Abstract

Subseasonal variability of Sea Surface Temperature (SST) in the central Benguela upwelling system is investigated using TMI satellite-derived data over the period 1999-2009. Spatial patterns and time-frequency characteristics of subseasonal variability are documented based on Empirical Orthogonal Functions (EOF) decomposition and wavelet analysis. Despite the land contamination of the TMI satellite data within approximately 100 km off the coast the first EOF of SST anomalies allows characterizing the coastal upwelling variability at the subseasonal scale. Two regimes of variability are evidenced: a submonthly (2-30 days) regime with a dominant 11 days oscillation and a lower frequency intraseasonal (30-90 days) regime with a dominant 61 days oscillation. Both regimes are modulated, to a large extent, by the local surface wind stress and are consistent with Ekman dynamics. The seasonality of the relationship between wind stress and SST for submonthly (intraseasonal) regime is characterized by a marked semiannual (annual) cycle, which is explained in terms of the impact of seasonal change of the upper ocean stratification on the vertical advection process. The wind-driven SST subseasonal variability is shown to be associated with eastward propagating disturbances in the midlatitudes corresponding to a wave number 4. The results also suggest an important role of the Antarctic Oscillation in modulating the intraseasonal wind-driven SST variability. The characteristics of the equatorial intraseasonal Kelvin waves are documented in order to discuss possible impact of remote oceanic forcing on SST variability along the coast in the Benguela upwelling system.