Skip to content

Cookies 🍪

This site uses cookies that need consent.

Learn more

Zur Powderguide-Startseite Zur Powderguide-Startseite
mountain knowledge

World of Science | ISSW Part 2 - Regional avalanche forecasting

Is significant always significant? The complicated path to consistent risk levels.

by Lea Hartl 01/04/2017
We continue with the International Snow Science Workshop 2016 and the studies presented there: Having already dealt with some topics in Part 1 that have more to do with the nature of snow and avalanches, we now turn our attention to people and how they deal with avalanches. Avalanche reports and the authorities that issue them are extremely important for us as winter sports enthusiasts, but their job is not always easy. On the one hand, of course, because the snow doesn't always do what you might expect, and on the other because very complex issues have to be communicated as simply and intuitively as possible. What are the current issues here and how do you approach them?

Continuing with the International Snow Science Workshop 2016 and the studies presented there: After we have already dealt with some topics in Part 1 that have more to do with the nature of snow and avalanches, we are now moving on to people and how they deal with avalanches. Avalanche reports and the authorities that issue them are extremely important for us as winter sports enthusiasts, but their job is not always easy. On the one hand, of course, because the snow doesn't always do what you might expect, and on the other because very complex issues have to be communicated as simply and intuitively as possible. What are the current issues here and how are they approached?

Danger level: consistency between warning services?

Avalanche warning services have the difficult task of summarizing a lot of different information in a single number: the danger level. Of course, the situation reports do not only contain the danger level, but for many it is the most important - although hopefully not the only - perceived content. Avalanche watchers from the USA have asked themselves the following question: Do avalanche warnings with the same information always come to the same conclusion, i.e. the same danger level? The study refers to the North American danger level scale, which differs slightly from the European scale and is primarily used in the USA, Canada and New Zealand. The definition of the hazard levels is somewhat less clear in the American version. To answer the initial question, 10 hypothetical snow scenarios were developed based on old situation reports from various regions. For each scenario, there was a brief summary of the weather in recent days, the snowpack, recent avalanche events, etc.

68 avalanche forecasters currently working in the public sector in the USA, Canada and New Zealand were asked to assign a danger level for the 10 scenarios. The evaluation shows that different danger levels are chosen, even if the results "roughly" match. In none of the 10 scenarios was the same level selected by all participants and in 9 of the 10 scenarios at least 3 different levels were specified. In most cases, the majority of participants selected the level that was actually issued. Deviating information was generally only one level higher or lower. The highest level of agreement was found for low and moderate scenarios. At higher hazard levels, the agreement is poorer, especially in scenarios that are difficult to forecast, in which a previously dry old snow cover is soaked or there is an old snow problem.

The authors conclude that although there is general consistency, the variability between individual countries, institutions and people still requires attention. Greater cooperation between warning services and joint training is needed for more consistent assessments. It is also suggested that more weight should be given to information other than the threat level in the situation report. It is emphasized that the survey cannot be directly compared with operations for various reasons. Avalanche forecasters typically benefit greatly from monitoring the snowpack throughout the season - a knowledge base that was inherently lacking in the survey scenarios.

Study: Brian Lazar, Simon Trautman, Mike Cooperstein, Ethan Greene, Karl Birkeland, 2016. North American Avalanche Danger Scale: Do Backcountry Forecasters apply it ocnsistently? Proceedings, International Snow Science Workshop, Breckenridge, Colorado.

mountain knowledge
presented by

Danger level: consistency among observers?

The researchers at the SLF are also dealing with a related problem: in order to compile a situation report, the warning services need detailed information about the snow cover and weather conditions in many different areas. As the warning service staff cannot be on site everywhere themselves, there are observer networks that report data and impressions to the warning services. Observers are typically members of avalanche commissions, local mountain guides or hut keepers who regularly report what is happening in their area. The reports from a region (region in the sense of the regions in the situation reports) are collected by the warning services and summarized in the bulletin for this region. The question arises: do observers from the same region also provide the same hazard assessments?

In order to find an answer, over 1000 cases were examined in which hazard assessments from different observers were available for the same region. In around 80% of cases, the assessments matched. In 10% of cases, observers were sometimes above or below the general danger level and in the remaining cases a bias (defined by statistical thresholds) was found: the observers always either assessed the situation as particularly critical or, conversely, saw less danger than other observers.

