Project Abstract
|
We propose a numerical study of rupture at bimaterial interfaces in 3D. Preliminary results from our recent finite difference 3D simulations motivate this study. These have revealed a previously unrecognized (so far as we know) mechanism for exciting the preferred-direction unilateral rupture on a bimaterial interface. This mechanism is purely 3D in nature, and applies
specifically to strike-slip ruptures that are very long compared with the seismogenic thickness. Therefore this excitation mechanism has not shown up previously in 2D simulations, nor in 3D simulations with fault length equal to a relatively small multiple of the seismogenic thickness. Excitation of the unilateral rupture mode occurs by this mechanism even with an elementary slipdependent friction law, and more realistic friction laws may reveal additional possibilities. Still open questions are whether the 3D bimaterial effect in this guise will lead to unilateral rupture over a significant range of prestress conditions, whether the unilateral mode will persist in the presence of reasonable along-fault material heterogeneities, and whether this effect can generate features (e.g., rise times, slip velocity functions) compatible with observational studies. Investigating these issues numerically requires 3D simulations on a large scale, because of the importance of the large fault aspect ratio, combined with the small scale lengths associated with
crack-tip normal stress perturbations. Large parameter studies on this scale have become feasible with the recent development of highly scalable 3D codes with well-quantified accuracy and convergence rates (e.g., Day et al, 2005), and their implementation on supercomputer clusters. The proposed SCEC DynaShake computational platform will further facilitate the proposed
numerical studies. |
