1Department of Civil Engineering, Sasi Institute of Technology & Engineering, Tadepalligudem
2Graduate Student, Department of Civil Engineering, Sasi Institute of Technology & Engineering, Tadepalligudem
Surface water quality assessment is essential for sustainable water resource management and environmental protection. In Andhra Pradesh, rapid growth in the agricultural, industrial, and economic sectors has caused an increase in industrial discharge and agricultural runoff, which may significantly affect surface water quality. Thus, the present study aims to evaluate the application of Machine Learning (ML) models for predicting and assessing surface water quality using physicochemical and biological parameters. A total of 74 surface water samples collected from rivers, lakes, canals, drains, marine regions, beaches, tanks, seas, and creeks were obtained from Central Pollution Control Board (CPCB) monitoring stations. Parameters such as Temperature, Dissolved Oxygen (DO), pH, Biochemical Oxygen Demand (BOD), nitrate, conductivity, fecal coliform, and total coliform were analysed. The surface water quality prediction was carried out using four ML models, and was evaluated using the Mean Squared Error (MSE), Root Mean Squared Error (RMSE), and coefficient of determination (R2). Among the models, Random Forest (RF) and Support Vector Machine (SVM) showed strong performance for physicochemical parameters, while Decision Tree (DT) performed better for microbiological parameters such as fecal and total coliform. Although Linear Regression (LR) produced perfect statistical results, the model indicated possible overfitting and data leakage issues. Correlation analysis indicated strong relationships between BOD, nitrate, and dissolved oxygen, highlighting their influence on water quality. The study used historical data from 2020–2022 to predict surface water quality for 2023. Comparison between actual and predicted values showed good agreement with nearly 90% prediction accuracy. The results demonstrate that ML models are effective tools for surface water quality prediction and environmental monitoring in Andhra Pradesh.
Machine learning, Support vector machine, Random Forest, Decision Tree, Linear regression
