Numerical Simulation Of Tsunami Research Articles

EXPERIMENTAL AND NUMERICAL STUDIES ON THE TSUNAMI WAVE CHARACTERISTICS

ABSTRACT The devastating effects of the great Indian ocean tsunami has forced researchers in focusing their attention more vigorously on understanding the behaviour during its propagation and its effects on structures. This can obviously be accomplished through numerical and physical model studies or combination of both. The characteristics of a tsunami wave can approximately be same as that of a solitary wave which is basically a shallow water wave. Hence, the studies on the characteristics of shallow water waves have become an emerging topic of interest. An important aspect of a mitigation effort is to predict the tsunami wave kinematics. A combination of experimental and numerical simulation of tsunami represented by the solitary wave was studied and their comparison is discussed in this paper. The details of the numerical approach, methodology, instrumentation and measurement adopted for the present study are reported. The disagreement of the experimental simulation of solitary wave elevations with that of numerical simulation has been addressed and the possible discrepancies are overcome, the procedure of which are briefly discussed. The dynamics of the tsunami wave propagation over an uneven topography is studied using the developed numerical model.

Characteristics of Propagation of Tsunami Caused by Seabed Deformation in Uniform Depth

The purpose of this paper was to make clear relationships between tsunami generating and its propagating and a seabed deformation. In particular, we have studied how the deformation velocity and the deformation width affect the process of tsunami propagation. For the purpose of our work, numerical simulations of tsunami generated by the seabed deformation were carried out in three kinds of seabed deformations. As a result, when the width of seabed deformation became to more than 100 km, tsunami heights decayed to about 40% than its tsunami maximum heights. Further, it was found that the decay rate of tsunami height of negative waves is lower than that of positive waves. Moreover, the negative wave propagated with the wave height almost equal to the maximum tsunami height.

Field Survey on 2006 Java Tsunami

This report is to summarize the result of our field survey for 2006 Java Tsunami. Major conclusions are as follows; 1) The area from 108.0 E to the eastward was affected by the tsunami with the wave height of 3-5 m. 2) Tsunami run-up and inundation around tombolo or sand dune are investigated. Some topographical or artificial factors in tsunami disasters are found out: tombolo, the places where river mouths cut across sand dunes, roads or paths leading to coast lines, and drifting fishery boats. 3) There is the area in oval shape with the water depth of over 3000 m off the south cost of Java Island. Tsunami numerical simulation proves that the tsunami refracts toward north direction when it propagates out of this area. Such bathymetry and refraction can cause the tsunami to reach to the distant area like Yogyakarta.

Numerical Simulation of Tsunamis on the Tamil Nadu Coast of India

The State of Tamil Nadu was the most affected region in India during the tsunami of December 26, 2004, in the Indian Ocean, in terms of loss of life and damage. Numerical simulation was made for three tsunamis, the December 26, 2004, event, the Sumatra tsunami of 1833, and a hypothetical tsunami originating in the Andaman-Nicobar region. Since inundation is not included in these simulations, the tsunami amplitudes were deduced at the 10m depth contour in the ocean, off several locations on the coast of Tamil Nadu. The computed amplitudes appear reasonable as compared to known tsunami amplitudes from past events.

Numerical simulation of tsunamis generated by caldera collapse during the 7.3 ka Kikai eruption, Kyushu, Japan

Abstract The relationship between tsunamis and scales of caldera collapse during a 7.3 ka eruption of the Kikai volcano were numerically investigated, and a hypothetical caldera collapse scale was established. Wave height, arrival time, and run-up height and distance were determined at some locations along the coastline around Kikai caldera, using non-linear long-wave equations and caldera collapse models using parameters showing the difference in geometry between pre- and post-collapse and the collapse duration. Whether tsunamis become large and inundations occur in coasts is estimated by the dimensionless collapse speed. Computed tsunamis were then compared with geological characteristics found in coasts. The lack of evidence of tsunami inundation at Nejime, 65 km from the caldera, suggests that any tsunamis were small; indicating that the upper limit of dimensionless caldera collapse speed was 0.01. On the other hand, on the coast of the Satsuma Peninsula, 50 km from the caldera, geological characteristics suggests that tsunamis did not inundate, or that even if tsunamis inundated the area, the traces of a tsunami have been eroded by a climactic pyroclastic flow or the tsunami itself and they have not been left. In numerical computations, when a dimensionless caldera collapse speed is more than 0.003, tsunami can inundate this area.

