Development and Analysis of a New Ionospheric Model

A new three-dimensional (3D) ionospheric model has been developed based on the tomography technique. This new ionospheric model employs ionospheric TEC data as input which are calculated from a network of dualfrequency GPS receivers. The 3D ionospheric model is fundamentally constructed by spherical harmonic functions and empirical orthogonal functions (EOFs). The spherical harmonic functions describe the horizontal ionospheric characteristics while the EOFs characterize vertical ionosphere properties. In this paper, the methodology to construct the ionospheric tomography model is first developed. Then the performance of the new model is assessed via the analysis of two data sets from two GPS reference networks with different numbers of GPS stations and network sizes. The model's accuracy is evaluated in two aspects. One is its ionospheric prediction accuracy. The prediction accuracy is evaluated by producing predicted TEC data using this new model and comparing them to the GPS measured TEC data. Another aspect is the application of these predicted TEC data to a real GPS positioning and navigation system. The TEC data predicted by the 3D ionosphere model are used in a precise point positioning (PPP) software package to correct the ionospheric refraction errors. As a comparison, two other ionospheric models are also used in the PPP. The final positioning accuracies corresponding to these ionospheric models are compared. The analysis of 24- hour data indicates that using the 3D tomography model within a medium size GPS network and under medium ionospheric activity condition, a short-term (5 minutes) ionosphere prediction can recover about 95% total ionospheric errors. Even under extremely variable ionospheric conditions (Kp=9), the test results using a large size GPS network confirm that over 80% total ionospheric error can be reproduced by employing the tomography model.