PDE foundation models are skillful AI weather emulators for the Martian atmosphere
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866908836783194112 |
|---|---|
| author | Schmude, Johannes Roy, Sujit Wang, Liping van Kessel, Theodore Klein, Levente Freitag, Marcus Bentivegna, Eloisa Manson-Sawko, Robert Lutjens, Bjorn Maskey, Manil Watson, Campbell Ramachandran, Rahul Bernabe-Moreno, Juan |
| author_facet | Schmude, Johannes Roy, Sujit Wang, Liping van Kessel, Theodore Klein, Levente Freitag, Marcus Bentivegna, Eloisa Manson-Sawko, Robert Lutjens, Bjorn Maskey, Manil Watson, Campbell Ramachandran, Rahul Bernabe-Moreno, Juan |
| contents | We show that AI foundation models that are pretrained on numerical solutions to a diverse corpus of partial differential equations can be adapted and fine-tuned to obtain skillful predictive weather emulators for the Martian atmosphere. We base our work on the Poseidon PDE foundation model for two-dimensional systems. We develop a method to extend Poseidon from two to three dimensions while keeping the pretraining information. Moreover, we investigate the performance of the model in the presence of sparse initial conditions. Our results make use of four Martian years (approx.~34 GB) of training data and a median compute budget of 13 GPU hours. We find that the combination of pretraining and model extension yields a performance increase of 34.4\% on a held-out year. This shows that PDEs-FMs can not only approximate solutions to (other) PDEs but also anchor models for real-world problems with complex interactions that lack a sufficient amount of training data or a suitable compute budget. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_15004 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | PDE foundation models are skillful AI weather emulators for the Martian atmosphere Schmude, Johannes Roy, Sujit Wang, Liping van Kessel, Theodore Klein, Levente Freitag, Marcus Bentivegna, Eloisa Manson-Sawko, Robert Lutjens, Bjorn Maskey, Manil Watson, Campbell Ramachandran, Rahul Bernabe-Moreno, Juan Machine Learning Atmospheric and Oceanic Physics We show that AI foundation models that are pretrained on numerical solutions to a diverse corpus of partial differential equations can be adapted and fine-tuned to obtain skillful predictive weather emulators for the Martian atmosphere. We base our work on the Poseidon PDE foundation model for two-dimensional systems. We develop a method to extend Poseidon from two to three dimensions while keeping the pretraining information. Moreover, we investigate the performance of the model in the presence of sparse initial conditions. Our results make use of four Martian years (approx.~34 GB) of training data and a median compute budget of 13 GPU hours. We find that the combination of pretraining and model extension yields a performance increase of 34.4\% on a held-out year. This shows that PDEs-FMs can not only approximate solutions to (other) PDEs but also anchor models for real-world problems with complex interactions that lack a sufficient amount of training data or a suitable compute budget. |
| title | PDE foundation models are skillful AI weather emulators for the Martian atmosphere |
| topic | Machine Learning Atmospheric and Oceanic Physics |
| url | https://arxiv.org/abs/2602.15004 |