Comparison of Two Different Earthquake Sources for the 26 December 2004 Aceh Tsunami Simulation

The tsunami wave propagation of the 26 December 2004 Aceh tsunami has been studied by performing a numerical tsunami simulation based on a method that was originally developed by the Tohoku University. The initial model was calculated based on the fault parameters proposed by GFZ Potsdam and Tohoku University. Despite the limitations in the numerical simulation, generally the calculated tsunami heights and arrival times show a relatively good agreement with the observed ones. Based on the simulation it can be estimated that the tsunami wave may reach the west and north coast of northern Sumatra at about 15 to 25 and 20 to 30 minutes after the earthquake, respectively. The maximum calculated tsunami heights in the west and north coast of Aceh were about 11 to 24 and 8 to 15 m, respectively. The calculated arrival times show that the tsunami wave may reach west coast of Thailand and east coast of India and Sri Lanka at about 90 to 100 and 100 to 120 minutes after the earthquake, respectively. The maximum calculated tsunami heights at Phuket of Thailand and Tricomalee of Sri Lanka were about 4 to 5 and 1.5 to 3.5 m, respectively.

Tsunami run-up heights of the 2003 Tokachi-oki earthquake

Tsunami height survey was conducted immediately after the 2003 Tokachi-oki earthquake. Results of the survey show that the largest tsunami height was 4 m to the east of Cape Erimo, around Bansei-onsen, and locally at Mabiro. The results also show that the tsunami height distribution of the 2003 Tokachi-oki earthquake is clearly different from that of the 1952 Tokachi-oki earthquake, suggesting the different source areas of the 1952 and 2003 Tokachioki earthquakes. Numerical simulation of tsunami is carried out using the slip distribution estimated by Yamanaka and Kikuchi (2003). The overall pattern of the observed tsunami height distribution along the coast is explained by the computed ones although the observed tsunami heights are slightly smaller. Large later phase observed at the tide gauge in Urakawa is the edge wave propagating from Cape Erimo along the west coast of the Hidaka area.

Development of the 2D/3D Hybrid Model for Tsunami Numerical Simulation

The 2D/3D hybrid tsunami numerical model was developed in which the conventional horizontally 2D model was adopted for the calculation in the wide region located far from the coastal structures while the 3D numerical model was used in the limited region adjacent to the structures. Applicability of the domain connection technique was examined by comparing the numerical results obtained by the present hybrid model with those obtained by applying the 3D model for the whole domain. Further, the results of the laboratory experiments were compared with the numerical results obtained by the hybrid model as well as the 2D model. It is shown that the present hybrid model reduces a calculation load significantly compared to the 3D model and is capable of reproducing the characteristics of three-dimensional and complicated flows around the structure, which cannot be reproduced by the 2D model alone.

Study of tsunami propagation in the Ligurian Sea

Abstract. Tsunami propagation is analyzed for the Ligurian Sea with particular attention on the French coasts of the Mediterranean. Historical data of tsunami manifestation on the French coast are analyzed for the period 2000 B.C.–1991 A.D. Numerical simulations of potential and historical tsunamis in the Ligurian Sea are done in the context of the nonlinear shallow water theory. Tsunami wave heights as well as their distribution function is calculated for historical tsunamis and it is shown that the log-normal distribution describes reasonably the simulated data. This demonstrates the particular role of bottom irregularities for the wave height distribution function near the coastlines. Also, spectral analysis of numerical tide-gauge records is done for potential tsunamis, revealing the complex resonant interactions between the tsunami waves and the bottom oscillations. It is shown that for an earthquake magnitude of 6.8 (averaged value for the Mediterranean Sea) the tsunami phenomenon has a very local character but with long duration. For sources located near the steep continental slope in the vicinity of the French-Italian Rivera, the tsunami tide-gauge records in the vicinity of Cannes – Imperia present irregular oscillations with a characteristic period of 20–30 min and a total duration of 10–20 h. For the western French coasts the amplitudes are significantly less with characteristic low-frequency oscillations (period of 40 min–1 h).

