Home

Registration

Preparation of papers

Presenters

Programme

Contact

Pierre De Mey¹
Julien Lamouroux², Matthieu Le Henaff¹ and Baptiste Mourre^1,3

(1) CNRS-LEGOS, Toulouse, France
(2) Noveltis, Parc Technologique du Canal 2, Ramonville-Saint-Agne, France
(3) Institut de Ciiences del Mar, Barcelona, Spain

Main authors appear in bold

In order to assess the value of coastal and shelf seas models and the well-posedness of their configuration (including boundary specification using larger-scale models), observation networks can be set up. It seems natural to plan the new observations in areas and on variables for which model errors are expected to be largest (if they are made permanent, such networks can subsequently be used to feed data assimilation). Our objectives in this talk are twofold:
(1) review statistical criteria, using model error estimates, to assess the relative value of observational network options;
(2) illustrate the power of ensemble methods (stochastic modelling, Ensemble Kalman filter) for providing suitable model (forecast) error estimates and objective criteria.
Examples in the adjacent seas of France will be given.

Oral presentation

Back