Climate conditions during the Upper Pleistocene and Holocene

Van Zinderen Bakker (1976) proposed that, during the Last Glacial Period, cool and dry conditions prevailed in north and east South Africa, while cold temperatures, strong winds, and wet winters prevailed in the southern section.

The same author maintains that, during the interglacial periods, pluvial conditions prevailed in the eastern part of South Africa in summer, while dry westerlies prevailed along the western and southwestern coasts. During winter, cyclonic storms penetrated the southwestern regions, while the high-pressure conditions caused dryness in the eastern part of South Africa.

Nicholson and Flohn (1980) suggested that, during the Last Glacial Maximum the ITCZ migrated southward, causing most of southern Africa, except the southern tip of the 30° S line, to be drier than at present. During the early part of the Holocene (10 ka to 8kaBP), the area north of the 25° S line was more humid than today, while the areas south of it were drier.

Heine and Geyh (1983) suggested that, during the Last Glacial Period, strengthening of the circulation pattern caused stronger trade winds to bring winter rains as far as the southern Kalahari, and summer rainfall over the Kalahari, but not as far as Namibia. During the Post Glacial Period (17 ka to 15 ka BP), eastern circulation caused most of South Africa to be semi-humid, and to benefit from summer as well as winter rains. At c. 12kaBP, the circulation weakened and semi-arid conditions prevailed in eastern South Africa, where rains fell only during summer. Western South Africa was arid. Some climatologists have pointed out the impact of the ENSO, in which the southern Indian Ocean high-pressure cell is displaced northeastward. This would have caused much drier conditions in the summer rainfall regions of southern Africa, and wetter conditions in the regions of winter rain. Such conditions may have prevailed during the Last Glacial Maximum.

Johnson et al. (1997), on the basis of ratios of carbon, nitrogen and oxygen isotopes from ostrich eggshell at Equus Cave, derived distribution of types of vegetation (C3 and C4) and thus rainfall and temperatures for the last 17 ka. They concluded that paleo-temperatures were at a minimum between 17 ka and 14 ka BP and reached their maximum in the Late Holocene. At 17 ka BP, mean annual precipitation was at a minimum. It increased steadily to modern values by c. 6 ka BP and remained relatively unchanged until the present.

Tyson (1986) suggested that cool periods are generally correlated with enhanced climatic instability, leading to greater extremes of climate. There appears to be general agreement that, during the colder periods of the Upper Pleistocene and Lower to Middle Holocene, the climate was wetter and windier in the winter rain regions, which extended further north and east of the present boundaries, while the summer rain regions were cooler and drier. Periods of warmer climates may have caused higher rainfall in the summer rain regions, even promoting some rain in the winter rain regions, which may have become warmer and drier.

However, the construction of a general paleo-climate curve for the sub-continent of South Africa is difficult to carry out because of contradictions in the conclusions drawn by different research groups. I assume that these contradictions arise through the many instances of subjective interpretation by the various investigators, especially where reconstruction was based on the evidence of paleo-ecological assemblages, such as bones or pollen. These assemblages may have been influenced by local factors as well as climate.

Partridge et al. (1990) reconstructed the climatic fluctuations for the different regions of southern Africa on the basis of paleo-environmental proxy-data. They divided the sub-continent into five paleo-climatic regions.

1. The southern and western Cape, coinciding with mainly "Mediterranean" or winter rain climate;

2. The Karroo semi-arid, mostly summer rain climate (80%);

3. The Kalahari arid summer rain region;

4. The Namib hyper-arid zone;

5. The high precipitation region of eastern South Africa.

Correlating the data from these different regions shows that during the Last Glacial Maximum, 21 ka to 17 ka BP, all the regions, except the Kalahari (which was humid), were characterized by low temperatures (about 5-6 °C lower than at present). Warming proceeded gradually from 17 ka to 12 ka BP, causing wetter conditions in all regions. From 12 ka to 10 ka BP, the warming trend continued. In the eastern and southeastern parts of the subcontinent, the humidity was at the same level as at present, but the Kalahari and Namib deserts became dryer.

