Upwelling and coastal current biases in climate models

The upwelling regions at the eastern edges of the world's subtropical oceans exhibit pronounced SST biases in leading global coupled climate models. Reducing the SST and upwelling biases is important to climate variability and predictability, as well as projections of how coastal ecosystems respond to changing climate, including fishery and other impacts. In the southeastern tropical Atlantic, the wind-stress curl associated with the Benguela atmospheric low-level coastal jet (LLCJ) plays a major role in determining the coastal circulation and the spatial distribution of SST. Realistic representation of the detailed structure of the Benguela LLCJ, including both the double core structure and coastal wind drop-off, is critical to reducing the coastal ocean bias in the southeastern tropical Atlantic and is highly dependent on model resolution. In recent experiments with coupled models, it is found that increasing atmospheric model resolution substantially improves the LLCJ core structure, whilst in order to get a realistic upwelling system, a high-resolution ocean model is also required. Further, coupled feedbacks between SST and wind and cloud are important. Hence, attempting to derive parameterizations to reduce the bias remains a significant challenge in low or standard resolution climate models, as it involves deficiencies in the atmosphere and ocean, and estimation of the feedbacks. However, on the bright side, there is substantial theoretical, numerical, and observational work that can be brought to address the problem. The presentation will hopefully motivate discussion of parameterizations in the light of this work.