Acoustic properties, shoaling behavior and diel vertical migration of Atlantic redfish (Sebastes spp.) were studied with the objective of enhancing the basis for acoustic assessment techniques. The acoustic target strength (TS) of redfish was measured using both ex situ and in situ methods. To conduct ex situ measurements, a technique was developed to maintain captive individuals alive for study in an experimental set-up using a split-beam echosounder. Techniques of in situ (direct) TS estimation were also studied in detail and approaches to minimize biases due to multiple scattering were proposed. Systematic bias in TS occurred above a certain fish density threshold. Furthermore, the measurement scale used to estimate density influenced the threshold level. Deep-tow dual beam and hull-mounted split beam systems were used to measure the TS of redfish at sea (in situ) under various environmental and biological conditions. After controlling for effects of thresholding, fish density, reverberation volume, range, and avoidance behavior, in situ and ex situ data did not differ and were pooled to obtain a TS-length model of TS = 20Log[length(cm)]-68.7.
Redfish exhibited strong patterns of diel vertical migration and many diverse types of shoaling behavior at several spatial scales, both of which can affect acoustic and trawl survey results. During vertical migration, TS did not change significantly. Combined with a 24-hr caged experiment on an immobilized fish, these results suggested that redfish have an endogenous cycle for the control of swimbladder gas. Catch data indicated increased mean fish length with depth, while length standard deviation was highest in shallow areas. Pelagic schools of various size and density were most evident in spring and summer. The extent of vertical migration appeared to be limited by differential hydrostatic pressure and related to light intensity and feeding conditions.
Significant differences between trawl and acoustic density estimates were observed. Acoustic assessments were biased downward during the day, because an increased proportion of fish occupied the near-bottom acoustic dead zone (DZ) and there was increased heterogeneity in distribution. This work suggests that the most accurate and least variable estimate of redfish abundance can be achieved using acoustics during the night, when fish are dispersed in the water column and most available to acoustic measurement.