4:15 PM - 4:30 PM
[SCG55-04] Development of coseismic fault model estimation system using real-time PPP
Keywords:GNSS, REGARD, MADOCA, PPP, GEONET
Geospatial Information Authority of Japan (GSI) has operated REGARD (REal-time GEONET Analysis system for Rapid Deformation monitoring), which rapidly estimates finite fault models using permanent displacement field based on the continuous GNSS observation network called GEONET. Since the real-time GNSS has stable sensitivity to large earthquakes with predominant long-wavelength components, the purpose of REGARD is to provide information on the magnitude of such large earthquakes and their associated fault dimensions, and to support tsunami prediction.
REGARD conducts real-time kinematic (RTK) analysis that constructs radial baselines between one reference station and other stations. If the observation at the reference station is stopped, it affects to the positioning for all of the baselines, and if physical fluctuation or analytical noise occurs at the reference station, the effect will spread to results of all baselines, which will adversely affect the fault estimation results. In order to overcome these weaknesses of relative positioning, it is desirable to improve the positioning method itself, which depends on one reference station.
Based on these backgrounds, we attempted to introduce Precise Point Positioning (PPP), which does not have the disadvantages of relative positioning because PPP can perform positioning alone. On the other hand, since there is no improvement in positioning accuracy by using baselines, precise satellite orbit/clock information are required to achieve real-time PPP with the same level of accuracy as the current REGARD. Therefore, the system under development includes a new process to generate the precise orbit/clock information with high temporal resolution and to create resultant correction information, whereas the current REGARD depends on external data. For the new calculation, we used MADOCA (Multi-GNSS Advanced Demonstration tool for Orbit and Clock Analysis) provided by Japan Aerospace Exploration Agency (JAXA).
We are currently conducting real-time PPP at GEONET stations, and are evaluating the positioning accuracy and stability of the fault model estimation. The accuracy of real-time PPP was comparable to that of RTK on the time scale on which REGARD estimates coseismic crustal deformation (about 5 minutes). On the other hand, the noise caused by the estimated precision information existed as systematic apparent variations in all stations. Since the noise does not differ much from region to region, we tried to automatically remove the trend in the estimation process of crustal deformation by estimating the trend using the variations at the stations far from the epicenter. In addition, we tried to introduce a process to extract stations with extreme noise using positioning errors. In this presentation, we will show the stability of the crustal deformation obtained by these processes and the results of their application to past offshore earthquakes (the 2003 Tokachi-Oki earthquake, the 2011 Tohoku-Oki earthquake, and the 2011 Ibaraki-Oki earthquake).
REGARD conducts real-time kinematic (RTK) analysis that constructs radial baselines between one reference station and other stations. If the observation at the reference station is stopped, it affects to the positioning for all of the baselines, and if physical fluctuation or analytical noise occurs at the reference station, the effect will spread to results of all baselines, which will adversely affect the fault estimation results. In order to overcome these weaknesses of relative positioning, it is desirable to improve the positioning method itself, which depends on one reference station.
Based on these backgrounds, we attempted to introduce Precise Point Positioning (PPP), which does not have the disadvantages of relative positioning because PPP can perform positioning alone. On the other hand, since there is no improvement in positioning accuracy by using baselines, precise satellite orbit/clock information are required to achieve real-time PPP with the same level of accuracy as the current REGARD. Therefore, the system under development includes a new process to generate the precise orbit/clock information with high temporal resolution and to create resultant correction information, whereas the current REGARD depends on external data. For the new calculation, we used MADOCA (Multi-GNSS Advanced Demonstration tool for Orbit and Clock Analysis) provided by Japan Aerospace Exploration Agency (JAXA).
We are currently conducting real-time PPP at GEONET stations, and are evaluating the positioning accuracy and stability of the fault model estimation. The accuracy of real-time PPP was comparable to that of RTK on the time scale on which REGARD estimates coseismic crustal deformation (about 5 minutes). On the other hand, the noise caused by the estimated precision information existed as systematic apparent variations in all stations. Since the noise does not differ much from region to region, we tried to automatically remove the trend in the estimation process of crustal deformation by estimating the trend using the variations at the stations far from the epicenter. In addition, we tried to introduce a process to extract stations with extreme noise using positioning errors. In this presentation, we will show the stability of the crustal deformation obtained by these processes and the results of their application to past offshore earthquakes (the 2003 Tokachi-Oki earthquake, the 2011 Tohoku-Oki earthquake, and the 2011 Ibaraki-Oki earthquake).