Authors: Baya Cherif, Marouane Temimi – Stevens Institute of Technology
Title: A SLSTR-Based Framework for Operational Monitoring of River Ice and Surface Conditions
Presentation Type: Poster Presentation
Abstract: The project evaluates the suitability of Sentinel-3 Sea and Land Surface Temperature Radiometer (SLSTR) imagery for river ice detection using deep learning. Despite its daily global coverage and multi-spectral capabilities, SLSTR remains underutilized for cryospheric monitoring tasks. Here, a deep learning-based framework is developed to exploit SLSTR observations, including visible, near-infrared, and thermal infrared channels, as well as satellite and solar geometry, for automated river ice mapping.
The generated SLSTR-based product is assessed qualitatively through visual comparison with RGB imagery and quantitatively against a well-established VIIRS-derived river ice dataset. The results show strong consistency with VIIRS observations, demonstrating the effectiveness of SLSTR for river ice detection. These findings highlight the potential for synergistic use of SLSTR and VIIRS to improve spatial and temporal monitoring of river ice conditions, particularly in remote northern regions.