Chapter 6

Tsunami Simulations and

Forecasting Systems

6.1 Tsunami Simulations

(1) Predicting tsunami behavior
Tsunami Downloaded from www.worldscientific.com

It would be needless to say that quicker and more accurate tsunami fore-
casting (information) is important for mitigating the loss of life and injury.
If when and where a tsunami will come can be known in advance, tak-
ing appropriate evacuation action would become easy. Conversely, without
this kind of information, mistaken actions can easily occur. Because the
behavior of tsunamis changes in a complex manner due to effect of coastal
topography, numerical models obtained analytically on the basis of simple
assumptions are inadequate. In order to know the actual behavior of a tsu-
nami and estimate the height of the tsunami, it is necessary to predict the
propagation of the tsunami in a way that fully incorporates the seafloor
topography. The only method for achieving this is a simulation using the
actual initial waveform of the tsunami and topographical data. Research
activities which are conscious of tsunami forecasting have been carried out
since around 1983, and the possibility of practical tsunami simulations was
first pointed out after seafloor topographical data was started to be col-
lected off Japan’s Sanriku coast. One work has shown that the maximum
water level distribution can be reproduced with accuracy within an error
of 20% when the rupture process (fault model) in past earthquakes is rela-
tively well understood (Shuto et al., 1988).
b.

Because the basic equation for tsunami simulations is a time-depen-

dent model, the behavior of the tsunami at the necessary time can be
predicted, when the initial conditions are given. It is possible to obtain a
variety of information on the behavior of a tsunami, including its arrival
time at object locations, height of the 1st wave, maximum wave height,

439

10968-master.indd 439 12-02-18 2:21:33 PM

 440 Chapter 6: Tsunami Simulations and Forecasting Systems

duration, velocity field, etc. However, in time-dependent models (par-

ticularly in explicit differential schemes), it is necessary to satisfy stability
conditions, and if these conditions are not satisfied, the simulation will
give unrealistic values. Depending on the selection of the computational
grid (mesh) and the differential scheme, discretization error may occur in
some cases, resulting in reduced predictive accuracy (Imamura et al., 1986;
Imamura and Goto, 1988). However, if the time required for the simula-
tion (numerical analysis) is longer than the time from the generation of a
tsunami to its arrival at the coast, this method is not useful in real-time
prediction of tsunamis when computations begin simultaneously with
the occurrence of an earthquake. In such cases, shortening the computa-
Tsunami Downloaded from www.worldscientific.com

tion time becomes an important study item, on the same level as accuracy.
Another approach to tsunami prediction is to perform computations for a
large number of cases giving initial values in advance, rather than attempt-
ing to predict the tsunami in real time, and to accumulate the results in a
database for future utilization. Because simulations are performed before
an earthquake occurs, this approach relieves the constraint of computa-
tional speed, and there is a plenty of time for calculation. Information on a
variety of behaviors can be obtained by simulations, including the time of
arrival at object areas, the height of the 1st wave, the maximum wave, the
duration of the tsunami, the flow velocity field, and so on. These informa-
tions are quite useful not only for evacuations but also for mitigating vari-
ous tsunami disasters.
On the other hand, the earthquake and tsunami observation network
has gradually been expanded, making it possible to carry out detailed
investigations of the actual condition of tsunamis. As a result, it has gradu-
ally been pointed out that the tsunami wave source estimated from a fault
model by seismic wave analysis does not necessarily coincide with the wave
source estimated from tsunami observation or investigation. Recently, the
importance of dynamic fault motion, which had been ignored until now in
tsunami simulations, has also been pointed out. Furthermore, precise esti-
b.

mation of the tsunami source area in real time is also a difficult task. Study
of a real-time method of tsunami forecasting which integrates simulation
and the tsunami observation network was begun in order to solve these
problems. Provided that the attacking tsunami can actually be captured
directly by offshore observation, it should be possible to use that data as

10968-master.indd 440 12-02-18 2:21:33 PM

 6.1: Tsunami Simulations 441

input information for a reliable tsunami simulation, even when the fault
motion and wave source have some unknown factors, and further even
when it is a tsunami caused by phenomena other than earthquakes. This
requires that numerical forecast is performed simultaneously with real-
time observation using an advanced tsunami monitoring system, where
much more improvement is made with regard to the location of tsunami
meters (GPS buoys), their number, and so forth.

