LEVDA: Latent Ensemble Variational Data Assimilation via Differentiable Dynamics

TL;DR

LEVDA introduces a novel latent-space variational data assimilation method using differentiable neural surrogates, enabling efficient, accurate, and uncertainty-aware geophysical forecasts with irregular observations and reduced computational costs.

Contribution

It develops LEVDA, a new ensemble variational smoother operating in a learned latent space, eliminating the need for adjoint models and handling irregular sampling effectively.

Findings

LEVDA outperforms existing latent filtering methods under sparse observations.

It achieves higher accuracy and better uncertainty quantification than full-state 4DEnVar.

LEVDA offers improved computational efficiency in geophysical data assimilation.

Abstract

Long-range geophysical forecasts are fundamentally limited by chaotic dynamics and numerical errors. While data assimilation can mitigate these issues, classical variational smoothers require computationally expensive tangent-linear and adjoint models. Conversely, recent efficient latent filtering methods often enforce weak trajectory-level constraints and assume fixed observation grids. To bridge this gap, we propose Latent Ensemble Variational Data Assimilation (LEVDA), an ensemble-space variational smoother that operates in the low-dimensional latent space of a pretrained differentiable neural dynamics surrogate. By performing four-dimensional ensemble-variational (4DEnVar) optimization within an ensemble subspace, LEVDA jointly assimilates states and unknown parameters without the need for adjoint code or auxiliary observation-to-latent encoders. Leveraging the fully differentiable, continuous-in-time-and-space nature of the surrogate, LEVDA naturally accommodates highly irregular sampling at arbitrary spatiotemporal locations. Across three challenging geophysical benchmarks, LEVDA matches or outperforms state-of-the-art latent filtering baselines under severe observational sparsity while providing more reliable uncertainty quantification. Simultaneously, it achieves substantially improved assimilation accuracy and computational efficiency compared to full-state 4DEnVar.