A Case Study on Modeling Water Quality Index of River using Multiple Linear Regression

Water pollution assessment remains a complex and critical task. To facilitate comprehensive evaluation, the Water Quality Index (WQI) serves as a reliable metric to quantify overall water quality in terms of type and extent. In this study, the WAWQI method was employed to calculate the WQI based on key parameters including pH, DO, COD, BOD, TS, TSS, TDS, turbidity, and hardness at the Omkareshwar Sampling Station, Pune district, Maharashtra, India, over a 4-year period (2019-2022). The calculated WQI values ranged between 55 and 57, classifying the river water quality as “poor”. To explore interrelationships among water quality parameters and their predictive utility, Pearson correlation and Multiple Linear Regression (MLR) analyses were conducted. A strong correlation was observed among the studied parameters, indicating potential for modelbased water quality assessment. The Regression analysis for the year 2019 yielded four MLR models: MLR1 (R² = 0.9799), MLR2 (R² = 0.9828), MLR3 (R² = 0.9837), and MLR4 (R² = 0.9953). These models were subsequently applied to estimate WQI for the year 2022. The predictive model explained 74% of the variance in WQI (R² = 0.74, p < 0.05), confirming the applicability of historical water quality data in forecasting future trends. The study highlights the utility of correlation and regression approaches for water quality monitoring and supports the development of data-driven WQI models for effective environmental management. Major Findings: The study assessed water quality of the Mutha River (2019-2022) using the WAWQI method, with WQI values (55-57) indicating poor water quality. Significant correlations among parameters were identified, and multiple linear regression models (R² up to 0.9953) were developed to predict WQI, explaining 74% of the variance for 2022.

View article page

Download PDF