On probabilities of avalanches triggered by alpine skiers. An empirically driven decision strategy for backcountry skiers based on these probabilities

摘要

Most fatal avalanche accidents in the Alps are caused by skiers and snow-boarders. It has been one aim from the beginning to give guidelines for backcountry skiers in order to avoid avalanche accidents. About 10 years ago, the mountain guide Werner Munter developed a strategy for backcountry skiers whether to go or not to go on a skiing tour. However, his decision strategy has a lack of empirical evidence because he does not take into account incidents without avalanche accidents. This article proposes a decision strategy for backcountry skiers based on probabilities of a logistic regression model using variables, such as danger level, incline of the slope and aspect of the slope, which turned out to be the most important ones. Additional information on frequencies of skiers on slopes under specific conditions is included in the model. We used accident data and avalanche forecasts in Tyrol reported by the Tyrolean avalanche information service within three seasons (1999-2002, 497 days of observations) for model building. Additionally we carried out a holdout validation using data of the same type within two seasons (2002-2004, 314 days of observation) in order to check the accuracy of the model. Our proposal shows a remarkable correlation with Munter's method.