The Study On Prediction Models And Methods Of Ionosphere Based On Ground-Based GNSS Over China

Ionospheric delay error has been one of the main error sources in Global Navigation Satellite System(GNSS)positioning.To eliminate the ionospheric delay error via the improvement of accuracy and reliability of the GNSS positioning has become an important research topic.Total Electron Content(TEC)is a key element to indicate the morphology and structure of ionosphere in satellite navigation.The research on its temporal changes and predictions can provide an important information in studying the morphology of ionosphere.In addition,the researches and predictions have great significance on high-accuracy survey in GNSS,timing navigation positioning and network Real Time Kinematic(RTK)as well.Under this circumstances,this paper will study on the temporal variation of the peak in TEC,the shorttime prediction of TEC and the method to improve the ionospheric delay model.The main contents and conclusions of the paper are the followings:1.This paper uses the Global Ionosphere Maps(GIM)data provided by the International Global Navigation Satellite System Service(IGS)center from 1999 to 2015 to calculate the asymmetry index of TEC peak values in the northern and southern hemisphere.Then,analyzing the correlation between the asymmetry index of TEC peak values and the sunspot number as well as geomagnetic activity index.At last,the variation of the peak values in different longitudes during a day are analyzed.After analyzing the data,it can be concluded that:(1)the asymmetry index of TEC peak values have seasonal cycle and annual cycle obviously,either in the western hemisphere or in the eastern hemisphere;(2)the asymmetry index of TEC peak values are much more related with sunspot in western hemisphere than that of in eastern hemisphere;while the asymmetry index of TEC peak values has a parallel relation with geomagnetic activity index in both eastern hemisphere and western hemisphere;and(3)other than unobvious peak value at 08:00 LT to 10:00 LT on both side of the equator,the rest hours have obvious peak values in every period,especially peaking at 12:00 LT ~ 16:00 LT.2.The 15 day's GIM data over China are provided by IGS center at quiet and active period.The previous 10 days' TEC values are taken as sample data,the prediction models are set up by three common time series analyze methods.Afterwards,the established models are used to forecast the latter 5 days' TEC data.The experimental results are manifested below:(1)generally speaking,Winter has the greatest anticipation on prediction,followed by Spring and Summer while Autumn has the worst one;(2)additive model has a greatest prediction in different grids and seasons whether in quiet or active period;and(3)there are significant differences in Root Mean Square Error(RMSE)and the relative accuracy of different latitudes reaching a maximum number in 22.5°N in all forecast results of the three models and the forecast results has a strong correlation with TEC in the longitude circle.3.The observation data of seventeen IGS stations within and around China from 2011 provided by the IGS center are used to calculate the TEC values using the Klobuchar model and the dual-frequency model.The Holt–Winters exponential smoothing model is used to forecast the error of the 7th day between the Klobuchar model and the dual-frequency model by using the error of the former six days.The forecast results are used to develop the sophisticated Klobuchar model when no epochs are missing,considering that certain reasons may result in some of the observation data being missing and weaken the relationship between each epoch in practical applications.We study the applicability of the sophisticated Klobuchar model when observation data are missing.This study deletes observation data of some epochs randomly and then calculates TEC values using the Klobuchar model.A cubic spline curve is used to restore the missing TEC values calculated in the Klobuchar mode.Finally,we develop the sophisticated Klobuchar model when N epochs are missing in China.The sophisticated Klobuchar model is compared with the dual-frequency model.The experimental results reveal the following:(1)the sophisticated Klobuchar model can correct the ionospheric delay more significantly than the Klobuchar model;(2)the sophisticated Klobuchar model can reflect the ionosphere temporal evolution,particularly at night,with the correct results increasing with increasing latitude;and(3)the sophisticated Klobuchar model can achieve remarkable correction results when N epochs are missing,with the correct results being nearly as good as that of the dual-frequency model when no epochs are missing.