SNOWstorm (v1.0) – a deep-learning based model for near-surface winds and drifting snow in mountain environments

Saigger, Manuel; Goger, Brigitta; Mölg, Thomas

doi:10.5194/gmd-19-6497-2026

Articles | Volume 19, issue 14

https://doi.org/10.5194/gmd-19-6497-2026

© Author(s) 2026. This work is distributed under
the Creative Commons Attribution 4.0 License.

https://doi.org/10.5194/gmd-19-6497-2026

© Author(s) 2026. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 19, issue 14

Model description paper

|

17 Jul 2026

Model description paper |

| 17 Jul 2026

SNOWstorm (v1.0) – a deep-learning based model for near-surface winds and drifting snow in mountain environments

Manuel Saigger, Brigitta Goger, and Thomas Mölg

Supplement

https://doi.org/10.5194/gmd-19-6497-2026-supplement

Data sets

Observational data and model output to "SNOWstorm (v1.0) - a deep-learning based model for near-surface winds and drifting snow in mountain environments" Manuel Saigger et al. https://doi.org/10.5281/zenodo.18670232

HEF-LES simulations Brigitta Goger https://doi.org/10.5281/zenodo.18206320

Model output for "SNOWstorm (v1.0) - a deep- learning based model for near-surface winds and drifting snow in mountain environments" Manuel Saigger et al. https://doi.org/10.5281/zenodo.18184973

Model code and software

SNOWstorm v1.0 Manuel Saigger et al. https://doi.org/10.5281/zenodo.17580746

SNOWstorm Manuel Saigger et al. https://doi.org/10.5281/zenodo.17580745

WRFsnowdrift Manuel Saigger https://doi.org/10.5281/zenodo.10837359

Short summary

We present a new model to predict near-surface winds and wind-driven transport of snow in mountain regions at high resolutions. With its deep-learning based design, it is several orders of magnitude less computationally expensive compared to traditional numerical methods, while being applicable over a wide range of topographic settings and atmospheric conditions. A first application case study on a glacier in the European Alps showed good agreement with numerical simulations and observations.