HR: 08:45h
AN: S51B-04    [PDF]
TI: Non-planar Fault Geometry and Stress Field around the Fault
AU: * Fukuyama, E
EM: fuku@bosai.go.jp
AF: Nat'l Res. Inst. Earth Sci. Disas. Prev., 3-1 Tennodai, Tsukuba, 305-0006 Japan
AB: The distribution of asperities (i.e.large slip area during an earthquake) should be correlated with its initial shear stress distribution (or possible stress drops during the earthquake). If the surrounding stress field around the fault is known and rather uniform, initial shear stress distribution on the fault could be controlled by its fault geometry. This suggests that in order to detect the distribution of asperity prior to the earthquake, estimation of initial tectonic stress field applied to the fault becomes important. In this study, we examine if the final slip distribution can constrain the initial stress field when the constitutive relation on the fault and fault geometry are provided. We used a boundary integral equation method for triangular elements, which allows us to model an arbitrary shaped fault system with curvatures, branches and jogs. First, using a fault model with a complicated geometry and homogeneous constitutive law, we tried to invert the final slip distribution for the principal initial stress direction and its stress ratio (R) by a grid search method. This numerical test suggests that the principal stress direction can be estimated within an error of 5 degrees, but stress ratio R cannot be well constrained. Then we applied this technique to the 2000 western Tottori earthquake (Mw 6.6). The principal stress direction of the surrounding stress has already been estimated using a stress tensor inversion of aftershock moment tensors. The fault geometry is also estimated using relocated aftershock distribution by double difference method. The constitutive relations are also estimated from the slip rate function estimated by a waveform inversion. Using this model, we conducted a forward modeling to estimate the dynamic rupture history. We found that the principal stress direction obtained by the stress tensor inversion is the easiest angle for the dynamic rupture to propagate and we think this initial stress distribution is constrained by the fault geometry of this earthquake.
DE: 3210 Modeling
DE: 5104 Fracture and flow
DE: 7209 Earthquake dynamics and mechanics
DE: 7260 Theory and modeling
DE: 8164 Stresses--crust and lithosphere
SC: Seismology [S]
MN: 2003 Fall Meeting