Synthetic Seismicity Model for the Xianshuihe Fault, Southwestern China

(1)构造模式。利用全球定位系统资料研究了中国的形变规律。假定中国由几个刚性块体组成，我们确定了这些块体的位移速率、断层上的滑动亏损和内部变形。结果表明，由印度碰撞引起的川滇地块的位移速率为7.0~9.5 mm/a，与地质估计值9 mm/a一致。鲜水河断裂的滑动赤字明显小于两个刚性块体之间的相对位移速率，表明断裂实际上是锁定的。(2)实验室实验。测量了三轴压缩条件下含节理岩样脆性变形过程中的声发射(AE)。对于不含节理的岩样，当外加压应力较小时，声发射的震源分散，震源分布的分维较低。另一方面，在岩石样本w…的情况下在较低的压应力下，更多的节理沿节理集中分布。声发射团簇在关节弯曲附近产生。(3)鲜水河断裂地震活动的数值模拟。我们建立了一个合成地震活动模型，假定断层的摩擦应力遵循实验室推导的速率和状态相关的摩擦定律，并使用从地质和大地测量观测估计的沿断层的长期平均滑动速率。通过选择合适的模型参数，我们可以模拟出断裂活动活跃期的重现，重现间隔为几百年，与历史记录的地震活动相一致。破裂在断层弯曲或分支区域或具有速度增强摩擦性质的区域被阻止。被阻止的破裂通常在几年后重新激活。这是因为无地震滑动在速度加强区传播缓慢。这解释了地震活动沿断层的迁移。较少

(1)Tectonic Model. We investigated deformation pattern of China using GPS data. Assuming that China consists of several rigid blocks, we determined displacement rates, slip deficit on faults, and internal deformation of the blocks. The results show that the displacement rate of the Sichuan-Yunnan block, which is caused by the collision of India, is 7.0 to 9.5 mm/yr, consistent with the geologically estimated value of 9 mm/yr. The slip deficit along the Xianshuihe fault is significantly smaller than the relative displacement rate between the two rigid blocks, indicating that the fault is virtually locked. (2)Laboratory Experiment. We measured acoustic emission (AE) during brittle deformation process of rock samples with preexisting joints under tri-axial compression. For rock samples without joints, the hypocenters of AEs are scattered and the fractal dimension of the hypocenter distribution is low when the applied compressional stress is low. On the other, in the case of rock samples w … More ith joints, AEs are concentrated along the joints under low compressional stress. AE clusters are generated near joint bends. (3)Numerical Simulation of Seismicity along the Xianshuihe Fault. We develop a synthetic seismicity model, assuming that the frictional stress of the fault follows a laboratory-derived rate and state-dependent friction law and using the long-term average slip rate along the fault estimated from geologically and geodetic observations. By selecting appropriate model parameters, we can simulate recurrence of active periods of fault motion with a recurrence interval of a few hundred years, being consistent with historically recorded seismicity. Rupture is arrested at regions of fault bend or branch or regions with velocity-strengthening frictional property. The arrested rupture is often reactivated after several years. This is because aseismic sliding propagates slowly in the velocity-strengthening regions. This explains migration of seismic activity along the fault. Less