10:45 AM - 12:15 PM
[STT39-P01] Monitoring ground deformation in eastern Japan using InSAR time series analysis with ALOS-2 data
Keywords:InSAR time series analysis, ALOS-2, Crustal deformation, Ground deformation
Introduction
Geospatial Information Authority of Japan (GSI) has been conducting InSAR analysis of ALOS-2 images to monitor crustal and ground deformation in Japan. However, it is difficult to detect slowly progressing deformation because SAR interferograms contain various errors. On the other hand, with the ALOS-2 observation data accumulated since its launch in 2014, we can perform InSAR time series analysis to improve detection accuracy by statistically processing a large number of interferograms. In addition, in the case of ALOS-4, the successor to ALOS-2, which is scheduled to be launched in the future, observation frequency will be about five times more than ALOS-2. Therefore, InSAR time series analysis using ALOS-4 data is expected to enable more accurate measurement.
Based on the above, GSI has been conducting InSAR time series analysis using ALOS-2 data for all of Japan for monitoring crustal and ground deformation nationwide, as well as for maintaining and managing survey control points. The results of the analysis for Hokkaido area have been released in 2022. In this presentation, we report on details of deformation detected by the results InSAR time series analysis for eastern Japan, released in January 2023.
Data and Analysis
For InSAR time series analysis, we use ALOS-2 data observed from descending and ascending orbits from 2014 to 2021.
First, pairs within 200 m perpendicular baseline and within 730 days (descending) or 1095 days (ascending) time interval were selected. Then, each interferogram was corrected for tropospheric delay using the numerical weather model (Kobayashi et al., 2014), and ionospheric delay with the Split-Spectrum method (Wegmüller et al., 2018), and long-wavelength noise using GNSS data (GNSS correction; Kobayashi et al., 2011; Tobita et al. 2005). In GNSS corrections, we estimated and reduced differences between SAR and GNSS, as well as wide-ranging and long-term tectonic deformation such as post-seismic deformation caused by the 2011 off the Pacific coast of Tohoku earthquake.
After that, InSAR time series analysis was conducted by phase optimization (SBAS method; Berardino et al., 2002) with temporal smoothing process using GSITSA (Kobayashi et al., 2018). Finally, 2.5D analysis (Fujiwara et al., 2000) was performed using displacement velocities in the LOS direction from descending and ascending orbits to obtain quasi-vertical and quasi-east-west displacement velocities.
Result
Figure 1 shows one of the results of the InSAR time series analysis, the distribution of displacement velocity in the LOS direction from the ascending orbit in eastern Japan. We successfully estimated the velocities over the entire area, not only in the flat area but also in the mountainous areas, suggesting that L-band SAR data are indispensable for nationwide monitoring. Thanks to the time series analysis, we succeeded in detecting slowly progressing deformation, some of which could not been clearly identified in conventional InSAR analysis results due to low S/N ratio. In particular, the detectability for local deformation is substantially improved.
Figure 2 shows a result of the 2.5D analysis. Significant subsidences are seen at the Tsugaru Plain and the Echigo Plain (Fig. 2-a, b). These are thought to indicate ground subsidence. On the other hand, displacements dominant in the east-west direction and along the topography are seen at Mt. Hakkoda and Mt. Gassan (Fig. 2-c, d). These may imply landslides. Noted that tectonic deformation is removed through the GNSS correction.
We will discuss the details such as temporal transition of the detected displacements by this analysis.
Acknowledgments
ALOS-2 data were provided based on the joint research agreement with JAXA and under a cooperative research contract between GSI and JAXA. The ownership of ALOS-2 data belongs to JAXA. The numerical weather model was provided by JMA based on the agreement. Some of GNSS data were provided by JMA and NIED.
Geospatial Information Authority of Japan (GSI) has been conducting InSAR analysis of ALOS-2 images to monitor crustal and ground deformation in Japan. However, it is difficult to detect slowly progressing deformation because SAR interferograms contain various errors. On the other hand, with the ALOS-2 observation data accumulated since its launch in 2014, we can perform InSAR time series analysis to improve detection accuracy by statistically processing a large number of interferograms. In addition, in the case of ALOS-4, the successor to ALOS-2, which is scheduled to be launched in the future, observation frequency will be about five times more than ALOS-2. Therefore, InSAR time series analysis using ALOS-4 data is expected to enable more accurate measurement.
Based on the above, GSI has been conducting InSAR time series analysis using ALOS-2 data for all of Japan for monitoring crustal and ground deformation nationwide, as well as for maintaining and managing survey control points. The results of the analysis for Hokkaido area have been released in 2022. In this presentation, we report on details of deformation detected by the results InSAR time series analysis for eastern Japan, released in January 2023.
Data and Analysis
For InSAR time series analysis, we use ALOS-2 data observed from descending and ascending orbits from 2014 to 2021.
First, pairs within 200 m perpendicular baseline and within 730 days (descending) or 1095 days (ascending) time interval were selected. Then, each interferogram was corrected for tropospheric delay using the numerical weather model (Kobayashi et al., 2014), and ionospheric delay with the Split-Spectrum method (Wegmüller et al., 2018), and long-wavelength noise using GNSS data (GNSS correction; Kobayashi et al., 2011; Tobita et al. 2005). In GNSS corrections, we estimated and reduced differences between SAR and GNSS, as well as wide-ranging and long-term tectonic deformation such as post-seismic deformation caused by the 2011 off the Pacific coast of Tohoku earthquake.
After that, InSAR time series analysis was conducted by phase optimization (SBAS method; Berardino et al., 2002) with temporal smoothing process using GSITSA (Kobayashi et al., 2018). Finally, 2.5D analysis (Fujiwara et al., 2000) was performed using displacement velocities in the LOS direction from descending and ascending orbits to obtain quasi-vertical and quasi-east-west displacement velocities.
Result
Figure 1 shows one of the results of the InSAR time series analysis, the distribution of displacement velocity in the LOS direction from the ascending orbit in eastern Japan. We successfully estimated the velocities over the entire area, not only in the flat area but also in the mountainous areas, suggesting that L-band SAR data are indispensable for nationwide monitoring. Thanks to the time series analysis, we succeeded in detecting slowly progressing deformation, some of which could not been clearly identified in conventional InSAR analysis results due to low S/N ratio. In particular, the detectability for local deformation is substantially improved.
Figure 2 shows a result of the 2.5D analysis. Significant subsidences are seen at the Tsugaru Plain and the Echigo Plain (Fig. 2-a, b). These are thought to indicate ground subsidence. On the other hand, displacements dominant in the east-west direction and along the topography are seen at Mt. Hakkoda and Mt. Gassan (Fig. 2-c, d). These may imply landslides. Noted that tectonic deformation is removed through the GNSS correction.
We will discuss the details such as temporal transition of the detected displacements by this analysis.
Acknowledgments
ALOS-2 data were provided based on the joint research agreement with JAXA and under a cooperative research contract between GSI and JAXA. The ownership of ALOS-2 data belongs to JAXA. The numerical weather model was provided by JMA based on the agreement. Some of GNSS data were provided by JMA and NIED.