Nava, G.; Lotto, G. C.; Dunham, E. M.
Tsunami propagation in the open ocean is most commonly modeled by solving the shallow water wave equations. These equations require two initial conditions: one on sea surface height and another on depth-averaged horizontal particle velocity or, equivalently, horizontal momentum. While most modelers assume that initial velocity is zero, Y.T. Song and collaborators have argued for nonzero initial velocity, claiming that horizontal displacement of a sloping seafloor imparts significant horizontal momentum to the ocean. They show examples in which this effect increases the resulting tsunami height by a factor of two or more relative to models in which initial velocity is zero. We test this claim with a “full-physics” integrated dynamic rupture and tsunami model that couples the elastic response of the Earth to the linearized acoustic-gravitational response of a compressible ocean with gravity; the model self-consistently accounts for seismic waves in the solid Earth, acoustic waves in the ocean, and tsunamis (with dispersion at short wavelengths). We run several full-physics simulations of subduction zone megathrust ruptures and tsunamis in geometries with a sloping seafloor, using both idealized structures and a more realistic Tohoku structure. Substantial horizontal momentum is imparted to the ocean, but almost all momentum is carried away in the form of ocean acoustic waves. We compare tsunami propagation in each full-physics simulation to that predicted by an equivalent shallow water wave simulation with varying assumptions regarding initial conditions. We find that the initial horizontal velocity conditions proposed by Song and collaborators consistently overestimate the tsunami amplitude and predict an inconsistent wave profile. Finally, we determine tsunami initial conditions that are rigorously consistent with our full-physics simulations by isolating the tsunami waves (from ocean acoustic and seismic waves) at some final time, and backpropagating the tsunami
Ryan, K. J.; Geist, E. L.; Oglesby, D. D.; Kyriakopoulos, C.
Motivated by the 2011 Mw 9 Tohoku-Oki event, we explore the effects of realistic fault dynamics on slip, free surface deformation, and the resulting tsunami generation and local propagation from a hypothetical Mw 9 megathrust earthquake along the Alaskan-Aleutian (A-A) Megathrust. We demonstrate three scenarios: a spatially-homogenous prestress and frictional parameter model and two models with rate-strengthening-like friction (e.g., Dieterich, 1992). We use a dynamic finite element code to model 3-D ruptures, using time-weakening friction (Andrews, 2004) as a proxy for rate-strengthening friction, along a portion of the A-A subduction zone. Given geometric, material, and plate-coupling data along the A-A megathrust assembled from the Science Application for Risk Reduction (SAFRR) team (e.g., Bruns et al., 1987; Hayes et al., 2012; Johnson et al., 2004; Santini et al., 2003; Wells at al., 2003), we are able to dynamically model rupture. Adding frictional-strengthening to a region of the fault reduces both average slip and free surface displacement above the strengthening zone, with the magnitude of the reductions depending on the strengthening zone location. Corresponding tsunami models, which use a finite difference method to solve the long-wave equations (e.g., Liu et al., 1995; Satake, 2002; Shuto, 1991), match sea floor displacement, in time, to the free surface displacement from the rupture models. Tsunami models show changes in local peak amplitudes and beaming patterns for each slip distribution. Given these results, other heterogeneous parameterizations, with respect to prestress and friction, still need to be examined. Additionally, a more realistic fault geometry will likely affect the rupture dynamics. Thus, future work will incorporate stochastic stress and friction distributions as well as a more complex fault geometry based on Slab 1.0 (Hayes et al., 2012).
Moore, G F; Bangs, N L; Taira, A; Kuramoto, S; Pangborn, E; Tobin, H J
Megasplay faults, very long thrust faults that rise from the subduction plate boundary megathrust and intersect the sea floor at the landward edge of the accretionary prism, are thought to play a role in tsunami genesis. We imaged a megasplay thrust system along the Nankai Trough in three dimensions, which allowed us to map the splay fault geometry and its lateral continuity. The megasplay is continuous from the main plate interface fault upwards to the sea floor, where it cuts older thrust slices of the frontal accretionary prism. The thrust geometry and evidence of large-scale slumping of surficial sediments show that the fault is active and that the activity has evolved toward the landward direction with time, contrary to the usual seaward progression of accretionary thrusts. The megasplay fault has progressively steepened, substantially increasing the potential for vertical uplift of the sea floor with slip. We conclude that slip on the megasplay fault most likely contributed to generating devastating historic tsunamis, such as the 1944 moment magnitude 8.1 Tonankai event, and it is this geometry that makes this margin and others like it particularly prone to tsunami genesis.
