8:35 - Relationship Between Tsunami Calculations and Physics: - Z. Kowalik, UAK
9:05 - The Momentum of Tsunami Waves - H. G. Loomis, Honolulu
9:35 - Volcanically Generated Tsunamis - G. Pararas-Carayannis, Honolulu
10:20 - Lituya Bay Mega-Tsunam i - H. Fritz, VAW, C. Mader, MCCO
10:50 - Water Cavity Generation - C. Mader, G. Gittings, LANL
11:20 - Asteroid Generated Tsunam is - R. Weaver, G. Gisler, M. Gittings, LANL
1:00 - NTHMP Inundation Mapping Program - F. Gonzalez, V. Titov, etc, PMEL
1:30 - Tsunami Hazards Maps of Alaska -R. Hansen, etc. - UAK
2:OO - Review of 1994 Skagwq and Future Plans -D. Nottingham, PN&D
2:20 - Locally Generated Hawaii Tsunamis - D. Walker, R. Cessaro, Honolulu
3:lO - NOAA Water Level Observations Network - J. Hubbard, S .Duncan, NOAA
3:40 - Deep Ocean Detection of Tsunamis - E. Bernard, F. Gonzalez, etc., PMEL
4:lO - Session Summary - E. Bernard, PMEL
8:30 - Introduction to Workshop - B. Keating, UH
9:00 - Storegga Submarine Landslides and Tsunami - J. Mienert, Norway
10:30 - Mass Flow in Marine Sediment - M. Max, MDS
11:OO- Global Distribution of Gas Hydrates - W. Sager, Texas A&M
11:30 - Gas Hydrate Destablieation on Oregon Margin - C. Goldfinger, etc, OSU
1:OO - Gas Hydrates in New Zealand - C. Helsley, UH
1:30 - Marine Methyl Hydrates - S. Masutani, UH
2:00 - Numerical Simulation Landslide Tsunam i - Y. Shigihara, F. Imamura, Japan
2:30 - Gas Hydrates on Western Coast of North America- C. Synolakis, USC
8:30 - Tsunam i Vulnerability in Greece- M. Papathoma, D. D-Howes, etc., UK
8:55 - Tsunami Hazards in Canada - J; Clague, T. Murty, etc, Canada
9:20 - Tsunami Hazard in Indonesian Region - J. Rym, CEBA
9:45 - Tsunamigenic Efficiency of Stratovolcanoes - G. Pararas-Carayannis, Honolulu
10:30 - TOPICS and Geowave - P. Watts, etc. - AFE
10:55 - Probabilistic Tsunami Hazard and Risk Assessment- R. Sewell, RTSAC
1:00 - Elevated Strandlines on Lanai - B. Keating, UH
1:25 - Tsunamis in Cyprus - F. Whelan, D. Kelletat, U Bamberg, Germany
1:50 - Tsunamis on Aruba - A. Scheffers, U Essen, Germany
2:15 - Megaclast Emplacement and Transport on Oabu - R. Noormets, etc, UH
RELATIONS BETWEEN TSUNAMI CALCULATIONS AND THEIR PHYSICS
Institute of Marine Science
University of Alaska, Fairbanks, AK 99775
Spatial and temporal resolution of tsunami waves through numerical computations is closely related to tsunami physics. Presently available bathymetry, for computation of the transoceanic tsunami propagation, is given on 5’ and 1’ grid. Since the approximation
errors are compounding with the tsunami travel distance, the amplitude and phase of the short period wave (5 to 10 min) is strongly distorted at the distances of propagation of about 5000 km. One possible solution includes the higher spatial resolution of the bathymetry data. Unfortunately even 1’ grided bathymetry data used to investigate the transoceanic propagation are of very poor quality. Therefore, the application of the higher order of approximation in the numerical solution seems to be the best approach for the present time. We have tested some of the schemes: they are quite simple, work well, and can be easily introduced in the older algorithms. Another domain where physics and numerics interact strongly includes the nonlinear processes occurring when tsunami travels upsloping bathymetry. In tsunami spectra, the shorter waves are first to be transformed through the nonlinear interactions. This process leads to dissipation and breaking of the shorter period waves and leaving in the spectra the longer period waves only. Even the best numerical scheme is not able to resolve the shortest breaking wave. Therefore, the computation while the breaking occurs can be only concerned with the remaining signal of the original tsunami signal. This signal can be recovered through the filter built into the nonlinear advective terms.
THE MOMENTUM OF TSUNAMI WAVES
Harold G. Loomis
Honolulu, Hawaii USA
In the generation and propagation of tsunamis, it seemed like the momentum might be a quantity of some usefulness. In many tsunami generating situations the source mechanism might impart significant initial velocity to the water in addition to surface displacement. In the cases of pyroclastic flow and landslides from land into the water this is surely the case. The property of momentum that is especially noteworthy is that, unlike energy, the momentum of a body of water is affected only by external forces and not by internal forces associated with turbulence or laminar flow. These latter aspects of wave propagation dissipate energy and have disappeared from the distant wave motions in which principally irrotational flow remains. The impulse, Fdt, where F are external forces on the body of water, result in a change of momentum, d(Mv) of the body of water. The momentum density of a column of water of dimensions dxdy and from the bottom to the surface corresponding to particle velocities v(z) is the quantity discussed in this paper. Once the pattern of horizontal momentum vectors is imparted to the ocean, it pretty much remains there, diminished only slightly by bottom friction and refracted by bathymetry much as wave crests are refracted, until reaching a distant shore at which
time momentum is a measure of the destructive capacity of the waves. The physics of momentum for shallow-water long waves turns out to have an interestingly simple form.
VOLCANICALLY GENERATED TSUNAMIS
Honolulu, Hawaii USA
Tsunami generation from a variety of volcanic source mechanisms and from the slope failures of underwater volcanic islands, is reviewed. Expected near and far-field tsunami effects are discussed. Seismo-tectonic magma and crustal movements, pyroclastic
emissions, nuees ardentes, and other kinematic processes resulting from the complex, eruptive activities of Shield, Strombolian, Plinian, Phreatomagmatic and composite, Polygenetic Stratovolcanoes are examined for tsunamigenic potential. Paroxysmal volcanic
activity, explosion/collapse, and underwater caldera formation are evaluated for specific tsunamigenic source characteristics and expected near and far-field tsunami effects - including the generation of atmospherically-induced events. Specific examples are
provided, based on the 1883 Krakatoa, the 1490 BC Santorin, the August 19, 1975 Hawaii, and other tsunami events. The potential of tsunami generation from volcanic island slope failure is also evaluated, in terms of inferred source dimensions, event time history, and other source characteristics. The near and far field tsunami effects from the postulated under water slope failure of the island of La Palma, in the Canary Islands of the Atlantic, are examined. Based on source dimensions and evaluation of different scenarios of failure mechanisms, conclusions are drawn as to the what may be expected - in terms of tsunami height distribution and degree of inundation along Atlantic coastlines - if such slope failure is induced by an eruption of the Cumbre Vieja volcano. The same analysis is extended to another, hypothetical, underwater slope failure at Reunion Island in the Indian Ocean, which also appears to be ”unstable” and may fail if induced by movements of magma underneath the intensively active, Piton De La Fournaise volcano. Finally, the prehistoric underwater Koolau Volcano slope failure and landslide off the East coast of Oahu Island, in Hawaii, is examined as to its triggering mechanism, whether it generated a large tsunami in the Pacific, and what may have been such an event’s near and far-field effects.
THE LITUYA BAY MEGA-TSUNAMI
Hermann M. Fritz
VAW, Zurich, Switzerland
Charles L. Mader
Mader Consulting Co., Honolulu, Hawaii USA
This report is a review of our current understanding of the Lituya Bay mega-tsunami that occurred July 8, 1958 in Lituya Bay, Alaska.