(2) Dissemination of simulation techniques

The activities to provide simulation techniques developed in Japan to other

Tsunami Downloaded from www.worldscientific.com

tsunami-prone countries would bring about significant benefit to these

countries. As efforts during the International Decade for Natural Disaster
Reduction (IDNDR; 1987–1997), the TIME (Tsunami Inundation Model-
ing Exchange) Project was launched with the support of a joint project of
the International Union of Geodesy and Geophysics (IUGG) and UNES-
CO’s Intergovernmental Oceanographic Commission (IOC) (Fig. 6.1).
In the TIME Project, simulation techniques were transferred to countries
which have suffered tsunami damage in the past, or are expected to suf-
fer such damage in the future. The source code was made available, and a
complete manual was prepared and distributed (IOC, 1997). In addition,
local researchers acquired Tohoku University’s tsunami numerical model
code (TUNAMI: Tohoku University Numerical Analysis Modeling for
Inundation), and are inputting topographical and land use data for areas
b.

Fig. 6.1  Logo of the TIME Project

10968-master.indd 441 12-02-18 2:21:33 PM

 442 Chapter 6: Tsunami Simulations and Forecasting Systems

of concern, data on assumed earthquakes, and other information, calculat-

ing expected tsunami inundation, and preparing hazard maps, etc., while
also conducting training and joint research. As of 2007, 39 organizations
from 21 countries have participated in this project. Resolutions expressing
appreciation for this project have repeatedly been passed at various confer-
ences, including those of the IUGG and groups involved in actual tsunami
warning work.

References

Imamura, F. and Goto, C. (1986): Truncation error in numerical simulation by the finite
Tsunami Downloaded from www.worldscientific.com

difference method, Journals of the Japan Society of Civil Engineers, No. 375/II-6,
pp. 241–250. (in Japanese)
Imamura, F. and Goto, C. (1988): Truncation error in numerical simulation by the finite
difference method, Coastal Engineering in Japan, Vol. 31, No. 2, pp. 245–263.
Shuto, N., Goto, C. and Imamura, F. (1988): Use of numerical simulation in tsunami fore-
casting and warning, Journals of the Japan Society of Civil Engineers, No. 393/II-9,
pp. 181–189. (in Japanese)
IOC (1997): Numerical method of tsunami simulation with the leap-frog scheme, IUGG/
IOC TIME Project, Manual and Guides 35, SC-97/WS-37.
b.

10968-master.indd 442 12-02-18 2:21:33 PM

 6.2: Features of  Tsunami Forecasting 443

6.2 Features of Tsunami Forecasting

(1) Necessities of tsunami forecasting

Tsunamis can be classified as “local” or “remote,” depending on the dis-

tance from the source of the tsunami to the coast concerned. This clas-
sification is based on relative distance and is not based on characteristics
of tsunamis themselves. For example, when the tsunami caused by the sea
bottom earthquake off the coast of Chile in South America arrived at the
coast of Chile, it was considered a “local tsunami,” but when the same tsu-
Tsunami Downloaded from www.worldscientific.com

nami propagated across the Pacific Ocean, reaching Japan on the opposite
side, it was a “remote tsunami.” Because local tsunamis occur close to the
focus of the earthquake, the ground motion caused by the earthquake can
frequently be felt by residents along the coast, whereas in remote tsunamis,
it is impossible to feel the ground motion due to the distance from the
center of the earthquake.
In local tsunamis, rapid evacuation is indispensable because the time
from occurrence to arrival of the tsunami is short. Therefore, in forecasts
for local tsunamis, the response from the occurrence of the earthquake to
issuance and transmission of the tsunami forecast must be rapid. The key
point is how to minimize the time required for this.
On the other hand, in the case of remote tsunamis, there is a consider-
able time between the occurrence of the tsunami and its arrival. Therefore,
in forecasts for remote tsunamis, the highest priority is not to shorten the
time required to issue the forecast. Rather, because remote tsunamis prop-
agate over an extremely large area, it is critically important to strengthen
the earthquake observation system and create a network that evaluates the
risk of tsunamis more accurately and based thereon, ensures proper trans-
mission of warnings by cooperation between coastal nations, correspond-
ing to the process of occurrence and propagation of tsunamis.
b.

