TSUNAMI OBSERVATIONS, MODELLING AND HAZARD
REDUCTION (IASPEI, IAPSO, IAVCEI, IUGG TSUNAMI
COMMISSION)
Location: Haworth Building 101LT
Location of Poster: Old Gym
Thursday 29 July AM
Presiding Chairs:V. Gusiakov (Inst. of Computational Mathematics
& Mathematical Geophysics, Russia)
N. Shuto (Iwate Prefectural University, Japan)
tsunami mitigation and hazard reduction
Introduction 0930
V. Gusiakov
JSS42/W/27-B4 0940
CONTRIBUTION TO THE IDNDR: TSUNAMI FLOODING MAPS
E.N. Bernard, Tsunami Commission
Since 1989, the IUGG Tsunami Commission has facilitated the dedicated efforts of tsunami scientists to produce over 30 tsunami inundation maps for mitigating the effects of tsunamis in 17 threatened countries. Through a process of planning, proposal acceptance, tsunami data collection, and numerical model development, an internationally accepted methodology for preparing tsunami flooding maps has been created as the tsunami community's contribution to the International Decade of Natural Disaster Reduction. More importantly, scientists in 17 nations were trained to use of this methodology enabling those countries to produce more maps in the future. The value of such maps is that they identify coastal areas that could be flooded by future tsunamis that require some evacuation procedure to remove people from tsunami danger. To accomplish this goal, the tsunami scientific community had to carefully measure the effects of tsunamis and use these data to construct numerical models that could realistically simulate tsunami dynamics. In the past decade 9 destructive tsunamis have killed over 3,200 coastal residents in 7 countries. Each of these tsunamis were surveyed by teams of international tsunami scientists to collect measurements on the extent of flooding along affected coastlines. All surveying scientists generously shared their data with all tsunami scientists quickly and efficiently through internet technology. In addition to new data ,a focussed modeling effort was required to convert tsunami flooding data into useful emergency management information. A partnership between the the IUGG Tsunami Commission, the United Nation's Intergovernmental Oceanographic Commission, and Tohoku University in Sendai, Japan was created to support the Tsunami Inundation Modeling Exchange (TIME) project. TIME is a modeling center at Tohoku University, under the leadership of F.Imamura and N. Shutothat trains scientists from other countries to use numerical models to estimate the extent of potential tsunami flooding for any threatened community. So far, these models have been transferred to scientists in seventeen countries, including: Australia, Canada, Chile, Columbia, Costa Rica, Ecuador, Greece, Indonesia, Italy, Japan, Korea, Mexico, New Zealand, Peru, Puerto Rico, Turkey, and the United States who have produced more than 30 inundation maps for tsunami threatened coastal communities. The global emergency management community now has a better tool to mitigate the impact of tsunamis.
JSS42/E/15-B4 1000
TRAFFIC HINDRANCE AFTER TSUNAMI
Nobuo SHUTO(Faculty of Policy Studies, Iwate Prefectural University, Takizawa, Iwate 020-0173, Japan, email shuto@iwate-pu.ac.jp)
It is important to open coastal traffic after a tsunami, in order to ensure a quick rescue operation within the "Golden 24" hours. Eighty five examples of coastal traffic hindrance were collected, classified and analyzed for 11 tsunamis in Japan.
First cause of the traffic hindrance is destruction of structures. (1)Destruction of embankment of roads and railways by overflowing sea water. Destruction might begin with the erosion of slopes of embankment. If an embankment is lower than 0.5 m, and if the thickness of overflowing water is smaller than 1 m, then the embankment is not damaged, wigh no regard that the road is paved or not paved. (2)Damage due to concentration of water flow. Erosion around a bridge abutment begins at the end of wing walls and develops to the complete destruction of embankment, even though the abutment can withstand. A bridge pier is tilted or turned due to erosion around it. (3)Damage caused by impact of floating materials. Concrete bridges can withstand. Wooden bridges and light steel plate girder bridges are destroyed. Railways is twisted due to movement of railway bridge. (4)Damage caused by buoyancy. Rails are moved due to buoyancy of wooden sleepers. Second cause is hindrance of function. (1)Standing sea water in low land. (2)Sand and mud depostion on roads and railways. (3)Depositions of logs, ships, houses and debris on roads. (4)Fires on or along roads.
JSS42/E/19-B4 1020
MITIGATION STRATEGIES BASED ON LOCAL TSUNAMI EFFECTS
JANE PREUSS (Urban Regional Research, 1809 7th Avenue, Suite 1000, Seattle, Washington, 98101 USA, tel.: (206) 624-1669, fax: (206) 626-5324, email: jpreuss@nwlink.com); Razvan Bidoae and Peter E. Raad (Mechanical Engineering Department, Southern Methodist University, Dallas, Texas, 75275-0337 USA, tel.: (214) 768-3043, fax: (214) 768-1473, email: peter@seas.smu.edu
When the local effects imposed on structures are better understood, the knowledge constitutes an important tool that planners and regulators can use when making decisions with respect to land use practices in coastal areas. An initial field inventory defined key vulnerability issues for analysis through numerical simulation. Structures meeting the following three criteria presented the most significant risks: a) potentiality for significant damage, b) high disruption or loss of life, and c) potentiality for interactive damage. In this work, the impact of a water wave with a single, tall structure was numerically simulated and the forces and moments on the structure were calculated. To assess the effects of upstream conditions on the forces experienced by the main structure, four simulations were considered. In the first simulation, the flow path of the wave toward the structure was unobstructed. The results revealed the existence of two large peaks in the force and moment fields. The initial impact (and force peak) occurs at the base of the structure, as expected. The unexpected secondary impact occurs due to the collapse of the water column that ascended along the front face of the structure. In the second simulation, a short, wide obstacle was placed in front of the main structure. The obstacle (dike, shorter building, etc.) causes the initial impact to occur at a higher elevation, giving rise to a significant increase in the moment experienced by the structure. In the third simulation scenario, single and multiple pylons were placed in front of the structure. The results indicated a reduction in the amount of water impacting the structure, but the decreases in impact pressure, force, and moment were not significant. In the fourth simulation, a groove (or trough) was prescribed in front of the structure (the containment dike used around tank farms). This final simulation indicates a significant decrease in the forces and moments exerted on the structure. Land use implications, pertain to structures fronting the water, and also the second and third tiers of buildings, which often differ in building quality (e.g., hotels in front and behind older one or two-story homes). Site planning implications include the layout of structures, relationships of primary to accessory structures such as beach houses, parking lots, and open spaces.
JSS42/E/02-B4 1040
A FRESH LOOK AT ELEMENTS OF THE TSUNAMI WARNING SYSTEM IN THE PACIFIC
Michael BLACKFORD (International Tsunami Information Center, US National Weather Service, 737 Bishop Street Suite 2200, Honolulu Hawaii 96813-3213, USA, email: michael.blackford@noaa.gov)
In the aftermath following the 1960 Chilean earthquake and tsunami the Tsunami Warning System in the Pacific, was established in 1968 under the auspices of the UNESCO Intergovernmental Oceanographic Commission. An event of similar magnitude has not recurred in the last 30 years, however, thousands of lives have been lost and millions of dollars of damage have been wrought since the System was established. These losses can largely be attributed to insufficient preparation in many areas for tsunami events on a regional or local scale. Although there are several exceptions, the Pacific States have generally moved slowly with regard to the development of adequate tsunami warning subsystems.
Are States overwhelmed by subsystem start up costs? Do they believe the centralized System fulfills their needs? This paper looks at recent advances in communications and data acquisition that can bring down dramatically the costs of operating a tsunami warning system. It proposes a fresh look at a System that emphasizes the development of distributed subsystems that can be much more responsive to the needs of those vulnerable to tsunamis.
JSS42/W/24-B4 1120
U.S. TSUNAMI HAZARD MITIGATION PROGRAM
Bernard, E.N., NOAA/Pacific Marine Environmental Laboratory, bernard@pmel.noaa.gov
The National Tsunami Hazard Mitigation Program is a Federal/State partnership (National Oceanic and Atmospheric Administration, Federal Emergency Management Agency, United States Geologic Survey/ Alaska, California, Hawaii, Oregon, and Washington) committed to reducing the vulnerability of U.S.coastal communities to tsunami hazards. Through three interdependent activities of tsunami flooding mapping for coastal communities, tsunami warnings, and education, the program is making progress on a broad range of emergency planning practices to mitigate the effects of tsunamis. A summary of the program's activities for the past 3 years will be given with emphasis on reducing coastal community vulnerability to tsunami hazards. A description of the program can be found at http:/ /www.pmel.noaa.gov/tsunami-hazard/.
JSS42/W/25-B4 1140
EARLY DETECTION AND REAL-TIME REPORTING OF DEEP-OCEAN TSUNAMIS
F.I. GONZALEZ, E.N. Bernard, H.B. Milburn and H.O. Mofjeld, (all at NOAA’s Pacific Marine Environmental Laboratory, 7600 Sand Point Way, NE, Seattle, WA, 98115)
Current tsunami warnings are based on seismic data and coastal tide gage observations. But neither provides direct measurement of tsunami energy propagating toward coastal communities. As a consequence, an understandably conservative tsunami warning philosophy has produced an unacceptably high false alarm rate -- approximately 75% since 1950. These false alarms are a serious problem because they are expensive, they undermine the credibility of the warning system, and they place citizens at physical risk of accidental injury or death during an evacuation. The speed and accuracy of tsunami warnings can be improved when real-time reports are made of deep ocean tsunami data collected near the source region within a few minutes of generation. Such data will enable a more direct and rapid assessment of the hazard and, when coupled with model forecasting tools, a more accurate prediction of the impact on specific coastal communities. Thus, destructive tsunamis will be identified more reliably and the number of false alarms will be reduced. An added benefit of the real-time data stream is continued offshore tsunami monitoring. Dangerous conditions can persist for several hours after the first wave strikes a community because very large tsunamis can have periods as long as an hour and the largest wave may arrive as late as the third or fourth in a series. So offshore tsunami monitoring will provide important guidance for decision- makers who must judge the risk of deploying rescue and recovery personnel and equipment and, when the area is safe for the return of residents, sound the "all clear." NOAA’s Deep-ocean Assessment and Reporting of Tsunamis (DART) Project is an effort of the U.S. National Tsunami Hazard Mitigation Program to develop this early tsunami detection and real- time reporting capability -- a formidable technological and logistical challenge. DART systems utilize bottom pressure recorders (BPRs) that are capable of detecting and measuring tsunamis with amplitude as small as 1 cm in 6000 m of water. The data are transmitted by acoustic modem to a surface buoy, which then relays the information to a ground station via satellite telecommunications. This design has been tested and refined through several deep ocean deployments of prototype systems. The latest prototype DART system was deployed on September 30, 1998 at Ocean Weather Station PAPA (50 N, 145 W). If this prototype system survives the winter, three additional systems will be deployed in 1999 off Alaska and Oregon as part of a planned six-buoy network in the north Pacific and equatorial region. The network is designed to provide early detection and measurement of tsunamis generated in source regions that threaten U.S. coastal communities: the Alaska Aleutian Subduction Zone, the Cascadia Subduction Zone, and the South American Seismic Zone.
JSS42/W/23-B4 1200
ELEMENTS OF A TSUNAMI WARNING SYSTEM FOR THE INTRA-AMERICAS SEA
GEORGE A. MAUL, Florida Institute of Technology, 150 West University Boulevard, Melbourne FL 32901 USA, and Douglas M. Martin, NOAA National Ocean Survey, 1305 East-West Highway, Silver Spring MD 20910 USA
At a series of workshops (Barbados, 1995; St. John, 1996; Puerto Rico, 1997; Miami, 1998) the case has been made for significant tsunami hazards in the Intra-Americas Sea (Gulf of Mexico, Caribbean Sea, Bahamas, and Guianas). For example, since the great 1867 US Virgin Islands earthquake and 9 meter-high tsunami, the population density of the region has increased 10-fold, infrastructure development has progressed without a notable natural hazards component, and governments seem to be oblivious to the risk. Accordingly, four essentials comprise the proposed Intra-Americas Sea Tsunami Hazards System: Education, Warning, Management, and Research.The first order of business is to better educate the populace through public information, K-12 student indoctrination, video and other multi-media products, workshops, and popular press articles. The warning component should capitalize on the recently established CPACC (Caribbean Planning for Climate Change) sea-level/weather GOES-reporting network, on existing seismic and meteorological reporting and warning systems, and on active participation with the ICG/ITSU (International Co-ordination Group for the Tsunami Warning System in the Pacific). Management issues include: integration with other natural hazards warning systems; exploration for funds; local warning and evacuation; search and rescue; fire suppression; emergency medical services; damage assessment; inter- and intra-governmental coordination; and, damage and hazard analysis. Research needs are: improved resolution bottom relief data; travel time maps for population centers; earthquake magnitude / depth thresholds; tsunami wave arrival amplitude estimation; potential for Kick’em Jenny and Soufriere (Montserrat) eruption; tsunami and earthquake history improvements; fault locations, activity, and tsunamigenic mechanisms; inundation maps; GPS stations for crustal motion monitoring; and loss estimation studies, amongst others.
