THE　REPORT ON THE TSUNAMI OF THE CHILEAN EARTHQUAKE, 1960

face

TECHNICAL REPORT OF THE JAPAN METEOROLOGICAL AGENCY

No. 26

THE　REPORT ON THE TSUNAMI OF THE

CHILEAN EARTHQUAKE, 1960

気象庁技術報告

箱26号

昭和35年5月24臼

チリ地震津波調査報告

昭和38年3月

気象庁

PUBLISHED　BY THE JAPAN METEOROLOGICAL AGENCY,　TOKYO

MARCH 1963

PREFACE

Before dawn of May 24, 1960, a big tsunami attacked suddenly on the Pacific coast of Japan and caused much loss of human life and much damage on properties. The tsunami was found to be generated by the big earthquake which occurred around 4 a. m. of the previous day near Chile.
J. M. A. (Japan Meteorological Agency) issued the tsunami warning to inhabitants in the coastal region of Japan. Considering the experience we had this time, J. M. A. is now planning to establish tsunami warning system in order to issue warning adequately for a tsunami caused by distant earthquake such as the Chilean one and ordered Dr. Y. Kawabata, chief of the Observation Division, to go to U. S. A. to strengthen the international cooperation for the rapid exchange of observed data.
Furthermore, to do field investigations on the occurrence and the feature of attacking of tsunami, J. M. A. ordered Dr. T. Hirono, chief of the Seismological Section, to go to Chile. He is responsible for the description of Chapter IV.
Each district meteorological observatory as well as weather stations all over Japan surveyed all the coast of Japan about the feature of the attack of tsunami and got valuable data for the tsunami warning and prevention of damage by tsunami in future.
This is the abbreviated translation of the Japanese edition " Report on the Tsunami of the Chilean Earthquake, 1960 " published on March, 1961 by J. M. A. In this English edition, which is compiled by the members of the Seismological Section of J. M. A., the report of the field investigation on the coastal region of Japan and the circumstances about the issue of tsunami warning are omitted and only the introductory explanation of tsunami warning system now in effect in Japan is given. Descriptions on all the other items are also shortened.
Our sincere thanks go to JOTI (Japanese Organization for Tsunami Investigation) and Universidad de Chile for their permitting us to use the data they gathered and to Dr. E. Kausel and Dr. E. Gajardo of Instituto de Geofisica y Sismologia, Universidad de Chile and the Japanese Embassy in Chile (Ambassador ; Mr. Rokuzo Yaguchi) for their kindness given to Dr. Hirono during his stay in Chile. Our cordial thanks are also due to members of various organizations and private persons who gave us valuable data and help.
Expenses for the dispatch to U. S. A. and Chile were paid by the "Expenditure Necessary for Urgent Countermeasure to Damages Caused by the Chilean Tsunami ".


March,1962


Kiyoo WADATI


Director-General
of J.M.A.

The Report on the Tsunami of the Chilean Earthquake, 1960
INTRODUCTION

A great earthquake occurred on May 22, 1960 in Chile. It was accompanied by a big tsunami and killed 909 persons and destroyed many cities and towns in southern Chile.
The tsunami which traveled over the Pacific Ocean hit the Pacific coast of Japan about 22 hours after the shock and caused much loss of life and big damage.
The origin time and the epicenter of the earthquake were determined by USC and GS (United States Coast and Geodetic Survey) as May 22, 19h 11m 20s (GMT), and 38 degrees South, 73 degrees 1/2 West, respectively. Its magnitude was estimated by the Seismological Observatory at Matsushiro as 8x(3/4), showing that it is one of the biggest earthquakes in the world.
A group of big fore-shocks began with a destructive earthquake near Concepcion on May 21, 10h 20m, and involved five shocks in all with magnitude of more than 7. But the after-shocks are not so conspicuous as the fore-shocks.
The tsunami was caused by a crustal deformation of a tremendous scale along the coast of southern Chile. It is about 600 km long between Isla Mocha and Isla Guafo, where the ground upheaval of 1-2 m took place and the intermediate region between them subsided, attaining at most about 2 m near Maullin.
It seems that the sea water flowed towards the central region of the crustal deformation from the both sides and this motion of water resulted in a big wave with a period of about one hour.
The tsunami was propagated over the Pacific Ocean. The energy of tsunami was estimated to be 3x10^23 ergs from the tide gauge record of Christmas Is. It is about the same with the energy of tsunami of the Kamchatka earthquake, 1952. Therefore, the present tsunami can not be thought to be widely different in the order of energy from those ever occurred. But, many strange movements of sea water were seen here and there along the coast of Japan because of its long period.
Many mareograms were gathered from stations around the Pacific Ocean, and the arrival times and amplitudes of the tsunami wave were read. The arrival times were found to be in good agreement with the theoretical ones estimated from the refraction diagram. The wave fronts drawn in it showed the clear convergency of the tsunami energy to the northeastern Honshu of Japan, as well as the coast between Okinawa and Philippine. This is principally due to the sphericity of the earth. This fact explains the observational result of lowest heights at islands in the midst of the Pacific. But the Hawaii Island is an exception, where the special submarine topography near the island made the wave height high enough to assault Hilo city killing 61 persons.
The average of the, maximum height of wave (from top to bottom) in, northeastern Honshu of Japan was 3 m and the highest one was observed at Hachinohe city attaining 5.8m. In the coast of Hokkaido the average value was 2 m and the highest one appeared at Kushiro city attaining 6.1 m. In western Japan the average height was 1-2 m and the highest 2-4 m at Omaezaki, Owase and Kocki.
The casualties in Japan caused by the tsunami were as follow; dead: 119, missing: 20, injured: 872, houses destroyed: 1,571, houses half destroyed: 2,183, houses washed away: 1,259, houses flooded over floors: 19,863.

