HClimRep-Ocean: A Global Ocean Emulator on an Unstructured Mesh

2026-09-23

AI Summary

This study addresses the limitation of existing ocean machine learning models that rely on latitude–longitude grids, which struggle to represent complex coastlines and unstructured meshes. To overcome this, we propose the first global ocean emulator built upon the native unstructured grid of FESOM2. Trained on AWI-CM3 data, the model achieves independent predictions without atmospheric forcing while effectively supporting local mesh refinement and complex boundary representation. Evaluated on the OceanBench benchmark, our approach outperforms all reference baselines in 30-day ocean current forecasting, achieving the lowest root mean square error (RMSE). These results validate the superiority and competitiveness of the native unstructured grid paradigm for data-driven ocean simulation.

Abstract

Machine-learning (ML) emulators for atmospheric processes have advanced rapidly in recent years, transforming weather forecasting. Although early ML ocean forecasting models now exist, they remain less developed than their atmospheric counterparts. Unlike the atmosphere, much of the ocean's kinetic energy resides in mesoscale eddies whose characteristic spatial scales are approximately an order of magnitude smaller than those of comparable atmospheric features. Moreover, complex coastlines, narrow straits, and ice-covered seas make boundary representation a central challenge that atmospheric models do not face. Consequently, numerical ocean simulations commonly use locally refined or even completely unstructured meshes. However, their data-driven counterparts have so far been built around latitude-longitude grids. We present HClimRep-Ocean, an ocean emulator that operates directly on the native unstructured mesh of FESOM2. The emulator is trained on a 209-year AWI-CM3 control integration and is run without atmospheric forcing, receiving the atmospheric state only at initialisation time, which isolates the predictability carried by the ocean state itself. Skill is strongly field-dependent: for currents, HClimRep-Ocean outperforms every reference at 30 day forecast, whereas for temperature and salinity a damped-anomaly persistence forecast remains the more accurate estimator. This behaviour is physically interpretable: current variability is largely geostrophic and internally generated, whereas sea-surface temperature and salinity fluctuations are driven by atmospheric forcing through weather state. Evaluated independently on the OceanBench benchmark, a reanalysis-trained variant of HClimRep-Ocean achieves the lowest RMSE against GLORYS reanalysis among all assessed systems, confirming the competitiveness of the native-mesh approach.

Problem

• ocean emulator
• unstructured mesh
• mesoscale eddies
• machine learning
• boundary representation

Innovation

• unstructured mesh
• ocean emulator
• machine learning
• FESOM2
• mesoscale eddies