An algorithm for retrieving vertical wind profiles from satellite-observed winds over the Indian ocean using complex EOF analysis

With an aim to exploit current satellite observations for determining vertical wind profiles, the authors have carried out a complex empirical orthogonal function (CEOF) analysis of a large number of radiosonde observations of wind fields over the Indian Ocean. This analysis suggests that the first two CEOFs explain more than 80% of the total variance. While the first principal component is highly correlated with the upper-level winds at 250 mb (r = 0.95), the second one is well correlated with the 800-mb winds (r = 0.82). This analysis leads to a retrieval algorithm that ensures the retrieval of vertical profiles of winds, using satellite-tracked cloud motion vector winds. Assuming that accurate measurements of wind are available at the above-mentioned levels, the rms error of retrieval for each component of wind is estimated to range between 2 and 6.5 m s(-1) at different levels, which is much lower than the natural variance of wind at these levels. To construct a better visualization of retrieval, the authors have provided retrieved and true wind profiles side by side for three typical synoptic conditions.