The magnitude of the threat posed by comets or asteroids that might collide with the Earth will be described. While the probability of collision is small, the effects including tsunamis could be devastating, suggesting that it should be carefully considered in relation to other natural disasters. It is one of the few natural disasters that could be averted by technical means. Although many more complex schemes are possible, the most cost-effective and the only currently-available means of disruption (deflection or pulverization) is a nuclear explosive.
The optimal tactics for terminal intercept and remote-interdiction scenarios will be described. The optimal mass ratio of an interceptor rocket carrying a nuclear explosive depends mainly on the ratio of the exhaust velocity to the object closing velocity. Nuclear explosives can be employed in three different modes depending on their location at detonation: (1) buried below the object's surface by a penetrating vehicle: (2) detonated at the object's surface; or (3) detonated some distance above the surface.
A model for gravitationally bound objects will be used to obtain the maximum non-fracturing deflection speed for a variety of object sizes and structures. For a single engagement, we conclude that the non-fracturing deflection speed obtainable with a stand-off device is about four times the speed obtainable with a surface-burst device. Furthermore, the non-fracturing deflection speed is somewhat dependent on the number of competent components of the object.
Generalizations indicate: (1) asteroids more than 3 km in diameter can
be the most efficiently deflected with a surface burst; (2) asteroids as
small as half km in diameter can be effectively deflected with a stand-off
device; (3) smaller asteroids are best pulverized.
Enigmatic geological features described from south Lanai, including assemblages of marine faunas, apparently in situ, at elevations of up to 329 meters, and the occurrence of gravel deposits containing coral clasts dated at 101-115 Ka have been attributed to a ``Giant Wave'' generated by a large submarine landslide off Hawaii approximately 105,000 years ago (Moore and Moore, 1984, 1988). It has also been suggested that this wave traveled across the Pacific Ocean and impacted the coast of southeastern Australia. the ``Lanai tsunami'' runup is an order of magnitude greater than tsunami runups in historic times. It is critical to assessments of tsunami risk to verify that such a wave did indeed occur.
Our review of evidence cited in support of the giant wave hypothesis, including ongoing field studies, leads us to question the validity of the hypothesis.
An examination of gravel stratigraphy at the type section has identified and characterized a complex sequence consisting of 15 beds, rather than 6 beds of alternating coral-rich and coral-poor basalt illustrated by Moore and Moore (1988), and interpreted by them to reflect the run-up and seaward return flow of each of three waves in a tsunami wave train. Our studies also show that deposition of the gravel sequence was punctuated by at least three periods of subaerial exposure of sufficient duration to allow soil formation, and several other erosional breaks. This is inconsistent with a wholly tsunamigenic origion for the deposit.
Other observational evidence cited in support of the giant wave hypothesis includes soil stripping from the land surface up to 375 meters, deposition of a continuous blanket of gravel (the Hulopoe Gravel) up the slopes of Lanai to 329 meters, and thinning and fining of this gravel landwards, could not be verified in the field.
A single giant wave event was suggested by Moore and Moore (1984, 1988) based on radiometric dating of a limited number of coral clasts from two locations, which yielded a narrow range of dates. Larger numbers of recently reported dates tend to cluster around 220 and 120 thousand years, periods of former high sea level stands. While these dates record ages of material in the deposits, they do not represent depositional age(s). Without stratigraphic control of sampling, there can be no assessment of whether dated material is eroded and re-deposited.
We suggest that the Hulopoe Gravel is a product of normal events and
processes occurring on a rocky, high-energy coast of a tropical oceanic
We investigate the generation, propagation, and probabilistic hazard of tsunamis spawned by oceanic asteroid impacts. The process first links the depth and diameter of parabolic impact craters to asteroid density, radius, and impact velocity by means of elementary energy arguments and crater scaling rules. Then, linear tsunami theory illustrates how these transient craters evolve into vertical sea surface waveforms at distant positions and times.
By measuring maximum wave amplitude at many distances from a variety of impactor sizes, we derive simplified attenuation relations that account both for geometrical spreading and frequency dispersion of tsunamis on uniform depth oceans. In general, the tsunami wavelengths contributing to the peak amplitude coincide closely with the diameter of the transient impact crater. For the moderate size impactors of interest here (those smaller than a few hundred meters diameter), crater widths are less than or comparable to mid-ocean depths. As a consequence, dispersion increases the 1/(sqrt r) long-wave decay rate to nearly 1/r for tsunamis from these sources. In the final step, linear shoaling theory applied at the wavelength associated with peak tsunami amplitude corrects for amplifications as the waves near land. By coupling this tsunami amplitude/distance information with the statistics of asteroid falls, the probabilistic hazard of impact tsunamis are assessed in much the same way as probabilistic seismic hazard by integrating contributions over all admissible impactor sizes and impact locations. In particular, tsunami hazard, expressed as the Poissonian probability of being inundated by waves 2 to 50 meter height in a 1000 year interval, is computed at both generic (generalized geography) and specific (real geography) sites.
For a conservative estimate of the impact flux, a typical generic site
with 180 degrees of ocean exposure and a 6,000 km reach admits a 1:23 chance
of an impact tsunami of 2 meter height or greater in 1000 years. The likehood
drops to 1:58 for a 5 meter wave, and to 1:476 for a 25 meter wave. Specific
sites of Tokyo and New York have 1:38 and 1:76 chances of suffering an
impact tsunami greater than 5 meters in the next millennium. We believe
that investigations of this style that merge proper tsunami theory with
rigorous probabilistic hazard analysis advance considerably the science
of impact tsunami forecasting.
A hazard is a potentially perilous event, such as a tsunami, while risk
is the probability that the hazard will occur repeatedly and affect a specified
population. Risk includes the frequency of occurrence, exposure, and magnitude.
