2005 年三月六日凌晨三點零六分，在台灣東北部宜蘭平原下方，深度7.58 公里處發生了芮氏規模5.9 的淺層地震，而在68 秒後，在主震西方約5 公里處，深度6.22 公里處，另一起芮氏規模5.9 的地震隨之發生，根據台灣寬頻地震網(BATS)利用地震矩張量逆推所得之震源機制解與本研究中餘震分布分析可得，該斷層為東西走向約270 度向北傾之走向滑移斷層；但由強震紀錄之P 波初動解卻得出正斷層的震源機制，震源機制解存在相當大的差異性，在本研究中可變動滑移角逆所得結果，震源起始位置出現正斷層的破裂滑移向量，支持初動解之結果。由於此對地震發生的時間十分接近，震央距稍遠的測站，儀器所紀錄到的兩個地震波形將混雜在一起而難以分別，所幸圍繞震央約20 公里範圍內，有超過10 處由中央氣象局所建置之強震站，儘管地震發生的時間與空間的分布如此接近，由強震紀錄我們可以看出兩個地震源之間有著不同的破裂分布。研究中我們假設斷層模型為有限斷層(Finite fault)，則每一子斷層視為一點震源，再利用全波形理論格林函數計算加入震源機制後所得到的理論位移波形，與測站記錄之觀測波形分別於速度場與位移場進行逆推，可得斷層面上每個子斷層最符合觀測波形的滑移情況。分析兩個斷層面上逆推所得滑移量分布，可以發現第二起地震發生於第一地震破裂時，斷層面上較少滑移量分布的區域，而第二主震最大錯動量出現的區域，緊鄰著第一主震最大錯動量出現的位置，代表此區域在第一主震錯動時，應力並未完全釋放。可知兩者間之破裂滑移分布於時間與空間上有著相當緊密的關連性，值得深入探討。 Two earthquakes, which were 68 sec apart, with similar magnitudes and depths occurred beneath I-Lan Plain in northeastern Taiwan on March 6, 2005. The first event was located at longitude of 120.84o E, latitude of 24.65o N, and a depth of 6.4 km. The second event was located at longitude o120.80 o E, latitude of 24.65o N, and depth of 7.0 km. The spatial separation of these two events is only within several kilometers. The CMT solutions from BATS (Broadband Array in Taiwan for Seismology) and the distributions of aftershocks of these two events suggest these two earthquakes with similar focal mechanisms of left lateral-strike slip. It is interesting for the occurrence of the earthquake doublet. In this study, we try to explore the mechanism of seismic source triggering of these two events. Other studies on doublet have suggested that earthquake doublet often occurred in the region where has the relatively large and isolated high-stress. We considered the finite-fault model to obtain the spatial slip distributions of these two events. The fault was divided into several subfault. Each subfault is considered as a single source for Green’s function calcluation. from the nearby strong motion stations as close as 4km deployed by CWB were used for waveform inversion. A linear least-square procedure was applied to obtain the best dislocation model, which can explain the best observation waveform with optimum misfit to the observations. The spatial slip distribution of these two events show that the event2 was initiated in the region where dislocations was less during the rupture of event 1. Further studies on the correlation of these two events will be carried out to understand the possible dynamic behavior on earthquake rupture.
