World of Science | Snowpack: Stability and variability

This time: Snowpack: Stability and variability (Session 10)

Snowpack stability and variability are essential factors for us winter sports enthusiasts. Both variables form the integral basis of avalanche forecasting. There, however, the factors are named slightly differently and are included in the EAWS matrix of avalanche danger levels as "probability of avalanche triggering" and "extent of danger spots". This is reason enough for researchers to keep looking into it. And so it is not surprising that Session 10 "Snowpack: Stability and variability" contains the most contributions. All of the 43 contributions presented cannot be summarized here in 3 pages, so this article is mainly limited to contributions on slab avalanches, weak layers and their distribution. And as a bonus this time with a do-it-yourself snowboard experiment!

Anti-crack: A fracture that closes

Three ingredients are needed to break a weak layer: A weak layer, a bound snow slab and a trigger. If a break in the weak layer, also known as a collapse, leads to a slab avalanche, a slope inclination of 28-30° or steeper is required. If such a snowpack is stepped on, for example, and individual (weak) connections in the weak layer break, then the overlying snow slab "closes" the resulting crack - nowadays this so-called "anti-crack model" is used as the basis for triggering a snow slab. If the energy released by this collapse is sufficient to cause neighboring connections in the weak layer to break, there is independent fracture propagation or fracture propagation, which leads to larger snow slab avalanches.

Several contributions present computer models to describe these fracture processes in the weak layer. These models calculate the complex interplay between the nature of the snow slab, the stability of the weak layer and the energy released, which leads to fracture propagation. On the one hand, there is the model that was developed for the movie "Frozen" (yes, this catcher is already worn out), on the other hand, two other types of models are presented here. Certainly outstanding is the analytical fracture model from TU Darmstadt, which combines knowledge from the field of structural mechanics with snow research (O10.6). In contrast to purely numerical models, this "Phillip&Phillip" model is not as flexible in terms of material properties, but it requires very little computing capacity and could also be calculated in real time on a smartphone. You can find out more about this new model in the ISSW2018 special edition of Berg&Steigen and the presentation from the conference is also online.

The fracture resistance of a particular layer is nowadays often tested using a PST. A fairly large block is exposed and sawn from the downhill side into the weak layer to be tested. The aforementioned critical crack length is then the distance that was sawn until the fracture propagates through the entire block on its own. Contribution P10.12 uses such a PST together with a high-speed camera to determine the crack propagation speed. A good Arte documentary was recently published about these experiments (and other current research), the images of the PST can be found from minute 17 onwards.

Unfortunately without a written legacy for posterity, there was an article about the validation of the so-called cross(-slope)-PST. The CPST is about combining the advantages of ECT and PST because, for example, ECT is unreliable in deep weak layers and the classic PST is quite complex to excavate (30x100cm but upslope, not parallel to the slope as with ECT). For the CPST test, a block the size of an ECT is simply exposed and the weak layer to be tested is penetrated parallel to the slope with the smooth side of the snow saw (see here).

Continue on the next page -->

Experiment for the next tour: "Do It Yourself" snowboard

Unaware of the existence of the Cross-PST, we recently carried out a similar stability experiment on an artificial snowboard. We found some wonderfully large surface frost near a stream (you can now also create this artificially, but the machine from the article P10.34 goes beyond the typical DIY equipment on a ski tour...), and placed a small snowboard on top of it. Here are the step-by-step instructions:

1) Find a spot with surface frost.

2) Take a layer of semolina-like snow and carefully pour a 10cm thick layer onto the frost (right-hand image).

3) Using a shovel, carefully cut off the sides of this artificial snow board to create a block the size of a typical ECT (30x90cm).

4) Use the snow saw (or alternatively a flat shovel) to drive sideways into the frost layer until the "house of cards made of frost" collapses (see video).

Notice that although the artificial snow slab is hardly bound (grit) and is only 10cm thick, there is fracture propagation in the surface frost. When the avalanche situation report speaks of snow-covered surface rime, all alarm bells must ring!

Professional experiments are also just DIY

Once again unknowingly, this article research revealed that two ISSW articles also examined precisely such DIY snowboards for surface maturity. In article O10.2, the authors prepared 30 such 10cm snowboards and tested them at different times using PST. They found that the critical crack length increases with time, i.e. it must be sawn further into the weak layer to generate fracture propagation, and thus find a corresponding snowpack stabilization with time. The second article O10.3 examines the mechanical stabilization of the 10cm snow planks over time in more detail.

Following on from such areal modeling of snow cover variability, one article deals with the comparability of snow profiles (O10.5). This refers to manually recorded snow profiles or simulated snow profiles. The technique used is the so-called Dynamic Time Warping method, which is widely used in speech recognition. The layers are not compared at absolute positions, but the layer sequences are compared - similar to how Alexa understands words regardless of whether they are spoken slowly or faster.

Furthermore, the article P10.28 examines how the results of the snowpack models can be communicated to practitioners and users. Snowpack models still do not play a major role in the avalanche situation report, not because they are wrong, but because they are very complex to use. The authors suggest that the key factors for the formation of avalanche patterns and problems should be presented and communicated.

Beech-Larch Theory

The climatic conditions of a region are generally reflected in the snowpack and also in the vegetation. For example, there are possible correlations between areas with larch-pine vegetation and old snow problems, while old snow is less of a problem in areas with fir and beech trees. Our snow pusher Lukas has written about this (P10.17), but his words put the knowledge about the tree species into perspective: "Risk management can be done with better tools".

Based on many years of snow pro recordings, further articles on the typical regional snow cover patterns in the Pyrenees (P10.13), Tromso area (P10.24), Japan (P10.29) and Eastern Canada (P10.37).

Conclusion

The size of the session alone, with 43 papers, shows how actively research is being conducted in the field of snowpack stability and variability. Something is definitely happening, but transferring the results into practice is another very complex step. Ultimately, the snowpack is similar to "Schrödinger's cat" - you only know how stable and variable it has been once you have dug up the entire snowpack.