FIGURE 7.20 Relationship between mean annual runoff and mean annual precipitation for areas with different mean annual temperatures (in degrees Celsius).Temperatures are weighted by dividing the sum of the products of monthly precipitation and temperature by the annual precipitation.The quotient gives a mean annual temperature in which the temperature of each month is weighted in accordance with the precipitation during that month. A weighted mean annual temperature greater than the mean, which would normally be computed indicates that precipitation is concentrated in warm months (and vice versa) (Langbein et o/„ 1949).

1000 1200 1400

FIGURE 7.21 Relationship between mean annual temperature and évapotranspiration loss in humid areas, based on data from the eastern United States (Langbein et al., 1949).

1000 1200 1400


FIGURE 7.21 Relationship between mean annual temperature and évapotranspiration loss in humid areas, based on data from the eastern United States (Langbein et al., 1949).

would be far less significant in tropical lakes, but could be important in midlatitude situations where relatively large lakes developed in the past (e.g., Lakes Bonneville and Lahontan). The formation and duration of ice cover would also significantly affect evaporation rates. At the other end of the spectrum, lakes that are drying up may have extremely high salt concentrations; as salinity increases, the evaporation rate will decline due to a lowering of vapor pressure. For example, in a lake with a salinity of 200%o, evaporation will be only 80% of that from a freshwater lake (Langbein, 1961). Such factors complicate any simple empirical relationship one might derive from instrumentally recorded data and point to the inherent difficulties involved in hydrological balance calculations for paleolakes.

Similar difficulties are encountered with precipitation-runoff relationships (see Table 7.2). Even if precise empirical relationships could be demonstrated, reliable paleotemperature estimates are required. Commonly, these too are fraught with uncertainty, and may, in fact, depend implicitly on assumptions about paleoprecipita-tion amounts. For example, paleotemperatures derived from studies of snowline depression depend on the assumption that precipitation amounts are similar to those of today. Any increase in precipitation would require a smaller fall in temperature to produce the same amount of snowline depression. Hence, the use of paleotemperature estimates based on snowline depression (Leopold 1951; Brakenridge, 1978) leads to suspiciously circular reasoning. Without accurate paleotemperatures, quite divergent conclusions may ensue. In Table 7.3, for example, three different studies of Paleolake Estancia, New Mexico are summarized. Although each used slightly different approaches and empirical relationships, the fundamental difference in their final paleoprecipitation estimates (ranging from 80-150% of today's values) lies in the different paleotemperatures assumed. The larger the change in temperature assumed, the smaller is the required increase in precipitation (Benson, 1981). Given a sufficiently large decrease in temperature, values of precipitation even smaller than today can be shown to balance the hydrological budget at times of relatively high lake levels (Galloway, 1970).

TABLE 7.3 Paleoprecipitation Estimates from Selected Western U.S. Hydrological Balance Studies

Paleotemperature change assumed

(°C) Paleoprecipitation/


Study area


July Annual precipitation

Lake Estancia, New Mexico Leopold (1951)

Lake Estancia, New Mexico Brakenridge (1978)

Lake Estancia, New Mexico Galloway (1970)

Spring Valley, Nevada Snyder and Langbein (1962)

Various, in Nevada Mifflin and Wheat (1979)


It is unlikely that controversies over paleoprecipitation estimates will be resolved until (a) more detailed studies of modern relationships between evaporation and temperature, precipitation, and runoff are undertaken, to provide more reliable empirical equations, and (b) better (independent) paleotemperature estimates are available. Paleotemperatures calculated from the extent of amino-acid epimeriza-tion in the shells of freshwater gastropods of known age may help to resolve this issue (McCoy, 1987b; Oviatt et al., 1994).

7.6.2 Hydrological-energy Balance Models

An alternative to the conventional hydrological balance models already described here has been proposed by Kutzbach (1980), who applied the method to Paleo-lake Chad in North Africa. Kutzbach utilized the climatonomic approach of Let-tau (1969) by considering the hydrological balance of a lake basin in terms of energy fluxes at the surface. In simple terms, a positive hydrological balance results when there is insufficient energy available to evaporate precipitation falling on the basin. Instead of calculating paleoprecipitation amounts from estimates of runoff and evaporation (via paleotemperature estimates) a hydrological-energy balance model utilizes estimates of net radiation and sensible and latent heat fluxes over the lake and tributary basin. Modern values of these components are used, based on measurements from locations thought to characterize the paleo-environments of the basin being studied. Paleotemperature estimates are thus implicit in this "analog" approach; in the Paleolake Chad study, for example, the changes in vegetation that were assumed for 5000-10,000 yr B.P. correspond to an area-weighted fall in mean annual temperature of 1.5 °C, by analogy with areas of similar vegetation today. Precipitation was estimated to have been almost double modern values (-650 mm vs 350 mm today), a result that is similar to previous estimates of precipitation for the area at that time, based on a variety of paleoenvironmental data. Using a similar approach, but with somewhat different assumptions about the magnitude of the important parameters, Tetzlaff and Adams (1983) conclude that precipitation over the Lake Megachad basin was at least three times modern values in order to produce the observed increase in lake size.

