Research On The Theory And Methods Of Determining Geochemical Background Values Of Soil Heavy Metals

Determining the geochemical background value of soil heavy metals is the prerequisite and basis for evaluating soil environmental quality,is the necessary data for judging the degree and trend of soil heavy metal pollution,and is of great significance for environmental protection and rational distribution of industrial and agricultural production.However,the geochemical background of heavy metals in soil is not fixed;it varies from place to place,which makes it difficult to determine the background value.This study takes 1:250,000 scale soil geochemical data from the central and western regions of Jilin Province as an example,analyzing the root cause of soil heavy metal background variation from the frequency distribution of heavy metal content and the correlation between heavy metal and soil chemical composition,respectively.Two methods for determining the geochemical background value of heavy metals in soil are proposed:the multiple population separation method and the chemical composition regression method.Practical examples were carried out to verify the effectiveness of the two methods,and the advantages and disadvantages and applicable conditions of the two methods were summarized to provide a reference for the background research of soil heavy metal geochemistry.The primary knowledge obtained in this paper is summarized as follows:（1）The soil samples in the central and western regions of Jilin Province were classified according to soil type,parent material,geological background,geomorphic characteristics,and eco-geological environment.The frequency distribution of heavy metal content in different types of soils was studied in detail,and the geochemical background values of heavy metals in different types of soils were calculated.The results showed that heavy metal background values of different types of soils varied greatly,and the background values of heavy metals in the same type of soil were not all fixed,as were the results of the other four indicators.The above research shows that soil type,soil parent material,geological background,geomorphic characteristics,and eco-geological environment are not the fundamental controlling factors for soil heavy metal background variation.（2）From the viewpoint of chemical composition,the fundamental controlling factors of soil heavy metal background variation were deeply analyzed,and the control effect of soil chemical composition was discussed in the frequency distribution of heavy metal content and the correlation between heavy metal and soil chemical composition.The result showed that soil chemical composition controls the content of heavy metals and is the underlying cause of heavy metal background variation.The control effect is reflected in two aspects:the frequency distribution of soil heavy metal content is skewed（left,right,or multi-modal distribution）,and some heavy metals are correlated with soil chemical composition（Al₂O₃,Fe₂O₃,La,Co,Ti,V,Y,Sc,etc.）.There was a significant correlation at 0.01 level.This research lays a theoretical foundation for solving the problem of soil heavy metal background variation.（3）Based on the study of the frequency distribution of soil heavy metal content,a method for determining soil heavy metal background value based on multiple population separation was proposed.Taking the surface soil geochemical data in the northern part of Changchun as an example,the k-means and the EM（Expectation-Maximization）algorithms were used to separate multiple populations by sample clustering.The background values of heavy metals were calculated in each sample cluster,and the heavy metal risk areas were delineated.The results showed that the background value of heavy metals determined after the separation of the population is more reasonable,which is mainly reflected in the identification of the subtle risk area of heavy metals in low-background areas and the elimination of the false risk area of heavy metals in high-background areas.In comparison,the population separation result of the EM algorithm is more reliable.（4）Based on the correlation analysis between heavy metals and soil chemical composition,a method for determining the background value of heavy metals in soil based on chemical composition regression was proposed.Taking the surface soil geochemical data of the Songhua River basin（Jilin Province）as an example,the partial least squares regression method and four machine learning regression algorithms were used to regress the background values of soil heavy metals.Compared with the traditional method of determining the uniform background value of the whole region,the chemical composition regression method can determine the background value of heavy metals in each sample more precisely,and the artificial input of heavy metals in each point can be calculated accordingly.The Gaussian process regression algorithm is the most outstanding among these regression methods.（5）Taking residual soil geochemical data in central Jilin Province as an example,this study analyzes the application effects of the EM and Gaussian process regression algorithms.The results showed that the background values obtained by the chemical composition regression method are more precise,but they are not suitable for those heavy metals that have no or only weak correlation with soil chemical composition.The multiple population separation method does not have the above problems,but it is only a population separation method and cannot directly calculate the background value of heavy metals.The two approaches do not conflict and can complement each other.In the subsequent background research work,the appropriate background value determination method can be selected according to the amount of data,the complexity of regional geology,the items tested,and the types of elements.Based on the study of the variation of soil heavy metal background,two methods for determining soil heavy metal background value were proposed in this study,namely the multiple population separation method and the chemical composition regression method,and the effectiveness of these two methods was verified by practical application.The research results will solve the problem of soil heavy metal background variation to a certain extent,improve the accuracy of background value calculation,and provide an essential basis for resource exploration,environmental investigation,and agricultural scientific research.