欢迎访问《环境科学学报》编辑部网站！

快速检索			高级检索

研究报告

A comparative study of different assessment methods of regional heavy metal pollution

摘要：根据昆山市2km×2km网格土壤采样重金属测试数据,选取As、Cd作为代表元素,以地累积指数为污染指数,运用简单数理统计、正态模糊数与核密度估计对区域土壤重金属总体污染程度进行了评价,系统比较了不同评价方法的结果差异.结果表明:评价的便捷性上,3种方法的排序是简单数理统计>>正态模糊数法≥核密度估计;结果的准确性上,与简单数理统计相比,正态模糊数法和核密度估计能较敏感地显示研究区域分布极少的污染等级区域,3种方法下研究区As总体污染评价的平均地累积指数相对于参照值的偏差分别为19.2%、19.2%、15.4%,Cd的分别为14.3%、14.3%、10.7%,准确度排序为核密度估计>正态模糊数法=简单数理统计,结果所包含信息的全面性上,为正态模糊数法>>核密度估计=简单数理统计,应用正态模糊数法评价能得到表征总体污染程度的区间数,核密度估计与简单数理统计只能得出唯一值.

Abstract：We obtained 2 km×2 km heavy metal concentration data in Kunshan by field sampling and indoor test. By selecting As and Hg as typical elements and geo-accumulation index as pollution index, we applied simple mathematical statistics, normal fuzzy numbers, and kernel density estimation to evaluate regional soil heavy metal contamination level, and made intercomparison between evaluation results of these methods. Simple statistical evaluation showed the highest evaluation convenience, followed by normal fuzzy numbers evaluation and kernel density estimation. Normal fuzzy numbers evaluation and kernel density estimation could more accurately display the pollution degrees. The difference between average geoaccumulation index and the reference was 19.2%, 19.2% and 15.4% for As and 14.3%, 14.3% and 10.7% for Cd with the three methods, respectively. The results of kernel density estimation had higher accuracy than normal fuzzy numbers evaluation and simple statistical evaluation, while normal fuzzy numbers evaluation contained more information than kernel density estimation and simple statistical evaluation. Normal fuzzy numbers evaluation can obtain the number of intervals characterizing the overall pollution levels, which kernel density estimation and simple statistical evaluation can only get a unique value.

Key words：evaluation of heavy metals simple mathematical statistics normal fuzzy numbers kernel density estimation systematic comparison