Comparison of Environmental DNA Metabarcoding and Trawl Data Highlights Gear Selectivity of Marine Demersal Fish Surveys on the Faroese Plateau

Environmental DNA (eDNA) has emerged as a promising tool for studying fish dynamics in aquatic environments. However, the quantitative capabilities of eDNA metabarcoding, especially in marine systems, remain contentious due to limited comparative studies. This study presents a comparison between a standardized trawl survey and eDNA metabarcoding (12S) in demersal fishing grounds around the Faroe Islands. Data were collected at 26 sites distributed across six regions. A total of 31 fish species were detected from trawls. Inclusion of Molecular Operational Taxonomic Units increased survey taxa richness by 151% to 47, with 21 species shared between methods, 11 unique to trawling, and 15 unique to eDNA. Unique trawl detections were linked to low biomass species (rarity) and technical limitations in species differentiation (e.g., Sebastes spp.). Unique eDNA detections were associated with rarity, size, gear-behavioral traits, and pelagic species. For taxa detected by both methods, total reads were posi-
tively correlated with trawl biomass over the survey area (r = 0.85, q < 0.001), but not by site or region. Accounting for stochastic amplification in technical replicates improved the survey-wide eDNA-biomass correlation (r = 0.91, q < 0.001). Regional distribution of sandeel eDNA agreed with stomach contents of Gadus morhua, providing information on its spatial distribution otherwise
missed by trawling. Overall, eDNA metabarcoding identified small species under-sampled by trawl gear, thus enhancing species richness when methods were combined. The study also highlighted the potential for eDNA to describe the spatial distribution of sandeel, a keystone species in the Faroese marine ecosystem, and generated reliable quantitative estimates of species rank bio-
mass. However, the study highlights limitations of both eDNA and trawling that are problematic in marine systems: (i) detection of low biomass species, (ii) variability over small spatial scales, and some limitations specific to eDNA: (iii) vertical mixing of eDNA signals from pelagic and demersal habitats.