Optical sub-pixel matching and active tectonics

Thesis

In this thesis I use sub-pixel optical matching, Interferometric Synthetic Aperture Radar (InSAR), and Light Detection and Ranging (lidar) spatial geodetic observations to produce reliable 3D displacement fields caused by co-seismic events and reliable earthquake source models with slip distribution on fault planes.

I produce horizontal displacement maps for the 2005 Dabbahu segment, Afar using SPOT4 satellite images. By combining InSAR descending data and range offsets with optical sub-pixel I produced a vertical displacement map of the event. I attempted to perform the inversion of the dataset obtained by sub-pixel matching but I found that datasets are not well suited for the typical numerical inversion, and I fit data with direct dislocation modelling instead.

I identify biases and errors that arise from optical sub-pixel matching of satellite images using many horizontal datasets constructed using SPOT5 images for the El Mayor-Cucapah earthquake. I develop algorithms for removal of some of these biases from horizontal displacement maps. Using sub-pixel matching I asses the quality of several DEMs available to me for study of the El Mayor-Cucapah earthquake.

I developed a novel technique for producing vertical displacement maps caused by an earthquake by combining archived pre-event satellite images with post event acquired lidar. I use this technique to produce a vertical displacement map of the El Mayor-Cucapah earthquake.

I produce a source model of the El Mayor-Cucapah earthquake by inverting InSAR datasets using the method. After attempts to do joint inversion of InSAR and optical sub-pixel matching I developed the code to use Bayesian inversion instead, because its advantages when it comes to joint modelling of datasets. I sucessfully invert four InSAR datasets on seven fault planes using the Bayesian approach. I found that the results of the Bayesian inversion are very similar to the results of the optimization inversion.

Authors: Barišin, I

• DOI: 10.5287/ora-7q1mbvvd1
• Type of award: DPhil
• Level of award: Doctoral
• Awarding institution: University of Oxford
• Language: English