De l'utilité des observations de vitesse de surface pour l'assimilation de données dans les systèmes opérationnels de prévision océanique

This article shows that new satellite sea surface velocities are a promising complementary dataset to satellite altimetry. The field experiment stands on performing Observing System Simulation Experiments (OSSE) following [1] that share a "simulated" observational assimilation dataset. The targeted measurement error on SAR-derived surface velocity should not exceed 7 cm/s to have a comparable weight than the one of the conventional altimetry observing system in the assimilation scheme and thus to have a measurable impact on the analysis and forecast fields of a 1/3° Atlantic basin operational system. An updated release of the assimilation scheme was developed to add new control variable as the two (u,v) component velocity or radial velocity (in a privileged direction) at any depth. This article draws the conclusion from a comparison of a twin experiment simulation performed with real data only that the updated scheme shows better performance than the conventional one on the assimilation of sea surface height in comparison with independent satellite altimetry records without the actual assimilation of velocity data.This outstanding improvement is related to the new ageostrophic correction and the increase on the cross-shelf velocity component that is a first step for improving the exchanges between coastal regions and open ocean with that system.