Understanding the geochemical footprints of red mud disposal to the neighboring environment: A data-driven case study through risk assessment, sensitivity analysis and source apportionment of heavy metals

Article Type

Research Article

Publication Title

Journal of Environmental Chemical Engineering

Abstract

Excessive generation of bauxite residue, commonly knowns as red mud (RM), leaves a harmful environmental footprint, leads to irreversible eco-geological consequences. The present study thoroughly understood how RM influences soil quality and human health in the immediate environment through statistical and machine learning approaches. Fifty soil and plant (rice) samples were collected from farmlands under 500 m (Zone 1) and over 500 m (Zone 2) of an RM-dumpsite. Zone 1 exhibited significantly greater concentrations (p < 0.001) of total (Cd 9.83 ± 0.63; Cr 410.38 ± 24.11, Pb 164.88 ± 10.37, Ni 109.12 ± 5.30, Cu 83.23 ± 2.95 mg kg−1) and bioavailable (DTPA-extractable) heavy metal (HM) (Cd 1.01 ± 0.11; Cr 16.03 ± 0.56, Pb 9.86 ± 0.45, Ni 6.14 ± 0.36, Cu 6.37 ± 0.27 mg kg−1) in soil. The plant accumulated HM (Cr 21.85 ± 0.90, Pb 13.98 ± 1.08, Ni 11.15 ± 0.44, Cd 0.46 ± 0.03, Cu 10.95 ± 0.68 mg kg−1) in rice grain, surpassing acceptable limits. Geostatistics and self-organizing map documented a concentric-localization pattern of HM at Zone 1. Further, PMF identified four probable pollution sources (industrial, 34.44 %; transportation, 27.33 %; anthropogenic, 18.44 %; and geogenic, 19.8 %). Sobol analysis declared both total and DTPA available metal significantly impacted the plant uptake, precisely grains were highly sensitive for Cr, Ni which finally leads to carcinogenic and non-carcinogenic risks to human health confirmed by FIAM-HQ and MCS probabilistic analysis. Lastly the machine learning model performance was evaluated to predict to rice grain metal based on soil metal availability input. Therefore, this study demonstrates the effectiveness of an integrated approach of statistics and machine learning in resolving source–risk–distribution linkages and guiding more precise pollution mitigation strategies.

DOI

10.1016/j.jece.2025.119750

Publication Date

12-1-2025

Link to Full Text

Share

COinS