1. An ML‐Based P3‐Like Multimodal Two‐Moment Ice Microphysics in the ICON Model.
- Author
-
Seifert, Axel and Siewert, Christoph
- Subjects
- *
NUMERICAL weather forecasting , *MELTWATER , *MACHINE learning , *MICROPHYSICS , *ATMOSPHERIC models - Abstract
Machine learning (ML) is used to build a bulk microphysical parameterization including ice processes. Simulations of the Lagrangian super‐particle model McSnow are used as training data. The ML performs a coarse‐graining of the particle‐resolved microphysics to multi‐category two‐moment bulk equations. Besides mass and number, prognostic particle properties (P3) like melt water, rime mass, and rime volume are predicted by the ML‐based bulk model. The ML‐based scheme is tested with simulations of increasing complexity. As a box model, the ML‐based bulk scheme can reproduce the simulations of McSnow quite accurately. In 3d idealized squall line simulations, the ML‐based P3‐like scheme provides a more realistic extended stratiform region when compared to the standard two‐moment bulk scheme in ICON. In a realistic case study, the ML‐based scheme runs stably, but can not significantly improve the results. This shows that ML can be used to coarse‐grain super‐particle simulations to a bulk scheme of arbitrary complexity. Plain Language Summary: Numerical weather prediction and climate models need a description of unresolved cloud microphysical processes. Such microphysical parameterizations are usually formulated as systems of equations for bulk variables that describe the time evolution of clouds and precipitation. In this study, we use machine learning (ML) techniques to build such a parameterization. As input or training data simulations of a very detailed cloud model are used. This detailed model provides information not only on the mass and number of cloud particles but also other properties like the degree of melting or the mass of liquid drops frozen on the ice particles called rime mass. The ML approach can successfully construct the necessary statistical relations that are needed for microphysical parameterization. This parameterization is then tested in simulations of increasing complexity. The new ML‐based scheme provides physically reasonable solutions and improves the simulation of a line of thunderstorms. Key Points: Machine learning (ML) is successfully applied to build a complex bulk ice microphysics scheme by coarse‐graining output of a Lagrangian particle microphysics modelThe ML‐based P3‐like microphysics scheme improves the representation of the stratiform region of an idealized squall line compared to a classic two‐moment schemeThe ML‐based P3‐like microphysics scheme runs stable and provides meaningful results in three‐dimensional real‐case simulations [ABSTRACT FROM AUTHOR]
- Published
- 2024
- Full Text
- View/download PDF