In the southern Cape Province, a region characterized by rains during all the seasons, a good sequence was derived for the upper part of the Holocene from a speleotheme taken from Cango Cave (Talma and Vogel, 1992). The paleo-temperatures were calculated from the 18O content of the carbonate. In order to do this, one must know the 518O of the water in the past. This was obtained by comparing the 518O content with that of the Uitenhage artesian aquifer, located about 350 km east of Cango Cave. The

Fig. 4.2. Climates during the Holocene in southern Africa. *Adjusted to scale and streamlined (3-5) points by the running average method.

information derived from the speleotheme regarding temperatures in the Uppermost Pleistocene areas conformed with what would have been expected. There was a temperature decrease from c. 30kaBP, reaching a minimum value between 19 ka and 17 kaBP (reaching about 10 °C compared with 18 °C for the most recent sample).

Afterwards, up to 13.8kaBP, the temperature increased (at about 14kaBP, it reached a maximum of about 14°C). At this time, precipitation of carbonates stopped and was not renewed until 5 ka BP. The range of temperatures from this period to the present is presented in Fig. 4.2.

The 13C contents of the stalagmites formed during the Holocene was higher than that during the Pleistocene. While the latter varied between —9%c and — 11%c (where minus indicates lower than a standard value), that of the Holocene (since 5 ka BP) started at

—9%c and increased quickly to reach a peak of —4%c at c. 2 ka BP and —6.4% at present. The low 13C content of the stalagmites during the Last Glacial Period is a function of the dominance of C3 type vegetation (18%) owing to lower temperatures. Today, the vegetation is composed of more than 60% C4. The slow increase of 13C from 5 kaBP (—9% to —4.9%) does not correspond with the temperature fluctuations curve. Talma and Vogel (1992) suggested that this difference between the 13C and temperature curves could be explained by the fact that the rainfall at 5 ka BP was more of a winter rainfall type than that of today, which is an all-season rainfall pattern.

Issar (1997) suggested that the absence of deposition of calcium carbonate from c. 13 ka to 5 ka BP was caused by winter rains starting to decrease and summer rains to increase as temperatures started rising. The vegetation, being dominantly C3, absorbed and transpired all the water infiltrating the subsurface. Only after the ecosystem changed into a floral assemblage, dominated by C4 plants, which do not have deep roots, did recharge to groundwater, and thus drip water, restart.

A similar cessation of stalagmite formation can be found in Boompleas Cave (Deacon et al., 1984; Deacon and Lancaster, 1988). This cave, which is in Cape Province, is an enlarged opening of a fissure that formed a large domed rock shelter. The sequence is rather continuous down to c. 80kaBP. The Last Glacial Maximum, dated between 22 ka and 17.8 ka BP, was characterized by more angular debris in the deposits, presumably as a result of front type erosion, and a pollen assemblage characterized by a low diversity of species. By comparison, the Holocene samples have a high diversity of species. The major change in vegetation from the Glacial Period to inter-glacial in this region was in species diversity, with a substantial increase occurring after c. 14kaBP.

A palynological and paleo-botanical study of the site (Scholtz, unpublished data, cited by Deacon and Lancaster, 1988) concluded that, during the Last Glacial Maximum, the climate was cold and dry throughout the year. The climate was very wet from c. 14ka to 12kaBP. Total annual precipitation was high and was distributed throughout the year. From 10 ka to 9 ka BP, the climate became a little drier, somewhat cooler and more seasonal than the preceding period. At c. 6.4kaBP, there were strong xeric conditions, probably with little or no rain in the warmer months, and with long, dry, hot summers. From 2 ka to 1 ka BP, there was a decrease in rainfall, especially during the warmer months (October to February). This resulted in a contraction of forest vegetation. Lower temperatures may also have been experienced.