CHAPTER I. INSTRUMENTAL OBSERVATIONS OF THE CHILEAN EARTHQUAKE

1.1 Introduction

A very large earthquake, rare in recent years, occurred in Chile on May 22, 1960 (GMT). It caused damages in large scale to cities in southern Chile. A big tsunami accompanying the earthquake devastated the coast of Chile, and after travelling through the Pacific Ocean it brought big damage to Japan. It has not occurred since 1877 that a tsunami originating from such a distant place reached Japan and caused such a great damage.

1. 2 Epicenter and magnitude

Among a series of Chilean earthquakes which started with a considerable big shock on May 21, 10 h 02 m (GMT), the largest one accompanied by a big tsunami giving heavy damage to the Japanese coast occurred on May 22, 19h 11 m (GMT=JST-9hrs) off the coast of the middle part of Chile. According to the preliminary report of USC and GS, its origin time and location of epicenter were 11 m 20s and 38ーS and 73x(1/2) degrees West (afterwards it was revised as 41 degrees South, 73x(1/2) degrees West). Its magnitude was reported from various sources as follows; M=8x(3/4) (Matsushiro, Santiago), 8x(1/2) (Berkeley), 8.4 (Moscow), 8x(1/4)-8x(1/2) (Pasadena).

1. 3 Fore-shocks and after-shocks

Many shocks were observed before and after the main shock, and those which occurred before the end of August, 1960 are listed in Table 1.1. We owe this list to the report of USC and GS togather with their origin times and epicenters. The times of occurrences at Matsushiro Observatory are concerned with the first phase. The values of magnitude are based on the report of USC and GS and other observatories.
There were considerable numbers of fore-shocks having large magnitudes, especially two which occurred 15 min and 30 sec before the main shock were estimated to have magnitudes of 8 and 7x(1/2)-7x(1/4) respectively.
After-shocks are reported to have occurred 24, 9, and 4 times in June, July and August respectively. Among these the number of those having M not less than 6 is shown in Table 1.2. The geographical distribution of these shocks is shown in Fig. 1.1 which shows that the seismic activities took place within about 4.0 x 10^5km^2 in area extending from 37 degrees South to 48 degrees South in latitude and from 72 degrees West to 77 degrees West in longitude.Symbols used in the last column are as follows. Mat: Matsushiro, Pas: Pasadena, Berk:Berkeley, Mos : Moscow, Sant : Santiago, Tif : Tiflis, Simf : Simferopol.

Symbols * and ** mean errors less than 1/10 min and 1/4 min respectively.