Hyvernaud, O.; Reymond, D.; Okal, E.; Hebert, H.; Clément, J.; Wong, K.
We present the field survey and observations of the Japan tsunami of March 2011, in Society and Marquesas islands. Without being catastrophic the tsunami produced some damages in the Marquesas, which are always the most prone to tsunami amplification in French Polynesia: 8 houses were destroyed and inundated (up to 4.5 m of run-up measured). Surprisingly, the maximum run-up was observed on the South-West coast of Nuku Hiva island (a bay open to the opposite direction of the wave-front). In Tahiti, the tsunami was much more moderate, with a maximum height observed on the North coast: about 3 m of run-up observed, corresponding to the highest level of the seasonal oceanic swell without damage (just the main road inundated). These observations are well explained and reproduced by the numerical modeling of the tsunami. The results obtained confirm the exceptional source dimensions. Concerning the real time aspect, the tsunami height has been also rapidly predicted during the context of tsunami warning, with 2 methods: the first uses a database of pre-computed numeric simulations, and the second one uses a formula giving the tsunami amplitude in deep ocean in function of the source parameters (coordinates of the source, scalar moment and fault azimuth) and of the coordinates of the receiver. The population responded responsibly to the evacuation order on the 19 islands involved, helped in part by a favourable arrival time of the wave (7:30 a.m., local time).
Moore, C. W.; Eble, M. C.; Rabinovich, A.; Titov, V. V.
The Mw = 9.3 megathrust earthquake of December 26, 2004 off the coast of Sumatra generated a catastrophic tsunami that crossed the Indian Ocean and was widespread in the Pacific and Atlantic oceans being recorded by a great number of coastal tide gauges located in 15-25 thousand kilometers from the source area. The data from these instruments throughout the world oceans enabled estimates of various statistical parameters and energy decay of this event. However, only very few open-ocean records of this tsunami had been obtained. A unique high-resolution record of this tsunami from DART 32401 located offshore of northern Chile, combined with the South American mainland tide gauge measurements and the data from three island stations (San Felix, Juan Fernandez and Easter) enabled us to examine far-field characteristics of the event in the southeastern Pacific and to compare the results of global numerical simulations with observations. The maximum wave height measured at DART 32401 was only 1.8 cm but the signal was very clear and reliable. Despite their small heights, the waves demonstrated consistent spatial and temporal structure and good agreement with DART 46405/NeMO records in the NE Pacific. The travel time from the source area to DART 32401 was 25h 55min in good agreement with the computed travel time (25h 45min) and consistent with the times obtained from the nearby coastal tide gauges. This agreement was much better than it followed from the direct travel time estimation based classical kinematic theory that gave the travel time approximately 1.5 hrs shorter than observed. The later actual arrival of the 2004 tsunami waves corresponds to the most energetically economic path along the mid-ocean ridge wave-guides, which is distinctly reproduced by the numerical model. Also, the numerical model described well the frequency content, amplitudes and general structure of the observed waves at this DART and the three island stations. Maximum wave heights in this
Murphy, Shane; Spagnuolo, Elena; Lorito, Stefano; Di Toro, Giulio; Scala, Antonio; Festa, Gaetano; Nielsen, Stefan; Piatanesi, Alessio; Romano, Fabrizio; Aretusini, Stefano
Seismological, tsunami and geodetic observations have shown that subduction zones are complex systems where the properties of earthquake rupture vary with depth. For example nucleation and high frequency radiation generally occur at depth but low frequency radiation and large tsunami-genic slip appear to occur in the shallow crustal depth. Numerical simulations used to describe these features predominantly use standardised theoretical equations or experimental observations often assuming that their validity extends to all slip-rates, lithologies and tectonic environments. However recent rotary-shear experiments performed on a range of diverse materials and experimental conditions highlighted the large variability of the evolution of friction during slipping pointing to a more complex relationship between material type, slip rate and normal stress. Simulating dynamic rupture using a 2D spectral element methodology on a Tohoku like fault, we apply experimentally derived friction laws (i.e. thermal slip distance friction law, Di Toro et al. 2011) Choice of parameters for the friction law are based on expected material type (e.g. cohesive and non-cohesive clay rich material representative of an accretionary wedge), the normal stress which is controlled by the interaction between the regional stress field and the fault geometry. The shear stress distribution on the fault plane is fractal with the yield stress dependent on the static coefficient of friction and the normal stress, parameters that are dependent on the material type and geometry. We use metrics such as the slip distribution, ground motion and fracture energy to explore the effect of frictional behaviour, fault geometry and stress perturbations and its potential role in tsunami generation. Preliminary results will be presented. This research is funded by the European Union’s Seventh Framework Programme (FP7/2007-2013) under grant agreement n° 603839 (Project ASTARTE – Assessment, Strategy and Risk Reduction
Titov, Vasily V.; Bernard, Eddie N.; Weinstein, Stuart A.; Kanoglu, Utku; Synolakis, Costas E.
