2015 GSA Annual Meeting in Baltimore, Maryland, USA (1-4 November 2015)

Paper No. 101-14
Presentation Time: 11:45 AM

LAKE VICTORIA MEGADROUGHT AND EQUATORIAL GRASSLAND EXPANSION DURING THE LATE PLEISTOCENE


BEVERLY, Emily J.1, WHITE, Joseph D.2, PEPPE, Daniel J.1, FAITH, J. Tyler3, BLEGEN, Nick4 and TRYON, Christian5, (1)Terrestrial Paleoclimatology Research Group, Department of Geology, Baylor University, One Bear Place #97354, Waco, TX 76798-7354, (2)Department of Biology, Baylor University, Waco, TX 76798, (3)School of Social Science, University of Queensland, Michie Building (#9), Brisbane, 4072, Australia, (4)Department of Anthropology, University of Connecticut, Charles Beach Hall, 354 Mansfield Road, Storrs, CT 06269, (5)Department of Anthropology, Harvard University, Cambridge, MA 02138, Emily_Beverly@baylor.edu

Lake Victoria is the largest freshwater lake in the tropics by surface area (68,800 km2) and is currently a biogeographic barrier between the eastern and western branches of the East African Rift. Lake Victoria has desiccated in the past at ~17 ka and again at ~15 ka, but little is known from this region prior to the Last Glacial Maximum. New data from thirteen Late Pleistocene paleosols from Karungu, Rusinga Island, and Mfangano Island indicate that mean annual precipitation (MAP) was ~56% less than modern between ~94 ka and >35 ka. These paleosols have been tephrostratigraphically correlated across seven sites, and sample a transect along the eastern margin of the modern lake ~55 km long. Using the CIA-K paleoprecipitation proxy, the MAP for these paleosols ranges from 0.552 to 1.000 m yr-1 with an average 0.800 m yr-1 and no statistical difference throughout the 11-m thick sequence. These paleoprecipitation estimates were applied to a water budget model to understand the effects of prolonged aridity Lake Victoria. Data from 1965 and 1978 were used to establish the validity of the model for the present day lake. This model was adapted to reconstruct lake levels during the Late Pleistocene using a runoff ratio (0.08), the fractional precipitation in relation to modern (0.44), changes in fractional vegetated area (0.25), and an adapted Penman equation for evaporation to adjust for changes in insolation (±40 W m-2 of modern) due to precessional forcing and temperature (±2°C of modern). Using a bathymetric analysis of Lake Victoria, changes in lake level can be related to changes in area and volume of Lake Victoria. According to the model, at 44% precipitation, Lake Victoria dries up within centuries regardless of the change in insolation, and loses the modern Nile outlet within decades. At +40 W m-2, the lake can be refilled but requires at least 95% of modern MAP. At -40 W m-2, Lake Victoria can be refilled at lower precipitation (60%), but over 1000s of years. Therefore, it is likely that for most, if not all, of the interval between 94 ka and >35 ka, Lake Victoria was desiccated and not present on the landscape. This would have removed a major barrier for the movement of fauna, likely including early modern humans, and providing a dispersal corridor between the rifts and across the equator.