Generating observation guided ensembles for data assimilation with denoising diffusion probabilistic model

TL;DR

This paper introduces a novel ensemble data assimilation approach utilizing denoising diffusion probabilistic models to generate pseudo ensembles, improving performance especially with imperfect simulation models by leveraging the model's ability to produce diverse ensembles close to observations.

Contribution

It proposes a new method that uses diffusion models to generate ensembles for data assimilation, enhancing robustness against model imperfections.

Findings

Outperforms traditional ensemble methods with imperfect models

Generates ensembles closely aligned with noisy observations

Leverages variance in generated ensembles for better accuracy

Abstract

This paper presents an ensemble data assimilation method using the pseudo ensembles generated by denoising diffusion probabilistic model. Since the model is trained against noisy and sparse observation data, this model can produce divergent ensembles close to observations. Thanks to the variance in generated ensembles, our proposed method displays better performance than the well-established ensemble data assimilation method when the simulation model is imperfect.