SBAS Ionospheric Modeling with the Quadratic Approach: Reducing the Risks

摘要

In principle, the quadratic approach can portray ionospheric total electron content (TEC) better than the linear approach by introducing the non-linear spatial variation terms into the model. The previous research [Rho et al., 2004] has shown that the most improvement in the quadratic approach is in the improvement of r.m.s. of residuals. This approach was tested in a batch process using data from 48 reference stations in North America. The algorithm ensures that the estimator is free from the main risks, such as uneven distribution of data and even data holes which can lead to spurious spikes and unphysical features in the resulting models. However, we are left with several issues to properly use the quadratic approach in SBAS. First, the quadratic model with a batch process requires significant computational power when more and more GPS stations providing more and more data are used to determine the model parameters at each epoch. Second, since the batch process uses all the collected ionospheric measurements from all the stations in the network, the modeled results are smoothed over the network and are less sensitive to the local ionospheric variations at each station. And third, as long as a limited number of reference stations are used to estimate the ionospheric vertical delays at the specific grid points of an SBAS ionospheric model, there exist some difficulties, such as a limited density of measurements and geometry issues.