International Tsunami Warning Excercise Pacific Wave 2008

An international exercise to test tsunami warning systems is currently taking place across the Pacific. Exercise Pacific Wave 08 is testing the distribution of tsunami advisories (bulletins, watches and warnings), by the international and national centres, including the Joint Australian Tsunami Warning Centre (JATWC).

The exercise scenario uses a fictional Tsunami from a magnitude 9.2 earthquake near Japan on October 29, 2008 at 0000UTC. There is a carefully scripted list of events scheduled at specific times in a detailed 132 page Exercise Pacific Wave 08 Technical Bulletin. The exercise materials are supposed to be marked "For Exercise Purposes Only" to distinguish them from real messages and if there is a major real event the exercise will be cancelled.

Pacific countries are being encouraged to exercise their own internal tsunami warning and emergency response procedures. While messages are issued by the international warning centres, it can take a long time for them to reach the population.

Dangerous Lack of Standardisation in Tsunami Message Format

While the exercise has been carefully planned, one problem with the warning systems remain: there is no standard international format for tsunami warning messages. The formats used for tsunami warnings predate the Internet and were designed for devices such as Teletype machines. As a result there are incompatibilities between the formats and Internet technology which could cause dangerous miscommunication. In addition there is no rigorously defined format for messages.

As an example the exercise document states that "All documentation and correspondence relating to this exercise is to be clearly identified as Exercise Pacific Wave 08 and For Exercise Purposes Only." However, only the first two of the supplied exercise messages include the text "Exercise Pacific Wave 08" and none of the messages include "For Exercise Purposes Only". Such imprecision can cause dangerous confusion, with the exercise messages being mistaken for real ones and real messages for exercises.

International tsunami authorities need to agree one standard format for tsunami messages which has a precisely defined format and wording and is compatible both with the Internet and backward compatible with older communication devices. Failure to do so could result in the loss of hundreds of thousands of lives.

The initial bulletin will be issued by the NWPTAC because the earthquake is located in the immediate vicinity of Japan. Initial bulletins from PTWC and WC/ATWC will follow, initially using the earthquake parameters from the NWPTAC. To avoid any possible misinterpretation, bulletins issued by the warning centers will be in a “dummy” exercise message format (Appendix I) that will refer participants to a specific scenario bulletin number in this exercise manual (in Appendices II – IV). Dummy messages will be issued for each simulated real message at the beginning of the exercise, but later PTWC and WC/ATWC dummy messages will be issued only once every four hours until the simulated tsunami has crossed the entire Pacific and the exercise concludes. ...

