Frontiers in Source Modeling for Near-Source Ground-Motion Prediction

Abstract

Accurate prediction of the intensity and variability of strong ground motions for future large earthquakes depends on our ability to simulate realistic earthquake source models. While there has been considerable progress in characterizing the complexity of earthquake ruptures, recent devastating earthquakes have exhibited rather unexpected behavior. Moderate-size events occurred with surprisingly large ground motions, in contrast to very large ruptures that showed relatively low ground motions in the frequency range of 0.1 – 3.0 sec. These observations are at odds with standard ground-motion attenuation relationships, and fundamentally challenge current strong-motion prediction methods. A related issue is the question about the upper limits of near-source ground motions. These topics can only be reconciled by considering the complexity of earthquake faulting and the dynamic processes of rupture nucleation, propagation and arrest. In this paper we discuss recent improvements to generate physically consistent earthquake rupture models for strong-motion simulation. We combine simulated slip distributions (realization of spatial random fields, consistent in scaling and spatial variability with slip distributions of past earthquakes) with constraints on the rupture nucleation and energy budget of earthquake rupture. Efficient stress-drop calculations in the spectral domain serve as input to estimate the temporal evolution of the rupture process through a set of empirical relationships derived from the analysis of spontaneous dynamic rupture models. Long-term earthquake-cycle simulations with realistic variability in rate-and-state friction parameters provide constraints on the generation, long-term behavior and characterization of earthquake source complexity for fault zones at different evolutionary stages. The pseudo-dynamic source characterization is inherently kinematic, but emulates the most important characteristics of dynamic rupture. While the relationships between dynamic source parameters are simplifications of the true complexity in rupture physics, they help identify the interaction between source properties that are relevant for strong ground motion prediction, and provide an improvement over purely kinematic models.