Please see attached document with review and comments.

The presented article describes an evaluation of a recently developed look-up table for the avalanche danger level during operational use. Results from 26 avalanche forecasting services were compiled and analyzed. The results stimulate ongoing discussion on (the use of) concepts in avalanche forecasting. Finally, suggestions to modify the table are made. There is no doubt that the present work is important for the European avalanche forecasting community – and may in the mid-term improve forecasting consistency and transparency. Some of the article’s key points:

• For most of the cases, the authors found good agreement of issued and matrix-suggested danger levels. However, this could be expected as again the “same” population, i.e. avalanche forecasters, were questioned – only this time, in a real-life forecasting setting.
• Results in some cells of the table, and in two cells in particular, were more dispersed than in other dry-snow situations. The authors provide hypotheses, but a clear explanation for the observed differences does not seem in reach, yet.
• More importantly, wet- and glide-snow avalanche situations led to considerable differences between issued and matrix-suggested danger levels. It seems that the look-up table, and in particular the concepts behind, could not be applied with sufficient rigor. Vague stability descriptions and other workarounds may be at the origin of the observed differences.
• A finer resolution for some cells is demonstrated based on evaluations from 2 forecasting services. The value of finer resolution remains unclear for risk management and applicability in forecasting services. The look-up table relies on a classification concept with discreet and clear-cut classes, after all.
• The suggested simplification of the look-up table can be regarded a practical outcome. Some cells were only rarely used during the 2 winters of testing. Some cells showed a majority vote for one choice of danger level. Modifications are suggested, but the final result need to be published.

Some questions arose during the review regarding methodology and interpretation of the results. Improved descriptions of stability and frequency classes seem of major importance for the success of the look-up table. I structured my comments into major comments, which address the analysis and the interpretation of the results, and minor comments related to presentation and writing. They are attached in a pdf file.

Overall, the study seems suitable for NHESS. I hope the comments are helpful in bringing the manuscript to publication.