有珠火山の山体崩壊による洞爺湖の津波シミュレーション

有珠火山の山体崩壊による洞爺湖の津波シミュレーション

Analysis of Factors Influencing Simulations of the 1993 Hokkaido Nansei-Oki and 1964 Alaska Tsunamis

The purpose of this paper is to examine factorsinfluencing numerical simulations of tsunamis, andtheir implications for hazard mitigation. We focus ona specific finite element hydrodynamic model, chosenfor its role in the systematic development ofinundation maps for regions threatened primarily byCascadia Subduction Zone (CSZ) tsunamis. However, inpart for generality and in part because of poorhistorical records for CSZ events, we discuss here theperformance of the model in the context of betterdocumented past events with epicenters locatedelsewhere: the July 12, 1993 Hokkaido Nansei-Oki andthe March 28, 1964 Alaska tsunamis. Our analysisincludes the influence of grid refinement,interactions between tides and tsunamis, artificialenergy loss, and numerical parameterization. We showthat while the ability exists to reproduce pastevents, limitations remain in the modeling processthat should be accounted for in translating modelingresults into information for tsunami mitigation andresponse.

Generation of Tsunamis in the Okhotsk Sea Caused by the 1994 Great Kuril Earthquake

—The 1994 great Kuril earthquake generated an unusual tsunami that was observed at five tide gauges on the Hokkaido coast of the Okhotsk Sea. The tsunami arrived at tide gauges considerably earlier than the expected time, calculated on the assumption that the tsunami source area coincides with the aftershock area. Numerical simulation of the tsunami shows that the first wave of the tsunami in the Okhotsk Sea was generated by the significant subsidence north of the Kuril Islands. It is assumed that this subsidence is due to the earthquake. The coseismic deformation area of the ocean bottom extended over a vastly larger area than the aftershock area or the rupture area for the Kuril earthquake. The numerical simulation also shows that the tsunami observed at Utoro during the first hour after the origin time of the earthquake was mainly generated by the horizontal movement of the sloping ocean bottom near the Shiretoko Peninsula.

Numerical simulation of far-field tsunamis using the linear Boussinesq equation - The 1998 Papua New Guinea Tsunami.

The dispersion effect is not negligible in the numerical simulation of far-field tsunamis propagating through deep oceans. Imamura et al. (1990) introduced a technique in which the discretization error in the finite difference equation of the linear long wave equation was used to approximate the physical dispersion term. The technique is widely accepted to compute trans-Pacific tsunamis caused by great earthquakes (Mw> 8). However, the technique has never been applied to compute tsunamis caused by smaller earthquakes (Mw< 7) because the approximation may break down. In order to compute the tsunami caused by the 1998 Papua New Guinea earthquake (Mw 7.1), we numerically solve the linear Boussinesq equation, which includes the physical dispersion term, using an implicit scheme. For comparison, we also compute the tsunami using Imamura's technique. The comparison of the computed waveforms at the ocean bottom pressure gauge off Boso (BS3-OBP) from the two numerical simulations indicates that the linear Boussinesq equation should be used to simulate the tsunami waveform more accurately, especially the later phase of tsunami waveforms. We also found that the observed tsunami that was originally generated by the 1998 Papua New Guinea earthquake and recorded at BS3-OBP was a ridge wave. The ridge wave was enhanced by the shallow water region around the Izu-Bonin Islands.

Finite-element numerical simulation of tsunamis generated by earthquakes near a circular island

Abstract A numerical model for the propagation of tsunamis generated by earthquakes is applied to study the impact of the water waves against the coast of a circular island. The hydraulic source is modeled by means of the static deformation of the seafloor produced by an offshore seismic fault. The initial sea-surface disturbance is assumed to equal the seabed displacement that is computed analytically through the classical dislocation theory of plane faults with uniform slip. The wave propagation is described by means of the shallow-water approximation of the Navier-Stokes equations and is computed in a radially symmetric domain for which a quasi-analytical solution is available. The governing equations are solved numerically by means of a finite-element (FE) method making use of the Galerkin procedure, the mesh consisting of triangular elements of variable size. After a preliminary test of the numerical model against the analytical results, several simulations are performed by varying the fault location, dimension, and dip as well as by varying the radial profiles of the basin bathymetry. Most attention is devoted to studying the amplification of tsunami waves impacting the island coast. What is seen is that, depending on the source characteristics (namely, fault length, dip, strike, and distance from the island) and on the basin bathymetry, high amplifications (in the range 1 to 5) are normally obtained not only in the front of the island as is expected but also on the lee side of the island, because of positive interference of waves traveling around the island in both directions. Under particular conditions (downwardly concave bathymetry), the largest amplifications are found neither on the front side nor on the lee side, but at intermediate places along the island coast.

