A Forecast Model of Ionospheric foF2 for HF Radio Wave Propagation Based on Echo State Network

Abstract

The critical frequency of the ionospheric F2 layer (foF2), one of the essential channel parameters for high-frequency (HF) global communications via sky waves, shows significant variability in the equatorial region. This is mainly due to its nonlinear and nonstationary, which poses a considerable challenge for making efficient and accurate short-term forecasts of foF2 in low-latitude areas. Neural network algorithms have shown promising results in forecasting foF2 time series. This communication first introduces a neural network model based on the echo state network (ESN) to forecast HF propagation ionospheric parameters foF2 time series. By modeling and analyzing the foF2 data collected in 2014 and 2017 at the Sanya (18.34°N, 109.42°E) station in China, the ESN model effectively captures the characteristic trends of the foF2 time series. The results indicate that the forecast accuracy of the ESN model surpasses that of the international reference ionosphere (IRI), long short-term memory (LSTM), and bidirectional LSTM (Bi-LSTM) models, with significantly higher computational efficiency than other deep learning models. Moreover, it accurately tracks the foF2 trend during the geomagnetic storm in May 2017.