The International Decade for Disaster Reduction has focussed attention
on assessing and mitigating the risk of tsunamis. Statistical and scenario
methods of determining risk for rare and more common events are discussed.
The problems of warning are considered, and a matrix illustrating the most
effective use of research, mitigation, and warning programs is presented.
Emphasis is on public safety, with due consideration of public property
factors. Examples of evaluation of relative risk are provided.
A review of historical data for locally generated tsunamis suggests average recurrence intervals of about 20 years for destructive tsunamis, with the last such tsunami occurring in 1975. Preliminary modeling indicates that a large tsunami generated on the Kona Coast could have significant destructive potential on other islands, especially on the south shore of Oahu. Unfortunately, the recurrence interval for such large tsunamis on the Kona Coast is not known. In evaluating local warning system capabilities and limitations, it should be noted that warnings based only on earthquake magnitudes will have an unacceptably high failure rate.
Incorporating conventional tide gauge readings into the decision making
process with magnitude determinations may moderately reduce this failure
rate. An acceptable warning system will require (1) many more wave recorders
than the three now present on the Big Island; (2) modeling studies of wave
heights or runups at instrumented sites for a suite of possible tsunamigenic
earthquakes along the Puna, Kau, and Kona coasts; (3) perhaps a new generation
of tsunami detectors; and (4) automated warnings for highly localized tsunamis.
Although the majority of the reported tsunamis are from littoral countries of the Pacific Ocean, there are a few cases of tsunamis in the Indian Ocean. The approximate length of the Indian coast is about 6000 kilometers. The coasts run from north to south and have two arms in the east and west with a tapering end at Kanyakumari. The tsunamigenic earthquakes occur mostly at the following three locations; (1) The Andaman sea, (2) Area about 400-500 kilometers SSW of Sri Lanka (Ceylon), (3) The Arabian Sea about 70-100 kilometers south of Pakistan Coast -- off Karachi and Baluchistan. The oldest record of tsunami is available from November 326 BC earthquake near the Indus delta/Kutch region. Alexander the Great was returning to Greece after his conquest and wanted to go back by a sea route. But an earthquake of large magnitude destroyed the mighty Macedonian fleet as reported by Lietzin (1974).
The earliest record of tsunami is reported to be about 1.5 meters at
Chennai (formerly Madras) which was created due to the August 8, 1883 Krakatoa
volcanic explosion in Indonesia. An earthquake of magnitude 8.25 occurred
about 70 kilometers south of Karachi (Pakistan) at 24.5 N and 63.0 E on
November 27, 1945. This created a large tsunami of about 11.0 to 11.5 meters
high on the coasts of India in the Kutchch region, as reported by Pendse
(1945). An earthquake of magnitude 8.1 occurred in the Andaman Sea at 12.9
N and 92.5 E on June 26, 1941 and a tsunami hit the east coast of India.
As per non-scientific/journalistic sources, the height of the tsunami was
of the order of 0.75 to 1.25 meters. At the time no tide gauge was in operation.
Mathematical calculations suggest that the height could be of the order
of 1.0 meter. There are a few more cases of earthquakes of magnitude less
than 8.0 which have given rise to some smaller tsunamis. Bapat, et al (1983)
have reported a few more earthquakes on the coast of Myanmar (formerly
Evidence of historic Cascadia subduction zone earthquakes and subsequent tsunamis have prompted hydrodynamic modeling efforts to identify potential flow patterns and coastal hazards for plausible future events. In this study we identify the methods used to derive potential seismic source scenarios and present a thorough evaluation of finite element simulations of the tsunamis associated with these scenarios. The first part of the paper deals with regional impacts of potential tsunamis, while the second part evaluates the fate of the modeled waves from the local perspectives of Seaside and Newport, Oregon. Both parts are composed of physical as well as numerical interpretations of the simulations.
Regional analyses of the simulations help identify the factors influencing the propagation of the waves from the source to the coastline. The local analyses will then evaluate the fate of the tsunami waves as they interact with the coast line and the topography of the land. Enough grid refinement is added to capture the dynamic intensity of the waves at local spatial scales. The physical interpretation of both these results provides clues to the determining factors in the fate of tsunami waves. The numerical interpretation, likewise, is a crucial component in helping to assess the usefulness of numerical models in evaluating and mitigating tsunami hazards. The identification of the limits of a numerical model and how those limits can be minimized in turn allows the physical mechanisms to be better represented and the mitigation to be more effective.
It is the mitigation, after all, which is the goal of a study such as this. For example, the state of Oregon is utilizing these results to estimate potential inundation patterns resulting from Cascadia tsunamis. These patterns have been translated into inundation maps identifying zones of low, medium, and high risk of flooding throughout coastal communities. The results presented for Seaside and Newport, Oregon identify the physical characteristics of the waves responsible for inundation in those communities. Such results are being used by the Oregon Department of Geology and Mineral Industries and NOAA to mitigate local tsunami hazards with inundation maps and community awareness.
A comprehensive version of this work is available on the Science of
Tsunami Hazards Web site at http://www.ccalmr.ogi.edu/STH/online/volume
17/ number 1/mbp/.
This paper provides information concerning an effective methodology of investigating past tsunami events developed following the geological investigation of recent, historical and paleohistoric tsunamis in the Aegean Sea region of Greece. The methodology described includes the use of contemporary and historical records, recent scientific publications, eye-witness accounts, geomorphological mapping and analyses and laboratory analysis of tsunami-deposited sediments. From the data presented, it will be seen that a multidisciplinary approach to the investigation of individual tsunami events is preferable because either insufficient or misleading evidence may be obtained when only one method is used.