TSUNAMI BULLETIN NUMBER 001
PACIFIC TSUNAMI WARNING CENTER/NOAA/NWS
ISSUED AT 0010Z 29 OCT 2008
THIS BULLETIN APPLIES TO AREAS WITHIN AND BORDERING THE PACIFIC
OCEAN AND ADJACENT SEAS...EXCEPT ALASKA...BRITISH COLUMBIA...
WASHINGTON...OREGON AND CALIFORNIA.
... A TSUNAMI WARNING AND WATCH ARE IN EFFECT ...
A TSUNAMI WARNING IS IN EFFECT FOR
JAPAN / RUSSIA / MARCUS IS.
A TSUNAMI WATCH IS IN EFFECT FOR
N. MARIANAS / GUAM / WAKE IS. / CHINESE TAIPEI / TAIWAN / YAP /
PHILIPPINES / MARSHALL IS. / CHUUK / MIDWAY IS. / POHNPEI /
BELAU / KOSRAE / INDONESIA / PAPUA NEW GUINEA / HAWAII
FOR ALL OTHER AREAS COVERED BY THIS BULLETIN... IT IS FOR
INFORMATION ONLY AT THIS TIME.
THIS BULLETIN IS ISSUED AS ADVICE TO GOVERNMENT AGENCIES. ONLY
NATIONAL AND LOCAL GOVERNMENT AGENCIES HAVE THE AUTHORITY TO MAKE
DECISIONS REGARDING THE OFFICIAL STATE OF ALERT IN THEIR AREA AND
ANY ACTIONS TO BE TAKEN IN RESPONSE.
AN EARTHQUAKE HAS OCCURRED WITH THESE PRELIMINARY PARAMETERS
ORIGIN TIME - 0000Z 29 OCT 2008
COORDINATES - 40.0 NORTH 143.0 EAST
DEPTH - 33 KM
LOCATION - OFF EAST COAST OF HONSHU JAPAN
MAGNITUDE - 8.5
EVALUATION
IT IS NOT KNOWN THAT A TSUNAMI WAS GENERATED. THIS WARNING IS
BASED ONLY ON THE EARTHQUAKE EVALUATION. AN EARTHQUAKE OF THIS
SIZE HAS THE POTENTIAL TO GENERATE A DESTRUCTIVE TSUNAMI THAT CAN
STRIKE COASTLINES NEAR THE EPICENTER WITHIN MINUTES AND MORE
DISTANT COASTLINES WITHIN HOURS. AUTHORITIES SHOULD TAKE
APPROPRIATE ACTION IN RESPONSE TO THIS POSSIBILITY. THIS CENTER
WILL MONITOR SEA LEVEL DATA FROM GAUGES NEAR THE EARTHQUAKE TO
DETERMINE IF A TSUNAMI WAS GENERATED AND ESTIMATE THE SEVERITY OF
THE THREAT.
ESTIMATED INITIAL TSUNAMI WAVE ARRIVAL TIMES AT FORECAST POINTS
WITHIN THE WARNING AND WATCH AREAS ARE GIVEN BELOW. ACTUAL
ARRIVAL TIMES MAY DIFFER AND THE INITIAL WAVE MAY NOT BE THE
LARGEST. A TSUNAMI IS A SERIES OF WAVES AND THE TIME BETWEEN
SUCCESSIVE WAVES CAN BE FIVE MINUTES TO ONE HOUR.
LOCATION FORECAST POINT COORDINATES ARRIVAL TIME
-------------------------------- ------------ ------------
JAPAN HACHINOHE 40.5N 141.7E 0033Z 29 OCT
IOC Technical Series No. 82 Appendix II
page 12
KUSHIRO 42.9N 144.3E 0043Z 29 OCT
KATSUURA 35.0N 140.3E 0102Z 29 OCT
SHIMIZU 32.8N 133.0E 0225Z 29 OCT
OKINAWA 26.2N 127.8E 0319Z 29 OCT
RUSSIA URUP IS 46.1N 150.5E 0118Z 29 OCT
PETROPAVLOVSK K 53.2N 159.6E 0226Z 29 OCT
SEVERO KURILSK 50.8N 156.1E 0229Z 29 OCT
UST KAMCHATSK 56.1N 162.6E 0248Z 29 OCT
MEDNNY IS 54.7N 167.4E 0249Z 29 OCT
MARCUS IS. MARCUS IS. 24.3N 154.0E 0225Z 29 OCT
N. MARIANAS SAIPAN 15.3N 145.8E 0326Z 29 OCT
GUAM GUAM 13.4N 144.7E 0342Z 29 OCT
WAKE IS. WAKE IS. 19.3N 166.6E 0350Z 29 OCT
CHINESE TAIPEI HUALIEN 24.0N 121.7E 0403Z 29 OCT
TAITUNG 22.7N 121.2E 0407Z 29 OCT
CHILUNG 25.2N 121.8E 0430Z 29 OCT
KAOHSIUNG 22.5N 120.3E 0442Z 29 OCT
TAIWAN HUALIEN 24.0N 121.6E 0404Z 29 OCT
YAP YAP IS. 9.5N 138.1E 0417Z 29 OCT
PHILIPPINES PALANAN 17.1N 122.6E 0425Z 29 OCT
LAOAG 18.2N 120.5E 0445Z 29 OCT
SAN FERNANDO 16.7N 120.2E 0504Z 29 OCT
LEGASPI 13.2N 123.8E 0507Z 29 OCT
DAVAO 6.8N 125.7E 0521Z 29 OCT
ZAMBOANGA 6.9N 122.1E 0605Z 29 OCT
MARSHALL IS. ENIWETOK 11.4N 162.3E 0428Z 29 OCT
KWAJALEIN 8.7N 167.7E 0511Z 29 OCT
MAJURO 7.1N 171.4E 0540Z 29 OCT
CHUUK CHUUK IS. 7.4N 151.8E 0432Z 29 OCT
MIDWAY IS. MIDWAY IS. 28.2N 177.4W 0443Z 29 OCT
POHNPEI POHNPEI IS. 7.0N 158.2E 0444Z 29 OCT
BELAU MALAKAL 7.3N 134.5E 0446Z 29 OCT
KOSRAE KOSRAE IS. 5.5N 163.0E 0511Z 29 OCT
INDONESIA GEME 4.6N 126.8E 0519Z 29 OCT
BEREBERE 2.5N 128.7E 0529Z 29 OCT
WARSA 0.6S 135.8E 0545Z 29 OCT
MANOKWARI 0.8S 134.2E 0552Z 29 OCT
PATANI 0.4N 128.8E 0554Z 29 OCT
MANADO 1.6N 124.9E 0601Z 29 OCT
SORONG 0.8S 131.1E 0604Z 29 OCT
JAYAPURA 2.4S 140.8E 0607Z 29 OCT
PAPUA NEW GUINE KAVIENG 2.5S 150.7E 0556Z 29 OCT
MANUS IS. 2.0S 147.5E 0556Z 29 OCT
VANIMO 2.6S 141.3E 0607Z 29 OCT
BULLETINS WILL BE ISSUED HOURLY OR SOONER IF CONDITIONS WARRANT.
THE TSUNAMI WARNING AND WATCH WILL REMAIN IN EFFECT UNTIL
FURTHER NOTICE.
THE JAPAN METEOROLOGICAL AGENCY MAY ALSO ISSUE TSUNAMI MESSAGES
FOR THIS EVENT TO COUNTRIES IN THE NORTHWEST PACIFIC AND SOUTH
CHINA SEA REGION. IN CASE OF CONFLICTING INFORMATION... THE
MORE CONSERVATIVE INFORMATION SHOULD BE USED FOR SAFETY.
THE WEST COAST/ALASKA TSUNAMI WARNING CENTER WILL ISSUE PRODUCTS
FOR ALASKA...BRITISH COLUMBIA...WASHINGTON...OREGON...CALIFORNIA. ...

From: Exercise Pacific Wave 08 Technical Bulliten, Commission océanographique intergouvernementale. Exercise Pacific Wave 08. A Pacific-wide Tsunami Warning and Communication Exercise Pacific, 28–30 October 2008. IOC Technical Series No. 82. Paris, UNESCO, 2008. (English)

Standard format needed for tsunami warning messages

This is to suggest that one global format is needed for messages issued by tsunami warning systems, along with a specific set of words to be used in them. The Pacific and Interim Indian Ocean systems use slightly different formats and words for their warning messages. These differences could result in dangerous misunderstandings by computer software and human readers.

As an example, Pacific warnings use dashes "-" to separate parameters, whereas Indian Ocean warnings use colons ":". Also the Pacific warnings say: "NO DESTRUCTIVE PACIFIC-WIDE TSUNAMI THREAT EXISTS BASED ON
HISTORICAL EARTHQUAKE AND TSUNAMI DATA", whereas the equivalent Indian Ocean message is "THERE IS A VERY SMALL POSSIBILITY OF A DESTRUCTIVE LOCAL TSUNAMI ".

