Multimodal Spatiotemporal Atmospheric Data Assimilation with Latent Flow-matching

2026-08-05

AI Summary

This work addresses the challenge of efficiently integrating sparse, multi-source atmospheric observations into temporally and spatially coherent full-state trajectories—a task where traditional data assimilation methods often fall short. The authors propose a unified data assimilation framework based on latent video flow matching, which, for the first time in this domain, leverages continuous trajectory generation and flow matching to propagate information naturally through observation gaps without requiring an explicit dynamical model. By constructing a spatiotemporal prior from ERA5 reanalysis data and fusing NOAA radiosonde and surface observations via Bayesian posterior sampling, the method seamlessly unifies diverse assimilation tasks such as filtering and smoothing. Experiments demonstrate that high-accuracy ensemble forecasts of the full atmospheric state can be generated from sparse observations alone, achieving performance comparable to state-of-the-art observation-to-forecast models.

Abstract

Data assimilation (DA) uses Bayesian inference to update the state of a numerical forecast model with observed data. In this study, we propose a fundamentally different, unified approach to atmospheric data assimilation. We use latent video flow-matching to sample temporally consistent trajectories from a prior trained using ERA5 reanalysis (69 variables over an 8-day window). We also use posterior sampling to assimilate real observation sources, such as those from the NOAA Integrated Global Radiosonde Archive and the Integrated Surface Database. Because the prior generates a continuous trajectory, it naturally propagates information between observed and unobserved frames. Therefore, we can perform various DA tasks, such as filtering and smoothing, simply by changing the observed frames. Moreover, we generate full-state ensemble forecasts directly from sparse observations, achieving performance competitive with state-of-the-art observation-to-forecast models.

Problem

• Data Assimilation
• Multimodal Observations
• Spatiotemporal Consistency
• Atmospheric Modeling
• Ensemble Forecasting

Innovation

• latent flow-matching
• data assimilation
• spatiotemporal modeling
• ensemble forecasting
• multimodal atmospheric data