A Rigorous Framework for the Statistical Characterization of Errors in Broadcast GNSS Ionospheric Model

Methodologies to perform a statistical characterization of errors of ionospheric mitigation strategies have been a long-standing issue in the Global Navigation Satellite System field. The present study aims to demonstrate the benefits of adopting a framework that isolates the ionospheric modeling error by using slant Total Electron Content reference data with centimeter-level accuracy. This reference data is obtained relying on carrier-phase measurements from a global network of double-frequency receivers processed with an Integer Ambiguity Resolution algorithm. After a Least Squares estimation made to determine the station and satellite biases associated with any other ionospheric model, the post-fit residuals are used to evaluate its accuracy. Such analyses should be carried out with the use of experimental Cumulative Distribution Functions to avoid the skewing of alternate simpler metrics by outliers. This framework has been implemented to test three major Ionospheric Correction Algorithms (Klobuchar, NeQuick G, NTCM-G) and the final product of the Global Ionospheric Map computed by CODE to assess the performance both in nominal conditions and during a perturbed geomagnetic event. Results show how NeQuick G outperforms NTCM-G in non-equatorial latitudes on the 95th percentile by more than 10%, and that CODE products register periods and regions in which their ionospheric error quadruples with respect to nominal conditions.