RETURN PERIODS OF TSUNAMIS IN CHILE”

COMPARISION WITH OTHER REGIONS ON THE GLOBE

T.S. Murty1, R.D. Scott1, and C. Fournier2

1
W.F. Baird & Associates Coastal Engineers Ltd.
1145 Hunt Club Road, Suite 1
Ottawa, Ontario K1V 0Y3, Canada
2
Atria Baird Consultores S.A.
Eliodoro Yanex 1893
Oficina 602A
Providencia
Santiago de Chile

Abstract: For the safety of population as well as for the protection of coastal
infrastructure and development of the coastal zone, a knowledge of the return periods of
tsunamis on the coast of Chile is useful. Based on historical tsunami data, return periods
and percentage probability of a tsunami (irrespective of intensity) occurring in various
latitudinal belts along the Chilean coast were determined. Comparison was made with
return periods of tsunamis for several other areas in the world. The resonance
characteristics of the coastal inlets is assessed through a simple analysis of the normal
modes. Finally, the intensity of secondary undulations in several inlets is studied. This
provides information on how long tsunami energy can get trapped inside an inlet before it
gets dissipated.

1
1.0 INTRODUCTION

It is well known that Chile is impacted by tsunamis originating locally as well as from
outside. Figure 1 shows the broad seismic features on the Pacific coast of South
America.

The various coastal regions of Chile can be conveniently divided into the following five
zones listed in Table 1 and shown in Figure 2 (Lockridge, 1985) from a tsunami point of
view. For development of coastal infrastructure as well as for safety of the population, it
is useful to know the return periods of tsunamis in the various coastal belts. Table 1 also
lists the duration of data from Lockridge (1985), number of events during that period,
percentage probability of tsunami event (irrespective of its intensity) occurring in any
given year and a return period in years for that belt. For comparison purposes, row one
lists similar data for Peru.
Table 1
Tsunami Data for Chile and Peru
Region Latitudinal No. of No. of Percentage Return
Belt ( in Deg Years of Tsunami Probability per Period in
South) Data Events Year Years
Peru 3 to 18 388 34 8.8 11.4
North 18 to 27 368 44 12.0 8.4
Chile
North 27 to 33 286 27 9.4 10.6
Central
Chile
Central 33 to 37 224 19 8.5 11.8
Chile
South 37 to 41 412 11 2.7 37.5
Central
Chile
South 41 to 60 337 11 3.3 30.6
Chile

2
2.0 COMPARISON WITH OTHER REGIONS

Table 2 lists similar data for some other parts of the globe for comparison purposes.

Table 2
Tsunami Data for Some Other Regions of the Globe
Region No. of No. of Percentage Return Reference
Years of Tsunami Probability Period in
Data Events per Year Years
Alaska 177 79 44.6 2.24 Lander and
Lockridge,
1989
American 150 69 46.0 2.17 Lander and
Samoa Lockridge,
1989
Australia 127 16 12.6 7.94 Davidson
and Rynn,
1998
Caribbean 500 100 20.0 5.00 Lander and
O’Laughlin
2000
East Coast of 209 11 5.3 19.00 Lander and
USA Lockridge,
1989
Hawaii 175 138 78.9 1.27 Lander and
Lockridge,
1989
India 2,300 25 1.09 92.0 Murty and
Bapat, 1999
Puerto Rico 228 12 5.3 19.00 Lander and
and Virgin Lockridge,
Islands 1989
Turkey 3,300 100 3.03 33.00 Altinok and
Ersoy, 1995
U.S. 167 37 22.2 4.51 Lander and
Possessions Lockridge,
in the Pacific 1989
West Coast 176 103 58.5 1.71 Lander and
of USA Lockridge,
1989

3
3.0 PREDICTIONS FOR FUTURE

As for future predictions, Mazova and Ramirez (1999) suggest that a catastrophic
tsunami is likely to occur in the Arica region of Northern Chile in the next 10 to 20 years.

Table 3 is based on Lockridge (2000, personal communication).

Table 3
Number of Tsunamis (at Least One Meter in Amplitude)
Expected During the
Next 100 Years

Region No. of Tsunami Events

Likely to Occur
Alaska 7
Australia 6
California (West Coast of USA) 6
Chile 9
Europe (Mediterranean Sea Coast) 5
East Cost of USA 0
Hawaii 12
Indonesia 13
Japan 38
New Guinea 15

4.0 RESONANCE CHARACTERISTICS OF CHILEAN INLETS

The resonance characteristics of various inlets on the Chilean coast are explored in terms
of normal modes of oscillations and secondary undulations.

