PLIO-PLEISTOCENE PALEOENVIRONMENTS IN WEST TURKANA (EAST AFRICAN RIFT SYSTEM, KENYA)

International audienceUnderstanding Plio-Pleistocene paleoenvironments in the East African Rift and how they have changed over time is fundamental across geosciences, the most emblematic being perhaps their role in hominin evolution. In the northern Turkana Depression, among the three main Formations of the Omo Group, the Nachukui Formation is exposed on the western shore of present-day Lake Turkana. It consists of a > 700 m succession of fluvial-deltaic-paralic-lacustrine sediments deposited between ca. 4.00 and ca. 0.75 Ma, which have been extensively investigated using depositional environments and sequence analyses during the RiLakS project. Additionally, we use δ13C isotopic ratios in soil carbonates as a proxy of paleovegetation.Facies and sequence analyses reveal that proto-lake Turkana experienced 7 high-amplitude Transgression-Regression (T-R) cycles between ca. 4.00 and 1.25 Ma, which are locally superimposed by lower amplitude, secondary T-R cycles. Paralic or offshore depositional environments are observed quasi-continuously, at least in basinward sections, which suggest a more prolonged persistence of lacustrine conditions than hitherto accepted. Remarkable match between reconstructed evolution of palaeolake level and evolution of woody cover in the Lower Omo Valley suggests that rainfall in the drainage basin of the Omo river, i.e. rainfall over the Ethiopian dome, controlled high-amplitude paleolake fluctuations during the Plio-Pleistocene whereas rainfall in West Turkana had only limited influence. Additionally, we reveal that variation in sediment supply from the rift shoulder caused two different sedimentary dynamics, i.e. accommodation-dominated and supply-dominated systems, to alternate during the deposition of the Nachukui Formation, likely due to successive pulses of activity of the border fault.Finally, a reconstruction of paleolake fluctuation in the Turkana Depression is proposed and successive sedimentary paleolandscapes are presented. Respective influences of climate (local to regional) and tectonic (i.e. border fault activity) on paleoenvironments are deconvoluted. Our work highlights that the continuous reconstruction of depositional environments at various localities in a well-constrained chronological context is essential to reliably reconstruct paleoenvironments. However, until now, such powerful approach remains rarely associated with traditional site by site investigations