Quaternary glaciation in Africa: key chronologies and climatic implications

Abstract

AbstractMultiple episodes of Quaternary glaciation are evidenced on >10 distinct mountain localities throughout Africa, with the best dated sites from Kilimanjaro and Mt Kenya in equatorial East Africa. A general paucity of radiogenic dates constrains the glacial chronology, and regional sequences have largely been based on correlations by relative weathering of features. Excellent glacial moraine preservation and other features of erosion put limits on the spatial extent of palaeoglaciers, and have facilitated advances in palaeo‐equilibrium line altitude (ELA) analyses. All radiocarbon dates (n < 30) provide minimum ages of deglaciation, and are consistent with a widespread synchronous advance during the Last Glacial Maximum (LGM, 21 ± 2 ka). K–Ar dates on Kilimanjaro lava flows initially recognised up to three glaciations prior to the LGM. Over 100 cosmogenic radionuclide (CRN) 36Cl exposure ages on glacial erratics and bedrock from Kilimanjaro and Mt Kenya date 16 landforms, confirming maximum glacial extent prior to Marine Isotope Stage (MIS) 10. No glacial geological features have been dated to MIS 6. While on Kilimanjaro the CRN ages confirm a clear LGM (MIS 2) maximum, Mt Kenya lacks definitive LGM moraines. Lateglacial advances or stillstands on both mountains are suggested by low‐relief moraines dating to 14.6 ± 1.2 ka (two boulders) on Mt Kenya and 16.3 ± 1.9 ka and 15.8 ± 2.5 ka on Kilimanjaro. Multiple moraines upvalley have early Holocene ages ranging from 11.2 ± 0.8 to 8.6 ± 0.2 ka. Isolated and equivocal dates on both mountains also provide speculative evidence that the maximum extent since MIS 5e may have occurred close to 30 ka, similar to sites in South America. New palaeoclimatic understanding from lake and ocean proxies suggests hypotheses for the source of low‐latitude climate changes causing the advances that can be tested with improved dates. A high degree of intra‐regional variance in ELA is observed, with differences by aspect reflecting precipitation gradients. Future work should include more CRN dates, improved mapping and inverse climate–mass balance modelling. Copyright © 2008 John Wiley & Sons, Ltd.