Total electron content

Quantum mechanics

Critical frequency

Universal Time

Computer science

Geology

Ionosphere

Physics

Electron density

Astronomy

International Reference Ionosphere

Artificial intelligence

Geophysics

Electron

Artificial neural network

Ionosonde

Ionogram

IEEE Geoscience and Remote Sensing Letters

Equatorial electrojet (EEJ) and the subsequent development of equatorial ionization anomaly (EIA) are responsible for the highly complex and nonlinear variability nature of the ionosphere. Prediction of ionospheric parameters like Ionospheric F2 layer Critical frequency (foF2) and peak height (hmF2) feature at low latitude regions is of significant interest in understanding the ionospheric weather effects on communication and navigation systems. The role of artificial intelligence-based machine learning algorithms is successful in the prediction of ionospheric variability. In this letter, a deep learning model based on Bidirectional long short-term memory (Bi-LSTM) technique is implemented for predicting foF2 and hmF2 parameters. The Bi-LSTM method was trained and tested on one-year (2015) ionospheric foF2 and hmF2 data from Canadian Advanced Digital Ionosonde (CADI) located at Hyderabad, India (17.47 °N, 78.57 °E). Bi-LSTM model captures time sequence processing features using past and present foF2 and hmF2 data samples. It is evident from the results that the Bi-LSTM model performs better than long short-term memory (LSTM), neural networks (NNs), and International Reference Ionosphere (IRI) 2016 models in predicting foF2 and hmF2 values. The performance of the Bi-LSTM model tested and found to better predict ionospheric foF2 and hmF2 features for two significant geomagnetic storms occurred in the year 2015 (March and June).

A Bidirectional Long Short-Term Memory-Based Ionospheric foF2 and hmF2 Models for a Single Station in the Low Latitude Region