Groundwater environmental forensic investigation combining multivariate statistical techniques and screening analyses

• Published in: Environmental forensics 2020-11, Vol.ahead-of-print (ahead-of-print), p.1-15
• Main Authors: Sutliff-Johansson, S., Pontér, S., Mäki, A., Engström, E., Rodushkin, I., Peltola, P., Widerlund, A.
• Format: Article
• Language: English
• Subjects: Applied Geochemistry; groundwater contamination; heavy metals; multi-elemental screening; multivariate statistical techniques; principle component analysis; Rönnskär; Tillämpad geokemi
• ISSN: 1527-5922; 1527-5930; 1527-5930
• DOI: 10.1080/15275922.2020.1850571

Heavy metal contamination was identified in groundwater monitoring wells surrounding a waste deposit facility at the Rönnskär Cu-Pb-Zn smelter in Skellefteå, northern Sweden. The main objective of this study is to identify the sources of contamination, utilizing element screening analyses and multivariate statistical techniques. A second objective is to determine the usefulness of these techniques in Environmental Forensics investigations of contaminated groundwater at a complex industrial site. Water samples were collected from four groundwater monitoring wells and six waste deposit cells surrounding the contaminated area. Seventy-two elements are statistically examined and the dataset is reduced to the variables representative of the contaminated source material from the smelting process. A three-component model is identified and explains 88% of the total variation in the dataset. Component 1 includes concentrations of Cd, Co, Ni, Rb, Re, and Zn. This component displays a high correlation with two of the deposit cells and their associated groundwater monitoring wells. Component 2 is comprised of Sb, Cu, and Mo. This component displays a correlation between all monitoring wells and deposits likely due to the high mobility of these elements as oxyanions. Component 3 is dominated by As and displays high correlation to three older deposit cells representing a completely different source than for Components 1 and 2. The application of screening analyses and multivariate statistics in this study has achieved a meaningful identification of sources of contamination in the investigated area. It was also shown to be useful as an initial survey aiming to optimize a full-scale monitoring program at the site.