AUTHORS: Garrett R. Johnson U.S. Fish & Wildlife Service; Benjamin J. Marcek U.S. Fish & Wildlife Service
ABSTRACT: Species-specific density estimates derived from hydroacoustic data often require fish community information to apportion hydroacoustic targets to species. However, the tools used to collect fish community data have biases that can affect species composition and size structure. The effect of these community gear biases on species-specific hydroacoustic density estimates is unknown. To address this knowledge gap, we collected fish community data using three gears (boat electrofishing, electrified dozer trawl, and gill nets) during hydroacoustic surveys in two navigation pools of the Ohio River (J.T. Myers and Newburgh pools) during fall 2021. Hydroacoustic data were collected using two side-looking, split-beam transducers offset to maximize water volume sampled. We used a Bayesian hierarchical model to estimate the probability of a fish being a Silver Carp given its length for each community gear. We also estimated the total length of individual fish targets from the hydroacoustic surveys using Love’s 1971 side-aspect equation and the probability of each of these hydroacoustic targets being a Silver Carp based on its estimated length. We then estimated site-specific Silver Carp densities and compared these densities among gears, pools, and habitats (main channel, side channel, tributary, and backwater). The probability of a fish being a Silver Carp given its length differed between community sampling gears. Further, gear affected the overall density estimates with gill nets producing the lowest estimates in all pool-habitat combinations. No apparent difference existed between density estimates apportioned with dozer trawl and boat electrofishing except in Newburgh Pool backwater and tributary habitats where the dozer trawl produced greater Silver Carp density estimates. These findings emphasize the need to consider community gear bias when collecting data to apportion hydroacoustic targets to species. Combining community data from multiple gears may reduce the effect of gear bias on the apportionment of hydroacoustic targets to species.