The paper describes how geomorphology may be used to infer the magnitude
and likely effects of coastal tsunami flooding. Additionally, the paper
considers the accuracy of historical records of individual tsunami flood
events and how inaccurate recording can have important implications for
disaster preplanning and coastal vulnerability reduction. The Aegean tsunamis
of September 29, 1650 and July 9, 1956 are used to illustrate the methodology
described. Data are also presented which indicate that it is possible to
distinguish episodes of coastal tsunami flooding within the long-term geological
record on the basis of the microfossil (Foraminifera) assemblage. Such
analysis may assist in the reconstruction of the number of paleotsunami
events with specific coastal areas providing a valuable predictive tool
for future tsunami recurrence.
As Japan is a nation small in area and surrounded by seas, a potential threat of a destructive tsunami becomes a national event. The Japan Meterological Agency, an agency of the national government, has the mandate to issue tsunami warnings. By using an archive of precalculated tsunami scenarios, the agency can forecast wave heights for all the coasts of Japan, when the magnitude and epicenter of the generating earthquake are known.
Tsunami warnings and forecasts start from the cabinet level of the national
government and are transmitted through the various layers of the national
government, to the prefecture governments and eventually, in a matter of
minutes, to the local governments. Transmissions of the warning and forecasts
from the local governments to the general public is done through a variety
of media. The response of the warning system to the Sea of Japan tsunami
of July 12, 1993, was well documented and showed successes and loopholes.
Lituya Bay, Alaska is a T-Shaped bay, 7 miles long and up to 2 miles wide. The two arms at the head of the bay, Gilbert and Crillon Inlets, are part of a trench along the Fairweather Fault. On July 8, 1958, a 7.5 Magnitude earthquake occurred along the Fairweather fault with an epicenter near Lituya Bay.
A mega-tsunami wave was generated that washed out trees to a maximum altitude of 520 meters at the entrance of Gilbert Inlet. Much of the rest of the shoreline of the Bay was denuded by the tsunami from 30 to 200 meters altitude.
The SWAN code which solves the nonlinear long wave equations was used to numerically model possible tsunami wave generation mechanisms.
A rockslide of about 30 million cubic meters was probably triggered by the earthquake. It has been assumed to have been the source of the tsunami wave even though it was difficult to correlate with the eye-witness observations. Numerical studies indicated that the tsunami wave generated by the rockslide gave tsunami wave inundations that were less than a tenth of those observed if the slide was assumed to lift a volume of water corresponding to the volume of the slide to above normal sea level.
Another possible source of the tsunami was a massive uplift of the sea floor along the Fairweather Fault that underlies the Gilbert and Crillon Inlets at the head of the bay. Even if all the water in the inlets was initially raised to above normal sea level, the observed tsunami inundations could not be numerically reproduced. Since it appeared that a much larger source of water than was available in the inlet was required, the one possible source was a partially subglacial lake near the sharp bend in Lituya glacier which flows down Gilbert inlet. The level of the lake was observed to have lowered 100 feet after the earthquake. After the earthquake the glacial front of Lituya Glacier had become a nearly straight wall and about 400 meters of ice had been sheared off of the glacier front.
Various models of water flowing from breaking glacial dams were studied but they did not reproduce the observations.
Dr. George Pararas-Caryannis suggested that a tsunami wave was formed
by a rockslide impact similar to an asteroid impact making a cavity to
the inlet ocean floor and a wave that splashed up to 520 meters height.
If the run-up was 50 to 100 meters thick, adequate water is available between
the slide and the run-up and the results are consistent with the observations.
Further studies will require full Navier-Stokes modeling similar to those
required for asteroid generated tsunami waves.
At 7:10 p.m. on November 3, 1994, a large tsunami generated by a massive landslide in the submerged Skagway River delta occurred near Skagway, Alaska, resulting in one fatality and damaging or destroying many harbor structures.
At first, it was theorized by some that construction activity in the harbor caused the initial landslide. However, this paper presents the findings of an in-depth scientific investigation that concludes that such a theory is impossible.
The findings paint a clear picture of the failure of the submerged Skagway River delta that was overloaded by flood sediments and exacerbated by river diking. Slide volumes estimated at over 20 million cubic yards that consisted of a massive initial slide and subsequent retrogressive earth slide produced the tsunami that caused one fatality and destroyed or damaged harbor structures.
The analysis relies on physical evidence and reconstructs the tsunami
on a second-by-second time-line that shows conclusively that the failure
of the submerged Skagway River delta was not caused by the harbor construction.
Each shred of evidence is examined and the event systematically reconstructed
on a step-by-step basis without interjecting supposition, speculation,
theory or hypotheses.
The most remarkable difference of tsunami characteristics due to a landslide and an earthquake is the movement of the source region. For a landslide, the source region moves horizontally. For an earthquake the source region (usually 100 kilometer or more wide) only moves vertically, this volume change converted to sea surface. So the long wave approximation is valid for tsunamis generated by earthquakes, but not for landslides.
A new simulaton method was developed for submarine landslide tsunamis based on the combination of analytical and numerical calculations. The landslide results showed strong directivity compared with tsunamis generated by earthquakes.
For the detection and early warning for these tsunamis, it is necessary
to observe not only tsunami wave heights but also its directivity. For
this purpose, the present status of tide gauge distribution even in the
Pacific region is not adequate. I propose a cable system using laser-tsunami
meters. The total cost of the cable system including installation will
be inexpensive compared with the cable systems currently deployed around
In spite of significant advances in our understanding of the science of tsunamis, the basic facts about the dangers of tsunami waves are not understood by the general public. Tsunamis are the most deadly natural disaster facing those living in the Hawaiian Islands, having resulted in some 291 fatalities since 1837. The town of Hilo in particular has suffered great destruction and loss of life with 177 victims, therefore making it an appropriate site for a museum focused on tsunamis.
