Machine Learning-Based Estimation of foF2 and MUF(3000)F2 Using GNSS Ionospheric TEC Observations

This study developed machine learning models using different algorithms, including support vector machine (SVM), random forest (RF), and backpropagation neural network (BPNN), to estimate the critical frequency of the F2 layer (foF2) and the maximum usable frequency of the F2 layer for a 3000 km cir...

Full description

Saved in:
Bibliographic Details
Main Authors: Yuhang Zhang, Ming Ou, Liang Chen, Yi Hao, Qinglin Zhu, Xiang Dong, Weimin Zhen
Format: Article
Language:English
Published: MDPI AG 2025-05-01
Series:Remote Sensing
Subjects:
Online Access:https://www.mdpi.com/2072-4292/17/10/1764
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849327053470433280
author Yuhang Zhang
Ming Ou
Liang Chen
Yi Hao
Qinglin Zhu
Xiang Dong
Weimin Zhen
author_facet Yuhang Zhang
Ming Ou
Liang Chen
Yi Hao
Qinglin Zhu
Xiang Dong
Weimin Zhen
author_sort Yuhang Zhang
collection DOAJ
description This study developed machine learning models using different algorithms, including support vector machine (SVM), random forest (RF), and backpropagation neural network (BPNN), to estimate the critical frequency of the F2 layer (foF2) and the maximum usable frequency of the F2 layer for a 3000 km circuit (MUF(3000)F2) based on the total electron content (TEC) observed by global navigation satellite system (GNSS) receivers. The ionospheric dataset used comprised TEC, foF2, and MUF(3000)F2 measurements from 11 stations in China during a solar activity period (2008–2020). The results indicate that all three machine learning models performed better than the IRI-2020 model, with varying levels of accuracy. For foF2 (MUF(3000)F2) estimation, the root mean square error (RMSE) values at Kunming and Xi’an stations were reduced by approximately 38% (26%) and 18% (11%), respectively, compared to IRI-2020. During geomagnetic disturbances, all three models were able to reproduce the variations in both foF2 and MUF(3000)F2 parameters. Nevertheless, the RF model showed significantly better performance in foF2 estimation compared to the SVM and BPNN models.
format Article
id doaj-art-fe97d118fbf5407bb06110c3fed8cda4
institution Kabale University
issn 2072-4292
language English
publishDate 2025-05-01
publisher MDPI AG
record_format Article
series Remote Sensing
spelling doaj-art-fe97d118fbf5407bb06110c3fed8cda42025-08-20T03:47:58ZengMDPI AGRemote Sensing2072-42922025-05-011710176410.3390/rs17101764Machine Learning-Based Estimation of foF2 and MUF(3000)F2 Using GNSS Ionospheric TEC ObservationsYuhang Zhang0Ming Ou1Liang Chen2Yi Hao3Qinglin Zhu4Xiang Dong5Weimin Zhen6China Research Institute of Radiowave Propagation, Qingdao 266100, ChinaChina Research Institute of Radiowave Propagation, Qingdao 266100, ChinaChina Research Institute of Radiowave Propagation, Qingdao 266100, ChinaChina Research Institute of Radiowave Propagation, Qingdao 266100, ChinaChina Research Institute of Radiowave Propagation, Qingdao 266100, ChinaChina Research Institute of Radiowave Propagation, Qingdao 266100, ChinaChina Research Institute of Radiowave Propagation, Qingdao 266100, ChinaThis study developed machine learning models using different algorithms, including support vector machine (SVM), random forest (RF), and backpropagation neural network (BPNN), to estimate the critical frequency of the F2 layer (foF2) and the maximum usable frequency of the F2 layer for a 3000 km circuit (MUF(3000)F2) based on the total electron content (TEC) observed by global navigation satellite system (GNSS) receivers. The ionospheric dataset used comprised TEC, foF2, and MUF(3000)F2 measurements from 11 stations in China during a solar activity period (2008–2020). The results indicate that all three machine learning models performed better than the IRI-2020 model, with varying levels of accuracy. For foF2 (MUF(3000)F2) estimation, the root mean square error (RMSE) values at Kunming and Xi’an stations were reduced by approximately 38% (26%) and 18% (11%), respectively, compared to IRI-2020. During geomagnetic disturbances, all three models were able to reproduce the variations in both foF2 and MUF(3000)F2 parameters. Nevertheless, the RF model showed significantly better performance in foF2 estimation compared to the SVM and BPNN models.https://www.mdpi.com/2072-4292/17/10/1764machine learningSVMRFBPNNTECfoF2
spellingShingle Yuhang Zhang
Ming Ou
Liang Chen
Yi Hao
Qinglin Zhu
Xiang Dong
Weimin Zhen
Machine Learning-Based Estimation of foF2 and MUF(3000)F2 Using GNSS Ionospheric TEC Observations
Remote Sensing
machine learning
SVM
RF
BPNN
TEC
foF2
title Machine Learning-Based Estimation of foF2 and MUF(3000)F2 Using GNSS Ionospheric TEC Observations
title_full Machine Learning-Based Estimation of foF2 and MUF(3000)F2 Using GNSS Ionospheric TEC Observations
title_fullStr Machine Learning-Based Estimation of foF2 and MUF(3000)F2 Using GNSS Ionospheric TEC Observations
title_full_unstemmed Machine Learning-Based Estimation of foF2 and MUF(3000)F2 Using GNSS Ionospheric TEC Observations
title_short Machine Learning-Based Estimation of foF2 and MUF(3000)F2 Using GNSS Ionospheric TEC Observations
title_sort machine learning based estimation of fof2 and muf 3000 f2 using gnss ionospheric tec observations
topic machine learning
SVM
RF
BPNN
TEC
foF2
url https://www.mdpi.com/2072-4292/17/10/1764
work_keys_str_mv AT yuhangzhang machinelearningbasedestimationoffof2andmuf3000f2usinggnssionospherictecobservations
AT mingou machinelearningbasedestimationoffof2andmuf3000f2usinggnssionospherictecobservations
AT liangchen machinelearningbasedestimationoffof2andmuf3000f2usinggnssionospherictecobservations
AT yihao machinelearningbasedestimationoffof2andmuf3000f2usinggnssionospherictecobservations
AT qinglinzhu machinelearningbasedestimationoffof2andmuf3000f2usinggnssionospherictecobservations
AT xiangdong machinelearningbasedestimationoffof2andmuf3000f2usinggnssionospherictecobservations
AT weiminzhen machinelearningbasedestimationoffof2andmuf3000f2usinggnssionospherictecobservations