Temperature and precipitation trends are reviewed in Annex A of this report. Rainfall trends-especially over the past 30 years or so-have had a very large bearing on socioeconomic development of the continent because most activities are agriculturally based (Serageldin, 1995).
Uncertainties in GCMs make deriving regional climate change predictions impossible (see Annex B of this report for an assessment of regional projections of climate change). Therefore, it is important to interpret model outputs in the context of their uncertainties and to consider them as potential scenarios of change for use in sensitivity and vulnerability studies. In IPCC (1990), IPCC (1996), and Hernes et al. (1995), the Sahel (10-20�N, 20-40�E) was selected as a study region for purposes of inter-comparing GCM outputs. Since then, several approaches to subregional climate projection have been developed at national levels (e.g., Joubert, 1995; Ringius et al., 1996); several others have been developed under the U.S. Country Studies Program (USCSP, 1996).
Although most initial climate change simulations used GCMs, an increasing number
of climate-modeling centers have used regional climate models (RCMs). RCMs rely
on similar physical representations of atmospheric processes as GCMs but operate
at a much finer spatial resolution-typically 50 km-over limited domains. Little
climate change work using RCMs nested within GCMs has been completed as yet
for Africa (Ringius et al., 1996), so it remains necessary to rely on extracting
regional results for Africa from GCM climate change experiments. A selection
of such results is summarized in Box 2-3.
With respect to temperature, land areas may warm by 2050 by as much as 1.6�C over the Sahara and semi-arid parts of southern Africa (Hernes et al., 1995; Ringius et al., 1996). Equatorial countries (Cameroon, Uganda, and Kenya) might be about 1.4�C warmer. This projection represents a rate of warming to 2050 of about 0.2�C per decade. Sea-surface temperatures in the open tropical oceans surrounding Africa will rise by less than the global average (i.e., only about 0.6-0.8�C); the coastal regions of the continent therefore will warm more slowly than the continental interior.
Rainfall changes projected by most GCMs are relatively modest, at least in relation to present-day rainfall variability. In general, rainfall is projected to increase over the continent-the exceptions being southern Africa and parts of the Horn of Africa; here, rainfall is projected to decline by 2050 by about 10%. Seasonal changes in rainfall are not expected to be large; Joubert and Tyson (1996) found no evidence for a change in rainfall seasonality among a selection of mixed-layer and fully coupled GCMs. Hewitson and Crane (1996) found evidence for slightly extended later summer season rainfall over eastern South Africa (though nowhere else), based on a single mixed-layer model prediction. Great uncertainty exists, however, in relation to regional-scale rainfall changes simulated by GCMs (Joubert and Hewitson, 1997). Parts of the Sahel could experience rainfall increases of as much as 15% over the 1961-90 average. Equatorial Africa could experience a small (5%) increase in rainfall. These rainfall results are not consistent: Different climate models, or different simulations with the same model, yield different patterns. The problem involves determining the character of the climate change signal on African rainfall against a background of large natural variability compounded by the use of imperfect climate models.
Projected temperature increases are likely to lead to increased open water and soil/plant evaporation. Exactly how large this increased evaporative loss will be would depend on factors such as physiological changes in plant biology, atmospheric circulation, and land-use patterns. As a rough estimate, potential evapotranspiration over Africa is projected to increase by 5-10% by 2050. Little can be said yet about changes in climate variability or extreme events in Africa. Rainfall may well become more intense, but whether there will be more tropical cyclones or a changed frequency of El Ni�o events remains largely in the realm of speculation.
Changes in sea level and climate in Africa might be expected by the year 2050. Hernes et al. (1995) project a sea-level rise of about 25 cm. There will be subregional and local differences around the coast of Africa in this average sea-level rise-depending on ocean currents, atmospheric pressure, and natural land movements-but 25 cm by 2050 is a generally accepted figure (Joubert and Tyson, 1996). For Africa south of the Equator, simulated changes in mean sea-level pressure produced by mixed-layer and fully coupled GCMs are small (~1 hPa)-smaller than present-day simulation errors calculated for both types of models (Joubert and Tyson, 1996). Observed sea-level pressure anomalies of the same magnitude as simulated changes are known to accompany major large-scale circulation adjustments associated with extended wet and dry spells over the subcontinent.
Other reports in this collection