Development of a length-structured and spatialised stock assessment model for the European hake (Merluccius merluccius)

• Main Author: Drouineau, Hilaire
• Other Authors: Écologie et Modèles pour l'Halieutique (IFREMER EMH), Institut Français de Recherche pour l'Exploitation de la Mer - Atlantique (IFREMER Atlantique), Institut Français de Recherche pour l'Exploitation de la Mer (IFREMER)-Institut Français de Recherche pour l'Exploitation de la Mer (IFREMER), Ecole Nationale d'Agronomie de Rennes, Dominique Pelletier
• Format: Doctoral or Postdoctoral Thesis
• Language: French
• Published: HAL CCSD 2008
• Subjects: Merluccius merluccius; [SDV.SA.STP]Life Sciences [q-bio]/Agricultural sciences/Sciences and technics of fishery; [SDV.BA.ZV]Life Sciences [q-bio]/Animal biology/Vertebrate Zoology; Hake; North East Atlantic
• Online Access: https://hal.archives-ouvertes.fr/tel-03216330; https://hal.archives-ouvertes.fr/tel-03216330/document; https://hal.archives-ouvertes.fr/tel-03216330/file/Drouineau_2008_D%C3%A9veloppement%20et%20ajustement%20d%27un%20mod%C3%A8le%20de%20dynamique%20des%20populations%20structur%C3%A9.pdf

Although the North-East Atlantic stock of European hake has been widely studied, certain characteristics of its biology are still relatively poorly known. This is particularly true of its seasonal migrations and growth, which a recent tagging campaign shows was probably largely underestimated. These uncertainties have several consequences both for the quality of the stock assessment using the usual population dynamics models and for management, particularly by limiting the possibilities of evaluating spatial management measures. To improve the estimation of the parameters of these processes, we have developed an integrated spatialized and length-structured population dynamics model based on a state model describing the dynamics of the population and the fishing activity, and on a model describing the observation processes, allowing the writing of a model of the population dynamics, allowing the writing of a likelihood function.A literature review of existing length-structured models highlighted the key role of growth process modelling in this type of model. To ensure the robustness of the model assumptions related to the discretisation of this continuous process and to individual growth variability, we performed a sensitivity analysis of the growth model. This method of numerical exploration of the model is based on (i) the development of experimental designs and (ii) the fitting of statistical models to the model outputs to quantify the impact of these assumptions. This generic approach is transposable to any sensitivity analysis of a discrete model describing a continuous process. In our case, it showed that for North-East Atlantic hake it was preferable to choose a quarterly time step, 1 cm classes, a gamma distribution of growth increments and a uniform distribution of individuals within the classes. During the development of the model, particular attention was paid to writing the likelihood function so that it would be robust and consistent with the observation processes. Different algorithms were also ...