Data-mining previous surveys to improve walleye pollock target strength estimates

Abstract

Acoustic-trawl surveys are widely used to measure the abundance and distribution of pelagic fishes. The echo integration method used in these surveys requires estimates of the target strength (TS, dB re 1 m2) of individuals to estimate abundance. In situ TS measurements are attractive to establish length-to-TS relationships because they are obtained under realistic conditions utilizing the same tools as those during surveys. However, these measurements are usually made in restricted circumstances, are limited in number, and data processing is often subjective. Here, we present a novel method to estimate TS from a large volume of previously collected survey data recorded at trawl sites. By applying a series of published filtering methods to TS data based on frequency response, packing density, and degree of co-location in overlapping beams, single fish echoes can be reliably isolated from existing survey data. The automated approach produced measurements representative of survey conditions (e.g., collected at survey speeds on free-swimming fish) from existing surveys. We applied this method to 30 surveys of walleye pollock (Gadus chalcogrammus) in Alaska and estimated a new length-to-TS relationship. The results were largely consistent with the relationship currently used for walleye pollock; depth and geographic area were significant covariates.