Dynamic Earthquake Source Modeling and the Study of Slab Effects

Abstract

Listen

Translate article

In this Thesis, I report my Ph.D. research on two major issues that are devoted towards constructing more realistic earthquake source model using computational tools: (1) constructing physically consistent dynamic rupture models that include complexities in fault geometry as well as heterogeneous stress and frictional properties inferred from observations; (2) study the effect of subducting slab structure on earthquakes that occur inside it with a special focus on the teleseismic waveforms. Fault step over is one of the most important geometric complexities that control the propagation and arrest of earthquake ruptures. In Chapter 2, we study the role of seismogenic depth and background stress on physical limits of earthquake rupture across fault step overs. We conclude that the maximum step over distance that a rupture can jump is approximately proportional to seismogenic depth. We also conclude that the pre-stress conditions have a fundamental effect on step over jump distance while the critical nucleation size has a secondary effect. Seismic wave carries information of source as well as structures along the path it travels. It was found that seismic waves generated by shallow events in subduction zones whose ray path coincide with the down going slab structure display waveform complexities that feature multipathing. In Chapter 3, we study deep earthquakes whose depth phases sample the slab structure on their way up to the surface. Differential travel time sP-P analysis shows a systematic decrease of up to 5 seconds from Europe to Australia and then to Pacific which is indicative of a dipping high velocity layer above the source region. Finite-difference simulations showed that a slab shaped structure that follows the Benioff zone at shallow depth and steepens beyond 400 km produces a model that can account for the sP-P differential travel times of 5 seconds for oceanic paths. In Chapter 4, we design a slab operator that can be applied on the 1D synthetics to generate 2D synthetics with slab structure. We hope this operator can be used for generating more accurate Green's functions that could potentially serve earthquake source inversion. In Chapter 5, we design a dynamic rupture model of the Mw 7.8 Gorkha, Nepal earthquake. We employ a novel approach of integrating kinematic inversion results which provide low frequency stress distribution and stochastic high frequency stress motivated by earthquake cycle models and observations. By doing this, we are able to reproduce the observed frequency dependent rupture processes, in particular the concentration of high-frequency radiation in the down-dip part of the rupture. In Chapter 6, I report my on going work on the spectral element method based earthquake cycle simulator. Large scale earthquake cycle simulation with consideration of complicated velocity structure and fault geometry is a great challenge for numerical modeling. I tried to push forward this boundary by extending the existing spectral element earthquake cycle simulator to enable cycle simulations on bi-material faults. This chapter includes a benchmark test in 2D that demonstrates the correctness of this new algorithm and an application of this method on bi-material fault earthquake cycle modeling.