In mid-1998 the Pacific Tsunami Museum opened in downtown Hilo. The museum has two goals: (1) to preserve the local history of tsunamis in Hawaii as a memorial to those lost, and (2) to prevent future loss of life from tsunami waves by fostering tsunami education, preparedness, and other mitigating measures. These two goals are compatible and produce a powerful synergism. The local history of tsunamis in Hawaii contains many true stories of tragedy, sacrifice and heroism, as well as accurate descriptions of the tsunami run-up phase. It is the power of these true stories as told by the survivors themselves which has the ability to capture the imagination and educate audiences of residents and visitors who most need to understand the danger of tsunamis.
Funded by private donations and a grant from FEMA, the museum is currently planning a dozen permanent exhibits to be installed by the end of 1999. An ambitious outreach program is already underway and includes development of curriculum packages for all public and private schools statewide, plus specialized literature targeted at specific groups including surfers, boaters, visitors, businesses occupying inundation areas, etc.
The museum has established an archive collection of photographs, films,
videos, and artifacts, which will be made available to other educational
organizations around the world, and has already assisted in the production
of television documentaries aired nationally including those produced by
the National Geographic Society, the Discovery Channel, and the History
Channel. A tsunami education special was produced by KGMB-TV and shown
in Hawaii during Tsunami Awareness Month. Future plans for exhibits, educational
programs, and possible alliances with the tsunami research community will
Tsunami run-up and recurrence may be reconstructed from sediment sequences in near-coastal lakes. We present a partial chronology for tsunamis generated at the Cascadia subduction zone from an analysis of sediments in Kanim, Catala, Deserted, and Kakawis lakes on the west coast of Vancouver Island, British Columbia. Basal marine sand, gravel and shell in these lakes are overlain successively by fine-grained lagoonal sediments and freshwater gyttja. The change from a marine environment through an intertidal environment to a freshwater one is the result of regional uplift at a rate of about 1 meter per thousand years.
Inferred tsunami deposits in the lagoonal sediments and gyttja consist
of massive to graded sand or gravel overlain by, or interbedded with, thin
layers of forest detritus. The tsunami deposits commonly contain marine
microfossil assemblages which are strikingly different from the microfossil
assemblages in the enclosing gyttja. As a result of regional uplift, the
lakes progressively emerge above the zone of tsunami influence. A complete
tsunami history for northern Cascadia, therefore, requires systematic sampling
of lakes across a range of elevations. Currently we have evidence of major
tsunami events dating from about 300, 1000-1400, 1600-1700 and 2700-2800
years ago. The overlap between these dates and the ages of inferred earthquakes
at the Cascadia subduction zone suggests that all of these tsunamis were
locally generated. Calculations of runup for these events are complicated
by local variations in shoreline configuration, and the fact that the distance
between the lake and the sea commonly increases as the lake emerges, but
initial estimates suggest that runup magnitudes on the outer coast average
less than 5 meters.
At about 12:30 p.m. (local time) on September 14, 1953, the city of Suva was devastated by an ML 6.5 earthquake and associated tsunami of local origin. The earthquake source was about 25 km SW of Suva and the tsunami generation was attributed to submarine landslides (turbidity currents). The main industrial area and shore and harbour facilites of Suva were severly damaged. As part of the UNDHA - South Pacific Programme Office ``South Pacific Disaster Reduction Programme'', within the auspices of the Pacific Region IDNDR and the 1994 Yokohoma Statement, the ``Suva Earthquake Risk Management Scenario Pilot Project'' (SERMP) was faciliated for the Government of the Republic of Fiji. SERMP considered mitigation measures for both earthquake and tsunami impacting upon the city of Suva, with the scenario event based on the real experience of the 1953 Suva earthquake and tsunami.
A specific tsunami mitigation methodology was developed involving a multidisciplinary approach with multi-agency cooperation to address in both quantitative and qualitative terms, the premise
RISK = HAZARD X VULNERABILITY
and then integrate the assessments in terms of Fiji's emergency management requirements.
The outcomes include hazard, vulnerability and risk zonation maps with
associated commentaries, estimates of relevant tsunami parameters and possible
damage situations. It was concluded that a significant risk of a local
tsunami does exist for the city of Suva and its harbour environs. Practical
applications of these results, in terms of community vulnerability and
reduction of potential losses, and including a simulated tsunami exercise,
have been a major element in this project. This information resource has
been implemented for Fiji's National Disaster Management Office in terms
of disaster planning, response actions, training and community education.
Currently, Fiji is developing its own regional tsunami warning system.
Recent tsunami disasters, like that in Papua New Guinea in July 1998, serve
to reinforce the vital need for mitigation measures in these vulnerable
coastal communities of Pacific Island nations.
The Australian continent has not been impacted by large or devastating tsunamis since European colonization over 200 years ago. As a consequence this country is normally viewed as largely protected from impact of, or too far removed from the source of these hazards.
Considerable evidence exists, however, to show that very large waves
have struck the coast of Australia in the relatively recent past. This
evidence occurs along the eastern, northern and western Australian shores
and also occurs within seemingly protected areas such as along the mainland
coast inside the Great Barrier Reef in northeast Queensland and within
the Gulf of Carpentaria which is an epicontinental sea with a maximum depth
of 60 meters. The evidence is in the form of shell and coral deposits on
top of headlands many tens of meters in height, sand deposits containing
large boulders, shell and coral 20 - 30 meters above modern sea level and
several kilometers inland, fields of large imbricated boulders across shore
platforms and sculptured bedrock forms. The size of the transported boulders
together with numerical modeling, and the heights above sea level of these
deposits suggests that tsunamis are responsible as opposed to large storm
waves. The orientation of boulders and bedform deposits provide paleowave
directions for much of the continent's coast allowing reasonable estimation
of the source and possible generating mechanism of the tsunami. Carbon
dating of these deposits show that at least two very large tsunami events
have occurred along this coast during the last millennium.
