Bayesian Inference analysis of Total Electron Content (TEC) across latitudinal regions

Autores
Argüelles, Noelia Beatriz; Villegas, Atuel; Namour, Jorge Habib; Asamoah, Eric; Cesaroni, Claudio; Molina, Maria Graciela
Año de publicación
2025
Idioma
inglés
Tipo de recurso
documento de conferencia
Estado
versión publicada
Descripción
To accurately characterize the ionosphere, we investigate its structure and variability (time and spatial) by analyzing Total Electron Content (TEC) derived from Global Navigation Satellite Systems (GNSS). We estimate the probability density function (PDF) of TEC using Bayesian techniques. The Bayesian approach enhances this analysis by not only identifying the most probable values but also quantifying the uncertainties associated with these estimates. Unlike the frequentist approach, which yields single-point estimates, Bayesian methods generate full posterior distributions for model parameters, offering deeper insights into the reliability of these estimates at each station. This probabilistic framework allows for a more comprehensive understanding of TEC across different latitudes and geographical regions (e.g. oceanic regions, at different meridional regions, etc.), improving the robustness of ionospheric modeling and forecasting using machine learning.We analyzed a large dataset of TEC measurements from 2005 to 2017, spanning nearly a full solar cycle, with a temporal resolution of two hours from GIM maps. The data included stations at low, mid, and high latitudes.Various probability distributions, such as gamma and lognormal, were evaluated to model TEC. Bayesian inference, implemented using the PyMC3 package, was employed to estimate the parameters of these distributions. Robust parameter estimation and model validation were ensured through posterior predictive checks and diagnostic tools such as trace plots.Our results indicate that the posterior predictive distributions closely align with observed data across all stations, demonstrating the effectiveness of Bayesian models in capturing TEC variability across different latitudes and regions. We found that TEC distributions vary with latitude, with low and mid latitude stations best described by a lognormal distribution, while high-latitude regions require a mixture model, obtaining a bimodal distribution.The performance metrics used to evaluate the models confirm that the Markov chain Monte Carlo sampling effectively estimated the parameters, with well-sampled and stable posterior distributions. This study highlightsthe potential of Bayesian methods in ionospheric modeling, where their ability to quantify uncertainties enhances both interpretability and confidence in model predictions. Furthermore, these results can help improve machine learning models by providing a better understanding of TEC behavior and a way to account for uncertainties in future predictions, including the integration of Bayesian priors to enhance model reliability and data efficiency.
Fil: Argüelles, Noelia Beatriz. Consejo Nacional de Investigaciones Científicas y Técnicas. Centro Científico Tecnológico Conicet - Tucumán; Argentina. Universidad Nacional de Tucumán. Facultad de Ciencias Exactas y Tecnología; Argentina
Fil: Villegas, Atuel. Universidad Nacional de Tucumán. Facultad de Ciencias Exactas y Tecnología; Argentina
Fil: Namour, Jorge Habib. Universidad Nacional de Tucumán. Facultad de Ciencias Exactas y Tecnología; Argentina
Fil: Asamoah, Eric. Istituto Nazionale di Geofisica e Vulcanologia; Italia
Fil: Cesaroni, Claudio. Istituto Nazionale di Geofisica e Vulcanologia; Italia
Fil: Molina, Maria Graciela. Consejo Nacional de Investigaciones Científicas y Técnicas. Centro Científico Tecnológico Conicet - Tucumán; Argentina. Universidad Nacional de Tucumán. Facultad de Ciencias Exactas y Tecnología; Argentina. Istituto Nazionale di Geofisica e Vulcanologia; Italia
Union Radio-Scientifique Internationale Asia-Pacific Radio Science Conference
Sidney
Australia
