On the predictability of subsurface sound speed from satellite-measured sea-surface temperature in the Mediterranean
The report investigates the extent to which subsurface sound speed may be determined from satellitemeasuredsea-surface temperature (SST). Using CTD data from four hydrographic surveys in the Mediterranean, it is shown that the vertical sound-speed profiles can be decomposed into Empirical Orthogonal Eigenfunctions (EOF). with the first three functions accounting for at least 88% of the variance. Significant correlation between EOF amplitudes and SST has been found for only one of the three EOFs, either the 1st EOF or the 2nd EOF. In order to determine subsurface sound-speed fromSST, this EOF has to represent important features of the sound-speed variability. This is the case for only one of the four surveys. Failures for the other surveys are due to a combination of low spatialvariability of SST, not visible in the satellite-retrieved values, and complex vertical sound-speedprofiles, which may not be represented by one EOF only.