Study of the impact of satellite data assimilation into a hydrodynamic model of the Ligurian Sea. Comparison between SST fields and SST satellite-based predicted fields

The verification of a 3D hydrodynamic model of the Ligurian Sea is presented. Two assimilation experiments have been carried out with this model: the assimilation of real SST, and the assimilation of SST forecasted by a statistical predictor. The aim of the study is to establish the skill of the model in these two configurations. The assimilation of predicted SST can help to increase the model skill when observations are not available, and preliminary results show that both approaches obtained similar results. The verification is done in a multi-scale approach, by decomposing the model results and the observations into several spatial scales, using 2D discrete wavelet transforms. At each scale the error between the model and the observations is calculated, and the scales where the biggest errors occur can be identified. The variability distribution of the model and the observations is also examined at each scale, to study the impact of the assimilation on the model variability. This methodology provides a scale-dependent insight in the study of the assimilation of SST and predicted SST. The differences between both assimilated data and how these differences affect the model results are examined.