Multidimensional characterization of the groundwater contamination and the health risk in an industrial park: A combined multivariate statistics, source apportionment, and Monte Carlo simulation approach

Groundwater contamination at decommissioned petrochemical sites represents a significant understudied environmental challenge. This study addresses the urgent need for pollution source identification and risk assessment through an innovative framework integrating multivariate receptor modeling (PCA-APCS-MLR), multi-index pollution evaluation (PI/NPI/HPI/HEI), and probabilistic health risk analysis. Hydrochemical analysis for 201 samples (Oct-Dec 2024) revealed 47.39 % heavy metal enrichment (Mn > Fe > Cr > Zn > As) above background levels, with pollution indices confirming moderate-to-high contamination (HPI: 44.75-209.69; HEI: 5.03-42.44). Random forest regression identified Ca²⁺, Mg²⁺, ORP, TOC, and Na⁺ as primary control governing eight groundwater heavy metals. The strong agreement between measured and simulated values confirmed robust model predictability (R² = 0.703-0.799). PCA-APCS-MLR quantification identified five key sources: heavy metal pollution (16.42 %), industrial salts (13.73 %), petroleum hydrocarbons (14.51 %), water-rock interaction (14.44 %), and anthropogenic copper (11.33 %) (R² = 0.61-0.87). Health risks exhibited critical age-dependency: children showed higher non-carcinogenic risk (HI=13.70 vs. adults' 7.27) while adults faced elevated carcinogenic risk (CR=2.54 ×10⁻³ vs. children's 1.40 ×10⁻³). Over 99 % of total risk originated from oral ingestion, primarily driven by Cr and As exposure, with Monte Carlo simulations validating deterministic risk estimates. This integrated approach provides essential insights for source-specific remediation and exposure control in legacy industrial zones.