Mining previous acoustic surveys to improve walleye pollock (Gadus chalcogrammus) target strength estimates

Abstract

Acoustic-trawl surveys are widely used to measure the abundance and distribution of pelagic fish. The echo integration method used in these surveys requires estimates of the target strength (TS, dB re 1 m2) of acoustic scatterers. Here, we present a new automated method to estimate TS from a large volume of previously collected acoustic survey data recorded near trawl sites. By applying a series of selection and filtering methods to echosounder data, single echo measurements representative of fish encountered during surveys can be objectively and reliably isolated from existing survey data. We applied this method to 30 surveys of walleye pollock (Gadus chalcogrammus) conducted in Alaska from 2007 to 2019 and estimated a new length-to-TS relationship. The resulting relationship was largely consistent with previous in situ estimates made during dedicated, mostly nighttime TS collection events. Analysis of this sizeable data set (n = 142) indicates that increased fish depth, lower ambient temperature, and summer months may increase pollock TS. The application of a new TS model incorporating these environmental covariates to historic surveys resulted in -16 to +21% changes in abundance relative to the model without environmental covariates. This study indicates that useful TS measurements can be uncovered from existing datasets.