Warning System Components
Home
System Components
Warning Message
Integration
Resources

In every populated region at risk from tsunamis, there must be a system that can warn the people of the approaching wave. This system should have the capacity to relay messages as clearly and quickly as possible and should be cost effective.

Utilizing Existing Technology

A simple place to start is by sending warnings through the communication infrastructure that is already in place: commercial radio, marine radio (for boats and ships), cable and local television, internet, conventional telephones, and cell phone text messaging. Except for conventional telephones, which require expensive automatic dialers (about $24,000 for installation costs, operating costs of $0.02 per telephone line per year [$45,000 per year for 100,000 lines], and $0.20 per completed call when the system needs to be activated1) in order to reach a large number of people quickly and even they can only reach at most 500 people per minute1, these are all fast and easy ways to convey relatively detailed messages. It is possible, therefore, to use these means to give out more information than just "evacuate now!" including tsunami advisories in addition to warnings.

In terms of the ease of implementation, all that is required for warnings to be sent out over cable or local television and commercial radio is an agreement with local and cable stations to allow an override in the case that information needs to be broadcasted, internet simply requires posting a web site, and cell phone text messaging only needs a list of numbers obtained from the cell phone providers. Using marine radio, like regular radio, would only require permission to override the frequencies. In Peru, this would involve contacting la Empresa National de Puertos which is in charge of getting tsunami messages to ships2. Posting information on the internet allows people who have access to computers to confirm advisories they may have heard from radio, television, text messages, or other people as well as allowing them to look at visual images such as maps of the zones at risk.

To keep the warning system cost effective, conventional telephones can only realistically be used to warn a few people. The logical choices are emergency response and evacuation officials and the heads of establishments such as schools, hospitals, nursing homes, hotels, and office buildings. This would still require automatic dialers, but at a much smaller scale; these few calls are definitely worth the cost since the receipt of this type of message does not depend on a radio or television happening to be on at any given moment and will reach a far greater number of people per call than calls to individual houses could. Also, some automatic dialers can even change the sound of the phone's ring so the person receiving the call knows immediately that it is an emergency3.

Another possibility is to send out warnings via email. This would certainly be as cheap and simple as sending out text messages, but is unlikely to be nearly as effective. As long as they are on, cell phones alert their owners of the fact that a text message has come in, whereas email needs to be checked. Among those who have computers and an internet connection, very few check their email frequently enough to be able to get this kind of a warning in a timely manner: according to the rules of probability, the chance that someone who checks their email three times a day will see an emailed warning within 30 minutes after the time it was issued is only about 6.06%. Someone who only checks their email once a day has only a 2.06% chance of seeing the email within half an hour of the time it was sent. Clearly, email is not going to be particularly useful for our purposes - time and effort should be focused on other things.

Methods of Warning Requiring Installation

So far, our plan has only dealt with existing communication infrastructure. One problem with this approach is that not everyone has access to the technologies necessary for these types of warnings. Micronesia, for example, has only about 10,100 phone lines, 1800 cell phones, 6 radio stations, 3 television stations (in addition to cable), and 6000 internet users for of a population of 108,105 people4. An obvious solution to this problem is to give every household a radio if it does not already have one. This, however, would require searching out and contacting a large number of families and/or individuals who are unlikely to receive any notice sent out through the media, not to mention the costs associated with the distribution and maintenance of an equally large number of radios. Also, even after the radios are distributed, they will only be effective when they are working properly and on at the time a warning or advisory is sent out. Advisories and warnings that come at night are especially likely to be missed. Distribution of tone-alert radios instead of the conventional type would solve the problem of warnings only being conveyed when the radios are turned on, but would be even more expensive (they cost between about $20 and $500 [high end includes features that allow message programming] each5,6 ).

