Studying the finiteness of large earthquakes with higher order moments of the moment tensor

Abstract

Finite fault parameters of the large earthquakes can be obtained using kinematic finite fault models which consist of a collection of subfaults. Yet calculating each subfault individually costs time and the fault model needs to be defined a priori for the inversion process. Alternative representation of the source is defining it as a moment tensor density distribution. In this case higher order moments of the distribution can be calculated. Higher order moments can also be used to estimate first order finite-fault parameters and it has the advantage of having less number of unknowns. This characteristic would allow for more rapid fault parameter solutions. The aim of this thesis is to develop a measure on the approximation of finite fault models using higher order moments up to degree two. Synthetic seismograms for both finite fault sources and their higher order moment approximations are generated using infinite homogeneous isotropic medium to identify the similarities between the waveforms. The effects of the receiver azimuths and distances are investigated using using different frequency ranges. It is found that higher order moments can give an approximation of waveform broadness or pulse width rather than the overall shape of the waveforms. Higher order moments also improved the point source approximations at frequencies that are beyond the corner frequency of the event. Fault type, fault strike direction and receiver azimuth influence the higher order moment solutions while distance is an insignificant factor at least for the whole-space medium which is considered in this study.