Multi-scale and multi-physics modeling of puzzling earthquake dynamics

Alice-Agnes Gabriel

Submitted August 21, 2019, SCEC Contribution #9920, 2019 SCEC Annual Meeting Talk on Tue 1330

Physics-based earthquake scenarios using modern numerical methods and hardware specific computational optimizations can shed light on the dynamics, and severity, of earthquake behavior. This is specifically useful in tectonic settings which are currently underrepresented in operational seismic or tsunami hazard assessment, as well as for bridging scales from subduction zones to geo-reservoirs. Additionally, exploiting expected exascale computing infrastructure will allow to go beyond scenario-driven approaches.

I will present recent examples of physics-driven interpretations that can be integrated synergistically with established data-driven efforts. A preferred model of the 2016 Mw 7.8 Kaikoura, New Zealand earthquake yields a rupture cascade across apparently weak crustal faults by combining observational constraints with mechanical viability. Bayesian dynamic source inversion of the 2016 Central Italy sequence allows subsequent analysis of the posterior samples to infer stable features of the results and their uncertainties. Modeling how faults slip requires, in general, numerical methods which are capable of spanning a large range of spatial and temporal scales. In addition, pronounced geometric and rheological complexity needs to be accounted for. However, initializing such models with self-consistent fault and surface geometry, fault stress and rheology, fluid pressures and subsurface lithology is challenging. This can be overcome in coupled frameworks such as developed in the ASCETE project connecting subduction dynamics over millions of years, seismic cycling and earthquake dynamics down to fractions of a second, as well as tsunami propagation and inundation.

Citation
Gabriel, A. (2019, 08). Multi-scale and multi-physics modeling of puzzling earthquake dynamics. Oral Presentation at 2019 SCEC Annual Meeting.


Related Projects & Working Groups
Stress and Deformation Over Time (SDOT)