Geometrically Accurate Earthquake Modeling

Abstract

Earthquakes are one of the great natural hazards. However, only a small portion of the earthquake cycle is captured in the modern observational record. Hence, understanding the physics and statistics of earthquakes requires sophisticated modeling to combine and interpret the data we do have. This thesis studies the full geometrically complex earthquake cycle in contrast to previous work that has assumed simple, flat representations of faults and the Earth. I use new powerful and efficient boundary element software and methododology to re-interpret the interseismic and coseismic geodetic record in the Longmen Shan and in Southern California. In addition, I demonstrate the complexity of fault behavior in the first geometrically accurate three-dimensional earthquake cycle models of the Cascadia subduction zone. By improving our ability to model faults as they exist in nature, we will begin to link together the myriad observational records to understand the physics of fault behavior and eventually, the current state of the earthquake cycle.