Characterizing the spatial signal of environmental DNA in river systems using a community ecology approach
AbstractEnvironmental DNA (eDNA) is gaining a growing popularity among scientists but its applicability to biodiversity research and management remain limited in river systems by the lack of knowledge about the spatial extent of the downstream transport of eDNA. Up to now, attempts to measure eDNA detection distance compared known species distributions to eDNA results, limiting therefore studies to a few intensively studied rivers. Here we developed a framework to measure the detection distance of eDNA in rivers based on the comparison of faunas across an increasing range of spatial extents, making it independent from knowledge on species distributions. We hypothesized that under short detection distance the similarity between fish faunas should peak between nearby sites, whereas under long detection distance each site should cumulate species from a large upstream area. Applying this framework to the fish fauna of two large and species rich Neotropical river basins (Maroni and Oyapock), we show that fish eDNA detection distance did not exceed a few kilometers. eDNA hence provided inventories of local species communities. Those results were validated by retrieving the distance decay of species similarity, a general pattern in ecology based on the decline of local species community similarity with spatial distance between them. We finally compared species distribution derived from eDNA to the known distribution of the species based on capture data, and this comparison also confirmed a global match between methods, testifying for a short distance of detection of the fauna by eDNA.