5:15 PM - 7:15 PM
[SCG61-P15] Steady coupling along the Nankai trough and Median Tectonic Line (MTL) estimated from GNSS and InSAR observation
Keywords:Nankai trough, Median Tectonic Line, coupling, GNSS (Global Navigation Satellite System), InSAR (Interferometric Synthetic Aperture Radar), crustal deformation
The Median tectonic Line (MTL) is one of the longest active fault zones in Japan. The MTL has continued right lateral slip caused by the oblique subduction of the Philippine Sea plate from the latter Quaternary (Okada,1973). However, its coupling state and fault geometry have not been revealed in spatially detail. Especially, it is still under debate whether the MTL is north-dipping or vertical fault. While GNSS-based interseismic displacements were explained well by a north-dipping MTL fault model (Tabei et al. 2002; 2007), slip tendency becomes higher when assuming a vertical fault rather than north-dipping fault (Uchide et al. 2022). In this study, we estimated the steady coupling state along the Nankai trough and MTL and the MTL dip angle based on interseismic displacement, which was obtained by InSAR (Interferometric Synthetic Aperture Radar) as well as GNSS.
First, we conducted GNSS and InSAR time-series analysis to detect interseismic crustal deformation. In GNSS analysis, we fitted a function composed of linear, trigonometric, logarithmic and offset terms to GEONET daily coordinate (F5) from March 12, 2011, to December 31, 2018, which is less affected by long-term SSEs along the Nankai trough. We assumed the interseismic coupling effect is included in the linear term. Logarithmic terms explain the postseismic effect of the 2011 Tohoku-oki earthquake (Tobita 2016). In InSAR analysis, we first made InSAR images for all possible pairs using SAR data in 2 frames (Descending: 21-2930, 21-2940) obtained by Stripmap mode observation by ALOS-2. We corrected ionospheric phase advance (Gomba et al. 2016), neutral atmospheric delay (Kinoshita 2022) and the postseismic effect of the Tohoku-oki earthquake. After that, we obtained the line of sight (LOS) velocity field though SBAS based time-series analysis using LiCSBAS2 software (Morishita et al. 2020). Finally, we converted the GNSS and InSAR velocities to the Amurian plate (AM)-fixed ones. The RMSE value between GNSS and InSAR velocities was less than 1.8 mm/year, which indicated they were highly consistent. Velocity discontinuities were not seen along the MTL, which suggested the upper part of the MTL would be almost fully locked.
Next, we estimated the steady coupling for three cases. We tried explaining GNSS and InSAR observations with only the Nankai trough (case 1), with the Nankai trough and north-dipping MTL (case 2) or with the Nankai trough and vertical MTL (case 3). We used 185 GEONET stations located in Shikoku or Chugoku district, 9 GPS-A stations and InSAR velocities in eastern Shikoku to estimate the coupling distribution. We performed the fault slip inversion using Okada dislocation model (Okada 1992) assuming elastic half-space with prior information that slip is spatially smooth. The strength of smoothing was determined based on ABIC (Yabuki and Matsuura 1992). Here, we assumed the MTL dip angle is uniform both in strike and dip direction. The estimated coupling distributions along the Nankai trough were similar among three cases, which suggested that the MTL coupling state would less affect the Nankai trough coupling state. Although residual velocities were smaller in case 2 and 3 than case 1, the estimated coupling distributions along the MTL were a little bit unrealistic. Some segments of the MTL showed negative coupling ratio, but we thought they should have positive value because seismic events such as SSE have not reported along the MTL. Estimated MTL block motion rates were 0.93 and 2.1 mm/year in case 2 and 3, respectively, which were significantly smaller than previous researches estimation. Next, assuming the MTL fault plates are completely flat, we estimated the MTL dip angle as well as the smoothing strength based on ABIC (Fukahata and Wright 2008). ABIC value was insensitive to the MTL dip angle, so we couldn’t obtain the robust solution for the MTL dip angle.
First, we conducted GNSS and InSAR time-series analysis to detect interseismic crustal deformation. In GNSS analysis, we fitted a function composed of linear, trigonometric, logarithmic and offset terms to GEONET daily coordinate (F5) from March 12, 2011, to December 31, 2018, which is less affected by long-term SSEs along the Nankai trough. We assumed the interseismic coupling effect is included in the linear term. Logarithmic terms explain the postseismic effect of the 2011 Tohoku-oki earthquake (Tobita 2016). In InSAR analysis, we first made InSAR images for all possible pairs using SAR data in 2 frames (Descending: 21-2930, 21-2940) obtained by Stripmap mode observation by ALOS-2. We corrected ionospheric phase advance (Gomba et al. 2016), neutral atmospheric delay (Kinoshita 2022) and the postseismic effect of the Tohoku-oki earthquake. After that, we obtained the line of sight (LOS) velocity field though SBAS based time-series analysis using LiCSBAS2 software (Morishita et al. 2020). Finally, we converted the GNSS and InSAR velocities to the Amurian plate (AM)-fixed ones. The RMSE value between GNSS and InSAR velocities was less than 1.8 mm/year, which indicated they were highly consistent. Velocity discontinuities were not seen along the MTL, which suggested the upper part of the MTL would be almost fully locked.
Next, we estimated the steady coupling for three cases. We tried explaining GNSS and InSAR observations with only the Nankai trough (case 1), with the Nankai trough and north-dipping MTL (case 2) or with the Nankai trough and vertical MTL (case 3). We used 185 GEONET stations located in Shikoku or Chugoku district, 9 GPS-A stations and InSAR velocities in eastern Shikoku to estimate the coupling distribution. We performed the fault slip inversion using Okada dislocation model (Okada 1992) assuming elastic half-space with prior information that slip is spatially smooth. The strength of smoothing was determined based on ABIC (Yabuki and Matsuura 1992). Here, we assumed the MTL dip angle is uniform both in strike and dip direction. The estimated coupling distributions along the Nankai trough were similar among three cases, which suggested that the MTL coupling state would less affect the Nankai trough coupling state. Although residual velocities were smaller in case 2 and 3 than case 1, the estimated coupling distributions along the MTL were a little bit unrealistic. Some segments of the MTL showed negative coupling ratio, but we thought they should have positive value because seismic events such as SSE have not reported along the MTL. Estimated MTL block motion rates were 0.93 and 2.1 mm/year in case 2 and 3, respectively, which were significantly smaller than previous researches estimation. Next, assuming the MTL fault plates are completely flat, we estimated the MTL dip angle as well as the smoothing strength based on ABIC (Fukahata and Wright 2008). ABIC value was insensitive to the MTL dip angle, so we couldn’t obtain the robust solution for the MTL dip angle.