Study of TEC fluctuation via stochastic models and Bayesian inversion

A. Bires, L. Roininen, B. Damtie, M. Nigussie, Heikki Antero Vanhamaki

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

We propose stochastic processes to be used to model the total electron content (TEC) observation. Based on this, we model the rate of change of TEC (ROT) variation during ionospheric quiet conditions with stationary processes. During ionospheric disturbed conditions, for example, when irregularity in ionospheric electron density distribution occurs, stationarity assumption over long time periods is no longer valid. In these cases, we make the parameter estimation for short time scales, during which we can assume stationarity. We show the relationship between the new method and commonly used TEC characterization parameters ROT and the ROT Index (ROTI). We construct our parametric model within the framework of Bayesian statistical inverse problems and hence give the solution as an a posteriori probability distribution. Bayesian framework allows us to model measurement errors systematically. Similarly, we mitigate variation of TEC due to factors which are not of ionospheric origin, like due to the motion of satellites relative to the receiver, by incorporating a priori knowledge in the Bayesian model. In practical computations, we draw the so-called maximum a posteriori estimates, which are our ROT and ROTI estimates, from the posterior distribution. Because the algorithm allows to estimate ROTI at each observation time, the estimator does not depend on the period of time for ROTI computation. We verify the method by analyzing TEC data recorded by GPS receiver located in Ethiopia (11.6°N, 37.4°E). The results indicate that the TEC fluctuations caused by the ionospheric irregularity can be effectively detected and quantified from the estimated ROT and ROTI values.

Original languageEnglish
Pages (from-to)1772-1782
Number of pages11
JournalRadio Science
Volume51
Issue number11
DOIs
Publication statusPublished - Nov 1 2016

Fingerprint

Stochastic models
inversions
ionospherics
Electrons
electrons
irregularities
estimates
receivers
Ethiopia
ionospheric electron density
Electronic density of states
stochastic processes
stochasticity
inverse problem
Measurement errors
Random processes
Inverse problems
estimators
Parameter estimation
Probability distributions

All Science Journal Classification (ASJC) codes

  • Condensed Matter Physics
  • Electrical and Electronic Engineering
  • Earth and Planetary Sciences(all)

Cite this

Bires, A., Roininen, L., Damtie, B., Nigussie, M., & Vanhamaki, H. A. (2016). Study of TEC fluctuation via stochastic models and Bayesian inversion. Radio Science, 51(11), 1772-1782. https://doi.org/10.1002/2016RS005959

Study of TEC fluctuation via stochastic models and Bayesian inversion. / Bires, A.; Roininen, L.; Damtie, B.; Nigussie, M.; Vanhamaki, Heikki Antero.

In: Radio Science, Vol. 51, No. 11, 01.11.2016, p. 1772-1782.

Research output: Contribution to journalArticle

Bires, A, Roininen, L, Damtie, B, Nigussie, M & Vanhamaki, HA 2016, 'Study of TEC fluctuation via stochastic models and Bayesian inversion', Radio Science, vol. 51, no. 11, pp. 1772-1782. https://doi.org/10.1002/2016RS005959
Bires A, Roininen L, Damtie B, Nigussie M, Vanhamaki HA. Study of TEC fluctuation via stochastic models and Bayesian inversion. Radio Science. 2016 Nov 1;51(11):1772-1782. https://doi.org/10.1002/2016RS005959
Bires, A. ; Roininen, L. ; Damtie, B. ; Nigussie, M. ; Vanhamaki, Heikki Antero. / Study of TEC fluctuation via stochastic models and Bayesian inversion. In: Radio Science. 2016 ; Vol. 51, No. 11. pp. 1772-1782.
@article{ae36ec6b551e4b3aa3187f4f2c66d7cc,
title = "Study of TEC fluctuation via stochastic models and Bayesian inversion",
abstract = "We propose stochastic processes to be used to model the total electron content (TEC) observation. Based on this, we model the rate of change of TEC (ROT) variation during ionospheric quiet conditions with stationary processes. During ionospheric disturbed conditions, for example, when irregularity in ionospheric electron density distribution occurs, stationarity assumption over long time periods is no longer valid. In these cases, we make the parameter estimation for short time scales, during which we can assume stationarity. We show the relationship between the new method and commonly used TEC characterization parameters ROT and the ROT Index (ROTI). We construct our parametric model within the framework of Bayesian statistical inverse problems and hence give the solution as an a posteriori probability distribution. Bayesian framework allows us to model measurement errors systematically. Similarly, we mitigate variation of TEC due to factors which are not of ionospheric origin, like due to the motion of satellites relative to the receiver, by incorporating a priori knowledge in the Bayesian model. In practical computations, we draw the so-called maximum a posteriori estimates, which are our ROT and ROTI estimates, from the posterior distribution. Because the algorithm allows to estimate ROTI at each observation time, the estimator does not depend on the period of time for ROTI computation. We verify the method by analyzing TEC data recorded by GPS receiver located in Ethiopia (11.6°N, 37.4°E). The results indicate that the TEC fluctuations caused by the ionospheric irregularity can be effectively detected and quantified from the estimated ROT and ROTI values.",
author = "A. Bires and L. Roininen and B. Damtie and M. Nigussie and Vanhamaki, {Heikki Antero}",
year = "2016",
month = "11",
day = "1",
doi = "10.1002/2016RS005959",
language = "English",
volume = "51",
pages = "1772--1782",
journal = "Radio Science",
issn = "0048-6604",
publisher = "American Geophysical Union",
number = "11",

