NOVA Home Find out what's coming up on air Listing of previous NOVA Web sites NOVA's history Subscribe to the NOVA bulletin Lesson plans and more for teachers NOVA RSS feeds Tell us what you think Program transcripts Buy NOVA videos or DVDs Watch NOVA programs online Answers to frequently asked questions

Ask the Expert Responses
Set 1: Posted March 29, 2005

Wave That Shook the World homepage

Next set of responses

NOTE: I wrote the answers to the questions in this set on Friday, March 25, 2005. On Monday, March 28, an earthquake with the preliminary magnitude of 8.5 struck the northern part of Sumatra. This earthquake appears to have been located on the next section of the subduction zone south of the December 26 earthquake in one of the areas Dr. Sieh [see first Q&A below] was concerned about. Monday's earthquake doesn't mean the risk is now gone. There is still concern about the next segments further south along the central and southern Sumatran coast.—Lori Dengler

Q: How likely is it, as some newspapers reported recently, that a different portion of the subduction zone off Sumatra, one further south and parallel to the southern coast of the island, might rupture in the near future? How big could such an earthquake be, and would it likely trigger another tsunami that would again devastate Sumatra?


A: Last week I was at a briefing of the Earthquake Engineering Research Institute in Pasadena. One of the speakers was Professor Kerry Sieh of Caltech. Dr. Sieh has been working in Sumatra for the past decade studying the geologic record of past great earthquakes on the Sunda trench. He and his coworkers have identified at least five past earthquakes on the segment of the trench along the south-central section of Sumatra with an average recurrence of about 230 years. The most recent event in 1833 is estimated to have a magnitude between 8.8 and 9.2.

Recently a team of scientists from Ulster University published an assessment of stress changes caused by the December earthquake in the magazine Nature, and they show increases both along the western Sumatran coast and on faults closer to Banda Aceh.

So both the paleoseismology and the stress analysis raise real concerns about another major earthquake in Sumatra that would not only produce strong ground shaking but also produce a sizeable tsunami. The central and southern portions of the Sumatra coast are more heavily populated than Aceh Province so it is imperative to assess the likely inundation zone and assist the Indonesian government in both education and warning efforts.

Q: Have you (or anyone else) been able to determine the amount of time between the water beginning to retreat and the arrival of the first destructive surge or wave? That is, if people know to react upon a sudden major retreat of the sea, how much time is there (generally) for warnings to be given, and more importantly, for those on shore to scurry to higher ground? Are we talking mere seconds or five to 10 minutes or more?

John Thomas
Sunnyvale, California

A: The time is variable, but it is on the order of minutes, not seconds. The time will depend on the wavelength of the tsunami, the location, and local coastal conditions. I talked to many eyewitnesses of the 2001 tsunami in Peru who estimated about 15 minutes between the water beginning to recede and the first positive waves. In Papua New Guinea there were only a few minutes. In the Indonesian tsunami, the time was variable but in areas far from the source region like Thailand and Sri Lanka the interval was generally at least 10 minutes—enough time for the 10-year-old British girl Tilly Smith to warn her family and many other tourists to get out of the water and to high ground safely.

There is no question that had people who observed the drawdown on December 26 in the Indian Ocean immediately evacuated, the death toll would have been much lower. However, a large drawdown is not always easy to recognize. On the Maldive Islands, eyewitnesses to the December tsunami suggest that it was difficult to observe and not nearly as obvious as in other parts of the Indian Ocean.

Q: I live and teach science in a tsunami danger zone. We felt mild shaking from a 7.0 earthquake 150 miles inland. What if that quake was out at sea? Could it have generated a tsunami? If a large subduction zone earthquake occurred 300-500 miles away, would we feel enough shaking to know that we have only an hour(?) to get to high ground? I know what to do in the event of extreme shaking, and very mild shaking, but what about in-between shaking? How will I know that there isn't a big event that is sending a wave in my direction, very fast?

