A case for multi-gear assessments: Detection probabilities of nearshore fish with eDNA and seine nets vary by functional traits

Ecological studies aim to understand species distributions, yet the methods used to sample affect which species are detected (i.e., gear selectivity) and may be influenced by species traits. Interactions between traditional gear selectivity for marine fishes and their species traits have been studied, but such studies have generally not extended to selectivity of environmental DNA (eDNA). Here, we investigated which functional species traits (scale type, schooling behaviour, and position-in-water-column) are selected for by eDNA metabarcoding and beach seines in nearshore eelgrass, mixed eelgrass, and understory kelp habitats. Using data from 35 sites across southeast Alaska, we applied occupancy modeling to estimate detection of species traits by each gear type. Detection probability with eDNA was 27 times greater for species with deciduous scales compared to species with non-deciduous scales, and lower for species with plates (rather than scales). Conversely, species with plates showed greater odds of detection with beach seines. Given the novelty of eDNA sampling, quantifying interactions between functional traits and gear selectivity will be important to accurately characterize species distributions across marine habitats.