11.2.1 Key Processes

The bulk of the African continent is tropical or subtropical with the central phenomenon being the seasonal migration of the tropical rain belts. Small shifts in the position of these rain belts result in large local changes in rainfall. There are also regions on the northern and southern boundaries of the continent with winter rainfall regimes governed by the passage of mid-latitude fronts, which are therefore sensitive to a poleward displacement of the storm tracks. This is evident from the correlation between South African rainfall and the Southern Annular Mode (Reason and Rouault, 2005) and between North African rainfall and the NAO (Lamb and Peppler, 1987). Troughs penetrating into the tropics from mid-latitudes also influence warm season rainfall, especially in southern Africa, and can contribute to a sensitivity of warm season rains to a displacement of the circulation (Todd and Washington, 1999). Any change in tropical cyclone distribution and intensity will affect the southeast coastal regions, including Madagascar (Reason and Keibel, 2004).

The factors that determine the southern boundary of the Sahara and rainfall in the Sahel have attracted special interest because of the extended drought experienced by this region in the 1970s and 1980s. The field has moved steadily away from explanations for rainfall variations in this region as primarily due to land use changes and towards explanations based on changes in sea surface temperatures (SSTs). The early SST perturbation Atmospheric GCM (AGCM) experiments (Palmer, 1986; Rowell, et al., 1995) are reinforced by the results from the most recent models (Giannini et al., 2003; Lu and Delworth, 2005; Hoerling et al., 2006). The north-south inter-hemispheric gradient, with colder NH oceans conducive to an equatorward shift and/or a reduction in Sahel rainfall, is important. This has created interest in the possibility that aerosol cooling localised in the NH could dry the Sahel (Rotstayn and Lohmann, 2002; see also Section 9.5.4.3.1). However, temperatures over other oceanic regions, including the Mediterranean (Rowell, 2003), are also important.

In southern Africa, changing SSTs are also thought to be more important than changing land use patterns in controlling warm season rainfall variability and trends. Evidence has been presented for strong links with Indian Ocean temperatures (Hoerling et al., 2006). The warming of the troposphere over South Africa, possibly a consequence of warming of the Indo-Pacific, has been linked with the increase in days with stable inversion layers over southern Africa (Freiman and Tyson, 2000; Tadross et al., 2005a, 2006) in the late 20th century.

In addition to the importance of ocean temperatures, vegetation patterns help shape the climatic zones throughout much of Africa (e.g., Wang and Eltahir, 2000; Maynard and Royer, 2004a; Paeth and Henre, 2004; see also Section 11.7, Box 11.4). In the past, land surface changes have primarily acted as feedbacks generated by the underlying response to SST anomalies, and vegetation changes are thought to provide a positive feedback to climate change. The plausibility of this positive feedback is enhanced by recent work suggesting that land surface feedbacks may also play an important role in both intra-seasonal variability and rainy season onset in southern Africa (New et al., 2003; Anyah and Semazzi, 2004; Tadross et al., 2005a,b).

The MMD models prescribe vegetation cover; they would likely respond more strongly to large-scale forcing if they predicted vegetation, especially in semi-arid areas. The possibility of multiple stable modes of African climate due to vegetation-climate interactions has been raised, especially in the context of discussions of the very wet Sahara during the mid-Holocene 6 to 8 ka (Claussen et al., 1999; Foley et al., 2003). One implication is that centennial time-scale feedbacks associated with vegetation patterns may have the potential to make climate changes over Africa less reversible.