JSS42/E/16-B4 1220
CONTEMPORARY ASSESSMENT OF TSUNAMI RISK AND IMPLICATIONS FOR EARLY WARNINGS FOR AUSTRALIA AND ITS ISLAND TERRITORIES
Jack RYNN (Centre for Earthquake Research in Australia, PO Box 276, Indooroopilly, Queensland 4068 Australia, email sally.brown@uq.net.au; Jim Davidson (Bureau of Meteorology, GPO Box 413, Brisbane, Queensland 4001, Australia, email jdavidson@bom.gov.au)
The natural hazard of tsunami relative to Australia and its Island Territories is perceived to be of little or no consequence, hence a small risk, when compared to our other more frequent natural disasters. The historical record shows that tsunami damage, although rare, occurred along the eastern seaboard (1877 and 1960 Chile earthquakes), and northwest coast (1983 Krakatoa volcanic eruption and the 1977 and 1994 Indonesian earthquakes). Tsunami mitigation is a need because the island nations of Australia depend on its coastal facilities for sustainable development, with 90% of the population domiciled in this environment. One Australian IDNDR project assessed the tsunami risk to its shorelines and island Territories. A specific methodology invoking a multidisciplinary approach was developed. More than 350 earthquakes and specific submarine volcanoes and landslide 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, tsunami data base, maps of potential tsunamigenic sources, tsunami travel time charts and relationships between relevant tsunami parameters. This provides practical applications to upgrade the Bureau's warning procedures and emergency services counter disaster planning, including development of a regional tsunami warning system.
Thursday 29 July PM
Presiding Chair: K. Satake (Seismotectonic Section, Geological Survey of Japan)
S. Tinti (University of Bologna, Italy)
Tsunami generation and seismotectonics
JSS42/W/08-B4 1400
MODELING FOR TSUNAMI GENERATED BY LANDSLIDING
Fumihiko Imamura; Kazumasa Hashi, Tomohiro Matsumoto, Nobuo Shuto Disaster Control
Research Center, School of Eng., Tohoku University, Aoba 06, Sendai 980-8579, Japan, imamura@tsunami2.civil.tohoku.ac.jp
The numerical model to simulate tsunami propagation and generation due to a marine as well as land sliding is developed by deriving the governing equation and boundary condition for two layer flows with different density. The governing equation is derived by integration of Euler equation with viscosity , and by using kinetic and dynamic boundary conditions at free surface, interface, and sea bottom. Numerical scheme and procedure is designed the same as the standard model ; TUNAMI code for applications. It is found that the front condition of the lower layer as landslide is very important for numerical stability. The form/drag force and interaction of momentum between two layers at the front are introduced to solve the unstability. The numerical model is applied to two tsunami cases; the 1741 Oshima-oshima tsunami caused by lansliding on the volcanic island and the 1998 Sissano tsunami in PNG which might be related with a marine landslide. Although a pyloprastic model by Aida could not reproduce the runup heights measured along the peninsula of Oshima, the present model assuming the volume of landslide estimated from the change before and after eruption well simulate the distribution which is energy concentration on Kumaishi in north and Era in south. In the another case for marine landslide, it is demonstrated the interaction between two tsunamis caused by fault motion and landslide, suggesting the energy focus on the Sissano lagoon.
JSS42/W/03-B4 1420
ENERGY BALANCE IN THE PROBLEM OF LANDSLIDE-INDUCED TSUNAMIS
StefanoTinti and Cinzia Chiavettieri (both at Dipartimento di Fisica, Settore di Geofisica, Università di Bologna, email: steve@ibogfs.df.unibo.it)
Tsunamis generated by the movement of submerged large masses of sediments accumulated on the sea bottom over long periods of time are not rare and can be even very severe. The body of water and the body of sediments can be physically seen as two interacting systems exchanging energy. The main interest is typically focused on the energy gained by the water and manifesting in form of gravity water waves (the tsunami), but in principle the two systems are fully coupled, with energy flowing in both directions: energy is transmitted by the moving mass to the water giving rise to waves and energy passes from the waves to the mass, which means that the motion of the landslide is influenced by the propagation of the waves. This two-way nonlinear exchange of energy is generally overlooked. Here it is explored and consequences on modeling of tsunamis induced by underwater landslides are discussed.
JSS42/E/06-B4 1440
LANDSLIDE TSUNAMI ‘GENERATION MECHANISM AND ITS DETECTION FOR EARLY TSUNAMI WARNING ISSUE’
S.I.IWASAKI(cuh@ess.bosai.go.jp) and S.Sakata (both at the National Research Institute for Earth Science and Disaster Prevention, 3-1 Tennodai Tsukuba 305-0006, JAPAN)
The cause of the Papua New Guinea Tsunami was probably landslide triggered by the original earthquake. As this example shows, these kinds of tsunamis generated by non-seismic causes sometimes gave heavy damages. Many reasons can be cited, such as the natural precursor of tsunamis, that is "shaking", were weak or nothing, generation regions were located in near shore regions, almost tsunami warning systems issue the tsunami warnings based on the Magnitude of earthquakes, generation mechanisms are still not fully clear and so on. In this presentation, at first, the simple landslide tsunami generation model will be introduced. Then, using a simplified topography of the ocean off Papua New Guinea coast, differences of wave forms of tsunami generated by a landslide and an earthquake will be discussed. A method to distinguish tsunami generated by an earthquake or a landslide will be shown. Finally, application of newly invented laser tsunami meter will be presented. The main conclusions are as follows; 1. Tsunami generated by a landslide shows strong directivity compared with that generated by an earthquake. 2. For the early tsunami warning issue, detection of tsunami informations at on site is essential. 3. For this aim, the cheaper tsunami warning systems are necessary and the newly invented laser tsunami meter is one of these.
JSS42/E/12-B4 1500
NUMERICAL SIMULATION OF THE LANDSLIDE-GENERATED TSUNAMI OF NOVEMBER 3, 1994 IN SKAGWAY HARBOR, ALASKA
Evgueni A. Kulikov and Alexander B. RABINOVICH (both at Tsunami Center, Shirshov Institute of Oceanology, 36 Nakhimovsky Prosp., Moscow 117851, Russia, e-mail abr@tsucen.msk.ru); Isaac V. Fine and Brian D. Bornhold (both at International Tsunami Research, Inc. 11321 Chalet Rd., Sidney, BC, V8L 5M1, e-mail itri@ii.ca); Richard E. Thomson (Institute of Ocean Sciences, 9860 W.Saanich Rd., Sidney, BC, V8L 4B2, Canada, e-mail thomsonr@dfo-mpo.gc.ca)
We examine the origin and behavior of the catastrophic tsunami that impacted Skagway Harbor, Alaska, on November 3, 1994. Geomorphic and tide gauge data, combined with a rigorous numerical modeling of the event, reveals that the tsunami was generated by an underwater landslide which also resulted in collapse of a cruise-ship dock undergoing construction. The three-dimensional shallow-water numerical model for a viscous landslide with full slide-wave interaction and subaerial slide was used to simulate tsunami waves in Skagway Harbor caused by the failure.The results of the numerical simulation closely agree with the NOAA tide gauge record and eyewitness observations. The landslide has been linked to critical overloading of the slope at a time of extreme low tide and is consistent with similar events in other coastal regions of the world. In general, we conclude that the numerical model of the submarine landslide associated with the dock failure in Skagway Harbor accounts for all aspects of the observed wave field.
JSS42/W/22-B4 1520
"RED", "GREEN" AND "BLUE" TSUNAMIGENIC EARTHQUAKES: DOES ANY PHYSICAL BASIS FOR THIS CLASSIFICATION EXIST?
V.K. GUSIAKOV (Institute of Computational Mathematics and Mathematical Geophysics, Siberian Division, Russian Academy of Sciences, Novosibirsk 630090, Russia, Email: gvk@omzg.sscc.ru)
In the paper (Chubarov, Gusiakov, 1985), the theoretical dependence of the tsunami intensity I (on the Soloviev-Imamura scale) on the moment-magnitude Mw of a submarine earthquake has been obtained (I = 3.55Mw - 27.1). This formula can be used for calculation of expected (on the basis of the "piston-like" model of the tsunami generation) tsunami intensity for a submarine earthquake with the known moment-magnitude. In this study, we introduce the formal classification of the Pacific tsunamigenic earthquakes on the basis of their dI parameter, that is the difference between the actually observed and the expected tsunami intensity (dI=Iobs-Iexp). Based on the dI value, we divide all tsunamigenic earthquakes with known I and Mw into the three groups: "red" (dI>1), "green" (-1<dI<1), and "blue" (dI<-1). From 293 tsunamigenic events that occurred in the Pacific from 1900 to 1998 and which have both I and Mw values, 90 events fall within the "red" group, 153 are within the "green" group and 50 events are within the "blue" group. The geographical distribution of these tsunamigenic events shows their clear correlation with the climatic and circumcontinetal zonation in the oceanic sedimentation as described in (Lisitzyn, 1996). Namely, two of the five major zones of the oceanic lithogenesis (northern humid zone and equatorial humid zone), that are characterized by the increased rates of the oceanic sedimentation, are clearly indicated by the increased number of the "red" tsunamigenic earthquakes (for instance, 6 of 7 events occurred within the Java trench are "red"). The circumcontinental zonation is clearly expressed by the fact that all the tsunamigenic events occurred in this century in the East China, the Yellow, the Japan, the Okhotsk and the Bering Seas are "red". On the other hand, all major submarine earthquakes that occurred in the remote (from continents) tsunamigenic zones like Guam, Tonga, New Zealand are "green" or "blue". Despite a considerable Mw value (greater than 7.5), many of these events have generated very minor tsunamis with the run-up heights less than 1 meter. The results of this study show that the earthquake-induced disturbance of the bottom sediments, resulted in submarine slumping, can be the leading factor controlling the tsunami generation mechanism, and these processes should be taken into account in the operational tsunami warning as well as in the coastal tsunamizoning.
JSS42/W/25-B4 1600
LIMITATIONS OF FAULT DISLOCATION MODELS FOR TSUNAMI SIMULATIONS ON THE CASCADIA SUBDUCTION ZONE, NORTHWEST COAST, NORTH AMERICA
GEORGE R. PRIEST (Oregon Department of Geology and Mineral Industries, Suite 965, 800 NE Oregon St., #28, Portland, OR 97232, email george.priest@state.or.us) Edward Myers and António M. Baptista (both at Oregon Graduate Institute of Science and Technology, PO Box 91000, Portland, OR, 97291-91000, email: emyers@ccalmr.ogi.edu; baptista@ccalmr.ogi.edu); Paul Flück (Swiss Federal Institute of Technology, Zürich, Switzerland); Kelin Wang (Geological Survey of Canada, Pacific Geoscience Centre, P.O. Box 6000, Sidney, B.C. V8L4B2, Canada, email: wang@pgc.emr.ca); Curt D. Peterson (Department of Geology, Portland State University, Portland, OR, 97207-0751, email: curt@ch1.ch.pdx.edu)
Three dimensional fault dislocation models (Okada’s (1985) point source solution for an elastic half space) are used here to simulate coseismic ruptures on the Cascadia subduction zone for tsunami simulations. Earthquake recurrence of ~450 years, slip of ~15-20 m, and rupture widths (140 km) extending well onshore fit paleoseismic data. Narrower ruptures (70 km) that occur mostly offshore in Oregon and northern California fit thermal and geodetic data. The wide ruptures produced tsunamis about 30 percent lower than narrow ruptures, all other factors equal. Tsunami height varied directly with total slip. A rupture 1,050 km long with 15-20 m slip produced 6-9 m tsunamis, about twice as high as tsunamis from segmented ruptures 450 km in length with 7-10 m slip. The latter ruptures are most consistent with coupling and aspect (length:width) ratios of world wide analogues to Cascadia. Distribution of paleotsunami deposits appears to favor the 6-9 m tsunamis, but uncertainties in the fault dislocation models preclude assuming that lower slip (450 km) ruptures are not reasonable sources. None of the fault dislocation scenarios accurately depict possible inelastic deformation in the accretionary wedge or deformation from local faults, asperities, and submarine landslides, any of which could strongly modify deformation (and tsunamis) relative to the simulations.