In order to investigate the response of inlets to tsunamis, the resonance characteristics of

some inlets on the coast of Chile have been deduced through simple analytical
considerations. A comparison is made with the inlets of southeast Alaska, the mainland
coast of British Columbia and Vancouver Island. It is shown that the general level of
intensity of secondary undulations is highest for Vancouver Island inlets, and least for
those of Chile and Alaska. It is also shown that tsunami forerunners are more common
than is generally believed.

The Pacific Coast of the South American continent can be conveniently visualized as two
regions of widely differing characteristics: from 9°N to 42°S, a distance along the coast
of about 5700 kilometers, the coastline is regular with a few indentations, while from
42°S to Tierra Del Fuego at 56°S, a coastal distance of 1800 kilometers, the coast is
penetrated by a complex system of fjords. These two regions have parallels on the

4
Pacific coast of North America: from Panama (9°N) to the Gulf of Alaska (59°N), a
coastal distance of 1600 km, the coast is highly indented. All of these regions have been
subject to severe tsunamis.

Tsunamis in the indented regions of both South and North America are modified by
similar physiographic features. Inlets penetrate deeply into the coast and are connected to
the open sea by complex passages between islands. Beyond the islands lies the
continental shelf, averaging 32 kilometers in width of central and southern Chile, and 40
kilometers off the coast of British Columbia and southeast Alaska.

A simple examination of the available tsunami data is made with a view to obtaining
qualitative and quantitative information on a number of aspects, including resonance
phenomena, secondary undulations, tsunami forerunner and initial withdrawal of water.
Detailed comparisons are made between the relevant features of the inlets of the Chilean
coast and those of the inlets of Alaska and British Columbia.

Inlets of the coast of Chile whose resonance characteristics may be defined from
presently available depth information are listed in Table 4 (central Chile) and Table 5
(southern Chile). Their locations and the sections assumed to be resonant are shown in
Figures 3 and 4. Unless otherwise indicated the dimensions of these inlets are those
listed by Picard (1971). The data from Picard as well as those extracted by the writers
from hydrographic charts were used to compute the period of the fundamental
longitudinal mode of the inlets and to determine the possible intensity of secondary
undulations: calculations were based on the lengths, mean widths and mean mid inlet
depths for each of these inlets. The period of the fundamental longitudinal mode of the
inlets was calculated using the Merian formula (for an inlet open at the mouth and closed
at the head), i.e.

Τ1 = 4 L / gH (1)

where T1 is the period, L is the length of the inlet, H is the average depth of the inlet and
g is gravity. The average depth of the inlet obviously would be somewhat less than the
mean midinlet depth and from (1) it can be seen that use of the latter in place of the
former provides a smaller estimate of the period. Thus, as using both, i.e. mean midinlet
depth and an approximate formula, produces opposite effects, the estimates of the periods
may be satisfactory.

Tables 4 and 5 list the identifications of the inlets, their length, mean (midinlet) depth,
period computed from the Merian formula and the ratio of the length to width. The
significance of the last column is explained in Section 5 dealing with secondary
undulations. It can be seen from these tables that the periods of the inlets vary from 15
minutes to 400 minutes, but predominantly they are in the range of 30 to 180 minutes
(Figure 5). The periods with which tsunamis enter from the deep ocean into coastal
waters usually lie in this range and thus the resonance in coastal inlets should be
considered in prediction of tsunami heights.

5
 Table 4
Dimensions and Periods of Fundamental Mode and Intensity of Secondary Undulations of Inlets of Central Chile

6
 Table 5
Dimensions and Periods of Fundamental Mode and Intensity of Secondary Undulations of Inlets of Central and Southern Chile

7
It can be seen from Figure 5 that as far as the resonance periods are concerned, the
distribution is similar for the inlets of the various regions. Thus, for understanding the
resonance phenomenon, a study of one of these systems of inlets enable one to make
deductions about other systems.

5.0 SECONDARY UNDULATIONS

Nakano (1932) defined secondary undulations as the oscillations on tide gauge records
that are distinct from tides. In a bay excited by a tsunami, these oscillations are
essentially the normal modes of the bay. From Nakano’s work it may be seen that the
intensity of secondary undulations is proportional to the length L of the bay and inversely
proportional to the width W and inversely proportional to H3/2, where H is the average
depth of the inlet. In the last column of tables 4 and 5 the calculated intensity of
secondary undulations for each inlet is listed. This intensity value provides in a limited
quantitative manner, a measure of the continued disturbance in an inlet following the
main tsunami wave. For inlets receiving a similar amount of energy from a tsunami,
those with a higher value for the secondary undulation should excite larger amplitudes of
oscillation relative to those with a smaller value.