A stalagmite formed in Boompleas Cave, starting during the Upper Pleistocene (later than 80 ka and before 40 ka BP). It ceased to form c. 14.2 ka BP and, unlike that in Cango Cave, did not rejuvenate. The higher percentage of trees during the shift from the Uppermost Pleistocene to the Lower Holocene could explain the absence of groundwater recharge of the sub-surface and thus into the cave (as in Cango Cave).

Another source of information on the climate in Cape Province during the Uppermost Pleistocene and Lower Holocene is to be found in the noble gas composition of the water in the Uitenhage artesian aquifer, northeast of Port Elizabeth and c. 350 km east of Cango Cave. As gas solubility is a function of the temperature of the water, Heaton et al. (1986) measured the concentration of nitrogen (N2) and argon in the water. The samples, dating between 28 ka and 15 ka BP, indicated mean temperatures of about 14 ± 1 °C. This means that they were 5 °C lower than contemporary mean temperatures in this region. The mean temperature for the period from 9 ka BP to the present was calculated as 19.5 °C.

Another paleo-climatic Holocene sequence was derived from a core into a peat deposit formed by a spring in Wonderkrater in the Transvaal, which is in the summer rain eco-zone (Scott and

Thackeray, 1987; Fig. 4.2). At present, the area receives between 400 and 600 mm of precipitation per annum. The palynological sequence from bottom to top was:

1. Earlier than c. 34kaBP: woodland with expanded montane forests;

2. After c. 34.4 ka BP: Kalahari-type savannah with restricted montane forest;

3. At c. 25 kaBP: open grassland with more podocarpus forest;

4. At c. 25 ka to 11 ka BP: mainly open grassland and restricted montane forest;

5. At c. 11 ka to 9.5 kaBP: open grassland with much reduced virtually absent montane forest;

6. 9.5 ka to 6kaBP: Kalahari-type savannah;

7. 6 ka to 4 ka BP: savannah with broad-leaf element;

8. 4ka to 2kaBP: upland bush land type with restricted montane forest;

10. 1 ka BP to present: bush land with restricted montane forests.

A multivariate analysis of the pollen assemblages produced two curves, one interpreting the pollen data for moisture changes and the other for temperatures. Comparing these data curves with that of Cango Cave (Talma and Vogel, 1992), one can say that there is a rather good correlation between dry periods on the Wonderkrater curve and cold periods on the Cango Cave curve, while there is a greater variance between the moisture curves from the two sites.

As the evidence from Cango Cave is more direct, it is suggested that only the dry-humidity curve from the Wonderkrater data should be adopted. We may conclude that during most of the Holocene - on the basis of the Cango Cave data - the warm periods were more humid in the Transvaal, while cold periods were dryer.

A more general conclusion can be drawn regarding the regions in summer rains in South Africa: during warm periods these regions will enjoy higher rates of precipitation and vice versa.

This conclusion is supported by an analysis of the stable isotope variations and layer structure of a section spanning 3000 years from a stalagmite at Cold Air Cave, in the Northern Province of the Republic of South Africa (Holmgren et al., 1999). The darker colored layers of the stalagmite, a product of mobilization of organic matter from the soil, were also characterized by a higher ratio of heavy 13C and 18O isotopes. This results from a warm humid climate of summer rains. Such layers characterized the period from 900 to 1300 AD. Lighter coloring and depleted isotopic composition, denoting cool and drier climates, were characteristic for the period from 1300 to c. 1800 AD, the Little Ice Age.

Support for this conclusion can also be found in the profile of the Pretoria saltpan, which is the infilling of a meteor impact crater (Partridge et al., 1993). The analysis of the pollen assemblages showed that warm and moderately cool conditions were wet, while a cool climate equated to dry conditions. The resolution for the Holocene was not detailed enough and no pollens were found between c. 30 ka and 8 ka BP, most probably because of bad preservation conditions. The few samples that were analyzed showed that rather cool conditions prevailed c. 7.2kaBP; from 4.6 ka to 4.4 ka BP, the climate was warm, and c. 2.3 ka BP, it was cool. The climate later warmed up, reaching its current maximum.

