摘要
Widespread and thick loess deposit in the northeastern Tibetan Plateau records paleoclimatic changes over the Quaternary period. In this study, we investigate a loess section located in a high terrace of the Huangshui River in the Xining Basin, and collect samples at 20-cm intervals by digging a well of 16-m depth （including paleosol S1 and SO and several loess interlayers over the last glacial-interglacial cycle and the Holocene）. Concentration of some active elements （Sr and Ca） and inert elements （Rb, Zr, Ti） in these loess samples are measured. The results show that ratios of those elements （oxide）, such as Rb/Sr, Ti/Sr, and Zr/CaO ratios, would indicate sensitively paleoclimatic changes, especially the history of precipitation over the last glacial-interglacial and the Holocene. Moreover, the curves of Rb/Sr, Ti/Sr, and Zr/CaO ratios had almost the same variable trends from the bottom to the top and also closely correlated with values of magnetic susceptibility and the global marine δ^18O record. But compared with the curves of magnetic susceptibility, the changes of ratios of these elements （oxide） are clearer and easy to understand. Therefore, the ratios of these typical chemical elements in the loess deposit may be used as an excellent index to investigate the history of paleoclimate, particularly precipitation changes in the northeastern Tibetan Plateau.
Widespread and thick loess deposit in the northeastern Tibetan Plateau records paleoclimatic changes over the Quaternary period. In this study, we investigate a loess section located in a high terrace of the Huangshui River in the Xining Basin, and collect samples at 20-cm intervals by digging a well of 16-m depth （including paleosol S1 and SO and several loess interlayers over the last glacial-interglacial cycle and the Holocene）. Concentration of some active elements （Sr and Ca） and inert elements （Rb, Zr, Ti） in these loess samples are measured. The results show that ratios of those elements （oxide）, such as Rb/Sr, Ti/Sr, and Zr/CaO ratios, would indicate sensitively paleoclimatic changes, especially the history of precipitation over the last glacial-interglacial and the Holocene. Moreover, the curves of Rb/Sr, Ti/Sr, and Zr/CaO ratios had almost the same variable trends from the bottom to the top and also closely correlated with values of magnetic susceptibility and the global marine δ^18O record. But compared with the curves of magnetic susceptibility, the changes of ratios of these elements （oxide） are clearer and easy to understand. Therefore, the ratios of these typical chemical elements in the loess deposit may be used as an excellent index to investigate the history of paleoclimate, particularly precipitation changes in the northeastern Tibetan Plateau.
