The chronicles of medieval geographers and historians, the journals of early European travelers, the archeological record, and other sources reveal that rainfall in the Sahel is generally low, unevenly distributed, and highly variable. Drought is an inherent feature of the region. These records indicate that major droughts, persisting from 12-15 years, evidently occurred in the 1680s, the mid-1700s, the 1820s and 1830s, the l910s, and since 1968. Generally arid conditions characterized the period from 1790-1850, and comparatively minor droughts apparently occurred in the 1640s, 1710s, 1810s, the beginning of the twentieth century, and the 1940s. Relatively wet periods occurred during the ninth through thirteenth centuries, the sixteenth through eighteenth centuries, from 1870-1895, and during the 1950s.
This chapter considers the interaction of climate and man in the Sahel by examining the rainfall characteristics to which human activities must be adapted. (A more detailed description of sub-Saharan climatic regimes appears in Appendix B.) The discussion attempts to underscore several points about the Sahel that are too rarely kept in mind. First, the region is part of a global climatic system, and it must be treated as such in any attempt to understand it. Second, it is an environment that can sustain only cautious use, as rainfall is low, spotty, and highly variable (Vermeer 1981); recurring droughts are inherent, and dry years are more prevalent than wet ones. Finally, the interplay of various parts of the climatic system are involved in producing the extreme rainfall fluctuations in the region; therefore, narrowly based attempts to understand and forecast the region's climate and to comprehend its relation to man and environment are generally unproductive. Throughout this chapter, remarks applicable to the true Sahel are appropriate for the entire sub-Saharan region from approximately latitude 10°N to latitude 20°N.
THE HISTORICAL PERIOD: 850-1900 A.D.
While it is apparent from modern meteorological records that the Sahel experiences marked changes of rainfall distribution on the scale of decades, other classes of data confirm this variability as an inherent characteristic of the region and illustrate that conditions differing significantly from the "average ones can persist over longer periods of time. It is, of course, more difficult to reconstruct the climate of earlier centuries, but historical accounts and geographical texts, together with various classes of environmental data, provide a fairly reliable picture of times before instrumental observations were made.
The types of information useful for a historical reconstruction of Sahelian climate are sketched in Table 1. By identifying independent indicators of the same event or trend, by comparing evidence from many areas, and by recording anomalous events, it is possible to determine a number of drier and wetter periods during the last three centuries and to broadly describe conditions during the last millenium.
Relatively little material on the Sahelian climate is available for periods prior to the sixteenth century, and what can be found is general and somewhat speculative. Nevertheless, the level of the groundwater table, the extent of lakes, archaeological studies, descriptions of landscapes and caravan routes from medieval Arabic sources, and climatic descriptions found in the journals of early European travelers all provide evidence of past conditions (Nicholson 1979). Tentative conclusions drawn from these sources are that the Sahel probably experienced wetter conditions in the ninth through thirteenth centuries, and that these conditions may have set in as early as the eighth century and declined sometime during the fourteenth century.
Several indicators suggest that the sixteenth through eighteenth centuries were generally wetter than subsequent centuries (Nicholson 1979). They include the level and extent of numerous lakes, historical accounts, and descriptions of landscape and climate. As indicated in Chapter 1, perhaps the best point of reference is the level of Lake Chad (Maley 1981), which often stood 4-5 m above modern levels, although within the present century fluctuations have been on the order of 1 or 2 m. Also, historical chronicles from Mauritania, Mali, Nigeria, and Chad indicate a relative absence of drought throughout the period (Nicholson 1979). Early geographical accounts tell of the verdure of many Sahelian regions, suggesting greater rainfall and higher groundwater levels. When such indicators are taken individually, the interpretation of each can often be challenged. Collectively, however, they present an accurate picture of the past and indicate that generally wetter conditions prevailed throughout much of the sixteenth, seventeenth, and eighteenth centuries.
