Model error covariance estimation in particle and ensemble Kalman filters using an online expectation-maximization algorithm

Autores
Cocucci, Tadeo Javier; Pulido, Manuel Arturo; Lucini, María Magdalena; Tandeo, Pierre
Año de publicación
2021
Idioma
inglés
Tipo de recurso
artículo
Estado
versión publicada
Descripción
Fil: Cocucci, Tadeo Javier. Universidad Nacional del Nordeste. Facultad de Ciencias Exactas y Naturales y Agrimensura; Argentina.
Fil: Pulido, Manuel Arturo. Universidad Nacional del Nordeste. Facultad de Ciencias Exactas y Naturales y Agrimensura; Argentina.
Fil: Pulido, Manuel Arturo. Consejo Nacional de Investigaciones Científicas y Técnicas; Argentina.
Fil: Lucini, María Magdalena. Universidad Nacional del Nordeste. Facultad de Ciencias Exactas y Naturales y Agrimensura; Argentina.
Fil: Lucini, María Magdalena. Consejo Nacional de Investigaciones Científicas y Técnicas; Argentina.
Fil: Tandeo, Pierre. Centre National de la Recherche Scientifique. Institut Mines-Télécom Atlantique. Laboratoire des Sciences et Techniques de l'Information, de la Communication et de la Connaissance; Francia.
The performance of ensemble-based data assimilation techniques that estimate the state of a dynamical system from partial observations depends crucially on theprescribeduncertaintyofthemodeldynamicsandoftheobservations.These are not usually knownandhavetobeinferred.Manyapproacheshavebeenproposed to tackle this problem, including fully Bayesian, likelihood maximization and innovation-based techniques. This work focuses on maximization of the likelihood function via the expectation–maximization (EM) algorithm to infer the model error covariance combined with ensemble Kalman filters and particle filters to estimate the state. The classical application of the EM algorithm in a data assimilation context involves filtering and smoothing a fixed batch of observations in order to complete a single iteration. This is an inconvenience whenusing sequential filtering in high-dimensional applications. Motivated by this, an adaptation of the algorithm that can process observations and update the parameters on the fly, with some underlying simplifications, is presented. The proposed technique was evaluated and achieved good performance in experiments with the Lorenz-63 and Lorenz-96 dynamical systems designed to represent some common scenarios in data assimilation such as nonlinearity, chaoticity and model mis-specification.
Fuente
Quarterly Journal of the Royal Meteorological Society, 2021, vol. 147, no. 734, p. 526-543.
Materia
Expectation-maximization
Model error
Parameter estimation
Uncertainty quantification
Nivel de accesibilidad
acceso abierto
Condiciones de uso
http://creativecommons.org/licenses/by-nc-nd/2.5/ar/
Repositorio
Repositorio Institucional de la Universidad Nacional del Nordeste (UNNE)
Institución
Universidad Nacional del Nordeste
OAI Identificador
oai:repositorio.unne.edu.ar:123456789/60144
Acceder
id RIUNNE_f846bc81034dfeb63578ab85e352d155
oai_identifier_str oai:repositorio.unne.edu.ar:123456789/60144
network_acronym_str RIUNNE
repository_id_str 4871
network_name_str Repositorio Institucional de la Universidad Nacional del Nordeste (UNNE)
spelling Model error covariance estimation in particle and ensemble Kalman filters using an online expectation-maximization algorithmCocucci, Tadeo JavierPulido, Manuel ArturoLucini, María MagdalenaTandeo, PierreExpectation-maximizationModel errorParameter estimationUncertainty quantificationFil: Cocucci, Tadeo Javier. Universidad Nacional del Nordeste. Facultad de Ciencias Exactas y Naturales y Agrimensura; Argentina.Fil: Pulido, Manuel Arturo. Universidad Nacional del Nordeste. Facultad de Ciencias Exactas y Naturales y Agrimensura; Argentina.Fil: Pulido, Manuel Arturo. Consejo Nacional de Investigaciones Científicas y Técnicas; Argentina.Fil: Lucini, María Magdalena. Universidad Nacional del Nordeste. Facultad de Ciencias Exactas y Naturales y Agrimensura; Argentina.Fil: Lucini, María Magdalena. Consejo Nacional de Investigaciones Científicas y Técnicas; Argentina.Fil: Tandeo, Pierre. Centre National de la Recherche Scientifique. Institut Mines-Télécom Atlantique. Laboratoire des Sciences et Techniques de l'Information, de la Communication et de la Connaissance; Francia.The performance of ensemble-based data assimilation techniques that estimate the state of a dynamical system from partial observations depends crucially on theprescribeduncertaintyofthemodeldynamicsandoftheobservations.These are not usually knownandhavetobeinferred.Manyapproacheshavebeenproposed