January 31st, 2015

Piz Vilan, GR, Switzerland: Avalanche kills 5 tourists


Eight people were swept backcountry touring in an avalanche in Switzerland. Five of them did not survive. The announced danger level was 3. The death toll on that day was the worst in five years.

Up to 1.2 meters of snow has fallen in the region in recent days, and authorities had made repeated warnings of high avalanche dangers in many parts of the Swiss Alps and Jura Mountains.

This avalanche went down the mountain’s eastern flank at an elevation of 2340 meters. The group of skiers caught was backcountry touring.

There was no immediate word on what triggered the avalanche.

  • What is the evolution of such accidents over time?
  • Is the exposure of the accident’s location a main cause?
  • Are some mountain activities riskier than others?
  • Is the danger level a good indicator of risk?

We will try to answer these questions through an interactive visualisation of the last 15 years of avalanche cases.

We know from the press release that the group was caught backcountry touring, but accidents can also happen off-piste skiing. Do both activities have the same exposal to danger?

We can observe that the deadliest levels of risk are 2 and 3 . An explanation of that phenomenon is that low levels of danger (1 ) is often synonym of a very small amount of snow and a high level (4 or 5 ) repeals skiers to expose themselves to hazard. These are outlier situations that will be excluded from our further analysis.

Backcountry tourers tend to be more experienced than off-piste skiers. But from the chart it seems that they are more easily caught and killed by a level 2 avalanche? Is there an explanation for this?

An assumption we can make is that backcountry tourers are more educated than off-piste skiers who do not check the announced danger level before going out. We did a statistical study over these two populations to find out whether the mean danger exposition is a significantly different between these two populations.

To make our study meaningful, we did a matching on the two populations with propensity score to avoid comparing a group of 14 alpinists tackling the Jungfrau summit (4’158 m) in July ‘17 with 2 people going off-piste at 1500m elevation in January. We want the two groups to be exposed to the same environmental conditions but differentiate only by their danger judgement. We found that backcountry touring deadly accidents happens at 2.61 danger level while off-piste accidents happens at 2.77 on average (with a significant difference, p=0.017). We showed that educated skiers tend to expose themselves to less dangerous situations. You can check out our statistical analysis in details in this iPython notebook.

The avalanche cases that make it to the front page of the news mostly report deaths. Let’s look at the more general picture. We can observe in the following plot that on average two to three people on average are caught per avalanche accident and the survival rate is around 2/3.

Once again, decision-making from the skiers has a big impact. The consequences of an avalanche can be catastrophic for a simple day of ski. As we are right in the winter season, think of our project during your vacations and please be safe.

Have fun in the EXPLORE tab and learn from the data by yourself.