【作者】 李思米；

【导师】 卞新民；

【作者基本信息】 南京农业大学 ， 生态学， 2005， 硕士

【摘要】 农田土壤重金属污染是与备受关注的食品安全问题、人类健康等密切相关的问题。本论文以空间插值方法为主要的研究手段,针对南通市随机布点的农田土壤样本进行空间插值方法和农田土壤重金属污染空间分布的研究。 本论文第二章详细介绍了本次采样的布点实施情况。第三章介绍了各种常用的空间插值技术。并深入论述了空间插值理论的基础——空间统计方法。在本论文中采用了两种空间插值技术:反距离插值方法和普通克吕格方法。其中,克吕格法中我们选用了球状和指数两种模型。并在第五章详细对比了这三种方法对南通市随机布点的数据的拟合效果。 第四章对实际数据进行数据探索发现5种重金属因子在南通市农田土壤中的含量差异突出,As、Cr、Pb三种重金属含量的数值范围广,数据离散程度大;Cd和Hg含量的数值范围小,数据离散程度较小。并且五种重金属在土壤中的含量分布都不符合正态分布。第五章为了验证这5种重金属的数据特性是否会影响空间插值的精度,并且为了最大程度的减小由随机布点和土壤性质等原因造成的系统误差,本论文采用3中权威模型,对各模型参数进行梯度调整,分别对数据源的各种重金属因子进行空间插值运算,所得的结果进行交叉检验。对原始样点和估测样点求其标准差、显著性概率并计算所用模型的拟合系数,结合Arcgis系统输出拟合后得到的块金值,对不同的重金属确定最优的拟合模型。 第六章运用测试所确定的数据拟合模型,分别对南通市的As、Cd、Cr、Hg、Pb五种重金属进行插值运算。输出图像可以直观的反映南通市各种农田重金属的空间分布情况,并运用插值得到南通市的农田土壤重金属含量的各级等值线。由等值线的分布情况,总结出各重金属在南通市的分布趋势。结合空间插值运算,本论文对南通市农田土壤重金属进行单因子评价和多因子综合评价。单因子评价使用的是标准污染指数法,多因子评价使用的是半集均方差方法。半集均方差方法,能够给予所有评价点上的大值以大的权重,选用的原因主要是考虑到土壤重金属污染是比较值得关注的农更多还原

【Abstract】 Heavy metal pulltion of farmland is in relation to the attended problems such as food safety and human health. In this thesis, the spatial interpolation method was mainly used to analyse the heavy metal content in the farmland soil samples came from Nantong set with the random sampling method.In the second chapter of the thesis, the sampling method was described in detail. And the interpolation techniques in common use was introduced in the third chapter, as well as the basic theory of spatial interpolation, the geostatistics. Two kinds of spatial interpolation technique were used in this thesis: the IDW( Inverse distance weighted) and Ordinary Kriging. And we choose two models of Ordinary Kriging, one is the Spherical and the other is Exponential model. Thses three model were particularly tested in the fifth chapter to compare with each other the effect of the agreement of Nantong sample datum.In the forth chapter 5 heavy metal factors’original datum was explored thoroughly. And we found the 5 heavy metal’s content in the soil was different from each other. As, Cr, and Pb has a wider scale of datum with a considerable discrete degree. Just the opposite the Cd and Hg’s content in soil is much lower than the others. The scale of data is small with little discrete degree. And the distributing of each heavy metal’s data disagree with normal distribution 。 It was proved in the fifth chapter, whether the datum’s structure will be a infulence to the model agreement’s precision. In order to make the system error be the least which bringed by the random sampling and something in clued in the soil structure, the thesis tested the IDW, Ordinary Kriging Spherical model and Ordinary Kriging Exponential model, in the parameters grads. And the agreement results were cross-validated. The Std.error, significance probability and the agree index were calculated after the cross validation. The Nugget was also considered with those index when we decide which is the best model with the proper parameter to each of the 5 heavy metal factors.Apply the test-decided model, interpolate to As, Cd, Cr, Hg, Pb in the Arcgis software. The results were output by the images, so the spatial distribution of each heavy metal’s content can be displayed directly. Also the content’s isoline can show the distribution tendence of each heavy metal in Nantong. Combined with the spatial interpolate techniques the thesis appraised the heavy metal pollution of Nantong farmland soil, both in the single factor and the multi-factors way. The single factorappraisal choose the standard pollution index method. In consideration of the importance of farmland soil polltion the thesis select the semi-variance as the multi-factors appraisal method. In this appraisal method the weighted heavy metal factor’s content will be given the bigger power in any one sampling point.更多还原