Because the issues related to tsunami forecasting differ in local and

remote tsunamis, as outlined above, this section will discuss the evolution
of tsunami forecasting systems separately for the two types of tsunamis.
In combination with this, this section will introduce the various
issues that became clear from the viewpoint of tsunami warnings for local

10968-master.indd 443 12-02-18 2:21:33 PM

 444 Chapter 6: Tsunami Simulations and Forecasting Systems

giant tsunamis in the Great East Japan Earthquake Disaster of March 2011,
which caused an unprecedented disaster, and will also introduce the efforts
to improve tsunami warning based on those issues.

(2) Local tsunami forecasting

An organization responsible for tsunami forecasting was created in Japan

in 1941 for the Sanriku coast of northeastern Japan, which had suffered
numerous tsunami attacks in the past. However, tsunami forecasting on a
nationwide scale based on the official legal framework began only in 1952
in the Central Meteorological Observatory (now the Japan Meteorological
Tsunami Downloaded from www.worldscientific.com

Agency). In the initial stage of forecasting work, the arrival times of earth-
quake P and S waves observed at various earthquake monitoring points
were transmitted by telephone or telegram to the central tsunami forecast-
ing office, where the focus and magnitude of the earthquake were deter-
mined manually using maps, charts, pencils, compasses, rulers, and certain
types of calculators. The person in charge then judged whether or not a
tsunami would occur by referring to a “Tsunami forecasting diagram” pre-
pared empirically based on past data in such a way that the possibility of a
tsunami could be judged from the relationship among the intensity of the
earthquake, duration of the initial tremor, and total amplitude. On aver-
age, the time until forecasts were issued was around 17 minutes. Further,
only means available to communicate the resultant judgment to the region
where the tsunami attack was expected were telegram and telephone.
To shorten the time required to issue tsunami forecasts, first, the Japan
Meteorological Agency (JMA) improved the communications network for
collecting earthquake observation data, automated data processing, etc. in
order to speed up the process of determining the focus and magnitude of
earthquakes as far as possible. As a result, the time required for tsunami
forecasts was reduced to approximately 14 minutes in 1980. Three years
later, when the Nihonkai-Chubu Earthquake occurred on May 26, 1983,
b.

the JMA succeeded in issuing a tsunami warning 14 minutes after the

earthquake, as planned. However, the tsunami had already arrived at the
coast before the forecast was issued. The 1st wave attacked only 7 minutes
after the earthquake, and 100 people died in the tsunami.
Therefore, with the aim of issuing forecasts within 7 minutes after

10968-master.indd 444 12-02-18 2:21:33 PM

 6.2: Features of  Tsunami Forecasting 445

an earthquake occurs, the JMA introduced a system that could automati-

cally detect earthquakes and automatically perform the entire series of
processing operations from calculation of the location and magnitude of
the earthquake to judgment of the possibility of a tsunami and issuance
of the judgment. At the same time, tsunami information was also issued
by emergency broadcasting by news organizations. This system actually
operated during the Hokkaido Nansei-oki Earthquake on July 12, 1993,
and issued a tsunami warning in only 5 minutes, which was faster than the
target time. Unfortunately, however, a tsunami with a maximum height of
10 m struck Okushiri Island off the Hokkaido coast only 3 minutes after
the earthquake, killing 142 persons.
Tsunami Downloaded from www.worldscientific.com

Following this disaster, the existing monitoring network was reviewed

in April 1994 targeting a further shortening of the time required to judge
whether a tsunami will occur or not. The tsunami earthquake early detec-
tion network was expanded by installing seismographs at approximately
180 locations throughout Japan, and the time required to issue tsunami
forecasts was reduced to 3 minutes. The current tsunami forecasting system
began operation in April 1999. The tsunami forecast system incorporated
achievements in tsunami propagation simulation techniques in recent
years. Approximately 100,000 different simulations for tsunami generation
and propagation after earthquakes were performed in advance for various
earthquake locations, depths, and magnitudes, and the simulation results
were accumulated in a database. When an earthquake occurs, the system
searches these simulation results for earthquakes which most resemble the
actual earthquake, and announces the height and other behavior of the
tsunami expected in various areas. In comparison with the conventional
technique of prediction from past examples based on empirical rules, this
system uses the results of actual model calculations that cover the entire
process from the occurrence of a tsunami caused by fault motion in an
earthquake to the transmission of the tsunami wave to the coastline. This
has made it possible to divide the forecasting areas into much smaller seg-
b.

ments and describe the height of predicted tsunamis more in detail.

