[S-SS11] Strong Ground Motion and Earthquake Disaster

[SSS11-09] On applicabiltiy of ground-motion simulation data to data-driven ground-motion prediction model

*Asako Iwaki¹, Hiroyuki Fujiwara¹, Nobuyuki Morikawa¹, Takahiro Maeda¹ (1.National Research Institute for Earth Science and Disaster Prevention)

Keywords:strong-motion prediction, seismic hazard assessment, ground-motion prediction equation, ground-motion simulation

Ground-motion simulation based on fault and wave-propagation models, often called physics-based simulation (PBS), has become a powerful tool for ground-motion prediction that can consider the effects of rupture propagation and wave propagation through complicated tectonics. On the other hand, ground-motion prediction equation (GMPE aka. GMM) is the most practical tool in probabilistic seismic hazard assessment (PSHA).

Our group has started a project to construct strong-motion database and data-driven GMPEs (Fujiwara et al. and Morikawa et al., 2021, this meeting), aiming to consider uncertainties in GMPEs to improve current PSHA in Japan. It has been pointed out that current GMPEs are not sufficient accurate at predicting ground motion with low probability, such as near-fault ground motion and ground motion from mega-earthquakes.

In this study, we attempt to merge PBS data with observation data in order to make up for the deficiency of the observation database especially in short distances and large magnitudes. We generated PBS data for scenario active fault earthquakes in Kanto area and performed statistical analysis to ensure the PBS data provide stable ground motion level and appropriate variability of ground motion.