SST-forced surface wind variability in the tropical Atlantic: An empirical model

[1] Simple dynamical models developed for the tropical Pacific surface winds may not be directly applicable in the Atlantic in view of the basin's different geometry and size and the potential impact of deep heating from over the adjacent African and South American continents. As simple dynamical models generate surface winds from sea-surface temperature (SST), it is worth ascertaining the extent to which monthly surface winds over the tropical (30degreesS-30degreesN) Atlantic are driven by this sector's own SST variations. A procedure is developed to separate this internally driven (i.e., by tropical Atlantic basin SSTs) component of the surface wind field from those induced by changes external to this tropical basin. Our analysis shows the internally driven component to dominate wind variability only within the equatorial (10degreesS-10degreesN) Atlantic. In contrast, Pacific surface winds driven by the tropical Pacific basin SSTs dominate variability across the tropical Pacific. The vertical structure of atmospheric diabatic heating linked with the leading modes of internally driven variability in the two basins (e. g., Pacific El Nino and the Atlantic Nino) is compared to assess the suitability of hypotheses underlying simple dynamical models. The potential limitations of these models, particularly, in the tropical Atlantic, motivated the development of an empirical model for Atlantic surface-wind variability. The model is based on rotated principal component analysis (RPCA) of the internally driven component of interannual variability in the tropical Atlantic basin. The empirical model is constructed from RPCA analysis of combined SST and surface wind anomalies of individual calendar months (from COADS data set) and outperforms the simple dynamical models. The model's performance is also compared to an empirical model based on the singular value decomposition analysis of SST and surface winds.