Union Radio-Scientifique Internationale
Materia
Bayesian Statistics
Total Electron Content
Space Weather
Nivel de accesibilidad
acceso abierto
Condiciones de uso
https://creativecommons.org/licenses/by-nc-sa/2.5/ar/
Repositorio
CONICET Digital (CONICET)
Institución
Consejo Nacional de Investigaciones Científicas y Técnicas
OAI Identificador
oai:ri.conicet.gov.ar:11336/284835
Acceder
id CONICETDig_6903e1faa476e51f3fc2c450928504ba
oai_identifier_str oai:ri.conicet.gov.ar:11336/284835
network_acronym_str CONICETDig
repository_id_str 3498
network_name_str CONICET Digital (CONICET)
spelling Bayesian Inference analysis of Total Electron Content (TEC) across latitudinal regionsArgüelles, Noelia BeatrizVillegas, AtuelNamour, Jorge HabibAsamoah, EricCesaroni, ClaudioMolina, Maria GracielaBayesian StatisticsTotal Electron ContentSpace Weatherhttps://purl.org/becyt/ford/1.5https://purl.org/becyt/ford/1To accurately characterize the ionosphere, we investigate its structure and variability (time and spatial) by analyzing Total Electron Content (TEC) derived from Global Navigation Satellite Systems (GNSS). We estimate the probability density function (PDF) of TEC using Bayesian techniques. The Bayesian approach enhances this analysis by not only identifying the most probable values but also quantifying the uncertainties associated with these estimates. Unlike the frequentist approach, which yields single-point estimates, Bayesian methods generate full posterior distributions for model parameters, offering deeper insights into the reliability of these estimates at each station. This probabilistic framework allows for a more comprehensive understanding of TEC across different latitudes and geographical regions (e.g. oceanic regions, at different meridional regions, etc.), improving the robustness of ionospheric modeling and forecasting using machine learning.We analyzed a large dataset of TEC measurements from 2005 to 2017, spanning nearly a full solar cycle, with a temporal resolution of two hours from GIM maps. The data included stations at low, mid, and high latitudes.Various probability distributions, such as gamma and lognormal, were evaluated to model TEC. Bayesian inference, implemented using the PyMC3 package, was employed to estimate the parameters of these distributions. Robust parameter estimation and model validation were ensured through posterior predictive checks and diagnostic tools such as trace plots.Our results indicate that the posterior predictive distributions closely align with observed data across all stations, demonstrating the effectiveness of Bayesian models in capturing TEC variability across different latitudes and regions. We found that TEC distributions vary with latitude, with low and mid latitude stations best described by a lognormal distribution, while high-latitude regions require a mixture model, obtaining a bimodal distribution.The performance metrics used to evaluate the models confirm that the Markov chain Monte Carlo sampling effectively estimated the parameters, with well-sampled and stable posterior distributions. This study highlightsthe potential of Bayesian methods in ionospheric modeling, where their ability to quantify uncertainties enhances both interpretability and confidence in model predictions. Furthermore, these results can help improve machine learning models by providing a better understanding of TEC behavior and a way to account for uncertainties in future predictions, including the integration of Bayesian priors to enhance model reliability and data efficiency.Fil: Argüelles, Noelia Beatriz. Consejo Nacional de Investigaciones Científicas y Técnicas. Centro Científico Tecnológico Conicet - Tucumán; Argentina. Universidad Nacional de Tucumán. Facultad de Ciencias Exactas y Tecnología; ArgentinaFil: Villegas, Atuel. Universidad Nacional de Tucumán. Facultad de Ciencias Exactas y Tecnología; ArgentinaFil: Namour, Jorge Habib. Universidad Nacional de Tucumán. Facultad de Ciencias Exactas