Abstract
Knowledge of the spatio-temporal occurrence of avalanche activity is critical for avalanche forecasting. We present a near-real time automatic avalanche monitoring system that outputs detected avalanche polygons within roughly 10 min after Sentinel-1 SAR data are download. Our avalanche detection algorithm has an average probability of detection (POD) of 67.2% with a false alarm rate (FAR) averaging 45.9, with a maximum POD of over 85% and a minimum FAR of 24.9% compared to manual detection of avalanches. The high variability in performance stems from the dynamic nature of snow in the Sentinel-1 data. After tuning parameters of the detection algorithm, we processed five years of Sentinel-1 images acquired over a 150 × 100 km large area in Northern Norway, with the best setup. Compared to a dataset of field-observed avalanches, 77.3% were manually detectable. Using these manual detections as benchmark, the avalanche detection algorithm achieved an accuracy of 79% with high POD in cases of medium to large wet snow avalanches. For the first time, we present a dataset of spatio-temporal avalanche activity over several winters from a large region. Currently, the Norwegian Avalanche Warning Service is using our processing system for pre-operational use in three regions in Norway.
Highlights
IntroductionSnow avalanches (hereafter called avalanches) are rapid mass movements of snow down a hillside or slope occurring in all snow-covered mountain areas worldwide
Snow avalanches are rapid mass movements of snow down a hillside or slope occurring in all snow-covered mountain areas worldwide
We compared the automatic detections to manual detections from the same dates
Summary
Snow avalanches (hereafter called avalanches) are rapid mass movements of snow down a hillside or slope occurring in all snow-covered mountain areas worldwide. Improving public avalanche forecasting to reduce fatalities and mitigate damages to infrastructure has a high socio-economic relevance to people living in snow-covered mountain areas. Conventional public avalanche forecasting (how many avalanches of which size are expected) is carried out by human experts that rely on diverse, incomplete data, with a convenient spatial sample, i.e., what can be observed/obtained within the forecast domain. The experts must deal with spatio-temporal scaling issues where data over short time frames from small and not always-representative areas are available. Data on avalanche activity are rarely available at a scale relevant for the entire forecast domain, despite its critical importance in forecasting avalanches [3]. Knowledge about the spatio-temporal occurrence of avalanche activity provides direct evidence of snow instability as well as quantitative data on the consequence factor in the avalanche risk equation (risk = likelihood × consequence)
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
