断层导波研究加利福尼亚兰德斯和海克特曼恩地震断层带的四维特征(英文)

美国加利福尼亚州兰德斯和海克特曼恩地区于1992年和1999年先后发生7.4级和7.1级地震,分别在地面产生80km和40km长的断裂带。震后在断裂带布置的密集地震站台记录到明显的断层导波(fault-zone guided waves)。这些导波由断层带内的余震和人工震源激发产生,走时在S波之后,但具有比体波更强的振幅和更长的波列,并具有频散特征。通过对2～7 Hz断层导波的定量分析和三维有限差分数字模拟,获得了震深区断裂带的高分辨内部构造图像以及岩石的物理特性。数字模拟结果表明这些断裂带上存在被严重破碎了的核心层,形成低速、低Q值地震波导。核心破碎带宽约100～200 m,其内地震波波速降为周围岩石的40％～50％,Q值约为10～50。根据岩石断裂力学观点,这一低速、低Q值带可被解释为地震过程中处于断层动态断裂前端的非弹性区(或称之为破碎区,相干过程区)。在兰德斯和海克特曼恩断裂带测得的破碎区宽度与断裂带长度之比约为0.005,基本上符合岩石断裂力学预期的结果。观察到的断层导波还显示兰德斯和海克特曼恩地震中多条断层发生滑移和破碎。兰德斯地震时多条阶梯形断层相继断裂;而在海克特曼恩地震中,断裂带南北两端均出现分枝断裂,深处的分枝断裂较地表出现的破裂状况更为复杂。由三维有限元模拟的动态断裂过程表明,?

CHARACTERIZATION OF RUPTURE ZONES AT LANDERS AND HECTOR MINE,CALIFORNIA IN 4-D BY FAULT-ZONE GUIDED WAVES

Linear seismic arrays deployed across and along rupture zones of the 1992 M7.4 Landers and 1999 M7.1 Hector Mine, California, earthquakes have registered prominent fault-zone guided (trapped) waves generated by aftershocks and explosions within the rupture zone. Observations and 3-D finite-difference simulations of 2～7 Hz trapped waves allowed a high-resolution delineation of the internal structure of rupture zone and physical properties of fault-zone rock at seismogenic depths. The Landers and Hector Mine rupture zones are marked by a low velocity and low Q waveguide about 100 to 200 m wide, in which S velocities are reduced by 40%～50% from wall-rock velocities and Q values are 10～50. We interpret the trapped wave inferred low-velocity waveguide as a remnant of process zone (damage zone) where inelastic deformation occurs around the propagating crack tip in crustal rock during dynamic rupture in the earthquake. The width of fault zone waveguide likely scales to the rupture length as predicted in published dynamic rupture models. Trapped-wave delineated low-velocity waveguides show the multiple-fault rupture pattern in the Landers and Hector Mine earthquakes. The Landers rupture segmented at stepovers between faults. At Hector Mine, the northern rupture bifurcated at depth with a more complicated pattern than the surface breakage. The generic finite-element model for dynamic rupture shows that this bifurcation in the Hector Mine rupture is physically plausible and consistent with trapped wave observations. Repeated surveys using explosions detonated within the Landers and Hector Mine rupture zone revealed post-seismic fault healing with time. The shear wave velocities measured within the Landers rupture zone showed an increase of ～1.2% between 1994 and 1996, and further increase of ～0.7% between 1996 and 1998, indicating the fault strength recovery with rock rigidity increase after the mainshock. It is most likely due to the closure of crustal cracks that opened during the 1992 earthquake. The observed velocity increase is consistent with a decrease of ～0.03 in apparent crack density within the rupture zone between 1994 and 1998. The ratio of traveltime decreases for P to S waves changed from 0.75 to 0.65 in this period, suggesting that cracks near the fault became more fluid saturated with time after the mainshock. The repeated surveys at the Hector Mine rupture zone revealed an increase of 0.65%～1.0% in shear velocity between 2000 and 2001, showing that the fault healing occurred not only on the Landers rupture zone. However, the healing rate on the Hector Mine rupture zone varied from one fault to another. The greater damage in rock was inflicted, and thus greater healing is observed, in regions with larger slip in the mainshock. We also found that the healing process on the Landers rupture zone was interrupted by the 1999 M7.1 Hector Mine earthquake which occurred 25 km away. The Hector Mine quake added damage of fault-zone rock as a temporal reversal of the healing on Landers faults by strong shaking and permanent strain. The post-Landers trend of fault strengthening then recovered after the Hector Mine earthquake. We have watched a fault zone evolving the damage and healing with time. The knowledge of spatial and temporal patterns of Landers and Hector Mine rupture zones allowed us further understand the physics of earthquake on the active faults.