Simulation of inelastic deformation in the rheologically weak leading edge of the accretionary wedge was attempted by decreasing slip there. Simulations produced anomalous "spikes" of uplift, because the elastic model treats the area of decreased slip as a strong barrier rather than (as intended) a weak mass. Slip was thus deflected upward at this "barrier," effectively reproducing the same total uplift as full slip, but with a different shape. Tsunami run-up therefore did not decrease as expected.
JSS42/W/18-B4 1620
THE DESTRUCTIVE SEISMIC ACTIVITY OF 1303 IN THE EASTERN MEDITERRANEAN: A POSSIBLE INTERPRETATION BASED ON REALISTIC SYNTHETIC WAVEFORMS
El-Sayed Attya (Geological Department Mansoura University, Mansoura, Egypt, The Abdus Salam International Center for Theoretical Physics, SAND group, Trieste, Italy and Dipartimento di Scienze della Terra – Università di Trieste, Via Weiss, 4 34127 Trieste,
email elsayeda@geosun0.univ.trieste.it)Fabio Romanelli (Gruppo Nazionale per la Difesa dai Terremoti - CNR, Rome, Italy and Dipartimento di Scienze della Terra – Università di Trieste, Via Weiss, 4 34127 Trieste, Italy, email romanel@geosun0.univ.trieste.it)Giuliano F. PANZA (Dipartimento di Scienze della Terra – Università di Trieste, Via Weiss, 4 34127 Trieste, Italy and The Abdus Salam International Center for Theoretical Physics, Miramar, Italy, email panza@geosun0.univ.trieste.it)
The Hellenic Arc is located about 500 km away from the Egyptian border, but earthquakes generated in this area can cause severe effects in Egypt. The observed shaking duration in Lower Egypt due to an event located in the Hellenic Arc is, on average, 3 minutes long. The seismic activity of 8th of August 1303 seems an exception to this pattern in terms of damage and duration of shaking. It was strongly felt in Lower Egypt for about 15 minutes and caused a widespread damage in Crete, Egypt, Rhodes, Jordan, Syria, Palestine, Turkey and Cyprus. The location of the seismic source(s) is ambiguous. Recent seismic activity, synthetic tsunami and ground motion modeling are used to infer the parameters of the possible source(s) from the available historical records. Using as constraint the available macroseismic and tsunami information we propose that two events occurred on the 8th of August of 1303. One earthquake is a shallow event with small to moderate size, located in Egypt, probably to the south of Cairo. This event can be the source of the extensive damage in the Nile Delta, Cairo and Upper Egypt. The other suggested event has a relatively larger size and is situated in the Hellenic Arc at shallow depth. This second event can explain the observed tsunami in the Eastern Mediterranean, the minor to moderate macroseismic effects in a wide area, including sites in, Jordan, Palestine, Syria and Turkey, and the extensive damage in Crete and Alexandria of Egypt.
JSS42/W/10-B4 1640
NUMERICAL STUDY OF THE SOURCE OF THE JULY 17, 1998 PNG TSUNAMI
Vasily TITOV (NOAA/PMEL, JISAO, 7600 Sand Point Way NE, Seattle WA, 98115; titov@pmel.noaa.gov) and Frank Gonzalez (NOAA/PMEL, 7600 Sand Point Way NE, Seattle WA, 98115; gonzalez@pmel.noaa.gov)
The July 17, 1998 Papua New Guinea earthquake produced a tsunami with very high amplitudes, localized in a small area near the earthquake source. The relatively small earthquake magnitude and the concentration of the tsunami impact have prompted questions about the tsunami source mechanism. Possible source scenarios are landslide, bottom deformation, or a combination of the two. A numerical study of the July 17, 1998 Papua New Guinea tsunami is conducted to estimate discrepancies between different source mechanisms. Several source scenarios are simulated using the MOST numerical model, including a landslide source modeled as a viscous sediment flow and a co-seismic bottom deformation. Sources of various sizes at different locations are tested. Computed runup estimates are compared with measurements obtained by the International Tsunami Survey Team. It is found that several computed scenarios produce the runup distributions along the PNG coast which are reasonably close to the observed runup heights. Both pure landslide and bottom deformation are among these possible generation mechanisms, so it is difficult to distinguish between them on the basis of the runup comparison only. Distinctions in the tsunami runup dynamics for different tsunami sources are discussed. In the case of the PNG tsunami, as well as for many other events, runup values alone present too little information to distinguish between different tsunami source mechanisms. Any additional information about tsunamis, such as current velocities, sediments, number of waves and direct measurements in the possible source area, can provide crucial additional information for the tsunami source determination.
JSS42/W/20-B4 1700
GENERATION OF GRAVITO-ACOUSTICAL WAVES BY BOTTOM DISPLACEMENTS OF FINITE DURATION
M.A. NOSOV, Moscow State University, Faculty of Physics , Russia
The linear potential theory was used for the sake of a comparative study of waves generation by bottom displacements of finite duration in the layer of an ideal incompressible and compressible homogeneous fluid of constant depth in the field of gravity (3-D problem). The analysis of exact analytical solutions to the problem has shown that a substantial difference exists between the behaviour of compressible and incompressible fluids under typical tsunami source conditions. In particular the results of this study indicate that the compressibility of water can not be neglected, if one pretends to give an adequate description of the tsunami generation by bottom displacements of a duration less than 10 H/c, where H is ocean depth in the source region, c is the sound velocity in water. Tsunami source radiates not pure gravitational but a system of gravito- acoustical waves. The further study of the features of this waves could be of interest for tsunami forecasting in terms of tsunami forerunners.
Friday July 23 AM
Presiding Chairs: E.N. Bernard (Pacific marine Environmental Lab, Seattle, USA)
F. Imamura (Disaster Control Res. Center, Tohoku Univ, Japan)
JSS42/W/14-B5 0830
STUDY OF TSUNAMI DEPOSITS IN HOKKAIDO, JAPAN: A PROGRESS REPORT
Kenji SATAKE, Futoshi Nanayama, Koichi Shimokawa (all at Earthquake Research Dept., Geological Survey of Japan, Tsukuba, 305-8567 JAPAN, email satake@gsj.go.jp); K. Shigeno (Meiji consultant co., ltd.)
We are studying tsunami deposits on both the Japan Sea and Pacific coasts of Hokkaido, Japan, to better understand the sedimentology and the history of earthquakes and volcanic eruptions in the region. Our sites include: Taisei town across from Okushiri Island, where we identified tsunami deposits from the 1993 Southwest Hokkaido (Okushiri) earthquake as well as sand deposited by a 1959 typhoon; Ayukawa coast, where we identified tsunami deposits from the 1741 Oshima-oshima eruption; Kiritappu marsh and Bettoga coast, where we found sand layers from at least six tsunamis. The 1993 tsunami deposited four layers at Taisei. We correlate these layers with landward and seaward flows from the two main tsunami waves; the flow directions are shown by imbrication of gravel, current ripples, and the remains of knocked-down plants. On the other hand, the typhoon deposit forms only one layer and its cross bedding records only landward flow. The tsunami's landward-flow layers are mostly marine sand, whereas the seaward flow layers are composed of non-marine sand and a mixture of soil, gravel, and plant remains. The typhoon deposit is composed of mostly marine sand. The 1741 tsunami was associated with the eruption of Oshima-oshima volcano, but the generation mechanism has not been known. We found several cobbles in the tsunami-sand layer in our trench. The Pacific coast of Hokkaido has been affected by tsunamis from subduction-zone earthquakes both nearby and across the Pacific. We found multiple sand layers interbedded with tephra layers: two sand layers above a 1739 tephra, three sand layers between 1694 and the 10th century, and a few more sand layers below the 10th century tephra.
JSS42/W/01-B5 0850
THE GEOMORPHOLOGY AND SKDIMENTOLOGY OF THE JULY 9TH 1956 AEGEAN SEA TSUNAMI, ASTYPALAEA ISLAND, GREECE
Dale Dominey-Howes (Coventry Centre for Disaster Management, School of the Built Environment, Coventry University, Coventry, CV1 5FB, UK, Kmail:apx124@coventry.ac.uk)
Sediments are described from Livadia aad Stavros, Astypalaea Island, Greece and are interpreted to be associated with the Aegean tsunami of9th July 1956. At Livadia, the marine provenance of the tsunami unit is inferred from two observations. Firstly, the similarity between the clasts comprising the tsunami unit and the contemporary beach sediments from which the sediments are derived, and secondly the inclusion of foraminiferal tests within the sediment matrix. Derivation from the 1956 tsunami, rather than from storm surge is inferred from the uniqueness of the deposits within the sedimentary record, their distinct imbrication and orientation, and 137Cs and 210Pb dating of overlying and underlying colluvium. The top of the tsunami sediment layer occurs at+2.00 meters above sea level indicating minimum flood level. At Stavros, the marine provenance of the tsunami sediments is demonstrated by the inclusion of marine mollusca, while the sediments here occur up to an elevation of+10.00 meters above sea level. The minimum flood levels indicated by these deposits are significantly lower than values previously reported. The results presented here represent the first systematic investigation into the geomorphology and sedimentology of modern Aegean tsunami, and, when combined with geomorphological evidence, indicate that previous reports of flooding from the 1956 tsunami were significantly over-estimated. Such over-estimation of flood levels may have important implications for tsunami hazard risk assessment and emergency pre-planning.
JSS42/P/02-B5 0910
GLOBAL DISTRIBUTIONS OF PEAK FREQUENCY AND THE AMPLITUDE TO THE BIGGEST THREE PACIFIC TSUNAMIS IN THIS CENTURY
Kuniaki Abe (Niigata Junior College, Nippon Dental University, Hamauracho 1-8, Niigata City, 951-8580, JAPAN, email:abeku@ngt.ndu.ac.jp)
Amplitude spectra were obtained for tide-gage records of the 1952 Kamchatka, 106 Chilean and 1964 Alaska tsunamis in all over the Pacific. Number of used stations are 38 (1952), 42 (1960) and 61(1964). The spectrum was calculated using Goertzel method for 6 hour's time history including the initial arrival from which tidal level was reduced and plotted for the frequency range from 0 to 2mHz(8.3min in period) at the interval of 0.0l mHz. In the spectrum we noticed a peak frequency and the amplitude. They were discussed in relation to the epicentral distance, which is approximately propagation distance, and azimuth angle of the station to the epicentre.
Average peak frequencies obtained are 0.46 (1952), 0.34 (1960) and 0.36 (1964) in mHz. The difference is mainly attributed to sea-depth at the source because of the proportionality to long wave velocity. The approximate average sea-depths are 2000m, 650m and 150m, respectively. In the amplitude-distance curve amplitude increased or kept a constant value for the distance larger than 10000km. The increase is noticeable for the azimuth direction normal to trench axis. This fact suggests a geometrical contraction of tsunami after passing the critical distance of a quarter of global circumference. The remarkable tendency in the 1952 Kamchatka tsunami is related to station distribution. Azimuth dependences of peak frequency and the amplitude were also observed and support a directivity of wave radiation from the source. Particularly, the largest value in all the peak frequencies for each tsunami is found in the direction approximately normal to the trench axis. If taking into consideration of curving wave ray, we conclude that tsunami including the highest frequency is radiated to the direction normal to the trench axis.