Nevertheless, drought did occur (Cissoko 1968). A major drought apparently occurred in the 1680s. It also appears that droughts at least as severe as that of 1968 occurred in the Sahel from the mid-1730s to the mid-1750s and again in the 1820s and 1830s-persisting for 12-15 years in many areas. The latter drought was continental in extent. Lake Chad was partially desiccated, and signs of increasing aridity appeared throughout much of southern Africa. Generally dry conditions prevailed during most of the period from around 1790 to 1850. Less significant droughts occurred in the 1640s, 1710s, and 1810s (Nicholson 1980).
TABLE 1 Types of Data Useful for Historical Climatic Reconstructions
I. Landscape Descriptions
(1) Forests and vegetation. Were they as they are today?
(2) Conditions of lakes and rivers
(a) Height of the annual flood, month of maximum flow of the river
(b) Villages directly along lakeshores
(c) Size of the lake (e.g., as indicated on maps)
(d) Navigability of rivers
(e) Desiccation of present-day lakes or the appearance of lakes that no longer exist
(g) Seasonality of flow condition in wet and dry seasons
(3) Wells, oases, bogs in presently dry areas-also,drying up of wells
(4) Flow of wadis
(5) Measured height of lake surface (frequently given in travel journals, but optimally some instrumental calibration or standard should accompany this).
II. Drought and Related Information
(1) References to famine or drought, preferably accompanied by the
(a) Where occurred-as precisely as possible
(b) When occurred-as accurately as possible
(c) Who reported it-whether the information is secondhand
(e) Cause of famine
(f) Localized or widespread
(Note: Some sources mentioning the "very severe drought" may simply be referring to the normal dry season.)
(2) Agricultural prosperity
(a) Condition of harvest
(b) What produced this condition
(c) Months of harvests-in both bad years and good years
(d) What crops are grown
(3) Rainfed agriculture in regions presently too arid
III. Climate and Meteorology
(1) Measurements of temperature, rainfall, etc.
(2) Weather diaries
(3) Descriptions of climate and the rainy season. When do the rains occur, what winds prevail?
(4) References to occurrence of rain, tornadoes, storms
(5) Seasonality and frequency of tornadoes, storms
(6) Snowfall. Is this clearly snow or may the reporter be mistakenly reporting frost, etc.?
(7) Freezing temperatures, frost, hail
(8) Duration (or absence) of snow cover on mountains
(9) References to dry or wet years, severe or mild winters
(10) References to wind. Particularly important is the prevalence of the harmattan or a steady northeast wind in areas south of the Sahara (in West Africa) because this is an unambiguous sign of a dry period. Conversely, steady southwest winds (associated with the rainy season) are also important indicators in this region. Important:Does the reporter suggest this is a common or uncommon occurrence?
(Note: Even very isolated references may be quite important. Also, "tornado" generally refers to the frequent West African wind storms or squalls.)
SOURCE: Nicholson 1979.
Such climatic indicators as lake levels and river flow, rainfall, and harvests (Figure 4) indicate extreme fluctuations during the late nineteenth and early twentieth century. A wet episode, comparable to or more extreme than the 1950s (Kimble 1962), lasted from about 1870-1895 (Nicholson 1978, 1981). The harvests of the agriculturalists associated with the nomadic Kel Tamacheq (Tuareg) were abundant during this period; droughts were local and of short duration. The Niger Bend region near Timbuktu yielded abundant crops and became the "breadbasket" of West Africa, whereas current annual rainfall in the area averages 228 mm. The discharge of the Niger and Senegal rivers was higher during this period (1870-1895) than during the twentieth century; and lowland areas of Senegal, Mali, and southern Africa, which are now dry, contained lakes or ponds. The surface of Lake Chad stood several meters above its present level.
The picture is similar for other parts of Africa. Throughout East Africa, Rift Valley lakes maintained levels several meters above modern ones, and the Nile carried considerably more water than at present. In southern Africa, Lake Ngami, now an expansive marsh, was a deep and extensive body of water. Harvests were consistently good in semiarid regions of Namibia, southern Angola, and some areas of South Africa.
