Bibliography

J. van der Voort, M. Verlaan and H. Kekkonen. Balancing Accuracy and Speed: A Multi-Fidelity Ensemble Kalman Filter with a Machine Learning Surrogate Model, 2025. [under review]  arxiv link

In this paper we investigate the use of a Multi-Fidelity Ensemble Kalman Filter (MF-EnKF) in which the low-fidelity model is a Machine Learning model, instead of a traditional low-resolution model. Results with the Lorenz-2005 and Quasi-Geostrophic model show that we can achieve improved accuracy within the same computational budget compared to the standard EnKF. Our method contributes to increasing the effective ensemble size in the EnKF, which improves the estimation of the initial condition in fields such as meteorology and oceanography.