宝鸡市城区地表灰尘重金属空间变异性研究

Spatial Heterogeneity and Interpolation Strategy of Heavy Metals in Urban Surface Dust of Baoji City

为探究宝鸡市城区地表灰尘重金属含量的空间变异特征,采用ArcGIS软件对宝鸡市城区地表灰尘重金属含量进行莫兰指数分析,运用GS+9.0软件对重金属含量进行变异性分析,并在ArcGIS地统计模块中对变异函数的分析结果进行插值分析.结果表明:测得的宝鸡市城区地表灰尘中Cd、Cu、Zn、Cr、As的平均值均大于陕西省背景值,说明宝鸡市城区存在重金属污染现象.As在空间上为强相关(Z=5.579),Cu、Zn、Cr在空间上均为相关(Z>1.65),Cd、Pb、Co、Ni在空间上均为弱相关(Z<1.65).变异函数拟合最优模型显示,Cr、Co、Zn为指数模型,Ni为高斯模型,Cd为球状模型.Cr、Co、Ni、Cd的变异主要受结构性因素影响,Zn的变异受结构性因素和随机因素共同影响.优化后的变异函数插值结果具有较高精度,确定性插值法与地统计法精度相当(ME接近于0).空间上相关且符合正态分布的数据选用地统计插值法,空间上相关但是不符合正态分布的数据选用确定性插值法,两种方法的精确度相当.

In order to explore the spatial variation characteristics of heavy metal content in urban surface dust of Baoji City,the paper analyzed the Moran’s I of heavy metal content in urban surface dust of Baoji City by using ArcGIS software,and analyzed the variation of heavy metal content by GS+9.0 software.The interpolation analysis of the analysis results of variation function was carried out in ArcGIS geostatistics module.The results show that the average values of Cd,Cu,Zn,Cr and As in urban surface dust of Baoji City are all higher than the background values of Shaanxi Province,which indicates that there is heavy metal pollution in the urban area of Baoji City.As is strongly correlated in space(Z=5.579),Cu,Zn and Cr are all correlated in space(Z>1.65),while Cd,Pb,Co and Ni are all weakly correlated in space(Z<1.65).The optimal fitting model of the variation function shows that Cr,Co,Zn fit exponential models,Ni fits Gaussian model,and Cd fits spherical model.The variation of Cr,Co,Ni and Cd is mainly affected by structural factors,while the variation of Zn is affected by both structural factors and random factors.The interpolation results of optimized variation function have high precision,and the accuracy of deterministic interpolation method is equivalent to that of geostatistical method(ME is close to 0).The geostatistical interpolation method is used for the data related to spatial and conforming to the normal distribution,the deterministic interpolation method is used for the data related to the spatial but not in accordance with the normal distribution.The accuracy of the two methods is equivalent.