Investigation into the Space Weather Event of September 2017 through GNSS Raw Samples Processing

Global Navigation Satellite System (GNSS) positioning quality, continuity and reliability can be severely impaired by ionospheric propagation, in particular during strong space weather events. Ionospheric amplitude and phase scintillation cause power fading and random phase fluctuations of the received signal, leading to a reduced performance of the receiver tracking loops. This can translate into significant errors in the final position solution, especially when higher accuracy is concerned. In the period between 4 and 10 September 2017, several space weather events, including two coronal mass ejections, triggered disturbed conditions of the near-Earth space. This interval was considered one of the most flare-productive of now-waning solar cycle. Significant disturbances on GPS L1 signals have been observed and automatically recorded by a network of GNSS monitoring stations based on customized Software Defined Radio data grabbers and receivers. Relevant and highly unique raw samples are being processed to deeply investigate the phenomenon. The analysis of the data collected in different stations, in terms of scintillation indices time-series, shows a connection between the recording in the different locations of the world. In addition, the impact of such a strong event in receiver has been analyzed, focusing on positioning algorithms employing phase measurements to smooth code measurements. Results show a good correlation between the values of the scintillation indices and the positioning error, showing that carrier-smoothing techniques are particularly sensitive to distortion induced by phase scintillation. Such an impact is automatically observed by means of the identification of clusters in the positioning solutions through a machine learning algorithm. The adoption of carrier smoothing hence reveals weaknesses in precise positioning during scintillations events. At the same time, a novel scintillation detection technique based on clustering of the position solution is suggested.