JSS42/E/05-B5 0930
DISTRIBUTION FUNCTIONS OF RUNUP HEIGHTS FOR RECENT TSUNAMIS BASED ON OBSERVED DATA
Efim PELINOVSKY (Laboratory of Hydrophysics and Nonlinear Acoustics, Institute of Applied Physics, 46 Uljanov Str., Nizny Novgorod, 603600, Russia, email enpeli@appl.sci-nnov.ru); Igor Ryabov (Department of Applied Mathematics, Nizhny Novgorod Technical State University, 24 Minin Str., Nizhny Novgorod, 603600, Russia, email igor@cced.kis.ru)
Studying the characteristics of the runup heights for several historic tsunamis, Van Dorn, Kajiura and Go have found that the distribution functions of the tsunami heights are described by the log-normal distribution. The aim of this paper is to obtain the distribution functions of the runup heights for recent tsunamis. The observed data of following tsunamis were analysed: Flores (12.12.92); Japan Sea (12.07.93); Java (02.06.94), Shikotan (04.10.94); Mindoro (14.11.94); Sulawesi (01.01.96); Irian Java (17.02.96), Peru (21.02.1996), Papua New Guinea (17.08.98). It is confirmed that the log-normal distribution describes the observed data well, and parameters of this distribution are found. An important problem of the quality of the prognostic characteristics of the tsunami distribution is discussed taking into account different numbers of the observed points.
JSS42/W/21-B5 0950
IMPACT OF LARGE TSUNAMIS IN THE MESSINA STRAITS, ITALY
Stefano Tinti, Alberto Armigliato and Francesca Fiorini (all at Dipartimento di Fisica, Settore di Geofisica, Università di Bologna, e-mail: steve@ibogfs.df.unibo.it, armigliato@ibogfs.df.unibo.it)
Among the Italian coasts, Messina Straits are one of the most exposed to tsunami attacks. Several historical events are reported for this region, generated both locally and in the surrounding regions, in particular in Thyrrenian Calabria to the north (the 1783 event) and in eastern Sicily to the south (1693). The 28 December 1908 earthquake generated tsunami was the last catastrophic event affecting this region: a large body of documents and reports exists for such event, as well as some tide-gauge records some hundred km away from the source. Nowadays, the Messina Straits coasts are characterised by a large number of urban, industrial and touristic settlements: in addition, towns like Messina and Reggio Calabria represent fundamental nodes for transports. The tsunami hazard assessment for the Messina Straits is then a very relevant problem, which may be faced on the basis of documents available for historical events, of the available geological and seismotectonic information and of numerical modelling and simulations.
JSS42/E/18-B5 1010
TSUNAMIS IN CENTRAL AMERICA
Mario Fernandez Acre, (Escuela Centroamericana de Geologia, Universidad de Costa Rica, San Jose, Costa Rica) Enrique Molina, (Instituto de Sismologia, Volcanologia, Hidrologia y Meteorologia de Guatemala, Guatemala) Jens HAVSKOV and Kuvvet Atakan (both at Institute of Solid Earth Physics, University of Bergen, Allegaten 41, 5007 Bergen, Norway, email jens@ifij.uib.no)
A catalog of tsunamis has been compiled for Central America containing 49 tsunamis for the time period 1539 to 1996. This is believed to be the most complete catalog made so far. Most of the information (43 events) is reported after 1850. 12 of the tsunamis occurred on the Caribbean side and 37 on the Pacific side. 47 tsunamis were related to local earthquakes while 2 were caused by earthquakes outside the region. On the Caribbean side, there only seem to be a tsunami risk due to local earthquakes in the north and the south while on the Pacific side, the whole coast can experience tsunamis. Here the highest risk is along the costs from Guatemala to Nicaragua and outside Central Costa Rica. About 500 persons are known to have been killed by tsunamis and the largest known tsunami occurred in Nicaragua in 1992 killing 170 persons. On the Caribbean side, all earthquakes near the coast and larger then Ms have generated a significant tsunami while on the Pacific side only 43% of coastal earthquakes larger then 7 have generated significant tsunamis. These facts will be used as criteria for the new tsunami warning center to be set up in Central America.
JSS42/E/17-B5 1030
DEVELOPMENT OF LASER TSUNAMIMETER
Shoji SAKATA and Shin'ichi Iwasaki (both at National Research Institute for Earth Science and Disaster Prevention, Ten'nodai 3-1, Tsukuba-shi, Ibaraki-ken, 305-0006, Japan, email: sakata@geo.bosai.go.jp)
Direct observation of tsunami heights in off-shore areas by tsunamimeters installed on the sea bottom is very effective for mitigation of tsunami disasters. We have been developing a new laser tsunamimeter based on the invention by Sakata. The inspilation for this tsunamimeter came from the concept of the borehole laser strainmeter, of which development Sakata has been engaged in. Later Sakata invented an ingenious device for cancellation of effects of temperature change of the sea water around the vessel. By adding this device to the original laser strainmeter the concept of the laser tsunamimeter is completed. This laser tsunamimeter has many advantages over the conventional tsunamimeter of the quartz oscillator type. The change of the diameter of the vessel of the undersea part caused by pressure change due to the tsunami propagation is detected as the change of resonance frequency of the Fabry-Perot interferometer. Due to the simple design of the undersea part, in which neither IC nor mechanically movable part is included, possibility of trouble will besmall. Electricity supply to the undersea part is not necessary since it is connected to the land part by only optical fibers. The slender cable needs smaller amount of cost for production and installation. Expectedresolution is between 0.1-1cm. Recently we completed the production of the undersea part .
JSS42/W/02-B5 1110
RECENT TSUNAMIS OF 1990-1998 IN THE PACIFIC: GENERAL OVERVIEW
V.K. GUSIAKOV (Institute of Computational Mathematics and Mathematical Geophysics, Siberian Division, Russian Academy of Sciences, Novosibirsk 630090, Russia, Email: gvk@omzg.sscc.ru); J.F.Lander (National Geophysical Data Center, Boulder, CO 90309, USA, Email: jfl@ngdc.noaa.gov)
A comprehensive list of recent tsunamis occurred in the Pacific during the last nine years has been compiled as part of the joint IUGG Tsunami Commission and ICG/ITSU Project "Basic Pacific Tsunami Catalog and Database, 47 B.C. - 2000 A.D." The list contains 77 tsunamigenic events occurred in the Pacific from 01.01.1990 to 31.12. 1998. From these events, 60 have validity 4 (definite tsunami), 10 - validity 3 (probable tsunami) and 7 events still have validity 2 (questionable tsunami). For all the events, having at least one run-up measurement, the tsunami intensity I (on the Soloviev - Imamura scale) has been calculated. In terms of this parameter, the smallest but still detectable event was the February 9, 1991 Solomon Islands tsunami (I=-4.5) with the reported maximum wave height of only 3 cm. The three strongest events within this period were the July 12, 1993 Okushiri tsunami (I=3.1) with the highest run-up of 30.6 m, the July 17, 1998 New Guinea tsunami (I=3.1) and the September 2, 1992, Nicaragua (I=2.8). Several major submarine earthquakes having Ms or Mw value greater than 7.6 generated very minor tsunamis (with the wave height less that 0.5 meter). During this period, 30 regional and 1 Pacificwide warnings were issued by the national and the regional tsunami warning centers. Among them, only 10 warnings can be considered as justified, in other cases the maximum run-up height observed at the coast was less than 1 meter. By different reasons, for 3 damaging tsunamis warnings were not issued in time. They are the September 2, 1992, Nicaragua, the December 12, 1992, Flores and the January 1, 1996 Sulawesi tsunamis. Analysis of historical tsunami occurrence in the Pacific in this century shows that the last nine years had the slightly increased level of perceptible (I=1) as well as damaging (I=2) tsunamis as compared to the long-term average. Their annual rates of occurrence in 1990-98 were 2.2 and 1.1 against 1.7and 0.75 in 1900-98, respectively.
JSS42/E/18-B5 1130
A NEW TSUNAMI CATALOGUE FOR EASTERN MEDITERRANEAN SEA
G. A. PAPADOPOULOS (Institute of Geodynamics, National Observatory of Athens, 118 10 Athens, Greece, g.papad@egelados.gein.noa.gr)
A new catalogue of tsunamis occurring in the area of Eastern Mediterranean Sea from the antiquity to 1998 has been compiled on the basis of original historical documents, previous catalogues, scientific papers as well as of geological and archaeological evidence. The format of the catalogue is similar to that arranged by the GITEC group for the new European Tsunami Catalogue. The data incorporated in the new catalogue have been elaborated as for their completeness and reliability described as a function of time and of the particular geographic regions covered.
JSS42/E/20-B5 1150
OFFSHORE GEOLOGICAL ASPECTS OF THE SISSANO TSUNAMI PAPUA NEW GUINEA
DAVID R TAPPIN, British Geological Survey, Keyworth, Nottingham, NG12 5GG UK Takeshi Matsomotu, Japan Marine Science and Technology Centre, Natsushima-Cho, Yokosuka, 237-0061, Japan and shipboard scientific party.
A survey to investigate the offshore aspects of the July 1998 Sissano tsunami was carried out in January 1999 by the Research Vessel Kairei. It was coordinated by the Japan Marine Science and Technology Centre and the South Pacific Applied Geoscience Commission and acquired multibeam bathymetry, seabed sonar, sub-bottom profiles and piston cores. Models of the tsunami event have so far been based upon a fault solution. However, other evidence suggests that seabed faulting may be only part of the cause.
The survey results reveal a complex seabed morphology that is attributable to subduction of the Pacific Plate southwards beneath PNG along the New Guinea Trench. A narrow shelf 10 kilometres wide lies offshore of Sissano and passes seaward into an inner trench slope dissected by submarine canyons. Sediment transported from the land mainly bypasses the upper inner trench slope and is transported out into lower slope basins, into the trench and out onto the Pacific Plate. Sediments on the inner trench wall are cohesive clays, whereas in the lower slope basins and the trench they are interbedded fine clays (hemipelagites) and silts (turbidites). The inner trench slope is faulted and bears large arcuate scars interpreted as signs of slope failure. This failure is interpreted as taking place through rotational faulting rather than as debris flows. Nearshore bathymetry off of Sissano suggests a subsided delta front that may provide a focusing mechanism for the Sissano tsunami.
JSS42/P/01-B5 1210
FLOWS STRENGTH ON LAND OF THE 1998 PAPUA NEW GUINEA TSUNAMI
Hideo MATSUTOMI (Dept. of Civil and Environmental Eng., Akita Univ., Akita, 010-0852, Japan, email matsu@ipc.akita-u.ac.jp); Yoshiaki Kawata, Yoshinobu Tsuji, Koji Fujima, Fumihiko Imamura, Masashi Matsuyama, Tomoyuki Takahashi, Norio Maki and Seh-Sub Han (Members of 1st and 2nd International Tsunami Survey Team); Nobuo Shuto (Faculty of Policy Studies, Iwate Prefectural Univ., Iwate, 020-0173, Japan, email shuto@iwate-pu.ac.jp)
Field survey was conducted two times: during August 3 to 6 and September 30 to October 3, 1998. Main items of the survey were: (1) spatial distribution of both tsunami height and inundation depth in Sissano, Wapapu, Arop 2, Arop 1 and Malol regions, (2) dimension of houses and their floor height, (3) degree of damage to houses and (4) sand erosion depth and sand grain size on the sand spits of Sissano lagoon. Based on these data, (1) current velocities on flat land and on the sand spits, (2) relation between inundation depth and degree of damage to houses, (3) dependency of sand erosion depth on current velocity and so on are discussed. Laboratory experiments were also carried out to confirm the state of flow on the sand spits.
Results indicate that: (1) the largest inundation depth of 10m was found at Arop 2, and the current velocity was estimated greater than 10 m/s, (2) current velocities on flat land had the same tendency as that in past tsunamis, (3) wooden houses which are church were destroyed when the inundation depth exceeded 1 m. This critical depth is very small compared with that in other places and (4) the dependency of sand erosion depth on current velocities on flat land is the same as that in past tsunamis. But, that on the sand spits was different. As a reason, it is considered that undulate jumps would be formed on the lagoon side.
