Scholarlink

Groundwater resources in Haryana, India, are increasingly threatened by agrochemical leaching driven by intensive agricultural practices, high fertiliser use, and shallow water tables. This systematic review synthesises recent research on groundwater vulnerability assessment using machine learning (ML) and geospatial approaches. A comprehensive literature search (2010–2025) was conducted across major scientific databases to identify studies integrating ML algorithms—such as Random Forest, Support Vector Machine, and Logistic Regression—with Geographic Information Systems (GIS) and Remote Sensing (RS) datasets. The review highlights that combining intrinsic vulnerability models (e.g., DRASTIC, SINTACS, and GOD) with ML-based predictive mapping significantly enhances accuracy and spatial interpretability. Key determinants influencing agrochemical leaching vulnerability include soil permeability, land use, groundwater depth, slope, and nitrate concentration. The findings emphasise the utility of ML–GIS integration for spatially explicit vulnerability mapping, decision support, and sustainable groundwater management. The study concludes by identifying research gaps in data resolution, model generalisation, and the need for hybrid frameworks tailored to Haryana’s diverse agro-ecological zones.

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

Groundwater Vulnerability to Agrochemical Leaching in Haryana, India—A Systematic Review of Machine Learning and Geospatial Approaches

  • 章节

  • 发表日期：

    2025年

查看全文
  • Udai Bhanu Pratap Singh Rathore,
  • Bhartendu Sajan,
  • Suraj Kumar Singh,
  • Shruti Kanga

摘要

Groundwater resources in Haryana, India, are increasingly threatened by agrochemical leaching driven by intensive agricultural practices, high fertiliser use, and shallow water tables. This systematic review synthesises recent research on groundwater vulnerability assessment using machine learning (ML) and geospatial approaches. A comprehensive literature search (2010–2025) was conducted across major scientific databases to identify studies integrating ML algorithms—such as Random Forest, Support Vector Machine, and Logistic Regression—with Geographic Information Systems (GIS) and Remote Sensing (RS) datasets. The review highlights that combining intrinsic vulnerability models (e.g., DRASTIC, SINTACS, and GOD) with ML-based predictive mapping significantly enhances accuracy and spatial interpretability. Key determinants influencing agrochemical leaching vulnerability include soil permeability, land use, groundwater depth, slope, and nitrate concentration. The findings emphasise the utility of ML–GIS integration for spatially explicit vulnerability mapping, decision support, and sustainable groundwater management. The study concludes by identifying research gaps in data resolution, model generalisation, and the need for hybrid frameworks tailored to Haryana’s diverse agro-ecological zones.