Author(s) NAKAGAWA, Ichiro; NODA, Hiromu; HAT.

Size: px

Start display at page:

Download "Author(s) NAKAGAWA, Ichiro; NODA, Hiromu; HAT."

Marion Dortha Hicks
5 years ago
Views:

the Geophy University (1966), 6: 255-265 ssue Date 1966-12 URL http://hdl.handle.

1 Title FREE OSCLLATONS OF THE EARTH OBSE GALTZN SESMOGRAPH AT ABUYAMA, JA Author(s) NAKAGAWA, chiro; NODA, Hiromu; HAT Citation Special Contributions of the Geophy University (1966), 6: ssue Date URL Right Type Departmental Bulletin Paper Textversion publisher Kyoto University

2 Special Contributions, Geophysical nstitute, Kyoto University, No. 6, 1966, FREE OSCLLATONS OF THE EARTH OBSERVED BY A GALTZN SESMOGRAPH AT ABUYAMA, JAPAN By chiro NAKAGAWA, Hiromu NonA and Asahi HATTOR (Received October 28, 1966) Abstract The great Chilean earthquake of May 22, 196 as observed ith a Galitzin seismograph installed at Abuyama Seismological Observatory of Kyoto University, Japan. Only a vertical component of record as used in the present investigation. The record as read at every minute during a period of about 5.5 days beginning on May 19, 11 h m, 196 (UT). The hole reading as divided into five intervals in order to investigate the free oscillations of the earth excited by both the main shock and one of the large foreshocks of the Chilean earthquake. A lo-cut filter as applied tice to all of the reading values, and then a Fourier analysis as made for every interval. t has come to be established that the free oscillations of the earth-spheroidal oscillations-could be detected by the present investigation and that their periods ere in excellent agreement ith periods obtained at other observation stations for the same earthquake and also ith those predicted by theoretical investigations. An analysis as also made using data reread at 15 second intervals in order to investigate a fine spectral structure for loer modes of the spheroidal oscillations, and a splitting as found in spectral peaks corresponding to modes osz, osa and os. 1. ntroduction Free oscillations of the earth have theoretically being studied since the latter part of the nineteenth century, hile they ere investigated for records obtained at the time of the great Chilean earthquake of May 22, 196. earth's free oscillations excited by this earthquake ere detected ith various instruments at several observation stations over the orld (Benioff et al. (1961J, Ness et al. (1961J, Alsop et al. (1961), Bogert (1961), Buchheim et al. (1961), Bolt et al. (1962), Nakagaa et al. (1964)), and obtained periods corresponding to the earth's free oscillations ere in excellent agreement ith periods theoretically predicted on the basis of an earth's model after Gutenberg and others. n Japan, the free oscillations of the earth ere observed ith an The

256. NAKAGAWA, H. NODA AND A. HATTOR Askania gravimeter orking for the purpose of observing the earth tides at Kyoto (Nishimura et al. (1961), Takeuchi et al.

3 256. NAKAGAWA, H. NODA AND A. HATTOR Askania gravimeter orking for the purpose of observing the earth tides at Kyoto (Nishimura et al. (1961), Takeuchi et al. (1962)), and periods obtained ere in good agreement ith those obtained at other observation stations, as ell as theoretical periods. A trial to detect free oscillations of the earth using the record (vertical component) obtained ith a Galitzin seismograph at Abuyama Seismological Observatory of Kyoto University, at that time, is described in the present article. Several large earthquakes ere folloed successively by the great Chilean earthquake before and after the main shock. is also made for one of them. 2. Data and method of analysis A similar investigation The position of observation station, characteristics of the instrument and epicentral data of the earthquakes under consideration are shon in Table 1. Table 1. Description of the observation station, instrument and epicentral data Observation station Longitude Latitude Altitude Depth Abuyama '51'24"N 22m 2m nstrument used Constant Recording speed Galitzin seismograph (vertical) T,= 8. sec Tu=81. sec Vmaa:=86 h,=1.4 hg=l.o a =.1 2mm/min Epicentral data Origin time Epicentre Foreshock i May 21. 1:2:5, 196 (UT) 37+" s. 73-}-' Magnitude Main shock May 22, 19:l1:2, 196 (UT) Reading of the original record as made at every minute during a period of about 5.5 days from May 19, 11 h m, 196 to May 25, h 19m, 196 CUT). t as made again at every 15 seconds for a hole day of May 23, 196 (UT). An accuracy of the reading as up to.1 mm on the recording paper. The hole reading period as divided into five intervals. They are

