A Probabilistic Framework for Learning Non‐Intrusive Corrections to Long‐Time Climate Simulations From Short‐Time Training Data
Name
J Adv Model Earth Syst - 2026 - Barthel Sorensen - A Probabilistic Framework for Learning Non‐Intrusive Corrections to.pdf
Description
Published version
Size
6.66 MB
Format
Adobe PDF
Checksum (MD5)
4fd489598d0b7c281d5a74d10553c40f
Author(s)
Barthel Sorensen, Benedikt
Zepeda‐Núñez, Leonardo
Lopez‐Gomez, Ignacio
Wan, Zhong Y
Carver, Rob
Sha, Fei
Sapsis, Themistoklis P
Date Issued
January 8, 2026
Journal
Journal of Advances in Modeling Earth Systems
Publisher
American Geophysical Union
Citation
Barthel Sorensen, B., Zepeda-Núñez, L., Lopez-Gomez, I., Wan, Z. Y., Carver, R., Sha, F., & Sapsis, T. P. (2026). A probabilistic framework for learning non-intrusive corrections to long-time climate simulations from short-time training data. Journal of Advances in Modeling Earth Systems, 18, e2024MS004755.
Version
Final published version
Abstract
Despite advances in high performance computing, accurate numerical simulations of global
atmospheric dynamics remain a challenge. The resolution required to fully resolve the vast range scales as well
as the strong coupling with—often not fully‐understood—physics renders such simulations computationally
infeasible over time horizons relevant for long‐term climate risk assessment. While data‐driven
parameterizations have shown some promise of alleviating these obstacles, the scarcity of high‐quality training
data and their lack of long‐term stability typically hinders their ability to capture the risk of rare extreme events.
In this work we present a general strategy for training variational (probabilistic) neural network models to non‐
intrusively correct under‐resolved long‐time simulations of turbulent climate systems. The approach is based on
the paradigm introduced by Barthel Sorensen et al. (2024, https://doi.org/10.1029/2023ms004122) which
involves training a post‐processing correction operator on under‐resolved simulations nudged toward a high‐
fidelity reference. Our variational framework enables us to learn the dynamics of the underlying system from
very little training data and thus drastically improve the extrapolation capabilities of the previous deterministic
state‐of‐the art—even when the statistics of that training data are far from converged. We investigate and
compare three recently introduced variational network architectures and illustrate the benefits of our approach
on an anisotropic quasi‐geostrophic flow. For this prototype model our approach is able to not only accurately
capture global statistics, but also the anistropic regional variation and the statistics of multiple extreme event
metrics—demonstrating significant improvement over previously introduced deterministic architectures.
Terms of Use
Creative Commons Attribution
Persistent DSpace Link
DOI of Published Version
https://doi.org/10.1029/2024MS004755
