Geographia Technica, Vol 21, Issue 1, 2026, pp. 46-60
REMOTE SENSING OF RESERVOIR WATER QUALITY: MODELLING OPTICALLY SENSITIVE AND NON-SENSITIVE PARAMETERS WITH LANDSAT
Suraj PAWAR 
, Rushikesh KULKARNI , Kanchan KHARE , Humera KHANUM
ABSTRACT: This study integrated in situ water quality measurements with Landsat 8 Operational Land Imager (OLI) data to assess the water quality of the Khadakwasala Reservoir in Pune, India. Stepwise regression analysis was applied to develop models correlating Landsat-derived spectral indices with key water quality parameters, including turbidity, chlorophyll-a, dissolved oxygen (DO), biochemical oxygen demand (BOD), and chemical oxygen demand (COD). The models achieved predictive accuracy with R2 values of 0.85 for turbidity, 0.90 for chlorophyll-a, 0.65 for DO, 0.64 for BOD, and 0.84 for COD. Turbidity was identified as a key predictor for the non-sensitive parameters (DO, BOD, COD). These models provide a scalable method for monitoring water quality, facilitating broader spatial assessments using satellite data. The dataset supports the reuse of remote sensing data for water quality management and environmental monitoring. The distributed images of water quality parameters can be obtained at the repository address mentioned in the data availability section.
Keywords: Water Quality Assessment; Remote Sensing of Lakes; Landsat 8/9 OLI; Surface Water Quality Parameters; Stepwise Regression Model; Chlorophyll, Turbidity; Khadakwasla Reservoir
Full article here