4 FREE OSCLLATONS OF THE EARTH 257 hereinafter referred to as intervals,,, V and V, respectively, and their details are shon in both Table 2 and Fig. 1. The values read at both every minute and every 15 seconds are available for the interval V, hile the values read at every minute are available for other intervals. Table 2. Different intervals adopted in analysis First time Last time Origin time Number of of analysis readings nterval May 19, 11: May 2, 1:59 nterval May 19, 13: May 2, 1:59 May 19, 23: May 19, 22 : nterval May 21, 12:3 May 22, 1:29 May 21, 23: nterval V May 23, : May 23, 23:59 May 23, 11: nterval V May 24, : May 24, 23:59 May 24, 11: (Time in UT) May 19, Origin time of Origin time of the foce sl1ock fhe main shock t m V Fig. 1. Different intervals adopted in analysis. Speaking in more detail about the intervals : nterval : Quiet 24 hours before hen the foreshock as occurred. nterval : 22 hours including in the interval ; that is, quiet 22 hours before hen the foreshock as occurred. Last time of this interval as the same as that of the interval. nterval : 22 hours just after hen the foreshock as occurred. The earth's free oscillations excited by the foreshock ere expected to be recorded in this interval, if they ere existing. nterval V : 24 hours just after hen the main shock (the great Chilean earthquake) as occurred. Free oscillations of the earth excited by the main shock ere expected to be recorded in this interval, if they ere existing. nterval V : 24 hours folloing the interval V. The earth's free oscillations ere also expected to be recorded in this interval, if they ere existing, but their amplitudes should be small in comparison ith those for the interval V. When free oscillations of the earth are discussed, periods extending several tens of minutes to several minutes must be taken into account. n order to

258. NAKAGAWA, H. NODA AND A. HATTOR remove aves ith periods longer than 2 hours, a lo-cut filter as applied tice to all of the values read at every minute.

132-119x2X2=844 for the intervals and. The values thus obtained ere analyzed by Fourier's method.

5 258. NAKAGAWA, H. NODA AND A. HATTOR remove aves ith periods longer than 2 hours, a lo-cut filter as applied tice to all of the values read at every minute. 119 reading values at each end of the intervals disappeared in each filtering process and consequently the total numbers of the values amounted to x2x2=964 for the intervals, V and V x2X2=844 for the intervals and. The values thus obtained ere analyzed by Fourier's method. The values read at every 15 seconds for the interval V ere analyzed by Fourier transform method, ithout applying any lo-cut filter, the number of the values used being Results of the analysis Results of the Fourier analysis are shon in Figs. 2 to 6. Figs. 2 to 6 correspond to the intervals to V, respectively. The values read at every minute ere used in the analysis. Results of the analysis by the Fourier transform method are also shon in Figs. 7 to 1. The values read at every 15 seconds ere used in this analysis. Theoretical periods of the earth's spheroidal oscillations calculated for the earth's model after Gutenberg-Bullen A are also shon in Figs. 4 to Discussion and conclusion As Fig. 2 is poer spectrum of the record for the quiet period (interval ) before the foreshock, any r emarkable peak should not originally be seen in NTERVAL PEROD N MNUTES n.--T 1 --T. 1 rr,-r-t--t---trrnttrrrtrnn-r 1 n j QO 15 Q2 Q25 FREQUENCY N CYCLES PER MNUTE Fig. 2. Poer spectrum corresponding to the interval.

FREE OSCLLATONS OF THE EARTH 259 NTERVAL T PEROD N MNUTES 6 5 [1111111111'-,-o-.-1 <! "' f-- (.) o._ UJ "' o._ _J ::: f-- <! _J "" Fig. 3.

