EUVIMEDEUVIMED entdecken
Crossref

Integrated machine learning and geochemical modeling reveal hydrogeochemical controls on fluoride and arsenic co-contamination in groundwater

Zahid Ullah, Waqar Ali, Ibad Ullah, Tauseef Ahmad, Sebnem Arslan, Shakeela Rauf, Peng Licheng

Environmental Geochemistry and Health · 2026-04 · 48 (6)

DOI: 10.1007/s10653-026-03161-4 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Environmental Geochemistry and Health, online erschienen am 08.04.2026. Verfasst von Zahid Ullah und 6 weiteren Autor:innen. Beteiligte Einrichtung: School of Environmental Science and Engineering, Hainan University, Haikou, 570228, People's Republic of China. Hauptthemen laut MEDLINE: Groundwater, Fluorides, Arsenic, Machine Learning. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 91 Literaturangaben.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Groundwater contamination by fluoride (F⁻) and arsenic (As) is a serious environmental issue and poses significant human health risks in many developing countries, including Pakistan, particularly in Sindh Province. To investigate a realistic situation of groundwater contamination, a total of 170 groundwater samples were collected and analyzed concerning F- and (As) along with other physicochemical parameters. The concentrations of fluoride (F⁻) and arsenic (As) in groundwater samples ranged from 0.5 to 6.35 mg/L and 0.5 to 22 µg/L, with mean values of 1.82 mg/L and 5.78 µg/L, respectively. Hydrochemical facies result show that water type of the groundwater resources of the study area belong to mixed CaNaHCO3 followed by CaHCO3 type while few samples fall into NaCl type. Gibbs diagrams indicate that rock-water interaction controls groundwater hydrochemistry, with saturation indices showing calcite, dolomite, fluorite, and goethite are saturated in the study area. Machine Learning (ML) models, including Random Forest (RF), Artificial Neural Network (ANN) and Logistic Regression (LR), were applied and the target variable was F- due to its higher concentration in groundwater samples as compared to As. In ML, models the permutation feature, as well as the mean decrease in impurity (MDI), was used to identify the variables affecting F- in the research region. Among the models, RF achieved the highest accuracy (0.94) and sensitivity (0.97), along with a relatively low error rate (0.06). ANN showed strong performance with an accuracy of 0.92 and sensitivity of 0.88, while LR demonstrated comparatively lower sensitivity (0.81) despite achieving an accuracy of 0.90. The study shows that groundwater in the Dadu Canal Command area is affected by dual contamination, and that integrating hydrogeochemical analysis with machine learning helps identify sources and spatial risk, supporting groundwater management in Sindh.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

GroundwaterFluoridesArsenicMachine LearningWater Pollutants, ChemicalEnvironmental MonitoringPakistanRandom ForestNeural Networks, ComputerLogistic Models

Stichwörter der Autor:innen

ArsenicDadu Canal CommandFluorideGeochemical modelsGroundwaterMachine learning

Substanzen

FluoridesArsenicWater Pollutants, Chemical

Autor:innen und Institutionen

  • School of Environmental Science and Engineering, Hainan University, Haikou, 570228, People's Republic of China.
  • Anhui Provincial Key Laboratory of Forest Resources and Silviculture. School of Forestry & Landscape Architecture, Anhui Agricultural University, Hefei, 230036, China.
  • School of Computer Science, Geological Resources and Geological Engineering, China University of Geosciences, Wuhan, 430074, China.
  • Faculty of Engineering, Department of Geological Engineering, Ankara University, 06830, Ankara, Turkey.
  • Department of Chemistry, School of Science, University of Management & Technology Lahore, Lahore, 54770, Pakistan.
  • School of Environmental Science and Engineering, Hainan University, Haikou, 570228, People's Republic of China. lcpeng@hainanu.edu.cn.

Zitieren

APA

Ullah, Z., Ali, W., Ullah, I., Ahmad, T., Arslan, S., Rauf, S., & Licheng, P. (2026). Integrated machine learning and geochemical modeling reveal hydrogeochemical controls on fluoride and arsenic co-contamination in groundwater. Environmental Geochemistry and Health, 48(6). https://doi.org/10.1007/s10653-026-03161-4

Vancouver

Ullah Z, Ali W, Ullah I, Ahmad T, Arslan S, Rauf S, et al. Integrated machine learning and geochemical modeling reveal hydrogeochemical controls on fluoride and arsenic co-contamination in groundwater. Environmental Geochemistry and Health. 2026;48(6). doi:10.1007/s10653-026-03161-4

BibTeX
@article{Ullah2026,
  author = {Ullah, Zahid and Ali, Waqar and Ullah, Ibad and Ahmad, Tauseef and Arslan, Sebnem and Rauf, Shakeela and Licheng, Peng},
  title = {Integrated machine learning and geochemical modeling reveal hydrogeochemical controls on fluoride and arsenic co-contamination in groundwater},
  journal = {Environmental Geochemistry and Health},
  year = {2026},
  volume = {48},
  number = {6},
  doi = {10.1007/s10653-026-03161-4}
}
Kostenlos starten →