Almost five years after the 26 December 2004 Indian Ocean tragedy, tsunami warnings are finally benefiting from decades of research toward effective model-based forecasts. Since the 2004 tsunami, two seminal advances have been (i) deep-ocean tsunami measurements with tsunameters and (ii) their use in accurately forecasting tsunamis after the tsunami has been generated. Using direct measurements of deep-ocean tsunami heights, assimilated into numerical models for specific locations, greatly improves the real-time forecast accuracy over earthquake-derived magnitude estimates of tsunami impact. Since 2003, this method has been used to forecast tsunamis at specific harbors for different events in the Pacific and Indian Oceans. Recent tsunamis illustrated how this technology is being adopted in global tsunami warning operations. The U.S. forecasting system was used by both research and operations to evaluate the tsunami hazard. Tests demonstrated the effectiveness of operational tsunami forecasting using real-time deep-ocean data assimilated into forecast models. Several examples also showed potential of distributed forecast tools. With IOC and USAID funding, NOAA researchers at PMEL developed the Community Model Interface for Tsunami (ComMIT) tool and distributed it through extensive capacity-building sessions in the Indian Ocean. Over hundred scientists have been trained in tsunami inundation mapping, leading to the first generation of inundation models for many Indian Ocean shorelines. These same inundation models can also be used for real-time tsunami forecasts as was demonstrated during several events. Contact Information Vasily V. Titov, Seattle, Washington, USA, 98115
González-Carrasco, J. F.; Gonzalez, G.; Aránguiz, R.; Catalan, P. A.; Cienfuegos, R.; Urrutia, A.; Shrivastava, M. N.; Yagi, Y.; Moreno, M.
Tsunami inundation maps are a powerful tool to design evacuation plans of coastal communities, additionally can be used as a guide to territorial planning and assessment of structural damages in port facilities and critical infrastructure (Borrero et al., 2003; Barberopoulou et al., 2011; Power et al., 2012; Mueller et al., 2015). The accuracy of inundation estimation is highly correlated with tsunami initial conditions, e.g. seafloor vertical deformation, displaced water volume and potential energy (Bolshakova et al., 2011). Usually, the initial conditions are estimated using homogeneous rupture models based in historical worst-case scenario. However tsunamigenic events occurred in central Chilean continental margin showed a heterogeneous slip distribution of source with patches of high slip, correlated with fully-coupled interseismic zones (Moreno et al., 2012). The main objective of this work is to evaluate the predictive capacity of interseismic coupling models based on geodetic data comparing them with homogeneous fault slip model constructed using scaling laws (Blaser et al., 2010) to estimate inundation and runup in coastal areas. To test our hypothesis we select a seismic gap of Maule, where occurred the last large tsunamigenic earthquake in the chilean subduction zone, using the interseismic coupling models (ISC) proposed by Moreno et al., 2011 and Métois et al., 2013. We generate a slip deficit distribution to build a tsunami source supported by geological information such as slab depth (Hayes et al., 2012), strike, rake and dip (Dziewonski et al., 1981; Ekström et al., 2012) to model tsunami generation, propagation and shoreline impact using Neowave 2D (Yamazaki et al., 2009). We compare the tsunami scenario of Mw 8.8, Maule based in coseismic slip distribution proposed by Moreno et al., 2012 with homogeneous and heterogeneous models to identify the accuracy of our results with sea level time series and regional runup data (Figure 1). The estimation of
Seto, S.; Takahashi, T.