From Tsunami bulletin number 001, issued by the Pacific Tsunami Warning Centre/NOAA/NWS at 0628Z 17 JUN 2005:

AN EARTHQUAKE HAS OCCURRED WITH THESE PRELIMINARY PARAMETERS

ORIGIN TIME - 0622Z 17 JUN 2005
COORDINATES - 40.6 NORTH 126.3 WEST
LOCATION - OFF COAST OF NORTHERN CALIFORNIA
MAGNITUDE - 6.9

EVALUATION

NO DESTRUCTIVE PACIFIC-WIDE TSUNAMI THREAT EXISTS BASED ON
HISTORICAL EARTHQUAKE AND TSUNAMI DATA. ...

From Tsunami bulletin number 001, issued by the Japan Meteorological Agency (JMA) at 1550 16 MAY 2006 (UTC):

1.EARTHQUAKE INFORMATION
ORIGIN TIME : 1528 16 MAY 2006 (UTC)
COORDINATES : 0.1 NORTH 97.0 EAST
LOCATION : NORTHERN SUMATRA, INDONESIA
MAGNITUDE : 6.9

2.EVALUATION
THERE IS A VERY SMALL POSSIBILITY OF A DESTRUCTIVE LOCAL
TSUNAMI IN THE INDIAN OCEAN. ...

I noticed the differences when attempting to forward the messages to a mobile telephone using SMS. The SMS system can only send short messages. So I had my mail system search for a key phrase in the message. This worked with "LOCATION", as this is common to Pacific and Indian ocean messages. However, if I attempted to have the system look for the punctuation after the word "LOCATION" the problem becomes much more difficult, as it has to allow for either a dash or a colon.

Also there does not appear an easy way for a computer program, or a human reader, to look at a message and quickly determine if it is for a destructive tsunami, or not. All the messages use the word "destructive". It would seem preferable to phrase the messages so they indicate what is predicted to happen, not what is predicted to not happen.

It would aid reading by both human operators and machines interpreting the information if the same format and wording was used by all international tsunami warning centres. This could be done by agreeing a specification for the format of the warning message and a controlled vocabulary to be used in the messages.

It might be argued that in the long term the messages should be formatted in a precise computer readable format, such as Common Alerting Protocol, or that the tsunami warnings are only intended for trained emergency staff not the general public. However, for the foreseeable future there will be a need to communicate tsunami warnings in natural language and so it would be desirable if that language was as precise and standardised as possible.

It should be noted that SMS is not necessarily the best technology for emergency communication and there are other options for Wireless Internet for Emergencies.

Wobbling cyber chair only earthquake warning

While I had given a talk to a earthquake warning center a few days ago, the only warning that an earthquake was happening here in Greece was when the chair I was sitting on started to wobble. At the time I was in Athens, using a computer in the cyber-cafe of the hotel I am staying at.

The chair I was sitting on seemed to be wobbling from side to side. Given it was made of solid steel, after a few seconds I realised we must be having an earthquake. This lasted a few seconds and as there was no damage, I thought no more about it. It was only later that I saw media reports of the earthquake killing two south west of Athens.

In my seminar to the Kandilli Observatory and Earthquake Research Institute, Turkey, on Internet based systems for emergency warning.I warned that t6he region is at risk as the Mediterranean and connected Seas Tsunami Warning System is still under construction. It is fortunate that the Greek earthquake did not involve a tsunami.

It is actually possible to issue a warning of earthquakes, but only a few seconds in advance. Such warnings are still useful, sent out electronically to switch off equipment automatically, which might otherwise be damaged or cause injury.

ATHENS: An earthquake with a preliminary magnitude of 6.5 rattled southern Greece on Sunday, killing two people, injuring dozens more and damaging scores of homes, officials said.

The earthquake, which struck at 3:25 p.m., was centered near the northern Peloponnesian city of Patras, about 120 miles west of Athens, officials said. Much of the damage occurred around Andravida, about 40 miles southwest of Patras.

The Greek interior minister, Prokopis Pavlopoulos, said a 43-year-old man was killed by a falling roof in a village near Andravida. Another man died of a heart attack while being taken to a hospital for treatment of injuries, Pavlopoulos said.

At least four people were treated for leg injuries after leaping from the upper floors of their shaking homes, said Panayiotis Efstathiou, a regional health official. ...

From: Earthquake kills 2 in southern Greece, International Herald Tribune, June 9, 2008

Kandilli Observatory and Earthquake Research Institute

Greetings from the Kandilli Observatory and Earthquake Research Institute, Turkey, I am here to give a Seminar on Internet based systems for emergency warning. The institute is the tsunami watch center for the Eastern Mediterranean region, as a part of ICG/North East Atlantic and connecting seas tsunami warning system.

Internet based systems for emergency warning, Seminar, Turkey, 23 May 2008

I will be giving a seminar on Internet based systems, for emergency warning and disaster recovery, at Kandilli Observatory and Earthquake Research Institute, Bogazici University, Turkey on 23 May 2008. The institute is the tsunami watch center for the Eastern Mediterranean region, as a part of ICG/North East Atlantic and connecting seas tsunami warning system. Visitors are welcome to attend the seminar:

Seminar

Kandilli Observatory and Earthquake Research Institute, Bogazici University, Turkey

Friday, 23rd of May 2008 at 15:30

Topic: Internet based systems, for emergency warning and disaster recovery

Speaker: Tom Worthington FACS HLM, Australian National University and the Australian Computer Society

This seminar will discuss the role of the Internet and the web in emergency warning systems, particularly for Tsunami and for disaster recovery. It looks at the optimization of the design of web based systems for emergencies, including compatibility with mobile telephones. The speaker argues that Internet and web technologies can be used for communication of emergency information to the public, but will require a change in message formats previously intended for emergency workers.