Figure 5 shows the histogram of the distribution of the number of inlets versus the
resonant period for inlets on the coasts of Chile, Alaska, British Columbia (mainland) and
Vancouver Island. Figure 6 shows a similar histogram for the intensity of secondary
undulations. The distribution is somewhat similar for Chilean and Alaskan inlets and is
drastically different for Vancouver Island inlets. The pattern for British Columbia
mainland coast (and Puget Sound) falls between that for Vancouver Island and that of
either Chile or Alaska. It may be noted that differences of this type are not obvious in
figure 5.

This similarity among the systems does not appear strongly when secondary undulation
activity is considered. Some of the inlets included in tables are separated from the open
ocean by passages so restricted that only minor excitation from tsunamis due to distant
causes (here by distant tsunamis is meant tsunamis which originate in the Pacific Ocean
from a distant earthquake) will be possible. However, those inlets are included since a
tsunami response to seismic activity occurring within their region is possible. Others of
the listed inlets are restricted, rather than closed at the inner end, and many have
bifurcations and confluences with other passages; higher modes and more complex
resonances are likely (Henry and Murty, 1972).

6.0 TSUNAMI FORERUNNER

Nanamura and Watanabe (1961) defined “tsunami forerunner” as a series of oscillations

of the water level preceding the arrival of the main tsunami waves. Since the forerunner
has periods and amplitudes less than the main tsunami wave, the forerunner could be

8
easily distinguished. Nakamura and Watanabe stated that there is no clear indication of a
tsunami forerunner in the data for either the North or South American coasts and they
ascribed the absence to the oblique nature of incidence of the initial wave upon these
coasts. They explained the existence of the forerunner at other places such as Japan as
due to the resonance in bays which could occur before the arrival of the main tsunami
wave. They remarked that no previous tsunamis on the Japanese coast showed
forerunners. They also stated that the 1960 Chilean earthquake tsunami did not generate
forerunners at any region other than Japan.

In an examination of Canadian and United States records for the 1960 tsunami it was
noted that initial waves of lower amplitude and simple form marked the arrival of the
disturbance at most ports. More complex waves of larger amplitude occurred after the
first two or three crests. As in the Japanese example, this pattern was not characteristic of
preceding and succeeding tsunamis. The existence of these low amplitude initial waves
might be useful in determining the potential destruction of the tsunami that follows the
forerunner.

7.0 INITIAL WITHDRAWAL OF WATER

Much has been said about initial withdrawal of water before the arrival of the main
tsunami waves. There are several instances of initial withdrawal and probably just as
many cases of no initial withdrawal. Here we present some information on the initial
withdrawal on the South American coast and it should be noted that this list (Table 6 is
not a complete one due to the inaccuracies of descriptions by people not specifically
trained.

Table 6
Instances of Initial Withdrawal on the Pacific Coast of South America

No Year Month Day Remarks

1 1687 October 20 Initial withdrawal at Callao
2 1751 May 25 Epicentre near Concepcion;
duration of initial withdrawal was
7 minutes.
3 1819 April 11 Epicentre near Copiapo
4 1859 October 5 Water level at Caldera was 14 feet
below average sea level
5 1877 May 9 At San Francisco (37° 40°S,
122°32°W) initial withdrawal.
Initial withdrawal also at Carrizal
Alto (28° 5°S, 71° 10° W)
Chanaral (29° 2°S, 71° 34°W)
Talcabuano, Penco, Aranco

9
 6 1960 May 22 Initial withdrawls at Ancud
Isla Mocna, Isla Guabo, Quellon,
Linao, Punta Corona, Maullin,
Puerto Sauvedra, Lota and Isla
Juan Fernandez

8.0 CONCLUSIONS

Return periods of tsunamis for various latitudinal belts on the coast of Chile have been
determined based on historical data and comparison was made with return periods for
other seismically active regions of the world.

A comparison of the fundamental periods of the inlets of Chile with those of southeast
Alaska, mainland British Columbia including Puget Sound, and Vancouver Island shows
that all areas have a similar distribution of resonance periods. However, a calculation of
the intensity of secondary undulations shows significant differences, with Chilean and
Alaskan inlets being similar, Vancouver Island inlets being markedly greater and British
Columbia mainland inlets intermediate.

A study of the response of inlets to the 1964 and 1960 tsunamis tend to confirm the
validity of the secondary undulation calculation, but shows major discrepancies in inlets
with side channels and inlets where the one-dimensional approximation may not be valid.
The large range of intensity in secondary undulations is not obvious “a priori” in the
tsunami records.

