Implications of Late Pliocene–Pleistocene Humidity Fluctuations in the Qaidam Paleolake (NE Tibetan Plateau) Deduced from Magnetic Susceptibility Measurements
Wednesday, 17 December 2014
The Qaidam Basin (NE Tibetan Plateau) contains a near-continuous, up to 12 km thick sequence of Cenozoic strata that offers a unique opportunity for studying long-term climate change. We investigate the 940‑m‑long drill core SG-1 from the western Qaidam Basin, which is characterized by a long-term transition from a semi-deep freshwater lake to nearly complete exsiccation of the water body, detected by several studies including geochemical and lithological observations. Based on magnetostratigraphy and optically stimulated luminescence dating, and refined by orbital tuning, the SG-1 core spans the interval from 2.69 to 0.1 Ma. Moisture availability in the western Qaidam Basin deduced from the pollen ratio Artemisia/Chenopodiaceae (A/C), suggests desert to steppe vegetation along core SG-1 as a long-term feature. Magnetic susceptibility (χ) is well suited for the high-resolution investigation of paleohumidity. The meaning of χ as a paleohydrology proxy is shown by comparing χ to other magnetic proxies for checking its relation to magnetic grain sizes and magnetic mineralogy as well as to pollen results. χ variations are analyzed to obtain regional information on the factors leading to the drying process of the Qaidam paleolake as well as potential driving factors for humidity fluctuations (e.g., insolation). An important topic that needs further investigation is the influence of monsoon in the Qaidam Basin. While the southern part of the Tibetan Plateau is directly affected by monsoon precipitation through the topographic barrier, its influence in the past is questionable in the hyper-arid Qaidam Basin. We check a potential coupling to the monsoon system in the western Qaidam Basin by comparing our χ record to reconstructions of the Asian monsoon system from other archives as well as searching evidence from orbital cyclicities found in the χ time series.