Geographic variations in particle size distribution of the last interglacial pedocomplex S1 across the Chinese Loess Plateau: Their chronological and pedogenic implications

Due to northwestward attenuation of the summer monsoon and northwestward intensification of loess deposition during the last interglacial, the last interglacial pedocomplex S1 gradually differentiated from the northwest to the southeast. The three paleosols (S1S1, S1S2, S1S3) corresponding to the marine isotope sub-stages 5a, 5c, and 5e and the two intercalated loess units (S1L1, S1L2) corresponding to the marine isotope sub-stages 5b and 5d are completely preserved at northwestern sections. Towards southeast, both the S1L1 and S1L2 were annexed by the subsequent paleosol development (S1S1 and S1S2) and the three paleosols (S1S1, S1S2, and S1S3) were partially welded. At the southeasternmost site, the three soil-forming events (S1S1, S1S2, and S1S3) repeatedly occurred in a single paleosol profile. The three observed orders of particle-size variations are interpreted to have imprinted important chronological and pedogenic signatures. First-order variations, i.e., a remarkable difference between the interglacial pedocomplex S1 and the glacial loess units (L1 and L2), implies that the S1 parent material was considerably finer in the source areas or/and the proximity to the source areas was much farther during interglacial period than during the preceding (L2) and following (L1) glacial periods. Second-order variation, i.e., the parenthetical trends in > 63 μm and < 10 μm fraction curves, might have resulted from the delayed response of the source material supplies to the climate changes. The third-order variations in < 10 μm fraction correspond well to the variations in frequency-dependent magnetic susceptibility, suggesting that pedogenically formed finer particles (i.e., < 10 μm fraction) and the associated ultra-fine paramagnetic minerals as expressed by the frequency-dependent susceptibility occurred only as a minor component of the < 10 μm fraction. The third-order variations seem to be obscured by soil welding and annexation at southeastern sections where the paleosols within the S1 were partially or completely welded. The laboratory data-indicated < 10 μm fraction peaks of the third-order and field-observed clay coatings on ped-faces, together with carbonate leaching and accumulation, in the paleosols within the S1 indicate occurrence of within-S1 material translocation. To sum up, it is unrealistic to reconstruct high-resolution climatic records from the S1 pedocomplex because a number of factors might have undermined the validity of the particle size as a winter monsoon proxy. These factors include weathering in the source areas, in situ post-depositional weathering and fine fraction translocation, and downward penetration of soil formation into underlying previously deposited materials on stable land surfaces.