Research on different time-scale prediction models for the total electron content

Abstract

In the solar-terrestrial space environment, the ionosphere couples tightly with the upper magnetic layer as well as the lower middle atmosphere in various forms. Meanwhile, the ionosphere can affect radio-communication and satellite navigations, so the research on ionosphere prediction model is very important. Now, the accuracy of statistic prediction mode is about 60%, but cannot meet the practical requirements. In order to solve the problem, the prediction model of total electron content (TEC) data is achieved in three major phases: decomposition of the spatiotemporal variability of the TEC data, noise reduction of the encoded space, and time variability and the prediction, by a nonlinear forecasting technique of the time variability. Experiments show that the new prediction model is better than traditional prediction model. The prediction data shows realistic features and a reliable physical distribution, and the relative accuracies of prediction for 1, 2, 4, and 7 d obtained by our method is 0.32, 0.48, 0.68 and 0.94 TECU.