Impact of assimilating satellite surface velocity observations in the Mercator Ocean International analysis and forecasting global 1/4°system Provisionally Accepted

Isabelle Mirouze^1* Elisabeth REMY² Jean-Michel Lellouche² Matthew J. Martin³ Craig J. Donlon⁴

• ¹Independent researcher, France
• ²Mercator Ocean (France), France
• ³Met Office, United Kingdom
• ⁴European Space Research and Technology Centre (ESTEC), Netherlands

Representing and forecasting global ocean velocities is challenging. Velocity observations are scarce and sparse, and are rarely assimilated in a global ocean configuration. Recently, different satellite mission candidates have been proposed to provide surface velocity measurements. To assess the impact of assimilating such data, Observing System Simulation Experiments (OSSEs) have been run in the Mercator Ocean International analysis and forecasting global 1/4°system.Results show that assimilating simulated satellite surface velocities in addition to classical observations has a positive impact on the predicted currents at the surface and below to some extent. Compared to an experiment that assimilates only the classical observations, the surface velocity root-mean-squared error (RMSE) is reduced, especially in the Tropics. From a certain depth depending on the region (e.g. 200 m in the Tropics) however, slight degradations can be spotted. Temperature and salinity RMSEs are generally slightly degraded except in the Tropics where there is a small improvement at the surface and sub-surface. Sea surface height results are mixed, with some areas having reduced RMSE and some increased. The OSSEs reported here constitute a first study and aim to provide first insights on the features that improve by assimilating surface velocity data, and those which need to be worked on.