About the speaker:

Tom Worthington is a member of the Project Management Committee of Sahana open source disaster management system used for the 2004 Indian Ocean tsunami. At the Australian National University he teaches the design of web based systems. As Director of Professional Development of the Australian Computer Society, responsible for our e-learning programs.

See:

• Tom Worthington <https://cs.anu.edu.au/people.php?StaffID=140274>
• Optimising Sahana <http://www.tomw.net.au/2005/wd/sahana3.shtml>
• Tsunami warning systems: <http://www.tomw.net.au/technology/it/webethics.shtml#tsunami>
• Australian Computer Society e-learning: <http://www.acs.org.au/cpeprogram/>.

Mediterranean Tsunami Warning System Needed

After the 2004 Indian Ocean Tsunami the United Nations coordinated the establishment of an Indian Ocean Tsunami Warning System. Less well known (and less well established) is the North Eastern Atlantic, the Mediterranean and connected Seas Tsunami Warning System (NEAMTWS). The need for such a facility became more urgent with recent research from the University of Cambridge showing that Tsunami in the Mediterranean of the magnitude of the 2004 Indian Ocean Tsunami are far more frequent than previously thought.

The NEAMTWS is intended to alert Europe, North African and Middle Eastern countries. As with the Indian Ocean system, there is an email list provided by UNESCO which can be subscribed to for receiving warnings. However, given the small size of areas such as the Mediterranean, the warning time for a Tsunami could be very short and faster means of communication should be used by government and official emergency management organizations.

There appears to be a dangerious lack of progress with the NEAMTWS . The latest meeting was 21 – 23 November 2007. The meeting was to consider Version 3.2 of the implementation plan (summary appended). However, all the meeting report says is: "Please check back here in the next few days for the meeting summary report.".

The plan appears to be a well researched, but it is no substitute for action. Professionals involved, including ICT professionals, need to keep in mind that failure to act on such a clear and present danger is unethical, and may in addition expose them to civil and criminal prosecution in the event of a disaster.

The complex system of regional, national and local systems does not make for an effective system and may need to be rethought. A system which sends warnings directly from those detecting them, to the citizens may be better. The bureaucrats and politicians who would object to this need to keep in mind they may face a lynch mob in the event of large loss of life.

This Implementation Plan for the North-East Atlantic, the Mediterranean and Connected Seas Tsunami Warning and Mitigation System (NEAMTWS) specifies detailed requirements of the design and implementation of the tsunami warning and mitigation system for the North-East Atlantic, the Mediterranean and Connected Seas. Following its initial phase, the warning system is intended to apply to other marine-related hazards in addition to tsunamis, in particular storm surges, both causing i.e coastal inundation. As the requirements are developing and therefore still are subject to changes or modifications, and as the implementation will progress, this Implementation Plan is a dynamic document. In constant use and development, it will only represent the status of the system at a specific time of viewing. As a living document it will be available on the Intergovernmental Oceanographic Commission (IOC) web site and subsequent versions will be distributed at Intergovernmental Coordination Group (ICG)/NEAMTWS meetings.

The Implementation Plan is structured to reflect the ICG and its Working Groups, WG 1 (Hazard Assessment, Risk and Modelling), WG 2 (Seismic and Geophysical Measurements), WG 3 (Sea Level Data Collection and Exchange, including Offshore Tsunami Detection and Instruments), and WG 4 (Advisory, Mitigation and Public Awareness).

After a status summary, details are condensed in Action Plans for all components of the system. Capacity building is explicitly addressed to highlight the importance of training and extend the basis of the people involved in operating the system at all levels. Reflecting the work as its progresses there are parts that are not yet as detailed as required.

The Implementation Plan is only one of the documents that describe the NEAMTWS and help in managing it. Others are, or will be available.

The NEAMTWS is a complex operation owned and operated by Member States through their designated agencies. Besides the national functions these agencies serve as conduits for information within the system that is amongst all participating partners. These are further augmented into international, mostly regional, functions that serve an agreed regional ensemble of member states. These functions need particular attention for the system to perform as a whole.

The performance of the NEAMTWS depends on the implementation of all its components, their sustained operation and the adherence to agreed common principles of operation, interaction and data policy. This performance needs to be monitored in order to improve the NEAMTWS, identify deficiencies and suggest remedial action. A real test of the NEAMTWS may happen only rarely. But it will then highlight the credibility of the system and all its participating partners. The public will only judge the performance or the success of the system from the impact, the loss of lives and the damages that occurred.

The timely and appropriate implementation of the NEAMTWS is crucial to its success. Recent events have shown that time to prepare, implement and train is short as the events are unpredictable. In its implementation priorities, requirements and details will change or have to be adapted. Member States are therefore asked to prioritize details they feel need special attention on both the national and international level. They also are invited to provide guidance as to further developing governance mechanisms for the NEAMTWS.

From: Implementation Plan, North-East Atlantic, the Mediterranean and Connected Seas Tsunami Warning and Mitigation System, NEAMTWS,
Version 3.2, Intergovernmental Oceanographic Commission Technical Series 73, 24 July 2007

Tsunami Warning System worked for latest Indonesian Earthquake

The Pacific Tsunami Warning Center issued a local tsunami watch for Indonesia 0822Z 20 FEB 2008. The copy forwarded by the interim Indian Ocean Center is timed at one minute later 08:23:21 GMT. The watch was canceled at 0947Z 20 FEB 2008.

The Pacific Tsunami Warning Center's system provided a map and details of the earthquake.

