Estimating abundance indices from the international 0-group fish survey in the Barents Sea

Annual abundance and distribution of 0-group fish in the Barents Sea have been recorded since 1965. Concern has been raised about the methods presently used to establish the abundance indices and about the catching efficiency of the trawl for the smaller-sized fish. The data have been reviewed for the period 1980-2002 and new abundance indices and length distributions of northeast Arctic cod (Gadus morhua), northeast Arctic haddock (Melanogrammus aeglefinus), capelin (Mallotus villosus), Norwegian spring spawning herring (Clupea harengus), and redfish (Sebastes spp.) were estimated. The abundance indices were estimated by two different statistical techniques, the method of stratified sample mean and a method based on the lognormal theory. The latter method was concluded to be the preferred one for this particular survey. The poor catching efficiency of smaller cod and haddock was corrected for and the results showed that length dependent selection contributes to a serious bias in the estimates when not corrected for, and it is likely that selection will bias the estimates for the other species as well. It is recommended that the technique based on the lognormal theory and with length corrections becomes the new standard method for estimating abundance indices from the 0-group survey in the Barents Sea.