The Japan Meteorological Agency began providing “Earthquake Early
Warnings (EEW)” which precedes the arrival of the main earthquake
motion. This “lead time to strong motion” is expected to be useful infor-
mation for disaster prevention. EEW are prepared by rapidly analyzing the

10968-master.indd 445 12-02-18 2:21:33 PM

 446 Chapter 6: Tsunami Simulations and Forecasting Systems

observation data obtained by seismographs located near the focus of an

earthquake and by estimating the arrival time/seismic intensity in various
areas before the main earthquake motion arrives. In some earthquake, it
has become possible to shorten the time to issuance of a tsunami forecast
to less than 2 minutes at the fastest by using the automatic processing tech-
nology of this EEW system. The tsunami forecasting function has been
operational since October 2006. In the Noto Peninsula Earthquake (M 6.9)
on March 25, 2007, a tsunami forecast was issued 1 min 40 sec after the
earthquake.

(3) Problems of tsunami warnings during the 2011 off the Pacific
Tsunami Downloaded from www.worldscientific.com

Coast of Tohoku Earthquake and improvement of tsunami

warnings

As described above, the quickness of tsunami warning announcement has

increased since the start of tsunami forecasting service. However, in the
2011 off the Pacific Coast of Tohoku Earthquake on March 11, it became
clear that there were a variety of problems with tsunami warnings for a
local giant tsunami (i.e., giant tsunami occurring in nearby waters).
Three minutes after the earthquake occurred, the Japan Meteoro-
logical Agency (JMA) issued tsunami warnings, including major tsunami
warnings for Iwate Pref., Miyagi Pref., and Fukushima Pref. Although the
original magnitude of the earthquake was thought to be M 7.9, the actual
magnitude (moment magnitude, Mw) was 9.0. In the initial tsunami fore-
cast, the tsunami height was seriously underestimated because it was based
on the early value of M 7.9.
The estimated tsunami height was repeatedly revised upward and tsu-
nami warnings were updated based on the rapid rise of the tide level at GPS
buoys (maintained by the Ports and Harbours Bureau, MLIT) installed
10–20 km off the Pacific Coast of Japan’s Tohoku region. Data were also
obtained from ocean-bottom pressure gauge (OBPG) tsunami sensors
b.

(Earthquake Research Institute, The University of Tokyo) off Kamaishi,

which were further from the coast than the GPS buoys and thus were able
to capture the tsunami about 10 min earlier than the buoys. However, the
technology was still inadequate to use the OBPG data in tsunami warn-
ings. Under ordinary conditions, moment magnitude can be calculated

10968-master.indd 446 12-02-18 2:21:34 PM

 6.2: Features of  Tsunami Forecasting 447

approximately 15 min after an earthquake occurs, but calculations were

not possible in this disaster because the giant earthquake exceeded the
capacity of most broadband seismographs in Japan. As a result, it was not
possible to update the tsunami warnings at an earlier stage.
The following actions have been taken to deal with the issues.

i) Regarding the 1st tsunami warning, for technical reasons, it is dif-

ficult to estimate the scale of a giant earthquake exceeding M 8 only
in 3 minutes after the earthquake occurs. Therefore, the JMA intro-
duced monitoring/judgment techniques which make it possible to
recognize whether the initial magnitude has been underestimated.
Tsunami Downloaded from www.worldscientific.com

When there is a possibility of underestimation, tsunami warnings

are issued based on the maximum magnitude assumed in the sea
areas concerned. In this case, the uncertainty of the estimation of
the tsunami scale is large. Therefore, qualitative expressions like
“giant” are used to describe the height of the expected tsunami,
communicating the fact that this is an abnormal situation and is not
like ordinary conditions.
ii) To contribute to appropriate follow-up tsunami warnings, the
moment magnitude should be obtained quickly and accurately, even
in a giant earthquake. Therefore, the JMA is installing new broad-
band seismographs which can measure seismic waves over a wide
frequency range without going off-scale. To prepare for tsunamis
caused by submarine landslides and other phenomena which are
difficult to estimate from earthquake data, application of offshore
ocean-bottom pressure gauges to tsunami prediction is also being
promoted.
iii) In combination with these actions, the content of the information
in tsunami warnings has been changed to urge evacuation more
strongly. The classification of predicted tsunami heights used in
announcements has also been changed, as shown in Fig. 6.2 (revi-
b.

sion of March 2013).

