Committee Members Information

• Ph.D. Advisor: Prof. Paul Martin Mai
• External Examiner: Prof. Fabrice Cotton
• Committee Chair: Prof. Frans Van Buchem
• 4th Committee Member: Prof. Sigurjón Jónsson
• 5th Committee Member: Prof. Olaf Zielke

Abstract

This dissertation develops and applies a physics-based framework for advancing probabilistic seismic hazard analysis in the Gulf of Aqaba, the southernmost extension of the Dead Sea Fault Zone. The region is tectonically active, has experienced large historical and instrumental earthquakes, and is now home to critical infrastructure and large-scale development projects such as NEOM. Yet, the shortness of the instrumental record, limited paleoseismic information, and sparse geophysical constraints present major challenges for traditional hazard approaches, which depend heavily on observational data and ergodic ground-motion models. To overcome these limitations, in this dissertation we use multi-cycle earthquake simulations using the MCQsim, to explore epistemic uncertainties in source characterization and to embed physically consistent rupture processes into hazard estimates. We develop different modelling realizations explicitly account for uncertainties in fault connectivity, segmentation, and seismogenic depth, thereby capturing the range of plausible rupture behaviours along the Gulf of Aqaba fault system. The resulting catalogues provide a robust basis for probabilistic hazard calculations, replacing statistical extrapolation from incomplete observations with physically motivated representations of long-term fault activity. To translate these catalogues into hazard estimates, a workflow was developed to integrate physics-based ruptures into the OpenQuake, a platform for hazard modelling. Two complementary pathways were pursued: (i) a stochastic (event-based) and (ii) a non-parametric implementation, in which individual ruptures from the synthetic seismic events are directly incorporated with their simulated occurrence rates. Moving beyond ergodic hazard analysis, the dissertation further integrates physics-based rupture with forward ground-motion simulations. Source time functions derived from MCQsim were used as input for AXITRA, a discrete wavenumber method to solve the seismic-wave equation in a layered velocity model. Our framework allowed the computation of synthetic waveforms up to 2 Hz and the derivation of ground-motion intensity measures, including PGA, PGV, and SA. By systematically varying frictional-strength properties, the simulations explicitly capture earthquake-to-earthquake variability, rupture directivity, and near-fault effects that are not well represented in empirical ground-motion models. Incorporating these simulations into hazard calculations represents a step toward partially non-ergodic frameworks, where both source processes and ground motions are described in a physically consistent manner rather than through purely statistical generalizations.