The natural hazard of tsunamis relative to Australia and its Island Territories has been perceived to be of little or no consequence -- and hence a small risk -- when compared to other more frequent natural disasters of meteorological origin, or even occasional earthquakes. The historical record shows that tsunami damage, although rare, has occurred along the eastern seaboard (from the 1877 and 1960 Chile earthquakes), and northwest coast (from the 1883 Krakatoa (Indonesia) volcanic eruption and the 1977 and 1994 (Indonesia) earthquakes) of the continent. Because of the infrequent occurrences of tsunamis, they are little known and, in some cases, have been forgotten. However there is a need for tsunami mitigation, because, as an island nation, Australia is totally dependent on its coastal facilities for sustainable development, with more than 90% of the population domiciled in this environment. Recent devastating tsunamis in the Pacific region emphasise this need.
As part of Australia's contribution to the United Nations IDNDR (1990-2000) program, Emergency Management Australia's IDNDR Coordination Committee specifically directed one project to access the risk of tsunamis on the shorelines of Australia and its island territories. A specific methodology was developed, invoking a multidisciplinary approach to quantitatively and qualitatively define the hazard and the vulnerability, and then integrate these elements into a comprehensive risk assessment. More than 350 earthquakes and specific submarine volcanoes and landslides were considered as possible tsunamigenic sources. In the period 1788 through 1995 more than 60 registrations on tide-gauge records were identified, together with anecdotal information. The outcomes have been presented as an ``information resource'' in terms of hazard, vulnerability and risk assessment maps and commentaries, comprehensive tsunami data base, maps of potential tsunamigenic sources, tsunami travel time charts and relationships between relevant tsunami parameters.
These outcomes have been delineated in terms of proactive applications
necessary to upgrade both tsunami warning procedures by the Bureau of Meteorology
and response actions through counter disaster planning by the emergency
service authorities. As such, Australia is currently developing its own
regional tsunami warning system.
Hazard was estimated from historical records, tide gauge records and similar sources of information. Between 1788 and 1995 a total of 65 tsunami events were identified. Most were minor. There were three reports of damage, including the tsunami resulting from the 1960 Chilean Earthquake which resulted in "Considerable [damage], from Brisbane to Eden; most severe at Sydney, Evans Head". Apparently there were no reports of casualties. Based on this assessment two coastlines were rated as having a "high" tsunami hazard: the New South Wales coast near Sydney and the Western Australian coast near Broome. No attempt was made to quantify the hazard due to the lack of data.
Vulnerability was estimated from an assessment of built environment, including residential communities and industries along the coast, and natural environment such as significant coastal geography. The East Coast of Australia between Wollongong (just south of Sydney) and Cooktown (Far North Queensland) was found to be highly vulnerable to tsunami.
The coastline between Wollongong, Sydney and Newcastle (the three major population centres of New South Wales) therefore has both a high hazard and a high vulnerability to tsunami. It is rated as having the highest tsunami risk in Australia.
Further research is needed to quantify the risk and to develop mitigation measures (discussed in the paper). It is evident that the Sydney coastline would make a logical starting point for any detailed study of tsunami risk.
Four potential sources of tsunami are described: earthquakes, undersea volcanoes, submarine landslides and asteroid/comet impacts. Earthquakes from anywhere in or around the Pacific and Indian Oceans are a potential threat, as are submarine volcanoes in Indonesia and Polynesia - based on the historical records presented by the authors, damaging tsunami (say 1 or 2 m run-up height) from these sources can be expected over intervals of several decades. Submarine landslides are a speculative source of tsunami in the Tasman Sea (East Coast of Australia) - research by Ted Bryant suggests an interval between devastating tsunami events (say a run-up height 10m or more) along the NSW South Coast of several hundred years and it is possible that submarine landslides are a source of these events (note that Rynn and Davidson advise caution when interpreting such pre-historic events). Tsunami from asteroid impacts into the South Pacific, Indian and Southern Oceans "cannot be discounted". Fortunately such events are thought to be very rare. My own rough estimates of the asteroid tsunami risk suggest that the average interval between events sufficient to cause a damaging 2 metre tsunami somewhere along the coastline of Australia is around 100,000 years (assuming a typical tsunami run-up factor of 5). A much more devastating 10m tsunami probably has an average interval of around one million years but it should be noted that impacts do not run like clockwork - there is mounting evidence that the Earth is subjected to a barrage of impacts from time to time. This may have occurred in the case of the tsunami events reported by Nott and Bryant since asteroid/comet impacts are a likely source of giant tsunami.
The Central American Coasts have been hit by nine destructive tsunamis during the last two centuries. Seven of these tsunamis are from the Pacific and two from the Caribbean. Reported damages range from coastal and ship damage to destruction of small towns. Almost 500 people have been killed by these tsunamis. The Pacific coast of Central America has higher tsunami hazard than the Caribbean Coast. Tectonic environments that generate tsunamigenic earthquakes are the Middle American Trench, the Polochic-Motagua Fault System and the North Panama Deformed Belt (NPDB).
A Tsunami Warning System for Central America has been designed. This
system uses earthquake magnitude as the trigger for tsunami warning. Three
institutions are involved in this system: The Instituto de Estudios Territoriales
de Nicaragua (INETER), the Central American Seismological Center (CASC)
and the National Emergency Office (NEO) of each country. CASC locates the
earthquake and determines the magnitude and sends the seismic information
to INETER. This institution evaluates the seismic information and decides
if the earthquake has potential to generate a tsunami. In the event of
a tsunamigenic earthquake INETER issues a tsunami warning which is sent
to the National Emergency Office (NEO). NEO activates the local emergency
plan and takes actions to protect coastal residents.
