摘要
采用长江水系约２６０个背景样点沉积物原样和小于６３μｍ细粒沉积物中１１种重金属元素（铜、铅、锌、镉、汞、钴、镍、砷、铬、锰和铁）含量的数据，研究了数据的频率分布特征和变异系数，比较了算术均值、几何均值和中位数与由Ｂｏｘ－Ｃｏｘ变换而得到的对称均值间的大小关系。结果表明，大多数重金属元素含量服从对数正态分布；采用算术均值表示均值会有较大的偏差，且这种偏差随变异系数的增大而增大；采用中位数偏差较小，但效率低；采用几何均值偏差也较小，在数据服从对数正态分布条件下可以使用，但不是最优；而采用对称均值来表示均值，由于分布的对称性，而最具有代表性，本研究采用稳健统计与Ｂｏｘ－Ｃｏｘ变换相结合的方法，得到稳健对称均值，并采用此法表示长江水系河流沉积物中重金属元素含量的均值。
Contents of 11 heavy metals(Cu,Pb,Zn,Cd,Hg,Co,Ni,As,Cr,Mn and Fe)inthe sediments of the Changjiang River system were used for background contents calcula-tion. About 260 locations were selected,for sampling of both raw sediments and the finegrained parts(<63 μm)。The frequency distribution and coeffieients of variation(Cv)werecalculated。Several methods were used and compared to calculate mean contents. Contentsof most heavy metals follow log-normal distribuiton,Geometrical means generated relativelysmall biases for log-normally distributed data,but it is not optimal. Symmetric means,based on Box-Cox transformation,are strongly recommended.In this study,a combinationof robust statistics and Box-Cox transformation were used to produce robust-symmetricmeans. The robust-symmetric means for heavy metal contents in sediments of theChangjiang River system are the following(in ppm):fot raw sediments,Cu 21.5,Pb21.4,Zn73.6,Cd0.148, Hg0.034,Co12.1, Ni26.4, As7.6,Cr52.3,Mn589,andFe 2.94 (%);and for the fine grained parts(<63 μm),Cu 27.0,Pb 26.1,Zn 85.9, Cd0.227,Hg 0.039,Co 13.8,Ni 29.1,As9.0,Cr 57.1,Mn 673,and Fe 3.31(%).
出处
《环境科学学报》
CAS
CSSCI
CSCD
北大核心
1995年第3期258-264,共7页
Acta Scientiae Circumstantiae
基金
"七.五"国家科技攻关环保项目
关键词
长江水系
沉积物
均值
重有色金属
计算方法
Changjiang River,sediments,heavy metals,mean,robust statistics,data transformation.
