Résumé
The Bay of Biscay is a highly productive ecosystem submitted to an intense exploitation. Concerns on mixed fishery issues in the context of catch limitations and of the multiannual mixed fisheries management plan encourages the development of new operational models capable of making projections at seasonal and fleet scales for this region. This is challenging because it implies the development of models that are both complex (to tackle ecosystem complexity) and reliable (to efficiently support scientific advice). In order to meet this second requirement, model evaluation has been recognised as an essential process to improve quality and reliability of scientific advice for ecosystem-based fisheries management. In this study, we present the calibration of a new ISIS-Fish simulation model for the Bay of Biscay demersal mixed fishery, built on recent biological knowledge on seven exploited species’ biology and a detailed description of the fishing activity, including the smallest vessels (less than 12 meters length). A thorough evaluation process was carried out in which model estimates were quantitatively compared to reference data at species and fleet scales. The quality of prediction was found dependent on species, fleets and variables. The model was particularly good at predicting landings profiles by métier and season for most species and for the most important fleets in total landings. The main utility of the evaluation process is to make the model’s strength and weaknesses transparent, in order to guide its future use for decision support.