Keywords: Differential Interferometry, Synthetic Aperture Radar, Surface Deformation, Earthquake
AbstractWe present a mapping of the surface deformation resulting from the 921 earthquake in Central Range of Taiwan using data acquired from ERS/2 high-resolution synthetic aperture radar(SAR) imagery.
SAR data sets acquired in January 21, May 6, September 23 and October 28 1999 were processed by differential interferometric technique to generate interferograms from which the deformation is assessed. The coherence between these pairs of interferograms were evaluated for quality check. In addition, the atmosphere effects on the coherence were reduced using sereral sets of interferograms. It is shown that the deformation derived from the interferograms are compared well with GPS measurements by Central Geological Survey (CGS) with interferometric baseline even more than 200m. Results also indicate that the subsidence more than 10cm on the west side of fault area in Taichung is detectable by differential interferometric SAR. It is concluded in this study that the InSAR technique provides a promising tool for detecting crustal deformation by earthquake
1.Assistant Researcher of CSRSR, National Central University
2.Technician of CSRSR, National Central University
3.Professor of CSRSR, National Central University
4.Professor of CSRSR, National Central University
1.Introduction
The purpose of this paper is to utilize the technique of interferometric synthetic aperture radar (IFSAR) to extract information about the crustal deformation after earthquake using the phase content of the complex SAR images, and to detect the land subsidence and crustal movement in Taiwan. SAR not only obtains the radar scattering coefficient of the land cover, but also carries the phase information about the target when a pair of data is conherently processed. The phase has two major parts: the character of the land cover itself and the distance between radar and the target. A single phase echoes essentially bears no information. Therefore, we must have at least one image pair with the same measuring direction and near the same measuring position to examine the difference of echos. If the character of the land cover changes little during the period of images taken, we can ignore the influence of phase caused by it, and then get the radar Interferometric map, produced by the difference of the measuring distance. We then calculate the difference of the distance by using the Interferometric image pairs, and furthermore we use the differential interferogram map with the same area but at different time to calculate the height and the crustal movement. On the development of the relative technology and application, many scholars abroad had researched in this aspect, and the result can be found in many academic periodical. Decorrection in interferometric radar echos has been demonstrated by Howard A. Zebker in 1992 [2]. S. N. Madsen uses airborne synthetic aperture radar interferometry to produce topographic mapping by NASA DC-8 plane in 1993 [3]; and he also presented that accuracy of topographic maps derived from ERS-1 interferometric radar in 1994 [4]. R. M. Goldsten announced satellite radar interferometry: two-dimensional phase unwrapping in 1988 [5]. D. J. Bone published Fourier fringe analysis: the two-dimensional phase unwrapping problem in 1991 [6]. While the application of surface displacement, A. K. Gabtiel ever presented mapping small elevation changes over large areas: differential radar interferometry in 1989 [7]. Massonnet used a map by radar interferometry to check the displacement field of the Landers earthquake in 1993 [8]. Zebker also used differential radar interferometry to derive coseismic displacement in 1994 [9]. Carnec explained two examples of the useage of SAR interferometry on displacement fields of small spatial extent in 1996 [10].
According a report of Central Weather Bureau ( CWB ), a major earthquake
( M
L =7.3 ) has occurred near the small town Chi-Chi in the Nantou County, Taiwan, at 1:47 a.m. local time on September 21, 1999 ( 17:47 GMT on 20 September ). The epicenter was 12.5 km northwest the Sun-Moon lake. The focal depth is 8 km. The death toll has exceeded 2100 and is mounting during the first week after the main shock. The Chi-Chi earthquake essentially was resulted from a major reverse fault, the Chelungpu Fault, trending N-S to NNE-SSW, one of the major trust faults in the deformation front of the fold-and-thrust belt of the Taiwan orogenic belt. The movement of the earthquake produced a linear zone of surface ruptures, extending about 80 km long from north to south. According to the in-situ t field investigations, the surface break shows large vertical as well as horizontal displacements, ranged from 1 to 8 meters, along the fault line as a consequence of the Chi-Chi earthquake.
When we reviewed the epicenter distribution during this century, we will get a conclusion that there is a main earthquake zone, which extends to the southeast direction from HouLong ( in MiaoLi County), passing through MiaoLi city, SanYi, ZhuoLan, DongShi, PuLi, SunMoon Lake, and then turning to the southwest direction, passing through ZhuShan, MeiShan, JiaYi, XinYing, Tainan, and forming a wide arc curve seismic zone. This arc curve seismic zone surrounds roughly a stable foundation, which hides in shallow underground of BeiGang, TaiXi, and PengHu. This zone is called BeiGang highlands; it is exactly located in the core area of the inner side in the arc curve seismic zone; relatively, the epicenter of the area distributes fewer, and the earthquake also happens fewer. When the Philippine plate subducts towards the North beneath the Eurasian plate, BeiGang highlands hinder Taiwan island from moving west, as a result, there are some activities of trust faults and earthquakes in the front of mountains; this is the reason that most active faults distribute in the front of mountains[1]. This paper reports the analyzed results for the case of 921 Great Earthquake using differential interferometric SAR technique. Qualitative and quantitative assessments were discussed and compared to field GPS survey.
2. Data Description
In respect of data sources, there are two ways to acquire interferometric SAR pairs: dual pass and dual antenna. Currently, satellite SAR belongs to dual pass, while airborne system such as JPL AIRSAR is kind of dual antenna. This study makes use of ERS SAR data. ERS1 and ERS2 are both earth source satellite, developed by European Space Agency (ESA). They pass over Taiwan at 10:28 am and 10:20 pm. The Center for Space and Remote Sensing Research at National Central University has received and archived quite a large volume of ERS1/2 radar image data since 1993m, providing a very valuable database for interferogram measurement in Taiwan.
We choose the central mountainous area of Taiwan for testing. The ERS satellite orbit no.232, frame 3123 (Figure. 1) was selected to form the image pairs according on the base line term for interferometry provided by ESA. Then, based on the date for the 921 Great Earthquake, data of 21 Jan?6 May?23 Sep and 28 Oct 1999 were singled out for use. The perpendicular and horizontal components of the base lines of these data sets is given in Table 1.
Figure. 1 : Coverage map of ERS2 over Taiwan
Table 1. : the image pair of ERS2 interferometry
| The date of the master image |
The date of the slave image |
The interval of the date |
Perpendicular base line |
Horizontal base line |
| 99/05/06 |
99/01/21 |
105 |
-96m |
-91m |
| 99/10/28 |
99/09/23 |
35 |
219m |
91m |
| 99/05/06 |
99/09/23 |
140 |
213m |
54m |
| 99/05/06 |
99/10/28 |
175 |
-6m |
37m |