An 8.3 magnitude earthquake triggered a landslide into Gilbert Inlet and created a megatsunami that caused forest destruction and erosion down to bedrock up to an altitude of 524 meters. A 1:675 scale laboratory model was built at VAW at the Swiss Federal Institute of Technology at Zurich, Switzerland. A pneumatic landslide generator was used to generate a high speed (110 meters/second) granular slide impact. The laboratory experiments indicate that the 1958 Lituya Bay 524 meter run-up on the spur ridge of Gilbert Inlet was caused by a landslide impact. The study was reported in Science of Tsunami Hazards, Vol 19, pages 3-22 (2001). The water on the ridge then resulted in the flooding of the rest of Lituya Bay up to 200 meters as described using numerical modeling in Science of Tsunami Hazards, Vol 17, pages 57-67 (1999). A PowerPoint presentation will be given and the status of our attempts to numerically model landslide impact tsunami generation will be described.
DYNAMICS OF WATER CAVITY GENERATION
Charles L. Mader
Los Alamos National Laboratory, Los Alamos, NM, USA
Science Applications International , Los Alamos, NM, USA
The hypervelocity impact (1.25 to 6 km/sec) of projectiles into water has been studied by Gault and Sonnet (“Laboratory Simulation of Pelagic Asteroid Impact,” Geological Society of America Special Paper 190 (1982) ). They observed quite different behavior of the water cavity as it expanded when the atmospheric pressure was reduced from one to a tenth atmosphere. Above about a third of an atmosphere, a jet of water formed above the expanding bubble and a stem developed below the bottom of the bubble. Similar results were observed by Kedrinskii (private communication) when the water cavity was generated by exploding bridge wires with jets and stems forming for normal atmospheric pressure and not for reduced pressures. Earlier B. G. Craig, (“Experimental Observations of Underwater Detonations Near the Water Surface”, LA-5548-MS (1974) ) observed the formation of jets and stems while the gas cavity was expanding by bubbles generated by small spherical explosives detonated near the water surface. These remarkable experimental observations have resisted all modeling attempts for over 25 years. The numerical simulations could not describe the thin water ejecta plumes formed above the cavity or the interaction with the atmosphere on the outside of the ejecta plume and the pressure inside the expanding cavity and plume. During the last decade a compressible Eulerian hydrodynamic code called SAGE has been under development by the Los Alamos National Laboratory and Science Applications International (SAIC) which has continuous adaptive mesh refinement (AMR) for following shocks and contact discontinuities with a very fine grid while using a coarse grid in smooth flow regions. This allows the code to devote the bulk of the computing resources to those areas where they are needed most. A version of the SAGE code that models explosives called NOBEL has been used to model the experimental geometries of Sonnet and of Craig described above. The experimental observations were reproduced as the atmospheric pressure was varied. When the atmospheric pressure was increased the difference between the pressure outside the ejecta plume above the water cavity and the decreasing pressure inside the water plume and cavity as it expanded resulted in the ejecta plume converging and colliding at the axis forming a jet of water proceeding above and back into the bubble cavity along the axis. The jet proceeding back thru the bubble cavity penetrates the bottom of the cavity and forms the stem observed experimentally. The complicated bubble collapse and resulting wave generation was also numerically modeled. Now that a code is available that can describe the experimentally observed features of projectile interaction with the ocean, we have a tool that can be used to evaluate impact landslide, projectile or asteroid interactions with the ocean and the resulting generation of tsunami waves.
ASTEROID GENERATED TSUNAMIS
Robert Weaver and Galen Giesler
Los Alamos National Laboratory, Los Alamos, NM USA
Science Applications International, Los Alamos, NM USA
We have performed a series of two-dimensional and three-dimensional simulations of asteroid impacts into an ocean using the SAGE code from Los Alamos National Laboratory and Science Applications International Corporation. The SAGE code is a compressible
Eulerian hydrodynamics code using continuous adaptive mesh refinement for following discontinuities with a fine grid while treating the bulk of the simulation more coarsely. We have used realistic equations of state for the atmosphere, sea water, and the oceanic crust. In two dimensions, we threw asteroid impactors at 20 km/s vertically through an exponential atmosphere into a 5 km deep ocean. The impactors were composed of mantle material (3.32 g/cc) with diameters of 250m, 500m, and 1000m, chosen to compare with the previous work of Crawford and Mader. We also performed some runs with asteroids composed of iron (7.8 g/cc). Because some of the iron asteroids produced craters that penetrated the basalt crust, we included a layer of mantle material in all simulations. A vertical impact produces a large underwater cavity with nearly vertical walls followed by a collapse starting from the bottom and subsequent vertical jetting. Tsunamis up to a kilometer in initial height were generated and followed out to 100 km from the point of
impact. In the three-dimensional run, an impactor of iron was thrown at 20 km/s at an angle of 45 degrees. Differences between this run and the vertical two-dimensional runs will be discussed.
THE NTHMP INUNDATION MAPPING PROGRAM
F. I. Gonzalez, V. Titov, H. Mofjeld, A. Venturato
Pacific Marine Environmental Laboratory (PMEL)
Center for Tsunami Inundation Mapping Efforts (TIME)
National Oceanic and Atmospheric Administration (NOAA), Seattle, Washington USA
A tsunami inundation map representing a source- and community-specific “credible worst case scenario” is a powerful planning and hazard mitigation tool. The inundation map is a scientific effort that provides fundamental guidance for the production of operational emergency management products, such as evacuation maps, that take social, political and economic factors into account. The NTHMP has conducted 19 inundation mapping efforts over a five-year period, covering 88 coastal communities with an estimated
at-risk population of more than a million. Not surprisingly, it has been found that two major technical issues dominate the development of a useful inundation map – creation of adequate bathymetric/topographic computational grids and specification of sources that correspond to credible worst case scenarios. Lessons learned as a result of past efforts will be presented, as will future plans for improving the Mapping Program and increasing the usefulness of inundation mapping products to emergency managers.
TSUNAMI HAZARD MAPS OF ALASKA COMMUNITIES
Roger Hansen, Elena Suleimani and Rod Combellick
Geophysical Institute, University of Alaska, Fairbanks, AK, USA
The Geophysical Institute of the University of Alaska Fairbanks and the Alaska Division of Geological and Geophysical Surveys participate in the National Tsunami Hazard Mitigation Program by evaluating and mapping potential inundation of selected coastal
communities in Alaska. The communities are selected in coordination with the Alaska Division of Emergency Services on the basis of location, infrastructure, availability of bathymetric and topographic data, and willingness for a community to use the results for hazard mitigation. We work in cooperation with the NOAA/PMEL Center for Tsunami Inundation Mapping Efforts, which assists in developing bathymetric and topographic data grids for the area of interest. Three communities in the vicinity of Kodiak were the first
for which we produced inundation maps. The work is under way for Homer, Seldovia, and possibly other communities along Kachemak Bay. We use numerical modeling as a primary research tool to study tsunami waves generated by earthquake sources. We consider several hypothetical tsunami scenarios with a potential to generate tsunami waves that can affect the coastal communities. The nonlinear shallow water wave equations are solved with a finite-difference method. We use embedded grids that increase in resolution from the source area to the target community. State and local emergency planners will use results of the numerical modeling combined with historical observations to develop evacuation plans and to educate the public for reducing risk from
REVIEW OF THE 1994 SKAGWAY, ALASKA TSUNAMI
AND FUTURE PLANS
Dennis Nottingham, P.E.
Peratrovich, Nottingham and Drage, Anchorage, AK USA
On November 3, 1994 a 30-foot amplitude submarine landslide-created tsunami with a resonate wave train lasting about 30 minutes struck the Skagway, Alaska water front causing extensive damage and loss of one life. Numerous scientists and engineerings have studied the 1994 tsunami and at a meeting on the subject in Seattle, Washington, on October 30-31, 2001, have generally concluded
that large down inlet submarine landslide(s) created the tsunami. A general plan under the National Tsunami Hazard Mitigation Program was developed to start a study which wil lead to mitigation measures at Skagway and possible adaptability to other parts of the world with similar problems. The paper briefly overviews the events preceeding the tsunami, reviews findings following the event and outlines plans relating to similar future expected tsunamis.
