黄土高原表土磁化率与气候要素的定量关系研究

黄土高原地区已建立多个表土低频磁化率-气候转换函数,为黄土古气候定量重建提供了关键方法。但是表土磁化率变化的气候控制因素的系统研究尚未开展; 同时,低频磁化率部分受到沉积作用的影响,需要寻找气候意义更明确的指标。我们系统采集了黄土高原及周边地区的表土样品,运用相关、回归和因子分析等方法,研究了不同气候要素及其季节变化对表土磁化率和频率磁化率的控制作用。结果显示,湿度是控制黄土高原表土磁化率的主要因素,温度的影响相对较小; 降水的季节分配也有影响,月降水变率较小的地区磁化率较高。在此基础上,选择反映成壤磁性颗粒组分的频率磁化率,建立了频率磁化率-年均降水量的转换函数,为黄土古气候定量估算提供了新的途径。

QUANTITATIVE RELATIONSHIPS BETWEEN MODERN SOIL MAGNETIC SUSCEPTIBILITY AND CLIMATIC VARIABLES OF THE CHINESE LOESS PLATEAU

Environmental magnetism has been widely employed to reconstruct past climate changes in the loess deposits on the Chinese Loess Plateau(CLP),and several climofunctions based on magnetic properties of loess have been developed.However,quantitative relationship between topsoil magnetic enhancement and modern meteorological parameters has not been systematically investigated,which is fundamental to decipher paleoclimate variations in loess-paleosol sequences using magnetic properties.The low-frequency magnetic susceptibility Xlf employed in most of published climofunctions comprises of magnetic signals of both detrital and pedogenic origins.It is still needed to develop climofunctions based on magnetic properties of pedogenic origin.Here we present statistic results of two magnetic parameters: low-frequency and frequency-dependent magnetic susceptibility,Xlf and Xfd,of modern soil samples on the CLP and eight meteorological variables,including mean annual precipitation(MAP),mean annual temperature(MAT),mean annual evaporation(MAE),annual range of temperature,monthly precipitation variability,MAP-MAE,MAP/MAE and MAE/MAP.The 257 modern soil samples,typically developed upon the L0 deposited on Middle Holocene paleosol S0,were collected across the main CLP region east to Lanzhou city and in adjacent regions.The sampling intervals are 2~5cm beneath the ground surface for each site,and any sites possibly disturbed or contaminated by human activities were carefully avoided.Xfd is better than Xlf in reflecting the ferrimagnetic minerals of pedogenic origin,both determined by measurements using a Bartington MS2 meter.Modern meteorological data were obtained by inverse distance interpolation based on the data of 648 meteorological stations all over China.77 of 257 modern soil samples were excluded in the statistic analysis for their extremely low Xfd and Xfd%,abnormal S -ratio,high altitude etc. Correlation analysis between magnetic susceptibility and meteorological parameters shows that MAP has the best correlations with Xlf,Xfd and Xfd%,with correlation coefficients are 0.62,0.64 and 0.76, respectively, and that MAT and monthly precipitation variability have rather good correlations as well.These indicate that the MAP,MAT and monthly precipitation variability are potential primary factors controlling the variation of Xlf,Xfd and Xfd%.Factor analysis on the modern meteorological parameters was used to extract non-collinear climatic factors,which shows that the total accumulative variance of first and second components’ reached 85.18%,high enough to represent overall climate features of CLP.These two factors are interpreted,by the component loadings,as "humidity" and "heat".Multiple regression analysis between the humidity and heat factors and Xlf,Xfd and Xfd% respectively shows that humidity factor(λ=5.075)has standardized coefficient values 0.726,0.738 and 0.799 with Xlf,Xfd and Xfd%,respectively,much higher than heat factor,0.409,0.422 and 0.392.The results indicate that the humidity factor mainly controls spatial distributions of magnetic susceptibility of pedogenic origin.Additionally,correlation analysis based on 64 soils in a region with similar rainfall range(550~600mm)shows that lower monthly precipitation variability corresponds to stronger magnetic signals,indicating that seasonal distribution of precipitation may also play a role.According to the analyses above,we developed a logarithmic MAP-Xfd transfer function(MAP=87.501×ln Xfd+360.636,R2=0.782,n=180),which is a better climofuction for paleorainfall estimates in the region with past MAP<1000mm.