Model error characterization for data assimilation in a regional ocean model of the Bay of Biscay

• Main Author: Broquet, Grégoire
• Other Authors: Laboratoire des Écoulements Géophysiques et Industriels Grenoble (LEGI), Université Joseph Fourier - Grenoble 1 (UJF)-Institut polytechnique de Grenoble - Grenoble Institute of Technology (Grenoble INP )-Centre National de la Recherche Scientifique (CNRS), Université Joseph-Fourier - Grenoble I, Jacques VERRON([e-mail])
• Format: Doctoral or Postdoctoral Thesis
• Language: French
• Published: HAL CCSD 2007
• Subjects: physical oceanography; data assimilation; régional modeling; model error; atmospheric forcings; open boundary conditions; Bay of Biscay; océanographie physique; assimilation de données; modélisation régionale; erreurs de modélisation; forçages atmosphériques; conditions aux frontières ouvertes; Golfe de Gascogne; [SDU.OCEAN]Sciences of the Universe [physics]/Ocean; Atmosphere; North Atlantic
• Online Access: https://theses.hal.science/tel-00143004; https://theses.hal.science/tel-00143004/document; https://theses.hal.science/tel-00143004/file/Thesegbroquet.pdf

A data assimilation system for ocean models, the SEEK (Singular Evolutive Extended Kalman) filter, is studied to control a Bay of Biscay configuration. This 1/15° configuration, nested in a 1/3° North Atlantic configuration, through the use of Open (sea) Boundaries Conditions, is developed using HYCOM (Hybrid Coordinate Ocean Model). This study focuses on the parametrization of the model error in the SEEK filter, and more generally in low rank Kalman filters, in order to control regional models. Classic parametrizations of these data assimilation systems, which have been developed initially for basin models, are not adapted to the regional dynamics complexity. Ensemble methods are used to get a realistic estimation of the model error due to bad determination of atmospheric and open boundary forcings. These forcings influence is supposed to be very important on regional dynamics. Model error statistics are characterized using the method of representers, which demonstrates the impact of the assimilation of various type of observations to control the oceanic state. The propagation of the error generated at open boundaries is weak. The use of the error due to atmospheric forcings to parameterize the SEEK filter for surface temperature assimilation experiments gives good results. Their comparison with those given by a more classical parametrization shows the benefits of this study on model error. Cette thèse porte sur l'application du filtre SEEK (Singular Evolutive Extended Kalman filter), un système d'assimilation de données pour les modèles océaniques, au contrôle d'une configuration du Golfe de Gascogne. Cette configuration au 1/15°, emboîtée dans une configuration au 1/3° de l'Atlantique Nord à travers l'emploi de Conditions aux Frontières Ouvertes (en mer), est développée à l'aide du modèle HYCOM (Hybrid Coordinate Ocean Model) à coordonnée verticale hybride. L'étude porte essentiellement sur la paramétrisation de l'erreur modèle dans le filtre SEEK, et plus généralement dans les filtres de Kalman de rangs ...