Scholarlink

<p>This study aims to evaluate key parameters of groundwater quality and associated health risks in three coastal aquifers of Cox’s Bazar, Bangladesh, with a focus on manganese contamination and geochemical processes. A total of 288 groundwater samples from 36 monitoring wells were analyzed to assess physicochemical parameters and calculate the Water Quality Index (WQI). Hydrogeochemical facies revealed distinct water types, with the Dupi Tila aquifer containing predominantly fresh Ca–HCO₃ type water, while the Tipam and Bokabil aquifers exhibited Na-Cl–SO₄ facies, indicating seawater intrusion and water–rock interactions. To predict WQI and identify key influencing parameters, four machine learning (ML) models, Random Forest (RF), Gradient Boosting Regressor (GBR), XGBoost, and Artificial Neural Network (ANN) were employed. Among these, XGBoost achieved the highest prediction accuracy (R<sup>2</sup> = 0.947, RMSE = 12.2, MAPE = 9.6%), followed by GBR and RF, while ANN showed lower performance. Feature importance analysis highlighted manganese (Mn), total dissolved solids (TDS), sodium (Na⁺), and chloride (Cl⁻) as dominant predictors. Health risk assessments using Hazard Quotient (HQ) analysis identified manganese as a significant threat, particularly for children, with over 50% of samples exceeding safe limits. The findings emphasize the need for regular monitoring and targeted mitigation in vulnerable aquifers. This study is novel in its integration of ML algorithms with geochemical analysis in a refugee-impacted coastal region, offering a predictive framework for sustainable groundwater management. The outcomes are broadly applicable to similar hydrogeological settings affected by salinization and trace metal contamination.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Prioritizing geochemical drivers of groundwater quality and health risks in coastal aquifers of Bangladesh using machine learning algorithms

  • 期刊论文

  • 发表日期：

    2025年10月13日

查看全文
  • Sumiya Sadeak,
  • Farjana Zareen Maria,
  • Md Al Amin,
  • Tareq Chowdhury,
  • Md Jahangir Alam,
  • Md Bodruddoza Mia,
  • Kazi Matin Ahmed,
  • Mahfuzur R. Khan

摘要

This study aims to evaluate key parameters of groundwater quality and associated health risks in three coastal aquifers of Cox’s Bazar, Bangladesh, with a focus on manganese contamination and geochemical processes. A total of 288 groundwater samples from 36 monitoring wells were analyzed to assess physicochemical parameters and calculate the Water Quality Index (WQI). Hydrogeochemical facies revealed distinct water types, with the Dupi Tila aquifer containing predominantly fresh Ca–HCO₃ type water, while the Tipam and Bokabil aquifers exhibited Na-Cl–SO₄ facies, indicating seawater intrusion and water–rock interactions. To predict WQI and identify key influencing parameters, four machine learning (ML) models, Random Forest (RF), Gradient Boosting Regressor (GBR), XGBoost, and Artificial Neural Network (ANN) were employed. Among these, XGBoost achieved the highest prediction accuracy (R2 = 0.947, RMSE = 12.2, MAPE = 9.6%), followed by GBR and RF, while ANN showed lower performance. Feature importance analysis highlighted manganese (Mn), total dissolved solids (TDS), sodium (Na⁺), and chloride (Cl⁻) as dominant predictors. Health risk assessments using Hazard Quotient (HQ) analysis identified manganese as a significant threat, particularly for children, with over 50% of samples exceeding safe limits. The findings emphasize the need for regular monitoring and targeted mitigation in vulnerable aquifers. This study is novel in its integration of ML algorithms with geochemical analysis in a refugee-impacted coastal region, offering a predictive framework for sustainable groundwater management. The outcomes are broadly applicable to similar hydrogeological settings affected by salinization and trace metal contamination.