利用同震級震後GPS位移探討2003年12月成功地震的同震震後錯移模型以及其庫倫應力轉移情形

Abstract

花東縱谷為歐亞大陸板塊與菲律賓海板塊碰撞交界，為地質構造活躍、地震頻仍之處，本研究中的池上斷層即為南段花東縱谷與海岸山脈的邊界，跨越池上斷層每年約有1.8到3.5公分的變形量 ( Yu et al.,2001 )，大致以地震或潛移的方式釋放。 2003年12月10日芮氏規模6.5的成功地震，由地震重定位及餘震分佈的情形，我們可以得知成功地震為池上斷層錯動所造成，並造成地表上數公分至數十公分的地表位移。由於以往對於池上斷層地表下的實際幾何位態，並不是十分清楚，本研究希望藉由地表上GPS觀測系統所量測到的同震位移，配合彈性半無限空間錯位模型，來推求池上斷層最佳的斷層幾何形貌，以及模擬成功地震在池上斷層面上滑移量的分佈情形;當斷層位態與滑移量分佈為已知的情況之下，可以藉由庫倫應力轉移模型，看出主震對於斷層週遭應力影響的情形，評估成功地震對於附近已存在的破裂面是否有觸發可能性，藉以了解斷層的活動性及未來可能發生較大地震的區域。 彈性半無限空間錯位模型配合地表位移資料過去廣泛應用於尋找斷層面的幾何位態，以及斷層面上可能的滑移量分布 ( Okada, 1985; Johnson et al., 2001 )，來了解斷層的行為和地震之間的關係。

The Mw=6.8 Chengkung Earthquake is almost a pure thrust event which is occurred on December 10th, 2003 in the Costal Range of eastern Taiwan. This earthquake is believed to rupture the NNE-striking Chihshang Fault in the Longitudinal Valley. Based on the relocation of aftershock sequences, the main shock mainly dislocates the east dipping high-angle Chihshang Fault plane. The maximum permanent vertical displacement shown by GPS data in the hanging wall side is about 30 cm. Remarkably, not only all the stations on the hanging wall of the Chihshang Fault were uplifted in this earthquake event, but also the stations on the Longitudinal Valley were still raised till the foothills of the Central Range. We employ GPS data around the Chihshang Fault by using the elastic half-space dislocation model to figure out the fault plane geometry and the distribution of co-seismic dislocations. I construct the Chihshang Fault, consisting of 6 segments, by the delineation of the aftershock distribution. The segments all strike N˚18E, because of the fault trace on the surface, and the segments of Chihshang fault, from top to bottom, are 60˚, 65˚, 60˚, 40˚, 20˚, 10˚, respectively, revealing a listric fault type. The dislocation model in term of GPS data reveals that the maximal dislocation is about 1 m along dip-slip and the dislocations gradually decrease to 10 cm near the surface. The average slip is 0.52 m along the fault surface, which yields a scalar moment of 2.0 × 1026 dyne-cm. These results are similar to that of the Harvard CMT, indicating a scalar moment of 2.0 × 1026 dyne-cm. The result shows the uplift of the foot wall highly affects the root mean square of vertical component. From the post-seismic displacements recorded by GPS 3 months after the main shock, the hanging wall of Chihshang fault still uplifted during this period. Contrast to the hanging wall, the footwall moved downward. Under the assumption that the postseismic deformation was due to the afterslip on the same fault, the dislocation model is also applied to investigate the postseismic afterslip distribution. The result shows the maximum afterslip is up to 15 cm and most slip are distributed to upper part of the fault. This phenomenon probably implies that coseismic dislocation near the fault trace may be partially locked or damped, thus the strain cumulated and caused the uplift of the footwall. After the mainshock, the strain released near the surface causes aseismic creeping and significant post-slip. To inspect the stress field after the Chengkeng earthquake, Coulomb failure criterion is employed to investigate the occurrence of aftershocks and future rupturing, based on the coseismic dislocation and fault geometry. Coulomb stress change, on specified orientation and optimally orientated faults, both reveal that the footwall should experience a significant stress increase up to 2 bars, a candidate for the occurrence of aftershocks. The comparison between the Coulomb stress change pattern and aftershock distribution shows there should be a pre-existing geological structure beneath the Longitudinal Valley and was triggered after the Chengkung earthquake.