6 FREE OSCLLATONS OF THE EARTH 259 NTERVAL T PEROD N MNUTES 6 5 [ '-,-o-.-1 <! "' f-- (.) o._ UJ "' o._ _J ::: f-- <! _J "" Fig. 3. FREQUENCY N CYCLES PER MNUTE Poer spectrum corresponding to the interval. <! Ct: t- NTERVAL ill Q. (/) "' "' Q. 2 _J > i= <! _J Ct: PEROD N MNUTES [ : lllili!illllll ill t l!,s,,s.,s. ose osso os5,s,,s"' ' :--H-i '-'-' ' M ' M \ - if r M Fig OJO FREQUENCY N CYCLES PER MNUTE Poer spectrum corresponding to the interval. the spectrum. But, several strong peaks are actually found in the spectrum, as shon in Fig. 2. Their periods are as follos : 35.6, 18.2, 12.2, 9.34, 7.23, 6.1 and 5.17 minutes. Taking the rotating period (about 36.3 minutes) of the recording drum of the seismograph into consideration, these periods recognized in Fig. 2 are due to the rotating period of the record. Except these strong peaks, Fig. 2 shos noise level of the record. On the contrary, Fig. 5 shos poer spectrum of the record for the

7 26. NAKAGAWA, H. NODA AND A. HATTOR NTERVAL V PEROD N MNUTES s;-,_-- 1 r 5 rr,-t 1 -T 9 8 r---rrntttnttnn""",... 4 (.) UJ Q. U) :: UJ,. Q....J " UJ,... 2: «...J UJ ::.s,,s. s.s os1o os1s oszo,s,,s., Fig OJ FREQUENCY N CYCLES PER MNUTE Poer spectrum corresponding to the interval V. NTERVAL V PEROD N MNUTES 4 «:: (.) UJ Q. U) 5 3 Hil--11--'-1,}11---'--\t-Htt+--rltt---lh Cl. " g H1--?:...J :: FREQUENCY N CYCLES PER MNUTE Fig. 6. Poer spectrum corresponding to the interval V. disturbed period (interval V) after the main shock of the great Chilean earthquake. Many remarkable peaks are found in the spectrum of this figure in comparison ith that of Fig. 2 and it is therefore perspicuous that oscillations corresponding to these peaks in Fig. 5 ere excited by the main shock. Theoretical periods calculated for the Gutenberg-Bullen A earth's model are also shon in Fig. 5. As can easily be seen from Fig. 5, the positions of the peaks obtained by the present analysis are in good agreement ith those theoretically predicted.

FREE OSCLLATONS OF THE EARTH 261 Theoretical os2 < g: 4 tl 8; "' 3: i' Theoretical 35.9min 352mln Benioff and others 36.34min Q. 3 f----------4- ="'--,L'----..J > >= < uj2f--------------j "' Fig. 7.

8 FREE OSCLLATONS OF THE EARTH 261 Theoretical os2 < g: 4 tl 8; "' 3: i' Theoretical 35.9min 352mln Benioff and others 36.34min Q. 3 f ="'--,L'----..J > >= < uj2f j "' Fig FREQUENCY N CYCLES PER MNUTE Fine structure of poer spectrum around os2 (interval V). Fig FREQUENCY N CYCLES PER MNUTE Fine structure of poer spectrum around os3 (interval V). <( "' 5 4 (L (/) "' :;: (L -' f-- <( -' "' 3 2 Theoretical os4 <( "' f-- l) 4 (L (/) "' :;: Q._ 3 -' > t= <( -' 2 a:: s. oss Theoretical 2.1min 19.19min Fig FREQUENCY N CYCLES PER MNUTE Fine structure of poer spectrum around os (interval V). Fig FREQUENCY N CYCLES PER MNUTE Fine structure of poer spectrum around oss (interval V). Several large earthquakes ere successively occurred before and after the main shock. n order to investigate hether free oscillations of the earth had been excited by such earthquakes or not, similar analysis as made for the earthquake occurred at May 21, 1 h 2 m 5s, 196 CUT). 132 reading values ere used in the analysis for the interval. For the comparison purpose, the same number of reading values as selected from the undisturbed interval. Poer spectrum of the record for the undisturbed period (interval ) before the foreshock and that for the disturbed period (interval ) after the

262. NAKAGAWA, H. NODA AND A. HATTOR foreshock are shon in Figs. 3 and 4, respectively. Fig. 3 shos noise level of the record, notithstanding the folloing periods can be picked up from this figure as spectral peaks, 35.

