A Study of Prediction Models for Ionosphere

Abstract The radio wave used in communications at distance and navigation by satellite is affected by ionosphere. Therefore study on the total electron content (TEC) in the ionosphere is very important. There are some ionospheric prediction models, but the accuracy of prediction by them is 50%~60% only. A new ionosphere model to predict total electron content is proposed. The total electron content given by IGS at each node of grid is fitted by spherical‐harmonic expansion. The fitted coefficients in different time form the time series, these time series are predicted by ARMA ( p, q ) model so that the prediction of total electron content in ionosphere is implemented. The TEC data of IGS in 2004 and 2005 are analyzed by our method in order to predict the TEC in geographic frame. The relative predicting accuracy for 5 days obtained by our method is about 90%.