Study Of Ionosphere TEC Prediction Models

The radio wave with 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. Firstly, the two most common TEC prediction models, the Klobuchar model and the calculating TEC part in international reference ionosphere, are detailed introducted and their prediction results are compared with the IGS ionosphere TEC products. Then, the principle of time series model is described and their prediction result is compared with the IGS ionosphere TEC products too. Using time series model to predicte TEC, different order determination methods have different prediction results. The FPE(Final prediction Erro), AIC (A-Information Criterion) and BIC(B-Information Criterion) order determination methods are used besides the order determination method proposed by Li Zhigang and Cheng Zongyi for time series prediction TEC in this paper.The results calculated by the four kinds of order determination methods are statistic compared.If the model order is not restricted, the prediction precision is equivalent. At last, the applicability of the three ionosphere TEC prediction models is analzed and compared. The results shows that time series is the most accurate method in case of short-term prediction of ionospheric TEC,the IRI model can be used in case of longer period TEC prediction values. If a simple TEC model is needed, Klobuchar model can be used.