LOCALLY GENERATED TSUNAMIS IN HAWAII:
A LOW COST, REAL TIME WARNING SYSTEM
WITH WORLD-WIDE APPLICATIONS
Daniel A. Walker and Robert K. Cessaro
Honolulu, Hawaii, USA
Cellular runup detectors have been installed along those coastlines of the island of Hawaii which have been frequently inundated by locally generated tsunamis. These devices provide for near instantaneous (i.e., less than 1 minute) warnings of locally generated tsunamis. Principal components of the detectors are cellular transceivers and water sensors. Because of the extensive use of these devices throughout the security industry, they are extremely reliable and cost effective. Such instruments could be useful in the measurement of tsunami runups in coastal areas of the world with cellular coverage. As satellite versions of the cellular transceivers become available, eventual coverage of all coastal areas threatened by tsunamis may be possible.
TSUNAMI HAZARD MITIGATION
AND THE NOAA NATIONAL WATER LEVEL
James R. Hubbard and Scott A. Duncan
NOAA, Silver Spring, Maryland USA
With the renewed interest in regional Tsunami Warning Systems and the potential tsunami threats throughout the Caribbean and West coast of the United States, the National Ocean Service (NOS), National Water Level Observation Network (NWLON) consisting of 150 primary stations, located for the most part in populated coastal regions, is well situated to play a major role in the National Hazard Mitigation effort. In addition, information regarding local mean sea level trends and GPS derived geodetic datum relationships at numerous low lying coastal locations is readily available for tsunami hazard assessment applications. Tsunami inundation maps and modeling are just two of the more important products which may be derived from NWLON data. In addition to the twelve water level gauges that are an integral part of the Alaskan and Pacific Tsunami Warning System, NOS has a significant number of gauges with real-time satellite telemetry capabilities located along the Pacific Northwest coastline, the Gulf of Mexico and the Caribbean. These gauges, in concert with near shore buoy systems, have the potential for increasing the effectiveness of the existing tsunami warning system. The recent expansion of the Caribbean Sea Level Gauge Network through NOS regional partnerships with Central American and Caribbean countries has opened an opportunity for a basin-wide tsunami warning network in a region which is ill prepared for a major tsunami event.
EARLY DETECTION AND REAL-TIME REPORTING
OF DEEP OCEAN TSUNAMIS
E. N. Bernard, F. I. Gonzalez, C. Meinig, and H. Milburn
National Oceanic and Atmospheric Administration (NOAA)
Pacific Marine Environmental Laboratory (PMEL)
Seattle, Washington USA
The National Oceanic and Atmospheric Administrations (NOAA) Deep-ocean Assessment and Reporting of Tsunamis (DART) Project is an effort of the U.S. National Tsunami Hazard Mitigation Program (NTHMP) to develop an early tsunami detection and real-time reporting capability. Although seismic networks and coastal tide gauges are indispensable for assessing the hazard during an actual event, an improvement in the speed and accuracy of real-time forecasts of tsunami inundation for specific sites requires direct tsunami measurement between the source and a threatened community. Currently, only a network of real-time reporting, deep-ocean bottom pressure (BPR) stations can provide this capability. Numerous NOAA deployments of everimproving prototype systems have culminated in the current operating network of 6 DART stations in the North and Equatorial Pacific. DART data can be viewed online
at http://tsunami.pmel.noaa.gov/dart/qc/WaveWatcher. Network coverage is presently limited to known tsunamigenic zones that threaten U.S. coastal communities. Because tsunamis can be highly directional in their energy projection, DART data must be
supplemented with numerical model output to accurately forecast tsunami wave heights. DART buoys have been tested by nontsunamigenic earthquakes and perform as designed. An example of one of these cases will be presented to illustrate the performance of DART. Future plans for maintaining and expanding the existing network of DART buoys will also be discussed.
THE STOREGGA SUBMARINE LANDSLIDES -
WHAT DO WE OBSERVE AND WHAT DO WE MODEL?
University of Tromso, Norway
Submarine landslides and gas hydrates are a global phenomenon. Factors such as earthquake loading have proven to be important, and gas hydrates in the pore space of marine sediments have been observed to be of variable importance for submarine slope failures. A detailed geological and geophysical evaluation of the Storegga gas-hydrate system off the Mid-Norwegian margin reveals that leakage from deep-seated Tertiary dome structures might provide a supply of hydrocarbons to the hydrate formation, while fluids from polygonal fault systems may contribute enough water for the hydrate formation (Berndt et al., 2002). The polygonal fault systems are widespread on the Mid-Norwegian margin. Gas hydrates on the other hand are much less frequent. The fact that we observe fluid flow-related pipes rising from the top of the polygonal fault systems, i.e. a boundary between a gas hydrate bearing lithofacies and the fluid expelling lithofacies, implies that a supply of pore fluids exists into the gas hydrate bearing sediments. There is no geophysical evidence for fluid flow from deeper stratigraphic levels. It suggests that the pore fluids that are expelled from polygonal fault systems may be the main source for the formation of gas hydrates at shallower stratigraphic levels. The Tertiary dome structures with proven hydrocarbon potential in the study area might contribute gases to the formation of gas hydrates. This poses an important question for slope stability. What role do deep-seated gas reservoirs and their leakage systems play in the shallow-seated gas hydrate formation of continental margins. Does major submarine sliding on passive margins coincide with
regions of deep-seated gas reservoirs? At Storegga, bottom simulating reflectors (BSR’s) indicate the existence of free gas at the base of the gas hydrate stability zone (GHSZ). Here, the gas concentrates beneath gas-hydrated sediments. Gas hydrates occur in restricted areas within the slide and close to the northern sidewall but have not been observed in the 300 km long headwall. As the occurrence of gas hydrates potentially poses a risk to deep-water hydrocarbon exploration, industry and academia in Norway have developed joined research programs. Understanding the distribution and the formation controls of present-day gas hydrates is crucial for an assessment of the evolution of slope stability in the Storegga region. The total area, where geophysical evidence for gas hydrates exists is about 4000 km2 (Buenz et al., 2002). The area can be much larger because gas hydrates may also occur in areas without a BSR. Within the slide area, the formation of gas hydrates must have taken place after the slide event, i.e. 8200 yrs B.P. (Haflidason et al., 2001). It underlines the advection of fluids from deeper sources. Some of the important questions, which will be addressed in this presentation are: Did a major gas hydrate dissociation process start in Holocene times and if so, did this trigger submarine slides off the Mid-Norwegian margin or did earthquake loading during postglacial Scandinavian uplift or a combination of both trigger the slides? Our observations and models, age dating of slides in the ocean and Tsunami deposits on land tell us a story - but are we really reading it well enough to understand the complex coupling between gas hydrates, submarine slides and tsunamis.
MASS FLOW IN MARINE SEDIMENT:
PHYSICAL CAUSES AND POSSIBLE EXAMPLES FROM THE GEOLOGICAL RECORD
Michael D. Max
MDS Research, Washington, DC, USA
Dissociation of gas hydrate in response to the lowering of pressure and/or the raising of temperature can lead to sediment mass flows in two principal manners. First, relatively rapid dissociation of hydrate in a volume of sediment can lead to formation of a thixotropic mass that will flow. This is most likely to happen in near surface sediments where melting of even small amounts of dispersed hydrate can produce enough dispersed gas throughout the body to substantially reduce shear strength within the mass as a whole. Second, gas can form in concentrated zones and lower the shear strength of those zones so that they can act as slip surfaces. This is likely at or near the base of the gas hydrate stability zone GHSZ where gas is known to pond. Where there is a strong contrast in shear strengths in sediment masses, for instance where higher strength hydrate-rich sediments directly overly sediments containing overpressured free gas, and especially where these incipient slip surfaces are nearly parallel to the seafloor, large scale slip and mass flow mass generation is likely. Evidence for mass flows should be found in the geological record because gas hydrate should be a geologically persistant phenomenon. A possible example of the one-time presence of abundant methane producing sediments associated with mobilized sediments occurs in Cambro-Ordovician continental slope Caledonian sediments in eastern Ireland. Mobilized beds occur on a wide scale. In a number of small bodies, injection of fluidized sediment into bounding bedded sediments occurs on both the upper and lower margins and indicates that the chaotic beds were fluidized in place. Although smaller fluidized beds may represent masses that did not break through to the seafloor, spatially related large olistostromes are probably true mass flow deposits that were generated in an environment where gas was widely produced, captured in hydrate, and released in such a manner as to generate sediment fluidization and mass flow deposits. Understanding of the relationship between climate change, which is reflected by ocean warming or sea level falls that can cause hydrate dissociation, and the geological record
of mass flow generation and distribution, should allow for more accurate prediction about the likelihood of modern submarine landslip that can result in a tsunami.