In the 2011 Tohoku earthquake tsunami disaster, the delay of understanding damage situation increased the human damage. To solve this problem, it is important to search the severe damaged areas. The tsunami numerical modeling is useful to estimate damages and the accuracy of simulation depends on the tsunami source. Seto and Takahashi (2017) proposed a method to estimate the characterized tsunami source model by using the limited observed data of GPS buoys. The model consists of Large slip zone (LSZ), Super large slip zone (SLSZ) and background rupture zone (BZ) as the Cabinet Office, Government of Japan (below COGJ) reported after the Tohoku tsunami. At the beginning of this method, the rectangular fault model is assumed based on the seismic magnitude and hypocenter reported right after an earthquake. By using the fault model, tsunami propagation is simulated numerically, and the fault model is improved after comparing the computed data with the observed data repeatedly. In the comparison, correlation coefficient and regression coefficient are used as indexes. They are calculated with the observed and the computed tsunami wave profiles. This repetition is conducted to get the two coefficients close to 1.0, which makes the precise of the fault model higher. However, it was indicated as the improvement that the model did not examine a complicated shape of tsunami source. In this study, we proposed an improved model to examine the complicated shape. COGJ(2012) assumed that possible tsunami source region in the Nankai trough consisted of the several thousands small faults. And, we use these small faults to estimate the targeted tsunami source in this model. Therefore, we can estimate the complicated tsunami source by using these small faults. The estimation of BZ is carried out as a first step, and LSZ and SLSZ are estimated next as same as the previous model. The proposed model by using GPS buoy was applied for a tsunami scenario in the Nankai Trough. As a result
Alvarez-Gomez, J. A.; Aniel-Quiroga, I.; Gonzalez, M.; Olabarrieta, Maitane; Carreno, E.
The tsunami impact on the Spanish and North African coasts of the Alboran Sea generated by several reliable seismic tsunamigenic sources in this area was modeled. The tectonic setting is complex and a study of the potential sources from geological data is basic to obtain probable source characteristics. The tectonic structures considered in this study as potentially tsunamigenic are: the Alboran Ridge associated structures, the Carboneras Fault Zone and the Yusuf Fault Zone. We characterized 12 probable tsunamigenic seismic sources in the Alboran Basin based on the results of recent oceanographical studies. The strain rate in the area is low and therefore its seismicity is moderate and cannot be used to infer characteristics of the major seismic sources. These sources have been used as input for the numerical simulation of the wave propagation, based on the solution of the nonlinear shallow water equations through a finite-difference technique. We calculated the Maximum Wave Elevations, and Tsunami Travel Times using the numerical simulations. The results are shown as maps and profiles along the Spanish and African coasts. The sources associated with the Alboran Ridge show the maximum potential to generate damaging tsunamis, with maximum wave elevations in front of the coast exceeding 1.5 m. The Carboneras and Yusuf faults are not capable of generating disastrous tsunamis on their own, although their proximity to the coast could trigger landslides and associated sea disturbances. The areas which are more exposed to the impact of tsunamis generated in the Alboran Sea are the Spanish coast between Malaga and Adra, and the African coast between Alhoceima and Melilla.
Grilli, S. T.; Insua, T. L.; Grilli, A. R.; Douglas, K. L.; Shelby, M. R.; Wang, K.; Gao, D.
Located in the heart of Vancouver Island, BC, Port Alberni has a well-known history of tsunamis. Many of the Nuu-Chah-Nulth First Nations share oral stories about a strong fight between a thunderbird and a whale that caused big waves in a winter night, a story that is compatible with the recently recognized great Cascadia tsunami in January, 1700. Port Alberni, with a total population of approximately 20,000 people, lies beside the Somass River, at the very end of Barkley Sound Inlet. The narrow canal connecting this town to the Pacific Ocean runs for more than 64 km ( 40 miles) between steep mountains, providing an ideal setting for the amplification of tsunami waves through funnelling effects. The devastating effects of tsunamis are still fresh in residents’ memories from the impact of the 1964 Alaska tsunami that caused serious damage to the city. In June 2016, Emergency Management BC ran a coastal exercise in Port Alberni, simulating the response to an earthquake and a tsunami. During three days, the emergency teams in the City of Port Alberni practiced and learned from the experience. Ocean Networks Canada contributed to this exercise with the development of preliminary simulations of tsunami impact on the city from a buried rupture of the Cascadia Subduction Zone, including the Explorer segment. Wave propagation was simulated with the long-wave model FUNWAVE-TVD. Preliminary results indicate a strong amplification of tsunami waves in the Port Alberni area. The inundation zone in Port Alberni had a footprint similar to that of the 1700 Cascadia and 1964 Alaska tsunamis, inundating the area surrounding the Somass river and preferentially following the Kitsuksis and Roger Creek river margins into the city. Several other tsunami source scenarios, including splay faulting and trench-breaching ruptures are currently being modeled for the city of Port Alberni following a similar approach. These results will be presented at the conference.
