ASSESSING WATER STRESS USING TRIPLE OXYGEN ISOTOPES IN SOIL CARBONATES IN THE AWASH AND TURKANA REGIONS OF EASTERN AFRICA DURING THE PLIOCENE AND PLEISTOCENE (Invited Presentation)

The transition to more open ecosystems in eastern Africa during the Pliocene and Pleistocene is often attributed to aridification. However, it has been difficult to test these links at and between landscapes where fossils are found. Carbon isotopes of soil carbonates from the Turkana and Awash Basins have been widely used to document the increase in open ecosystems in eastern Africa since 4 Ma. However, the oxygen isotope records from these same soil carbonates (δ¹⁸O_sc) suggest different histories of water stress between the basins. The Turkana record shows a significant increase in δ¹⁸O_sc values (~8 ‰) after 2 Ma. In contrast, δ¹⁸O_sc values from the Awash region are relatively steady for the last 4 Ma. Testing links between grassland expansion and aridity using δ¹⁸O_sc records is challenging due to the difficulty in separating the influence of temperature, precipitation δ¹⁸O values and soil water evaporation.

Here, we use triple oxygen isotope (∆'¹⁷O) distributions of soil carbonates to assess water stress in the Turkana and Awash Basins. We sampled soil carbonates from both basins (n = 18), spanning ~4.0 to 0.4 Ma. ∆'¹⁷O values of reconstructed soil water (∆'¹⁷O_rsw) for all samples range from -26 to +26 per meg and suggest formation of soil carbonates from soil water with varying degrees of evaporation. These values are consistent with ∆'¹⁷O_rsw values from modern soil carbonates from eastern Africa (Serengeti -29 to +33 per meg; Awash -24 to +4 per meg). The ∆'¹⁷O_rsw values from Plio-Pleistocene soils also overlap with the range in ∆'¹⁷O values of river and groundwater samples from Turkana and the Awash today (+5 to +35 per meg, n = 8). Temporal trends from these initial data indicate overlap in ∆'¹⁷O_rsw values from the Awash and Turkana before ~2 Ma. After ~2 Ma, ∆'¹⁷O_rsw values from Turkana are lower than those from Awash, indicating more evaporation. Furthermore, the distribution of data in ∆'¹⁷O ~ δ'¹⁸O space suggests variation in precipitation δ¹⁸O values at both Turkana and the Awash since 4 Ma. Our results support the hypothesis that both variation in soil water evaporation and the isotopic composition of meteoric water drive variation in δ¹⁸O values of soil carbonates. These data show promise for directly comparing water stress of ancient landscapes associated with records of both grassland expansion and human evolution.