Item

Abstract

This paper presents a theoretical study on ionospheric reconstruction using GNSS data obtained from Low Earth Orbit (LEO) satellites in a PNT (Position, Navigation, and Timing) configuration, where the LEO satellites not only receive but also transmit GNSS signals, which can be tracked by ground or mobile receivers. The study is intended to pave the way for incorporating slant Total Electron Content (TEC) data from ESA’s upcoming LEO-PNT into ionospheric reconstructions. We generate synthetic slant TEC for three observation scenarios: Ground-GNSS, ground-LEO, and LEO-GNSS links. As ground-truth, the IRI-20 model with the Ozhogin plasmasphere extension is used. An inversion to recover the electron density from slant TEC observations is performed using an Extended Kalman Filter (EKF) in the information-filter formulation for all possible combinations of observation scenarios. As the LEO constellation, we will utilize existing LEO satellites that were available in May 2020, including Swarm, COSMIC-2, GRACE-FO, Jason-3, Sentinel-1, Sentinel-2, and Sentinel-3, as well as several Spire satellites. They cover a variety of altitudes between 400 km and 1350 km. For this study, we assume they could transmit dual-frequency GNSS-like signals like a PNT mission, which is not the case for any of the satellites mentioned. We only consider relative slant TEC to be insensitive to calibration biases that may reach a few TEC units. Given a real global ground-station network, LEO and GNSS satellites, we show that 15-min reconstruction solutions, only containing ground stations, cannot compete with solutions including LEO satellites. Furthermore, our results show that the joint use of LEO-POD (Precise Orbit Determination Antenna) and LEO-PNT (RMSE at 500 km: ) provides superior performance compared to configurations where either is substituted by ground-based GNSS (ground-GNSS and PNT: ; Gound-GNSS and POD ). We also show that the reconstruction error roughly doubles when radio occultation measurements are omitted. The dependency of the error on the distribution of the ground stations is also shown. Areas with only a few or no ground stations show the lowest correlation between IRI-20 and the reconstructions, e.g., near Point Nemo, where the correlation drops to 0.5.