Groundwater Suitability Analysis in Punjab: A Critical Component for Sustainable Agriculture

Strategic Insights -

Researchers from Punjab Agricultural University have published a new study on integrating water quality indices with machine learning algorithms to determine groundwater suitability for drinking and irrigation purposes. The study highlights the importance of evaluating groundwater quality for sustainable agricultural practices to mitigate soil health risks and safeguard long-term farming viability. The researchers used machine learning models to analyze water quality indices, achieving high predictive accuracy for various water quality parameters.

Key Takeaways:

  • The study focused on designing and assessing data-driven machine learning frameworks for groundwater suitability analysis in Punjab using water quality indices (WQIs).
  • Five machine learning models, including Random Forest (RF), Support Vector Machine (SVM), Linear Regression (LR), Naïve Bayes, and Nearest Neighbors (NNH), were trained and evaluated using statistical performance metrics.
  • The SVM model consistently performed well for most WQIs, achieving high R2 values (0.99, 0.98, 0.96, 0.99, and 0.99) and relatively low RMSE values (1.7, 0.4, 0.21, 0.61, and 0.17) for Soluble Sodium Percentage (SSP), Magnesium Hazard (MH), Sodium Adsorption Ratio (SAR), Permeability Index (PI), and Percentage Sodium (% Na), respectively.
  • The LR model performed best for the Kelly Ratio (KR), achieving an R2 of 0.92 and an RMSE of 0.19.
  • The RF and SVM models effectively predicted the composite Ground Water Quality Index (GWQI), with R2 = 0.84 and RMSE = 7.0 for GWQIi (irrigation), and R2 = 0.76 and RMSE = 9.33 for GWQId (drinking), respectively.

Statistics:

  • The SVM model achieved R2 values of 0.99, 0.98, 0.96, 0.99, and 0.99 for SSP, MH, SAR, PI, and % Na, respectively.
  • The SVM model achieved RMSE values of 1.7, 0.4, 0.21, 0.61, and 0.17 for SSP, MH, SAR, PI, and % Na, respectively.
  • The LR model achieved an R2 value of 0.92 and an RMSE value of 0.19 for the Kelly Ratio (KR).
  • The RF and SVM models achieved R2 values of 0.84 and 0.76, and RMSE values of 7.0 and 9.33 for GWQIi (irrigation) and GWQId (drinking), respectively.

Sources:

  • NewsRx. Findings from Punjab Agricultural University Update Understanding of Sustainable Water (Integrating Water Quality Indices With Machine Learning Algorithms for Groundwater Suitability In Punjab: Implications for Drinking and Sustainable ...). Ecology, Environment & Conservation. August 15, 2025; p 124.
  • Journal of Water Process Engineering. Integrating Water Quality Indices With Machine Learning Algorithms for Groundwater Suitability In Punjab: Implications for Drinking and Sustainable Irrigation. Volume 76, 2025.