Abstract

The 2004 Mw 6.0 Parkfield, California, earthquake took place in a very well‐instrumented area, producing a substantial quantity of high‐quality near‐field recordings. Taking advantage of the rare luxury of having a large number of near‐field ground‐motion recordings distributed around the fault zone and the availability of various slip models as well as an Earth structure model of the region, we study the effects of various kinematic rupture parameters to derive implications for strong ground motion simulation in engineering applications. We model the 3D wave propagation resulting from this earthquake using the 3D staggered‐grid finite‐difference method. Using a grid spacing of 100 m in our fourth‐order explicit finite‐difference code, we could properly resolve frequencies of up to 1 Hz with a minimum of eight grids per wavelength for shear waves, except in the immediate vicinity of the fault where fault‐trapped waves dominate the records. We assess the effects of various simulation parameters such as slip model, rise time (constant or variable), rupture velocity, and the earth model (1D versus 3D) on the resulting waveforms. We also investigate the distribution of engineering parameters such as peak ground velocities, peak ground displacements, and spectral accelerations at specific periods on the Earth’s surface. An outstanding feature is that at high frequencies fault‐normal components near the edge of fault segments dominate the ground‐motion field. Fault‐parallel components are dominated by lower frequencies. The difference between fault‐parallel and fault‐normal components is clearly observed in such engineering parameters as peak ground velocity and peak ground displacement.

You do not currently have access to this article.