[SCG58-07] Long-lasting afterslip associated with the 2016 moderate earthquakes along the Chaman fault, Pakistan
キーワード：チャマン断層、アフタースリップ、干渉合成開口レーダー、センチネル1
The Chaman fault system forms a ~900 km long major sinistral strike-slip fault, and is a transform boundary between Indian and Eurasian plates in Pakistan and Afghanistan. Previous geological surveys find the relative velocity of the Indian plate to the Eurasian plate at about 40 mm/yr. However, no large earthquakes (M>7) have been observed along the fault (Lawrence, 1992). Low seismicity results from either a long recurrence interval or aseismic creeping (Ambrasseys and Bilham, 2003). Recent geodetic studies reveal a heterogeneous distribution of locked and creeping segments along the Chaman fault (Fattahi and Amelung, 2016; Barnhalt, 2016).Meanwhile, Furuya and Satyabala (2008) showed a long lasting and spreading afterslip following a moderate earthquake (M5.0) in 2005 on the fault , where no creeping signals were detected prior to the earthquake.
We focus on the two moderate earthquakes occurred on May 13 and July 10 in 2016 on the fault, which were the first M5-class events since the 2005 M5.0 event examined by Furuya and Satyabala (2008). Ascending 11 SAR data of C-band Sentinel-1 from 2015 to 2017 are used to generate 55 interferograms. To separate deformation signals from other temporally random errors, we perform InSAR time series analysis based on a linear least squares inversion, which translates the set of interferograms into a range change time series. We use Small Baseline Subset (SBAS) method, following Schmidt and Bürgmann (2003), to detect cumulative LOS displacements. Our results indicate long-lasting afterslip for both of the two events. In particular, the second event has a sharper phase discontinuity across the fault, whereas the LOS changes are smaller than the first event.
Based on the cumulative LOS changes derived by SBAS, we estimate fault source models in an elastic half-space. We use Triangular Dislocation Element (TDE) to build non-planar fault. Our model indicates cumulative slip distribution which expands(?)on the fault plane. The maximum slip for the first event increased from 0.12 m to 0.53 m at a depth of 4-5 km during the 293 days after the event, while that for the second event increased from 0.11 m to 0.26 m at a depth of 2 km during the 235 days after the event.
We focus on the two moderate earthquakes occurred on May 13 and July 10 in 2016 on the fault, which were the first M5-class events since the 2005 M5.0 event examined by Furuya and Satyabala (2008). Ascending 11 SAR data of C-band Sentinel-1 from 2015 to 2017 are used to generate 55 interferograms. To separate deformation signals from other temporally random errors, we perform InSAR time series analysis based on a linear least squares inversion, which translates the set of interferograms into a range change time series. We use Small Baseline Subset (SBAS) method, following Schmidt and Bürgmann (2003), to detect cumulative LOS displacements. Our results indicate long-lasting afterslip for both of the two events. In particular, the second event has a sharper phase discontinuity across the fault, whereas the LOS changes are smaller than the first event.
Based on the cumulative LOS changes derived by SBAS, we estimate fault source models in an elastic half-space. We use Triangular Dislocation Element (TDE) to build non-planar fault. Our model indicates cumulative slip distribution which expands(?)on the fault plane. The maximum slip for the first event increased from 0.12 m to 0.53 m at a depth of 4-5 km during the 293 days after the event, while that for the second event increased from 0.11 m to 0.26 m at a depth of 2 km during the 235 days after the event.