In the heart of Odisha, India, a groundbreaking study led by Abhijeet Das, a researcher from the Department of Civil Engineering at C.V. Raman Global University (CGU), is revolutionizing the way we assess and manage surface water quality. Das and his team have harnessed the power of machine learning models to predict water quality with unprecedented accuracy, offering a beacon of hope for sustainable water management in the face of climate change and urbanization.
The Brahmani Basin, a vital river system in Odisha, is under threat from increasing water demand and decreasing supply. “The growing demand and decreasing water supply led to scientific research into surface water availability and sustainable management as a substitute,” Das explains. His study, published in the journal ‘Desalination and Water Treatment’ (which translates to ‘Water Purification and Treatment’), employs a sophisticated blend of geospatial techniques and machine learning models to assess water quality and delineate surface water potential zones.
The research team collected 15 physicochemical parameters from seven monitoring locations over a span of six years, from 2019 to 2025, during the pre-monsoon season. They utilized a Geographical Information System (GIS) integrated with a multi-criteria decision-making (MCDM) methodology to manage water resources and analyze spatial data. The study revealed that the water quality in the Brahmani Basin is largely poor, with only 42.85% of the area falling under the ‘good’ category.
Das and his team developed an Entropy (E)-based classification model for water quality (WQ) prediction, overcoming the limitations of traditional single prediction models. They also employed the Technique of Order of Preference by Similarity to Ideal Solution (TOPSIS) to identify the most polluted sites. The findings indicated that sites N-(1), (2), (7), and (3) are the most polluted, primarily due to extensive land-use changes, urbanization, and industrial activity.
The study also highlighted the degradation of water quality at seven sites, exacerbated by the disposal of domestic and industrial solid waste. “The degradation of WQ at 7 sites, is exacerbated by the disposal of domestic and industrial solid waste, including plastics, polythene bags, paper waste and food waste,” Das noted.
The team’s innovative approach also involved the use of the Synthetic Minority Over-Sampling Technique (SMOTE) and Support Vector Machine (SVM) investigation. The SVM model demonstrated impressive accuracy, with root mean square error (RMSE) values ranging from 0.08 to 2.75 for training and 0.04 to 5.79 for testing, and a coefficient of determination (R2) ranging from 0.91 to 1.0.
The implications of this research are vast, particularly for the energy sector, which relies heavily on water for cooling and other processes. Accurate water quality assessment can help energy companies make informed decisions about water usage and management, ultimately leading to more sustainable and efficient operations.
Das’s study also provides valuable insights for machine learning-based water quality assessment, aiding researchers and professionals in decision-making and management. The outcomes of this study will hold significant value for upcoming researchers as it will reduce the duration and expenses of analysis through the utilization of these algorithms in predicting the sustainability of surface water potability.
As we grapple with the challenges posed by climate change and urbanization, Das’s research offers a promising path forward. By combining Entropy, TOPSIS, SMOTE, and SVM modeling methodologies, we can improve water quality forecasting and surface water appropriateness, ensuring a more sustainable future for all.