Dynamic monitoring and mapping of water quality indicators using multi-modal and multi-scale satellite imagery, UAVs, and open-source cloud computing platform

Algal blooms, resulting from rapidly growing algae in freshwater and marine environments, pose serious risks to biodiversity, ecosystems, and human health. Algal blooms are predicted to increase due to temperature, nutrient availability, and alien species invasions. Satellite remote sensing products provide a versatile monitoring tool that complements in-situ sampling. This study used remote sensing products to investigate algal bloom dynamics across inland water bodies in England (Lake Windermere) and India (Yamuna and Ganga rivers). Specifically, Google Earth Engine was used to characterise BGA, MCI, and NDCI from high-resolution satellite data to investigate algal blooms in Windermere from 2020 to 2025 and Yamuna and Ganga rivers in 2021 to 2024. By integrating data from Sentinel-2 and PlanetScope, enhanced by UAV sensor technology for high-resolution data collection, we establish predictive models for assessing water quality parameters. To analyze the data, we implement ML algorithms. Our findings indicate that RF outperforms other ML algorithms when using Sentinel-2 data, achieving an overall accuracy of 70.71% with a Kappa statistic of 0.79. Integrating a similar methodology on PlanetScope and high-resolution drone imagery to improve and increase performance boost is an ongoing task. To assess phytoplankton blooms using satellite data, we are analyzing imagery from sources like Landsat-8, 9, MODIS, and Sentinel-2 to quantify the number of blooms based on chlorophyll-A concentrations. The effectiveness of this monitoring depends on the spatial resolution, which influences the detection of smaller blooms (high-resolution imagery captures more detail), and the temporal resolution, which affects the ability to monitor ephemeral events (daily data is optimal) and thus to actually quantify is a challenge per se. Even with the performance metrics, we establish correlations between band indices and in-situ field-based measurements (pH, temperature, salinity, conductivity, turbidity, etc.). An online dashboard application will be developed to visualize results through spectral band wavelength charts, time-series data, and spatial distribution maps by integrating UK and India’s environmental agency open-source data. The future scope of our methodology can incorporate advanced techniques such as SAM, spectral feature fitting, and continuum band removal for quantitative hyperspectral data analysis. This comparative analysis emphasizes the urgent need for continuous monitoring to protect ecosystems and public health in both regions. We advocate innovative, sustainable water resource management approaches by uniting advanced remote sensing technologies with traditional methods. Ultimately, our findings aim to inform interventions to improve water quality and ecological health, benefiting local communities and the ecosystems they depend on. Through collaborative efforts, this study aspires to enhance understanding of the intricate connections between water quality dynamics, paving the way for policymakers to adhere to comprehensive management strategies that address the needs of future generations by focusing on SDG-6 (clean water sanitation) and SDG-14 (life below water).