A statistical evaluation of Sentinel 2 and Landsat 9 sensors in water quality retrieval in mesotrophic and hypereutrophic reservoirs of India

The present study compares the performance of Sentinel 2 and Landsat 9 sensors in estimating water quality parameters in two reservoirs in India that have contrasting trophic statuses. In-situ measurements and reflections captured by the two sensors were regressed to derive the concentrations of Chlorophyll-a, electrical conductivity, and total dissolved solids from the satellite images. Analysed the spatial and temporal variations of these parameters and the trophic state index of the reservoirs. The study reports that Bhavani Sagar Reservoir (BSR) is mesotrophic and has lower nutrient levels than the Krishnagiri Reservoir Project (KRP), which is hypereutrophic and has higher algal blooms. Sentinel 2 outperforms Landsat 9 in the study for parameters such as Chlorophyll-a, which shows higher spatial variability of SD = 58.95 % in the BSR reservoir (R² = 0.71 for Sentinel 2 and R² = 0.61 for Landsat 9). For parameters like EC and TDS, Landsat 9 matches the performance of Sentinel in both KRP and BSR as their variability doesn’t change much. The study also finds the use of different spectral resolutions particularly red edge bands towards improving the accuracy of the models. From the study, it is evident that Sentinel 2’s better spatial and spectral resolution facilitates the monitoring of water quality indicators with a wider spatial distribution, whereas Landsat 9’s performance is best suited for small reservoirs with less regional fluctuation in water quality.