REFERENCES

Altinok, Y. and S. Erosy (1995). Tsunamis in the Aegean Sea and Near Surroundings,
Proceedings of the International Earth Sciences Colloquim on the Aegean Region, 9-14
October 1995, Izmir-Gulluk, Turkey, edited by O. Piskin, M. Ergun, M.Y. Savascin and
G. Tarcan, Vol. 1, pages 215-226.

Berkman, S.C. and J.M. Symons 1960. The tsunami of May 22, 1960 as recorded at tide
stations. U.S. Coast and Geodetic Survey: 79pp.

Davidson, J. and J. Rynn (1998). Contemporary Assessment of Tsunami Risk and

Implications for Early Warnings for Australia and its Island Territories, Summary Report
for Emergency Management Australia, Australian IDNDR Coordination Committee,
Project 11/94, June 1998, Brisbane, Australia, 25 pages.

Henry, R.F. and T.S. Murty 1972. Resonance periods of multi-branched inlets with
tsunami amplification. Marine Sciences Directorate, Department of Environment,
Ottawa, Canada. Report Series 28, 47-79.

10
Lander, J. and K. O’Laughlin (2000). Personal communication.

Lander, J.F. and P.A. Lockridge (1989). United States Tsunamis (including United States
Possessions) 1690-1988, National Geophysical Data Center, Boulder, Colorado, USA,
265 pages.

Lockridge, P.A. (1985). Tsunamis in Peru-Chile, Report SE-39, World Data Center A
for Solid Earth Geophysics, National Geophysical Data Center, NOAA, Boulder,
Colorado, USA, July 1985, 97p.

Lockridge, P.A. (2000). Personal communication.

Mazova, R.K. and J. Ramirez (1999). Tsunami Waves with an Initial Negative Wave on
the Chilean Coast, Natural Hazards, Vol 20, 83-92.

Murty, T.S. and A. Bapat (1999). Tsunamis on the Coastlines of India, Science of
Tsunami Hazards, Vol. 17, No. 3, 167-172.

Murty, T.S., S.O. Wigen and R. Chawla (1975). Some features of tsunamis on the
Pacific coast of South and North America, manuscript report series, No. 36, Marine
Sciences Directorate, Department of the Environment, Ottawa, Canada, 37 pages.

Nakamura, K. and H. Watanabe 1961. Tsunami forerunner observed in case of the

Chilean tsunami of 1960. Report on the Chilean Tsunami Field Investigation Committee
for the Chilean Tsunami, _______.

Nakano, M. 1932. Preliminary note on the accumulation and dissipation of energy of the
secondary undulations in a bay. Proc. of the Physico-mathematical Soc. of Japan. Vol.
14: 44-56.

NGDC (National Geophysical Data Center) (1999). Tsunami Runup Database, NOAA,
Boulder Colorado, USA

Pickard, G.L. 1961. Oceanogrpahic features of inlets in the British Columbia mainland
coast. J. Fish Res. Bd. Can. 18 (6): 907-999.

Pickard, G.L. 1963. Oceanographic characteristics of inlets of Vancouver Island British

Columbia. J. Fish. Res. Bd. Can. 20 (5): 1109-1144.

Pickard. G.L. 1971. Some physical oceanographic features of inlets of Chile. J. Fish.
Res. Bd. Can 28 (8): 1077-1106.

Ramirez, J., H. Titichoca, J.F Lander and L.S. Whiteside (1997). The Minor Destructive
Tsunami Occurring Near Antofagasta, Northern Chile, July 30, 1995, Science of Tsunami
Hazards, Vol. 15, No. 1, 3-22.

11
Republica De Chile, Departmento De Navegacion E Ilidrografia De La Armada. 1961.
El Maremoto Del 22 De Mayo De 1960 En Las Costos De Chile. Publicacion No. 3012.

Salsman, G.G. 1959. The tsunami of March 9, 1957, as recorded at tide stations U.S.
Coast and Geodetic Survey, Tech. Bull. No. 6: 18 pp.

Spaeth, M.G. and S.C. Berkman. 1967. The tsunami of March 28, 1964, as recorded at
tide stations. U.S. Coast and Geodetic Survey. Tech. Rept. 33: 86 pp.

12
 Figure 1
Geologic Features of Peru-Chile
(from Lockridge, 1985)

13
 Figure 2
From Lockridge (1985)

14
 Figure 3
Inlets of Central Chile

15
 Figure 4
Inlets of Southern Chile

16
 Figure 5

The distribution of the resonant periods of the inlets studied, in minutes.

17
 Figure 6
The distribution of the intensity of secondary undulations

18