PACIFIC TSUNAMI WARNING CENTER/NOAA/NWS
ISSUED AT 0822Z 20 FEB 2008

THIS BULLETIN IS FOR ALL AREAS OF THE INDIAN OCEAN.

... A LOCAL TSUNAMI WATCH IS IN EFFECT ...

A TSUNAMI WATCH IS IN EFFECT FOR

INDONESIA

FOR OTHER AREAS OF THE INDIAN OCEAN REGION...THIS MESSAGE IS FOR
INFORMATION ONLY AT THIS TIME.

THIS BULLETIN IS ISSUED AS ADVICE TO GOVERNMENT AGENCIES.  ONLY
NATIONAL AND LOCAL GOVERNMENT AGENCIES HAVE THE AUTHORITY TO MAKE
DECISIONS REGARDING THE OFFICIAL STATE OF ALERT IN THEIR AREA AND
ANY ACTIONS TO BE TAKEN IN RESPONSE.

AN EARTHQUAKE HAS OCCURRED WITH THESE PRELIMINARY PARAMETERS

ORIGIN TIME -  0809Z 20 FEB 2008
COORDINATES -   2.8 NORTH   96.0 EAST
LOCATION    -  OFF W COAST OF NORTHERN SUMATRA
MAGNITUDE   -  7.2

EVALUATION

A DESTRUCTIVE WIDESPREAD TSUNAMI THREAT DOES NOT EXIST BASED ON 
HISTORICAL EARTHQUAKE AND TSUNAMI DATA.                         
                                                             
HOWEVER - THERE IS THE POSSIBILITY OF A LOCAL TSUNAMI THAT      
COULD AFFECT COASTS LOCATED USUALLY NO MORE THAN A HUNDRED      
KILOMETERS FROM THE EARTHQUAKE EPICENTER. AUTHORITIES FOR THE   
REGION NEAR THE EPICENTER SHOULD BE AWARE OF THIS POSSIBILITY.  
AREAS FURTHER FROM THE EPICENTER COULD EXPERIENCE SMALL SEA     
LEVEL CHANGES AND STRONG OR UNUSUAL COASTAL CURRENTS.           

DUE TO ONLY LIMITED SEA LEVEL DATA FROM THE REGION IT MAY NOT
BE POSSIBLE FOR THIS CENTER TO RAPIDLY CONFIRM NOR EVALUATE THE
STRENGTH OF A TSUNAMI IF ONE HAS BEEN GENERATED.

ESTIMATED INITIAL TSUNAMI WAVE ARRIVAL TIMES AT FORECAST POINTS
WITHIN THE WARNING AND WATCH AREAS ARE GIVEN BELOW. ACTUAL
ARRIVAL TIMES MAY DIFFER AND THE INITIAL WAVE MAY NOT BE THE
LARGEST. A TSUNAMI IS A SERIES OF WAVES AND THE TIME BETWEEN
SUCCESSIVE WAVES CAN BE FIVE MINUTES TO ONE HOUR.

LOCATION         FORECAST POINT     COORDINATES     ARRIVAL TIME
--------------------------------    ------------    ------------
INDONESIA        SIMEULUE            2.5N  96.0E    0827Z 20 FEB

THIS WILL BE THE ONLY BULLETIN ISSUED BY THE PACIFIC TSUNAMI
WARNING CENTER FOR THIS EVENT UNLESS ADDITIONAL INFORMATION
BECOMES AVAILABLE.

THE JAPAN METEOROLOGICAL AGENCY MAY ISSUE ADDITIONAL INFORMATION
FOR THIS EVENT. IN THE CASE OF CONFLICTING INFORMATION...
TSUNAMI BULLETIN NUMBER 001

MORE CONSERVATIVE INFORMATION SHOULD BE USED FOR SAFETY.

Tsunami Warning Spam

The Pacific Tsunami Warning Center Issued a Tsunami Watch for Indonesia at 2114z 24 Oct 2007. Fortunately there was no Tsunami, but the warning message was rated 1.4 by the Spam Assen Spam filter. As I noted previously, there is a risk such messages may be blocked in transmission by Spam filters. The format of such messages need to be changed to make them look less like spam. This could be done at the same time as other changes to make the messages more readable.

Subject: {SPAM 01.4} [Tsunami Message - IOC] Local Tsunami Watch Bulletin
X-Spam: spam
X-Spam-score: 1.4
X-Spam-hits: BAYES_00 -2.599, DCC_CHECK 2.5, UPPERCASE_75_100 1.528, BAYES_USED user
Date: Wed, 24 Oct 2007 21:16:09 GMT
To: iobulletin@ptwc.noaa.gov
From: tsunami-information-ioc@lists.unesco.org
Message-id:
X-Spam-orig-subject: [Tsunami Message - IOC] Local Tsunami Watch Bulletin
Errors-To: tsunami-information-ioc-owner@lists.unesco.org
Precedence: list
X-no-archive: yes
List-Id:
List-Archive: <https://lists.unesco.org/wws/arc/tsunami-information-ioc>
List-Help: <mailto:sympa@lists.unesco.org?subject=help>
List-Owner: <mailto:tsunami-information-ioc-request@lists.unesco.org>
List-Post: <mailto:tsunami-information-ioc@lists.unesco.org>
List-Subscribe: <mailto:sympa@lists.unesco.org?subject=subscribe%20tsunami-information-ioc>
List-Unsubscribe: <mailto:sympa@lists.unesco.org?subject=unsubscribe%20tsunami-information-ioc>
X-Virus-Scanned: by amavisd-new-20030616-p10 (Debian) at ans.com.au

TSUNAMI BULLETIN NUMBER 001
PACIFIC TSUNAMI WARNING CENTER/NOAA/NWS
ISSUED AT 2114Z 24 OCT 2007

THIS BULLETIN IS FOR ALL AREAS OF THE INDIAN OCEAN.

