干旱区生态环境知识资源中心

A 12 m sediment core from the Aiding Lake basin in the today hyper-arid Turpan depression (Autonomous Xinjiang Province, NW-China) has been studied. The core, which was drilled near the deepest point of the endorheic Turpan basin (154 m below sea level), revealed several climatic and palaeoenvironmental changes during the Late Pleistocene. The sediments which were analyzed relating to sedimentology (grain size), geochemistry (Ti/Al ratio, chemical index of weathering, total organic and inorganic carbon, sulfur, nitrogen), and magnetic susceptibility, are of fluvial, eolian, and lacustrine origin. Age control was done by five C-14 dates. Reservoir effects for the stratigraphic units were considered. Less saline conditions lasted until the end of the Pleistocene, forced by stronger westerlieS and low, but increasing temperatures. Sometime after Dansgaard-Oeschger event 1 (14.6 ka) up to present, environmental conditions were arid and Aiding became a salt lake. High sand contents indicate eolian input during a period of cold, windy and arid conditions, associated with Heinrich event 2 (-24.8-23.4 ka cal BP; HEMMING 2004), whereas high sand input of the uppermost 2 m and salt accumulation reflects a general change to drier conditions since the Pleistocene/Holocene transition. In summary, the Lake decreased from its maximum size of around 3000 km(2) during the MIS3 to a seasonal small lake today.

The measured Rb/Sr ratio, generally used as an indicator for weathering conditions of lake catchments, with reduced Sr influx during arid periods due to weak chemical weathering, has to be interpreted differently for Aiding Lake: Very low Rb/Sr ratios in the uppermost 2.7m similar to that of unweathered loess represent the recent drought and Sr accumulation during the Holocene. Contrary to other studies, we found e.g. that Rb/Sr dynamics do not indicate the intensity of past chemical weathering, but allows the distinction of different source areas and the decoupling of weathering and subsequent transport to the sink. This is consistent to the westerly index, according to AN et al. (2012), which was calculated to separate eolian from fluvial influx. Our study shows that the interpretation of palaeoclimate proxies depends on the understanding of the entire process-response-system within a specific catchment and environment. In summary and in comparison to other regional studies, we can distinguish six palaeoclimatic phases: A) with strong westerlies and cold-wet climate (c. 28.5-24.8 ka BP); B) with very strong winter monsoon and very cold climate (24.8-23.4 ka BP); C) with strong winter monsoon and cold climate, merging to cold more humid conditions due to enhancing westerlies (23.4 - similar to 20.5 ka BP). Phase D showed strongly reduced eolian dynamics afterwards and was a period of cool-wet conditions with weakening westerlies and a geomorphological stability in the Turpan basin (until similar to 16 ka). Phase E was dominated by the cold-dry spell of Heinrich event 1(15.8-14.6 ka BP), eolian transport was reactivated by high winter monsoon wind speeds, and fluvial discharge decreased. During the D/O1 warming (after 14.6 ka BP), material was merely transported from the melting mountain glaciers to the lake. In the Turpan basin, warm and dry conditions led to a sealing of surfaces by biocrusts, which again suppressed eolian processes. Phase F started with the transition to hot-very dry Holocene climate. It experienced enhanced seasonality (stronger winter and summer monsoon) and marks a period when the propagation of moist westerly air masses to the Turpan basin was prevented.