The above has presented an outline of the history of development of

the tsunami forecasting system and the improvements in tsunami warn-
ings based on the 2011 off the Pacific Coast of Tohoku Earthquake. Fig. 6.3

10968-master.indd 447 12-02-18 2:21:34 PM

 448 Chapter 6: Tsunami Simulations and Forecasting Systems

Previous After improvement of tsunami warnings (March 2013)

Classification of Expression
Announced Class of tsunami
warning/advisory Qualitative
tsunami height Numerical value height forecast
expression
Exceeding 10 m Height (h) > 10 m
Major tsunami 10 m or more, 8 m,
10 m Giant 5 m < h ≦10 m
warning 6 m, 4 m, 3 m
5m 3 m < h ≦5 m
Tsunami warning 2 m, 1 m 3m High 1 m < h ≦3 m
Tsunami advisory 0.5m 1m (None) 20 cm < h ≦1 m

Number of classes of tsunami height used in Height of tsunami is expressed by a qualitative Higher value in the forecast class is used as the
tsunami warnings was changed from 8 to 5 (not numerical) expression when there is large forecast tsunami height.
levels. uncertainty in estimation of earthquake scale.

Fig. 6.2  Classification of tsunami warnings and expressions/classes used in tsunami height forecasts
Tsunami Downloaded from www.worldscientific.com

Japan Meteorological Agency meteorological stations Multi-functional transport satellite

(Himawari 7)

Approx. 620 locations

Seismic intensity data
Symbols
Transmission of information

Transmission of data
Seismic intensity data
Data processing/transmission system
Seismic intensity observation facilities ○ Collection of observation data
○ Preparation/announcement of 　
Approx. 240 locations 　 emergency earthquake bulletins
Seismic

Broadcasting stations, etc.

Weather satellite transmissions
waveform

Emergency earthquake

Seismological observation facilities bulletins

Tsunami
Approx. 80 locations wave height
Cities, towns, and villages
Meteorological offices

News organizations, prefectural governments, police,

Tide gauge stations

Seismic intensity data Tsunami warnings/
Tide level/offshore tsunami observation facilities advisories Residents, etc.
Tsunami
wave height
Approx. 3690 locations Tsunami information
Japan Coast Guard, NTT

Seismic intensity observation facilities Earthquake information

Local governments/National Research Institute for Seismic
Earth Science and Disaster Prevention waveform

Approx. 140 locations ○ Check/analysis of observation data

○ PPreparation/announcement of tsunami warnings and advisories
Tide gauge stations ○ Preparation/announcement of earthquake information,
　 tsunami information, etc.

Tide level/offshore tsunami observation facilities Supervised by Japan Meteorological Agency

Universities and
other related institutions
MLIT Ports and Harbours Bureau, Geospatial Information Seismic Intensity SI≧6 lower (in 23 wards of Tokyo Metropolitan area, SI≧5 lower): Meeting of Prime Minister’s Office
Authority of Japan, Japan Coast Guard, National Research emergency team is held.
Institute for Earth Science and Disaster Prevention, universities,
Japan Agency for Marine-Earth Science and Technology, etc. SI≧4: Start of Cabinet Office DIS earthquake early damage estimate.
SI≧4-5: Start of survey of condition of damage by aircraft of Ministry of Defense, Japan Coast Guard,
National Police Agency, etc.
SI≧3: Issuance of bulletins by television and radio

Fig. 6.3  Flow from preparation to transmission of earthquake/tsunami information

shows the flow of preparation/transmission of earthquake/tsunami warn-

ings at present.

(4) Remote tsunami forecasting

b.

