Hydrogen isotopic compositions along a precipitation gradient of Chinese Loess Plateau: Critical roles of precipitation/evaporation and vegetation change as controls for leaf wax δD

The hydrogen isotopic compositions (δD) of long-chain plant leaf waxes can reflect changes of continental hydrology and thus have been increasingly utilized for paleoclimate reconstruction. One of the unresolved major issues is whether variations of leaf wax δD (δDwax) signals along a precipitation gradient reflect changes of precipitation δD, precipitation amount, and/or evapo-transpiration. This ambiguity limits our interpretation of δDwax in geological records as well as the quantitative reconstruction of paleohydrological variability. Here we systematically investigated δD of soil water and waxes extracted from soil and plant leaves in the Chinese Loess Plateau (CLP) and its surrounding areas along a precipitation gradient with mean annual precipitation (MAP) varying from 140 mm to 676 mm. The results showed that while the variation of modeled precipitation δD has no significant correlation with MAP, soil water δD, soil δDwax, and individual plant δDwax exhibit negative correlations with MAP. In relatively arid areas, the δD values of soil water and plant leaf waxes are significantly more positive due to much lower precipitation relative to evapo-transpiration, suggesting that effective precipitation (P:E ratio) has played a crucial role in the D-enrichment of soil and plant leaf waxes in this region. While parallel decreasing trends in δDwax are found along the increase of precipitation gradient among dominant plants (including Artemisia spp., Aster hispidus, Stipa bungeana, and Cleistogenes squarrosa) in the region with similar slopes, the large δD offsets between plant groups suggest that plant type is an important factor in controlling plant hydrogen isotope fractionations. Our results indicate that δDwax preserved in paleosols can be used to infer past conditions of water availability in arid and semi-arid inland regions, but with a mechanism that is different from the influence of “amount effect” in humid areas. Moreover, vegetation changes should be constrained by independent paleobotanical data before monsoon related paleohydrology in the CLP and its surrounding areas can be quantitatively reconstructed using the plant wax δD proxy