JSS42/W/15-B5 1230
CLUSTER ANALYSIS OF TSUNAMI MARIGRAMS
Gerald T. Hebenstreit (SAIC, Ocean Systems Group,1710 Goodridge Drive, McLean, Virginia, USA, 22102,703-827-4975,gerald.t.hebenstreit@cpmx.saic.com)
Teletsunamis and some local tsunamis on long coastlines are often marked by surprising differences in the intensity of tsunamis seen in marigrams. One section of a coastline may receive clearly apparent waves, while another section close by may recieve much less apparent effects. Cluster analysis is a tool which uses the characteristics of the data to delineate patterns that may not be apparent to visual observation. The analysis makes no assumptions about the data other than that the various pieces can be directly compared. Marigrams from two tsunamis - Chile 1985 and Andreanof Islands 1996 - were examined to identify possible systematic (as opposed to local) patterns in the dispersion of tsunami intensity on coastlines.
Friday 30 July PM
Presiding Chairs: F. Gonzalez (Pacific Marine Environmental Lab, Seattle, USA)
Y. Nishimura (Inst. of Seismology & Volcanology, Hokkaido University, Japan)
physical and numerical modelling
JSS42/W/06-B5 1400
TSUNAMI RUNUP INTERACTIONS WITH A TEST STRUCTURE
Harry YEH, Halldor Arnarson, and Catherine Petroff (all at Department of Civil & Environmental Engineering, University of Washington, Seattle, WA 98195-2700, USA, email: harryeh@u.washington.edu); Razvan Bidoae and Peter Raad (both at Department of Mechanical Engineering, Southern Methodist University , Dallas, TX 75275-0337, USA, email: peter@seas.smu.edu)
The hydrodynamics of local tsunami runup interactions with a structure were investigated both experimentally and numerically. A tall, vertical column with a square footprint was used as a model of the structure, and the tsunami runup was assumed to take the form of a uniform turbulent bore. In the laboratory, the impact and transient load forces on the column were measured with a load cell, while detailed temporal and spatial variations of the tsunami runup profiles were obtained with a laser-induced fluorescent technique. Furthermore, water velocities at various locations around the column were measured with a laser Doppler anemometer. The numerical simulations were obtained by the use of the recently developed three-dimensional surface marker and micro cell (3DSMMC) technique, which is a fully three-dimensional finite-volume approach that treats the free-surface dynamics by the use of Lagrangian surface markers defined in more highly resolved surface cells. The transient numerical results are in good agreement with the laboratory results in terms of the net force exerted on the column, the water-surface variations, and the flow velocities around the column. The favorable agreement of the numerical model with the laboratory data demonstrates that the numerical capability can be used with confidence in practical field applications to assess local effects of tsunami runup.
JSS42/E/21-B5 1420
CHARACTERISTICS OF ON-SHELF TSUNAMIS AND THE ACCURACY OF THE NUMERICAL SIMULATION
Shun-ichi KOSHIMURA (DCRC, School of engineering, Tohoku University, Aoba 06, Sendai 980-8579, JAPAN, email: kossy@tsunami2.civil.tohoku.ac.jp), Fumihiko Imamura ( DCRC, School of engineering, Tohoku University, Aoba 06, Sendai 980-8579, JAPAN,
(email: imamura@tsunami2.civil.tohoku.ac.jp), Nobuo Shuto (Faculty of Policy Studies, Iwate Prefectural University, Morioka, Japan, (email: shuto@poly.iwate-pu.ac.jp)
The waveform and amplitude of obliquely incident tsunamis on a slope such as a continental shelf vary according to the conditions of incidence. However, there are cases that numerical simulation can not predict the characteristics of these on-shelf tsunamis with sufficient accuracy. Especially, if the tsunami is incident almost parallel to the slope, numerical result differ significantly depending on the grid size because of the error associated with the refraction of shallow water waves. So far, there is no criterion for the option of appropriate grid size for the numerical computation except for the specific case such as tsunamis around islands proposed by Fujima et al. (1998).
In the present study,the condition to obtain the accurate result in the numerical simulation of on-shelf tsunamis are discussed by taking the following procedure.
(1) We derive the analytical solutions of on-shelf tsunamis which is obliquely incident on a constant slope with arbitrary waveform.
(2) Based on the solutions obtained,the fundamental characteristics of on-shelf tsunamis are discussed for various types of incident waves.
(3) The numerical solutions of on-shelf tsunamis by the Leap-Frog finite difference method with a variety of grid sizes are compared with the analytical ones obtained by the present study,and the magnitudes of the errors are examined.
(4) The criterion of the grid size to obtain the sufficiently accurate result in the numerical simulation of on-shelf tsunamis is proposed.
JSS42/W/26-B5 1440
THE TSUNAMI COMMUNITY MODELING ACTIVITY (TCMA)
F.I. GONZALEZ, E.N. Bernard and S. Hankin (all at NOAA’s Pacific Marine Environmental Laboratory, 7600 Sand Point Way, NE, Seattle, WA, 98115)
Tsunami modeling is a critical community activity that contributes to the overriding goal of saving lives and property. In particular, our community conducts R&D that is focused on model improvement and the development of effective tools for hazard mitigation. A Tsunami Community Modeling Activity (TCMA) is a way to facilitate this R&D process. Community models are well-established as valuable tools in other scientific communities. Good examples are the ocean circulation models maintained at the Institute of Marine and Coastal Sciences at Rutgers University, the Princeton Ocean Model (POM) maintained by Princeton’s Program in Atmospheric & Oceanic Sciences, and the Australian Community Ocean Model (ACOM) maintained by the CSIRO Division of Marine Research in Hobart, Tasmania. The value of a community model rests on the following assumption. If easier and more widespread access to models and associated databases is provided, then the R&D to improve model science and develop more effective applications will be accelerated by broadening the user base and providing a mechanism to incorporate and document subsequent innovations and advances. The TCMA will provide a much-needed focal point for the tsunami research community to improve model physics and numerical techniques. Multiple users of the same model and databases will create regular, systematic, and invaluable feedback that will be exploited to improve the models and provide a high degree of quality control on both the model and related databases. A "community memory" will be created and institutionalized, so that important advances are not lost, redundancy and "re-inventing the wheel" is minimized, and the next generation of tsunami modelers will not be obliged to "start from scratch." Finally, and perhaps most importantly from a practical point of view, an optimal environment will be created for the systematic development and improvement of hazard mitigation products. A first step in developing an effective TCMA and enhancing the tsunami modeling R&D environment is the implementation of a distributed facility for the analysis and comparison of tsunami simulations (FACTS) that will allow global, transparent sharing of databases and model results that are resident on servers of participating institutions. FACTS will exploit the World Wide Web and additional existing technologies: Live Access Server software to seamlessly link each server; the FERRET applications package for manipulation, comparison, and analysis capabilities; the NSF/NASA supported Distributed Ocean Data System (DODS) from the University of Rhode Island for networked scientific data communications. Users of participating institutions will access and interact with a fused database that will appear to reside locally but, in fact, be distributed globally. As a primary tool of the Tsunami Community Modeling Activity (TCMA) this capability will expand access to model results and associated databases and accelerate the improvement of tsunami model physics and applications.
JSS42/W/07-B5 1500
OFFSHORE FORECASTING OF ALASKIAN TSUNAMIS IN HAWAII
V.V. TITOV, H.O. Mofjeld, F.I. Gonzalez and J.C. Newman (all at NOAA/PMEL, JISAO/UW, 7600 Sand Point Way NE, Bldg.3, Seattle, WA, 98115, USA, e-mail: titov@pmel.noaa.gov)
This study is an R&D activity conducted in an effort to develop tsunami forecasting tools for the Pacific Disaster Center (PDC). The research includes analytical and numerical sensitivity
studies of tsunami wave characteristics offshore of Hawaii, for ranges of earthquake source parameters in the Alaska-Aleutian Subduction Zone (AASZ). The results demonstrate that, for a fixed location and magnitude, there is very weak dependence of the offshore tsunami characteristics on a reasonable range of other AASZ earthquake parameters. This sensitivity study provided guidance for the construction of a database consisting of pre-computed tsunami scenarios. The simulation results are stored as an online database with a WWW interface. A database user can very quickly obtain a model prediction of the tsunami offshore wave heights at chosen locations, for a wide variety of AASZ earthquake scenarios. Direct tsunami observations can be combined with the forecast database to improve the accuracy of the tsunami forecast methodology.Potentially, the best approach to forecasting the impact of a particular tsunami event on a specific coastal site is to perform real-time tsunami model simulations that include real-time data assimilation. As a community, we should work toward this difficult "Holy Grail" of tsunami forecasting. However insufficient real-time earthquake data and a lack of confidence in tsunami generation and inundation modeling, presently render the implementation of such an operational capability imprudent.Fortunately, however, existing numerical models are good enough to provide useful guidance, if exercised with care by an experienced tsunami modeler. Event- and site-specific inundation estimates can be computed far in advance of the earthquake, thoroughly tested and scrutinized for reasonableness and sensitivity to errors, then stored as a database. When an event occurs, the appropriate pre-computed results can be recalled, modified by data assimilation schemes that utilize a user-friendly interface to incorporate real-time tsunami measurements, and then made available to aid hazard assessment and evacuation decision-making.
JSS42/W/09-B5 1520
TSUNAMI SCOUR MECHANISMS AROUND A CYLINDER
Harry YEH (Department of Civil & Environmental Engineering, University of Washington, Seattle, WA 98195-2700, USA, email: harryeh@u.washington.edu); nji Sato and Norifumi Kato (both at Public Works Research Institute, 1 Asahi, Tsukuba, Japan, email: sato@pwri.go.jp)
It is known that tsunamis cause substantial erosion and scours on shore. The scour mechanisms around on-shore structures are expected to be different from the present understanding of bridge or pier type scour processes in a river or coastal environment. Flows associated with tsunami runup are not steady nor uniform, and the scouring occurs in a short duration, less than half an hour. In order to understand fundamental mechanisms of the scouring, a series of experiments was performed in a 135 m long, 2 m wide, 5 m deep sediment tank; this large facility is necessary because the substantial scale effects are anticipated for the sediment motions. The maximum stroke of the wave paddle is 2.4 m and it was tested and proven to generate, at least, a clean solitary wave of 40 cm high in 3 m deep of water. The circular column is used as a structure model, and is placed upright at three different locations: 1) 4m on-shore from the shoreline, 2) at the shoreline, and 3) 4 m offshore from the shoreline. The cylinder is 50 cm in diameter, made of 1 cm thick Plexiglas. Because of its transparent cylinder wall, the scouring process is conveniently recorded with three miniature CCD video cameras equipped inside of the cylinder, which cover more than 180 degrees of the circumference view. Pore pressure transducers are placed at four locations around the cylinder in the sediment depths of 10 cm, 20 cm and 30 cm. Also placed are the capacitance type wave gages and an electromagnetic flow meter. It was found that the maximum scour occurs during the tsunami drawdown stage and is approximately twice the depth of the final scour configuration. Based on the pore pressure data as well as the video images, sediments around the cylinder appear to become liquefied during the drawdown process.
JSS42/W/17-B5 1600
NUMERICAL SIMULATION OF 1998 PNG TSUNAMI ON A NEW BATHMETRY BY KAIREI
Masafumi MATSUYAMA, Central Research Institute of Electric Power Industry, 1646 Abiko, Abiko-shi, Chiba-ken, Japan.
The 1998 Papua New Guinea (PNG) tsunami caused devastating damage and casualty along the north coast of PNG, in particular near Sissano lagoon. Surveyed tsunami effects were much greater than the estimated based on the available seismic and bathymetry data. The similar discrepancy is often attributed to the phenomenon so called "the tsunami earthquake." This PNG event does not fit to the characteristics of the tsunami earthquake. The survey also found that only narrow coastal area, approximately 30 km long, around Sissano lagoon was significantly affected by the tsunami; the tsunami effects drastically decrease away from the region. In order to explain the discrepancies, Japan Marine Science and Technology Center(JAMSTEC) and South Pacific Applied Geo-Science Commission(SOPEC) conducted the marine geophysics survey in January, 1999. During the cruise, detailed bathymetry data and sub-bottom profiles are obtained in the offshore area adjacent to Sissano lagoon. With the obtained bathymetry data and the newly discovered fault location, Numerical simulations for the PNG tsunami were performed using the standard seismic model. It was found that significant tsunami-energy focusing toward Sissano lagoon is caused by the bathymetry effects. Furthermore, a possible submarine landslide site was identified, and the initial condition for the landslide generated tsunami is estimated. Even the landslide source area is small (5 km x 7 km), the computed runup heights around Sissano lagoon is significant (approximately one half the measured values). More importantly, the runup height distribution along the shore is qualitatively in good agreement with the simulation. The results demonstrate that the acurate bathymetry and the source location are critical for the predictions for tsumani runup.