... A LOCAL TSUNAMI WATCH IS IN EFFECT ...

A TSUNAMI WATCH IS IN EFFECT FOR

INDONESIA

FOR OTHER AREAS OF THE INDIAN OCEAN REGION, THIS MESSAGE IS AN
ADVISORY ONLY.

AN EARTHQUAKE HAS OCCURRED WITH THESE PRELIMINARY PARAMETERS

ORIGIN TIME - 2103Z 24 OCT 2007
COORDINATES - 4.1 SOUTH 100.8 EAST
LOCATION - SOUTHWEST OF SUMATRA INDONESIA
MAGNITUDE - 7.0

EVALUATION

A DESTRUCTIVE WIDESPREAD TSUNAMI THREAT DOES NOT EXIST BASED ON
HISTORICAL EARTHQUAKE AND TSUNAMI DATA.

HOWEVER - THERE IS THE POSSIBILITY OF A LOCAL TSUNAMI THAT
COULD AFFECT COASTS LOCATED USUALLY NO MORE THAN A HUNDRED
KILOMETERS FROM THE EARTHQUAKE EPICENTER. AUTHORITIES FOR THE
REGION NEAR THE EPICENTER SHOULD BE AWARE OF THIS POSSIBILITY.
AREAS FURTHER FROM THE EPICENTER COULD EXPERIENCE SMALL SEA
LEVEL CHANGES AND STRONG OR UNUSUAL COASTAL CURRENTS.

DUE TO ONLY LIMITED SEA LEVEL DATA FROM THE REGION IT MAY NOT
BE POSSIBLE FOR THIS CENTER TO RAPIDLY CONFIRM NOR EVALUATE THE
STRENGTH OF A TSUNAMI IF ONE HAS BEEN GENERATED.

ESTIMATED INITIAL TSUNAMI WAVE ARRIVAL TIMES. ACTUAL ARRIVAL TIMES
MAY DIFFER AND THE INITIAL WAVE MAY NOT BE THE LARGEST. THE TIME
BETWEEN SUCCESSIVE TSUNAMI WAVES CAN BE FIVE MINUTES TO ONE HOUR.

LOCATION COORDINATES ARRIVAL TIME
-------------------------------- ------------ ------------
INDONESIA BENGKULU 3.9S 102.0E 2134Z 24 OCT

THIS WILL BE THE ONLY BULLETIN ISSUED BY THE PACIFIC TSUNAMI
WARNING CENTER FOR THIS EVENT UNLESS ADDITIONAL INFORMATION
BECOMES AVAILABLE.

THE JAPAN METEOROLOGICAL AGENCY MAY ISSUE ADDITIONAL INFORMATION
FOR THIS EVENT.

Joint Australian Tsunami Warning Centre

Daniel Jaksa, Operations Team Leader of the Joint Australian Tsunami Warning Centre (JATWC) gave a presentation at the Geoscience Australia Open day in Canberra Sunday 26 August 2007 entitled "Earthquakes to Evacuate".

Boxing Day Tsunami

Geoscience Australia is part of the Australian and international tsunami warning systems. Daniel played a sound recording of the 26 December 2004 earthquake, from a hydrophone at Diego Garcia, 2,600 kilometers from the epicenter. This was a deep rubling which seemed to go on forever.

The tsunami caused by the earthquake hit Indonesia 15 minutes later. Daniel showed video of the tsunami in Phuket, Thailand, 30 minutes after the first wave. He pointed out that the first wave is not necessarily the largest. The Tsunami stuck Kenya eight hours later. In total 230,000 people were killed and 45,000 missing.

In response the Australian government provided $68.9 M for a four year program to establish a warning system. Involved are GA, Australian Bureau of Meteorology (BoM), Emergency Management Australia (EMA), and the State Emergency Services (SES).

A tsunami with a 9m runup hit Steep Point WA at about 7:30pm. This destroyed a campsite, but the family were able to hold onto their car and avoid being swept out to sea.

Establishing an Australian Tsunami Warning Centre

The Joint Australian Tsunami Warning Centre (JATWC) is jointly run by GA (earthquake detection), and Bureau of Meteorology, (sea level monitoring). The Operations Hub for GA's part of AusTWC was officially opened on 1 December 2006 in Canberra

GA detects earthquakes, evaluates them for tsunamigenic potential, passes the information on to BoM and evaluates the results. The GA person on duty (called the "watch stander") evaluates an earthquake within five to ten minutes of detection. The watch stander presses a red button to issue an electronic alert to BoM. The BoM only have one tsunami buoy currently (others being installed) , plus tide gauges to detect a tsunami. Tsunami propagation models are used to evaluate the size and timing of a potential tsunami.

Daniel showed news reports from North Queensland of evacuations due to an earthquake in the Solomon Islands 2 April 2007 and later beach closures along the east coast of Australia. He showed a simulation of how an earthquake in New Zealand could cause a tsunami inundating Wollogong. This was a chilling demonstration. But the high risk areas in Australia are the North West shelf and around Hobart.

Dangerous Confusion in Australian Tsunami Warning System

The AusTWC computer system at GA is mirrored at the BoM, in case of a failure at GA in Canberra. Also other warning centers around the region can assist if the Australian system is not available. However, the major problem with the Australian system is not the technical design, or competence of the scientific staff, but the lack of effort put into communicating warnings to the public.

Like other regional warning centers, the JATWC does not directly issue public alerts, this is left to state governments via EMA. As a result there is a potential delay and lack of standard terminology in the messages issued. With the most recent Solomon Islands tsunami warning there was a lack of preparedness, particularly by the Queensland government. Resulting delays and confusion could result in a large loss of life in a future tsunami.

