Quantification of climate change for the last 20,000years from Wonderkrater, South Africa: Implications for the long-term dynamics of the Intertropical Convergence Zone

International audience; In southeast Africa – a region for which few palaeoenvironmental records are available – the fossil pollen record from the Wonderkrater spring mound has contributed substantially to our understanding of past vegetation change since the Last Glacial Maximum (LGM; 21 ka). Multivariate analysis of the pollen data by Scott and Thackeray (1987) provided environmental reconstructions suggesting relatively mesic LGM conditions, with warm and dry conditions during the early Holocene (11–6 cal kBP). This conforms to predicted patterns of precipitation change in the southern African tropics in response to Northern Hemisphere cooling and or-bital forcing. Subsequent data from the Cold Air Cave speleothems and a sea-surface temperature record from the Mozambique Channel, however, indicate that conditions during the early to mid-Holocene may have been wetter than present in the Wonderkrater region. To explore this question further, we have created a series of botanical–climatological transfer functions based on a combination of modern climate and plant distribution data from southern Africa. Applying these to the Wonderkrater fossil pollen sequence, we have derived quantitative estimates for temperatures during the cold and warm quarters, as well as precipitation during the wet and dry quarters. In addition, a species-selection method based on Bayesian statistics is outlined, which provided a parsimonious choice of likely plant species from what are otherwise taxonomical-ly broad pollen-types. We do not propose that our findings invalidate the previous principal component analyses, but they do have the advantage of being based more clearly on the relationship between modern plant distributions and individual climatic variables. Results indicate that temperatures during both the warm and cold seasons were 6 ± 2 °C colder during the LGM and Younger Dryas, and that rainy season precipitation during the Last Gla-cial Maximum was ~ 50% of that during the mid-Holocene. Our results also imply that changes in precipitation at Wonderkrater generally track changes in Mozambique Channel sea-surface temperatures, with a steady increase following the Younger Dryas to a period of maximum water availability at Wonderkrater ~ 3–7 ka. These findings argue against a dominant role of a shifting Intertropical Convergence Zone in determining long-term environmental trends, and indicate that the northern and southern tropics experienced similar climatic trends during the last 20 kyr.