A NEW PROXY FOR REGIONAL PALEO‑ARIDITY USING CLUMPED AND TRIPLE OXYGEN ISOTOPES OF MODERN SOIL CARBONATES FROM THE SERENGETI ECOSYSTEM, TANZANIA (Invited Presentation)

Paleosol carbonates are used extensively in the rock record to reconstruct vegetation using carbon isotopes, but poor constraints on temperature and soil evaporation has limited the use of oxygen isotopes in these reconstructions. Advances in carbonate clumped isotope thermometry allow for independent controls on temperature (T_∆47), but proxies for evaporation remain elusive. The sensitivity of ¹⁸O-¹⁷O-¹⁶O distributions to kinetic fractionation makes it possible to track evaporation using triple oxygen isotope distributions (∆ʹ¹⁷O). Recent work on lacustrine carbonates shows that ∆ʹ¹⁷O is sensitive to evaporation, but little is known about variation of ∆ʹ¹⁷O in soil carbonates or their potential to track evaporation. Here we report isotopic data from modern soil carbonates collected from 11 sites along a transect in the Serengeti Ecosystem where Mean Annual Precipitation (MAP) and Aridity Index (AI = MAP/PET) range from 499 to 846 mm yr^-1 and 0.33 to 0.55, respectively. δ¹³C values range from ‑2.7 to 1.8‰ and reflect C₄ grasslands. δ¹⁸O values vary by ~8‰ along the transect and increase with increased aridity (lower MAP, higher AI). Mean T_∆47 from these soil carbonates is 23 ± 4˚C and is within error of mean annual air temperature (MAAT) or 1-2˚C warmer at all sites. T_∆47 does not vary significantly across sites or with depth, and soil temperature measurements and modeling suggest that this invariability is likely due to small changes in temperature at the equator. Using T_∆47 temperatures, reconstructed soil water δ¹⁸O values are up to 6‰ greater than δ¹⁸O values of local precipitation and springs, indicating considerable evaporation in the soils. The ∆ʹ¹⁷O values of these soil carbonates range from -161 to -111 per meg and decrease with increasing aridity and increasing δ¹⁸O. This clear trend in ∆ʹ¹⁷O values across an aridity gradient supports the hypothesis that soil water evaporation drives the variance in δ¹⁸O of soil carbonate and indicates the promise of using ∆ʹ¹⁷O values of soil carbonates to track paleoaridity in eastern Africa.