A novel integration of self-organizing maps and NETPATH inverse modeling to trace uranium and toxic metal contamination risks in West Mecsek, Hungary
Mohamed Hamdy Eid, Zsolt Pinjung, Viktória Mikita, Czímer Bence, Péter Szűcs
Abstract
This study presents a novel integration of Self-Organizing Maps (SOM) and NETPATH inverse modeling to assess uranium (U) and toxic metal contamination risks in the West Mecsek, Hungary, a historically significant uranium mining area. Groundwater samples were analyzed from five aquifers (Basement, Boda Claystone Formation (BCF), Permian-Triassic (PT), Karst, and Neogene) to identify contamination sources, quantify aquifer mixing via fault/fractures-driven connectivity, and evaluate non-carcinogenic health risks in two different age groups. Hydrochemical facies from Piper and Chadha diagrams reveal dominant Ca-Mg-HCO₃ water types in the Neogene, karst, and evaporite aquifers, attributed to carbonate dissolution. In contrast, Na-HCO₃ and Na-Cl types in BCF/PT aquifers suggest silicate weathering, ion exchange, and anthropogenic contamination from mining activities. NETPATH mixing models quantify significant recharge from karst (≤66 %) and BCF (up to 82 %) into the Neogene aquifer, facilitated by fault-controlled flow paths. Isotopic analysis (δ¹⁸O–δ²H) differentiates modern recharge in Neogene and karst aquifers from paleo-waters in BCF and basement units. Water Quality Index (WQI) analysis classifies PT and evaporite waters as excellent quality, while BCF and karst aquifers show localized high-risk zones. Health risk assessments identify arsenic, uranium, and fluoride as major threats in the aquifer systems. • SOM and NETPATH modeling assessed contamination risks and aquifer connectivity in West Mecsek region, Hungary. • Hydrochemical, isotopic, and statistical methods to differentiate geogenic and anthropogenic sources of contamination. • NETPATH revealed structural mixing: 66 % karst and 82 % BCF recharge into Neogene aquifer via fault-controlled pathways. • SOM clustering classified hydrochemical facies and contamination hotspots, corroborated by PCA, Gibbs plots, ion correlations. • Health risk assessment identified uranium, arsenic, fluoride as major threats in BCF and PT aquifers, especially for children.