Scholarlink

<p>Groundwater, which accounts for approximately 98% of available freshwater resources, is vital in providing drinking water and supporting agriculture and industry, especially in arid regions. The Rmel aquifer in northwestern Morocco is a strategic resource increasingly threatened by agricultural pollution, salinization, and overexploitation. This study integrates advanced geostatistical methods, machine learning models, and spatiotemporal analyses to evaluate the aquifer’s hydrochemical evolution and forecast future trends. Data from 2010 and 2021 were analyzed using techniques such as kriging, principal component analysis (PCA), hierarchical clustering analysis (HCA), and self-organizing maps (SOM). Predictive models, including ANFIS (Adaptive Neuro-Fuzzy Inference System), kernel-based Gaussian Process Regression (k-GPR), and decision trees (DT), were applied to estimate electrical conductivity (EC) by 2030 under climatic and anthropogenic pressures. Results reveal a progressive increase in salinization, particularly in southern areas, mainly driven by the intensification of agricultural activities and return flows from irrigation. ANFIS exhibited the highest performance (R<sup>2</sup> = 0.95). The prediction maps identified zones most vulnerable to salinization and provided concrete recommendations: artificial recharge, optimization of agricultural practices, and regulation of groundwater abstraction. This integrated approach underscores the aquifer’s fragility and provides a robust framework for the sustainable management of groundwater resources.</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

Assessment and modeling of the hydrochemical evolution of the Rmel aquifer (NW Morocco): geostatistical approaches and machine learning for sustainable management

  • 期刊论文

  • 发表日期：

    2025年5月23日

查看全文
  • Morad Chahid,
  • Ismail Hilal,
  • Khalid En-Nagre,
  • Chaimaa Et-Takaouy,
  • Jamal Eddine Stitou El Messari,
  • Mourad Aqnouy

摘要

Groundwater, which accounts for approximately 98% of available freshwater resources, is vital in providing drinking water and supporting agriculture and industry, especially in arid regions. The Rmel aquifer in northwestern Morocco is a strategic resource increasingly threatened by agricultural pollution, salinization, and overexploitation. This study integrates advanced geostatistical methods, machine learning models, and spatiotemporal analyses to evaluate the aquifer’s hydrochemical evolution and forecast future trends. Data from 2010 and 2021 were analyzed using techniques such as kriging, principal component analysis (PCA), hierarchical clustering analysis (HCA), and self-organizing maps (SOM). Predictive models, including ANFIS (Adaptive Neuro-Fuzzy Inference System), kernel-based Gaussian Process Regression (k-GPR), and decision trees (DT), were applied to estimate electrical conductivity (EC) by 2030 under climatic and anthropogenic pressures. Results reveal a progressive increase in salinization, particularly in southern areas, mainly driven by the intensification of agricultural activities and return flows from irrigation. ANFIS exhibited the highest performance (R2 = 0.95). The prediction maps identified zones most vulnerable to salinization and provided concrete recommendations: artificial recharge, optimization of agricultural practices, and regulation of groundwater abstraction. This integrated approach underscores the aquifer’s fragility and provides a robust framework for the sustainable management of groundwater resources.