Noise variance decomposition of regional GNSS network with missing observations using structured additive predictor

Abstract This study evaluates noise variance characteristics at configured GNSS stations using a combination of Factor Analysis (FA), stacked Gauss Markov Random Field (GMRF), and Structured Additive Predictor (SAP). FA is employed to decompose the variance in CGNSS time series into station-specific...

Full description

Saved in:
Bibliographic Details
Main Authors: Mohammed Ouassou, Halfdan Pascal Kierulf, Tor-Ole Dahlø
Format: Article
Language:English
Published: Springer 2024-12-01
Series:Discover Applied Sciences
Subjects:
Online Access:https://doi.org/10.1007/s42452-024-06363-6
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1846100985973309440
author Mohammed Ouassou
Halfdan Pascal Kierulf
Tor-Ole Dahlø
author_facet Mohammed Ouassou
Halfdan Pascal Kierulf
Tor-Ole Dahlø
author_sort Mohammed Ouassou
collection DOAJ
description Abstract This study evaluates noise variance characteristics at configured GNSS stations using a combination of Factor Analysis (FA), stacked Gauss Markov Random Field (GMRF), and Structured Additive Predictor (SAP). FA is employed to decompose the variance in CGNSS time series into station-specific and common components. Shorter correlation times are dominated by specific variance, which accounts for over 60% of the total variance, while longer correlation times are more influenced by common variance, contributing up to 40%. Prediction errors using stacked GMRF and SAP were reduced by 15% compared to standard time series models, highlighting the effectiveness of these methods for spatial data imputation and prediction. Additionally, our analysis revealed that the choice of geodetic GNSS processing software (GAMIT vs. BERNESE) has a minimal impact on variance characteristics, with differences below 5%. Some stations demonstrated serial correlation, indicative of colored noise patterns. The analysis of data from fifty reference receivers demonstrates the practical viability and effectiveness of the proposed methodologies, offering a robust approach that enhances both the accuracy and interpretability of GNSS time series analysis.
format Article
id doaj-art-806b543bc1a64b499cbf418812ca73d3
institution Kabale University
issn 3004-9261
language English
publishDate 2024-12-01
publisher Springer
record_format Article
series Discover Applied Sciences
spelling doaj-art-806b543bc1a64b499cbf418812ca73d32024-12-29T12:40:50ZengSpringerDiscover Applied Sciences3004-92612024-12-017113410.1007/s42452-024-06363-6Noise variance decomposition of regional GNSS network with missing observations using structured additive predictorMohammed Ouassou0Halfdan Pascal Kierulf1Tor-Ole Dahlø2Norwegian Mapping Authority, Geodetic InstituteNorwegian Mapping Authority, Geodetic InstituteNorwegian Mapping Authority, Geodetic InstituteAbstract This study evaluates noise variance characteristics at configured GNSS stations using a combination of Factor Analysis (FA), stacked Gauss Markov Random Field (GMRF), and Structured Additive Predictor (SAP). FA is employed to decompose the variance in CGNSS time series into station-specific and common components. Shorter correlation times are dominated by specific variance, which accounts for over 60% of the total variance, while longer correlation times are more influenced by common variance, contributing up to 40%. Prediction errors using stacked GMRF and SAP were reduced by 15% compared to standard time series models, highlighting the effectiveness of these methods for spatial data imputation and prediction. Additionally, our analysis revealed that the choice of geodetic GNSS processing software (GAMIT vs. BERNESE) has a minimal impact on variance characteristics, with differences below 5%. Some stations demonstrated serial correlation, indicative of colored noise patterns. The analysis of data from fifty reference receivers demonstrates the practical viability and effectiveness of the proposed methodologies, offering a robust approach that enhances both the accuracy and interpretability of GNSS time series analysis.https://doi.org/10.1007/s42452-024-06363-6CME (common mode error)EM (expectation-maximization)FA (factor analysis)GMRF (Gauss Markov random field)PC (principal component)PPP (precise point positioning)
spellingShingle Mohammed Ouassou
Halfdan Pascal Kierulf
Tor-Ole Dahlø
Noise variance decomposition of regional GNSS network with missing observations using structured additive predictor
Discover Applied Sciences
CME (common mode error)
EM (expectation-maximization)
FA (factor analysis)
GMRF (Gauss Markov random field)
PC (principal component)
PPP (precise point positioning)
title Noise variance decomposition of regional GNSS network with missing observations using structured additive predictor
title_full Noise variance decomposition of regional GNSS network with missing observations using structured additive predictor
title_fullStr Noise variance decomposition of regional GNSS network with missing observations using structured additive predictor
title_full_unstemmed Noise variance decomposition of regional GNSS network with missing observations using structured additive predictor
title_short Noise variance decomposition of regional GNSS network with missing observations using structured additive predictor
title_sort noise variance decomposition of regional gnss network with missing observations using structured additive predictor
topic CME (common mode error)
EM (expectation-maximization)
FA (factor analysis)
GMRF (Gauss Markov random field)
PC (principal component)
PPP (precise point positioning)
url https://doi.org/10.1007/s42452-024-06363-6
work_keys_str_mv AT mohammedouassou noisevariancedecompositionofregionalgnssnetworkwithmissingobservationsusingstructuredadditivepredictor
AT halfdanpascalkierulf noisevariancedecompositionofregionalgnssnetworkwithmissingobservationsusingstructuredadditivepredictor
AT toroledahlø noisevariancedecompositionofregionalgnssnetworkwithmissingobservationsusingstructuredadditivepredictor