Appa: Bending Weather Dynamics with Latent Diffusion Models for Global Data Assimilation

Saved in:
Bibliographic Details
Title: Appa: Bending Weather Dynamics with Latent Diffusion Models for Global Data Assimilation
Authors: Andry, Gérôme, Lewin, Sacha, Rozet, François, Rochman, Omer, Mangeleer, Victor, Pirlet, Matthias, Faulx, Elise, Grégoire, Marilaure, Louppe, Gilles
Source: Machine Learning and the Physical Sciences Workshop (NeurIPS 2025), San Diego, United States - California [US-CA], 06/12/2025
Publication Year: 2025
Subject Terms: Computer Science - Learning, Physics - Atmospheric and Oceanic Physics, Engineering, computing & technology, Computer science, Ingénierie, informatique & technologie, Sciences informatiques
Description: Deep learning has advanced weather forecasting, but accurate predictions first require identifying the current state of the atmosphere from observational data. In this work, we introduce Appa, a score-based data assimilation model generating global atmospheric trajectories at 0.25\si{\degree} resolution and 1-hour intervals. Powered by a 565M-parameter latent diffusion model trained on ERA5, Appa can be conditioned on arbitrary observations to infer plausible trajectories, without retraining. Our probabilistic framework handles reanalysis, filtering, and forecasting, within a single model, producing physically consistent reconstructions from various inputs. Results establish latent score-based data assimilation as a promising foundation for future global atmospheric modeling systems.
Document Type: conference paper not in proceedings
http://purl.org/coar/resource_type/c_18cp
conferencePaper
peer reviewed
Language: English
DOI: 10.48550/arXiv.2504.18720
Access URL: https://orbi.uliege.be/handle/2268/342181
Rights: open access
http://purl.org/coar/access_right/c_abf2
info:eu-repo/semantics/openAccess
Accession Number: edsorb.342181
Database: ORBi
Description
Abstract:Deep learning has advanced weather forecasting, but accurate predictions first require identifying the current state of the atmosphere from observational data. In this work, we introduce Appa, a score-based data assimilation model generating global atmospheric trajectories at 0.25\si{\degree} resolution and 1-hour intervals. Powered by a 565M-parameter latent diffusion model trained on ERA5, Appa can be conditioned on arbitrary observations to infer plausible trajectories, without retraining. Our probabilistic framework handles reanalysis, filtering, and forecasting, within a single model, producing physically consistent reconstructions from various inputs. Results establish latent score-based data assimilation as a promising foundation for future global atmospheric modeling systems.
DOI:10.48550/arXiv.2504.18720