JSS42/W/01-B5 1620
NON-REFLECTIVE OPEN BOUNDARY CONDITION FOR TSUNAMI NUMERICAL SIMULATION
Tomoyuki Takahashi (DRS, DPRI, Kyoto University, Uji, Kyoto 611-0011, Japan, email tomo@drs.dpri.kyoto-u.ac.jp)Hiroya Ishihara (Fac. of Engineering, Kansai University, Suita, Osaka 564-8680, Japan, email gj50010@edu.kansai-u.ac.jp)Yoshiaki Kawata (DRS, DPRI, Kyoto University, Uji, Kyoto 611-0011, Japan, email kawata@drs.dpri.kyoto-u.ac.jp)
In order to carry out numerical simulation stably and with high precision, the setup of suitable boundary condition is indispensable. The boundary condition in tsunami numerical simulation is divided into the land side and the offing side. As for the boundary condition on the land side, there are a vertical wall condition and a runup condition, and a lot of studies about it have been made so far. Especially, much knowledge about the runup condition has been obtained theoretically and experimentally. On the other hand, as for the boundary condition on the offing side, the method of the long wave approximation, the characteristics method and the method of the virtual non-reflecting plane boundary are used well. However, they have been used without sufficient examination about their accuracy and stability. Although the free penetration of the progressive wave generated inside the computational area needs to be carried out, it turns out that a reflected wave occurs according to open boundary conditions. The reflected wave returns to the computational area and will cause instability and lower accuracy. Especially, the reflected wave occurs mostly by the case when the progressive wave diagonally approach to the boundary and the case where depth is shallow or bottom slope is large in the vicinity of the boundary. To reduce the influence of the reflected wave, the computational area is taken widely. But this makes a problem that the amount of the computation and the computation time become large.Moreover, it is abundant to compute without investigating what reflected wave has occurred at the open boundary. Then, by this research, each non-reflected open boundary condition was compared about the various depth, bottom slope and incident angle, and the applicable condition in tsunami numerical simulation was examined.
JSS42/E/09-B5 1640
COMPUTATIONAL EXPERIMENTS FOR SIMULATION OF TSUNAMI EFFECTS NEAR SOUTH KURIL ISLANDS
TATYANA IVELSKAYA, Sakhalin Tsunami Warning Center, 78, Zapadnaya str., Yuzhno-Sakhalinsk, 693000, Russia, email: TWC@Sakhalin.ru. Vasily hramushin, Special Research Bureau for Automation of Marine Researches, Far East Branch of Russian Academy of Sciences, 25 Gorky str., Yuzhno-Sakhalinsk, 693013, Russia, email: Khram@Sakhalin.ru;
The tsunamis and storm surges are dangerous sea phenomena, however, these admits the possibilities to organize the actions for the evacuation of peoples, will have earliness from a few minutes or a next hours.
The decision of a problem of numerical simulation with the purpose of operative forecasting of wave processes in a coastal zone for specific tsunami is described in this work.
As initial data for modeling a tsunami parameters of earthquakes, and also the information on displays a tsunami on coast are used the experience of work on realization of similar computational experiments at modeling of real tsunami has shown good results, which give sufficient conformity with records a tsunami.
Presiding Chairs: F. Gonzalez (Pacific Marine Environmental Lab., Seattle,USA)
Y. Nishimura (Inst. of Seismology & Volcanology, Hokaido University, Japan)
JSS42/W/28-B5 Poster 1700-01
TSUNAMIS BY VOLCANIC EXPLOSIONS: DEPOSITS AND RUNUP DATA FROM THE 1996 ERUPTION IN KARYMSKOYE LAKE, KAMCHATKA, RUSSIA
Alexander BELOUSOV and Marina Belousova (Institute of Volcanic Geology and Geochemistry, Petropavlovsk-Kamchatsky, Russia, email: Belousov@g23.relcom.ru) Barry Voight (The Pennsylvania State University, USA, email: voight@ems.psu.edu)
The 1996 subaquatic explosive eruption in Karymskoye lake (4 km across, maximal depth 70 m) generated multiple tsunamis. The steep northern shore of the lake adjacent to the crater was violently eroded by the tsunamis, with all plants and soil more than 1.5 m thick stripped off, and poorly consolidated bedrock exposed. Along the rest of the shoreline, more than 1.3 km from the crater, the tsunamis carved new cliffs up to 2 - 3 m high, eroded the upper layer of frozen soil up to 50 cm thick, and inundated adjacent areas as shown by vegetation removal, or destruction of alder bushes. In the band of devastation the tsunamis deposited several discontinuous, finely laminated layers of sand and gravel up to 35 cm thick, with scattered pebbles, fragments of plants and clots of soil. The distinctive band of devastation, and the tsunami deposits, allowed us to measure the runup height of the tsunamis at 24 points around the lake. These data enable determination of a law of attenuation of runup (wave) height for "explosive" tsunamis, which is compared with theoretical modelling. For the proximal zone, to radial distances (r) up to 1.3 km, the runup height (R) shows rapid attenuation (from >30m to 8 m) with distance as log R = -0.56 log[r] + 5.8. For the distal zone, r > 1.3 km, R decays more slowly (from 8m to 3m) as log R= -1.98 log[r] + 16.3. Rapid decay in proximal zone suggests that in the near field of the explosion the tsunami propagate as a collapsing wave with discontinuous change in height. The break-in-slope of the runup plot at 1.3 km suggests that the tsunami propagate further as a decaying one-dimensional wave in a channel of approximately constant width.
JSS42/W/13-B5 Poster 1700-02
LOCAL TSUNAMIS GENERATED BY STORM WAVES
GERARD J. FRYER (Hawaii Institute of Geophysics & Planetology, University of Hawaii at Manoa, 2525 Correa Rd, Honolulu, HI 96822, USA, email: gerard@hawaii.edu)
Many orphan tsunamis (events without identifiable sources) have occurred in the two centuries of recorded Hawaiian history. Such "tsunamis" have been commoner in the winter months, when surf can reach 10 m, so they are dismissed in the catalogs as "waves of meteorological origin." Unexplained, however, is occasional inundation far larger than expected even from 10-m surf. Inundation and damage have been so extensive that a true tsunami is the only possible explanation. Without a local earthquake, such an event is assumed to be "a tsunami from a distant but unknown source." The assumption is reasonable: flooding reports are typically from north-facing coastlines, and to the north lie the sparsely-populated Aleutian Islands, where tsunamigenic earthquakes (until the modern era of worldwide instrumentation) might pass unreported. It is now clear, however, that at least three events (in 1860, 1878, and 1903) were not from distant sources. Each of the major Hawaiian Islands has a broad, north-facing bay prone to extreme tsunami runup: Hilo on the Big Island, Kahului on Maui, Waialua on Oahu, and Hanalei on Kauai. A teletsunami producing high runup in one of these bays invariably produces high runup in all. The orphan events, however, affect only Kahului and Waialua Bays, the adjacent coastlines of Maui and Oahu, and the north coast of Molokai. Since the orphans occur only when surf is exceptionally high, it seems probable that by pumping up pore pressures, storm-generated waves drive bottom sediments to failure. The resulting landslide generates a tsunami whose effects are added to the already-large waves. The limited geographic extent of the anomalous flooding clearly identifies the locus of the landslides: the north slope of Molokai.
JSS42/W/16-B5 Poster 1700-03
PALEO-TSUNAMIS NOVEL FINDINGS FROM NORTHWESTERN EUROPE
Nils-Axel Mörner Paleogeophysics & Geodynamics, S-10691 Stockholm, Sweden,
email: morner@pog.su.se
Tsunamis are able to cause disasterous damage to coasts and coastal habitation. They are generated by coastal and submarine vertical fault movements or huge submarine slides. It is a well known phenomenon in the Pacific region. The Lisbon 1755 event is classic. In recent years we have also started to appriciate their imprints in our paleorecords. Past events are documented in the Mediterranian region. More surprising are the records from NW Europe. First we learned about the tsunami of the 7200 BP Storegga submarine slide outside Norway and its effect along the coasts of
Norway and Scotland. The latest news come from Sweden where we have been able to document a major earthquake (generating liquefactoin over 320 km) occurring in the autumn of varve 10,430 BP and a subsequent event occurring in varve 9663 BP (generating seismites over 210 km). Both events generated tsunamis; the first one washed the strait between the Baltic and the North Sea free of ice so that marine water suddenly could invade the Baltic basin (by this creating the Yoldia Sea stage sensu stricto). The second event set up a wave recorded in the sea bed sedimentology and ice marginal shore morphology. A triple sedimentary sequence was recorded indicating sucking motions towards the epicenter, a deformational wave away from the epicenter, and a strong back-swash wave depositing high energy accumulations, deforming beds underneath and setting up a turbidite which is recorded in the varve chronology over an area of 210 km (in varve -424 = varve 9663 BP). Other deglacial events are under investigation. An event of more local character has been recorded in a lake system and dated at about 3500 BP. This gives a totally new picture of the long-term seismic hazard and actual occurrence of tsunami events in Sweden and surrounding regions.
JSS42/E/14-B5 Poster 1700-04
IDENTIFICATION OF TSUNAMI DEPOSIT AND ITS APPLICATION FOR EVALUATING HISTORIC TSUNAMI HAZARDS IN HOKKAIDO, NORTHERN JAPAN: A REVIEW
YUICHI NISHIMURA (Institute of Seismology and Volcanology, Hokkaido University, Sobetsu 59, Usu-gun, Hokkaido 052-0103, Japan, nishi@eos.hokudai.ac.jp) Naomichi Miyaji (Shizuoka Prefectural Agricultural Experiment Station, Godo, Hamaoka-cho, Ogasa-gun, Shizuoka 437-1613, Japan, miyaji@ss.sand.agri-exp.pref.shizuoka.jp)
We have studied geological evidence for recent and historic tsunamis which a ttacked Hokkaido, northern Japan. At various kind of land beside the beach, we made pits to take core samples or excavated trenches, and identified tsunami deposits overlies volcanic ash soil, old dune and peat. These tsunami deposits are found to show some common features: (1) Deposit thickness and mean grain sizes decrease with distance from the sea. (2) The thickness and lithofacies vary according to the original surface undulation. (3) Graded bedding reflecting tsunami runup and backwash is present in thick deposits. It was possible to find coarse tsunami deposits sandwiched by fine dune sands, but hard to specify the original tsunami event except that known volcanic tephra lies closely. On the other hand, the age could be estimated for tsunami deposits in homogeneous peat deposit based on growth rate of the peat. When the upper boundary of a tsunami deposit is traced, we can estimate the minimum height of its corresponding tsunami runup. For the 1640 tsunami caused by an eruption of Hokkaido Komagatake, which killed more than 700 people around Uchiura Bay, we examined documentary descriptions, made numerical computations and traced tsunami deposits. It is useful to combine these independently obtained data to estimate height distribution of a historic tsunami and evaluate its disaster.
JSS42/E/23-B5 Poster 1700-05
1952 NORTH KURIL TSUNAMI: A NEW DATA FROM ARCHIVES
Victor KAISTRENKO and Valentina Sedaeva (both at Institute of Marine Geology and Geophysics, Far Eastern Division of the Russian Academy of Sciences; Nauki St., Yuzhno-Sakhalinsk, 693002 Russia, email: tsunami@sakhmail.sakhalin.ru)
During the long time until 1990th most of data related to the catastrophic 1952 tsunami were secret and last years these data can be investigated. Most essential information is contained in the survey report made by Hydrographic Service of the Russian Navy. Several reports of different commissions were found in Sakhalin archives. Interesting information was received from eye-witness of this event.
