SIX MILLION YEARS OF TEMPERATURE AND RAINFALL HISTORY ON THE CHINESE LOESS PLATEAU: EVIDENCE FROM A NOVEL APPLICATION OF TWO MAGNETIC GRAIN SIZE PROXY PARAMETERS TO THE CHINESE LOESS RECORD

Rock magnetic proxies are arguably the most widely-used paleoclimate proxies in Chinese loess studies. However, most previous studies emphasize using magnetic concentration proxies to reconstruct paleoclimate history from loess. Magnetic grain size variations are also sensitive paleoclimate proxies but their climatic implications are not entirely clear due to a lack of calibration to modern soils. In this study, we report relationships between two magnetic grain size proxies (χ_ARM/χ_LF and χ_ARM/SIRM) and two climate parameters (annual mean temperature and rainfall) for surface soils across the Chinese Loess Plateau (CLP). We find that χ_ARM/SIRM is sensitive to rainfall variations, but not to temperature variations. In contrast, χ_ARM/χ_LF is sensitive to both rainfall and temperature variations. Thus, applying these two proxies to the loess and the underlying red-clay sequences has the potential to obtain both the paleo-rainfall and paleo-temperature history on the CLP since the late Miocene.

Indeed, such reconstructions indicate that the paleo-temperature of the CLP area is coherent with ice volume changes since 6 Ma. The interval ~5.5-~5 Ma is a warm period on the CLP and a cooling trend occurs from ~4.6 to ~2.7 Ma. In contrast, although rainfall follows ice volume changes during the interval 6 to ~4.6 Ma, rainfall has an intensification trend from ~4.6 to ~2.7 Ma, likely resulting from tectonic forcing of the Asian monsoon. During the last 2.7 Myr, rainfall and temperature basically co-vary with ice volume. Considering the coherent relationship between low latitude sea surface temperature and global ice volume, we attribute the coherent relationship between global ice volume and CLP temperature during the last 6 Myr to atmospheric CO₂ forcing. However, ice volume probably had a direct effect on CLP rainfall via modulating sea level and thus the position of the East Asian summer monsoon. Our studies demonstrate that contrasting/comparing the χ_ARM/χ_LF and χ_ARM/SIRM proxy records from Chinese loess can resolve both the paleo-rainfall and paleo-temperature history of the CLP region. This result has implications for long-term paleoclimate studies on the CLP and loess paleoclimate studies in other parts of the world.