Robust Ocean Subgrid-Scale Parameterizations Using Fourier Neural Operators

TL;DR

This paper introduces Fourier Neural Operators for ocean subgrid-scale parameterizations, improving accuracy and generalizability in climate simulations by effectively modeling small-scale processes.

Contribution

It presents a novel neural network approach using Fourier Neural Operators to enhance ocean subgrid-scale parameterizations in climate models.

Findings

Fourier Neural Operators outperform traditional methods in accuracy.

The approach demonstrates strong generalization across different ocean conditions.

Neural networks in the frequency domain show promising potential for climate modeling.

Abstract

In climate simulations, small-scale processes shape ocean dynamics but remain computationally expensive to resolve directly. For this reason, their contributions are commonly approximated using empirical parameterizations, which lead to significant errors in long-term projections. In this work, we develop parameterizations based on Fourier Neural Operators, showcasing their accuracy and generalizability in comparison to other approaches. Finally, we discuss the potential and limitations of neural networks operating in the frequency domain, paving the way for future investigation.