5:15 PM - 7:15 PM
[SSS11-P11] Estimation of group velocity using seismic interferometry of coda waves observed from strong ground motion records of local earthquakes
Keywords:coda waves, seismic interferometry, group velocity, Kanto Basin, MeSO-net
Accurate ground motion simulation is indispensable for predicting damage from earthquake disasters. Although geophysical subsurface models have been proposed in various locations to simulate ground motions, there is not enough data available for nationwide subsurface models, and many assumptions remain to be made. In estimating subsurface structural models, seismic interferometry has recently attracted much attention. Seismic interferometry can estimate the subsurface structure model more easily than active seismic surveys because it is based on long-term microtremors. However, since not many stations continuously measure microtremors, there is a limit to the amount of data that can be used for seismic interferometry. Therefore, the use of coda waves, which have similar properties to microtremors, has been considered, and there are several studies of applying seismic coda waves to cross-correlation functions (Campillo and Paul, 2003; Emoto et al., 2015). The purpose of this study is to investigate group velocity estimation by applying seismic interferometry to coda waves of local earthquake observed by MeSO-net, where continuous observations and strong ground motion records are also available.
In this study, group velocity estimation was examined for five conditions: coda wave range, long-period band spectrum, epicenter region, number of earthquakes, and distance between stations. We found that two conditions, the number of earthquakes and the distance between stations, significantly affect the group velocity estimation by seismic interferometry using coda waves of local earthquakes. The number of earthquakes is more than 100 with M4.5 or greater, and the distance between stations is about 10 km, then, the values were close to the theoretical group velocity for most of the stations. However, by refereeing Emoto et al. (2015), the coda wave range was set to 300 to 600 s for the strong ground motion records, so in order to examine the possibility of applying the method to strong ground motion records with a data length of approximately several minutes, we estimated the group velocity for 60 s in the coda wave range 120 to 180 s between stations with close distances between them. The group velocity was estimated for 60 seconds in the coda wave range of 120 to 180 seconds between sites with close distances between them. As a result, we were able to estimate the group velocity even with the coda wave range of 120 to 180 seconds for 60 seconds, which indicates that the strong ground motion records are feasible when the distance between sites is short. If it is possible to estimate group velocity by seismic interferometry using strong ground motion records, it will be possible to estimate group velocity not only by MeSO-net but also at other seismic networks in Japan. However, there are still some points where the accuracy of group velocity estimation is low, and it is a future issue to clarify other factors that cause the low accuracy of group velocity estimation.
In this study, group velocity estimation was examined for five conditions: coda wave range, long-period band spectrum, epicenter region, number of earthquakes, and distance between stations. We found that two conditions, the number of earthquakes and the distance between stations, significantly affect the group velocity estimation by seismic interferometry using coda waves of local earthquakes. The number of earthquakes is more than 100 with M4.5 or greater, and the distance between stations is about 10 km, then, the values were close to the theoretical group velocity for most of the stations. However, by refereeing Emoto et al. (2015), the coda wave range was set to 300 to 600 s for the strong ground motion records, so in order to examine the possibility of applying the method to strong ground motion records with a data length of approximately several minutes, we estimated the group velocity for 60 s in the coda wave range 120 to 180 s between stations with close distances between them. The group velocity was estimated for 60 seconds in the coda wave range of 120 to 180 seconds between sites with close distances between them. As a result, we were able to estimate the group velocity even with the coda wave range of 120 to 180 seconds for 60 seconds, which indicates that the strong ground motion records are feasible when the distance between sites is short. If it is possible to estimate group velocity by seismic interferometry using strong ground motion records, it will be possible to estimate group velocity not only by MeSO-net but also at other seismic networks in Japan. However, there are still some points where the accuracy of group velocity estimation is low, and it is a future issue to clarify other factors that cause the low accuracy of group velocity estimation.