**Luke Phillipson \* and Ralf Toumi**

Space and Atmospheric Physics Group, Department of Physics, Imperial College London, London SW7 2AZ, UK; r.toumi@imperial.ac.uk

**\*** Correspondence: l.phillipson14@imperial.ac.uk

Received:13 November 2019; Accepted:15 December 2019; Published: 18 December 2019 -

**Abstract:** Satellite salinity data from the Soil Moisture and Ocean Salinity (SMOS) mission was recently enhanced, increasing the spatial extent near the coast that eluded earlier versions. In a pilot attempt we assimilate this data into a coastal ocean model (ROMS) using variational assimilation and, for the first time, investigate the impact on the simulation of a major river plume (the Congo River). Four experiments were undertaken consisting of a control (without data assimilation) and the assimilation of either sea surface height (SSH), SMOS and the combination of both, SMOS SSH. Several metrics specific to the plume were utilised, including the area of the plume, distance to the centre of mass, orientation and average salinity. The assimilation of SMOS and combined SMOS SSH consistently produced the best results in the plume analysis. Argo float salinity profiles provided independent verification of the forecast. The SMOS or SMOS SSH forecast produced the closest agreement for Argo profiles over the whole domain (outside and inside the plume) for three of four months analysed, improving over the control and a persistence baseline. The number of samples of Argo floats determined to be inside the plume were limited. Nevertheless, for the limited plume-detected floats the largest improvements were found for the SMOS or SMOS SSH forecast for two of the four months.

**Keywords:** SMOS; data assimilation; 4D-Var; Congo River plume; satellite salinity; Angola Basin; ROMS
