The analysis of the coda oscillations of earthquake motions has been carried out in Chapters 1 to 8 with respect to 54 earthquakes observed at Hongo, Tokyo, whose epicentral distances ranged from 33 km to almost 18,000 km. Some examples of seismograms used for the analysis of the coda oscillations will be seen in Fig. 10. In the first place, (1) frequency distribution diagrams of all the periods of coda waves found in every successive one minute were prepared. The peaks of such diagrams were not sharp enough to enable one to determine the predominant periods of coda oscillations. In the second place (2) the mean period of every one minute has been calculated. The mean period of each successive one minute was, so far as a single earthquake was concerned, fairely uniform throughout the whole duration of the coda oscillations, and this mean period agreed well with the predominant period determined from the frequency diagrams of the same earthquake. In the case of extremely distant earthquakes the coda oscillations were found to continue even after the large surface waves propagated along the major arc of the earth's surface had died away. This part of the coda waves has been named ""the second coda waves"", and in contrast to it the coda waves that precede the large waves propagated along the earth's surface have been called ""the first coda waves."" The predominant periods of the second coda waves were seen to be constantly 16 sec. with all the earthquakes. These were described in Chapters 1～4. Then (3) in Chapter 5, with a view to establishing more definitely the periodicity of coda oscillations, a method of periodgram analysis has been resorted to with eight selected earthquakes having different epicentral distances. The periodgrams of the second coda waves showed a single very high peak at about TA=16 sec. Those of the first coda waves showed two high peaks, one of which always appeared at a period that agreed well with the predominant period of the same earthquake determined by the method described in (1) and (2), while the other peak was seen at the period of about 20 sec. Lastly (4) in Chapter 6, for the purpose of making clear the oscillation period of coda waves as accurately as possible, an expanded Takahasi-Husimi's method of analysis has been applied for the study of the period of coda oscillations of ten selected earthquakes. We obtained an oscillation curve for each portion of the respective earthquakes-a composite curve made up of two sine curves of a damping nature. After elaborate calculations we decomposed each sine curve and thus were able to find out separately the period of each of these two constituent curves. From the studies described in (3) and (4), it has become clear that three types of waves of different periods T1, T2 and T3, are existent in the first coda waves of earthquake motions. T1 agreed well with the predominant period decided by (1) and (2), while T2 and T3 were constant, being 20 sec. and 16 sec. respectively. Moreover, it was remarkable to notice that T1, when it was larger than 19 sec., was accompanied by T2, and when much smaller than 19 sec., by Tn. In the second coda waves, however, we found only one period, i.e. T3 (T3,=16 sec.), and this agreed well with the result obtained in (1) and (2). These studies made it clear that three types of waves of different periods exist in the coda oscillations of earthquake motions. T1 increases with the epicentral distance, while T2 and T3 have constant values with all the earthquakes irrespective of their difference in the epicentral distances. In Chapter 8 it was concluded that T1 represents a period of waves propagated from the origin, and that the length of the period of those waves increases with the epicentral distances ⊿, the relation between T1 and ⊿ being given by T1=■式■, where α is a constant connected with the coefficient η of the internal friction of the earth's crust. From the observed values of T1 and ⊿ we get η=0.4×1010 C.G.S. Unlike T1, the periods of T2 and T3 were considered to represent the periods proper to the vibrating systems upon which the observations of earthquake motions have been made. As for the vibrating systems it will be most reasonable to think of the earth's surface layers, and it was assumed that these periods T2 and T3 might represent the free oscillation periods of the standing shear waves that take place in the surface layers. For the purpose of corroborating this assumption, in Chapter 7, the periods of the standing waves that take place in surface layers have been calculated by Rayleigh-Ritz's method. On the basis of the most reasonable numerical constants for the physical properties of the surface layers, it was calculated that, in order that the period of the standing shear waves taking place in the first and second layers might be equal to 16.0 sec. and that in all the three layers equal to 20 sec., the thicknesses of the three layers must of necessity be 0.83 km, 3.95 km and 17.1 km respectively. These figures agree very well with the result obtained from seismological investigations. From these facts it may be reasonably concluded that two predominant periods T2 and T3 that appear in the coda oscillations of earthquake motions represent the free oscillation periods of the earth's surface layers.
雑誌名
東京大學地震研究所彙報 = Bulletin of the Earthquake Research Institute, University of Tokyo
巻
29
号
1
ページ
116 - 141
発行年
1951-06-25
ISSN
00408972
書誌レコードID
AN00162258
フォーマット
application/pdf
日本十進分類法
453
出版者
東京大学地震研究所
出版者別名
Earthquake Research Institute, University of Tokyo
アイテムへのリンクの際はタイトル右肩にある「Permalink」をご利用ください！
Please use "Permalink" at the top right of the title when linking to items!
On the Coda Waves of Earthquake Motions. (Part 6)
ji0291010.pdf
(1.38MB)