y Tecnología; ArgentinaFil: Asamoah, Eric. Istituto Nazionale di Geofisica e Vulcanologia; ItaliaFil: Cesaroni, Claudio. Istituto Nazionale di Geofisica e Vulcanologia; ItaliaFil: Molina, Maria Graciela. Consejo Nacional de Investigaciones Científicas y Técnicas. Centro Científico Tecnológico Conicet - Tucumán; Argentina. Universidad Nacional de Tucumán. Facultad de Ciencias Exactas y Tecnología; Argentina. Istituto Nazionale di Geofisica e Vulcanologia; ItaliaUnion Radio-Scientifique Internationale Asia-Pacific Radio Science ConferenceSidneyAustraliaUnion Radio-Scientifique InternationaleUnion Radio-Scientifique Internationale2025info:eu-repo/semantics/publishedVersioninfo:eu-repo/semantics/conferenceObjectConferenciaBookhttp://purl.org/coar/resource_type/c_5794info:ar-repo/semantics/documentoDeConferenciaapplication/pdfapplication/pdfhttp://hdl.handle.net/11336/284835Bayesian Inference analysis of Total Electron Content (TEC) across latitudinal regions; Union Radio-Scientifique Internationale Asia-Pacific Radio Science Conference; Sidney; Australia; 2025; 1-1CONICET DigitalCONICETenginfo:eu-repo/semantics/altIdentifier/url/https://www.ursi.org/proceedings/procAP25/papers/0917.pdfinfo:eu-repo/semantics/altIdentifier/doi/10.46620/URSIAPRASC25/HWNE6003Internacionalinfo:eu-repo/semantics/openAccesshttps://creativecommons.org/licenses/by-nc-sa/2.5/ar/reponame:CONICET Digital (CONICET)instname:Consejo Nacional de Investigaciones Científicas y Técnicas2026-08-25T14:33:54Zoai:ri.conicet.gov.ar:11336/284835instacron:CONICETInstitucionalhttp://ri.conicet.gov.ar/Organismo científico-tecnológicoNo correspondehttp://ri.conicet.gov.ar/oai/requestdasensio@conicet.gov.ar; lcarlino@conicet.gov.arArgentinaNo correspondeNo correspondeNo correspondeopendoar:34982026-08-25 14:33:54.97CONICET Digital (CONICET) - Consejo Nacional de Investigaciones Científicas y Técnicasfalse
dc.title.none.fl_str_mv Bayesian Inference analysis of Total Electron Content (TEC) across latitudinal regions
title Bayesian Inference analysis of Total Electron Content (TEC) across latitudinal regions
spellingShingle Bayesian Inference analysis of Total Electron Content (TEC) across latitudinal regions
Argüelles, Noelia Beatriz
Bayesian Statistics
Total Electron Content
Space Weather
title_short Bayesian Inference analysis of Total Electron Content (TEC) across latitudinal regions
title_full Bayesian Inference analysis of Total Electron Content (TEC) across latitudinal regions
title_fullStr Bayesian Inference analysis of Total Electron Content (TEC) across latitudinal regions
title_full_unstemmed Bayesian Inference analysis of Total Electron Content (TEC) across latitudinal regions
title_sort Bayesian Inference analysis of Total Electron Content (TEC) across latitudinal regions
dc.creator.none.fl_str_mv Argüelles, Noelia Beatriz
Villegas, Atuel
Namour, Jorge Habib
Asamoah, Eric
Cesaroni, Claudio
Molina, Maria Graciela
author Argüelles, Noelia Beatriz
author_facet Argüelles, Noelia Beatriz
Villegas, Atuel
Namour, Jorge Habib
Asamoah, Eric
Cesaroni, Claudio
Molina, Maria Graciela
author_role author
author2 Villegas, Atuel
Namour, Jorge Habib
Asamoah, Eric
Cesaroni, Claudio
Molina, Maria Graciela
author2_role author
author
author
author
author
dc.subject.none.fl_str_mv Bayesian Statistics
Total Electron Content
Space Weather
topic Bayesian Statistics
Total Electron Content
Space Weather
purl_subject.fl_str_mv https://purl.org/becyt/ford/1.5
https://purl.org/becyt/ford/1
