Incorporation of Habitat Data in Modeling Acoustic Dead Zone Correction and Survey Bottom Trawl Efficiency Parameters for Semi-Pelagic Species

            The population abundance of semi-pelagic fishes is typically estimated with acoustic-trawl or bottom trawl surveys, both of  which sample a limited area.  Acoustic instruments are effective in the water column, but have a near-bottom acoustic dead zone (ADZ), in which fish near the seafloor cannot be detected.  Bottom trawl surveys cannot account for fish that are located above so called effective fishing height (EFH), which is difficult to assess. Here, we present a novel modeling method that combines acoustic and bottom trawl abundance measurements  and habitat data (e.g. grain size, temperature, depth, light levels) to derive ADZ correction and survey bottom trawl efficiency parameters.  We used bottom trawl and acoustic measurements of walleye pollock (Theragra chalcograma)  abundance and available habitat data from the eastern Bering sea to present practical application of this method. Our results show that predictions of fish abundance in the ADZ can be significantly improved by incorporation of bottom habitat features such as depth and sediment particle size, as well as pelagic habitat features such as temperature and current velocity. Additionally, we show that by modeling bottom trawl catches as a function of acoustic measurements and the environmentally driven ADZ correction we can obtain improved predictions for trawl efficiency parameters such as EFH, density dependent trawl efficiency, and proportionality coefficient between trawl and acoustic data. In light of these results we concluded that the detectability of acoustic trawl surveys and well as catchability of bottom trawl surveys for walleye pollock in the EBS are spatially and temporarily variable. Our results can be directly applied in the eastern Bering sea walleye pollock stock assessment in three ways. First, ADZ correction derived from the model can be used to assess detectability the acoustic trawl surveys in relation to the habitat and environmental factors. Second, environmental effects on bottom trawl survey catchability can be assessed using estimated trawl efficiency parameters. Third, these estimates can lead to derivation of abundance estimates that are corrected for the habitat specific or environmentally dependent catchability of both bottom trawl or acoustic surveys. Additionally these results will be useful in walleye pollock spatial dynamics studies allowing for better understanding of pollock distribution in relation to environmental factors.

The modeling method used in this study can be easily extended to other surveys of semi-pelagic species in the world where ADZ is of a concern or bottom trawl survey parameters are unknown.