Challenges and Opportunities in Deep Learning and Geostationary Satellite Remote Sensing for Volcanic Cloud Monitoring

Volcanic eruptions pose significant hazards to aviation, public health, and the environment due to the complex mix of materials ejected, including ash and sulfur dioxide (SO<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf>). Detecting and tracking volcanic clouds is crucial for issuing timely warnings and advisories. In this study, we propose a deep learning model trained on multispectral images from geostationary satellite sensors, specifically the Advanced Baseline Imager (ABI) on GOES and the Spinning Enhanced Visible and InfraRed Imager (SEVIRI) on MSG. The model, trained on Ash RGB images, enables automated detection and tracking of volcanic clouds, facilitating timely responses to volcanic activities worldwide, such as at Mt. Etna (Italy) and the Aleutian Arc (Alaska).