enabling human occupancy and cultural activities to take place in Saharan Africa at a scale almost inconceivable today (see Section 9.7.4). Since -4500 yr B.P. there has been a relatively steady decrease in lake size, leading to a situation over the last 1000 yr in which lakes over most of the Tropics and lower midlatitudes are virtually all at low stages (Fig. 7.25). In many areas, it appears that lake levels have rarely been lower during the past 20,000 yr.
In spite of the excellent spatial coherence displayed in the maps shown, caution must be used in interpreting lake-level data in this way. Once a lake has desiccated, there is no way of knowing just how much drier the conditions may have been during the period of desiccation. Dry periods are thus likely to be underestimated. Comparison of lake basins of vastly different sizes can lead to erroneous conclusions. Small volume lakes respond much more rapidly to hydrological variations than large deepwater lakes and are likely to record higher frequency climatic variations than large volume lakes. A good analogy to this would be the problems encountered in trying to correlate fluctuations of a small alpine glacier with those of a large ice sheet; clearly the response times of the two systems are different by perhaps an order of magnitude. The problem is not quite as profound in lake systems, however, as mass turnover rates in lakes are rarely longer than a few decades even for very large lakes (Langbein, 1961). Thus, providing a coarse enough time interval is used, and major low-frequency components of hydrological changes are being considered, it should be possible to make broad regional comparisons. When regional patterns show little spatial coherence for a particular interval, it may be that the climate was fluctuating fairly rapidly over a wide range so that no major low-frequency signal dominates the record. Finally, it should be noted that the 15 and 70% category boundaries (for low and high stages) used by Street and Grove (1976, 1979) may represent quite different surface area states, depending on the morphol-
ogy of the basin. Consider, for example a lake that overflows from a deep narrow basin to a broad shallow plain at some level; the increase in lake depth represents a vastly greater change in surface area (and hence in evaporation from the surface) than a fall in lake level of comparable magnitude. When evaporation from the expanded lake area balances inflow, a new equilibrium is reached. Thus lake surface area is the critical variable controlling lake depth, and this in turn is a function of the basin morphology. Unfortunately, there are not yet enough reliable data on lake area changes to make a global-scale study feasible.
In spite of these caveats, lake-level data from Africa, together with palynologi-cal, geomorphological, and archeological data, have enabled a fairly detailed picture of paleoclimatic fluctuations over the last 20,000 yr to be obtained. This led Nicholson and Flohn (1980) to speculate on what the major circulation features over the continent were like at different periods in the past. Figures 7.26 and 7.27 show the principal differences in circulation that they envision during the main arid phase (20,000-12,000 yr B.P.) and the subsequent period of high lake levels (10,000-8000 yr B.P.). Major changes in position of the subtropical high pressure centers are evident in their reconstructions; at 20,000-12,000 yr B.P. a much stronger Hadley cell circulation would have resulted in increased subsidence in the subtropical high pressure cells and intensified upwelling of cooler equatorial water, thereby reducing oceanic evaporation rates in those regions. The seasonal migration of the intertropical convergence zone (ITCZ) would have been greatly reduced, preventing moisture-bearing winds from the Gulf of Guinea reaching southern Saha-ran regions. Displaced westerly flow (due to a strong baroclinic zone along the ice-sheet margin over northern Europe) would have brought relatively frequent depressions and, hence, relatively moist conditions, to North Africa. By contrast, from 10,000 to 8000 yr B.P. subtropical high-pressure zones may have been displaced poleward as the ice sheet over Scandinavia diminished in size and Equator-Pole temperature gradients (in both hemispheres) were reduced. An increase in interhemispheric temperature differences could have resulted in a northward dis-
quasi- , permanent quasi-permanent
FIGURE 7.26 Conceptual model of atmospheric circulation at ~I8,000 yr B.P. (and prevailing circulation pattern from 20,000-12,000 yr B.P) based on geological and palynological data. Dark shading = areas more humid than today; light shading = areas drier than today. Inset (top right) shows present position of intertropical convergence zone (ITCZ) in summer and winter months (Nicholson and Flohn, 1980).
FIGURE 7.27 Conceptual model of atmospheric circulation at 10,000-8000 yr B.P. Dark shading = areas more humid than today; light shading = areas drier than today. ITCZwin refers only to the position over southern Africa (Nicholson and Flohn, 1980).
placement of the ITCZ and increased moisture flux to the continent (facilitated by a warmer equatorial ocean). Evaporation rates from the ocean may have increased by as much as 50% in areas where upwelling of cool water was no longer occurring. Finally, the interaction of upper level troughs with low-level tropical disturbances may have led to increased cyclogenesis and a significant contribution to Sahara rainfall totals from the resultant Sudano-Saharan depressions (Flohn, 1975; Nicholson and Flohn, 1980).
General circulation model experiments by Kutzbach and Otto-Bliesner (1982), Kutzbach and Street-Perrott (1985), and Kutzbach and Guetter (1986) largely support the conceptual models of Nicholson and Flohn. However, their studies indicate that the most important driver of the observed hydrological changes is orbital forcing. At 9000 yr B.P. July radiation from 0° to 30° N was 7% higher (at the top of the atmosphere) compared to today, resulting in an increase in net radiation at the continental surface of -11%. This led to an increase in the pressure gradient between the oceans and land areas, causing enhanced monsoonal airflow and an increase in precipitation of >4 mm/day in July (compared to today) over much of the Saharan region (Kutzbach, 1983; Street-Perrott and Perrott, 1993). The effectiveness of the moisture increase was further enhanced by reduced winter radiation receipts, and hence less evaporation. Positive feedbacks due to the change in vegetation from desert to grasslands (and associated changes in soils) may also have led to increased precipitation at the boundaries of the arid zone (Kutzbach et al., 1996). These large-scale changes in forcing seem to explain the first-order changes in lake levels and environment recorded by the sedimentary and archeological evidence from Africa. However, there is also much evidence for very rapid changes in lake levels that occurred at a higher frequency than can be accounted for simply by orbital forcing (Street-Perrott and Roberts, 1983; Gasse and Van Campo, 1994). Such changes may be related to sea-surface temperature changes in the Atlantic (Street-Perrott and Perrott, 1990), but further studies of such linkages are needed to fully understand the mechanisms involved.
Was this article helpful?