On the Spatial Correlation of Earthquake Source Parameters
SCHMEDES, J., email@example.com; ARCHULETA, R.J., firstname.lastname@example.org; LAVALLÉE, D., Institute for Crustal Studies, UC Santa Barbara, Santa Barbara, CA 93106, email@example.com
To predict seismic ground motions an accurate method for their computation is needed. The basis of such a method is an understanding of the physical source parameters that control the level of ground-shaking. A correlation between rupture velocity and slip amplitude on the fault was observed in different studies. Based on these observations a kinematic source description based on the spatial correlation of slip amplitude and average rupture velocity, and slip amplitude and rise time on the fault, was constructed. To get a more physical based understanding of the spatial correlation between earthquake source parameters and their amplitude distributions, dynamic modeling of the earthquake rupture process is necessary. To get good statistics, many dynamic models have to be computed. The basis of such computations is the choice of realistic initial models for stress and strength. Several models have been proposed in the literature to describe the spatial correlation ‚Äìor the power spectrum- associated with the slip or the stress spatial variability. In these models, the power spectrum is attenuated according to a power law with exponent . In this study, we consider two values: =2 (short correlation length) and =4 (long correlation length). Furthermore, we use amplitudes distributed according to a Gauss or a Cauchy distribution to construct different sets of initial models. Of special interest in this study is the resulting amplitude distribution of the rupture velocity, and the correlation between rupture velocity and other earthquake source parameters for the different sets of initial conditions used. We address the question under what conditions supershear rupture occurs. Furthermore, we analyze the effect of different rupture velocity distributions on the resulting ground motion. We implement the resulting distributions for rupture velocity and its correlation with slip into the kinematic description and compute broadband ground motion for each distribution.