Application of Multivariate Statistical Approach to Identify Heavy Metal Sources in Sediment and Waters: a Case Study in Yangzhong, China

Jian Zhou,Dongsheng Ma,Jiayong Pan,Wenming Nie,Kai Wu
DOI: https://doi.org/10.1007/s00254-007-0824-5
2007-01-01
Environmental Geology
Abstract:Multivariate statistical approach is used to identify the sources of heavy metals (Bi, Cd, Co, Cr, Mn, Pb, U, V, and Zn) in surface water and freshly deposited riverine sediment samples in Yangzhong city, China. The metal concentration data for the water and sediment samples are reported in terms of basic statistical parameters and metal-to-metal correlations. In both surface water and sediment samples, significant correlations are observed between some metals. Principal component analysis and cluster analysis distinguishes factors of lithogenic and anthropogenic origin. Bismuth, Cd, Co, and Pb (Co only for water samples) contents are controlled by the regional lithogenic high background factor; Co, Mn, U, and V (Co only for sediment samples) are interpreted to be mainly inherited from soil parent materials, while Cr, Zn, and Mn in the two kinds of samples are recognized as the tracer of industrial pollution. Obvious similarity between factor loadings of the two kinds of samples is observed, evidencing that metal variability in the two kinds of samples is controlled by the same sources. Statistical analysis agrees with discussion based on background value and field survey of point-source pollutant affected sediment, making this statistical discussion more convincing.
What problem does this paper attempt to address?