Tsunamis are an ever-present threat to lives and property along the coasts of most of the world's oceans. As the Sumatra tsunami of 26 December 2004 reminded the world, we must be more proactive in developing ways to reduce their impact on our global society. This article provides an overview of the state of knowledge of tsunamis, presents some challenges confronting advances in the field and identifies some promising frontiers leading to a global warning system. This overview is then used to develop guidelines for advancing the science of forecasting, hazard mitigation programmes and the development of public policy to realize a global system. Much of the information on mitigation and forecasting draws upon the development and accomplishments of a joint state/federal partnership that was forged to reduce tsunami hazards along US coastlines—the National Tsunami Hazard Mitigation Programme. By integrating hazard assessment, warning guidance and mitigation activities, the programme has created a roadmap and a set of tools to make communities more resilient to local and distant tsunamis. Among the tools are forecasting, educational programmes, early warning systems and design guidance for tsunami-resilient communities. Information on international cooperation is drawn from the Global Earth Observing System of Systems (GEOSS). GEOSS provides an international framework to assure international compatibility and interoperability for rapid exchange of data and information.
1. Knowledge of tsunamis
A tsunami is a set of ocean waves caused by any large, abrupt disturbance of the sea surface. If the disturbance is close to a coastline, local tsunamis can demolish nearby coastal communities within minutes. A very large disturbance can cause local devastation and export destruction thousands of miles to distant coastal communities. A major tsunami can bring destruction to the coastal residents of an entire ocean, making the hazard inherently international. The word tsunami is Japanese, represented by two characters: tsu, meaning ‘harbour’, and nami, meaning ‘wave’. It is thought that from ancient times, when the Japanese observed unusual wave activity in a harbour, they were aware that a tsunami could be approaching. Tsunamis rank high on the scale of natural disasters. Since 1850 alone, tsunamis have been responsible for the loss of over 420 000 lives and billions of dollars of damage to coastal structures and habitats throughout the world. Most of these casualties were caused by local tsunamis that occur about once per year somewhere in the world. Predicting when and where the next tsunami will strike is currently impossible. Once generated, however, forecasting tsunami arrival and impact is possible through modelling and measurement technologies.
Tsunamis are most commonly generated by earthquakes in marine and coastal regions. Major tsunamis are produced by large (greater than magnitude 7 on the Richter scale), shallow focus (less than 30 km beneath the surface) earthquakes associated with the movement of oceanic and continental tectonic plates. They frequently occur in the Pacific, where dense oceanic plates slide under the lighter continental plates. When these plates fracture they produce vertical movement of the seafloor that allows a transfer of substantial energy from the solid earth to the ocean. When a powerful earthquake struck the coastal region of Indonesia in 2004, the movement of the seafloor produced a tsunami with an amplitude in excess of 30 m (100 feet) along the adjacent coastline, killing more than 168 000 people (Red Cross 2005). The tsunami radiated outward from this source and within 2 hours had claimed over 63 000 lives in Thailand, Sri Lanka and India (figure 1).
Underwater landslides associated with smaller earthquakes are also capable of generating destructive tsunamis. The tsunami that devastated the northwestern coast of Papua New Guinea on 17 July 1998 was generated by an earthquake that registered Ms 7.0 on the Richter scale and apparently triggered a large underwater landslide. Three waves measuring more than 7 m high struck a 10 km stretch of coastline within 10 minutes of the earthquake/slump. Three coastal villages were swept completely clear by the deadly attack, leaving nothing but sand and 2200 people dead (Synolakis et al. 2002). Other large-scale disturbances of the sea surface that can generate tsunamis are explosive volcanoes and asteroid impacts. The eruption of the volcano Krakatoa in Indonesia on 27 August 1883 produced a 30 m tsunami that killed over 36 000 people (Simkin & Fiske 1983). In 1988, scientists discovered evidence of a 10 km diameter asteroid that landed in the Gulf of Mexico approximately 65 Myr ago and produced a huge tsunami that swept over portions of the southern US (Bourgeois et al. 1988).
