TIMING AND DURATION OF LATE QUATERNARY ARID EPISODES IN THE KALAHARI: RESOLVING UNCERTAINTIES IN THE DATED AEOLIAN RECORD

Since the mid 1990s published luminescence chronologies (principally OSL (e.g. Stokes et al 1997, Thomas et al 2000); also TL (e.g. Eitel and Blumel 1998)) have been contributing a directly-dated record of dune building phases in the Mega Kalahari during the last glacial cycle. Multiple phases of dune construction have been proposed, at c.16-9 ka, 26-20 ka, 36-29 ka, 46-41 ka, 60-55 ka, 115-95 ka ago with regional variations that indicate a complex spatial and temporal history of aridity since 100 ka ago. A number of potentially problematic issues, however, have arisen within this record and its interpretation (Thomas and Shaw 2002, Bateman et al 2003), including record preservation in the sediment limited SW Kalahari, constraints on the general depth of sampling within aeolian units, the potential for clustering of apparent dune development phases to be an artefact of sampling strategies, and the compatibility of independently-derived arid and humid chronologies. We present an examination of recent and new data sources that help a reassessment of these issues. Data sources presented include integrated arid and humid chronologies from the N Kalahari, deep mine samples from the SE Kalahari, and linear and lunette full-depth cored chronologies from the SW Kalahari. We also consider the implications of recent offshore core data (Shi et al 2000) for interpretation of the terrestrial Kalahari record. Together these sources allow a better resolved picture of the duration and dynamics of last glacial cycle arid conditions in the Kalahari to be made.