The researchers used radar data from the Sentinel-1 satellites to map where avalanches have occurred in recent years. This technology provides high-resolution images that can be used to analyze changes over time and identify avalanche-prone areas. The mapping shows that far more avalanches occur than previously thought.
The researchers use deep learning and data-driven models fed with historical avalanche data, topography, and meteorological variables to train a machine learning model that will eventually help us predict avalanches and reduce damage to people and infrastructure.
You can get the details in this webinar, organized by the Geoscience and Remote Sensing Society. Here, Jakob Grahn explains what an avalanche is, how SAR radar data is used for mapping, and how the researchers are working to predict future avalanches.