Extrapolating oceanic signals from surface data to deeper layers: Application to different datasets

Different approaches can be followed to extract as much information as possible from sea surface and vertically integrated measurements, potentially allowing to infer vertical profiles from sea surface level measured by satellite, coupled to other remotely sensed data. Here, the direct analysis of the sole observations and of their covariances, demonstrates its ability to identify what is the information content of the data and how this can be extracted more efficiently. In particular, two recently proposed methods to project surface measurements on the vertical, i.e. Coupled Pattern Reconstruction (CPR) and multivariate Empirical Orthogonal Function Reconstruction (mEOF-R) have been applied to different datasets. The more relevant results of these analyses are presented.