There is a lack of standard terminology in the Australian approach to tsunami warning. As an example the BoM refer to the Australian Tsunami Warning System (ATWS), Australian Tsunami Alert System (ATAS), and the Joint Australian Tsunami Warning Centre (JATWC). GA refer to the Australian Tsunami Warning Centre but use a different acronym for it to BoM: AusTWC. Research by one of my students at ANU has shown that in an emergency the credibility of emergency information, is very important in the public accepting and acting on the information. If there are different names for the same organization, this will cause dangerous confusion.

The Australian Government has recently decided to intervene in state matters in several areas. Providing tsunami warnings direct to the public is one areas in which such intervention would be justified. The "last century" approach of GA passing a tsunami warning to EMA who pass it to SESs who then decide to pass it on (depending on the whim of the state premier) is not satisfactory. Instead each party should pool the information, interpretation and advice they have available. All those involved, including the general public, can then make an informed decision as to what to do.

During the 2004 Tsunami, the staff of the Pacific Tsunami Warning Centre attempted to warn the Indian Ocean region, but there was no system or procedure by which to do so. These staff could be forgiven for this unanticipated failing. However, this does not apply to the staff of the Joint Australian Tsunami Warning Centre. The scientists and public servants involved in the centre have an obligation to ensure that they can issue clear warnings to the Australian public. The excuse that it is someone else's job to pass on the warnings is not acceptable.

Problems with Indian Ocean Tsunami Warning System

In addition to administrative problems with the system in Australia, there are also problems with the electronic format of messages used for tsunami warnings in the region. The result of this could be that even when a timely message is issued, it may not be received or understood. AusTWC and similar organizations need to work together on better formats for the information.

Tsunami Warning System communication test shows Indian Ocean problem

The Pacific Tsunami Warning Center (part of the US NOAA, based in Hawaii) conducted a communications test on 11 July 2007. This showed a possible problem with the Indian Ocean system.

This is a test to verify communication links and determine transmission times involved in the dissemination of operational tsunami advice products from the pacific tsunami warning center to designated 24-hour tsunami warning focal points of the pacific tsunami warning system. ...

From: COMMUNICATIONS TEST, PTWC, 0153Z, 11 JUL 2007

The official addressees were asked to report back how and when the message was received. The centre will then calculate the transmission time. Messages are sent by fax and email, as well as several specialized networks: Global Telecommunications System (GTS), Aeronautical Fixed Telecommunication Network (AFTN) , Emergency Managers Weather Information Network (EMWIN), and Ranet.

The official recipients are designated 24-hour emergency centers in countries of the Pacific region. The message was also forwarded by the interim Indian Ocean warning system sponsored by UNESCO.

While not part of the official test, receipt on my system indicated:

From	Time
PTWC	1:53:39
unesco.org	2:19:44
Spam Filter	2:19:46
Local Host	2:19:51

This indicates there was a twenty six minute delay in the UNESCO system forwarding the message to the Indian Ocean network. Also there are six seconds delay within the local host machine, mostly due to the spam filtering. As I previously noted there is also a risk the message may not be received at all due to spam filters.

The PTWC have also instituted a colour coded four level system for grading tsunami warnings, with green, yellow, amber and red, indicating the increasing severity. A dial with pointer (similar to Australian bushfire Fire Danger Meters) is used to indicate the severity on the colored scale. This is useful for those who are unable to perceive color. However, there are no text equivalents for the colors and the text description are incomplete and ambiguous.

The description says: "Colors correspond to the message severity (red=most, green=least)". However, the other two levels are not mentioned. Also the text descriptions "most" and "least" do not match those used to label the corresponding dial images: "low", "moderate", "high", and "severe". In addition the levels on the description page are in the reverse order (high to low) to those on the indicator dial (low to high), which is confusing. Also those less familiar with English will have difficulty distinguishing which is higher priority: "high" or "severe".

PTWC should list all four severity levels in order of severity on the description page and include standard text captions. The same text captions should be used adjacent to the dial on warning pages for those who cannot see the images (as the ACT Emergency Services Agency does for bush fires). PTWC should also consider including a numeric ranking from 1 to 4.

Ranking	Severity	Color	Pointer
1	Low	Green
2	Moderate	Yellow
3	High	Amber
4	Severe	Red

Tsumani waring system sending out Spam

A number of national and international Tsunami warning centers around the world issue advice on potential threats via electronic systems. These are being issued via the Internet, web and other modern systems, using the same format previously used for teletype. This may stop the messages being easily interpreted or even received at all.

Before the Internet Tsunami warning centers issued warnings via teletype. Some centers are experimenting with web based formats, with XML machine readable message formats, images and HTML formatting to enhance the message. But the basic warning messages are still being issued in the same text format used for teletypes.

Teletypes have a limited character set and upper case was used for messages. This is still used, even when the message is sent by e-mail or other Internet based system. The result is a message which is harder for the human reader to interpret. In addition a spam filter will interpret the all upper case message as being potential unsolicited mail. These is a risk that the messages will therefore be blocked by a Spam filter.

This problem was reported to the UNESCO interim Indian Ocean System and the Japan Meteorological Agency (JMA), in 2005. An item on it appeared in Computerworld magazine, the same month.

The same problem occurred with a recent Tsunami Bulletin from the Pacific Tsunami Warning Centre (number 001 issued 0307Z 28 JUN 2007). Spamassassin (a popular open source spam filter) rated the message at 2.7. Wile this is a long way from a rating of definitely being Spam (around 10 on a 0 to 10 scale), it ideally should be zero. Spamassassin's tests applied were: BAYES_40 -0.148, DNS_FROM_RFC_WHOIS 0.6, NO_REAL_NAME 0.961, UPPERCASE_75_100 1.371.

