Previous Abstract Return to Session B3 Next Abstract

Session B3: Atmospheric Effects

Mitigation of High Latitude Ionospheric Scintillation Effects on Precise Point Positioning (PPP) During the September 2019 Geomagnetic Storm
Kai Guo, Marcio Aquino, Sreeja Vadakke Veettil, Chris Hill, Brian Weaver, Nottingham Geospatial Institute, University of Nottingham, UK
Location: Pavilion Ballroom West
Alternate Number 1

Radio frequency (RF) signals can suffer from rapid and random fluctuations, known as ionospheric scintillation, when they pass through plasma density irregularities in the ionosphere. As a result, signal quality and receiver performance of Global Navigation Satellite System (GNSS), such as Global Positioning System (GPS), GLONASS, Galileo and Beidou, can be adversely affected, thereby leading to increased positioning errors. Scintillation presents a higher probability of occurrence at high latitudes and equatorial regions (Basu et al. 1988). Over high latitudes, scintillation is more frequent during solar maximum and during geomagnetic storms. The S4 and Phi60 indices are generally used to represent amplitude and phase scintillation levels, which relate to the fluctuations of amplitude and phase in RF signals, respectively. This study analyses the scintillation events observed at high latitudes during the geomagnetic storm that occurred in September 2019. Effects of scintillation on Precise Point Positioning (PPP) are investigated. The Phase Locked Loop (PLL) and the Delay Locked Loop (DLL) tracking jitter in the presence of scintillation is estimated at a higher rate. Scintillation effects on GNSS positioning are mitigated using the method described in Aquino et al. (2009).
The scintillation data processed in this study are collected at Longyearbyen (Lat. 78°13?N, Long. 15°38?E) in Svalbard, denominated as LYB0 station, where a Septentrio PolaRxS receiver is operational. The PolaRxS is a specialised Ionospheric Scintillation Monitoring Receiver (ISMR) capable of outputting scintillation indices and high frequency (50 Hz) measurements. Scintillation data recorded on the GPS L1 frequency during the geomagnetic storm from 29 August to 3 September 2019 are chosen to carry out the analysis in this work.
PPP calculation is performed by the POINT software, developed for research at the University of Nottingham. Code and carrier phase measurements on the GPS L1 and L2 frequencies at a rate of 1 s are used to form the ionosphere-free linear combination. International GNSS Service (IGS) Multi-GNSS Experiment (MGEX) Precise orbits and clock products estimated by the Centre for Orbit Determination in Europe (CODE) were constrained in the PPP model. Differential Code Biases (DCB) and satellite and receiver antenna corrections from IGS are used in the processing. The tropospheric delay is estimated through a random walk process. A satellite elevation angle mask of 7 degrees is set in the PPP processing.
The occurrence of amplitude and phase scintillation is firstly presented. For the scintillation occurrence study, only signals from satellites with elevation higher than 30 degrees are considered in order to remove non-scintillation related effects, such as the multipath. Additionally, only scintillation events with scintillation index S4 or Phi60 higher than 0.3 are counted. Results show that a total number of 364 phase scintillation events are observed on the GPS L1 signal from 29 August to 3 September 2019. The occurrence increases from 1 to 117 between 29 and 31 August, when it reaches its maximum, followed by a gradual decrease until 3 September. This trend agrees well with the variation of geomagnetic activity levels as characterised by the Kp index. Moreover, strong phase scintillation with Phi60 higher than 0.7 is frequently observed on 31 August and 1 September, which are exactly the most disturbed days during the geomagnetic storm. By contrast, only a total number of 53 amplitude scintillation events with S4 lower than 0.4 are observed over these 6 days. This indicates that phase scintillation is much more frequent and stronger than amplitude scintillation in this region.
To investigate the effects of scintillation on GPS positioning, PPP processing is carried out over these six days. As there is no obvious scintillation observed on 29 August, the precise coordinates obtained by static PPP, which solves for the station coordinates based on 24 hours data on this day, are set as the reference coordinates. The root mean square (RMS) of positioning errors obtained using kinematic PPP (solving for the coordinates epoch per epoch) in the east, north and up directions are calculated using a satellite elevation based weighting strategy. The RMS of 3D errors are also computed. It should be noted that due to the convergence period of PPP processing, the first hours of the positioning error time series on every day are not considered in the analysis. Results show that on 29 August, the positioning errors in all the three directions are less than 0.05 m over the day, while, they can be as large as 10.60 m in the up direction on 31 August during strong scintillation occurrence. Additionally, the RMS of the 3D error increases gradually from 0.04 m on 29 August to 0.36 m on 31 August and to 0.59 m on 1 September, followed by a decrease to 0.04 m on 3 September. This agrees with the variation of geomagnetic activity levels. Thus, it can be concluded that GPS positioning errors may increase significantly under scintillation related to geomagnetic storms. Furthermore, it can be seen that the positioning errors in the up direction show more prominent fluctuations compared with those in the east and north directions, indicating that the positioning accuracy in the up direction is more affected by scintillation.
