WATER QUALITY PARAMETERS PREDICTION OF TIGRIS RIVER USING SENTINEL-2 DATA AND LASSO REGRESSION

Abstract

Water-related issues have become a growing concern due to the impact of human activities and climate change unpredictability, and Iraq faces a significant challenge with freshwater scarcity and poor quality. Field measurements are expensive, while remote sensing is a cost-effective alternative. This study aimed to monitor the water quality of the Tigris River in Baghdad using Sentinel-2 satellite images. The study employed the least absolute shrinkage and selection operator (LASSO) and field data from 14 different stations during 2018 and 2019 to measure water quality parameters, including temperature (Temp), electrical conductivity (Cond), total dissolved solids (TDS), potential of hydrogen (pH), turbidity (Turb), Chlorophyll-a (Chl_a), Blue-Green Algae (BGA), and Dissolved Oxygen (DO). The incorporation of spectral indices significantly improved the models’ effectiveness, with R2 greater than 0.8, except for Cond, which was 0.73. Water Quality Index (WQI) based on Iraqi standards was estimated and showed that the river water quality was categorized as poor and very poor. This approach can enhance water resource management and decisionmaking in areas where traditional monitoring approaches may be challenging and costly.