Low-field magnetic susceptibility has been widely used to determine the pedostratigraphy of the Chinese loess/paleosol sequences. However, uncertainties remain in correlating between the loess magnetic susceptibility and the marine oxygen isotope records because susceptibility variations are affected by both global and local paleoclimatic changes. To provide a more sound paleoclimatic interpretation of magnetic susceptibility variations, age models across Marine Oxygen Isotope Stage (MIS) 5 for the Jiuzhoutai (JZT) and Yuanbao (YB) sections, western Chinese Loess Plateau, were constructed through an integrated approach by linking the major pedostratigraphic boundaries of the loess profiles to the SPECMAP oxygen isotope curve, and by correlating relative magnetic paleointensity records with both the SINT800 global paleointensity stack from marine sediments and 36Cl records from the GRIP ice core. Results indicate good correlation of SIRM 60 mT (a residual remanence of saturation isothermal remanent magnetization after a 60 mT alternating field demagnetization) variations between these two sites, which agree well with fluctuations in subtropical Atlantic sea surface temperatures. All cooling events recorded by ice-core and Atlantic marine sediments within MIS5 have counterparts in SIRM 60 mT. SIRM 60 mT is partially controlled by the degree of low-temperature oxidation, which is strongly temperature dependent. However, strong pedogenesis can decrease SIRM 60 mT due to further oxidation of partially oxidized magnetites above some critical points. Therefore, we propose that SIRM 60 mT is best suited to record paleotemperature changes in loess profiles from the western Chinese Loess Plateau, where pedogenesis is the weakest. Furthermore, by inter-profile correlation between the YB and JZT sections, we note that the seemingly uniform sub-paleosol unit with a broad susceptibility peak (previously assigned to MIS5c) between ∼34.4 and ∼37.4 m in the YB profile actually consists of two independent units (lower part of S1L1/MIS5b and S1S2/MIS5c). This indicates that susceptibility values can be strongly affected by local factors (e.g., mainly precipitation). Therefore, beside the simplistic traditional paleoclimatic interpretation of variations in loess susceptibility involving only cold/dry and warm/humid scenarios, cold/humid and warm/dry scenarios should also be considered.
Bibliographical noteFunding Information:
We are grateful to Dr. B. Carter-Stiglitz for collecting samples, and to Professor F.H. Chen for providing us with samples. We thank Dr. J. Beer for providing us with 36 Cl data sets. Thanks go to Professor A. Roberts and Dr. J. Stoner for their thoughtful and helpful reviews and discussion on this topic. This study is supported by National Science Foundation of China (NSFC) Grant 40221402, NSF Grant Nos. EAR 0003421 and EAR/IF 9818704. Fieldwork and sampling is partly supported by NSFC (Nos. 40325011 and 40125001). All rock magnetic measurements were made at the Institute for Rock Magnetism (IRM), which is supported by the W.M. Keck Foundation, the US National Science Foundation, and the University of Minnesota. This is IRM publication #0406.