Can model-based avalanche forecasts match the discriminatory skill of human danger-level forecasts? A comparison from Switzerland

Techel, Frank; Purves, Ross S.; Mayer, Stephanie; Schmudlach, Günter; Winkler, Kurt

doi:https://doi.org/10.5194/nhess-25-3333-2025

Articles | Volume 25, issue 9

https://doi.org/10.5194/nhess-25-3333-2025

© Author(s) 2025. This work is distributed under
the Creative Commons Attribution 4.0 License.

https://doi.org/10.5194/nhess-25-3333-2025

© Author(s) 2025. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 25, issue 9

Research article

|

11 Sep 2025

Research article |

| 11 Sep 2025

Can model-based avalanche forecasts match the discriminatory skill of human danger-level forecasts? A comparison from Switzerland

Frank Techel, Ross S. Purves, Stephanie Mayer, Günter Schmudlach, and Kurt Winkler

Data sets

Evaluating the predictive performance of human avalanche forecasts and model predictions in Switzerland F. Techel and G. Schmudlach https://doi.org/10.16904/envidat.535

Can model-based avalanche forecasts match the discriminatory skill of human danger-level forecasts? A comparison from Switzerland (data, code) F. Techel et al. https://doi.org/10.5281/zenodo.17060394

Short summary

We tested how well fully data- and model-driven avalanche forecasts compare to human-made forecasts, which also integrate added context like field observations or model output. Using data from Switzerland over three winters, we found that models – even without this extra input – performed nearly as well. While human forecasts still have a slight edge, model predictions already offer reliable support for daily avalanche forecasting.