Decadal variability and trends of the Benguela upwelling system as simulated in a high-resolution ocean simulation

Abstract. Detecting the atmospheric drivers of the Benguela upwelling systems is essential to understand its present variability and its past and future changes. We present a statistical analysis of a high-resolution (0.1°) ocean-only simulation driven by observed atmospheric fields over the last 60 years with the aim of identifying the large-scale atmospheric drivers of upwelling variability and trends. The simulation is found to reproduce well the seasonal cycle of upwelling intensity, with a maximum in the June–August season in North Benguela and in the December–February season in South Benguela. The statistical analysis of the interannual variability of upwelling focuses on its relationship to atmospheric variables (sea level pressure, 10 m wind, wind stress). The relationship between upwelling and the atmospheric variables differ somewhat in the two regions, but generally the correlation patterns reflect the common atmospheric pattern favouring upwelling: southerly wind/wind stress, strong subtropical anticyclone, and an ocean–land sea level pressure gradient. In addition, the statistical link between upwelling and large-scale climate variability modes was analysed. The El Niño–Southern Oscillation and the Antarctic Oscillation exert some influence on austral summer upwelling velocities in South Benguela. The decadal evolution and the long-term trends of simulated upwelling and of ocean-minus-land air pressure gradient do not agree with Bakun's hypothesis that anthropogenic climate change should generally intensify coastal upwelling.