dc.description.none.fl_txt_mv To accurately characterize the ionosphere, we investigate its structure and variability (time and spatial) by analyzing Total Electron Content (TEC) derived from Global Navigation Satellite Systems (GNSS). We estimate the probability density function (PDF) of TEC using Bayesian techniques. The Bayesian approach enhances this analysis by not only identifying the most probable values but also quantifying the uncertainties associated with these estimates. Unlike the frequentist approach, which yields single-point estimates, Bayesian methods generate full posterior distributions for model parameters, offering deeper insights into the reliability of these estimates at each station. This probabilistic framework allows for a more comprehensive understanding of TEC across different latitudes and geographical regions (e.g. oceanic regions, at different meridional regions, etc.), improving the robustness of ionospheric modeling and forecasting using machine learning.We analyzed a large dataset of TEC measurements from 2005 to 2017, spanning nearly a full solar cycle, with a temporal resolution of two hours from GIM maps. The data included stations at low, mid, and high latitudes.Various probability distributions, such as gamma and lognormal, were evaluated to model TEC. Bayesian inference, implemented using the PyMC3 package, was employed to estimate the parameters of these distributions. Robust parameter estimation and model validation were ensured through posterior predictive checks and diagnostic tools such as trace plots.Our results indicate that the posterior predictive distributions closely align with observed data across all stations, demonstrating the effectiveness of Bayesian models in capturing TEC variability across different latitudes and regions. We found that TEC distributions vary with latitude, with low and mid latitude stations best described by a lognormal distribution, while high-latitude regions require a mixture model, obtaining a bimodal distribution.The performance metrics used to evaluate the models confirm that the Markov chain Monte Carlo sampling effectively estimated the parameters, with well-sampled and stable posterior distributions. This study highlightsthe potential of Bayesian methods in ionospheric modeling, where their ability to quantify uncertainties enhances both interpretability and confidence in model predictions. Furthermore, these results can help improve machine learning models by providing a better understanding of TEC behavior and a way to account for uncertainties in future predictions, including the integration of Bayesian priors to enhance model reliability and data efficiency.
Fil: Argüelles, Noelia Beatriz. Consejo Nacional de Investigaciones Científicas y Técnicas. Centro Científico Tecnológico Conicet - Tucumán; Argentina. Universidad Nacional de Tucumán. Facultad de Ciencias Exactas y Tecnología; Argentina
Fil: Villegas, Atuel. Universidad Nacional de Tucumán. Facultad de Ciencias Exactas y Tecnología; Argentina
Fil: Namour, Jorge Habib. Universidad Nacional de Tucumán. Facultad de Ciencias Exactas y Tecnología; Argentina
Fil: Asamoah, Eric. Istituto Nazionale di Geofisica e Vulcanologia; Italia
Fil: Cesaroni, Claudio. Istituto Nazionale di Geofisica e Vulcanologia; Italia
Fil: Molina, Maria Graciela. Consejo Nacional de Investigaciones Científicas y Técnicas. Centro Científico Tecnológico Conicet - Tucumán; Argentina. Universidad Nacional de Tucumán. Facultad de Ciencias Exactas y Tecnología; Argentina. Istituto Nazionale di Geofisica e Vulcanologia; Italia
Union Radio-Scientifique Internationale Asia-Pacific Radio Science Conference
Sidney
Australia
Union Radio-Scientifique Internationale
description To accurately characterize the ionosphere, we investigate its structure and variability (time and spatial) by analyzing Total Electron Content (TEC) derived from Global Navigation Satellite Systems (GNSS). We estimate the probability density function (PDF) of TEC using Bayesian techniques. The Bayesian approach enhances this analysis by not only identifying the most probable values but also quantifying the uncertainties associated with these estimates. Unlike the frequentist approach, which yields single-point estimates, Bayesian methods generate full posterior distributions for model parameters, offering deeper insights into the reliability of these estimates at each station. This probabilistic framework allows for a more comprehensive understanding