Assessment of the monsoonal impact of air pollutants and meteorological factors on physicochemical water quality parameters using remote sensing

Abstract

Abstract With growing urbanization, water contamination has become a problem. The water quality is assessed using physicochemical parameters and requires manual collection. Moreover, physicochemical parameters are insufficient for water quality monitoring as heavy rainfalls and abundance of air pollutants cause water pollution. Thus, considering natural factors as influencing parameters and the latest technology for easy and global coverage for sampling, water quality monitoring is modified. This study investigates Rawal watershed with (a) physicochemical, (b) air pollutants like nitrogen dioxide (NO2), and (c) meteorological variables like wind speed for June 2018 to September 2022. Correlation and regression analysis are performed. The results show negative correlations for NO2 with total dissolved solids (TDS) (ranging, 0.51–0.85), turbidity (range, 0.53–0.65), pH (range, 0.5–0.75), and dissolved oxygen (DO) (range, 0.5–0.82), and positive correlation with electric conductivity (EC) (range, 0.54–0.85). The regression analysis with LightGBM, multi-layer perceptron (MLP), and support vector machine (SVM) is applied with air pollutants, and meteorological parameters taken as independent variables giving root-mean-square error (RMSE) (ranging, 0.015–0.18). MLP gave an RMSE of 0.18 and 0.003 for TDS and pH, respectively. SVM performed well for DO, turbidity, and EC with RMSE ranging from 0.015 to 0.027. Moreover, floods on August 2022 are taken as a case study.