Towards an Improved Single-Frequency Ionospheric Correction: Focus on Mid-Latitudes

The modelling of the Total Electron Content (TEC) plays an important role in global satellite navigation systems (GNSS) accuracy, especially for single frequency receivers, the most common ones constituting the mass market. For the latter and in the framework of Galileo, the NeQuick model has been chosen for correcting the ionospheric error contribution. It has been designed to calculate the electron density at a given point of the ionosphere according to the time conditions and the solar activity. This electron density can be integrated along the path from the receiver to the considered satellite to provide the TEC. For Galileo, a parameter Az (“effective ionisation level”) will be provided to the model as solar activity information and will be daily updated by the ground stations.
In order to reach the ionosphere error correction level objective (70% or 20 TECu whichever is larger), the model itself as well as its use for Galileo are investigated. Different situations have to be considered: different latitude regions (space conditions), different hours, seasons and years (time conditions) and specific phenomena appearance (magnetic storms, Travelling Ionospheric Disturbances – TIDs). In addition the results can be compared to different data sets among which GPS slant or vertical TEC (sTEC or vTEC) measurements, Global Ionospheric Maps, ionosonde profiles, topside soundings but also other ionosphere models results such as IRI.
In our comparison process, we take benefit of various ionosphere data from several European stations (Chilton in UK, Dourbes in Belgium, El Arenosillo and Roquetes in Spain, Pruhonice in Czech Republic) where ionosonde and GPS TEC data are available for different solar activity levels. These data allow us to study NeQuick representation of the ionosphere at mid-latitudes. We investigate the difference between GPS-derived vTEC and corresponding values from NeQuick for the latest years (between solar maximum in 2000 and minimum in 2007) in order to observe the temporal dependencies towards Universal Time, season and solar activity. On the one hand, we use ionosonde data to constrain the model so that we can concentrate on its formulation of the profile only. We especially highlight the improvements from the latest (second) version of NeQuick and show the critical importance of the topside formulation. On the other hand, we analyse the model residual errors for the same situations computing vTEC through the Galileo algorithm.