Efficient and stable coupling of the SuperdropNet deep-learning-based cloud microphysics (v0.1.0) with the ICON climate and weather model (v2.6.5)

Arnold, Caroline; Sharma, Shivani; Weigel, Tobias; Greenberg, David S.

doi:https://doi.org/10.5194/gmd-17-4017-2024

Articles | Volume 17, issue 9

https://doi.org/10.5194/gmd-17-4017-2024

© Author(s) 2024. This work is distributed under
the Creative Commons Attribution 4.0 License.

Special issue:

Particle-based methods for simulating atmospheric aerosol...

https://doi.org/10.5194/gmd-17-4017-2024

© Author(s) 2024. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 17, issue 9

Development and technical paper

|

16 May 2024

Development and technical paper |

| 16 May 2024

Efficient and stable coupling of the SuperdropNet deep-learning-based cloud microphysics (v0.1.0) with the ICON climate and weather model (v2.6.5)

Caroline Arnold, Shivani Sharma, Tobias Weigel, and David S. Greenberg

Data sets

Data set for: Efficient and Stable Coupling of the SuperdropNet Deep Learning-based Cloud Microphysics (v0.1.0) to the ICON Climate and Weather Model (v2.6.5) C. Arnold et al. https://doi.org/10.5281/zenodo.8348266

Model code and software

DKRZ-AIM/dkrz-hereon-icon-superdropnet: Integrating SuperdropNet (v0.1.0) C. Arnold et al. https://doi.org/10.5281/zenodo.10069121

ICON Code v 2.6.5 including coupling schemes for integrating SuperdropNet C. Arnold et al. https://doi.org/10.5281/zenodo.8348256

Short summary

In atmospheric models, rain formation is simplified to be computationally efficient. We trained a machine learning model, SuperdropNet, to emulate warm-rain formation based on super-droplet simulations. Here, we couple SuperdropNet with an atmospheric model in a warm-bubble experiment and find that the coupled simulation runs stable and produces reasonable results, making SuperdropNet a viable ML proxy for droplet simulations. We also present a comprehensive benchmark for coupling architectures.