A tsunami forecasting system for remote tsunamis was established as a

response to the tsunami following the Aleutian Earthquake near Alaska on
April 1, 1946, which caused heavy damage in the Hawaiian Islands, includ-
ing 173 deaths. Although the tsunami attacked Hawaii approximately

10968-master.indd 448 12-02-18 2:21:35 PM

 6.2: Features of  Tsunami Forecasting 449

5 hours after the earthquake, residents did not feel the earthquake in any
way, and did not know (i.e., were not informed) that a large earthquake had
occurred in the Aleutian Islands. Children playing along the coast sensed
that something was strange about the sea and immediately told their par-
ents, but because it was April Fool’s Day, their parents did not heed the
warning. It is said that this contributed to the heavy loss of life.
As a result of this experience, the United States established the Pacific
Tsunami Warning Center (PTWC) in 1949 as a tsunami warning orga-
nization, centering on the Coast and Geodetic Survey (CGS). However,
because this was a domestic system in the US, there was no cooperation
among the various countries around the Pacific Ocean during the Chilean
Tsunami Downloaded from www.worldscientific.com

Earthquake Tsunami in 1960. As a result, many of the nations around the

Pacific suffered heavy damage, beginning with the Hawaiian Islands and
also including Japan. Based on this, in 1965, the International Tsunami
Information Center (ITIC) was established under UNESCO’s Intergov-
ernmental Oceanographic Commission (IOC) as an organization which
provides information on tsunami warning systems and supports the con-
struction of systems and tsunami disaster mitigation planning in each
nation, with the aim of alleviating damage due to tsunamis in the Pacific
coastal nations. In 1968, the International Coordination Group for the
Tsunami Warning System in the Pacific (ICG/ITSU) was established under
the same UNESCO IOC as a warning organization for remote tsuna-
mis occurring in the Pacific Ocean. (The name of this organization has
now been changed to the Intergovernmental Coordination Group for the
Pacific Tsunami Warning System (ICG/PTWS).) This organization was
established to strengthen the tsunami disaster prevention systems of the
Pacific nations by exchanges and sharing of information on earthquakes
and tsunamis occurring in the Pacific Ocean among the member nations.
The group has 45 countries and districts as its members, including Japan,
as of October 2011. The ICG/ITSU holds meetings every 2 years in order to
coordinate efforts among the member nations and give recommendations
b.

on tsunami warning systems. The previously-mentioned Pacific Tsunami

Warning Center in Hawaii is conducting the operation for actually provid-
ing tsunami information to the member nations.
Following this, a proposal by the Japan Meteorological Agency (JMA)
to establish a tsunami warning center for the northwestern Pacific region

10968-master.indd 449 12-02-18 2:21:35 PM

 450 Chapter 6: Tsunami Simulations and Forecasting Systems

was approved at the 1999 meeting of the ICG/ITSU. Since March 2005, the
JMA has provided “northwestern Pacific tsunami information” to the vari-
ous related countries around the northwestern Pacific Ocean as the North-
west Pacific Tsunami Advisory Center (NWPTAC). NWPTAC provides
information to 11 nations as of August 2012. This “northwestern Pacific
tsunami information” is issued when an earthquake of magnitude 6.5 or
greater occurs in the northwestern Pacific region, and comprises the time
when the earthquake occurred, the location of the earthquake focus, the
magnitude of the earthquake, the estimated possibility of a tsunami, and
when there is a possibility of a tsunami, the estimated arrival time and
predicted height of the tsunami at points on the coast. The “northwest-
Tsunami Downloaded from www.worldscientific.com

ern Pacific tsunami information” issued by the JMA is used by the recipi-
ent nations to take emergency tsunami disaster prevention action, such as
issuing domestic tsunami warnings for the predicted tsunami in the coun-
try concerned, advising residents to evacuate, and similar actions. In this
connection, it may be noted that the tsunami information issued by the
Pacific Tsunami Warning Center does not include the predicted height of
tsunamis.
The flow of the provision of tsunami information in the northwestern
Pacific Ocean and Indian Ocean, for example which the Japan Meteoro-
logical Agency is responsible, as discussed in the following, is shown in
Fig. 6.4.
The Great Indian Ocean Tsunami, which was caused by the 2004
Indian Ocean Earthquake on December 26, 2004, was an unprecedented
disaster, claiming more than 200,000 lives in coastal countries across the
entire Indian Ocean region. As one factor in the large-scale damage, it has
been pointed out that no tsunami warning system like that in the Pacific
Ocean had been created for the Indian Ocean. Utilizing its experience in
issuing tsunami information in the Pacific region, the Japan Meteorological
Agency, in cooperation with the Pacific Tsunami Warning Center, began
providing tsunami monitoring information in March 2005 to nations
b.

