Catches of Atlantic cod (Gadus morhua) from Greenland gill-net surveys were analyzed by a two-step approach. In the initial step the form of the selection curve was identified as binormal, which was caused by fish being gilled or caught by the maxillae. Both capture processes could be described by normal distributions and could be related to mesh size in accordance with the principle of geometrical similarity. In the second step the selection parameters were estimated by a nonlinear least squares fit. The model also estimated the relative efficiency of the two capture processes and the fishing powers of different netting sections. Gilling was about three times as efficient as maxillae catching and fishing power could be related to the ratio of twine diameter to mesh size. It is proposed that information on how fish are caught should be included when modelling gill-net selectivity, as lack of such data may lead to the formulation of unrealistic selection models.
|Journal||Canadian Journal of Fisheries and Aquatic Sciences|
|Publication status||Published - 1996|