Persistent warmth in the Turkana Basin, Kenya, from the Late Miocene until present day

Hypotheses for hominid evolution variously call on paleoenvironmental change as a driver of adaptation. Records of Late Cenozoic paleoclimate on global and regional scales are important for evaluating these hypotheses, but they do not necessarily reflect the ancient environmental conditions on continents, where orography, continentality, and ecology can amplify or dampen global forcing in the proxy record. The climate of the Turkana Basin in northwestern Kenya and southern Ethiopia, a focal point of hominid fossil discovery, has appeared to have been hot for the last ~4 million years, despite major global climate change and regional evidence for the canonical C ₃ /C ₄ photosynthesis transition. Here we provide additional paleotemperature data from the Nawata and Nachukui Formations, exposed in the Lothagam uplift, that extend the Plio-Pleistocene record of pedogenic carbonate clumped isotopes back to 8.5 Ma. The average clumped isotope paleotemperature (T(Δ ₄₇ )) from the Lothagam sequence is 32 ± 5 °C (4-8.5 Ma), which is indistinguishable from the Plio-Pleistocene average, and there is no apparent trend with time in the combined record. The relationship between T(Δ ₄₇ ) and ancient soil water δ ¹⁸ O (δ ¹⁸ O _slw ), calculated assuming carbonate δ ¹⁸ O equilibrium, has similarities with the modern relationship between surface water δ ¹⁸ O and mean annual air temperature in East Africa, with exceptions to this trend in the composite Turkana dataset typically having higher calculated δ 18 O _slw . This is readily explained by models of soil water evaporation in arid settings. The temporal trend in Turkana soil carbonate δ ¹⁸ O is thus driven by changes in δ ¹⁸ O _slw and not temperature, and is highest at 6.5 Ma, lowest at 3 Ma, and increases through the Pleistocene. Gaps in the record from ~4-6 Ma preclude further interpretation, but may be filled with new samples.