Evaluation of three ionospheric delay computation methods for ground-based GNSS receivers

GNSS observables for ionospheric estimation are commonly based on carrier-to-code leveling (CCL) and precise point positioning (PPP) methods. The CCL method is a geometry-free method which uses carrier phase to level pseudorange observation for decreasing multipath error and observation noise. However, the ionospheric observable based on the CCL has been proven to be affected by leveling errors. The leveling errors are caused by pseudorange multipath and intraday variation of receiver DCB. To obtain more accurate ionospheric observable, the PPP method takes advantage of precise satellite-to-ground range for retrieving slant total electron content and is less affected by the leveling errors. Previous studies have only proven that the ionospheric observables extracted by the two methods are affected by the leveling errors. The influence on ionospheric observable by the pseudorange inter-receiver satellite bias (IRSB) of the receiver has not been taken into consideration. Also, the magnitude of the differences between the ionospheric observables extracted by the two methods has also not been given. In this work, three methods, namely, the CCL, the conventional ionospheric-free PPP method which uses the ionospheric-free Hatch–Melbourne–Wubbena (HMW) function, and the University of Calgary (UOFC) PPP method, are selected to analyze and compare the differences of ionospheric observables and the global ionospheric maps, using a large number of measured data from international GNSS service global stations. Experimental results show that the accuracy of ionospheric observables obtained by the three methods is not only related to the leveling error, but also pseudorange IRSB. The IRSB of the receiver exerts a major effect on the ionospheric observables obtained by the CCL method and a minor effect on the ionospheric observables obtained by the HMW and UOFC methods. The accuracies in the latter case are similar and superior to those obtained by the CCL. The differences of the ionospheric observables obtained by the CCL and UOFC methods, or the CCL and HMW methods, are at decimeter level, whereas the difference of the ionospheric observables obtained by the UOFC and HMW methods is at centimeter level. The UOFC method presented the highest single-frequency pseudorange positioning accuracy using estimated global ionospheric products, followed by the HMW and the CCL methods which presented the lowest positioning accuracy.