Multifrequency analyses of fish distributions in the northwest Atlantic

Abstract

Routine acoustical surveys for estimating Atlantic herring (Clupea harengus) population abundance have been conducted on Georges Bank during the autumn spawning season from 1998 to present. Acoustical data are collected with a Simrad EK500 scientific echo sounder operating at 12 or 18, 38, and 120 kHz, and split-beam (the 12-kHz system is a single beam) transducers. Biological measurements and verification of acoustical scatterers are obtained with a pelagic trawl. Acoustical data are evaluated (scrutinized) manually to remove noise, faulty bottom detections, and to classify acoustical backscattering to species. Species classification is currently subjective, and is based on the experience of the scientists and trawl catches. Objective species classification and automated fish density and abundance estimates are an obvious goal for fisheries surveys using advanced technologies. Classification methods using relationships among frequency-dependent volume backscattering strengths, such as presence–absence and combination–permutation, are described and presented. Results indicate that while classification using these methods and acoustical information alone is not robust, these methods highlight backscattering patterns within aggregations and have the potential to characterize backscattering patterns observed in fisheries acoustics data. [Work supported by NOAA Fisheries and ONR.]