Scholarlink

<p>This study evaluates the capabilities of extreme gradient boosting (XGBoost) in using water quality parameters of <i>aflaj</i> systems to predict the concentrations of coliform and <i>E. coli.</i> It uses seasonal datasets to determine the most influential predictors for estimating coliform and <i>E. coli</i> concentrations. The XGBoost model analyses physicochemical, climatological, hydrological, topographical, soil, geological, land use and land cover (LULC) parameters to predict the concentrations of winter and summer coliforms and <i>E. coli</i>. The study uses the area under the curve (AUC) to determine the predictive accuracy of the coliform and <i>E. coli</i> potential maps generated by the XGBoost model. The AUC values for the validation datasets are 87%, 85%, 93% and 93% for winter coliform, winter <i>E. coli</i>, summer coliform and summer <i>E. coli</i>, respectively. The XGBoost model exhibits robust predictive performance for the validation datasets. The coliform and <i>E. coli</i> potential maps for winter and summer produced by XGBoost illustrate the distribution of high and very high concentrations across the <i>aflaj</i> system, particularly around the <i>wadi</i> networks. It is crucial to employ advanced technologies, such as machine learning techniques, for spatial and temporal monitoring for effective surveillance <i>of aflaj</i> systems. These approaches are vital in ensuring the sustainable management of groundwater resources from <i>aflaj</i> aquifers, which is critical to safeguarding public health.</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

The potential of spatial machine learning to detect and map Escherichia coli and coliform bacteria concentrations in groundwater

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年6月2日

查看全文
  • Khalifa M. Al-Kindi,
  • Ghanim Salim Said AAl-Thani

摘要

This study evaluates the capabilities of extreme gradient boosting (XGBoost) in using water quality parameters of aflaj systems to predict the concentrations of coliform and E. coli. It uses seasonal datasets to determine the most influential predictors for estimating coliform and E. coli concentrations. The XGBoost model analyses physicochemical, climatological, hydrological, topographical, soil, geological, land use and land cover (LULC) parameters to predict the concentrations of winter and summer coliforms and E. coli. The study uses the area under the curve (AUC) to determine the predictive accuracy of the coliform and E. coli potential maps generated by the XGBoost model. The AUC values for the validation datasets are 87%, 85%, 93% and 93% for winter coliform, winter E. coli, summer coliform and summer E. coli, respectively. The XGBoost model exhibits robust predictive performance for the validation datasets. The coliform and E. coli potential maps for winter and summer produced by XGBoost illustrate the distribution of high and very high concentrations across the aflaj system, particularly around the wadi networks. It is crucial to employ advanced technologies, such as machine learning techniques, for spatial and temporal monitoring for effective surveillance of aflaj systems. These approaches are vital in ensuring the sustainable management of groundwater resources from aflaj aquifers, which is critical to safeguarding public health.