What drives natural variability of African rainfall? Continent-wide analysis of decadal to multidecadal correlation patterns

Abstract:

African rainfall shows significant year-to-year natural fluctuations that in part are linked to teleconnections associated with modes of variability in the Atlantic, Pacific and Indian oceans. Several of these relations have already been documented in the literature, e.g. the link between Sahel rains and the Atlantic Multidecadal Oscillation (AMO) or the connection between Moroccan precipitation and the North Atlantic Oscillation (NAO). A better understanding of African rainfall variability and potential natural drivers would help to better prepare African societies for anticipated droughts and floods by taking early precautionary action. In this contribution we are presenting the first continent-wide analysis of African rainfall variability on a month-by-month and country-by-country basis. We have calculated Pearson r values for smoothed monthly rainfall data of 49 African countries over the period 1901-2017 which we compared to six potential climatic drivers of natural variability, namely AMO, NAO, ENSO (El Niño Southern Oscillation), Pacific Decadal Oscillation (PDO), Indian Ocean Dipole (IOD) and solar activity changes. In the search for the best correlations we allowed time lags of up to 11 months for each potential driver (120 months for solar activity). The Pearson coefficients were regionally mapped out across Africa separately for each of the 12 months of the year. The strongest of the identified relationships were schematically summarized on results maps that detail the regional extent, seasonal occurrence and intensity of the linkeage. The dynamic temporal-spatial evolution of the correlations was mapped out across the continent, tracking the gradual or abrupt expansion, displacement and subsequent waning of the various effects over the course of the year. Relationships are complicated by characteristic time lags, non-stationary correlations and occasional phase shifts, as evidenced by time series plots and literature reports. Our empirical results may help to further improve short- to midterm rainfall prognoses in Africa and and provide important calibration data for the further improvement of climate models.