GLOBAL DISTRIBUTION AND SIGNIFICANCE
OF NATURAL GAS HYDRATE
Alexei V. Milkov
Texas A & M University, College Station, TX, USA
Gas hydrate is a mineral composed of water and gas that occurs worldwide on shorein polar regions and offshore at water depths greater than about 200-600 m, depending on seafloor temperature and gas composition. Hydrate- bound gases are mainly methane
but also include ethane through butanes hydrocarbons and non-hydrocarbon gases such as carbon dioxide and hydrogen sulfide. Gas hydrate samples are recovered in 21 offshore areas and one onshore area. In addition, geophysical (e.g. BSR and well-log anomalies), geochemical (e.g. chlorinity anomalies), and geological (e.g. slumps) evidence of gas hydrate are reported from 54 offshore and six onshore localities. Offshore gas hydrate occurs mainly on continental slopes and rises, and in inland seas. Three types of gas hydrate accumulations are distinguished based on the mode of fluid migration and gas hydrate concentration within the gas hydrate stability zone (GHSZ). Structural accumulations occur where fault systems, mud volcanoes and other geologic structures favor rapid fluid transport from depth into the GHSZ. In stratigraphic accumulations, gas hydrate crystallizes in relatively permeable strata from bacterial gas generated in situ or which is slowly supplied from great depth. These are end-members, and combination accumulations controlled both by structures and stratigraphy may occur. The study of natural gas hydrate is important because it may serve as a future energy resource, be a geohazard, and be an agent of global change. The global volume of hydrate-bound is uncertain. Offshore gas hydrate may contain from 0.2 x 1015 to 7, 600 x 1015 m3 of gas at STP. Specific gas hydrate accumulations are estimated to contain from 108 to 1013 m3 of gas. The economic potential of well-studied gas hydrate accumulations and provinces varies widely. Gas hydrate recovery may be economic only where concentrated gas hydrate deposits occur in relatively permeable sediments (mainly structural accumulations). Natural gas hydrate may be transitionally unstable in shallow (first meters below seafloor) sediments in areas where seafloor temperature or pressure changes over time. Repetitive gas hydrate formation and decomposition may cause sediment deformation, slumps, gas blowout craters, and increase the rate of gas and oil venting to the water column. Such events may impact sub-sea petroleum exploitation. It has been hypothesized that the release of hydrate-bound greenhouse gases and their subsequent oxidation to carbon dioxide may lead to appreciable climate change at global scale and may impact the global carbon cycle. However, the models relating gas hydrate and global change are yet poorly constrained and remain largely speculative.
IS THERE A CONNECTION BETWEEN SUBDUCTION ZONE EARTHQUAKES, SUBMARINE LANDSLIDES, AND THE
DESTABILIZATION OF GAS HYDRATES ON THE CENTRAL OREGON MARGIN?
Chris Goldfinger, Joel F. Johnson
Oregon State University, Corvallis, OR, USA
C. Hans Nelson
Texas A & M University, College Station, TX, USA
Current gas hydrate research is focused on the physical, chemical, and biological aspects of gas hydrate systems, the dynamics of gas hydrate stabilization, and perhaps more importantly, factors that influence their destabilization. Although, there are several sedimentological and geochemical mechanisms believed to influence the stability of gas hydrate systems on continental margins worldwide, the influence of the great earthquake cycle on the destabilization of gas hydrate systems in active margin settings is unknown. Current examininaton of the interaction between recurrent subduction zone earthquakes, submarine landsides, and gas hydrates on the Cascadia accretionary prism suggests linkages independent of climatologic factors. Hydrate Ridge, a well-studied gas hydrate province offshore central Oregon, is an ideal location to determine this interaction because it lies within a portion of the margin that has a well constrained Holocene paleoseismic record of great subduction zone earthquakes and contains a record of Holocene submarine landslide deposits derived from the ridge. Core data from the basin west of this ridge contain a record of cyclic turbidites similar to that recorded on the Cascadia abyssal plain, suggesting that denudation of the ridge, and thus destabilization of the hydrate system by submarine landsliding may be controlled by the subduction zone earthquake cycle (Johnson et al., 2001; 2002). The great earthquake cycle in Cascadia over the past 10,000 years is now known via margin wide correlation of turbidite events and testing of alternative turbidite triggers (Goldfinger et al., 2001; 2002). To evaluate the link between earthquakes and landslides on Hydrate Ridge, we will collect new piston cores from Hydrate Ridge Basin-West (HRB-W) and use AMS radiocarbon techniques to establish Holocene datums throughout the cores. These data will help to identify the occurrence, distribution, and recurrence interval for Holocene submarine landslides derived from Hydrate Ridge. Comparison of this record with the Holocene subduction zone earthquake record, from abyssal plain turbidites, will then help determine the influence of subduction zone earthquakes on submarine landslides on Hydrate Ridge. We postulate that earthquake-triggered submarine landslides are a dominant
mechanism that could have a short-term recurrent effect on the destabilization of gas hydrate in an active margin setting. Alternatively, denudation of the ridge and subsequent destabilization of gas hydrate may be controlled by both earthquakes and additional processes such as rapid fluid expulsion or degassing of the ridge driven by local deformation. Regardless of the landslide
trigger, establishing the frequency of the submarine landslides derived from the ridge is the first step in identifying the influence of margin-wide tectonic processes on the stability of slopes and gas hydrates on an active margin. On a larger scale, mega-landslides identified on the Oregon margin and other localities may also serve as a vehicle for injection of massive amounts of methane into the water column, and thus the atmosphere, on a less frequent basis. The great Oregon slides most likely have little to do with climate change, as they are deep seated, and most likely related to tectonic factors such as seamount subduction. They do however represent catastrophic tectonically driven events that destabilized large tracts of seafloor, and most likely introduced large volumes of both hydrate and free gas into the water column.
GAS HYDRATES ON THE CHATHAM RISE, NEW ZEALAND,
AND THEIR POTENTIAL OF TSUNAMI GENERATION
Charles E. Helsley
University of Hawaii, Honolulu, HI USA
During a set of research cruises undertaken to identify deep-water fishery resources within the EEZ of New Zealand in 1994, a series of large arcuate structures were observed to the east and southeast of the Chatham Islands in water depths in excess of 1000 meters.
