Ionospheric TEC Forecasting by Exploiting the Ionosonde with the LSTM-Piecewise Model in China

Muhammad Usama, Kai Guo, Yiming Wang, Zhipeng Wang

Peer Reviewed

Abstract: In global navigation satellite system (GNSS), ionospheric delay is one of the significant positioning errors. Ionospheric total electron content (TEC) is one of the varying components of the atmosphere that causes delay. The fluctuations in TEC, arising from diverse environmental, solar, and geomagnetic factors, necessitate proactive measures to predict and monitor these variations. This study demonstrates the correlation analysis, based on which we assessed varying critical parameters of relevant phenomena, causing variation in TEC, in the selected regions of China. This analysis serves to quantify the interdependence between Vertical TEC (VTEC) and the diverse features from the dataset. This analysis also serves as a comparative investigation of the correlation of TEC with the ionosonde parameters along with solar and geomagnetic parameters. The results of this analysis are then fed to a piecewise long short-term memory (LSTMp) model to conduct the TEC forecasting in this study. The developed model underwent training using data drawn from a comprehensive dataset spanning one year, 2015, from various latitudes in China (17.5° to 40°N). The model and prediction performance are benchmarked using mean square error (MSE) and root mean square error (RMSE), respectively. The best-trained models for different stations are evaluated based on Mean Squared Error (MSE). The MSE values for all stations are minor. All stations have predicted Root Mean Squared Error (RMSE) less than 0.1 TEC Unit. The TEC is predicted from 00:00:00 UTC 1 Jan 2016, to 23:00:00 UTC 21 Jan 2016, during the time span of a given period there are at least 4 significant peaks in TEC. These peaks are due to corresponding geomagnetic or solar variance. The predicted TEC values are compared with the TEC values of the empirical model of international reference for ionosphere (IRI). It was found that the results are accurately predicted in comparison with Reference TEC and TEC from the IRI model.
Published in: Proceedings of the 2024 International Technical Meeting of The Institute of Navigation
January 23 - 25, 2024
Hyatt Regency Long Beach
Long Beach, California
Pages: 146 - 160
Cite this article: Usama, Muhammad, Guo, Kai, Wang, Yiming, Wang, Zhipeng, "Ionospheric TEC Forecasting by Exploiting the Ionosonde with the LSTM-Piecewise Model in China," Proceedings of the 2024 International Technical Meeting of The Institute of Navigation, Long Beach, California, January 2024, pp. 146-160. https://doi.org/10.33012/2024.19528
Full Paper: ION Members/Non-Members: 1 Download Credit
Sign In