Biodiversity Assessment Using Environmental DNA

Abstract

Environmental DNA (eDNA) metabarcoding -- the detection and identification of species from genetic material shed into the environment (water, sediment, air) and sequenced using amplicon metabarcoding -- has emerged as a transformative tool for biodiversity assessment, capable of detecting dozens to thousands of species simultaneously from a single water sample without physical capture or observation of individual organisms. Despite rapid methodological advances, the quantitative relationship between eDNA detection probability, species abundance, and environmental conditions -- essential for translating eDNA presence-absence and read count data into ecologically meaningful biodiversity metrics -- remains insufficiently calibrated for most freshwater and marine taxa. This study presents a rigorous comparative validation of eDNA metabarcoding against conventional survey methods across 184 freshwater and 84 marine sites in 18 countries, simultaneously deploying eDNA water samples (18S rRNA + COI metabarcoding), standardised conventional surveys (electrofishing; kick-netting; trawling; SCUBA transects), and environmental occupancy models to quantify detection probability, false negative rate, and species richness estimation performance across 847 focal taxa. eDNA detected a mean of 84.7% +- 8.4% of species confirmed by conventional surveys (sensitivity), while conventional surveys confirmed a mean of 74.7% +- 9.4% of eDNA-detected taxa as truly present (positive predictive value). eDNA detected 18.4 +- 7.4 additional taxa not found by conventional surveys per site -- including rare, cryptic, and invasive species -- providing systematic added value. The ratio of eDNA reads to conventional abundance was significantly correlated (r = 0.68; p < 0.001) with species biomass, enabling semi-quantitative abundance estimation from read counts for 73.4% of tested taxa.