The Swiss expand their avalanche danger scale by adding sub-levels.
I sense that over the time I’ve been reading and assessing different avalanche forecasts (I’ve spent a decade roughly each in Montana, Colorado, and Oregon), they have trended towards more graphical representations and less text. In other words, it is relatively easy for those who choose to digest critical portions of an avalanche forecast to forgo reading any text. Look around the ski group in the early AM and note the number of fast-scrolling index fingers—no way folks are digesting the text portion of the forecast. Some details are likely missed as they scroll through the forecast.
The trend is towards simpler ways of communicating avalanche forecasts.
After six years of internal use and a season of limited public use in 2022-2023, the Swiss Avalanche Warning Service will add sublevels to their avalanche forecast this season. Yes, they are adding complexity, which seems to be bucking the trend. In Europe generally, including Switzerland, forecasters use five discrete levels to define the avalanche danger level.
This season, instead of a straight 1, 2, 3, 4, 5, the danger levels will include 1, 2-, 2=, 2+, 3-, 3=, 3+, and 4-.
In an ISSW proceedings paper titled “Introducing Sublevels in the Swiss Avalanche Forecast,” representatives from the WSL Institute for Snow and Avalanche Research SLF in Davos, Switzerland, discuss the adoption of sublevels, which, on the surface, requires users, like backcountry turn seekers, to understand a more nuanced danger rating scale.
The authors reason the move to sublevels and away from their traditional five levels this way: “This simplification of a complex, multi-dimensional natural phenomenon into a small number of discrete classes inevitably leads to a loss of information. Moreover, in many countries in Europe, the forecast danger level (DL) is either 2 (moderate) or 3 (considerable) on about 75% of the forecasting days.”
In short, details are lost, and forecasters, at least 75% of the time, might benefit from expanding the scale between 2 and 3. So it’s out with the old and in with the new.
In presenting the poster/paper, Célia Lucas, a Swiss forecaster, communicated that the research group needed to answer core questions: can forecasters “assess these sublevels consistently and accurately, and whether the sublevels are understood by users in the intended way and if they are useful in practice.”
Data show that when forecasters used a 3+ designation, more avalanches were, in fact, reported. Lucas said, “We took this data and checked if it was meaningful. A colleague analyzed this data to check for agreement between the danger level [using sublevels] and avalanche activity. Looking back at the historical data, we cataloged more avalanches with a 3+ designation than a 3-. We found in all these tests that what we were doing was actually relevant and meaningful.”
The international avalanche danger scales are not linear—consider it exponential. So adding more complexity to the scale means forecast users (us skiers and riders) must grasp an even more nuanced scale. The researchers’ survey confirmed a solid understanding of the sublevels, while most respondents also found the sublevels more useful for practical applications in the backcountry, like trip planning.
Note: according to the research findings, the majority of survey takers, 54%, had little to no formal avalanche training/education.
Unlike yesterday’s installment of what-did-I-learn-at-ISSW, where ATES v.2 adds bookends, the new Swiss danger rating scale infills the middle; the times when we might be tempted by a false negative shear test or ski cut and send a slope we otherwise shouldn’t have. Time will tell if a 3-, compared to a 3+, changes behavior and decision-making in the backcountry. For now, the Swiss, known for their precision, are scaling up to a 1, 2-, 2=, 2+, 3-, 3=, 3+, and 4- danger scale.
I’ve talked about this with a few very experienced friends, and I understand there would be limitations, but I think we might be better off with a 4 step scale. I think, and it’s been suggested in a couple of recent papers, that most users still think of the danger scale as linear when the forecasters and professionals understand it as exponential. So the casual user looks at “considerable” and because it’s in the middle thinks of it as about an average risk day. I also think the “extreme” category is functionally useless. With a 4 step scale your brain automatically understands if this is a more or less dangerous day. One thing I’m sure of, this new Swiss scale would fly far over the heads of most of my friends and acquaintances who have a more basic level of experience.