Returning to the original plan, another problem is that people who are outside (on beaches, streets, fields, in small boats close to shore, or in any other remote location), asleep, or simply not listening to the radio, watching television, or surfing the web, even where all of these technologies are ubiquitous, will not hear the warnings. Our solution is to have a loud outdoor warning signal (sirens) that can be heard over long distances covering all areas at risk from tsunamis; they would be able to alert people both inside and out. Sirens have the inherent collateral advantage of attracting attention and increasing the sense of urgency for people who may already have heard about the tsunami through the technologies previously mentioned. Also, sirens, especially when they can be remotely activated through the receipt of a shortwave radio signal, can send messages out in a matter of minutes. Shortwave signals are advantageous because they can carry over much greater distances than can long wave signals and therefore do not require a chain of transmitters to carry the signal along. Not only would transmitters slow the signal down, but if one failed, the chain would be broken and the message would, at best, be delayed or, at worst, be prevented from reaching sirens and tone-alert radios at the end of the chain.

An option in deciding what type of siren signals to use is to make sirens give several different levels of warning through the use of different colored flashing lights and/or different sounding alarms. This, however, greatly increases the probability that people will be confused and not know to evacuate when they need to. Even if people are taught the differences between the alarms, differences in sounds can easily be confused or forgotten. Also, even if people could always distinguish between different warnings, having the sirens go off in the case of tsunami advisories in addition to warnings will make people more accustomed to hearing them, fostering complacency. This effect is similar to that of having too many false alarms; it is like "crying wolf."7 In light of these considerations, we propose just one type of siren to be activated only in the case of a warning (when evacuation is necessary). In this case, as soon as people hear the siren, there will be no ambiguity as to its meaning.

The remaining hole in this siren system is that, in areas that lack or have very few telephones, radios, televisions, and computers, nobody will have specific information about the tsunami. How long should they stay evacuated? How far away from the water do they need to go? To ameliorate this problem, we propose distributing tone-alert radios that receive shortwave signals to one or more local officials (depending on how large the town is) who will receive specific information and help with the evacuation. (We will also give two-way tone-alert radios to coastal and navigational organizations, emergency response organizations, and national government officials so they get speedy alerts and in case some telephone lines are down.) This will also ensure that at least one person will know the tsunami is coming in case automatic sirens fail. For that situation, there needs to be some back-up way for this person or group of people to quickly get out the warning. Depending on what resources are available and on local customs, this could be done with loudspeakers, loudspeakers mounted on vehicles or planes, church bells, a manually activated siren, or, if necessary, by simply knocking on doors. When the siren is working, these tools can be used to help direct the evacuation and reinforce the siren's message, making sure people know they need to evacuate, where they need to go to be safe, and how long they need to stay there. Tone-alert radios are portable, so the officials will be able to receive updated information if they cannot stay inside.

This brings us to another important point - how do we prevent siren failure? One way to approach this problem is to have some sort of back-up system like the one previously discussed for places that lack a developed communication infrastructure. In larger towns and cities it is impractical to rely on loudspeakers or church bells and next to impossible (not to mention slow) to reach everyone by knocking on doors. More than one siren could be installed, but this gets expensive (sirens cost between $10,000 and $40000 each1) and does not guarantee that the second siren will not fail for the same reason the first did. A more feasible solution is to design a more robust siren. Siren failure is likely to be caused by power outages, damage to the siren itself during an earthquake, or problems with the mechanism that have gone unnoticed. Sirens with solutions to some of these problems already exist: the AHAB system (all hazard alert broadcasting system) has self-sufficient power8 and thus would not be disabled by a power outage and other systems of sirens have regularly scheduled tests to make sure that they are still working. The Siuslaw National Forest in the United States, for example, has a 3-minute solid tone-out tsunami siren test on the second Monday of every month.9 Sirens could also be designed so they could be activated manually if necessary and have cheap, sturdy parts that make what little maintenance is necessary easy and inexpensive. Ideally, this type of engineering could also make sirens less expensive in the long term. The development of a system with a much smaller chance of failure than the average siren is simply a matter of bringing these features together.