to tackle this problem, including fully Bayesian, likelihood maximization and innovation-based techniques. This work focuses on maximization of the likelihood function via the expectation–maximization (EM) algorithm to infer the model error covariance combined with ensemble Kalman filters and particle filters to estimate the state. The classical application of the EM algorithm in a data assimilation context involves filtering and smoothing a fixed batch of observations in order to complete a single iteration. This is an inconvenience whenusing sequential filtering in high-dimensional applications. Motivated by this, an adaptation of the algorithm that can process observations and update the parameters on the fly, with some underlying simplifications, is presented. The proposed technique was evaluated and achieved good performance in experiments with the Lorenz-63 and Lorenz-96 dynamical systems designed to represent some common scenarios in data assimilation such as nonlinearity, chaoticity and model mis-specification.Royal Meteorological Society2021info:eu-repo/semantics/articleinfo:eu-repo/semantics/publishedVersionhttp://purl.org/coar/resource_type/c_6501info:ar-repo/semantics/articuloapplication/pdfp. 526-543application/pdfCocucci, Tadeo Javier, et al., 2021. Model error covariance estimation in particle and ensemble Kalman filters using an online expectation-maximization algorithm. Quarterly Journal of the Royal Meteorological Society. Reading: Royal Meteorological Society, vol. 147, no. 734, p. 526-543. E-ISSN 1477-870X. DOI https://doi.org/10.1002/qj.39310035-9009http://repositorio.unne.edu.ar/handle/123456789/60144Quarterly Journal of the Royal Meteorological Society, 2021, vol. 147, no. 734, p. 526-543.reponame:Repositorio Institucional de la Universidad Nacional del Nordeste (UNNE)instname:Universidad Nacional del Nordesteenghttps://doi.org/10.1002/qj.3931info:eu-repo/semantics/openAccesshttp://creativecommons.org/licenses/by-nc-nd/2.5/ar/Atribución-NoComercial-SinDerivadas 2.5 Argentina2026-09-24T12:45:32Zoai:repositorio.unne.edu.ar:123456789/60144instacron:UNNEInstitucionalhttp://repositorio.unne.edu.ar/Universidad públicaNo correspondehttp://repositorio.unne.edu.ar/oaiososa@bib.unne.edu.ar;sergio.alegria@unne.edu.arArgentinaNo correspondeNo correspondeNo correspondeopendoar:48712026-09-24 12:45:32.928Repositorio Institucional de la Universidad Nacional del Nordeste (UNNE) - Universidad Nacional del Nordestefalse
dc.title.none.fl_str_mv Model error covariance estimation in particle and ensemble Kalman filters using an online expectation-maximization algorithm
title Model error covariance estimation in particle and ensemble Kalman filters using an online expectation-maximization algorithm
spellingShingle Model error covariance estimation in particle and ensemble Kalman filters using an online expectation-maximization algorithm
Cocucci, Tadeo Javier
Expectation-maximization
Model error
Parameter estimation
Uncertainty quantification
title_short Model error covariance estimation in particle and ensemble Kalman filters using an online expectation-maximization algorithm
title_full Model error covariance estimation in particle and ensemble Kalman filters using an online expectation-maximization algorithm
title_fullStr Model error covariance estimation in particle and ensemble Kalman filters using an online expectation-maximization algorithm
title_full_unstemmed Model error covariance estimation in particle and ensemble Kalman filters using an online expectation-maximization algorithm
title_sort Model error covariance estimation in particle and ensemble Kalman filters using an online expectation-maximization algorithm
dc.creator.none.fl_str_mv Cocucci, Tadeo Javier
Pulido, Manuel Arturo
Lucini, María Magdalena
Tandeo, Pierre
author Cocucci, Tadeo Javier
author_facet Cocucci, Tadeo Javier
Pulido, Manuel Arturo
Lucini, María Magdalena
Tandeo, Pierre
author_role author
author2 Pulido, Manuel Arturo
Lucini, María Magdalena
Tandeo, Pierre
author2_role author
author
author
dc.subject.none.fl_str_mv Expectation-maximization
Model error
Parameter estimation
Uncertainty quantification
topic Expectation-maximization
Model error
Parameter estimation
Uncertainty quantification
dc.description.none.fl_txt_mv Fil: Cocucci, Tadeo Javier. Universidad Nacional del Nordeste. Facultad de Ciencias Exactas y Naturales y Agrimensura; Argentina.
Fil: Pulido, Manuel Arturo. Universidad Nacional del Nordeste. Facultad de Ciencias Exactas y Naturales y Agrimensura; Argentina.
