Physically Consistent Global Atmospheric Data Assimilation with Machine Learning in Latent Space

TL;DR

This paper introduces Latent Data Assimilation (LDA), a novel machine learning framework that performs Bayesian data assimilation in a learned latent space, improving atmospheric analysis accuracy and robustness over traditional methods.

Contribution

LDA leverages autoencoders to capture nonlinear physical relationships, enabling flow-dependent physical consistency without explicit physical modeling, and enhances forecast skill.

Findings

LDA produces more balanced and accurate atmospheric analyses.

LDA improves forecast skill compared to traditional model-space DA.

LDA remains effective even with autoencoders trained on imperfect data.

Abstract

Data assimilation (DA) integrates observations with model forecasts to produce optimized atmospheric states, whose physical consistency is critical for stable weather forecasting and reliable climate research. Traditional Bayesian DA methods enforce these nonlinear, flow-dependent physical constraints through empirical and tunable covariance structures, but with limited accuracy and robustness. Here, we introduce Latent Data Assimilation (LDA), a framework that performs Bayesian DA in a latent space learned from multivariate global atmospheric data via an autoencoder. We demonstrate that the autoencoder can largely capture nonlinear physical relationships, enabling LDA to produce balanced analyses without explicitly modeling physical constraints. Assimilation in latent space also improves both analysis quality and forecast skill compared to traditional model-space DA, under both idealized and real observational settings. Furthermore, LDA exhibits strong robustness across latent dimensions and remains effective even when the autoencoder is trained on inaccurate but physically realistic forecasts, highlighting its flexibility for real-world applications.