Séminaire
Séminaire SAMA (Groupe Statistiques pour l’Analyse, la Modélisation et l’Assimilation).
Stochastic parameterizations have been successfully used to represent the uncertainty associated with unresolved scale processes for ensemble forecasting and data assimilation systems. In order to accurately describe the uncertainty associated to the dynamical model and data assimilation system, stochastic parameters have to be optimized. Such parameters are related to the stochastic perturbations amplitude and their spatial covariance structure. A novel technique based on hierarchical ensemble Kalman filters is introduced, aiming to infer these type of parameters. The technique is proposed to be applied offline as part of an a priori optimization of the data assimilation system and could in principle be extended to the estimation of other hyperparameters of a data assimilation system.
bruno.deremble@ens.fr