9 262. NAKAGAWA, H. NODA AND A. HATTOR foreshock are shon in Figs. 3 and 4, respectively. Fig. 3 shos noise level of the record, notithstanding the folloing periods can be picked up from this figure as spectral peaks, 35.1, 18.3, 12., 8.96, 7.39, 6.1 and 5.17 minutes. These are due to the rotating period of the recording drum. f the earth's free oscillations ere excited by the foreshock, spectra. peaks corresponding to them should be recognized in Fig. 4. Examining in detail the spectrum of Fig. 4 compared ith that of Fig. 3, some of the spectral peaks are found commonly in both Figs. 3 and 4, hile many other peaks are also recognized only in the spectrum of Fig. 4. The spectral peaks found commonly in Figs. 3 and 4 are, as described above, due to the rotation of the recording drum and those only in Fig. 4 can be regarded as the earth's free oscillations excited by the foreshock. The spectral peaks obtained for the interval (Fig. 4) are neither so remarkable as those for the interval V (Fig. 5) nor alays coincide ith positions for the theoretical ones. Taking magnitude of the earthquakes into consideration, it is natural that the free oscillations of the earth excited by the foreshock are smaller than those by the main shock. t should be significant that similar investigation ill, for the future, be made for records of earthquakes ith magnitude of about 7+, because a great earthquake hose magnitude is more than 8 ill be of rare occurrence. n Fig. 6 is shon poer spectrum of the record for the interval V. The spectral peaks shon in Fig. 6 are not so remarkable as those in Fig. 5, oing to rapid attenuation of the seismograph. Periods corresponding to the spectral peaks obtained by the present analyses are shon in Table 3. n this table, theoretical periods calculated for the earth's model after Gutenberg-Bullen A are also shon. n case of large ave number n, phase velocity C of surface aves is expressed by the folloing formula, 2rr:a C= (n+.5) T' here T and a are period for each mode and radius of the earth, respectively. The surface aves corresponding ith spheroidal oscillations are Rayleigh aves in the case of the earth's free oscillations. Phase velocities ere calculated using the periods for each mode of the earth's free oscillations obtained for the interval V. They are shon in Fig. 11. As can easily be seen from Fig. 11, the phase velocities thus obtained are slightly smaller than those theoretically calculated for the Gutenberg-Bullen A earth's model.

10 FREE OSCLLATONS OF THE EARTH 263 Table 3. Observed and theoretical periods of free oscillations of the earth Theoretical periods (Gut.-Bul. A) Periods observed at Abuyama nterval nterval V nterval V (Unit in minutes) Magnification curve of the seismograph is shon in Fig. 12. An amplitude of the vertical motion of ground corresponding to the spheroidal oscillations in the interval V is calculated as Table 4. As shon in Table 4, the amplitude of the vertical motion for the mode os2 is 17.2 mm. On the other hand, its value has been calculated to be 5.2 mm from data obtained ith an Askania gravimeter at Kyoto. The amplitude for the

$i J \ \ - l \ \ /, ' i :r i i \ Period n seconds l -- - Fig. 11. Observed and theoretical (Gutenberg Bullen A) phase velocity. 1' 1' 1 l<r' 1 N /i r---- f--- 1/ 1 1 Period n seconds Fig. 12.$

