Depth Extent Analysis of the 1981 October 16 Chile Earthquake

概要

論文の詳細を見る
We present a method to retrieve the vertical rupture extent of a large shallow earthquake whose rupture mode is assumed to be simple. In this method, we calculate GREEN's functions of vertically distributed point sources which mimic upward or downward rupture propagation with a constant velocity. Using these GREEN's functions we deconvolve P-wave data to obtain source time functions for the assumed depth extent and rupture direction. First we estimate the maximum source duration for point sources by letting the length of the source time function be as long as that of the data. We then deconvolve the source time function limiting its maximum length to be shorter than the time window of the data but long enough to cover the maximum duration for the point sources. We take the depth extent and rupture direction which gives the minimum residual as the optimum solution. We test this method by deconvolving the synthetic data using kernels slightly different from those used for the synthetics. We find that the deconvolution is robust to the rupture velocity. However, the crustal structure near the source, in particular a water layer thickness different from that used for the synthetics, has a significant effect on the deconvolution. The deconvolution is moderately sensitive to the focal mechanism. We applied this method to a large outer-rise (Ms=7.2) earthquake which occurred west of the Chile Trench on October 16, 1981. We determined the water layer thickness to be 4.5km, which gives the smallest residuals. We obtained a 30 or 40km deep point source as an optimum for this event, using 10km grid spacing, i.e., a 10km depth interval between point sources. Using 5km grid spacing, we obtained a rupture from 37 to 32km depth; however, the difference between the residuals for different rupture modes is so small that this rupture extent does not seem meaningful. In order to obtain more information on the depth extent of this event, we conducted further synthetic data/deconvolution experiments. In the calculation of synthetics, we restricted the shape of the source time function to be an isosceles triangle, whose width is determined by matching the total source duration observed for point sources at 20-40km depth. We compared the source time functions deconvolved from the synthesized data with those obtained from the observed data. To explain the observed source time functions and residual patterns, a centroid depth at 35km and depth extent less than 15km are required. Upward rupture propagation is preferred. We consider the solution with rupture from 41 to 29km (12km depth extent with a centroid at 35km) to be the optimum solution for the 1981 Chile earthquake. This depth extent coincides with the lower half of the mechanical plate of 40Ma age, which is the age of the ocean floor at the epicenter, and is consistent with the compressional stress in the deeper portion of the plate associated with its bending prior to subduction beneath the Chilean coast.浅い大地震の断層の深さ方向の広がりを推定する方法を提案した.この方法では,有限の広がりは,深さ方向に等間隔で置かれた点震源の重ね合わせで表現される.これらの点震源は同じメカニズムを持ち,速度3km/secで上,あるいは下方向に伝播するものとした.様々な断層の拡がりに対しグリーン関数を計算し,それらを用いて観測波形のdeconvolutionを行い,震源時間関数を求める.観測波形と合成波形の残差が最小となる深さの拡がりを解とする.
東京大学地震研究所,Earthquake Research Institute, University of Tokyo,Institute of Geology and Mineralogy, Faculty of Science, Hiroshima Universityの論文
1990-06-30

著者

Seno T
Univ. Tokyo Tokyo Jpn

Depth Extent Analysis of the 1981 October 16 Chile Earthquake

スポンサーリンク

概要

著者

関連論文

スポンサーリンク