Potability assessment of surface water of two Indus basin rivers using information enhanced linear and non-linear data-driven frameworks
- 1Department of Soil Science, Punjab Agricultural University
Abstract
Satluj and Beas, Indus basin Himalayan glaciers fed rivers, have seen astronomic magnitude of human interference due to agricultural intensification, population pressure, urbanization and industries which subsequently deteriorates the water quality of the river systems. It is therefore, important to realize the dynamics and the need to regularly monitor the river water bodies to explore the influence of changes on the collective water quality using robust analytical approaches that allow accurate predictive modeling of water quality for better river water management. Such approaches provide framework for better comprehension of complex river water quality data, enable understanding of data structure/pattern and also identifies the key parameters influencing the water quality. This work describes the application of two such approaches: multiple linear regression (MLR) and Random Forest regression (RFR) for predicting the water quality index (WQI) for the river bodies of Indian Punjab over the conventional indexing procedure using fixed set of parameters. MLR and RFR provided non-linear and linear wrapper framework around computed CCME WQI to establish the relation between WQI and variables. This analysis besides confirming the importance of variables involved in the CCME WQI computation such as Biological Oxygen Demand (BOD), Dissolved Oxygen (DO) and Total coliform (T.Coli), also identified other variables important for predicting the status of water quality. Models were able to predict the WQI with RFR had a higher predictability (R2 = 0.97) than MLR model (R2 = 0.64). The approach presented here offers quick alternative to procedural WQI computation, which allows data exploration for prediction compared to following the set steps for sub index formulation of the constituent water quality variables and their aggregation.
Related
- Safety and risk assessment frameworks for offshore CO2 storage
- Towards regulatory readiness: evaluating frameworks for microplastic health risk assessment
- First-arrival time detection based on data-driven enhanced Akaike information criterion
- Energy, environmental, and economic assessment of sustainable fuels from wet biomass via hydrothermal, sorbent-enhanced gasification, and Fischer-Tropsch route
- Comprehensive assessment of historical surface air temperature simulations over iran using 51 CMIP6 models
BibTeX
@article{kaur2026potability,
title = {{Potability assessment of surface water of two Indus basin rivers using information enhanced linear and non-linear data-driven frameworks}},
author = {Harsimran Kaur and Dinesh Kumar Benbi and Sumita Chandel and Manpreet Kaur and Dhanwinder Singh and Kuldip Singh},
journal = {Discover Water},
year = {2026},
month = oct,
doi = {10.1007/s43832-026-00474-4},
url = {https://doi.org/10.1007/s43832-026-00474-4},
note = {CC BY},
}
RIS
TY - JOUR
TI - Potability assessment of surface water of two Indus basin rivers using information enhanced linear and non-linear data-driven frameworks
AU - Harsimran Kaur
AU - Dinesh Kumar Benbi
AU - Sumita Chandel
AU - Manpreet Kaur
AU - Dhanwinder Singh
AU - Kuldip Singh
PY - 2026
DA - 2026/10/25
JO - Discover Water
DO - 10.1007/s43832-026-00474-4
UR - https://doi.org/10.1007/s43832-026-00474-4
AB - Satluj and Beas, Indus basin Himalayan glaciers fed rivers, have seen astronomic magnitude of human interference due to agricultural intensification, population pressure, urbanization and industries which subsequently deteriorates the water quality of the river systems. It is therefore, important to realize the dynamics and the need to regularly monitor the river water bodies to explore the influence of changes on the collective water quality using robust analytical approaches that allow accurate predictive modeling of water quality for better river water management. Such approaches provide framework for better comprehension of complex river water quality data, enable understanding of data structure/pattern and also identifies the key parameters influencing the water quality. This work describes the application of two such approaches: multiple linear regression (MLR) and Random Forest regression (RFR) for predicting the water quality index (WQI) for the river bodies of Indian Punjab over the conventional indexing procedure using fixed set of parameters. MLR and RFR provided non-linear and linear wrapper framework around computed CCME WQI to establish the relation between WQI and variables. This analysis besides confirming the importance of variables involved in the CCME WQI computation such as Biological Oxygen Demand (BOD), Dissolved Oxygen (DO) and Total coliform (T.Coli), also identified other variables important for predicting the status of water quality. Models were able to predict the WQI with RFR had a higher predictability (R2 = 0.97) than MLR model (R2 = 0.64). The approach presented here offers quick alternative to procedural WQI computation, which allows data exploration for prediction compared to following the set steps for sub index formulation of the constituent water quality variables and their aggregation.
SN - 2730-647X
ER -