At the greatest depths these features were simply smooth arcs, some with a radius of several kilometers and were considered to be entrenched meanders of some past river system although the individual arc segments were never seen to connect with adjacent
arcs. The area inside the arc was about 100 meters deeper than the adjacent area outside the arc and the arcuate trench itself was about 200 meters deep in a few cases. As the survey proceeded to shaller depth, the features became more completely circular with the largest fully circular feature being about 1500 meters across. Again, the interior was about 100 meters deeper than the surrounding terrain. As the features became smaller, the trench like,aspects of the arcuate border became less pronounced. At still shallower depth the craters became more numerous and took on the typical morphology of pockmarks at depths less than 700 meters. This pockmocked area also had raised cones, some with summit craters. At first these were interpreted as volcanic features but in hindsight, they were probably small mud volcanoes. All evidence of pockmarks and mud volcanoes ceased at depths of about 500 meters. Upon reflection, all of the observed features can best be explained as a gas hydrate field developed in 500 to 1000 meters of water in an area where the biological productivity of the sea is known to be large, i.e. there is a large amount of biological carbon available to be incorporated into the sediment. But gas hydrates are common throughout the world so why is this observation of significance. These are some of the largest crater like gas hydrate features ever observed. The question then becomes, how did such large flat-bottomed crater-like features form? My preferred hypothesis is that they are the gas hydrate equivalent of volcanic caldera collapse features. Thus would imply that the surficial gas hydrate sealed the gas reservoir beneath the hydrate until sufficient overpressure built up to require the fracturing of the gas hydrate seal along a ring fracture exactly equivalent to the ring fracture associated with volcanic caldera collapse features. If such a mechanism is shown to be viable, then it could imply that when the gas reservoir vented or released to the overlying sea, the central portion of the gas hydrate cemented lid could no longer be supported and that therefore it could suddenly collapse due to the loss of reservoir pressure that had been supporting it. Conceivably, this sudden drop of the seafloor could result in the formation of a tsunami. Initial calculations suggest that a run-up of 5 meters could
occur on the nearby islands.
MARINE METHANE HYDRATES
Stephen M. Masutani
University of Hawaii, Honolulu, HI USA
Hydrates are crystalline solids comprising water molecules linked by hydrogen bonds in a tight polyhedral cage structure. Guest molecules, which can include various hydrocarbons found in natural gas mixtures, reside in the interstices of this lattice. According to recent assessments, natural gas (methane) hydrates represent an enormous untapped hydrocarbon resource. The primary known repositories of methane hydrates are arctic permafrost zones and undersea basins on the continental margins. Sediment layers in deep ocean basins also may contain large deposits of hydrates. Estimates of the total volume of hydrocarbons locked in hydrate deposits worldwide range widely from about 105 trillion standard cubic feet (TCF) to 2.7x108 TCF (i.e., 2.8x1015 to 7.6x1018 cubic meters). Even at the lower end of this range, this resource could potentially satisfy the energy needs of the world for centuries, provided that practicable and environmentally responsible recovery techniques can be devised. At the same time, it must be recognized that oxidation of the methane currently sequestered in these hydrates for energy could produce in excess of 103 gigatonnes of carbon (in the form of CO2 ). This associated greenhouse gas burden would profoundly impact global climate if sequestration or other emissions control technologies are not pursued. While offering tremendous opportunities as a future primary energy resource, marine hydrate deposits also represent an immediate and formidable nuisance to offshore oil and gas operations. This problem has become more critical as these commercial activities move into increasingly deeper waters. From a defense perspective, there is a need to characterize the geoacoustic properties of hydrate sediments and to assess their potential as an in situ offshore energy source, since this information is relevant to Naval operations. The promise and peril associated with methane hydrates has led to recent national research and development programs in Japan, India, and the U.S. Research funding for marine methane hydrates has also increased in a number of European Union and Asian nations, with significant growth anticipated in the near term. This paper provides an overview of research challenges and priorities for programs focusing on hydrate energy production, pipeline blockage and seafloor stability issues confronting offshore oil and gas operations, and defense interests. The collaborative research and development program on marine methane hydrates between the University of Hawaii and the Naval Research Laboratory will also be summarized.
NUMERICAL SIMULATION WITH TWO-LAYER MODEL
FOR A LANDSLIDE-INDUCED TSUNAMI
Yoshinori Shigihara and Fumihiko Imamura
Tohoku University, Japan
Tsunamis generated by landslide are known in Japan but studied less than tsunamis generated by fault motion because of the low frequency of events. However, historical documents report that landslide-induced historical tsunamis have killed many people and
caused serious damage; about 1,500 casualties in the 1741 Oshima-oshima and 15,000 in the 1792 Ariake event. Once this type of a tsunami is generated, a local wave amplification and energy concentration and/or radiation can be significant, causing much damage. It is thought that the landslide-induced tsunami is triggered by strong ground motion, failure of slope, volcanic activity or increasing ground water. Gas-hydrates on the continental margin have been recognized to be another source that can generate a tsunami. Melting gas-hydrates and its explosion in a wide area can disturb the water surface. A giant tsunami by the second Storegga slide in the Norwegian continental slope 7000 years BP is reported to be caused by gas-hydrate excess pore water pressure (Harbitz, 1991). Since a small number of reports on these phenomena and less study on the mechanism of the generation and propagation are available, we need historical and fundamental research on landslide induced tsunamis. In this study, some results of the analysis of historical tsunamis and hydraulic experiments on landslide-induced tsunami are presented. Hirakawa et al. (2000) newly discovered a tsunami deposit on the shore of Tokachi in the southern part of Hokkaido, indicating tsunami runup heights of 4 -12 m above the sea level. Aging analysis suggest the event in the 17th century when a giant tsunami in 1771 occurred and affected the Sanriku coast in the Northeastern Japan. No information in Hokkaido is available except for a new finding by Hirakawa et al. (2000). We carried out the numerical simulation for the 1771 tsunami assuming a fault offshore of Sanriku and compared it with the measured data (Aida, 1977, Hirakawa, et. al. 2000). Good agreement between computed and measured runup were obtained in Sanriku, whereas a smaller computed than measured wave was obtained in Hokkaido, suggesting a new tsunami source is needed such as a marine landslide or explosion of gas-hydrate offshore Tokachi where a large amount of marine sediment is reported and gas-hydrate is buried under the sea bottom. We developed a numerical model with two layers for landslide-induced tsunami and
carried out hydraulic experiments in the open channel with a slope. We clarified the processes of tsunami generation caused by landslide. First, the effect of the interactive force occurring between the debris flow layer and the tsunami is significant. Second,
the continuous flowing of the debris in the water makes the wave period of the tsunami short. Numerical simulations, using the model of two-layer with shear stress models on the bottom and interface,was performed and compared with the results of experiment. Simulated debris flow shows good agreement with the measured results. The model required a Manning coefficient of 0.01 for the smooth slope, 0.015 for the rough slope, and a horizontal viscosity of 0.01 for landsides. However, modeling of interactive forces will require further analysis to determine its coefficient and the acting duration.
PROBABILITY DISTRIBUTIONS AND SENSITIVITY ANALYSIS OF GAS HYDRATE INDUCED LANDSLIDES AND TSUNAMIS
Applied Fluids Engineering, Inc., Long Beach, CA, USA
A Monte-Carlo technique is developed to evaluate the landslide and tsunami hazards posed by gas hydrate induced sediment instability. The technique produces thousands of random tests of sediment stability that cover the full parameter space anticipated for most continental slopes. Probability distributions of landslide dimensions and tsunami features such as amplitude and wavelength are predicted by the model. Sensitivity analyses of the input parameters yield an indication of those conditions whereby gas hydrate induced instability can be expected to dominate the morphology of continental slopes. This information can provide some indication of the risks associated with coastal development and with offshore industrial activity.
TSUNAMI HAZARD IN CANADA
John J. Clague
Simon Fraser University, Vancouver, B.C, Canada
National Hazards and River Services, Waikato, New Zealand
Baird and Associates, Ottawa, Ontario, Canada
Canada has experienced many tsunamis, but few have caused damage. The only disaster included in Jones’(2000) exhaustive compendium of Canadian disasters is the tsunami triggered by the Grand Banks earthquake of November 18, 1929, which claimed twentyeight lives, all except for one on Newfoundlands Burin Peninsula. Interestingly, Jones excludes from this catalogue the tsunami caused by the Halifax Harbour explosion of December 6, 1917, even though perhaps as many as two hundred people drowned during this catastrophe (Halifax Herald, December 7, 1917). If this statistic is correct, the Halifax Harbour tsunami is, by far, the most devastating tsunami in Canadian history. Canada experienced a third, large, destructive tsunami in 1964, on the west coast of Vancouver Island. Several other, smaller tsunamis have occurred on both the Atlantic and Pacific coasts, and it is possible that much larger tsunamis than those in 1917, 1929 and 1964 could strike Canada in the future. A tsunami larger than any of the historical
period could be triggered by a great earthquake at the Cascadia subduction zone, which underlies the sea floor of the eastern North Pacific Ocean, or by a large landslide from the submerged flank of one of the volcanoes on the Hawaiian Islands. A large landslide at the Atlantic continental margin could trigger a tsunami that would devastate Canada’s east coast. These very large tsunamis are rare events, but if one were to occur, it would have a catastrophic effect on coastal infrastructure and might kill and injure large numbers of people. This paper summarizes what is known about tsunami hazard and risk in Canada.