John Harwood
Westport, Washington

A: Washington State has an active program of tsunami education and inundation zone mapping, and I am glad to hear that you are aware of living in a tsunami hazard zone. A magnitude 7 earthquake won't cause a significant tsunami on its own—the fault size and slip are too small. However, a 7 in rare cases may trigger a submarine slump that in some cases will cause a damaging tsunami. In 1998, a magnitude 7.1 earthquake in Papua New Guinea caused a tsunami that swept over four villages and killed at least 2,200 people. Most scientists agree that the main culprit was not the earthquake itself, but a large slide that it triggered.

A magnitude 7 earthquake might also cause a tsunami if it steps over from one fault trend to another. Such stepovers can concentrate significant uplift or downdropping in a small area again in rare cases causing a very localized but potentially large tsunami.

How can you tell if the ground shaking is strong enough to put you in harm's way? First, find out how hazardous is the area you live and work in. Tsunami hazard maps are now available for most of the populated areas of Washington State. What is the ground elevation and how far are you from the open coast? If you are on the waterfront, you need to set your alarm level much higher than if you are a mile inland. Get a NOAA weather radio. Tsunami warning messages are broadcast on NOAA Weather Radio as soon as they are issued. Just as important, messages of non-tsunamigenic earthquakes that are likely to be felt will also be broadcast. The tsunami warning centers are working on getting such messages out within minutes of the event. But don't assume you will get any official notification. Get into the habit of counting the duration of shaking. If you count only a few seconds, the earthquake is small and unlikely to be of concern. If you experience 20 seconds or longer of strong shaking, I would take nature's warning seriously.

As to the second part of your question, you might feel a magnitude 8.5 to 9 earthquake several hundered miles away if you were indoors. Quite a few people felt the Indonesian earthquake from Thailand, especially people sitting or lying down. It won't feel very strong but it will last a long time. I talked to a man who was in a beach bungalow in Thailand on December 26. He was in bed and definitely noticed the earthquake and recognized that it was a very big event. He had been in San Francisco during the 1989 Loma Prieta earthquake and recognized from the duration of shaking that this earthquake was probably bigger. Unfortunately he didn't associate shaking with a possible tsunami and he just rolled over and went back to sleep. He was very fortunate that when the tsunami arrived two hours later he was in the restaurant on the second floor of one of the few buildings that survived.

But to answer your question, there is a good chance you won't feel this earthquake from several hundred miles away. So how will you be warned? The most likely scenario is that the earthquake is on the southern protion of the Cascadia subduction zone (where I live). The West Coast Alaska Tsunami Warning Center (responsible for sending messages to Washington as well as Alaska, Oregon, California, and Canada) should be able to locate the event and send a bulletin out in a few minutes. If you have a NOAA Weather Radio with the alert option, it will automatically signal you when that bulletin is issued. Since it will take close to an hour for the tsunami to travel from me to you, you should have plenty of time to respond appropriately. Washington also has other ways of notifying you including the EAS radio system and the All Hazard Alert Broadcasting (AHAB) sirens in some areas.

Q: My family and I live on an island in the northwest part of Florida called the Panhandle. I hear about earthquakes in Central America all the time. Can a tsunami happen in the northern Gulf of Mexico?

William Bates
St. George Island, Florida

A: Yes. The Caribbean includes a subduction zone capable of producing large tsunamigenic earthquakes. It also has a number of active volcanoes that on extremely rare occasions may explode and also cause tsunamis. There aren't as many faults in the Caribbean as in the Pacific so the tsunami risk is much lower. But tsunamis have occurred in the past and will do so in the future. Nine credible tsunamis have been observed in the Caribbean in historic times, although none of these were detected on Florida's Panhandle. I am sure the Panhandle will experience many hurricanes before a significant tsunami occurs.

