K-Netデータから抽出した地震動の統計的性質とそれを用いた波形合成用グリーン関数の生成

概要

論文の詳細を見る
We extract statistical characteristics of seismic ground motions from K-net records of 25 earthquakes observed at 142 stations in the Kyushu region in order to establish a synthesis method of statistical Green's functions. For the statistical envelope characteristics first we express them by Boore's envelope function and identify its model parameters using the least square method. Then we express these model parameters as a function of the magnitude M and the hypocentral distance X using two step regression analysis. For the statistical spectral characteristics we separate source, path, and site effects from the observed Fourier spectra and express them also as a function of the magnitude M and the hypocentral distance X. Synthetic waveforms based on these estimated statistical characteristics match quite well with the observed records. Thus we can use these statistical Green's functions to predict strong motions for future large earthquakes through the so-called semi-empirical method.
2001-05-30