The giant waves that rose to a maximum height of 1720 feet at the
head of Lituya Bay on July 9, 1958 were generated by a combination of
disturbances triggered by a large, 8.3 magnitude earthquake along the
Fairweather fault. Several mechanisms for the generation of the giant waves
have been proposed, none of which can be conclusively supported by the data
on hand. Factors that contributed to the giant waves in Lituya Bay were the
result of cumulative effects rather than those from a single source.
Possible generative causes include a combination of tectonic movements
associated with the earthquake, movements of a tidal glacier front, a major
subaerial rockslide/landslide in Gilbert Inlet, other subaerial or
submarine sliding at the head of the Bay, and the possible sudden drainage
of a subglacial lake on the Lituya Glacier. These factors are examined, as
well as the near field strong ground motions associated with the
Dynamic earthquake ground motions lasting 40-60 seconds or more,
rather than net crustal displacements, may have contributed significantly
to the generation of the giant waves, particularly because of the
orientation of the seismic disturbance and the upper Lituya Bay's physical
dimensions, geometrical configuration and orientation with the Fairweather
fault. Upper Lituya Bay response and the associated secondary phenomena,
contributing to the giant slushing wave action in Gilbert Inlet, depended
on the earthquake's energy release, proximity to the epicenter, physical
rupture along the fault, propagation path of surface seismic waves, and the
magnitude and duration of the dynamic, near-field, strong motions.
Earthquake ground motions of high intensity (XI, XII on the Modified
Mercalli scale) could have resulted in vertical accelerations of up to
0.75g and horizontal accelerations of as much as 2.0g. In the absence of
adequate data for the Lituya Bay event, analogies are drawn from recorded
recent large earthquakes elsewhere, such as the 17 January 1994 Northridge
earthquake in California, for their characteristics of near field ground
motions, duration, and the extent of vertical and horizontal accelerations.
Additionally, the tectonic setting of the Fairweather fault is examined as
characterized by past events as the September 4, 1899, Cape Yakataga
The following mechanism can account for the giant 1720 foot wave
runup at the head and for the wave along its main body of Lituya Bay:
Almost immediately, the strong ground motions of the earthquake triggered a
giant landslide/rockslide at the headland of the bay. Almost,
instantaneously, this rockslide/landslide, acting as a monolith and thus
resembling an asteroid, impacted with great force the bottom of Gilbert
Inlet. The impact created a crater which displaced and folded recent and
Tertiary deposits and sedimentary layers. The
displaced water and the folding of sediments broke and uplifted 1300 feet
of ice along the entire front of the Lituya Glacier. Also, the impact
resulted in water splashing action that reached the 1720 foot elevation on
the other side of inlet. The same landslide impact, in combination with
strong ground movements, the net vertical crustal uplift of about 1 meter,
and an overall tilting seaward of the entire crustal block on which Lituya
Bay was situated, generated a solitary gravity wave which swept as an edge
wave the main body of the bay. An analytical solution based on this
proposed impulsive mechanism can further support the 1720-foot runup.
Mathematical modeling studies conducted by Dr. Charles Mader, support this
mechanism as there is a sufficient volume and an adequately deep layer of
water in the Lituya Bay inlet to account for the giant wave runup. Dr.
Mader has suggested full Navier-Stokes modeling, as with asteroid generated
Necessary focus of future research in understanding mega-tsunamis in
enclosed bodies of water, such as the Lituya Bay, should be directed
towards the examination and modeling of the elements relative to the
earthquake energy release, the empirical analysis of earthquake source and
seismic energy propagation processes, the near-field ground motions from
finite fault sources of past mega-thrust earthquake events, and the
systematic studies of resulting secondary effects. Additionally, measurable
input and output parameters
derived from mathematical modeling and analysis of the Lituya Bay event can
be further applied to models of asteroid tsunami generation for purposes of
calibration, verification and validation.
Shortly after 7 PM local time on July 17, 1998, more than 10 km ofthe
northern PNG coastline home to at least 10,000 people was swept clean by
water approximately 10 m high. More than 2,200 people perished in the
torrent or shortly thereafter. The scale of the PNG tragedy coupled with
unexpectedly large water wave amplitudes for the earthquake size and the
local geological complexity motivated an intense international scientific
effort to assess if the tsunami was triggered by coseismic displacement or
by mass movements.
The July 17, 1998, tsunami that struck Sissano, Sandaun Province, Papua New
Guinea (PNG) is the first major tsunami linked directly to a giant mass
movement. This event is also a milestone in that modeling efforts have been
simultaneously informed by marine surveys and geological analyses carried
out on the Kairei (KR98-13) and Natsushima (NT99-02) joint Japan Marine
Science and Technology Center (JAMSTEC) and South Pacific Applied
Geoscience Commission (SOPAC) cruises. We describe the first effort to
model tsunami generation, propagation, and coastal interaction based on
recent bathymetric data, geological interpretation, and tethered ROV
(Dolphin 3K) investigations of the seafloor.
The Alaska Tsunami Warning Center (ATWC) was established in Palmer, Alaska in 1967 as a direct result of the great Alaskan earthquake that occurred in Prince William Sound on March 27, 1964. In 1996, the responsibility was expanded to include all Pacific-wise tsunamigenic sources which could affect California, Oregon, Washington, British Columbia and Alaska coasts and the center became the West Coast/Alaska Tsunami Warning Center (WC/ATWC).
