Earth’s surface continuously regulates the planet’s energy balance by reflecting a portion of incoming solar radiation back into space, a process commonly described through surface albedo. Variations in albedo influence climate processes, ecosystem functioning, urban heat dynamics, agricultural productivity, and overall planetary health. However, existing global satellite-derived albedo products are generally available at coarse spatial resolutions, limiting their ability to capture the fine-scale heterogeneity of landscapes such as cities, forests, croplands, wetlands, and water bodies.
In this talk, Mr. Varun Tiwari will present a new artificial intelligence–driven framework for generating high-resolution (10 m) surface albedo products using multispectral satellite imagery from Sentinel-2. By integrating Earth observation data with machine learning techniques, including Random Forest and neural network models, the framework learns relationships between satellite reflectance and surface energy properties to produce detailed albedo maps at unprecedented spatial scales. The study demonstrates that AI-based approaches can accurately reproduce satellite-derived albedo while significantly improving the representation of local landscape variability and ecosystem patterns.
Beyond methodological advances, the presentation will highlight how satellite remote sensing and AI are transforming our understanding of planetary health by enabling scalable monitoring of ecosystem services, land–atmosphere interactions, urban heat mitigation, and climate resilience. The talk will also discuss the broader role of geospatial AI in supporting sustainable environmental management, agricultural decision-making, and future climate adaptation strategies through globally accessible, high-resolution Earth observation systems.
Speaker
Varun Tiwari
Research Fellow, Geospatial Analytics for Water and Agriculture