1.4 Reading of seismograms

The seismic records either original or copied were gathered from stations in Japan, New Zealand, Canada and U. S. A. The number on earthquakes on which readings were performed amounted to ten involving the main shock (May 23, 04 h 11 m), five foreshocks and four after-shocks (May 21, 19 h 02 m ; 22, 19 h 30 m, 19 h 32 m ; 23, 03 h 55 m ; 04h lOm; 25, 17h 34m; June 6, 14h 55m; 20, llh 01m; 21h 59m (JST)).
The recording instruments are mostly Wiechert tpye in Japan except at Matsushiro Observatory where Benioff, Galitzin and other sensitive types of seismographs are working. The records from foreign countries are of Wood-Anderson, Milne-Shaw, Benioff, Wilmore, and Galitzin type seismographs. The readings were carried out with all records we could gather for the main shock, but those for fore-shocks we did with seismograms of foreign stations and some of Japanese stations. As for the readings of after-shocks we did only with selected Japanese data.
The maximum amplitudes were measured from Wiecherts seismograms and corrected by the dynamical magnification. In the cases of fore-shocks and after-shocks the maximum amplitudes occurred after the onset of L phase, but in the case of the main shock it occurred earlier at about the time of onset of SS phase. This is true for almost all stations. For the convenience of comparison, we read another large amplitudes occurring near the L phase.
As already stated, J. M. Symons and B. D. Zetler of USC and GS gave another epicenter than that of the preliminary report of USC and GS, but they did not give the origin time. Further, the data from the stations near the epicenter could not be available. Therefore, there is no way of determining the location and origin time accurately. But the examination of the travel time curves of P and P(dash) suggests that the epicenter of the fore-shock of 03h 55m is located at 38 degrees South rather than 41 degrees South. As for the fore-shock of 04h 10m, Japanese records were not good for the reading of initial motion because of being overlapped by the coda of the 03h 55m shock, but foreign data favoured the epicenter of 38 degrees South. As for the main shock, initial motions are masked by a shock which occurred just 15 minutes before, hence their behaviour is not clear, but mostly by foreign records we could rather say that the epicenter seems to be located at 38 degrees South. Thus we adopted 38 degrees South, 73x(1/2) degrees West as the epicenter and May 23, 04h 11m 20s as the origin time. The epicentral distances are given for both epicenters, that is, 38 degrees South, 73x(1/2) degrees West and 41 degrees South, 73x(1/2) degrees West in the table.

(1) Data of the main shock and its travel time

The results of reading of records of all stations are given in Table 1.3. In Figs. 1.2 and 1.3, travel time curves are shown by plotting these data, of which Fig. 1.2 contains both Japanese and foreign data, and Fig. 1.3 contains only Japanese data. As the standard travel time for surface waves we employed Gutenberg-Richter (25 km) Table and forbody waves Jeffreys-Bullen (0 km) Table. Foreign seismological data fall between 79 degrees and 85 degrees and Japanese ones between 150 degrees and 160 degrees in epicentral distance. In either case P or P(dash) of the main earthquake is, as mentioned before, obscured in general by overlapping coda of fore-shocks of 03h 55m and 04h 10m. Appearance of phases on records of Japanese stations was rather clear in Z- and E-components than N-component and SS phase was observed at Japanese stations with large amplitude arriving 30-60 sec later than the standard travel time table. In the cases of Chilean earthquakes of 1933 (M: 7.6) and 1939 (M: 7.8), just a similar retardation of SS wave was observed in Japan. In Japan the largest amplitude occurred in the vicinity of SS phase with the values between 1,000 and 5,000 mu, and period of 30-50 sec. The amplitude of E-component is generally larger than the other ones. The reason why the maximum amplitude happened to occur near SS phase seems to be that the surface waves of the 03h 55m shock, and the SS phase of the 04h 10m shock arrived Japan simultaneously around this time. Anyway, it is very difficult to discriminate these phases.
L phase began to appear at 42-44 min after the time of P(dash) in Japan. This is earlier than Macelwane Table and somewhat later than G-R Table. Accoring to USC and GS report, the magnitude of the 03h 55m shock is 7x(1/2) , and that of 04h 10m is 7x(1/2)-7x(3/4) and the latter has occurred scores of seconds before the main shock making the record complex and the phases indistinct.

(2) Data of the fore-shocks and after-shocks

The seismometric readings of the important fore-shocks and after-shocks which occurred in the period of about one month are shown in Tables 1.4-1.12. Fig. 1.4 shows the travel time of the Oh 55m shock of May 23. G-R Travel Time for surface wave and J-B Travel Time for body waves are drawn for reference.
The initial motions of all shocks, whenever we could discriminate the sense, started in push sense at all the stations in Japan and foreign countries. At some Japanese stations a phase observed emerging on record next to P_1(dash) is likely to be P(dash)_2 but we expressed them in the tables as simply i or e phase. The next phase to P(dash)_2 appearing after a time lapse equal to about P_2(dash)-P_1(dash) is presumably PP, though its arrival is late by scores of seconds as compared with J-B Table. A similar phenomena could be seen in the cases of the Chilean earthquakes of 1933 and 1939.
Surface wave began at about 50 min after P(dash) for all fore- and after-shocks. Maximum waves were found in the sequence of the surface wave train with amplitude of 50-300 mu and period of 17-50 sec, much smaller than that of the main shock.

