Attention-based Convolutional Autoencoders for 3D-Variational Data Assimilation

We propose a new ‘Bi-Reduced Space’ approach to solving 3D Variational Data Assimilation using Convolutional Autoencoders. We prove that our approach has the same solution as previous methods but has significantly lower computational complexity; in other words, we reduce the computational cost witho...

Celý popis

Uložené v:
Podrobná bibliografia
Vydané v:Computer methods in applied mechanics and engineering Ročník 372; s. 113291
Hlavní autori: Mack, Julian, Arcucci, Rossella, Molina-Solana, Miguel, Guo, Yi-Ke
Médium: Journal Article
Jazyk:English
Vydavateľské údaje: Elsevier B.V 01.12.2020
Predmet:
Attention networks
Convolutional Autoencoders
Variational Data Assimilation
Convolutional Autoencoders
Attention networks
Variational Data Assimilation
ISSN:0045-7825, 1879-2138
On-line prístup:Získať plný text
Tagy: Pridať tag
Žiadne tagy, Buďte prvý, kto otaguje tento záznam!
Popis
Shrnutí:We propose a new ‘Bi-Reduced Space’ approach to solving 3D Variational Data Assimilation using Convolutional Autoencoders. We prove that our approach has the same solution as previous methods but has significantly lower computational complexity; in other words, we reduce the computational cost without affecting the data assimilation accuracy. We tested our proposal with data from a real-world application: a pollution model of a site in Elephant and Castle (London, UK) and found that we could (1) reduce the size of the background covariance matrix representation by O(103), and (2) increase our data assimilation accuracy with respect to existing reduced space methods. •We propose a new ‘Bi-Reduced Space’ approach to solving 3D Variational Data Assimilation.•We use Convolutional Autoencoders to create the reduced space for solving 3D Var.•We prove that our approach has the same solution as previous methods for reducing 3D Var space.•We prove that our approach has significantly lower computational complexity of previous methods.•We tested the new method with data from a real-world application: a pollution model in London.
ISSN:0045-7825
1879-2138
DOI:10.1016/j.cma.2020.113291