AN ASSESSMENT OF TSUNAMI HAZARD AND RISK IN THE INDONESIAN REGION
Centre for Earthquake Research in Australia, Brisbane, Australia
The natural hazard of tsunami has, for too long, been underrated as a potential cause of major disasters. However, several devastating tsunamis in and around the Pacific Ocean Basin over the last decade - all claiming significant loss of life, major property and environmental damage and severe socio-economic losses - have heightened the awareness to this natural hazard. As a consequence, the approach to mitigation strategies and measures in disaster management is today being addressed in several countries around the world. Indonesia is a region with a considerable record of tsunami occurrence dating back hundreds of years. In a large number of these instances, devastating effects have been reported. However, reliable and complete data bases, scientific analyses and integration of relevant information into disaster planning has been lacking. This paper reviews the status of available tsunami information (with particular reference to Indonesian and Japanese studies and the recently published Historical Tsunami Data Base for the US Pacific Coast HTDB/US), presents a comprehensive data base (the central element to any tsunami hazard assessment) and reports on quantitative hazard and risk assessments relative to the site-specific location of Halmahera Island. Comments are made on the integration of these assessments into engineering design and risk management for a major commercial
TSUNAMIGENIC EFFICIENCY OF MAGMA-INDUCED KINEMATIC CRUSTAL CHANGES, ASSOCIATED WITH “SILENT” VOLCANIC EARTHQUAKES AND MASSIVE SLOPE FAILURES OF OCEANIC ISLAND STRATOVOLCANOES
Computer modeling of tsunami generation triggered by ”silent” volcanic earthquakes and postulated massive slope failures of island stratovolcanoes - such as Kilauea in Hawaii, and Cumbre Vieja in La Palma, Canary Islands - has been based on incorrect assumptions of initial input parameters of source mechanism. Size and dimensions of the disturbing source events, their time history, duration, speed, volume of water displacement and effective energy transfer to the body of water, have been overestimated. Source ground motion and coupling characteristics have been correlated incorrectly to maximum elevation of water surface of the leading waves. Erroneous input functions have led to inaccurate output estimates. Slope failures of oceanic island stratovolcanoes occur, but in phases, over a period of time, and not necessarily as single, sudden, large-scale events. Although such failures have
generated destructive local tsunamis in the past - as demonstrated by the 1868 and 1975 events along Kilauea’s southern flank, in Hawaii - greater source dimensions and longer wave periods are required for the generation of tsunamis with significant, far field effects. Recent, high-resolution bathymetric and sidescan surveys of offshore areas of Kilauea do not reveal any evidence of recent or prehistoric massive slope failures. The caldera collapses and large slope failures associated with the volcanic eruptions of Krakatau in 1883 and of Santorin in 1490 B.C., generated catastrophic local tsunamis, but no waves of any significance at distant locations. The threat of mega-tsunamis from postulated massive slope failures of Cumbre Vieja or Kilauea has been greatly overstated.
VALIDATION OF TOPICS AND CASE STUDIES WITH GEOWAVE
Applied Fluids Engineering, Inc., Long Beach, CA, USA
S. T. Grilli
University of Rhode Island, Narragansett, RI, USA
J. T. Kirby
University of Delaware, Newark, Delaware, USA
G. J. Fryer
University of Hawaii, Honolulu, HI, USA
British Geological Survey, Kewsworth, Nottingham, UK
There has been a proliferation of landslide tsunami generation and propagation models in recent time, spurred largely by the 1998 Papua New Guinea event. However, few of these models or techniques have been carefully validated. Moreover, few of these models have proven capable of integrating the best available geological data and interpretations into convincing case studies. The Tsunami Open and Progressive Initial Conditions System (TOPICS) rapidly provides approximate landslide tsunami sources for propagation models. We present 3D laboratory experiments and 3D Boundary Element Method simulations that validate the tsunami sources given by TOPICS, to within the anticipated degree of accuracy. Geowave is a combination of parts of TOPICS with parts of the fully nonlinear Boussinesq model FUNWAVE, subject of extensive testing and validation over the course of the last decade. Geowave is currently a tsunami community model made available to all tsunami researchers. We validate Geowave with case studies of the 1946 Unimak, Alaska, the 1994 Skagway, Alaska, and the 1998 Papua New Guinea events. The benefits of Boussinesq wave propagation over traditional nonlinear depth-averaged models is very apparent for these relatively steep and nonlinear waves. These benefits accrue with little additional computational expense. We encourage other tsunami researchers to acquire and test Geowave alongside their existing models.
PROBABILISTIC TSUNAMI HAZARD
AND RISK ASSESSMENT
Dr. Robert T. Sewell
R. T. Sewell Associates, Consulting
Louisville, Colorado, USA
Well-established approaches for probabilistic seismic hazard analysis (PSHA) are adapted to the problem of probabilistic quantification of tsunami hazards. Whereas the PSHA approaches result in estimates of annual exceedance frequencies (or, equivalently, return periods) of earthquake ground motions, their adaptation to probabilistic tsunami hazard analysis (PTHA) result in estimates of annual exceedance frequencies (or return periods) of tsunami wave heights, run-up elevations, or other wave characteristics of scientific and engineering interest. The methodological adaptation is sufficiently general to address all tsunamigenic sources, not just earthquakes. As for PSHA, the format of PTHA results is particularly compatible for use in the development of appropriate zoning criteria and other input for codified siting and design of buildings, and has applications (among others) in design of harbor facilities, offshore structures, and other important facilities often located on coasts (e.g., nuclear power plants, natural gas plants, etc.). Logic-tree methodology, which is useful to formally account for multiple expert opinions in evaluating confidence bounds on probabilistic hazard results - and which has been successfully applied on extensive basis for conveying the scientific uncertainty in PSHA results for important projects - is also presented and adapted to probabilistic tsunami
hazard evaluation. The framework for incorporating PTHA results (and related scientific uncertainties) in to overall risk assessment/management and related safety decision-making is also discussed.
ELEVATED STRANDLINES - NOT TSUNAMI DEPOSITS ON LANAI, HAWAII
Barbara H. Keating & Charles E. Helsley
University of Hawaii, Honolulu, Hawaii, USA
Outcrops of shell hash and coral at elevations of up to 190 m have been previously described from the south coast of the island of Lanai, Hawaii (Stearns, 1978) and have been cited as evidence for deposits left by giant tsunami waves (Moore and Moore, 1984
and 1988). Recent detailed mapping of these units at elevations between 100 and 200 m by the authors has identified ledge-like deposits and wave-cut notches containing marine fossils at specific elevations that are best explained by elevated strandlines rather than scattered deposits swept up by giant wave activity. Fossil assemblages, in-situ coral heads, and internal stratigraphy within well exposed portions of these units, are best explained by normal sedimentologic processes in reef, lagoon and beach environment. Moreover, the consistent elevation of these units along several kms of the south flank of Lanai leads to the unescapable conclusion that the units between 98 and 200 m are uplifted marine strandline deposits rather that deposits of a chaotic process such as transport by giant waves. We interpret these deposits as elevated units originally formed at sea level. Lanai is currently at the same distance from the active Hawaiian hotspot as is the arch to the east and west of the hotspots. Thus we interpret the strandlines to be ancient sea level lines that were formed during glacial and interglacial periods during the last 400,000 years. The limited age dates available are consistent with this interpretation and the maximum elevation is consistent with the uplift of the seafloor along the arch in the deep oceans surrounding the island of Hawaii.
