Climate Change 2001:
Working Group II: Impacts, Adaptation and Vulnerability
Other reports in this collection

10.2.6.3. Climatic Factors in Desertification

Precipitation and temperature determine the potential distribution of terrestrial vegetation and constitute principal factors in the genesis and evolution of soil. Extended droughts in certain arid lands have initiated or exacerbated desertification. In the past 25 years, the Sahel has experienced the most substantial and sustained decline in rainfall recorded anywhere in the world within the period of instrumental measurements (Hulme and Kelly, 1997). Linear regression of 1901-1990 rainfall data from 24 stations in the west African Sahel yields a negative slope amounting to a decline of 1.9 standard deviations in the period 1950-1985 (Nicholson and Palao, 1993). Since 1971, the average of all stations fell below the 89-year average and showed a persistent downward trend since 1951.

Because evapotranspiration constitutes the only local input to the hydrological cycle in areas without surface water, reduction in vegetative cover may lead to reduced precipitation, initiating a positive feedback cycle. Degradation of vegetation cover in moister areas south of the Sahel may have decreased continental evapotranspiration and reduced precipitation in the Sahel (Xue, 1997).

A positive feedback mechanism between vegetation cover and albedo may help to explain the Sahel drought (Charney, 1975). Some research supports an albedo-precipitation feedback mechanism (Otterman, 1974; Cunnington and Rowntree, 1986; Xue et al., 1990; Diedhiou and Mahfouf, 1996; Zheng and Eltahir, 1997; Zeng et al., 1999), although other research disputes the importance of albedo (Jackson and Idso, 1974; Ripley, 1976; Wendler and Eaton, 1983; Gornitz and NASA, 1985; Nicholson et al., 1998; Nicholson, 2000).

Degraded land also may increase atmospheric dust and aerosols, which influence precipitation (see Section 10.1.3.2).

SST anomalies, often related to ENSO or NAO, also contribute to rainfall variability in the Sahel (Lamb, 1978; Folland et al., 1986; Hulme and Kelly, 1997; Nicholson and Kim, 1997; Hulme et al., 1999). Lamb (1978) observes that droughts in west Africa correlate with warm SST in the tropical south Atlantic. Examining oceanographic and meteorological data from the period 1901-1985, Folland et al. (1986) found that persistent wet and dry periods in the Sahel were related to contrasting patterns of SST anomalies on a near-global scale. When northern hemisphere oceans were cold, rainfall in the Sahel was low.

Street-Perrott and Perrott (1990) demonstrate that injections of freshwater into the north Atlantic (such as from glacial melt) decrease salinity—stabilizing the water column, inhibiting deep convection, and reducing northern transport of heat by the Atlantic thermohaline circulation, which is driven by a north-south SST gradient. This decreases evaporation from the ocean surface, causing drought in the Sahel and Mexico. From 1982 to 1990, Mynemi et al. (1996) found a correlation between ENSO-cycle SST anomalies and vegetative production in Africa. They found that warmer eastern equatorial Pacific waters during ENSO episodes correlated with rainfall of <1,000 mm yr-1 over certain African regions.

A combination of factors—including vegetation cover, soil moisture, and SST—best explains the reduction in rainfall in the Sahel. Diedhiou and Mahfouf (1996) modeled changes in albedo, soil moisture, land surface roughness, and SST anomalies and calculated a rainfall deficit over the Sahel similar to observed patterns. Eltahir and Gong (1996) suggest that a meridional distribution of boundary-layer entropy regulates the dynamics of monsoon circulation over west Africa, explaining observed correlations of SST to rainfall and the sensitivity of monsoon circulation to land-cover changes. A coupled surface-atmosphere model indicates that—whether anthropogenic factors or changes in SST initiated the Sahel drought of 1968-1973—permanent loss of Sahel savanna vegetation would permit drought conditions to persist (Wang and Eltahir, 2000). Zeng et al. (1999) compared actual rainfall data from the period 1950-1998 with the output of a coupled atmosphere-land-vegetation model incorporating SST, soil moisture, and vegetative cover. Their results indicate that actual rainfall anomalies are only weakly correlated to SST by itself. Only when the model includes variations in vegetative cover and soil moisture does it come close to matching actual rainfall data. Modeling the importance of SST, sea ice, and vegetative cover to the abrupt desertification of the Sahara 4,000-6,000 years ago, Claussen et al. (1999) show that changes in vegetative cover best explain changes in temperature and precipitation.

10.2.6.4. Linkages and Feedbacks between Desertification and Climate

CO2-induced climate change might exacerbate desertification through alteration of spatial and temporal patterns in temperature, rainfall, solar insolation, and winds. Conversely, desertification aggravates CO2-induced climate change through the release of CO2 from cleared and dead vegetation and through the reduction of the carbon sequestration potential of desertified land.

Areas that experience reduced rainfall and increased temperature as a result of CO2-induced climate change also could experience declines in agricultural yields, livestock yields, and tree cover, placing local people at risk of famine.

Lower soil moisture and sparser vegetative cover also would leave soil more susceptible to wind erosion. Reduction of organic matter inputs and increased oxidation of soil organic matter (SOM) could reduce the long-term water-retention capacity of soil, exacerbating desertification. Sample plots in Niger lost 46 t ha-1 in just four windstorms in 1993 (Sterk et al., 1996), releasing 180 ± 80 kg ha-1 yr-1 of soil carbon (Buerkert et al., 1996). Moreover, increased wind erosion increases wind-blown mineral dust, which may increase absorption of radiation in the atmosphere (Nicholson and Kim, 1997).

Desertification from anthropogenic and climatic factors in Senegal caused a fall in standing-wood biomass of 26 kg C ha-1 yr-1 in the period 1956-1993, releasing carbon at the rate of 60 kg C cap-1 yr-1 (Gonzalez, 1997).

Although altered surface albedo may increase surface air temperatures locally (Williams and Balling, 1996), the effect of desertification on global mean temperature is unlikely to have exceeded 0.05°C in the past century (Hulme and Kelly, 1997).



Other reports in this collection