As a result of the NOAA Tsunami Hazard Mitigation effort, the flow of
real-time seismic data to
the WC/ATWC has been significantly increased with the implementation of the Earthworm system. This front-end system provides the WC/ATWC the ability to receive/transmit digit al seismic data with
others who have an Earthworm system. Integrating this front-end with the WC/ATW C processing
system permits over 80 channels of vertical short-period and long-period seismic data, and broadband
seismic data to be recorded and processed at the WC/ATWC. To process this inform ation, the
automatic processing system has been enhanced to provide a geophysicist with many other useful results and displays to aid in the issuing and concluding warnings.
An on-going project at WC/ATWC is the prediction of tsunami amplitudes outside the tsunami generating area described in Science of Tsunami Hazards 14, 147-166 (1996). The basic idea behind this technique is that pre-computed tsunami models can be scaled by recorded tsunami amplitudes during an earthquake to give a reasonable amplitude estimate outside the source zone. Tsunami models for moment magnitude 7.5, 8.2, 9.0 earthquakes have been computed along the Pacific plate boundary from Honshu, Japan to the Cascadia subduction zone. The modeling technique was verified by comparison to historic tsunamis from different regions. At present, the results from the scaled models are not distributed to the emergency officials during warnings, but are used only internally as an aid in canceling or extending warnings.
Another current project at WC/ATWC is to receive tsunami data from the Pacific-wide tsunami sites via a satellite phone system. Due to the current delay of 1 to 3 hours in receiving tide data from NOS gauges, selected windows of data can be received from these tide sites using a satellite phone, antenna, PC computer and special hardware inserted into the current NOS field packages. This will permit obtaining data from selected tide sites nearest the tsunami source.
WC/ATWC conducts a community preparedness program which provides advice
and training sesssions to coastal citizens and emergency managers to aid
in pre-event planning. The aim of the program is to educate the public
to help themselves if they are caught in the middle of a violent earthquake
and/or tsunami, and to be aware of the safety procedures, safe areas, and
the limitation of the Tsunami Warning System.
Of the 140 impact craters known on the surface of Earth, the most famous was created about 65 million years ago when a 10 km asteroid or comet came down in shallow water near the present day town of Chicxulub, Mexico. With a kinetic energy equivalent to 100 trillion tons of TNT, the impact event lofted enough debris onto globe-straddling trajectories to flash heat much of the surface of the Earth and then darken the skies for several years.
Numerous investigations have demonstrated that such an event, which happens, on average, every 100 million years, caused extreme stress on Earth's climate and most likely led to the extinction of many species. Recent high fidelity computational simulations demonstrate that more numerous asteroids or comets as small as 1-2 km in diameter, impacting, on average, every 300,000 years may be globally catastrophic-producing large tsunamis and lofting debris to high altitudes worldwide. Indeed, the odds of an individual dying from a relatively frequent 1-2 km impacting object (about 1 in 10,000) are substantially greater than from the impact of an infrequent dinosaur killer (1 in 1,000,000).
What can we do to reduce the hazard from impacting comets and asteroids? Recent computational investigations by Asphaug et al, suggest that weakly bound asteroids (little more than rubble piles) are easier to break than deflect (E. Asphaug, S. J. Ostro, R. S. Hudson, D. J. Scheeres and W. Benz (1998), Nature, Vol. 393, pp. 437-440.). Is this an advantage or disadvantage? Clearly, the mechanical and compositional properties of asteroids and comets need to be better understood if viable deflection technologies are to be developed. Because the detection time prior to impact may be months (long period comet) to years (asteroid or short period comet several orbits prior to impact), it is possible that we may not have much time to perform such studies if faced with an actual threat.
Related web sites:
In the U.S.A. tsunami warnings are issued by the West Coast/Alaska Tsunami Warning Center and the Pacific Tsunami Warning Center, but the practical task of informing the general populace and evacuating people from potential tsunami inundation areas are the responsibility of state authorities. Of the various Pacific states, Hawaii has the most developed and proven tsunami warning and evacuation system. The warning systems of other states -- Alaska, California, Oregon and Washington -- are at various stages of approaching the Hawaii model.
A federal agency, the Pacific Marine Environmental laboratory, a division
of NOAA, has established the National Tsunami Mitigation Program to assist
The Asteroid Tsunami Program has as one of its objectives the evaluation of the inundation of major coastal cities of the world to be expected from mega-tsunamis generated by asteroids. The regions studied include Japan, the U.S. East Coast, Los Angeles, San Francisco, the Oregon coast, the Hawaiian Islands, Iceland and the European coast.
Some of the inundation studies have been published. Computer animations generated as part of the studies are also available. Many of the animations are included in the directory TSUNAMI.MVE on the Science of Tsunami Hazards CD-ROM included with the Tsunami Symposium Abstracts.
The following publications describe some of the inundation modeling
shown in the animations.
"Asteroid Tsunami Inundation of Hawaii,'' Science of Tsunami Hazards, 14, 85-88 (1996)
"Asteroid Tsunami Inundation of Japan,'' Science of Tsunami Hazards, 16, 11-16 (1998)
The tsunami inundation of the U.S. East Coast is described in part in
"Tsunami Produced by the Impacts of Small Asteroids ,'' Jack G. Hills and Charles L. Mader, New York Academy of Sciences 822, 381-394 (1997).
Greater numbers of civil defense and other pertinent personnel are becoming aware of the nature of the tsunami hazard in the Caribbean. Interest is growing for establishing a warning system and creating plans for mitigation of damages, search and rescue operations, and education of the public and key officials, coordinated among the numerous political divisions in the area, and perhaps land use planning, engineering, and insurance programs. A history of prior occurrences and effects is key to understanding the local nature of the hazard and for designing the most effective plan for warning and mitigation systems. We have submitted a paper, ``Caribbean Tsunamis: A 500 Year History, 1498 to 1998,'' with data on 88 Caribbean Tsunamis for publication.