Conker et al. (2003) derived formulas to evaluate the Phase Locked Loop (PLL) and the Delay Locked Loop (DLL) tracking jitter variance under scintillation, using 1-minute scintillation indices and CN0 values as inputs. The inverse of these estimated tracking jitter variance is exploited by Aquino et al. (2009) to define weights used to modify the least square stochastic model in order to mitigate scintillation effects on GPS positioning. In this study, the approach described by Aquino et al. (2009) is applied, however, by obtaining scintillation parameters and CN0 at a 1-second rate, rather than 1-minute, and subsequently using them to calculate the tracking jitter. The 1-second tracking jitter can provide more detail about the signal fluctuations within 1 min, thus better representing the signal noise level in the presence of scintillation. With this approach, PPP is performed from 29 August to 3 September 2019. Positioning errors estimated using the tracking jitter weighting strategy are then compared with those using the satellite elevation based weighing strategy. Results show that when using the tracking jitter weighting, the RMS of 3D errors is generally improved over all six days. On 29 August, when there is no scintillation, an improvement of 1.82 % is seen, while, on the more scintillation active days, i.e. 31 August and 1 September, significant improvements of 51.98 % and 93.02 % are found, respectively. Therefore, it can be concluded that using the 1-second based tracking jitter to modify the least square stochastic model can successfully mitigate the scintillation effects on GPS positioning. The PPP results applying 1-min based tracking jitter to modify the stochastic model will be presented and compared with the results presented herein.
The following conclusions can be drawn based on the analysis:
By studying the occurrence of scintillation recorded at high latitudes, it is found that the occurrence of phase scintillation increases dramatically during a geomagnetic storm. In the case analysed in the study, it peaks on 31 August with 117 occurrence events of Phi60 higher than 0.3. However, only a few amplitude scintillation events with S4 lower than 0.4 are observed. This indicates that high latitude phase scintillation was observed to be much more frequent and stronger than amplitude scintillation during the geomagnetic storm.
PPP errors in the presence of scintillation are analysed. Results show that the RMS of the 3D error increases significantly on 31 August and 1 September, which agrees with the variation of geomagnetic activity levels. Additionally, the positioning errors in the up direction show more prominent fluctuations compared with those in the east and north directions, which means that the positioning accuracy in the up direction is more affected by scintillation.
The approach described in Aquino et al. (2009) is exploited in this study to mitigate scintillation effects on GPS positioning. However, in this new study 1-second based tracking jitter is used instead of its 1-minute counterpart to modify the least square stochastic model. It is found that using the 1-second based tracking jitter to modify the least square stochastic model can successfully mitigate the scintillation effects on GPS positioning. In the presence of strong scintillation, the RMS of 3D PPP errors is significantly improved by up to 93.02 % compared with the satellite elevation based weighting solution. PPP processing using 1-min tracking jitter to modify the least square stochastic model will also be presented and compared with the results herein.
This work is relevant for a better understanding of the high latitude scintillation. It is also beneficial for developing scintillation mitigation tools for GNSS positioning.
Acknowledgement
The author thanks the TREASURE project (www.treasure-gnss.eu), which has received funding from the European Union’s Horizon 2020 research and innovation programme under the Marie Sk?odowska-Curie grant agreement No. 722023. The authors want to thank Dr. Claudio Cesaroni and Juliana Damaceno in National Institute of Geophysics and Volcanology (INGV) in Italy, who kindly provided the scintillation data for analysis. The authors used historical Kp index data provided by German Research Centre for Geosciences (ftp://ftp.gfz-potsdam.de/pub/home/obs/kp-ap/tab/).
Reference
Basu, S., MacKenzie, E., & Basu, S. (1988). Ionospheric constraints on VHF/UHF communications links during solar maximum and minimum periods. Radio Science, 23(3), 363-378. https://doi.org/10.1029/RS023i003p00363
Conker, R. S., El-Arini, M. B., Hegarty, C. J., & Hsiao, T. (2003). Modelling the effects of ionospheric scintillation on GPS/satellite-based augmentation system availability. Radio Science, 38(1), 1-1-1-23. https://doi.org/10.1029/2000RS002604
Aquino, M., Monico, J. F. G., Dodson, A. H., Marques, H., De Franceschi, G., Alfonsi, L., et al. (2009). Improving the GNSS positioning stochastic model in the presence of ionospheric scintillation. Journal of Geodesy, 83(10), 953–966. https://doi.org/10.1007/s00190-009-0313-6



Previous Abstract Return to Session B3 Next Abstract