Lancaster (1979) found evidence for a widespread Late Pleistocene humid period in the Kalahari desert from 21 ka to 14kaBP and a sub-humid period from c. 9.7 ka to 6.5 kaBP.

Baker et al. (1995) concluded that the period from 25 ka to 16kaBP was a period of desiccation of the deserts of southern Africa. In the Kalahari, a humid period started c. 17kaBP and persisted also during the early Holocene; theNamib desert became dry during the Holocene.

Tyson and Lindesay (1992) gave the pattern of climate changes for 0-1810 AD in South Africa:


100-200: cool;


200-600: warm;


600-900: cool;


900-1300: warm;


1300-1500: cool;


1500-1675: warm;


1675-1780: cool;


1790-1810: warm.

Based on archaeological evidence, Huffman (1996) found that the warm period, which was also humid, extended from 200 to 600 AD (local Early Iron Age). It led to the extension of settlements in the Central Transvaal and in Botswana, reaching the edge of the Kalahari desert. Another phase of occupation of the fringes of the Kalahari was c. 850 to 1350 AD (local Middle Iron Age). The Little Ice Age, which started c. 1300 and lasted until 1780 AD, with a warmer interval between 1500 and 1675 AD, caused the desertion of the agricultural settlements in the west and shifted the settlement to Zimbabwe's southeast escarpment, which had relatively more rains than the western region because of its warm and wet climate.

A record of the changes of the vegetation of the Kalahari was obtained from investigating a core drilled into a speleotheme in Drotsky's Cave, Botswana. The core, which contained pollen, was dated by the uranium-series method (Burney et al., 1994). It was found that from c. 10 ka to 7 ka BP the site was surrounded by arid grassland with trees, which were adapted to dryness. From c. 7 ka to 6 ka BP, the assemblage of pollen indicated wetter conditions, which prevailed until 3 ka BP although a dryer interlude occurred between 4 ka and 5 ka BP. In general, it can be concluded that the changes in Kalahari vegetation during the Holocene were slight relative to other regions in Africa.

As discussed above, the aridity of southwestern Africa is connected with the upwelling of cold waters of the Benguela system. In its southern part, this system is dominated by a strongly annual regime. In the more extreme south, the sea surface temperatures are also influenced by episodic warm Agulhas current intrusions from the Indian Ocean. Cohen et al. (1992) reconstructed the Holocene history of this upwelling through an analysis of 18 O content and the calcite to aragonite ratio in the shells of Patella, found in shell middens in archaeological sites on the coast of southwestern Africa. The calcite to aragonite ratio increases with an increase in temperature. They found three discrete episodes of 18O enrichment in the shells, corresponding with lower aragonite ratios: evidence of glaciation episodes as well as of colder water. The times of these episodes were between 11 ka and 10 ka, between 4ka and 2ka and between 0.75ka and 0.4kaBP. These authors correlate the first episode with the Younger Dryas cold period, the second with a period of glacier expansion in the northern hemisphere, observed by Lamb (1982), and the third episode with the Little Ice Age. If the diagram by Cohen et al. (1992) is compared with the Levant data, it can be said that the first episode can be correlated with the cold period of the Early Neolithic, the second with that of the Late Bronze and Early Iron Age, while the last one can be correlated with the Crusader and Little Ice Age cold periods.

A group of scientists from the Climatic Research Unit of the University of East Anglia at Norwich, UK has carried out a comprehensive investigation on the potential impacts of the expected climatic warming on southern African environments, natural as well as human (Hulme, 1996). The study included an investigation of the impact of past climates as well as of recent trends of change. These data were used in a simulation using a composite computered climate model, which linked results from a general circulation model with a simple climatic model. The running of the model enabled a forecast to be made of a future "core" scenario for South African climate in the 2050s.

In general the "core" scenario forecast is in accord with my conclusions and it can serve as a conclusion for this chapter. This scenario sees modest drying over large parts of the southern part of Africa (south of latitude 15° S) of c. 5%, except for the southwestern Cape Province, Zimbabwe and the Transvaal, where the drying may even reach 10%.

0 0

Post a comment