around the Indian Ocean which requested this service (information was
being provided to 24 nations as of August 2012). This is being carried out
as a provisional action until an actual Indian Ocean tsunami warning sys-
tem is constructed. “Indian Ocean tsunami information” is issued when
an earthquake of magnitude 6.5 or greater occurs in the Indian Ocean

10968-master.indd 450 12-02-18 2:21:35 PM

 6.2: Features of  Tsunami Forecasting 451

M6.5 or higher
Domestic Japanese earthquake/tsunami observation data Japan Meteorological
Agency (JMA) Northwestern Pacific Northwestern
tsunami information Pacific Ocean and
Indian Ocean countries
Indian Ocean tsunami
monitoring information

• Estimation of earthquake focus and Northwestern Pacific Ocean

magnitude (Kamchatka Peninsula to Solomon Islands)
Overseas seismograph data • Evaluation of possibility of tsunami
• Prediction of tsunami arrival time
• Prediction of tsunami height (only
northwestern Pacific tsunami
information)

Overseas tide level data Information exchange

· Earthquake focus
· Magnitude
· Tsunami information
Tsunami Downloaded from www.worldscientific.com

Indian Ocean: 26 countries

Pacific Tsunami Warning
Center (Hawaii)

Fig. 6.4  Flow of northwestern Pacific tsunami information and Indian Ocean tsunami monitoring
information

region, and comprises the time, location of the focus, and magnitude of the
earthquake, the possibility of a tsunami, and when a tsunami is possible,
the predicted value of the time until arrival of the tsunami at designated
coastlines.
In June 2005, the Intergovernmental Coordination Group for the
Indian Ocean Tsunami Warning and Mitigation System (ICG/IOTWS)
was established with the participation of the Indian Ocean nations, the
members of UNESCO/IOC, and others. ICG/IOTWS started to study for
constructing the Indian Ocean tsunami warning system. Through this
study, operation of the tsunami warning system in the Indian Ocean began
in October 2011. As of August 2012, three organizations, the Bureau of
Meteorology of Australia, the Indian National Centre for Ocean Informa-
tion Services (INCOIS), and Indonesia’s Agency for Meteorology, Clima-
tology and Geophysics (BMKG) were providing information to nations
b.

around the Indian Ocean.

10968-master.indd 451 12-02-18 2:21:36 PM

 452 Chapter 6: Tsunami Simulations and Forecasting Systems

6.3 Future of Tsunami Forecasting Systems

Although the importance of information in disaster prevention is steadily
increasing, the importance of information in connection with tsunamis
is particularly large. As also mentioned at the beginning of this chapter,
although the time margin between an earthquake and a tsunami attack is
small, lives can be saved by evacuation during this interval. On the other
hand, the possibility of a maximum of more than 320,000 victims, if evacu-
ation is delayed, was also been pointed out in a study by the Cabinet Office
of Japan published in August 2012. However, not all earthquakes which are
Tsunami Downloaded from www.worldscientific.com

accompanied by tremors cause tsunamis, and there is a type of earthquake

that causes large tsunamis even though the tremors are small. Thus, it is
difficult for residents to judge that evacuation is necessary based simply
on earthquake tremors. For this reason, the Japan Meteorological Agency
issues tsunami information and predicted values, and this serves as a stan-
dard for judgments by residents to take action to evacuate in case of a tsu-
nami. This is why tsunami information is extremely important.
Up to the present, higher speed has been achieved in monitoring and
prediction, as seen in the JMA’s Earthquake Early Warning (EEW) system.
In Japan, tsunami information can now be issued approximately 3 min-
utes after an earthquake. Although this information system is the most
advanced of its type in the world, challenges still remain on both the infor-
mation providing side and the information receiving side.
First, one issue on the information providing side is how far it is pos-
sible to issue detailed, and highly accurate information. At present, the
country is divided into 66 districts, quantitative information is issued to
each prefectural level government. However, considering the local nature
of tsunamis and the response in the area, this is not adequate. There are
also phenomena that cannot be predicted by the standard techniques. For
example, there are tsunami earthquakes that cause large tsunamis in spite of
b.

