NeuralOM: Neural Ocean Model for Subseasonal-to-Seasonal Simulation

TL;DR

NeuralOM introduces a neural operator framework with residual correction and physics-guided graph networks to improve long-term ocean simulation accuracy and stability, outperforming existing models in subseasonal-to-seasonal forecasting.

Contribution

The paper presents NeuralOM, a novel neural operator architecture combining residual correction and physics-guided graph networks for enhanced long-term physical system simulation.

Findings

NeuralOM achieves 13.3% lower RMSE at 60-day lead time.

It surpasses state-of-the-art models in forecast accuracy and stability.

NeuralOM effectively simulates extreme ocean events.

Abstract

Long-term, high-fidelity simulation of slow-changing physical systems, such as the ocean and climate, presents a fundamental challenge in scientific computing. Traditional autoregressive machine learning models often fail in these tasks as minor errors accumulate and lead to rapid forecast degradation. To address this problem, we propose NeuralOM, a general neural operator framework designed for simulating complex, slow-changing dynamics. NeuralOM's core consists of two key innovations: (1) a Progressive Residual Correction Framework that decomposes the forecasting task into a series of fine-grained refinement steps, effectively suppressing long-term error accumulation; and (2) a Physics-Guided Graph Network whose built-in adaptive messaging mechanism explicitly models multi-scale physical interactions, such as gradient-driven flows and multiplicative couplings, thereby enhancing physical consistency while maintaining computational efficiency. We validate NeuralOM on the challenging task of global Subseasonal-to-Seasonal (S2S) ocean simulation. Extensive experiments demonstrate that NeuralOM not only surpasses state-of-the-art models in forecast accuracy and long-term stability, but also excels in simulating extreme events. For instance, at a 60-day lead time, NeuralOM achieves a 13.3% lower RMSE compared to the best-performing baseline, offering a stable, efficient, and physically-aware paradigm for data-driven scientific computing. Code link: https://github.com/YuanGao-YG/NeuralOM.