CHAPTER II. TSUNAMI

The Chilean tsunami of 1960 was recorded at many tide gauge stations in Japan,not only along its Pacific coast but also along the Okhotsk Sea and Japan Sea coasts.This tsunami brought big damage to Japan, especially to the Pacific coast of Hokkaido and northeastern Honshu. In order to investigate this tsunami, we gathered the records of tide gauges around the Pacific Ocean, of which foreign data we owe to the USC and GS report and the collection of JOTI (Japanese Organization of Tsunami Investigation). These data show that the higher amplitude of tsunami reached Japan than any other places. Some of the records of tsunami are reproduced in the appendices and their readings are given in Tables 2.1-2.2. The distribution of tide gauge stations is shown in Fig. 2.1.

2.1 The method of reading mareograms

(1) First motion (A)

Generally the arrival time of tsunami is taken from the very point on the record where the sea level began to deflect from the tide prediction curve. The initial motion was distinguished by either (+) or (-) according as it moves upwards or downwards. Further, according at the speed with which the first motion moves vertically is larger or less than 1 cm/min, i or e is adopted as an indication of the clearness of initial motion.

(2) Maximum amplitude (M)

The maximum amplitude and its period are measured by the method shown in Fig. 2.2. If the wave crest giving the maximum value is solitary one, we characterize it bysymbol " s " and the period is given by doubling the time lapse from peak (trough) to trough (peak) of the wave.

(3) Maximum sea level

The highest sea level attained by the tsunami above the mean sea level of the Tokyo Bay is shown by M_T. Instead of that, when the local mean sea level is used as the base line it is designated by " m ". Both levels are practically not very different from each other. When the highest sea level is measured above some particular datum line, it is expressed by M_D. When recording pen went out of scale and the amplitude was estimated roughly, its value is shown in parentheses.

(4) The type of tsunami

The types of tsunami records were classified into 6 groups as shown in Table 2.3.

2.2 Tsunami at coasts around the Pacific

In Fig. 2.3 the data about initial motion and maximum amplitude are shown. All through the Pacific coast except regions not yet investigated, tsunami began with upward movements. The largest amplitude happened in Japan and the next is 335 cm, in Aleutian Isls., then followed, 332 cm in Canada, 293 cm in Hawaii, and 219 cm in Chile. At otherplaces it is lower than 200 cm, especially very small in Chile. The arrival times of tsunami along the east coast of Australia are not systematically distributed and the maximum amplitude is comparatively small except the central part where it attains 162 cm.

