Inferring fault rupture dynamics from strong motion data

An inversion method is presented that combines a spontaneous rupture code and a linear programming algorithm to image earthquake rupture dynamics consistent with ground motion data. The spontaneous rupture code is converted into an inversion algorithm by imposing the slip required by the data as a boundary condition to the solution of the elastodynamics equation.; This inversion procedure is used to investigate the resolution of parameters of a simple slip-weakening friction law: strength excess, and the slip distance over which the stress is released, D_c, that may be obtained from the analysis of strong-ground motions. Band-limited waveform inversion cannot uniquely resolve both strength excess and D _c. This nonuniqueness derives from the trade-off between strength excess and D_c in controlling rupture velocity. However, fracture energy, G_c, might be relatively stably estimated from waveform inversions. If the stress drop is fixed by the slip distribution, the rupture velocity is controlled by fracture energy.; The dynamic rupture of the 1995 Kobe earthquake is investigated to determine its consistency with laboratory-derived rate and state variable friction laws. The Dieterich-Ruina friction law, with values of d_c = 1–5 cm for critical slip displacement, fits the stress change time series well. This range of d_c is 10–20 times smaller than D_c. The fracture energy for the 1995 Kobe earthquake is G_c ≈ 1.5 × 106 Jm−2, in agreement with estimates of previous earthquakes.; The effect of low absolute stress on earthquake dynamics is considered. The main difference between the dynamics of high and low stress events is the amount of coseismic temporal rake rotations occurring at given points on the fault. Temporal rake rotations are low stress phenomena and allow to infer the initial stress (Spudich, 1992) before the earthquake. Inference of low initial stress for the 1995 Kobe earthquake is presented.; Strong motion seismograms are calculated with dynamic models of hypothetical earthquakes consistent with the statistical properties of slip found in finite-source models of past events. These simulations show that including complexity in the rupture process consistent with dynamic rupture models has the potential to improve current strong ground motion prediction procedures.