Baglione, Enrico; Armigliato, Alberto; Pagnoni, Gianluca; Tinti, Stefano
The fact that ruptures on the generating faults of large earthquakes are strongly heterogeneous has been demonstrated over the last few decades by a large number of studies. The effort to retrieve reliable finite-fault models (FFMs) for large earthquakes occurred worldwide, mainly by means of the inversion of different kinds of geophysical data, has been accompanied in the last years by the systematic collection and format homogenisation of the published/proposed FFMs for different earthquakes into specifically conceived databases, such as SRCMOD. The main aim of this study is to explore characteristic patterns of the slip distribution of large earthquakes, by using a subset of the FFMs contained in SRCMOD, covering events with moment magnitude equal or larger than 6 and occurred worldwide over the last 25 years. We focus on those FFMs that exhibit a single and clear region of high slip (i.e. a single asperity), which is found to represent the majority of the events. For these FFMs, it sounds reasonable to best-fit the slip model by means of a 2D Gaussian distributions. Two different methods are used (least-square and highest-similarity) and correspondingly two “best-fit” indexes are introduced. As a result, two distinct 2D Gaussian distributions for each FFM are obtained. To quantify how well these distributions are able to mimic the original slip heterogeneity, we calculate and compare the vertical displacements at the Earth surface in the near field induced by the original FFM slip, by an equivalent uniform-slip model, by a depth-dependent slip model, and by the two “best” Gaussian slip models. The coseismic vertical surface displacement is used as the metric for comparison. Results show that, on average, the best results are the ones obtained with 2D Gaussian distributions based on similarity index fitting. Finally, we restrict our attention to those single-asperity FFMs associated to earthquakes which generated tsunamis. We choose few events for which tsunami
Miller, Garrett S.; Andy Take, W.; Mulligan, Ryan P.; McDougall, Scott
In this experimental study, granular material is released down slope to investigate landslide-generated waves. Starting with a known volume and initial position of the landslide source, detailed data are obtained on the velocity and thickness of the granular flow, the shape and location of the submarine landslide deposit, the amplitude and shape of the near-field wave, the far-field wave evolution, and the wave runup elevation on a smooth impermeable slope. The experiments are performed on a 6.7 m long 30° slope on which gravity accelerates the landslides into a 2.1 m wide and 33.0 m long wave flume that terminates with a 27° runup ramp. For a fixed landslide volume of 0.34 m3, tests are conducted in a range of still water depths from 0.05 to 0.50 m. Observations from high-speed cameras and measurements from wave probes indicate that the granular landslide moves as a long and thin train of material, and that only a portion of the landslide (termed the “effective mass”) is engaged in activating the leading wave. The wave behavior is highly dependent on the water depth relative to the size of the landslide. In deeper water, the near-field wave behaves as a stable solitary-like wave, while in shallower water, the wave behaves as a breaking dissipative bore. Overall, the physical model observations are in good agreement with the results of existing empirical equations when the effective mass is used to predict the maximum near-field wave amplitude, the far-field amplitude, and the runup of tsunamis generated by granular landslides.
Fujii, Y.; Satake, K.