producing only small tremors, and there are non-seismic tsunamis which
attack even when no tremors occur in the area, as exemplified by remote
tsunamis. In order to increase accuracy and reliability of techniques for
predicting these phenomena, it is important to monitor actual tsunamis
quickly further out to sea, and to use the results to correct the predicted

10968-master.indd 452 12-02-18 2:21:36 PM

 6.3: Future of  Tsunami Forecasting Systems 453

values. Thus, high expectations are placed on improvement of observation

by the existing monitoring data networks in coastal regions, installation of
GPS tsunami-monitoring buoys, and similar systems.
On the other hand, there are two issues on the information receiving
side. First, there are questions as to whether the need to evacuate can be
judged, and coastal damage can be predicted, from the present tsunami
information (height and arrival time of the tsunami). That is, information
concerning tsunamis is only a necessary condition, but it is not a suffi-
cient condition for disaster mitigation measure and emergency response.
Damage does not necessarily occur simply because a tsunami is large, and
conversely, even small tsunamis sometimes cause serious damage. Damage
Tsunami Downloaded from www.worldscientific.com

can be viewed as the result of the discrepancy between the magnitude of

an external force, that is, a tsunami, and the response capability poten-
tial for preventing damage by that force, or disaster prevention capabili-
ties. In other words, damage cannot be judged from tsunami information
which concerns only external force. It is necessary to fill the gap between
the present tsunami information and damage. For this, an evaluation of
disaster prevention capabilities is necessary. Furthermore, because there is
no time margin to confirm disaster prevention capabilities after an earth-
quake occurred, preparations must be made in advance. Regional inun-
dation maps, disaster prevention maps, and similar tools are necessary
information for bridging this gap. The second issue on the information
receiving side is the strong tendency of each individual to make judgments
about tsunamis based on his or her own personal standards. Even if vari-
ous information is provided, people judge for themselves that they are safe,
and as a result, they fail to take action to avoid disaster. Thus, it is neces-
sary to understand information about tsunamis correctly, and to have a
knowledge of the correct response for disaster mitigation. Each individual
must be able to understand the conditions that are occurring, and must
heighten his or her ability to select the proper response corresponding to
those conditions.
b.

References

Japan Meteorological Agency (2007): Earthquake Early Warning System and Forecasts/
Warnings, http://www.seisvol.kishou.go.jp/eq/EEW/kaisetsu/eew_naiyou/html (accessed
August 22, 2012) (in Japanese)

10968-master.indd 453 12-02-18 2:21:36 PM

 454 Chapter 6: Tsunami Simulations and Forecasting Systems

Japan Meteorological Agency (2011a): Directions for Improvement of Tsunami Warnings

Based on Tsunami Damage in the Off the Pacific Coast of Tohoku Earthquake, http://
www.jma.go.jp/jma/press/1109/12a/tsunami_kaizen_matome.html (accessed August
22, 2012) (in Japanese)
Japan Meteorological Agency (2011b): Recommendations on Criteria for Issuance of Tsu-
nami Warmings, Etc. and Proper Form of Information, http://www.jma/go.jp/jma/
press/1202/07a/tsunami_keihou_teigen.html. (accessed January 10, 2012) (in Japanese)
Kusano, F. and Yokota, T. (2011): History of Tsunami Warning Services in Japan, Quarterly
Journal of Seimology, 74, 35–91. (in Japanese)
Shuto, N., Imamura, F., Koshimura, S., Satake, K. and Matsutomi, H. (2007): Tsunami Ency-
clopedia. Asakura Publishing Co., Ltd., pp. 20–39. (in Japanese)
Tatehata, H. (1998): Application of tsunami numerical simulation technique to tsunami
warning, Kaiyo Monthly, Special Issue No. 15, pp. 23–30. (in Japanese)
Tsunami Downloaded from www.worldscientific.com
b.

10968-master.indd 454 12-02-18 2:21:36 PM