正定矩阵因子法解析破罡湖底泥重金属源研究

Application of Positive Matrix Factorization on Source Apportionment of Heavy Metals in Sediments of Pogang Lake

为将正定矩阵因子模型(PMF)应用于湖泊底泥重金属源解析,探究其污染源识别方面存在的问题,并进一步提高其源解析的科学性,文章基于PMF模型,辅以主成分分析及污染源调查,对安徽破罡湖底泥重金属源解析进行了研究。结果表明:PMF解析所得各因子对该湖底泥中各重金属的贡献率几乎相同,不能根据贡献率差异进行污染源识别。污染源调查结果显示该湖底泥重金属污染源有3种,分别是农业、交通及工业污染源,而PMF解析显示该湖底泥重金属应有4种污染源,PMF解析可弥补人为划分污染源类型的不足。由PMF可知,农业、交通、自然母质源及工业源对破罡湖底泥总重金属的贡献率依次分别是18.2%、41.3%、10.5%及30.0%。PMF预测的Cu、Cd和Pb含量与实测值线性拟合R^2>99%,其余2种重金属Zn、Cr拟合R^2>75%。

In order to apply positive matrix factorization (PMF)on source apportionment towards heavy metal in lake sediment,explore problems of PFM on source identification and further improve PFM scientificity,PMF,principle component analysis (PCA)and pollutant source investigation were comprehensively applied on source apportionment of heavy metals towards sediment in Pogang Lake,Anhui Province.The results indicated that the factor contribution rates to each heavy metal calculated from PMF were similar,and it was hard to identify the heavy metal source according to them.Heavy metals in sediments of Pogang Lake mainly came from agriculture,transportation and industry based on the pollution sources investigation, while PMF calculation showed that heavy metal sources in sediments of the Lake should be ascribed to four species.PMF can offset the negligence of anthropogenic heavy metals source recognition.The contribution rates of agricultural sources,traffic sources,natural source and industrial source to the total heavy metals were 18.2%,41.3%,10.5% and 30%,respectively,using PMF.At the same time,the correlations of Cr,Cu,Zn,Cd,and Pb between the predicted value and the measured value were high,with more than 99% of R^2 for Cu,Cd and Pb and more than 75% of R^2 for Cr and Zn.