2.3 Propagation of tsunami

The tsunami wave originating at the coast of Chile on May 22, 1960 was propagated over the Pacific Ocean as a colossal undulation and gave big damage to Japan as well as Hawaii. Many tide gauges installed around the Pacific Ocean recorded it presenting the valuable data for investigation on the phenomena of tsunami.
The travelling speed of tsunami is approximately expressed by square_root(gxh), where g is the gravity acceleration, h the depth of sea. It is well known that this formula gives good basis for calculation of travel time of tsunami. Practically, we draw the wave fronts on a map using Huygens principle. This map is called a refraction diagram. From this wecan estimate the travel time of tsunami. Moreover by drawing orthogonal trajectories to the wave fronts, we can find out the places where the divergence or convergence of wave energy takes place. The refraction diagram is indispensable as a reference when we have to issue a tsunami warning against distant tsunami. We made a new stereographic map of Pacific Ocean convenient for drawing a refraction diagram of tsunami generating anywhere in the Ocean.
A stereographic projection has two advantages as a basic map for drawing refraction diagram on it. The first is that a circle on a sphere is projected also as a circle on the map, the second is that the angle between two intersecting circles on the sphere is projected as such with the same angle. It is shown in another paper(T. Hirono and S.Hisamoto (1952): A Method of Drawing the Wave Fronts of Tsunami on a Chart. Geophys. Mag., 23, 399-406) how to draw refraction diagram on a stereographic map. In that paper we used the map with the center of projection at the North or the South Pole and felt it very inconvenient as wecould not see the whole part of Pacific Ocean. We need two sheets of such a map when a tsunami is propagated from one hemisphere to the other. To improve this defect, the
center of projection is chosen at a point in the Pacific (0 degrees lat., 150 degrees West) in the new map so that the whole Pacific area can be shown in one sheet.
In the case of the map with the North Pole as a center of projection the radius of elementary circle is given by r=R sin (square_root(g X h X delta t) / (cos (square_root(g X h X delta t)) +sin phi)), where phi is the latitude of the center of an elementary wave, R the distance from the center to the periphery of equatorial circle, delta t the time difference from the present to the next front. In the new map phi is displaced by a distance from the central point (0 degrees South, 150 degrees West) and delta t is taken to be 30 min and we drew fronts with 30 min intervals.
As the tsunami origin, we did not employ the epicenter given by USC and GS in its preliminary report but employed the epicenter 41 degrees South, 73x(1/2) degrees West given by Simons and Zetler of USC and GS. This revised epicenter is located inland. The first 30 min circular front around the epicenter was drawn assuming the sea depth to be 2,500m. The final result is shown in Fig. 2.4. For the reference sake, great circles joining the epicenter to its antipode are drawn, making ten degrees with each other at the epicenter.
It is difficult to continue drawing the front when it approaches the coast, because the depth becomes shallow rapidly and the velocity of tsunami drops abruptly. We coveredthe lack of travel time in this part by numerical estimation. Lastly we compared the resultant whole travel time with observed ones. The comparison is shown in Table 2.4,
and the order of difference between observation and calculation is illustrated in Fig. 2.5. The figure shows good agreement between them at islands amid the Pacific Ocean, North America, and Japan, while at South America observed value is less than the calculated one and at Australia the differences mentioned above are at random.
As will be explained in Chapter V, there is a good reason to think that the first wave of tsunami emitted to the north direction from the epicentral area must have started from the location of 38 degrees South rather than 41 degrees South. Consequently, it is reasonable to have earlier arrival of tsunami than theory along the coast of South America. As for the time difference at Australia, it seems to depend on that the tsunami wave, before reaching the Australian coast, was disturbed by the islands such as New Zealand and New Caredonia, resulting in the ambiguous beginning of the record.
It is notable in Fig. 2.5 that the wave front of the tsunami rushed almost parallel to the coast of north-eastern Japan and that the tsunami converged towards the north- eastern coast of Japan. As other places where tsunami front impinged normally to the coast, we can mention the coasts of Okinawa Islands and New Zealand. Okinawa was a place suffered from considerably big damage.

CHAPTER III. GENERAL DESCRIPTION OF TSUNAMI

The tsunami reached Japan at about 02 h a.m. on May 24 th (JST). The earliest arrival was observed at Oshima, south of Tokyo as 02h 33m and the second one at Hanasaki, eastern Hokkaido as 02h 38m. The first arrival time was earlier on northeastern coast of Japan and getting to the west it was later and was 05h 50m at Makurazaki, southern
Kyushu. On the coast of Japan Sea, the arrival times of tsunami were between 06h-13h on May 24th.
Maximum amplitudes appeared 1.5-5 hours after the arrival of tsunami and were large on the coast of southern Hokkaido, north-eastern Honshu, Tokaido, Kii Peninsula and southern Shikoku, where damages were also severe. Maximum amplitudes were small (about 20-50 cm) on the coast of Japan Sea and Seto Inland Sea. At places where they were large, they appeared 2-3 hours after the arrival of tsunami.
We compare the tsunami this time with that of March 3rd, 1933 (the Sanriku Tsunami) (JST) for the Pacific coast of north-eastern Japan.

(1) In open sea, the heights of Sanriku tsunami vary from place to place and were 5-10 m on the coast of Iwate Pref. and northern part of Miyagi Pref. and about 2 m on the other regions of north-eastern Honshu, while for the Chilean tsunami they are nearly constant at about 2-3 m all along the coast. This difference may be due to the ifference
in distances from the wave source.

(2) In bays, the wave height reached up to more than 20 m in case of 1933 tsunami, while it was lower than 5 m this time. In big bays such as Miyako, Ofunato and Hirota bays, the wave height was higher at the mouth than at inner part in case of 1933 tsunami, while it was higher at inner part than at the mouth this time. This phenomenon may depend on the relation between the period of tsunami and proper period of bay. The periods of tsunami were about 10-20 min for the 1933 tsunami and about 50 min for the Chilean tsunami.

(3) Almost no effect was reported for the Mutsu Bay, the extreme northern part of Honshu, in case of 1933 tsunami. The wave height there attained to about 2 m this time, causing slight damages on Aomori and Mutsu cities.