Q: How much of the energy of a tsunami is dissipated as it travels across the open ocean? I've read "not much." If that is true, why is that? Thanks.


A: The main loss of energy in a tsunami is due to geometric spreading of the wave front. If the tsunami source were a point, the wave would spread out in ever greater circles and the energy would be spread over a larger and larger area (because the spreading is essentially two-dimensional, the decay is proportional to 1 over the square root of the distance). A landslide or volcano is essentially a point source and the energy of the tsunami decays fairly quickly away from the source. A very large tsunami source like the Indonesian earthquake is not a point but a line and the decay rate is less.

Of course, the spreading is also limited by the size of the ocean. For very large tsunamis, the wave paths actually reconverge on the far side! For example the 1960 tsunami generated by the Chilean earthquake was particularly devastating in Japan in part because the wave front that had spread across the central Pacific was concentrated in the far west. Frictional attenuation of tsunamis in the open ocean is very small because the wavelength is so long. Once a tsunami is flowing on land, turbulence and friction do play a significant role and the tsunami energy quickly dissipates.

Q: Lori, you asked for any zones and other preparations in Oregon and California plan on just a Cascadia event or for all likely sources?

George Curtis
Tsunami expert
Hilo, Hawaii

A: George, you would ask me a hard one! In the Pacific Northwest, Cascadia is taken as the worst case event. However, there are many possible variations on Cascadia. George Priest at Oregon's Department of Geology and Mineral Industry has run a number of models with asperities—patches of greater fault slip—in various locations. These asperities can make a very big difference in runup. But Oregon has also looked at the likely impacts of distant tsunami sources and has prepared maps with several inundation scenarios. Washington State has looked at the potential for tsunamis generated within Puget Sound. In California, I am most familiar with planning in Humboldt and Del Norte Counties. The distant source evacuation plan is based largely on the 1964 inundation from the Alaska earthquake. In Southern California, the design tsunami is a submarine landslide.

Q: I am sure I do not understand enough about tsunamis and hence this question: why can't we build systems that prevent tsunamis, not just a warning system?

Conceptually, a tunsami is nothing but a wave. Any wave has two basic characteristics —it has a certain frequency and a certain wavelength. We already know that when a wave meets another wave that is out-of-phase with itself both waves can potentially cancel each other out. The same principle is used in many devices, for example in noise-canceling headphones, etc.

So, in theory, after detecting a tsunami this system can generate another wave which is out-of-phase with the tsunami and send it towards the source, which might, perhaps, completely cancel the tsunami wave. Since we already have tsunami detection systems in place, the hardest thing to do would be to generate an "anti-tsunami" wave.

Does this make sense or is it science fiction?


A: I'm not all that familiar with noise-canceling headphones, but I think they work by inputting essentially an equal signal that is out of phase with the annoying waves. So by this reasoning, you would need to produce a tsunami with the same energy and frequency as the real tsunami and somehow direct it in exactly the precise orientation to be out of phase everywhere with the real tsunami in order to cancel it. Probably impossible to do, don't you think? I'm afraid I have a hard time even making a credible science fiction story out of this, but Hollywood might be interested!

Q: I read that there was a possibility that another earthquake could happen soon in the same region as the recent tsunami and could create an equally powerful wave. Is there any new tsunami detection system in place since December that would help people in the region from suffering again?

Jared Falen
Los Angeles, California

A: No reputable sources have suggested a repeat of the December 26 earthquake in the same location soon. It will probably be hundreds of years before that particular section of the plate boundary accumulates enough strain to snap again. However, there is concern that a large and possibly tsunamigenic earthquake could happen either further north from the Andamans extending into Myanmar or to the south along the central Sumatra coast. No tsunami warning system has been established in the Indian Ocean since the December earthquake. However, planning for such a system is underway, and there is a move for existing tsunami warning centers to provide some information to Indian Ocean countries in the interim.

