Analysis of the 2018 Hualien earthquake (Taiwan) by using SAR interferometry and pixel offset techniques

Chih Heng Lu, Fabio Bovenga, Jiun Yee Yen, Chung Pai Chang

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review


The local magnitude (ML) scale 6.22 earthquake struck eastern Taiwan on 6 February 2018. The epicenter was on the coastline near Hualien and caused a maximum shaking intensity of 7 in Hualien City, which was the most severely affected area. The crushed buildings and main surface cracks are distributed along the Milun fault, which is a oblique-slip fault. The earthquake is the largest of a sequence of events that affected the area over a period of days. In order to identify the surface deformation from Milun fault to coastal mountain range of Taiwan, we processed C-band Sentinel-1 and L-band ALOS-2 interferometric pairs acquired across the earthquake date, from both ascending and descending orbits. The Generic Mapping Tools Synthetic Aperture Radar (GMTSAR), and SNAPHU tools were used. According to the interferometric line-of-sight (LOS) displacement maps, the northwestern Milun fault moved of about 40 cm in the descending geometry, and about 0-10 cm in the ascending one, while the southwestern part of the fault moved 15-20 cm in the ascending geometry and is stable in the descending one. By combining interferometric LOS displacements derived from both ascending and descending acquisition geometries, just E-W and vertical displacement components can be derived. However, GPS and field survey results show northwards movements ranging from 50 and 70 cm over the Milun tableland, which is located on the eastern side of Milun fault. In order to provide SAR-based displacement measurements along north-south direction, we adopted the pixel-off set technique. The offsets along the azimuth direction were derived by computing the maximum of the amplitude cross-correlation of two SAR images, by using estimation windows of 128 x 128 pixel size. The 3D displacement was then computed by combining through least-square method the interferometric LOS displacements from ascending and descending acquisition geometry, and azimuth displacement component derived though pixel offset. Results indicate that, in the northern part of Milun tableland the maximum of E-W displacement reaches 60 cm, and that the whole Milun tableland moves northwards (up to about 60 cm) and upwards (up to about 40 cm). Moreover, the western side of Milun fault shows westwards movement, while Hualien city is affected by subsidence. The correlation coefficients in three directions between GPS data and 3D displacement are 0.81 (E-W), 0.89 (N-S), and 0.91 (U-D).

Original languageEnglish
Title of host publicationActive and Passive Microwave Remote Sensing for Environmental Monitoring II
EditorsEmanuele Santi, Claudia Notarnicola, Fabio Bovenga, Nazzareno Pierdicca
ISBN (Electronic)9781510621596
StatePublished - 2018
EventActive and Passive Microwave Remote Sensing for Environmental Monitoring II 2018 - Berlin, Germany
Duration: 12 Sep 201813 Sep 2018

Publication series

NameProceedings of SPIE - The International Society for Optical Engineering
ISSN (Print)0277-786X
ISSN (Electronic)1996-756X


ConferenceActive and Passive Microwave Remote Sensing for Environmental Monitoring II 2018


  • 3D displacement
  • DInSAR
  • Hualien earthquake
  • Pixel offset


Dive into the research topics of 'Analysis of the 2018 Hualien earthquake (Taiwan) by using SAR interferometry and pixel offset techniques'. Together they form a unique fingerprint.

Cite this