Development and evaluation of a new 3-D ionospheric modeling method

Described in this paper is a new three-dimensional (3-D) ionospheric modeling method developed on the basis of tomographic techniques and GPS network data. This model is constructed with spherical harmonic and empirical orthogonal functions. The methods for constructing the model and evaluating its performance are first developed. Performance assessment results are then presented using datasets from two different types of GPS reference networks. Performance is evaluated in terms of ionospheric total electron content (TEC) prediction accuracy and the impact on improving single-point positioning accuracy. Results indicate that short-term (5 min) predictions can typically recover over 95 percent of the total ionospheric errors when modeling a local-area GPS network under medium ionospheric activity conditions. Under extremely active ionospheric conditions (Kp = 8-9), recovery can achieve an average level of about 72 percent using data from a wide-area GPS network.