Decadal variability and predictability of Senegalo-mauritanian upwelling system

Coastal countries in West Africa heavily rely on the ocean, which is a major source of food and employment. This is mainly due to the presence of coastal upwelling, upward motion of sea water bringing the nutrient-rich deeper waters into the illuminated surface layers in the coastal zone, where they become available for photosynthesis. The resulting phytoplankton production, the base of the food chain, render coastal upwelling the most productive of large marine ecosystems in the world’s oceans. Recently, decadal variability and predictability of coastal upwelling systems has received a lot of attention, since near-term changes of upwelling systems could have a strong impact of living marine resources and hence surrounding countries economy. On this aspect, recent progress has been made in generating near-term (“decadal”) predictions of physical using Earth system models (ESMs). Initialized forecasts have shown significant predictability from 1 to 10 years in advance for climate events showing substantial decadal variability.

Our objective here is two-fold: first we investigate the decadal variability of the Senegalo-mauritanian upwelling system (SMUS) in the reanalysis and historical simulations from eleven climate models using indices based on the SST and wind stress and also identify the processes controlling this variability. Second, we exploit the decadal prediction experiments of CMIP6 (DCPP-A), to investigate this upwelling predictability. Our results show that the SMUS is characterized by a strong decadal variability, in part linked to the Atlantic Multidecadal Variability Consequently, the DCPP- A experiment shows strong and generally significant correlation prediction scores at various lead times for the dynamical indices (Ekman transport and Ekman pumping). However, even though coastal SST are also significantly predictable, non-significant ACC scores are found for the thermal upwelling indices. The analysis concludes on trying to qualify and quantify the predictable components of the SMUS and possible applications.