特定サイトにおける強震動予測 PREDICTION OF SITE SPECIFIC GROUND MOTION FOR LARGE EARTHQUAKE
Prediction of strong ground motion at a site near a fault is an important subject for the earthquake resistant design of a structure such as a reactor building. In this paper, we apply the semi-empirical synthesis method by IRIKURA (1983, 1986) to the estimation of site specific ground motion using accelerograms observed at Kumatori in Osaka prefecture. Target earthquakes used here are a comparatively distant earthquake (Δ = 95 km, M = 5.6) caused by the YAMASAKI fault and a near earthquake (Δ =27 km, M = 5.6). Then we discuss some problems to apply this synthesis method to the estimation of strong ground motions. The results obtained are as follows. 1) The accelerograms from the distant earthquake (M = 5.6) are synthesized using the aftershock records (M = 4.3) for 1983 YAMASAKI fault earthquake whose source parameters have been obtained by other authors from the hypocentral distribution of the aftershocks. The resultant synthetic motions show a good agreement with the observed ones. 2) The synthesis for a near earthquake (M = 5.6, we call this target earthquake) are made using a small earthquake which occured in the neighborhood of the target earthquake. Here, we apply two methods for giving the parameters for synthesis. One method is to use the parameters of YAMASAKI fault earthquake which has the same magnitude as the target earthquake, and the other is to use the parameters obtained from several existing empirical formulas. The resultant synthetic motion with the former parameters shows a good agreement with the observed one, but that with the latter parameters does not show so good agreement. These results suggest that the scaling relation between moment and magnitude may vary regionally. Therefore, we should adopt the empirical formulas for synthesis parameters in the region where we predict the strong ground motions because the existing empirical formulas predict the average of regional variation. 3) We estimate the source parameters from the source spectra of several earthquakes which have been observed in this site. Consequently we find that the small earthquakes (M <4) as Green's functions should be carefully used because the stress drops are not constant. 4) We propose that we should designate not only the magnitudes but also seismic moments of the target earthquake and the small earthquake to apply this synthesis method to the prediction of strong ground motions. Additionally, it is desirable to examine the scaling law of the earthquake sequences ocurring near the site based on the earthquake observation in advance.
日本建築学会構造系論文報告集 (409), p11-25, 1990-03