Assessment of global and regional ionospheric corrections in multi-GNSS PPP

The general objective of this study is to assess and analyze the performance of Global Ionospheric Maps (GIM) and ionospheric corrections from nearby stations through stochastic constraints in Precise Point Positioning (PPP) initialization and convergence. Key differences between the estimated slant ionospheric term and GIM delays were presented and discussed. One of the objectives was to answer to the question: How precise does GIM or regional corrections have to be in order to further improve PPP float and fixed solutions in comparison to the estimated slant ionospheric delays? The answer, though satellite dependent, ranges from a few decimetres to the metre-level. The correlation between the choice of GIM product and how it impacts multi-GNSS PPP convergence was also investigated using predicted and post-processed products, either combined from different Analysis Centers (ACs) or standalone. It was deduced that the choice of products, whether they are predicted, combined or post-processed, impacts PPP convergence when used to constrain the solution. There are over 20 existing different IONEX post-processed products and the choice of selection PPP ionospheric solution constraints in either dual- or triple-frequency multi-GNSS processing modes. Another specific objective of this study was to review the performance of these post-processed products using globally distributed network of multi-GNSS stations. The solutions from the stations were constrained using each product with a horizontal error threshold of 10 cm and convergence time of 10 minutes in dual-and triple-frequency modes. Results indicated centimetre to millimetre differences between the different products which can either be critical or indifferent depending on the user application. The key intended message however, implies differences between the IONEX generated products and their impact on multi-GNSS PPP performance. The correlation between different latitudes and the effect of ionospheric corrections in multi-GNSS PPP was also investigated. Using homogenous groups of stations in upper, middle and lower latitudes, ionospheric constraints were applied to PPP solutions in dual- and triple-frequency processing modes. Considering a 10-minute convergence period, 40%, 50% and 18% of stations reached a 20 cm horizontal threshold in the upper, mid and lower latitudes, respectively, for dual-frequency processing. In the triple-frequency case, 46%, 48% and 36% of stations reached the horizontal threshold in the upper, middle and lower latitudinal regions, respectively. For regional analysis, slant ionospheric corrections from nearby stations relative to a reference site, with varying baseline distances, were used to constrain the reference station. The research question intended to be answered with this analysis was how close does a reference station have to be or how densely close does a network of stations have to be, in order to obtain PPP-AR like performance in multi-GNSS PPP processing? GNSS stations with baseline distances ranging from 50 to 200 km were analyzed. It was concluded that within a network of 100 km spaced stations, it is possible to obtain 10 to 15 minutes of convergence when considering a horizontal error threshold of 10 cm.