121-19 Using Quantitative Acoustic Backscatter to Improve Basin-Scale Models for Eastern Bering Sea Groundfish

Robert A. McConnaughy , RACE Division, Alaska Fisheries Science Center, National Marine Fisheries Service, NOAA, Seattle, WA
Stephen E. Syrjala , RACE Division, Alaska Fisheries Science Center, National Marine Fisheries Service, NOAA, Seattle, WA
Cynthia Yeung , RACE Division, Alaska Fisheries Science Center, National Marine Fisheries Service, NOAA, Seattle, WA
Groundfish and benthic invertebrates are not randomly distributed over the continental shelf of the eastern Bering Sea (EBS).  Annual trawl surveys reveal patterns of distribution that vary according to species.  Substantial interannual variation in these patterns suggests some degree of environmental control. 

We are developing statistical models to explain the distribution and abundance of EBS groundfish.  Simple models based on readily available data (temperature, depth) are informative, but our objective is to further explain observed variability.  Earlier research in the EBS indicates that surficial sediments affect the distribution and abundance of groundfish.  However, traditional sampling (grabs, cores) is impractical over large areas, indicating a need for a different and more efficient sampling strategy.  Acoustic tools are suitable for broad-scale surveying and therefore promising, but it is generally unknown if they measure the relevant sediment properties.

We have collected seabed returns from a calibrated single beam echosounder during a 17,000 km survey covering the EBS shelf.  A generalized additive model (GAM) analysis with ten species showed statistically significant contributions of the echoreturns to the best models for each species.  The full models explained 28-77% of variability in abundance, with 2–13% of that total contributed by the acoustic predictors (QTCView principal components).  The results presented here are similar to, but less compelling than, another recent study using a sidescan sonar in the EBS where 9-54% of variability in abundance was explained by quantitative acoustic predictors.  Taken together, we suggest there are important differences in the relative costs and benefits of different acoustic systems and these should be considered when developing plans for broad-scale (EEZ) habitat assessments.