Analysis and Forcast by Computer Simulation. Tsunami Numerical Simulation and Its Application.

Analysis and Forcast by Computer Simulation. Tsunami Numerical Simulation and Its Application.

A smoothing algorithm to enhance finite-element tsunami modelling: An application to the 5 February 1783 Calabrian case, Italy

Numerical simulations are essential tools for studying tsunami generation and evolution and finite-element (FE) methods are widely used, especially because of their capability in modeling water waves in basins with complex bathymetry and irregular coastlines. This paper presents the numerical simulation of an historical Italian tsunami that affected the Tyrrhenian coasts of Calabria and Sicily on 5 February 1783 following a strong destructive earthquake that was the first of a terrible sequence of seismic shocks terrifying the Calabrian population for more than two months. The numerical model is an FE model based on the nonlinear nondispersive shallow-water approximation of the Navier-Stokes equations. Since FE discretization schemes may lead to solutions undesirably affected by noise over coarse grids, in this study numerical noise is controlled by suitably smoothing the FE solution at regular time steps Δts. The performance of our smoothing algorithm is tested for significant linear cases for which an analytical solution is available.

Fault parameters and tsunami excitation of the May 13, 1993, Shumagin Islands earthquake

The Shumagin Islands earthquake of May 13, 1993, occurred in a previously identified seismic gap where a large subduction earthquake is expected. We analyzed long‐period surface waves and P waves recorded on the IRIS stations to estimate the fault parameters. The Centroid Moment Tensor solution shows that the focal mechanism is a thrust type with the strike parallel to the Aleutian trench. The seismic moment is 2.0×1019 Nm and the corresponding moment magnitude is 6.8. The Moment Tensor Rate Function inversion from P waves also yields a similar focal mechanism and seismic moment. In addition, this computation provides estimates of 10 s for the duration of the source time function and 35 km for the best point source depth. These seismological analyses indicate that the fault mechanism of the 1993 earthquake was as expected, but that the magnitude was too small to fill the gap. This earthquake did not generate a tsunami large enough to be observed at the Sand Point, Alaska tide gauge or at an ocean bottom pressure gauge, at distances of 100 and 300 km, respectively. Numerical tsunami simulations result in amplitudes at both stations that are within the background noise level. Additional numerical experiments also suggest that the small tsunami amplitudes are due to the location of the source area in the shallow shelf region.

Numerical simulation of tsunamis ? Its present and near future

Hindcasting of a tsunami by numerical simulations is a process of lengthy and complicated deductions, knowing only the final results such as run-up heights and tide records, both of which are possibly biased due to an insufficient number of records and due to hydraulic and mechanical limitation of tide gauges. There are many sources of error. The initial profile, determined with seismic data, can even be different from the actual tsunami profile. The numerical scheme introduces errors. Nonlinearity near and on land requires an appropriate selection of equations. Taking these facts into account, it should be noted that numerical simulations produce satisfactory information for practical use, because the final error is usually within 15% as far as the maximum run-up height is concerned.

Truncation Error in Numerical Tsunami Simulation by the Finite Difference Method

ABSTRACTIt is well known that the truncation error in long wave simulations by the numerical scheme appears as the artificial dispersion or dissipation effects and causes the phase shift or the num...

USE OF NUMERICAL SIMULATION AS A MEANS OF TSUNAMI WARNING

Possibility of quantitative forecasting of local tsunamis sufficiently before their arrivals, by use of a super computer, is examined. Provided that the initial tsunami profile can be determined within a few minutes after an earthquake by such a method as proposed by lzutani and Hirasawa, the tsunami numerical simulation based upon the linear long wave theory can be finished within two minutes and give practically accurate tsunami heights at every 200m along the castline of the Sanriku district extending from 41°30′N. to 36°20′N.