UTokyo Repository 東京大学

UTokyo Repository >
131 地震研究所 >
東京大学地震研究所彙報 >

このページ(論文)をリンクする場合は次のURLを使用してください: http://hdl.handle.net/2261/10605

タイトル: 21. 地震動の尾部にこついて(其の1)
その他のタイトル: 21. On the Coda Wares of the Earthquake Motions(Part 1)
著者: 表, 俊一郎
著者(別言語): Omote, Syun'itiro
発行日: 1944年8月15日
出版者: 東京帝国大学地震研究所
掲載誌情報: 東京帝国大学地震研究所彙報. 第21冊第3/4号, 1944.8.15, pp.458-500
抄録: The earthquake motion due to a distant origin consists generally of a series of different sections or stages. F. Omori divided the motion into three main parts, namely, "the preliminary tremor," "the principal portion," and "the end portion" purely on the bases of its appearance. The preliminary tremor, which consists principally of vibrations of small amplitudes and short periods, is divided into two parts, the first part bing a series of P waves and the sceond that of S waves. The principal portion, or that of "the large waves" as some researchers call this section of the motion is considerded to consist mostly of surface waves, since the travel time of the waves in it is approximately proportional to the arcual distance from the epicenter to the station where the observation is made. The large waves of the principal portion are usually followed up by a feeble finishing part of the earthquake motion, which is called the end portion or the coda waves. The oscillations of the coda waves frequently last three or more hours with an almost uniform amplitude and period. Both the P and S waves of the preliminary tremor and the large waves of the principal portion have been exhaustively studied by numerous investigators, but as yet little is known even to day, on the coda waves. In Chapter I of this report have been studied the oscillation periods of the coda waves which are seen in the seismograms of near earthquakes. In the case of near earthquakes whose epicentral distances are less than 1000km, the mode of oscillations somewhat differs from that of distant earthquakes. Records of near earthquakes are generally characterized by the great complexity of form and the relatively short duration of the entire motion. Such records begin with P waves of relatively small amplitudes and short periods and then proceed to S waves of large amplitudes. These are sometimes followed by large surface waves, though sometimes they are followed up by none at all. After them come the coda waves with an almost uniform amplitude and period. In studying the poriod of coda waves, the records taken by means of the Omori seismograph of the horizontal pendulum type have been used, the free oscillation period of the seismograph being throughout kept up at about 60 seconds. Figs.5~36 show the frequency distribution of different periods of coda waves as recorded in the seismogram. In preparing these figures, the relative frequency N_p has been used, which is given by the formula N_p={n_p/n_p'}×10, n_p'=60/(p+0.5') where n_p and p denote the frequency and period of the wave whose period is p second, and N_p the relative frefquency of that period. The peaks of the diagrams in Figs.5~36 are not sharp enough to enable one to determine the predominant periods from these figures. In order to determine more exactly the predominant period of the coda waves the mean period T_0 of every one minute has been determined by the formula T_0=(Σt_im_i)/(Σm_t) where t_i denotes all the periods that is seen duringthe one minite from the beginning of the (i-1)'th minute to the beginning of the i'ch minute after the commencement of the P waves, and m_i=t_i/60, T_0 the mean period of that one minute. In other words, the value of the weighted mean of all the periods that are found in any one minute has been defined as the mean period of that one minute. In this the length of the respective period itself has been taken as the weight of the respective period. The mean period of every one minute of coda waves has been determined with 29 earthquakes, the results of which are shown in Table III and in Figs.40-69. These figures show that the mean period curve of any earthquake can be re presented by a straight line parallel to the abscissa. This means that with any earthquake the mean period of every one minute of coda waves is almost uniform. This constant period has been defined as the predominant period of the coda waves of that earthquake. The determined predominant periods of the coda waves of the twenty-nine near earthquakes are tabulated in Table TV. In the map of Fig.70. are given the epicenters of these earthquakes together with the predominant periods of the coda waves of the respective earthquakes. It will be seen from these figures that the epicentral distance is an important factor in determining the period of coda waves. The nearer the epicenters are, the shorter are the predominant periods; the greater the distance, the longer the periods. In the Teganuma earthquake of August 21, 1938, the nearest, of the 29 earthquakes (Δ=33km), the coda waves had a predominant period of 5.0 seconds, while the Kusiro earthquake of (Δ=893km), September 11, 1938, the most distant of the 29 the coda period was found to be 7.4 seconds. The curve in Fig.71 shows the relation between the predominant period and the epicental distance. In Chapter II is studied the coda period of the Sanriku earthquake which occurred on March 3rd, 1936, an earthquake whose epicenter was located in the ocean depth off the Sanriku coast. The waves proceeding from it were recorded at the ten stations, that had about that time been temporarily established in Tokyo and Yokohama. Since each of these stations had been equipped with two horizontal components of Ishimoto's acceleration seismograhps of the same type, very reliable records could be obtained. The displacements of the earthquake motions have been computed from the actual acceleration seismograms, and the twice numerical integrations had been laboured out according to Simpson's rule with the original acceleration curves. The displacement curves obtained in this method are shown in Figs.64, 65. From these curves the predominant periods of coda waves as observed by the respective stations have been calculated. The determined predominant periods are tabulated in Table IX. From this it will be clear that the periods of coda waves are not dependent on the conditions near the observing stations i.e., the nature of the ground on which the station stands. The periods of coda waves of the Sanriku earthquake were determined to be 7.36±0.02sec in Tokyo districts and 7.32±0.02sec. in Yokohama. As the epicentral distances of Tokyo and Yokohama only differs 30km, it is but natural that the coda periods should be almost equal.
URI: http://hdl.handle.net/2261/10605
ISSN: 00408972


ファイル 記述 サイズフォーマット
ji0213010.pdf2.49 MBAdobe PDF見る/開く



Valid XHTML 1.0! DSpace Software Copyright © 2002-2010  Duraspace - ご意見をお寄せください