Even within the same region (the regions in Switzerland are on average 225km2 in size), snow conditions can of course differ and then deviations in the observers' assessments are to be expected. However, there are also some observers who always tend to have a comparatively high or low risk assessment. It is therefore important that joint and uniform training courses take place and that the warning services know their observers so that they can assess their assessments accordingly. According to Rudi Mair, the Tyrolean avalanche warning service has found that the observers are often influenced by the forecasters: Only when the situation report goes down with the danger level do the observers do the same in their assessments.

Study: Frank Techel, Lukas Du?rr and Ju?rg Schweizer, 2016. Variations in individual danger level estimates within the same forecast region. Proceedings, International Snow Science Workshop, Breckenridge, Colorado.

Harmonization of the hazard assessment

In order to counteract the discrepancies in the hazard assessments and the corresponding hazard level output, more precise definitions and uniform schemes are needed that can be adhered to. The so-called Bavarian Matrix (BM, see figure above) has become established in the European warning services, although there are still discussions as to how the individual cells are to be interpreted exactly, as some terms are not clear. The BM consists of two parts: the left part refers to human-triggered avalanches, the right part to spontaneous avalanches. In North America, there is the Conceptual Model of Avalanche Hazard (CMAH). The CMAH is more of a conceptual workflow and does not produce a hazard level at the end. The figure on the right shows an example application for two different cases (e.g. storm slab - drifting snow, persistent slab - old snow). The size of the ellipses shows the forecast uncertainty inherent in the avalanche problem. In contrast to the BM, strong reference is made here to the potential size of the avalanches.

Some European avalanche forecasters have now proposed a combination of the two methods: the Avalanche Danger Assessment Matrix (ADAM). The ADAM consists of two parts: the first relates to the probability of triggering and the spatial distribution of the danger spots. The second part shows the relationship between the spatial trigger probability resulting from the first part and the avalanche size. At the end there is an explicit hazard level.

This variant also offers a certain potential for conceptual ambiguity, but the authors hope that ADAM can contribute to more standardized hazard levels - both internationally and within the warning service. They emphasize that uniform terminology would also be desirable across languages and that clear translations should be agreed upon accordingly. The main aim of regional avalanche forecasting is generally to inform the public. The authors suggest assigning short, intuitive phrases to the hazard levels, which are simpler than the definitions of the hazard level scale (see table). These could also be easily derived from ADAM.

Study: Karsten Mu?ller, Christoph Mitterer, Rune Engeset, Ragnar Ekker and Solveig Kosberg, 2016. Combining the Conceptual Model of Avalanche Hazard with the Bavarian Matrix. Proceedings, International Snow Science Workshop, Breckenridge, Colorado.

mountain knowledge
presented by
Ad

Whining at a high level

The discussion about terminology and matrices is in many ways a classic "first world problem". This is illustrated in an article that deals with avalanche warnings in Afghanistan, Pakistan and Tajikistan. Time and again, catastrophic avalanches occur here, burying entire villages. A development aid organization and avalanche warning experts in Montana have identified 571 villages in particularly critical locations. Local observers have been trained to enter snow and weather data into a central online system and know what to do in an emergency. Professional avalanche forecasters in the USA have access to the system and are on hand to advise the observers on site via the internet. If, for example, a lot of precipitation is reported, the professional in the USA can evaluate radar data and weather models and pass on relevant information.

The entire project works with very simple methods: Weather monitoring, basic avalanche education and awareness on the ground, and some rules of thumb on when to evacuate a village. Since implementation, villages have been successfully evacuated a few hours before massive avalanches. A hopeful result of international and interdisciplinary cooperation and a clear indication that even simple methods can save lives.

Study: Doug Chabot and Aysha Kaba, 2016. Avalanche Forecasting in the Central Asian Countries of Afghanistan, Pakistan and Tajikistan. Proceedings, International Snow Science Workshop, Breckenridge, Colorado.

Photo gallery

This article has been automatically translated by DeepL with subsequent editing. If you notice any spelling or grammatical errors or if the translation has lost its meaning, please write an e-mail to the editors.

Show original (German)

Related articles

Comments

mountain knowledge
presented by