17:15 〜 19:15
[STT40-P05] 17-years Land subsidence history obtained from ALOS, ALOS-2, and Sentinel-1 InSAR time series analysis at Jakarta, Indonesia
キーワード：InSAR、地盤沈下、SAR時系列解析
Land subsidence in Jakarta, Indonesia, is an ongoing environmental issue primarily driven by anthropogenic factors such as excessive groundwater extraction and natural processes such as soil compaction. Most previous studies tried to detect subsidence in Jakarta using only one SAR satellite system, such as Envisat, ALOS, and Sentinel-1. Now, we have an opportunity to use several SAR satellite observation archives, which enables us to investigate the subsidence history of more than tens of years. This study investigates the development of land subsidence in Jakarta, examining its spatial and temporal variations using multiple SAR satellite data from ALOS, ALOS-2, and Sentinel-1. Using these three satellite SARs enables us to investigate the history of land subsidence over 17 years.
For InSAR analysis of ALOS and ALOS-2, we used the GAMMA software. In ALOS, we used 20 stripmap mode SLCs with ascending orbit. For ALOS-2, we used 46 SM3 SLCs with descending orbit. Orbital and topographic fringes were modelled and removed using precise orbit information and the 1-arcsecond SRTM DEM. After phase filtering with the GW spectral filter, wrapped interferograms were unwrapped with the MCF method. Since ALOS and ALOS-2 SAR operate in L-band microwave sensitive to ionospheric disturbance, we applied the Split Spectrum Method to all ALOS and ALOS-2 InSAR images to mitigate ionospheric phase contamination. Then, we used the LiCSBAS software to perform the time series analysis. For Sentinel-1 SAR, we used LiCSAR products, including 243 ascending and 185 descending images for the LiCSBAS time series analysis. By using both ascending and descending observations from Sentinel-1, we performed the 2.5 dimensional (2.5D) analysis to estimate quasi-eastward and quasi-vertical displacement velocity.
The results showed that the ALOS time series detected subsidence signals in Jakarta, with the maximum LOS change reaching 27 mm/year in northern Jakarta. For ALOS-2, the maximum LOS velocity of approximately 10 mm/year was detected at the western boundary of Jakarta. The Sentinel-1 2.5D analysis showed that a maximum vertical velocity of 38.9 mm/year was detected at the western boundary of the city, the same area where the ALOS-2 also detected the maximum LOS velocity. We did not entangle the reason for the discrepancy between ALOS-2 and Sentinel-1 maximum velocities, we will continue to investigate it. Comparing the Sentinel-1 time series result with that of ALOS, the area of the maximum LOS velocity changed from the northern coastal area to the western boundary of Jakarta, although at the western boundary, both ALOS and Sentinel-1 detected significant LOS changes in the sense of subsidence. Due to the shorter wavelength of Sentinel-1 (C-band), interferometric coherence was lost in the northern coastal area, especially for the ascending result. The Sentinel-1 time series result also revealed that subsidence velocity has slowed in some regions in recent years, likely due to improved groundwater management practices and changing urban development patterns.
The research identifies the spatial and temporal subsidence patterns across Jakarta, showing that excessive groundwater use, especially in areas where the water table has significantly dropped, is a major contributor.
For InSAR analysis of ALOS and ALOS-2, we used the GAMMA software. In ALOS, we used 20 stripmap mode SLCs with ascending orbit. For ALOS-2, we used 46 SM3 SLCs with descending orbit. Orbital and topographic fringes were modelled and removed using precise orbit information and the 1-arcsecond SRTM DEM. After phase filtering with the GW spectral filter, wrapped interferograms were unwrapped with the MCF method. Since ALOS and ALOS-2 SAR operate in L-band microwave sensitive to ionospheric disturbance, we applied the Split Spectrum Method to all ALOS and ALOS-2 InSAR images to mitigate ionospheric phase contamination. Then, we used the LiCSBAS software to perform the time series analysis. For Sentinel-1 SAR, we used LiCSAR products, including 243 ascending and 185 descending images for the LiCSBAS time series analysis. By using both ascending and descending observations from Sentinel-1, we performed the 2.5 dimensional (2.5D) analysis to estimate quasi-eastward and quasi-vertical displacement velocity.
The results showed that the ALOS time series detected subsidence signals in Jakarta, with the maximum LOS change reaching 27 mm/year in northern Jakarta. For ALOS-2, the maximum LOS velocity of approximately 10 mm/year was detected at the western boundary of Jakarta. The Sentinel-1 2.5D analysis showed that a maximum vertical velocity of 38.9 mm/year was detected at the western boundary of the city, the same area where the ALOS-2 also detected the maximum LOS velocity. We did not entangle the reason for the discrepancy between ALOS-2 and Sentinel-1 maximum velocities, we will continue to investigate it. Comparing the Sentinel-1 time series result with that of ALOS, the area of the maximum LOS velocity changed from the northern coastal area to the western boundary of Jakarta, although at the western boundary, both ALOS and Sentinel-1 detected significant LOS changes in the sense of subsidence. Due to the shorter wavelength of Sentinel-1 (C-band), interferometric coherence was lost in the northern coastal area, especially for the ascending result. The Sentinel-1 time series result also revealed that subsidence velocity has slowed in some regions in recent years, likely due to improved groundwater management practices and changing urban development patterns.
The research identifies the spatial and temporal subsidence patterns across Jakarta, showing that excessive groundwater use, especially in areas where the water table has significantly dropped, is a major contributor.