(b) Wave propagation
Earth movements associated with large earthquakes are thousands of square kilometres in area, and the vertical movement of the seafloor immediately deforms the sea surface. The resulting tsunami propagates as a set of waves whose energy is concentrated at wavelengths corresponding to the dominant length-scale of Earth movements (approx. 100 km), at wave heights determined by vertical displacement (approx. 1 m), and at wave directions determined by the geometry of the source and adjacent coastline. Because each earthquake is unique, every tsunami has unique wavelengths, wave heights and directionality. However, a characteristic of such waves is that they all travel at a speed proportional to the square root of the water depth. Thus, in the open ocean, these waves propagate at the speed of a jet airplane but slow to the speed of an automobile as they flood land. Application of the mathematical rules governing this behaviour has allowed scientists to predict the arrival time of a tsunami by knowing the source location and the variation of water depth encountered by a wave as it propagates to a specific coastal location. The 26 December 2004 Sumatra tsunami propagated globally from the Indian Ocean into the Pacific and Atlantic Oceans and was recorded at most major harbours in the world (Titov et al. 2005a). Figure 2 shows the distribution of wave energy inferred from a propagation model for the Sumatra tsunami. Two features of this distribution are particularly dramatic and informative. Near the source in the Indian Ocean, the distribution pattern was governed primarily by the shape of the seismic disturbance that created two beams of energy perpendicular to the main fault axis, coupled with reflection from the adjacent Sumatra coastline. Far from the source, outside of the Indian Ocean, mid-ocean ridges acted as waveguides that focused and steered the waves as they propagated throughout the global world oceans.
(c) Tsunami inundation
As a tsunami approaches the shoreline, its energy is progressively concentrated into a smaller volume of water. To conserve energy, the wave height and current speed increase dramatically. The resulting amplified tsunami floods the coastline with flow velocities as high as 14 m s−1 (about 40 mph), destroying everything in its path (Titov & Synolakis 1997). The 26 December 2004 Sumatra tsunami was not only the largest, but also the best video-documented in recorded history. The videos, via televised broadcasts, vividly showed to the world the power of tsunami inundation (figure 3) and the effects of repeated tsunami attacks on the coastline. The first wave destroys most structures, except reinforced concrete, and the debris from these structures, along with automobiles and other unattached objects, become battering rams swept inland by subsequent waves. Each wave creates more debris for the following wave to use for further destruction. Major tsunamis are composed of 6–12 large waves that repeatedly attack the coastline at intervals of 30–90 minutes. For example, during the Sumatra tsunami, the Sumatra coastline was attacked for 12 hours by eight waves about 90 minutes apart (Titov et al. 2005a).
Our quantitative knowledge of tsunamis at the coast is limited to observations of the integrated damage inflicted on the coastline and to tide gauge records in harbours. These observations have provided only limited insights into tsunami dynamics. Numerical models of inundation have been developed that simulate the flooding potential of tsunamis and are presently used to plan evacuation procedures (figure 4) for threatened coastal communities (González et al. 2005a). In the past decade, however, we have been able to measure tsunamis in the deep ocean, which gives us an important new observational tool to forecast their impact. Still, the complicated source and inundation dynamics will remain elusive until we develop additional tools to observe, model and understand these complex processes.
(d) A short history of international tsunami warning systems
The history of hazard mitigation tracks well with the history of destructive tsunamis and the history of coastal hydrodynamics (Synolakis & Bernard 2006). Following the 1946 Alaska-generated tsunami that killed 173 people in Hawaii, the Pacific Tsunami Warning Center (PTWC) was established in Hawaii to warn against distant tsunamis affecting the United States. In response to the 1960 Chilean tsunami that killed 1000 people in Chile, 61 in Hawaii and 199 in Japan, the international community formed two international bodies, representing science and governments, to provide distant warnings to nations of the Pacific Ocean. The International Union of Geodesy and Geophysics formed the Tsunami Commission (science) and the United Nations Educational, Scientific and Cultural Organization/Intergovernmental Oceanographic Commission (governments) formed the Intergovernmental Coordinating Group for Tsunami Warnings in the Pacific (ITSU). The United States offered PTWC as the operations centre for the entire Pacific Basin and offered to support the creation and maintenance of the International Tsunami Information Center. Thus, the international effort to mitigate the impacts of tsunamis began over 40 years ago.