GEOMORPHIC EVIDENCE AND C14 DATING RESULTS FOR TSUNAMIS IN CYPRUS
University of Bamberg, Bamberg, Germany
University of Essen, Essen, Germany
For the Mediterranean area, almost 100 tsunamis were recorded in historical sources from Antiquity until today, and through measurements in the 20th century. Recordings often describe the consequences for human lives and buildings in coastal areas. However, evidence for the geomorphic effects of tsunamis has not been collected yet in this region. Literature mentions only four local hints for small correlated deposits. Recently extensively dispersed tsunami sediments were found in southwestern and
southeastern Cyprus. This report describes these sediments, their morphologic characteristics, and possibilities of relative and absolute dating. Tsunami run-up may destroy soil and vegetation. Tsunamis may further move extremely large volumes of coarse
clastic material including individual boulders weighing more than 20 t. Trottoirs, supralittoral cliffs, and tafoni may also be destroyed. Deepwater foraminifers deposited on land also provide evidence for Tsunami action. Field collected evidence proves tsunami action
for over 60 km of coastline and about 100-150 m inland. Coastal areas up to 15 m a.s.l., sometimes up to a maximum height of 30-50 m a.s.l., have been influenced by tsunami action on Cyprus island. Clues for relative age determination are provided by soil and vegetation, tafoning, karstification on displaced boulders, as well as by post-tsunami cliff and beach rock development. Field evidence suggests that tsunamis occurred during the last few centuries. This time estimate was also supported by the absolute 14C dating of vermetids and calcareous algae crusts on displaced boulders, and by the dating of relocated wood and charcoal. Overall, strong tsunami action can be assumed for the time 1700 to 1750 AD.
EVIDENCES OF TSUNAMIS ON CURACAO, BONAIR AND ARUBA
University of Essen, Essen, Germany
Tsunamis are one of the major natural hazards in the Caribbean - the historical record lists 88 tsunamis, from local events to teletsunamis, in the time period from 1489 till 1998. This study focuses on the spatial distribution and geomorphologic evidences related to coarse littoral sediment deposition by tsunami events of Holocene age in the Southern Caribbean. On a worldwide scale, these debris deposits represent the most extensive and impressive records of Holocene paleo-tsunamis so far studied. Hitherto, the Leeward Lesser Antilles, consisting of the islands of Aruba, Curarao and Bonaire, were unknown for tsunamis affecting their coastlines. The islands are located north of the Venezuelan coast in the Caribbean Sea within the tectonically active Caribbean - South American plate boundary zone and the West Indies Island Arc, an active subduction zone with frequent occurrence of earthquakes, volcanic eruptions and slides, situated to the northeast. The nature of the tsunami deposits of the ABC-islands include extensive rubble ridges, rampart formations and boulder assemblages situated from sea-level to a height of + 12 a.s.l. and reaching up to 200 m inland. The largest boulder is 5.9 x 5.3 x 3.6 m in size with a volume of 112 cubic meters and a weight of 281 tons. Situmetric measurements and hydrodynamic calculations have been applied to distinguish between hurricane- or tsunami-induced depositional processes; the results relating the sediments unambiguously to tsunami events. The spatial distribution is presented in detailed maps of all debris deposits. Predominantly, the debris deposits have been accumulated on the northeastern sides of the islands, suggesting a tsunami approach out of a 0 to 90 degree sector. In addition, the extent and amount of tsunami debris diminishes from east to west with the highest energy impact on Bonaire in the east and a considerable lower impact on Aruba, the most westerly island. Forty-five samples of coral and mollusks species were dated with the Radiocarbon method indicating a young Holocene age of the deposits. The ages cluster clearly around three time periods with the oldest event dating back to 3500 BP, the second impact taking place around 1500 BP and a third occurring at 400-500 BP, presumably just before the occupation by the Dutch in 1634.
HYDRODYNAMICS OF MEGACLAST EMPLACEMENT AND TRANSPORT ON A SHORE PLATFORM, OAHU, HAWAII
University of Stockholm, Sweden
Keith A. W. Crook and E. Anne Felton
University of Hawaii, Honolulu, Hawaii USA
Limestone megaclasts, emplaced onto a shore platform by large waves, have been documented on the North Shore of Oahu, Hawaii since 1905. Emplacement and movements of the largest, 96 ton, megaclast have been dated using aerial photographs. Theoretical
hydrodynamic forces at the submerged cliff that forms the seaward edge of the shore platform are computed using design wave characteristics based on linear wave theory and laboratory results considering the local wave climate and near-shore bottom topography. Those forces are compared with the minimum forces required for the emplacement and transport of the largest megaclast currently on the terrace. The analysis shows that tsunami as well as large swell waves are capable of emplacing and transporting the largest megaclast. The rough estimates of maximum pressures exerted on the shoreline cliff by breaking and broken waves lie in the range of 1.4 MPa and 230 kPa, respectively. Both exceed the minimum pressure required for the emplacement of the largest megaclast (59 kPa) assuming that the only two forces that the wave force has to overcome are gravity and friction forces, i.e. the megaclast is essentially detached from the parent limestone body. Maximum dynamic shock pressures are most likely produced by relatively short and steep deep-water waves. Longer waves, however, produce higher breakers at the submerged cliff and have therefore a capability of transporting megaclasts farther toward the rear of the shore platform. Minimum height and celerity of a bore resulting from a tsunami capable of emplacing the largest megaclast onto the terrace were estimated to be H=5.1 m above still water level and u=7.1 m/s, respectively. Minimum height of a bore capable of transporting the largest megaclast on the terrace is approximately H=1.6 m at celerity of u=4 m/s. Transport mechanisms seem to vary depending on megaclast shape. Sliding is probably a common mechanism of transport of larger megaclasts of irregular shape whereas somewhat smaller and platy megaclasts are occasionally found in overturned positions. The swell waves are capable of transporting the megaclast as much as 110 m inland from the shoreline cliff.
THE LANDSLIDE ORIGIN OF THE ALEUTIAN TSUNAMI OF 1946: IMPLICATIONS FOR TSUNAMI WARNING
Gerard J. Fryer
University of Hawai`i at Manoa
Honolulu, Hawaii, USA
With glacial advance and retreat, the Aleutian Shelf is alternately exposed and eroded. On the upper slope of the forearc, the prograding wedge of glacial debris of eustatic low stands becomes a zone of mere reworking of ne-grained sediments during high stands. So now a thick wedge of weak material extends from the shelf edge 50km downslope to a midslope terrace 4km deep. Large earthquakes repeatedly shake the entire structure. Failure is inevitable. Failure occurred on April 1, 1946. Evidence for the landslide origin of the 1946 tsunami is overwhelming. No earthquake dislocation can simultaneously explain the near- eld travel time, the long source duration, the phase of the initial wave, and the rapid azimuthal variation of wave heights in both the near and far eld; all have been successfully modeled assuming a landslide source. The landslide (the Ugamak Slide) has only been crudely mapped so far, but will be the subject of intense eldwork later this summer. Because of the shallow depth of only 110 m, the heads of landslides along the Aleutian Shelf edge accelerate to more than 60% of celerity within a few minutes of failure. The closer a slide approaches the phase velocity of a tsunami, the stronger the excitation of surface gravity waves in the direction of slide motion and the narrower the radiation pattern. In 1946 a narrow beam of large waves was projected the length of the Paci c, from the Aleutians to Antarctica. In Hawai`i, runup on open coasts averaged 6 m, with ampli cation to 18m from local topography. Hawai`i, however, was o the direction of slide motion; the largest waves actually passed east of the islands. Those waves hit the Marquesas, where, despite being twice the distance from the source as Hawai`i, runup was much larger, averaging 8m on open coasts, and reaching 20m in narrow valleys. If such a tsunami were to hit Hawai`i squarely, average runup would exceed 10 m, with maxima over 25 m|far larger than anticipated in
any evacuation plans. With existing instrumentation, a large retrogressive failure of the Ugamak Slide would unquestionably trigger a tsunami warning, because one of the sea floor (dart) tsunami gauges lies within the expected radiation pattern. But if an earthquake triggered a submarine slide farther west, where the trench normal points more directly at Hawai`i, the central beam of the tsunami would pass between dart instruments such that, if the measured open-ocean tsunami were interpreted as coming from an earthquake, no warning would be issued. Conditions for landslides to produce large, narrow-beam tsunamis (thick glacial debris,
a long gentle slope, a shallow shelf edge) exist along the eastern Aleutians from the Alaska Peninsula to Umnak Island, a distance of 600 km. Future landslides in this area are inevitable, making two actions imperative: (1) detailed mapping of the upper Aleutian forearc to determine size, number, and frequency of submarine landslides, and (2) rapid di erentiation of landslides from tsunamis in seismic data so that dart data can be interpreted correctly.
