An atmospheric general circulation model has been forced by observed global sea surface temperature data from ten individual years. The results support the idea that oceanic effects tend to dominate the forcing of both interannual and interdecadal variability of Sahel rainfall, with different local atmospheric mechanisms dominating in different years. Further experiments indicate that skilful predictions of July--September Sahel rainfall are possible using the persistence of June SST Anomalies. ¿ American Geophysical Union 1992