The 1964 Alaska earthquake and tsunami killed 120 in Alaska and the west coast of the United States. This led to the creation of the Alaska Tsunami Warning Center for local warnings in Alaska. Similar local and regional warning centres were established in Japan, USSR, Tahiti and Chile in the following decades. All of these Pacific warning centres coordinated their efforts with PTWC in Hawaii. The international effort to mitigate tsunamis continued to improve as the seismological community successfully installed a global network of instruments to detect earthquakes throughout the world. Advances in satellite communications made detecting earthquakes faster and more accurate. Unfortunately, the technology to directly measure tsunamis lagged.
The warning centres became rich in seismic information while remaining poor in tsunami generation determination. Assessing the latter involves more than just determination of the location and characteristics of the earthquake (see §1a), yet the warning centres had no choice but to use the earthquake fault solutions alone to assess tsunami potential along with very limited data from coastal tide gauges. This situation led to over-warnings, i.e. warnings that were unnecessary because the observed tsunami was non-destructive. From 1949 to 2000, 75% of the warnings turned out to be unnecessary. The problem of over-warning led to credibility issues among the populations being served and to economic studies of the cost of over-warning. Following the 1986 ‘false alarm’ in Hawaii when Waikiki Beach was evacuated, the State of Hawaii estimated an economic loss of US$ 40 million due to the interruption of business and the closure of State government offices (Bernard 2005a,b).
The decade of the 1990s elevated the awareness of tsunamis, as over 4800 people died from 11 destructive events. Serendipitously, the United Nations established International Decade of Natural Disaster Reduction (1990–2000). The tsunami science community responded by developing technology to produce inundation maps at the community level. During the decade, 73 maps were produced in nine countries (Bernard 1999). As expected, the impact of the mapping was overwhelmingly positive. The maps provided the necessary information for local communities to plan for tsunamis, including evacuation procedures and recovery plans.
One small tsunami occurred in 1992 in northern California that killed no one. However, this was the first recorded earthquake on the Cascadia subduction zone, and seismologists feared that a larger tsunamigenic earthquake was imminent. In response to this concern, the US created the National Tsunami Hazard Mitigation Programme to address two concerns: (i) a Cascadia earthquake/tsunami and (ii) over-warning. By 2001, this programme had developed the technology to detect tsunamis in the deep ocean, paving the way for forecasting which reduced over-warning (see §5 for details), produced inundation maps for 130 communities in the US and introduced the concept of tsunami-resilient communities (Bernard 2005a). A special issue of Natural Hazards was published to document the accomplishments of the programme (Bernard 2005b).
The 26 December 2004 Sumatra tsunami killed 232 000 and shocked the world with the impact that a destructive tsunami can have on the international community. People from 55 nations perished during the attack in the Indian Ocean. As a result, there is now demand for a global tsunami warning system. The efforts of the tsunami community since 1946 will serve as a solid foundation upon which to build the warning system of the future and will reduce the loss of life from future events (Bernard 2001).
2. Challenges facing tsunami scientists and emergency managers
The greatest challenge facing practitioners of tsunami research and mitigation is change in the ‘perception’ that tsunamis are rare events. Data clearly show that every year a damaging tsunami will cause death and destruction somewhere along our global oceans. That said, there are four specific challenges that can be addressed immediately: observations, standards, vulnerability and education.
