10:45 AM - 12:15 PM
[SCG53-P05] Feasibility study of wave gradiometry for real-time ground motion prediction
Keywords:real time ground motion prediction, seismic array analysis
We are developing a scheme to predict ground motion from current wavefield in real time. In the scheme, current wavefield, i.e., amplitude distribution and propagation direction of ground motion, is an important factor. Hoshiba and Aoki (2015) and Ogiso et al. (2018) estimated current wavefield using observed amplitude distribution, simulation of wave propagation, and optimal interpolation. In this way, propagation direction was implicitly estimated. In contrast, seismic array analyses enable us to estimate propagation direction directly. In this presentation, we focus on wave gradiometry (e.g., Langston, 2007a, b, c). The method seems to be able to estimate not only slowness vectors but also apparent geometrical spreading that is important for ground motion prediction. We conducted a feasibility study of the wave gradiometry using synthetic waveforms.
First, we assumed a point source at the hypocenter of the Northern Nagano earthquake (Mw6.3) on 22 November, 2014, then calculated synthetic wavefield using the OpenSWPC (Maeda et al., 2017). The station network consisted of strong motion stations of the K-NET, KiK-net, and MeSo-net operated by the National Research Institute of Earth Science and Disaster Resilience (NIED), stations of the Japan Meteorological Agency (JMA) networks, and stations operated by the local governments.
When we applied the wave gradiometry to the synthetic waveforms calculated with a simple 1-D structure, we success fully estimated amplitude distribution and slowness vectors at the area where stations were densely distributed. However, apparent geometrical spreading terms showed the increase of amplitude at some regions, which was not expected from the synthesized wavefield. Langston (2007b) proposed to estimate apparent geometrical spreading terms after the correction of slowness terms. From a viewpoint of calculation in real time, the correction of slowness terms should increase the computational cost. We will keep on investigating the feasibility of wave gradiometry for not only estimating wavefield features but also computational cost.
First, we assumed a point source at the hypocenter of the Northern Nagano earthquake (Mw6.3) on 22 November, 2014, then calculated synthetic wavefield using the OpenSWPC (Maeda et al., 2017). The station network consisted of strong motion stations of the K-NET, KiK-net, and MeSo-net operated by the National Research Institute of Earth Science and Disaster Resilience (NIED), stations of the Japan Meteorological Agency (JMA) networks, and stations operated by the local governments.
When we applied the wave gradiometry to the synthetic waveforms calculated with a simple 1-D structure, we success fully estimated amplitude distribution and slowness vectors at the area where stations were densely distributed. However, apparent geometrical spreading terms showed the increase of amplitude at some regions, which was not expected from the synthesized wavefield. Langston (2007b) proposed to estimate apparent geometrical spreading terms after the correction of slowness terms. From a viewpoint of calculation in real time, the correction of slowness terms should increase the computational cost. We will keep on investigating the feasibility of wave gradiometry for not only estimating wavefield features but also computational cost.