Application Of Synchrotron Radiation X Ray Fluorescence Spectroscopy And ?-fourier Transform Infrared Spectroscopy In Reconstruction Of Paleoenvironment Of Maar Lake Sihailongwan

Synchrotron radiation X-ray fluorescence spectroscopy(SRXRF)analytical method was established for analyzing contents of Fe?Mn?Br and another ten kinds of elements in the sediments.Different normalization methods of elements quantifying were compared,results show that the normalization method of ROI 0?195 is the best method for analyzing K,Ca,Ti,Cr,Mn,Fe,Ni,Cu,Zn,and ROI of rayleigh is the best method for analyzing As,Br,Rb,Pb.The quantitative equations of Fe,Mn,Br and another ten kinds of elements were obtained in the study,R2 is above 0.74,relative standard deviation(RSD)is between 7%?25%.According to the high resolution data of element contents by SRXRF analyzing of varved sediments of Sihailongwan Maar Lake,we extracted the Factor of 1(F1)as an indictor of dry-wet climatic condition and Factor of 2(F2)as an indictor of warm-cold climatic condition using PCA method.The scoring curves of the two factors revealed the climate change during 10.8?15.2kaB.P.in Sihailongwan Maar Lake.The paleoclimate was characterized by cold and dry during the last glacial period(15.2?14.5kaB.P.),warm and wet in the Bolling-Allerod period(14.5?12.5kaB.P.),cold and wet in Younger Dry as period(12.5?11.5kaB.P.),warm and dry in the early Holocene(11.5?10.8kaB.P.).High resolution data can reveal paleoclimate events of high frequency(10?100a scale),the law of seasonal change of elements is also discovered.Relied on the high sensibility of Mid-infrared(MIR)spectral analysis and the advantage of chemometrics,a?-infrared spectrum method(?-FTIR)for analyzing the contents and indexes in varved sediments was established.We compared the different sampling methods and scanning modes,and came to a conclusion that sampling method of no compression and MIR scanning mode is propitious for constructing the models of indexes of GDGTs.The effects of analysis of GDGTs using Offset,Derivative correction,Spectral smoothing,Multiple scattering correction,and other spectral pretreatments were compared,and the numbers of factors of principal component regression(PCR)were also compared,the MIR models of MBT,CBT,BIT,i-GDGT-0 considering the parameters of calibration models and validation models were established.R2 of calibration models is above 0.62,Standard Error of prediction(SEP)of validation models is between 0.04?8.2,the average relative error of the validations of external samples is between3.9%?13.8%.We predicted MBT,CBT,BIT and any other indexes in the varved sediments of Maar lake Sihailongwan between 10.8?15.2kaB.P..This study shows the perspective of SRXRF and?-FTIR analysis in high-resolution research of paleoclimate,paleoenvironment changes.