Reevaluation of carbonate concentration and oxygen isotope records from Lake Qinghai, the northeastern Tibetan Plateau

Lake Qinghai is the largest lake on the Tibetan Plateau, and it is also one of the important sites for studying global environmental changes. Over the past 30 years, many studies have used oxygen isotope of authigenic carbonates from the lake as the proxy to infer past environmental and climate changes on the Plateau. However, debate on interpretations of isotopic data and their environmental implications still exist, largely due to the complex arid environment settings and multiple sources/species for carbonate minerals within the lake. In this study, we systematically analyze δ18O values in different-type carbonates collected [i.e. bulk carbonates, ostracode shells, Chara encrustations, and fine-grain (<63 μm) carbonate minerals] from modern lake sediments and surrounding soils, as well as the down-core δ18O values of bulk/fine-grain carbonates since the Last Glacial Maximum. Together with previously published δ18O records from ostracode shells, we try to re-evaluate the controlling factors of variations in lacustrine carbonate δ18O data and to infer environmental changes on the northeastern Tibetan Plateau since the Last Glacial Maximum. Our results show that the lake depth, or the size of the water body, is an important factor to influence the lake water and carbonate δ18O values. A shallow and small lake would be more easily influenced by precipitation δ18O which is characterized by negative values at Lake Qinghai region, while a deep and large lake would be better to reflect environmental changes such as the precipitation-evaporation balance. The “lake volume” effect might be an explanation for the negative carbonate δ18O values during the early Holocene, which was likely caused by an increased influence of negative δ18O values in precipitation and glacial melt water under a small and shallow water body. The δ18O values of ostracode shells and bulk carbonates show similar variations since both of them are dominated by lake water oxygen isotopic composition, but they still have distinct geochemical information. The isotopic differences between ostracode and bulk carbonates probably reflect the temperature differences between the surface and the bottom of lake water. In addition, the δ18O values of evaporative induced carbonates may correlate with carbonate contents, while those of Chara encrustations do not show any correlation with carbonate contents. Our results suggest that special caution would be necessary when using lacustrine δ18O values of authigenic carbonates to infer past hydrological and climate changes in an arid environment.