Forecasting of ionospheric vertical total electron content (TEC) using LSTM networks

Ionosphere is an important space environment near the earth. Its disturbance would result in severe propagation effects to radio information system, thus causing bad influences on communication, navigation, radar and so on. The total electron content (TEC) is an important parameter to present the disturbance of ionosphere, so TEC forecast is very meaningful in scientific research field. In this paper, we propose a long short-term memory (LSTM) based model to predict ionospheric vertical TEC of Beijing. The input of our model is a time sequence consisting of the vector of daily TECs and other closely related parameters. The output is TECs of future 24 hours. The result shows the root of mean square (RMS) error of test data can reach 3.50 and RMS error is less than this number during the period of low solar activity. Compared to multilayer perceptron network, LSTM is more promising and reliable to forecast ionospheric TEC.