As described in §1, our ability to understand tsunami dynamics is closely tied to our ability to accurately characterize a source and observe generation, propagation, inundation and interactions with structures. Among the observations needed, those of the source motion will be the most difficult. Source observations are complex because tsunami generation is probably a combination of many effects due to an earthquake, including intense solid earth shaking that couples to the ocean, deformation of the crust, underwater landslides and activation of splay faults. Any one of these processes would be difficult to measure, but to measure all processes simultaneously will be very challenging. Propagation observations are currently provided by real-time bottom pressure recorders—tsunameters. These instruments, coupled with a real-time reporting Deep-ocean Assessment and Reporting of Tsunamis (DART) system (figure 5), have demonstrated their value in the early cancellation of a distant and a local warning by measuring a non-destructive tsunami in the deep ocean and relaying these data to warning centres in the US (González et al. 2005b).
A sufficiently dense network of DARTs will be required to accurately forecast tsunamis. The challenge will be to deploy and maintain the network of these instruments over vast areas of the world's oceans. The payoff is a global society better prepared for the next tsunami. By placing the DART network in the Global Earth Observing System of Systems (GEOSS), international cooperation and interoperability will be assured and a global tsunami warning system can be realized.
Inundation observations are required to verify existing inundation models and must include measurements of inundation flow velocities, friction and withdrawal flow velocities. Tsunami/structure interactions must be observed to enable the design of structures resilient to tsunami forces, including measurements of strain, acceleration and other impact parameters. To develop the measurement systems to accurately document tsunami flooding dynamics should be a high priority.
As in any science, standards must be established for observation and modelling of tsunamis to provide a credible basis for advances in the field. Tide gauges have no standards for tsunami application. US tsunameters are standardized because they use a common pressure sensor designed for high-pressure use and are manufactured under careful quality control. As the global network of tsunameters develops, a sensor standard must be established. A step in this direction has already been taken by the International Coordinating Group for the Indian Ocean Tsunami Warning System (ICG/IOTWS), who have established a DART Operator's Group to set standards for deep-ocean measurements of tsunamis. These standards should suffice for current and future measurement methods.
Because models will be used in the future to evaluate the potential flooding risks and public safety of any community, modelling standards are also required. Coastal communities are under constant physical development, which alters the coastline position, bathymetry and topography. All of these affect the accuracy of inundation models. Using standard models provides some assurance that the latest version of the model represents responsible science and is derived from legacy models that represented the state of the science at the time they were developed. The tsunami modelling community has held a series of workshops to benchmark these models. The latest workshop was held in 2004 and compared inundation models (Liu et al. in preparation). One recommendation from this workshop is to establish a web-based community model as a standard, with researchers encouraged to improve this community model.
The present practice of establishing inundation maps is focused on the evacuation of people to save lives, i.e. human vulnerability. Economic vulnerability, i.e. the vulnerability of coastal infrastructure to destructive waves, is another important aspect of tsunami mitigation. Ports and harbours transport a large percentage of any nation's gross domestic product. Should a tsunami interrupt port operations, the economic impact could be devastating both locally and across the nation. A recent study found that if the ports of Long Beach and Los Angeles were closed for 1 year due to tsunami damage, the impact on the US economy would be US$ 43.5 billion (Borrero et al. 2005). At the local level, impaired port operations would impact the local labour force and local commerce. At the distant level, businesses that rely on ‘just in time’ components would be disrupted and possibly unable to produce products until an alternate supply is established. These studies require economic expertize to quantify the true vulnerability of a nation's risk to the tsunami hazard. Such studies provide the necessary economic justification for such evaluations and pave the way for a process to make the infrastructure more resilient to tsunamis. Saving both lives and communities is vital and essential. The challenge is to engage and support economists in tsunami studies.
(d) Educational programmes
Knowledge of tsunamis saves lives. Several examples from past events indicate that populations aware of tsunami hazards and physical indicators of their presence can significantly reduce fatalities. In 1993, Japan experienced a local tsunami that killed about 15% of the exposed population on Okushiri Island, while a similar event occurred in Papua New Guinea in 1998 with a resultant 40% loss of life at a small village. The difference was that the Japanese people were educated about tsunamis and knew how to respond to the strong shaking of an earthquake, while the people in New Guinea did not. In the most recent example of the 2004 Boxing Day tsunami, the indigenous inhabitants of the islands of the eastern Indian Ocean had ancestral knowledge to flee to high ground upon feeling the strong shaking from a large earthquake. The challenge is to effectively educate a wide range of communities that vary from subsistence villages to major metropolitan ports. The educational tools are available (e.g. signs as shown in figure 6), but a global plan needs to be developed and implemented in tsunami-prone communities to save lives in the future. International organizations could lead this effort by supplying the educational material, while each nation could accept responsibility for implementation at the community level in the appropriate visual forms and languages.
