Uncertainty-aware data assimilation through variational inference

TL;DR

This paper introduces a variational inference extension to data assimilation models, allowing for uncertainty quantification with Gaussian predictions, demonstrated on Lorenz-96 dynamics to improve calibration and performance over longer data windows.

Contribution

It presents a novel variational inference-based approach for uncertainty-aware data assimilation, extending deterministic models to produce probabilistic Gaussian predictions.

Findings

Achieves nearly perfectly calibrated predictions.

Enhances data assimilation performance with longer windows.

Integrates seamlessly into existing variational pipelines.

Abstract

Data assimilation, consisting in the combination of a dynamical model with a set of noisy and incomplete observations in order to infer the state of a system over time, involves uncertainty in most settings. Building upon an existing deterministic machine learning approach, we propose a variational inference-based extension in which the predicted state follows a multivariate Gaussian distribution. Using the chaotic Lorenz-96 dynamics as a testing ground, we show that our new model enables to obtain nearly perfectly calibrated predictions, and can be integrated in a wider variational data assimilation pipeline in order to achieve greater benefit from increasing lengths of data assimilation windows. Our code is available at https://github.com/anthony-frion/Stochastic_CODA.