Navy’s survey team started to work in Severo-Kurilsk next day after the tsunami. Their report contains 10 maps of flooding areas around the settlements on North Kuril Islands and good description of damages due to tsunami and many photos.
The reports in Sakhalin archives give many descriptions of tsunami behavior and contain the lists of damages and lost persons. So, the Okeanskaya settlement was destroyed almost fully by 20-meter tsunami and 460 persons were lost, i.e. about 50 % of population, and 1753 persons were killed by tsunami in this region.
Several descriptions and photos were given by eye-witness. All these data are planned to publish as a book.
JSS42/E/13-B5 Poster 1700-06
RUNUP OF TSUNAMI WAVES ON A VERTICAL WALL AND GENTLE BEACH IN A BASIN OF COMPLEX TOPOGRAPHY
Efim PELINOVSKY (Laboratory of Hydrophysics and Nonlinear Acoustics, Institute of Applied Physics, 46 Uljanov Str., Nizhny Novgorod, 603600, Russia, email enpeli@appl.sci-nnov.ru) Elena Troshina (Geophysical Institute, University of Alaska Fairbanks, Fairbanks, AK 99775, USA, email etroshin@images.alaska.edu) Vladimir Golinko and Nataly Osipenko (both at Department of Applied Mathematics, Nizhny Novgorod Technical University, 24 Minin Str., Nizhny Novgorod, 603600, Russia)
Being able to estimate the flooding area of the coastal zone caused by the tsunami waves is essential for tsunami hazard mitigation. Much progress has been made in applying the one-dimensional nonlinear shallow water theory to the runup height calculation. It was based on the analytical results of the well-known paper by Carrier and Greenspan. This method has been used by many authors to calculate the tsunami wave runup. The exact estimation of the size of the area flooded by tsunami waves requires the solution of two-dimensional shallow-water equations, taking into account the geometry of the coastal line (bays, straits, estuaries, etc.). We study runup of tsunami waves on a vertical wall and gentle beach in a bay of different cross-section using shallow water equations. The characteristics of the nonlinear deformed waves in bays of different cross-section are investigated. The nonlinearity causes the runup height to increase. Several examples of beach topography are considered. The criteria of the wave breaking are discussed.
JSS42/E/11-B5 Poster 1700-07
NUMERICAL SIMULATION OF POTENTIAL SUBMARINE SLOPE FAILURES AND ASSOCIATED TSUNAMIS ALONG THE COAST OF BRITISH COLUMBIA
Alexander B. RABINOVICH, Evgueni A. Kulikov (both at Tsunami Center, Shirshov Institute of Oceanology, 36 Nakhimovsky Prosp., Moscow 117851, Russia, e-mail abr@tsucen.msk.ru)
Richard E. Thomson (Institute of Ocean Sciences, 9860 W.Saanich Rd., Sidney, BC, V8L 4B2, Canada, e-mail thomsonr@dfo-mpo.gc.ca) Brian D. Bornhold, and Isaac V. Fine (both at International Tsunami Research, Inc. 11321 Chalet Rd., Sidney, BC, V8L 5M1, e-mail itri@ii.ca)
Studies in coastal areas of British Columbia indicate the high instability of deltaic and nearshore sediments in many areas. Buildings, coastal facilities, pipelines, telecommunication and electrical transmission cable lines are at significant risk of direct damage from subaerial and submarine landslides. Tsunamis generated by these events probably pose an even greater threat in terms of damage and loss of life. We concentrate our attention on two areas of potentially high risk of slope failure: (1) Malaspina Strait, a narrow (5 to 10 km) channel separating Texada Island from the mainland of British Columbia; and (2) the Fraser River delta front in the southern portion of the Strait of Georgia. The main reasons for our interest in these areas are: (1) large volumes of unconsolidated sediments, accumulated there; (2) high risk of instability of these sediments; and (3) location of significant coastal infrastructure, which would be at risk should a failure and tsunami occur. We have developed two types of numerical model to simulate landslide-generated tsunamis: (1) viscous flow, and (2) rigid body. These models bracket the true response of the sea surface to the landslide, with the viscous slide model underestimating the wave response and the rigid body model overestimating the response. Parameters of slide bodies were defined from existing geological and geophysical investigations, and direct video observations. In the case of Malaspina Strait, the failure of the northern lobe of a perched sediment mass on the slope off eastern Texada Island, yielded trough-to-crest waves from 7 to 9 m. Two cases were…
JSS42/E/25-B5 Poster 1700-08
ANALYSIS OF THE ALGERIAN TSUNAMIS ON THE MEDITERANEAN COAST OF SPAIN
Alexander B. Rabinovich (IOS, Sidney BC, Canada, Email: RabinovichA@pac.dfo-mpo.gc.ca), Sebastian MONSERRAT (CSIC-UIB, Palma de Malorca, Spain, Email: dfssmt4@clust.uib.es)
Although tsunamis in the Mediterranean are not so frequent and powerful as in the Pacific, they still present serious threat for the coastal areas. The major risk for the coast of Spain is related to the earthquakes in the northern part of Algeria. The Orleansville Earthquake (1954) with magnitude M = 6.6 and El Asnam Earthquake (1980) with M = 7.3 generated significant tsunamis observed all along the Mediterranean coast of Spain and in the Balearic Islands. These tsunamis were recorded by several tide gauges (Malaga, Ceuta, Alicante, Almeria, Algeciras, etc.). The number of tsunami records in the Mediterranean basin is extremely limited, so examination of these two tsunamis simultaneously measured by the instruments situated in various sites of the Western Mediterranean is of high scientific and applied interest. The method recently proposed by Rabinovich (1997) to separate the influence of source and topography in tsunami spectra was to analyze the data. The main attention was paid to Alicante records, which were of high quality, especially to the 1980 case when the corresponding tsunami was recorded by two gauges (inside and outside of the harbor). The response characteristics of the harbors were constructed. Several cases of atmospherically generated seiches have been also examined and the results have been compared with results of tsunami analysis.
JSS42/W/28-B5 Poster 1700-09
THE AD1650 MT. COLUMBO (THERA ISLAND) ERUPTION AND TSUNAMI, AEGEAN SEA, GREKCE
Dale Dominey-Howes (Coventry Centre for Disaster Management, School of the Built Environment, Coventry University, Coventry, CV1 5FB, UK, Kmail:apx124(R)coventry.ac.uk)
This presentation reviews evidence for the eruption and tsunami reported to have occurred in AD1650 in the area of Mt. Columbo, Thera Island, Greece. The tsunami is believed to have been generated as consequence of the eruption of Mt. Columbo 6.5 km NE of Thera Island. Historical documents state that the tsunami flooded up to 2 miles inland and destroyed many engineered structures. Lithostratigraphic evidence from one abandoned trench and two trench excavations close to sea level in the villages of Kamari and Perissa respectively, which lie well within the reported inundation zone of the tsunami. The results presented show that _® marine- (tsunami) sediments were deposited at these locations. Alternative hypotheses of discontinuous sediment deposition and overestimation of the event magnitude are considered to explain the observations presented here. The data may have important implications for the development of hazard zone maps, risk assessment, vulnerability reduction and for emergency management officials.
JSS42/E/08-B5 Poster 1700-10
MITIGATION OF TSUNAMI RISK FOR THE CITY OF SUVA, FIJI
Jack RYNN (Centre for Earthquake Research in Australia, PO Box 276, Indooroopilly, Brisbane, Queensland 4068 Australia, email sally.brown@uq.net.au) Gajendra Prasad (Mineral Resources Department, Government of the Republic of Fiji, Private Mail Bag, Suva, Fiji, email gajen@mrd.gov.fj); Atu Kaloumaira (Mitigation Advisor, South Pacific Disaster Reduction Management Office, c/o UNDP, Private Mail Bag, Suva, Fiji, email atu@sopac.com.org)
As part of the UNDHA - South Pacific Programme Office "South Pacific Disaster Reduction Programme", within the auspices of Pacific Region IDNDR and the 1994 Yokohama Statement, the "Suva Earthquake Risk Management Scenario Pilot Project" (SERMP) was facilitated 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 (ML 6.5) and tsunami, which devastated the city, its industrial and harbour facilities. A specific 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 commentaries, tsunami parameters and possible damage situations. It was concluded that a significant risk of 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. Fiji is now developing a regional tsunami warning system.
JSS42/W/05-B5 Poster 1700-11
TSUNAMI WAVE SCATTERING IN THE NORTH PACIFIC
H.O. Mofjeld, V.V. Titov, F.I. González and J.C. Newman NOAA/Pacific Marine Environmental Laboratory Seattle, WA, USA
A theoretical study is being carried out to understand how escarpments, ridges and seamounts
affect deep-water tsunami propagation in the Pacific Ocean. The study is also designed to determine the accuracy and resolution of bottom topography that are needed in numerical models to accurately simulate the effects of small-scale (*100 km) topographic features on tsunami propagation. The initial focus of the work is on tsunamis that are generated in the Alaska/Aleutian Subduction Zone (AASZ) and propagate southward to Hawaii. Analytic theory shows that main effects of these features is to scatter energy from the tsunami waves. The amount of scattering depends on the heights of the features relative to the total depth, their spatial extent compared with tsunami wavelengths, and their orientation relative to the direction of wave propagation. 2-D wavelet analyses of the Sandwell/Smith topography (Topo 6.2) are used to identify the spatial scales and locations of scattering features in the North Pacific. Numerical simulations based on the MOST model (with and without small-scale topography) show that deep-water scattering produces only a small amplitude decrease in the first waves of AASZ tsunamis propagating to Hawaii (a few percent in energy). This is compared with the geometric spreading and broad-scale refraction that occurs when the tsunami waves propagate southward from their sources. The primary reason for relatively little deep-water scattering is that the tsunami-scale topographic features between Alaska and Hawaii do not extend vertically over a substantial fraction of the total water depth. However, significant focusing of tsunami waves does occur near Hawaii; this is due to the seamount field north of Kauai, the Emperor Seamount Chain and the Hawaiian Ridge. While deep-water scattering has only a minor influence on the first waves in AASZ tsunamis, it does contribute to the complicated temporal patterns of the waves that occur later in Pacific-wide tsunamis. Scattering processes are much stronger on the upper continental slope and shelf; hence there is a need for accurate topography in these regions. Future work will extend the analysis to shallower water, to other source regions (e.g., Kamchatka, Kuril/Japan and Cascadia) and to other impact regions, especially the U.S. West Coast.
JSS42/E/03-B5 Poster 1700-12
THE TSUNAMI OF 1722.12.27, ALGARVE–PORTUGAL
M.A. Baptista (Instituto Superior de Engenharia de Lisboa, Lisboa Portugal, email: mariaana@fc.ul.pt, fax + 3511 3953327) C.Lemos (Instituto Hidrográfico – Portugal) J.M.Miranda (Centro de Geofísica da Universidade de Lisboa)
The earthquake of 1772.12.27 generated a tsunami that affect the southern coast of Portugal, mainly in the area of Tavira. The maximum earthquake intensity was X and the magnitude 7.5 (given by Instituto Geografico Nacional, Madrid Spain). The historical data on this event is rather scarce. However, the earthquake was felt along the whole south Portuguese coast and in Faro and Tavira a rather strong river withdrawal were observed (Mendonça, 1758). The epicentre of the earthquake was located, from isoseismal analysis offshore the south coast of Portugal but the precise location is largely unknown. The aim of this work is the compilation of all geophysical data, and the use of tsunami hydrodynamic modeling to evaluate among a set of fault candidates which is the most probable source of the Tavira earthquake.
This work was financially supported by project RIMAR.
JSS42/E/07-B5 Poster 1700-13
TSUNAMI RUN-UP
C. CORELA and L.A. Mendes Victor( both at ICTE, Instituto de Ciências da Terra e do Espaço, Rua da Escola Politécnica 58, 1200 Lisboa, Portugal,ccorela@ correio.fc.ul.pt )
The description of the Tsunami run-up is of greastest importance for Tsunami zonation and evaluation of Tsunami Hazard.The difficulties arising here are evident: the complexity of coastal zone morphology and the variety of underlying surfaces changing due to their interaction with the water flows caused by Tsunamis, the possibility of waves breaking and competition of nonlinear and disperson effects. A numerical method based on the upstream formulation for the advective term is applied to compute the run-up.
