Concept: Absolute dating
The loess stratigraphic boundary at the Pleistocene/Holocene transition defined by the magnetic susceptibility (MS) has previously been assumed to be synchronous with the Marine Isotope Stage (MIS) 2/1 boundary, and approximately time-synchronous at different sections across the Chinese Loess Plateau (CLP). However, although this assumption has been used as a basis for proxy-age model of Chinese loess deposits, it has rarely been tested by using absolute dating methods. In this study, we applied a single-aliquot regenerative-dose (SAR) protocol to the 45-63 μm quartz grain-size fraction to derive luminescence ages for the last glacial and Holocene sections of three loess sections on a transect from southeast to northwest across the CLP. Based on the 33 closely spaced optically stimulated luminescence (OSL) samples from the three sections, OSL chronologies were established using a polynomial curve fit at each section. Based on the OSL chronology, the timing of the Pleistocene/Holocene boundary, as defined by rapid changes in MS values, is dated at ~10.5 ka, 8.5 ka and 7.5 ka in the Yaoxian section, Jingchuan and Huanxian sections respectively. These results are clearly inconsistent with the MIS 2/1 boundary age of 12.05 ka, and therefore we conclude that the automatic correlation of the Pleistocene/Holocene transition, as inferred from the MS record, with the MIS 2/1 boundary is incorrect. The results clearly demonstrate that the marked changes in MS along the southeast to northwest transect are time-transgressive among the different sites, with the timing of significant paleosol development as indicated by the MS record being delayed by 3-4 ka in the northwest compared to the southeast. Our results suggest that this asynchronous paleosol development during the last deglacial was caused by the delayed arrival of the summer monsoon in the northwest CLP compared to the southeast.