Combining data assimilation and machine learning to emulate a dynamical model from sparse and noisy observations: A case study with the Lorenz 96 model

• Published in: Journal of computational science Vol. 44; p. 101171
• Main Authors: Brajard, Julien, Carrassi, Alberto, Bocquet, Marc, Bertino, Laurent
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.07.2020; Elsevier
• Subjects: Data assimilation; Dynamical model; Emulator; Geophysics; Machine learning; Observations; Physics; Dynamical model; Emulator; Observations; Machine learning; Data assimilation

•We address the problem of dynamics emulation from sparse and noisy observations.•An algorithm combining data assimilation and machine learning is applied.•The approach is tested on the chaotic 40-variables Lorenz 96 model.•The output of the algorithm is a data-driven surrogate numerical model.•The...