Kutzbach's approach to the quantification of past climatic conditions from paleolake studies could be applied to many other lake systems. However, whether it represents a major improvement over conventional hydrological studies is debatable, as it involves at least as many assumptions, and may involve a good deal more (Benson, 1981). Nevertheless, this kind of modeling has the merit of being able to identify, via sensitivity tests, which climatic variables are likely to have been of most significance in bringing about the observed lake level changes.

7.6.3 Regional Patterns of Lake-Level Fluctuations

A number of relatively well-dated stratigraphic and geomorphological studies of lake-level fluctuations have enabled maps of relative lake levels to be constructed for selected time intervals for many parts of the world (Street-Perrott and Harrison,

1985a, 1985b; Harrison, 1993). Although individual errors in dating lake levels no doubt occur, mapping relative lake levels for discrete time periods has the advantage that regionally coherent patterns may be discerned, even if isolated "anomalies" occur. Thus, relative lake levels in much of the arid and semiarid world can be mapped for selected time intervals over the last 25,000 yr. Street-Perrott and Grove (1976, 1979) established a basic methodology that has been widely followed ever since. They identified the total range of lake-level fluctuation at each site, from the level of complete desiccation to the known maximum level (or overflow) and defined three categories: low levels were when lakes were at no more than 15% of their maximum range; intermediate levels were when lake levels fluctuated between 15 and 70% of their range; high levels were when lake levels exceeded 70% of the total altitudinal range. Maps were then prepared to show the spatial distribution of low, intermediate, or high lake levels, so defined, at each time interval (Figs. 7.22 and 7.23). These maps demonstrate remarkable coherence in the spatial and temporal patterns of lake-level fluctuations. During the Last Glacial Maximum (18,000-17,000 yr B.P.) most of the evidence from the intertropical zone (largely dominated by data from Africa) indicates that the area was relatively dry (see Fig. 7.22); only in extratropical regions (the North American Great Basin, in particular) is there an abundance of evidence for extensive lake stages at that time, related to

FIGURE 7.22 Lake-level status at -18,000 yr B.P. Lakes were low over most of Africa, but high in the western United States. See text for definition of high, intermediate, and low lake status (Street-Perrott and Harrison, 1985a).
FIGURE 7.23 Lake-level status at -6000 yr B.P. Lakes were high over much of Africa, as they had been for most of the preceding -3000 yr. See text for definition of high, intermediate, and low lake status (Street-Perrott and Harrison, 1985a).

the displacement of storm tracks around the southern margins of the Laurentide ice sheet (Webb et al., 1993a). Thus, the so-called "pluvial" climate of the western United States at the glacial maximum is not a viable model for tropical and equatorial regions (Butzer et al., 1972; Nicholson and Flohn, 1980; see Section 9.7.4). During this arid phase, dune systems on the equatorward margins of the Sahara and Kalahari deserts greatly expanded (Sarnthein, 1978); in the southwestern Sahara, for example, dunes even blocked the much-reduced flow of the Senegal River in Mali (Michel, 1973). Low lake levels persisted in these regions until -12,000 yr B.P., when many lakes began to fill, commonly spilling over from their enclosed catchment basins and initiating new drainage systems. Although maximum lake development throughout the intertropical zone peaked in the early Holocene (Rognon and Williams, 1977; Harrison, 1993; Figs. 7.23 and 7.24), it is of interest that lake levels were generally low in midlatitudes at this time, suggesting a poleward displacement of the Westerlies. In sub-Saharan Africa lake expansion was particularly spectacular, with Lake Chad expanding in area to a size comparable with the Caspian Sea today (Grove and Warren, 1968; Rognon, 1976; Street and Grove, 1976). The high lake phase continued until 4500-4000 yr B.P., interrupted in most of Africa, at least, by an episode of increased aridity and somewhat lower lake levels (though still relatively high) around 7000 yr B.P. (Nicholson and Flohn, 1980). Profound ecological changes accompanied these periods of more effective precipitation,

Global Intertropical (IT)

Number of basins

Was this article helpful?

0 0

Post a comment