CHAPTER IV. FIELD INVESTIGATION IN CHILE

4. 1 Distribution of seismic intensity

Making use of the data that IGS (Instituto de Geofisica y Sismologia, Universidad de Chile) gathered from many localities through the country by questionaire cards, and kindly permitted us to use, the distribution of seismic intensities for a fore-shock (the Concepcion Earthquake of May 22) and the principal earthquake were drawn as shown in Fig. 4.1 where Chilean intensity scale, derived from M. M. scale by dividing by 2, was used. Thecenter of the meizoseismal area was located at the place half way between Valdivia and Ancud. This is just the place three degrees south in latitude from the instrumental epicenter in the preliminary report of USC and GS, but it coincides with the epicenter revised afterwards by Symons and Zetler of USC and GS (41 degrees South, 73x(1/2) degrees West).
The radius of the extreme boundary of the perceptible area extends to the north about 1,000 km from the meizoseismal center. With the lack of data, the circumstances to the
east and south are obscure. But the shapes of isoseismal lines seem to be prolonged in the north-south direction.

4. 2 Crustal deformation

According to IGS, they could not find superficially any new fault in the vicinity of the epicenter. But we cannot but think there should have occurred a kind of breakage such as causing a fault under the ground, whether it may appear on the surface or not, because the crustal deformation produced by the shock has so large extension ranging
from 38 degrees South to 43 degrees South in north to south direction along the Chilean coast, and it must have finished forming in a relatively short time.
The meizoseismal area already mentioned is situated in the neighbourhood of the center of that area. The ground upheaved one to two meters at Isla Mocha and Isla Guafo which is located at north and south limits of the deformed area (Fig. 4.2). At Corral and Maullin, the middle part of the region, ground subsided about two meters. Therefore, the half way from the both ends to the center has, so to speak, inclined towards the center. Along the east to west line, also the maximum of vertical displacement seems to have occurred near the coastal line diminishing rapidly toward the inland.
This range of 38 degrees to 43 degrees South agrees approximately with that of the area within which fore-shocks and after-shocks occurred in the same day with the principal shock. The area of after-shocks also stretched in great distances from north to south with narrow width in E-W direction and the central axis of this area seems to run along 100 km off the continental coast somewhat aslant to the coastal line. That the after-shock region of a big earthquake nearly agrees with that of crustal deformation caused by the same earthquake, and consequently, almost agrees with the area of tsunami origin has been made sure of Japanese earthquakes. The same rule seems to hold to this Chilean earthquakes. Therefore, it is reasonable to think that the tsunami was generated under the sea bottom of the area bounded by 38 degrees South and 43 degrees South lines in the north-south direction and possibly having an axis along about 100 km off the continental coast.

4. 3 Tsunami

Big tsunamis assaulted the coast stretching between 38 degrees South and 43 degrees South which coincides with the coast of the crustal deformation, and they were reported to have attacked in a similar manner all through the coast. The first wave reached the coast between 15 to 20min after the earthquake (Fig. 4.2). There were three big waves and the last one was the largest. But it is strange that the wave height along the Chilean coast north of the
region just mentioned was not so high remaining only one to two meters. Only exception is the Talcahuano Bay which is famous for tsunami from an old time and about 4 m height (peak to trough) of the recent tsunami was recorded at the Talcahuano Port (see Fig. 4.3). As already stated in the preceding section the sea bottom between Isla Mocha in the north and Isla Guafo in the south drooped like a catenary. Consequently sea water may have flown from both sides into the central region, finally transforming into a huge wave and travelling to the open sea. The first wave which can be seen in the mareogram of Talcahuano seems to have come from the upheaved region near Isla Mocha, while the second wave, from far distant region near Isla Guafo. This is deduced by assuming that the mean depth of the open sea is 1,000 m, and that it takes 30 minutes in the Talcahuano Bay to reach the innermost station from its mouth.
Topography of the sea bottom near Isla Mocha forms a kind of submarine bank ex-tending to the west into the ocean from the continent. The bank may reflect partially the wave coming from the south hindering its propagation to the north. The attached sea
chart does not show clearly the topography, but if there is a similar submarine bank near Isla Guafo a part of the wave coming from north will be reflected too to the north again. Though the both banks do not seem to extend further than 100 km from the coast, it may be enough to reflect the first wave because the motion of the sea water just after the earthquake must have occurred along the axis of crustal deformation near the coast (depth is about 1000 m) as mentioned before. But it may be reasonable to think that the reflected waves would pass now the deeper part in the ocean (depth is about 2,000-3,000 m) diverting from the original pass. Then about one hour is taken for the wave to travel through the distance between 38 degrees South to 44 degrees South. This seems to explain one hour period of the tsunatni. Big three waves which are said to have assaulted the coast between 38 degrees South and 44 degrees South can be explained by only one barrier near Isla Mocha.
Energy of the whole tsunami may be calculated by the formulae

