Electrofishing Capture Efficiencies for Common Stream Fish Species to Support Watershed-Scale Studies in the Central Appalachians

As watershed-scale studies of stream fishes become increasingly common, there is a need for more accurate estimation of fish population abundance and size structure with single-pass electrofishing techniques. Capture efficiencies are known to vary considerably across habitats and species, yet few studies have quantified or provided a logical framework for accounting for this variability. Consequently, our objectives were to (1) determine which species and which size-classes within species exhibited significant site-to-site variation in electrofishing efficiency and (2) construct models to predict species-and size-specific capture probability from physicochemical parameters. We used three-pass removal sampling to capture fishes from 40 study sites located in wadeable streams in the Cheat River and Tygart Valley River watersheds of West Virginia. The program MARK was used to estimate capture probabilities for 12 commonly sampled fishes, to assess among-site variability in capture probability, and to model variability in capture probability as a function of fish size and readily measured environmental covariates. Our results indicated that for most species sampled (9 of 12), some combination of these covariates-mean stream width, gradient, specific conductivity, and species size-class-could be used to produce a model of capture probability superior to the most complex model that accounted for site-to-site variability in capture probability. One species demonstrated constant capture probability among sites, and two others exhibited significant site-to-site variation in capture probability that could not be accounted for using any of the measured covariates. Our results provide a basis for using single-pass electrofishing data to infer population-level phenomena (e. g., relative abundance, size structure) in watershed-scale studies of stream fish assemblages.