aceDetermining groundwater quality and its associated human health risk using hydrochemical signatures and some machine learning techniques
Abstract: Groundwater is a major source of domestic water in coastal Ghana, but its quality is increasingly threatened by salinisation, nutrient enrichment, geogenic mineralisation, and localised anthropogenic contamination. In addition to hydrochemical indices and multivariate statistics that have been used over the years to assess coastal water quality, this study has incorporated nonlinear machine learning and probabilistic risk assessment to enhance source discrimination and uncertainty-based health risk characterisat…
Distribution of wells in the central part of the western San Joaquin Valley, California by Gronberg, Jo Ann M Belitz, Kenneth Phillips, Steven P Geological Survey (U.S.) San Joaquin Valley Drainage Program. Public domain
On 2026-08-05, a peer-reviewed study reported analysis of 432 groundwater wells along Ghana's central coastal zone, combining hydrochemical indices with PCA, Self-Organising Maps and Monte Carlo probabilistic risk assessment. The work documented pronounced salinisation and mineralisation, with EC from 94.6 to 52,700 uS/cm and chloride up to 22,433 mg/l, and used machine learning to discriminate geogenic versus anthropogenic controls.
The findings matter because the same assessment flagged low non-carcinogenic risk for adults but elevated risk for children and infants, with fluoride as the key variability driver and some nitrate and fluoride exceedances of WHO guidelines. Uncertainty remains about temporal trends, exposure pathways beyond ingestion, and effectiveness of proposed mitigation such as defluoridation and nutrient-source controls in high-risk communities.
- 432 groundwater samples from wells in central coastal Ghana were analysed for physicochemical parameters, major ions, nutrients and trace metals.
- Weighted Average Water Quality Index classified 73.49% as excellent to good, while Comprehensive Pollution Index found 76.10% slightly to heavily polluted and Water Nutrient Pollution Index found 70.48% considerably to extremely polluted.
- PCA extracted five components explaining 64.25% of variance; SOM identified two clusters reflecting water-rock interaction versus anthropogenic and redox effects.
- Probabilistic risk assessment with Monte Carlo simulation found mean HI 0.38 for adults, 1.04 for children, and 1.10 for infants, with fluoride as the most important factor in risk variability.
Probabilistic assessment of the sampled wells indicated high non-carcinogenic risk for children and infants, with 32.11% and 36.49% exceedance and fluoride identified as the dominant driver, alongside nitrate and fluoride above WHO guidelines at some sites.
The rundown
Researchers examined 432 well samples from the central coastal zone of Ghana for physicochemical parameters, major ions, nutrients and trace metals, motivated by increased reliance on groundwater due to fluctuations in national water delivery. They applied hydrochemical indices, PCA, Self-Organising Maps, and Monte Carlo-based probabilistic health risk assessment.
Results showed extreme ranges in EC, TDS and chloride, with 73.49% rated excellent to good by WAWQI but 76.10% slightly to heavily polluted by CPI and 70.48% considerably to extremely polluted by WNPI, while 91.37% had low heavy metal contamination by HMEI. Risk modelling indicated low mean hazard for adults but mean HI above 1 for children and infants, prompting recommendations for routine monitoring, nutrient-source control, wellhead protection and targeted defluoridation.
Sources
- Peer-reviewedEnvironmental Geochemistry and Health2026-08-05
How should this claim be treated?
The debate