}

TY - JOUR

T1 - Study of TEC fluctuation via stochastic models and Bayesian inversion

AU - Bires, A.

AU - Roininen, L.

AU - Damtie, B.

AU - Nigussie, M.

AU - Vanhamaki, Heikki Antero

PY - 2016/11/1

Y1 - 2016/11/1

N2 - We propose stochastic processes to be used to model the total electron content (TEC) observation. Based on this, we model the rate of change of TEC (ROT) variation during ionospheric quiet conditions with stationary processes. During ionospheric disturbed conditions, for example, when irregularity in ionospheric electron density distribution occurs, stationarity assumption over long time periods is no longer valid. In these cases, we make the parameter estimation for short time scales, during which we can assume stationarity. We show the relationship between the new method and commonly used TEC characterization parameters ROT and the ROT Index (ROTI). We construct our parametric model within the framework of Bayesian statistical inverse problems and hence give the solution as an a posteriori probability distribution. Bayesian framework allows us to model measurement errors systematically. Similarly, we mitigate variation of TEC due to factors which are not of ionospheric origin, like due to the motion of satellites relative to the receiver, by incorporating a priori knowledge in the Bayesian model. In practical computations, we draw the so-called maximum a posteriori estimates, which are our ROT and ROTI estimates, from the posterior distribution. Because the algorithm allows to estimate ROTI at each observation time, the estimator does not depend on the period of time for ROTI computation. We verify the method by analyzing TEC data recorded by GPS receiver located in Ethiopia (11.6°N, 37.4°E). The results indicate that the TEC fluctuations caused by the ionospheric irregularity can be effectively detected and quantified from the estimated ROT and ROTI values.

AB - We propose stochastic processes to be used to model the total electron content (TEC) observation. Based on this, we model the rate of change of TEC (ROT) variation during ionospheric quiet conditions with stationary processes. During ionospheric disturbed conditions, for example, when irregularity in ionospheric electron density distribution occurs, stationarity assumption over long time periods is no longer valid. In these cases, we make the parameter estimation for short time scales, during which we can assume stationarity. We show the relationship between the new method and commonly used TEC characterization parameters ROT and the ROT Index (ROTI). We construct our parametric model within the framework of Bayesian statistical inverse problems and hence give the solution as an a posteriori probability distribution. Bayesian framework allows us to model measurement errors systematically. Similarly, we mitigate variation of TEC due to factors which are not of ionospheric origin, like due to the motion of satellites relative to the receiver, by incorporating a priori knowledge in the Bayesian model. In practical computations, we draw the so-called maximum a posteriori estimates, which are our ROT and ROTI estimates, from the posterior distribution. Because the algorithm allows to estimate ROTI at each observation time, the estimator does not depend on the period of time for ROTI computation. We verify the method by analyzing TEC data recorded by GPS receiver located in Ethiopia (11.6°N, 37.4°E). The results indicate that the TEC fluctuations caused by the ionospheric irregularity can be effectively detected and quantified from the estimated ROT and ROTI values.

UR - http://www.scopus.com/inward/record.url?scp=85000751817&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85000751817&partnerID=8YFLogxK

U2 - 10.1002/2016RS005959

DO - 10.1002/2016RS005959

M3 - Article

AN - SCOPUS:85000751817

VL - 51

SP - 1772

EP - 1782

JO - Radio Science

JF - Radio Science

SN - 0048-6604

IS - 11

ER -