Asako Iwaki

Seismic‐Hazard Analysis of Long‐Period Ground Motion of Megathrust Earthquakes in the Nankai Trough Based on 3D Finite‐Difference Simulation

The 3D wave propagation simulation is suitable for the long‐period ground‐motion hazard analysis, because the realistic seismic source and velocity structure models are applicable and ground motion at any point can be obtained. However, large‐scale simulations for various scenarios are needed to adopt uncertainty due to the source model of a large earthquake. Here we show our approach to overcome a problem of heavy computational expense. We improve computation performance of the 3D finite‐difference method by using discontinuous grids and by adapting to multigraphics processing units technique. Then, we simulate long‐period ground motion of many scenarios of a hypothetical megathrust earthquake in the Nankai trough. We use the characterized source models based on the “recipe” in the simulation. Finally, we apply the simulation results based on about 400 scenarios to a technical framework of the probabilistic seismic‐hazard analysis. The hazard maps of velocity response spectra for a given conditional exceedance probability show that the long‐period ground motion is more significant with respect to the underground structure than is the distance from the source area.

Validation of the Recipe for Broadband Ground‐Motion Simulations of Japanese Crustal Earthquakes

The robustness of broadband ground‐motion simulation can promote seismic‐hazard assessment. A broadband ground‐motion simulation technique called “the recipe” is used in the scenario earthquake shaking maps of the National Seismic Hazard Maps for Japan. The recipe represents a fault rupture based on a multiple asperity model referred to as the characterized source model. Broadband ground‐motion time histories on the engineering bedrock are computed by a hybrid approach of the 3D finite‐difference method and the stochastic Green’s function method for the long‐ (>1  s) and short‐period ( 70  km); such conditions are outside of the target range of the recipe. Simulations using a 1D velocity structure model were also examined. The simulation results for the 1D and 3D velocity structure models indicated that the 3D velocity structure models are important in reproducing PGV and the later phases with long duration, especially on deep sediment sites.

Spatial Distribution of Ground‐Motion Variability in Broadband Ground‐Motion SimulationsSpatial Distribution of Ground‐Motion Variability in Broadband Ground‐Motion Simulations

Ground-motion prediction for a scenario earthquake requires evaluation of both the average ground-motion level and ground-motion variability due to model uncertainties. This study aims to evaluate the ground-motion variability due to aleatory variability of the source parameters by modeling ground motion of the 2000 Tottori earthquake (strike-slip type) and the 2004 Chuetsu earthquake (reverse-fault type). The source models are based on the characterized source model by the “recipe” (HERP, 2016) with fault location, size, and geometry as given parameters. Aleatory variability for the three source parameters is considered: (1) asperity location, (2) rupture initiation point, and (3) seismic moment. Two asperities are randomly located on the fault with no overlapping. A rupture initiation point is chosen randomly from the 2 km grids on the fault. Seismic moment M0 is sampled from a normal distribution in which the mean value is given by the M0-S relation (S being the fault area) by Irikura and Miyake (2001) and mean+2σ equals to 2M0. Short-period level A, another important parameter in the characterized source model, is derived from A-M0 relation by Dan et al. (2001). Ground motion for each earthquake is simulated by a hybrid approach; 3D FDM (Aoi and Fujiwara, 1999) for long periods (> 1 s) and the stochastic Green’s function method (Dan and Sato, 1998) for short periods (< 1 s), using a set of 50 source models and a 3D velocity model of J-SHIS v2 (Fujiwara et al., 2012). For the 2004 Chuetsu earthquake, simulations using a simple 1D stratified velocity model are also conducted in order to exclude the effects of the complicated subsurface structure around the source area. From the ground-motion simulation results with 50 source models for each earthquake, standard deviation (SD) of ground-motion indexes, ln of 5% damped acceleration response (Sa), PGA, and PGV, are analyzed at 10 km interval mesh. Distance and azimuthal dependence of SD are observed; the characteristics of the spatial distribution of SD differ from short periods to long periods. It is also found that the spatial distribution of SD is largely distorted by the complicated subsurface velocity structure for the Chuetsu earthquake. As a step toward constructing a model of ground-motion variability in ground-motion prediction for a scenario earthquake, we attempt to fit the SD, each for strike-slip type and reverse-fault type, with a simple regression model using the fault distance and directivity parameters. Effects of variability in other source parameters, such as rupture velocity and source time function, should be studied in our future works. Modeling variabilities in such source parameters requires investigation in physicsor empirical-based criteria.