Characterization of heavy-metal-contaminated sediment by using unsupervised multivariate techniques and health risk assessment

Abstract

This study characterized the sediment quality of the severely contaminated Erjen River in Taiwan by using multivariate analysis methods—including factor analysis (FA), self-organizing maps (SOMs), and positive matrix factorization (PMF)—and health risk assessment. The SOMs classified the dataset with similar heavy-metal-contaminated sediment into five groups. FA extracted three major factors—traditional electroplating and metal-surface processing factor, nontraditional heavy-metal-industry factor, and natural geological factor—which accounted for 80.8% of the variance. The SOMs and FA revealed the heavy-metal-contaminated-sediment hotspots in the middle and upper reaches of the major tributary in the dry season. The hazardous index value for health risk via ingestion was 0.302. PMF further qualified the source apportionment, indicating that traditional electroplating and metal-surface-processing industries comprised 47% of the health risk posed by heavy-metal-contaminated sediment. Contaminants discharged from traditional electroplating and metal-surface-processing industries in the middle and upper reaches of the major tributary must be eliminated first to improve the sediment quality in Erjen River. The proposed assessment framework for heavy-metal-contaminated sediment can be applied to contaminated-sediment river sites in other regions.