The tsunami generation process of the 2004 Sumatra-Andaman earthquake were estimated from the tsunami waveforms recorded on tide gauges and sea surface heights captured by satellite altimetry measurements over the Indian Ocean. The earthquake (0:58:53, 26, Dec., 2004, UTC), the largest in the last 40 years, caused devastating tsunami damages to the countries around the Indian Ocean. One of the important questions is the source length; the aftershocks were distributed along the Sunda trench for 1000 to 1200 km, from off northwestern part of Sumatra island through Nicobar islands to Andaman island, while seismic wave analyses indicate much shorter source length (several hundred km). We used instrumental data of this tsunami, tide gauges and sea surface heights. Tide gauge data have been collected by Global Sea Level Observing System (GLOSS). We have also used another tide gauges data for tsunami simulation analysis. Tsunami propagation was captured as sea surface heights of Jason-1 satellite altimetry measurements over the Indian Ocean for the first time (Gower, 2005). We numerically compute tsunami propagation on actually bathymetry. ETOPO2 (Smith and Sandwell, 1997), the gridded data of global ocean depth from bathymetry soundings and satellite gravity data, are less reliable in the shallow ocean. To improve the accuracy, we have digitized the charts near coasts and merged the digitized data with the ETOPO2 data. The long-wave equation and the equation of motion were numerically solved by finite-difference method (Satake, 1995). As the initial condition, a static deformation of seafloor has been calculated using rectangular fault model (Okada, 1985). The source region is divided into 22 subfaults. We fixed the size and geometry of each subfault, and varied the slip amount and rise time (or slip duration) for each subfault, and rupture velocity. Tsunami waveforms or Greens functions for each subfault were calculated for the rise times of 3, 10, 30 and 60 minutes
Wang, D.; Becker, N. C.; Weinstein, S.; Duputel, Z.; Rivera, L. A.; Hayes, G. P.; Hirshorn, B. F.; Bouchard, R. H.; Mungov, G.
The Pacific Tsunami Warning Center (PTWC) began forecasting tsunamis in real-time using source parameters derived from real-time Centroid Moment Tensor (CMT) solutions in 2009. Both the USGS and PTWC typically obtain W-Phase CMT solutions for large earthquakes less than 30 minutes after earthquake origin time. Within seconds, and often before waves reach the nearest deep ocean bottom pressure sensor (DARTs), PTWC then generates a regional tsunami propagation forecast using its linear shallow water model. The model is initialized by the sea surface deformation that mimics the seafloor deformation based on Okada’s (1985) dislocation model of a rectangular fault with a uniform slip. The fault length and width are empirical functions of the seismic moment. How well did this simple model perform? The DART records provide a very valuable dataset for model validation. We examine tsunami events of the last decade with earthquake magnitudes ranging from 6.5 to 9.0 including some deep events for which tsunamis were not expected. Most of the forecast results were obtained during the events. We also include events from before the implementation of the WCMT method at USGS and PTWC, 2006-2009. For these events, WCMTs were computed retrospectively (Duputel et al. 2012). We also re-ran the model with a larger domain for some events to include far-field DARTs that recorded a tsunami with identical source parameters used during the events. We conclude that our model results, in terms of maximum wave amplitude, are mostly within a factor of two of the observed at DART stations, with an average error of less than 40% for most events, including the 2010 Maule and the 2011 Tohoku tsunamis. However, the simple fault model with a uniform slip is too simplistic for the Tohoku tsunami. We note model results are sensitive to centroid location and depth, especially if the earthquake is close to land or inland. For the 2016 M7.8 New Zealand earthquake the initial forecast underestimated the
Volcanic tsunamis are generated by a variety of mechanisms, including volcano-tectonic earthquakes, slope instabilities, pyroclastic flows, underwater explosions, shock waves and caldera collapse. In this review, we focus on the lessons that can be learnt from past events and address the influence of parameters such as volume flux of mass flows, explosion energy or duration of caldera collapse on tsunami generation. The diversity of waves in terms of amplitude, period, form, dispersion, etc. poses difficulties for integration and harmonization of sources to be used for numerical models and probabilistic tsunami hazard maps. In many cases, monitoring and warning of volcanic tsunamis remain challenging (further technical and scientific developments being necessary) and must be coupled with policies of population preparedness. © 2015 The Author(s).
Harbitz, C. B.; Frauenfelder, R.; Kaiser, G.; Glimsdal, S.; Sverdrup-thygeson, K.; Løvholt, F.; Gruenburg, L.; Mc Adoo, B. G.
