The age and palæoenvironmental significance of the Kalahari Sands in western Zimbabwe: A thermoluminescence reconnaissance study

Thermoluminescence (TL) dating, has been applied to 10 samples from the æolian Kalahari Sands of western Zimbabwe using the total bleach (additive dose) and regeneration techniques on coarse-grain quartz separates. The results suggest that the main phases of sand accumulation occurred between 10 and 96 ka. This Upper Pleistocene age is consistent with current opinions and argues against an important period of Holocene æolian activity. The oldest sands were found in the Victoria Falls area where they overlie a ferricrete horizon. Based on our age determinations, we were able to constrain the age of this horizon to between 96 ± 8 and 160 ± 23 ka, which approximately coincides with the Last Interglaciation in the high latitudes. The youngest ages were obtained for sands from the southeastern part of the Hwange dune field but our limited sampling permits assigning only a minimum age of ca 20 ka to the ferricrete horizon observed in this area. Thus, we cannot ascertain if this ferricrete horizon is isochronous with the one at Victoria Falls or if it developed during a younger humid phase. Dune building activity most probably occurred in recurrent cycles of aridity interspersed with periods during which æolian activity was limited or non-existent. Unfortunately, besides the ferricrete horizons, no other stratigraphical evidence of climatic fluctuation has been observed in the study area. However, it is possible that palæosols have been entirely eroded before the subsequent deposition of new sediment, resulting in an incomplete stratigraphical record. The possibilities and limitations of luminescence dating for elucidating phases of climatic fluctuation in such situations are discussed briefly. In addition, we noticed serious discrepancies between the results yielded by the different techniques that were used for determining the dose rate. Hence, further work is needed to validate the accuracy of our dose rate measurements that may cause our TL ages to be about 30% too young.