bay, an estuary in Washington, USA. Our replicated design and hierarchical statistical model assesses the abundance, biomass, and DNA concentration at two spatial scales (site- and population-) over five months. We find both eDNA- and seine-derived abundance indices reflect the seasonal migration of salmon; at the population-scale, eDNA and seines provide virtually identical quantitative information. At the site scale, the methods are less correlated, suggesting the methods reveal different information about a patchily distributed species. Environmental DNA may act to smooth otherwise patchy biological signals in space and time. Reduced within-site variability for eDNA relative to seines suggests that eDNA may offer more precise population estimates. We partition sources of variability in space and time and compare eDNA and seine surveys – a first, to our knowledge – and so reveal the behavior of eDNA in the field. Our results underscore the value of using eDNA in conjunction with traditional surveys. Combining eDNA and seine estimates should improve the population data on which management of threatened species depends.