Advances in tsunami forecasting can profoundly influence the development of mitigation strategies. At the heart of any mitigation programme is an accurate assessment of the hazard. Some elements of tsunami hazard assessment include the frequency of occurrence, the extent, forces, and duration of flooding, the impact of flooding on structures, facilities and infrastructure, and an assessment of the use of the potentially affected coastal areas. In the United States, one study (Eisner 2005) codified the way to design a tsunami-resilient community through seven principles:
know your community's tsunami risk, hazard, vulnerability and exposure;
avoid new development in inundation areas to minimize future losses;
locate and configure new developments that are built in inundation areas to minimize future losses;
design and construct new buildings to minimize damage;
protect existing developments through redevelopment, retrofit, and land reuse plans;
take special precautions in locating and designing infrastructure and critical facilities to minimize damage; and
plan for evacuation.
Note that to put these principles to use, accurate technical information must be available. The implication is that actions must be guided by a connection between quantitative science and hazard reduction practices.
4. Frontiers for the future
The scale of the 2004 Sumatra tsunami has reinforced the need to examine frontiers from the need to provide continuity based on past best practices and to examine new opportunities revealed by the event itself. Certainly, the frontier of tsunami forecasting and human response simulation have roots from the past. Based on experience with the Sumatra tsunami, real-time remote sensing has emerged as an excellent frontier.
(a) Forecasting tsunami impacts
Among the most promising frontiers for societal benefit is the accurate forecasting of tsunami impacts. Not only would such forecasts save lives during a tsunami, but accurate forecasting information in a scenario mode can be used to design, build and maintain tsunami-resilient communities. Although a major advance has occurred in tsunami forecasting, as described in the following section, this is only the beginning of this new frontier. To achieve accurate forecasting will require the application of three-dimensional models that can simulate tsunami forces and their interactions with structures and movable objects. To verify the models and wave tank experiments, precise observations of these forces will also be required. Such information is necessary to design structures to withstand tsunami forces and help guide building practices in threatened areas.
Owing to the international nature of tsunamis, these forecast products should be developed and sustained as part of the global tsunami warning system that could be embedded in the GEOSS framework to assure compatibility and interoperability (Lautenbacher 2005).
(b) Human response simulations
Little research has been done on the human response to tsunamis. These studies would represent a new frontier in modelling the response of humans to this threat. In the United States, educational programmes have been effective in raising awareness of tsunamis, but have not been effective in modifying human behaviour. For example, recent surveys of Washington state residents who live along the coast revealed that residents were aware of tsunamis, but were not taking any concrete action to prepare for them (Johnston et al. 2005). Hence, there is some social element missing in the existing educational approach to motivate change. This Washington survey is a clue that other aspects of our educational approach are deficient. One approach would be to analyse social behaviour of recent disasters to determine how human response can be simulated. Simulations of response to warnings, evacuations, preparation and recovery are a few of the experiments that could guide the development of more effective educational and mitigation programmes.
(c) Real-time remote sensing
The incredible satellite images taken before and during the Sumatra tsunami illustrated the value of remotely sensed information (figure 7). If this information were available in near real time, such data would be valuable for rescue and recovery operations. These data would also be useful for before and after damage imagery to quickly evaluate the extent of damage caused by the tsunami as well as other natural hazards. The recent hurricanes in the US and the 2004 earthquake in Pakistan and, in 2006, the Philippines landslide and the Tonga and Yogyakarta earthquakes vividly demonstrate the need for real-time remote-sensed imagery.
