杏吧原创

鈥淵ou could have a weakness in one place, and 2 metres away it could be the opposite鈥�

How do you go about assessing a slope for avalanche risk?
You take a sample of the snow by digging down to ground level so you can see all the layers. We are really interested in the interfaces between the layers, because that is where the snow is going to fail. This will provide 15 to 20 per cent of our information. The rest comes from travelling to different parts of the mountain so we can see how the hazard is distributed across the landscape. We can then present a map of avalanche risk for a region. But there鈥檚 always uncertainty because of the weather, the way the snow lies and the way it鈥檚 distributed by the wind.

We know a lot about the dynamics of avalanches, yet people are still regularly killed by them. Why is that?
When I first started forecasting avalanches, we were looking for an answer in the snow. We would tap it, pull it, test its stability, then we鈥檇 predict whether it would avalanche or not. Today we know that we can鈥檛 rely on a sample applying to the whole slope, because there is great spatial variability. You could have a weakness in one place, and 2 metres away it could be the opposite. It鈥檚 a bit of a lottery. A number of people can go down a slope and it can be OK, and then another person goes down and they trigger an avalanche because the wind has moved some snow or there are local differences in the layers.

How do you factor in human behaviour, the way people act on your information?
The human factor is probably the most important element when people venture into avalanche terrain. If you deconstruct avalanches that involve people, you can see that often their decisions were determined by what they鈥檇 already invested in the trip. It may have fulfilled a plan they made days or even months ago. They may have driven a long way. These sorts of things influence people greatly in how they make decisions on the mountain. It is astonishing what people ignore.

Give me an example.
Recently in the Cairngorms in Scotland a big avalanche had fallen and a rescue helicopter was taking people out. Another party walked right past it and went climbing right next to it. Of course they got caught in an avalanche too. People think it鈥檚 not going to happen to them.

What鈥檚 driving that kind of behaviour?
We鈥檙e dealing with the invisible. In the mountains, there are no traffic lights like on a road. There may be a known hazard, but there鈥檚 nothing in the slope to reinforce that, so people are more likely to gamble. Another issue is that mountain activities are very goal-orientated: you can鈥檛 have any more of a goal than a summit.

Do people with more experience or training take fewer risks?
None of us is immune. of avalanche victims have shown that even experienced people are strongly influenced by heuristics 鈥� behavioural biases 鈥� when they make decisions about the hazard. They may ski a slope because they鈥檝e done it before, or because other people are skiing it, or because they made a prior commitment to do it. Whether or not they have avalanche training makes no difference.

Read more: 鈥�Chance: How randomness rules our world鈥