One point I wish to emphasize is that a tsunami warning system is only helpful for communities at some distance from the source region. If a warning system had been in place on December 26 (and coastal communities had a way to receive the information, and coastal communities understood what areas where at risk, and evacuation routes and safe areas had been designated, and people understood the tsunami hazard and the importance of following evacuation orders), it could have significantly reduced losses in regions an hour or more away from the source like Thailand, India, and Sri Lanka. It wouldn't have helped the people in Banda Aceh, because the first waves were arriving just about the time the warning message was being issued.

For people in the near source region, education is the primary tool for saving lives. People need to understand that the ground shaking beneath their feet is the warning. They need to immediately respond. I think the December 26 tsunami raised awareness of the tsunami hazard and if another major Indonesian earthquake were to occur soon, many people would flee to higher ground. However, Indonesia could really use assistance in developing hazard maps, emergency response plans, and education programs for an effective response and to institutionalize such programs so that future generations will know what to do even if many years pass without another event.

Q: Does the strength of the earthquake affect the speed or intensity of the tsunami? How strong does a earthquake have to be to produce a deadly tsunami? How does the ocean's topography affect the direction the tsunami goes?


A: The strength of the earthquake does not affect the speed of a tsunami, but will have a big effect on its size. A two-inch tsunami travels just as fast as a two-foot one—speed is only a function of the depth of the water. To produce a tsunami capable of affecting areas far from the source, an earthquake needs to be at least a magnitude 7.5. There are two situations where a smaller earthquake may cause a potentially damaging tsunami: first, when the shaking triggers a submarine slump, and second, when the earthquake rupture involves two or more faults that step over, creating a zone of uplift or subsidence in between. In either case the tsunami is likely to be very localized—but significantly damaging if the area is heavily populated or developed. Seafloor topography (bathymetry) can focus or defocus a tsunami—roughness, presence of sea stacks, and slope are all parameters tsunami modelers consider in their numerical models.

Q: Given that most ocean tsunamis are initiated by underwater earthquakes, is there a dynamic upper limit to the size or intensity of the waves propagated by these events? Are there any geological markers that may give us examples of what the upper limits could be and where it's most likely to occur? What effect if any do ocean tides have on the overall destructiveness of tsunami waves?

Marv Zellmer

A: In looking at historic large earthquake-generated tsunamis along the open coast, peak tsunami wave heights may reach 25-50 feet, with some areas on the order of 100 feet. There are a few cases where water heights are greater, but these usually involve a bay or inlet where the energy is focused. The 1960 Chilean earthquake had a magnitude of 9.5 and is considered just about as big as a tectonic earthquake can get on our planet; it produced tsunami heights in this range.

But earthquakes aren't the only tsunami source—massive volcanic island collapse events and asteroid impacts are capable of producing higher waves than the earthquake source, although just how high is still a matter of research and debate. The best studied evidence of past tsunamis are sand deposits, and many paleotsunami studies of sand deposits have and are being conducted. Several scientists have suggested certain erosional features may also be used to estimate the strength and size of past tsunamis, but there is no consensus at this time on how to interpret such features.

Q: I'm interested in the "bathtub effect." I live Seattle, which is located on Puget Sound. While most homes along the shores of Puget Sound tend to be located away from immediate shoreline, there are a great many homes located very close to the shorelines of large lakes, Lake Washington and Lake Sammamish. Is the possibility of a large wave created by either lateral or vertical displacement of concern to the creation of waves on these bodies of water? If the answer is yes, would these waves behave in a typical manner, or would they possess characteristics unique to a closed body of water like a lake? "Sloshing" around comes to mind.

I know that far distant earthquakes have caused damage to property located on the lakes around Seattle due to P waves, and I am wondering what the causational factors would need to be in order to create a devastating wave on inland bodies of water.