JSS42/E/19-B5 Poster 1700-14
THE EXPERIMENTAL ITALIAN TSUNAMI WARNING SYSTEM AT AUGUSTA (EASTERN SICILY)
Alessandra Maramai, Alessandro Piscini and Giuseppe D Anna (Istituto Nazionale di Geofisica (ING), Via di Vigna Murata 605, Rome - Italy
In the frame of the GITEC-TWO (Genesis and Impact of Tsunamis on the European Coasts) project, in the Eastern Sicily coasts a "pilot" station of the first Italian Tsunami Warning System (TWS) has been installed. Most of the tsunamis affecting the Italian coasts are generated by local earthquakes with focal region close to the coast, both in land or in the sea. As regards the Eastern Sicily coast, usually the historical tsunamis are characterized by a first withdrawal of the sea, followed within a few minutes by a more or less violent flooding. Taking into account this characteristics, the experimental TWS is conceived for the coastlines affected by tsunamis locally generated, where the waves invest the coast just a few minutes after the earthquake occurrence, so what is needed is an immediate detection of the generated tsunamis. In this context, the Italian TWS is an integrated system designed for the simultaneous acquisition of the seismic signal and the sea level oscillations, for processing both signals, for detecting dangerous conditions and for issuing an alarm message to specific units of the local Civil Protection service. As regards the sea level data, a pressure gauge has been installed in a suitable wharf close to the Augusta harbour. The water level sensor is connected to an automatic station to collect and process the sea level data and many other intersting parameters and to send the data (via radio link) to a TWS "control center". As for the seismic data, the three components (S-13) Augusta seismic station is used and a spectral analysis of the background seismic noise has been done. The TWS control center is equipped with two PC units, connected in LAN network, to collect and process in real time the seismic and the sea level data and to identify anomalous conditions in the two signals, in order to send a useful alarm message. One PC unit is responsible for the seismic signal acquisition and processing, working in a Windows NT environment. The second unit, in UNIX environment, is for both the seismic and the sea level data and it contains specific routines for the detection of sea level anomalous oscillation and for the alarm message transmission code. This unit is also equipped with a specific software for the data management, that is data automatic acquisition, real time data validation, data collection displaying, monitoring of the station activity, etc. The station has been working since July 1998 and at present some preliminary analysis of the registered data sets has been done.
JSS42/W/11-B5 Poster 1700-15
TSUNAMI INUNDATION MODELING FOR SELECTED COMMUNITIES IN KODIAK ISLAND, ALASKA
Elena N. Suleimani and Roger Hansen (both at Geophysical Institute, University of Alaska Fairbanks, Fairbanks, AK 99775-7320, USA, email: etroshin@images.alaska.edu)
The 28 March 1964 earthquake produced a destructive tsunami in Alaska and on the West Coast of the United States. The event showed avulnerability of Kodiak, a region with a number of relatively high populated communities and significant fishing resources. Over the last ten years there had been substantial growth of Kodiak City and other island communities, a number of engineering projects, and significant changes to the harbor. As part of an ongoing component of the National Tsunami Hazard Mitigation Program and the Alaska Science and Technology Foundation we are providing assistance to threatened communities with potential tsunami inundation assessment. Three of the Kodiak communities were identified as high-priority zones for inundation mapping.
The numerical model is based on the nonlinear shallow-water equations which are solved by a finite-difference method. The historically documented wave runup, mainly from the 1964 Alaska earthquake, is used to calibrate the model. First, the actual ground displacement of the 1964 earthquake is assumed, then, the hypothetical sources are modeled. The results will be presented using a web-based interface. They will allow us to evaluate the potential tsunami hazard for the Kodiak communities and to provide them with the tsunami inundation maps for emergency planning of evacuation routes. Finally, about 10 additional communities are planned for evaluation over the next four years as part of the federal/state mitigation program.
JSS42/W/12-B5 Poster 1700-16
DISPERSIVE TSUNAMI GENERATION BY BOTTOM DISPLACEMENTS WITH COMPLICATED TIME-SPATIAL HISTORY
M.A. NOSOV, S.V. Mironyuk
The linear potential theory was used to obtain the exact analytical solution of the 3-D problem of gravitational waves generation by bottom displacements of finite duration as double Fourier integrals over components of the wave vector. The integrals were evaluated numerically. In particular it was demonstrated that the bottom displacement should be considered as a process of a finite duration, not as an impulse one. Using some models of bottom displacements of finite duration (piston, membrane-like, traveling, with alternating signs) a certain influence of bottom displacements time-spatial history on amplitude, energy and directivity of tsunami radiation from asymmetric source was revealed. It was also established that as the distance from the source increases, the directivity of energy radiation is preserved, whereas the distribution of wave amplitude characteristics in the azimuthal angle becomes isotropic.
JSS42/E/04-B5 Poster 1700-17
SPECTRAL ANALYZIS OF OCTOBER 5, 1994 SHIKOTAN TSUNAMI RECORDS
TATYANA IVELSKAYA, Sakhalin Tsunami Warning Center, 78, Zapadnaya St. Youzno-Sakhalinsk, 693000, Russia, George Shevchenko, Institute of Marine Geology and Geophysics, Far East Branch of Russian Academy of Sciences, Nauky str., Yuzhno-Sakhalinsk, 693002, Russia, email: Tsunami@Sakhalin.ru
Spectral characteristics of wave field after strong earthquake that occurred at the October 5, 1994 in the region of Shikotan Island were analyzed. It was shown that main energy of tsunami at the South Kuril stations was concentrated in the relatively high frequency band (with periods about 21 min). Probably, it was induced by the particularities of the seismic processes in the epicentral area of underwater earthquake.
At the same time the energy of tsunami on the coast of Japan was concentrated in the relatively low frequency band (with periods about 28, 36, 48, 60 min). Probably, it was connected not only with particularities of tsunami source, but with the influence of sea bottom topography in this region.
Calculation of spectral-time diagrams shows the complicated structure of wave records that includes five splashes of intensity on the main part of stations. Very likely that these splashes were induced by aftershocks in the epicentral area of underwater earthquake. Spectral-time diagrams detected the early wave groups with wide spectra, and resonance particularities of basins selectively amplify some spectral components.
JSS42/W/19-B5 Poster 1700-18
A FAST AND SIMPLE DIAGNOSTIC METHOD FOR IDENTIFYING
TSUNAMIGENIC EARTHQUAKES
Nikolai. M. SHAPIRO, Shri Krishna Singh, and Javier Pacheco (all at Instituto de Geofisica, Universidad National, Autonoma de Mexico, Mexico D.F., Mexico, e-mail: shapiro@ollin.igeofcu.unam.mx)
An analysis of regional broadband seismograms of moderate and large subduction-zone earthquakes of Mexico shows that the earthquakes which occur near the trench are abnormally depleted in high-frequency radiation. This observation leads to a simple and fast method to assess regional tsunami potential from earthquakes which occur along the Pacific coast of Mexico. A significant advantage of the method is that a single broadband seismograph is sufficient for the purpose. The method is based on the ratio of the total energy to the high-frequency energy (between 1 and 5 Hz), ER, computed from the seismograms. For earthquakes of same seismic moment, ER is an order of magnitude greater for a source area near the trench as compared to that for areas near the coast. The same seismograms are used to compute energy magnitude, Me, which is tied to the moment magnitude, Mw. A regional tsunami may be expected along the coast if Me > 6.5 and ER corresponds to a source near the trench. If, however, ER corresponds to a near-coast source, then Me may have to be greater than about 7.3 for a tsunami of similar size to occur along the coast. The method holds promise for a fast regional tsunami warning in many Pacific basin countries which lack an adequate seismic network.
JSS42/E/21-B5 Poster 1700-19
FUNDAMENTAL STUDY ON TSUNAMIS GENERATED ON A SHELF
Koji FUJIMA (Department of Civil Engineering, National Defense Academy, Yokosuka 239-8686, Japan, email:fujima@cc.nda.ac.jp)
Based on the linear long wave theory, theoretical solution is obtained for the tsunami propagation >from an arbitral tsunami source generated on a straight beach with an uniform slope. The solution indicates that the behaviors of tsunami generated on a shelf is governed by the conditions of initial tsunami profile. The edge wave is generated considerably in the case when a tsunami source locates near a coastline. On the contrary, in the case when a tsunami source is generated in a region far from a coastline, the edge wave is not generated and tsunami energy is concentrated in a narrow area. The empirical formulas are derived which evaluate the characteristics of tsunami by those of tsunami source such as the lengths of long-axis and short-axis, the position and direction of tsunami source, and so on. Further, effect of the Colioris force is also discussed.
JSS42/E/20-B5 Poster 1700-20
THE THREE-DIMENSIONAL TSUNAMI NUMERICAL ANALYSIS WITHOUT ASSUMING THE STEP-TYPE BOTTOM BOUNDARY
Kenji MASAMURA and Koji Fujima (both at Department of Civil Engineering, National Defense Academy, 1-10-20 Hashirimizu, Yokosuka 239-8686, Japan, email: kenji@cc.nda.ac.jp, fujima@cc.nda.ac.jp) Chiaki Goto (Department of Civil Engineering, Tokai University, 1117 Kitakaname, Hiratsuka 259-1292, Japan, email: goto@keyaki.cc.u-tokai.ac.jp)
The conventional numerical analysis of Tsunami based on the shallow water theory cannot reproduce three-dimensional flow. To improve this deficiency, three-dimensional numerical model without using the long wave approximation was developed. This model can compute the propagation of non-linear dispersive waves, because the hydrostatic pressure distribution is not assumed in the model. However, when the complicated geographical features are adopted for the calculation, submarine bottom boundaries are approximated as the stair-shaped step. Therefore, accuracy of calculation tends to become worse at the tip of runup. Thus, in this study, a "permeation rate" is introduced as a new parameter in each surface of the cell. The "permeation rate" is defined by he ratio of the section area of the portion where a water particle can pass through and the area of the surface of the cell. Therefore, the bottom boundary is approximated by a slope crossing the inside of the cell. As the result, runup height can be reproduced with good precision by the present model.
JSS42/L/01-B5 Poster 1700-21
COMBINATION OF BATHYMETRY AND TOPOGRAPHY DATA FOR TSUNAMI MODELING
R.A. KAMPHAUS (NOAA/PMEL, TIME, 2115 SE OSU Drive, Newport, OR 97365, USA; e-mail: kamphaus@pmel.noaa.gov), D. Divins (NOAA/ NGDC, CIRES, 325 Broadway, Boulder, CO 80303, USA; e-mail: ddivins@ngdc.noaa.gov) and F.I. Gonzalez (NOAA/PMEL, 7600 Sand Point Way NE, Bldg.3, Seattle, WA, 98115, USA, e-mail: gonzalez@pmel.noaa.gov)
Grids that integrate bathymetric and topographic data are needed for
effective tsunami inundation modeling. Topographic data in the form of Digital Elevation Models (DEMs) have been available for the majority of the United States for several years from the US Geological Survey (USGS). Bathymetric data, on the other hand, has been more sparse and not available digitally for all areas. As more digital data becomes available through the efforts of data centers-such as NOAA's National Geophysical Data Center (NGDC) -- techniques for efficiently incorporating these data into numerical grids are being developed. Using techniques and databases developed at NGDC, the Center for Tsunami Inundation Mapping Efforts (TIME) has created several high quality gridded data sets for tsunami inundation modeling along the coast of California. The grids are assembled by gridding the NOS Hydrogaphic Database soundings, and supplemental soundings, to the same 3 arc-second (~90 m) resolution and registration as the USGS's 3-arc-second DEMs. In the United States the topographic and bathymetric data are collected in different reference datums; therefore, the data are adjusted to a common datum and a correction for tides is calculated. By adjusting the data to a common vertical datum, and gridding across the shoreline, a smooth transition from ocean bottom to land surface is achieved. The two data sets are then spliced at the NOS medium-resolution vector shoreline. The result is a 3-arc-second elevation grid in which elevations are resolved to 1/10 of a meter. Similar techniques can be used by researchers for different resolution grids and in different areas depending on the availability of data.