Satellite altimetric detection of Somali eddies by signal processing technique

Abstract

Signal processing technique for detecting oceanic mesoscale eddies using dynamic topography data derived by satellite altimeters is described. This technique can be applied to single tracks of a satellite altimeter and consists of 3 steps. Firstly, the data time series is treated as an autoregressive sequence of finite order. The order best supported by the data, and the related autoregressive parameters are found by the modified covariance technique coupled with Cholesky decomposition. An analytical noise power spectral density which is used in the second step, in conjunction with a generic height signature due to an anticyclonic eddy, to design a matched filter is obtained by this method. Finally, the data are processed by the filter and the output is compared with a predetermined threshold to detect the presence (or, absence) of an eddy signature. This technique has been used to process several Geosat tracks near the Somali coast region of the Arabian Sea during the monsoon months of 1988. A few examples presented in the paper show that the technique is able to detect the great whirl, the southern gyre, the Socotra eddy and a few others. This approach alongwith sea surface temperature measurements by satellite radiometers can be a powerful tool for detecting oceanic mesoscale features.