formula_046.jpg
where E is the energy in erg involved between two wave rays making one degree with each other at the point of wave origin, p the density of sea water, g the acceleration of gravity, A the wave amplitude, lambda the wave length, and S the breadth between the two adjacent rays measured at island where A is measured. From the record of tide gaugeat Christmas and Johnston Isls., we got A and, from the refraction diagram shown in Chapter II of this report, lambda and S assuming the peri u of tsunami to be 60 min, and the energy of tsunami was computed as 1.3X10^23 (Christmas Is.) and 3x10^23 (Johnston Is.) respectively which is the same order with that of the Kamchatka Earthquake of 1947 (M= 8.5). This fact indicates that the energy of Chilean tsunami is not so extraordinarily large for the earthquake with M=8.5, and for the reason why it was so large as to have brought damage to Japan, the spherical surface of the glove seems to be responsible. It gathers the wave energy that once diverged. This fact may be recognized only by taking a glance at the refraction diagram.
Fig. 4.5 shows tide-gauge records at Valparaiso and Coquimbo. We can see in them the tsunami accompanying the fore-shock of May 22. Next, we can see some rather large amplitude of oscillation far after the beginning. Considering the travel time and the condition of easy reflection on the coast of the other side of the Pacific, this seems to be reflected waves from the New Zealand (10 hours) and Japan (22 hours) respectively (see the refraction diagram).

4. 4 Damage and casualties

According to the Proceedings of Chilean Congress held on May 30, 1960, casualties are as follows ; The dead were 909, the missing 834, seriously wounded 235, slightly wounded 432. In the proceedings, the damage of the house, was expressed by percentage of the number of damaged houses to the whole (see Fig. 4.6). Of course there was much damage of roads, bridges, rail roads, and communication lines by the earthquake, but what produced by tsunami was greater. Pto. Saavedra, Mehuin, Corral, Maullin and Ancud were the cities suffured heavily by tsunami. By the earthquake itself, Concepcion, Valdivia, Pto. Montt and Ancud were considerably damaged. The pattern of the damage of these cities were not uniform but the constructions built on artifical ground along the river side, sea side or on depression ground, were the only victims of the shock. But we could see everywhere that fire wall and wall around houses that were made by bricks or something like that were most breakable. In these earthquake suffered cities, even a modern building of good appearance has a poor foundation without any pilings, only constructed on a reinforced concrete plate. Consequently they are very weak for irregular depression of ground under the building. But we think the greatest damage by the recent great earthquake was the lost of vast area by ground subsidence which deprived cultivated field and houses of people.

CHAPTER V. TSUNAMI WARNING SYSTEM IN JAPAN

To mitigate damage caused by tsunami accompanying large earthquakes occurring near Japan as much as possible, the tsunami warning system of Japan Meteorological Agency has been in effect since 1952. The outline of the system is as follows.

(1) The system was established supposing that we have at least 30 min before the arrival of the first wave of tsunami on the coast after the occurrence of an earthquake.

(2) The system consists of 5 tsunami warning centers (weather station attached to J.M.A.) which issue tsunami warning in case of need and of 40 communication centers which send warning to inhabitants in coastal region.

(3) Works to be shared by J.M.A. are to judge the occurrence and magnitude of tsunami and for that purpose the special system has been established in J.M.A. so that it can be made the most of available facilities of J.M.A.

(4) Works to be shared by the communication centers are to send tsunami warning issued by the tsunami warning center to inhabitants in the coastal region within as short
time as possible and for that purpose each communication center has established its own system by which the tsunami warning is given the highest priority and sent to its smallest units.

(5) Mayors of cities, towns and villages who have received the tsunami warning have to let inhabitants know it by the most convenient method and let them retire to safer places.

The tsunami warning system of J. M. A.

I. The tsunami warning system of J.M.A.

Works which should be done by J.M.A. are divided into the following three parts.

a) Seismological observation and rapid communication of the observed data.

b) Judgement of the occurrence and magnitude of tsunami.

c) Communication of tsunami warning to the communication centers.

Details of these works are described as follows.

a) Seismological observation and rapid communication of the observed data.

