Physics-informed machine learning-based cloud microphysics parameterization for earth system models

Conference item

In this study, we develop a physics-informed machine learning (ML)-based cloud microphysics parameterization for the ICON model. By training the ML parameterization on high-resolution simulation data, we aim to improve Earth System Models (ESMs) in comparison to traditional parameterization schemes. We investigate the usage of a multilayer perceptron (MLP) with feature engineering and physics-constraints, and use explainability techniques to understand the relationship between input features and model output. Our novel approach yields promising results, with the physics-informed ML-based cloud microphysics parameterization achieving an R² score up to 0.777 for an individual feature. Additionally, we demonstrate a notable improvement in the overall performance in comparison to a baseline MLP, increasing its average R² score from 0.290 to 0.613 across all variables. This approach to improve the representation of cloud microphysics in ESMs promises to enhance climate projections, contributing to a better understanding of climate change.

Authors: Sarauer, E; Schwabe, M; Weiss, P; Lauer, A; Stier, P

• Publication status: Published
• Peer review status: Peer reviewed
• Publisher: International Conference on Learning Representations
• Host title: 12th International Conference on Learning Representations (ICLR 2024)
• Article number: 35
• Publication date: 2024-03-01
• Acceptance date: 2024-03-01
• Event title: 12th International Conference on Learning Representations (ICLR 2024): Tackling Climate Change with Machine Learning workshop
• Event location: Vienna, Austria
• Event website: https://iclr.cc/Conferences/2024
• Event start date: 2024-05-07
• Event end date: 2024-05-11
• Language: English