Assessment of the Decadal Prediction Skill of Sahel Rainfall in CMIP5 and CMIP6

Accurately predicting the decadal variations in Sahel rainfall has important implications for the lives and economy in the Sahel. Previous studies found that the decadal variations in sea surface temperature (SST) in the Atlantic, Mediterranean Sea, Indian Ocean, and Pacific contribute to those in Sahel rainfall. This study evaluates the decadal prediction skills of Sahel rainfall from all the available hindcasts contributing to phases 5 and 6 of the Coupled Model Intercomparison Project (CMIP5 and CMIP6), in comparison with the related uninitialized simulations. A majority of the prediction systems show high skill with regard to Sahel rainfall. The high skill may be partly attributed to external forcings, which are reflected in good performance of the respective uninitialized simulations. The decadal prediction skills of the key SST drivers and their relationships with the Sahel rainfall are also assessed. Both the hindcasts and the uninitialized simulations generally present high skill for the Atlantic multidecadal variability (AMV) and Mediterranean Sea SST indices and low skill for the Indian Ocean basin mode (IOBM) and interdecadal Pacific variability (IPV) indices. The relationship between the Sahel rainfall and the AMV or Mediterranean Sea SST index is reasonably captured by most prediction systems and their uninitialized simulations, while that between the Sahel rainfall and the IOBM or IPV index is captured by only a few systems and their uninitialized simulations. The high skill of the AMV and Mediterranean Sea SST indices as well as the reasonable representations of their relationships with the Sahel rainfall by both the hindcasts and uninitialized simulations probably plays an important role in predicting the Sahel rainfall successfully.