11 264. NAKAGAWA, H. NODA AND A. HATTOR a :i >.6 54 l :-Tl-1 T------,- 1-:--t. --- o Observed! f1 - Theoret icol -- - i ! i J \ \ - l \ \ /, ' i :r i i \ Period n seconds l -- - Fig. 11. Observed and theoretical (Gutenberg Bullen A) phase velocity. 1' 1' 1 l<r' 1 N /i r---- f--- 1/ 1 1 Period n seconds Fig. 12. Magnification curve of the Galitzin seismograph. mode os2 obtained by the present Table 4. Amplitude of the vertical motion of the ground excited by the great Chilean earthquake analysis is therefore about three times larger than that obtained by Amplitude the Askania gravimeter. Q values (millimetres) for higher modes (n=93) calculated from amplitude decay for the os os a os 4 os s intervals V and V are hundreds. The Q values obtained are much osts larger than those obtained at some os os observation stations at the time of.192 the same earthquake. Generally speaking, spectral peaks of higher modes (n=93) are considerably clear in comparison ith those of loer modes, as can easily be seen from Fig. 5. This is due to the fact that the number of aves of higher modes included in a certain interval is much larger than that of loer modes. n the case hen the reading from the record as done at every minute, the position of the spectral peaks corresponding to loer modes is unreliable because of roughness in period difference of adjoining ave numbers. The record of the interval V as then reread at every 15 seconds and the reading values obtained ere analyzed by Fourier transform method, ithout applying any lo-cut filter. Results of the analysis near the modes os2, os3, os4 and os5 are shon in Figs. 7 to 1, respectively. Spectral splitting is recognized for

12 FREE OSCLLATONS OF THE EARTH 265 the modes os2, osa and os,, as shon in Figs. 7 to 9, although the splitting around a period corresponding to the mode S 3 is less clear than that for the modes os2 and os4. Such a separation in spectrum is also recognized in Fig. 1. Fig. 1 has maxima at 2. and 19.2 minutes. The former may be the mode oso and the latter oss. The spectral splitting has been observed in America at the time of the same earthquake (Benioff et al. (1961], Ness et al. (1961]). Results for the modes os2 and os 3 obtained by Benioff and others Table 5. Observed periods of the spectral splitting Observed period 54.6 and and and 25.3 Benioff et al and and 35.2 (Unit in minutes) Acknoledgements are also shon in Figs. 7 and 8, respectively. Obtained periods of the spectral splitting are shon in Table 5. The periods of the splitting obtained by the present investigation are slightly shorter than those obtained by Benioff and others. The authors ish to thank Professor Haruo Miki of the Abuyama Seismological Observatory of Kyoto University for his generous permission to use the record. The programme made by Dr. Takeshi Mikumo as used in the analysis by Fourier transform method. This research as made through UNCON using an BM 79 computer. References Alsop, L. E., G. H. Sutton and M. Eing, 1961 ; Free oscillations of the earth observed on strain and pendulum seismographs, Journal of Geophysical Research, 66, Benioff, H., F. Press and S. Smith, 1961 ; Excitation of the free oscillations of the earth by earthquakes, ibid., 66, Bogert, B. P., 1961 ; An observation of free oscillations of the earth, ibid., 66, Bolt, B. A. and A. Marussi, 1962 ; Eigenvibrations of the earth observed at Trieste, The Geophysical Journal, 6, Buchheim, W. and S. W. Smith, 1961 ; The earth's free oscillations observed on earth tide instruments at Tiefenort, East Germany, Journal of Geophysical Research, 66, Nakagaa,., P. Melchior and H. Takeuchi, 1964; Free oscillations of the earth observed by a gravimeter at Brussels, Communications de l'observatoire Royal de Belgique 236, Serie Geophysique 69, Ness, N. F., J. C. Harrison and L. B. Slichter, 1961 ; Observations of the free oscillations of the earth. Journal of Geophysical Research, 66, Nishimura, E.,. Nakagaa, K. Hosoyama, M. Saito and H. Takeuchi, 1961 ; Free oscillations of the earth observed on gravimeters, Journal of the Seismological Society of Japan, Series 2, 14, Takeuchi, H., M. Saito, N. Kobayashi and. Nakagaa, 1962 ; Free oscillations of the earth observed on gravimeters. ibid., Series 2, 15,

Possible Cosmic Influences on the 1966 Tashkent Earthquake and its Largest Aftershocks

Geophys. J. R. mtr. Sac. (1974) 38, 423-429. Possible Cosmic Influences on the 1966 Tashkent Earthquake and its Largest Aftershocks G. P. Tamrazyan (Received 1971 August 13)* Summary I present evidence