Abstract
Size selectivity of a new sorting section combining a sorting grid and a square mesh panel was tested for haddock (Melanogrammus aeglefinus) in the Barents Sea demersal trawl fishery. Sampling data for a wide size range enabled investigating the selection process for this species in detail, both for the grid and the square mesh panel. Contrary to earlier studies modelling size selectivity for grids and square mesh panels, which assume that the escape behaviour of all sizes of fish is equal, we applied a model that accounted for that haddock of different sizes can show different escape behaviour. Our results demonstrated that this model could describe the experimental data collected better than existing models. Specifically, our results showed that the likelihood for smaller haddock to seek escape through the grid and the square mesh panel was higher than for bigger haddock that still would manage to escape through the devices if they attempted. The new modelling approach presented in this study may be applicable to other species, selection devices and fisheries.