• 关键词：土壤污染调查  地统计条件模拟  污染概率  局部空间变异  污染区范围  布点优化

• 基金项目：国家自然科学基金(No.41301229);国家环境保护公益性行业科研专项(No.201409047)

• 谢云峰
• 中国环境科学研究院, 环境基准与风险评估国家重点实验室, 北京 100012

• 曹云者
• 中国环境科学研究院, 环境基准与风险评估国家重点实验室, 北京 100012

• 杜晓明
• 中国环境科学研究院, 环境基准与风险评估国家重点实验室, 北京 100012

• 徐竹
• 中国环境科学研究院, 环境基准与风险评估国家重点实验室, 北京 100012

• 柳晓娟
• 中国环境科学研究院, 环境基准与风险评估国家重点实验室, 北京 100012

• 陈同斌
• 中国科学院地理科学与资源研究所环境修复中心, 北京 100101

• 李发生
• 中国环境科学研究院, 环境基准与风险评估国家重点实验室, 北京 100012

• 杜平
• 中国环境科学研究院, 环境基准与风险评估国家重点实验室, 北京 100012

• 摘要：鉴于传统调查布点方法对土壤污染物平均含量的估计精度较高,但对污染区范围的估计精度不能满足修复治理需求的问题,同时,为了准确估计土壤污染区面积及其空间位置,提出了土壤污染详查样点优化方法.在初步调查的基础上,首先利用地统计条件模拟方法预测土壤污染概率,基于污染概率和土壤污染物含量局部空间变异确定加密布点的优先区域,并根据污染物含量的空间变化趋势布设样点,然后根据优化后的土壤污染调查布点方案估计污染区面积和空间位置.最后,以某Cd污染场地为例,验证了布点优化方法的效果.结果表明,土壤污染调查加密布点方法显著提高了污染区面积及其空间位置的估计精度.案例场地土壤Cd污染区面积的预测误差为4.10%,污染区空间位置的精度为86.10%.土壤污染调查布点优化方法在保证土壤污染调查精度的同时,相比于传统调查方法可显著降低土壤污染调查的样本量.

• Abstract：Accurate estimation of soil contaminated area and its spatial distribution is one of the main purposes in soil pollution investigation. Traditional sampling design methods (random sampling, systematic sampling and so on) have good performance in estimating the average concentration of contaminants in soil. However, the accuracy of estimated contaminated area is insufficient for site remediation. In order to improve the estimate accuracy of contaminated area, a sampling design optimization method was proposed. Based on the preliminary soil investigation data, probability of soil contamination was predicted using geostatistical conditional simulation (i.e., sequential gaussian simulation), while local spatial variation of the soil contamination was depicted by the coefficient of variation. Priority areas for detailed soil investigation were delimited according to the contamination probability and local spatial variation of the pollution concentration. Finally, additional samples were placed according to spatial trend of the soil contaminants determined by spatial pattern analysis. The method was then evaluated and verified at a cadmium contaminated field. Results indicated that the proposed sampling optimization method improved the estimation accuracy of polluted area and spatial location significantly. Relative error of the predicted area of the Cd contamination soil was 4.1%. Precision of the spatial location of the contaminated area was 86.10%. Compared to traditional sampling, the optimized method needs 57.1% less of the number of samples.

• Key words：soil pollution investigation  geostatistical conditional simulation  probability of contamination  local spatial variability  extent of contaminated soil  sampling design optimization

• 
  摘要点击次数： 1645 全文下载次数： 2795