Fil: Pulido, Manuel Arturo. Consejo Nacional de Investigaciones Científicas y Técnicas; Argentina.
Fil: Lucini, María Magdalena. Universidad Nacional del Nordeste. Facultad de Ciencias Exactas y Naturales y Agrimensura; Argentina.
Fil: Lucini, María Magdalena. Consejo Nacional de Investigaciones Científicas y Técnicas; Argentina.
Fil: Tandeo, Pierre. Centre National de la Recherche Scientifique. Institut Mines-Télécom Atlantique. Laboratoire des Sciences et Techniques de l'Information, de la Communication et de la Connaissance; Francia.
The performance of ensemble-based data assimilation techniques that estimate the state of a dynamical system from partial observations depends crucially on theprescribeduncertaintyofthemodeldynamicsandoftheobservations.These are not usually knownandhavetobeinferred.Manyapproacheshavebeenproposed to tackle this problem, including fully Bayesian, likelihood maximization and innovation-based techniques. This work focuses on maximization of the likelihood function via the expectation–maximization (EM) algorithm to infer the model error covariance combined with ensemble Kalman filters and particle filters to estimate the state. The classical application of the EM algorithm in a data assimilation context involves filtering and smoothing a fixed batch of observations in order to complete a single iteration. This is an inconvenience whenusing sequential filtering in high-dimensional applications. Motivated by this, an adaptation of the algorithm that can process observations and update the parameters on the fly, with some underlying simplifications, is presented. The proposed technique was evaluated and achieved good performance in experiments with the Lorenz-63 and Lorenz-96 dynamical systems designed to represent some common scenarios in data assimilation such as nonlinearity, chaoticity and model mis-specification.
description Fil: Cocucci, Tadeo Javier. Universidad Nacional del Nordeste. Facultad de Ciencias Exactas y Naturales y Agrimensura; Argentina.
publishDate 2021
dc.date.none.fl_str_mv 2021
dc.type.none.fl_str_mv info:eu-repo/semantics/article
info:eu-repo/semantics/publishedVersion
http://purl.org/coar/resource_type/c_6501
info:ar-repo/semantics/articulo
format article
status_str publishedVersion
dc.identifier.none.fl_str_mv Cocucci, Tadeo Javier, et al., 2021. Model error covariance estimation in particle and ensemble Kalman filters using an online expectation-maximization algorithm. Quarterly Journal of the Royal Meteorological Society. Reading: Royal Meteorological Society, vol. 147, no. 734, p. 526-543. E-ISSN 1477-870X. DOI https://doi.org/10.1002/qj.3931
0035-9009
http://repositorio.unne.edu.ar/handle/123456789/60144
identifier_str_mv Cocucci, Tadeo Javier, et al., 2021. Model error covariance estimation in particle and ensemble Kalman filters using an online expectation-maximization algorithm. Quarterly Journal of the Royal Meteorological Society. Reading: Royal Meteorological Society, vol. 147, no. 734, p. 526-543. E-ISSN 1477-870X. DOI https://doi.org/10.1002/qj.3931
0035-9009
url http://repositorio.unne.edu.ar/handle/123456789/60144
dc.language.none.fl_str_mv eng
language eng
dc.relation.none.fl_str_mv https://doi.org/10.1002/qj.3931
dc.rights.none.fl_str_mv info:eu-repo/semantics/openAccess
http://creativecommons.org/licenses/by-nc-nd/2.5/ar/
Atribución-NoComercial-SinDerivadas 2.5 Argentina
eu_rights_str_mv openAccess
rights_invalid_str_mv http://creativecommons.org/licenses/by-nc-nd/2.5/ar/
Atribución-NoComercial-SinDerivadas 2.5 Argentina
dc.format.none.fl_str_mv application/pdf
p. 526-543
application/pdf
dc.publisher.none.fl_str_mv Royal Meteorological Society
publisher.none.fl_str_mv Royal Meteorological Society
dc.source.none.fl_str_mv Quarterly Journal of the Royal Meteorological Society, 2021, vol. 147, no. 734, p. 526-543.
reponame:Repositorio Institucional de la Universidad Nacional del Nordeste (UNNE)
instname:Universidad Nacional del Nordeste
reponame_str Repositorio Institucional de la Universidad Nacional del Nordeste (UNNE)
collection Repositorio Institucional de la Universidad Nacional del Nordeste (UNNE)
instname_str Universidad Nacional del Nordeste
repository.name.fl_str_mv Repositorio Institucional de la Universidad Nacional del Nordeste (UNNE) - Universidad Nacional del Nordeste
repository.mail.fl_str_mv ososa@bib.unne.edu.ar;sergio.alegria@unne.edu.ar
_version_ 1877230452972978176
score 13.265058

Publicaciones similares