The message being in upper case contributed more than half of the spam rating. While organizations are reluctant to change long established formats for emergency messages, perhaps it is time the tsunami format was changed.

Tsunami warnings need to be practiced

Last week I gave a lecture at the ANU on emergency management using the web. In this I explained how tsunami warnings could be received on a mobile phone. But I cautioned the volume of warnings intended for emergency experts could cause confusion. This seems to have happened on Monday, with the earthquake in the Solomon Islands and resulting tsunami. The Queensland Government in particular seems to have caused panic due to poorly planned emergency warnings.

"Homes, hospitals, schools and beaches were evacuated as panic gripped parts of Australia's east coast after warnings a destructive tsunami could be on its way.

The quake sent a tsunami crashing into some nearby low-lying islands and sparked a Pacific-wide tsunami alert.

At the height of the crisis, the Australian Bureau of Meteorology issued a tsunami alert for Australia's entire east coast, including Tasmania, warning dangerous waves and currents may strike."

From: "Tsunami warning sparks panic in Qld, NSW", The Age, April 2, 2007 - 10:44AM.

The tsunami alert was soon canceled, but there were unnecessary evacuations along the Queensland coast. Fortunately there appear to have been no deaths or serious injuries reported as a result.

There are ways to avoid this panic, but the techniques need to be practiced and tested. One approach I suggested were better visual and multimedia representations of the data. Emergency workers are used to looking at a simple text reports and tables of numbers. But these are hard to interpret for non-experts. Instead maps can be used to show the likely direction, timing and effect of a tsunami. Animation can also be used.

The Pacific Tsunami Warning Center issued eight bulletins for the earthquake at 01 Apr 2007 20:40 UTC. The first warning was at 01 Apr 2007 20:55 UTC for the Solomon Islands and Papua New Guinea. The first mention of Australia was in message 2, at 01 Apr 2007 21:32 UTC, giving an Estimated Time of Arrival in Cairns as 23:49 UTC. Message 3 gave the first observed measurement of the size of the tsunami as 15CM and Honiara. Message 8 canceled the warning at 02 Apr 2007 04:05 UTC.

In addition to the text bulletins, the PTWC provided a computer generated map of the location of the earthquake. The the West Coast/Alaska Tsunami Warning Center provided a tsunami travel time map.This shows when the tsunami is expected to arrive across the Pacific. It is part of an experiment to improve the text based information.Also they have an XML based interface, which provided the warning in machine readable format.

Tsunami Warnings in Australia

According to the Australian Bureau of Meteorology, they issue tsunami warnings, on advice from the PTWC:

In Australia, the Bureau of Meteorology has responsibility for issuing tsunami warnings. For areas covered by the PTWC the warnings are issued by the Bureau under guidance from the PTWC. At the present time the Australian Tsunami Alert System (ATAS) has been established by the Bureau of Meteorology, Geoscience Australia (GA) and Emergency Management Australia (EMA).

From Tsunami Information, Bureau of Meteorology, 2007

The Australian Tsunami Alert System (ATAS) issues bulletins in a similar format to those of the international warning centers.

Tsunami Alerts Issued in Queensland

The Queensland Department of Emergency Services issued three media releases for Monday's tsunami:

1. Tsunami alert 2/4/2007 ... Emergency Management Queensland (EMQ) is urging North Queensland residents north of Mackay to take safety precautions....
2. Updated Tsunami Alert 2/4/2007 ... BoM have advised that Willis Island Meteorological station reported NO noticeable affect of waves at 9am by which time the Tsunami should have passed. They are checking to see if there has been any affect to beaches. At this stage the threat seems to have eased, however the BoM will maintain the warnings until after any wave may pass near Cooktown. ...
3. Eased Tsunami Threat 2/4/2007 The tsunami warning issued this morning has now eased. This was a serious warning and Emergency Services have congratulated Queenslanders on the way they reacted to the warnings. ...

Unfortunately these warnings appear to have been an ad hoc response, rather than based on a plan. The warnings do not use a standard format, or terminology. There are no times on the messages making it easy for confusion to be caused. The use of the term "Erased" to indicate that the danger had passed is non-standard and inappropriate. The Queensland Government needs to adopt a standard format and wording, similar to that used by other states and national agencies for Tsunami warnings and needs to rehearse its use.

Tsunami Advice Issued in NSW

The New South Wales State Emergency Service issued two bulletins:

1. SES Monitoring NSW Tsunami Warning, 2 April 2007, 9:17am: A Tsunami Warning has been issued by the Bureau of Meteorology for people in coastal areas of New South Wales, Lord Howe Island and Norfolk Island. The NSW SES are monitoring the situation along with the Bureau and urge people to keep listening to their local news programs for further advice, updates and information. ...
2. SES Advise of NSW Tsunami Cancellation, 2 April 2007, 2:17pm: The Bureau of Meteorology has cancelled the tsunami warning that was issued earlier today for people in coastal areas of New South Wales, Lord Howe Island and Norfolk Island. ...

The NSW warnings were better than the Queensland ones, being time stamped and using similar terminology to the BOM bulletins. They also give more detailed and useful advice than the Queensland bulletins. However, NSW emergency information is difficult to find on the web, being at least three menu levels down from the NSW Government Home Page. This lack of clear emergency information was noted in 2004. In a major emergency this deficiency in web design may result in considerable loss of life.

Use of Smartphones to Enhance Response to Emergencies

The use of RIM Blackberry and other hand held devices combining phone and PDA functions ("Smartphones) is now becoming widespread in government agencies and companies. These devices could be used to enhance the response to emergencies, such as the recent tsunami, by providing better information to decision makers quickly.

Web pages designed for these devices would also provide a more efficient way to communicate information to emergency workers, the media and to the general public. Smartphones can display simplified maps, such as those for bushfires.