Estimating the biomass of a mixed species complex using hydroacoustics and catch data from the Bay of Fundy and Scotian Shelf summer ecosystem survey
An acoustic-based index of biomass was developed due to concerns that the existing bottom-trawl index for western component pollock (Pollachius virens) was inaccurate and imprecise. Acoustic data were recorded during annual summer ecosystem surveys. Pollock and other common groundfish could not be r...
| Published in: | Canadian Journal of Fisheries and Aquatic Sciences |
|---|---|
| Main Author: | |
| Format: | Article in Journal/Newspaper |
| Language: | English |
| Published: |
Canadian Science Publishing
2020
|
| Subjects: | |
| Online Access: | http://dx.doi.org/10.1139/cjfas-2019-0152 http://www.nrcresearchpress.com/doi/full-xml/10.1139/cjfas-2019-0152 http://www.nrcresearchpress.com/doi/pdf/10.1139/cjfas-2019-0152 |
| Summary: | An acoustic-based index of biomass was developed due to concerns that the existing bottom-trawl index for western component pollock (Pollachius virens) was inaccurate and imprecise. Acoustic data were recorded during annual summer ecosystem surveys. Pollock and other common groundfish could not be readily identified from echo-traces within echograms. Survey catch data were analyzed concurrently with acoustic data to estimate biomass from echo-integrals contributed by the six most commonly caught fish species: pollock, Atlantic cod (Gadus morhua), haddock (Melanogrammus aeglefinus), redfish (Sebastes fasciatus), Atlantic herring (Clupea harengus), and spiny dogfish (Squalus acanthias). A cluster analyses that used a Bayesian conditional probability classified homogeneous regions based on species composition. Species composition, nautical area scatter coefficient, and backscattering cross-section of fishes for regions were used to calculate the biomass of fishes. Acoustic indices were considered as appropriate as bottom-trawl indices for fisheries management advice because the bias and variance were similar between the indices when analyzed using repeated K-fold cross-validation. |
|---|