66 out of weather stations attached to J.M.A. have duties to do seismological observations for the tsunami warning. In case of earthquakes with intensity more than IV (J.M.A. intensity scale) or with double trace amplitude of more than 10 mm (The amplitude should be corrected for the ground conditions) on the record of strong motion seismograph of magnification 1, each station expresses intensity, direction of the initial phase, maximum double amplitude, occurrence time and P-S duration in a code consisted of 14 numerals and send them to the corresponding tsunami warning centers without delay by telegram exclusively used by J.M.A. These worksshould be finished within 5 min after the occurrence of earthquake.

b) Judgement of .occurrence and magnitude of tsunami

We divide coastal regions around Japan into 17 (Fig. 5.1). Five tsunami warning centers (Sapporo, Sendai, Tokyo, Osaka and Fukuoka) share these 17 regions and each center judge the occurrence and magnitude of tsunami in coastal regions under its charge. At each center, they deduce the epicenter using intensity and seismograms obtained at
that center as well as observed data sent from weather stations under its supervision and then give judgement on the occurrence and magnitude of tsunami referring to the above deduced epicenter and plotting the amplitude of the ground motion on the diagram made for the tsunami warning (Fig. 5.2).
Based on their judgement, they choose one of the tsunami warning among the next four according to expected severity.

1. No tidal wave is expected.

2. Tsunami is expected, wave height can not be predicted.

3. Minor tsunami is expected.

4. Major tsunami is expected.

c) Communication of tsunami warning to the communication center.

Tsunami warning chosen by the tsunami warning center is handed to person (s) in charge of communication and it is sent to the communication centers without delay by telegram and (or) telephone exclusively used between them.

II. Warning of tsunami due to distant earthquakes.

From the experience of tsunami caused by the Chilean earthquake of May 24th, 1960, the international cooperation is keenly realized as to the warning of tsunami due to distant earthquakes. Hitherto, we were at a loss in issuing tsunami warning for large earthquakes occurring in Kamchatka and Aleutian Islands regions, because although tsunamis reach Japan in a short time after the earthquake, we have not established the international cooperation system and we can not deduce the epicenter.
We are planning to improve seismometers in Japan and establish international cooperation system for the better issuing of the tsunami warning.

CHAPTER VI. TABLE OF TSUNAMIS CAUSED BY EARTHQUAKES IN THE PACIFIC REGION EXCEPT NEAR JAPAN

Tsunamis caused by earthquakes which occurred in Pacific region except near Japan (ranging from Hokkaido to Kyushu) are summarized as Table 6.3. There are 75 tsunamis in total. The locations of wave sources are clear for 70 out of them. For the other 5 tsunamis, the locations of wave sources are not clear or they remain unknown whether they are tsunamis or high tides.
The " Rika-nempyo " was referred as to the magnitudes M of older earthquakes in the table. "Intensity of tsunami in Japan " does not mean the magnitude of tsunami, but the grade of influence of tsunami on the coast of Japan. They are as follows :Table 6.3 is classified as Table 6.2 which shows the relation between the district of earthquake origin and the effect of tsunamis on Japan.

APPENDICES

1. Table of Damage Caused by the Chilean Tsunami in Japan

2. Mareograms

3. Seismograms

4. Photographs of Tsunami and Damage

Table of Damage Caused by the Chilean Tsunami in Japan

2. Mareograms

Some of the mareograms of the Chilean tsunami are reproduced here for reference. Microfilms of these records were taken and are kept in the Seismological Section, Japan Meteorological Agency. Anybody who wants to use them should apply to the above mentioned section.
Mareograms except those of Chile are collected by JOTI (Japanese Organization for Tsunami Investigation). In the figures, the symbol A means the initial motion of tsunami and B and C are the first and second peaks (or troughs) respectively. The numbers correspond to those in Table 2.1.

3. Seismograms

Seismograms of the Chilean earthquake obtained at various stations are reproduced here. Types and constants of seismographs are shown in the reproductions. Japan Standard Time is used instead of G.M.T. The relation between them is

(J. S. T.) - 9hrs = (G.M.T.)

Our thanks go to stations which offered us the seismograms and allowed us to reproduce them.

4. Photographs of Tsunami and Damage

Photographs of the tsunami and damage caused by the tsunami were reproduced here. Photos No. 26-No. 31 were taken and offered by the Japan Ground Safety Defense Force. Our thanks go to those who took these pictures and allowed us to reproduce them.

backside

昭和38年3月26日印刷

昭和38年3月30日発行

編集兼

発行人

気象庁

東京都千代田区大手町一ノ七

印刷者

花崎実

東京都中央区新富町ニノ十九

印刷所

大東印刷工芸株式会社

東京都中央区新富町ニノ十九