Committee Members Information

• Ph.D. Advisor: Prof. Paul Martin Mai
• External Examiner: Prof. Henriette Sudhaus
• Committee Chair: Prof. Thomas Finkbeiner
• 4th Committee Member: Prof. Sigurjón Jónsson

Abstract

Understanding the earthquake rupture process is critical for constraining seismic hazard and advancing knowledge of fault behavior. Earthquake rupture may range from relatively simple to inherently complex, particularly for large earthquakes, with complexity governed by fault geometry and heterogeneous stress conditions, making its estimation by inversion challenging. Inversion of rupture parameters from surface observations remains an ill-posed problem, strongly dependent on noise in the data, observational resolution, unknown subsurface properties, and modeling assumptions. This dissertation addresses these challenges by integrating an advanced Bayesian framework for finite fault inference, back projection imaging, a quantitative comparison technique, and seismic network optimization to improve the resolution and reliability of rupture imaging. The first study investigates the 2021 Mw7.4 Maduo earthquake in Qinghai, China, which ruptured a previously unmapped fault. Bayesian finite-fault inversion of teleseismic and geodetic data, complemented by back-projection imaging, resolves a bilaterally propagating rupture with asymmetric complexity. Comparison across published slip models shows that differences in assumed fault geometry and inversion strategy produce substantial variability in slip models and Coulomb stress transfer, underscoring how model uncertainties may influence interpretations. The second study examines the 2023 Mw7.8 and Mw7.6 Türkiye earthquake doublet. A Bayesian inversion of static GPS, InSAR, strong-motion, and teleseismic data reveals rupture across a geometrically complex, multi-segment fault. The Mw7.8 event ruptured bilaterally with sustained subshear velocities, while the Mw7.6 event transitioned between subshear and supershear. Variability among single-dataset and joint inversions highlights limited data resolution and emphasizes the value of integrating diverse observations. This study further demonstrates that geometric barriers control rupture termination and complexity. The third study explores seismic monitoring strategies for the Gulf of Aqaba, where station deployment is restricted to the Saudi Arabian margin. Synthetic tests show that station distribution and azimuthal coverage are more critical than array shape or density. Practical line arrays positioned at scaled distances relative to rupture length can adequately capture rupture parameters under one-sided observational constraints. Together, these studies demonstrate the value of uncertainty quantification for reliable rupture characterization, highlight the sensitivity of stress-transfer estimates to slip-model variability, and establish principles for network design in regions of limited access.