Towards a model-based danger level prediction for forecasting regions

Ciao @fherla, @Christoph.Mitterer,

just brainstorming and collecting info here on how to integrate Flo's hazard chart in the dashboard and maybe extend it. Main idea: Use the input parameters to the EAWS matrix, which we get from the gridded model chain and directly assign danger levels for each avalanche problem and forecasting region.

Visualization: What about this echart of a scatter plot with bar plots, where we could define the bins of the bar plots according to categories that go into the EAWS matrix? Then we directly get the distribution of the stability (needed for EAWS matrix) and for the avalanche size to decide on the relevant category. Avalanche size: 1-Low, 2-Moderate, or 3-Considerable, 4-High which we have to translate from the depth of the buried layer Stability: Good, Fair, Poor, Very poor. We would have to define these categories based on the stability index we want to use. Does this make sense for Punstable, for example, to define these categories? Original definitions are different for most indices. Maybe we can define these categories for the "combination index" from Fabiano's student...

This is a paper by Frank and colleagues (On the importance of snowpack stability, the frequency distribution of snowpack stability, and avalanche size in assessing the avalanche danger level). Maybe this could be a basis for defining thresholds in the distribution to decide which category to use for assigning the danger level