Generative Lagrangian data assimilation for ocean dynamics under extreme sparsity
Niloofar Asefi, Leonard Lupin-Jimenez, Tianning Wu, Ruoying He, Ashesh Chattopadhyay · 2025
Read Generative Lagrangian data assimilation for ocean dynamics under extreme sparsity by Niloofar Asefi, Leonard Lupin-Jimenez, Tianning Wu, Ruoying He, Ashesh Chattopadhyay (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2507-06479.txt.
Description
Asefi and four co-authors present a study from twenty twenty-five on rebuilding ocean states from very sparse drift data. Ocean measures are sparse and scattered, most of all below the surface. That sparsity blocks forecasts of eddies and rogue waves. Their method pairs a neural operator with a generative diffusion model. The operator gives a first draft, and the generator refines it. The pair capture fine, fast-changing patterns even at extreme sparsity. Tests cover lab benchmarks, fake float data, and real satellite records.
The method holds up at ninety-nine percent missing data on fake sets. On real satellite data it copes with even higher sparsity. It beats other deep learning baselines under those harsh limits. The authors aim at better ocean forecasts from thin data.
Catalog record
| CREATOR | Niloofar Asefi, Leonard Lupin-Jimenez, Tianning Wu, Ruoying He, Ashesh Chattopadhyay |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 39 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2507-06479