One more possible feature of a warning system is the use of flares to warn boats that do not have marine radios of an approaching Tsunami. There are a few problems with this idea. First, there is no way to ensure that people in boat will even see the flares, especially on a bright day. The flares could be made to make noise as they go up into the air, but if a boat is close enough to shore to hear that noise, it will probably already be in range of the sirens and the flare will be unnecessary. Also, boats that are far enough away from the shore to be out of ear-shot of the sirens (about ½ - 1 mile from shore1) will almost certainly be in deep enough water to be safe from a tsunami. In the unlikely event that the point from which the tsunami propagates is very close to the location of a boat, it may not be so safe, but flares still would not help since they take time to set off. Furthermore, provided the boat does see the flare in time, there is still the problem of making sure its meaning is clear: what is to distinguish a tsunami warning from an ordinary firework? In short, the work and money required to buy, distribute, and set off flares would not do much good and could even decrease the resources and manpower available for other, more urgent and effective warning and evacuation procedures.

Integration with Existing Warning Systems

Another consideration is what to do if a country already has a warning system. Peru, for example, has el Sistema National de Alerta de Tsunamis (National Tsunami Alert system) which includes a generalized plan to use means such as telephone, radio, and the Aeronautical Fixed Telecommunication Network get word of a tsunami from the Pacific Tsunami Warning Center in Hawaii to officials and boats/ships with radios2. This plan, however, does not include specifics on getting that warning to the people. In this case, it is not necessary to start from scratch in devising the route of the message or in setting up the radio alerts because these things have already been started, but it is necessary to ensure the existence of the other modes of warning we have decided to use (such as television and siren warnings) throughout the country. In short, if an existing system is able to give a quick, clear alert to people everywhere from modernized, urban areas to rural, and remote outdoor locations, give specific information at least to local officials, and has a guard against system failure, nothing new needs to be added; however, if the system consists of only pieces of what is outlined above as essential for an optimal warning system, the gaps need be filled in.

Our plan is general rather than specific to Peru and Micronesia due to the great range of differences even inside those two countries in levels of modernization, availability of resources, and culture. Our plan does not attempt to outline exactly what devices, messages, or communication networks should be used in all the cities, villages, towns, ships, boats, and beaches of the two countries because that would be impossible too do accurately without more specific information than is currently available to us. Instead, our plan explains what is necessary everywhere for an optimal warning system and where different types of systems will be most needed.

Sirens Tone Alert Radios Commercial Media Mass Communication Flares Small Airplanes
Cities      
Towns    
Villages    
Sparsely-populated Areas
Tourist Areas      
Beaches        
Boats      

denotes that a warning method would be useful in selected area
     denotes that a warning method would be useful, but the necessary infrastructure may not be available

Warning Devices Considered

The following list consists of every warning device we considered for inclusion in the warning system, and their most important attributes.
Sirens1
  • can be heard up to half a mile away
  • operate under self-sufficient power
  • can be activated remotely (by radio frequency) or locally
  • emit prerecorded alert message or siren
  • most useful for warning people who are outdoors

Tone Alert Radios1
  • activated by radio frequencies (do not need to be already on to broadcast alert)
  • must be within 40 miles of broadcasting source to be activated by standard radio frequencies
  • short wave tone alert radios can be activated from a much further distance
  • can emit prerecorded voice messages, sirens, live voice messages

Commercial Media (Radio and Television)10
  • infrastructure already in place
  • will work best during the day and evening
  • more information can be conveyed, such as pictures
  • not universally available
  • protocols and standard procedures must be in place beforehand

Mass Communication (Text Messages, Email and Automatic Dialers)1
  • infrastructure already in place
  • not universally accessible
  • may not work if towers are damaged, such as in an earthquake
  • automatic dialers cannot reach everyone who has a phone in the available time
  • email messages are not likely to be received in the available time

Flares
  • can be seen from offshore
  • must be clearly distinguishable from ordinary fireworks
  • would work better at night
  • cannot convey instructions

Small airplanes11
  • can carry banner or loudspeaker to warn convey warning
  • useful in isolated areas and regions difficult to access
  • prohibitively expensive


1 - Oregon Emergency Management, 2001
2 - Direccion de Hidrografia y Navicagion, 2005
3 - Mileti, 1990
4 - CIA, 2005
5 - National Weather Service, 2005
6 - Federal Signal Informer, 2005
7 - Tierney, 2000
8 - Washington State Emergency Management Division, 2004
9 - Sandlake Recreational Area, 2005
10 - Federal Communications Commission, 2003
11 - Platform for the Promotion of Early Warning, 2005



Questions? Comments? Contact Tsunami 6.