Subsurface geostrophic velocities inference from altimeter data: Application to the Sicily Channel (Mediterranean Sea)

Abstract

Coupled Pattern Reconstruction (CPR) and multivariate Empirical Orthogonal Function Reconstruction (mEOF‐R) techniques have been applied to infer the vertical structure of the sea from surface/integrated measurements in the Sicily Channel (Mediterranean Sea). The data used to train the methods are the CTD casts present in the MEDAR/MEDATLAS climatology. The independent test and validation data set consisted of Topex/Poseidon sea level measurements and CTD profiles collected during the SYnoptic Mesoscale PLankton EXperiment (SYMPLEX) 1996 and 1998 cruises. mEOF‐R was found to give the more accurate and promising results, being able to discriminate and reconstruct the main processes that drive the variability inside the channel. In particular, the analysis of the covariance of steric heights, temperature and salinity profiles confirmed that mesoscale processes are predominant. In fact, the first mEOF steric height pattern closely resembles the first baroclinic normal mode estimated from the mean stratification profile. A critical discussion on the signals contributing to the sea level variability has been done before applying the mEOF‐R to altimetric data. The steric contribution of the upper 300 dbar has then been estimated from Topex/Poseidon measurements. The effects of the adjustment of altimeter data with a synthetic mean dynamic topography have also been quantified.