Abstract/Summary

In Equatorial Africa, millions of people rely on rain-fed agriculture as their only source of livelihood and are vulnerable to varying precipitation patterns, especially extreme precipitation. This thesis investigates, using observations, reanalysis, a state-of-the-art Africa-wide convection permitting climate model (CP4A) and its corresponding global simulation (G25), the precipitation relationship between Western Equatorial Africa (WEA) and Eastern Equatorial Africa (EEA) and the associated physical processes. CP4A and G25 are evaluated against observations and ERA-Interim, with a focus on precipitation and Kelvin wave activity. Lead/lag correlation and spatio-temporal correlation patterns over various sub-regions reveal a synoptic-scale relationship in precipitation between WEA and EEA in which precipitation over EEA lags precipitation over WEA by 1–2 days. Composites on anomalous precipitation events and an equatorial wave dataset show an apparent connection between eastward/northeastward propagating anomalous precipitation and Kelvin wave low-level convergence, suggesting an influence of Convectively Coupled Kelvin Waves (CCKWs). Observations/reanalysis and simulations show that the two important processes through which CCKWs modulate the eastward propagation of convection and precipitation across Equatorial Africa are: 1) low-level westerly anomalies that lead to increased low-level convergence, and 2) westerly moisture flux anomalies that amplify the lower-to-mid-tropospheric specific humidity. CP4A and G25 generally simulate the key horizontal structure of CCKWs, with anomalous low-level westerlies in phase with positive precipitation anomalies. On days with an observed high-amplitude Kelvin wave, an extreme precipitation episode is up to twice as likely to occur compared to climatology, and precipitation intensity increases by up to 4 mm day-1, although extreme precipitation patterns associated with intense Kelvin waves vary somewhat from case study to case study. These results reveal the important processes that need to be satisfactorily represented in models, and they suggest that monitoring the propagation characteristics of CCKWs may be important in synoptic-timescale forecasting over Equatorial Africa.