Numerical Simulation of Earthquake Dynamic Rupture along Curved/ Branched fault Segments

用边界积分方程法模拟了非平面断层上的地震动态破裂。我们发现，在模拟中，断层的几何形状、施加在断层上的剪应力和断层的本构关系对动态破裂的计算是重要的。基于这些信息，我们计算了2000年鸟取县西部地震、1891年野壁地震和1995年神户地震的动态破裂。对于鸟取县西部地震，假定已有的断层几何形状和均匀本构关系，我们估计了断层周围的应力场，其结果与应力张量反演法估计的结果一致。对于Nobi地震，人们一直在争论地震中埋藏的断层是否滑动。为了加速讨论，我们利用已有的断层几何进行了动态破裂模拟，发现隐伏的断层是否滑动取决于与相邻断层的连通性。最后，我们调查了神户地震。根据得到的断层几何形状，断裂不能双向传播。我们发现，要进行双侧破裂，就需要一个小的走滑断层就在Akashi鲑鱼的正下方，这是在Akashi桥用GPS数据进行反演分析所必需的。此外，利用鸟取县西部地震和2002年阿拉斯加德纳利地震的近断层地震图，我们发现滑动减弱的距离与大小有关。我们在2004年和2005年美国地球物理联合会秋季会议上组织了两次特别会议。我们在这些会议上总共有150多篇摘要。我们认识到，这是许多人感兴趣的热门话题之一。我们现在正在准备一本书，以出版会议上的主要问题。

We have conducted simulations of earthquake dynamic rupture on non-planar faults using boundary integral equation method. We found that in the simulation, fault geometry, shear stress applied on the fault and fault constitutive relation are important for the computation of dynamic rupture. Based on these information, we computed the dynamic ruptures of the 2000 western Tottori earthquake, 1891 Nobi earthquake, and 1995 Kobe earthquake. For the western Tottori earthquake, assuming the preexisting fault geometry and uniform constitutive relation, we estimated the stress field around the fault, which was found to be consistent with that estimated by the stress tensor inversion. For the Nobi earthquake there has been a debate whether the buried fault slipped during the earthquake or not. To accelerate the discussion, we conducted a dynamic rupture simulation using the preexisting fault geometry and found that whether or not the buried fault slipped depends on the connectivity to the adjacent fault. Finally, we investigated the Kobe earthquake. Using the fault geometry obtained, the rupture could not propagate bi-laterally. We found that to make a bi-lateral rupture, a small strike slip fault is required just beneath the Akashi trout, which was required by the inversion analysis with GPS data at Akashi bridge. In addition, using the near-fault seismograms observed at the western Tottori earthquake and the 2002 Denali, Alaska earthquake, we showed that the slip-weakening distance is size dependent.We organized two special sessions in 2004 and 2005 American Geophysical Union Fall Meeting. And we had more than 150 abstracts in total for these sessions. We recognized that this is one of the high-lighted topics that many people are interested in. We are now preparing a book to publish the major issues in the sessions.