The hope is to establish a warning system and general education on the nature of the hazard before the next disaster. Plans to establish a region-wide Tsunami Warning System received a boost with the involvement of the IOC-IOCARIBE in planning, coordination, and other assistance, beginning in 1996, at the May Caribbean Tsunami Workshop, held on St. John, U.S. Virgin Islands. Because most tsunamis are quite small, relatively rare, and do little damage, they have often been overlooked as a natural hazard until a disastrous event occurs. The region has averaged about one damaging tsunami every 26 years, but since they have not had one in 53 years, a destructive tsunamis is overdue. With the increase in population, tourism, coastal development, and also rising sea levels, the hazard is greater. The 500-year history shows that tsunamis in the Caribbean have the potential to produce major regional or local disasters. These can be mitigated through proper preparation.
The state of tsunami preparedness in the Caribbean today is similar
to that in the Pacific prior to the establishment of the Pacific Tsunami
Warning System. Without a warning system little or nothing could be done
to mitigate disaster. As many as 9600 fatalities have been reported as
due to tsunamis and tsunamigenic earthquakes in the Caribbean. The upcoming
area-wide workshop organized by the IOC-IOCARIBE, ``Intra-American Sea
Tsunami Warning System,'' on April 24-26, 1999, at San Jose, Costa Rica,
will set the course of the Caribbean Warning System, helping to determine
the tsunami hazard mitigation needs.
We are preparing a global historical tsunami catalog for the period from 1983 through the end of 1999. This catalog will bring the previous Pacific Catalog by Soloviev up to date to the end of the millennium. It will, however, also include tsunamis from all worldwide sources available to make it the first global tsunami catalog.
We have collected data on 136 tsunamis so far, including 119 from the Pacific, 9 in the Mediterranean, 5 in the Caribbean, and one each from the South China Sea, the Seychelles Islands in the Indian Ocean, and the Gulf of Aqaba near the Red Sea. Efforts to promote accuracy and completeness include contacts with local sources to see if smaller earthquakes caused unreported small tsunamis and a call for the preservation of maregrams and other data to fortify the record.
The Millennial Tsunami Catalog manuscript will be posted on a website
for review before being published by the National Geophysical Data Center
early in the year 2000. After publication, additional tsunami reports,
updates, and corrections will be added to the online catalog. An up-to-date
global historical catalog of tsunamis could thus be published directly
Operational activities of the Pacific Tsunami Warning Center can be classified into four major categories: collection and processing of seismic data, collection and processing of water level data, decision making, and dissemination of message products. Over the past few years significant improvements have been made or are underway in all these areas that are enabling PTWC to be faster, more accurate, more reliable, and more effective. The amount and quality of continuous seismic waveform data received at PTWC has increased substantially with the acquisition from the US Geological Survey of an Earthworm seismic data collection and processing system, and establishment of high-speed dedicated digital links between PTWC, the West Coast / Alaska Tsunami Warning Center, the National Earthquake Information Center, and the Hawaii Volcanoes Observatory. Water level data in near real time are now being received from five stations along the Pacific coast of Japan; two new stations are scheduled for installation this summer in the Kuril-Kamchatka region; and ten new stations are being installed along the coast of Chile.
Development by the Pacific Marine Environmental Laboratory of real-time-reporting
deep-ocean pressure gauges is continuing, with up to six gauges in the
north and northwest Pacific planned for deployment over the next few years.
Better and faster techniques for estimating tsunamigenic potential and
predicting impacts using seismic and water level data have recently been
developed and are being implemented in the decision-making process for
warnings and cancellations. Message dissemination continues over longstanding
dedicated circuits such as GTS and AFTN, while new opportunities for sending
both text and graphical information over higher bandwidth and more widely
available links such as the Internet and the Emergency Managers Weather
Information Network are being developed.
Tsunamis may be the most devastating source of economic damage caused by asteroid impacts. The worldwide darkness, which may last several months, caused by large asteroid impacts, such as occurred after the KT impact, may kill more people by mass starvation, especially in developing countries, than tsunami, but the dust should not severely affect the economic infrastructure. The tsunami may even kill more people in developed countries with a large coastal population, such as the United States, than would worldwide darkness.
At Los Alamos we are in the middle of a systematic study of asteroid tsunami. The study is divided into three parts: a determination of those regions of the world that are most susceptible to asteroid tsunami by simulating the effect of an asteroid impact into mid-ocean, the simulation of the formation of the initial crater by use of an SPH code, and a Monte Carlo study of the accumulative effect of many small impactors on some of the more strategically valuable regions that we find to be particularly vulnerable in the first part of this study. The first part of the study is well underway. Progress has been made on the other two.
The critical factor in the third part of the study is to accurately determine the dispersion in the waves produced by the smaller impactors. Dispersion may greatly reduce the effectiveness of the smaller impactors at large distances from the impact point. We wish to understand this effect thoroughly before going to the Monte Carlo study. We have modeled mid-Atlantic impacts with craters 150 and 300 km in diameter. We are proceeding to Pacific impacts. The code has been progressively improved to eliminate problems at the domain boundaries, so it now runs until the tsunami inundation is finished. We find that the tsunami generated by such impacts will travel to the Appalachian mountains in the Eastern USA. We find that the larger of these two impacts would engulf the entire Florida Peninsula. The smaller one would cover the Eastern third of the Peninsula while a wave passing through the Gulf of Cuba would cause the inundation of the west coast of Florida.
Top of page
This page prepared by Michael Paine,
The Planetary Society Australian Volunteers
21 April 1999