Use of neural networks for stable, accurate and physically consistent parameterization of subgrid atmospheric processes with good performance at reduced precision

Abstract

A promising approach to improve climate-model simulations is to replace traditional subgrid parameterizations based on simplified physical models by machine learning algorithms that are data-driven. However, neural networks (NNs) often lead to instabilities and climate drift when coupled to an atmospheric model. Here we learn an NN parameterization from a high-resolution atmospheric simulation in an idealized domain by coarse graining the model equations and output. The NN parameterization has a structure that ensures physical constraints are respected, and it leads to stable simulations that replicate the climate of the high-resolution simulation with similar accuracy to a successful random-forest parameterization while needing far less memory. We find that the simulations are stable for a variety of NN architectures and horizontal resolutions, and that an NN with substantially reduced numerical precision could decrease computational costs without affecting the quality of simulations.

arXiv comment23 pages, 6 figures (2 figures in main file)

References in corpus (5)

• Deep learning to represent sub-grid processes in climate models
• Stable machine-learning parameterization of subgrid processes for climate modeling at a range of resolutions
• Using machine learning to parameterize moist convection: potential for modeling of climate, climate change and extreme events
• Spatially Extended Tests of a Neural Network Parametrization Trained by Coarse-graining
• Interpreting and Stabilizing Machine-learning Parametrizations of Convection

Cited by in corpus (12)

• Bridging observation, theory and numerical simulation of the ocean using Machine Learning
• Deep Learning Based Cloud Cover Parameterization for ICON
• Machine learning emulation of gravity wave drag in numerical weather forecasting
• Non-local parameterization of atmospheric subgrid processes with neural networks
• Differentiable Programming for Earth System Modeling
• Neural Closure Models for Dynamical Systems
• MieAI: A neural network for calculating optical properties of internally mixed aerosol in atmospheric models
• Interpretable multiscale Machine Learning-Based Parameterizations of Convection for ICON
• Generalizability of reservoir computing for flux-driven two-dimensional convection
• Reconstructing Historical Climate Fields With Deep Learning
• Interpretable structural model error discovery from sparse assimilation increments using spectral bias-reduced neural networks: A quasi-geostrophic turbulence test case
• Reduced Data-Driven Turbulence Closure for Capturing Long-Term Statistics