The 2011 Tōhoku tsunami caused a high number of fatalities and massive destruction. Data collected after the event allow for retrospective analyses. Since 2009, NGI has developed a generic GIS model for local analyses of tsunami vulnerability and mortality risk. The mortality risk convolves the hazard, exposure, and vulnerability. The hazard is represented by the maximum tsunami flow depth (with a corresponding likelihood), the exposure is described by the population density in time and space, while the vulnerability is expressed by the probability of being killed as a function of flow depth and building class. The analysis is further based on high-resolution DEMs. Normally a certain tsunami scenario with a corresponding return period is applied for vulnerability and mortality risk analysis. Hence, the model was first employed for a tsunami forecast scenario affecting Bridgetown, Barbados, and further developed in a forecast study for the city of Batangas in the Philippines. Subsequently, the model was tested by hindcasting the 2009 South Pacific tsunami in American Samoa. This hindcast was based on post-tsunami information. The GIS model was adapted for optimal use of the available data and successfully estimated the degree of mortality.For further validation and development, the model was recently applied in the RAPSODI project for hindcasting the 2011 Tōhoku tsunami in Sendai and Ishinomaki. With reasonable choices of building vulnerability, the estimated expected number of fatalities agree well with the reported death toll. The results of the mortality hindcast for the 2011 Tōhoku tsunami substantiate that the GIS model can help to identify high tsunami mortality risk areas, as well as identify the main risk drivers.The research leading to these results has received funding from CONCERT-Japan Joint Call on Efficient Energy Storage and Distribution/Resilience against Disasters (http://www.concertjapan.eu; project RAPSODI – Risk Assessment and design of
Li, Linlin; Switzer, Adam D.; Wang, Yu; Chan, Chung-Han; Qiu, Qiang; Weiss, Robert
Current tsunami inundation maps are commonly generated using deterministic scenarios, either for real-time forecasting or based on hypothetical “worst-case” events. Such maps are mainly used for emergency response and evacuation planning and do not include the information of return period. However, in practice, probabilistic tsunami inundation maps are required in a wide variety of applications, such as land-use planning, engineer design and for insurance purposes. In this study, we present a method to develop the probabilistic tsunami inundation map using a stochastic earthquake source model. To demonstrate the methodology, we take Macau a coastal city in the South China Sea as an example. Two major advances of this method are: it incorporates the most updated information of seismic tsunamigenic sources along the Manila megathrust; it integrates a stochastic source model into a Monte Carlo-type simulation in which a broad range of slip distribution patterns are generated for large numbers of synthetic earthquake events. When aggregated the large amount of inundation simulation results, we analyze the uncertainties associated with variability of earthquake rupture location and slip distribution. We also explore how tsunami hazard evolves in Macau in the context of sea level rise. Our results suggest Macau faces moderate tsunami risk due to its low-lying elevation, extensive land reclamation, high coastal population and major infrastructure density. Macau consists of four districts: Macau Peninsula, Taipa Island, Coloane island and Cotai strip. Of these Macau Peninsula is the most vulnerable to tsunami due to its low-elevation and exposure to direct waves and refracted waves from the offshore region and reflected waves from mainland. Earthquakes with magnitude larger than Mw8.0 in the northern Manila trench would likely cause hazardous inundation in Macau. Using a stochastic source model, we are able to derive a spread of potential tsunami impacts for earthquakes
Koh, H. L.; Teh, S. Y.; Tan, W. K.; Kh’ng, X. Y.
A standard set of benchmark problems, known as OAR-PMEL-135, is developed by the US National Tsunami Hazard Mitigation Program for tsunami inundation model validation. Any tsunami inundation model must be tested for its accuracy and capability using this standard set of benchmark problems before it can be gainfully used for inundation simulation. The authors have previously developed an in-house tsunami inundation model known as TUNA-RP. This inundation model solves the two-dimensional nonlinear shallow water equations coupled with a wet-dry moving boundary algorithm. This paper presents the validation of TUNA-RP against the solutions provided in the OAR-PMEL-135 benchmark problem set. This benchmark validation testing shows that TUNA-RP can indeed perform inundation simulation with accuracy consistent with that in the tested benchmark problem set.
Yong, Wei; Newman, Andrew V.; Hayes, Gavin P.; Titov, Vasily V.; Tang, Liujuan
Correctly characterizing tsunami source generation is the most critical component of modern tsunami forecasting. Although difficult to quantify directly, a tsunami source can be modeled via different methods using a variety of measurements from deep-ocean tsunameters, seismometers, GPS, and other advanced instruments, some of which in or near real time. Here we assess the performance of different source models for the destructive 11 March 2011 Japan tsunami using model–data comparison for the generation, propagation, and inundation in the near field of Japan. This comparative study of tsunami source models addresses the advantages and limitations of different real-time measurements with potential use in early tsunami warning in the near and far field. The study highlights the critical role of deep-ocean tsunami measurements and rapid validation of the approximate tsunami source for high-quality forecasting. We show that these tsunami measurements are compatible with other real-time geodetic data, and may provide more insightful understanding of tsunami generation from earthquakes, as well as from nonseismic processes such as submarine landslide failures.