Skillful Global Ocean Emulation and the Role of Correlation-Aware Loss
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2026
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866911608994791424 |
|---|---|
| author | Agarwal, Niraj Smith, Timothy A. Frolov, Sergey Slivinski, Laura C. |
| author_facet | Agarwal, Niraj Smith, Timothy A. Frolov, Sergey Slivinski, Laura C. |
| contents | Machine learning emulators have shown extraordinary skill in forecasting atmospheric states, and their application to global ocean dynamics offers similar promise. Here, we adapt the GraphCast architecture into a dedicated ocean-only emulator, driven by prescribed atmospheric conditions, for medium-range predictions. The emulator is trained on NOAA's UFS-Replay dataset. Using a 24 hour time step, single initial condition, and without using autoregressive training, we produce an emulator that provides skillful forecasts for 10-15 day lead times. We further demonstrate the use of Mahalanobis distance as loss that improves the forecast skill compared to the Mean Squared Error loss by explicitly accounting for the correlations between tendencies of the target variables. Using spatial correlation analysis of the forecasted fields, we also show that the proposed correlation-aware loss acts as a statistical-dynamical regularizer for the slow, correlated dynamics of the global oceans, offering a better background forecast for downstream tasks like data assimilation. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2604_18727 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Skillful Global Ocean Emulation and the Role of Correlation-Aware Loss Agarwal, Niraj Smith, Timothy A. Frolov, Sergey Slivinski, Laura C. Atmospheric and Oceanic Physics Artificial Intelligence Chaotic Dynamics Machine learning emulators have shown extraordinary skill in forecasting atmospheric states, and their application to global ocean dynamics offers similar promise. Here, we adapt the GraphCast architecture into a dedicated ocean-only emulator, driven by prescribed atmospheric conditions, for medium-range predictions. The emulator is trained on NOAA's UFS-Replay dataset. Using a 24 hour time step, single initial condition, and without using autoregressive training, we produce an emulator that provides skillful forecasts for 10-15 day lead times. We further demonstrate the use of Mahalanobis distance as loss that improves the forecast skill compared to the Mean Squared Error loss by explicitly accounting for the correlations between tendencies of the target variables. Using spatial correlation analysis of the forecasted fields, we also show that the proposed correlation-aware loss acts as a statistical-dynamical regularizer for the slow, correlated dynamics of the global oceans, offering a better background forecast for downstream tasks like data assimilation. |
| title | Skillful Global Ocean Emulation and the Role of Correlation-Aware Loss |
| topic | Atmospheric and Oceanic Physics Artificial Intelligence Chaotic Dynamics |
| url | https://arxiv.org/abs/2604.18727 |