Matthew Rinearson
Seattle, Washington

A: There has been considerable interest in the potential tsunami hazard of Puget Sound. A workshop was held in 2001 addressing the potential for landslide tsunamis ( and numerical models of the likely tsunami to be caused by a rupture of the Seattle fault can also be viewed online ( The tsunami hazard in Puget Sound is not a seiching problem (the rhythmic oscillations of water in a bay or harbor), because the complex geometry and bathymetry of the sound makes it difficult to produce a coherent resonant signal.

The situation in the lakes is different. The 2002 Denali magnitude 7.9 earthquake in Alaska produced oscillations in a number of lakes, including both Lake Washington and Lake Union in the Seattle area. These oscillations were caused by the very large amplitude seismic surface waves produced by the earthquake and were not tsunamis. These oscillations, however, can be destructive. A number of boats in Lake Union were torn from moorings. In 1964 a similar phenomenon was observed in Louisiana from the 1964 Alaska earthquake. To produce a true tsunami in one of these lakes you would need either fault rupture beneath the lake that caused vertical deformation or a large landslide within or into the lake.

Q: Which company is the leader in early warning systems technology, and how would one get in touch?


A: Tsunami warning systems are not commercial products and can't be purchased retail. The National Oceanic and Atmospheric Administration is the government agency in the United States that is responsible for the tsunami warning system. Components include seismic networks, a coastal tide gauge network, and the tsunameter network of the DART system and the communication links. Most of the system is custom-made. There are a number of companies that produce siren systems that are interested in working with coastal communities to establish local warning systems. A good contact point is the Partnership for Public Warning at

Q: Within the past decade science journals carried an article about a rift that developed in the Indian Ocean crust. Might this in any way have a relationship with what happened to cause the December 2004 tsunami?

Martha Rinderle
Erie, Pennsylvania

A: I don't recall the articles you are referring to and haven't been able to find any specific references, so my answer is going to be more general than perhaps you would like. There are three main oceanic rift systems in the Indian Ocean that join up at the Rodriguez triple junction east of Madagascar and the triple junction is actively migrating. The African, Indian, and Antarctic plates are all moving away from each other. One of those plates, the Indian, is spreading to the north- northeast. The December 26 earthquake was the result of the subduction of the northeastward-moving Indian (sometimes called ther Indo-Australian plate) beneath the Eurasian plate (or Burma subplate). The crustal dynamics that drive this subduction process have been around for a very long time and are unlikely to be affected by the development of a new rift system.

Q: I am interested as to why the distance between the crests of tsunami waves is so great.


A: The distance is so long because the entire water column from the seafloor to the surface of the ocean is involved with the wave motion. A tsunami is produced when the seafloor is displaced. Even the smallest tsunami sources like a submarine landslide displace a very large volume of water. It's didficult for those of us familiar with wind waves and swells to imagine these very long tsunami wavelengths. Typical ocean waves have a period between successive waves of about eight seconds. The very longest ocean waves may reach 14 seconds. In contrast, the very shortest tsunami waves have periods of about five minutes and are much more often in the 30-minute to one-hour range. As a result, a relatively modest tsunami will contain a much larger volume of water than a higher wind wave.

Next set of responses

Back to top

Wave That Shook the World
Wave of the Future

Wave of the Future
What will it take to be ready for the next major tsunami?

Ask the Expert

Ask the Expert
Tsunami expert Lori Dengler answers viewers' e-mailed questions.

Anatomy of a Tsunami

Anatomy of a Tsunami
Follow the life cycle of the 2004 tsunami on this interactive map.

Once and Future Tsunamis

Once and Future Tsunamis
Learn about ten deadly tsunamis—and where the next could strike.

Send Feedback Image Credits
NOVA Home Find out what's coming up on air Listing of previous NOVA Web sites NOVA's history Subscribe to the NOVA bulletin Lesson plans and more for teachers NOVA RSS feeds Tell us what you think Program transcripts Buy NOVA videos or DVDs Watch NOVA programs online Answers to frequently asked questions