Accurate deep learning-based filtering for chaotic dynamics by identifying instabilities without an ensemble
Chaos: An Interdisciplinary Journal of Nonlinear Science · 01.09.2024 · 34 (9)
DOI: 10.1063/5.0230837 ↗Kurz gefasst
Fachartikel in der Zeitschrift Chaos: An Interdisciplinary Journal of Nonlinear Science, online erschienen am 30.09.2024. Verfasst von Marc Bocquet und 7 weiteren Autor:innen. Beteiligte Einrichtung: CEREA, École des Ponts and EDF R&D, Île-de-France, France. Die Arbeit wurde bisher 16-mal zitiert (Crossref) und nennt 65 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Schmidt Sciences.
Abstract (Originalsprache)
We investigate the ability to discover data assimilation (DA) schemes meant for chaotic dynamics with deep learning. The focus is on learning the analysis step of sequential DA, from state trajectories and their observations, using a simple residual convolutional neural network, while assuming the dynamics to be known. Experiments are performed with the Lorenz 96 dynamics, which display spatiotemporal chaos and for which solid benchmarks for DA performance exist. The accuracy of the states obtained from the learned analysis approaches that of the best possibly tuned ensemble Kalman filter and is far better than that of variational DA alternatives. Critically, this can be achieved while propagating even just a single state in the forecast step. We investigate the reason for achieving ensemble filtering accuracy without an ensemble. We diagnose that the analysis scheme actually identifies key dynamical perturbations, mildly aligned with the unstable subspace, from the forecast state alone, without any ensemble-based covariances representation. This reveals that the analysis scheme has learned some multiplicative ergodic theorem associated to the DA process seen as a non-autonomous random dynamical system.
Autor:innen und Institutionen
- CEREA, École des Ponts and EDF R&D, Île-de-France, France.
- Department of Meteorology and National Centre for Earth Observation, University of Reading, Earley Gate, PO Box 243, Reading RG6 6BB, United Kingdom.
- Department of Physics and Astronomy, University of Bologna, Viale Carlo Berti Pichat, 6/2, Bologna 40127, Italy.
ORCID-Profile: Marc Bocquet, Alban Farchi, Tobias S Finn, Charlotte Durand, Sibo Cheng, Yumeng Chen, Ivo Pasmans, Alberto Carrassi
Förderung
- Schmidt Sciences
Zitieren
APA
Bocquet, M., Farchi, A., Finn, T. S., Durand, C., Cheng, S., Chen, Y., Pasmans, I., & Carrassi, A. (2024). Accurate deep learning-based filtering for chaotic dynamics by identifying instabilities without an ensemble. Chaos: An Interdisciplinary Journal of Nonlinear Science, 34(9). https://doi.org/10.1063/5.0230837
Vancouver
Bocquet M, Farchi A, Finn TS, Durand C, Cheng S, Chen Y, et al. Accurate deep learning-based filtering for chaotic dynamics by identifying instabilities without an ensemble. Chaos: An Interdisciplinary Journal of Nonlinear Science. 2024;34(9). doi:10.1063/5.0230837
BibTeX
@article{Bocquet2024,
author = {Bocquet, Marc and Farchi, Alban and Finn, Tobias S and Durand, Charlotte and Cheng, Sibo and Chen, Yumeng and Pasmans, Ivo and Carrassi, Alberto},
title = {Accurate deep learning-based filtering for chaotic dynamics by identifying instabilities without an ensemble},
journal = {Chaos: An Interdisciplinary Journal of Nonlinear Science},
year = {2024},
volume = {34},
number = {9},
doi = {10.1063/5.0230837}
}
EUVIMED