5. Science of forecasting
One of the greatest contributions of science to society is to serve it purposefully, as when providing forecasts to allow communities to respond before a disaster strikes. The Extreme Natural Hazards Symposium and the Royal Society in 2004 focused on the role of science in reducing societal impact from a set of natural hazards. This section lays out the ground-breaking scientific advances in forecasting tsunamis that give us the confidence that science makes a difference.
(a) Forecasting technology
Since 1946, the Pacific tsunami warning system has provided warnings of potential danger in the Pacific basin by monitoring earthquake activity and the passage of waves at coastal tide gauges. However, neither seismometers nor tide gauges provide data that allow accurate prediction of the impact of a tsunami at a particular coastal location. Monitoring earthquakes gives a good estimate of the potential for tsunami generation, based on earthquake size and location, but gives no direct information about the tsunami itself. The variation in local bathymetry and harbour shapes severely limits the effectiveness of harbour tide gauges in providing useful measurements. Partly because of these data limitations, 15 of 20 tsunami warnings issued since 1946 were considered false alarms because the waves that arrived were too weak to cause damage. Recently developed real-time, deep-ocean tsunami detectors provide the data necessary for models to make forecasts (González et al. 2005b). Recent distant and local tsunamis illustrate the accuracy of this observation/model technology.
(b) Distant tsunami example
The 17 November 2003 Rat Islands tsunami in Alaska provides the most comprehensive test for the forecast methodology for distant tsunamis to date. The Mw 7.8 earthquake on the shelf near Rat Islands, Alaska, generated a tsunami that was detected by two tsunameters located along the Aleutian Trench—the first such detection by the newly developed real-time tsunameter system. These real-time data, combined with a model database, were then used to produce the real-time model tsunami forecast. For the first time, tsunami model predictions were obtained during the tsunami propagation before the waves reached most coastlines. The initial offshore forecast was obtained immediately after preliminary earthquake parameters (location and magnitude Ms=7.5) became available from the West Coast/Alaska Tsunami Warning Center, about 15–20 minutes after the earthquake. The model estimates provided expected tsunami time series at tsunameter locations. When the closest tsunameter recorded the first tsunami wave about 80 minutes after generation (figure 8), the model predictions were compared with the deep-ocean data to refine the model offshore predictions using data assimilation methods.
This offshore model prediction was then used as input for the high-resolution inundation model for Hilo Bay (Titov et al. 2005b). The model computed tsunami dynamics on several telescoping grids, with the highest spatial resolution of 30 m inside Hilo Bay. The tsunami did not produce inundation at Hilo, but a nearly 0.5 m (peak-to-trough) tsunami was recorded at the Hilo gauge. A comparison of model forecast and tide gauge observations (figure 9) demonstrates that amplitudes, arrival time and periods of several first waves of the tsunami wave train were correctly forecast (Titov et al. 2005b). More tests are required to ensure that inundation forecasts are reliable. When implemented, such a forecast will be obtained even faster and will provide enough lead time for potential evacuation or warning cancellation for US coasts threatened by distant-source tsunamis.
(c) Local tsunami example
On 14 June 2005, a magnitude 7.2 earthquake generated a non-destructive tsunami near Crescent City, California. A deep-ocean tsunameter/DART system detected the 0.5 cm tsunami 43 minutes after the earthquake, providing information to cancel the warning. A 3 cm tsunami with a period of 20 minutes was forecast for Crescent City, using tsunameter data ingested into a set of nested tsunami models similar to the example above. The observed tsunami at the tide gauge was masked in a noisy tide record, but post-event processing revealed that the signal was about 8 cm at a period of 20 minutes, which is the natural oscillation period of Crescent City harbour.
(d) International implications
Tsunami forecast products could be provided to all nations of the world that have a tsunami hazard. By establishing a global network of DART stations along with a set of tsunami models, global forecasts are within technical reach. The GEOSS framework would provide an appropriate home for the global tsunami warning system.
