Timing and structure of the weak Asian Monsoon event about 73,000 years ago

Oxygen isotope (delta O-18) records of cave speleothem have played an important role in the past two decades in characterizing the Asian Monsoon (AM) variability and correlating the monsoonal events with other global climate events on millennial timescales. Of a series of millennial events occurred during the last glacial period, the Chinese Stadial-20 (CS-20, corresponding to the Greenland Stadial-20, GS-20) event around similar to 73 kyr BP (thousand years before present, where present = 1950 AD) is distinctive, since it is the weakest AM event during the last glacial period and is likely linked to the Toba volcanic super-eruption. While Greenland ice core records are commonly used to correlate the last glacial millennial events, yet their absolute age uncertainties around the GS-20 is larger than 1000 years. This prohibits precise correlations of the event between global climate archives from different climate systems to investigate the underlying climatic dynamics. Here, we present three Chinese cave stalagmite delta O-18 records from the AM region, covering a period from 76 to 71 kyr BP, across the CS-20. All stalagmites have high uranium contents and relatively fast growth rates, allowing acquisition of high-resolution (similar to 10 years) delta O-18 records with precise Th-230 age controls (<= 200 years, 2 sigma) to precisely characterize the structure and timing of the CS-20. Our results demonstrate that the onset and termination of the CS-20 are more gradual in Chinese cave records relative to the GS-20 event in Greenland ice core records. As such, we suggest a new 'break-point' approach to correlate CS-20 with GS-20, at either the initial onset or initial termination shifts of the events, instead of the conventional 'mid-point' match. We dated the initial onset and initial termination of CS-20 to similar to 74.0 +/- 0.2 and 72.5 +/- 0.2 kyr BP, respectively, confirming the Greenland ice core chronology well within the quoted uncertainty. The 'break-point' correlation at the GS-20/CS-20 initial termination suggests a lagged onset of CS-20 relative to the onset of GS-20. The lagged onset is in line with a northern high-latitude forcing mechanism triggering the event and a central role of oceanic reorganizations in the propagation of the climate signal. An alternative 'break-point' correlation at the GS-20/CS-20 initial onset suggests that the CS-20 initial termination leads the GS-20 initial termination by a few hundred years. This apparent paradox thus calls for further empirical and theoretical studies to better understand the underlying climatic dynamics and in turn the correlation strategy.