of TEC across different latitudes and geographical regions (e.g. oceanic regions, at different meridional regions, etc.), improving the robustness of ionospheric modeling and forecasting using machine learning.We analyzed a large dataset of TEC measurements from 2005 to 2017, spanning nearly a full solar cycle, with a temporal resolution of two hours from GIM maps. The data included stations at low, mid, and high latitudes.Various probability distributions, such as gamma and lognormal, were evaluated to model TEC. Bayesian inference, implemented using the PyMC3 package, was employed to estimate the parameters of these distributions. Robust parameter estimation and model validation were ensured through posterior predictive checks and diagnostic tools such as trace plots.Our results indicate that the posterior predictive distributions closely align with observed data across all stations, demonstrating the effectiveness of Bayesian models in capturing TEC variability across different latitudes and regions. We found that TEC distributions vary with latitude, with low and mid latitude stations best described by a lognormal distribution, while high-latitude regions require a mixture model, obtaining a bimodal distribution.The performance metrics used to evaluate the models confirm that the Markov chain Monte Carlo sampling effectively estimated the parameters, with well-sampled and stable posterior distributions. This study highlightsthe potential of Bayesian methods in ionospheric modeling, where their ability to quantify uncertainties enhances both interpretability and confidence in model predictions. Furthermore, these results can help improve machine learning models by providing a better understanding of TEC behavior and a way to account for uncertainties in future predictions, including the integration of Bayesian priors to enhance model reliability and data efficiency.
publishDate 2025
dc.date.none.fl_str_mv 2025
dc.type.none.fl_str_mv info:eu-repo/semantics/publishedVersion
info:eu-repo/semantics/conferenceObject
Conferencia
Book
http://purl.org/coar/resource_type/c_5794
info:ar-repo/semantics/documentoDeConferencia
status_str publishedVersion
format conferenceObject
dc.identifier.none.fl_str_mv http://hdl.handle.net/11336/284835
Bayesian Inference analysis of Total Electron Content (TEC) across latitudinal regions; Union Radio-Scientifique Internationale Asia-Pacific Radio Science Conference; Sidney; Australia; 2025; 1-1
CONICET Digital
CONICET
url http://hdl.handle.net/11336/284835
identifier_str_mv Bayesian Inference analysis of Total Electron Content (TEC) across latitudinal regions; Union Radio-Scientifique Internationale Asia-Pacific Radio Science Conference; Sidney; Australia; 2025; 1-1
CONICET Digital
CONICET
dc.language.none.fl_str_mv eng
language eng
dc.relation.none.fl_str_mv info:eu-repo/semantics/altIdentifier/url/https://www.ursi.org/proceedings/procAP25/papers/0917.pdf
info:eu-repo/semantics/altIdentifier/doi/10.46620/URSIAPRASC25/HWNE6003
dc.rights.none.fl_str_mv info:eu-repo/semantics/openAccess
https://creativecommons.org/licenses/by-nc-sa/2.5/ar/
eu_rights_str_mv openAccess
rights_invalid_str_mv https://creativecommons.org/licenses/by-nc-sa/2.5/ar/
dc.format.none.fl_str_mv application/pdf
application/pdf
dc.coverage.none.fl_str_mv Internacional
dc.publisher.none.fl_str_mv Union Radio-Scientifique Internationale
publisher.none.fl_str_mv Union Radio-Scientifique Internationale
dc.source.none.fl_str_mv reponame:CONICET Digital (CONICET)
instname:Consejo Nacional de Investigaciones Científicas y Técnicas
reponame_str CONICET Digital (CONICET)
collection CONICET Digital (CONICET)
instname_str Consejo Nacional de Investigaciones Científicas y Técnicas
repository.name.fl_str_mv CONICET Digital (CONICET) - Consejo Nacional de Investigaciones Científicas y Técnicas
repository.mail.fl_str_mv dasensio@conicet.gov.ar; lcarlino@conicet.gov.ar
_version_ 1874774185026256896
score 13.265058

Publicaciones similares