新疆天山黄土GDGTs重建的全新世温度逐步升高及其可能意义

Holocene temperature changes in arid Central Asia revealed by GDGTs of loess-paleosol sequence in Tianshan Mountains

全新世温度变化是理解现代变暖和预测未来温度变化的重要依据,但目前根据地质记录重建的温度变化趋势与气候模拟重建的温度变化趋势存在显著矛盾。依据已经建立了可靠年代的新疆天山北麓鹿角湾全新世黄土—古土壤序列(LJW10剖面),选取4个重要时段的黄土—古土壤样品,开展支链GDGTs(br-GDGTs)测量,采用国际上新提出的GDGTs分析方法和MBT'5ME指标,重建了该剖面全新世以来年均温度变化的初步框架。结果发现,采用新的MBT'5ME指标重建的表层样品的温度更接近现代气象记录的实际温度,利用国际新的GDGTs分析及计算方法重建的全新世温度自早全新世以来逐步升高,与已有气候模拟重建的全新世温度变化趋势一致。应当指出的是,本文只是使用新GDGTs分析方法和MBT'5ME指标重建全新世温度变化的初步探讨,今后不但需要开展研究方法的进一步探讨,而且需要开展更高时间分辨率和不同区域黄土—古土壤记录的GDGTs温度重建研究。

The study of temperature changes during Holocene plays an important role in understanding current global warming and forecasting temperature changes in the future. However, the temperature change trend during Holocene inferred from proxy index reconstructions are contradicted with those from climate model stimulation. In this study, four samples from a loess-paleosol sequence （LJW10） in Tianshan Mountains, arid Central Asia, were used to reconstruct annual temperature trend during Holocene. The proxy index of MBT＇ 5ME and trans- fer function derived from GDGTs of those samples using newly proposed analytical approach were used for temperature reconstruction. As a result, we achieved separation of 6-methyl br-GDGTs using improved chromatogra- phy, the temperature inferred from MBT＇ 5ME index of the four samples was -6.3 ℃, 1.9 ℃,4.3 ℃ and 6.6 ℃ respectively at the corresponding depth of 2.2 m, 1.6 m, 1.2 m and 0.6 m. The reconstructed temperature at the depth of 0.6 m was in accordance with the modem meteorological record, which indicated our temperature recon- struction is reliable. Importantly, we find the low temperature phenomenon in mid-Holocene and the gradual tem- perature increasing trend since Holocene, which are consistent with results from the climatic modeling. It helps understand Holocene temperature conundrum induced by the records and the climate model. However, there is a significant difference on temperature trend during Holocene between our results and the results based on br- GDGTs using Chinese Loess Plateau （CLP）. This may be due to different distribution of seasonal rainfall, which caused the temperature recorded by br-GDGTs was about summer in CLP but annual average temperature in Tianshan Mountains. The seasonal response ofbr-GDGTs to temperature still can not be ignored. It should be also noted that this study is only a preliminary step in the research of using GDGTs and MBT＇ 5ME index to reconstruct the temperature change in Holocene and the next step is to continue with this model using higher temporal resolution.