Driving factor, source identification, and health risk of PFAS contamination in groundwater based on the self-organizing map

Jingwen Zeng,Kai Liu,Xiao Liu,Zhongen Tang,Xiujuan Wang,Renchuan Fu,Xiaojun Lin,Na Liu,Jinrong Qiu
DOI: https://doi.org/10.1016/j.watres.2024.122458
2024-09-15
Abstract:The complex interactions between groundwater chemical environments and PFAS present challenges for data analysis and factor assessment of the spatial distribution and source attribution of PFAS in groundwater. This study employed spatial response analysis combining self-organizing maps (SOM), K-means clustering, Spearman correlation, positive matrix factorization (PMF) and risk quotient (RQ), to uncover the spatial characteristics, driving factors, sources, and human health risks of groundwater PFAS in the Pearl River Basin. The results indicated that the characteristics of PFAS in groundwater were classified into 16 neurons, which were further divided into 6 clusters (I-VI). This division was due to the contribution of industrial pollution (33.2 %) and domestic pollution (31.5 %) to the composition of PFAS in groundwater. In addition, the hydrochemical indicators such as pH, dissolved organic carbon (DOC), chloride (Cl-), and calcium ions (Ca2+) might also affect the distribution pattern of PFAS. The potential human health risk in the area was minimal, with cluster Ⅱ presenting the highest risk (RQ value 0.25) which is closely related to PFOA emissions from fluoropolymer industry. This study provides a theoretical basis and data support for applying of SOM to the visualization and control of PFAS contamination in groundwater.
What problem does this paper attempt to address?