Kun-Shan Chen(陳錕山)and Zhi-Tian Wang(王志添)

(Center for Space and Remote Sensing Research,Central University,Taoyuan County 32001)

Capability of mapping surface deformation offered by satellite radar interferometry has been successfully demonstrated and app lied since early 90s,from short to long term smeasuring and tracing of deformation patterns up to centimeter scale,both qualitatively and quantitatively.There remain limits in mapping earthquake triggered deformations in temporal and spatial scales,however,as already indicated in case of Chi-Chi Earthquake.This report focuses on the analysis of p re-and post-seismic deformation patterns derived from series of ALOS-PALSAR data.A novel InSAR processing flow was applied.First,an automatic registration method starts with the estimation of initial offset between slave and master images,selection of tie points,and follow ed by a refined matching by means of spatial feature analysis and pyramidal matching.A newly developed noise filter was then applied to the interferograms before phase unwrapping.To ensure the accuracy and robustness of the processing chain,results between In SAR derived and those from leveling measurements over a test site in Taiwan was compared and found to be very satisfactory.Deformation pattern mapped by ALOS/PALSAR differential interferometry with Fine Beam Single(FBS)and Fine Beam Daul(FBD)data pairs along the Longmenshan fault zone(LMSFZ)about 400 km from Wenchuan to Beichuan,Sichuan.The co-seismic interferogram was generated with 42 scenes(FBS,3frames/track,total 7 tracks from Pass 470 to Pass 476,since 20 Jul.,2007 to 7 Feb.,2009),while the pre-seismic and post-seismic PALSAR interferogram with 21 scenes(Pass 475,both FB Sand FBD,since 20 Jul.,2007 to 7 Feb.,2009)on the Wenchuan epicenter area were produced.It is suggested from the preliminary analysis that the deformation did occur before earthquake and recovered elastically after the earthquake.In the future,these results should be confirmed by comparing with the ground survey and GPS observation data.