On the Influence of Wind Forcing and Resolution on the Modelled Warm Bias of the Benguela Upwelling System

Most ocean and climate models exhibit a warm sea surface temperature (SST) bias in upwelling areas. This bias is especially prominent in the Benguela upwelling system, which is the major eastern boundary upwelling system in the south-east Atlantic Ocean along the African coast. We analyze the effects of different wind forcings and resolutions on the warm bias in order to improve the models. We are using a global 1/2° ocean-only model with a 1/10° nest around Africa. This resolution turns out to be necessary and sufficient to resolve the coastal upwelling. The model is forced by different satellite wind products with high temporal and spatial resolution, mostly based on QuikSCAT. As the theory predicts, the upwelling is directly proportional to the wind stress and therefore the wind forcing is crucial for modeling the upwelling. Other parameters turn out to be almost irrelevant. Increasing the model resolution by inserting a second nest with 1/30° horizontal resolution in the Benguela upwelling system did not show any significant effect on the amount of upwelling and the warm bias. The same holds for increasing the number of vertical levels from 46 to 76, and thereby doubling the number of levels in the upper ocean. Moreover, also the resolution of the wind forcing, which was tested by downscaling the wind data, and the temporal resolution of the wind are less important for the upwelling. Consequently, mainly the alongshore wind and its gradient are relevant for modeling the temperature in an eastern boundary upwelling system, but the usage of realistic satellite winds mainly improves the SST close to the coast. The offshore SST warm bias still remains, probably due to other reasons than the wind.