Fast Land Deformation and Damage Assessment After the 2018 Hualien Earthquake in Taiwan Using Multi-Source Remote Sensing Data
Abstract
A Magnitude-6.4 (on the moment magnitude) earthquake hit the east coast city of Hualien, Taiwan on February 6, 2018. The earthquake along with its 11 foreshocks of Magnitude-4 or greater starting from February 3, 2018 caused 17 casualties and injured more 280 people, in addition to server damages to many buildings and infrastructures such as bridges and highways. After the earthquake, multiple remotely sensed data from different platforms and with assorted characteristics were utilized to investigate and assess the land deformation and damages of the affected areas. Two sets of Differential Interferometric Synthetic Aperture Radar (DInSAR) analysis results from four Sentinel-1 images reveal that a foreshock (with a Magnitude of 6.1) on February 4 has already caused a slight (-7 mm to 7 mm) LOS (line of sight) displacement along the Milun fault. The main shock on February 6 further increased the displacement to -150 mm to 150 mm along the satellite line of sight. Comparing the DInSAR results with collected reports, most of the collapsed buildings and ground raptures were located at the west of the fault where the subsidence was most significant.
In addition to the DInSAR analyses for land deformation estimation, high resolution and very high resolution satellite images were also used to investigate the damages in the built-up areas and infrastructures. The satellite images provided not only an overview of the affected regions but also detailed damages of some buildings and structures. In some locations, unmanned aerial vehicles (UAV) were deployed to collect more detailed information. The collected UAV data and cloud-source photographs uploaded by residents and others on-site were then processed using structure-from-motion (SfM) and dense matching algorithms to generate three-dimensional (3D) point clouds and digital surface models of collapsed or damaged buildings for damage assessment and decision support for crucial tasks such as search-and-rescue, evacuation, demolition and the like. The examples exhibited in this presentation demonstrate that multi-source remote sensing is effective for investigating and assessing the aftermath of natural hazards as well as to provide useful information for hazard mitigation.- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2019
- Bibcode:
- 2019AGUFMNH31F0912T
- Keywords:
-
- 4355 Miscellaneous;
- NATURAL HAZARDS