An abrupt centennial-scale drought event and mid-holocene climate change patterns in monsoon marginal zones of East Asia.

The mid-latitudes of East Asia are characterized by the interaction between the Asian summer monsoon and the westerly winds. Understanding long-term climate change in the marginal regions of the Asian monsoon is critical for understanding the millennial-scale interactions between the Asian monsoon and the westerly winds. Abrupt climate events are always associated with changes in large-scale circulation patterns; therefore, investigations into abrupt climate changes provide clues for responses of circulation patterns to extreme climate events. In this paper, we examined the time scale and mid-Holocene climatic background of an abrupt dry mid-Holocene event in the Shiyang River drainage basin in the northwest margin of the Asian monsoon. Mid-Holocene lacustrine records were collected from the middle reaches and the terminal lake of the basin. Using radiocarbon and OSL ages, a centennial-scale drought event, which is characterized by a sand layer in lacustrine sediments both from the middle and lower reaches of the basin, was absolutely dated between 8.0-7.0 cal kyr BP. Grain size data suggest an abrupt decline in lake level and a dry environment in the middle reaches of the basin during the dry interval. Previous studies have shown mid-Holocene drought events in other places of monsoon marginal zones; however, their chronologies are not strong enough to study the mechanism. According to the absolutely dated records, we proposed a new hypothesis that the mid-Holocene dry interval can be related to the weakening Asian summer monsoon and the relatively arid environment in arid Central Asia. Furthermore, abrupt dry climatic events are directly linked to the basin-wide effective moisture change in semi-arid and arid regions. Effective moisture is affected by basin-wide precipitation, evapotranspiration, lake surface evaporation and other geographical settings. As a result, the time scales of the dry interval could vary according to locations due to different geographical features.