UC Davis

Interferometric synthetic-aperture radar (InSAR) interferograms provide maps of the surface deformation of the earth. Hidden in these deformation maps is valuable information about the fault systems hidden beneath the surface of the earth. InSAR interferograms are commonly used in inverse problems to determine the characteristics of earthquake sources from the deformation that they caused, which tell us about the faults upon which the earthquake occurred. These inversions typically use models composed of rectangular fault planes, but many studies suggest that faults are much more complicated. In this thesis, we present a new method for the inversion of InSAR interferograms using a model composed of a distribution of seismic point sources. We argue that point sources provide a much greater degree of flexibility to the model at less computational cost compared to rectangular sources.

We start our analysis with an interferogram containing the ground deformation caused by the 2015 Mw 7.8 Gorkha earthquake. After defining a point source model, we explore several different methods of performing the inversion for the source parameters: a genetic algorithm, a multiple linear regression and a nonlinear least-squares solver. During our analysis, we describe the positive and negative aspects of each inversion method, as well as possible interpretations of the results. We conclude our analysis with a case study of the 2019 Mw 7.1 Ridgecrest earthquake mainshock, which serves to expand the preceding analysis to fully explore the capabilities of each method using ground deformation in multiple directions. To extend our analysis of the point source distributions from our inversions, we calculate their fractal dimensions and compare them to the dimensions of the aftershock distribution and fault traces.