THE TWENTIETH CENTURY
Around 1895, conditions became markedly more arid throughout the tropics. A serious drought occurred at the turn of the century. Although Senegal, Mali, and Niger were particularly hard hit (Sidikou 1973), the Ahaggar region of the Sahara was also influenced, and Lake Chad was reduced considerably (Alexander 1907). Lakes, rivers, and rainfall throughout Africa progressively diminished, and harvests were fair to poor. The "desiccation" culminated in severe drought, which was most intense in 1911 and 1914 and which persisted for about a decade (Bernus and Savonnet 1973, Sircoulon 1976, Kates 1981). At Freetown, Sierra Leone, in coastal West Africa, annual rainfall was 30-35 percent lower during the period 1910-1940 than between 1880 and 1895. The discharge of the Nile was reduced by 35 percent, and the mean depth of Lake Chad was reduced by about 50 percent. West African rivers also carried less water. Lake levels dropped rapidly in many areas of the continent. A less intense and less widespread drought occurred in the 1940s. Following a period of greater rainfall in the 1950s, drier conditions recommenced in the early 1960s. They set in first around 1960 in the most northerly areas and spread progressively southward. By 1968 the impact of drought was felt throughout the region. The drought continued, with some amelioration in 1974 and 1975, into the 1980s.
The contrast between the 1950s and the subsequent drought is extreme. It would appear that the sequence of these two episodes was responsible for much of the tragedy associated with the drought, as the wet conditions of the 1950s promoted use of more marginal lands. Rainfall was consistently high throughout the 1950s (Figure 2); as much as 50-60 percent above the mean in the Sahelo-Saharan zone, 20-30 percent above the mean in the Sahel proper, and 10-20 percent above the mean in the Sudanian zone to the south.
Conditions changed abruptly around 1960, and rainfall continually declined until 1973. During that year rainfall was about 60, 40, and 30 percent below the mean in the Sahelo-Saharan zone, the Sahel proper, and in the Sudanian zone respectively. Thus, in some marginal areas, tropical rainfall in the 1950s was more than twice that for the period 1968-1973; in the Sahel proper, rainfall averaged nearly 350 mm/yr in the 1950s but only 200 mm/yr from 1968-1973. The drought continually spread and intensified from 1968-1973. Even 1969, considered by some to be a year of relatively good rainfall, was abnormally dry in all areas except the far western portion of the Sahel. In the 1950s, rainfall in northern Africa was below average in only 7 of 37 regions but was above average in only 5 regions during the period 1968-1973 (Figure 5). Thus, the drought extended well beyond the Sahel.
It is commonly thought that the drought ended in the mid-1970s, but a recent analysis of more current data contradicts this. The years 1974 and 1975 were wetter than the preceding ones, but rainfall totals were still below normal. The years 1976-1980 were consistently dry, with certain years apparently matching the worst ones earlier in the decade. In the Sahelo-Saharan zone, rainfall was more than 50 percent below normal in 1977, 1978, and 1980, and 40 percent below normal in 1976 (Figure 6). The Sahel proper and the Sudanian zone were extremely dry in all 5 years (1976-1980), with deficits ranging from 15-35 percent in the Sahel and 10-25 percent in the Sudan. At individual stations, even larger deficits were recorded. These relatively dry conditions have persisted for approximately 2 decades, and it now appears that the present century may be the driest one in the Sahel and its borderlands in over 1,000 years (Figure 7).
Perhaps the most satisfactory explanation for the prolongation of the 1968 drought is to be found in a hypothesis initially proposed by J. G. Charney (1975). This hypothesis maintains that Sahelian rainfall is strongly influenced by "biogeophysical feedback." According to the hypothesis, drought would be reinforced either through the changes it evokes in the Sahelian land surface, usually through devegetation, or through similar changes produced by human impact on Sahelian ecosystems. Charney's hypothesis is discussed by Nicholson in Appendix B.