THE LATE MINOAN TSUNAMI IN THE EASTERN MEDITERRANEAN: A RE-EXAMINATON
Kingston University, Kingston, United Kingdom
It has long been speculated that the paroxysmal eruption of Santorini volcano (circa 3,500 BP) referred to as the Late Minoan (LM) eruption resulted in the generation of a tsunami. The occurrence and probable impacts of this tsunami have frequently been cited
in the text of scientific papers and articles and appear to have been assumed as proven. This presentation reviews the arguments previously forwarded to imply the occurrence of the tsunami and summarises the archaeological and geological evidence. The presentation then re-examines the original arguments presented for an LM tsunami in the light of recent volcanological investigations. This indicates that previously published arguments may be challenged because the assumptions on which they were based are flawed. The re-analysis of the original tsunami hypothesis indicates that there is insufficient evidence to demonstrate that a large tsunami propagated throughout the eastern Mediterranean circa 3,500 years BP.
USING TSUNAMI DEPOSITS TO IMPROVE UNDERSTANDING OF THE JUNE 23, 2001 PERU TSUNAMI
Bruce E. Jaffe, Dave M. Rubin and Robert Peters
U. S. Geological Survey, Santa Cruz, CA
Guy Gelfenbaum, and Roberto Anima
U. S. Geological Survey, Menlo Park, CA
Matt Swensson, University of Southern California, Los Angeles, CA
Daniel Olcese and Luis Bernales Anticona
Direccion de Hidrografia y Navegacion de la Marina de Guerra del Peru
Juan Carlos Gomez, Instituto Geoffisico de Peru
Percy Colque Riega, University of San Agustin, Peru
The June 23, 2001 Peru tsunami was devastating in the Camana region of southern Peru, killing dozens of people and destroying approximately 2000 buildings. The tsunami left sedimentary deposits (tsunami deposits) that can be interpreted to better understand flow conditions during the event. A team of 16 scientists from Peru and the United States conducted a field investigation from September 6-15, 2001 to study these tsunami deposits. The team measured sediment characteristics and topography along shore-normal transects at 6 locations from Playa Chira (20 km north of Camana, the hardest hit area) to Pampa Grande (25 km south of Camana). Transect locations were chosen during a field reconnaissance guided by inundation and runup data collected in early July by the 1st International Tsunami Survey Team (1st ITST). One transect was located where the 1st ITST did not make measurements. At each location, sediments were trenched, described, photographed, and sampled. Topographic profiles and indicators of tsunami
flow direction, inundation, and runup were also measured. Tsunami deposits were found at all transects. Some tsunami deposits were easy to identify, such as in the agricultural fields where tsunami deposits were sandy sediments with mud rip-up clasts found overlying mud. However, most were difficult to identify during the early part of the field survey, but after a week in the field, the team’s knowledge had increased enough to make positive identification of tsunami deposits in difficult depositional settings. For example, on beaches, tsunami deposits were sand overlying sand and sedimentary features such as trample structures in the underlying beach sand, thick tsunami layers, and vertical grain size variation were used to distinguish between tsunami and beach deposits. Some tsunami deposits had multiple layers, and in some of these deposits the number of layers equaled the number of tsunami waves reported. Estimates of inundation and runup using tsunami deposits mads more than two months after the tsunami were similar to those measured by the 1st ITST at the same sites. Inland inundation ranged from 230 m on the steeply sloping beach at Playa la Chira to about 1 km at the gently sloping coastal plain of La Quinta. The landward limit of tsunami sedimentation at all locations was slightly less (within 25 m) than the limit of inundation Runup elevation ranged from 3.0 m at Pampa Grande to 8.2 m at Playa la Chira (both within 0.5 m of measurements by the 1st ITST). Sediment was not deposited at the highest runup elevations where the tsunami reflected off a cliff (in these cases runup is from wave splash and a detectable deposit would not be expected), but was deposited near the elevation of runup in areas where the tsunami was not reflected off a cliff. New information on tsunami,
including estimates of runup and inundation, was obtained at the location the 1st ITST did not visit. In addition to being able to estimate inundation and runup, it is possible to estimate other wave characteristics from tsunami deposits. We present estimates of tsunami flow velocity made using deposit data collected in Peru and a newly developed tsunami sedimentation numerical model.
QUEENS BEACH TSUNAMI DEPOSIT BASELINE STUDY
Barabara H. Keating
University of Hawaii, Honolulu Hawaii USA
Queens Beach is located on the SE tip of Oahu, Hawaii. The beach is the newly designated site of a regional park, providing access to Makapu Lighthouse, scenic overlook, and hiking trail. Topographic features within the Queens Beach coastal zone occur in a 3-
tier configuration. The modern beach and near-shore dunes are associated with the current sea level and consist of sands and coral clasts that are a conspicuous bleached-white color. Inland of the modern beach is a terrace consisting of volcanic outcrops and yellow sands, with abundant coral clasts and isolated pockets of red soil, truncated at an elevation of 5.7-6 m above sea level. This terrace is probably equivalent to wave cut notches at 6.6 and 8.1 m elevation at nearby Kailua, that are dated at 120,000- 125,000 years old. The third topographic tier is boulder debris bulldozed into an area adjacent to the highway area during 1972-1975. (The area modified in the 1970’s will become a parking lot for the regional park.) The Queens Beach has been inundated by 3 tsunamis within the last 100 years, with inundations of 5m, 3m, and 4m due to the 1946 Aleutian Tsunami, the 1952 Kamchatka Tsunami, and the 1960 Chile Tsunami (runup above sea level, Walker, 1994). The 1946 tsunami destroyed a group of houses built on the coastal plain, destroyed the highway (constructed in 1932), eroded and truncated dunes and left a steep sand beach (opposed to the modern coral clast beach). An Army Corps of Engineers analysis suggests Queens Beach is likely to be inundated every 25 years by tsunami waves of 2.4 m and every 100 years by tsunami waves of 6.6 m. In order to better understand the redistribution of sediments due to tsunami inundation, sedimentary clasts were collected in a profile at Queens Beach, to establish a baseline of clast sizes distribution in the modern back-beach setting. Following future tsunami inundations, this site can be reoccupied (using GPS positioning) to establish profiles of clast distribution, which can be used to place constraints on tsunami wave carrying-capacity,
and characteristics of tsunami deposition on ocean islands. The Queens Beach profile is situated NE of the current Alan Davis beach parking area. Mr. Davis resided in one of the beach homes destroyed by the 1946 tsunami. He reported walking through chest-high water, during the initial tsunami wave inundation, suggesting the wave was not very turbulent. Photographs documenting the event show the Queens Beach area was covered by thick sand deposits. Currently the area is deprived of sands, with only 6 inches of sand on a well-cemented red clay substrate. Clasts were dominantly subangular pahoehoe basalt and a’a basalt but also included clasts of ash, pumice, and coral. These were collected and measured. The roundness, circumference, and length of the primary (a) axis all decrease landward. This pattern is interpreted as a decreasing volume of rounded (beached-derived) rocks away from the surf zone and a decrease in clast size as waves moved inland and lost carrying-capacity.