Pollen-corrected leaf wax D/H reconstructions of northeast African hydrological changes during the late Miocene

Plant leaf wax hydrogen isotope (δDwax) reconstructions are increasingly being used to reconstruct hydrological change. This approach is based upon the assumption that variations in hydroclimatic variables, and in particular, the isotopic composition of precipitation (δDP), dominate δDwax. However modern calibration studies suggest that offsets between plant types may bias the δDwax hydrological proxy at times of vegetation change. In this study, I pair leaf wax analyses with published pollen data to quantify this effect and construct the first vegetation-corrected hydrogen isotopic evidence for precipitation (δDcorrP). In marine sediments from Deep Sea Drilling Program Site 231 in the Gulf of Aden spanning 11.4–3.8 Ma (late Miocene and earliest Pliocene), I find 77‰ swings in δDwax that correspond to pollen evidence for substantial vegetation change. Similarities between δDP and δDcorrP imply that the hydrological tracer is qualitatively robust to vegetation change. However, computed vegetation corrections can be as large as 31‰ indicating substantial quantitative uncertainty in the raw hydrological proxy. The resulting δDcorrP values quantify hydrological change and allow us to identify times considerably wetter than modern at 11.09, 7.26, 5.71 and 3.89 Ma. More generally, this novel interpretative framework builds the foundations of improved quantitative paleohydrological reconstructions with the δDwax proxy, in contexts where vegetation change may bias the plant-based proxy. The vegetation corrected paleoprecipitation reconstruction δDcorrP, represents the best available estimate as proof-of-concept, for an approach that I hope will be refined and more broadly applied.