A Novel Remote Sensing Based Approach to Estimate the Water Quality Index Using Sentinel-2 Multispectral Data

The constant disposal of solid wastes in open dumpsites and fresh water bodies severely affects the water quality that causes health concerns especially in developing countries. Hence, in order to determine the contamination rate of a specific water body, water quality index (WQI), a tool to provide the overall water quality in numerical scale, was used for better understanding to the public. The primary objective of this present study is to analyse and map-observed WQI values over Ukkadam lake, Coimbatore, India by utilizing multispectral Sentinel-2 satellite imagery, so as to generate a remote sensing based WQI model for this particular lake. The final results showed that, from the eight water leaving reflectance (rhow) bands analysed, a combination of three bands, i.e., blue (B2), coastal aerosol (B1), red edge (B5) showed good correlation with observed water quality index values yielding R2=0.778 and low p-value (.011). The validations of generated step wise regression (SWR) models were also done using 25% of the observed WQI results, showing a strong correlation with R
²
value more than 0.8.