Yoshihiro Kaneko (UCSD)
Study of the earthquake source brings about a set of fascinating interdisciplinary problems characterized by nonlinearity, a broad range of spatial and temporal scales, rare but catastrophic events, competing physical mechanisms, remote observations, inverse problems, non-uniqueness, and substantial societal significance. The ultimate challenge is to understand and quantify factors controlling the spatio-temporal behavior of active faults, including earthquake nucleation, seismic patterns, and the interaction of seismic and aseismic fault slip. My research aims to address this challenge by developing realistic physical models of earthquake source over several seismic cycles that rely on recent dramatic advances in observations, computational resources, and laboratory experiments. The goal is to use the models in conjunction with seismic, geodetic, and geological observations to constrain earthquake-source properties in terms of experimentally-derived constitutive laws, and then to study the potential set of future behaviors. Here, I will present two examples that illustrate this approach. In the first one, numerical modeling is used to establish the relation between variations in fault friction properties, the pattern of interseismic coupling (which characterizes the degree of fault locking between seismic events), the properties of earthquake sequences, and the observable characteristics of individual seismic events. The second example presents an innovative method for inferring fault friction properties based on comparison of numerical simulations and geodetic observations, which is applied to the central section of the North Anatolian fault (Turkey).