SLABS: An improved probabilistic method to assess the avalanche risk on backcountry ski tours

Abstract

The main danger to which backcountry skiers are exposed are dry slab avalanches triggered by skiers themselves. In this paper we present a new probabilistic method to assess the avalanche risk when planning a ski tour: SLABS (Screening the Likelihood of Avalanches on Backcountry Ski tours). The SLABS method is the first fully statistically derived probabilistic method for backcountry skiing. We used a data set with GPS (Global Positioning System) tracks of backcountry ski tours (57.8 thousand km) and a data set of 1250 accidents recorded in Switzerland during the last 20 winter seasons. A GAM (Generalised Additive Model) with binomial link discriminating between accidents and non-accidents was fitted to these data. As predictors the model uses a non-linear relation with the slope angle, and a linear relation with the danger level, the elevation and the aspect. Conditions with elevated and high risk were defined such that the accident prevention rate is 80% and 60%, respectively. A comparison of the SLABS method with the Graphical Reduction Method (GRM), the Professional Reduction Method (PRM) and the Quantitative Reduction Method (QRM) shows that it offers more freedom of movement for a given accident prevention rate. It is the first time that probabilistic methods are compared in terms of their trade-off between accident prevention and freedom of movement. Because the SLABS method is not suited for mental arithmetic it is implemented in the website www.skitourenguru.ch. This website evaluates on a daily basis the avalanche risk along thousands of backcountry tours throughout the Alps to help skiers plan their next tour. For a single slope the risk assessment of SLABS provides a starting point that has to be supplemented with local observations.