17:15 〜 19:15
[SGD03-P07] Creep characteristics along the Philippines Fault on Leyte Island from ALOS-2 time-series analysis
キーワード：Philippines Fault, crustal deformation, fault creep, slip modeling, InSAR
Observations indicate that slip characteristics differ across various segments of active faults. Some fault sections may be seismically locked, whereas others may exhibit creep. As a result, the earthquake hazards in certain sections of the fault system could be comparatively higher or lower than in others. By analyzing slip distributions along active faults, we can create more accurate fault models for the region and enhance the development of effective earthquake disaster mitigation strategies.
The detection and detailed analysis of creeping faults on land have advanced significantly due to the increased number of observations made possible by synthetic aperture radar interferometry (InSAR). This study presents results on slip rate distributions along the Philippine Fault on Leyte Island, using descending data sets from ALOS-2/PALSAR-2 collected between March 2016 and July 2023. We processed six different ALOS-2/PALSAR-2 Stripmap SM3 frames using a small baseline subset algorithm for InSAR time series analysis (DeNTiS package). The radar observation geometry of the descending path was well-suited for measuring displacements along the northwestern-southeastern oriented left-lateral strike-slip fault. In order to reduce the phase decorrelation noise, we used a multilook of 12 by 22 in range and azimuth, respectively, corresponding to approximately 90-meter pixel resolution. We also performed a loop closure analysis (Biggs et al., 2007; Morishita et al., 2020) to detect possible unwrapping errors in the interferometric network. Additionally, in order to reduce the contribution of ionospheric noise, we employed split-spectrum analysis (Wegmüller et al., 2018), and for tropospheric delays, we applied the correction term from the GACOS product (Yu, Penna, and Li 2017; Yu, Li, and Penna 2018; Yu et al. 2018). Using GPS data, we also corrected InSAR mean line-of-sight velocities by removing the bilinear ramp between InSAR and GPS mean velocity fields.
We utilized the corrected mean line-of-sight InSAR velocities to estimate surface creep rates using the method adopted to derive the surface creep rates of 2006 – 2011 using ALOS data (Fukushima et al., 2019). The method consists of forming pairs of 2 km length (fault normal direction) and 1 km wide (fault strike direction) rectangles separated by 500 meters on each side of the fault and calculating the velocity difference between two rectangles facing each other. We also performed aseismic slip rate inversion using a fault model previously adopted by Dianala et al. (2020) for the Philippine Fault. We compared the results of the ALOS-2 period and previously published surface creep rates between October 2006 and January 2011 using ALOS-1 data and offset cultural features (roads, buildings, and curbs) by Tsutsumi and Perez 2024. Our results showed variable surface creep rates of about 20-60 mm/yr, excluding the region affected by the 6 July 2017 Leyte Earthquake. From the aseismic slip rate inversion, similar creep rate distribution of the shallowest portion less than 1 km depth was obtained. In addition, we spotted two possible regions with shallow slip deficits in the northernmost and central parts of Leyte Island. At the northernmost part, a shallow (1 km) depth Mw4.7 earthquake with ~8-km-long surface rupture reaching up to 2 cm of displacements precisely on the surface trace of the Philippine fault occurred on 15 July 2023. The shallow slip deficit, detected in our analysis, may have resulted in the earthquake.
The detection and detailed analysis of creeping faults on land have advanced significantly due to the increased number of observations made possible by synthetic aperture radar interferometry (InSAR). This study presents results on slip rate distributions along the Philippine Fault on Leyte Island, using descending data sets from ALOS-2/PALSAR-2 collected between March 2016 and July 2023. We processed six different ALOS-2/PALSAR-2 Stripmap SM3 frames using a small baseline subset algorithm for InSAR time series analysis (DeNTiS package). The radar observation geometry of the descending path was well-suited for measuring displacements along the northwestern-southeastern oriented left-lateral strike-slip fault. In order to reduce the phase decorrelation noise, we used a multilook of 12 by 22 in range and azimuth, respectively, corresponding to approximately 90-meter pixel resolution. We also performed a loop closure analysis (Biggs et al., 2007; Morishita et al., 2020) to detect possible unwrapping errors in the interferometric network. Additionally, in order to reduce the contribution of ionospheric noise, we employed split-spectrum analysis (Wegmüller et al., 2018), and for tropospheric delays, we applied the correction term from the GACOS product (Yu, Penna, and Li 2017; Yu, Li, and Penna 2018; Yu et al. 2018). Using GPS data, we also corrected InSAR mean line-of-sight velocities by removing the bilinear ramp between InSAR and GPS mean velocity fields.
We utilized the corrected mean line-of-sight InSAR velocities to estimate surface creep rates using the method adopted to derive the surface creep rates of 2006 – 2011 using ALOS data (Fukushima et al., 2019). The method consists of forming pairs of 2 km length (fault normal direction) and 1 km wide (fault strike direction) rectangles separated by 500 meters on each side of the fault and calculating the velocity difference between two rectangles facing each other. We also performed aseismic slip rate inversion using a fault model previously adopted by Dianala et al. (2020) for the Philippine Fault. We compared the results of the ALOS-2 period and previously published surface creep rates between October 2006 and January 2011 using ALOS-1 data and offset cultural features (roads, buildings, and curbs) by Tsutsumi and Perez 2024. Our results showed variable surface creep rates of about 20-60 mm/yr, excluding the region affected by the 6 July 2017 Leyte Earthquake. From the aseismic slip rate inversion, similar creep rate distribution of the shallowest portion less than 1 km depth was obtained. In addition, we spotted two possible regions with shallow slip deficits in the northernmost and central parts of Leyte Island. At the northernmost part, a shallow (1 km) depth Mw4.7 earthquake with ~8-km-long surface rupture reaching up to 2 cm of displacements precisely on the surface trace of the Philippine fault occurred on 15 July 2023. The shallow slip deficit, detected in our analysis, may have resulted in the earthquake.