Climatologists fear prolonged drought in Sahelian zone

Throughout human history, rainfall in sub-Saharan Africa has been variable from year to year and from place to place. There were recurrent droughts during the first 60 years of this century, but the rain has usually returned within one to four years. For this reason, agricultural planners have become confident, even in years of low rainfall, that "normal" rains would return within a few years.

Since 1960, however, rainfall in the Sahel has been declining. More precisely, the year-to -year variations in rainfall in any particular place have been superimposed on a continuing long-term decline. The same is true of overall rainfall in the western Sahel as reflected in the statistical indices constructed by climatologists. Through out this region rainfall has become sparser and more variable. To cite one dramatic example, rainfall in Dakar has declined from 500-800 mm in the 1950s (that typical of the mid western United States) to 130-200 mm in the 1970s (that typical of the American arid southwest). The effect of this long-term decline is exacerbated by a shorter-term drought that extends over a huge crescent-shaped area that stretches across the Sahel and doubles back down through southern and southeastern Africa, an area twice the size of the United States.

The historical precedents for this kind of prolonged Sahelian drought have made climatologists seriously concerned that the current drought will persist for another decade or so. The last period comparable to this one took place in the 1820s and 1830s. According to the rather meagre records available, rainfall declined for a period of 20 to 30 years and the level of Lake Chad fell to about where it is today. The geological record shows several similar drops in the level of the lake in the past 600 years. (The Sahel has also seen much wetter periods. The area near Timbuktu, which now has 100 mm of rainfall a year, exported wheat to other regions of West Africa in the late nineteenth century, taking advantage of then higher rainfall and river flow.)

Climatologists still do not know the extent to which the drought and its persistence are due to human activities. Any climatic phenomenon that extends over an entire continent and which has occurred intermittently in the past cannot be due to human activities alone, but must be related to the global atmospheric circulation that controls the climate of the entire earth.

Other areas of the tropics are affected by periodic droughts, but these last for one or, at most, two seasons. Many of these are known to be caused by the El Nino/Southern Oscillation (ENSO) phenomenon. A global change in the oceans and atmosphere that causes drought in India, Australia, South America, and the Pacific Islands. (See Ceres, Nov.-Dec. 1985).

The year-to-year variability caused by the ENSO phenomenon can be seen in African climatic records. Of the last 28 ENSO years, 22 have been unusually dry in southeast Africa. Since ENSO-related changes take place a few months before the rainy season, further research may make it possible to use them to give a few months' notice of some particularly bad drought years in southeast Africa, south of Zimbabwe. However, their usefulness for this purpose in the Sahel is doubtful.

In the Sahel, by contrast, some dry years are correlated with a warming of the surface of the Atlantic just north of the equator. This change may be related to the movements of the Intertropical Convergence Zone (ITCZ), the great band of equatorial clouds whose north-south migrations bring rain to the humid tropics and the Sahel.

But unlike the situation in other parts of the world, Africa has had many dry years that are not correlated with ENSO or other sea surface phenomena. And the trend line of its rainfall regime in the Sahel has gotten steadily worse for 20 years. It is the persistence of the Sahelian drought that sets it apart from droughts in other parts of the world.

While the cause of this persistence is not known, the obvious suspect is the large-scale changes in the land surface in Africa. Some of these changes are probably due to human activities, such as deforestation. But the most important changes are probably caused by the drought itself. In other words, there is probably a natural feedback mechanism that causes the drought to feed on itself.

Both human and natural changes involve major changes in surface hydrology, reflectance (albedo), and dust - all of which are major influences on climate. In particular, surface runoff increases, and percolation and soil moistture decrease, decreasing in turn the amount of water available for evapotransportation back into the atmosphere. In addition, crops and grasses evapotranspire less than trees because they have shallower roots. These changes can have a serious effect on local precipitation, since in many areas rainfall is derived from water evapotranspired locally. Another possible contribution comes from the coaling effect caused by the increase in reflectivity and dust. This draws dry air down from the upper atmosphere, reinforcing the effect of decreased soil moisture.

From the scientific point of view, the problem is, first, that no one knows the relative contribution of human and natural causes to the persistence of the drought. (To be sure, erosion and deforestation caused by humans seriously affect the availability of soil and water for agriculture, but this is another matter.) There is no empirical evidence that even large-scale changes in land surface (covering an area of, say, more than 500 km²) can affect climate. But they are not yet capable of modelling climate reliably at the continental level, or of modelling the details of surface hydrology or the relations between plant cover, soil, and atmosphere. Hence, scientists are unwilling to rely on their ability to predict the climatic effects of even large-scale changes in land surface in the absence of empirical data. These research techniques are now rapidly improving.

Rainfall data in Africa, both current and historical, is meagre. Local research capacity is almost non-existent. The network of operational meteorological stations is poor and deteriorating. Upper air measurements of wind and temperature, which are essential for research, are few and dwindling over the Sahel. Even routine rainfall observations are often not communicated to international networks or are subject to long delay. Historical records are difficult of access; one researcher reported that she had to visit the station in person and copy the records by hand in order to obtain satisfactory historical records. Satellite data help, but are still mostly qualitative and extend back at most ten years. Well-trained African meteorologists do exist, but they are employed mostly in routine forecasting and administrative duties.

Two conclusions can be drawn. First of all, rainfall atlases based on the historical records of the last three or four decades are clearly not applicable to the next decade or two. Specifically, agricultural and other projects now being planned for the Sahel should not use farm models based on the assumption that "normal rains will return". The causes of the drought are not understood, and there is no assurance that the rain will return soon. On the contrary, it is prudent to assume that the pattern of the last decade may persist for another decade or more.

Charles Weiss