Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
4676677 | Cold Regions Science and Technology | 2007 | 10 Pages |
Snow profile interpretation has developed in the last few years from being based on experience into a semi-quantitative scientific method. Emphasizing structural rather than mechanical instability, threshold values were developed for key parameters such as weak layer grain size and hardness, and differences in grain size and hardness between layers. Despite promising attempts so far it has not been shown that this method works to quantitatively interpret snow profiles, in particular if the principal weakness is unknown. Our aim was to provide an easy and robust method based on a threshold sum approach to assess snowpack stability based on layer properties. Second, we investigated whether that method was also suited to find the principal weakness (in case it is unknown) and assess the probability for a skier-triggered avalanche on this weakness. Our data set consisted of 500 manual snow profiles observed over 16 years on skier-tested and skier-triggered avalanche slopes from both western Canada and Switzerland. A weighted threshold sum with the failure layer depth as independent variable scored highest (77% for the learning data set, 65% for the test data set). Detection of potentially critical layers proved to be less successful, in particular for the Swiss profiles. If the principal weakness was unknown, the stability classification for the potentially critical layers agreed with the observed stability for the Swiss profiles in about 53% and for the Canadian profiles in about 62% of the cases. The results emphasize that stability assessment should include – besides stability tests that help locate the principal weakness – analysis of snow layer properties, in particular grain size, type and hardness. The proposed threshold sum considering seven variables is well suited for profile analysis of manual profiles by practitioners. Stability classification of snow profiles simulated by snow cover models such as SNOWPACK will need further adaptation, in particular for application in transitional snow climates.