|
AirSAR/MASTER
|
A Study of Surface Deformation From Earthquake by Differential Radar Interferometry
4. Results and Discussions
From the coherence maps, there is no interference fringe in the sea due to the lack of coherence. In plain area including town and the suburb area, the vertical component of the base line is about 6 meters with average coherence of above 0.2, in spite of the longest time interval of 175 days ( between 6 May and 28 Oct ). Similarly, for the time interval of 140 days between May 6 and Sep 23, the vertical component of the base line is about 213 meters with average coherence of above 0.2. These may be acceptable for purpose of interferometric detection. By differential interferometry we further get a series of subsidence maps in Taichung city. A maximum scale of about 18 cm after earthquake was estimated, which compare fairly well with GPS measurement [1].
To make sense of the comparison, we carefully investigated and checked the GPS measurement along fault broken zone. From the great earthquake resulted in the obvious bulge along the Chelungpu fault, we calculated the measurement of the fault broken zone which has both historical meaning and practical value that leads the Central Geological Survey(CGS) to give every effort to survey the fault belt ( zone ). In order to make sure the position of the surface rupture of fault, CGS has confirmed the position of outcrop by in-situ field measurements. Trimble 4000 SSi receiver was used in the survey of the fault broken line(zone) by means of fast static survey method, with one or two the GPS stake set by CGS as reference station. Processing along the fault zone by 4 or 5 moving GPS station was conducted with receiving signals sampled at every 5 seconds for 8 to 15 minutes. Depending on the field of view of the GPS station, each critical outcrop had been measured except in mountainous area where it was difficult, if not impossible, to be tracked. After data taken, the base line precision was confirmed to be under 10 cm [10] for the distance in between 3 and 60 kilometers from reference station.
The GPS network measurement shows that between the north Tunzijiao fault and Yuanli it arises 4 to 8 cm in vertical displacement component; it descended 11 to 22 cm in the west of ChangHua fault. Among the east of ChungHua fault, the west of Chelungpu fault, and the south of Caotun, a 18 to 22 cm decending was observed. And in the north of Caotun, there was a 8 to 18 cm descending, while between the east of Chelungpu fault and Shuangdong fault, it arises 1.5 to 4 meters. It was even greater near the Chelungpu fault; in the east part of Shuangdong fault appears about 1 meter arising; and the position of geography center of Taiwan moved 2.3 meters in northwest direction with descending of 63 cm in vertical
To this end, it may be recalled that the differential interferometry is measuring movemen in centimeter scale. When the vertical displacement is too large such as those analyzed above, the interferometric coherence was almost lost, making the deformation detection very difficult. The Chelungpu fault, for example, presented a displacement amount to meters. C band system like ERS is not appropriate. We must consider the radar system of the L band like JERS-1 having longer wavelength.
5. Summary
Some conclusions are reached based the case study in this paper:
- The most severe land subsidence of Taichung city was estimated to be about 18 cm. Comparison with the GPS surveying data confirms that the InSAR and GPS results have in the same sinking tendency, suggesting that the differential interferometry has great potential for detecting the displacement in large area.
- If the displacement is too large in vertical part, the coherence of the image pairs is too low, then we can't detect possible surface deformation. A system with longer wavelength might be a solution. Besides, the phase noise due to the vegetation cover can be reduced.
- The influence of atmospheric and ionospheric effects in interferogram should be taken into account. One solution is using GPS network observations to remove such effetcs. Another way is to increase the numbers of monitoring. In this paper, the ERSs' data are all in descending mode, the change in distance being in slant range. Thus, it is not possible to separate the horizontal and perpendicular component. If we acquire the data in ascending mode data in the future, we may analyze the components of change.
- The complete detecting task of long-term and large area of surface deformation depends greatly on continuity of the image data receiving. Also, a large volume of SAR data archive is essential for long term displacement monitoring as well as short term displacement by natural disaster such as earthquake. The GPS survey is good for a point by point, but not cost effective for extensive area. The interferometric SAR technique and GPS are complementary and supplementary to each other.
- The ground resolution of ERS satellite series is 20 meters now, and its incident angle is fixed to 23°, so image of mountainous area has many layovers and shadows. We should have choice for higher resolution and flexible incident angle, such as Radersat-1 or Radersat-2, to minimize the amount of layovers and shadows.
Acknowledgement
We would like to thank Dr. Kuo-Fong Ma ( Institute of Geophysics, National Central University, Chung-Li, Taiwan, ROC ), Dr. Shui-Bei Yu ( Institute of Earth Science, Academia Sinica, Nankang, Taiwan, ROC ), the GPS surveying data after earthquake was supplied by the Central Geological Survey ( CGS ). We are also indebted to all the staff members of CSRSR who processed ERS and SPOT images used in this study.
6. References
- Zhao-Xia Chen(1999), the survey reports of 921 earthquake and geology, Central Geological Survey, Department of Economy, Taiwan.
- Goldsten, R.M., Zebker, H. A., and Werner, C. L.(1988),Satellite radar interferometry: Two-dimensional phase unwrapping,Radio Science,Vol. 23,No. 4,pp. 713-720.
- Gabtiel, A. K., Goldstein, R. M., and Zebker, H. A.(1989), Mapping small elevation changes over large areas: Differential radar Interferometry, j. Geophys. Res., Vol. 94, No. B7, pp. 9183-9191.
- Bone, D. J.(1991), Fourier fringe analysis: the two-dimensional phase unwrapping problem, Applied Optics, Vol. 30, No. 25, pp. 3627-3632.
- Zekber, H. A., and Villasenor, J.(1992), Decorrelation in interferometric radar echoes, IEEE Trans. Geosci. Remote Sensing, Vol.30., No. 5, pp.950-959.
- Madsen, S. N., Zebker, H. A., and Martin, J.(1993), Topographic mapping using radar interferometry: Processing techniques, IEEE Trans. Geosci. Remote Sensing, Vol. 31, No. 1, pp. 246-256.
- Massonnet, D., Rossi, M., Carmona, C., Adragna, F., Peitzer, G., Feigi, K., and Rabaute, T.(1993), The displacement field of the Landers earthquake mapped by radar interferometry, Nature, Vol. 364, No. 8, pp. 138-142.
- Zekber, H. A., Werner, C. L., Rosen, P. A., and Hensley, S.(1994), Accuracy of topographic maps derived from ERS-1 interferometric radar, IEEE Trans. Geosci. Remote Sensing, Vol. 32, No. 4, pp. 823-836.
- Zekber, H. A., Rosen, P. A., Goldstein,R. M., Gabriel,A., and Werner,C. L. (1994), On the derivation of coseismic displacement fields using differential radar interferometry: The Landers earthquake, J, Geophy., Vol. 99, No. B10, pp. 19.617-19.634.
- Carnec. C.,Massonnet D. and King C.(1996), Two examples of the use of SAR interferometry on displacement fields of small spatial extent, Geophy. Res., Vol. 23, No. 24, pp. 3579-3582.
- Hui-Zheng Zhang(1999), the survey reports of 921 earthquake and geology, Central Geological Survey, Department of Economy, Taiwan.
|
|
|
|
|