How might the preceding discussion on tsunamis be distilled into some guidelines for policy-makers?
First, it is important to recognize that tsunamis are but one hazard facing society, with many elements in common with other rapid-onset hazards. This leads to the first policy implication: tsunamis should be considered in the context of a global all-hazards programme.
Second, the public deserves state-of-the-science technology that adheres to scientifically accepted standards. Standards are not easy to establish but are required to protect the public's interest. Tsunamis also involve public safety over an extended period of time. Therefore, a second policy implication is: scientific standards for tsunami forecast products must be established and maintained by credible scientific organizations.
Third, since tsunamis are an infrequent event, it is natural to place emphasis on the hazard immediately following an event but then allow the effort to languish once interest wanes—until another tsunami repeats this destructive pattern. Therefore, a third policy implication is: comprehensive tsunami preparedness strategies that include protection, mitigation, recovery, training, exercises and research and development must be institutionalized. One promising institution for the global tsunami warning system is GEOSS.
The Sumatra tsunami of 26 December 2004 reminded the world that the tsunami hazard can be devastating for coastal communities unaware and unprepared for coastal flooding by multiple tsunami waves. Recent advances in the science of tsunami forecasting, communications technology and tsunami-resilient community development hold promise for a global tsunami mitigation system that would greatly reduce the impact of future tsunamis. There are no technical barriers restraining the development, implementation and maintenance of a global tsunami mitigation system that supports tsunami-resilient communities.
Before the next Sumatra-type tsunami strikes, we must resolve to create a world that can coexist with the tsunami hazard.
M. J. Purvis (University of Bristol, UK). Can the relationships of ‘p’ and ‘s’ wave arrival be used to determine the threat of development of a tsunami? Is there a threshold of earthquake magnitude that will tend to the formation of a tsunami?
E. N. Bernard. To date, no one has successfully linked the seismic ‘p’ and ‘s’ waves' arrival to tsunami formation. The earthquake threshold for ocean-wide tsunami formation has been Ms 7.5 or greater, with shallow (less than 50 km) hypocentre. However, smaller earthquakes are capable of generating local tsunamis. For example, in 1992 and 1998, Ms 7.0 generated killer tsunamis in Nicaragua and New Guinea.
J. E. Sharpe (Beal High School, Essex, UK). Would the building of artificial ‘reefs’, beach nourishment and other coastal protection schemes be a useful way of mitigating for another tsunami in the Indian Ocean, bearing in mind the likelihood of a future large earthquake on the ‘rocked’ parts of the fault? Perhaps this would be a cost-effective method of reducing mortality, coupled with education and signage in these areas, i.e. Podang in Indonesia, which has a population of 800 000.
E. N. Bernard. Coastal protection measures have been effective in Japan, but are extremely expensive. For the Indian Ocean, socially acceptable protection devices would help, but education is probably the most effective and least expensive mitigation tool available.
G. C. Mayberry (USAID Office of Foreign Disaster Assistance, Washington, DC, USA). The human response to tsunami warnings in the northwest USA seems to be pretty good, as was tested by the cancelled tsunami warning on 14 June. Has the human response in less developed countries threatened by tsunamis in the Pacific been satisfactory or is there a need for better information dissemination and education for at-risk communities?
E. N. Bernard. Any tsunami threatened coastal community would benefit from improved education and better information dissemination.
R. E. A. Robertson (University of the West Indies, St Augustine, Trinidad). In the context of plans for a global tsunami warning system and also GEOSS, does NOAA have any mechanisms in its modus operandi for working with local/regional scientific organizations that may already be involved in monitoring activities in the areas where installations are planned?
E. N. Bernard. GEOSS is a multi-national effort represented by many nations. The operating principles of GEOSS are for local scientific organizations to be part of their national effort and bring their plans to the international GEOSS meetings for coordination. NOAA, per se, represents the US efforts for GEOSS and welcomes participation by all nations of the globe.
One contribution of 20 to a Discussion Meeting Issue ‘Extreme natural hazards’.
- © 2006 The Royal Society